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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ecc7a9a9-2c74-438c-9099-373a4c17328f | sequential-neural-networks-for-noetic-end-to | 2003.02126 | null | https://arxiv.org/abs/2003.02126v1 | https://arxiv.org/pdf/2003.02126v1.pdf | Sequential Neural Networks for Noetic End-to-End Response Selection | The noetic end-to-end response selection challenge as one track in the 7th Dialog System Technology Challenges (DSTC7) aims to push the state of the art of utterance classification for real world goal-oriented dialog systems, for which participants need to select the correct next utterances from a set of candidates for the multi-turn context. This paper presents our systems that are ranked top 1 on both datasets under this challenge, one focused and small (Advising) and the other more diverse and large (Ubuntu). Previous state-of-the-art models use hierarchy-based (utterance-level and token-level) neural networks to explicitly model the interactions among different turns' utterances for context modeling. In this paper, we investigate a sequential matching model based only on chain sequence for multi-turn response selection. Our results demonstrate that the potentials of sequential matching approaches have not yet been fully exploited in the past for multi-turn response selection. In addition to ranking top 1 in the challenge, the proposed model outperforms all previous models, including state-of-the-art hierarchy-based models, on two large-scale public multi-turn response selection benchmark datasets. | ['Wen Wang', 'Qian Chen'] | 2020-03-03 | null | null | null | null | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-3.22128274e-02 1.57379642e-01 -1.42250508e-01 -9.30031478e-01
-1.04805136e+00 -5.79479814e-01 7.58098483e-01 1.79328531e-01
-4.86635268e-01 7.55045056e-01 6.66040838e-01 -3.71932566e-01
2.27076504e-02 -2.60622501e-01 3.20568353e-01 -9.96288434e-02
1.77727640e-01 1.18165600e+00 5.89978814e-01 -1.29666603e+00
4.31083888e-01 1.61583144e-02 -1.39351583e+00 1.11026645e+00
5.16931117e-01 8.71233344e-01 1.88927069e-01 1.07092941e+00
-4.19847280e-01 1.12552190e+00 -6.01584673e-01 -3.50843102e-01
-1.31745920e-01 -6.13107562e-01 -1.61274970e+00 -3.63352329e-01
4.28045243e-01 -5.17058492e-01 -1.39939174e-01 3.86902034e-01
8.30277383e-01 5.43920159e-01 3.75133425e-01 -1.09937286e+00
-1.16767786e-01 1.02122438e+00 3.66915315e-01 -2.37127282e-02
1.03929448e+00 2.21985310e-01 1.39447248e+00 -9.26862895e-01
7.84560382e-01 1.80913544e+00 4.24583048e-01 1.08612108e+00
-1.13294244e+00 -1.97366297e-01 5.00658393e-01 3.85739356e-01
-9.07200933e-01 -7.57048547e-01 6.34089053e-01 -3.18316013e-01
1.55608368e+00 7.48536527e-01 5.67802228e-03 1.09443724e+00
1.87655110e-02 1.07751799e+00 1.11098111e+00 -5.63432038e-01
1.86437652e-01 2.85768390e-01 7.26490617e-01 4.89129871e-01
-1.15379822e+00 -1.36874199e-01 -7.23567009e-01 -5.21247387e-01
-8.84885043e-02 -4.11065876e-01 -6.56296983e-02 1.58231899e-01
-1.01056945e+00 1.11934030e+00 1.57162949e-01 3.75627190e-01
-1.31642029e-01 -4.79725331e-01 7.13119924e-01 7.64176071e-01
5.67111969e-01 5.91961682e-01 -7.40679085e-01 -2.89223135e-01
-4.08904016e-01 8.70508313e-01 1.42908180e+00 1.02684855e+00
4.82798755e-01 -3.48883808e-01 -8.90389800e-01 1.26153386e+00
2.97825724e-01 -9.24894512e-02 5.12089014e-01 -9.45913076e-01
7.44570553e-01 7.93899298e-01 2.99594551e-01 -3.61287951e-01
-8.42955172e-01 2.39762008e-01 -3.42973530e-01 -2.44764030e-01
6.54121816e-01 -3.00463319e-01 -4.11323607e-01 1.62646878e+00
4.31769013e-01 -3.31214786e-01 1.77083507e-01 1.01469684e+00
1.45010555e+00 3.63432258e-01 2.89189547e-01 -2.45035082e-01
1.56584191e+00 -1.31528580e+00 -6.46729112e-01 -1.64681464e-01
9.01354074e-01 -8.89761388e-01 1.29072797e+00 2.98596650e-01
-1.09704483e+00 -5.22288740e-01 -7.56785691e-01 -2.27015808e-01
-4.43498105e-01 -1.30377010e-01 5.73458016e-01 5.62489867e-01
-1.23039091e+00 1.66018575e-01 -5.75215034e-02 -5.04696429e-01
-7.13014305e-01 5.58671236e-01 -1.50431708e-01 1.86468750e-01
-1.87706816e+00 1.22795713e+00 1.77799854e-02 -1.74064577e-01
-7.37151325e-01 -3.79082263e-01 -6.88089490e-01 5.90299293e-02
6.20249569e-01 -3.58969420e-01 2.10838675e+00 -3.57089043e-01
-2.19124246e+00 6.92363143e-01 -3.22418571e-01 -6.18514299e-01
3.80074114e-01 -2.66880929e-01 -2.40503371e-01 9.33394879e-02
-1.63859650e-01 7.03800499e-01 3.59061062e-01 -9.92103815e-01
-1.04794383e+00 -1.93188950e-01 4.47604805e-01 6.51737213e-01
-1.60318129e-02 5.79677224e-01 -9.93822590e-02 8.16184059e-02
-9.17161629e-02 -1.07884252e+00 -4.43980008e-01 -9.99899626e-01
-3.20560336e-01 -1.15123665e+00 7.11516440e-01 -3.74926776e-01
1.47882462e+00 -1.56085324e+00 2.65103400e-01 -3.41185212e-01
1.93495780e-01 1.43856749e-01 3.11261583e-02 1.06430602e+00
3.93336803e-01 1.35765418e-01 2.55097687e-01 -9.24790978e-01
3.41889858e-01 -1.19109869e-01 -3.66032928e-01 -1.40795484e-01
-3.03438604e-01 6.01355255e-01 -8.85065734e-01 -3.95561904e-01
3.98973137e-01 -1.53897911e-01 -4.83092219e-01 8.43555033e-01
-6.78554714e-01 5.67482889e-01 -4.35869694e-01 2.92887509e-01
1.71513587e-01 1.18333861e-01 5.87082922e-01 7.70195350e-02
-1.86258748e-01 1.04675174e+00 -6.75098836e-01 1.57492721e+00
-6.38380826e-01 2.55782932e-01 2.83245653e-01 -5.15168011e-01
8.88003767e-01 6.98552430e-01 2.57196218e-01 -6.95354998e-01
-9.01913922e-03 2.74121314e-01 2.03317747e-01 -5.36452174e-01
1.03379023e+00 1.02795046e-02 -7.71728337e-01 4.89955574e-01
1.56708077e-01 -2.95442641e-01 6.55843467e-02 5.24467885e-01
1.00582588e+00 -1.57896966e-01 3.58232945e-01 -2.44664833e-01
9.86483276e-01 -6.14040755e-02 2.75275856e-01 1.19499362e+00
-3.76297116e-01 4.77408022e-01 3.83204132e-01 -5.02935648e-01
-4.42254573e-01 -3.53819013e-01 2.61330813e-01 2.21218324e+00
1.47208869e-02 -4.13069844e-01 -8.72450173e-01 -1.02998674e+00
-2.32573971e-01 8.42367172e-01 -2.34206453e-01 1.75509468e-01
-8.19930017e-01 -3.25454086e-01 7.14192033e-01 1.23842373e-01
3.75688612e-01 -1.30895936e+00 -4.88719255e-01 5.63240647e-01
-6.41974032e-01 -1.31654382e+00 -6.55465305e-01 1.91632658e-01
-2.61670619e-01 -8.69146883e-01 -2.35425249e-01 -7.28447080e-01
-2.57633090e-01 1.05656832e-01 1.47936630e+00 2.82537103e-01
1.68675348e-01 5.37626147e-01 -8.37365746e-01 -1.62066266e-01
-6.82483137e-01 4.43105340e-01 1.59773096e-01 -2.23304033e-02
6.35907710e-01 3.48976105e-02 -4.27148759e-01 8.09570253e-01
-2.43721351e-01 -9.54614300e-03 -8.13046619e-02 1.05269158e+00
-2.09419593e-01 -6.28356040e-01 1.08693993e+00 -1.00820661e+00
1.24876451e+00 -3.83345336e-01 -1.72421470e-01 5.57486951e-01
-5.51936090e-01 -1.43409729e-01 6.59151793e-01 -2.25269601e-01
-1.45151699e+00 6.26936480e-02 -5.98143280e-01 3.54671538e-01
-4.70217377e-01 4.36246186e-01 -3.09254006e-02 3.07452045e-02
6.72712982e-01 1.10034170e-02 -2.72795618e-01 -5.89294493e-01
4.24447179e-01 1.11022210e+00 1.51816577e-01 -7.91067123e-01
-2.62034208e-01 -1.74093813e-01 -4.64173257e-01 -8.17268252e-01
-6.97055697e-01 -1.01052582e+00 -6.82346225e-01 -3.55441868e-01
1.05007386e+00 -6.31761491e-01 -1.23007202e+00 4.60269839e-01
-1.31114602e+00 -6.73962831e-01 3.69852960e-01 1.50378197e-01
-5.30053735e-01 2.99083322e-01 -1.14336956e+00 -1.37351608e+00
-7.21782327e-01 -1.46652091e+00 9.17197347e-01 1.68075219e-01
-6.52837098e-01 -9.30715024e-01 9.75978896e-02 7.76905239e-01
6.03630602e-01 -4.81655896e-01 9.58899558e-01 -1.51330197e+00
-2.47515649e-01 -1.61907047e-01 3.23172480e-01 -8.91424567e-02
-2.34341428e-01 -4.13127393e-01 -1.00624549e+00 -2.09168211e-01
3.04578304e-01 -9.61569369e-01 7.28008926e-01 6.84942752e-02
5.05476654e-01 -2.86720335e-01 -1.46858647e-01 -3.19308817e-01
7.04738677e-01 5.77112794e-01 2.57014364e-01 8.49965885e-02
3.00713003e-01 1.27117014e+00 1.03268456e+00 3.81090909e-01
1.09458375e+00 1.34483564e+00 3.67938370e-01 3.87947142e-01
2.24380419e-01 -9.33161676e-02 3.49482626e-01 6.80945396e-01
4.73868340e-01 -5.30862093e-01 -9.80053186e-01 5.50374091e-01
-2.22093725e+00 -7.56623387e-01 -2.76380539e-01 2.17159033e+00
1.03447461e+00 3.44682008e-01 5.98455429e-01 -3.88617516e-01
5.04217803e-01 2.10439160e-01 -1.80171594e-01 -8.67962003e-01
1.60841748e-01 7.91246593e-02 -2.57940680e-01 1.21413076e+00
-1.18384874e+00 1.39358008e+00 5.75329065e+00 5.81712842e-01
-8.71694028e-01 1.63808748e-01 7.44904101e-01 1.45462543e-01
-1.58898812e-02 1.91834480e-01 -1.48184991e+00 2.21836582e-01
1.27284777e+00 1.11575373e-01 3.01253498e-01 8.47950518e-01
1.25952929e-01 -1.50512308e-01 -1.51112676e+00 3.96562010e-01
3.52062434e-02 -1.17969668e+00 -2.71734565e-01 -3.78446907e-01
1.79148257e-01 -9.47361514e-02 -1.77456409e-01 1.14041007e+00
6.77751243e-01 -1.04276907e+00 4.10838276e-01 2.46137470e-01
3.30421060e-01 -3.56648713e-01 1.01861203e+00 6.67654574e-01
-1.10326159e+00 -2.29505211e-01 -1.83281362e-01 -1.85454428e-01
4.45958912e-01 -1.52991936e-01 -1.40983248e+00 5.89991689e-01
4.22682196e-01 -3.25157754e-02 -3.39662492e-01 4.28418845e-01
1.12844907e-01 5.35169184e-01 -2.01873913e-01 -6.34330392e-01
5.21177471e-01 1.58412158e-01 6.48241997e-01 1.54795861e+00
-3.56365263e-01 5.33404589e-01 8.69730711e-01 2.01610848e-01
-5.73427305e-02 2.65591651e-01 -1.95751891e-01 4.44757789e-01
8.71508241e-01 1.15029562e+00 -1.95763394e-01 -3.70777130e-01
-2.67447025e-01 6.86082780e-01 2.60887235e-01 4.70780954e-02
-3.19334418e-01 -9.12825912e-02 4.33057040e-01 -3.06640804e-01
-2.14327648e-01 6.35779053e-02 -1.72219664e-01 -8.43173862e-01
-2.63633609e-01 -1.32607114e+00 6.85330331e-01 -1.28827512e-01
-1.37680328e+00 1.10353827e+00 2.17079893e-01 -9.29759622e-01
-8.89274061e-01 -4.04342324e-01 -7.06434250e-01 9.92109597e-01
-9.73357260e-01 -1.36006439e+00 1.70885324e-02 5.34206688e-01
1.45568991e+00 -4.01868999e-01 1.34135211e+00 1.40502363e-01
-3.18427145e-01 7.80795336e-01 -4.01150882e-01 -2.06769723e-02
1.14465022e+00 -1.60052001e+00 4.73095447e-01 6.44814372e-02
-2.66873837e-01 7.51421988e-01 8.71052146e-01 -4.40853745e-01
-1.14517629e+00 -3.30277026e-01 1.44091237e+00 -7.92201102e-01
4.73138630e-01 -6.27019227e-01 -8.88987243e-01 4.33105171e-01
5.89228332e-01 -6.92924500e-01 7.32077420e-01 6.14156842e-01
-1.40890971e-01 1.58378527e-01 -1.11490655e+00 6.65600717e-01
6.78996027e-01 -6.33223951e-01 -7.33861804e-01 5.05514801e-01
1.06835854e+00 -7.87818730e-01 -8.07858050e-01 4.03388649e-01
6.90107763e-01 -1.16607833e+00 9.05129373e-01 -1.02287734e+00
3.02304178e-01 2.95593202e-01 -3.74234140e-01 -1.17736399e+00
8.35020468e-02 -9.28251088e-01 -4.21953946e-02 1.31808615e+00
6.13368869e-01 -2.93978244e-01 8.23048770e-01 1.09289753e+00
-4.52411205e-01 -9.51853454e-01 -1.09190071e+00 -2.39830494e-01
2.20886573e-01 -2.00487778e-01 5.95670581e-01 6.78166449e-01
7.13167369e-01 1.18529272e+00 -7.92051613e-01 -3.72223347e-01
4.64641452e-02 2.22690418e-01 9.70087588e-01 -1.13124359e+00
-2.84761757e-01 -6.41608596e-01 1.34033903e-01 -1.72234321e+00
3.70617092e-01 -5.75974405e-01 3.93843591e-01 -1.38916647e+00
-2.55224466e-01 -4.75451082e-01 2.54662223e-02 2.78430253e-01
-4.50541496e-01 -5.09733200e-01 5.38218737e-01 1.32785812e-01
-1.00220573e+00 5.63902259e-01 5.85738957e-01 -8.78389031e-02
-6.38614118e-01 5.91639400e-01 -4.76038724e-01 5.17828166e-01
6.36009634e-01 -2.15598673e-01 -5.26967347e-01 -8.92270580e-02
-2.11469263e-01 9.47337925e-01 -1.81679934e-01 -5.79785407e-01
6.48164988e-01 -3.33704621e-01 -5.37354469e-01 -7.56802261e-01
7.45456636e-01 -5.22054911e-01 -5.68723440e-01 1.76005229e-01
-1.24569678e+00 -2.68548001e-02 -8.44587833e-02 3.50569397e-01
-1.78211838e-01 -4.99429196e-01 5.61947525e-01 -3.97091478e-01
-9.02525783e-01 -1.07312918e-01 -7.71106958e-01 5.20929024e-02
5.72534621e-01 3.46340805e-01 -6.86356187e-01 -1.02750731e+00
-9.23048794e-01 7.31183112e-01 -1.88046262e-01 8.11805665e-01
7.02878416e-01 -8.10897291e-01 -9.44373965e-01 -4.23877567e-01
2.45100304e-01 -4.36444163e-01 5.15604556e-01 7.24529505e-01
2.48982408e-03 9.29038346e-01 8.67832601e-02 -5.67246318e-01
-1.73745346e+00 2.23061144e-01 4.75084901e-01 -1.04532874e+00
-2.45813772e-01 1.21627533e+00 -5.65160923e-02 -1.19860816e+00
5.43675005e-01 4.26473096e-02 -7.69712329e-01 1.61488846e-01
6.96765006e-01 2.09070057e-01 2.80889988e-01 -5.91424108e-01
-4.15878385e-01 -2.85597414e-01 -5.98134100e-01 -5.66514015e-01
7.01294839e-01 -4.61940140e-01 -1.68227613e-01 5.44388056e-01
9.29823458e-01 -2.63598025e-01 -6.56516790e-01 -6.06882811e-01
5.23492515e-01 -1.23616196e-01 -4.13139462e-01 -1.19495499e+00
-7.60411993e-02 7.68949628e-01 4.67754096e-01 7.14220583e-01
6.38724685e-01 -2.21200168e-01 9.51294124e-01 8.98526430e-01
7.71386802e-01 -1.26966727e+00 2.34971389e-01 1.33936310e+00
1.10295498e+00 -1.56307232e+00 -2.93904424e-01 -4.60570872e-01
-1.31118309e+00 9.29649055e-01 1.24088717e+00 3.49442303e-01
4.69487876e-01 -1.29177049e-01 3.20580602e-01 -1.51074141e-01
-1.49632537e+00 -2.26425067e-01 2.35427037e-01 3.53112161e-01
9.58956599e-01 1.91664785e-01 -5.76130271e-01 7.33931720e-01
-2.08930880e-01 -5.77017486e-01 4.21583056e-01 8.13846231e-01
-5.91043293e-01 -1.53992617e+00 -5.50591983e-02 4.31678891e-01
-4.36458081e-01 -2.41593108e-01 -1.14040732e+00 4.88373458e-01
-5.37361383e-01 1.78280675e+00 -5.23566306e-01 -9.75994647e-01
6.11645997e-01 6.70094311e-01 -1.27363533e-01 -1.03085411e+00
-1.66806495e+00 -4.56902273e-02 9.81817603e-01 -4.26996440e-01
-2.07738727e-01 -3.90032768e-01 -9.97344077e-01 -2.22537011e-01
-3.23450714e-01 5.27544856e-01 4.12103146e-01 1.10806191e+00
1.94364801e-01 1.88280910e-01 8.49983335e-01 -1.00747812e+00
-1.15133882e+00 -1.45863378e+00 -2.85514002e-03 4.24143255e-01
1.53205737e-01 -5.59461176e-01 -1.75583348e-01 -4.37280029e-01] | [12.711976051330566, 7.887463092803955] |
7bf63a7f-d47b-4bcd-9687-d4c8d6fb62ab | textattack-a-framework-for-adversarial | 2005.05909 | null | https://arxiv.org/abs/2005.05909v4 | https://arxiv.org/pdf/2005.05909v4.pdf | TextAttack: A Framework for Adversarial Attacks, Data Augmentation, and Adversarial Training in NLP | While there has been substantial research using adversarial attacks to analyze NLP models, each attack is implemented in its own code repository. It remains challenging to develop NLP attacks and utilize them to improve model performance. This paper introduces TextAttack, a Python framework for adversarial attacks, data augmentation, and adversarial training in NLP. TextAttack builds attacks from four components: a goal function, a set of constraints, a transformation, and a search method. TextAttack's modular design enables researchers to easily construct attacks from combinations of novel and existing components. TextAttack provides implementations of 16 adversarial attacks from the literature and supports a variety of models and datasets, including BERT and other transformers, and all GLUE tasks. TextAttack also includes data augmentation and adversarial training modules for using components of adversarial attacks to improve model accuracy and robustness. TextAttack is democratizing NLP: anyone can try data augmentation and adversarial training on any model or dataset, with just a few lines of code. Code and tutorials are available at https://github.com/QData/TextAttack. | ['Eli Lifland', 'Yanjun Qi', 'John X. Morris', 'Jake Grigsby', 'Jin Yong Yoo', 'Di Jin'] | 2020-04-29 | null | https://aclanthology.org/2020.emnlp-demos.16 | https://aclanthology.org/2020.emnlp-demos.16.pdf | emnlp-2020-11 | ['adversarial-text'] | ['adversarial'] | [-1.33320883e-01 2.68157005e-01 -3.51545811e-01 -5.50052524e-01
-9.21482325e-01 -1.40838718e+00 7.08966076e-01 1.14646144e-01
-3.26410383e-02 5.53535104e-01 1.74502805e-01 -8.21563482e-01
3.46779883e-01 -9.17805314e-01 -8.05963278e-01 -2.87179470e-01
1.12442479e-01 6.77173793e-01 -5.45758195e-02 -2.79626966e-01
-1.00249104e-01 7.11983562e-01 -6.41506851e-01 4.16469663e-01
5.97288728e-01 6.02647781e-01 -6.40204430e-01 1.02690208e+00
-2.71083474e-01 9.40935671e-01 -8.79861712e-01 -8.11904073e-01
5.94653785e-01 7.11663663e-02 -8.85758936e-01 -9.14800167e-01
5.81387699e-01 -2.86450356e-01 -5.08318484e-01 8.99841785e-01
7.80385852e-01 -7.22420737e-02 2.30958298e-01 -1.83545160e+00
-9.03135061e-01 9.56932247e-01 -4.56109847e-04 1.92943126e-01
5.38191557e-01 7.13192761e-01 9.43063676e-01 -5.37052512e-01
3.19442391e-01 1.34101176e+00 9.01029885e-01 1.01399648e+00
-1.35991812e+00 -1.32945395e+00 -6.79909438e-02 -1.48718804e-01
-9.82637465e-01 -3.75137389e-01 6.38963699e-01 -3.98791254e-01
1.12375200e+00 5.63530266e-01 4.42521811e-01 1.80636656e+00
1.99004769e-01 8.20336878e-01 7.06894875e-01 -1.60302162e-01
1.45331532e-01 2.07083941e-01 2.52812117e-01 5.54979444e-01
-1.21244393e-01 5.53121209e-01 -1.95629746e-01 -7.87997842e-01
4.30337846e-01 -3.30831856e-01 8.36787373e-02 -6.07330315e-02
-6.37570322e-01 9.88456786e-01 5.29957891e-01 2.54008416e-02
1.58498645e-01 3.87140483e-01 5.98910987e-01 2.97135323e-01
2.70053953e-01 9.38196599e-01 -8.18461776e-01 -1.03871353e-01
-3.80785704e-01 7.28780270e-01 1.12819338e+00 1.07042074e+00
3.35196435e-01 3.83260280e-01 -1.10947929e-01 5.44615984e-01
2.61779487e-01 6.41409934e-01 3.33316714e-01 -8.39259744e-01
6.36101365e-01 4.14514840e-01 -9.58192423e-02 -9.55876410e-01
-5.23916543e-01 -1.06333129e-01 -2.89543778e-01 2.50118732e-01
4.68721658e-01 -6.94368303e-01 -9.38598394e-01 1.81721294e+00
4.06330347e-01 1.97403178e-01 2.49716178e-01 5.21068513e-01
9.32169199e-01 7.52501369e-01 5.49716830e-01 5.31719685e-01
1.10563409e+00 -8.48381817e-01 -4.43097740e-01 -6.10507429e-01
7.03531921e-01 -1.01935685e+00 1.30580664e+00 2.81523585e-01
-1.21812344e+00 -1.59387976e-01 -9.02826250e-01 -3.35202575e-01
-8.28020334e-01 -4.73552793e-01 1.02127779e+00 1.01179695e+00
-6.48221791e-01 4.43752348e-01 -9.32963014e-01 -4.13682424e-02
5.47611833e-01 3.87339860e-01 -3.22425097e-01 9.72855389e-02
-1.63932598e+00 9.47487175e-01 4.58991915e-01 -3.02517951e-01
-1.05499923e+00 -1.26308036e+00 -1.08212006e+00 -2.16920618e-02
1.05386235e-01 -6.97508574e-01 1.45595062e+00 -7.62067258e-01
-1.38530600e+00 4.39641088e-01 2.94376254e-01 -7.03368902e-01
3.80425155e-01 -3.50125432e-01 -6.13541067e-01 -2.24324971e-01
-2.18628764e-01 4.91011053e-01 7.16940224e-01 -9.29299593e-01
-4.62818816e-02 -7.85445645e-02 3.98729056e-01 -1.22631453e-01
-4.33678687e-01 3.92570138e-01 -1.72088444e-01 -9.14779603e-01
-5.02903759e-01 -8.82003725e-01 -3.36389959e-01 3.52877006e-02
-6.61356926e-01 -7.61331525e-03 1.07987571e+00 -6.15211189e-01
1.04984164e+00 -1.95752990e+00 -3.22946906e-01 3.76443565e-01
1.19596362e-01 6.90295696e-01 -3.60189259e-01 4.84390855e-01
-5.47436953e-01 7.31188357e-01 -1.83456764e-01 -1.14709683e-01
4.65537637e-01 3.29840332e-01 -7.20641434e-01 2.75526702e-01
3.41328770e-01 1.20813870e+00 -8.17946911e-01 -2.39493668e-01
4.05606449e-01 3.62571150e-01 -9.56196904e-01 2.34389827e-01
-6.53512836e-01 3.80219609e-01 -4.99931008e-01 7.89386809e-01
8.92695487e-01 1.79022238e-01 1.22542664e-01 9.06312019e-02
2.69476622e-01 6.23698831e-01 -1.06438196e+00 1.33798039e+00
-3.70355248e-01 4.96764094e-01 1.39303312e-01 -6.58667505e-01
7.63084233e-01 3.16979527e-01 3.58200133e-01 -2.38240942e-01
1.47253990e-01 -6.01348802e-02 -1.07464664e-01 -3.53626043e-01
4.14538503e-01 -7.20861703e-02 -4.48898762e-01 5.85391700e-01
7.66428784e-02 -5.86799443e-01 1.62143055e-02 3.96767616e-01
1.50538754e+00 7.68534616e-02 2.93534636e-01 1.09736890e-01
2.85985649e-01 9.88242477e-02 5.93039989e-01 7.93636620e-01
-2.83806980e-01 2.26696655e-01 6.24466598e-01 -6.47482812e-01
-1.00379622e+00 -1.41803133e+00 -2.33198583e-01 1.31547868e+00
-5.44291735e-01 -7.53144741e-01 -5.62825322e-01 -1.02468717e+00
4.24121588e-01 9.22686279e-01 -5.34978151e-01 -3.81627411e-01
-5.88392556e-01 -7.08059967e-01 1.60868657e+00 7.51758039e-01
3.57491106e-01 -1.25483859e+00 3.36294025e-01 3.57201509e-02
-3.41893472e-02 -1.02760613e+00 -6.12449408e-01 3.00109595e-01
-3.69740635e-01 -1.22058880e+00 2.91239977e-01 -7.92803049e-01
4.36689824e-01 -5.34760892e-01 1.21136248e+00 -1.01925611e-01
-1.89690977e-01 2.88753688e-01 -3.03275794e-01 -9.96861041e-01
-1.04441798e+00 1.57913536e-01 9.10876691e-02 -7.28543639e-01
4.06068444e-01 -6.97525620e-01 2.25237776e-02 1.91160381e-01
-1.03873134e+00 -4.98263687e-01 9.65602621e-02 6.96105123e-01
1.68607920e-01 -7.63284788e-02 4.62346822e-01 -1.14983475e+00
7.40752697e-01 -6.16337061e-01 -9.28743303e-01 1.16988188e-02
-4.61554229e-01 -1.97365850e-01 1.07529175e+00 -5.36478102e-01
-6.32852435e-01 1.19375534e-01 -7.93819129e-01 -4.62264150e-01
-7.84079283e-02 3.04856777e-01 -4.56878632e-01 -3.14900339e-01
1.22609043e+00 -1.55311301e-01 -3.44908647e-02 -3.29617321e-01
7.11670637e-01 4.17905301e-01 4.18697089e-01 -8.84250164e-01
1.32048023e+00 -1.09170657e-02 -3.77440304e-01 -3.26774508e-01
-6.60105050e-01 1.07951857e-01 -2.53822386e-01 3.21517080e-01
5.44166028e-01 -6.01880729e-01 -9.92956221e-01 4.81510073e-01
-1.13439572e+00 -6.51276827e-01 -5.90383351e-01 1.99014947e-01
-3.00396711e-01 2.47289047e-01 -8.90287101e-01 -3.37374568e-01
-6.41306162e-01 -1.05156159e+00 4.00187075e-01 1.13410130e-01
-4.16772068e-01 -1.25843179e+00 1.42457098e-01 4.83094573e-01
4.21700180e-01 4.22966659e-01 8.95616531e-01 -1.39161253e+00
-3.74459714e-01 -6.20529592e-01 2.27854222e-01 6.88513935e-01
-5.14048673e-02 3.26186538e-01 -1.07680571e+00 -2.47602865e-01
-6.74994811e-02 -7.09795892e-01 1.99266210e-01 8.48915130e-02
1.34124839e+00 -9.66065228e-01 -3.92345935e-01 9.59464908e-01
8.75045061e-01 2.38964006e-01 6.70051157e-01 3.51649076e-01
7.44815886e-01 1.25083253e-01 2.14114010e-01 2.79392719e-01
1.72088087e-01 4.12734210e-01 7.05994368e-01 -8.56622979e-02
2.79343367e-01 -4.42599058e-01 6.85266435e-01 2.06948936e-01
6.48175597e-01 -4.00505453e-01 -1.36882627e+00 3.32507312e-01
-1.27760756e+00 -1.06990814e+00 8.11855793e-02 1.75426877e+00
1.27354276e+00 2.33172029e-01 -3.71594355e-02 -7.78572932e-02
3.37382317e-01 3.05432767e-01 -7.20697105e-01 -1.03010380e+00
-1.57478768e-02 6.73790693e-01 5.63634276e-01 8.28394830e-01
-1.40264952e+00 1.39953113e+00 7.51541328e+00 6.57122493e-01
-1.08013046e+00 2.19409596e-02 2.34939113e-01 -2.37229839e-01
-5.11256397e-01 1.15455508e-01 -9.00096834e-01 4.45121408e-01
9.75120842e-01 -6.59973919e-01 6.00773871e-01 1.02658224e+00
-7.04660341e-02 6.23904526e-01 -1.18482900e+00 2.81114042e-01
-9.88573059e-02 -1.42167175e+00 2.99035668e-01 -1.86844423e-01
5.34878910e-01 3.04503918e-01 1.84872538e-01 7.13405311e-01
1.40548575e+00 -1.34804118e+00 4.25703317e-01 3.13764811e-02
3.44443828e-01 -8.03165674e-01 5.00029802e-01 1.25248507e-01
-8.49849999e-01 -2.92890459e-01 -1.41176268e-01 -2.46642325e-02
-9.79021937e-02 2.20635235e-01 -1.12541604e+00 4.05428976e-01
5.97118735e-01 6.37179077e-01 -5.89889646e-01 5.28379738e-01
-5.80156565e-01 9.24182534e-01 -5.61500669e-01 2.03136921e-01
1.40289634e-01 2.97950834e-01 8.10865939e-01 1.27998698e+00
-3.64306927e-01 -2.71107507e-04 3.98747921e-01 9.16322589e-01
-3.65292668e-01 -5.41081168e-02 -6.65371060e-01 -4.68344063e-01
9.07238066e-01 1.23909760e+00 8.60106796e-02 -1.41523257e-01
-2.37346768e-01 5.34055114e-01 2.20922992e-01 2.35198259e-01
-1.25419390e+00 -4.75646377e-01 1.24617040e+00 -4.27291356e-02
-2.64646918e-01 -2.70484447e-01 -5.27901590e-01 -1.06186295e+00
-9.86269191e-02 -1.59941208e+00 6.92175865e-01 -7.07701027e-01
-1.52939188e+00 3.79835188e-01 -1.66669235e-01 -8.26121628e-01
-2.50136763e-01 -6.29647791e-01 -8.41660380e-01 1.07817531e+00
-9.66417193e-01 -1.43526244e+00 -1.20123610e-01 9.16347146e-01
1.05241165e-01 -3.81506532e-01 1.33162093e+00 3.44891697e-01
-8.47292244e-01 1.14770341e+00 -1.13134570e-01 7.03625977e-01
9.21164334e-01 -1.37180257e+00 1.24741983e+00 1.01658177e+00
-3.83984707e-02 8.75272512e-01 7.84671903e-01 -7.45035172e-01
-1.30718911e+00 -1.29371679e+00 5.31111538e-01 -1.10196435e+00
1.14509630e+00 -7.94186234e-01 -8.95344496e-01 1.51551056e+00
4.97392416e-02 2.74986237e-01 8.89306962e-01 1.99717015e-01
-8.59401166e-01 -5.38223907e-02 -1.55087996e+00 8.23371589e-01
6.44781172e-01 -6.58430934e-01 -5.79618573e-01 8.29577804e-01
9.95725513e-01 -9.99504566e-01 -1.15283144e+00 2.37086311e-01
5.18975198e-01 -3.75596613e-01 1.29106152e+00 -1.06568742e+00
3.15334409e-01 -6.94241896e-02 -7.67345205e-02 -1.26504898e+00
-1.16497874e-01 -1.02572656e+00 -2.28928924e-01 1.40261149e+00
6.87496901e-01 -1.15167797e+00 1.02006233e+00 1.10139573e+00
-1.68738469e-01 -5.54915726e-01 -5.72922587e-01 -7.28014827e-01
8.21033537e-01 -9.00494635e-01 9.96168673e-01 1.33830452e+00
8.12797174e-02 2.29926437e-01 -2.45305240e-01 4.69228923e-01
3.20438325e-01 -5.38373053e-01 1.19972646e+00 -7.13157296e-01
-3.57434571e-01 -1.96971327e-01 -2.56328851e-01 -5.81556439e-01
3.78901273e-01 -1.37809467e+00 -3.16129684e-01 -1.18284714e+00
-3.58157218e-01 -8.56402218e-01 7.44309351e-02 1.39933026e+00
-1.36279210e-01 8.04781467e-02 3.88296068e-01 3.81730571e-02
8.58769566e-02 3.50469172e-01 9.79545653e-01 -4.27703291e-01
-3.66326600e-01 4.92720790e-02 -1.13136733e+00 8.70321810e-01
1.35898995e+00 -7.10312068e-01 -3.24836642e-01 -5.43496907e-01
1.92608982e-01 -4.44258690e-01 4.94791508e-01 -9.80904162e-01
2.35369071e-01 -3.07662070e-01 6.30787849e-01 -2.96607912e-01
4.39091653e-01 -8.12406659e-01 7.25495517e-02 3.82173359e-01
-4.90319401e-01 2.70783871e-01 9.63671029e-01 1.19094573e-01
1.88580096e-01 -1.69496596e-01 9.25166070e-01 1.16080102e-02
-2.80739456e-01 3.49064112e-01 -2.59622365e-01 3.82662982e-01
1.17481232e+00 2.75054693e-01 -8.67639482e-01 -2.44861394e-01
-9.48559105e-01 5.86581409e-01 3.40463936e-01 7.10538208e-01
3.63629133e-01 -1.16062546e+00 -6.13055110e-01 5.46893001e-01
-1.19996034e-01 2.62698391e-04 -3.09324265e-03 7.72100165e-02
-7.09574342e-01 1.93029478e-01 -1.81557626e-01 4.51954454e-02
-1.27325392e+00 7.83480525e-01 6.68337822e-01 -5.45991659e-01
-4.02331561e-01 8.64044845e-01 -4.65229154e-03 -1.15389657e+00
1.99284628e-01 -2.93831915e-01 1.79945588e-01 -3.00377131e-01
6.24717832e-01 2.71259367e-01 -6.44522086e-02 -9.34188515e-02
-4.82871592e-01 3.84919457e-02 -3.08839113e-01 5.83493076e-02
1.09661365e+00 5.38114071e-01 -7.45920688e-02 -8.35504979e-02
9.89016175e-01 3.85066897e-01 -9.09331977e-01 -3.88134234e-02
-3.37153763e-01 -1.88436851e-01 -3.66045535e-01 -1.35866511e+00
-9.84667361e-01 5.58192849e-01 1.39088079e-01 2.26206720e-01
9.23569620e-01 -1.77338138e-01 9.37815011e-01 4.51971292e-01
3.06303315e-02 -6.86379313e-01 5.67588247e-02 8.98487687e-01
1.15724385e+00 -8.72877419e-01 6.80366084e-02 -4.53599960e-01
-5.45019984e-01 7.71635771e-01 8.11866581e-01 -1.59886301e-01
8.03709686e-01 9.70725477e-01 3.67445111e-01 -9.15747974e-03
-7.29784369e-01 3.08537066e-01 -1.36939436e-02 8.89657617e-01
2.84578115e-01 8.06501955e-02 1.20068580e-01 9.11567628e-01
-8.23685408e-01 -1.42378792e-01 3.24600965e-01 9.55417216e-01
1.29401073e-01 -1.68745649e+00 -5.57201326e-01 3.52612108e-01
-8.64918172e-01 -4.13995177e-01 -6.81432903e-01 1.02955580e+00
4.77082394e-02 9.80936527e-01 -1.41229093e-01 -6.53436303e-01
4.28337574e-01 2.94903994e-01 1.61506295e-01 -7.77150571e-01
-1.36588871e+00 -6.82913482e-01 3.26916844e-01 -7.30163455e-01
2.55019933e-01 -4.62141424e-01 -1.26036108e+00 -9.81105268e-01
2.29579909e-03 2.61744231e-01 6.80709243e-01 4.42678154e-01
4.02534127e-01 4.06912953e-01 5.38607001e-01 -5.40428102e-01
-5.55815399e-01 -8.40074003e-01 -1.45417348e-01 1.87454581e-01
1.28314421e-01 -1.07684389e-01 -4.67578471e-01 -1.30648270e-01] | [6.023875713348389, 8.096360206604004] |
13fa0700-046a-4440-8249-493eea3f9b32 | adversarial-disentanglement-of-speaker | 2012.04454 | null | https://arxiv.org/abs/2012.04454v3 | https://arxiv.org/pdf/2012.04454v3.pdf | Adversarial Disentanglement of Speaker Representation for Attribute-Driven Privacy Preservation | In speech technologies, speaker's voice representation is used in many applications such as speech recognition, voice conversion, speech synthesis and, obviously, user authentication. Modern vocal representations of the speaker are based on neural embeddings. In addition to the targeted information, these representations usually contain sensitive information about the speaker, like the age, sex, physical state, education level or ethnicity. In order to allow the user to choose which information to protect, we introduce in this paper the concept of attribute-driven privacy preservation in speaker voice representation. It allows a person to hide one or more personal aspects to a potential malicious interceptor and to the application provider. As a first solution to this concept, we propose to use an adversarial autoencoding method that disentangles in the voice representation a given speaker attribute thus allowing its concealment. We focus here on the sex attribute for an Automatic Speaker Verification (ASV) task. Experiments carried out using the VoxCeleb datasets have shown that the proposed method enables the concealment of this attribute while preserving ASV ability. | ['Andreas Nautsch', 'Jean-François Bonastre', 'Titouan Parcollet', 'Driss Matrouf', 'Mohammad Mohammadamini', 'Paul-Gauthier Noé'] | 2020-12-08 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 3.24145645e-01 3.42511922e-01 -1.40864491e-01 -5.19044876e-01
-4.22126591e-01 -6.66190863e-01 6.87870800e-01 3.76499385e-01
-4.83168542e-01 5.56720197e-01 3.74402076e-01 -3.85349602e-01
4.15596403e-02 -6.79693878e-01 -4.94210124e-01 -8.62061262e-01
-1.08979858e-01 1.50699854e-01 -2.38545671e-01 1.03774041e-01
1.54631892e-02 7.70199180e-01 -1.51802945e+00 -3.86237502e-02
3.64473403e-01 1.02225757e+00 -4.48344976e-01 5.45992374e-01
-1.72513500e-01 2.87422359e-01 -9.02161121e-01 -6.33839011e-01
2.22399890e-01 -2.80684888e-01 -6.11833334e-01 -1.00144826e-01
1.51607066e-01 -2.63349980e-01 -2.95911431e-01 1.23101032e+00
6.09038711e-01 1.31101599e-02 6.51229560e-01 -1.32429838e+00
-4.49399024e-01 7.86045253e-01 -5.68966046e-02 7.45675862e-02
4.81056035e-01 -1.09029643e-01 9.17995691e-01 -4.73223776e-01
3.92191648e-01 1.29242289e+00 3.26817662e-01 9.38747704e-01
-1.33214641e+00 -8.09349000e-01 -1.20239541e-01 6.04390278e-02
-1.72294450e+00 -8.11568201e-01 1.07260942e+00 -4.47728544e-01
1.96376326e-03 7.76438653e-01 4.78872895e-01 1.25176728e+00
-5.14026247e-02 5.82833171e-01 1.04152548e+00 -3.22970241e-01
4.17486995e-01 9.64565635e-01 5.25546223e-02 4.23092574e-01
8.97249952e-02 1.47630125e-01 -3.24473709e-01 -7.39294231e-01
2.58804351e-01 -6.12100251e-02 -5.54003298e-01 -4.34143066e-01
-9.05483544e-01 1.02270782e+00 2.86611319e-01 2.54015744e-01
-2.90906012e-01 -1.53500319e-03 4.54315215e-01 2.39578798e-01
3.37515891e-01 2.68459439e-01 -1.30735993e-01 2.14235872e-01
-6.19913280e-01 2.24275336e-01 9.28671837e-01 4.84795362e-01
5.59508324e-01 1.54708341e-01 -1.90126270e-01 6.28026664e-01
5.64535677e-01 3.85980964e-01 8.07280779e-01 -3.51577967e-01
3.09397459e-01 2.71822661e-01 -5.82490452e-02 -1.11777020e+00
8.89353603e-02 -1.41652361e-01 -9.06422496e-01 2.84930438e-01
1.53845325e-01 -1.29752234e-01 -6.43947601e-01 2.02322936e+00
5.06040096e-01 1.37480900e-01 3.05014879e-01 8.21923256e-01
7.04388916e-01 5.70153832e-01 1.11065403e-01 -8.53360370e-02
1.75462794e+00 -2.96905071e-01 -8.36145818e-01 2.66719097e-03
9.83918607e-02 -5.66496670e-01 7.34030187e-01 2.32048139e-01
-6.07701719e-01 -1.98694468e-01 -1.09666348e+00 1.03231788e-01
-6.00035727e-01 -8.60140473e-02 4.34057504e-01 1.44555616e+00
-8.31565976e-01 3.57859612e-01 -4.54760551e-01 -2.13030159e-01
3.89792562e-01 7.14087903e-01 -7.51087546e-01 2.79595286e-01
-1.54094803e+00 3.95175964e-01 3.12199026e-01 -2.31325049e-02
-7.32509434e-01 -5.98282993e-01 -1.12559533e+00 3.00485641e-01
-5.44120185e-02 -3.87511194e-01 8.65600467e-01 -9.59584117e-01
-1.67886686e+00 9.41877782e-01 -6.78489311e-03 -7.50874281e-01
4.67345148e-01 2.80199826e-01 -7.63093948e-01 -1.66801251e-02
-2.54348725e-01 2.57590413e-01 1.61734462e+00 -8.30653965e-01
-2.29127541e-01 -7.36249328e-01 -1.65070266e-01 -8.45671818e-02
-5.40223837e-01 1.87413782e-01 1.44829139e-01 -8.00873041e-01
-1.32664934e-01 -8.31500292e-01 2.15765402e-01 6.46657571e-02
-6.98150277e-01 -2.29873788e-03 1.14528060e+00 -8.73945773e-01
1.01048505e+00 -2.61335564e+00 2.27916703e-01 3.87967110e-01
1.06497951e-01 3.57969522e-01 2.72515327e-01 2.13430911e-01
-3.03617924e-01 1.94907516e-01 -3.41880202e-01 -5.60326993e-01
5.00967950e-02 3.45652066e-02 -4.82494414e-01 9.66590583e-01
7.38120526e-02 2.67863721e-01 -2.89146423e-01 -3.61573815e-01
1.36227250e-01 1.02443135e+00 -6.59473419e-01 2.39327312e-01
1.68125898e-01 3.24425340e-01 -6.19228303e-01 3.32779080e-01
7.85031199e-01 6.48105323e-01 -9.62955877e-04 7.68246204e-02
1.95949107e-01 3.72322440e-01 -1.28782916e+00 1.15933740e+00
-4.44071233e-01 5.84973931e-01 6.88192248e-01 -6.39590740e-01
1.09060693e+00 7.32542872e-01 1.88851923e-01 2.85078194e-02
4.24090236e-01 2.57806424e-02 1.26895532e-01 -1.36110842e-01
4.01652515e-01 -1.94843203e-01 -1.35278150e-01 3.76715869e-01
-1.80773169e-01 2.75816321e-01 -6.28787458e-01 -2.99765989e-02
6.24064207e-01 -5.56048632e-01 5.23929298e-01 -2.36756101e-01
1.10779226e+00 -8.54554772e-01 3.77046525e-01 1.18006818e-01
-3.80874127e-01 4.43348914e-01 4.34598833e-01 1.67674329e-02
-7.03742802e-01 -8.27741027e-01 -2.82285273e-01 6.73600554e-01
-4.03237343e-01 -1.54136017e-01 -8.56497884e-01 -7.51896739e-01
2.67112136e-01 7.48330414e-01 -7.15050876e-01 -5.14212728e-01
-3.71966928e-01 -2.16625273e-01 8.47888410e-01 7.31790513e-02
1.51878998e-01 -8.21746230e-01 -3.34933788e-01 5.44798076e-02
9.90707651e-02 -8.45804870e-01 -7.56684661e-01 -7.52942264e-02
-4.71271724e-01 -5.66435397e-01 -7.69108474e-01 -5.00888586e-01
5.20596385e-01 -3.30561191e-01 3.11437100e-01 -7.31267929e-02
-2.25856945e-01 4.67469931e-01 -1.31548956e-01 -5.38094699e-01
-9.23270881e-01 1.55543908e-01 4.35805947e-01 1.02025390e+00
2.57087648e-01 -5.84069610e-01 -4.10626739e-01 5.25440983e-02
-9.66642737e-01 -6.85011685e-01 1.82839558e-01 6.01281583e-01
1.63310170e-02 1.59159377e-01 5.35443902e-01 -9.67665851e-01
8.52411211e-01 -4.35726911e-01 -5.09040356e-01 4.50467989e-02
-4.72805530e-01 2.69999683e-01 5.14684498e-01 -6.54462814e-01
-7.50830948e-01 3.29006106e-01 -4.31094110e-01 -3.64446014e-01
-3.41938138e-01 1.04169138e-01 -9.11722004e-01 -2.03291371e-01
2.30191663e-01 5.26435912e-01 3.23072910e-01 -6.73478246e-01
3.91821712e-01 1.23742318e+00 4.37428355e-01 -2.45157748e-01
8.23591232e-01 1.81276843e-01 -1.24213345e-01 -1.22568738e+00
1.32512569e-01 -3.95824522e-01 -3.28847468e-01 4.62056026e-02
8.92943561e-01 -6.09941304e-01 -1.11145067e+00 4.40183729e-01
-1.15543437e+00 6.40395463e-01 -8.25367048e-02 4.57674056e-01
-4.15849507e-01 3.78498346e-01 -2.41007313e-01 -1.12175953e+00
-4.84473228e-01 -1.15423954e+00 8.20083559e-01 5.93023263e-02
-2.45706096e-01 -9.11237419e-01 -1.14829086e-01 2.76810378e-01
5.10683537e-01 1.37364209e-01 1.03766239e+00 -1.33372676e+00
-2.80256301e-01 -7.55695820e-01 3.54448259e-01 5.92555821e-01
4.66224700e-01 -2.91219771e-01 -1.40594316e+00 -3.61178488e-01
3.29091132e-01 1.91846982e-01 6.17703080e-01 4.02189977e-02
1.11089599e+00 -1.01658607e+00 -9.48455557e-02 7.56363928e-01
1.02540839e+00 1.33750409e-01 6.52023911e-01 -8.66690874e-02
5.99092841e-01 8.87136281e-01 2.59299204e-02 4.82018203e-01
-4.62360866e-02 9.09198165e-01 4.89056289e-01 2.59398341e-01
8.19520876e-02 -4.40465868e-01 5.35267115e-01 3.96723121e-01
2.93142885e-01 -2.33659595e-01 -3.34069937e-01 4.01211023e-01
-1.21614444e+00 -8.99471462e-01 3.41533005e-01 2.66645551e+00
7.05313981e-01 -1.74208179e-01 1.90176770e-01 5.78728557e-01
8.97538066e-01 4.25548583e-01 -3.11005652e-01 -8.04677963e-01
9.51148495e-02 3.79144028e-02 5.28823137e-01 7.25120425e-01
-1.16462326e+00 6.59868121e-01 4.74640417e+00 6.62853420e-01
-1.40195990e+00 7.45694935e-02 3.99169207e-01 1.31231353e-01
-3.44928980e-01 -2.44013578e-01 -7.32810795e-01 4.96001363e-01
1.02885330e+00 -4.48513627e-01 7.27018952e-01 8.75521362e-01
-1.52191464e-02 5.15851140e-01 -1.20715845e+00 9.38756526e-01
3.27988088e-01 -8.61295044e-01 8.68279934e-02 4.70750093e-01
-1.96487710e-01 -5.82481802e-01 4.39148009e-01 4.10436541e-02
-2.65619487e-01 -1.08274519e+00 5.76803744e-01 1.81663081e-01
7.96056628e-01 -1.04995608e+00 4.90073323e-01 2.49699682e-01
-9.14410174e-01 -2.06084475e-01 -1.96822688e-01 3.33562821e-01
-3.52620371e-02 1.67851210e-01 -1.27361691e+00 2.82961249e-01
2.70347029e-01 2.69857273e-02 -2.67676562e-01 6.68469906e-01
-1.82676271e-01 6.56128943e-01 -3.42289031e-01 -1.44879654e-01
-1.31766170e-01 -7.72060156e-02 1.04112470e+00 9.40573812e-01
4.52754408e-01 6.24118634e-02 -3.69853348e-01 8.96819353e-01
-3.77426505e-01 4.21344638e-01 -1.06389558e+00 -3.69889855e-01
5.50390661e-01 9.29125845e-01 -1.35105729e-01 4.46398519e-02
1.45402059e-01 1.31047666e+00 -5.97463623e-02 1.80363938e-01
-4.21540171e-01 -6.39871180e-01 1.32967925e+00 3.00113350e-01
2.64416784e-01 3.84464748e-02 2.84362845e-02 -8.48495841e-01
-8.21693912e-02 -1.00828564e+00 2.31474355e-01 -8.65030214e-02
-8.43664348e-01 7.23660171e-01 -3.56856465e-01 -1.09261978e+00
-4.15767550e-01 -4.40046281e-01 -4.63486671e-01 1.17420483e+00
-1.22290242e+00 -1.07592857e+00 5.14940023e-01 8.58551323e-01
1.86927617e-01 -6.54585719e-01 1.14986289e+00 2.95221150e-01
-5.26528120e-01 9.53881621e-01 -7.99406543e-02 5.42581499e-01
5.22065461e-01 -9.82195079e-01 2.05415264e-01 7.13183343e-01
3.00328612e-01 8.57232451e-01 9.39752400e-01 -3.35203260e-01
-1.40169203e+00 -8.41317713e-01 1.17077041e+00 -2.44931489e-01
4.00954247e-01 -7.68679619e-01 -1.02995646e+00 7.23829508e-01
1.19265974e-01 2.99518518e-02 9.39558923e-01 -1.64039180e-01
-4.79718834e-01 -3.56782645e-01 -1.72816694e+00 4.03873712e-01
1.03846632e-01 -9.65356886e-01 -5.15253067e-01 2.26943586e-02
8.89544487e-01 1.04605556e-01 -7.95260489e-01 -1.92493021e-01
5.63749969e-01 -7.24268079e-01 1.07036483e+00 -8.07776392e-01
-1.79189160e-01 -2.02849820e-01 -4.24118042e-01 -1.22760773e+00
1.08756632e-01 -7.75090694e-01 -4.50484306e-02 1.89909375e+00
2.38166064e-01 -1.00040936e+00 8.19280565e-01 9.80012774e-01
5.19387305e-01 -1.62435487e-01 -1.47414994e+00 -5.24490356e-01
-1.37861401e-01 -4.21049386e-01 1.16581953e+00 1.02132583e+00
-1.29240513e-01 2.52440900e-01 -8.27605367e-01 7.27332234e-01
8.28688920e-01 -2.42487237e-01 6.73738420e-01 -1.38154006e+00
-1.64112970e-01 -2.62732923e-01 -9.76539850e-01 -5.77348888e-01
5.26623011e-01 -9.45095062e-01 -1.71341836e-01 -6.27613485e-01
-2.70138830e-01 -2.65204191e-01 -3.62734526e-01 1.50963902e-01
1.72636852e-01 -1.55799910e-02 4.03468050e-02 -3.59594673e-01
4.33180004e-01 7.58046567e-01 6.39762759e-01 -3.25845480e-01
-1.91391602e-01 7.63873041e-01 -7.55824745e-01 5.04994333e-01
7.87600219e-01 -6.76784754e-01 -4.29742038e-01 6.67484775e-02
-3.31106454e-01 2.49366343e-01 4.08622772e-01 -6.74775004e-01
1.73178509e-01 1.87922910e-01 3.25166397e-02 9.77298804e-03
6.76246405e-01 -1.32141387e+00 -7.48647600e-02 5.70689261e-01
-4.67115819e-01 -2.01301560e-01 6.00975826e-02 7.04070330e-01
-3.10682476e-01 -3.33873838e-01 7.84885287e-01 2.56754190e-01
-2.71612316e-01 2.79889524e-01 -4.80387449e-01 -4.30993885e-01
9.47999597e-01 6.59205019e-02 3.44696455e-02 -6.83803201e-01
-7.62467027e-01 -2.54881442e-01 3.12974691e-01 5.96329749e-01
7.79316664e-01 -1.31716549e+00 -6.96037591e-01 7.53965557e-01
1.62019432e-01 -6.60589814e-01 1.62548304e-01 3.45700800e-01
-1.59994707e-01 4.21685696e-01 -1.92394167e-01 -1.31183103e-01
-1.87883174e+00 8.71646523e-01 3.44342828e-01 4.01511669e-01
-3.85887176e-01 7.47307360e-01 5.86269647e-02 -4.28222090e-01
3.71072173e-01 -1.84864327e-01 -4.73787546e-01 1.67011470e-01
7.39110231e-01 2.19298273e-01 4.09420282e-02 -1.14976513e+00
-6.10868514e-01 3.28809470e-01 -2.11011004e-02 -3.28759193e-01
9.09554362e-01 -2.78160453e-01 -1.05682239e-01 2.98086971e-01
1.69064748e+00 6.34776890e-01 -6.48427427e-01 -1.50563449e-01
-1.31577969e-01 -6.42203987e-01 2.34456956e-01 -4.03333485e-01
-1.15600705e+00 9.89942253e-01 8.84615660e-01 5.03651917e-01
8.86541069e-01 -1.02959521e-01 6.17870450e-01 3.36111411e-02
2.80731678e-01 -5.57109356e-01 -7.80589163e-01 -4.22280142e-03
8.61938775e-01 -1.07959425e+00 -2.57506728e-01 -2.83354133e-01
-8.02668869e-01 8.50126088e-01 -9.13829654e-02 3.05067062e-01
1.03198028e+00 -2.66236626e-02 4.00760859e-01 1.91882476e-01
-3.29672158e-01 3.00003171e-01 4.03544068e-01 8.47150803e-01
1.44725591e-01 3.87975931e-01 -1.89768225e-01 7.34891534e-01
-5.18072844e-01 -5.17160058e-01 5.59172392e-01 5.80728650e-01
-1.75066918e-01 -1.34639513e+00 -7.16938555e-01 5.74406460e-02
-8.29577208e-01 -1.61471404e-02 -6.65909469e-01 2.89717317e-01
-9.08230022e-02 8.76967311e-01 5.50958291e-02 -4.49387908e-01
2.45172247e-01 4.63372469e-01 2.03214046e-02 -4.36463863e-01
-6.39015436e-01 -4.24809188e-01 1.66890863e-02 -1.97829738e-01
-2.69388109e-02 -7.69617200e-01 -9.03872192e-01 -3.76544684e-01
-6.09005950e-02 4.39092427e-01 1.18493211e+00 6.89734757e-01
2.17148095e-01 1.49560586e-01 1.06206262e+00 -5.11068523e-01
-8.03676844e-01 -6.92182243e-01 -1.02044475e+00 3.37786168e-01
1.01053727e+00 -3.77358794e-01 -6.03629112e-01 -1.97206095e-01] | [14.023775100708008, 5.879519462585449] |
e28026e0-52e2-4e75-8ee6-5c2cce1b05a2 | desta-a-framework-for-safe-reinforcement-1 | 2110.14468 | null | https://arxiv.org/abs/2110.14468v3 | https://arxiv.org/pdf/2110.14468v3.pdf | DESTA: A Framework for Safe Reinforcement Learning with Markov Games of Intervention | Reinforcement learning (RL) involves performing exploratory actions in an unknown system. This can place a learning agent in dangerous and potentially catastrophic system states. Current approaches for tackling safe learning in RL simultaneously trade-off safe exploration and task fulfillment. In this paper, we introduce a new generation of RL solvers that learn to minimise safety violations while maximising the task reward to the extent that can be tolerated by the safe policy. Our approach introduces a novel two-player framework for safe RL called Distributive Exploration Safety Training Algorithm (DESTA). The core of DESTA is a game between two adaptive agents: Safety Agent that is delegated the task of minimising safety violations and Task Agent whose goal is to maximise the environment reward. Specifically, Safety Agent can selectively take control of the system at any given point to prevent safety violations while Task Agent is free to execute its policy at any other states. This framework enables Safety Agent to learn to take actions at certain states that minimise future safety violations, both during training and testing time, while Task Agent performs actions that maximise the task performance everywhere else. Theoretically, we prove that DESTA converges to stable points enabling safety violations of pretrained policies to be minimised. Empirically, we show DESTA's ability to augment the safety of existing policies and secondly, construct safe RL policies when the Task Agent and Safety Agent are trained concurrently. We demonstrate DESTA's superior performance against leading RL methods in Lunar Lander and Frozen Lake from OpenAI gym. | ['Changmin Yu', 'Xiuling Zhang', 'Yaqi Sun', 'Usman Islam', 'Jun Wang', 'Ziyan Wang', 'Yaodong Yang', 'Aivar Sootla', 'Joel Jennings', 'David Mguni'] | 2021-10-27 | desta-a-framework-for-safe-reinforcement | https://openreview.net/forum?id=ht61oVsaya | https://openreview.net/pdf?id=ht61oVsaya | null | ['safe-exploration'] | ['robots'] | [ 2.64315784e-01 8.34923446e-01 -2.33434498e-01 1.26781896e-01
-7.77944386e-01 -7.05753505e-01 5.46390891e-01 -9.59305018e-02
-5.96368968e-01 1.28327596e+00 -1.50249854e-01 -4.93566602e-01
-4.47791159e-01 -7.10833013e-01 -7.87688375e-01 -1.07806265e+00
-6.58915579e-01 3.97886425e-01 7.74004590e-03 -2.97964990e-01
-8.66589174e-02 3.67922992e-01 -1.66234553e+00 -2.82488972e-01
7.98083246e-01 6.45942032e-01 -4.73015346e-02 7.94192851e-01
7.82886922e-01 1.30433536e+00 -4.34862196e-01 3.83939445e-01
5.45803905e-01 -4.38244611e-01 -1.07203805e+00 -1.99244678e-01
-2.40678415e-01 -2.40128875e-01 9.14850235e-02 1.03305614e+00
3.60285461e-01 5.63196242e-01 2.61620760e-01 -1.74209559e+00
1.78908750e-01 6.76705182e-01 -2.32264817e-01 -2.20185935e-01
3.37415397e-01 9.75909472e-01 7.08344042e-01 1.80570796e-01
2.76153982e-01 1.13477921e+00 1.70334712e-01 1.08895350e+00
-1.39603484e+00 -6.88025594e-01 5.06297171e-01 -2.76858091e-01
-1.09212136e+00 -3.30803066e-01 3.41871768e-01 -1.94820628e-01
1.30934441e+00 4.88776445e-01 8.06775272e-01 1.26848567e+00
8.07184756e-01 5.79310656e-01 1.37243271e+00 -2.92373925e-01
8.64559352e-01 -6.44803420e-02 -3.73004675e-01 5.26949108e-01
4.52196114e-02 1.04179454e+00 -3.73577148e-01 -3.37070227e-01
4.64583158e-01 -6.13582551e-01 -6.85082674e-02 -5.56910455e-01
-1.10659122e+00 7.42570996e-01 1.09890483e-01 -1.52511913e-02
-5.04622877e-01 6.60704672e-01 6.38493717e-01 6.63487494e-01
-1.16928048e-01 1.09390044e+00 -5.94888866e-01 -2.52765507e-01
-4.25942570e-01 6.44325137e-01 8.31905901e-01 6.66783094e-01
3.12466174e-01 5.33531606e-01 -2.21498296e-01 -2.04731920e-03
3.50528121e-01 3.95018190e-01 1.81922317e-01 -1.36342883e+00
6.33260161e-02 4.25062120e-01 5.15642643e-01 -1.58727497e-01
-5.14801800e-01 -3.81312758e-01 -3.66011679e-01 1.25702417e+00
2.58100718e-01 -6.22067034e-01 -6.52848721e-01 2.12924886e+00
6.33179426e-01 2.90173084e-01 6.82969809e-01 8.23297381e-01
-1.64089903e-01 7.65491664e-01 3.93749952e-01 -8.19960415e-01
9.76260364e-01 -5.05303741e-01 -3.57978314e-01 -4.48646039e-01
8.84441793e-01 -7.97263347e-03 1.02612650e+00 8.14171314e-01
-1.39757836e+00 -2.63685673e-01 -1.19841921e+00 7.55015433e-01
2.34305009e-01 -6.47763252e-01 3.67310464e-01 4.81351197e-01
-9.79424536e-01 6.64247632e-01 -1.00907195e+00 8.33915323e-02
1.36243120e-01 8.26099038e-01 -1.03686722e-02 5.89648902e-01
-1.31215179e+00 1.23378122e+00 8.92808795e-01 2.69214250e-03
-1.96875679e+00 -5.91734231e-01 -9.88637626e-01 -2.30984036e-02
1.04348767e+00 -5.00307620e-01 1.79419768e+00 -9.23669219e-01
-1.77427089e+00 4.27176267e-01 6.84927404e-01 -8.31283033e-01
7.27623582e-01 -3.23793769e-01 -1.45160198e-01 -3.52006286e-01
1.45171240e-01 5.80054104e-01 9.17937100e-01 -1.28254938e+00
-8.32709491e-01 -3.37187648e-02 4.98491049e-01 5.73686898e-01
7.62532279e-02 -2.06826732e-01 4.71674830e-01 -1.44708589e-01
-6.57081306e-01 -1.25996614e+00 -6.80543780e-01 -4.34003472e-01
-2.79347897e-01 -3.95350546e-01 5.42177677e-01 1.93319619e-02
1.02304387e+00 -2.00483298e+00 3.72963399e-01 3.73682082e-01
-1.82649896e-01 1.33623645e-01 -2.09868386e-01 3.19798738e-01
-1.09795347e-01 -2.02162385e-01 -4.00616884e-01 1.44089997e-01
3.12980056e-01 7.04676151e-01 -5.90051353e-01 5.90417206e-01
2.39881743e-02 7.77443469e-01 -1.23530734e+00 -3.38316202e-01
8.55024233e-02 5.87216429e-02 -5.05127132e-01 5.46800315e-01
-7.66559720e-01 7.63905406e-01 -6.16624057e-01 1.55906528e-01
1.01665862e-01 5.23879230e-01 5.28703511e-01 7.60089576e-01
-3.80686283e-01 3.55743825e-01 -1.30351818e+00 1.10621417e+00
-4.19567555e-01 -5.75516485e-02 2.71020502e-01 -7.37843990e-01
7.43815064e-01 5.38772762e-01 5.77378273e-01 -9.60847378e-01
2.57283807e-01 1.17572963e-01 1.84995532e-01 -2.77532399e-01
1.44862607e-01 -5.63445449e-01 -5.47099411e-01 7.96386242e-01
-3.71686488e-01 -3.85102391e-01 -6.79962663e-03 -3.41732949e-02
1.30500495e+00 8.49038780e-01 4.60603625e-01 -7.39378452e-01
7.22881615e-01 8.64272490e-02 7.77195096e-01 8.94653499e-01
-4.60091501e-01 -5.72887719e-01 9.87093091e-01 -5.85786521e-01
-8.07882190e-01 -9.63158846e-01 2.33545899e-01 1.40969884e+00
8.34508911e-02 -1.44723654e-01 -6.49299741e-01 -1.11180782e+00
-6.93008676e-02 1.38680208e+00 -7.28477657e-01 -7.56426036e-01
-1.05662334e+00 -2.29073122e-01 6.17626965e-01 3.27502042e-01
3.39584500e-01 -1.76049781e+00 -1.80874538e+00 1.43091127e-01
1.68605551e-01 -3.28681946e-01 -5.30980468e-01 8.59047055e-01
-6.19803905e-01 -1.07768357e+00 4.57053632e-02 -6.00546658e-01
5.22075236e-01 -4.52296615e-01 8.27894032e-01 9.08343941e-02
-4.38821353e-02 3.99296135e-01 2.25574966e-03 -5.56143343e-01
-7.28872120e-01 -8.82662162e-02 5.55212796e-01 -5.56071401e-01
-1.90625042e-01 -4.19498235e-01 -2.50501484e-01 3.00697535e-01
-7.87594259e-01 4.17851247e-02 1.86865434e-01 7.16050208e-01
5.54447949e-01 4.61981326e-01 6.56616688e-01 -5.37192643e-01
7.52558649e-01 -4.03425008e-01 -1.00554967e+00 2.95720488e-01
-9.36903894e-01 6.32531106e-01 7.59774387e-01 -6.46239996e-01
-1.05817235e+00 4.14945960e-01 1.06730320e-01 -1.29717708e-01
-5.60342781e-02 -6.15835097e-03 -3.18391860e-01 1.19423844e-01
8.77109051e-01 2.93297827e-01 2.15905845e-01 2.03392848e-01
1.76497132e-01 1.08784288e-01 4.68167096e-01 -1.25913942e+00
9.31992054e-01 -7.97438398e-02 2.39197701e-01 -2.01911926e-01
-7.19433010e-01 1.33140445e-01 -2.73323864e-01 -4.52237487e-01
5.92534304e-01 -7.29050457e-01 -1.53092039e+00 -3.99679840e-02
-4.36333597e-01 -1.09166455e+00 -9.62649107e-01 4.39286307e-02
-1.38159430e+00 -7.87704363e-02 1.06460989e-01 -1.45287716e+00
-3.78035635e-01 -1.08012652e+00 8.97618473e-01 2.59144127e-01
-5.10906219e-01 -7.31988072e-01 4.57887441e-01 -1.53253853e-01
7.45114535e-02 7.14862645e-01 7.06662655e-01 -3.80762905e-01
-3.63706291e-01 5.28823555e-01 7.52596200e-01 1.48362309e-01
-9.22619551e-02 -4.13135469e-01 -7.89976001e-01 -8.69766057e-01
3.70733052e-01 -8.08505118e-01 2.57650048e-01 2.46595383e-01
7.04025090e-01 -9.79777098e-01 -1.75264910e-01 3.85590106e-01
1.18556499e+00 7.04140842e-01 5.20117462e-01 9.57098484e-01
6.63519353e-02 7.47004867e-01 1.18179524e+00 5.34391880e-01
3.12138647e-02 5.46522021e-01 1.00647879e+00 3.79816532e-01
6.18116260e-01 -2.33206600e-01 1.02722073e+00 -1.77630141e-01
-5.46041615e-02 1.70346782e-01 -8.48909795e-01 4.66289312e-01
-2.34687352e+00 -9.62537050e-01 4.43011045e-01 2.43679500e+00
1.33143151e+00 5.13664722e-01 4.55604345e-01 8.53748098e-02
5.93457967e-02 -9.46548805e-02 -1.03748119e+00 -1.01445615e+00
4.25236851e-01 1.74700230e-01 4.44023520e-01 8.46635342e-01
-9.75357056e-01 1.00404823e+00 5.89659739e+00 5.29262662e-01
-7.15972483e-01 -1.76588729e-01 3.80293936e-01 -5.71120143e-01
-2.08801866e-01 1.75395384e-01 -6.04319751e-01 1.70168996e-01
1.33721185e+00 -4.46704775e-01 6.99895144e-01 1.09583640e+00
4.09843504e-01 -3.45968693e-01 -1.34581292e+00 1.48530230e-01
-4.94075269e-01 -6.84957683e-01 -5.81468284e-01 6.94034398e-02
6.67825282e-01 -4.34665740e-01 5.19628450e-02 8.17622960e-01
9.53327894e-01 -1.28838313e+00 9.52617466e-01 3.55267137e-01
6.52857244e-01 -1.57683897e+00 3.62087905e-01 1.02390277e+00
-8.87773991e-01 -4.98818338e-01 1.23185560e-01 -4.53330606e-01
-7.30036665e-03 -2.92833656e-01 -9.30095851e-01 2.58511126e-01
6.08607054e-01 1.37226626e-01 -4.94697355e-02 5.37301540e-01
-7.64735281e-01 2.57849276e-01 -2.55173087e-01 -6.68955073e-02
7.13805795e-01 -1.39487222e-01 6.42645776e-01 6.21069968e-01
-5.96595369e-02 5.91050759e-02 5.57462633e-01 7.28425801e-01
6.63074374e-01 -3.26160878e-01 -8.70165348e-01 2.91373491e-01
1.91768065e-01 1.04542136e+00 -4.35205817e-01 -1.09100729e-01
3.96973193e-01 5.28884411e-01 4.57239538e-01 1.81713760e-01
-1.07499754e+00 -4.60872762e-02 1.04048157e+00 -9.97077897e-02
-2.08362401e-01 -1.56124488e-01 -2.21938536e-01 -4.88058895e-01
-3.31452042e-01 -1.30122542e+00 6.69500232e-01 -5.36598980e-01
-7.03698575e-01 4.20006365e-01 2.76565224e-01 -9.75684702e-01
-7.01023102e-01 -1.74156621e-01 -6.23878002e-01 6.78539157e-01
-1.32656062e+00 -7.20054746e-01 2.94982284e-01 7.57530093e-01
6.29910827e-01 -1.35354012e-01 8.58880699e-01 -5.79040229e-01
-6.57662570e-01 2.76000410e-01 -3.32926661e-01 -6.40939236e-01
4.23705667e-01 -1.49594367e+00 1.11395985e-01 7.61511445e-01
-3.77057225e-01 4.02416378e-01 9.66666579e-01 -9.69692707e-01
-1.37268579e+00 -1.20233548e+00 3.71153057e-01 -3.83891910e-01
5.84232211e-01 -2.09758401e-01 -6.97504342e-01 9.39389467e-01
2.05370143e-01 -2.28641510e-01 6.68465570e-02 -1.42895073e-01
-1.18715391e-02 4.94148806e-02 -1.17348242e+00 9.87674832e-01
8.01615417e-01 -1.63455516e-01 -6.62656069e-01 2.80058086e-01
7.27786899e-01 -5.61399937e-01 -6.22947216e-01 4.39647377e-01
3.16879839e-01 -7.59000361e-01 7.00347424e-01 -9.68518734e-01
2.59541031e-02 -5.48559964e-01 2.89630383e-01 -1.43004704e+00
-3.19008321e-01 -1.38462889e+00 -5.84499776e-01 5.01746595e-01
1.81240916e-01 -5.77854335e-01 6.73147917e-01 8.40857804e-01
-2.45390221e-01 -8.53115499e-01 -1.43054569e+00 -1.13448656e+00
5.35267293e-01 -3.87188762e-01 4.95257884e-01 5.97152531e-01
5.33999860e-01 1.41587913e-01 -6.80611432e-01 2.75343627e-01
7.86974251e-01 -1.51084870e-01 4.65514779e-01 -6.46570146e-01
-4.54293340e-01 -2.12615117e-01 3.43115181e-01 -2.69204050e-01
6.27364814e-01 -6.23823524e-01 7.26794302e-01 -1.10136700e+00
-6.69000670e-02 -5.07142425e-01 -3.48099411e-01 1.09996891e+00
-1.78654008e-02 -4.12441373e-01 1.70008257e-01 -1.95520073e-02
-8.64821613e-01 5.60646474e-01 1.44726360e+00 -1.43023450e-02
-6.77550316e-01 1.03030339e-01 -6.75383508e-01 6.53303921e-01
1.10037446e+00 -5.20529270e-01 -8.43550861e-01 1.63256943e-01
6.05056107e-01 2.72387713e-01 3.12849939e-01 -9.16056812e-01
5.09866960e-02 -8.87189865e-01 -9.01628807e-02 1.91748291e-02
2.78598070e-03 -1.01724803e+00 3.79333109e-01 1.33483744e+00
-7.08449244e-01 -1.39688188e-03 4.10836786e-01 4.15883690e-01
4.34861422e-01 -1.79394767e-01 1.06381989e+00 -1.22446209e-01
-6.36595249e-01 -1.03817724e-01 -8.37708712e-01 8.49019364e-02
1.97464991e+00 4.42918465e-02 -1.53059661e-01 -1.84167281e-01
-7.90711582e-01 1.08240569e+00 2.51913160e-01 3.34941000e-01
5.06086230e-01 -1.08729088e+00 -5.04126608e-01 2.46604547e-01
-1.14870526e-01 1.19677432e-01 4.83582132e-02 7.06232786e-01
-1.31574273e-01 2.08193392e-01 -5.32775164e-01 -2.13456631e-01
-1.22478461e+00 8.49413514e-01 7.92180598e-01 -6.01406991e-01
-8.60338032e-01 6.68139935e-01 2.67094761e-01 -4.81223971e-01
4.80248094e-01 -1.86966226e-01 -2.08841830e-01 -3.10117483e-01
5.69805264e-01 4.57486153e-01 -2.95647711e-01 -1.63904697e-01
-4.78091180e-01 1.89278021e-01 6.69820532e-02 -3.28080744e-01
1.25230300e+00 1.50753468e-01 8.64422843e-02 1.96775213e-01
3.63547713e-01 -4.84156996e-01 -1.97946048e+00 2.53960699e-01
1.29956290e-01 -7.20571578e-02 3.24877687e-02 -1.08634424e+00
-4.21592057e-01 2.72738844e-01 4.47319686e-01 3.23089361e-01
1.28015471e+00 -3.18770111e-01 3.93073738e-01 4.52808768e-01
7.81959116e-01 -1.48104632e+00 1.81604028e-01 5.90117693e-01
9.58726168e-01 -9.01222646e-01 8.20740238e-02 2.67347723e-01
-1.20134699e+00 7.86691546e-01 1.17103958e+00 -2.84716517e-01
-2.05087848e-02 6.57158494e-01 -2.43748009e-01 -2.97208969e-02
-1.29882085e+00 -1.57067969e-01 -1.18118346e-01 8.15268338e-01
-2.86463380e-01 1.47596613e-01 -4.33953814e-02 2.46962801e-01
-2.15418860e-01 -1.90728739e-01 5.42376697e-01 1.32404828e+00
-8.84286284e-01 -1.14316213e+00 -5.81384480e-01 6.40122592e-03
-1.01050943e-01 3.02776486e-01 -2.22261384e-01 1.00589287e+00
2.84493834e-01 7.24738896e-01 -1.33688450e-01 -5.05354442e-02
3.70054811e-01 1.51608542e-01 5.82154453e-01 -7.05694795e-01
-9.50281262e-01 1.97787061e-01 3.29314470e-01 -1.11263621e+00
-8.38465896e-03 -8.51135254e-01 -1.73570573e+00 -2.64700148e-02
2.43586712e-02 4.18016613e-01 8.35001543e-02 9.71252859e-01
-1.75673231e-01 8.41387570e-01 7.74119854e-01 -4.40864354e-01
-1.25805247e+00 -3.16625535e-01 -5.72987199e-01 -1.14543550e-01
7.54973054e-01 -5.57695210e-01 -3.01863849e-01 -1.26398057e-01] | [4.510750770568848, 2.092385768890381] |
934f78ff-4dc4-435e-be35-cddab87c11c9 | multilingual-corpora-with-coreferential | null | null | https://aclanthology.org/L14-1701 | https://aclanthology.org/L14-1701.pdf | Multilingual corpora with coreferential annotation of person entities | This paper presents three corpora with coreferential annotation of person entities for Portuguese, Galician and Spanish. They contain coreference links between several types of pronouns (including elliptical, possessive, indefinite, demonstrative, relative and personal clitic and non-clitic pronouns) and nominal phrases (including proper nouns). Some statistics have been computed, showing distributional aspects of coreference both in journalistic and in encyclopedic texts. Furthermore, the paper shows the importance of coreference resolution for a task such as Information Extraction, by evaluating the output of an Open Information Extraction system on the annotated corpora. The corpora are freely distributed in two formats: (i) the SemEval-2010 and (ii) the brat rapid annotation tool, so they can be enlarged and improved collaboratively. | ['Pablo Gamallo', 'Marcos Garcia'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['open-information-extraction'] | ['natural-language-processing'] | [-3.73689175e-01 4.60723013e-01 -1.71043769e-01 -1.56072333e-01
-5.24038792e-01 -1.16594076e+00 9.26568210e-01 5.91577530e-01
-1.02711940e+00 1.35760665e+00 1.16291106e+00 -1.79612413e-01
-3.68975908e-01 -5.04206717e-01 -1.08861551e-02 -4.21347708e-01
6.87783584e-02 1.17119312e+00 3.20813864e-01 -6.49866879e-01
1.86609849e-01 5.69173336e-01 -1.42500079e+00 3.22276264e-01
5.81376672e-01 1.00746555e-02 3.64671201e-01 2.91420043e-01
-1.40194014e-01 4.09602463e-01 -6.15197778e-01 -1.07286692e+00
-1.55987024e-01 2.62129635e-01 -1.46822643e+00 -7.78527617e-01
1.31946981e-01 3.77342373e-01 1.87415406e-02 1.20688534e+00
6.52848184e-01 1.57887056e-01 4.53667104e-01 -7.88578451e-01
-1.75382555e-01 1.17424583e+00 -6.28803372e-02 4.50900257e-01
1.04521370e+00 -2.79723853e-01 1.37810540e+00 -6.58339441e-01
1.41595721e+00 1.61802495e+00 5.84834039e-01 5.81949592e-01
-7.57465720e-01 -5.84568202e-01 -5.18507779e-01 2.45050564e-01
-1.08666694e+00 -4.21455860e-01 7.25905076e-02 -5.37422478e-01
1.62956190e+00 6.53734446e-01 3.94828647e-01 1.07780087e+00
-1.89573050e-01 4.40874785e-01 1.04062796e+00 -8.06502342e-01
-3.19446921e-01 3.42101246e-01 5.96121550e-01 9.03961286e-02
6.32105410e-01 1.53392345e-01 -3.35778087e-01 -4.81830120e-01
2.72327811e-01 -5.70815146e-01 -4.05561030e-01 3.03208888e-01
-1.16536784e+00 6.73245370e-01 -8.06942135e-02 1.14301181e+00
-4.25659537e-01 -5.67880869e-01 8.89751077e-01 1.73674613e-01
8.47211480e-02 6.17012918e-01 -7.80158043e-01 -3.65764707e-01
-3.89002800e-01 6.10124230e-01 1.31262755e+00 1.21898770e+00
4.01177704e-01 -9.18811619e-01 -6.59853145e-02 9.29916084e-01
4.15415347e-01 5.40458560e-01 4.03093934e-01 -9.10802484e-01
7.99613237e-01 6.49031401e-01 5.67684472e-01 -7.04100013e-01
-7.63769984e-01 2.80748069e-01 -1.38448760e-01 -3.93744558e-01
6.72923148e-01 -3.82271290e-01 7.02386945e-02 1.71407735e+00
4.42497164e-01 -4.99767959e-01 5.04983306e-01 8.43328595e-01
1.51672971e+00 3.89703661e-01 6.82327211e-01 -7.31531382e-01
2.10669327e+00 -3.53601724e-01 -1.40967572e+00 1.40354484e-01
5.50263405e-01 -1.08679056e+00 5.51861882e-01 -2.58669141e-03
-1.36502671e+00 -1.14997447e-01 -7.09155142e-01 -2.72417843e-01
-6.27545297e-01 -2.97926873e-01 5.32758057e-01 4.11270827e-01
-5.26895523e-01 5.23919106e-01 -7.70064175e-01 -9.36031282e-01
-2.28053331e-01 5.61487019e-01 -7.53156900e-01 4.52088326e-01
-1.55668390e+00 1.30576074e+00 1.23799872e+00 -3.64753485e-01
2.00502068e-01 -3.59885216e-01 -9.45341349e-01 -2.00620055e-01
1.38929591e-01 -4.51819211e-01 1.11668658e+00 -7.71307290e-01
-8.21929753e-01 1.74506247e+00 4.94324714e-02 -3.16541970e-01
2.74670333e-01 -4.63375121e-01 -9.97526407e-01 8.85808170e-02
7.01279163e-01 2.94556171e-01 -1.99430659e-01 -8.89610350e-01
-1.22123969e+00 -6.90602601e-01 2.61954159e-01 4.35066909e-01
-1.17188632e-01 1.31772280e+00 -2.81194389e-01 -5.17286718e-01
-3.76948446e-01 -7.73732066e-01 2.84456104e-01 -1.01542592e+00
-5.00929773e-01 -9.92433131e-01 5.73938847e-01 -1.11510015e+00
1.58744490e+00 -1.87029564e+00 1.74988076e-01 1.09766610e-02
-1.73043281e-01 3.67334962e-01 3.59947056e-01 8.03468823e-01
-3.28287065e-01 6.15851358e-02 3.91313434e-02 3.15042853e-01
2.88719237e-01 7.07474411e-01 3.68041843e-02 4.67456758e-01
-1.50537983e-01 6.55937672e-01 -1.22662771e+00 -8.69790912e-01
-7.71437064e-02 1.82800576e-01 -2.35532224e-01 -1.13488004e-01
1.52363339e-02 1.61367133e-01 -4.30903584e-01 3.60611379e-01
4.05264646e-01 3.92166644e-01 8.48070383e-01 -4.40629125e-01
-8.71904314e-01 9.41510379e-01 -1.34632683e+00 1.34901536e+00
-8.20803270e-02 5.03679574e-01 3.64599913e-01 -4.62370425e-01
5.49399018e-01 1.04485786e+00 1.73648998e-01 -4.07350242e-01
3.47752988e-01 5.34061015e-01 -3.00980136e-02 -7.41972148e-01
8.93423438e-01 -1.47826508e-01 -4.19634044e-01 3.47469717e-01
4.16442186e-01 5.22355378e-01 9.76611495e-01 1.66290566e-01
6.54368758e-01 4.03694838e-01 1.02302206e+00 -8.65140021e-01
9.94861424e-01 3.39638829e-01 9.57157493e-01 -6.40638471e-02
1.06984593e-01 3.97804938e-02 4.42871273e-01 -2.55225569e-01
-1.22156370e+00 -8.58211100e-01 -9.42693472e-01 1.10802615e+00
-1.15282305e-01 -6.50573552e-01 -7.32422054e-01 -6.02859676e-01
-7.71247000e-02 9.63864028e-01 -3.22434902e-01 7.10349977e-01
-1.17837656e+00 -5.63666523e-01 7.53008723e-01 4.73309010e-01
4.89697903e-02 -1.59559250e+00 -8.09082568e-01 2.46477112e-01
-6.30302727e-01 -1.10350192e+00 -3.81453037e-01 2.21046820e-01
-4.00113642e-01 -1.40299392e+00 -4.55649085e-02 -1.09591758e+00
2.82185972e-01 -7.20587552e-01 1.46243858e+00 3.19778621e-01
3.31283584e-02 3.36977184e-01 -4.50205296e-01 -5.06527126e-01
-5.20140827e-01 3.94898281e-02 2.61880696e-01 -7.70486891e-01
1.19039226e+00 -3.98579925e-01 2.53101327e-02 -5.84151521e-02
-5.68465829e-01 -4.73669976e-01 1.01674736e-01 8.90805781e-01
8.98540318e-02 -6.03282213e-01 3.00342679e-01 -1.42345631e+00
7.60617852e-01 -5.16166270e-01 -4.14253086e-01 2.57982165e-02
-3.49366814e-01 -5.04435450e-02 1.06758766e-01 3.18593271e-02
-1.49728417e+00 -3.87248278e-01 -5.53952694e-01 4.50465888e-01
-4.30371642e-01 5.05529463e-01 -6.48414075e-01 6.26963377e-01
6.68893337e-01 -4.67397094e-01 -3.49471062e-01 -7.35166073e-01
2.11551085e-01 8.80470037e-01 1.11184764e+00 -8.98668528e-01
5.46044111e-01 1.80292577e-01 -3.02717447e-01 -8.67673397e-01
-7.14205623e-01 -1.07500577e+00 -1.30702603e+00 2.27340939e-03
1.15082288e+00 -8.64515007e-01 -9.92773652e-01 -1.13234803e-01
-1.51727903e+00 2.60610223e-01 -5.13177097e-01 5.69854140e-01
-2.25092083e-01 4.79097992e-01 -8.58345568e-01 -6.45226777e-01
-7.04496145e-01 -4.89556789e-01 6.36964381e-01 4.88145590e-01
-1.02782953e+00 -1.42606699e+00 4.36480999e-01 2.87184596e-01
-2.25257412e-01 3.41482669e-01 8.89668941e-01 -1.47397947e+00
4.06092793e-01 -7.19342008e-02 -7.28017539e-02 -2.09447339e-01
-2.18223155e-01 -8.65405574e-02 -5.77614546e-01 -1.83482900e-01
-4.11394417e-01 -2.35211805e-01 2.08936661e-01 -3.05807441e-01
-2.87977189e-01 -7.62739420e-01 -6.67606533e-01 2.58522958e-01
1.45930290e+00 8.76286849e-02 8.87106419e-01 8.74960065e-01
1.95043832e-01 8.49648058e-01 8.78605485e-01 2.61157513e-01
5.11390150e-01 8.12634110e-01 -6.59976527e-02 6.90873921e-01
-1.86230496e-01 1.68115608e-02 3.02737534e-01 9.60228562e-01
-7.88441718e-01 1.83322743e-01 -9.32328105e-01 1.04535472e+00
-1.83125484e+00 -1.46130943e+00 -8.19951475e-01 2.05670571e+00
1.35168850e+00 -6.14068918e-02 1.08085051e-01 -7.66551262e-03
8.54217172e-01 -1.39420703e-01 4.79731381e-01 -5.68993926e-01
-6.13773167e-01 5.17759144e-01 3.93444002e-01 7.71282136e-01
-1.22168326e+00 1.07472610e+00 5.66419363e+00 4.23628420e-01
-3.50272208e-01 3.59456480e-01 -5.40932596e-01 3.33985806e-01
-3.31510931e-01 1.59709603e-01 -1.28572512e+00 3.95003259e-01
1.03041458e+00 -4.72166806e-01 5.62161170e-02 4.43999588e-01
-3.71868052e-02 -1.56629190e-01 -1.06663775e+00 6.73687041e-01
-5.40039651e-02 -1.20773435e+00 -3.05235207e-01 6.16038069e-02
5.57978511e-01 2.70052254e-01 -9.95358944e-01 3.29502672e-01
4.39383030e-01 -5.83271444e-01 7.87262619e-01 4.40340459e-01
5.77548563e-01 -6.94816649e-01 1.33415604e+00 4.68070731e-02
-1.09963262e+00 -3.88272516e-02 -3.45159501e-01 1.30500972e-01
4.54052180e-01 -4.24088910e-02 -3.88159424e-01 8.70112121e-01
7.86516786e-01 4.59964454e-01 -3.74246687e-01 9.93059397e-01
-6.27335846e-01 3.87973398e-01 -3.82972300e-01 -2.42660850e-01
6.90131187e-02 -4.18304116e-01 1.23920691e+00 1.89648950e+00
-4.80447002e-02 3.93061697e-01 -2.17938740e-02 5.76982439e-01
2.18863949e-01 5.41477978e-01 -2.94890732e-01 1.82233632e-01
9.92681682e-01 1.62136471e+00 -4.34002876e-01 -1.83644220e-01
-5.52061021e-01 6.16154611e-01 2.67264843e-01 -4.46021967e-02
-1.51555017e-01 -5.57794094e-01 4.90312010e-01 1.71023577e-01
1.42268538e-01 2.06889719e-01 2.29477301e-01 -7.88550377e-01
-3.14624727e-01 -7.49586701e-01 1.24433219e+00 -5.26171565e-01
-1.29713714e+00 4.75339025e-01 4.95180249e-01 -6.69607341e-01
-7.51271665e-01 -7.98749983e-01 -5.68954289e-01 1.00345325e+00
-9.87955332e-01 -1.38943481e+00 2.22792625e-01 7.59723604e-01
-1.16252162e-01 -3.43917876e-01 1.27981138e+00 7.01964974e-01
-2.83179432e-01 3.28119338e-01 -1.38965443e-01 6.45482361e-01
8.81877720e-01 -1.57895362e+00 -1.35367766e-01 4.77160931e-01
-2.81816244e-01 8.34547400e-01 1.17086089e+00 -7.42747843e-01
-8.25801253e-01 -6.17404878e-01 2.16191936e+00 -6.87945306e-01
9.97089446e-01 1.64131343e-01 -6.82485759e-01 1.10686719e+00
9.34408486e-01 -4.95757490e-01 9.14333463e-01 4.94735390e-01
-2.71933246e-02 4.30197507e-01 -1.38157773e+00 4.42173719e-01
1.06315565e+00 -3.01248550e-01 -1.72243023e+00 4.46364999e-01
5.71759403e-01 -5.38728952e-01 -1.45675254e+00 2.30079234e-01
4.23167109e-01 -6.93648577e-01 7.99024820e-01 -8.68433356e-01
-1.47433132e-02 -1.58588365e-01 -8.18611756e-02 -9.37812269e-01
-3.40193182e-01 -7.06149399e-01 1.93268657e-01 1.87430751e+00
5.63439965e-01 -6.05394602e-01 8.48786756e-02 5.06689847e-01
-3.85992408e-01 3.84508550e-01 -1.14527738e+00 -3.66359949e-01
1.89805165e-01 -1.18087351e-01 4.65646505e-01 1.33932757e+00
9.76474762e-01 9.44664955e-01 2.48406798e-01 -9.92038175e-02
5.88376760e-01 2.60455329e-02 3.15135568e-01 -1.86422288e+00
1.00197196e-01 -3.50636095e-01 -3.31097960e-01 -3.08657527e-01
5.35916030e-01 -1.13239193e+00 -4.55706328e-01 -1.38758707e+00
3.56861889e-01 -3.83002430e-01 2.02604353e-01 6.18121147e-01
-4.42869812e-02 -6.10958086e-03 4.21077162e-02 4.41621065e-01
-6.00466847e-01 -1.56938270e-01 7.94853091e-01 1.82654589e-01
-7.86948353e-02 -2.98748821e-01 -6.19527817e-01 1.15652061e+00
3.81276459e-01 -5.77821732e-01 6.51968300e-01 -1.17286304e-02
5.87401152e-01 -2.79829782e-02 1.00408338e-01 -4.76434320e-01
3.16935211e-01 -3.31617482e-02 -1.73766375e-01 -7.56046414e-01
-7.94230849e-02 -6.91359639e-01 1.91603869e-01 4.82936233e-01
6.66519552e-02 4.29362535e-01 3.19560021e-01 -7.59157687e-02
-2.57597506e-01 -9.87510145e-01 4.54699099e-01 -4.35680926e-01
-8.56966496e-01 -1.96687713e-01 -2.83110052e-01 5.86150467e-01
9.13323760e-01 6.20092861e-02 -4.60400939e-01 1.84218884e-01
-1.00960624e+00 2.98079848e-01 3.68065804e-01 3.78464580e-01
-9.25912112e-02 -1.43322396e+00 -9.55756247e-01 -3.73822987e-01
3.40687595e-02 -4.04321671e-01 -1.22702628e-01 9.09642637e-01
-5.18081844e-01 7.30790794e-01 -5.37464142e-01 4.05722260e-02
-1.69275296e+00 6.61074579e-01 2.54576188e-02 -6.20626926e-01
-5.76535583e-01 3.44323814e-01 -3.68562132e-01 -7.89346218e-01
2.03617066e-01 3.19422126e-01 -1.25296140e+00 6.76430106e-01
5.92851937e-01 4.96597201e-01 -8.23745355e-02 -1.55209410e+00
-6.64377391e-01 1.60776749e-01 1.98165819e-01 -1.51707128e-01
1.60379636e+00 -3.61386955e-01 -9.41791892e-01 4.38469917e-01
9.28996682e-01 6.80271327e-01 -1.05998717e-01 -1.84113309e-01
8.48907650e-01 3.30281854e-02 -4.50488240e-01 -7.20405161e-01
-4.34508950e-01 5.13701886e-02 2.33711526e-01 3.59177887e-01
5.49406469e-01 4.32476938e-01 4.72785443e-01 2.49944463e-01
4.33408886e-01 -1.55421293e+00 -8.50332379e-01 9.49754179e-01
1.06459355e+00 -5.62883437e-01 1.78742960e-01 -6.51185215e-01
-5.93542874e-01 1.29083323e+00 3.71888876e-01 1.51980715e-02
6.45030797e-01 6.74460292e-01 -5.40927537e-02 -6.25888586e-01
-7.11838841e-01 -5.77453673e-01 1.59756467e-01 7.60664940e-01
1.06959641e+00 2.40921438e-01 -1.72963059e+00 1.07614458e+00
-7.62039065e-01 -3.34825426e-01 5.30420601e-01 6.63306653e-01
-7.69462511e-02 -1.51303613e+00 -6.88955307e-01 9.47944969e-02
-1.22201312e+00 -3.26852888e-01 -3.43992859e-01 1.45210552e+00
4.72343117e-01 8.36557865e-01 4.26627874e-01 3.02821040e-01
6.11396253e-01 1.06760040e-01 6.62516236e-01 -7.05885291e-01
-1.31005323e+00 -2.50077981e-04 1.22089863e+00 -1.37572229e-01
-1.03180337e+00 -1.47272444e+00 -1.55695879e+00 -3.13344538e-01
-4.16106999e-01 7.31589973e-01 2.38405496e-01 1.18903017e+00
-3.60496938e-01 2.46102940e-02 -2.91695505e-01 -5.17939985e-01
-2.98268497e-01 -1.41484261e+00 -5.77891350e-01 1.01515424e+00
-4.69502032e-01 -6.46510363e-01 -2.39338979e-01 2.51090880e-02] | [9.339009284973145, 9.55928897857666] |
0b11a1b3-6840-4e8e-8995-ae2e64e7a185 | category-level-pose-retrieval-with | 2208.06195 | null | https://arxiv.org/abs/2208.06195v3 | https://arxiv.org/pdf/2208.06195v3.pdf | Category-Level Pose Retrieval with Contrastive Features Learnt with Occlusion Augmentation | Pose estimation is usually tackled as either a bin classification or a regression problem. In both cases, the idea is to directly predict the pose of an object. This is a non-trivial task due to appearance variations between similar poses and similarities between dissimilar poses. Instead, we follow the key idea that comparing two poses is easier than directly predicting one. Render-and-compare approaches have been employed to that end, however, they tend to be unstable, computationally expensive, and slow for real-time applications. We propose doing category-level pose estimation by learning an alignment metric in an embedding space using a contrastive loss with a dynamic margin and a continuous pose-label space. For efficient inference, we use a simple real-time image retrieval scheme with a pre-rendered and pre-embedded reference set of renderings. To achieve robustness to real-world conditions, we employ synthetic occlusions, bounding box perturbations, and appearance augmentations. Our approach achieves state-of-the-art performance on PASCAL3D and OccludedPASCAL3D and surpasses the competing methods on KITTI3D in a cross-dataset evaluation setting. The code is currently available at https://github.com/gkouros/contrastive-pose-retrieval. | ['Tinne Tuytelaars', 'Punarjay Chakravarty', 'Sushruth Nagesh', 'Cédric Picron', 'Shubham Shrivastava', 'Georgios Kouros'] | 2022-08-12 | null | null | null | null | ['pose-retrieval'] | ['computer-vision'] | [ 1.00892812e-01 -3.47581357e-01 9.53067765e-02 -4.93692309e-01
-1.24827087e+00 -7.58913159e-01 7.73601353e-01 4.77007702e-02
-5.02334893e-01 3.08969706e-01 -1.62900105e-01 8.99266526e-02
1.49251848e-01 -4.22327727e-01 -8.11875105e-01 -7.01789796e-01
9.26906317e-02 7.51807690e-01 3.94235611e-01 -7.02268332e-02
2.44173259e-01 5.53687513e-01 -1.69467437e+00 -6.43548593e-02
4.67121124e-01 1.18030894e+00 8.19857717e-02 5.87190211e-01
1.32500395e-01 2.54335284e-01 -5.49009323e-01 -5.58843434e-01
5.80404639e-01 -1.39636293e-01 -5.40164471e-01 1.32590979e-01
1.07917833e+00 -3.14377546e-01 -2.41634235e-01 8.82548153e-01
7.23337233e-01 4.26719822e-02 7.62609065e-01 -1.23093259e+00
-1.73988014e-01 -4.89317089e-01 -8.31244946e-01 -1.70810148e-01
6.00338340e-01 4.77339141e-02 9.96644497e-01 -9.96125698e-01
6.60374463e-01 1.31215119e+00 4.98255104e-01 3.96523476e-01
-1.63488352e+00 -6.55058861e-01 1.11796685e-01 4.05140743e-02
-1.54983044e+00 -4.41252887e-01 7.41880774e-01 -5.08392811e-01
5.75421929e-01 4.31803197e-01 5.45912802e-01 1.11480403e+00
1.75466865e-01 7.23857760e-01 1.29460883e+00 -2.42761165e-01
1.51921377e-01 -3.56490794e-03 -3.38920116e-01 6.86697721e-01
-9.94618982e-02 7.74751604e-02 -6.20343804e-01 -1.30643368e-01
6.58782721e-01 -6.49842322e-02 -3.50552708e-01 -1.05379212e+00
-1.04341078e+00 5.97109437e-01 4.34012026e-01 -3.03728819e-01
-2.15830669e-01 1.56460568e-01 3.11605036e-01 3.11143547e-01
7.48334527e-01 2.84473181e-01 -3.61243397e-01 -1.38284460e-01
-8.20564568e-01 5.74792802e-01 6.61145151e-01 7.11371601e-01
6.76651597e-01 -4.37718928e-01 1.36063611e-02 7.90792406e-01
3.22362751e-01 5.80647767e-01 1.96309343e-01 -9.38225031e-01
4.27189708e-01 2.73683637e-01 1.51076421e-01 -1.18327856e+00
-1.63222522e-01 -2.97471642e-01 -4.07633781e-01 5.41410565e-01
5.98502874e-01 3.02199602e-01 -9.69500363e-01 1.62565279e+00
7.90037811e-01 4.72790934e-02 -3.78731310e-01 1.18592596e+00
4.61492449e-01 5.26817977e-01 -3.54619697e-02 1.41787067e-01
1.24596560e+00 -9.55099702e-01 -3.96531105e-01 -3.70440990e-01
2.96836197e-01 -1.23701644e+00 1.15640378e+00 4.88156587e-01
-8.93219233e-01 -5.22332072e-01 -1.05399311e+00 -3.75500768e-01
-3.14321399e-01 2.34962463e-01 2.99303800e-01 5.88837147e-01
-7.44387925e-01 6.39296174e-01 -9.14707541e-01 -3.35675001e-01
1.67590037e-01 3.53615433e-01 -5.84056556e-01 -2.15134934e-01
-6.73141122e-01 9.81337249e-01 1.13410972e-01 7.97325000e-02
-7.46092439e-01 -5.98228931e-01 -8.95264506e-01 -3.90952080e-01
5.45895636e-01 -4.96472567e-01 1.14608860e+00 -7.33782291e-01
-1.59564161e+00 1.24403894e+00 3.57819162e-02 -1.19140735e-02
9.27221000e-01 -7.39693046e-01 -2.42188806e-04 -3.17518562e-02
-7.36857355e-02 6.96737468e-01 1.07282853e+00 -1.29054677e+00
-2.17311069e-01 -6.44868493e-01 1.04957521e-01 4.93124098e-01
-5.42132407e-02 -1.25015259e-01 -9.23899949e-01 -6.09151602e-01
4.46492940e-01 -1.31879008e+00 9.70985293e-02 7.50151813e-01
-3.17618221e-01 -1.61281265e-02 9.08005536e-01 -7.41172135e-01
6.93961024e-01 -2.20813513e+00 3.73032212e-01 5.82955293e-02
7.35558346e-02 2.34067500e-01 -2.07082540e-01 3.28282535e-01
1.32781984e-02 -3.21026623e-01 -6.56092167e-02 -7.77957559e-01
1.60610050e-01 3.34137082e-02 -2.44673833e-01 9.36329663e-01
3.42075765e-01 6.01972818e-01 -8.94755602e-01 -4.84546691e-01
5.57787418e-01 7.92816162e-01 -5.19351840e-01 4.32211012e-01
-2.47813210e-01 6.14580095e-01 -2.74100959e-01 8.12252939e-01
7.06270397e-01 3.28090154e-02 -1.14837755e-02 -4.16810781e-01
2.77005024e-02 3.40788901e-01 -1.45924866e+00 2.04027534e+00
-5.84798276e-01 5.90879142e-01 -1.77901797e-02 -7.03949988e-01
8.68129790e-01 -4.11338881e-02 2.42445067e-01 -5.53718567e-01
1.78585291e-01 3.83134156e-01 -2.62714416e-01 -2.02091798e-01
3.42625648e-01 3.18379477e-02 -2.97965505e-03 1.80190742e-01
-9.58068967e-02 -6.51176214e-01 3.10863066e-03 -1.62151814e-01
9.50113773e-01 7.20176101e-01 1.13473780e-01 -2.35047601e-02
3.62770975e-01 -2.46097207e-01 4.24180716e-01 3.02761108e-01
-5.62931783e-02 1.16064584e+00 3.60481679e-01 -3.73052895e-01
-9.81591105e-01 -1.32829964e+00 -2.69552827e-01 8.31385791e-01
4.36450511e-01 -4.68639791e-01 -4.77802038e-01 -6.88806832e-01
1.40113726e-01 3.22569817e-01 -5.25258660e-01 7.64834210e-02
-5.54196537e-01 -3.82850707e-01 9.32074338e-02 2.86811948e-01
2.80048072e-01 -5.76938689e-01 -6.70744777e-01 9.01898649e-03
-8.78641456e-02 -1.10632277e+00 -6.58683658e-01 1.60319787e-02
-6.68346524e-01 -1.13893175e+00 -7.82464981e-01 -5.65116107e-01
7.03095019e-01 1.60878330e-01 1.22477484e+00 -1.41034842e-01
-5.98116755e-01 3.49096984e-01 -1.64293110e-01 -2.71971047e-01
-3.54082435e-02 -5.70080392e-02 6.14238083e-02 -1.45287625e-02
1.41092658e-01 -4.88637447e-01 -8.82092476e-01 6.29454195e-01
-7.77596354e-01 7.58309588e-02 5.64341009e-01 7.24442065e-01
9.33443844e-01 -3.90321106e-01 -1.75004482e-01 -4.48211432e-01
2.28196472e-01 -6.82092272e-03 -8.80224824e-01 2.27924958e-01
-3.10200453e-01 3.38320956e-02 2.87751138e-01 -5.29613018e-01
-6.80112600e-01 4.55119133e-01 5.03911264e-03 -7.45242536e-01
-2.34881174e-02 4.65219356e-02 -2.97646344e-01 -2.30773523e-01
5.53917289e-01 -2.58402769e-02 8.86059701e-02 -5.89526296e-01
3.47603619e-01 5.28395176e-01 4.71463054e-01 -7.62443304e-01
1.08584476e+00 5.69703221e-01 1.84081554e-01 -7.43761599e-01
-8.31464231e-01 -5.21554291e-01 -7.81252444e-01 -3.55559021e-01
6.45896733e-01 -9.56908703e-01 -5.97429276e-01 5.24673343e-01
-1.13456368e+00 -3.84689212e-01 1.12024378e-02 3.94726336e-01
-5.96412182e-01 3.88554960e-01 -3.62210155e-01 -6.29465580e-01
-1.66173249e-01 -1.28467846e+00 1.77679133e+00 -6.20162748e-02
-1.99744716e-01 -5.40585876e-01 2.70818889e-01 5.64565003e-01
1.49605229e-01 5.87218165e-01 5.62461674e-01 -1.95706084e-01
-6.94443882e-01 -5.89056253e-01 -1.87222987e-01 3.48896325e-01
1.05319612e-01 1.43705621e-01 -1.07403576e+00 -4.98293668e-01
-1.67309389e-01 -7.06266522e-01 5.47043920e-01 -9.06972662e-02
1.30744445e+00 5.29914126e-02 -1.85504004e-01 6.29022002e-01
1.46197045e+00 -1.95780039e-01 5.81496358e-01 2.40511343e-01
6.21019363e-01 8.00526798e-01 1.04443073e+00 2.65158653e-01
3.30066144e-01 1.18317497e+00 4.47952151e-01 -3.60778794e-02
-1.51861548e-01 -2.20450073e-01 1.95853189e-01 5.72037339e-01
-9.76937730e-03 -2.34591857e-01 -9.32392299e-01 2.87665099e-01
-1.68733406e+00 -6.35479450e-01 9.42604244e-02 2.68571925e+00
8.40604424e-01 1.94328874e-01 1.11274548e-01 -1.74166802e-02
4.05908167e-01 2.46260241e-01 -5.04459798e-01 4.91365194e-02
1.96897149e-01 6.92067668e-02 3.65929991e-01 5.09082258e-01
-1.26428759e+00 8.04537356e-01 4.83969593e+00 8.06536853e-01
-1.38797212e+00 6.79904073e-02 6.87970996e-01 -3.27333570e-01
1.22725502e-01 -1.47391602e-01 -6.34311914e-01 3.09227288e-01
4.32650685e-01 3.58344764e-01 2.27381036e-01 8.32289457e-01
-1.16812572e-01 -3.35899949e-01 -1.24194717e+00 1.20864332e+00
1.79164857e-01 -7.11388707e-01 -2.67813683e-01 1.26385791e-02
4.62007046e-01 9.09870043e-02 1.65087208e-01 1.69501290e-01
-1.43540978e-01 -9.13623989e-01 8.34814966e-01 2.64922053e-01
9.18030739e-01 -5.15873134e-01 5.39802432e-01 1.39073834e-01
-1.21466637e+00 3.44394207e-01 -1.83510795e-01 9.21986625e-02
2.13233698e-02 3.73929888e-01 -5.54001212e-01 4.28997278e-01
7.87103415e-01 4.75920200e-01 -6.11761034e-01 1.20467341e+00
-3.75442266e-01 5.40622547e-02 -5.19716918e-01 1.25915945e-01
5.86524466e-03 -2.31665060e-01 5.37966073e-01 9.22352493e-01
1.55912191e-01 -1.93299085e-01 3.71863961e-01 4.84388918e-01
5.66961803e-02 1.69203505e-01 -4.49450612e-01 2.97755957e-01
2.71534681e-01 1.25581062e+00 -7.22516477e-01 -2.58472990e-02
-2.67543912e-01 1.25942218e+00 3.64579171e-01 1.03156529e-01
-9.60764587e-01 -2.93825954e-01 8.57229948e-01 3.79567236e-01
3.73965323e-01 -3.58448148e-01 3.87049690e-02 -1.08418441e+00
5.49675167e-01 -8.69142890e-01 1.23499051e-01 -7.50598252e-01
-1.16026092e+00 6.31206512e-01 2.30473325e-01 -1.54816568e+00
-1.63791388e-01 -6.93566322e-01 -2.37402558e-01 7.37858474e-01
-1.45194495e+00 -1.13417780e+00 -5.22567391e-01 2.88277209e-01
6.28568292e-01 3.29674482e-01 8.89364958e-01 5.21840990e-01
-3.65000635e-01 5.52461267e-01 1.68316886e-01 4.34047170e-03
1.09585023e+00 -1.32957578e+00 4.21092570e-01 5.77574492e-01
2.31064633e-01 3.48309219e-01 9.71267939e-01 -4.02561396e-01
-1.53821993e+00 -9.31910455e-01 6.17651761e-01 -7.12736845e-01
5.22251904e-01 -8.07810128e-01 -8.39999497e-01 3.95651817e-01
-1.24390878e-01 4.38007832e-01 3.76267344e-01 8.03349819e-03
-6.13656521e-01 -3.10854822e-01 -1.03204572e+00 7.02364266e-01
1.07027960e+00 -6.14130914e-01 -3.26318711e-01 5.40529847e-01
3.13632429e-01 -9.31400597e-01 -8.49896371e-01 4.29354966e-01
8.70898068e-01 -9.03984666e-01 1.13424802e+00 -9.46999118e-02
3.37357968e-01 -5.68344593e-01 -2.67792374e-01 -1.09602714e+00
1.50314942e-01 -5.94494402e-01 -4.13248837e-02 1.04932940e+00
1.15428470e-01 -5.03991485e-01 8.99919033e-01 6.10209644e-01
2.05520391e-01 -1.04565477e+00 -9.33738232e-01 -1.06298339e+00
-2.06015319e-01 -3.12661916e-01 2.81702280e-01 5.48990905e-01
-5.76785743e-01 2.84900069e-01 -3.95554960e-01 3.07376534e-01
8.75746846e-01 2.54404932e-01 1.19290245e+00 -9.83674347e-01
-4.37576354e-01 -2.24470690e-01 -7.96395421e-01 -1.18416703e+00
9.87223536e-02 -5.06645441e-01 3.06233734e-01 -1.10965526e+00
-2.80925669e-02 -4.62832242e-01 -4.68289554e-02 3.53855044e-01
-3.48317213e-02 7.18694687e-01 2.81863213e-01 2.60341585e-01
-5.84716320e-01 6.77097619e-01 1.05110967e+00 -1.27660900e-01
-1.21045820e-01 -4.81773168e-02 3.19754668e-02 5.98018110e-01
6.75197184e-01 -5.99419177e-01 -4.40194845e-01 -3.95711571e-01
1.19593680e-01 -8.03801641e-02 6.83620930e-01 -1.10933626e+00
-1.85421035e-02 8.43263641e-02 3.90265584e-01 -7.90897787e-01
1.07983041e+00 -8.25347960e-01 2.42361858e-01 4.80110168e-01
-1.32319316e-01 1.39625922e-01 1.76599368e-01 5.80299675e-01
-1.64374664e-01 -3.30320746e-02 7.74762809e-01 1.29238024e-01
-6.28664494e-01 4.63970631e-01 2.78481960e-01 1.62491336e-01
1.02093983e+00 -3.08264583e-01 -1.55447414e-02 -3.30042660e-01
-3.96671712e-01 9.65061486e-02 8.40535223e-01 6.24718845e-01
5.31687260e-01 -1.28063893e+00 -6.80220842e-01 1.04883313e-01
3.73414516e-01 2.58978277e-01 7.17953146e-02 8.66034567e-01
-7.78536916e-01 6.47037998e-02 -1.13986293e-02 -1.04122519e+00
-1.72498012e+00 3.07027668e-01 3.95154595e-01 -7.72013292e-02
-3.65360707e-01 9.72734332e-01 2.99524158e-01 -6.78723633e-01
5.20796776e-01 4.27462719e-02 3.54492188e-01 6.41751066e-02
4.53028917e-01 1.56292692e-01 2.90209264e-01 -6.96238339e-01
-4.67160523e-01 9.66203809e-01 -1.86744601e-01 -1.07679948e-01
1.22439182e+00 -2.37516332e-02 7.53842518e-02 4.71464306e-01
1.58388853e+00 -1.56137750e-01 -1.52125788e+00 -2.90641725e-01
-2.52153009e-01 -1.02511311e+00 3.91990226e-03 -6.08734369e-01
-9.85342145e-01 8.83659720e-01 9.68630373e-01 -1.69001773e-01
9.18813407e-01 8.06542709e-02 6.37593091e-01 2.59285450e-01
4.56540525e-01 -1.02884543e+00 3.54087174e-01 1.54612049e-01
1.22153318e+00 -1.43790662e+00 4.05005753e-01 -5.72047353e-01
-2.57385850e-01 9.89180565e-01 7.03177452e-01 -2.02205926e-01
4.57489252e-01 -2.80800480e-02 3.07115465e-01 -1.71175733e-01
-4.32326108e-01 -9.25591309e-03 6.60950899e-01 3.44448030e-01
5.17527699e-01 -5.35758026e-02 -7.09704906e-02 -3.77791792e-01
-1.44728199e-01 -4.31352615e-01 -5.11044860e-02 1.03442025e+00
1.25286598e-02 -1.24609077e+00 -5.81268132e-01 2.23239645e-01
-3.93921912e-01 1.36636972e-01 -3.80934536e-01 9.25157607e-01
-3.95911839e-03 5.87275982e-01 -5.00993058e-02 -3.14682007e-01
4.86826032e-01 -8.18392560e-02 7.39137352e-01 -5.18860817e-01
-2.26342902e-01 2.66676068e-01 -6.32150546e-02 -9.01132047e-01
-3.75468224e-01 -9.22261119e-01 -6.44338548e-01 -1.86940670e-01
-4.05267149e-01 -3.74938190e-01 9.43402410e-01 5.78654110e-01
3.15960497e-01 1.87666029e-01 6.13012910e-01 -1.31143904e+00
-7.47067511e-01 -6.24068260e-01 -2.38090008e-01 6.27284646e-01
2.03762487e-01 -8.87014806e-01 -2.89318621e-01 -1.74811468e-01] | [7.830636978149414, -2.603100299835205] |
750ae449-9600-4224-8605-f3eec7497af9 | adversarial-de-confounding-in-individualised | 2210.10530 | null | https://arxiv.org/abs/2210.10530v3 | https://arxiv.org/pdf/2210.10530v3.pdf | Adversarial De-confounding in Individualised Treatment Effects Estimation | Observational studies have recently received significant attention from the machine learning community due to the increasingly available non-experimental observational data and the limitations of the experimental studies, such as considerable cost, impracticality, small and less representative sample sizes, etc. In observational studies, de-confounding is a fundamental problem of individualised treatment effects (ITE) estimation. This paper proposes disentangled representations with adversarial training to selectively balance the confounders in the binary treatment setting for the ITE estimation. The adversarial training of treatment policy selectively encourages treatment-agnostic balanced representations for the confounders and helps to estimate the ITE in the observational studies via counterfactual inference. Empirical results on synthetic and real-world datasets, with varying degrees of confounding, prove that our proposed approach improves the state-of-the-art methods in achieving lower error in the ITE estimation. | ['David A. Clifton', 'Tingting Zhu', 'Anshul Thakur', 'Marzia Hoque Tania', 'Soheila Molaei', 'Vinod Kumar Chauhan'] | 2022-10-19 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 3.66653591e-01 1.79253951e-01 -1.13110888e+00 -2.33519703e-01
-6.75080538e-01 -2.43328854e-01 5.47836781e-01 -3.15198377e-02
-3.69217068e-01 1.34198141e+00 6.49247766e-01 -5.75597525e-01
-6.13851130e-01 -7.84194767e-01 -6.99349821e-01 -8.80869448e-01
-4.10706103e-01 3.70533884e-01 -5.74069381e-01 2.25967705e-01
6.91843182e-02 3.53128046e-01 -9.86613750e-01 -1.02485031e-01
1.19733965e+00 4.54006433e-01 -5.80151618e-01 2.57696718e-01
3.64445060e-01 5.20161629e-01 -4.01400179e-01 -4.09025490e-01
3.63620847e-01 -4.15786386e-01 -1.51135996e-01 -3.87306064e-01
4.37035382e-01 -3.32961231e-01 -6.03615820e-01 9.37261045e-01
9.13061321e-01 7.70174200e-03 1.13807297e+00 -1.49577963e+00
-9.69608605e-01 8.46314907e-01 -1.05739093e+00 3.03675503e-01
-1.14606120e-01 4.25249606e-01 8.70180666e-01 -3.54198009e-01
2.34882876e-01 1.42342365e+00 6.20748103e-01 7.14091778e-01
-1.51402378e+00 -1.41940546e+00 1.73312932e-01 3.16253118e-02
-1.09714389e+00 -2.70473838e-01 6.78539574e-01 -6.47558749e-01
2.59109676e-01 1.71395019e-01 3.70398760e-01 1.73381495e+00
3.92159402e-01 4.42517400e-01 1.48214519e+00 -2.26572111e-01
4.59522784e-01 4.84005101e-02 -8.47020298e-02 3.35387677e-01
8.25378418e-01 1.12727928e+00 1.16095603e-01 -5.78736305e-01
1.01084113e+00 3.75832647e-01 -2.93027937e-01 -5.15562952e-01
-1.09003174e+00 1.27615774e+00 5.86044192e-01 -1.33358300e-01
-8.73715281e-01 1.96954891e-01 6.19460762e-01 1.30140647e-01
5.91967583e-01 5.20077288e-01 -5.08783877e-01 3.78122836e-01
-7.35150039e-01 4.95914012e-01 3.05989027e-01 6.08241618e-01
4.09609601e-02 2.26279646e-01 -5.58239639e-01 4.80493516e-01
-2.11417135e-02 6.71826720e-01 6.01809680e-01 -8.35703611e-01
7.61289954e-01 5.74988961e-01 3.62892389e-01 -8.94144475e-01
-4.19740289e-01 -4.76166844e-01 -1.62851799e+00 3.46777141e-01
4.99954551e-01 -5.65597117e-01 -8.63710940e-01 2.30056834e+00
4.67279762e-01 6.38754368e-01 -1.22662567e-01 8.88466895e-01
5.81712425e-01 2.29247421e-01 7.75441051e-01 -6.05725110e-01
1.34729195e+00 -3.97504568e-01 -9.30576622e-01 -2.45987147e-01
4.97342110e-01 -2.34664738e-01 8.90545845e-01 2.37571988e-02
-9.22593594e-01 -2.48114154e-01 -8.15964758e-01 3.89310449e-01
-2.14243829e-01 -1.73530038e-02 9.47970510e-01 8.95632327e-01
-2.18271911e-01 5.61120510e-01 -4.41690147e-01 1.49034336e-01
1.04705679e+00 5.96114814e-01 -4.07737464e-01 7.23153725e-02
-1.64644277e+00 6.79854095e-01 1.17409945e-01 1.21923955e-02
-1.12888467e+00 -1.48757482e+00 -7.93794870e-01 3.45476478e-01
6.27343476e-01 -1.22135687e+00 9.61159408e-01 -8.80063772e-01
-1.27415359e+00 6.75238073e-01 3.58686626e-01 -5.15467942e-01
8.38847458e-01 -6.25746921e-02 -2.45191738e-01 -2.44046688e-01
2.00668983e-02 2.96852112e-01 7.34850347e-01 -9.32352424e-01
-3.36579293e-01 -6.62020802e-01 4.21626642e-02 3.72246467e-02
3.42014581e-02 -5.57583869e-02 6.21828854e-01 -1.12515485e+00
-5.90305805e-01 -9.41321135e-01 -6.53694272e-01 -3.30810063e-03
-5.03212631e-01 -1.74697340e-01 2.48277918e-01 -4.51196611e-01
1.22947717e+00 -2.10436535e+00 1.75555885e-01 -2.01575041e-01
5.52486122e-01 2.44708836e-01 -5.07499576e-02 1.95798174e-01
-7.15617001e-01 4.18474525e-01 -1.99879855e-01 -5.45790382e-02
-1.38446810e-02 -1.89524181e-02 -3.64620000e-01 8.72338891e-01
7.78861865e-02 7.92652190e-01 -1.09302354e+00 -5.55172384e-01
3.14377606e-01 1.49920076e-01 -6.93685830e-01 4.27359819e-01
1.76703371e-02 7.11532176e-01 -7.74430037e-01 7.28650913e-02
8.84831429e-01 -9.11987051e-02 2.29038864e-01 5.79463616e-02
6.80645406e-02 1.77930847e-01 -1.13551474e+00 1.00975895e+00
-5.80838799e-01 2.29677588e-01 -1.21704504e-01 -1.46447372e+00
3.28419924e-01 6.66583240e-01 4.02116567e-01 -5.67273200e-01
3.89316827e-01 -3.15787643e-02 3.49991560e-01 -6.27537429e-01
-1.44496471e-01 -7.49906003e-01 -4.40966517e-01 3.34139168e-01
-2.61527866e-01 2.05091298e-01 -3.24313968e-01 -4.45291027e-02
7.26475596e-01 -3.53316218e-01 9.40150619e-01 -3.89770865e-01
2.71833301e-01 -4.50040609e-01 8.65874350e-01 8.70668650e-01
-3.63241851e-01 3.87425423e-01 7.56447911e-01 -3.64756852e-01
-1.01252615e+00 -1.22469640e+00 -3.04376394e-01 6.10791445e-01
-3.16277206e-01 5.19847810e-01 -4.24706817e-01 -8.35468590e-01
5.07664025e-01 8.94913256e-01 -1.42839897e+00 -6.22457147e-01
-5.90810001e-01 -1.22616851e+00 5.84626675e-01 8.23149979e-01
1.84216067e-01 -1.01444876e+00 -2.09541485e-01 -5.22559620e-02
2.01333240e-01 -4.77819383e-01 -4.90773559e-01 -5.25628030e-02
-1.05517745e+00 -1.36249864e+00 -9.83563483e-01 -3.49527806e-01
6.41376793e-01 -5.71252313e-04 9.03546512e-01 -3.74443918e-01
-1.41909763e-01 -2.97315180e-01 3.40344280e-01 -7.21720815e-01
-4.82123077e-01 -1.70306861e-01 1.84035599e-01 -1.48419261e-01
2.75932461e-01 -7.34865248e-01 -1.22108161e+00 1.97801411e-01
-7.81515121e-01 -1.51718691e-01 6.91464782e-01 1.19598269e+00
2.97463447e-01 -1.91774294e-01 1.17831802e+00 -1.45355356e+00
6.50719523e-01 -1.03031635e+00 -8.19184780e-01 -3.38118374e-02
-8.07722747e-01 1.34970665e-01 7.63467610e-01 -1.03815925e+00
-1.16423571e+00 -6.64269805e-01 4.53370839e-01 -3.64486694e-01
-2.11934552e-01 3.36004108e-01 -1.71973705e-01 4.31444317e-01
7.87585795e-01 -3.34913105e-01 -4.26950529e-02 -2.98792392e-01
4.47645664e-01 6.79915547e-01 1.58647060e-01 -4.81912374e-01
5.88127792e-01 3.46152574e-01 4.11991894e-01 -3.18945274e-02
-9.56967592e-01 6.21320158e-02 -8.55991170e-02 4.04181093e-01
6.61493003e-01 -1.00356197e+00 -1.08513165e+00 1.31498188e-01
-6.15837991e-01 -2.52868921e-01 -4.74561840e-01 9.59656477e-01
-6.15726233e-01 7.89646581e-02 -3.22717816e-01 -8.38996589e-01
-4.79551703e-01 -1.19462490e+00 7.51774073e-01 2.82342672e-01
-3.89569819e-01 -1.07590103e+00 3.90684247e-01 2.63471901e-01
2.50892788e-01 6.56776726e-01 1.46238899e+00 -5.07894397e-01
-1.61409944e-01 -4.33896333e-01 -2.91908175e-01 7.82848299e-02
4.62226152e-01 -1.82898983e-01 -8.41722906e-01 -2.81999737e-01
-8.64988640e-02 -2.17503533e-01 5.54485381e-01 1.18108892e+00
1.47687769e+00 -7.34487712e-01 -4.74314839e-01 4.44749862e-01
1.32700932e+00 2.87369996e-01 5.33591032e-01 -5.42391986e-02
3.69302630e-01 4.45675164e-01 3.91611040e-01 5.76794386e-01
1.35534003e-01 4.89412367e-01 4.91623282e-01 -2.59517133e-01
6.74279630e-02 -3.57891589e-01 -1.82606816e-01 2.73189768e-02
-1.64405212e-01 -2.89704204e-01 -4.21371549e-01 6.88670099e-01
-1.77231729e+00 -1.17296326e+00 -1.14105090e-01 2.66989326e+00
1.12429595e+00 -3.40779573e-02 3.12602490e-01 1.75820645e-02
9.64962661e-01 1.10878699e-01 -6.70570314e-01 -5.79913437e-01
9.07419771e-02 3.12943906e-01 8.10435772e-01 2.04519093e-01
-1.11317694e+00 3.29087973e-01 6.74823093e+00 6.46213412e-01
-1.02583206e+00 2.18870282e-01 8.58898401e-01 -5.61492369e-02
-2.85309970e-01 -2.00473472e-01 -2.03662038e-01 7.56150186e-01
9.56571162e-01 -7.27156043e-01 1.58848166e-01 4.73714769e-01
7.13201523e-01 9.97580364e-02 -1.27690077e+00 5.51369071e-01
-5.49943030e-01 -1.18011832e+00 3.21479514e-03 2.17487603e-01
1.17751825e+00 -2.92037100e-01 4.97303694e-01 4.84776407e-01
7.08133042e-01 -1.32112229e+00 1.72766119e-01 4.72947478e-01
1.02114379e+00 -7.16750562e-01 1.02849734e+00 3.22212070e-01
-3.86704355e-01 -3.81943375e-01 -2.72010475e-01 -2.43931085e-01
-1.48309708e-01 5.78721762e-01 -6.04041696e-01 5.29607594e-01
2.26369143e-01 4.23103243e-01 2.77065467e-02 9.53621984e-01
-4.47323978e-01 9.44923520e-01 -6.29399046e-02 4.27367948e-02
2.01893166e-01 -2.38705888e-01 2.89985627e-01 7.20954716e-01
1.19682513e-01 2.56343931e-01 -1.92301631e-01 9.00632441e-01
-4.62213218e-01 9.45184380e-02 -8.50689411e-01 3.52126248e-02
5.23377120e-01 6.48914814e-01 -9.61584228e-05 -2.88713574e-01
-3.74514461e-01 1.32914752e-01 1.99965641e-01 4.75022078e-01
-1.00647461e+00 1.53422654e-01 7.73402214e-01 1.13569811e-01
-8.29871818e-02 5.30163944e-01 -5.57175338e-01 -9.48265195e-01
-4.03319329e-01 -1.18017745e+00 8.17045212e-01 -4.40543681e-01
-1.58802378e+00 -1.17113642e-01 3.58978719e-01 -1.11671162e+00
-3.49442422e-01 -3.75477433e-01 -6.49032116e-01 1.13834798e+00
-1.18959296e+00 -7.71942258e-01 1.13010742e-01 3.13563049e-01
3.18405956e-01 -2.36106571e-02 9.22012508e-01 4.78785694e-01
-9.70551074e-01 8.54068160e-01 3.33381414e-01 -2.38929112e-02
7.81031013e-01 -1.20326805e+00 -6.23604394e-02 2.96613157e-01
-5.84936142e-01 6.48457944e-01 8.15264344e-01 -7.34844863e-01
-1.01814342e+00 -9.94545341e-01 6.54570997e-01 -2.28771254e-01
8.40439498e-01 -1.89811170e-01 -6.38977170e-01 7.70145178e-01
1.34687990e-01 1.35910660e-01 7.82816172e-01 3.47987354e-01
-4.59277302e-01 -1.90980300e-01 -1.42967272e+00 1.06798148e+00
8.92788708e-01 -1.66961670e-01 -7.96039283e-01 4.18807477e-01
6.48876309e-01 -1.09139115e-01 -8.20007503e-01 6.52531505e-01
6.85209930e-01 -4.24623191e-01 1.19817626e+00 -1.51667225e+00
7.96349049e-01 3.63241613e-01 2.56627440e-01 -1.67181015e+00
-5.36780417e-01 -4.80335414e-01 3.75404395e-02 9.81919944e-01
2.97140181e-01 -8.71981025e-01 8.66433203e-01 6.25526309e-01
5.27834475e-01 -6.81369305e-01 -1.13940740e+00 -4.94354069e-01
6.88836932e-01 9.90781561e-02 8.34378600e-01 1.21992993e+00
-2.18467668e-01 2.96101838e-01 -7.97037661e-01 4.22253281e-01
9.26882863e-01 3.39659542e-01 6.00823879e-01 -1.24239647e+00
-2.37780184e-01 -6.19245350e-01 -3.88564795e-01 -2.54274458e-01
4.18957353e-01 -6.72562003e-01 -4.07475382e-01 -1.10188293e+00
6.27389550e-01 -5.21403670e-01 -6.10607028e-01 2.72165179e-01
-7.89423048e-01 -2.86803871e-01 -2.01070815e-01 -3.76166344e-01
1.71265647e-01 8.60133350e-01 1.51713312e+00 -3.82496566e-01
-2.68190298e-02 3.36666852e-01 -9.86760557e-01 5.08322060e-01
7.22990096e-01 -1.07349288e+00 -6.61917448e-01 1.28806308e-02
-1.69154592e-02 4.58229601e-01 5.05241513e-01 -2.56477714e-01
-3.74390215e-01 -6.25490546e-01 3.19490939e-01 3.86457741e-02
-9.61480439e-02 -9.16640162e-01 2.48627499e-01 8.28781009e-01
-8.05681109e-01 2.17342582e-02 3.12223136e-01 7.73024499e-01
1.58224806e-01 1.22046359e-02 8.52275908e-01 4.31624763e-02
1.79352105e-01 5.15707910e-01 -1.11172415e-01 4.74798232e-01
1.02646470e+00 2.32917562e-01 -5.02668977e-01 -3.43141973e-01
-4.74393964e-01 2.98637986e-01 -8.89831558e-02 3.15489918e-01
2.96930730e-01 -1.33223104e+00 -1.09651208e+00 6.87418804e-02
2.29053870e-02 -2.96272516e-01 6.32635415e-01 9.39511657e-01
1.18291102e-01 3.16583574e-01 -2.58775890e-01 8.53919517e-03
-9.94826436e-01 1.22802556e+00 3.70690644e-01 -6.87043667e-01
-4.91745800e-01 3.49800318e-01 1.07652080e+00 -4.25581932e-01
1.26721948e-01 -1.96952820e-01 -1.78445950e-01 -1.13447569e-01
4.16979671e-01 6.92178071e-01 -3.77340704e-01 -1.85735390e-01
-2.04580173e-01 2.16129914e-01 -3.83684002e-02 2.74941832e-01
1.45974374e+00 1.62590876e-01 2.64686763e-01 2.65015692e-01
1.13115966e+00 -9.33975503e-02 -1.08883572e+00 -1.51363403e-01
-4.58769768e-01 -6.58207476e-01 2.59466738e-01 -8.83475006e-01
-1.20875454e+00 8.22985649e-01 8.08064461e-01 1.61044583e-01
1.03476858e+00 -4.11125451e-01 9.92872790e-02 -1.90871805e-01
1.87414065e-01 -5.59081078e-01 -3.98465335e-01 -4.35045749e-01
9.82469976e-01 -1.33222640e+00 4.13562089e-01 -4.13961917e-01
-4.55034554e-01 3.35541070e-01 4.70003366e-01 -5.80620944e-01
7.57253110e-01 1.99503064e-01 -4.71992604e-02 -3.83459218e-02
-6.03373230e-01 4.57699418e-01 3.38898182e-01 5.56382835e-01
4.69979227e-01 5.52021265e-01 -7.98107743e-01 7.91243613e-01
1.59354761e-01 3.03183377e-01 3.70447725e-01 3.37598473e-01
6.01489782e-01 -8.96879017e-01 -3.85585934e-01 7.69216895e-01
-8.82920921e-01 -1.30973786e-01 -1.25289802e-02 1.32701945e+00
2.71036401e-02 7.68472493e-01 -4.37809117e-02 2.44967297e-01
6.66127443e-01 -1.71148479e-01 3.74077290e-01 -3.06892663e-01
-6.77006602e-01 -3.06121975e-01 -6.61931559e-02 -4.36518133e-01
-4.73831624e-01 -5.71121812e-01 -6.75053775e-01 -5.00787735e-01
-3.79750252e-01 2.22434253e-01 8.05727765e-02 7.81157732e-01
4.65470076e-01 7.67798007e-01 1.04267430e+00 -4.46499467e-01
-1.28254247e+00 -1.07884383e+00 -6.13627017e-01 7.13185132e-01
5.12675822e-01 -1.05569470e+00 -7.61258841e-01 -5.07843494e-01] | [8.042593955993652, 5.388043403625488] |
8a406eaf-c131-4072-9e3f-59e822e1c8d7 | synthesis-of-realistic-ecg-using-generative | 1909.09150 | null | https://arxiv.org/abs/1909.09150v1 | https://arxiv.org/pdf/1909.09150v1.pdf | Synthesis of Realistic ECG using Generative Adversarial Networks | Access to medical data is highly restricted due to its sensitive nature, preventing communities from using this data for research or clinical training. Common methods of de-identification implemented to enable the sharing of data are sometimes inadequate to protect the individuals contained in the data. For our research, we investigate the ability of generative adversarial networks (GANs) to produce realistic medical time series data which can be used without concerns over privacy. The aim is to generate synthetic ECG signals representative of normal ECG waveforms. GANs have been used successfully to generate good quality synthetic time series and have been shown to prevent re-identification of individual records. In this work, a range of GAN architectures are developed to generate synthetic sine waves and synthetic ECG. Two evaluation metrics are then used to quantitatively assess how suitable the synthetic data is for real world applications such as clinical training and data analysis. Finally, we discuss the privacy concerns associated with sharing synthetic data produced by GANs and test their ability to withstand a simple membership inference attack. For the first time we both quantitatively and qualitatively demonstrate that GAN architecture can successfully generate time series signals that are not only structurally similar to the training sets but also diverse in nature across generated samples. We also report on their ability to withstand a simple membership inference attack, protecting the privacy of the training set. | ['Anne Marie Delaney', 'Eoin Brophy', 'Tomas E. Ward'] | 2019-09-19 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 6.07901394e-01 5.67898750e-01 4.39634055e-01 -3.14180225e-01
-1.01101291e+00 -8.99112403e-01 3.46263558e-01 3.10010258e-02
-1.53685868e-01 1.11662138e+00 -1.45539343e-02 -2.40513399e-01
8.82610828e-02 -8.63931417e-01 -6.98020995e-01 -6.57298803e-01
-3.14180791e-01 2.01788485e-01 -5.06341279e-01 6.95009157e-02
-2.66791046e-01 4.97768670e-01 -1.30738890e+00 2.87018895e-01
8.29557061e-01 8.18401456e-01 -8.13115478e-01 7.19215214e-01
6.17377162e-01 3.60196769e-01 -1.57829106e+00 -6.27947569e-01
6.71711206e-01 -9.12688494e-01 -4.60692734e-01 -2.78884798e-01
1.70709938e-01 -3.61381412e-01 -1.44084230e-01 8.62462878e-01
1.06191647e+00 -2.48290703e-01 4.82056588e-01 -1.57577634e+00
-3.84936899e-01 5.59461892e-01 1.05297215e-01 -1.06770553e-01
4.61288601e-01 4.42296743e-01 3.98451030e-01 3.44609283e-02
6.64145529e-01 5.88296771e-01 1.02424467e+00 1.00844145e+00
-1.50634980e+00 -1.00672483e+00 -9.43334997e-01 -4.87791598e-01
-1.65262651e+00 -4.90449399e-01 8.27475846e-01 -2.19981149e-01
2.33781949e-01 8.03436399e-01 7.43821800e-01 1.69722986e+00
3.37093920e-01 2.21313596e-01 1.39556932e+00 -8.61097351e-02
4.29535121e-01 3.17930043e-01 -4.20158803e-01 1.96055725e-01
3.71347904e-01 2.59539366e-01 -3.88412535e-01 -7.68330097e-01
7.19733059e-01 -3.81756037e-01 -5.16560674e-01 -1.95650826e-03
-1.27622998e+00 8.64507198e-01 5.75415604e-03 1.59739256e-01
-4.13887441e-01 2.39817247e-01 5.87779522e-01 5.97114980e-01
2.27474377e-01 9.93588388e-01 -5.39478995e-02 -1.93628073e-02
-8.76459062e-01 2.25787625e-01 9.40414190e-01 9.48447764e-01
7.78772384e-02 4.36573386e-01 -3.53629112e-01 3.58023793e-01
-2.57784605e-01 5.84815443e-01 7.29328096e-01 -8.50740671e-01
2.29666948e-01 5.97575456e-02 -8.34109727e-03 -1.14437652e+00
-6.44175708e-02 -3.68910521e-01 -1.23292327e+00 2.42842689e-01
5.63700378e-01 -6.79681778e-01 -7.77682960e-01 2.03705859e+00
1.48448944e-01 3.20450932e-01 5.21008193e-01 4.68394876e-01
8.37797403e-01 3.03439945e-01 -1.22496508e-01 -2.41218716e-01
1.24090326e+00 8.16249251e-02 -7.38301277e-01 4.21510190e-01
4.58990723e-01 -5.77288806e-01 6.73876882e-01 2.00581297e-01
-1.26264775e+00 -4.07786667e-01 -1.20614231e+00 3.57806206e-01
-2.01374769e-01 -3.04325491e-01 3.03581238e-01 1.43668640e+00
-1.02852595e+00 6.41096652e-01 -7.23372161e-01 -5.73054142e-03
8.67177665e-01 3.35461140e-01 -5.84043682e-01 2.73204535e-01
-1.69097269e+00 5.12404323e-01 1.35027021e-01 -1.93307623e-01
-7.43158281e-01 -1.10762513e+00 -6.92772090e-01 -1.01686761e-01
-3.20222467e-01 -9.48682845e-01 7.02927291e-01 -9.96660531e-01
-1.20209146e+00 9.83681619e-01 4.06167507e-01 -9.61310744e-01
1.02079701e+00 3.62280965e-01 -8.68233204e-01 1.33182645e-01
-3.10004950e-02 4.21626598e-01 8.45574081e-01 -1.22631168e+00
3.42593482e-03 -2.88627446e-02 -2.64099181e-01 -2.86454886e-01
-2.08939984e-01 -2.15752378e-01 2.62276679e-01 -1.18789744e+00
-3.93790781e-01 -1.00128365e+00 -2.05156982e-01 -1.31898567e-01
-8.51116240e-01 6.85519755e-01 5.72622120e-01 -6.40599549e-01
9.85374272e-01 -2.13264561e+00 -4.24076706e-01 6.84958637e-01
1.95981383e-01 4.12025958e-01 3.03896293e-02 5.81079960e-01
-1.54521585e-01 6.47984147e-01 -5.05137682e-01 -1.72343299e-01
-1.70397028e-01 1.31355435e-01 -5.67676783e-01 5.60300648e-01
2.34725788e-01 1.02064335e+00 -5.98079741e-01 -1.59842029e-01
-2.22139001e-01 8.61085236e-01 -4.30389524e-01 3.37779254e-01
1.18617311e-01 1.03920650e+00 -3.44870567e-01 3.28247607e-01
5.59933007e-01 1.27060488e-02 6.30565882e-02 -2.51128048e-01
5.73166370e-01 -7.59611875e-02 -8.07560086e-01 1.27926576e+00
-1.59695745e-01 6.24281228e-01 -8.12774152e-02 -6.29573524e-01
9.88038480e-01 8.03112090e-01 5.64376354e-01 -4.79659230e-01
2.39781857e-01 1.78823888e-01 3.53529394e-01 -1.53693840e-01
-1.09888781e-02 -3.33243966e-01 -4.48107272e-01 8.00521374e-01
-3.50310743e-01 -3.36030394e-01 -3.22208464e-01 -1.32209599e-01
1.18331349e+00 -3.68929595e-01 1.64238930e-01 -2.52417326e-01
3.39343190e-01 -2.09590435e-01 5.12619674e-01 7.56400406e-01
1.78670194e-02 1.10622168e+00 5.12511730e-01 -3.26382369e-01
-1.30509818e+00 -1.16562474e+00 -2.57451445e-01 -1.01130284e-01
-2.80903786e-01 -2.49111846e-01 -9.45685387e-01 -6.32054806e-01
-5.85817173e-02 8.46387565e-01 -8.53022695e-01 -5.30070186e-01
-4.56507415e-01 -8.43702137e-01 1.67471361e+00 4.67591047e-01
3.81476730e-01 -1.11599481e+00 -1.02771115e+00 1.26146972e-01
-2.36184552e-01 -8.92067015e-01 -5.61601758e-01 -6.47553578e-02
-6.92963123e-01 -9.63495612e-01 -8.40186536e-01 -3.62850964e-01
8.24689448e-01 -6.69546664e-01 1.11848462e+00 1.17631212e-01
-7.37856150e-01 5.77945471e-01 -2.29854748e-01 -8.79761994e-01
-1.21377373e+00 -1.76943138e-01 5.79831377e-02 2.78145939e-01
-8.21894482e-02 -9.74057376e-01 -8.03119481e-01 4.18816894e-01
-1.27581835e+00 -2.50123411e-01 1.54824376e-01 8.98288190e-01
5.84874451e-01 9.50736031e-02 1.05310607e+00 -1.31741416e+00
1.06347978e+00 -5.76544702e-01 -1.34850562e-01 1.48758113e-01
-5.59808254e-01 -2.47881800e-01 9.64454591e-01 -5.49511731e-01
-4.69334096e-01 -1.93026945e-01 -2.35508084e-01 -5.18097281e-01
-1.95372984e-01 3.08335662e-01 -1.37588292e-01 -8.38124976e-02
9.35675144e-01 2.94434935e-01 4.10598069e-01 -2.55950466e-02
1.41136810e-01 6.38410151e-01 6.39236987e-01 -4.53095645e-01
9.15168464e-01 5.46421826e-01 1.68233335e-01 -6.67719960e-01
-1.72669187e-01 3.69938701e-01 -2.04095677e-01 5.76587506e-02
8.06799352e-01 -7.95884371e-01 -7.27214694e-01 6.78038359e-01
-7.38238752e-01 -1.83808431e-01 -8.29865813e-01 2.41449952e-01
-7.70756364e-01 1.80795744e-01 -4.44397718e-01 -6.75155222e-01
-7.51267076e-01 -6.84831440e-01 7.42170155e-01 -1.00647785e-01
-6.29602313e-01 -9.28216457e-01 -1.56055409e-02 2.87585020e-01
6.57546759e-01 1.40115941e+00 7.81742692e-01 -9.10390139e-01
-1.64435253e-01 -5.54029047e-01 5.30066669e-01 4.77441132e-01
5.04828990e-01 -2.46656731e-01 -1.16096878e+00 -6.11560702e-01
3.69460762e-01 -2.10766152e-01 1.14790760e-01 1.27606928e-01
1.15140641e+00 -7.44568169e-01 -1.03402235e-01 9.71346438e-01
1.05374217e+00 3.47854793e-01 1.16641486e+00 -2.19039395e-01
4.93425220e-01 6.72758043e-01 3.02666903e-01 3.94765556e-01
-2.70579338e-01 3.80008250e-01 1.47802159e-01 -2.89611131e-01
9.82908681e-02 -3.22315931e-01 -2.83535756e-02 5.66714168e-01
6.97758198e-02 -3.57041448e-01 -7.17998028e-01 5.64417839e-01
-1.22174466e+00 -1.16147780e+00 7.34007731e-02 2.48157716e+00
1.10320640e+00 -1.02607027e-01 1.91059798e-01 3.29449028e-01
6.97855949e-01 -2.07243830e-01 -6.62306666e-01 -4.92188781e-01
-3.58218879e-01 7.69947290e-01 6.10572696e-01 -1.74072325e-01
-8.92904282e-01 1.17800921e-01 6.60577536e+00 4.61093009e-01
-1.26959133e+00 -8.35943744e-02 9.87272978e-01 -1.70961916e-01
-4.91683662e-01 -5.10229588e-01 7.32731596e-02 7.07476199e-01
1.23655021e+00 -6.95213497e-01 2.06135988e-01 4.42606032e-01
1.44936843e-03 5.67799568e-01 -1.23358858e+00 8.99089396e-01
-1.44402593e-01 -1.49484527e+00 4.69903015e-02 1.47716895e-01
7.06330538e-01 -4.39950556e-01 3.86137813e-01 -2.28538319e-01
2.64186472e-01 -1.83772707e+00 4.54608768e-01 6.05500638e-01
1.43909061e+00 -1.02354181e+00 8.32880676e-01 1.55268505e-01
-6.53711140e-01 2.76623219e-01 8.92421696e-04 5.09489715e-01
1.50963113e-01 4.88050401e-01 -1.09736967e+00 6.45677269e-01
4.80917633e-01 1.13184392e-01 -4.78343517e-01 8.38503301e-01
8.72559473e-02 6.94295406e-01 -4.18655187e-01 1.82894200e-01
-2.84844995e-01 7.54970908e-02 6.72872543e-01 7.38915205e-01
6.73216224e-01 6.62372559e-02 -2.67663091e-01 1.08419478e+00
-1.20481387e-01 2.47452129e-02 -8.93203259e-01 -2.75898397e-01
6.57392204e-01 6.86075330e-01 -4.12845045e-01 5.66114718e-03
2.00376809e-01 9.43812013e-01 -3.23851854e-01 1.61965504e-01
-7.74302602e-01 -5.64227402e-01 7.80212760e-01 2.92935252e-01
4.25121523e-02 2.85424173e-01 -3.00181359e-01 -8.80688846e-01
1.49177209e-01 -1.35959744e+00 4.87670213e-01 -6.82745457e-01
-1.63684893e+00 1.04248857e+00 -1.64193958e-01 -1.61852074e+00
-7.61997163e-01 1.54471666e-01 -6.84396446e-01 1.25098479e+00
-6.91430032e-01 -1.19504094e+00 -9.59552154e-02 7.96709836e-01
-3.07634741e-01 -3.18489254e-01 1.52989793e+00 1.44461557e-01
-1.15558609e-01 1.23060381e+00 6.90958500e-02 4.54506457e-01
4.71757829e-01 -1.03597462e+00 6.48747146e-01 8.07804465e-01
1.41785532e-01 9.42246079e-01 8.14671576e-01 -7.00504363e-01
-1.18619192e+00 -1.43843555e+00 2.57698298e-01 -4.45125639e-01
-1.77528635e-02 -4.47026134e-01 -1.03789628e+00 5.78459978e-01
1.32108375e-01 1.35724887e-01 1.19157457e+00 -6.38858616e-01
-3.63644660e-01 -4.84211408e-02 -1.99962258e+00 6.20817125e-01
4.88566220e-01 -5.98228157e-01 -3.15110385e-01 9.99581162e-03
5.12150764e-01 -4.46169883e-01 -1.48210311e+00 4.63605970e-01
5.92806399e-01 -8.97735238e-01 1.01506388e+00 -6.41822696e-01
2.05148280e-01 -2.97848582e-01 9.92557257e-02 -1.38822782e+00
2.92235970e-01 -1.30048108e+00 1.61865175e-01 1.52025604e+00
5.29038370e-01 -1.05700970e+00 8.47552359e-01 7.95741379e-01
4.66277540e-01 -4.48549688e-01 -1.03426099e+00 -7.68817484e-01
2.14155331e-01 -1.87272102e-01 1.12777042e+00 1.31830788e+00
-1.78132743e-01 -2.10120946e-01 -6.01419210e-01 1.80727869e-01
7.18519330e-01 -5.62307276e-02 9.30473268e-01 -9.71142173e-01
-3.51051956e-01 -8.08781013e-02 -8.42303813e-01 8.76406059e-02
-1.52912259e-01 -8.74234736e-01 -2.30380610e-01 -8.40327024e-01
-4.43051457e-01 -7.79675901e-01 -1.50358275e-01 2.95876384e-01
4.88455109e-02 7.45469093e-01 1.44599155e-01 4.92698811e-02
5.34195304e-01 1.75624207e-01 1.05351639e+00 2.57419012e-02
-1.47643864e-01 3.39761645e-01 -8.90948117e-01 2.23414466e-01
1.07225120e+00 -6.54979348e-01 -8.10809672e-01 1.97028905e-01
8.34826306e-02 3.32093626e-01 5.98413229e-01 -1.30747437e+00
-7.22618699e-02 2.81399369e-01 4.59771752e-01 -2.93992851e-02
2.74760246e-01 -9.16123211e-01 1.21443534e+00 5.93142450e-01
-5.26531219e-01 1.69950519e-02 3.82723302e-01 4.85039055e-01
-1.26012787e-01 3.85827571e-02 7.22001135e-01 -7.47699738e-02
2.75938839e-01 3.39383870e-01 -2.50661105e-01 4.09557879e-01
1.32159913e+00 -2.66028881e-01 -1.45079926e-01 -8.44459236e-01
-8.92997205e-01 -7.71220624e-02 7.51002550e-01 2.09309757e-01
5.29448271e-01 -1.37271285e+00 -1.04407823e+00 5.74543834e-01
1.94787830e-02 -1.45841092e-01 3.28929335e-01 4.78370845e-01
-6.71419859e-01 -5.12146764e-02 -5.47872186e-01 -4.36160684e-01
-1.38733399e+00 6.89772666e-01 4.70317274e-01 1.73318554e-02
-8.26744676e-01 4.57924455e-01 6.90349936e-02 -2.46762633e-01
8.39740187e-02 -1.15306720e-01 2.05631465e-01 -1.35653645e-01
4.61084336e-01 5.44876829e-02 1.10195696e-01 -4.29044664e-01
-2.68116146e-01 2.13061184e-01 4.25403386e-01 -1.43452570e-01
1.21324754e+00 1.34912610e-01 1.00689009e-01 2.13076442e-01
1.34013212e+00 3.58605236e-01 -8.69885802e-01 2.83295780e-01
-5.36627948e-01 -4.83654737e-01 -6.90115333e-01 -8.70433331e-01
-1.31986153e+00 6.49492860e-01 6.34779453e-01 8.05013120e-01
1.36364353e+00 -5.26347697e-01 8.65690172e-01 -1.46493524e-01
5.21368146e-01 -5.28651237e-01 -4.52289134e-01 -4.24120247e-01
8.08140457e-01 -6.48383319e-01 -1.51648402e-01 -4.25413936e-01
-7.28871167e-01 7.99561918e-01 2.07148138e-02 -1.60742253e-01
5.95177412e-01 5.99751472e-01 4.85100269e-01 2.20104642e-02
-4.59804386e-01 5.50895989e-01 2.44082630e-01 1.26642478e+00
2.29271084e-01 1.02704935e-01 -2.07962021e-01 6.78215444e-01
-7.87767291e-01 -2.30600610e-02 8.58614266e-01 8.26622486e-01
7.24234045e-01 -1.17488492e+00 -5.36758780e-01 7.10996270e-01
-1.11759651e+00 4.03249040e-02 -4.97238457e-01 6.72744155e-01
2.75225453e-02 9.21248496e-01 2.73532309e-02 -3.23040396e-01
1.43604293e-01 4.97719571e-02 1.81141019e-01 -2.37874657e-01
-1.23110688e+00 -3.92648876e-01 2.41881579e-01 -2.98846602e-01
-2.99902469e-01 -7.30652094e-01 -9.52490985e-01 -3.54112118e-01
3.00663356e-02 2.28893042e-01 4.42230850e-01 3.52540642e-01
6.95851564e-01 6.81804001e-01 7.52522230e-01 -8.09328184e-02
-5.78508377e-01 -4.97645259e-01 -5.43240368e-01 8.78077149e-01
3.35754216e-01 1.23070024e-01 -3.77358347e-01 2.36940876e-01] | [14.304494857788086, 3.0460662841796875] |
f8cdca36-aad2-431c-997a-608b71b1bb4f | ai-outperformed-every-dermatologist-improved | 2003.02597 | null | https://arxiv.org/abs/2003.02597v2 | https://arxiv.org/pdf/2003.02597v2.pdf | AI outperformed every dermatologist: Improved dermoscopic melanoma diagnosis through customizing batch logic and loss function in an optimized Deep CNN architecture | Melanoma, one of most dangerous types of skin cancer, re-sults in a very high mortality rate. Early detection and resection are two key points for a successful cure. Recent research has used artificial intelligence to classify melanoma and nevus and to compare the assessment of these algorithms to that of dermatologists. However, an imbalance of sensitivity and specificity measures affected the performance of existing models. This study proposes a method using deep convolutional neural networks aiming to detect melanoma as a binary classification problem. It involves 3 key features, namely customized batch logic, customized loss function and reformed fully connected layers. The training dataset is kept up to date including 17,302 images of melanoma and nevus; this is the largest dataset by far. The model performance is compared to that of 157 dermatologists from 12 university hospitals in Germany based on MClass-D dataset. The model outperformed all 157 dermatologists and achieved state-of-the-art performance with AUC at 94.4% with sensitivity of 85.0% and specificity of 95.0% using a prediction threshold of 0.5 on the MClass-D dataset of 100 dermoscopic images. Moreover, a threshold of 0.40858 showed the most balanced measure compared to other researches, and is promisingly application to medical diagnosis, with sensitivity of 90.0% and specificity of 93.8%. | ['Antoine Doucet', 'Dung Van Hoang', 'Cong Tri Pham', 'Mai Chi Luong'] | 2020-03-05 | null | null | null | null | ['melanoma-diagnosis'] | ['computer-vision'] | [ 1.85181394e-01 1.22932464e-01 -4.50693816e-01 -1.01368301e-01
-4.28648770e-01 -1.60546497e-01 2.89379507e-01 4.69907165e-01
-9.30429161e-01 8.41986060e-01 -4.62371618e-01 -3.94923240e-01
-4.57503229e-01 -9.33322787e-01 -1.43020555e-01 -8.41138482e-01
1.04740009e-01 9.41731557e-02 -2.92967185e-02 -7.12629482e-02
3.08962137e-01 6.30583167e-01 -1.33621216e+00 4.24797148e-01
1.27963448e+00 1.16443658e+00 4.41906713e-02 6.58082724e-01
6.35665804e-02 5.27144372e-01 -5.42094886e-01 -6.80148304e-01
1.44208014e-01 -1.71131864e-01 -7.22541034e-01 -4.44013737e-02
3.20933193e-01 -2.70939648e-01 -1.39641374e-01 9.15363014e-01
6.73279226e-01 -5.91449976e-01 9.63495612e-01 -7.52218544e-01
-6.12523079e-01 2.95788139e-01 -6.68625176e-01 -6.60660043e-02
5.24862222e-02 2.47860506e-01 6.19400084e-01 -4.42341596e-01
6.80325270e-01 4.89025563e-01 8.64918053e-01 8.72618854e-01
-8.13842595e-01 -4.44946170e-01 -3.53220463e-01 2.28026628e-01
-1.42187405e+00 9.68668684e-02 1.69345200e-01 -6.19929016e-01
6.90524638e-01 4.60458428e-01 9.47627246e-01 8.71300161e-01
5.90981960e-01 2.47752711e-01 1.42916703e+00 -5.45174420e-01
1.54873908e-01 6.16966784e-01 1.93330213e-01 9.42100227e-01
5.53891182e-01 6.85353298e-03 -8.65056664e-02 1.37621641e-01
5.86014211e-01 5.63451909e-02 -1.59082055e-01 3.03669035e-01
-7.42630541e-01 8.03055882e-01 5.23885250e-01 4.12890524e-01
-4.28495526e-01 -2.64904439e-01 3.41845334e-01 3.56221735e-01
2.73363292e-01 3.70941967e-01 -2.86814064e-01 3.98600727e-01
-5.91794014e-01 -2.81943053e-01 7.69552886e-01 1.47918448e-01
1.42590418e-01 -2.33209506e-01 -1.98906824e-01 9.14654553e-01
7.58756772e-02 2.79601246e-01 6.28815711e-01 -3.42414677e-01
-1.08063467e-01 1.09599900e+00 -2.76309490e-01 -7.30521381e-01
-7.30753481e-01 -5.21351933e-01 -1.30402040e+00 4.88435030e-01
5.58873832e-01 -4.02149647e-01 -1.12023222e+00 1.18108416e+00
1.11841215e-02 -4.10123646e-01 1.77289322e-01 8.94963980e-01
9.59376514e-01 1.90299422e-01 8.90437961e-02 -2.36071814e-02
1.40506887e+00 -6.19648397e-01 -4.43595856e-01 1.85745806e-01
5.72919488e-01 -5.86617708e-01 6.30757391e-01 8.80702376e-01
-8.64275157e-01 -3.41392040e-01 -1.10212004e+00 3.21167320e-01
-4.04928803e-01 6.22290134e-01 8.18792701e-01 1.03147233e+00
-1.12614787e+00 4.73265290e-01 -5.44634402e-01 -8.97158265e-01
5.91799319e-01 5.60970604e-01 -6.04857326e-01 1.05586857e-01
-1.09799981e+00 1.10574949e+00 4.25416857e-01 5.28761037e-02
-6.79798782e-01 -6.17680728e-01 -4.02973294e-01 -3.65610719e-01
7.71442577e-02 -6.53821945e-01 8.36802125e-01 -1.10715294e+00
-1.27637315e+00 1.31929624e+00 8.55510607e-02 -6.81262672e-01
7.01362491e-01 2.62954056e-01 -7.05223322e-01 3.66721720e-01
-2.65286297e-01 4.95919734e-01 4.42790389e-01 -8.50726008e-01
-9.31168675e-01 -5.82711756e-01 8.36876705e-02 2.04428341e-02
-6.72302067e-01 -2.12577611e-01 -2.20774248e-01 -1.24496445e-01
-2.15205878e-01 -6.82744026e-01 -4.94588733e-01 2.98027992e-01
-5.25414526e-01 -1.39819548e-01 2.36603305e-01 -8.95152450e-01
1.24106538e+00 -1.80757332e+00 -2.41809279e-01 4.13304478e-01
4.45632607e-01 6.91349983e-01 1.12652026e-01 4.21649069e-01
-1.59868896e-01 3.59147161e-01 -3.58733023e-03 2.07184717e-01
-3.51298362e-01 -2.79565006e-01 4.77413177e-01 5.13383687e-01
5.83790898e-01 6.57982886e-01 -6.38196111e-01 -5.31606317e-01
5.25496542e-01 6.34354711e-01 -4.85597849e-02 -1.51977018e-01
1.04476340e-01 1.02910809e-02 -3.36547792e-01 1.01870406e+00
6.93999350e-01 -1.88753709e-01 2.94320732e-01 -8.72480199e-02
-2.64597163e-02 -5.33123910e-01 -9.50423241e-01 1.23878062e+00
-3.95663112e-01 7.70947695e-01 -1.57357901e-01 -6.44465804e-01
1.06187177e+00 3.03686172e-01 3.58687252e-01 -6.12672389e-01
3.28281850e-01 2.56780744e-01 2.25218847e-01 -9.66088891e-01
-1.43519506e-01 -1.19096279e-01 3.15727860e-01 -1.46805048e-01
-8.89758542e-02 7.81286228e-03 2.63353795e-01 -1.15158677e-01
1.18064070e+00 -5.17263710e-01 6.09104812e-01 -1.21227607e-01
7.91454136e-01 2.28360042e-01 2.54541993e-01 6.09019935e-01
-2.60217816e-01 3.70874912e-01 8.41362953e-01 -6.01315737e-01
-7.61931419e-01 -9.56072152e-01 -6.03655159e-01 2.06892967e-01
-1.08577572e-01 -3.82829388e-03 -1.00630164e+00 -5.48526347e-01
6.71956465e-02 4.06648308e-01 -9.19630885e-01 -8.47131610e-02
4.19237614e-02 -1.16498733e+00 6.95696175e-01 2.09526479e-01
7.64850855e-01 -7.64027178e-01 -5.23216367e-01 -1.04579456e-01
2.58500844e-01 -6.96663737e-01 3.72054666e-01 4.32591289e-02
-6.37660682e-01 -1.63734937e+00 -8.61206055e-01 -8.71168613e-01
8.42207909e-01 -2.88410783e-01 8.07138622e-01 3.60986263e-01
-1.10804331e+00 -1.93964243e-01 -2.10713461e-01 -7.65073121e-01
-6.07679784e-01 1.57179281e-01 -2.29381070e-01 2.22780973e-01
5.57737708e-01 7.08733201e-02 -6.96356595e-01 7.55667686e-02
-9.63231981e-01 -8.75948519e-02 1.27351332e+00 1.03646898e+00
4.50973511e-01 2.08533153e-01 5.56802154e-01 -9.67152596e-01
6.66481256e-01 -3.32996756e-01 -4.28404510e-01 2.67084152e-01
-8.61279309e-01 -5.47984660e-01 5.95250309e-01 -1.83076128e-01
-7.68878698e-01 1.81173220e-01 -3.16674858e-01 1.33680522e-01
-4.61594075e-01 4.28358465e-01 3.14244986e-01 -4.27098870e-01
8.77278030e-01 -5.60554788e-02 4.99040663e-01 -6.69100359e-02
-3.87581348e-01 9.64481473e-01 1.77406371e-01 2.37156436e-01
3.14601123e-01 4.03381467e-01 4.81100291e-01 -1.01919353e+00
-6.38305962e-01 -3.80770326e-01 -5.31105340e-01 -3.19930106e-01
7.91419029e-01 -7.02267289e-01 -9.16456282e-01 1.08626640e+00
-7.55269051e-01 -8.89589190e-02 -9.83050242e-02 5.01038730e-01
-4.38145623e-02 2.72422642e-01 -6.24074757e-01 -9.57739472e-01
-6.41286552e-01 -1.04031241e+00 6.60465837e-01 4.82940882e-01
-1.27244547e-01 -1.13734078e+00 -2.27063060e-01 3.25392723e-01
3.79602134e-01 7.33933091e-01 8.17080915e-01 -6.24250710e-01
3.36869471e-02 -7.74787903e-01 -3.37275982e-01 6.78123772e-01
2.18320325e-01 5.22956312e-01 -1.16106200e+00 -2.57721424e-01
-2.67525434e-01 -2.53736645e-01 1.08351958e+00 5.03155172e-01
1.26418543e+00 4.66130935e-02 -4.73691076e-01 5.36890149e-01
1.98190629e+00 5.50848186e-01 7.53986239e-01 3.13752860e-01
2.39871919e-01 6.31484866e-01 4.88272041e-01 2.29511201e-01
1.49315253e-01 9.08845589e-02 8.50347817e-01 -4.80117738e-01
-1.26897216e-01 1.70478553e-01 -1.84757367e-01 2.44282156e-01
-4.01663333e-01 -2.40375280e-01 -1.12277925e+00 5.56489706e-01
-1.25475776e+00 -7.24700332e-01 -4.28079158e-01 2.35966516e+00
6.90894067e-01 3.31808925e-01 1.37754425e-01 4.74937171e-01
9.22049701e-01 -3.25936794e-01 -4.86824840e-01 -5.60438514e-01
-1.91457391e-01 4.47532207e-01 6.05423033e-01 3.54207963e-01
-1.18098938e+00 5.16485035e-01 6.08290243e+00 9.01156485e-01
-1.50305986e+00 -2.59385198e-01 1.01925647e+00 -3.26077715e-02
2.78342158e-01 -5.11575818e-01 -6.38519287e-01 5.26765049e-01
9.64306533e-01 4.83615249e-02 5.78841344e-02 3.67982745e-01
2.53695585e-02 -4.52374965e-01 -7.48874903e-01 7.70184875e-01
2.72898912e-01 -1.33548260e+00 -1.90294639e-03 2.68012226e-01
6.85754955e-01 -3.54476959e-01 4.13571805e-01 -8.04596469e-02
-1.17303818e-01 -1.60969341e+00 -1.93520948e-01 9.92632091e-01
1.21793306e+00 -8.47315073e-01 1.39065361e+00 2.80995280e-01
-5.27127862e-01 -2.44174391e-01 -2.88865060e-01 -1.48929104e-01
-3.94370764e-01 8.28109860e-01 -1.30646527e+00 6.06822848e-01
6.17514789e-01 5.28294683e-01 -6.35456443e-01 1.24109423e+00
-1.18527688e-01 5.95720410e-01 -2.09616840e-01 -7.62541056e-01
3.94963682e-01 -3.20171826e-02 6.97194114e-02 1.12004828e+00
1.88654348e-01 -3.10555696e-01 -5.74942589e-01 4.06433910e-01
2.55893081e-01 4.78918463e-01 -5.13021111e-01 -1.88655946e-02
2.37241149e-01 1.50813234e+00 -4.76887196e-01 3.43786106e-02
-2.90123135e-01 6.65963113e-01 -2.71596052e-02 1.87488273e-02
-7.81522334e-01 -7.30609596e-01 2.60025620e-01 1.78975046e-01
-1.66630194e-01 5.91651142e-01 -4.32882518e-01 -5.41864872e-01
-6.28863201e-02 -6.87139213e-01 5.43200135e-01 -3.64829808e-01
-1.36034894e+00 6.89714730e-01 -4.57331091e-01 -1.20083737e+00
7.09305182e-02 -1.29354751e+00 -7.63874531e-01 1.01873076e+00
-1.51501143e+00 -1.33357453e+00 -8.16027403e-01 4.18588191e-01
1.96581766e-01 -4.89175856e-01 1.08170259e+00 1.06806792e-02
-8.48364949e-01 8.30711782e-01 1.54935271e-01 2.44117573e-01
8.37555170e-01 -1.45854580e+00 -1.98004618e-01 3.91639054e-01
-6.87184036e-01 3.09792042e-01 5.37481457e-02 -4.59304094e-01
-1.04849744e+00 -1.02513719e+00 8.94058347e-01 -1.41600922e-01
6.69083893e-01 3.03585589e-01 -3.81094456e-01 -2.15546023e-02
2.67320842e-01 -3.18590760e-01 9.64349926e-01 -2.18196794e-01
-2.61547640e-02 -4.23395067e-01 -1.73068762e+00 5.73330522e-01
5.46860516e-01 -7.06579089e-02 -4.73541804e-02 5.24278164e-01
1.95112944e-01 -3.52217376e-01 -1.26415575e+00 5.70762873e-01
8.51101816e-01 -1.21664238e+00 8.00287545e-01 -5.17875254e-01
6.63059771e-01 1.99648723e-01 1.73570827e-01 -1.11861181e+00
-1.50243416e-01 -2.26188511e-01 2.60344803e-01 8.06888044e-01
6.77202165e-01 -9.18942034e-01 9.07683671e-01 2.96688348e-01
1.82255685e-01 -1.55517936e+00 -7.36027002e-01 -3.72928113e-01
2.94207186e-01 8.84171054e-02 1.92574099e-01 7.45210648e-01
-2.68898308e-01 1.40142692e-02 1.12787195e-01 -7.09219947e-02
5.64150274e-01 -3.03897232e-01 4.14021343e-01 -1.43672001e+00
1.29847556e-01 -6.71868980e-01 -8.44131410e-01 1.07313342e-01
-2.22531423e-01 -7.73290455e-01 -6.56598330e-01 -1.76892066e+00
2.94180512e-01 -2.80950487e-01 -4.86798644e-01 6.11931622e-01
1.13943033e-02 4.27964211e-01 -1.42728373e-01 -1.18183926e-01
1.80696592e-01 -3.76368016e-01 1.36931407e+00 -3.98260713e-01
-4.69601601e-02 3.14691126e-01 -8.67729425e-01 8.74480963e-01
1.27389467e+00 1.25317797e-01 -1.77465335e-01 1.26667060e-02
-8.08742549e-03 -2.89907549e-02 4.08006519e-01 -1.16930342e+00
4.00197774e-01 -2.41867632e-01 9.09460306e-01 -5.63159525e-01
4.17839050e-01 -7.50906348e-01 2.08811373e-01 1.21375906e+00
-2.49122694e-01 -3.97158533e-01 1.67455435e-01 3.08164209e-01
-1.72907069e-01 -5.66501379e-01 1.04541469e+00 -2.29103029e-01
-9.90368962e-01 2.33490050e-01 -3.83523226e-01 -4.13284749e-01
1.75700366e+00 -5.74606240e-01 -5.46680570e-01 9.39880908e-02
-7.08340883e-01 -4.10389341e-02 4.58012164e-01 6.02559978e-03
6.50860786e-01 -8.30027044e-01 -9.92793083e-01 1.82698771e-01
2.27111295e-01 -2.41572738e-01 4.95915115e-01 1.00608885e+00
-1.11814642e+00 4.80692178e-01 -6.26716912e-01 -5.13691068e-01
-1.53223026e+00 1.34856895e-01 6.76105201e-01 -2.96068460e-01
-2.83521920e-01 8.98136199e-01 -4.28442389e-01 -3.25970173e-01
3.67131621e-01 -1.26986831e-01 -5.46105623e-01 1.43348962e-01
5.43663979e-01 5.68356097e-01 3.26760799e-01 -7.36350566e-02
-3.41452032e-01 7.57272363e-01 -2.76950359e-01 2.75327086e-01
1.06174910e+00 4.48481202e-01 -3.52980405e-01 1.76365390e-01
9.08826530e-01 -1.39257669e-01 -6.32420361e-01 2.77071595e-01
-3.95965457e-01 -3.44110966e-01 4.12412174e-03 -1.50099158e+00
-1.07839775e+00 8.21949422e-01 1.03051805e+00 2.94761539e-01
1.44675314e+00 -4.49840695e-01 4.07048225e-01 4.10916507e-01
2.36991286e-01 -1.12598097e+00 -1.09308317e-01 1.08302474e-01
7.05266237e-01 -1.38423347e+00 3.19791250e-02 -5.54370582e-01
-7.25178242e-01 1.31893861e+00 8.67800593e-01 -1.96659043e-01
5.91498852e-01 2.25103974e-01 2.91819990e-01 -5.33903912e-02
-7.06365108e-01 -1.91388533e-01 3.40720028e-01 7.91932166e-01
4.43548054e-01 3.01029474e-01 -9.41364825e-01 5.12482643e-01
8.38553011e-02 3.43138933e-01 5.02942383e-01 6.92003548e-01
-3.34711611e-01 -8.77312183e-01 -2.30570868e-01 1.06780946e+00
-7.29219735e-01 -9.82323661e-03 -7.29541659e-01 1.34013116e+00
4.64873821e-01 9.00625587e-01 1.84145663e-02 -5.38421214e-01
2.13767499e-01 -1.93340853e-01 4.32174295e-01 -2.78893501e-01
-8.79279554e-01 -3.10688466e-01 2.24494144e-01 -2.40896448e-01
-3.31837147e-01 -2.65369833e-01 -1.01659846e+00 -3.00160170e-01
-3.97886634e-01 -9.20088217e-02 9.63303685e-01 5.27462006e-01
-2.39113625e-02 5.78293324e-01 6.33075297e-01 -3.02837160e-03
-6.89322352e-01 -9.70852792e-01 -9.35078859e-01 1.12372093e-01
2.26810172e-01 -2.12318420e-01 -3.01558584e-01 -2.29621992e-01] | [15.636204719543457, -2.987696409225464] |
11070a0c-ecd3-4c8b-a8b3-1af98d89af47 | limit-bert-linguistics-informed-multi-task | null | null | https://aclanthology.org/2020.findings-emnlp.399 | https://aclanthology.org/2020.findings-emnlp.399.pdf | LIMIT-BERT : Linguistics Informed Multi-Task BERT | In this paper, we present Linguistics Informed Multi-Task BERT (LIMIT-BERT) for learning language representations across multiple linguistics tasks by Multi-Task Learning. LIMIT-BERT includes five key linguistics tasks: Part-Of-Speech (POS) tags, constituent and dependency syntactic parsing, span and dependency semantic role labeling (SRL). Different from recent Multi-Task Deep Neural Networks (MT-DNN), our LIMIT-BERT is fully linguistics motivated and thus is capable of adopting an improved masked training objective according to syntactic and semantic constituents. Besides, LIMIT-BERT takes a semi-supervised learning strategy to offer the same large amount of linguistics task data as that for the language model training. As a result, LIMIT-BERT not only improves linguistics tasks performance but also benefits from a regularization effect and linguistics information that leads to more general representations to help adapt to new tasks and domains. LIMIT-BERT outperforms the strong baseline Whole Word Masking BERT on both dependency and constituent syntactic/semantic parsing, GLUE benchmark, and SNLI task. Our practice on the proposed LIMIT-BERT also enables us to release a well pre-trained model for multi-purpose of natural language processing tasks once for all. | ['Shuailiang Zhang', 'Hai Zhao', 'Zhuosheng Zhang', 'Junru Zhou'] | 2020-11-01 | null | null | null | findings-of-the-association-for-computational | ['semantic-role-labeling'] | ['natural-language-processing'] | [-1.86673969e-01 3.32682222e-01 -3.04721326e-01 -7.11641431e-01
-1.17019093e+00 -4.54836756e-01 2.50658900e-01 3.11327815e-01
-6.42431021e-01 6.35122538e-01 5.21843851e-01 -3.44326317e-01
3.19054484e-01 -5.51327050e-01 -6.36637747e-01 -4.42832887e-01
5.29546402e-02 7.00286865e-01 3.87572020e-01 -5.09388626e-01
-9.27830562e-02 1.41630843e-01 -1.23365843e+00 5.90206504e-01
7.74414957e-01 4.98079687e-01 7.32041359e-01 2.61247128e-01
-3.62311840e-01 6.15792096e-01 -3.84826392e-01 -4.14760739e-01
-1.85872599e-01 3.49238925e-02 -9.20179367e-01 -1.89175099e-01
8.10015574e-02 1.71240106e-01 3.65419000e-01 7.52008557e-01
6.81711435e-01 3.43790323e-01 4.08748955e-01 -8.17795575e-01
-1.07596922e+00 1.26904631e+00 -6.53126538e-01 2.76507556e-01
6.53151497e-02 -2.33547874e-02 1.62377119e+00 -1.10154188e+00
3.91284585e-01 1.82786667e+00 4.00322855e-01 8.37511182e-01
-1.08999455e+00 -6.30413949e-01 5.67613780e-01 2.93049682e-02
-8.57264519e-01 -3.53386343e-01 8.65694940e-01 -1.49604633e-01
1.44712722e+00 -2.07817093e-01 -4.62289937e-02 1.10884023e+00
6.16746023e-02 1.18379509e+00 1.02267265e+00 -6.93946302e-01
6.38823435e-02 -1.12852827e-01 6.34606957e-01 6.70714378e-01
8.89088288e-02 -3.31100464e-01 -3.92920077e-01 1.99904293e-01
5.16499519e-01 -5.93419731e-01 2.15168774e-01 1.45588309e-01
-1.27634966e+00 8.80326450e-01 1.86103791e-01 7.17612565e-01
-1.40533760e-01 2.63993621e-01 8.80813718e-01 1.98525444e-01
7.78926730e-01 3.02258223e-01 -1.12269378e+00 1.47131234e-01
-4.12303507e-01 8.10089409e-02 5.98017931e-01 8.81637692e-01
9.31342244e-01 4.22560066e-01 -2.19670802e-01 1.31962466e+00
5.10690749e-01 4.78553414e-01 7.94124186e-01 -6.21743023e-01
7.37224221e-01 6.45270586e-01 -4.78845149e-01 -4.00093585e-01
-8.05125356e-01 -2.03446999e-01 -6.54268503e-01 -1.85743332e-01
1.96214437e-01 -3.08172584e-01 -8.69069040e-01 2.03981113e+00
1.33769348e-01 -8.70179161e-02 3.10321182e-01 6.52389586e-01
1.19335115e+00 6.73355341e-01 6.68506444e-01 -5.40542565e-02
1.83490014e+00 -1.18757260e+00 -8.21983993e-01 -8.97064865e-01
1.10726035e+00 -5.49651325e-01 1.57233572e+00 1.19915850e-01
-1.09859729e+00 -7.39301860e-01 -7.70735443e-01 -4.91179436e-01
-5.57170093e-01 2.90008724e-01 8.63650322e-01 5.04952073e-01
-1.00234365e+00 4.34698574e-02 -7.75705099e-01 -3.06888312e-01
1.76385626e-01 2.81870753e-01 -3.31631601e-01 -1.40642330e-01
-1.55192327e+00 8.87740910e-01 8.10471117e-01 1.31612569e-01
-6.76579177e-01 -5.36994815e-01 -1.47540510e+00 5.96787296e-02
4.11670893e-01 -2.71128774e-01 1.26932371e+00 -7.65207052e-01
-1.36719489e+00 1.32521009e+00 -3.03872168e-01 -4.20906454e-01
-2.76082218e-01 -2.13176399e-01 -3.66396070e-01 -1.34472534e-01
4.36155140e-01 8.46424758e-01 4.97598231e-01 -1.09871590e+00
-5.94290853e-01 -3.33432615e-01 7.57912323e-02 1.45897776e-01
-4.00446951e-01 5.17090321e-01 -3.40142816e-01 -7.64568985e-01
-1.25268534e-01 -6.81229889e-01 -2.71892041e-01 -7.14755356e-01
-1.52895719e-01 -1.05172706e+00 6.34770215e-01 -5.29366076e-01
1.20635879e+00 -2.14588428e+00 1.80655301e-01 -5.64916253e-01
-3.48098204e-02 4.37974244e-01 -6.31223917e-01 3.17301273e-01
-2.05607340e-01 2.72000194e-01 -1.60047904e-01 -8.50305259e-01
7.59069026e-02 8.39733124e-01 -1.92668103e-02 1.81481257e-01
4.28949684e-01 1.31639516e+00 -9.47655976e-01 -5.64159095e-01
1.40806749e-01 1.69820458e-01 -4.59230125e-01 1.21223852e-01
-5.91513455e-01 4.37341124e-01 -4.69451904e-01 8.19404483e-01
5.55317044e-01 -2.10950106e-01 1.91314474e-01 -1.77585278e-02
-3.15619372e-02 6.59250438e-01 -8.13166261e-01 2.00946689e+00
-1.09038854e+00 5.59872799e-02 1.62714571e-01 -1.24069941e+00
1.07381082e+00 5.39378405e-01 2.63221294e-01 -7.78063953e-01
1.19974367e-01 3.54877979e-01 3.04983616e-01 -4.61496741e-01
2.69246519e-01 -2.44544923e-01 -7.57735908e-01 6.26886368e-01
4.94638085e-01 -4.12237197e-02 4.19186950e-01 -9.92842987e-02
8.80314767e-01 3.35557997e-01 5.23994386e-01 -6.03231192e-01
8.19567144e-01 -2.91335016e-01 1.00548840e+00 3.34935516e-01
-3.28796804e-01 2.53777325e-01 4.16692793e-01 -3.88889819e-01
-6.88627124e-01 -7.44710624e-01 -2.72880435e-01 2.10566807e+00
-2.15524063e-01 -2.40221754e-01 -4.95360345e-01 -9.24163818e-01
1.02384640e-02 9.86395895e-01 -4.93168384e-01 5.98486252e-02
-1.16937530e+00 -9.64981794e-01 6.64523423e-01 6.50791824e-01
4.83792722e-01 -1.81213045e+00 -2.18074515e-01 6.24986529e-01
-1.28485516e-01 -1.51693654e+00 -4.63945836e-01 6.20445132e-01
-6.00157440e-01 -7.10363567e-01 -4.55815464e-01 -1.37797093e+00
2.44891644e-01 7.60544837e-02 1.28734994e+00 3.82088218e-03
1.76102832e-01 6.33510426e-02 -6.67444348e-01 -4.91251588e-01
-5.22555709e-01 3.62058014e-01 -1.59984723e-01 -8.46339762e-02
7.39857316e-01 -5.74176431e-01 2.83370055e-02 1.27543941e-01
-7.82069266e-01 -1.29201978e-01 6.80300891e-01 9.14048791e-01
5.86053967e-01 -3.71626377e-01 1.09425223e+00 -1.17389607e+00
7.52162516e-01 -4.97742385e-01 -3.37713748e-01 3.80677640e-01
-3.09265018e-01 2.31254995e-01 7.30815649e-01 -3.22541386e-01
-1.32653916e+00 6.48762798e-04 -4.13579166e-01 1.99590567e-02
-2.35694379e-01 7.76941717e-01 -5.29570460e-01 4.47351098e-01
5.57969153e-01 -1.65860038e-02 -2.62558252e-01 -7.79928565e-01
7.37982452e-01 5.60725152e-01 3.61074239e-01 -9.80520129e-01
3.54466885e-01 5.95415831e-02 -2.64166653e-01 -8.67940009e-01
-1.40516424e+00 -5.49141467e-01 -1.05705905e+00 3.42077285e-01
1.17776108e+00 -1.14910972e+00 -5.17761946e-01 4.73460704e-01
-1.45957005e+00 -4.67250705e-01 -5.03284447e-02 1.57539696e-01
-2.49225870e-01 4.67938691e-01 -8.73112261e-01 -7.32506037e-01
-5.13791323e-01 -1.10140562e+00 1.50757647e+00 -3.95674109e-02
-1.04555391e-01 -1.62175739e+00 -1.18310668e-01 4.66498524e-01
2.22110748e-01 4.40856405e-02 1.35596442e+00 -9.72657621e-01
-1.09601051e-01 2.80977815e-01 -2.72217631e-01 5.15290737e-01
1.41560525e-01 -4.95443612e-01 -1.05836046e+00 -2.47762010e-01
-2.87060943e-02 -8.08207750e-01 1.11663973e+00 4.96848375e-01
8.42510581e-01 -2.80970577e-02 -7.84744546e-02 4.87146437e-01
1.26865494e+00 -1.39448196e-01 1.21239059e-01 3.47254902e-01
8.93657446e-01 8.31461072e-01 7.39284456e-01 1.12069830e-01
1.01742220e+00 5.18248618e-01 3.07951033e-01 -1.92819074e-01
-1.96239755e-01 6.59486232e-03 6.62321866e-01 1.18875039e+00
2.70509452e-01 -2.19947770e-01 -1.09045362e+00 7.46942163e-01
-2.07195592e+00 -3.73141676e-01 -2.30203599e-01 1.57699716e+00
1.14272904e+00 1.92906290e-01 5.76814860e-02 -2.46058241e-01
6.83882058e-01 3.84248286e-01 -5.28737962e-01 -7.90574908e-01
-4.04537261e-01 4.67380553e-01 3.47104251e-01 4.56124485e-01
-1.35407364e+00 1.75915098e+00 5.23164272e+00 9.54105139e-01
-8.78412127e-01 6.30899608e-01 4.51820016e-01 4.56903458e-01
-4.29140478e-01 -1.08408593e-01 -1.36994684e+00 6.23895787e-02
1.17402732e+00 -1.92116648e-01 -9.75933149e-02 9.66477215e-01
1.90075815e-01 -7.15817660e-02 -9.16568637e-01 6.32858515e-01
3.87412198e-02 -1.01148176e+00 1.27811253e-01 -2.02821985e-01
2.88263500e-01 4.26732957e-01 -2.23896831e-01 7.71893442e-01
6.62958860e-01 -9.34290290e-01 6.36806130e-01 -2.82459438e-01
8.65604162e-01 -5.61886072e-01 8.60619366e-01 5.76638877e-01
-1.53228807e+00 -1.02864310e-01 -5.90102613e-01 -2.09630430e-01
6.18352115e-01 6.12251699e-01 -5.97458839e-01 5.70361078e-01
3.75636190e-01 9.86358643e-01 -6.66734278e-01 2.30828106e-01
-8.01386893e-01 9.26611602e-01 -5.71174882e-02 -1.35028422e-01
5.36217988e-01 4.33153771e-02 3.75674337e-01 1.59672379e+00
-3.10627967e-01 -1.23343535e-01 7.96703696e-01 6.93412423e-01
5.42529337e-02 4.80829626e-01 -4.51771826e-01 -3.45026180e-02
4.83055055e-01 1.41166198e+00 -7.48553932e-01 -2.97030061e-01
-6.27615690e-01 6.44296646e-01 7.57274210e-01 6.99714646e-02
-4.88103926e-01 -6.34883940e-02 5.58891416e-01 -2.02542558e-01
2.75320232e-01 -5.33782125e-01 -4.41346347e-01 -1.13455939e+00
-1.97159529e-01 -6.79489911e-01 7.24477828e-01 -4.42055076e-01
-1.53655398e+00 9.15220380e-01 1.14370778e-01 -7.54408658e-01
-2.28284225e-01 -9.50984418e-01 -4.70326573e-01 1.07748294e+00
-2.18873453e+00 -2.03107643e+00 5.14285803e-01 4.67990935e-01
1.18842888e+00 -4.22661096e-01 1.19238901e+00 2.29265273e-01
-7.57339716e-01 6.45874560e-01 -4.40825522e-01 4.20137018e-01
7.57425666e-01 -1.54630435e+00 7.65604377e-01 8.15006554e-01
1.59785837e-01 5.88893890e-01 1.47324249e-01 -6.30949914e-01
-1.12986612e+00 -1.23864448e+00 1.39407563e+00 -4.36467916e-01
9.66115296e-01 -6.82257712e-01 -1.18627477e+00 8.66912067e-01
3.22981894e-01 8.14023167e-02 8.20780456e-01 6.23615146e-01
-3.87849122e-01 -2.20268555e-02 -8.46019506e-01 3.83976668e-01
1.01686823e+00 -4.08038974e-01 -1.09134829e+00 6.86187506e-01
1.32589519e+00 -2.58598149e-01 -6.58268332e-01 2.60877699e-01
7.55057037e-02 -5.97900152e-01 8.08099508e-01 -6.83550775e-01
4.54367697e-01 1.27173394e-01 -1.78873822e-01 -1.35251570e+00
-5.03213346e-01 -5.39031863e-01 1.95697278e-01 1.54048538e+00
7.69987226e-01 -5.92454910e-01 2.92247295e-01 2.54860014e-01
-6.30010843e-01 -6.92832172e-01 -9.80864346e-01 -7.41086066e-01
6.38425052e-01 -7.18458712e-01 4.66714919e-01 7.79615045e-01
-1.33496106e-01 1.02374160e+00 -2.24465877e-01 1.67081743e-01
5.27503967e-01 7.68649727e-02 2.47486189e-01 -1.35840809e+00
-3.01287025e-01 -3.04558814e-01 1.36870623e-01 -1.08676803e+00
1.10628366e+00 -1.37977111e+00 8.77050534e-02 -1.46548510e+00
-1.88168600e-01 -5.21756709e-01 -4.62174296e-01 1.14392507e+00
-4.06824380e-01 -2.04345658e-02 1.07306130e-01 2.62200236e-02
-7.01408982e-01 5.84734619e-01 1.34487176e+00 1.82150647e-01
-2.70692378e-01 5.27521446e-02 -9.73905325e-01 8.72212470e-01
8.42336535e-01 -5.68575263e-01 -2.09104404e-01 -7.87199557e-01
1.92088410e-01 -1.61186874e-01 -8.67116153e-02 -4.64893758e-01
-1.02725059e-01 -9.75746289e-02 -2.27924794e-01 -4.53062892e-01
1.15320489e-01 -4.47407633e-01 -8.61666560e-01 1.56557530e-01
-2.59620309e-01 2.30944872e-01 4.00809139e-01 3.53626192e-01
-3.02869350e-01 -4.97541487e-01 7.82559633e-01 -6.31091535e-01
-1.20285773e+00 1.11188792e-01 -4.60116267e-01 3.44205976e-01
7.11284399e-01 5.96410111e-02 -3.16061050e-01 1.59477159e-01
-8.20174158e-01 7.17969894e-01 -2.44793370e-01 7.75814891e-01
4.35117334e-01 -1.12435353e+00 -9.32524264e-01 2.81576276e-01
2.97556818e-01 2.85871089e-01 1.58572167e-01 4.87919509e-01
-4.83389832e-02 4.77770895e-01 6.44532815e-02 -5.05818367e-01
-1.09264529e+00 5.42096138e-01 6.18293062e-02 -7.83814847e-01
-6.60792172e-01 1.04526520e+00 3.44111353e-01 -9.87082362e-01
1.24576792e-01 -6.38961077e-01 -6.41472459e-01 2.71745563e-01
3.31549644e-01 -1.90939978e-01 1.86519846e-01 -8.50092947e-01
-4.84057814e-01 4.41568941e-01 -2.73094863e-01 -2.04387218e-01
1.61940622e+00 -2.41933525e-01 -3.82361293e-01 7.61105716e-01
8.95091474e-01 7.07484270e-03 -9.97408152e-01 -5.64974189e-01
7.59482801e-01 1.91966057e-01 2.18399859e-04 -9.14991438e-01
-7.91769803e-01 1.16086757e+00 -1.09896913e-01 -4.44471948e-02
8.97791326e-01 3.62808406e-01 1.04997289e+00 5.05600095e-01
4.04197693e-01 -1.25155425e+00 2.53291845e-01 1.15932012e+00
9.60424066e-01 -1.33530486e+00 -3.65134865e-01 -4.42880183e-01
-9.96420920e-01 1.02690959e+00 8.38184178e-01 -2.16560289e-01
8.78455102e-01 4.87482727e-01 4.26947415e-01 -2.39182219e-01
-9.48138475e-01 -6.79078162e-01 2.86737770e-01 6.92563474e-01
9.51699555e-01 7.63635710e-02 -5.62045157e-01 1.04750395e+00
-1.32178038e-01 -4.83416706e-01 2.60842413e-01 7.14110792e-01
-6.42964780e-01 -1.62951171e+00 -4.35176753e-02 -8.91010091e-03
-3.99906427e-01 -3.99778306e-01 -2.12864708e-02 8.36508632e-01
2.33851954e-01 8.88966382e-01 -1.16043568e-01 8.29755589e-02
3.70635599e-01 4.15954709e-01 1.71800241e-01 -1.54412353e+00
-9.17904019e-01 4.62853163e-01 3.64570111e-01 -3.71410936e-01
-6.36600494e-01 -5.94291151e-01 -1.65855718e+00 3.34218293e-01
-1.91246539e-01 1.46908849e-01 4.67529237e-01 1.26879227e+00
1.90425634e-01 6.83371842e-01 3.73862952e-01 -7.74143875e-01
-6.62411094e-01 -1.40849328e+00 -6.02291048e-01 3.38270396e-01
2.58504786e-02 -7.38662958e-01 -1.94406256e-01 -7.43447170e-02] | [10.441217422485352, 9.438273429870605] |
dd0a07c6-3069-4ee4-919a-887097f048b0 | vaxformer-antigenicity-controlled-transformer | 2305.11194 | null | https://arxiv.org/abs/2305.11194v1 | https://arxiv.org/pdf/2305.11194v1.pdf | Vaxformer: Antigenicity-controlled Transformer for Vaccine Design Against SARS-CoV-2 | The SARS-CoV-2 pandemic has emphasised the importance of developing a universal vaccine that can protect against current and future variants of the virus. The present study proposes a novel conditional protein Language Model architecture, called Vaxformer, which is designed to produce natural-looking antigenicity-controlled SARS-CoV-2 spike proteins. We evaluate the generated protein sequences of the Vaxformer model using DDGun protein stability measure, netMHCpan antigenicity score, and a structure fidelity score with AlphaFold to gauge its viability for vaccine development. Our results show that Vaxformer outperforms the existing state-of-the-art Conditional Variational Autoencoder model to generate antigenicity-controlled SARS-CoV-2 spike proteins. These findings suggest promising opportunities for conditional Transformer models to expand our understanding of vaccine design and their role in mitigating global health challenges. The code used in this study is available at https://github.com/aryopg/vaxformer . | ['Javier Antonio Alfaro', 'Diego A. Oyarzún', 'Ajitha Rajan', 'Achille Fraisse', 'Michał Kobiela', 'Aryo Pradipta Gema'] | 2023-05-18 | null | null | null | null | ['protein-language-model'] | ['medical'] | [ 1.58387423e-01 -3.25290263e-01 7.88621828e-02 -3.36214811e-01
-7.10347116e-01 -6.38336897e-01 2.49397010e-01 2.09926665e-01
-1.90287888e-01 1.04437089e+00 1.46655366e-01 -8.12503934e-01
2.38900572e-01 -4.73957360e-01 -1.00686979e+00 -1.05442119e+00
-1.70862213e-01 9.14898217e-01 -2.88328439e-01 -4.96473163e-01
-2.68788040e-01 3.76369327e-01 -9.81373787e-01 4.17954594e-01
1.24594104e+00 4.21249896e-01 5.50552130e-01 7.53430605e-01
1.51097164e-01 3.22716057e-01 -4.59884167e-01 -1.30319446e-02
1.90225109e-01 -5.11926591e-01 -2.87325561e-01 -8.09383154e-01
-9.50635299e-02 -3.58152568e-01 2.39330038e-01 7.69175291e-01
3.85768831e-01 -5.95465116e-02 8.48993540e-01 -1.01383531e+00
-9.12842095e-01 -2.65542581e-03 -2.29390085e-01 2.69564360e-01
1.40517861e-01 4.35380578e-01 8.59597504e-01 -7.94386625e-01
9.93873477e-01 1.05319965e+00 8.31519365e-01 8.06675076e-01
-1.49634671e+00 -6.87820494e-01 -1.94725260e-01 4.33113426e-02
-1.25997484e+00 -4.48988006e-02 3.69890422e-01 -1.01708269e+00
1.55286705e+00 1.24711975e-01 7.08827913e-01 1.15021217e+00
1.10819542e+00 5.97391367e-01 1.05297267e+00 1.76832944e-01
2.92644471e-01 -3.80895346e-01 5.38036600e-02 7.54200995e-01
4.26469356e-01 5.11699378e-01 3.36727756e-03 -9.70978975e-01
5.20095408e-01 6.06136262e-01 -3.39849353e-01 -3.11617851e-01
-7.79452562e-01 1.44811416e+00 1.74055204e-01 2.17840523e-01
-1.07788980e+00 -2.29510635e-01 4.75453958e-02 9.49927419e-02
4.34367448e-01 3.38120729e-01 -1.05316234e+00 1.50433511e-01
-9.42205489e-01 6.33311927e-01 3.48195583e-01 3.71506184e-01
6.17045343e-01 4.56306100e-01 -1.52936846e-01 3.86118412e-01
4.02933776e-01 9.36241388e-01 -1.74465366e-02 -4.22790080e-01
-4.30511594e-01 2.48149931e-01 2.84050882e-01 -5.04121482e-01
-4.68967766e-01 -7.04617649e-02 -5.47287464e-01 -7.28503987e-02
-1.99235767e-01 -5.02698362e-01 -1.05383682e+00 1.99827838e+00
5.53420305e-01 8.97585824e-02 2.72914916e-01 7.12298155e-01
7.12920606e-01 1.22726977e+00 5.64539611e-01 -4.59215969e-01
1.53545392e+00 -4.40104365e-01 -6.05606139e-01 2.39762306e-01
3.67809653e-01 -5.25290906e-01 4.23511088e-01 -5.16027212e-02
-6.30213737e-01 -3.34764481e-01 -8.86527002e-01 3.12688231e-01
-2.21200481e-01 -2.36966401e-01 4.87874985e-01 4.73558545e-01
-1.04405594e+00 3.45987469e-01 -1.34751308e+00 -2.62890100e-01
1.74833298e-01 3.63624722e-01 -2.24241987e-01 9.71146747e-02
-1.48318672e+00 1.02334189e+00 4.23350126e-01 -4.89292778e-02
-1.27665901e+00 -1.07569134e+00 -8.08732271e-01 -6.20317906e-02
-2.77282089e-01 -9.68028307e-01 9.69607592e-01 -5.26066482e-01
-1.15280259e+00 5.36431134e-01 -3.68180752e-01 -6.61166131e-01
-3.30239236e-01 -1.22794062e-01 -3.77592355e-01 6.01507947e-02
-1.68063715e-01 8.12679291e-01 6.22973323e-01 -1.10573196e+00
-5.01522198e-02 -3.86784136e-01 -4.36392039e-01 -4.65980262e-01
6.71037793e-01 4.03395712e-01 4.90454644e-01 -8.37220252e-01
-8.45952392e-01 -1.05875921e+00 -4.56025630e-01 -5.59401870e-01
3.34160589e-02 -3.36959869e-01 5.39373934e-01 -1.03743029e+00
6.27984583e-01 -1.46178734e+00 2.45608360e-01 -2.39164867e-02
2.70137668e-01 8.12627077e-01 -4.56932902e-01 9.01172578e-01
-9.69178528e-02 -1.62691679e-02 -5.75691700e-01 4.39929307e-01
-3.52560490e-01 5.91241196e-02 -4.86283094e-01 5.62602341e-01
6.77811086e-01 1.24320042e+00 -7.42626071e-01 3.48403491e-02
1.82894934e-02 1.32231474e+00 -8.04177284e-01 6.64299011e-01
-8.84412885e-01 8.20729256e-01 -6.95630372e-01 6.85350001e-01
1.00806928e+00 -3.26581329e-01 6.46919668e-01 -2.75348080e-03
-4.02249619e-02 -4.99542952e-02 1.10669732e-01 9.62973058e-01
2.79239774e-01 1.55575782e-01 5.28132953e-02 -5.27881563e-01
8.08963776e-01 3.22969794e-01 7.14854777e-01 -6.67947829e-01
4.07614917e-01 -2.25342423e-01 1.95945017e-02 -3.25126112e-01
5.68755306e-02 -6.00168347e-01 2.91038692e-01 1.72567770e-01
-3.52756791e-02 3.69109035e-01 -6.67322502e-02 6.70462474e-02
7.53040075e-01 4.20222312e-01 1.22599468e-01 -5.21562397e-01
3.57219815e-01 5.19095182e-01 9.98521924e-01 2.95748919e-01
-3.68459582e-01 5.07169999e-02 6.01380803e-02 -6.36788309e-01
-1.10295689e+00 -1.24976325e+00 -3.61611009e-01 1.18070149e+00
-4.61234748e-01 -3.92769307e-01 -9.60158885e-01 -1.93851024e-01
-4.76526432e-02 7.55783916e-01 -6.20144904e-01 7.79391974e-02
-6.10824704e-01 -9.90190804e-01 6.55292213e-01 3.01064521e-01
-3.16557080e-01 -9.81785715e-01 -6.84975922e-01 3.37429076e-01
-3.00537646e-01 -5.26755750e-01 -5.24610877e-01 3.26865329e-03
-7.70944238e-01 -1.31827724e+00 -9.19471323e-01 -9.52065706e-01
5.60401857e-01 -7.45715201e-02 8.83845568e-01 -8.47910345e-02
-1.77742422e-01 1.36453822e-01 -3.38592350e-01 -6.79189384e-01
-7.64018118e-01 -4.86154258e-01 5.07111609e-01 -3.35194856e-01
9.86112177e-01 -1.27764076e-01 -9.59588468e-01 -8.86467174e-02
-9.38796878e-01 -9.38466638e-02 4.29625690e-01 9.21734273e-01
1.11239004e+00 -4.38502014e-01 5.46990156e-01 -6.04653716e-01
7.54437745e-01 -7.55944490e-01 -9.41352606e-01 2.63045996e-01
-7.52058029e-01 2.74296790e-01 7.25881994e-01 -3.12257141e-01
-1.03386164e+00 -7.25616515e-02 -7.04370499e-01 -2.09954664e-01
1.93997212e-02 5.42904973e-01 3.54579419e-01 2.51032203e-01
3.73460174e-01 5.02847016e-01 3.87046814e-01 -4.98852432e-01
1.72392115e-01 4.49506253e-01 1.61264956e-01 -4.59298939e-01
5.36819458e-01 2.94520199e-01 -2.41187274e-01 -9.32727814e-01
-1.48869812e-01 -2.95125455e-01 1.31930619e-01 -2.15070741e-03
1.33778024e+00 -1.10052025e+00 -1.06240427e+00 4.84666646e-01
-1.19413960e+00 -3.83469909e-01 4.79422241e-01 4.39277500e-01
-4.47700173e-01 1.80793211e-01 -1.01643395e+00 -6.10325158e-01
-1.02745938e+00 -1.46199250e+00 1.05797124e+00 6.25669956e-03
-2.97010869e-01 -8.44923913e-01 1.15563345e+00 5.68134248e-01
7.15698540e-01 6.21973336e-01 1.22515035e+00 -8.36933434e-01
-5.02548635e-01 3.22879046e-01 1.71142876e-01 3.26049149e-01
1.65526918e-03 7.07022622e-02 -3.98944110e-01 -8.10313582e-01
1.45660222e-01 -3.58505607e-01 9.08720434e-01 9.02064681e-01
2.08384827e-01 -5.47529995e-01 -2.34438151e-01 5.73228776e-01
1.54823411e+00 7.48286605e-01 5.44080377e-01 -1.54385954e-01
4.52393502e-01 4.83767211e-01 3.14945519e-01 2.96944171e-01
4.16538596e-01 4.12473798e-01 1.18744612e-01 3.06851435e-02
6.00405276e-01 -3.47876221e-01 5.34100592e-01 9.73744631e-01
-1.86473519e-01 -2.83595353e-01 -8.94099891e-01 3.91100943e-01
-1.33357108e+00 -1.05372429e+00 9.91922915e-02 1.67698872e+00
9.41651583e-01 -6.42082095e-01 1.18361764e-01 -9.24694419e-01
4.78109866e-01 1.86817437e-01 -5.78103423e-01 -7.37461090e-01
-2.79932827e-01 5.57927668e-01 2.42043942e-01 6.97699189e-01
-8.05419683e-01 8.85979414e-01 6.75492048e+00 5.46345234e-01
-1.29347908e+00 1.80346489e-01 5.12798190e-01 1.62109688e-01
-7.70975053e-01 5.62370084e-02 -6.25675917e-01 3.88194382e-01
1.54694068e+00 1.21400706e-01 3.82494718e-01 5.95722437e-01
4.16189492e-01 5.71644366e-01 -4.33768332e-01 4.38761026e-01
-6.91826418e-02 -1.86761069e+00 8.79352167e-02 7.50601143e-02
6.56837046e-01 7.02710748e-01 -8.12292993e-02 3.14693749e-01
4.56737816e-01 -9.81376112e-01 4.82270904e-02 5.87214053e-01
8.36367369e-01 -9.00822401e-01 7.54078269e-01 1.80644765e-02
-1.09372151e+00 5.18420160e-01 -2.25493804e-01 3.63328516e-01
5.49670577e-01 2.49832481e-01 -9.75003481e-01 9.11976174e-02
4.47438031e-01 3.76197360e-02 8.98396820e-02 3.57370019e-01
4.41642627e-02 7.03300238e-01 -3.34223025e-02 -5.56987487e-02
1.71795547e-01 -4.37121183e-01 4.86070424e-01 1.27783060e+00
1.19619459e-01 3.51735771e-01 3.55228364e-01 1.01396966e+00
2.34146059e-01 1.97336562e-02 -3.85194182e-01 -6.52262032e-01
2.79609859e-01 5.15138030e-01 -2.11444795e-01 -2.35946104e-01
-2.30534330e-01 7.51200318e-01 4.65803519e-02 5.43589830e-01
-9.77161467e-01 -1.76067740e-01 1.48955369e+00 1.07220843e-01
9.06830847e-01 -1.77439362e-01 6.25144362e-01 -1.01325166e+00
-4.84688967e-01 -1.24747467e+00 2.83977687e-01 -8.45826864e-01
-1.09681213e+00 9.73041952e-01 7.46606216e-02 -6.42099798e-01
-4.93524075e-01 -8.53625536e-01 -4.35082972e-01 9.20697570e-01
-1.19118321e+00 -1.27457166e+00 4.94742841e-01 3.11251521e-01
3.81721944e-01 -1.87245131e-01 1.16662240e+00 -1.30187213e-01
-3.38756114e-01 2.67506748e-01 6.91658139e-01 -1.71202198e-01
2.62269169e-01 -5.29490113e-01 7.58446634e-01 7.22409129e-01
-3.80750537e-01 1.21709323e+00 1.27619338e+00 -1.26690650e+00
-1.67364383e+00 -1.29127002e+00 1.11645806e+00 -6.66313887e-01
4.31361139e-01 -2.59138227e-01 -1.03460586e+00 8.57506812e-01
4.59428996e-01 -8.04043710e-01 1.26727235e+00 -3.55578274e-01
-5.52390277e-01 4.17310357e-01 -1.49550796e+00 5.02479851e-01
4.15926129e-01 -6.23443484e-01 -7.37931192e-01 2.96352565e-01
1.28066337e+00 7.32029881e-03 -9.82090890e-01 7.34438539e-01
6.82966113e-01 -6.76326752e-01 9.82239664e-01 -1.27448845e+00
4.65503074e-02 -5.64641297e-01 -2.26571485e-01 -1.25462830e+00
-7.55785525e-01 -5.06588697e-01 -2.43471786e-01 3.85779351e-01
4.60638553e-01 -8.04332197e-01 4.83845592e-01 1.84123740e-01
9.72939003e-03 -1.03062129e+00 -9.03704047e-01 -6.26146019e-01
4.08561409e-01 -3.39789875e-02 7.30898201e-01 7.28301883e-01
-5.08070111e-01 1.40477985e-01 -6.90852702e-01 1.52793959e-01
4.95572448e-01 2.06610665e-01 3.24343979e-01 -1.14450002e+00
-4.90931869e-01 -1.09321708e-02 -6.55950829e-02 -7.10876346e-01
1.42239749e-01 -8.02732468e-01 -6.71976805e-02 -1.42740428e+00
3.83838326e-01 1.78217009e-01 -6.03472054e-01 3.60416263e-01
-2.37266779e-01 -2.86346544e-02 6.52545616e-02 -1.96885224e-03
-2.74712332e-02 8.07532966e-01 8.52194726e-01 -3.22231442e-01
-2.81992227e-01 -4.19192582e-01 -7.27830887e-01 2.00420156e-01
1.18858337e+00 -6.56763315e-01 -4.79806453e-01 -2.43216589e-01
1.55423582e-01 9.54031050e-02 1.03514664e-01 -3.86723816e-01
-4.13187385e-01 -7.97434568e-01 3.09693962e-01 -9.93698418e-01
3.35995376e-01 -3.99394482e-01 9.03592348e-01 1.06857157e+00
2.66176194e-01 6.51995599e-01 4.62011695e-01 5.06294906e-01
2.27277100e-01 5.46438336e-01 7.00440168e-01 5.97954057e-02
-2.91627198e-01 5.31990349e-01 -1.11262214e+00 5.84283508e-02
8.09737921e-01 -1.17615022e-01 -4.23372895e-01 2.92894710e-02
-3.79668683e-01 1.30083233e-01 6.93691254e-01 5.56457341e-01
9.20202672e-01 -9.72623229e-01 -1.22136140e+00 5.85202873e-01
1.87708944e-01 -7.30916917e-01 7.20201373e-01 6.97888017e-01
-1.08425117e+00 9.40511167e-01 -5.45575202e-01 -6.21767342e-01
-1.23506606e+00 1.14756095e+00 2.28196219e-01 -2.90013343e-01
-2.46099949e-01 7.59994745e-01 5.09985685e-01 -5.37139833e-01
-3.41578901e-01 -8.50265324e-02 -2.22396571e-02 -2.98663586e-01
7.37481177e-01 1.86529271e-02 -7.63228312e-02 -1.03031707e+00
-7.65727699e-01 3.36310506e-01 -1.23037741e-01 5.44134498e-01
1.58053100e+00 4.63380307e-01 -3.31860900e-01 -1.56562150e-01
9.99605298e-01 -6.48592338e-02 -1.10227752e+00 2.38651127e-01
-2.92525947e-01 2.66040862e-01 -1.94577366e-01 -7.86188960e-01
-7.26295114e-01 2.83075601e-01 1.08604515e+00 -5.73131621e-01
9.59602475e-01 -4.53484058e-02 1.38849890e+00 2.16026053e-01
5.54171979e-01 -5.01078367e-01 -6.10733509e-01 6.51015818e-01
8.38325799e-01 -1.07189095e+00 -3.15932482e-01 3.29785556e-01
-8.74959111e-01 3.49904001e-01 1.05201334e-01 -2.52464917e-02
6.95888102e-01 1.77256599e-01 4.33285534e-01 -5.11247396e-01
-1.06700051e+00 5.10785170e-02 2.39079371e-01 1.00080121e+00
6.14739656e-01 6.77081764e-01 -5.93426406e-01 5.35150766e-01
9.43476632e-02 4.66494635e-02 -9.73012298e-02 9.87060189e-01
-4.41195041e-01 -1.49515414e+00 -3.49241674e-01 3.81724566e-01
-5.04623055e-01 -5.11781394e-01 -4.97231483e-01 1.39623553e-01
1.92653656e-01 8.92181933e-01 5.31056197e-03 -5.01566470e-01
-1.36810616e-01 3.82053167e-01 3.52131635e-01 -1.42845616e-01
-7.40076721e-01 2.49576718e-01 -6.60612285e-02 -3.26812893e-01
-4.54923034e-01 -5.07353246e-01 -1.29872000e+00 -6.63804352e-01
9.72551554e-02 4.80790079e-01 5.46019197e-01 5.96015513e-01
1.15037870e+00 2.00475916e-01 2.08952829e-01 -5.74156821e-01
-2.88325042e-01 -7.64790595e-01 -1.08160801e-01 3.09057593e-01
6.33925080e-01 -5.17385721e-01 1.41446054e-01 4.10706475e-02] | [4.902958869934082, 5.439722537994385] |
0fe29de9-6e46-4196-8179-44dcb18e0db9 | openood-v1-5-enhanced-benchmark-for-out-of | 2306.09301 | null | https://arxiv.org/abs/2306.09301v2 | https://arxiv.org/pdf/2306.09301v2.pdf | OpenOOD v1.5: Enhanced Benchmark for Out-of-Distribution Detection | Out-of-Distribution (OOD) detection is critical for the reliable operation of open-world intelligent systems. Despite the emergence of an increasing number of OOD detection methods, the evaluation inconsistencies present challenges for tracking the progress in this field. OpenOOD v1 initiated the unification of the OOD detection evaluation but faced limitations in scalability and usability. In response, this paper presents OpenOOD v1.5, a significant improvement from its predecessor that ensures accurate, standardized, and user-friendly evaluation of OOD detection methodologies. Notably, OpenOOD v1.5 extends its evaluation capabilities to large-scale datasets such as ImageNet, investigates full-spectrum OOD detection which is important yet underexplored, and introduces new features including an online leaderboard and an easy-to-use evaluator. This work also contributes in-depth analysis and insights derived from comprehensive experimental results, thereby enriching the knowledge pool of OOD detection methodologies. With these enhancements, OpenOOD v1.5 aims to drive advancements and offer a more robust and comprehensive evaluation benchmark for OOD detection research. | ['Hai Li', 'Yiran Chen', 'Ziwei Liu', 'Yixuan Li', 'Wayne Zhang', 'Kaiyang Zhou', 'Xuefeng Du', 'Yiyou Sun', 'Haoran Zhang', 'Yueqian Lin', 'Haoqi Wang', 'Pengyun Wang', 'Jingkang Yang', 'Jingyang Zhang'] | 2023-06-15 | null | null | null | null | ['out-of-distribution-detection'] | ['computer-vision'] | [-2.38989487e-01 2.03385353e-01 -4.22972262e-01 -2.27183446e-01
-4.58368838e-01 -7.46540904e-01 8.65638673e-01 4.72596377e-01
3.31858806e-02 7.77667537e-02 1.42579526e-01 -1.86883047e-01
-2.00649232e-01 -7.32008636e-01 -1.29058942e-01 -1.87252492e-01
-4.59082693e-01 4.88287538e-01 5.15656114e-01 -5.31180799e-02
2.59918749e-01 4.03707504e-01 -2.17336488e+00 1.84040785e-01
9.87021446e-01 1.04638660e+00 1.73679247e-01 5.50638735e-01
-4.58305150e-01 7.13310897e-01 -1.09095204e+00 -1.08565442e-01
3.38918060e-01 2.79436231e-01 -8.31126571e-01 1.27684757e-01
7.69850910e-01 -4.71983731e-01 -2.78952122e-01 1.06055510e+00
7.64698148e-01 -6.18378446e-02 6.83503270e-01 -1.87028444e+00
-1.02198613e+00 2.50271112e-01 -3.10749531e-01 4.71201748e-01
6.54354751e-01 1.61299676e-01 1.34350646e+00 -7.60414004e-01
1.26919937e+00 1.26083624e+00 2.76653677e-01 2.28525296e-01
-9.17167187e-01 -4.80511039e-01 6.07749112e-02 2.77160823e-01
-1.23751664e+00 -2.93235689e-01 3.47891748e-01 -4.02038097e-01
1.37319541e+00 4.30431426e-01 5.93815982e-01 8.95328164e-01
4.70953062e-02 1.31051695e+00 7.91743636e-01 -6.31062090e-01
3.05214405e-01 3.76675099e-01 5.35003841e-01 6.29116952e-01
4.69302714e-01 -5.94742708e-02 -4.82994884e-01 -2.43256599e-01
2.96792507e-01 -1.22435562e-01 -2.86567044e-02 -6.94703281e-01
-1.06625438e+00 6.56831324e-01 3.99509221e-01 3.06897461e-01
-1.97894901e-01 -4.01459485e-01 5.90285182e-01 5.16848505e-01
5.13757408e-01 4.65973705e-01 -6.39039800e-02 -3.92030776e-01
-7.11867332e-01 5.50431967e-01 1.14057994e+00 1.39751709e+00
5.22714138e-01 -1.76754683e-01 -1.74068421e-01 9.49449062e-01
4.08899456e-01 3.96755546e-01 3.06254655e-01 -1.08355439e+00
4.87672806e-01 9.59181130e-01 2.29979947e-01 -1.03286242e+00
-4.61097777e-01 -3.94250542e-01 -2.83024967e-01 3.04850876e-01
2.88321763e-01 2.56465256e-01 -7.68629372e-01 1.18750024e+00
7.00716138e-01 -5.22184730e-01 -7.82838836e-02 8.83951843e-01
1.15422463e+00 5.46119630e-01 -1.95193931e-01 1.54090166e-01
1.70145249e+00 -1.15908051e+00 -1.13486481e+00 -2.03667566e-01
8.93472552e-01 -9.62957561e-01 1.13871825e+00 4.43741322e-01
-6.79107130e-01 -2.17937216e-01 -1.32694840e+00 -1.70706615e-01
-9.21899498e-01 -3.94163162e-01 8.01846921e-01 8.86471450e-01
-7.85245478e-01 -1.13001056e-01 -5.30234039e-01 -9.10932660e-01
5.20360112e-01 -1.17032848e-01 -1.83478132e-01 -2.49969795e-01
-1.24101663e+00 9.67314541e-01 2.37287164e-01 -5.99057496e-01
-9.95820880e-01 -8.04839551e-01 -9.22948539e-01 -2.62612849e-01
7.10644424e-01 -3.99040759e-01 1.37088025e+00 -4.24119920e-01
-9.74651992e-01 1.19059932e+00 1.33851245e-01 -5.18399835e-01
6.63790822e-01 -3.09622765e-01 -7.90859759e-01 1.48718283e-01
6.36148870e-01 5.92595220e-01 5.47481656e-01 -1.04158807e+00
-8.63768697e-01 -5.13956964e-01 1.17164187e-01 5.46041690e-02
-7.31770933e-01 4.72949475e-01 -6.45019352e-01 -6.83915138e-01
-2.97092032e-02 -6.31690025e-01 1.21670544e-01 1.37858286e-01
-5.73930144e-01 -7.28772044e-01 1.01338327e+00 -3.09608787e-01
1.60499656e+00 -2.16174555e+00 -3.57815832e-01 1.01522140e-01
9.03511226e-01 1.48286819e-01 -1.55984715e-01 6.85985267e-01
1.94021240e-01 1.34886608e-01 2.74081200e-01 -1.25235662e-01
7.53548026e-01 8.85478407e-02 -1.07606038e-01 5.91195405e-01
1.43667581e-02 5.05348146e-01 -9.76150692e-01 -5.86896896e-01
3.86651069e-01 1.96199283e-01 -3.58897448e-01 2.71658629e-01
-1.43723235e-01 -5.45109630e-01 -4.68366355e-01 1.33246815e+00
6.37897909e-01 -3.56153965e-01 -1.87080413e-01 -2.85677286e-03
-5.41224718e-01 3.36893260e-01 -1.62366569e+00 1.40079939e+00
-1.62874982e-01 1.04435682e+00 1.27299577e-01 -6.23975456e-01
9.44911778e-01 1.30393177e-01 5.53547919e-01 -9.03172195e-01
3.15939516e-01 2.61398107e-01 -3.08056891e-01 -6.14476085e-01
8.66397798e-01 7.54633904e-01 -1.58202067e-01 5.64572453e-01
1.24925472e-01 -1.66046411e-01 8.13498497e-01 4.93599951e-01
1.35935986e+00 -3.37413490e-01 6.84035003e-01 -4.70168799e-01
2.06591532e-01 1.44407630e-01 2.10703686e-01 1.03349793e+00
-8.79592478e-01 3.13635439e-01 4.91606414e-01 -5.02171099e-01
-7.41281152e-01 -1.22720397e+00 -7.53797889e-01 1.21336532e+00
4.12124634e-01 -1.02970088e+00 -9.75409925e-01 -7.79655457e-01
4.38142985e-01 7.60706484e-01 -5.04562914e-01 2.70369500e-01
6.60906313e-03 -6.91324532e-01 7.21190751e-01 2.42531136e-01
3.75508577e-01 -9.18633401e-01 -6.97658598e-01 2.29647562e-01
-2.69320905e-01 -1.37912500e+00 -1.99622497e-01 1.33019447e-01
-5.40589392e-01 -1.24690330e+00 -5.71663082e-01 -9.22587335e-01
3.59141827e-01 6.57612920e-01 1.22889984e+00 -6.97327331e-02
-9.16226268e-01 5.53758204e-01 -4.42921370e-01 -9.52754796e-01
-3.60914260e-01 2.97812462e-01 1.15186431e-01 -4.45016146e-01
9.44057465e-01 2.29279678e-02 -2.31253877e-01 6.77135885e-01
-7.18998432e-01 -5.56998730e-01 2.69286305e-01 4.76433486e-01
3.15535277e-01 2.13629484e-01 4.52599198e-01 -6.49800420e-01
9.02528644e-01 -6.68668628e-01 -7.29524791e-01 -1.18072825e-02
-1.09645832e+00 -1.74612865e-01 1.73102662e-01 -1.54340059e-01
-1.04987359e+00 -6.66022897e-01 -5.72069474e-02 2.19835155e-02
-4.90829706e-01 1.54205054e-01 -2.09894180e-01 1.16565071e-01
8.72800946e-01 -2.34013215e-01 -1.71131104e-01 -7.09299505e-01
5.45083642e-01 1.24857032e+00 3.85478288e-01 -4.05143350e-01
6.82128727e-01 5.61707437e-01 -7.13590562e-01 -1.27881789e+00
-6.64151192e-01 -8.60268295e-01 -3.39324027e-01 -3.31066817e-01
5.26016593e-01 -9.26606715e-01 -2.56683946e-01 5.26187241e-01
-9.72110629e-01 -1.98560096e-02 -1.88533545e-01 6.56407699e-02
-1.74924850e-01 4.61015493e-01 -4.66975272e-01 -7.77172208e-01
-2.25716740e-01 -1.14171433e+00 9.85485494e-01 7.36380070e-02
-5.45055330e-01 -8.66221726e-01 1.27080336e-01 4.75450188e-01
2.75013477e-01 8.88386220e-02 5.30931234e-01 -8.41192901e-01
-5.80113113e-01 -5.89435995e-01 -4.83492583e-01 8.87110233e-02
8.56647119e-02 1.97412968e-01 -9.63764846e-01 -1.74133778e-01
-4.40250248e-01 -1.59398243e-01 4.95502144e-01 1.05596289e-01
7.32253909e-01 1.22688755e-01 -6.34541690e-01 3.01952213e-01
1.14181340e+00 5.25711775e-02 4.48219448e-01 1.25415719e+00
2.90073186e-01 8.34810257e-01 9.44272459e-01 7.16896713e-01
8.00340474e-01 6.79914176e-01 7.06944883e-01 -7.83186108e-02
-4.53714639e-01 4.97267656e-02 2.82283962e-01 6.14567339e-01
1.85382739e-01 -5.74057221e-01 -1.07395363e+00 7.34721661e-01
-1.45908999e+00 -7.76662707e-01 -4.34444636e-01 1.93594372e+00
4.76288855e-01 3.94186914e-01 2.41664141e-01 2.41892368e-01
6.96529925e-01 4.27229613e-01 -6.39485061e-01 -3.38149548e-01
-1.43400639e-01 -2.69811451e-01 3.20888847e-01 9.84372869e-02
-1.28262722e+00 7.83272803e-01 7.38718653e+00 7.85543263e-01
-5.27776241e-01 2.55812883e-01 -8.56489092e-02 -5.35935201e-02
-1.65542945e-01 -2.47935534e-01 -1.08704555e+00 4.62344497e-01
6.30604208e-01 -3.14663500e-01 1.89315483e-01 1.25814497e+00
8.87809973e-03 -1.25929117e-01 -9.65776563e-01 8.93399537e-01
2.46493906e-01 -1.25649726e+00 -1.84723988e-01 2.79473811e-01
7.51930058e-01 6.53026462e-01 -1.34216040e-01 4.14203703e-01
3.92656744e-01 -4.65784609e-01 9.60777223e-01 -7.71050006e-02
3.14806402e-01 -4.62913454e-01 8.63019943e-01 1.72208890e-01
-1.31194365e+00 -2.70127237e-01 -2.69585460e-01 -1.26431599e-01
1.42650738e-01 7.59768367e-01 -6.71814919e-01 4.33311433e-01
1.25578690e+00 7.97215819e-01 -7.33456969e-01 1.08065522e+00
-2.43708104e-01 2.12045088e-01 -4.34280723e-01 -2.01413527e-01
1.15463063e-01 2.08442673e-01 8.96946490e-01 1.29212785e+00
-1.17655337e-01 -4.15517151e-01 5.37918627e-01 8.62539411e-01
1.19501362e-02 3.52489889e-01 -7.78585851e-01 -3.02076340e-01
9.09300268e-01 1.16208339e+00 -6.69077754e-01 -4.37266350e-01
-7.31200516e-01 5.68157494e-01 3.18375051e-01 5.30896895e-02
-7.31613755e-01 -9.36650336e-01 1.01027310e+00 8.89358751e-04
1.43624544e-01 9.60945431e-03 -2.11313263e-01 -1.04838443e+00
2.79802643e-02 -9.46931303e-01 7.47499108e-01 -5.91399074e-01
-1.39318776e+00 5.61083853e-01 8.35084356e-03 -1.32102525e+00
-4.27462682e-02 -6.34430230e-01 -3.12340170e-01 2.25253180e-01
-1.61894739e+00 -8.11891556e-01 -5.97529411e-01 1.10773340e-01
8.01190197e-01 -2.11678669e-01 7.97077179e-01 5.97635746e-01
-8.20600569e-01 5.30790746e-01 2.46910125e-01 2.02352763e-03
1.07619715e+00 -1.50868237e+00 6.84563220e-01 7.79492259e-01
-1.78431258e-01 5.37483513e-01 9.81135547e-01 -5.04357874e-01
-1.49369323e+00 -8.71239364e-01 9.11962628e-01 -8.10723305e-01
9.63573456e-01 -4.29226965e-01 -5.71395934e-01 6.50870681e-01
2.41364554e-01 5.00042662e-02 5.13078034e-01 1.80472925e-01
-3.90921980e-01 -1.52196630e-03 -1.19362164e+00 4.93292540e-01
1.26971030e+00 -4.63267118e-01 -9.84129906e-01 3.41065407e-01
6.94187164e-01 -3.81139606e-01 -1.30443251e+00 1.15472302e-01
6.66179240e-01 -1.21787834e+00 1.14082241e+00 -1.94619387e-01
-1.15343286e-02 -2.04205021e-01 -2.80201644e-01 -9.05885041e-01
-1.44198790e-01 -6.58172190e-01 -5.06378293e-01 1.53771341e+00
2.92301625e-02 -9.02475238e-01 5.48446417e-01 3.15849453e-01
-2.99066156e-01 -5.08728206e-01 -5.87986648e-01 -1.24776113e+00
-3.00170481e-01 -6.78521931e-01 8.21902871e-01 9.09705937e-01
3.49174976e-01 1.85861252e-02 6.50002330e-04 2.36704022e-01
1.02607858e+00 -1.31935095e-02 1.17673850e+00 -1.49647653e+00
2.71389037e-01 -6.25538707e-01 -6.96115196e-01 -1.26308763e+00
-1.59248009e-01 -9.52186704e-01 -1.77293375e-01 -1.67742598e+00
1.77256629e-01 -2.85666198e-01 -1.38425827e-01 3.13839495e-01
1.10930622e-01 3.35091531e-01 3.07496578e-01 3.75338554e-01
-1.17147005e+00 3.79129946e-01 1.06391072e+00 -4.02895272e-01
-8.35152492e-02 -2.39763826e-01 -6.17064536e-01 7.96154678e-01
5.84449232e-01 -5.32214701e-01 -3.73723358e-01 -2.41677836e-01
-1.38100823e-02 -6.16614461e-01 1.71593398e-01 -9.22777176e-01
1.10180184e-01 1.37437493e-01 -3.27112749e-02 -9.15079951e-01
-1.58095092e-01 -6.38011396e-01 -5.59581578e-01 5.20978153e-01
-5.09065799e-02 9.64895636e-02 2.56337170e-02 4.61334795e-01
-2.26862684e-01 -2.82856733e-01 4.44808006e-01 -9.95525047e-02
-1.41876924e+00 1.65789366e-01 -7.94805527e-01 5.74171841e-01
1.22530496e+00 -5.48229098e-01 -7.19792306e-01 3.77076454e-02
-4.20310169e-01 6.07564330e-01 5.46610653e-01 1.00689864e+00
1.99410230e-01 -1.19249976e+00 -3.93658608e-01 3.50092620e-01
9.17475998e-01 -3.62615317e-01 -8.64105150e-02 5.13647139e-01
-5.71239829e-01 5.78053176e-01 -2.12026492e-01 -7.17087567e-01
-1.22898936e+00 5.48050761e-01 7.69737586e-02 -1.60918757e-01
-7.92091250e-01 4.29962307e-01 4.54106256e-02 -7.14465916e-01
8.86289716e-01 -2.22929642e-01 -2.04433128e-01 2.89517701e-01
8.78004909e-01 1.10232985e+00 2.12165385e-01 -4.29484636e-01
-5.35376251e-01 1.17342658e-01 -2.25084707e-01 2.61011362e-01
1.20856392e+00 -5.16275764e-01 -1.89781815e-01 5.02767324e-01
1.00790954e+00 -2.65269615e-02 -7.30326235e-01 -2.39508763e-01
3.20343167e-01 -6.31801307e-01 2.55246431e-01 -9.22460437e-01
-5.76374352e-01 4.97683734e-01 8.22525620e-01 9.19638276e-01
5.91286361e-01 3.76605958e-01 7.87447512e-01 4.23953474e-01
5.78993320e-01 -1.44706726e+00 1.82293624e-01 3.63951236e-01
7.73456454e-01 -1.39644814e+00 1.25632718e-01 -6.63940966e-01
-1.77311972e-01 1.02256894e+00 9.23566818e-01 2.44567737e-01
6.85621381e-01 5.41086257e-01 2.76895642e-01 -7.12347507e-01
-6.18511140e-01 -1.97115660e-01 2.20554933e-01 9.20033991e-01
3.45896661e-01 8.21045786e-02 -2.56344229e-01 4.37756360e-01
-2.08233044e-01 -1.96725771e-01 4.37422484e-01 1.18734181e+00
-6.95086062e-01 -1.03714252e+00 -6.24989927e-01 4.84147817e-01
-1.00679278e-01 1.67990774e-01 -3.56495559e-01 1.09260380e+00
2.74805110e-02 1.11420119e+00 9.93848369e-02 -7.62742534e-02
6.79455400e-01 -1.47893935e-01 1.63855627e-01 -5.58099151e-01
-3.02776903e-01 -2.23129883e-01 3.67589533e-01 -8.48550856e-01
-1.69525936e-01 -8.00580382e-01 -1.22320557e+00 -4.16764319e-01
-6.83035731e-01 -1.03731595e-01 8.34061742e-01 6.38902247e-01
7.10457802e-01 6.44536853e-01 3.26120377e-01 -5.56912541e-01
-7.02294648e-01 -8.05424154e-01 -8.60878646e-01 4.02409345e-01
2.25556448e-01 -1.05839396e+00 -5.20928264e-01 -2.26923436e-01] | [9.231376647949219, 3.0950820446014404] |
8044280b-ec43-4962-8a3d-9eb9a18a9dec | conformal-prediction-intervals-with-temporal | 2205.12940 | null | https://arxiv.org/abs/2205.12940v3 | https://arxiv.org/pdf/2205.12940v3.pdf | Conformal Prediction Intervals with Temporal Dependence | Cross-sectional prediction is common in many domains such as healthcare, including forecasting tasks using electronic health records, where different patients form a cross-section. We focus on the task of constructing valid prediction intervals (PIs) in time series regression with a cross-section. A prediction interval is considered valid if it covers the true response with (a pre-specified) high probability. We first distinguish between two notions of validity in such a setting: cross-sectional and longitudinal. Cross-sectional validity is concerned with validity across the cross-section of the time series data, while longitudinal validity accounts for the temporal dimension. Coverage guarantees along both these dimensions are ideally desirable; however, we show that distribution-free longitudinal validity is theoretically impossible. Despite this limitation, we propose Conformal Prediction with Temporal Dependence (CPTD), a procedure that is able to maintain strict cross-sectional validity while improving longitudinal coverage. CPTD is post-hoc and light-weight, and can easily be used in conjunction with any prediction model as long as a calibration set is available. We focus on neural networks due to their ability to model complicated data such as diagnosis codes for time series regression, and perform extensive experimental validation to verify the efficacy of our approach. We find that CPTD outperforms baselines on a variety of datasets by improving longitudinal coverage and often providing more efficient (narrower) PIs. | ['Jimeng Sun', 'Shubhendu Trivedi', 'Zhen Lin'] | 2022-05-25 | null | null | null | null | ['prediction-intervals', 'time-series-regression'] | ['miscellaneous', 'time-series'] | [ 2.05940351e-01 5.24666570e-02 -7.71057308e-01 -5.11217356e-01
-9.04530466e-01 -4.95026797e-01 2.59080321e-01 5.33619106e-01
-1.37629151e-01 9.16823864e-01 3.20828855e-01 -8.08298469e-01
-8.40027273e-01 -8.21955323e-01 -9.46626663e-01 -4.64401275e-01
-6.81177974e-01 5.24527073e-01 -4.51161377e-02 1.69190675e-01
-2.70150572e-01 2.71205813e-01 -1.16439009e+00 1.73816249e-01
1.00180352e+00 1.09546494e+00 -3.32944006e-01 1.79195464e-01
4.80342060e-01 6.04523242e-01 -2.63855100e-01 -5.46262860e-01
2.24591091e-01 -2.33340681e-01 -8.40906441e-01 -2.63093710e-01
-1.03454210e-01 -9.18135792e-02 2.02195823e-01 6.39144361e-01
2.33219564e-01 -1.56488732e-01 7.34081507e-01 -1.38182271e+00
-2.68716812e-01 9.45197999e-01 -4.88042235e-01 3.08564454e-02
2.43592143e-01 -1.37871787e-01 9.98743415e-01 -3.23749065e-01
4.08409685e-01 8.70888352e-01 1.45134902e+00 3.11638862e-01
-1.45900309e+00 -6.78045332e-01 1.67631894e-01 -3.49575073e-01
-1.03243983e+00 -1.73420340e-01 4.89704818e-01 -6.67026699e-01
6.95906341e-01 4.67418045e-01 5.84548891e-01 1.14848125e+00
7.56319463e-01 3.80559921e-01 1.14864123e+00 -1.76664889e-01
4.19298440e-01 -7.26795867e-02 5.10566294e-01 1.21870026e-01
3.08877528e-01 5.34584880e-01 -5.95617071e-02 -7.98516273e-01
6.24629319e-01 4.55966890e-01 -2.25193128e-01 -3.44649941e-01
-1.27522218e+00 9.12563682e-01 1.74473166e-01 3.15651655e-01
-5.60764074e-01 -1.86697260e-01 6.14489615e-01 3.95938009e-01
6.52011395e-01 4.29300934e-01 -8.33553195e-01 5.82703110e-03
-1.05837059e+00 2.56024063e-01 7.61800289e-01 9.17560935e-01
-2.33543874e-03 -4.05233294e-01 -4.31953281e-01 7.95922816e-01
-3.63720477e-01 3.49697977e-01 4.06189859e-01 -8.89886796e-01
3.77406090e-01 4.79112625e-01 2.97978699e-01 -5.61079681e-01
-8.53490114e-01 -7.48984158e-01 -1.25375342e+00 -2.52577662e-01
5.17717779e-01 -3.23330015e-01 -5.34915328e-01 2.22602248e+00
3.61370057e-01 6.94368184e-02 1.16150483e-01 3.45451921e-01
1.68796375e-01 3.62390161e-01 2.21169606e-01 -9.49905753e-01
1.35072923e+00 -3.29945415e-01 -6.07727468e-01 2.53073573e-01
9.30056036e-01 -2.37207815e-01 6.60367370e-01 3.89110655e-01
-1.18128872e+00 -1.36351809e-01 -7.10243642e-01 3.24744463e-01
-9.11791176e-02 -2.31156692e-01 5.26881099e-01 4.52446669e-01
-7.29937792e-01 7.72731602e-01 -8.25006723e-01 -2.18645558e-01
2.92097509e-01 2.34308258e-01 -3.94219846e-01 -3.55589315e-02
-1.48282444e+00 7.36227214e-01 2.93842912e-01 -3.31116289e-01
-2.59970963e-01 -1.33397150e+00 -8.33761871e-01 1.21035725e-01
2.80144244e-01 -8.89302433e-01 1.30491650e+00 -8.15783381e-01
-6.05566800e-01 5.24041295e-01 -2.13865548e-01 -7.77511299e-01
8.61353278e-01 1.24765664e-01 -6.56034708e-01 -2.35877544e-01
2.94638425e-01 7.61741325e-02 3.60310256e-01 -9.15997446e-01
-6.86060846e-01 -4.94349897e-01 4.95570432e-03 -2.85278827e-01
-2.32520804e-01 -1.02412403e-01 -2.40906924e-01 -8.61768305e-01
1.28317410e-02 -9.87903893e-01 -4.08719867e-01 -3.35321903e-01
-4.71241415e-01 -4.23270553e-01 2.36017302e-01 -7.38680601e-01
1.72062504e+00 -1.91781998e+00 -3.64078760e-01 3.66027236e-01
1.28367543e-01 -2.13510334e-01 2.30511025e-01 5.73526442e-01
-5.29552400e-01 8.58457014e-02 -5.84661961e-01 -2.23409444e-01
-2.73368627e-01 1.85503244e-01 -3.92565638e-01 5.78906119e-01
1.74794018e-01 8.41896534e-01 -6.36768818e-01 -5.04547238e-01
-2.27973580e-01 1.98520347e-01 -6.02352500e-01 3.57053690e-02
-1.34296760e-01 3.19319427e-01 -2.88846821e-01 4.55666572e-01
5.47405243e-01 -7.85794735e-01 4.88171697e-01 9.52133909e-02
-7.02815056e-02 2.71472961e-01 -8.85342538e-01 1.07877076e+00
-5.08697987e-01 2.86458880e-02 -4.55502868e-01 -1.26671052e+00
6.91118002e-01 5.22649229e-01 1.03061199e+00 -6.23301983e-01
-1.29450694e-01 1.34946004e-01 -9.25032794e-02 -4.17632222e-01
8.47478583e-02 -4.94225770e-01 -2.05808342e-01 5.17152011e-01
-5.73088408e-01 6.62048995e-01 1.27759337e-01 -2.41531923e-01
1.28687203e+00 -1.07528687e-01 8.34248781e-01 -5.50345957e-01
1.33896798e-01 2.47279331e-02 9.89883304e-01 7.36607552e-01
-1.89163685e-02 4.60435420e-01 8.55507314e-01 -4.16611731e-01
-1.10830224e+00 -1.00201857e+00 -9.41843510e-01 6.48031294e-01
-2.02518597e-01 -1.70710877e-01 -3.75588357e-01 -8.91517341e-01
3.16977739e-01 6.26729250e-01 -1.00755394e+00 4.84734476e-02
-3.89820009e-01 -8.47040892e-01 3.06392729e-01 8.92626524e-01
-1.48285136e-01 -6.51757360e-01 -6.82368398e-01 4.62034225e-01
-1.29433244e-01 -8.40501666e-01 -4.45810556e-01 3.75223607e-01
-1.12273109e+00 -1.31681323e+00 -8.54490519e-01 -5.50765455e-01
5.50471425e-01 -1.31947726e-01 1.37929475e+00 -1.65749595e-01
2.96889156e-01 2.40207538e-01 -1.42073974e-01 -4.14569169e-01
-4.75963801e-01 -1.22786254e-01 1.26315400e-01 -2.36656889e-01
2.53237844e-01 -6.69562936e-01 -6.49631381e-01 5.06147087e-01
-9.44419026e-01 6.23348951e-02 3.95231575e-01 1.12563837e+00
8.14480126e-01 1.39484793e-01 1.09848893e+00 -1.21292734e+00
4.60958600e-01 -8.87165666e-01 -7.53634036e-01 4.58644569e-01
-1.06204331e+00 4.83157597e-02 7.53224492e-01 -6.77859068e-01
-5.79771280e-01 -1.59244448e-01 1.44114671e-02 -3.61240745e-01
2.12432873e-02 1.02291453e+00 2.20914081e-01 5.84712982e-01
5.01578689e-01 4.25108410e-02 1.85836881e-01 -3.18282574e-01
-2.28881180e-01 5.17208517e-01 3.88614327e-01 -4.99567509e-01
3.55817467e-01 3.78959328e-01 3.36788893e-01 -2.62683898e-01
-8.09088290e-01 -5.16780317e-01 -3.53331834e-01 8.89278576e-02
3.53539854e-01 -9.49056804e-01 -1.16503954e+00 -1.64817035e-01
-7.33389139e-01 -2.51962930e-01 -3.42834622e-01 6.10920608e-01
-6.20022893e-01 6.67883158e-02 -4.19863284e-01 -1.02977073e+00
-4.68500853e-01 -9.13517237e-01 8.54865849e-01 -4.23995465e-01
-6.50658727e-01 -1.37680209e+00 3.15620095e-01 -1.97821915e-01
7.83567578e-02 7.87791550e-01 1.31134880e+00 -9.39102292e-01
1.04975522e-01 -3.55422705e-01 -1.94368213e-01 1.09562851e-01
1.42336667e-01 -1.48438781e-01 -6.41832113e-01 -4.17303890e-01
3.46797565e-03 9.11275670e-03 7.24993408e-01 9.01325405e-01
1.50968802e+00 -8.00368190e-01 -6.37820899e-01 4.03064132e-01
1.36012924e+00 2.85995632e-01 2.18374312e-01 1.61532819e-01
7.67851174e-02 9.65355873e-01 7.10371375e-01 7.01478004e-01
6.40001237e-01 7.20412433e-01 1.23551473e-01 -2.67873019e-01
5.98370552e-01 -2.23547012e-01 -7.86501318e-02 3.62051189e-01
1.48933232e-01 -2.57812105e-02 -1.08016074e+00 8.15803111e-01
-1.99018896e+00 -1.03952491e+00 -1.47352412e-01 2.74417281e+00
1.21539569e+00 2.25020632e-01 7.49390781e-01 4.73285377e-01
5.24149597e-01 -4.09380436e-01 -6.95563018e-01 -2.10346639e-01
-3.11959046e-03 -8.75439644e-02 5.76145828e-01 2.24531978e-01
-1.10107529e+00 -2.93890834e-01 6.55274296e+00 6.65879548e-01
-1.04252040e+00 1.46835670e-01 1.19423866e+00 2.08389014e-02
-4.90086734e-01 -2.09067717e-01 -4.66102868e-01 6.47943497e-01
1.21583521e+00 -2.81911463e-01 -1.12026215e-01 7.94735670e-01
3.25260371e-01 7.28676394e-02 -1.46223438e+00 6.36704922e-01
-4.89303142e-01 -1.30663681e+00 -4.65170473e-01 2.52445698e-01
8.72228801e-01 -1.31347075e-01 1.23269238e-01 3.45791370e-01
3.77959311e-01 -1.06159186e+00 4.47718650e-01 4.85996455e-01
1.19035125e+00 -8.04247975e-01 7.84368932e-01 5.45777678e-01
-9.01575685e-01 -3.93603295e-01 1.29632419e-02 1.28649652e-01
1.15887560e-01 1.12012434e+00 -6.51380122e-01 8.83919477e-01
6.74378097e-01 9.05643642e-01 -4.64152694e-02 9.94875491e-01
6.38216257e-01 7.43206143e-01 -4.96217042e-01 3.66983622e-01
-4.75965030e-02 -4.65840362e-02 1.21133998e-01 9.97980952e-01
5.88856936e-01 1.57883242e-01 9.11968127e-02 6.39043331e-01
8.68379250e-02 7.43420720e-02 -7.03723967e-01 3.06810170e-01
4.29684639e-01 5.79823971e-01 -3.15906942e-01 -3.55482548e-01
-5.26071787e-01 1.11358292e-01 3.74546871e-02 2.37045541e-01
-9.91017342e-01 1.86646342e-01 4.42249984e-01 4.06718791e-01
8.69303793e-02 2.49348685e-01 -6.55483782e-01 -8.80477130e-01
2.06686035e-01 -9.07056212e-01 1.00728619e+00 -2.35462442e-01
-1.68709564e+00 6.00766003e-01 2.63787627e-01 -1.69850039e+00
-5.76192021e-01 -2.80477703e-01 -2.39556521e-01 8.41425776e-01
-1.45984590e+00 -1.20950902e+00 1.86204597e-01 6.44683719e-01
1.12847671e-01 3.59720945e-01 8.10586035e-01 2.14678541e-01
-5.25634766e-01 8.28420579e-01 3.48019242e-01 -1.65047839e-01
6.53692186e-01 -1.08840334e+00 1.41531944e-01 5.53634167e-01
-4.45597291e-01 7.32727706e-01 5.60221195e-01 -9.32528138e-01
-1.02311587e+00 -1.43833828e+00 1.24904525e+00 -5.28515995e-01
5.27162433e-01 -2.53494065e-02 -1.04162598e+00 8.04188311e-01
-3.93105388e-01 6.79772347e-02 8.48221421e-01 7.41343677e-01
-3.98494244e-01 -3.08484048e-01 -1.10256577e+00 3.65758359e-01
9.39839959e-01 -2.71330863e-01 -4.19511795e-01 4.20186609e-01
8.42333078e-01 -2.91110486e-01 -1.60371578e+00 1.03411341e+00
8.95735025e-01 -9.14114535e-01 9.23262894e-01 -8.75311017e-01
6.54659748e-01 2.56045461e-01 1.00375600e-02 -9.42789793e-01
-3.92655760e-01 -6.05410278e-01 -7.30620474e-02 9.75388229e-01
6.83813035e-01 -8.71303260e-01 3.65510762e-01 8.16191733e-01
1.49303302e-01 -1.27849579e+00 -1.06222916e+00 -1.28028977e+00
5.07857025e-01 -7.18864799e-01 1.04177964e+00 1.17459154e+00
4.51685727e-01 1.53985862e-02 -5.24354339e-01 3.06412369e-01
4.38176572e-01 5.51352620e-01 2.76430666e-01 -1.49445677e+00
-3.90776515e-01 -3.54244977e-01 4.70021889e-02 -5.26021123e-01
-3.97233255e-02 -5.81645012e-01 -1.89046264e-01 -1.17372227e+00
4.38353866e-01 -9.55548823e-01 -5.55924416e-01 5.80840528e-01
-1.65096894e-01 -1.30278319e-01 -3.95947039e-01 3.39190960e-01
-3.15107912e-01 3.04943502e-01 9.26765561e-01 7.15404600e-02
-3.80021900e-01 7.33802199e-01 -5.67140996e-01 4.49468076e-01
7.16546714e-01 -5.98392487e-01 -4.94609624e-01 2.22238526e-01
3.06326717e-01 9.16221797e-01 2.41516396e-01 -7.70969212e-01
-9.02116746e-02 -4.27263141e-01 2.94092655e-01 -6.62984371e-01
-1.21840164e-01 -1.15786564e+00 6.83445096e-01 7.52979159e-01
-6.62264884e-01 3.66014928e-01 1.74618959e-01 6.37691617e-01
-1.00260310e-01 1.88113779e-01 6.40395463e-01 3.46512735e-01
7.04085156e-02 4.57240850e-01 1.72909394e-01 1.14322245e-01
1.16056705e+00 -9.76421461e-02 -1.98688194e-01 -2.91483879e-01
-6.14013910e-01 4.10368055e-01 3.23424190e-01 2.66562283e-01
2.38387391e-01 -1.39018941e+00 -7.25516856e-01 1.98926613e-01
5.36228359e-01 -2.17651054e-01 3.13810974e-01 1.36700714e+00
2.21420869e-01 7.21675217e-01 2.52227664e-01 -7.28935003e-01
-1.09021199e+00 1.20092177e+00 2.02586994e-01 -6.67733610e-01
-7.32600689e-01 1.95196360e-01 5.64971566e-01 -2.23948047e-01
1.96047783e-01 -7.32463658e-01 -2.28371918e-01 1.46860315e-03
5.27481019e-01 2.99030304e-01 1.80754662e-01 -3.16139728e-01
-3.40709925e-01 3.47366810e-01 2.08210871e-01 9.88299623e-02
1.37481737e+00 -8.22300613e-02 9.85359475e-02 6.29485071e-01
1.15264738e+00 -2.17645317e-01 -1.12889552e+00 -4.92926717e-01
4.36928049e-02 6.60282839e-03 -2.89481610e-01 -9.78407025e-01
-7.00132668e-01 4.00961488e-01 2.93884844e-01 6.46147907e-01
1.36114192e+00 -5.87303340e-02 6.35434449e-01 -1.79778352e-01
3.79202694e-01 -6.54938638e-01 -5.74220181e-01 2.21965641e-01
8.92738938e-01 -1.16183245e+00 -6.30147457e-02 -2.50200748e-01
-6.05691135e-01 7.46084750e-01 1.12977691e-01 1.84438601e-01
9.43677366e-01 3.19264829e-01 -3.67518395e-01 4.48129326e-02
-1.20848572e+00 1.35913625e-01 5.42468250e-01 4.33212012e-01
4.98105228e-01 3.79248738e-01 -5.25393724e-01 1.05581439e+00
-1.74030945e-01 3.57234269e-01 1.79410875e-01 6.80005789e-01
2.65734106e-01 -8.33506644e-01 -3.90686095e-01 8.77991378e-01
-7.51174986e-01 1.42143934e-03 3.25886905e-01 9.39762414e-01
-1.84636787e-02 1.00400543e+00 1.94612175e-01 -2.56217152e-01
4.71786052e-01 1.10046059e-01 3.08683906e-02 -1.69034377e-01
-4.60531145e-01 1.54008344e-01 1.84591159e-01 -5.09629250e-01
-4.10708249e-01 -8.91544819e-01 -9.09774899e-01 -4.91847634e-01
-2.46870786e-01 2.58008927e-01 3.11670065e-01 1.07496834e+00
4.30729955e-01 4.82356280e-01 8.94452512e-01 1.36667520e-01
-8.81609738e-01 -8.42291713e-01 -5.37684619e-01 3.57661903e-01
7.53394783e-01 -5.61194420e-01 -2.11034268e-01 5.96110560e-02] | [7.814303398132324, 5.003946781158447] |
0701be8c-6fea-493f-9026-b4fa65a5e104 | causal-imitation-learning-under-temporally | 2202.01312 | null | https://arxiv.org/abs/2202.01312v1 | https://arxiv.org/pdf/2202.01312v1.pdf | Causal Imitation Learning under Temporally Correlated Noise | We develop algorithms for imitation learning from policy data that was corrupted by temporally correlated noise in expert actions. When noise affects multiple timesteps of recorded data, it can manifest as spurious correlations between states and actions that a learner might latch on to, leading to poor policy performance. To break up these spurious correlations, we apply modern variants of the instrumental variable regression (IVR) technique of econometrics, enabling us to recover the underlying policy without requiring access to an interactive expert. In particular, we present two techniques, one of a generative-modeling flavor (DoubIL) that can utilize access to a simulator, and one of a game-theoretic flavor (ResiduIL) that can be run entirely offline. We find both of our algorithms compare favorably to behavioral cloning on simulated control tasks. | ['Zhiwei Steven Wu', 'J. Andrew Bagnell', 'Sanjiban Choudhury', 'Gokul Swamy'] | 2022-02-02 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [-1.91335287e-02 1.02934606e-01 -9.83162969e-02 2.18458489e-01
-8.30618739e-01 -1.02501571e+00 7.72730768e-01 -1.94498330e-01
-6.44900143e-01 1.17640638e+00 -4.98994105e-02 -1.00451946e+00
-1.95881113e-01 -6.04641855e-01 -9.23599899e-01 -7.60700881e-01
-2.93838203e-01 5.41457772e-01 1.11603059e-01 -7.03722909e-02
6.05628975e-02 2.58346021e-01 -1.20998311e+00 -2.56105363e-01
6.39460742e-01 4.10048097e-01 7.48699531e-02 1.05398750e+00
1.94013372e-01 1.44216168e+00 -7.38875985e-01 2.46854331e-02
4.61132824e-01 -6.78646207e-01 -5.45294344e-01 -6.42636837e-03
-3.83973688e-01 -5.50284564e-01 -5.16108096e-01 1.06789505e+00
3.00060630e-01 4.13532227e-01 6.65338576e-01 -1.20153892e+00
-2.13745236e-01 6.58100903e-01 -3.84496987e-01 1.08636357e-01
3.27833980e-01 7.63988972e-01 4.87555355e-01 -9.10047591e-02
6.43804014e-01 1.35695767e+00 4.60080624e-01 5.80249369e-01
-1.66017330e+00 -7.56548405e-01 2.54610538e-01 -2.31103674e-01
-1.08384991e+00 -4.40714568e-01 4.95981783e-01 -5.97277105e-01
7.98868716e-01 -9.78082791e-03 8.15164685e-01 1.63217902e+00
4.98877853e-01 7.31228411e-01 1.62777090e+00 -2.07343176e-01
7.72960544e-01 -1.01384252e-01 -3.12531948e-01 5.94600320e-01
2.27198228e-01 8.22640538e-01 -1.31420359e-01 -4.92067218e-01
1.27664196e+00 1.62999087e-03 -6.29455522e-02 -3.41527671e-01
-7.62739360e-01 8.64997029e-01 -2.55275071e-01 -9.53052863e-02
-6.89570844e-01 7.11760879e-01 1.58176839e-01 7.86570251e-01
2.63163686e-01 5.68440318e-01 -5.03888965e-01 -7.16626823e-01
-5.32928646e-01 5.41655898e-01 9.82368588e-01 8.18678558e-01
5.55144906e-01 4.17525530e-01 -1.21552251e-01 1.74161643e-01
1.61628112e-01 5.49967408e-01 3.43144774e-01 -1.56161332e+00
2.25743309e-01 -1.18849007e-02 6.65173054e-01 -3.14007550e-01
-3.38285595e-01 -1.05548881e-01 -2.89996505e-01 7.56529272e-01
6.33401632e-01 -1.04136252e+00 -9.05682206e-01 1.86668122e+00
1.12433404e-01 4.57849681e-01 3.45472097e-02 5.36014795e-01
-1.16663486e-01 4.76195157e-01 -6.36497438e-02 -4.11108941e-01
8.10760975e-01 -5.36138952e-01 -5.87149680e-01 -1.53350398e-01
4.99548018e-01 -2.60155380e-01 8.60039771e-01 5.12849271e-01
-1.14065015e+00 -1.51699856e-01 -7.83228874e-01 5.52620053e-01
-7.75405541e-02 -5.22820413e-01 5.63148558e-01 4.91599739e-01
-1.14134502e+00 9.82899368e-01 -1.35751772e+00 -3.52522358e-02
7.50650764e-02 4.24671918e-01 -1.11787535e-01 3.75752717e-01
-9.94862258e-01 8.95366311e-01 1.11300826e-01 -4.38207507e-01
-1.60259938e+00 -4.02442217e-01 -6.76726699e-01 -1.11355364e-01
1.04258680e+00 -5.28539240e-01 1.89078951e+00 -1.01985383e+00
-2.09294677e+00 5.15218265e-03 2.15708036e-02 -5.15839159e-01
7.53243268e-01 -2.41523340e-01 1.13196068e-01 2.55400315e-02
-8.28149468e-02 -6.99639618e-02 1.06296909e+00 -1.15235591e+00
-3.43699604e-01 -9.64059159e-02 1.95301265e-01 2.13524863e-01
3.49008679e-01 -3.59588042e-02 -1.67827442e-01 -4.00442183e-01
-4.50990766e-01 -1.24589682e+00 -7.65706599e-01 -5.08252680e-01
-2.33877450e-01 -1.43194810e-01 4.99318898e-01 -5.25596201e-01
1.03935099e+00 -1.75203085e+00 3.37948442e-01 3.20978254e-01
1.70265973e-01 1.41131520e-01 -7.50913545e-02 7.86052823e-01
-9.01169553e-02 2.39256769e-01 -1.51938181e-02 -3.45719814e-01
-1.10092759e-02 5.36654651e-01 -5.14395475e-01 6.52252316e-01
-3.97741906e-02 9.20608997e-01 -1.24049199e+00 -1.22591801e-01
1.82489187e-01 -2.25065663e-01 -7.16695189e-01 4.06822830e-01
-5.57443798e-01 9.33091640e-01 -7.82645762e-01 2.40090370e-01
-3.75671312e-02 -9.60421711e-02 5.41175783e-01 9.33552325e-01
-3.33468407e-01 5.33831596e-01 -1.16810584e+00 1.31607866e+00
-4.55801964e-01 3.98397982e-01 1.16437830e-01 -8.74343216e-01
2.90811658e-01 6.18013442e-01 5.70802033e-01 -3.68783027e-01
3.82337302e-01 -1.40993491e-01 5.91754206e-02 -4.47198242e-01
1.88986167e-01 -9.87251997e-02 -1.51664227e-01 9.08713698e-01
2.96700038e-02 -2.30318055e-01 -5.73662408e-02 2.09899455e-01
1.74911976e+00 5.14185488e-01 4.74943131e-01 -9.15228799e-02
-2.31464639e-01 2.25770175e-01 6.10143125e-01 1.48220837e+00
-2.05209240e-01 -7.79309124e-02 1.10213566e+00 -1.24071166e-01
-1.29217386e+00 -1.02542031e+00 4.50330734e-01 9.65892792e-01
-2.10476279e-01 -3.56389016e-01 -6.40033782e-01 -4.41870749e-01
2.15918735e-01 8.26666772e-01 -8.27464819e-01 -1.22409895e-01
-3.71996045e-01 -3.14718992e-01 6.25480294e-01 5.64986885e-01
6.79901764e-02 -1.02679062e+00 -7.58974910e-01 4.79408681e-01
2.71046937e-01 -5.90475321e-01 -3.99100602e-01 6.85952246e-01
-8.38180661e-01 -9.41313565e-01 -3.39783221e-01 5.88130355e-02
3.74644786e-01 7.99514130e-02 1.03434217e+00 -2.20000222e-01
-6.63316762e-03 7.30107486e-01 -2.55960703e-01 -5.66923261e-01
-7.49406040e-01 -2.15643317e-01 5.03829598e-01 -5.09521127e-01
1.21163636e-01 -8.25492740e-01 -3.41723591e-01 4.92667221e-02
-6.24620259e-01 -1.85710996e-01 4.79677558e-01 9.12677467e-01
2.60220796e-01 -1.12854894e-02 2.61944443e-01 -8.77357125e-01
1.03455508e+00 -7.64401495e-01 -1.19521749e+00 -1.00517735e-01
-6.46637440e-01 4.98636335e-01 8.45294654e-01 -9.53553021e-01
-7.78068185e-01 2.65879165e-02 4.13208187e-01 -9.09331262e-01
-1.53103724e-01 3.80350620e-01 2.65336514e-01 1.01244815e-01
6.00430965e-01 3.30960006e-01 3.08286875e-01 -4.70772594e-01
4.41792637e-01 3.18691134e-01 3.98619741e-01 -9.22533095e-01
9.63925362e-01 5.29833734e-02 -1.55445099e-01 -6.32156849e-01
-2.39576235e-01 -2.63789874e-02 -2.23365828e-01 -2.05303222e-01
7.01294720e-01 -9.89513874e-01 -1.16283154e+00 2.86787212e-01
-6.94431365e-01 -1.04101861e+00 -4.41211283e-01 6.38118446e-01
-9.59006369e-01 -6.53636549e-03 -7.98521042e-01 -1.32287443e+00
4.58992153e-01 -1.11815536e+00 4.82375711e-01 2.60639012e-01
-2.70520717e-01 -9.50860739e-01 6.23171151e-01 -1.77940160e-01
1.88266158e-01 2.36747526e-02 6.58257246e-01 -7.84095168e-01
-6.40548468e-01 -7.25843161e-02 4.40020919e-01 1.15382865e-01
5.12243249e-02 1.38193876e-01 -6.71840847e-01 -5.27210236e-01
1.51515588e-01 -4.10713017e-01 3.88012618e-01 4.82589722e-01
9.29694712e-01 -5.59529603e-01 -2.57729024e-01 4.02076662e-01
1.11102760e+00 7.86816955e-01 3.90314877e-01 3.17671388e-01
3.36439252e-01 1.73199579e-01 3.80356133e-01 6.64677382e-01
2.55248517e-01 4.96919721e-01 1.88609585e-01 8.64174888e-02
6.16735399e-01 -7.40335882e-01 7.79300570e-01 3.85275036e-01
-4.45723623e-01 -1.78340927e-01 -7.63356209e-01 4.33462620e-01
-2.14475298e+00 -1.23897016e+00 1.89126804e-01 2.38209391e+00
8.80270541e-01 1.57075569e-01 5.59236526e-01 -5.17286956e-01
3.16636324e-01 2.02644635e-02 -9.12057161e-01 -6.22888744e-01
4.69418347e-01 5.08550107e-01 1.09696496e+00 5.88346660e-01
-7.03288794e-01 9.65105712e-01 7.79103231e+00 6.41303420e-01
-7.13875353e-01 1.91667408e-01 2.25921601e-01 -2.92587578e-01
-2.49708027e-01 3.34454924e-01 -3.36068392e-01 5.26713789e-01
1.26871371e+00 -3.81816238e-01 1.13525021e+00 8.14093649e-01
7.39570796e-01 -4.03742343e-01 -1.08292854e+00 6.08619809e-01
-6.37390137e-01 -7.96995759e-01 -5.15745461e-01 4.80281204e-01
8.24593842e-01 9.70917568e-02 2.07000170e-02 7.19384193e-01
1.75270391e+00 -1.11770213e+00 5.75011909e-01 4.12186414e-01
4.60336924e-01 -8.37925971e-01 2.75664866e-01 9.06326771e-01
-9.22578812e-01 -3.23851526e-01 -2.50896424e-01 -5.26312768e-01
-1.22280262e-01 -1.31085575e-01 -8.10288787e-01 1.96538761e-01
2.98743963e-01 2.71053731e-01 -4.19873628e-04 7.01677561e-01
-5.87173820e-01 1.04224980e+00 -2.47693986e-01 -7.72450715e-02
4.23570067e-01 -6.19826674e-01 7.93256879e-01 6.50772870e-01
1.99278891e-01 2.68538833e-01 4.34206575e-01 1.01692080e+00
1.25324741e-01 -4.43264604e-01 -1.17492640e+00 -3.64332348e-01
5.23966074e-01 7.90576458e-01 -4.44541365e-01 -5.03946364e-01
-4.40683454e-01 9.09147084e-01 3.14797938e-01 7.43477404e-01
-9.34367657e-01 8.95673409e-03 1.06170034e+00 -1.15083233e-01
3.42519432e-01 -6.51960313e-01 -4.32931222e-02 -1.28264439e+00
-3.74277472e-01 -1.25028920e+00 -2.85479259e-02 -5.01079381e-01
-8.54160309e-01 -5.74826300e-02 1.66387945e-01 -1.02001929e+00
-1.01011360e+00 -1.67172030e-01 -7.36578345e-01 1.01852298e+00
-7.36560643e-01 -2.20521465e-01 5.03267109e-01 7.82345533e-01
4.65913951e-01 -1.66046277e-01 6.13218963e-01 -2.59333789e-01
-6.67653322e-01 2.35529363e-01 4.80030119e-01 3.46554331e-02
3.71330678e-01 -1.53596151e+00 5.23817480e-01 8.16679180e-01
9.27301869e-02 7.41085529e-01 1.04892778e+00 -8.48077893e-01
-1.72235143e+00 -8.36782157e-01 -8.52297097e-02 -6.81624055e-01
1.12460864e+00 -1.61570266e-01 -7.69812524e-01 1.22166038e+00
7.44183958e-02 -4.50393528e-01 2.89386243e-01 1.04508467e-01
-2.56881602e-02 4.39133584e-01 -7.31886029e-01 9.13127065e-01
9.07222569e-01 -7.86193848e-01 -6.53989971e-01 1.68096036e-01
7.72104979e-01 -3.75524789e-01 -6.53869867e-01 -1.74950287e-01
4.90202904e-01 -7.15606689e-01 5.86690784e-01 -1.27225769e+00
1.69533700e-01 -2.39327610e-01 1.49340972e-01 -1.62852788e+00
-2.53754616e-01 -1.44678390e+00 -1.65326491e-01 7.42146611e-01
2.93498874e-01 -9.05963600e-01 4.80730295e-01 8.48659933e-01
2.18782604e-01 -3.49729866e-01 -8.28357399e-01 -1.13813829e+00
3.68063331e-01 -4.90903497e-01 3.82995367e-01 7.19186544e-01
1.58338472e-01 1.61251381e-01 -6.44835174e-01 2.00629294e-01
6.75846815e-01 -7.42492452e-02 1.01377332e+00 -8.83253753e-01
-9.21463609e-01 -3.10390830e-01 8.48505273e-02 -1.17127180e+00
2.48598412e-01 -3.60019743e-01 4.05932605e-01 -8.70291293e-01
1.68429434e-01 -2.15742171e-01 -3.44933510e-01 6.17195785e-01
-8.31165686e-02 -3.99324298e-01 1.48016736e-01 1.72771826e-01
-5.88624597e-01 5.99240303e-01 1.00005138e+00 3.16111654e-01
-7.16798127e-01 2.44807437e-01 -4.96089667e-01 8.35495472e-01
7.38053381e-01 -8.91069949e-01 -6.25726223e-01 -5.06862737e-02
1.19982585e-01 9.20457840e-01 4.93553579e-01 -8.10047209e-01
2.79118627e-01 -6.57747269e-01 1.28789112e-01 -7.74153545e-02
5.30606247e-02 -5.89943051e-01 4.96425450e-01 7.63154864e-01
-3.98860246e-01 2.24936023e-01 1.85133472e-01 8.45151544e-01
2.47971147e-01 -1.30017906e-01 6.64014280e-01 -3.38262558e-01
-2.99106747e-01 2.17994317e-01 -1.24356532e+00 1.54494569e-01
1.08891761e+00 2.67471641e-01 -2.15831444e-01 -9.44045722e-01
-6.40434504e-01 3.99935991e-01 7.72916675e-01 1.45546287e-01
2.00802535e-01 -9.04834270e-01 -3.34230304e-01 -1.16327792e-01
-5.32603562e-01 -3.27608526e-01 -2.27650687e-01 8.16964746e-01
-1.95788145e-01 3.32680285e-01 -4.95791286e-02 -1.74519554e-01
-8.63316238e-01 7.73095310e-01 4.38943416e-01 -3.08031887e-01
-7.36250579e-01 5.32312214e-01 6.36854470e-02 -2.79742211e-01
3.22754830e-02 -2.50110209e-01 1.49374560e-01 -2.06850663e-01
4.41414982e-01 3.98454666e-01 -5.79201102e-01 1.14371613e-01
-7.04062581e-02 -5.62333921e-03 1.96103558e-01 -8.15981865e-01
1.16551411e+00 -2.35438924e-02 2.21626073e-01 8.78656626e-01
5.89329541e-01 6.60696849e-02 -1.96333110e+00 -2.40655318e-01
-8.64311159e-02 -2.16523111e-01 4.97779958e-02 -5.62061310e-01
-5.40885568e-01 4.69667524e-01 3.41848135e-01 3.93815607e-01
6.77129924e-01 -3.40279967e-01 3.81768376e-01 6.27031684e-01
8.60816121e-01 -1.14187133e+00 -1.18539073e-01 5.45944989e-01
2.90844738e-01 -9.95039642e-01 -2.14847490e-01 4.63231444e-01
-8.30576241e-01 7.22463131e-01 2.57613242e-01 -5.21295130e-01
4.38723177e-01 6.62794530e-01 -3.82135808e-02 -3.19617949e-02
-1.52169919e+00 -4.56263274e-01 -2.68519938e-01 6.12983823e-01
-8.23659152e-02 3.07135880e-01 -9.89592150e-02 3.80023569e-01
9.69469920e-02 2.34241694e-01 7.63152301e-01 1.26942515e+00
-4.43569392e-01 -1.20119977e+00 -4.69268620e-01 4.68446910e-01
-5.26515722e-01 -5.87288588e-02 -3.26107949e-01 1.00922251e+00
-2.80305296e-01 1.04236901e+00 7.74205178e-02 -3.68761152e-01
1.40305683e-01 2.39361450e-01 4.17380661e-01 -5.53987324e-01
-5.40019929e-01 2.73091614e-01 4.21292558e-02 -1.12560713e+00
-5.64855244e-03 -7.47056365e-01 -9.56878662e-01 -5.30591547e-01
4.35330207e-03 3.65041494e-01 4.59123880e-01 1.14573944e+00
1.08561523e-01 6.33039176e-01 8.57996047e-01 -9.52936709e-01
-1.09698987e+00 -8.78017247e-01 -9.02238905e-01 -3.08664329e-02
4.86447632e-01 -8.31701815e-01 -4.69213784e-01 -1.99252039e-01] | [4.1044921875, 2.1130425930023193] |
c60d7608-251f-4bd9-84cf-58df8d4451c7 | universal-instance-perception-as-object | 2303.06674 | null | https://arxiv.org/abs/2303.06674v1 | https://arxiv.org/pdf/2303.06674v1.pdf | Universal Instance Perception as Object Discovery and Retrieval | All instance perception tasks aim at finding certain objects specified by some queries such as category names, language expressions, and target annotations, but this complete field has been split into multiple independent subtasks. In this work, we present a universal instance perception model of the next generation, termed UNINEXT. UNINEXT reformulates diverse instance perception tasks into a unified object discovery and retrieval paradigm and can flexibly perceive different types of objects by simply changing the input prompts. This unified formulation brings the following benefits: (1) enormous data from different tasks and label vocabularies can be exploited for jointly training general instance-level representations, which is especially beneficial for tasks lacking in training data. (2) the unified model is parameter-efficient and can save redundant computation when handling multiple tasks simultaneously. UNINEXT shows superior performance on 20 challenging benchmarks from 10 instance-level tasks including classical image-level tasks (object detection and instance segmentation), vision-and-language tasks (referring expression comprehension and segmentation), and six video-level object tracking tasks. Code is available at https://github.com/MasterBin-IIAU/UNINEXT. | ['Huchuan Lu', 'Zehuan Yuan', 'Ping Luo', 'Dong Wang', 'Jiannan Wu', 'Yi Jiang', 'Bin Yan'] | 2023-03-12 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yan_Universal_Instance_Perception_As_Object_Discovery_and_Retrieval_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yan_Universal_Instance_Perception_As_Object_Discovery_and_Retrieval_CVPR_2023_paper.pdf | cvpr-2023-1 | ['object-discovery', 'referring-expression', 'visual-tracking', 'multiple-object-tracking', 'referring-expression-segmentation', 'video-instance-segmentation', 'referring-video-object-segmentation', 'visual-object-tracking', 'multi-object-tracking-and-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.93212825e-01 4.32367437e-02 -5.58755457e-01 -6.02738678e-01
-9.21862006e-01 -7.77587891e-01 7.00547814e-01 2.40318567e-01
-4.24818844e-01 5.38749754e-01 -1.02657877e-01 -8.76701996e-03
-3.55386026e-02 -6.13611162e-01 -8.71651471e-01 -5.08709550e-01
2.66105741e-01 5.60517073e-01 4.44402933e-01 -8.14110041e-02
2.51387596e-01 1.57287672e-01 -1.94061112e+00 7.13461101e-01
7.02859640e-01 1.36480880e+00 6.50229156e-01 4.54495519e-01
-4.13869053e-01 8.80113840e-01 -5.52880287e-01 -2.97476023e-01
6.37490079e-02 1.56583749e-02 -1.19571841e+00 3.44280273e-01
8.43118429e-01 -3.18994932e-02 -9.71459821e-02 1.21764088e+00
3.83924007e-01 2.66245872e-01 5.41259825e-01 -1.53111410e+00
-1.07383716e+00 6.10232651e-01 -2.23861039e-01 1.94100231e-01
4.20629889e-01 3.65539700e-01 1.23263943e+00 -9.43707705e-01
6.20927930e-01 1.51490486e+00 2.36025415e-02 6.96226835e-01
-1.31564891e+00 -5.58569074e-01 7.28587806e-01 5.27831733e-01
-1.41827476e+00 -3.76138330e-01 4.57733899e-01 -4.77877825e-01
9.24179792e-01 4.06327277e-01 2.63469130e-01 1.02573586e+00
-5.62523067e-01 1.32659626e+00 1.05938447e+00 -1.60967767e-01
-4.87116259e-03 8.10819119e-02 5.21445036e-01 7.69137561e-01
1.39172167e-01 -2.56898552e-01 -3.83211255e-01 2.45427832e-01
6.97522819e-01 1.54492393e-01 -5.17354250e-01 -2.34139144e-01
-1.70954037e+00 5.13586104e-01 7.49924839e-01 3.58426213e-01
-1.37051627e-01 3.84586155e-01 5.61987340e-01 2.84676254e-01
1.58919007e-01 4.95831430e-01 -6.70115232e-01 2.06759408e-01
-3.64865094e-01 2.93165147e-01 6.57446027e-01 1.48819447e+00
9.01054144e-01 -1.46826744e-01 -6.01640582e-01 9.80263233e-01
9.13441852e-02 4.72156495e-01 4.05816674e-01 -1.14338887e+00
4.61092442e-01 6.97862208e-01 2.10336745e-01 -6.04585052e-01
-3.70072186e-01 -3.50608557e-01 -6.34911120e-01 -1.65438831e-01
3.96650136e-01 3.46102238e-01 -1.20943558e+00 1.76657045e+00
3.69995981e-01 1.78870991e-01 7.96554238e-02 1.09239995e+00
1.68183041e+00 5.85338414e-01 4.94894385e-01 1.59569565e-04
1.87501204e+00 -1.31714606e+00 -7.29447365e-01 -3.93994480e-01
3.15098763e-01 -5.10910094e-01 1.23150158e+00 3.04757625e-01
-1.09465241e+00 -1.04756796e+00 -7.91007638e-01 -6.37429416e-01
-8.70999455e-01 1.37138218e-01 7.64708817e-01 5.03769778e-02
-1.09493589e+00 -1.72580266e-03 -5.01611114e-01 -2.68046111e-01
6.61821008e-01 4.09157813e-01 -2.66998351e-01 -4.09336597e-01
-1.09026122e+00 5.61458647e-01 7.39328265e-01 -3.32934111e-02
-9.98665988e-01 -8.17656040e-01 -9.57842112e-01 6.09229803e-02
9.20046747e-01 -9.80487585e-01 1.45720160e+00 -1.06257689e+00
-1.14606059e+00 1.43185711e+00 -3.43570441e-01 -2.24710613e-01
3.02931309e-01 -2.31343433e-01 -1.90296009e-01 7.00204149e-02
3.74319077e-01 1.18383145e+00 9.05590653e-01 -1.46458316e+00
-8.19651306e-01 -4.09725368e-01 6.00385010e-01 2.86274523e-01
2.62695193e-01 -4.92504686e-02 -1.13509047e+00 -3.93310070e-01
2.44174317e-01 -6.75317645e-01 -2.73635853e-02 2.04664841e-01
-4.88104045e-01 -9.07612145e-01 6.71035349e-01 -1.70881525e-01
6.95457995e-01 -2.24874258e+00 3.52644354e-01 -4.93408978e-01
3.80461842e-01 2.32027248e-01 -2.97024220e-01 1.12925164e-01
5.28144240e-02 7.04494342e-02 -9.36320052e-02 -3.50782543e-01
2.43489593e-01 5.65144181e-01 -4.03936625e-01 4.16454934e-02
2.43845895e-01 1.32556939e+00 -1.14374101e+00 -5.52625835e-01
2.30042025e-01 2.06991374e-01 -1.94085449e-01 4.00352269e-01
-9.38662410e-01 5.50931633e-01 -7.04350054e-01 1.13135731e+00
5.67816257e-01 -6.45945668e-01 -1.25106037e-01 -5.18869221e-01
1.37322575e-01 1.70709804e-01 -1.00382149e+00 1.83526170e+00
-3.94466788e-01 5.35012782e-01 6.30112812e-02 -1.18472731e+00
7.04876661e-01 2.89054096e-01 2.10253134e-01 -9.21261072e-01
-5.00549898e-02 1.57392576e-01 -3.01960111e-01 -8.68615687e-01
2.52751619e-01 2.36341879e-01 -3.15612376e-01 3.47142518e-02
3.03015918e-01 -2.87675917e-01 4.60813791e-01 2.12522194e-01
6.29388928e-01 1.82457119e-01 4.45775628e-01 -1.03315324e-01
5.48569322e-01 1.55684829e-01 5.12242913e-01 1.03781402e+00
-3.24355423e-01 4.06209916e-01 3.92904192e-01 -4.58583117e-01
-3.72572154e-01 -1.23377335e+00 -2.64878631e-01 1.78912747e+00
6.39260590e-01 -3.01491201e-01 -2.74001688e-01 -5.85393608e-01
1.95340335e-01 6.47862673e-01 -6.57889724e-01 2.36592352e-01
-2.80502111e-01 -3.89650136e-01 3.20655584e-01 7.01513588e-01
7.10053265e-01 -1.40318620e+00 -5.65770805e-01 -6.35127872e-02
-4.53944862e-01 -1.58362031e+00 -3.04164916e-01 1.19075887e-01
-7.23962665e-01 -1.26819611e+00 -4.87635642e-01 -9.27315295e-01
6.32766545e-01 3.89099479e-01 1.58698976e+00 2.32284650e-01
-4.54398990e-01 8.15249383e-01 -4.10289705e-01 -6.99815691e-01
4.01900597e-02 -1.40374243e-01 -3.04534584e-01 4.42758203e-02
4.27967310e-01 -4.92995679e-02 -5.23817778e-01 4.58338112e-01
-9.87306118e-01 1.71333015e-01 5.30710280e-01 6.81448460e-01
1.11119318e+00 -3.93461555e-01 5.78132272e-01 -9.74070847e-01
3.53043586e-01 -4.22530442e-01 -6.55437231e-01 5.49927950e-01
-1.22295022e-01 -1.98239908e-01 3.32376957e-01 -5.59632003e-01
-9.44207072e-01 9.19812918e-03 1.60396904e-01 -5.21181285e-01
-5.44572353e-01 2.02912852e-01 -4.30544436e-01 1.36663422e-01
5.23369014e-01 4.76082534e-01 -2.28068143e-01 -4.20306236e-01
7.36381650e-01 4.49589491e-01 7.56383002e-01 -8.20484996e-01
4.17485505e-01 4.63396460e-01 -2.65272230e-01 -7.01584816e-01
-1.51325977e+00 -8.09831023e-01 -7.71404147e-01 -2.03965023e-01
1.13233042e+00 -1.09286284e+00 -9.73588109e-01 3.33115220e-01
-1.29489362e+00 -4.91268307e-01 -1.86731324e-01 3.12303584e-02
-7.09420919e-01 2.83461995e-02 -3.53272349e-01 -6.01039588e-01
-1.39785692e-01 -1.32918513e+00 1.40918028e+00 4.25765425e-01
2.40925238e-01 -8.14717114e-01 -5.35040259e-01 6.75040781e-01
3.75370204e-01 2.85755068e-01 9.36634123e-01 -6.85955405e-01
-1.16457152e+00 3.43373194e-02 -7.15517342e-01 2.59001911e-01
1.00065313e-01 -2.60077804e-01 -1.04428160e+00 -1.82424128e-01
-3.41314226e-01 -8.04767013e-01 1.19866955e+00 3.20528507e-01
1.77734244e+00 -2.22991139e-01 -5.58602214e-01 6.43501639e-01
1.41777205e+00 6.26811311e-02 2.84791142e-01 1.04512744e-01
8.09266627e-01 5.56795895e-01 8.28220069e-01 2.19939519e-02
6.22069657e-01 8.24216127e-01 6.94703579e-01 -1.97219644e-02
-3.24523896e-01 7.58481920e-02 8.75188559e-02 3.93298298e-01
-2.39162117e-01 -3.12289447e-01 -7.34280050e-01 7.79615939e-01
-2.17168427e+00 -9.35309649e-01 -1.36283383e-01 1.89136541e+00
7.54508674e-01 -9.91552919e-02 8.44708607e-02 -3.23819727e-01
6.14501059e-01 2.67601967e-01 -9.28601444e-01 -3.28545906e-02
-2.33181249e-02 -3.58961299e-02 2.32222706e-01 2.52171755e-01
-1.29577601e+00 1.20145929e+00 5.40465927e+00 7.75894761e-01
-9.14855838e-01 3.98243725e-01 5.24970531e-01 2.57788114e-02
-4.09356467e-02 -2.11200237e-01 -8.50643575e-01 2.31058747e-01
1.72245562e-01 -1.02503173e-01 1.70304134e-01 8.78511488e-01
-2.46505603e-01 -6.53278753e-02 -1.51475036e+00 1.20699298e+00
1.70802921e-01 -1.26638246e+00 2.79554069e-01 -4.85278845e-01
4.83873844e-01 2.49717087e-01 1.70437053e-01 7.19851911e-01
1.52113006e-01 -1.05226743e+00 6.94213867e-01 6.50300503e-01
8.64546180e-01 -1.53630659e-01 4.94468927e-01 2.39927933e-01
-1.28001285e+00 -1.27973124e-01 -5.61393142e-01 9.47388187e-02
8.36048201e-02 2.11501613e-01 -7.43713677e-01 6.72014713e-01
8.27349305e-01 7.86495626e-01 -5.50240278e-01 1.06411099e+00
-3.34521890e-01 3.90065879e-01 -2.33831555e-01 -3.60928080e-03
3.26906592e-01 9.20933392e-03 4.57870036e-01 1.41348958e+00
-2.41158485e-01 3.13608944e-01 6.36326075e-01 1.06754851e+00
-3.15856934e-01 -1.45290509e-01 -3.53253603e-01 1.58763066e-01
5.03045678e-01 1.23637772e+00 -5.51032662e-01 -6.34553432e-01
-5.62696397e-01 8.51398766e-01 4.54233348e-01 7.20179856e-01
-8.05237114e-01 -1.45553932e-01 8.62365901e-01 -1.65703014e-01
4.69202012e-01 -1.26898870e-01 -1.93685200e-02 -1.19779193e+00
-1.38380099e-02 -7.66922295e-01 7.56381691e-01 -1.02888572e+00
-1.40999627e+00 5.39305866e-01 2.28534907e-01 -9.98024106e-01
1.00267567e-01 -1.04512346e+00 -3.15558821e-01 5.20793796e-01
-1.81148803e+00 -1.32650590e+00 -8.05536270e-01 8.43089521e-01
1.11963391e+00 1.10227421e-01 7.58138299e-01 2.30772421e-01
-5.83435476e-01 3.44895959e-01 -5.82195759e-01 2.45241135e-01
5.80698431e-01 -1.41127789e+00 -9.14381146e-02 4.15799260e-01
3.79979610e-01 5.92198431e-01 4.64635491e-01 -2.18364760e-01
-1.41415131e+00 -1.21420717e+00 6.57658398e-01 -7.48527110e-01
5.64445019e-01 -3.91121715e-01 -1.11757398e+00 9.21837687e-01
1.61471128e-01 3.73706788e-01 5.33094287e-01 1.24945208e-01
-7.11695969e-01 -6.90866262e-02 -8.94867599e-01 4.64394629e-01
1.38785052e+00 -5.99714398e-01 -7.27918804e-01 9.08361197e-01
1.03498566e+00 -6.02960169e-01 -8.70662510e-01 5.60657799e-01
2.61664659e-01 -7.26077735e-01 1.29551268e+00 -9.80633199e-01
2.12189585e-01 -5.13884842e-01 -5.55441618e-01 -6.96460247e-01
-2.97854483e-01 -2.47575566e-01 -2.61245310e-01 1.14662158e+00
3.87328625e-01 -5.77911258e-01 2.13715658e-01 6.19298041e-01
-2.04420164e-01 -9.52379048e-01 -7.41891623e-01 -6.92673862e-01
-2.94849813e-01 -5.81399798e-01 5.49135566e-01 8.54972243e-01
-3.66964370e-01 5.28843760e-01 2.50970311e-02 3.94602358e-01
6.80508077e-01 6.87058747e-01 7.60343373e-01 -1.08025730e+00
-2.57554233e-01 -6.30318463e-01 -5.52412152e-01 -1.56855929e+00
2.93825865e-01 -1.18059516e+00 3.87800410e-02 -1.81532121e+00
3.58628958e-01 -4.30225879e-01 -5.71225584e-01 9.59873259e-01
-4.15834963e-01 1.35487974e-01 4.96195018e-01 3.87969643e-01
-1.26598096e+00 4.10293102e-01 1.52909100e+00 -4.95509923e-01
9.81835127e-02 1.30667612e-01 -7.52634585e-01 6.52901471e-01
6.04013741e-01 -1.85550481e-01 -5.52573264e-01 -8.44289362e-01
1.70829520e-02 -2.63198391e-02 8.11326444e-01 -7.69319594e-01
2.17427641e-01 -2.77170002e-01 1.02724425e-01 -5.96224010e-01
5.45234561e-01 -6.47062659e-01 -3.28606457e-01 9.48765650e-02
-5.03506184e-01 -1.86705023e-01 2.74047554e-01 7.58842409e-01
-5.07176757e-01 -9.52648968e-02 5.61582267e-01 -4.59981233e-01
-1.51865256e+00 5.15228093e-01 1.90843448e-01 2.12245375e-01
1.05269730e+00 -9.17816311e-02 -6.81833684e-01 -1.63478836e-01
-1.02784956e+00 6.67950690e-01 2.13888392e-01 8.01881313e-01
6.32624090e-01 -1.04541636e+00 -7.02965498e-01 -1.97429046e-01
6.87981844e-01 4.67701465e-01 3.23519975e-01 8.17749202e-01
-2.03692112e-02 6.25860572e-01 -3.10894717e-02 -1.16327477e+00
-1.16748786e+00 8.94117296e-01 3.66293818e-01 1.01232350e-01
-4.15083498e-01 1.30176175e+00 7.68963873e-01 -5.83709002e-01
4.69510913e-01 -4.62896824e-01 -2.19610751e-01 1.15303118e-02
5.76919615e-01 -3.33408825e-02 -1.90111220e-01 -5.50501943e-01
-4.69650477e-01 5.86243927e-01 -1.04162633e-01 4.70490128e-01
1.05345607e+00 -1.45153090e-01 -2.02769503e-01 6.82628870e-01
1.06068337e+00 -6.02686167e-01 -1.21986079e+00 -5.34877479e-01
9.60099623e-02 -4.08498138e-01 -2.30974838e-01 -9.69918132e-01
-1.01371551e+00 8.62157524e-01 4.60406452e-01 3.44614059e-01
1.22852683e+00 7.72484183e-01 4.28650200e-01 6.95793927e-01
5.62453568e-01 -7.61018395e-01 2.49724746e-01 3.79056424e-01
1.16515720e+00 -1.63719475e+00 -1.42360181e-01 -6.77355826e-01
-7.08150327e-01 7.63093352e-01 1.06194711e+00 1.21230572e-01
4.34032679e-01 -2.48289883e-01 8.40576887e-02 -4.90828902e-01
-9.74430740e-01 -8.30435038e-01 5.61873674e-01 6.58636987e-01
4.53331709e-01 2.66877145e-01 -5.70596978e-02 6.27168596e-01
3.26130360e-01 -1.65449083e-01 9.36154649e-02 6.64028823e-01
-5.21676540e-01 -9.08582330e-01 -2.01167524e-01 4.73814815e-01
-2.12362885e-01 -1.80257000e-02 -2.99550980e-01 8.85499477e-01
3.79983097e-01 8.93514097e-01 1.44100562e-01 1.64236501e-01
4.33432460e-01 5.99785224e-02 5.24187446e-01 -9.40025628e-01
-3.58122081e-01 -2.84526914e-01 8.71253908e-02 -9.33444619e-01
-8.63317788e-01 -4.61080849e-01 -1.31696129e+00 4.44103152e-01
-1.29928157e-01 -2.15139002e-01 4.83880699e-01 9.81834352e-01
4.95877922e-01 6.83331668e-01 2.97305882e-02 -8.64954114e-01
-3.22685778e-01 -8.38848054e-01 -1.78521514e-01 7.05511153e-01
4.33867157e-01 -9.55365837e-01 -1.25720324e-02 3.65994245e-01] | [10.166075706481934, 1.5883451700210571] |
54d7e811-c428-446b-9c34-bcce6c60e158 | remote-sensing-image-change-detection-towards | 2305.14722 | null | https://arxiv.org/abs/2305.14722v1 | https://arxiv.org/pdf/2305.14722v1.pdf | Remote Sensing Image Change Detection Towards Continuous Bitemporal Resolution Differences | Most contemporary supervised Remote Sensing (RS) image Change Detection (CD) approaches are customized for equal-resolution bitemporal images. Real-world applications raise the need for cross-resolution change detection, aka, CD based on bitemporal images with different spatial resolutions. Current cross-resolution methods that are trained with samples of a fixed resolution difference (resolution ratio between the high-resolution (HR) image and the low-resolution (LR) one) may fit a certain ratio but lack adaptation to other resolution differences. Toward continuous cross-resolution CD, we propose scale-invariant learning to enforce the model consistently predicting HR results given synthesized samples of varying bitemporal resolution differences. Concretely, we synthesize blurred versions of the HR image by random downsampled reconstructions to reduce the gap between HR and LR images. We introduce coordinate-based representations to decode per-pixel predictions by feeding the coordinate query and corresponding multi-level embedding features into an MLP that implicitly learns the shape of land cover changes, therefore benefiting recognizing blurred objects in the LR image. Moreover, considering that spatial resolution mainly affects the local textures, we apply local-window self-attention to align bitemporal features during the early stages of the encoder. Extensive experiments on two synthesized and one real-world different-resolution CD datasets verify the effectiveness of the proposed method. Our method significantly outperforms several vanilla CD methods and two cross-resolution CD methods on the three datasets both in in-distribution and out-of-distribution settings. The empirical results suggest that our method could yield relatively consistent HR change predictions regardless of varying resolution difference ratios. Our code will be public. | ['Zhenwei Shi', 'Zhengxia Zhou', 'Song Chen', 'Chenyao Zhou', 'Keyan Chen', 'Haotian Zhang', 'Hao Chen'] | 2023-05-24 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 6.98556244e-01 -3.28058839e-01 -2.16777310e-01 -4.96398062e-01
-1.10599029e+00 -4.19760227e-01 7.11143672e-01 -6.43636048e-01
-1.68808997e-01 8.08095932e-01 3.79640192e-01 -2.62093879e-02
-3.24966937e-01 -1.25346327e+00 -9.34295177e-01 -9.29085553e-01
-2.45936766e-01 -2.95389265e-01 1.47469983e-01 -4.23290163e-01
4.64107059e-02 7.26695359e-01 -1.87086666e+00 4.24251050e-01
9.64330554e-01 9.25895274e-01 6.57852352e-01 7.36865640e-01
3.19844663e-01 5.67080438e-01 -2.44251832e-01 9.63581502e-02
8.10565591e-01 -3.25018287e-01 -5.19917786e-01 1.78962901e-01
9.87014472e-01 -6.62284374e-01 -3.63167346e-01 1.14369249e+00
4.85093504e-01 6.03796262e-03 5.84017992e-01 -2.97023088e-01
-1.31560564e+00 2.78053969e-01 -1.11413515e+00 5.14678121e-01
2.44079366e-01 1.53194845e-01 8.21806490e-01 -9.52780306e-01
7.08017647e-01 1.44040966e+00 7.98163950e-01 2.55484372e-01
-1.60540938e+00 -5.48177421e-01 1.35421008e-01 2.65617222e-01
-1.53350627e+00 -5.89438200e-01 7.43310988e-01 -3.87203038e-01
7.63077915e-01 5.53603411e-01 2.68153161e-01 9.83559191e-01
2.91830570e-01 1.42149866e-01 1.69116008e+00 -4.19331700e-01
1.86393205e-02 -1.93749234e-01 -4.35658216e-01 4.16310549e-01
1.78172737e-01 5.96868157e-01 -2.81095684e-01 8.66070092e-02
1.23016298e+00 7.47397840e-02 -8.32244694e-01 -5.33689447e-02
-1.02401280e+00 8.27974796e-01 7.06682205e-01 3.50358993e-01
-5.09912193e-01 -4.19912115e-02 -1.53237075e-01 4.48525637e-01
7.95499325e-01 2.95848817e-01 -5.40689826e-01 5.17816007e-01
-9.85667944e-01 5.90903237e-02 6.20438680e-02 4.94933754e-01
1.13082540e+00 -7.97056630e-02 -3.13798547e-01 1.22791898e+00
6.51805550e-02 9.18290615e-01 4.37863469e-01 -9.78960454e-01
4.60758328e-01 1.45790130e-01 3.48640561e-01 -1.01784098e+00
-9.55115482e-02 -2.03935295e-01 -1.04624546e+00 4.61511463e-01
4.33497392e-02 2.47069940e-01 -9.72269535e-01 1.65344059e+00
3.28026593e-01 -3.93513255e-02 2.49855295e-01 1.06097281e+00
4.49059278e-01 9.32089150e-01 -9.88553688e-02 -4.46421772e-01
1.23666131e+00 -6.36120200e-01 -7.12540567e-01 -4.16382849e-01
-1.22269362e-01 -6.66759133e-01 1.20569181e+00 -1.04924724e-01
-7.24740565e-01 -9.15037215e-01 -1.03522444e+00 -1.77155942e-01
-3.05677772e-01 4.89839345e-01 2.19694093e-01 3.63140851e-01
-1.02408719e+00 6.45659685e-01 -6.32599354e-01 -3.30192178e-01
3.79249930e-01 -2.64066935e-01 -3.75377595e-01 -3.39399993e-01
-1.36154389e+00 1.08302546e+00 -2.92394031e-02 4.68266934e-01
-7.53991663e-01 -6.51820660e-01 -1.00044358e+00 -1.35216087e-01
-7.01922476e-02 -3.86903197e-01 7.08785772e-01 -1.23799801e+00
-1.45612347e+00 8.95243168e-01 -9.34778079e-02 -2.68511146e-01
6.10962391e-01 -3.37167263e-01 -6.94267988e-01 1.74605578e-01
3.63159716e-01 7.39953518e-01 1.21910083e+00 -1.43041050e+00
-7.28899539e-01 -3.79188925e-01 -1.27450551e-03 4.70934331e-01
3.04451793e-01 -1.31849915e-01 -4.10554633e-02 -7.60057628e-01
3.82680297e-01 -4.66574818e-01 -1.05619639e-01 4.00450826e-01
9.00889561e-02 3.54863405e-01 6.85375273e-01 -9.31144834e-01
1.03850162e+00 -2.24128342e+00 -1.40120462e-01 -2.62160063e-01
-3.41882229e-01 -1.03223715e-02 -3.49873632e-01 8.54396299e-02
-2.37885579e-01 -2.26930771e-02 -6.77009821e-01 2.16693208e-01
-3.29630703e-01 2.61777729e-01 -6.82292700e-01 8.84065092e-01
5.14016390e-01 4.94575709e-01 -7.64851391e-01 -2.93404460e-01
4.81674731e-01 7.91139722e-01 -2.08196416e-01 4.33887661e-01
3.91637161e-02 5.00771165e-01 -1.62944049e-01 4.68214303e-01
1.30962956e+00 9.74148810e-02 1.38832675e-02 -6.41029954e-01
-6.15059495e-01 3.07851564e-02 -9.82864857e-01 1.31865144e+00
-7.75825083e-01 7.64099002e-01 -5.28808497e-02 -5.52326083e-01
1.22236502e+00 -1.44830868e-01 2.02200651e-01 -1.01424360e+00
-5.62171936e-01 9.74638760e-02 -3.91677201e-01 -7.05972910e-01
6.19309545e-01 -2.69707263e-01 2.89422631e-01 -9.53285489e-03
-5.14226198e-01 -1.92203104e-01 -3.61877441e-01 -6.39134705e-01
6.73344374e-01 4.01673019e-01 6.03984773e-01 -9.43930298e-02
5.96210003e-01 -2.12563381e-01 3.77879739e-01 7.80486763e-01
-1.18204914e-01 1.15737653e+00 -1.13724068e-01 -5.67007244e-01
-1.10909188e+00 -1.24898207e+00 -8.18394542e-01 9.59145486e-01
3.50755095e-01 4.75255936e-01 -4.11914021e-01 -2.46241823e-01
-2.03786924e-01 6.10707283e-01 -8.65114331e-01 6.16261587e-02
-6.76042557e-01 -1.12105429e+00 3.23168546e-01 2.19829246e-01
1.13842440e+00 -9.21252191e-01 -9.09339130e-01 2.75000557e-02
-3.44060928e-01 -1.05543399e+00 -3.09031308e-01 3.46966088e-02
-7.26516783e-01 -9.58026588e-01 -7.72518754e-01 -6.22932673e-01
4.60840434e-01 6.42327547e-01 8.36381257e-01 -5.59228599e-01
-3.65033180e-01 6.83408529e-02 -3.36282730e-01 5.00764608e-01
-3.54766071e-01 -3.05153519e-01 -2.37690151e-01 1.91723391e-01
4.23044572e-03 -4.23388839e-01 -7.86956131e-01 3.02749693e-01
-1.05027556e+00 2.23398298e-01 7.34584570e-01 1.02042675e+00
7.55064845e-01 2.23975509e-01 3.45817626e-01 -4.79410499e-01
2.61937171e-01 -4.37645406e-01 -6.09489560e-01 4.58234221e-01
-6.68266833e-01 1.15519151e-01 3.99744451e-01 -3.73831064e-01
-1.52460968e+00 6.16938546e-02 8.33234563e-03 -2.43967250e-01
-3.33226830e-01 1.81239441e-01 -4.32282202e-02 9.61537957e-02
1.11541140e+00 4.62334424e-01 -2.61457413e-01 -5.80053866e-01
5.73093593e-01 8.79265308e-01 8.30342770e-01 -2.16344565e-01
9.34135437e-01 8.47719371e-01 -3.02265853e-01 -8.79230201e-01
-8.58507335e-01 -1.59328744e-01 -5.42932808e-01 -7.99169913e-02
9.16668415e-01 -1.38632107e+00 5.90687729e-02 5.08654416e-01
-9.32567775e-01 -4.33389783e-01 -2.86813766e-01 4.30525124e-01
-6.06583416e-01 3.12292337e-01 -4.77560639e-01 -7.97344029e-01
-3.66019607e-01 -9.08118069e-01 1.41938782e+00 3.17432195e-01
3.67357701e-01 -6.24386191e-01 3.75928968e-01 8.46371129e-02
7.69708872e-01 3.05986047e-01 7.23016441e-01 5.16253710e-01
-6.05552256e-01 3.39623690e-01 -6.66486740e-01 3.74328375e-01
7.57571101e-01 -5.96891418e-02 -1.20357168e+00 -3.22583646e-01
1.12694785e-01 -3.25943418e-02 1.07980239e+00 7.10926950e-01
1.13518476e+00 -5.75183034e-01 -7.76228607e-02 8.93229008e-01
1.91756165e+00 -7.62342587e-02 1.01417899e+00 6.08564019e-01
5.24471402e-01 5.51679850e-01 9.42076147e-01 4.28542048e-01
2.24475980e-01 9.64291751e-01 3.93279433e-01 -4.21872616e-01
-5.34964263e-01 -1.88176483e-01 6.65227771e-01 -3.11357398e-02
-3.02170247e-01 1.88859746e-01 -3.33209187e-01 5.84884405e-01
-1.46660292e+00 -1.47557700e+00 1.77214980e-01 2.31342673e+00
1.33614671e+00 -5.19797683e-01 -4.95344192e-01 -1.97617814e-01
9.96200800e-01 7.38255799e-01 -7.31248856e-01 2.02911105e-02
-6.47066236e-01 3.49506080e-01 9.43649888e-01 8.70577812e-01
-1.31549847e+00 9.56967831e-01 6.13190508e+00 7.86462545e-01
-1.54262924e+00 1.37689143e-01 7.53883421e-01 1.49378017e-01
-3.86494130e-01 -3.24653119e-01 -6.94474399e-01 3.36744606e-01
5.85573018e-01 3.29781234e-01 7.80602992e-01 6.52678967e-01
5.80345154e-01 -3.25425476e-01 -7.15609729e-01 1.05441022e+00
-2.94310991e-02 -1.20243549e+00 -3.02278046e-02 -2.55104750e-01
9.95332897e-01 2.81736493e-01 2.49064490e-01 3.52595001e-02
2.03595430e-01 -1.04002643e+00 7.17951357e-01 8.09511542e-01
1.52009213e+00 -3.32854301e-01 4.47614044e-01 -6.55028522e-02
-1.34513462e+00 -1.41885132e-01 -8.21627796e-01 1.31279975e-01
-5.55912331e-02 5.49474418e-01 -4.38207716e-01 6.10382855e-01
1.17570555e+00 8.69033217e-01 -6.77398384e-01 4.39080864e-01
-1.51041031e-01 2.30542392e-01 -1.47694007e-01 7.97757924e-01
-1.72126859e-01 -3.94490749e-01 3.49723548e-01 1.18811274e+00
6.10695720e-01 2.24970266e-01 -3.56280893e-01 1.04291034e+00
3.27014148e-01 -3.35771173e-01 -6.60677731e-01 6.23200774e-01
6.66436851e-01 9.25102711e-01 -7.44994432e-02 -1.27409294e-01
-4.26529258e-01 1.19153428e+00 1.33520335e-01 6.15350008e-01
-9.62195396e-01 -7.60189444e-02 9.63051379e-01 1.34212568e-01
7.65396237e-01 1.77606121e-02 -2.60973454e-01 -1.14488590e+00
1.14518322e-01 -8.13706040e-01 2.40723938e-01 -1.12532604e+00
-1.35083008e+00 7.24456310e-01 2.07729667e-01 -1.55336714e+00
-9.99634564e-02 -2.94162601e-01 -1.39223486e-01 1.09254014e+00
-2.15709758e+00 -1.19467902e+00 -6.37668312e-01 4.77393806e-01
7.98821807e-01 2.83875257e-01 5.83846450e-01 1.93387717e-01
-3.15382957e-01 1.46697879e-01 4.65095490e-01 -1.87383965e-01
8.54797721e-01 -8.85883868e-01 2.49990016e-01 1.03968418e+00
-1.94812030e-01 2.59871870e-01 7.64873683e-01 -5.00364006e-01
-9.96113837e-01 -1.57279015e+00 6.42281651e-01 -9.87563506e-02
3.22005630e-01 1.47212178e-01 -1.10984719e+00 5.33405960e-01
-7.55639151e-02 3.07059318e-01 2.07950212e-02 -4.54965323e-01
-5.23812056e-01 -6.00608110e-01 -1.42505574e+00 5.01907110e-01
1.08859766e+00 -8.95746469e-01 -6.79972589e-01 4.93959747e-02
6.10641539e-01 -4.38659400e-01 -8.68946731e-01 7.02235222e-01
6.52521193e-01 -1.24728537e+00 1.28275907e+00 6.81571513e-02
8.38170707e-01 -7.15940118e-01 -7.87409067e-01 -1.23633218e+00
-6.86121047e-01 3.03716026e-02 1.74884722e-01 1.01485276e+00
3.38758342e-02 -6.84754491e-01 -3.87168117e-02 6.21999390e-02
6.84893578e-02 -4.24879014e-01 -9.99552608e-01 -6.36066854e-01
6.83776215e-02 -3.98554653e-02 6.77703857e-01 1.06709683e+00
-6.46264315e-01 1.40234604e-02 -5.98716021e-01 9.77659285e-01
5.81089079e-01 6.22571170e-01 4.59381014e-01 -8.38367879e-01
-2.62511939e-01 -2.60105669e-01 -8.20948854e-02 -1.14340425e+00
-1.56985730e-01 -3.91967684e-01 2.35348374e-01 -1.42049432e+00
1.80957913e-01 -4.73652869e-01 -6.85303137e-02 4.26347107e-01
-3.85935545e-01 4.96328712e-01 -2.67139137e-01 5.66076815e-01
1.70210183e-01 6.88992620e-01 1.23688757e+00 -2.82588273e-01
-2.95550913e-01 -2.82592803e-01 -4.91130233e-01 3.45228314e-01
6.92475677e-01 -2.74967790e-01 -2.07184345e-01 -5.78154862e-01
-3.03008370e-02 6.91880435e-02 5.39967418e-01 -9.57997561e-01
-3.16165537e-01 -6.17127776e-01 5.94325542e-01 -5.10878921e-01
2.77399153e-01 -7.83945441e-01 6.34999752e-01 3.12652677e-01
-4.87918556e-01 -2.78444260e-01 -7.52942190e-02 5.46452105e-01
-1.15021706e-01 6.66600689e-02 1.42352867e+00 -5.93877584e-02
-7.88386285e-01 2.19456881e-01 -1.43666476e-01 -3.61543775e-01
6.49633110e-01 -3.08532089e-01 -7.87609935e-01 -1.47863865e-01
-3.08907449e-01 -3.98049384e-01 7.32511759e-01 5.38060844e-01
5.79149246e-01 -1.27408826e+00 -9.52180266e-01 4.16630745e-01
2.53307253e-01 3.05951089e-02 5.02748668e-01 5.65709114e-01
-6.44664705e-01 1.03783086e-01 -4.72536892e-01 -6.67649925e-01
-1.00998521e+00 4.12881851e-01 7.62836874e-01 2.31024716e-02
-7.77178168e-01 4.79482025e-01 5.36081791e-01 -3.23915541e-01
-4.20543134e-01 -4.75361675e-01 -2.50441372e-01 -1.37932971e-01
7.44172156e-01 2.53220826e-01 -7.31047336e-03 -8.26121867e-01
-1.35572508e-01 1.07967138e+00 1.96443826e-01 -9.27408934e-02
1.49878430e+00 -7.09971905e-01 -1.07434243e-02 4.13631141e-01
1.20630980e+00 6.82211742e-02 -1.90451205e+00 -4.74896133e-01
-4.81791854e-01 -1.07063925e+00 2.76945800e-01 -8.06459546e-01
-1.15822983e+00 4.84088212e-01 1.44187653e+00 -4.32176180e-02
1.54171467e+00 7.99970105e-02 1.61477000e-01 1.29289478e-01
3.12688619e-01 -9.02298868e-01 -2.34595135e-01 2.05543175e-01
1.38148272e+00 -1.46695089e+00 2.81715244e-01 -2.64919102e-01
-4.57867235e-01 1.09904277e+00 4.59888250e-01 -4.74899784e-02
3.58168215e-01 -9.83680487e-02 2.77800769e-01 -5.89324236e-02
-5.90731621e-01 -3.28190029e-01 1.04602069e-01 8.44522357e-01
3.31596673e-01 1.64219305e-01 -3.07636887e-01 -8.98731351e-02
-1.47592083e-01 -2.14664176e-01 4.87532318e-01 4.16979641e-01
-6.56562984e-01 -4.18573976e-01 -8.32845032e-01 1.51630893e-01
-6.96340352e-02 -3.26341659e-01 1.55090332e-01 7.09805965e-01
2.97631294e-01 9.01989341e-01 3.91919672e-01 -1.85702026e-01
2.83685803e-01 -4.02498454e-01 6.07420862e-01 -2.03925371e-01
8.78110901e-02 3.31627727e-02 -1.58848584e-01 -6.40391469e-01
-7.50154257e-01 -8.07985961e-01 -7.81540036e-01 -1.24455482e-01
-1.47797257e-01 -4.99592006e-01 4.14829850e-01 6.40352309e-01
2.28926376e-01 1.85854092e-01 1.20176136e+00 -1.02975798e+00
-6.77565455e-01 -1.26773500e+00 -6.80205584e-01 2.18482643e-01
9.24810708e-01 -6.61497951e-01 -6.69254661e-01 3.02883357e-01] | [10.738204956054688, -2.0406689643859863] |
21b56e9f-685a-4cb5-8f47-15238647ff4a | uncertainty-quantification-for-atlas-level | 2211.03793 | null | https://arxiv.org/abs/2211.03793v1 | https://arxiv.org/pdf/2211.03793v1.pdf | Uncertainty Quantification for Atlas-Level Cell Type Transfer | Single-cell reference atlases are large-scale, cell-level maps that capture cellular heterogeneity within an organ using single cell genomics. Given their size and cellular diversity, these atlases serve as high-quality training data for the transfer of cell type labels to new datasets. Such label transfer, however, must be robust to domain shifts in gene expression due to measurement technique, lab specifics and more general batch effects. This requires methods that provide uncertainty estimates on the cell type predictions to ensure correct interpretation. Here, for the first time, we introduce uncertainty quantification methods for cell type classification on single-cell reference atlases. We benchmark four model classes and show that currently used models lack calibration, robustness, and actionable uncertainty scores. Furthermore, we demonstrate how models that quantify uncertainty are better suited to detect unseen cell types in the setting of atlas-level cell type transfer. | ['Fabian Theis', 'Malte Luecken', 'Lisa Sikkema', 'Giovanni Palla', 'Leon Hetzel', 'Jan Engelmann'] | 2022-11-07 | null | null | null | null | ['type'] | ['speech'] | [ 1.72301427e-01 -3.17877382e-02 -1.67267933e-01 -3.76944035e-01
-1.36373520e+00 -1.11993098e+00 7.53221512e-01 7.33406007e-01
-2.77841240e-01 1.38482249e+00 -1.26848677e-02 1.06583335e-01
2.73515675e-02 -7.23524034e-01 -9.02485371e-01 -9.94036019e-01
1.92135751e-01 9.00455236e-01 2.08617762e-01 2.41691977e-01
1.26036704e-01 6.18870318e-01 -1.14353526e+00 1.71737581e-01
5.46363771e-01 8.26860249e-01 -2.90208280e-01 7.84840345e-01
-7.68281296e-02 1.64954975e-01 -4.83801901e-01 -5.06582744e-02
4.53571156e-02 -3.74366969e-01 -5.69817781e-01 -4.96018648e-01
5.53757191e-01 9.16579589e-02 2.17718959e-01 9.49124634e-01
5.48188210e-01 -3.40762585e-01 1.24931228e+00 -1.41086662e+00
-2.36500815e-01 2.81842142e-01 -4.42886651e-02 8.52518976e-02
9.81328934e-02 4.27931756e-01 7.29942560e-01 -6.05713427e-01
1.25940323e+00 9.50530231e-01 8.69662523e-01 7.31198430e-01
-1.90724957e+00 -7.39401817e-01 -1.53766498e-01 -6.73271835e-01
-1.59863865e+00 -5.40859699e-01 2.95395777e-02 -1.13298786e+00
5.41317046e-01 3.46462637e-01 5.50563812e-01 1.04463935e+00
6.93310559e-01 1.80086717e-01 1.49270797e+00 -3.23178321e-02
6.82134449e-01 2.30253279e-01 -1.42944343e-02 4.17087436e-01
3.53945494e-01 4.02650498e-02 -3.33342403e-01 -3.33890200e-01
8.01003098e-01 -1.89651057e-01 -2.55839735e-01 -4.91850942e-01
-1.58552015e+00 4.47925836e-01 1.32198483e-01 3.45485955e-01
8.83877501e-02 2.84465134e-01 5.48862219e-01 -3.55944969e-02
5.43872476e-01 8.18698525e-01 -9.34145868e-01 -1.68037433e-02
-9.79622841e-01 1.47480011e-01 5.61262310e-01 1.07269382e+00
6.72516584e-01 -2.95770079e-01 -5.95980465e-01 6.47700787e-01
7.56656528e-02 4.97723997e-01 3.99575144e-01 -1.10835528e+00
-3.77598882e-01 5.80378175e-01 1.33771047e-01 -4.66066688e-01
-7.30426073e-01 -5.23560584e-01 -8.06747794e-01 1.22894317e-01
8.71518195e-01 -1.01077557e-02 -1.03224945e+00 1.95702338e+00
3.76914859e-01 2.15731293e-01 -9.68692005e-02 4.55908954e-01
4.42641526e-01 1.67403013e-01 3.72939587e-01 -2.23568022e-01
1.40250254e+00 7.07223639e-02 -4.14117187e-01 2.05961078e-01
1.08345306e+00 -3.39335889e-01 8.32377732e-01 2.65293452e-03
-6.07210636e-01 8.13490376e-02 -7.21693575e-01 2.93512475e-02
-8.33855629e-01 -1.81837067e-01 4.33820158e-01 7.49172926e-01
-7.94521987e-01 5.00605106e-01 -7.64480233e-01 -8.18290770e-01
7.30024874e-01 3.89338940e-01 -4.53999966e-01 2.10982069e-01
-8.83969605e-01 7.81862259e-01 2.50258267e-01 -1.05351329e-01
-1.07200980e+00 -1.40144110e+00 -4.86149192e-01 6.14696136e-03
-1.91584617e-01 -7.89936721e-01 1.02097380e+00 -2.85592407e-01
-1.45819092e+00 1.27218699e+00 4.73609455e-02 -1.53216392e-01
3.62122625e-01 6.53380513e-01 -1.29697725e-01 -1.33597165e-01
1.77220374e-01 9.57832098e-01 3.24240737e-02 -1.44881606e+00
-6.22419536e-01 -5.81307352e-01 -4.41378146e-01 -3.91645402e-01
2.40197465e-01 -1.85190946e-01 1.88775007e-02 -2.79998422e-01
8.42307732e-02 -9.63615656e-01 -1.22296557e-01 1.50705546e-01
-1.63646728e-01 4.20255154e-01 2.44019195e-01 -4.68122125e-01
7.49527156e-01 -1.89785659e+00 4.47586402e-02 3.68313342e-01
1.46946818e-01 -3.23868781e-01 7.94816539e-02 5.19787073e-02
2.23611817e-01 5.59666157e-01 -3.82804692e-01 5.05780429e-03
9.95911285e-02 2.11643297e-02 -2.61393562e-03 6.77180409e-01
1.87947869e-01 1.00654066e+00 -8.55235934e-01 -5.68485022e-01
5.22039160e-02 3.22039336e-01 -4.34228361e-01 -5.90272434e-02
-3.54951143e-01 9.49310064e-01 -7.33481199e-02 9.11437750e-01
4.62621868e-01 -1.37271628e-01 7.16794357e-02 -2.17633918e-01
-2.87662055e-02 -1.19774126e-01 -6.09131932e-01 1.52079070e+00
-3.31260026e-01 1.94549158e-01 3.52248490e-01 -4.26575303e-01
7.70377159e-01 1.36629239e-01 6.19391084e-01 -2.26934031e-01
2.48163089e-01 5.20528555e-01 4.03335271e-03 2.62673318e-01
3.66964824e-02 -6.73337281e-01 -3.15412402e-01 -2.82493811e-02
4.24705505e-01 -5.97989440e-01 -3.49162556e-02 6.97775558e-02
1.08707690e+00 -8.84668250e-03 1.43336087e-01 -9.61858690e-01
2.82897234e-01 7.65915215e-02 9.67813909e-01 5.84902704e-01
-5.30509233e-01 9.57599103e-01 7.93116212e-01 -1.43940017e-01
-1.29122639e+00 -1.11426103e+00 -8.19223106e-01 6.50791466e-01
-1.66988760e-01 -6.80728927e-02 -6.49029315e-01 -4.69515651e-01
3.90844375e-01 4.84547824e-01 -8.49816144e-01 -1.32397637e-01
6.45935386e-02 -1.22799623e+00 1.07243156e+00 3.67124498e-01
-1.99057207e-01 -3.15770537e-01 -4.96826917e-02 2.92430341e-01
7.74541870e-02 -9.93911505e-01 -3.65471542e-01 4.84168649e-01
-7.11996913e-01 -1.24600339e+00 -8.35691631e-01 -4.77360427e-01
8.34977627e-01 -5.67701042e-01 9.46118116e-01 -1.70846120e-01
-4.42059547e-01 5.39345741e-01 1.54313385e-01 -4.63303328e-01
-6.40889049e-01 2.07829401e-01 2.04002514e-01 -2.37124905e-01
4.90455851e-02 -4.35731024e-01 -4.85918969e-01 4.60517585e-01
-8.14098418e-01 -2.79069215e-01 3.18196207e-01 1.00681686e+00
1.21743190e+00 -3.37503493e-01 8.17187369e-01 -1.10999370e+00
2.11466774e-01 -4.91892189e-01 -9.73529756e-01 4.94019240e-01
-7.23994553e-01 9.46108699e-02 5.05745351e-01 7.46164704e-03
-1.02252233e+00 2.41720647e-01 -6.35419553e-03 -3.57424878e-02
-4.35988516e-01 2.03751728e-01 -1.02278486e-01 -3.32045406e-01
8.67617667e-01 3.07414588e-02 1.83721557e-01 -7.72158951e-02
8.65417123e-02 3.77355605e-01 6.42016292e-01 -9.73767757e-01
3.81230563e-01 5.69039822e-01 6.54057264e-01 -3.42686474e-01
-5.48859835e-01 -3.40117246e-01 -9.90430772e-01 -1.90229356e-01
8.81758273e-01 -1.01896286e+00 -7.45966017e-01 4.55738366e-01
-7.29138136e-01 -7.78516293e-01 -3.23740751e-01 3.35047036e-01
-8.67963374e-01 -1.11077748e-01 -7.61595070e-01 -5.73194683e-01
-1.34749347e-02 -1.06083202e+00 1.54730296e+00 1.20836601e-01
-3.71131033e-01 -1.18919563e+00 3.78708541e-01 -6.25015721e-02
3.71286124e-01 5.64727068e-01 1.03682601e+00 -6.89790905e-01
-7.56131828e-01 -3.12483728e-01 9.79694072e-04 -2.24832937e-01
2.24239990e-01 4.27031904e-01 -1.22287762e+00 -2.50565827e-01
-6.26648486e-01 -1.87014803e-01 7.30554283e-01 6.77151024e-01
1.02446210e+00 -6.45134076e-02 -7.76516020e-01 7.61428297e-01
1.42159951e+00 -2.72591952e-02 4.60046232e-01 8.25043097e-02
3.96080136e-01 6.01732194e-01 3.54089588e-01 1.41313761e-01
9.71561596e-02 6.56117499e-01 -1.10890284e-01 4.21091504e-02
-1.02765962e-01 -1.57532945e-01 -1.95226688e-02 5.17628372e-01
3.31676394e-01 -2.56875962e-01 -1.26311266e+00 5.65830112e-01
-1.54056537e+00 -7.74326801e-01 -1.38545975e-01 2.46259880e+00
9.99227822e-01 7.02742264e-02 7.92807043e-02 -3.70954245e-01
6.46717370e-01 -5.30727327e-01 -7.07748532e-01 -8.77917279e-03
-4.36782569e-01 -2.54667699e-01 9.15516973e-01 6.30087376e-01
-6.88962162e-01 7.35359728e-01 7.60540438e+00 7.82719791e-01
-8.96514893e-01 -1.54876895e-02 1.22382367e+00 -1.17370270e-01
-4.61350590e-01 1.67241186e-01 -1.00837600e+00 6.38532162e-01
9.57857430e-01 -4.42963600e-01 2.51071662e-01 2.99338490e-01
8.98706261e-03 -1.66296631e-01 -1.43575549e+00 7.15186298e-01
-5.15928864e-01 -1.52595019e+00 -3.26493084e-01 3.05099726e-01
9.73184466e-01 1.82417899e-01 -1.64227948e-01 2.68996119e-01
7.24196196e-01 -1.13989663e+00 3.29471499e-01 7.83691823e-01
1.09188831e+00 -4.61407810e-01 1.04243314e+00 1.29476294e-01
-8.21093798e-01 4.32153583e-01 -3.52755666e-01 3.77949148e-01
1.03309333e-01 9.74222660e-01 -1.02496767e+00 2.97275424e-01
3.01284820e-01 2.78262585e-01 -6.70797169e-01 1.05348134e+00
5.03316581e-01 4.99404103e-01 -7.58749306e-01 2.50118434e-01
-4.05971497e-01 -6.42719492e-03 2.70124227e-01 1.26316869e+00
5.97753286e-01 8.28541741e-02 -1.92647446e-02 1.11150467e+00
-2.58581489e-01 -1.22954305e-02 -6.43189073e-01 -3.25801186e-02
6.26582086e-01 1.56340420e+00 -1.15026271e+00 -2.83869624e-01
-4.59235953e-03 5.83856642e-01 3.32448214e-01 2.72604227e-01
-5.00864923e-01 1.18271731e-01 8.22840214e-01 1.81230500e-01
-2.49599412e-01 7.55235031e-02 -6.49293065e-01 -9.94930029e-01
-5.81669986e-01 -4.98870164e-01 5.70759058e-01 -6.37665331e-01
-1.65214193e+00 -2.53867246e-02 -5.24959080e-02 -8.60000432e-01
-6.54317960e-02 -8.95705581e-01 -1.48245022e-01 6.92555428e-01
-1.09219360e+00 -1.17661607e+00 -1.41246781e-01 -1.46848232e-01
-7.99733177e-02 1.00092039e-01 1.08886576e+00 2.89488267e-02
-5.22114277e-01 6.05474651e-01 5.65061510e-01 -4.38583344e-02
9.85672414e-01 -1.42180395e+00 -1.03590280e-01 3.78464162e-01
-2.23692775e-01 6.61971867e-01 8.83763194e-01 -8.60047400e-01
-1.07735372e+00 -1.24525273e+00 5.04837215e-01 -1.12082267e+00
5.59194922e-01 -7.39529729e-01 -8.97180438e-01 8.41395319e-01
-4.50459182e-01 3.56697172e-01 9.49841857e-01 3.73163462e-01
-3.48966032e-01 -2.40982488e-01 -1.54125667e+00 4.79800463e-01
8.45003426e-01 -6.88003600e-01 1.24963999e-01 3.11808586e-01
2.61546820e-01 -6.12839520e-01 -1.76021433e+00 4.31343108e-01
8.86944354e-01 -5.07583261e-01 5.07838905e-01 -5.99251032e-01
5.84914386e-02 -4.95538741e-01 -2.78746843e-01 -1.45863903e+00
-4.45517600e-01 -3.18782985e-01 5.17113805e-01 1.68226469e+00
6.71658456e-01 -6.13752484e-01 6.52076781e-01 1.06175852e+00
-1.52344584e-01 -3.13085645e-01 -1.23570287e+00 -9.67482328e-01
7.69730330e-01 1.57844722e-01 8.70305419e-01 1.04504144e+00
3.78022343e-01 9.28670689e-02 2.18899071e-01 1.47061935e-02
7.00775206e-01 -3.06906611e-01 7.76752174e-01 -1.60289288e+00
-2.19556484e-02 -6.91686451e-01 -6.97113872e-01 -1.89418659e-01
2.86201447e-01 -9.66138184e-01 4.46753085e-01 -1.07343578e+00
5.20444453e-01 -6.17955327e-01 -5.91668546e-01 2.76735455e-01
1.00581618e-02 2.38064453e-01 -1.35102734e-01 8.37038979e-02
-4.64980662e-01 3.38222235e-01 1.00150776e+00 -1.66951492e-01
1.10467717e-01 -4.53283280e-01 -5.69778860e-01 3.84606570e-01
8.52728963e-01 -6.00442111e-01 -2.13017389e-02 1.56598330e-01
2.32532367e-01 -7.61978030e-02 5.62830806e-01 -1.18693364e+00
1.79980844e-01 -3.41804117e-01 6.37940824e-01 -2.98924088e-01
3.04148924e-02 -7.65964031e-01 7.24328458e-01 3.05083930e-01
-6.86905980e-01 -3.25062543e-01 5.04733980e-01 7.55141854e-01
8.96005705e-02 6.89381808e-02 1.12055397e+00 -2.69994617e-01
-1.13326512e-01 4.25022036e-01 -5.77564597e-01 2.14106783e-01
9.63387132e-01 -1.51992768e-01 -6.15957737e-01 -7.07308948e-02
-6.80145323e-01 2.60315120e-01 1.14425778e+00 -7.13361204e-02
-2.18783990e-01 -1.47366440e+00 -6.73240602e-01 -1.10345259e-01
6.44944608e-01 1.47594690e-01 3.47222447e-01 8.65723968e-01
-4.07543689e-01 6.17866755e-01 -3.72497469e-01 -9.26918209e-01
-9.76084352e-01 4.66499418e-01 7.20830142e-01 -2.72138417e-01
1.44424826e-01 7.52418876e-01 5.64990103e-01 -9.60367858e-01
-3.99383187e-01 -4.78765905e-01 2.38875896e-01 -2.41559036e-02
4.65633236e-02 1.21370651e-01 1.83653578e-01 -6.88477695e-01
-5.50059259e-01 5.82567811e-01 1.40661299e-01 -8.51325840e-02
9.25729036e-01 -1.60468891e-01 -2.50829637e-01 8.64321172e-01
9.91410673e-01 -7.78631717e-02 -1.32036662e+00 2.69117624e-01
1.53213292e-01 -1.78248197e-01 -5.28167784e-02 -1.12231708e+00
-7.03699648e-01 5.23372948e-01 6.90448701e-01 -1.81016296e-01
6.50734365e-01 7.35150129e-02 2.48307601e-01 5.09505980e-02
6.22404635e-01 -1.27495503e+00 -7.60931373e-01 2.95351595e-01
5.82234561e-01 -1.08326042e+00 1.86066218e-02 -4.86171424e-01
-8.07499811e-02 7.87502170e-01 7.66109109e-01 3.35028499e-01
6.67958379e-01 6.09027267e-01 1.22825086e-01 -1.59677178e-01
-9.87397075e-01 1.23406745e-01 -1.75369486e-01 9.78959084e-01
5.39892435e-01 1.41853586e-01 -1.45820424e-01 6.39667034e-01
-4.12097154e-03 4.79835868e-01 5.02709329e-01 5.45707405e-01
-4.22428012e-01 -9.87532616e-01 -2.86093622e-01 6.25874579e-01
-4.11454111e-01 1.74755856e-01 -5.33078313e-01 6.68842018e-01
-5.21699004e-02 3.95612687e-01 1.57950655e-01 -4.29430492e-02
9.06101838e-02 5.88126600e-01 6.16458237e-01 -5.54113507e-01
-4.06260461e-01 -4.01443355e-02 -6.59074411e-02 -2.67794639e-01
-1.77908465e-01 -8.39586735e-01 -1.44552469e+00 -1.91123366e-01
-3.16844583e-01 -5.63830882e-02 4.74312097e-01 8.46218646e-01
6.59158409e-01 3.80799621e-01 2.08868921e-01 -5.51294744e-01
-1.50842279e-01 -8.28059137e-01 -9.04778302e-01 5.04635990e-01
9.46724042e-02 -5.93275726e-01 -4.28337604e-01 3.59447271e-01] | [14.64203929901123, -3.0564961433410645] |
e4ea3c7c-eef8-458b-a8df-fbf3fdb59502 | dynamic-linear-transformer-for-3d-biomedical | 2206.00771 | null | https://arxiv.org/abs/2206.00771v2 | https://arxiv.org/pdf/2206.00771v2.pdf | Dynamic Linear Transformer for 3D Biomedical Image Segmentation | Transformer-based neural networks have surpassed promising performance on many biomedical image segmentation tasks due to a better global information modeling from the self-attention mechanism. However, most methods are still designed for 2D medical images while ignoring the essential 3D volume information. The main challenge for 3D transformer-based segmentation methods is the quadratic complexity introduced by the self-attention mechanism \cite{vaswani2017attention}. In this paper, we propose a novel transformer architecture for 3D medical image segmentation using an encoder-decoder style architecture with linear complexity. Furthermore, we newly introduce a dynamic token concept to further reduce the token numbers for self-attention calculation. Taking advantage of the global information modeling, we provide uncertainty maps from different hierarchy stages. We evaluate this method on multiple challenging CT pancreas segmentation datasets. Our promising results show that our novel 3D Transformer-based segmentor could provide promising highly feasible segmentation performance and accurate uncertainty quantification using single annotation. Code is available https://github.com/freshman97/LinTransUNet. | ['Ulas Bagci', 'Zheyuan Zhang'] | 2022-06-01 | null | null | null | null | ['3d-medical-imaging-segmentation', 'pancreas-segmentation'] | ['medical', 'medical'] | [ 3.04778852e-02 6.15759313e-01 -2.11318612e-01 -6.61477745e-01
-1.32804024e+00 -1.12979837e-01 9.32833850e-02 5.10662556e-01
-4.78607029e-01 4.77730423e-01 2.09058687e-01 -3.08568716e-01
4.48089391e-02 -5.89867234e-01 -9.47886527e-01 -6.94261611e-01
3.30150574e-02 7.03260660e-01 3.01141948e-01 2.34166637e-01
-2.00104136e-02 4.65569571e-02 -8.10339868e-01 2.61227876e-01
1.21014881e+00 1.24715137e+00 3.69481653e-01 4.54183847e-01
-2.22072482e-01 8.23790789e-01 -5.17468691e-01 -3.57989907e-01
7.56627694e-02 -4.52452153e-01 -1.00379431e+00 -2.93784291e-01
2.05830216e-01 -6.56320691e-01 -2.51798719e-01 1.24038768e+00
7.93390751e-01 -1.98883131e-01 4.70915943e-01 -7.46684372e-01
-5.05632937e-01 1.41946173e+00 -3.58539730e-01 4.53061938e-01
-9.25085098e-02 5.73682785e-02 7.62178481e-01 -4.51632679e-01
3.69405717e-01 8.42885554e-01 8.52572858e-01 5.32861412e-01
-9.62897480e-01 -5.62318325e-01 6.39099702e-02 2.34392732e-01
-1.30299461e+00 -1.03074662e-01 5.97862542e-01 -2.90963054e-01
1.11002433e+00 9.93117318e-02 8.13166142e-01 8.46485853e-01
7.27238953e-01 1.18477750e+00 9.24117923e-01 -4.10598740e-02
2.03097209e-01 -4.65527549e-03 2.11525783e-01 8.02304983e-01
8.15380961e-02 -2.48528458e-02 -1.52398676e-01 2.85836369e-01
9.33585286e-01 -5.48735335e-02 -3.21192801e-01 -2.31855363e-01
-1.17465973e+00 7.22271085e-01 1.00135279e+00 3.76873374e-01
-2.77404517e-01 5.27280748e-01 7.55186319e-01 -2.02138707e-01
6.68326855e-01 4.38066870e-01 -4.90971893e-01 -1.64523110e-01
-1.10388589e+00 -1.11546822e-01 4.40427065e-01 1.18164015e+00
2.85597444e-01 -6.37927353e-02 -6.03198171e-01 6.02087259e-01
4.22809154e-01 3.51067275e-01 6.92410231e-01 -7.95411170e-01
1.11602306e-01 4.64582503e-01 -5.27461290e-01 -5.02485812e-01
-8.83457720e-01 -7.99991608e-01 -9.68537092e-01 -1.09016120e-01
3.99766356e-01 7.18664303e-02 -1.42014110e+00 1.51518047e+00
2.57989824e-01 3.19403633e-02 -3.27363938e-01 9.29517269e-01
1.20441079e+00 2.93102831e-01 7.29878098e-02 5.12400381e-02
1.46037042e+00 -1.05405450e+00 -7.95109332e-01 1.15532331e-01
8.67712438e-01 -4.36427295e-01 8.15485060e-01 3.30872864e-01
-1.39186800e+00 -2.70421177e-01 -1.05142903e+00 -5.16432106e-01
-1.59289259e-02 -1.33367358e-02 5.74963033e-01 7.36231565e-01
-1.16287410e+00 7.39519238e-01 -1.35867357e+00 7.06007928e-02
9.74825680e-01 4.52142149e-01 8.98379385e-02 4.90912162e-02
-1.25773704e+00 1.02529395e+00 4.59089041e-01 2.67947018e-01
-1.07312989e+00 -1.01014435e+00 -1.07381439e+00 1.38945937e-01
3.84376138e-01 -7.46748984e-01 1.64147007e+00 -3.93094033e-01
-1.63905501e+00 8.79942417e-01 1.35364726e-01 -7.99090624e-01
6.89633846e-01 -7.41327927e-02 3.80752355e-01 2.58966029e-01
1.15202934e-01 9.35072601e-01 4.22755450e-01 -9.73000348e-01
-2.02115953e-01 -4.06041890e-01 1.83807723e-02 2.39320457e-01
-1.12677598e-02 -3.77221256e-01 -6.35736406e-01 -6.13066435e-01
1.76664814e-01 -6.96166098e-01 -5.66649914e-01 -5.98076358e-02
-6.03556156e-01 -1.43592343e-01 1.82228982e-01 -6.99696720e-01
1.04420376e+00 -1.99090207e+00 1.02889195e-01 -3.37388963e-02
5.66395164e-01 4.35367674e-02 2.28422374e-01 -4.46868002e-01
9.17540267e-02 2.65920579e-01 -6.29016340e-01 -4.18290764e-01
-3.85688692e-02 2.25727648e-01 3.69600803e-01 3.74565691e-01
1.67105809e-01 1.25559711e+00 -8.63836825e-01 -9.46016550e-01
4.18046594e-01 6.33542240e-01 -8.26817989e-01 6.77463114e-02
-2.23531440e-01 6.03642046e-01 -4.67965931e-01 7.89524794e-01
7.25908518e-01 -5.35495102e-01 -2.29083039e-02 -4.79897618e-01
7.13291205e-03 4.59681123e-01 -4.47657198e-01 2.40381885e+00
-4.39885408e-01 2.58396775e-01 9.05139893e-02 -1.06495690e+00
4.83736962e-01 4.28173065e-01 8.26876700e-01 -8.69994760e-01
7.78968751e-01 3.11538428e-01 7.77482614e-02 -2.95670092e-01
2.42230579e-01 -4.01318580e-01 -2.41680309e-01 8.34424049e-02
4.23227131e-01 -4.69211906e-01 5.97272813e-02 1.84328616e-01
1.00764179e+00 2.25090802e-01 2.68618166e-01 -6.10163629e-01
2.31015906e-01 1.45522192e-01 5.85256636e-01 8.13367605e-01
-6.13829315e-01 8.16490471e-01 5.70566237e-01 -3.48324567e-01
-9.33993578e-01 -1.03214049e+00 -6.07120574e-01 5.57298303e-01
2.06394896e-01 -3.89025718e-01 -1.02788651e+00 -7.33271122e-01
-1.86778411e-01 6.59307837e-01 -7.69401968e-01 -1.50994837e-01
-4.08508658e-01 -9.43450451e-01 7.71285295e-01 7.51614153e-01
4.48586226e-01 -6.66791081e-01 -8.92954230e-01 2.84106612e-01
-3.24876755e-01 -1.07102358e+00 -5.85959077e-01 5.34348428e-01
-1.25311279e+00 -7.46501625e-01 -1.11658740e+00 -6.28554761e-01
5.87026834e-01 -3.47612381e-01 1.09004414e+00 -1.96258441e-01
-5.31004250e-01 1.76291957e-01 -1.50584534e-01 -5.19720912e-01
-2.98920393e-01 2.71494627e-01 -3.70157063e-01 -7.79798150e-01
2.13692769e-01 -2.91852534e-01 -8.01977515e-01 7.88841620e-02
-7.31150448e-01 3.38330060e-01 5.92679679e-01 9.25428152e-01
9.04506147e-01 -1.96048722e-01 2.84553409e-01 -7.80670583e-01
3.23130399e-01 -3.59188199e-01 -5.62093019e-01 1.15661241e-01
-7.49687195e-01 3.48198295e-01 3.79988790e-01 -2.61109192e-02
-1.04034412e+00 -9.26162526e-02 -6.39489055e-01 -4.34136361e-01
1.17147870e-01 5.34698784e-01 1.16020240e-01 1.09280497e-01
3.39261591e-01 1.97476044e-01 1.28720060e-01 -3.91713083e-01
1.37526095e-01 2.98112750e-01 2.79929698e-01 -5.53429723e-01
-4.72111143e-02 3.71552140e-01 -1.24820553e-01 -3.04553986e-01
-8.47631931e-01 -1.44293666e-01 -5.03777266e-01 -1.57503694e-01
1.14658821e+00 -1.00547266e+00 -8.75856340e-01 4.76392239e-01
-1.13948643e+00 -4.32473302e-01 -6.90142810e-01 6.70663893e-01
-6.09337568e-01 4.12651896e-01 -1.08351934e+00 -4.29738104e-01
-8.38779986e-01 -1.79491031e+00 1.15192246e+00 2.38181174e-01
9.01972204e-02 -9.19023514e-01 -1.34292513e-01 4.06857729e-01
6.23930454e-01 1.51037559e-01 8.89646173e-01 -6.76451385e-01
-7.86887109e-01 -1.42445518e-02 -4.58628416e-01 3.76914412e-01
-5.30666672e-02 -4.93530482e-01 -8.05220127e-01 5.07470332e-02
2.93794364e-01 -2.99799234e-01 1.08702052e+00 1.12977946e+00
1.52861166e+00 2.50882190e-02 -2.50815451e-01 8.56730282e-01
1.41244757e+00 2.22567633e-01 4.50853288e-01 3.33101451e-02
8.68810713e-01 6.41967207e-02 1.38568759e-01 4.69125330e-01
6.74835324e-01 3.47097754e-01 6.26251101e-01 -2.20734015e-01
-2.95048773e-01 4.95711640e-02 2.55649071e-02 1.25655520e+00
1.51573112e-02 -3.54452908e-01 -1.21080470e+00 6.00762248e-01
-1.65483952e+00 -4.36178207e-01 -5.65090179e-02 1.94943857e+00
9.74559128e-01 3.00801247e-01 -3.28204185e-01 -2.36788079e-01
4.93888110e-01 -3.35730761e-02 -7.83040941e-01 -4.20611441e-01
3.09635401e-01 3.24794412e-01 6.92012429e-01 5.58801651e-01
-1.16307282e+00 7.04178393e-01 6.03221369e+00 1.02009797e+00
-9.77821887e-01 4.48610127e-01 9.76610541e-01 -2.20012978e-01
-4.47109848e-01 -3.52752626e-01 -7.06883490e-01 4.13960904e-01
9.08596337e-01 -1.24060631e-01 -1.32320374e-01 6.99164867e-01
-1.76568497e-02 -2.59471446e-01 -1.13243032e+00 9.09461617e-01
7.09286779e-02 -1.33555162e+00 -2.54318025e-02 -1.12648576e-01
4.76008475e-01 5.37431598e-01 4.24155705e-02 3.25273633e-01
2.00981736e-01 -1.07789063e+00 6.28880203e-01 4.35756654e-01
7.73863196e-01 -5.13617456e-01 1.07683885e+00 2.50560567e-02
-9.21937406e-01 2.45464817e-01 -2.09956586e-01 4.65058059e-01
4.68277544e-01 1.03880548e+00 -9.97650981e-01 7.54228592e-01
8.31368625e-01 6.93704724e-01 -2.92480707e-01 1.28229856e+00
-2.24073394e-03 4.32107806e-01 -6.25014782e-01 2.87354849e-02
3.85560840e-01 1.21262364e-01 5.15099645e-01 1.23334026e+00
4.35470998e-01 1.67627364e-01 -2.67341398e-02 1.05907476e+00
-2.82636851e-01 6.44185394e-02 -1.12646312e-01 1.07767597e-01
-1.33804068e-01 1.03421652e+00 -1.09010983e+00 -4.80355084e-01
-1.86907709e-01 8.62996697e-01 2.29662023e-02 -1.76039353e-01
-1.35319471e+00 -2.07513452e-01 1.25289828e-01 6.54816851e-02
2.01812610e-01 -2.70036589e-02 -6.90228641e-01 -1.25314772e+00
-8.69195163e-02 -4.27498013e-01 4.04496074e-01 -5.85435092e-01
-1.03082061e+00 8.96207869e-01 4.42308746e-02 -9.39936221e-01
8.78887847e-02 -5.71389735e-01 -1.34034291e-01 5.23611665e-01
-1.55262733e+00 -1.03044403e+00 -2.52366394e-01 3.16356659e-01
6.11638486e-01 4.14461821e-01 7.64578938e-01 5.85218847e-01
-5.02173066e-01 6.94446266e-01 7.06964880e-02 1.50622785e-01
4.79674518e-01 -1.58144891e+00 2.34858871e-01 5.40109515e-01
-2.81300485e-01 2.37399444e-01 3.25295508e-01 -6.19561434e-01
-1.12123477e+00 -8.49767029e-01 5.99696815e-01 -3.88838261e-01
3.44620824e-01 -2.55276531e-01 -6.85367703e-01 7.15344906e-01
3.37448120e-01 9.10074636e-02 6.05646670e-01 -1.22449346e-01
5.01365438e-02 1.10239564e-02 -1.36019349e+00 2.08052903e-01
8.98470163e-01 -1.18107989e-01 -5.09965122e-01 3.12014908e-01
1.09740627e+00 -1.07750714e+00 -1.37635434e+00 7.44057238e-01
4.37548012e-01 -8.22942257e-01 8.39084506e-01 -2.16025729e-02
6.83313310e-01 1.24589540e-02 9.79412571e-02 -1.07955265e+00
-2.20365450e-01 -1.87554449e-01 -5.45961112e-02 8.15916002e-01
5.51869273e-01 -5.07764339e-01 7.33623385e-01 8.45893145e-01
-7.85340369e-01 -1.02792847e+00 -1.36269855e+00 -5.44206202e-01
4.71890301e-01 -5.48489869e-01 3.97552431e-01 6.51477873e-01
3.09105724e-01 4.37657833e-02 -1.64261371e-01 -1.44716144e-01
8.34995985e-01 -9.93723422e-02 -8.31939057e-02 -7.77782440e-01
-1.46716386e-01 -6.21859491e-01 -3.89745533e-01 -1.11082375e+00
-2.99897403e-01 -1.36141145e+00 4.23181176e-01 -1.77303028e+00
4.20100600e-01 -4.79707927e-01 -5.82506001e-01 5.17352998e-01
3.78856026e-02 1.11192711e-01 1.63353726e-01 -1.45561859e-01
-7.40969479e-01 7.52338409e-01 1.67004502e+00 -3.20258379e-01
2.79807718e-03 -1.48228511e-01 -5.60115695e-01 5.24053454e-01
9.63157356e-01 -6.31007314e-01 -3.71841103e-01 -8.50330472e-01
-5.63476495e-02 2.28832185e-01 2.53822535e-01 -1.05224371e+00
3.76125127e-01 4.09430623e-01 3.88428569e-01 -8.29877496e-01
1.28721207e-01 -7.45380402e-01 -2.02868164e-01 6.94983721e-01
-4.00563419e-01 -1.27556100e-01 6.08753324e-01 3.10566694e-01
-1.44251272e-01 -2.80853212e-01 9.09643292e-01 -5.01601160e-01
-3.12211633e-01 5.47138751e-01 -3.96679044e-01 3.34030300e-01
8.91519904e-01 1.68465241e-03 -5.99143505e-02 6.28566071e-02
-9.89789426e-01 4.87022191e-01 1.52267843e-01 1.19210526e-01
6.45529270e-01 -1.05142295e+00 -6.77105367e-01 5.04988730e-02
-1.25175565e-01 6.59403980e-01 7.69523740e-01 1.26485288e+00
-7.94232190e-01 7.03181386e-01 -2.46198684e-01 -1.12826407e+00
-7.36115515e-01 4.75290060e-01 7.04035461e-01 -6.62729502e-01
-9.28835928e-01 1.22706008e+00 3.35994691e-01 -4.52651352e-01
2.10238740e-01 -1.18730748e+00 1.13739841e-01 -5.02951480e-02
2.60626614e-01 7.41821900e-02 2.54616201e-01 -2.46156022e-01
-6.16055071e-01 3.81077528e-01 -3.61305296e-01 9.17205289e-02
1.23808968e+00 -4.53574955e-02 -1.39349820e-02 2.29065925e-01
1.25786698e+00 -6.04010522e-01 -1.22057545e+00 -1.80875137e-01
-7.83014148e-02 -1.20656081e-01 5.35248876e-01 -9.69722450e-01
-1.44130492e+00 1.06226301e+00 8.21417630e-01 -1.38400018e-01
1.16076958e+00 1.56435713e-01 1.06487167e+00 -1.01145715e-01
3.35075557e-01 -1.06123173e+00 -2.32478738e-01 3.94722819e-01
6.11290038e-01 -1.51366210e+00 -5.29085286e-03 -2.54243016e-01
-5.99772274e-01 8.85947347e-01 5.59785068e-01 2.17052195e-02
8.54212463e-01 6.31444395e-01 -5.03905714e-02 -3.65955144e-01
-4.92807180e-01 -9.23514813e-02 1.74265519e-01 2.67386615e-01
6.91591144e-01 1.19907603e-01 -3.24542642e-01 7.97232032e-01
6.70081452e-02 3.16169947e-01 4.76461262e-01 6.62367105e-01
-1.54829875e-01 -7.94940412e-01 3.78645696e-02 6.15008533e-01
-7.71933317e-01 -4.56000924e-01 3.15753788e-01 6.09819889e-01
5.71518578e-02 3.97791535e-01 4.67170700e-02 -4.68085781e-02
1.16469368e-01 7.66784372e-03 8.45465481e-01 -5.18072426e-01
-7.91439950e-01 3.33701193e-01 -2.01392233e-01 -6.54078901e-01
-2.92368472e-01 -5.07211447e-01 -1.67421269e+00 -3.09957433e-02
-4.17279124e-01 1.43951043e-01 7.96627760e-01 7.65191495e-01
3.33243906e-01 1.17664921e+00 4.49474566e-02 -7.85429895e-01
-4.49042290e-01 -9.23340678e-01 -2.87238955e-01 -9.99862179e-02
2.38128364e-01 -5.25441229e-01 -4.96117510e-02 -2.39657089e-01] | [14.57255744934082, -2.4575862884521484] |
159e5eb6-3873-46e5-83f0-148705f71564 | deep-learning-for-robust-motion-segmentation | 2102.10929 | null | https://arxiv.org/abs/2102.10929v1 | https://arxiv.org/pdf/2102.10929v1.pdf | Deep Learning for Robust Motion Segmentation with Non-Static Cameras | This work proposes a new end-to-end DCNN based approach for motion segmentation, especially for video sequences captured with such non-static cameras, called MOSNET. While other approaches focus on spatial or temporal context only, the proposed approach uses 3D convolutions as a key technology to factor in, spatio-temporal features in cohesive video frames. This is done by capturing temporal information in features with a low and also with a high level of abstraction. The lean network architecture with about 21k trainable parameters is mainly based on a pre-trained VGG-16 network. The MOSNET uses a new feature map fusion technique, which enables the network to focus on the appropriate level of abstraction, resolution, and the appropriate size of the receptive field regarding the input. Furthermore, the end-to-end deep learning based approach can be extended by feature based image alignment as a pre-processing step, which brings a gain in performance for some scenes. Evaluating the end-to-end deep learning based MOSNET network in a scene independent manner leads to an overall F-measure of 0.803 on the CDNet2014 dataset. A small temporal window of five input frames, without the need of any initialization is used to obtain this result. Therefore the network is able to perform well on scenes captured with non-static cameras where the image content changes significantly during the scene. In order to get robust results in scenes captured with a moving camera, feature based image alignment can implemented as pre-processing step. The MOSNET combined with pre-processing leads to an F-measure of 0.685 when cross-evaluating with a relabeled LASIESTA dataset, which underpins the capability generalise of the MOSNET. | ['Markus Bosch'] | 2021-02-22 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 6.51357025e-02 -2.03122348e-01 1.47688106e-01 -3.71002376e-01
-2.81866193e-01 -4.36935723e-01 6.33751690e-01 6.11157306e-02
-1.15166509e+00 4.28704977e-01 -1.36079133e-01 6.43853918e-02
-2.51711041e-01 -6.21089816e-01 -6.01751924e-01 -7.05295205e-01
-1.65835753e-01 1.16538763e-01 7.50359535e-01 -9.81402621e-02
3.11960187e-02 7.40492344e-01 -1.70195556e+00 3.66329968e-01
2.48679399e-01 9.88561392e-01 4.94659334e-01 9.55173612e-01
2.42886562e-02 5.96355796e-01 -5.44452310e-01 1.27117969e-02
4.06515777e-01 -2.61753470e-01 -7.90132940e-01 1.54575512e-01
6.93561614e-01 -3.34056228e-01 -1.74009278e-01 7.95217812e-01
4.44099844e-01 3.93929213e-01 4.78428781e-01 -8.13506663e-01
3.73399436e-01 3.36592585e-01 -9.48898047e-02 4.41277266e-01
1.17120452e-01 4.48193878e-01 5.58495998e-01 -5.14346063e-01
7.91066587e-01 9.47297096e-01 4.97394919e-01 3.18601191e-01
-1.04275465e+00 -2.66248047e-01 2.75505101e-03 3.33570480e-01
-1.13267934e+00 -2.58104205e-01 7.26604223e-01 -5.18334389e-01
1.02789879e+00 5.43562472e-02 8.34959745e-01 8.95595133e-01
1.50721848e-01 3.92555684e-01 9.05796468e-01 -3.51564705e-01
2.61749804e-01 1.01537323e-02 -8.93321261e-02 3.69460762e-01
1.04526252e-01 8.85806754e-02 -2.08377421e-01 6.34221017e-01
7.43770957e-01 1.04869768e-01 -8.03004876e-02 -2.68637300e-01
-1.31309617e+00 5.76142073e-01 5.75499177e-01 9.30752039e-01
-5.27220190e-01 2.79172122e-01 6.90420389e-01 2.39217073e-01
1.71245396e-01 2.66207188e-01 -5.26135445e-01 -1.29350826e-01
-1.41568387e+00 9.95247588e-02 4.65553254e-01 3.41825396e-01
7.68628597e-01 2.56516844e-01 -1.52665898e-01 4.46399182e-01
1.45141661e-01 3.32074553e-01 5.92250049e-01 -8.96646857e-01
2.78152913e-01 5.56367040e-01 -1.04968242e-01 -8.76870632e-01
-7.70278275e-01 -5.32739937e-01 -7.78466880e-01 7.81032622e-01
7.09468842e-01 -2.11005583e-01 -1.21158707e+00 1.69608068e+00
4.00051504e-01 1.00836687e-01 2.03063533e-01 1.06799507e+00
7.10825145e-01 5.99159420e-01 2.31419858e-02 -4.89119068e-02
1.42668140e+00 -8.39193046e-01 -4.31489915e-01 -2.90072355e-02
5.03980637e-01 -7.68135428e-01 9.36368704e-01 5.05694091e-01
-8.63388002e-01 -9.25761700e-01 -1.11013901e+00 1.00808799e-01
-5.85919380e-01 2.42307216e-01 1.94272995e-01 4.74026591e-01
-1.24865019e+00 6.12878978e-01 -8.02815080e-01 -7.12099314e-01
1.94938228e-01 6.50095522e-01 -7.26344347e-01 7.39986310e-03
-9.66014028e-01 9.17536616e-01 9.01388466e-01 2.39192039e-01
-9.42799985e-01 -3.65000784e-01 -6.48613095e-01 6.87369332e-02
3.09207141e-01 -6.89502060e-01 6.78003728e-01 -1.40758920e+00
-1.50937855e+00 7.14297295e-01 -4.07837294e-02 -6.99014366e-01
7.50785291e-01 1.91447337e-03 -1.94792882e-01 6.16173089e-01
-7.51587376e-02 9.57105696e-01 1.04836893e+00 -9.46886003e-01
-8.96051049e-01 -4.42301035e-02 3.22199643e-01 2.28723720e-01
-2.57494122e-01 1.00704812e-01 -4.82105583e-01 -5.23057580e-01
-2.15547234e-01 -8.71135056e-01 -1.72857881e-01 -2.75054157e-01
9.65254605e-02 1.06877968e-01 1.05448878e+00 -6.17927909e-01
8.14066768e-01 -2.05696058e+00 1.96198851e-01 1.59795150e-01
-3.84444866e-05 6.77322328e-01 -1.55497402e-01 1.66372329e-01
-1.83597833e-01 -1.87064931e-01 -2.52607197e-01 -2.69502431e-01
-4.14833963e-01 -2.47804094e-02 3.39338511e-01 4.20575082e-01
3.03141981e-01 6.31860793e-01 -6.70816720e-01 -3.91799361e-01
8.94255638e-01 5.57979405e-01 -4.42663819e-01 1.54390782e-01
-1.29277706e-01 5.92590272e-01 -1.47152301e-02 1.08906880e-01
5.18810511e-01 2.37867028e-01 -1.56363145e-01 -2.93751955e-01
-3.18865597e-01 -2.71371514e-01 -1.36927092e+00 1.87819922e+00
-4.63636339e-01 9.28089976e-01 7.93135092e-02 -1.22852731e+00
9.34185326e-01 4.58664775e-01 6.54849291e-01 -5.59987724e-01
4.53102350e-01 2.02395573e-01 7.36722127e-02 -6.04982376e-01
4.00039852e-01 8.05859268e-02 1.50854543e-01 4.17546555e-02
4.56734031e-01 1.59101367e-01 5.51325083e-01 -1.89775869e-01
8.96692514e-01 3.39258641e-01 3.77456099e-02 -3.26671243e-01
9.32765007e-01 9.97961964e-03 2.45486870e-01 4.85736728e-01
-9.84520540e-02 7.84353554e-01 2.26465955e-01 -7.45044529e-01
-1.11900687e+00 -6.57173872e-01 6.49133995e-02 7.66516268e-01
1.77071188e-02 -5.86038902e-02 -9.51559663e-01 -6.76752150e-01
-5.86017907e-01 2.88451284e-01 -6.33848786e-01 -5.07480837e-03
-7.17493415e-01 -4.38417524e-01 4.43591028e-01 3.06112438e-01
8.40393066e-01 -1.25975156e+00 -1.17883801e+00 3.98236692e-01
-8.82931054e-02 -1.32865393e+00 2.94749923e-02 4.40158606e-01
-1.07018578e+00 -9.56151366e-01 -8.69010746e-01 -6.81039512e-01
4.42788452e-01 -9.66221502e-04 8.11576903e-01 -1.68621510e-01
-1.65688977e-01 2.42195189e-01 -4.67337757e-01 -9.83498096e-02
-2.04818770e-01 2.65049994e-01 -1.78582862e-01 5.32756932e-02
3.07087034e-01 -4.65671688e-01 -6.68733358e-01 2.71294862e-01
-1.20056796e+00 -1.76944304e-02 5.85626304e-01 6.52388513e-01
3.77759874e-01 1.01999298e-01 3.07990134e-01 -3.00398737e-01
7.55183026e-02 -1.34867057e-01 -7.06241965e-01 -2.06068642e-02
-1.83117539e-01 1.72521006e-02 8.63612950e-01 -4.48847204e-01
-8.06035399e-01 4.25351709e-01 -3.40562493e-01 -4.65931177e-01
-6.82764947e-01 3.53418052e-01 -1.74223274e-01 -1.07497834e-01
6.06532812e-01 -2.87454687e-02 2.29276996e-02 -2.67037153e-01
1.92552701e-01 2.75286645e-01 5.85313439e-01 -5.34655675e-02
7.24904537e-01 5.11876047e-01 2.50078976e-01 -9.40165401e-01
-3.41003448e-01 -6.03986382e-01 -9.99083757e-01 -4.79360372e-01
1.42143416e+00 -8.54347765e-01 -5.80924511e-01 7.18307436e-01
-1.06472421e+00 -4.46923077e-01 -3.09524596e-01 8.91535163e-01
-5.87566733e-01 1.61519840e-01 -2.83301532e-01 -5.67738950e-01
-3.19471419e-01 -1.27316070e+00 6.67642117e-01 3.90665084e-01
-1.12154678e-01 -1.03260565e+00 -1.71203673e-01 2.46861905e-01
3.77802432e-01 4.73468006e-01 4.27078605e-01 -7.00169623e-01
-6.59580946e-01 -2.52595216e-01 -1.94695622e-01 6.77640140e-01
-8.78643021e-02 1.51610374e-01 -1.09480059e+00 -1.85969904e-01
-9.14446078e-03 -9.17856675e-03 1.08633924e+00 6.87129080e-01
6.73087001e-01 2.00166345e-01 1.58794224e-01 6.94134653e-01
1.67760503e+00 3.15022826e-01 7.33911395e-01 6.71224415e-01
7.42577910e-01 5.64430416e-01 8.20028961e-01 1.41049072e-01
8.59020352e-02 8.15358222e-01 7.56070971e-01 -4.87744838e-01
-2.82084584e-01 3.62057894e-01 4.31139827e-01 3.23589236e-01
-2.32818425e-01 -2.25704357e-01 -7.78760672e-01 7.24225760e-01
-1.78275657e+00 -1.04736125e+00 -1.68000847e-01 2.19014525e+00
2.10935846e-01 3.56027156e-01 3.52999568e-01 4.26071376e-01
6.25983179e-01 2.73997545e-01 -2.01944917e-01 -6.41413510e-01
-1.54648155e-01 3.40908647e-01 7.14988053e-01 4.56199378e-01
-1.29780614e+00 9.59349453e-01 5.00516987e+00 9.10140097e-01
-1.68215764e+00 -1.61962789e-02 3.95626009e-01 -1.25328451e-01
3.93638343e-01 -7.39114434e-02 -6.93403840e-01 5.67938328e-01
1.08334196e+00 4.59899783e-01 2.62202591e-01 6.68065488e-01
6.10079885e-01 -4.32453305e-01 -8.76152873e-01 1.05220866e+00
4.39621620e-02 -1.22408402e+00 -1.69543847e-01 7.11658373e-02
6.50626063e-01 1.35162637e-01 -1.56859607e-01 2.19815820e-02
-4.38764602e-01 -1.05484569e+00 8.59787285e-01 6.75910950e-01
5.24152040e-01 -8.67108047e-01 1.23338699e+00 2.64104217e-01
-1.28704345e+00 -1.83354020e-01 -2.03166828e-01 -1.11954518e-01
3.11752051e-01 4.56926912e-01 -7.26426005e-01 7.00256288e-01
9.05401051e-01 5.10929286e-01 -5.29095054e-01 1.20952225e+00
-3.19001414e-02 4.01093066e-01 -4.80550557e-01 1.45999596e-01
6.85777903e-01 -4.59399261e-02 5.76777756e-01 1.59973669e+00
4.44381893e-01 -2.89763302e-01 -9.13777128e-02 4.17433947e-01
2.72046089e-01 1.02470972e-01 -5.20415068e-01 2.75323004e-01
-1.66742295e-01 1.49741912e+00 -1.12876701e+00 -4.19319421e-01
-1.96228728e-01 1.01517689e+00 -1.63531721e-01 3.93353164e-01
-8.21182489e-01 -5.48290968e-01 4.64955777e-01 2.01902077e-01
8.08691561e-01 -3.87439519e-01 8.89725704e-03 -8.73657584e-01
4.15547453e-02 -7.67478585e-01 1.82115957e-01 -6.59654796e-01
-6.37793183e-01 8.48183632e-01 2.93354928e-01 -1.30216205e+00
-5.75199127e-01 -8.30419481e-01 -7.35594630e-01 8.60709250e-01
-1.35241234e+00 -1.17668211e+00 -5.16449034e-01 7.55125761e-01
5.80797493e-01 -1.91537648e-01 4.40097809e-01 4.35153246e-01
-3.85897160e-01 2.24536836e-01 -1.03617325e-01 2.07536489e-01
5.80308676e-01 -1.17014861e+00 8.90851244e-02 1.26670778e+00
8.58353898e-02 2.25125790e-01 6.51941478e-01 -2.40296155e-01
-8.09356928e-01 -1.12627459e+00 7.25158870e-01 -1.44203037e-01
2.01537952e-01 -2.13915050e-01 -8.41871738e-01 1.11311615e-01
3.87519866e-01 1.75701246e-01 1.54599816e-01 -4.05509233e-01
2.53921062e-01 -4.57385242e-01 -1.20381594e+00 2.31301919e-01
6.12553537e-01 -3.61410111e-01 -5.24097919e-01 -1.12792358e-01
5.14745295e-01 -2.31641591e-01 -8.25764954e-01 4.62434649e-01
5.55109382e-01 -1.32011735e+00 8.53392839e-01 -1.52549222e-01
2.96621561e-01 -6.91978872e-01 -2.92711537e-02 -9.77478564e-01
-6.74814358e-02 -3.06077421e-01 3.27737123e-01 1.19087470e+00
2.47144371e-01 -3.94594967e-01 7.98628330e-01 2.21524104e-01
-4.47627679e-02 -4.21589553e-01 -9.90440071e-01 -7.08028793e-01
-3.05770665e-01 -6.00907385e-01 2.02493787e-01 6.69813633e-01
-6.40939116e-01 1.48092702e-01 -2.97838748e-01 -2.13210322e-02
3.73260498e-01 -3.53036046e-01 8.89123619e-01 -1.21073627e+00
-1.95716143e-01 -5.20375729e-01 -1.02594459e+00 -6.16590738e-01
3.04259658e-02 -6.24212980e-01 -1.19613171e-01 -1.46691716e+00
-1.83523715e-01 -1.05165228e-01 -4.20289397e-01 3.25470209e-01
1.31636947e-01 6.15359128e-01 4.20875132e-01 -1.89570412e-01
-3.94285947e-01 1.16035223e-01 9.54685569e-01 8.31875112e-03
-3.61300737e-01 -4.33029532e-02 1.69205412e-01 7.78789818e-01
8.50727856e-01 -3.78957152e-01 -3.97106290e-01 -3.05628598e-01
-1.07528448e-01 -1.91795588e-01 6.10965192e-01 -1.53625667e+00
1.78241163e-01 1.42304808e-01 6.31021857e-01 -6.79623604e-01
3.91526431e-01 -1.27961791e+00 3.10667336e-01 6.40954256e-01
-4.40316536e-02 1.42824957e-02 4.32304174e-01 2.07160518e-01
-4.83582705e-01 -4.74919409e-01 9.93435860e-01 -2.25778997e-01
-1.14438701e+00 9.23343003e-02 -5.68807125e-01 -2.98651248e-01
1.13793826e+00 -8.41845930e-01 6.34116083e-02 -2.63201773e-01
-9.25992966e-01 -1.41880333e-01 4.90277410e-01 2.80223370e-01
3.88650537e-01 -9.32368517e-01 -4.74545002e-01 2.23154142e-01
-1.11930653e-01 4.45070341e-02 4.68210667e-01 1.00967991e+00
-9.42869365e-01 4.01927978e-01 -6.78812861e-01 -9.27188277e-01
-1.55877686e+00 3.98775727e-01 4.99895751e-01 -2.10306630e-01
-3.70913267e-01 5.17023802e-01 -7.45888278e-02 2.72027198e-02
1.71847761e-01 -6.68314278e-01 -6.02609813e-01 4.19262588e-01
3.17618728e-01 3.16705614e-01 2.66494721e-01 -9.03734326e-01
-4.51980293e-01 9.49472308e-01 2.47800350e-01 -4.15791303e-01
1.32421505e+00 -1.18557207e-01 8.70787948e-02 3.85145843e-01
1.39146459e+00 -1.52529165e-01 -1.49635673e+00 1.30938262e-01
-6.77906573e-02 -3.43567371e-01 2.82652944e-01 -5.94585299e-01
-1.27077627e+00 1.01065171e+00 1.06229877e+00 -1.36232123e-01
1.42626965e+00 -4.16743189e-01 3.83906782e-01 1.94382414e-01
1.33077011e-01 -1.07793045e+00 2.83342097e-02 6.44529104e-01
5.53814113e-01 -1.23535168e+00 -1.19792625e-01 1.71549127e-01
-5.71293354e-01 1.49203384e+00 4.37921375e-01 -2.18372568e-01
4.03959602e-01 8.81095901e-02 1.73948184e-01 -7.22765028e-02
-3.30083132e-01 -6.81626797e-01 4.43179846e-01 5.53341091e-01
2.72043675e-01 -2.45272681e-01 -3.31066161e-01 1.17697060e-01
3.26186977e-02 1.46165311e-01 3.92434925e-01 7.88999379e-01
-4.34833497e-01 -1.08009517e+00 -2.96950907e-01 4.30941582e-02
-6.64331496e-01 5.99226542e-02 -8.36429521e-02 1.09507954e+00
6.42799973e-01 8.63212883e-01 1.83742747e-01 -5.08391321e-01
2.74838269e-01 -1.81622542e-02 3.81023824e-01 -2.65972376e-01
-1.05385733e+00 1.78254515e-01 -5.66283874e-02 -6.94571793e-01
-8.98209333e-01 -5.33627570e-01 -1.21773565e+00 -8.63473788e-02
-2.04917774e-01 -1.19896978e-01 9.73987818e-01 1.05310023e+00
3.61243635e-02 7.23220766e-01 4.22396541e-01 -1.26481211e+00
1.76495656e-01 -9.63419259e-01 -2.49318287e-01 4.42614436e-01
4.71677005e-01 -4.32423562e-01 -2.37736344e-01 3.05087686e-01] | [8.946513175964355, -0.712940514087677] |
a100fda1-305b-4ea1-9ba8-cdfd71f8e60f | cross-modal-retrieval-augmentation-for-multi-1 | 2104.08108 | null | https://arxiv.org/abs/2104.08108v1 | https://arxiv.org/pdf/2104.08108v1.pdf | Cross-Modal Retrieval Augmentation for Multi-Modal Classification | Recent advances in using retrieval components over external knowledge sources have shown impressive results for a variety of downstream tasks in natural language processing. Here, we explore the use of unstructured external knowledge sources of images and their corresponding captions for improving visual question answering (VQA). First, we train a novel alignment model for embedding images and captions in the same space, which achieves substantial improvement in performance on image-caption retrieval w.r.t. similar methods. Second, we show that retrieval-augmented multi-modal transformers using the trained alignment model improve results on VQA over strong baselines. We further conduct extensive experiments to establish the promise of this approach, and examine novel applications for inference time such as hot-swapping indices. | ['Austin Reiter', 'Douwe Kiela', 'Ser-Nam Lim', 'Chris Stauffer', 'Natalia Neverova', 'Shir Gur'] | 2021-04-16 | cross-modal-retrieval-augmentation-for-multi | https://aclanthology.org/2021.findings-emnlp.11 | https://aclanthology.org/2021.findings-emnlp.11.pdf | findings-emnlp-2021-11 | ['multi-modal-classification'] | ['miscellaneous'] | [ 2.19966888e-01 1.12426095e-01 -3.28868866e-01 -4.10167098e-01
-1.83640385e+00 -9.45577800e-01 8.63964856e-01 3.84237431e-02
-4.87116247e-01 4.24049020e-01 7.62459934e-01 -2.43404999e-01
2.62503326e-01 -6.04182899e-01 -1.10685968e+00 -4.81580198e-01
3.30304623e-01 5.40497839e-01 2.84470022e-01 -2.51530975e-01
2.94325143e-01 2.24886268e-01 -1.16992772e+00 8.22251201e-01
4.11599070e-01 7.00728178e-01 1.19565316e-02 5.84640920e-01
-2.13181198e-01 9.67096567e-01 -5.73728204e-01 -8.51774693e-01
2.42417380e-01 -2.09788188e-01 -1.14583254e+00 -2.55060429e-03
1.02448630e+00 -6.51815772e-01 -6.07954562e-01 8.54488850e-01
4.44533348e-01 1.78942233e-01 7.43992507e-01 -1.17170966e+00
-1.51381135e+00 3.34096074e-01 -6.49664223e-01 3.98037672e-01
2.84469336e-01 4.73228544e-01 1.38818336e+00 -1.06297410e+00
1.04106784e+00 1.57125902e+00 1.50042325e-01 6.23191953e-01
-1.38534689e+00 -3.26355338e-01 1.26713872e-01 4.86038864e-01
-1.26906645e+00 -5.38460493e-01 4.83234346e-01 -2.53547251e-01
1.15942097e+00 2.54593462e-01 1.78688094e-01 1.05172992e+00
-1.70961708e-01 1.29425919e+00 8.99165034e-01 -4.75174665e-01
-2.25400925e-01 3.34209293e-01 1.08431228e-01 7.65957773e-01
1.16006322e-02 -2.97511965e-01 -5.68595171e-01 -1.68645188e-01
6.47954106e-01 -4.53158319e-01 -1.88147068e-01 -5.47131598e-01
-1.32307923e+00 1.01885116e+00 8.30993235e-01 1.02770939e-01
-1.36173606e-01 7.09849417e-01 4.65844363e-01 2.14541063e-01
6.03548646e-01 7.59043396e-01 -2.32758358e-01 2.99112678e-01
-8.40292454e-01 1.80146843e-01 3.08666468e-01 1.08009553e+00
6.63941205e-01 -1.15156189e-01 -8.99307966e-01 8.67315888e-01
2.36051336e-01 9.56688762e-01 1.94717422e-01 -1.38305259e+00
7.64848948e-01 2.84904838e-01 1.18557051e-01 -9.55034375e-01
1.13244683e-01 -9.63053405e-02 -4.30723280e-01 -2.18472242e-01
2.30115995e-01 4.51121211e-01 -1.46983016e+00 1.82931828e+00
1.17383242e-01 -2.34124675e-01 3.01139832e-01 9.42775249e-01
9.19652104e-01 1.03430688e+00 4.83147591e-01 9.89733636e-02
1.61005795e+00 -1.46209443e+00 -6.49205029e-01 -4.15394187e-01
5.84663928e-01 -8.47990036e-01 1.22051060e+00 -8.69637206e-02
-1.37098205e+00 -4.61087197e-01 -6.10593319e-01 -8.11131001e-01
-4.59779471e-01 2.18282044e-01 5.31996489e-01 2.63731092e-01
-1.55703402e+00 -5.85033558e-02 -6.44279778e-01 -3.73739421e-01
6.44734621e-01 -5.38878255e-02 -4.22101498e-01 -4.74297225e-01
-1.17683113e+00 1.04542661e+00 2.00283695e-02 -6.43054545e-02
-1.38179469e+00 -7.77575970e-01 -9.11066175e-01 4.20639999e-02
1.25336766e-01 -1.09966123e+00 1.51770747e+00 -9.89642560e-01
-9.23976719e-01 1.30494988e+00 -6.40358090e-01 -4.76456702e-01
1.88963339e-01 -4.93498117e-01 -1.54802918e-01 9.25962329e-01
3.89684141e-01 1.31801987e+00 9.95480657e-01 -1.52077174e+00
-1.91516653e-01 -3.77329975e-01 3.59053403e-01 4.22139704e-01
-5.76984048e-01 1.35008857e-01 -1.06723201e+00 -5.27489305e-01
-3.14710349e-01 -7.54868925e-01 -2.34882399e-01 3.22983980e-01
-1.95076808e-01 -2.85387725e-01 4.71564233e-01 -9.42231774e-01
8.56398821e-01 -2.03248215e+00 2.51922309e-01 1.13637723e-01
7.93993399e-02 2.69936591e-01 -8.52192163e-01 4.97981012e-01
1.37133330e-01 3.67690474e-01 -3.18631879e-03 -2.94795126e-01
-2.86525162e-03 2.03971148e-01 -9.82934177e-01 1.99423246e-02
6.34304225e-01 1.87829411e+00 -8.91755581e-01 -8.86935115e-01
-5.13829477e-02 6.23702824e-01 -4.39754069e-01 1.93887785e-01
-4.84491676e-01 6.81165159e-02 -4.97269899e-01 6.44961417e-01
4.08879459e-01 -7.85254002e-01 -2.07609355e-01 -5.56093097e-01
4.02251750e-01 1.39598802e-01 -1.71661556e-01 1.82939184e+00
-3.74461114e-01 9.03340876e-01 -1.27345189e-01 -7.00111866e-01
4.89459276e-01 3.35947692e-01 1.07479148e-01 -1.09579289e+00
-1.88257262e-01 -1.53721809e-01 -5.59987664e-01 -5.44063866e-01
6.99160039e-01 5.62362978e-03 6.00530319e-02 5.00115216e-01
2.92557806e-01 -4.24954951e-01 1.17020085e-01 1.07832003e+00
1.06655788e+00 1.71081573e-01 -1.39197201e-01 1.71241254e-01
2.20868379e-01 3.88945818e-01 -3.35627556e-01 1.04781449e+00
-2.95697063e-01 7.77732491e-01 2.47800827e-01 -1.04445994e-01
-1.53851557e+00 -1.39544165e+00 7.41180079e-03 1.22356224e+00
1.83443964e-01 -3.14695418e-01 -5.28579235e-01 -7.25086570e-01
2.75490470e-02 5.60855925e-01 -7.26919174e-01 -2.76411831e-01
-4.83841360e-01 -4.59063411e-01 7.46953607e-01 7.82193422e-01
3.05736333e-01 -9.98514414e-01 -2.66734749e-01 -1.35350525e-01
-6.11315310e-01 -1.46884096e+00 -4.81480420e-01 -3.17594647e-01
-8.73734355e-01 -7.24131942e-01 -1.20325887e+00 -9.45731997e-01
6.90080345e-01 7.22533047e-01 1.59715080e+00 2.99958616e-01
-2.66020119e-01 1.32571912e+00 -2.95032352e-01 -2.29575142e-01
-1.47422269e-01 1.44113330e-02 -4.79672849e-01 -3.16547126e-01
3.89003366e-01 -8.77454318e-03 -7.92534888e-01 1.31674722e-01
-1.14386332e+00 2.13166531e-02 6.69959068e-01 7.22338080e-01
7.18832970e-01 -9.98394549e-01 5.05304873e-01 -6.05919719e-01
6.75258994e-01 -3.22932541e-01 -4.90653396e-01 8.15307796e-01
-3.09848279e-01 4.62746084e-01 2.10557748e-02 -1.95003465e-01
-1.16252482e+00 -8.28650519e-02 -5.69702648e-02 -7.25687981e-01
1.24264695e-01 2.88920015e-01 1.64115503e-01 -5.69473282e-02
6.97650969e-01 2.51560360e-01 -5.85513860e-02 -1.53840587e-01
1.21518433e+00 3.71401697e-01 7.25409567e-01 -7.20819533e-01
1.05550575e+00 8.33122075e-01 -3.35826427e-01 -5.14772952e-01
-1.17930782e+00 -7.87567317e-01 -2.92794287e-01 -5.20216078e-02
1.26818442e+00 -1.34424078e+00 -3.87265712e-01 1.26330154e-02
-1.51011622e+00 -2.47394606e-01 -3.72680396e-01 6.19107746e-02
-6.69483006e-01 4.81058002e-01 -7.47685671e-01 -5.70926070e-01
-7.39725471e-01 -1.11008680e+00 1.66749716e+00 2.93174777e-02
1.74271435e-01 -7.75469840e-01 2.84857333e-01 1.01800978e+00
4.23188299e-01 -2.69202590e-01 9.42646384e-01 -4.26936954e-01
-1.29996586e+00 9.18614194e-02 -8.13132107e-01 3.00894737e-01
-2.71412492e-01 -3.02259922e-01 -1.06485474e+00 -3.30078632e-01
-4.41765428e-01 -9.54981804e-01 1.26479125e+00 1.90447435e-01
1.05756736e+00 -1.17101692e-01 -4.13362324e-01 3.73762399e-01
1.60731649e+00 -2.40258351e-01 1.02045476e+00 4.15190578e-01
6.91012383e-01 5.97591102e-01 5.47686338e-01 -1.34010017e-01
5.30083477e-01 6.68990672e-01 3.65650743e-01 -4.25616652e-01
-5.39307535e-01 -4.07703757e-01 3.59651804e-01 8.40674162e-01
2.02183649e-01 -4.27972704e-01 -8.75122666e-01 1.28911650e+00
-1.68878543e+00 -1.02584350e+00 1.16343856e-01 1.73733354e+00
1.03848898e+00 -2.99583077e-01 -2.14361086e-01 -7.81548738e-01
5.48107386e-01 3.48529458e-01 -6.28311396e-01 -3.95368099e-01
-4.13363993e-01 3.55246574e-01 4.38414067e-01 5.54653704e-01
-9.30547357e-01 1.38170755e+00 7.20736074e+00 7.53946126e-01
-5.86510360e-01 4.48658258e-01 7.93787003e-01 -1.74293935e-01
-8.29356909e-01 1.19664751e-01 -6.05512261e-01 -1.85144201e-01
7.63993204e-01 -3.97154391e-02 3.97182018e-01 5.93424261e-01
-2.39644870e-01 3.74666192e-02 -9.99525130e-01 9.55116868e-01
6.42893255e-01 -1.60049462e+00 7.98065066e-01 -2.86098540e-01
9.20344830e-01 3.77044171e-01 3.60915601e-01 2.64196336e-01
5.52030563e-01 -9.66745198e-01 4.81264561e-01 3.98772925e-01
8.75226319e-01 -4.02497768e-01 4.68967080e-01 -3.99030536e-01
-8.96022260e-01 1.83579698e-01 -4.57108349e-01 4.82253373e-01
5.89017630e-01 1.13091901e-01 -7.47218192e-01 4.59565371e-01
7.28973389e-01 4.52441603e-01 -9.01639819e-01 8.60580862e-01
-4.27789718e-01 4.75248307e-01 -1.21449724e-01 2.13168591e-01
4.81845081e-01 2.49009356e-01 2.73166209e-01 1.16552854e+00
-7.34737813e-02 2.46664226e-01 -3.58732671e-01 8.15352082e-01
-4.45285648e-01 2.17439696e-01 -8.77671838e-01 -2.85629719e-01
1.15155004e-01 1.13193762e+00 -3.21713746e-01 -6.50739431e-01
-5.74203312e-01 1.10936475e+00 6.94909573e-01 8.08075070e-01
-7.97824264e-01 -1.50304407e-01 4.59586322e-01 -2.24491417e-01
4.82413471e-01 -9.90237594e-02 1.06726393e-01 -1.48340333e+00
1.07078753e-01 -9.53333139e-01 6.01517737e-01 -1.58495080e+00
-1.54249680e+00 5.58339357e-01 1.09251112e-01 -9.33705926e-01
-1.54377371e-01 -6.27509654e-01 -9.30204988e-02 7.03632534e-01
-1.92274904e+00 -1.46490300e+00 -1.01844549e-01 8.63703787e-01
6.64906085e-01 -7.73390755e-03 6.92785442e-01 1.38526335e-01
2.92808525e-02 7.30964363e-01 2.31260002e-01 2.85200089e-01
1.19260609e+00 -1.04848146e+00 5.32025158e-01 9.17991936e-01
7.65290201e-01 6.33905411e-01 4.67022210e-01 -5.12888491e-01
-1.59848332e+00 -1.01652753e+00 6.36062741e-01 -1.26856542e+00
7.87174940e-01 -2.71449357e-01 -8.61427844e-01 1.09660804e+00
8.49040329e-01 -8.64150524e-02 5.68515003e-01 -5.74733242e-02
-1.03367853e+00 4.79460955e-02 -7.30679274e-01 6.91689372e-01
8.95365119e-01 -1.03051925e+00 -8.91512632e-01 6.88275158e-01
1.21157253e+00 -1.87710732e-01 -7.84099519e-01 2.39626765e-01
4.26266760e-01 -3.16029370e-01 1.66359222e+00 -1.02863336e+00
9.25746143e-01 -2.05015033e-01 -2.92567581e-01 -9.98660982e-01
-2.44570166e-01 -1.97162300e-01 -6.90126196e-02 1.12338841e+00
5.30231178e-01 -1.81185603e-01 6.15843892e-01 8.21743548e-01
6.38967007e-02 -3.28255653e-01 -8.14559639e-01 -5.51883817e-01
3.29795957e-01 -1.92947030e-01 2.02808037e-01 7.19391882e-01
-1.49371266e-01 7.18443811e-01 -4.90723968e-01 6.73917606e-02
6.54935896e-01 2.99664706e-01 8.60299170e-01 -4.16337281e-01
-2.00997338e-01 -6.22158125e-02 -2.64632404e-01 -1.32339382e+00
1.90660521e-01 -1.10379517e+00 1.14436366e-01 -2.05209637e+00
8.38550925e-01 -4.60703000e-02 -4.00547028e-01 6.67200804e-01
-3.74256402e-01 8.58647287e-01 5.11196494e-01 4.17300195e-01
-1.25679731e+00 6.29150569e-01 1.49933386e+00 -7.52411842e-01
3.32663774e-01 -8.12253833e-01 -7.89584339e-01 2.12482557e-01
5.26977420e-01 -2.95514524e-01 -5.04931450e-01 -1.38563740e+00
4.66946006e-01 -7.27930143e-02 6.02042854e-01 -4.03591245e-01
1.50151923e-01 7.47721195e-02 4.03597265e-01 -5.96141160e-01
5.35238445e-01 -6.57836735e-01 -3.85907382e-01 1.07059553e-01
-7.51498282e-01 3.73403281e-01 5.08101106e-01 7.49753833e-01
-4.30040956e-01 -1.36621162e-01 4.26802069e-01 -3.76729161e-01
-8.64647031e-01 1.77856699e-01 -1.75849810e-01 4.08246100e-01
7.85253525e-01 1.95610344e-01 -1.05354822e+00 -7.92895734e-01
-5.91622293e-01 5.53297758e-01 4.28170532e-01 6.63115621e-01
8.78477991e-01 -1.50624204e+00 -7.59535968e-01 -4.24532533e-01
6.29039049e-01 -5.05950272e-01 1.80812016e-01 6.10014677e-01
-4.04560357e-01 8.35967898e-01 -1.27061844e-01 -6.38291299e-01
-1.23713887e+00 9.39528883e-01 -9.36540309e-03 -4.81088758e-01
-3.17058295e-01 8.74333978e-01 6.84714794e-01 -2.12628245e-01
6.37531057e-02 7.40710944e-02 -3.28957550e-02 -6.05080165e-02
6.27947867e-01 -1.75969377e-01 -1.29047468e-01 -7.18962312e-01
-3.35522234e-01 7.05761492e-01 -3.72878313e-01 -6.27441525e-01
1.20549786e+00 -3.01160961e-01 -2.65430450e-01 1.12616651e-01
1.51570570e+00 -1.76622853e-01 -9.24008369e-01 -4.42309260e-01
-1.27816245e-01 -5.38684905e-01 -4.44892421e-02 -9.61692929e-01
-1.04678237e+00 1.14368963e+00 7.07804084e-01 -2.57764757e-01
1.09283471e+00 6.50921166e-01 8.19594085e-01 6.27924085e-01
2.56366938e-01 -8.30750525e-01 5.94492853e-01 2.91007072e-01
1.06368291e+00 -1.42007267e+00 2.25769784e-02 -2.02232432e-02
-8.86688054e-01 6.89966679e-01 6.73662484e-01 6.88794032e-02
1.16775744e-01 -4.38524127e-01 4.77219224e-01 -5.36121190e-01
-9.56197917e-01 -3.99813712e-01 5.90098619e-01 5.20019710e-01
1.50923952e-01 -2.55792886e-01 -1.59157440e-01 -2.06611916e-01
3.07133257e-01 -3.10781728e-02 2.17681453e-01 7.69781649e-01
-2.14007720e-01 -1.02309155e+00 -3.61455232e-01 3.83568525e-01
-5.12164056e-01 -6.49367690e-01 -4.00895059e-01 5.93752325e-01
-6.29955292e-01 8.24966371e-01 1.55814484e-01 4.48535495e-02
2.35572308e-01 1.87031120e-01 7.53849566e-01 -5.38933396e-01
-4.79807645e-01 -7.11114183e-02 1.27988577e-01 -8.72486174e-01
-6.75041318e-01 -4.06802356e-01 -8.85291755e-01 2.30736598e-01
-3.15906882e-01 1.41714141e-01 6.31111562e-01 7.09397256e-01
6.41029418e-01 6.68825880e-02 9.85671803e-02 -4.46822762e-01
-4.24927682e-01 -5.97814500e-01 1.29189372e-01 9.07067955e-01
3.71341228e-01 -2.56516159e-01 -2.85237759e-01 4.03788030e-01] | [10.905160903930664, 1.5714211463928223] |
e4178098-5175-477b-bf97-1144d7df48b6 | frame-interpolation-for-dynamic-scenes-with | 2209.13284 | null | https://arxiv.org/abs/2209.13284v2 | https://arxiv.org/pdf/2209.13284v2.pdf | Frame Interpolation for Dynamic Scenes with Implicit Flow Encoding | In this paper, we propose an algorithm to interpolate between a pair of images of a dynamic scene. While in the past years significant progress in frame interpolation has been made, current approaches are not able to handle images with brightness and illumination changes, which are common even when the images are captured shortly apart. We propose to address this problem by taking advantage of the existing optical flow methods that are highly robust to the variations in the illumination. Specifically, using the bidirectional flows estimated using an existing pre-trained flow network, we predict the flows from an intermediate frame to the two input images. To do this, we propose to encode the bidirectional flows into a coordinate-based network, powered by a hypernetwork, to obtain a continuous representation of the flow across time. Once we obtain the estimated flows, we use them within an existing blending network to obtain the final intermediate frame. Through extensive experiments, we demonstrate that our approach is able to produce significantly better results than state-of-the-art frame interpolation algorithms. | ['Nima Khademi Kalantari', 'Avinash Paliwal', 'Pedro Figueirêdo'] | 2022-09-27 | null | null | null | null | ['video-frame-interpolation'] | ['computer-vision'] | [ 1.17528908e-01 -2.26594359e-01 4.22173366e-02 -3.60378355e-01
-1.75494403e-01 -4.86112744e-01 7.44555056e-01 -7.42753521e-02
-3.99736643e-01 8.42650771e-01 1.87367629e-02 -1.15897596e-01
1.23130456e-01 -8.10394168e-01 -8.09453428e-01 -3.41345966e-01
-6.58547506e-02 7.68352151e-02 4.48446423e-01 -5.09920716e-02
3.49732637e-01 7.78734922e-01 -1.62799501e+00 1.15324728e-01
7.38068581e-01 9.38437879e-01 -4.80051003e-02 9.48725700e-01
-2.00955763e-01 1.05589449e+00 -2.58100778e-01 -2.77109265e-01
5.32438874e-01 -7.49825656e-01 -1.02702832e+00 1.73907265e-01
6.10704303e-01 -6.03933215e-01 -4.57031757e-01 7.23672867e-01
9.14752763e-03 4.69479084e-01 2.88575441e-01 -1.17728794e+00
-3.26120436e-01 1.59965053e-01 -4.59862918e-01 2.13329449e-01
5.51738858e-01 3.03606719e-01 5.58804452e-01 -7.72092223e-01
9.78896737e-01 1.29084826e+00 6.90140903e-01 4.35311407e-01
-1.48977172e+00 -3.41152281e-01 1.02874123e-01 2.55297482e-01
-1.05162752e+00 -5.92639148e-01 9.30244923e-01 -4.40599889e-01
5.66573799e-01 2.30234955e-02 9.70967829e-01 6.41826630e-01
-5.02060764e-02 4.67477798e-01 8.68768156e-01 -5.79598546e-01
7.27524236e-02 -5.45801967e-02 -2.83146948e-01 6.46830618e-01
-2.10792005e-01 1.49451166e-01 -5.03074288e-01 6.40506223e-02
1.10586345e+00 -1.49658099e-02 -6.29980028e-01 -5.45741022e-01
-1.33233511e+00 5.23949087e-01 6.84602618e-01 4.65537429e-01
-4.66044158e-01 3.73057961e-01 2.16284752e-01 2.15753287e-01
6.53988123e-01 2.59809434e-01 -1.68345552e-02 -1.59812167e-01
-1.31799209e+00 3.55525166e-01 9.59714949e-01 4.75949287e-01
1.07667732e+00 -5.08196577e-02 -8.18991438e-02 4.98554200e-01
2.54292905e-01 -5.05874790e-02 1.13222688e-01 -1.52513611e+00
4.38470751e-01 2.91460067e-01 4.58142638e-01 -1.22762299e+00
-9.80460718e-02 -6.44402727e-02 -7.06654847e-01 5.32473326e-01
9.43174541e-01 -1.42377675e-01 -7.21717358e-01 1.61202216e+00
5.18965423e-01 7.34436154e-01 -3.75091210e-02 9.16112065e-01
1.43455356e-01 8.16322982e-01 -2.84282118e-02 -1.42369851e-01
8.67336392e-01 -1.01142383e+00 -7.06226945e-01 2.52675638e-02
3.03412229e-01 -9.65938866e-01 6.30313039e-01 2.22886324e-01
-1.64890373e+00 -7.83282757e-01 -1.01984012e+00 -3.28980386e-01
-2.34722883e-01 -2.04602331e-01 3.08789015e-01 3.09073180e-01
-1.34237409e+00 9.24722373e-01 -1.07195091e+00 -3.62356126e-01
2.92108536e-01 3.08653265e-01 -3.93200457e-01 -2.49630697e-02
-1.03324604e+00 9.67177331e-01 3.45406741e-01 3.95054638e-01
-7.59275019e-01 -8.66098702e-01 -8.53710651e-01 -1.01585304e-02
9.48467702e-02 -1.00191844e+00 1.21827984e+00 -1.48947799e+00
-1.80634642e+00 4.71565962e-01 -4.64514077e-01 -4.01230812e-01
9.67678070e-01 -1.73472762e-02 5.19417822e-02 2.80603468e-01
-2.57488221e-01 9.67141330e-01 7.20135629e-01 -1.41436958e+00
-8.04675877e-01 5.73391132e-02 4.87889171e-01 8.41988847e-02
-2.51471847e-01 -7.42547140e-02 -5.33070326e-01 -4.54199523e-01
-1.77211672e-01 -9.63010490e-01 -2.85001546e-01 6.66053355e-01
-2.44981229e-01 1.15640610e-01 1.03166211e+00 -6.29659653e-01
9.20502722e-01 -2.00020266e+00 2.75110751e-01 1.27470404e-01
7.09894523e-02 4.55939084e-01 -5.51945623e-03 2.73212552e-01
-8.12537000e-02 -2.21518397e-01 -4.48476553e-01 -7.89007545e-01
-4.05470759e-01 2.98712045e-01 -2.46390969e-01 5.85283160e-01
1.64862022e-01 6.69560492e-01 -1.19871998e+00 -4.56387252e-01
5.87539136e-01 9.24082100e-01 -6.62076950e-01 4.73312736e-01
-2.60527790e-01 9.88093615e-01 -1.33337051e-01 -5.27390540e-02
7.69920170e-01 -4.45806794e-02 3.73591818e-02 -2.58765936e-01
-4.52815473e-01 1.31759658e-01 -1.21361208e+00 2.05794287e+00
-7.46816039e-01 9.86574471e-01 2.29873061e-02 -1.07170105e+00
7.41426349e-01 4.62828368e-01 6.89984798e-01 -4.05398846e-01
1.21689178e-01 2.64513940e-01 -1.38536051e-01 -4.34623718e-01
5.99286914e-01 -6.15133382e-02 5.77692330e-01 4.06593293e-01
1.24832848e-02 -1.82154134e-01 4.77120161e-01 -7.23249139e-03
7.49840260e-01 5.50367773e-01 -6.16422147e-02 1.36321962e-01
1.04645634e+00 -1.76490262e-01 4.46990818e-01 4.81609672e-01
-1.81962103e-01 8.43104541e-01 5.31588197e-01 -9.32660043e-01
-1.29886830e+00 -8.74610841e-01 9.52354893e-02 4.40125972e-01
1.54267401e-01 -1.97682053e-01 -8.64165664e-01 -4.86246735e-01
-1.26676977e-01 3.07594597e-01 -5.26862741e-01 1.81259841e-01
-9.50394630e-01 -1.55747235e-01 2.18072653e-01 4.37293828e-01
8.21930468e-01 -9.50329423e-01 -8.58594060e-01 4.67473209e-01
-3.11831445e-01 -1.32262862e+00 -4.60563093e-01 -2.94186741e-01
-7.92758524e-01 -9.55494106e-01 -9.78789628e-01 -7.52570033e-01
6.47028506e-01 1.70708895e-01 1.02153122e+00 3.25632781e-01
-7.56101087e-02 2.83061296e-01 -8.49230494e-03 9.12644714e-02
-5.62716067e-01 9.66181140e-03 -3.48677069e-01 4.75851029e-01
-1.84737980e-01 -7.69806206e-01 -8.70250821e-01 1.10000469e-01
-1.21867561e+00 3.06797564e-01 -1.07102424e-01 7.53444016e-01
3.82343352e-01 -2.65669972e-01 2.63498038e-01 -7.12466478e-01
3.46784949e-01 -2.53303885e-01 -8.74259830e-01 9.80603769e-02
-2.88981080e-01 1.29350632e-01 9.01314795e-01 -2.26058215e-01
-1.27616310e+00 4.34372157e-01 -2.15730980e-01 -5.38993716e-01
-2.33450145e-01 2.46215835e-01 2.66512156e-01 -3.99409562e-01
3.73341978e-01 -1.08380608e-01 9.21473727e-02 -1.25873223e-01
6.37085259e-01 4.14712489e-01 7.77279437e-01 -3.51186872e-01
8.99319232e-01 8.85304809e-01 2.22960219e-01 -6.39057279e-01
-7.14006066e-01 -3.24456125e-01 -9.37417984e-01 -5.05755007e-01
1.01631629e+00 -6.10035419e-01 -7.41241455e-01 5.60785115e-01
-1.58072364e+00 -4.65060800e-01 -3.36539626e-01 5.64050794e-01
-8.69594395e-01 4.78742719e-01 -5.53502977e-01 -6.30157530e-01
-2.65318956e-02 -1.37535012e+00 7.01193631e-01 3.74427229e-01
-1.51552493e-02 -1.44149482e+00 2.62666434e-01 -6.19833767e-02
6.03918731e-01 5.45368075e-01 5.47340631e-01 1.55272424e-01
-8.74497890e-01 6.64912164e-02 -2.02988520e-01 3.46279800e-01
3.87745917e-01 4.12448406e-01 -8.67159665e-01 3.09347045e-02
-4.16890308e-02 4.73386981e-02 7.23598897e-01 4.50471908e-01
1.19923770e+00 -1.25314221e-01 -1.92447975e-01 8.25810432e-01
1.45994723e+00 1.23716548e-01 8.23106587e-01 3.87980461e-01
7.73573577e-01 6.88499808e-01 1.97554708e-01 2.70997912e-01
5.19781470e-01 8.52393150e-01 5.68012655e-01 -2.93142796e-01
-3.63925368e-01 -1.99515328e-01 1.02908224e-01 4.68753874e-01
-4.28181976e-01 -1.36792377e-01 -6.88620389e-01 7.00404465e-01
-1.92559814e+00 -1.07590878e+00 -1.22058220e-01 2.24870563e+00
7.86855996e-01 -8.02661777e-02 1.06857717e-02 4.40992787e-02
6.35492325e-01 3.47606391e-01 -3.43337715e-01 -5.05518198e-01
1.76452011e-01 1.51601747e-01 4.08694595e-01 8.59472215e-01
-9.73850250e-01 7.82157779e-01 6.46232271e+00 8.05295259e-02
-1.42815506e+00 -6.32295460e-02 7.02268720e-01 1.25123002e-02
-3.55443001e-01 3.86249363e-01 -2.55167842e-01 4.48722363e-01
8.85474503e-01 -8.42766091e-02 6.87854290e-01 3.01472187e-01
5.62931418e-01 -2.46131450e-01 -1.22359931e+00 7.77869225e-01
6.01754673e-02 -1.62723792e+00 -2.00925171e-02 -2.03653291e-01
8.75028312e-01 -2.79864758e-01 -9.05491933e-02 -3.27228010e-01
2.17227340e-01 -7.93117642e-01 7.03574300e-01 7.81969488e-01
4.13527787e-01 -6.06504500e-01 3.86898577e-01 2.61672854e-01
-1.17696953e+00 1.97425932e-01 -1.61115170e-01 -1.36279985e-01
6.74839675e-01 5.71263433e-01 -4.74825382e-01 7.21283853e-01
4.96036887e-01 9.95856524e-01 -2.95098275e-01 1.32856834e+00
-2.94396996e-01 2.79883534e-01 -2.46418908e-01 5.27232170e-01
2.99503118e-01 -3.53666991e-01 4.01080221e-01 9.67382133e-01
4.03938025e-01 -2.72751927e-01 1.03521541e-01 1.00585139e+00
-2.39133954e-01 -1.65952966e-01 -6.40312254e-01 4.24445003e-01
1.44159645e-02 1.20795858e+00 -7.07280159e-01 -5.21358907e-01
-6.37575686e-01 1.13253534e+00 3.82576704e-01 5.95649898e-01
-7.76624799e-01 -3.53041351e-01 7.39385962e-01 7.62639046e-02
2.29797840e-01 -4.93995935e-01 1.88041304e-03 -1.35492003e+00
9.68394428e-02 -4.56703812e-01 -3.36541655e-03 -9.52817798e-01
-9.71369505e-01 6.49706900e-01 -8.24542157e-03 -1.19284904e+00
-5.36498606e-01 -2.08998859e-01 -7.37548411e-01 1.12734628e+00
-2.05644798e+00 -9.48841214e-01 -5.63793898e-01 6.44517541e-01
5.58924556e-01 4.55914021e-01 5.42481601e-01 3.96851808e-01
-2.06925705e-01 7.50983283e-02 -8.68444592e-02 1.77644610e-01
8.32581460e-01 -1.10261130e+00 5.04473805e-01 1.15946341e+00
6.63590431e-02 4.41551685e-01 7.21832991e-01 -3.28847826e-01
-1.09592044e+00 -9.39001322e-01 9.81621563e-01 -3.51123847e-02
6.46698296e-01 -1.22853681e-01 -1.01137710e+00 9.05886471e-01
6.22423887e-01 7.37629890e-01 6.15581870e-02 -5.34019530e-01
-1.22922882e-01 -3.36554110e-01 -9.19503689e-01 5.62651634e-01
7.66350448e-01 -4.55969840e-01 -2.31155947e-01 1.06921405e-01
4.62816119e-01 -6.43292904e-01 -8.39704037e-01 4.52947952e-02
6.90882683e-01 -1.30939996e+00 1.03792012e+00 -3.17565769e-01
6.99445188e-01 -5.75599670e-01 2.94708967e-01 -1.52992630e+00
-2.90000602e-03 -8.66408110e-01 -4.57764007e-02 1.13115299e+00
1.05493389e-01 -7.34115362e-01 7.66310215e-01 8.12731802e-01
6.65007830e-02 -4.36816931e-01 -9.37196314e-01 -5.67635119e-01
3.81100997e-02 -1.93378583e-01 6.40170813e-01 8.22262108e-01
-1.48084253e-01 -1.60992697e-01 -5.92181444e-01 3.80373336e-02
5.79245210e-01 9.44337398e-02 8.96275580e-01 -1.07920086e+00
-1.31289586e-01 -3.94798100e-01 -5.17366886e-01 -1.24288750e+00
4.42181587e-01 -5.66375136e-01 2.08984077e-01 -1.64352763e+00
-2.45209798e-01 -5.30884147e-01 -1.39841754e-02 2.47423947e-01
-1.98953211e-01 4.89976615e-01 5.20540476e-01 1.75982773e-01
-1.50832742e-01 3.78673494e-01 1.45608175e+00 8.77730176e-02
-2.94090480e-01 -1.46919280e-01 -5.86552136e-02 7.74464011e-01
6.19695961e-01 -2.54348725e-01 -3.49953920e-01 -7.24235296e-01
-1.10048555e-01 3.66270274e-01 4.73694324e-01 -1.13526189e+00
4.47312653e-01 -8.66794214e-02 4.81239587e-01 -2.37853020e-01
3.86592895e-01 -1.03676867e+00 4.45475966e-01 3.91018450e-01
-4.72491354e-01 1.56505182e-01 4.08368520e-02 3.22534353e-01
-4.17158812e-01 -3.00013542e-01 9.26397979e-01 -2.15634003e-01
-4.32941467e-01 3.58466446e-01 -2.26908743e-01 -3.03911656e-01
1.04891014e+00 -2.09697038e-01 -2.55186617e-01 -5.34796238e-01
-7.09919453e-01 7.05428375e-03 7.81598449e-01 3.10393035e-01
4.07635570e-01 -1.23996711e+00 -3.47938120e-01 3.03102791e-01
-3.04246604e-01 1.13533296e-01 8.63746479e-02 9.15053904e-01
-9.91161346e-01 1.88224345e-01 -3.68443996e-01 -5.94070435e-01
-8.97139370e-01 4.30471331e-01 6.52409554e-01 -2.49699399e-01
-5.78201413e-01 2.93079883e-01 1.89663726e-03 -4.74965088e-02
8.23101327e-02 -4.73108590e-01 -1.39406517e-01 -1.44499332e-01
4.41053599e-01 4.11747485e-01 -7.76047558e-02 -8.37852538e-01
-1.96530707e-02 7.66544521e-01 2.99854100e-01 -3.06891680e-01
1.27764976e+00 -2.94481575e-01 -1.79541573e-01 4.18472737e-01
1.39881754e+00 -1.38265669e-01 -1.90897131e+00 5.55532724e-02
-2.38666967e-01 -8.80977869e-01 -1.76521819e-02 -2.51244187e-01
-1.49888420e+00 9.27417576e-01 3.37636888e-01 4.29413617e-01
1.16929138e+00 -4.87050414e-01 1.02492976e+00 -1.85841084e-01
1.25102714e-01 -6.30673468e-01 -1.12506598e-01 4.91823435e-01
4.60405946e-01 -1.02105606e+00 -1.64721400e-01 -4.40311968e-01
-4.54353802e-02 1.55364978e+00 2.42065370e-01 -3.74324888e-01
5.91830611e-01 6.88893422e-02 2.02180028e-01 3.03248703e-01
-5.71240664e-01 -1.22696422e-01 1.07667372e-01 5.80541313e-01
5.21071732e-01 -4.91474867e-01 -3.35539013e-01 -8.31168890e-01
1.17499553e-01 4.92783904e-01 9.19733942e-01 8.42719257e-01
-1.13510698e-01 -1.47921610e+00 -3.21589023e-01 -1.46979555e-01
-4.65091377e-01 2.45922968e-01 8.48504752e-02 6.39317214e-01
6.21671453e-02 8.27076316e-01 3.16012710e-01 2.13967219e-01
3.71508896e-01 2.16668136e-02 6.85468376e-01 -1.53340787e-01
-6.10042989e-01 -1.56514540e-01 -1.81418449e-01 -8.29865813e-01
-1.15187514e+00 -6.33845329e-01 -1.12171626e+00 -3.06443036e-01
3.27582024e-02 3.18180658e-02 7.65755236e-01 9.79858816e-01
1.85462207e-01 4.49850231e-01 7.14273155e-01 -1.44410968e+00
5.28677963e-02 -3.75645101e-01 -1.85093164e-01 6.48806155e-01
7.43312180e-01 -5.57123244e-01 -4.28537220e-01 6.11153543e-01] | [10.664812088012695, -1.3347513675689697] |
47448e2d-97d5-4833-a882-892df7041ee6 | weighted-risk-minimization-deep-learning | 1812.03372 | null | https://arxiv.org/abs/1812.03372v3 | https://arxiv.org/pdf/1812.03372v3.pdf | What is the Effect of Importance Weighting in Deep Learning? | Importance-weighted risk minimization is a key ingredient in many machine learning algorithms for causal inference, domain adaptation, class imbalance, and off-policy reinforcement learning. While the effect of importance weighting is well-characterized for low-capacity misspecified models, little is known about how it impacts over-parameterized, deep neural networks. This work is inspired by recent theoretical results showing that on (linearly) separable data, deep linear networks optimized by SGD learn weight-agnostic solutions, prompting us to ask, for realistic deep networks, for which many practical datasets are separable, what is the effect of importance weighting? We present the surprising finding that while importance weighting impacts models early in training, its effect diminishes over successive epochs. Moreover, while L2 regularization and batch normalization (but not dropout), restore some of the impact of importance weighting, they express the effect via (seemingly) the wrong abstraction: why should practitioners tweak the L2 regularization, and by how much, to produce the correct weighting effect? Our experiments confirm these findings across a range of architectures and datasets. | ['Zachary C. Lipton', 'Jonathon Byrd'] | 2018-12-08 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [ 4.52753782e-01 3.81813288e-01 -5.41517496e-01 -5.52419841e-01
-5.08257747e-01 -4.28164274e-01 5.63807666e-01 2.88849890e-01
-6.85296893e-01 9.11795139e-01 6.48388743e-01 -5.42798042e-01
-6.12859845e-01 -5.70510149e-01 -9.22660649e-01 -8.41362596e-01
-1.59009427e-01 4.52752978e-01 1.46831572e-01 6.39634877e-02
2.03317165e-01 5.36819816e-01 -1.18614137e+00 1.85199846e-02
7.21216559e-01 8.00693572e-01 -3.96139205e-01 4.15397912e-01
1.11941911e-01 1.02307630e+00 -4.01512504e-01 -6.14995956e-01
1.32269651e-01 -3.95858973e-01 -7.21766949e-01 -2.50730991e-01
6.22169614e-01 -3.50696236e-01 -4.70571965e-01 1.00358653e+00
4.47269261e-01 2.24135146e-01 9.92765844e-01 -1.11774492e+00
-8.39387238e-01 1.04280949e+00 -4.76373583e-01 5.79897463e-01
-4.90314633e-01 3.89557034e-01 1.37640297e+00 -4.42012697e-01
4.36507374e-01 1.27830362e+00 7.94386864e-01 5.92620015e-01
-1.58177769e+00 -7.24925160e-01 4.89530265e-01 -6.93855435e-02
-9.02365088e-01 -5.28759778e-01 6.79966450e-01 -5.13371825e-01
6.16408467e-01 5.87298609e-02 3.79693180e-01 1.42781556e+00
4.80128974e-02 6.88654482e-01 8.76879632e-01 -1.60382032e-01
3.47357601e-01 7.52218962e-02 3.40343118e-01 4.42059726e-01
7.25137830e-01 2.35698625e-01 -5.68177044e-01 -5.24294317e-01
6.91133261e-01 -5.78609481e-02 -2.39053264e-01 -3.83125842e-01
-9.28232908e-01 9.15554523e-01 7.98767284e-02 1.23004109e-01
-3.35371524e-01 7.29564607e-01 4.20357198e-01 3.77049834e-01
5.81185162e-01 5.70179880e-01 -7.98202872e-01 -1.88638210e-01
-6.27253890e-01 5.24073064e-01 6.62258983e-01 5.35733640e-01
2.12893024e-01 3.16382557e-01 -3.72261852e-01 7.45488286e-01
1.25805333e-01 2.63758570e-01 3.67608815e-01 -9.58311617e-01
4.70893800e-01 5.02064943e-01 8.22569057e-02 -6.87774658e-01
-7.27394164e-01 -6.07936203e-01 -8.66611898e-01 2.78181463e-01
1.10337400e+00 -6.34592950e-01 -6.28044248e-01 2.40550423e+00
5.00827245e-02 1.14494018e-01 -2.53158927e-01 1.03075039e+00
2.22959086e-01 1.19773723e-01 6.52806401e-01 -1.50542781e-02
1.33650386e+00 -3.26640785e-01 -3.10220063e-01 -5.85542381e-01
5.33076763e-01 -4.40282881e-01 1.26053929e+00 3.68016273e-01
-1.02259195e+00 -1.29076481e-01 -1.00647271e+00 -1.31642252e-01
-1.12094060e-01 -4.01917547e-01 1.05150604e+00 6.62474096e-01
-7.15202093e-01 1.02512610e+00 -5.34049690e-01 -1.08610705e-01
8.31919789e-01 4.25237298e-01 -2.04376400e-01 -4.24873680e-02
-1.47866130e+00 1.01807976e+00 2.36655429e-01 -1.17987148e-01
-8.46998751e-01 -1.26170707e+00 -3.42342466e-01 4.32864368e-01
4.37344700e-01 -6.98984087e-01 1.05494606e+00 -1.43754435e+00
-1.20275044e+00 5.68860233e-01 1.92113549e-01 -6.88564897e-01
5.68668664e-01 -1.74034953e-01 -1.14018247e-01 -1.94402933e-01
-4.44240212e-01 4.17407393e-01 1.05702996e+00 -8.89036298e-01
-5.66156685e-01 -5.71386397e-01 2.46421307e-01 1.31936952e-01
-4.24266458e-01 -9.83683094e-02 1.79772556e-01 -7.20733702e-01
-2.98550010e-01 -7.20724165e-01 -3.59434932e-01 5.92560470e-02
-1.30198210e-01 -3.82948607e-01 2.55398184e-01 -4.38169658e-01
8.80729795e-01 -2.22591686e+00 -4.47715595e-02 2.05689177e-01
5.55706263e-01 1.74131542e-02 -3.34949791e-01 5.15346527e-02
-4.25790936e-01 2.13380471e-01 -4.28050041e-01 -3.38217840e-02
3.61898959e-01 3.19118053e-02 -4.71608907e-01 5.95602095e-01
5.36886394e-01 8.00182700e-01 -8.65318000e-01 -1.89792544e-01
-1.72421023e-01 5.53447545e-01 -9.19442236e-01 -1.05049238e-01
-3.69482934e-01 4.01879326e-02 -4.53413069e-01 9.66664925e-02
3.98989677e-01 -4.14096564e-01 3.64976078e-01 -6.51817694e-02
2.20660657e-01 4.04235989e-01 -1.25292325e+00 1.09582436e+00
-5.09693623e-02 7.32940733e-01 5.02517298e-02 -1.41632354e+00
4.77595836e-01 3.16866487e-01 3.94442171e-01 -6.33688688e-01
1.09804787e-01 1.54322416e-01 4.24972147e-01 -4.03385431e-01
2.69070357e-01 -5.72305143e-01 1.08721986e-01 6.47112846e-01
1.94351338e-02 3.23839098e-01 -3.77005786e-02 1.22962445e-01
1.34648848e+00 -3.14989040e-04 2.42259186e-02 -4.83274996e-01
-1.76055282e-01 -2.17261724e-02 6.81806326e-01 1.03621268e+00
-2.69339561e-01 7.12099075e-01 1.22465956e+00 -3.74225229e-01
-1.12616730e+00 -8.91314924e-01 -3.15354168e-01 1.48157787e+00
-2.91227013e-01 1.80023283e-01 -5.42614102e-01 -6.97154284e-01
4.24093723e-01 9.46968555e-01 -8.42479706e-01 -6.82434797e-01
-5.44097543e-01 -1.46248889e+00 6.33500099e-01 6.47692680e-01
-5.03313057e-02 -9.34750438e-01 -5.76482058e-01 2.87915558e-01
2.80044824e-01 -6.14736915e-01 -4.43495452e-01 5.78199387e-01
-1.02332914e+00 -1.13014817e+00 -6.83609664e-01 -7.29325190e-02
5.24280012e-01 -1.14850327e-01 1.48556781e+00 1.77166581e-01
-2.64402833e-02 2.89972812e-01 1.64753675e-01 -6.43654346e-01
-2.08894536e-01 2.10316733e-01 5.95443919e-02 -1.14955708e-01
6.25113070e-01 -9.16229665e-01 -6.35272324e-01 -9.51025113e-02
-9.13409650e-01 -4.39457089e-01 7.16287673e-01 9.55588400e-01
2.64208857e-02 -6.64016232e-02 8.86070609e-01 -1.35744107e+00
9.00488377e-01 -5.74946463e-01 -7.73714960e-01 4.80014160e-02
-8.22292268e-01 6.63255811e-01 5.45410514e-01 -8.54926586e-01
-9.82230604e-01 -3.91477048e-01 -1.27567202e-01 -1.46464512e-01
-4.95601259e-02 3.98056388e-01 -9.06643644e-02 2.48106226e-01
8.85069251e-01 -3.28854501e-01 -1.96718603e-01 -5.73771775e-01
3.76972944e-01 8.81939307e-02 1.15890905e-01 -8.55346143e-01
6.80810809e-01 3.27381253e-01 6.98856264e-02 -4.63262558e-01
-1.23455536e+00 1.88305438e-01 -2.76697934e-01 2.29735941e-01
7.85129189e-01 -8.48810077e-01 -9.22801256e-01 1.05519786e-01
-9.46001887e-01 -7.18804002e-01 -6.87256396e-01 5.81328571e-01
-2.14016050e-01 -1.62315294e-01 -3.18633199e-01 -7.99405992e-01
-4.24484648e-02 -1.01291621e+00 4.89665240e-01 7.01528490e-02
-4.66742814e-01 -1.11374903e+00 -1.57461613e-01 1.19872801e-01
5.10809898e-01 2.73498464e-02 1.67622817e+00 -1.01773322e+00
-2.79026985e-01 -8.49267989e-02 -4.06542212e-01 3.94817531e-01
-1.61466941e-01 -1.96204744e-02 -1.08546627e+00 1.06118836e-01
-1.05138212e-01 -3.79012376e-01 1.26293755e+00 5.58120608e-01
1.33545840e+00 -3.99515957e-01 5.54795712e-02 4.07287955e-01
1.14719296e+00 -6.86551630e-02 3.21041018e-01 1.09677218e-01
7.51489103e-01 6.20744050e-01 -5.14324121e-02 3.15379322e-01
3.75264913e-01 4.19610143e-01 5.46898425e-01 -1.85900092e-01
-4.65853624e-02 -1.42885029e-01 2.85189897e-01 6.56929240e-02
5.06784488e-03 3.46089229e-02 -7.88322747e-01 5.03874540e-01
-1.65452588e+00 -9.93215621e-01 -1.32011920e-01 2.40288997e+00
1.37080693e+00 6.21196985e-01 2.13262185e-01 8.97196960e-03
3.02477986e-01 4.34015960e-01 -9.07163322e-01 -5.52208781e-01
-2.04260811e-01 2.19692454e-01 8.93258214e-01 4.60353911e-01
-8.91929567e-01 6.22158945e-01 6.48906851e+00 9.33376908e-01
-1.28518653e+00 2.97501653e-01 9.82221484e-01 -5.00143111e-01
-6.28460467e-01 4.50918786e-02 -6.34190023e-01 3.44784319e-01
1.05237782e+00 -1.34202257e-01 3.99135828e-01 8.30902815e-01
1.94344401e-01 1.37147397e-01 -1.34384429e+00 3.60806316e-01
-3.00672144e-01 -1.39001167e+00 -2.59664148e-01 5.82841448e-02
6.42664969e-01 1.21168092e-01 1.59341201e-01 3.63005757e-01
8.98396432e-01 -1.20916915e+00 8.12556446e-01 2.39340782e-01
4.06028777e-01 -8.74354720e-01 5.34082234e-01 1.56337619e-01
-1.43065572e-01 -2.95216322e-01 -6.35242820e-01 -1.79912463e-01
-1.50812104e-01 1.14681113e+00 -5.71459353e-01 -2.03373656e-01
2.93265551e-01 2.00141460e-01 -4.36291128e-01 6.57836914e-01
-3.87425333e-01 1.15776920e+00 -2.37129122e-01 5.08879162e-02
2.52351493e-01 5.84130324e-02 3.11085880e-01 1.10035765e+00
-8.60268250e-02 5.25813848e-02 -5.16887605e-01 1.00442064e+00
-4.44314063e-01 -2.53042549e-01 -4.36882406e-01 -2.81956971e-01
4.58855331e-01 8.97213161e-01 -4.21930104e-01 -2.68713355e-01
-3.27379197e-01 3.58351141e-01 5.98240674e-01 5.79622090e-01
-7.50507236e-01 -1.21233180e-01 9.78167474e-01 3.36525738e-01
9.02045369e-02 -6.30505383e-02 -7.14451909e-01 -8.79156113e-01
-1.42924532e-01 -8.53226423e-01 4.73493040e-01 -3.03532928e-01
-1.36777472e+00 -2.39422303e-02 -9.38317105e-02 -4.74453926e-01
-4.13388386e-02 -5.86806357e-01 -5.36413431e-01 8.48570228e-01
-1.65942574e+00 -5.59009731e-01 3.93172532e-01 2.98998922e-01
2.10925549e-01 1.00725606e-01 5.76503873e-01 4.96103287e-01
-8.16841483e-01 5.98969877e-01 2.68231899e-01 5.06352447e-02
8.78482342e-01 -1.38248992e+00 1.92173287e-01 5.19064069e-01
3.95135023e-02 7.54918635e-01 9.15072441e-01 -3.60256225e-01
-1.32677913e+00 -1.06789780e+00 9.11450624e-01 -7.45841444e-01
8.06002378e-01 -4.04718399e-01 -1.08353925e+00 6.16176724e-01
-3.38052250e-02 1.67392492e-01 5.66955209e-01 6.72276020e-01
-7.56819904e-01 -2.10535482e-01 -1.13211536e+00 7.23344624e-01
1.00256395e+00 -2.51544863e-01 -6.08002365e-01 3.06279778e-01
8.94575119e-01 3.05703040e-02 -7.70175397e-01 2.59105980e-01
6.84691429e-01 -8.67097497e-01 9.33096051e-01 -1.19113839e+00
8.54584813e-01 2.82116443e-01 -9.82018281e-03 -1.45197856e+00
-4.23361450e-01 -3.47631156e-01 -1.73636481e-01 1.12803638e+00
7.16984689e-01 -7.87957013e-01 8.19160759e-01 1.12647450e+00
2.26225942e-01 -6.23204410e-01 -9.58956003e-01 -5.21647334e-01
6.28379524e-01 -5.74788153e-01 6.02849841e-01 1.20352519e+00
-2.29893997e-01 5.03414989e-01 -2.95504719e-01 1.22700855e-01
9.27407622e-01 -1.92680329e-01 3.52429450e-01 -1.48973596e+00
-3.50890785e-01 -7.87137389e-01 1.03621952e-01 -6.90971732e-01
2.37781420e-01 -6.89494908e-01 -2.70330817e-01 -9.46312308e-01
1.50049478e-01 -6.89757466e-01 -6.01482928e-01 5.81095517e-01
-5.43658555e-01 -2.05826890e-02 1.14886910e-01 -4.21007350e-02
-3.83207768e-01 3.92212749e-01 1.00091136e+00 -3.66232358e-02
1.68946274e-02 5.54112382e-02 -1.26215076e+00 7.92295992e-01
7.36734867e-01 -8.76354992e-01 -5.41473567e-01 -6.05209768e-01
7.29365528e-01 -1.94694072e-01 5.34043193e-01 -5.30827463e-01
-1.62422117e-02 -6.62998497e-01 4.65054959e-01 1.25261649e-01
7.67045394e-02 -8.54013443e-01 -2.73127228e-01 4.21575159e-01
-9.39672470e-01 -1.77502498e-01 2.51612753e-01 4.62293237e-01
2.63768762e-01 -4.01296407e-01 9.28949714e-01 -7.79347271e-02
-2.77860224e-01 2.69785523e-01 -3.61444801e-01 7.33574271e-01
2.92164356e-01 1.73332453e-01 -3.79713356e-01 -2.60688007e-01
-2.65799582e-01 -6.14861585e-02 6.89395666e-02 4.40221876e-01
1.75170094e-01 -1.25319886e+00 -9.33757901e-01 1.43183693e-01
-1.29924625e-01 2.89905909e-02 2.56212115e-01 9.47048366e-01
1.89931035e-01 2.15782404e-01 1.68595538e-01 -7.76780024e-02
-6.90782368e-01 2.71619141e-01 3.10636908e-01 -2.58356839e-01
-4.77271706e-01 1.06512642e+00 5.54454446e-01 -1.96318150e-01
4.90472764e-01 -3.75541091e-01 2.04625353e-02 1.87613115e-01
6.03631973e-01 2.13238999e-01 8.52588415e-02 2.13796690e-01
-1.73004597e-01 -1.41412513e-02 -1.83685586e-01 1.00782767e-01
1.59051096e+00 2.45195642e-01 2.63720781e-01 4.93592978e-01
8.99492860e-01 -1.34781674e-01 -1.69782543e+00 -2.81816065e-01
2.87870377e-01 -8.22391212e-02 2.39066258e-01 -8.23861897e-01
-1.19291997e+00 1.13989115e+00 5.18445551e-01 1.72281384e-01
7.75773168e-01 -1.87790319e-01 2.77473181e-01 3.15737098e-01
-1.22720726e-01 -1.12527561e+00 -5.96183874e-02 5.47378957e-01
8.10182869e-01 -1.19996762e+00 1.91019639e-01 2.96721339e-01
-5.83563149e-01 7.25082040e-01 3.55754197e-01 -2.67496705e-01
6.87745214e-01 3.44966352e-01 -3.42666745e-01 -2.04581872e-01
-1.17272615e+00 8.96122605e-02 7.38428012e-02 5.74219584e-01
5.41558802e-01 -3.93782407e-02 -1.52776182e-01 8.78053486e-01
4.59434651e-02 8.63886625e-02 4.07155603e-01 4.86063540e-01
-3.99593830e-01 -8.62523735e-01 -2.69036800e-01 8.31355214e-01
-8.40924144e-01 -4.84534562e-01 -2.57753193e-01 9.09145117e-01
4.41075489e-02 4.01905298e-01 3.23330402e-01 6.51939139e-02
4.12092060e-01 2.68730879e-01 3.68836641e-01 -4.94936228e-01
-6.95081115e-01 -4.67251986e-01 1.39293298e-01 -3.19098860e-01
-6.56480193e-02 -6.37978315e-01 -1.09608245e+00 -5.16665578e-01
-4.28505950e-02 -6.07834831e-02 5.73775053e-01 1.13207197e+00
2.89435983e-01 6.89842165e-01 2.97773242e-01 -4.83762443e-01
-1.13246143e+00 -9.41081822e-01 -6.86808050e-01 5.07970095e-01
6.73173726e-01 -6.95407808e-01 -8.69915605e-01 -1.90924034e-01] | [8.496439933776855, 4.077672958374023] |
37e1d529-6863-46b9-bc2d-c001525f105b | motion-segmentation-using-frequency-domain | 2004.08638 | null | https://arxiv.org/abs/2004.08638v1 | https://arxiv.org/pdf/2004.08638v1.pdf | Motion Segmentation using Frequency Domain Transformer Networks | Self-supervised prediction is a powerful mechanism to learn representations that capture the underlying structure of the data. Despite recent progress, the self-supervised video prediction task is still challenging. One of the critical factors that make the task hard is motion segmentation, which is segmenting individual objects and the background and estimating their motion separately. In video prediction, the shape, appearance, and transformation of each object should be understood only by predicting the next frame in pixel space. To address this task, we propose a novel end-to-end learnable architecture that predicts the next frame by modeling foreground and background separately while simultaneously estimating and predicting the foreground motion using Frequency Domain Transformer Networks. Experimental evaluations show that this yields interpretable representations and that our approach can outperform some widely used video prediction methods like Video Ladder Network and Predictive Gated Pyramids on synthetic data. | ['Hafez Farazi', 'Sven Behnke'] | 2020-04-18 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 5.45475006e-01 1.59415916e-01 -5.27653098e-01 -5.30716240e-01
-3.10041994e-01 -2.27293164e-01 4.25677150e-01 -3.00882101e-01
1.47148535e-01 5.03910244e-01 3.82266343e-01 8.67854804e-02
3.58429849e-01 -5.24450123e-01 -9.94570673e-01 -6.33324325e-01
-4.36641015e-02 3.89131069e-01 7.90080130e-01 1.34868219e-01
7.07729533e-02 1.88007265e-01 -1.40224910e+00 7.56759584e-01
7.30478764e-01 1.15860903e+00 6.08895302e-01 9.14533973e-01
-6.00615069e-02 1.67050338e+00 5.90210259e-02 -7.72882923e-02
1.82462156e-01 -7.98825622e-01 -8.77579272e-01 6.15877926e-01
6.75575554e-01 -5.74599385e-01 -6.68222845e-01 8.91954064e-01
-1.59099445e-01 1.52835593e-01 6.41258836e-01 -1.10400736e+00
-3.74224067e-01 5.41023552e-01 -5.98467648e-01 3.81931633e-01
5.97032756e-02 1.81795537e-01 9.39175487e-01 -6.86868429e-01
6.53770149e-01 1.06557739e+00 5.89247465e-01 7.73621261e-01
-1.40865850e+00 -3.29699099e-01 4.60730284e-01 5.73859394e-01
-1.01892185e+00 -5.55056393e-01 9.76505637e-01 -7.11117446e-01
6.98431790e-01 3.33303884e-02 7.52084911e-01 1.12729597e+00
2.19226360e-01 1.14665902e+00 5.48605084e-01 -2.91394323e-01
1.27849400e-01 -1.44248381e-01 -2.99455207e-02 1.00949419e+00
-2.32667804e-01 -8.81232396e-02 -6.37522399e-01 1.14058793e-01
1.02382433e+00 2.76432931e-01 -4.74743992e-01 -7.19840825e-01
-1.23271120e+00 4.96551692e-01 2.17000857e-01 1.36705354e-01
-4.94359434e-01 6.11515224e-01 1.62182808e-01 -7.67300278e-02
5.97793043e-01 -1.14448152e-01 -6.14208758e-01 -1.39815480e-01
-1.42055273e+00 9.10272226e-02 5.97190201e-01 7.94164062e-01
9.21141326e-01 5.08562684e-01 -7.77436420e-02 4.38967913e-01
3.24807018e-01 1.26733094e-01 5.39993942e-01 -1.39272571e+00
2.43907407e-01 5.27813196e-01 3.00412238e-01 -1.05493689e+00
-2.34097749e-01 -1.01349212e-01 -9.41782117e-01 1.10596260e-02
5.25360465e-01 -1.63089827e-01 -1.10107923e+00 1.59349108e+00
5.54379895e-02 8.59821022e-01 -1.01437204e-01 9.04062569e-01
6.43134952e-01 1.27833402e+00 3.12721670e-01 -4.37379688e-01
7.87598491e-01 -1.41445959e+00 -4.67034340e-01 -5.35553873e-01
2.27668405e-01 -5.08033216e-01 5.26963651e-01 2.13797703e-01
-1.22815335e+00 -1.10469484e+00 -7.38536596e-01 6.20418303e-02
4.56292957e-01 3.09969991e-01 5.61066389e-01 2.67413378e-01
-1.05238342e+00 9.92392838e-01 -1.29011333e+00 -2.30037957e-01
5.44209659e-01 4.46576178e-01 -1.40716806e-01 -3.62590477e-02
-6.77456975e-01 4.93737131e-01 3.90757531e-01 6.51562866e-03
-1.07221115e+00 -6.77798092e-01 -7.85112381e-01 3.36735457e-01
2.26665244e-01 -7.74000168e-01 1.28619802e+00 -1.93164492e+00
-1.44474697e+00 6.18670464e-01 -6.37201667e-01 -9.35719490e-01
4.86893117e-01 -3.43059480e-01 -2.01259926e-01 3.75386447e-01
-2.91198865e-02 1.13598061e+00 1.32399261e+00 -1.14477086e+00
-9.69108999e-01 -1.49123475e-01 -4.31907475e-01 1.30004615e-01
-1.90752611e-01 -1.52613536e-01 -6.46714091e-01 -7.50706315e-01
1.19094416e-01 -9.37233567e-01 -5.71368039e-01 1.21398889e-01
-2.05641553e-01 -2.96520572e-02 1.13715827e+00 -9.48782742e-01
9.13878858e-01 -2.13278365e+00 3.61255199e-01 -2.13376701e-01
3.65558833e-01 1.99544668e-01 -4.04955409e-02 -8.04445669e-02
-1.41548738e-01 -2.10216090e-01 -1.44575328e-01 -3.60353112e-01
-4.67382848e-01 2.94546306e-01 -5.65093279e-01 2.82136917e-01
4.05943066e-01 9.23349440e-01 -8.46662164e-01 -5.54563642e-01
2.72498727e-01 5.82272291e-01 -7.02196419e-01 4.11769032e-01
-7.01812983e-01 7.71773994e-01 -3.47932130e-01 3.85040462e-01
2.32552052e-01 -7.13702977e-01 3.94789666e-01 -4.37205106e-01
6.22350313e-02 6.23043552e-02 -9.51685607e-01 1.41486061e+00
1.79549400e-02 1.18516862e+00 -3.06707621e-01 -1.34950352e+00
7.69832551e-01 3.26194614e-01 8.64738166e-01 -3.74361485e-01
-1.40148699e-01 -2.25407884e-01 -1.49975702e-01 -6.67239010e-01
4.28572357e-01 -5.69301378e-03 4.31860477e-01 3.14994395e-01
1.64988831e-01 3.98269475e-01 -1.88654810e-02 8.12480599e-02
8.41386318e-01 6.84374273e-01 2.21277922e-02 -7.10966811e-02
4.13287371e-01 2.38771737e-01 1.07986653e+00 4.06499058e-01
-1.90959930e-01 9.01564002e-01 3.91823769e-01 -8.83174717e-01
-1.12335432e+00 -8.61576617e-01 5.64397156e-01 1.24276185e+00
3.05106401e-01 -3.16078424e-01 -7.50251114e-01 -6.02985799e-01
-3.45421165e-01 5.54322839e-01 -4.97851938e-01 -2.56888308e-02
-9.47163522e-01 -4.94507700e-01 -2.44269013e-01 8.79197955e-01
4.21358943e-01 -8.85506153e-01 -8.36607099e-01 4.46235716e-01
-6.33385122e-01 -1.56422091e+00 -5.24998426e-01 1.09407119e-01
-1.18495989e+00 -1.05005872e+00 -6.38233960e-01 -1.14151394e+00
6.51361227e-01 4.22677517e-01 1.20351398e+00 4.53873873e-02
-1.32250935e-01 3.76595080e-01 -1.36342779e-01 1.53422877e-01
-5.45589626e-01 -2.36053869e-01 -1.31026998e-01 5.69142282e-01
-1.93676539e-02 -5.49985647e-01 -8.78524065e-01 3.12983423e-01
-6.20367825e-01 6.72874868e-01 3.50390047e-01 6.46809697e-01
8.81951153e-01 2.41391644e-01 3.51515338e-02 -9.13870156e-01
-3.49258095e-01 -3.89180511e-01 -5.61499953e-01 3.16191137e-01
-1.27786174e-01 -1.21011399e-03 8.40300083e-01 -5.77455461e-01
-1.27444589e+00 8.77427876e-01 2.36534163e-01 -7.12390602e-01
-1.77111655e-01 -1.34707680e-02 -3.43773216e-02 3.03500649e-02
3.54903013e-01 6.92208827e-01 -2.80396193e-02 -3.59532773e-01
1.60828710e-01 6.36132509e-02 7.39329696e-01 -8.91612321e-02
6.22135520e-01 7.38029361e-01 -3.60190384e-02 -1.01614130e+00
-9.65767860e-01 -4.17083055e-01 -8.64224315e-01 -4.27288175e-01
1.22071421e+00 -1.22194862e+00 -2.90511578e-01 2.49415606e-01
-1.15981340e+00 -6.52407169e-01 -3.84446174e-01 3.12655032e-01
-8.94141734e-01 6.12935483e-01 -7.11044788e-01 -7.20112264e-01
-1.50384352e-01 -1.04730892e+00 9.05095935e-01 1.61710367e-01
-3.20128411e-01 -1.05714691e+00 -1.72330797e-01 5.05585432e-01
1.94387272e-01 2.06894502e-01 8.71885121e-01 -1.63086593e-01
-1.16021430e+00 6.11476824e-02 -1.66842565e-01 3.62502456e-01
1.12338319e-01 3.71686041e-01 -7.63528109e-01 2.33649537e-02
8.42223167e-02 -1.26333654e-01 1.10314345e+00 8.23352516e-01
1.47417772e+00 -4.05621052e-01 -4.48746711e-01 6.98108673e-01
1.28330469e+00 1.68330058e-01 6.92534506e-01 -4.46581393e-02
1.01851511e+00 4.67432588e-01 3.41022670e-01 3.88802439e-01
4.57948893e-01 5.86098731e-01 4.43229347e-01 2.22274065e-02
-2.69189686e-01 -5.17650902e-01 5.31934738e-01 6.59881890e-01
-2.50155598e-01 -1.83424622e-01 -8.74327183e-01 4.80447441e-01
-2.14169478e+00 -1.39873219e+00 -1.40841752e-01 1.75666153e+00
5.21549582e-01 1.78878486e-01 2.21416026e-01 -4.10222672e-02
7.30704367e-01 2.06830770e-01 -7.20320106e-01 1.94872767e-01
-3.93917300e-02 -1.54441580e-01 3.52626652e-01 4.17485476e-01
-1.33260965e+00 1.26591873e+00 6.61122847e+00 5.43122828e-01
-1.31716442e+00 -1.83709994e-01 1.30378783e+00 1.91101700e-01
-3.55173312e-02 1.46766394e-01 -8.89799297e-01 4.98450369e-01
8.97036374e-01 6.85059130e-02 2.87119597e-01 9.21885669e-01
5.14030099e-01 -9.17720199e-02 -1.34385812e+00 9.59623992e-01
4.83009070e-02 -1.75351894e+00 2.81118214e-01 -3.06773305e-01
1.03358185e+00 1.59518817e-03 -1.43140396e-02 3.85624208e-02
1.27433524e-01 -9.44106042e-01 8.14156950e-01 7.89936304e-01
4.05477226e-01 -3.23479474e-01 1.23431705e-01 5.59355497e-01
-1.28667402e+00 -3.25344235e-01 -4.04798597e-01 -2.60686994e-01
2.15095207e-01 2.94324517e-01 -6.85239613e-01 -8.56715813e-02
6.36593461e-01 1.17364013e+00 -3.96419287e-01 1.02562547e+00
-1.68655533e-02 1.15115654e+00 -2.44077332e-02 3.88675451e-01
2.95648962e-01 -1.62902564e-01 2.89300293e-01 1.07374990e+00
2.50738233e-01 1.41320467e-01 4.20516372e-01 8.87827516e-01
5.90656102e-02 -1.84521452e-01 -2.67959177e-01 -7.67029747e-02
-8.41934606e-02 9.14837837e-01 -9.77383375e-01 -4.89445895e-01
-4.80746537e-01 1.35479450e+00 2.30743274e-01 5.79620063e-01
-9.12835002e-01 5.45439065e-01 5.75712204e-01 3.90303940e-01
8.92295897e-01 -3.27189654e-01 -1.10332154e-01 -1.33642018e+00
-2.07606494e-01 -5.82946539e-01 1.85377136e-01 -8.71215820e-01
-9.77514148e-01 4.67278928e-01 -3.69939238e-01 -1.43330407e+00
-5.49225152e-01 -5.85063696e-01 -6.66352391e-01 2.87409484e-01
-1.39538968e+00 -1.15925515e+00 -3.30536216e-01 5.74792624e-01
1.16113341e+00 -2.28998676e-01 5.93721211e-01 -8.43195468e-02
-5.29123187e-01 2.88539361e-02 1.92806542e-01 2.76040137e-01
2.06040129e-01 -1.03352749e+00 4.17561859e-01 8.58101189e-01
5.24102867e-01 -3.55990045e-02 7.32334495e-01 -6.10080302e-01
-1.35957289e+00 -1.41933501e+00 7.76479661e-01 -2.99454510e-01
6.44004047e-01 -2.41086707e-01 -1.02771592e+00 8.09434712e-01
7.79204145e-02 4.99897033e-01 5.90848863e-01 -4.96576786e-01
-1.86632015e-02 -2.48973235e-01 -5.86015940e-01 6.47479057e-01
1.02288842e+00 -3.44095886e-01 -2.32955217e-01 4.70606685e-01
5.76147497e-01 -3.05141568e-01 -4.17772114e-01 3.02758485e-01
4.57970351e-01 -1.26770198e+00 9.69643652e-01 -7.43175685e-01
9.22183037e-01 -2.52981186e-01 -2.04242960e-01 -9.50067580e-01
-5.61327219e-01 -7.58008957e-01 -6.13599002e-01 8.45219254e-01
2.18046129e-01 1.66573837e-01 1.55847144e+00 7.65046358e-01
1.93784028e-01 -6.57373726e-01 -5.88182449e-01 -3.57516855e-01
-3.11985224e-01 -3.36509794e-01 5.74856363e-02 7.37265885e-01
-3.25825900e-01 3.55669469e-01 -9.22041118e-01 1.98909178e-01
7.15291500e-01 5.43931961e-01 7.07871616e-01 -1.10712588e+00
-4.67409670e-01 -1.95198640e-01 -6.96331263e-01 -1.64243877e+00
4.13076639e-01 -4.01377797e-01 1.30340740e-01 -1.46358478e+00
3.02833915e-01 -7.94677623e-03 -1.48550019e-01 2.56404936e-01
-2.25146025e-01 2.17170447e-01 3.09551477e-01 4.23885286e-01
-9.07329857e-01 5.81665754e-01 1.08606517e+00 -2.92270690e-01
-1.53902128e-01 2.08975345e-01 -3.31177294e-01 1.20322239e+00
6.55117631e-01 -1.76397517e-01 -5.10346830e-01 -6.39540792e-01
-1.90222517e-01 5.65367281e-01 5.15496433e-01 -1.23557949e+00
3.69212210e-01 -2.92580456e-01 8.72434795e-01 -5.87344408e-01
4.86461312e-01 -9.57359195e-01 1.15089379e-01 5.92186272e-01
-4.46749359e-01 -7.85357133e-02 5.16774226e-03 9.02517021e-01
-3.43470037e-01 1.98808566e-01 8.41633260e-01 -1.86878741e-01
-1.42593205e+00 4.98732477e-01 -7.42209554e-01 -6.74135014e-02
1.06950080e+00 -4.98538941e-01 2.55898863e-01 -7.48996973e-01
-1.06611097e+00 9.10354480e-02 3.12639326e-01 3.10095757e-01
8.89565527e-01 -1.06415749e+00 -6.65348828e-01 1.85915187e-01
-3.13995361e-01 -8.49260166e-02 3.52297187e-01 5.27766764e-01
-5.90880334e-01 1.97922468e-01 -2.96836019e-01 -8.86909246e-01
-1.31363976e+00 4.55936939e-01 5.10880172e-01 -7.81209618e-02
-7.79071808e-01 8.63563478e-01 6.68703973e-01 3.02076966e-01
3.36997986e-01 -2.93220878e-01 -3.27410102e-01 -2.76052266e-01
6.02277935e-01 1.81417674e-01 -6.76996529e-01 -9.74476039e-01
7.82469362e-02 5.25602043e-01 5.88340219e-03 4.62599546e-01
1.61063194e+00 -2.86884815e-01 4.36767675e-02 5.68412960e-01
1.00493550e+00 -5.41099787e-01 -2.15535712e+00 -2.07054570e-01
2.03046706e-02 -3.54353189e-01 8.62717927e-02 -1.73475429e-01
-1.41949201e+00 9.60194051e-01 5.61221957e-01 7.92704895e-02
1.15058076e+00 -1.91051647e-01 9.35965300e-01 2.58065492e-01
1.38229862e-01 -9.44854081e-01 3.93137187e-01 4.58041251e-01
4.16369379e-01 -1.37415278e+00 -8.13826472e-02 -6.91242218e-01
-7.14803040e-01 1.34283030e+00 8.78953397e-01 -3.63630205e-01
6.76385283e-01 4.85777929e-02 -1.91395417e-01 1.53700098e-01
-9.33652103e-01 -2.62155477e-02 4.93398458e-01 5.52802563e-01
4.86865759e-01 -2.74803162e-01 5.48150480e-01 4.74565983e-01
2.86917776e-01 3.05203438e-01 3.67232829e-01 6.69003248e-01
-5.94292343e-01 -9.01938915e-01 -1.95658103e-01 4.26568031e-01
-5.08628607e-01 1.00801131e-02 -1.55650094e-01 2.37685025e-01
1.45099238e-01 6.15695596e-01 2.41552770e-01 -3.46038103e-01
-2.16293469e-01 8.66727307e-02 4.93931383e-01 -5.09919226e-01
-1.50847733e-01 2.36274272e-01 -1.90503180e-01 -6.79253757e-01
-7.49313414e-01 -6.44724309e-01 -1.25089264e+00 7.82907680e-02
5.24123758e-02 -1.24837287e-01 1.68924138e-01 1.00443065e+00
3.28774482e-01 5.03599167e-01 6.36092842e-01 -1.31273270e+00
-1.95519581e-01 -4.95996594e-01 -3.05378437e-01 5.74747741e-01
4.93299216e-01 -4.57858205e-01 -3.49597819e-03 9.95994627e-01] | [8.729410171508789, 0.22664016485214233] |
46083dd3-77b5-45f6-b803-d64971af0b49 | monitoring-model-deterioration-with | 2201.11676 | null | https://arxiv.org/abs/2201.11676v3 | https://arxiv.org/pdf/2201.11676v3.pdf | Monitoring Model Deterioration with Explainable Uncertainty Estimation via Non-parametric Bootstrap | Monitoring machine learning models once they are deployed is challenging. It is even more challenging to decide when to retrain models in real-case scenarios when labeled data is beyond reach, and monitoring performance metrics becomes unfeasible. In this work, we use non-parametric bootstrapped uncertainty estimates and SHAP values to provide explainable uncertainty estimation as a technique that aims to monitor the deterioration of machine learning models in deployment environments, as well as determine the source of model deterioration when target labels are not available. Classical methods are purely aimed at detecting distribution shift, which can lead to false positives in the sense that the model has not deteriorated despite a shift in the data distribution. To estimate model uncertainty we construct prediction intervals using a novel bootstrap method, which improves upon the work of Kumar & Srivastava (2012). We show that both our model deterioration detection system as well as our uncertainty estimation method achieve better performance than the current state-of-the-art. Finally, we use explainable AI techniques to gain an understanding of the drivers of model deterioration. We release an open source Python package, doubt, which implements our proposed methods, as well as the code used to reproduce our experiments. | ['Dan Saattrup Nielsen', 'Carlos Mougan'] | 2022-01-27 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 3.27472165e-02 1.87458947e-01 -2.97139436e-02 -2.75891244e-01
-7.80410409e-01 -8.89064074e-01 5.23746669e-01 6.02550328e-01
-1.60098448e-01 1.02698350e+00 -3.20802718e-01 -6.60158157e-01
-4.02159393e-01 -5.94014704e-01 -9.60995734e-01 -6.40685201e-01
-3.27715665e-01 7.01494277e-01 3.19204390e-01 3.27515274e-01
2.53371328e-01 5.65961540e-01 -1.56778109e+00 -1.32833287e-01
9.18826044e-01 1.28242207e+00 -2.02085778e-01 7.70178378e-01
1.59959093e-01 6.60275161e-01 -7.56225467e-01 -2.74081826e-01
1.15464114e-01 -1.80260271e-01 -5.97623706e-01 -2.25638554e-01
-8.60215351e-02 -2.74772674e-01 3.37261796e-01 8.32281947e-01
2.02065900e-01 -1.68068141e-01 8.49623501e-01 -1.57839477e+00
-1.09479822e-01 9.05676842e-01 -5.41394293e-01 3.03819329e-01
-6.43375739e-02 6.30299598e-02 3.07450861e-01 -2.86800325e-01
-4.06558393e-03 9.46999073e-01 9.17344570e-01 2.42769510e-01
-1.30147457e+00 -7.09178329e-01 1.06600478e-01 1.67243958e-01
-1.40304887e+00 -4.54797655e-01 5.14283240e-01 -7.20832646e-01
6.33159161e-01 2.38549367e-01 2.00991735e-01 9.32827532e-01
2.30677828e-01 3.76243293e-01 1.14352906e+00 -5.13863266e-01
7.25973904e-01 3.77074599e-01 4.09210697e-02 3.03497404e-01
4.96017843e-01 4.47331816e-01 -3.84120405e-01 -5.54787755e-01
3.40401620e-01 -1.98133156e-01 2.46010423e-02 -3.74289781e-01
-7.41253376e-01 6.95522845e-01 1.25910178e-01 9.49658751e-02
-6.30997121e-01 2.16353804e-01 1.58267602e-01 1.10912956e-01
7.27143109e-01 4.73149210e-01 -6.84654176e-01 -6.60125732e-01
-1.20559824e+00 1.54626325e-01 9.12744105e-01 7.40149260e-01
4.93631691e-01 -5.16938493e-02 -2.58809835e-01 3.62237394e-01
2.49823496e-01 3.69352907e-01 4.02107947e-02 -9.50366795e-01
1.05237350e-01 4.22012925e-01 7.01069951e-01 -7.19988108e-01
-4.06038642e-01 -6.48938060e-01 -4.27227348e-01 1.36194646e-01
4.69770581e-01 -3.94408047e-01 -6.72909558e-01 1.70317185e+00
2.71780729e-01 4.51158375e-01 -1.90083478e-02 3.37662339e-01
1.63504332e-01 3.63336205e-01 1.44326866e-01 -5.13681948e-01
7.67965317e-01 -5.63647687e-01 -4.75020796e-01 -2.31609896e-01
5.78520477e-01 -4.32225764e-01 8.58162403e-01 5.96048295e-01
-9.07927811e-01 -3.37588549e-01 -1.12132680e+00 9.23301041e-01
-1.60828680e-01 -2.04482913e-01 5.43102622e-01 9.20114875e-01
-7.66681790e-01 8.70499969e-01 -1.29880631e+00 -2.72642553e-01
3.70717645e-01 7.77211338e-02 5.16849309e-02 1.94439292e-01
-9.21289086e-01 1.05530417e+00 4.85245198e-01 1.29942164e-01
-1.10394919e+00 -7.16011107e-01 -6.50517046e-01 1.01799536e-02
6.07633770e-01 -1.68528646e-01 1.51978767e+00 -4.87071365e-01
-1.21389604e+00 3.83098304e-01 1.52196318e-01 -7.45749593e-01
7.89100051e-01 -1.45682096e-01 -3.06192935e-01 -4.28711742e-01
-2.61735111e-01 2.15842500e-01 7.69780934e-01 -1.34107637e+00
-6.57833397e-01 -3.73865098e-01 -1.09871095e-02 -2.72376209e-01
-2.42995813e-01 2.46334951e-02 -7.66030177e-02 -9.72439125e-02
-1.70332789e-01 -9.42966878e-01 -2.61019498e-01 -3.76746207e-01
-3.58861864e-01 1.20185472e-01 5.29608190e-01 -6.01415753e-01
1.64289379e+00 -1.86328983e+00 -1.49675697e-01 1.74408749e-01
-8.19713101e-02 2.48200521e-01 2.63815463e-01 4.94602144e-01
1.44047081e-01 4.20829684e-01 -7.27503479e-01 -3.43378901e-01
1.14554904e-01 2.82503307e-01 -3.23062122e-01 4.23216969e-01
2.26598635e-01 3.53823572e-01 -7.49672413e-01 -2.97412217e-01
2.77333736e-01 1.03117190e-01 -1.47707850e-01 3.50192964e-01
-3.12536091e-01 5.57372808e-01 -1.27420276e-01 6.28437161e-01
8.30930293e-01 -6.56972965e-03 -1.29397800e-02 1.38314277e-01
-2.33210459e-01 3.22354678e-03 -1.24734581e+00 9.58568871e-01
-5.92682064e-01 3.48926455e-01 -7.94435441e-02 -9.59947288e-01
9.30495799e-01 -1.60943046e-02 2.33237237e-01 -1.15106091e-01
1.59843698e-01 4.00295965e-02 -4.67442274e-02 -2.88449705e-01
2.62551546e-01 -1.34269118e-01 -1.09445423e-01 4.45358127e-01
-4.04393375e-01 -3.81281048e-01 4.37143892e-02 -1.79935172e-01
1.29054332e+00 2.82054722e-01 4.84206676e-01 -2.86992192e-01
2.05277756e-01 1.00973912e-01 5.34522712e-01 1.00498211e+00
-2.78707027e-01 2.34293103e-01 7.61863828e-01 -1.32748634e-01
-8.82224798e-01 -1.09587121e+00 -5.46703339e-01 7.36030281e-01
-1.32936791e-01 -1.27409935e-01 -8.93145621e-01 -7.01208413e-01
2.40968123e-01 1.32511353e+00 -7.62195706e-01 -3.48415464e-01
-3.57172801e-03 -7.14264095e-01 4.51147765e-01 5.62911272e-01
2.69834846e-01 -8.30464065e-01 -8.56997430e-01 2.34480008e-01
2.20094472e-02 -8.48079503e-01 1.98881283e-01 3.93816322e-01
-8.21303010e-01 -1.11994839e+00 -1.53591320e-01 8.52130055e-02
4.98340905e-01 -2.48243734e-01 1.17666852e+00 2.45630573e-02
5.17225042e-02 3.27587843e-01 -3.85319233e-01 -1.00997961e+00
-7.45794952e-01 1.59517005e-01 2.62160391e-01 -3.26831043e-01
2.14623794e-01 -3.87852281e-01 -2.52174586e-01 5.86772501e-01
-9.71352220e-01 -2.64018357e-01 3.27080399e-01 5.63699663e-01
5.87539315e-01 5.50508261e-01 7.45322943e-01 -8.25840592e-01
8.43141437e-01 -8.42967808e-01 -9.94163156e-01 5.45682669e-01
-1.10849798e+00 2.68761218e-01 5.20654857e-01 -3.75546783e-01
-8.79127085e-01 -1.28290877e-01 1.15657240e-01 -4.34794247e-01
-2.47621655e-01 8.74367297e-01 1.55827448e-01 2.19177157e-01
8.15003455e-01 -3.48914504e-01 -1.54803157e-01 -4.89767283e-01
3.06630582e-02 9.25614715e-01 4.46954846e-01 -5.77892840e-01
8.31049383e-01 2.02260002e-01 6.59156516e-02 -3.39898050e-01
-9.64015007e-01 -2.47355163e-01 -7.09868371e-01 -3.79608393e-01
2.01363325e-01 -5.71028054e-01 -5.12524366e-01 3.58211488e-01
-9.30419087e-01 -6.43683732e-01 -1.86860859e-01 3.33495468e-01
-5.20217597e-01 2.75513798e-01 4.53573046e-03 -1.49210441e+00
-1.44016594e-01 -8.86319578e-01 9.10727918e-01 2.69712001e-01
-1.59799978e-01 -1.08022225e+00 2.20737934e-01 -5.79887256e-02
6.49119616e-01 6.44362688e-01 7.36703575e-01 -7.97537148e-01
-2.87791323e-02 -4.00789201e-01 8.40149224e-02 4.17227060e-01
2.34025978e-02 3.48190635e-01 -9.36112344e-01 -4.22253817e-01
1.48779556e-01 -1.40781567e-01 5.38144946e-01 5.36833584e-01
1.34577477e+00 -2.84553856e-01 -3.46005708e-01 1.83236182e-01
1.09419096e+00 1.52885020e-01 4.56888527e-01 5.22251248e-01
4.23259661e-02 7.96290100e-01 1.08160150e+00 9.06336427e-01
2.98988134e-01 6.08678579e-01 8.20427477e-01 2.34648049e-01
6.14967048e-01 -2.34811544e-01 2.13496491e-01 2.64309257e-01
1.38781846e-01 -2.06426576e-01 -1.29690921e+00 4.42710936e-01
-1.98150706e+00 -7.90693045e-01 -6.79248497e-02 2.85967064e+00
7.02136576e-01 5.71858287e-01 3.15630823e-01 1.97174013e-01
6.28130138e-01 -2.63900608e-01 -7.97938764e-01 -4.39711452e-01
3.88275951e-01 -1.41826347e-01 6.67988181e-01 4.43494141e-01
-9.26622927e-01 6.10169232e-01 6.73385239e+00 6.02993190e-01
-8.78561437e-01 4.73162122e-02 7.12924540e-01 -2.56814919e-02
-1.29703686e-01 2.13423133e-01 -6.36880577e-01 6.22679412e-01
1.51409709e+00 -2.06951573e-01 5.27712107e-01 1.03274882e+00
3.50387126e-01 -4.96686757e-01 -1.25474501e+00 7.53222406e-01
-1.57462358e-01 -8.75695944e-01 -8.17707598e-01 -6.20406121e-02
4.77173805e-01 1.71095263e-02 -1.37292564e-01 5.27579606e-01
5.93009770e-01 -9.82473612e-01 8.61909688e-01 8.82022083e-01
5.42743266e-01 -9.98458445e-01 1.07014489e+00 7.94268072e-01
-7.36870706e-01 -2.41980299e-01 -3.76684517e-01 -2.81136543e-01
5.19196764e-02 9.02051985e-01 -1.18847680e+00 4.69117850e-01
8.45812857e-01 2.56360322e-01 -8.05215180e-01 1.26451075e+00
-2.55747765e-01 1.19295692e+00 -6.26140952e-01 1.46034732e-01
-1.32211953e-01 1.19643658e-01 4.47837114e-01 8.93542409e-01
4.12626594e-01 -3.96911502e-01 -5.18268496e-02 1.01433802e+00
3.09140384e-01 -3.40853274e-01 -4.15859997e-01 -7.20387995e-02
9.01736796e-01 1.12684798e+00 -7.15723217e-01 -8.67862403e-02
1.63153872e-01 4.45166111e-01 2.13308647e-01 1.33847728e-01
-1.08337069e+00 -1.65577680e-01 2.73967505e-01 1.50827672e-02
6.08079508e-02 -1.89007655e-01 -2.24367544e-01 -7.80098081e-01
1.77880228e-02 -6.06902003e-01 4.42808390e-01 -5.97327769e-01
-1.28924227e+00 5.30193090e-01 5.27220309e-01 -1.17378259e+00
-7.89476871e-01 -3.05228978e-01 -6.69363976e-01 7.11100817e-01
-1.32390213e+00 -8.44113052e-01 -3.96723330e-01 -2.10042764e-02
2.03971013e-01 2.38829389e-01 7.70658493e-01 -2.21304178e-01
-7.33659208e-01 4.86066699e-01 4.10038978e-01 -6.44393325e-01
5.48277140e-01 -1.22093916e+00 2.60131299e-01 9.97579992e-01
-1.15816571e-01 3.71572793e-01 1.17885983e+00 -7.98968673e-01
-9.97883677e-01 -1.02280867e+00 3.44414800e-01 -7.13956594e-01
8.28239501e-01 -3.94896835e-01 -9.08148944e-01 6.85102046e-01
-2.84485936e-01 -5.92207760e-02 7.76416957e-01 3.19442064e-01
3.16753760e-02 -1.82827711e-01 -1.44705021e+00 2.12045848e-01
6.75581813e-01 -1.33888870e-01 -2.48351559e-01 1.65562481e-01
5.64702809e-01 -3.14688325e-01 -9.98698771e-01 7.26842463e-01
5.29028177e-01 -1.14605761e+00 4.15207714e-01 -6.47250354e-01
-3.57483998e-02 -3.61864448e-01 -3.81392568e-01 -1.52215624e+00
-4.87581342e-02 -5.48658311e-01 -4.70794588e-01 1.66381454e+00
7.16470182e-01 -6.97536945e-01 4.53827798e-01 9.43386793e-01
4.70239073e-02 -6.94447279e-01 -9.67617750e-01 -1.06912374e+00
1.39680922e-01 -8.97203982e-01 9.84122515e-01 6.24657452e-01
-5.95952906e-02 -1.28811955e-01 -3.50043088e-01 5.95440567e-01
7.96102166e-01 -1.42444029e-01 7.74595141e-01 -1.32330918e+00
-4.29708660e-01 -2.15586677e-01 -4.04491097e-01 -2.17729539e-01
1.65684745e-01 -2.97735065e-01 2.22730979e-01 -1.34261394e+00
1.68781266e-01 -6.77726030e-01 -4.96988893e-01 5.36060035e-01
-5.52846678e-02 2.11037118e-02 4.64019105e-02 3.45131904e-01
-5.98789513e-01 4.39047664e-01 3.56091976e-01 1.77117944e-01
-2.76474386e-01 5.86279035e-01 -7.55239546e-01 7.03528285e-01
9.38453794e-01 -7.28454947e-01 -4.04211402e-01 -1.49987832e-01
2.97496885e-01 9.46495533e-02 3.75032246e-01 -1.29544723e+00
5.59992017e-03 -3.30231637e-01 3.01775277e-01 -6.49249315e-01
-1.10124752e-01 -1.05335021e+00 5.61326742e-01 5.03851652e-01
-3.00896108e-01 3.79435569e-02 6.32430017e-01 5.92572749e-01
1.60988450e-01 -6.19810820e-01 6.20300651e-01 2.94633806e-01
-1.91878185e-01 1.67758256e-01 -2.59104013e-01 -2.28395358e-01
1.33672667e+00 5.66877872e-02 -4.13580596e-01 -4.25562501e-01
-6.47531211e-01 4.72387791e-01 6.24472439e-01 3.18443000e-01
1.31890818e-01 -9.88426447e-01 -7.18013287e-01 -1.62027091e-01
2.47080788e-01 -2.05463469e-02 1.70865297e-01 8.98118794e-01
-3.23285460e-01 3.18922728e-01 6.65409863e-02 -7.27913439e-01
-9.52973843e-01 6.86804831e-01 4.51919973e-01 -3.51266682e-01
4.02611382e-02 5.20931542e-01 -3.73384416e-01 -4.81358439e-01
3.47717226e-01 -4.15800512e-01 -6.65863603e-02 -1.55944154e-01
5.82195103e-01 5.77366948e-01 2.81127840e-01 -1.20487101e-01
-3.91917884e-01 5.75159583e-03 1.60894588e-01 -8.50714222e-02
1.32248855e+00 -3.07550013e-01 7.06615672e-02 9.07534540e-01
5.15289247e-01 -1.70514524e-01 -1.62538075e+00 -1.20411955e-01
2.41070539e-01 -3.81340116e-01 1.28879339e-01 -1.30714381e+00
-4.72730458e-01 7.88388371e-01 9.20648515e-01 6.95810795e-01
1.14044464e+00 1.39452517e-01 1.90779378e-04 2.83851102e-02
5.84069371e-01 -1.02093446e+00 -4.31380421e-01 1.30234659e-01
9.29365635e-01 -1.42183554e+00 8.35740641e-02 -2.72163339e-02
-5.88243246e-01 8.53063405e-01 5.66722393e-01 4.31705773e-01
7.60661721e-01 4.35305923e-01 -2.41171181e-01 7.06373975e-02
-8.77291739e-01 -6.93501979e-02 4.16486599e-02 7.80881107e-01
-6.33337675e-03 3.99065703e-01 6.85430169e-02 8.98220241e-01
-3.58097702e-01 8.88346583e-02 6.20243430e-01 9.18239594e-01
-4.47448343e-01 -9.06266987e-01 -6.44258738e-01 7.08422780e-01
-4.20913875e-01 3.71717475e-02 -4.68206286e-01 6.44344866e-01
-2.14775037e-02 1.38770139e+00 5.17907031e-02 -4.76723641e-01
4.20228243e-01 1.01175793e-01 2.99484998e-01 -4.41093087e-01
-1.98111385e-01 -2.06569731e-01 2.48529464e-01 -2.52777040e-01
-2.83188224e-01 -8.71371925e-01 -9.75037396e-01 -3.98450732e-01
-7.20915854e-01 3.56034935e-01 1.15563309e+00 9.32425201e-01
4.06999260e-01 4.46430951e-01 7.72677720e-01 -7.28966713e-01
-9.48711038e-01 -1.33884716e+00 -6.11559451e-01 -5.60419336e-02
2.05601141e-01 -1.18827987e+00 -6.93153858e-01 -3.13973546e-01] | [7.358674049377441, 3.7839975357055664] |
0732672d-166a-4fc3-b732-47d6a0442fe1 | a-neural-network-approach-to-missing-marker | 1803.02665 | null | http://arxiv.org/abs/1803.02665v4 | http://arxiv.org/pdf/1803.02665v4.pdf | A Neural Network Approach to Missing Marker Reconstruction in Human Motion Capture | Optical motion capture systems have become a widely used technology in
various fields, such as augmented reality, robotics, movie production, etc.
Such systems use a large number of cameras to triangulate the position of
optical markers.The marker positions are estimated with high accuracy. However,
especially when tracking articulated bodies, a fraction of the markers in each
timestep is missing from the reconstruction. In this paper, we propose to use a
neural network approach to learn how human motion is temporally and spatially
correlated, and reconstruct missing markers positions through this model. We
experiment with two different models, one LSTM-based and one time-window-based.
Both methods produce state-of-the-art results, while working online, as opposed
to most of the alternative methods, which require the complete sequence to be
known. The implementation is publicly available at
https://github.com/Svito-zar/NN-for-Missing-Marker-Reconstruction . | ['Hedvig Kjellström', 'Taras Kucherenko', 'Jonas Beskow'] | 2018-03-07 | null | null | null | null | ['missing-markers-reconstruction'] | ['computer-vision'] | [-4.08291705e-02 -3.66374642e-01 -1.40469864e-01 -4.10036631e-02
-4.16405767e-01 -4.05053020e-01 5.06585956e-01 -1.94114298e-01
-6.35733724e-01 7.71794021e-01 -1.51781753e-01 1.14892967e-01
2.44109437e-01 -5.88082910e-01 -9.39496577e-01 -5.49388409e-01
4.72316220e-02 4.54657048e-01 3.83382857e-01 1.76152319e-01
1.49871722e-01 5.48838794e-01 -1.37110817e+00 -1.17107391e-01
3.92647415e-01 8.04194450e-01 5.24285316e-01 8.83689046e-01
1.66655630e-01 8.59402776e-01 -1.81943029e-01 -2.12633580e-01
2.10271791e-01 -3.16841841e-01 -4.47058886e-01 4.16136719e-02
3.23474467e-01 -5.28660178e-01 -6.96950018e-01 1.07156050e+00
3.91723156e-01 3.56889158e-01 1.08258367e-01 -1.11447191e+00
-2.83544213e-01 2.27075785e-01 -5.53973496e-01 2.07523219e-02
5.80680907e-01 -2.98494883e-02 2.94018775e-01 -9.58876550e-01
9.43415105e-01 9.56184924e-01 9.00291264e-01 6.48347735e-01
-7.34101593e-01 -4.89956737e-01 -2.29021218e-02 3.37744832e-01
-1.39704680e+00 -6.33183658e-01 7.88894713e-01 -5.58687627e-01
6.81769729e-01 1.10798813e-01 1.00997293e+00 1.13516545e+00
2.20138550e-01 8.66148829e-01 7.42514968e-01 -4.35479850e-01
6.98420033e-02 -1.98025569e-01 -1.42462477e-01 9.03932989e-01
2.52161354e-01 2.46008709e-01 -6.16057694e-01 -1.51523769e-01
1.42154229e+00 5.85040450e-01 -3.99521470e-01 -6.06783092e-01
-1.79011500e+00 5.10237634e-01 2.33176485e-01 2.68054873e-01
-6.49854004e-01 5.33804476e-01 1.74373627e-01 -1.18396796e-01
5.09747684e-01 -1.87283230e-03 -3.39630216e-01 -5.02971470e-01
-1.10908246e+00 2.47051999e-01 6.30690873e-01 9.78542566e-01
6.41654551e-01 1.03400029e-01 3.57347816e-01 7.05589116e-01
4.73608881e-01 4.81989235e-01 6.29311502e-01 -1.41630948e+00
3.63162786e-01 1.59682676e-01 5.41782737e-01 -1.01903152e+00
-5.28811812e-01 6.17424026e-02 -9.96691883e-01 2.55439252e-01
5.54175675e-01 -5.28056741e-01 -9.17642593e-01 1.50004387e+00
5.38494468e-01 9.87209380e-01 -2.80968487e-01 1.22677553e+00
4.76228535e-01 7.15036094e-01 -3.79674047e-01 -4.87141348e-02
9.87227678e-01 -1.13506591e+00 -9.25374925e-01 -3.25496942e-01
4.57779944e-01 -8.49771917e-01 4.71658826e-01 4.98752385e-01
-1.24023247e+00 -6.00895286e-01 -1.02116263e+00 -1.43638119e-01
-4.22093496e-02 3.03130150e-01 5.62782168e-01 1.92257375e-01
-1.05867839e+00 9.63615358e-01 -1.48026407e+00 -3.91089022e-01
6.19861707e-02 3.23697746e-01 -5.52430809e-01 -1.60500910e-02
-9.33694363e-01 9.94069397e-01 2.36976385e-01 4.92831498e-01
-7.22421527e-01 -2.03362927e-01 -1.00312483e+00 -3.44627649e-01
2.23504722e-01 -9.01678145e-01 1.52670991e+00 -7.11035013e-01
-1.91310751e+00 6.59481406e-01 -4.55247074e-01 -3.62586260e-01
7.71534443e-01 -7.52115428e-01 -1.33597463e-01 -3.11387563e-03
-1.91395685e-01 6.98467374e-01 5.84665537e-01 -1.23293221e+00
-5.98367333e-01 -2.44858801e-01 -1.15431465e-01 2.17320785e-01
1.77285537e-01 -9.48155392e-03 -7.96379685e-01 -5.11094213e-01
2.70051777e-01 -1.39087224e+00 -5.19777179e-01 4.40194160e-01
-4.07862037e-01 2.51335174e-01 6.42484009e-01 -8.06485951e-01
8.34342360e-01 -1.88141274e+00 3.16887230e-01 -1.97138041e-01
3.19083221e-02 4.20826226e-01 1.61130235e-01 3.61742198e-01
1.97449520e-01 -3.87450069e-01 -1.20585211e-01 -8.79902124e-01
-2.20670253e-01 2.39366502e-01 -9.61403251e-02 8.06340277e-01
-1.38296634e-01 6.83241963e-01 -1.04375505e+00 -3.57614696e-01
6.44006073e-01 8.37570429e-01 -1.09288990e-01 9.60618407e-02
-1.10916585e-01 7.16429591e-01 -2.70223528e-01 4.68096137e-01
4.57410455e-01 -2.73830980e-01 2.04306036e-01 6.97513670e-02
-2.78951138e-01 5.99809289e-02 -1.40850747e+00 2.22570944e+00
-3.78749192e-01 7.94320285e-01 1.15707470e-03 -6.38983488e-01
7.08905816e-01 6.16519868e-01 6.49053156e-01 -2.99881458e-01
2.51544744e-01 1.87292024e-01 -2.78571606e-01 -4.25652891e-01
7.97564328e-01 1.07465453e-01 2.61274874e-01 2.27816790e-01
-2.72235781e-01 1.84889331e-01 9.37808082e-02 -2.26172537e-01
1.03830040e+00 7.57783949e-01 3.21323961e-01 5.10821819e-01
2.33215913e-01 1.74955845e-01 7.46038496e-01 4.23776209e-01
-2.36271709e-01 9.61528361e-01 -1.71453550e-01 -5.84999263e-01
-1.11711693e+00 -6.88908517e-01 2.43738726e-01 4.17447478e-01
4.60488200e-01 -1.33465394e-01 -6.07431531e-01 4.38423753e-02
-1.39460579e-01 3.69907528e-01 -2.22396106e-01 1.01925522e-01
-1.03011036e+00 -2.25637913e-01 3.98705781e-01 6.25918329e-01
3.45907211e-01 -1.10719717e+00 -1.16990662e+00 3.22627187e-01
-2.01072514e-01 -1.21804678e+00 -4.37266499e-01 -1.91737503e-01
-1.14589894e+00 -8.87824178e-01 -1.04935789e+00 -6.01463974e-01
7.05904007e-01 4.34584796e-01 8.59336019e-01 2.51246691e-01
1.76559994e-03 1.91713065e-01 -2.47887954e-01 -3.10103387e-01
-1.00162126e-01 -2.25930706e-01 2.56387711e-01 -1.54124230e-01
1.36397824e-01 -4.56667423e-01 -5.57424307e-01 2.66328812e-01
-5.83955288e-01 2.93554634e-01 4.17724133e-01 6.96494639e-01
6.57469392e-01 -5.01390338e-01 -1.37977391e-01 -7.05314577e-01
2.23977238e-01 -3.85320961e-01 -7.26958394e-01 -1.31166399e-01
1.07490316e-01 -2.27851078e-01 3.90140563e-01 -7.93254197e-01
-6.94500446e-01 6.49836898e-01 -1.21522613e-01 -1.10887837e+00
-2.46761158e-01 5.35380602e-01 1.70491457e-01 4.99693537e-03
1.42808154e-01 7.39500523e-02 2.18067002e-02 -6.17484689e-01
2.34874427e-01 2.26310492e-01 7.16690063e-01 -2.27439359e-01
5.73645175e-01 6.65383041e-01 -6.99258521e-02 -6.95517540e-01
-4.47221100e-01 -6.51565015e-01 -8.71850133e-01 -4.52870399e-01
5.97657740e-01 -9.99935389e-01 -7.21492171e-01 8.31186473e-01
-1.65746188e+00 -4.92268711e-01 -1.58214092e-01 9.42260981e-01
-7.06624091e-01 4.64423865e-01 -9.54765618e-01 -8.96394372e-01
-1.33717418e-01 -1.09605324e+00 1.14131129e+00 3.50059241e-01
-3.07970226e-01 -9.18004453e-01 5.20391285e-01 -5.91656901e-02
-5.24529964e-02 5.05023599e-01 -9.61138234e-02 -1.75945655e-01
-8.56495559e-01 -4.32838231e-01 1.61713973e-01 -3.09516657e-02
1.18127070e-01 1.10295616e-01 -7.63022542e-01 -2.33395740e-01
-4.80335839e-02 8.45364407e-02 5.63199341e-01 6.89094663e-01
7.04474092e-01 -2.38061950e-01 -5.85646570e-01 6.43206000e-01
1.28831935e+00 3.85706961e-01 7.67381728e-01 4.64522213e-01
9.53350663e-01 3.61346036e-01 8.14937413e-01 5.97462296e-01
5.13648272e-01 1.01074564e+00 3.93454254e-01 6.92132786e-02
4.93595041e-02 -2.58194119e-01 5.25048852e-01 1.08582008e+00
-3.74031901e-01 -2.70805776e-01 -9.52790737e-01 6.81348145e-01
-2.32110333e+00 -1.04293036e+00 -2.86640942e-01 2.33884501e+00
6.14357293e-01 -1.18350536e-01 4.98360358e-02 3.47883776e-02
1.00194657e+00 3.35871965e-01 -5.64757764e-01 2.37436290e-03
1.70632899e-01 -1.64895982e-01 5.10226190e-01 6.53933406e-01
-1.14900601e+00 9.79442239e-01 5.78343344e+00 1.90753743e-01
-1.32261646e+00 2.15465337e-01 6.64910078e-02 -3.08452755e-01
3.02113444e-01 -2.79790926e-04 -7.34515011e-01 7.20510304e-01
1.06092453e+00 2.04666391e-01 3.81833792e-01 6.77966475e-01
4.33612674e-01 -3.90986294e-01 -9.36465859e-01 1.19144905e+00
-7.00395554e-03 -1.48216021e+00 -4.46524948e-01 2.54277252e-02
7.03980565e-01 2.56410241e-01 -3.30627114e-01 -3.03565770e-01
3.29117596e-01 -6.20234489e-01 1.00799537e+00 1.02404833e+00
6.05853140e-01 -5.24779618e-01 7.02516198e-01 6.57274902e-01
-1.07691514e+00 3.59404892e-01 -4.63014960e-01 -3.87528569e-01
7.72580802e-01 7.84815788e-01 -5.75124443e-01 5.07802725e-01
6.80526495e-01 8.84196579e-01 -4.65486385e-02 1.42780507e+00
-2.14139298e-01 2.10330576e-01 -4.66264665e-01 -2.75920872e-02
-2.08227616e-02 -3.73646140e-01 6.13315642e-01 7.55557001e-01
7.69037962e-01 -5.51479980e-02 3.45178097e-01 4.89274591e-01
5.85421249e-02 -4.87971067e-01 -8.60835373e-01 2.13137582e-01
6.37406886e-01 1.16833425e+00 -6.91820502e-01 -5.24434388e-01
-4.37389642e-01 1.20422745e+00 1.71504363e-01 1.92900240e-01
-9.18655396e-01 -1.85477644e-01 5.59877753e-01 4.93629985e-02
2.53157675e-01 -9.95787501e-01 1.26600377e-02 -1.25797999e+00
2.34884724e-01 -5.36531806e-01 -9.43136737e-02 -1.13780677e+00
-7.84124196e-01 4.44683254e-01 -7.97376111e-02 -1.59529543e+00
-7.25558996e-01 -4.88576382e-01 -3.95881623e-01 6.77120805e-01
-1.31110168e+00 -1.00266898e+00 -3.66746873e-01 4.47170079e-01
6.08616173e-01 2.75619090e-01 6.35887861e-01 6.20952487e-01
-6.03341579e-01 -6.38097599e-02 3.39666158e-01 3.04814517e-01
6.83889389e-01 -9.08676445e-01 7.22200394e-01 1.00625002e+00
5.02656341e-01 5.75493813e-01 7.15124071e-01 -7.92403221e-01
-1.52002490e+00 -9.45567071e-01 8.14758837e-01 -5.38224757e-01
4.96221066e-01 -7.00625330e-02 -8.57729316e-01 1.04033399e+00
9.81620401e-02 4.79083359e-01 2.19927311e-01 -4.06850964e-01
3.09493750e-01 2.19085261e-01 -8.05457413e-01 5.92355788e-01
1.04519844e+00 -3.83125663e-01 -4.01013196e-01 2.55546361e-01
5.28246462e-01 -1.11798251e+00 -7.80766249e-01 1.77983910e-01
8.08831632e-01 -1.04313231e+00 9.97486293e-01 -1.43210977e-01
2.74361402e-01 -5.82366407e-01 8.26501846e-03 -1.23457110e+00
-1.88041762e-01 -6.10119343e-01 -6.11217082e-01 9.42254007e-01
3.65870781e-02 -5.34334481e-01 1.03294575e+00 5.79165637e-01
-1.45703226e-01 -6.48170054e-01 -9.82513428e-01 -6.82080388e-01
-5.00642538e-01 -2.90642887e-01 3.96949351e-01 9.48909521e-01
-2.69000977e-01 1.42339878e-02 -8.37406874e-01 3.87103885e-01
5.29990852e-01 -1.45901106e-02 9.58215773e-01 -1.14738524e+00
-1.27983153e-01 -6.61696047e-02 -4.58575517e-01 -1.32503462e+00
9.34485421e-02 -2.34731197e-01 4.00978863e-01 -1.64675760e+00
-1.50666580e-01 -3.00601482e-01 -6.72008246e-02 3.48645777e-01
1.67634040e-02 3.35505724e-01 3.96171778e-01 3.98923665e-01
-6.04127705e-01 3.79037976e-01 9.89783049e-01 9.61646065e-02
-1.85431808e-01 1.06882498e-01 1.89106718e-01 1.22042716e+00
9.09490347e-01 -6.27258599e-01 -6.79438487e-02 -8.48041236e-01
-2.67600268e-03 4.86965537e-01 6.84485853e-01 -1.29113305e+00
6.24556601e-01 -2.39050880e-01 6.10836029e-01 -9.18596327e-01
9.33899999e-01 -8.57073128e-01 8.50485861e-01 6.49953187e-01
7.82515854e-02 6.25101149e-01 1.02455579e-01 5.02191305e-01
-2.05459863e-01 -4.04829979e-01 5.58353961e-01 -4.16143417e-01
-9.18925107e-01 4.34286684e-01 -2.90136904e-01 -3.89408678e-01
8.10925126e-01 -3.97520572e-01 -2.63137549e-01 -4.09981042e-01
-5.05596578e-01 -1.31958872e-01 9.17822897e-01 4.53487992e-01
7.66237020e-01 -1.41311061e+00 -3.33650559e-01 -5.83423376e-02
-3.20128053e-01 3.64140749e-01 3.55474234e-01 9.58026350e-01
-9.96328592e-01 3.89697433e-01 -2.25687295e-01 -7.73970544e-01
-1.13013029e+00 5.10713577e-01 2.33519897e-01 -2.68034399e-01
-8.73121083e-01 5.85609972e-01 -3.19409877e-01 -3.91201705e-01
1.00281239e-01 -2.79009163e-01 -6.05197763e-03 -2.40364447e-01
5.16499698e-01 6.04227483e-01 -1.36110231e-01 -8.53105366e-01
-2.83210099e-01 9.73548532e-01 3.68256330e-01 -3.64705652e-01
1.30843103e+00 -2.15708748e-01 -5.30628413e-02 9.14307058e-01
9.29419935e-01 -5.60393669e-02 -1.52124560e+00 -2.25295231e-01
1.73260167e-01 -6.08248353e-01 -3.81293744e-02 -2.90380090e-01
-1.10918498e+00 1.01893377e+00 5.97045720e-01 -1.03159830e-01
6.42185330e-01 -5.43601453e-01 1.09179330e+00 4.31844473e-01
7.62901247e-01 -7.85940409e-01 -2.26648122e-01 6.21199250e-01
6.53651536e-01 -1.11707306e+00 1.57664135e-01 -2.77874082e-01
-4.89193022e-01 9.11222160e-01 4.82916892e-01 -3.41972321e-01
4.36118603e-01 3.93688411e-01 2.33298317e-01 7.47541264e-02
-6.26069844e-01 -6.78760409e-02 -1.75078828e-02 5.69847167e-01
3.91455501e-01 -5.06106839e-02 8.17828774e-02 7.70478845e-02
-7.28316456e-02 3.96646589e-01 6.02614045e-01 1.31040561e+00
-1.85041785e-01 -1.10300052e+00 -6.49539590e-01 1.22704677e-01
-3.44634533e-01 6.63997680e-02 -1.25568837e-01 8.04751873e-01
-1.19968250e-01 6.78979576e-01 2.47100666e-01 -3.08890253e-01
3.26742947e-01 -5.94110191e-02 5.97724319e-01 -4.27625775e-01
-3.07103068e-01 1.73290879e-01 1.26415625e-01 -7.10737884e-01
-7.76122272e-01 -8.17452013e-01 -1.37177801e+00 -5.08632481e-01
-3.97550702e-01 -1.50197595e-01 7.05799699e-01 7.05797970e-01
4.92338598e-01 3.25804681e-01 2.50712037e-01 -1.67963052e+00
-8.61553177e-02 -8.69191170e-01 -4.99171138e-01 2.15141848e-01
6.11287773e-01 -8.64845157e-01 2.21793149e-02 3.48803490e-01] | [7.498161315917969, -1.1025842428207397] |
191c9e22-08dd-4c5a-ab2a-9040863f38a3 | diva-deep-unfolded-network-from-quantum | 2301.00247 | null | https://arxiv.org/abs/2301.00247v1 | https://arxiv.org/pdf/2301.00247v1.pdf | DIVA: Deep Unfolded Network from Quantum Interactive Patches for Image Restoration | This paper presents a deep neural network called DIVA unfolding a baseline adaptive denoising algorithm (De-QuIP), relying on the theory of quantum many-body physics. Furthermore, it is shown that with very slight modifications, this network can be enhanced to solve more challenging image restoration tasks such as image deblurring, super-resolution and inpainting. Despite a compact and interpretable (from a physical perspective) architecture, the proposed deep learning network outperforms several recent algorithms from the literature, designed specifically for each task. The key ingredients of the proposed method are on one hand, its ability to handle non-local image structures through the patch-interaction term and the quantum-based Hamiltonian operator, and, on the other hand, its flexibility to adapt the hyperparameters patch-wisely, due to the training process. | ['Denis Kouamé', 'Bertrand Georgeot', 'Adrian Basarab', 'Sayantan Dutta'] | 2022-12-31 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [ 4.26471174e-01 9.03461874e-02 2.54655808e-01 -8.94262642e-02
-5.97674251e-01 -1.85826659e-01 6.01624608e-01 -3.32891941e-03
-3.11719298e-01 8.06403100e-01 2.13931590e-01 1.29030775e-02
-3.26409608e-01 -6.83935881e-01 -9.74155545e-01 -1.22607386e+00
1.53241068e-01 2.46715844e-01 -4.39208671e-02 -6.10494137e-01
2.85353601e-01 5.23226678e-01 -1.32169592e+00 -7.83005804e-02
1.05644643e+00 8.23625863e-01 -2.31257230e-02 5.08862078e-01
4.78071719e-01 7.58761883e-01 -4.29095328e-02 -3.38626891e-01
3.25704992e-01 -8.04917634e-01 -8.31857979e-01 8.56462419e-02
7.23530710e-01 -4.84987438e-01 -7.39398420e-01 1.20312357e+00
5.49962640e-01 4.60449487e-01 4.47600037e-01 -5.86944580e-01
-9.10311818e-01 1.51669502e-01 -2.96668202e-01 8.82827342e-02
7.45484382e-02 2.32492417e-01 8.92328262e-01 -7.15952754e-01
6.34880185e-01 1.10954118e+00 8.82613242e-01 7.07589328e-01
-1.38791239e+00 -1.91334799e-01 -5.73650062e-01 5.81619084e-01
-1.16706872e+00 -3.50131392e-01 9.30439532e-01 -2.80168593e-01
7.69483864e-01 9.46130604e-02 4.83390719e-01 1.00374472e+00
6.61389410e-01 1.90532476e-01 1.46422982e+00 -6.60000086e-01
3.33805978e-01 -2.65466541e-01 2.10648715e-01 7.60303378e-01
1.21819474e-01 3.15732241e-01 -5.52326381e-01 -3.21312308e-01
8.60522330e-01 -2.50131667e-01 -5.05355358e-01 -4.74648923e-01
-1.08111989e+00 8.23741019e-01 6.61373675e-01 2.37341478e-01
-6.28362894e-01 3.33995700e-01 4.13921028e-01 3.50494713e-01
4.37496036e-01 4.24377590e-01 -1.40579790e-01 3.28982800e-01
-8.68546546e-01 4.51693296e-01 6.55046582e-01 3.47101271e-01
9.57443416e-01 1.47289321e-01 -2.34680787e-01 4.30432409e-01
8.77257735e-02 3.86364251e-01 3.49535435e-01 -1.23241162e+00
-3.69996317e-02 -7.70426840e-02 3.32007825e-01 -1.09264863e+00
-4.99190390e-01 -5.13074517e-01 -1.36055720e+00 3.94404173e-01
2.13234544e-01 1.71805304e-02 -8.52052569e-01 2.04498911e+00
3.68119419e-01 4.27238107e-01 -1.42582401e-03 1.04879034e+00
6.22631192e-01 6.10803604e-01 -8.22576806e-02 -3.93898517e-01
1.31596482e+00 -9.12042737e-01 -1.01842713e+00 -2.34018341e-02
2.74754643e-01 -5.74745059e-01 8.18219602e-01 3.52582991e-01
-1.25327957e+00 -4.67743695e-01 -1.21446860e+00 -5.52807748e-01
-3.32455859e-02 -2.71895230e-01 6.54702485e-01 3.72067839e-01
-1.39612103e+00 1.31512046e+00 -8.54818225e-01 -3.53605986e-01
1.97291538e-01 4.28924292e-01 -3.51978213e-01 -1.29155785e-01
-1.37539399e+00 1.06666386e+00 2.26184413e-01 2.60164112e-01
-7.66805172e-01 -5.83855748e-01 -5.61698079e-01 2.30780721e-01
1.86953947e-01 -1.28954935e+00 8.48545611e-01 -9.74165618e-01
-1.86700726e+00 6.29991293e-01 -1.57367349e-01 -7.04370022e-01
4.86118942e-01 -9.27724242e-02 -9.72626060e-02 5.63814282e-01
-1.87303409e-01 2.80491412e-01 1.17060256e+00 -1.03689432e+00
2.33981520e-01 -4.11867499e-01 9.12617296e-02 1.54057771e-01
-1.42584801e-01 -2.20650807e-01 -9.08781141e-02 -7.41386354e-01
1.78797454e-01 -9.22893524e-01 -4.72685099e-01 5.22756763e-02
-2.33466193e-01 1.83470458e-01 4.51280504e-01 -9.70386684e-01
6.34106815e-01 -2.06000900e+00 9.22062457e-01 4.35922369e-02
3.51538718e-01 1.87725797e-01 -3.02219301e-01 6.57131612e-01
-4.75605786e-01 -1.57988116e-01 -5.88165522e-01 -3.74711663e-01
-1.11925773e-01 3.65325987e-01 -2.86101401e-01 7.68003762e-01
1.72600731e-01 8.72216105e-01 -6.92987561e-01 -9.35533494e-02
4.44123596e-01 9.18309510e-01 -6.42623663e-01 -1.59280032e-01
-3.18895429e-02 1.03552425e+00 -2.42194206e-01 -1.79563705e-02
1.00091767e+00 -1.75930858e-01 -4.05757539e-02 -5.16903698e-01
-1.08338773e-01 1.97310999e-01 -1.10054064e+00 1.80090582e+00
-4.97153342e-01 3.82813513e-01 4.98521537e-01 -1.37465191e+00
3.43911737e-01 4.14830625e-01 5.33155143e-01 -8.65573347e-01
8.11292976e-02 2.85756290e-01 6.83193887e-03 -3.81516188e-01
5.09980679e-01 -4.97529656e-01 3.95404905e-01 4.17060077e-01
2.89624482e-01 3.80652621e-02 3.27032283e-02 2.44358942e-01
1.17579615e+00 5.18413484e-02 2.55275160e-01 -4.44192439e-01
7.07771778e-01 -2.35087827e-01 3.75650972e-01 9.99753654e-01
-2.09956944e-01 8.00933003e-01 1.88704953e-01 -7.14407146e-01
-1.40092409e+00 -7.86122382e-01 -4.50214326e-01 7.75213361e-01
1.71781421e-01 -1.94387734e-01 -1.06499124e+00 -7.76335746e-02
-3.77544731e-01 5.60819328e-01 -6.45618558e-01 -3.20495367e-01
-7.75039375e-01 -1.03195298e+00 1.28840730e-01 4.01369436e-03
6.57017410e-01 -1.13937926e+00 -3.62682343e-01 2.77674049e-01
-3.33728731e-01 -1.10298634e+00 -2.71055102e-01 9.82774198e-02
-9.53396022e-01 -8.58446002e-01 -8.05407822e-01 -3.50505024e-01
3.90600741e-01 3.11847866e-01 1.03417909e+00 2.43889257e-01
-1.97628334e-01 2.96844453e-01 -2.12462425e-01 2.32378826e-01
-6.75314426e-01 -8.29353034e-02 1.94330774e-02 1.32449880e-01
-2.29561299e-01 -1.08978140e+00 -7.12794006e-01 -1.17432855e-01
-1.34916890e+00 1.10527925e-01 5.09891629e-01 1.29340076e+00
6.16440535e-01 1.90879345e-01 3.48696768e-01 -6.28732443e-01
4.68384594e-01 -1.82610914e-01 -4.38777745e-01 6.84501976e-02
-4.89320219e-01 3.80469412e-01 7.40074873e-01 -1.23418316e-01
-1.07536077e+00 -5.82332388e-02 -4.91706222e-01 -8.14306736e-02
-7.17755258e-02 3.89575601e-01 -3.56701612e-02 -7.39848018e-01
6.99139476e-01 3.19260359e-01 1.21308394e-01 -4.40698773e-01
4.92532820e-01 1.35067627e-01 7.28632212e-01 -3.93431991e-01
1.00854301e+00 8.19296360e-01 6.56856656e-01 -8.44088972e-01
-1.00834179e+00 -2.02252597e-01 -7.92205095e-01 7.73367211e-02
1.02544856e+00 -8.99206042e-01 -6.92743838e-01 7.57548153e-01
-1.28871560e+00 -1.62956193e-01 -3.09895158e-01 2.93339849e-01
-8.14085305e-01 8.11485112e-01 -1.02762246e+00 -5.30443847e-01
-5.02723813e-01 -1.09264696e+00 9.16553140e-01 2.35661596e-01
3.28345478e-01 -1.13621044e+00 8.36381763e-02 4.33242887e-01
5.12940288e-01 3.50332081e-01 1.09668088e+00 -1.53063208e-01
-6.80594027e-01 8.33097566e-03 -1.68611407e-01 6.41048968e-01
-1.67982683e-01 -4.11476851e-01 -9.30593789e-01 -4.75214005e-01
6.00819767e-01 -3.36716443e-01 1.19531989e+00 6.41292274e-01
1.06535375e+00 -3.00507724e-01 2.57437497e-01 8.65900874e-01
1.55086386e+00 -4.10931081e-01 8.93848598e-01 3.05067778e-01
8.13660920e-01 4.05467927e-01 -1.52484715e-01 2.42500275e-01
2.06620127e-01 9.63792145e-01 6.58048153e-01 -1.42892525e-01
-3.69071245e-01 2.22405314e-01 3.03049713e-01 9.09863412e-01
-4.00102258e-01 6.41834140e-02 -5.36864460e-01 1.93673313e-01
-1.96186733e+00 -1.22078395e+00 -4.19691980e-01 2.17050481e+00
6.87227845e-01 -3.14198911e-01 -2.48147190e-01 -5.52196838e-02
5.19305468e-01 2.87523776e-01 -7.31393933e-01 -4.70430046e-01
-4.96607274e-01 5.83596826e-01 5.58391452e-01 6.17951274e-01
-1.08895624e+00 7.69949496e-01 6.12824774e+00 8.65544975e-01
-9.57630336e-01 5.31210124e-01 3.23114723e-01 3.64516199e-01
-1.89917609e-01 2.35580700e-03 -2.07324117e-01 1.11334383e-01
8.74140263e-01 1.25307590e-01 1.06226182e+00 3.77072155e-01
3.40909719e-01 6.15159683e-02 -8.05180490e-01 8.88617754e-01
1.52733132e-01 -1.44239020e+00 8.67485777e-02 1.38176993e-01
7.97363818e-01 1.80950850e-01 1.54790848e-01 -8.40701312e-02
-2.28752583e-01 -9.44584191e-01 6.02771580e-01 6.88028157e-01
3.56094509e-01 -7.92521715e-01 8.91161680e-01 3.16442043e-01
-4.48453367e-01 -2.47843955e-02 -4.91032511e-01 -2.53417909e-01
2.86426783e-01 7.85142064e-01 1.85839951e-01 8.78609061e-01
6.48141384e-01 7.77153552e-01 -4.63428825e-01 8.85131657e-01
-2.77222455e-01 5.37436962e-01 -9.54435319e-02 5.08357227e-01
2.29020238e-01 -7.83387303e-01 8.15287292e-01 6.95755005e-01
4.15212005e-01 2.74060339e-01 -2.34513685e-01 1.01980340e+00
-4.87616621e-02 -1.29383549e-01 -3.07540834e-01 2.50744253e-01
-1.96804047e-01 1.38875556e+00 -4.03069824e-01 -9.99267474e-02
-2.97158480e-01 1.35811174e+00 2.59997457e-01 5.13984680e-01
-7.36241341e-01 -2.05476835e-01 4.03965712e-01 5.58542684e-02
4.35265779e-01 -2.83047259e-01 -2.04270273e-01 -1.40864003e+00
6.84376210e-02 -8.67668867e-01 -5.30690067e-02 -8.77112448e-01
-1.43760502e+00 4.53856617e-01 -2.31191456e-01 -8.36137414e-01
-8.63975883e-02 -6.75045371e-01 -4.90634114e-01 1.03330863e+00
-1.59088135e+00 -1.15554810e+00 -7.20481277e-02 5.89480281e-01
-1.07118666e-01 3.33334595e-01 9.79484916e-01 2.92843282e-01
-6.03953123e-01 2.44150117e-01 7.80292332e-01 -1.74969122e-01
5.36743820e-01 -1.22793710e+00 3.96432281e-01 9.04468000e-01
-5.90070561e-02 7.84266591e-01 1.22590959e+00 -2.56239295e-01
-1.61136913e+00 -7.92556286e-01 5.88501632e-01 -1.25496909e-01
7.44088471e-01 -3.56246233e-01 -1.11998260e+00 4.93671507e-01
5.80850720e-01 1.92191735e-01 3.05786937e-01 -7.87712410e-02
-2.90305555e-01 -3.39125283e-02 -1.30898523e+00 4.93548632e-01
8.79078269e-01 -6.75123334e-01 -4.47384596e-01 7.91479766e-01
4.65482891e-01 -4.65805590e-01 -7.97014892e-01 1.07683264e-01
3.40498686e-01 -1.30487883e+00 1.22438109e+00 -4.81980950e-01
5.64239442e-01 -1.09392896e-01 8.85048732e-02 -1.42673767e+00
-5.72428882e-01 -1.00489318e+00 -1.77076980e-01 8.23958695e-01
-1.73333675e-01 -6.45162702e-01 4.03489739e-01 4.96214807e-01
-3.08146298e-01 -4.42782640e-01 -1.39724159e+00 -5.49997032e-01
3.29706132e-01 -6.96829632e-02 1.15013346e-01 7.17893600e-01
-4.24088061e-01 4.76830304e-01 -9.99518514e-01 3.55418533e-01
9.55317378e-01 1.91014297e-02 4.05250698e-01 -1.14556170e+00
-6.03593290e-01 -7.10109845e-02 -3.14538687e-01 -9.46171343e-01
3.08120728e-01 -7.37485707e-01 -2.27981973e-02 -1.14914191e+00
3.49146843e-01 3.83025147e-02 -3.58425975e-01 2.17484608e-01
-2.29907178e-04 5.37813485e-01 1.09313190e-01 3.84366184e-01
-4.39195812e-01 8.57848644e-01 1.41492367e+00 -6.06354773e-02
-5.34444302e-02 -1.82215169e-01 -5.07054627e-01 8.13279986e-01
5.66018164e-01 -5.14926374e-01 7.35971108e-02 -4.78628337e-01
5.19106448e-01 1.77739441e-01 9.98989582e-01 -1.03531444e+00
1.34426087e-01 1.96353093e-01 1.02423295e-01 -1.44690216e-01
4.69367832e-01 -5.32473266e-01 1.00414358e-01 6.40442371e-01
-2.78827846e-01 -2.95566142e-01 6.27178550e-02 7.27140188e-01
-1.19257860e-01 -4.97591704e-01 1.16385591e+00 -2.14589760e-01
-4.64774728e-01 1.26377672e-01 -3.75480980e-01 -2.12343097e-01
4.65711772e-01 5.76664284e-02 -2.84612685e-01 -5.26107311e-01
-7.70884752e-01 -4.77855593e-01 5.88721871e-01 -1.77111700e-01
2.55524009e-01 -1.27913451e+00 -6.30899549e-01 -4.32165787e-02
-1.75973296e-01 -2.24020645e-01 8.94192100e-01 1.22386038e+00
-7.70417273e-01 2.63162106e-01 -3.88223708e-01 -4.68213260e-01
-8.09616923e-01 6.25556529e-01 6.00673795e-01 -3.70846719e-01
-8.71745586e-01 5.24466395e-01 1.65614381e-01 -3.97104770e-01
-2.36084342e-01 -1.08242452e-01 1.05055764e-01 -4.83104438e-01
6.98635399e-01 3.79395962e-01 3.69859576e-01 -9.35518444e-01
-9.42338854e-02 7.73542941e-01 6.62312284e-02 1.62881128e-02
1.62046051e+00 -4.08476204e-01 -5.91323972e-01 4.11161073e-02
1.03717053e+00 -8.50066096e-02 -1.30389655e+00 -2.42943019e-01
-2.91619807e-01 3.39084445e-03 4.96681809e-01 -6.31196320e-01
-1.07005417e+00 9.32748675e-01 7.96288967e-01 2.81945258e-01
1.30986345e+00 -2.18028203e-01 1.01319635e+00 5.58706582e-01
2.11441129e-01 -8.69399846e-01 -6.51474297e-02 5.06293595e-01
8.90391469e-01 -1.06562591e+00 1.51461482e-01 -3.07654619e-01
-1.22912869e-01 1.21335244e+00 -8.64368975e-02 -5.25954902e-01
5.41595995e-01 -1.39179781e-01 -2.21655339e-01 -1.19428441e-01
-3.05291891e-01 -1.04330136e-02 1.87862664e-01 5.43858767e-01
3.41598690e-01 -2.57810950e-01 -5.67250550e-01 1.02069356e-01
2.42534816e-01 3.36141773e-02 7.63713837e-01 3.76845092e-01
-2.07703993e-01 -1.22897017e+00 -3.67166013e-01 2.88598090e-02
-4.54078346e-01 -2.09347516e-01 -1.13359690e-01 6.98241353e-01
1.77598596e-01 8.17908049e-01 -2.74797678e-01 8.60843286e-02
7.73440823e-02 -1.46079034e-01 6.74098611e-01 -2.83008039e-01
-6.06801569e-01 7.16508254e-02 -3.36270481e-01 -7.19403207e-01
-9.20549870e-01 -6.05785489e-01 -9.39371407e-01 -5.58381557e-01
-3.21239531e-01 -1.06620699e-01 5.66977262e-01 1.24014211e+00
6.08958304e-01 5.13860583e-01 3.42894882e-01 -1.35994160e+00
-8.03593218e-01 -9.47813034e-01 -7.10846245e-01 7.40022540e-01
5.39302289e-01 -6.40508831e-01 -4.29847449e-01 -8.44056532e-02] | [11.74710464477539, -2.425461769104004] |
af966639-7259-4011-aab8-4abcdeb64274 | dexmv-imitation-learning-for-dexterous | 2108.05877 | null | https://arxiv.org/abs/2108.05877v5 | https://arxiv.org/pdf/2108.05877v5.pdf | DexMV: Imitation Learning for Dexterous Manipulation from Human Videos | While significant progress has been made on understanding hand-object interactions in computer vision, it is still very challenging for robots to perform complex dexterous manipulation. In this paper, we propose a new platform and pipeline DexMV (Dexterous Manipulation from Videos) for imitation learning. We design a platform with: (i) a simulation system for complex dexterous manipulation tasks with a multi-finger robot hand and (ii) a computer vision system to record large-scale demonstrations of a human hand conducting the same tasks. In our novel pipeline, we extract 3D hand and object poses from videos, and propose a novel demonstration translation method to convert human motion to robot demonstrations. We then apply and benchmark multiple imitation learning algorithms with the demonstrations. We show that the demonstrations can indeed improve robot learning by a large margin and solve the complex tasks which reinforcement learning alone cannot solve. More details can be found in the project page: https://yzqin.github.io/dexmv | ['Xiaolong Wang', 'Yang Fu', 'Ruihan Yang', 'Hanwen Jiang', 'Shaowei Liu', 'Yueh-Hua Wu', 'Yuzhe Qin'] | 2021-08-12 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [-2.38660812e-01 -1.29375979e-01 -1.56928986e-01 6.73952000e-03
-3.95130664e-01 -7.74234176e-01 5.01542687e-01 -9.52147722e-01
-4.16517526e-01 6.75023139e-01 -3.09946865e-01 -3.65642399e-01
2.05634329e-02 1.53403925e-02 -1.08639669e+00 -5.24521470e-01
-2.35518903e-01 8.57700288e-01 4.20858115e-01 -1.30699977e-01
2.60053009e-01 8.21309626e-01 -1.39810824e+00 2.17786990e-02
5.98407090e-01 4.49518234e-01 9.42475557e-01 1.08262050e+00
6.29962385e-01 9.51868296e-01 -2.94532925e-01 3.66820157e-01
7.62719989e-01 -2.53855586e-01 -8.05330813e-01 1.27328439e-02
3.33267421e-01 -9.98391509e-01 -9.96914029e-01 9.34359968e-01
4.84596610e-01 1.35167226e-01 6.61380172e-01 -1.91374528e+00
-2.69450724e-01 6.09429002e-01 -4.12632883e-01 -4.02778596e-01
5.52454054e-01 8.57125044e-01 6.82166636e-01 -5.94925463e-01
1.00635517e+00 1.52759564e+00 3.33236635e-01 7.39049792e-01
-8.54376793e-01 -7.93913305e-01 -6.67887926e-02 2.73527563e-01
-1.03044868e+00 -1.54195800e-01 3.36284578e-01 -5.75042129e-01
1.24057090e+00 -1.89300850e-01 6.75980330e-01 1.56183743e+00
3.95893455e-01 1.15592170e+00 7.70295799e-01 -3.20748687e-01
-1.85975999e-01 -3.83093953e-01 -4.18555945e-01 9.86306190e-01
1.34513471e-02 4.60011512e-01 -7.86890232e-05 3.53766799e-01
1.45856965e+00 3.53177577e-01 -3.00772578e-01 -9.83212709e-01
-1.87306035e+00 4.17809814e-01 6.26069427e-01 7.51630217e-02
-5.10504186e-01 7.50079989e-01 1.71808496e-01 7.84335971e-01
-7.63514578e-01 6.36916518e-01 -7.28548348e-01 -4.46641713e-01
-1.35693640e-01 7.69126177e-01 1.24798346e+00 1.71461225e+00
4.21260655e-01 -1.26069158e-01 1.12876102e-01 3.06392848e-01
2.35444918e-01 7.82028139e-01 2.13169500e-01 -1.69731033e+00
5.67093313e-01 2.74722755e-01 5.72370648e-01 -2.54257321e-01
-5.48584580e-01 3.78798455e-01 -2.99338490e-01 1.08212328e+00
5.87121546e-01 -2.39942580e-01 -7.33408511e-01 1.32943261e+00
3.45933765e-01 -2.18949944e-01 4.37970683e-02 1.24593055e+00
4.54906255e-01 4.00152862e-01 -2.45657608e-01 2.87646025e-01
1.05037785e+00 -1.58877313e+00 -4.65154439e-01 -1.57748550e-01
6.44870043e-01 -7.06628501e-01 1.14746070e+00 6.26414657e-01
-8.21827173e-01 -6.15781665e-01 -1.02033925e+00 -1.65557310e-01
6.83505237e-02 4.09176469e-01 6.77240312e-01 -3.06382924e-01
-5.86647391e-01 1.06256557e+00 -1.44202125e+00 -4.77090865e-01
1.49398580e-01 5.53607821e-01 -4.81608242e-01 -2.82575525e-02
-6.05929017e-01 1.42987299e+00 4.82098639e-01 -1.70866475e-02
-1.56520915e+00 -2.48796850e-01 -8.35534036e-01 -3.25106919e-01
7.53678381e-01 -7.04130054e-01 1.82616615e+00 -6.01970315e-01
-1.80601525e+00 6.21183097e-01 4.77220982e-01 -8.92127827e-02
1.03784919e+00 -6.20931566e-01 2.39776462e-01 3.53098094e-01
-2.64575463e-02 1.02628326e+00 1.10622299e+00 -1.27162814e+00
-6.90631926e-01 -2.26319820e-01 3.13448757e-01 3.17159712e-01
2.83932954e-01 -1.08394712e-01 -4.68277335e-01 -4.84575748e-01
-1.92114353e-01 -1.68907905e+00 -7.93011710e-02 6.61046445e-01
-2.57550895e-01 -4.91124511e-01 1.07042754e+00 -6.25941217e-01
1.62530355e-02 -1.96023262e+00 9.87303138e-01 -2.99987167e-01
1.50636107e-01 2.34996170e-01 -4.19720918e-01 5.69921851e-01
2.77460456e-01 -6.93801284e-01 3.71591523e-02 -3.99771929e-02
2.93493301e-01 4.64067340e-01 -3.12206566e-01 4.66886222e-01
-8.55832249e-02 1.20020258e+00 -1.28213286e+00 -2.95837075e-01
4.32936460e-01 2.49567777e-01 -3.88532788e-01 6.20414615e-01
-6.48178518e-01 8.84259284e-01 -5.22193372e-01 8.46068501e-01
1.83464512e-01 -9.06358510e-02 3.54962885e-01 -1.00813642e-01
-6.00032061e-02 -6.24240786e-02 -1.03667367e+00 2.19623256e+00
-4.82308835e-01 6.02250218e-01 5.66021025e-01 -7.72996604e-01
5.21138549e-01 2.94673264e-01 3.09600949e-01 -8.32219124e-02
2.34146699e-01 4.96141493e-01 4.14269149e-01 -1.09014237e+00
2.23770216e-01 1.84288025e-01 8.18007737e-02 3.87784600e-01
2.83129990e-01 -9.53302085e-01 1.59417257e-01 -1.08427249e-01
1.40450203e+00 1.16036332e+00 1.50902018e-01 2.68177688e-01
5.46050668e-02 5.22445023e-01 -2.04401836e-02 7.31895030e-01
-2.97876507e-01 3.20580333e-01 2.27263451e-01 -6.56992942e-02
-1.44491255e+00 -1.34635758e+00 4.51526344e-01 9.57806528e-01
2.09485695e-01 3.25337797e-02 -4.02810931e-01 -5.81497550e-01
6.60228074e-01 4.74139720e-01 -2.30490834e-01 1.60183102e-01
-8.25967073e-01 3.48403871e-01 4.34359968e-01 8.30048323e-01
3.07406336e-01 -1.69962931e+00 -1.09205365e+00 -9.14917812e-02
1.39355764e-01 -1.12289476e+00 -4.82358038e-01 1.09717973e-01
-7.56731033e-01 -1.34199286e+00 -1.03354776e+00 -1.28185904e+00
5.07118464e-01 4.95182931e-01 4.14462298e-01 -8.17725137e-02
-7.52565682e-01 7.39382982e-01 -5.64606428e-01 -3.33504885e-01
-7.48626113e-01 6.21292405e-02 3.85458440e-01 -1.05510259e+00
-2.09824815e-02 -7.10593343e-01 -4.38682079e-01 3.97173345e-01
-4.62749988e-01 9.95696932e-02 1.15411127e+00 9.11531389e-01
9.83108282e-02 -4.95185643e-01 1.40507072e-01 1.47723600e-01
4.09050882e-01 -1.29805818e-01 -9.62058425e-01 5.33654019e-02
-2.40670532e-01 8.23189020e-02 6.32641435e-01 -9.15436506e-01
-7.14105308e-01 5.72664976e-01 3.05848406e-03 -1.03567219e+00
-2.44272545e-01 6.27452694e-03 1.43734977e-01 -2.54035652e-01
5.54223835e-01 1.50059775e-01 6.05824709e-01 -4.78714675e-01
5.45657396e-01 8.91017795e-01 7.22759128e-01 -7.48608947e-01
8.84695768e-01 2.83796787e-01 4.40095738e-02 -5.98698258e-01
-1.52355116e-02 -3.07525575e-01 -1.13730991e+00 -1.91660911e-01
6.84597015e-01 -9.97816682e-01 -1.44023657e+00 6.06752038e-01
-1.25585234e+00 -1.25340343e+00 5.68047538e-02 1.04045177e+00
-1.34465384e+00 4.59019303e-01 -7.97589719e-01 -3.50525558e-01
-1.08614497e-01 -1.45426226e+00 1.16981590e+00 4.87661175e-02
-3.74676406e-01 -3.01300883e-01 1.17827365e-02 1.01011902e-01
-4.37456649e-03 1.07831270e-01 4.36456174e-01 -2.37694040e-01
-8.67253244e-01 -1.32084742e-01 -2.24216253e-01 3.51651341e-01
1.08760558e-01 -5.52766398e-02 -3.40134114e-01 -8.01587164e-01
-2.60478050e-01 -1.05636263e+00 5.10310531e-01 1.08757265e-01
9.22424614e-01 -2.36725174e-02 -4.79302555e-01 5.25989830e-01
1.02764344e+00 3.35179180e-01 4.22463864e-01 3.95103753e-01
8.69991601e-01 5.39663970e-01 1.07795095e+00 3.49092484e-01
1.78348601e-01 8.42671633e-01 5.68358481e-01 5.87122083e-01
-2.77114421e-01 -3.22821736e-01 7.65525401e-01 6.08330667e-01
-4.51606452e-01 2.08715349e-01 -8.53842378e-01 3.70950490e-01
-2.08707190e+00 -7.57669806e-01 7.24657923e-02 1.66580963e+00
7.28309035e-01 -1.69900447e-01 1.87276021e-01 -3.06671411e-01
2.48258367e-01 -2.68833131e-01 -1.07973695e+00 1.25674978e-01
4.79891717e-01 -1.03142887e-01 4.15397197e-01 5.46096385e-01
-9.04301882e-01 1.12981868e+00 5.38595772e+00 3.53242457e-01
-1.06722450e+00 -2.24372059e-01 -7.95723557e-01 -2.78893471e-01
5.88953257e-01 -3.01616117e-02 -4.82424587e-01 2.68858969e-01
2.42303401e-01 3.70495282e-02 9.09960210e-01 1.16387689e+00
5.36525398e-02 -1.07641153e-01 -1.71849644e+00 1.15162742e+00
-7.21911043e-02 -6.87046349e-01 -2.42140353e-01 -6.64660409e-02
5.45829654e-01 4.99864072e-01 -2.99660087e-01 5.38087547e-01
6.55630231e-01 -9.96379137e-01 6.42132998e-01 3.46857131e-01
7.21054792e-01 -1.46417916e-01 1.54513150e-01 8.43262315e-01
-8.38523746e-01 -4.05710965e-01 -3.07630897e-01 -3.59180570e-01
2.43743017e-01 -4.34985757e-01 -9.48521376e-01 2.95064986e-01
7.09124207e-01 9.34455276e-01 6.96118921e-02 8.90996099e-01
-8.20343912e-01 -2.02507272e-01 -2.38230005e-01 -3.70265424e-01
2.47333348e-01 -2.63706714e-01 7.27749765e-01 8.37424040e-01
2.90411353e-01 1.39259413e-01 4.77886856e-01 8.61513257e-01
-6.81418605e-05 -5.56882441e-01 -8.71964216e-01 -2.09524855e-01
2.66268700e-01 1.32292295e+00 -4.54671711e-01 -2.76801854e-01
-2.37589464e-01 1.47875345e+00 4.34404165e-01 2.73086667e-01
-9.34539676e-01 -7.87926435e-01 7.68275380e-01 -4.83484626e-01
5.83910406e-01 -1.00396991e+00 3.38521630e-01 -1.38537860e+00
3.10981572e-01 -1.21486962e+00 -2.65298307e-01 -1.20455956e+00
-9.08721507e-01 1.24740846e-01 6.31270111e-02 -1.53651536e+00
-5.71172416e-01 -1.27669847e+00 -1.99529395e-01 5.86385012e-01
-1.30383193e+00 -1.05181623e+00 -6.42994523e-01 6.22373581e-01
9.74216163e-01 -2.67514139e-01 6.97444916e-01 -4.91916575e-02
-4.60295491e-02 4.53202799e-02 1.44155428e-01 2.28359699e-01
1.04454303e+00 -1.18406308e+00 4.42923874e-01 1.73505083e-01
-2.55946934e-01 5.50280631e-01 7.15081215e-01 -6.54497683e-01
-2.26032996e+00 -6.10506535e-01 4.98942882e-02 -7.78362691e-01
9.65141594e-01 -4.03102875e-01 -4.77764636e-01 1.36040688e+00
1.96962267e-01 -1.52513608e-01 -3.05035353e-01 -4.47810680e-01
-3.27360004e-01 2.81091392e-01 -9.07381356e-01 8.76127839e-01
1.51437116e+00 -3.00642729e-01 -1.17376125e+00 8.12093616e-01
6.45244598e-01 -8.69405210e-01 -9.46580827e-01 3.10519576e-01
1.34889781e+00 -2.65387058e-01 1.00844371e+00 -7.78607488e-01
7.09395647e-01 -5.46243608e-01 -6.14587814e-02 -1.48010230e+00
-2.47255936e-01 -6.76286936e-01 -4.19351280e-01 2.57101715e-01
1.91611908e-02 -5.08944631e-01 5.76053321e-01 7.80736133e-02
-9.71105993e-02 -5.06170392e-01 -5.90475023e-01 -1.50778246e+00
1.92036733e-01 3.14984284e-02 1.90648146e-03 6.02433443e-01
4.77959484e-01 1.59274444e-01 -4.87937480e-01 2.40854502e-01
6.90329015e-01 4.62740064e-01 1.54409790e+00 -8.82787824e-01
-7.30879247e-01 -4.42292362e-01 -1.90571710e-01 -1.55295753e+00
4.97828484e-01 -8.51877749e-01 5.62861562e-01 -1.54108930e+00
2.28375822e-01 -2.42100209e-01 3.84218216e-01 6.36544943e-01
2.75215626e-01 -2.19351202e-01 6.19280756e-01 7.58299232e-01
-5.82968771e-01 5.41645110e-01 1.83969140e+00 -1.91953108e-01
-1.85453638e-01 -4.65709670e-03 2.22337320e-01 5.95522940e-01
8.45014453e-01 -2.67821491e-01 -2.08592519e-01 -6.36328816e-01
-2.79625177e-01 3.56618524e-01 7.92262971e-01 -1.01603818e+00
3.46680939e-01 -2.72503167e-01 3.72891128e-01 -4.07247096e-01
3.83537889e-01 -8.84792745e-01 1.03929467e-01 1.24393785e+00
-3.13212425e-01 1.01387724e-01 1.24865130e-01 3.67510498e-01
2.31059387e-01 -1.94940463e-01 4.67876405e-01 -4.41353858e-01
-9.89628553e-01 1.44274130e-01 -3.07994246e-01 -2.85151154e-01
1.35909092e+00 1.40963048e-01 -3.41409266e-01 -3.64043683e-01
-7.90737033e-01 7.64178932e-01 8.31174135e-01 8.62747610e-01
4.93055195e-01 -1.02656066e+00 -6.25068426e-01 -5.87978996e-02
3.62107274e-03 -6.60140580e-03 -2.99457815e-02 9.20911372e-01
-1.00573087e+00 3.82482678e-01 -8.02148700e-01 -6.79066718e-01
-1.36538064e+00 7.46270180e-01 1.96655542e-01 2.81034678e-01
-1.02194393e+00 5.09797454e-01 -4.66697626e-02 -1.07909346e+00
5.20856798e-01 -6.22838855e-01 4.12945390e-01 -6.25089228e-01
1.26429811e-01 5.40734947e-01 -5.69221258e-01 -1.47173136e-01
-1.74278334e-01 5.83930373e-01 4.04417515e-04 -7.13042393e-02
1.42660463e+00 1.64471880e-01 -8.19544867e-02 3.88887525e-01
1.03875875e+00 -5.08263648e-01 -2.00095630e+00 2.09415983e-02
-2.58043110e-01 -4.42835629e-01 -6.46484375e-01 -8.76356959e-01
-4.79314685e-01 8.35441291e-01 3.43885064e-01 -4.45330799e-01
4.68863308e-01 2.55450666e-01 6.41234457e-01 1.55618072e+00
9.01172519e-01 -1.20144379e+00 4.59208339e-01 8.14592361e-01
1.54759574e+00 -1.44245005e+00 1.14446953e-01 -2.24018320e-01
-6.18048608e-01 1.37603426e+00 7.87633002e-01 -5.33039749e-01
4.35078561e-01 5.01885474e-01 1.44918203e-01 -4.17927876e-02
-5.48231483e-01 -9.29352343e-02 -1.51317194e-01 7.83794403e-01
-1.92930833e-01 1.60340175e-01 7.22255185e-02 8.52826163e-02
-5.69653250e-02 4.34245318e-01 5.21305144e-01 1.35559261e+00
-5.11448801e-01 -9.60510790e-01 -2.59780377e-01 1.41540259e-01
2.41712198e-01 4.44985986e-01 -2.73271859e-01 1.22749448e+00
-3.39478046e-01 4.53107506e-01 -2.67642438e-01 -3.30002785e-01
5.89579642e-01 -7.70759732e-02 1.42109227e+00 -6.45751655e-01
-1.50524750e-01 -1.63825244e-01 -1.90392956e-01 -9.65885520e-01
-2.53301173e-01 -8.20876122e-01 -1.53213048e+00 -1.68464825e-01
1.37939965e-02 -4.34997350e-01 7.93728590e-01 7.89778173e-01
2.08383530e-01 5.28220654e-01 2.91628182e-01 -1.91540420e+00
-1.43219972e+00 -1.21771896e+00 -6.77398682e-01 3.19780141e-01
4.65111464e-01 -1.06220520e+00 -3.61758888e-01 -7.02589238e-03] | [4.697535514831543, 0.646843671798706] |
6078181b-7d57-4ff2-932a-e29224f34b99 | hypergraph-neural-networks-for-hypergraph | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Liao_Hypergraph_Neural_Networks_for_Hypergraph_Matching_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Liao_Hypergraph_Neural_Networks_for_Hypergraph_Matching_ICCV_2021_paper.pdf | Hypergraph Neural Networks for Hypergraph Matching | Hypergraph matching is a useful tool to find feature correspondence by considering higher-order structural information. Recently, the employment of deep learning has made great progress in the matching of graphs, suggesting its potential for hypergraphs. Hence, in this paper, we present the first, to our best knowledge, unified hypergraph neural network (HNN) solution for hypergraph matching. Specifically, given two hypergraphs to be matched, we first construct an association hypergraph over them and convert the hypergraph matching problem into a node classification problem on the association hypergraph. Then, we design a novel hypergraph neural network to effectively solve the node classification problem. Being end-to-end trainable, our proposed method, named HNN-HM, jointly learns all its components with improved optimization. For evaluation, HNN-HM is tested on various benchmarks and shows a clear advantage over state-of-the-arts. | ['Haibin Ling', 'Yong Xu', 'Xiaowei Liao'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['hypergraph-matching'] | ['graphs'] | [ 3.23635072e-01 5.21192074e-01 -4.56934094e-01 -1.69917852e-01
-6.02763116e-01 -2.97986180e-01 4.16794449e-01 1.95527956e-01
-3.56187858e-02 3.28004092e-01 -6.13736436e-02 -4.05017823e-01
-5.25126994e-01 -1.16554439e+00 -6.55766129e-01 -5.21790266e-01
-1.52644619e-01 7.67544985e-01 -8.48717541e-02 -1.55186549e-01
-7.73876160e-02 5.47762096e-01 -1.55658913e+00 8.45955387e-02
6.55596614e-01 9.27626014e-01 8.82611424e-02 3.21201354e-01
-3.82215492e-02 6.04868293e-01 -8.69450346e-02 -7.23992646e-01
5.37155747e-01 -2.77319789e-01 -1.26417017e+00 8.39678273e-02
1.02100146e+00 -1.09520227e-01 -1.08554196e+00 1.09086084e+00
4.29212332e-01 2.26215109e-01 4.19251561e-01 -1.55701602e+00
-9.89338398e-01 1.07743716e+00 -5.40918648e-01 1.01146102e-01
5.06265700e-01 1.27416268e-01 1.77321815e+00 -4.28429306e-01
5.91531277e-01 1.17873061e+00 7.00559556e-01 3.55968028e-01
-1.29118717e+00 -6.82155371e-01 -1.95508776e-03 3.56627166e-01
-1.33149469e+00 2.41599474e-02 9.08451498e-01 -1.29239604e-01
1.12027311e+00 1.98615000e-01 7.43559301e-01 8.63101959e-01
6.68364093e-02 8.27638149e-01 6.14055037e-01 -2.77153075e-01
-1.94977626e-01 -3.13768864e-01 4.09475923e-01 9.88268077e-01
4.86599416e-01 4.06955153e-01 3.79210338e-02 -1.31369159e-01
5.95230997e-01 7.79023115e-03 -3.67107451e-01 -4.91835296e-01
-8.20446610e-01 6.64306164e-01 1.24253690e+00 4.44826782e-01
-2.46796563e-01 4.15043801e-01 4.84785080e-01 4.69905376e-01
9.65684652e-02 5.39074600e-01 -3.10487151e-02 5.72471499e-01
-4.05043274e-01 1.08717307e-01 9.62808371e-01 1.09695470e+00
9.42297220e-01 -5.51375821e-02 -2.57634312e-01 6.96931899e-01
1.27005905e-01 2.78799057e-01 2.37740085e-01 -4.73324329e-01
6.00221694e-01 1.05357838e+00 -7.59942651e-01 -1.48187423e+00
-7.63232946e-01 -5.46941519e-01 -1.32160187e+00 -2.84969091e-01
1.91780254e-01 2.26916090e-01 -9.29134548e-01 1.73382497e+00
1.93843722e-01 7.03826070e-01 1.03002846e-01 8.40090632e-01
1.38151777e+00 4.76750642e-01 -4.30009030e-02 2.74985224e-01
1.18022919e+00 -1.03667569e+00 -3.61605376e-01 -1.68767408e-01
6.44072056e-01 -2.32026264e-01 7.12373793e-01 -1.16097726e-01
-1.10613573e+00 -5.72319627e-01 -1.17853618e+00 -5.73765077e-02
-3.85652363e-01 -3.94328445e-01 9.68567967e-01 4.73149389e-01
-1.28405964e+00 9.62566793e-01 -4.35368478e-01 -5.63290477e-01
3.90618205e-01 7.56620347e-01 -6.06960952e-01 -1.16089165e-01
-1.38426876e+00 6.16541624e-01 7.67998755e-01 2.05933422e-01
-5.65755785e-01 -8.09430778e-01 -1.20581353e+00 6.25776768e-01
5.14196992e-01 -1.01019216e+00 1.15283453e+00 -6.56704605e-01
-1.16950059e+00 1.12862909e+00 3.47560525e-01 -4.38062340e-01
1.79210603e-02 4.61301804e-01 -4.45967853e-01 6.13511726e-03
-1.79734930e-01 5.12445748e-01 3.76310855e-01 -1.30970311e+00
-5.19123077e-01 -4.12273675e-01 3.60699296e-01 -2.68037748e-02
-5.51545501e-01 -3.41189444e-01 -8.23429406e-01 -2.61014879e-01
2.01767296e-01 -1.09216404e+00 -2.32457608e-01 -4.02386725e-01
-7.26688504e-01 -4.45224583e-01 3.38579774e-01 -4.29677367e-01
1.32880270e+00 -1.75997210e+00 3.39457870e-01 7.32514024e-01
9.57554162e-01 3.90780360e-01 -5.36787271e-01 5.51543951e-01
-3.27292234e-01 -1.25616789e-01 -3.18806529e-01 -1.78038687e-01
3.42642576e-01 7.00073689e-02 1.00228637e-01 5.42025566e-01
-1.04340442e-01 1.50631380e+00 -8.83547962e-01 -4.24902171e-01
-1.02570718e-02 3.67715865e-01 -4.08127993e-01 2.06118777e-01
-6.41254857e-02 -2.09436119e-01 -3.19784641e-01 5.48203111e-01
8.87527049e-01 -8.70226085e-01 6.51054323e-01 -3.00356954e-01
4.94624168e-01 -1.93913318e-02 -1.01228690e+00 1.68875682e+00
-4.65213001e-01 7.52697945e-01 -1.75527707e-01 -1.06026661e+00
1.08803988e+00 1.79985419e-01 6.22664392e-01 -9.60641146e-01
2.43199751e-01 2.14764386e-01 1.36507794e-01 -3.94081324e-01
4.38345969e-01 3.34173292e-01 -9.61807463e-03 3.77555639e-01
2.74897069e-01 3.63542944e-01 2.58258611e-01 3.42638463e-01
1.39245474e+00 -4.01868284e-01 3.33865345e-01 -1.48287594e-01
6.77778661e-01 -3.91225606e-01 2.38066673e-01 8.04899693e-01
-1.22617409e-02 3.39916766e-01 5.28314352e-01 -7.15998530e-01
-8.94185901e-01 -7.63492644e-01 7.60188028e-02 7.42906809e-01
3.88368279e-01 -4.52378064e-01 -7.54122019e-01 -7.94794858e-01
4.18647259e-01 3.82483192e-02 -7.32163906e-01 -3.94823551e-01
-8.51818919e-01 -5.28681815e-01 5.47583759e-01 6.55216932e-01
3.79709780e-01 -1.38293242e+00 -6.00609742e-02 3.24122131e-01
8.16279203e-02 -1.30574822e+00 -7.10724533e-01 6.02929518e-02
-6.52926087e-01 -1.39491451e+00 -4.82052982e-01 -1.29691315e+00
6.34352982e-01 4.80718136e-01 1.64583778e+00 7.84431517e-01
-5.18140495e-01 2.46484622e-01 -1.71861336e-01 2.13139430e-01
-4.08512950e-01 7.59616494e-01 -3.97869468e-01 -8.08534771e-02
2.96626806e-01 -8.65110755e-01 -4.07326818e-01 2.53703892e-01
-8.83197784e-01 -1.28969327e-01 8.42978656e-01 7.89515674e-01
4.39318895e-01 2.19327621e-02 3.71487588e-01 -1.49316442e+00
5.23686171e-01 -4.34271395e-01 -9.62049663e-01 4.17517424e-01
-8.72217953e-01 3.70756745e-01 7.63861239e-01 -4.47208285e-02
-5.68162799e-01 1.64675698e-01 -1.19073860e-01 -4.71863538e-01
-6.40739202e-02 8.51722360e-01 -4.45748478e-01 -5.51005185e-01
4.48282510e-01 1.91689245e-02 -1.96585819e-01 -4.02621597e-01
4.41109747e-01 2.21627146e-01 9.40250218e-01 -4.33849782e-01
1.19292748e+00 1.81787148e-01 6.23502851e-01 -3.07811052e-01
-7.29591250e-01 -6.71780109e-01 -7.35322952e-01 -5.07984273e-02
4.67904568e-01 -4.57668573e-01 -1.28851032e+00 3.26686502e-01
-1.08895266e+00 -2.81171471e-01 1.69760361e-01 7.43399113e-02
-4.98353481e-01 4.30718094e-01 -5.98817050e-01 -1.97946995e-01
-6.98383808e-01 -1.03227758e+00 9.58274961e-01 2.34714821e-01
1.46397695e-01 -1.26983464e+00 3.75957519e-01 2.46709466e-01
2.43788913e-01 3.52343082e-01 1.21675122e+00 -9.06748116e-01
-1.03161263e+00 -3.68423551e-01 -7.56515265e-01 -4.05905813e-01
2.14384533e-02 -2.97598004e-01 -8.38427901e-01 -6.74969256e-01
-8.56699169e-01 -1.92048684e-01 1.12168908e+00 1.18585847e-01
1.26860106e+00 -4.18151021e-01 -7.41389453e-01 1.15863550e+00
1.74486351e+00 -3.25461514e-02 7.59250700e-01 4.67101216e-01
1.13666821e+00 4.69249576e-01 4.56272997e-03 1.53239980e-01
6.00329280e-01 7.83804476e-01 6.56061053e-01 -5.66910207e-01
-2.75783092e-01 -3.41895282e-01 -3.47644627e-01 8.05251598e-01
1.38471916e-01 -6.96189582e-01 -8.43747437e-01 3.67585421e-01
-1.96809542e+00 -7.59445012e-01 -2.11218968e-01 1.96969569e+00
3.13566417e-01 4.41576215e-03 1.15347303e-01 -2.42386181e-02
8.89819920e-01 3.21289957e-01 -5.74226439e-01 -1.11267626e-01
-4.24322709e-02 2.68533498e-01 5.58398962e-01 3.62887412e-01
-1.26165736e+00 8.79090250e-01 5.23744822e+00 6.69278443e-01
-7.44449854e-01 -3.68760586e-01 3.61923218e-01 4.97039616e-01
-5.22759616e-01 2.65853293e-02 -6.87799692e-01 1.78747892e-01
6.70405626e-01 -4.30336773e-01 7.42040217e-01 6.23685658e-01
-6.61408246e-01 6.94857538e-01 -1.49507165e+00 1.03585839e+00
1.03863567e-01 -1.46481419e+00 1.55109957e-01 8.18936154e-02
7.40631759e-01 -9.09834132e-02 1.36138290e-01 5.04184663e-01
3.90032679e-01 -1.11734653e+00 -3.03538702e-02 3.50225121e-01
6.31430030e-01 -9.19321597e-01 9.35314715e-01 -1.57073513e-01
-1.77656913e+00 -1.40776679e-01 -4.90822554e-01 4.47028905e-01
-3.27483122e-03 2.93983459e-01 -7.21153140e-01 1.15160048e+00
4.02288854e-01 7.07753956e-01 -6.64477766e-01 1.46927297e+00
-6.26516938e-02 2.25867853e-01 5.73958829e-02 1.28006473e-01
4.51212376e-01 -1.41660482e-01 4.54056770e-01 1.07071471e+00
1.85674801e-01 7.90148526e-02 5.20398855e-01 9.68871236e-01
-7.58432150e-01 1.59326106e-01 -7.09122419e-01 -1.63043588e-01
3.74873132e-01 1.51745510e+00 -7.22706914e-01 -3.28098908e-02
-4.87432212e-01 8.23968232e-01 9.45738733e-01 2.32725441e-01
-7.33379781e-01 -6.05588317e-01 4.53145295e-01 -1.43454209e-01
1.47697747e-01 4.74766761e-01 1.03230946e-01 -7.78542340e-01
-1.90304443e-02 -9.82352316e-01 9.73072350e-01 -4.26648766e-01
-1.53258038e+00 1.01402950e+00 -1.66147426e-01 -9.21796322e-01
-1.81476325e-01 -6.95064545e-01 -6.30326867e-01 6.42008901e-01
-1.52110398e+00 -1.21989834e+00 -8.13047767e-01 3.02722484e-01
-2.61769518e-02 -4.27759111e-01 7.16964722e-01 5.59062302e-01
-6.10663176e-01 1.18002748e+00 -4.59095351e-02 4.52838719e-01
2.95142829e-01 -1.44214761e+00 1.06148446e+00 5.61989903e-01
3.27853113e-01 3.36092055e-01 3.38760793e-01 -4.97043759e-01
-1.76283705e+00 -1.25279331e+00 8.96750510e-01 -1.43797165e-02
6.63313687e-01 -3.22014749e-01 -1.22653079e+00 6.57710195e-01
3.64339620e-01 3.52562554e-02 2.61452675e-01 3.24146360e-01
-6.33096814e-01 -3.01295102e-01 -1.02416468e+00 5.22379458e-01
1.31537151e+00 -5.85162163e-01 -1.21992193e-01 4.11847502e-01
1.15847421e+00 -6.50810778e-01 -1.04650247e+00 6.24375284e-01
6.26397848e-01 -6.81013763e-01 1.14127505e+00 -8.03158343e-01
3.34908903e-01 6.86080977e-02 7.26676583e-02 -1.40632427e+00
-8.34219873e-01 -7.94093430e-01 -1.84945285e-01 1.15102732e+00
5.77367187e-01 -5.73962092e-01 1.36540234e+00 5.14826596e-01
-2.21008405e-01 -8.68890941e-01 -5.03831029e-01 -8.39424610e-01
-1.26192912e-01 8.67962837e-02 1.14928436e+00 1.14381874e+00
1.23837024e-01 3.97502512e-01 -3.04049671e-01 3.17996651e-01
8.07813466e-01 5.69710255e-01 7.90388107e-01 -1.57120061e+00
-5.91296017e-01 -8.42285514e-01 -6.83716238e-01 -8.66693616e-01
9.81877506e-01 -1.63674724e+00 -1.23734161e-01 -1.51765370e+00
6.42588258e-01 -3.60704362e-01 -4.80534941e-01 5.60868144e-01
-4.11832154e-01 1.60691440e-01 2.62308449e-01 -1.88904166e-01
-5.66722572e-01 3.54944706e-01 1.35172415e+00 -6.30008042e-01
-2.50880837e-01 7.57010728e-02 -6.83321476e-01 8.41001496e-02
7.71767259e-01 -4.39812988e-01 -3.80898625e-01 -4.75586832e-01
3.50108415e-01 2.65843332e-01 3.26910406e-01 -8.82473350e-01
5.86029887e-01 1.57841906e-01 -2.27082074e-02 -5.82685947e-01
1.86524279e-02 -1.01932633e+00 3.52782637e-01 4.60482448e-01
-6.45302832e-01 4.05496299e-01 2.46278256e-01 6.33896232e-01
-1.53724849e-01 -2.79723316e-01 6.97904348e-01 -2.05822233e-02
-6.97557092e-01 9.92667258e-01 1.98028713e-01 1.29580125e-01
7.11647809e-01 -1.63507313e-01 -5.98287463e-01 -3.08262855e-01
-4.37557280e-01 5.05781651e-01 2.08377853e-01 2.69423544e-01
6.70955598e-01 -1.54597807e+00 -7.30892539e-01 2.09106296e-01
4.44684952e-01 -2.24055126e-01 2.90860206e-01 4.98115569e-01
-3.28812093e-01 7.01829255e-01 -1.29698947e-01 -3.18032771e-01
-1.32746506e+00 1.05570233e+00 4.78508532e-01 -6.95714176e-01
-9.35485780e-01 7.92650402e-01 3.62968475e-01 -7.48933554e-01
6.23263001e-01 -1.06460527e-01 -3.40321571e-01 -3.16477209e-01
2.45110154e-01 5.49848914e-01 3.03139299e-01 -6.32399857e-01
-2.12386861e-01 5.46258450e-01 -2.87213176e-01 5.76836407e-01
1.18698156e+00 2.66998857e-01 -3.37178707e-01 -3.43027174e-01
1.72033334e+00 -6.21665180e-01 -6.63521409e-01 -6.77866340e-01
2.63081223e-01 -4.23016936e-01 -1.53260054e-02 -4.01738584e-01
-1.73771513e+00 4.75873291e-01 3.35346669e-01 3.88170600e-01
1.11617577e+00 1.21071853e-01 1.24019468e+00 9.23866570e-01
1.60355508e-01 -5.28027296e-01 -6.54003024e-03 4.29635376e-01
6.88857734e-01 -1.25286615e+00 -6.57341406e-02 -6.56369746e-01
-1.30619407e-01 1.10647333e+00 9.77624416e-01 -1.66097552e-01
6.25804186e-01 7.35491738e-02 -2.22735345e-01 -6.51423156e-01
-7.11849451e-01 -5.63311934e-01 8.20505261e-01 5.87616146e-01
1.91790402e-01 1.25776917e-01 1.03653543e-01 1.90146387e-01
-1.53238654e-01 -2.32365936e-01 1.70396492e-01 3.62220049e-01
-2.05600366e-01 -1.20957410e+00 1.18095070e-01 5.97382665e-01
-1.07969992e-01 -2.08929569e-01 -6.51542604e-01 9.18950260e-01
-2.96760261e-01 6.55086637e-01 -7.03162374e-03 -6.56783164e-01
6.06347442e-01 -4.49118704e-01 5.60839951e-01 -4.06533003e-01
-8.92058849e-01 -4.75444853e-01 7.79914781e-02 -7.33912349e-01
-4.80105616e-02 -2.04579368e-01 -1.16409612e+00 -5.82649291e-01
-4.49663222e-01 7.67630488e-02 1.95230633e-01 7.84259260e-01
2.84109861e-01 6.75157785e-01 8.07508290e-01 -7.13688135e-01
-5.20550430e-01 -5.59669197e-01 -6.02657318e-01 7.26024568e-01
2.57464230e-01 -4.95720446e-01 -1.86113283e-01 -6.46913052e-01] | [7.084822177886963, 6.3451385498046875] |
8a016444-d2b3-4322-92f5-af864f9019ec | chinese-named-entity-recognition-via-adaptive | null | null | https://aclanthology.org/2020.ccl-1.86 | https://aclanthology.org/2020.ccl-1.86.pdf | Chinese Named Entity Recognition via Adaptive Multi-pass Memory Network with Hierarchical Tagging Mechanism | Named entity recognition (NER) aims to identify text spans that mention named entities and classify them into pre-defined categories. For Chinese NER task, most of the existing methods are character-based sequence labeling models and achieve great success. However, these methods usually ignore lexical knowledge, which leads to false prediction of entity boundaries. Moreover, these methods have difficulties in capturing tag dependencies. In this paper, we propose an Adaptive Multi-pass Memory Network with Hierarchical Tagging Mechanism (AMMNHT) to address all above problems. Specifically, to reduce the errors of predicting entity boundaries, we propose an adaptive multi-pass memory network to exploit lexical knowledge. In addition, we propose a hierarchical tagging layer to learn tag dependencies. Experimental results on three widely used Chinese NER datasets demonstrate that our proposed model significantly outperforms other state-of-the-art methods. | ['Jun Zhao', 'Kang Liu', 'Yubo Chen', 'Pengfei Cao'] | null | null | null | null | ccl-2020-10 | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-2.19731495e-01 -6.39679432e-02 -2.84092903e-01 -3.16772074e-01
-6.13306642e-01 -5.54386020e-01 9.10318121e-02 2.13705033e-01
-7.97406197e-01 9.58939314e-01 3.03881019e-01 -2.96193928e-01
4.22880828e-01 -9.21634793e-01 -2.16743991e-01 -2.12683469e-01
2.48443082e-01 2.19697654e-01 6.71280742e-01 1.56610936e-01
4.29602504e-01 2.59584546e-01 -7.02559054e-01 2.03264669e-01
1.28533292e+00 5.82792044e-01 3.19497555e-01 1.91030070e-01
-9.22373891e-01 8.88097942e-01 -5.98313212e-01 -5.23334026e-01
-3.40945482e-01 -4.07383561e-01 -1.06060588e+00 -3.59082520e-01
-1.76953495e-01 -1.93131138e-02 -2.33817130e-01 1.16507077e+00
4.21221286e-01 1.47920206e-01 5.01061499e-01 -6.24407351e-01
-8.67007792e-01 9.08405364e-01 -3.81314456e-01 1.96835458e-01
-3.84905934e-02 -5.30725002e-01 1.06147802e+00 -8.38626385e-01
4.29930568e-01 7.06048369e-01 9.92895126e-01 8.14804077e-01
-6.50192797e-01 -6.50670826e-01 2.77949303e-01 1.34157002e-01
-1.52694952e+00 -2.43542105e-01 3.72049809e-01 -3.44237775e-01
1.04805064e+00 -1.10751644e-01 6.25770241e-02 5.68898201e-01
5.38961031e-02 7.97341287e-01 8.95883739e-01 -5.57755172e-01
1.20088555e-01 1.05029205e-02 5.85689843e-01 7.20008373e-01
3.68924260e-01 -4.73331183e-01 -1.23664878e-01 -7.80758560e-02
6.47548258e-01 6.44060448e-02 -7.71284327e-02 1.87461600e-01
-9.59685087e-01 6.09811425e-01 2.27935523e-01 9.73063946e-01
-3.42281193e-01 -2.04346418e-01 4.34426248e-01 -3.17237198e-01
4.71463054e-01 4.32703346e-01 -8.45445514e-01 -6.93521351e-02
-8.72519910e-01 -6.32810354e-01 1.02003682e+00 1.12930644e+00
7.91001797e-01 -1.12324931e-01 -1.12875581e-01 8.97372067e-01
2.59720922e-01 1.19895697e-01 8.57479870e-01 -2.99655408e-01
5.77909589e-01 8.42539608e-01 2.32931018e-01 -8.26987147e-01
-7.39090443e-01 -3.00851256e-01 -7.71054983e-01 -5.75173497e-01
2.43090197e-01 -6.04087651e-01 -1.00677180e+00 1.48741472e+00
1.88839927e-01 4.50214326e-01 4.01389629e-01 6.63924515e-01
9.07900631e-01 8.52721870e-01 7.43676603e-01 -1.87409922e-01
1.45154488e+00 -1.09986162e+00 -9.13199365e-01 -3.63270313e-01
1.09717286e+00 -7.96689391e-01 6.41690314e-01 -1.64653480e-01
-5.82240939e-01 -5.14858246e-01 -7.44421542e-01 -2.02195887e-02
-5.06120443e-01 4.60155427e-01 5.92184246e-01 9.24286962e-01
-6.43261492e-01 4.69391912e-01 -9.63286817e-01 -3.76171231e-01
1.02902604e-02 3.42673212e-01 -2.88888931e-01 9.63407829e-02
-1.59793472e+00 8.18614662e-01 1.05860233e+00 2.92188436e-01
-3.59158635e-01 -3.94458860e-01 -8.41785014e-01 2.78874785e-01
2.96532154e-01 -1.77167863e-01 1.18766713e+00 -6.59525037e-01
-1.28081346e+00 5.75980902e-01 -4.10411775e-01 -2.55732447e-01
-1.27503738e-01 -5.01502812e-01 -8.16160858e-01 -8.00961182e-02
1.77441895e-01 3.42101753e-01 -2.56777704e-02 -9.78215754e-01
-9.29940164e-01 -1.38028964e-01 -9.89230648e-02 5.08141331e-02
-8.14632535e-01 3.22710603e-01 -4.86131370e-01 -6.00260377e-01
3.48294377e-02 -6.07168078e-01 -5.96491575e-01 -9.12471175e-01
-4.65665251e-01 -5.24345875e-01 5.46626985e-01 -8.17542195e-01
1.85886550e+00 -2.10464215e+00 -4.56675798e-01 -3.07498034e-02
9.86301750e-02 7.12400615e-01 -2.50183437e-02 4.87228245e-01
1.27042636e-01 7.14942932e-01 -1.55533984e-01 -2.34080791e-01
-1.31648198e-01 2.41739482e-01 -1.57669038e-01 -6.53882325e-03
2.78422117e-01 7.12542892e-01 -8.76224816e-01 -9.48891521e-01
-1.55145824e-01 4.24604267e-01 -1.03749536e-01 3.03356230e-01
-7.81615674e-02 4.19747382e-01 -8.26369405e-01 2.97505319e-01
7.04263866e-01 -2.42320135e-01 5.57123661e-01 -5.50649241e-02
-6.33588731e-01 5.93006372e-01 -9.47054625e-01 1.49627841e+00
-4.07481641e-01 1.16892733e-01 -4.05296654e-01 -8.84175181e-01
1.07839882e+00 5.90610743e-01 6.67765439e-02 -4.18671072e-01
1.12117521e-01 3.77181381e-01 -1.54072925e-01 -5.60792685e-01
6.32522881e-01 -2.10815743e-01 -2.93102682e-01 9.69635770e-02
4.68868995e-03 8.38872910e-01 1.65985823e-01 -8.20288733e-02
1.21188664e+00 9.01045650e-02 3.99243087e-01 8.48154351e-03
7.66263902e-01 1.57035366e-01 1.39998209e+00 5.70004880e-01
-2.71587789e-01 3.91984820e-01 3.73111635e-01 -4.18098778e-01
-8.24223936e-01 -5.81370652e-01 9.77670923e-02 1.10949039e+00
1.81424618e-01 -4.81100470e-01 -1.07439864e+00 -1.17435944e+00
-5.87044418e-01 7.89051533e-01 -3.20808858e-01 1.71469003e-01
-1.04011810e+00 -8.06395471e-01 1.09943104e+00 9.44916487e-01
8.33993733e-01 -1.29027212e+00 -1.37710050e-01 6.35690510e-01
-5.40435851e-01 -1.26890361e+00 -6.37168467e-01 2.03589588e-01
-9.04257715e-01 -8.13415349e-01 -6.98001206e-01 -1.40605295e+00
6.95394874e-01 1.44049302e-01 7.76671827e-01 3.24550390e-01
2.11402893e-01 -1.86514437e-01 -8.46824288e-01 -5.22640944e-02
-2.18210101e-01 7.49487579e-01 -1.35069326e-01 -2.63032764e-01
8.98829103e-01 -2.42203370e-01 -3.28424364e-01 4.50294554e-01
-8.04600716e-01 -6.47449344e-02 8.20445001e-01 8.00073624e-01
5.32315910e-01 2.37554237e-01 8.26360345e-01 -1.21050572e+00
2.11676568e-01 -4.83879000e-01 -4.15749639e-01 7.11911023e-01
-6.28646851e-01 2.37302810e-01 9.66304123e-01 -3.83493990e-01
-1.68256795e+00 2.32219294e-01 -4.50393498e-01 2.70124376e-01
-5.00355422e-01 8.43432605e-01 -5.96041262e-01 1.29671723e-01
8.76403153e-02 2.92941064e-01 -9.61670101e-01 -8.40952992e-01
2.04239562e-01 9.40443337e-01 3.82496059e-01 -5.41122675e-01
4.49418455e-01 -8.09591711e-02 -4.16099101e-01 -6.34473085e-01
-1.26276016e+00 -6.03914797e-01 -1.07235110e+00 2.53607750e-01
1.29634631e+00 -8.90449822e-01 -4.00527239e-01 4.92105335e-01
-1.37919486e+00 -1.62682593e-01 3.40513587e-01 7.16053665e-01
-3.49794775e-02 6.30213141e-01 -1.10078859e+00 -8.02925408e-01
-4.78036910e-01 -5.59021294e-01 5.14417768e-01 8.74030411e-01
6.26457781e-02 -1.19347072e+00 1.29762083e-01 2.04212949e-01
3.13279003e-01 -1.74658164e-01 9.41410899e-01 -1.24947011e+00
-1.98485449e-01 -1.32426426e-01 -3.81657422e-01 1.21911466e-01
1.04080684e-01 -3.82115275e-01 -5.82869768e-01 1.54845089e-01
-3.33842516e-01 3.76801565e-02 7.92311192e-01 -3.44761014e-02
6.68010056e-01 -1.90656006e-01 -6.56260610e-01 2.85847127e-01
1.43310261e+00 5.60964525e-01 6.95360363e-01 5.84382713e-01
9.51880038e-01 4.77351815e-01 7.41971374e-01 3.35417777e-01
6.73623383e-01 1.92428291e-01 -1.06476322e-01 -9.89146307e-02
2.53001243e-01 -4.70942408e-01 1.26859590e-01 1.27168286e+00
6.04256503e-02 -4.07179207e-01 -1.19730067e+00 6.94380879e-01
-1.91868174e+00 -7.38463998e-01 -3.18495870e-01 1.91029465e+00
1.05349672e+00 2.32767314e-01 -2.47663617e-01 -2.78226465e-01
1.34177399e+00 -1.01270340e-01 -2.43499666e-01 -1.96550235e-01
-1.06399201e-01 2.27725223e-01 6.96226239e-01 1.33960024e-01
-1.39760172e+00 1.50769329e+00 5.63424301e+00 8.93506110e-01
-8.73460948e-01 3.59037101e-01 3.20762962e-01 9.03223634e-01
-1.11974828e-01 3.52548808e-01 -1.51646042e+00 3.84692550e-01
1.10576189e+00 -2.68218610e-02 -2.86229283e-01 8.06319475e-01
-1.54648097e-02 2.28919595e-01 -5.28303802e-01 4.74806309e-01
-9.16568860e-02 -1.00042474e+00 -1.63308531e-01 -1.98487043e-01
6.26916945e-01 -5.23916818e-02 -6.26968920e-01 5.09705842e-01
5.46409369e-01 -6.21548653e-01 3.97553742e-01 6.17542684e-01
5.23752272e-01 -8.65632653e-01 1.13951015e+00 5.50218940e-01
-1.58160686e+00 3.76974009e-02 -6.65938318e-01 2.10789423e-02
4.61590052e-01 6.17947459e-01 -5.30534983e-01 7.17688859e-01
4.88444030e-01 4.78551894e-01 -4.03238952e-01 1.38154411e+00
-6.23970866e-01 1.04135656e+00 -9.75155309e-02 -3.22656304e-01
3.30710024e-01 5.15787452e-02 1.70311786e-03 1.65929925e+00
3.90715390e-01 5.82913280e-01 3.84346724e-01 3.44569951e-01
-3.13186795e-01 5.24784327e-01 -7.71928802e-02 -3.73289824e-01
8.09269249e-01 1.28851831e+00 -1.06097662e+00 -4.73900467e-01
-8.17928791e-01 1.11334455e+00 7.78684020e-01 8.89414921e-02
-7.69243658e-01 -1.05432808e+00 2.30757102e-01 -3.12106282e-01
5.55412412e-01 -5.20186484e-01 -2.83391923e-01 -1.41280818e+00
-3.05116951e-01 -2.63938963e-01 6.66321516e-01 -4.13344920e-01
-1.30868173e+00 6.87489390e-01 -5.63930452e-01 -9.98897493e-01
1.98121756e-01 -3.66032690e-01 -6.86942756e-01 7.62238383e-01
-1.58286226e+00 -1.10850358e+00 1.04655996e-02 2.01193422e-01
4.45754409e-01 1.59544408e-01 9.89596725e-01 6.71329737e-01
-1.00336969e+00 6.95320606e-01 2.01398984e-01 1.08065784e+00
8.33725274e-01 -1.12060416e+00 6.51634991e-01 1.08583736e+00
8.66087377e-02 9.34066594e-01 1.38513088e-01 -1.02244818e+00
-6.65201843e-01 -1.29424453e+00 1.59818721e+00 -3.94784771e-02
6.21392190e-01 -1.53643847e-01 -1.17493451e+00 8.25937629e-01
2.07668528e-01 -1.56062260e-01 1.15182471e+00 1.54670492e-01
-3.99165839e-01 1.98693529e-01 -7.97556698e-01 3.35556269e-01
9.43668485e-01 -2.61245489e-01 -9.28086936e-01 -8.22804496e-02
8.30432832e-01 -1.74906746e-01 -1.03859401e+00 2.98029721e-01
4.33084458e-01 -5.69519877e-01 3.97223026e-01 -6.64469302e-01
9.05493181e-03 -4.48442936e-01 6.91708475e-02 -1.11589134e+00
-5.69948673e-01 -2.25789502e-01 1.25454843e-01 2.05271816e+00
6.83495522e-01 -6.41883135e-01 7.84689307e-01 7.60802329e-01
-3.43449861e-01 -2.96031237e-01 -6.49325907e-01 -8.82503390e-01
1.59020081e-01 1.56851262e-02 5.83178282e-01 1.32128978e+00
2.79484242e-01 6.69613361e-01 -4.59587425e-01 4.09578472e-01
2.89675593e-01 -7.56936222e-02 1.41310140e-01 -1.30428219e+00
1.56967565e-01 -1.76556617e-01 -2.12051526e-01 -1.24679887e+00
5.09899914e-01 -5.23306966e-01 4.26989347e-01 -1.74125814e+00
3.24537545e-01 -8.38403165e-01 -5.40394902e-01 8.64869356e-01
-6.07563615e-01 -6.35332838e-02 -9.39964782e-03 4.36282307e-01
-1.07337129e+00 2.92601019e-01 5.67193627e-01 2.52468824e-01
-3.23880523e-01 -2.60670513e-01 -5.57165980e-01 8.02877367e-01
1.12489295e+00 -8.65423858e-01 2.89793968e-01 -7.35518396e-01
1.71971872e-01 1.22130616e-02 -3.70023489e-01 -1.04834521e+00
5.18540859e-01 -2.99944878e-01 4.95312363e-01 -6.58348680e-01
-4.16943312e-01 -5.98389447e-01 -4.37013321e-02 3.83277118e-01
-3.02147239e-01 5.25780246e-02 -2.55690329e-03 3.93346906e-01
-2.96678454e-01 -7.48332560e-01 6.06150985e-01 -2.04764739e-01
-1.14624548e+00 2.02995956e-01 -5.94272196e-01 2.76840061e-01
8.19580793e-01 2.03063697e-01 -4.58391070e-01 2.07237959e-01
-6.55373454e-01 4.44836974e-01 2.19936416e-01 3.31563652e-01
3.04992288e-01 -1.17451823e+00 -4.46368039e-01 -2.24814162e-01
3.26767452e-02 -1.54466465e-01 3.42312664e-01 5.33565938e-01
-5.27355254e-01 6.26984954e-01 -7.90659105e-05 1.47874504e-01
-1.03044331e+00 6.50701582e-01 1.23950891e-01 -7.47167408e-01
-4.03184384e-01 6.52090311e-01 1.40671581e-01 -6.13251090e-01
-3.64219062e-02 1.73637092e-01 -8.50492120e-01 -4.26591299e-02
6.18824363e-01 2.32912764e-01 -8.92913342e-02 -7.78184056e-01
-5.33324182e-01 5.26835382e-01 -2.50773340e-01 1.97632551e-01
1.09238362e+00 -3.58257592e-01 -2.83353150e-01 2.39210352e-01
8.93642366e-01 2.43537247e-01 -7.41911054e-01 -4.45907325e-01
8.68449926e-01 9.65227485e-02 -1.63739011e-01 -6.68271244e-01
-8.81525815e-01 1.00876057e+00 2.15230644e-01 1.44703746e-01
9.88558710e-01 -1.83851555e-01 1.40788198e+00 5.70541859e-01
4.32897449e-01 -1.16127896e+00 -4.83640701e-01 1.08243299e+00
-1.99943513e-01 -9.56714571e-01 -4.82263714e-01 -5.63824892e-01
-5.78158677e-01 1.21641660e+00 1.06589019e+00 1.22495919e-01
5.54069698e-01 2.87678719e-01 3.61274719e-01 3.44559312e-01
-4.22359049e-01 -5.68791986e-01 -5.50686195e-03 4.72177923e-01
8.70055616e-01 -1.05526865e-01 -1.11065078e+00 1.11239314e+00
3.45941156e-01 1.68385841e-02 4.20595497e-01 1.03966463e+00
-8.53364050e-01 -1.49386907e+00 -2.12613538e-01 1.76995978e-01
-1.12233388e+00 -5.21052480e-01 -3.09949279e-01 5.72121739e-01
2.18141168e-01 1.18417370e+00 -8.44960064e-02 -4.69848633e-01
3.02395999e-01 3.26904625e-01 -9.56935212e-02 -8.19727361e-01
-6.83840692e-01 1.78876072e-01 3.15069705e-01 4.84004878e-02
-5.41496575e-01 -3.64905715e-01 -1.88081992e+00 6.59346431e-02
-7.91803539e-01 9.53857601e-01 3.97906482e-01 1.10124207e+00
4.15549695e-01 2.54736423e-01 4.89483058e-01 -1.42694548e-01
-1.81975067e-01 -1.03767920e+00 -5.76737225e-01 1.31535605e-01
-3.15888703e-01 -4.36134249e-01 -2.15751901e-02 1.02832973e-01] | [9.757789611816406, 9.723267555236816] |
3cb61058-a3db-4b9c-8261-93c150db471a | nmt5-is-parallel-data-still-relevant-for-pre | 2106.02171 | null | https://arxiv.org/abs/2106.02171v1 | https://arxiv.org/pdf/2106.02171v1.pdf | nmT5 -- Is parallel data still relevant for pre-training massively multilingual language models? | Recently, mT5 - a massively multilingual version of T5 - leveraged a unified text-to-text format to attain state-of-the-art results on a wide variety of multilingual NLP tasks. In this paper, we investigate the impact of incorporating parallel data into mT5 pre-training. We find that multi-tasking language modeling with objectives such as machine translation during pre-training is a straightforward way to improve performance on downstream multilingual and cross-lingual tasks. However, the gains start to diminish as the model capacity increases, suggesting that parallel data might not be as essential for larger models. At the same time, even at larger model sizes, we find that pre-training with parallel data still provides benefits in the limited labelled data regime. | ['Linting Xue', 'Rami Al-Rfou', 'Melvin Johnson', 'Noah Constant', 'Aditya Siddhant', 'Mihir Kale'] | 2021-06-03 | null | null | null | null | ['multilingual-nlp'] | ['natural-language-processing'] | [-2.10958764e-01 3.40582654e-02 -5.88839889e-01 -4.14498806e-01
-1.55687332e+00 -8.67991269e-01 6.87810779e-01 3.23776118e-02
-6.15079284e-01 8.90803337e-01 3.77038956e-01 -9.54038978e-01
2.53852993e-01 -1.70986623e-01 -8.35861981e-01 -1.47403508e-01
2.18825623e-01 7.99426377e-01 -1.92926794e-01 -4.07980293e-01
-3.20271015e-01 5.00869341e-02 -6.51638448e-01 5.32843411e-01
1.05864072e+00 1.73389226e-01 3.92317832e-01 4.80870336e-01
-2.44620055e-01 4.85677063e-01 -1.44796595e-01 -7.16587186e-01
3.72971952e-01 -1.58524200e-01 -9.48027968e-01 -1.51646227e-01
4.74281788e-01 -9.20951813e-02 -2.69521952e-01 7.10018933e-01
5.16979694e-01 -1.49409413e-01 4.98787493e-01 -9.14498925e-01
-6.22496009e-01 9.99402046e-01 -5.47310174e-01 1.88470319e-01
3.58065516e-02 2.36448914e-01 1.31858039e+00 -1.12156487e+00
7.70634055e-01 1.37886310e+00 7.27077484e-01 2.42851108e-01
-1.48220015e+00 -5.95854938e-01 1.27816856e-01 -8.47104043e-02
-1.13902116e+00 -8.39450657e-01 2.23151639e-01 -4.07668203e-01
1.51105142e+00 -1.04432464e-01 6.96992874e-02 1.02924800e+00
5.31377494e-01 8.21854770e-01 1.30093694e+00 -7.85597444e-01
-6.23415828e-01 1.27526402e-01 -1.13319919e-01 5.18029392e-01
3.84138227e-02 -4.47575599e-02 -4.55209434e-01 -1.02356531e-01
3.01507175e-01 -5.01922369e-01 -1.40111623e-02 -2.72667464e-02
-1.50699353e+00 8.62314522e-01 7.76903778e-02 5.62769771e-01
-1.34225696e-01 7.52941817e-02 5.85038960e-01 8.20191026e-01
1.10926557e+00 5.96116841e-01 -1.11477172e+00 -2.25659639e-01
-1.00460267e+00 -1.28649492e-02 7.45005965e-01 8.27770412e-01
8.27213466e-01 -6.37747068e-03 3.08569334e-02 1.11119461e+00
7.89927244e-02 6.42199993e-01 3.92825603e-01 -7.97936440e-01
1.24763715e+00 3.60911936e-01 -9.08938274e-02 -1.05806351e-01
-3.92739952e-01 -6.01117492e-01 -4.06489998e-01 -2.29986906e-01
6.99200869e-01 -6.31343484e-01 -8.42137694e-01 1.90116191e+00
8.20261240e-02 -3.42206866e-01 2.91348040e-01 5.75781167e-01
1.85402006e-01 8.13140869e-01 2.47022107e-01 -1.31939262e-01
1.14699149e+00 -1.28778911e+00 -5.85669041e-01 -7.74014831e-01
1.50446081e+00 -1.39275312e+00 1.34706974e+00 7.68098012e-02
-1.03478074e+00 -5.16669571e-01 -7.20815539e-01 -4.37212527e-01
-3.46566290e-01 2.66688675e-01 5.77347517e-01 4.42791045e-01
-1.20908332e+00 4.59939301e-01 -8.23938489e-01 -5.77378988e-01
-1.40250206e-01 1.44368336e-01 -5.12837887e-01 -5.91426253e-01
-1.28944707e+00 1.50414062e+00 3.86999279e-01 -1.72128066e-01
-8.06473076e-01 -7.06687808e-01 -8.53404403e-01 -1.17047466e-01
1.57874212e-01 -4.85646158e-01 1.53428066e+00 -7.87627578e-01
-1.24998248e+00 8.98743629e-01 -4.31880951e-01 -3.05049360e-01
5.66586375e-01 -4.50058073e-01 -3.76802832e-01 -5.57700574e-01
2.66837299e-01 8.67227852e-01 3.34293634e-01 -1.05204642e+00
-7.57755160e-01 -2.47604221e-01 -2.17991918e-01 6.90602899e-01
-4.29150641e-01 4.19893086e-01 -4.17324930e-01 -5.20047665e-01
-2.11411864e-01 -1.20025671e+00 -3.19455266e-01 -6.98327780e-01
-1.41694218e-01 -3.03904116e-01 4.99785334e-01 -1.09744406e+00
1.11660409e+00 -1.87803006e+00 1.79862380e-01 -2.36372665e-01
-9.73746553e-02 2.31656045e-01 -4.81264114e-01 7.69235790e-01
7.58624524e-02 3.76366585e-01 8.40916634e-02 -7.71362782e-01
-7.19819516e-02 3.03817749e-01 -1.11603215e-01 3.10560167e-01
2.53885984e-01 1.20994294e+00 -6.77791297e-01 -5.62388837e-01
7.99101591e-02 2.23695844e-01 -3.26933771e-01 -2.10590795e-01
-3.79667461e-01 7.52438068e-01 -2.16372818e-01 4.53628063e-01
3.40789437e-01 -2.01403335e-01 4.37577069e-01 2.47015402e-01
-2.37826422e-01 7.60697186e-01 -4.66030657e-01 2.03270674e+00
-9.28659916e-01 7.71145403e-01 1.94621176e-01 -7.15002656e-01
4.68031883e-01 5.45845687e-01 3.24628919e-01 -8.95949543e-01
-8.68608877e-02 7.15033174e-01 5.12567997e-01 -3.15261334e-01
6.60343349e-01 -3.18768680e-01 -2.34855264e-01 6.71617210e-01
2.28520319e-01 -1.12089299e-01 2.50942707e-01 7.05853924e-02
7.14085460e-01 3.35445762e-01 2.47941971e-01 -5.40280700e-01
1.74107492e-01 4.29164231e-01 4.87227261e-01 4.90711212e-01
7.36304522e-02 1.24055006e-01 1.45544350e-01 -1.74454689e-01
-1.39069819e+00 -8.80481720e-01 -2.24578053e-01 1.58552790e+00
-6.06931090e-01 -3.95190805e-01 -5.75112224e-01 -8.28648925e-01
4.39092405e-02 7.88973629e-01 2.48009572e-03 2.37944558e-01
-8.51216137e-01 -9.04157460e-01 7.96791732e-01 2.04647154e-01
-1.03981659e-01 -7.57109046e-01 2.77554810e-01 5.39644480e-01
-5.50423026e-01 -1.48970258e+00 -6.51706040e-01 4.88319308e-01
-9.86173272e-01 -5.34856677e-01 -8.43590617e-01 -9.88692939e-01
3.33157212e-01 2.70489156e-01 1.27746296e+00 -3.37521732e-01
3.47068250e-01 -7.29845315e-02 -2.42800742e-01 -2.34221652e-01
-8.93328369e-01 7.39043832e-01 4.75071147e-02 -4.93457735e-01
2.58207232e-01 -3.82677048e-01 3.96087300e-03 2.22260803e-01
-4.18141454e-01 2.58425832e-01 8.34185958e-01 6.95967853e-01
2.85651326e-01 -4.47923362e-01 7.82017529e-01 -8.55555534e-01
6.58791423e-01 -4.16899383e-01 -2.85720527e-01 5.41556656e-01
-7.22945809e-01 3.50690126e-01 7.30729103e-01 -4.81269270e-01
-1.10036910e+00 -2.78139085e-01 -1.06137179e-01 -1.55298546e-01
6.54781908e-02 9.33415055e-01 -5.24068996e-02 2.90295780e-02
6.33487046e-01 2.34948732e-02 -2.28516489e-01 -8.61426950e-01
7.39683330e-01 7.90325701e-01 1.56964213e-01 -8.14784169e-01
5.70907712e-01 -2.61440165e-02 -3.02720785e-01 -4.61889267e-01
-8.96546543e-01 -4.89377975e-01 -7.69548297e-01 2.12043971e-01
7.20367253e-01 -1.44465220e+00 5.79523928e-02 3.05689424e-01
-1.16904128e+00 -9.29622352e-01 1.95733115e-01 7.80951619e-01
-2.92483777e-01 2.63125420e-01 -1.19844711e+00 -3.43394369e-01
-2.95533717e-01 -1.38612807e+00 1.03283155e+00 -4.76371616e-01
-1.49451971e-01 -1.35053945e+00 2.07611367e-01 6.62073016e-01
4.23621267e-01 -3.83351117e-01 1.28449285e+00 -8.03757191e-01
-3.66703928e-01 1.27581999e-01 -1.67278722e-01 2.96904504e-01
3.02665740e-01 -3.70490372e-01 -8.03893864e-01 -5.26533008e-01
-2.64934033e-01 -6.23446524e-01 6.44382000e-01 2.42841646e-01
3.56996030e-01 -1.88176632e-01 -3.78591537e-01 4.67962265e-01
1.35275149e+00 -7.60868415e-02 2.28391796e-01 3.55796218e-01
8.70149791e-01 6.58353865e-01 6.28583610e-01 -2.29661062e-01
8.65463555e-01 7.53637195e-01 -2.68761933e-01 -4.01665300e-01
-3.13616276e-01 -3.93479526e-01 6.88357055e-01 1.49341488e+00
4.58894819e-01 -3.06803823e-01 -1.44390345e+00 7.60434270e-01
-1.74964464e+00 -3.02272469e-01 -2.89033473e-01 1.96281683e+00
1.30880189e+00 1.94638208e-01 -4.67564538e-02 -5.61519980e-01
5.67416310e-01 2.46330574e-02 -3.00295681e-01 -5.29908895e-01
-2.75314748e-01 7.35200942e-02 8.02795529e-01 8.84637058e-01
-8.19221318e-01 1.68899941e+00 6.76961756e+00 9.88721251e-01
-1.41752470e+00 7.70075738e-01 7.66225994e-01 -1.27359048e-01
-4.03551400e-01 3.67070198e-01 -1.11801493e+00 3.20020735e-01
1.46953988e+00 -1.91949248e-01 4.22794640e-01 4.56974119e-01
4.57010657e-01 -1.51199223e-02 -1.19431520e+00 5.11679471e-01
-1.57561451e-01 -9.88897622e-01 8.08281302e-02 3.78204763e-01
1.02240443e+00 1.02136874e+00 -9.34717730e-02 6.39212906e-01
7.71938980e-01 -1.00518441e+00 6.85325801e-01 -2.59726733e-01
1.08357406e+00 -5.94049871e-01 6.19926274e-01 7.44529188e-01
-1.01048875e+00 2.05492228e-01 -3.08466583e-01 -7.60041699e-02
5.96353233e-01 6.23388112e-01 -1.03409791e+00 8.39756310e-01
1.78051025e-01 5.04253745e-01 -6.96954548e-01 4.92103428e-01
-1.06099106e-01 8.11345279e-01 -4.48808670e-01 4.28034216e-01
6.21131480e-01 -9.73047316e-02 2.56297946e-01 1.55364692e+00
4.71193165e-01 -6.26312733e-01 4.78857040e-01 1.16891719e-01
-3.63644212e-01 4.88442272e-01 -8.82579148e-01 -1.52752370e-01
3.96236569e-01 1.12554646e+00 -2.94965684e-01 -4.08683211e-01
-8.10314298e-01 8.00988197e-01 8.01666737e-01 4.16088194e-01
-5.64406753e-01 1.32763952e-01 4.30733025e-01 -1.70590659e-03
-5.33904247e-02 -7.70006478e-01 -4.30308908e-01 -1.39969051e+00
1.13589708e-02 -1.28163075e+00 3.34056020e-01 -6.26203179e-01
-1.07773876e+00 5.29166520e-01 -6.25458360e-02 -8.59946609e-01
-5.49731314e-01 -6.02472961e-01 -1.89121932e-01 1.23259544e+00
-1.76908183e+00 -1.65057445e+00 6.45103633e-01 4.69916165e-01
8.15350056e-01 -1.22830190e-01 6.99146867e-01 5.56751847e-01
-3.95458043e-01 7.44911015e-01 5.95364094e-01 6.01663478e-02
1.31322730e+00 -1.05960369e+00 8.48614335e-01 1.04144442e+00
5.56532979e-01 5.63292384e-01 4.55436289e-01 -7.77074993e-01
-1.29486346e+00 -1.10940814e+00 1.60037494e+00 -7.54026890e-01
1.04225564e+00 -6.45439267e-01 -7.99379170e-01 1.33175564e+00
5.85684955e-01 -3.43755841e-01 5.53067923e-01 7.26351082e-01
-2.59782970e-01 1.16138145e-01 -5.23017228e-01 7.36100376e-01
9.71162617e-01 -8.78676713e-01 -5.24354696e-01 6.02859855e-01
1.00416064e+00 -3.08561087e-01 -1.18181264e+00 4.48956579e-01
3.29727232e-01 -2.13144064e-01 7.06007838e-01 -7.11181104e-01
3.86194199e-01 1.49587154e-01 -3.04369628e-01 -1.82595825e+00
-2.42292479e-01 -6.99660897e-01 3.50979388e-01 1.29024220e+00
1.17992711e+00 -8.11783433e-01 3.85844320e-01 1.69489101e-01
-4.66775984e-01 -5.87770164e-01 -1.05566657e+00 -9.78589237e-01
8.57526183e-01 -4.43478674e-01 2.51855016e-01 1.32941544e+00
2.48469159e-01 9.45166945e-01 -4.74343926e-01 -4.68009561e-02
5.05978227e-01 -2.05412149e-01 7.30642915e-01 -9.49670136e-01
-3.19719374e-01 -2.72378981e-01 3.36996377e-01 -1.10543704e+00
5.64195156e-01 -1.44578815e+00 -1.11177191e-02 -1.57685304e+00
1.68032140e-01 -7.61608362e-01 -1.40744880e-01 7.06724226e-01
-4.02833998e-01 1.05177406e-02 3.57394218e-01 5.03827870e-01
-3.56213391e-01 2.84655243e-01 1.35335326e+00 1.66959852e-01
-1.10847138e-01 -3.17403018e-01 -6.38677239e-01 3.29207361e-01
8.44708264e-01 -6.83135688e-01 -3.12425762e-01 -1.34659266e+00
6.57477677e-02 2.20853955e-01 -2.81969041e-01 -5.76903820e-01
9.15705413e-02 -1.76477075e-01 7.28953332e-02 -4.30820942e-01
3.13965797e-01 -4.66683537e-01 -4.20123339e-02 3.54820698e-01
-3.16719592e-01 6.45980000e-01 5.04156411e-01 7.59517848e-02
-2.56955206e-01 1.50039550e-02 5.65281391e-01 -1.78355962e-01
-4.65047419e-01 1.01265989e-01 -3.99001598e-01 3.55600029e-01
5.27916551e-01 3.57259154e-01 -5.12813807e-01 -2.73574740e-01
-5.33986151e-01 5.49297631e-01 5.36146820e-01 6.98156834e-01
-3.41349304e-01 -1.14320028e+00 -1.16633534e+00 -1.55120818e-02
1.67126253e-01 -2.48733327e-01 -1.15943171e-01 1.01235271e+00
-2.43811458e-01 1.02454245e+00 1.70167595e-01 -6.07290149e-01
-1.03579676e+00 2.11274356e-01 3.00831616e-01 -9.10349011e-01
-3.92332226e-01 5.35959840e-01 2.47436002e-01 -1.04015279e+00
-1.54572800e-01 -3.40370655e-01 3.62434804e-01 -1.17527500e-01
-5.58000989e-02 4.73497435e-02 3.90397131e-01 -7.32841611e-01
-2.47744173e-01 4.46115792e-01 -3.93460602e-01 -7.29558289e-01
1.25010931e+00 -4.50670630e-01 9.54996720e-02 6.11778140e-01
1.21292949e+00 2.69405186e-01 -1.05361342e+00 -6.08591914e-01
3.86368811e-01 -5.37856705e-02 1.31685305e-02 -1.10728419e+00
-5.81543922e-01 1.05166662e+00 6.87018782e-02 -3.73021483e-01
5.06043315e-01 1.66468173e-01 9.71330285e-01 5.83049119e-01
5.49767792e-01 -1.16524768e+00 -4.16957438e-01 1.01106143e+00
5.84957719e-01 -1.50762975e+00 -1.48335189e-01 -8.51974115e-02
-7.42832839e-01 7.33013213e-01 4.34793770e-01 4.45282459e-01
4.55000013e-01 4.43564087e-01 4.06271368e-01 1.05255902e-01
-1.10848367e+00 -1.35583401e-01 2.60667026e-01 1.96426004e-01
9.29561436e-01 2.92301774e-01 -3.23612452e-01 1.42110318e-01
-3.48317862e-01 -2.09971428e-01 2.15252608e-01 7.59097278e-01
-3.45557511e-01 -1.60541093e+00 -2.16040522e-01 3.17514449e-01
-9.15647686e-01 -7.73600698e-01 -2.29634613e-01 9.65324581e-01
-1.13624737e-01 1.09950292e+00 -1.46126762e-01 -1.47152215e-01
7.45099932e-02 6.23110294e-01 5.42921662e-01 -8.48690629e-01
-7.76877940e-01 4.04009908e-01 6.77819788e-01 -2.45055586e-01
-1.07467674e-01 -6.41202450e-01 -8.39624763e-01 -5.17868698e-01
-3.67348343e-01 2.02476621e-01 8.84568810e-01 1.21436918e+00
4.35703665e-01 1.02306411e-01 3.48973453e-01 -6.14434659e-01
-6.93809927e-01 -1.44636214e+00 -1.27433315e-02 6.80343360e-02
2.46403262e-01 -1.97144270e-01 -2.51195490e-01 4.53460366e-02] | [11.334486961364746, 10.205280303955078] |
fd9a02fe-ce03-4a50-a145-3061cc0a43df | syngen-a-syntactic-plug-and-play-module-for | 2302.13032 | null | https://arxiv.org/abs/2302.13032v1 | https://arxiv.org/pdf/2302.13032v1.pdf | SynGen: A Syntactic Plug-and-play Module for Generative Aspect-based Sentiment Analysis | Aspect-based Sentiment Analysis (ABSA) is a sentiment analysis task at fine-grained level. Recently, generative frameworks have attracted increasing attention in ABSA due to their ability to unify subtasks and their continuity to upstream pre-training tasks. However, these generative models suffer from the neighboring dependency problem that induces neighboring words to get higher attention. In this paper, we propose SynGen, a plug-and-play syntactic information aware module. As a plug-in module, our SynGen can be easily applied to any generative framework backbones. The key insight of our module is to add syntactic inductive bias to attention assignment and thus direct attention to the correct target words. To the best of our knowledge, we are the first one to introduce syntactic information to generative ABSA frameworks. Our module design is based on two main principles: (1) maintaining the structural integrity of backbone PLMs and (2) disentangling the added syntactic information and original semantic information. Empirical results on four popular ABSA datasets demonstrate that SynGen enhanced model achieves a comparable performance to the state-of-the-art model with relaxed labeling specification and less training consumption. | ['Yujiu Yang', 'Xingyu Bai', 'Jiayi Li', 'Taiqiang Wu', 'Chengze Yu'] | 2023-02-25 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-1.43914018e-02 2.61537611e-01 -9.40310583e-02 -6.47333443e-01
-7.51876891e-01 -5.30633271e-01 5.52488267e-01 -1.83765799e-01
1.38546675e-02 5.00517607e-01 5.50971329e-01 -3.09073180e-01
3.11125219e-02 -1.07329047e+00 -7.58958578e-01 -5.90487480e-01
4.54001546e-01 3.63726884e-01 8.87808576e-02 -5.20502388e-01
-2.48555522e-02 -1.83063578e-02 -1.26996160e+00 4.98783767e-01
9.24550712e-01 7.45672703e-01 2.52185285e-01 1.20830022e-01
-5.17597616e-01 7.27385521e-01 -3.94885898e-01 -8.64159226e-01
2.10972549e-03 -5.87040246e-01 -8.12184751e-01 3.16753983e-02
3.87147143e-02 1.11052066e-01 1.04919143e-01 8.29750717e-01
4.01938111e-01 -6.17092885e-02 4.37101662e-01 -1.29843915e+00
-1.02208352e+00 1.20644367e+00 -6.10315859e-01 -1.67594235e-02
1.84648465e-02 -3.02511062e-02 1.62578678e+00 -1.05591452e+00
3.81024092e-01 1.30227172e+00 5.80282569e-01 5.31129420e-01
-1.00479031e+00 -6.64565146e-01 8.46228361e-01 -2.56259032e-02
-9.95751321e-01 -1.74842671e-01 8.99195731e-01 -3.29800874e-01
1.21765554e+00 1.02663152e-01 6.84636831e-01 1.15831149e+00
1.56498387e-01 1.07813835e+00 9.67098117e-01 -3.13427806e-01
2.16238648e-01 4.75929268e-02 4.39479321e-01 5.90001702e-01
3.08539450e-01 -3.79427820e-01 -5.81236124e-01 1.07056618e-01
2.66712040e-01 7.63816461e-02 1.05931275e-02 -1.96499571e-01
-8.95470858e-01 1.03520572e+00 4.36555177e-01 3.00386131e-01
-3.08184773e-01 2.25997865e-01 2.73553103e-01 1.01628646e-01
4.92272913e-01 5.74971080e-01 -7.14735925e-01 -4.74264324e-02
-5.76463342e-01 2.07367748e-01 7.77771711e-01 1.10553837e+00
8.36501300e-01 2.84465998e-01 -3.33498865e-01 6.33541286e-01
5.56498587e-01 3.82737249e-01 7.64537275e-01 -3.10974002e-01
4.31007415e-01 9.94111776e-01 -5.00393033e-01 -7.27073908e-01
-2.82597393e-01 -8.78888190e-01 -5.00132620e-01 -1.95904225e-01
-1.05621092e-01 -3.17255080e-01 -9.17156160e-01 1.89706421e+00
1.64993063e-01 -1.54932234e-02 7.67980441e-02 6.15482152e-01
8.15825880e-01 6.15389168e-01 2.07681656e-01 1.19491823e-01
1.64140582e+00 -1.35967195e+00 -5.97877741e-01 -6.38408184e-01
6.39272153e-01 -8.09633434e-01 1.27394700e+00 1.78120956e-01
-9.17509675e-01 -5.30423641e-01 -1.13168883e+00 -1.12925611e-01
-5.35772860e-01 4.92526889e-02 9.83857989e-01 7.16803968e-01
-9.66042459e-01 1.87017053e-01 -7.76969016e-01 -3.20365071e-01
4.36594158e-01 2.10143715e-01 -5.74743375e-02 5.92261292e-02
-1.23179185e+00 5.25175631e-01 3.20509255e-01 1.04900338e-01
-6.10714555e-01 -9.41366017e-01 -1.07740128e+00 2.01943725e-01
6.10784650e-01 -1.17146325e+00 1.26293278e+00 -9.71813679e-01
-1.59684598e+00 6.00621939e-01 -3.49555939e-01 -3.25198442e-01
-7.50029385e-02 -5.51276684e-01 -1.99414268e-01 -5.00062525e-01
3.64148945e-01 5.16679049e-01 6.27810180e-01 -1.14129686e+00
-5.83275855e-01 -5.00998557e-01 3.48617047e-01 2.63096392e-01
-4.43788856e-01 -3.34930532e-02 -4.54903364e-01 -8.43190312e-01
-7.78587982e-02 -8.54367375e-01 -1.87822655e-01 -8.89887035e-01
-5.55047095e-01 -2.86526054e-01 6.75785244e-01 -2.95940399e-01
1.43835485e+00 -2.09915853e+00 2.63928056e-01 -3.88191193e-02
1.23157009e-01 1.06538571e-01 -1.87854350e-01 4.54983294e-01
-1.29828289e-01 2.67559022e-01 -1.44063696e-01 -6.64612055e-01
3.50091875e-01 1.94446251e-01 -5.69658816e-01 -1.57617956e-01
3.35885108e-01 1.42896521e+00 -8.38998318e-01 -2.06716046e-01
-1.51925778e-03 4.60772306e-01 -8.66984367e-01 1.62688658e-01
-4.11031246e-01 5.11944406e-02 -7.04437554e-01 5.94497919e-01
4.20808762e-01 -2.61908144e-01 1.52668908e-01 -3.78153056e-01
-2.76886411e-02 6.92173302e-01 -8.30848157e-01 1.80929148e+00
-7.81301737e-01 2.01581538e-01 -1.47613034e-01 -8.26016307e-01
9.65940773e-01 1.00241654e-01 1.45309791e-01 -5.95031023e-01
3.03499132e-01 6.21841326e-02 8.49660188e-02 -1.64346755e-01
6.65130019e-01 -3.56052428e-01 -5.05318940e-01 5.56881607e-01
3.03094953e-01 2.73090117e-02 3.46965194e-01 4.05828774e-01
7.43791878e-01 4.19849634e-01 3.58825415e-01 -3.80929232e-01
3.71396631e-01 -2.46248052e-01 7.75288343e-01 5.22218645e-01
2.21098408e-01 5.80318630e-01 6.94728673e-01 -2.46291861e-01
-7.47165978e-01 -8.13164651e-01 9.29029882e-02 1.43097472e+00
-1.66648179e-01 -8.89278471e-01 -8.30138862e-01 -8.78186524e-01
-2.17955410e-01 8.64692152e-01 -9.28215086e-01 -2.16788024e-01
-5.33694744e-01 -1.15486014e+00 3.22474778e-01 8.90263259e-01
4.79930818e-01 -1.12078571e+00 -1.49501771e-01 2.02462822e-01
-5.66662736e-02 -1.08548903e+00 -5.90902984e-01 2.24850148e-01
-7.14655101e-01 -8.11237514e-01 -2.68372357e-01 -8.63458097e-01
5.96146703e-01 2.70065337e-01 1.22377360e+00 -3.36501956e-01
1.50626838e-01 1.42766044e-01 -5.65074146e-01 -5.38033426e-01
-1.68093055e-01 6.54312134e-01 -3.89841557e-01 1.60364330e-01
6.09954000e-01 -7.66095579e-01 -5.44791341e-01 1.98212117e-01
-9.59503293e-01 3.87632728e-01 7.24693537e-01 6.40814364e-01
6.53385758e-01 -5.40182143e-02 6.52702868e-01 -1.36490119e+00
8.17652285e-01 -5.70645094e-01 -3.93036962e-01 1.61353812e-01
-7.93540418e-01 2.13447869e-01 6.50365472e-01 -1.36372279e-02
-1.19519925e+00 -1.52164564e-01 -3.11913311e-01 -1.71913594e-01
-1.26286998e-01 7.69170463e-01 -8.00246179e-01 5.08816063e-01
2.00317681e-01 2.49354750e-01 -3.31378788e-01 -4.54927802e-01
6.54346168e-01 3.20081383e-01 2.29532063e-01 -7.29757249e-01
6.66256785e-01 3.34909499e-01 -1.98901042e-01 -3.94317120e-01
-1.32410729e+00 -3.41835231e-01 -3.64347190e-01 2.14534208e-01
8.59161496e-01 -1.08185542e+00 -4.36196357e-01 4.75606620e-01
-1.03852499e+00 -6.68525174e-02 -3.21294367e-01 8.31174850e-02
-1.61642641e-01 1.16276614e-01 -4.96721119e-01 -4.45666224e-01
-6.62925839e-01 -1.28956759e+00 1.14872885e+00 3.65312368e-01
-2.65566885e-01 -1.08227849e+00 3.77851948e-02 4.45060700e-01
5.94209790e-01 3.40842530e-02 8.66874933e-01 -9.06186998e-01
-4.27882433e-01 5.38732857e-03 -1.11630246e-01 3.70902032e-01
3.27301025e-01 -1.46841183e-01 -1.17704380e+00 -1.35534108e-01
1.14864320e-01 -5.47975115e-02 1.00856268e+00 3.61197382e-01
8.33721459e-01 -3.92899454e-01 -1.66860476e-01 6.85429871e-01
1.30251813e+00 1.43257901e-01 5.34626544e-01 5.32861531e-01
1.10336614e+00 4.60942060e-01 3.65364313e-01 2.35302284e-01
7.09147453e-01 5.67954481e-01 4.46912915e-01 -1.32853970e-01
-2.90952951e-01 -5.94624043e-01 6.13129079e-01 1.25004303e+00
2.42541835e-01 -3.60822529e-01 -7.53131926e-01 7.15659320e-01
-1.93554318e+00 -7.09785283e-01 -1.93313017e-01 1.68906236e+00
8.52480352e-01 1.91508278e-01 6.72267750e-02 -1.96814537e-01
5.12535274e-01 4.15071547e-01 -2.91428357e-01 -4.37202394e-01
-1.55102998e-01 3.47150683e-01 8.71588755e-03 5.28503299e-01
-1.02403057e+00 1.21532857e+00 4.99938345e+00 9.19206798e-01
-8.66207242e-01 2.70557851e-01 6.08895898e-01 -1.33521378e-01
-7.81540990e-01 3.13125014e-01 -1.23473334e+00 4.53772396e-01
7.24213839e-01 -2.89293140e-01 9.87803265e-02 1.09434795e+00
-1.25905260e-01 4.25983042e-01 -1.04723775e+00 5.04945993e-01
3.00901383e-01 -1.21885955e+00 5.29269338e-01 5.77691011e-02
8.90045106e-01 -4.37695123e-02 1.33773968e-01 6.30707204e-01
5.99217832e-01 -1.01787484e+00 7.41794527e-01 2.02960268e-01
3.16613406e-01 -1.03169668e+00 1.00861847e+00 -1.43859491e-01
-1.24835026e+00 1.31196618e-01 -3.09392124e-01 -5.62498122e-02
4.35513824e-01 8.00438821e-01 -5.49191475e-01 7.27269650e-01
4.61076707e-01 8.77189338e-01 -6.04921043e-01 6.01661921e-01
-7.31138825e-01 7.74434507e-01 1.09921038e-01 -1.01618841e-01
6.12376392e-01 -1.73749313e-01 4.39651608e-01 1.31167102e+00
2.10918561e-01 -1.54972911e-01 2.83195097e-02 1.08053923e+00
-1.76635966e-01 2.68596321e-01 -3.98977816e-01 -1.12431422e-01
2.37719551e-01 1.47305012e+00 -6.67173326e-01 -1.02859303e-01
-6.71243846e-01 7.83676267e-01 3.13908100e-01 3.68177831e-01
-9.65474665e-01 -4.85312432e-01 8.02942812e-01 1.03044204e-01
6.97302103e-01 6.52174205e-02 -4.84157294e-01 -1.27367079e+00
-3.42163481e-02 -9.26981509e-01 3.50976795e-01 -7.09555864e-01
-1.30898476e+00 9.26941812e-01 -1.90814629e-01 -6.91261530e-01
-1.96596757e-01 -5.16579628e-01 -7.74830043e-01 9.55233335e-01
-1.47747946e+00 -1.87571537e+00 -1.78525895e-01 4.84708160e-01
7.63793528e-01 -2.12101981e-01 7.50717163e-01 1.90197513e-01
-7.52301157e-01 7.27798700e-01 -2.69161820e-01 1.82176292e-01
5.55173218e-01 -1.35588241e+00 7.32984364e-01 1.15283883e+00
2.66572833e-01 9.37731504e-01 4.55608487e-01 -5.68942189e-01
-1.37526870e+00 -1.32138145e+00 1.07983744e+00 -6.48574769e-01
9.28599358e-01 -6.96544349e-01 -8.58009577e-01 1.00395656e+00
4.87876832e-01 -5.23239911e-01 9.80542600e-01 6.70927405e-01
-5.98933697e-01 -2.36569047e-01 -4.38735187e-01 5.32886207e-01
8.87518406e-01 -3.67178261e-01 -7.23294675e-01 -1.26194522e-01
1.12850821e+00 -2.00627625e-01 -4.42344368e-01 4.62991506e-01
3.56237561e-01 -9.71095443e-01 7.04300225e-01 -5.83914816e-01
8.10595155e-01 -3.00183624e-01 -2.40677238e-01 -1.39420438e+00
-4.59980071e-01 -6.30639791e-01 3.94767299e-02 1.75544763e+00
8.05733740e-01 -7.25395858e-01 5.24191260e-01 4.24436837e-01
-4.83414739e-01 -9.71336246e-01 -5.19511580e-01 -5.11721194e-01
1.84070528e-01 -6.25894725e-01 9.90362704e-01 8.15246344e-01
7.92753771e-02 1.01216805e+00 -2.08133206e-01 5.23066856e-02
2.58147269e-01 5.48814297e-01 7.43396103e-01 -9.49765623e-01
-6.86458468e-01 -6.39179468e-01 -8.13678354e-02 -9.84917641e-01
1.77386925e-01 -1.13543987e+00 -1.18826643e-01 -1.67598665e+00
4.45838720e-01 -3.61829668e-01 -2.94662803e-01 8.99165809e-01
-4.42118496e-01 1.48071021e-01 2.02950940e-01 -2.03650650e-02
-6.43881857e-01 8.56933296e-01 1.16581929e+00 5.38414568e-02
-1.15801685e-01 9.80850216e-03 -1.56251609e+00 7.05904543e-01
9.74629104e-01 -4.97785240e-01 -7.61194110e-01 -6.68279350e-01
8.43432903e-01 -5.63039243e-01 7.18789920e-02 -4.56977993e-01
1.49403112e-02 -8.89313035e-03 -2.00129688e-01 -4.22191739e-01
-4.16436978e-02 -4.46980745e-01 2.81124804e-02 7.65726939e-02
-1.06815659e-01 1.71973869e-01 3.90930563e-01 4.34713721e-01
-4.07611251e-01 -1.95039123e-01 4.19252485e-01 -1.14593133e-01
-6.17718816e-01 2.10214958e-01 -1.78706303e-01 2.30073348e-01
8.38508129e-01 -4.86846594e-03 -4.57449883e-01 -1.73695967e-01
-3.38361382e-01 6.80270791e-02 3.33865851e-01 7.62643218e-01
2.11191222e-01 -1.27823985e+00 -5.90231895e-01 2.99115390e-01
9.61310118e-02 2.81291753e-01 2.24267244e-01 6.91979110e-01
-1.78519147e-03 6.56629264e-01 1.71027362e-01 -2.03334346e-01
-9.42247450e-01 4.83178705e-01 1.04936928e-01 -5.80121756e-01
-4.19498950e-01 9.39573824e-01 8.26803267e-01 -4.53657925e-01
-1.38001785e-01 -3.69919300e-01 -1.71162486e-01 2.05935597e-01
3.77651840e-01 7.18723238e-02 2.68557459e-01 -4.98358607e-01
-3.88877869e-01 4.79193717e-01 -5.16866088e-01 6.17148355e-02
1.46194303e+00 -1.74746335e-01 -2.07490206e-01 3.08170706e-01
9.88683283e-01 2.63653189e-01 -9.25230980e-01 -1.61481082e-01
-3.46023172e-01 -1.17170021e-01 1.71027914e-01 -9.11818624e-01
-1.41835797e+00 9.02006805e-01 -9.26833227e-02 1.95103332e-01
1.14425802e+00 3.12851846e-01 8.30962718e-01 -4.77361418e-02
1.71101049e-01 -7.71295428e-01 9.76372138e-02 6.23798311e-01
7.41022527e-01 -8.79556119e-01 -2.44700998e-01 -5.03610194e-01
-8.90901506e-01 6.88780546e-01 8.61128569e-01 -1.01575218e-01
5.40288389e-01 4.39268857e-01 -1.05407452e-02 -3.34170073e-01
-9.33411717e-01 -2.28564441e-01 3.24280888e-01 3.67923230e-01
7.49685168e-01 3.89408506e-02 -4.48845118e-01 1.46126235e+00
-5.81750512e-01 -2.54704595e-01 2.83506632e-01 7.70921290e-01
-1.99630812e-01 -1.37054312e+00 1.06716000e-01 1.53550684e-01
-6.45109534e-01 -6.24085546e-01 -3.26203138e-01 8.75027299e-01
1.19873755e-01 8.31514776e-01 -1.23177536e-01 -2.93578118e-01
4.03750122e-01 2.20426440e-01 1.80995777e-01 -8.15054059e-01
-7.84182191e-01 2.50627905e-01 4.77960557e-02 -4.84730989e-01
-3.64575684e-01 -5.76201975e-01 -1.16579962e+00 -1.06291696e-01
-2.77372956e-01 2.92758286e-01 5.75884938e-01 1.10842681e+00
8.22584271e-01 1.06116772e+00 3.98897022e-01 -3.79980326e-01
-1.34529278e-01 -9.73431349e-01 -5.01428723e-01 2.51134038e-01
-1.11748919e-01 -3.96053195e-01 -2.08275333e-01 -3.65387648e-02] | [11.473362922668457, 6.697690010070801] |
53791eda-f67c-4fba-bbda-6937c9a4dacd | tuple-oriented-compression-for-large-scale | 1702.06943 | null | http://arxiv.org/abs/1702.06943v3 | http://arxiv.org/pdf/1702.06943v3.pdf | Tuple-oriented Compression for Large-scale Mini-batch Stochastic Gradient Descent | Data compression is a popular technique for improving the efficiency of data
processing workloads such as SQL queries and more recently, machine learning
(ML) with classical batch gradient methods. But the efficacy of such ideas for
mini-batch stochastic gradient descent (MGD), arguably the workhorse algorithm
of modern ML, is an open question. MGD's unique data access pattern renders
prior art, including those designed for batch gradient methods, less effective.
We fill this crucial research gap by proposing a new lossless compression
scheme we call tuple-oriented compression (TOC) that is inspired by an unlikely
source, the string/text compression scheme Lempel-Ziv-Welch, but tailored to
MGD in a way that preserves tuple boundaries within mini-batches. We then
present a suite of novel compressed matrix operation execution techniques
tailored to the TOC compression scheme that operate directly over the
compressed data representation and avoid decompression overheads. An extensive
empirical evaluation with real-world datasets shows that TOC consistently
achieves substantial compression ratios by up to 51x and reduces runtimes for
MGD workloads by up to 10.2x in popular ML systems. | ['Yijing Zeng', 'Fengan Li', 'Jeffrey F. Naughton', 'Xi Wu', 'Jignesh M. Patel', 'Lingjiao Chen', 'Arun Kumar'] | 2017-02-22 | null | null | null | null | ['text-compression'] | ['natural-language-processing'] | [ 6.20480441e-02 -2.58067340e-01 -4.07558084e-01 -6.26357019e-01
-9.12787497e-01 -1.70901865e-01 4.13520306e-01 7.22722769e-01
-5.31729281e-01 5.30731082e-01 4.80728686e-01 -8.89285982e-01
-6.17779605e-02 -9.26529527e-01 -8.29426348e-01 -3.18520933e-01
-3.49577993e-01 7.88901806e-01 2.34131262e-01 -3.70167911e-01
4.98722047e-01 6.95346057e-01 -1.78432345e+00 9.03172553e-01
5.39536417e-01 1.23738647e+00 7.63627589e-02 8.37378979e-01
-5.18142343e-01 1.14714694e+00 -3.72967392e-01 -7.65762150e-01
7.53873169e-01 -2.22013310e-01 -7.83327103e-01 -2.57422328e-01
4.36113089e-01 -8.33517849e-01 -4.14710432e-01 5.80655992e-01
4.70316470e-01 5.08381017e-02 1.65650353e-01 -8.20774138e-01
1.07064553e-01 1.00794852e+00 -5.14475107e-01 1.42812937e-01
1.56828001e-01 1.60022918e-02 1.07346344e+00 -9.53437865e-01
7.59864509e-01 1.38534904e+00 8.30244482e-01 -1.88147724e-02
-1.41876924e+00 -2.62819678e-02 -2.93867201e-01 2.11667061e-01
-1.20707083e+00 -6.79169118e-01 3.66452456e-01 4.15738225e-02
1.59393179e+00 8.73009980e-01 5.11417389e-01 2.59632677e-01
1.04939736e-01 1.21809828e+00 7.17971325e-01 -4.82381284e-01
5.02720952e-01 2.03130528e-01 1.18697047e-01 5.28136790e-01
5.37710309e-01 -9.10909567e-03 -9.48922575e-01 -6.12272561e-01
6.55376613e-02 3.63463551e-01 5.33429114e-03 -6.33392990e-01
-8.00304413e-01 9.76861775e-01 9.32326391e-02 4.56498452e-02
-6.22763753e-01 1.41571745e-01 1.05217040e+00 7.00632095e-01
5.96743345e-01 6.74453154e-02 -5.34820139e-01 -7.99067438e-01
-1.51588666e+00 7.48829603e-01 1.60853541e+00 1.14031804e+00
8.21300626e-01 1.79045927e-02 -4.20368016e-01 7.99337327e-01
9.99174714e-02 3.07937473e-01 5.27888775e-01 -9.98880267e-01
8.87939990e-01 7.21957624e-01 -2.30623722e-01 -8.55212450e-01
-4.63102832e-02 -2.75057882e-01 -8.87531698e-01 -1.24679655e-01
-4.12230231e-02 2.57574171e-01 -2.71880358e-01 1.08443284e+00
4.56020474e-01 -4.50226873e-01 -1.27995297e-01 5.22326767e-01
1.03864692e-01 6.58816934e-01 -3.78498912e-01 -6.51234806e-01
7.85197854e-01 -8.83102059e-01 -5.44101894e-01 -4.37765904e-02
1.11350369e+00 -8.02735269e-01 1.30041718e+00 5.95378876e-01
-1.66131032e+00 -2.81887800e-01 -1.19975197e+00 -2.16277003e-01
-1.44535050e-01 -2.41413742e-01 8.93801808e-01 8.55167091e-01
-1.20526934e+00 9.08554375e-01 -1.20236242e+00 -3.43063027e-01
5.12908220e-01 2.42798150e-01 7.70813972e-03 -3.45774591e-01
-4.38078761e-01 6.02901638e-01 4.91582721e-01 -3.56504977e-01
-4.02601331e-01 -9.88702774e-01 -3.58989865e-01 3.06976378e-01
5.09546936e-01 -5.86942673e-01 1.36511338e+00 1.75808184e-02
-1.40241730e+00 8.60196412e-01 -4.04386282e-01 -1.30021489e+00
7.36704171e-01 -6.83178306e-01 2.07804948e-01 1.07364893e-01
-4.89140958e-01 1.43305704e-01 7.44404256e-01 -8.82390022e-01
-5.35148382e-01 -6.31102979e-01 -5.70758164e-01 5.80365509e-02
-6.47889555e-01 3.07000317e-02 -5.46918511e-01 -6.20252609e-01
1.71562895e-01 -6.83251023e-01 -9.74065736e-02 -1.91966400e-01
-4.39497560e-01 -2.63760030e-01 8.13882709e-01 -6.23946965e-01
1.91139603e+00 -1.93698013e+00 -5.71845472e-02 2.57752895e-01
2.98491895e-01 3.89210105e-01 3.16599369e-01 1.06823742e+00
4.08979565e-01 -1.05202734e-01 -5.86583197e-01 -6.82534873e-01
2.66700953e-01 3.21056336e-01 -8.30105126e-01 3.40985775e-01
-2.48976111e-01 9.36676443e-01 -4.98229831e-01 -5.02031982e-01
6.02443442e-02 3.41648199e-02 -8.97800386e-01 6.19198680e-01
-4.68743801e-01 -7.16919303e-01 -1.58203002e-02 8.51317286e-01
8.41272414e-01 -2.00907499e-01 2.98137039e-01 -2.55104840e-01
-2.63205886e-01 5.48503101e-01 -1.14866281e+00 1.56899047e+00
-9.47796926e-02 3.77277941e-01 1.67679042e-01 -1.23815143e+00
1.05229187e+00 -2.86970228e-01 7.38506019e-01 -7.10889935e-01
-3.50653440e-01 5.73074698e-01 -6.15618706e-01 -5.06737947e-01
9.51287627e-01 2.92715818e-01 2.55302370e-01 8.07623565e-01
-3.40757996e-01 -4.75218236e-01 6.14349902e-01 6.25987530e-01
1.47982216e+00 -2.20111176e-01 1.24484770e-01 -1.29901871e-01
2.03792825e-01 1.20810471e-01 2.84207314e-01 1.20234823e+00
1.10519104e-01 5.02300739e-01 3.53163868e-01 -6.15354896e-01
-1.40474808e+00 -8.51740897e-01 -9.87627953e-02 1.44437420e+00
-4.07001525e-01 -1.11434507e+00 -5.64157963e-01 -3.35161418e-01
6.01393640e-01 8.64482462e-01 -5.37391566e-02 -1.92428291e-01
-9.16780531e-01 -9.60078418e-01 4.82033759e-01 3.15798163e-01
5.73607862e-01 -7.45319366e-01 -9.16968644e-01 5.58060288e-01
3.45046461e-01 -8.58295739e-01 -3.02380890e-01 4.41195011e-01
-1.67819035e+00 -6.94412827e-01 -5.54711595e-02 -2.70786673e-01
2.74088178e-02 1.38595983e-01 1.58522105e+00 -6.23941198e-02
-4.95569259e-01 5.56487329e-02 -4.03202236e-01 -4.00247842e-01
-5.63973427e-01 1.79987535e-01 -1.34138271e-01 -2.00581297e-01
7.90630817e-01 -8.55930865e-01 -6.33832335e-01 -1.71138123e-01
-1.35730577e+00 6.86468631e-02 7.85617232e-01 8.14870298e-01
9.97825801e-01 -1.53377220e-01 -6.64196685e-02 -1.49470925e+00
9.45460260e-01 -5.91053545e-01 -4.92485076e-01 1.82721123e-01
-1.58550560e+00 3.87641668e-01 5.83316207e-01 -7.56908879e-02
-8.08642387e-01 -1.74881026e-01 -2.99131453e-01 -5.86568058e-01
2.80599475e-01 6.65265977e-01 2.71875679e-01 1.47675902e-01
7.07024217e-01 4.77138042e-01 3.15094680e-01 -8.94734979e-01
4.26636606e-01 7.24051118e-01 6.64360523e-01 -6.22279882e-01
4.50583041e-01 7.99161553e-01 1.88789107e-02 -7.57404387e-01
-5.13952374e-01 -5.30101001e-01 -2.00913444e-01 3.61508876e-01
1.08352110e-01 -5.28471172e-01 -7.24617302e-01 2.57297754e-01
-7.68175364e-01 -3.34352970e-01 -8.52998555e-01 4.21446741e-01
-7.65986323e-01 4.47765678e-01 -1.11186624e+00 -6.24952674e-01
-9.62454259e-01 -7.91743755e-01 1.01148355e+00 -3.69879305e-01
-1.34602144e-01 -4.85368669e-01 3.26406568e-01 2.16083393e-01
9.40525830e-01 -2.86415927e-02 1.23133349e+00 -8.99932802e-01
-6.80283844e-01 -4.85523045e-01 -2.67850935e-01 5.83222508e-01
-3.44503403e-01 -1.32789999e-01 -4.32660878e-01 -7.02896833e-01
5.54315329e-01 -6.31992221e-01 1.10235691e+00 9.23697352e-02
1.49987924e+00 -6.80984974e-01 -1.67087600e-01 1.08609951e+00
1.74080765e+00 -1.48057222e-01 7.07607329e-01 2.50936121e-01
4.41988796e-01 4.34504449e-01 6.24615371e-01 1.14427960e+00
2.88238734e-01 4.72403198e-01 3.76134396e-01 3.17699999e-01
6.61394447e-02 -4.49411362e-01 3.90613645e-01 1.38810718e+00
2.60555655e-01 -3.37360650e-02 -7.32427359e-01 3.25733781e-01
-1.86468446e+00 -8.11827183e-01 -1.66869357e-01 2.54661250e+00
1.24752784e+00 3.58369976e-01 1.59074366e-01 5.15275776e-01
1.52121872e-01 2.13062003e-01 -7.75249004e-01 -9.50002611e-01
-2.94232547e-01 4.80796993e-01 7.79108524e-01 1.56229392e-01
-6.98006094e-01 7.65125096e-01 6.13867092e+00 1.05260921e+00
-7.06882715e-01 -5.62080331e-02 7.33388245e-01 -6.21904969e-01
-3.55285674e-01 8.99897367e-02 -1.25643611e+00 4.87258166e-01
1.54339707e+00 -3.08214903e-01 7.43222773e-01 1.13056898e+00
5.37363626e-02 -2.08227471e-01 -1.31544173e+00 1.28811574e+00
-1.04217399e-02 -1.82214928e+00 1.94453388e-01 6.97013363e-02
7.22117126e-01 3.45825940e-01 -3.29921134e-02 4.81996030e-01
1.67215973e-01 -8.00812185e-01 3.58291417e-01 4.13294315e-01
5.49103379e-01 -7.59226501e-01 5.60235441e-01 3.96250755e-01
-8.13128650e-01 -1.10769838e-01 -8.07973266e-01 1.10043297e-02
1.98091656e-01 1.26479459e+00 -1.00792229e+00 3.77558917e-02
1.04339552e+00 3.62743139e-01 -7.14044392e-01 9.23376501e-01
7.45980442e-01 1.01714396e+00 -7.33820558e-01 -7.51800165e-02
1.94065809e-01 -2.69109458e-01 4.94433522e-01 1.57168460e+00
1.80896237e-01 -3.47092628e-01 -8.25310946e-02 6.42351687e-01
-3.63000453e-01 5.00362217e-01 -5.90713322e-01 -1.69753000e-01
7.47787118e-01 6.85138047e-01 -2.27676779e-01 -6.29995406e-01
-3.75641882e-01 8.00940156e-01 4.48642850e-01 1.55287623e-01
-2.86235332e-01 -5.35468519e-01 5.44949234e-01 3.03099364e-01
7.20016599e-01 -2.65785843e-01 -5.98636568e-01 -9.92500067e-01
6.43133461e-01 -1.18609893e+00 5.54384947e-01 -5.30120842e-02
-1.23088551e+00 3.79682094e-01 9.24177021e-02 -6.97935462e-01
-6.67962909e-01 -2.34699875e-01 -1.23331666e-01 5.54230690e-01
-1.19090331e+00 -3.67994994e-01 -1.13642909e-01 6.44179761e-01
4.58603114e-01 -3.12282383e-01 5.58660030e-01 3.59911978e-01
-1.97589427e-01 8.95950735e-01 9.77907538e-01 -5.48275113e-01
5.52519858e-01 -1.26669145e+00 5.17671943e-01 5.94862223e-01
1.73886791e-01 7.73297846e-01 7.55112827e-01 -6.58069015e-01
-2.37896562e+00 -9.45604622e-01 9.80710745e-01 -1.69715602e-02
3.26204747e-01 -4.20394570e-01 -1.26779807e+00 5.18980324e-01
-5.33609279e-02 -2.14487419e-01 4.01761562e-01 -1.19857155e-01
-4.23178345e-01 -8.07818413e-01 -1.30115902e+00 3.90098363e-01
9.18586075e-01 -7.01663256e-01 -4.64675277e-01 7.66059637e-01
9.79411006e-01 -4.00471956e-01 -1.05698955e+00 6.11750722e-01
3.65506202e-01 -1.50357258e+00 1.08426082e+00 -7.21482515e-01
4.31414455e-01 2.32049644e-01 -7.28037238e-01 -5.57142138e-01
2.76139557e-01 -1.06728554e+00 -1.18407476e+00 1.08888650e+00
-1.35047911e-02 -4.02666539e-01 1.29556310e+00 8.09399724e-01
-9.79045108e-02 -1.37387431e+00 -7.01946616e-01 -7.46653557e-01
-4.81174365e-02 -7.03082442e-01 7.02119291e-01 4.72448468e-01
-8.50808397e-02 -3.42722535e-02 -3.29832733e-01 -6.77422762e-01
7.34444380e-01 4.18524057e-01 1.00497496e+00 -7.14394450e-01
-7.51408160e-01 -5.04050970e-01 -2.19642386e-01 -1.30799437e+00
-4.24473912e-01 -9.46721196e-01 -3.61963809e-01 -9.82726693e-01
7.70473406e-02 -5.03915846e-01 -1.08249411e-01 1.39519619e-02
1.02317475e-01 -4.49770354e-02 7.20131472e-02 6.72310591e-01
-6.80469394e-01 5.19213915e-01 4.30556566e-01 -1.62187573e-02
-4.93360549e-01 1.03115655e-01 -4.22241032e-01 1.81185186e-01
5.95586002e-01 -4.17818099e-01 -3.90665859e-01 -5.12031615e-01
5.72787404e-01 6.81273192e-02 -2.60616750e-01 -9.94352698e-01
6.54651642e-01 1.47178382e-01 -2.52421126e-02 -1.06735718e+00
1.04739808e-01 -4.44396019e-01 8.74503404e-02 6.59546793e-01
-6.90800846e-01 5.11595964e-01 8.84406418e-02 5.04885912e-01
-4.56708610e-01 -2.29757160e-01 5.20754337e-01 -1.56923860e-01
-2.23989949e-01 3.53137076e-01 -9.21154246e-02 4.78208095e-01
6.26279175e-01 -1.34020731e-01 -1.52042463e-01 -3.52212042e-01
-3.60911191e-02 2.34734546e-02 5.01858592e-01 -2.12748185e-01
9.79809403e-01 -1.06498992e+00 -1.07054973e+00 3.77780944e-01
-7.95401931e-02 2.61769444e-02 -8.10667872e-02 8.16395640e-01
-1.08222246e+00 7.63401926e-01 3.30317467e-01 -7.17641413e-01
-1.24625933e+00 7.27031827e-01 -1.81369767e-01 -7.16903269e-01
-7.17592835e-01 1.09140730e+00 -6.60821497e-01 -3.24238330e-01
6.48257911e-01 -2.23268658e-01 6.52428508e-01 -1.17865816e-01
5.52861154e-01 8.10089827e-01 7.67757714e-01 1.84001207e-01
3.74740362e-02 -2.50188977e-01 -5.75127542e-01 6.23194361e-03
1.78597403e+00 -1.06014870e-01 -6.67106450e-01 6.43362045e-01
1.80395114e+00 6.42693415e-02 -9.10472870e-01 -6.32049978e-01
3.64997834e-01 -8.40217352e-01 1.35801267e-02 -3.69408399e-01
-9.01071429e-01 6.88152492e-01 5.89100122e-01 4.18815911e-01
1.30616260e+00 -3.04476470e-01 1.40327501e+00 9.64802682e-01
4.24368501e-01 -1.21494186e+00 -2.18885392e-01 3.91887367e-01
5.76941669e-01 -1.02132666e+00 3.50286752e-01 1.23594373e-01
-2.11521894e-01 1.12000012e+00 2.25918442e-01 -7.05147237e-02
7.42642045e-01 6.07818961e-01 -5.02005935e-01 -1.50206432e-01
-1.30939853e+00 4.57661122e-01 -2.39285916e-01 -1.52475405e-02
4.71809447e-01 1.15155205e-02 -7.64958858e-01 -1.29994629e-02
-5.01866996e-01 1.49599612e-01 5.71790449e-02 1.68714988e+00
-5.32205582e-01 -1.28863883e+00 -2.03810841e-01 1.39324307e+00
-4.85423952e-01 -3.39840591e-01 -3.55882123e-02 4.11714733e-01
-5.64134777e-01 6.74953401e-01 1.05576180e-01 -6.85386181e-01
2.27174357e-01 3.25048804e-01 1.79439962e-01 -2.26118520e-01
-9.00012374e-01 -1.86955854e-01 -6.21376969e-02 -1.17479396e+00
7.51382858e-02 -7.63516009e-01 -8.59473228e-01 -1.12245369e+00
9.02255550e-02 3.70992243e-01 9.42839503e-01 4.80247468e-01
7.44959116e-01 -1.22072510e-01 7.15133548e-01 -4.12549853e-01
-1.42996526e+00 -7.41598308e-01 -5.83619535e-01 6.61418200e-01
1.23599596e-01 1.85812458e-01 -2.79518366e-01 -4.94327769e-02] | [8.516111373901367, 3.4466450214385986] |
2014a509-4be2-4471-86d8-12d38d38271f | lane-graph-as-path-continuity-preserving-path | 2303.08815 | null | https://arxiv.org/abs/2303.08815v1 | https://arxiv.org/pdf/2303.08815v1.pdf | Lane Graph as Path: Continuity-preserving Path-wise Modeling for Online Lane Graph Construction | Online lane graph construction is a promising but challenging task in autonomous driving. Previous methods usually model the lane graph at the pixel or piece level, and recover the lane graph by pixel-wise or piece-wise connection, which breaks down the continuity of the lane. Human drivers focus on and drive along the continuous and complete paths instead of considering lane pieces. Autonomous vehicles also require path-specific guidance from lane graph for trajectory planning. We argue that the path, which indicates the traffic flow, is the primitive of the lane graph. Motivated by this, we propose to model the lane graph in a novel path-wise manner, which well preserves the continuity of the lane and encodes traffic information for planning. We present a path-based online lane graph construction method, termed LaneGAP, which end-to-end learns the path and recovers the lane graph via a Path2Graph algorithm. We qualitatively and quantitatively demonstrate the superiority of LaneGAP over conventional pixel-based and piece-based methods. Abundant visualizations show LaneGAP can cope with diverse traffic conditions. Code and models will be released at \url{https://github.com/hustvl/LaneGAP} for facilitating future research. | ['Xinggang Wang', 'Chang Huang', 'Wenyu Liu', 'Qian Zhang', 'Tianheng Cheng', 'Bo Jiang', 'Shaoyu Chen', 'Bencheng Liao'] | 2023-03-15 | null | null | null | null | ['graph-construction', 'trajectory-planning'] | ['graphs', 'robots'] | [-1.65695533e-01 3.16006541e-01 -4.75826293e-01 -5.00613391e-01
-3.70751888e-01 -7.13177383e-01 4.15058374e-01 -1.54837351e-02
1.31854847e-01 4.41332757e-01 9.12136137e-02 -8.75858486e-01
-1.11838222e-01 -1.03182006e+00 -8.50192845e-01 -3.31580848e-01
-1.78579167e-02 2.18890697e-01 6.92347825e-01 -4.07653242e-01
4.38565582e-01 7.08985031e-01 -1.46761227e+00 -1.74587041e-01
1.05480278e+00 5.09229958e-01 1.60888761e-01 6.34845614e-01
-3.25754732e-01 4.04063135e-01 3.67731154e-01 -3.12839001e-01
4.95096415e-01 -3.33229810e-01 -3.61162156e-01 3.00690204e-01
5.88384628e-01 -3.91948730e-01 -9.48972642e-01 1.05005360e+00
-9.25974250e-02 5.35070663e-03 1.95554227e-01 -1.76713920e+00
-7.44053423e-02 2.61840940e-01 -6.37771070e-01 -2.61735935e-02
1.95671618e-01 5.72412431e-01 9.94525492e-01 -6.46748185e-01
9.31083202e-01 1.26085997e+00 6.04704022e-01 1.15984499e-01
-9.55860853e-01 -6.45038486e-01 3.62049848e-01 5.86350739e-01
-1.33932436e+00 -3.57465804e-01 1.23929155e+00 -4.83322412e-01
2.91250229e-01 3.85103747e-02 8.11581910e-01 5.85363567e-01
4.64846402e-01 7.65840471e-01 9.64318275e-01 -7.47043714e-02
-7.28116557e-02 -1.98495999e-01 4.11945581e-01 1.27035117e+00
1.96954474e-01 5.28018832e-01 -4.78490829e-01 3.08569729e-01
8.67691338e-01 -8.10267462e-04 -1.17418498e-01 -8.63725305e-01
-1.28570235e+00 6.76031590e-01 6.03409648e-01 -3.27854067e-01
-2.06695452e-01 3.12675625e-01 1.87410533e-01 1.20847799e-01
5.53087443e-02 -1.26520634e-01 1.34042859e-01 -3.17394406e-01
-7.36504197e-01 4.93381232e-01 6.99097157e-01 1.49192941e+00
1.41338909e+00 -1.75681710e-01 -5.41888028e-02 3.23648542e-01
3.47894967e-01 5.71218431e-01 -3.76198530e-01 -1.50116181e+00
5.88026643e-01 6.97278559e-01 1.88542917e-01 -1.38572407e+00
-5.47206581e-01 -2.84615517e-01 -5.17733455e-01 3.30445349e-01
6.55187309e-01 -4.79799956e-02 -7.88780093e-01 1.44169104e+00
5.19160032e-01 5.91327906e-01 -1.50363699e-01 9.86731231e-01
5.82360446e-01 6.74899995e-01 -2.69407183e-01 1.92290545e-01
1.16761196e+00 -1.54709876e+00 -8.38840306e-01 -4.47158188e-01
7.38987863e-01 -2.67524183e-01 1.12889612e+00 1.15762934e-01
-9.61059451e-01 -4.99035954e-01 -1.12120509e+00 -3.76524329e-01
-3.90095562e-01 2.61743460e-02 4.15249735e-01 3.51307839e-01
-9.76126075e-01 2.43597314e-01 -7.23107040e-01 -1.10613421e-01
5.00537336e-01 -2.52548099e-01 -4.50418025e-01 -4.97227699e-01
-1.09271491e+00 6.24708891e-01 3.37078691e-01 2.63746232e-01
-6.96743488e-01 -6.70237958e-01 -1.11821806e+00 -1.65327355e-01
6.55193865e-01 -4.91604805e-01 1.07194865e+00 -2.22848922e-01
-1.23302603e+00 5.79783738e-01 -4.99838173e-01 -5.52007914e-01
9.02474642e-01 1.14745103e-01 -3.66599143e-01 1.23195879e-01
1.55542806e-01 8.61068249e-01 6.19345009e-01 -1.56831110e+00
-1.00484192e+00 -1.57928467e-01 1.96785063e-01 3.55666727e-02
4.54838336e-01 -7.20932245e-01 -9.78340447e-01 -4.13801372e-02
5.32263577e-01 -9.97533202e-01 -2.80505687e-01 4.56135988e-01
-6.26047373e-01 -2.34186351e-01 1.17318416e+00 -5.98894477e-01
1.51530743e+00 -2.17715836e+00 -3.89595509e-01 3.57232034e-01
5.08202493e-01 -1.42101422e-01 -2.73034275e-01 6.41123235e-01
2.53393829e-01 -1.18126586e-01 -6.20340884e-01 -1.56522557e-01
1.59926400e-01 3.74917924e-01 -4.13605005e-01 5.91154158e-01
-1.26121938e-01 1.15238225e+00 -1.21610069e+00 -4.15419310e-01
3.03078145e-01 2.53995031e-01 -3.23235393e-01 -1.87187761e-01
-1.98978156e-01 5.20091474e-01 -4.44336236e-01 4.76772130e-01
1.01250410e+00 2.11084142e-01 -7.44263381e-02 -1.45329773e-01
-7.21805274e-01 1.96956530e-01 -1.04481065e+00 1.76154757e+00
-2.17349261e-01 1.05963433e+00 8.31993595e-02 -5.90664089e-01
1.17297661e+00 -3.34796190e-01 3.54895562e-01 -9.58157659e-01
-7.76129367e-04 8.84681866e-02 -1.15240075e-01 -4.41533953e-01
6.40064180e-01 2.37645105e-01 -2.80264586e-01 3.19670647e-01
-6.30581737e-01 -1.43843234e-01 2.79571295e-01 3.02492052e-01
9.40457523e-01 3.27995986e-01 -2.58085765e-02 -1.40442371e-01
3.89072537e-01 5.88271916e-01 6.56543672e-01 6.23236179e-01
-5.78145206e-01 4.25703347e-01 7.97306716e-01 -4.21262294e-01
-1.06723952e+00 -1.24366629e+00 -1.82865595e-03 5.54303527e-01
8.73568058e-01 -3.75498265e-01 -8.05258512e-01 -5.61505258e-01
3.08700055e-01 8.63993645e-01 -3.71455282e-01 -1.29677251e-01
-8.05313170e-01 2.78505474e-01 2.53272742e-01 2.94651538e-01
8.09906483e-01 -7.35907733e-01 -5.46252787e-01 2.73522824e-01
-2.04955950e-01 -1.17498136e+00 -7.69653201e-01 -4.34185386e-01
-6.16108894e-01 -1.24969733e+00 -8.69111642e-02 -6.71582818e-01
8.35781157e-01 9.30480957e-01 5.94750941e-01 2.13195458e-01
-6.62796840e-04 3.64701569e-01 -1.07857466e-01 -2.35385790e-01
-2.16374367e-01 -2.89066490e-02 -4.04710025e-01 1.09403089e-01
1.87186077e-01 -6.14177287e-01 -8.13368857e-01 6.32102132e-01
-3.51376951e-01 7.11156011e-01 3.57014984e-01 3.89556408e-01
1.00895536e+00 3.36592942e-01 4.04555470e-01 -1.01782906e+00
4.01453257e-01 -6.30498528e-01 -9.79742169e-01 -1.68765962e-01
-5.98540068e-01 -7.93153197e-02 4.87380683e-01 1.52253300e-01
-7.78508484e-01 3.41888130e-01 -5.90208992e-02 -6.45919383e-01
-2.41870239e-01 2.69737393e-01 -4.42290723e-01 -7.83652663e-02
2.35593617e-01 2.60657161e-01 2.10606754e-01 -1.68238938e-01
8.91799688e-01 3.00371051e-01 8.82816970e-01 -2.10153937e-01
1.11389089e+00 8.84534121e-01 2.77827501e-01 -5.74223995e-01
-5.66305757e-01 -6.34409010e-01 -1.05739701e+00 -7.40432441e-01
6.97315454e-01 -6.31634176e-01 -7.55300343e-01 3.15608591e-01
-1.11404824e+00 -7.88259208e-01 -1.46062598e-01 1.86447576e-01
-8.21666002e-01 4.17674392e-01 -2.15343446e-01 -6.84690714e-01
3.21142107e-01 -1.01207292e+00 9.12296474e-01 2.29218215e-01
3.46588194e-01 -1.10908878e+00 -7.95096830e-02 1.92094237e-01
2.40995642e-03 5.63169599e-01 6.07472301e-01 1.46464065e-01
-1.22857916e+00 -4.21275169e-01 -6.39832377e-01 -2.77847528e-01
6.47301301e-02 8.13809857e-02 -8.03019524e-01 3.43027903e-04
-4.95547086e-01 4.78723466e-01 9.77445960e-01 3.21131378e-01
1.04814184e+00 -3.70721251e-01 -5.90655088e-01 7.83949614e-01
1.35485792e+00 1.91984221e-01 8.47342253e-01 4.57905859e-01
1.14856136e+00 1.10745227e+00 8.43691587e-01 1.08956710e-01
1.12472200e+00 5.24991691e-01 5.58956742e-01 -4.35102619e-02
-3.90478015e-01 -1.04852796e+00 2.07549706e-01 5.96493423e-01
4.33348984e-01 -1.44516081e-01 -1.06875575e+00 6.92290843e-01
-2.15366316e+00 -8.52948606e-01 -7.38287449e-01 2.09936762e+00
1.76847190e-01 4.51590359e-01 3.24706852e-01 8.96974206e-02
7.39765525e-01 2.97076046e-01 -7.12396264e-01 -4.28545624e-01
8.81073847e-02 -6.84885323e-01 1.05127025e+00 1.11910355e+00
-8.05803239e-01 1.25661254e+00 5.86112356e+00 7.71863461e-01
-1.02865827e+00 -3.19807120e-02 3.76072913e-01 4.03634667e-01
-5.91200590e-01 4.56296504e-01 -9.15782213e-01 4.34788138e-01
8.01893234e-01 -4.55399960e-01 4.66873676e-01 7.49665916e-01
8.30708504e-01 -2.79432327e-01 -8.00547302e-01 8.76731455e-01
-2.52861291e-01 -1.46600866e+00 -1.40652344e-01 2.52949178e-01
6.35130525e-01 -1.26291767e-01 -1.22913659e-01 9.14154723e-02
1.75934449e-01 -6.97222412e-01 9.56930101e-01 8.56442153e-01
7.45228648e-01 -7.85302460e-01 6.52939528e-02 6.66770995e-01
-1.68898892e+00 3.54772992e-02 -2.24221215e-01 8.43973979e-02
6.71394825e-01 4.67119217e-01 -8.46391201e-01 6.46575153e-01
2.25290403e-01 1.16349876e+00 -5.63219607e-01 1.25638878e+00
-4.53010887e-01 5.03839374e-01 -6.54466748e-02 2.78277963e-01
5.16405523e-01 -7.59322703e-01 8.53050411e-01 1.19872975e+00
2.57484853e-01 7.94429779e-02 3.82671922e-01 1.11748374e+00
1.78190202e-01 -2.39128947e-01 -1.07979691e+00 1.26837894e-01
7.63059914e-01 1.15335405e+00 -6.90168798e-01 -1.27866834e-01
-4.70932811e-01 5.89125514e-01 1.29799068e-01 4.46760625e-01
-9.62004602e-01 -5.90640962e-01 7.12060690e-01 6.39756203e-01
1.70377091e-01 -8.72697353e-01 -5.54944158e-01 -2.84538478e-01
1.45714775e-01 -1.77846462e-01 -6.61080405e-02 -8.51004064e-01
-7.51860142e-01 2.86394000e-01 1.58392772e-01 -1.65027606e+00
2.12997243e-01 -4.31984216e-01 -8.30329478e-01 6.10859752e-01
-2.07258129e+00 -1.19191301e+00 -7.72612929e-01 4.95547205e-01
5.79920530e-01 2.39660874e-01 -4.53860015e-02 7.96868429e-02
-5.04756689e-01 3.74812543e-01 -1.32864177e-01 -9.93855670e-02
2.90114552e-01 -9.32737410e-01 6.93797946e-01 9.46595907e-01
-3.10024589e-01 2.73728281e-01 5.98030329e-01 -8.27639699e-01
-1.70892692e+00 -1.42483139e+00 8.05611849e-01 -3.65918964e-01
8.13287616e-01 -3.74353766e-01 -1.04774499e+00 6.93258703e-01
-3.72843146e-02 2.51522046e-02 -1.11277089e-01 -4.73495752e-01
-1.98902577e-01 -3.09119225e-01 -8.83026898e-01 8.92056584e-01
1.56898057e+00 -4.90174770e-01 -1.27434358e-01 2.17559218e-01
8.77799273e-01 -4.97852176e-01 -2.92674750e-01 6.12146258e-02
5.69422066e-01 -1.13498938e+00 7.20844507e-01 -7.90496692e-02
1.72742873e-01 -8.65103483e-01 2.49275014e-01 -1.10831726e+00
-3.93435180e-01 -8.71805489e-01 -4.65142317e-02 8.69693160e-01
3.60312939e-01 -9.03104901e-01 8.23595762e-01 3.53306621e-01
-6.63830698e-01 -7.28349507e-01 -8.85976493e-01 -7.65261173e-01
-2.22521648e-02 -8.27307045e-01 7.59789705e-01 5.95041931e-01
-1.12636842e-01 -4.84727286e-02 -1.92222476e-01 3.83558184e-01
7.87035346e-01 1.80319190e-01 1.12780118e+00 -1.01198792e+00
3.64630312e-01 -6.92419708e-01 -4.31246698e-01 -1.62709439e+00
2.75699228e-01 -9.68942821e-01 4.38119233e-01 -1.94928384e+00
-3.42635125e-01 -7.49883592e-01 1.68165892e-01 3.87881935e-01
1.82832092e-01 -8.64152685e-02 1.26006514e-01 1.54996812e-01
-4.78581756e-01 4.82463747e-01 1.93667161e+00 -1.05586000e-01
-3.68924856e-01 -5.59574887e-02 -5.21979988e-01 5.42440116e-01
9.80277836e-01 -9.19527039e-02 -7.87502587e-01 -3.20640117e-01
-1.90570801e-02 2.87552387e-01 4.65453207e-01 -8.78740549e-01
6.04799807e-01 -4.41183090e-01 -4.92013603e-01 -1.09253728e+00
3.76880914e-01 -8.26270461e-01 1.39049798e-01 3.62061530e-01
-1.28834158e-01 1.57810003e-01 2.97148973e-01 9.91523147e-01
-6.75402731e-02 1.45235226e-01 5.44992387e-01 1.29855976e-01
-1.27693617e+00 7.00292706e-01 -2.42438167e-01 -2.71093011e-01
1.26029253e+00 -8.10101628e-01 -5.53846478e-01 -4.95829195e-01
-4.95039642e-01 8.63424659e-01 6.71791792e-01 5.50364971e-01
8.19154441e-01 -1.44896793e+00 -7.23300636e-01 4.24181014e-01
2.64448524e-01 1.12694733e-01 4.51394379e-01 1.06766272e+00
-8.95116866e-01 4.58660334e-01 -2.78772146e-01 -6.53107643e-01
-8.82607579e-01 5.60412467e-01 4.11122173e-01 2.43462846e-01
-1.24001813e+00 3.65238339e-01 4.40505475e-01 -3.17790627e-01
2.99197454e-02 -3.13964128e-01 -9.15809274e-02 -2.63977051e-01
4.44951147e-01 7.40522742e-01 -3.17305773e-01 -9.06239212e-01
-8.78652781e-02 8.26787472e-01 3.71695131e-01 -1.77616879e-01
7.85307050e-01 -7.43321359e-01 2.99603101e-02 5.93648612e-01
9.85134900e-01 5.34830131e-02 -1.92718351e+00 2.08596867e-02
-5.55791445e-02 -8.00172210e-01 8.22090209e-02 -3.47844422e-01
-1.09825265e+00 1.01619720e+00 3.64879906e-01 1.64267302e-01
7.14620650e-01 -2.63595283e-01 1.21493554e+00 5.56640886e-02
6.33874536e-01 -9.71936047e-01 -3.66412669e-01 5.93072534e-01
7.86490202e-01 -1.08293784e+00 -2.50981718e-01 -9.82858658e-01
-5.18226445e-01 1.20545208e+00 6.36730433e-01 -2.66057909e-01
7.90100098e-01 6.61371946e-02 -4.09220308e-02 -2.89853334e-01
-5.55277824e-01 -3.99335086e-01 2.80116677e-01 6.84539616e-01
-1.94626555e-01 3.74296665e-01 -3.20285618e-01 2.12633878e-01
-4.80071336e-01 -2.17579827e-01 6.11571968e-01 8.58880341e-01
-9.41060066e-01 -9.75998819e-01 -2.39622250e-01 2.09618241e-01
7.19592571e-01 3.04414392e-01 -1.11251630e-01 8.38432431e-01
2.82019645e-01 1.01793075e+00 2.01449990e-01 -4.71126735e-01
5.90390682e-01 -1.10563844e-01 1.24615081e-01 -2.58222163e-01
3.66971701e-01 -1.88134313e-01 2.56423414e-01 -9.82575834e-01
1.76182330e-01 -6.46678448e-01 -1.69135118e+00 -7.67469108e-01
1.49388716e-01 -1.22709945e-01 6.77977264e-01 7.21637785e-01
6.15366459e-01 3.71201783e-01 8.88968229e-01 -8.25803339e-01
2.70706683e-01 -2.37979963e-01 -5.02190769e-01 2.62632668e-01
5.75946569e-01 -8.40910137e-01 -2.92737365e-01 1.40414713e-02] | [8.103949546813965, -1.5379945039749146] |
ed1cc550-4972-4ed3-9933-f437354b911a | og-sgg-ontology-guided-scene-graph-generation | 2202.10201 | null | https://arxiv.org/abs/2202.10201v3 | https://arxiv.org/pdf/2202.10201v3.pdf | OG-SGG: Ontology-Guided Scene Graph Generation. A Case Study in Transfer Learning for Telepresence Robotics | Scene graph generation from images is a task of great interest to applications such as robotics, because graphs are the main way to represent knowledge about the world and regulate human-robot interactions in tasks such as Visual Question Answering (VQA). Unfortunately, its corresponding area of machine learning is still relatively in its infancy, and the solutions currently offered do not specialize well in concrete usage scenarios. Specifically, they do not take existing "expert" knowledge about the domain world into account; and that might indeed be necessary in order to provide the level of reliability demanded by the use case scenarios. In this paper, we propose an initial approximation to a framework called Ontology-Guided Scene Graph Generation (OG-SGG), that can improve the performance of an existing machine learning based scene graph generator using prior knowledge supplied in the form of an ontology (specifically, using the axioms defined within); and we present results evaluated on a specific scenario founded in telepresence robotics. These results show quantitative and qualitative improvements in the generated scene graphs. | ['Luis Merino', 'Natalia Díaz-Rodríguez', 'Fernando Caballero', 'Fernando Amodeo'] | 2022-02-21 | null | null | null | null | ['scene-graph-generation'] | ['computer-vision'] | [ 3.59990120e-01 8.57849419e-01 2.71500777e-02 -3.88871789e-01
5.72079141e-03 -3.85861456e-01 8.84987652e-01 4.73075986e-01
-3.28161865e-01 6.78236663e-01 8.83835405e-02 -3.89995694e-01
-3.42074603e-01 -1.11141455e+00 -7.37260699e-01 -1.91671252e-01
7.96466880e-03 8.80500019e-01 5.72175920e-01 -6.85563326e-01
2.63423473e-01 7.13450432e-01 -2.12816334e+00 3.64392966e-01
8.31865489e-01 8.13746154e-01 4.04466063e-01 3.63055676e-01
-3.93405676e-01 1.12552750e+00 -4.64700669e-01 -1.79408103e-01
9.31036845e-02 -4.98106062e-01 -1.18769383e+00 5.22307992e-01
1.51456356e-01 -8.83250013e-02 -6.57731667e-02 1.10506845e+00
-1.74555734e-01 2.47358426e-01 6.60855293e-01 -1.41862190e+00
-1.26477525e-01 3.77730817e-01 1.60521135e-01 -1.98267713e-01
6.31383777e-01 -1.13340942e-02 7.14608908e-01 -1.99749902e-01
1.00138867e+00 1.31988347e+00 9.57411528e-02 5.83906054e-01
-1.05487764e+00 1.06141374e-01 2.76544750e-01 4.34738755e-01
-1.13349462e+00 -2.86879420e-01 7.96758294e-01 -5.84727764e-01
8.34758222e-01 1.92404538e-01 6.85397625e-01 6.61985755e-01
-6.81643635e-02 7.28140116e-01 1.01232755e+00 -8.43035579e-01
6.19821906e-01 5.09933889e-01 -1.77517176e-01 5.89602172e-01
3.90636176e-01 -1.19464025e-01 -3.45899194e-01 1.33018285e-01
8.33982468e-01 -3.83384287e-01 -2.74633318e-01 -1.20538104e+00
-1.20004702e+00 8.12959254e-01 7.35853434e-01 5.91432750e-01
-3.02292854e-01 -4.54176143e-02 3.65189344e-01 4.32994664e-01
1.87586904e-01 7.28386283e-01 -1.77146360e-01 5.60688495e-04
-3.63051116e-01 4.01405036e-01 1.00859153e+00 1.14327240e+00
1.04070461e+00 1.56863760e-02 2.92335689e-01 4.10971075e-01
2.02096894e-01 2.37035215e-01 2.58257240e-01 -1.00780380e+00
2.14792669e-01 1.22904706e+00 2.85739332e-01 -9.60794032e-01
-5.72434306e-01 -5.23994956e-03 -3.51290047e-01 4.82010543e-01
5.02969861e-01 7.15746656e-02 -8.66918862e-01 1.36857641e+00
4.35099125e-01 -2.30938345e-01 2.29148060e-01 1.02529728e+00
8.86678040e-01 4.03779984e-01 7.24662617e-02 -2.09181923e-02
1.21934760e+00 -6.63690984e-01 -6.27995133e-01 -5.77497482e-01
8.81225824e-01 -3.15008521e-01 1.23640990e+00 4.19115633e-01
-7.25401342e-01 -4.76780385e-01 -9.61400747e-01 -2.58712005e-02
-8.06782603e-01 -1.13121606e-01 8.07184637e-01 5.95585287e-01
-1.02217579e+00 1.80716947e-01 -5.10770380e-01 -9.26928580e-01
1.11020371e-01 2.39432529e-01 -5.48945248e-01 -2.71374375e-01
-1.12255442e+00 1.39912677e+00 9.12367284e-01 1.48515493e-01
-7.69114673e-01 -6.67048022e-02 -1.20625019e+00 -1.82732195e-02
8.35219741e-01 -6.46747231e-01 1.32184458e+00 -1.11289942e+00
-1.45267260e+00 1.01503408e+00 1.52295321e-01 -6.93519413e-01
5.74877739e-01 -7.35628679e-02 -4.32163030e-02 2.30501294e-01
-6.24620495e-03 9.31940615e-01 4.28849012e-01 -1.53441405e+00
-5.96718311e-01 -2.66447216e-01 9.18826044e-01 4.10815001e-01
-4.40877303e-03 -1.34712189e-01 -4.03353274e-01 6.76305443e-02
8.16975087e-02 -8.37599158e-01 -5.12185454e-01 4.74561453e-02
-1.55640870e-01 -6.18689917e-02 7.14074194e-01 -3.23288918e-01
6.35157347e-01 -1.94576347e+00 1.84203580e-01 4.47545379e-01
-1.99427322e-01 2.59065896e-01 1.19809844e-01 6.84474409e-01
2.64258254e-02 -2.29878172e-01 -2.49407083e-01 9.82975438e-02
1.90362200e-01 6.05964899e-01 -6.97409511e-02 2.23787427e-01
2.26519387e-02 6.28057599e-01 -1.10835207e+00 -5.54448783e-01
8.41369987e-01 9.90817845e-02 -6.08821332e-01 2.02344134e-01
-8.14017594e-01 4.52994227e-01 -5.71043313e-01 7.42727071e-02
1.14454769e-01 -7.10696876e-02 5.34116507e-01 -5.48848175e-02
3.80755216e-03 1.19023688e-01 -1.31140459e+00 1.85535109e+00
-5.25912762e-01 5.72511554e-01 -2.79679537e-01 -1.18217289e+00
1.12240279e+00 2.40833119e-01 4.55198497e-01 -7.49626696e-01
2.45377988e-01 2.36440683e-03 1.07014455e-01 -7.39439666e-01
4.89330649e-01 -1.50159195e-01 -1.10092834e-01 1.24789350e-01
-4.13244776e-03 -7.07165718e-01 5.93832433e-01 2.27368101e-01
9.23238397e-01 5.59587181e-01 6.83148086e-01 -1.59641355e-01
7.25910008e-01 6.91191137e-01 2.52133105e-02 4.33178633e-01
1.01965725e-01 4.75568473e-01 7.99192727e-01 -5.24284184e-01
-1.11336887e+00 -6.33352399e-01 2.25290388e-01 6.11021280e-01
3.17680031e-01 -3.77305299e-01 -9.42079425e-01 -5.90420783e-01
-3.75637949e-01 1.00679612e+00 -3.91716212e-01 -7.72215277e-02
-2.26862356e-01 -7.16836229e-02 8.98479149e-02 1.37527645e-01
4.66147244e-01 -1.60926092e+00 -1.32277763e+00 1.73562884e-01
-1.19035751e-01 -1.66487360e+00 4.90151137e-01 1.09750812e-03
-7.73614943e-01 -1.43932438e+00 -1.76028639e-01 -6.70126081e-01
9.01443720e-01 1.16616346e-01 1.06266499e+00 3.56336944e-02
-9.19820592e-02 7.06694305e-01 -8.12259674e-01 -6.34725392e-01
-7.11547136e-01 -5.18146455e-02 -4.37990338e-01 -1.70241650e-02
3.39995682e-01 -3.31131130e-01 -5.28310612e-02 1.60853952e-01
-1.24874616e+00 2.80535877e-01 5.00217736e-01 4.60038841e-01
1.76293731e-01 3.88776034e-01 2.15788394e-01 -1.15564537e+00
4.72806096e-01 -1.32804409e-01 -8.00331831e-01 2.62393206e-01
-2.68551797e-01 1.55492023e-01 4.67257380e-01 -8.45546275e-02
-1.21033478e+00 3.49995315e-01 3.14391442e-02 -1.83081433e-01
-7.19862103e-01 7.15824604e-01 -5.55031538e-01 -2.62381602e-03
9.02291715e-01 -4.97955941e-02 7.14316890e-02 -2.33850643e-01
6.27286613e-01 4.37335849e-01 4.29451942e-01 -5.52962363e-01
6.69882834e-01 5.45363724e-01 2.26376295e-01 -8.11204135e-01
-4.40280855e-01 -5.45769036e-01 -7.30652869e-01 -4.30825621e-01
7.43200481e-01 -6.04840696e-01 -6.21155143e-01 6.75317496e-02
-1.20250106e+00 -5.21254838e-01 -4.73784208e-01 3.62185359e-01
-1.00227702e+00 2.98967451e-01 8.24848786e-02 -9.71251547e-01
2.90946662e-01 -1.13577378e+00 1.11346662e+00 1.18177518e-01
-2.24904045e-01 -9.97109532e-01 -2.05664083e-01 3.19992244e-01
1.50153041e-01 5.24526834e-01 1.08963573e+00 -5.57760954e-01
-6.89953625e-01 -1.38280883e-01 -1.97931454e-01 4.00514454e-01
1.92177653e-01 -2.89411455e-01 -8.47544670e-01 8.72295871e-02
-3.67843658e-01 -2.46795952e-01 1.30252510e-01 1.25413775e-01
7.68073618e-01 -1.41803414e-01 -1.97069839e-01 -1.63934141e-01
1.58353829e+00 2.29511872e-01 9.51360703e-01 5.68452537e-01
4.73051131e-01 1.31747246e+00 9.28446591e-01 3.17962706e-01
5.30369222e-01 9.63463962e-01 8.18954229e-01 -6.70908839e-02
-1.97348744e-01 -3.74138296e-01 1.89135522e-01 4.48543616e-02
-2.15217426e-01 -1.34624392e-01 -1.22699654e+00 8.00985873e-01
-2.18937683e+00 -8.38391900e-01 -3.78245890e-01 2.20505714e+00
2.37756595e-01 2.13075459e-01 5.69964834e-02 3.40218574e-01
5.81887305e-01 -1.05660103e-01 -2.70031899e-01 -5.64291179e-01
1.33325741e-01 5.80648752e-03 3.26242447e-01 4.67447132e-01
-8.47242594e-01 1.08825707e+00 5.54840279e+00 3.16982985e-01
-9.29326892e-01 -1.05695710e-01 -2.17862613e-02 6.64115965e-01
-2.48952389e-01 4.52590644e-01 -2.09955305e-01 1.17656207e-02
7.52524257e-01 -1.40036434e-01 3.64584148e-01 1.17921114e+00
3.95089269e-01 -4.07308131e-01 -1.06407702e+00 8.86012554e-01
1.80668741e-01 -1.14547992e+00 1.86380491e-01 1.47795632e-01
3.73582929e-01 -2.45644733e-01 -5.34618676e-01 2.06828222e-01
2.98611641e-01 -8.45372140e-01 8.67160976e-01 4.81915861e-01
3.17948282e-01 -6.08100057e-01 8.99000168e-01 7.42934465e-01
-8.74506533e-01 -2.87072007e-02 -4.18971479e-01 -2.61669993e-01
3.23489577e-01 4.26746011e-01 -1.34097838e+00 9.12635326e-01
5.99659443e-01 2.83858597e-01 -6.78220809e-01 1.25377846e+00
-4.72020239e-01 7.85065815e-02 -2.06221029e-01 -2.31884211e-01
1.60683140e-01 -1.35180980e-01 5.32082558e-01 8.67700398e-01
1.30768031e-01 -9.36677381e-02 2.53970593e-01 7.19344974e-01
3.48935604e-01 3.08195949e-01 -1.10018790e+00 9.74993221e-04
-2.32696850e-02 1.08797204e+00 -9.97224212e-01 -3.31875205e-01
-3.72971833e-01 6.49195135e-01 2.03762531e-01 3.45154166e-01
-5.30160964e-01 -3.51175636e-01 1.24645032e-01 4.50527132e-01
8.77487510e-02 -3.02285910e-01 1.25081852e-01 -1.03652096e+00
5.82309254e-02 -9.60091412e-01 2.71072507e-01 -1.34862459e+00
-7.72577286e-01 6.35997415e-01 5.01352668e-01 -1.37709248e+00
-7.58259654e-01 -9.99097466e-01 -8.15541893e-02 3.69721085e-01
-1.52205253e+00 -1.39322853e+00 -6.65408194e-01 5.92027843e-01
3.69979352e-01 1.13891214e-01 9.13110554e-01 -6.99351309e-03
2.08039522e-01 -2.48179123e-01 -6.48293436e-01 -3.20260748e-02
3.38542193e-01 -1.31186426e+00 1.40722483e-01 8.26800704e-01
3.43074411e-01 4.06697154e-01 1.06643617e+00 -3.98764729e-01
-1.23269892e+00 -9.39006567e-01 8.34650755e-01 -3.62059891e-01
4.94044244e-01 -3.94750655e-01 -6.29119277e-01 7.59387612e-01
-8.31562001e-03 -6.87809289e-02 2.06259534e-01 9.76402909e-02
-3.31867002e-02 -5.53387962e-02 -1.20909166e+00 6.09582961e-01
1.11689579e+00 -4.35084313e-01 -7.98356414e-01 4.36062127e-01
4.69291508e-01 -5.12514949e-01 -6.04002774e-01 6.03094399e-01
1.79325238e-01 -1.21288693e+00 6.02754951e-01 -6.38008535e-01
1.38328642e-01 -6.80872500e-01 -2.00611055e-01 -1.18579531e+00
1.69180349e-01 -3.38100761e-01 2.00510889e-01 9.37186062e-01
3.60608011e-01 -5.39921999e-01 9.97506082e-01 6.75561011e-01
-3.80354851e-01 -1.82625890e-01 -6.40473425e-01 -6.47521257e-01
-4.49915111e-01 -5.60878396e-01 7.38820553e-01 7.93599308e-01
1.83210388e-01 3.33134025e-01 -3.16442214e-02 2.02069119e-01
2.52993345e-01 5.05379476e-02 1.31290567e+00 -1.36407268e+00
-1.81174532e-01 -2.28984490e-01 -1.13481939e+00 -7.43966520e-01
2.60636061e-01 -5.54808378e-01 2.74252206e-01 -2.20258403e+00
-3.37329596e-01 -3.94589603e-01 1.23369500e-01 5.64029098e-01
2.73085058e-01 -8.75792950e-02 2.84846395e-01 -1.42107546e-01
-6.51624680e-01 4.62140441e-01 1.27062607e+00 -6.29802495e-02
-7.02115223e-02 -1.98090404e-01 -5.02928019e-01 8.93338323e-01
6.84064686e-01 -1.34974167e-01 -9.94234920e-01 -2.27324665e-01
4.73879755e-01 -9.42464992e-02 5.53093612e-01 -1.17657268e+00
2.43880972e-01 -4.06484932e-01 -2.92283803e-01 -4.44309935e-02
3.52352113e-01 -1.25255430e+00 3.73324454e-01 3.51953059e-01
-3.10737863e-02 -2.72527099e-01 1.12262577e-01 2.48958960e-01
-4.69792902e-01 -5.31670153e-01 4.78583157e-01 -3.92520398e-01
-1.48544133e+00 -2.60029048e-01 -3.98208171e-01 -3.43845248e-01
1.29830945e+00 -4.52118874e-01 -2.68602371e-01 -6.67405128e-01
-6.57173336e-01 9.70802978e-02 8.07186127e-01 5.25998831e-01
5.08586228e-01 -9.48606789e-01 -2.86353171e-01 -2.19083093e-02
7.45878935e-01 1.52397409e-01 1.04785569e-01 3.91584009e-01
-8.49133670e-01 5.06840050e-01 -4.45611626e-01 -5.08108258e-01
-9.29288566e-01 7.78226972e-01 3.79014254e-01 -1.50775850e-01
-6.05497062e-01 1.85441151e-01 2.47038469e-01 -6.38567805e-01
1.15127340e-01 -4.64298815e-01 -6.57433987e-01 -2.86743194e-01
2.64470577e-01 -2.10282076e-02 2.57129610e-01 -7.77865767e-01
-1.58823177e-01 3.58775347e-01 4.59479213e-01 -1.27228081e-01
1.16145551e+00 -4.15775292e-02 -9.16227177e-02 3.54018599e-01
4.60229784e-01 -2.05198735e-01 -8.74774754e-01 -4.43181442e-03
4.81538385e-01 -3.82844478e-01 -1.87933147e-01 -5.26277542e-01
-6.52350307e-01 7.77360976e-01 2.54678190e-01 5.50435603e-01
1.11150038e+00 2.85767645e-01 1.60971448e-01 8.32244635e-01
1.06030703e+00 -1.03430450e+00 -8.36541057e-02 2.85877913e-01
1.00710440e+00 -1.06285357e+00 6.74795508e-02 -8.50545824e-01
-8.69791985e-01 1.14640701e+00 6.81212664e-01 -5.76624423e-02
3.33597302e-01 -4.19700220e-02 1.37233824e-01 -5.07092953e-01
-4.51772422e-01 -8.89917374e-01 1.84459448e-01 1.00537694e+00
8.50077793e-02 -1.07892066e-01 -2.76630461e-01 -8.88603702e-02
-1.93910316e-01 1.30503371e-01 7.05270529e-01 1.11423469e+00
-8.23751986e-01 -1.36052573e+00 -2.23301977e-01 1.37978867e-01
1.32752791e-01 3.64343703e-01 -6.54929340e-01 1.22291887e+00
3.87287587e-01 1.10142791e+00 -8.26382861e-02 -1.00195929e-01
7.40358651e-01 -1.11972310e-01 7.94608712e-01 -1.04152679e+00
-2.08653405e-01 -3.49036574e-01 6.54659688e-01 -5.51114082e-01
-9.83461320e-01 -5.11670887e-01 -1.39293432e+00 3.59417312e-02
-1.47633061e-01 3.15643638e-01 8.80991161e-01 1.17801344e+00
9.22804773e-02 3.45920444e-01 1.01077631e-01 -7.98058271e-01
6.59432635e-02 -7.95448482e-01 -3.74086320e-01 7.47041047e-01
-1.77720040e-01 -8.44180346e-01 -5.67110218e-02 4.14428711e-01] | [4.726789951324463, 0.860319197177887] |
6fb33d9d-8fd3-4d69-b155-deb6c46b1223 | effective-deep-learning-models-for-automatic | null | null | https://ieeexplore.ieee.org/document/9274427 | https://ieeexplore.ieee.org/document/9274427 | Effective Deep Learning Models for Automatic Diacritization of Arabic Text | While building a text-to-speech system for the Arabic language, we found that the system synthesized speeches with many pronunciation errors. The primary source of these errors is the lack of diacritics in modern standard Arabic writing. These diacritics are small strokes that appear above or below each letter to provide pronunciation and grammatical information. We propose three deep learning models to recover Arabic text diacritics based on our work in a text-to-speech synthesis system using deep learning. The first model is a baseline model used to test how a simple deep learning model performs on the corpora. The second model is based on an encoder-decoder architecture, which resembles our text-to-speech synthesis model with many modifications to suit this problem. The last model is based on the encoder part of the text-to-speech model, which achieves state-of-the-art performances in both word error rate and diacritic error rate metrics. These models will benefit a wide range of natural language processing applications such as text-to-speech, part-of-speech tagging, and machine translation. | ['Ali Mustafa Qamar', 'Mokthar Ali Hasan Madhfar'] | 2020-11-01 | null | null | null | null | ['arabic-text-diacritization'] | ['natural-language-processing'] | [ 5.17874807e-02 1.51807457e-01 1.82114001e-02 -3.74689102e-01
-8.60350609e-01 -5.46985209e-01 6.58952713e-01 -7.10362270e-02
-3.73136550e-01 5.75422466e-01 4.89791244e-01 -7.89816797e-01
6.35448456e-01 -7.34365404e-01 -7.21617341e-01 -4.91739333e-01
3.51966232e-01 6.70901656e-01 1.76558346e-01 -8.35961819e-01
9.75077525e-02 3.05611044e-01 -9.40015256e-01 8.63871157e-01
9.62114632e-01 4.86565232e-01 2.58845866e-01 8.18533659e-01
-2.99837470e-01 9.49884295e-01 -1.20085573e+00 -7.24754035e-01
7.98679516e-02 -6.78309798e-01 -8.08021188e-01 -6.27990812e-02
1.16991654e-01 -5.74824035e-01 -4.99547213e-01 8.14551353e-01
5.12467682e-01 -2.84247119e-02 7.26846755e-01 -7.68011272e-01
-1.05285203e+00 1.04024887e+00 -2.04417817e-02 4.17546220e-02
3.64679098e-01 -7.41934404e-02 1.09440684e+00 -1.27689481e+00
4.02997673e-01 1.48478687e+00 4.54717308e-01 6.71054780e-01
-7.57490695e-01 -2.89190233e-01 -1.84325978e-01 -3.63785513e-02
-1.17055547e+00 -7.96505570e-01 3.27136576e-01 -1.24820158e-01
1.50900447e+00 4.05511148e-02 2.02030376e-01 1.18874121e+00
5.38182676e-01 1.17371213e+00 7.17006087e-01 -9.99857545e-01
2.06291489e-02 -8.33370760e-02 -1.05560541e-01 8.40773880e-01
-2.26202726e-01 -1.07175328e-01 -5.55022538e-01 1.72042117e-01
5.53432643e-01 -3.32394451e-01 4.16863076e-02 5.14083624e-01
-1.17838585e+00 1.06446970e+00 9.24524665e-02 3.10384005e-01
-1.10771552e-01 -9.59818531e-03 5.70616484e-01 5.72678745e-01
3.13697189e-01 3.75310183e-01 -6.50187135e-01 -3.97212625e-01
-8.31903875e-01 -7.14634359e-02 1.05503154e+00 8.96498859e-01
3.06074858e-01 5.77207446e-01 -1.95438951e-01 1.22918379e+00
5.19015193e-01 8.54934335e-01 8.49462986e-01 -6.16460085e-01
7.18592465e-01 2.88289428e-01 -5.32278344e-02 -4.43534911e-01
-2.04859301e-01 -1.39692528e-02 -4.35050607e-01 1.38163850e-01
5.52567184e-01 -5.56991458e-01 -1.16306782e+00 1.29806304e+00
-3.49583983e-01 -5.97632349e-01 4.46362883e-01 7.35517323e-01
9.03631270e-01 1.38364291e+00 -2.12775320e-01 4.78436276e-02
1.44181752e+00 -1.48157191e+00 -1.08132064e+00 -6.57476306e-01
9.01051998e-01 -1.33076048e+00 1.40241742e+00 2.01694831e-01
-1.37188637e+00 -4.98211980e-01 -1.34956992e+00 -3.19147021e-01
-6.32592976e-01 6.85234666e-01 1.80747017e-01 7.81694591e-01
-1.08185136e+00 4.40543503e-01 -9.56614912e-01 -3.79269123e-01
-2.18503490e-01 1.49495915e-01 -2.76833922e-01 2.00768203e-01
-1.41782379e+00 1.39951956e+00 3.27328265e-01 3.73510756e-02
-6.19173169e-01 9.95533392e-02 -1.08632338e+00 2.15152726e-01
-6.38945252e-02 -1.47627309e-01 1.78918350e+00 -1.27815568e+00
-2.20638323e+00 8.10887337e-01 -4.28067923e-01 -4.36069608e-01
1.03742912e-01 -1.64600521e-01 -7.28114605e-01 -7.37292273e-03
-1.39089022e-02 5.06662250e-01 9.09499526e-01 -9.61381018e-01
-6.80947900e-01 -9.12643299e-02 -3.16453695e-01 2.78852463e-01
-7.30087906e-02 5.74944496e-01 -1.67325735e-01 -1.08076429e+00
-8.04703236e-02 -1.02908731e+00 2.29359090e-01 -4.43928957e-01
-3.29392940e-01 -3.64764094e-01 7.32343733e-01 -1.21949077e+00
1.39159954e+00 -1.83275127e+00 -5.59443980e-02 -1.51903138e-01
-4.90590125e-01 7.32750952e-01 -4.75018620e-01 1.00771689e+00
-1.21480413e-02 -1.63595565e-02 -4.27129269e-01 -4.36433345e-01
9.65411291e-02 3.72754961e-01 -6.02289379e-01 1.64888188e-01
7.88624406e-01 9.20770347e-01 -9.67347503e-01 -1.97423901e-03
6.70691729e-02 3.34990412e-01 -1.66824341e-01 3.03250402e-01
-3.03260624e-01 -5.64399883e-02 1.24416515e-01 7.97067225e-01
3.81500900e-01 1.79418474e-01 1.62795454e-01 2.75132477e-01
-9.98474658e-02 1.42945111e+00 -5.09040177e-01 1.49177599e+00
-5.98532856e-01 8.48309100e-01 -4.74218167e-02 -8.87486994e-01
9.92543578e-01 6.10842526e-01 -2.16928214e-01 -8.14178348e-01
2.64137506e-01 7.91478574e-01 4.28717315e-01 -2.65565962e-01
8.56758595e-01 -8.03164318e-02 -2.09093764e-01 6.09662056e-01
3.84146065e-01 -3.87472540e-01 4.37809110e-01 1.18023574e-01
8.94066691e-01 1.18946157e-01 4.14511740e-01 -1.47283822e-01
4.58166689e-01 7.67197609e-02 1.64661303e-01 4.16636825e-01
7.39335790e-02 9.11558688e-01 5.21555185e-01 -4.17281181e-01
-1.18106413e+00 -1.07098889e+00 3.51395905e-02 1.50202668e+00
-4.90038663e-01 -5.25482595e-01 -1.00456321e+00 -8.11743259e-01
-4.59753335e-01 1.07223535e+00 -3.30396414e-01 -1.02149233e-01
-9.21404302e-01 -6.42071307e-01 1.05724168e+00 5.04424870e-01
2.19019428e-01 -1.60252738e+00 3.96124274e-02 5.02205968e-01
-3.59601378e-01 -1.00454438e+00 -6.74853384e-01 4.75984603e-01
-3.96291167e-01 -6.29762471e-01 -7.78884828e-01 -1.40087712e+00
4.67589468e-01 -6.10598400e-02 1.05887210e+00 8.18244740e-02
3.43998462e-01 -2.66989291e-01 -8.46008539e-01 -6.49661243e-01
-1.40720928e+00 3.21541578e-01 -1.06399031e-02 -1.81960687e-01
7.82690585e-01 8.20425078e-02 1.46641225e-01 2.27133602e-01
-9.32842255e-01 -1.21829547e-01 6.79002941e-01 9.96651173e-01
1.14250325e-01 -5.10524154e-01 6.17813110e-01 -7.36615956e-01
8.86510432e-01 -2.92318165e-01 -4.12520021e-01 2.28366658e-01
-1.48108423e-01 2.14114919e-01 1.00767446e+00 -1.73306346e-01
-8.36559772e-01 6.28689900e-02 -1.00130332e+00 3.59594077e-01
1.24816503e-03 7.49628305e-01 -2.89989620e-01 4.99837160e-01
7.53009200e-01 4.14115012e-01 2.27637380e-01 -3.90386194e-01
3.04912239e-01 1.41875589e+00 5.06666720e-01 -3.48561585e-01
3.92631412e-01 -1.54212385e-01 -5.45350671e-01 -1.09325910e+00
-8.40761900e-01 -1.63982704e-01 -6.08767152e-01 3.24143857e-01
6.57704175e-01 -9.65416670e-01 -3.38293046e-01 7.11338878e-01
-1.65289652e+00 -6.50955439e-01 -1.26045972e-01 2.94252932e-01
-3.88205916e-01 4.23947603e-01 -1.11563587e+00 -4.07401741e-01
-3.83782059e-01 -1.36154187e+00 1.25699270e+00 3.67100239e-02
-2.22627833e-01 -1.01646721e+00 2.20499896e-02 -2.34704074e-02
4.43964869e-01 -5.84664345e-01 1.09070420e+00 -8.89151156e-01
2.92089712e-02 -6.29328489e-02 -1.49644047e-01 7.43261218e-01
3.53124082e-01 1.86801642e-01 -8.85136485e-01 -2.63522804e-01
-2.49346435e-01 -3.72286975e-01 7.49233603e-01 1.41144797e-01
6.05188012e-01 -5.78855276e-01 2.18550965e-01 3.07442397e-01
7.73235738e-01 5.29974461e-01 1.10925162e+00 3.55636358e-01
5.73589981e-01 4.41333950e-01 3.48635733e-01 1.57850906e-01
5.68857253e-01 6.22281730e-01 2.74493337e-01 -4.04337309e-02
-4.49428052e-01 -3.21954310e-01 1.15560341e+00 1.64735007e+00
4.26832199e-01 -6.85903013e-01 -1.03174591e+00 6.99002504e-01
-1.62706721e+00 -7.22375035e-01 -2.56127477e-01 1.89958322e+00
1.22545612e+00 1.72084980e-02 -5.95296323e-02 8.72640833e-02
5.12308717e-01 1.95467323e-01 4.19166461e-02 -1.31116080e+00
-2.46397823e-01 5.12284577e-01 3.71640354e-01 8.31377625e-01
-9.80852544e-01 1.68766475e+00 6.54505634e+00 8.47803116e-01
-1.30434072e+00 5.43272719e-02 3.12592685e-01 5.44796765e-01
-1.89538464e-01 -9.69539210e-02 -1.09625649e+00 4.45131958e-01
1.48507214e+00 1.87601477e-01 6.27284944e-01 5.17627239e-01
1.37432069e-01 1.20771050e-01 -1.00701857e+00 5.09913266e-01
3.08274984e-01 -1.25212610e+00 4.98355567e-01 -3.76105100e-01
6.58724427e-01 1.91272750e-01 -1.05979770e-01 5.28915763e-01
6.15889490e-01 -1.18057978e+00 8.69342148e-01 8.41682032e-02
9.35231626e-01 -7.30537057e-01 8.59798074e-01 1.95134595e-01
-7.74600744e-01 1.78113565e-01 -5.06066084e-01 -1.88163668e-01
1.23061344e-01 1.92713320e-01 -1.30461943e+00 1.60630360e-01
2.32854024e-01 5.01961052e-01 -4.18954253e-01 6.32795155e-01
-7.44122565e-01 1.05539930e+00 -4.16502208e-02 -4.68398452e-01
6.59918129e-01 -2.35087171e-01 4.22031432e-01 1.71481180e+00
5.55750370e-01 -2.25544661e-01 -4.18884009e-02 4.43934292e-01
-2.72879034e-01 3.36997300e-01 -4.02443975e-01 -4.71012890e-01
5.91586053e-01 8.68279874e-01 -6.87847793e-01 -4.06415373e-01
-6.02500558e-01 1.27888572e+00 3.20280731e-01 2.16825381e-01
-4.69063878e-01 -1.05106997e+00 5.77436984e-01 -9.73209646e-03
3.82132232e-01 -6.11520290e-01 -3.75908375e-01 -1.06488168e+00
-8.39429349e-03 -1.48948991e+00 -1.01511069e-01 -7.47800708e-01
-1.39476228e+00 9.69902515e-01 -5.63769400e-01 -1.15280426e+00
-8.06907177e-01 -1.07747817e+00 -5.71060061e-01 1.15516627e+00
-1.44764388e+00 -1.30810642e+00 2.31336713e-01 2.42469311e-01
1.04018235e+00 -8.28769922e-01 1.27315569e+00 1.59180120e-01
-3.35105151e-01 7.45484710e-01 3.37021679e-01 8.27257156e-01
1.05143249e+00 -1.27941179e+00 1.30316842e+00 1.26033998e+00
2.61093259e-01 5.57557225e-01 3.74902010e-01 -5.62358022e-01
-1.18955469e+00 -9.83988822e-01 1.62957609e+00 -5.65012693e-01
8.29081595e-01 -6.13879144e-01 -9.08570290e-01 8.20194423e-01
6.48634553e-01 -2.43216842e-01 5.53061068e-01 -7.99288675e-02
-1.04624771e-01 1.64396673e-01 -6.65877938e-01 9.64538813e-01
4.04186010e-01 -6.47273540e-01 -9.62687135e-01 4.21849489e-01
9.99284625e-01 -6.25439882e-01 -3.58297378e-01 -6.13596737e-02
4.86673206e-01 -5.21875203e-01 4.67718601e-01 -6.54837489e-01
8.70251238e-01 -2.04845563e-01 -3.82263511e-01 -1.96478605e+00
-1.71540543e-01 -9.44416046e-01 -4.56665941e-02 9.50345218e-01
9.41030502e-01 -4.90856647e-01 2.73623049e-01 5.37452549e-02
-8.60907912e-01 -3.41858387e-01 -7.56222069e-01 -7.84169018e-01
5.56176484e-01 -1.44840062e-01 6.02604210e-01 8.13993096e-01
2.55599618e-01 7.92651772e-01 -2.33716652e-01 3.85725610e-02
-2.96557307e-01 -3.54317665e-01 5.53058922e-01 -8.86489391e-01
-2.37149671e-01 -5.01099229e-01 6.36060014e-02 -1.52778220e+00
3.29020441e-01 -9.33541656e-01 5.36468923e-01 -1.73731685e+00
-5.48547745e-01 -1.93345711e-01 3.38588923e-01 8.09258759e-01
-6.79204017e-02 1.38832703e-01 2.51663417e-01 -1.17130697e-01
-3.47874984e-02 7.10294485e-01 1.05793834e+00 -1.81358874e-01
-3.14170569e-01 1.75855253e-02 -4.32768703e-01 6.29046857e-01
8.56901228e-01 -4.99674767e-01 1.25486225e-01 -1.02660954e+00
1.59990534e-01 1.24705516e-01 -3.18722457e-01 -6.79129601e-01
5.52530326e-02 -1.33412350e-02 1.52061060e-01 -5.68959117e-01
3.35138857e-01 -4.29230273e-01 -7.96611845e-01 4.61082011e-01
-1.68619215e-01 5.24422050e-01 3.35265040e-01 -3.08605134e-01
-5.24713039e-01 -5.97738206e-01 8.91355336e-01 -1.59182057e-01
-6.19261920e-01 -1.37742266e-01 -1.21943486e+00 1.11493409e-01
3.86512130e-01 1.93607472e-02 -6.52288735e-01 -4.96568769e-01
-5.04688561e-01 -2.70834565e-01 2.53343850e-01 7.59036541e-01
6.24538124e-01 -1.23223579e+00 -1.15571237e+00 6.28691077e-01
1.04286067e-01 -4.29928273e-01 -7.40810752e-01 3.96994144e-01
-1.18453634e+00 7.91912138e-01 -3.19123060e-01 -2.10006088e-01
-8.85347903e-01 2.96449903e-02 4.12011236e-01 -1.17845669e-01
-2.72170037e-01 8.18217695e-01 -1.62298784e-01 -8.38553905e-01
2.33108342e-01 -5.51672876e-01 -1.90677512e-02 -1.26218170e-01
7.03512788e-01 2.19955280e-01 7.18749225e-01 -8.90034556e-01
-3.51579845e-01 -6.95484802e-02 -2.39197284e-01 -5.39388418e-01
1.14342499e+00 -5.36995195e-02 -1.58064663e-01 5.04919171e-01
9.82036352e-01 4.58967924e-01 -9.86523986e-01 -1.27010927e-01
-6.80068694e-03 1.23120088e-03 -7.44228065e-02 -1.09233665e+00
-6.28516912e-01 1.29196870e+00 7.24117979e-02 2.59444803e-01
6.78690970e-01 -5.14741540e-01 1.30524218e+00 8.78949404e-01
6.22037612e-02 -1.43021321e+00 8.06966871e-02 1.72328818e+00
9.71694231e-01 -1.29906321e+00 -5.68745255e-01 4.79263701e-02
-9.33344424e-01 1.49303615e+00 4.30497825e-01 -2.02740967e-01
5.00672281e-01 6.00506365e-01 5.14287114e-01 2.49002978e-01
-6.20143950e-01 -2.73742884e-01 2.93870032e-01 7.09613860e-01
1.08573937e+00 1.99269161e-01 -3.24613154e-01 5.95360637e-01
-6.60848975e-01 -4.78474885e-01 9.65629935e-01 8.59628081e-01
-5.67119956e-01 -1.42961586e+00 -3.19498897e-01 1.95548818e-01
-5.96334219e-01 -6.37948930e-01 -7.51935422e-01 5.38942516e-01
-3.48773271e-01 1.31288731e+00 2.62083739e-01 -2.76159942e-01
3.14342469e-01 3.46922666e-01 3.60217690e-01 -9.58405495e-01
-7.54177868e-01 1.19604662e-01 3.78396124e-01 -1.20956235e-01
1.54696971e-01 -5.33501327e-01 -1.58436739e+00 -3.91598433e-01
-2.89919287e-01 9.92460325e-02 1.02165091e+00 1.07182515e+00
1.55786425e-01 5.01477778e-01 5.85234106e-01 -7.44703114e-01
-7.28683293e-01 -1.43457830e+00 -5.46191931e-01 -3.14676948e-02
4.94472027e-01 -9.35954228e-03 -3.09310313e-02 1.87486604e-01] | [10.888806343078613, 10.294568061828613] |
b1fdd6d6-efd2-4d2d-a950-c014f509fcee | plane-based-content-preserving-warps-for | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Zhou_Plane-Based_Content_Preserving_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Zhou_Plane-Based_Content_Preserving_2013_CVPR_paper.pdf | Plane-Based Content Preserving Warps for Video Stabilization | Recently, a new image deformation technique called content-preserving warping (CPW) has been successfully employed to produce the state-of-the-art video stabilization results in many challenging cases. The key insight of CPW is that the true image deformation due to viewpoint change can be well approximated by a carefully constructed warp using a set of sparsely constructed 3D points only. However, since CPW solely relies on the tracked feature points to guide the warping, it works poorly in large textureless regions, such as ground and building interiors. To overcome this limitation, in this paper we present a hybrid approach for novel view synthesis, observing that the textureless regions often correspond to large planar surfaces in the scene. Particularly, given a jittery video, we first segment each frame into piecewise planar regions as well as regions labeled as non-planar using Markov random fields. Then, a new warp is computed by estimating a single homography for regions belong to the same plane, while inheriting results from CPW in the non-planar regions. We demonstrate how the segmentation information can be efficiently obtained and seamlessly integrated into the stabilization framework. Experimental results on a variety of real video sequences verify the effectiveness of our method. | ['Zihan Zhou', 'Yi Ma', 'Hailin Jin'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['video-stabilization'] | ['computer-vision'] | [ 3.65668148e-01 8.84661302e-02 -7.36062899e-02 1.08911343e-01
-5.39102674e-01 -7.85580099e-01 3.98611009e-01 -2.17397362e-01
1.96042418e-01 5.96873820e-01 -5.59461154e-02 1.77205756e-01
1.77664638e-01 -6.39428616e-01 -1.03894484e+00 -9.56402123e-01
2.59476304e-01 2.93715984e-01 6.46674693e-01 -2.67394215e-01
2.08541095e-01 5.24323225e-01 -1.28779590e+00 -1.17332317e-01
8.54385078e-01 6.50251627e-01 -1.55354291e-02 2.53489494e-01
1.00948162e-01 2.47189060e-01 -2.10621282e-01 -9.71078500e-02
5.47547936e-01 -4.60594058e-01 -5.93669951e-01 8.00521195e-01
6.59532607e-01 -4.40917224e-01 -2.01103643e-01 1.09150910e+00
3.07601877e-02 3.60512465e-01 2.27452204e-01 -9.63831246e-01
-1.26488969e-01 1.17763104e-02 -9.24217761e-01 -3.28316778e-01
6.20188594e-01 -1.15051130e-02 5.33574343e-01 -9.34846044e-01
1.16155446e+00 1.06039035e+00 5.16695678e-01 3.14433217e-01
-1.40551960e+00 -3.63609076e-01 3.86703908e-01 -1.08197071e-01
-1.41654980e+00 -3.64701211e-01 1.35308957e+00 -4.49216634e-01
2.92549968e-01 3.83941680e-01 8.41307521e-01 6.34987056e-01
3.21738303e-01 5.35778344e-01 9.64837909e-01 -5.00278234e-01
1.60014480e-01 -1.90742359e-01 -6.60496429e-02 8.20552349e-01
1.31695703e-01 9.97941568e-03 -3.80695909e-01 -2.77459592e-01
1.22625470e+00 1.93681821e-01 -8.24505150e-01 -1.00366724e+00
-1.50111544e+00 3.21788907e-01 2.24746183e-01 2.45373309e-01
-2.75411636e-01 -4.96729650e-03 7.88593106e-03 -1.02228865e-01
5.66521466e-01 -1.41252633e-02 -1.25832438e-01 3.51735204e-02
-1.06289387e+00 3.27462703e-01 6.03591919e-01 9.80770171e-01
1.01672149e+00 1.52430445e-01 1.50178373e-01 4.83390838e-01
1.67840481e-01 4.99301910e-01 5.33340573e-02 -1.37596047e+00
4.60400343e-01 4.41964835e-01 5.41432261e-01 -1.28546977e+00
2.05827709e-02 -1.02349363e-01 -8.64677370e-01 2.85149097e-01
5.17720878e-01 1.33500189e-01 -7.74862647e-01 1.67026436e+00
8.44195426e-01 3.14799607e-01 -2.84649223e-01 8.48780751e-01
2.50196427e-01 8.07568371e-01 -6.58833444e-01 -6.09237492e-01
1.23259032e+00 -9.65891838e-01 -7.81999886e-01 7.38041699e-02
1.51228756e-01 -8.61065626e-01 6.76255107e-01 3.51479620e-01
-1.30889976e+00 -5.30959010e-01 -8.75433922e-01 8.28329548e-02
3.27157885e-01 -2.87564218e-01 -4.05666083e-02 2.27765635e-01
-1.01081467e+00 6.75150514e-01 -1.07413757e+00 -2.13835567e-01
-4.16990668e-02 2.00852513e-01 -5.29073060e-01 -2.30761200e-01
-6.92174196e-01 5.24969578e-01 6.58350214e-02 1.46198943e-01
-8.17962170e-01 -6.12109065e-01 -8.11300635e-01 -1.15175899e-02
7.16990113e-01 -7.27798402e-01 8.94657910e-01 -1.18018258e+00
-1.80248809e+00 6.89344943e-01 -4.71058547e-01 5.26342410e-05
7.36056745e-01 -2.01678216e-01 2.36169286e-02 4.36183661e-01
-8.80476907e-02 2.75910228e-01 1.11232507e+00 -1.79656327e+00
-2.59164631e-01 -2.88010836e-01 3.12693603e-02 2.62839407e-01
1.02724954e-01 -1.16122149e-01 -7.50919044e-01 -8.20452511e-01
5.87443948e-01 -1.24423265e+00 -1.84287265e-01 2.25636050e-01
-5.19425631e-01 2.47521594e-01 1.11338198e+00 -7.27549493e-01
1.01374710e+00 -2.19148397e+00 5.49965739e-01 3.68278027e-01
2.54328609e-01 1.29502401e-01 1.09046161e-01 4.45378929e-01
-6.87216967e-02 -1.20167486e-01 -5.04245341e-01 -2.76728451e-01
-4.71080124e-01 4.00391638e-01 -4.21577722e-01 7.67252266e-01
1.16855362e-02 5.09923518e-01 -9.84625757e-01 -4.24916744e-01
4.75648105e-01 6.14182949e-01 -6.64793313e-01 1.69069186e-01
-1.06807187e-01 9.65163291e-01 -6.08150780e-01 6.29432261e-01
1.01547945e+00 4.44058888e-02 1.34781271e-01 -3.08401674e-01
-5.28707683e-01 -2.20001966e-01 -1.33220243e+00 1.88482714e+00
-3.89413759e-02 3.06296706e-01 4.57747072e-01 -8.27304304e-01
8.27142537e-01 3.06318730e-01 8.03204596e-01 3.60436291e-02
1.48596615e-01 1.31579071e-01 -5.21740437e-01 -3.80386472e-01
5.18653333e-01 3.22458856e-02 2.64154971e-01 1.40529767e-01
-2.64772087e-01 -3.79190803e-01 -6.11473061e-02 -8.79712999e-02
6.93035781e-01 4.53128785e-01 2.46278509e-01 -3.38829756e-01
7.43829012e-01 -3.00829057e-02 9.52687919e-01 2.13096589e-01
-2.89592296e-01 1.17284560e+00 3.42357486e-01 -4.45277780e-01
-1.14692152e+00 -9.71792102e-01 2.59315199e-03 2.54434109e-01
7.02068627e-01 -4.13206130e-01 -1.10297823e+00 -3.06405067e-01
-4.27286744e-01 3.06021839e-01 -4.33503836e-01 7.88482502e-02
-9.92019355e-01 -2.76695371e-01 -1.89575166e-01 1.74228936e-01
6.26196325e-01 -5.86801767e-01 -6.44604027e-01 3.88441265e-01
-4.50122595e-01 -1.22250605e+00 -8.88206542e-01 -3.97359729e-01
-1.09885597e+00 -1.12980115e+00 -1.03380203e+00 -8.45751107e-01
9.86681283e-01 9.31857228e-01 8.50141346e-01 1.14852034e-01
1.41261190e-01 3.08602124e-01 -2.47713119e-01 4.04384583e-01
-3.23140740e-01 -4.18008983e-01 4.70160358e-02 5.55386722e-01
-4.66757745e-01 -5.74369431e-01 -6.72019303e-01 7.83765256e-01
-1.01965129e+00 2.96045721e-01 -4.35171798e-02 8.37330699e-01
1.03872848e+00 1.28758177e-01 -2.63831347e-01 -6.12537622e-01
-1.13339126e-01 -9.91097838e-02 -8.18726957e-01 2.48548582e-01
-1.00136839e-01 -1.12011656e-01 4.29944277e-01 -5.42991221e-01
-1.11305630e+00 3.77323657e-01 1.88555121e-01 -8.62671316e-01
1.78081896e-02 1.50175661e-01 -2.83886611e-01 -4.56004560e-01
1.29562259e-01 3.58504772e-01 1.13177709e-01 -3.62003922e-01
3.89366150e-01 -6.51952531e-03 7.05929816e-01 -7.24107623e-01
1.18115115e+00 1.06555176e+00 1.79181054e-01 -9.37844515e-01
-4.61758167e-01 -3.25583607e-01 -7.66773939e-01 -5.27571440e-01
8.75228226e-01 -7.22151339e-01 -4.46856201e-01 6.10215843e-01
-1.21488309e+00 -1.36458084e-01 -3.11471760e-01 4.70769197e-01
-6.53472304e-01 8.67489338e-01 -4.30595607e-01 -5.03715873e-01
-1.09200627e-01 -1.41137171e+00 1.08407831e+00 1.97081417e-01
-4.67708670e-02 -8.59926403e-01 2.66833514e-01 3.14258844e-01
5.51252142e-02 7.10334480e-01 7.95131743e-01 1.28404245e-01
-1.05811524e+00 -1.74910650e-01 2.69411623e-01 2.43159831e-01
2.00458065e-01 5.58589995e-01 -6.22894645e-01 -4.12251860e-01
3.37823540e-01 2.85407424e-01 4.15365666e-01 5.40138721e-01
7.82395720e-01 -1.36084437e-01 -4.13572043e-01 8.58005106e-01
1.52011526e+00 2.26876527e-01 6.37913048e-01 1.49909988e-01
9.19995964e-01 5.84363401e-01 6.73479795e-01 3.45053345e-01
2.60912955e-01 1.03165066e+00 5.71406245e-01 -5.86365126e-02
1.15073379e-02 -2.62523562e-01 4.78016287e-01 9.71660674e-01
-5.53883791e-01 -1.92056328e-01 -6.22940481e-01 4.49763745e-01
-1.91555643e+00 -8.43671978e-01 -2.53236890e-01 2.53163028e+00
6.19107068e-01 -1.47104800e-01 -1.71171129e-01 1.73291430e-01
9.39272463e-01 4.14610893e-01 -2.59882778e-01 -1.88785177e-02
-1.49241179e-01 -1.56646296e-01 4.64660048e-01 8.01898062e-01
-9.03026760e-01 8.67456198e-01 5.90748930e+00 6.97394907e-01
-1.09986579e+00 -7.02030808e-02 2.90892750e-01 1.54674709e-01
-4.15495366e-01 3.84587348e-01 -5.31608760e-01 3.52912903e-01
2.39760622e-01 -1.36301592e-01 3.59642029e-01 6.57004416e-01
5.45405984e-01 -2.68609107e-01 -7.83397079e-01 8.70920777e-01
1.32263064e-01 -1.28829122e+00 2.31816862e-02 -5.78133576e-02
1.11439610e+00 -3.67916435e-01 -4.53002118e-02 -4.10437793e-01
-5.74088618e-02 -2.33647153e-01 1.02268243e+00 3.87667060e-01
6.80222034e-01 -6.99123442e-01 2.65822649e-01 2.75305450e-01
-1.29877198e+00 3.87385815e-01 -2.89767653e-01 1.55291989e-01
4.54470009e-01 5.94712317e-01 -2.53175586e-01 8.59397709e-01
5.42455792e-01 8.76125455e-01 -7.75080845e-02 8.58939350e-01
-2.52527356e-01 3.66273671e-01 -3.54495466e-01 7.99140155e-01
9.29729193e-02 -8.03484082e-01 1.05898690e+00 4.43639159e-01
5.64068675e-01 4.60473448e-01 2.64059961e-01 7.51595855e-01
9.88938287e-02 -6.09519333e-02 -6.69753313e-01 5.04283249e-01
2.44976148e-01 1.14122903e+00 -9.06857312e-01 -3.79840374e-01
-4.88738030e-01 1.08274531e+00 -9.66610834e-02 6.65272355e-01
-9.49940503e-01 7.06898198e-02 6.88357413e-01 5.22666991e-01
3.79562020e-01 -6.24310493e-01 -1.03836479e-02 -1.67825437e+00
2.77494848e-01 -6.81943238e-01 -1.43091246e-01 -8.52741063e-01
-6.93360329e-01 6.22459829e-01 1.94270536e-01 -1.71019411e+00
-1.39051691e-01 -1.01962648e-01 -6.33322120e-01 6.29105210e-01
-1.43896985e+00 -1.15982211e+00 -5.82301497e-01 8.95827651e-01
6.41355574e-01 4.33848768e-01 5.21870971e-01 4.40870970e-02
-5.39188504e-01 1.11467354e-01 4.00305957e-01 -1.19660370e-01
8.16387832e-01 -7.24749446e-01 1.24910705e-01 1.19735765e+00
-1.32125944e-01 5.12916744e-01 7.20094681e-01 -7.03890562e-01
-1.56394315e+00 -1.02359581e+00 6.17598176e-01 -1.74736485e-01
3.88756216e-01 -1.15613528e-01 -1.20988393e+00 7.74176955e-01
2.41047725e-01 2.22452879e-01 4.38982397e-02 -6.94557667e-01
-1.41449615e-01 -5.28443120e-02 -1.12772286e+00 6.22120619e-01
9.29963708e-01 -2.11838469e-01 -5.38881958e-01 1.97796151e-01
6.91959977e-01 -8.63474727e-01 -7.88824499e-01 4.14631665e-01
3.76903921e-01 -1.07555151e+00 1.08119631e+00 -8.57500732e-02
4.51238334e-01 -8.12249660e-01 -9.41667035e-02 -1.21378469e+00
-2.28225216e-01 -1.15366650e+00 -1.15380667e-01 1.33095634e+00
-3.21224123e-01 -5.86620271e-01 7.48720467e-01 6.97819591e-01
-3.07588309e-01 -5.28008044e-01 -1.00553465e+00 -7.75806606e-01
-1.51696712e-01 1.66592165e-03 3.98233652e-01 1.03664684e+00
-2.75214642e-01 -1.88852683e-01 -5.36344945e-01 3.91110539e-01
8.88496578e-01 4.84331250e-01 7.94682562e-01 -9.72953439e-01
-1.43319875e-01 6.41338229e-02 -3.79669785e-01 -1.21360147e+00
5.10868914e-02 -3.69924307e-01 1.74386904e-01 -9.86733496e-01
5.51809147e-02 -3.09713751e-01 2.71461666e-01 1.64692342e-01
-1.45356804e-01 2.04237580e-01 2.96818554e-01 5.13756931e-01
-5.32777868e-02 6.10336065e-01 1.66196489e+00 9.83547047e-03
-3.09381723e-01 -5.15434481e-02 -6.81329146e-02 1.07019150e+00
3.97987276e-01 -3.44896227e-01 -3.27299446e-01 -5.74136257e-01
-1.17837578e-01 5.49346864e-01 3.39961916e-01 -8.86220098e-01
4.21864837e-02 -4.41909999e-01 7.28523731e-03 -7.81402349e-01
4.55301881e-01 -1.15461206e+00 6.38095498e-01 3.42287898e-01
1.73491806e-01 8.88880715e-02 3.42155509e-02 7.19683588e-01
-4.40852880e-01 -4.72903438e-02 1.03317559e+00 -7.63769224e-02
-4.21684325e-01 3.76973897e-01 -2.18845099e-01 -8.40790793e-02
1.31202924e+00 -4.52756226e-01 -6.87694252e-02 -2.89620727e-01
-6.82924211e-01 -3.73167507e-02 1.24012303e+00 1.58232808e-01
5.87439716e-01 -1.34420431e+00 -4.26153898e-01 5.18340349e-01
-1.17878824e-01 5.58340192e-01 4.11679536e-01 8.83575916e-01
-9.56538796e-01 5.71608357e-02 -1.32793412e-01 -9.81094241e-01
-1.36769843e+00 6.35634184e-01 2.41967037e-01 -1.58760175e-01
-9.11403179e-01 4.61796016e-01 5.37894070e-01 1.18403204e-01
-8.84068683e-02 -6.00226939e-01 1.48732424e-01 -1.95284516e-01
4.60238963e-01 3.72050434e-01 -1.16989471e-01 -9.72928703e-01
-1.70238107e-01 1.45528495e+00 1.79051846e-01 -1.25801712e-01
1.22423208e+00 -4.98175204e-01 -2.86507130e-01 1.69494361e-01
1.07980621e+00 2.61994213e-01 -1.83494210e+00 -2.58016646e-01
-5.19865215e-01 -9.29757357e-01 -1.65549517e-01 6.84560910e-02
-1.39820337e+00 6.15502179e-01 2.45902926e-01 4.45634872e-02
1.21260905e+00 -2.39743441e-01 9.67303216e-01 -1.42773718e-01
6.41114950e-01 -7.31900096e-01 -1.23071678e-01 2.79721051e-01
9.15023088e-01 -8.14409554e-01 8.20397735e-02 -9.56679165e-01
-4.78845268e-01 1.32925487e+00 2.24326298e-01 -5.10259390e-01
5.55823863e-01 1.37614876e-01 -1.53397188e-01 1.27888456e-01
-3.74939382e-01 3.38368535e-01 1.95466354e-01 4.72506911e-01
7.81290904e-02 -1.92923158e-01 -2.32119471e-01 1.82326734e-02
2.35177949e-01 2.97881179e-02 8.10357451e-01 1.09350371e+00
-3.77124727e-01 -1.19848990e+00 -8.10385168e-01 -3.13177764e-01
-2.54952550e-01 1.87166661e-01 -6.00037239e-02 8.66096497e-01
1.18772797e-02 6.02466285e-01 -3.54686636e-03 -2.77933836e-01
3.12190980e-01 -1.99235246e-01 6.23459101e-01 -3.50937247e-01
-4.34087515e-01 6.05173528e-01 -2.73376107e-01 -9.93053019e-01
-7.91252136e-01 -8.87287498e-01 -1.29271138e+00 -3.17372888e-01
-2.13083655e-01 -4.58277762e-02 3.38771015e-01 7.53979743e-01
2.38171265e-01 2.56162345e-01 9.22245979e-01 -1.27787375e+00
-2.66204894e-01 -3.85878533e-01 -5.21757185e-01 5.28632343e-01
5.33094943e-01 -8.22354972e-01 -4.58485156e-01 6.22412860e-01] | [9.263564109802246, -2.2664644718170166] |
dddb6ec3-2481-4b51-b84e-ee51395c74d4 | what-makes-data-to-text-generation-hard-for | 2205.11505 | null | https://arxiv.org/abs/2205.11505v1 | https://arxiv.org/pdf/2205.11505v1.pdf | What Makes Data-to-Text Generation Hard for Pretrained Language Models? | Expressing natural language descriptions of structured facts or relations -- data-to-text generation (D2T) -- increases the accessibility of structured knowledge repositories. Previous work shows that pre-trained language models(PLMs) perform remarkably well on this task after fine-tuning on a significant amount of task-specific training data. On the other hand, while auto-regressive PLMs can generalize from a few task examples, their efficacy at D2T is largely unexplored. Furthermore, we have an incomplete understanding of the limits of PLMs on D2T. In this work, we conduct an empirical study of both fine-tuned and auto-regressive PLMs on the DART multi-domain D2T dataset. We consider their performance as a function of the amount of task-specific data and how these data are incorporated into the models: zero and few-shot learning, and fine-tuning of model weights. In addition, we probe the limits of PLMs by measuring performance on subsets of the evaluation data: novel predicates and abstractive test examples. To improve the performance on these subsets, we investigate two techniques: providing predicate descriptions in the context and re-ranking generated candidates by information reflected in the source. Finally, we conduct a human evaluation of model errors and show that D2T generation tasks would benefit from datasets with more careful manual curation. | ['Mark Dredze', 'Adrian Benton', 'Moniba Keymanesh'] | 2022-05-23 | null | null | null | null | ['data-to-text-generation'] | ['natural-language-processing'] | [ 2.77950615e-01 5.55345178e-01 -1.72370553e-01 -3.71826917e-01
-9.33647037e-01 -5.15344739e-01 1.10371935e+00 3.67750406e-01
-4.62620944e-01 1.01143861e+00 6.00396097e-01 -1.60017744e-01
-1.54927656e-01 -8.97793829e-01 -7.48559535e-01 -1.12896644e-01
-6.42579570e-02 9.74516153e-01 4.83460307e-01 -5.06543577e-01
8.88120010e-02 1.40189379e-01 -1.60152972e+00 6.69043422e-01
1.03900671e+00 7.22671866e-01 1.13669679e-01 4.29584861e-01
-4.82940137e-01 8.12891722e-01 -8.72163057e-01 -6.13279104e-01
1.98484614e-01 -2.60406673e-01 -8.46006632e-01 -7.13221952e-02
3.21231097e-01 7.65066314e-03 -8.87906775e-02 5.69230556e-01
5.22636294e-01 3.71750981e-01 7.65515327e-01 -1.12368274e+00
-8.85485888e-01 1.05544114e+00 -1.05807096e-01 4.66882855e-01
5.81528544e-01 3.96561056e-01 1.10683417e+00 -1.08395183e+00
1.04186714e+00 1.39097345e+00 6.17732346e-01 5.84635794e-01
-1.51619661e+00 -5.91642737e-01 1.18552625e-01 1.01549596e-01
-1.30056012e+00 -6.63054764e-01 4.90596324e-01 -6.18022203e-01
1.51558650e+00 2.26846384e-03 3.39056194e-01 1.42091548e+00
-2.52971321e-01 6.16987348e-01 9.49231207e-01 -6.96072519e-01
2.98132062e-01 5.26639044e-01 1.69849932e-01 4.51079756e-01
4.92615521e-01 1.67863324e-01 -7.54886568e-01 -3.30647677e-01
4.23541248e-01 -6.43441498e-01 -2.47422412e-01 -2.93634176e-01
-1.26842189e+00 8.41886997e-01 6.15402013e-02 5.07708788e-01
-3.25743467e-01 -1.61603391e-01 4.85136747e-01 3.21146816e-01
8.37400496e-01 1.20206356e+00 -7.55963504e-01 -1.83955401e-01
-8.86328578e-01 5.27917027e-01 9.47929382e-01 1.21833193e+00
7.52098918e-01 -8.32775086e-02 -7.35827267e-01 9.96510267e-01
-3.11955869e-01 1.60503954e-01 7.59074032e-01 -5.34792006e-01
8.60137820e-01 6.92279518e-01 3.20188850e-01 -6.80430770e-01
-3.49148512e-01 -3.43102634e-01 -1.73504516e-01 -1.02131411e-01
3.66094947e-01 -4.47240531e-01 -9.44587946e-01 1.80551004e+00
4.66892682e-02 -2.47013435e-01 4.57348466e-01 5.75083971e-01
8.13266873e-01 6.89928234e-01 4.50222254e-01 -1.18772581e-01
1.15540862e+00 -8.17098737e-01 -4.94923025e-01 -5.14703333e-01
8.63931060e-01 -4.95276242e-01 1.34685409e+00 5.39243408e-02
-1.06467605e+00 -6.68417692e-01 -8.34129870e-01 -1.07617706e-01
-6.60196185e-01 -6.90122470e-02 5.76137424e-01 3.76966000e-01
-8.00536871e-01 7.32004464e-01 -5.60224354e-01 -7.40427315e-01
4.12581772e-01 -3.77269200e-04 -3.36159110e-01 -1.53174788e-01
-1.54456711e+00 1.28053880e+00 9.04645324e-01 -6.05909586e-01
-8.79977703e-01 -1.10112691e+00 -8.77735913e-01 1.95351169e-01
6.16091728e-01 -7.97165096e-01 1.37951589e+00 -5.68654478e-01
-1.05625403e+00 8.48881602e-01 -1.22383364e-01 -6.48165047e-01
3.82383078e-01 -1.86377272e-01 -4.74654764e-01 -9.66761559e-02
3.65913600e-01 7.31805503e-01 5.82754910e-01 -1.20021248e+00
-7.23516941e-01 -7.00551718e-02 1.46010771e-01 8.93105417e-02
-3.97648513e-01 -1.25556812e-01 -1.23717345e-01 -6.22223973e-01
-4.27851021e-01 -7.15171456e-01 -1.22147918e-01 -5.23017347e-01
-3.86908889e-01 -5.81919909e-01 4.49195355e-01 -4.45324212e-01
1.30420399e+00 -2.00243926e+00 -1.07279226e-01 -5.80393076e-02
-7.88395405e-02 4.61349607e-01 -4.32853550e-01 6.94993079e-01
-1.51684910e-01 5.04365206e-01 -3.28794234e-02 -1.38188362e-01
8.31118226e-02 2.15359360e-01 -4.30265546e-01 -3.92672807e-01
6.47673905e-01 1.02771509e+00 -1.07113695e+00 -5.15537322e-01
-1.30354509e-01 1.30303413e-01 -5.02204061e-01 2.35768966e-02
-7.30453968e-01 4.45950627e-02 -5.80061615e-01 3.52064282e-01
8.60239565e-03 -2.79618740e-01 -2.48548039e-03 -1.83438733e-01
-7.07056448e-02 7.42888510e-01 -7.43802309e-01 1.64293110e+00
-5.48211098e-01 5.16672015e-01 -7.59893835e-01 -8.25103402e-01
9.50663030e-01 4.07188863e-01 1.27190351e-01 -7.49274373e-01
-1.89647198e-01 2.56319344e-01 8.55618119e-02 -7.21383274e-01
6.73020244e-01 -4.54038054e-01 -1.25780657e-01 5.49619615e-01
3.59224439e-01 -1.76519841e-01 7.21155047e-01 2.62336761e-01
1.33118379e+00 1.21677183e-01 4.21680212e-01 -2.11673751e-01
3.69490273e-02 6.12431049e-01 3.37425411e-01 9.63732779e-01
2.43302539e-01 4.82050687e-01 4.61284578e-01 -3.47012222e-01
-1.08758461e+00 -8.06456506e-01 -1.78420886e-01 1.26637995e+00
-3.07387531e-01 -6.75252616e-01 -4.45681274e-01 -8.48687112e-01
2.21721739e-01 1.64498103e+00 -7.16441333e-01 -3.61686140e-01
-3.62121850e-01 -7.19475091e-01 5.89932501e-01 5.53821623e-01
2.85684109e-01 -1.25822210e+00 -6.39956355e-01 3.64710957e-01
-2.27175668e-01 -1.23968124e+00 -8.67649466e-02 3.09656858e-01
-7.30593681e-01 -8.96378398e-01 -4.75320131e-01 -4.65594441e-01
4.80270177e-01 2.59566437e-02 1.65540516e+00 -3.33434284e-01
-1.11775547e-01 3.95078003e-01 -6.15149498e-01 -6.79520249e-01
-7.72228837e-01 3.85959357e-01 -7.88860768e-02 -4.78869140e-01
7.39340663e-01 -6.32745802e-01 9.14613996e-03 1.20492145e-01
-7.46737242e-01 2.05608770e-01 6.56387210e-01 7.79146373e-01
3.13475400e-01 1.69572830e-01 6.79490030e-01 -1.24264562e+00
1.18251002e+00 -5.55577755e-01 -3.63703400e-01 4.80115980e-01
-6.48328066e-01 5.60194016e-01 4.55287755e-01 -6.48614466e-01
-1.35559464e+00 -3.29019517e-01 2.42923364e-01 -4.22391891e-01
-6.36752546e-02 8.20938468e-01 2.80966684e-02 3.82400095e-01
1.40906131e+00 4.79297154e-02 -3.95916015e-01 -4.50943768e-01
5.81638336e-01 3.97456974e-01 3.02537501e-01 -9.73206460e-01
8.93343925e-01 -1.70143154e-02 -4.47850496e-01 -6.65540278e-01
-1.26230657e+00 -1.65349290e-01 -3.91677260e-01 2.13893816e-01
6.26779377e-01 -8.50987673e-01 2.94329505e-02 -2.33418807e-01
-1.08711958e+00 -5.65652251e-01 -8.11148524e-01 3.01607668e-01
-4.53891307e-01 -1.17280148e-01 -2.71859497e-01 -6.08277261e-01
-3.20917130e-01 -7.07549095e-01 9.72787738e-01 2.22566142e-03
-7.83058822e-01 -1.07962155e+00 1.66420504e-01 2.04832241e-01
4.84385818e-01 2.24160090e-01 1.49472702e+00 -1.25664485e+00
-3.03637594e-01 -1.97562292e-01 -8.52458403e-02 2.24408761e-01
5.43646961e-02 -2.92148560e-01 -9.96894181e-01 1.41669903e-02
-2.16070265e-01 -6.73468232e-01 8.70609045e-01 3.63567658e-02
8.85497808e-01 -4.13174868e-01 -5.39634705e-01 5.86361531e-03
1.25890982e+00 8.73650163e-02 3.70447218e-01 3.96420896e-01
2.83813238e-01 8.60430717e-01 7.82314181e-01 4.50847387e-01
3.80074054e-01 6.48807883e-01 -1.77558735e-01 2.45984182e-01
-3.09866548e-01 -6.96390688e-01 1.47699639e-01 2.71420836e-01
-1.46447703e-01 -3.09340388e-01 -1.21262991e+00 8.55210364e-01
-1.64913142e+00 -9.82735336e-01 2.97012061e-01 2.00914502e+00
1.22443771e+00 6.03916645e-01 -3.87813672e-02 -3.95490117e-02
5.58433533e-01 -3.27457897e-02 -4.63046759e-01 -2.15928495e-01
-1.78613588e-01 4.06708539e-01 1.78160742e-01 2.70721704e-01
-7.98565805e-01 1.12500429e+00 6.40654421e+00 9.60702121e-01
-7.61742294e-01 1.08512603e-01 3.61844867e-01 -2.57125139e-01
-5.55802107e-01 1.84657320e-01 -1.00072670e+00 2.60091215e-01
1.00452006e+00 -6.91632330e-01 2.35564858e-01 9.99086440e-01
5.75309061e-02 -5.54166362e-02 -1.54706621e+00 6.16154969e-01
1.21258259e-01 -1.43439186e+00 4.90949899e-01 -1.09770209e-01
8.19496214e-01 7.47804940e-02 -2.56113857e-01 1.03443432e+00
6.01382375e-01 -8.91737759e-01 6.62482381e-01 4.32449698e-01
7.45925725e-01 -3.85184288e-01 5.98413587e-01 4.84118283e-01
-7.33978987e-01 -1.87616304e-01 -3.91132444e-01 -1.51742101e-01
1.28498152e-01 6.06294930e-01 -1.32213914e+00 3.85483116e-01
3.92689794e-01 4.67108727e-01 -9.15520906e-01 6.72626138e-01
-3.21528018e-01 5.84807098e-01 -1.90438926e-01 -1.44771829e-01
1.29504934e-01 3.87214363e-01 4.84218627e-01 1.36686456e+00
3.19418728e-01 1.67272300e-01 1.31947398e-01 1.18327653e+00
-8.85901228e-02 -3.13971564e-02 -6.71438336e-01 -3.94980580e-01
6.74008846e-01 1.01077461e+00 -2.19926834e-01 -5.93599856e-01
-4.45386708e-01 3.39059204e-01 6.09393716e-01 4.70345885e-01
-4.10454094e-01 -1.82157964e-01 4.83580112e-01 4.40780371e-01
1.62016541e-01 4.29322831e-02 -1.64578199e-01 -1.24772227e+00
3.19367945e-02 -9.29230928e-01 3.52882445e-01 -1.03706431e+00
-1.63495207e+00 6.16972208e-01 4.23478305e-01 -7.92668819e-01
-6.08533323e-01 -3.61273259e-01 -3.55863184e-01 9.45250809e-01
-1.35499227e+00 -9.50869560e-01 -1.77902505e-01 3.98746014e-01
7.27145016e-01 -3.39078903e-01 1.00525486e+00 2.32088678e-02
-3.33463073e-01 4.15505469e-01 -3.46780390e-01 8.43545347e-02
7.37325191e-01 -1.20428765e+00 7.99725652e-01 6.59590185e-01
2.48321623e-01 8.32541227e-01 9.45837736e-01 -8.75536859e-01
-8.33955944e-01 -1.16568410e+00 1.21107960e+00 -8.51761222e-01
7.71390557e-01 -3.79421502e-01 -1.08472550e+00 8.08678865e-01
6.67494088e-02 -1.39508367e-01 7.46262252e-01 6.35736585e-01
-4.13216949e-01 1.87157303e-01 -9.45016742e-01 6.24854267e-01
1.18085182e+00 -5.10192513e-01 -1.16534400e+00 4.47398484e-01
9.42250013e-01 -1.83670312e-01 -1.04241967e+00 4.25091147e-01
2.45206639e-01 -6.91694021e-01 7.50643969e-01 -8.43471646e-01
6.69038296e-01 -3.21094543e-02 -6.41544014e-02 -1.62245858e+00
-2.92155117e-01 -4.53025907e-01 -4.68008369e-02 1.39735174e+00
1.02168214e+00 -2.89492130e-01 5.92897177e-01 9.17600334e-01
-2.93551207e-01 -7.94695437e-01 -6.86097920e-01 -9.28092062e-01
9.01659727e-02 -3.87106031e-01 5.82075894e-01 9.59851325e-01
2.75397122e-01 8.86021078e-01 -8.21962673e-03 -3.20873469e-01
2.63160527e-01 -9.20680687e-02 8.65419090e-01 -1.33846450e+00
-3.08936149e-01 -2.38287702e-01 -2.30417363e-02 -5.54731667e-01
2.29002625e-01 -1.01542735e+00 -1.20743372e-01 -1.65954971e+00
2.63231516e-01 -5.46571136e-01 -2.82419771e-02 8.08417201e-01
-2.92598605e-01 -4.16452438e-01 1.72890574e-01 1.73170164e-01
-4.78554398e-01 5.90910077e-01 1.14560854e+00 -6.23713396e-02
-4.15037066e-01 -1.78251520e-01 -8.71430755e-01 4.87613410e-01
7.17648864e-01 -5.40491045e-01 -7.40433693e-01 -5.93183815e-01
2.41725385e-01 2.74107456e-02 2.58464098e-01 -9.78268623e-01
2.99579259e-02 -1.71941146e-01 3.51294547e-01 -3.05967957e-01
2.76696086e-01 -4.26816702e-01 -3.61169986e-02 7.92032555e-02
-8.44886422e-01 -1.04458645e-01 4.37329739e-01 3.89516026e-01
-1.47346631e-01 -3.05599362e-01 4.31106508e-01 -4.35012311e-01
-8.00162971e-01 -6.08042814e-03 -2.51241505e-01 6.48501635e-01
7.62952387e-01 -1.86177805e-01 -6.01188421e-01 -2.04262987e-01
-7.12935567e-01 1.95181802e-01 2.83046663e-01 7.30960071e-01
3.87652099e-01 -1.19491279e+00 -8.87555003e-01 9.81182382e-02
6.93476140e-01 -4.59667109e-02 -1.02593899e-01 4.57964271e-01
6.96265921e-02 7.42321610e-01 -1.09267153e-01 -2.30190203e-01
-7.48050869e-01 7.40242660e-01 1.56022102e-01 -6.18480206e-01
-6.23588324e-01 7.14106381e-01 1.17294669e-01 -2.53492206e-01
-3.43113169e-02 -3.76616865e-01 -2.90482670e-01 2.04362109e-01
4.06445384e-01 6.53815120e-02 1.07972257e-01 -1.96081892e-01
-9.72380955e-03 -4.17890735e-02 -3.67404580e-01 -2.64540702e-01
1.30103016e+00 1.88043922e-01 3.14718336e-01 5.77164531e-01
7.42767990e-01 -1.73912719e-01 -1.02881169e+00 -5.47453701e-01
5.24632335e-01 -2.13061005e-01 -2.63224959e-01 -1.12690961e+00
-3.18763196e-01 6.15120530e-01 2.56384462e-02 2.76446611e-01
7.60856450e-01 2.00533986e-01 5.17303884e-01 6.15131736e-01
4.92321044e-01 -1.08494687e+00 2.67483801e-01 5.87510169e-01
1.03997254e+00 -1.19509721e+00 1.03133088e-02 -4.04549628e-01
-7.83969343e-01 7.62442231e-01 8.70314538e-01 1.24429412e-01
2.35085651e-01 8.07245523e-02 -2.07013562e-01 -2.74134845e-01
-1.28813624e+00 -3.41551453e-01 2.40995482e-01 7.71131396e-01
5.70191443e-01 -1.08571909e-01 -2.52475530e-01 8.37117255e-01
-5.89588761e-01 1.93332642e-01 3.70741755e-01 9.07454073e-01
-4.00907069e-01 -9.43109572e-01 -9.54017639e-02 7.00532317e-01
-1.49745032e-01 -4.25696164e-01 -5.56898177e-01 1.02952552e+00
1.00587554e-01 9.59054053e-01 -2.07747415e-01 -2.74180144e-01
7.47226119e-01 5.45314252e-01 3.32970500e-01 -1.30342770e+00
-4.97568995e-01 -3.84252906e-01 7.90273011e-01 -2.90459573e-01
-2.87007958e-01 -6.38341904e-01 -8.61703277e-01 1.18681550e-01
-3.23191524e-01 3.60374928e-01 3.98417443e-01 1.09691036e+00
4.58173692e-01 6.75191462e-01 -5.80017045e-02 -6.29606009e-01
-7.49340832e-01 -1.36627805e+00 -2.81807452e-01 6.39073431e-01
-8.84448290e-02 -7.94442892e-01 -2.98433542e-01 5.48931286e-02] | [10.766088485717773, 8.536148071289062] |
4c98969a-666e-4822-925c-ee99b3d6d784 | semantic-modeling-for-food-recommendation | 2105.01269 | null | https://arxiv.org/abs/2105.01269v1 | https://arxiv.org/pdf/2105.01269v1.pdf | Semantic Modeling for Food Recommendation Explanations | With the increased use of AI methods to provide recommendations in the health, specifically in the food dietary recommendation space, there is also an increased need for explainability of those recommendations. Such explanations would benefit users of recommendation systems by empowering them with justifications for following the system's suggestions. We present the Food Explanation Ontology (FEO) that provides a formalism for modeling explanations to users for food-related recommendations. FEO models food recommendations, using concepts from the explanation domain to create responses to user questions about food recommendations they receive from AI systems such as personalized knowledge base question answering systems. FEO uses a modular, extensible structure that lends itself to a variety of explanations while still preserving important semantic details to accurately represent explanations of food recommendations. In order to evaluate this system, we used a set of competency questions derived from explanation types present in literature that are relevant to food recommendations. Our motivation with the use of FEO is to empower users to make decisions about their health, fully equipped with an understanding of the AI recommender systems as they relate to user questions, by providing reasoning behind their recommendations in the form of explanations. | ['Deborah L. McGuinness', 'Daniel Gruen', 'Shruthi Chari', 'Oshani Seneviratne', 'Ishita Padhiar'] | 2021-05-04 | null | null | null | null | ['food-recommendation', 'knowledge-base-question-answering'] | ['miscellaneous', 'natural-language-processing'] | [ 8.94145668e-02 1.12562120e+00 -5.49357355e-01 -9.33754742e-01
2.05106005e-01 -4.87553507e-01 1.55602947e-01 8.95900369e-01
2.51734555e-01 1.87512681e-01 1.07041347e+00 -4.99796659e-01
-8.12162280e-01 -9.84676301e-01 -4.54379737e-01 1.93112284e-01
1.21762864e-01 4.75258887e-01 8.82341340e-03 -9.00714159e-01
1.86878487e-01 4.81191128e-02 -2.05410099e+00 8.77342343e-01
1.23318923e+00 5.61261535e-01 1.90933138e-01 7.23125517e-01
-6.26733422e-01 9.40438390e-01 -1.06475232e-02 -4.45847154e-01
-1.72860906e-01 -6.36490107e-01 -9.60225284e-01 -8.90323371e-02
2.16078460e-01 -6.11368775e-01 5.54570332e-02 7.28193879e-01
-1.96705133e-01 4.22951907e-01 7.50865936e-01 -1.26407504e+00
-1.51722407e+00 1.25988579e+00 7.62234211e-01 -1.90801755e-01
1.07649589e+00 -2.14884073e-01 1.01731527e+00 -4.55317259e-01
4.58430320e-01 1.17725003e+00 4.91625011e-01 1.13953316e+00
-9.79676604e-01 -1.45147189e-01 3.45275581e-01 4.62211609e-01
-5.52726388e-01 -2.44943455e-01 1.71111077e-01 -5.35184085e-01
1.38542485e+00 9.63999212e-01 9.65025902e-01 5.88835597e-01
1.15577757e-01 3.51250976e-01 4.64077026e-01 -6.15388274e-01
5.72490156e-01 6.76742911e-01 9.75436389e-01 4.05587584e-01
8.62781882e-01 2.84910023e-01 -3.47746819e-01 -2.66784430e-01
3.46580118e-01 4.50646609e-01 -1.82484269e-01 -3.56965691e-01
-6.72736347e-01 1.20949519e+00 6.55073524e-01 4.64493543e-01
-7.48049021e-01 -1.43833309e-01 5.94310183e-03 3.89321685e-01
2.67947197e-01 8.67296696e-01 -7.21831143e-01 6.59860849e-01
-8.50344375e-02 5.89184463e-01 1.48803782e+00 8.55403304e-01
4.34129775e-01 -2.15787590e-01 -1.70954183e-01 5.42717338e-01
1.14605927e+00 5.12697518e-01 6.51287556e-01 -1.05916834e+00
-4.28893775e-01 1.00388300e+00 4.73368734e-01 -1.27233565e+00
-5.74869931e-01 -3.41252148e-01 -4.44972143e-02 4.54381071e-02
3.09592575e-01 9.84464809e-02 -6.17873609e-01 1.51208842e+00
5.56418717e-01 -6.29267335e-01 6.96817696e-01 1.08884418e+00
1.51973045e+00 3.57456118e-01 7.89524257e-01 6.50138184e-02
1.85982180e+00 -7.55773365e-01 -8.88976514e-01 -3.36086929e-01
8.58914852e-01 -4.79337931e-01 9.71965969e-01 3.09806347e-01
-6.00626767e-01 -6.67247593e-01 -9.21662033e-01 -3.56180787e-01
-8.91870618e-01 -1.67363599e-01 1.02693331e+00 6.42994106e-01
-7.75963604e-01 6.28467619e-01 -3.76682162e-01 -1.21950328e+00
-3.21386337e-01 4.88577783e-01 -2.79336013e-02 -3.15170676e-01
-1.47273517e+00 1.14206660e+00 5.43926179e-01 -2.98961431e-01
-8.61766711e-02 -9.82639968e-01 -1.18638349e+00 4.84009564e-01
2.35760108e-01 -1.40086019e+00 1.47016883e+00 -1.12607813e+00
-1.30768657e+00 6.08668998e-02 2.49850258e-01 -4.68923479e-01
-3.88768107e-01 -4.45089340e-01 -9.78671551e-01 8.43110830e-02
2.42323056e-01 9.51262891e-01 6.83507472e-02 -8.69404614e-01
-6.15586519e-01 -3.09016168e-01 6.78489447e-01 2.67372757e-01
-1.51572078e-01 -8.42163786e-02 5.38882434e-01 -3.17393661e-01
-1.77846923e-02 -5.35126626e-01 -5.06849051e-01 8.34089816e-02
-4.31210101e-02 -3.61418396e-01 2.90065706e-01 -6.83250010e-01
1.14292431e+00 -1.63306093e+00 -3.14779311e-01 2.13886574e-01
2.95544654e-01 6.99544046e-03 -7.51997828e-02 7.67895460e-01
-1.49282604e-01 1.39521658e-01 5.16410291e-01 6.52478933e-01
7.42751896e-01 6.99613452e-01 -2.11068809e-01 -3.88441443e-01
-1.64828882e-01 5.79031408e-01 -1.21083522e+00 8.63701552e-02
4.17195112e-01 5.06836414e-01 -1.16814327e+00 3.06026936e-01
-1.02157402e+00 7.54616782e-02 -9.21389639e-01 8.99448991e-02
1.60081893e-01 -2.84768343e-01 3.55300546e-01 -4.95299757e-01
-3.41793895e-02 5.80773175e-01 -1.12399173e+00 1.52872467e+00
2.33819149e-03 -3.14344913e-01 -5.05370259e-01 -5.58179259e-01
8.59817028e-01 5.57666421e-01 3.64373535e-01 -5.17120779e-01
3.77742015e-02 3.23459297e-01 1.87034979e-01 -1.21646488e+00
5.72872877e-01 -1.70151293e-01 2.00641498e-01 6.27051115e-01
-1.33051053e-01 7.60219619e-02 5.28425753e-01 4.10581380e-01
7.96220362e-01 4.73501652e-01 8.60224664e-01 -7.27711618e-01
6.18920863e-01 5.72558463e-01 8.93279817e-03 6.92194641e-01
4.89521623e-01 -5.82432449e-02 -4.18500274e-01 -9.06024992e-01
-6.96369052e-01 -4.96955931e-01 4.83125895e-02 1.14967501e+00
-1.75841630e-01 -8.55720758e-01 -6.20260477e-01 -6.42001331e-01
1.33578166e-01 1.52744138e+00 -6.89696610e-01 -2.64236957e-01
-6.07059635e-02 -2.22131625e-01 -3.00929785e-01 5.24205446e-01
6.56483397e-02 -1.06851113e+00 -1.19295359e+00 5.64537168e-01
-2.71386564e-01 -4.24297303e-01 -4.73156869e-01 1.40068516e-01
-9.28385556e-01 -1.44725049e+00 -2.87203312e-01 -3.68821651e-01
6.00001812e-01 3.84187341e-01 1.47777379e+00 7.53430367e-01
2.68075671e-02 1.07307553e+00 -9.69159901e-01 -6.89749122e-01
-9.31067884e-01 -1.90553412e-01 -3.02598178e-01 -6.01709127e-01
1.12403357e+00 -3.44821334e-01 -1.06740856e+00 3.09664279e-01
-1.29210973e+00 3.30902636e-01 7.42037669e-02 1.54453561e-01
4.61607069e-01 -3.57202590e-01 7.29168832e-01 -1.26905060e+00
5.94420314e-01 -1.15991139e+00 -8.45645070e-02 2.99664795e-01
-1.09126043e+00 4.44844693e-01 5.17015874e-01 -3.43123466e-01
-1.01602888e+00 2.00385996e-03 -5.50079465e-01 7.23986447e-01
-7.16328084e-01 6.55011773e-01 -1.48938820e-01 2.46539772e-01
1.30919552e+00 -4.28381324e-01 -1.06181525e-01 -6.34027779e-01
1.14634013e+00 4.43763942e-01 3.22237313e-01 -3.48585486e-01
1.87561601e-01 -2.20705703e-01 -2.90620297e-01 -3.75356048e-01
-9.25839424e-01 -4.93084341e-01 1.20078102e-01 -5.91879636e-02
9.45457697e-01 -1.71991304e-01 -8.27037454e-01 -9.01612580e-01
-7.20295429e-01 -6.93295151e-02 -8.89579117e-01 7.34062612e-01
-6.29983604e-01 1.59270197e-01 -3.29926580e-01 -5.67423582e-01
-6.35573328e-01 -7.75346458e-01 4.62781131e-01 3.12734783e-01
-1.13784838e+00 -1.03030682e+00 -1.11351348e-02 4.92124528e-01
8.08920741e-01 1.06202208e-01 1.64931345e+00 -1.24496675e+00
-5.61415106e-02 1.47144690e-01 1.47434264e-01 -3.05586159e-01
3.09748948e-01 -4.86002237e-01 -6.29075825e-01 4.93213743e-01
-3.18442518e-03 2.84892499e-01 1.50305867e-01 4.59580481e-01
6.99331403e-01 -8.28416646e-01 -3.12132865e-01 -1.36627913e-01
1.38227308e+00 4.41349298e-01 6.23287439e-01 3.01096320e-01
1.03378683e-01 1.05218887e+00 6.21420622e-01 6.79698065e-02
8.97728682e-01 7.11089313e-01 4.27832693e-01 -1.28786996e-01
-3.51743996e-01 -3.42285275e-01 1.79568738e-01 2.70411134e-01
1.78019240e-01 -2.15773672e-01 -2.52446532e-01 2.17827946e-01
-2.05856752e+00 -8.69200349e-01 -6.31743193e-01 1.88755584e+00
4.98573184e-01 -6.99271500e-01 2.64252454e-01 9.90557745e-02
1.15605332e-01 -7.82788932e-01 -7.65652239e-01 -8.81181121e-01
4.93541628e-01 2.39870623e-02 3.11332252e-02 7.23152280e-01
-4.50529873e-01 4.89278078e-01 6.81375742e+00 -2.64743149e-01
-3.02783579e-01 -9.02716294e-02 2.06282422e-01 5.02561986e-01
-1.10281563e+00 -1.92603059e-02 -4.79496062e-01 -1.45861395e-02
1.31764960e+00 -2.90097594e-01 6.60455287e-01 8.15920770e-01
3.38409632e-01 2.21255332e-01 -1.48005068e+00 2.00625360e-01
2.83589941e-02 -1.39103591e+00 4.43627387e-01 -2.57299483e-01
2.68758297e-01 -5.38763881e-01 -3.83034766e-01 2.57428855e-01
7.63108790e-01 -8.95120323e-01 6.71100974e-01 7.27021754e-01
3.24092768e-02 -1.91811204e-01 6.92937970e-01 2.79493034e-01
-9.90124881e-01 -3.53935480e-01 -7.08387017e-01 -3.94464552e-01
-1.30217746e-01 3.48294854e-01 -9.07445192e-01 7.32026994e-01
5.69916189e-01 6.52266383e-01 -3.12973082e-01 8.86881888e-01
-3.63320261e-01 2.04712644e-01 -3.23075891e-01 -1.61267594e-01
-1.68619320e-01 -1.19296998e-01 4.99422811e-02 1.10331893e+00
7.60934114e-01 1.00050569e+00 1.73476875e-01 8.50919008e-01
6.77265525e-01 9.53733623e-01 -4.48688954e-01 -7.57973418e-02
1.40738308e-01 9.61385489e-01 -6.22016907e-01 -5.24118364e-01
-5.02196431e-01 6.16577864e-01 -3.42774302e-01 2.23894894e-01
-4.13483739e-01 4.04037312e-02 7.50887692e-01 4.44502711e-01
7.64033897e-03 6.40814841e-01 1.64665237e-01 -9.74862993e-01
-7.19311774e-01 -1.33798850e+00 9.10973251e-01 -1.29876947e+00
-1.33604395e+00 6.11454010e-01 3.33775669e-01 -5.94961047e-01
-7.25484073e-01 -5.48397183e-01 -2.08076343e-01 7.13106573e-01
-1.16845953e+00 -1.18465817e+00 -3.71780038e-01 4.44123566e-01
5.55762231e-01 3.62543881e-01 1.43336689e+00 3.84904414e-01
4.57275599e-01 -1.19142719e-01 -4.80684608e-01 -7.55804360e-01
3.74428689e-01 -1.35650218e+00 4.51886579e-02 2.87233710e-01
9.39363465e-02 9.37827408e-01 1.36022842e+00 -8.84831369e-01
-1.34901714e+00 -8.05631459e-01 1.24738753e+00 -4.87465501e-01
3.64586264e-02 4.82287437e-01 -1.03010798e+00 8.28714609e-01
2.39091709e-01 -6.98993146e-01 1.57383442e+00 2.14901805e-01
-2.27983087e-01 1.24353707e-01 -1.61833704e+00 4.75308716e-01
9.93343472e-01 1.42241446e-02 -1.17863572e+00 6.48647785e-01
8.87130380e-01 -1.68458968e-01 -1.11183429e+00 6.09392375e-02
7.96783566e-01 -7.44953752e-01 1.29889262e+00 -1.15027654e+00
3.56165498e-01 -6.58373654e-01 -4.22919512e-01 -1.36806524e+00
-7.76963711e-01 -5.98154888e-02 -1.98977336e-01 8.86745453e-01
5.42027235e-01 -6.41071379e-01 4.40329164e-01 1.40517843e+00
-4.81397599e-01 -1.98174730e-01 7.82185718e-02 5.57348505e-02
-2.08421573e-01 -2.31179416e-01 1.47603464e+00 8.73983979e-01
7.28703856e-01 3.10072839e-01 -1.18093207e-01 4.46363479e-01
3.40365589e-01 1.68353572e-01 3.65770102e-01 -1.54570341e+00
-6.40159786e-01 -1.20005667e-01 -1.24836769e-02 -9.51945901e-01
-5.19067883e-01 -1.24927962e+00 -2.29727905e-02 -2.31646681e+00
3.70022692e-02 -1.15131766e-01 -2.00763822e-01 7.30832458e-01
-6.75542131e-02 6.49263039e-02 1.99290112e-01 1.21291988e-02
-2.71564245e-01 -9.76649001e-02 9.70449626e-01 6.75838515e-02
-2.71873325e-01 -1.72206417e-01 -1.70509601e+00 9.09743905e-01
7.05924988e-01 -8.43160927e-01 -9.89871264e-01 -3.82381171e-01
8.39685082e-01 5.22128977e-02 6.89501613e-02 -7.86446869e-01
3.21701318e-02 -3.29737902e-01 4.45812881e-01 -9.10415500e-02
-2.24028677e-01 -1.32680488e+00 1.12429500e+00 9.67855334e-01
-1.00505209e+00 -6.56687608e-03 1.25236496e-01 2.30052531e-01
3.38263988e-01 -7.56360769e-01 9.65650007e-03 -1.75510079e-01
-7.66297698e-01 -1.50482088e-01 -7.72465646e-01 -6.16690516e-01
4.70765322e-01 -2.64947116e-01 -3.74862105e-01 -4.90087420e-01
-1.42494714e+00 3.02552730e-01 2.54602015e-01 6.64811671e-01
4.38352674e-01 -1.27888966e+00 -3.52201283e-01 2.34123141e-01
5.12491465e-01 -8.92305136e-01 3.17378134e-01 1.91090316e-01
-5.00437021e-01 7.97349989e-01 -3.47567856e-01 9.18941498e-02
-9.59630728e-01 1.01734209e+00 5.38986921e-01 -8.03480577e-03
-9.84540284e-01 1.95343748e-01 3.32750410e-01 -7.29089320e-01
1.02195524e-01 -9.30666089e-01 -1.12433183e+00 -4.31240290e-01
1.16754901e+00 1.88798800e-01 -2.52612650e-01 -1.25895500e-01
-1.07301615e-01 2.39121377e-01 2.54907370e-01 4.65663970e-01
1.30735648e+00 -3.20724159e-01 2.28061691e-01 -9.73080937e-03
8.46454576e-02 -1.14823692e-01 -4.94180977e-01 6.37260154e-02
1.21313818e-01 -3.17120105e-01 -1.48888230e-01 -1.71597147e+00
-7.76952803e-01 3.84389281e-01 8.13483298e-01 8.19848835e-01
9.72209156e-01 2.14626640e-01 6.97930813e-01 5.35461485e-01
8.35478604e-02 -5.02311647e-01 -3.29564244e-01 -1.51142150e-01
1.10693109e+00 -7.68280149e-01 -5.97817218e-03 -8.92287076e-01
-2.83355057e-01 1.67452168e+00 2.31623277e-01 2.50515640e-01
6.90535069e-01 -2.15260565e-01 2.82696903e-01 -7.23585725e-01
-5.78536808e-01 -2.14928344e-01 7.23256946e-01 1.01341701e+00
9.18320239e-01 2.99493968e-01 -9.15169179e-01 1.23366809e+00
-2.24815503e-01 6.83827758e-01 2.99816459e-01 4.25834358e-01
-1.04110980e+00 -1.53959155e+00 -2.81795144e-01 6.03639007e-01
-2.94836015e-01 -1.22811802e-01 -4.01985109e-01 5.91776013e-01
3.99533719e-01 1.64479971e+00 -4.49109524e-01 -5.98211139e-02
6.79268062e-01 1.21538967e-01 3.87875944e-01 -9.83481169e-01
-1.20979345e+00 -2.27946103e-01 5.89121640e-01 -6.33156538e-01
-5.44574738e-01 -3.93485546e-01 -1.68602359e+00 -3.88952434e-01
-2.41496816e-01 8.56846035e-01 6.41003907e-01 1.12807655e+00
4.92939860e-01 6.79878414e-01 -2.85311848e-01 -1.12191737e-01
-4.71591622e-01 -6.48303807e-01 -3.69580597e-01 8.58254671e-01
-1.71713203e-01 -4.37843353e-01 -1.95211880e-02 2.22080350e-01] | [11.45789623260498, 4.564875602722168] |
123be388-0fe3-4ff7-b60e-bfe61a0e1e6f | adaptivepose-human-parts-as-adaptive-points | 2112.13635 | null | https://arxiv.org/abs/2112.13635v1 | https://arxiv.org/pdf/2112.13635v1.pdf | AdaptivePose: Human Parts as Adaptive Points | Multi-person pose estimation methods generally follow top-down and bottom-up paradigms, both of which can be considered as two-stage approaches thus leading to the high computation cost and low efficiency. Towards a compact and efficient pipeline for multi-person pose estimation task, in this paper, we propose to represent the human parts as points and present a novel body representation, which leverages an adaptive point set including the human center and seven human-part related points to represent the human instance in a more fine-grained manner. The novel representation is more capable of capturing the various pose deformation and adaptively factorizes the long-range center-to-joint displacement thus delivers a single-stage differentiable network to more precisely regress multi-person pose, termed as AdaptivePose. For inference, our proposed network eliminates the grouping as well as refinements and only needs a single-step disentangling process to form multi-person pose. Without any bells and whistles, we achieve the best speed-accuracy trade-offs of 67.4% AP / 29.4 fps with DLA-34 and 71.3% AP / 9.1 fps with HRNet-W48 on COCO test-dev dataset. | ['Mingshu He', 'Qian Zhang', 'Guoli Wang', 'Dongdong Yu', 'Xiaojuan Wang', 'Yabo Xiao'] | 2021-12-27 | null | null | null | null | ['multi-person-pose-estimation'] | ['computer-vision'] | [-6.24615066e-02 -1.65431201e-02 1.26831636e-01 -3.36443305e-01
-7.76046574e-01 -2.02783167e-01 3.57366502e-01 -3.41299586e-02
-7.11363077e-01 5.26060104e-01 6.84113875e-02 4.15046871e-01
-3.22273485e-02 -5.43322206e-01 -6.89520895e-01 -5.09982705e-01
3.61228213e-02 7.05228984e-01 3.13440025e-01 -3.33736598e-01
-1.29656419e-01 4.42118526e-01 -1.46033359e+00 -2.11846396e-01
8.28013361e-01 9.51005757e-01 -2.70175725e-01 6.79966509e-01
6.05868816e-01 5.96035048e-02 -6.33023858e-01 -8.51054430e-01
3.52158755e-01 -4.72399481e-02 -3.57740581e-01 -4.02594805e-02
8.48911881e-01 -4.38355476e-01 -2.93420702e-01 7.37375915e-01
8.59634876e-01 2.67450184e-01 5.16218305e-01 -1.18824732e+00
-1.42053798e-01 1.56410366e-01 -1.06597579e+00 -1.14309773e-01
5.14293849e-01 9.96654406e-02 7.99645662e-01 -9.31065619e-01
4.31959063e-01 1.62135851e+00 1.05094743e+00 5.17319679e-01
-1.16411781e+00 -7.13466704e-01 1.43751442e-01 6.25400171e-02
-1.59916198e+00 -3.34924340e-01 7.45951176e-01 -3.42493892e-01
5.87867081e-01 2.33371899e-01 9.97997761e-01 1.13488472e+00
3.88677150e-01 7.34174967e-01 7.87605345e-01 -1.23190701e-01
-2.23714665e-01 -4.64216739e-01 5.91143779e-02 8.71597886e-01
6.29892588e-01 -9.12723765e-02 -5.09521067e-01 1.62459258e-02
1.02568460e+00 2.66059726e-01 5.66700846e-02 -3.88224512e-01
-1.13674116e+00 6.46567464e-01 5.91079533e-01 -1.87962815e-01
-4.46462035e-01 3.71058702e-01 4.62185055e-01 -3.78530920e-01
2.74904847e-01 5.72227202e-02 -5.05776048e-01 -8.46044794e-02
-9.47806835e-01 8.33594620e-01 5.68308592e-01 9.73059595e-01
5.25747061e-01 4.45176549e-02 -4.71928209e-01 8.25515389e-01
4.40446794e-01 5.94067037e-01 2.42043346e-01 -8.10725272e-01
6.73473299e-01 5.93078434e-01 1.87384978e-01 -1.21040404e+00
-8.36673141e-01 -7.31330574e-01 -1.02869511e+00 -8.74620602e-02
7.18245804e-01 -2.91150987e-01 -9.77121890e-01 1.71860909e+00
6.79817021e-01 9.45989490e-02 -4.60167319e-01 1.14504969e+00
6.79176807e-01 3.38615924e-01 2.01231673e-01 1.55692846e-01
1.92735004e+00 -1.31416214e+00 -4.66435820e-01 -4.23302025e-01
1.20319717e-01 -4.63365823e-01 9.38706100e-01 3.72900307e-01
-1.34236705e+00 -9.06680942e-01 -1.12754846e+00 -4.14840281e-01
4.34704199e-02 4.30947661e-01 5.86188316e-01 5.71091175e-01
-4.84084278e-01 7.37361491e-01 -1.06878936e+00 -1.92429483e-01
4.71778274e-01 5.89884222e-01 -4.15987134e-01 8.88599381e-02
-1.03089023e+00 7.46243477e-01 1.73251539e-01 6.12634778e-01
-5.76580346e-01 -7.10533977e-01 -9.53635871e-01 4.49991375e-02
6.74658418e-01 -1.20763993e+00 8.74997437e-01 -4.41819936e-01
-1.66149855e+00 6.04468703e-01 -8.61825421e-02 -4.12668765e-01
9.81971204e-01 -9.23814952e-01 -2.36147314e-01 1.69681385e-01
9.60569903e-02 7.62626529e-01 9.16460335e-01 -9.83057141e-01
-6.34900928e-01 -8.27009320e-01 -1.78000972e-01 4.32179511e-01
-1.90772116e-01 -1.16754539e-01 -1.12552428e+00 -8.75146985e-01
3.15395057e-01 -1.43835509e+00 -2.04404324e-01 2.77414739e-01
-5.17610908e-01 -1.57673657e-01 4.08416063e-01 -9.18387890e-01
1.14467478e+00 -1.78478074e+00 4.10160124e-01 8.52519721e-02
3.77794355e-01 2.66292274e-01 1.26006141e-01 6.11388497e-02
1.42295375e-01 -2.29632780e-01 -4.63131256e-02 -8.15026939e-01
1.20346360e-01 -3.77122983e-02 3.77718866e-01 7.51494527e-01
1.44778743e-01 1.10223532e+00 -6.05156958e-01 -6.40877903e-01
3.18092018e-01 8.99389088e-01 -7.92628288e-01 8.75253081e-02
1.87639520e-01 4.56790954e-01 -4.17032748e-01 6.90938890e-01
7.23242342e-01 -4.61864434e-02 -1.07700795e-01 -5.55348217e-01
1.49129227e-01 -1.46078542e-01 -1.43725097e+00 2.09725618e+00
-1.46766976e-01 4.60723937e-02 3.45201120e-02 -5.48191547e-01
8.69436204e-01 8.81227255e-02 4.80038315e-01 -3.93009573e-01
2.90548265e-01 -3.75739438e-03 4.47752811e-02 -2.44683072e-01
5.00411570e-01 -6.27303496e-02 -3.28744769e-01 -1.79434270e-01
3.87389399e-03 4.22934592e-01 9.44440588e-02 -2.64872849e-01
6.23170197e-01 5.93194604e-01 3.30126047e-01 -2.04696044e-01
6.01245463e-01 -4.65793848e-01 8.98707390e-01 4.91975397e-01
-2.98221648e-01 7.91331708e-01 2.62434900e-01 -4.84361142e-01
-8.87869000e-01 -1.00915849e+00 6.94441274e-02 1.11730027e+00
1.54901907e-01 -5.98101318e-01 -8.08799803e-01 -3.24478030e-01
2.32452706e-01 5.58294877e-02 -6.03019476e-01 -6.10724241e-02
-1.01271045e+00 -6.73134208e-01 7.78384864e-01 7.69099951e-01
6.37220144e-01 -7.13984251e-01 -8.14418018e-01 1.09210223e-01
-2.21530020e-01 -1.25980043e+00 -7.57138312e-01 -2.20950738e-01
-7.29786634e-01 -1.03978920e+00 -1.12838638e+00 -5.07851422e-01
5.19395351e-01 -1.32626504e-01 8.97362888e-01 -6.84563369e-02
-2.91417867e-01 -7.11668879e-02 -1.41283616e-01 -1.62120596e-01
2.98468173e-01 1.78456992e-01 3.40782672e-01 3.64341699e-02
4.37056785e-03 -5.31692147e-01 -1.01932621e+00 3.81687433e-01
-1.96925864e-01 2.42382124e-01 7.20559120e-01 7.88677275e-01
7.31054187e-01 -2.89603710e-01 2.82460332e-01 -5.51007032e-01
2.60957479e-01 9.90532455e-04 -3.21550906e-01 1.73485562e-01
-3.53646338e-01 -8.69262293e-02 6.48752987e-01 -6.06714189e-01
-8.84526789e-01 3.91472012e-01 -3.40965420e-01 -5.07714212e-01
-1.55088693e-01 3.25739831e-02 -3.51331949e-01 -1.34649679e-01
4.66185391e-01 7.10164905e-02 1.24365138e-02 -6.52039886e-01
2.76337564e-01 1.82698905e-01 9.17468131e-01 -8.31427634e-01
9.37682092e-01 4.26293433e-01 2.21416593e-01 -6.96017563e-01
-8.26404154e-01 -4.23123032e-01 -9.44997370e-01 -3.14294249e-01
1.08239257e+00 -1.10521114e+00 -1.27740991e+00 6.71910763e-01
-1.11791480e+00 1.01661690e-01 -4.50198613e-02 4.16429996e-01
-4.77563620e-01 5.28871715e-01 -7.69200206e-01 -7.05936968e-01
-8.47289443e-01 -1.21655202e+00 1.46954775e+00 2.83971041e-01
-3.26307952e-01 -3.06381911e-01 -1.76323965e-01 4.80208129e-01
4.45435382e-02 7.04178631e-01 2.50994712e-01 -3.88088465e-01
-1.63752049e-01 -4.05344516e-01 -1.84892058e-01 -1.54212549e-01
-2.37572148e-01 -2.28473008e-01 -6.61350846e-01 -6.31527305e-01
-3.06879878e-01 -1.49258181e-01 4.78359163e-01 4.21313822e-01
1.09193540e+00 -2.88929284e-01 -2.93088496e-01 1.07898831e+00
1.16447520e+00 -2.26832956e-01 4.51696455e-01 2.94512361e-01
1.21352780e+00 5.90072274e-01 6.07699037e-01 5.38133979e-01
7.59529471e-01 1.01839817e+00 2.41115957e-01 -8.95112008e-02
-5.40756918e-02 -5.66151440e-01 1.19805098e-01 6.18906617e-01
-6.25227690e-01 8.02351236e-02 -7.15284109e-01 8.65523592e-02
-1.86824250e+00 -8.00344706e-01 -1.56077191e-01 2.10249472e+00
5.80363393e-01 2.98290342e-01 7.03538537e-01 2.41766628e-02
7.53885150e-01 9.27877277e-02 -7.24606574e-01 2.48208061e-01
3.15776587e-01 1.18084006e-01 5.04027307e-01 3.05252016e-01
-1.15807414e+00 8.81053865e-01 5.25688982e+00 6.99452817e-01
-7.12394774e-01 -1.45821348e-01 4.41635072e-01 -1.99394256e-01
3.71710032e-01 -3.20286870e-01 -1.17008007e+00 4.95714664e-01
7.53565788e-01 2.41447538e-01 1.27443925e-01 7.04615891e-01
2.35470265e-01 8.75190124e-02 -1.12189269e+00 1.37494528e+00
3.54091674e-02 -9.58154678e-01 1.10611496e-02 3.33508998e-02
2.66227782e-01 -4.54571456e-01 -2.51948029e-01 3.77094865e-01
-1.88973352e-01 -9.74544764e-01 1.12133229e+00 7.36659884e-01
7.21177340e-01 -9.65506554e-01 5.73590279e-01 4.46632773e-01
-1.59765244e+00 -2.13562716e-02 -1.30610749e-01 -2.24440843e-02
3.42293352e-01 1.30633056e-01 -1.58386722e-01 8.36933970e-01
8.98213208e-01 5.60672343e-01 -6.59459710e-01 9.68817353e-01
-2.54450738e-02 7.78435729e-04 -4.59765941e-01 8.76389965e-02
-2.21938983e-01 -1.47706687e-01 4.64772403e-01 1.27976143e+00
2.07383975e-01 1.12467729e-01 4.98282790e-01 5.04838347e-01
8.85190070e-02 -5.14363050e-02 2.49698147e-01 5.42527258e-01
3.89963359e-01 1.34764028e+00 -6.28350437e-01 -2.69415885e-01
-1.38275921e-01 1.16304278e+00 3.79874915e-01 5.69586502e-03
-1.23293769e+00 -3.46454710e-01 6.28223360e-01 3.40503067e-01
2.79744536e-01 -4.35827315e-01 -1.73839167e-01 -1.26995623e+00
2.27012232e-01 -1.04168701e+00 4.12909687e-01 -5.08422792e-01
-1.13388729e+00 4.66395736e-01 2.38188028e-01 -1.19531512e+00
-2.41936877e-01 -5.15239894e-01 -2.26114079e-01 8.70119035e-01
-1.04986989e+00 -1.63605618e+00 -6.26245499e-01 6.68293536e-01
4.88492519e-01 1.96266904e-01 6.49558961e-01 5.14724076e-01
-9.38652217e-01 1.07820582e+00 -5.95341146e-01 3.83726865e-01
7.56546915e-01 -1.08671379e+00 5.58078647e-01 7.89117992e-01
-9.37018916e-02 8.68001699e-01 7.86092937e-01 -6.35177910e-01
-1.47469640e+00 -8.92689884e-01 6.06281400e-01 -5.01076698e-01
3.75713110e-01 -5.59968650e-01 -6.66551232e-01 5.97100675e-01
-3.72681528e-01 1.85926348e-01 3.97478223e-01 2.27921024e-01
-3.12739283e-01 -2.94363499e-01 -1.00740004e+00 8.16238225e-01
1.15379882e+00 -4.31111045e-02 -5.41453660e-01 2.68299989e-02
5.84233463e-01 -8.98981631e-01 -1.09674466e+00 4.77111250e-01
1.02169013e+00 -8.87657225e-01 1.40206504e+00 -2.89278507e-01
2.89480597e-01 -3.70952696e-01 1.14352174e-01 -7.46035576e-01
-4.99117732e-01 -7.02298582e-01 -4.78519708e-01 1.13734281e+00
-5.95959574e-02 -4.71472681e-01 9.52566087e-01 5.98565519e-01
1.73961930e-03 -1.04237747e+00 -9.39438879e-01 -5.52062631e-01
-5.25419787e-02 -2.68284619e-01 5.12744129e-01 4.12694782e-01
-4.64170277e-01 4.74627793e-01 -8.75344038e-01 1.51533246e-01
9.84330416e-01 -1.45652652e-01 1.13958108e+00 -1.33370543e+00
-4.31570828e-01 -3.41186792e-01 -4.64031488e-01 -1.33542407e+00
-9.63361785e-02 -4.01294440e-01 9.36853960e-02 -1.25970948e+00
2.52178848e-01 -1.93179831e-01 -1.16137534e-01 4.80178535e-01
-3.76713037e-01 4.20171797e-01 4.34524119e-01 2.69714177e-01
-5.48908412e-01 6.28305018e-01 1.40100825e+00 1.78466737e-01
-1.87786400e-01 2.19395012e-01 -6.88086450e-01 9.69923913e-01
3.62273157e-01 -3.41409028e-01 -9.66004059e-02 -3.14156175e-01
-1.17718473e-01 2.99459517e-01 7.86220789e-01 -1.31133068e+00
3.28475177e-01 1.45387217e-01 9.61341143e-01 -8.22817564e-01
8.40330303e-01 -6.78012788e-01 3.52333844e-01 6.78973019e-01
-5.16238399e-02 2.82600135e-01 1.27123162e-01 7.37154007e-01
7.66489282e-02 2.62728333e-01 9.44119573e-01 -8.02625641e-02
-4.84630615e-01 6.98355377e-01 3.82463813e-01 -5.57347201e-02
1.08627307e+00 -5.05930841e-01 -2.55462117e-02 -7.89732262e-02
-8.38705778e-01 3.72077286e-01 3.41343701e-01 5.72447598e-01
4.19370711e-01 -1.34320259e+00 -7.73327589e-01 8.21732283e-02
-1.20937593e-01 2.99607903e-01 6.47165060e-01 8.66587996e-01
-5.68401337e-01 2.41502061e-01 -2.92748749e-01 -6.78367436e-01
-1.35961473e+00 1.57477409e-01 4.71198291e-01 -4.63411212e-01
-9.78718042e-01 8.20046842e-01 2.53755581e-02 -2.56334573e-01
1.21450663e-01 -1.10488340e-01 -1.45361021e-01 1.62944555e-01
3.30955058e-01 6.99935257e-01 -1.63606614e-01 -1.12793136e+00
-6.31934404e-01 1.01444578e+00 3.68364118e-02 1.70673653e-02
1.28890538e+00 6.97253942e-02 1.78366542e-01 1.96460828e-01
1.14818335e+00 -7.83679411e-02 -1.67425394e+00 6.69858009e-02
-2.36265317e-01 -5.20549834e-01 -2.46403620e-01 -6.65037096e-01
-1.00886631e+00 7.11912274e-01 6.75448954e-01 -4.24450785e-01
8.48515928e-01 -2.02229962e-01 1.02994573e+00 6.97393017e-03
3.45325530e-01 -1.08458769e+00 1.55111775e-01 3.40985954e-01
9.11815286e-01 -1.03468704e+00 4.71314162e-01 -5.65201342e-01
-4.76447940e-01 1.00683761e+00 9.26204979e-01 -4.62838560e-01
3.17479491e-01 1.18509285e-01 -3.21947336e-01 -5.64342029e-02
-2.20460281e-01 -4.59987745e-02 9.06000614e-01 3.73965174e-01
4.45991337e-01 1.59375757e-01 -4.15170491e-01 9.72735643e-01
-4.31275457e-01 -9.03945044e-02 -1.30525142e-01 7.40123451e-01
-1.40406087e-01 -8.23662341e-01 -5.35679877e-01 2.70502418e-01
-6.33028865e-01 2.30869651e-01 1.15159966e-01 1.13322473e+00
2.73717791e-01 5.70471466e-01 -2.02254858e-02 -4.48920846e-01
8.17047477e-01 -1.17726564e-01 5.82383335e-01 -4.05446738e-01
-8.44117463e-01 4.35186207e-01 -1.93391666e-02 -7.50877142e-01
-3.70606095e-01 -7.71118641e-01 -1.24602365e+00 -5.34021974e-01
-2.85783201e-01 -2.72617042e-01 3.39272857e-01 8.83053839e-01
4.02870625e-01 6.30572498e-01 -5.24090952e-04 -1.50116825e+00
-5.81948280e-01 -9.90687609e-01 -3.70311141e-01 5.43246865e-01
1.81851581e-01 -8.86303365e-01 3.40084471e-02 -1.89837620e-01] | [7.14849328994751, -0.7889328598976135] |
56bc0d93-f518-48b0-a55e-35847b82349e | cross-modal-clinical-graph-transformer-for-1 | 2206.01988 | null | https://arxiv.org/abs/2206.01988v1 | https://arxiv.org/pdf/2206.01988v1.pdf | Cross-modal Clinical Graph Transformer for Ophthalmic Report Generation | Automatic generation of ophthalmic reports using data-driven neural networks has great potential in clinical practice. When writing a report, ophthalmologists make inferences with prior clinical knowledge. This knowledge has been neglected in prior medical report generation methods. To endow models with the capability of incorporating expert knowledge, we propose a Cross-modal clinical Graph Transformer (CGT) for ophthalmic report generation (ORG), in which clinical relation triples are injected into the visual features as prior knowledge to drive the decoding procedure. However, two major common Knowledge Noise (KN) issues may affect models' effectiveness. 1) Existing general biomedical knowledge bases such as the UMLS may not align meaningfully to the specific context and language of the report, limiting their utility for knowledge injection. 2) Incorporating too much knowledge may divert the visual features from their correct meaning. To overcome these limitations, we design an automatic information extraction scheme based on natural language processing to obtain clinical entities and relations directly from in-domain training reports. Given a set of ophthalmic images, our CGT first restores a sub-graph from the clinical graph and injects the restored triples into visual features. Then visible matrix is employed during the encoding procedure to limit the impact of knowledge. Finally, reports are predicted by the encoded cross-modal features via a Transformer decoder. Extensive experiments on the large-scale FFA-IR benchmark demonstrate that the proposed CGT is able to outperform previous benchmark methods and achieve state-of-the-art performances. | ['Xiaojun Chang', 'Xiaodan Liang', 'Shirui Pan', 'Karin Verspoor', 'Wenjia Cai', 'Mingjie Li'] | 2022-06-04 | cross-modal-clinical-graph-transformer-for | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Cross-Modal_Clinical_Graph_Transformer_for_Ophthalmic_Report_Generation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Cross-Modal_Clinical_Graph_Transformer_for_Ophthalmic_Report_Generation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['medical-report-generation', 'clinical-knowledge'] | ['medical', 'miscellaneous'] | [ 4.13668811e-01 5.39630592e-01 -3.99838001e-01 -4.12672430e-01
-8.08037579e-01 -2.41070479e-01 4.68415946e-01 8.53230283e-02
7.33701587e-02 9.49140370e-01 4.71503764e-01 -4.74077940e-01
-3.40878874e-01 -8.00531149e-01 -6.76376700e-01 -3.39402080e-01
2.75590539e-01 2.32194901e-01 -1.63567260e-01 -1.63188286e-03
4.90403920e-02 1.79321140e-01 -1.45227790e+00 7.16945648e-01
1.24416471e+00 9.75623250e-01 1.31616920e-01 3.96484911e-01
-4.32185203e-01 1.25084734e+00 -7.01119184e-01 -7.70683944e-01
7.73274750e-02 -6.20811582e-01 -7.65804589e-01 3.74818772e-01
1.45606652e-01 -3.75044912e-01 -3.40239406e-01 9.90300536e-01
4.77140278e-01 -4.17725891e-01 7.54942298e-01 -9.51849520e-01
-9.97712731e-01 6.53356612e-01 -4.30032760e-01 -5.55164507e-03
4.63714302e-01 2.65794247e-01 7.63847947e-01 -7.95245826e-01
1.02679336e+00 1.02343869e+00 2.51509696e-01 6.11252487e-01
-8.83176029e-01 -6.23816192e-01 1.90558299e-01 2.69498199e-01
-1.59524584e+00 -4.51304644e-01 7.10360765e-01 -6.31613195e-01
1.01849353e+00 3.26718777e-01 7.61760652e-01 9.43321943e-01
3.77652884e-01 4.27442104e-01 9.67103302e-01 -2.74206549e-01
2.47753449e-02 3.38293970e-01 -2.33044758e-01 1.04794025e+00
3.40502709e-01 -6.89351708e-02 -6.80250108e-01 -1.72666162e-01
7.18798816e-01 -3.43027443e-01 -5.24599254e-01 1.04297847e-01
-1.14251375e+00 6.83998287e-01 6.98444426e-01 -8.33969563e-03
-4.41822648e-01 -3.33292812e-01 1.45690486e-01 1.07334599e-01
3.99468124e-01 3.47595990e-01 -8.12932700e-02 3.06738406e-01
-8.66185844e-01 4.87434939e-02 6.60994112e-01 1.26259887e+00
3.84496093e-01 -1.71459451e-01 -5.18792033e-01 7.37534881e-01
3.97053570e-01 3.45230609e-01 5.51810622e-01 -5.87659121e-01
6.91628635e-01 1.08952010e+00 -1.03859536e-01 -9.96485114e-01
-3.01786125e-01 -6.85546041e-01 -1.04415655e+00 -2.45644853e-01
5.55633008e-02 -1.18698455e-01 -1.32465577e+00 1.27093673e+00
2.47155339e-01 1.38581201e-01 3.61221135e-01 9.71168995e-01
1.54459405e+00 2.94682592e-01 1.33550033e-01 -4.36145812e-01
1.60667074e+00 -5.44243097e-01 -8.42136383e-01 -1.71919227e-01
6.79050148e-01 -8.39006722e-01 5.97001731e-01 1.83240756e-01
-9.99023736e-01 -3.84598345e-01 -9.94232059e-01 -2.74962157e-01
-2.58082062e-01 5.90268433e-01 6.38859808e-01 1.99012473e-01
-8.80797923e-01 3.50890425e-03 -6.29916906e-01 -1.85579553e-01
7.41334021e-01 3.82481039e-01 -4.34964538e-01 -3.22320133e-01
-1.12008440e+00 8.84902716e-01 4.76496130e-01 3.26963991e-01
-5.55813372e-01 -7.95327187e-01 -1.01531720e+00 -1.62267655e-01
6.66184962e-01 -1.36281264e+00 1.12279737e+00 -6.18588507e-01
-1.35879350e+00 9.19389188e-01 -3.73882294e-01 -2.82965362e-01
4.39202398e-01 1.56407475e-01 -6.23644412e-01 1.93805560e-01
3.23837101e-02 7.74726510e-01 7.72142291e-01 -1.18917882e+00
-5.83432734e-01 -1.85204700e-01 1.92903683e-01 2.08689183e-01
-9.06313658e-02 -1.32669404e-01 -7.85666227e-01 -6.90571606e-01
1.48441479e-01 -8.14222634e-01 -1.86278805e-01 -4.11805511e-02
-1.04494274e+00 -1.36695251e-01 2.62804270e-01 -7.29647875e-01
1.38886774e+00 -1.96155524e+00 -2.99358647e-03 2.64045626e-01
7.42342114e-01 2.10511878e-01 -5.84395565e-02 2.78241783e-01
-1.58896789e-01 1.20981336e-01 -1.17952548e-01 -2.04432160e-01
-3.54774445e-01 2.47525558e-01 -2.18040571e-01 1.23833194e-01
6.68136120e-01 1.02954996e+00 -6.62542701e-01 -8.20253193e-01
-1.24849238e-01 5.38877130e-01 -6.64868057e-01 3.35104227e-01
-3.58137727e-01 4.76595998e-01 -6.63783491e-01 8.06536973e-01
5.13716161e-01 -7.60298193e-01 2.43420020e-01 -6.55507326e-01
1.81091830e-01 5.48643291e-01 -6.79312885e-01 1.57311523e+00
-3.82830083e-01 2.70864159e-01 -5.78228295e-01 -6.06588662e-01
7.92671978e-01 4.20474529e-01 3.16083461e-01 -5.56250215e-01
1.77092791e-01 2.77543790e-03 2.34948725e-01 -9.06261325e-01
2.17347860e-01 -2.74988979e-01 3.20257455e-01 1.07414410e-01
1.22798339e-01 4.33831140e-02 1.76362261e-01 4.56001252e-01
1.11107385e+00 9.44954678e-02 5.25933027e-01 4.09418911e-01
6.33344591e-01 2.35501423e-01 7.24660695e-01 3.93252552e-01
2.44829834e-01 7.29278386e-01 7.48987138e-01 -3.33092481e-01
-5.12687981e-01 -9.88477290e-01 -9.63344574e-02 2.02463895e-01
-1.92811981e-01 -6.36335194e-01 -4.74871486e-01 -9.60906267e-01
-8.76030996e-02 6.66211784e-01 -6.68566644e-01 -3.69748950e-01
-1.70264795e-01 -7.79633284e-01 3.81775260e-01 5.08262634e-01
3.34942430e-01 -9.02416289e-01 -1.79280519e-01 2.14864999e-01
-3.63074034e-01 -1.45086050e+00 -4.38982815e-01 -3.68660092e-01
-7.61376262e-01 -1.26211190e+00 -6.36876047e-01 -6.74575984e-01
1.26733506e+00 -1.81426913e-01 9.03966725e-01 8.79386719e-03
-4.65539277e-01 1.07309923e-01 -2.53367603e-01 -5.33910513e-01
-7.04594135e-01 -9.58541483e-02 -2.80616999e-01 1.64921761e-01
5.45969307e-01 -2.64530897e-01 -6.98289633e-01 -7.78420791e-02
-9.57905471e-01 5.65004170e-01 1.02491093e+00 8.65420759e-01
7.65665412e-01 -8.99994224e-02 6.09686375e-01 -1.03659391e+00
7.86748648e-01 -4.09558445e-01 -4.98915642e-01 3.85906398e-01
-7.24774301e-01 2.05613360e-01 4.55716848e-01 -2.18396425e-01
-1.20956004e+00 1.30696923e-01 6.64159209e-02 -4.36756581e-01
7.85644725e-02 9.80067372e-01 -9.02606025e-02 1.32605687e-01
5.86320579e-01 3.92543823e-01 1.06717743e-01 -3.12449336e-01
4.22013879e-01 7.99988210e-01 4.13997293e-01 -1.89581990e-01
6.43392742e-01 1.82663321e-01 4.85776849e-02 -3.38487476e-01
-1.08708858e+00 -1.41362742e-01 -3.45424294e-01 -8.16710368e-02
7.59990811e-01 -1.06351542e+00 -6.04354024e-01 -2.49604173e-02
-1.29137146e+00 4.54778820e-01 -5.52254990e-02 6.76306963e-01
-2.97127455e-01 8.11572447e-02 -5.66140652e-01 -4.56691384e-01
-4.27674621e-01 -1.41612911e+00 9.35864508e-01 2.87985682e-01
-1.44904658e-01 -7.04459786e-01 -2.75992364e-01 6.23902321e-01
-2.47528758e-02 1.08840920e-01 1.41154337e+00 -3.83486390e-01
-8.19542289e-01 -5.41395135e-02 -5.71342349e-01 3.10473800e-01
5.39518178e-01 -1.48498621e-02 -8.79500270e-01 1.94145143e-02
-2.36836344e-01 -7.05852807e-02 7.77985632e-01 2.47429535e-01
1.13503873e+00 -6.46121025e-01 -6.21838331e-01 7.16251493e-01
1.34943318e+00 3.50754619e-01 6.48407102e-01 6.78653345e-02
8.44691038e-01 5.50198138e-01 4.25801188e-01 3.92954856e-01
8.80757630e-01 4.50157553e-01 2.63383299e-01 -7.66064152e-02
-4.26786363e-01 -4.24565226e-01 1.11924171e-01 1.06077504e+00
-3.27382416e-01 -2.68472344e-01 -9.60088670e-01 7.38859773e-01
-1.55363965e+00 -3.33377033e-01 1.08388342e-01 2.03386211e+00
1.40732133e+00 6.85742050e-02 -2.95884311e-01 -2.26943091e-01
4.08038497e-01 -1.93981245e-01 -5.54146945e-01 -9.73161906e-02
2.24679406e-03 1.13653764e-01 3.67848694e-01 2.85854757e-01
-6.89736009e-01 8.41949284e-01 5.31797123e+00 6.18650615e-01
-1.22451067e+00 -9.23233107e-03 4.91197944e-01 -1.88168973e-01
-3.23771983e-01 -1.22002274e-01 -9.27502573e-01 3.59804749e-01
7.71556318e-01 -4.17045325e-01 -7.84956813e-02 4.42765146e-01
2.59676069e-01 9.12009329e-02 -1.19037938e+00 9.83884871e-01
1.78903773e-01 -1.71443486e+00 7.33267188e-01 2.32282981e-01
6.15722537e-01 -3.06928992e-01 1.07930966e-01 1.44275188e-01
1.64134994e-01 -1.14550436e+00 1.89769343e-01 9.54971135e-01
1.28070068e+00 -6.32946789e-01 9.26464558e-01 -6.19229414e-02
-9.00170326e-01 1.20728247e-01 -3.32638711e-01 3.74738336e-01
4.70942743e-02 7.40555108e-01 -1.61326885e+00 1.03838861e+00
2.47509092e-01 7.62727141e-01 -7.07269967e-01 1.06487012e+00
-3.44081491e-01 6.00022197e-01 -5.54842837e-02 1.96467191e-01
7.79292956e-02 7.30571523e-02 4.76440370e-01 9.77926433e-01
5.15420198e-01 3.03836465e-01 8.24061595e-03 1.14707434e+00
-3.26725155e-01 3.50009739e-01 -7.56076694e-01 -3.22593480e-01
1.55916587e-01 9.58671391e-01 -3.96696925e-01 -4.85443115e-01
-7.07391679e-01 6.76355779e-01 2.15110913e-01 5.72408736e-01
-6.80505216e-01 -2.92432547e-01 4.70883250e-01 3.55793923e-01
2.55584866e-01 3.02608281e-01 -3.69833976e-01 -1.14359021e+00
3.37446362e-01 -9.92927313e-01 4.32637632e-01 -1.01905572e+00
-1.24293149e+00 9.65331912e-01 -1.88671350e-01 -1.83965385e+00
-4.53206599e-01 -4.36102003e-01 7.05062523e-02 9.01768744e-01
-1.79726338e+00 -1.46922922e+00 -3.40829343e-01 5.64545929e-01
2.87829906e-01 -4.64560449e-01 8.30655158e-01 2.29667813e-01
-5.84405124e-01 7.83196747e-01 -5.72942674e-01 2.77869403e-01
7.73957968e-01 -1.05181265e+00 6.18707947e-02 6.79296553e-01
-1.01568252e-01 9.23898995e-01 3.48967016e-01 -1.02159822e+00
-1.24504447e+00 -1.37978494e+00 1.04259312e+00 -2.94034153e-01
5.60908556e-01 6.99245557e-02 -9.21315849e-01 6.57084584e-01
5.39045893e-02 1.17209129e-01 8.48679841e-01 -2.23354101e-01
-3.30979258e-01 -1.08404964e-01 -1.05371428e+00 7.12246656e-01
1.09887457e+00 -5.67360461e-01 -6.37088954e-01 5.66110194e-01
9.20370698e-01 -7.48725355e-01 -1.13848758e+00 5.62185526e-01
4.22064573e-01 -4.91235912e-01 7.23192334e-01 -8.82804513e-01
8.71660352e-01 -5.07523417e-01 1.90168411e-01 -1.22654450e+00
-4.18192111e-02 -6.10624552e-01 -1.87349126e-01 1.16189170e+00
7.25844979e-01 -6.39613986e-01 4.79309291e-01 5.24344206e-01
-8.23260099e-02 -1.13174736e+00 -8.86678159e-01 -2.86947489e-01
-3.15682858e-01 -4.03096527e-01 7.44086146e-01 9.69086409e-01
9.41489562e-02 3.68054152e-01 -1.28927395e-01 4.90328550e-01
4.06195730e-01 6.93102106e-02 3.88225526e-01 -9.99664307e-01
-9.59142148e-02 -1.27147287e-01 -6.89780474e-01 -8.05437803e-01
-1.21463917e-01 -1.18432617e+00 -1.70717359e-01 -1.97691703e+00
3.58187437e-01 -3.54322195e-01 -3.26007307e-01 8.60056341e-01
-3.42219681e-01 1.92526862e-01 -3.67496572e-02 1.30209431e-01
-1.72730207e-01 3.91682029e-01 1.71940243e+00 -2.99904823e-01
-1.68628186e-01 -1.10604376e-01 -1.04969633e+00 6.86160922e-01
3.77154231e-01 -5.86560905e-01 -7.08552539e-01 -3.61075819e-01
4.56208110e-01 3.75984371e-01 3.63465846e-01 -7.33945608e-01
4.27853405e-01 -1.39468357e-01 2.97761232e-01 -4.91976768e-01
1.89502090e-01 -5.34475923e-01 1.89416558e-02 1.56012028e-01
-2.47175679e-01 -1.79031879e-01 2.47883394e-01 4.30713832e-01
-5.72793663e-01 2.37106174e-01 3.59417617e-01 -1.91967204e-01
-3.83658856e-01 2.36228913e-01 -1.51446372e-01 3.74144465e-02
8.10000658e-01 -2.74803400e-01 -7.18947828e-01 -3.36646497e-01
-9.45946693e-01 1.03999741e-01 4.73998562e-02 5.75331688e-01
1.06658363e+00 -1.13630974e+00 -9.04604852e-01 3.85470420e-01
5.89414895e-01 1.76138133e-01 4.60389525e-01 1.10551274e+00
-4.05161440e-01 7.25742817e-01 8.08758587e-02 -5.64297736e-01
-1.26923776e+00 4.49325711e-01 3.94107640e-01 -3.04129511e-01
-7.21679270e-01 7.41905332e-01 4.13883299e-01 -7.48278573e-02
1.01339154e-01 -7.53667116e-01 -4.10364091e-01 2.30627675e-02
6.48682594e-01 -3.67175072e-01 3.36147636e-01 -6.12896502e-01
-3.81209821e-01 6.01277292e-01 -5.85069060e-01 1.71546489e-01
1.25705838e+00 1.12251742e-02 -1.89103574e-01 1.22336354e-02
9.36013758e-01 4.90460396e-02 -7.41773665e-01 -4.31106210e-01
-2.13660687e-01 -2.55633980e-01 2.76772022e-01 -1.32243156e+00
-1.28267431e+00 5.24027228e-01 4.66335416e-01 -2.29057282e-01
1.35509229e+00 7.08753392e-02 6.89718127e-01 2.68410742e-01
1.28470317e-01 -5.95507801e-01 -2.45489493e-01 -1.37901017e-02
1.13051271e+00 -1.26858830e+00 2.12503403e-01 -1.01069033e+00
-9.75501657e-01 9.67364371e-01 6.19073331e-01 4.25542176e-01
4.83332932e-01 3.14731240e-01 3.49538773e-01 -4.29771483e-01
-9.77501035e-01 -4.13889080e-01 9.14332151e-01 6.69497311e-01
4.17006642e-01 -2.02583261e-02 -3.82174611e-01 8.40293825e-01
-3.42949629e-01 6.85910285e-01 7.19774127e-01 7.64769971e-01
1.20813824e-01 -9.78145421e-01 -1.34413064e-01 8.38370502e-01
-5.16322732e-01 -4.48087692e-01 -4.20350522e-01 6.31493270e-01
4.53838617e-01 8.91339540e-01 -2.38989651e-01 -5.70686340e-01
4.12237614e-01 9.47683603e-02 4.19535667e-01 -9.72385585e-01
-2.96739876e-01 9.68787223e-02 3.84462655e-01 -5.32321095e-01
-4.10046518e-01 -2.09023505e-01 -1.27579510e+00 2.34057441e-01
-1.40410617e-01 -9.51641202e-02 4.51873988e-01 1.11037481e+00
8.16496789e-01 1.06505299e+00 2.03334048e-01 4.35753204e-02
-1.14650056e-01 -9.37614083e-01 -2.74145216e-01 8.94299522e-02
4.98168439e-01 -6.99931562e-01 -1.59800686e-02 5.52082717e-01] | [15.059718132019043, -1.3923909664154053] |
680fee36-dd0e-464e-8dd4-7ccd082807cd | an-ensemble-of-convolutional-neural-networks | 2007.07966 | null | https://arxiv.org/abs/2007.07966v2 | https://arxiv.org/pdf/2007.07966v2.pdf | An Ensemble of Convolutional Neural Networks for Audio Classification | In this paper, ensembles of classifiers that exploit several data augmentation techniques and four signal representations for training Convolutional Neural Networks (CNNs) for audio classification are presented and tested on three freely available audio classification datasets: i) bird calls, ii) cat sounds, and iii) the Environmental Sound Classification dataset. The best performing ensembles combining data augmentation techniques with different signal representations are compared and shown to outperform the best methods reported in the literature on these datasets. The approach proposed here obtains state-of-the-art results in the widely used ESC-50 dataset. To the best of our knowledge, this is the most extensive study investigating ensembles of CNNs for audio classification. Results demonstrate not only that CNNs can be trained for audio classification but also that their fusion using different techniques works better than the stand-alone classifiers. | ['Michelangelo Paci', 'Gianluca Maguolo', 'Loris Nanni', 'Sheryl Brahnam'] | 2020-07-15 | null | null | null | null | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 3.98088574e-01 -2.69709498e-01 4.80091959e-01 -2.17835858e-01
-4.74641502e-01 -3.76762122e-01 4.38590139e-01 4.05474007e-01
-8.33694637e-01 5.49660742e-01 8.13604817e-02 -2.49343708e-01
-2.05883607e-01 -5.15892267e-01 -4.16187346e-01 -6.13363445e-01
-4.86610174e-01 1.67735800e-01 3.52001250e-01 -5.72546959e-01
-5.41789085e-02 6.65845394e-01 -2.38986421e+00 7.16786683e-01
3.03453237e-01 1.35894263e+00 -3.79264653e-01 9.99358833e-01
6.72396868e-02 5.27014434e-01 -8.39584529e-01 -2.28139982e-02
9.85150598e-03 -1.02776304e-01 -7.57385135e-01 -6.35403454e-01
6.33191824e-01 6.85276836e-02 -9.32647064e-02 4.17177975e-01
8.82234335e-01 4.36855614e-01 3.63405585e-01 -1.26108217e+00
-2.38069639e-01 1.09010661e+00 6.96493983e-02 8.06983173e-01
3.57247680e-01 3.16532813e-02 1.05329013e+00 -5.26818037e-01
-2.35404167e-03 9.59044456e-01 1.12137914e+00 4.31869984e-01
-1.24764776e+00 -1.09339809e+00 -1.55485690e-01 3.29310656e-01
-1.24411571e+00 -5.16185343e-01 9.29420412e-01 -4.56549942e-01
1.35738945e+00 5.18931925e-01 7.65217006e-01 1.28006721e+00
-2.58412600e-01 4.85016048e-01 1.22923136e+00 -6.56037509e-01
2.56783396e-01 6.30609542e-02 2.07579359e-01 2.99454600e-01
-5.37396856e-02 4.21995163e-01 -7.99993038e-01 -4.06290591e-01
2.14671627e-01 -4.73993212e-01 -1.48795128e-01 2.36489087e-01
-1.18151879e+00 6.79756880e-01 3.96938384e-01 1.07194746e+00
-5.17631710e-01 4.03094769e-01 8.92429709e-01 5.19290388e-01
6.98625982e-01 5.41081667e-01 -7.47277856e-01 -3.96621048e-01
-9.86792386e-01 4.35825974e-01 7.61978924e-01 3.06470782e-01
2.37067744e-01 8.13164532e-01 5.07420637e-02 9.97769237e-01
1.11329094e-01 3.85097191e-02 7.48915434e-01 -4.08376783e-01
3.27826798e-01 2.08884016e-01 -3.41091901e-01 -6.95505559e-01
-6.70837700e-01 -9.35313821e-01 -8.73074234e-01 2.91938603e-01
3.02944005e-01 -2.39849061e-01 -7.06454098e-01 1.71313035e+00
-6.07693009e-02 6.54475749e-01 9.76535007e-02 5.25879860e-01
1.44602406e+00 4.75018024e-01 1.60838842e-01 3.48468363e-01
1.18100405e+00 -6.90768540e-01 -6.21442318e-01 3.89274247e-02
4.31609094e-01 -8.04109991e-01 6.28045142e-01 7.17760861e-01
-8.18771482e-01 -1.15050828e+00 -1.40688777e+00 3.50787967e-01
-8.85114610e-01 4.79875416e-01 8.54043007e-01 1.09144604e+00
-1.17580962e+00 7.13972449e-01 -6.03413522e-01 -3.00036013e-01
3.75573486e-01 6.60555363e-01 -4.74191964e-01 5.40699363e-01
-1.29411197e+00 4.22879815e-01 5.03207147e-01 1.79919712e-02
-9.27466929e-01 -8.81360233e-01 -6.79422855e-01 2.25455567e-01
1.45437807e-01 -4.16397929e-01 1.48386395e+00 -1.08596706e+00
-1.69864726e+00 6.10925794e-01 5.08185267e-01 -1.07114673e+00
2.73234248e-01 -4.36277092e-01 -7.04822898e-01 -2.97303405e-02
-3.30895454e-01 7.96436310e-01 6.90314651e-01 -1.06685901e+00
-5.00139892e-01 -4.24134359e-02 -3.64782065e-02 -3.81047368e-01
-5.63493252e-01 3.14707935e-01 5.29606700e-01 -1.00805938e+00
-2.40185738e-01 -8.58164489e-01 -2.48323213e-02 -4.33297038e-01
-3.31448495e-01 -1.95048660e-01 7.67688274e-01 -5.26831448e-01
1.22511303e+00 -2.22969484e+00 5.01453765e-02 4.69693169e-02
-1.37053490e-01 5.20056903e-01 -2.78212696e-01 4.41855222e-01
-6.33478403e-01 1.62894443e-01 -2.89333344e-01 -5.98116040e-01
5.74459210e-02 2.15850547e-01 -3.93429756e-01 5.49613275e-02
2.82034934e-01 3.40841800e-01 -5.71687520e-01 -3.30937542e-02
2.88623184e-01 7.12797344e-01 -6.21061027e-01 1.48924710e-02
1.17612384e-01 4.43956584e-01 7.24573880e-02 4.60354835e-01
4.73459959e-01 5.37550211e-01 -2.44823217e-01 -1.76762000e-01
-3.41647118e-01 4.29025501e-01 -1.23047519e+00 1.59096384e+00
-6.24272764e-01 8.93403411e-01 -7.35177621e-02 -1.17476487e+00
1.03892493e+00 9.37236249e-01 4.41762060e-01 -5.95813990e-02
4.36156213e-01 4.71073359e-01 6.27573192e-01 -2.72995830e-01
3.29014331e-01 -1.56235486e-01 5.31518310e-02 2.65438288e-01
9.17272925e-01 -1.09026484e-01 1.91612229e-01 -2.54017234e-01
1.01667583e+00 -4.75154519e-02 5.05608469e-02 -1.72536805e-01
7.66912222e-01 -5.15254736e-01 1.95604637e-01 8.35830152e-01
-1.02855362e-01 4.99935597e-01 -3.42100137e-03 -5.94802976e-01
-8.19243252e-01 -5.91184556e-01 -1.81160763e-01 1.54553223e+00
-9.50910509e-01 -4.93665487e-01 -7.26344109e-01 -2.97106922e-01
-9.61641744e-02 5.57362616e-01 -9.46812689e-01 -3.75959575e-02
-3.00824761e-01 -7.99933851e-01 1.46941185e+00 8.58819962e-01
5.51250935e-01 -1.50647545e+00 -7.59860754e-01 4.82480377e-01
-1.41830686e-02 -1.20921576e+00 3.36679041e-01 6.82354510e-01
-8.62397194e-01 -9.21504140e-01 -4.78233337e-01 -5.35212278e-01
-1.97304115e-01 -1.65194914e-01 9.96634185e-01 2.10884556e-01
-3.00050318e-01 5.95517218e-01 -7.82579899e-01 -1.17072606e+00
-5.94799101e-01 3.12458992e-01 2.90085196e-01 1.77178949e-01
1.90724254e-01 -1.14503217e+00 -6.16199858e-02 -4.19685710e-03
-9.52248514e-01 -4.15447861e-01 3.66707921e-01 7.26921201e-01
1.53607219e-01 3.46565470e-02 9.14371729e-01 -3.07765484e-01
5.66009164e-01 -2.87200689e-01 -3.13640684e-01 -2.98155814e-01
-2.44098194e-02 -4.22168702e-01 6.26680374e-01 -5.73189437e-01
-6.67153716e-01 1.65754393e-01 -7.66149640e-01 -3.58252823e-01
-1.12387621e+00 4.14532065e-01 3.31090391e-01 -3.39739382e-01
8.57015371e-01 1.60966426e-01 -1.99268341e-01 -9.73637819e-01
7.81238228e-02 7.21795440e-01 6.46832824e-01 -4.45650965e-01
4.74925190e-01 3.93564582e-01 9.11381543e-02 -1.13858366e+00
-5.07407010e-01 -5.41839123e-01 -9.55899715e-01 -4.26881164e-01
7.34919429e-01 -6.49298191e-01 -5.79737604e-01 7.52922654e-01
-1.03546023e+00 -9.43710357e-02 -4.91727322e-01 6.75683379e-01
-4.56384927e-01 -3.29629444e-02 -5.39407730e-01 -1.20951116e+00
-3.27301085e-01 -9.47615027e-01 9.90889668e-01 -5.63465245e-02
-2.30862513e-01 -7.15415299e-01 3.37339491e-01 1.12088867e-01
7.82336593e-01 3.21933776e-01 5.14843404e-01 -1.25888002e+00
8.76480266e-02 -3.47792000e-01 3.61745298e-01 7.38785028e-01
-8.41970742e-02 3.13053191e-01 -1.91821694e+00 -6.15600869e-02
-3.46353173e-01 -3.40162426e-01 1.29891121e+00 3.28658223e-01
1.34468246e+00 -8.31638090e-03 6.34984076e-02 3.39104474e-01
9.32037354e-01 5.47466457e-01 5.31514764e-01 5.28327823e-01
9.72852781e-02 4.58385110e-01 4.46367897e-02 4.26681697e-01
-2.73590028e-01 7.42202640e-01 6.25768840e-01 -6.25049695e-02
-8.82400274e-02 2.44986102e-01 1.93804994e-01 8.91944468e-01
-4.95449275e-01 -1.91895977e-01 -1.06787348e+00 6.31616414e-01
-1.49973810e+00 -1.00259614e+00 -1.16752900e-01 1.89351940e+00
4.46428210e-01 1.36941224e-01 4.91019189e-01 1.30377996e+00
3.28495264e-01 3.38103436e-02 2.37855464e-01 -5.90333283e-01
-1.85539588e-01 1.12275136e+00 -1.91050097e-01 -3.62626240e-02
-1.78629053e+00 4.64581251e-01 6.85497284e+00 5.88292360e-01
-1.12763500e+00 3.43935996e-01 2.35643610e-01 -1.55486166e-01
5.58405817e-01 -4.27430034e-01 -4.37431931e-01 2.04245523e-01
1.44694245e+00 4.28961873e-01 2.44789362e-01 7.10134208e-01
-2.67076135e-01 2.19583318e-01 -9.32050109e-01 8.71521473e-01
1.31688684e-01 -1.18231165e+00 -9.15862620e-02 -1.54194981e-01
3.92952889e-01 3.21975023e-01 2.48135239e-01 4.92955655e-01
5.88997360e-03 -9.87925947e-01 8.65275443e-01 5.35794854e-01
3.16679209e-01 -9.69696403e-01 1.33887053e+00 -5.94956726e-02
-1.33107793e+00 -4.37373549e-01 3.81642133e-02 -2.59105891e-01
-1.21676847e-01 2.51835436e-01 -5.85792780e-01 8.81886959e-01
1.32316101e+00 5.88971019e-01 -9.75379229e-01 1.40406549e+00
1.97842479e-01 1.18219805e+00 -6.51347339e-01 -1.61268204e-01
4.94968474e-01 4.14789736e-01 7.47176647e-01 1.73421299e+00
5.14120460e-01 -3.02616686e-01 -1.82512328e-01 5.92231810e-01
1.13518029e-01 3.52083981e-01 -6.69973731e-01 -2.84875035e-01
1.87604874e-01 1.38206971e+00 -7.05934644e-01 -4.18017864e-01
-1.47043094e-01 2.38281384e-01 -1.10562466e-01 5.45714647e-02
-6.01948321e-01 -5.95654011e-01 6.65980935e-01 -3.11568528e-01
4.19024557e-01 4.64388952e-02 -3.03681232e-02 -5.35026312e-01
-3.19318354e-01 -7.74860561e-01 5.21874368e-01 -8.69267941e-01
-1.23367000e+00 1.09617507e+00 1.80092573e-01 -1.37011576e+00
-2.52647042e-01 -7.97861636e-01 -7.81534255e-01 5.12834132e-01
-1.31995106e+00 -1.06499636e+00 -4.09388602e-01 3.00481409e-01
2.80969232e-01 -8.83597434e-01 1.27751648e+00 6.37412190e-01
-3.23664308e-01 5.04274845e-01 -6.42966554e-02 2.53380775e-01
4.37477440e-01 -1.18209648e+00 2.61850506e-01 3.95172447e-01
7.44106591e-01 2.08466887e-01 6.47308528e-01 8.48717168e-02
-5.67903340e-01 -9.04314280e-01 4.76708651e-01 -1.25181720e-01
7.71519601e-01 -3.20319295e-01 -1.05370224e+00 5.19418895e-01
6.16627276e-01 6.56142365e-05 1.36115491e+00 2.36879990e-01
-5.21019101e-01 -3.04159462e-01 -9.86215472e-01 -8.85357484e-02
5.32096088e-01 -4.96637344e-01 -7.44550467e-01 -2.02033110e-02
4.88535643e-01 -2.41100103e-01 -9.04793561e-01 6.29919231e-01
7.45575309e-01 -1.17906249e+00 1.10253060e+00 -8.56240630e-01
3.58442396e-01 -1.93628088e-01 -3.24023664e-01 -1.57784796e+00
-2.39989031e-02 -2.94698805e-01 -2.84022273e-04 1.39715672e+00
4.05051678e-01 -6.84132695e-01 2.09456757e-01 -5.21568596e-01
-4.93025541e-01 -5.43449998e-01 -1.49923146e+00 -8.43706548e-01
2.96054203e-02 -1.07691228e+00 8.55012476e-01 9.44158733e-01
-3.52723897e-01 2.57878631e-01 -1.18826434e-01 -5.96356131e-02
2.26262316e-01 -3.79650414e-01 6.95295870e-01 -1.89524782e+00
-2.57884264e-01 -7.53767967e-01 -1.03596210e+00 4.62471768e-02
2.40387633e-01 -9.13335323e-01 -1.50649950e-01 -9.73059177e-01
-3.47253233e-01 -3.91834915e-01 -9.07514751e-01 8.54305744e-01
3.75669748e-01 7.56448090e-01 2.60188520e-01 -5.22728026e-01
-1.55402601e-01 4.81166035e-01 4.23802376e-01 -2.35697389e-01
-1.50092110e-01 1.48877457e-01 -4.35452461e-01 7.58831620e-01
1.23467290e+00 -4.25537258e-01 -1.49093330e-01 -2.88995914e-02
1.15110502e-01 -3.87210429e-01 8.54913771e-01 -1.94303679e+00
-2.76475623e-02 7.79781520e-01 3.75307769e-01 -6.98829353e-01
7.38635540e-01 -7.87704825e-01 2.11225197e-01 4.83399540e-01
-6.40161216e-01 -5.58040254e-02 1.07645416e+00 3.61284971e-01
-5.41263342e-01 -3.64510030e-01 7.56913245e-01 8.43661204e-02
-3.72916192e-01 -6.65163323e-02 -9.26216006e-01 -4.32142913e-01
5.81043661e-01 -3.49398237e-03 -4.43884470e-02 -4.24289763e-01
-1.31149769e+00 -5.22506237e-01 -5.20542681e-01 7.82553017e-01
4.14141297e-01 -1.46936202e+00 -8.06045175e-01 2.56772697e-01
1.63948491e-01 -4.94196355e-01 4.86985266e-01 9.73979235e-01
-2.45535389e-01 6.30549788e-01 -5.26144147e-01 -6.93198621e-01
-1.80481291e+00 2.76368439e-01 6.45468652e-01 -7.78047442e-02
-3.31230223e-01 9.76860106e-01 -3.81113887e-01 -7.78363049e-01
4.46804821e-01 -7.81452894e-01 -7.79335737e-01 3.89220119e-01
5.87655187e-01 3.81580204e-01 6.03195608e-01 -8.56411397e-01
-3.12601298e-01 2.14041084e-01 3.14588457e-01 -1.85299352e-01
1.76654458e+00 6.73744321e-01 1.13239707e-02 8.85348320e-01
8.96986902e-01 -3.68874162e-01 -3.42539251e-01 -1.27265617e-01
-1.09171875e-01 -1.04495943e-01 2.93075770e-01 -9.53500628e-01
-1.15953887e+00 1.41790485e+00 1.09942520e+00 6.95063174e-01
1.35021615e+00 -3.93935591e-01 3.21937501e-01 5.69587588e-01
1.06848396e-01 -9.98463929e-01 5.94494045e-02 5.51708102e-01
1.24055219e+00 -7.43159533e-01 -2.51830578e-01 -1.05955876e-01
-3.60896349e-01 1.52200532e+00 6.33395314e-01 -1.66360825e-01
9.53505516e-01 4.93032515e-01 -4.21484821e-02 -8.08857288e-03
-9.64299500e-01 -5.56555390e-01 5.05694509e-01 8.64465714e-01
8.07559848e-01 -1.12084173e-01 2.38142852e-02 8.86792004e-01
-5.84794104e-01 3.32147436e-04 2.80883193e-01 9.59393680e-01
-2.27634549e-01 -9.67035294e-01 -5.42820752e-01 4.44777846e-01
-8.25225651e-01 -3.99884544e-02 -8.06858063e-01 9.88431275e-01
5.76537132e-01 1.10454023e+00 1.77663997e-01 -6.76799655e-01
4.63779181e-01 6.55295610e-01 4.55746680e-01 -3.88119787e-01
-1.51873267e+00 5.18406462e-03 3.62028748e-01 -1.96866497e-01
-1.02571952e+00 -6.29628718e-01 -6.98692203e-01 2.41945595e-01
-5.74734151e-01 3.36914748e-01 8.36013496e-01 7.84061015e-01
1.49180591e-01 1.35898173e+00 3.83388430e-01 -1.45999575e+00
-2.33987212e-01 -1.57551897e+00 -5.59278190e-01 1.89982876e-01
4.53775108e-01 -9.58921134e-01 -5.04908383e-01 -6.85438588e-02] | [15.242955207824707, 5.2586588859558105] |
687f00c3-e07d-4363-9655-a64a42c14fd3 | cross-lingual-visual-pre-training-for | 2101.10044 | null | https://arxiv.org/abs/2101.10044v2 | https://arxiv.org/pdf/2101.10044v2.pdf | Cross-lingual Visual Pre-training for Multimodal Machine Translation | Pre-trained language models have been shown to improve performance in many natural language tasks substantially. Although the early focus of such models was single language pre-training, recent advances have resulted in cross-lingual and visual pre-training methods. In this paper, we combine these two approaches to learn visually-grounded cross-lingual representations. Specifically, we extend the translation language modelling (Lample and Conneau, 2019) with masked region classification and perform pre-training with three-way parallel vision & language corpora. We show that when fine-tuned for multimodal machine translation, these models obtain state-of-the-art performance. We also provide qualitative insights into the usefulness of the learned grounded representations. | ['Lucia Specia', 'Aykut Erdem', 'Erkut Erdem', 'Pranava Madhyastha', 'Mustafa Sercan Amac', 'Menekse Kuyu', 'Ozan Caglayan'] | 2021-01-25 | null | https://aclanthology.org/2021.eacl-main.112 | https://aclanthology.org/2021.eacl-main.112.pdf | eacl-2021-2 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [-8.30747038e-02 1.36455745e-01 -4.09982026e-01 -2.47048736e-01
-1.44317794e+00 -8.50952566e-01 1.15494001e+00 -1.91855356e-02
-2.69872665e-01 4.92758811e-01 4.37033892e-01 -5.44922709e-01
5.84071755e-01 -3.17535877e-01 -8.72732103e-01 -3.61471832e-01
1.58914149e-01 5.89460313e-01 9.68364417e-04 -2.79149473e-01
1.67460278e-01 4.63325709e-01 -1.10672855e+00 8.43445659e-01
6.55092597e-01 4.53446060e-01 2.10077867e-01 7.28104830e-01
-2.60411978e-01 4.01956826e-01 -2.27455199e-01 -4.88306344e-01
5.42268790e-02 -6.87155902e-01 -9.54525054e-01 1.77238449e-01
6.04853332e-01 -2.56838016e-02 -2.15625241e-01 7.73680568e-01
3.56043458e-01 -9.18377638e-02 8.08705866e-01 -9.24981117e-01
-1.14870429e+00 4.31829363e-01 -9.78016555e-01 1.29980311e-01
3.22378755e-01 3.27631503e-01 9.14817095e-01 -1.25050521e+00
8.93186390e-01 1.54604435e+00 5.88744879e-01 5.21157563e-01
-1.60334301e+00 -2.20517248e-01 1.12493701e-01 -2.95538250e-02
-1.23478806e+00 -7.32624590e-01 5.74865520e-01 -5.70777476e-01
1.26836944e+00 -1.71711415e-01 3.81376892e-01 1.25320959e+00
3.38610768e-01 7.91249633e-01 1.56919563e+00 -1.05440319e+00
-2.05099285e-01 4.56821024e-01 -3.30967724e-01 9.90403473e-01
-1.19917974e-01 3.51119578e-01 -6.26255453e-01 9.21283513e-02
8.69493425e-01 -3.16595912e-01 -2.96792034e-02 -8.27571094e-01
-1.50478601e+00 8.95487309e-01 7.36351132e-01 4.72263068e-01
-4.09871228e-02 1.75515071e-01 4.17984515e-01 2.01738179e-01
7.03552008e-01 4.74116802e-01 -1.95685238e-01 1.68266237e-01
-1.14628220e+00 -8.23345706e-02 2.97736019e-01 9.02947962e-01
8.70751500e-01 -1.84045564e-02 -1.91420853e-01 8.28472078e-01
5.12938321e-01 4.67205524e-01 5.46015918e-01 -8.01416874e-01
6.35581136e-01 5.04181206e-01 -1.67527288e-01 -6.16876483e-01
-3.15086246e-01 -4.49436188e-01 -5.09923518e-01 3.61701638e-01
5.86940050e-01 9.00964513e-02 -1.19620967e+00 1.50950515e+00
-6.18740693e-02 -4.99820352e-01 2.83666015e-01 8.58114481e-01
4.93277550e-01 6.72372460e-01 2.47521177e-01 4.24352027e-02
1.27903438e+00 -1.23273349e+00 -3.99651468e-01 -4.12367404e-01
7.69602358e-01 -1.09503901e+00 1.43172574e+00 4.46784273e-02
-1.08794641e+00 -5.57597458e-01 -8.09958994e-01 -4.32440639e-01
-4.76523787e-01 3.25028419e-01 5.07216871e-01 4.09226775e-01
-1.22785413e+00 9.90530923e-02 -9.08418238e-01 -9.36228096e-01
4.05980617e-01 7.25596547e-02 -5.47936320e-01 -1.50359631e-01
-7.70013273e-01 1.38518381e+00 3.76524441e-02 -7.21449107e-02
-1.10183609e+00 -1.99798480e-01 -9.25498962e-01 -3.12699407e-01
1.17900297e-01 -7.91293442e-01 1.18560314e+00 -1.32189751e+00
-1.34174943e+00 1.49244750e+00 -3.36734384e-01 -3.33092809e-01
4.73114073e-01 4.31785956e-02 -1.19545750e-01 3.10793370e-01
1.45264313e-01 1.29772532e+00 5.94077229e-01 -1.71351981e+00
-2.57685781e-01 -4.49880838e-01 9.87467989e-02 3.55364978e-01
-1.38496058e-02 3.52711886e-01 -6.73968554e-01 -7.10790157e-01
-1.05527453e-01 -9.14241314e-01 -9.34240371e-02 2.32950315e-01
-1.56800717e-01 4.74837348e-02 4.58042890e-01 -8.27821314e-01
4.72625107e-01 -1.91607356e+00 4.14857537e-01 -7.74059445e-02
3.92691605e-02 -1.07614540e-01 -4.47455734e-01 5.65522552e-01
2.55834013e-02 1.10720702e-01 -2.88307518e-01 -6.48176074e-01
-2.44620256e-02 2.00409830e-01 -3.28308851e-01 7.08598256e-01
3.65600318e-01 1.32842040e+00 -7.09784210e-01 -7.10881948e-01
3.18453193e-01 7.26075470e-01 -3.86382312e-01 8.24637711e-02
-2.33306438e-01 6.17112279e-01 -2.01263130e-02 7.13321865e-01
2.41531685e-01 -1.76891252e-01 9.76249650e-02 -1.26004502e-01
-2.25889146e-01 3.29513758e-01 -4.10081774e-01 2.10447168e+00
-7.27975607e-01 1.06155491e+00 1.53805256e-01 -9.43059206e-01
6.65738642e-01 3.31263155e-01 6.27581626e-02 -9.64382946e-01
9.77245718e-03 2.02193618e-01 1.82776095e-03 -4.60111022e-01
3.32135469e-01 -3.80768508e-01 9.11110044e-02 3.98625016e-01
2.33864039e-01 -2.96184659e-01 -4.43208143e-02 1.65431008e-01
3.38686705e-01 8.13458920e-01 3.80680412e-01 -3.18818748e-01
3.18644941e-01 6.15432501e-01 -8.80532712e-02 4.03379947e-01
-3.09409767e-01 9.69438314e-01 3.05019885e-01 -2.44937614e-01
-1.28662598e+00 -1.10594237e+00 -1.24578290e-01 1.38686025e+00
-5.92160150e-02 -2.08873779e-01 -8.88098538e-01 -6.73596621e-01
-2.28736460e-01 6.77016973e-01 -7.70509124e-01 5.68958446e-02
-5.46341062e-01 -4.84872848e-01 6.53722227e-01 6.70534134e-01
1.53010428e-01 -1.06170189e+00 -4.97846723e-01 -2.20677391e-01
-8.11132789e-02 -1.20846260e+00 -3.99633884e-01 1.34432197e-01
-7.76511073e-01 -7.03217030e-01 -1.15385306e+00 -1.13738811e+00
7.76446223e-01 3.89300138e-01 1.35228074e+00 -1.05504222e-01
-3.89421612e-01 6.74082220e-01 -2.43799731e-01 -1.35738164e-01
-5.86925149e-01 3.80804129e-02 2.54217535e-02 -2.37763807e-01
3.08424950e-01 -1.32361889e-01 -3.28557014e-01 2.27734476e-01
-5.70066512e-01 3.78440708e-01 7.60028064e-01 8.33197474e-01
7.30569839e-01 -8.67604971e-01 2.57962435e-01 -5.00194728e-01
6.15669847e-01 -1.17988445e-01 -5.37035406e-01 5.17597556e-01
-5.43306530e-01 1.55063793e-01 3.33203852e-01 -4.26353991e-01
-8.24636042e-01 1.53157845e-01 1.67079225e-01 -5.25561213e-01
-3.95250231e-01 5.16976893e-01 2.16568649e-01 -2.50467211e-01
8.97409081e-01 2.33016446e-01 -1.15726963e-02 -3.64479661e-01
9.23433304e-01 5.45268297e-01 5.03041267e-01 -8.51022720e-01
7.78546154e-01 4.19343412e-01 -2.84360230e-01 -7.33702660e-01
-4.96759593e-01 -2.46091351e-01 -1.14423776e+00 -6.82626888e-02
1.12701559e+00 -1.16908658e+00 -7.95709491e-02 -1.47905976e-01
-1.19381166e+00 -4.70676631e-01 -6.67351261e-02 5.31490684e-01
-7.48007894e-01 3.12218815e-01 -4.82765824e-01 -7.42179155e-01
-7.93435350e-02 -1.28578103e+00 1.39793646e+00 -1.05677895e-01
-2.01694921e-01 -1.20288956e+00 3.98204476e-01 5.49649596e-01
2.85991341e-01 -1.06320232e-02 9.77244973e-01 -3.35995823e-01
-5.67561030e-01 1.28691196e-01 -5.41556954e-01 -5.96397221e-02
-6.06080815e-02 -9.54589769e-02 -1.17032087e+00 -4.06132251e-01
-4.17640358e-01 -6.64873898e-01 9.50766563e-01 4.22058254e-01
5.15283227e-01 7.73274824e-02 -4.84782398e-01 6.62933469e-01
1.30187845e+00 -2.16967702e-01 5.34085512e-01 4.81496632e-01
8.03943634e-01 1.00062764e+00 3.30559045e-01 -4.37770486e-01
6.04625344e-01 9.75580931e-01 1.16504498e-01 -6.89039528e-01
-4.50768977e-01 -2.85175592e-01 5.91097236e-01 6.27006590e-01
-2.86457479e-01 -6.09624535e-02 -1.12046421e+00 7.46276796e-01
-1.85361183e+00 -8.44847441e-01 1.37114227e-01 1.93488467e+00
8.41692805e-01 -8.74103531e-02 3.53719175e-01 -5.22044480e-01
4.45689589e-01 -7.98987076e-02 -2.03502223e-01 -6.21436775e-01
-2.52728224e-01 7.77631812e-03 3.66806239e-01 9.00185764e-01
-9.04964089e-01 1.52721035e+00 6.96240282e+00 4.69813526e-01
-1.26055515e+00 3.61763865e-01 7.80291617e-01 -1.06751136e-02
-4.68471199e-01 4.27269116e-02 -5.69767594e-01 -2.10746825e-01
8.00951838e-01 2.01839775e-01 5.34417868e-01 4.69008982e-01
1.91130280e-01 -2.09479734e-01 -1.28099346e+00 9.70048070e-01
4.66886222e-01 -1.35641253e+00 4.92002703e-02 2.40543500e-01
8.90644431e-01 3.88635397e-01 1.85752273e-01 3.70144367e-01
3.38981718e-01 -1.28906345e+00 7.42354631e-01 2.45176733e-01
1.06508112e+00 -4.69366938e-01 2.72740871e-01 2.42133200e-01
-1.08267927e+00 2.90890694e-01 -3.04915220e-01 8.86597186e-02
1.34654894e-01 2.97081694e-02 -7.86158681e-01 5.11899531e-01
5.44065833e-01 5.54171443e-01 -7.97346115e-01 6.64547384e-01
-1.75693735e-01 4.24747616e-01 -7.62205496e-02 2.51938283e-01
5.25301635e-01 -1.60311282e-01 2.24053472e-01 1.57183003e+00
9.28112939e-02 -4.36965197e-01 3.93804312e-01 1.00652778e+00
-2.48549581e-02 2.39260003e-01 -9.14931118e-01 -2.04384491e-01
-1.35811001e-01 1.14073050e+00 -7.06203878e-01 -3.51807535e-01
-8.65412951e-01 1.10233855e+00 6.41615689e-01 8.06493342e-01
-6.37045860e-01 1.08394548e-01 5.07339776e-01 1.74633816e-01
2.00396806e-01 -5.49229145e-01 -5.47395945e-01 -1.45896339e+00
-1.99387953e-01 -8.84258628e-01 8.91810283e-02 -1.07468033e+00
-1.15102923e+00 7.03816533e-01 8.25610235e-02 -1.10779023e+00
-4.74813372e-01 -9.60159838e-01 -3.64532411e-01 1.21950984e+00
-1.53339005e+00 -1.93116403e+00 1.48452267e-01 4.86048907e-01
8.84719193e-01 -3.75787914e-01 1.06436384e+00 -3.88636850e-02
-3.61259609e-01 5.51438928e-01 9.60118100e-02 2.82727122e-01
9.06156480e-01 -1.06209528e+00 5.34618795e-01 9.09946501e-01
7.20905781e-01 7.52331555e-01 5.03627598e-01 -5.80913842e-01
-1.59441400e+00 -7.73411870e-01 9.52887535e-01 -7.09442258e-01
7.15314150e-01 -6.99661016e-01 -7.96783090e-01 1.06985521e+00
1.02289402e+00 1.15958907e-01 3.80930901e-01 1.96983740e-01
-6.75225377e-01 2.45100200e-01 -8.83973241e-01 7.27388680e-01
8.05436552e-01 -9.68180716e-01 -7.50599921e-01 4.31976467e-01
3.81786615e-01 -3.13125879e-01 -5.48225403e-01 7.48903230e-02
5.71229637e-01 -8.33369017e-01 1.14112914e+00 -6.58897161e-01
5.20653844e-01 -2.82530129e-01 -2.83655405e-01 -1.38722444e+00
-2.18201473e-01 -2.95725465e-01 2.27514967e-01 1.09541154e+00
7.27479160e-01 -3.08364600e-01 3.48897696e-01 6.86734989e-02
-9.97022614e-02 -6.45933628e-01 -8.20014298e-01 -5.44753611e-01
5.60754836e-01 -1.88930035e-01 -2.02590570e-01 1.05845594e+00
1.36896551e-01 7.16133773e-01 -2.78654158e-01 4.32567708e-02
4.63187009e-01 2.64625430e-01 8.30028892e-01 -7.93596387e-01
-2.81245232e-01 -6.50018752e-01 -1.72381103e-01 -1.00093687e+00
2.66599208e-01 -1.11551046e+00 1.50250062e-01 -1.85975206e+00
4.40073311e-01 5.45601919e-02 -3.65422256e-02 6.61164820e-01
1.69540927e-01 5.94868720e-01 1.81270257e-01 3.96654814e-01
-3.94731164e-01 4.11825359e-01 1.24151480e+00 -2.49374703e-01
-9.14695710e-02 -5.16431391e-01 -5.82421422e-01 4.77554739e-01
6.66376829e-01 -1.91886425e-01 -3.30447406e-01 -8.50266039e-01
6.03049211e-02 -2.37974763e-01 5.15904248e-01 -5.24937689e-01
-7.28754774e-02 -2.07746431e-01 6.93519056e-01 -4.55096722e-01
3.89886171e-01 -5.85339189e-01 -3.97107989e-01 2.98220962e-01
-5.36950231e-01 4.43977714e-01 5.51631868e-01 3.29125196e-01
-2.46579319e-01 1.18118994e-01 8.68742645e-01 -3.63278389e-01
-5.93114853e-01 -2.02915639e-01 -4.33526456e-01 6.08390905e-02
8.18720222e-01 -2.13850439e-01 -4.67730910e-01 -3.21409911e-01
-7.41639435e-01 1.14237666e-01 9.02013898e-01 5.43044448e-01
4.46907789e-01 -1.34901929e+00 -7.74820745e-01 4.37085219e-02
4.18986648e-01 -3.78202319e-01 -2.39274636e-01 1.09546232e+00
-5.94112635e-01 6.76532269e-01 -3.40788841e-01 -1.04747617e+00
-1.20807326e+00 7.61336803e-01 4.68452036e-01 -7.73619115e-02
-6.43984675e-01 6.39264047e-01 6.01403773e-01 -6.59916103e-01
7.42482841e-02 -3.33402395e-01 7.26550259e-03 -3.94667648e-02
1.93015054e-01 -1.73789382e-01 -1.44296855e-01 -1.04185534e+00
-5.77528298e-01 1.09135127e+00 1.66870020e-02 -8.42422962e-01
9.79564309e-01 -3.13858360e-01 -2.12238263e-03 6.85189962e-01
1.23729742e+00 1.72222242e-01 -1.13147032e+00 -2.78668642e-01
-1.03646845e-01 -1.89511567e-01 9.27085057e-02 -1.03906310e+00
-7.70684361e-01 1.46651077e+00 7.59484172e-01 -2.39977926e-01
8.56897235e-01 4.14177507e-01 1.53699517e-01 2.92959005e-01
2.81902432e-01 -8.48229170e-01 5.62539697e-02 4.86041039e-01
1.03140604e+00 -1.58825183e+00 2.05173753e-02 -1.31416023e-01
-9.21981394e-01 1.02521670e+00 5.33332825e-01 5.01275212e-02
1.14698894e-01 1.44960731e-01 6.03557467e-01 -1.24134928e-01
-6.08311594e-01 -4.98868972e-01 8.43888283e-01 8.37906480e-01
9.88777816e-01 3.93917710e-02 5.19182496e-02 -6.77088946e-02
5.27220406e-02 -2.24751800e-01 1.41112059e-01 7.51108885e-01
-4.20483381e-01 -9.99689817e-01 -6.39325917e-01 -3.14819068e-01
-1.29027694e-01 -4.60947752e-01 -7.34040141e-01 9.40013826e-01
-1.79600716e-01 7.59774983e-01 1.25696182e-01 -6.67586774e-02
2.00578094e-01 4.91796166e-01 1.04204571e+00 -7.16384411e-01
-5.84767878e-01 6.46084070e-01 4.87182699e-02 -3.42624098e-01
-5.32647491e-01 -5.94705760e-01 -1.23622465e+00 -2.66714543e-02
-1.64151266e-02 -2.08268002e-01 7.66643465e-01 1.08565152e+00
4.25723881e-01 3.64599526e-01 9.05040577e-02 -1.18779850e+00
-1.78586259e-01 -1.03961623e+00 8.18933770e-02 1.79823652e-01
4.54512805e-01 -5.51024556e-01 -6.41001761e-02 4.37555850e-01] | [11.358719825744629, 1.4969853162765503] |
bd503b63-f705-430e-a603-7056438d45e8 | attention-flows-analyzing-and-comparing | 2009.07053 | null | https://arxiv.org/abs/2009.07053v1 | https://arxiv.org/pdf/2009.07053v1.pdf | Attention Flows: Analyzing and Comparing Attention Mechanisms in Language Models | Advances in language modeling have led to the development of deep attention-based models that are performant across a wide variety of natural language processing (NLP) problems. These language models are typified by a pre-training process on large unlabeled text corpora and subsequently fine-tuned for specific tasks. Although considerable work has been devoted to understanding the attention mechanisms of pre-trained models, it is less understood how a model's attention mechanisms change when trained for a target NLP task. In this paper, we propose a visual analytics approach to understanding fine-tuning in attention-based language models. Our visualization, Attention Flows, is designed to support users in querying, tracing, and comparing attention within layers, across layers, and amongst attention heads in Transformer-based language models. To help users gain insight on how a classification decision is made, our design is centered on depicting classification-based attention at the deepest layer and how attention from prior layers flows throughout words in the input. Attention Flows supports the analysis of a single model, as well as the visual comparison between pre-trained and fine-tuned models via their similarities and differences. We use Attention Flows to study attention mechanisms in various sentence understanding tasks and highlight how attention evolves to address the nuances of solving these tasks. | ['Matthew Berger', 'Joseph F DeRose', 'Jiayao Wang'] | 2020-09-03 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [ 7.57305399e-02 1.08268380e-01 8.00229907e-02 -2.89345324e-01
-2.42965564e-01 -6.56176329e-01 6.08404219e-01 7.08609045e-01
-3.57361227e-01 6.03022315e-02 5.93990862e-01 -7.03352928e-01
3.22830714e-02 -6.04850173e-01 -2.78923303e-01 -1.80508852e-01
1.72158495e-01 3.98813754e-01 8.91543031e-02 -3.34752649e-01
6.01527750e-01 6.61008358e-01 -1.21485376e+00 7.39045739e-01
7.38644719e-01 4.69652444e-01 3.78059119e-01 9.66759145e-01
-8.68878841e-01 9.36427116e-01 -8.18510413e-01 -4.23945427e-01
-3.01000595e-01 -3.64092261e-01 -1.16136205e+00 -3.30789387e-01
4.05633807e-01 2.80789167e-01 1.90858450e-02 8.10324252e-01
2.56345421e-01 2.69106179e-01 6.08097136e-01 -1.10402918e+00
-1.25019908e+00 7.21087456e-01 -5.86496115e-01 9.17070389e-01
2.56350368e-01 3.48507106e-01 1.08269131e+00 -1.01819229e+00
4.01380539e-01 1.55764616e+00 5.64935625e-01 5.45714080e-01
-1.33376944e+00 -3.51704031e-01 5.43962896e-01 2.68997610e-01
-8.94627690e-01 -2.55361676e-01 5.97234249e-01 -9.62971807e-01
1.41820276e+00 2.06261128e-01 5.68005562e-01 8.30688834e-01
3.68716925e-01 9.02316391e-01 7.04435527e-01 -5.96320331e-01
1.65536418e-03 2.10461438e-01 7.43450105e-01 5.25378466e-01
2.11376697e-02 -3.14469218e-01 -7.27554440e-01 -1.94517840e-02
6.34239674e-01 1.01652734e-01 -2.74931163e-01 -2.80568510e-01
-9.86209154e-01 9.15323853e-01 7.74172783e-01 7.04365671e-01
-5.23109615e-01 8.29226300e-02 5.07830977e-01 1.51055962e-01
4.84574795e-01 8.96082282e-01 -5.60888827e-01 -1.56693339e-01
-9.00175989e-01 -6.74213320e-02 4.42129791e-01 7.13322461e-01
8.69085550e-01 -2.15981621e-02 -7.32688010e-01 8.14827025e-01
4.11765516e-01 -2.10540384e-01 6.06040120e-01 -4.60664630e-01
5.32777190e-01 1.05518830e+00 -1.56725138e-01 -1.02406108e+00
-4.97877836e-01 -4.61422652e-01 -6.95826709e-01 3.29588085e-01
5.01727939e-01 1.00811020e-01 -1.04663539e+00 1.62876177e+00
1.92873254e-02 -4.66098160e-01 -4.40618433e-02 6.73726797e-01
8.96717846e-01 6.98795736e-01 7.37455010e-01 2.99528748e-01
1.66848052e+00 -1.02107680e+00 -5.74694037e-01 -6.42162919e-01
8.58107328e-01 -5.71213603e-01 1.72457194e+00 3.35943252e-02
-1.25874758e+00 -8.98127377e-01 -8.19856346e-01 -7.07419336e-01
-6.84960663e-01 -2.81520754e-01 1.44294709e-01 3.93108666e-01
-1.18552220e+00 4.96247053e-01 -7.54400492e-01 -7.13359118e-01
8.54890823e-01 1.60026968e-01 -2.03556865e-02 1.82993367e-01
-1.11936247e+00 1.31591380e+00 1.86332554e-01 -7.50592276e-02
-6.34019971e-01 -1.14870238e+00 -7.85796463e-01 7.10220277e-01
4.46911305e-02 -8.52724552e-01 1.41727984e+00 -1.19565248e+00
-9.75799143e-01 9.10161078e-01 -4.57197607e-01 -3.74014437e-01
8.21179152e-02 -3.22286367e-01 -1.28174528e-01 -7.27166533e-02
-5.71896881e-03 7.87752807e-01 6.14023030e-01 -1.13393080e+00
-5.39099872e-01 -3.93125683e-01 3.88771653e-01 7.75058866e-02
-3.35444361e-01 3.99303436e-01 -3.98130715e-01 -7.95716286e-01
-2.13729769e-01 -3.60768825e-01 -1.77022740e-01 1.87772200e-01
-2.48971716e-01 -3.64856064e-01 8.96284223e-01 -6.59714282e-01
1.45035899e+00 -2.06282759e+00 1.68998688e-01 -1.40300483e-01
4.82814282e-01 3.70820850e-01 -2.60401011e-01 6.71487808e-01
-2.63597697e-01 7.11310983e-01 -2.60330796e-01 -4.73744959e-01
2.83996481e-02 1.09536871e-02 -5.51009059e-01 -5.16002998e-02
5.00195980e-01 1.15140271e+00 -9.36036587e-01 -3.48375082e-01
1.99708939e-01 6.82878017e-01 -5.69660187e-01 3.09211910e-01
-2.15519041e-01 3.63694817e-01 -2.05376580e-01 3.12483996e-01
2.75339037e-01 -5.59999943e-01 -2.93459464e-02 -2.90769994e-01
-1.95950583e-01 4.42249388e-01 -5.68506539e-01 1.52501953e+00
-6.49579942e-01 1.24096823e+00 7.90847242e-02 -7.95675576e-01
9.07711864e-01 1.90038636e-01 -3.14424545e-01 -7.88649559e-01
1.02323785e-01 -4.18779284e-01 4.25535262e-01 -5.99271297e-01
7.46249557e-01 -1.12761334e-01 1.80177644e-01 1.02606678e+00
8.10490847e-02 1.95694059e-01 2.95843899e-01 5.91988921e-01
8.15602124e-01 -1.68871865e-01 4.38314021e-01 -5.37815094e-01
5.68924010e-01 1.06936797e-01 1.72929037e-02 8.41811359e-01
-2.69167811e-01 4.22339261e-01 7.88068950e-01 -6.51195347e-01
-9.87749577e-01 -6.13054514e-01 8.52597430e-02 1.92206943e+00
-2.37297982e-01 -6.14960015e-01 -6.08707488e-01 -6.43656611e-01
-2.94045955e-02 9.70697045e-01 -1.04891360e+00 -2.19499454e-01
-5.37294984e-01 -4.34598505e-01 3.55530232e-01 8.83809805e-01
2.16755867e-01 -1.59190118e+00 -1.04536045e+00 5.63652534e-03
4.88226339e-02 -5.53656399e-01 -4.19113696e-01 1.44831076e-01
-7.74283588e-01 -1.09785342e+00 -6.78398192e-01 -7.76541710e-01
7.59705424e-01 1.58313915e-01 1.58999324e+00 3.21582109e-01
-2.91302592e-01 5.90789080e-01 -2.34994292e-01 -7.46827006e-01
-5.03615022e-01 3.01037163e-01 -5.05849779e-01 -2.61786953e-02
6.35951161e-01 -1.09684408e-01 -4.02934343e-01 2.21417341e-02
-8.65103543e-01 2.04292834e-01 3.40261757e-01 5.32045007e-01
1.58913225e-01 -6.51666105e-01 2.46173456e-01 -9.57243562e-01
1.28395188e+00 -7.23586619e-01 -3.11799198e-01 6.45242751e-01
-3.43558669e-01 4.15313780e-01 5.62801361e-01 -4.62636679e-01
-9.77731287e-01 -4.06791627e-01 7.67939240e-02 -3.22202653e-01
-1.64977551e-01 6.14139795e-01 4.09234054e-02 3.21754515e-01
7.47629106e-01 -9.79430452e-02 -2.56539106e-01 -5.19927621e-01
5.56283772e-01 4.10975426e-01 3.33541751e-01 -5.52040040e-01
4.04152960e-01 2.33966887e-01 -5.73885441e-01 -8.44707608e-01
-8.46085131e-01 -3.65638077e-01 -8.97564888e-01 -6.88914061e-02
9.46491897e-01 -3.71235132e-01 -8.47941756e-01 1.12585597e-01
-1.40072322e+00 -5.66423059e-01 -4.67318028e-01 -6.50695190e-02
-5.52229919e-02 8.21499974e-02 -4.21948403e-01 -7.24079847e-01
-5.05704522e-01 -1.09042895e+00 8.24413776e-01 3.63142967e-01
-8.57335150e-01 -1.52556038e+00 2.27617055e-01 8.89427885e-02
7.19109595e-01 -2.96944469e-01 1.41129935e+00 -8.84751081e-01
-2.57986486e-01 2.80925333e-01 -5.81444323e-01 -1.90257728e-01
-2.39984561e-02 1.90574184e-01 -1.25287175e+00 -1.91998303e-01
-1.19854681e-01 -3.08570057e-01 8.74173820e-01 3.11220855e-01
1.34103107e+00 -2.61664778e-01 -4.42294687e-01 3.98832500e-01
9.57384288e-01 1.85719803e-01 4.80048060e-01 4.07031626e-01
7.64967978e-01 9.11234558e-01 1.57447040e-01 4.03617322e-03
2.99989522e-01 3.65240276e-01 4.32255864e-01 -4.63468194e-01
7.52853900e-02 -2.77078986e-01 -5.69050089e-02 4.34524059e-01
2.78737098e-01 -4.37366545e-01 -1.40241992e+00 7.06158817e-01
-1.72418833e+00 -8.49107742e-01 -4.17515375e-02 1.90836895e+00
6.75779223e-01 1.78473741e-01 -7.71576911e-02 -9.52078216e-03
6.27933919e-01 2.67011434e-01 -6.55054808e-01 -1.16327977e+00
1.42987892e-01 2.16875985e-01 -2.78349549e-01 7.40581274e-01
-7.80913711e-01 1.04320502e+00 6.53052807e+00 3.46105635e-01
-1.39800262e+00 4.87823738e-03 7.92535365e-01 3.94834243e-02
-4.77772921e-01 2.92344522e-02 -6.14727199e-01 1.03635482e-01
8.47941995e-01 -4.46482211e-01 1.31020159e-01 5.57667673e-01
3.33961844e-01 7.84401782e-03 -1.34154308e+00 8.16642940e-01
4.73240279e-02 -1.54078948e+00 4.95387852e-01 -1.98274031e-01
3.12445432e-01 2.10740026e-02 3.59697842e-05 5.05504549e-01
4.33430135e-01 -1.31548309e+00 6.31142616e-01 5.66673875e-01
5.60463369e-01 -4.84256595e-01 3.23515624e-01 3.15702409e-01
-1.13741469e+00 -1.51835650e-01 -1.35644272e-01 -3.46571654e-01
1.84164658e-01 3.19514461e-02 -1.05909574e+00 -4.62272242e-02
8.81448507e-01 5.81597984e-01 -7.87074804e-01 7.37329841e-01
-3.86799067e-01 4.17493582e-01 2.41681322e-01 -1.38572127e-01
3.54973316e-01 2.69199908e-01 2.31415197e-01 1.69349837e+00
-1.22419707e-01 1.52040020e-01 -8.95716995e-02 1.29509580e+00
-4.11386676e-02 1.87834248e-01 -3.67033482e-01 -3.78094167e-01
3.32584232e-01 1.36220765e+00 -7.67019689e-01 -5.05784392e-01
-3.84790689e-01 6.30111277e-01 8.14430296e-01 5.80381393e-01
-5.52243054e-01 -4.08484042e-01 9.46709335e-01 3.04590225e-01
2.60537297e-01 -3.56384039e-01 -5.53128719e-01 -9.13280785e-01
-3.89153421e-01 -7.30292261e-01 5.38032353e-01 -1.26653790e+00
-1.08688378e+00 7.14246094e-01 -1.04127042e-01 -4.30359334e-01
2.07207855e-02 -7.19451964e-01 -1.11161792e+00 1.49692357e+00
-1.25940824e+00 -9.64229167e-01 -3.65689009e-01 4.13064986e-01
9.40814257e-01 6.36512488e-02 7.55203903e-01 -1.27015650e-01
-7.01808870e-01 4.12614554e-01 -3.56103480e-01 2.96213180e-01
4.30918306e-01 -1.24371588e+00 9.25702393e-01 9.00673866e-01
2.86239594e-01 1.01308239e+00 5.99628568e-01 -4.52931494e-01
-9.08962548e-01 -8.12714219e-01 1.19096601e+00 -7.04362333e-01
7.69164085e-01 -5.09450734e-01 -1.62357891e+00 7.48961985e-01
6.40407741e-01 -2.44673818e-01 8.26859236e-01 4.07959223e-01
-4.39412147e-01 2.42512107e-01 -7.34223962e-01 6.50539815e-01
6.75668180e-01 -7.75519609e-01 -9.70870137e-01 4.57021222e-02
5.56391180e-01 -1.72427535e-01 -4.48938459e-01 -5.79003617e-02
4.20616329e-01 -8.85640979e-01 8.70328069e-01 -1.29615414e+00
5.29857457e-01 -1.68195769e-01 4.05478179e-01 -1.33628523e+00
-7.53810704e-01 -2.83275276e-01 6.01578224e-03 1.28629363e+00
5.97870708e-01 -4.95721340e-01 4.66190368e-01 7.81798065e-01
-1.60196230e-01 -7.57835984e-01 -1.75220087e-01 -7.71933123e-02
3.19210082e-01 -5.43423116e-01 4.36055690e-01 1.07080972e+00
1.19796060e-01 6.79244399e-01 2.49634907e-01 3.74563560e-02
-1.19991293e-02 4.70409077e-03 7.03495562e-01 -1.30549729e+00
-7.62782246e-02 -8.27806234e-01 1.96795519e-02 -1.00784278e+00
-7.35251680e-02 -9.48324263e-01 -2.68010974e-01 -1.96719790e+00
1.96475923e-01 -2.46681705e-01 -2.94549465e-01 8.33768249e-01
-3.71030837e-01 -3.41402367e-02 6.16115272e-01 3.51086259e-01
-3.30790907e-01 2.26351708e-01 1.14809561e+00 -2.65886635e-01
-3.17167640e-01 -3.48262072e-01 -1.15651309e+00 8.73084247e-01
7.42777526e-01 -3.07959706e-01 -4.72063452e-01 -1.03422332e+00
5.46154976e-01 -5.45871675e-01 3.70269775e-01 -7.07905829e-01
5.94826996e-01 -1.46020710e-01 6.23326004e-01 -6.85341060e-01
6.68004975e-02 -4.41156775e-01 -3.01928073e-01 4.90853697e-01
-7.44720042e-01 5.52342236e-01 6.84706986e-01 1.76071420e-01
-1.58854887e-01 -1.86081111e-01 7.53887773e-01 -3.63974512e-01
-7.83486545e-01 2.15260033e-02 -5.12627184e-01 5.21090269e-01
6.51236475e-01 -3.28501016e-01 -5.54777145e-01 -3.23937595e-01
-1.11222267e+00 4.07993466e-01 3.37224960e-01 7.49603868e-01
4.30471390e-01 -8.50826979e-01 -6.38904333e-01 3.95189732e-01
1.62335396e-01 -1.53867640e-02 2.85594255e-01 5.55509508e-01
-5.33921003e-01 5.60164690e-01 -2.24267945e-01 -6.89018667e-01
-1.35864830e+00 7.14089811e-01 4.81987357e-01 -4.18903530e-01
-3.57960314e-01 1.09949052e+00 8.24403167e-01 -2.82972872e-01
2.28532881e-01 -6.91494584e-01 -6.96147263e-01 3.20021451e-01
7.91450799e-01 2.36082539e-01 -5.40082343e-02 -4.64262784e-01
-3.69287372e-01 7.11768866e-01 -3.02774876e-01 6.34494126e-02
1.02540410e+00 -1.69827595e-01 5.58613129e-02 7.91885734e-01
9.58147168e-01 -1.68745413e-01 -1.19232535e+00 -1.84400305e-01
1.99963629e-01 -1.23561561e-01 -1.87633038e-01 -1.03445995e+00
-8.41964543e-01 1.73462129e+00 2.26492688e-01 6.18467212e-01
8.63089859e-01 1.78810805e-01 2.97926903e-01 1.98120534e-01
-4.58111674e-01 -7.13191986e-01 1.88537464e-01 9.60406184e-01
1.34609962e+00 -1.08467913e+00 -6.05771616e-02 7.13852569e-02
-6.70055866e-01 1.33352160e+00 8.57976377e-01 1.08448386e-01
4.81419116e-01 2.97750592e-01 2.37813011e-01 -4.24434364e-01
-9.71548975e-01 -1.08635180e-01 4.59597677e-01 4.55886424e-01
8.70313525e-01 -3.67317051e-01 2.19171330e-01 5.01306117e-01
-1.95534840e-01 -1.54032156e-01 1.66605011e-01 9.08470869e-01
-5.27978539e-01 -8.53138149e-01 -3.82345676e-01 2.53485650e-01
-3.41346473e-01 -5.19746661e-01 -7.89155126e-01 8.72463048e-01
-2.48691402e-02 7.41913021e-01 5.75131476e-01 7.29355663e-02
3.86295825e-01 3.69544715e-01 1.55842200e-01 -1.03811979e+00
-1.27084017e+00 -4.16922092e-01 -2.70311385e-01 -3.20219010e-01
-1.83285296e-01 -3.33422601e-01 -1.24626172e+00 -1.54185802e-01
-1.32168802e-02 2.17261314e-01 4.45681542e-01 1.02113593e+00
6.76960528e-01 9.11857724e-01 -1.30762622e-01 -8.03728521e-01
-9.53132138e-02 -1.08498347e+00 7.71221519e-02 3.78943354e-01
4.77736562e-01 -3.08660746e-01 -2.86755770e-01 1.12734176e-01] | [10.117263793945312, 7.761754035949707] |
fe9cd5f8-a96b-4b4b-83da-f9bbdea1012e | focus-and-detect-a-small-object-detection | 2203.12976 | null | https://arxiv.org/abs/2203.12976v1 | https://arxiv.org/pdf/2203.12976v1.pdf | Focus-and-Detect: A Small Object Detection Framework for Aerial Images | Despite recent advances, object detection in aerial images is still a challenging task. Specific problems in aerial images makes the detection problem harder, such as small objects, densely packed objects, objects in different sizes and with different orientations. To address small object detection problem, we propose a two-stage object detection framework called "Focus-and-Detect". The first stage which consists of an object detector network supervised by a Gaussian Mixture Model, generates clusters of objects constituting the focused regions. The second stage, which is also an object detector network, predicts objects within the focal regions. Incomplete Box Suppression (IBS) method is also proposed to overcome the truncation effect of region search approach. Results indicate that the proposed two-stage framework achieves an AP score of 42.06 on VisDrone validation dataset, surpassing all other state-of-the-art small object detection methods reported in the literature, to the best of authors' knowledge. | ['Behçet Uğur Töreyin', 'İbrahim Batuhan Akkaya', 'Reyhan Kevser Keser', 'Onur Can Koyun'] | 2022-03-24 | null | null | null | null | ['object-detection-in-aerial-images', 'small-object-detection'] | ['computer-vision', 'computer-vision'] | [ 5.50578296e-01 -2.39500299e-01 1.24946401e-01 3.96204926e-02
1.83591004e-02 -6.24472797e-01 2.82327294e-01 2.01454043e-01
-3.77833217e-01 4.85639900e-01 -3.00422281e-01 -3.07058636e-02
-2.16783598e-01 -7.39428818e-01 -3.88430655e-01 -8.56419563e-01
2.20589037e-03 2.29497299e-01 9.94279027e-01 3.36391218e-02
5.37866294e-01 7.89704561e-01 -1.61966252e+00 2.98963040e-01
7.41432965e-01 9.96127963e-01 7.07178593e-01 6.45527482e-01
1.83140829e-01 5.91041207e-01 -5.27472019e-01 1.37894407e-01
6.11601532e-01 -1.77821696e-01 -4.50708628e-01 4.16538388e-01
3.41375977e-01 -5.93822241e-01 1.07209429e-01 1.17510605e+00
3.88517886e-01 2.01294228e-01 7.94800937e-01 -8.48006070e-01
-3.46028477e-01 2.98639536e-01 -1.21712565e+00 5.45381606e-01
-7.11347684e-02 -1.97541416e-02 6.15808785e-01 -9.72703457e-01
3.06170791e-01 1.02825141e+00 6.41174257e-01 3.38612586e-01
-1.18235743e+00 -8.39534044e-01 2.01709360e-01 -1.99496925e-01
-1.76605999e+00 -1.08978249e-01 5.00080466e-01 -6.40493572e-01
6.25605404e-01 4.72173125e-01 3.69546086e-01 8.46538618e-02
1.70270249e-01 6.89622998e-01 9.98454332e-01 -4.08434957e-01
1.57759279e-01 4.71404314e-01 1.47921786e-01 5.39993227e-01
5.74496686e-01 5.74900620e-02 6.15602620e-02 -1.21987782e-01
6.78630829e-01 2.98463672e-01 -2.82279402e-01 -4.20726269e-01
-9.36197758e-01 7.14845181e-01 5.62154830e-01 2.29769826e-01
-6.66467428e-01 -3.64502996e-01 8.74189213e-02 -2.76947856e-01
4.87244368e-01 2.42494777e-01 -2.71890461e-01 6.83296800e-01
-1.20699549e+00 3.76888424e-01 4.01156664e-01 9.40562248e-01
3.60419303e-01 -8.64233598e-02 -3.74794453e-01 6.28186405e-01
5.83762348e-01 5.14285684e-01 1.39317796e-01 -2.59439886e-01
4.70960349e-01 1.09706235e+00 3.49849343e-01 -9.96696115e-01
-4.79644686e-01 -7.86590397e-01 -8.16269815e-01 5.88207603e-01
2.95050561e-01 -1.82558820e-01 -9.81023252e-01 1.20450675e+00
7.85144448e-01 7.92884603e-02 8.82274061e-02 1.10323739e+00
1.01159704e+00 8.77079904e-01 1.39411092e-01 -1.78963333e-01
1.49687803e+00 -9.49456453e-01 -3.15849215e-01 -3.64470959e-01
2.65623331e-01 -1.03749216e+00 3.46535772e-01 3.59167755e-01
-8.35038900e-01 -7.09299803e-01 -1.00810742e+00 4.64483202e-01
-3.67946416e-01 9.84467030e-01 5.25870979e-01 5.48532605e-01
-5.68344176e-01 2.05268100e-01 -4.57682967e-01 -6.71960413e-01
6.88684523e-01 4.85838920e-01 -1.65110707e-01 6.47410974e-02
-2.69454151e-01 7.16791809e-01 8.95402074e-01 3.14901024e-01
-1.08739364e+00 -5.04887998e-01 -4.33967352e-01 4.34770249e-02
7.51964986e-01 -3.71263802e-01 9.34398830e-01 -1.02443027e+00
-9.71075416e-01 8.35047662e-01 4.47533578e-02 -4.40152764e-01
3.62778813e-01 -2.62995809e-01 -2.45609000e-01 7.12683648e-02
1.37267098e-01 6.04706049e-01 9.65540946e-01 -1.27995324e+00
-1.23574293e+00 -5.33324361e-01 -7.92008359e-03 4.32033151e-01
-7.45148733e-02 5.76920271e-01 -1.53335676e-01 -5.81056118e-01
3.09106320e-01 -8.78667414e-01 -2.95716852e-01 2.12385997e-01
-5.15244365e-01 -1.55376896e-01 1.23615670e+00 -5.77300072e-01
1.32571733e+00 -2.12786746e+00 3.41303237e-02 1.38286129e-01
1.43496662e-01 6.08453035e-01 1.32799909e-01 2.53348619e-01
-1.55766964e-01 -2.08893180e-01 -2.61377215e-01 -2.54282821e-02
-4.80244875e-01 -4.20055807e-01 -3.32670331e-01 6.81888878e-01
3.01287621e-01 1.42873153e-01 -6.82617724e-01 -6.74601972e-01
4.18193102e-01 4.17272262e-02 -2.04789817e-01 2.44173199e-01
9.91654694e-02 2.60638241e-02 -5.95101357e-01 1.10914958e+00
1.21957445e+00 -1.76831149e-02 -1.79069161e-01 -3.78221631e-01
-7.49168098e-01 -5.52084565e-01 -1.61477304e+00 1.08312476e+00
3.60539824e-01 4.65693206e-01 4.25311625e-01 -8.54710996e-01
9.46445286e-01 2.38635659e-01 1.81416661e-01 1.02845430e-01
3.36161256e-01 3.15961577e-02 2.50010163e-01 -5.43613434e-01
7.68580854e-01 -8.72770026e-02 3.67542565e-01 -2.74075083e-02
-4.37982902e-02 -8.65736529e-02 5.24711192e-01 3.06288395e-02
8.44666660e-01 2.08538070e-01 6.36525095e-01 -4.14327621e-01
4.84439075e-01 3.75682205e-01 4.77956504e-01 7.90090084e-01
-2.81118572e-01 7.07596362e-01 1.61710292e-01 -3.92121345e-01
-7.91894674e-01 -7.69562900e-01 -2.33428389e-01 9.74229693e-01
5.28080702e-01 -7.70028010e-02 -7.72359073e-01 -6.14754260e-01
8.02519694e-02 3.86124074e-01 -6.17790341e-01 8.90881270e-02
-1.66052669e-01 -1.16784227e+00 2.13167742e-01 4.31967407e-01
6.49110436e-01 -1.10491955e+00 -1.08854926e+00 7.14862645e-02
6.48405924e-02 -1.00210166e+00 -1.77940667e-01 3.06170881e-01
-1.15631187e+00 -1.13449156e+00 -8.94214451e-01 -7.42603600e-01
9.08553004e-01 8.78975987e-01 7.96081305e-01 1.82213753e-01
-8.33400369e-01 -2.83194035e-01 -6.31439745e-01 -8.63561988e-01
-3.66076902e-02 -1.70018792e-01 -5.54600582e-02 2.15525478e-01
3.18562835e-01 2.65845470e-02 -7.34279871e-01 5.17743051e-01
-1.02555990e+00 2.40006000e-02 1.09006464e+00 6.54814720e-01
5.22445500e-01 6.11950099e-01 1.01260275e-01 -6.25691712e-01
1.45522058e-01 -5.23139894e-01 -9.86723244e-01 2.91939259e-01
-2.06117496e-01 -4.74071592e-01 5.15987456e-01 -5.56012034e-01
-1.02807426e+00 4.85130906e-01 4.49560851e-01 -2.08401188e-01
-6.28468156e-01 1.76792711e-01 5.34905382e-02 -3.61997008e-01
7.17679918e-01 3.89833719e-01 -2.22386494e-01 -4.72880453e-01
-2.39986822e-01 8.55368972e-01 3.14553052e-01 8.65702257e-02
8.64145577e-01 6.50152564e-01 7.60171637e-02 -1.10106683e+00
-8.00679505e-01 -1.00790441e+00 -8.92712116e-01 -2.94455349e-01
8.78196776e-01 -1.01668584e+00 -3.10654193e-01 5.01987338e-01
-1.11358631e+00 -4.78336401e-03 -7.85669312e-02 5.26749790e-01
1.00846447e-01 2.68408269e-01 -1.44616470e-01 -1.32772923e+00
-4.97663349e-01 -9.24151897e-01 1.07273507e+00 8.20658743e-01
1.74228713e-01 -2.96830446e-01 -2.62767643e-01 3.80955040e-01
1.87442794e-01 3.01791191e-01 4.01044816e-01 -6.25914156e-01
-7.90038407e-01 -5.40235281e-01 -6.12513244e-01 2.90991843e-01
5.64229451e-02 1.50750935e-01 -7.93267608e-01 -2.90436506e-01
-1.93406641e-01 -1.06658479e-02 9.06220615e-01 7.44730651e-01
9.43355203e-01 1.88758507e-01 -8.81327987e-01 4.22649652e-01
1.69447386e+00 5.57399392e-01 5.37687898e-01 2.34927580e-01
3.55063468e-01 6.63785636e-01 1.11911774e+00 6.52671993e-01
-3.16259056e-01 4.57927078e-01 6.90552831e-01 -3.95776898e-01
-2.03238204e-01 1.20519243e-01 -1.64298397e-02 -2.05394566e-01
-3.84285957e-01 -4.78382081e-01 -7.26102412e-01 5.34975231e-01
-1.76175654e+00 -1.00411475e+00 -3.76601845e-01 2.08701158e+00
2.26246804e-01 2.32980028e-01 1.74560368e-01 9.03968439e-02
9.70067024e-01 1.45038869e-02 -4.65025932e-01 2.58607000e-01
-8.87134671e-02 -5.19077852e-02 9.49257672e-01 -7.12003633e-02
-1.61171925e+00 1.10635662e+00 6.22145414e+00 1.01598418e+00
-8.29416573e-01 -1.25979483e-01 3.49217296e-01 1.21307231e-01
6.60926640e-01 5.24795875e-02 -1.28110898e+00 4.58520710e-01
1.50957882e-01 5.18838704e-01 6.70937123e-03 1.03273284e+00
3.62472206e-01 -6.80842876e-01 -4.80975032e-01 7.57372200e-01
3.05939257e-01 -8.36044967e-01 1.36887386e-01 1.61292814e-02
6.34847581e-01 -2.65126765e-01 -2.71623433e-01 2.36287322e-02
-4.98274639e-02 -7.61954665e-01 8.19338024e-01 3.59771878e-01
4.78680849e-01 -7.36349225e-01 1.06116927e+00 5.90757847e-01
-1.56622303e+00 -5.55613577e-01 -8.92601192e-01 -1.66130468e-01
7.15374723e-02 4.93142992e-01 -8.73232424e-01 4.16977495e-01
7.83799887e-01 4.12281007e-01 -7.71466792e-01 1.66403329e+00
9.83450338e-02 5.69051743e-01 -3.82804781e-01 -2.29488894e-01
3.08750004e-01 -3.64325702e-01 8.22026789e-01 1.12907648e+00
4.69641298e-01 5.75755358e-01 4.18019950e-01 1.06956089e+00
2.84782231e-01 2.40219682e-01 -6.57270908e-01 -1.29089594e-01
3.71493071e-01 1.64712930e+00 -1.49060547e+00 -4.25272107e-01
-1.88484266e-01 8.51524532e-01 -1.67132184e-01 3.42928804e-02
-7.28582561e-01 -5.49477696e-01 1.41007816e-02 2.74641126e-01
6.13088250e-01 -2.90059205e-02 -8.90262648e-02 -7.40316212e-01
-7.85593390e-02 -6.40327334e-01 4.02391911e-01 -8.92462373e-01
-6.39178216e-01 4.88066018e-01 2.96324164e-01 -1.33574343e+00
4.95811135e-01 -8.08517158e-01 -7.53521979e-01 6.62217438e-01
-1.12549496e+00 -1.38444459e+00 -8.14718664e-01 2.83216476e-01
1.03731239e+00 -3.36080521e-01 6.04212582e-01 1.48173764e-01
-7.58709967e-01 -2.66514681e-02 1.81544796e-01 3.33120953e-03
4.53126609e-01 -9.79466856e-01 -2.40363732e-01 1.28755498e+00
-1.64047256e-01 2.77776241e-01 8.22824240e-01 -8.27098370e-01
-1.05162680e+00 -1.07429647e+00 3.75849932e-01 -1.48349702e-01
1.94643855e-01 -1.72351003e-01 -6.13712907e-01 3.32033247e-01
-9.21037272e-02 9.02579352e-02 4.58334208e-01 -3.58182222e-01
2.65480220e-01 -1.95105299e-01 -1.19767845e+00 1.86067328e-01
6.65180445e-01 2.65769869e-01 -4.53392804e-01 3.89683992e-01
2.21559733e-01 -3.73727858e-01 -2.92040557e-01 6.88280404e-01
6.20684028e-01 -1.09163427e+00 1.17161572e+00 -2.56241679e-01
2.28259251e-01 -8.83763731e-01 -2.46120438e-01 -9.40373063e-01
-6.55435383e-01 -1.58879027e-01 1.48025632e-01 1.22609329e+00
2.07643881e-01 -2.62306422e-01 7.24508703e-01 1.27171829e-01
-1.58991486e-01 -4.41473335e-01 -4.17036593e-01 -6.64624095e-01
-6.73142612e-01 2.45284766e-01 3.13780546e-01 4.55628008e-01
-6.54266357e-01 1.76903665e-01 -2.95601964e-01 6.74831092e-01
8.19122851e-01 3.94318670e-01 7.32883990e-01 -1.49392295e+00
-1.19964056e-01 -3.27776551e-01 -3.48254889e-01 -8.04196775e-01
-3.57881665e-01 -4.18290883e-01 3.73711646e-01 -1.47932267e+00
4.93604034e-01 -4.36835885e-01 -8.07681829e-02 3.25792670e-01
-2.54913747e-01 5.46421587e-01 1.37349322e-01 2.47018158e-01
-5.10901809e-01 2.17435196e-01 8.16952348e-01 -9.91350636e-02
-4.76006895e-01 2.35562205e-01 -5.78072131e-01 9.37232792e-01
8.62435341e-01 -5.84758699e-01 -2.80253530e-01 -2.00463548e-01
-4.01692480e-01 -2.78599381e-01 4.81547564e-01 -1.48853552e+00
2.38202378e-01 -3.32698315e-01 8.31860423e-01 -1.21178031e+00
3.03678215e-01 -9.55738842e-01 1.70148075e-01 7.14673877e-01
5.62295616e-02 -1.76854268e-01 4.63775992e-01 5.79680085e-01
-1.29226502e-02 -7.56456554e-01 1.11185634e+00 -2.68574446e-01
-9.08541083e-01 4.66718338e-02 -5.05231857e-01 -4.47821409e-01
1.68164515e+00 -5.57320416e-01 -1.32758483e-01 3.73580962e-01
-3.31220806e-01 3.06544434e-02 1.16238542e-01 1.30052090e-01
7.11367190e-01 -8.72835338e-01 -8.96960557e-01 2.85851777e-01
2.17392355e-01 2.99917042e-01 3.43483120e-01 9.19251800e-01
-8.17487776e-01 3.40103626e-01 -3.00982326e-01 -5.59436619e-01
-1.98034453e+00 5.03028214e-01 5.45939021e-02 -1.09594457e-01
-4.13918346e-01 1.05351973e+00 5.24657130e-01 1.48986587e-02
8.58029574e-02 -3.39390397e-01 -6.56041026e-01 1.28815517e-01
6.66560352e-01 5.71826935e-01 -1.56348795e-01 -7.45268643e-01
-4.69600499e-01 7.23890841e-01 -1.76062062e-02 3.47529650e-01
1.17834854e+00 -4.74236235e-02 -9.07663181e-02 6.88579530e-02
3.57429624e-01 -1.51563898e-01 -1.07147074e+00 1.30008115e-02
-2.73394585e-01 -6.95213854e-01 3.15586567e-01 -8.00565541e-01
-7.21337795e-01 6.16337299e-01 1.10405493e+00 2.21541673e-01
1.19916630e+00 -1.90956235e-01 2.66578078e-01 2.77190894e-01
1.08191758e-01 -1.16459119e+00 -1.11805238e-01 2.07942292e-01
8.64708960e-01 -1.45217931e+00 5.05025983e-01 -6.47110820e-01
-4.95868474e-01 1.16271019e+00 7.64472783e-01 -3.75051588e-01
5.00320256e-01 2.43909419e-01 -2.81452477e-01 -2.08557397e-01
-3.30649644e-01 -5.12925744e-01 4.90024865e-01 4.70105767e-01
1.81154892e-01 9.82276201e-02 -3.78191352e-01 4.68346477e-01
4.06545013e-01 8.95684287e-02 3.22606117e-01 1.17292750e+00
-1.07524753e+00 -5.02174914e-01 -1.03498578e+00 6.73072755e-01
-5.75706720e-01 -6.32596537e-02 -5.10681510e-01 9.45262909e-01
5.81817865e-01 9.52422917e-01 7.86195993e-02 -3.44150364e-01
2.91684389e-01 -4.19352144e-01 1.63607001e-01 -7.77655125e-01
-5.29752970e-01 2.52720445e-01 -2.49185711e-01 -1.97527647e-01
-5.54993391e-01 -5.78907192e-01 -9.42357242e-01 1.63255602e-01
-1.20403504e+00 5.93297891e-02 6.92145586e-01 6.92066312e-01
8.85871127e-02 4.30608839e-01 3.36189181e-01 -9.99603093e-01
-5.50150096e-01 -1.30170012e+00 -9.62515831e-01 4.96344827e-02
1.81217238e-01 -8.57238829e-01 -2.82930702e-01 1.95829600e-01] | [8.674727439880371, -0.7913186550140381] |
e457d162-bdc3-4e58-bbbb-af2af93d7ef9 | a-cnn-based-super-resolution-technique-for | 1906.10413 | null | https://arxiv.org/abs/1906.10413v1 | https://arxiv.org/pdf/1906.10413v1.pdf | A CNN-Based Super-Resolution Technique for Active Fire Detection on Sentinel-2 Data | Remote Sensing applications can benefit from a relatively fine spatial resolution multispectral (MS) images and a high revisit frequency ensured by the twin satellites Sentinel-2. Unfortunately, only four out of thirteen bands are provided at the highest resolution of 10 meters, and the others at 20 or 60 meters. For instance the Short-Wave Infrared (SWIR) bands, provided at 20 meters, are very useful to detect active fires. Aiming to a more detailed Active Fire Detection (AFD) maps, we propose a super-resolution data fusion method based on Convolutional Neural Network (CNN) to move towards the 10-m spatial resolution the SWIR bands. The proposed CNN-based solution achieves better results than alternative methods in terms of some accuracy metrics. Moreover we test the super-resolved bands from an application point of view by monitoring active fire through classic indices. Advantages and limits of our proposed approach are validated on specific geographical area (the mount Vesuvius, close to Naples) that was damaged by widespread fires during the summer of 2017. | ['Daniele Riccio', "Domenico Antonio Giuseppe Dell'Aglio", 'Massimiliano Gargiulo', 'Giuseppe Ruello', 'Antonio Iodice'] | 2019-06-25 | null | null | null | null | ['fire-detection'] | ['time-series'] | [ 4.81061071e-01 -4.77768987e-01 6.41266033e-02 -1.58799259e-04
-5.93173385e-01 -3.95340174e-01 7.81845689e-01 -7.96774626e-02
-7.65457034e-01 1.13757980e+00 -7.55493864e-02 -2.79844284e-01
-8.14116895e-01 -1.47055447e+00 -1.17502145e-01 -9.62106049e-01
-5.59916854e-01 -6.11092970e-02 4.93006743e-02 -8.50438416e-01
-2.41018385e-01 9.20649350e-01 -2.06669259e+00 3.15027535e-01
9.26035285e-01 1.01968193e+00 4.19981897e-01 4.90904123e-01
2.15369299e-01 1.55416921e-01 -3.43801945e-01 4.35165435e-01
7.64722645e-01 6.71050996e-02 -5.19298732e-01 -1.92496583e-01
4.38446850e-01 -6.31902397e-01 3.85767698e-01 1.10899973e+00
3.23490411e-01 1.22422338e-01 2.33635545e-01 -7.99420118e-01
-4.61948253e-02 6.17385447e-01 -7.64173806e-01 4.72050697e-01
-1.85338259e-01 4.65684384e-02 6.57665551e-01 -5.66442370e-01
3.41658056e-01 7.96701849e-01 8.55686128e-01 -1.40611947e-01
-1.21654832e+00 -7.02238798e-01 3.11177783e-02 2.41944015e-01
-1.76341724e+00 -2.01362282e-01 3.97898197e-01 -5.23402274e-01
8.42376292e-01 5.04419506e-01 7.20099747e-01 6.64217710e-01
2.23572012e-02 -1.46885350e-01 1.42810440e+00 -2.51610696e-01
1.66957125e-01 -3.78719807e-01 -1.70040309e-01 2.83037368e-02
6.59215152e-01 7.06961334e-01 -4.63962108e-02 -1.61803383e-02
8.15236568e-01 3.90640646e-01 -4.94371176e-01 2.15119958e-01
-1.04300463e+00 8.98192704e-01 1.03245795e+00 8.76822114e-01
-9.77491617e-01 -2.57675767e-01 1.23069182e-01 1.96777046e-01
8.45640421e-01 2.83074766e-01 -5.42636454e-01 3.68821949e-01
-1.47480667e+00 3.27604711e-01 9.68710482e-02 1.47903144e-01
9.82520819e-01 2.66289920e-01 1.48447871e-01 6.66746378e-01
3.04360360e-01 8.46441865e-01 -1.36813968e-01 -7.02352464e-01
2.80292571e-01 5.69360614e-01 4.41386402e-01 -9.86821830e-01
-7.47453749e-01 -9.16488409e-01 -1.57645142e+00 9.14660037e-01
6.16612583e-02 -2.99037546e-01 -6.62756979e-01 1.25727713e+00
7.19037652e-02 -9.09226984e-02 1.09721847e-01 1.29502881e+00
4.93213475e-01 8.53329539e-01 1.82200044e-01 -4.42217648e-01
1.24388540e+00 -1.38542011e-01 -6.17648780e-01 -2.50748307e-01
2.16066375e-01 -3.21796566e-01 4.38273102e-01 2.25354895e-01
-3.72695208e-01 -7.12613940e-01 -1.11573017e+00 7.58438349e-01
-7.98023045e-01 3.27300072e-01 4.34630841e-01 4.45449829e-01
-1.14640200e+00 7.17366397e-01 -5.68661034e-01 -5.13401270e-01
3.43123883e-01 -3.15748192e-02 -4.35280889e-01 1.95490301e-01
-1.46190548e+00 8.60399365e-01 6.94870591e-01 8.37335467e-01
-4.90423799e-01 -6.46847844e-01 -5.89600146e-01 2.86980212e-01
7.76458085e-02 -2.90194571e-01 4.26552027e-01 -9.95720446e-01
-8.42618883e-01 5.33879638e-01 3.82496566e-01 -6.92416072e-01
3.91367704e-01 -7.33809769e-02 -6.31054223e-01 2.84554809e-01
1.47728726e-01 2.45410398e-01 4.53573972e-01 -1.29206920e+00
-1.08822715e+00 -7.53849745e-01 3.02529573e-01 4.62752394e-03
-2.90435821e-01 5.43619990e-02 1.02740955e+00 -5.63399732e-01
3.12009662e-01 -4.97973263e-01 -3.19052696e-01 -1.46325603e-01
1.17453381e-01 3.73774052e-01 8.76239181e-01 -7.07959294e-01
9.96995926e-01 -2.01471257e+00 -1.50091469e-01 2.33683348e-01
-1.38722688e-01 6.65063977e-01 3.49407271e-02 4.73641157e-01
-2.65283525e-01 3.29112738e-01 -7.41435051e-01 3.32241088e-01
-4.91325587e-01 3.23669642e-01 -1.92042664e-01 4.35409576e-01
1.22346707e-01 2.60856569e-01 -6.13556623e-01 2.81952266e-02
6.04966640e-01 7.80311942e-01 2.86597580e-01 -9.22219679e-02
5.63751012e-02 4.61404771e-01 -2.62716055e-01 4.03668284e-01
1.53557408e+00 3.10329825e-01 -1.46533728e-01 -2.63556421e-01
-1.07770967e+00 -2.22995728e-01 -1.35291898e+00 1.41439271e+00
-3.77073377e-01 3.00773740e-01 6.44669533e-01 -8.92985761e-01
1.20965219e+00 5.33659518e-01 4.87536997e-01 -7.82606304e-01
-2.46717617e-01 2.07922757e-01 -2.08492726e-01 -4.74592328e-01
3.16880643e-01 -4.02682066e-01 5.81620455e-01 2.33527452e-01
-5.07482231e-01 1.23217747e-01 1.02781467e-01 -4.02419895e-01
4.92250562e-01 3.69853564e-02 4.70009744e-01 -5.58786690e-01
6.31786942e-01 2.25876123e-01 4.38039094e-01 6.09613121e-01
-1.90314159e-01 4.13780928e-01 -1.07730441e-01 -9.31003273e-01
-9.51133072e-01 -6.62251592e-01 -4.18988615e-01 7.65435934e-01
-2.09030405e-01 2.35464483e-01 -2.47016922e-01 -9.81005505e-02
-4.00497243e-02 3.40838909e-01 -7.03698218e-01 3.89017135e-01
-3.34161781e-02 -1.33692491e+00 6.16599321e-01 3.01521987e-01
1.34487772e+00 -9.55203891e-01 -1.12498593e+00 1.81137115e-01
-3.07089746e-01 -7.99699128e-01 7.77692437e-01 1.15219370e-01
-1.05060673e+00 -1.01765788e+00 -8.07331979e-01 -3.39463353e-02
-6.93303049e-02 6.88299119e-01 8.55198145e-01 -8.23008195e-02
-2.89632112e-01 -1.69966131e-01 -7.15184093e-01 -2.26538703e-01
-1.22575775e-01 1.32213175e-01 -2.95667410e-01 2.10953921e-01
1.68176472e-01 -1.00080335e+00 -5.34244418e-01 6.09742571e-03
-9.89620090e-01 7.54050687e-02 6.79815233e-01 4.12910610e-01
1.81384966e-01 4.38250452e-01 3.81707251e-01 -2.96594054e-01
1.50150225e-01 -4.74539906e-01 -8.54950488e-01 6.53560832e-02
-4.75058019e-01 -3.80085379e-01 3.14521641e-01 3.11912864e-01
-1.17174256e+00 1.91167220e-01 -1.32098168e-01 1.68163002e-01
-9.30410326e-01 8.60583663e-01 1.21696763e-01 -4.85182814e-02
1.19230735e+00 2.12864608e-01 -1.79758385e-01 -7.46336699e-01
-2.89363172e-02 9.37860370e-01 4.30025399e-01 -4.10957355e-03
1.13038850e+00 9.80087936e-01 1.51002198e-01 -1.29062045e+00
-6.50245428e-01 -5.60566485e-01 -8.56372356e-01 -4.86438483e-01
7.87631392e-01 -1.15762520e+00 -5.75066626e-01 5.35794079e-01
-1.02009308e+00 -1.85985044e-02 1.63436178e-02 6.69539511e-01
-1.40126899e-01 1.66188166e-01 -1.72515258e-01 -1.36683965e+00
-6.45810902e-01 -5.82234681e-01 9.24808919e-01 1.26623422e-01
3.35573345e-01 -5.08498669e-01 3.32238406e-01 9.81459841e-02
1.01505804e+00 9.09788251e-01 4.31415379e-01 2.54206121e-01
-4.50687021e-01 -1.16955161e-01 -6.56630576e-01 4.15427566e-01
3.67821395e-01 1.59571722e-01 -1.21417129e+00 -2.85659134e-01
-1.42317653e-01 3.55636537e-01 1.23720324e+00 6.72565401e-01
5.96078634e-01 -8.55498835e-02 -3.29193883e-02 7.86591887e-01
2.18859768e+00 1.11604415e-01 8.13929141e-01 7.75664508e-01
1.34971321e-01 5.33353090e-01 7.33022153e-01 8.19006145e-01
-1.16594881e-01 5.41408479e-01 1.27838528e+00 -4.92215395e-01
8.65092129e-02 6.77492917e-01 -4.63711424e-03 -2.94733852e-01
-1.17126322e+00 2.92356312e-01 -9.54229176e-01 6.82746530e-01
-1.83708346e+00 -1.60043395e+00 -6.26882792e-01 2.33599067e+00
1.82551920e-01 -2.18989015e-01 1.14253744e-01 6.87144756e-01
8.48378062e-01 6.07189476e-01 3.00390236e-02 1.84758797e-01
-6.25660717e-01 5.06165981e-01 1.04028344e+00 5.91264606e-01
-1.53648305e+00 4.25863862e-01 5.46871376e+00 2.96765208e-01
-1.25854564e+00 2.55984634e-01 -2.38316171e-02 9.16826501e-02
-6.08155653e-02 -6.03418984e-02 -3.81518424e-01 8.29682872e-02
8.34345579e-01 3.27832639e-01 4.68870461e-01 5.43578446e-01
8.51024210e-01 -4.46866274e-01 1.41252384e-01 7.41345227e-01
-4.47409362e-01 -1.37177229e+00 -1.05522700e-01 2.60478675e-01
5.81149340e-01 5.35549283e-01 -2.72903383e-01 -1.67193681e-01
-2.55590212e-02 -8.19455981e-01 4.13069457e-01 7.73191631e-01
7.89057255e-01 -1.01887298e+00 1.04651177e+00 5.41883707e-01
-1.51468968e+00 -4.28286225e-01 -5.66276193e-01 -5.81918240e-01
-1.11086711e-01 1.01735032e+00 -4.12078470e-01 1.17784178e+00
1.09803832e+00 7.83092797e-01 -2.74382621e-01 1.09271324e+00
-1.26997232e-01 4.76290017e-01 -4.57087696e-01 4.87369567e-01
5.30114174e-01 -2.60061264e-01 6.22124672e-01 1.06824708e+00
7.48635411e-01 3.29803199e-01 -2.62242090e-02 9.96040285e-01
7.19942391e-01 -5.50109372e-02 -8.67821991e-01 3.10686499e-01
1.56837687e-01 1.49143577e+00 -3.78810376e-01 -3.02965920e-02
-1.17489785e-01 4.85191792e-01 -5.09524047e-01 4.71224755e-01
-4.03378010e-01 -2.96810836e-01 9.19928849e-01 3.34695369e-01
1.43654048e-01 -4.63306129e-01 -1.30824119e-01 -9.43392158e-01
-2.74690222e-02 -5.08927703e-01 3.07973415e-01 -9.69374895e-01
-8.72622728e-01 7.72480726e-01 1.60558209e-01 -1.60565197e+00
-1.53581485e-01 -5.60245574e-01 -3.87656450e-01 1.30117416e+00
-2.07070231e+00 -1.37313783e+00 -1.04493380e+00 5.31310081e-01
1.27757415e-01 -1.93305127e-02 1.30978084e+00 3.29888999e-01
-2.26007223e-01 -4.90573883e-01 3.21894556e-01 -4.03220505e-02
1.89965323e-01 -8.90295982e-01 7.49774873e-02 1.08062124e+00
-4.24312055e-01 -8.90058931e-03 6.04860306e-01 -5.41564167e-01
-6.57650471e-01 -1.23692453e+00 8.96713257e-01 5.81493437e-01
3.34800929e-01 2.63977766e-01 -7.65550494e-01 4.38349038e-01
2.31059175e-03 5.96269257e-02 4.16729778e-01 7.58062303e-02
-2.85091013e-01 -6.38239622e-01 -1.51158941e+00 8.07652175e-02
7.32720017e-01 -4.35313225e-01 -2.23420456e-01 3.21949214e-01
1.29246607e-01 3.29639584e-01 -9.72487152e-01 8.90963137e-01
7.69643009e-01 -1.56399536e+00 9.91056025e-01 -1.32577658e-01
3.54733109e-01 -7.13380933e-01 -5.32378078e-01 -1.39127862e+00
-8.18229735e-01 1.23307310e-01 7.06703126e-01 7.46209264e-01
1.34116158e-01 -7.72975385e-01 2.50798434e-01 -3.38765025e-01
1.18596449e-01 1.17275946e-01 -1.15951097e+00 -7.61780858e-01
-1.90853670e-01 -4.84125704e-01 8.96477818e-01 9.57207501e-01
-3.80571842e-01 -2.46174172e-01 -2.99024731e-01 8.99930537e-01
6.63072407e-01 1.93629742e-01 3.63367230e-01 -1.99772310e+00
2.52656400e-01 -4.18096960e-01 -5.11200368e-01 -9.17173573e-04
-3.09933513e-01 -4.33129162e-01 -2.71365494e-01 -1.60639703e+00
-8.60053822e-02 -3.34950596e-01 -4.88262147e-01 7.08972812e-01
2.35320285e-01 6.41032994e-01 1.04409263e-01 1.61258921e-01
4.24126208e-01 4.62009162e-01 6.82302713e-01 -3.33081484e-01
1.68526918e-02 1.72009826e-01 -2.37052694e-01 5.47340810e-01
8.60300064e-01 -2.99374759e-01 7.94927850e-02 -5.67051351e-01
4.78020668e-01 1.58122107e-01 8.13646197e-01 -1.57620621e+00
-1.34563044e-01 -3.68019551e-01 3.91758233e-01 -8.58926117e-01
3.15048903e-01 -1.24722373e+00 6.58785999e-01 5.81366122e-01
1.34937122e-01 -2.89413244e-01 4.02474642e-01 -6.11674264e-02
-3.54777783e-01 -8.51875544e-02 1.03264713e+00 -3.33142608e-01
-8.61268103e-01 3.27214599e-01 -6.41289115e-01 -9.42353725e-01
7.81991601e-01 -2.17176050e-01 -5.63147068e-01 -1.76471278e-01
-8.31177831e-01 -1.36743978e-01 2.84943163e-01 1.85849398e-01
4.01871562e-01 -1.12452292e+00 -1.21227741e+00 4.67811376e-01
2.35078841e-01 -1.96610153e-01 5.97501099e-01 8.64558697e-01
-7.28749275e-01 3.73509914e-01 -8.63834798e-01 -5.81878304e-01
-1.23060071e+00 1.99849233e-01 7.76771426e-01 -7.81168267e-02
-5.14046609e-01 3.40860426e-01 -4.04674679e-01 -4.05274421e-01
-4.11933124e-01 -3.43056589e-01 -6.20693088e-01 4.95085359e-01
8.74770820e-01 4.32683706e-01 3.90553564e-01 -8.70698214e-01
-5.27630091e-01 6.54452384e-01 8.09419632e-01 -3.07294875e-01
1.77388871e+00 -1.67535603e-01 -3.38543445e-01 2.91521311e-01
5.87530255e-01 -3.52916092e-01 -1.05879927e+00 -9.89725143e-02
-2.13948622e-01 -6.38345122e-01 6.45846248e-01 -9.61908221e-01
-1.17946279e+00 9.13815737e-01 1.19139361e+00 6.79979861e-01
1.55957675e+00 -6.05039060e-01 1.75484613e-01 4.10143256e-01
4.02754098e-01 -8.21443379e-01 -1.13614595e+00 4.29291636e-01
1.03040040e+00 -1.17493546e+00 6.91392571e-02 -2.44695097e-01
2.43490394e-02 1.48301780e+00 -3.04475203e-02 2.50839032e-02
7.81359076e-01 2.18395337e-01 2.36491889e-01 -1.57806218e-01
-2.17161626e-01 -9.95558918e-01 -1.83951721e-01 8.17075968e-01
3.08872998e-01 5.67228556e-01 -3.40917468e-01 -4.84635942e-02
6.35322407e-02 3.20509374e-01 3.63230944e-01 6.69568896e-01
-9.73612905e-01 -4.78158236e-01 -1.02428317e+00 3.37809086e-01
-7.49015063e-02 -1.38501808e-01 -6.79961294e-02 8.82684708e-01
5.88355780e-01 1.19188905e+00 1.90209404e-01 -3.70903283e-01
2.67174304e-01 -6.84339851e-02 1.53823689e-01 -2.37119913e-01
-5.73728204e-01 1.83054078e-02 -7.83876702e-02 -4.89706397e-01
-1.11212623e+00 -5.82571030e-01 -6.78456903e-01 -5.80469072e-01
-2.12423488e-01 2.57518709e-01 9.55304265e-01 9.75420415e-01
2.49029025e-02 3.45912933e-01 8.06547642e-01 -1.19101369e+00
-2.34053507e-01 -1.55054355e+00 -1.27308166e+00 -1.65722340e-01
7.11961448e-01 -6.41267180e-01 -5.22047043e-01 -3.54947984e-01] | [9.752008438110352, -1.7273523807525635] |
179ac34b-d4c1-4ca9-8b59-c51969a467b2 | fugashi-a-tool-for-tokenizing-japanese-in | 2010.06858 | null | https://arxiv.org/abs/2010.06858v1 | https://arxiv.org/pdf/2010.06858v1.pdf | fugashi, a Tool for Tokenizing Japanese in Python | Recent years have seen an increase in the number of large-scale multilingual NLP projects. However, even in such projects, languages with special processing requirements are often excluded. One such language is Japanese. Japanese is written without spaces, tokenization is non-trivial, and while high quality open source tokenizers exist they can be hard to use and lack English documentation. This paper introduces fugashi, a MeCab wrapper for Python, and gives an introduction to tokenizing Japanese. | ['Paul McCann'] | 2020-10-14 | null | https://aclanthology.org/2020.nlposs-1.7 | https://aclanthology.org/2020.nlposs-1.7.pdf | emnlp-nlposs-2020-11 | ['multilingual-nlp'] | ['natural-language-processing'] | [-6.34735107e-01 -1.47923797e-01 -2.22923070e-01 -2.46739507e-01
-9.79689777e-01 -9.92017686e-01 2.33780533e-01 1.76442079e-02
-8.00012648e-01 1.29865158e+00 4.20638204e-01 -5.54116726e-01
3.30449015e-01 -4.36074078e-01 -2.21981823e-01 -3.93538743e-01
3.01569104e-01 4.08978492e-01 1.73470914e-01 -2.00045660e-01
3.18215191e-02 1.36367515e-01 -8.02023172e-01 8.40045735e-02
1.08898878e+00 -4.21526879e-02 4.79558259e-01 5.26346266e-01
-7.17449903e-01 5.99061489e-01 -7.44602799e-01 -4.22698528e-01
5.51021658e-02 -7.43146427e-03 -1.30490863e+00 -3.22281837e-01
2.35411599e-01 6.13930412e-02 1.66547626e-01 1.18977368e+00
4.21861231e-01 -1.39056072e-01 1.00854255e-01 -1.02658772e+00
-8.01107109e-01 1.11158061e+00 -1.65342063e-01 1.57271609e-01
6.33201957e-01 8.36607441e-02 1.04381216e+00 -7.02346206e-01
1.08324504e+00 1.03505218e+00 7.56105423e-01 4.25176382e-01
-1.10487378e+00 -5.29480696e-01 -5.70769198e-02 -2.46709406e-01
-1.48563206e+00 -1.76544949e-01 5.47059216e-02 -3.63335758e-01
1.67952883e+00 1.79145083e-01 5.35317302e-01 7.43763745e-01
1.12508163e-01 6.99570775e-01 1.07001173e+00 -7.66844690e-01
-3.17877144e-01 3.41019392e-01 2.84323901e-01 4.30041492e-01
2.71593153e-01 -5.13048410e-01 -2.32536659e-01 -2.11385876e-01
4.60896909e-01 -4.22175705e-01 -6.43234402e-02 2.24727839e-01
-1.66499150e+00 6.01011753e-01 -7.62895495e-02 1.06001127e+00
-2.57873349e-03 -1.67070446e-03 8.99965644e-01 4.81667608e-01
2.84836113e-01 6.07269347e-01 -8.21864069e-01 -7.44182229e-01
-9.50865030e-01 3.87012243e-01 1.21929705e+00 1.09695065e+00
9.92554605e-01 1.45964205e-01 3.87242496e-01 9.45603251e-01
1.20421350e-01 4.82127845e-01 6.81469440e-01 -9.90536511e-01
6.51568770e-01 3.81736606e-01 2.02336133e-01 -5.42068839e-01
-2.67506987e-01 1.55027434e-01 -3.31037901e-02 -1.40023351e-01
7.62765348e-01 -5.40859699e-01 -5.81345499e-01 1.36330473e+00
1.48508027e-01 -7.69723952e-01 3.37990910e-01 5.99866271e-01
6.45014048e-01 8.48971426e-01 4.81463850e-01 -5.16241090e-03
1.43779314e+00 -9.11657274e-01 -8.39782178e-01 -2.25170955e-01
1.11809111e+00 -1.36814046e+00 1.22328949e+00 4.08944637e-01
-1.07879877e+00 7.81141445e-02 -8.03763270e-01 -6.31114781e-01
-9.01255429e-01 5.54000996e-02 9.10257936e-01 9.04641747e-01
-1.00289249e+00 2.75011569e-01 -8.50233018e-01 -8.19810927e-01
-2.97626704e-01 2.32360154e-01 -6.27711177e-01 -6.03462160e-02
-1.42533112e+00 1.32612729e+00 7.42429733e-01 9.35566723e-02
-2.57522106e-01 -5.82002580e-01 -9.61605906e-01 -1.97211012e-01
3.54875684e-01 -3.06046695e-01 1.44447637e+00 -9.73757684e-01
-1.45430028e+00 1.05497956e+00 -2.15837225e-01 -1.63567718e-02
3.78831446e-01 -5.09351134e-01 -6.63082123e-01 -2.57563978e-01
6.16258144e-01 6.37089491e-01 1.43055797e-01 -6.77246451e-01
-9.54314113e-01 2.35707164e-02 8.08174536e-02 2.95243263e-01
-3.80824655e-01 1.12128568e+00 -4.53135759e-01 -5.85483968e-01
-4.28258896e-01 -7.36199856e-01 -3.40689272e-01 -4.95746583e-01
-3.49984318e-01 -4.02106643e-01 6.14768684e-01 -9.82612133e-01
1.41765606e+00 -2.21565437e+00 -2.87162066e-01 -6.45849779e-02
-5.64964861e-02 1.92682743e-01 1.17914371e-01 7.95835018e-01
1.60457432e-01 6.60297692e-01 -8.16505849e-02 4.16233554e-04
2.95543432e-01 6.77862108e-01 -1.24958664e-01 1.88343346e-01
-4.09615785e-02 6.73772216e-01 -1.05723023e+00 -8.20939839e-01
1.37643993e-01 2.81014979e-01 -3.07534575e-01 -3.29088897e-01
-9.96713638e-02 2.77025178e-02 -2.76944190e-01 9.68976438e-01
4.26311940e-01 1.86547160e-01 5.95921338e-01 4.17542726e-01
-1.13832700e+00 7.08493650e-01 -1.24566221e+00 1.60949469e+00
-6.06000125e-01 6.25674844e-01 4.07742590e-01 -5.29920578e-01
3.95615697e-01 7.66619503e-01 2.68848777e-01 -5.56376688e-02
-9.50518772e-02 8.17793190e-01 5.11245616e-02 -5.28162718e-01
1.22741485e+00 -1.95454925e-01 -5.70583701e-01 4.99501884e-01
1.67351365e-01 -4.95034099e-01 9.07394230e-01 2.84242123e-01
6.64033175e-01 4.39538270e-01 7.57582009e-01 -5.22517562e-01
2.09249035e-01 5.50934374e-01 8.93873155e-01 3.00477445e-01
-1.53587684e-01 3.72334063e-01 5.21408975e-01 -3.35316390e-01
-1.11981058e+00 -6.76902771e-01 -3.83663476e-01 1.17351437e+00
-5.18947363e-01 -6.02279842e-01 -8.21523309e-01 -4.90077585e-01
-1.77946031e-01 4.90776598e-01 3.91688105e-03 8.24887097e-01
-9.14727807e-01 -5.85516036e-01 1.11694491e+00 3.72645646e-01
5.79615273e-02 -1.32275498e+00 -3.51700813e-01 4.57946420e-01
-4.81196672e-01 -1.33395481e+00 -5.36449790e-01 2.92904228e-01
-5.42592704e-01 -6.91380978e-01 -6.66902065e-01 -1.42253757e+00
5.04808784e-01 4.62315790e-03 1.12873983e+00 -1.03382342e-01
-1.86593458e-01 1.12052843e-01 -3.36091906e-01 -5.48031330e-01
-3.45544279e-01 4.77040023e-01 2.23667454e-02 -8.58480096e-01
8.82441342e-01 -1.46188110e-01 2.68062562e-01 -9.17762294e-02
-7.70813286e-01 -3.76548201e-01 3.34322751e-01 6.28201663e-01
3.34928274e-01 -1.50965691e-01 3.76548082e-01 -1.05629373e+00
7.17133105e-01 -3.77487630e-01 -5.42345285e-01 2.64990866e-01
-3.59396935e-01 -3.47042121e-02 6.37810290e-01 -1.28646076e-01
-1.08394575e+00 -2.12766118e-02 -3.95692527e-01 4.68762755e-01
-3.91179889e-01 1.03339779e+00 -3.31817240e-01 -8.14702287e-02
3.37394863e-01 -1.12614155e-01 -3.90761942e-01 -6.74329221e-01
2.67056167e-01 7.72055447e-01 3.38372350e-01 -8.98473024e-01
3.76947254e-01 3.42048556e-02 -9.58409548e-01 -1.28573668e+00
-2.84439206e-01 -6.23064041e-01 -5.73818386e-01 8.35416615e-02
7.95242250e-01 -9.93415654e-01 -4.13566053e-01 4.02187049e-01
-1.21930218e+00 -5.21264672e-01 -1.74322009e-01 6.47836149e-01
-3.99955213e-02 7.25753069e-01 -1.16725123e+00 -3.14127892e-01
-2.70678818e-01 -1.13476288e+00 4.92604673e-01 2.05256075e-01
-7.32396066e-01 -1.28514850e+00 3.74438196e-01 1.60432860e-01
4.51477975e-01 -1.74967766e-01 9.68130052e-01 -6.28073454e-01
-2.66866654e-01 -2.31766015e-01 -1.23007640e-01 3.03644121e-01
1.52175993e-01 6.46652818e-01 -6.21274054e-01 1.55284345e-01
-4.38179851e-01 -4.49470907e-01 8.96974578e-02 1.82801247e-01
4.08027917e-01 -3.97935271e-01 -3.72510999e-01 1.94655269e-01
1.49833345e+00 -5.16764745e-02 4.39788043e-01 8.30228806e-01
6.39701486e-01 6.32489920e-01 5.19565403e-01 1.65327340e-01
7.74264514e-01 3.61414105e-01 -3.36000681e-01 -1.74443945e-02
1.77463278e-01 -7.21177310e-02 6.91835046e-01 1.63850939e+00
7.83401716e-04 4.41015922e-02 -1.27298725e+00 9.67240989e-01
-1.69424617e+00 -6.73301995e-01 -5.18243432e-01 1.78358877e+00
1.36490285e+00 -6.89573139e-02 9.60089862e-02 -2.44728819e-01
5.57758689e-01 -1.56140819e-01 1.51273608e-01 -8.72362375e-01
-4.64167416e-01 1.87598631e-01 5.96652389e-01 7.44583607e-01
-9.00809407e-01 1.49768341e+00 7.79241800e+00 6.30497158e-01
-1.23809278e+00 3.33169520e-01 -1.59185946e-01 1.78495839e-01
-3.70597482e-01 4.16457146e-01 -1.14016712e+00 3.71481866e-01
9.14793193e-01 -5.15532553e-01 2.54502714e-01 9.20170903e-01
1.53848901e-01 -2.75613427e-01 -7.33687878e-01 5.50934732e-01
-1.79425314e-01 -1.05324900e+00 -2.15387568e-01 -1.29314186e-03
5.84830523e-01 6.18588090e-01 -4.38875973e-01 5.28450072e-01
8.99804890e-01 -8.70779574e-01 7.60332823e-01 -6.95395768e-02
5.79541206e-01 -7.30025172e-01 8.82133007e-01 1.63691476e-01
-1.14027262e+00 3.37692857e-01 -5.66275775e-01 -2.33174264e-01
5.86975098e-01 7.07301497e-01 -4.51020956e-01 6.26735449e-01
8.86365294e-01 3.98686975e-01 -2.45226845e-01 1.08290458e+00
-4.32896793e-01 6.56867743e-01 -5.09949625e-01 -2.84184098e-01
5.79175651e-01 -3.12109828e-01 5.65633416e-01 1.94360471e+00
2.23975480e-01 -2.22982764e-01 6.04253590e-01 1.85531020e-01
1.53918877e-01 6.30058050e-01 -5.87054193e-01 -3.98436278e-01
4.72526550e-01 1.50899875e+00 -7.77532399e-01 -3.95688891e-01
-8.91605914e-01 7.48268843e-01 3.71508509e-01 3.70451808e-01
-6.57988310e-01 -9.62548375e-01 7.97674119e-01 -1.63724139e-01
3.03088874e-01 -6.95541084e-01 -1.37008175e-01 -1.29520142e+00
-8.40827078e-03 -1.20008922e+00 5.03728628e-01 -6.04577303e-01
-1.17889559e+00 4.81177986e-01 -3.78187262e-02 -8.75791848e-01
-3.26219112e-01 -9.12369788e-01 -3.36556077e-01 1.03487229e+00
-1.33142054e+00 -1.19933259e+00 4.59180564e-01 4.19707716e-01
2.50343323e-01 2.19551280e-01 1.11760008e+00 7.29702175e-01
-5.14070749e-01 4.37090009e-01 2.83059210e-01 4.32705075e-01
1.04857564e+00 -1.41435945e+00 2.76716977e-01 8.68767560e-01
7.70329013e-02 1.12454283e+00 7.84391165e-01 -6.72139585e-01
-1.23118496e+00 -6.23240232e-01 1.90472686e+00 -4.55003053e-01
1.39003325e+00 -4.82718766e-01 -8.27143848e-01 1.02191520e+00
8.28724802e-01 -2.22445577e-01 8.95475686e-01 4.25063252e-01
-2.55260736e-01 1.09354571e-01 -9.31731582e-01 7.26873398e-01
4.35388476e-01 -7.00713336e-01 -7.45986640e-01 4.12450552e-01
5.66972911e-01 -4.35998529e-01 -1.14857709e+00 -3.71268123e-01
5.86269736e-01 -3.49536210e-01 4.06813055e-01 -3.68515015e-01
-8.58803838e-03 -4.96621788e-01 1.35479406e-01 -1.19094992e+00
-9.66207236e-02 -9.93670404e-01 7.26155996e-01 1.56413162e+00
8.34016979e-01 -8.40245605e-01 6.07110322e-01 8.17918062e-01
-3.56260419e-01 1.65184997e-02 -8.95172715e-01 -1.04862201e+00
5.18948317e-01 -4.97549564e-01 2.44519770e-01 1.54365420e+00
9.89565492e-01 2.22311139e-01 -2.73193657e-01 -2.72533268e-01
1.21221408e-01 -1.33900836e-01 5.01646578e-01 -9.07022953e-01
-1.33675620e-01 -4.47293639e-01 -2.20177233e-01 -7.64648736e-01
7.11760446e-02 -9.57170188e-01 2.25043148e-01 -1.66021740e+00
1.24729229e-02 -5.64712942e-01 1.21998198e-01 1.01707137e+00
1.22709875e-03 2.40706325e-01 2.38662377e-01 2.43606232e-02
-4.53080595e-01 -1.06025599e-01 7.45961428e-01 3.31794322e-02
-3.07556391e-01 -5.36034346e-01 -5.77458560e-01 7.86663055e-01
9.62260187e-01 -7.85238862e-01 3.04089636e-01 -8.10297549e-01
6.65221095e-01 -3.81928533e-01 -2.48760149e-01 -7.29369819e-01
4.35886607e-02 -3.91216010e-01 -1.06225006e-01 -4.10823435e-01
-1.84099957e-01 -5.28280735e-01 1.05045833e-01 3.62332225e-01
1.12801142e-01 5.33687532e-01 4.62790072e-01 -3.98204654e-01
-5.51587224e-01 -6.38681650e-01 3.84981394e-01 -6.19362533e-01
-9.70534682e-01 -4.60157283e-02 -1.11535764e+00 3.80052179e-01
8.81729603e-01 -3.70393485e-01 -2.61391997e-01 1.14981115e-01
-4.35067058e-01 3.72580171e-01 7.10002065e-01 1.94644019e-01
-3.82000357e-02 -8.94721925e-01 -4.93093044e-01 -2.73994088e-01
6.33218884e-02 -2.20212668e-01 -1.02407791e-01 7.82452166e-01
-1.10477674e+00 7.50779748e-01 -2.47728571e-01 3.41085829e-02
-1.08529508e+00 4.44023579e-01 -5.75068295e-02 -4.07165885e-01
-6.84722424e-01 5.63071847e-01 -2.36129075e-01 -9.37580168e-01
-7.82710221e-03 -5.85370839e-01 -2.22848490e-01 1.02416210e-01
4.96386945e-01 2.95075595e-01 1.60213545e-01 -6.25577331e-01
-5.53455114e-01 2.30434611e-01 -2.18657792e-01 -5.40069044e-01
1.12427664e+00 -1.08774163e-01 -7.30298996e-01 8.19882214e-01
8.86386275e-01 7.97909319e-01 -4.00228500e-01 5.63885644e-02
4.38355625e-01 -6.40971884e-02 -4.57518220e-01 -6.90649033e-01
-4.78748173e-01 6.99352622e-01 -1.23135939e-01 3.20398360e-01
4.51229006e-01 -2.30475500e-01 1.04582393e+00 7.15435088e-01
4.36843514e-01 -1.76149023e+00 -6.69032156e-01 1.34980845e+00
4.61502492e-01 -8.23965609e-01 -9.32965279e-02 -3.80625725e-01
-7.34803319e-01 1.16678202e+00 5.29744148e-01 1.05795048e-01
6.48802936e-01 6.10243857e-01 7.10070372e-01 -1.50137851e-02
-3.98172021e-01 -9.43743438e-02 -5.64296484e-01 8.19523573e-01
1.13857722e+00 3.26394200e-01 -9.46704566e-01 4.98914331e-01
-5.47485590e-01 3.11494116e-02 7.39595473e-01 1.26271677e+00
-2.93148547e-01 -1.68094480e+00 -5.38519263e-01 8.11405629e-02
-1.13841856e+00 -6.99965894e-01 -2.55906254e-01 1.13676822e+00
-9.01788753e-03 7.26519167e-01 -5.65449633e-02 3.18673372e-01
-9.02706534e-02 3.51802975e-01 3.88173521e-01 -1.04013538e+00
-1.01697624e+00 9.59340110e-02 7.06544578e-01 -1.89341009e-01
-4.46291506e-01 -7.64336824e-01 -1.43595362e+00 -6.10113859e-01
-2.50619918e-01 6.71071172e-01 6.94899917e-01 9.39346910e-01
-1.68415543e-03 -6.53882399e-02 -2.22084269e-01 -4.60141420e-01
-1.40660301e-01 -9.15378511e-01 -7.34002888e-01 -3.25027555e-01
-1.20990559e-01 -1.23738669e-01 -1.74698785e-01 2.12270319e-01] | [10.341728210449219, 10.019989013671875] |
93312071-efa4-4059-ba83-77ab28ec59d7 | attresdu-net-medical-image-segmentation-using | 2306.14255 | null | https://arxiv.org/abs/2306.14255v1 | https://arxiv.org/pdf/2306.14255v1.pdf | AttResDU-Net: Medical Image Segmentation Using Attention-based Residual Double U-Net | Manually inspecting polyps from a colonoscopy for colorectal cancer or performing a biopsy on skin lesions for skin cancer are time-consuming, laborious, and complex procedures. Automatic medical image segmentation aims to expedite this diagnosis process. However, numerous challenges exist due to significant variations in the appearance and sizes of objects with no distinct boundaries. This paper proposes an attention-based residual Double U-Net architecture (AttResDU-Net) that improves on the existing medical image segmentation networks. Inspired by the Double U-Net, this architecture incorporates attention gates on the skip connections and residual connections in the convolutional blocks. The attention gates allow the model to retain more relevant spatial information by suppressing irrelevant feature representation from the down-sampling path for which the model learns to focus on target regions of varying shapes and sizes. Moreover, the residual connections help to train deeper models by ensuring better gradient flow. We conducted experiments on three datasets: CVC Clinic-DB, ISIC 2018, and the 2018 Data Science Bowl datasets and achieved Dice Coefficient scores of 94.35%, 91.68% and 92.45% respectively. Our results suggest that AttResDU-Net can be facilitated as a reliable method for automatic medical image segmentation in practice. | ['Md. Hasanul Kabir', 'Md. Bakhtiar Hasan', 'Fahim Shahriar Khan', 'Alif Ashrafee', 'Akib Mohammed Khan'] | 2023-06-25 | null | null | null | null | ['medical-image-segmentation'] | ['medical'] | [ 2.34228417e-01 3.23772877e-01 -1.67923272e-01 -2.70343244e-01
-5.30634403e-01 -5.20159602e-01 1.15642855e-02 5.34180403e-01
-5.65008938e-01 4.14619714e-01 1.11856669e-01 -6.25870407e-01
-7.33782426e-02 -7.05328405e-01 -5.43440461e-01 -6.97179615e-01
-1.49044633e-01 5.66423163e-02 4.18671280e-01 -8.95667239e-04
3.16902369e-01 4.25478101e-01 -7.84796715e-01 3.98967087e-01
1.38386869e+00 8.09192479e-01 3.50486100e-01 8.23926926e-01
-1.40373215e-01 6.32557690e-01 -2.10364714e-01 -3.74210447e-01
2.76368976e-01 -4.84596103e-01 -8.97350907e-01 -1.21482514e-01
4.35951382e-01 -3.34089339e-01 -3.88852209e-01 1.28274333e+00
3.88341159e-01 4.89045829e-02 6.66104794e-01 -4.54566985e-01
-1.06778824e+00 4.88622874e-01 -9.18291509e-01 4.76112574e-01
-5.85140623e-02 2.01754645e-01 6.59839809e-01 -2.68078119e-01
4.39967543e-01 9.40039098e-01 8.91142428e-01 6.98778749e-01
-7.98093140e-01 -3.54277939e-01 2.74972200e-01 -1.03970811e-01
-1.04523337e+00 2.32503042e-01 2.75196493e-01 -4.48219508e-01
6.92804217e-01 5.44182003e-01 7.37353206e-01 6.92778885e-01
4.74020362e-01 1.05011058e+00 6.04358375e-01 -3.46576244e-01
-1.23333745e-01 3.70888300e-02 2.89959222e-01 9.55318272e-01
6.21607304e-01 -1.94321573e-02 3.82767826e-01 1.48797885e-01
1.11220396e+00 4.03003812e-01 -4.50435132e-01 -2.62342334e-01
-1.15205717e+00 8.57956052e-01 1.40251911e+00 3.21773976e-01
-3.64839107e-01 6.05960563e-02 5.61663032e-01 -1.72763929e-01
2.62604743e-01 6.79757237e-01 -3.09987336e-01 2.41050228e-01
-7.56868601e-01 -3.13326299e-01 4.05852050e-01 9.53202963e-01
1.69282690e-01 -3.85167688e-01 -4.82425660e-01 6.38241470e-01
2.28662267e-01 5.14161512e-02 7.89560914e-01 -5.99945962e-01
2.75120497e-01 1.02941084e+00 1.41458660e-02 -9.51679289e-01
-7.11954951e-01 -6.86277866e-01 -1.13432837e+00 -4.88507003e-02
4.09528881e-01 -1.33727342e-01 -1.57396793e+00 1.22565472e+00
2.32715219e-01 -2.52888538e-02 -2.32642338e-01 1.08191848e+00
1.11638105e+00 2.57422328e-01 3.00602913e-01 3.47410828e-01
1.49274123e+00 -1.56625366e+00 -5.47153592e-01 -2.62965798e-01
8.50467563e-01 -6.33043468e-01 9.60874498e-01 2.08674103e-01
-1.08561277e+00 -4.66678977e-01 -1.09173977e+00 -4.93510306e-01
-3.27380151e-01 4.25832570e-01 8.44226241e-01 6.75724149e-01
-1.05542040e+00 5.46946585e-01 -1.20518398e+00 -3.65955234e-01
9.63177562e-01 4.29368526e-01 -7.86576644e-02 -2.65265077e-01
-8.53603661e-01 6.57412469e-01 1.66629285e-01 4.06747669e-01
-7.25366771e-01 -8.51111770e-01 -1.02464437e+00 1.48903236e-01
7.44318068e-02 -7.25750566e-01 1.05328715e+00 -9.83560741e-01
-1.21923649e+00 1.05851042e+00 1.91070229e-01 -5.98714828e-01
7.65029490e-01 -1.23383671e-01 6.36522472e-02 3.48457187e-01
1.51349798e-01 9.99526680e-01 2.82071650e-01 -8.45879018e-01
-6.87656879e-01 -3.64125520e-01 1.07441388e-01 2.57726133e-01
-2.50241775e-02 -3.49206746e-01 -7.34437525e-01 -7.44176030e-01
1.89976156e-01 -8.76911700e-01 -7.98826694e-01 3.30138534e-01
-6.04660809e-01 1.30057931e-01 4.20592099e-01 -9.39233065e-01
1.30987573e+00 -2.19208407e+00 -5.24144955e-02 6.40876740e-02
2.42634624e-01 4.40458328e-01 -1.18781388e-01 -3.24030370e-01
2.52346369e-03 4.04467106e-01 -3.47860247e-01 9.57874805e-02
-5.09011090e-01 -5.16890958e-02 2.73146063e-01 6.44482255e-01
2.18025148e-01 1.15932035e+00 -1.14819217e+00 -6.06647134e-01
3.33625734e-01 4.55259651e-01 -8.47038805e-01 -1.86803881e-02
4.85218987e-02 5.00464022e-01 -4.50368255e-01 8.37010145e-01
7.44668365e-01 -5.34091830e-01 -4.10570810e-03 -1.79117158e-01
-6.52394295e-02 7.06717819e-02 -5.91749251e-01 1.92690074e+00
-3.88291389e-01 5.62671065e-01 2.60360360e-01 -9.22741413e-01
5.80611169e-01 6.75963163e-02 4.47711349e-01 -7.59083807e-01
4.37466711e-01 3.03668261e-01 5.01663208e-01 -7.06060469e-01
3.20506692e-01 1.00754306e-01 7.04274401e-02 -6.69726580e-02
-1.98836297e-01 3.31141055e-02 3.45873415e-01 1.26264304e-01
8.50083411e-01 -1.06738202e-01 3.59237581e-01 -6.32302225e-01
5.76211929e-01 9.41689089e-02 4.23522919e-01 7.57770479e-01
-6.66154504e-01 8.53639305e-01 4.24974263e-01 -7.18825936e-01
-7.28069603e-01 -8.39499176e-01 -3.64196867e-01 5.88313758e-01
5.17806172e-01 1.22855522e-01 -6.75951362e-01 -1.05047023e+00
-1.42235935e-01 3.69333982e-01 -1.19608366e+00 -1.66586667e-01
-5.92811406e-01 -8.23022783e-01 3.01801145e-01 7.98118114e-01
5.87707579e-01 -1.04771745e+00 -7.92016447e-01 1.15250424e-01
-1.20430537e-01 -5.61259091e-01 -8.69470179e-01 1.85536548e-01
-1.00913835e+00 -1.32559752e+00 -1.25392747e+00 -1.24345791e+00
1.33884025e+00 3.47196758e-01 1.09691870e+00 3.18874568e-01
-9.08577204e-01 -1.66333765e-01 -2.38261998e-01 -4.53858912e-01
-3.38037945e-02 3.37148488e-01 -7.68774092e-01 -3.77093911e-01
3.56186271e-01 6.97218701e-02 -1.32721150e+00 2.54809737e-01
-9.97668922e-01 1.52405232e-01 7.61286855e-01 1.12006032e+00
7.38838732e-01 -3.19411784e-01 2.56734639e-01 -1.07332253e+00
5.11460125e-01 -5.13899148e-01 -4.89735574e-01 1.91716284e-01
-1.43474951e-01 -1.18719533e-01 5.70984423e-01 -3.45436484e-01
-8.99694383e-01 5.73365539e-02 -2.65811175e-01 -2.49074295e-01
4.35544131e-03 3.27591181e-01 4.50527072e-01 -8.59665871e-02
6.37485147e-01 -1.42315058e-02 -3.28516923e-02 -2.31251732e-01
1.05378322e-01 4.80006993e-01 4.44512725e-01 -4.43002917e-02
7.90299103e-02 5.62165082e-01 -3.01878065e-01 -4.66666818e-01
-9.43549097e-01 -6.56215370e-01 -4.75168973e-01 2.52630144e-01
1.23308909e+00 -7.06682801e-01 -5.90464532e-01 3.40810686e-01
-8.94144177e-01 -4.28930461e-01 -2.34328598e-01 5.58062136e-01
4.22630832e-02 3.75062734e-01 -1.10841548e+00 -3.25042903e-01
-7.23833084e-01 -1.50134027e+00 8.89397621e-01 8.57650280e-01
-2.10022137e-01 -1.21878827e+00 -1.92222625e-01 8.20901394e-02
6.22861028e-01 3.80277336e-01 9.19488788e-01 -5.18233478e-01
-5.21132648e-01 -3.49587917e-01 -7.00756848e-01 9.01377127e-02
3.84236574e-01 1.00826755e-01 -5.98359108e-01 -3.85492742e-01
-2.23586261e-01 -6.81290701e-02 1.25330055e+00 1.18570042e+00
1.66573524e+00 -7.32676014e-02 -6.02900922e-01 1.11408377e+00
1.47718036e+00 4.38041359e-01 5.71438313e-01 3.17130387e-01
5.98601818e-01 3.90194267e-01 3.94530356e-01 -1.76590886e-02
2.08363920e-01 3.26722697e-03 7.11558759e-01 -6.96194053e-01
-3.20357025e-01 -2.33283881e-02 -3.72001499e-01 6.05482221e-01
1.76681757e-01 -1.48573935e-01 -9.02946591e-01 1.00781655e+00
-1.41252506e+00 -4.19468462e-01 -3.03491116e-01 2.03982425e+00
8.53178263e-01 6.08139485e-02 -2.39230067e-01 -3.69025230e-01
7.96800017e-01 -2.42482379e-01 -7.44759202e-01 -5.05970895e-01
2.70920664e-01 1.59774125e-01 7.87098706e-01 5.88335752e-01
-1.50551224e+00 6.06619120e-01 5.81542158e+00 6.21664643e-01
-1.28136194e+00 -1.44848377e-01 1.21227658e+00 7.40097044e-03
-9.92020443e-02 -4.70778197e-01 -6.57746851e-01 4.63639110e-01
3.28412294e-01 2.84808069e-01 -9.86820906e-02 6.79193199e-01
-1.46600813e-01 -2.15938643e-01 -8.34567368e-01 7.67396629e-01
-1.35447308e-01 -1.35814524e+00 5.18397987e-02 -1.13873228e-01
1.11331451e+00 1.80743054e-01 2.81929791e-01 2.69608349e-01
3.88021141e-01 -1.30471647e+00 1.08533300e-01 3.59484136e-01
1.00613236e+00 -5.79393029e-01 1.10106528e+00 2.47588996e-02
-9.86559987e-01 7.21727237e-02 -4.57870841e-01 3.56503785e-01
-1.49277866e-01 4.34917033e-01 -8.17401826e-01 3.58642340e-01
7.07355440e-01 7.43060231e-01 -7.01334476e-01 1.56392825e+00
-1.00125059e-01 3.22880685e-01 -1.25144616e-01 -2.62466162e-01
7.96177506e-01 -1.87072083e-01 1.69559583e-01 1.42276371e+00
3.39478999e-01 -1.31022260e-01 1.79975647e-02 8.22892845e-01
-1.73248500e-01 2.04041079e-01 -2.31149226e-01 2.57565171e-01
-1.64751187e-02 1.46017063e+00 -1.19934845e+00 -2.98966765e-01
-3.72447938e-01 1.04411829e+00 1.02889568e-01 2.46865734e-01
-9.60889876e-01 -7.48448014e-01 2.38425508e-01 -1.87388388e-03
2.83796251e-01 3.13837320e-01 -5.97907960e-01 -1.02387130e+00
-2.61309713e-01 -6.60037577e-01 5.80697179e-01 -2.53731817e-01
-1.22011459e+00 8.52488816e-01 -5.34563720e-01 -1.13875270e+00
3.75482887e-01 -8.14266682e-01 -7.88599193e-01 9.57899511e-01
-1.84619391e+00 -9.96457756e-01 -6.41866982e-01 3.51567209e-01
5.70587814e-01 3.21035355e-01 6.65269136e-01 3.63284379e-01
-6.79703355e-01 6.62254393e-01 1.63639277e-01 4.34322625e-01
6.34123087e-01 -1.57519293e+00 3.59053612e-01 6.65720761e-01
-4.40428436e-01 8.42515707e-01 2.05718935e-01 -6.45951211e-01
-8.72337401e-01 -1.23801994e+00 3.46591085e-01 -1.39856832e-02
3.79507393e-01 2.48512011e-02 -7.99431145e-01 6.84279621e-01
3.67379516e-01 3.91471505e-01 6.84663713e-01 -1.92687258e-01
1.23191111e-01 1.68131664e-01 -1.38285577e+00 7.14955270e-01
6.94559693e-01 9.89247411e-02 -2.37946510e-01 3.34627301e-01
6.63978696e-01 -8.42547178e-01 -7.05895483e-01 3.72541904e-01
4.31269407e-01 -8.05136681e-01 1.00819266e+00 -4.59013373e-01
7.30815053e-01 2.79128011e-02 4.29765821e-01 -1.23650503e+00
-5.58021069e-01 -2.99113542e-01 2.79480606e-01 4.51537728e-01
6.14890814e-01 -5.16481757e-01 9.13672030e-01 6.10631943e-01
-5.07591903e-01 -1.18023145e+00 -5.53886652e-01 -1.32764429e-01
4.69883412e-01 1.85510159e-01 2.92024225e-01 8.56738448e-01
-1.69529319e-01 -1.45075038e-01 1.74931422e-01 8.68447945e-02
4.49450672e-01 2.76080985e-02 1.50176719e-01 -9.83556271e-01
-7.86543824e-03 -8.07559669e-01 -7.55686015e-02 -1.28182411e+00
-6.09889984e-01 -1.02499783e+00 -4.73797396e-02 -1.91662788e+00
3.54351997e-01 -4.23185557e-01 -4.34694231e-01 3.86304259e-01
-5.89413285e-01 2.88954198e-01 -1.73122257e-01 -3.14382166e-02
-5.09798110e-01 1.59557268e-01 1.97134733e+00 -3.64421666e-01
-3.73809397e-01 2.20722035e-01 -1.01338840e+00 8.26948643e-01
8.96366358e-01 -1.39843524e-01 -2.82898068e-01 -6.51354194e-01
-2.38741189e-02 -6.37918636e-02 1.59708604e-01 -7.15356767e-01
3.50794047e-01 1.27739131e-01 7.43434429e-01 -6.79329634e-01
-3.55796605e-01 -6.99044585e-01 -4.88692880e-01 1.12860441e+00
-4.77349073e-01 -2.15484038e-01 4.99342322e-01 3.96395028e-01
-2.58111358e-01 -3.89795750e-01 1.08573186e+00 -5.79144418e-01
-3.84185225e-01 4.18982029e-01 -1.97695568e-01 6.21396862e-02
1.23339128e+00 -2.20107600e-01 -2.76872456e-01 3.70695293e-02
-8.18955898e-01 5.85898221e-01 2.06041396e-01 2.06568241e-01
6.84549093e-01 -1.07450283e+00 -7.27976978e-01 2.87248969e-01
-6.02427758e-02 7.06004977e-01 6.92441642e-01 1.23490930e+00
-1.37232518e+00 7.07109809e-01 -1.81209743e-01 -6.43238544e-01
-1.02995610e+00 6.40327275e-01 6.91219270e-01 -6.19843662e-01
-7.43431270e-01 1.39489079e+00 6.53491378e-01 -4.64615673e-01
3.35909873e-01 -9.99204338e-01 -2.74809420e-01 -1.50112793e-01
5.30321002e-01 2.10960180e-01 5.19705266e-02 -1.07205525e-01
-2.66950488e-01 4.38903570e-01 -6.19947672e-01 6.14447773e-01
1.16473413e+00 -2.01273128e-01 -1.01694793e-01 -2.23876745e-01
1.20991921e+00 -1.36906296e-01 -1.38952386e+00 -6.04907535e-02
-1.92620426e-01 -3.93621624e-01 2.14854434e-01 -1.14837372e+00
-1.48690879e+00 1.07614076e+00 7.26050556e-01 2.05241665e-01
1.15361881e+00 -3.02315176e-01 8.85923207e-01 -2.19270095e-01
-2.32540816e-01 -7.47260571e-01 -6.98243603e-02 2.55135208e-01
8.70965242e-01 -1.45949793e+00 -9.79521275e-02 -3.74568164e-01
-5.57804704e-01 1.15541410e+00 9.06656802e-01 -4.00217623e-01
5.37592530e-01 1.77824512e-01 1.65136904e-01 -2.03217685e-01
-2.01910749e-01 2.27130651e-02 6.18327200e-01 5.06988764e-01
8.16099882e-01 1.12413391e-01 -4.56577420e-01 6.23381197e-01
1.39726937e-01 -1.73390985e-01 4.88013893e-01 7.72634268e-01
-3.74926388e-01 -5.45088291e-01 -1.63135886e-01 8.22380483e-01
-9.43930149e-01 -4.01573330e-01 -1.44822016e-01 8.58432531e-01
9.11619514e-02 4.77846205e-01 1.78502753e-01 2.29078040e-01
1.70339167e-01 -5.62493742e-01 3.37516636e-01 -5.97696185e-01
-8.99078786e-01 2.72761017e-01 -3.49748045e-01 -3.76254648e-01
-1.25450060e-01 -3.36916476e-01 -1.43621719e+00 1.07799746e-01
-3.07143241e-01 -2.93060690e-02 4.61409539e-01 3.35003763e-01
2.26010174e-01 1.13362622e+00 2.07005948e-01 -5.95367432e-01
-5.02161562e-01 -1.00431085e+00 -3.26963753e-01 4.64399636e-01
5.23599267e-01 -2.11887017e-01 -1.66729629e-01 -9.85896140e-02] | [14.61913776397705, -2.73463773727417] |
9c998118-272b-4912-a2b2-95def8335950 | feature-learning-for-stock-price-prediction | 2103.09106 | null | https://arxiv.org/abs/2103.09106v1 | https://arxiv.org/pdf/2103.09106v1.pdf | Feature Learning for Stock Price Prediction Shows a Significant Role of Analyst Rating | To reject the Efficient Market Hypothesis a set of 5 technical indicators and 23 fundamental indicators was identified to establish the possibility of generating excess returns on the stock market. Leveraging these data points and various classification machine learning models, trading data of the 505 equities on the US S&P500 over the past 20 years was analysed to develop a classifier effective for our cause. From any given day, we were able to predict the direction of change in price by 1% up to 10 days in the future. The predictions had an overall accuracy of 83.62% with a precision of 85% for buy signals and a recall of 100% for sell signals. Moreover, we grouped equities by their sector and repeated the experiment to see if grouping similar assets together positively effected the results but concluded that it showed no significant improvements in the performance rejecting the idea of sector-based analysis. Also, using feature ranking we could identify an even smaller set of 6 indicators while maintaining similar accuracies as that from the original 28 features and also uncovered the importance of buy, hold and sell analyst ratings as they came out to be the top contributors in the model. Finally, to evaluate the effectiveness of the classifier in real-life situations, it was backtested on FAANG equities using a modest trading strategy where it generated high returns of above 60% over the term of the testing dataset. In conclusion, our proposed methodology with the combination of purposefully picked features shows an improvement over the previous studies, and our model predicts the direction of 1% price changes on the 10th day with high confidence and with enough buffer to even build a robotic trading system. | ['Matloob Khushi', 'Jaideep Singh'] | 2021-03-13 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-7.04551935e-01 3.03591460e-01 2.60448316e-03 -7.34519884e-02
-2.91586101e-01 -9.07846689e-01 8.37007940e-01 1.77225605e-01
-1.75948516e-01 7.37327576e-01 -1.94223821e-01 -5.97974777e-01
-5.41217923e-01 -1.05147362e+00 -5.79408765e-01 -4.77054089e-01
-4.46537673e-01 5.71909606e-01 3.36284071e-01 -4.84054297e-01
7.84402966e-01 5.63316703e-01 -1.63324726e+00 2.17699125e-01
3.90360802e-01 1.77762294e+00 -3.03264827e-01 1.97957844e-01
9.83061865e-02 1.01657903e+00 -9.25715625e-01 -6.58380151e-01
9.53269124e-01 -5.78718744e-02 -2.00390935e-01 -3.05212319e-01
-1.48467598e-02 -4.61602658e-01 3.73624921e-01 5.54467976e-01
1.33719102e-01 -3.98368716e-01 4.05745029e-01 -1.23891342e+00
-4.23240930e-01 9.66203332e-01 -4.00565624e-01 3.71066272e-01
2.11142734e-01 -3.54959033e-02 1.53010213e+00 -8.66315007e-01
3.89284730e-01 6.39705122e-01 5.72900176e-01 -3.60326380e-01
-9.84403789e-01 -1.06600320e+00 -1.14484102e-01 -1.03976801e-01
-8.38246346e-01 2.24565864e-01 6.29911780e-01 -6.40064180e-01
9.29297149e-01 4.42145497e-01 1.11024046e+00 5.16290545e-01
4.15702164e-01 9.98182520e-02 1.50708103e+00 -3.91134083e-01
1.06374159e-01 7.04391897e-01 6.13384321e-02 6.88389689e-02
6.74963295e-01 5.72246611e-01 -3.17466676e-01 -2.85811096e-01
5.68380475e-01 -3.02763190e-02 2.54460514e-01 5.16820885e-02
-1.25664365e+00 1.12255657e+00 1.97670698e-01 6.79683924e-01
-4.66133773e-01 -4.14971769e-01 2.32633844e-01 7.55943179e-01
5.13264537e-01 9.08906579e-01 -9.59499478e-01 -6.89706057e-02
-9.22797918e-01 2.91171998e-01 1.13323343e+00 3.14120442e-01
4.36251581e-01 2.14020446e-01 -2.12204047e-02 1.93661943e-01
5.85434437e-02 4.60222185e-01 7.40152359e-01 -5.76859295e-01
3.66058618e-01 8.85784388e-01 2.24666670e-01 -1.03149593e+00
-5.27942300e-01 -9.10177350e-01 -2.31672123e-01 6.15471601e-01
7.00391352e-01 -3.38942111e-01 -4.27418321e-01 1.34777367e+00
-1.41410947e-01 -1.36045083e-01 8.04208219e-02 5.83498120e-01
-7.12627843e-02 4.74480122e-01 -5.53429544e-01 -2.96754152e-01
1.22295427e+00 -4.55218554e-01 -3.58259648e-01 3.32113802e-01
3.49694669e-01 -7.98665881e-01 5.33095419e-01 9.09588337e-01
-9.04044092e-01 -6.05728686e-01 -1.27814353e+00 9.43444133e-01
-6.25980318e-01 -1.99806578e-02 8.43694508e-01 6.48193836e-01
-7.82838762e-01 9.56335545e-01 -3.52033913e-01 3.40544999e-01
-1.67847738e-01 4.09620106e-01 3.35838506e-03 9.59878325e-01
-1.44365585e+00 1.11909890e+00 5.02776504e-01 -7.28302300e-02
-4.01273459e-01 -7.28739560e-01 -9.97138098e-02 2.70079702e-01
1.20926112e-01 9.11861211e-02 7.50198901e-01 -1.01609349e+00
-1.17906702e+00 5.47812581e-01 7.86119878e-01 -1.18774152e+00
8.33235562e-01 -1.21404178e-01 -8.29598665e-01 -1.74346492e-01
1.30349526e-03 -3.55541036e-02 6.61789477e-01 -7.09818602e-01
-1.14503682e+00 -3.53343397e-01 6.99843168e-02 -3.58883947e-01
-2.50609398e-01 -4.84806411e-02 4.77397919e-01 -1.00944710e+00
1.18154176e-01 -9.69307959e-01 2.68390954e-01 -8.57727706e-01
-1.07827876e-02 3.08561269e-02 4.66309518e-01 -7.22995162e-01
1.29299271e+00 -1.84856164e+00 -5.92483401e-01 8.18650901e-01
-1.68498550e-02 -4.06657085e-02 5.52373886e-01 6.16130054e-01
-6.15685761e-01 1.79395825e-01 2.32704163e-01 5.75802386e-01
1.77379385e-01 -2.71148443e-01 -8.92406285e-01 1.72213972e-01
2.01115355e-01 7.11525083e-01 -2.91898519e-01 3.07749122e-01
1.49307713e-01 -7.65573457e-02 -3.60761106e-01 4.79850322e-02
8.52213502e-02 2.13706549e-02 -1.81381762e-01 6.38209045e-01
4.43394721e-01 -2.12495960e-02 1.27752367e-02 1.30756646e-02
-3.81255776e-01 2.10299119e-01 -1.23658574e+00 3.29957485e-01
-2.22012013e-01 5.55378914e-01 -5.02467692e-01 -8.69205832e-01
1.43351305e+00 3.04151416e-01 6.21650100e-01 -1.02216315e+00
8.81842524e-02 7.68522143e-01 5.05380034e-01 1.25885263e-01
2.99643397e-01 -2.52387881e-01 -2.03179076e-01 6.57265544e-01
-2.07589194e-01 9.81953591e-02 7.32378289e-02 -2.27509081e-01
8.40865552e-01 -1.99947640e-01 1.51320904e-01 -5.12615860e-01
4.15479869e-01 -2.39159111e-02 4.15006787e-01 4.32957202e-01
2.04709604e-01 2.60603786e-01 8.04388344e-01 -6.11652970e-01
-8.47866476e-01 -8.92855704e-01 -2.53091127e-01 7.58318007e-01
-3.14219296e-01 1.59370854e-01 -3.29240561e-01 -5.52955210e-01
6.68285191e-01 1.04093194e+00 -7.22153008e-01 3.59889902e-02
2.51596477e-02 -1.09330034e+00 3.79323751e-01 1.55436382e-01
3.00718904e-01 -1.01949441e+00 -1.03705096e+00 2.63184547e-01
4.12809342e-01 -5.87532699e-01 2.84888178e-01 5.27967036e-01
-8.20536554e-01 -1.32006538e+00 -6.24218822e-01 -1.42966673e-01
3.25387090e-01 -4.40549642e-01 1.10669386e+00 -8.94609466e-02
5.13073504e-02 -9.85437483e-02 -2.98123926e-01 -1.11419880e+00
-3.52936655e-01 -2.30482504e-01 1.12569615e-01 2.07670823e-01
3.36543739e-01 -4.93928194e-01 -4.72222745e-01 5.21036804e-01
-6.26985788e-01 -5.79601288e-01 6.59961879e-01 6.05910182e-01
8.54784101e-02 4.61339623e-01 1.31213725e+00 -6.64596677e-01
6.98815644e-01 -5.16035199e-01 -1.21609139e+00 8.95160437e-02
-1.38508117e+00 1.49911523e-01 5.23706794e-01 -3.86749446e-01
-8.32414150e-01 -4.05909747e-01 3.10339570e-01 -1.87854752e-01
3.76872778e-01 5.77022552e-01 4.38989282e-01 2.01530963e-01
4.20206964e-01 3.80197503e-02 1.34958312e-01 -6.24334872e-01
-2.03464404e-01 4.07216877e-01 1.31216735e-01 -2.17426017e-01
8.96693468e-01 3.04416448e-01 1.46039143e-01 -1.22615062e-01
-6.14658654e-01 -1.95467994e-01 -4.64710802e-01 -2.31360555e-01
2.18315929e-01 -9.10256743e-01 -9.05071974e-01 5.70663691e-01
-4.78677332e-01 4.32169661e-02 -4.23884183e-01 8.54944885e-01
-3.57987404e-01 -1.81206495e-01 -5.31340480e-01 -1.16495907e+00
-2.43453428e-01 -7.83959508e-01 1.94228232e-01 -6.33101687e-02
-5.45060933e-01 -9.00981247e-01 1.58489242e-01 3.15672457e-01
6.52770936e-01 6.41599417e-01 8.33902717e-01 -1.66437972e+00
-3.77912998e-01 -6.31724417e-01 -5.12074456e-02 5.18817008e-01
3.01655471e-01 1.50765613e-01 -8.50697219e-01 -2.43793175e-01
3.88714463e-01 -5.81062250e-02 5.92524290e-01 9.39726904e-02
2.23422647e-01 -4.12623763e-01 1.99636266e-01 -1.34744691e-02
1.42032981e+00 6.56272650e-01 3.86651069e-01 1.05185926e+00
-1.29498109e-01 8.61497045e-01 1.00273418e+00 7.06613779e-01
-1.88970447e-01 5.94183445e-01 5.85207641e-01 1.25215992e-01
5.73967993e-01 -7.00543225e-02 6.27520442e-01 4.57671374e-01
-3.10722709e-01 3.17086846e-01 -6.92684710e-01 4.51142430e-01
-1.39438617e+00 -1.03568232e+00 -1.12361573e-01 2.53240609e+00
3.95245701e-01 1.02567589e+00 4.90939051e-01 4.46478903e-01
4.29646879e-01 -9.20373574e-02 -3.57736409e-01 -4.98009443e-01
-2.36559838e-01 2.57923633e-01 9.66705859e-01 9.24283937e-02
-7.82918215e-01 1.93706021e-01 6.50513649e+00 4.54906970e-01
-1.40053308e+00 -4.95869160e-01 8.84802163e-01 -2.01882094e-01
-3.72383147e-01 1.64095789e-01 -9.26061511e-01 8.75796139e-01
1.34634519e+00 -4.08004254e-01 2.21786961e-01 8.72314990e-01
2.18380243e-01 -2.00451761e-01 -8.06924045e-01 4.01195973e-01
-9.73443538e-02 -1.22727370e+00 -1.99900046e-02 6.31349742e-01
6.20489776e-01 -5.77191189e-02 2.21942633e-01 3.40115458e-01
1.06157035e-01 -8.63655031e-01 1.21972227e+00 8.55677843e-01
7.92373046e-02 -9.00244951e-01 1.18365407e+00 4.31579322e-01
-7.76305795e-01 -5.00340879e-01 -7.98288174e-03 -4.85737085e-01
-1.76705882e-01 7.24417031e-01 -9.73049581e-01 6.27968431e-01
6.86694443e-01 2.51327306e-01 -7.57647216e-01 7.04413116e-01
2.14282289e-01 5.48490703e-01 -5.90064406e-01 7.85479695e-02
3.04330081e-01 -4.39604342e-01 2.97024727e-01 6.65422082e-01
9.48542595e-01 -2.81829089e-01 -5.79876423e-01 6.97732449e-01
2.79852867e-01 2.22921833e-01 -5.51899195e-01 5.19811846e-02
2.70579338e-01 1.07527113e+00 -8.29569101e-01 -2.53553182e-01
-5.86525500e-01 2.41599232e-01 -2.83770293e-01 -8.23118016e-02
-9.41009402e-01 -3.90167177e-01 1.50076374e-01 5.69470584e-01
5.45650899e-01 3.50625068e-01 -4.13791895e-01 -8.71732712e-01
2.60849565e-01 -9.08908367e-01 4.70501959e-01 -4.36852396e-01
-1.11284792e+00 4.79300052e-01 -8.38710666e-02 -1.62530494e+00
-6.83671772e-01 -7.58072674e-01 -6.82347417e-01 1.00732362e+00
-1.39868474e+00 -5.82726717e-01 2.96503037e-01 2.39521861e-01
8.25568363e-02 -8.37904751e-01 4.87104148e-01 -1.19207818e-02
-7.53704682e-02 2.92506814e-01 3.97851676e-01 2.10571468e-01
5.34702599e-01 -1.48191941e+00 1.47913501e-01 6.44067407e-01
3.31080317e-01 6.43060267e-01 6.40749931e-01 -7.67456532e-01
-6.88051105e-01 -5.01022577e-01 8.72535110e-01 -5.60994744e-01
1.25734961e+00 6.59342557e-02 -7.35977471e-01 5.89351773e-01
2.57991046e-01 -5.66201389e-01 6.01170182e-01 -9.06821042e-02
-2.10098147e-01 -3.71580303e-01 -1.18669403e+00 5.97266108e-02
7.18523115e-02 -5.87189719e-02 -1.02050769e+00 -1.79197282e-01
2.15119705e-01 8.42540637e-02 -1.38176692e+00 5.04043937e-01
9.64420557e-01 -1.61816919e+00 8.17229211e-01 -2.67546356e-01
1.03969298e-01 -9.76190269e-02 -2.61137616e-02 -1.17914093e+00
-1.71434537e-01 -4.13934380e-01 -1.48590310e-02 1.30393076e+00
9.16929424e-01 -1.25209606e+00 6.98210120e-01 6.79324627e-01
3.92768562e-01 -8.30193400e-01 -8.78976345e-01 -9.27188933e-01
1.21664330e-01 -2.37682328e-01 8.56487334e-01 8.21842730e-01
1.24255523e-01 -3.37705947e-02 -2.78133988e-01 -1.87236607e-01
3.97097826e-01 5.82355857e-01 6.27738118e-01 -1.73936236e+00
-4.89602178e-01 -7.78646111e-01 -3.41481417e-01 -1.13375224e-01
-2.16176942e-01 -6.73554838e-01 -7.56649911e-01 -7.18776584e-01
-1.49832666e-01 -3.93451244e-01 -9.19363976e-01 2.39776015e-01
2.49598667e-01 1.20134890e-01 3.75332057e-01 5.47728658e-01
2.82064110e-01 -9.72317457e-02 7.86131322e-01 1.64554119e-02
-1.43494338e-01 5.12802720e-01 -8.95099938e-01 6.52031660e-01
8.27440679e-01 -9.94848832e-02 -2.57674694e-01 4.84591901e-01
6.12467289e-01 9.05802697e-02 2.33054742e-01 -8.25679839e-01
-2.14494660e-01 -1.27317756e-03 7.14087069e-01 -7.24086046e-01
7.01896772e-02 -1.03905809e+00 6.57456279e-01 8.02651763e-01
-1.80409983e-01 3.56428832e-01 1.15379021e-01 2.66023993e-01
-5.37492692e-01 -2.64674187e-01 3.24009448e-01 -1.24220051e-01
-4.70210046e-01 -3.77587676e-01 -1.38344482e-01 -2.38080874e-01
1.50639343e+00 -2.95139909e-01 -1.02548100e-01 -4.77034569e-01
-7.65003026e-01 1.17967382e-01 3.85156602e-01 6.08655751e-01
5.86451516e-02 -9.70750511e-01 -7.86773980e-01 2.69711554e-01
-6.89771306e-03 -7.83975363e-01 -3.88037980e-01 6.85372889e-01
-6.19638383e-01 8.54040980e-01 -4.30777550e-01 -7.73666650e-02
-8.71328115e-01 5.03440022e-01 4.29501444e-01 -7.03956008e-01
-2.97646254e-01 3.57290894e-01 -3.97190154e-02 -2.69550681e-02
8.06701332e-02 -6.34370387e-01 -5.13993740e-01 1.02288187e+00
4.12865758e-01 4.68560159e-01 4.06273037e-01 -4.83768612e-01
-2.51832902e-01 4.58421469e-01 5.46390489e-02 -3.49852771e-01
1.65340054e+00 1.42422661e-01 -2.73961183e-02 7.56026804e-01
8.46076071e-01 4.54876065e-01 -9.49293792e-01 2.89253503e-01
6.53683841e-01 -3.68679762e-01 -3.33303720e-01 -1.03093565e+00
-1.33808422e+00 3.99435461e-01 5.21314800e-01 1.07484150e+00
8.99421453e-01 -2.78250754e-01 4.51198608e-01 1.95728630e-01
6.14900231e-01 -1.29273570e+00 -2.76182294e-01 9.97980088e-02
9.43331361e-01 -1.08245683e+00 5.83863743e-02 1.54818952e-01
-7.17410266e-01 1.31332922e+00 -2.50755753e-02 -5.05621016e-01
9.74916875e-01 1.97962061e-01 1.19224042e-01 -2.92300045e-01
-8.04983377e-01 2.05814138e-01 1.40999720e-01 3.04577667e-02
2.57956296e-01 2.28388578e-01 -6.13175511e-01 9.38651562e-01
-7.81654119e-01 -5.39893061e-02 6.92696214e-01 5.71380138e-01
-4.40630555e-01 -9.35993135e-01 -5.13788164e-01 8.08751941e-01
-1.14386415e+00 8.37932304e-02 -4.46955144e-01 1.34441829e+00
3.31534632e-02 7.31497526e-01 4.56054449e-01 -4.98564065e-01
6.06887221e-01 2.98199415e-01 -2.41026774e-01 -1.91394478e-01
-1.01391673e+00 1.59606099e-01 1.40057549e-01 -3.25165808e-01
-2.18549475e-01 -9.51749384e-01 -1.04039645e+00 -2.94564277e-01
-5.38930118e-01 5.40298879e-01 4.74916130e-01 8.07678580e-01
4.96682562e-02 3.46159965e-01 1.33631420e+00 -4.75075066e-01
-1.04015708e+00 -1.09220648e+00 -1.09524453e+00 4.96075720e-01
1.72023192e-01 -7.23405957e-01 -1.08481228e+00 -1.66238785e-01] | [4.564206600189209, 4.177685260772705] |
d6ea87da-ce26-44ea-8364-b2f6bea5ffa7 | cyberwalle-at-semeval-2020-task-11-an | 2008.09859 | null | https://arxiv.org/abs/2008.09859v1 | https://arxiv.org/pdf/2008.09859v1.pdf | CyberWallE at SemEval-2020 Task 11: An Analysis of Feature Engineering for Ensemble Models for Propaganda Detection | This paper describes our participation in the SemEval-2020 task Detection of Propaganda Techniques in News Articles. We participate in both subtasks: Span Identification (SI) and Technique Classification (TC). We use a bi-LSTM architecture in the SI subtask and train a complex ensemble model for the TC subtask. Our architectures are built using embeddings from BERT in combination with additional lexical features and extensive label post-processing. Our systems achieve a rank of 8 out of 35 teams in the SI subtask (F1-score: 43.86%) and 8 out of 31 teams in the TC subtask (F1-score: 57.37%). | ['Verena Blaschke', 'Sam Tureski', 'Maxim Korniyenko'] | 2020-08-22 | null | https://aclanthology.org/2020.semeval-1.192 | https://aclanthology.org/2020.semeval-1.192.pdf | semeval-2020 | ['propaganda-technique-identification', 'propaganda-detection', 'propaganda-span-identification'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-2.60534827e-02 1.59935161e-01 -2.07665831e-01 -1.34255186e-01
-1.01177216e+00 -6.68951094e-01 1.13414443e+00 3.42895836e-01
-9.38356221e-01 6.13701284e-01 6.84088171e-01 -5.60343623e-01
9.99230295e-02 -4.43669170e-01 -4.83890384e-01 -2.38435254e-01
-3.28948200e-02 3.77762526e-01 -2.34155264e-02 -2.73968399e-01
5.82490981e-01 2.60017902e-01 -9.25921440e-01 8.04784179e-01
7.96183884e-01 1.15418613e+00 -5.28186381e-01 8.77851009e-01
-1.29830092e-01 1.60997486e+00 -9.37107384e-01 -5.66102326e-01
7.57939219e-02 -2.91074514e-01 -1.10741687e+00 -5.02094507e-01
5.14513195e-01 -1.33516237e-01 -3.53123009e-01 8.56152058e-01
4.07505512e-01 8.03314373e-02 7.48211145e-01 -7.06820011e-01
-6.65771902e-01 1.04947507e+00 -6.23441100e-01 7.10417271e-01
4.38231289e-01 -1.97749972e-01 1.01151109e+00 -1.10939062e+00
8.35273445e-01 1.01524305e+00 8.97967994e-01 3.30700964e-01
-1.30626523e+00 -7.94321358e-01 -1.34358257e-01 3.54411960e-01
-9.02827740e-01 -6.09102190e-01 6.82249308e-01 -9.21980083e-01
1.21664906e+00 -2.03869417e-01 3.64771217e-01 1.85187685e+00
4.74253297e-01 9.01193023e-01 1.39495111e+00 -4.59138721e-01
5.36985137e-02 2.34673858e-01 8.57268631e-01 7.89256155e-01
-5.67337573e-02 1.57326609e-01 -7.38137901e-01 -2.13978201e-01
4.56092179e-01 -4.49016213e-01 1.13691941e-01 8.39918256e-01
-1.34361708e+00 1.24947786e+00 3.24959755e-01 5.15472531e-01
-5.24125695e-01 1.22870788e-01 8.98160100e-01 5.79793215e-01
9.96024191e-01 9.85402107e-01 -3.32162261e-01 -3.37494612e-01
-1.19036448e+00 4.01600778e-01 1.00956845e+00 4.37400639e-01
1.37686178e-01 1.73343793e-01 -6.15200937e-01 9.62927580e-01
-4.60018426e-01 1.46956667e-01 4.06295002e-01 -7.77197301e-01
7.28459060e-01 3.91888052e-01 6.85047358e-02 -8.90533745e-01
-7.62070596e-01 -8.86521041e-01 -6.35842025e-01 3.25227194e-02
4.29082006e-01 -5.30957460e-01 -1.03160727e+00 1.59552598e+00
-1.04435809e-01 -2.14659825e-01 -4.63167071e-01 5.21140218e-01
9.42136586e-01 7.14757383e-01 2.91172594e-01 -3.73765796e-01
1.39680493e+00 -1.20177543e+00 -6.24591708e-01 -2.03205407e-01
8.32706988e-01 -8.26144576e-01 5.18238604e-01 4.84599203e-01
-8.51361334e-01 -7.16524303e-01 -7.83406615e-01 -7.75062442e-02
-2.86418021e-01 2.68640637e-01 1.63350955e-01 2.49556735e-01
-7.94076800e-01 3.83702427e-01 -4.42474961e-01 -2.17934519e-01
4.48163271e-01 -9.56180170e-02 -3.78105491e-01 3.15794468e-01
-1.32315505e+00 1.27251065e+00 4.78563696e-01 -2.36704305e-01
-1.19230592e+00 -7.48577535e-01 -6.66029930e-01 -2.78576817e-02
2.95483232e-01 -2.01882794e-01 1.14318967e+00 -7.53329813e-01
-1.09974480e+00 1.13926136e+00 2.02395305e-01 -6.50961757e-01
5.70605457e-01 -6.36005998e-01 -7.35497475e-01 1.08758137e-01
3.07474583e-01 8.48856643e-02 7.26841509e-01 -7.67804265e-01
-7.62338817e-01 -2.07238421e-01 -3.33076604e-02 -1.14514157e-01
-3.20586205e-01 8.05094063e-01 3.36727351e-01 -8.41657579e-01
-2.93438017e-01 -8.07207942e-01 1.51994139e-01 -7.21012115e-01
-6.42505467e-01 -6.82669342e-01 4.75151330e-01 -1.53932214e+00
1.61256588e+00 -1.76215971e+00 2.77061135e-01 -4.32115868e-02
6.47492170e-01 2.83471912e-01 -2.38497570e-01 6.85563684e-01
-2.37113684e-02 2.11028665e-01 4.02767956e-01 -4.38058853e-01
-1.11352563e-01 -4.75713879e-01 -5.24673104e-01 4.46322262e-01
2.90621072e-01 6.95728183e-01 -7.42878199e-01 -2.89456278e-01
-2.69674897e-01 2.72096023e-02 -3.25563729e-01 7.92059377e-02
-1.67238951e-01 3.41277093e-01 -3.44920099e-01 5.49585044e-01
1.86460733e-01 -5.65092452e-02 -1.35118350e-01 9.42577645e-02
-3.93391132e-01 8.70807350e-01 -4.13130611e-01 1.42635024e+00
-5.12864053e-01 1.04312432e+00 1.31323695e-01 -8.72644663e-01
8.08609664e-01 3.31295788e-01 2.39387408e-01 -6.69089973e-01
4.46324438e-01 2.83465147e-01 3.11234266e-01 -4.46980000e-01
6.55767858e-01 -1.54264525e-01 -4.31688219e-01 9.39623892e-01
3.53869587e-01 6.27698600e-01 4.04267699e-01 3.11325371e-01
1.51070142e+00 -8.85935947e-02 2.04831883e-01 -4.16367769e-01
3.52575302e-01 7.10732862e-02 3.98508936e-01 1.03435731e+00
-3.45565915e-01 2.09978282e-01 8.47844303e-01 -9.37328637e-01
-1.08629620e+00 -7.53726959e-01 -9.03736651e-02 1.74011099e+00
-6.40683174e-01 -3.74312758e-01 -6.48321509e-01 -1.20807910e+00
-1.57610074e-01 9.69183803e-01 -1.11038446e+00 2.02892125e-02
-9.66676116e-01 -4.49494481e-01 9.48925138e-01 6.11525059e-01
4.82164592e-01 -1.34559464e+00 -5.10911167e-01 5.13851285e-01
-2.94939846e-01 -1.10973215e+00 -2.22351089e-01 4.25519139e-01
-3.01537693e-01 -1.03213310e+00 -5.33690095e-01 -8.84237111e-01
4.14434500e-04 -2.93091863e-01 1.06200325e+00 -2.19242230e-01
-7.38242194e-02 -3.67234975e-01 -6.23507619e-01 -5.14606237e-01
-6.77195609e-01 5.33667624e-01 2.46351957e-01 -2.65474826e-01
5.36005259e-01 -2.16723010e-01 -1.38508692e-01 -5.65579981e-02
-3.21122080e-01 1.65802196e-01 2.95667917e-01 1.07927847e+00
-1.14205524e-01 -4.09623295e-01 9.18740273e-01 -8.94372582e-01
8.70731652e-01 -6.10128701e-01 -2.64297754e-01 1.74729586e-01
-3.45034719e-01 -2.35634565e-01 6.76457465e-01 -5.27099311e-01
-8.98939312e-01 -4.89480704e-01 -3.49817187e-01 -2.87275892e-02
4.63957600e-02 7.13789761e-01 6.55658722e-01 3.20785522e-01
1.28137577e+00 -5.25661521e-02 -1.20684654e-01 -6.91203117e-01
7.12399185e-02 9.57514524e-01 3.65509152e-01 -2.80450225e-01
6.38175309e-01 -1.96242735e-01 -5.46716928e-01 -5.92125475e-01
-1.72564077e+00 -4.10816848e-01 -3.56044620e-01 -2.96866775e-01
1.10227239e+00 -8.79902840e-01 -3.57843876e-01 5.74129462e-01
-1.38449991e+00 -3.48370701e-01 -2.81825271e-02 5.88929415e-01
-1.91074565e-01 -1.60115361e-02 -1.13946056e+00 -8.83672237e-01
-7.19739854e-01 -7.98143446e-01 7.21323490e-01 -3.16936560e-02
-5.47088921e-01 -1.01169801e+00 2.30485782e-01 5.20819664e-01
4.35649127e-01 4.19890970e-01 8.90124977e-01 -1.35581398e+00
4.33894247e-01 -4.58663821e-01 -3.93271476e-01 4.44729626e-01
-3.37883949e-01 -3.11970562e-01 -1.06539857e+00 -1.84464380e-01
5.19969277e-02 -7.12878883e-01 1.40485263e+00 5.23841977e-01
1.02280653e+00 -4.09280509e-01 -1.49415955e-01 7.55981728e-02
9.29252505e-01 1.92353740e-01 3.90972972e-01 5.41803896e-01
6.44594312e-01 6.53330982e-01 1.75109446e-01 1.71809450e-01
1.81425121e-02 6.68303490e-01 -2.40713507e-02 2.45299369e-01
-1.44655779e-01 -4.13998872e-01 7.73359179e-01 8.54179204e-01
-1.45622060e-01 -1.95078775e-01 -1.26109254e+00 5.70464134e-01
-1.67493927e+00 -1.46211481e+00 -3.73752385e-01 1.66576946e+00
9.24943566e-01 4.01864290e-01 5.17338872e-01 1.51696324e-01
7.37949073e-01 4.49011147e-01 7.49685094e-02 -1.00584269e+00
8.67383406e-02 2.42683411e-01 2.62133420e-01 5.36505103e-01
-1.69963205e+00 1.02273142e+00 6.80790424e+00 1.16263545e+00
-6.97641551e-01 5.69652081e-01 5.87424397e-01 -2.09297538e-01
8.53140950e-02 -2.38274679e-01 -9.16231453e-01 6.70298636e-01
1.39428937e+00 6.54642135e-02 2.94341832e-01 6.37376487e-01
-1.44796833e-01 6.19665300e-03 -9.27165031e-01 6.41286492e-01
4.10470605e-01 -1.68476450e+00 -3.20979685e-01 2.09245130e-01
8.81909311e-01 4.26095843e-01 -1.02444932e-01 8.75884354e-01
6.47426367e-01 -1.19616377e+00 9.39066112e-01 3.89744610e-01
8.49489033e-01 -7.23537445e-01 9.72273231e-01 7.90365994e-01
-4.87743109e-01 -3.87966305e-01 -8.66742805e-03 -3.97798896e-01
5.46058118e-01 7.49627292e-01 -9.22977567e-01 4.49341357e-01
4.72548455e-01 6.29221916e-01 -5.38287640e-01 7.24416614e-01
-6.68992043e-01 1.22369671e+00 -2.03354120e-01 -1.75783649e-01
4.51737255e-01 3.11109602e-01 8.10814023e-01 1.60242105e+00
-2.28378903e-02 -2.75826007e-01 3.37919712e-01 6.00005865e-01
-5.02247334e-01 -1.98010579e-02 -2.08268031e-01 -3.19611728e-01
4.78311479e-01 1.18945897e+00 -3.47991973e-01 -5.19416511e-01
1.78992108e-01 4.71342891e-01 7.79815733e-01 1.32726803e-01
-8.94611478e-01 -5.55873692e-01 8.34769905e-02 -5.95581718e-02
1.21243760e-01 -1.19598418e-01 -5.71181417e-01 -1.02963066e+00
-3.70823711e-01 -8.84295046e-01 4.84415054e-01 -2.71711916e-01
-1.53115201e+00 9.08107221e-01 -1.78196937e-01 -6.55703843e-01
-3.89992803e-01 -7.28547573e-01 -7.81907082e-01 7.85676241e-01
-8.46879363e-01 -1.43384469e+00 1.48969397e-01 -6.36227876e-02
7.87674546e-01 -7.07867324e-01 8.66657317e-01 2.37185866e-01
-6.57140791e-01 5.16989172e-01 1.25240445e-01 5.32761693e-01
7.00429559e-01 -1.14374495e+00 4.07913059e-01 6.31890416e-01
1.28641948e-01 5.88968396e-01 8.01580250e-01 -6.66018009e-01
-5.33675194e-01 -1.04774749e+00 1.51993334e+00 -7.34645307e-01
1.07246161e+00 -3.75262916e-01 -5.76186299e-01 8.37777793e-01
2.99879134e-01 -5.86074412e-01 8.26533377e-01 8.15615296e-01
-9.54277277e-01 5.19664645e-01 -8.74060929e-01 2.02749968e-01
9.33453918e-01 -9.19060707e-01 -9.91552651e-01 6.56717598e-01
4.83800352e-01 -1.84535578e-01 -9.54954326e-01 2.01898478e-02
8.65472078e-01 -5.78282833e-01 4.79249358e-01 -1.04610705e+00
1.17693484e+00 3.37237448e-01 1.20487280e-01 -1.14789903e+00
-6.91773415e-01 -2.04559848e-01 -3.88690293e-01 9.97267127e-01
6.08116090e-01 -3.21889222e-01 3.47410679e-01 -2.48782083e-01
-2.77201563e-01 -8.17665875e-01 -1.10928285e+00 -8.10870051e-01
5.70081413e-01 -4.26445663e-01 -2.14912325e-01 1.12523723e+00
6.54151365e-02 8.50064754e-01 -6.69061840e-01 -3.94481093e-01
5.05402625e-01 6.25965223e-02 3.92914146e-01 -1.32150972e+00
-2.66363442e-01 -6.86309278e-01 -2.41329763e-02 -5.79604864e-01
4.69748408e-01 -1.02695680e+00 -2.28839874e-01 -1.74426413e+00
4.87948596e-01 8.34368542e-02 -4.59376842e-01 4.93823022e-01
-8.46035965e-03 3.08376759e-01 8.85284021e-02 3.50393504e-01
-7.33708501e-01 1.65091082e-01 5.06898999e-01 -2.13318273e-01
-4.22400236e-02 -3.27123642e-01 -6.12451911e-01 8.41289639e-01
1.06525719e+00 -7.51309514e-01 4.27389801e-01 -5.93560159e-01
2.59002924e-01 -2.41614267e-01 3.68159622e-01 -9.05678988e-01
-2.51986310e-02 -4.93610613e-02 5.54136813e-01 -6.80453479e-01
1.26702413e-01 9.54687819e-02 -3.06239218e-01 7.49226689e-01
-7.20925927e-01 -5.28485365e-02 1.46279829e-02 3.89741063e-01
-4.45640786e-03 -4.07864422e-01 4.47648525e-01 1.10427871e-01
-3.40117425e-01 -2.19997019e-01 -7.94948161e-01 2.37518281e-01
7.49207437e-01 1.19749792e-01 -9.21391487e-01 -2.57559627e-01
-8.17291915e-01 1.84007585e-01 -3.48932631e-02 5.05865872e-01
1.53401569e-01 -1.13210547e+00 -1.17951143e+00 -6.99648708e-02
-2.23777629e-02 -6.79438591e-01 2.94160217e-01 1.23235261e+00
-5.18901050e-01 3.46449703e-01 -2.81902254e-01 1.04053803e-02
-1.06364655e+00 2.35877931e-01 -8.86508822e-02 -9.74407315e-01
-3.80432785e-01 1.22025669e+00 -2.59748936e-01 -2.20004112e-01
6.13922104e-02 4.11659956e-01 -6.41484082e-01 5.95716596e-01
8.27055275e-01 7.13940084e-01 -8.23943934e-04 -6.95144773e-01
-3.52667987e-01 7.19833374e-02 -6.82805240e-01 -2.70478845e-01
1.54483366e+00 5.31592190e-01 7.72766620e-02 4.25488174e-01
1.24597073e+00 2.90048391e-01 -5.39190888e-01 -2.64095008e-01
3.13528538e-01 1.58629358e-01 4.10651118e-01 -1.37913382e+00
-4.48397756e-01 7.70065188e-01 1.41717389e-01 5.52769005e-01
5.16874015e-01 -1.80857517e-02 8.13526154e-01 2.42746085e-01
2.88468063e-01 -1.61419427e+00 9.39141288e-02 1.13800001e+00
1.23181248e+00 -9.24074769e-01 1.17381424e-01 1.36075914e-01
-6.60942197e-01 9.69806135e-01 3.25736523e-01 -5.45575202e-01
3.72031242e-01 1.15235873e-01 -9.95286107e-02 -6.08474851e-01
-9.36056972e-01 -3.80536877e-02 4.74244475e-01 1.05129294e-01
6.18204474e-01 9.60436314e-02 -8.71010423e-01 7.12973475e-01
-2.73778707e-01 -1.54708788e-01 2.16933444e-01 7.24320889e-01
-6.96892560e-01 -5.49182594e-01 -1.23270363e-01 8.76071930e-01
-9.77958798e-01 -9.36654210e-02 -6.90580726e-01 5.87711573e-01
1.92810848e-01 1.07375801e+00 1.03536196e-01 -9.19000685e-01
1.25969335e-01 4.48145688e-01 2.39778385e-01 -7.00402379e-01
-1.34021282e+00 4.01581712e-02 1.02223682e+00 -4.14269865e-01
-2.43896648e-01 -4.50077534e-01 -5.59273183e-01 -5.20108223e-01
-3.80242497e-01 2.38598302e-01 6.00591540e-01 9.40389276e-01
8.91486332e-02 6.16486132e-01 5.55822968e-01 -7.87731290e-01
-1.07038331e+00 -1.76946104e+00 -3.05886060e-01 1.82649001e-01
2.56653368e-01 -6.50313377e-01 -4.65561360e-01 -1.54539615e-01] | [8.470608711242676, 10.694857597351074] |
4aac3cf3-1d32-4898-8592-700842e5021d | document-structure-extraction-for-forms-using | 1911.12170 | null | https://arxiv.org/abs/1911.12170v2 | https://arxiv.org/pdf/1911.12170v2.pdf | Document Structure Extraction using Prior based High Resolution Hierarchical Semantic Segmentation | Structure extraction from document images has been a long-standing research topic due to its high impact on a wide range of practical applications. In this paper, we share our findings on employing a hierarchical semantic segmentation network for this task of structure extraction. We propose a prior based deep hierarchical CNN network architecture that enables document structure extraction using very high resolution(1800 x 1000) images. We divide the document image into overlapping horizontal strips such that the network segments a strip and uses its prediction mask as prior for predicting the segmentation of the subsequent strip. We perform experiments establishing the effectiveness of our strip based network architecture through ablation methods and comparison with low-resolution variations. Further, to demonstrate our network's capabilities, we train it on only one type of documents (Forms) and achieve state-of-the-art results over other general document datasets. We introduce our new human-annotated forms dataset and show that our method significantly outperforms different segmentation baselines on this dataset in extracting hierarchical structures. Our method is currently being used in Adobe's AEM Forms for automated conversion of paper and PDF forms to modern HTML based forms. | ['Milan Aggarwal', 'Mausoom Sarkar', 'Hiresh Gupta', 'Balaji Krishnamurthy', 'Arneh Jain'] | 2019-11-27 | document-structure-extraction-using-prior | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6393_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123730647.pdf | eccv-2020-8 | ['table-detection'] | ['miscellaneous'] | [ 7.05624998e-01 2.04158053e-01 5.27308555e-03 -4.38358963e-01
-9.54088271e-01 -7.71617949e-01 6.24663472e-01 4.20514680e-02
-4.38089490e-01 4.17963535e-01 2.75635421e-01 -3.58162075e-01
-2.49440849e-01 -8.76778841e-01 -7.49068141e-01 -1.41431957e-01
2.90154010e-01 6.31623745e-01 5.91563106e-01 -2.39604171e-02
4.99957651e-01 6.94554150e-01 -1.14787877e+00 9.44190741e-01
7.45865345e-01 8.29612911e-01 1.34112403e-01 9.42597568e-01
-4.77271110e-01 6.11204863e-01 -1.09184051e+00 -4.77874547e-01
4.01426107e-01 -1.95689008e-01 -1.39545143e+00 4.21187460e-01
1.00278282e+00 -6.76164806e-01 -2.00137094e-01 8.41204345e-01
2.70543456e-01 -9.99779254e-02 6.21677339e-01 -6.82378292e-01
-6.98797107e-01 9.56168354e-01 -8.08916211e-01 -3.97913866e-02
3.97509962e-01 -3.05754870e-01 1.22454655e+00 -7.72947311e-01
9.58490789e-01 1.18567252e+00 8.82495344e-01 2.54520386e-01
-1.20687664e+00 -4.10638720e-01 1.59388468e-01 -5.25120795e-02
-1.30879819e+00 -2.91548193e-01 5.95075846e-01 -4.99602318e-01
1.30743957e+00 3.64251919e-02 5.53322196e-01 8.83169889e-01
1.26593143e-01 1.32086945e+00 9.06319857e-01 -8.32611561e-01
-4.26884033e-02 -2.75774509e-01 5.42493343e-01 8.26316297e-01
2.68510044e-01 -5.13955534e-01 -3.05281639e-01 2.55423218e-01
9.29729819e-01 -4.28013116e-01 -1.59854487e-01 -6.42901361e-02
-9.05440629e-01 6.79596364e-01 3.86621356e-01 5.67251623e-01
-8.44760016e-02 6.21041879e-02 1.19600862e-01 -1.07704200e-01
2.07142130e-01 5.82057714e-01 -3.00550461e-01 -1.10269682e-02
-1.68370140e+00 3.72358054e-01 9.27867472e-01 1.19812536e+00
4.95249897e-01 -2.26384476e-01 -2.91370541e-01 1.00576174e+00
1.85838535e-01 -2.37398043e-01 2.12458327e-01 -9.93299842e-01
7.74186611e-01 7.66987383e-01 -7.33286738e-02 -6.91049397e-01
-4.46911961e-01 -2.14581773e-01 -5.71234465e-01 1.07906070e-02
5.03936350e-01 -1.70733244e-03 -1.44550037e+00 1.08132482e+00
-1.41580924e-01 -5.55548012e-01 -2.52040535e-01 6.66423321e-01
8.96381199e-01 8.15905273e-01 -1.16494089e-01 2.90151954e-01
1.43052948e+00 -1.06562161e+00 -6.81923568e-01 -1.66125640e-01
2.88717210e-01 -1.05529559e+00 1.24838281e+00 7.72876382e-01
-1.34398997e+00 -4.82245207e-01 -1.47885323e+00 -5.79463422e-01
-5.22092104e-01 2.73509681e-01 5.80435395e-01 5.34604788e-01
-1.27401674e+00 8.25002968e-01 -7.47483611e-01 -7.37463713e-01
8.30308855e-01 5.67120492e-01 -1.57646820e-01 4.42210659e-02
-7.30600059e-01 4.38167155e-01 6.03029728e-01 -5.52100800e-02
-4.35755312e-01 -3.12282681e-01 -6.21702909e-01 2.98457712e-01
2.92991966e-01 -5.64347923e-01 1.52079129e+00 -6.86968029e-01
-1.33175325e+00 1.04296505e+00 -3.53530273e-02 -5.76410294e-01
4.83358532e-01 -4.45483863e-01 2.84159440e-03 6.12487495e-01
2.19361618e-01 1.08631325e+00 7.26598918e-01 -1.28184545e+00
-7.86005020e-01 -3.78659874e-01 1.45036489e-01 1.12449914e-01
-4.77089226e-01 1.82271361e-01 -1.14032888e+00 -8.04659426e-01
1.85095832e-01 -7.75589466e-01 1.03320107e-01 -2.92182475e-01
-9.50195193e-01 -2.17952952e-01 8.51316988e-01 -9.20098245e-01
1.22107625e+00 -1.82546711e+00 -1.78733930e-01 3.62153649e-01
1.82563052e-01 2.20779374e-01 -1.63701445e-01 1.65684104e-01
3.32841843e-01 6.53155923e-01 -5.50195873e-01 -4.41913188e-01
8.05379301e-02 -2.69760527e-02 -1.05779745e-01 -1.35052189e-01
1.32487446e-01 1.05173206e+00 -1.24010071e-01 -9.30214822e-01
-2.05405764e-02 5.13002038e-01 -4.87101197e-01 9.24169868e-02
-3.26456934e-01 -2.47114167e-01 -2.49127299e-01 5.94709873e-01
7.27333724e-01 -3.74767452e-01 3.00039947e-01 -3.18242908e-01
-1.31980002e-01 4.14520174e-01 -9.38842237e-01 1.80275857e+00
-1.67662010e-01 9.79220688e-01 2.86707968e-01 -6.14015698e-01
6.85968697e-01 8.65784809e-02 2.11168215e-01 -8.04949522e-01
3.11628506e-02 -1.06473386e-01 1.86440889e-02 -2.41962180e-01
9.04740989e-01 2.69907534e-01 -6.75837249e-02 4.11840349e-01
1.92226648e-01 -4.21763092e-01 7.41206527e-01 5.07795632e-01
1.11923373e+00 3.94975334e-01 1.32779859e-03 -2.52475828e-01
3.02021652e-01 2.41490439e-01 1.96868896e-01 9.94245768e-01
5.99555410e-02 1.04255390e+00 6.94740891e-01 -3.44788074e-01
-1.31773984e+00 -8.63010526e-01 -3.12143385e-01 1.10036767e+00
-1.05538212e-01 -5.97290039e-01 -1.34732664e+00 -8.14278185e-01
-2.22898751e-01 5.49110293e-01 -5.36501110e-01 6.86274946e-01
-9.95425224e-01 -8.20476830e-01 6.71732545e-01 8.14914703e-01
9.48955059e-01 -1.45606279e+00 -3.80022526e-01 2.55049884e-01
1.25393808e-01 -1.62404346e+00 -3.51365775e-01 3.55933368e-01
-7.88140059e-01 -1.05082214e+00 -7.26527154e-01 -1.04251695e+00
5.82912445e-01 9.32264552e-02 1.18198919e+00 1.24882840e-01
-2.99867958e-01 3.51159275e-01 -2.16205373e-01 -4.05188859e-01
-4.05909717e-01 7.65608728e-01 -7.25179493e-01 -4.03681070e-01
4.40185636e-01 -2.95514047e-01 -5.40245473e-01 -1.04122669e-01
-1.21413493e+00 3.07029337e-01 7.18821406e-01 3.76765132e-01
6.16040349e-01 1.96555078e-01 1.09342277e-01 -1.27992392e+00
1.05257559e+00 3.70435715e-01 -6.98499560e-01 2.09369317e-01
-4.61190969e-01 -2.18250584e-02 5.93816876e-01 1.96963206e-01
-1.31405485e+00 1.10558152e-01 -1.29879162e-01 4.04003233e-01
-4.34467643e-01 3.83429706e-01 -2.96295315e-01 2.37291515e-01
4.73253816e-01 -1.37962952e-01 -4.24380004e-01 -6.74901783e-01
3.60467315e-01 8.77255976e-01 8.46218824e-01 -6.49248242e-01
6.44581854e-01 5.19890547e-01 -1.94008276e-01 -9.35388505e-01
-1.13320935e+00 -3.91671479e-01 -1.17567682e+00 2.03760847e-01
1.22234344e+00 -4.79673922e-01 -2.25054339e-01 6.12969995e-01
-1.40465403e+00 -5.47214508e-01 -5.96155226e-03 -1.92309141e-01
-2.55059302e-01 5.00077546e-01 -1.24299121e+00 -2.47658387e-01
-5.16060054e-01 -1.07395041e+00 1.32740271e+00 3.47226948e-01
-4.06024247e-01 -8.92016172e-01 -1.11991189e-01 7.55858958e-01
-4.01059119e-03 2.14974433e-01 1.18317986e+00 -5.68415642e-01
-8.36092830e-01 -5.49135245e-02 -7.70659924e-01 2.90438443e-01
-4.81272563e-02 3.84911358e-01 -9.27535534e-01 -2.95069348e-02
-5.13755918e-01 -1.55368999e-01 1.23917568e+00 4.60067689e-01
1.42168188e+00 -3.06725264e-01 -3.23918253e-01 7.21017778e-01
1.47164929e+00 2.99522400e-01 9.28937495e-01 9.21390653e-01
1.10185623e+00 6.82119548e-01 2.55533010e-01 4.72371690e-02
2.02950105e-01 4.72381413e-01 9.42113549e-02 -3.71912986e-01
-2.77413070e-01 -9.30799171e-02 3.22302021e-02 6.17161095e-01
1.44529313e-01 -5.08760452e-01 -1.09536028e+00 7.27642775e-01
-1.58110964e+00 -5.89863718e-01 -2.53701568e-01 1.72448492e+00
1.09165275e+00 6.82268023e-01 5.34356125e-02 2.28971452e-01
7.12561965e-01 7.15863705e-02 -2.35669255e-01 -7.57356703e-01
-1.79200456e-01 7.36428618e-01 4.10120815e-01 4.45825934e-01
-1.49141037e+00 1.37613904e+00 6.52779245e+00 8.26251745e-01
-6.89288616e-01 -3.88566464e-01 7.40234077e-01 1.77997798e-02
-2.24305317e-02 -9.70762968e-02 -1.20252061e+00 1.20874457e-01
8.22934031e-01 3.26538384e-01 1.53350115e-01 5.01419663e-01
1.12249784e-01 -2.36514881e-01 -1.21662140e+00 6.34027421e-01
1.77309643e-02 -1.51709366e+00 5.29694438e-01 2.24311009e-01
9.11315024e-01 -2.23923758e-01 -8.53989646e-02 4.99465354e-02
4.80398387e-01 -1.40934932e+00 6.64248347e-01 2.07119644e-01
8.23362112e-01 -8.45174968e-01 6.50453687e-01 5.87985665e-02
-9.82822418e-01 2.52547950e-01 -3.00146610e-01 2.63078809e-01
5.37821949e-02 4.98615116e-01 -1.12175763e+00 2.77864307e-01
7.79192567e-01 4.18079972e-01 -9.97455239e-01 9.19220507e-01
-2.78059363e-01 7.26253748e-01 -4.20686245e-01 1.66499585e-01
6.36435568e-01 -2.82116145e-01 8.11955705e-02 1.76182044e+00
8.32128748e-02 1.01591111e-03 1.66463017e-01 1.03853285e+00
-5.26361406e-01 1.28480449e-01 -3.71907800e-01 -1.80660859e-01
3.95823419e-01 1.35943592e+00 -1.36648738e+00 -6.74425483e-01
-4.40626889e-01 9.89731014e-01 2.48468727e-01 2.99875855e-01
-3.71668100e-01 -1.05969858e+00 -1.54994845e-01 3.15561354e-01
6.73163652e-01 -3.20492476e-01 -6.72418952e-01 -8.41173053e-01
1.89818829e-01 -1.01373112e+00 4.09016192e-01 -9.40596402e-01
-7.46710896e-01 5.71478903e-01 -9.29350257e-02 -7.83779681e-01
-1.39234200e-01 -8.18818808e-01 -4.49342459e-01 6.07449412e-01
-1.44806015e+00 -1.41417801e+00 -4.82081443e-01 3.21212858e-01
8.40608180e-01 -1.49246693e-01 6.74669981e-01 1.76407509e-02
-6.84396684e-01 5.19739747e-01 9.12871435e-02 8.49208295e-01
6.13964856e-01 -1.72593701e+00 8.43141675e-01 1.00357938e+00
4.56265599e-01 8.40627670e-01 2.27617055e-01 -6.36832595e-01
-7.63941944e-01 -1.03536034e+00 5.72792292e-01 -3.96436423e-01
3.93194079e-01 -5.64861894e-01 -1.02291572e+00 8.60165417e-01
6.95344031e-01 -6.04282320e-01 4.88241374e-01 -1.40786618e-01
-3.40712279e-01 1.00700356e-01 -1.02679908e+00 7.13523746e-01
8.97494256e-01 -2.80275106e-01 -8.63029778e-01 2.60943979e-01
6.34199858e-01 -4.28778529e-01 -8.35950494e-01 3.27852368e-01
6.97260559e-01 -1.33996010e+00 7.14398742e-01 -2.88921773e-01
8.69647145e-01 4.78524454e-02 9.78706703e-02 -7.05956221e-01
-2.56746650e-01 -7.41740227e-01 -9.80807319e-02 1.43131411e+00
4.84278828e-01 -4.42413759e-04 1.16640222e+00 4.45716798e-01
-1.95105568e-01 -7.67381072e-01 -3.86402667e-01 -5.43953061e-01
4.96489316e-01 -2.74748951e-01 4.80930269e-01 5.86240888e-01
-4.09513831e-01 5.94791710e-01 2.59942502e-01 -1.43256366e-01
5.11368752e-01 3.10126811e-01 7.94053555e-01 -1.18991411e+00
-2.41960660e-01 -7.21343219e-01 1.62513126e-02 -1.42364788e+00
4.01737206e-02 -9.21026468e-01 -2.25474965e-02 -2.25109935e+00
3.87159497e-01 2.09997520e-01 -6.57355553e-03 4.59118664e-01
7.93260336e-02 5.57375312e-01 3.54846001e-01 2.90867448e-01
-8.28444958e-01 -8.32416117e-02 1.16769743e+00 -4.78459060e-01
-3.04002970e-01 -2.95464963e-01 -8.32525492e-01 9.35214818e-01
7.10685909e-01 -2.13647991e-01 -1.13026358e-01 -6.52041435e-01
1.81543410e-01 -3.53794992e-01 -8.47701821e-03 -9.84874666e-01
-7.00121652e-03 3.51939857e-01 1.04420877e+00 -1.20782983e+00
2.12060404e-03 -4.65011120e-01 -7.00362563e-01 -1.11666275e-02
-5.35703719e-01 -9.14091319e-02 4.51460361e-01 -3.16206813e-02
-1.83168128e-01 -6.07385814e-01 7.44923532e-01 -4.48517263e-01
-6.18342280e-01 1.40825421e-01 -5.81131995e-01 1.71774223e-01
4.69558746e-01 -6.98683560e-01 -4.37785447e-01 -3.55652332e-01
-7.36914277e-01 1.16422690e-01 5.40157139e-01 3.81695360e-01
6.89242423e-01 -9.33075547e-01 -5.63380182e-01 -7.87315294e-02
-1.54385179e-01 4.39221740e-01 -3.67538035e-01 2.83879846e-01
-1.16123199e+00 8.03745925e-01 -7.72975683e-02 -6.77340269e-01
-1.47415006e+00 1.31671071e-01 9.52306688e-02 -7.39475310e-01
-8.52098823e-01 6.69144690e-01 2.79936790e-01 -2.20667303e-01
2.84405053e-01 -5.15707791e-01 -4.05577660e-01 1.35962605e-01
4.04266298e-01 2.57059246e-01 3.75092804e-01 -3.38555217e-01
-1.18760176e-01 8.49529922e-01 -7.59544492e-01 -4.25176561e-01
1.50563955e+00 1.82282209e-01 -2.65793622e-01 4.16811705e-02
1.33413327e+00 1.26096278e-01 -1.41640949e+00 1.83616411e-02
5.01134992e-01 -4.12717536e-02 3.65528353e-02 -9.05550957e-01
-7.94709742e-01 7.94178069e-01 3.47948283e-01 4.23684895e-01
1.00854135e+00 -1.04415365e-01 1.05759859e+00 6.41425490e-01
5.36687614e-04 -1.56290519e+00 1.95370004e-01 7.84626544e-01
9.24441457e-01 -1.07256699e+00 4.10946727e-01 -5.19580424e-01
-1.37710765e-01 1.38552666e+00 5.88519573e-01 -1.39687791e-01
2.13169709e-01 6.87876642e-01 1.29986301e-01 -4.39989060e-01
-3.21904689e-01 -1.85852766e-01 4.41228151e-01 6.09676778e-01
8.56327832e-01 -2.73267418e-01 -1.86540082e-01 5.12062907e-01
-7.02937305e-01 -2.31101543e-01 8.10657203e-01 1.27432156e+00
-4.41025108e-01 -1.15678811e+00 -4.57646251e-01 6.16212308e-01
-1.03822005e+00 -3.73965561e-01 -1.10661960e+00 1.16997814e+00
-1.64814696e-01 6.67231560e-01 2.80017525e-01 1.67679682e-01
2.99913883e-01 2.52059221e-01 7.83431947e-01 -9.26749647e-01
-7.30664313e-01 4.75743294e-01 2.23854035e-01 -4.52512473e-01
-3.80181670e-01 -6.54679656e-01 -1.46558762e+00 -1.42443534e-02
-9.59146619e-02 -1.97803110e-01 6.12873256e-01 8.93339574e-01
3.89119834e-01 8.60121548e-01 -2.00179592e-02 -8.44817996e-01
-2.43670657e-01 -9.91419137e-01 -2.86737591e-01 5.99037290e-01
7.86727816e-02 -4.88068387e-02 1.05599016e-01 5.22071958e-01] | [11.689555168151855, 2.6828339099884033] |
d7756fc6-6fac-467e-ab06-57e1241dca15 | personalized-pricing-with-invalid | 2302.12670 | null | https://arxiv.org/abs/2302.12670v1 | https://arxiv.org/pdf/2302.12670v1.pdf | Personalized Pricing with Invalid Instrumental Variables: Identification, Estimation, and Policy Learning | Pricing based on individual customer characteristics is widely used to maximize sellers' revenues. This work studies offline personalized pricing under endogeneity using an instrumental variable approach. Standard instrumental variable methods in causal inference/econometrics either focus on a discrete treatment space or require the exclusion restriction of instruments from having a direct effect on the outcome, which limits their applicability in personalized pricing. In this paper, we propose a new policy learning method for Personalized pRicing using Invalid iNsTrumental variables (PRINT) for continuous treatment that allow direct effects on the outcome. Specifically, relying on the structural models of revenue and price, we establish the identifiability condition of an optimal pricing strategy under endogeneity with the help of invalid instrumental variables. Based on this new identification, which leads to solving conditional moment restrictions with generalized residual functions, we construct an adversarial min-max estimator and learn an optimal pricing strategy. Furthermore, we establish an asymptotic regret bound to find an optimal pricing strategy. Finally, we demonstrate the effectiveness of the proposed method via extensive simulation studies as well as a real data application from an US online auto loan company. | ['Lin Lin', 'Cong Shi', 'Zhengling Qi', 'Rui Miao'] | 2023-02-24 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [-1.07937925e-01 -5.32417744e-02 -8.33072126e-01 -1.98559940e-01
-6.85905814e-01 -6.14553750e-01 -1.16705030e-01 -1.28294110e-01
-3.12617958e-01 1.02667284e+00 1.09682018e-02 -6.49641752e-01
-6.15014911e-01 -8.74006033e-01 -8.93792570e-01 -8.96865070e-01
-6.39212728e-02 3.41732562e-01 -9.21859384e-01 2.09881544e-01
4.21234697e-01 7.21288323e-02 -6.80693865e-01 -6.23215914e-01
1.38930774e+00 1.02694261e+00 1.77077562e-01 3.81374620e-02
9.74116400e-02 3.63862336e-01 -2.30768591e-01 -4.95604306e-01
5.29965103e-01 -2.53268033e-01 -3.28874081e-01 4.12436724e-02
-1.89600557e-01 -7.85138130e-01 -1.58749312e-01 1.14676976e+00
5.76239347e-01 3.59987803e-02 6.11783743e-01 -1.40797055e+00
-7.95871496e-01 8.97269607e-01 -8.62260044e-01 4.40870784e-02
1.34125212e-02 -5.14471792e-02 1.37026870e+00 -4.79962200e-01
2.98587412e-01 1.31733298e+00 4.28180605e-01 3.08000147e-01
-1.58057034e+00 -9.95953977e-01 3.46126854e-01 -3.26748006e-02
-1.06129920e+00 -1.11954048e-01 8.11913669e-01 -3.60564947e-01
1.63012818e-01 -1.42153101e-02 3.50149840e-01 1.09874463e+00
3.72841358e-01 9.03892756e-01 1.15169024e+00 -4.17892039e-01
3.67351651e-01 2.29377955e-01 -4.91943359e-02 2.79469728e-01
5.30210078e-01 4.31114733e-01 4.00583167e-03 -5.06322026e-01
1.00502169e+00 2.85168141e-01 -2.34563768e-01 -3.38647991e-01
-7.66931355e-01 1.29822135e+00 6.79908367e-03 -3.56894642e-01
-8.58494997e-01 2.29610682e-01 3.27874035e-01 2.25986794e-01
3.51891428e-01 1.56351164e-01 -4.58091766e-01 7.97477514e-02
-4.65171039e-01 3.53257209e-01 8.59200537e-01 1.01261353e+00
3.63581330e-01 2.93090761e-01 -3.93443793e-01 6.58497453e-01
2.82486200e-01 1.07302749e+00 2.06591532e-01 -1.00805008e+00
6.11648917e-01 -1.30575998e-02 6.40313387e-01 -9.33684886e-01
-1.22990377e-01 -5.30711770e-01 -6.12022877e-01 -2.38065809e-01
3.90719265e-01 -6.61518812e-01 -2.31596664e-01 2.01083016e+00
1.27669573e-01 5.99260926e-01 -6.76558018e-02 9.11545753e-01
5.16624050e-03 5.13193190e-01 2.89583117e-01 -8.42118144e-01
1.18317509e+00 -3.72089833e-01 -8.45906019e-01 1.77170113e-01
6.01245701e-01 -3.95558149e-01 9.48280215e-01 4.08877522e-01
-8.93637180e-01 9.04752165e-02 -5.69271088e-01 4.71854985e-01
2.37127393e-01 -5.89384791e-03 9.50399458e-01 7.19791293e-01
-2.78347909e-01 5.08595526e-01 -4.06716764e-01 9.56037194e-02
2.47140303e-01 4.67380971e-01 1.50110751e-01 1.71619534e-01
-1.37297940e+00 3.73974502e-01 2.79013570e-02 -1.00416280e-02
-9.03949618e-01 -1.02990997e+00 -5.76250970e-01 4.18099493e-01
9.22677398e-01 -6.99412584e-01 1.32862914e+00 -1.01343453e+00
-1.58131611e+00 1.26306951e-01 -2.04116497e-02 -7.00770497e-01
7.80559301e-01 -7.77322352e-02 -1.33191764e-01 -1.46935850e-01
5.56338787e-01 -1.97888553e-01 8.07774484e-01 -1.13055813e+00
-8.25388074e-01 -3.87898862e-01 3.28654051e-01 9.17430874e-03
-2.60515183e-01 -3.20232749e-01 -1.37814268e-01 -8.10713828e-01
-2.34083906e-01 -8.50218892e-01 -3.04588288e-01 -6.16648197e-01
-6.24570787e-01 -4.00280915e-02 2.19756842e-01 -8.58963907e-01
1.21401393e+00 -2.12482405e+00 -2.36694124e-02 5.06061316e-01
-1.75953969e-01 -5.94370544e-01 4.38008681e-02 3.89021933e-01
-1.28827557e-01 3.23792368e-01 -1.43928856e-01 -2.23094121e-01
5.01891673e-01 2.59196758e-01 -6.98271453e-01 5.73916793e-01
-3.00158322e-01 7.02870965e-01 -6.78034186e-01 -3.00875396e-01
4.67477292e-02 -1.70465097e-01 -6.49614811e-01 -1.12870531e-02
5.88324219e-02 2.91604191e-01 -8.17140341e-01 6.46059752e-01
9.19137478e-01 -4.93146386e-03 4.20673937e-01 1.42695427e-01
-1.85873434e-01 -5.23361936e-02 -1.32318544e+00 1.07434177e+00
-7.79916406e-01 -2.02607006e-01 1.71519220e-01 -1.41400063e+00
4.70276684e-01 3.27311277e-01 7.26794243e-01 -8.67785335e-01
2.87920892e-01 2.08693206e-01 -2.00807229e-01 -4.65614647e-01
1.39888287e-01 -6.19810641e-01 -4.39025342e-01 5.46441674e-01
-1.46651074e-01 6.54814959e-01 6.27485244e-03 7.05840737e-02
5.07305741e-01 -6.03690036e-02 3.04711074e-01 -3.45821023e-01
3.22651774e-01 -4.45999950e-01 1.27382588e+00 1.12244296e+00
-1.64490938e-02 -3.14970687e-02 9.55137968e-01 1.83318928e-01
-9.65916038e-01 -8.15697372e-01 -2.41340607e-01 8.58794630e-01
2.28495911e-01 3.65986764e-01 -4.41127270e-01 -6.25734329e-01
7.25664377e-01 1.03775549e+00 -7.17906654e-01 1.30530144e-03
-1.40342221e-01 -8.52569878e-01 1.97539330e-01 4.60137486e-01
3.88714612e-01 -6.79077804e-01 1.69275310e-02 3.51536244e-01
-1.87434196e-01 -7.17196226e-01 -6.36751413e-01 -4.75742780e-02
-7.62377262e-01 -9.46573138e-01 -6.46998703e-01 -3.24413896e-01
5.21521509e-01 3.07133287e-01 5.27833045e-01 -4.49573189e-01
1.21811010e-01 5.11911154e-01 8.07638317e-02 -5.66541076e-01
2.54353276e-03 -4.63064462e-02 2.56633341e-01 3.66655946e-01
5.67623563e-02 -4.17105138e-01 -7.81661272e-01 2.16507092e-01
-6.57530487e-01 -3.78565580e-01 6.12808883e-01 9.94862735e-01
6.32039964e-01 2.51855046e-01 1.27612233e+00 -1.21786511e+00
8.55098724e-01 -9.89472568e-01 -1.27832890e+00 3.67180884e-01
-1.03070962e+00 1.29388481e-01 6.65883243e-01 -4.57978278e-01
-1.42542195e+00 -3.03609341e-01 3.25369537e-01 -5.63703656e-01
3.65091890e-01 8.94686341e-01 -6.16596639e-01 2.68670142e-01
-3.02850902e-02 1.14471838e-01 8.98876265e-02 -5.49676836e-01
3.37207824e-01 4.58787292e-01 2.87061810e-01 -9.10930037e-01
8.19585383e-01 5.58931947e-01 2.55749404e-01 -2.43721113e-01
-5.59774756e-01 -2.60310382e-01 -9.62017626e-02 1.83001533e-01
4.87219691e-01 -1.09417081e+00 -1.52993786e+00 1.76188722e-02
-6.00047350e-01 -8.81178211e-03 -1.34148851e-01 9.27167654e-01
-8.33206236e-01 3.87122035e-01 -5.48969209e-01 -1.18698311e+00
-1.83443338e-01 -8.78478467e-01 5.61619520e-01 2.18500614e-01
1.74550787e-01 -1.30749702e+00 -2.08159283e-01 2.10425079e-01
8.81797671e-02 5.26950769e-02 9.54610467e-01 -4.09462094e-01
-6.51400328e-01 -2.56050944e-01 -8.69809836e-02 9.37944055e-02
9.99379084e-02 -2.41317451e-01 -3.78505349e-01 -3.62878293e-01
1.04861796e-01 2.41287008e-01 5.31203508e-01 9.85902846e-01
1.21914732e+00 -7.61728525e-01 -4.28448886e-01 6.40718877e-01
1.65476477e+00 4.58514482e-01 4.15510982e-01 5.30994952e-01
4.78770673e-01 4.77825344e-01 1.05498266e+00 1.20193338e+00
4.49922442e-01 3.33167553e-01 4.79223281e-01 9.20482650e-02
1.20998263e+00 -5.36560237e-01 1.91381693e-01 4.66995686e-02
-5.42446636e-02 -2.53243238e-01 -2.47469887e-01 4.68078345e-01
-2.07356358e+00 -1.13541996e+00 7.54512940e-03 2.74018145e+00
6.75255001e-01 -8.98241550e-02 2.58503556e-01 -2.14384541e-01
8.78673792e-01 -4.68078971e-01 -8.80562842e-01 -5.08724749e-01
-1.26923665e-01 1.33936079e-02 1.42341626e+00 4.10161883e-01
-8.61818075e-01 5.63893437e-01 5.61417198e+00 9.75604892e-01
-1.05839348e+00 9.40635577e-02 6.76020980e-01 -6.88089207e-02
-7.44978666e-01 3.77520621e-01 -8.23464036e-01 8.05540860e-01
6.52103007e-01 -7.33631432e-01 6.16974533e-01 8.18190336e-01
1.06098330e+00 -6.38296902e-02 -8.09856474e-01 7.80843377e-01
-5.50593555e-01 -9.49253261e-01 -3.45953315e-01 6.82333767e-01
8.00767064e-01 -7.28144526e-01 3.51405203e-01 4.12145495e-01
4.35468525e-01 -6.21656001e-01 6.74464345e-01 6.09138250e-01
5.47666252e-01 -1.16907012e+00 6.79768741e-01 4.17106301e-01
-5.85077941e-01 -7.32042074e-01 -3.68298262e-01 -1.31037487e-02
3.48510534e-01 5.36795974e-01 -5.30525565e-01 5.84455073e-01
9.29827839e-02 3.64270061e-01 2.67683655e-01 1.06986773e+00
-3.72170091e-01 9.03028786e-01 -1.30176485e-01 3.35321426e-01
-2.66952417e-03 -6.84610069e-01 3.96049470e-01 7.01193035e-01
5.47556758e-01 1.83018655e-01 1.05197847e-01 1.03265619e+00
-2.91626215e-01 5.02386391e-01 -4.39623088e-01 -2.50853211e-01
6.98274612e-01 1.00680232e+00 -3.13682675e-01 -1.38358638e-01
-6.64697170e-01 6.58790112e-01 -1.62221372e-01 6.01261139e-01
-1.15519416e+00 -1.03227727e-01 5.58974981e-01 -3.88919786e-02
3.35833699e-01 1.61283538e-01 -4.88568455e-01 -1.11082113e+00
-4.93149422e-02 -6.69606388e-01 5.40311515e-01 -1.73386976e-01
-1.37232292e+00 -7.94677258e-01 6.75692260e-02 -1.12612855e+00
-3.12974483e-01 -7.73840770e-02 -7.30985880e-01 1.13074541e+00
-1.63055229e+00 -7.84406424e-01 2.70660222e-01 5.39596558e-01
1.53747365e-01 -6.08811490e-02 5.06240070e-01 4.04773444e-01
-8.69380295e-01 8.19833696e-01 7.63320684e-01 -2.33293176e-02
6.57249629e-01 -1.30144465e+00 -3.82273763e-01 6.06921315e-01
-5.13271511e-01 8.20270717e-01 6.63912296e-01 -8.39396179e-01
-1.71406198e+00 -8.23328853e-01 4.91304219e-01 3.36678773e-01
9.11463916e-01 -1.45666286e-01 -5.97229660e-01 9.10179615e-01
2.99777500e-02 -3.68221462e-01 6.70958221e-01 4.59926248e-01
-5.70113361e-02 -3.68178695e-01 -1.21608019e+00 5.48678041e-01
7.39251733e-01 -3.09876978e-01 -2.86605675e-02 3.71282846e-01
8.16453993e-01 -5.20013869e-02 -9.20475304e-01 2.54375607e-01
5.69291770e-01 -3.14221740e-01 7.71023273e-01 -6.32928550e-01
2.94512510e-01 2.00117901e-01 -8.11054930e-03 -1.19848859e+00
-5.21435320e-01 -7.52935827e-01 2.30117425e-01 1.55335057e+00
2.35314816e-01 -1.17391372e+00 4.81156826e-01 7.42231548e-01
3.66154909e-01 -3.37506235e-01 -8.71362805e-01 -9.10138845e-01
2.73856908e-01 -2.61165351e-01 6.81524754e-01 1.06980276e+00
7.44383829e-03 1.84073284e-01 -7.53639936e-01 3.52664769e-01
1.05879521e+00 5.02229214e-01 6.71724916e-01 -1.02963388e+00
-6.19287193e-01 -1.46415949e-01 2.57334143e-01 -8.57801199e-01
7.70369947e-01 -4.84521657e-01 -1.20472044e-01 -1.00716650e+00
5.23569822e-01 -7.36837745e-01 -5.21062791e-01 2.12441921e-01
-6.71636239e-02 -5.03323019e-01 5.78151941e-02 1.48085371e-01
-1.39439896e-01 6.47217929e-01 1.15183711e+00 -1.33855462e-01
-3.44090492e-01 4.22400683e-01 -1.12809098e+00 2.58547217e-01
8.27433288e-01 -3.69596571e-01 -6.67281747e-01 6.30401075e-02
2.82047004e-01 8.21348488e-01 4.13399726e-01 1.07167520e-01
-3.04192245e-01 -6.96351647e-01 -6.36858568e-02 -2.84800678e-01
-1.32909849e-01 -9.13611591e-01 2.12176636e-01 3.57917041e-01
-2.73825496e-01 -2.96691626e-01 -1.69490054e-02 1.01860797e+00
7.91686848e-02 -5.36435544e-01 3.56378555e-01 2.59340014e-02
-4.15265739e-01 5.37291467e-01 -3.15808505e-01 -5.47680371e-02
1.09575641e+00 2.25811392e-01 -5.24812639e-02 -7.96439111e-01
-8.64237905e-01 6.52725995e-01 9.61689726e-02 2.29572281e-01
2.92467207e-01 -1.32865560e+00 -5.77598274e-01 -1.91120535e-01
-2.55896062e-01 -5.58565676e-01 4.67996836e-01 1.03823555e+00
1.95962116e-01 7.70417035e-01 1.91722170e-01 -2.13606536e-01
-8.93818498e-01 8.54914129e-01 2.18040779e-01 -6.51356578e-02
-1.72721058e-01 2.12213069e-01 5.86802602e-01 1.22538395e-02
-4.49394695e-02 -3.92508917e-02 -1.22055702e-01 3.24185282e-01
-4.46095131e-02 4.24652666e-01 -4.42846149e-01 -2.83076316e-01
4.21291664e-02 1.47540554e-01 4.16850261e-02 -3.04065585e-01
1.21973717e+00 -5.97714305e-01 2.31212545e-02 3.52843225e-01
1.05901694e+00 2.75969028e-01 -1.30070150e+00 -3.46570075e-01
-2.93890964e-02 -7.86835968e-01 5.42757548e-02 -4.28263426e-01
-1.16393077e+00 5.13111949e-01 6.16080463e-01 3.58626515e-01
9.56375360e-01 -4.80861515e-01 6.42470360e-01 -1.25563011e-01
4.09666926e-01 -1.26628566e+00 -5.01866639e-01 -1.47191463e-02
4.74936426e-01 -1.28432262e+00 -2.04199180e-01 -3.72210681e-01
-6.93514168e-01 6.66028917e-01 1.63549140e-01 -1.28050327e-01
7.33042121e-01 -1.00178793e-01 -2.85443246e-01 2.52609640e-01
-5.24887383e-01 -2.45331079e-02 -2.15014145e-01 2.46281564e-01
1.66429594e-01 6.77916348e-01 -1.17486763e+00 1.05709600e+00
1.79072440e-01 1.15605548e-01 6.80445611e-01 6.89755857e-01
-2.48563793e-02 -1.17645979e+00 -4.80481207e-01 3.90403092e-01
-8.89018595e-01 -1.62273884e-01 2.37802863e-01 9.23004150e-01
-1.93438068e-01 8.38083088e-01 -2.34443843e-02 2.55494416e-01
3.97121191e-01 -1.30321994e-01 3.77595067e-01 -2.57199705e-01
4.65462580e-02 6.03274822e-01 -1.94268346e-01 -1.81499854e-01
-7.85794035e-02 -9.15717125e-01 -1.23692167e+00 -6.53079569e-01
-7.47638881e-01 3.48653585e-01 5.75313091e-01 9.59107816e-01
2.60155946e-01 2.84167111e-01 1.38735044e+00 -2.22318500e-01
-1.20597494e+00 -7.25018084e-01 -1.27856934e+00 3.24569017e-01
-2.84631620e-03 -9.24690008e-01 -5.37949264e-01 -3.34444493e-01] | [8.282209396362305, 5.112167835235596] |
540686f4-6f88-45a0-af5a-69cb079f48a5 | dsgn-deep-stereo-geometry-network-for-3d | 2001.03398 | null | https://arxiv.org/abs/2001.03398v3 | https://arxiv.org/pdf/2001.03398v3.pdf | DSGN: Deep Stereo Geometry Network for 3D Object Detection | Most state-of-the-art 3D object detectors heavily rely on LiDAR sensors because there is a large performance gap between image-based and LiDAR-based methods. It is caused by the way to form representation for the prediction in 3D scenarios. Our method, called Deep Stereo Geometry Network (DSGN), significantly reduces this gap by detecting 3D objects on a differentiable volumetric representation -- 3D geometric volume, which effectively encodes 3D geometric structure for 3D regular space. With this representation, we learn depth information and semantic cues simultaneously. For the first time, we provide a simple and effective one-stage stereo-based 3D detection pipeline that jointly estimates the depth and detects 3D objects in an end-to-end learning manner. Our approach outperforms previous stereo-based 3D detectors (about 10 higher in terms of AP) and even achieves comparable performance with several LiDAR-based methods on the KITTI 3D object detection leaderboard. Our code is publicly available at https://github.com/chenyilun95/DSGN. | ['Yilun Chen', 'Shu Liu', 'Xiaoyong Shen', 'Jiaya Jia'] | 2020-01-10 | dsgn-deep-stereo-geometry-network-for-3d-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_DSGN_Deep_Stereo_Geometry_Network_for_3D_Object_Detection_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_DSGN_Deep_Stereo_Geometry_Network_for_3D_Object_Detection_CVPR_2020_paper.pdf | cvpr-2020-6 | ['vehicle-pose-estimation', '3d-object-detection-from-stereo-images'] | ['computer-vision', 'computer-vision'] | [-1.73001096e-01 -5.49577586e-02 3.98595759e-04 -3.37213606e-01
-8.58819187e-01 -5.60541034e-01 5.34928977e-01 5.21224104e-02
-2.12213621e-01 -1.55832738e-01 -1.35667995e-01 -3.38914901e-01
3.29777628e-01 -8.87798488e-01 -8.90090406e-01 -9.63236690e-02
7.01964572e-02 9.42416012e-01 8.39837372e-01 1.06759496e-01
4.98341441e-01 9.95472968e-01 -1.74710011e+00 -3.24421301e-02
4.79139417e-01 1.29998279e+00 5.90697348e-01 5.99392831e-01
-3.48296404e-01 2.57675529e-01 -1.24179181e-02 9.47455689e-02
8.69427145e-01 2.77998567e-01 -2.88039595e-01 1.32074714e-01
9.51758802e-01 -7.87065625e-01 -5.50453484e-01 8.40269387e-01
6.04892731e-01 -2.36124977e-01 8.60302329e-01 -1.07520473e+00
-3.16349030e-01 -5.37072644e-02 -9.45742726e-01 1.50317907e-01
4.39369828e-01 3.68003845e-01 8.67524445e-01 -1.44318950e+00
5.17996967e-01 1.89009571e+00 8.25246513e-01 4.33111280e-01
-1.03767872e+00 -6.76975608e-01 6.15883656e-02 -2.16810256e-01
-1.51284766e+00 -1.58973932e-01 8.09922218e-01 -6.56672418e-01
1.30817389e+00 -2.33163491e-01 7.17210829e-01 7.02878237e-01
-3.98655534e-02 1.07181883e+00 9.34913933e-01 -2.14053586e-01
-6.05413050e-04 -1.17063142e-01 -8.33157822e-02 9.48240638e-01
3.79222721e-01 4.96277332e-01 -6.64055765e-01 8.06737393e-02
1.15264928e+00 2.11425871e-01 1.37638345e-01 -1.01392663e+00
-9.78388011e-01 8.13100100e-01 8.27205360e-01 -3.54986191e-01
-1.50628388e-01 3.42797160e-01 2.44929358e-01 -1.13627747e-01
8.58221710e-01 -5.91989025e-04 -3.54214191e-01 1.67223081e-01
-6.78388894e-01 5.33356130e-01 3.32593739e-01 1.13461006e+00
9.57132578e-01 -1.22302890e-01 2.50571128e-02 5.45008659e-01
8.66736531e-01 1.10099268e+00 -2.35234380e-01 -1.08678937e+00
6.26479268e-01 1.01208520e+00 4.65494767e-02 -6.29245877e-01
-4.48833257e-01 -2.64674872e-01 -2.54121065e-01 6.20989203e-01
1.95034549e-01 3.94085646e-01 -1.28450119e+00 1.01988983e+00
7.77613461e-01 8.69040638e-02 -3.40916932e-01 1.25379193e+00
1.20312107e+00 4.04441208e-01 -2.68196732e-01 6.07730687e-01
1.12930465e+00 -6.46070838e-01 2.79893488e-01 -6.13171220e-01
6.18948221e-01 -7.03787386e-01 9.16959524e-01 6.49692118e-02
-1.08193088e+00 -5.79994440e-01 -1.01099944e+00 -4.97915745e-01
-2.07683846e-01 2.39117183e-02 5.51073015e-01 4.77580339e-01
-1.13260102e+00 2.76750833e-01 -8.27974200e-01 -5.35242140e-01
8.51800263e-01 2.61866897e-01 -2.20564306e-01 -3.19864452e-01
-4.87246692e-01 9.66335237e-01 4.67262417e-01 -2.73875952e-01
-1.14198315e+00 -9.86439645e-01 -1.12710452e+00 -2.17812210e-01
3.42284888e-01 -9.46115553e-01 1.47676969e+00 5.27142249e-02
-1.04169130e+00 1.61823237e+00 -1.50782704e-01 -4.81057674e-01
7.29929149e-01 -3.37576389e-01 4.03559327e-01 9.85422581e-02
3.09739798e-01 1.15693521e+00 6.14610136e-01 -1.24811399e+00
-7.62747526e-01 -8.96039665e-01 -4.86042984e-02 4.21175092e-01
2.29445532e-01 -2.45751500e-01 -6.43997490e-01 -9.64427460e-03
6.39244080e-01 -6.88032269e-01 -2.87751347e-01 8.03082287e-01
-4.42627788e-01 -4.80684996e-01 9.15082455e-01 2.34729759e-02
3.25597733e-01 -2.00969315e+00 -2.54052281e-01 -7.31454119e-02
5.11627734e-01 2.09180266e-01 -1.62582085e-01 1.96360022e-01
4.37621713e-01 -4.22219187e-02 -2.07659930e-01 -7.60987043e-01
9.72841159e-02 -1.29924566e-02 -3.65639925e-01 6.82576001e-01
4.30077642e-01 9.13304090e-01 -9.13188756e-01 -4.84220952e-01
8.26152861e-01 7.52331972e-01 -5.11599064e-01 1.48433819e-01
-4.12587851e-01 3.00754368e-01 -7.94277668e-01 1.15083933e+00
1.28634930e+00 -3.66867408e-02 -4.62791085e-01 -7.47711584e-02
-4.49456245e-01 5.74765980e-01 -9.82078671e-01 2.03089404e+00
-3.00415516e-01 6.76681221e-01 8.56298655e-02 -6.85618758e-01
1.31289482e+00 -2.88161248e-01 4.31092829e-01 -6.48593545e-01
3.04528385e-01 3.56965333e-01 -6.14288509e-01 -9.27625597e-02
4.66198027e-01 5.35575785e-02 -3.32229249e-02 6.93839565e-02
-8.25641453e-02 -9.26165402e-01 -1.53393790e-01 1.80183232e-01
1.10873914e+00 5.18171370e-01 1.87556863e-01 -1.37686521e-01
2.20590174e-01 2.82203317e-01 4.95064765e-01 8.56448412e-01
-3.11533064e-01 8.82749498e-01 1.55818224e-01 -4.37479705e-01
-1.24775374e+00 -1.42895627e+00 -4.02220815e-01 4.15452451e-01
5.49635828e-01 -3.63231719e-01 -1.80038318e-01 -4.72540885e-01
8.84891450e-01 3.93447250e-01 -3.23781103e-01 1.58052519e-01
-4.28833812e-01 8.04656651e-03 3.66677672e-01 7.06796944e-01
4.46330130e-01 -5.62849224e-01 -9.45383370e-01 -6.73385663e-03
4.16200370e-01 -1.27211416e+00 -1.25384286e-01 2.99972326e-01
-1.33318222e+00 -1.03319955e+00 -4.25864458e-01 -4.46601242e-01
3.80994439e-01 1.10096812e+00 1.13862371e+00 -1.67613909e-01
-6.33505762e-01 4.83967036e-01 -1.56686574e-01 -8.79453063e-01
-9.08820033e-02 2.74618436e-02 8.53435844e-02 -7.52398968e-01
7.46325135e-01 -4.35095519e-01 -7.67733335e-01 1.90693125e-01
-3.79093647e-01 2.01552242e-01 5.17700970e-01 1.65117383e-01
9.41489458e-01 -4.82346445e-01 -2.92090744e-01 -3.48141342e-01
-6.76583350e-02 -2.88734585e-01 -1.23636556e+00 -3.83091420e-01
-4.52859968e-01 8.41647293e-03 -8.37590247e-02 6.83082640e-02
-6.86275125e-01 6.35600507e-01 -6.30800426e-02 -1.02218330e+00
-3.28689367e-01 -1.00438520e-01 3.68100815e-02 -2.78277457e-01
6.95736051e-01 2.19539758e-02 -4.75567169e-02 -8.20459664e-01
3.81634176e-01 6.22481406e-01 3.77522767e-01 -3.13369811e-01
1.17834866e+00 9.12261546e-01 4.39622581e-01 -7.62090147e-01
-1.08064795e+00 -9.51484382e-01 -1.01040089e+00 -4.19999570e-01
7.00712144e-01 -1.50585294e+00 -7.39176929e-01 5.04899681e-01
-1.48347437e+00 -2.87028968e-01 -1.51149020e-01 4.37868208e-01
-5.60851395e-01 2.71012753e-01 -4.74404842e-01 -1.20022428e+00
-3.12191784e-01 -9.98345852e-01 1.91393876e+00 5.12213632e-02
2.85406023e-01 -5.07723153e-01 -3.09661888e-02 3.78786325e-01
1.40080480e-02 4.42078203e-01 3.63063455e-01 1.07077474e-03
-1.36762977e+00 -2.96946436e-01 -8.01863372e-01 5.46367653e-02
-2.61188358e-01 -2.10407615e-01 -1.24475336e+00 -9.43520516e-02
-3.00819874e-01 -3.14678133e-01 1.27772880e+00 5.45213044e-01
1.08320248e+00 6.35883868e-01 -5.81117630e-01 7.44178474e-01
1.50398278e+00 -2.29519650e-01 2.08376005e-01 2.59383887e-01
9.11374331e-01 3.70267034e-01 8.12181592e-01 5.85585415e-01
6.22344613e-01 6.33552194e-01 1.02103782e+00 -1.11459352e-01
-3.48931342e-01 -6.29286349e-01 2.54006207e-01 1.00854352e-01
-4.70789336e-03 -5.26870117e-02 -1.18612194e+00 4.51315403e-01
-1.71690309e+00 -6.22590125e-01 -5.46996951e-01 2.16692042e+00
2.91592836e-01 4.98931885e-01 2.47156739e-01 -1.50336966e-01
5.99951744e-01 6.74646646e-02 -8.80681098e-01 5.87583240e-03
-5.31501025e-02 1.36872485e-01 9.27262187e-01 4.81142491e-01
-1.19024694e+00 1.23450732e+00 5.50196791e+00 5.82753778e-01
-9.24435556e-01 -4.49034804e-03 2.35343382e-01 -3.80363107e-01
-1.37811020e-01 -5.10168150e-02 -1.31539917e+00 7.45270262e-03
4.13834929e-01 8.02965164e-02 5.05421907e-02 1.22877026e+00
4.15534556e-01 -2.13523760e-01 -1.22844207e+00 1.38271010e+00
-5.81380771e-03 -1.39693868e+00 6.46160468e-02 3.07161748e-01
4.36868817e-01 9.22645211e-01 -9.08095464e-02 1.56370133e-01
2.15751678e-01 -7.73161113e-01 1.23643637e+00 1.71890646e-01
8.54281366e-01 -5.42724848e-01 3.50635737e-01 5.92284799e-01
-1.58365011e+00 9.83075872e-02 -7.45558858e-01 -2.71654725e-01
1.99236587e-01 8.06297600e-01 -1.16635299e+00 1.44218996e-01
9.23552513e-01 8.60364437e-01 -5.54526567e-01 1.41055310e+00
-3.74467760e-01 1.80185750e-01 -7.11735964e-01 -5.01028402e-03
4.58293319e-01 4.50740941e-02 9.01688159e-01 1.07956362e+00
5.18217146e-01 -2.80264746e-02 4.81697887e-01 1.24238098e+00
-1.92243814e-01 -3.47608536e-01 -1.10719669e+00 3.42805922e-01
7.37542748e-01 1.09483397e+00 -7.06869900e-01 -9.66993868e-02
-3.25387746e-01 7.20726073e-01 3.84522378e-01 -2.51490980e-01
-6.50444984e-01 -4.71749119e-02 8.25779438e-01 5.29817402e-01
5.67577720e-01 -6.17585897e-01 -4.75864351e-01 -9.25069511e-01
2.66057342e-01 4.63138409e-02 4.69933636e-02 -1.01892447e+00
-1.31096399e+00 1.08946905e-01 -4.24902812e-02 -1.43444300e+00
5.98296113e-02 -1.02570450e+00 -2.46719658e-01 8.86988401e-01
-1.95302844e+00 -1.29423678e+00 -6.95106208e-01 3.08220178e-01
6.50268137e-01 1.85645401e-01 3.12961817e-01 8.95726755e-02
-5.84120415e-02 4.94948290e-02 -2.28552222e-01 -3.65004651e-02
4.54907864e-01 -1.26129150e+00 9.72945333e-01 6.81378186e-01
2.13764068e-02 5.59076443e-02 3.41159046e-01 -7.12830186e-01
-1.74417591e+00 -1.25812423e+00 7.01969564e-01 -9.44762826e-01
3.43225300e-01 -7.98833370e-01 -6.79335773e-01 4.13061380e-01
-4.96524602e-01 1.47273049e-01 4.59514260e-02 -1.46208778e-01
-6.98676527e-01 -3.38321626e-02 -1.20536602e+00 2.69781053e-01
1.83969319e+00 -5.67678690e-01 -6.17319703e-01 4.40382212e-01
8.90753150e-01 -8.46795261e-01 -5.05964160e-01 6.28579855e-01
4.06771004e-01 -1.23597574e+00 1.32623041e+00 -2.51390152e-02
2.82868505e-01 -7.18365908e-01 -3.20451289e-01 -8.30196440e-01
-2.56050378e-01 -1.48159444e-01 -2.64091969e-01 7.07836092e-01
1.18556559e-01 -4.11195517e-01 1.14551890e+00 2.06653699e-01
-5.36719263e-01 -5.83508134e-01 -1.12411547e+00 -1.01876605e+00
-2.55777650e-02 -8.50701511e-01 4.86194223e-01 4.52716827e-01
-6.09195113e-01 3.50271493e-01 1.59094408e-01 3.34296137e-01
1.15465128e+00 4.26299453e-01 1.05435002e+00 -1.69095516e+00
1.15319423e-01 -6.27471209e-01 -8.68194580e-01 -1.71164942e+00
-7.06030726e-02 -1.16250205e+00 2.28224725e-01 -1.70581377e+00
2.01378822e-01 -5.55624366e-01 3.45762260e-02 2.70308346e-01
2.14638039e-01 3.82783294e-01 2.61829346e-01 2.93568641e-01
-5.20642996e-01 6.94693685e-01 1.14944756e+00 -6.51517585e-02
-1.90273061e-01 -1.61199644e-01 -3.47412527e-01 8.20274055e-01
6.29715085e-01 -4.58993882e-01 -1.08695820e-01 -8.21378231e-01
-1.05164386e-01 -2.52278894e-02 8.57487261e-01 -1.08550322e+00
1.37085140e-01 4.99762893e-02 5.79527915e-01 -1.66406417e+00
7.59070814e-01 -5.98782897e-01 -4.27665025e-01 5.45837522e-01
1.39660522e-01 -3.52211297e-01 2.69438714e-01 5.82568169e-01
2.32336566e-01 1.24242619e-01 9.15883362e-01 -3.85373890e-01
-1.02339578e+00 7.27829278e-01 -3.25960107e-02 -1.17589898e-01
9.05376613e-01 -6.44935548e-01 -2.10003421e-01 8.55417252e-02
-1.02957658e-01 5.34860492e-01 7.40183353e-01 4.90479350e-01
1.06443119e+00 -1.23180950e+00 -8.04420590e-01 1.98095351e-01
4.64568883e-01 7.90424526e-01 6.40560463e-02 5.63281655e-01
-7.71850049e-01 6.88748538e-01 6.17829747e-02 -1.30584240e+00
-1.07233369e+00 2.82598078e-01 4.23006147e-01 2.35133126e-01
-9.96872365e-01 1.06657052e+00 4.49588120e-01 -7.18106568e-01
3.62112850e-01 -5.56212068e-01 4.12797004e-01 -2.91260362e-01
2.57842600e-01 3.17817509e-01 1.03885934e-01 -5.15119016e-01
-7.24244177e-01 9.46456015e-01 6.52174726e-02 1.90772548e-01
1.32779527e+00 2.74353735e-02 2.43662789e-01 3.44258875e-01
1.00283635e+00 -4.54470575e-01 -1.64271879e+00 -3.97155821e-01
-2.02608630e-01 -9.41588223e-01 5.55590272e-01 -3.86435360e-01
-8.06355000e-01 1.19759691e+00 8.50684941e-01 -1.58080503e-01
5.56405008e-01 6.62112355e-01 7.17657864e-01 4.57886517e-01
6.22658074e-01 -7.63922513e-01 1.07050791e-01 8.50792229e-01
8.16215932e-01 -1.59121561e+00 2.10956708e-01 -7.29420602e-01
1.35392342e-02 1.13876581e+00 9.12688196e-01 -4.61264133e-01
5.38148761e-01 1.62322760e-01 -1.33233383e-01 -5.40375113e-01
-4.19693202e-01 -6.58485413e-01 3.29882532e-01 8.05272758e-01
1.84988007e-01 8.32388699e-02 3.78721625e-01 -3.80751453e-02
-4.15940844e-02 -2.48036906e-01 1.39982238e-01 8.86587083e-01
-9.05366540e-01 -8.50276411e-01 -5.92857242e-01 3.08717430e-01
2.90843129e-01 2.31220834e-02 -6.38118684e-01 8.91328335e-01
1.40630215e-01 6.35265172e-01 4.12985086e-01 -2.93861866e-01
5.89923680e-01 -3.33049625e-01 8.38537037e-01 -9.98041093e-01
7.20556378e-02 6.73678368e-02 -1.05272517e-01 -9.39904392e-01
-2.73031890e-01 -8.70383382e-01 -1.35639846e+00 -3.12478453e-01
-2.69348651e-01 -6.58900023e-01 9.26062107e-01 6.68300033e-01
5.92982829e-01 -8.80125314e-02 6.43510580e-01 -1.68958104e+00
-4.56118822e-01 -6.95268810e-01 -5.22489488e-01 2.47542039e-02
2.65866339e-01 -1.08681095e+00 -2.96964228e-01 -5.13972878e-01] | [7.771086692810059, -2.6630361080169678] |
70c7f60b-a114-495f-9a95-fac2ca78f95f | implicit-discourse-relation-classification | 1603.02776 | null | http://arxiv.org/abs/1603.02776v1 | http://arxiv.org/pdf/1603.02776v1.pdf | Implicit Discourse Relation Classification via Multi-Task Neural Networks | Without discourse connectives, classifying implicit discourse relations is a
challenging task and a bottleneck for building a practical discourse parser.
Previous research usually makes use of one kind of discourse framework such as
PDTB or RST to improve the classification performance on discourse relations.
Actually, under different discourse annotation frameworks, there exist multiple
corpora which have internal connections. To exploit the combination of
different discourse corpora, we design related discourse classification tasks
specific to a corpus, and propose a novel Convolutional Neural Network embedded
multi-task learning system to synthesize these tasks by learning both unique
and shared representations for each task. The experimental results on the PDTB
implicit discourse relation classification task demonstrate that our model
achieves significant gains over baseline systems. | ['Xiaodong Zhang', 'Yang Liu', 'Zhifang Sui', 'Sujian Li'] | 2016-03-09 | null | null | null | null | ['implicit-discourse-relation-classification'] | ['natural-language-processing'] | [ 3.28918636e-01 8.29496562e-01 -6.30235493e-01 -3.45706195e-01
-7.79621601e-01 -3.99749458e-01 1.05752492e+00 1.81399584e-01
-5.25467619e-02 8.75976264e-01 7.91320384e-01 -7.04118371e-01
2.14437321e-01 -8.19283664e-01 -5.53841293e-01 -3.43431830e-01
9.38469023e-02 5.37702799e-01 3.86367440e-01 -6.18949413e-01
1.93031311e-01 -3.70058417e-01 -9.70917523e-01 1.01172686e+00
8.62594068e-01 7.91248262e-01 1.14716344e-01 7.42879570e-01
-4.65146452e-01 1.98181641e+00 -1.31005955e+00 -4.00635689e-01
-3.08174103e-01 -6.03876054e-01 -1.66264999e+00 -2.30438843e-01
6.89029843e-02 -5.66631496e-01 -3.78553718e-01 6.95305288e-01
3.06096196e-01 1.32843912e-01 7.51143336e-01 -9.20382142e-01
-8.99353921e-01 1.29121208e+00 -2.87057549e-01 3.54085624e-01
5.04618883e-01 -2.52814591e-01 1.20773411e+00 -1.63909286e-01
8.58659744e-01 1.42844188e+00 5.28025329e-01 5.51240981e-01
-1.06773639e+00 -3.89585167e-01 4.11067992e-01 5.37248135e-01
-5.90503573e-01 -4.74354804e-01 9.09358680e-01 -6.11058652e-01
1.29847515e+00 5.50162852e-01 9.50674787e-02 1.63930452e+00
-3.06819886e-01 1.00002432e+00 8.74736905e-01 -4.42331314e-01
-2.49062851e-01 -1.51095092e-01 6.98300779e-01 1.10593438e-01
-1.18271463e-01 -4.46774811e-01 -2.24176675e-01 8.74169171e-02
4.24698174e-01 -7.74196148e-01 -4.34534520e-01 3.07414323e-01
-1.04956102e+00 8.29927504e-01 6.09530151e-01 5.97526431e-01
1.30606517e-01 -2.68402547e-02 1.10266113e+00 5.29758632e-01
7.89168775e-01 7.23190069e-01 -1.50820732e-01 -3.06384176e-01
-2.91910291e-01 4.15650398e-01 1.19880903e+00 9.19891477e-01
2.22763553e-01 -2.73330748e-01 -4.53166038e-01 1.03544080e+00
2.22524285e-01 -1.63712785e-01 5.14032364e-01 -1.02907836e+00
1.10680211e+00 9.22263384e-01 -2.21497297e-01 -1.15505910e+00
-5.41538000e-01 1.49477184e-01 -8.42288077e-01 -1.37271285e-01
5.26508212e-01 -3.03942293e-01 -2.20936775e-01 1.61854196e+00
2.14235485e-01 -5.96951647e-03 4.83608276e-01 7.98046172e-01
1.59352338e+00 6.76275909e-01 1.25824198e-01 -2.37955272e-01
1.39814401e+00 -1.43095493e+00 -1.18752778e+00 -2.06526741e-01
1.17077506e+00 -5.80341280e-01 8.04248273e-01 -9.01651159e-02
-1.00689089e+00 -3.83222729e-01 -1.46332073e+00 -7.11347759e-01
-1.49805844e-01 8.80701244e-02 6.22199893e-01 2.98839748e-01
-5.59377968e-01 3.23572457e-01 -7.46543586e-01 1.40276983e-01
6.65173531e-01 -1.27727404e-01 -3.38549400e-03 3.46000820e-01
-1.32862234e+00 1.40089345e+00 8.23716402e-01 9.50995460e-03
-4.83684957e-01 -4.31447268e-01 -1.00917435e+00 7.12260753e-02
6.20523810e-01 -4.75503117e-01 1.56643975e+00 -1.02444673e+00
-1.69382620e+00 9.62260723e-01 -1.72856245e-02 -8.19144845e-01
5.38424015e-01 -4.20192689e-01 -6.91948384e-02 -9.44662243e-02
4.48046140e-02 7.73173049e-02 3.33839953e-01 -9.60805535e-01
-4.15175706e-01 1.15136512e-01 7.47710824e-01 2.99360156e-01
-2.75157243e-01 3.58085126e-01 2.08682403e-01 -4.87269282e-01
-2.60952860e-01 -4.48299497e-01 -3.43338326e-02 -6.34171426e-01
-6.93686187e-01 -1.06475282e+00 1.20433474e+00 -7.78583467e-01
1.37595141e+00 -2.01839328e+00 5.15884936e-01 -5.82237065e-01
4.68925357e-01 5.94895542e-01 -1.22586638e-02 3.27119201e-01
1.97524428e-02 2.50106126e-01 -7.24382792e-03 -3.00354362e-01
-8.52410346e-02 3.01948220e-01 -5.49950242e-01 7.93369859e-02
5.74764192e-01 9.44043934e-01 -1.08184397e+00 -5.75538039e-01
-8.51864740e-03 1.20994195e-01 -2.88006395e-01 5.44850528e-01
-8.89022589e-01 7.12625325e-01 -6.71947122e-01 2.72440910e-01
2.16132268e-01 -7.34750867e-01 1.00041389e+00 -4.12922120e-03
4.80456389e-02 1.50137639e+00 -5.76034188e-01 1.60617173e+00
-4.88345414e-01 1.29236042e+00 4.94561829e-02 -1.62595093e+00
1.01216769e+00 7.94156313e-01 1.59174070e-01 -5.92540741e-01
4.47555631e-01 2.09951147e-01 4.71240193e-01 -9.09432292e-01
6.81187510e-01 1.60084546e-01 -1.83635250e-01 6.97976410e-01
4.31907438e-02 3.28118294e-01 2.88558811e-01 5.64357899e-02
1.34686768e+00 2.13826090e-01 5.82301319e-01 -1.20013475e-01
7.44596243e-01 1.59455717e-01 5.48622251e-01 3.34894568e-01
-9.47332159e-02 5.61382294e-01 1.16325569e+00 -4.64173526e-01
-7.94307113e-01 -4.09067124e-01 -2.12883204e-01 1.30543613e+00
2.30353594e-01 -6.40771210e-01 -5.14027119e-01 -7.22943127e-01
-1.74947053e-01 4.90011364e-01 -6.05507612e-01 3.14845949e-01
-1.45467770e+00 -7.65637100e-01 7.85897672e-01 6.58917010e-01
8.17446947e-01 -1.19147515e+00 -6.43955708e-01 3.48525941e-01
-6.37123764e-01 -1.26918256e+00 2.29027152e-01 4.38351072e-02
-3.50237459e-01 -1.62771380e+00 -3.53859872e-01 -1.06473398e+00
6.13344647e-02 1.43658116e-01 1.30961967e+00 4.35296088e-01
4.86011744e-01 -2.71967381e-01 -6.57836258e-01 -2.00790182e-01
-9.72050726e-01 7.39926219e-01 -7.53931820e-01 -5.57340801e-01
2.34681651e-01 -4.20456499e-01 -1.83695048e-01 -1.38321623e-01
-2.90417939e-01 4.31667328e-01 -2.17894334e-02 1.19258940e+00
-1.26595676e-01 -3.75755608e-01 1.01930201e+00 -1.30704248e+00
1.18173563e+00 -7.00483739e-01 -1.80565462e-01 3.91747296e-01
-1.52481925e-02 -1.84122160e-01 3.17492098e-01 -5.44013321e-01
-1.35702157e+00 -7.77882099e-01 -1.94628462e-01 7.21113309e-02
4.04445529e-02 8.13390136e-01 -2.85697430e-01 6.09049082e-01
8.37285042e-01 -5.97271085e-01 1.48842456e-02 -4.02122051e-01
6.02791667e-01 9.13116872e-01 4.20355022e-01 -9.11910355e-01
8.77556279e-02 -9.39214528e-02 -3.47622365e-01 -7.69262433e-01
-1.27094293e+00 -2.00394467e-01 -6.26201630e-01 -2.73039229e-02
1.30396605e+00 -1.00766456e+00 -9.09028113e-01 2.51929551e-01
-1.90665162e+00 -7.20311344e-01 -8.63399059e-02 2.98139099e-02
-4.28653717e-01 1.62664548e-01 -8.93333554e-01 -7.31939495e-01
-3.67399812e-01 -1.28113472e+00 7.02336907e-01 1.12364978e-01
-8.78706574e-01 -1.21957004e+00 1.95937842e-01 6.94217741e-01
2.87103385e-01 6.37836277e-01 1.14689541e+00 -1.11168098e+00
-6.10384405e-01 4.03240621e-01 -4.60932732e-01 2.80691624e-01
3.00591409e-01 -5.53560443e-02 -1.09258640e+00 1.42859016e-02
5.97294606e-02 -8.47870052e-01 8.64540458e-01 1.41261995e-01
1.07955623e+00 -6.88858628e-01 -5.97353220e-01 3.91652822e-01
6.48617089e-01 1.53348669e-01 6.35486603e-01 8.12344968e-01
8.56640637e-01 7.75691569e-01 3.34264457e-01 -7.92246163e-02
9.10106957e-01 7.21583545e-01 1.08981676e-01 1.94102600e-01
-4.80994403e-01 1.88991383e-01 3.06366622e-01 1.22033095e+00
-2.98611432e-01 -2.84798831e-01 -1.09769881e+00 6.41512036e-01
-2.21447349e+00 -1.07665861e+00 -3.62962902e-01 1.18537700e+00
1.39229846e+00 1.72206700e-01 -3.79555374e-02 1.99788868e-01
5.89319825e-01 7.29084074e-01 -2.83212960e-01 -5.74007988e-01
-3.85005325e-01 1.99025467e-01 -2.75796235e-01 4.23776597e-01
-1.49508667e+00 7.22280681e-01 6.16827297e+00 5.60505211e-01
-1.04999149e+00 3.83805037e-01 6.34310842e-01 1.47629574e-01
-7.63879940e-02 -4.22457755e-02 -5.19889355e-01 3.45319778e-01
9.07566667e-01 -4.67501909e-01 -1.50376782e-01 8.27365875e-01
-6.08645320e-01 1.51335523e-01 -1.47455716e+00 4.65251774e-01
-3.38467360e-02 -1.93368685e+00 -1.74151853e-01 -9.98745933e-02
6.81368947e-01 -4.31452245e-02 -2.91371226e-01 7.64049232e-01
6.56788230e-01 -1.34008443e+00 4.09177899e-01 -6.21109642e-02
5.68388224e-01 -5.96454740e-02 8.09938192e-01 4.18034226e-01
-9.65993345e-01 8.01380444e-03 -7.51238922e-03 -7.24979997e-01
2.11297184e-01 -7.85080250e-03 -1.05698466e+00 6.93960130e-01
3.38579088e-01 9.24821794e-01 -2.41692573e-01 3.00502807e-01
-6.35645270e-01 5.77426255e-01 -4.72726747e-02 -1.60371006e-01
8.68068188e-02 -1.01033272e-02 5.58565736e-01 1.24420738e+00
-2.10438013e-01 2.08412081e-01 3.47677350e-01 8.23956192e-01
-5.05934060e-01 -6.77430853e-02 -5.74191928e-01 -1.60913646e-01
6.98312461e-01 1.02461803e+00 -2.43041337e-01 -4.70186919e-01
-5.70226669e-01 2.54990101e-01 9.23215151e-01 1.70417473e-01
-7.14378715e-01 -9.58629847e-02 6.34813964e-01 -2.68710077e-01
1.28773615e-01 -1.09346539e-01 -5.54662526e-01 -1.13291395e+00
8.52656364e-02 -1.10664535e+00 3.91799986e-01 -2.32029840e-01
-1.42946672e+00 7.53636003e-01 1.14535712e-01 -9.42750454e-01
-6.74555719e-01 -6.25141740e-01 -1.00213873e+00 8.03101599e-01
-1.45624828e+00 -1.39123487e+00 -2.68611252e-01 5.49831279e-02
8.49681735e-01 -3.39007050e-01 1.16257310e+00 2.54521251e-01
-6.70382023e-01 4.88458425e-01 -2.09977537e-01 8.22697818e-01
6.81823611e-01 -1.38611698e+00 3.22921038e-01 3.42334569e-01
-1.95070028e-01 4.80476111e-01 4.42373276e-01 -4.26973075e-01
-7.73991704e-01 -7.82300055e-01 8.15267682e-01 -6.99431479e-01
9.83474374e-01 -2.20260784e-01 -1.44248736e+00 1.11410701e+00
1.00028694e+00 -4.72348452e-01 9.19628739e-01 9.42894459e-01
-4.71983939e-01 4.73837048e-01 -5.42453468e-01 5.39516389e-01
9.28191781e-01 -7.28280842e-01 -1.32512176e+00 5.31702042e-01
1.18575847e+00 -1.21517873e+00 -1.21525359e+00 6.02845550e-01
1.85013458e-01 -6.89979613e-01 9.15399015e-01 -9.85388696e-01
1.27926552e+00 1.56427156e-02 -1.41299535e-02 -9.97780979e-01
2.18508378e-01 -5.46959639e-01 -6.92496777e-01 1.70932484e+00
5.02429783e-01 -4.55743700e-01 4.14390385e-01 4.18292195e-01
-3.60896409e-01 -8.59720767e-01 -1.06118536e+00 -4.58971202e-01
6.46433115e-01 -4.54700850e-02 6.28274202e-01 1.44724631e+00
7.24516392e-01 1.25445592e+00 -1.66447274e-02 -2.65910923e-01
-8.12890232e-02 4.97185469e-01 1.01084006e+00 -1.15857267e+00
-2.24222094e-01 -8.62384617e-01 -4.31555994e-02 -1.37786603e+00
7.56443679e-01 -1.20131826e+00 -1.40744835e-01 -1.66326261e+00
-1.61897186e-02 -8.02767515e-01 1.08675309e-01 3.41157347e-01
-4.45168138e-01 -3.72603416e-01 2.68885940e-01 3.40333998e-01
-7.17938185e-01 6.54970229e-01 1.33378696e+00 -6.27547860e-01
-1.86937571e-01 -1.91383973e-01 -7.37886190e-01 8.80847991e-01
6.64312780e-01 -7.05485642e-02 -4.37920481e-01 -1.00010109e+00
1.77655473e-01 2.76466578e-01 6.57565445e-02 -4.46593314e-01
1.56846792e-01 -6.62052631e-02 -8.42416584e-02 -6.18759811e-01
4.77847964e-01 -7.49989897e-02 -4.02910769e-01 5.26313297e-02
-8.27225745e-01 -2.87380248e-01 1.19070180e-01 2.27941766e-01
-5.98843932e-01 -3.12183440e-01 2.19497323e-01 -7.87869617e-02
-4.42358583e-01 -4.14483458e-01 -2.85149872e-01 4.63993818e-01
9.53980386e-01 4.04678226e-01 -1.43890953e+00 -1.35505438e-01
-5.72281539e-01 6.00721776e-01 4.88686264e-02 5.01232743e-01
2.00690955e-01 -1.38482463e+00 -9.25255775e-01 -4.46081579e-01
-2.18207985e-01 8.08920383e-01 -1.89683020e-01 6.77082419e-01
-4.83380884e-01 4.02867049e-01 -2.20718518e-01 -6.24379754e-01
-1.44006848e+00 1.98306546e-01 3.06845605e-01 -8.49242091e-01
-9.65604961e-01 9.09530997e-01 2.68271744e-01 -3.31926346e-01
1.55022398e-01 -4.20008540e-01 -7.87792802e-01 3.83190840e-01
6.17385745e-01 1.44428745e-01 -2.11184785e-01 -6.37024522e-01
1.71885844e-02 -1.78353518e-01 -3.44378293e-01 2.49479607e-01
1.32640040e+00 -5.21406680e-02 -3.57544482e-01 6.42205477e-01
9.73668039e-01 -3.06484580e-01 -1.03247976e+00 -3.55335593e-01
6.15184546e-01 -9.21591893e-02 -3.70690525e-01 -4.31459785e-01
-5.54948628e-01 9.15989399e-01 -4.04403865e-01 9.24299717e-01
7.93085992e-01 1.81982115e-01 7.49408662e-01 5.81464052e-01
-1.82241313e-02 -9.14085329e-01 1.99015602e-01 1.16387093e+00
1.26951766e+00 -1.45132017e+00 1.68474317e-01 -6.75828099e-01
-5.63905358e-01 1.19363213e+00 1.09251440e+00 -3.15077484e-01
3.84446234e-01 3.32756251e-01 2.95451105e-01 -3.23542118e-01
-1.12544680e+00 -6.30574524e-02 2.48019561e-01 7.37385571e-01
1.12694848e+00 1.03780255e-01 -4.49244618e-01 9.14473355e-01
-3.97387594e-01 -2.83664137e-01 4.84112084e-01 8.26272070e-01
-1.05328657e-01 -1.48067105e+00 -9.33069084e-03 2.84020573e-01
-4.25282031e-01 1.91472784e-01 -6.13633275e-01 1.09404826e+00
-1.70412436e-01 9.89404678e-01 3.70414197e-01 -2.88037330e-01
2.07991481e-01 1.16155915e-01 6.41817153e-01 -8.41063142e-01
-9.98758435e-01 -2.92386562e-01 1.10985041e+00 -2.06397682e-01
-1.12696946e+00 -5.05317748e-01 -1.35733414e+00 -3.44957948e-01
-3.65522206e-01 -3.56934853e-02 5.20993322e-02 1.33610308e+00
6.03956729e-02 1.31918764e+00 1.71811059e-01 -5.48547208e-01
-4.33816820e-01 -1.58623862e+00 2.49254823e-01 6.00836992e-01
4.97539371e-01 -8.50456595e-01 1.55093789e-01 1.06529869e-01] | [10.813200950622559, 9.290236473083496] |
e42dbf7e-67e6-4486-85b2-162ab971904a | on-degeneracy-issues-in-multi-parametric | 2304.00435 | null | https://arxiv.org/abs/2304.00435v1 | https://arxiv.org/pdf/2304.00435v1.pdf | On Degeneracy Issues in Multi-parametric Programming and Critical Region Exploration based Distributed Optimization | This paper focuses on two aspects of interest related to multi-parametric linear/quadratic programming (mpLP/QP). First, we study degeneracy issues of mpLP/QP. A novel approach to deal with degeneracies is proposed to find all critical regions containing the given parameter. Our method leverages properties of the multi-parametric linear complementary problem, vertex searching technique, and complementary basis enumeration. Second, an improved critical region exploration (CRE) method to solve distributed LP/QP is proposed under a general mpLP/QP-based formulation. The improved CRE incorporates the proposed approach to handle degeneracies. A cutting plane update and an adaptive stepsize scheme are also integrated to accelerate convergence under different problem settings. The computational efficiency is verified on multi-area tie-line scheduling problems with various testing benchmarks and initial states. | ['Hongbin Sun', 'Hao liu', 'Ye Guo', 'Haitian Liu'] | 2023-04-02 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [ 2.80865222e-01 3.85906957e-02 -7.94877112e-01 7.14834630e-02
-1.19444883e+00 -7.79732406e-01 -2.91455626e-01 2.89035946e-01
1.88218147e-01 1.20517170e+00 -4.80869770e-01 -3.36434573e-01
-8.95201385e-01 -6.90696239e-01 -8.68636787e-01 -8.13246667e-01
-3.20141286e-01 6.17101967e-01 1.84278443e-01 -2.46398017e-01
7.27743626e-01 6.08905971e-01 -9.78698373e-01 6.29374087e-02
1.32441759e+00 1.05310631e+00 2.82421649e-01 3.34268123e-01
-2.63783839e-02 -9.42162871e-02 -3.86875987e-01 5.83414361e-02
6.41413271e-01 -2.99214087e-02 -8.26326549e-01 2.63911575e-01
2.10269205e-02 -1.17933219e-02 4.39941764e-01 8.78277540e-01
3.70546699e-01 5.29239215e-02 5.08536220e-01 -2.09106565e+00
-4.19041738e-02 4.20351684e-01 -1.46973217e+00 3.39150578e-01
2.61885315e-01 -6.40035495e-02 9.98340011e-01 -7.97998965e-01
4.71049964e-01 1.30853689e+00 6.19189382e-01 -2.22929537e-01
-1.41798496e+00 -5.89338064e-01 2.81077415e-01 -6.62457049e-02
-1.74927676e+00 1.81022640e-02 7.73811042e-01 -1.19310699e-01
1.00414920e+00 2.31945157e-01 4.21450585e-01 1.80954427e-01
8.27203095e-01 4.83152747e-01 1.10226929e+00 -5.33869922e-01
2.74820924e-01 -2.93035451e-02 3.47008079e-01 6.89459920e-01
7.56165385e-01 2.33803783e-02 -3.12566936e-01 -7.05636978e-01
6.74085557e-01 -4.93533552e-01 -4.04981449e-02 -9.82679904e-01
-9.02190149e-01 1.07628393e+00 -6.40028343e-02 -3.24117243e-01
-1.05215110e-01 1.35500059e-01 5.37417948e-01 1.35036111e-01
2.52257973e-01 6.10747397e-01 -5.71133614e-01 -3.75709422e-02
-1.02445316e+00 6.03904843e-01 9.69612002e-01 1.40206611e+00
8.15947652e-01 -1.46907538e-01 -4.38869685e-01 8.52939248e-01
2.89526612e-01 3.96409005e-01 -3.98286045e-01 -9.34633017e-01
1.03691041e+00 5.58924139e-01 1.23757392e-01 -1.35513783e+00
-6.23628199e-01 -5.35167038e-01 -4.90179271e-01 1.94859877e-02
7.71122426e-02 -2.37829268e-01 -4.02377695e-01 1.35561705e+00
6.24324501e-01 -9.44646820e-02 -1.93136707e-01 5.85778236e-01
-8.76156837e-02 1.06624007e+00 -5.46436310e-01 -7.86522388e-01
1.15655947e+00 -1.23420262e+00 -5.91529727e-01 -6.23791385e-03
7.10924387e-01 -5.14036417e-01 5.94741940e-01 5.89819431e-01
-1.39892435e+00 -6.66908100e-02 -1.34845352e+00 1.71481103e-01
-7.36633688e-02 1.67066813e-01 5.74944675e-01 8.94602239e-01
-1.04591179e+00 1.20615780e-01 -5.10399520e-01 6.35407567e-02
2.32896090e-01 7.44130135e-01 1.60593390e-01 -2.93031603e-01
-7.17319787e-01 6.16107702e-01 5.54508030e-01 2.79775202e-01
-6.09221220e-01 -1.11647701e+00 -8.49304378e-01 1.66901320e-01
1.06701148e+00 -4.64496374e-01 8.13535154e-01 -3.87483686e-01
-1.47811985e+00 3.08485985e-01 -3.37669015e-01 -9.20897499e-02
2.52074242e-01 2.59961009e-01 -5.04410453e-02 7.35895634e-02
4.17243093e-01 2.59088516e-01 5.01444936e-01 -1.45206118e+00
-6.06549084e-01 -1.87886894e-01 1.45593137e-01 2.62704641e-01
-3.50971855e-02 -9.78061184e-02 -6.26134217e-01 -5.32612264e-01
9.07154754e-02 -1.00669515e+00 -7.54673719e-01 -2.06572086e-01
-5.00042319e-01 2.38959100e-02 6.87248945e-01 -4.27104324e-01
1.66758907e+00 -1.73482215e+00 4.57594812e-01 1.01657951e+00
-1.01065598e-01 -2.21427605e-01 -8.78691152e-02 8.51298988e-01
2.57623941e-03 1.34054288e-01 -5.29758394e-01 4.26879115e-02
1.20357595e-01 3.66579354e-01 -2.40445033e-01 6.58740640e-01
2.85959780e-01 5.95164239e-01 -7.53022671e-01 -7.52893150e-01
-1.46080866e-01 -1.56000793e-01 -1.00187469e+00 -3.74465406e-01
-1.12390473e-01 -5.06705903e-02 -6.50659144e-01 1.04380524e+00
1.25569129e+00 8.55907947e-02 4.36586857e-01 -2.51341015e-01
-5.60742021e-01 -4.19550538e-01 -1.70920658e+00 1.49242043e+00
-5.66101611e-01 1.73882291e-01 6.32275343e-01 -1.11741209e+00
8.38769436e-01 -1.31961539e-01 7.41795003e-01 -5.93494833e-01
6.17539510e-02 2.92403936e-01 -2.41148695e-01 -7.76909739e-02
7.75842011e-01 2.18844369e-01 -4.29364055e-01 4.55090821e-01
-3.76754552e-01 -2.88271725e-01 7.08016455e-01 4.97147329e-02
9.26290393e-01 1.70914710e-01 4.90559012e-01 -1.11050284e+00
7.43040502e-01 4.18464839e-01 1.24144661e+00 4.86307114e-01
-7.26069212e-02 1.64026216e-01 1.42831910e+00 -7.59122744e-02
-8.23941648e-01 -9.20901716e-01 -4.69244033e-01 8.33567441e-01
5.02656519e-01 -4.48772579e-01 -4.31662381e-01 -3.48352462e-01
1.91371351e-01 6.37752414e-01 -3.63250941e-01 8.96933824e-02
-6.74453795e-01 -1.21666586e+00 1.18894704e-01 3.94261837e-01
6.88883588e-02 -2.32345849e-01 -6.16329908e-01 5.32580256e-01
9.37783569e-02 -9.32978988e-01 -5.67738414e-01 3.37439567e-01
-1.01858652e+00 -1.13306177e+00 -4.33146000e-01 -8.52359176e-01
8.95395160e-01 2.40565181e-01 9.14478242e-01 -2.06455708e-01
-5.29767454e-01 2.29372174e-01 -6.58694729e-02 8.83422643e-02
-1.35640174e-01 1.41552433e-01 -1.50910720e-01 -3.79784912e-01
-5.24980128e-01 -3.17422956e-01 -1.58840731e-01 7.91670263e-01
-7.05424547e-01 -1.03209078e-01 6.93200350e-01 9.73891556e-01
1.26673508e+00 5.33538163e-01 7.08008766e-01 -7.13225245e-01
7.76092649e-01 -7.10969746e-01 -1.26291060e+00 6.72170699e-01
-7.22008705e-01 3.19899768e-01 5.47266662e-01 -1.83013126e-01
-9.36512113e-01 4.16089818e-02 4.67974365e-01 -3.87156546e-01
6.19822085e-01 5.87725043e-01 -4.16581243e-01 -6.19113386e-01
4.67788279e-02 -1.74059629e-01 -2.45813131e-01 4.26258855e-02
1.83768705e-01 4.59013917e-02 1.88999146e-01 -1.28120506e+00
8.20988834e-01 2.99902171e-01 6.81022465e-01 -5.01566172e-01
-3.83907527e-01 -4.74584609e-01 -4.49288934e-01 4.97429632e-02
1.66919976e-01 -3.05715173e-01 -9.43956375e-01 -5.57560846e-02
-1.09214568e+00 -9.39594060e-02 -1.29398732e-02 -9.13341790e-02
-8.13626766e-01 6.46784604e-01 -2.92581975e-01 -9.00670767e-01
-2.12089449e-01 -1.52332020e+00 9.83714044e-01 1.25986543e-02
7.51578510e-02 -8.86551559e-01 3.00460756e-01 2.52817363e-01
1.07018590e-01 7.25984514e-01 1.20721757e+00 -9.02631357e-02
-6.00386858e-01 3.07139784e-01 -4.71434623e-01 -2.49579325e-01
-6.07109070e-02 4.33018565e-01 -7.65793696e-02 -7.19214737e-01
-1.81862995e-01 1.69279240e-02 1.73813701e-01 7.33023942e-01
1.23444641e+00 -4.65062946e-01 -8.98924232e-01 8.08027267e-01
1.97548437e+00 3.70204687e-01 2.57541031e-01 3.87334526e-01
3.72826934e-01 4.53528225e-01 1.30945516e+00 8.32266986e-01
4.94572401e-01 8.36032331e-01 4.42130506e-01 -6.95486888e-02
6.38585925e-01 2.05047473e-01 1.36951163e-01 4.91457820e-01
2.39126325e-01 -4.14460838e-01 -1.09020162e+00 7.57221997e-01
-1.86285877e+00 -3.36498946e-01 -2.77533114e-01 2.11264610e+00
7.64693499e-01 1.95595399e-01 2.56103635e-01 2.70163596e-01
9.47972655e-01 -1.72730640e-01 -5.80687582e-01 -1.19673026e+00
8.75649154e-02 2.39927188e-01 1.16651654e+00 5.24595320e-01
-9.40979898e-01 5.04621506e-01 6.40525007e+00 1.17518210e+00
-7.19438791e-01 -6.94400892e-02 6.73152924e-01 -1.82633717e-02
-4.01810616e-01 8.26151222e-02 -1.04279625e+00 2.43959993e-01
7.04514563e-01 -6.72886193e-01 4.61966604e-01 7.66484797e-01
4.14460331e-01 -5.48648119e-01 -8.76040637e-01 6.76433802e-01
3.44955176e-02 -1.30160344e+00 -3.64139646e-01 2.96204031e-01
1.27791321e+00 -6.13830686e-01 2.10704684e-01 9.83704701e-02
3.06295529e-02 -8.05179238e-01 4.62235272e-01 6.97211875e-03
7.30302155e-01 -1.52366936e+00 4.65391248e-01 7.32700154e-02
-1.58957708e+00 -4.46726829e-01 -3.06780159e-01 1.44480944e-01
5.48578382e-01 5.89388669e-01 -7.66291261e-01 1.14331257e+00
5.57236075e-01 3.77868116e-01 -1.60116360e-01 1.44421482e+00
2.14184031e-01 1.62556767e-01 -5.05376875e-01 2.37410739e-01
4.50079918e-01 -5.60560942e-01 7.68867373e-01 9.73279774e-01
2.57913411e-01 -4.86436933e-02 3.87873113e-01 1.09270298e+00
4.64799583e-01 3.47737193e-01 -3.00189346e-01 1.67664692e-01
5.12676179e-01 1.36558783e+00 -1.04911447e+00 3.84668648e-01
-3.11366886e-01 3.20635498e-01 1.00242548e-01 3.78072292e-01
-1.30579495e+00 -4.62653607e-01 4.45903480e-01 -1.22254387e-01
3.68481010e-01 -4.74023789e-01 -6.87353969e-01 -4.07652527e-01
1.59125656e-01 -6.89043701e-01 5.31959593e-01 -2.11283386e-01
-8.90561879e-01 1.78021505e-01 6.32032871e-01 -9.74328399e-01
7.28957541e-03 -4.76014048e-01 -7.15688169e-01 7.24331856e-01
-1.77779603e+00 -8.62061262e-01 7.81700462e-02 1.86413571e-01
6.50640607e-01 2.50796825e-01 2.33973250e-01 3.28177184e-01
-1.07703722e+00 7.43508875e-01 2.89713502e-01 -8.57462883e-01
2.95134127e-01 -1.00575042e+00 -1.75149918e-01 9.18212175e-01
-8.45052361e-01 4.71401453e-01 6.16412520e-01 -6.73797786e-01
-1.93179381e+00 -9.64718819e-01 2.77387023e-01 2.27375120e-01
7.23164916e-01 -2.30600893e-01 -6.43021286e-01 4.54395503e-01
4.52838019e-02 -2.52968132e-01 5.51032364e-01 -1.01163946e-01
2.37239063e-01 -3.26258540e-01 -1.33484328e+00 3.87807369e-01
8.11431885e-01 1.63904354e-01 -9.82094482e-02 4.94189829e-01
6.87199593e-01 -7.13451445e-01 -9.48442400e-01 5.89034557e-01
2.73714274e-01 -3.09702188e-01 1.04660141e+00 3.44955325e-02
1.89181328e-01 -3.56510550e-01 -1.22154236e-01 -1.12224901e+00
-3.45093548e-01 -1.10666203e+00 -1.15953706e-01 1.17831862e+00
6.17338598e-01 -7.86673248e-01 5.52224100e-01 5.80772638e-01
-4.36487615e-01 -1.32539833e+00 -1.12652934e+00 -1.08578694e+00
1.76770300e-01 -7.20908344e-02 4.71448660e-01 7.10968554e-01
1.06021084e-01 -8.76740739e-02 -2.46193111e-01 5.87344050e-01
7.51207292e-01 4.20419067e-01 4.69282866e-01 -7.64157712e-01
-4.33590472e-01 -2.52845526e-01 7.90377408e-02 -6.64495289e-01
8.42480659e-02 -7.70161152e-01 -3.75517011e-02 -1.49728942e+00
1.05585523e-01 -1.00518191e+00 -8.63176808e-02 3.67073923e-01
1.56900197e-01 -1.55951604e-01 1.81463473e-02 -2.21115530e-01
-6.18415594e-01 5.21866143e-01 1.25952220e+00 3.17933559e-02
-5.49048722e-01 9.21326056e-02 -5.15874863e-01 1.92098469e-01
8.37817550e-01 -5.30396879e-01 -6.04078948e-01 -2.06972688e-01
7.11593449e-01 6.33690000e-01 -4.15302766e-03 -7.79073775e-01
3.33616257e-01 -7.16117620e-01 -3.37438136e-01 -1.18099892e+00
1.81598350e-01 -9.23450232e-01 3.12031031e-01 5.78520834e-01
7.41520077e-02 5.98101020e-01 7.67951846e-01 6.33145511e-01
-7.52759725e-02 -4.66530532e-01 6.73750818e-01 4.83727813e-01
-6.94188118e-01 1.57120213e-01 -2.78662205e-01 6.42462149e-02
1.77847803e+00 -5.24925470e-01 -3.10973674e-01 1.08714849e-01
-3.11385661e-01 1.06950557e+00 3.70112747e-01 -8.85376707e-03
4.71663892e-01 -1.05368388e+00 -4.70460713e-01 6.69288635e-02
-6.25066366e-03 6.19162433e-02 2.38408878e-01 1.27661622e+00
-8.58671606e-01 7.78009176e-01 -2.65035212e-01 -6.43798947e-01
-1.04890954e+00 7.58695185e-01 1.59480959e-01 -8.18500519e-01
-1.04333207e-01 6.55988038e-01 3.74645703e-02 -2.71913230e-01
-1.46700561e-01 -6.23170614e-01 2.63116866e-01 7.71553665e-02
-8.85298178e-02 1.03252196e+00 -1.54051613e-02 -3.05832148e-01
-5.78199565e-01 8.79717231e-01 1.99925583e-02 -2.52713934e-02
1.28643036e+00 -2.44695887e-01 -4.85299855e-01 8.60385373e-02
1.16060674e+00 7.53187388e-02 -1.03384590e+00 1.89289749e-01
-3.12283393e-02 -5.29647112e-01 2.68297285e-01 -4.53330547e-01
-1.15328240e+00 2.08850250e-01 1.25205904e-01 3.08784973e-02
1.37735605e+00 -4.52090055e-01 6.51619911e-01 1.86267197e-01
7.20000803e-01 -1.41832876e+00 -3.56247336e-01 4.23003405e-01
7.92913198e-01 -7.90125489e-01 5.77195406e-01 -1.24650013e+00
-5.43315768e-01 1.28164601e+00 5.91455996e-01 -6.97582811e-02
4.48799193e-01 7.31582940e-01 -8.09112370e-01 -4.82131019e-02
-7.91189790e-01 2.58463800e-01 9.89715680e-02 2.70720690e-01
-1.47005230e-01 -1.24863312e-01 -1.06240726e+00 5.13433814e-01
2.57850915e-01 -1.71688497e-01 5.83370268e-01 1.31280470e+00
-1.46027699e-01 -1.26840353e+00 -7.96873629e-01 2.90080041e-01
-1.33699417e-01 9.53663364e-02 1.28172591e-01 9.11402464e-01
-8.11068788e-02 9.27253425e-01 5.12529463e-02 1.93499863e-01
1.70626312e-01 -4.47370291e-01 5.90642035e-01 -4.77459878e-01
-5.42253792e-01 1.76779583e-01 1.76292941e-01 -9.05928731e-01
-1.73383251e-01 -7.25520790e-01 -1.31036174e+00 -8.63747150e-02
-5.95459521e-01 2.32559115e-01 6.20242476e-01 6.65631115e-01
3.54512453e-01 7.03179061e-01 9.19267595e-01 -6.38551831e-01
-4.74963605e-01 6.43805116e-02 -6.50768459e-01 -7.76578426e-01
2.59868633e-02 -1.09608364e+00 -1.95631638e-01 -5.41300595e-01] | [5.285336971282959, 3.0373449325561523] |
dbbf3274-96bf-4729-919b-c1172f306404 | gpteval-nlg-evaluation-using-gpt-4-with | 2303.16634 | null | https://arxiv.org/abs/2303.16634v3 | https://arxiv.org/pdf/2303.16634v3.pdf | G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment | The quality of texts generated by natural language generation (NLG) systems is hard to measure automatically. Conventional reference-based metrics, such as BLEU and ROUGE, have been shown to have relatively low correlation with human judgments, especially for tasks that require creativity and diversity. Recent studies suggest using large language models (LLMs) as reference-free metrics for NLG evaluation, which have the benefit of being applicable to new tasks that lack human references. However, these LLM-based evaluators still have lower human correspondence than medium-size neural evaluators. In this work, we present G-Eval, a framework of using large language models with chain-of-thoughts (CoT) and a form-filling paradigm, to assess the quality of NLG outputs. We experiment with two generation tasks, text summarization and dialogue generation. We show that G-Eval with GPT-4 as the backbone model achieves a Spearman correlation of 0.514 with human on summarization task, outperforming all previous methods by a large margin. We also propose preliminary analysis on the behavior of LLM-based evaluators, and highlight the potential issue of LLM-based evaluators having a bias towards the LLM-generated texts. The code is at https://github.com/nlpyang/geval | ['Chenguang Zhu', 'Ruochen Xu', 'Shuohang Wang', 'Yichong Xu', 'Dan Iter', 'Yang Liu'] | 2023-03-29 | null | null | null | null | ['dialogue-generation', 'dialogue-generation'] | ['natural-language-processing', 'speech'] | [ 7.62736723e-02 6.68885708e-01 6.54956251e-02 -8.69581699e-02
-1.24003410e+00 -7.04688132e-01 9.04307604e-01 3.91017765e-01
-4.41921681e-01 9.63984907e-01 7.83514440e-01 -2.48372361e-01
9.52699110e-02 -6.87353373e-01 -3.08484435e-01 -1.26295686e-01
4.02490020e-01 7.88174331e-01 -1.99727178e-01 -4.03154939e-01
6.40463352e-01 -1.91964090e-01 -1.03236938e+00 5.73265553e-01
1.43261182e+00 4.00308102e-01 1.47977918e-01 1.11581326e+00
-9.80754718e-02 8.47835124e-01 -1.18499601e+00 -8.13510299e-01
-1.51034966e-01 -8.62726986e-01 -1.05455053e+00 -2.76865661e-01
6.08608305e-01 -1.73706919e-01 -4.83704247e-02 7.26582825e-01
9.83865023e-01 1.58920899e-01 8.49234760e-01 -9.60398018e-01
-9.32421446e-01 1.00770247e+00 -5.04413359e-02 4.32781316e-02
7.62445331e-01 4.43501562e-01 1.14555883e+00 -8.18845749e-01
7.42860198e-01 1.66454089e+00 6.92737639e-01 8.01726341e-01
-1.36224806e+00 -2.62005925e-01 -2.38927320e-01 -2.23050490e-01
-8.43141913e-01 -6.00195229e-01 3.48491967e-01 -4.01342452e-01
1.37649655e+00 3.35702747e-01 5.70510924e-01 1.18860793e+00
3.70256811e-01 8.79352629e-01 9.00309265e-01 -4.98430967e-01
1.50282934e-01 1.88404843e-01 1.62071735e-01 6.58865213e-01
4.47708547e-01 -2.27593392e-01 -8.18334222e-01 -3.45897675e-01
4.13235426e-01 -8.59025300e-01 -5.25518060e-01 3.37603867e-01
-1.37883699e+00 1.11241353e+00 1.33727640e-01 4.21000898e-01
-4.58847225e-01 2.27113813e-01 4.68893021e-01 3.49818677e-01
7.61348188e-01 1.37592196e+00 -2.98749059e-01 -4.75794405e-01
-1.17559862e+00 4.38837886e-01 1.23172438e+00 8.11796188e-01
2.07531318e-01 1.67977393e-01 -9.18769300e-01 1.01141405e+00
1.38541877e-01 3.98326844e-01 1.15613341e+00 -1.14672875e+00
6.62583530e-01 6.15352392e-01 1.60142332e-01 -9.36965406e-01
-4.22901154e-01 -3.80333334e-01 -7.84121335e-01 -8.23612437e-02
3.81781936e-01 -4.13488716e-01 -4.14362192e-01 1.85588527e+00
-1.55455530e-01 -6.66842639e-01 2.66142190e-01 5.25708616e-01
1.04553199e+00 8.61513913e-01 5.08161671e-02 -3.46010447e-01
1.00218987e+00 -1.05871177e+00 -6.04239762e-01 -2.70492315e-01
9.83293116e-01 -8.67120743e-01 1.45741224e+00 4.45023268e-01
-1.53638542e+00 -5.84717274e-01 -8.55944037e-01 -1.75950497e-01
-1.48605883e-01 2.93063492e-01 4.13280457e-01 7.42600739e-01
-1.47330844e+00 9.85867798e-01 -4.80084091e-01 -5.30733407e-01
1.36708021e-01 1.18618309e-02 -7.21752495e-02 3.73231977e-01
-1.26510763e+00 1.25120223e+00 5.47812283e-01 -7.32358024e-02
-6.30446970e-01 -3.54190826e-01 -6.89543426e-01 -2.74008159e-02
-4.47651632e-02 -1.09013581e+00 1.64250469e+00 -7.51362562e-01
-1.53840518e+00 7.54075527e-01 -1.85317829e-01 -5.18588483e-01
6.69217646e-01 -3.56154829e-01 -4.09255661e-02 6.39877394e-02
1.92858934e-01 7.80458093e-01 3.20943654e-01 -9.85275328e-01
-3.02461654e-01 -2.29947567e-02 -2.76639573e-02 5.64930439e-01
-2.86717683e-01 -8.55136439e-02 1.11365303e-01 -5.13889849e-01
-2.92790890e-01 -7.23633349e-01 -1.36438146e-01 -6.92434251e-01
-6.16797328e-01 -7.17238426e-01 -1.00047462e-01 -8.38308334e-01
1.71051610e+00 -1.36429572e+00 8.30869675e-02 -2.27612108e-01
2.22488970e-01 4.91415620e-01 -5.33170044e-01 8.10408592e-01
3.58434558e-01 6.43172622e-01 -7.03267232e-02 -4.41858113e-01
1.90529063e-01 -3.68176371e-01 -1.05503447e-01 -1.49123788e-01
2.64029980e-01 1.28448296e+00 -1.18882716e+00 -6.20240033e-01
-1.94827244e-01 1.25097468e-01 -4.57844257e-01 1.99498162e-01
-4.47862118e-01 8.97164717e-02 -2.73369819e-01 1.75361603e-01
-2.82421224e-02 -3.62103581e-01 2.49098856e-02 2.68031597e-01
-1.12771295e-01 7.38284051e-01 -7.11088538e-01 1.67824483e+00
-5.38645744e-01 8.35410535e-01 -3.21183830e-01 -2.52823800e-01
9.50405121e-01 5.11059523e-01 -2.62092829e-01 -5.89274764e-01
2.17401031e-02 4.34835970e-01 2.72155076e-01 -2.21301630e-01
1.07286787e+00 -1.45571185e-02 -1.18492335e-01 9.83314276e-01
2.59124637e-01 -6.91973209e-01 8.35270703e-01 7.02631712e-01
1.11719239e+00 -8.47113580e-02 5.11924028e-01 -3.40241373e-01
3.24825466e-01 4.41928171e-02 2.06978634e-01 1.06679463e+00
9.60547328e-02 6.09188735e-01 6.77541316e-01 8.75983164e-02
-1.13956106e+00 -7.17667103e-01 2.50638008e-01 9.80151474e-01
-4.59813505e-01 -8.27700019e-01 -1.22112477e+00 -4.42575067e-01
-3.51599246e-01 1.44730604e+00 -3.13786328e-01 -2.24649921e-01
-4.94496047e-01 -9.10902619e-01 9.04228628e-01 2.75662929e-01
1.98283449e-01 -1.47271359e+00 -6.24065578e-01 4.82845545e-01
-6.52241826e-01 -6.16721809e-01 -4.82041121e-01 -2.32334152e-01
-1.07309985e+00 -7.09470510e-01 -8.76609623e-01 -4.78698730e-01
4.27635878e-01 -1.90807998e-01 1.64360297e+00 2.67118007e-01
1.86109185e-01 3.36015314e-01 -4.79731172e-01 -4.39535826e-01
-1.12386489e+00 4.51959491e-01 -1.11137420e-01 -6.94139481e-01
2.02268854e-01 -2.84415781e-01 -5.15637994e-01 1.07509278e-01
-7.31730998e-01 4.68036294e-01 6.99600399e-01 9.63636696e-01
6.31984547e-02 -5.16957402e-01 1.05842137e+00 -9.82598484e-01
1.72932160e+00 -2.01678976e-01 4.25931066e-02 4.10005689e-01
-1.00836515e+00 1.79765359e-01 6.50658071e-01 -3.66559088e-01
-1.01638699e+00 -6.82423651e-01 -8.25337023e-02 3.32468599e-01
5.87804876e-02 6.38420939e-01 5.95415756e-02 3.95428449e-01
1.26061010e+00 6.63446113e-02 -7.70659149e-02 -2.07591355e-01
4.82219219e-01 7.37176001e-01 2.42273405e-01 -6.04064047e-01
4.21102017e-01 -3.45475733e-01 -4.50503558e-01 -8.32482874e-01
-8.76135468e-01 -1.36703938e-01 -3.28531563e-01 -2.77287662e-01
6.02555633e-01 -8.13025594e-01 -3.36782485e-01 3.43856394e-01
-1.54139531e+00 -6.66956842e-01 -4.18589205e-01 4.47015166e-01
-7.35212028e-01 5.16016185e-01 -9.26096499e-01 -8.95405352e-01
-1.07849634e+00 -7.88231790e-01 1.01556802e+00 3.65555763e-01
-1.19260097e+00 -1.10122383e+00 3.47421974e-01 5.45864582e-01
4.88329977e-01 2.33707443e-01 9.51265335e-01 -9.77274120e-01
-1.53270394e-01 -1.79154783e-01 4.17448245e-02 4.20533746e-01
-1.93147853e-01 -6.45926893e-02 -7.89830387e-01 -1.00610845e-01
-1.66866314e-02 -6.72051728e-01 8.85776758e-01 4.48626131e-01
4.64093149e-01 -7.06207752e-01 5.76913282e-02 9.55532864e-02
1.03741503e+00 -6.82606399e-02 6.89370811e-01 1.88608214e-01
5.17153263e-01 8.12504828e-01 5.20188987e-01 4.16383743e-01
3.45659316e-01 1.88951567e-01 -1.99661866e-01 1.71785071e-01
-2.34500423e-01 -6.27278626e-01 6.68775916e-01 1.28652143e+00
-1.40494883e-01 -9.50515032e-01 -8.06790709e-01 4.62869227e-01
-1.76146412e+00 -9.71676350e-01 -3.48547250e-01 2.01998401e+00
1.16995466e+00 1.54446438e-01 3.62748876e-02 -1.29857719e-01
7.13676453e-01 2.29266405e-01 -2.83088207e-01 -8.88968527e-01
-3.50412667e-01 2.01318964e-01 3.54383998e-02 6.34818971e-01
-4.80943501e-01 1.00304210e+00 6.27875280e+00 9.59415197e-01
-5.02290070e-01 5.56885190e-02 8.10529709e-01 -1.84052378e-01
-5.34812033e-01 1.08547388e-02 -6.75172448e-01 3.89467955e-01
1.40799916e+00 -7.50425339e-01 3.17865849e-01 5.57678819e-01
5.46493232e-01 -2.45764434e-01 -1.29328120e+00 7.65244901e-01
3.61173511e-01 -1.02396441e+00 3.49786431e-01 -8.13211948e-02
1.08328950e+00 -9.37226266e-02 -2.45764539e-01 4.61683333e-01
4.22715157e-01 -1.10355365e+00 8.19214523e-01 5.52404642e-01
7.14374483e-01 -4.65490431e-01 7.92398393e-01 5.62243700e-01
-5.54887950e-01 2.24827975e-01 -5.08860946e-01 -1.22394703e-01
4.38255996e-01 5.84975779e-01 -1.16432691e+00 3.04428518e-01
-4.03361954e-02 2.25725397e-01 -8.32720995e-01 7.98986137e-01
-5.37541449e-01 7.56612778e-01 6.41851053e-02 -6.67529345e-01
1.56415328e-01 -4.89858128e-02 7.10140884e-01 1.48831606e+00
5.55230916e-01 1.33837461e-02 -1.65193945e-01 9.81798232e-01
-2.45948434e-01 5.33214331e-01 -5.10952115e-01 -5.23231030e-01
4.06302899e-01 1.26696908e+00 -5.15096486e-01 -6.06420934e-01
4.59184200e-02 9.98279810e-01 3.11875790e-01 1.18738279e-01
-3.46105516e-01 -4.74762201e-01 -8.69938135e-02 9.71441045e-02
-4.46266413e-01 -9.34357122e-02 -4.37175095e-01 -1.13399231e+00
-8.10060725e-02 -1.30363786e+00 1.91500977e-01 -1.06532073e+00
-1.24161339e+00 8.17929149e-01 -1.46738410e-01 -8.70511115e-01
-8.85292947e-01 -4.66235489e-01 -7.80469239e-01 1.02915239e+00
-8.89091909e-01 -8.65730345e-01 -1.91929489e-01 -2.40563788e-02
8.51969600e-01 -3.00225258e-01 8.55640352e-01 -3.00775111e-01
-2.47100443e-01 5.84437430e-01 5.09623848e-02 -4.25589783e-03
9.16700900e-01 -1.47756326e+00 7.73223519e-01 6.42785966e-01
1.61560997e-01 8.52769315e-01 8.63470912e-01 -9.01035726e-01
-8.34296286e-01 -9.30228412e-01 1.43875051e+00 -8.57256651e-01
4.24722522e-01 -6.61172718e-02 -7.19701767e-01 3.10873508e-01
6.27491295e-01 -1.16694796e+00 7.11603999e-01 2.15370566e-01
8.07967931e-02 5.39050758e-01 -8.95863712e-01 6.74204767e-01
9.72411454e-01 -3.51170480e-01 -8.72088313e-01 6.13789201e-01
7.63902128e-01 -1.49681926e-01 -7.64454246e-01 7.47043863e-02
5.20454049e-01 -9.32456195e-01 4.63990390e-01 -4.77643132e-01
9.17176306e-01 -3.40269580e-02 1.87035486e-01 -1.98670280e+00
-3.61662954e-01 -8.90868723e-01 -9.78555903e-02 1.46069872e+00
8.67148399e-01 -6.31750107e-01 4.59440947e-01 7.03482866e-01
-2.81271935e-01 -6.02282405e-01 -4.26645190e-01 -7.38958061e-01
4.28146601e-01 -1.89216271e-01 3.19868416e-01 6.27456427e-01
3.73461455e-01 1.04157281e+00 -2.12914988e-01 -7.16622472e-01
2.44613200e-01 -1.68861330e-01 8.12317133e-01 -1.35723948e+00
-3.86496127e-01 -8.36184204e-01 8.41227695e-02 -7.62839615e-01
1.51061103e-01 -1.15538585e+00 2.40190223e-01 -2.14286828e+00
3.39128166e-01 1.70809045e-01 2.45698974e-01 1.95466116e-01
-3.85277331e-01 7.09386617e-02 3.52585971e-01 2.20066339e-01
-6.73987806e-01 6.34269416e-01 1.32409668e+00 -1.60191208e-02
-4.76355195e-01 -1.51463136e-01 -1.16946959e+00 7.02303708e-01
1.08276129e+00 -4.19099689e-01 -3.25157046e-01 -4.57823187e-01
6.31082416e-01 1.03669398e-01 7.03046247e-02 -8.82492661e-01
1.60960406e-01 5.98484427e-02 4.15111572e-01 -3.29039395e-01
1.78104837e-03 3.95209402e-01 -1.25831505e-02 3.68966073e-01
-8.96719038e-01 3.85269552e-01 1.52415875e-02 1.20518304e-01
-1.39115870e-01 -7.51178265e-01 4.41191226e-01 -6.13139927e-01
-9.04780850e-02 -1.62267953e-01 -5.33691883e-01 5.65989137e-01
2.66253352e-01 -9.11306515e-02 -7.48388529e-01 -8.11067224e-01
-2.37498268e-01 2.90248156e-01 5.36969662e-01 2.15610042e-01
4.60619301e-01 -1.05488241e+00 -1.33536422e+00 -4.62335855e-01
2.06664158e-03 -2.24366322e-01 1.11680165e-01 6.73356056e-01
-5.73183835e-01 8.03868830e-01 3.42670567e-02 -3.80522348e-02
-1.06571090e+00 -2.31086990e-04 1.30375728e-01 -8.18423569e-01
-1.95401385e-01 8.27366650e-01 9.24930349e-02 -5.19077182e-01
-1.13163382e-01 -2.09840074e-01 -2.63195425e-01 3.06454599e-01
5.39380431e-01 8.12356830e-01 8.61076231e-04 -5.08302152e-01
1.84060201e-01 6.39249533e-02 1.06337592e-01 -6.43009305e-01
1.00341260e+00 -3.34460326e-02 -2.12935910e-01 5.88675857e-01
7.92239249e-01 6.22048303e-02 -4.96359587e-01 1.20283313e-01
2.21600100e-01 -6.35118037e-02 -3.01969379e-01 -1.16329086e+00
-2.39350319e-01 8.71818542e-01 -1.59399305e-02 3.46035659e-01
7.09864438e-01 -1.56410277e-01 7.73923278e-01 6.59437120e-01
1.75883472e-01 -1.51418340e+00 3.47439378e-01 7.87102759e-01
1.31358027e+00 -1.03381526e+00 -5.06630875e-02 1.79600582e-01
-9.45169508e-01 1.08174670e+00 6.57353103e-01 1.43583968e-01
-2.18145147e-01 -2.90465087e-01 -9.11815464e-03 -9.60902348e-02
-1.28377664e+00 1.83816440e-02 4.59487677e-01 3.55702162e-01
1.16823936e+00 3.06632012e-01 -9.91845906e-01 6.55053020e-01
-9.96667981e-01 -5.46868630e-02 8.91086996e-01 4.77473676e-01
-6.02252901e-01 -1.04672003e+00 -2.17020690e-01 7.65576959e-01
-5.44077277e-01 -4.61620748e-01 -9.54292297e-01 5.21201849e-01
-3.54543656e-01 1.42208290e+00 -2.29460433e-01 -1.85353652e-01
2.80909121e-01 4.56974506e-01 4.76955414e-01 -1.02927148e+00
-1.02257907e+00 -7.99999535e-02 7.21547306e-01 -3.11314166e-01
-1.59837931e-01 -7.26321042e-01 -9.01486278e-01 -3.37341964e-01
-4.89408106e-01 5.05809665e-01 4.05661494e-01 6.08501017e-01
3.67418498e-01 3.47383499e-01 1.94916323e-01 -7.03984439e-01
-8.94374013e-01 -1.76204741e+00 -4.11750793e-01 4.12044585e-01
-1.12656772e-01 2.14196667e-02 -4.95790243e-01 -3.95030752e-02] | [11.976853370666504, 9.12154483795166] |
43f8f5e4-66ae-4840-934a-ebc813a3dfb8 | ctrlstruct-dialogue-structure-learning-for | 2303.01094 | null | https://arxiv.org/abs/2303.01094v1 | https://arxiv.org/pdf/2303.01094v1.pdf | CTRLStruct: Dialogue Structure Learning for Open-Domain Response Generation | Dialogue structure discovery is essential in dialogue generation. Well-structured topic flow can leverage background information and predict future topics to help generate controllable and explainable responses. However, most previous work focused on dialogue structure learning in task-oriented dialogue other than open-domain dialogue which is more complicated and challenging. In this paper, we present a new framework CTRLStruct for dialogue structure learning to effectively explore topic-level dialogue clusters as well as their transitions with unlabelled information. Precisely, dialogue utterances encoded by bi-directional Transformer are further trained through a special designed contrastive learning task to improve representation. Then we perform clustering to utterance-level representations and form topic-level clusters that can be considered as vertices in dialogue structure graph. The edges in the graph indicating transition probability between vertices are calculated by mimicking expert behavior in datasets. Finally, dialogue structure graph is integrated into dialogue model to perform controlled response generation. Experiments on two popular open-domain dialogue datasets show our model can generate more coherent responses compared to some excellent dialogue models, as well as outperform some typical sentence embedding methods in dialogue utterance representation. Code is available in GitHub. | ['Zhaochun Ren', 'Piji Li', 'Congchi Yin'] | 2023-03-02 | null | null | null | null | ['dialogue-generation', 'response-generation', 'dialogue-generation'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [ 2.95311004e-01 1.00337493e+00 -1.12015940e-01 -5.83193541e-01
-8.06799412e-01 -5.24817586e-01 1.02855051e+00 1.63681477e-01
1.79945305e-01 1.06023371e+00 1.05255568e+00 -2.07856223e-01
3.01057458e-01 -9.10858512e-01 -1.68100595e-01 -4.73109573e-01
8.80902410e-02 8.52839530e-01 7.86652975e-03 -8.27213347e-01
3.41978312e-01 -3.80039126e-01 -9.43270564e-01 6.34079456e-01
1.07480395e+00 3.24496031e-01 1.70490116e-01 9.65915501e-01
-8.29928219e-01 1.07964993e+00 -1.06837976e+00 -3.88116717e-01
-4.34844971e-01 -9.91760254e-01 -1.44998813e+00 4.07879859e-01
-2.67421514e-01 -2.44989812e-01 -2.75675893e-01 6.22215450e-01
6.10406160e-01 5.31639338e-01 9.52738106e-01 -1.01265788e+00
-6.95644498e-01 1.19020116e+00 -1.92643121e-01 -8.03500265e-02
8.89524579e-01 1.26136824e-01 1.21383119e+00 -7.06581712e-01
6.96014822e-01 1.74807811e+00 1.74644083e-01 1.19444525e+00
-1.22707772e+00 -2.59655476e-01 2.77939141e-01 1.32722422e-01
-5.47920585e-01 -3.44404310e-01 1.16257250e+00 -2.09399119e-01
9.68707561e-01 3.72830510e-01 5.43368638e-01 1.36914575e+00
-2.47309264e-02 1.12264025e+00 8.32857788e-01 -4.81851041e-01
-5.40174618e-02 5.34558952e-01 3.93866330e-01 4.99066949e-01
-5.66011965e-01 -4.58686471e-01 -3.58125448e-01 -4.08732116e-01
4.81843770e-01 -3.35878938e-01 -3.94556582e-01 -9.19991210e-02
-1.30064380e+00 1.50342417e+00 3.92109603e-01 2.33854592e-01
-2.49564856e-01 -4.07682478e-01 6.81301296e-01 5.04507363e-01
6.55829370e-01 7.81615674e-01 -3.66974652e-01 -2.92464465e-01
-1.46142140e-01 4.03896958e-01 1.35181332e+00 7.91087806e-01
8.17760289e-01 1.97316632e-01 -8.73016477e-01 1.31525540e+00
2.46233627e-01 -3.13056298e-02 8.28795195e-01 -7.38820732e-01
6.49090171e-01 9.26223576e-01 -6.62243664e-02 -1.10859406e+00
-4.32236403e-01 3.04628640e-01 -9.59314585e-01 -5.00433028e-01
2.22233981e-01 -8.50769341e-01 -3.76697600e-01 1.45178759e+00
4.69254076e-01 -3.62968608e-03 5.75148821e-01 7.31908917e-01
1.44523048e+00 1.27553606e+00 3.00528947e-02 -2.96668530e-01
1.46763337e+00 -1.27526772e+00 -1.10485077e+00 -4.73792106e-03
8.43589187e-01 -5.56213796e-01 1.17504287e+00 -1.17398135e-01
-8.93615186e-01 -5.54221392e-01 -5.41173041e-01 -3.85159906e-03
-2.20534101e-01 -1.02646530e-01 5.79398692e-01 4.20338780e-01
-9.15071189e-01 1.63522363e-01 -3.12804699e-01 -3.10349733e-01
-1.60672236e-02 2.01892450e-01 7.13024475e-03 2.89750814e-01
-1.80541909e+00 8.44985008e-01 5.05115390e-01 -1.40316904e-01
-6.82810724e-01 -4.66962308e-01 -1.31382895e+00 -2.26402748e-02
4.87850308e-01 -5.71973205e-01 1.60695565e+00 -4.43967432e-01
-2.16320395e+00 5.71914673e-01 -1.34890556e-01 -4.79079515e-01
1.54793158e-01 8.15313980e-02 -8.06022435e-02 1.06749162e-01
-7.38303885e-02 8.85690689e-01 5.83600879e-01 -1.40361905e+00
-4.95009631e-01 -9.48622078e-02 4.17948872e-01 8.64552498e-01
-4.85161841e-01 -2.18641981e-01 -1.07650489e-01 -2.85658240e-01
-2.68954098e-01 -6.90572619e-01 -6.06437802e-01 -7.80644834e-01
-8.16191018e-01 -1.08470273e+00 8.66252363e-01 -6.27594411e-01
1.28448975e+00 -1.51823437e+00 3.07798117e-01 -2.51943856e-01
3.94707054e-01 1.38938874e-01 -3.62281725e-02 9.88120556e-01
2.04951018e-01 1.40801251e-01 -1.08370900e-01 -1.57173663e-01
2.88579196e-01 2.02963546e-01 -4.93657380e-01 -2.39607975e-01
3.42021585e-01 8.36402416e-01 -1.08132327e+00 -4.99902874e-01
4.09645647e-01 4.12175171e-02 -5.21969199e-01 9.39058959e-01
-6.61664367e-01 7.65154660e-01 -7.83974648e-01 -5.98009769e-03
2.26510540e-01 -1.69793531e-01 3.25652540e-01 2.91675385e-02
1.75623000e-01 7.07384944e-01 -6.64048672e-01 1.57019258e+00
-8.98852110e-01 5.93349218e-01 -1.62572134e-02 -1.22553229e+00
1.52501202e+00 6.23492837e-01 1.96471632e-01 -3.30647588e-01
8.16686451e-02 -4.34358776e-01 1.78895041e-01 -6.70291662e-01
9.19261754e-01 1.68592501e-02 -5.80634713e-01 8.70832264e-01
1.68655127e-01 -6.04428470e-01 5.34771830e-02 5.85793018e-01
6.70519173e-01 -1.78475738e-01 3.03551823e-01 -3.72063108e-02
7.98015416e-01 1.11130819e-01 1.36478677e-01 6.57964587e-01
-3.31021585e-02 2.59873718e-01 9.47260141e-01 -2.53891408e-01
-7.85901725e-01 -7.40257025e-01 7.11868778e-02 1.29031360e+00
-7.10505769e-02 -5.57813346e-01 -8.19572031e-01 -7.49740183e-01
-2.95145929e-01 8.95225108e-01 -5.65183938e-01 -2.20650077e-01
-5.45907497e-01 -5.50295115e-01 5.21297634e-01 1.52712107e-01
5.85342765e-01 -1.38623893e+00 -5.49900159e-02 4.64707464e-01
-5.77330172e-01 -7.10843384e-01 -5.28403282e-01 -1.77133501e-01
-5.51435649e-01 -8.71365905e-01 -7.72339225e-01 -1.06768131e+00
5.24712503e-01 4.51388620e-02 1.13869584e+00 -1.96899697e-01
-1.42080113e-01 5.16547680e-01 -6.25910938e-01 -2.30530947e-01
-8.99568617e-01 2.88761079e-01 -2.53828526e-01 1.75387934e-02
3.37567568e-01 -1.87987506e-01 -4.84893948e-01 2.42030650e-01
-5.98471642e-01 2.72177249e-01 5.00646420e-02 1.44781280e+00
-9.55538228e-02 -4.07024145e-01 1.17848992e+00 -1.32854187e+00
1.85131192e+00 -8.15979421e-01 2.89524496e-02 2.97965288e-01
-3.89451891e-01 1.90168396e-01 7.65639961e-01 -4.19987440e-01
-1.64395356e+00 -1.59537822e-01 -2.64826536e-01 1.71766981e-01
-3.27400297e-01 5.91456652e-01 -4.83888201e-02 6.27577305e-01
7.04538465e-01 4.20642763e-01 2.63190717e-01 -1.21142551e-01
8.14210713e-01 1.07567418e+00 9.48979408e-02 -7.78148353e-01
3.05811197e-01 -1.61670461e-01 -6.98121727e-01 -1.10203171e+00
-6.03295207e-01 -4.09843385e-01 -4.73854035e-01 -3.11656594e-01
9.11860645e-01 -6.12600565e-01 -7.94043183e-01 5.99866807e-02
-1.46307623e+00 -4.20443356e-01 -1.74307436e-01 1.70769081e-01
-5.63798785e-01 3.55098903e-01 -8.16862941e-01 -1.02591479e+00
-5.17943203e-01 -9.27307606e-01 9.01609004e-01 5.13948321e-01
-6.50522113e-01 -1.73458099e+00 4.05472070e-01 5.14344513e-01
2.95941532e-01 1.88227281e-01 1.05274129e+00 -1.34938204e+00
-9.72618088e-02 -4.30163555e-02 5.72490916e-02 7.14235380e-02
3.70044857e-01 -1.67284220e-01 -7.76175559e-01 8.94861072e-02
1.32109925e-01 -8.79382551e-01 5.74814856e-01 1.63299054e-01
8.74019742e-01 -8.60198319e-01 -3.42797339e-01 -1.03187583e-01
5.54651260e-01 3.44032228e-01 4.56848919e-01 -1.53289393e-01
6.67665601e-01 1.33550525e+00 7.77129292e-01 6.78983033e-01
7.45024264e-01 6.67286158e-01 -4.37096767e-02 -1.13305837e-01
2.02278644e-01 -4.57231045e-01 3.45323503e-01 1.29334581e+00
3.65279287e-01 -4.27267075e-01 -7.78115630e-01 5.23055077e-01
-2.01142621e+00 -9.94390428e-01 -3.35939318e-01 1.62698829e+00
1.44485438e+00 -1.38228551e-01 1.66670755e-01 -3.17637950e-01
8.63364220e-01 6.63460255e-01 -4.03402328e-01 -8.51979554e-01
1.69467658e-01 -3.66224870e-02 -3.67386848e-01 8.00284088e-01
-7.97480643e-01 1.16724038e+00 5.45107508e+00 7.04686642e-01
-6.60532594e-01 -6.34726062e-02 8.11763048e-01 4.32619333e-01
-7.01152742e-01 1.46429380e-02 -7.35307455e-01 2.22286910e-01
9.39962268e-01 -6.78724945e-01 -1.03681572e-02 7.70913601e-01
2.78694451e-01 2.47849882e-01 -1.04771566e+00 6.59590065e-01
2.77602255e-01 -1.40706551e+00 2.71197110e-01 -2.44847402e-01
7.67630517e-01 -8.30637217e-01 -2.18440801e-01 9.38257277e-01
8.41741621e-01 -1.04257429e+00 -3.75045627e-01 3.09994191e-01
3.62292588e-01 -7.24337995e-01 6.84395790e-01 5.57578683e-01
-1.04732287e+00 2.17107460e-01 -6.18403614e-01 -1.51132300e-01
3.34047288e-01 8.56564660e-03 -1.82068372e+00 6.48481846e-01
-6.21125586e-02 6.94882929e-01 -2.43317455e-01 5.05400598e-01
-3.03712696e-01 7.66879201e-01 3.00247669e-01 -7.39677370e-01
4.00745332e-01 -3.50870639e-01 4.59826678e-01 1.25898969e+00
-5.39618172e-02 3.67418796e-01 5.15963554e-01 8.29378188e-01
-9.76015329e-02 4.29546326e-01 -8.94077182e-01 -1.37312233e-01
5.37011027e-01 1.38562870e+00 -3.12869519e-01 -3.71614397e-01
-2.25179493e-01 8.77103686e-01 3.16106111e-01 1.65043026e-01
-5.55859506e-01 -5.54770827e-01 3.60711634e-01 -3.20079505e-01
-2.09735245e-01 1.59136847e-01 3.06364223e-02 -1.13371515e+00
-4.65092570e-01 -1.03522170e+00 3.96870822e-01 -4.39653307e-01
-1.41525328e+00 8.34352851e-01 9.57617909e-03 -9.74997103e-01
-1.02957678e+00 -1.82748571e-01 -1.29693162e+00 9.39386845e-01
-1.00987709e+00 -9.21743333e-01 -1.58642262e-01 7.53990650e-01
1.48217666e+00 -4.99907047e-01 1.41314614e+00 -4.20321345e-01
-6.19933784e-01 5.89369118e-01 -1.92443281e-02 3.11246872e-01
7.74504423e-01 -1.52518141e+00 4.54094499e-01 5.30501157e-02
-1.07553728e-01 5.60824871e-01 6.41450584e-01 -6.60982788e-01
-9.84485805e-01 -8.78490031e-01 8.45877767e-01 -2.98030913e-01
7.40436196e-01 -6.03231311e-01 -1.06667590e+00 3.67756635e-01
1.07166171e+00 -9.33038116e-01 1.03215837e+00 3.40386599e-01
2.13119239e-01 4.08945978e-01 -7.70813942e-01 8.35316598e-01
4.70067084e-01 -4.10342574e-01 -1.16410184e+00 7.76860297e-01
1.15083516e+00 -6.03896797e-01 -8.72490525e-01 -5.44218272e-02
-3.62846367e-02 -5.94785333e-01 6.66908085e-01 -1.03004682e+00
8.74591708e-01 4.38562363e-01 3.83934259e-01 -1.79955840e+00
9.62419361e-02 -1.14539671e+00 -4.18271236e-02 1.60582066e+00
6.24560535e-01 -5.00377536e-01 7.47956276e-01 6.77621841e-01
-3.66999626e-01 -7.92027831e-01 -4.71249551e-01 -4.81748767e-02
2.15371788e-01 2.74585754e-01 5.03511906e-01 1.19786048e+00
8.58319938e-01 1.26920569e+00 -6.34193301e-01 -4.00967896e-01
2.07367912e-01 2.55600721e-01 1.18919992e+00 -1.15866601e+00
-1.88391060e-01 -2.39577308e-01 1.59984410e-01 -1.55414844e+00
6.40340567e-01 -7.98597991e-01 3.30884278e-01 -1.93848515e+00
-3.85808833e-02 -3.06729108e-01 5.15035629e-01 1.05026647e-01
-5.91848314e-01 -5.66404104e-01 -6.62661642e-02 4.68244916e-03
-5.73746443e-01 1.12844932e+00 1.64726686e+00 -3.97585928e-01
-5.42081833e-01 4.26331192e-01 -8.15108895e-01 5.39119899e-01
8.80030513e-01 -1.45129263e-01 -8.40541482e-01 8.08664411e-02
-4.17929232e-01 9.53468382e-01 -1.66832313e-01 -2.71877319e-01
2.38400072e-01 -2.82208830e-01 -6.41170815e-02 -4.93089914e-01
5.56511104e-01 -1.95769250e-01 -5.68140328e-01 3.20922792e-01
-1.21852040e+00 -2.03656390e-01 -1.12873271e-01 6.97275996e-01
-4.94696677e-01 -5.23737252e-01 2.65144438e-01 -3.37683856e-01
-4.60254014e-01 1.28143594e-01 -7.04157293e-01 3.43185157e-01
1.05696714e+00 -1.02863275e-01 -5.27275920e-01 -1.27677441e+00
-8.82665813e-01 7.65867174e-01 -1.67066470e-01 7.01487422e-01
8.41423869e-01 -1.31040931e+00 -1.07615316e+00 -1.90434724e-01
5.81478290e-02 1.17817372e-01 5.06433368e-01 2.57761627e-01
-1.11364998e-01 6.07342303e-01 -8.35415497e-02 -5.36475539e-01
-1.27340758e+00 1.31292507e-01 1.24483310e-01 -6.88316047e-01
-4.84782457e-01 9.31864977e-01 3.49268854e-01 -1.04268062e+00
3.07325721e-01 6.74452186e-02 -1.03101766e+00 3.10393333e-01
4.76863176e-01 7.10431188e-02 -4.57263708e-01 -4.54077095e-01
2.03453273e-01 -9.88430902e-02 -4.33726907e-01 -2.75914073e-01
9.68178988e-01 -3.70622396e-01 -1.95995495e-01 7.03992784e-01
1.15346599e+00 -1.81235492e-01 -1.15519881e+00 -3.43947083e-01
-3.90108079e-02 -1.88717559e-01 -5.57582557e-01 -5.06512940e-01
-5.39965987e-01 1.08351803e+00 1.02834282e-02 9.17081475e-01
4.83620703e-01 9.23214182e-02 8.56007576e-01 7.31461167e-01
-2.83101220e-02 -1.16267514e+00 7.17070282e-01 9.14440930e-01
1.13945687e+00 -1.35172200e+00 -3.34075093e-01 -3.43507618e-01
-1.54431903e+00 1.24320018e+00 1.13751304e+00 -5.80737591e-02
2.52029121e-01 -2.18476936e-01 2.40117982e-01 -1.93945155e-01
-1.19889963e+00 3.19796391e-02 1.17607720e-01 8.83290410e-01
8.24617624e-01 1.25792203e-02 -4.04507786e-01 7.65518725e-01
-4.69546109e-01 -7.82692909e-01 9.13158536e-01 5.13665617e-01
-5.84304690e-01 -1.26626730e+00 -1.27396062e-01 6.05379760e-01
-1.00706078e-01 -1.11713395e-01 -9.03109133e-01 4.80380028e-01
-7.35333800e-01 1.46472788e+00 1.29903033e-01 -5.39004326e-01
1.98604375e-01 3.18271458e-01 -2.30892394e-02 -1.34049952e+00
-9.13675666e-01 -9.83455852e-02 7.80959964e-01 -1.45216705e-02
-2.74294794e-01 -3.96519989e-01 -1.29534805e+00 -6.11538030e-02
-4.87512529e-01 8.13770711e-01 2.22096041e-01 7.89760947e-01
2.30826467e-01 7.93587029e-01 1.17419755e+00 -6.78684056e-01
-8.96615505e-01 -1.48828018e+00 -2.48382434e-01 5.52820444e-01
-2.03216495e-03 -3.42950433e-01 -1.55895993e-01 8.03186670e-02] | [12.707254409790039, 8.066283226013184] |
e98b1b5f-52d7-4b4f-8574-0d6879bad931 | approximating-interactive-human-evaluation | 1906.09308 | null | https://arxiv.org/abs/1906.09308v2 | https://arxiv.org/pdf/1906.09308v2.pdf | Approximating Interactive Human Evaluation with Self-Play for Open-Domain Dialog Systems | Building an open-domain conversational agent is a challenging problem. Current evaluation methods, mostly post-hoc judgments of static conversation, do not capture conversation quality in a realistic interactive context. In this paper, we investigate interactive human evaluation and provide evidence for its necessity; we then introduce a novel, model-agnostic, and dataset-agnostic method to approximate it. In particular, we propose a self-play scenario where the dialog system talks to itself and we calculate a combination of proxies such as sentiment and semantic coherence on the conversation trajectory. We show that this metric is capable of capturing the human-rated quality of a dialog model better than any automated metric known to-date, achieving a significant Pearson correlation (r>.7, p<.05). To investigate the strengths of this novel metric and interactive evaluation in comparison to state-of-the-art metrics and human evaluation of static conversations, we perform extended experiments with a set of models, including several that make novel improvements to recent hierarchical dialog generation architectures through sentiment and semantic knowledge distillation on the utterance level. Finally, we open-source the interactive evaluation platform we built and the dataset we collected to allow researchers to efficiently deploy and evaluate dialog models. | ['Rosalind Picard', 'Natasha Jaques', 'Noah Jones', 'Judy Hanwen Shen', 'Craig Ferguson', 'Agata Lapedriza', 'Asma Ghandeharioun'] | 2019-06-21 | approximating-interactive-human-evaluation-1 | http://papers.nips.cc/paper/9519-approximating-interactive-human-evaluation-with-self-play-for-open-domain-dialog-systems | http://papers.nips.cc/paper/9519-approximating-interactive-human-evaluation-with-self-play-for-open-domain-dialog-systems.pdf | neurips-2019-12 | ['dialogue-evaluation', 'open-domain-dialog'] | ['natural-language-processing', 'natural-language-processing'] | [-3.52932364e-01 5.00535905e-01 1.47284597e-01 -8.46793652e-01
-8.82522166e-01 -1.02922821e+00 1.20870328e+00 1.08938232e-01
-2.98991203e-01 9.52294767e-01 1.01031017e+00 -1.18667968e-01
6.47127395e-04 -5.95900655e-01 -1.72743574e-02 -8.20450410e-02
2.82837097e-02 1.20023429e+00 2.11104661e-01 -7.85395086e-01
1.42911941e-01 -1.16799735e-01 -1.36015546e+00 5.77481151e-01
6.57683074e-01 8.48726630e-01 -5.91113679e-02 9.31598961e-01
-1.92892119e-01 1.08358037e+00 -1.20613825e+00 -8.17143977e-01
-9.67411995e-02 -6.17849588e-01 -1.41729736e+00 1.91518605e-01
4.31974158e-02 -7.69233227e-01 9.59491804e-02 5.18987119e-01
5.58129847e-01 2.82014459e-01 4.83539492e-01 -1.42296219e+00
-4.80071634e-01 1.08726013e+00 5.20457327e-01 -1.24819376e-01
9.59101081e-01 5.01639843e-01 1.36826408e+00 -5.70171177e-01
9.30455804e-01 1.49184179e+00 6.19299829e-01 7.36946464e-01
-1.19620132e+00 -2.37079978e-01 4.90220450e-02 -1.19215295e-01
-6.47254467e-01 -6.55900002e-01 4.76307154e-01 -5.56285679e-01
1.07033455e+00 3.63196552e-01 5.31595469e-01 1.53944135e+00
-3.98495972e-01 5.98600507e-01 1.39006639e+00 -2.15324730e-01
3.63305271e-01 6.53196871e-01 3.54553640e-01 4.32059735e-01
-4.27290142e-01 -4.82238233e-02 -6.66638255e-01 -4.13760185e-01
2.62239307e-01 -4.84359920e-01 -1.04483053e-01 -1.29575506e-01
-1.22097647e+00 1.03971922e+00 2.05143094e-01 4.49221820e-01
-2.16083482e-01 -4.08345729e-01 4.44591224e-01 6.41791046e-01
5.47939062e-01 9.81614709e-01 -5.54431975e-01 -8.08469117e-01
-5.11538565e-01 6.35913491e-01 1.75093102e+00 8.29294622e-01
4.88039881e-01 -4.67420101e-01 -4.73748922e-01 1.07142854e+00
2.55781740e-01 1.45099893e-01 5.61765850e-01 -1.59712517e+00
1.53388798e-01 8.60778451e-01 3.75091791e-01 -6.26296878e-01
-5.56232572e-01 1.50270104e-01 -1.03492461e-01 2.06457719e-01
7.82162726e-01 -5.03932834e-01 -5.88306859e-02 1.86705434e+00
2.22806320e-01 -4.56363231e-01 4.46866274e-01 7.85206020e-01
1.16708219e+00 4.19897377e-01 2.50023734e-02 -2.76835561e-01
1.49791968e+00 -1.28291559e+00 -8.16857517e-01 9.75426938e-03
6.34935737e-01 -7.63262272e-01 1.62007964e+00 3.26242745e-01
-1.22390103e+00 -3.74063045e-01 -1.02067196e+00 6.14229366e-02
-4.33981448e-01 -4.64879364e-01 6.34740889e-01 7.07107842e-01
-1.36584640e+00 5.40931106e-01 -3.60310882e-01 -6.03404224e-01
-4.80427176e-01 2.70966757e-02 -1.35739401e-01 4.47541803e-01
-1.36845613e+00 1.41673255e+00 5.10707311e-03 -5.13063848e-01
-8.70888531e-01 -4.35153186e-01 -6.42993808e-01 1.50700919e-02
2.20662564e-01 -7.20621169e-01 2.23979092e+00 -7.98356414e-01
-2.12749434e+00 8.74314547e-01 2.63986308e-02 -4.53335494e-01
6.58532262e-01 -1.05156407e-01 -2.39482343e-01 -7.44423941e-02
-2.49481108e-02 7.07090437e-01 2.10122485e-02 -1.47253489e+00
-5.41668892e-01 -1.12196438e-01 9.77282882e-01 5.72242081e-01
-4.26815957e-01 1.30328611e-01 -6.84468076e-02 -7.49825612e-02
-4.94296879e-01 -9.64205801e-01 -1.92480952e-01 -4.87518549e-01
-2.90636808e-01 -4.28710073e-01 5.44280887e-01 -5.03049076e-01
1.26770926e+00 -1.65213549e+00 1.06416985e-01 -6.22834899e-02
3.81030411e-01 6.40157014e-02 -2.89312415e-02 8.52571309e-01
5.66747963e-01 7.87905827e-02 -1.56138688e-01 -8.40581894e-01
4.93644565e-01 2.35266760e-02 -7.57317767e-02 -2.39353910e-01
-5.26035950e-02 8.98906052e-01 -1.02647817e+00 -5.40321290e-01
2.74592936e-01 1.71492040e-01 -7.27217078e-01 7.83587098e-01
-7.58228481e-01 4.63821709e-01 -2.15672761e-01 2.41725743e-01
6.09103665e-02 -3.31357181e-01 4.19903308e-01 4.67554778e-02
-9.93093997e-02 8.04986179e-01 -7.66687572e-01 1.79021323e+00
-7.55900085e-01 7.17980146e-01 1.90107450e-01 -2.58379638e-01
1.03263903e+00 5.29609621e-01 2.49267042e-01 -4.73997921e-01
2.79534400e-01 -1.02971472e-01 3.84002998e-02 -3.40372413e-01
7.94941366e-01 -1.42506301e-01 -4.25014794e-01 9.17722046e-01
2.43278086e-01 -5.78628719e-01 2.88573414e-01 5.75883985e-01
1.28061020e+00 -2.90544569e-01 3.32142353e-01 -2.95159519e-01
4.77558047e-01 2.36413345e-01 -1.05854154e-01 6.90046191e-01
-4.58182096e-01 3.42129618e-01 6.92609310e-01 -6.26580492e-02
-9.09177482e-01 -1.04156661e+00 7.42126778e-02 1.49925590e+00
5.31234145e-02 -6.96547151e-01 -1.18543530e+00 -7.76382983e-01
-2.41805255e-01 1.04201567e+00 -4.64959115e-01 1.36957899e-01
-2.33315006e-01 -4.47622061e-01 7.00276792e-01 2.88993210e-01
5.68769097e-01 -1.19272184e+00 -3.51767331e-01 2.80694664e-01
-6.83600366e-01 -1.29716861e+00 -3.38296294e-01 -1.27883941e-01
-4.78398740e-01 -8.98741841e-01 -2.24916548e-01 -4.62095886e-01
-1.06039261e-02 -5.02577424e-02 1.71954024e+00 9.41511318e-02
5.06175041e-01 6.55613065e-01 -6.83142781e-01 -1.81210414e-01
-1.08754277e+00 1.32369712e-01 -2.41149142e-02 -4.68344688e-01
5.32284439e-01 -4.45043474e-01 -6.37463808e-01 7.84911156e-01
-5.22627175e-01 1.05464393e-02 1.57026142e-01 8.85933459e-01
-4.41089332e-01 -4.04315144e-01 8.89601886e-01 -8.97699416e-01
1.63285458e+00 -4.56276089e-01 -1.36464849e-01 2.15729922e-01
-6.88553691e-01 4.47669066e-02 4.64916378e-01 -1.51570097e-01
-1.39997172e+00 -3.50220770e-01 -3.39730084e-01 2.12730289e-01
-2.24716023e-01 4.14848417e-01 -3.09269782e-02 3.33981335e-01
9.25877154e-01 -3.20343018e-01 2.68579841e-01 -2.50988454e-01
7.65224695e-01 9.26291287e-01 5.34668744e-01 -7.30246067e-01
3.13875556e-01 1.68560565e-01 -8.29980075e-01 -5.07630467e-01
-7.85707116e-01 -4.47439075e-01 -3.71985793e-01 -6.01416469e-01
8.75023246e-01 -8.69191408e-01 -1.25213695e+00 2.08934560e-01
-1.27501869e+00 -9.16622937e-01 -2.17544585e-01 2.12291464e-01
-7.52763331e-01 2.16674551e-01 -8.88719320e-01 -1.00900269e+00
-3.75744432e-01 -1.19121015e+00 9.02477682e-01 5.76369353e-02
-1.07191014e+00 -1.29536903e+00 5.73062062e-01 7.92606235e-01
7.23424435e-01 1.68185547e-01 5.04498601e-01 -1.21875060e+00
-1.47836491e-01 -6.04108684e-02 -5.09271510e-02 3.96057874e-01
-6.91486476e-03 6.13840483e-02 -1.23200107e+00 -1.81462206e-02
1.58900559e-01 -9.19060111e-01 1.70762941e-01 -3.30619030e-02
3.30025613e-01 -5.82610846e-01 1.54302269e-02 -3.20054054e-01
6.70448720e-01 1.51187316e-01 2.74423629e-01 3.53407830e-01
4.70037498e-02 1.03577268e+00 6.27451956e-01 6.33936942e-01
9.90284443e-01 9.02941883e-01 3.04937661e-02 1.61037385e-01
-7.49685755e-03 -2.59675175e-01 5.01088619e-01 9.55101252e-01
5.49112186e-02 -5.01343012e-01 -8.22918773e-01 3.26113611e-01
-1.89010620e+00 -9.35752392e-01 1.41468555e-01 1.83769310e+00
1.17452133e+00 4.36109841e-01 5.81432581e-01 -1.86712563e-01
4.01171088e-01 2.16831550e-01 -1.54861107e-01 -7.61813343e-01
-1.05595812e-01 -8.93077254e-02 -2.06060365e-01 1.06429088e+00
-6.97110355e-01 9.96506929e-01 7.01165342e+00 2.40773648e-01
-5.38235009e-01 2.38258079e-01 7.10236788e-01 4.84175719e-02
-4.17569697e-01 2.23546159e-02 -5.42222738e-01 2.99058855e-01
1.41634059e+00 -3.94981563e-01 5.43890297e-01 1.05149436e+00
2.11374119e-01 -1.14170395e-01 -1.62155604e+00 5.50421000e-01
1.71284899e-02 -1.17907393e+00 -2.55236179e-01 -1.54909017e-02
6.16956592e-01 -1.58037305e-01 -2.63513893e-01 8.47923517e-01
1.18261838e+00 -9.53017473e-01 5.03996789e-01 3.47009063e-01
3.30090195e-01 -3.01770329e-01 8.77633095e-01 3.26477766e-01
-7.60029733e-01 1.68377012e-02 2.53620654e-01 -3.42897862e-01
5.57323694e-01 1.20343417e-01 -1.48865283e+00 1.31468967e-01
4.69234288e-01 1.57567799e-01 -3.99831712e-01 3.06782246e-01
-4.76003624e-03 2.61377752e-01 -1.22485973e-01 -4.56907034e-01
3.58818740e-01 -4.21917215e-02 5.67356408e-01 1.41007197e+00
-1.98032465e-02 3.01488370e-01 2.70339161e-01 7.64226019e-01
-1.47176579e-01 8.25964212e-02 -6.55251086e-01 4.28129509e-02
7.89896548e-01 1.33783233e+00 -3.33537281e-01 -6.23497248e-01
-3.26782286e-01 7.32037544e-01 4.13431287e-01 -6.55214563e-02
-6.66331887e-01 -6.11605402e-03 8.71871412e-01 -2.36267760e-01
-3.08802664e-01 -1.40487492e-01 -3.98069292e-01 -9.20072615e-01
-9.61832553e-02 -1.27053905e+00 3.67870957e-01 -6.64268672e-01
-1.50092757e+00 1.09747827e+00 7.29906037e-02 -9.51885939e-01
-1.00617349e+00 -5.04572272e-01 -7.99181104e-01 5.34650803e-01
-8.26972902e-01 -9.12665784e-01 -7.47768819e-01 4.47432429e-01
8.20249498e-01 -2.28359342e-01 1.30199051e+00 -1.70265380e-02
-1.88990027e-01 5.57256758e-01 -1.80272266e-01 -4.16607689e-03
1.06183660e+00 -1.58574212e+00 4.54037994e-01 1.20345585e-01
-1.50165230e-01 6.54963374e-01 1.17141974e+00 -4.95502591e-01
-9.94394362e-01 -4.90392178e-01 9.15784299e-01 -1.17921126e+00
9.49770689e-01 -3.92036706e-01 -6.92419887e-01 5.85454226e-01
7.62725890e-01 -1.03725660e+00 8.17082644e-01 5.74034870e-01
-2.48573482e-01 2.17755407e-01 -1.30146658e+00 6.71846688e-01
1.00509846e+00 -6.26158059e-01 -1.01020908e+00 4.58341062e-01
1.05773234e+00 -3.74337494e-01 -1.16906977e+00 4.24619913e-01
7.73649454e-01 -1.44118786e+00 8.28029871e-01 -5.52620888e-01
5.95613182e-01 1.34609833e-01 -3.27765822e-01 -1.58832741e+00
1.53221250e-01 -8.79926562e-01 5.61954416e-02 1.49833727e+00
7.86755919e-01 -4.34654415e-01 7.04062641e-01 1.11671710e+00
-2.83376068e-01 -4.99936640e-01 -5.65765142e-01 -5.77540040e-01
8.51193145e-02 -3.35637867e-01 6.94597065e-01 9.02624309e-01
1.01513672e+00 9.16410029e-01 -2.32334346e-01 -3.47405851e-01
1.38607770e-01 -6.49418831e-02 1.23634720e+00 -1.19020545e+00
-3.83585483e-01 -7.81813860e-01 -6.15495034e-02 -1.14870203e+00
1.81253731e-01 -5.45257449e-01 2.60922223e-01 -1.60021400e+00
1.19110502e-01 -3.02628964e-01 2.39577964e-01 -4.02532592e-02
-1.27802491e-01 -3.25707830e-02 2.13587940e-01 1.56000495e-01
-9.62188005e-01 6.83375239e-01 1.03814161e+00 -2.21138466e-02
-4.63528693e-01 -1.83498226e-02 -8.88173938e-01 6.78636312e-01
8.62617671e-01 3.55201922e-02 -5.36560297e-01 -1.37380973e-01
-1.60445273e-01 3.00081372e-01 1.87252015e-01 -9.36032891e-01
4.07157570e-01 3.97983119e-02 -2.57014364e-01 -2.75288701e-01
8.32283497e-01 -4.28478390e-01 -2.17775330e-02 -1.81325281e-03
-8.17198813e-01 1.53479129e-01 -4.81414013e-02 2.81133592e-01
-3.52988929e-01 -2.09190011e-01 4.57854927e-01 -2.45415226e-01
-5.15011311e-01 -3.59181970e-01 -5.07370949e-01 1.78496540e-01
7.81608999e-01 1.46679401e-01 -8.25993836e-01 -1.15657508e+00
-6.68395340e-01 5.45801103e-01 4.99097288e-01 5.90048432e-01
2.15014085e-01 -1.04984891e+00 -8.35821688e-01 -4.65125769e-01
3.50902379e-01 -5.02859473e-01 5.72601408e-02 5.11982977e-01
-3.60809922e-01 4.48127151e-01 -2.19739303e-01 -4.65982944e-01
-1.26682341e+00 1.53546676e-01 2.36372918e-01 -5.17585218e-01
-1.28840491e-01 6.12192094e-01 -2.17904616e-02 -9.22324777e-01
4.99584228e-01 -2.20583186e-01 -3.31272244e-01 2.58444607e-01
5.81701815e-01 3.12852859e-01 6.97015449e-02 -4.40621912e-01
-1.22948751e-01 -1.95718631e-01 8.94762129e-02 -8.65091443e-01
9.32965994e-01 -4.59483504e-01 5.22358827e-02 7.86090672e-01
9.53637660e-01 -1.41170276e-02 -1.10007906e+00 -3.73388320e-01
1.63422912e-01 -1.38657957e-01 -4.87646073e-01 -1.28281236e+00
-3.20347995e-01 4.13171202e-01 2.77284861e-01 9.99743342e-01
5.61887145e-01 3.98794651e-01 6.18263960e-01 8.72789025e-01
3.13763976e-01 -1.26109421e+00 5.75771809e-01 8.18597794e-01
1.18724954e+00 -1.54561603e+00 -2.97214389e-01 -1.38841227e-01
-1.36011279e+00 8.71568322e-01 9.74799037e-01 2.68372267e-01
5.07455587e-01 3.50883901e-01 5.90497732e-01 -4.01592225e-01
-1.35413587e+00 -2.36335143e-01 -2.78643798e-02 4.47623998e-01
8.42955232e-01 3.01659793e-01 -1.75607204e-01 7.52284765e-01
-1.05946398e+00 -2.83843666e-01 5.53748071e-01 6.89814329e-01
-5.61892331e-01 -1.08395278e+00 -2.46722884e-02 2.20297351e-01
-2.86419541e-02 4.77294773e-02 -1.09831715e+00 7.77268112e-01
-5.04921734e-01 1.77722263e+00 -5.68004623e-02 -8.07802975e-01
6.36034012e-01 4.15576786e-01 2.85431463e-02 -5.35197556e-01
-1.12689185e+00 -4.08677429e-01 1.08297980e+00 -5.04773080e-01
-6.06176317e-01 -5.84322512e-01 -8.20824921e-01 -7.35929310e-01
-2.97369301e-01 5.59705973e-01 7.04663455e-01 9.06412303e-01
1.57599062e-01 3.33228171e-01 8.82117808e-01 -8.44519973e-01
-6.92559063e-01 -1.54907525e+00 -5.12727015e-02 8.20565701e-01
-6.05622167e-03 -6.53584838e-01 -5.37805736e-01 6.38568355e-03] | [12.76809310913086, 8.027803421020508] |
abcbce6a-0bca-4176-b1ea-35be95dfc333 | an-improved-graph-model-for-chinese-spell | null | null | https://aclanthology.org/W14-6825 | https://aclanthology.org/W14-6825.pdf | An Improved Graph Model for Chinese Spell Checking | null | ['Zhongye Jia', 'Yang Xin', 'Yuzhu Wang', 'Hai Zhao'] | 2014-10-01 | null | null | null | ws-2014-10 | ['chinese-spell-checking'] | ['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] | [-7.467456817626953, 3.6108603477478027] |
77c1864b-b98e-4da3-91ac-546b141ccf97 | joint-optimization-of-multi-objective | 2105.14125 | null | https://arxiv.org/abs/2105.14125v1 | https://arxiv.org/pdf/2105.14125v1.pdf | Joint Optimization of Multi-Objective Reinforcement Learning with Policy Gradient Based Algorithm | Many engineering problems have multiple objectives, and the overall aim is to optimize a non-linear function of these objectives. In this paper, we formulate the problem of maximizing a non-linear concave function of multiple long-term objectives. A policy-gradient based model-free algorithm is proposed for the problem. To compute an estimate of the gradient, a biased estimator is proposed. The proposed algorithm is shown to achieve convergence to within an $\epsilon$ of the global optima after sampling $\mathcal{O}(\frac{M^4\sigma^2}{(1-\gamma)^8\epsilon^4})$ trajectories where $\gamma$ is the discount factor and $M$ is the number of the agents, thus achieving the same dependence on $\epsilon$ as the policy gradient algorithm for the standard reinforcement learning. | ['Vaneet Aggarwal', 'Mridul Agarwal', 'Qinbo Bai'] | 2021-05-28 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [-1.55578256e-01 1.54280230e-01 -9.66353714e-02 -1.49157450e-01
-6.65671170e-01 -2.43436009e-01 3.04673985e-02 3.65527242e-01
-1.18243742e+00 1.31545103e+00 -6.63151324e-01 -4.04670298e-01
-5.13824582e-01 -6.77229583e-01 -5.92725098e-01 -8.63405466e-01
-4.00139719e-01 3.99742663e-01 1.26962855e-01 -1.55871391e-01
7.11656988e-01 1.90049559e-01 -1.22525477e+00 -4.58001405e-01
1.07996619e+00 1.44719791e+00 2.76357055e-01 8.83588374e-01
8.27302933e-02 6.17905974e-01 -5.73541522e-01 -2.44020954e-01
2.65367359e-01 -6.33536518e-01 -6.36362851e-01 1.03981204e-01
-2.06628188e-01 -2.83164263e-01 2.16867298e-01 1.31894886e+00
3.88730317e-01 5.30310810e-01 6.70789599e-01 -1.11876810e+00
2.70603932e-02 4.34671082e-02 -5.80237985e-01 2.95222223e-01
-8.27244669e-02 2.21564040e-01 8.48575592e-01 -4.76128846e-01
1.40285447e-01 9.77240622e-01 2.99158096e-01 2.52338260e-01
-7.76875913e-01 -5.36053121e-01 5.28477490e-01 -2.90028542e-01
-1.18752646e+00 -7.08582178e-02 2.28939593e-01 -3.91770691e-01
9.55263376e-01 -4.20946628e-03 6.13144219e-01 6.46985546e-02
7.04739451e-01 2.93106526e-01 1.17755806e+00 -5.56023479e-01
7.17243969e-01 1.98136508e-01 -2.40207404e-01 8.90300333e-01
3.18070829e-01 4.44504648e-01 -2.35129729e-01 -1.23593353e-01
1.02139437e+00 -1.22833334e-01 7.80054331e-02 -6.02638796e-02
-7.43400872e-01 1.09376895e+00 8.20641890e-02 4.36504222e-02
-6.85230136e-01 8.16616058e-01 -1.09257318e-01 4.72123325e-01
4.86442477e-01 6.03179097e-01 -4.30601060e-01 -4.51334506e-01
-6.74936354e-01 6.01009130e-01 6.48372948e-01 6.27930045e-01
6.33764684e-01 6.20933592e-01 1.58567861e-01 5.87577879e-01
5.65347850e-01 7.07078338e-01 2.13336170e-01 -1.46921372e+00
4.32062417e-01 3.44675690e-01 8.40631008e-01 -6.38450384e-01
-1.79609656e-01 -5.49023330e-01 -3.54408920e-01 8.41521561e-01
6.62850440e-01 -7.80387759e-01 -5.00429690e-01 1.66583872e+00
2.51159370e-01 -5.16408563e-01 -2.79750049e-01 5.29394209e-01
-5.20004094e-01 7.08730817e-01 -4.97376136e-02 -7.96611369e-01
7.70088732e-01 -6.32478833e-01 -5.55322587e-01 -5.76619029e-01
1.02063879e-01 -6.64116204e-01 9.66295183e-01 3.79021645e-01
-1.47159350e+00 -1.57325834e-01 -1.00099123e+00 8.84047866e-01
8.72168690e-03 -2.32240319e-01 4.86244649e-01 5.96055686e-01
-9.09432232e-01 8.75241578e-01 -7.29900777e-01 2.66018212e-01
1.08086735e-01 6.21458113e-01 4.20622855e-01 2.98319221e-01
-7.58008122e-01 9.61574554e-01 4.19968665e-01 8.53709728e-02
-8.38035882e-01 -4.04036045e-01 -5.60162902e-01 -7.30033591e-02
6.88860774e-01 -1.60365865e-01 1.39953351e+00 -8.84344041e-01
-1.89352763e+00 2.29887918e-01 3.38378474e-02 -3.67350847e-01
5.56546867e-01 -7.15892538e-02 -3.79172452e-02 1.99139327e-01
1.33184344e-01 2.35848814e-01 1.02459455e+00 -9.81781006e-01
-9.75615442e-01 -3.98327172e-01 2.61792094e-01 2.59670824e-01
-1.48152530e-01 3.05229146e-02 1.31800652e-01 -4.69078243e-01
-1.27030343e-01 -1.06034565e+00 -6.00752115e-01 -1.65328771e-01
3.27313751e-01 5.99446334e-02 2.74331212e-01 -5.60182571e-01
1.32586968e+00 -1.83572769e+00 3.85549217e-02 4.84727740e-01
-1.52211696e-01 5.16734421e-02 2.61210382e-01 3.46571237e-01
2.50841916e-01 1.18299965e-02 -4.12301987e-01 1.93457655e-03
3.74351367e-02 3.66969593e-02 -1.34991594e-02 5.96665263e-01
-9.03540254e-02 3.61305624e-01 -1.08094168e+00 -2.62454122e-01
9.51218605e-02 9.32081491e-02 -4.28589940e-01 1.18030226e-02
-4.07226175e-01 1.35251969e-01 -7.65016675e-01 3.94490629e-01
3.04760784e-01 -2.54893273e-01 3.54698628e-01 6.60037935e-01
-4.99149233e-01 -1.45446792e-01 -1.65663338e+00 1.05161762e+00
-4.86222655e-01 1.50347175e-02 5.86877286e-01 -1.13392603e+00
8.10154974e-01 3.06384176e-01 8.03395689e-01 -5.58439314e-01
4.61250275e-01 5.00440359e-01 -1.75282404e-01 -1.07251860e-01
2.95441478e-01 -6.61406994e-01 6.06065132e-02 6.23848319e-01
-1.33281827e-01 -3.66940439e-01 4.18728113e-01 -3.44247788e-01
9.13619518e-01 1.07567430e-01 3.09961528e-01 -6.65051877e-01
3.97884965e-01 -2.52695233e-01 7.88021386e-01 8.01408768e-01
-4.50012952e-01 -3.01841348e-01 9.13815498e-01 -2.73205400e-01
-9.10885513e-01 -7.63429880e-01 3.36918165e-03 1.15201735e+00
2.73042560e-01 3.59141119e-02 -7.38038778e-01 -6.44109368e-01
1.79404393e-01 9.53695357e-01 -3.67582321e-01 -6.70560747e-02
-4.96123850e-01 -9.42747951e-01 3.60349528e-02 4.52974260e-01
5.23834050e-01 -1.18197739e+00 -1.27556741e+00 5.48688531e-01
2.83416212e-01 -4.59928215e-01 -5.13899744e-01 3.62799972e-01
-9.18046117e-01 -8.15749228e-01 -6.89417005e-01 -3.90489250e-01
9.13709283e-01 -2.46667370e-01 7.34276772e-01 -1.12506337e-01
-1.13024645e-01 5.38449883e-01 5.32272384e-02 -7.52343357e-01
-2.86122859e-01 -3.16866100e-01 2.34770060e-01 -1.29787311e-01
-3.50200832e-02 -4.38985407e-01 -7.26952791e-01 2.34670788e-01
-6.81812227e-01 -4.37342256e-01 3.87269318e-01 7.58433580e-01
6.55645430e-01 1.76066160e-01 8.52737904e-01 -4.32556570e-01
9.97512043e-01 -4.07784730e-02 -1.47215831e+00 1.68913871e-01
-1.12558222e+00 4.37839121e-01 7.22380280e-01 -4.57786769e-01
-9.41042542e-01 -8.52084160e-02 1.84308454e-01 -2.09213421e-01
2.78285623e-01 4.45922613e-01 3.87827903e-01 -3.04532379e-01
3.89369190e-01 1.43168136e-01 2.76345313e-01 -1.80763707e-01
1.34193540e-01 2.64288753e-01 4.84443177e-03 -7.92242944e-01
2.54387408e-01 3.70130688e-02 1.77583724e-01 -4.93413121e-01
-5.85856855e-01 2.01167166e-01 3.03906146e-02 -5.13966858e-01
6.40735090e-01 -3.37235481e-01 -1.36456752e+00 2.06845790e-01
-6.47956252e-01 -5.24788797e-01 -5.52278996e-01 7.19892681e-01
-1.02563643e+00 6.81598559e-02 -4.05974329e-01 -1.41884756e+00
-2.38292590e-01 -1.01186347e+00 1.41611084e-01 5.04985273e-01
1.44047707e-01 -8.76281798e-01 -4.01044562e-02 6.21185936e-02
5.12447000e-01 2.61869282e-01 7.71487653e-01 -1.06843270e-01
-4.12167192e-01 -4.96428579e-01 1.43013999e-01 6.79921031e-01
-5.89013286e-02 -2.05087706e-01 -1.99222252e-01 -6.59886241e-01
3.45336735e-01 -4.27261651e-01 4.25233096e-01 7.88474679e-01
7.82566905e-01 -6.16829693e-01 4.60052378e-02 1.44802883e-01
1.61052668e+00 1.00203562e+00 2.84252048e-01 4.33961451e-01
-9.62022766e-02 2.11441964e-01 9.20876563e-01 1.10871065e+00
1.91627629e-02 7.30352774e-02 7.21560597e-01 3.21676910e-01
7.49781370e-01 8.25025812e-02 3.61964911e-01 3.16795111e-01
-1.32676750e-01 1.12087503e-02 -7.81817079e-01 6.16807103e-01
-1.92636430e+00 -7.87311137e-01 4.57033128e-01 2.55545187e+00
7.72175252e-01 5.33496618e-01 2.26000279e-01 -1.46165103e-01
7.50285268e-01 5.97457998e-02 -8.40127766e-01 -8.44685197e-01
4.47050542e-01 5.94514549e-01 8.56134951e-01 8.76540542e-01
-7.24108100e-01 6.79299414e-01 6.24322557e+00 8.14931095e-01
-1.08626163e+00 -1.32216007e-01 6.78095341e-01 -4.15969849e-01
4.97962572e-02 2.92565167e-01 -6.96287155e-01 7.97490776e-01
1.03566885e+00 -5.48041940e-01 8.66854787e-01 9.90641356e-01
4.25607651e-01 -7.85683870e-01 -6.11737549e-01 5.14934540e-01
-3.79239947e-01 -9.16475892e-01 -6.26909077e-01 3.92958373e-01
9.40088212e-01 -5.39828777e-01 2.27652416e-01 3.36524189e-01
7.01689482e-01 -8.27563345e-01 7.30876505e-01 3.59529883e-01
5.13561368e-01 -1.26071870e+00 5.65182030e-01 7.20590889e-01
-1.12612152e+00 -6.91693783e-01 -1.59210563e-01 -2.93308109e-01
3.47220778e-01 5.59667468e-01 -5.77692807e-01 2.21433237e-01
5.91944456e-01 -1.98986694e-01 2.16681510e-01 8.89203668e-01
-2.25240737e-01 5.00165559e-02 -5.03177166e-01 -4.82220173e-01
5.99021554e-01 -5.42995870e-01 4.33954418e-01 7.37471044e-01
4.90051389e-01 3.16534728e-01 4.13872778e-01 6.38386846e-01
9.12945569e-02 9.43496302e-02 -2.12975722e-02 -2.25056306e-01
3.61457527e-01 8.06254983e-01 -8.93826663e-01 -1.55056551e-01
-3.60856429e-02 4.91397560e-01 2.81940281e-01 3.26529443e-01
-9.37164783e-01 -7.80793011e-01 4.44331825e-01 1.45192951e-01
4.72790867e-01 -5.72838187e-01 -1.70984924e-01 -6.78249419e-01
1.34799689e-01 -6.82034373e-01 3.49227697e-01 -1.59619898e-01
-7.46821642e-01 3.01162809e-01 2.16198757e-01 -8.78217161e-01
-6.54688120e-01 -4.96947706e-01 -4.71565634e-01 9.79111135e-01
-1.25007749e+00 -1.83664456e-01 3.83180201e-01 4.13151354e-01
2.07893893e-01 -4.03727710e-01 5.33010423e-01 -2.19413824e-02
-5.02648592e-01 2.52161175e-01 5.99733472e-01 -3.65529120e-01
1.71130627e-01 -1.14791799e+00 -4.62702274e-01 7.80960143e-01
-7.31952250e-01 2.95563400e-01 9.52938914e-01 -6.74930513e-01
-1.06596255e+00 -6.30782664e-01 6.93657100e-01 2.70243376e-01
5.97358525e-01 2.63246834e-01 -3.05343390e-01 5.35088778e-01
1.94399267e-01 -4.73024957e-02 1.55702770e-01 -3.67044836e-01
4.48769271e-01 -3.80130857e-01 -1.45005155e+00 5.20860136e-01
4.04902488e-01 -8.91076103e-02 -7.83832371e-02 1.02256984e-01
2.20066383e-01 -3.06285590e-01 -1.00363398e+00 3.18531722e-01
4.83306795e-01 -9.08564985e-01 5.32457769e-01 -4.08336878e-01
-4.27049994e-02 -4.36674505e-02 -1.80792600e-01 -1.21320450e+00
-1.05068674e-02 -9.13838685e-01 -2.46033594e-01 4.75058883e-01
4.09114063e-01 -8.64334524e-01 8.39194357e-01 8.07935059e-01
1.81793928e-01 -1.04011440e+00 -1.35028160e+00 -9.33002234e-01
2.91590750e-01 -9.15335640e-02 -1.14703827e-01 3.96182358e-01
5.42688109e-02 -1.24326594e-01 -3.59501392e-01 -9.00384560e-02
9.38957036e-01 -1.21219799e-01 1.10132443e-02 -7.14147270e-01
-6.01961315e-01 -4.76531267e-01 1.76695153e-01 -7.86378801e-01
-4.58360799e-02 -2.28924781e-01 2.16784060e-01 -1.21215940e+00
-1.17325321e-01 -5.30906498e-01 -5.83773434e-01 3.61914217e-01
1.51797971e-02 -6.17681146e-01 2.48741239e-01 -4.15423453e-01
-5.68047822e-01 8.14617813e-01 1.10199618e+00 1.29343659e-01
-3.64256978e-01 3.87893170e-01 -4.63250041e-01 9.23234940e-01
9.73684072e-01 -7.00162232e-01 -5.08710861e-01 -3.44934106e-01
4.13294822e-01 7.91934013e-01 3.20166275e-02 -7.66323507e-01
3.70705724e-02 -9.78685558e-01 2.32723239e-03 -3.29010457e-01
4.47741389e-01 -5.07886350e-01 -7.45647326e-02 7.99298584e-01
-3.80771905e-01 5.74353218e-01 3.27548832e-02 6.22708261e-01
2.46896110e-02 -8.09765577e-01 1.18864870e+00 -5.66207230e-01
-1.47835076e-01 5.04875965e-02 -5.86436689e-01 1.18468277e-01
1.23032570e+00 1.38629824e-01 7.21235722e-02 -5.95524013e-01
-5.16800821e-01 4.77266282e-01 2.34220654e-01 -1.57007128e-01
5.06608129e-01 -8.24519575e-01 -3.99067432e-01 -2.95370936e-01
-6.43588901e-01 -2.14669362e-01 7.80856833e-02 6.11264527e-01
-5.92105448e-01 2.91559339e-01 -1.64160132e-01 -1.42216787e-01
-7.81107605e-01 2.23236203e-01 6.56773329e-01 -5.85333109e-01
1.29195839e-01 9.52664673e-01 -4.92912322e-01 -1.68563612e-02
2.45338485e-01 -1.20184436e-01 3.50611657e-01 -1.30019709e-01
3.20918471e-01 8.82613897e-01 -3.08267862e-01 8.49889740e-02
-3.26967001e-01 2.76789159e-01 1.04817124e-02 -8.36938679e-01
1.31938529e+00 -7.88497999e-02 -1.20188445e-01 2.73612529e-01
8.91023755e-01 -3.64636809e-01 -1.80970323e+00 1.14750199e-01
-1.98099390e-02 -4.96578157e-01 1.47725224e-01 -7.77131557e-01
-7.30303884e-01 5.98716080e-01 8.16519320e-01 2.65310764e-01
7.97956765e-01 -6.07536793e-01 2.85807788e-01 3.43347609e-01
5.73370099e-01 -1.91395140e+00 4.12134379e-01 6.05897009e-01
7.47893453e-01 -9.17993069e-01 3.64288628e-01 2.98920065e-01
-7.78034508e-01 1.00492740e+00 6.74430311e-01 -4.87517655e-01
7.81795740e-01 9.58394781e-02 -3.27677965e-01 1.96257293e-01
-6.44580007e-01 7.52789155e-02 -2.06703722e-01 5.68356179e-02
3.07750791e-01 3.64863202e-02 -9.86592174e-01 2.16556653e-01
1.60480395e-01 1.72709778e-01 4.36977476e-01 1.67253542e+00
-9.66828883e-01 -1.20912433e+00 -2.27681860e-01 4.54681724e-01
-7.10061491e-01 2.00610653e-01 1.66419208e-01 6.93423271e-01
-3.32824960e-02 1.09937906e+00 -1.04166165e-01 3.05609465e-01
1.23661898e-01 1.63738534e-01 5.92287421e-01 -9.38851312e-02
-4.82267797e-01 3.12049538e-01 2.38173287e-02 -4.47209924e-01
-1.78821459e-01 -5.94701409e-01 -1.40290892e+00 -4.11639780e-01
-1.17704451e-01 4.08561319e-01 9.05775547e-01 9.01701033e-01
1.34614885e-01 3.19783241e-01 9.42396045e-01 -6.51618004e-01
-1.21214032e+00 -5.90243459e-01 -8.32409620e-01 -1.68011561e-01
1.07421398e-01 -7.73792148e-01 -4.62413132e-01 -4.18598771e-01] | [4.277583122253418, 2.637983560562134] |
34978fb7-cad8-48a0-a893-a2fcbc12af76 | multi-objective-hyperparameter-optimization-1 | 2209.04340 | null | https://arxiv.org/abs/2209.04340v1 | https://arxiv.org/pdf/2209.04340v1.pdf | Multi-objective hyperparameter optimization with performance uncertainty | The performance of any Machine Learning (ML) algorithm is impacted by the choice of its hyperparameters. As training and evaluating a ML algorithm is usually expensive, the hyperparameter optimization (HPO) method needs to be computationally efficient to be useful in practice. Most of the existing approaches on multi-objective HPO use evolutionary strategies and metamodel-based optimization. However, few methods have been developed to account for uncertainty in the performance measurements. This paper presents results on multi-objective hyperparameter optimization with uncertainty on the evaluation of ML algorithms. We combine the sampling strategy of Tree-structured Parzen Estimators (TPE) with the metamodel obtained after training a Gaussian Process Regression (GPR) with heterogeneous noise. Experimental results on three analytical test functions and three ML problems show the improvement over multi-objective TPE and GPR, achieved with respect to the hypervolume indicator. | ['Gonzalo Nápoles', 'Inneke Van Nieuwenhuyse', 'Alejandro Morales-Hernández'] | 2022-09-09 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 3.64558287e-02 -2.47196808e-01 1.53757175e-02 -7.32097402e-02
-1.07364452e+00 -9.00361165e-02 2.42024288e-01 4.14937526e-01
-2.54154652e-01 1.23306191e+00 -4.37079817e-01 5.23362719e-02
-8.76097441e-01 -7.84744084e-01 -4.38166887e-01 -1.28662097e+00
3.07350438e-02 1.08077121e+00 -3.89934108e-02 3.20114017e-01
5.34431815e-01 4.83772963e-01 -1.55668747e+00 -2.53530949e-01
1.01669419e+00 1.06821454e+00 6.44008666e-02 8.18473995e-01
7.95493796e-02 1.03862099e-01 -9.37838793e-01 -2.98935801e-01
2.31457874e-02 -2.85252362e-01 -3.20157588e-01 -2.35606208e-02
-3.34789485e-01 4.11495566e-01 4.96399850e-01 9.70810831e-01
9.46499050e-01 4.04116899e-01 1.10939932e+00 -1.28393888e+00
6.20550737e-02 6.23637140e-01 -3.59249651e-01 -7.40030482e-02
-1.70789137e-01 2.81161994e-01 8.19685996e-01 -7.18129396e-01
2.20077425e-01 1.24911165e+00 7.14017510e-01 -3.00315097e-02
-1.53924978e+00 -3.91448587e-01 -5.00498533e-01 1.17352985e-01
-1.68168223e+00 -1.01816766e-01 6.23174906e-01 -5.82753778e-01
7.98854053e-01 3.23665470e-01 5.90661645e-01 5.23421049e-01
7.26540148e-01 3.19706410e-01 1.19699907e+00 -6.34345949e-01
8.81756008e-01 2.40838006e-01 2.16568485e-02 3.22105020e-01
9.52537358e-01 3.34088981e-01 -3.75380963e-01 -5.37560463e-01
5.99131107e-01 -7.24043548e-01 -4.87194434e-02 -5.68087578e-01
-7.40175426e-01 1.01663280e+00 -2.54726529e-01 2.32038274e-01
-5.35603702e-01 1.87905088e-01 3.04822922e-01 2.23538801e-02
3.70395839e-01 1.09162307e+00 -5.91765821e-01 -2.65595317e-01
-9.99988019e-01 3.28268617e-01 1.10178578e+00 4.56275970e-01
6.06719017e-01 3.31926435e-01 -3.99659157e-01 1.05038381e+00
4.94155586e-01 6.30835593e-01 3.39743584e-01 -9.83331501e-01
1.90080896e-01 4.47762042e-01 2.60348648e-01 -7.00229824e-01
-5.10458648e-01 -7.27690458e-01 -5.11945665e-01 4.85968500e-01
3.03162515e-01 -5.24232328e-01 -6.34140193e-01 1.04982769e+00
3.36772829e-01 -2.20312029e-01 1.53274119e-01 5.10724425e-01
5.02626121e-01 8.14270377e-01 1.66227117e-01 -6.68270648e-01
8.38069379e-01 -6.46196127e-01 -8.25978994e-01 1.35210797e-01
5.22392213e-01 -7.84945726e-01 7.98032582e-01 8.31900418e-01
-1.11992919e+00 -3.57915610e-01 -1.13202655e+00 8.82240891e-01
-1.50753006e-01 1.52948245e-01 2.20307007e-01 1.22231567e+00
-3.58438045e-01 7.76051581e-01 -6.23132467e-01 1.64139755e-02
1.55493036e-01 6.22295916e-01 1.44553676e-01 3.19585294e-01
-9.08751309e-01 1.08596265e+00 7.90547788e-01 2.08784625e-01
-9.10648346e-01 -8.04443896e-01 -4.87259269e-01 2.89435744e-01
5.49800992e-01 -5.64526796e-01 1.05532253e+00 -5.20901442e-01
-2.12575078e+00 2.10670233e-01 2.32577071e-01 -2.24742502e-01
6.11803710e-01 -1.35426119e-01 -1.56542405e-01 -5.94357140e-02
-4.45426553e-01 -8.07873011e-02 8.61827672e-01 -1.43934619e+00
-4.29759532e-01 -1.41535997e-01 -6.40853465e-01 4.31348942e-02
1.43423855e-01 -5.54390885e-02 -3.36766765e-02 -4.69839483e-01
-3.62367071e-02 -7.90336072e-01 -4.33188796e-01 -8.45881879e-01
-6.10609114e-01 -3.18729021e-02 3.31506193e-01 -6.12040937e-01
1.45541656e+00 -1.71385109e+00 3.38438720e-01 7.92682827e-01
-3.68689030e-01 -2.02744175e-02 3.06831241e-01 5.58412611e-01
1.78655326e-01 3.44939381e-01 -4.65356618e-01 -1.45594656e-01
2.70796865e-01 1.09727904e-01 3.64794552e-01 5.23452342e-01
-8.04091692e-02 4.37708169e-01 -5.07425368e-01 -7.03843415e-01
3.97059798e-01 3.40929776e-01 -3.58505845e-01 2.20251858e-01
-3.49000573e-01 3.29438567e-01 -6.37369871e-01 9.00597990e-01
4.99228388e-01 -1.75619900e-01 7.21843466e-02 -2.16762885e-01
-8.56886432e-02 -5.62889993e-01 -1.69658589e+00 7.62355447e-01
-6.68121576e-01 2.96014816e-01 1.09353952e-01 -9.52547491e-01
1.06976223e+00 4.45459098e-01 7.71092653e-01 -1.37336433e-01
3.74111533e-01 4.95770574e-01 8.64338577e-02 -4.43278044e-01
2.15898365e-01 -3.28595847e-01 6.45375252e-02 1.72315538e-01
2.86440134e-01 -6.32431269e-01 3.54199797e-01 -6.86001241e-01
7.19630659e-01 3.42882760e-02 5.83211899e-01 -5.92771411e-01
7.59382248e-01 -3.11909476e-03 6.71603739e-01 7.33975828e-01
1.29029661e-01 5.77676773e-01 6.21827960e-01 1.96476802e-02
-1.13485003e+00 -8.44152987e-01 -6.02512181e-01 6.23092890e-01
-1.41599655e-01 -1.02074020e-01 -7.80234516e-01 -1.00221328e-01
6.43607676e-02 1.20519912e+00 -2.46851921e-01 -1.21245258e-01
-4.76842880e-01 -1.57319331e+00 2.54029661e-01 1.84155464e-01
2.46483743e-01 -7.54367828e-01 -5.96538723e-01 6.56052947e-01
3.00874114e-01 -7.00255096e-01 2.44820416e-01 3.57130408e-01
-1.06224620e+00 -1.08462667e+00 -5.57916939e-01 -1.28733531e-01
3.73791426e-01 -1.02484512e+00 1.04266346e+00 -2.72465527e-01
-3.97474885e-01 5.03591716e-01 -9.86504480e-02 -6.59214437e-01
-6.93976700e-01 5.28121814e-02 -2.04383031e-01 1.13958633e-02
-6.76290989e-02 -3.13027978e-01 -1.58174172e-01 6.26349390e-01
-6.47260368e-01 -6.19523227e-01 7.66833246e-01 9.79465246e-01
1.05442476e+00 7.79258192e-01 6.11010671e-01 -6.86640203e-01
8.89141262e-01 -2.43590593e-01 -1.21795619e+00 6.70479119e-01
-1.10124505e+00 3.53775233e-01 3.26651752e-01 -4.56499606e-01
-1.32233822e+00 -7.28577003e-02 1.72192708e-01 -3.41555029e-01
3.65880807e-03 6.63175523e-01 -3.71067047e-01 -4.09473807e-01
5.94418764e-01 -2.52857149e-01 -1.35419562e-01 -4.65497613e-01
-2.18087688e-01 4.32742715e-01 3.59698795e-02 -1.07551432e+00
6.09237254e-01 -1.17889836e-01 7.83676684e-01 -9.99166071e-01
-3.26857775e-01 -2.58622587e-01 -2.97849625e-01 -5.35839021e-01
7.84157336e-01 -2.37860486e-01 -9.42092359e-01 4.46825296e-01
-7.31919706e-01 -2.50200629e-01 -4.65031117e-01 8.58475626e-01
-9.28025305e-01 1.11643836e-01 -6.30890776e-04 -1.29765141e+00
-4.00281429e-01 -1.44371104e+00 8.21137488e-01 3.93870264e-01
-1.98990628e-01 -1.28540790e+00 1.89886659e-01 3.51492137e-01
5.09361088e-01 5.25164962e-01 1.13956308e+00 -8.12922239e-01
-4.68307495e-01 -5.37857592e-01 4.83852655e-01 4.52777267e-01
-4.24635500e-01 5.41578054e-01 -8.68653536e-01 -1.86340153e-01
3.65517437e-01 -7.32572600e-02 3.36788058e-01 1.17186272e+00
1.08467257e+00 -8.47041905e-02 -3.31138998e-01 5.42802334e-01
1.93349993e+00 4.17944908e-01 4.88892108e-01 7.19028592e-01
8.04392770e-02 5.12494504e-01 7.73384452e-01 8.26323330e-01
-3.78653526e-01 5.06878436e-01 2.37204537e-01 4.06160861e-01
4.13466066e-01 1.53238669e-01 1.38926640e-01 4.11517888e-01
-3.10604602e-01 -5.20005345e-01 -1.23704982e+00 1.03118464e-01
-1.72995389e+00 -6.13319337e-01 -1.59229398e-01 2.67878723e+00
6.34252191e-01 1.65487275e-01 -1.20546155e-01 3.30413073e-01
9.89229202e-01 -2.95307040e-01 -4.71575648e-01 -6.38746619e-01
-2.28935525e-01 2.43380457e-01 7.84091055e-01 6.90827072e-01
-9.49102581e-01 2.33173847e-01 6.84427452e+00 1.23007500e+00
-7.66233802e-01 -2.90025640e-02 7.27590919e-01 -1.55224577e-01
-7.88875893e-02 -1.04969524e-01 -1.07931650e+00 3.40151280e-01
1.11938405e+00 -5.52820265e-01 3.48084301e-01 8.29114974e-01
4.78513539e-01 -6.26288652e-01 -8.13886404e-01 9.26473439e-01
-2.80648530e-01 -9.45647299e-01 -3.67386222e-01 -2.28324812e-03
1.07627034e+00 -5.20352423e-01 -3.96021828e-02 3.03594410e-01
1.28984883e-01 -1.34227753e+00 3.94625157e-01 8.32094729e-01
1.71850726e-01 -1.12452996e+00 1.15476656e+00 1.67015806e-01
-8.62281263e-01 -3.63339722e-01 -5.06623983e-01 5.68786085e-01
3.20327669e-01 1.11621940e+00 -8.91783953e-01 7.43947983e-01
6.01739705e-01 -2.37101555e-01 -4.22735512e-01 1.89341247e+00
4.51860093e-02 7.80017912e-01 -6.30052149e-01 -4.96854305e-01
-2.31284555e-02 -6.34908617e-01 1.07401550e+00 8.55008543e-01
7.12427795e-01 -4.19543386e-01 -1.14271186e-01 1.00991940e+00
4.56006646e-01 5.70407808e-01 9.44688395e-02 -1.12057455e-01
5.49193203e-01 1.05079472e+00 -6.26603484e-01 -2.31151748e-02
6.16822951e-02 6.88560158e-02 -3.23813558e-01 5.63066125e-01
-7.84559429e-01 -2.90108860e-01 2.51197487e-01 -6.32046461e-02
2.22285807e-01 8.42013163e-04 -6.90017521e-01 -4.18368697e-01
-6.46665841e-02 -8.40820134e-01 5.19125581e-01 -5.53804934e-01
-1.29908156e+00 2.37325490e-01 5.83486497e-01 -9.75798368e-01
-4.04479295e-01 -8.08024168e-01 -5.76687455e-01 9.85643029e-01
-1.01612067e+00 -6.29423201e-01 -1.36562549e-02 -1.24225013e-01
3.66371840e-01 -4.20503139e-01 5.87604761e-01 -7.42060915e-02
-8.04720104e-01 3.46089453e-01 7.85830736e-01 -7.92840183e-01
3.60786080e-01 -1.15496123e+00 -7.22661197e-01 3.32380414e-01
-6.02064848e-01 3.22596282e-02 1.24330127e+00 -7.43393719e-01
-1.25396073e+00 -7.17325449e-01 4.08854485e-01 -5.42379171e-02
5.16861379e-01 2.49292180e-01 -8.69562507e-01 1.31102130e-01
-4.04785899e-03 -4.31341439e-01 7.07850099e-01 1.22760415e-01
6.12598956e-01 -2.23660380e-01 -1.44596851e+00 3.38513583e-01
1.00073688e-01 1.90223947e-01 -3.30872416e-01 3.91240716e-01
9.42354873e-02 -1.95821300e-01 -1.57956839e+00 9.02261138e-01
2.86123872e-01 -6.17040396e-01 9.99051630e-01 -2.41282210e-01
-8.65791291e-02 -1.79393098e-01 -1.31383926e-01 -1.48481131e+00
6.18608594e-02 -6.68009996e-01 -1.03481919e-01 1.37418211e+00
6.20880365e-01 -6.65809572e-01 7.40565360e-01 8.73339713e-01
1.48768187e-01 -9.49929774e-01 -1.20718968e+00 -1.15142965e+00
2.26929709e-01 -2.85001367e-01 6.03584826e-01 3.16762269e-01
-4.60110009e-01 1.43095031e-01 -1.14329413e-01 3.06360215e-01
1.10335886e+00 -1.32129729e-01 5.09246230e-01 -1.66798925e+00
-4.74327594e-01 -7.78168082e-01 -2.67877489e-01 5.45183793e-02
-2.27050893e-02 -4.00104493e-01 1.92758411e-01 -1.28091693e+00
-1.57306418e-02 -4.44083989e-01 2.91538239e-03 -2.08326146e-01
-2.36653417e-01 -4.78764534e-01 -1.19977385e-01 8.67520273e-02
-2.18016520e-01 8.08082998e-01 1.14639091e+00 1.08116917e-01
-4.48640138e-01 4.10698920e-01 8.35083500e-02 7.01653361e-01
1.07155788e+00 -8.04980338e-01 -2.81677157e-01 3.45588565e-01
3.74401867e-01 2.97799885e-01 7.00206263e-03 -1.29590619e+00
1.79403469e-01 -3.04081708e-01 4.02894229e-01 -6.76067293e-01
3.37588876e-01 -7.52122819e-01 7.93654978e-01 4.36439008e-01
-1.97464511e-01 -1.15133978e-01 2.26972863e-01 5.92332065e-01
-3.57970983e-01 -1.20250249e+00 1.10818708e+00 2.51615923e-02
-2.29652256e-01 -4.05169651e-02 -5.21565020e-01 -1.26589350e-02
1.18176055e+00 -3.07421684e-01 -2.40312759e-02 -6.78875446e-02
-8.14023256e-01 2.43982613e-01 3.84988874e-01 -3.29735965e-01
3.86537910e-01 -9.41129744e-01 -7.20602572e-01 -1.65265471e-01
-2.75858521e-01 -1.09022766e-01 1.13543846e-01 8.22517216e-01
-7.42095053e-01 3.67709249e-01 5.05097657e-02 -5.81432402e-01
-1.12505996e+00 2.46172622e-01 8.87743771e-01 -6.37573421e-01
-3.36905220e-03 7.17992425e-01 -4.22500610e-01 -4.84970957e-01
1.98318094e-01 6.29730523e-02 -2.90119141e-01 7.37935379e-02
-4.77590524e-02 1.16675460e+00 1.91429958e-01 -2.62293458e-01
-1.57282818e-02 8.32607329e-01 7.60728896e-01 -4.31150705e-01
1.31753886e+00 4.15866636e-02 -1.92791998e-01 6.01062655e-01
8.93596888e-01 -1.56363472e-01 -9.58003402e-01 1.97406113e-01
3.59791964e-01 -3.69461328e-01 5.18671095e-01 -8.28201652e-01
-8.29579771e-01 5.14624417e-01 6.54380143e-01 5.22328168e-02
1.00543201e+00 -4.09970671e-01 -4.71589342e-02 6.12760723e-01
1.48457974e-01 -1.83732462e+00 4.71994979e-03 1.35794222e-01
1.05272853e+00 -1.01217377e+00 3.94672483e-01 -4.54684347e-01
-6.00546181e-01 1.47815895e+00 4.61343795e-01 2.30432898e-01
9.83002245e-01 4.25269902e-01 -3.15675825e-01 -1.55726805e-01
-6.16941392e-01 1.98780790e-01 4.17915136e-01 4.05174524e-01
1.37943685e-01 7.96214864e-02 -5.66484034e-01 6.87177896e-01
7.63276741e-02 -4.42292467e-02 2.08646208e-01 9.05423224e-01
-6.78618193e-01 -1.07704914e+00 -1.05725861e+00 5.85196853e-01
-5.24790704e-01 1.55695692e-01 1.77104369e-01 1.02961802e+00
-1.37451261e-01 9.32998121e-01 -2.48218477e-01 8.84927213e-02
3.15097362e-01 2.51967758e-01 4.63955283e-01 -3.37773055e-01
-6.42309010e-01 2.10674897e-01 4.66505885e-01 -2.35536054e-01
-1.58749312e-01 -9.78653491e-01 -1.03793180e+00 -9.04433429e-02
-7.15407312e-01 4.17151898e-01 1.01010132e+00 8.25947821e-01
-2.07530439e-01 6.72036111e-01 5.37957311e-01 -6.22470200e-01
-9.19756293e-01 -9.06322300e-01 -9.31850672e-01 -2.35218912e-01
-3.39277238e-01 -1.04547954e+00 -8.28798473e-01 -3.79893661e-01] | [6.0888872146606445, 3.620873212814331] |
1f770eb8-8bb4-46ca-bf4f-70d771863452 | learnable-acoustic-frontends-in-bird-activity | 2210.00889 | null | https://arxiv.org/abs/2210.00889v1 | https://arxiv.org/pdf/2210.00889v1.pdf | Learnable Acoustic Frontends in Bird Activity Detection | Autonomous recording units and passive acoustic monitoring present minimally intrusive methods of collecting bioacoustics data. Combining this data with species agnostic bird activity detection systems enables the monitoring of activity levels of bird populations. Unfortunately, variability in ambient noise levels and subject distance contribute to difficulties in accurately detecting bird activity in recordings. The choice of acoustic frontend directly affects the impact these issues have on system performance. In this paper, we benchmark traditional fixed-parameter acoustic frontends against the new generation of learnable frontends on a wide-ranging bird audio detection task using data from the DCASE2018 BAD Challenge. We observe that Per-Channel Energy Normalization is the best overall performer, achieving an accuracy of 89.9%, and that in general learnable frontends significantly outperform traditional methods. We also identify challenges in learning filterbanks for bird audio. | ['Naomi Harte', 'Mark Anderson'] | 2022-10-03 | null | null | null | null | ['bird-audio-detection', 'activity-detection'] | ['audio', 'computer-vision'] | [ 1.90282956e-01 -5.75048149e-01 3.72905701e-01 -2.64268249e-01
-1.05974603e+00 -9.41627681e-01 1.31982192e-01 1.36337131e-01
-9.64022517e-01 3.93055141e-01 3.52591932e-01 2.92035133e-01
-6.64676726e-02 -2.73905277e-01 -5.31061411e-01 -7.47719944e-01
-7.67034054e-01 -2.68930614e-01 2.16949105e-01 -1.51884884e-01
-1.63153291e-01 3.06698591e-01 -1.68867958e+00 6.47957772e-02
-9.01374221e-03 8.34772348e-01 -1.04339354e-01 1.46904373e+00
7.42220044e-01 4.29088682e-01 -1.31930554e+00 -1.63785979e-01
4.10894424e-01 -3.00132722e-01 -3.19423974e-01 -6.59096122e-01
9.57059026e-01 -5.37491441e-01 -2.16839779e-02 8.62228930e-01
1.10967970e+00 3.87974709e-01 5.06181479e-01 -9.26727355e-01
3.83959524e-02 1.09107995e+00 -4.93028730e-01 9.21287537e-01
2.59452254e-01 4.33620721e-01 1.30108273e+00 -2.12963238e-01
-4.49808568e-01 9.57311749e-01 1.32777941e+00 5.60739338e-01
-1.56375766e+00 -1.17763853e+00 1.91062000e-02 2.92945970e-02
-1.44350207e+00 -9.99068141e-01 5.42556822e-01 -5.88451982e-01
1.08490133e+00 4.12769109e-01 6.89880967e-01 1.03388619e+00
-1.99315667e-01 2.56977916e-01 7.91966975e-01 6.90860115e-03
2.12399498e-01 -1.06192656e-01 -6.59909546e-02 3.78385544e-01
1.05357148e-01 5.17287374e-01 -1.16117561e+00 -6.75101340e-01
2.93449670e-01 -5.80159664e-01 -3.70823354e-01 4.59615827e-01
-9.45701420e-01 5.78674853e-01 4.22411174e-01 7.53725991e-02
-1.28972784e-01 7.20873773e-01 5.77320039e-01 3.99055928e-01
3.88934016e-01 1.00221443e+00 -5.88039696e-01 -4.38245028e-01
-1.07547104e+00 1.83629498e-01 1.08486915e+00 6.02487087e-01
3.41199249e-01 7.06528664e-01 9.61024463e-02 1.09041381e+00
2.68563241e-01 8.38989437e-01 4.48381722e-01 -8.67415667e-01
-2.31007300e-02 -2.45057508e-01 8.96098018e-02 -7.60374010e-01
-6.35764480e-01 -9.19441044e-01 -2.04769552e-01 -7.68334642e-02
4.03377146e-01 -6.83241010e-01 -5.54636776e-01 1.94057381e+00
2.52522826e-01 5.64099610e-01 -3.36740971e-01 9.05219316e-01
9.16255176e-01 7.52802849e-01 2.34418869e-01 3.90106477e-02
1.46192741e+00 -5.96144497e-01 -3.95419061e-01 -5.46383679e-01
1.54849261e-01 -8.11217844e-01 9.66155648e-01 5.96520007e-01
-5.64328492e-01 -5.33534229e-01 -1.45965171e+00 3.41093719e-01
-2.24611789e-01 1.28638640e-01 5.46685159e-01 1.23544574e+00
-1.02996325e+00 2.80570984e-01 -1.05212545e+00 -9.41525698e-02
1.63720091e-04 5.43755889e-01 -1.17961355e-01 9.96070862e-01
-8.33115518e-01 3.93213242e-01 -3.52097116e-03 1.03633903e-01
-1.57121706e+00 -1.17895794e+00 -6.44706547e-01 7.11198822e-02
1.92868784e-01 -7.19672441e-02 1.97175968e+00 -7.21581280e-01
-1.65867221e+00 6.94060445e-01 5.48707962e-01 -1.04261744e+00
1.53788671e-01 -5.19431710e-01 -6.57444596e-01 -1.78954704e-03
5.07774064e-04 6.25686646e-01 1.06156158e+00 -6.75471008e-01
-1.11907279e+00 -7.19022974e-02 -2.32804805e-01 7.69482180e-02
-6.00381553e-01 2.42281407e-01 2.04744533e-01 -8.01417351e-01
-4.68771696e-01 -7.80548990e-01 -5.70771433e-02 -8.25465992e-02
1.24021351e-01 2.62789220e-01 4.88240898e-01 -6.93618834e-01
1.11941206e+00 -2.36282301e+00 -2.47464865e-01 -7.76570141e-02
1.74009860e-01 1.42534673e-01 -2.04684928e-01 1.39784053e-01
3.82727027e-01 -2.67792866e-02 -1.13708548e-01 -2.04665780e-01
7.68329948e-02 -1.29748419e-01 -2.84525633e-01 7.97150254e-01
-8.47347528e-02 2.28720248e-01 -8.28101754e-01 -1.79772139e-01
9.39298570e-02 6.23438835e-01 -1.03610945e+00 5.33282280e-01
2.52365232e-01 1.57157928e-01 1.22307204e-01 6.91117227e-01
2.81190306e-01 7.02044129e-01 -3.67501318e-01 -1.60594136e-01
-3.25391263e-01 7.23831415e-01 -1.17713892e+00 1.50247943e+00
-7.52842546e-01 1.06395280e+00 8.95361602e-01 -6.71370268e-01
6.66318297e-01 3.54331136e-01 3.98193806e-01 -1.04030088e-01
9.34749395e-02 1.12564795e-01 4.92752284e-01 -1.58278406e-01
5.01207829e-01 -2.79196233e-01 -2.56012082e-01 2.58515298e-01
6.54138803e-01 -4.31707114e-01 -1.95545241e-01 -3.62610400e-01
1.18080795e+00 -4.70265448e-01 2.84538358e-01 -5.91446698e-01
5.34382351e-02 -2.87550986e-01 5.97957015e-01 1.03527582e+00
-5.49028456e-01 2.05814004e-01 -1.44544110e-01 -1.77524328e-01
-2.17731833e-01 -1.20127749e+00 -4.21832472e-01 2.00258088e+00
-2.98829645e-01 -6.47679329e-01 -9.59324062e-01 1.35674244e-02
-1.88499987e-02 4.60502446e-01 -6.50911570e-01 -4.25858438e-01
-4.33589488e-01 -1.12405980e+00 1.59310579e+00 5.91896653e-01
2.06898838e-01 -7.74546087e-01 -1.08909452e+00 4.75356340e-01
5.46093658e-02 -1.07662761e+00 -7.59938955e-01 8.43825936e-01
-4.43397552e-01 -5.93602002e-01 -1.35266259e-01 -6.37632251e-01
-9.63064209e-02 3.06400001e-01 1.03911197e+00 -2.14699849e-01
-6.32987559e-01 6.57911241e-01 -1.52445644e-01 -9.92003500e-01
-3.67552102e-01 2.95495450e-01 7.01317430e-01 -1.05700873e-01
3.06702763e-01 -8.25143874e-01 -5.77471614e-01 3.41125011e-01
-5.15369475e-01 -8.43518555e-01 2.89899021e-01 7.43844211e-01
2.96117187e-01 1.95452422e-02 8.00678849e-01 -2.65286118e-01
6.98413610e-01 -3.30150813e-01 -9.14712429e-01 -3.60693008e-01
-8.60009342e-02 -1.37637466e-01 7.43873596e-01 -7.57631719e-01
-5.80449581e-01 2.95889527e-01 -5.53148508e-01 1.87179744e-01
-3.98925513e-01 -2.45922115e-02 -1.33433407e-02 -3.34563017e-01
1.19108284e+00 -1.08293757e-01 -4.80618149e-01 -6.70518637e-01
5.60310483e-02 7.20902145e-01 1.05871880e+00 -3.37035298e-01
8.45376134e-01 1.25660479e-01 -3.86476040e-01 -1.45828164e+00
-5.99476099e-01 -7.14072108e-01 -1.86663687e-01 -2.43490130e-01
7.41912246e-01 -1.32580364e+00 -9.54015076e-01 3.89963269e-01
-5.99560678e-01 -4.78776366e-01 -3.85778308e-01 8.55240405e-01
-3.43576610e-01 -2.74880707e-01 -5.65152645e-01 -1.04336393e+00
-6.28999114e-01 -9.10809338e-01 1.08526623e+00 2.37846002e-01
-3.00456285e-01 -6.72689021e-01 5.64854264e-01 1.80418909e-01
5.67227840e-01 1.23034067e-01 4.95576188e-02 -6.69350266e-01
3.09252515e-02 -2.78984606e-01 5.85689783e-01 4.31748867e-01
2.59073526e-01 5.90549745e-02 -1.88793361e+00 -4.03099179e-01
1.04773819e-01 -3.21953923e-01 8.84478390e-01 7.26297855e-01
7.99011767e-01 -5.39563715e-01 -2.87528820e-02 9.54153299e-01
8.74510109e-01 2.59623360e-02 -2.31384605e-01 -5.90012670e-02
3.19953024e-01 4.49157268e-01 1.81371957e-01 5.13624370e-01
4.59033959e-02 6.53228581e-01 3.09784949e-01 -2.10046545e-02
-4.36746813e-02 -4.91777033e-01 8.20406973e-01 7.52418458e-01
9.98504758e-02 -2.28193775e-01 -7.47811675e-01 6.90217793e-01
-1.18028879e+00 -1.07308483e+00 2.20210806e-01 2.35114813e+00
9.56226170e-01 5.27053438e-02 7.41281152e-01 2.33147070e-01
4.16835964e-01 7.86533430e-02 -3.18589985e-01 -5.36779702e-01
2.57564969e-02 3.05370450e-01 1.08565724e+00 6.82560742e-01
-1.48531866e+00 6.41167581e-01 7.39304638e+00 7.23824322e-01
-1.09168291e+00 2.18246639e-01 -6.78728372e-02 -8.39348912e-01
4.57653046e-01 -4.88495767e-01 -1.03168654e+00 3.13425630e-01
1.55569458e+00 -7.11727366e-02 7.21355796e-01 8.82682562e-01
1.49254620e-01 -5.04955202e-02 -1.21774173e+00 1.14536643e+00
8.83601457e-02 -6.09391749e-01 -6.29070878e-01 -2.55725998e-02
2.70205170e-01 5.12745142e-01 2.29007453e-01 3.37782800e-01
4.90020573e-01 -1.09528363e+00 9.63340700e-01 -5.24274260e-02
5.35400569e-01 -7.85762548e-01 5.32726586e-01 3.17666203e-01
-1.62592328e+00 -1.57512933e-01 -3.89547318e-01 -3.65025461e-01
-8.49926174e-02 1.54526129e-01 -1.21096206e+00 -2.85183668e-01
1.30124521e+00 3.96527678e-01 -8.18275094e-01 1.44068491e+00
-8.84345267e-03 1.67402077e+00 -1.06968236e+00 -3.17388415e-01
9.45389569e-02 3.89142245e-01 8.88863564e-01 1.95492435e+00
1.27571160e-02 -4.49406207e-02 -1.74550414e-02 5.72651923e-01
-2.92433262e-01 -6.80292696e-02 -5.58573484e-01 4.78546359e-02
5.95818341e-01 1.47405863e+00 -4.07519221e-01 2.34237224e-01
-1.76801369e-01 3.27407807e-01 -1.04883812e-01 -5.87873310e-02
-8.91273499e-01 -5.28528631e-01 1.35235012e+00 -1.14533000e-01
2.20701531e-01 -1.52703017e-01 2.45335579e-01 -7.41814435e-01
-3.34721357e-01 -9.58389223e-01 5.69952667e-01 -5.75634055e-02
-1.27276158e+00 5.52919149e-01 2.15254128e-02 -1.34945726e+00
-2.82838643e-01 -3.93310726e-01 -4.84938562e-01 5.35064042e-01
-1.11080670e+00 -7.26326764e-01 -2.17199102e-01 2.50981122e-01
5.61602592e-01 -1.09654032e-02 9.27573085e-01 6.01712942e-01
-3.72117072e-01 1.11453712e+00 -3.93696204e-02 1.63180709e-01
8.70653629e-01 -1.47764945e+00 6.01482332e-01 9.81214046e-01
7.56853342e-01 3.33203852e-01 1.11326146e+00 -8.18237066e-02
-1.27093220e+00 -1.08170795e+00 -1.39958009e-01 -6.11955702e-01
8.80406380e-01 -1.05113530e+00 -6.82200372e-01 4.82214004e-01
9.08038914e-02 -1.29951745e-01 1.38383234e+00 2.48737916e-01
-3.71528596e-01 -5.68272889e-01 -9.02484655e-01 3.78394455e-01
9.01072323e-01 -7.28685856e-01 -5.26442528e-01 4.50406931e-02
6.36191726e-01 -2.25422084e-01 -9.10138845e-01 2.45748281e-01
1.02391922e+00 -4.92847562e-01 1.09390295e+00 -4.81842220e-01
-4.33492869e-01 -4.78336602e-01 -1.73772782e-01 -1.66266191e+00
-4.07459229e-01 -1.00380087e+00 2.00111568e-01 1.46451926e+00
5.92698991e-01 -5.30357778e-01 4.00976479e-01 7.32522681e-02
-8.39314610e-02 1.80459008e-01 -9.18026924e-01 -9.67005193e-01
-1.18191421e-01 -6.41360164e-01 3.58470172e-01 4.20221537e-01
-1.41663268e-01 5.90730190e-01 -6.48110569e-01 7.30959892e-01
7.50415385e-01 -2.50616521e-01 9.34454083e-01 -1.16174102e+00
-8.42223406e-01 -4.41795319e-01 -7.78483927e-01 -1.05348051e+00
4.10201680e-03 -4.74313706e-01 6.00999773e-01 -4.21764463e-01
-4.26258981e-01 -9.68473870e-03 -3.74737591e-01 4.44518924e-01
-1.72027871e-01 6.11292660e-01 -1.44507140e-01 -3.35311055e-01
-3.76832753e-01 2.61117071e-01 1.90714464e-01 -2.03372642e-01
-4.38113779e-01 3.98532838e-01 -6.89210653e-01 9.00292754e-01
8.29670548e-01 -7.81735361e-01 -2.58624732e-01 -6.04120195e-01
-7.01970607e-02 -2.65060306e-01 3.96280169e-01 -1.56138027e+00
4.52036202e-01 1.56194791e-01 2.65604019e-01 -7.46788904e-02
6.29976869e-01 -7.39292979e-01 -2.44393162e-02 5.34583628e-01
-7.08950818e-01 -1.59972191e-01 8.29374611e-01 6.83485091e-01
-1.46205842e-01 -1.12846665e-01 9.57475543e-01 2.24973440e-01
-4.94530797e-01 -1.39073700e-01 -1.02772808e+00 4.69019234e-01
2.79815018e-01 3.58877517e-02 -2.11384282e-01 -4.08498228e-01
-3.35383952e-01 7.27796778e-02 -1.39930919e-01 5.09551585e-01
4.02251855e-02 -6.89936876e-01 -9.75429952e-01 3.61641377e-01
2.52870470e-01 -3.99948746e-01 6.19849898e-02 5.88332713e-01
-5.39432108e-01 1.49028689e-01 -1.81856513e-01 -7.60996878e-01
-1.53444588e+00 -1.33220688e-01 8.92513514e-01 5.10204554e-01
-2.47034892e-01 1.48897219e+00 2.85277069e-01 -2.99513042e-01
7.06552267e-01 -5.76767027e-01 -8.05090889e-02 9.89260674e-02
8.13005328e-01 4.95089859e-01 2.69707441e-01 -6.09579206e-01
-5.57221949e-01 3.99028242e-01 1.92790896e-01 -4.03003573e-01
1.31522346e+00 1.56367153e-01 4.12580371e-01 6.56939030e-01
9.93175924e-01 4.25048918e-01 -1.33841097e+00 2.54766382e-02
-1.50181353e-01 -2.60760009e-01 5.70238233e-01 -6.57509208e-01
-8.49334896e-01 9.68621790e-01 1.09619975e+00 5.03573596e-01
1.18684196e+00 -1.00792401e-01 5.76930940e-01 7.56301343e-01
2.64952809e-01 -9.79532301e-01 -2.84130931e-01 1.00868411e-01
6.58256292e-01 -1.12510133e+00 -6.23921677e-02 1.76892564e-01
-3.54673505e-01 6.94007218e-01 6.48320198e-01 -1.23866677e-01
7.18958855e-01 1.05844402e+00 4.02600259e-01 -5.39585724e-02
-9.17443573e-01 -2.80214190e-01 3.34613323e-01 8.70158613e-01
4.22793984e-01 3.48750859e-01 2.18764037e-01 8.17912638e-01
-1.13264537e+00 -8.19964230e-01 3.02993983e-01 6.49111271e-01
-6.15434408e-01 -3.70726615e-01 -5.41379273e-01 4.01260257e-01
-9.37461495e-01 -1.75338894e-01 -8.29437673e-01 3.06990236e-01
-9.32702504e-04 1.43885958e+00 -1.38385147e-02 -7.07604408e-01
5.55157661e-01 -1.81780934e-01 1.88783288e-01 -7.58259594e-01
-1.43855274e+00 3.25708389e-01 1.60977125e-01 -1.27989039e-01
-4.19246137e-01 -5.64623177e-01 -7.65786648e-01 -3.90989557e-02
-8.12068880e-01 5.51972330e-01 7.55602479e-01 1.96942419e-01
1.72537770e-02 5.26898563e-01 5.76321125e-01 -9.12343025e-01
-7.85263419e-01 -9.95936573e-01 -7.22856462e-01 -2.54122585e-01
8.93330932e-01 -4.23004776e-01 -9.81757581e-01 3.16839099e-01] | [15.210835456848145, 5.315835952758789] |
0c1bf3b0-9d4e-45e7-97b5-041c1b1cf035 | effect-of-source-language-on-amr-structure | null | null | https://aclanthology.org/2022.law-1.12 | https://aclanthology.org/2022.law-1.12.pdf | Effect of Source Language on AMR Structure | The Abstract Meaning Representation (AMR) annotation schema was originally designed for English. But the formalism has since been adapted for annotation in a variety of languages. Meanwhile, cross-lingual parsers have been developed to derive English AMR representations for sentences from other languages—implicitly assuming that English AMR can approximate an interlingua. In this work, we investigate the similarity of AMR annotations in parallel data and how much the language matters in terms of the graph structure. We set out to quantify the effect of sentence language on the structure of the parsed AMR. As a case study, we take parallel AMR annotations from Mandarin Chinese and English AMRs, and replace all Chinese concepts with equivalent English tokens. We then compare the two graphs via the Smatch metric as a measure of structural similarity. We find that source language has a dramatic impact on AMR structure, with Smatch scores below 50% between English and Chinese graphs in our sample—an important reference point for interpreting Smatch scores in cross-lingual AMR parsing. | ['Nathan Schneider', 'Yifu Mu', 'Wai Ching Leung', 'Shira Wein'] | null | null | null | null | lrec-law-2022-6 | ['amr-parsing'] | ['natural-language-processing'] | [ 2.48891443e-01 4.79807228e-01 -1.78814694e-01 -4.75612015e-01
-6.60112858e-01 -8.68073940e-01 6.43186510e-01 4.40305740e-01
-2.88725704e-01 3.81308317e-01 6.78684950e-01 -5.41860640e-01
2.70039827e-01 -7.86337793e-01 -2.97368228e-01 -1.44785717e-01
3.38544399e-01 3.61171216e-01 4.03269082e-02 -3.05578411e-01
6.36012554e-02 1.61294088e-01 -7.69884884e-01 4.63057131e-01
6.93919241e-01 2.00718492e-01 2.48775348e-01 5.42210758e-01
-6.96926951e-01 1.05806518e+00 -1.02162039e+00 -9.72525716e-01
-6.86863298e-03 -6.90652490e-01 -1.10737693e+00 9.46542695e-02
4.37866330e-01 4.50786859e-01 -3.25617999e-01 1.20817697e+00
1.09468378e-01 -1.97455540e-01 6.20740831e-01 -8.47393453e-01
-8.57007444e-01 1.50433433e+00 -5.96158743e-01 3.01594913e-01
6.45619452e-01 -2.01438770e-01 1.53785789e+00 -5.13868332e-01
9.62566853e-01 1.45892251e+00 7.14937150e-01 6.23100698e-01
-1.30184829e+00 -1.40726537e-01 1.91339672e-01 -3.92554909e-01
-1.54640734e+00 -3.07902485e-01 6.78359151e-01 -3.21399540e-01
1.34305716e+00 2.14256600e-01 4.31608707e-01 8.32090139e-01
2.19561636e-01 4.94185299e-01 1.00545442e+00 -9.83501971e-01
4.28379215e-02 5.80895739e-03 4.29342330e-01 5.54724991e-01
5.22482038e-01 -6.88708544e-01 -3.16897154e-01 -8.27083960e-02
5.22095382e-01 -4.59488899e-01 -5.69366515e-02 1.78049505e-01
-9.96341586e-01 1.06518221e+00 1.47153586e-01 8.02302182e-01
-8.77728872e-03 3.38709384e-01 6.63452864e-01 4.81509596e-01
4.79950666e-01 4.78831977e-01 -5.08491576e-01 -4.73569334e-02
-3.41251731e-01 -1.11620054e-01 9.24181461e-01 1.21884549e+00
5.74270964e-01 1.65551111e-01 1.99344069e-01 1.09269118e+00
1.78025052e-01 2.58236289e-01 7.21719623e-01 -9.12416816e-01
6.48029208e-01 6.89648390e-01 -3.37881625e-01 -8.28591168e-01
-2.36894578e-01 -2.34053433e-01 -4.30608392e-01 -3.89119357e-01
6.14082813e-01 -1.28963813e-01 -5.87797284e-01 1.80401373e+00
8.36730003e-02 -8.74627173e-01 5.25978565e-01 5.08270800e-01
7.90431440e-01 5.93320370e-01 4.38809633e-01 -1.72673762e-02
1.74488008e+00 -7.04662323e-01 -7.22104609e-01 -6.72655046e-01
1.54939520e+00 -7.71958232e-01 1.44889522e+00 -1.87194511e-01
-1.00315619e+00 -4.04178679e-01 -9.19068933e-01 -3.69082004e-01
-4.75627512e-01 8.53493288e-02 5.93934596e-01 9.05400217e-01
-9.71428216e-01 3.76134932e-01 -7.35536516e-01 -5.69688797e-01
-5.52848093e-02 -9.52237919e-02 -4.86996680e-01 -2.44606614e-01
-1.19840908e+00 1.19429469e+00 7.21496165e-01 -2.30588734e-01
-7.81906620e-02 -2.40882173e-01 -1.18001616e+00 -1.34531841e-01
2.61025280e-01 -3.57192874e-01 1.32644129e+00 -1.27926779e+00
-1.03093505e+00 1.29312074e+00 -5.98611124e-02 -3.60946596e-01
2.61786152e-02 1.27865029e-02 -6.96063936e-01 -1.33362472e-01
2.44675666e-01 3.34928125e-01 4.31201220e-01 -9.22247827e-01
-5.61965704e-01 -5.57998121e-01 2.70430952e-01 3.34011763e-01
8.13867450e-02 5.97152829e-01 -2.98063159e-01 -8.45459878e-01
2.56145656e-01 -8.51579607e-01 -1.25226155e-01 -5.88510811e-01
-1.00300789e-01 -5.40031314e-01 4.30610448e-01 -8.43382239e-01
1.37043428e+00 -2.16745925e+00 1.33471027e-01 1.02569371e-01
2.44043246e-01 -1.71244606e-01 -2.47061834e-01 6.08417392e-01
-2.00821698e-01 6.70487940e-01 -4.73510712e-01 -1.83791801e-01
1.50533855e-01 8.18169773e-01 -1.31757349e-01 3.85530412e-01
3.34264904e-01 9.78005290e-01 -8.31770778e-01 -4.94687438e-01
-1.94974780e-01 3.62484753e-01 -2.23005354e-01 4.62057143e-02
-1.11139171e-01 1.44609615e-01 -3.12921911e-01 5.03148675e-01
1.51530102e-01 4.88536097e-02 9.09113050e-01 1.22052684e-01
-8.33967328e-02 8.93758833e-01 -7.19869256e-01 1.75630271e+00
-6.89964831e-01 5.96509218e-01 -1.07232779e-01 -8.17572057e-01
1.03511488e+00 2.81797200e-01 -1.88190877e-01 -6.49648249e-01
-6.20971434e-02 4.95160282e-01 6.40857637e-01 7.40630552e-02
8.60424340e-01 -2.46399134e-01 -7.51121819e-01 3.67889971e-01
5.40180206e-02 -5.21060824e-01 3.54782313e-01 3.25661272e-01
1.22840261e+00 1.10107459e-01 1.00092423e+00 -5.17334104e-01
5.43010473e-01 2.32469901e-01 5.20441532e-01 5.59348166e-01
2.13347688e-01 4.70846742e-01 8.97934854e-01 -5.75468898e-01
-1.16475403e+00 -9.08060849e-01 -2.30617166e-01 1.01396418e+00
-2.61861086e-01 -9.57642019e-01 -8.91243339e-01 -9.63645101e-01
-3.40924799e-01 1.46172965e+00 -3.82216066e-01 -9.60512236e-02
-1.11526358e+00 -6.36902511e-01 8.23754728e-01 6.92067027e-01
2.91169174e-02 -9.29437399e-01 -3.69421691e-01 2.98165590e-01
-1.14910074e-01 -1.47646308e+00 -3.51676375e-01 2.11046144e-01
-6.72003865e-01 -7.18885183e-01 -2.75326520e-01 -7.66705394e-01
5.96924961e-01 -1.81157216e-01 1.56671035e+00 3.94995451e-01
2.36253347e-02 4.05025840e-01 -6.75468564e-01 -2.99295962e-01
-1.12215912e+00 2.49832794e-01 -2.80185789e-01 -5.48748910e-01
7.30275035e-01 -1.57446861e-01 9.71447453e-02 1.39192045e-01
-9.19690311e-01 -1.91637173e-01 2.86530316e-01 3.68111640e-01
5.61175942e-01 -1.68851554e-01 2.96283692e-01 -1.45212221e+00
5.70867121e-01 -4.00359154e-01 -5.27534187e-01 2.11625293e-01
-3.99244547e-01 3.02721739e-01 7.54722893e-01 -1.04051776e-01
-1.01580215e+00 -2.14095607e-01 -4.51404512e-01 3.47299308e-01
-7.57240877e-02 6.63789392e-01 -3.38464379e-01 3.11845541e-01
5.22229195e-01 -1.76607549e-01 -4.34134692e-01 -4.59736139e-01
6.76766276e-01 8.29624653e-01 4.32321995e-01 -9.15449202e-01
4.53954130e-01 7.56321615e-03 -1.82393163e-01 -1.12153769e+00
-1.01054251e+00 -2.52366483e-01 -6.93006516e-01 2.49020547e-01
1.19967711e+00 -9.55080688e-01 8.56240317e-02 -2.48574857e-02
-1.44062603e+00 -3.37846398e-01 -5.73898673e-01 3.72560948e-01
-3.24798763e-01 6.03249729e-01 -9.98608291e-01 -5.26171863e-01
-2.10875288e-01 -9.81143057e-01 9.80553210e-01 -2.97440380e-01
-8.35450470e-01 -1.27970362e+00 1.43975988e-01 1.14593379e-01
1.41910851e-01 1.63533852e-01 1.45440686e+00 -1.00470436e+00
2.03955695e-02 -1.58859432e-01 -1.82632446e-01 1.57553405e-01
2.24462405e-01 -2.02668443e-01 -6.95217252e-01 -1.36309549e-01
6.81506768e-02 -2.57127844e-02 5.14612615e-01 5.67899793e-02
6.79021776e-01 -3.34122270e-01 -1.15229256e-01 1.71023235e-01
1.40587103e+00 8.45364109e-02 7.68534601e-01 2.64454961e-01
9.77481544e-01 7.48283684e-01 2.83948541e-01 -2.77847230e-01
4.27262068e-01 5.95856249e-01 -3.87362689e-02 1.96926802e-01
-4.65153575e-01 -4.11780804e-01 7.56875336e-01 1.59597683e+00
-7.49957561e-03 -4.34773266e-01 -1.17164040e+00 4.06459242e-01
-1.51956177e+00 -3.91612381e-01 -3.95691067e-01 2.22376466e+00
7.03131557e-01 2.74082452e-01 -1.64868355e-01 -1.43535405e-01
6.63092017e-01 3.75706851e-01 2.46537402e-01 -9.90209937e-01
-3.24631095e-01 3.27037603e-01 4.38803285e-01 6.27583921e-01
-5.99042892e-01 1.26902628e+00 6.39745760e+00 6.62508905e-01
-6.29770637e-01 2.16939285e-01 4.30911303e-01 7.16250718e-01
-6.53584063e-01 4.62251723e-01 -7.31486619e-01 2.07441300e-01
1.31367123e+00 -1.92564294e-01 2.23299220e-01 7.89306939e-01
-3.19197059e-01 1.93111256e-01 -1.20585048e+00 6.64607882e-01
1.29768122e-02 -9.54463959e-01 1.72620699e-01 -8.11968669e-02
3.30449641e-01 2.62399733e-01 -5.66426039e-01 2.18923315e-01
6.97140098e-01 -9.23817515e-01 8.82183790e-01 7.67288380e-04
9.72120404e-01 -6.35454714e-01 9.69489515e-01 -7.39457905e-02
-1.49353898e+00 3.48251343e-01 -7.22024024e-01 -2.80524045e-01
1.56346038e-01 2.78009474e-01 -5.53286910e-01 6.22404635e-01
3.82511497e-01 5.55154026e-01 -8.82844269e-01 2.60073654e-02
-6.58462822e-01 7.13355303e-01 -1.82859246e-02 -1.36736453e-01
2.12212473e-01 -5.27000844e-01 6.27258420e-01 1.57675195e+00
2.02533916e-01 -4.40541096e-02 1.98773190e-01 7.33636618e-01
-1.76265016e-01 5.03064454e-01 -9.09547746e-01 -6.27712607e-01
3.71384054e-01 1.16935074e+00 -1.14047611e+00 -4.46490675e-01
-9.91527557e-01 9.45102692e-01 7.56260335e-01 1.23884611e-01
-4.87421185e-01 -4.92302507e-01 6.54237211e-01 1.63214192e-01
3.90698351e-02 -5.94872415e-01 -7.41655678e-02 -1.35636449e+00
-1.21884190e-01 -9.30021346e-01 6.05105817e-01 -8.26103270e-01
-1.46842468e+00 8.16150546e-01 4.77724941e-03 -6.55172348e-01
-3.93273592e-01 -9.07056034e-01 -4.61266190e-01 9.19451535e-01
-1.10101497e+00 -1.05248535e+00 2.43604273e-01 7.34452233e-02
6.16348505e-01 -5.69339693e-02 1.14757943e+00 -7.34315068e-02
-4.94783670e-01 4.04607832e-01 -3.22273254e-01 6.80933356e-01
2.57381707e-01 -1.62787116e+00 1.07946372e+00 1.05653548e+00
7.07761526e-01 7.95838118e-01 7.23308146e-01 -6.49469614e-01
-1.24544621e+00 -1.07222962e+00 1.11382377e+00 -7.78821170e-01
1.06254685e+00 -5.26817083e-01 -1.05152261e+00 1.16449869e+00
5.24128377e-01 -3.60132800e-03 8.08000088e-01 3.23285908e-01
-6.65885627e-01 2.91650206e-01 -8.43152344e-01 8.64400625e-01
1.25251663e+00 -7.95430124e-01 -1.05014145e+00 1.63602054e-01
1.20870793e+00 -1.77155316e-01 -1.23144472e+00 -3.94545719e-02
2.06031591e-01 -6.46308362e-01 3.50626975e-01 -7.43644595e-01
3.55192810e-01 -3.70995551e-01 -5.91769278e-01 -1.31047344e+00
-3.70790631e-01 -6.17264450e-01 3.94315213e-01 1.41559994e+00
7.75971472e-01 -7.53303826e-01 3.99503767e-01 5.35264552e-01
-4.48967755e-01 -1.66811198e-01 -7.32061446e-01 -7.80755758e-01
3.88745904e-01 -6.54313803e-01 5.02414167e-01 1.06131029e+00
2.32909381e-01 1.09780252e+00 2.48599723e-01 -8.78570899e-02
4.46774691e-01 -1.56399067e-02 4.74128217e-01 -1.11966181e+00
-3.46454620e-01 -1.57054976e-01 -6.90781832e-01 -6.06643260e-01
4.86178637e-01 -1.39159381e+00 -1.59023032e-01 -1.60929096e+00
7.47149587e-02 -2.49147236e-01 1.61486492e-01 6.59104764e-01
3.82624157e-02 1.06925979e-01 2.62849003e-01 2.21231744e-01
-3.65610480e-01 6.79537728e-02 6.98477268e-01 7.07690716e-02
-1.72459334e-02 -4.57721353e-01 -7.78116286e-01 1.07906508e+00
9.06551242e-01 -8.35953891e-01 -2.10144311e-01 -7.30107486e-01
5.53120911e-01 -1.71369076e-01 -2.63007522e-01 -6.25099242e-01
-3.03283244e-01 2.40401868e-02 1.48500189e-01 2.20016856e-02
-1.82105809e-01 -8.10348094e-01 2.76508540e-01 4.08275962e-01
-4.37775493e-01 7.30909050e-01 3.22054960e-02 2.56429613e-01
-2.87914276e-01 -7.66103804e-01 5.44371724e-01 -4.55933243e-01
-8.10258925e-01 -2.26172939e-01 -6.08623624e-01 6.37534261e-01
5.41371644e-01 -2.44485006e-01 -3.46240401e-01 -2.16151029e-01
-4.68411475e-01 -3.14944655e-01 7.96717882e-01 3.70199710e-01
2.26482525e-01 -1.18446100e+00 -6.28014624e-01 -3.11230160e-02
4.51673299e-01 -2.20912844e-01 -4.34006244e-01 4.66459006e-01
-6.27890110e-01 1.12295114e-01 9.59020033e-02 -2.76531577e-01
-1.19549656e+00 5.07293165e-01 1.10352769e-01 -3.17720145e-01
-6.98796332e-01 2.17750818e-01 1.47864327e-01 -7.03035474e-01
-3.36206853e-01 -5.36600471e-01 -8.29005390e-02 -1.80651799e-01
2.38867164e-01 1.65573999e-01 4.44970187e-03 -1.11660397e+00
-2.99893200e-01 5.12206018e-01 -1.04753561e-01 -1.15458049e-01
9.83143389e-01 -3.75868678e-01 -5.44900239e-01 6.69217646e-01
1.11697686e+00 8.08717251e-01 -3.76921058e-01 -6.81174621e-02
7.20807254e-01 -5.32362275e-02 -3.55628610e-01 -3.13661426e-01
-6.10850096e-01 5.94926894e-01 -1.06299333e-01 5.99523842e-01
7.34120667e-01 5.70046902e-01 5.73689044e-01 4.12662178e-01
3.34843636e-01 -1.09701252e+00 -2.96960145e-01 8.17074895e-01
7.50618398e-01 -7.66413629e-01 -1.00697696e-01 -7.25398481e-01
-9.07725453e-01 1.12649226e+00 2.72717237e-01 -1.49330482e-01
2.28583515e-01 3.95229697e-01 2.25716054e-01 -4.30750340e-01
-3.90275627e-01 -1.75471514e-01 1.06523372e-01 5.93394458e-01
1.15478337e+00 3.65545690e-01 -7.82787740e-01 6.48645341e-01
-7.09461927e-01 -6.79455817e-01 8.49918902e-01 9.08364117e-01
-2.67924666e-01 -1.39689660e+00 4.22514975e-02 2.93824792e-01
-1.01370597e+00 -3.58524710e-01 -8.25730443e-01 1.28347623e+00
-2.01442540e-01 8.25670898e-01 3.57039601e-01 -1.15069285e-01
2.80393451e-01 3.48694354e-01 6.08542383e-01 -1.19367266e+00
-5.23037255e-01 4.84181717e-02 8.02430868e-01 -2.26440012e-01
-5.77464044e-01 -5.12574673e-01 -1.60491347e+00 -2.27096118e-02
-3.22835147e-01 4.77902681e-01 4.69129175e-01 9.88683224e-01
-3.29371821e-03 3.44996035e-01 1.26434267e-01 -6.56435341e-02
-2.79239476e-01 -1.00681329e+00 -7.72102714e-01 4.53368008e-01
-3.89472634e-01 -2.64093969e-02 -3.55707586e-01 -4.10995297e-02] | [10.494686126708984, 9.459127426147461] |
ddd3ed40-bedf-4051-b042-2653117776fb | ginius-lt-edi-acl2022-aasha-transformers | null | null | https://aclanthology.org/2022.ltedi-1.43 | https://aclanthology.org/2022.ltedi-1.43.pdf | giniUs @LT-EDI-ACL2022: Aasha: Transformers based Hope-EDI | This paper describes team giniUs’ submission to the Hope Speech Detection for Equality, Diversity and Inclusion Shared Task organised by LT-EDI ACL 2022. We have fine-tuned the Roberta-large pre-trained model and extracted the last four decoder layers to build a classifier. Our best result on the leaderboard achieve a weighted F1 score of 0.86 and a Macro F1 score of 0.51 for English. We have secured a rank of 4 for the English task. We have open-sourced our code implementations on GitHub to facilitate easy reproducibility by the scientific community. | ['Basavraj Chinagundi', 'Harshul Surana'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-4.27652866e-01 5.52972376e-01 -3.62920642e-01 -3.91041040e-01
-1.52110207e+00 -4.18706447e-01 6.66636586e-01 3.85310985e-02
-8.01628530e-01 9.78569567e-01 1.21943676e+00 -5.25896549e-01
1.88311911e-04 7.45138004e-02 -3.14316601e-01 -1.05307989e-01
-2.75915228e-02 5.47630548e-01 -1.08625285e-01 -3.23063344e-01
2.61279851e-01 -1.71142146e-01 -1.10957003e+00 5.54219961e-01
7.25611389e-01 4.09181982e-01 3.50165479e-02 1.13091385e+00
4.70185876e-01 1.04109943e+00 -7.70190299e-01 -9.89252329e-01
1.60872206e-01 -5.06380200e-01 -1.22022188e+00 -3.66505623e-01
5.03944337e-01 -1.72614932e-01 -6.18361473e-01 8.98243666e-01
1.06959367e+00 -5.94729409e-02 1.75625205e-01 -9.06946301e-01
-3.63030404e-01 1.26058066e+00 -1.37521952e-01 5.09517848e-01
4.32946444e-01 3.15759450e-01 1.31682718e+00 -9.95675623e-01
8.50277662e-01 1.23021984e+00 6.89530849e-01 6.98205113e-01
-8.81677508e-01 -5.87877274e-01 -2.62407631e-01 2.28372410e-01
-1.27711082e+00 -1.51398718e+00 6.84928745e-02 -3.31813335e-01
1.41489148e+00 3.86279345e-01 6.37563944e-01 1.38516557e+00
-4.95128855e-02 7.34814703e-01 1.16748309e+00 -2.72307664e-01
-2.05238625e-01 3.27186137e-02 -1.11232445e-01 5.26264489e-01
-2.60560811e-01 7.60169253e-02 -9.27473128e-01 -1.15410201e-01
1.31735250e-01 -9.93733287e-01 -4.06533599e-01 5.55726826e-01
-1.56946719e+00 6.20323837e-01 1.56375974e-01 2.84000874e-01
-1.66081980e-01 9.54389200e-02 8.40900838e-01 5.53579450e-01
3.40336442e-01 3.35467517e-01 -3.74903232e-01 -1.02210927e+00
-1.09583008e+00 2.39423886e-01 1.09207726e+00 9.28622365e-01
-1.40390947e-01 -1.24801859e-01 -3.69410157e-01 1.29248726e+00
2.92626619e-01 2.28571787e-01 4.97337461e-01 -1.24299109e+00
6.93107903e-01 2.62427647e-02 -2.46503532e-01 -3.89802784e-01
-4.96062368e-01 -7.03007519e-01 -6.48709953e-01 -1.34677470e-01
6.34863615e-01 -1.05377391e-01 -4.44510072e-01 1.81295991e+00
-1.94666937e-01 -4.70869303e-01 2.77177691e-01 8.33588123e-01
1.05272722e+00 2.84865469e-01 6.78891763e-02 -2.53365219e-01
1.07951868e+00 -1.10913205e+00 -6.35955393e-01 -1.80692032e-01
9.36336100e-01 -1.06803453e+00 8.25354636e-01 4.34372753e-01
-1.46695960e+00 -3.32056910e-01 -9.75966692e-01 -2.56677747e-01
9.67536345e-02 3.40302944e-01 2.38790497e-01 6.83415771e-01
-1.40550196e+00 4.38037455e-01 -5.06972969e-01 -6.10861838e-01
2.45070979e-01 3.60392720e-01 -5.73959768e-01 1.55766934e-01
-1.12682176e+00 1.04431021e+00 1.74411923e-01 -2.87688207e-02
-8.23105216e-01 -3.17185551e-01 -4.90580767e-01 -3.18270326e-01
7.77600557e-02 -6.45239770e-01 1.35975063e+00 -5.83692968e-01
-1.35721672e+00 1.68027735e+00 -4.77936305e-02 -6.71093583e-01
7.84714639e-01 -5.12242056e-02 -5.71398616e-01 -2.05605641e-01
2.73840606e-01 5.98881125e-01 2.03959525e-01 -3.85212004e-01
-6.43520594e-01 -1.66878641e-01 -2.00280324e-01 4.45569545e-01
-4.86881882e-02 6.96440220e-01 -1.68722853e-01 -4.67773706e-01
-1.09590799e-01 -7.84635723e-01 2.27679163e-01 -5.53709924e-01
-6.67117953e-01 -2.90107727e-01 -7.82430992e-02 -1.27467561e+00
1.46337008e+00 -2.20412397e+00 3.45746517e-01 -2.74092644e-01
6.31408215e-01 1.39711156e-01 -9.22150686e-02 5.01169980e-01
9.13857669e-02 2.70366460e-01 -3.94228175e-02 -6.28068268e-01
3.99702609e-01 -5.46337403e-02 1.54453479e-02 8.16709816e-01
-2.20636725e-01 6.42854333e-01 -8.43431175e-01 -3.58162731e-01
-2.17719764e-01 3.53618145e-01 -7.44058251e-01 1.12004273e-01
3.57531279e-01 3.66275072e-01 3.01367998e-01 3.38818312e-01
3.86069477e-01 1.15346141e-01 2.23801032e-01 3.40213090e-01
-6.56517565e-01 1.18701124e+00 -8.73138487e-01 1.81389022e+00
-3.18965167e-01 1.06304455e+00 4.57638204e-01 -6.69765353e-01
8.14689815e-01 3.18247676e-01 -9.02859271e-02 -5.76058567e-01
9.75823924e-02 7.54186809e-01 7.64053583e-01 -2.86171019e-01
6.93529904e-01 -1.40726238e-01 -2.51207769e-01 1.42910227e-01
2.09672317e-01 -1.32709414e-01 1.28229067e-01 5.21209657e-01
1.15548134e+00 -1.44289896e-01 2.20182210e-01 -6.57969296e-01
5.70834875e-01 -2.68494189e-01 4.39223140e-01 7.64842212e-01
-6.66908920e-01 7.72446692e-01 6.15362763e-01 -2.14512616e-01
-1.33288550e+00 -7.29567766e-01 -2.85898715e-01 1.29747295e+00
-8.06813419e-01 -1.05119240e+00 -7.31317997e-01 -4.55459774e-01
-4.52216208e-01 8.97146225e-01 -3.74104381e-01 -1.63180158e-02
-5.27426243e-01 -3.47468972e-01 1.15620983e+00 2.22937271e-01
4.44593251e-01 -9.20184314e-01 -2.67033391e-02 1.01536244e-01
-5.85810244e-01 -1.20633650e+00 -7.27587223e-01 1.50023654e-01
-8.08273256e-02 -7.32622325e-01 -8.28331470e-01 -7.54289269e-01
-3.25659841e-01 -4.28294808e-01 1.10074377e+00 6.14354163e-02
-1.18944511e-01 3.89698297e-02 -2.16075063e-01 -2.75396675e-01
-8.09223354e-01 5.09496570e-01 2.69412041e-01 -4.79927868e-01
1.87480956e-01 -1.26440018e-01 -2.74150640e-01 1.10207401e-01
1.46122292e-01 2.82256603e-01 1.22822985e-01 9.79641736e-01
1.95835084e-01 -6.59662962e-01 7.13950634e-01 -3.41669172e-01
7.62384534e-01 -2.22301826e-01 -2.03033462e-01 -1.50200799e-01
-2.92327315e-01 -3.25233012e-01 1.75771296e-01 2.71894574e-01
-6.33025706e-01 -3.70249063e-01 -8.02519739e-01 2.32155949e-01
1.48257345e-01 1.99751601e-01 -4.91392389e-02 3.30361038e-01
5.30881584e-01 -2.58655176e-02 -1.62574068e-01 -4.06368107e-01
4.42030936e-01 1.32077301e+00 8.45772326e-01 -4.18021917e-01
1.63582310e-01 -9.65979844e-02 -6.22983217e-01 -1.07503355e+00
-9.55671012e-01 -3.03914964e-01 -4.16276455e-01 -1.73972279e-01
9.47011173e-01 -1.42253578e+00 -7.64995694e-01 4.92342591e-01
-1.11431026e+00 -6.49545431e-01 -3.54630142e-01 6.02479398e-01
-5.90560615e-01 5.23356758e-02 -5.45939088e-01 -7.95392811e-01
-5.48005342e-01 -1.02778435e+00 8.68453085e-01 3.57886404e-02
-7.48517036e-01 -8.61129105e-01 4.64352250e-01 8.09202194e-01
4.63384181e-01 -3.01737577e-01 3.17982972e-01 -9.22767699e-01
-4.58501168e-02 1.05650462e-01 -2.11285815e-01 4.26818341e-01
-4.85941470e-01 1.34244293e-03 -1.08146501e+00 -2.82684892e-01
-1.44458696e-01 -7.39457428e-01 1.04701304e+00 2.71573395e-01
8.11065674e-01 -4.07082796e-01 -2.85029504e-02 5.39708376e-01
6.37513459e-01 -4.32480723e-01 6.54101729e-01 4.36902583e-01
4.60274309e-01 5.74168563e-01 1.95721373e-01 6.78460240e-01
8.10293853e-01 8.68198574e-01 -1.46148860e-01 3.45355839e-01
-7.21120000e-01 -4.07644421e-01 6.00946963e-01 1.46042097e+00
-7.45220706e-02 -3.33915770e-01 -1.15308714e+00 8.14528883e-01
-1.70191538e+00 -1.11122489e+00 -5.84518552e-01 2.13334727e+00
9.32885408e-01 3.31931889e-01 5.48881650e-01 -6.49570599e-02
5.55778980e-01 1.23649836e-01 1.64608523e-01 -8.96171570e-01
-5.00291407e-01 1.10830136e-01 4.23042953e-01 1.05761671e+00
-7.42219150e-01 1.27671015e+00 7.38017845e+00 9.00663316e-01
-7.31754422e-01 5.16449571e-01 6.77759230e-01 -4.55970645e-01
-1.92384332e-01 1.08197466e-01 -1.07800686e+00 4.44114119e-01
1.70229840e+00 -5.33042967e-01 5.40626943e-01 3.79545391e-01
2.06027582e-01 -3.34344804e-02 -8.61508191e-01 1.02068961e+00
4.96158153e-02 -1.10015917e+00 -6.40851438e-01 5.18262267e-01
6.14032507e-01 9.99074161e-01 4.80735637e-02 4.88709718e-01
5.73096395e-01 -1.35006392e+00 8.42187226e-01 4.56908315e-01
1.10373366e+00 -7.23919928e-01 5.47184646e-01 4.22255963e-01
-6.46470964e-01 -1.01771854e-01 -3.97246718e-01 -2.98622817e-01
3.15464251e-02 3.89877170e-01 -8.66543055e-01 3.86838555e-01
3.46659869e-01 5.76145649e-01 -4.01494354e-01 1.18471289e+00
-2.93934405e-01 8.79836559e-01 -5.09628117e-01 9.92256254e-02
2.57295698e-01 4.00577337e-01 9.36157346e-01 1.55141294e+00
1.31968588e-01 -1.95280001e-01 1.69048328e-02 3.94148707e-01
-5.67292869e-01 3.75294924e-01 -4.08074379e-01 -1.68970466e-01
4.84149009e-01 1.01785231e+00 -5.53391762e-02 -2.73886204e-01
-3.23793352e-01 1.04883385e+00 5.45417249e-01 -1.83245465e-01
-5.67963243e-01 -4.49683964e-01 5.90639710e-01 6.33310750e-02
2.57172566e-02 -2.70248055e-01 -2.68109828e-01 -1.17308438e+00
-3.11774939e-01 -1.26123869e+00 3.56653094e-01 -5.70561826e-01
-1.05334699e+00 8.06872964e-01 -3.65508586e-01 -7.04069436e-01
-4.75427032e-01 -4.55571413e-01 -4.90232587e-01 1.15481961e+00
-8.96988034e-01 -9.88810122e-01 1.37732461e-01 2.27698341e-01
5.49875915e-01 -6.59799218e-01 9.03199136e-01 4.32015926e-01
-7.51924336e-01 1.07055497e+00 1.04710475e-01 1.33260623e-01
1.04448891e+00 -1.29554498e+00 6.24331713e-01 7.36207902e-01
8.98971260e-02 4.95318264e-01 7.34503686e-01 -4.45856422e-01
-1.04651153e+00 -5.92036307e-01 1.92687237e+00 -6.50959790e-01
9.52276707e-01 -3.96044850e-01 -3.69270205e-01 8.26503694e-01
7.28942335e-01 -4.40297186e-01 6.95867479e-01 3.59508723e-01
-2.51656979e-01 2.42886603e-01 -8.46226037e-01 4.25753653e-01
1.33571768e+00 -7.79102981e-01 -5.78282297e-01 6.49611950e-01
5.58418334e-01 -5.84747016e-01 -9.27137196e-01 1.68472286e-02
6.69031918e-01 -1.16404915e+00 5.17908037e-01 -6.11807644e-01
5.38921654e-01 1.85297489e-01 -4.59734470e-01 -1.35918975e+00
-2.79124409e-01 -1.12553287e+00 2.16294944e-01 1.26544774e+00
7.49420106e-01 -5.89932442e-01 3.36231351e-01 2.22640231e-01
-4.98635530e-01 -5.49447834e-01 -1.41964865e+00 -5.88412404e-01
4.57769006e-01 -5.85937738e-01 5.58472909e-02 8.46414208e-01
5.03922403e-01 7.28326082e-01 -5.44777095e-01 -3.77928495e-01
5.13936579e-01 -5.17137229e-01 9.53785777e-01 -8.53236258e-01
-6.13933384e-01 -8.70612562e-01 -3.28389972e-01 -7.15661764e-01
2.24392489e-01 -1.52345538e+00 -1.82679191e-01 -1.47912383e+00
4.12495077e-01 -1.77450299e-01 2.49125306e-02 4.82589930e-01
-1.77282622e-04 4.47938859e-01 2.78854549e-01 3.14591616e-01
-7.92310596e-01 4.44068730e-01 9.26646650e-01 1.32728085e-01
5.96329533e-02 -6.52331561e-02 -1.26464653e+00 4.85369146e-01
8.77038777e-01 -5.18403471e-01 1.60517529e-01 -3.62697870e-01
2.27731422e-01 -8.89723822e-02 2.87383467e-01 -1.07592034e+00
1.40866205e-01 2.40548074e-01 5.40756732e-02 -3.82515132e-01
3.13354731e-01 1.82250351e-01 -2.71190908e-02 5.23248792e-01
-6.04937553e-01 -8.63287821e-02 2.67039299e-01 -2.12230593e-01
-9.01525840e-02 -1.86761111e-01 6.90341890e-01 -1.16118580e-01
-1.52271494e-01 -1.42596319e-01 -4.53582257e-01 7.26724744e-01
3.85972559e-01 1.95579559e-01 -5.89219511e-01 -6.65112019e-01
-8.76979768e-01 3.91442567e-01 3.66544753e-01 3.45692098e-01
1.69086531e-01 -1.04177260e+00 -1.57091987e+00 1.77182648e-02
4.70223688e-02 -8.00173402e-01 1.53595820e-01 1.24040055e+00
-3.87305647e-01 5.54772377e-01 -9.78737175e-02 -2.73672551e-01
-1.42249012e+00 -1.52995437e-01 5.67773342e-01 -2.06016228e-01
-6.38347268e-01 1.28576422e+00 -3.56185049e-01 -8.16592336e-01
3.49933475e-01 2.74847746e-01 1.49850085e-01 8.07526261e-02
7.81094432e-01 6.38797939e-01 6.64525107e-02 -1.02571452e+00
-6.03005409e-01 6.64432719e-02 -1.36913791e-01 -7.66813278e-01
1.21952999e+00 -9.92530733e-02 -3.01055983e-02 5.03388107e-01
1.32238030e+00 5.14887214e-01 -5.69284976e-01 1.51405558e-01
-9.37118083e-02 -2.15987921e-01 3.81955266e-01 -1.16098726e+00
-4.34930831e-01 8.55949342e-01 2.83283234e-01 1.05246797e-01
6.39975250e-01 1.98029622e-01 5.03569901e-01 1.03610635e-01
-8.52474421e-02 -1.31284809e+00 -3.37187231e-01 1.01830542e+00
1.24808145e+00 -9.80459154e-01 -2.51365956e-02 3.14734988e-02
-8.87046456e-01 9.18159783e-01 2.90010422e-01 1.82163760e-01
1.43236279e-01 3.42319965e-01 6.66124448e-02 -4.55199443e-02
-1.11609161e+00 -3.80274713e-01 1.69453889e-01 5.69422185e-01
1.04683709e+00 3.96292329e-01 -7.51025438e-01 7.25467741e-01
-9.65061486e-01 -1.88276783e-01 5.28385341e-01 2.20697552e-01
-4.65526938e-01 -1.15047050e+00 7.53562227e-02 2.35858500e-01
-8.30120683e-01 -4.47263837e-01 -7.87068069e-01 6.02742434e-01
-1.21831134e-01 9.16547894e-01 -8.94680619e-03 -6.21369898e-01
3.25771302e-01 3.77827376e-01 5.90655327e-01 -5.29805481e-01
-9.75308716e-01 2.08508343e-01 1.09434938e+00 -4.46464479e-01
3.97670362e-03 -9.20958042e-01 -8.97345424e-01 -8.44302952e-01
1.04316892e-02 1.29477039e-01 7.70725787e-01 6.91196799e-01
3.32526237e-01 4.16909903e-01 2.73213416e-01 -6.09939218e-01
-6.42717183e-01 -1.40576112e+00 -3.23618442e-01 -1.29657641e-01
2.06602737e-01 -8.13263655e-02 -4.35754299e-01 -4.10442472e-01] | [9.461379051208496, 10.723828315734863] |
933591fa-41b8-402f-801e-27575ee5182f | the-dark-side-of-explanations-poisoning | 2305.00574 | null | https://arxiv.org/abs/2305.00574v1 | https://arxiv.org/pdf/2305.00574v1.pdf | The Dark Side of Explanations: Poisoning Recommender Systems with Counterfactual Examples | Deep learning-based recommender systems have become an integral part of several online platforms. However, their black-box nature emphasizes the need for explainable artificial intelligence (XAI) approaches to provide human-understandable reasons why a specific item gets recommended to a given user. One such method is counterfactual explanation (CF). While CFs can be highly beneficial for users and system designers, malicious actors may also exploit these explanations to undermine the system's security. In this work, we propose H-CARS, a novel strategy to poison recommender systems via CFs. Specifically, we first train a logical-reasoning-based surrogate model on training data derived from counterfactual explanations. By reversing the learning process of the recommendation model, we thus develop a proficient greedy algorithm to generate fabricated user profiles and their associated interaction records for the aforementioned surrogate model. Our experiments, which employ a well-known CF generation method and are conducted on two distinct datasets, show that H-CARS yields significant and successful attack performance. | ['Gabriele Tolomei', 'Yongfeng Zhang', 'Jia Wang', 'Fabrizio Silvestri', 'Ziheng Chen'] | 2023-04-30 | null | null | null | null | ['counterfactual-explanation', 'logical-reasoning'] | ['miscellaneous', 'reasoning'] | [ 1.91603541e-01 4.82540309e-01 -3.10023546e-01 -3.95068079e-01
-3.08541715e-01 -7.65955925e-01 6.84092700e-01 -4.07063682e-03
1.78149551e-01 6.67434931e-01 3.89739662e-01 -9.72579062e-01
-3.12840581e-01 -9.75846350e-01 -9.48434055e-01 -2.34625414e-01
4.98251952e-02 2.46154070e-01 -4.10346031e-01 -3.30800682e-01
4.80603933e-01 4.39351201e-01 -1.37449777e+00 6.72203660e-01
8.92846763e-01 7.88376868e-01 -4.20361042e-01 5.17958343e-01
8.81004333e-02 8.63312781e-01 -6.39195442e-01 -1.08633661e+00
2.42244542e-01 -3.60593766e-01 -6.96273923e-01 -1.99777052e-01
1.13735124e-01 -6.08314157e-01 -3.37677956e-01 9.19730246e-01
1.35292023e-01 7.00674504e-02 6.29083455e-01 -1.57193613e+00
-9.58006322e-01 1.33040726e+00 -3.29207569e-01 -2.30645776e-01
4.92257893e-01 1.53132498e-01 1.42098522e+00 -5.87231874e-01
1.74460217e-01 1.18863475e+00 4.87129539e-01 8.05427909e-01
-1.25776327e+00 -8.72363210e-01 3.35252672e-01 1.78353578e-01
-7.96871901e-01 -1.45129830e-01 8.35148335e-01 -2.00273797e-01
5.97656548e-01 6.62039697e-01 4.99956459e-01 1.42058301e+00
7.81991631e-02 9.72372115e-01 9.44808841e-01 -3.24407995e-01
4.87032413e-01 4.81239289e-01 3.16781163e-01 4.34173137e-01
7.79288590e-01 4.73288447e-01 -3.92711431e-01 -6.00502729e-01
2.64820963e-01 3.91543597e-01 -2.61643827e-01 -3.14644307e-01
-7.29198575e-01 1.16743040e+00 4.40569162e-01 7.64204040e-02
-6.49912357e-01 1.43313840e-01 2.01147661e-01 2.06103981e-01
1.41176939e-01 1.06729436e+00 -5.41482866e-01 4.13224846e-02
-5.97559214e-01 5.44504941e-01 1.05505288e+00 5.11384666e-01
2.92074353e-01 1.51456475e-01 -2.23875403e-01 6.55127317e-02
5.03885806e-01 3.89365584e-01 1.91957414e-01 -9.07284498e-01
1.21373944e-01 6.23244643e-01 4.40428019e-01 -1.10951483e+00
1.41729889e-02 -6.08173430e-01 -5.45398653e-01 1.00535117e-01
4.50466007e-01 -3.64638537e-01 -3.60523820e-01 1.66743875e+00
1.88338697e-01 4.08102840e-01 1.59498394e-01 9.06920075e-01
4.67347234e-01 3.40213507e-01 1.91009082e-02 4.36301157e-02
1.14670730e+00 -6.53385520e-01 -3.86044621e-01 1.06947646e-01
6.66179299e-01 -3.10705870e-01 1.33695924e+00 8.44934762e-01
-8.18113148e-01 -2.07115039e-01 -1.26494324e+00 5.09066880e-01
-2.27703944e-01 5.36078587e-02 9.56784785e-01 1.03765261e+00
-4.31950808e-01 7.30179548e-01 -3.73225391e-01 -1.52332550e-02
4.48115140e-01 4.26046550e-01 3.30520384e-02 8.59197751e-02
-1.45074546e+00 4.90634948e-01 7.73563534e-02 1.39695600e-01
-1.01544595e+00 -7.61535525e-01 -4.91498560e-01 5.49224257e-01
7.09973693e-01 -5.94892085e-01 1.40121412e+00 -8.10603559e-01
-1.53532326e+00 6.67837039e-02 2.50075847e-01 -9.13521171e-01
4.18231338e-01 -5.29532075e-01 -4.69578534e-01 -2.41098553e-01
-3.01890999e-01 2.16329932e-01 9.93701279e-01 -1.51215076e+00
-4.87134993e-01 -1.34887248e-01 7.48433650e-01 -2.58129090e-01
-4.76072401e-01 -7.53305182e-02 3.16379905e-01 -4.99683499e-01
-5.42697072e-01 -1.08321381e+00 -4.48511809e-01 -3.92684340e-01
-1.04800844e+00 -7.68038034e-02 5.14345407e-01 -4.20318663e-01
1.41264153e+00 -1.81605816e+00 -3.69285733e-01 5.96787751e-01
4.06056613e-01 5.38072586e-01 -5.71055189e-02 3.42309982e-01
-2.08023056e-01 5.25648594e-01 2.57891595e-01 9.37696919e-02
3.89887691e-01 -6.01926930e-02 -1.12573838e+00 9.37684551e-02
1.68332960e-02 7.94566453e-01 -7.99682915e-01 2.23023146e-01
1.79167226e-01 3.20317835e-01 -1.11946714e+00 4.53697026e-01
-6.31301463e-01 3.09586078e-01 -6.36474729e-01 7.45769069e-02
5.40307999e-01 -3.94704074e-01 5.63375592e-01 -1.75592661e-01
2.84948319e-01 5.31861305e-01 -9.85604465e-01 1.14043581e+00
-5.79969406e-01 2.03869462e-01 -4.50952351e-01 -7.10581064e-01
6.88630462e-01 2.18728706e-01 1.38582304e-01 -3.14829111e-01
3.76740098e-01 -7.14326724e-02 2.53632694e-01 -3.78234863e-01
5.09368479e-01 -3.89198363e-02 -1.86813787e-01 1.05781758e+00
-5.58857381e-01 4.84237731e-01 -2.14104444e-01 4.87546653e-01
1.18601298e+00 -6.55558631e-02 3.24682891e-01 1.30831404e-02
6.05326891e-01 -6.11027181e-02 3.42063904e-01 1.21450496e+00
1.59883708e-01 2.34593287e-01 6.32152677e-01 -7.42264390e-01
-7.47179627e-01 -7.65878320e-01 4.27219063e-01 9.25992429e-01
5.14690056e-02 -5.88828921e-01 -9.31352735e-01 -1.23042476e+00
7.61786774e-02 1.52342629e+00 -6.40073836e-01 -4.97212082e-01
-3.19998831e-01 -5.42665720e-01 6.46473706e-01 3.64798307e-01
2.48147607e-01 -1.03638566e+00 -6.29375994e-01 1.40886098e-01
-2.26701438e-01 -7.04995751e-01 -3.64965618e-01 -2.40907386e-01
-3.96637022e-01 -1.27146840e+00 -1.32435068e-01 2.38097683e-01
6.53479695e-01 5.83664238e-01 1.01286685e+00 6.65397763e-01
1.58646852e-01 7.03050345e-02 -4.63568211e-01 -3.90810996e-01
-7.60152698e-01 1.06439309e-03 4.40137476e-01 1.86850369e-01
5.37485063e-01 -5.71889102e-01 -8.38871956e-01 3.41683030e-01
-8.16108048e-01 -4.02728580e-02 5.77162266e-01 7.63575852e-01
1.39597163e-01 1.65874884e-01 7.12778986e-01 -1.42728996e+00
1.10398269e+00 -7.68340111e-01 -5.90364873e-01 4.92526680e-01
-1.11879110e+00 3.34597111e-01 1.26943612e+00 -6.10012770e-01
-1.18116248e+00 -2.78610498e-01 -1.82320029e-01 -3.17775458e-01
-3.77391249e-01 4.90094483e-01 -4.77254450e-01 2.09348992e-01
1.05595028e+00 8.63539129e-02 -3.28185976e-01 -3.72950077e-01
8.63366008e-01 7.27762163e-01 3.78020346e-01 -7.34072983e-01
1.05962360e+00 4.15169626e-01 -3.52387130e-01 -1.24491036e-01
-1.19900680e+00 1.59398675e-01 6.97290376e-02 -1.92372166e-02
5.33320963e-01 -3.19389224e-01 -1.40463614e+00 -1.79856777e-01
-1.11750937e+00 -3.95607986e-02 -1.19579151e-01 4.04908031e-01
-3.64780903e-01 2.17588961e-01 -3.33840430e-01 -9.66387212e-01
-6.22091174e-01 -1.03580546e+00 4.89787936e-01 2.15212449e-01
-5.56466103e-01 -7.21054971e-01 -4.60156761e-02 6.03629529e-01
4.28364128e-01 8.47231969e-02 1.15547383e+00 -1.46148431e+00
-5.87147832e-01 -3.32970887e-01 -1.53136775e-01 1.61227882e-01
-7.91617855e-02 6.19774237e-02 -9.31919932e-01 -1.87305018e-01
-8.53715166e-02 -1.10367358e-01 1.44384012e-01 8.54663327e-02
1.32123005e+00 -1.09804761e+00 -1.97535470e-01 2.78229415e-01
1.12290931e+00 2.71961838e-01 5.95566392e-01 3.17303777e-01
3.69750500e-01 5.58546245e-01 6.42118037e-01 5.71925879e-01
2.02075362e-01 6.48116350e-01 5.56524932e-01 6.75415471e-02
3.55363041e-01 -8.57243717e-01 4.35219258e-01 6.17057830e-02
-1.36224157e-03 -3.43362302e-01 -4.59142745e-01 9.12306830e-02
-1.87574518e+00 -1.23749709e+00 -1.76666260e-01 2.18177867e+00
4.42619383e-01 4.66689318e-01 1.59279093e-01 2.32132703e-01
5.76409101e-01 -2.41618708e-01 -7.86894858e-01 -6.37682855e-01
2.58840382e-01 6.23476394e-02 3.28777432e-01 2.07848296e-01
-5.89548528e-01 7.30560541e-01 5.83040380e+00 5.64380765e-01
-9.59640920e-01 -5.28617501e-02 6.40483081e-01 -5.07405736e-02
-9.77642298e-01 1.64726168e-01 -6.04627192e-01 4.31240380e-01
1.23553932e+00 -4.94418979e-01 7.80211151e-01 1.07442260e+00
3.85677516e-01 5.05773723e-01 -1.25756896e+00 5.78035653e-01
-1.33339226e-01 -1.73968446e+00 4.42245960e-01 3.75733018e-01
4.79037911e-01 -5.09418070e-01 3.25933695e-01 3.91666740e-01
8.00481141e-01 -9.54544008e-01 7.75185764e-01 3.33710611e-01
4.43892032e-01 -8.51520300e-01 5.59672892e-01 4.45377022e-01
-5.22658050e-01 -5.85623920e-01 -2.52090424e-01 -2.37334520e-01
-2.95886602e-02 4.22031999e-01 -1.02475584e+00 4.85321909e-01
3.92511010e-01 2.71295130e-01 -2.44122699e-01 6.84146166e-01
-7.34531045e-01 1.04683733e+00 7.24610547e-03 -1.89103648e-01
-9.79909394e-03 -8.35371390e-02 4.45669830e-01 6.37944520e-01
2.90424198e-01 2.05194548e-01 -2.81588078e-01 1.22728479e+00
-2.68790454e-01 -2.02050075e-01 -6.63500130e-01 -2.42632151e-01
6.31923497e-01 1.38810837e+00 -3.55966210e-01 -2.53356844e-01
-4.19847161e-01 6.38773799e-01 9.15921554e-02 2.46950433e-01
-1.20241106e+00 -1.51277512e-01 8.26595426e-01 3.70345503e-01
1.00808233e-01 3.82097125e-01 -4.58393216e-01 -1.32770157e+00
-2.46724397e-01 -1.49101663e+00 3.42683285e-01 -6.86519444e-01
-1.26570940e+00 6.93295062e-01 -1.62467986e-01 -1.09569800e+00
-4.83035117e-01 -3.97590578e-01 -9.37594533e-01 4.94192928e-01
-1.39438641e+00 -1.12011099e+00 -1.79605186e-02 4.91579384e-01
2.66531378e-01 -4.27366316e-01 8.55427504e-01 7.93776438e-02
-5.98461509e-01 9.99360323e-01 -1.45853639e-01 -8.40442404e-02
3.41447473e-01 -1.14301813e+00 7.17149019e-01 9.36583757e-01
5.43571532e-01 1.29153502e+00 1.07509959e+00 -5.66063464e-01
-1.48465383e+00 -9.14465129e-01 5.74506462e-01 -5.40177882e-01
6.59867048e-01 -3.27805519e-01 -8.48068833e-01 5.96022427e-01
2.90116221e-02 -5.25902927e-01 1.10239947e+00 4.64452833e-01
-6.48102701e-01 5.26457205e-02 -1.17139506e+00 9.45381761e-01
8.78825963e-01 -3.87286216e-01 -6.50141180e-01 3.40047359e-01
8.64934683e-01 1.13232031e-01 -5.48382401e-01 1.40432939e-01
7.79039383e-01 -1.08296919e+00 9.30356801e-01 -1.34306073e+00
5.52628219e-01 -3.05257738e-01 -3.29659656e-02 -1.28581643e+00
-1.85367748e-01 -1.14172161e+00 -6.56318128e-01 1.03396249e+00
6.40146375e-01 -4.98591870e-01 1.15685225e+00 1.23125470e+00
1.45551637e-01 -7.54390895e-01 -2.23391861e-01 -5.82801819e-01
-7.23008290e-02 -6.97631717e-01 1.33794534e+00 8.49018872e-01
3.16039413e-01 3.50087643e-01 -7.22959816e-01 4.08730000e-01
7.63106763e-01 4.88226742e-01 1.10524559e+00 -1.24675465e+00
-7.87311912e-01 -2.02399373e-01 2.68324286e-01 -9.19420362e-01
2.90061265e-01 -7.10496604e-01 -3.29687327e-01 -1.07094991e+00
1.10418983e-01 -4.00357306e-01 -3.41620237e-01 5.44817686e-01
-8.30168650e-02 -1.47714287e-01 2.54188538e-01 2.28181452e-01
-4.33225393e-01 3.13056052e-01 8.67733121e-01 4.42171544e-02
-1.78066865e-01 5.44348359e-01 -1.60668933e+00 7.28157341e-01
9.13295865e-01 -6.37400627e-01 -7.30625033e-01 1.09531051e-02
6.98928714e-01 9.52154547e-02 3.21587712e-01 -4.91465300e-01
1.16896391e-01 -3.49967360e-01 -8.38591829e-02 -1.73473075e-01
-6.74511716e-02 -9.22500908e-01 2.98899710e-01 6.37910247e-01
-8.30477059e-01 -1.46077961e-01 -2.72787601e-01 7.60552287e-01
2.37296879e-01 -1.81077749e-01 3.45746309e-01 1.01380542e-01
-2.45546713e-01 1.70459270e-01 -4.04328734e-01 -3.08619261e-01
9.24392283e-01 -1.06660919e-02 -5.47774911e-01 -7.77141809e-01
-4.79031622e-01 -2.16005016e-02 2.14807123e-01 4.93575096e-01
7.59832680e-01 -1.15972936e+00 -5.05407095e-01 2.00514361e-01
2.18019132e-02 -7.50662148e-01 1.52222320e-01 4.00958389e-01
-4.80908789e-02 5.94463289e-01 -5.34111373e-02 1.92929700e-01
-9.38623071e-01 1.05172133e+00 3.05928826e-01 -4.21367675e-01
-5.48059583e-01 4.99470770e-01 3.00298423e-01 -5.13565719e-01
1.69466943e-01 -1.09140955e-01 -2.86657989e-01 -5.23380816e-01
8.35276306e-01 3.40113580e-01 -1.48440629e-01 -3.42246220e-02
-1.24697655e-01 -4.07552660e-01 -4.57416207e-01 2.03634631e-02
1.21696770e+00 9.48499218e-02 3.35356325e-01 -2.85423070e-01
6.31617725e-01 3.13701451e-01 -1.16284132e+00 -8.62742513e-02
7.88240805e-02 -6.77901387e-01 -1.44570440e-01 -1.11363661e+00
-9.87018764e-01 8.65728259e-01 1.18960790e-01 6.97340071e-01
8.20992768e-01 -3.85437161e-01 1.02394664e+00 6.08761787e-01
4.18141574e-01 -5.80246985e-01 2.68581714e-02 -6.56197965e-03
6.76053107e-01 -9.45931971e-01 -4.91189258e-03 -3.26949328e-01
-8.14326763e-01 8.95845652e-01 6.78622067e-01 -4.53865938e-02
4.93581593e-01 3.83530557e-02 -6.25033826e-02 -2.16169536e-01
-9.54058647e-01 1.97883859e-01 1.96122140e-01 4.19644505e-01
3.02836478e-01 2.39879057e-01 -2.03713313e-01 1.47549796e+00
-2.85680711e-01 2.01562524e-01 8.89531195e-01 5.21548688e-01
-1.23707302e-01 -1.19906449e+00 -2.09016830e-01 5.94847441e-01
-6.88430667e-01 -1.65644079e-01 -6.40616357e-01 5.63966155e-01
-1.67526320e-01 1.38111007e+00 -5.84664166e-01 -1.01715457e+00
3.02054197e-01 -2.26772249e-01 -5.70374429e-02 -4.97009128e-01
-9.05315042e-01 -3.87061626e-01 2.71331370e-01 -6.74261570e-01
-1.27593093e-02 -4.23151076e-01 -1.26922929e+00 -7.68668592e-01
-5.06385088e-01 7.22028732e-01 5.16527951e-01 1.11164653e+00
6.45053148e-01 2.87971526e-01 8.92693102e-01 -4.31279153e-01
-1.18305314e+00 -5.33011496e-01 -4.64080542e-01 5.76442897e-01
-1.26600713e-01 -6.12757504e-01 -4.47740078e-01 -1.97325975e-01] | [9.530865669250488, 5.704881191253662] |
bbd0987b-0d84-47de-92f0-79a850f4955c | large-margin-mechanism-and-pseudo-query-set | 2005.09218 | null | https://arxiv.org/abs/2005.09218v1 | https://arxiv.org/pdf/2005.09218v1.pdf | Large Margin Mechanism and Pseudo Query Set on Cross-Domain Few-Shot Learning | In recent years, few-shot learning problems have received a lot of attention. While methods in most previous works were trained and tested on datasets in one single domain, cross-domain few-shot learning is a brand-new branch of few-shot learning problems, where models handle datasets in different domains between training and testing phases. In this paper, to solve the problem that the model is pre-trained (meta-trained) on a single dataset while fine-tuned on datasets in four different domains, including common objects, satellite images, and medical images, we propose a novel large margin fine-tuning method (LMM-PQS), which generates pseudo query images from support images and fine-tunes the feature extraction modules with a large margin mechanism inspired by methods in face recognition. According to the experiment results, LMM-PQS surpasses the baseline models by a significant margin and demonstrates that our approach is robust and can easily adapt pre-trained models to new domains with few data. | ['Ping-Chia Huang', 'Yi-Rong Chen', 'Bing-Chen Tsai', 'Hsin-Ying Lee', 'Jia-Fong Yeh', 'Winston H. Hsu'] | 2020-05-19 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 3.86461854e-01 6.83268011e-02 -3.96148831e-01 -7.24699855e-01
-7.67923176e-01 2.75109746e-02 8.05148244e-01 -2.01532707e-01
-2.58850336e-01 7.51027286e-01 5.41659929e-02 4.46385175e-01
-3.96042109e-01 -9.77459490e-01 -5.02816379e-01 -3.90347242e-01
1.10004835e-01 7.20092773e-01 7.77361035e-01 -4.52193171e-01
1.19495869e-01 1.75757423e-01 -1.93325460e+00 3.16042483e-01
6.38776004e-01 8.78028512e-01 2.52424538e-01 3.24342430e-01
-1.98829949e-01 6.42117977e-01 -5.17426968e-01 -4.81939584e-01
2.68814236e-01 -6.85141683e-01 -8.79015684e-01 3.66245180e-01
2.86294788e-01 -2.77891397e-01 -2.86416829e-01 9.64546859e-01
9.19822872e-01 4.40957010e-01 8.47235382e-01 -1.24413753e+00
-8.12265754e-01 4.35533762e-01 -4.56206828e-01 2.60857463e-01
1.34321764e-01 1.39205322e-01 6.13167703e-01 -1.06477225e+00
8.83168161e-01 1.27159178e+00 6.69378817e-01 1.14150274e+00
-1.18457484e+00 -5.47898173e-01 -2.58964509e-01 5.99840939e-01
-1.46129298e+00 -6.28916740e-01 7.24379003e-01 -4.45534885e-01
7.75609195e-01 -3.40343863e-02 1.48055047e-01 1.27630568e+00
-5.81787005e-02 4.56493080e-01 9.59548593e-01 -5.90425193e-01
8.43989551e-01 4.29286242e-01 -8.25834274e-03 3.94915521e-01
-1.03910351e-02 1.86642811e-01 -6.29197001e-01 -2.19614744e-01
4.18967456e-01 2.69356728e-01 -1.14138357e-01 -8.09451401e-01
-7.86237240e-01 1.25196362e+00 5.50139435e-02 5.02846003e-01
-1.88553303e-01 -5.05062342e-01 5.85125089e-01 5.17191589e-01
4.70868737e-01 4.11798656e-01 -5.77939868e-01 -5.74315004e-02
-1.01093876e+00 3.28154191e-02 6.44467831e-01 1.06154156e+00
1.17224455e+00 4.90460619e-02 -5.87401569e-01 1.31981194e+00
-6.41063899e-02 1.52809024e-01 1.13908112e+00 -3.38734955e-01
1.78842112e-01 5.59018433e-01 -1.47817820e-01 -5.95533788e-01
-2.92593539e-01 -1.65400673e-02 -7.82489419e-01 7.20339417e-02
2.80377790e-02 -3.27168941e-01 -1.22381687e+00 1.53757429e+00
6.37861967e-01 7.72630751e-01 1.41038150e-01 7.19658136e-01
1.06198204e+00 5.76003373e-01 1.36762276e-01 -3.64486217e-01
1.17042112e+00 -1.04775178e+00 -5.22795141e-01 -4.73791867e-01
4.08720851e-01 -4.37145948e-01 1.23673904e+00 -7.46712983e-02
-5.64962626e-01 -9.51252520e-01 -1.22096097e+00 4.80271965e-01
-6.22892320e-01 -4.71858293e-01 2.62918562e-01 7.84915686e-01
-7.40488112e-01 8.98992658e-01 -2.63906926e-01 -9.86515403e-01
6.00355208e-01 1.45743832e-01 -3.65973592e-01 -4.78219599e-01
-1.46913397e+00 9.64613378e-01 5.76005697e-01 -6.12412989e-01
-1.26120269e+00 -8.47322345e-01 -1.01601446e+00 1.23218648e-01
6.75878167e-01 -4.92915362e-01 1.25801015e+00 -8.82166684e-01
-1.68554032e+00 9.36356485e-01 2.33232126e-01 -2.70839125e-01
3.64765167e-01 4.83951010e-02 -7.72629559e-01 4.86796126e-02
1.85281336e-01 5.24285555e-01 1.43506324e+00 -7.60205150e-01
-5.00234365e-01 -4.57111359e-01 -2.45667547e-01 -5.14601171e-02
-6.79177225e-01 1.74393468e-02 -6.00543201e-01 -6.09109282e-01
-6.27710283e-01 -5.82791030e-01 -2.53886491e-01 -1.48605555e-01
1.43072411e-01 -2.86281049e-01 9.08594847e-01 -4.12028693e-02
1.10007966e+00 -2.15522289e+00 -4.59860219e-03 -1.14841565e-01
-3.07106853e-01 7.38835037e-01 -4.54677016e-01 4.35264885e-01
-2.26463363e-01 -2.80928254e-01 -4.14267480e-01 -4.37762439e-02
-1.33765982e-02 3.67068708e-01 -1.50882855e-01 2.22688615e-01
5.13192475e-01 8.58574510e-01 -1.00274324e+00 -6.73678815e-01
2.51309514e-01 2.10437253e-01 -3.26659560e-01 4.41546679e-01
-2.30453953e-01 1.09245785e-01 -2.98302174e-01 7.99887776e-01
7.17311323e-01 -3.65687877e-01 2.54013777e-01 5.83330989e-02
4.06745523e-01 -6.47252679e-01 -1.48140550e+00 2.03595996e+00
-3.01779330e-01 1.20001860e-01 -4.99348849e-01 -1.22411144e+00
1.03453588e+00 3.86301816e-01 4.20439452e-01 -8.93778801e-01
1.23266593e-01 4.13303934e-02 -3.09521586e-01 -8.68743300e-01
1.00899331e-01 -5.49218416e-01 -2.19160423e-01 5.01884758e-01
8.98829877e-01 -2.02478364e-01 2.89095402e-01 -1.58196211e-01
1.05412138e+00 1.34145498e-01 8.14933777e-01 4.33065146e-02
5.38765252e-01 3.96357998e-02 8.23131800e-01 7.41572082e-01
-7.26371527e-01 8.59958947e-01 7.33799934e-02 -6.18356228e-01
-1.12452459e+00 -8.15751493e-01 -4.22099978e-01 1.80402219e+00
1.47124052e-01 -2.87369549e-01 -8.04171264e-01 -8.79785299e-01
1.48201406e-01 6.17046893e-01 -1.04922938e+00 -5.37318289e-01
-1.13504596e-01 -8.93410981e-01 2.64119953e-01 4.50795084e-01
4.94585663e-01 -1.12652910e+00 -7.20403194e-01 3.96063119e-01
2.35362411e-01 -8.71168554e-01 -3.68163675e-01 2.81554699e-01
-8.18734884e-01 -1.19624281e+00 -1.15108299e+00 -9.73673284e-01
3.93605709e-01 2.91421771e-01 1.14798164e+00 -4.02032435e-01
-7.77441263e-01 2.82989472e-01 -7.05023766e-01 -4.21206385e-01
-2.10617125e-01 3.25249769e-02 2.22547516e-01 2.90339887e-01
8.42233300e-01 -4.72488135e-01 -2.37480119e-01 5.41181803e-01
-1.13136125e+00 -4.45896596e-01 6.34830475e-01 1.31702590e+00
4.60356802e-01 -1.24842145e-01 9.32177544e-01 -1.30253422e+00
5.35638928e-01 -7.34823167e-01 -3.27345669e-01 6.74874187e-01
-7.54247010e-01 1.67267159e-01 4.72527087e-01 -7.54755139e-01
-1.23095584e+00 8.54797959e-02 1.19215630e-01 -5.64787686e-01
-4.06103402e-01 2.28989989e-01 -2.34237716e-01 -1.28159940e-01
1.28383684e+00 2.47525230e-01 6.49698591e-03 -5.41108787e-01
4.66395050e-01 8.22764695e-01 3.62266392e-01 -2.74447829e-01
9.27722633e-01 4.62487459e-01 -3.13089550e-01 -8.95846069e-01
-8.47758353e-01 -4.93529081e-01 -8.48242521e-01 -9.06444415e-02
6.00475371e-01 -8.52341890e-01 6.93059340e-02 4.10408497e-01
-6.33809507e-01 -2.66175777e-01 -7.28583634e-01 2.88310736e-01
-5.99271715e-01 1.49569795e-01 -1.78821713e-01 -3.96244586e-01
-3.60695273e-01 -7.37378716e-01 1.03162432e+00 5.80123007e-01
-1.71927810e-01 -8.73395860e-01 5.30382097e-01 9.26811527e-03
5.65687835e-01 1.37934119e-01 6.99716687e-01 -1.01481760e+00
1.72819287e-01 -1.08629711e-01 -1.14308499e-01 1.75349370e-01
3.21780980e-01 -4.28151399e-01 -1.34139121e+00 -3.41723233e-01
2.41557866e-01 -8.09664547e-01 8.30774188e-01 1.27368197e-01
9.35510099e-01 -6.53302968e-02 -2.79882103e-01 5.80380678e-01
1.70700109e+00 2.75448382e-01 6.96594417e-01 4.12118882e-01
6.95692971e-02 6.26443207e-01 9.61550236e-01 7.60063708e-01
-1.91691853e-02 6.32593036e-01 2.65395045e-01 2.45927289e-01
-9.76785794e-02 -5.67897260e-02 1.90715417e-01 3.28630567e-01
1.73170581e-01 3.25132757e-01 -8.07701409e-01 8.37886572e-01
-1.97008979e+00 -1.21805918e+00 7.61021793e-01 1.97778225e+00
1.03855467e+00 4.62412462e-02 1.63729981e-01 -7.95834959e-02
9.11858261e-01 2.86050022e-01 -9.45518970e-01 -2.78825432e-01
7.85869807e-02 5.91523409e-01 7.66070709e-02 -3.23267025e-03
-1.29323685e+00 1.07458746e+00 6.35757065e+00 1.15103495e+00
-1.03039122e+00 5.07077515e-01 4.72062349e-01 -1.28212139e-01
1.30883411e-01 -1.90754458e-01 -8.85314643e-01 4.50756311e-01
8.94351363e-01 -2.50297785e-01 1.99799240e-01 1.37866926e+00
-3.57110351e-01 1.44496262e-01 -1.19159818e+00 1.18112779e+00
5.74971914e-01 -1.46909261e+00 3.78252715e-02 -3.98323536e-01
1.09399247e+00 3.46082896e-02 -8.18863213e-02 9.48285639e-01
1.86295778e-01 -8.43211889e-01 1.84574813e-01 4.95522141e-01
1.08253980e+00 -6.27423048e-01 7.20245957e-01 4.48004544e-01
-9.13885653e-01 -3.10382307e-01 -1.00971293e+00 1.57183930e-01
-4.79914024e-02 3.21728051e-01 -1.12419546e+00 4.97061193e-01
7.69289374e-01 6.38716459e-01 -8.08030844e-01 1.17887247e+00
5.70717975e-02 4.17091221e-01 2.90928364e-01 7.13178515e-02
7.76261911e-02 1.42028913e-01 1.43888786e-01 1.04864776e+00
3.00168902e-01 1.31304473e-01 6.67605624e-02 5.49194753e-01
4.70915549e-02 2.22870097e-01 -8.17347109e-01 -1.16207302e-01
3.68193686e-01 1.25791049e+00 -4.47954059e-01 -6.29339457e-01
-6.99482024e-01 1.06375086e+00 3.08123559e-01 1.22788824e-01
-7.42612660e-01 -8.55120361e-01 5.72201967e-01 -8.96910280e-02
7.11914957e-01 5.03391504e-01 1.03808500e-01 -1.25591147e+00
-3.02652091e-01 -9.08066273e-01 7.26630986e-01 -4.94681269e-01
-1.68370891e+00 7.07524240e-01 -5.78658693e-02 -1.57848322e+00
-5.55701017e-01 -4.78194714e-01 -7.44309068e-01 4.98215199e-01
-1.76626289e+00 -1.11661983e+00 -4.44894075e-01 1.00742567e+00
1.02213597e+00 -7.96674073e-01 1.20694721e+00 2.95792013e-01
-5.09343088e-01 6.26947522e-01 4.35298741e-01 1.77500889e-01
1.27782726e+00 -8.57068956e-01 3.52835745e-01 5.26007533e-01
9.48836207e-02 3.48421812e-01 6.34245574e-01 -5.99783421e-01
-1.25632310e+00 -1.36473858e+00 6.43423498e-01 -1.69075817e-01
5.66632807e-01 -1.60756752e-01 -1.15768409e+00 3.37886631e-01
2.83198595e-01 4.09987062e-01 1.10517430e+00 -1.15724895e-02
-4.74240243e-01 -3.86423260e-01 -1.44080055e+00 1.63624108e-01
9.30895984e-01 -4.16033328e-01 -1.04885292e+00 3.35244298e-01
5.18769741e-01 -4.76282611e-02 -7.59084523e-01 5.20799994e-01
2.52771407e-01 -8.58088315e-01 8.40514600e-01 -1.17374992e+00
2.44837672e-01 -3.83376554e-02 -4.20800149e-01 -1.57866096e+00
-5.45925856e-01 -5.14423966e-01 -2.45468006e-01 1.31370842e+00
-3.79167359e-05 -2.40630656e-01 7.46179342e-01 3.96551013e-01
1.70657992e-01 -4.09453005e-01 -9.67746973e-01 -1.13607574e+00
-1.08136021e-01 -1.24164611e-01 9.08213973e-01 1.15237093e+00
1.16088875e-01 6.22882068e-01 -7.51637995e-01 -2.56558150e-01
7.14710474e-01 4.35118616e-01 8.26106966e-01 -1.56812596e+00
-3.77886981e-01 -2.20281184e-02 -5.94023883e-01 -2.72478938e-01
1.30519405e-01 -7.58254826e-01 1.49530262e-01 -1.22723174e+00
4.86098975e-01 -1.35436710e-02 -4.06553298e-01 6.86741114e-01
-2.37351879e-01 4.37732160e-01 1.62126705e-01 8.40819106e-02
-7.83635437e-01 8.26730788e-01 9.67962205e-01 -4.56739247e-01
-2.34798700e-01 4.59682420e-02 -7.56344438e-01 6.60616994e-01
4.87434536e-01 -4.84636128e-01 -5.58216095e-01 -1.82919409e-02
-2.94004589e-01 -1.34137541e-01 -1.04905032e-02 -1.36105072e+00
3.36559832e-01 -5.33588290e-01 6.46234035e-01 -8.50794911e-02
3.88427109e-01 -7.59562492e-01 -7.86131322e-02 4.26108867e-01
-1.20059587e-01 -5.74212670e-01 5.76982386e-02 7.91397989e-01
-2.26569638e-01 -5.55422485e-01 1.26887274e+00 -3.19635123e-01
-1.81148338e+00 5.90063334e-01 -2.25784723e-02 4.49549705e-01
1.68286550e+00 -3.76014113e-01 -2.41037831e-01 5.20375371e-02
-9.30442870e-01 1.03000134e-01 2.91845351e-01 7.29104161e-01
8.31445754e-01 -1.64249194e+00 -6.96978629e-01 3.96112740e-01
7.99011171e-01 -4.28791642e-01 6.32873535e-01 2.83869684e-01
1.04270592e-01 2.41238311e-01 -7.19536901e-01 -4.61811185e-01
-1.04477489e+00 1.03844571e+00 1.30250573e-01 4.32635657e-02
-4.80025768e-01 9.10295129e-01 -1.09056905e-01 -7.12906420e-01
2.67059773e-01 3.26867074e-01 -5.18419981e-01 5.87945521e-01
1.00370741e+00 4.21295404e-01 3.11488193e-02 -5.27455926e-01
-3.73686403e-01 6.54414952e-01 -1.17636211e-01 1.05128445e-01
1.83201659e+00 -5.16526448e-03 4.22809899e-01 6.09391212e-01
1.11036265e+00 -6.20018065e-01 -1.42481530e+00 -8.44703197e-01
2.07728297e-01 -6.34151757e-01 -4.64526683e-01 -7.69766986e-01
-8.50388765e-01 9.18175995e-01 1.02225077e+00 -1.54115602e-01
1.11717379e+00 9.22053307e-02 6.57480419e-01 3.94614190e-01
5.62636673e-01 -1.73559761e+00 6.36437654e-01 5.46878278e-01
5.64454973e-01 -1.73290801e+00 -8.53355005e-02 1.23721011e-01
-1.00488520e+00 1.06523788e+00 9.56648350e-01 -1.42375836e-02
9.19246018e-01 -9.93856341e-02 -2.21714661e-01 -7.33426493e-03
-7.94815063e-01 -5.02551556e-01 2.68312782e-01 1.06489468e+00
-2.80008931e-02 -1.21660374e-01 -4.37289523e-03 7.93962061e-01
4.21129793e-01 5.50039649e-01 3.05422276e-01 1.08445859e+00
-8.91342163e-01 -1.23032200e+00 -3.01227093e-01 4.76542234e-01
-2.44170353e-02 1.43925130e-01 -2.52390176e-01 5.22763252e-01
4.19425100e-01 8.62915576e-01 -1.09861076e-01 -5.41642666e-01
4.25167471e-01 4.68315959e-01 2.10361972e-01 -1.23405182e+00
-2.42408529e-01 -2.87904352e-01 -3.15138787e-01 -4.96049762e-01
-4.18817431e-01 -4.56730485e-01 -8.21766257e-01 1.08328223e-01
-2.36544997e-01 2.31430173e-01 1.88786939e-01 1.02478456e+00
5.28075695e-01 3.44144374e-01 8.00731719e-01 -7.50220895e-01
-1.03400111e+00 -1.19606030e+00 -1.03157151e+00 6.36420310e-01
6.12928830e-02 -8.78514886e-01 -1.07255690e-01 1.31281912e-01] | [10.03745174407959, 3.055036783218384] |
26d8ef36-19db-408e-a85b-1da375252171 | towards-optimal-energy-management-strategy | 2305.12365 | null | https://arxiv.org/abs/2305.12365v1 | https://arxiv.org/pdf/2305.12365v1.pdf | Towards Optimal Energy Management Strategy for Hybrid Electric Vehicle with Reinforcement Learning | In recent years, the development of Artificial Intelligence (AI) has shown tremendous potential in diverse areas. Among them, reinforcement learning (RL) has proven to be an effective solution for learning intelligent control strategies. As an inevitable trend for mitigating climate change, hybrid electric vehicles (HEVs) rely on efficient energy management strategies (EMS) to minimize energy consumption. Many researchers have employed RL to learn optimal EMS for specific vehicle models. However, most of these models tend to be complex and proprietary, making them unsuitable for broad applicability. This paper presents a novel framework, in which we implement and integrate RL-based EMS with the open-source vehicle simulation tool called FASTSim. The learned RL-based EMSs are evaluated on various vehicle models using different test drive cycles and prove to be effective in improving energy efficiency. | ['Marco F. Huber', 'Christof Nitsche', 'Elisabeth Wedernikow', 'Xinyang Wu'] | 2023-05-21 | null | null | null | null | ['energy-management'] | ['time-series'] | [-3.34274232e-01 -3.17039788e-02 -5.34737229e-01 -1.53147563e-01
-1.46666959e-01 -2.81402111e-01 5.16944230e-01 1.13224760e-01
-3.19586903e-01 1.16493678e+00 -5.46189487e-01 -4.07865673e-01
-3.33520681e-01 -1.00824082e+00 -4.78885919e-01 -7.89020360e-01
-4.98492606e-02 3.98348302e-01 1.06697232e-01 -3.70604008e-01
3.27072740e-01 7.69250751e-01 -2.04465246e+00 -4.18931395e-01
1.14910507e+00 5.65932333e-01 5.41627765e-01 3.54425013e-01
-1.74854189e-01 6.03766680e-01 -6.47362471e-01 2.65766621e-01
-2.19792128e-02 -4.07930076e-01 -5.51941216e-01 -3.39762032e-01
-6.43672228e-01 -4.97436002e-02 -3.82155389e-01 8.28421831e-01
2.68958896e-01 2.55625397e-01 5.90020061e-01 -1.66032934e+00
2.86483586e-01 4.21581894e-01 -2.43221208e-01 7.29341358e-02
-1.31133407e-01 4.42482591e-01 4.74307030e-01 -3.78279865e-01
3.82191479e-01 9.07113552e-01 2.72752851e-01 6.53125405e-01
-9.30723548e-01 -9.26819563e-01 -1.89638197e-01 1.07198215e+00
-1.39950001e+00 -2.67351836e-01 1.10340810e+00 -6.41503185e-02
1.38948441e+00 1.55158564e-01 1.00478816e+00 5.05861998e-01
7.42298543e-01 8.96335661e-01 1.35495830e+00 -5.06408811e-01
6.75290823e-01 5.88871896e-01 -7.91060776e-02 4.51741248e-01
3.97940397e-01 5.25899887e-01 1.42373502e-01 2.42658854e-01
1.11631602e-01 -3.27342123e-01 1.59492210e-01 -4.97321755e-01
-6.62689924e-01 1.09761524e+00 3.36311609e-01 4.97515768e-01
-4.24351692e-01 2.59529352e-01 4.05347914e-01 1.14430465e-01
6.49772063e-02 5.19965172e-01 -3.63134623e-01 -3.20660651e-01
-5.37463725e-01 4.31859702e-01 6.79885566e-01 6.03432655e-01
9.92354155e-01 5.69193602e-01 2.99809575e-01 7.82134831e-01
5.28592706e-01 7.02155471e-01 3.89788777e-01 -1.06310225e+00
-6.57258555e-02 7.24086225e-01 1.25670791e-01 -8.46712768e-01
-7.03035176e-01 -2.53738344e-01 -7.94186473e-01 6.38440788e-01
-3.09071809e-01 -3.18771243e-01 -4.38061625e-01 1.55102062e+00
2.34455198e-01 2.41151541e-01 3.94678295e-01 4.76971209e-01
4.08869505e-01 1.13964987e+00 4.52651232e-01 -4.28474277e-01
6.90368772e-01 -6.92352295e-01 -1.06107128e+00 -2.20518842e-01
6.22824192e-01 -2.02705801e-01 4.80805844e-01 4.89482522e-01
-8.82712066e-01 -5.82975626e-01 -1.61811376e+00 6.87851667e-01
-7.75545180e-01 -9.59674791e-02 6.73169851e-01 8.01041007e-01
-8.84771824e-01 6.19659722e-01 -8.51703048e-01 -3.34923059e-01
1.82861716e-01 5.96760511e-01 1.45783469e-01 1.69143438e-01
-1.41035116e+00 1.41826582e+00 7.63647258e-01 -1.47249922e-01
-9.48868990e-01 -5.98734915e-01 -7.87802100e-01 -3.93646285e-02
4.47350413e-01 -3.25091243e-01 1.14664793e+00 -6.71423733e-01
-1.98271227e+00 1.31275773e-01 1.89126819e-01 -8.00900757e-01
1.99392289e-01 8.46565366e-02 -8.71865749e-01 9.45471600e-02
-4.86015171e-01 6.16759956e-01 5.70403755e-01 -1.35685849e+00
-8.50194871e-01 5.75191602e-02 -1.48481339e-01 5.46241067e-02
-4.69926953e-01 -2.89023012e-01 1.63823202e-01 -2.23703250e-01
-6.85767591e-01 -9.82104838e-01 -2.72288650e-01 -5.75974464e-01
1.78683490e-01 -7.42183328e-01 1.33266902e+00 -3.71354431e-01
1.39690816e+00 -1.64546227e+00 1.30461589e-01 4.53568071e-01
-2.64176250e-01 7.10640311e-01 8.34256336e-02 5.46184301e-01
3.24381709e-01 -3.80142555e-02 3.47849950e-02 1.61121905e-01
1.05456345e-01 7.18409777e-01 1.49245396e-01 1.65057972e-01
2.50159472e-01 7.18772769e-01 -8.25729489e-01 -4.72167194e-01
1.05967093e+00 4.71586138e-01 -3.34308386e-01 2.67452121e-01
-4.96376276e-01 2.07169712e-01 -7.18170524e-01 3.53264362e-01
5.46892583e-01 3.01019609e-01 3.25869143e-01 -6.99811354e-02
-3.63543123e-01 -3.90678167e-01 -1.10290301e+00 1.03536844e+00
-7.66191423e-01 7.65117049e-01 -1.09874839e-02 -1.44566929e+00
9.23307121e-01 1.46272242e-01 8.13364089e-01 -1.16810930e+00
3.95469666e-01 1.76362813e-01 -1.42345373e-02 -7.55286098e-01
4.00990158e-01 8.78546834e-02 -1.70086268e-02 6.93715215e-02
-4.37286161e-02 -4.51718241e-01 5.58069587e-01 -2.45473087e-01
8.79860818e-01 2.75550485e-01 3.62539887e-01 -2.81580001e-01
1.06110191e+00 4.06867206e-01 7.66179800e-01 2.44448900e-01
-1.77026317e-01 -5.03912687e-01 -2.92659327e-02 -3.33157361e-01
-1.11220205e+00 -8.50312233e-01 -4.01340835e-02 3.38743836e-01
3.55464458e-01 -1.71665400e-01 -8.04257095e-01 -2.75276601e-01
-3.99151705e-02 1.46445835e+00 8.24367329e-02 -5.12603819e-01
-6.36553466e-01 -6.00448847e-01 1.83593586e-01 3.60416532e-01
8.75637889e-01 -1.22305238e+00 -1.12663496e+00 5.82615852e-01
2.68212259e-01 -9.44357216e-01 3.36319327e-01 1.70279965e-01
-8.17915142e-01 -1.06337726e+00 -1.83986157e-01 -5.58865905e-01
3.39273959e-01 2.56293088e-01 1.05681038e+00 2.68171042e-01
-4.09943819e-01 4.52070981e-01 -2.76529074e-01 -8.78221929e-01
-9.89671350e-01 2.08247632e-01 1.26581073e-01 -3.78379166e-01
4.55512643e-01 -2.55846858e-01 -4.77799654e-01 4.68266726e-01
-7.71031857e-01 1.55342966e-01 7.56265879e-01 7.22552121e-01
5.81473172e-01 9.66005981e-01 1.27546132e+00 -4.52483952e-01
6.62445426e-01 -6.43257737e-01 -1.15260303e+00 3.07708055e-01
-1.29689217e+00 7.90446922e-02 8.99397254e-01 -8.53364095e-02
-1.10826015e+00 -1.18365856e-02 -2.31393620e-01 -1.90865815e-01
-4.34800297e-01 4.22711372e-01 -3.74590755e-01 -2.50727385e-01
-4.45121750e-02 1.82633325e-01 3.42041373e-01 -2.75048204e-02
-1.48140010e-03 6.48321033e-01 1.71266481e-01 -3.69860709e-01
9.14465606e-01 -2.54017174e-01 3.71767998e-01 -9.59616542e-01
5.65247573e-02 -3.50465886e-02 5.18067628e-02 -7.65767097e-01
8.73755455e-01 -6.10630929e-01 -1.04672956e+00 5.34576237e-01
-6.24999702e-01 -4.61242497e-01 4.68858145e-03 5.87758124e-01
-6.01247966e-01 -1.24132745e-01 -1.39145538e-01 -9.77488041e-01
-3.36436778e-01 -1.26695120e+00 1.54894575e-01 9.33079481e-01
-9.65506211e-02 -1.02577066e+00 6.19264096e-02 1.68982834e-01
7.74394393e-01 1.73536748e-01 1.34216249e+00 -2.69916177e-01
-6.24794960e-01 -2.05055490e-01 2.92662710e-01 5.41817665e-01
-1.08741209e-01 2.42928892e-01 -7.18332946e-01 -3.85619283e-01
-2.71835536e-01 -1.36805192e-01 1.50746480e-01 2.43300483e-01
1.24467349e+00 -4.31722999e-02 -7.31350899e-01 4.91827354e-02
1.83996975e+00 9.53127146e-01 7.76024103e-01 6.28262520e-01
1.72178045e-01 3.52516323e-01 9.25646424e-01 3.87255758e-01
4.54629809e-01 6.64819658e-01 5.83882213e-01 2.24022120e-02
1.93967804e-01 -8.00324678e-02 4.00839150e-01 9.55365181e-01
3.83237824e-02 -2.71084040e-01 -8.92917514e-01 3.23032022e-01
-1.67705083e+00 -1.04728544e+00 2.43649527e-01 1.91236985e+00
4.80381906e-01 3.38090092e-01 7.32502267e-02 4.86716896e-01
3.13196808e-01 -2.04642445e-01 -6.77342296e-01 -8.63741577e-01
-2.13216431e-03 2.85228312e-01 6.52586758e-01 2.33906746e-01
-7.14576900e-01 7.12750912e-01 6.18639946e+00 8.50957155e-01
-1.17682815e+00 -2.15779260e-01 2.46326879e-01 2.94287175e-01
-2.91933268e-01 -1.41650647e-01 -6.51977837e-01 4.77198184e-01
1.56802762e+00 -7.49401748e-01 6.89195216e-01 9.04777765e-01
7.38671780e-01 -5.32590210e-01 -7.28844166e-01 7.49616563e-01
-3.39843333e-01 -1.40591061e+00 -1.49237528e-01 -1.19673528e-01
7.94858813e-01 -1.34662390e-01 -3.17367345e-01 7.68021524e-01
1.44631475e-01 -1.02944541e+00 3.58391732e-01 5.78922689e-01
2.21621752e-01 -1.62871945e+00 1.02003372e+00 5.08987069e-01
-1.10703683e+00 -4.37841892e-01 -2.54258484e-01 1.11161835e-01
2.78210416e-02 1.58684984e-01 -8.77322316e-01 6.63255215e-01
7.19471812e-01 4.88808364e-01 -4.87778276e-01 1.02557063e+00
2.48232409e-02 7.22018898e-01 -1.96906447e-01 -8.26823831e-01
1.67677209e-01 -4.36350584e-01 4.07459617e-01 7.39290357e-01
4.03767824e-01 -1.03468649e-01 8.63692723e-03 7.32670844e-01
3.20770264e-01 -7.07000569e-02 -7.21709549e-01 -2.38634333e-01
5.88675857e-01 1.47425103e+00 -6.41498387e-01 -1.67233720e-01
-4.67236876e-01 1.26704246e-01 -2.61062950e-01 2.92239368e-01
-1.14250445e+00 -6.09921873e-01 8.73932838e-01 -6.81945160e-02
6.85579702e-02 -2.95130759e-01 -8.91044587e-02 -2.56249040e-01
-6.27678454e-01 -8.07770312e-01 3.08790500e-03 -7.50375986e-01
-6.47915304e-01 2.77751118e-01 5.83422601e-01 -1.25914478e+00
-6.66970968e-01 -6.80756092e-01 -5.80769837e-01 4.64069515e-01
-1.83907437e+00 -5.98165095e-01 -4.08874869e-01 2.96980023e-01
7.57137775e-01 -6.57673538e-01 9.02533174e-01 4.31376696e-01
-8.17599654e-01 3.35967839e-01 4.23743516e-01 -5.19521475e-01
7.93921128e-02 -9.46320534e-01 -3.97711426e-01 4.78592485e-01
-4.51676518e-01 -6.53356239e-02 1.17339945e+00 -2.68719345e-01
-2.18078375e+00 -9.74488795e-01 2.24589720e-01 3.55279863e-01
4.61834997e-01 1.92703195e-02 -8.07298601e-01 1.35562077e-01
8.30753744e-01 -4.15640175e-01 2.77119488e-01 -7.07458198e-01
6.40892804e-01 -6.43784761e-01 -1.41956139e+00 7.08144546e-01
2.52461880e-01 4.09019887e-02 -3.10206115e-01 -1.91279545e-01
2.79042780e-01 1.48968780e-02 -9.63747263e-01 8.12538207e-01
2.69878954e-01 -7.51889288e-01 6.78158820e-01 -1.40622944e-01
-8.54984596e-02 -5.99571645e-01 2.43281558e-01 -1.82536840e+00
-1.32587716e-01 -4.91310865e-01 -4.36672330e-01 1.38303590e+00
2.03859597e-01 -8.18127453e-01 7.24680662e-01 4.91655290e-01
-1.92850471e-01 -8.51613998e-01 -9.08279538e-01 -8.21623385e-01
1.29211962e-01 -4.80662018e-01 7.41308153e-01 4.60954875e-01
-5.61102806e-03 2.05731496e-01 -3.63017097e-02 3.20388526e-02
6.30521953e-01 -2.19703585e-01 6.70071602e-01 -1.18363702e+00
4.66263667e-02 -6.67004526e-01 -4.24657881e-01 -9.48855579e-02
6.74416840e-01 -6.56366050e-01 4.65626381e-02 -1.61201799e+00
-6.49084225e-02 -4.15047258e-01 -5.85915267e-01 2.43426442e-01
2.36846223e-01 -2.49850973e-01 7.65788322e-03 -3.43414992e-01
-4.84453231e-01 8.18744242e-01 8.46414804e-01 -4.74189401e-01
-2.13316575e-01 1.14928447e-01 -6.85756356e-02 6.59408927e-01
1.37766314e+00 -4.07023996e-01 -8.11197519e-01 2.10076481e-01
1.89149112e-01 1.20591611e-01 1.13323741e-01 -1.45753336e+00
1.61324814e-01 -4.03795779e-01 2.76303828e-01 -8.71107996e-01
-5.86838983e-02 -1.39257753e+00 6.00017548e-01 9.41685617e-01
3.57186757e-02 2.23695725e-01 6.47606134e-01 4.28072751e-01
-2.42423043e-01 -2.67277122e-01 9.67375040e-01 2.94432878e-01
-1.39306247e+00 -2.73285080e-02 -9.81342196e-01 -3.31888258e-01
1.85710871e+00 -1.71745688e-01 -1.25443861e-01 -1.09077640e-01
-6.85178116e-02 7.69440234e-01 2.74072856e-01 8.11591506e-01
6.36734009e-01 -1.25967836e+00 -3.14891696e-01 1.22562133e-01
6.14998490e-02 -4.75633800e-01 3.03662002e-01 4.55790013e-01
-6.12511635e-01 5.49028158e-01 -5.46465695e-01 -4.66596305e-01
-1.20599830e+00 7.32361853e-01 5.31302571e-01 -2.97577888e-01
-4.86049533e-01 -2.51047760e-01 -6.03370905e-01 -2.12014332e-01
9.99351516e-02 5.14486879e-02 -4.77985084e-01 -3.08805853e-01
3.32300395e-01 9.01749849e-01 4.99748737e-02 -3.27871382e-01
-2.69762903e-01 5.21550655e-01 3.42725039e-01 9.16850939e-02
1.37508583e+00 -2.11720273e-01 2.47127295e-01 3.15879494e-01
8.99675131e-01 -4.80340064e-01 -1.10377634e+00 2.77332157e-01
3.71281952e-01 8.02953467e-02 4.21603382e-01 -8.26366067e-01
-1.21183014e+00 5.56715250e-01 1.04082918e+00 1.84985489e-01
1.31799042e+00 -5.05054653e-01 6.97260022e-01 6.29636824e-01
8.68182003e-01 -1.89916527e+00 -2.32702851e-01 3.19798172e-01
5.81005275e-01 -1.14669895e+00 6.30976930e-02 1.52193934e-01
-3.55094314e-01 1.27087772e+00 8.87267709e-01 3.32969166e-02
7.38661885e-01 5.69964349e-01 -7.56407678e-02 2.12929517e-01
-9.01424050e-01 2.66031502e-03 -5.85829057e-02 7.05715358e-01
3.93997729e-02 1.13752954e-01 -6.15980327e-01 2.05191389e-01
2.55115539e-01 2.57409722e-01 4.45436656e-01 1.07007742e+00
-6.33307099e-01 -1.27302384e+00 -2.90974021e-01 3.38635534e-01
-1.10353298e-01 7.99594223e-01 2.24106967e-01 1.19626212e+00
1.70473561e-01 1.22541451e+00 -1.62031189e-01 -5.76267958e-01
2.71543413e-01 3.60324886e-03 3.68368864e-01 9.91683826e-02
-3.97943944e-01 -4.73577410e-01 1.22617804e-01 -5.15460968e-01
-5.47580063e-01 -4.93068546e-01 -1.90103483e+00 -4.16754752e-01
-1.83823518e-02 5.44359803e-01 1.09612358e+00 1.04064608e+00
3.46397996e-01 8.72734666e-01 1.08976865e+00 -7.20113575e-01
-4.73328561e-01 -6.80265069e-01 -5.76562107e-01 -7.39446431e-02
-1.81847796e-01 -9.86833394e-01 -1.98954687e-01 -4.13154870e-01] | [5.514489650726318, 2.2643649578094482] |
e41409fe-5921-4263-86a5-2faea849f6fe | generative-cooperative-net-for-image | 1705.02887 | null | http://arxiv.org/abs/1705.02887v3 | http://arxiv.org/pdf/1705.02887v3.pdf | Generative Cooperative Net for Image Generation and Data Augmentation | How to build a good model for image generation given an abstract concept is a
fundamental problem in computer vision. In this paper, we explore a generative
model for the task of generating unseen images with desired features. We
propose the Generative Cooperative Net (GCN) for image generation. The idea is
similar to generative adversarial networks except that the generators and
discriminators are trained to work accordingly. Our experiments on hand-written
digit generation and facial expression generation show that GCN's two
cooperative counterparts (the generator and the classifier) can work together
nicely and achieve promising results. We also discovered a usage of such
generative model as an data-augmentation tool. Our experiment of applying this
method on a recognition task shows that it is very effective comparing to other
existing methods. It is easy to set up and could help generate a very large
synthesized dataset. | ['Tao Wan', 'Qiangeng Xu', 'Zengchang Qin'] | 2017-05-08 | null | null | null | null | ['facial-expression-generation'] | ['computer-vision'] | [ 4.89943266e-01 7.18679488e-01 2.70230681e-01 -3.59939367e-01
-6.52372181e-01 -4.29275244e-01 1.17568564e+00 -8.63783717e-01
1.72522366e-02 1.04657042e+00 1.29079506e-01 6.81459764e-03
3.65351558e-01 -1.00775504e+00 -7.27909088e-01 -9.72104967e-01
1.73731208e-01 5.85378230e-01 -1.77728966e-01 -5.19592702e-01
1.50887847e-01 5.57649255e-01 -1.58715522e+00 3.87705952e-01
7.20590293e-01 6.16585255e-01 7.88232759e-02 7.54432440e-01
8.55205208e-02 1.02359724e+00 -1.08869565e+00 -5.36087632e-01
3.32437515e-01 -9.70867574e-01 -7.38152683e-01 4.38529253e-01
3.02296340e-01 -1.93460330e-01 -1.89953908e-01 9.36835885e-01
9.08261836e-01 -2.76054750e-04 9.73516583e-01 -1.76189578e+00
-1.18628585e+00 6.27574861e-01 -1.71548262e-01 -3.06647211e-01
3.62184346e-01 2.29482606e-01 3.14821571e-01 -7.79123425e-01
8.77670467e-01 1.25119567e+00 4.96255398e-01 1.38864124e+00
-1.22911978e+00 -6.54796183e-01 -4.13584948e-01 -3.48553360e-01
-1.24828923e+00 -5.65782964e-01 8.37567329e-01 -3.27137023e-01
6.01525307e-01 3.48444432e-01 5.78295231e-01 1.71251202e+00
-3.22787799e-02 8.41705322e-01 1.14700520e+00 -8.26370239e-01
1.91253290e-01 2.79224485e-01 -6.15889966e-01 5.69850981e-01
2.66219247e-02 4.48808104e-01 -2.78470159e-01 -6.48561269e-02
1.05242574e+00 -1.64028883e-01 -1.81849480e-01 -3.77142251e-01
-1.00411201e+00 1.16279650e+00 4.75411713e-01 4.92751241e-01
-2.31421322e-01 6.47352457e-01 2.23084353e-02 3.97186488e-01
4.26306784e-01 6.66079700e-01 1.97925702e-01 1.57373160e-01
-8.64336610e-01 4.33482200e-01 8.47488225e-01 1.03994346e+00
4.79122788e-01 6.74276769e-01 -4.79030520e-01 7.70592630e-01
1.02528088e-01 4.19041604e-01 8.48323405e-01 -9.20551896e-01
8.79198164e-02 3.55432779e-01 1.89346466e-02 -7.44310141e-01
1.14619620e-01 -3.14445972e-01 -1.06957257e+00 6.37636960e-01
2.57141709e-01 -6.70841694e-01 -1.38455343e+00 1.87112617e+00
-9.58651602e-02 7.30247721e-02 2.01976642e-01 6.33463144e-01
9.45831299e-01 7.89457619e-01 7.73971975e-02 4.62454483e-02
8.21648002e-01 -9.28838313e-01 -7.57524967e-01 -4.57688048e-02
3.08509558e-01 -7.95431435e-01 7.57329881e-01 3.31309497e-01
-1.17635930e+00 -9.83382642e-01 -8.86496246e-01 3.51188868e-01
-4.23949301e-01 2.52522767e-01 1.00559568e+00 1.04750562e+00
-1.46496832e+00 4.97039974e-01 -4.26101595e-01 -3.07034194e-01
3.92644137e-01 4.77982312e-01 -3.81864518e-01 3.87140475e-02
-1.07138598e+00 7.12055624e-01 5.22069633e-01 5.37224440e-03
-1.35596609e+00 2.64451504e-02 -8.92487466e-01 -2.65746891e-01
-6.28828406e-02 -1.01196826e+00 1.21746814e+00 -1.40273654e+00
-1.72930837e+00 9.06896710e-01 2.73101360e-01 -4.44951713e-01
6.41661763e-01 1.02685630e-01 -3.15858662e-01 -1.92803979e-01
-6.97831884e-02 1.37616003e+00 1.10115385e+00 -1.58107805e+00
-1.87141657e-01 -9.46548767e-04 -1.46473691e-01 -4.50405926e-02
-2.37213328e-01 -2.00574279e-01 -1.37935102e-01 -1.03120244e+00
-2.07955405e-01 -9.28603590e-01 -5.65505743e-01 -2.86642015e-01
-6.28393888e-01 -1.55863062e-01 8.41394305e-01 -2.03992218e-01
7.10920990e-01 -1.85047030e+00 1.57689050e-01 1.69646561e-01
-6.24699146e-03 4.11663234e-01 -3.33382010e-01 6.89634919e-01
-3.77845973e-01 2.88340271e-01 -2.65984863e-01 -2.97109574e-01
-1.02787819e-02 3.75353873e-01 -5.35306156e-01 -3.38070653e-02
5.75537145e-01 1.37503469e+00 -8.12866747e-01 -2.32575461e-01
3.79110090e-02 5.91001272e-01 -5.89907110e-01 4.48638827e-01
-3.25467676e-01 5.45845389e-01 -3.60362619e-01 5.47024727e-01
4.70826089e-01 2.62177258e-04 5.90283051e-02 1.95904136e-01
3.81555259e-01 -4.12210941e-01 -9.31034327e-01 1.54834020e+00
-3.12728494e-01 7.45599627e-01 -3.69979382e-01 -1.16241932e+00
1.57127988e+00 4.99152601e-01 -2.52978411e-02 -4.50389147e-01
2.76564866e-01 1.85295776e-01 5.26969694e-02 -3.90859485e-01
3.25505495e-01 -4.30999219e-01 -1.39957279e-01 4.94467676e-01
5.20880640e-01 -6.14855945e-01 8.88302401e-02 1.14462920e-01
9.84183311e-01 3.84251833e-01 2.01487139e-01 -1.31977975e-01
4.41979885e-01 -1.78578705e-01 1.34562001e-01 9.47218776e-01
2.56217152e-01 9.87334490e-01 4.12568569e-01 -4.63887095e-01
-1.29504395e+00 -9.03098166e-01 4.20247108e-01 6.47271276e-01
-2.80091673e-01 -8.10926482e-02 -1.12066305e+00 -7.54516780e-01
-3.21672201e-01 7.68074393e-01 -8.50536883e-01 -3.54994476e-01
-4.32917833e-01 -9.20189857e-01 6.72213197e-01 6.16618931e-01
6.31093383e-01 -1.66884291e+00 -1.97173312e-01 3.32825005e-01
-3.43202949e-02 -8.28370810e-01 -1.52347371e-01 1.50320292e-01
-6.05992258e-01 -6.91485882e-01 -1.14376497e+00 -1.14363170e+00
9.74427402e-01 -2.74122119e-01 1.19833624e+00 9.28137228e-02
-3.81280780e-01 2.31666028e-01 -5.48980296e-01 -7.14009643e-01
-1.08607924e+00 6.47183927e-03 -2.74676412e-01 8.10226202e-02
-2.50785816e-02 -6.75620139e-01 -3.42501163e-01 2.81024426e-01
-1.25195563e+00 1.39578834e-01 5.31222522e-01 9.95964825e-01
2.30359018e-01 -1.41400933e-01 9.76694167e-01 -1.15623140e+00
1.05090415e+00 -1.96977749e-01 -4.75739509e-01 1.56895623e-01
-4.56636459e-01 2.35780135e-01 6.60564244e-01 -6.18521035e-01
-1.00191879e+00 3.95788938e-01 -3.50540340e-01 -3.75155866e-01
-3.82381141e-01 -5.08577228e-02 -3.22980881e-01 -1.89659163e-01
1.02477455e+00 3.99768740e-01 2.25661278e-01 -2.59803742e-01
6.64046109e-01 6.11879587e-01 6.75783932e-01 -4.46447164e-01
9.80740488e-01 2.51203448e-01 5.48116341e-02 -4.53109443e-01
-3.12170058e-01 3.08681577e-01 -3.77864361e-01 -1.72992945e-01
7.06316769e-01 -7.99457192e-01 -3.61591697e-01 6.44076645e-01
-1.26658893e+00 -4.71785665e-01 -7.71733463e-01 1.08852655e-01
-1.02154756e+00 7.26856617e-03 -3.80766571e-01 -8.14622760e-01
-3.62992972e-01 -1.01718187e+00 1.12154973e+00 3.98436069e-01
-1.08387724e-01 -1.06887007e+00 2.82602042e-01 1.63301155e-01
6.40693486e-01 8.17918479e-01 4.86936212e-01 -6.61455750e-01
-6.51478946e-01 -2.72638559e-01 1.60206482e-01 6.85640693e-01
2.42899165e-01 1.46667331e-01 -1.04777169e+00 -1.74556911e-01
2.87619531e-02 -6.74475610e-01 9.19996202e-01 1.49588048e-01
1.19422770e+00 -4.25124586e-01 -3.47187996e-01 5.84113479e-01
1.44333684e+00 5.08423746e-01 1.41610622e+00 -3.03233732e-02
5.53453684e-01 4.18416977e-01 3.29458088e-01 1.78449482e-01
-2.66639292e-01 7.26308167e-01 3.90691638e-01 -4.79573727e-01
-4.30898547e-01 -4.04326767e-01 4.59444761e-01 3.85412604e-01
-4.02114600e-01 -7.56146371e-01 -6.67924523e-01 3.92230272e-01
-1.77585781e+00 -1.28095925e+00 -4.03645858e-02 1.77512765e+00
9.04405475e-01 -6.61904812e-02 5.55090308e-02 2.00558156e-01
7.93992698e-01 -1.45865783e-01 -1.85018759e-02 -7.25594759e-01
-2.51582921e-01 8.30811620e-01 9.64611769e-02 3.00389200e-01
-1.01421177e+00 1.10563469e+00 7.57099819e+00 1.14761817e+00
-1.14602780e+00 2.03953101e-03 1.00525951e+00 2.75004655e-01
-3.30651134e-01 -1.66999459e-01 -7.82424390e-01 4.82936323e-01
8.26793611e-01 3.63857672e-03 2.21763164e-01 9.34598088e-01
-1.65487170e-01 1.92560092e-01 -1.07672596e+00 9.82283115e-01
3.79406154e-01 -1.46526384e+00 6.00690544e-01 1.25088960e-01
1.12204790e+00 -5.72288930e-01 1.96025804e-01 3.08580697e-01
7.80250371e-01 -1.60263908e+00 6.37134194e-01 6.76102221e-01
9.60756540e-01 -7.85277784e-01 7.65157938e-01 3.98429155e-01
-5.58340728e-01 1.20314151e-01 -4.24591720e-01 4.10779454e-02
1.15449198e-01 2.88111866e-01 -1.05674779e+00 4.46722388e-01
2.04492118e-02 2.56540120e-01 -6.61429405e-01 9.48058546e-01
-5.82848191e-01 4.21172172e-01 9.70648304e-02 -1.11725584e-01
3.25100392e-01 -5.63511811e-02 2.94292659e-01 1.02110946e+00
6.05993569e-01 -1.47440761e-01 -1.71961084e-01 1.18529904e+00
-1.78676397e-01 -1.15157679e-01 -1.15267372e+00 -9.99719128e-02
-5.02980426e-02 1.25628853e+00 -8.07819963e-01 -4.70046908e-01
1.37116745e-01 1.27388299e+00 1.21212617e-01 2.78308481e-01
-9.62362349e-01 -4.38578814e-01 9.77226943e-02 1.38498154e-02
3.57500643e-01 1.23322867e-01 3.38903740e-02 -1.00370312e+00
-2.87069678e-01 -1.05548525e+00 -1.76575020e-01 -1.23489392e+00
-1.33021510e+00 1.17769241e+00 -1.55316561e-01 -1.33462310e+00
-1.01996350e+00 -5.23184597e-01 -7.80948877e-01 9.17115331e-01
-9.75287080e-01 -1.44105804e+00 -5.85411012e-01 7.70849109e-01
4.83297735e-01 -6.18101120e-01 1.12826514e+00 -1.65816658e-04
-1.89115658e-01 7.48074532e-01 -1.75127655e-01 3.52641523e-01
4.72294331e-01 -1.18503928e+00 6.80215001e-01 8.65349770e-01
5.65310478e-01 2.77024597e-01 7.15451181e-01 -5.11548221e-01
-8.31786036e-01 -1.19813132e+00 7.81207442e-01 -5.68061411e-01
1.05164438e-01 -5.58198631e-01 -4.88745481e-01 5.48760056e-01
5.78358471e-01 -1.01417854e-01 5.83355188e-01 -5.22340536e-01
-9.74503458e-02 4.41326723e-02 -1.32377505e+00 5.27817011e-01
1.25124896e+00 -8.45387429e-02 -4.49995756e-01 4.69445735e-01
4.36921179e-01 -2.44203925e-01 -4.11000907e-01 3.69887412e-01
2.46238291e-01 -1.10527575e+00 8.69589150e-01 -6.88606322e-01
7.70970643e-01 -1.12400807e-01 4.54675369e-02 -1.64264667e+00
-2.02628553e-01 -1.07150352e+00 2.58432925e-01 1.49228656e+00
4.04243022e-01 -6.29987359e-01 8.63410592e-01 3.02508771e-01
-2.72536781e-02 -5.54130197e-01 -4.96847808e-01 -1.02288651e+00
7.26254875e-05 -8.35810155e-02 7.12794542e-01 8.66273582e-01
-4.11567211e-01 5.66290736e-01 -6.46781683e-01 -4.18135434e-01
4.47961777e-01 6.16638064e-02 1.13816404e+00 -1.02876651e+00
-3.93079579e-01 -1.82076842e-01 -6.69245243e-01 -6.29343867e-01
1.20815828e-01 -9.43633795e-01 -9.53978568e-04 -1.44234586e+00
1.04427956e-01 -4.69370186e-01 2.53446579e-01 6.66027069e-01
-1.96698438e-02 7.15966403e-01 3.02707404e-01 -1.28744477e-02
-1.16164982e-01 6.05355918e-01 1.58391559e+00 -1.21362180e-01
-9.15901586e-02 1.51095152e-01 -1.01117647e+00 4.31922525e-01
9.46767032e-01 -4.81889755e-01 -6.21633410e-01 -2.07169168e-02
-1.53753355e-01 3.38647626e-02 5.42919695e-01 -1.08428037e+00
-8.99616256e-02 1.07783824e-01 6.65535688e-01 -1.38417661e-01
4.02903527e-01 -4.53048229e-01 7.06871986e-01 5.44574797e-01
-2.88202345e-01 -2.01956704e-01 -2.85706623e-03 1.74432695e-01
-5.57413518e-01 -4.56211567e-01 8.57280374e-01 -4.59661603e-01
-6.08647287e-01 1.14825912e-01 -3.76190186e-01 -1.76754802e-01
1.20767701e+00 -3.37572962e-01 -1.93890989e-01 -8.51315618e-01
-9.40073609e-01 -3.00948977e-01 4.17381823e-01 5.34627140e-01
7.55295753e-01 -1.72371197e+00 -9.23938036e-01 4.89492238e-01
-9.54059809e-02 -2.79194653e-01 -2.81461626e-01 -1.57385990e-02
-5.88775814e-01 2.98420668e-01 -6.30931795e-01 -3.60063165e-01
-1.01635957e+00 6.39577806e-01 4.71043706e-01 -3.43329102e-01
-1.41116291e-01 9.56150770e-01 3.77986819e-01 -2.49931663e-01
-1.26599167e-02 -7.05871591e-03 -2.09930912e-01 -6.70412853e-02
4.11923736e-01 -1.14261553e-01 -1.33339405e-01 -4.70501393e-01
9.80922803e-02 4.00805682e-01 2.65174210e-01 -2.35593900e-01
1.41937029e+00 3.77380371e-01 -9.56835076e-02 -6.56866804e-02
1.01207221e+00 -1.34187296e-01 -1.10065508e+00 3.42037082e-01
-4.30149347e-01 -2.61449784e-01 -4.44441140e-01 -7.78862536e-01
-1.44124687e+00 7.72459149e-01 5.93744755e-01 4.11821842e-01
1.22649372e+00 -5.47589250e-02 3.50758523e-01 1.43022001e-01
5.46862602e-01 -6.95547700e-01 4.55116391e-01 1.83502331e-01
1.32227659e+00 -1.01065350e+00 -2.66454101e-01 -4.76131618e-01
-6.74048603e-01 1.20614469e+00 6.58234596e-01 -4.69375163e-01
3.96106422e-01 4.09236997e-01 1.28668100e-01 -1.69433549e-01
-7.76729763e-01 -3.30486178e-01 2.65939683e-01 1.16675818e+00
4.98988837e-01 -2.67083142e-02 -2.68202752e-01 2.47973129e-01
-4.50155377e-01 2.77373761e-01 7.21411765e-01 8.22877049e-01
-2.17929348e-01 -1.70509088e+00 -3.37561220e-01 1.00318931e-01
-3.85091960e-01 -4.50636372e-02 -7.65408993e-01 9.94452536e-01
3.82422417e-01 8.22253466e-01 3.69131938e-02 -5.15576899e-01
9.50219333e-02 2.67150313e-01 7.38757432e-01 -7.50091195e-01
-6.54637814e-01 -9.89520829e-03 5.53280711e-02 -3.96187812e-01
-6.83346510e-01 -4.38474178e-01 -6.59924090e-01 -5.54156415e-02
-2.99933612e-01 1.50022134e-01 5.80847204e-01 4.69648093e-01
1.87106147e-01 4.26947564e-01 9.41190660e-01 -9.16096330e-01
-4.42237556e-01 -1.09212279e+00 -6.19556367e-01 6.13476455e-01
-7.59804621e-02 -4.72426802e-01 -2.71467328e-01 6.46904349e-01] | [11.749236106872559, -0.226897731423378] |
2fc4efee-0262-4501-8b6e-f43c0a54f4cf | verifying-global-neural-network | 2306.12495 | null | https://arxiv.org/abs/2306.12495v1 | https://arxiv.org/pdf/2306.12495v1.pdf | Verifying Global Neural Network Specifications using Hyperproperties | Current approaches to neural network verification focus on specifications that target small regions around known input data points, such as local robustness. Thus, using these approaches, we can not obtain guarantees for inputs that are not close to known inputs. Yet, it is highly likely that a neural network will encounter such truly unseen inputs during its application. We study global specifications that - when satisfied - provide guarantees for all potential inputs. We introduce a hyperproperty formalism that allows for expressing global specifications such as monotonicity, Lipschitz continuity, global robustness, and dependency fairness. Our formalism enables verifying global specifications using existing neural network verification approaches by leveraging capabilities for verifying general computational graphs. Thereby, we extend the scope of guarantees that can be provided using existing methods. Recent success in verifying specific global specifications shows that attaining strong guarantees for all potential data points is feasible. | ['Stefan Leue', 'David Boetius'] | 2023-06-21 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [ 1.94539309e-01 4.66533929e-01 -6.16760075e-01 -4.01048183e-01
-4.69966561e-01 -8.93493354e-01 3.52610856e-01 3.04811716e-01
-2.16842238e-02 6.67698026e-01 -1.98380008e-01 -7.70248950e-01
-5.58660090e-01 -1.00338078e+00 -1.19307625e+00 -3.46981764e-01
-3.87535363e-01 -8.63407701e-02 4.65018362e-01 -7.06486106e-02
-8.09267014e-02 8.61365795e-01 -1.81333959e+00 1.93995222e-01
1.95368960e-01 8.84903491e-01 -8.43557358e-01 7.46923089e-01
3.38154674e-01 5.94535530e-01 -6.52409434e-01 -1.59245148e-01
4.09716427e-01 -1.78315192e-01 -6.93435848e-01 -4.66489732e-01
7.16411471e-01 -5.09287298e-01 -3.89516987e-02 1.27782559e+00
-1.01800628e-01 -1.89085811e-01 7.21114427e-02 -2.12556171e+00
-5.45523226e-01 1.18417776e+00 1.06963843e-01 -3.28079164e-01
6.89615235e-02 4.28826809e-01 1.27445340e+00 1.87822774e-01
4.75716352e-01 8.57340038e-01 8.80908668e-01 1.03815806e+00
-1.40386271e+00 -6.29812539e-01 5.06887913e-01 -5.69785357e-01
-1.36582983e+00 -6.46027565e-01 4.40665126e-01 -3.62230450e-01
1.07964003e+00 7.36156762e-01 2.26094261e-01 8.49776566e-01
2.97960222e-01 5.97670078e-01 5.94630003e-01 -9.41187367e-02
3.88892621e-01 1.49575204e-01 7.20951319e-01 5.18096209e-01
7.70225525e-01 5.26425242e-01 -2.89084405e-01 -3.22026670e-01
7.12101400e-01 -4.34906632e-01 -4.11137074e-01 -3.99394363e-01
-1.18113184e+00 4.73598719e-01 1.31426930e-01 2.59704649e-01
2.64711797e-01 7.36920238e-01 6.61698997e-01 6.28304422e-01
-1.52091682e-01 5.93698978e-01 -1.01332951e+00 -2.81165875e-02
-7.02010453e-01 3.53947312e-01 9.82587993e-01 1.31198406e+00
3.90827894e-01 3.22267950e-01 -1.19112983e-01 -3.21506202e-01
2.94054002e-01 3.93971443e-01 -1.93441957e-01 -1.00433218e+00
1.95291698e-01 6.09149873e-01 1.10990606e-01 -7.75871575e-01
-3.40704292e-01 -4.59747314e-01 -4.48368579e-01 5.47598243e-01
5.68076253e-01 -3.50763440e-01 -6.46091759e-01 2.30413866e+00
1.19500205e-01 2.04743370e-01 2.93621480e-01 5.99293888e-01
1.89418867e-01 2.88413852e-01 -1.18506253e-01 -2.79206216e-01
7.00990379e-01 -2.87900865e-01 -4.35735673e-01 1.02019690e-01
6.65161431e-01 2.79674202e-01 9.77096736e-01 5.03076255e-01
-1.26960671e+00 -1.40533194e-01 -1.36955237e+00 2.00633958e-01
-4.05121803e-01 -2.97741264e-01 6.94303334e-01 1.21571517e+00
-1.24281251e+00 6.06144190e-01 -9.82425809e-01 5.24101928e-02
8.44170153e-02 1.03363431e+00 -4.70287561e-01 4.64028984e-01
-9.62266028e-01 5.67125320e-01 5.61793268e-01 9.84407067e-02
-1.22196686e+00 -1.03754890e+00 -9.65531647e-01 1.53039664e-01
3.06424052e-01 -4.46363926e-01 1.32356083e+00 -9.88094628e-01
-1.34499526e+00 2.74373651e-01 4.17487293e-01 -6.59100175e-01
4.57642406e-01 3.50258112e-01 -5.34083903e-01 -3.07005018e-01
-6.57495499e-01 2.96295464e-01 7.10768163e-01 -1.22421610e+00
-4.90147322e-01 -1.87739894e-01 6.93925500e-01 -4.89353120e-01
-5.86917758e-01 1.40672565e-01 1.48668706e-01 1.31796738e-02
-2.98855960e-01 -6.79788232e-01 -1.43215939e-01 3.68756205e-01
-4.61796582e-01 1.60688281e-01 7.24947095e-01 5.84223680e-02
9.44888473e-01 -2.19783473e+00 -2.08961651e-01 8.72769952e-01
2.12438226e-01 2.37870947e-01 -1.71513051e-01 9.30327028e-02
-6.36020750e-02 6.47468746e-01 -2.44620387e-02 1.31999344e-01
6.03877962e-01 3.09333444e-01 -5.94716311e-01 7.07154214e-01
2.34477952e-01 1.02942753e+00 -6.14938378e-01 3.88871059e-02
-1.07016258e-01 3.70540261e-01 -7.74567187e-01 -1.07839182e-01
-6.72853887e-01 -1.79922462e-01 -2.23642617e-01 5.13567865e-01
4.68688697e-01 -3.68221961e-02 3.90737772e-01 1.93970963e-01
1.74138993e-02 3.50262195e-01 -1.40732694e+00 1.02225554e+00
-3.76374632e-01 3.49273860e-01 4.96232122e-01 -5.39651871e-01
6.55035198e-01 5.60246646e-01 5.46876527e-02 -1.84433967e-01
4.46351349e-01 6.99505955e-02 3.03439088e-02 1.16052136e-01
1.61675483e-01 -6.68633357e-02 -2.99738497e-01 6.40910983e-01
-1.60352737e-01 4.86852974e-02 -9.23683569e-02 -2.99504790e-02
1.22510934e+00 -3.55272107e-02 -9.70493350e-03 -5.87499022e-01
4.23956245e-01 -2.18078882e-01 7.41245747e-01 8.96459222e-01
-1.11761846e-01 2.64546067e-01 1.25414610e+00 -7.82947540e-02
-1.12978625e+00 -1.04747617e+00 8.43840018e-02 6.38738632e-01
-1.00285724e-01 -4.74883318e-01 -8.77945602e-01 -7.86838830e-01
2.46708184e-01 5.99194586e-01 -6.61587059e-01 -5.26608348e-01
-1.03618823e-01 5.16581833e-02 1.24869943e+00 9.27668691e-01
-2.35936921e-02 -7.54669309e-01 -7.76445270e-01 -1.34464398e-01
8.58927667e-01 -1.00721323e+00 -4.56163555e-01 2.53543854e-01
-9.07432079e-01 -1.05746150e+00 1.85695216e-01 -4.14389491e-01
9.66577530e-01 -3.56962442e-01 6.97942734e-01 6.12776339e-01
1.75347760e-01 3.67841154e-01 3.74886811e-01 -2.94325471e-01
-6.05231583e-01 -1.20192528e-01 2.46905342e-01 -1.89400032e-01
-1.24086842e-01 -5.89843750e-01 1.66014686e-01 2.38793120e-01
-9.95264411e-01 -2.76432246e-01 1.18527882e-01 5.49705565e-01
2.96649516e-01 3.36275250e-01 6.35393679e-01 -1.12883556e+00
8.33485067e-01 -1.29372820e-01 -1.37617886e+00 5.09858370e-01
-9.23858762e-01 3.32225025e-01 9.95573521e-01 -7.72484899e-01
-5.96632779e-01 4.45408607e-03 1.61805257e-01 -4.11618054e-01
-3.81945133e-01 4.91307020e-01 -6.13175809e-01 -3.03734452e-01
8.84362638e-01 -1.93281978e-01 1.34798050e-01 1.74839586e-01
2.92552322e-01 3.40885639e-01 5.47216237e-01 -1.24252820e+00
1.07994437e+00 1.37664229e-01 3.11807930e-01 -2.44209915e-01
-2.25558981e-01 2.02765077e-01 -4.00714017e-02 -4.50100116e-02
1.33318976e-01 -4.39214319e-01 -1.69997060e+00 4.61805284e-01
-7.43800461e-01 -5.55311501e-01 -4.85760093e-01 9.68546271e-02
-5.05436361e-01 -6.76653814e-03 -5.15064776e-01 -1.15459144e+00
-2.03957632e-01 -1.28044677e+00 4.95022237e-01 -5.41355796e-02
-5.27193785e-01 -9.68010426e-01 -3.47945601e-01 -4.53133970e-01
7.85338819e-01 2.80704558e-01 9.54096377e-01 -9.02464390e-01
-8.27682137e-01 -4.61029768e-01 -1.42898783e-02 5.13543010e-01
-5.68910278e-02 7.26135075e-01 -9.71865654e-01 -2.86087543e-01
-3.74270767e-01 -2.46439099e-01 8.11866149e-02 6.59141466e-02
8.95436287e-01 -8.17454278e-01 -6.20257435e-03 4.77445990e-01
1.42179632e+00 -3.28574590e-02 4.00094420e-01 -1.15709780e-02
2.93885887e-01 4.09135431e-01 1.57780781e-01 3.24443907e-01
1.97422970e-02 3.56154710e-01 6.41966403e-01 3.16387713e-01
3.57739747e-01 -2.26001039e-01 4.39071596e-01 -1.00480951e-01
1.90409422e-01 -1.90582313e-02 -8.75812292e-01 8.91664326e-01
-1.93069768e+00 -7.99910843e-01 -9.11135692e-03 2.51813078e+00
1.21535432e+00 6.18222594e-01 2.14776427e-01 5.62451363e-01
4.34082627e-01 -2.99159259e-01 -5.64733088e-01 -8.85552466e-01
1.31641611e-01 1.76614314e-01 9.07833338e-01 8.11394393e-01
-6.40537679e-01 5.13261974e-01 7.29100275e+00 1.76069260e-01
-1.06806886e+00 -1.57006413e-01 2.32327089e-01 -3.78565975e-02
-9.65048134e-01 1.83258295e-01 -8.91981244e-01 2.96897884e-03
1.34715927e+00 -5.27328014e-01 5.83222747e-01 1.04188979e+00
-9.87362266e-02 3.98779392e-01 -1.58301461e+00 1.68499842e-01
-2.63195276e-01 -1.24117768e+00 -2.58359224e-01 -7.78540075e-02
7.66001999e-01 -4.29834038e-01 2.39490688e-01 1.17570162e-01
7.29882300e-01 -1.16184258e+00 8.33445311e-01 4.33629215e-01
8.30795109e-01 -1.27015913e+00 4.94551003e-01 4.09963906e-01
-8.02242696e-01 -2.25915387e-01 -5.04631363e-02 -1.83968619e-02
-3.69473279e-01 4.91047770e-01 -6.07277274e-01 3.31804276e-01
1.18125007e-01 2.83316404e-01 -2.89695412e-01 8.75126719e-01
-5.57404578e-01 2.92743415e-01 -6.61837339e-01 -4.18681130e-02
-1.95327803e-01 4.16256905e-01 3.11469048e-01 9.23278630e-01
1.65786877e-01 -2.23376572e-01 -2.49643713e-01 1.17249966e+00
-9.01721641e-02 -5.03599465e-01 -8.81634891e-01 -2.79530495e-01
4.02608305e-01 8.67597222e-01 -2.87321568e-01 -1.67100858e-02
-2.75503397e-01 1.17489733e-01 9.16768610e-02 3.99615109e-01
-9.75971341e-01 -4.60988075e-01 1.20209002e+00 -1.73155785e-01
-7.66220242e-02 -4.40622300e-01 -6.66116714e-01 -9.42555606e-01
1.83824196e-01 -1.16528618e+00 5.32397807e-01 -4.60585028e-01
-8.32103968e-01 2.98154980e-01 2.60411173e-01 -7.07644582e-01
-2.15978280e-01 -5.63475907e-01 -4.17946368e-01 6.52554512e-01
-1.17808533e+00 -1.38942480e+00 2.54520059e-01 8.02640855e-01
-4.78367001e-01 2.21852779e-01 8.91774774e-01 6.70458600e-02
-4.29025441e-01 1.27106166e+00 -7.48424232e-01 -5.23800738e-02
1.22806281e-01 -9.88113046e-01 4.84715223e-01 1.45165324e+00
-1.05112165e-01 1.29703391e+00 8.76263261e-01 -5.72765887e-01
-1.88017762e+00 -1.16259646e+00 7.89603472e-01 -3.52329075e-01
8.44201565e-01 -4.54871535e-01 -9.17249620e-01 1.28554893e+00
-2.06293538e-01 3.84783894e-01 1.32871240e-01 2.19298527e-01
-9.87320125e-01 -3.77923846e-01 -1.33699417e+00 7.25489378e-01
1.04028106e+00 -8.50425005e-01 -3.47352773e-02 1.13065965e-01
1.06302655e+00 -5.62292278e-01 -1.20336950e+00 5.71643531e-01
6.49550676e-01 -6.75997674e-01 7.30836153e-01 -1.06535244e+00
-1.37689203e-01 -6.70082211e-01 -2.21692428e-01 -7.42443740e-01
1.47789836e-01 -1.07440746e+00 -5.23943126e-01 1.35095787e+00
8.94728720e-01 -9.47537601e-01 7.36522615e-01 1.62156618e+00
-8.22206810e-02 -2.68943518e-01 -8.50254893e-01 -1.23424661e+00
2.58600295e-01 -1.01460826e+00 1.30151057e+00 9.58718240e-01
6.98937893e-01 -3.47435474e-01 -1.50640234e-01 7.31128573e-01
4.38810021e-01 -8.76190811e-02 5.68225563e-01 -1.02931607e+00
-4.87130195e-01 -6.78760231e-01 -7.01113880e-01 -3.90701383e-01
5.33892453e-01 -9.20243919e-01 1.52095497e-01 -1.12251306e+00
-1.81836843e-01 -5.69970191e-01 -6.19610608e-01 1.20255470e+00
6.10829234e-01 4.36921977e-03 -9.34438854e-02 -5.53073049e-01
-2.58572161e-01 -1.14765830e-01 8.29463840e-01 -1.47799700e-01
-1.33879900e-01 2.20919445e-01 -8.89633894e-01 3.20988446e-01
9.80437338e-01 -3.38013381e-01 -7.18000352e-01 -2.32348457e-01
7.81466663e-01 3.46108899e-02 5.33686578e-01 -1.11991537e+00
5.89043856e-01 -7.57974446e-01 -3.17296982e-01 9.88003463e-02
5.23590762e-03 -1.28071570e+00 6.67868555e-01 3.98040712e-01
-8.61004472e-01 -1.62501007e-01 4.35208410e-01 2.66835928e-01
1.72188684e-01 -9.02986228e-02 5.56792498e-01 6.19327307e-01
-3.05768818e-01 4.02472466e-01 -2.10800841e-01 -1.39814854e-01
1.28582227e+00 1.15508055e-02 -7.89983392e-01 -2.74779052e-01
-6.50607228e-01 2.23527744e-01 8.06513548e-01 3.01658034e-01
7.36996710e-01 -1.02320290e+00 -2.43848056e-01 5.91244876e-01
1.50332615e-01 -1.95060238e-01 -2.78718978e-01 7.51536250e-01
-1.02169171e-01 2.92244405e-01 -2.24935532e-01 -1.97689861e-01
-1.49418366e+00 9.92108226e-01 6.89336061e-01 3.79793555e-01
-4.26218331e-01 5.04862905e-01 -3.33318502e-01 -6.92370772e-01
5.51069379e-01 -1.09635186e+00 3.77294570e-01 -6.22645676e-01
5.85600793e-01 -1.14148155e-01 2.54520446e-01 5.24199530e-02
-6.30382836e-01 2.95401871e-01 3.14969182e-01 -3.06030124e-01
1.05828547e+00 4.79822993e-01 -2.71331608e-01 1.42111495e-01
9.13202643e-01 1.70064196e-01 -1.19475067e+00 2.20693156e-01
1.66611578e-02 -2.73190588e-01 9.36096534e-03 -7.84575403e-01
-1.21206605e+00 8.12106490e-01 8.40409473e-02 3.25975895e-01
1.25658834e+00 -2.99787730e-01 3.93786132e-01 6.76562667e-01
6.00474298e-01 -6.37819290e-01 -8.23915184e-01 5.35977781e-01
4.05768096e-01 -3.94748867e-01 -2.21339419e-01 -2.34945729e-01
-1.72692269e-01 1.18449974e+00 5.85779965e-01 -1.99336130e-02
4.98571873e-01 1.26113176e+00 -2.07994893e-01 2.03213498e-01
-1.25467229e+00 1.94498271e-01 1.85174838e-01 7.39912808e-01
3.81211713e-02 9.92794931e-02 6.44264789e-03 8.83970797e-01
8.89019389e-03 1.54455438e-01 9.01286602e-01 1.10628343e+00
-3.01544696e-01 -1.23977983e+00 -1.50406763e-01 2.74138093e-01
-6.65812433e-01 -3.52558121e-02 -3.56857955e-01 1.05179977e+00
-1.52833670e-01 8.60514760e-01 -1.12896025e-01 -5.49093366e-01
2.40770385e-01 -5.33244349e-02 6.73691630e-01 -5.05100310e-01
-8.71922910e-01 -3.82094353e-01 4.85897124e-01 -6.73070788e-01
-1.06391132e-01 -5.27692974e-01 -1.68989921e+00 -5.04392803e-01
-4.51297820e-01 9.01020691e-02 7.19038546e-01 7.96320140e-01
3.12480330e-01 5.21269381e-01 2.61249572e-01 1.05219878e-01
-7.81518757e-01 2.69807875e-02 -5.33105671e-01 -1.65966406e-01
8.30978572e-01 8.96550119e-02 -5.04144192e-01 3.93884107e-02] | [6.141655445098877, 7.565188884735107] |
45a0a996-6cea-4f60-98f7-358813001a6f | building-chinese-affective-resources-in | null | null | https://aclanthology.org/N16-1066 | https://aclanthology.org/N16-1066.pdf | Building Chinese Affective Resources in Valence-Arousal Dimensions | null | ['Xue-jie Zhang', 'Lung-Hao Lee', 'Liang-Chih Yu', 'K. Robert Lai', 'Jun Hu', 'Shuai Hao', 'Yunchao He', 'Jin Wang'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['twitter-sentiment-analysis'] | ['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] | [-7.391371726989746, 3.774693489074707] |
8dcc72b1-22d3-4793-b849-3a2814e7e7b3 | spherical-view-synthesis-for-self-supervised | 1909.08112 | null | https://arxiv.org/abs/1909.08112v1 | https://arxiv.org/pdf/1909.08112v1.pdf | Spherical View Synthesis for Self-Supervised 360 Depth Estimation | Learning based approaches for depth perception are limited by the availability of clean training data. This has led to the utilization of view synthesis as an indirect objective for learning depth estimation using efficient data acquisition procedures. Nonetheless, most research focuses on pinhole based monocular vision, with scarce works presenting results for omnidirectional input. In this work, we explore spherical view synthesis for learning monocular 360 depth in a self-supervised manner and demonstrate its feasibility. Under a purely geometrically derived formulation we present results for horizontal and vertical baselines, as well as for the trinocular case. Further, we show how to better exploit the expressiveness of traditional CNNs when applied to the equirectangular domain in an efficient manner. Finally, given the availability of ground truth depth data, our work is uniquely positioned to compare view synthesis against direct supervision in a consistent and fair manner. The results indicate that alternative research directions might be better suited to enable higher quality depth perception. Our data, models and code are publicly available at https://vcl3d.github.io/SphericalViewSynthesis/. | ['Federico Alvarez', 'Antonis Karakottas', 'Dimitrios Zarpalas', 'Petros Daras', 'Nikolaos Zioulis'] | 2019-09-17 | null | null | null | null | ['3d-depth-estimation'] | ['computer-vision'] | [ 6.27994686e-02 2.06248879e-01 -2.53971875e-01 -4.10493642e-01
-4.40001220e-01 -6.39010847e-01 7.75048971e-01 -3.89872193e-01
-4.45467621e-01 7.28169262e-01 3.42006087e-01 -3.43809485e-01
7.34746456e-02 -6.50793076e-01 -6.64627492e-01 -6.07869387e-01
3.53797346e-01 3.82690020e-02 -1.23272367e-01 3.54889706e-02
4.78202999e-01 5.32335520e-01 -1.86653125e+00 9.26708207e-02
5.60481369e-01 8.46055925e-01 4.07941669e-01 8.55669498e-01
3.84916395e-01 6.73846245e-01 -2.12452397e-01 -2.84052223e-01
5.73362470e-01 -1.75643921e-01 -6.90200865e-01 1.20862149e-01
9.48212326e-01 -8.40931416e-01 -4.37411845e-01 7.67485023e-01
7.42641211e-01 -1.33745342e-01 3.18851650e-01 -1.10518348e+00
-3.10856313e-01 -8.14478844e-02 -3.88616949e-01 1.46487162e-01
5.85157990e-01 3.56761217e-01 8.71472418e-01 -8.84244919e-01
8.39595795e-01 8.94916356e-01 3.05687398e-01 6.57463253e-01
-8.25481832e-01 -3.66925448e-01 2.67800003e-01 2.54481792e-01
-9.51878369e-01 -7.75163710e-01 7.01994061e-01 -3.89712602e-01
1.24590480e+00 -2.01436952e-02 7.60436058e-01 1.39218688e+00
1.33795857e-01 7.83260047e-01 1.39524329e+00 -5.66161990e-01
1.52796388e-01 2.10325763e-01 -3.23859453e-01 6.27786577e-01
3.10100287e-01 5.82935750e-01 -7.68609047e-01 5.03222704e-01
1.01427674e+00 -2.45704204e-01 -5.19148409e-01 -9.60814655e-01
-1.11411536e+00 6.56171739e-01 5.55201113e-01 8.95864889e-03
-1.11128300e-01 2.44646724e-02 1.71006233e-01 3.58614743e-01
6.01379097e-01 5.03296196e-01 -5.22525012e-01 -4.96960759e-01
-7.65560329e-01 3.06750834e-01 6.89689815e-01 9.14367080e-01
7.61995733e-01 1.44244403e-01 4.53384668e-01 5.16965508e-01
3.89152437e-01 4.99949515e-01 2.71021336e-01 -1.50081825e+00
6.24836028e-01 2.21868619e-01 1.66667085e-02 -8.20924461e-01
-3.70746702e-01 -4.62357521e-01 -4.76129502e-01 8.46405327e-01
7.90911198e-01 -1.53214991e-01 -7.36844122e-01 1.58454621e+00
2.92054862e-01 -9.54903141e-02 2.65544266e-01 1.01672542e+00
7.10007191e-01 1.18644908e-01 -5.99604070e-01 1.02502823e-01
9.90946174e-01 -1.05644190e+00 -2.56585658e-01 -4.38728750e-01
6.66166842e-01 -7.40484059e-01 1.15615535e+00 9.12895381e-01
-1.05810511e+00 -3.77784014e-01 -1.03556168e+00 -4.96696174e-01
-3.08068275e-01 7.51801729e-02 8.93533170e-01 6.13571763e-01
-1.35922337e+00 4.06523854e-01 -9.50759411e-01 -5.04551172e-01
3.60056549e-01 1.64328337e-01 -5.54184139e-01 -4.32460994e-01
-8.28824341e-01 9.62466359e-01 2.00165734e-01 1.36219531e-01
-7.65622437e-01 -5.93674302e-01 -1.19166744e+00 -3.89453441e-01
3.80232602e-01 -1.01954544e+00 1.40891898e+00 -7.58099020e-01
-1.65218699e+00 1.01276445e+00 -2.60015666e-01 -4.20246989e-01
8.45796585e-01 -2.84589708e-01 2.83351094e-01 2.96247363e-01
-1.07528521e-02 1.05327845e+00 5.54082394e-01 -1.42937553e+00
-5.81408918e-01 -6.49214268e-01 6.84404433e-01 7.00810969e-01
1.28829316e-03 -6.21945679e-01 -2.74476886e-01 -5.96375205e-02
3.73174787e-01 -7.62510061e-01 3.51381190e-02 1.45940974e-01
-3.20315212e-01 1.92830473e-01 5.07507801e-01 -4.35989439e-01
5.78766406e-01 -1.96038949e+00 1.71813086e-01 -2.29959786e-01
3.92523706e-01 -9.34433267e-02 3.25844467e-01 4.83796120e-01
2.75175292e-02 -2.60591090e-01 8.89686942e-02 -7.20814645e-01
-1.36116087e-01 1.49598256e-01 -1.60076380e-01 5.42149544e-01
-7.31075034e-02 8.57146144e-01 -9.14147139e-01 -2.49607176e-01
8.00659359e-01 4.89483029e-01 -8.31172228e-01 1.58655241e-01
-7.34282956e-02 6.88524187e-01 2.08546370e-02 8.10625672e-01
5.53660572e-01 -1.84085727e-01 -9.86300930e-02 -6.09197542e-02
-5.00691950e-01 4.20612246e-01 -1.02897000e+00 2.03023481e+00
-8.53052855e-01 1.03493798e+00 2.14973152e-01 -6.90703988e-01
6.78761661e-01 1.04048744e-01 3.06353092e-01 -9.51191545e-01
6.31640851e-02 3.97778660e-01 -1.63278505e-01 -4.02615428e-01
6.39046788e-01 -8.09139088e-02 5.31363487e-01 4.87240590e-02
-3.34109622e-03 -6.47305369e-01 -1.08585032e-02 1.65913906e-02
7.08978355e-01 7.25285113e-01 5.58014393e-01 5.44636212e-02
8.42648968e-02 -1.66321784e-01 9.67688337e-02 6.13940299e-01
-2.20722467e-01 9.69837964e-01 2.79701501e-01 -2.45529473e-01
-1.10578310e+00 -1.09016955e+00 -2.86019385e-01 3.50167811e-01
5.85125424e-02 -2.17060447e-01 -5.38152874e-01 -2.71076351e-01
-1.29073322e-01 5.05244970e-01 -5.46516180e-01 3.59811515e-01
-1.83009982e-01 -1.20639749e-01 2.85687894e-01 5.55375695e-01
6.47035658e-01 -9.04646993e-01 -1.16500008e+00 -1.72353655e-01
-1.52950853e-01 -1.20856225e+00 6.60878420e-02 3.85695130e-01
-1.03794622e+00 -1.16589189e+00 -9.35546219e-01 -4.33171660e-01
4.33948845e-01 6.27954245e-01 9.96299803e-01 -3.30513328e-01
-9.54694301e-02 8.25530589e-01 -1.98910415e-01 -4.99546707e-01
1.82124227e-01 2.16649830e-01 -3.07525340e-02 -3.70985031e-01
3.66591871e-01 -8.20591927e-01 -1.05670130e+00 6.86296150e-02
-5.91517627e-01 4.28000510e-01 4.77871150e-01 5.18939495e-01
3.70321989e-01 -5.06332934e-01 1.34992853e-01 -7.35137820e-01
1.21477231e-01 -3.54409188e-01 -8.32112372e-01 -3.78009439e-01
-9.57911551e-01 -1.55964792e-01 4.46876764e-01 1.01674356e-01
-1.12858546e+00 -1.01922154e-01 -1.35485485e-01 -4.39499706e-01
-5.20448446e-01 3.42507988e-01 -1.66780710e-01 -1.16709985e-01
8.28755975e-01 2.39967674e-01 9.44053754e-02 -1.82685003e-01
4.04235333e-01 5.87994993e-01 3.81039262e-01 -3.04779232e-01
7.04919040e-01 1.09061742e+00 6.17636293e-02 -1.05507016e+00
-7.86683738e-01 -4.83460277e-01 -9.09650266e-01 -3.50251406e-01
8.48450124e-01 -1.25760198e+00 -6.79657638e-01 6.84654295e-01
-9.01884794e-01 -5.91988385e-01 -1.39993727e-01 7.85069764e-01
-9.56277013e-01 4.36379939e-01 -4.50006098e-01 -7.82841802e-01
8.65271688e-02 -1.06306469e+00 1.33166099e+00 2.17302844e-01
-1.04086408e-02 -1.27518392e+00 1.05383217e-01 6.64345503e-01
2.02528030e-01 2.90746152e-01 3.60606015e-01 4.18815538e-02
-1.02936387e+00 1.86362475e-01 -2.72043347e-01 4.83037174e-01
1.54972970e-01 -1.36505395e-01 -1.51315486e+00 -4.01066959e-01
1.41970277e-01 -6.00000203e-01 8.37999105e-01 5.78923464e-01
8.30111086e-01 -3.59234698e-02 2.34756209e-02 1.07026136e+00
1.51441717e+00 1.32391855e-01 6.56043828e-01 8.16372335e-01
9.43722188e-01 9.39185381e-01 6.93114221e-01 4.31551427e-01
7.90866971e-01 6.99708939e-01 8.17826152e-01 -2.59369165e-01
-3.04873824e-01 -3.05687428e-01 6.40480816e-02 4.88901228e-01
-2.14916110e-01 -2.80881226e-01 -9.32198346e-01 5.55742323e-01
-1.44817805e+00 -1.01267135e+00 6.26236899e-03 2.17724586e+00
6.07160389e-01 2.00242892e-01 -3.10151037e-02 3.38977993e-01
-5.09922020e-02 3.67743939e-01 -5.86118162e-01 -4.52351242e-01
-2.14522243e-01 6.63295612e-02 6.75893664e-01 8.13944876e-01
-8.08207870e-01 7.49595284e-01 5.56569576e+00 7.13547766e-02
-1.43282890e+00 -2.38186121e-01 3.96775097e-01 -4.15584147e-01
-4.07909542e-01 8.96995887e-02 -7.26121306e-01 1.26386926e-01
6.65410221e-01 1.28126845e-01 5.65810561e-01 8.32091630e-01
5.20120025e-01 -5.60257316e-01 -1.37422013e+00 1.31731284e+00
2.01263383e-01 -1.01350820e+00 -4.12509978e-01 3.87627125e-01
8.48652482e-01 3.56227696e-01 3.24376762e-01 -1.03667364e-01
-5.00798114e-02 -1.05943239e+00 6.31090045e-01 4.37499613e-01
7.89496064e-01 -4.05938864e-01 5.28616548e-01 4.49599147e-01
-7.25867569e-01 4.87139709e-02 -2.34869391e-01 -7.17149496e-01
-2.54311133e-03 4.79248106e-01 -1.06384087e+00 9.11838055e-01
6.84146821e-01 1.14493024e+00 -6.44376338e-01 1.03517079e+00
-3.06674510e-01 1.85628220e-01 -3.22890908e-01 2.29100451e-01
2.19687134e-01 -7.97708482e-02 2.97354907e-01 7.81920195e-01
2.18990698e-01 -2.44269207e-01 -1.97705925e-01 5.24841011e-01
2.32924268e-01 -2.42627338e-02 -1.24271142e+00 3.01461190e-01
2.47390851e-01 9.77217078e-01 -5.27382731e-01 9.83687118e-02
-7.64268279e-01 8.37779284e-01 4.21826750e-01 4.43709731e-01
-3.66909117e-01 -4.75595631e-02 6.78340435e-01 2.50768751e-01
2.69724905e-01 -5.70307493e-01 -6.88992381e-01 -1.51725304e+00
3.39983612e-01 -8.70549798e-01 1.97374392e-02 -1.14109921e+00
-8.92949164e-01 4.14617091e-01 2.42519245e-01 -1.34440923e+00
-5.82816839e-01 -9.25081372e-01 -1.64007291e-01 7.80454934e-01
-2.07029009e+00 -1.10680580e+00 -7.81454921e-01 5.41289926e-01
6.59689963e-01 9.86207277e-02 6.61531031e-01 1.52318865e-01
-2.49831770e-02 2.68604189e-01 -1.82614941e-02 -2.67122805e-01
8.19364130e-01 -1.45468128e+00 3.02500486e-01 9.92794573e-01
2.14127868e-01 5.19839585e-01 7.72532701e-01 -2.49898598e-01
-1.35459220e+00 -4.93141741e-01 7.11789131e-01 -8.12288821e-01
2.99636245e-01 -3.77125949e-01 -5.21650374e-01 7.58436620e-01
2.88901150e-01 -1.58190921e-01 2.86362857e-01 7.28935599e-02
-2.30528533e-01 -5.27302325e-02 -8.77089381e-01 8.37620199e-01
1.22645509e+00 -6.66903615e-01 -4.01735336e-01 1.02777429e-01
4.37082291e-01 -8.43277991e-01 -6.12745225e-01 3.75728935e-01
8.24360073e-01 -1.71558559e+00 8.95316839e-01 9.18639898e-02
8.75090718e-01 -1.35088518e-01 -3.90206516e-01 -1.22374094e+00
2.46664464e-01 -2.01664805e-01 -7.66125098e-02 7.82061219e-01
1.77614972e-01 -9.20369744e-01 1.07513893e+00 2.58269191e-01
-2.58282483e-01 -8.63959312e-01 -9.59469259e-01 -6.20482743e-01
1.77351385e-01 -6.08259141e-01 5.18221706e-02 7.45992422e-01
-1.66556910e-01 1.76537514e-01 -3.88011754e-01 1.56209633e-01
6.70393229e-01 8.87249112e-02 1.14502740e+00 -9.80955422e-01
-5.51521957e-01 -3.63220662e-01 -4.79639143e-01 -1.56025267e+00
-3.01971789e-02 -6.65640354e-01 -5.64682931e-02 -1.69112480e+00
-1.59759119e-01 -1.82865068e-01 2.18875557e-01 8.53241086e-02
2.16851771e-01 3.44423950e-01 1.62972152e-01 1.43138722e-01
-1.61977172e-01 4.01432842e-01 1.49260032e+00 2.94824064e-01
-1.06074750e-01 1.38757616e-01 -7.24118292e-01 6.97567701e-01
9.68813956e-01 6.81611821e-02 -9.71251070e-01 -7.80489504e-01
2.35904351e-01 1.23057127e-01 6.34523451e-01 -1.07576489e+00
2.80043483e-01 -5.77453114e-02 3.63562018e-01 -6.28413498e-01
7.66052246e-01 -6.38661921e-01 -2.00949565e-01 1.64545745e-01
-5.79331703e-02 4.74530421e-02 1.12911142e-01 3.77198994e-01
-2.84587055e-01 -1.22417182e-01 5.24629116e-01 -4.26920205e-01
-8.61556113e-01 2.33029872e-01 8.13628286e-02 1.26396954e-01
8.25967193e-01 -9.19374347e-01 -3.58236641e-01 -6.97067201e-01
-5.22715569e-01 2.22106930e-02 1.03474748e+00 3.89260113e-01
8.15134466e-01 -8.93310487e-01 -4.44119573e-01 3.64316165e-01
3.20068955e-01 2.70350009e-01 2.40555018e-01 7.47391760e-01
-7.03465939e-01 8.68938386e-01 -3.28560680e-01 -8.38701844e-01
-1.13126242e+00 2.41700992e-01 5.05010426e-01 1.65160477e-01
-6.79983497e-01 7.68155217e-01 4.60203856e-01 -4.33230191e-01
3.14882129e-01 -5.18232405e-01 8.32033996e-03 -1.62355736e-01
2.04388142e-01 2.46270373e-01 1.94027469e-01 -2.05598518e-01
-2.79983759e-01 6.60441756e-01 1.17498219e-01 -4.58391339e-01
1.22631764e+00 -5.05462766e-01 3.64398807e-01 6.81936324e-01
1.16961205e+00 2.20834136e-01 -1.95821822e+00 1.05256289e-02
-2.60516465e-01 -9.19664264e-01 3.69897373e-02 -7.13758826e-01
-7.98358262e-01 1.13052893e+00 5.43177426e-01 -4.81289700e-02
1.15366077e+00 -7.93460757e-02 1.64214611e-01 2.68449396e-01
5.73374689e-01 -7.82856107e-01 1.69953868e-01 5.44674814e-01
7.06168234e-01 -1.65949476e+00 -3.91945653e-02 -3.84473264e-01
-5.06505132e-01 1.14183021e+00 8.04780245e-01 9.97404754e-02
3.89326781e-01 2.55164832e-01 3.45869064e-01 -2.80081272e-01
-7.91864157e-01 -3.39772820e-01 2.02679429e-02 9.89004433e-01
5.78138173e-01 -1.80771261e-01 1.36136070e-01 -3.30487758e-01
-6.05497420e-01 -3.91361713e-02 8.64433706e-01 9.71656024e-01
-2.40505815e-01 -1.05449641e+00 -4.34099585e-02 1.53016999e-01
-2.78680563e-01 -2.04641432e-01 -1.91530541e-01 1.07850051e+00
-3.23490351e-02 9.67487097e-01 -1.01948775e-01 -1.23755999e-01
3.81349295e-01 -1.84366688e-01 9.68746424e-01 -8.75066161e-01
-4.23199385e-02 -1.47646353e-01 1.45003825e-01 -7.62089550e-01
-6.00499928e-01 -7.52284050e-01 -7.88927674e-01 -3.25418770e-01
4.71889935e-02 -4.29703444e-01 8.74753594e-01 8.83618951e-01
2.09639564e-01 3.40669185e-01 5.09854376e-01 -1.14920437e+00
-4.16151911e-01 -6.69687092e-01 -2.87458211e-01 1.13321086e-02
6.92676365e-01 -7.08828390e-01 -6.19717717e-01 -1.02496371e-01] | [8.680229187011719, -2.4554920196533203] |
9aef4bf7-d320-4294-a25f-e47bb79e4a80 | sequence-to-sequence-data-augmentation-for | 1807.01554 | null | http://arxiv.org/abs/1807.01554v1 | http://arxiv.org/pdf/1807.01554v1.pdf | Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding | In this paper, we study the problem of data augmentation for language
understanding in task-oriented dialogue system. In contrast to previous work
which augments an utterance without considering its relation with other
utterances, we propose a sequence-to-sequence generation based data
augmentation framework that leverages one utterance's same semantic
alternatives in the training data. A novel diversity rank is incorporated into
the utterance representation to make the model produce diverse utterances and
these diversely augmented utterances help to improve the language understanding
module. Experimental results on the Airline Travel Information System dataset
and a newly created semantic frame annotation on Stanford Multi-turn,
Multidomain Dialogue Dataset show that our framework achieves significant
improvements of 6.38 and 10.04 F-scores respectively when only a training set
of hundreds utterances is represented. Case studies also confirm that our
method generates diverse utterances. | ['Yutai Hou', 'Ting Liu', 'Yijia Liu', 'Wanxiang Che'] | 2018-07-04 | sequence-to-sequence-data-augmentation-for-1 | https://aclanthology.org/C18-1105 | https://aclanthology.org/C18-1105.pdf | coling-2018-8 | ['text-augmentation'] | ['natural-language-processing'] | [ 5.43528855e-01 8.61567914e-01 8.08678493e-02 -8.51183474e-01
-7.89923370e-01 -6.60748482e-01 7.99578547e-01 2.07742956e-02
-4.21259731e-01 1.23269892e+00 8.22372437e-01 -1.89666748e-01
3.12575877e-01 -4.95255291e-01 -3.78169030e-01 -3.22583735e-01
3.07047516e-01 1.06460774e+00 7.28496686e-02 -1.10898602e+00
2.82560080e-01 -3.22264105e-01 -1.29765284e+00 7.62207210e-01
1.02970147e+00 5.52857995e-01 2.79535621e-01 9.76751029e-01
-5.44908702e-01 9.92905021e-01 -9.34051335e-01 -2.02489078e-01
1.95912704e-01 -6.48773730e-01 -1.44977784e+00 5.55221915e-01
2.02222124e-01 -5.05324304e-01 -1.71776186e-03 6.34940088e-01
5.80267966e-01 6.29502594e-01 5.60628891e-01 -1.16343141e+00
-5.95760643e-01 9.85772252e-01 -1.64544180e-01 1.56245410e-01
6.65178180e-01 -1.73133567e-01 9.71038163e-01 -5.71818113e-01
8.51803541e-01 1.69696879e+00 3.12012345e-01 9.55848873e-01
-9.26816106e-01 -2.25575760e-01 2.31833324e-01 -6.71957061e-03
-6.49953008e-01 -5.86346149e-01 7.02660859e-01 -1.98259905e-01
1.48213935e+00 1.13928996e-01 1.46519944e-01 1.05241811e+00
-4.94446814e-01 1.07541442e+00 9.29354727e-01 -8.48116398e-01
-7.92688727e-02 2.72593737e-01 6.61401451e-01 4.43576753e-01
-4.73063558e-01 -2.72790462e-01 -4.07980025e-01 -5.46907745e-02
3.96470010e-01 -6.02869868e-01 3.02114710e-02 3.29995342e-02
-9.90141034e-01 1.23234916e+00 -1.54353902e-01 3.09466779e-01
-4.03779268e-01 -2.82955974e-01 9.51067090e-01 6.09220505e-01
6.63426161e-01 7.43742168e-01 -7.52503157e-01 -5.36843359e-01
-1.35398999e-01 6.04210794e-01 9.90354836e-01 1.09028220e+00
5.71912587e-01 1.47515774e-01 -3.72699231e-01 1.43392706e+00
4.92434837e-02 7.03472272e-02 8.74642432e-01 -1.04355538e+00
8.10449481e-01 8.29414010e-01 2.09612489e-01 -4.92484093e-01
-5.36058068e-01 1.67104434e-02 -4.43832397e-01 -4.56522763e-01
3.83344203e-01 -5.07233143e-01 -9.49400127e-01 1.80678523e+00
5.74590206e-01 -2.24091575e-01 9.18924153e-01 6.79949701e-01
1.11083746e+00 7.08091736e-01 1.81405023e-01 -1.76079109e-01
1.35565352e+00 -1.41726077e+00 -1.14056420e+00 9.01848823e-02
1.20705092e+00 -7.27975130e-01 1.29279816e+00 2.81611294e-01
-9.27603781e-01 -6.99883401e-01 -8.11757386e-01 -2.36423627e-01
-3.47468048e-01 2.10489123e-03 5.43315291e-01 7.76473343e-01
-7.77978241e-01 -1.71300814e-01 -1.81615561e-01 -3.71585518e-01
-1.33625627e-01 2.92951912e-01 -2.60267943e-01 -5.52918576e-02
-1.55650997e+00 9.76059496e-01 8.88752520e-01 -5.40865779e-01
-6.82953894e-01 -5.60102582e-01 -1.05569732e+00 -2.86800176e-01
4.91396636e-01 -5.38953602e-01 1.64695776e+00 -9.69447732e-01
-1.81882155e+00 4.76442426e-01 -2.60100514e-01 -8.25323403e-01
3.13569576e-01 -2.66291916e-01 -3.27296227e-01 6.58558961e-03
1.12868488e-01 1.19404542e+00 3.16277653e-01 -1.24017692e+00
-8.88443828e-01 -1.90426603e-01 6.37757540e-01 8.40324104e-01
-1.70964688e-01 -1.91720143e-01 -3.39372717e-02 -5.42340159e-01
3.00202575e-02 -1.03991926e+00 -3.13954294e-01 -1.06272781e+00
-5.21936238e-01 -5.87123990e-01 9.35080230e-01 -6.99728072e-01
1.05552936e+00 -1.68098319e+00 1.61733985e-01 -2.14879483e-01
-1.93504646e-01 2.91309625e-01 -3.52724493e-01 6.60298765e-01
9.96783152e-02 8.88103619e-02 -3.34187418e-01 -4.25600946e-01
4.58322093e-02 5.15277982e-01 -1.82455301e-01 -3.40947002e-01
3.48013282e-01 7.81722486e-01 -8.69860172e-01 -1.65378273e-01
2.78646052e-01 8.17314461e-02 -8.24796855e-01 4.15556461e-01
-6.78864300e-01 7.56212533e-01 -4.96270776e-01 2.47020021e-01
4.45612967e-01 1.32605538e-01 2.39091918e-01 -5.99835664e-02
2.06208736e-01 6.25767648e-01 -7.90518999e-01 2.12071180e+00
-6.68735802e-01 3.89198720e-01 -3.62981170e-01 -9.46438730e-01
1.15122283e+00 6.43799543e-01 8.47703516e-02 -8.04096580e-01
4.08764035e-02 -5.50372526e-02 4.14384484e-01 -8.13602984e-01
1.18132114e+00 -3.26497436e-01 -4.68464673e-01 6.50483072e-01
2.14835718e-01 -1.33424118e-01 4.62709248e-01 4.61934149e-01
9.27475870e-01 2.68837437e-02 2.99676478e-01 -8.62946659e-02
7.61518478e-01 3.76476705e-01 3.05028468e-01 6.86735690e-01
-1.25860527e-01 4.16553497e-01 3.94465864e-01 -2.88002104e-01
-1.32864344e+00 -4.40263927e-01 1.51605662e-02 1.70641148e+00
-1.20053113e-01 -1.75169915e-01 -8.38501632e-01 -1.08184087e+00
-1.91564351e-01 1.16792893e+00 -5.45725465e-01 1.39619380e-01
-5.87275326e-01 -5.99855900e-01 9.98582661e-01 3.74749601e-01
8.48038554e-01 -1.01604915e+00 -2.37237811e-01 4.86157000e-01
-5.32814205e-01 -1.31315649e+00 -2.39161327e-01 -7.07509890e-02
-6.09240413e-01 -9.66921031e-01 -3.94261271e-01 -8.25191319e-01
5.74017227e-01 3.90986092e-02 1.10324192e+00 -3.16891484e-02
1.73328847e-01 4.26000863e-01 -9.70596015e-01 -4.28099632e-01
-1.18971908e+00 2.67671674e-01 -5.06325886e-02 -3.12766880e-01
5.90190232e-01 -1.23452269e-01 4.57522161e-02 1.13356628e-01
-9.04671967e-01 4.70121652e-01 9.14160311e-02 1.50569773e+00
-5.79192191e-02 -1.40913382e-01 1.47387910e+00 -1.32324505e+00
1.12217498e+00 -6.59450829e-01 -4.42039408e-02 2.67239034e-01
-2.28470922e-01 2.96015531e-01 4.12459970e-01 -2.77578443e-01
-1.75928903e+00 9.67288688e-02 -4.22503650e-01 1.55655533e-01
-5.22316217e-01 4.56689417e-01 -2.34220147e-01 4.40406889e-01
5.45443475e-01 2.36920625e-01 1.55432999e-01 -4.94220912e-01
7.14802504e-01 9.41337049e-01 3.32799286e-01 -8.01380336e-01
1.18932657e-01 -1.44938380e-01 -5.81444085e-01 -9.95009840e-01
-6.22274876e-01 -6.40273929e-01 -8.87718201e-01 -8.87670591e-02
7.64611363e-01 -8.62339139e-01 -4.90080416e-01 1.30917192e-01
-1.39340806e+00 -3.12220365e-01 -1.31717965e-01 2.02231050e-01
-5.58695018e-01 2.72390783e-01 -4.47753727e-01 -1.08397281e+00
-3.03067625e-01 -1.27399886e+00 9.37714636e-01 1.83292374e-01
-6.20951414e-01 -1.22543848e+00 1.18661903e-01 9.20595169e-01
2.83579677e-01 1.10832825e-01 1.14670694e+00 -1.60854471e+00
3.45407724e-02 1.12996101e-01 9.97442156e-02 4.75632340e-01
4.89102781e-01 -5.25848269e-01 -1.08890080e+00 -6.75616637e-02
3.28932912e-03 -9.73222375e-01 4.27918166e-01 5.17436340e-02
5.80574870e-01 -6.11612260e-01 -7.47993030e-03 -2.79456079e-01
9.21974838e-01 7.15829372e-01 4.66585606e-01 3.27414423e-01
5.74035704e-01 1.11452687e+00 1.11008167e+00 4.94964540e-01
7.11145878e-01 6.60199046e-01 -4.07294556e-03 1.26085237e-01
-1.91218019e-01 -2.52196103e-01 -1.92628112e-02 1.07267213e+00
3.32673341e-01 -5.86679935e-01 -8.50639641e-01 7.61667252e-01
-1.73905814e+00 -7.78719425e-01 -1.22142397e-01 1.60253549e+00
1.26509583e+00 -3.77313830e-02 2.50081420e-01 8.12149495e-02
6.41770065e-01 3.38495634e-02 -5.02292097e-01 -8.61292005e-01
-1.97701663e-01 -5.27862012e-02 8.46119076e-02 9.39654052e-01
-9.40783262e-01 1.31238544e+00 6.35104227e+00 6.38219714e-01
-6.42959118e-01 1.51172906e-01 7.68324733e-01 1.09934986e-01
-5.33131957e-01 -2.35622615e-01 -9.06312108e-01 3.11542839e-01
1.03608429e+00 -3.84282708e-01 4.43433598e-02 7.58123875e-01
2.42023870e-01 -2.49211743e-01 -1.04170024e+00 4.18975562e-01
2.28686258e-01 -1.35939503e+00 4.58576381e-01 -2.43624121e-01
8.04689229e-01 -3.17503363e-01 -1.78103283e-01 7.43108094e-01
7.34819174e-01 -1.00282192e+00 1.43992737e-01 2.77130604e-02
3.24603647e-01 -1.04657996e+00 9.84538138e-01 5.89387774e-01
-6.38506830e-01 8.06110054e-02 -2.15480641e-01 -2.61113793e-01
4.56051767e-01 -8.69666561e-02 -2.04562712e+00 7.51593649e-01
6.63837269e-02 3.22194427e-01 -2.93554127e-01 1.99542180e-01
7.58157894e-02 6.24700427e-01 -5.38772866e-02 -6.53856918e-02
6.42371953e-01 6.20842651e-02 6.61414027e-01 1.38410413e+00
-2.66267002e-01 5.73050797e-01 3.68662924e-01 3.97407889e-01
-2.06821129e-01 2.39217162e-01 -8.11100543e-01 -8.44942853e-02
7.45464325e-01 8.11267674e-01 -2.75636524e-01 -7.51858890e-01
-3.13470781e-01 9.07704592e-01 1.91426322e-01 1.92439690e-01
-3.89930099e-01 -3.24399889e-01 4.43057388e-01 -4.54555064e-01
-9.48314294e-02 -1.01585671e-01 -2.36327395e-01 -6.90763772e-01
-2.26748243e-01 -1.29966640e+00 2.70789295e-01 -5.42206526e-01
-1.15339839e+00 9.61863935e-01 1.24527022e-01 -9.04604077e-01
-1.04735410e+00 -2.16659591e-01 -3.01166683e-01 8.94281864e-01
-1.29747033e+00 -1.28883553e+00 -5.85041568e-02 3.10085803e-01
1.66817796e+00 -8.02215457e-01 1.12343466e+00 8.15245956e-02
-4.14840519e-01 6.45222902e-01 -4.11966182e-02 1.64851680e-01
6.32977545e-01 -1.39387238e+00 5.38200557e-01 4.06197041e-01
-1.31239727e-01 5.11453152e-01 9.24038470e-01 -8.03900421e-01
-1.12549460e+00 -9.12878871e-01 8.10276568e-01 -5.33100784e-01
5.03038168e-01 -1.00842319e-01 -1.20956421e+00 6.85350895e-01
6.70199692e-01 -8.72001350e-01 1.02341175e+00 1.41126633e-01
-5.60956225e-02 6.82147324e-01 -1.39610350e+00 3.73574644e-01
8.27119172e-01 -2.83156455e-01 -1.25464690e+00 3.99745941e-01
1.19959533e+00 -7.32206345e-01 -9.14067507e-01 4.49445724e-01
2.65676439e-01 -6.23998761e-01 7.81911016e-01 -1.16875160e+00
4.94995117e-01 1.83653180e-02 -4.78329927e-01 -1.49376440e+00
3.53639960e-01 -3.92163128e-01 2.39999130e-01 1.50673842e+00
6.10273182e-01 -4.59190816e-01 7.41537511e-01 7.96203732e-01
-6.02790952e-01 -5.34920514e-01 -9.03838456e-01 -6.20888710e-01
1.29722938e-01 -1.63433045e-01 7.43726134e-01 1.17773819e+00
2.25144982e-01 9.80996430e-01 -5.06482184e-01 -6.11778945e-02
7.13463202e-02 -2.53895819e-01 1.05451775e+00 -6.89816058e-01
-1.96645916e-01 1.16170660e-01 -9.24528986e-02 -1.37571824e+00
4.81533855e-01 -7.24839330e-01 2.25980669e-01 -1.37146568e+00
-9.36685409e-03 -6.92816854e-01 2.79800475e-01 4.60492134e-01
-3.55379641e-01 -1.53349832e-01 8.77769589e-02 -1.53848037e-01
-5.95771074e-01 8.93503308e-01 1.12366247e+00 -1.08704589e-01
-4.89999622e-01 -1.54386729e-01 -6.16442323e-01 5.63983381e-01
9.28601205e-01 -1.64832577e-01 -9.47673619e-01 -4.41381633e-01
-5.04830480e-01 2.50246376e-01 -2.33856887e-01 -4.11587954e-01
-3.58939730e-02 -3.60046439e-02 -9.30235684e-02 -6.99842215e-01
6.82613075e-01 -4.32164609e-01 -3.65418255e-01 2.72954345e-01
-9.85566080e-01 7.82318711e-02 5.07724226e-01 4.47862178e-01
-5.58026373e-01 -5.53801417e-01 5.55799186e-01 -3.06576878e-01
-1.07592070e+00 -4.50859070e-01 -5.46267211e-01 3.68003219e-01
8.60171556e-01 -2.22483680e-01 -5.41780949e-01 -6.65421605e-01
-7.64591753e-01 5.44502258e-01 9.30973515e-02 8.25176239e-01
4.04018551e-01 -1.34686863e+00 -9.72028077e-01 7.87508190e-02
2.53331929e-01 1.01818174e-01 3.88832867e-01 2.19447449e-01
-2.40150332e-01 7.94447422e-01 -3.16799462e-01 -6.22361660e-01
-1.60179651e+00 3.30953486e-02 1.14409015e-01 -3.34581435e-01
-3.61235082e-01 9.47693050e-01 2.49736786e-01 -9.59279180e-01
2.18044788e-01 -2.70559251e-01 -6.24058366e-01 1.63114145e-01
4.30828184e-01 3.74432713e-01 -1.21032223e-02 -7.81773329e-01
2.15795543e-02 -1.05802625e-01 -5.61981320e-01 -5.48111677e-01
1.10760617e+00 -5.93385696e-01 2.95681626e-01 7.29313195e-01
1.06523383e+00 -2.48735979e-01 -9.76088464e-01 -4.66779441e-01
3.57095242e-01 -3.41318667e-01 -3.38947386e-01 -1.23659718e+00
-4.57099199e-01 5.96814036e-01 3.09531361e-01 4.22571182e-01
8.58783841e-01 -8.77252519e-02 8.34299564e-01 8.73333514e-01
2.44577110e-01 -1.24732137e+00 3.71588081e-01 1.23219025e+00
1.14306545e+00 -1.59747601e+00 -3.49937320e-01 -4.31209654e-01
-1.49348533e+00 8.39204669e-01 9.54590499e-01 3.00420135e-01
-1.18489787e-01 -8.77339207e-03 3.21396679e-01 -1.10306905e-03
-1.12783372e+00 -1.69976920e-01 5.20859547e-02 7.27742970e-01
7.91293263e-01 -5.62050678e-02 -4.30497855e-01 6.25771821e-01
-5.44279456e-01 -2.53720701e-01 7.83915579e-01 1.01205921e+00
-7.35644758e-01 -1.50887656e+00 -1.24802478e-01 3.89996290e-01
-2.20291451e-01 -2.40691930e-01 -5.83328366e-01 9.14148331e-01
-1.56285867e-01 1.31042266e+00 1.41024798e-01 -3.15792739e-01
3.84329796e-01 6.62786365e-01 1.92197025e-01 -1.11825240e+00
-7.52937853e-01 -3.90044041e-02 1.19354582e+00 1.98633876e-02
-6.16074979e-01 -4.53797013e-01 -1.63590825e+00 7.32798800e-02
-4.14224714e-01 6.04414105e-01 6.68634534e-01 1.12960577e+00
2.64843404e-01 7.58518338e-01 6.85153484e-01 -3.66004080e-01
-4.65987980e-01 -1.61482513e+00 -1.93674177e-01 6.14473760e-01
2.97149599e-01 -6.93216622e-01 1.38039142e-01 2.82136858e-01] | [12.741643905639648, 7.987917423248291] |
465dd8c5-5233-4030-9976-105da2018a67 | grammarshap-an-efficient-model-agnostic-and | null | null | https://aclanthology.org/2022.lnls-1.2 | https://aclanthology.org/2022.lnls-1.2.pdf | GrammarSHAP: An Efficient Model-Agnostic and Structure-Aware NLP Explainer | Interpreting NLP models is fundamental for their development as it can shed light on hidden properties and unexpected behaviors. However, while transformer architectures exploit contextual information to enhance their predictive capabilities, most of the available methods to explain such predictions only provide importance scores at the word level. This work addresses the lack of feature attribution approaches that also take into account the sentence structure. We extend the SHAP framework by proposing GrammarSHAP—a model-agnostic explainer leveraging the sentence’s constituency parsing to generate hierarchical importance scores. | ['Georg Groh', 'Fabio Raffagnato', 'Luca Mülln', 'Defne Demirtürk', 'Edoardo Mosca'] | null | null | null | null | lnls-acl-2022-5 | ['constituency-parsing'] | ['natural-language-processing'] | [ 3.26367348e-01 8.19006503e-01 -6.70925081e-01 -6.98544800e-01
-5.42070210e-01 -5.27122617e-01 7.08475053e-01 5.12804747e-01
2.97720850e-01 7.46409595e-01 7.28847563e-01 -6.47489786e-01
-4.66774367e-02 -7.46780515e-01 -6.75419807e-01 4.22154292e-02
2.07349196e-01 4.03922111e-01 2.95137316e-01 -3.40375721e-01
3.25554788e-01 2.00542107e-01 -1.55866516e+00 6.45178080e-01
7.85694420e-01 6.57691479e-01 8.99017379e-02 5.67770123e-01
-4.33756381e-01 1.20725083e+00 -3.62296045e-01 -6.63807988e-01
-2.76401132e-01 -2.86158532e-01 -8.85752678e-01 -1.66898087e-01
9.50684473e-02 -3.51370708e-03 -7.65433758e-02 8.11353445e-01
-4.05670971e-01 -2.90184975e-01 5.76054096e-01 -1.32803476e+00
-9.19048011e-01 1.43240929e+00 5.92905432e-02 4.05856013e-01
4.13518220e-01 2.10661367e-02 1.96067548e+00 -6.97298288e-01
6.27162993e-01 1.12395132e+00 6.74870610e-01 7.64417410e-01
-1.43201005e+00 -3.43293041e-01 4.82859999e-01 3.70896101e-01
-6.63885176e-01 -3.45381111e-01 1.05666256e+00 -3.31622928e-01
1.65969038e+00 4.88672495e-01 5.02810955e-01 1.40367591e+00
4.47189659e-01 8.12232673e-01 9.84581053e-01 -6.01357818e-01
7.91817084e-02 1.77467346e-01 7.90213227e-01 5.33836782e-01
3.65630805e-01 1.69149280e-01 -8.17723572e-01 -1.08580627e-01
2.24708095e-01 -2.67715305e-01 8.37097019e-02 -5.13464026e-03
-7.93208718e-01 9.98943329e-01 1.65829569e-01 4.62961257e-01
-4.10354614e-01 2.67563581e-01 2.24511892e-01 2.44331643e-01
6.20802343e-01 9.88545716e-01 -1.11195290e+00 -2.52723992e-01
-7.48727560e-01 2.09142789e-01 1.02196896e+00 8.72163117e-01
7.90973425e-01 -1.82172190e-02 -2.01695696e-01 1.48036003e-01
4.96657223e-01 1.27560288e-01 4.17874604e-01 -6.63184166e-01
4.05077964e-01 1.06773627e+00 -1.66456595e-01 -1.02668166e+00
-5.16697705e-01 -7.09411979e-01 -1.09345883e-01 -3.33535671e-01
-1.03636444e-01 2.12624341e-01 -7.08414257e-01 1.89327943e+00
1.17344216e-01 -3.87816206e-02 1.33400783e-01 5.44616759e-01
7.99510658e-01 3.93461287e-01 6.05535388e-01 -1.00979514e-01
1.28103971e+00 -9.47336137e-01 -6.43125296e-01 -6.36857152e-01
7.61537373e-01 -4.22003269e-01 1.22584045e+00 1.16549045e-01
-9.19444978e-01 -3.15088928e-01 -1.03331125e+00 -2.76911318e-01
-2.74915993e-01 -2.37298340e-01 1.06537652e+00 5.87829113e-01
-9.26991522e-01 6.18837416e-01 -8.59062612e-01 -2.47488633e-01
1.74701691e-01 2.23037168e-01 -3.20076644e-01 4.60246295e-01
-1.42134607e+00 1.30897748e+00 6.37269914e-01 -3.36949438e-01
-5.36200821e-01 -8.52297366e-01 -9.76407170e-01 3.92038167e-01
3.32273513e-01 -8.29926968e-01 1.18826234e+00 -9.03544903e-01
-1.30818295e+00 4.13343370e-01 -5.37748814e-01 -7.17693388e-01
-5.06285094e-02 -1.68861896e-01 -4.05158371e-01 -1.62884504e-01
9.31994095e-02 3.83183688e-01 5.58485091e-01 -9.89687920e-01
-5.33617496e-01 -2.09565610e-01 3.70752126e-01 -2.26003572e-01
-4.27969605e-01 -1.24838902e-02 8.01826715e-02 -5.76436579e-01
9.40645039e-02 -6.31453216e-01 -3.92421782e-01 -6.01202607e-01
-8.66473138e-01 -1.70088217e-01 4.24038410e-01 -6.58810437e-01
1.52213347e+00 -1.68419278e+00 1.95397720e-01 4.05905172e-02
3.48542154e-01 -2.32788160e-01 -1.64374039e-01 7.88573086e-01
-1.04496650e-01 5.66822350e-01 -1.40922189e-01 -4.38563883e-01
3.54077667e-01 4.46051151e-01 -8.36386144e-01 -2.79727906e-01
7.05466270e-01 1.14949286e+00 -6.61708951e-01 -3.55221987e-01
1.24254428e-01 4.04564112e-01 -1.00570166e+00 5.59595749e-02
-6.30120695e-01 1.94433719e-01 -7.39265859e-01 5.47484040e-01
2.19436467e-01 -4.64532673e-01 4.98054385e-01 -5.72967976e-02
-1.92796066e-01 1.07087958e+00 -4.97497380e-01 1.12514770e+00
-5.57574272e-01 5.11984348e-01 -4.56797004e-01 -7.47730732e-01
8.06386113e-01 1.71824872e-01 8.56341794e-02 -1.81371883e-01
-2.05235794e-01 1.45863265e-01 1.68343052e-01 -3.03198338e-01
7.33154655e-01 -5.88686109e-01 -3.71519834e-01 4.99636322e-01
9.05257016e-02 -5.32067232e-02 -1.76980458e-02 2.65213341e-01
1.48505330e+00 3.29399437e-01 8.96787763e-01 -2.15846643e-01
5.00809371e-01 3.03052783e-01 7.10101008e-01 7.54631698e-01
7.48836249e-02 4.53751266e-01 6.58509970e-01 -6.26129746e-01
-1.20275414e+00 -8.30432832e-01 -2.21149087e-01 1.25012410e+00
-1.05003327e-01 -9.26493943e-01 -6.62145793e-01 -9.98836577e-01
-1.64814405e-02 1.45658791e+00 -9.21251357e-01 -2.48385504e-01
-3.89253646e-01 -5.26965737e-01 3.95599365e-01 8.74486268e-01
-2.16257095e-01 -1.32798934e+00 -4.91988331e-01 4.51311290e-01
-3.03749025e-01 -1.17505872e+00 8.02673548e-02 5.30621469e-01
-8.55323493e-01 -1.09084070e+00 5.92517078e-01 -1.70364812e-01
4.39940184e-01 -3.53116006e-01 1.52978051e+00 3.17248076e-01
1.13643184e-01 2.31181473e-01 -6.59610391e-01 -6.06608987e-01
-9.22174871e-01 4.90168095e-01 -2.19818801e-02 -5.25660813e-01
7.80151248e-01 -7.98384011e-01 1.27667524e-02 -3.60617451e-02
-6.90195203e-01 4.29383487e-01 6.01506889e-01 7.05548108e-01
4.19672698e-01 -1.96919650e-01 5.06625354e-01 -1.32369745e+00
6.11359596e-01 -5.65548480e-01 -1.68050691e-01 4.09747720e-01
-9.45949197e-01 6.56028628e-01 7.62065649e-01 -1.34430856e-01
-9.85248089e-01 -2.16702998e-01 -2.87224859e-01 1.05501011e-01
-3.04835260e-01 7.61003315e-01 -2.41993666e-01 3.86354595e-01
4.18814272e-01 2.87351996e-01 -4.63105679e-01 -4.84842628e-01
3.10915619e-01 1.25124380e-01 7.38524646e-02 -6.07545078e-01
8.96451116e-01 5.89606985e-02 9.34673008e-03 -3.36368620e-01
-1.55043805e+00 -1.36729613e-01 -6.88293874e-01 1.40850306e-01
6.79976344e-01 -3.29206139e-01 -5.52024007e-01 -5.52758574e-01
-1.47164929e+00 -9.95493121e-03 -4.39659536e-01 2.47908100e-01
-6.32397115e-01 8.58895108e-02 -5.57422340e-01 -8.61561537e-01
-4.37114954e-01 -9.59992766e-01 8.62396002e-01 -1.96805988e-02
-9.19901848e-01 -1.17527163e+00 7.78320953e-02 3.61189306e-01
5.10701656e-01 1.00748293e-01 1.51210666e+00 -1.21771562e+00
-4.72962230e-01 -1.42007574e-01 6.40897229e-02 1.43377990e-01
-8.16115811e-02 -4.48975638e-02 -1.12303042e+00 4.78257805e-01
-7.77252426e-04 -3.90501739e-03 8.01909924e-01 1.34439275e-01
8.99176836e-01 -7.35301971e-01 -3.75788987e-01 1.78664222e-01
1.17479348e+00 -3.84023786e-01 3.62705469e-01 5.64002693e-01
6.02790177e-01 8.71417165e-01 2.67094582e-01 3.76089573e-01
7.13874519e-01 5.36256492e-01 7.89447427e-01 2.57374525e-01
-1.51324585e-01 -8.54035556e-01 3.46147239e-01 8.81371558e-01
-5.11734448e-02 -2.46291667e-01 -1.01501632e+00 4.21501964e-01
-1.82622099e+00 -1.02946806e+00 -4.51793373e-01 1.53984225e+00
8.35791945e-01 4.16489661e-01 -4.66124117e-01 1.74211219e-01
4.48454767e-01 3.25505227e-01 -4.35250908e-01 -6.65772319e-01
-2.12418675e-01 1.41432166e-01 1.79544240e-01 6.87999606e-01
-8.68127406e-01 1.15375710e+00 7.29861498e+00 3.01220655e-01
-5.53739786e-01 1.17802784e-01 4.16643351e-01 2.61125863e-01
-9.74838555e-01 5.51275790e-01 -1.01827252e+00 2.14834988e-01
1.28305948e+00 -2.23568678e-01 1.52388662e-01 1.12766016e+00
1.21387683e-01 2.15599090e-01 -1.50273502e+00 1.64944738e-01
-6.21763989e-02 -1.44733357e+00 5.14723420e-01 1.81647986e-01
3.75968218e-01 -1.02086961e-01 -3.37076038e-02 4.75170791e-01
3.39419305e-01 -9.96441364e-01 1.02620590e+00 7.25154400e-01
8.54124203e-02 -5.97818971e-01 7.32972324e-01 6.06691301e-01
-1.01545238e+00 -1.85988113e-01 -3.39883238e-01 -5.68442225e-01
4.04125363e-01 6.00995898e-01 -1.05229664e+00 2.50821471e-01
2.38539889e-01 6.54706061e-01 -7.38904834e-01 2.18118027e-01
-8.71668398e-01 1.10151541e+00 6.66806847e-02 -1.68351948e-01
1.19301721e-01 2.67993003e-01 8.66332412e-01 1.18918622e+00
2.44449154e-01 -7.41219753e-03 8.87184888e-02 1.07248485e+00
3.91665511e-02 6.78004399e-02 -5.27620196e-01 -2.21568301e-01
4.98279244e-01 9.97867882e-01 -4.23423916e-01 -2.91225225e-01
-4.63693082e-01 7.46187449e-01 5.19935489e-01 -1.59886882e-01
-8.74332488e-01 2.44141042e-01 7.94734478e-01 4.52743500e-01
2.40271762e-01 -7.59317130e-02 -7.21823931e-01 -1.26272571e+00
1.07207343e-01 -6.62500322e-01 2.77968377e-01 -6.95636690e-01
-1.38512647e+00 7.33382523e-01 -9.11982432e-02 -7.33156502e-01
-5.99954128e-01 -6.88422680e-01 -6.45399749e-01 6.46814227e-01
-1.67644083e+00 -1.72857726e+00 2.86405742e-01 1.45737873e-02
8.07313323e-01 4.71932255e-03 1.19418466e+00 -5.17613173e-01
-3.23739886e-01 3.92519832e-01 -5.85595310e-01 -1.54636383e-01
2.04009548e-01 -1.45256162e+00 7.71028101e-01 9.65379655e-01
4.71317142e-01 9.99709785e-01 1.34887207e+00 -7.34567404e-01
-1.31180501e+00 -9.00695384e-01 1.73284864e+00 -1.16569209e+00
1.08270347e+00 -3.25364381e-01 -1.02600181e+00 9.83755708e-01
1.16308138e-01 -1.03987403e-01 1.01504600e+00 8.82980466e-01
-8.06432962e-01 4.12382364e-01 -8.92856359e-01 4.77644771e-01
1.14922857e+00 -6.06804907e-01 -1.14826858e+00 3.67790870e-02
1.17814493e+00 -1.47845015e-01 -8.25504780e-01 4.18489337e-01
4.60922122e-01 -1.03300893e+00 5.93568325e-01 -1.21848047e+00
1.04686487e+00 -5.44347540e-02 -4.65552211e-01 -1.17354167e+00
-4.72788036e-01 -3.19254756e-01 -5.95137596e-01 1.26973224e+00
1.21286643e+00 -7.66665161e-01 6.97190821e-01 1.20025289e+00
-4.37154949e-01 -7.06800222e-01 -7.18959033e-01 -6.07257605e-01
3.36754844e-02 -1.10561812e+00 9.64448810e-01 8.32728684e-01
4.85704720e-01 4.98047918e-01 -3.53202432e-01 2.94648498e-01
4.27230418e-01 2.48827592e-01 4.41970885e-01 -1.48616111e+00
-6.33089364e-01 -3.40849340e-01 -3.72308403e-01 -4.10187632e-01
7.22560942e-01 -1.07363987e+00 -2.17107683e-02 -1.53764832e+00
5.21142900e-01 -2.67219782e-01 -3.80843252e-01 8.82408440e-01
-2.46729866e-01 2.32052300e-02 2.96301246e-01 2.45260045e-01
-5.48504055e-01 3.94929051e-01 6.92027390e-01 7.73040503e-02
1.87091425e-01 -1.18247107e-01 -1.14718294e+00 1.04761779e+00
9.41990376e-01 -7.81122983e-01 -2.97501117e-01 -4.76224214e-01
7.37641752e-01 -1.98201239e-01 3.76478434e-01 -8.15242708e-01
-1.53279863e-02 -4.39839691e-01 7.09604993e-02 -2.99429089e-01
1.72571898e-01 -7.50851631e-01 1.59760594e-01 3.14933121e-01
-9.50864851e-01 1.64754838e-01 1.57435417e-01 6.69410110e-01
-3.16108644e-01 -2.43568823e-01 1.77438051e-01 -8.17698985e-02
-6.86938286e-01 9.68118533e-02 -3.98332894e-01 -1.54171109e-01
7.25022495e-01 -5.17441146e-02 -4.89070177e-01 -2.06655145e-01
-7.23099887e-01 -1.97067842e-01 5.80238998e-01 6.47285163e-01
5.16225517e-01 -1.12984896e+00 -5.27614594e-01 1.83561608e-01
4.33777690e-01 -6.98066294e-01 -3.01707834e-02 6.14311993e-01
6.33314699e-02 7.10415840e-01 1.34092391e-01 -1.28885001e-01
-1.01122928e+00 5.65239787e-01 3.90679762e-02 -6.48146510e-01
-6.66359365e-01 5.61501026e-01 2.34759867e-01 -5.71188927e-01
-3.88371587e-01 -4.41624969e-01 -4.22987401e-01 -2.13743314e-01
6.55272663e-01 -9.48772952e-02 1.19967768e-02 -6.82260394e-01
-5.79321444e-01 1.62233084e-01 -1.18243873e-01 -1.76968887e-01
1.62221777e+00 -1.69342980e-01 -1.92883477e-01 5.57301939e-01
6.67578220e-01 2.68393666e-01 -1.01189435e+00 -1.34421229e-01
7.20404863e-01 -1.81108981e-01 1.16351269e-01 -1.08666909e+00
-4.19392139e-01 8.99812818e-01 -3.40483546e-01 5.14863491e-01
7.60122359e-01 4.48128313e-01 6.80088520e-01 3.17492068e-01
5.68754077e-01 -8.82791817e-01 -1.17653809e-01 8.47982645e-01
8.88594568e-01 -1.12690139e+00 -2.10177541e-01 -6.49997056e-01
-8.31492901e-01 1.27204669e+00 5.53011715e-01 1.94419175e-01
4.44821984e-01 5.56740642e-01 -1.10566445e-01 -1.62395939e-01
-1.44626558e+00 -2.04953045e-01 4.04045254e-01 5.61000586e-01
9.56200242e-01 2.85145551e-01 -4.47752237e-01 1.36017478e+00
-5.50819635e-01 -3.13950419e-01 6.11827075e-01 7.40584493e-01
-6.25457644e-01 -1.46300995e+00 8.80424231e-02 5.05269408e-01
-6.21337175e-01 -7.12822556e-01 -6.04897678e-01 7.21693933e-01
-5.11742979e-02 1.04340172e+00 -1.55768454e-01 -6.46458030e-01
2.15713084e-01 4.51610327e-01 2.00385764e-01 -1.12395990e+00
-7.96194732e-01 -4.73714590e-01 5.86359859e-01 -6.07854009e-01
-3.51833820e-01 -6.21591032e-01 -1.45504880e+00 -2.43690342e-01
-2.17918828e-01 2.49146700e-01 7.42900789e-01 1.29138052e+00
3.48550498e-01 5.26533544e-01 3.71957779e-01 -3.98775667e-01
-4.84392017e-01 -9.83336985e-01 -3.22982401e-01 2.61489838e-01
2.34422758e-01 -4.96093601e-01 -4.42764282e-01 3.10995616e-02] | [9.557624816894531, 6.961639881134033] |
c5294764-eea7-4df2-add2-9d6960354a95 | query-structure-modeling-for-inductive | 2305.13585 | null | https://arxiv.org/abs/2305.13585v1 | https://arxiv.org/pdf/2305.13585v1.pdf | Query Structure Modeling for Inductive Logical Reasoning Over Knowledge Graphs | Logical reasoning over incomplete knowledge graphs to answer complex logical queries is a challenging task. With the emergence of new entities and relations in constantly evolving KGs, inductive logical reasoning over KGs has become a crucial problem. However, previous PLMs-based methods struggle to model the logical structures of complex queries, which limits their ability to generalize within the same structure. In this paper, we propose a structure-modeled textual encoding framework for inductive logical reasoning over KGs. It encodes linearized query structures and entities using pre-trained language models to find answers. For structure modeling of complex queries, we design stepwise instructions that implicitly prompt PLMs on the execution order of geometric operations in each query. We further separately model different geometric operations (i.e., projection, intersection, and union) on the representation space using a pre-trained encoder with additional attention and maxout layers to enhance structured modeling. We conduct experiments on two inductive logical reasoning datasets and three transductive datasets. The results demonstrate the effectiveness of our method on logical reasoning over KGs in both inductive and transductive settings. | ['Xuanjing Huang', 'Qi Zhang', 'Haijun Shan', 'Zhihao Fan', 'Meng Han', 'Zhongyu Wei', 'Siyuan Wang'] | 2023-05-23 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 1.10946439e-01 6.51011646e-01 -3.41506451e-01 -6.54815793e-01
-4.84371036e-01 -7.14707494e-01 3.11651260e-01 5.59704661e-01
2.28496715e-02 5.23933828e-01 3.35943878e-01 -7.12827981e-01
-1.59111053e-01 -1.54755962e+00 -1.35641158e+00 2.37198353e-01
-1.39995456e-01 6.65759683e-01 3.29677463e-01 -3.81774127e-01
2.27966215e-02 2.70207644e-01 -1.46195602e+00 7.55356312e-01
1.13626087e+00 1.11315703e+00 -8.02937821e-02 6.51868403e-01
-6.60471559e-01 1.98960149e+00 -4.16045308e-01 -6.98464155e-01
-2.96341572e-02 -2.77387332e-02 -1.36653280e+00 -4.20151949e-01
5.85198164e-01 -5.88854015e-01 -5.63123167e-01 6.91892028e-01
3.78000960e-02 1.37172699e-01 5.60239613e-01 -1.10875773e+00
-1.15472770e+00 1.25450957e+00 1.98994093e-02 -2.13065669e-01
8.06891143e-01 -3.33619155e-02 1.54137468e+00 -7.22317576e-01
6.87924743e-01 1.55893958e+00 5.69224596e-01 1.90918595e-01
-1.19993246e+00 -3.05946589e-01 4.29239452e-01 5.57284176e-01
-1.44092572e+00 -2.49984533e-01 6.56591177e-01 -4.34024520e-02
1.34968925e+00 1.38849840e-01 5.70936680e-01 4.72462237e-01
-1.04477599e-01 1.33274281e+00 3.99466902e-01 -4.00909781e-01
6.91487268e-02 4.29484695e-02 2.76585668e-01 1.25881863e+00
2.32429922e-01 -3.10223192e-01 -5.91250718e-01 -1.35061041e-01
6.40282989e-01 -1.90250412e-01 1.52179047e-01 -5.40595353e-01
-9.54972327e-01 9.49886203e-01 9.07163262e-01 -1.61539391e-01
-1.30838767e-01 4.80949521e-01 4.13467735e-01 4.34617013e-01
1.16382092e-01 7.60451436e-01 -6.05976582e-01 2.61619121e-01
-4.98129159e-01 3.56341004e-01 1.24132264e+00 1.47753620e+00
8.47798109e-01 -1.55850664e-01 -3.05486172e-01 5.92276096e-01
3.44560981e-01 6.24236524e-01 -3.59241605e-01 -9.01111662e-01
7.90913641e-01 1.37076151e+00 -2.11214289e-01 -1.08276570e+00
-4.05514061e-01 -3.14980537e-01 -5.55284500e-01 -6.00356281e-01
6.60472661e-02 3.35413106e-02 -6.27317250e-01 1.71891141e+00
2.89753526e-01 -8.11785161e-02 4.53109801e-01 4.68668699e-01
1.24796188e+00 6.96574152e-01 3.03033143e-01 1.71252191e-01
1.09591854e+00 -7.89906025e-01 -4.12210852e-01 -1.88156843e-01
1.20947576e+00 -1.70504376e-01 1.18090904e+00 1.06377803e-01
-1.22007930e+00 -3.68822813e-01 -9.24001098e-01 -8.01509857e-01
-6.10850513e-01 1.14547059e-01 1.07409871e+00 -6.71096817e-02
-9.38485146e-01 1.22295044e-01 -6.91925406e-01 -1.23849563e-01
3.19593310e-01 4.69935834e-01 7.05877040e-03 -5.00124693e-01
-1.64085674e+00 7.87498593e-01 8.16753209e-01 2.43358493e-01
-4.71738547e-01 -9.63000119e-01 -1.48563874e+00 3.73478889e-01
8.94051850e-01 -1.05781603e+00 1.34868503e+00 -2.17258096e-01
-1.00273967e+00 7.92434454e-01 -2.60337681e-01 -5.24651945e-01
1.07325055e-02 -2.84199476e-01 -2.52195269e-01 1.18320575e-02
7.19312131e-02 7.74320066e-01 2.62307912e-01 -1.25120044e+00
-4.13191974e-01 -3.24507505e-01 9.98899281e-01 2.84172237e-01
-7.06173256e-02 -4.13850427e-01 -5.69617152e-01 -3.43322545e-01
4.21025634e-01 -5.99944949e-01 4.65646908e-02 -2.27516145e-01
-8.38255167e-01 -6.93826377e-01 4.25260842e-01 -3.83510500e-01
1.39817512e+00 -1.96394122e+00 1.93409860e-01 1.58339202e-01
2.57698059e-01 -1.41911328e-01 7.30700269e-02 5.11438131e-01
1.82944506e-01 1.88746825e-01 6.89155515e-03 8.25648755e-02
4.77391541e-01 5.45503259e-01 -9.12519455e-01 -3.63790542e-01
6.28487289e-01 1.64280176e+00 -1.04757583e+00 -8.61463189e-01
2.18797587e-02 -1.39352530e-01 -9.67647731e-01 3.89343292e-01
-1.23447680e+00 -2.96219140e-01 -6.25608265e-01 9.72199738e-01
5.00062168e-01 -5.84982395e-01 3.30743045e-01 -5.99406719e-01
3.87488812e-01 6.18392169e-01 -9.44171965e-01 1.78655326e+00
-6.91534638e-01 1.80834964e-01 -4.17431206e-01 -9.25917268e-01
7.16380537e-01 -1.15227558e-01 3.65969837e-02 -8.56770158e-01
-2.56923825e-01 2.65283715e-02 -3.02266181e-01 -6.99387372e-01
8.19491029e-01 -1.30236179e-01 -6.50603712e-01 3.70251626e-01
7.16548711e-02 -5.93959332e-01 3.23801905e-01 8.25410843e-01
1.18307900e+00 2.65915871e-01 2.15505123e-01 1.42406836e-01
6.40262544e-01 2.61482984e-01 2.11830035e-01 8.80558372e-01
5.41823506e-01 -2.21712768e-01 7.51920938e-01 -6.06546700e-01
-4.20509785e-01 -1.34684706e+00 2.65510827e-01 1.45180333e+00
3.05878282e-01 -6.35774672e-01 -2.82825828e-01 -7.51471758e-01
3.50712657e-01 1.07915390e+00 -3.43292207e-01 -5.72162688e-01
-7.38657475e-01 -2.07850859e-01 8.78209710e-01 8.96374106e-01
6.29679263e-01 -1.17506754e+00 -3.87045979e-01 1.42070293e-01
-3.51984113e-01 -1.39249182e+00 6.51659165e-03 3.22951138e-01
-8.41625750e-01 -1.18972480e+00 3.28453511e-01 -9.86925840e-01
9.14345741e-01 -1.67270958e-01 1.63424385e+00 3.05777073e-01
-1.15162134e-01 5.38065434e-01 -2.63351291e-01 -4.20188755e-01
-2.97256440e-01 4.24639136e-01 -4.01649237e-01 -5.48425257e-01
4.55952138e-01 -2.26611868e-01 -1.41164228e-01 6.59424055e-04
-9.34202194e-01 4.16836590e-01 6.34918630e-01 5.42076051e-01
6.01807952e-01 3.12851459e-01 2.62633175e-01 -1.23354053e+00
6.17467403e-01 -4.85158265e-01 -5.94656587e-01 8.67749870e-01
-3.21953416e-01 7.48960018e-01 7.88849950e-01 -1.05249785e-01
-1.13316905e+00 -1.45426944e-01 2.32771575e-01 -2.96626091e-01
9.98854712e-02 1.13732195e+00 -1.98919028e-01 2.30755031e-01
5.14034808e-01 2.07290828e-01 -5.94797611e-01 5.31072132e-02
7.35635102e-01 9.31879058e-02 6.01998210e-01 -1.39334345e+00
8.07878017e-01 1.45350307e-01 2.26198435e-01 -4.28303510e-01
-1.34386337e+00 -1.47288516e-01 -4.76892442e-01 1.44593880e-01
4.75522429e-01 -8.89334321e-01 -1.19996917e+00 1.61509261e-01
-1.35129666e+00 -5.77788949e-01 -4.23425078e-01 2.59214789e-02
-6.65188432e-01 6.75752461e-02 -7.62670994e-01 -7.44694769e-01
-2.86807209e-01 -8.83063257e-01 1.13756394e+00 -2.28895321e-02
-2.91487306e-01 -1.15228057e+00 -2.81493962e-01 4.47705954e-01
1.65624559e-01 2.00950518e-01 1.74753809e+00 -4.57621068e-01
-1.26995647e+00 -1.41526505e-01 -5.32078326e-01 5.01884520e-02
-4.45590653e-02 -2.08297372e-01 -5.36369443e-01 2.31802851e-01
-3.77801985e-01 -1.02925289e+00 7.63255060e-01 -1.40006408e-01
1.38797057e+00 -5.59821248e-01 -3.09818387e-01 5.94922960e-01
1.29811442e+00 -3.19375433e-02 5.62655628e-01 -1.43405944e-01
9.25512493e-01 5.44614792e-01 5.49390674e-01 2.15152517e-01
1.15040338e+00 1.72411665e-01 4.99588042e-01 1.75494671e-01
2.16132492e-01 -9.58271861e-01 1.12979747e-01 5.94307065e-01
2.66287357e-01 -3.05826813e-01 -1.10385954e+00 4.22460675e-01
-1.79178715e+00 -8.40218663e-01 6.72891587e-02 1.73262715e+00
1.18778419e+00 2.69774169e-01 -4.93829817e-01 -4.65921164e-02
7.91559368e-02 9.61310938e-02 -9.13427591e-01 -4.57590997e-01
-1.69588447e-01 3.85842025e-01 3.61459643e-01 6.85888886e-01
-9.31726396e-01 1.44458508e+00 5.87887716e+00 4.41246986e-01
-8.30595493e-01 -6.25884891e-01 3.81391257e-01 -1.11482710e-01
-6.82080150e-01 7.32328296e-02 -8.52541745e-01 -1.82313085e-01
6.08245611e-01 -7.51916617e-02 6.98099613e-01 7.96302915e-01
-3.89249742e-01 -1.92676812e-01 -1.60564661e+00 7.33542144e-01
3.82159352e-02 -1.51633215e+00 6.14953697e-01 -3.80867928e-01
4.81599301e-01 -3.87136936e-01 -8.51645768e-02 1.07957721e+00
7.84693718e-01 -1.16547000e+00 7.93247998e-01 6.80277944e-01
8.24560821e-01 -6.40594840e-01 4.27600294e-01 2.75959939e-01
-1.49427593e+00 -1.10482730e-01 -2.95029461e-01 -7.33783767e-02
2.95337327e-02 2.30944097e-01 -1.03882515e+00 8.14674735e-01
3.28831911e-01 7.57409275e-01 -7.13369489e-01 2.81537980e-01
-8.09739292e-01 4.09228206e-01 -4.09943879e-01 -2.66535252e-01
1.79576218e-01 1.89494923e-01 -9.11289528e-02 9.79829967e-01
-1.07002169e-01 3.75961483e-01 2.22441643e-01 1.32791770e+00
-3.83722335e-01 -1.14112839e-01 -6.63371861e-01 -3.78646880e-01
4.75250095e-01 8.52845371e-01 -2.13536486e-01 -5.10317683e-01
-4.73942280e-01 6.15626037e-01 9.30408120e-01 4.89637673e-01
-9.15421426e-01 -3.46112162e-01 2.84770012e-01 2.26146609e-01
2.26294070e-01 -1.92573801e-01 -5.31596720e-01 -1.16300976e+00
2.86908299e-01 -7.19887674e-01 8.99173677e-01 -1.13388634e+00
-1.06069803e+00 2.00062841e-01 4.18841004e-01 -4.83131438e-01
-3.97670329e-01 -6.29167140e-01 -3.67372990e-01 7.16103315e-01
-1.39671338e+00 -1.42079413e+00 -3.21458548e-01 5.52865505e-01
1.02677673e-01 3.00689906e-01 8.20943832e-01 -1.43853938e-02
-2.23665997e-01 5.41202188e-01 -6.96572185e-01 4.14075464e-01
3.00964773e-01 -1.45191681e+00 3.81567806e-01 4.57967520e-01
4.08978015e-02 1.23603880e+00 2.54139692e-01 -6.93964958e-01
-2.04984474e+00 -1.15922511e+00 9.94706869e-01 -5.15181720e-01
7.67194450e-01 -5.01316965e-01 -9.81248915e-01 1.19897008e+00
-8.45085382e-02 -8.06596421e-04 7.64939249e-01 5.90362430e-01
-7.93555617e-01 -8.81558955e-02 -8.09097767e-01 8.06584179e-01
1.42489398e+00 -8.55583429e-01 -8.73290598e-01 3.14263552e-01
1.20889688e+00 -6.89447522e-01 -9.99288559e-01 7.60278881e-01
3.84735823e-01 -5.41125655e-01 1.10009515e+00 -1.02745020e+00
6.99176252e-01 -4.11475092e-01 -3.34134370e-01 -9.14405763e-01
-9.17237774e-02 -1.62440047e-01 -8.18725646e-01 8.09149563e-01
6.56924188e-01 -4.14869517e-01 1.02113783e+00 9.35572565e-01
-1.36465058e-01 -1.21513462e+00 -3.47282082e-01 -3.84011418e-01
-1.08026221e-01 -7.87108660e-01 7.35981226e-01 7.87463427e-01
4.65527833e-01 7.53015280e-01 2.45704025e-01 4.10696059e-01
3.82038802e-01 7.62879252e-01 9.00038540e-01 -1.14341545e+00
-2.62704819e-01 -1.85350507e-01 -2.50918478e-01 -1.44925094e+00
6.10147834e-01 -1.19417298e+00 -1.20034650e-01 -2.01121664e+00
-9.59971733e-03 -5.90894580e-01 1.04288407e-01 8.67778361e-01
-1.37973025e-01 -5.44528604e-01 -7.21497387e-02 -1.84967980e-01
-1.00284493e+00 5.54708302e-01 1.29919100e+00 -6.41184211e-01
-5.07501960e-02 -4.39161867e-01 -8.85907769e-01 5.12984037e-01
5.58392942e-01 3.33571471e-02 -9.65681255e-01 -8.08260798e-01
1.14523339e+00 1.64643154e-01 4.93018001e-01 -8.20063353e-01
6.49963081e-01 -3.03305477e-01 2.99962759e-01 -8.45880091e-01
2.71491170e-01 -6.64508283e-01 -1.59975752e-01 2.38329411e-01
-7.23439038e-01 -1.47597462e-01 4.71912324e-01 4.55249697e-01
-4.41280454e-01 -4.34736088e-02 1.77589938e-01 -2.46157840e-01
-9.60559249e-01 3.20504099e-01 1.39096022e-01 5.02938867e-01
6.72280610e-01 5.51414071e-03 -4.10229743e-01 -3.79121006e-01
-7.52581358e-01 8.52196336e-01 2.79940218e-01 2.84702182e-01
8.42859089e-01 -1.27095532e+00 -3.15378696e-01 1.10421628e-01
5.23568630e-01 9.40470874e-01 1.47132993e-01 5.09509325e-01
-7.71121025e-01 7.49217510e-01 3.92821759e-01 -2.30726078e-01
-8.90111148e-01 8.28440607e-01 3.47496390e-01 -5.65545380e-01
-1.64251164e-01 9.70642865e-01 3.60421747e-01 -8.68656695e-01
4.48804855e-01 -1.14416754e+00 9.07800943e-02 -3.23311865e-01
2.26738110e-01 2.39057541e-02 -7.15992600e-02 -1.27023056e-01
-3.21874887e-01 2.12769538e-01 -2.73064464e-01 4.63290602e-01
1.08620656e+00 2.42700815e-01 -5.09111345e-01 4.20770377e-01
9.78846192e-01 -1.50016472e-01 -5.43801963e-01 -7.22262383e-01
2.63805628e-01 1.15224449e-02 -2.90544629e-01 -9.15665984e-01
-5.79823375e-01 6.46949708e-01 -6.05498552e-01 1.66499630e-01
1.01840007e+00 4.86624569e-01 6.44187629e-01 1.26251674e+00
5.08096457e-01 -8.03905785e-01 1.76323786e-01 1.08195627e+00
1.04800951e+00 -1.07733083e+00 -6.88373670e-03 -9.35308754e-01
-3.72880518e-01 1.07103181e+00 9.29987907e-01 1.97115183e-01
3.80481958e-01 2.52758205e-01 -4.32564795e-01 -6.33167088e-01
-1.06849623e+00 -1.70184121e-01 2.99058408e-01 2.84689337e-01
4.92121965e-01 1.05740368e-01 3.96177053e-01 5.51568687e-01
-4.06754196e-01 2.37858519e-01 5.58483712e-02 1.20748997e+00
-3.23442310e-01 -7.37333953e-01 -7.89958239e-02 6.27128780e-01
1.41529709e-01 -4.67183471e-01 -7.32653916e-01 7.35574722e-01
3.38030048e-02 7.19656765e-01 -2.00928934e-02 -4.05303746e-01
7.21520424e-01 3.01795095e-01 9.02959466e-01 -9.34545755e-01
-2.95005977e-01 -9.75255251e-01 3.69982749e-01 -4.43509489e-01
-2.92423479e-02 -3.08965236e-01 -1.95837736e+00 -4.15957540e-01
-1.01860404e-01 1.06506683e-01 5.67554571e-02 9.44825888e-01
4.00754124e-01 6.73760414e-01 7.24034607e-02 1.24813221e-01
-6.11958861e-01 -7.39150465e-01 -3.92079502e-01 5.07098138e-01
1.31049752e-02 -5.29944718e-01 1.50563970e-01 1.06501035e-01] | [9.286738395690918, 7.5770111083984375] |
41f331ad-29db-418b-8e09-c03eaff76fa9 | not-all-voxels-are-equal-semantic-scene | 2112.12925 | null | https://arxiv.org/abs/2112.12925v2 | https://arxiv.org/pdf/2112.12925v2.pdf | Not All Voxels Are Equal: Semantic Scene Completion from the Point-Voxel Perspective | We revisit Semantic Scene Completion (SSC), a useful task to predict the semantic and occupancy representation of 3D scenes, in this paper. A number of methods for this task are always based on voxelized scene representations for keeping local scene structure. However, due to the existence of visible empty voxels, these methods always suffer from heavy computation redundancy when the network goes deeper, and thus limit the completion quality. To address this dilemma, we propose our novel point-voxel aggregation network for this task. Firstly, we transfer the voxelized scenes to point clouds by removing these visible empty voxels and adopt a deep point stream to capture semantic information from the scene efficiently. Meanwhile, a light-weight voxel stream containing only two 3D convolution layers preserves local structures of the voxelized scenes. Furthermore, we design an anisotropic voxel aggregation operator to fuse the structure details from the voxel stream into the point stream, and a semantic-aware propagation module to enhance the up-sampling process in the point stream by semantic labels. We demonstrate that our model surpasses state-of-the-arts on two benchmarks by a large margin, with only depth images as the input. | ['Jiaxiang Tang', 'Xiaokang Chen', 'Gang Zeng', 'Jingbo Wang'] | 2021-12-24 | null | null | null | null | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 2.84140468e-01 6.84184358e-02 2.38688394e-01 -5.44568539e-01
-4.74508405e-01 -1.59762591e-01 5.57477891e-01 2.19859660e-01
-3.39369923e-01 2.73930132e-01 1.41015828e-01 1.18875057e-02
7.91474506e-02 -1.13652849e+00 -9.69088614e-01 -4.91751194e-01
9.00969654e-02 4.51153815e-01 7.23384917e-01 1.13473525e-02
3.32145929e-01 8.26866448e-01 -1.61347008e+00 4.61133033e-01
8.68181825e-01 1.19133425e+00 4.73788589e-01 1.68775082e-01
-8.60757351e-01 5.68429410e-01 -2.23455742e-01 7.45452642e-02
3.71742487e-01 5.58680110e-02 -7.14573264e-01 2.69937038e-01
4.73147005e-01 -5.07310569e-01 -5.09576142e-01 1.02614605e+00
1.73482597e-01 3.49038213e-01 4.60114360e-01 -9.84117866e-01
-1.59823686e-01 1.85996771e-01 -7.28501618e-01 -1.79688260e-01
2.76618656e-02 3.20564657e-01 6.54503703e-01 -1.08448470e+00
5.91608226e-01 1.41178143e+00 5.11328757e-01 2.42047101e-01
-1.09308839e+00 -6.60548687e-01 5.95579624e-01 -9.72317234e-02
-1.35790777e+00 -2.17280492e-01 1.13410306e+00 -4.71826583e-01
9.49500442e-01 2.47038335e-01 7.92424262e-01 6.03612661e-01
-3.14173400e-02 6.05979383e-01 8.56748402e-01 6.23882115e-02
4.71537560e-01 -2.49036252e-01 2.11448282e-01 5.73513865e-01
-7.26409778e-02 -1.45767882e-01 -5.92768848e-01 -9.58412439e-02
1.18927944e+00 5.01402259e-01 -1.88411847e-01 -5.87259412e-01
-1.09482491e+00 6.14891052e-01 8.70146394e-01 -9.76200253e-02
-4.17255253e-01 5.36934018e-01 4.21911567e-01 -2.11854234e-01
8.14820111e-01 1.42026424e-01 -3.27933878e-01 1.94299534e-01
-9.54712391e-01 2.71096945e-01 3.12097281e-01 1.09099066e+00
1.05311990e+00 -2.35504776e-01 -2.34649524e-01 6.92595065e-01
3.69937241e-01 3.73359084e-01 -7.37193823e-02 -1.13136232e+00
6.31758571e-01 9.50612903e-01 -4.36051227e-02 -9.60470796e-01
-4.79060709e-01 -6.30233765e-01 -1.00026262e+00 2.93693393e-01
9.69928950e-02 5.68804622e-01 -1.23068655e+00 1.31267977e+00
7.45975196e-01 5.17922699e-01 -3.64854038e-01 1.17947245e+00
8.65193844e-01 8.91293883e-01 1.68816268e-01 7.88459256e-02
1.22136605e+00 -1.27374732e+00 -2.23284051e-01 -2.33209953e-01
3.96142483e-01 -5.02808571e-01 1.24439406e+00 2.91175783e-01
-1.13616574e+00 -5.42925656e-01 -9.54377353e-01 -6.52858377e-01
-5.09511195e-02 -3.89497519e-01 8.87855053e-01 1.87907875e-01
-7.64238954e-01 7.18425333e-01 -1.10978556e+00 2.40353476e-02
8.74689877e-01 8.80457535e-02 -9.33588371e-02 -4.24637884e-01
-6.74527705e-01 2.31464744e-01 1.49211094e-01 7.27645829e-02
-9.89756405e-01 -1.08748424e+00 -8.79401028e-01 3.27687353e-01
2.43113428e-01 -8.78347516e-01 1.03830433e+00 -5.70717692e-01
-1.18240249e+00 7.37144530e-01 -3.87405962e-01 -8.50954801e-02
5.67541242e-01 -9.47749764e-02 9.57927778e-02 1.69282287e-01
2.61641771e-01 7.33584166e-01 6.50380492e-01 -1.47944999e+00
-6.48267746e-01 -6.75963700e-01 1.83985069e-01 3.38613927e-01
7.26130977e-03 -4.31775808e-01 -8.54956865e-01 -4.98613805e-01
7.76369154e-01 -4.12585527e-01 -7.17783332e-01 5.19904494e-01
-4.74589825e-01 -8.81064683e-02 6.10557139e-01 -3.86680335e-01
8.24590206e-01 -2.55078077e+00 4.24668156e-02 3.25760335e-01
6.62591100e-01 -3.62642080e-01 -1.03731006e-01 1.95014868e-02
2.17165262e-01 -2.04506680e-01 -6.97783768e-01 -9.07723963e-01
9.15268958e-02 3.41251791e-01 -4.74607855e-01 5.93174994e-01
6.17958866e-02 7.05408454e-01 -9.68112350e-01 -4.43708360e-01
5.83795428e-01 7.54726708e-01 -8.84692073e-01 1.12912253e-01
-5.77020824e-01 6.29672587e-01 -8.41588497e-01 4.44265127e-01
1.24878848e+00 -3.22043270e-01 -4.92817402e-01 -2.85375118e-01
-3.59887272e-01 5.43530107e-01 -1.13850284e+00 2.40778017e+00
-5.34641564e-01 1.30597442e-01 3.20040703e-01 -7.38310695e-01
9.15990829e-01 -1.86751634e-01 7.70527124e-01 -7.92185545e-01
9.82946623e-03 1.90858126e-01 -6.10907376e-01 1.03330025e-02
5.60916662e-01 -4.88721803e-02 7.25227520e-02 1.59920260e-01
-4.45967138e-01 -5.89626253e-01 -3.79401773e-01 1.68683037e-01
9.98250484e-01 2.32049704e-01 -3.49867314e-01 -1.18075982e-01
3.73535514e-01 3.16050798e-01 5.73649704e-01 4.09637481e-01
1.38675302e-01 8.72115314e-01 4.22177255e-01 -5.86209834e-01
-1.10137403e+00 -1.28075802e+00 -2.02228159e-01 6.80221081e-01
6.99642539e-01 -4.68333900e-01 -8.16460133e-01 -3.02087456e-01
6.37848154e-02 6.81567550e-01 -2.96318829e-01 -7.82051906e-02
-5.58056951e-01 -4.43360627e-01 -3.77139449e-02 5.21873415e-01
6.33542538e-01 -7.90368259e-01 -7.26520717e-01 4.35145080e-01
-1.01492137e-01 -1.24094570e+00 -4.44954157e-01 2.48173669e-01
-1.20022345e+00 -8.63445640e-01 -5.04388869e-01 -6.41431749e-01
8.02867413e-01 7.06722438e-01 1.01785386e+00 1.35094479e-01
-1.42948866e-01 -1.00371502e-01 -1.60138860e-01 -7.46563375e-02
5.37416115e-02 8.29855800e-02 -3.37803364e-01 -1.70758680e-01
1.16639584e-01 -9.69849944e-01 -9.77960825e-01 1.47555977e-01
-1.12148333e+00 7.95329213e-01 2.48381868e-01 4.15684819e-01
1.16743696e+00 3.05463076e-01 -1.21807933e-01 -8.45337510e-01
-9.54685062e-02 -4.39837724e-01 -6.49665356e-01 -3.15216303e-01
-1.15095027e-01 1.77425276e-02 5.92563093e-01 5.31451814e-02
-9.98757362e-01 3.49250495e-01 -4.41479415e-01 -8.22239935e-01
-5.79004735e-02 1.29959568e-01 -3.60416293e-01 1.34867713e-01
2.75401682e-01 2.38715947e-01 -2.82437056e-01 -8.05371165e-01
5.06101549e-01 2.72733301e-01 5.79072177e-01 -6.05054140e-01
7.15957105e-01 1.13130486e+00 3.60275596e-01 -7.15251803e-01
-8.96125734e-01 -5.33393443e-01 -6.39145017e-01 -1.19907886e-01
9.09094632e-01 -1.07077384e+00 -4.88922417e-01 4.82825428e-01
-1.50676179e+00 -5.24295032e-01 -5.86680770e-01 3.11854839e-01
-5.24229825e-01 3.36207598e-01 -6.01906776e-01 -4.96959597e-01
-2.02744395e-01 -1.40095055e+00 1.55439436e+00 -8.45848322e-02
2.43852735e-01 -4.22139913e-01 -1.66053995e-01 2.18725547e-01
1.64650589e-01 3.56089026e-01 1.08612347e+00 1.48021668e-01
-1.29088914e+00 2.04019621e-01 -7.04919934e-01 8.63430426e-02
-8.12394693e-02 -2.65810996e-01 -1.30228436e+00 -2.04387028e-03
2.16376215e-01 1.68286517e-01 1.01278532e+00 4.70788807e-01
1.88599026e+00 3.80713865e-02 -4.18955386e-01 1.22437775e+00
1.45439172e+00 -3.00299942e-01 6.40950739e-01 1.94562703e-01
1.18699646e+00 7.01751232e-01 3.31871867e-01 5.71595669e-01
5.39523482e-01 5.78515232e-01 9.82718110e-01 -2.31168091e-01
-4.69332069e-01 -5.77635288e-01 -1.52553901e-01 7.99053311e-01
1.75272927e-01 -1.39533713e-01 -7.19154954e-01 3.22337657e-01
-1.76144803e+00 -4.51669484e-01 -5.80291212e-01 2.23844147e+00
4.61856782e-01 3.27613562e-01 -2.81805068e-01 4.77768108e-03
4.69312012e-01 3.98342222e-01 -7.39627898e-01 -2.07896046e-02
6.17251843e-02 3.52131814e-01 6.46562099e-01 5.76563776e-01
-9.03522968e-01 1.04058516e+00 4.93408918e+00 1.00500381e+00
-1.02837074e+00 2.85976022e-01 7.46395290e-01 -3.65887761e-01
-7.34801173e-01 4.71336357e-02 -5.85264146e-01 4.74290997e-01
3.74891430e-01 2.77152926e-01 3.71524662e-01 9.20670450e-01
2.82939792e-01 -8.51922780e-02 -1.18316603e+00 1.19786787e+00
-2.83627898e-01 -1.41887140e+00 1.73037305e-01 1.34680957e-01
7.44908273e-01 3.15930098e-01 -1.32832646e-01 5.60466833e-02
7.06524998e-02 -9.74177778e-01 9.62960660e-01 4.38330173e-01
8.72868598e-01 -7.91014731e-01 1.79665640e-01 3.66330326e-01
-1.49227870e+00 1.53718427e-01 -7.52087593e-01 -7.78618455e-02
3.84922862e-01 1.24204302e+00 -1.63466468e-01 5.06380737e-01
7.23279119e-01 6.62923157e-01 -3.17961663e-01 1.22575080e+00
3.83464503e-03 2.07062751e-01 -6.72083974e-01 3.39297473e-01
5.69901705e-01 -2.06548452e-01 4.28761274e-01 9.21338856e-01
3.18526298e-01 1.26411214e-01 3.16375434e-01 1.27796376e+00
-1.58757851e-01 1.22478027e-02 -4.03835207e-01 5.63222528e-01
4.06374186e-01 1.03867114e+00 -1.04734874e+00 -3.51314992e-01
-3.96941036e-01 1.25605106e+00 4.18950796e-01 3.05192411e-01
-7.40840852e-01 -1.56991750e-01 8.43333960e-01 6.02232456e-01
7.51932487e-02 -5.45957029e-01 -8.83310378e-01 -9.77089703e-01
2.36212879e-01 -1.17279723e-01 -6.54400513e-02 -8.56357813e-01
-1.08726799e+00 5.37067175e-01 -1.18393496e-01 -1.18394780e+00
4.91892487e-01 -3.45350832e-01 -5.11037230e-01 9.48763490e-01
-1.87585092e+00 -8.86985838e-01 -8.65618408e-01 5.55697322e-01
6.89292431e-01 5.51981926e-01 2.05625474e-01 5.37393808e-01
-2.50845760e-01 4.16737013e-02 -6.01501763e-02 -2.16126725e-01
2.37203360e-01 -9.18307722e-01 7.66748488e-01 4.99961525e-01
-2.04414174e-01 3.33165199e-01 2.42967978e-01 -7.46292055e-01
-1.40948772e+00 -1.44044483e+00 5.27377963e-01 -2.16448888e-01
2.27471381e-01 -6.56747758e-01 -1.04644239e+00 3.60562444e-01
-4.47305441e-01 2.54285365e-01 3.45860869e-02 -2.60967046e-01
-3.33385199e-01 -6.89177215e-02 -1.10020828e+00 6.67624295e-01
1.63136899e+00 -4.99332339e-01 -3.03974837e-01 4.48455751e-01
1.42282546e+00 -6.31675661e-01 -5.94778597e-01 3.99657369e-01
1.41126707e-01 -1.00807369e+00 1.33780646e+00 -1.52774945e-01
7.64231324e-01 -6.18947625e-01 -3.27867746e-01 -1.13391566e+00
-4.90272075e-01 -1.33573323e-01 -5.61155379e-02 8.67493212e-01
-5.11190332e-02 -3.35505754e-01 1.07621360e+00 6.24782741e-01
-8.74288201e-01 -8.61914217e-01 -1.09202313e+00 -5.66514730e-01
-3.33996862e-02 -9.58941758e-01 1.11441290e+00 6.88092589e-01
-5.79435408e-01 2.18173176e-01 1.45876050e-01 3.46993387e-01
9.11486626e-01 3.11427325e-01 8.59138787e-01 -1.21956408e+00
1.19743142e-02 -6.56399846e-01 -3.32354099e-01 -1.68146837e+00
-2.42034234e-02 -1.08001292e+00 1.40918344e-01 -1.91343927e+00
1.98134795e-01 -9.07478631e-01 -6.48334771e-02 1.40692145e-01
-4.64453585e-02 2.38907650e-01 9.91210341e-02 3.88806373e-01
-4.43035424e-01 1.01757300e+00 1.67769432e+00 -1.45558029e-01
-3.09597909e-01 -2.16637447e-01 -4.18041468e-01 8.39202583e-01
5.40645659e-01 -3.80472392e-01 -5.29828966e-01 -9.66870964e-01
8.06565732e-02 -1.17293678e-01 7.85973191e-01 -1.08122480e+00
2.60347247e-01 -2.78168976e-01 4.26507354e-01 -1.14287674e+00
7.76842296e-01 -1.04582298e+00 1.87223762e-01 2.37267643e-01
-6.76483288e-02 -3.25356156e-01 1.04775973e-01 6.26962125e-01
3.13900709e-02 1.42672854e-02 8.27099741e-01 -2.55002439e-01
-6.82439387e-01 9.41755235e-01 8.54828805e-02 -1.87268317e-01
8.49679708e-01 -3.42009336e-01 -7.35485554e-02 2.51507729e-01
-4.65039372e-01 4.19402361e-01 8.60812306e-01 2.28816584e-01
8.53985369e-01 -1.27057779e+00 -4.85454917e-01 4.28705394e-01
1.25392616e-01 1.22376859e+00 7.29277790e-01 6.45817280e-01
-1.00807083e+00 2.99097151e-01 1.11768149e-01 -9.26375628e-01
-6.28525317e-01 6.19703054e-01 2.89545029e-01 -7.81212002e-02
-1.36190820e+00 8.80063713e-01 1.07532668e+00 -4.81434643e-01
3.39528143e-01 -7.69830406e-01 2.85780221e-01 -4.14145142e-01
3.38505685e-01 2.48563781e-01 1.72275022e-01 -4.92181301e-01
-2.66086757e-01 6.52140498e-01 1.31329149e-01 2.36900877e-02
1.43888736e+00 -2.10528910e-01 -2.74653316e-01 3.46426934e-01
1.31023240e+00 -2.01506004e-01 -1.67953992e+00 -2.99933940e-01
-4.68345106e-01 -8.71180892e-01 5.00679791e-01 -1.62042901e-01
-1.27427101e+00 1.12818158e+00 3.44031632e-01 -1.68836519e-01
9.06774282e-01 1.82143062e-01 1.17585981e+00 1.11151330e-01
5.97400367e-01 -8.66004765e-01 -1.78669482e-01 4.69789356e-01
8.02573025e-01 -8.91958892e-01 1.38795823e-01 -1.09662592e+00
-1.87020287e-01 8.36192071e-01 7.31015742e-01 -2.59250551e-01
8.57380748e-01 9.33075845e-02 -4.80872452e-01 -6.01501048e-01
-4.15243626e-01 3.23670506e-02 -1.58090815e-02 5.91108978e-01
-4.51932959e-02 8.56364369e-02 1.07758969e-01 5.49701929e-01
-1.50021791e-01 -1.14390813e-01 2.04070419e-01 7.60079741e-01
-6.24117076e-01 -6.42399728e-01 -3.89042467e-01 5.07510245e-01
1.03982471e-01 -2.45664984e-01 4.54452448e-02 3.21845680e-01
1.82464615e-01 5.28801143e-01 5.64364851e-01 -2.01639906e-01
6.13471687e-01 -3.46855372e-01 4.11597818e-01 -8.76650155e-01
-3.19828719e-01 2.07446754e-01 -4.21236008e-01 -1.01710081e+00
-1.63266584e-01 -3.73973459e-01 -1.66794574e+00 -3.59066278e-01
1.36719227e-01 -9.57171097e-02 8.38975370e-01 6.61186397e-01
3.77552241e-01 8.69945288e-01 8.09724271e-01 -1.14434552e+00
-7.02944696e-02 -5.81720710e-01 -7.48172343e-01 2.98914403e-01
2.16238558e-01 -7.15612650e-01 -3.34327698e-01 -2.47220606e-01] | [8.475542068481445, -2.95529842376709] |
9bf22516-9838-4228-bbc1-457e57dda973 | a-practical-framework-for-roi-detection-in | 2103.01584 | null | https://arxiv.org/abs/2103.01584v1 | https://arxiv.org/pdf/2103.01584v1.pdf | A Practical Framework for ROI Detection in Medical Images -- a case study for hip detection in anteroposterior pelvic radiographs | Purpose Automated detection of region of interest (ROI) is a critical step for many medical image applications such as heart ROIs detection in perfusion MRI images, lung boundary detection in chest X-rays, and femoral head detection in pelvic radiographs. Thus, we proposed a practical framework of ROIs detection in medical images, with a case study for hip detection in anteroposterior (AP) pelvic radiographs. Materials and Methods: We conducted a retrospective study which analyzed hip joints seen on 7,399 AP pelvic radiographs from three diverse sources, including 4,290 high resolution radiographs from Chang Gung Memorial Hospital Osteoarthritis, 3,008 low to medium resolution radiographs from Osteoarthritis Initiative, and 101 heterogeneous radiographs from Google image search engine. We presented a deep learning-based ROI detection framework utilizing single-shot multi-box detector (SSD) with ResNet-101 backbone and customized head structure based on the characteristics of the obtained datasets, whose ground truths were labeled by non-medical annotators in a simple graphical interface. Results: Our method achieved average intersection over union (IoU)=0.8115, average confidence=0.9812, and average precision with threshold IoU=0.5 (AP50)=0.9901 in the independent test set, suggesting that the detected hip regions have appropriately covered main features of the hip joints. Conclusion: The proposed approach featured on low-cost labeling, data-driven model design, and heterogeneous data testing. We have demonstrated the feasibility of training a robust hip region detector for AP pelvic radiographs. This practical framework has a promising potential for a wide range of medical image applications. | ['Chien-Hung Liao', 'Shann-Ching Chen', 'Chih-Chi Chen', 'Feng-Yu Liu'] | 2021-03-02 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [ 1.73697978e-01 3.10301065e-01 -3.47117245e-01 -7.98763856e-02
-1.27645683e+00 -1.91400405e-02 1.63587496e-01 -5.75699797e-03
-6.96500719e-01 7.47624636e-01 2.25070357e-01 -6.48165792e-02
-2.39940539e-01 -8.24868739e-01 -6.46197081e-01 -5.20429254e-01
-3.65457058e-01 8.19339395e-01 8.91626596e-01 1.38189495e-01
1.83957279e-01 5.69314182e-01 -1.12343347e+00 2.12814033e-01
3.94582212e-01 9.13465381e-01 5.75527847e-01 7.23022044e-01
5.02358973e-01 9.28436756e-01 -5.40203631e-01 -1.64168507e-01
2.41937578e-01 -3.20423692e-01 -6.83009624e-01 -5.17667085e-02
7.32972473e-02 -5.00384271e-01 -3.71867269e-01 6.20555341e-01
1.02681267e+00 -1.31261125e-01 7.86762178e-01 -8.26317191e-01
-3.25717270e-01 6.05158329e-01 -1.03023458e+00 8.05562437e-01
-5.30058425e-03 2.94852316e-01 7.31899083e-01 -1.01520264e+00
8.75320375e-01 6.31957650e-01 9.76823926e-01 4.97020841e-01
-6.31945670e-01 -7.94792712e-01 -9.24444437e-01 3.13571282e-02
-1.43931639e+00 -1.30783496e-02 1.26060858e-01 -5.14297605e-01
6.68387473e-01 3.68311137e-01 6.16840780e-01 7.39972949e-01
9.37230468e-01 5.06499887e-01 1.01124406e+00 -4.09009933e-01
1.42838508e-01 -1.94636285e-01 3.38437594e-02 9.76673067e-01
3.75606477e-01 -1.34149864e-01 -1.63433716e-01 -3.71682972e-01
1.24942768e+00 8.48145112e-02 -1.29612744e-01 -3.73076111e-01
-1.39296710e+00 8.17037106e-01 6.48793459e-01 5.35499513e-01
-5.76547205e-01 1.96456611e-01 4.96086150e-01 -3.39584917e-01
1.27687871e-01 1.56782344e-01 1.79251060e-02 4.91942525e-01
-9.67639625e-01 4.39281277e-02 2.94076532e-01 6.74760461e-01
4.73266505e-02 -1.78385109e-01 -5.10983884e-01 8.38646352e-01
5.07433116e-01 6.30762160e-01 1.05561996e+00 -9.48035955e-01
1.94164038e-01 4.71568406e-01 -2.82827139e-01 -7.37010121e-01
-8.54157031e-01 -4.20246094e-01 -7.71627605e-01 1.78642690e-01
4.82175857e-01 -1.02132499e-01 -1.16520309e+00 1.27697098e+00
4.65000689e-01 -1.10997282e-01 -3.19315374e-01 1.19360983e+00
1.07030594e+00 2.07235724e-01 3.56315255e-01 -6.49421215e-02
2.03127575e+00 -7.92337656e-01 -3.41475546e-01 -9.57806110e-02
7.37991273e-01 -5.34488857e-01 1.07169127e+00 1.41340485e-02
-1.06877148e+00 -3.69401813e-01 -1.10903442e+00 1.73928499e-01
1.56441376e-01 4.39548045e-01 2.79844195e-01 7.22308576e-01
-8.82669926e-01 4.28884625e-01 -9.43886757e-01 -4.04933721e-01
6.10186756e-01 6.80448711e-01 -4.00051862e-01 -4.84272651e-02
-1.18691099e+00 1.02019215e+00 3.30108225e-01 -1.47094086e-01
-9.53768790e-01 -6.48467064e-01 -5.06864071e-01 -1.02817737e-01
5.54336369e-01 -8.98352504e-01 1.25625813e+00 -2.83662051e-01
-9.12978530e-01 1.26753235e+00 4.41589743e-01 -6.31835401e-01
6.64782226e-01 -2.25323793e-02 -2.69315511e-01 5.89425027e-01
6.60788834e-01 6.17091596e-01 3.47736329e-01 -7.83769906e-01
-5.68772256e-01 -6.27802312e-01 -4.93398637e-01 2.78188437e-01
1.93631813e-01 4.68994230e-01 -6.01626873e-01 -7.24716067e-01
1.24729238e-01 -9.17511344e-01 -5.21948040e-01 2.04176411e-01
-3.37485224e-01 -2.81049535e-02 4.04135436e-01 -9.75450814e-01
1.17840528e+00 -1.90659571e+00 -5.40564597e-01 3.90310913e-01
4.77194518e-01 -8.29387531e-02 5.95258772e-01 -4.14100766e-01
-9.67350304e-02 1.46556139e-01 -4.79827225e-01 2.50872970e-01
-5.14316916e-01 -8.77586603e-02 4.07607377e-01 8.68310571e-01
-1.69863641e-01 8.84577274e-01 -6.48734272e-01 -1.31227756e+00
2.86989510e-01 2.73520410e-01 -3.53808433e-01 5.89573495e-02
3.96384984e-01 4.02739555e-01 -5.44239581e-01 9.47993934e-01
2.50923544e-01 -5.43687463e-01 1.04691252e-01 -2.35242978e-01
1.10076323e-01 -3.64840120e-01 -1.28787017e+00 1.71788359e+00
-2.50855744e-01 2.25266039e-01 -7.28310132e-03 -5.69180965e-01
7.24171400e-01 5.23333371e-01 1.07195377e+00 -1.00086617e+00
2.69073635e-01 5.45141757e-01 1.30788907e-01 -5.66762269e-01
7.88023025e-02 -2.54687190e-01 -1.43448204e-01 5.27079642e-01
-3.67762288e-03 1.62404984e-01 1.26652777e-01 1.03337906e-01
1.29532743e+00 -8.06095004e-02 5.90348721e-01 -3.55747372e-01
5.88067949e-01 2.44861051e-01 4.50772345e-01 1.02644169e+00
-4.88546282e-01 1.33863449e+00 3.43501508e-01 -3.38454127e-01
-1.07250488e+00 -1.14945388e+00 -5.42947769e-01 8.95961761e-01
3.46561708e-02 2.99406201e-02 -7.38202214e-01 -6.49041235e-01
-2.66243279e-01 1.05616763e-01 -7.49928892e-01 1.29864350e-01
-1.09997845e+00 -1.00889313e+00 7.23964810e-01 6.36626422e-01
5.72521210e-01 -1.24832773e+00 -1.27796721e+00 1.98978692e-01
7.60119706e-02 -8.08514416e-01 -3.83859485e-01 2.01716661e-01
-1.08049357e+00 -1.48817635e+00 -1.61139083e+00 -7.73914516e-01
4.39976603e-01 -5.43340035e-02 1.18095660e+00 1.67891964e-01
-1.04094279e+00 4.17234868e-01 1.88476861e-01 -6.37837574e-02
-3.59579444e-01 1.28419157e-02 -1.91587478e-01 -6.54472530e-01
-1.50128543e-01 -5.72600998e-02 -1.07963300e+00 6.16995454e-01
-6.76826239e-01 -2.01266557e-01 1.03225255e+00 8.04591775e-01
8.95503342e-01 -3.80622268e-01 5.20871222e-01 -1.01140404e+00
2.79362053e-01 -7.56131828e-01 -2.94197768e-01 4.48807925e-01
-5.76530039e-01 -3.70861471e-01 1.49480179e-01 -4.87532131e-02
-9.68003750e-01 1.93528503e-01 -1.84515476e-01 -2.08654299e-01
1.57477930e-02 2.01553106e-01 3.94976914e-01 -3.54639627e-02
9.92254138e-01 -5.77564053e-02 2.56566912e-01 -1.55482933e-01
4.15581241e-02 6.10677302e-01 9.08107400e-01 -4.77921516e-01
-2.47465707e-02 4.96187717e-01 2.89018810e-01 -6.96884274e-01
-4.57927734e-01 -6.57101035e-01 -6.55842066e-01 -4.77753341e-01
1.11725783e+00 -1.03923857e+00 -3.51947248e-01 1.19487658e-01
-6.59075201e-01 1.95237454e-02 -3.78336638e-01 7.84251571e-01
-5.39201975e-01 4.34258968e-01 -8.33335638e-01 -3.38442713e-01
-9.53168571e-01 -1.56167233e+00 1.14092481e+00 3.32285255e-01
-3.99509430e-01 -6.15844548e-01 1.38941213e-01 4.49217349e-01
4.85176384e-01 4.28431153e-01 7.88376927e-01 -9.18587208e-01
-3.63551944e-01 -3.77130061e-01 -3.59648585e-01 -3.17985304e-02
-1.28423750e-01 -1.76844254e-01 -7.22119331e-01 9.47997868e-02
1.01689726e-01 -1.69910848e-01 5.53125501e-01 9.54344392e-01
1.13698697e+00 3.91454518e-01 -5.00117600e-01 4.14768338e-01
1.45498705e+00 4.04793084e-01 5.49771607e-01 6.58050001e-01
4.78794634e-01 6.22478202e-02 7.19506681e-01 4.95142847e-01
1.50768913e-03 5.46296000e-01 4.91212249e-01 -3.23882133e-01
-5.42753458e-01 1.77182183e-01 -1.36062279e-01 3.88715327e-01
-3.61528635e-01 6.95258528e-02 -1.49829292e+00 5.92847049e-01
-1.36119509e+00 -5.07918239e-01 -4.82982755e-01 2.07512593e+00
6.54998362e-01 3.64438832e-01 2.85942525e-01 -1.82689026e-01
1.17362785e+00 -2.37036511e-01 -5.02491713e-01 3.61982107e-01
2.55298615e-01 5.67633212e-01 9.03264880e-01 -1.44641250e-02
-1.26237333e+00 2.02637389e-01 6.33084774e+00 7.41020918e-01
-8.31134498e-01 6.35683119e-01 8.28552783e-01 -8.70825574e-02
2.73970157e-01 -3.40737760e-01 -5.74635208e-01 4.66400534e-01
9.19570982e-01 -1.60820410e-01 -5.22074282e-01 1.10447049e+00
-1.20742442e-02 -5.56582510e-01 -6.86558783e-01 6.72876596e-01
3.02449055e-02 -1.48317635e+00 -3.62355977e-01 3.93132791e-02
3.50222707e-01 3.71767133e-01 -9.75827873e-02 3.76062751e-01
-3.84042300e-02 -1.01309013e+00 3.19859475e-01 4.17327791e-01
1.10921490e+00 -4.54804480e-01 1.00916088e+00 1.98593512e-01
-1.02563417e+00 2.08079237e-02 -2.25210994e-01 4.96649355e-01
3.72935608e-02 3.34796369e-01 -1.29520082e+00 3.94793212e-01
7.30073214e-01 2.05388710e-01 -6.59977555e-01 1.35503983e+00
2.03217044e-01 4.99898374e-01 -4.77125913e-01 1.54419243e-01
-1.18586782e-03 2.57200152e-01 3.11879337e-01 1.15518832e+00
3.76263231e-01 2.09448949e-01 -6.47988394e-02 5.88582337e-01
-2.21039385e-01 4.21512574e-01 -5.04788220e-01 7.38463879e-01
4.02807117e-01 1.28424764e+00 -1.44361413e+00 -4.91903663e-01
-4.84843463e-01 2.34872282e-01 -3.61635476e-01 -2.37989724e-01
-1.28726900e+00 -4.71502654e-02 -4.65775847e-01 6.16453946e-01
-1.24257535e-01 3.53143692e-01 -4.84262407e-01 -7.40597785e-01
-1.88564435e-01 -6.57499611e-01 9.26395059e-01 -8.36954951e-01
-9.20676768e-01 4.16496158e-01 1.96315005e-01 -1.24911880e+00
-1.99177563e-01 -3.69081616e-01 -5.29631615e-01 6.82136774e-01
-8.54146600e-01 -9.17477846e-01 -3.62002224e-01 5.59366584e-01
6.91964328e-01 -1.48734003e-01 4.76724148e-01 6.47030056e-01
-5.54797888e-01 2.97897816e-01 -7.99085423e-02 3.74895930e-01
6.35953128e-01 -1.11073196e+00 6.35182187e-02 6.28328145e-01
-3.38302851e-01 5.24479568e-01 2.85768807e-01 -8.98483574e-01
-1.12293136e+00 -8.98157239e-01 3.50935876e-01 -2.44307861e-01
4.56178904e-01 4.69523907e-01 -6.93960369e-01 5.91056526e-01
-6.25569969e-02 5.49042225e-01 6.23845994e-01 -7.38602161e-01
5.37683010e-01 2.43854389e-01 -1.49215949e+00 2.79396683e-01
5.19168556e-01 2.02460389e-04 -6.48349583e-01 6.09304070e-01
2.50946075e-01 -9.28569615e-01 -1.29662991e+00 8.15266013e-01
6.84422612e-01 -8.88691664e-01 1.37483382e+00 -2.70359099e-01
5.31139731e-01 -8.13917220e-02 -7.29061440e-02 -2.24465787e-01
-2.40024954e-01 1.91159204e-01 -4.16072980e-02 6.32907629e-01
2.91209817e-01 -2.30570570e-01 1.07780623e+00 5.85264564e-01
-3.73332560e-01 -1.01174772e+00 -1.27588916e+00 -2.64090687e-01
-5.78256175e-02 -3.49548697e-01 3.38948146e-02 6.30172431e-01
-5.18621385e-01 -5.58762141e-02 -1.08056396e-01 5.35286479e-02
7.20343888e-01 -3.10068130e-01 4.80888158e-01 -1.09642458e+00
-2.83417493e-01 -3.18479687e-01 -4.61191803e-01 -3.10605913e-01
-4.48880404e-01 -1.00587761e+00 -1.72140256e-01 -1.73134506e+00
8.00035954e-01 -4.85248983e-01 -4.86172169e-01 2.77016491e-01
-6.53116032e-02 6.11324608e-01 -1.36278838e-01 7.16908216e-01
-4.87231344e-01 -5.92636457e-03 1.46473646e+00 1.69749379e-01
2.10926697e-01 3.11469227e-01 -2.39727825e-01 8.81618381e-01
5.26730835e-01 -8.93774033e-01 -1.38704211e-01 1.37271136e-01
-1.21913873e-01 5.84516406e-01 5.26475847e-01 -1.44025671e+00
2.25061864e-01 2.43426889e-01 7.52051413e-01 -7.43719339e-01
9.47541147e-02 -6.47471964e-01 2.83129811e-01 1.09588027e+00
-1.51008070e-01 -4.34514415e-03 -1.20843470e-01 4.86567646e-01
1.07594142e-02 -5.35598993e-01 1.02486360e+00 -8.81682336e-01
-8.14468145e-01 9.31040198e-02 -3.35997760e-01 2.75027335e-01
1.41815245e+00 -4.22105968e-01 -1.62234798e-01 2.24675223e-01
-1.13615119e+00 9.74266306e-02 1.72721654e-01 8.94046351e-02
7.20096767e-01 -1.14156783e+00 -6.11479938e-01 8.64787959e-04
-1.27527609e-01 1.56081513e-01 2.27413818e-01 1.31875634e+00
-1.14117992e+00 4.20022160e-01 -4.28530782e-01 -1.09415781e+00
-1.24842751e+00 1.79935798e-01 5.99756479e-01 -6.52710736e-01
-1.19805300e+00 7.35406101e-01 1.97993055e-01 -2.60012746e-01
1.05999820e-01 -4.22262102e-01 -2.89932489e-01 -1.29251834e-02
2.71318316e-01 4.21260774e-01 2.95833290e-01 -7.21547008e-01
-6.42367184e-01 5.62251866e-01 -5.25617413e-02 1.43380379e-02
1.18519461e+00 4.40615974e-02 2.34179735e-01 2.66067594e-01
9.92639959e-01 -1.89638421e-01 -7.22026229e-01 -1.09760411e-01
1.04162142e-01 -2.50031590e-01 3.32455933e-02 -6.52025163e-01
-1.17033863e+00 4.60505486e-01 1.31600332e+00 -2.26458624e-01
8.36532712e-01 3.43476415e-01 7.49964356e-01 6.14304952e-02
5.06145418e-01 -1.00151193e+00 1.11702479e-01 7.04307458e-04
6.25798583e-01 -1.29723668e+00 5.81825256e-01 -2.06254259e-01
-9.23977256e-01 1.22167873e+00 6.76325321e-01 -2.76886433e-01
5.13509274e-01 2.88539648e-01 -1.33584097e-01 -7.03076780e-01
-2.67838299e-01 1.34977266e-01 1.71629921e-01 3.18165720e-01
6.68938696e-01 -2.97038052e-02 -5.68974137e-01 6.71042919e-01
2.30271578e-01 2.52332389e-01 5.02467096e-01 1.27739477e+00
-7.67663419e-01 -4.20858204e-01 -7.34737396e-01 7.71864414e-01
-1.06552315e+00 1.62471011e-01 2.14888781e-01 1.46747518e+00
2.06855312e-02 1.02783903e-01 -1.91796944e-01 2.19524235e-01
3.80967706e-01 -6.39290512e-02 2.28586257e-01 -7.26075232e-01
-6.96670532e-01 2.09633216e-01 -3.81976292e-02 -4.40356255e-01
-3.78336012e-01 -5.31284630e-01 -1.71527803e+00 4.21236992e-01
-3.67624134e-01 -1.72173753e-01 4.99993294e-01 9.01688576e-01
1.69810690e-02 8.10199976e-01 3.61735225e-01 -6.34776235e-01
-3.92926604e-01 -9.96613026e-01 -6.64421737e-01 1.82225332e-01
-1.44918919e-01 -9.58457470e-01 -1.60329744e-01 3.40641662e-02] | [15.067214012145996, -2.2918591499328613] |
2fba2663-d453-4ce8-b0b3-8045c2ea96d1 | evaluating-the-role-of-language-typology-in | 2004.13939 | null | https://arxiv.org/abs/2004.13939v2 | https://arxiv.org/pdf/2004.13939v2.pdf | Evaluating Transformer-Based Multilingual Text Classification | As NLP tools become ubiquitous in today's technological landscape, they are increasingly applied to languages with a variety of typological structures. However, NLP research does not focus primarily on typological differences in its analysis of state-of-the-art language models. As a result, NLP tools perform unequally across languages with different syntactic and morphological structures. Through a detailed discussion of word order typology, morphological typology, and comparative linguistics, we identify which variables most affect language modeling efficacy; in addition, we calculate word order and morphological similarity indices to aid our empirical study. We then use this background to support our analysis of an experiment we conduct using multi-class text classification on eight languages and eight models. | ['William Yang Wang', 'Sharon Levy', 'Aesha Parekh', 'Sophie Groenwold', 'Samhita Honnavalli', 'Lily Ou', 'Diba Mirza'] | 2020-04-29 | null | null | null | null | ['multilingual-text-classification'] | ['miscellaneous'] | [-3.18818152e-01 -3.20679367e-01 -9.89187419e-01 -4.90918234e-02
-2.77560383e-01 -9.48056042e-01 7.20312953e-01 4.50941861e-01
-7.14897394e-01 4.52961206e-01 7.35134602e-01 -1.16109014e+00
-2.65599161e-01 -6.03793085e-01 -1.99950129e-01 1.48398831e-01
4.22021985e-01 5.32500505e-01 -2.14397013e-01 -1.81804687e-01
7.09471226e-01 1.90512046e-01 -1.28900504e+00 4.07382101e-01
1.12134469e+00 2.78958648e-01 2.81256586e-01 1.02539614e-01
-7.67912686e-01 5.85302234e-01 -4.70416844e-01 -5.92266977e-01
8.78034830e-02 -2.66568750e-01 -7.92266726e-01 -3.45434159e-01
4.21039850e-01 1.73414037e-01 -1.83032930e-01 9.97822225e-01
3.19575310e-01 -7.83890206e-03 9.77840900e-01 -8.08430552e-01
-1.16891015e+00 1.18030012e+00 -2.70273685e-01 4.15959030e-01
7.00922549e-01 3.98831964e-02 1.39645827e+00 -9.49806452e-01
1.13647747e+00 1.62977302e+00 8.30836058e-01 2.22169548e-01
-1.35078585e+00 -7.28087485e-01 1.46447808e-01 1.12841174e-01
-1.38015985e+00 -3.17132354e-01 7.11451650e-01 -7.13364244e-01
1.13111031e+00 1.08286940e-01 7.60671377e-01 1.10303509e+00
5.99094510e-01 3.91273230e-01 1.44541800e+00 -8.34770799e-01
-8.04219991e-02 4.14214969e-01 4.21037376e-01 2.33971477e-01
5.80276608e-01 -7.01389313e-02 -3.00819874e-01 -2.74404109e-01
3.35794389e-01 -3.08612049e-01 1.62604511e-01 2.82720953e-01
-1.15172327e+00 1.17605686e+00 -2.48379141e-01 9.55761731e-01
-1.08003885e-01 -3.02117378e-01 5.33528864e-01 2.98160672e-01
6.67831004e-01 8.03759933e-01 -4.76535261e-01 -2.23575354e-01
-9.74979937e-01 3.37088853e-01 1.08225942e+00 6.73672140e-01
3.48731905e-01 -7.01765623e-03 3.74765575e-01 1.47105789e+00
6.34257734e-01 5.56897163e-01 9.50211167e-01 -6.70606494e-01
4.13447142e-01 6.99584782e-01 -2.07238466e-01 -1.14961278e+00
-4.99661595e-01 -8.10635388e-02 -1.04562506e-01 -2.61109501e-01
4.83724892e-01 7.90300779e-03 -6.63226187e-01 1.22912073e+00
-1.58980653e-01 -7.87341595e-01 -1.41434580e-01 3.26024681e-01
8.44230413e-01 8.23429704e-01 6.43533051e-01 -2.92429984e-01
1.65913188e+00 -5.39850354e-01 -5.40570080e-01 -4.22155112e-01
1.12240565e+00 -1.18219352e+00 1.65732491e+00 1.60696551e-01
-9.93361056e-01 -3.37185800e-01 -6.69096291e-01 -3.16171438e-01
-7.35919774e-01 -4.69931625e-02 7.29636133e-01 9.15488422e-01
-7.86086321e-01 1.71126023e-01 -4.60863531e-01 -8.28521609e-01
-3.23012508e-02 -1.81616485e-01 -5.69021441e-02 1.68676153e-01
-1.12508261e+00 1.42710161e+00 6.71569824e-01 -4.98063177e-01
1.77590385e-01 -8.04631054e-01 -7.03823626e-01 -2.55034298e-01
5.62453829e-02 -4.70072329e-01 1.11779702e+00 -1.04438257e+00
-1.03233600e+00 1.09192157e+00 -2.16968581e-01 -2.10946929e-02
2.43475184e-01 1.85022280e-01 -8.34525883e-01 -3.56710851e-01
3.03771764e-01 4.49103117e-01 5.72509244e-02 -1.32528400e+00
-9.29969311e-01 -2.29380548e-01 -1.05237141e-01 2.26881638e-01
-5.89616179e-01 8.98761153e-01 9.07053500e-02 -9.98661160e-01
1.68454379e-01 -7.04821646e-01 8.90002325e-02 -3.91753972e-01
3.08473296e-02 -5.13057709e-01 2.25382909e-01 -7.49796391e-01
1.72592020e+00 -1.90903020e+00 -1.60452768e-01 3.14175963e-01
1.31279603e-01 -1.28212735e-01 6.00498728e-03 8.55296969e-01
2.49890849e-01 1.01362264e+00 1.41899154e-01 -1.82768315e-01
2.64525115e-01 2.57126957e-01 -4.34999973e-01 2.24567965e-01
-4.13268775e-01 8.65271866e-01 -7.50169218e-01 -8.42112124e-01
1.13544047e-01 3.01125199e-01 -5.64374328e-01 -6.29419327e-01
5.62647209e-02 -1.82342082e-01 -1.14722975e-01 8.18813622e-01
2.30138332e-01 2.14502290e-01 6.98830128e-01 6.08181991e-02
-7.93250620e-01 9.78059947e-01 -6.90940380e-01 1.00085175e+00
-7.42453575e-01 8.13257515e-01 -4.31092352e-01 -5.04601002e-01
7.92336226e-01 2.96310820e-02 2.00899735e-01 -7.15347707e-01
2.65777737e-01 7.22793400e-01 8.62808883e-01 -3.60748291e-01
8.66377234e-01 -3.74304235e-01 -1.68259189e-01 5.26974440e-01
-7.65956864e-02 -2.30023980e-01 6.50484920e-01 -1.11674383e-01
6.76083386e-01 -1.24353811e-01 6.38473213e-01 -9.28708136e-01
1.91062078e-01 5.02743840e-01 5.09976625e-01 5.40589392e-01
2.03987933e-03 8.58938694e-02 3.07270586e-01 -4.34753597e-01
-1.33639896e+00 -7.97294140e-01 -7.77007639e-01 1.10516357e+00
-1.53282121e-01 -6.16334438e-01 -4.02207822e-01 -3.71566713e-01
2.34164208e-01 1.28655839e+00 -3.56967121e-01 3.05842996e-01
-6.45012498e-01 -7.49212027e-01 7.72282720e-01 3.64942551e-01
-6.82840198e-02 -1.37144196e+00 -3.75758231e-01 2.35012606e-01
-3.52086090e-02 -9.15783644e-01 -2.50196397e-01 -1.68958649e-01
-8.79306614e-01 -1.01546061e+00 -3.24837446e-01 -1.03220105e+00
3.51290792e-01 2.06358209e-01 1.24627566e+00 5.06649539e-02
2.19365180e-01 2.46014789e-01 -6.01641953e-01 -6.15941882e-01
-9.47951317e-01 4.58241373e-01 1.72303230e-01 -8.32221508e-01
9.23067987e-01 -3.76527011e-01 1.89278334e-01 3.16906497e-02
-8.89107823e-01 -3.31722945e-02 3.79870474e-01 4.91055429e-01
1.77358091e-01 1.42773554e-01 4.27275926e-01 -1.00261784e+00
1.10115409e+00 -6.95898831e-01 -3.58305275e-01 4.47942883e-01
-8.92290711e-01 -3.81218791e-01 7.48087585e-01 -7.62041867e-01
-8.14618051e-01 -7.40444005e-01 -1.78666875e-01 3.46730679e-01
3.94107290e-02 1.27719867e+00 5.64857163e-02 4.68666181e-02
7.37849534e-01 4.57210615e-02 -5.18477708e-02 -5.14715791e-01
2.87176281e-01 8.75805497e-01 -6.55542910e-02 -6.96129501e-01
7.70574629e-01 5.02450615e-02 -5.28618217e-01 -1.35498941e+00
-3.68546158e-01 -2.44219229e-01 -6.99353456e-01 -8.09317678e-02
4.13288742e-01 -6.77069306e-01 -5.35391808e-01 2.37880751e-01
-1.07864654e+00 -3.78666878e-01 -7.70878419e-02 7.82411933e-01
-1.60716429e-01 3.87721688e-01 -7.82481790e-01 -7.07602978e-01
-1.12291865e-01 -8.76590014e-01 4.10548389e-01 -1.23256996e-01
-1.12525070e+00 -1.76638877e+00 1.78937301e-01 3.14621389e-01
2.77847767e-01 -2.63814330e-01 1.75147045e+00 -9.76169407e-01
2.77853370e-01 1.47286862e-01 -7.60086849e-02 -3.74403000e-02
1.47427410e-01 3.23862404e-01 -3.66298079e-01 1.16840236e-01
-1.47069439e-01 -2.80040558e-02 3.21906954e-01 3.77510875e-01
5.97856283e-01 -5.24378181e-01 -3.88283491e-01 3.75078261e-01
1.51990855e+00 4.49325651e-01 1.71631619e-01 7.70733058e-01
7.21322179e-01 9.58198607e-01 4.53647107e-01 6.74464926e-02
6.68062270e-01 4.56881374e-01 -5.67598641e-01 3.76794875e-01
-3.93533260e-02 -5.23907602e-01 4.75016385e-01 1.43666911e+00
2.69521534e-01 -3.86042386e-01 -1.80321896e+00 5.93190312e-01
-1.43632483e+00 -6.66270137e-01 -1.31390169e-01 1.90221715e+00
9.78604972e-01 2.51222491e-01 2.49549136e-01 -1.12791534e-03
5.20685911e-01 2.17015579e-01 7.84259960e-02 -8.07359993e-01
-3.30079436e-01 2.59017516e-02 4.92018610e-01 7.57246912e-01
-5.25522649e-01 1.29491615e+00 7.55736828e+00 9.45183575e-01
-1.04189897e+00 -5.01660928e-02 4.10970300e-01 2.73947448e-01
-9.87119615e-01 1.17392376e-01 -9.44457889e-01 6.75957203e-01
9.45705175e-01 -6.18706763e-01 4.04834270e-01 6.70175791e-01
3.63079965e-01 -1.83374912e-01 -9.47122991e-01 8.13900292e-01
2.78091937e-01 -1.24818766e+00 4.00793582e-01 1.76446080e-01
5.72028577e-01 -1.15633063e-01 -5.59718981e-02 3.38898987e-01
4.57890630e-01 -1.08954346e+00 1.12441814e+00 -7.85110816e-02
7.59045780e-01 -5.39093018e-01 4.54012930e-01 3.06487471e-01
-8.48484337e-01 -2.56874144e-01 -4.81669873e-01 -5.81177294e-01
3.51981103e-01 3.15559119e-01 -4.96437222e-01 2.79099625e-02
4.75912601e-01 4.66731131e-01 -5.60814619e-01 7.08255768e-01
-1.34501368e-01 1.05692434e+00 -2.76764750e-01 -4.84352469e-01
1.00399703e-01 -3.32166702e-01 6.09514952e-01 1.44854701e+00
2.23210886e-01 -2.03823168e-02 2.57761896e-01 6.93822980e-01
1.43300653e-01 8.35260630e-01 -8.02628636e-01 -5.19001961e-01
1.11042750e+00 8.52176070e-01 -1.24940991e+00 -3.02702546e-01
-7.91467547e-01 2.26079151e-01 1.95001021e-01 2.81059176e-01
-4.67790574e-01 -2.23427936e-01 7.24412680e-01 4.86459523e-01
-3.50275934e-01 -7.05788314e-01 -1.04346263e+00 -1.08365870e+00
-1.60193130e-01 -1.24944711e+00 2.15060011e-01 -4.08730000e-01
-1.68234277e+00 3.04178596e-01 4.36666697e-01 -6.17495954e-01
-1.02095157e-01 -1.05730653e+00 -4.26138163e-01 9.20846343e-01
-9.95450616e-01 -1.22065473e+00 3.76579583e-01 -1.72504276e-01
6.35658205e-01 -3.58927578e-01 7.76514292e-01 2.39513442e-01
-3.78547311e-01 5.22215068e-01 2.79713392e-01 1.13331102e-01
7.05203533e-01 -9.38008130e-01 7.24805832e-01 5.58146060e-01
3.08336824e-01 1.04418099e+00 6.26924634e-01 -1.18602109e+00
-1.05369258e+00 -4.75593626e-01 1.73121119e+00 -6.79065764e-01
1.40252423e+00 -3.95688713e-01 -6.46082759e-01 7.85379231e-01
2.96111047e-01 -8.34784448e-01 1.05369771e+00 6.59348726e-01
-3.21734667e-01 3.50064993e-01 -8.95973682e-01 1.17884624e+00
1.16843963e+00 -6.09626293e-01 -1.03081095e+00 2.51632094e-01
5.57012141e-01 -8.32981393e-02 -9.81448412e-01 1.15690634e-01
9.18300807e-01 -3.66862029e-01 6.14845872e-01 -7.17272520e-01
5.39052606e-01 9.40847099e-02 -2.60158211e-01 -1.26456952e+00
-6.27727747e-01 -3.05187911e-01 5.45060575e-01 1.42296612e+00
8.42913806e-01 -1.06672823e+00 2.89521188e-01 5.45496106e-01
-7.86899924e-02 -6.15209222e-01 -6.81285918e-01 -9.23676908e-01
9.75242138e-01 -6.75885260e-01 4.60814714e-01 1.38241017e+00
4.91232216e-01 3.13783348e-01 3.83404046e-01 -4.94905889e-01
2.25453585e-01 -1.03669107e-01 3.97332132e-01 -1.44099784e+00
1.94561724e-02 -1.04077649e+00 -2.97905475e-01 -5.87804675e-01
6.56154871e-01 -1.19347966e+00 -1.97094277e-01 -1.46349931e+00
3.03520709e-01 -6.33852243e-01 1.98849022e-01 4.87798899e-01
1.40593052e-02 2.66502112e-01 4.23191190e-01 4.91177827e-01
8.87724459e-02 6.67444170e-02 6.63295090e-01 2.83560790e-02
-5.07482111e-01 -5.42610943e-01 -9.13864136e-01 1.16617465e+00
9.72059488e-01 -5.72573304e-01 -2.59655595e-01 -4.67375666e-01
7.37582862e-01 -5.68550706e-01 -7.16731921e-02 -5.80987036e-01
-4.52501886e-02 -7.30391979e-01 2.35465661e-01 -1.29085794e-01
-2.69337147e-01 -7.02023745e-01 1.15199767e-01 4.77877080e-01
-3.26153368e-01 7.08384693e-01 4.91006464e-01 -3.42420995e-01
-7.93458670e-02 -6.76821589e-01 4.31364179e-01 -1.05853453e-02
-6.19131088e-01 -1.13467038e-01 -1.00426388e+00 4.53058362e-01
7.25463986e-01 -5.36523819e-01 -3.73590887e-01 -7.89784733e-03
-2.44841754e-01 -4.62377968e-04 1.08865273e+00 7.55549550e-01
2.87120230e-02 -1.23426688e+00 -6.62411034e-01 -1.14047147e-01
8.39675441e-02 -7.88927197e-01 -2.02708036e-01 4.66839075e-01
-8.64854693e-01 6.71448350e-01 -1.29059166e-01 -1.73528511e-02
-1.13977075e+00 4.88318294e-01 4.68241535e-02 -1.54106289e-01
-4.90246713e-01 2.65877753e-01 3.53587456e-02 -7.24212527e-01
-8.55473131e-02 -4.81644183e-01 -3.45645010e-01 4.47406262e-01
-1.59658026e-02 6.58790350e-01 -4.09475327e-01 -9.98568714e-01
-3.03960174e-01 4.90982980e-01 -3.26840095e-02 -6.77840173e-01
8.69784117e-01 -2.22168565e-01 -3.64874691e-01 1.23294759e+00
8.73491764e-01 5.20193577e-01 -2.27392733e-01 1.03427581e-01
4.50171947e-01 -3.65208894e-01 -1.98431492e-01 -8.77666056e-01
-1.06981210e-01 4.16955590e-01 -3.89918648e-02 4.07724559e-01
4.59473699e-01 -1.55057147e-01 5.91878355e-01 1.85631528e-01
1.45475745e-01 -1.49823523e+00 -5.87260425e-01 1.00895822e+00
6.89932942e-01 -8.05788636e-01 1.71448171e-01 -4.99012858e-01
-6.33726239e-01 9.82501984e-01 3.98247540e-01 1.93299100e-01
8.77795875e-01 3.46060783e-01 4.19771910e-01 -7.32778534e-02
-5.94152272e-01 1.02611214e-01 3.41213524e-01 5.01186788e-01
1.12358308e+00 2.23090187e-01 -1.50378144e+00 5.93628109e-01
-9.64732587e-01 -3.03785384e-01 4.41894829e-01 6.21718764e-01
-3.98707956e-01 -1.46481049e+00 -4.64421272e-01 7.03551114e-01
-7.10406482e-01 -6.10401094e-01 -6.53323948e-01 1.40568161e+00
8.81869867e-02 1.02598548e+00 3.61826777e-01 -3.79885316e-01
1.31039005e-02 4.37406689e-01 3.62628043e-01 -8.54165792e-01
-9.00287986e-01 -4.84312847e-02 4.58153039e-01 1.65277161e-02
-2.63741046e-01 -1.11957681e+00 -1.05118525e+00 -9.02806520e-01
-3.02900642e-01 2.03317955e-01 8.12726557e-01 1.04019964e+00
2.51917411e-02 4.73037884e-02 8.91573131e-02 -4.07244831e-01
-4.01655853e-01 -1.30162799e+00 -6.43695474e-01 2.77541876e-02
-4.50113088e-01 -5.30840755e-01 -5.02666652e-01 -1.32055655e-01] | [10.330680847167969, 9.927830696105957] |
e3c45c61-c47b-4049-a28e-df5aab3b9cc0 | tree-based-optimization-a-meta-algorithm-for | 1809.09284 | null | http://arxiv.org/abs/1809.09284v1 | http://arxiv.org/pdf/1809.09284v1.pdf | Tree-Based Optimization: A Meta-Algorithm for Metaheuristic Optimization | Designing search algorithms for finding global optima is one of the most
active research fields, recently. These algorithms consist of two main
categories, i.e., classic mathematical and metaheuristic algorithms. This
article proposes a meta-algorithm, Tree-Based Optimization (TBO), which uses
other heuristic optimizers as its sub-algorithms in order to improve the
performance of search. The proposed algorithm is based on mathematical tree
subject and improves performance and speed of search by iteratively removing
parts of the search space having low fitness, in order to minimize and purify
the search space. The experimental results on several well-known benchmarks
show the outperforming performance of TBO algorithm in finding the global
solution. Experiments on high dimensional search spaces show significantly
better performance when using the TBO algorithm. The proposed algorithm
improves the search algorithms in both accuracy and speed aspects, especially
for high dimensional searching such as in VLSI CAD tools for Integrated Circuit
(IC) design. | ['Hoda Mohammadzade', 'Saeed Sharifian', 'Benyamin Ghojogh'] | 2018-09-25 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 1.41446874e-01 -3.52006137e-01 -2.38259360e-01 -3.94910611e-02
4.71571796e-02 -2.96800554e-01 -8.05819482e-02 1.34708226e-01
-3.46970320e-01 1.04954016e+00 -1.76245525e-01 -1.63517877e-01
-9.76720870e-01 -9.65551972e-01 -1.81553382e-02 -9.42162812e-01
2.11258885e-02 6.67559981e-01 1.64440498e-01 -3.35052639e-01
1.01280105e+00 6.41651273e-01 -1.61385119e+00 -2.18726918e-01
1.16327190e+00 9.64418828e-01 3.69929850e-01 2.84285337e-01
-5.23212433e-01 -8.74324143e-03 -6.47760868e-01 -7.72883892e-02
2.40754366e-01 -4.96467799e-01 -6.20349169e-01 -9.82654318e-02
-7.06863642e-01 6.59124732e-01 1.83973536e-01 1.19577909e+00
6.77440882e-01 1.79778576e-01 3.60210627e-01 -1.23015428e+00
-1.68350041e-01 3.74448180e-01 -6.90708041e-01 4.17548865e-01
3.01107287e-01 -8.86865854e-02 7.39090979e-01 -4.23536479e-01
6.11829519e-01 1.41460800e+00 5.80578923e-01 2.58888096e-01
-8.23177099e-01 -7.03161299e-01 -1.90588191e-01 5.54007411e-01
-1.50176752e+00 -4.67839576e-02 9.35172677e-01 1.04756549e-01
1.18242550e+00 7.63391614e-01 5.57449639e-01 1.38583556e-01
7.51673460e-01 5.04077256e-01 9.36798096e-01 -6.79229558e-01
6.04263484e-01 -3.13678347e-02 3.26272786e-01 5.57416439e-01
7.32919931e-01 2.62508482e-01 -3.19006622e-01 -4.27164137e-01
3.21421832e-01 -1.74737707e-01 -1.35811314e-01 -4.92901266e-01
-8.56103599e-01 8.51326644e-01 3.20309013e-01 8.15841675e-01
-4.91808921e-01 9.65191424e-03 2.53672391e-01 1.80359453e-01
-1.29000261e-01 9.71722841e-01 -3.98516178e-01 -2.13108867e-01
-8.66918325e-01 2.81221986e-01 7.44803727e-01 6.21984303e-01
4.46107835e-01 1.38811544e-01 -1.69646833e-02 8.19535017e-01
3.09956163e-01 1.71221510e-01 8.78637075e-01 -5.02564311e-01
4.87105638e-01 1.09448338e+00 1.54764235e-01 -1.36230099e+00
-5.85430145e-01 -8.66793513e-01 -7.33575642e-01 2.94081569e-01
-2.96399772e-01 -7.16451257e-02 -1.00095582e+00 1.05584955e+00
6.10735416e-01 -1.46795854e-01 8.23841542e-02 6.97465360e-01
5.80589414e-01 1.02163589e+00 -3.56409639e-01 -8.39703679e-01
9.85575438e-01 -1.21444106e+00 -1.14501262e+00 -1.66020473e-03
5.62635958e-01 -1.08644807e+00 2.70507246e-01 5.72747409e-01
-1.05495000e+00 -3.98868889e-01 -1.44085467e+00 6.62583172e-01
-5.92094898e-01 -1.53615013e-01 6.76393926e-01 1.11759961e+00
-7.72577345e-01 8.33375514e-01 -7.45365024e-01 -4.12045151e-01
1.35648534e-01 8.74249518e-01 2.16846928e-01 -6.34421185e-02
-8.46654117e-01 7.77561545e-01 7.57448375e-01 3.35701197e-01
-1.76976696e-01 -3.57388973e-01 -4.48123544e-01 9.02893096e-02
5.50971091e-01 -4.89260852e-01 4.11201090e-01 -6.19920552e-01
-1.50685859e+00 4.02807117e-01 -5.36319315e-01 -3.36145103e-01
1.72530502e-01 2.76934266e-01 -5.93616009e-01 -1.72273129e-01
-5.31968335e-03 5.87923303e-02 6.02260947e-01 -1.02228546e+00
-6.52193427e-01 -4.21234995e-01 -4.25974578e-01 1.08063243e-01
-3.00844967e-01 2.18321346e-02 -3.32346767e-01 -5.65363050e-01
4.78669584e-01 -9.03557420e-01 -7.91054547e-01 -4.97096628e-01
-3.06218803e-01 -2.53450274e-01 1.17508602e+00 -2.43995428e-01
2.00136781e+00 -1.75592852e+00 5.03926635e-01 8.37775171e-01
-1.78306982e-01 4.92037952e-01 1.22267336e-01 6.10100925e-01
-1.05198666e-01 2.81484425e-01 -2.38705441e-01 3.59789312e-01
-1.93618491e-01 2.88405001e-01 4.98746574e-01 2.96977699e-01
-1.88995674e-01 7.72089124e-01 -6.05870724e-01 -6.09212637e-01
1.93573594e-01 -2.88453419e-02 -5.68266451e-01 -2.93346852e-01
-9.93678123e-02 9.16307271e-02 -1.10342324e+00 9.40980256e-01
8.29984486e-01 -2.10396107e-02 2.97319353e-01 -2.92803049e-01
-5.44156730e-01 -3.01782429e-01 -1.67182136e+00 1.23974669e+00
-3.92443240e-01 3.75088066e-01 1.82321057e-01 -1.27292490e+00
1.27662265e+00 9.91229117e-02 7.87082374e-01 -8.84524822e-01
5.84527075e-01 4.44956303e-01 3.68680716e-01 -4.62125838e-01
4.60345209e-01 2.21932814e-01 2.27827817e-01 3.60722005e-01
-4.09162909e-01 3.29934736e-03 6.26898110e-01 -4.52350169e-01
9.35135901e-01 -2.22456440e-01 4.97899324e-01 -7.37071157e-01
1.15898895e+00 3.57569784e-01 8.09154093e-01 3.92466366e-01
-3.22520621e-02 6.25659674e-02 1.79177567e-01 -5.26027083e-01
-7.60577321e-01 -4.40220296e-01 -2.98811495e-01 3.51189584e-01
5.95443368e-01 -4.84728962e-01 -5.16439557e-01 -2.21939951e-01
-1.13842279e-01 7.17676401e-01 -4.91047978e-01 -1.89461350e-01
-8.79974365e-01 -1.31221747e+00 -1.15716900e-03 -2.19541118e-02
6.58121347e-01 -1.09477174e+00 -1.06835902e+00 6.57433331e-01
1.04841180e-01 -5.92073381e-01 6.00759201e-02 2.50690192e-01
-1.26523721e+00 -8.11457694e-01 -4.06115621e-01 -8.98803890e-01
7.03451276e-01 -9.42637399e-03 7.98410654e-01 6.54462337e-01
-8.02531242e-01 -9.07237679e-02 -5.75921595e-01 -3.71468663e-01
-1.75003514e-01 1.87558919e-01 -2.50616789e-01 -6.81880414e-02
2.05445010e-02 -5.66236019e-01 -3.23561221e-01 8.47088575e-01
-6.23983264e-01 -4.73699868e-01 7.04370320e-01 8.88517559e-01
8.41504335e-01 1.01909423e+00 4.93629903e-01 -4.28241193e-01
7.36008108e-01 -3.85129929e-01 -1.05280387e+00 4.68032598e-01
-1.03263283e+00 6.76984847e-01 4.94040698e-01 -2.46568963e-01
-7.63355732e-01 -2.28382032e-02 7.84601718e-02 -2.56033987e-01
4.31181014e-01 4.98714417e-01 -2.97965318e-01 -5.52002072e-01
2.49977946e-01 2.84982264e-01 -1.53993979e-01 -8.05126429e-01
-5.09082258e-01 5.41672707e-01 7.44871795e-02 -4.39532548e-01
8.49055529e-01 2.04198673e-01 7.73170114e-01 -6.69540405e-01
-1.74029380e-01 -6.36222720e-01 -2.49441177e-01 -1.37617379e-01
8.53854775e-01 2.54918873e-01 -8.09745133e-01 1.24807946e-01
-7.70992815e-01 4.50468332e-01 2.36758962e-01 3.57847035e-01
-2.29706854e-01 2.55060673e-01 2.90080696e-01 -1.08645105e+00
-6.67362094e-01 -1.46785271e+00 5.78640699e-01 6.53191090e-01
-1.77472815e-01 -1.04988694e+00 -7.99600706e-02 8.48750770e-02
4.66726333e-01 6.46857321e-01 1.18875062e+00 -4.41069603e-01
-6.02719307e-01 -1.89008519e-01 3.03841472e-01 -3.84926915e-01
1.11666135e-01 1.73633650e-01 -1.22026265e-01 -2.48704717e-01
2.66412854e-01 2.35734105e-01 2.15379253e-01 5.60758471e-01
9.29377675e-01 -1.21600546e-01 -1.00054443e+00 7.27398753e-01
1.99187291e+00 1.16634154e+00 8.01980853e-01 1.03805292e+00
1.52832987e-02 4.06097978e-01 1.38770080e+00 5.65213799e-01
-4.56069708e-01 8.60506237e-01 6.35442615e-01 1.45178273e-01
2.37248212e-01 2.58661330e-01 -1.79196507e-01 5.22217035e-01
-3.25510022e-03 -4.69800532e-01 -9.70864117e-01 4.60299909e-01
-2.00922394e+00 -8.72010887e-01 -4.24494684e-01 2.09003639e+00
4.02858436e-01 3.27152222e-01 -4.08108681e-02 7.24621177e-01
1.04283834e+00 -1.11521058e-01 -4.18680400e-01 -1.17574894e+00
7.58565292e-02 5.72981596e-01 7.58541584e-01 4.88660097e-01
-7.30739892e-01 6.24338686e-01 5.97227001e+00 1.15281653e+00
-1.30041635e+00 -3.01327139e-01 3.98099720e-01 -1.80797234e-01
-1.07181691e-01 -8.19275621e-03 -9.12595630e-01 5.27011275e-01
6.41333222e-01 -6.02313936e-01 5.62609017e-01 7.24797130e-01
1.84120774e-01 -2.71078885e-01 -6.21077538e-01 1.28775477e+00
-1.29592931e-02 -1.47369242e+00 -2.09845930e-01 1.48869455e-01
1.14018953e+00 -6.81496024e-01 4.92522605e-02 -1.07274540e-01
-2.32238695e-01 -1.33327150e+00 3.50574881e-01 1.91420153e-01
6.15108525e-03 -1.31153619e+00 1.06575537e+00 2.13061720e-01
-1.31676042e+00 -4.47469771e-01 -4.23530608e-01 1.54873610e-01
3.15125942e-01 4.39653158e-01 -4.22446966e-01 1.07177973e+00
8.80360246e-01 4.23890889e-01 -5.82394660e-01 1.82590020e+00
5.55419385e-01 -5.51979849e-03 -7.01984346e-01 -8.32022250e-01
6.41692400e-01 -5.32507598e-01 9.52148736e-01 5.51171601e-01
4.15380895e-01 7.09504187e-02 4.23794873e-02 6.90592349e-01
4.53675359e-01 3.77340734e-01 -6.57561421e-02 -1.09908856e-01
5.76658726e-01 1.03512013e+00 -1.11125290e+00 2.34385282e-02
2.68489331e-01 4.49297547e-01 -3.29860985e-01 -5.82174882e-02
-1.01784575e+00 -1.02630842e+00 1.42579198e-01 -1.75810933e-01
4.91097361e-01 -5.78309894e-02 -7.51859963e-01 -1.52028129e-01
-6.28014728e-02 -8.52031946e-01 5.43119907e-01 -3.32842469e-01
-5.04606128e-01 7.13495970e-01 1.47040680e-01 -1.15374899e+00
-9.90656093e-02 -6.89311147e-01 -7.55023897e-01 8.27609658e-01
-1.19785750e+00 -4.48032349e-01 -1.53083727e-01 3.44013840e-01
8.36353898e-01 -3.48769516e-01 2.79769450e-01 2.88794905e-01
-7.47798264e-01 5.62804043e-01 5.57618737e-01 -7.70550311e-01
3.51273827e-02 -6.60212338e-01 -2.70779669e-01 6.06961131e-01
-3.65082026e-01 6.25218809e-01 1.09344232e+00 -6.00096762e-01
-1.77378035e+00 -4.48314458e-01 8.05419743e-01 1.35337085e-01
3.50730151e-01 2.08501860e-01 -4.70501631e-01 -9.99629647e-02
9.86673757e-02 -6.38000190e-01 2.88509756e-01 -1.24160396e-02
7.67876804e-01 -2.83044755e-01 -1.53816628e+00 5.36054254e-01
9.96920824e-01 6.10172212e-01 -4.20553237e-01 2.30199918e-01
1.62012562e-01 -3.34490657e-01 -7.29024351e-01 7.63592303e-01
4.86339629e-01 -1.02087450e+00 1.15548813e+00 -7.45069161e-02
-4.53120917e-02 -3.99833202e-01 3.39884683e-02 -8.99789155e-01
-3.77742678e-01 -9.33012187e-01 1.44725904e-01 1.25333500e+00
4.67171729e-01 -7.59759724e-01 9.36383605e-01 2.19677567e-01
-7.24150985e-02 -1.36536443e+00 -1.10608065e+00 -1.10956693e+00
-3.05744678e-01 1.73200384e-01 1.04868197e+00 6.19213641e-01
-3.40884149e-01 2.16982052e-01 -3.40618640e-02 -2.77262926e-02
6.29586756e-01 3.99837822e-01 3.58361810e-01 -1.25170255e+00
-2.05866009e-01 -7.80103505e-01 -5.73273361e-01 -3.56528074e-01
-2.05539405e-01 -5.11513293e-01 -3.16004932e-01 -1.59106719e+00
-1.89445332e-01 -7.11019695e-01 -2.62788773e-01 4.14058082e-02
-8.12023133e-03 3.40872183e-02 2.76473351e-02 -2.16064360e-02
-4.01986003e-01 5.06643355e-01 1.34610283e+00 -8.60833153e-02
-6.04100943e-01 2.18253568e-01 -2.91339159e-01 4.90353733e-01
9.00315464e-01 -8.63575697e-01 -3.90871346e-01 -1.87508270e-01
1.75778419e-01 1.73786998e-01 -3.38190734e-01 -1.21795356e+00
1.75146312e-01 -4.63941127e-01 1.95605919e-01 -1.01121414e+00
2.24562600e-01 -1.23581910e+00 6.21591032e-01 1.21583772e+00
7.43946582e-02 5.23909032e-01 3.99566412e-01 4.18596447e-01
-4.29416209e-01 -1.02143729e+00 9.90712285e-01 1.48220463e-02
-7.24787235e-01 -1.32561356e-01 -8.81606191e-02 -3.30700994e-01
1.43973374e+00 -1.00543785e+00 6.91512972e-02 8.78560320e-02
-5.75284302e-01 5.04929245e-01 2.85121888e-01 2.77668148e-01
3.76556426e-01 -1.20532656e+00 -2.36622632e-01 -7.52349198e-02
-4.17641073e-01 -4.49350417e-01 -6.01512454e-02 8.25908780e-01
-9.95360196e-01 8.07197630e-01 -4.29476917e-01 -5.54333627e-01
-1.64448655e+00 7.76437819e-01 4.00538504e-01 -7.08101273e-01
-1.03068195e-01 6.69576943e-01 -5.05178809e-01 -9.19896439e-02
1.10808790e-01 -5.73866516e-02 -3.97507578e-01 -1.80657908e-01
1.35277957e-01 1.02056730e+00 2.01518714e-01 -3.93093497e-01
-8.61102223e-01 1.18822360e+00 2.18048215e-01 8.42293575e-02
1.32919955e+00 -4.42224666e-02 -5.85676193e-01 -1.50331542e-01
1.30492449e+00 1.07212983e-01 -9.80709642e-02 4.25017178e-01
4.18672413e-01 -7.16972649e-01 2.82543153e-01 -9.82769012e-01
-1.08273661e+00 3.88435632e-01 8.40788662e-01 6.57783002e-02
1.54428327e+00 -5.52042067e-01 8.70377600e-01 4.86648351e-01
5.13618648e-01 -1.29822147e+00 -1.12698257e-01 4.47052836e-01
7.80173242e-01 -6.80598378e-01 4.86294001e-01 -6.13978028e-01
-1.04814239e-01 1.37676752e+00 6.52395129e-01 -1.55374661e-01
6.37834370e-01 2.65124202e-01 -4.06140745e-01 -1.80572808e-01
-6.16414309e-01 1.33095803e-02 4.02457178e-01 3.82383853e-01
1.50205478e-01 -4.69507992e-01 -1.24377489e+00 4.60039228e-01
-2.83922315e-01 -2.18253717e-01 4.45457082e-03 1.40040243e+00
-6.71948671e-01 -1.55590856e+00 -9.40001249e-01 3.32089186e-01
-5.04475474e-01 4.39924970e-02 -3.83438498e-01 8.51675153e-01
2.39703998e-01 1.12994683e+00 -2.57876664e-01 -2.34230295e-01
1.49392888e-01 -2.35897318e-01 5.67603350e-01 1.49988579e-02
-9.28667188e-01 -9.29468051e-02 1.11796632e-01 -6.71966612e-01
-2.03034818e-01 -4.64065671e-01 -1.34628868e+00 -2.02494800e-01
-7.28031516e-01 9.41273630e-01 9.56593275e-01 9.05943096e-01
2.10237473e-01 7.52550602e-01 7.59495676e-01 -5.57111561e-01
-3.00317883e-01 -2.88136274e-01 -1.47088900e-01 -3.32979679e-01
-4.22335006e-02 -1.03039443e+00 -1.34633213e-01 -5.32036960e-01] | [5.746366500854492, 3.5353596210479736] |
27f053a0-a6ae-4cbd-a367-eb2f5b022c1e | multi-temporal-scene-classification-and-scene | 2006.02176 | null | https://arxiv.org/abs/2006.02176v1 | https://arxiv.org/pdf/2006.02176v1.pdf | Multi-Temporal Scene Classification and Scene Change Detection with Correlation based Fusion | Classifying multi-temporal scene land-use categories and detecting their semantic scene-level changes for imagery covering urban regions could straightly reflect the land-use transitions. Existing methods for scene change detection rarely focus on the temporal correlation of bi-temporal features, and are mainly evaluated on small scale scene change detection datasets. In this work, we proposed a CorrFusion module that fuses the highly correlated components in bi-temporal feature embeddings. We firstly extracts the deep representations of the bi-temporal inputs with deep convolutional networks. Then the extracted features will be projected into a lower dimension space to computed the instance-level correlation. The cross-temporal fusion will be performed based on the computed correlation in CorrFusion module. The final scene classification are obtained with softmax activation layers. In the objective function, we introduced a new formulation for calculating the temporal correlation. The detailed derivation of backpropagation gradients for the proposed module is also given in this paper. Besides, we presented a much larger scale scene change detection dataset and conducted experiments on this dataset. The experimental results demonstrated that our proposed CorrFusion module could remarkably improve the multi-temporal scene classification and scene change detection results. | ['Bo Du', 'Lixiang Ru', 'Chen Wu'] | 2020-06-03 | null | null | null | null | ['scene-change-detection'] | ['computer-vision'] | [ 2.89538622e-01 -7.28648305e-01 2.04171106e-01 -8.00661564e-01
-4.35824990e-01 -1.40319273e-01 7.39572227e-01 3.06361429e-02
-7.20285833e-01 3.89982790e-01 1.47405773e-01 -8.87054950e-02
-2.19030946e-01 -1.11293948e+00 -6.01196289e-01 -8.80983531e-01
-4.81526107e-01 -6.03320837e-01 3.25983465e-01 -2.36909777e-01
-9.01444852e-02 4.90420431e-01 -1.66506350e+00 2.76037008e-01
7.83614397e-01 1.15834069e+00 4.65965748e-01 8.52953613e-01
-2.33180039e-02 7.47794211e-01 -6.08842410e-02 4.16373126e-02
2.73715109e-01 -2.15324685e-01 -4.96649712e-01 2.10520312e-01
5.77388167e-01 -3.91867250e-01 -6.84354246e-01 1.21330214e+00
2.96779364e-01 3.32423538e-01 2.80982345e-01 -1.21764040e+00
-5.44862211e-01 3.77887040e-01 -5.37808359e-01 7.00012088e-01
-8.05379376e-02 2.03102037e-01 7.70606577e-01 -1.01590598e+00
4.81658220e-01 1.29256797e+00 8.25298309e-01 -2.30890363e-01
-8.63462448e-01 -5.73967516e-01 6.81606948e-01 6.53856277e-01
-1.41230011e+00 -2.26162776e-01 9.20438170e-01 -6.25193477e-01
1.02392399e+00 4.58457559e-01 9.72653687e-01 5.14138639e-01
3.57623190e-01 8.96124005e-01 1.11825657e+00 -5.00297686e-03
-1.84822872e-01 -2.27527663e-01 3.99974972e-01 4.86052781e-01
9.23357531e-02 5.15144095e-02 1.25687838e-01 3.60143572e-01
4.37272221e-01 6.67057037e-01 -2.21749559e-01 -1.37031108e-01
-1.51764023e+00 7.53066838e-01 1.36642540e+00 8.35812926e-01
-3.67023647e-01 3.14859807e-01 4.36797649e-01 2.44659185e-01
6.99884117e-01 6.40463736e-03 -5.19213319e-01 2.18769968e-01
-8.62929821e-01 -9.43626240e-02 4.50753458e-02 6.67918801e-01
1.09472656e+00 -1.14171684e-01 -3.00802201e-01 8.20442677e-01
3.43602538e-01 5.55260777e-01 6.20573521e-01 -3.18200737e-01
5.11393845e-01 7.64382184e-01 -7.39945844e-02 -1.41608977e+00
-6.70231700e-01 -5.10009527e-01 -1.13253665e+00 -4.88736890e-02
-1.90585747e-01 -1.24613352e-01 -1.06747997e+00 1.58770359e+00
3.55789036e-01 4.89269197e-01 1.03067517e-01 9.44548666e-01
8.49529982e-01 9.97828841e-01 2.23303631e-01 -1.22392468e-01
1.29647815e+00 -9.05267894e-01 -8.98456216e-01 -9.52853113e-02
7.08461702e-01 -4.87646759e-01 8.04550469e-01 -5.09175599e-01
-5.26149631e-01 -9.43973601e-01 -1.15247869e+00 -2.03790426e-01
-9.22604501e-01 4.47280645e-01 8.02903473e-01 -4.38499264e-02
-9.80672717e-01 6.07106984e-01 -1.05388689e+00 -8.11697900e-01
6.99318707e-01 6.29833564e-02 -4.77287650e-01 -8.04395825e-02
-1.37257171e+00 8.21435213e-01 6.31708741e-01 8.23740542e-01
-5.03358126e-01 -4.78445858e-01 -1.15137100e+00 1.73859373e-02
-1.75680101e-01 -4.77673024e-01 9.11187172e-01 -1.04319251e+00
-1.12692702e+00 8.74455929e-01 -1.92291439e-01 -3.94942671e-01
4.24829453e-01 -1.93622380e-01 -9.41357136e-01 -1.30727574e-01
4.04783130e-01 7.53526151e-01 5.23852706e-01 -9.46674764e-01
-1.13189852e+00 -3.14047545e-01 -5.54372258e-02 5.09739876e-01
-3.63895506e-01 -1.05347499e-01 -3.69235873e-01 -5.65181553e-01
5.36880076e-01 -7.23928034e-01 -4.09783721e-01 6.84658885e-02
3.03360932e-02 -5.58063574e-02 1.14884293e+00 -6.19478226e-01
1.29392397e+00 -2.44707751e+00 -1.97888762e-01 -1.31380677e-01
-3.32378969e-02 1.69993892e-01 -1.55563056e-01 3.05994570e-01
-2.74767309e-01 2.30181171e-03 -6.19875431e-01 -6.90893903e-02
-1.67202830e-01 6.38980493e-02 -3.30284595e-01 7.98122227e-01
4.17557299e-01 1.02768052e+00 -1.02313340e+00 -3.53970617e-01
7.31082499e-01 4.85971659e-01 -2.53530204e-01 1.07393697e-01
1.79437786e-01 3.48759033e-02 -2.90130496e-01 3.01821113e-01
1.17787373e+00 8.17898661e-02 -1.71634734e-01 -5.28034329e-01
-5.20571291e-01 -1.81565024e-02 -1.00730515e+00 1.71562374e+00
-4.44087446e-01 1.11141467e+00 -4.59482223e-01 -1.01456857e+00
9.40858901e-01 -1.93375364e-01 7.54997075e-01 -7.82728553e-01
1.42567575e-01 3.53790820e-02 -4.17018048e-02 -7.97401249e-01
6.64869428e-01 -5.07807471e-02 -1.74471840e-01 -2.91667134e-01
-1.57810435e-01 -1.79930091e-01 2.00535972e-02 -2.82447845e-01
8.77052486e-01 1.31040318e-02 2.47594580e-01 -1.70815632e-01
7.33155608e-01 1.43627211e-01 5.11941433e-01 3.46670508e-01
-5.43941200e-01 2.88026035e-01 -2.09447425e-02 -7.83710361e-01
-6.34972632e-01 -9.14942980e-01 -4.14856255e-01 7.97892988e-01
5.64877212e-01 -1.36523575e-01 -9.95258465e-02 -6.33818984e-01
1.59924194e-01 6.02938175e-01 -7.04491377e-01 -4.52121466e-01
-5.92398822e-01 -1.25428128e+00 3.13610286e-01 4.96346802e-01
1.28005040e+00 -9.06492233e-01 -5.87245226e-01 3.05364609e-01
-1.19398624e-01 -1.17140710e+00 -2.92540580e-01 2.69733310e-01
-8.35947573e-01 -9.41670060e-01 -4.99554873e-01 -1.15425730e+00
5.64696372e-01 8.17690194e-01 3.66131306e-01 -4.28215027e-01
-3.09437573e-01 1.80874206e-03 -2.67511696e-01 -9.53500271e-02
2.16226116e-01 -1.85892954e-01 -1.68060109e-01 3.81924450e-01
5.94372332e-01 -7.72161722e-01 -6.91810250e-01 8.14206824e-02
-9.14493263e-01 2.57824808e-01 5.97093165e-01 6.91013992e-01
4.78687555e-01 5.00685871e-01 6.65574968e-02 -4.71179932e-01
2.87394524e-01 -4.16574746e-01 -5.53028405e-01 1.05603449e-01
-2.44622976e-01 -8.46111998e-02 2.85656273e-01 -2.87866712e-01
-1.28647375e+00 1.38409391e-01 -2.67723054e-01 -2.83314288e-01
-2.91739881e-01 8.59280944e-01 -1.39125094e-01 2.16835544e-01
3.83567512e-01 4.23169643e-01 -7.94790268e-01 -3.83282632e-01
5.39898813e-01 6.94402397e-01 6.16940439e-01 1.64068595e-01
9.30265248e-01 7.07727373e-01 -3.90864164e-01 -7.43034542e-01
-7.13507533e-01 -7.80013919e-01 -8.89740765e-01 -2.50955790e-01
1.04361451e+00 -1.38622391e+00 -3.38137448e-01 1.00057685e+00
-1.08416045e+00 -2.40312383e-01 -2.70414501e-01 9.05467808e-01
-3.96009356e-01 2.11337566e-01 -3.49156260e-01 -5.33983409e-01
-1.24643110e-01 -7.30065346e-01 1.22145271e+00 3.71238559e-01
3.89510691e-01 -8.26543450e-01 2.93186605e-01 -2.09678382e-01
5.05865514e-01 4.68034685e-01 6.12641692e-01 4.74067181e-02
-4.43242639e-01 -2.94312716e-01 -6.45155907e-01 1.34048119e-01
5.75674713e-01 1.34823933e-01 -1.13126552e+00 -1.93931356e-01
8.81876871e-02 2.38998294e-01 1.51770842e+00 5.98198295e-01
1.22517252e+00 -1.69209167e-01 -5.26462138e-01 1.10450315e+00
1.83964503e+00 3.28049481e-01 6.30715966e-01 5.85869551e-01
1.07973015e+00 2.82682121e-01 7.95108557e-01 3.72837037e-01
5.26456594e-01 6.37549400e-01 3.83165419e-01 -4.54726785e-01
-1.82837263e-01 -2.05285847e-01 5.63159585e-01 9.13518429e-01
9.48211625e-02 6.28269911e-02 -8.05576026e-01 8.06803346e-01
-2.18381071e+00 -1.33172953e+00 -4.05484170e-01 1.70814872e+00
5.82062483e-01 1.56304747e-01 -3.20299655e-01 -3.54146138e-02
8.92685115e-01 5.84221959e-01 -6.26395166e-01 -7.40344003e-02
-4.64242667e-01 -2.17453346e-01 7.85881221e-01 4.79172558e-01
-1.97437823e+00 1.18748236e+00 5.56196976e+00 4.40449536e-01
-1.55276430e+00 2.18080148e-01 2.41967812e-01 5.66411316e-02
-7.64519051e-02 -1.33287251e-01 -5.01218259e-01 3.93767625e-01
6.38323843e-01 -1.76072046e-01 -8.96474123e-02 7.50828981e-01
4.18206722e-01 -8.72644708e-02 -7.07606375e-01 1.18985343e+00
-1.72627985e-01 -1.03322268e+00 1.03744045e-02 -2.89726615e-01
8.82268429e-01 5.88161349e-01 -8.05253386e-02 5.31181157e-01
4.07830179e-01 -6.40571833e-01 6.12581134e-01 8.04108202e-01
6.60361946e-01 -4.32884216e-01 8.48227620e-01 9.86490324e-02
-1.67867351e+00 -3.03195655e-01 -4.52677131e-01 -1.95915848e-01
5.84286451e-02 7.03995228e-01 -5.08894563e-01 7.36044407e-01
8.78923118e-01 1.85043609e+00 -8.83557737e-01 1.17768049e+00
-1.49761140e-02 4.75977540e-01 -3.30899924e-01 1.78399414e-01
4.94819522e-01 -1.52876541e-01 2.75684059e-01 1.67823780e+00
4.18281168e-01 1.47776948e-02 2.42660135e-01 6.53751314e-01
1.62842572e-01 -1.19957224e-01 -6.69808388e-01 7.38876536e-02
7.57207423e-02 1.52153063e+00 -5.81751943e-01 -4.59323734e-01
-4.28447604e-01 1.27078462e+00 1.73656121e-01 4.85485584e-01
-9.60940778e-01 -5.90766847e-01 1.00807691e+00 -3.31874222e-01
4.40600276e-01 -4.80164200e-01 -2.45063484e-01 -1.56786001e+00
2.13740528e-01 -6.02475069e-02 4.52830493e-01 -8.23398888e-01
-1.25156736e+00 5.86530805e-01 5.44721819e-02 -1.57071900e+00
3.28342944e-01 -4.38128561e-01 -8.12903464e-01 9.04313087e-01
-1.99485278e+00 -1.33353674e+00 -7.84116626e-01 7.72009015e-01
6.05256677e-01 2.48699307e-01 6.52414858e-01 4.80257154e-01
-9.41813469e-01 1.61255598e-01 3.32552999e-01 2.81386882e-01
4.22422051e-01 -1.05363369e+00 5.56143284e-01 1.44633627e+00
-1.74670830e-01 1.07775815e-01 2.73621410e-01 -4.67492431e-01
-1.00106251e+00 -1.71914244e+00 9.27994847e-01 2.02130288e-01
8.03968489e-01 -1.58585027e-01 -8.76362443e-01 7.05774248e-01
2.26649996e-02 3.07005793e-01 2.23134562e-01 -7.99859911e-02
-1.80976704e-01 -5.91958940e-01 -1.01247263e+00 4.82711732e-01
1.32165349e+00 -8.14835489e-01 -5.95537305e-01 2.89177716e-01
9.28220570e-01 -1.86862051e-01 -7.75412142e-01 6.47447228e-01
2.83323437e-01 -5.17873228e-01 7.43732631e-01 -5.08923769e-01
3.46923053e-01 -8.24599564e-01 -6.06999636e-01 -1.33488894e+00
-1.03572011e+00 1.89711899e-01 3.36256683e-01 1.03177750e+00
-1.24742545e-01 -6.01170361e-01 1.95224717e-01 1.08844936e-01
-4.57266480e-01 -3.84677291e-01 -8.50632548e-01 -5.90973794e-01
-2.06242457e-01 -6.71672106e-01 6.53425515e-01 1.10057986e+00
-2.88033336e-01 2.82955915e-01 -1.71219498e-01 6.43899739e-01
1.80912822e-01 3.90933633e-01 5.92884183e-01 -1.05020237e+00
3.15466553e-01 -6.48608088e-01 -9.47514653e-01 -9.93987381e-01
-4.43447009e-02 -1.25326991e+00 1.97985843e-01 -1.74059427e+00
2.36370370e-01 -2.48675987e-01 -7.09599555e-01 4.92585778e-01
-5.96885562e-01 1.04543611e-01 7.31930137e-02 3.23379815e-01
-4.43353981e-01 1.04499292e+00 1.12840569e+00 -5.79662919e-01
-2.28426382e-01 -1.99946374e-01 -2.81238467e-01 5.38601875e-01
8.21072519e-01 -3.00117910e-01 -2.08730146e-01 -5.74856997e-01
-1.25360742e-01 -4.97456759e-01 6.72907293e-01 -1.48392892e+00
1.56434804e-01 -3.65405977e-01 6.16812110e-01 -9.91976023e-01
1.99290663e-01 -1.17162418e+00 2.53789395e-01 6.16293311e-01
-2.33685538e-01 -2.02531174e-01 4.62947875e-01 5.36745071e-01
-5.61311722e-01 3.71784657e-01 7.42397845e-01 1.03003718e-01
-1.45190048e+00 5.90532959e-01 -4.35537338e-01 -5.94842374e-01
1.19648933e+00 -2.00402156e-01 -1.60375789e-01 -2.91066710e-02
-6.37616098e-01 3.76013458e-01 8.97955000e-02 8.97719443e-01
5.79142570e-01 -1.81960142e+00 -7.24327326e-01 3.58138829e-01
5.03319502e-01 -1.68138653e-01 5.01437843e-01 8.16039383e-01
-4.86370116e-01 3.80156219e-01 -5.28812408e-01 -8.96566451e-01
-1.17503631e+00 6.30519927e-01 8.67361128e-01 -8.05650428e-02
-6.20864153e-01 6.84559822e-01 2.44635344e-01 -6.27836108e-01
-2.79150188e-01 -9.57635522e-01 -3.74911666e-01 3.64027590e-01
2.94720113e-01 5.07392772e-02 -3.38949710e-02 -9.68975365e-01
-6.71476960e-01 7.37803578e-01 1.56106576e-01 4.12110910e-02
1.76281583e+00 -3.64807516e-01 -5.99718243e-02 7.11525559e-01
1.81756020e+00 -6.29311502e-01 -1.29535854e+00 -6.48902833e-01
-1.15881547e-01 -6.18752420e-01 4.49617833e-01 -4.84659165e-01
-1.26452231e+00 8.93665850e-01 1.28001261e+00 1.17674701e-01
1.47357559e+00 -3.12041700e-01 7.89868116e-01 4.89896834e-01
1.39882073e-01 -9.41823006e-01 -4.12187517e-01 6.72031522e-01
9.28504586e-01 -1.52675641e+00 -5.79200797e-02 -2.48902470e-01
-3.76376957e-01 1.20266593e+00 6.03823304e-01 -2.34344214e-01
1.20024943e+00 -1.25928596e-01 -5.61475940e-02 -4.27925080e-01
-4.03612792e-01 -8.02471220e-01 2.82211691e-01 2.60228992e-01
2.22316295e-01 5.27948737e-01 -2.18845695e-01 3.12245131e-01
-8.02641809e-02 1.25894323e-01 1.50493965e-01 8.50698054e-01
-5.56386530e-01 -5.56176245e-01 -8.69641751e-02 4.11671758e-01
2.53888041e-01 -1.17755741e-01 -4.29001302e-02 5.60047865e-01
7.37211287e-01 9.52575684e-01 4.21728104e-01 -7.01012909e-01
6.00150287e-01 -2.09221572e-01 1.05474040e-01 -3.38167846e-01
-3.88393492e-01 -6.44599050e-02 -2.28262499e-01 -6.11316979e-01
-7.47476101e-01 -9.56025481e-01 -1.12426519e+00 -1.99533775e-01
-2.06166372e-01 -3.23107153e-01 6.29996657e-01 7.33966410e-01
2.72057623e-01 7.36298144e-01 1.04698896e+00 -9.20727253e-01
1.19316362e-01 -1.22929466e+00 -5.71115673e-01 7.24878252e-01
6.34594440e-01 -5.23246825e-01 -2.46722311e-01 2.18291476e-01] | [9.697660446166992, -1.3029162883758545] |
f1644e74-deb3-4bc2-a03b-3fdd8a465c74 | ultrastereo-efficient-learning-based-matching | null | null | http://openaccess.thecvf.com/content_cvpr_2017/html/Fanello_UltraStereo_Efficient_Learning-Based_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Fanello_UltraStereo_Efficient_Learning-Based_CVPR_2017_paper.pdf | UltraStereo: Efficient Learning-Based Matching for Active Stereo Systems | Efficient estimation of depth from pairs of stereo images is one of the core problems in computer vision. We efficiently solve the specialized problem of stereo matching under active illumination using a new learning-based algorithm. This type of 'active' stereo i.e. stereo matching where scene texture is augmented by an active light projector is proving compelling for designing depth cameras, largely due to improved robustness when compared to time of flight or traditional structured light techniques. Our algorithm uses an unsupervised greedy optimization scheme that learns features that are discriminative for estimating correspondences in infrared images. The proposed method optimizes a series of sparse hyperplanes that are used at test time to remap all the image patches into a compact binary representation in O(1). The proposed algorithm is cast in a PatchMatch Stereo-like framework, producing depth maps at 500Hz. In contrast to standard structured light methods, our approach generalizes to different scenes, does not require tedious per camera calibration procedures and is not adversely affected by interference from overlapping sensors. Extensive evaluations show we surpass the quality and overcome the limitations of current depth sensing technologies.
| ['Sean Ryan Fanello', 'Julien Valentin', 'Philip Davidson', 'Vladimir Tankovich', 'Shahram Izadi', 'Christoph Rhemann', 'Adarsh Kowdle'] | 2017-07-01 | null | null | null | cvpr-2017-7 | ['stereo-matching'] | ['computer-vision'] | [ 8.12301695e-01 6.44635707e-02 4.78587952e-03 -6.15093887e-01
-7.50012696e-01 -4.64011341e-01 5.70321560e-01 -1.71324238e-01
-6.01253331e-01 6.98217750e-01 5.72841801e-02 -6.55367374e-02
-2.17478380e-01 -6.58717573e-01 -7.26671755e-01 -1.09815574e+00
5.02757668e-01 6.16394937e-01 3.42004150e-01 -3.64050530e-02
8.85179639e-01 4.65880990e-01 -1.99522948e+00 1.80290952e-01
7.53526628e-01 1.00117385e+00 6.71737731e-01 5.77107906e-01
-1.13343596e-01 8.19073319e-01 6.01480715e-02 1.37500139e-02
7.81219780e-01 4.78362739e-02 -7.56877363e-01 4.27852452e-01
1.03371918e+00 -5.48338652e-01 -2.60072947e-01 1.04616499e+00
3.90988946e-01 -1.76621526e-02 2.77399957e-01 -8.65074754e-01
-1.52884889e-02 -3.39993715e-01 -6.09382033e-01 8.54143053e-02
9.08002257e-01 -3.27512138e-02 7.92398989e-01 -9.42691147e-01
6.01156473e-01 1.00955021e+00 5.67332149e-01 5.06791353e-01
-1.43481910e+00 -3.39089960e-01 -2.80584782e-01 2.87478596e-01
-8.95732343e-01 -8.51298869e-01 8.47572505e-01 -5.59276879e-01
9.76350605e-01 3.57383549e-01 7.76708663e-01 8.38637948e-01
1.31411016e-01 5.58583915e-01 1.40895212e+00 -8.50315452e-01
4.59240854e-01 1.83088481e-01 1.78828984e-01 7.54495859e-01
3.11283231e-01 4.36303675e-01 -1.10205054e+00 -3.21069926e-01
8.73664379e-01 2.84454614e-01 -5.92090607e-01 -9.79634821e-01
-1.16725731e+00 7.94899523e-01 1.52259544e-01 -2.51247972e-01
-6.54166862e-02 -1.16252989e-01 -1.72685862e-01 2.09834501e-01
5.05224109e-01 3.44360977e-01 -1.44226417e-01 2.50059038e-01
-8.05082977e-01 1.53098637e-02 6.22066736e-01 8.19328547e-01
1.45302093e+00 -3.03431243e-01 5.81769168e-01 8.66243243e-01
4.67990637e-01 5.07420957e-01 1.91987589e-01 -1.48409271e+00
5.34980297e-01 6.98492110e-01 1.00801960e-01 -7.62988567e-01
-1.60086140e-01 9.36565921e-02 -5.58853865e-01 9.09340382e-01
3.24409246e-01 2.02809840e-01 -5.32001138e-01 9.22511995e-01
4.51892346e-01 1.34756610e-01 -1.06392853e-01 8.98829401e-01
5.40585160e-01 5.16145468e-01 -8.47871959e-01 -2.87708700e-01
5.34105182e-01 -8.39094877e-01 -1.90181404e-01 -5.52495420e-01
2.34748334e-01 -1.00349212e+00 7.31737435e-01 7.52391219e-01
-1.01199257e+00 -5.22673190e-01 -1.22518194e+00 -1.95161358e-01
-9.11740493e-03 -4.69482839e-01 7.92272866e-01 7.45175064e-01
-1.23563242e+00 6.20663285e-01 -5.47520936e-01 -4.11021173e-01
1.12231284e-01 7.19920218e-01 -7.75813043e-01 -3.57611865e-01
-3.47343862e-01 6.20223224e-01 2.87612945e-01 -1.52137484e-02
-6.05730414e-01 -6.52266145e-01 -9.73114789e-01 -4.82385099e-01
2.12422490e-01 -7.78119683e-01 9.23501253e-01 -1.06698310e+00
-1.64809346e+00 1.40212226e+00 -5.30213177e-01 -1.79025471e-01
2.90632844e-01 -2.97218710e-01 1.48237467e-01 2.77403474e-01
7.16498196e-02 6.66262388e-01 9.79012311e-01 -1.30286336e+00
-7.58671820e-01 -7.41506398e-01 2.09766820e-01 5.42782843e-01
-1.16348386e-01 -2.14153945e-01 -4.17800903e-01 2.24829372e-03
8.28310430e-01 -9.21432376e-01 -4.42325085e-01 3.54666919e-01
-2.16945007e-01 2.78603256e-01 7.76122212e-01 -1.57613620e-01
3.14409047e-01 -1.87534344e+00 2.71605879e-01 2.47298241e-01
3.50794584e-01 -2.04747915e-01 1.71534374e-01 3.94736290e-01
-7.88983926e-02 -8.77434790e-01 -2.47930706e-01 -6.38905108e-01
-4.35314685e-01 2.85375834e-01 -2.59244174e-01 9.25919473e-01
-2.74206996e-01 2.63231814e-01 -8.75915527e-01 -3.31907272e-01
7.03707874e-01 3.23442012e-01 -6.34326100e-01 4.39499050e-01
1.41010165e-01 6.97088122e-01 3.92933674e-02 6.67263389e-01
8.53755891e-01 -1.05901554e-01 -4.60775048e-02 -1.29730701e-01
-6.35829568e-01 2.74653405e-01 -1.46145225e+00 2.13769746e+00
-4.51574683e-01 7.64658630e-01 1.16258197e-01 -1.16079295e+00
1.18829846e+00 1.57293186e-01 6.87917173e-01 -1.00589967e+00
-2.32916653e-01 3.84654731e-01 -7.04449713e-01 -4.01561886e-01
4.09941077e-01 -5.65977916e-02 3.82034034e-01 4.21643227e-01
2.65313429e-03 -3.98043901e-01 -2.16276973e-01 -2.22143769e-01
1.11035478e+00 2.57700086e-01 4.47318316e-01 -4.81328011e-01
5.84775507e-01 5.39416112e-02 5.59115529e-01 6.97013199e-01
1.22169428e-01 8.96706879e-01 -2.52675503e-01 -6.53613746e-01
-1.16672790e+00 -9.95505929e-01 -4.35738325e-01 5.41155517e-01
5.31154931e-01 -2.18725111e-02 -4.91173834e-01 -1.88244611e-01
-5.49425855e-02 1.03426158e-01 -2.81662643e-01 2.45398402e-01
-5.60925901e-01 -4.75982040e-01 -2.40317672e-01 1.13549009e-01
5.86257637e-01 -6.51033103e-01 -1.06104243e+00 4.02427353e-02
-6.82070553e-02 -1.02783287e+00 -1.24102823e-01 4.12521631e-01
-1.33346677e+00 -1.27933991e+00 -6.92855656e-01 -8.94630492e-01
7.55527139e-01 1.02813721e+00 1.09422994e+00 -2.23589584e-01
-4.93504554e-01 9.28506970e-01 -1.81330502e-01 -1.81778625e-01
2.41229787e-01 -4.34162617e-01 8.98594037e-02 2.69924641e-01
4.76792842e-01 -7.43863881e-01 -7.58146822e-01 3.26080114e-01
-5.71285069e-01 2.88426399e-01 3.61557275e-01 9.28593040e-01
6.68345988e-01 -2.22046688e-01 -5.83841443e-01 -9.92427111e-01
-2.67826855e-01 -3.79758282e-03 -1.26942813e+00 -1.30939990e-01
-6.23372614e-01 5.43406568e-02 3.28434110e-01 7.43460506e-02
-1.25253224e+00 7.56465375e-01 2.75508642e-01 -2.45786682e-01
-2.47813880e-01 -2.02603161e-01 -1.43153206e-01 -9.15867984e-01
7.58594096e-01 2.26287276e-01 -4.70120348e-02 -4.10001516e-01
-1.38340980e-01 6.79313540e-01 6.61122918e-01 -3.71582419e-01
1.01927578e+00 1.15749931e+00 3.49289834e-01 -1.07071865e+00
-8.86789262e-01 -1.25784683e+00 -9.63341355e-01 -3.72950494e-01
6.21547043e-01 -1.05577779e+00 -7.57015705e-01 4.90653276e-01
-1.13278818e+00 -2.19960168e-01 -2.53580213e-01 6.74118459e-01
-9.90103483e-01 7.15280950e-01 -2.17225820e-01 -8.33087504e-01
-1.23586752e-01 -9.81475115e-01 1.44991803e+00 1.93666518e-01
4.35390472e-02 -1.10085154e+00 6.61449432e-01 8.64072204e-01
7.24653304e-02 3.01139932e-02 4.54374075e-01 4.16380823e-01
-1.23065770e+00 2.59628869e-03 -1.75526783e-01 2.61956990e-01
9.45149586e-02 -3.52421433e-01 -1.48825073e+00 -4.19835031e-01
6.17090404e-01 -3.65209788e-01 8.38087380e-01 3.97233784e-01
9.11124945e-01 -5.70920445e-02 -1.88572392e-01 1.09875917e+00
2.09574127e+00 3.40070397e-01 9.81808126e-01 7.55402803e-01
7.01900303e-01 9.22158837e-01 5.80478728e-01 4.85209942e-01
7.74662942e-02 9.68581378e-01 6.08438909e-01 -3.22939396e-01
2.64621764e-01 8.56765136e-02 3.41878355e-01 5.60149252e-01
-1.89121559e-01 2.73435295e-01 -1.03093886e+00 2.75807440e-01
-1.65810013e+00 -1.13820481e+00 -3.35800588e-01 2.81183314e+00
5.50327301e-01 5.30615151e-02 -2.07319409e-01 6.00802064e-01
5.26525855e-01 9.00339261e-02 -4.53210831e-01 -1.87647924e-01
-2.92453468e-01 4.05548066e-01 8.14236999e-01 9.27880883e-01
-1.07199740e+00 4.61810589e-01 6.36416340e+00 1.22727200e-01
-1.01515377e+00 1.53534124e-02 2.01827258e-01 -8.90595913e-02
-3.15841407e-01 4.57132399e-01 -7.99781144e-01 2.11510539e-01
2.37386346e-01 1.64999709e-01 5.46936154e-01 7.23436296e-01
1.47379637e-01 -6.46626770e-01 -1.38804913e+00 1.64675725e+00
3.96493495e-01 -1.29827821e+00 -2.41734281e-01 2.93497711e-01
8.58885109e-01 1.64622605e-01 3.00608873e-02 -6.33729160e-01
-2.36965679e-02 -6.99724555e-01 5.59146643e-01 5.08481741e-01
6.63269877e-01 -2.58947521e-01 3.67670178e-01 4.61871684e-01
-8.17211330e-01 -1.71463847e-01 -7.27720916e-01 -3.65953743e-01
7.76504204e-02 7.73164928e-01 -6.55289292e-01 2.87967026e-01
8.45633805e-01 9.20533180e-01 -3.33908081e-01 1.32486331e+00
2.52877593e-01 -1.18081290e-02 -3.61559629e-01 4.37207937e-01
-3.57789099e-02 -6.36714339e-01 6.34840965e-01 6.76090002e-01
3.07232529e-01 -2.81802211e-02 5.24177626e-02 5.07570505e-01
4.96341079e-01 -2.75474433e-02 -1.07329476e+00 6.70043647e-01
1.34271353e-01 9.38946247e-01 -4.56875533e-01 -9.19741541e-02
-7.47075200e-01 1.31635821e+00 1.64525449e-01 2.02813402e-01
-1.87072933e-01 -7.95389563e-02 2.76493877e-01 3.35043967e-01
2.10484490e-02 -2.51141340e-01 -4.29383069e-01 -1.43426025e+00
1.83796674e-01 -6.16326213e-01 7.46820644e-02 -9.34004009e-01
-1.09296298e+00 2.36948222e-01 -1.45813495e-01 -1.54878771e+00
-1.54561654e-01 -8.38313699e-01 -3.29347253e-01 9.09694314e-01
-1.64818621e+00 -6.94323599e-01 -9.11176980e-01 1.08539867e+00
7.35675931e-01 -2.54111230e-01 9.10353303e-01 2.61016488e-01
-1.48541659e-01 8.67766440e-02 4.08793718e-01 -3.74458820e-01
6.97227836e-01 -1.25771308e+00 1.20633520e-01 8.53910267e-01
2.22957462e-01 3.67979914e-01 6.84772015e-01 -2.78696895e-01
-1.62223005e+00 -5.24021089e-01 8.91743779e-01 -4.21074301e-01
2.48517692e-01 -4.89557832e-01 -4.74869698e-01 4.74113703e-01
1.01040691e-01 -9.54855978e-02 6.23945296e-01 8.06982890e-02
-4.27214712e-01 -5.43364167e-01 -1.05748904e+00 1.65429682e-01
1.22766721e+00 -8.44004929e-01 -6.24343932e-01 5.13677597e-01
7.55447969e-02 -4.01880682e-01 -2.68623978e-01 2.79245347e-01
6.29744709e-01 -1.90896320e+00 1.10463369e+00 1.78903975e-02
3.50186646e-01 -2.34200761e-01 -5.48198223e-01 -7.51424432e-01
-3.20147544e-01 -6.68889403e-01 2.79997826e-01 7.31548727e-01
-1.97538137e-02 -8.28371763e-01 1.33483398e+00 4.66208935e-01
-2.47557536e-01 -3.21360856e-01 -1.04117739e+00 -6.34739101e-01
-7.53438115e-01 -1.24751821e-01 -1.00561798e-01 8.68461668e-01
-7.92992041e-02 3.03003013e-01 -3.09328437e-01 2.41349697e-01
1.48505580e+00 3.03831846e-01 9.13205862e-01 -1.58870232e+00
-7.83685386e-01 1.29889920e-01 -8.79910707e-01 -1.32972527e+00
2.38463450e-02 -4.74210471e-01 6.44379854e-02 -1.21175468e+00
4.87120956e-01 -5.93699157e-01 1.34299695e-01 1.32640958e-01
2.60214418e-01 4.83761728e-01 -1.04419969e-01 4.56649154e-01
-3.00895274e-01 2.35769466e-01 8.07709992e-01 -1.96932897e-01
-1.90066863e-02 1.62143521e-02 -1.22787848e-01 7.58695900e-01
4.50839430e-01 -3.30779344e-01 -4.96517986e-01 -6.65941119e-01
4.23586667e-01 3.64626795e-01 5.19343078e-01 -1.26215982e+00
6.70439243e-01 -2.04940692e-01 1.80845171e-01 -5.47539771e-01
8.37817729e-01 -1.05320370e+00 3.98549765e-01 2.40940303e-01
-6.20463453e-02 -1.99199975e-01 -2.87246525e-01 5.48887432e-01
-3.78913015e-01 -3.95176917e-01 1.11204910e+00 -3.79469216e-01
-9.82736707e-01 1.66171521e-01 -3.50060798e-02 -3.77805382e-01
8.20948422e-01 -1.02792728e+00 -1.44707188e-01 -1.47492573e-01
-2.85646766e-01 -3.31947714e-01 1.05589855e+00 -6.59466535e-03
7.93063760e-01 -1.09656596e+00 -3.26217860e-01 5.97449541e-01
3.78877997e-01 2.72866547e-01 1.81259468e-01 7.29978144e-01
-9.88582969e-01 6.39998496e-01 -4.50885266e-01 -1.22779274e+00
-1.54614639e+00 3.17323714e-01 3.78195614e-01 2.86808848e-01
-8.61303389e-01 9.44658041e-01 5.23307443e-01 -3.71790677e-01
3.64955515e-01 2.76816398e-01 1.36080217e-02 -2.05180556e-01
5.44127345e-01 5.40988684e-01 4.02623713e-01 -5.39271235e-01
-1.37024432e-01 1.16217268e+00 1.12884857e-01 -1.87114716e-01
1.42545223e+00 -3.39629859e-01 -3.34124535e-01 6.17878079e-01
1.23757672e+00 1.10088080e-01 -1.41657352e+00 -2.86529779e-01
-1.49068221e-01 -1.04131556e+00 4.18299943e-01 -1.31843865e-01
-9.02672648e-01 9.02100444e-01 9.32449520e-01 -1.80888608e-01
1.24812281e+00 -2.07357317e-01 4.81396407e-01 6.75362706e-01
1.01570368e+00 -1.04273438e+00 7.39238858e-02 3.59268099e-01
4.99430180e-01 -1.51027238e+00 1.22732982e-01 -6.50427818e-01
1.74649768e-02 1.36573255e+00 3.56754214e-01 -2.44268909e-01
3.79442334e-01 3.14764649e-01 2.38220975e-01 -2.65097708e-01
-5.91236115e-01 -1.61367387e-01 7.04556927e-02 7.33580709e-01
1.95096940e-01 -4.31689978e-01 1.61644131e-01 -8.37518334e-01
3.60851064e-02 -9.17923674e-02 5.23113608e-01 1.01825154e+00
-7.14967668e-01 -1.31595647e+00 -7.54044890e-01 1.73233449e-01
1.63310483e-01 4.26224209e-02 -2.77369678e-01 2.69289643e-01
1.90969855e-01 9.07377005e-01 1.86451972e-01 -1.75517336e-01
1.95021033e-01 -3.10826391e-01 9.85010028e-01 -8.18571866e-01
-8.10709149e-02 -5.58178648e-02 -1.88286006e-01 -1.02169001e+00
-9.22287703e-01 -9.26212549e-01 -5.53675711e-01 -1.62126034e-01
-3.08425158e-01 -1.40939608e-01 6.11807406e-01 8.36031139e-01
-7.99011886e-02 -4.80210841e-01 1.10351098e+00 -1.23385561e+00
-2.14150414e-01 -2.87930965e-01 -7.15876400e-01 3.55444342e-01
5.20284116e-01 -7.04634309e-01 -6.77576423e-01 1.22703612e-01] | [9.050328254699707, -2.525481700897217] |
636cbf95-2943-4a8d-a49b-fedf54fc00b9 | language-model-pre-training-for-hierarchical | 1901.09128 | null | http://arxiv.org/abs/1901.09128v1 | http://arxiv.org/pdf/1901.09128v1.pdf | Language Model Pre-training for Hierarchical Document Representations | Hierarchical neural architectures are often used to capture long-distance
dependencies and have been applied to many document-level tasks such as
summarization, document segmentation, and sentiment analysis. However,
effective usage of such a large context can be difficult to learn, especially
in the case where there is limited labeled data available. Building on the
recent success of language model pretraining methods for learning flat
representations of text, we propose algorithms for pre-training hierarchical
document representations from unlabeled data. Unlike prior work, which has
focused on pre-training contextual token representations or context-independent
{sentence/paragraph} representations, our hierarchical document representations
include fixed-length sentence/paragraph representations which integrate
contextual information from the entire documents. Experiments on document
segmentation, document-level question answering, and extractive document
summarization demonstrate the effectiveness of the proposed pre-training
algorithms. | ['Kristina Toutanova', 'Ming-Wei Chang', 'Kenton Lee', 'Jacob Devlin'] | 2019-01-26 | language-model-pre-training-for-hierarchical-1 | https://openreview.net/forum?id=rygnfn0qF7 | https://openreview.net/pdf?id=rygnfn0qF7 | iclr-2019-5 | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 6.07541263e-01 1.93900615e-01 -4.79473501e-01 -5.81708133e-01
-9.71078336e-01 -5.66029966e-01 4.75188494e-01 8.96998048e-01
-6.13461912e-01 6.81337714e-01 8.13910842e-01 -5.99104583e-01
1.97750956e-01 -6.03049457e-01 -5.26338220e-01 -3.22319269e-01
1.92865089e-01 2.18099177e-01 -5.74378222e-02 3.28035988e-02
6.05768085e-01 1.51976928e-01 -1.10554779e+00 5.72975397e-01
9.81032789e-01 5.86957753e-01 2.08401963e-01 1.00415945e+00
-7.72016287e-01 7.28340387e-01 -9.64017272e-01 -1.06268629e-01
-4.06815171e-01 -4.99700904e-01 -1.16205108e+00 2.34831631e-01
4.17255461e-01 -3.45119089e-01 3.44116241e-02 7.82974362e-01
3.00611138e-01 4.01074648e-01 8.60775292e-01 -2.54916281e-01
-8.00103903e-01 1.06012416e+00 -6.91435874e-01 3.91712338e-01
1.61897942e-01 -5.28945565e-01 1.41448474e+00 -7.52054155e-01
4.01281536e-01 1.24317908e+00 3.71688753e-01 6.10534668e-01
-9.11278188e-01 -1.52752250e-01 6.69406533e-01 -1.87058941e-01
-8.52183521e-01 -1.86338544e-01 6.94354832e-01 -3.10005486e-01
1.40017152e+00 8.57927203e-02 2.04431936e-01 1.10999894e+00
1.01119541e-01 1.42490113e+00 4.46440011e-01 -8.31045389e-01
1.30655199e-01 1.02811446e-03 9.23439324e-01 2.91285515e-01
3.68081778e-01 -7.72764862e-01 -1.53905630e-01 -3.07290461e-02
4.27126557e-01 6.79297224e-02 -7.81128928e-02 2.22167432e-01
-8.14873397e-01 1.07976782e+00 3.14255327e-01 7.64956713e-01
-3.68775815e-01 1.31452098e-01 9.15348232e-01 -2.59436388e-03
6.84480190e-01 5.40865719e-01 -5.08355379e-01 -2.54686847e-02
-1.27347302e+00 -1.02912448e-01 7.62607217e-01 8.79803002e-01
5.55941045e-01 3.49014699e-01 -4.47801590e-01 9.90046144e-01
7.32795820e-02 9.08155274e-03 8.59541476e-01 -4.74772334e-01
8.39678347e-01 7.42486417e-01 -3.29084806e-02 -6.64600074e-01
-4.08178210e-01 -2.47629583e-01 -9.48802352e-01 -6.61106229e-01
-2.16838997e-02 -4.96323794e-01 -1.05130100e+00 1.41171134e+00
-7.86867067e-02 -3.02134126e-01 4.32026386e-01 3.00133079e-01
1.02724493e+00 1.07802510e+00 3.54571909e-01 -3.80729347e-01
1.10123098e+00 -1.06597364e+00 -7.42269099e-01 -4.18445438e-01
9.47461545e-01 -5.87744296e-01 1.02625871e+00 8.41054544e-02
-9.15109336e-01 -5.73474765e-01 -9.71540451e-01 -4.38102365e-01
-5.19358158e-01 2.16387168e-01 6.27087653e-01 5.48421681e-01
-7.81191766e-01 3.09568584e-01 -6.72504246e-01 -4.65249181e-01
4.93571222e-01 4.01859850e-01 -8.10407624e-02 -1.08772129e-01
-9.83017266e-01 5.74655414e-01 6.75444722e-01 2.81950742e-01
-6.39306903e-01 -2.53438711e-01 -1.07205760e+00 4.72030491e-01
2.81265378e-01 -6.33886874e-01 1.46727014e+00 -1.01366913e+00
-1.45123315e+00 6.00071549e-01 -3.96153420e-01 -5.85321665e-01
-4.76017073e-02 -6.60255134e-01 5.57075255e-02 3.03723305e-01
2.45520975e-02 5.98028064e-01 7.65947342e-01 -1.05642056e+00
-4.98097807e-01 -3.52669895e-01 3.03941727e-01 4.80299592e-01
-6.95796967e-01 1.64629698e-01 -3.95613074e-01 -6.62338793e-01
-1.59776464e-01 -4.97313291e-01 -5.92824459e-01 -1.02090883e+00
-7.62965977e-01 -5.01148224e-01 8.86133313e-01 -6.79971397e-01
1.25704384e+00 -1.84926629e+00 7.30635151e-02 -9.84753966e-02
-2.85029054e-01 2.44805560e-01 -2.09868297e-01 5.94053864e-01
1.29948705e-01 3.88077319e-01 -4.11751211e-01 -5.57088614e-01
-9.03710574e-02 2.12454572e-01 -5.73679864e-01 -1.05836928e-01
2.71151155e-01 9.03712630e-01 -7.62443781e-01 -6.13271952e-01
4.45618033e-02 2.69981831e-01 -4.88993526e-01 1.99262768e-01
-5.10621369e-01 2.53161818e-01 -6.73325956e-01 4.16533202e-01
1.24347575e-01 -1.73108816e-01 3.05428535e-01 3.02528411e-01
6.03672750e-02 6.43857479e-01 -7.66816437e-01 1.62623441e+00
-7.33287930e-01 7.86171794e-01 -7.73664042e-02 -1.34341931e+00
7.51724541e-01 4.40806240e-01 1.87529057e-01 -4.49842244e-01
2.22544476e-01 -1.19457632e-01 -2.18606889e-01 -4.55521703e-01
1.20989215e+00 -5.34214042e-02 -4.76320982e-01 5.80315590e-01
1.50777668e-01 -2.96418160e-01 6.12436533e-01 5.40745735e-01
8.31096470e-01 -3.46384794e-02 4.34325546e-01 -1.04789793e-01
6.81216538e-01 -5.73020279e-02 2.78626859e-01 8.58169675e-01
3.24474514e-01 7.13512361e-01 6.88890755e-01 1.01008220e-03
-8.27508330e-01 -3.46628726e-01 -1.79413967e-02 1.61745691e+00
-1.69367909e-01 -7.09382474e-01 -1.00667977e+00 -9.01900947e-01
-3.34858507e-01 8.53113592e-01 -5.32827139e-01 1.76506624e-01
-8.66327882e-01 -7.84610569e-01 3.44260365e-01 9.91957426e-01
2.68030494e-01 -1.36613929e+00 -4.70673770e-01 4.12897319e-01
4.21480238e-02 -1.17520058e+00 -3.92194092e-01 3.70550722e-01
-1.29200387e+00 -8.36417317e-01 -7.02328444e-01 -1.11352336e+00
9.31134820e-01 1.60745680e-01 1.00082624e+00 1.81757882e-01
-6.24853969e-02 6.42459810e-01 -5.81131577e-01 -4.13055867e-01
-3.71127933e-01 7.74690807e-01 -2.20395193e-01 -2.96098918e-01
2.66248226e-01 -2.36317307e-01 -4.61661518e-01 -2.56572098e-01
-1.26658702e+00 -4.23997827e-02 7.34027147e-01 8.85154963e-01
5.34087479e-01 -9.78431851e-02 1.06111658e+00 -1.49264789e+00
1.19601262e+00 -3.24003279e-01 -1.96092546e-01 3.13886911e-01
-2.27431908e-01 1.11449279e-01 8.58803272e-01 -2.91212410e-01
-1.27657914e+00 -2.92723715e-01 -3.03163707e-01 2.43407756e-01
-4.25483763e-01 9.83379304e-01 -4.70904522e-02 6.44278288e-01
4.66903418e-01 1.16980240e-01 -6.10494554e-01 -4.15843368e-01
5.75937152e-01 8.04667473e-01 4.09367591e-01 -6.47341549e-01
3.38388443e-01 6.91215247e-02 -4.56064939e-01 -1.26922894e+00
-1.21155393e+00 -7.45684385e-01 -9.66371000e-01 3.50659847e-01
1.10926139e+00 -5.42699039e-01 -1.93075642e-01 1.01403818e-01
-1.20345676e+00 -2.65017688e-01 -4.55647677e-01 3.82797658e-01
-2.46982753e-01 7.49329984e-01 -7.19112396e-01 -8.11743915e-01
-8.23702097e-01 -8.77199411e-01 1.00238001e+00 5.27092576e-01
-4.38303828e-01 -1.32976520e+00 -5.99388266e-03 4.01945025e-01
1.99492633e-01 8.32929835e-02 1.40127110e+00 -1.05878520e+00
-1.58079952e-01 -5.83149433e-01 -1.46888187e-02 6.81329429e-01
2.68350512e-01 9.08380970e-02 -9.70367253e-01 -2.28549168e-01
-1.09516844e-01 -6.40416443e-01 1.19706082e+00 6.47125721e-01
1.25489128e+00 -4.71871853e-01 -2.50387162e-01 1.32635161e-02
1.16095352e+00 -3.82592343e-02 1.95465580e-01 1.46962613e-01
8.33033264e-01 7.56358504e-01 4.15965348e-01 1.92778498e-01
3.10900897e-01 -8.00744891e-02 7.40419328e-02 -7.93870687e-02
2.89131016e-01 -9.43935066e-02 2.78145909e-01 1.25860798e+00
1.91693187e-01 -4.58687961e-01 -7.43234515e-01 8.38348210e-01
-1.84545720e+00 -7.24120736e-01 1.24863461e-01 1.85405767e+00
8.54825377e-01 4.30631906e-01 -1.08588405e-01 5.04521765e-02
6.33088708e-01 4.55752999e-01 -4.73945767e-01 -9.97431755e-01
1.13732629e-01 1.90077677e-01 1.61729887e-01 5.40207505e-01
-1.26364803e+00 1.28183925e+00 6.36486387e+00 5.15054107e-01
-9.61308420e-01 -1.83288172e-01 7.24869728e-01 -1.25019820e-02
-4.00093436e-01 -2.52584442e-02 -1.02021945e+00 6.81034178e-02
1.16968048e+00 -1.10306621e-01 -3.48806441e-01 9.39286351e-01
1.37406752e-01 -3.53298962e-01 -1.22300613e+00 5.81487238e-01
3.01033795e-01 -1.38593554e+00 4.58958507e-01 -1.79996803e-01
1.06023502e+00 -1.86851069e-01 -1.08927958e-01 7.78079867e-01
2.42703453e-01 -1.02395177e+00 2.16769651e-01 -8.44587106e-03
4.38228309e-01 -8.99038255e-01 8.78747046e-01 6.03749394e-01
-1.01646459e+00 -8.41851905e-02 -6.13532901e-01 -9.06071886e-02
2.02564985e-01 4.42600280e-01 -9.92207170e-01 6.67884767e-01
3.28664213e-01 6.46063209e-01 -5.00298321e-01 7.64944017e-01
-4.69614297e-01 9.12849724e-01 -1.13257207e-01 -2.67788947e-01
7.32860506e-01 -1.59567952e-01 1.26975924e-01 1.72607458e+00
-1.31363779e-01 1.27069309e-01 3.00378323e-01 1.81338862e-01
-5.58884561e-01 5.36164284e-01 -4.62605864e-01 -3.99851382e-01
6.42391890e-02 1.24310255e+00 -9.94342148e-01 -6.53654039e-01
-5.86937428e-01 8.34630489e-01 4.01828289e-01 5.62276065e-01
-2.60211200e-01 -6.37023270e-01 4.81479242e-02 -2.14597285e-01
6.21737063e-01 -5.30837059e-01 -3.79860103e-01 -1.15921187e+00
-8.14715475e-02 -6.98941410e-01 6.84889734e-01 -3.58282000e-01
-9.71680403e-01 6.34144723e-01 8.31117854e-02 -5.97246766e-01
-7.39021897e-01 -5.63024163e-01 -1.03865850e+00 5.62602758e-01
-1.51382613e+00 -1.19381297e+00 1.60320312e-01 2.60237992e-01
1.25564933e+00 -7.55490661e-02 1.06724906e+00 -2.11818308e-01
-7.69159436e-01 3.60177219e-01 4.49690521e-01 4.12777185e-01
4.32604551e-01 -1.44122863e+00 3.68050873e-01 9.04140651e-01
4.29395229e-01 1.01372039e+00 4.28671926e-01 -6.17207050e-01
-1.06922257e+00 -9.94037509e-01 7.46153831e-01 -3.26754332e-01
5.84304154e-01 -5.02432823e-01 -1.09576726e+00 9.64787781e-01
7.36867428e-01 -3.87718827e-01 1.10253263e+00 5.35116494e-01
-1.20911144e-01 1.66592494e-01 -6.99585021e-01 5.18758237e-01
3.42699438e-01 -6.16918743e-01 -1.05549467e+00 4.71778780e-01
7.95176148e-01 -2.01901406e-01 -5.02712727e-01 2.61979133e-01
1.97501346e-01 -3.89224350e-01 5.82654119e-01 -9.76152956e-01
4.61010009e-01 2.27698982e-01 3.37036476e-02 -1.09507203e+00
2.57761013e-02 -4.14483607e-01 -1.83811530e-01 1.59100127e+00
5.63395321e-01 -1.01969868e-01 8.03929985e-01 3.71762305e-01
-3.47777516e-01 -8.48677039e-01 -4.95358706e-01 -3.20652038e-01
4.66893673e-01 -3.52233022e-01 1.49792060e-01 6.65697157e-01
4.81943488e-02 9.90002871e-01 -1.64258361e-01 -4.87110429e-02
1.56881303e-01 5.01310110e-01 6.15592480e-01 -1.10649908e+00
-9.42947641e-02 -5.35783887e-01 9.88185704e-02 -1.47122014e+00
6.21510684e-01 -8.20478797e-01 4.09718454e-01 -2.23307610e+00
2.19196707e-01 -6.51042536e-02 -2.79901505e-01 3.30880880e-01
-4.47923094e-01 -1.86831757e-01 -1.02069832e-01 1.03310116e-01
-9.43199396e-01 6.24613762e-01 9.14420664e-01 -3.75946373e-01
-3.87218267e-01 2.13522330e-01 -9.98266280e-01 9.25778747e-01
8.44398618e-01 -4.07016009e-01 -5.72171390e-01 -6.38131380e-01
1.33551151e-01 -1.58769533e-01 -4.48817015e-01 -7.02508032e-01
2.07851678e-01 -1.60057202e-01 5.14825463e-01 -9.89855230e-01
8.91834348e-02 -4.16967750e-01 -1.02611184e+00 3.37580947e-04
-8.26164722e-01 -1.56174898e-01 2.91715711e-01 6.30775452e-01
-4.86255705e-01 -8.83654892e-01 4.77154374e-01 -4.25614119e-01
-5.94521463e-01 7.39794225e-02 -7.79706955e-01 2.66821504e-01
5.41129410e-01 -2.52639860e-01 -2.22267464e-01 -3.67013156e-01
-5.32552779e-01 4.73076642e-01 3.70429568e-02 4.43205714e-01
6.06800437e-01 -7.25159526e-01 -6.15156114e-01 -2.41125688e-01
-1.71536654e-01 5.41293025e-01 2.20739156e-01 2.86942959e-01
-4.32580441e-01 8.51086199e-01 2.19796136e-01 -4.84630793e-01
-1.11514044e+00 4.92412716e-01 -1.93993971e-01 -7.56174505e-01
-6.25369251e-01 8.03615868e-01 5.12819052e-01 -2.17165112e-01
4.22041595e-01 -7.76602209e-01 -7.64347434e-01 4.37644631e-01
5.45687973e-01 4.57070163e-03 9.21141580e-02 -4.03201133e-01
-6.06748182e-03 4.64590579e-01 -7.92165697e-01 -1.04201481e-01
1.43946338e+00 -1.93359464e-01 -9.38491300e-02 7.19893754e-01
1.00423777e+00 1.03772553e-02 -8.28470111e-01 -1.76539764e-01
5.64276874e-01 2.03666598e-01 -2.57570427e-02 -1.99230164e-01
-5.14723539e-01 1.34018016e+00 -1.38191432e-01 4.72692132e-01
8.19985032e-01 -1.15098715e-01 8.29610884e-01 9.66775358e-01
-2.49709412e-01 -1.36146283e+00 3.55170637e-01 1.00130129e+00
8.10628474e-01 -1.14663410e+00 2.35690460e-01 9.17533189e-02
-6.77580714e-01 1.22963262e+00 7.74322808e-01 -1.74708098e-01
3.07108343e-01 -4.16927971e-02 -9.88455191e-02 2.99414266e-02
-6.00439012e-01 -1.45514861e-01 2.25026235e-01 3.07300985e-01
8.91693294e-01 -2.78400093e-01 -1.95957050e-01 6.66846812e-01
-8.08422491e-02 -3.94662172e-01 7.25203931e-01 1.31641710e+00
-7.13719726e-01 -1.10558558e+00 -1.11679338e-01 5.75938046e-01
-9.20289695e-01 -3.67662162e-01 -5.79959333e-01 5.96807599e-01
-4.81972694e-01 8.82812202e-01 1.25054404e-01 3.07283610e-01
1.53296426e-01 4.36696351e-01 2.82961726e-01 -1.42867613e+00
-8.14418077e-01 1.90267891e-01 2.28221506e-01 1.14785053e-01
-5.02451718e-01 -3.87981474e-01 -1.56447971e+00 3.26386750e-01
-3.26084733e-01 5.57812095e-01 6.56811476e-01 1.19225228e+00
1.17824890e-01 7.67776132e-01 3.92003089e-01 -9.23698127e-01
-4.10859227e-01 -1.39824402e+00 -3.34830970e-01 1.87044606e-01
5.14802277e-01 5.03740646e-02 2.98145256e-04 1.82812959e-01] | [12.53037166595459, 9.478259086608887] |
8348836e-c8fe-4647-8735-433910096961 | accelerating-diffusion-models-for-inverse | 2305.16965 | null | https://arxiv.org/abs/2305.16965v1 | https://arxiv.org/pdf/2305.16965v1.pdf | Accelerating Diffusion Models for Inverse Problems through Shortcut Sampling | Recently, diffusion models have demonstrated a remarkable ability to solve inverse problems in an unsupervised manner. Existing methods mainly focus on modifying the posterior sampling process while neglecting the potential of the forward process. In this work, we propose Shortcut Sampling for Diffusion (SSD), a novel pipeline for solving inverse problems. Instead of initiating from random noise, the key concept of SSD is to find the "Embryo", a transitional state that bridges the measurement image y and the restored image x. By utilizing the "shortcut" path of "input-Embryo-output", SSD can achieve precise and fast restoration. To obtain the Embryo in the forward process, We propose Distortion Adaptive Inversion (DA Inversion). Moreover, we apply back projection and attention injection as additional consistency constraints during the generation process. Experimentally, we demonstrate the effectiveness of SSD on several representative tasks, including super-resolution, deblurring, and colorization. Compared to state-of-the-art zero-shot methods, our method achieves competitive results with only 30 NFEs. Moreover, SSD with 100 NFEs can outperform state-of-the-art zero-shot methods in certain tasks. | ['Yujiu Yang', 'Fei Yin', 'Jiayi Li', 'Haoze Sun', 'Gongye Liu'] | 2023-05-26 | null | null | null | null | ['colorization', 'deblurring'] | ['computer-vision', 'computer-vision'] | [ 3.58100206e-01 -9.33437124e-02 2.89040685e-01 -2.03032956e-01
-8.18136513e-01 -1.55185461e-01 6.48686707e-01 -6.89162076e-01
-3.65094602e-01 4.77844685e-01 2.92565465e-01 1.17238760e-01
-1.47964448e-01 -6.48705781e-01 -5.67361474e-01 -8.67779195e-01
6.10735118e-01 2.81989992e-01 4.48032051e-01 -1.05315998e-01
4.63371009e-01 3.04917246e-01 -1.21219206e+00 3.39530975e-01
1.21929288e+00 6.17260456e-01 7.50390410e-01 4.15919781e-01
-2.14799568e-01 6.54217362e-01 -5.20658374e-01 -2.44544670e-01
1.18016280e-01 -7.90252268e-01 -5.52339494e-01 9.55406169e-04
3.93486112e-01 -7.12791324e-01 -6.23258829e-01 1.49927104e+00
6.90500855e-01 3.40437084e-01 5.66072762e-01 -7.87463129e-01
-1.36693656e+00 5.30647159e-01 -1.08741307e+00 3.93020689e-01
1.97776228e-01 2.79406577e-01 6.03085279e-01 -1.40229416e+00
1.02348053e+00 1.28275585e+00 3.07784379e-01 7.37668037e-01
-1.33996034e+00 -6.61347330e-01 2.92097092e-01 3.92374367e-01
-1.40735281e+00 -7.55596220e-01 8.25102746e-01 -3.21266621e-01
8.03119957e-01 4.77937683e-02 5.62735915e-01 1.09772623e+00
4.33134846e-02 9.38093066e-01 1.12720251e+00 -2.11875454e-01
3.34281564e-01 -2.19414920e-01 -2.64881528e-03 6.07475817e-01
-1.78828955e-01 1.84773892e-01 -8.26279819e-01 1.86544940e-01
9.47643518e-01 -5.87841831e-02 -6.06642485e-01 -7.16096535e-02
-1.24841297e+00 4.06322241e-01 4.54598308e-01 1.35634765e-01
-4.63275731e-01 1.26475841e-01 -7.44404867e-02 1.86916180e-02
7.20636427e-01 1.79430261e-01 -3.59634571e-02 -2.69340575e-02
-1.31873918e+00 4.65661585e-02 4.00694877e-01 8.26295137e-01
6.96329832e-01 5.92065118e-02 -7.19823599e-01 9.26143825e-01
2.49432921e-01 4.03971046e-01 3.21986020e-01 -1.18791008e+00
3.42309564e-01 1.49789795e-01 2.48477206e-01 -7.39058495e-01
-5.64374886e-02 -5.40776730e-01 -1.07968163e+00 1.91816315e-01
2.67594188e-01 -2.75433566e-02 -1.22356284e+00 1.52398551e+00
5.05075634e-01 5.79550385e-01 4.52846522e-03 1.27419853e+00
7.17068136e-01 9.38490570e-01 -4.05535161e-01 -3.71097326e-01
1.12215805e+00 -1.30762649e+00 -9.94152546e-01 -4.08342957e-01
6.35263696e-02 -8.14142883e-01 1.16756582e+00 5.97187996e-01
-1.36202657e+00 -3.65135103e-01 -9.65252817e-01 -5.91005683e-01
1.99269503e-01 2.29647636e-01 3.68641347e-01 2.62962610e-01
-9.91117477e-01 7.79430449e-01 -1.04774642e+00 -6.02425113e-02
5.11940777e-01 -8.62323046e-02 -5.95257618e-02 -5.27048886e-01
-9.10750866e-01 7.87873209e-01 -3.05173188e-01 3.23902518e-01
-1.23410881e+00 -8.69270205e-01 -6.36730969e-01 1.41984463e-01
4.77778375e-01 -7.82765388e-01 8.66421342e-01 -6.53980672e-01
-1.81137252e+00 6.63562000e-01 -5.17860889e-01 -1.36557132e-01
6.95088089e-01 -5.47858238e-01 -2.76167423e-01 3.16031218e-01
1.96299389e-01 6.28398061e-01 1.21782768e+00 -1.29048324e+00
-2.64547110e-01 -4.53790158e-01 -2.00405791e-01 1.93503261e-01
-2.55487710e-01 1.14688337e-01 -1.01723731e+00 -8.27714682e-01
6.34531140e-01 -7.46317744e-01 -2.76238501e-01 4.05586362e-01
-5.31427562e-01 1.38895303e-01 9.19944823e-01 -8.98300827e-01
1.21417987e+00 -2.36002111e+00 6.60794914e-01 -7.27643892e-02
3.16540837e-01 2.00690985e-01 -1.05136909e-01 8.11291933e-02
9.34734344e-02 -5.47508560e-02 -6.79124057e-01 -6.59589648e-01
-7.98949227e-02 3.40547934e-02 -4.47839081e-01 6.24229670e-01
-1.31564569e-02 8.60396564e-01 -8.88386726e-01 -3.39120686e-01
2.00205475e-01 7.65192509e-01 -5.34080327e-01 1.61014706e-01
-1.30680263e-01 6.16897881e-01 -2.38115773e-01 4.35184389e-01
1.07086945e+00 -4.04197961e-01 -6.05826871e-03 -3.84239286e-01
-2.21296757e-01 2.80543957e-02 -1.20236242e+00 2.25148225e+00
-3.30201685e-01 6.08561516e-01 3.03709328e-01 -6.01656258e-01
8.16177547e-01 4.53814380e-02 1.34195670e-01 -8.02019894e-01
-1.84003726e-01 5.10495454e-02 -1.07728966e-01 -5.55912614e-01
5.61209679e-01 4.68441844e-02 4.96678382e-01 6.97419703e-01
-4.68477197e-02 -2.69092351e-01 2.28164330e-01 5.46713054e-01
8.33369136e-01 5.02137244e-01 -2.27327555e-01 -1.32630587e-01
4.25319433e-01 -3.63786608e-01 6.65550292e-01 7.01628983e-01
-1.33025497e-01 1.27786052e+00 5.28272390e-01 3.96698005e-02
-1.04164732e+00 -1.25604534e+00 1.18278295e-01 7.67258942e-01
5.47599256e-01 -3.55506808e-01 -1.07497716e+00 -3.34469795e-01
-4.41620201e-01 9.24537301e-01 -7.05884874e-01 -1.79254055e-01
-6.02497101e-01 -9.14058983e-01 2.94796556e-01 3.69748175e-01
7.29186356e-01 -8.40208054e-01 -2.54287690e-01 1.41544551e-01
-3.60514522e-01 -1.02339041e+00 -7.68642724e-01 -3.37998122e-01
-8.04962039e-01 -7.12760389e-01 -1.24342310e+00 -4.02810305e-01
8.80543292e-01 6.72576129e-01 7.59778559e-01 2.30225711e-03
-2.82753050e-01 -1.00549432e-02 -1.17995411e-01 2.31068850e-01
-9.95011181e-02 -2.31081218e-01 -2.19132572e-01 3.12007159e-01
-1.10431410e-01 -7.00965345e-01 -1.03762627e+00 3.57826471e-01
-9.40432191e-01 3.89076799e-01 6.10686839e-01 8.04148078e-01
8.51098478e-01 -1.19642258e-01 2.72319347e-01 -7.78412402e-01
6.43923402e-01 -2.73266524e-01 -4.08698350e-01 3.64528507e-01
-6.83042943e-01 2.56903321e-01 3.91077280e-01 -6.04524612e-01
-1.61609006e+00 -1.68979689e-01 -1.14097655e-01 -7.79824972e-01
2.18093559e-01 2.11695567e-01 -1.92523018e-01 7.78906569e-02
6.01252973e-01 5.20460784e-01 -2.23233383e-02 -8.65558624e-01
5.81734180e-01 3.19856107e-01 6.28476679e-01 -4.74319369e-01
6.83544278e-01 1.03668880e+00 -2.41231456e-01 -8.22987795e-01
-9.98590887e-01 -1.18492708e-01 -5.70464671e-01 -1.97422639e-01
9.96415138e-01 -8.16964686e-01 -2.93842286e-01 8.20746362e-01
-1.23390388e+00 -4.24468547e-01 -2.77671337e-01 2.83386678e-01
-3.73559743e-01 5.64170301e-01 -8.42800319e-01 -6.48301542e-01
-3.97083431e-01 -1.25693178e+00 1.04064298e+00 4.04974371e-01
1.60802737e-01 -5.53717017e-01 -1.48959517e-01 3.26732218e-01
5.91686368e-01 -3.26370120e-01 5.62971652e-01 1.62687019e-01
-8.70329440e-01 2.49855772e-01 -6.15876079e-01 2.83934683e-01
-6.31449521e-02 9.87209156e-02 -9.19377625e-01 -2.64505416e-01
2.56569058e-01 1.03546992e-01 1.28629959e+00 5.85531831e-01
1.20209515e+00 3.95944901e-03 -3.79508227e-01 1.13554120e+00
1.19677711e+00 -4.48719859e-02 1.09225070e+00 1.14521012e-01
7.20654726e-01 4.36769277e-01 4.46658254e-01 3.67884785e-01
3.61540437e-01 6.74662888e-01 1.72050953e-01 -1.41016439e-01
-6.94734335e-01 -2.75844514e-01 3.84816617e-01 8.10834169e-01
-5.09191640e-02 -1.90016270e-01 -6.67782307e-01 5.02325892e-01
-1.71919739e+00 -9.49187219e-01 -2.45418340e-01 2.01529360e+00
7.97888219e-01 7.00649992e-02 -5.29690027e-01 -1.81598827e-01
8.12854886e-01 5.06362855e-01 -8.61566305e-01 2.21785963e-01
-2.47920752e-01 8.27130973e-02 8.97574052e-02 7.78987110e-01
-6.57949090e-01 1.14892554e+00 5.67460251e+00 9.63902235e-01
-9.93098021e-01 4.49143380e-01 4.93784487e-01 -3.90176207e-01
-4.95727748e-01 1.94125623e-01 -7.24172115e-01 5.17692685e-01
2.88436294e-01 8.99532624e-03 7.77311802e-01 3.58644396e-01
4.07050610e-01 -2.95773715e-01 -7.80443311e-01 8.54383767e-01
2.32508391e-01 -1.30306053e+00 -8.02827328e-02 -1.16378777e-01
1.07248962e+00 -3.70320468e-03 2.59761751e-01 -1.50387570e-01
1.45739645e-01 -7.82702744e-01 8.59694302e-01 9.80087221e-01
9.42132294e-01 -4.72522795e-01 6.98638111e-02 4.66827661e-01
-9.08522248e-01 4.56509702e-02 -5.67271590e-01 2.37058714e-01
7.59800196e-01 1.03521717e+00 3.75461690e-02 4.79754329e-01
7.99886227e-01 8.49358976e-01 -2.59293437e-01 9.88750458e-01
-7.30211079e-01 3.92595053e-01 -1.91677198e-01 5.42642772e-01
1.63725261e-02 -5.70475578e-01 7.57527709e-01 1.02029574e+00
6.85110331e-01 4.85050231e-01 -2.85693586e-01 1.29374456e+00
-2.29427926e-02 -3.05880904e-01 -1.47624046e-01 7.26612061e-02
4.04553533e-01 1.11747289e+00 -8.07473958e-01 -5.17931342e-01
-2.21228182e-01 1.65703666e+00 2.64113873e-01 8.31540883e-01
-8.56136620e-01 -3.57283890e-01 5.39339960e-01 1.58950826e-03
5.60753405e-01 -3.46888453e-01 -3.59559298e-01 -1.38630509e+00
5.96175753e-02 -7.15405464e-01 -1.14429807e-02 -1.19039536e+00
-1.14171743e+00 4.96099502e-01 -2.50913918e-01 -1.10467899e+00
1.86699435e-01 -1.13594286e-01 -6.01147592e-01 1.15387881e+00
-1.44245863e+00 -9.74173427e-01 -5.65905631e-01 5.14112890e-01
8.09082806e-01 2.47130573e-01 3.21551889e-01 3.64363760e-01
-7.55176663e-01 1.62038401e-01 3.81119430e-01 -2.25976408e-01
9.10342872e-01 -9.50346231e-01 7.81132877e-01 1.27967930e+00
2.96984296e-02 7.54618049e-01 6.87269866e-01 -9.24470305e-01
-1.38709557e+00 -7.09192634e-01 4.74447280e-01 -5.00702485e-02
5.37771225e-01 -3.05694312e-01 -1.26433063e+00 3.23867142e-01
1.86372057e-01 1.99140891e-01 2.85076704e-02 -3.22428346e-01
-3.30307215e-01 -9.83044729e-02 -9.96706307e-01 7.40640819e-01
1.38904560e+00 -6.32779360e-01 -3.51665080e-01 2.03457415e-01
7.69066453e-01 -6.99153304e-01 -3.83545190e-01 8.53600875e-02
4.85020101e-01 -1.16981542e+00 1.25836599e+00 -1.99705765e-01
9.39755082e-01 -5.20015240e-01 1.58898860e-01 -1.35766423e+00
-4.02964711e-01 -7.82143593e-01 -5.05168557e-01 1.22017670e+00
2.82751471e-01 -3.86814535e-01 6.19781613e-01 5.45245469e-01
-2.31334969e-01 -6.28912091e-01 -7.88474679e-01 -4.99457568e-01
-1.34319991e-01 -2.42247671e-01 4.56849486e-01 7.29588151e-01
-5.92407882e-01 2.74994284e-01 -6.85193896e-01 1.95933253e-01
1.00778484e+00 1.09273285e-01 5.64785957e-01 -8.11821342e-01
-5.70876479e-01 -3.48439455e-01 3.53444904e-01 -1.67263889e+00
-4.43844460e-02 -6.15728736e-01 1.88987672e-01 -1.93829727e+00
2.48319417e-01 -1.89885795e-01 -1.58356786e-01 -3.92176919e-02
-3.85823846e-01 2.76092142e-01 2.47271463e-01 4.96131122e-01
-3.80079418e-01 8.68721187e-01 1.77414107e+00 -1.30902022e-01
-1.95155725e-01 -2.41998225e-01 -6.55462325e-01 7.29491770e-01
3.42873365e-01 -6.01465106e-01 -4.68265951e-01 -1.04181838e+00
1.87801898e-01 9.95682254e-02 3.18800569e-01 -6.95664406e-01
6.97967649e-01 -4.91180979e-02 3.48925203e-01 -7.12707996e-01
6.68408334e-01 -2.98582882e-01 1.62519231e-01 2.45696843e-01
-2.62605309e-01 -1.73974514e-01 -1.92914501e-01 7.24612415e-01
-4.34009545e-02 -4.52944219e-01 8.69493306e-01 -1.58743605e-01
-6.23425722e-01 3.65848362e-01 -1.26960114e-01 1.69857163e-02
8.33496988e-01 -1.03246778e-01 -5.26731670e-01 -2.87003666e-01
-7.93416262e-01 9.67460126e-02 6.59028590e-01 2.20784292e-01
9.04113650e-01 -8.58110249e-01 -7.74238467e-01 3.43471646e-01
-4.27811116e-01 1.75123364e-01 8.15635622e-01 1.11435485e+00
-5.00172734e-01 -2.11692557e-01 3.08335163e-02 -4.70188946e-01
-9.68036592e-01 6.09554052e-01 4.11235154e-01 -4.45849635e-03
-1.17559063e+00 1.18904352e+00 5.68878710e-01 -8.41525123e-02
7.43198320e-02 3.98377292e-02 3.76578160e-02 -2.12015379e-02
8.58701169e-01 6.12531662e-01 -2.47828767e-01 -3.03405493e-01
-1.76775500e-01 7.19518483e-01 -2.66004741e-01 -6.64637506e-01
1.41491449e+00 -4.93495792e-01 -3.96346837e-01 1.84914514e-01
8.60407591e-01 8.11727419e-02 -1.81184781e+00 -2.62216866e-01
-3.30874950e-01 -8.47251475e-01 4.46339071e-01 -7.98207164e-01
-1.42940986e+00 1.26276767e+00 6.16750777e-01 -1.75764859e-01
1.17858267e+00 -1.25369281e-02 8.90204310e-01 -1.58243049e-02
1.02860034e-01 -1.15421999e+00 1.66525692e-01 3.89082998e-01
1.07971609e+00 -9.84146416e-01 2.85686553e-01 -5.86611509e-01
-6.34018779e-01 7.90138006e-01 6.12014890e-01 -6.48565665e-02
5.25583148e-01 4.11888748e-01 -4.57781218e-02 -2.15505674e-01
-7.06510365e-01 -6.57636747e-02 2.11434066e-01 5.57230175e-01
1.49082333e-01 -3.94865125e-01 -3.37042272e-01 4.46635425e-01
3.01598936e-01 2.18440622e-01 6.00925684e-01 6.85350657e-01
-4.10632968e-01 -8.30351591e-01 -3.59081239e-01 2.10336462e-01
-3.15496176e-01 -4.51973110e-01 -1.75206929e-01 2.37860039e-01
-2.29380727e-01 7.56481290e-01 7.78881647e-03 -7.77337030e-02
1.48172036e-01 -3.43944818e-01 5.37896991e-01 -4.41174567e-01
-2.64497489e-01 3.77911061e-01 -1.95436522e-01 -7.46047676e-01
-2.73925126e-01 -6.29320085e-01 -1.23079526e+00 -1.67045400e-01
-3.53402257e-01 -1.68753982e-01 6.38354480e-01 9.95616913e-01
5.52520156e-01 7.03796804e-01 3.75344127e-01 -1.06421816e+00
-5.30451238e-01 -9.72640157e-01 -5.83929002e-01 2.24022046e-01
2.74643600e-01 -6.83649838e-01 -4.48259205e-01 7.04582185e-02] | [11.418567657470703, -2.1658918857574463] |
82f4cf3a-e503-4576-a39a-712bbb399e13 | joint-entity-and-relation-extraction-based-on | null | null | https://aclanthology.org/2022.spnlp-1.2 | https://aclanthology.org/2022.spnlp-1.2.pdf | Joint Entity and Relation Extraction Based on Table Labeling Using Convolutional Neural Networks | This study introduces a novel approach to the joint extraction of entities and relations by stacking convolutional neural networks (CNNs) on pretrained language models. We adopt table representations to model the entities and relations, casting the entity and relation extraction as a table-labeling problem. Regarding each table as an image and each cell in a table as an image pixel, we apply two-dimensional CNNs to the tables to capture local dependencies and predict the cell labels. The experimental results showed that the performance of the proposed method is comparable to those of current state-of-art systems on the CoNLL04, ACE05, and ADE datasets.Even when freezing pretrained language model parameters, the proposed method showed a stable performance, whereas the compared methods suffered from significant decreases in performance. This observation indicates that the parameters of the pretrained encoder may incorporate dependencies among the entity and relation labels during fine-tuning. | ['Naoaki Okazaki', 'Tatsuya Hiraoka', 'Youmi Ma'] | null | null | null | null | spnlp-acl-2022-5 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [-1.48198143e-01 4.27914113e-01 -3.28496397e-01 -6.32375300e-01
-7.17415273e-01 -5.84752381e-01 6.53637648e-01 3.28278959e-01
-4.27286595e-01 8.42705667e-01 7.97893330e-02 -2.77802140e-01
2.69021332e-01 -1.18813026e+00 -1.10928237e+00 -2.31075436e-01
-2.59075135e-01 7.51181483e-01 1.09913625e-01 -1.30741537e-01
-1.68296427e-01 6.00339055e-01 -1.30575514e+00 6.79518878e-01
6.04260564e-01 1.39561057e+00 -3.59351426e-01 4.32514101e-01
-4.97110635e-01 1.41067743e+00 -8.75498354e-01 -1.12972271e+00
1.20581783e-01 -1.66528404e-01 -9.43369269e-01 1.02839256e-02
4.67491984e-01 -2.46103272e-01 -6.22277141e-01 9.99893546e-01
5.59719093e-02 -1.03408739e-01 4.79099900e-01 -1.12020063e+00
-1.16891384e+00 1.29895294e+00 -3.89202148e-01 1.11586340e-01
3.43893096e-02 -5.42173624e-01 1.36342001e+00 -9.29339707e-01
7.47351527e-01 1.35704029e+00 5.33433020e-01 2.01993480e-01
-1.06693733e+00 -8.16205859e-01 3.29903007e-01 4.24674660e-01
-1.71741748e+00 -3.73476565e-01 4.01500940e-01 -3.53982866e-01
1.35575306e+00 -9.08966884e-02 2.80004591e-01 6.64906025e-01
2.81725198e-01 6.18028760e-01 5.97656667e-01 -3.99115920e-01
-5.19435480e-02 4.92242217e-01 2.99037427e-01 1.03794444e+00
6.09919310e-01 -4.66921449e-01 -5.98789454e-01 9.93606895e-02
7.47854233e-01 -4.10309583e-01 1.60718188e-01 -4.37515944e-01
-1.23557639e+00 6.39823437e-01 8.75566483e-01 3.07790220e-01
-3.12905669e-01 2.13890031e-01 5.61868727e-01 1.05918527e-01
4.94662613e-01 3.14522237e-01 -6.37334943e-01 3.47333997e-01
-5.00219166e-01 6.84838369e-02 9.95450675e-01 1.51932454e+00
8.76101315e-01 -2.10333392e-01 -2.70737022e-01 6.20018244e-01
5.63807115e-02 9.39638987e-02 -8.91727805e-02 -5.22063136e-01
1.05362236e+00 1.07279193e+00 2.64475998e-02 -1.13637650e+00
-3.40922862e-01 -5.63953221e-01 -9.78143573e-01 -3.12563032e-01
2.81617254e-01 -1.47604004e-01 -1.02724051e+00 1.52222204e+00
-8.71008821e-03 -1.64191186e-01 2.96977073e-01 4.48172778e-01
1.12411118e+00 7.04592943e-01 2.09991142e-01 -1.29819836e-03
1.43377018e+00 -1.18147385e+00 -1.32560945e+00 -2.96567380e-01
8.57717514e-01 -5.28071046e-01 7.54684210e-01 6.74581081e-02
-1.17848587e+00 -7.33575284e-01 -1.31431675e+00 -4.29993570e-01
-1.04983604e+00 6.55881882e-01 8.18500757e-01 3.32281917e-01
-9.83727872e-01 3.35057765e-01 -7.97987044e-01 -3.25479567e-01
5.31405032e-01 6.51134968e-01 -5.45828521e-01 1.13125004e-01
-1.56114602e+00 1.20463717e+00 6.81779981e-01 5.46698570e-01
-5.52736402e-01 -2.87373751e-01 -1.17330420e+00 5.45781374e-01
5.05828619e-01 -4.22143787e-01 1.07922244e+00 -4.75063920e-01
-1.02650523e+00 1.02022171e+00 -3.12334955e-01 -8.98627341e-01
4.71052863e-02 -4.12557155e-01 -6.74015939e-01 -1.61530584e-01
6.75467923e-02 7.57303953e-01 9.50873345e-02 -1.35593843e+00
-6.31395102e-01 -1.96912408e-01 3.53654683e-01 9.44215953e-02
-1.78578466e-01 4.82896566e-02 -9.95077789e-01 -4.80299979e-01
1.53255969e-01 -7.23783314e-01 3.68776098e-02 -2.90491045e-01
-8.13654721e-01 -3.23647678e-01 6.15025103e-01 -6.90766215e-01
1.65184307e+00 -2.04143882e+00 -3.46980654e-02 7.94430375e-02
1.73622102e-01 2.20688701e-01 1.41054556e-01 3.87487859e-01
-2.23049775e-01 4.64474231e-01 -6.15395717e-02 -3.66437376e-01
2.38390230e-02 4.31454450e-01 -1.12154283e-01 1.10631429e-01
5.66005647e-01 1.08178890e+00 -4.37056333e-01 -7.11650491e-01
-1.03512509e-02 5.41455448e-01 -3.52366954e-01 2.25981876e-01
-3.84396702e-01 -2.26844624e-01 -2.84673631e-01 6.64103150e-01
6.92945361e-01 -5.63010335e-01 6.20643258e-01 -5.48155904e-01
2.31779620e-01 7.44264543e-01 -1.13490987e+00 1.25066364e+00
-4.51378971e-01 6.42975032e-01 -1.61016062e-01 -8.68213952e-01
1.11674881e+00 2.35239536e-01 1.68921173e-01 -7.08061278e-01
2.33634114e-01 1.35007709e-01 1.15241356e-01 -3.27829540e-01
4.66474801e-01 3.01862478e-01 -2.88312197e-01 1.38251493e-02
3.87380034e-01 2.11209282e-01 5.61498940e-01 2.55306333e-01
7.94179916e-01 1.16871089e-01 3.35236698e-01 -1.60336122e-01
9.13590491e-01 1.26869544e-01 4.64306772e-01 5.51700115e-01
2.45177329e-01 2.70276070e-01 8.13067615e-01 -8.43228936e-01
-8.22488785e-01 -9.22539651e-01 -1.16517350e-01 9.87005115e-01
5.14568090e-02 -8.45963538e-01 -8.07909966e-01 -7.98719525e-01
-1.73630510e-02 6.68910444e-01 -9.00596559e-01 5.32432087e-02
-7.69922614e-01 -6.71606541e-01 4.80947971e-01 8.70989263e-01
9.22460020e-01 -1.18508387e+00 -4.21492979e-02 2.83461779e-01
-2.64683068e-01 -1.76894557e+00 -1.42070144e-01 7.29923308e-01
-5.03065586e-01 -9.91495013e-01 -1.20386332e-01 -1.24520516e+00
7.11887896e-01 -3.92438650e-01 1.58068752e+00 4.58696391e-03
2.07873121e-01 -3.45713317e-01 -1.70341507e-01 -4.17534471e-01
-2.72319406e-01 6.27709985e-01 -2.50149697e-01 -1.15706056e-01
6.97970927e-01 -1.54898718e-01 -5.47168553e-02 7.63380900e-02
-6.81380332e-01 8.10178928e-03 5.27550519e-01 7.98143685e-01
7.87661612e-01 2.91753322e-01 2.73057401e-01 -1.48098421e+00
5.19410670e-01 -1.92633465e-01 -7.39632726e-01 6.93705380e-01
-6.16971016e-01 3.12856525e-01 6.15098476e-01 -2.04191450e-02
-1.05484974e+00 3.59198421e-01 6.10455163e-02 -7.04803467e-02
-7.23834187e-02 6.50691926e-01 -5.72477758e-01 3.24252248e-01
3.37971747e-01 -1.66331500e-01 -6.26809835e-01 -3.81888330e-01
5.08540332e-01 4.84712064e-01 6.25092447e-01 -5.57627201e-01
6.33964777e-01 2.27017477e-01 -1.52257696e-01 -1.52014568e-01
-1.26499581e+00 9.50802118e-03 -1.14703810e+00 3.63532096e-01
1.21992707e+00 -1.30093908e+00 -7.40390539e-01 2.74379820e-01
-1.47696328e+00 -1.84670165e-01 -4.95366640e-02 2.14351550e-01
-2.32765362e-01 -3.15770000e-01 -1.00496769e+00 -4.42512512e-01
-1.08097717e-01 -1.23007321e+00 8.53881180e-01 7.69564062e-02
-1.34650052e-01 -9.69237924e-01 -3.80274177e-01 3.69743437e-01
6.23084791e-02 1.61891147e-01 1.36484218e+00 -9.30614710e-01
-9.44890380e-01 -3.31069618e-01 -6.31553352e-01 2.02717453e-01
2.46647388e-01 8.71668980e-02 -9.24204230e-01 2.18600765e-01
-4.10765082e-01 -4.09016669e-01 8.75028491e-01 4.99006845e-02
1.30915868e+00 -3.46448451e-01 -4.34844732e-01 7.47668326e-01
1.39091766e+00 5.07671118e-01 6.82679594e-01 5.93719006e-01
9.04123604e-01 5.10538042e-01 4.50905979e-01 1.48897693e-01
8.50473523e-01 5.70393980e-01 2.58260757e-01 -3.94819736e-01
-9.51993316e-02 -4.33471352e-01 7.59077445e-02 7.70120561e-01
2.03157052e-01 -6.43671155e-01 -1.01524734e+00 4.62346137e-01
-1.82258105e+00 -6.01472199e-01 -9.54438299e-02 1.73273897e+00
9.15451407e-01 5.32090604e-01 -2.76617110e-01 3.04892496e-03
7.84139633e-01 1.66517034e-01 -4.33327556e-01 -6.27586007e-01
-4.62920815e-01 2.56826997e-01 6.96387470e-01 4.33998495e-01
-1.34420609e+00 1.37852538e+00 6.96516323e+00 3.85506004e-01
-8.58793736e-01 -1.78822279e-01 1.01334620e+00 2.32716545e-01
-1.06371082e-02 -2.40455166e-01 -1.21380615e+00 -9.06104445e-02
1.10639620e+00 1.88022456e-03 1.64074555e-01 6.94229245e-01
-4.05922353e-01 -6.62886947e-02 -1.36604583e+00 8.43692541e-01
-4.14496325e-02 -1.64162910e+00 2.78438538e-01 -7.97626153e-02
7.02365518e-01 -1.75237030e-01 -8.15763250e-02 5.20281017e-01
5.68367243e-01 -1.26546967e+00 6.99416101e-01 3.37568730e-01
8.40430439e-01 -9.73065794e-01 1.18545246e+00 -4.46414202e-02
-1.37983179e+00 1.44375861e-01 -3.02229404e-01 -1.81151018e-01
-1.53249159e-01 3.57675284e-01 -9.11593616e-01 5.36238551e-01
8.31121743e-01 6.11020207e-01 -8.66485775e-01 3.89244318e-01
-4.63468194e-01 3.71022075e-01 -8.08529928e-02 3.62561084e-02
2.37361297e-01 -9.98040736e-02 -3.90143394e-01 1.44482994e+00
8.68889224e-03 -7.00201327e-03 -7.99634904e-02 9.83713031e-01
-7.01864243e-01 -4.11310717e-02 -4.40158755e-01 -2.60344774e-01
5.54142118e-01 1.17274296e+00 -8.97787154e-01 -8.15083742e-01
-7.15599000e-01 7.18141675e-01 8.08933139e-01 3.63640070e-01
-9.87415016e-01 -4.55382615e-01 4.91432488e-01 -4.35008109e-02
7.17195213e-01 -2.39579812e-01 -7.80142665e-01 -1.22392118e+00
2.41753399e-01 -7.56623507e-01 4.58971620e-01 -7.56555676e-01
-9.80223179e-01 1.16957414e+00 4.31435779e-02 -8.18651974e-01
-2.45590121e-01 -7.09356129e-01 -8.42886046e-02 8.79299819e-01
-1.30511439e+00 -1.03326666e+00 -1.54335231e-01 3.57989550e-01
1.65044799e-01 -1.40444681e-01 1.07472920e+00 4.68626827e-01
-9.01152134e-01 8.07430148e-01 -1.34688899e-01 8.87902737e-01
6.58492208e-01 -1.34047210e+00 8.03511739e-01 7.97035813e-01
4.42183226e-01 8.22812080e-01 3.73995155e-01 -5.99712431e-01
-1.07833576e+00 -1.09371781e+00 1.55839288e+00 -4.36313838e-01
8.38200569e-01 -9.60628390e-01 -9.89606917e-01 1.29925132e+00
6.03347778e-01 3.16144496e-01 4.97903585e-01 3.95586699e-01
-5.74813366e-01 -2.98137993e-01 -9.27939594e-01 4.25990462e-01
1.03164124e+00 -7.58440614e-01 -4.30204451e-01 3.48416328e-01
8.93821776e-01 -7.62380600e-01 -1.12583530e+00 4.66011047e-01
4.35986876e-01 -7.95390785e-01 8.87652397e-01 -4.77283925e-01
7.12089181e-01 -3.40950608e-01 -2.72186100e-01 -1.07126558e+00
-4.62953925e-01 7.43857026e-02 -5.13527811e-01 1.60613382e+00
1.13642395e+00 -2.85593688e-01 8.73753488e-01 6.84153080e-01
1.68945596e-01 -9.85682726e-01 -3.45622212e-01 -4.90956157e-01
4.65432592e-02 -1.22619800e-01 9.03385818e-01 9.44562972e-01
-1.84254557e-01 7.35940695e-01 -2.17365161e-01 4.48934287e-01
2.07358092e-01 7.34903812e-02 6.22232199e-01 -1.07403696e+00
1.13499127e-01 -1.28297120e-01 -2.26975948e-01 -8.71664762e-01
4.08818424e-01 -8.27615261e-01 -3.36675942e-02 -1.91181827e+00
5.89176118e-02 -3.86963010e-01 -5.41544139e-01 6.22109354e-01
-1.65027916e-01 7.83334896e-02 2.14922100e-01 -1.08416229e-01
-6.04837835e-01 2.28783667e-01 1.15538836e+00 -4.18445885e-01
-8.65353737e-03 -2.31426656e-01 -8.55076551e-01 5.73046923e-01
6.31876111e-01 -4.81330127e-01 -3.67753118e-01 -7.65033066e-01
3.54218483e-01 1.66950226e-01 -1.04480281e-01 -9.97509897e-01
5.39630532e-01 2.09699973e-01 7.04814374e-01 -1.02612484e+00
2.23327518e-01 -9.10766780e-01 4.15581800e-02 1.45747364e-01
-7.01424360e-01 3.97449911e-01 4.11682874e-01 1.81710377e-01
-7.78419435e-01 2.44337648e-01 4.63869274e-01 -2.02961490e-01
-7.38761544e-01 2.09385291e-01 -1.83924735e-01 -5.71916662e-02
7.75121570e-01 1.64779052e-01 -3.30880642e-01 -1.73340380e-01
-9.02655900e-01 2.98261911e-01 6.55390620e-02 4.64186788e-01
3.27909589e-01 -1.38630795e+00 -3.98399025e-01 3.14704359e-01
1.93981379e-01 5.06040990e-01 -9.17018130e-02 2.51800358e-01
-6.70230091e-01 8.36555541e-01 -1.75216228e-01 -2.63766944e-01
-1.08199060e+00 6.73466086e-01 4.67771977e-01 -8.75850320e-01
-1.87994227e-01 1.16925180e+00 2.79331595e-01 -5.29400587e-01
7.00651824e-01 -7.78430283e-01 -3.63198906e-01 1.76485747e-01
2.53413081e-01 -1.71598524e-01 4.71978873e-01 -6.11946583e-01
-5.89653492e-01 2.73559332e-01 -4.95545685e-01 2.26359874e-01
1.20868134e+00 -9.71951634e-02 -2.51361400e-01 5.19603372e-01
1.24309564e+00 -1.35180384e-01 -8.13426197e-01 -4.71760243e-01
3.68329883e-01 1.17442794e-01 -8.19171742e-02 -1.02606881e+00
-1.21327746e+00 9.12074506e-01 1.62967980e-01 1.07330501e-01
1.01090598e+00 1.31253541e-01 6.85541630e-01 8.18579376e-01
3.40695649e-01 -8.54652882e-01 -3.23187143e-01 8.41426313e-01
7.00870216e-01 -1.28050518e+00 3.83401923e-02 -8.30666184e-01
-7.84298062e-01 1.19785798e+00 1.03220499e+00 -1.60505489e-01
7.45182753e-01 8.30006778e-01 2.19887301e-01 -3.11178803e-01
-1.01806319e+00 -2.32621118e-01 4.71292764e-01 2.90391296e-01
8.66111815e-01 7.44472221e-02 -2.21368168e-02 7.10457504e-01
-4.11975831e-01 4.32937816e-02 3.48534167e-01 7.43855238e-01
-4.28828262e-02 -1.16165900e+00 -8.46029148e-02 3.45749795e-01
-5.75622320e-01 -3.89873505e-01 -7.40280151e-01 1.18130744e+00
4.93894517e-01 6.76336586e-01 4.60433632e-01 -6.42274201e-01
6.32174790e-01 3.20814252e-01 2.98573375e-01 -7.42828786e-01
-8.30896556e-01 -1.21314697e-01 4.77325052e-01 -3.36393803e-01
-4.32021588e-01 -2.68627197e-01 -1.47456729e+00 -2.18684688e-01
-3.60951006e-01 1.79810181e-01 3.09930623e-01 8.45949471e-01
3.49098176e-01 1.01324880e+00 3.05187702e-01 -1.30637676e-01
-6.89368695e-02 -9.74392653e-01 -5.92724085e-01 2.86891013e-01
2.10261852e-01 -7.32957780e-01 7.68634975e-02 2.13805839e-01] | [9.512981414794922, 8.054248809814453] |
631f4696-7fd1-46a3-a695-a3b8396daad6 | deep-relevance-ranking-using-enhanced | 1809.01682 | null | http://arxiv.org/abs/1809.01682v2 | http://arxiv.org/pdf/1809.01682v2.pdf | Deep Relevance Ranking Using Enhanced Document-Query Interactions | We explore several new models for document relevance ranking, building upon
the Deep Relevance Matching Model (DRMM) of Guo et al. (2016). Unlike DRMM,
which uses context-insensitive encodings of terms and query-document term
interactions, we inject rich context-sensitive encodings throughout our models,
inspired by PACRR's (Hui et al., 2017) convolutional n-gram matching features,
but extended in several ways including multiple views of query and document
inputs. We test our models on datasets from the BIOASQ question answering
challenge (Tsatsaronis et al., 2015) and TREC ROBUST 2004 (Voorhees, 2005),
showing they outperform BM25-based baselines, DRMM, and PACRR. | ['Ion Androutsopoulos', 'Georgios-Ioannis Brokos', 'Ryan McDonald'] | 2018-09-05 | deep-relevance-ranking-using-enhanced-1 | https://aclanthology.org/D18-1211 | https://aclanthology.org/D18-1211.pdf | emnlp-2018-10 | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [ 1.72570154e-01 -1.28650561e-01 -2.07860976e-01 -4.72422540e-01
-1.37520051e+00 -8.38894069e-01 1.27615535e+00 6.26792669e-01
-7.00192213e-01 4.67788607e-01 8.71820450e-01 -5.33475637e-01
-5.15011013e-01 -5.67959070e-01 -6.58485830e-01 5.43118939e-02
-2.31042400e-01 5.48769355e-01 5.63116670e-01 -7.32038200e-01
7.11890757e-01 9.75107476e-02 -1.38336265e+00 9.00759399e-01
5.23371994e-01 9.44147110e-01 1.90990753e-02 1.15464735e+00
1.12836407e-02 7.79522717e-01 -4.79472935e-01 -7.66879678e-01
-9.06230211e-02 1.94473758e-01 -1.65580070e+00 -8.33650231e-01
9.06088948e-01 -2.82998562e-01 -5.78257501e-01 5.53328276e-01
5.00078678e-01 3.96491319e-01 8.54047596e-01 -6.74502075e-01
-1.32287657e+00 6.71595097e-01 -2.34198809e-01 3.88856858e-01
7.99818218e-01 -5.99629916e-02 1.72038579e+00 -1.22751272e+00
5.41666508e-01 1.48510432e+00 3.97980422e-01 4.16832507e-01
-9.62804794e-01 -2.50872642e-01 2.75932938e-01 5.61809003e-01
-1.07964551e+00 -3.68355960e-01 1.09682910e-01 -1.15918167e-01
1.46059549e+00 7.38658369e-01 1.41476288e-01 1.24906266e+00
1.06511161e-01 1.07194722e+00 5.64612150e-01 -8.02375376e-01
9.14149433e-02 -4.95486200e-01 5.46631634e-01 1.46974847e-01
-1.39293345e-02 2.55067889e-02 -2.80338645e-01 -9.18228567e-01
1.74599379e-01 4.24784124e-02 -1.21361233e-01 1.62935408e-03
-1.06067491e+00 7.63718367e-01 6.40150428e-01 3.69751275e-01
-3.04283023e-01 4.81028050e-01 4.79481101e-01 5.36357641e-01
3.61293018e-01 8.23436856e-01 -5.74645698e-01 1.39675438e-01
-4.86296773e-01 9.30410862e-01 7.00137138e-01 1.11353517e+00
4.84132737e-01 -9.02702689e-01 -1.18027866e+00 9.85695422e-01
6.77038252e-01 4.15675938e-01 6.08326614e-01 -1.02773821e+00
5.31582654e-01 3.19911391e-01 2.67601401e-01 -7.70210683e-01
-2.64376879e-01 -4.36469793e-01 -3.83919686e-01 -6.21332407e-01
8.91523436e-02 3.69341880e-01 -9.79571223e-01 1.62678361e+00
-4.83060405e-02 -3.66656005e-01 9.33861881e-02 8.27662528e-01
1.10575569e+00 5.11087894e-01 4.61659938e-01 2.76035428e-01
1.59215271e+00 -1.16137493e+00 -4.90584433e-01 -3.33986640e-01
8.90610218e-01 -1.04482377e+00 1.16317141e+00 1.25835791e-01
-1.23808718e+00 -3.65447015e-01 -6.15622759e-01 -7.82856226e-01
-5.23080826e-01 -2.74984300e-01 9.52150404e-01 1.99441016e-01
-1.59075224e+00 3.98236096e-01 -2.58328974e-01 -5.18915653e-01
5.55869378e-02 1.42810270e-01 -1.01079397e-01 -5.73101938e-01
-1.82092643e+00 9.78813231e-01 2.00078804e-02 5.12251034e-02
-8.68962049e-01 -4.75154459e-01 -5.41638792e-01 2.01301649e-01
2.13295326e-01 -1.01387155e+00 1.74845862e+00 -3.88283998e-01
-1.08331156e+00 7.47277915e-01 -3.90795261e-01 -4.11916286e-01
-2.64589712e-02 -5.58143675e-01 -3.99541438e-01 1.64174527e-01
-6.76132068e-02 7.57393122e-01 2.52020717e-01 -7.45448828e-01
-4.51419652e-01 -3.24180335e-01 6.76579952e-01 5.18063605e-01
-1.65129051e-01 6.39431596e-01 -7.31927812e-01 -5.24798393e-01
-1.10583395e-01 -6.98203087e-01 -3.10574174e-01 -1.91612691e-01
-8.12728405e-02 -6.52895451e-01 2.88277864e-01 -7.79654920e-01
1.38344467e+00 -1.43160081e+00 7.92315509e-03 1.78112179e-01
3.94806154e-02 3.71014714e-01 -8.41752589e-01 1.01880455e+00
-6.01200722e-02 5.57889819e-01 -2.26973891e-01 -5.08918762e-01
2.00410590e-01 -7.14866370e-02 -5.64439237e-01 -1.86546013e-01
2.21605331e-01 1.59970975e+00 -1.10560918e+00 -4.44440007e-01
-2.82544762e-01 3.41608584e-01 -5.00389516e-01 1.72136620e-01
-5.07554114e-01 -1.30658552e-01 -4.50620592e-01 7.72106946e-01
3.46036196e-01 -3.22854221e-01 9.73852426e-02 6.93639964e-02
3.47433031e-01 1.05809844e+00 -4.57586467e-01 1.78632855e+00
-5.16265452e-01 4.45476949e-01 -1.91948995e-01 -5.44956803e-01
5.69684565e-01 6.63010240e-01 1.58606946e-01 -1.11723161e+00
-3.37928832e-01 1.95064992e-01 -4.58225071e-01 -3.50504130e-01
1.11560345e+00 4.65765029e-01 -2.47303829e-01 6.52852774e-01
1.03090322e-02 -2.07115695e-01 4.29336041e-01 8.59714031e-01
1.52979016e+00 -2.21692584e-02 2.64984369e-02 -3.11277866e-01
6.54690862e-01 -2.36187026e-01 -2.06505775e-01 1.18837750e+00
-1.63027033e-01 8.53973508e-01 2.49199301e-01 -1.40880793e-01
-6.69910014e-01 -9.74903524e-01 -1.48546875e-01 1.69137824e+00
-9.38486159e-02 -7.89621592e-01 -3.43080491e-01 -9.13907051e-01
-7.35254446e-03 4.84183520e-01 -9.04877782e-01 -3.14965814e-01
-6.88226163e-01 -4.66480970e-01 4.77502286e-01 4.44906861e-01
-6.27788082e-02 -1.17868805e+00 7.98905082e-03 1.78879142e-01
-5.45345664e-01 -7.07335711e-01 -7.93182015e-01 4.63732285e-03
-9.29889321e-01 -9.39904809e-01 -7.60649383e-01 -4.85036641e-01
8.93630832e-02 5.03217757e-01 1.85638630e+00 4.14053351e-01
-1.21997513e-01 7.93255031e-01 -8.31683576e-01 -2.04287082e-01
-5.31923957e-02 2.21154496e-01 -5.12653351e-01 -7.31717229e-01
4.97258782e-01 -2.96930730e-01 -9.79329824e-01 1.92061067e-01
-1.24573457e+00 -5.59090197e-01 5.46721697e-01 8.54572713e-01
5.14610171e-01 -8.67422879e-01 7.80157328e-01 -7.53514647e-01
1.23178041e+00 -6.98728025e-01 -2.73160487e-01 8.43124270e-01
-8.44020307e-01 9.23409536e-02 1.39472216e-01 -1.17639840e-01
-8.26246679e-01 -7.11971104e-01 -4.14503932e-01 9.81248319e-02
1.61543921e-01 6.05500400e-01 8.23478922e-02 7.71313980e-02
8.79947960e-01 -1.81885198e-01 -7.87611425e-01 -6.38809621e-01
7.34216332e-01 7.27574944e-01 4.83134806e-01 -7.76715755e-01
5.11254251e-01 4.54257801e-02 -2.92521477e-01 1.59302995e-01
-9.55337703e-01 -1.07096934e+00 -3.93203944e-01 1.50815532e-01
6.90580726e-01 -9.10793006e-01 -5.62633812e-01 3.87614481e-02
-1.53932238e+00 9.20717940e-02 -1.15026668e-01 1.98947370e-01
-2.59000957e-01 5.19860566e-01 -1.00746262e+00 -6.32053256e-01
-8.45712900e-01 -7.88496614e-01 1.37920856e+00 1.60908047e-02
-4.42445874e-01 -1.16149843e+00 5.53821981e-01 4.75477457e-01
8.00798595e-01 -2.76994705e-01 9.94019926e-01 -7.95720339e-01
-5.94520986e-01 -4.11554515e-01 -3.35882425e-01 1.36884004e-01
-1.42747670e-01 -1.95548937e-01 -1.09292483e+00 -3.09078991e-01
-3.28134120e-01 -7.18738556e-01 1.51026094e+00 1.25453696e-01
1.25557315e+00 -4.97892797e-01 -2.82608658e-01 1.60240054e-01
1.13776100e+00 2.74858568e-02 8.62866104e-01 5.86336851e-01
3.06799382e-01 6.71483099e-01 7.57860065e-01 -4.20819484e-02
9.15613770e-01 6.73527956e-01 6.91608429e-01 2.12891608e-01
1.02372259e-01 -1.87615514e-01 2.29195237e-01 6.63491428e-01
-8.97965208e-02 -7.96723008e-01 -7.02053666e-01 6.86484873e-01
-2.08850408e+00 -1.02338517e+00 -2.02965811e-01 2.22221518e+00
1.15477037e+00 -1.82817623e-01 -1.90319940e-01 -2.92210817e-01
4.09062922e-01 2.35597506e-01 -2.24234357e-01 -7.75354266e-01
-2.17178091e-01 5.61690152e-01 1.75481871e-01 7.80624092e-01
-1.17042053e+00 8.67911577e-01 6.84361553e+00 7.14172781e-01
-2.10901290e-01 2.07092427e-02 5.87549806e-01 -1.07936196e-01
-9.85643566e-01 2.08095700e-01 -6.13771021e-01 1.65935919e-01
1.26396036e+00 -3.17459740e-02 3.63680035e-01 4.32437867e-01
-6.94116890e-01 -1.31706059e-01 -1.20801556e+00 4.44196135e-01
2.61888802e-01 -1.14684284e+00 2.38192603e-01 -2.12638736e-01
6.53078258e-01 3.45560491e-01 1.95983574e-01 8.22342038e-01
8.73018742e-01 -1.26411688e+00 3.39146525e-01 7.95582235e-01
1.76520988e-01 -4.42939401e-01 7.66598523e-01 -2.31507514e-03
-8.64148557e-01 -1.34380117e-01 -5.73914587e-01 1.09773159e-01
-6.39732704e-02 5.36505818e-01 -6.49609089e-01 7.85708249e-01
8.33026826e-01 4.09353256e-01 -1.17199278e+00 9.36831176e-01
-2.60026604e-01 6.54173017e-01 -4.51168232e-02 -4.25939374e-02
2.61846751e-01 3.49841356e-01 3.58350992e-01 1.37432122e+00
1.35261901e-02 -4.33785096e-02 -2.00030297e-01 5.17657936e-01
-3.28687400e-01 2.14712203e-01 -2.01705173e-01 -2.07652319e-02
5.90941131e-01 1.38754761e+00 4.22813557e-02 -4.98641521e-01
-3.25511515e-01 7.82475412e-01 5.28183579e-01 6.34060204e-01
-2.57952154e-01 -4.56536263e-01 4.10892040e-01 -2.16379225e-01
1.08329758e-01 1.15433723e-01 6.91265538e-02 -1.02653110e+00
2.24939212e-01 -9.49633002e-01 8.32809150e-01 -7.70591080e-01
-1.63441777e+00 6.33947551e-01 1.45669192e-01 -8.43587935e-01
-6.94547296e-01 -3.09251457e-01 -6.07078195e-01 1.50938964e+00
-1.84720874e+00 -1.34279525e+00 5.86853288e-02 3.85936201e-01
2.94227481e-01 1.69633359e-01 1.01651943e+00 1.19624682e-01
5.74588962e-02 7.25960255e-01 3.38129848e-01 -1.62159041e-01
1.10979378e+00 -1.28466058e+00 1.30858588e+00 4.20747876e-01
2.04488754e-01 1.40501380e+00 3.13045025e-01 -5.12102783e-01
-1.38805938e+00 -1.04162741e+00 1.53458965e+00 -1.18211913e+00
4.73122388e-01 -2.76874572e-01 -1.02416086e+00 7.15868890e-01
6.31742358e-01 -1.25144750e-01 6.81480527e-01 5.48166931e-01
-9.29462671e-01 2.23276302e-01 -9.43510592e-01 6.75416887e-01
9.78456378e-01 -9.01991904e-01 -7.28420794e-01 6.33603811e-01
1.20062506e+00 -4.55750525e-02 -8.45882475e-01 8.43311131e-01
7.38536060e-01 -7.98911572e-01 1.38418710e+00 -9.82627749e-01
5.81635892e-01 2.11035401e-01 -5.29290676e-01 -1.13141477e+00
-6.47914052e-01 -1.95042416e-01 -4.88273382e-01 1.12043226e+00
6.20653152e-01 -3.65188658e-01 2.37145483e-01 7.47011185e-01
-1.90314427e-01 -8.93349349e-01 -6.69925153e-01 -5.43094337e-01
6.11781299e-01 -2.90670127e-01 8.14783037e-01 7.25419641e-01
1.47127211e-01 5.03210425e-01 1.22077040e-01 -2.20215037e-01
9.03615579e-02 -7.53106102e-02 3.98398280e-01 -1.20502388e+00
-6.34549022e-01 -5.98028898e-01 1.72690392e-01 -1.08712113e+00
-1.68136824e-02 -1.03292084e+00 7.40749985e-02 -2.00855398e+00
5.43973386e-01 -2.62078911e-01 -9.14881766e-01 2.82364875e-01
-7.80181885e-01 1.77212223e-01 7.58887529e-02 2.36610904e-01
-1.24931991e+00 4.15654331e-01 1.13029397e+00 -3.37597907e-01
2.24169284e-01 -5.83199523e-02 -9.56997573e-01 1.82052955e-01
5.91725290e-01 -2.39368752e-01 -5.19515991e-01 -1.06019294e+00
8.09032023e-01 4.93436269e-02 3.96756828e-01 -3.24050963e-01
4.17050362e-01 -1.25374928e-01 4.12735909e-01 -7.04497278e-01
3.78134608e-01 -3.69360685e-01 -4.08937722e-01 6.85144961e-02
-1.19118607e+00 4.54999357e-01 2.25169718e-01 5.74461997e-01
-2.49598011e-01 -5.42628646e-01 1.38863221e-01 -3.49704891e-01
-2.85009831e-01 3.73312145e-01 -2.53133714e-01 3.33631605e-01
9.03293565e-02 3.55381399e-01 -9.59339380e-01 -6.72032535e-01
-1.59793139e-01 7.89436638e-01 1.74488679e-01 7.59373903e-01
5.61623335e-01 -1.40530157e+00 -7.81365395e-01 -4.09420639e-01
6.24846697e-01 -1.17570519e-01 2.97166646e-01 4.71035779e-01
4.20821942e-02 1.16401482e+00 4.54598516e-01 -2.11459577e-01
-1.05583012e+00 4.96461600e-01 -2.06129178e-02 -9.56525862e-01
1.86254576e-01 1.09094644e+00 2.12010950e-01 -7.27391124e-01
1.86010018e-01 -2.03128844e-01 -5.80915749e-01 -1.28249481e-01
9.14668500e-01 2.16716722e-01 5.56816518e-01 -1.75516680e-01
-4.35169846e-01 2.68309861e-01 -6.83406591e-01 -5.10408461e-01
9.23730850e-01 -2.89592683e-01 -2.65994668e-01 -1.91136336e-04
1.49663377e+00 -1.75394759e-01 -2.98601538e-01 -7.43133843e-01
6.24961615e-01 -3.48967463e-01 3.60656250e-03 -1.39007926e+00
-3.59923363e-01 9.48900700e-01 4.35448617e-01 7.49945343e-02
9.86726701e-01 1.71257436e-01 7.19141901e-01 9.37879801e-01
2.04853490e-01 -8.52365971e-01 1.81116492e-01 1.00783110e+00
1.38625228e+00 -1.22890770e+00 -2.40649693e-02 2.28167772e-01
-3.66821229e-01 8.29387426e-01 5.54207146e-01 2.57082790e-01
4.15450543e-01 -3.88273150e-01 -1.12193324e-01 -3.06922406e-01
-1.62282503e+00 -4.49170500e-01 9.09376383e-01 4.99450445e-01
9.02559757e-01 -2.00968087e-01 -6.03513479e-01 4.03067857e-01
9.67985094e-02 -8.37310702e-02 5.80229126e-02 1.17130768e+00
-5.06396949e-01 -1.46928275e+00 -2.22028449e-01 5.70743263e-01
-6.26363695e-01 -8.66352081e-01 -7.68858612e-01 3.63068759e-01
-5.20746648e-01 1.24160886e+00 -5.60922436e-02 -4.67648357e-01
3.61063838e-01 3.07158917e-01 3.42616677e-01 -8.82056475e-01
-1.16853249e+00 -2.35730171e-01 3.92974794e-01 -5.38290918e-01
-2.14976504e-01 -5.02129555e-01 -7.70841420e-01 5.29151559e-02
-4.21296895e-01 4.50492799e-01 5.80050588e-01 7.75754988e-01
5.99451244e-01 2.73946047e-01 6.22729123e-01 -3.23608100e-01
-7.75148451e-01 -1.53494370e+00 -6.93919361e-02 4.71462220e-01
3.56112808e-01 -2.39267528e-01 -3.66966337e-01 -3.23023677e-01] | [11.513283729553223, 7.7100019454956055] |
ff279b30-c2a8-456f-850e-a723efdb4b70 | weakly-supervised-attentional-model-for-low | null | null | https://aclanthology.org/D19-6129 | https://aclanthology.org/D19-6129.pdf | Weakly Supervised Attentional Model for Low Resource Ad-hoc Cross-lingual Information Retrieval | We propose a weakly supervised neural model for Ad-hoc Cross-lingual Information Retrieval (CLIR) from low-resource languages. Low resource languages often lack relevance annotations for CLIR, and when available the training data usually has limited coverage for possible queries. In this paper, we design a model which does not require relevance annotations, instead it is trained on samples extracted from translation corpora as weak supervision. This model relies on an attention mechanism to learn spans in the foreign sentence that are relevant to the query. We report experiments on two low resource languages: Swahili and Tagalog, trained on less that 100k parallel sentences each. The proposed model achieves 19 MAP points improvement compared to using CNNs for feature extraction, 12 points improvement from machine translation-based CLIR, and up to 6 points improvement compared to probabilistic CLIR models. | ['Zhongqiang Huang', 'Damianos Karakos', 'Lingjun Zhao', 'Zhuolin Jiang', 'Rabih Zbib'] | 2019-11-01 | null | null | null | ws-2019-11 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-2.08476558e-01 7.25376457e-02 -6.32129848e-01 -3.80049318e-01
-2.00586033e+00 -6.24309421e-01 9.08332348e-01 2.97898561e-01
-9.27444041e-01 9.69346881e-01 5.13274550e-01 -3.14817876e-01
-2.72461064e-02 -3.82088900e-01 -9.81649995e-01 -1.97655812e-01
3.94827157e-01 1.16714597e+00 -4.61529009e-03 -5.43932796e-01
-9.04001296e-02 1.88058838e-01 -8.67372036e-01 6.02344811e-01
6.53813004e-01 8.43458235e-01 3.20936203e-01 4.72098023e-01
-7.16141909e-02 9.01015341e-01 -4.25003588e-01 -4.71326768e-01
1.08680546e-01 2.57920586e-02 -1.30208087e+00 -6.08038187e-01
6.58394873e-01 -2.32005447e-01 -2.40431264e-01 8.24009955e-01
5.64940810e-01 -3.34909856e-02 7.45263875e-01 -4.37254488e-01
-1.33087313e+00 9.89822149e-01 -3.68481427e-01 4.14028317e-01
2.99546480e-01 -4.78298426e-01 1.51958334e+00 -1.17147779e+00
8.47857714e-01 1.12408364e+00 5.15328288e-01 6.42939150e-01
-1.00589669e+00 -4.17155653e-01 -6.39540702e-02 1.75194740e-01
-1.55571938e+00 -5.08690119e-01 4.83623564e-01 -1.02084376e-01
1.46665621e+00 1.92500606e-01 -1.12100281e-01 1.11904824e+00
4.03240956e-02 1.12857234e+00 6.59096539e-01 -8.78623188e-01
-2.69951522e-01 4.75322872e-01 2.09816173e-01 3.12113672e-01
1.21121526e-01 -2.09626853e-01 -5.23152888e-01 -2.67369777e-01
3.88089806e-01 -3.36179823e-01 1.03514373e-01 1.65402904e-01
-1.17491925e+00 9.30295348e-01 4.88998532e-01 5.25245428e-01
-2.99092621e-01 1.15769759e-01 3.52646083e-01 5.47611177e-01
1.08484781e+00 7.23272562e-01 -9.46891308e-01 3.39419127e-01
-8.30842197e-01 1.14066094e-01 6.03661418e-01 1.32571673e+00
8.69792044e-01 -5.16842365e-01 -5.18559933e-01 1.19172120e+00
3.54558408e-01 8.73133600e-01 4.87093121e-01 -4.69469696e-01
7.93813407e-01 3.25021476e-01 2.23215565e-01 -4.65948313e-01
-9.94035378e-02 -5.57339609e-01 -5.59485614e-01 -6.81721389e-01
3.92483287e-02 -1.99738175e-01 -6.91775739e-01 1.43035305e+00
-1.09633505e-01 -6.62813842e-01 3.77187103e-01 9.79451835e-01
9.01277602e-01 7.19463825e-01 -3.10132746e-02 -2.29702275e-02
1.29997766e+00 -1.18609142e+00 -6.39566720e-01 -4.60886955e-01
8.39534938e-01 -1.14755905e+00 1.24472487e+00 5.48779517e-02
-1.13579965e+00 -3.33797336e-01 -7.45629728e-01 -5.54433525e-01
-4.62060750e-01 4.54307854e-01 3.68751466e-01 -6.17144555e-02
-1.43073618e+00 1.00274213e-01 -4.99676257e-01 -4.43015307e-01
-3.46307009e-02 3.50002348e-01 -4.29060280e-01 -4.65705514e-01
-1.50957644e+00 1.02736032e+00 2.11792529e-01 3.29774432e-02
-6.98823094e-01 -3.70766252e-01 -8.67241740e-01 1.32711694e-01
1.81348339e-01 -3.91897380e-01 1.41960275e+00 -1.15082586e+00
-1.15935874e+00 1.00615418e+00 -4.23124522e-01 -5.12873232e-01
2.90271789e-01 -6.19186044e-01 -3.17760855e-01 6.48542717e-02
4.40823466e-01 6.78616285e-01 4.09503371e-01 -7.83233821e-01
-5.34731925e-01 -2.43980184e-01 4.79709655e-02 4.51008230e-01
-3.75681460e-01 9.32972610e-01 -1.08226204e+00 -3.64861786e-01
-2.84572482e-01 -8.23436975e-01 -1.31649479e-01 -6.04138970e-01
-3.26913655e-01 -7.28852749e-01 4.58803624e-01 -9.14624333e-01
9.07578468e-01 -1.76197362e+00 -1.89268172e-01 -1.42712981e-01
-2.98244596e-01 2.00762570e-01 -4.73764300e-01 5.39362133e-01
3.80316049e-01 3.15547705e-01 1.09763227e-01 -3.16222221e-01
-7.65944496e-02 4.34124470e-02 -4.75515366e-01 8.27869400e-02
7.44961560e-01 1.32942653e+00 -9.11014676e-01 -5.52136123e-01
-3.21804255e-01 5.95208406e-01 -3.69656295e-01 1.90757260e-01
-4.04750079e-01 2.90652424e-01 -6.56645954e-01 6.32952511e-01
2.45914608e-01 -4.00331765e-01 2.57694572e-01 6.12003319e-02
2.43966673e-02 8.69567156e-01 -1.39636308e-01 1.78330469e+00
-8.09850335e-01 7.49042153e-01 -1.56743392e-01 -8.87952924e-01
1.00765371e+00 5.67109048e-01 2.00449005e-01 -1.14516270e+00
-1.08721562e-01 4.05236363e-01 -4.46451545e-01 -1.97581396e-01
8.00435185e-01 2.40093648e-01 -3.18909168e-01 5.47466457e-01
2.31333613e-01 1.90610960e-01 1.40340492e-01 1.73289031e-01
1.06610847e+00 2.20509306e-01 3.87685522e-02 -5.17890692e-01
6.11238241e-01 3.48810822e-01 3.85205418e-01 9.18336749e-01
1.31471917e-01 5.12248099e-01 -1.27687484e-01 -6.33246779e-01
-1.10456121e+00 -6.60298049e-01 -3.68666679e-01 1.47365713e+00
-1.33280531e-01 -1.66498646e-01 -4.59338218e-01 -8.62649381e-01
-4.71053272e-01 4.95972961e-01 -3.79066080e-01 1.35879666e-01
-8.04493129e-01 -7.72570789e-01 5.01721025e-01 4.46981788e-01
2.61110485e-01 -1.27096057e+00 1.38739854e-01 1.92639798e-01
-5.78685105e-01 -1.36338520e+00 -7.31248617e-01 3.30639154e-01
-6.45318687e-01 -4.84470218e-01 -8.66862535e-01 -1.11288416e+00
5.28861165e-01 2.41334006e-01 1.62293875e+00 -5.69530614e-02
-5.93423508e-02 2.55073249e-01 -4.63256925e-01 -4.03508544e-01
-2.34184667e-01 6.89289808e-01 1.02747612e-01 -4.84867305e-01
1.21255124e+00 2.27140367e-01 -2.62234837e-01 1.23783022e-01
-5.62838793e-01 -2.96763420e-01 8.63882661e-01 1.13883817e+00
8.70940983e-01 -4.57637817e-01 6.25863075e-01 -8.01088214e-01
6.94795549e-01 -5.25136292e-01 -6.64299846e-01 6.44419551e-01
-7.09050655e-01 3.37524801e-01 5.46831071e-01 -2.75571167e-01
-1.16707397e+00 -1.12497844e-01 2.02419739e-02 -3.38240564e-01
-6.04758188e-02 9.38991249e-01 1.36645645e-01 3.18639129e-01
8.73160005e-01 2.39797365e-02 -6.00509286e-01 -7.69349992e-01
3.63083988e-01 1.13190639e+00 1.80135608e-01 -8.16620052e-01
5.44689953e-01 9.58044901e-02 -6.48592114e-01 -6.86624885e-01
-1.28446293e+00 -7.36855388e-01 -7.97771156e-01 3.51619333e-01
8.52974176e-01 -1.34259522e+00 -1.88200966e-01 -3.39700699e-01
-1.37279546e+00 -2.68857032e-01 2.02386137e-02 7.53293514e-01
-3.32639813e-01 -1.74582861e-02 -1.07028639e+00 -6.38421655e-01
-9.38026309e-01 -8.83567870e-01 1.44106054e+00 -9.79434550e-02
-8.71421248e-02 -1.09271002e+00 3.69262546e-01 6.17527902e-01
6.29157603e-01 -6.83042467e-01 7.96413660e-01 -1.06401992e+00
-6.59285009e-01 -4.68646139e-01 -4.32267755e-01 4.22483563e-01
1.10283487e-01 -4.31113094e-01 -1.19278109e+00 -3.47945660e-01
-3.91685456e-01 -8.31309617e-01 8.10083508e-01 1.51868269e-01
7.10917890e-01 -6.97885156e-01 -2.48689547e-01 8.98256674e-02
1.40845811e+00 -1.68048859e-01 3.19937080e-01 3.77796561e-01
6.20455861e-01 7.72436321e-01 6.96592927e-01 -3.13024461e-01
4.82444316e-01 1.06511045e+00 -2.71159768e-01 -2.62629539e-01
-8.47381055e-02 -3.60148609e-01 5.55162847e-01 1.20001209e+00
1.40057415e-01 -2.28725985e-01 -9.79441345e-01 9.32719290e-01
-1.93138671e+00 -7.17253327e-01 1.31973878e-01 2.17116427e+00
1.26210129e+00 -1.81758985e-01 -3.34954262e-01 -6.86733127e-01
4.51950163e-01 -3.90994072e-01 -2.96257049e-01 -3.91952872e-01
-3.27473789e-01 3.44486177e-01 5.20691574e-01 8.99992108e-01
-1.13039005e+00 1.45523047e+00 6.66463375e+00 8.65819216e-01
-1.17690969e+00 4.31539088e-01 6.84800982e-01 6.70960546e-02
-4.22802389e-01 -1.51677176e-01 -1.24315655e+00 -1.67693183e-01
1.09401405e+00 -1.07021339e-01 3.49072635e-01 7.99426258e-01
1.35560511e-02 3.34111542e-01 -1.09687591e+00 6.19622946e-01
4.35654581e-01 -1.21298444e+00 1.05710439e-01 2.93804053e-03
8.44087660e-01 9.57892776e-01 6.12333044e-02 4.93750781e-01
6.86793864e-01 -1.00560904e+00 4.52861190e-01 3.14018548e-01
9.23766851e-01 -7.45315731e-01 1.21336591e+00 5.24550736e-01
-7.50692725e-01 4.39775437e-01 -8.88182521e-01 2.31762245e-01
-2.12104812e-01 2.76719600e-01 -1.28524601e+00 6.44858658e-01
8.54485333e-01 7.08450258e-01 -5.90294659e-01 3.92490596e-01
-3.58637393e-01 7.14515448e-01 -2.37049252e-01 -2.17687055e-01
4.70531791e-01 1.64500743e-01 2.03934759e-01 1.48952484e+00
2.15935796e-01 -5.20010412e-01 2.60204285e-01 7.25866139e-01
-4.69728023e-01 7.03207612e-01 -8.30323875e-01 -1.20321266e-01
1.95485726e-01 1.22803462e+00 -6.82414547e-02 -4.71472859e-01
-7.17163146e-01 1.10710168e+00 7.79964685e-01 5.25449336e-01
-2.42708758e-01 -3.18572223e-01 3.11401725e-01 -5.86286895e-02
1.96969375e-01 4.38754894e-02 5.00900373e-02 -1.50307786e+00
1.98657170e-01 -8.51095796e-01 5.18701911e-01 -8.13819945e-01
-1.77456343e+00 1.01694441e+00 -1.27563655e-01 -1.05392122e+00
-8.73741448e-01 -6.28763676e-01 -1.09631736e-02 1.53663146e+00
-1.95557141e+00 -1.64135194e+00 5.40613115e-01 5.83035827e-01
7.45472729e-01 -3.49040061e-01 1.11057711e+00 3.94760489e-01
-1.46514744e-01 8.15560162e-01 3.69469941e-01 5.06511390e-01
1.00428569e+00 -1.02003849e+00 4.68383759e-01 6.95992529e-01
5.73728502e-01 9.34734881e-01 2.84009427e-01 -4.76571679e-01
-1.35124218e+00 -1.27495730e+00 2.00822091e+00 -9.54767585e-01
6.49554968e-01 -4.48154747e-01 -8.60113502e-01 9.59781229e-01
5.69563746e-01 4.21046615e-02 6.60416961e-01 8.03599775e-01
-6.78001702e-01 -1.33014262e-01 -9.09333050e-01 4.01917845e-01
5.17803311e-01 -1.06834853e+00 -7.45617986e-01 7.66033173e-01
1.10887313e+00 -1.71396900e-02 -7.57237673e-01 3.69029701e-01
3.58411014e-01 -1.40530998e-02 1.02636945e+00 -9.20791268e-01
4.61374909e-01 -1.15152150e-01 -4.38088477e-01 -1.07999992e+00
-4.10985410e-01 -2.04786777e-01 1.84148714e-01 1.00805545e+00
1.13672137e+00 -1.41580120e-01 3.54260862e-01 5.27812123e-01
-2.36056317e-02 -3.72737765e-01 -8.73483658e-01 -8.14747214e-01
5.72298527e-01 -5.67936301e-01 3.45816553e-01 1.06918955e+00
2.41493136e-01 1.13969076e+00 -4.08130586e-01 1.49638295e-01
4.60436642e-01 -6.01025186e-02 4.95102018e-01 -1.03623939e+00
-5.10834038e-01 -2.36790776e-01 1.07387252e-01 -1.04071510e+00
5.67175865e-01 -1.23709118e+00 2.88201630e-01 -1.55069554e+00
5.76457620e-01 -5.44940770e-01 -5.50139308e-01 7.73361444e-01
-2.19810247e-01 4.80661839e-01 -1.83957368e-01 5.09216845e-01
-9.04130697e-01 3.84152234e-01 7.96580911e-01 -5.18622398e-01
1.70274085e-04 -2.90449321e-01 -7.74655819e-01 3.45714748e-01
6.03669822e-01 -4.92903471e-01 -1.25267521e-01 -1.30356550e+00
5.33022881e-01 2.16609333e-03 -8.86936262e-02 -4.14743513e-01
1.20700613e-01 -7.00926930e-02 4.18056011e-01 -5.38391531e-01
1.75553903e-01 -6.28315926e-01 -4.90377814e-01 -4.70236354e-02
-1.01983070e+00 1.26048729e-01 2.26978809e-01 2.69380301e-01
-4.67418820e-01 -1.56053558e-01 2.92351365e-01 -3.26433569e-01
-3.48264545e-01 3.66830677e-01 -2.94706523e-01 8.22052956e-02
2.30464637e-02 7.92670190e-01 -2.84425735e-01 -5.47895312e-01
-4.32480782e-01 2.59008884e-01 2.80357838e-01 7.69820452e-01
3.16721231e-01 -1.38403273e+00 -1.17090881e+00 -2.37703566e-02
7.18514621e-01 -1.45059541e-01 -2.26779610e-01 5.02910495e-01
-1.69512838e-01 1.26393259e+00 3.85875136e-01 -4.34441984e-01
-1.03365052e+00 6.61653280e-01 -5.01108989e-02 -7.44203091e-01
-3.27496439e-01 7.81660557e-01 3.00911754e-01 -9.82248306e-01
-1.75769217e-02 5.39355762e-02 -3.84479672e-01 -2.74622440e-01
8.47784936e-01 -2.85855144e-01 5.45684516e-01 -7.75299788e-01
-2.90121168e-01 3.57249141e-01 -5.17959118e-01 -6.05523705e-01
1.11710489e+00 -3.45704556e-01 -2.74106711e-01 5.89301467e-01
1.25472617e+00 8.48082975e-02 -5.01698017e-01 -1.06138694e+00
5.64711750e-01 -1.37893230e-01 2.49495998e-01 -1.14832163e+00
-6.21881843e-01 9.86856282e-01 3.74944776e-01 -2.28598565e-01
8.38509202e-01 5.53143799e-01 6.22088134e-01 1.15279651e+00
6.70363128e-01 -1.14448643e+00 -1.85114190e-01 1.11620653e+00
1.05669582e+00 -1.49444711e+00 -4.42827553e-01 1.52546927e-01
-5.30143678e-01 8.41630101e-01 4.17481273e-01 3.30305519e-03
4.65347081e-01 7.45311454e-02 4.49261934e-01 -2.00543225e-01
-1.19504368e+00 -3.93163621e-01 8.53844464e-01 4.44731474e-01
1.05875790e+00 1.07693173e-01 -3.61515015e-01 4.55549806e-01
-3.27812284e-02 -1.43442884e-01 6.34729266e-02 6.62398696e-01
-2.78919160e-01 -1.24035966e+00 -9.33981463e-02 2.48555616e-01
-9.25257444e-01 -8.62332821e-01 -6.03002489e-01 3.84956151e-01
-4.60702330e-01 1.07310116e+00 1.49814665e-01 -2.04800859e-01
-3.34573351e-02 2.55986691e-01 1.92503586e-01 -8.35509181e-01
-6.23542964e-01 3.79731774e-01 3.71832401e-01 -3.63461554e-01
-3.53432029e-01 -5.21772981e-01 -7.39046574e-01 1.78527653e-01
-3.72348458e-01 5.15152931e-01 7.89913058e-01 1.09163165e+00
4.67721492e-01 -7.64787421e-02 5.99704683e-01 -2.94559687e-01
-4.15753990e-01 -1.41614282e+00 -3.69812958e-02 7.19519258e-02
2.90488839e-01 -2.22107600e-02 -1.11783236e-01 -3.53856198e-02] | [11.422146797180176, 9.83120346069336] |
2b97490e-26c8-4bdd-950d-8105d8442bbd | multi-modal-multi-level-fusion-for-3d-single | 2305.06794 | null | https://arxiv.org/abs/2305.06794v1 | https://arxiv.org/pdf/2305.06794v1.pdf | Multi-modal Multi-level Fusion for 3D Single Object Tracking | 3D single object tracking plays a crucial role in computer vision. Mainstream methods mainly rely on point clouds to achieve geometry matching between target template and search area. However, textureless and incomplete point clouds make it difficult for single-modal trackers to distinguish objects with similar structures. To overcome the limitations of geometry matching, we propose a Multi-modal Multi-level Fusion Tracker (MMF-Track), which exploits the image texture and geometry characteristic of point clouds to track 3D target. Specifically, we first propose a Space Alignment Module (SAM) to align RGB images with point clouds in 3D space, which is the prerequisite for constructing inter-modal associations. Then, in feature interaction level, we design a Feature Interaction Module (FIM) based on dual-stream structure, which enhances intra-modal features in parallel and constructs inter-modal semantic associations. Meanwhile, in order to refine each modal feature, we introduce a Coarse-to-Fine Interaction Module (CFIM) to realize the hierarchical feature interaction at different scales. Finally, in similarity fusion level, we propose a Similarity Fusion Module (SFM) to aggregate geometry and texture clues from the target. Experiments show that our method achieves state-of-the-art performance on KITTI (39% Success and 42% Precision gains against previous multi-modal method) and is also competitive on NuScenes. | ['Zheng Fang', 'Zuoxu Gu', 'Yubo Cui', 'Zhiheng Li'] | 2023-05-11 | null | null | null | null | ['3d-single-object-tracking'] | ['computer-vision'] | [-2.32480720e-01 -6.55609190e-01 1.67514514e-02 -4.92090778e-03
-8.88799965e-01 -7.35961258e-01 8.73466134e-01 -9.76343974e-02
-1.24425255e-01 -1.47433728e-01 -2.28229612e-01 -3.86131257e-02
-1.77844077e-01 -6.92090571e-01 -6.54318750e-01 -7.44239748e-01
3.01913828e-01 4.98751849e-01 8.85212481e-01 -1.68877855e-01
2.30926245e-01 7.90567756e-01 -1.72688985e+00 -9.62158963e-02
6.48736238e-01 1.24763489e+00 4.37649220e-01 3.32999915e-01
-3.52989346e-01 9.32886153e-02 -3.38349849e-01 -1.71470076e-01
5.26690483e-01 -5.03388196e-02 -1.52813777e-01 2.45619655e-01
7.16655910e-01 -1.70911402e-02 -2.55381703e-01 1.36135602e+00
4.51191634e-01 -4.41587120e-02 5.83433390e-01 -1.36139154e+00
-3.16889077e-01 -9.52468589e-02 -9.60127771e-01 2.22881243e-01
3.61419469e-01 3.92242432e-01 6.59278512e-01 -1.06760585e+00
2.08549291e-01 1.39810348e+00 7.94747114e-01 2.73674130e-01
-8.78138244e-01 -8.99996161e-01 2.20441729e-01 8.43031481e-02
-1.44360507e+00 -2.13711664e-01 8.14420760e-01 -3.33693087e-01
5.34531534e-01 4.25697178e-01 7.94500113e-01 5.18434405e-01
2.01400921e-01 5.61459363e-01 9.65026498e-01 -1.86224297e-01
-1.79631874e-01 -1.11076072e-01 -8.04339945e-02 5.47685921e-01
4.48141694e-01 5.27333975e-01 -6.99094594e-01 -1.05710655e-01
9.69976604e-01 2.71611422e-01 -2.26672053e-01 -8.25750351e-01
-1.59023583e+00 6.03535652e-01 8.03653479e-01 3.02972645e-01
-3.82919103e-01 1.17172629e-01 -5.01188152e-02 -4.14076336e-02
2.56367534e-01 -1.62391569e-02 -6.61994293e-02 1.87473878e-01
-6.21325731e-01 2.65287220e-01 1.76929966e-01 1.24153066e+00
8.49895358e-01 -1.55928954e-01 -3.00171167e-01 4.14368063e-01
8.79097283e-01 1.23310375e+00 2.61269599e-01 -6.40023887e-01
3.39721859e-01 9.29414690e-01 1.69684380e-01 -1.03971374e+00
-4.71425056e-01 -6.44523859e-01 -5.17771065e-01 4.43839431e-01
1.92396536e-01 3.90508205e-01 -9.42034602e-01 1.17870474e+00
1.01662672e+00 4.07557160e-01 -5.24940901e-02 1.37697887e+00
1.02261090e+00 2.80358911e-01 -2.39169132e-02 9.22240410e-03
1.73487282e+00 -6.55541003e-01 -1.98496014e-01 -1.98524222e-01
5.01232982e-01 -1.11812174e+00 6.81412399e-01 -2.19173217e-03
-8.66053104e-01 -8.55925322e-01 -9.45121706e-01 2.73058563e-01
-2.54240602e-01 2.24807605e-01 4.94275987e-01 4.92066860e-01
-6.70834184e-01 -3.37978341e-02 -8.57065737e-01 -5.26409566e-01
1.74617305e-01 4.47300225e-01 -3.66438985e-01 7.82033652e-02
-7.84879267e-01 9.07037318e-01 5.71340740e-01 1.05765201e-01
-7.01082528e-01 -1.00277209e+00 -8.04293275e-01 -2.79340357e-01
4.53154683e-01 -9.73315001e-01 1.06951642e+00 -3.69816542e-01
-1.30022621e+00 8.63837600e-01 -3.94204520e-02 -4.85996790e-02
2.38613918e-01 -1.80994704e-01 -5.39479852e-01 -1.53416712e-02
1.58863485e-01 5.97252607e-01 8.43111932e-01 -1.32455814e+00
-1.27120483e+00 -6.89102411e-01 -1.16143398e-01 6.48346007e-01
1.02324791e-01 1.11862995e-01 -9.97020364e-01 -3.93495560e-01
8.16414595e-01 -9.26948369e-01 -9.85611305e-02 1.62283719e-01
-2.14552656e-01 -5.22274852e-01 1.15559733e+00 -1.32413432e-01
7.72031248e-01 -2.31074071e+00 1.32628158e-01 2.59179592e-01
3.13513011e-01 1.54247329e-01 5.65406457e-02 -1.87346399e-01
1.71409816e-01 -6.51305914e-01 1.75020501e-01 -2.71926612e-01
7.27832168e-02 4.12125289e-02 -1.50684133e-01 7.66135812e-01
4.11103368e-02 1.03922272e+00 -7.98520565e-01 -7.00059652e-01
6.92777395e-01 5.13470948e-01 -2.39964515e-01 -1.05257802e-01
-2.86147892e-01 5.51938117e-01 -8.67970824e-01 1.13974178e+00
1.09455001e+00 -1.98793039e-01 -3.49223226e-01 -7.67097652e-01
-6.14772081e-01 -1.84493586e-01 -1.39398563e+00 1.93005133e+00
-1.91013694e-01 -7.44104162e-02 1.34371042e-01 -4.40992206e-01
1.09878755e+00 4.15286198e-02 6.31973267e-01 -5.37419677e-01
3.76062214e-01 2.32426316e-01 -2.16959432e-01 -2.73895152e-02
5.40114462e-01 -9.54602263e-04 -2.15642110e-01 2.63420376e-03
-6.23892657e-02 -3.81450921e-01 -2.56661564e-01 6.22888794e-03
8.60369921e-01 2.93240666e-01 -5.20224012e-02 -1.51111692e-01
8.93912554e-01 3.84148151e-01 6.88224077e-01 4.87702966e-01
-2.26374075e-01 5.37752450e-01 -3.91731381e-01 -3.52176845e-01
-7.31851280e-01 -1.40636718e+00 -9.88423675e-02 6.50503814e-01
9.39678848e-01 -2.87949890e-01 -2.15411037e-01 -5.97007930e-01
2.98721492e-01 2.20515132e-02 -3.51830512e-01 -1.86186403e-01
-4.05986607e-01 -5.23491681e-01 1.52026176e-01 5.20526588e-01
6.41315401e-01 -5.54503381e-01 -8.48515391e-01 1.33308051e-02
-1.21017277e-01 -1.14945745e+00 -5.06502092e-01 -1.03013866e-01
-8.60541046e-01 -1.09045196e+00 -4.95063275e-01 -5.03184140e-01
3.75634819e-01 1.06249678e+00 8.39657366e-01 3.37810926e-02
-1.18047044e-01 4.91441131e-01 -4.07255352e-01 -4.15850580e-01
1.39696345e-01 -2.18651265e-01 3.64425659e-01 9.46378261e-02
6.29772544e-01 -3.07093441e-01 -5.90471685e-01 6.83599889e-01
-6.66653514e-01 3.74464616e-02 8.68244410e-01 3.78953993e-01
9.52773511e-01 -2.84659918e-02 -2.72780716e-01 6.89639477e-03
-1.50093749e-01 -1.24177933e-01 -1.21488857e+00 2.60093600e-01
-2.02638149e-01 -1.53413847e-01 4.39620055e-02 -5.08690894e-01
-8.53499711e-01 4.88027930e-01 2.62397259e-01 -1.19218457e+00
-7.22797439e-02 7.81739429e-02 -4.66404319e-01 -7.70376801e-01
5.73446572e-01 3.98402035e-01 -3.58215906e-02 -6.48865402e-01
4.08797324e-01 5.24079382e-01 9.08619523e-01 -5.17599225e-01
1.50115085e+00 7.55264103e-01 2.83121616e-01 -4.45657879e-01
-9.56658542e-01 -8.98961604e-01 -8.13674033e-01 -5.26163638e-01
1.03044140e+00 -1.27918196e+00 -1.00125360e+00 3.99289072e-01
-1.02499473e+00 2.93020993e-01 -1.14012577e-01 8.77825618e-01
-4.05265272e-01 2.98311859e-01 -5.30772023e-02 -5.70928991e-01
-2.73572803e-01 -1.05663466e+00 1.59382021e+00 6.35842085e-01
5.37555873e-01 -5.36418796e-01 1.30575269e-01 2.79725164e-01
2.17180714e-01 2.64530540e-01 5.79076819e-02 -1.88131809e-01
-1.05736005e+00 -1.80214614e-01 -5.53101242e-01 -3.22945595e-01
1.74345776e-01 -2.19744295e-01 -7.72684216e-01 -3.28915924e-01
3.14747877e-02 1.50509760e-01 5.72600842e-01 4.03959572e-01
6.26279056e-01 6.46721303e-01 -8.21108818e-01 8.37247968e-01
1.40262544e+00 1.09734222e-01 3.45901996e-01 6.06823027e-01
1.02689767e+00 2.81708270e-01 1.23362625e+00 3.18943650e-01
5.19848406e-01 1.14541471e+00 5.74010313e-01 7.36019090e-02
-3.76383990e-01 -1.20681763e-01 4.50642467e-01 5.62650263e-01
-1.31523132e-01 5.22951603e-01 -8.04678380e-01 3.14153016e-01
-1.86371446e+00 -6.79284513e-01 -5.08212864e-01 2.25357795e+00
2.72822052e-01 2.90739387e-01 2.24043265e-01 -2.47855589e-01
8.70051503e-01 -1.34394035e-01 -5.31802416e-01 5.37663281e-01
-1.98608905e-01 -1.00117318e-01 7.79506087e-01 2.47862950e-01
-1.30777323e+00 1.06728339e+00 4.81314373e+00 1.19071341e+00
-1.09201074e+00 2.89715677e-01 -4.16519672e-01 1.02505842e-02
-8.17535073e-02 1.18740864e-01 -1.38614452e+00 4.86858130e-01
4.40466255e-01 -1.03811566e-02 -4.35512401e-02 8.65155160e-01
-2.33179465e-01 -4.15034108e-02 -8.25292468e-01 1.29854929e+00
-2.36981884e-02 -1.13712168e+00 -1.58765286e-01 2.11881518e-01
3.02158147e-01 3.09460163e-01 -5.71517050e-02 1.08212367e-01
1.80986330e-01 -3.43980998e-01 9.78676975e-01 5.39061606e-01
6.04000747e-01 -6.41361713e-01 5.51756144e-01 3.69092673e-01
-2.09775758e+00 1.66427776e-01 -5.82567751e-01 4.46696669e-01
1.21352889e-01 4.74149346e-01 -4.69312698e-01 1.21443558e+00
1.00434649e+00 6.77430451e-01 -7.93424010e-01 1.52914393e+00
-2.84530967e-03 -1.73358291e-01 -7.30362356e-01 2.06971854e-01
2.59991884e-01 -6.15543239e-02 7.71913648e-01 7.57979810e-01
5.87957680e-01 1.56630754e-01 5.78674316e-01 6.77747607e-01
4.42080706e-01 -2.86080599e-01 -4.04103190e-01 4.46422607e-01
7.16616869e-01 1.56591749e+00 -7.54630566e-01 -2.19757080e-01
-5.80148935e-01 7.41528690e-01 -1.19805507e-01 -1.45231843e-01
-1.03887808e+00 -7.27485716e-02 8.32983673e-01 -1.33587599e-01
4.58243400e-01 -4.14901376e-01 -1.13694951e-01 -1.12382150e+00
1.16186269e-01 -5.07078826e-01 3.76256764e-01 -9.72467005e-01
-1.23546219e+00 5.38830340e-01 -1.73614733e-02 -2.06779504e+00
2.22816125e-01 -4.91047591e-01 -4.28249091e-01 1.13876081e+00
-1.54589641e+00 -1.84193456e+00 -7.05561399e-01 1.05216074e+00
3.40262979e-01 -6.49416149e-02 3.34100574e-01 3.77842247e-01
-2.00290173e-01 3.64115268e-01 -2.94441760e-01 -7.71120191e-02
6.40451312e-01 -8.76753747e-01 3.36143553e-01 9.34988022e-01
2.78800040e-01 3.40273678e-01 5.00779808e-01 -8.69209886e-01
-2.04145217e+00 -1.23874521e+00 5.01723468e-01 -8.15872192e-01
6.56357229e-01 -2.94442087e-01 -7.06780195e-01 4.77248639e-01
-3.34404051e-01 2.35146999e-01 1.74834073e-01 -3.71999294e-02
-3.90095502e-01 -2.60738283e-01 -9.49661374e-01 3.17150414e-01
1.14138699e+00 -4.52615470e-01 -7.97662854e-01 3.98431361e-01
8.77684891e-01 -9.09012496e-01 -9.19333994e-01 8.27568889e-01
4.46633697e-01 -7.24648476e-01 1.39990401e+00 1.64169773e-01
-5.91450453e-01 -1.39759982e+00 -3.43717009e-01 -9.88179147e-01
-7.10881054e-01 -2.34865263e-01 -1.79097623e-01 1.15410137e+00
-1.75244614e-01 -4.80690867e-01 9.70177948e-01 9.28953588e-02
-5.21860600e-01 -2.46906221e-01 -1.10266614e+00 -1.16583645e+00
-4.87572640e-01 -5.04021406e-01 8.04864585e-01 7.51966834e-01
-5.51910043e-01 1.60495564e-01 -4.11302857e-02 8.53436172e-01
1.14093268e+00 7.30275869e-01 1.05736125e+00 -1.57727253e+00
-9.29436013e-02 -4.86368239e-01 -6.27593935e-01 -1.14188695e+00
-1.94354221e-01 -8.13740194e-01 1.18256822e-01 -1.13886333e+00
1.13475330e-01 -8.45045686e-01 -2.93639690e-01 2.12599397e-01
-8.05834085e-02 6.52314305e-01 5.24411082e-01 6.01287723e-01
-9.00887132e-01 6.51655674e-01 1.21097040e+00 -8.30964074e-02
-1.27792284e-01 1.09317407e-01 -4.87285912e-01 5.23738027e-01
2.87691414e-01 -3.47323895e-01 7.45925754e-02 -4.79092062e-01
-1.68081567e-01 -6.75527304e-02 8.06443334e-01 -1.55732441e+00
6.50032103e-01 -1.11352332e-01 6.67111099e-01 -1.49229300e+00
5.87849379e-01 -1.16198123e+00 5.59664786e-01 5.75635612e-01
5.37452400e-01 4.36724462e-02 2.70692408e-01 5.61102331e-01
-2.30624303e-01 2.46973813e-01 6.65310085e-01 9.30276662e-02
-1.05909789e+00 7.00348794e-01 3.22068483e-01 -5.08740127e-01
1.22219193e+00 -4.50550646e-01 -3.29595745e-01 4.04648215e-01
-4.98018831e-01 6.19166434e-01 7.29331613e-01 7.67478108e-01
6.80871904e-01 -1.89418960e+00 -5.17367005e-01 3.35767299e-01
5.59906840e-01 3.13384086e-01 4.11311239e-01 1.23637486e+00
-2.34738871e-01 6.25579476e-01 -1.59516245e-01 -1.37740397e+00
-1.45881784e+00 6.05675697e-01 2.93417484e-01 9.10163447e-02
-6.65122390e-01 8.84369314e-01 4.54619855e-01 -4.24613595e-01
8.60099345e-02 -4.16314155e-01 -4.04324830e-02 -9.18344855e-02
5.61548769e-01 -2.25446895e-02 2.00515807e-01 -1.03105092e+00
-6.98447824e-01 1.45543885e+00 2.24745527e-01 1.00395247e-01
9.09872174e-01 -3.10324103e-01 8.41365904e-02 1.64560661e-01
6.84881032e-01 -5.10042459e-02 -1.41467154e+00 -5.59846222e-01
-9.16660950e-02 -7.34416902e-01 2.07350329e-01 -4.04516697e-01
-1.02128088e+00 6.88720763e-01 1.03279185e+00 1.04484577e-02
1.10353613e+00 4.38921005e-01 7.08536267e-01 1.62766241e-02
6.84293687e-01 -4.65793431e-01 3.08535658e-02 5.37455678e-01
5.21021247e-01 -1.27636433e+00 -3.41591388e-02 -5.72760403e-01
-3.51693839e-01 8.91976714e-01 8.55843663e-01 -1.20213129e-01
5.79208612e-01 1.35231614e-01 1.04743680e-02 -6.42765880e-01
-3.45017433e-01 -6.84863985e-01 7.60153890e-01 7.26041973e-01
-1.52929470e-01 -1.71185490e-02 5.89597970e-02 4.00810599e-01
-6.84978664e-02 -2.36278832e-01 -3.18031698e-01 9.24460232e-01
-8.10499966e-01 -1.08771098e+00 -1.10788631e+00 8.95799249e-02
1.97446123e-01 1.50193170e-01 -1.84621260e-01 8.23410034e-01
4.37376231e-01 7.32746363e-01 2.13949159e-01 -8.75908673e-01
5.59756517e-01 -3.35239261e-01 8.34855199e-01 -5.06801605e-01
-7.57658660e-01 5.18784165e-01 -3.12052876e-01 -9.35853779e-01
-6.70141339e-01 -8.60162556e-01 -1.15439785e+00 -1.89560577e-01
-6.62489593e-01 3.29115838e-02 8.82317364e-01 9.44461644e-01
3.27896565e-01 4.01640534e-01 6.06667936e-01 -1.25299633e+00
-2.54097104e-01 -7.76212811e-01 -4.34805870e-01 2.38909721e-01
1.58978790e-01 -1.29484963e+00 -1.12817995e-01 -2.29375124e-01] | [6.553342819213867, -2.32308030128479] |
0bb0c09c-660d-4b9e-8dd8-3183b6b499a5 | mocapact-a-multi-task-dataset-for-simulated | 2208.07363 | null | https://arxiv.org/abs/2208.07363v3 | https://arxiv.org/pdf/2208.07363v3.pdf | MoCapAct: A Multi-Task Dataset for Simulated Humanoid Control | Simulated humanoids are an appealing research domain due to their physical capabilities. Nonetheless, they are also challenging to control, as a policy must drive an unstable, discontinuous, and high-dimensional physical system. One widely studied approach is to utilize motion capture (MoCap) data to teach the humanoid agent low-level skills (e.g., standing, walking, and running) that can then be re-used to synthesize high-level behaviors. However, even with MoCap data, controlling simulated humanoids remains very hard, as MoCap data offers only kinematic information. Finding physical control inputs to realize the demonstrated motions requires computationally intensive methods like reinforcement learning. Thus, despite the publicly available MoCap data, its utility has been limited to institutions with large-scale compute. In this work, we dramatically lower the barrier for productive research on this topic by training and releasing high-quality agents that can track over three hours of MoCap data for a simulated humanoid in the dm_control physics-based environment. We release MoCapAct (Motion Capture with Actions), a dataset of these expert agents and their rollouts, which contain proprioceptive observations and actions. We demonstrate the utility of MoCapAct by using it to train a single hierarchical policy capable of tracking the entire MoCap dataset within dm_control and show the learned low-level component can be re-used to efficiently learn downstream high-level tasks. Finally, we use MoCapAct to train an autoregressive GPT model and show that it can control a simulated humanoid to perform natural motion completion given a motion prompt. Videos of the results and links to the code and dataset are available at https://microsoft.github.io/MoCapAct. | ['Matthew Hausknecht', 'Ching-An Cheng', 'Ricky Loynd', 'Felipe Vieira Frujeri', 'Andrey Kolobov', 'Nolan Wagener'] | 2022-08-15 | null | null | null | null | ['humanoid-control'] | ['robots'] | [-4.53201294e-01 4.50330041e-03 -1.38418943e-01 3.17965150e-01
-5.61779320e-01 -5.80733478e-01 6.45047724e-01 -3.92809391e-01
-6.36017084e-01 9.65701699e-01 1.95568234e-01 -2.51842439e-01
-9.96530429e-02 -6.26352072e-01 -1.08838308e+00 -8.05924773e-01
-3.40432346e-01 5.89474440e-01 4.46455657e-01 -5.95537066e-01
9.04385448e-02 6.11088991e-01 -1.64346707e+00 -1.15053728e-01
6.41082704e-01 3.43671441e-01 5.68039000e-01 1.19240427e+00
6.81885958e-01 1.11085832e+00 -3.10886979e-01 4.85256463e-01
3.43924195e-01 -4.55844611e-01 -6.35619223e-01 -8.26148912e-02
1.30475670e-01 -4.33324575e-01 -8.10349345e-01 3.83835256e-01
7.27626324e-01 8.84081662e-01 5.82035840e-01 -1.59468830e+00
-2.49464914e-01 1.40230104e-01 3.95237282e-03 3.78921703e-02
6.32914543e-01 1.18994498e+00 4.80174869e-01 -4.15900528e-01
8.52726460e-01 1.24153554e+00 5.19087374e-01 9.99919474e-01
-1.10454953e+00 -4.11211997e-01 -1.05053641e-01 1.24741606e-01
-1.05671859e+00 -3.43666047e-01 1.75572112e-01 -5.67407787e-01
1.38494444e+00 -6.33903891e-02 1.27496982e+00 1.67166328e+00
5.36325753e-01 8.97761226e-01 9.05650795e-01 1.98680852e-02
6.11619592e-01 -6.53522730e-01 -3.45344841e-01 1.04662693e+00
-3.51224951e-02 7.33120859e-01 -5.65967202e-01 -5.50712645e-02
1.06704009e+00 -1.10008024e-01 -3.60385060e-01 -6.84701145e-01
-1.35926867e+00 4.93395269e-01 5.50682724e-01 -1.03554524e-01
-6.10944390e-01 8.80422473e-01 4.01274376e-02 1.52269676e-01
-5.41076243e-01 8.13456059e-01 -5.81977129e-01 -7.85594165e-01
-2.75559366e-01 1.10302186e+00 1.03595555e+00 1.04240990e+00
3.80689800e-01 3.58915508e-01 2.38351226e-02 4.25904840e-01
4.32846099e-02 7.82443106e-01 4.13500696e-01 -1.84781837e+00
2.42556557e-01 2.63117522e-01 5.42694926e-01 -7.88151383e-01
-8.35958302e-01 1.85528323e-01 -3.91666293e-01 1.03484976e+00
5.01784027e-01 -6.45046234e-01 -9.70927775e-01 1.78638697e+00
5.63262820e-01 2.01308563e-01 1.66929573e-01 1.34382069e+00
3.99810791e-01 9.68013704e-01 1.94565967e-01 3.20658028e-01
1.17996871e+00 -1.18162453e+00 -2.91304022e-01 -3.93454432e-01
6.30432427e-01 -1.67169392e-01 1.28174531e+00 4.95857209e-01
-1.32236791e+00 -5.05144536e-01 -9.68249977e-01 1.76123139e-02
-1.86737791e-01 -4.19054180e-01 6.49610460e-01 -1.02529839e-01
-1.03587174e+00 9.53171968e-01 -1.74256063e+00 -5.40255010e-01
4.76350449e-02 4.77571487e-01 -4.42151874e-01 1.10345818e-01
-1.01233363e+00 1.29722881e+00 3.36626202e-01 -3.86938483e-01
-1.70380759e+00 -8.18864405e-01 -9.60976660e-01 -9.80098769e-02
6.32809460e-01 -1.07750237e+00 1.72285140e+00 -3.49023581e-01
-2.03561878e+00 1.25972509e-01 4.80771512e-01 -3.45912039e-01
5.87606728e-01 -3.87914568e-01 -4.26834170e-03 3.61404210e-01
3.36349122e-02 1.15965486e+00 5.50720751e-01 -1.04964745e+00
-7.36127436e-01 1.44643486e-01 1.59646317e-01 5.35056710e-01
1.06158294e-01 -3.65470767e-01 -4.60243016e-01 -4.59037304e-01
-3.77504438e-01 -1.43489957e+00 -5.34715474e-01 2.17875183e-01
-4.91350070e-02 -2.62938980e-02 5.64746380e-01 -4.13749903e-01
6.78836823e-01 -1.74717915e+00 8.31599534e-01 -2.73220718e-01
3.56113091e-02 1.74263835e-01 -1.27763331e-01 7.23559558e-01
4.68979150e-01 -1.17749304e-01 -3.69056404e-01 6.71569898e-04
2.60005414e-01 4.86528903e-01 1.02631599e-02 3.29801887e-01
1.65916849e-02 1.08492839e+00 -1.19236147e+00 -3.76072347e-01
3.37120265e-01 5.85183918e-01 -8.77127111e-01 3.30188692e-01
-6.13991439e-01 9.25921917e-01 -7.52712965e-01 3.39515269e-01
1.77016184e-02 -2.88160890e-02 9.33786184e-02 3.97035420e-01
-3.34013730e-01 8.33115876e-02 -1.01522446e+00 1.84786665e+00
-5.17958820e-01 4.04849797e-01 3.31683725e-01 -6.32413745e-01
4.53789651e-01 2.72368431e-01 7.18308330e-01 -5.97533584e-01
7.35689923e-02 1.14426434e-01 8.82615745e-02 -8.23159754e-01
4.90495116e-01 -4.12074700e-02 -3.79860818e-01 2.75856763e-01
1.56116067e-02 -8.20820808e-01 2.67243475e-01 3.28493714e-02
1.69979501e+00 7.79927671e-01 2.42837109e-02 -1.67557135e-01
1.24269262e-01 8.10876369e-01 4.39520180e-01 7.95411468e-01
-3.50954771e-01 3.76133889e-01 2.13236317e-01 -2.84766734e-01
-1.42167497e+00 -1.16588604e+00 3.78489077e-01 1.17179441e+00
6.64509907e-02 -2.56314039e-01 -9.67446804e-01 -7.05591440e-02
2.11145520e-01 6.16808593e-01 -3.93722832e-01 -2.99290329e-01
-1.15118802e+00 -2.67832577e-01 5.61263740e-01 7.30711281e-01
4.72758591e-01 -1.69792068e+00 -1.22626066e+00 3.95923972e-01
-3.81825939e-02 -9.85071898e-01 -5.00926793e-01 2.28932649e-02
-6.54829383e-01 -1.14322937e+00 -7.03906059e-01 -8.52637768e-01
3.53443623e-01 7.90855363e-02 9.16626155e-01 2.69192457e-01
-2.44462162e-01 7.66261399e-01 -2.08190188e-01 3.09105236e-02
-4.87425864e-01 1.15559198e-01 4.87391055e-01 -8.49794745e-01
-1.99569911e-01 -6.58278644e-01 -7.52556264e-01 4.26053494e-01
-6.64989471e-01 3.19820642e-01 3.96231860e-01 7.95690179e-01
4.23726320e-01 -3.33509445e-01 2.54140854e-01 -3.21685337e-02
5.59309781e-01 -5.34635842e-01 -7.52592266e-01 -2.69908875e-01
-4.99615036e-02 8.94709602e-02 8.15525830e-01 -7.13958204e-01
-8.47216010e-01 3.49804342e-01 -7.63263404e-02 -3.81926745e-01
-3.19748670e-01 3.32161993e-01 1.25151068e-01 -1.18979923e-02
7.79261887e-01 2.09790289e-01 2.04658061e-01 -1.42517552e-01
4.52088475e-01 1.53850570e-01 8.35299194e-01 -9.82580483e-01
6.92341328e-01 3.03230941e-01 1.50574788e-01 -7.56295741e-01
-2.91665494e-01 -3.49399745e-02 -4.23856914e-01 -3.67147624e-01
1.23242581e+00 -8.12301695e-01 -1.58803248e+00 5.62472403e-01
-7.23928511e-01 -1.46122956e+00 -1.22906432e-01 7.71647751e-01
-1.49476182e+00 -1.07611164e-01 -8.55260670e-01 -5.82166016e-01
1.50713995e-01 -1.27037954e+00 8.03747237e-01 3.66133511e-01
-3.54119211e-01 -7.70531237e-01 5.42645693e-01 1.98116973e-01
3.07728410e-01 5.45648754e-01 5.41517735e-01 1.86318979e-02
-6.94792390e-01 6.31478056e-02 5.61240435e-01 -3.00086945e-01
-7.31860548e-02 1.82361528e-02 -3.86339784e-01 -6.18062258e-01
-3.10298175e-01 -7.65293300e-01 2.96007872e-01 4.92828876e-01
8.97904038e-01 -3.32569808e-01 -4.72107977e-01 5.96504748e-01
1.07811654e+00 1.70208454e-01 4.24959660e-01 5.93342781e-01
6.28469229e-01 4.65732723e-01 7.28079677e-01 4.17925477e-01
6.93594396e-01 8.51466835e-01 4.18088585e-01 2.90302634e-01
-7.51217231e-02 -5.07193685e-01 7.87658453e-01 7.96264887e-01
-4.10339177e-01 -1.67350963e-01 -1.20495617e+00 5.31659603e-01
-2.18714404e+00 -1.23535609e+00 9.22416002e-02 1.82626963e+00
8.24931800e-01 5.56323826e-02 4.54370022e-01 -2.14966297e-01
3.36696029e-01 -2.03583539e-01 -8.34539354e-01 -2.16512948e-01
3.05931747e-01 1.69266120e-01 4.87168133e-01 7.28978455e-01
-9.63016689e-01 1.25386345e+00 6.04139900e+00 5.07313609e-01
-9.59818184e-01 -3.10678095e-01 -1.12821959e-01 -4.12492305e-01
3.34684819e-01 -1.43498883e-01 -5.47913611e-01 4.91899997e-01
1.16602600e+00 -2.78245479e-01 9.06769037e-01 1.05467629e+00
7.47423410e-01 -2.71289021e-01 -1.14618027e+00 5.33431113e-01
-4.24981356e-01 -1.22067726e+00 -4.56716239e-01 8.96672830e-02
8.57013226e-01 2.15915009e-01 1.37389163e-02 8.47804546e-01
1.25772142e+00 -1.22088552e+00 7.90318191e-01 6.59259975e-01
3.29457760e-01 -7.41955280e-01 2.39756610e-02 8.05810452e-01
-1.08826172e+00 -2.74488628e-01 -3.33365232e-01 -4.71413076e-01
4.57273871e-01 -4.02981639e-01 -6.37830436e-01 1.41136035e-01
6.76958859e-01 5.91241896e-01 -8.16798881e-02 1.02309752e+00
-4.31887358e-01 3.92957300e-01 -4.75588441e-01 -4.62177515e-01
4.76026684e-01 -1.12241887e-01 6.53806210e-01 7.45200574e-01
3.31747979e-01 6.06548250e-01 6.58605278e-01 8.05701792e-01
3.40968311e-01 -4.87596124e-01 -7.82580197e-01 4.07784842e-02
3.04165304e-01 1.15093613e+00 -4.43848461e-01 -3.39216560e-01
1.38214082e-02 8.62950444e-01 4.61414576e-01 4.60352927e-01
-1.05775273e+00 -1.35595024e-01 1.18349111e+00 1.76854338e-02
3.23845565e-01 -1.09821820e+00 4.38338965e-02 -1.12450159e+00
-2.52249151e-01 -1.08670425e+00 9.34285969e-02 -1.10758400e+00
-8.98388982e-01 1.65096909e-01 1.85889229e-01 -1.25797737e+00
-8.55545878e-01 -6.61499143e-01 -3.87947053e-01 5.20501852e-01
-9.47259247e-01 -8.25072408e-01 -6.66060984e-01 7.09523976e-01
5.89025676e-01 2.69391164e-02 7.31350303e-01 -2.06103828e-02
-4.64092463e-01 -6.13417178e-02 -1.66669130e-01 3.58508229e-02
4.52511668e-01 -1.23807228e+00 6.99284554e-01 5.08727193e-01
-3.17300171e-01 3.94643486e-01 1.00425243e+00 -8.40998948e-01
-2.07547593e+00 -8.20946753e-01 -1.51004121e-02 -8.82718027e-01
8.10764730e-01 -2.67414570e-01 -8.66019547e-01 7.35924006e-01
5.83296195e-02 1.59047857e-01 -6.45587370e-02 -6.04499578e-01
3.46605986e-01 3.92912865e-01 -8.72888625e-01 1.27338266e+00
1.43363893e+00 -1.10030904e-01 -6.79492772e-01 1.68533906e-01
7.56855249e-01 -8.63682091e-01 -9.07498300e-01 1.76380396e-01
6.40137851e-01 -6.47754848e-01 1.03127348e+00 -1.03735900e+00
7.67660260e-01 -5.13877332e-01 -1.23902988e-02 -1.77868927e+00
-2.83613235e-01 -7.32924461e-01 -5.02840877e-01 4.71759349e-01
1.96573764e-01 -3.11834604e-01 8.74066710e-01 5.38916230e-01
-2.90805936e-01 -6.13619208e-01 -6.37375772e-01 -9.38973308e-01
5.48511207e-01 -3.70792121e-01 3.37153167e-01 6.51430190e-01
2.97330618e-01 1.31282769e-02 -4.81185913e-01 2.91244954e-01
4.70476359e-01 -2.92336553e-01 1.28902364e+00 -6.01431608e-01
-6.76661670e-01 -4.06322151e-01 -1.77705884e-01 -1.27758038e+00
3.33664447e-01 -6.86376631e-01 6.36640191e-01 -1.64895058e+00
-2.61165977e-01 -4.46310490e-01 2.08811417e-01 6.29698336e-01
-8.96872394e-03 -2.13412613e-01 5.25731981e-01 2.51185060e-01
-3.80472392e-01 8.30696940e-01 1.69873464e+00 1.16811708e-01
-4.30820405e-01 -1.32271454e-01 -5.34736402e-02 6.66972518e-01
9.87811267e-01 -3.44223440e-01 -4.61120337e-01 -4.96331453e-01
-1.11227371e-01 6.38329506e-01 6.81889474e-01 -1.61128485e+00
4.92338806e-01 -7.26343215e-01 3.14377427e-01 -1.47628188e-01
6.79326653e-01 -5.47779739e-01 2.23252252e-01 9.69522417e-01
-2.95930088e-01 3.80546540e-01 3.53349924e-01 4.82868701e-01
2.07390517e-01 2.28329375e-01 6.16207778e-01 -3.71926248e-01
-9.11877155e-01 3.46034229e-01 -1.00037932e+00 2.70404458e-01
1.23369634e+00 4.71833870e-02 -3.79088104e-01 -4.60567534e-01
-9.16199744e-01 6.79184318e-01 8.01667869e-01 4.21217680e-01
5.03720641e-01 -1.25102055e+00 -5.76588690e-01 -2.06415787e-01
-2.03931063e-01 9.94870439e-02 1.38018146e-01 6.77952766e-01
-9.61388052e-01 2.05095321e-01 -8.48725557e-01 -5.71067512e-01
-6.39735341e-01 6.83273137e-01 4.94615495e-01 -8.52168910e-03
-1.10171807e+00 3.12607318e-01 6.30782619e-02 -7.74959505e-01
-1.54654169e-02 -4.46628898e-01 7.36423135e-02 -7.09343791e-01
3.50492507e-01 5.65896988e-01 -3.57108891e-01 -2.27485090e-01
-2.95886874e-01 5.28034687e-01 6.13194346e-01 -7.01474965e-01
1.45679188e+00 2.18958497e-01 4.11426276e-01 1.85947716e-01
5.91545939e-01 -3.76410693e-01 -2.26515508e+00 5.28174341e-01
-9.33353603e-02 -1.59300655e-01 -2.33521804e-01 -5.86542428e-01
-7.64060676e-01 6.81766391e-01 1.72425821e-01 -1.01277731e-01
5.78638136e-01 -1.92880839e-01 7.84858227e-01 8.56042027e-01
7.81742871e-01 -1.32381403e+00 5.04031360e-01 8.46935987e-01
1.11541557e+00 -1.01722753e+00 -2.10242450e-01 6.94232360e-02
-8.53259444e-01 8.95615339e-01 1.15441084e+00 -5.44314742e-01
3.37952405e-01 6.15328193e-01 1.64973214e-02 -1.48528710e-01
-1.22389948e+00 -2.10240722e-01 -1.37058198e-01 9.98400688e-01
4.80003376e-03 5.99386245e-02 5.84428273e-02 2.55832374e-01
-5.70559859e-01 1.40925825e-01 9.41051185e-01 1.45912087e+00
-5.93849421e-01 -7.24439204e-01 -3.54380935e-01 -1.38612762e-01
7.84958974e-02 4.01254267e-01 -5.15178218e-02 1.11632133e+00
-3.14022779e-01 7.83109069e-01 7.93803670e-03 -3.32227081e-01
5.70242524e-01 -2.15337649e-01 6.61800861e-01 -5.10171056e-01
-5.73152244e-01 -4.45845872e-01 1.07071914e-01 -1.01699066e+00
-2.49886531e-02 -7.51975954e-01 -1.81670666e+00 -8.37970972e-01
4.94252443e-01 -1.14265077e-01 5.68232715e-01 5.77106893e-01
4.10616934e-01 6.63610101e-01 1.36859372e-01 -1.64332402e+00
-6.17786944e-01 -7.35388219e-01 1.82005912e-02 4.02569711e-01
5.16880751e-01 -1.07636523e+00 -1.25624612e-01 1.00324817e-01] | [4.741871356964111, 0.9454472661018372] |
fe3f49cc-040a-495b-8938-355fcfa59ed8 | the-tetrazole-analogue-of-the-auxin-indole-3 | 1808.08842 | null | http://arxiv.org/abs/1808.08842v1 | http://arxiv.org/pdf/1808.08842v1.pdf | The tetrazole analogue of the auxin indole-3-acetic acid binds preferentially to TIR1 and not AFB5 | Auxin is considered one of the cardinal hormones in plant growth and
development. It regulates a wide range of processes throughout the plant.
Synthetic auxins exploit the auxin-signalling pathway and are valuable as
herbicidal agrochemicals. Currently, despite a diversity of chemical scaffolds
all synthetic auxins have a carboxylic acid as the active core group. By
applying bio-isosteric replacement we discovered that indole-3-tetrazole was
active by surface plasmon resonance (SPR) spectrometry, showing that the
tetrazole could initiate assembly of the TIR1 auxin co-receptor complex. We
then tested the tetrazole's efficacy in a range of whole plant physiological
assays and in protoplast reporter assays which all confirmed auxin activity,
albeit rather weak. We then tested indole-3-tetrazole against the AFB5
homologue of TIR1, finding that binding was selective against TIR1, absent with
AFB5. The kinetics of binding to TIR1 are contrasted to those for the herbicide
picloram, which shows the opposite receptor preference as it binds to AFB5 with
far greater affinity than to TIR1. The basis of the preference of
indole-3-tetrazole for TIR1 was revealed to be a single residue substitution
using molecular docking, and assays using tir1 and afb5 mutant lines confirmed
selectivity in vivo. Given the potential that a TIR1-selective auxin might have
for unmasking receptor-specific actions, we followed a rational design, lead
optimisation campaign and a set of chlorinated indole-3-tetrazoles was
synthesised. Improved affinity for TIR1 and the preference for binding to TIR1
was maintained for 4- and 6-chloroindole-3-tetrazoles, coupled with improved
efficacy in vivo. This work expands the range of auxin chemistry for the design
of receptor-selective synthetic auxins. | [] | 2018-08-27 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 9.72914696e-01 1.56734928e-01 -5.18453360e-01 6.93614632e-02
-2.95914680e-01 -1.13460159e+00 6.24158919e-01 5.97707868e-01
-3.93296421e-01 1.15813982e+00 -5.12609445e-02 -9.47760224e-01
-2.06120938e-01 -5.78692555e-01 -5.92943847e-01 -1.12139547e+00
7.10431859e-02 1.65628359e-01 4.24090892e-01 -5.31853259e-01
5.38915157e-01 9.87371266e-01 -1.43255436e+00 -1.08876703e-02
8.26980412e-01 2.72119790e-01 7.25394428e-01 7.47907579e-01
-1.27546728e-01 1.45073622e-01 -4.43013668e-01 2.42295533e-01
-6.14860393e-02 -4.84483153e-01 -9.77949202e-01 -3.40551853e-01
-5.11050299e-02 -1.07234128e-01 7.29461432e-01 3.07563722e-01
3.55994701e-01 -9.69624668e-02 8.48618329e-01 -9.04759049e-01
-4.65181887e-01 -1.84725635e-02 -8.27274442e-01 -4.90057170e-01
3.63503814e-01 -1.53321698e-01 8.21396470e-01 -1.09587979e+00
6.73362970e-01 1.13287723e+00 2.76592046e-01 5.18588185e-01
-1.62136745e+00 -2.32794777e-01 -1.17619082e-01 -4.06216145e-01
-9.03949559e-01 -5.17176509e-01 -1.60306007e-01 -1.50920182e-01
7.35564530e-01 5.69573104e-01 3.54343474e-01 7.29635417e-01
5.74678600e-01 4.18784678e-01 9.47608173e-01 -4.77689743e-01
3.67413849e-01 -4.78180677e-01 -6.63454473e-01 1.39116570e-01
3.84392470e-01 -1.86320394e-01 -1.42031401e-01 -5.55233121e-01
6.05665743e-01 -1.12441406e-01 -2.31093094e-01 -6.88231230e-01
-7.80082822e-01 6.42067850e-01 5.74936979e-02 4.25512791e-01
-6.18690491e-01 -2.10994765e-01 5.37680268e-01 -4.22750026e-01
-2.49329239e-01 7.21538186e-01 -9.88390803e-01 1.09228864e-01
3.28797917e-03 5.93132898e-02 7.50500023e-01 3.34262222e-01
4.17110562e-01 -3.37904394e-02 1.90048993e-01 1.02144921e+00
2.82651603e-01 4.60447669e-01 1.20862193e-01 -8.63014102e-01
-5.03139019e-01 4.62044835e-01 6.63569510e-01 -3.63989919e-01
-5.56159854e-01 2.78031081e-01 -4.91461575e-01 4.58300769e-01
1.16868234e+00 -2.70926952e-02 -7.95506656e-01 1.95145214e+00
3.74862373e-01 -2.41030440e-01 2.40126997e-01 7.22084045e-01
2.03099325e-01 6.32811844e-01 8.09024155e-01 -4.52258378e-01
1.64136016e+00 5.15631922e-02 -2.51156777e-01 1.79526687e-01
8.90571237e-01 -1.15962005e+00 9.03672218e-01 3.49266201e-01
-7.77774811e-01 3.76381688e-02 -8.44791353e-01 3.13601077e-01
-5.93955219e-01 2.02544272e-01 8.50826681e-01 9.03762698e-01
-5.84902585e-01 6.47170782e-01 -6.75362229e-01 -7.70676255e-01
2.13284381e-02 5.63878059e-01 -7.66204417e-01 -4.57972512e-02
-8.59540462e-01 1.02365446e+00 4.21319008e-01 8.79554749e-02
-7.37851083e-01 -4.27613109e-01 -4.02098477e-01 6.28164560e-02
1.15958363e-01 -3.97080809e-01 1.20432186e+00 -2.81947732e-01
-1.98907292e+00 6.58726931e-01 -4.00727123e-01 -2.08361447e-01
-3.80404741e-01 2.08409488e-01 -2.34277204e-01 -4.72422838e-02
4.87473518e-01 6.76130772e-01 8.31780806e-02 -1.35951078e+00
-7.11437404e-01 -5.00605226e-01 -8.33005458e-02 -3.81541229e-03
4.75207776e-01 1.23562925e-01 5.42806864e-01 -5.13537228e-02
4.83184963e-01 -1.14880097e+00 -4.91850287e-01 -4.35095251e-01
-3.30117255e-01 1.50280267e-01 1.03439224e+00 2.35737085e-01
4.10302043e-01 -1.63863504e+00 -3.78397703e-01 3.16455394e-01
-1.43904209e-01 7.72803068e-01 -3.37861538e-01 1.11228728e+00
-5.96443772e-01 5.47661364e-01 -7.92419910e-02 9.06321585e-01
-5.44007301e-01 1.06189087e-01 -5.74017107e-01 6.67522788e-01
4.09100592e-01 5.95445991e-01 -1.25427520e+00 2.11638615e-01
2.34301195e-01 4.67812836e-01 5.39933331e-02 -5.89262605e-01
-4.02845144e-01 6.93805039e-01 -8.32663953e-01 1.09830594e+00
1.20727420e+00 7.13766292e-02 8.28610659e-01 -4.20887955e-02
-8.25940669e-01 3.20482701e-01 -5.36176920e-01 1.04156017e+00
-6.65621236e-02 5.69927432e-02 3.68839130e-02 -5.98729849e-01
1.22013509e+00 5.18586516e-01 6.13662839e-01 -6.15126729e-01
-1.83220372e-01 7.04265296e-01 1.94654211e-01 4.32869554e-01
2.44467288e-01 -4.14332040e-02 2.18529254e-01 3.61371152e-02
-3.36523086e-01 -3.59932296e-02 5.28352201e-01 1.05721720e-01
1.05691111e+00 4.87220854e-01 6.00483835e-01 -6.41577065e-01
7.55585432e-01 -4.34921123e-03 3.51556361e-01 5.01093507e-01
-6.28222770e-04 1.24956422e-01 8.09413731e-01 -2.79249489e-01
-9.40916836e-01 -7.87589371e-01 -5.37235260e-01 1.50257015e+00
1.93589941e-01 -4.01474446e-01 -4.31133777e-01 -1.55024692e-01
-8.93904194e-02 8.23229015e-01 -5.27082920e-01 4.14468944e-02
-1.27413288e-01 -8.50354314e-01 8.66929054e-01 -2.07200602e-01
1.08691439e-01 -7.43971109e-01 -8.59099746e-01 5.98167539e-01
4.04718459e-01 -5.91380239e-01 2.83692896e-01 7.87414730e-01
-6.04953229e-01 -1.11790836e+00 -1.00425315e+00 -3.20538402e-01
2.48723775e-01 6.19764149e-01 4.21207279e-01 -6.10674694e-02
-1.02011085e-01 -1.12914480e-01 -3.60736042e-01 -1.23058546e+00
-3.94579291e-01 3.97662930e-02 -4.02227640e-02 -5.68236589e-01
1.79972902e-01 -9.09984112e-01 -5.92805862e-01 5.55978656e-01
-7.63204277e-01 -2.52743125e-01 7.16518819e-01 9.62203622e-01
6.06170833e-01 -6.10702097e-01 6.71614408e-01 -1.00582421e+00
4.79400039e-01 -4.40692425e-01 -7.43288755e-01 2.28594109e-01
-3.09486449e-01 -2.39640832e-01 7.14960039e-01 -2.58736461e-01
-1.02304995e+00 2.67927319e-01 1.80194050e-01 8.29615533e-01
-5.47731340e-01 7.21942306e-01 -5.57577968e-01 -1.40016824e-01
7.51483619e-01 -3.23426761e-02 2.54800841e-02 -1.61128283e-01
1.96686313e-01 9.84853506e-02 5.95435858e-01 -9.66410339e-01
4.37337369e-01 6.14857852e-01 8.14928234e-01 -1.51418924e+00
-4.24811006e-01 -5.02534568e-01 -2.81690031e-01 1.69531554e-01
7.66524315e-01 -5.36470771e-01 -1.57241797e+00 3.09165239e-01
-1.28717458e+00 -8.93644020e-02 3.82178351e-02 5.64122736e-01
-5.60221195e-01 1.37012631e-01 -1.61431432e-01 -7.04974830e-01
9.97298583e-02 -1.14905584e+00 9.60272789e-01 3.82778883e-01
-3.45766932e-01 -5.73879540e-01 3.48352373e-01 -2.97257006e-01
1.70312524e-01 7.29233861e-01 1.39308703e+00 -5.78980446e-01
-3.82581949e-01 -1.85163006e-01 -5.55302680e-01 -4.80026543e-01
4.51137543e-01 7.07443357e-01 -9.06968117e-01 1.84696943e-01
-8.33329260e-01 -4.38709438e-01 6.03439510e-01 6.28528833e-01
5.49906015e-01 2.35549018e-01 -6.53420269e-01 4.09443974e-01
1.23351264e+00 9.21407223e-01 1.03101921e+00 8.24485838e-01
4.83511001e-01 3.19130808e-01 1.11018717e+00 5.20006835e-01
-4.66109693e-01 4.67896551e-01 8.34909856e-01 -4.05288815e-01
7.40342140e-01 -1.05009235e-01 -6.69204444e-02 -6.11917317e-01
-6.69458807e-01 -5.07903874e-01 -7.74973929e-01 3.14472705e-01
-1.57598996e+00 -9.38808322e-01 -7.09853053e-01 2.46261477e+00
9.22039092e-01 8.62316415e-03 2.46391073e-01 -5.80964759e-02
4.12231475e-01 -2.18699768e-01 -5.61293781e-01 -1.10810328e+00
-5.64400375e-01 5.76383412e-01 1.32698059e+00 3.83305460e-01
-8.33214998e-01 1.24796832e+00 6.61657667e+00 6.14432156e-01
-1.21342194e+00 -8.41288328e-01 1.96572870e-01 9.02230978e-01
-2.71668464e-01 1.07046962e+00 -5.29945254e-01 -7.54058957e-02
9.12497163e-01 -3.49703938e-01 -9.14588645e-02 5.92728376e-01
6.59294844e-01 -8.93133283e-01 -7.45139897e-01 -1.02793902e-01
-1.17798936e+00 -1.56868494e+00 8.29389989e-02 5.65567076e-01
6.75504267e-01 -2.44670689e-01 -1.33710921e-01 -4.03716892e-01
3.88357252e-01 -1.00754869e+00 5.88587597e-02 2.60332197e-01
8.42517495e-01 -9.57920909e-01 3.27638566e-01 1.37865663e-01
-9.80922222e-01 1.06685728e-01 -4.83133286e-01 -5.67729101e-02
2.23013889e-02 3.58763993e-01 -1.38123500e+00 3.68089080e-01
2.82584459e-01 -9.45139024e-03 -2.52142996e-01 7.75282204e-01
-1.44817576e-01 3.82147133e-01 -1.82199135e-01 1.44907802e-01
4.49970692e-01 -5.93656600e-01 3.76979262e-01 8.83049488e-01
8.74274671e-02 3.13277245e-01 1.37312025e-01 3.56279492e-01
2.19006389e-01 2.87540913e-01 -8.26290429e-01 -4.33852583e-01
5.70707202e-01 1.13749421e+00 -1.17810273e+00 1.71473268e-02
-2.82484084e-01 8.09161961e-01 -4.32876945e-01 5.77938259e-01
-2.13147283e-01 -6.34676814e-01 9.17839348e-01 -1.32924259e-01
3.09860557e-01 -5.90652786e-02 -1.17513373e-01 3.97472233e-02
-8.32514465e-01 -9.15922225e-01 -9.87389237e-02 -7.41345882e-01
-4.31643158e-01 -1.87193587e-01 1.46352034e-02 -5.73842466e-01
-1.02892011e-01 -6.64283395e-01 -5.81993937e-01 1.38925290e+00
-9.15914059e-01 -1.33506620e+00 1.36647269e-01 -1.12202615e-01
7.30157420e-02 1.04203276e-01 1.86750197e+00 -3.32296014e-01
-2.08594650e-01 -7.58277252e-02 4.51167643e-01 -7.55341947e-01
1.15427184e+00 -1.12115073e+00 2.41176113e-01 3.60159099e-01
-6.02470219e-01 1.00982678e+00 9.94807780e-01 -7.70904779e-01
-1.21289980e+00 -6.88761890e-01 8.23194444e-01 -1.86614782e-01
4.64843631e-01 1.19390622e-01 -6.74508035e-01 4.59798843e-01
-2.38870196e-02 -6.00052178e-01 9.82328773e-01 9.09078196e-02
-2.33227864e-01 7.13499606e-01 -1.10465646e+00 1.13410044e+00
5.16277611e-01 -3.73716801e-01 4.29814994e-01 3.12028855e-01
4.22513306e-01 -3.68950754e-01 -8.40005219e-01 5.67251146e-01
1.17298067e+00 -9.68673110e-01 1.05243182e+00 -7.41251886e-01
2.05970719e-01 -8.24243069e-01 -2.88259476e-01 -1.05086160e+00
-5.55261970e-01 -8.77157867e-01 5.30188143e-01 9.19851422e-01
4.16616499e-01 -4.98860747e-01 3.28779697e-01 4.02180731e-01
5.12702093e-02 -5.40548086e-01 -6.80257738e-01 -6.37360990e-01
-6.10579327e-02 4.08486247e-01 5.40900826e-01 6.99566066e-01
8.83935541e-02 7.27223039e-01 -2.29342148e-01 -5.95375672e-02
-5.40108941e-02 -2.47603267e-01 9.02699709e-01 -1.51313198e+00
3.92679930e-01 -1.98926702e-01 -2.77451485e-01 -5.24844825e-01
4.03174050e-02 -4.59528029e-01 -5.37537754e-01 -1.37576497e+00
-3.15068722e-01 -2.52654672e-01 1.18131116e-01 5.18132627e-01
3.75661701e-01 -2.06032202e-01 -1.25392169e-01 -3.51242684e-02
1.87374920e-01 1.98742710e-02 1.13701057e+00 3.11290249e-02
-8.39402199e-01 3.44986111e-01 -1.11242676e+00 7.65178621e-01
1.02361488e+00 -4.47918743e-01 -6.43014669e-01 3.13037217e-01
3.75204176e-01 -3.06002088e-02 -4.18347456e-02 -1.57042369e-01
-4.32515532e-01 -8.02793026e-01 6.10942960e-01 -4.56978709e-01
5.14291644e-01 -5.73723614e-01 2.27841511e-01 4.88718182e-01
2.26255536e-01 -2.30221480e-01 6.76550090e-01 3.08929533e-01
3.86267930e-01 -2.16020167e-01 9.01559711e-01 -1.72277540e-01
-7.07971871e-01 -4.12693545e-02 -1.43278122e+00 -5.01010716e-01
8.22221696e-01 -8.05930853e-01 -4.21151906e-01 -2.18545824e-01
-6.05205417e-01 -3.52310479e-01 7.99633622e-01 5.67625687e-02
2.53835142e-01 -8.00016940e-01 -3.52358192e-01 -1.52732760e-01
5.19624889e-01 -4.43441004e-01 -1.16068095e-01 1.11089110e+00
-8.08043957e-01 1.02153730e+00 -7.38460958e-01 -7.41431892e-01
-1.33476388e+00 4.91085380e-01 2.36052200e-01 8.34334493e-02
2.51047105e-01 5.54956555e-01 7.26096869e-01 -4.07000929e-01
-3.93074870e-01 -5.69085181e-02 -3.20005536e-01 -1.77665487e-01
3.74995202e-01 3.91856879e-01 5.44953682e-02 -7.67483413e-01
-7.51689136e-01 2.64088452e-01 -1.77937821e-01 1.53716981e-01
1.17554462e+00 5.87451980e-02 -3.45145792e-01 1.79271892e-01
6.64972842e-01 2.30872095e-01 -7.28158772e-01 4.89678234e-01
4.27835047e-01 -5.17450392e-01 -3.57883722e-01 -8.43888521e-01
-2.47084182e-02 4.48990464e-02 5.34442246e-01 -1.00447319e-01
8.29945624e-01 -2.58660614e-01 1.26197964e-01 5.12008727e-01
3.96984935e-01 -8.90597522e-01 -4.95676845e-01 6.61800921e-01
7.66243100e-01 -6.57683849e-01 -3.83402593e-02 -5.72350562e-01
-3.62983286e-01 9.02434707e-01 5.25890887e-01 2.43474320e-01
2.20787197e-01 9.75575149e-02 1.35312036e-01 -4.80649918e-02
-9.58187819e-01 -2.78089851e-01 -2.82541394e-01 1.02455473e+00
1.29060102e+00 1.03632465e-01 -1.16744387e+00 -6.72778487e-02
2.14819983e-01 5.87096438e-02 8.17395091e-01 1.11351085e+00
-4.98316199e-01 -1.95807874e+00 -4.68790621e-01 -1.55382603e-01
-5.67766428e-01 4.67178747e-02 -9.69843447e-01 1.11330354e+00
8.87970775e-02 8.03819478e-01 -4.75478679e-01 1.46014601e-01
5.50630033e-01 1.89494148e-01 5.09242773e-01 -4.45579261e-01
9.26587954e-02 7.87362278e-01 1.91088334e-01 -3.69584918e-01
-3.51841241e-01 -5.70610404e-01 -1.58475924e+00 -2.91617602e-01
-6.79562271e-01 4.18269932e-01 9.64395523e-01 8.76118481e-01
3.24211121e-01 2.51026809e-01 4.73041266e-01 -7.87662268e-01
7.74053782e-02 -4.93991941e-01 -1.02274823e+00 -4.11163360e-01
2.78041124e-01 -6.74904287e-01 2.83334553e-01 1.70428798e-01] | [4.678316593170166, 5.125249862670898] |
70a3c145-76ca-48ae-811f-644d093d7b5a | learning-to-smiles | 1602.06289 | null | http://arxiv.org/abs/1602.06289v2 | http://arxiv.org/pdf/1602.06289v2.pdf | Learning to SMILE(S) | This paper shows how one can directly apply natural language processing (NLP)
methods to classification problems in cheminformatics. Connection between these
seemingly separate fields is shown by considering standard textual
representation of compound, SMILES. The problem of activity prediction against
a target protein is considered, which is a crucial part of computer aided drug
design process. Conducted experiments show that this way one can not only
outrank state of the art results of hand crafted representations but also gets
direct structural insights into the way decisions are made. | ['Damian Leśniak', 'Stanisław Jastrzębski', 'Wojciech Marian Czarnecki'] | 2016-02-19 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 9.45867360e-01 6.38144672e-01 -4.69653904e-01 -2.33641893e-01
-4.01926488e-01 -6.79622948e-01 9.16863561e-01 1.05380774e+00
-3.06839705e-01 1.48249102e+00 4.39093232e-01 -9.66462076e-01
-2.57187486e-01 -6.54906809e-01 -4.61355656e-01 -7.59444773e-01
-8.27411637e-02 4.62079227e-01 -1.37611106e-01 -2.18655705e-01
8.80451798e-01 8.82133365e-01 -1.04927003e+00 6.66392028e-01
6.54471695e-01 5.89258015e-01 1.68520167e-01 4.52243745e-01
-4.49789733e-01 1.17111099e+00 -5.67061484e-01 -1.25593439e-01
-1.61862552e-01 -3.84327114e-01 -1.37274444e+00 -2.19716862e-01
1.31538168e-01 5.99167407e-01 3.99222553e-01 1.12251794e+00
2.74645567e-01 -3.82169373e-02 1.07976246e+00 -4.74991351e-01
-5.82459390e-01 4.63965237e-01 -3.99732977e-01 1.58585042e-01
1.08731413e+00 1.45475613e-02 9.37025011e-01 -8.11338007e-01
8.03431273e-01 1.36365819e+00 1.32304952e-01 2.68952042e-01
-1.32824886e+00 -1.92605644e-01 8.40237141e-02 1.98064357e-01
-9.55823898e-01 -6.72157854e-02 4.55360591e-01 -7.08081484e-01
1.36383271e+00 3.93082350e-01 4.00479227e-01 8.38156879e-01
8.01579237e-01 5.28874040e-01 1.33309650e+00 -4.95790213e-01
4.69267279e-01 2.86694527e-01 6.39241874e-01 5.39000511e-01
2.26100489e-01 2.94104312e-03 -3.67702097e-01 -4.34534699e-01
2.01360449e-01 -1.47026464e-01 -3.06522369e-01 -1.50359228e-01
-1.00604200e+00 1.00330269e+00 3.39212388e-01 4.57291484e-01
-4.07643974e-01 -3.65410000e-01 4.81548935e-01 3.16317558e-01
1.15767702e-01 7.87970424e-01 -7.09950566e-01 6.22953996e-02
-6.50918603e-01 1.91589966e-01 8.80527914e-01 5.39954066e-01
2.99586654e-01 -1.19213402e-01 -1.41904354e-01 2.95304358e-01
5.09502530e-01 2.34839600e-02 5.91467917e-01 -8.74786451e-02
1.18710153e-01 7.39878654e-01 1.77346155e-01 -7.60104835e-01
-3.46877456e-01 -1.22273311e-01 -4.93202657e-01 5.62032640e-01
3.54493529e-01 1.50942300e-02 -1.07941377e+00 1.16044104e+00
2.94537276e-01 -3.87376249e-01 4.18971181e-01 4.74886000e-01
8.28518093e-01 9.62202251e-01 8.38014424e-01 -6.18433475e-01
1.61610126e+00 -5.03082573e-01 -4.91221189e-01 1.44748790e-02
6.73524976e-01 -9.44504976e-01 5.52725911e-01 9.49957073e-01
-6.06456518e-01 -3.48041385e-01 -1.45427251e+00 -1.41902134e-01
-1.06721163e+00 -1.09076865e-01 1.01435316e+00 6.31026804e-01
-4.07108009e-01 1.09536469e+00 -4.18492407e-01 -2.89566427e-01
5.03572822e-01 6.51616275e-01 -6.96712971e-01 9.42705944e-02
-1.00297093e+00 1.41972017e+00 6.77991450e-01 -2.41358325e-01
-5.40419877e-01 -6.36025131e-01 -6.49776399e-01 -1.33704364e-01
3.49662602e-01 -4.50358510e-01 1.04111099e+00 -8.99242580e-01
-1.40779769e+00 8.95103812e-01 -4.02213454e-01 -6.60591424e-01
3.14007968e-01 -1.48749739e-01 -3.25542152e-01 -9.51655954e-02
-2.17029065e-01 3.03011477e-01 4.98278797e-01 -7.94750392e-01
-2.04732254e-01 -4.56699312e-01 -1.72564402e-01 -1.41146379e-02
1.76707894e-01 2.44577140e-01 5.36637068e-01 -4.56531405e-01
-3.30994874e-01 -3.91594380e-01 -5.56111634e-01 -2.21712694e-01
-6.05516851e-01 -5.49885690e-01 5.49898922e-01 -4.75858450e-01
1.17449176e+00 -1.60737216e+00 3.27717125e-01 5.01539767e-01
2.09105149e-01 4.56316203e-01 2.10350975e-01 8.38342965e-01
-9.53564882e-01 3.66120905e-01 -1.25020355e-01 6.67560399e-01
-2.25829333e-01 9.42368358e-02 -4.20754790e-01 4.23173815e-01
3.39317858e-01 6.53210878e-01 -7.86772728e-01 -3.70563179e-01
3.10839593e-01 3.74305934e-01 -2.28179634e-01 -1.57853231e-01
-6.71852231e-01 5.36192656e-01 -9.85541284e-01 5.91133416e-01
4.78437871e-01 -2.69937158e-01 5.98259568e-01 -1.14959754e-01
-3.97499979e-01 5.82807779e-01 -8.09397101e-01 1.32410014e+00
9.19824243e-02 3.15595090e-01 -6.57199979e-01 -1.25029576e+00
8.23752105e-01 4.82026935e-01 3.72149974e-01 -3.40117395e-01
2.72460878e-01 -5.39696105e-02 7.42589831e-02 -4.98691827e-01
7.80031458e-02 -6.49922550e-01 1.95932820e-01 2.36477867e-01
-4.48499508e-02 2.81287849e-01 2.58106261e-01 1.27012916e-02
9.40044522e-01 3.84855658e-01 1.48892617e+00 -5.80267906e-01
1.04959166e+00 4.40638661e-01 8.81713852e-02 2.99074739e-01
2.63111770e-01 8.43236670e-02 8.01321507e-01 -7.53215313e-01
-9.50413883e-01 -7.76520491e-01 -4.39194262e-01 8.43634367e-01
-5.20433486e-01 -5.84985912e-01 -6.44809961e-01 -5.34421921e-01
-7.44198859e-02 5.05175114e-01 -8.47089350e-01 9.13804397e-02
-2.65281349e-01 -9.63958442e-01 9.28527415e-02 1.91006929e-01
-1.61910996e-01 -1.16452932e+00 -4.40963745e-01 4.98595983e-01
4.60536569e-01 -6.57574773e-01 2.88232088e-01 6.91691101e-01
-1.05995023e+00 -1.37761939e+00 -6.83157444e-01 -7.65911877e-01
7.06425250e-01 -2.85149872e-01 1.09276104e+00 -1.05699830e-01
-4.88608420e-01 -5.19674420e-01 -2.12872162e-01 -9.54102457e-01
-6.95971906e-01 -2.02529326e-01 -1.58764601e-01 -3.18407178e-01
8.32739174e-01 -6.80417717e-01 -3.53254616e-01 -2.21524224e-01
-8.92402053e-01 -1.18563809e-01 6.59764171e-01 6.13661945e-01
5.61294973e-01 -1.37504593e-01 5.81475914e-01 -1.38766944e+00
9.94441569e-01 -3.61792535e-01 -5.19651890e-01 5.83009362e-01
-6.49769247e-01 7.60738850e-01 7.79155195e-01 -3.20383042e-01
-7.87123442e-01 4.44242567e-01 -3.33650678e-01 5.34054935e-01
-5.27943194e-01 8.11895728e-01 -3.50672126e-01 -2.36832425e-01
9.34922457e-01 2.27930710e-01 -1.87048048e-01 -5.78863800e-01
3.59665930e-01 6.70118570e-01 2.29645863e-01 -5.97915888e-01
1.68704644e-01 -1.16737820e-02 4.02813792e-01 -1.05239785e+00
-4.21472520e-01 -4.65917677e-01 -5.61361909e-01 3.46177667e-01
8.73860240e-01 -4.42745388e-01 -1.24992907e+00 -3.10242236e-01
-1.31220901e+00 3.90271366e-01 -5.96188940e-04 2.53066629e-01
-6.28308415e-01 5.21969497e-01 -2.80068934e-01 -8.08567941e-01
-4.88079041e-01 -8.61341715e-01 8.10318530e-01 2.08054949e-02
-6.76581144e-01 -9.04174089e-01 3.60087365e-01 4.03757185e-01
-1.64760485e-01 6.54422581e-01 1.83423698e+00 -1.14621246e+00
-2.03460708e-01 -3.62993598e-01 1.41312152e-01 2.88658142e-02
3.12588871e-01 -4.39163223e-02 -9.70744550e-01 2.06450149e-01
6.77141175e-02 -2.90868700e-01 7.02975929e-01 2.17394561e-01
9.79811847e-01 -5.19110501e-01 -5.11438012e-01 -1.31206065e-01
1.77570915e+00 6.06334269e-01 9.94644463e-01 1.42738983e-01
2.25898564e-01 7.82457113e-01 7.77268171e-01 3.10072809e-01
-6.13283932e-01 4.47858244e-01 1.01083130e-01 5.27779572e-02
3.05885971e-01 -3.06254834e-01 1.86687395e-01 2.93789729e-02
-4.17884678e-01 -2.72018880e-01 -8.52742851e-01 -5.63176982e-02
-1.74959517e+00 -9.03702319e-01 -8.34894478e-02 2.11580992e+00
1.15818536e+00 1.69305444e-01 4.73168828e-02 4.51687902e-01
3.48755747e-01 -1.20706044e-01 -5.01345515e-01 -8.35565388e-01
4.44671623e-02 8.53118658e-01 4.72988009e-01 6.55083299e-01
-1.05500007e+00 9.77157891e-01 7.12047625e+00 1.01229453e+00
-1.20078254e+00 -5.08831143e-01 7.61810899e-01 3.46965462e-01
4.41167392e-02 -2.64924634e-02 -6.34972870e-01 2.21951142e-01
1.21366084e+00 -3.64396900e-01 -2.18039472e-03 8.25938344e-01
5.66959560e-01 -3.59794736e-01 -1.60510921e+00 7.84110725e-01
-2.82947749e-01 -1.77130616e+00 7.44510531e-01 9.40770656e-02
3.18002820e-01 -3.93514007e-01 -2.23839372e-01 -2.69279629e-01
3.30454499e-01 -1.82556534e+00 1.89987496e-01 4.84068304e-01
2.71625340e-01 -8.54639053e-01 5.99894345e-01 3.84182155e-01
-8.53771389e-01 1.50723364e-02 -5.07079840e-01 -2.01066479e-01
4.31498401e-02 4.60305631e-01 -1.24654818e+00 7.70393789e-01
-7.63659701e-02 7.20245481e-01 -4.08694744e-01 1.13054299e+00
-3.30328375e-01 2.45137721e-01 -9.91626158e-02 -4.63674128e-01
4.18579757e-01 -1.52545437e-01 2.28086546e-01 1.23892617e+00
-1.21892191e-01 4.17614371e-01 8.99425894e-02 5.77528477e-01
1.53949514e-01 5.60513258e-01 -1.06830096e+00 -6.11932933e-01
-1.65666774e-01 8.06103528e-01 -7.10798621e-01 -5.14619827e-01
-1.25900820e-01 6.20172977e-01 -1.34946331e-02 1.56089544e-01
-2.51477242e-01 -2.95139670e-01 3.85727078e-01 2.33491510e-01
6.90275207e-02 -7.40926415e-02 -2.83185124e-01 -7.18410075e-01
-3.74020100e-01 -1.19340372e+00 3.18340600e-01 -7.30327666e-01
-1.02815557e+00 4.78176892e-01 -1.28477775e-02 -1.06595135e+00
-2.91676912e-02 -1.38919818e+00 -6.51022911e-01 9.53098834e-01
-1.17614937e+00 -7.81151772e-01 6.10347331e-01 3.26116920e-01
6.09328270e-01 -2.16837347e-01 1.32106328e+00 9.30944365e-03
-3.08052272e-01 -1.02970526e-01 1.61880478e-01 -2.60009140e-01
4.52161223e-01 -1.34676802e+00 2.58118939e-02 3.06578279e-01
1.21439561e-01 1.05992436e+00 1.20254672e+00 -5.62362015e-01
-1.21957839e+00 -6.17181599e-01 1.13666129e+00 -4.55385774e-01
5.81214488e-01 1.52619742e-03 -8.13020051e-01 6.16437532e-02
4.55963463e-01 -4.04078305e-01 1.23978364e+00 -1.08155206e-01
-1.92926079e-01 5.45267940e-01 -1.32095647e+00 4.56314445e-01
4.62970465e-01 -2.80701011e-01 -9.21987176e-01 8.26716542e-01
3.98718327e-01 6.19176850e-02 -8.97515655e-01 2.31242314e-01
4.45055783e-01 -5.02376139e-01 1.19159579e+00 -1.75463724e+00
5.00454128e-01 -3.47221732e-01 1.03554301e-01 -9.92844224e-01
2.68175360e-02 -6.81440353e-01 1.36697456e-01 5.34354925e-01
9.83585477e-01 -4.42453861e-01 9.00329947e-01 4.98844981e-01
2.98322916e-01 -9.98277783e-01 -6.03021741e-01 -3.79232287e-01
4.27185208e-01 5.76313958e-02 2.02740520e-01 9.22913909e-01
8.19563508e-01 1.05233979e+00 -1.73701912e-01 -2.50510693e-01
2.18460917e-01 5.23822792e-02 2.79185832e-01 -1.50030541e+00
-3.67045760e-01 -3.81955236e-01 -8.15831482e-01 -8.69216502e-01
-1.72994196e-01 -8.70537102e-01 -3.37125212e-01 -1.44187796e+00
2.98245013e-01 2.60628104e-01 -2.66587853e-01 5.27344167e-01
3.33660513e-01 -2.00199559e-01 -1.44094020e-01 -1.75948918e-01
-3.01779836e-01 7.74577558e-02 1.13885653e+00 -2.64299244e-01
-1.05504043e-01 -8.16006735e-02 -1.08038378e+00 8.87927830e-01
8.29848945e-01 -6.32411659e-01 -5.10961413e-01 3.25585067e-01
4.93331552e-01 1.30586192e-01 -1.05991215e-01 -4.70193833e-01
6.59678206e-02 -6.60335779e-01 5.82619429e-01 -5.03760934e-01
1.35595411e-01 -7.96562433e-01 1.13643445e-01 6.92388535e-01
-6.46948993e-01 -9.82854292e-02 2.18641222e-01 8.30187023e-01
-3.06601286e-01 -5.17205954e-01 7.76542544e-01 -6.02851212e-01
-4.51533914e-01 -5.79974204e-02 -8.43793571e-01 -3.66944313e-01
1.17238402e+00 -4.85150933e-01 -3.04014295e-01 1.28174067e-01
-1.12029016e+00 -1.52940497e-01 -3.88517830e-04 1.66319549e-01
5.77932537e-01 -8.90908062e-01 -3.39908749e-01 -2.41922900e-01
2.27355868e-01 -3.43522578e-01 -4.21991974e-01 7.15981126e-01
-9.95459855e-01 1.14447808e+00 -2.53323406e-01 -1.09206267e-01
-1.93719935e+00 8.27960908e-01 1.17999248e-01 -4.63546902e-01
-2.77391434e-01 5.07374525e-01 2.25587621e-01 1.01991393e-01
3.21018845e-01 -5.01183331e-01 -8.84058833e-01 2.51983013e-02
8.71128142e-01 -4.53795642e-02 2.55450308e-01 -3.83016169e-01
-5.48382044e-01 5.57011425e-01 -5.68960130e-01 2.14438811e-01
1.54870844e+00 6.75791860e-01 -1.23757809e-01 1.61916137e-01
1.10231042e+00 -2.82359034e-01 -4.52746958e-01 9.22662988e-02
7.10882723e-01 -1.29004508e-01 -2.03245804e-01 -1.28541517e+00
-1.17704719e-01 1.10430026e+00 4.75089639e-01 -9.75639001e-03
6.58397794e-01 -4.14094739e-02 2.57718056e-01 1.19692338e+00
2.52877265e-01 -9.34848726e-01 -6.51881546e-02 2.82517940e-01
9.98190224e-01 -1.00286841e+00 7.22269416e-01 -7.06569731e-01
-5.51760375e-01 1.49063253e+00 -1.51887298e-01 -1.51173040e-01
5.81272066e-01 1.67857990e-01 -3.60263437e-01 -6.24010444e-01
-8.02758515e-01 6.13412298e-02 2.43886590e-01 5.67323983e-01
1.14383090e+00 6.84693232e-02 -9.32543516e-01 3.53028446e-01
8.56337175e-02 1.83720320e-01 3.41121703e-01 1.48296738e+00
-9.94242430e-01 -1.78393102e+00 -5.79695284e-01 9.97499228e-02
-1.02649665e+00 -4.27501708e-01 -1.38267827e+00 5.86184740e-01
-1.09348424e-01 8.27978253e-01 -8.36877108e-01 -7.52471983e-02
1.36455178e-01 3.86531860e-01 7.79490829e-01 -9.02665377e-01
-6.48722708e-01 -9.90777910e-02 3.96889836e-01 -1.95456475e-01
-7.13970959e-01 -1.84135363e-01 -1.45470583e+00 -1.60646096e-01
-1.47890866e-01 7.35903502e-01 6.17939055e-01 1.17649949e+00
1.40690237e-01 3.25316936e-01 1.18082508e-01 -5.24339199e-01
-4.37589884e-01 -9.80812013e-01 -2.69589961e-01 4.02144752e-02
3.54955256e-01 -4.59683597e-01 1.37105629e-01 4.52278197e-01] | [5.066187858581543, 5.899891376495361] |
af0da6ae-d729-497f-be52-7c4da71a0dee | semiconductor-defect-detection-by-hybrid-1 | 2208.03514 | null | https://arxiv.org/abs/2208.03514v1 | https://arxiv.org/pdf/2208.03514v1.pdf | Semiconductor Defect Detection by Hybrid Classical-Quantum Deep Learning | With the rapid development of artificial intelligence and autonomous driving technology, the demand for semiconductors is projected to rise substantially. However, the massive expansion of semiconductor manufacturing and the development of new technology will bring many defect wafers. If these defect wafers have not been correctly inspected, the ineffective semiconductor processing on these defect wafers will cause additional impact to our environment, such as excessive carbon dioxide emission and energy consumption. In this paper, we utilize the information processing advantages of quantum computing to promote the defect learning defect review (DLDR). We propose a classical-quantum hybrid algorithm for deep learning on near-term quantum processors. By tuning parameters implemented on it, quantum circuit driven by our framework learns a given DLDR task, include of wafer defect map classification, defect pattern classification, and hotspot detection. In addition, we explore parametrized quantum circuits with different expressibility and entangling capacities. These results can be used to build a future roadmap to develop circuit-based quantum deep learning for semiconductor defect detection. | ['Min Sun', 'YuanFu Yang'] | 2022-08-06 | semiconductor-defect-detection-by-hybrid | http://openaccess.thecvf.com//content/CVPR2022/html/Yang_Semiconductor_Defect_Detection_by_Hybrid_Classical-Quantum_Deep_Learning_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yang_Semiconductor_Defect_Detection_by_Hybrid_Classical-Quantum_Deep_Learning_CVPR_2022_paper.pdf | cvpr-2022-1 | ['defect-detection'] | ['computer-vision'] | [ 3.25896293e-01 1.00847455e-02 -3.38438782e-03 -2.94775724e-01
-7.04833329e-01 -4.01963472e-01 2.20604181e-01 3.43431056e-01
5.73683381e-02 4.06640738e-01 -6.26668990e-01 -5.35494804e-01
4.90596369e-02 -1.43315589e+00 -4.65661168e-01 -8.44437063e-01
4.52889711e-01 5.95153272e-01 8.42663366e-03 -4.48632002e-01
7.34795034e-01 4.92712826e-01 -1.39081550e+00 4.81369764e-01
8.77097070e-01 1.14315510e+00 1.15316324e-01 5.48296928e-01
3.14492770e-02 4.09920216e-01 -3.48372519e-01 -2.29026064e-01
4.01578516e-01 -5.18590689e-01 -4.86786216e-01 -2.06954196e-01
-1.11126915e-01 -3.20045352e-01 -6.67816877e-01 1.47989261e+00
3.76071900e-01 -3.78164232e-01 7.22307146e-01 -8.86202157e-01
-4.63885039e-01 4.65086281e-01 -2.81563461e-01 -4.51391190e-02
7.00740665e-02 4.47131872e-01 1.08034003e+00 -3.95472378e-01
5.57042837e-01 1.08655977e+00 1.50745198e-01 5.23919880e-01
-9.96756077e-01 -7.14640856e-01 -7.70337641e-01 2.35935822e-01
-1.25252533e+00 -2.00005546e-01 6.91010118e-01 -1.68326482e-01
1.34670556e+00 -2.24205121e-01 4.06008720e-01 6.32674158e-01
1.12826216e+00 2.09479094e-01 7.20762014e-01 -6.45161271e-01
5.92769921e-01 9.32799354e-02 1.18720770e-01 1.17202687e+00
6.37365401e-01 6.27180517e-01 -7.00918972e-01 1.06405996e-01
5.04379272e-01 2.11501107e-01 3.89001161e-01 -1.36914879e-01
-8.62729728e-01 8.61386180e-01 5.43278635e-01 3.91215920e-01
-5.85138239e-02 5.85722744e-01 5.75112283e-01 6.79572999e-01
2.09309682e-01 6.62843108e-01 -1.06970847e-01 -1.20825008e-01
-5.54171920e-01 1.10478185e-01 6.14575744e-01 7.36431420e-01
1.15863526e+00 -2.79564619e-01 -2.94315862e-03 7.39133358e-03
2.16732115e-01 7.23284364e-01 2.14982163e-02 -9.50703859e-01
7.90353566e-02 7.67144918e-01 1.84548534e-02 -6.16992533e-01
-6.47813916e-01 -1.53354704e-01 -8.17579865e-01 1.98703632e-01
-6.59736842e-02 -4.41273451e-02 -1.03605425e+00 1.08480644e+00
-1.05351463e-01 -4.45749968e-01 2.09044978e-01 6.04114115e-01
2.40177453e-01 9.55892146e-01 -4.58215773e-01 2.57080458e-02
9.98156309e-01 -5.93215108e-01 -7.26045549e-01 1.50878534e-01
1.27866244e+00 -3.82949173e-01 5.09782672e-01 6.84803903e-01
-7.83204675e-01 -5.11696637e-01 -1.70288467e+00 -2.55257905e-01
-5.31402469e-01 6.94223046e-02 1.15243459e+00 1.24921358e+00
-8.10174882e-01 1.09990799e+00 -1.19878244e+00 -2.32325613e-01
4.29806083e-01 8.10673654e-01 -4.17743996e-03 -6.26327634e-01
-9.80445683e-01 6.39938712e-01 2.26947471e-01 9.17763114e-02
-1.11070538e+00 -4.31468308e-01 -5.75877964e-01 1.64531127e-01
4.82502244e-02 -7.48323619e-01 1.05148447e+00 -1.08638637e-01
-1.58873951e+00 7.15024412e-01 1.41733345e-02 -4.34379250e-01
-7.73072094e-02 4.29411590e-01 -5.23836434e-01 8.11813921e-02
5.31455204e-02 5.85862160e-01 6.07197821e-01 -4.76977676e-01
-5.74801028e-01 -7.33618140e-01 5.06921969e-02 -1.60412759e-01
-2.40071148e-01 -2.70578027e-01 5.55046089e-02 6.57610774e-01
6.27704918e-01 -9.18420434e-01 -4.32770044e-01 -4.16011870e-01
-3.36439699e-01 -3.00183117e-01 7.34861970e-01 1.40205160e-01
7.42066920e-01 -2.22285151e+00 1.26774274e-02 2.34192267e-01
1.85527146e-01 -2.45464832e-01 -2.24724878e-02 6.85773253e-01
2.29670674e-01 1.63253114e-01 -1.36110052e-01 -3.44319753e-02
6.87609017e-02 3.70627828e-02 -1.02534793e-01 5.94626129e-01
4.39537913e-01 8.66375208e-01 -8.23757708e-01 7.14252070e-02
2.00166970e-01 -1.62887648e-01 -7.89695203e-01 -2.37555087e-01
-3.30719888e-01 3.37886781e-01 -6.34969413e-01 1.08427119e+00
8.71667624e-01 -2.17572138e-01 4.06710878e-02 -9.10381228e-02
-1.63644895e-01 4.94440436e-01 -5.56067288e-01 1.78852510e+00
-3.41049910e-01 3.39797467e-01 -1.85130104e-01 -8.88523698e-01
1.04732490e+00 -1.15869194e-01 3.07878613e-01 -1.16760540e+00
2.58817673e-01 7.06975520e-01 3.57117206e-01 -3.65965605e-01
6.43636286e-01 -5.76583922e-01 -5.67262769e-01 3.63429695e-01
1.43039189e-02 -9.16605592e-01 1.05131865e-01 3.51185173e-01
1.67875624e+00 -3.50561678e-01 -2.31064320e-01 -3.12930375e-01
2.79254556e-01 1.65666610e-01 4.10394907e-01 5.96743405e-01
-3.86761785e-01 6.46822602e-02 8.85570228e-01 -3.66546571e-01
-1.28805351e+00 -1.17021668e+00 -3.06737751e-01 4.92258549e-01
5.89928150e-01 -3.15454394e-01 -6.00800991e-01 -4.58580285e-01
4.70649675e-02 5.75118005e-01 6.16741739e-02 -8.86116982e-01
-1.50664449e-01 -1.13052297e+00 4.96967137e-01 2.74745733e-01
4.14346486e-01 -6.88542664e-01 -3.52860808e-01 1.59178197e-01
5.20976245e-01 -9.11508918e-01 1.94956243e-01 8.51963401e-01
-9.26623464e-01 -7.72808015e-01 1.01621464e-01 -6.78240538e-01
7.73810565e-01 1.80186242e-01 7.41618693e-01 -9.64712724e-02
-8.68537188e-01 -3.98617424e-02 -2.03268290e-01 -3.73760849e-01
-8.54710996e-01 9.13100913e-02 3.65056485e-01 -4.13725048e-01
7.42491186e-01 -6.96471557e-02 -5.19800544e-01 -1.15428016e-01
-7.01448917e-01 -1.38572693e-01 8.24058115e-01 8.58982861e-01
4.11732078e-01 1.09590530e+00 5.68091273e-01 -8.69792879e-01
2.79402077e-01 -3.27588737e-01 -9.82898831e-01 -9.45856720e-02
-5.88840008e-01 3.91693980e-01 5.36394536e-01 6.00844383e-01
-8.80263448e-01 5.22400960e-02 -1.30638063e-01 1.76299900e-01
2.51189768e-02 3.81078541e-01 -2.10967571e-01 -4.59634930e-01
7.11683929e-01 -1.76540583e-01 -2.38057286e-01 2.31946558e-01
4.83749919e-02 7.93390036e-01 -8.44989195e-02 -2.21037716e-01
6.88068271e-01 5.56943893e-01 7.13618457e-01 -8.95827413e-01
-5.30590236e-01 -2.49409154e-01 -4.66261119e-01 -3.20274532e-02
9.54650164e-01 -8.26826453e-01 -9.51317906e-01 4.44751173e-01
-1.23274899e+00 1.05294801e-01 7.02773109e-02 3.09270650e-01
-4.41577733e-01 2.19066337e-01 -8.39923382e-01 -9.90871370e-01
-1.14297517e-01 -1.45186818e+00 1.24393320e+00 1.50021449e-01
2.90005475e-01 -6.76903605e-01 -1.32872254e-01 4.81610775e-01
2.42268905e-01 -5.29008061e-02 1.47542989e+00 -2.06427336e-01
-1.39337456e+00 -5.47284544e-01 -4.67908472e-01 4.64121044e-01
1.31241798e-01 -2.46766657e-01 -1.00514627e+00 -4.65210319e-01
2.83046842e-01 -6.02206707e-01 1.15713179e+00 2.86180854e-01
1.11685181e+00 8.35403979e-01 -5.52356541e-01 4.33280051e-01
1.56792080e+00 4.10602748e-01 6.53033733e-01 -3.12842250e-01
5.56827128e-01 4.19662178e-01 6.66506112e-01 3.55874717e-01
1.85560454e-02 1.64155319e-01 6.87685132e-01 4.50525224e-01
2.15774789e-01 -1.00357629e-01 4.18696254e-01 1.16475976e+00
2.91139513e-01 -3.07465196e-01 -1.17368639e+00 1.78196251e-01
-1.32489383e+00 -5.00961542e-01 -1.48712680e-01 2.11533237e+00
2.53774285e-01 5.35299242e-01 -6.46734297e-01 1.15199447e-01
6.49960876e-01 -1.08333178e-01 -8.95907223e-01 -6.46735489e-01
2.36524746e-01 5.26834846e-01 8.70102465e-01 1.27820864e-01
-6.60597205e-01 1.06693506e+00 5.89392614e+00 7.62697041e-01
-1.06074131e+00 2.43230805e-01 7.22313583e-01 1.10047810e-01
-5.18112123e-01 2.62449235e-01 -9.57009673e-01 1.15760714e-01
1.41286767e+00 -6.16185553e-02 6.11618936e-01 5.10194182e-01
-2.16156155e-01 -4.86379147e-01 -1.48825455e+00 1.30084586e+00
-1.94801629e-01 -1.55678070e+00 -2.64443278e-01 3.13636512e-01
9.09200907e-01 3.02087933e-01 2.11371973e-01 5.46499729e-01
-1.00152522e-01 -9.21476483e-01 6.53076693e-02 1.98894486e-01
6.93828881e-01 -1.07975614e+00 8.72010469e-01 1.24498293e-01
-5.84930778e-01 -4.59812433e-01 -1.00367868e+00 -3.61558050e-01
-1.40813351e-01 1.03568912e+00 -7.69395471e-01 7.14578629e-01
2.54686326e-01 6.24900103e-01 -1.43076718e-01 5.83272576e-01
1.15475222e-01 3.06441367e-01 -1.37354285e-01 -4.73191410e-01
2.36839369e-01 -5.37065804e-01 9.55296382e-02 4.37256068e-01
6.14322424e-01 -4.95465994e-02 4.94166976e-03 1.34247148e+00
-4.23134029e-01 -3.83180052e-01 -9.64872062e-01 -8.20947766e-01
3.36189985e-01 1.19863582e+00 -9.20331955e-01 1.33327069e-02
-1.53007448e-01 1.19709146e+00 -1.09498814e-01 -2.14511737e-01
-5.96952319e-01 -6.36153519e-01 5.86642504e-01 -2.94675957e-02
-8.59106146e-03 -4.26954657e-01 -7.44477034e-01 -8.16330791e-01
-2.18975142e-01 -3.30923527e-01 -3.28659832e-01 -2.13512778e-01
-9.91891682e-01 -8.56471956e-02 -7.08180070e-01 -7.71296680e-01
1.42649010e-01 -1.15702307e+00 -6.88696802e-01 5.81020176e-01
-1.47310579e+00 -4.36284721e-01 1.47178471e-01 -3.00392717e-01
-2.80090980e-02 -1.92814410e-01 8.89185369e-01 2.54199862e-01
-6.62778080e-01 1.89205617e-01 5.51914275e-01 -3.16688955e-01
5.04389346e-01 -1.08185649e+00 6.67635083e-01 5.68518162e-01
-1.02479607e-01 1.79927379e-01 4.05613571e-01 -8.00621212e-01
-2.34377480e+00 -1.04908800e+00 4.86707658e-01 -5.55629253e-01
8.34860325e-01 -7.85371661e-01 -2.51706094e-01 1.91675201e-01
-1.28855169e-01 -2.22935379e-01 4.12128568e-01 1.57810599e-01
-4.29577343e-02 -4.41064984e-01 -1.36427045e+00 2.42079124e-01
8.49169552e-01 -9.87331033e-01 1.51908964e-01 8.15653503e-01
8.74044597e-01 -2.42237166e-01 -6.22774601e-01 5.68385780e-01
2.26509333e-01 -1.06765151e+00 3.00703257e-01 -2.56772161e-01
5.39705217e-01 -6.79757744e-02 -3.51339847e-01 -1.04286849e+00
-2.27548927e-01 -5.99864364e-01 3.96325231e-01 5.51997781e-01
5.68915963e-01 -6.85678601e-01 1.19580722e+00 3.44926178e-01
-6.09290302e-01 -5.25646865e-01 -1.15212035e+00 -8.36831212e-01
2.26884514e-01 -4.79496539e-01 3.65308732e-01 3.46767962e-01
3.04474056e-01 2.96157420e-01 4.62174006e-02 4.66579109e-01
5.46088874e-01 9.06795040e-02 1.67231634e-01 -1.24246931e+00
-3.70905221e-01 -2.91128665e-01 -1.02333057e+00 -6.78884208e-01
2.70117037e-02 -1.15055692e+00 1.25677839e-01 -1.23933315e+00
3.33009899e-01 -4.82953995e-01 -5.35346925e-01 3.90653200e-02
4.00338233e-01 1.47573873e-01 -2.20601901e-01 -2.04847902e-01
-7.32873976e-01 6.54728591e-01 1.22465491e+00 -4.85771954e-01
-5.99218439e-03 -8.18388388e-02 -3.74902874e-01 3.21149290e-01
7.94166446e-01 -5.88289320e-01 -1.66050538e-01 -1.51288256e-01
8.86618555e-01 2.97567934e-01 5.53064793e-02 -1.51231241e+00
2.81423807e-01 3.30749989e-01 1.86189488e-01 -6.61170602e-01
4.09359813e-01 -4.42578048e-01 -3.84187281e-01 1.06661880e+00
8.59129727e-02 -4.57590610e-01 -5.37610352e-02 7.04376161e-01
-1.15864038e-01 -4.76584822e-01 8.60324442e-01 -1.41414419e-01
-3.98584843e-01 2.81214863e-01 -5.02265155e-01 -6.92223668e-01
1.10458624e+00 3.46930474e-02 -5.01980007e-01 9.09782350e-02
-2.38141879e-01 1.80573806e-01 7.97391772e-01 2.82551825e-01
6.13233566e-01 -1.02670240e+00 -2.67326832e-01 5.23894012e-01
3.60550284e-01 -1.36122584e-01 6.55420661e-01 5.39692283e-01
-8.43124509e-01 9.73355114e-01 -2.91409910e-01 -6.27196372e-01
-4.16357309e-01 4.95833308e-01 2.75018603e-01 -3.26886445e-01
-2.58790731e-01 9.07474518e-01 1.20783590e-01 -5.52406609e-01
-1.99860170e-01 -6.30258977e-01 7.67138600e-01 -3.45237166e-01
2.87397057e-01 3.58877867e-01 6.01858020e-01 1.27339527e-01
-1.82603046e-01 3.33021373e-01 -3.99340659e-01 1.42697111e-01
9.44329679e-01 1.88842371e-01 -5.14942884e-01 5.79950571e-01
1.11982524e+00 -3.20073545e-01 -6.68759167e-01 3.28221977e-01
1.98685095e-01 2.34550498e-02 5.53223670e-01 -5.51204026e-01
-6.72875404e-01 1.13302445e+00 1.06394553e+00 3.16942245e-01
8.14176083e-01 7.10638016e-02 1.09127641e+00 1.14730525e+00
1.07869744e+00 -1.21801722e+00 1.81632251e-01 5.99937201e-01
-7.96510652e-02 -1.42626977e+00 -1.76867902e-01 -3.65882248e-01
1.84924845e-02 1.28732824e+00 4.29888517e-01 -8.21367577e-02
5.20739019e-01 2.68290639e-01 -3.20272893e-01 -6.18691862e-01
-7.62825847e-01 -1.04916640e-01 -3.75555843e-01 2.69798398e-01
3.23653400e-01 3.20129991e-01 -4.14945967e-02 1.09782025e-01
8.68828874e-03 -2.01373592e-01 8.02321136e-01 1.07249475e+00
-9.02066290e-01 -1.55815017e+00 -9.96794105e-02 8.33696187e-01
-5.08435704e-02 -1.03170991e-01 -1.59178883e-01 1.36488602e-01
3.90438765e-01 8.94738197e-01 2.35234514e-01 -5.96815467e-01
-6.82980865e-02 1.09714955e-01 9.94924605e-01 -1.12809062e+00
-6.68813735e-02 -5.11827290e-01 -1.12935729e-01 -5.50756395e-01
1.02725528e-01 -5.09021044e-01 -1.72801173e+00 -3.01187277e-01
-6.95962012e-01 -1.03277206e-01 1.38043237e+00 9.82022047e-01
5.22840321e-01 8.24100077e-01 8.24447513e-01 -6.06950402e-01
-8.12700689e-01 -6.85102344e-01 -1.13158512e+00 -4.72607195e-01
2.51499228e-02 -6.36853397e-01 -2.18472794e-01 -7.47587800e-01] | [5.554671287536621, 4.9677510261535645] |
dbffaf15-1d15-416e-877f-cbd32289f149 | question-answering-via-web-extracted-tables | 1903.07113 | null | http://arxiv.org/abs/1903.07113v2 | http://arxiv.org/pdf/1903.07113v2.pdf | Question Answering via Web Extracted Tables and Pipelined Models | In this paper, we describe a dataset and baseline result for a question
answering that utilizes web tables. It contains commonly asked questions on the
web and their corresponding answers found in tables on websites. Our dataset is
novel in that every question is paired with a table of a different signature.
In particular, the dataset contains two classes of tables: entity-instance
tables and the key-value tables. Each QA instance comprises a table of either
kind, a natural language question, and a corresponding structured SQL query. We
build our model by dividing question answering into several tasks, including
table retrieval and question element classification, and conduct experiments to
measure the performance of each task. We extract various features specific to
each task and compose a full pipeline which constructs the SQL query from its
parts. Our work provides qualitative results and error analysis for each task,
and identifies in detail the reasoning required to generate SQL expressions
from natural language questions. This analysis of reasoning informs future
models based on neural machine learning. | ['Anthony Tomasic', 'Bhavya Karki', 'Zihua Liu', 'Suhail Barot', 'Matthias Grabmair', 'Lucile Callebert', 'Nithin Haridas', 'Fan Hu'] | 2019-03-17 | null | null | null | null | ['table-retrieval'] | ['natural-language-processing'] | [ 8.15573633e-02 6.86082959e-01 -2.08490863e-01 -8.00417960e-01
-1.50294578e+00 -1.16183829e+00 5.50245285e-01 8.67596447e-01
8.08214098e-02 5.35306990e-01 4.51659143e-01 -7.36563504e-01
-2.06076071e-01 -1.52769506e+00 -1.13476205e+00 2.64682233e-01
2.51732826e-01 7.99824357e-01 7.01002717e-01 -4.59545344e-01
2.05780610e-01 6.01008469e-05 -1.50609851e+00 1.12220788e+00
8.44710946e-01 1.59051192e+00 -2.25091144e-01 5.36125958e-01
-9.82509613e-01 1.67102063e+00 -6.90626323e-01 -1.15496731e+00
1.29384607e-01 -3.83830816e-01 -1.33689916e+00 -2.22942099e-01
7.30251849e-01 -2.05278769e-01 1.11956829e-02 7.57119954e-01
4.35646325e-02 -1.14603244e-01 4.94087756e-01 -1.39924300e+00
-6.10035121e-01 9.80082870e-01 2.53205359e-01 1.50898555e-02
1.02931714e+00 8.58451054e-03 1.52296376e+00 -7.06138909e-01
7.74921775e-01 1.16765440e+00 5.29947281e-01 2.59482205e-01
-9.82248843e-01 -1.38607889e-01 -9.05467942e-02 2.28536069e-01
-5.89722037e-01 -2.29666770e-01 4.37354684e-01 -3.97797376e-01
1.02667916e+00 6.40008390e-01 2.65394300e-01 7.57523298e-01
1.83089584e-01 7.81360924e-01 8.95921588e-01 -3.73572409e-01
5.71683407e-01 6.43096447e-01 5.62727511e-01 8.32349658e-01
4.68326211e-01 -5.91393650e-01 -4.23094183e-01 -2.57987320e-01
5.69678023e-02 -3.39631975e-01 8.78167301e-02 -4.42540824e-01
-1.02237546e+00 7.79851973e-01 4.58404392e-01 1.08347036e-01
-3.90372992e-01 -3.40248365e-03 4.14273918e-01 4.65099752e-01
-1.24663591e-01 7.76869595e-01 -8.93497884e-01 6.29506633e-02
-1.91993564e-01 8.46365809e-01 1.67605805e+00 1.30114424e+00
1.05558097e+00 -8.13073635e-01 -5.26213467e-01 5.09714425e-01
-4.12079766e-02 3.59850645e-01 9.09692049e-02 -1.38730121e+00
1.01922524e+00 1.43339610e+00 1.83266580e-01 -1.12701786e+00
-2.47648552e-01 1.03691481e-02 -7.52916485e-02 -4.55248088e-01
7.83115804e-01 5.19512817e-02 -3.11973602e-01 1.39754581e+00
4.36053991e-01 -8.28228235e-01 2.40092129e-01 3.67216468e-01
1.31857073e+00 5.18265069e-01 -4.25784886e-02 2.91463107e-01
2.06455731e+00 -8.14286411e-01 -9.48515236e-01 -2.93109953e-01
7.49815047e-01 -3.65596861e-01 1.44927204e+00 2.00147390e-01
-1.20372701e+00 -3.73699993e-01 -7.30538905e-01 -6.77533448e-01
-1.10468352e+00 -1.36609569e-01 5.68383873e-01 5.99460065e-01
-7.92296112e-01 3.41188274e-02 -2.78037399e-01 -5.13690233e-01
8.84086192e-02 -1.05446726e-01 -2.29038298e-01 -7.44403824e-02
-1.50149071e+00 7.83741295e-01 3.44169229e-01 -4.17491704e-01
-3.24703962e-01 -9.44110394e-01 -1.05083728e+00 3.30508947e-01
9.58913267e-01 -8.16674411e-01 1.60808671e+00 -3.35342705e-01
-8.52182746e-01 1.09416270e+00 -3.22477192e-01 -6.52544916e-01
1.78918853e-01 1.90794300e-02 -3.69068682e-01 1.78851798e-01
4.98052597e-01 3.50238949e-01 1.43471569e-01 -1.29805398e+00
-7.37959027e-01 -5.47540903e-01 7.83388734e-01 -3.31568360e-01
-3.80947851e-02 4.88325581e-02 -6.22095227e-01 -2.74239391e-01
1.18108913e-01 -3.30795676e-01 3.79097797e-02 -1.66257937e-02
-6.02503300e-01 -4.74573970e-01 1.19369820e-01 -8.86981606e-01
1.51842010e+00 -1.66712081e+00 -4.78336424e-01 3.25573832e-01
3.85339558e-01 -4.88091797e-01 1.65525258e-01 7.98804760e-01
9.29536223e-02 4.53500122e-01 -3.21942091e-01 4.73650575e-01
5.17824173e-01 1.02652885e-01 -7.78207541e-01 -6.22584641e-01
5.35823286e-01 1.29409015e+00 -5.83449721e-01 -6.58958018e-01
-5.09264469e-01 -2.68551439e-01 -8.13308179e-01 5.11969328e-01
-1.17261684e+00 -4.56828266e-01 -5.76314211e-01 8.50130796e-01
5.39025962e-01 -5.43828070e-01 2.87357301e-01 -5.33936024e-01
1.64125368e-01 1.03437686e+00 -1.16404092e+00 1.27256703e+00
-3.85215014e-01 1.49223313e-01 -7.73454159e-02 -5.28992414e-01
1.00058973e+00 5.11473045e-02 2.05173582e-01 -9.03414130e-01
-1.23648860e-01 2.80392319e-01 -4.33879256e-01 -1.11091590e+00
5.08131742e-01 4.01132196e-01 -4.45223808e-01 6.81186557e-01
7.35737849e-03 -4.21023637e-01 7.81126678e-01 5.12591600e-01
1.62315774e+00 8.29892159e-02 3.53805661e-01 -2.24816680e-01
7.86453605e-01 4.89045322e-01 6.30037785e-02 9.52407479e-01
1.22837231e-01 1.50822103e-01 1.16249490e+00 -6.61930859e-01
-7.45067835e-01 -1.32759011e+00 -5.32013476e-02 1.19023526e+00
-8.84914771e-02 -8.79245579e-01 -8.44190896e-01 -8.75363648e-01
3.89342576e-01 9.17371869e-01 -8.15906048e-01 2.46382952e-02
-7.02190757e-01 -2.33213007e-01 3.86780560e-01 5.76046407e-01
5.19022584e-01 -1.14442563e+00 -6.41323864e-01 1.29166827e-01
-6.61216319e-01 -1.27142644e+00 -1.87497720e-01 3.47250342e-01
-6.03182971e-01 -1.64386714e+00 2.51682758e-01 -8.49281371e-01
3.59810233e-01 -2.81294048e-01 1.97403669e+00 -2.53824145e-02
5.29550612e-02 6.99926853e-01 -3.31260562e-01 -6.33428752e-01
-5.18765271e-01 4.70852643e-01 -7.36899674e-01 2.74193473e-02
8.76615644e-01 -1.78346798e-01 -2.92970747e-01 2.35542670e-01
-1.31730258e+00 -3.09025049e-01 3.79407704e-01 2.89018661e-01
4.86405104e-01 -2.77865119e-03 3.27135593e-01 -1.27969396e+00
8.68569613e-01 -7.97753632e-01 -6.98387742e-01 8.65382075e-01
-4.21195626e-01 5.52544534e-01 6.11021280e-01 2.80066133e-01
-9.86986935e-01 -8.03934634e-02 -1.68063715e-01 5.56521714e-01
-1.35084823e-01 8.09480011e-01 -6.24214470e-01 5.10176718e-01
1.02395594e+00 1.19935863e-01 -5.63870743e-02 -5.65649986e-01
4.71898586e-01 4.05493647e-01 4.50524807e-01 -1.05247712e+00
1.02564549e+00 2.82342583e-01 -2.15819687e-01 -2.13256985e-01
-1.20750523e+00 -1.21874250e-01 -4.02308106e-01 -9.61092636e-02
8.62205088e-01 -5.51964939e-01 -1.17512476e+00 -1.35156080e-01
-1.03943837e+00 -1.63372099e-01 -5.81987739e-01 -2.41832539e-01
-5.42821527e-01 -1.25340432e-01 -5.27850747e-01 -7.15441108e-01
-9.08701569e-02 -8.52614641e-01 1.00935245e+00 2.61260439e-02
-3.06078374e-01 -7.79839218e-01 7.29722753e-02 7.99561799e-01
4.61549252e-01 6.54924586e-02 1.73963737e+00 -1.21427393e+00
-1.09623170e+00 -2.63066351e-01 -2.98217773e-01 2.77789552e-02
-1.65307894e-02 -1.97493106e-01 -6.81365192e-01 4.79777455e-01
1.16482615e-01 -7.81714916e-01 5.41246533e-01 -3.62715364e-01
1.30994594e+00 -6.79047942e-01 1.04515972e-02 1.51475057e-01
1.50240743e+00 2.81293154e-01 6.38838410e-01 7.88108885e-01
2.32003555e-01 1.34866953e+00 5.06959796e-01 1.32966399e-01
1.20735800e+00 2.87843108e-01 4.17669386e-01 4.09535527e-01
2.51757681e-01 -5.83585560e-01 1.13878377e-01 4.36765492e-01
7.03768492e-01 -1.54703587e-01 -1.20870697e+00 5.09992123e-01
-1.48385727e+00 -9.26520646e-01 -3.26375633e-01 1.91720939e+00
1.21810985e+00 1.18433744e-01 1.78249046e-01 2.66101241e-01
2.90381908e-01 4.87724831e-03 -4.65253830e-01 -4.00052518e-01
-1.42090127e-01 3.07351172e-01 2.77851499e-03 4.87379342e-01
-8.90253663e-01 5.55794835e-01 6.29702616e+00 2.18038768e-01
-3.62979352e-01 -4.35982555e-01 5.71190298e-01 3.60344738e-01
-1.03527904e+00 1.45840749e-01 -9.29314137e-01 2.09740117e-01
1.12735927e+00 -3.00178289e-01 5.13380110e-01 8.54578733e-01
-4.79140460e-01 -2.80574441e-01 -1.49870539e+00 3.82001460e-01
5.80575056e-02 -1.56276023e+00 6.11055195e-01 -3.39077055e-01
1.20971099e-01 -6.58100963e-01 -1.84829056e-01 5.88974178e-01
3.95927161e-01 -8.33232880e-01 6.34660661e-01 6.59726620e-01
3.29663813e-01 -3.98200184e-01 6.44696057e-01 1.38819054e-01
-1.06449509e+00 -3.68192106e-01 -1.54390298e-02 1.54616311e-01
-3.82375121e-01 4.41708684e-01 -6.97187483e-01 6.15538836e-01
1.08931351e+00 9.44443047e-02 -1.23847258e+00 5.34583211e-01
-1.30901739e-01 5.40984571e-01 -2.32917354e-01 -3.73558998e-01
-1.10271394e-01 -2.01185003e-01 -6.47384599e-02 9.82231319e-01
3.06532029e-02 2.94454489e-02 -1.75237879e-02 1.22637165e+00
-4.63167191e-01 2.98884809e-01 -3.99368584e-01 -9.07430723e-02
5.23557305e-01 1.17728007e+00 -4.42863941e-01 -6.36228323e-01
-6.16917431e-01 2.96522349e-01 4.32896197e-01 2.92290241e-01
-6.85723484e-01 -9.30477798e-01 4.04248804e-01 5.16088247e-01
2.49291688e-01 3.02495897e-01 -5.24983406e-01 -1.30105543e+00
8.75094473e-01 -1.33497524e+00 8.84697139e-01 -1.05703044e+00
-1.35014737e+00 5.85026383e-01 5.06243259e-02 -6.66808486e-01
-5.43044686e-01 -7.74501741e-01 -1.51033044e-01 7.23632038e-01
-1.39415169e+00 -8.40720415e-01 -6.96410418e-01 6.20207489e-01
7.96767622e-02 3.45892534e-02 8.94094110e-01 3.22178751e-01
-3.92768204e-01 4.90248471e-01 -4.59442943e-01 6.64265811e-01
7.05266058e-01 -1.71052873e+00 7.45384514e-01 4.67216909e-01
2.34132279e-02 9.37996209e-01 6.00110769e-01 -5.08483946e-01
-1.89205825e+00 -1.03475392e+00 1.57603586e+00 -1.18918991e+00
9.43686545e-01 -7.46852815e-01 -1.12567210e+00 1.07680607e+00
2.78812170e-01 -1.46422371e-01 7.40297139e-01 4.53147739e-02
-9.08170819e-01 -5.56012988e-01 -1.25537467e+00 4.52631056e-01
7.24007070e-01 -8.91459823e-01 -1.02144670e+00 4.78495091e-01
1.16090381e+00 -4.02140349e-01 -9.69095767e-01 2.83950329e-01
4.81628180e-01 -1.15078115e+00 8.03886533e-01 -1.06705868e+00
9.54232156e-01 -2.96234041e-01 -6.34102345e-01 -6.03395641e-01
1.09395355e-01 -2.59137809e-01 -4.37494874e-01 1.42576098e+00
1.01564991e+00 -6.11061811e-01 9.10898149e-01 1.13793623e+00
2.76243567e-01 -9.88302767e-01 -3.91800106e-01 -4.76196468e-01
1.62795499e-01 -4.44733441e-01 1.08747065e+00 6.74990296e-01
3.27585414e-02 5.87752223e-01 6.23557985e-01 7.45700225e-02
3.77329797e-01 5.41569769e-01 9.40549195e-01 -1.19847393e+00
-2.73695648e-01 -1.77669778e-01 1.61396205e-01 -9.33445811e-01
-1.72939509e-01 -1.03764153e+00 -7.55392089e-02 -1.99774075e+00
-3.83028761e-02 -3.02714676e-01 1.11762248e-01 4.45572287e-01
-2.61646837e-01 -4.28131133e-01 9.81754959e-02 -1.68665335e-01
-7.20401525e-01 4.14523780e-02 7.71435440e-01 -3.06419641e-01
1.24818131e-01 -5.29577769e-02 -1.21546364e+00 3.03081542e-01
6.60937488e-01 -4.44013834e-01 -3.68046254e-01 -4.55043763e-01
1.12962139e+00 3.82635713e-01 2.68065959e-01 -9.79845703e-01
4.15819883e-01 -9.44232270e-02 2.66907990e-01 -6.42388582e-01
6.10031672e-02 -1.02313721e+00 -2.81237096e-01 2.19567552e-01
-9.82668340e-01 4.77558196e-01 9.89127625e-03 1.91622660e-01
-5.30003488e-01 -2.20816135e-01 2.24296123e-01 -3.91577780e-01
-2.98127174e-01 -6.11720197e-02 -2.83745825e-01 8.54173362e-01
6.87130213e-01 1.99914515e-01 -5.84084034e-01 -5.39141655e-01
-3.36872071e-01 4.70549673e-01 1.39438659e-01 4.32449341e-01
3.70218158e-01 -1.17209089e+00 -5.05258739e-01 1.68765575e-01
6.40406728e-01 -4.77415323e-02 -2.33829841e-01 2.60577857e-01
-6.02991998e-01 7.07939208e-01 3.11859436e-02 -1.73945233e-01
-5.42243302e-01 8.08850825e-01 4.66224700e-01 -6.45459354e-01
8.04411471e-02 5.25654197e-01 -1.22467309e-01 -1.17445469e+00
4.66860890e-01 -8.60262036e-01 -3.83997560e-01 2.38044545e-01
6.46858573e-01 1.85512528e-01 4.51490521e-01 2.61143833e-01
-2.62752563e-01 1.98053926e-01 3.21129411e-02 -9.83204618e-02
9.81601775e-01 1.32415071e-01 -7.19287992e-01 6.55400455e-01
1.16454172e+00 2.28582859e-01 -4.13757682e-01 -3.94078672e-01
7.16605842e-01 -1.64823651e-01 -8.12423944e-01 -1.36311591e+00
-6.44718766e-01 5.87683797e-01 -7.85284117e-02 8.05016518e-01
1.07418883e+00 4.09702837e-01 9.11793947e-01 1.14403331e+00
4.26723331e-01 -7.30939507e-01 2.05085605e-01 7.24174142e-01
1.08244395e+00 -1.17003548e+00 -2.11705595e-01 -6.26076221e-01
-3.05725783e-01 1.12441826e+00 9.02525008e-01 1.18626341e-01
5.14628708e-01 4.29135084e-01 3.13750476e-01 -6.61320508e-01
-1.15555310e+00 -2.15026602e-01 2.60499746e-01 3.41395378e-01
4.33285445e-01 -4.33606774e-01 -1.69575214e-01 1.03576922e+00
-7.35049307e-01 -5.43984771e-03 4.22979534e-01 1.24046564e+00
-4.60916430e-01 -1.24780130e+00 -4.01911378e-01 9.07490790e-01
-5.19261062e-01 -1.28018767e-01 -8.03874195e-01 7.52442718e-01
-1.84875727e-01 1.18080986e+00 1.61410257e-01 -4.78561252e-01
8.98323774e-01 5.13502002e-01 2.13036120e-01 -5.71897626e-01
-1.19763815e+00 -9.99927521e-01 4.66171116e-01 -8.41008902e-01
1.08930305e-01 -4.11973059e-01 -1.32878029e+00 -2.34057590e-01
3.09205443e-01 5.78360260e-01 5.78383923e-01 6.69901133e-01
4.92121309e-01 3.25738251e-01 4.15857852e-01 4.66653109e-01
-7.30689585e-01 -5.38132370e-01 -1.48231372e-01 5.95794857e-01
2.59636968e-01 -7.04653785e-02 -3.08655977e-01 9.84759629e-02] | [9.976139068603516, 7.890412330627441] |
1f3505f7-48e4-4760-9612-1aeeaf266b8a | enhancing-few-shot-text-to-sql-capabilities | 2305.12586 | null | https://arxiv.org/abs/2305.12586v1 | https://arxiv.org/pdf/2305.12586v1.pdf | Enhancing Few-shot Text-to-SQL Capabilities of Large Language Models: A Study on Prompt Design Strategies | In-context learning (ICL) has emerged as a new approach to various natural language processing tasks, utilizing large language models (LLMs) to make predictions based on context that has been supplemented with a few examples or task-specific instructions. In this paper, we aim to extend this method to question answering tasks that utilize structured knowledge sources, and improve Text-to-SQL systems by exploring various prompt design strategies for employing LLMs. We conduct a systematic investigation into different demonstration selection methods and optimal instruction formats for prompting LLMs in the Text-to-SQL task. Our approach involves leveraging the syntactic structure of an example's SQL query to retrieve demonstrations, and we demonstrate that pursuing both diversity and similarity in demonstration selection leads to enhanced performance. Furthermore, we show that LLMs benefit from database-related knowledge augmentations. Our most effective strategy outperforms the state-of-the-art system by 2.5 points (Execution Accuracy) and the best fine-tuned system by 5.1 points on the Spider dataset. These results highlight the effectiveness of our approach in adapting LLMs to the Text-to-SQL task, and we present an analysis of the factors contributing to the success of our strategy. | ['Dragomir Radev', 'Arman Cohan', 'Ellen Zhang', 'Jaesung Tae', 'Narutatsu Ri', 'Weijin Zou', 'Yilun Zhao', 'Linyong Nan'] | 2023-05-21 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 1.62799746e-01 -2.37315655e-01 -2.92256087e-01 -6.00079656e-01
-1.12098348e+00 -6.43231034e-01 6.98984563e-01 5.05906463e-01
-6.26685381e-01 3.17819446e-01 2.76243627e-01 -7.94221997e-01
-1.34952545e-01 -5.98824263e-01 -8.77602875e-01 9.46574435e-02
1.07830197e-01 2.22341537e-01 5.98042130e-01 -3.61082315e-01
7.76081622e-01 5.47503412e-01 -1.81742394e+00 7.86767900e-01
9.21695113e-01 8.65674555e-01 5.67497015e-01 8.90788376e-01
-6.07847095e-01 1.23361552e+00 -8.38270247e-01 -8.51493776e-02
-6.53441921e-02 -1.24699064e-02 -9.32119906e-01 -5.47802806e-01
8.85887206e-01 -4.72812414e-01 -5.75257875e-02 3.88876766e-01
3.49245191e-01 2.82349497e-01 3.09103996e-01 -1.19464672e+00
-3.90305907e-01 4.50048685e-01 -8.17825571e-02 3.44858110e-01
9.57821727e-01 3.31356287e-01 9.01462674e-01 -9.86415744e-01
4.92905408e-01 1.20990896e+00 4.99781549e-01 3.19589913e-01
-1.21317923e+00 -3.63964200e-01 1.49217606e-01 4.41873670e-01
-1.09392154e+00 -6.21578693e-01 6.44634366e-01 -4.79804695e-01
1.61823356e+00 2.20523074e-01 1.37663737e-01 1.04625773e+00
2.64764875e-01 1.05340552e+00 1.21644938e+00 -1.01595855e+00
1.91812873e-01 3.56884420e-01 6.28703952e-01 7.43206501e-01
1.67833990e-03 2.46019647e-01 -9.97827590e-01 -3.47561985e-01
4.11314994e-01 -6.39967546e-02 1.64276846e-02 -3.19109976e-01
-1.15193319e+00 7.80242205e-01 6.11843579e-02 1.83194220e-01
-2.57953852e-01 -5.56606874e-02 5.14668405e-01 4.95397598e-01
-1.54561907e-01 9.42920327e-01 -6.97330415e-01 -5.36377311e-01
-7.08741665e-01 4.00331318e-01 1.25716281e+00 1.26643121e+00
7.86103547e-01 -9.23225656e-02 -4.25131649e-01 7.48566926e-01
-3.25876959e-02 2.14081720e-01 6.60880685e-01 -1.03460693e+00
5.52988350e-01 7.66693354e-01 1.52800053e-01 -5.94478607e-01
-3.03334683e-01 -2.48767547e-02 2.70580351e-01 -8.02325550e-03
5.09906411e-01 -7.24569848e-03 -5.44482946e-01 1.70626795e+00
2.25424275e-01 -3.36823881e-01 1.20591491e-01 3.97363156e-01
8.34698558e-01 5.28370917e-01 4.10274446e-01 3.08563299e-02
1.28395307e+00 -1.06500113e+00 -5.17000973e-01 -3.66228342e-01
1.08819818e+00 -8.22898626e-01 2.07521796e+00 4.23992038e-01
-9.07987416e-01 -8.74890625e-01 -1.00303411e+00 -2.65528381e-01
-6.08562112e-01 8.37646425e-02 6.14391923e-01 4.72198904e-01
-9.41449583e-01 5.87619662e-01 -6.77829623e-01 -6.52862966e-01
-9.67676342e-02 1.11515597e-01 5.30035570e-02 -1.81483671e-01
-8.09070110e-01 1.02458465e+00 3.57011795e-01 -5.72941065e-01
-7.95040965e-01 -9.01923716e-01 -7.74503827e-01 1.73069656e-01
9.15371776e-01 -4.62152094e-01 1.65434515e+00 -3.33488345e-01
-1.40303135e+00 5.13560176e-01 -3.69446844e-01 -3.42344493e-01
-2.90710125e-02 -6.55690134e-01 -2.64647424e-01 2.14834362e-01
1.94100633e-01 5.32649696e-01 5.29799283e-01 -1.00296760e+00
-7.34694719e-01 -3.01450133e-01 2.96490580e-01 9.92512554e-02
-3.65306795e-01 1.56375021e-01 -4.58653003e-01 -4.25729752e-01
-2.60892242e-01 -8.14068317e-01 -4.07991558e-02 -4.18689907e-01
-1.62191197e-01 -6.49476409e-01 6.96410537e-01 -5.40605724e-01
1.61410224e+00 -1.98184049e+00 -3.43483418e-01 -6.06847107e-02
-8.35227221e-02 2.69954383e-01 -3.27234119e-01 8.45763862e-01
1.93891063e-01 -1.07953874e-02 1.35519981e-01 -1.13350563e-02
2.59124696e-01 1.09820753e-01 -5.17130792e-01 -4.19931918e-01
2.92867929e-01 1.06487656e+00 -7.45510101e-01 -8.12167466e-01
2.11763635e-01 -5.22511713e-02 -7.40781367e-01 7.77016997e-01
-6.22518361e-01 -4.81122471e-02 -5.66935122e-01 7.88956046e-01
-8.67634714e-02 -2.86309391e-01 7.47858435e-02 -1.71686381e-01
-1.92938626e-01 7.84480512e-01 -8.91285121e-01 1.83553362e+00
-7.52459764e-01 5.37665308e-01 -2.89711714e-01 -7.33047009e-01
9.00904596e-01 -6.26580119e-02 7.12172911e-02 -8.65613401e-01
-3.70700270e-01 1.86854646e-01 -2.21064426e-02 -1.05506253e+00
6.81953251e-01 2.49221429e-01 -1.59743711e-01 4.44285721e-01
4.91761276e-03 -2.28862882e-01 3.91862869e-01 4.30660963e-01
1.30293131e+00 4.45451915e-01 4.21196818e-01 -1.95606902e-01
3.83239388e-01 4.10013318e-01 2.10697800e-02 1.21308625e+00
-2.96546131e-01 3.37537490e-02 4.23054963e-01 -2.00851619e-01
-6.89722359e-01 -9.23074901e-01 1.84166774e-01 1.79791355e+00
-2.88309157e-01 -7.01246738e-01 -7.08180308e-01 -8.40371788e-01
1.17483303e-01 1.39216340e+00 -1.72391772e-01 -9.14507359e-02
-7.15540469e-01 -1.36140540e-01 5.18505752e-01 7.50768363e-01
4.35364187e-01 -1.39564574e+00 -1.01037598e+00 1.79321513e-01
-8.61086845e-02 -1.28010261e+00 -4.79118258e-01 3.69205773e-01
-1.03638566e+00 -1.17904174e+00 9.53716785e-02 -8.07094038e-01
1.78043395e-01 1.30407214e-01 1.50375068e+00 8.93594250e-02
-3.25780749e-01 1.05134952e+00 -4.12096918e-01 -4.67025697e-01
-5.26520848e-01 2.48389065e-01 -1.63850501e-01 -6.37695968e-01
8.73661399e-01 -3.46610010e-01 -3.14972281e-01 2.47920170e-01
-9.70779717e-01 -8.29324573e-02 7.79007673e-01 6.98461473e-01
4.70314324e-01 -4.09860998e-01 8.81240964e-01 -1.00366592e+00
1.18616450e+00 -2.43487582e-01 -6.11954570e-01 6.52933657e-01
-1.00078821e+00 5.04317880e-01 6.34691954e-01 -4.86979216e-01
-1.14671171e+00 4.02578115e-02 -8.37704539e-02 -2.52937526e-01
-5.09878516e-01 8.06525350e-01 -3.00765224e-02 -7.02412575e-02
9.96674955e-01 1.92634463e-01 -5.96289299e-02 -4.39355820e-01
4.11404014e-01 5.38503230e-01 4.27193999e-01 -1.19475591e+00
2.41573885e-01 -2.44031101e-01 -1.86684787e-01 -8.11007857e-01
-6.66216314e-01 -5.06299615e-01 -5.18830180e-01 8.21376592e-03
5.84688425e-01 -6.43119872e-01 -9.05462325e-01 -1.37582282e-03
-9.74610388e-01 -6.09849632e-01 -2.22084746e-01 2.64316618e-01
-7.00593770e-01 4.59832072e-01 -8.01015258e-01 -7.79625475e-01
-3.07830900e-01 -1.27679002e+00 8.84961665e-01 2.22764909e-01
-5.49180269e-01 -7.14658797e-01 -1.15820691e-02 5.25296271e-01
6.21348619e-01 -8.90120715e-02 1.40985882e+00 -1.23827136e+00
-8.14643383e-01 -1.47155568e-01 -6.13921769e-02 2.90294290e-01
-8.26301202e-02 -2.01082960e-01 -7.98296869e-01 -3.51668037e-02
1.22493975e-01 -7.42268324e-01 5.10027111e-01 -8.67353287e-03
1.42431200e+00 -2.30563223e-01 -3.26805949e-01 2.12592125e-01
1.42577100e+00 3.65050644e-01 2.83471197e-01 5.93671739e-01
4.35112745e-01 7.34535933e-01 1.06999755e+00 1.61111370e-01
4.79288757e-01 6.66234195e-01 1.12649173e-01 4.11471963e-01
-1.48073256e-01 -5.82219958e-01 4.43891793e-01 7.00774908e-01
5.19502103e-01 1.83146194e-01 -1.10025001e+00 6.52382493e-01
-1.73463786e+00 -6.38095856e-01 8.73641372e-02 2.34944940e+00
1.08503890e+00 3.14216316e-01 -2.38452144e-02 -3.43080759e-01
7.93404132e-02 1.49867488e-02 -6.43093288e-01 -5.60951710e-01
2.74971575e-01 4.75479364e-01 2.37843357e-02 4.74374563e-01
-7.85683453e-01 1.12707496e+00 6.76077795e+00 8.84763539e-01
-8.36036205e-01 -3.42986882e-01 1.75435737e-01 -7.46869296e-02
-8.90781134e-02 3.78055498e-02 -1.12401986e+00 7.35682696e-02
1.34782088e+00 -7.63475671e-02 4.28098649e-01 1.19274080e+00
7.60232583e-02 -4.46269184e-01 -1.72370362e+00 6.13434315e-01
1.83399513e-01 -1.26630020e+00 1.90733537e-01 -3.92279357e-01
3.77270252e-01 -1.25483304e-01 -1.03682868e-01 9.93250310e-01
8.05424824e-02 -7.52270818e-01 3.49765211e-01 4.70667064e-01
5.22069991e-01 -4.81037825e-01 3.59565675e-01 6.84349477e-01
-1.02757096e+00 -2.47730508e-01 -7.12660030e-02 -1.11558810e-01
-1.13835499e-01 -6.39088675e-02 -1.35999572e+00 3.19125533e-01
7.57913768e-01 8.23228806e-02 -1.08197010e+00 7.07188487e-01
-1.84181094e-01 8.07153761e-01 -2.79089093e-01 -6.03738248e-01
-4.53195162e-02 2.61249304e-01 2.19336718e-01 1.56437266e+00
1.01292744e-01 1.80698454e-01 5.21167874e-01 8.87809157e-01
1.45940900e-01 3.51440787e-01 -7.17844307e-01 -2.10751340e-01
7.28845537e-01 9.05534446e-01 -1.23326987e-01 -6.83407426e-01
-6.12237871e-01 6.11522019e-01 5.98019779e-01 5.31636775e-01
-5.68244576e-01 -7.95429349e-01 2.99607605e-01 2.71347851e-01
2.23701626e-01 -3.52825493e-01 -4.46459442e-01 -9.62726712e-01
2.57432908e-01 -1.39021266e+00 4.62174356e-01 -1.09480584e+00
-1.15697503e+00 5.07475197e-01 4.99434471e-01 -7.46101439e-01
-6.30422056e-01 -7.03456998e-01 -5.31541348e-01 9.56781447e-01
-1.54777431e+00 -8.79719198e-01 -2.71018058e-01 5.10162413e-01
8.95777404e-01 -3.20626795e-01 9.64294791e-01 2.86360439e-02
-2.79852629e-01 6.73351645e-01 -2.36988604e-01 -2.07772359e-01
1.00664973e+00 -1.31808591e+00 2.61315882e-01 7.22407937e-01
1.43953398e-01 1.25319791e+00 5.39240420e-01 -7.79138029e-01
-1.81225145e+00 -6.98891699e-01 1.15714633e+00 -9.39577222e-01
7.97897100e-01 -1.63211361e-01 -1.18337643e+00 8.37387323e-01
3.44899327e-01 -4.28964913e-01 9.22859430e-01 3.06870878e-01
-5.89931786e-01 -3.05394661e-02 -9.20566380e-01 7.94719577e-01
8.65828037e-01 -9.41306353e-01 -1.21165383e+00 2.40184873e-01
1.22541273e+00 -4.63252872e-01 -9.20400023e-01 4.82160866e-01
4.71957743e-01 -9.46687996e-01 1.00573099e+00 -9.32541013e-01
7.34073162e-01 -5.04979268e-02 -5.01380861e-01 -9.11349356e-01
-4.00500698e-03 -3.82990628e-01 -3.60205293e-01 1.01803267e+00
3.52318794e-01 -4.57715362e-01 7.38370001e-01 9.11601186e-01
-3.74719381e-01 -9.98141170e-01 -4.64857608e-01 -7.00565577e-01
-1.11701535e-02 -7.28015423e-01 4.89475965e-01 6.44261122e-01
2.78967559e-01 6.12077415e-01 1.48519456e-01 -1.78954005e-02
3.48030776e-01 3.59549254e-01 1.11905110e+00 -8.23292911e-01
-5.17336309e-01 -4.17948455e-01 2.81870574e-01 -1.66791451e+00
1.36998534e-01 -6.90254569e-01 8.55753496e-02 -1.18537474e+00
1.15983695e-01 -3.58517796e-01 -2.46507004e-01 4.56671894e-01
-5.65173686e-01 -6.88312769e-01 5.02999842e-01 1.37732262e-02
-7.84286618e-01 2.50581741e-01 8.69369328e-01 6.93035871e-02
-5.33173084e-01 1.38769494e-02 -6.64033890e-01 4.45757329e-01
7.09721208e-01 -1.46429569e-01 -5.20052731e-01 -3.25005859e-01
8.13289359e-02 2.74050087e-01 2.40014538e-01 -1.00159395e+00
5.20954669e-01 -2.54136354e-01 1.58363372e-01 -8.56218517e-01
3.01537722e-01 -7.25803256e-01 -4.98691618e-01 2.83208698e-01
-9.46769476e-01 3.38682860e-01 7.36486733e-01 4.92174327e-01
-1.60924420e-01 -6.56075656e-01 1.72039986e-01 -1.64480194e-01
-1.12829030e+00 -2.46873200e-01 -4.22986418e-01 1.80332989e-01
7.08019197e-01 -8.64013210e-02 -4.94084656e-01 -2.18820974e-01
-3.77295077e-01 4.11335826e-01 2.42902175e-01 5.68233609e-01
6.85332417e-01 -1.00920689e+00 -3.03770512e-01 2.41653293e-01
4.74149555e-01 -2.33296305e-01 -9.14368182e-02 4.95082289e-01
-3.54817361e-01 9.53054368e-01 1.26180639e-02 -6.63422823e-01
-1.35000837e+00 8.04356098e-01 -1.20388173e-01 -4.64893460e-01
-2.40348220e-01 6.68430448e-01 2.77944449e-02 -6.52912259e-01
6.75781608e-01 -6.52170956e-01 -1.77511409e-01 -3.43934625e-01
4.09590036e-01 3.50320160e-01 2.08331570e-01 1.87344536e-01
-1.75445423e-01 2.41313592e-01 -3.03090185e-01 -1.01316437e-01
1.00265968e+00 1.33260563e-01 1.77168116e-01 6.97873056e-01
1.15562737e+00 1.62534401e-01 -9.63055849e-01 -5.22479892e-01
5.92757106e-01 -3.65243822e-01 -3.25325817e-01 -1.28829288e+00
-2.46323287e-01 7.81310201e-01 4.15013045e-01 2.08889425e-01
1.09927988e+00 3.38468351e-03 6.69465303e-01 1.11846197e+00
5.76230764e-01 -1.07246411e+00 5.46133697e-01 6.93099141e-01
9.56934035e-01 -1.52672923e+00 -9.61267389e-03 -2.96518266e-01
-6.51895165e-01 1.22407579e+00 1.16276920e+00 1.64144292e-01
3.93398166e-01 1.71927601e-01 -5.50317951e-03 -2.74456561e-01
-1.15610993e+00 1.16394404e-02 1.57824397e-01 5.84153056e-01
7.58406818e-01 -1.96816158e-02 -2.41053179e-01 6.80213094e-01
-2.31315106e-01 1.74676076e-01 3.96002620e-01 1.55913746e+00
-5.76429904e-01 -1.28446460e+00 -5.21318495e-01 7.76258349e-01
-1.74849480e-01 -4.39127624e-01 -4.24438357e-01 1.02570426e+00
-4.99355942e-01 1.15811384e+00 -2.20898375e-01 -3.75895321e-01
8.45438778e-01 8.75689864e-01 4.63594466e-01 -9.00728106e-01
-1.06568694e+00 -1.44559667e-01 2.83448577e-01 -8.44200015e-01
-3.31302248e-02 -5.59882104e-01 -1.25966859e+00 -7.66537245e-03
-2.55999565e-01 1.18979774e-01 5.49441695e-01 7.28115916e-01
6.78278625e-01 5.13963819e-01 2.22319916e-01 -3.45968425e-01
-1.11076796e+00 -9.76252496e-01 2.10782848e-02 3.98500621e-01
2.16571108e-01 -4.81202394e-01 -1.43090814e-01 9.22935382e-02] | [10.42761516571045, 8.075784683227539] |
271aad28-ad27-4964-bc81-8349fe02af70 | designing-deep-networks-for-scene-recognition | 2303.07402 | null | https://arxiv.org/abs/2303.07402v1 | https://arxiv.org/pdf/2303.07402v1.pdf | Designing Deep Networks for Scene Recognition | Most deep learning backbones are evaluated on ImageNet. Using scenery images as an example, we conducted extensive experiments to demonstrate the widely accepted principles in network design may result in dramatic performance differences when the data is altered. Exploratory experiments are engaged to explain the underlining cause of the differences. Based on our observation, this paper presents a novel network design methodology: data-oriented network design. In other words, instead of designing universal backbones, the scheming of the networks should treat the characteristics of data as a crucial component. We further proposed a Deep-Narrow Network and Dilated Pooling module, which improved the scene recognition performance using less than half of the computational resources compared to the benchmark network architecture ResNets. The source code is publicly available on https://github.com/ZN-Qiao/Deep-Narrow-Network. | ['Xiaohui Yuan', 'Zhinan Qiao'] | 2023-03-13 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [-4.13199626e-02 1.43848360e-01 -1.14286445e-01 -5.01723468e-01
5.46545744e-01 -4.27129924e-01 2.94687331e-01 -3.95549119e-01
-6.86409771e-01 7.38053501e-01 -1.50270201e-02 -5.40352523e-01
-2.35296607e-01 -9.88222361e-01 -5.49360931e-01 -6.29809976e-01
-5.55944219e-02 -1.57535970e-01 4.93545532e-01 -2.40521759e-01
5.43281972e-01 8.38899672e-01 -1.22306466e+00 2.11871967e-01
4.46359366e-01 9.35104609e-01 3.81011248e-01 4.04104114e-01
-1.87403768e-01 9.92474020e-01 -7.36030161e-01 -3.96153033e-01
6.41925812e-01 -2.37864509e-01 -8.24558675e-01 7.31090491e-04
3.50660115e-01 -4.01221961e-01 -9.30415332e-01 1.01535475e+00
6.20350480e-01 -6.29695728e-02 1.40179098e-01 -1.34222662e+00
-5.17344177e-01 7.94812560e-01 -3.20464343e-01 5.80507100e-01
-1.54659420e-01 3.15963805e-01 8.64520967e-01 -7.57311165e-01
4.32837427e-01 1.07750380e+00 6.84843004e-01 3.82912397e-01
-1.07026100e+00 -1.02525699e+00 3.06101650e-01 3.65842849e-01
-1.51326907e+00 -5.76877475e-01 8.36678267e-01 -1.12157911e-01
6.44538760e-01 8.87855589e-02 6.94733143e-01 1.05874193e+00
3.75550300e-01 3.32875341e-01 8.33415389e-01 -3.26817453e-01
-1.40884919e-02 1.66115835e-01 8.43893439e-02 6.76381886e-01
6.82847381e-01 -6.55899849e-03 -3.83313745e-01 3.22477132e-01
1.15053487e+00 4.80495170e-02 -2.57929921e-01 -2.02491358e-01
-1.11171842e+00 7.06857443e-01 7.93164909e-01 4.21381086e-01
-1.53244197e-01 2.88400769e-01 4.35185134e-01 5.38028657e-01
1.31911337e-01 7.13881552e-01 -4.39837217e-01 8.08259025e-02
-6.43731952e-01 6.33897632e-03 7.87381828e-01 9.00634229e-01
9.24400330e-01 1.77030280e-01 2.08892096e-02 7.81250715e-01
1.95203468e-01 -6.08653985e-02 4.37712997e-01 -9.38577950e-01
2.51152664e-01 7.49985993e-01 -5.64547837e-01 -1.27720547e+00
-5.25506914e-01 -7.60565221e-01 -9.71497476e-01 1.83903262e-01
4.49230552e-01 -4.35011894e-01 -8.23069930e-01 1.58948052e+00
6.72697574e-02 1.79368377e-01 -9.95805413e-02 7.47652471e-01
1.05719221e+00 2.68040240e-01 -3.95242833e-02 4.13380384e-01
1.31300783e+00 -1.14725852e+00 -5.18792450e-01 7.67377168e-02
6.50941908e-01 -8.72372448e-01 1.13613927e+00 1.63867638e-01
-8.54338348e-01 -7.55530894e-01 -1.15647566e+00 9.54346433e-02
-5.61649680e-01 -5.86208738e-02 8.43177259e-01 7.41869211e-01
-1.22729194e+00 5.57894766e-01 -5.20287693e-01 -7.78150797e-01
5.94998538e-01 5.19370794e-01 -3.00919056e-01 5.28398044e-02
-1.10949588e+00 6.14271164e-01 7.33731925e-01 2.21709937e-01
-7.75489450e-01 -7.14711905e-01 -5.05109906e-01 1.18838623e-01
2.34245658e-01 -5.29949069e-01 1.24360108e+00 -1.09898186e+00
-1.56401145e+00 5.44972241e-01 2.50791371e-01 -3.59055191e-01
5.20307779e-01 4.48843278e-02 -4.14140880e-01 2.62952089e-01
-2.32442126e-01 8.29559326e-01 3.82676870e-01 -1.05454719e+00
-4.41383719e-01 9.04861279e-03 3.94631237e-01 9.31318402e-02
-5.79928517e-01 2.58813370e-02 -4.07093525e-01 -7.92635739e-01
1.10654518e-01 -7.96947241e-01 -4.41209763e-01 1.41996920e-01
-6.34099722e-01 -4.83217612e-02 1.07854390e+00 -9.64848921e-02
1.14716709e+00 -2.17991877e+00 -4.70605731e-01 4.92984146e-01
3.76582086e-01 4.10211861e-01 -4.32723284e-01 4.83372480e-01
-3.51251721e-01 3.83243769e-01 6.99668303e-02 7.07005616e-03
-3.39669615e-01 2.01956764e-01 -1.77863359e-01 4.05935168e-01
4.13688496e-02 7.74030209e-01 -6.08058870e-01 -4.47590500e-01
1.73206016e-01 4.19369370e-01 -6.44094408e-01 1.25268787e-01
2.69767553e-01 3.29507738e-01 -4.69555348e-01 5.98632395e-01
7.81695008e-01 -2.69072473e-01 2.22250760e-01 -5.76442301e-01
-2.96567082e-01 2.02119097e-01 -1.09901571e+00 1.54059446e+00
-7.17123821e-02 9.52045083e-01 -7.93184042e-02 -1.02362120e+00
1.11873996e+00 2.31851339e-01 4.15159255e-01 -8.66758108e-01
2.77641326e-01 7.86246136e-02 5.73998630e-01 -4.05663699e-01
4.72731650e-01 2.23213464e-01 2.79034525e-01 2.27585465e-01
6.15829974e-02 1.92329049e-01 2.62955278e-01 9.11157727e-02
1.37054384e+00 -2.66143918e-01 -5.61817433e-04 -5.20625353e-01
4.48827028e-01 1.17690295e-01 7.94166386e-01 1.01503348e+00
-5.17843664e-01 5.10475218e-01 7.00051665e-01 -5.92790484e-01
-1.11175680e+00 -9.15829659e-01 -2.93610215e-01 8.46523702e-01
2.51419783e-01 -4.26989973e-01 -7.75264919e-01 -5.90624809e-01
-2.16428027e-01 3.48178983e-01 -4.22915697e-01 -1.24602005e-01
-6.23190999e-01 -7.53895402e-01 1.11776590e+00 2.72594631e-01
1.28664684e+00 -1.17068493e+00 -5.57925403e-01 4.62852754e-02
3.54747474e-01 -1.18593264e+00 -2.27225482e-01 1.45173296e-01
-9.59361792e-01 -1.30131817e+00 -4.26165909e-01 -9.95263040e-01
9.62533474e-01 6.10451519e-01 8.73320997e-01 5.36265731e-01
-2.09007770e-01 8.38276520e-02 -3.76983613e-01 -2.60124147e-01
-1.36519754e-02 6.43747568e-01 -1.58944383e-01 -1.24297187e-01
3.75745744e-01 -9.62319195e-01 -8.84694099e-01 4.98559117e-01
-1.07544696e+00 3.07698399e-01 8.62272739e-01 5.04359365e-01
1.11145288e-01 2.82526910e-01 5.84799588e-01 -1.03559041e+00
7.42306173e-01 -4.78426069e-01 -4.73755956e-01 1.29931420e-01
-4.71500099e-01 -5.48194572e-02 6.92196608e-01 -3.12990487e-01
-9.73925292e-01 -2.35486314e-01 -6.76108301e-02 -1.65953577e-01
-3.63584876e-01 3.82805347e-01 -2.51548469e-01 -4.30013269e-01
5.12646854e-01 6.25176430e-02 3.93741168e-02 -4.72617805e-01
1.04636468e-01 5.33331513e-01 1.09119102e-01 -6.43986106e-01
7.87060678e-01 5.17219305e-01 4.50823270e-02 -1.02029526e+00
-4.56625283e-01 -9.32826698e-02 -7.50647604e-01 -2.45419919e-01
5.27630627e-01 -8.08520019e-01 -6.46073401e-01 6.21647060e-01
-1.17591238e+00 -6.23407602e-01 -1.75444424e-01 5.35784125e-01
-1.65631682e-01 1.78578004e-01 -6.74335063e-01 -4.97718304e-02
-1.09309763e-01 -1.29167867e+00 2.84005012e-02 3.79581273e-01
-9.64241684e-04 -1.11432731e+00 -1.10582851e-01 1.44409940e-01
7.26916492e-01 1.42429411e-01 7.52732933e-01 -6.86492145e-01
-7.78832018e-01 1.70265347e-01 -7.19963908e-01 6.36756659e-01
1.91970959e-01 3.27108443e-01 -9.82352674e-01 -3.41187507e-01
-1.04407147e-01 -4.02626060e-02 7.86801875e-01 2.70982027e-01
1.65083015e+00 -5.59177175e-02 -9.90387648e-02 1.02751315e+00
1.61323726e+00 3.63162011e-01 1.00209141e+00 6.96561873e-01
6.93202198e-01 5.44691563e-01 -2.74272598e-02 3.36864501e-01
2.65146375e-01 3.63582850e-01 5.33798218e-01 -4.12123471e-01
-3.75770301e-01 -1.42854273e-01 1.81212530e-01 7.78054953e-01
-5.70413470e-02 -1.65188611e-01 -9.22658682e-01 4.51811314e-01
-1.67036080e+00 -8.38796198e-01 2.14610957e-02 1.68926346e+00
5.32037020e-01 2.89686084e-01 -1.53642952e-01 -1.96534336e-01
8.85178685e-01 3.52403671e-01 -5.35983086e-01 -3.47162306e-01
-2.64427722e-01 1.30521923e-01 8.74622524e-01 1.40963286e-01
-7.38894224e-01 1.08279717e+00 6.46601105e+00 6.53798997e-01
-1.40487480e+00 -6.22564293e-02 7.23426878e-01 -1.28349632e-01
-5.81902526e-02 9.25549492e-02 -6.84751749e-01 4.78652596e-01
8.17844927e-01 -3.22244503e-02 1.20252430e-01 7.23397791e-01
4.31227386e-01 7.91164562e-02 -8.64417136e-01 7.93099165e-01
-3.40166867e-01 -1.69619596e+00 1.95876986e-01 2.41456926e-01
5.60228765e-01 4.37096119e-01 -2.63298340e-02 -3.24478000e-02
1.82906136e-01 -8.67993891e-01 6.50084555e-01 2.84975320e-01
5.77316821e-01 -6.13506854e-01 6.66549265e-01 -1.03392720e-01
-1.11833537e+00 -5.52003533e-02 -5.94083905e-01 -4.61708426e-01
-3.11461568e-01 7.00727642e-01 -6.46852076e-01 3.21524203e-01
9.28474247e-01 7.57484436e-01 -8.56245875e-01 1.25158536e+00
-3.03979605e-01 7.43019760e-01 -1.84805885e-01 2.56743319e-02
3.04655284e-01 -7.86824375e-02 3.60564947e-01 1.35975790e+00
1.62486523e-01 -1.56430751e-01 -3.35394591e-02 8.86644781e-01
-4.16621685e-01 -8.40150341e-02 -8.15702736e-01 1.34872198e-01
6.65495336e-01 1.39101768e+00 -1.10178721e+00 -1.27722502e-01
-6.30225420e-01 6.12226903e-01 3.49823564e-01 6.11899018e-01
-7.99127817e-01 -7.45144963e-01 9.72902536e-01 1.12831689e-01
3.68275851e-01 -2.72039592e-01 -3.37993503e-01 -1.09151614e+00
-1.48009449e-01 -7.96451628e-01 1.26026645e-01 -5.56161940e-01
-1.02487957e+00 6.62418127e-01 -9.06858146e-02 -8.50823700e-01
6.73188925e-01 -8.67073238e-01 -9.90590215e-01 7.25061297e-01
-1.69537437e+00 -6.83588684e-01 -6.68426454e-01 8.24457943e-01
3.57799262e-01 -4.40739840e-01 4.05087620e-01 5.49430192e-01
-1.15193295e+00 7.20643938e-01 -1.96894929e-02 6.30138457e-01
6.23687983e-01 -7.36704886e-01 4.43420589e-01 1.04184568e+00
-2.12077826e-01 7.17534721e-01 3.91267031e-01 -2.80363590e-01
-9.60570395e-01 -1.07139039e+00 5.03490388e-01 8.00894424e-02
7.21158028e-01 -4.14847851e-01 -8.40284288e-01 6.91186726e-01
6.48337364e-01 4.45174687e-02 8.42013001e-01 -1.65164974e-02
-2.49364853e-01 -5.74149728e-01 -1.10967386e+00 8.18033159e-01
1.34066343e+00 -1.99097544e-01 -2.04401299e-01 2.35254150e-02
7.71772861e-01 -1.40508354e-01 -8.12637866e-01 3.70373964e-01
4.76815969e-01 -1.26889503e+00 7.82988727e-01 -5.02949536e-01
4.14972097e-01 -2.91124165e-01 -2.12879807e-01 -1.14749503e+00
-4.92636293e-01 -5.24906874e-01 2.48364955e-01 1.08733058e+00
3.57767224e-01 -1.04730618e+00 1.16971409e+00 2.59638309e-01
-2.17242420e-01 -8.71491909e-01 -6.36398017e-01 -6.75120592e-01
-1.15232579e-01 -3.95811528e-01 8.73590529e-01 9.12137926e-01
-4.35231000e-01 1.10622630e-01 8.74373782e-03 1.76521927e-01
5.75364828e-01 -2.78643101e-01 9.28237677e-01 -1.01775694e+00
6.84090555e-02 -5.44519722e-01 -6.46001935e-01 -1.10687613e+00
6.38851300e-02 -7.04633117e-01 -4.13197994e-01 -1.26691806e+00
-6.44968078e-02 -7.78059483e-01 -6.52700901e-01 5.69500566e-01
3.53448421e-01 4.17555392e-01 4.34033960e-01 3.01460803e-01
-4.88934368e-01 4.11862612e-01 1.58462811e+00 8.21435750e-02
-8.13270360e-02 -1.49469957e-01 -1.15570879e+00 7.85441577e-01
1.54481864e+00 -4.17629451e-01 -8.13412786e-01 -8.15074921e-01
1.30275309e-01 -6.42655194e-01 4.16155964e-01 -1.20770884e+00
5.40439546e-01 -2.09946916e-01 3.13801736e-01 -2.95976639e-01
-4.97196987e-02 -1.03440011e+00 8.27160850e-02 9.03201997e-01
-4.46588039e-01 3.51751894e-01 2.89362341e-01 5.35883345e-02
-2.85130829e-01 -2.01580450e-01 7.94758558e-01 -2.72051066e-01
-1.17219579e+00 4.19516176e-01 -3.82988244e-01 -1.63198560e-01
9.88710701e-01 -6.17176414e-01 -6.73829556e-01 -1.03340439e-01
-3.33086222e-01 7.69829452e-02 2.83678621e-01 3.54559064e-01
6.32571638e-01 -1.22562444e+00 -4.98033434e-01 3.84802222e-01
-1.71114907e-01 7.05134571e-02 2.11069643e-01 7.15795755e-01
-1.13277853e+00 5.19489825e-01 -6.56681657e-01 -3.53319317e-01
-1.10009158e+00 5.03288768e-02 7.76954830e-01 3.55519261e-03
-5.26524007e-01 8.67976785e-01 3.58388960e-01 -5.71987629e-01
4.49374080e-01 -3.48073035e-01 -9.43051949e-02 -1.44617647e-01
3.38616163e-01 4.35945213e-01 3.15032713e-02 -2.34917417e-01
-3.25092047e-01 2.52805799e-01 -3.70168805e-01 3.37093234e-01
1.41576958e+00 -1.09113030e-01 -3.90923798e-01 4.40669842e-02
1.23526180e+00 -4.07667339e-01 -1.38820350e+00 -3.80786359e-01
-1.16998166e-01 -4.89074439e-01 2.12531045e-01 -4.65302140e-01
-1.71191227e+00 6.46005034e-01 5.50764680e-01 2.67321706e-01
1.35107577e+00 -2.49061540e-01 5.96841216e-01 6.87215865e-01
1.91506714e-01 -1.05140245e+00 2.46525168e-01 6.75975382e-01
7.08399296e-01 -1.03610492e+00 3.81429344e-02 -3.79741758e-01
-1.95240438e-01 1.32822645e+00 1.12953937e+00 -3.43358636e-01
1.05377340e+00 2.96968371e-01 2.10407659e-01 -4.27674949e-01
-6.01551235e-01 2.86072940e-02 -2.13190347e-01 5.04493773e-01
5.33493459e-01 -1.76924035e-01 -1.07306950e-01 9.62998345e-02
-3.71470481e-01 -7.38037676e-02 7.23552763e-01 8.73381615e-01
-3.85219961e-01 -1.26115716e+00 -1.96136624e-01 4.14425343e-01
-4.41315085e-01 -2.38207221e-01 -2.69108921e-01 1.09088683e+00
3.19427013e-01 8.17013383e-01 2.90586501e-01 -5.15749454e-01
5.01581848e-01 -3.97698164e-01 2.33474776e-01 -6.43498302e-01
-6.74217463e-01 -4.70686316e-01 1.68589745e-02 -7.03333020e-01
-5.97240508e-01 -1.20692544e-01 -1.03717232e+00 -1.08157742e+00
-2.71024741e-03 3.55059956e-03 5.95078528e-01 6.04932189e-01
5.82315147e-01 8.30659628e-01 6.04208231e-01 -6.88183367e-01
-1.88685775e-01 -8.53838265e-01 -5.83674848e-01 2.52015859e-01
7.90705755e-02 -4.17491376e-01 -3.06174219e-01 -1.75289735e-01] | [9.081787109375, 2.2958738803863525] |
8265b408-1425-4735-a9cd-fdb0769da0e6 | designing-discontinuities | 2305.08559 | null | https://arxiv.org/abs/2305.08559v1 | https://arxiv.org/pdf/2305.08559v1.pdf | Designing Discontinuities | Discontinuities can be fairly arbitrary but also cause a significant impact on outcomes in social systems. Indeed, their arbitrariness is why they have been used to infer causal relationships among variables in numerous settings. Regression discontinuity from econometrics assumes the existence of a discontinuous variable that splits the population into distinct partitions to estimate the causal effects of a given phenomenon. Here we consider the design of partitions for a given discontinuous variable to optimize a certain effect previously studied using regression discontinuity. To do so, we propose a quantization-theoretic approach to optimize the effect of interest, first learning the causal effect size of a given discontinuous variable and then applying dynamic programming for optimal quantization design of discontinuities that balance the gain and loss in the effect size. We also develop a computationally-efficient reinforcement learning algorithm for the dynamic programming formulation of optimal quantization. We demonstrate our approach by designing optimal time zone borders for counterfactuals of social capital, social mobility, and health. This is based on regression discontinuity analyses we perform on novel data, which may be of independent empirical interest in showing a causal relationship between sunset time and social capital. | ['Lav R. Varshney', 'Ting-Yi Wu', 'Suyoung Park', 'Ibtihal Ferwana'] | 2023-05-15 | null | null | null | null | ['econometrics'] | ['miscellaneous'] | [ 2.73637861e-01 5.94454110e-01 -8.53146493e-01 -2.40725100e-01
-6.05051756e-01 -1.44321369e-02 4.73901063e-01 5.03000498e-01
-4.43323165e-01 1.14916611e+00 5.02875566e-01 -6.66726947e-01
-6.69128895e-01 -1.03348887e+00 -1.02546358e+00 -5.95141888e-01
-4.93391484e-01 4.77763712e-01 -2.42817104e-01 -1.38178207e-02
3.60612959e-01 9.97450575e-02 -1.08759201e+00 -2.35139236e-01
1.15729868e+00 4.24167365e-01 6.33585304e-02 4.47888553e-01
1.76788539e-01 6.59348667e-01 -5.43269694e-01 -3.01561654e-01
2.37834796e-01 -7.38739729e-01 -6.32039726e-01 6.15265928e-02
-5.60170859e-02 -2.73240566e-01 4.69129831e-02 6.74831569e-01
3.88019145e-01 1.41322225e-01 1.03118610e+00 -1.44594896e+00
-6.72650099e-01 9.24433053e-01 -7.79801250e-01 1.62122235e-01
2.26611912e-01 1.37737885e-01 1.33634865e+00 8.80595371e-02
5.69934368e-01 1.51793683e+00 6.58643544e-01 1.41713560e-01
-1.86521637e+00 -6.24237061e-01 2.28836060e-01 -7.78866187e-02
-1.02376807e+00 -3.20636123e-01 6.45926952e-01 -6.57772660e-01
5.65542400e-01 2.15349913e-01 9.61458027e-01 8.20513129e-01
5.23323238e-01 4.60046053e-01 1.05329299e+00 -6.23624206e-01
5.54960430e-01 -1.88505486e-01 1.77593790e-02 5.51990390e-01
4.48807329e-01 3.94559175e-01 -2.06873775e-01 -3.86699259e-01
7.97487199e-01 3.22822988e-01 2.94846967e-02 -1.89834356e-01
-9.29055631e-01 1.15827215e+00 4.48092818e-01 2.35732142e-02
-6.73431516e-01 6.07848108e-01 1.74425274e-01 4.34236228e-01
4.54168230e-01 4.48485881e-01 -3.38384241e-01 1.37364296e-02
-7.47641265e-01 5.26382625e-01 6.12135887e-01 3.61253560e-01
6.35240316e-01 -3.91236752e-01 -4.81922776e-01 5.30190468e-01
3.12179267e-01 6.65759027e-01 -6.20476902e-02 -1.36880374e+00
5.22484720e-01 3.86803716e-01 5.00095248e-01 -8.73037457e-01
-4.76126015e-01 1.43598169e-01 -6.92793965e-01 2.14979321e-01
7.54969180e-01 -6.42374814e-01 -5.41236103e-01 2.11961913e+00
3.80423576e-01 2.28702620e-01 -2.46412367e-01 4.94584382e-01
-3.00160497e-01 5.05694747e-01 3.02421212e-01 -6.40738308e-01
1.15117157e+00 -1.27057880e-01 -8.55555296e-01 1.95157170e-01
6.70192897e-01 -1.27655670e-01 1.00358748e+00 -1.35471476e-02
-1.13881516e+00 -1.91322416e-01 -6.07199252e-01 2.56454915e-01
-2.18077544e-02 -5.11547029e-01 5.91732562e-01 8.13665271e-01
-8.76042664e-01 8.29658091e-01 -8.80223930e-01 -1.40223607e-01
4.89844888e-01 5.00180602e-01 4.80692178e-01 2.66517401e-01
-1.37209141e+00 5.71661711e-01 -2.26172537e-01 -3.20008516e-01
-4.42991167e-01 -9.95581329e-01 -5.66886008e-01 4.27827597e-01
5.12393177e-01 -9.94133592e-01 1.15121007e+00 -1.20620620e+00
-1.29967082e+00 5.19201458e-01 -1.13116197e-01 -8.28695655e-01
6.96251750e-01 2.07405329e-01 -1.56736612e-01 -2.67928429e-02
6.73802614e-01 4.96826857e-01 7.69930780e-01 -8.62922192e-01
-7.25706041e-01 -4.71551031e-01 1.67144760e-01 2.00893193e-01
3.10765114e-02 -3.25701565e-01 2.26134405e-01 -6.27675474e-01
-3.21265966e-01 -8.68270397e-01 -6.23391867e-01 -2.33848616e-01
-3.87336582e-01 -3.24981898e-01 -1.04135685e-02 -5.95097601e-01
1.37195027e+00 -1.98319256e+00 3.47979292e-02 3.17323387e-01
1.15386024e-01 -6.92128181e-01 1.91025436e-01 4.26148236e-01
2.05285490e-01 5.20336509e-01 -5.13218701e-01 -1.41949832e-01
7.73535818e-02 7.42704719e-02 -1.83657244e-01 6.89939201e-01
2.16798916e-01 7.56708026e-01 -9.28709686e-01 -5.32910943e-01
1.50690317e-01 5.00336327e-02 -1.06660473e+00 -7.98552111e-02
-1.90344617e-01 2.78052628e-01 -5.73235989e-01 1.90350890e-01
4.81754988e-01 -1.25556231e-01 5.62330842e-01 6.15948617e-01
-3.30427349e-01 2.46865228e-01 -1.06897879e+00 1.04317653e+00
-2.37793252e-01 5.46132326e-01 -1.95769116e-01 -1.29520488e+00
4.19525087e-01 2.13991836e-01 7.86456704e-01 -8.55811119e-01
2.90988721e-02 -4.33733724e-02 9.62004736e-02 -7.75458336e-01
3.76296282e-01 -8.37871492e-01 -3.29201102e-01 5.57827234e-01
-6.34832859e-01 7.84262195e-02 6.39862120e-02 -4.38050181e-02
1.12958455e+00 -2.88054168e-01 4.87020224e-01 -5.05355775e-01
-3.39786299e-02 -8.31957087e-02 6.81948900e-01 9.08999920e-01
-2.61464924e-01 1.61700338e-01 1.27955115e+00 -9.44877714e-02
-1.10623455e+00 -1.18644726e+00 -4.23904270e-01 7.74747372e-01
6.97658509e-02 4.09312993e-01 -7.37117589e-01 -5.04786313e-01
5.56916714e-01 9.23915923e-01 -1.07511759e+00 -1.46541744e-01
-4.38768476e-01 -8.74227762e-01 3.15880090e-01 1.84407517e-01
4.00721878e-01 -9.01263833e-01 -8.64859581e-01 1.59039080e-01
-1.62593991e-01 -4.29014862e-01 -4.52304840e-01 2.37718374e-02
-9.72139657e-01 -1.23105061e+00 -6.78943694e-01 -3.50776702e-01
6.14020586e-01 -1.26624731e-02 9.31953728e-01 7.56119564e-02
1.99371949e-01 3.93749386e-01 1.97785124e-02 -4.80627626e-01
-3.64897907e-01 1.39431596e-01 2.97850482e-02 3.26560661e-02
1.41413480e-01 -4.14782286e-01 -7.24226654e-01 -4.51637898e-03
-8.00118744e-01 -2.29310051e-01 2.63144523e-01 8.33703637e-01
3.39250505e-01 1.74709424e-01 1.11825967e+00 -1.15793335e+00
7.79183805e-01 -1.11386788e+00 -8.14530551e-01 2.39767596e-01
-8.07392955e-01 3.38931143e-01 3.83784950e-01 -4.26707655e-01
-9.34617877e-01 -3.09033632e-01 2.24228904e-01 9.64208394e-02
7.12520406e-02 5.89167356e-01 -1.36386110e-02 6.59802675e-01
6.20021880e-01 -4.12773967e-01 6.53426424e-02 -1.54550090e-01
3.25078517e-01 6.89979970e-01 -6.89281076e-02 -7.23518491e-01
1.47782207e-01 6.57403231e-01 2.15311989e-01 -8.08709502e-01
-3.63701254e-01 -1.25466675e-01 -1.72429860e-01 -9.05996934e-02
9.43691432e-01 -8.71086597e-01 -1.22656584e+00 -2.18310710e-02
-7.32254982e-01 -8.31698537e-01 -6.69935584e-01 5.19536316e-01
-9.87858415e-01 -2.12752730e-01 -1.87982053e-01 -1.03163242e+00
4.35441077e-01 -8.93015563e-01 8.02028835e-01 2.44147815e-02
-4.14466858e-01 -1.37922561e+00 4.37097162e-01 1.54210076e-01
-9.85203404e-03 6.83912516e-01 1.16417027e+00 5.89814372e-02
-6.67895555e-01 2.71898508e-01 5.64248450e-02 -3.48387629e-01
2.90140808e-01 1.07790880e-01 -4.30996418e-01 -1.67388380e-01
-2.78541178e-01 6.52740151e-02 5.83496153e-01 1.33302581e+00
8.52751434e-01 -6.82566583e-01 -5.83117008e-01 1.97954208e-01
1.40310609e+00 4.25772011e-01 4.60975498e-01 3.81497055e-01
3.43392402e-01 8.71710598e-01 5.59302866e-01 8.00317407e-01
6.60323620e-01 7.08231151e-01 1.45766139e-01 -2.42864311e-01
3.70946258e-01 -4.25362825e-01 1.62970975e-01 -8.22244063e-02
8.79736058e-03 -5.47412075e-02 -7.86014378e-01 8.73447359e-01
-1.81267583e+00 -1.18968225e+00 -6.12563379e-02 2.47039819e+00
8.89307916e-01 2.00106949e-01 7.79645205e-01 8.29880461e-02
8.20433140e-01 -9.61884260e-02 -6.32396340e-01 -6.67190552e-01
1.81002498e-01 -1.13266781e-01 9.66117024e-01 9.30885851e-01
-5.71853995e-01 5.17906129e-01 6.97414351e+00 4.15612310e-01
-9.53024983e-01 -8.69860426e-02 1.07907021e+00 -1.63586661e-01
-7.25041866e-01 1.31816521e-01 -4.88425791e-01 6.44818723e-01
1.29935694e+00 -4.37473089e-01 5.08844078e-01 2.07008645e-01
1.14825416e+00 -4.65459585e-01 -1.23825014e+00 2.58584321e-01
-8.14709306e-01 -1.30466032e+00 -2.47150645e-01 4.53022778e-01
1.01489401e+00 -3.95262152e-01 2.02960670e-01 2.85938323e-01
7.41098046e-01 -9.21614230e-01 8.21890831e-01 5.81251025e-01
6.98632061e-01 -1.02469265e+00 1.64944947e-01 3.26490909e-01
-5.99692523e-01 -3.79381776e-01 -1.47089630e-01 -6.19971931e-01
1.86758697e-01 7.67053425e-01 -9.10763562e-01 -3.28040905e-02
3.64054084e-01 5.31019986e-01 4.35949937e-02 7.88570225e-01
-1.16121419e-01 8.07498336e-01 -2.61644751e-01 -9.31138769e-02
1.80580780e-01 -3.79822999e-01 2.74863422e-01 8.14773679e-01
2.18313992e-01 1.75015643e-01 -9.71710384e-02 1.11888289e+00
-2.67410636e-01 4.17615548e-02 -8.89024794e-01 6.94282912e-03
6.01469219e-01 2.60996222e-01 -9.59935784e-01 -1.88106030e-01
-4.34342295e-01 5.59099376e-01 2.48158947e-01 6.44150913e-01
-9.63381529e-01 -2.30825916e-01 8.04110944e-01 5.03053904e-01
1.36258110e-01 1.82438940e-01 -7.00806260e-01 -8.34274173e-01
-2.30493933e-01 -7.03958452e-01 5.27673781e-01 -1.67959288e-01
-9.39134896e-01 -5.55795670e-01 3.04551005e-01 -7.39907563e-01
-3.64754558e-01 1.49639651e-01 -4.76770639e-01 8.26601982e-01
-1.44377792e+00 -3.37665915e-01 5.39631426e-01 3.46113116e-01
3.89333278e-01 4.15525466e-01 2.10290417e-01 5.02501093e-02
-7.25586474e-01 3.30755234e-01 4.59776998e-01 -1.73305362e-01
4.01658416e-01 -1.50995433e+00 3.56154323e-01 5.24620056e-01
-3.15573752e-01 4.97386873e-01 7.19030738e-01 -1.02148962e+00
-8.89562786e-01 -7.67341077e-01 9.52333808e-01 -2.66614348e-01
6.77324235e-01 -1.62165284e-01 -4.98125732e-01 8.63489330e-01
1.25386827e-02 -6.43340468e-01 4.58373547e-01 4.85394090e-01
2.59789616e-01 -1.59738109e-01 -1.46035624e+00 1.01864386e+00
8.34054828e-01 -3.21680486e-01 -4.13583457e-01 1.12658329e-01
1.01023281e+00 4.30585556e-02 -9.00755703e-01 -4.69644777e-02
5.19561946e-01 -9.40391839e-01 9.61446881e-01 -6.71826661e-01
7.61760235e-01 1.96425974e-01 2.62993593e-02 -1.38561070e+00
-2.52053112e-01 -6.52683735e-01 2.55029052e-01 1.06372297e+00
5.21932006e-01 -7.50199735e-01 6.73478425e-01 8.01178038e-01
4.14532065e-01 -7.23657787e-01 -1.02621877e+00 -6.77554607e-01
5.07105649e-01 -2.27556273e-01 8.27913046e-01 1.23648953e+00
2.78676361e-01 2.96621501e-01 -2.68309265e-01 1.73682824e-01
8.37406516e-01 2.02892452e-01 6.02743745e-01 -9.38695967e-01
-4.48470026e-01 -5.35614550e-01 2.85905730e-02 -7.38897204e-01
1.02842018e-01 -5.47828853e-01 -9.17949006e-02 -1.37122881e+00
3.42603326e-01 -8.72282565e-01 -1.24765031e-01 1.34774834e-01
-4.57894564e-01 -4.50696498e-01 1.11160010e-01 3.21879014e-02
-1.66447520e-01 5.38082659e-01 1.24328732e+00 -1.48484632e-01
-9.05742943e-01 2.69688904e-01 -8.85766685e-01 3.62168968e-01
7.59687304e-01 -7.66650736e-01 -5.83174706e-01 -2.49831676e-02
3.79242212e-01 8.56623471e-01 4.36698914e-01 -4.85349834e-01
-1.83893397e-01 -8.94295454e-01 -4.70428132e-02 -2.51503428e-03
-2.43378237e-01 -8.21911037e-01 2.30926961e-01 1.01051986e+00
-8.67084563e-01 -1.53882861e-01 -4.82134633e-02 8.00875187e-01
1.86626866e-01 -4.01419997e-02 6.86484516e-01 6.07466251e-02
6.25796244e-02 2.04317793e-02 -6.51669443e-01 3.81523401e-01
1.08423173e+00 1.90736502e-02 6.63918331e-02 -5.41853011e-01
-7.26261318e-01 5.22230804e-01 4.98823196e-01 1.19411252e-01
2.83629149e-01 -1.20318794e+00 -7.43652821e-01 -1.98710188e-01
-3.56329978e-01 -4.11311001e-01 1.71260864e-01 7.79391885e-01
-2.97643960e-01 3.05612355e-01 -1.62043691e-01 -2.80314803e-01
-8.20801198e-01 8.00113201e-01 4.39090639e-01 -4.56543773e-01
-4.68073159e-01 1.71594933e-01 3.36460143e-01 -1.36690596e-02
-3.28748412e-02 -6.33118093e-01 6.08560927e-02 3.15870106e-01
1.98271334e-01 8.76690269e-01 -5.71807802e-01 -8.20592344e-02
-6.85542300e-02 3.21195215e-01 3.90036881e-01 -5.59253156e-01
1.34851086e+00 -5.75518608e-01 2.05158278e-01 7.55807936e-01
1.09124637e+00 8.61764178e-02 -1.64655209e+00 1.70262530e-01
2.08460480e-01 -2.86814243e-01 -9.65371057e-02 -4.78838563e-01
-6.18131697e-01 5.18470585e-01 5.29214084e-01 8.28985929e-01
1.01746690e+00 -1.29165083e-01 4.43635494e-01 -1.32059902e-02
1.37350753e-01 -1.12165725e+00 -1.48360237e-01 -2.29444891e-01
6.10297561e-01 -1.23586869e+00 -2.94858031e-02 -7.71673694e-02
-3.72536123e-01 4.19590682e-01 -8.01073909e-02 -3.14759612e-01
7.69426227e-01 -1.36457250e-01 -3.18473190e-01 -1.54798955e-01
-7.10087180e-01 -2.17453822e-01 -2.38076195e-01 4.87381399e-01
3.44812423e-01 6.62940025e-01 -9.22282219e-01 3.28477435e-02
-9.25408080e-02 9.22967419e-02 7.57078826e-01 6.93650961e-01
-2.55160421e-01 -9.78151202e-01 -6.29228354e-01 5.61371982e-01
-6.96621656e-01 -2.55868658e-02 -1.25494227e-01 1.02295518e+00
2.22204417e-01 8.67295146e-01 4.32470620e-01 1.39359266e-01
3.77225429e-01 -1.45003885e-01 3.65566671e-01 -3.36969614e-01
-3.42152566e-01 2.02367245e-03 6.98623061e-02 -2.69809753e-01
-5.92602789e-01 -1.20273602e+00 -1.39904344e+00 -8.07134569e-01
5.56217916e-02 2.52483010e-01 2.76637733e-01 1.08229101e+00
1.41137272e-01 6.62801862e-01 1.00484216e+00 -4.87660080e-01
-4.37359005e-01 -5.64301908e-01 -8.68457615e-01 4.44478184e-01
7.68320382e-01 -6.50470555e-01 -3.90743494e-01 -6.28341138e-02] | [8.049784660339355, 5.301637649536133] |
98bd9bd0-f738-4549-a572-a9b7e4162c9f | a-horizon-detection-algorithm-for-maritime | 2110.13694 | null | https://arxiv.org/abs/2110.13694v3 | https://arxiv.org/pdf/2110.13694v3.pdf | A vectorized sea horizon edge filter for maritime video processing tasks | The horizon line is a fundamental semantic feature in several maritime video processing tasks, such as digital video stabilization, camera calibration, target tracking, and target distance estimation. Visible range Electro-Optical (EO) sensors capture richer information in the daytime, which often comes with challenging clutter. The best methods rely on tailored filters to keep, ideally, only horizon edge pixels. These methods work well but often fail in the case of edge-degraded horizons. Our first aim is to solve this problem while taking the real-time constraint into account; we propose a tailored edge filter that relies on growing line segments with a low edge threshold and filters them based on their slope, length, and relative position. Next, we build the filtered edge map by computing Cartesian coordinates of pixels across line segments that survived the filter. We infer the horizon from the filtered edge map using line fitting techniques and simple temporal information. We consider the real-time constraint by vectorizing the computations and proposing a better way to leverage image downsizing. Extensive experiments on 26,125 visible range frames show that the proposed method achieves significant robustness while satisfying the real-time constraint. | ['Astito Abdelali', 'Boulaala Mohammed', 'Yassir Zardoua'] | 2021-10-26 | null | null | null | null | ['video-stabilization'] | ['computer-vision'] | [ 4.16157603e-01 -4.52693939e-01 5.29463515e-02 -1.66372329e-01
-3.29548597e-01 -8.55332732e-01 4.64524835e-01 6.86196983e-02
-6.57084346e-01 3.43415141e-01 -1.12311497e-01 -2.40535840e-01
-3.74040037e-01 -6.66859210e-01 -6.15267277e-01 -7.19259024e-01
-8.61308947e-02 -2.84856886e-01 6.40842974e-01 -2.47113943e-01
4.51855779e-01 6.96969688e-01 -1.18664122e+00 -1.91730782e-01
6.66298389e-01 1.39769077e+00 -8.91792700e-02 6.75743520e-01
4.18091863e-01 4.58141923e-01 -2.79816657e-01 -3.42058539e-01
8.59012842e-01 -9.34091508e-02 8.80159717e-03 5.12803018e-01
5.19380927e-01 -4.34493512e-01 -6.11329615e-01 1.31487620e+00
2.69403428e-01 3.87553304e-01 1.63358465e-01 -1.19852591e+00
-6.45668060e-02 1.66537724e-02 -9.09282327e-01 4.94185746e-01
2.46277124e-01 9.42006856e-02 6.13438725e-01 -8.48909080e-01
7.01413870e-01 6.01547062e-01 1.12603807e+00 5.27861752e-02
-6.61842167e-01 -2.99925387e-01 2.65732497e-01 3.81043017e-01
-1.46057045e+00 -5.09961069e-01 9.07086909e-01 -4.60719049e-01
4.73401099e-01 4.02030498e-01 6.19121730e-01 4.98109311e-01
3.06528211e-01 2.21733615e-01 8.25131178e-01 -2.83883482e-01
9.59314555e-02 -1.56737819e-01 8.38569999e-02 4.39124972e-01
2.35191286e-01 2.33195186e-01 -5.58660388e-01 -7.32715577e-02
7.31617093e-01 1.91292271e-01 -8.70495200e-01 -5.32589316e-01
-1.22054720e+00 6.23373747e-01 2.89382577e-01 1.23752272e-02
-3.58520299e-01 -1.57684430e-01 7.10953847e-02 3.76367599e-01
4.98947412e-01 2.18386218e-01 -2.68756121e-01 8.67316574e-02
-1.06972957e+00 -6.52367771e-02 6.09586060e-01 9.96372581e-01
7.46707797e-01 8.00111145e-02 -8.45590755e-02 3.16435993e-01
1.12059824e-01 5.04575014e-01 -2.97283735e-02 -8.76545668e-01
3.96224231e-01 2.67125487e-01 4.25849646e-01 -1.18262994e+00
-5.20525157e-01 -5.19023776e-01 -4.99069303e-01 5.26541352e-01
5.76349020e-01 -3.95705134e-01 -7.36598372e-01 1.13201833e+00
6.05944693e-01 5.21640599e-01 3.26941498e-02 1.11884129e+00
2.86430031e-01 5.27734518e-01 -4.77166742e-01 -5.97452283e-01
1.23090184e+00 -7.92549193e-01 -6.71270192e-01 -4.50931191e-01
4.05534655e-01 -8.16520035e-01 3.22268456e-01 4.94564325e-01
-7.33218431e-01 -2.52704144e-01 -1.11961865e+00 2.91310281e-01
-1.68293670e-01 2.71112114e-01 5.06475687e-01 6.47779524e-01
-7.80138075e-01 5.88834405e-01 -6.53863847e-01 -1.20594777e-01
-2.26265550e-01 4.07520272e-02 -3.76820683e-01 -3.93959098e-02
-9.61197972e-01 6.44667625e-01 2.54254371e-01 4.95170504e-01
-3.08730751e-01 -6.35767877e-01 -1.07261121e+00 -7.75533617e-02
9.72930253e-01 -5.56916952e-01 8.65239859e-01 -8.80167425e-01
-1.49175382e+00 4.94792014e-01 -1.29016787e-01 -7.36347318e-01
8.91558588e-01 -4.02994692e-01 -5.51917136e-01 3.60426575e-01
-2.63212740e-01 1.18893273e-01 1.24743462e+00 -9.85674679e-01
-9.73841846e-01 -4.20148045e-01 1.66393548e-01 4.29958582e-01
-1.94837168e-01 -1.27491727e-01 -7.55884290e-01 -5.82312763e-01
5.50163984e-01 -1.06309450e+00 -2.44230643e-01 3.98841649e-01
-2.13105068e-01 4.71876770e-01 8.63284528e-01 -7.26897120e-01
1.31866443e+00 -2.31649852e+00 -2.46073619e-01 2.52778471e-01
1.45039409e-01 7.94267803e-02 3.00009698e-01 1.73008114e-01
1.83849528e-01 -5.69734752e-01 -2.67329395e-01 -1.81657761e-01
-3.37602526e-01 -2.93019861e-01 -3.81335199e-01 9.97947633e-01
-1.18055753e-01 3.67266268e-01 -8.80484879e-01 -2.07556173e-01
5.62225878e-01 4.65614915e-01 -4.46581364e-01 -1.74880072e-01
-3.29651637e-03 4.84161884e-01 -2.46980175e-01 6.12605512e-01
9.19505596e-01 1.81904629e-01 -2.75239050e-01 -6.19405627e-01
-7.09352911e-01 -3.66192997e-01 -1.53057456e+00 1.47449636e+00
-7.95102045e-02 1.07799172e+00 2.29433388e-01 -4.80890334e-01
1.02838767e+00 -8.71088803e-02 8.03744137e-01 -5.23237467e-01
2.72144109e-01 6.01649843e-02 -2.37445340e-01 -4.96190041e-01
9.00601566e-01 4.70871031e-02 7.49080032e-02 -1.55496821e-01
-7.02962041e-01 -7.65912086e-02 5.34422845e-02 -1.70352161e-01
1.01927078e+00 1.56227440e-01 2.61805147e-01 -7.89755024e-03
5.12863815e-01 1.55599028e-01 6.94156408e-01 6.42469049e-01
-3.13133776e-01 7.72183061e-01 -1.12987403e-02 -4.57756102e-01
-8.36658418e-01 -6.39097512e-01 -8.56188834e-02 5.96806586e-01
7.04399109e-01 -2.96733648e-01 -5.55093288e-01 -2.66556710e-01
-1.87006906e-01 4.40488160e-01 -4.22448635e-01 -1.59545198e-01
-5.31832635e-01 -6.28570020e-01 1.87307850e-01 2.77230442e-01
6.74662113e-01 -1.38556942e-01 -1.11521971e+00 3.70782614e-01
-9.25466791e-02 -1.55248356e+00 -6.88287139e-01 -5.70643842e-02
-6.55170679e-01 -1.01987743e+00 -7.41210163e-01 -3.43130678e-01
6.45451367e-01 9.18743670e-01 4.69276160e-01 -1.55021161e-01
-1.94460332e-01 3.64946127e-01 -2.58949310e-01 -1.65454462e-01
2.44280234e-01 -5.06382048e-01 1.31976847e-02 5.37990570e-01
4.93690148e-02 -1.30862370e-01 -7.42043197e-01 4.82115656e-01
-7.69862056e-01 1.08964667e-01 2.55074799e-01 5.91563821e-01
5.66960335e-01 3.42249781e-01 -1.64499581e-01 -5.43875933e-01
1.34746596e-01 -1.10934511e-01 -1.26465929e+00 1.97543427e-01
-4.17890102e-01 -2.85930544e-01 4.81813997e-01 -4.31617141e-01
-9.62781787e-01 3.55770826e-01 2.30511487e-01 -8.37877214e-01
2.67851144e-01 4.92218882e-01 4.54893187e-02 -5.70531905e-01
5.87704599e-01 1.19231194e-01 -2.04957858e-01 -9.17885900e-02
1.06163770e-01 4.91294801e-01 9.39152122e-01 2.13482141e-01
1.03263164e+00 1.07800210e+00 2.25969285e-01 -1.14645123e+00
-8.53855610e-01 -8.11970770e-01 -5.32984436e-01 -5.61938226e-01
6.51147246e-01 -1.04623365e+00 -6.04894340e-01 5.21326840e-01
-9.42253351e-01 -2.47789741e-01 -2.61521041e-02 7.75941551e-01
-3.30615252e-01 7.70257533e-01 -3.89469951e-01 -9.01998818e-01
-1.66664019e-01 -7.97037363e-01 9.52812850e-01 4.62660998e-01
3.86859626e-01 -9.66237247e-01 -8.21026415e-02 -5.05988207e-03
3.21365237e-01 5.49608648e-01 1.28576672e-02 3.41321193e-02
-7.30815411e-01 -3.87973398e-01 -2.25553542e-01 -2.82941144e-02
-7.88657963e-02 1.30793110e-01 -7.90897429e-01 -3.18991274e-01
2.06149131e-01 6.50611103e-01 1.00424445e+00 6.92899227e-01
6.56143785e-01 -4.52783005e-03 -3.68009806e-01 1.23247457e+00
1.58879244e+00 1.55179694e-01 5.81597447e-01 6.71529531e-01
6.45745456e-01 5.72049737e-01 1.06550789e+00 7.82247722e-01
1.25459030e-01 8.29702556e-01 5.32419920e-01 -3.98471719e-03
1.61284775e-01 1.96100771e-01 3.89157265e-01 3.17602187e-01
-4.92952615e-02 -2.92911530e-01 -6.98388755e-01 3.06924105e-01
-1.92128277e+00 -9.64704335e-01 -3.64508510e-01 2.54645634e+00
2.30663195e-01 4.37118173e-01 -1.17432855e-01 1.73325121e-01
8.71650934e-01 1.11590318e-01 -5.52254081e-01 1.08345523e-01
-2.34146789e-01 -5.31610966e-01 1.47665215e+00 6.81417525e-01
-1.30200076e+00 6.68984652e-01 5.44753122e+00 5.12328327e-01
-1.49101698e+00 -2.70287722e-01 1.21064298e-01 8.31369385e-02
-1.53652981e-01 2.69533038e-01 -1.11456466e+00 3.76920104e-01
5.66037476e-01 -3.11687946e-01 4.32524502e-01 5.75021148e-01
4.27237630e-01 -3.63807082e-01 -7.01399326e-01 1.20453334e+00
3.67737174e-01 -1.21408212e+00 -5.12867868e-01 -6.83320835e-02
6.19680345e-01 -4.84361686e-02 -8.84755403e-02 -1.74107060e-01
-3.16282302e-01 -3.04839104e-01 8.97211313e-01 7.42682219e-01
5.21577239e-01 -5.31655312e-01 4.32682365e-01 3.18925261e-01
-1.38339114e+00 -1.91302121e-01 -4.23719078e-01 -1.48094207e-01
4.28254843e-01 8.70333970e-01 -6.49629951e-01 7.48807251e-01
5.98814666e-01 8.45594704e-01 -4.54013914e-01 1.72160494e+00
1.85841825e-02 6.77499622e-02 -6.52206182e-01 5.00015020e-01
2.01075599e-01 -5.98951519e-01 1.00106645e+00 9.90328610e-01
6.32633269e-01 2.12682575e-01 2.25168616e-01 3.31089854e-01
3.91061336e-01 -2.06014633e-01 -7.16219008e-01 2.60568380e-01
4.38656241e-01 1.32336175e+00 -1.02754891e+00 -9.78954732e-02
-7.01586783e-01 9.54840183e-01 -2.46291190e-01 3.49220783e-01
-1.08505785e+00 -4.06514406e-01 5.57509482e-01 1.76254407e-01
3.46258372e-01 -6.36687279e-01 -1.96955591e-01 -1.18622017e+00
2.35288709e-01 -2.94189125e-01 3.78384680e-01 -7.69015849e-01
-7.09383667e-01 5.78411698e-01 -7.15336353e-02 -1.91547465e+00
-5.94870672e-02 -6.40368164e-01 -5.22871792e-01 4.09451246e-01
-1.87718737e+00 -9.11950350e-01 -7.85812259e-01 6.76470578e-01
5.60486495e-01 2.97468513e-01 1.24484032e-01 3.94739449e-01
-4.26930755e-01 2.98288882e-01 2.33210117e-01 1.42574310e-04
7.20077395e-01 -7.27184355e-01 2.28668839e-01 1.37673283e+00
-7.01933131e-02 2.20843434e-01 1.02559912e+00 -6.83621466e-01
-1.64508164e+00 -9.39558327e-01 6.09960616e-01 -1.81716248e-01
7.71303236e-01 -3.35500985e-02 -7.66913831e-01 6.85822606e-01
-1.77840635e-01 3.95086795e-01 1.80614457e-01 -6.58111095e-01
2.69828085e-02 -6.73624873e-02 -8.50041866e-01 5.94928801e-01
1.03430343e+00 -2.86519110e-01 -3.07749271e-01 2.80196697e-01
5.48238456e-01 -8.91621292e-01 -6.76036656e-01 6.17604315e-01
4.99565303e-01 -1.05535507e+00 1.09944487e+00 8.47456232e-02
-1.84331179e-01 -8.67683232e-01 1.33634925e-01 -1.09989977e+00
-1.35649815e-01 -9.69365716e-01 1.47790477e-01 9.02222157e-01
1.38211310e-01 -6.10298514e-01 7.88680851e-01 6.75184548e-01
-4.41834062e-01 -2.85336554e-01 -9.90179777e-01 -9.24030960e-01
-8.54443550e-01 -5.72511494e-01 1.66540518e-01 9.89287972e-01
-2.15262666e-01 -1.53859660e-01 -6.36836588e-01 8.17028522e-01
8.15526247e-01 3.41586083e-01 8.34394753e-01 -1.19448042e+00
-3.27300400e-01 -2.10940406e-01 -6.62871778e-01 -1.16822088e+00
-2.06034139e-01 -1.76313162e-01 2.14796260e-01 -1.29001939e+00
-3.79587233e-01 -2.02026919e-01 7.63276368e-02 1.55044608e-02
-8.50235671e-02 3.88352424e-01 2.07759410e-01 1.93066075e-01
-5.06916463e-01 3.19017470e-01 9.77979064e-01 1.15706483e-02
-3.63368630e-01 2.13914931e-01 -1.40294448e-01 1.05852866e+00
6.02608144e-01 -2.20296726e-01 -2.01382115e-01 -5.75556278e-01
2.91733176e-01 3.30992401e-01 3.64185512e-01 -1.30280662e+00
8.54181528e-01 -3.31217438e-01 4.26162124e-01 -6.21941030e-01
5.35970211e-01 -1.23596644e+00 4.22496349e-01 3.18750024e-01
8.54308307e-02 1.43476024e-01 1.58388004e-01 8.02848816e-01
-2.11138994e-01 -3.63633066e-01 8.09441030e-01 1.33000880e-01
-1.34842539e+00 4.06556576e-01 -1.41521588e-01 -1.63160086e-01
1.47774434e+00 -7.29266346e-01 -3.21819007e-01 -5.21524847e-01
-6.74139977e-01 2.52065480e-01 7.59963751e-01 3.10475767e-01
6.57908618e-01 -9.52576756e-01 -6.01653636e-01 3.73071373e-01
2.09885344e-01 -2.06129521e-01 2.53327280e-01 1.23919678e+00
-7.49302149e-01 1.47714928e-01 -1.14991460e-02 -7.24146783e-01
-1.21496904e+00 5.92217088e-01 5.14105320e-01 1.32372677e-01
-9.76394236e-01 6.91010594e-01 9.12115201e-02 4.96673822e-01
2.76427031e-01 -2.99775183e-01 -2.95211047e-01 2.20824465e-01
7.16051042e-01 3.68123412e-01 2.26451516e-01 -6.90046728e-01
-3.19897085e-01 1.06092405e+00 2.63559639e-01 -7.53241777e-02
1.08942115e+00 -6.68823898e-01 3.62159997e-01 1.62857160e-01
8.86649311e-01 3.94267052e-01 -1.96394193e+00 -3.20973039e-01
8.60503465e-02 -8.06829870e-01 3.65253389e-01 8.84177425e-05
-1.20735705e+00 5.30309200e-01 5.87892234e-01 2.09839150e-01
1.25488412e+00 -6.62360251e-01 7.89578438e-01 1.94362044e-01
3.82796794e-01 -1.03612030e+00 -3.17648113e-01 6.23079360e-01
3.01407784e-01 -1.14493227e+00 1.59457192e-01 -8.05391133e-01
-4.38243985e-01 1.51764822e+00 2.18079925e-01 -1.42533407e-01
5.21077812e-01 3.20522547e-01 1.39815897e-01 -1.40711674e-02
-4.30992663e-01 -4.70544636e-01 2.10710898e-01 2.87185848e-01
-1.25397176e-01 -2.21037865e-01 -1.40122563e-01 1.18405476e-01
1.91466305e-02 -7.53964856e-02 7.78959572e-01 1.12484133e+00
-7.45117843e-01 -4.07894015e-01 -8.00010264e-01 1.50285199e-01
-3.40414971e-01 -7.72562847e-02 1.94300655e-02 6.06066525e-01
-1.71926692e-01 9.94992733e-01 1.19931981e-01 -3.69715303e-01
4.49637353e-01 -3.03268403e-01 2.99344510e-01 -1.10277697e-01
-3.06501210e-01 3.35224867e-01 8.75626951e-02 -4.39223439e-01
-3.76827449e-01 -7.20605969e-01 -1.10930121e+00 -1.66535199e-01
-6.52912199e-01 9.04889498e-03 7.71713972e-01 7.17147171e-01
1.95029318e-01 3.00995916e-01 7.71355212e-01 -1.02560985e+00
-4.64974612e-01 -4.01184291e-01 -6.04615927e-01 3.59525710e-01
7.01037586e-01 -6.35251820e-01 -7.94604778e-01 2.77911156e-01] | [7.800045490264893, -1.7048269510269165] |
c6d153bf-6010-4ab9-bb49-944b9dd00b23 | an-analysis-of-the-transfer-learning-of | 2011.02727 | null | https://arxiv.org/abs/2011.02727v2 | https://arxiv.org/pdf/2011.02727v2.pdf | An analysis of the transfer learning of convolutional neural networks for artistic images | Transfer learning from huge natural image datasets, fine-tuning of deep neural networks and the use of the corresponding pre-trained networks have become de facto the core of art analysis applications. Nevertheless, the effects of transfer learning are still poorly understood. In this paper, we first use techniques for visualizing the network internal representations in order to provide clues to the understanding of what the network has learned on artistic images. Then, we provide a quantitative analysis of the changes introduced by the learning process thanks to metrics in both the feature and parameter spaces, as well as metrics computed on the set of maximal activation images. These analyses are performed on several variations of the transfer learning procedure. In particular, we observed that the network could specialize some pre-trained filters to the new image modality and also that higher layers tend to concentrate classes. Finally, we have shown that a double fine-tuning involving a medium-size artistic dataset can improve the classification on smaller datasets, even when the task changes. | ['Saïd Ladjal', 'Yann Gousseau', 'Nicolas Gonthier'] | 2020-11-05 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 2.76479006e-01 1.77561417e-01 2.83569068e-01 -2.92715639e-01
3.78660336e-02 -7.52096176e-01 9.06856298e-01 8.44442844e-02
-5.82195878e-01 6.91885233e-01 5.43935522e-02 2.13028282e-01
-1.47486389e-01 -8.96896183e-01 -9.39658821e-01 -8.21538210e-01
2.52690312e-04 3.85165930e-01 4.06074405e-01 -3.69538724e-01
5.35712838e-01 7.54571617e-01 -1.56644928e+00 5.27682602e-01
3.56924564e-01 8.30863655e-01 -1.93996932e-02 4.66508329e-01
-7.77709717e-03 6.14200532e-01 -9.25882161e-01 -2.53070444e-01
2.24597678e-01 -2.94354439e-01 -7.69336641e-01 -8.99432600e-02
5.50663590e-01 1.18260011e-01 1.46733522e-02 9.20326531e-01
2.34556347e-01 1.79918438e-01 8.61922443e-01 -8.82213235e-01
-7.11173534e-01 6.58391178e-01 -1.37739390e-01 4.23928529e-01
1.19990977e-02 2.70651221e-01 7.04763234e-01 -7.64552772e-01
8.32212090e-01 1.14124537e+00 5.39819479e-01 5.27830422e-01
-1.61095929e+00 -6.35955572e-01 -7.01379031e-02 2.17941284e-01
-1.05193186e+00 -2.89592505e-01 9.13067043e-01 -8.90068531e-01
4.53906864e-01 1.18476771e-01 7.47315228e-01 1.09771323e+00
-4.44548838e-02 4.04969782e-01 1.38307309e+00 -6.19331777e-01
2.85686642e-01 7.18595982e-01 1.40430108e-01 3.82403016e-01
2.38967478e-01 1.65118754e-01 -3.24097365e-01 2.82062322e-01
9.67984617e-01 -1.46939486e-01 -2.42929459e-01 -6.61325812e-01
-1.06639385e+00 8.79602969e-01 7.87513733e-01 9.24896419e-01
-3.67367297e-01 1.92540243e-01 4.29552108e-01 3.75995040e-01
4.60813671e-01 9.96180117e-01 -6.10501170e-01 1.35239139e-01
-8.08041096e-01 -2.85261720e-01 4.81592625e-01 9.83668268e-02
1.13499117e+00 -7.17766955e-02 -1.60619780e-01 5.58818460e-01
-1.42931402e-01 -6.44135699e-02 6.21338367e-01 -8.31170022e-01
7.84433782e-02 9.02319252e-01 -1.77626491e-01 -8.96266997e-01
-2.95592010e-01 -7.88293302e-01 -6.42185986e-01 6.76573336e-01
8.96933198e-01 -4.20153625e-02 -7.21856713e-01 1.79257417e+00
-7.28202164e-02 -7.65757263e-02 -4.77862246e-02 5.94971061e-01
3.45109016e-01 3.99850667e-01 2.05230460e-01 2.25481182e-01
1.05906141e+00 -6.67692661e-01 -4.47293580e-01 -1.85945556e-01
5.25528967e-01 -4.32191074e-01 1.45056593e+00 3.51389080e-01
-9.15104210e-01 -9.14027989e-01 -1.10762250e+00 1.67859465e-01
-8.92282248e-01 2.06062883e-01 3.29228997e-01 5.26174784e-01
-1.04961610e+00 1.15217757e+00 -6.04912937e-01 -5.86242080e-01
7.11110592e-01 4.28820640e-01 -2.86200196e-01 2.95791894e-01
-9.78473365e-01 9.62342620e-01 5.40104628e-01 -7.96067566e-02
-7.62591958e-01 -7.09880829e-01 -4.53124613e-01 4.41136390e-01
1.70473859e-01 -3.97711992e-01 6.83684707e-01 -1.59764862e+00
-1.48806059e+00 1.18013990e+00 4.70235050e-01 -3.92368853e-01
4.07441497e-01 -9.42371115e-02 -9.90158170e-02 1.47638008e-01
-2.70264566e-01 6.98970973e-01 1.02303636e+00 -1.41861415e+00
-3.53360265e-01 -4.94712740e-01 1.45762727e-01 -2.31639907e-01
-7.29313612e-01 -2.69986331e-01 -5.08495467e-03 -7.09033370e-01
-2.75064468e-01 -8.77187967e-01 1.03602082e-01 -4.34842445e-02
-8.03914443e-02 1.46167846e-02 6.69226408e-01 -2.82107025e-01
9.59386230e-01 -2.40521359e+00 4.02049035e-01 4.77858692e-01
9.34148282e-02 4.51563954e-01 -1.24567620e-01 2.08572537e-01
-4.54537988e-01 1.94785461e-01 -1.09013863e-01 1.21815071e-01
-2.66546398e-01 6.05192631e-02 -2.04255864e-01 2.23314136e-01
3.19625497e-01 8.73816073e-01 -5.85035026e-01 -1.75603777e-01
2.76162505e-01 4.45718378e-01 -4.18088734e-01 -1.52876023e-02
-1.44890979e-01 6.53629005e-01 -2.05875620e-01 -7.45565519e-02
1.91080749e-01 -3.32112432e-01 1.69541985e-02 -4.48074102e-01
-1.65656153e-02 -7.46024251e-02 -7.65454531e-01 1.57621002e+00
-4.35005575e-01 1.11585426e+00 -2.19825476e-01 -1.20318890e+00
9.26729143e-01 5.29484451e-02 2.24620610e-01 -8.06765735e-01
3.91246557e-01 4.96671312e-02 3.03207040e-01 -3.31035465e-01
5.07033840e-02 -2.12143138e-01 2.89330631e-01 4.53518748e-01
2.97196776e-01 1.15025632e-01 2.93097883e-01 -1.74313083e-01
8.09132516e-01 5.20144068e-02 1.24358930e-01 -4.28142697e-01
6.66614711e-01 -8.46438035e-02 -2.96296906e-02 5.19523919e-01
6.28304332e-02 3.38735759e-01 5.78058720e-01 -6.78165793e-01
-9.07419264e-01 -8.20422530e-01 -1.19305171e-01 1.40968108e+00
-2.17585072e-01 -8.66087973e-02 -1.00289905e+00 -5.90921044e-01
-1.24678060e-01 6.86262131e-01 -1.34249866e+00 -6.62258148e-01
-6.62910938e-01 -6.46720171e-01 3.88930559e-01 4.43453074e-01
4.27609295e-01 -1.58547688e+00 -1.00778461e+00 -8.33624527e-02
3.22442412e-01 -9.00030971e-01 -2.28598975e-02 4.32389915e-01
-1.24690652e+00 -1.16472673e+00 -7.88213432e-01 -7.28973985e-01
7.47109532e-01 -2.58678734e-01 1.12434995e+00 5.57658561e-02
-4.49334115e-01 2.84887910e-01 -1.98324591e-01 -2.15687349e-01
-5.87320805e-01 3.53938401e-01 -3.48049939e-01 3.21440727e-01
1.71944216e-01 -8.62634599e-01 -4.89560097e-01 2.17756927e-01
-1.07272625e+00 -1.96959302e-01 8.41473758e-01 6.05864584e-01
1.66764989e-01 -9.44874212e-02 2.91398674e-01 -1.07190168e+00
7.08760679e-01 2.41295341e-02 -4.63283837e-01 2.82415301e-01
-6.13595903e-01 6.41139209e-01 5.53546429e-01 -5.79387426e-01
-9.78370905e-01 -2.07562931e-03 3.61731768e-01 -3.02288443e-01
-2.71990776e-01 3.03477198e-01 5.09946048e-02 -1.54172763e-01
1.15615380e+00 6.56738877e-02 -1.25251055e-01 -4.91917193e-01
4.03405070e-01 1.36077553e-01 4.16161805e-01 -5.38435996e-01
8.30320001e-01 5.04122257e-01 -4.51537147e-02 -6.83100879e-01
-7.94469774e-01 -1.01152971e-01 -9.98363495e-01 -2.79002845e-01
1.00562072e+00 -3.25177073e-01 -6.07469201e-01 3.68107498e-01
-9.56032991e-01 -7.56933272e-01 -8.04941893e-01 2.72606015e-01
-4.42358553e-01 -1.43549919e-01 -3.30828428e-01 -3.37382197e-01
-7.04471469e-02 -1.03839755e+00 6.25147104e-01 2.50026047e-01
-2.61656791e-01 -1.22692049e+00 3.10265362e-01 3.46587226e-02
5.20366192e-01 3.09187174e-01 1.29026699e+00 -6.89074695e-01
-2.91699380e-01 -7.08409101e-02 -2.37721369e-01 5.14549196e-01
2.24616647e-01 1.05415486e-01 -1.32533801e+00 -1.57790259e-01
-1.78023055e-02 -1.94232032e-01 1.00461864e+00 3.45189899e-01
1.32043552e+00 -1.15200929e-01 -2.16634795e-01 4.93430257e-01
1.36587000e+00 5.82123958e-02 7.77436972e-01 7.37891555e-01
3.60915899e-01 9.28950489e-01 1.12388775e-01 8.03943276e-02
-4.37510580e-01 6.33972406e-01 4.53266323e-01 -1.38399675e-01
-2.90892154e-01 -6.31528199e-02 2.64418602e-01 1.90318286e-01
-5.10913908e-01 1.87289938e-01 -7.42819309e-01 3.65041047e-01
-1.46599817e+00 -8.28427196e-01 2.17091903e-01 2.32332373e+00
7.49504566e-01 5.30354023e-01 1.42710358e-01 2.54648089e-01
7.12407827e-01 -6.36999011e-02 -4.45128709e-01 -3.43489081e-01
-1.58352554e-01 5.12759209e-01 3.33709121e-01 3.46425593e-01
-8.56163800e-01 9.86188829e-01 6.36219931e+00 6.26237929e-01
-1.54944789e+00 -7.12079853e-02 5.88383853e-01 2.30778791e-02
-6.22011945e-02 -1.09057359e-01 -2.62785137e-01 3.55597407e-01
8.52045596e-01 1.55130640e-01 4.98846889e-01 8.31252873e-01
-1.55046016e-01 -1.01016134e-01 -1.41163707e+00 8.87630403e-01
2.13721581e-02 -1.42718542e+00 1.85543448e-01 2.42940634e-01
6.05715394e-01 -2.18470246e-01 2.09413648e-01 3.55990916e-01
-1.56480595e-02 -1.03730917e+00 5.52346706e-01 7.38635719e-01
5.67974031e-01 -4.59052652e-01 6.27027333e-01 2.61878878e-01
-4.85133439e-01 -1.73781931e-01 -1.40671462e-01 3.07568628e-02
-3.79769981e-01 3.75498891e-01 -8.78220618e-01 -2.68005989e-02
7.90052354e-01 5.43441951e-01 -1.20124888e+00 7.50820160e-01
-2.53481895e-01 5.95092833e-01 -3.84182297e-02 6.77854121e-02
-8.11801758e-03 -2.08964691e-01 3.07371676e-01 1.11536264e+00
1.18425205e-01 -3.83445203e-01 -4.19979870e-01 1.01683986e+00
-1.84477344e-01 -2.27918345e-02 -4.96365368e-01 -2.55585819e-01
-8.96201208e-02 1.39378810e+00 -9.10796940e-01 -4.09674764e-01
-3.83068249e-02 8.89003932e-01 5.27081966e-01 4.10145611e-01
-5.31735182e-01 -3.49311918e-01 4.12958175e-01 4.43158954e-01
4.12394822e-01 -2.29847664e-03 -3.98433298e-01 -9.83968675e-01
-1.71406597e-01 -6.39776409e-01 1.38302177e-01 -9.64859307e-01
-8.91017139e-01 6.49796844e-01 -6.26810640e-02 -7.57764637e-01
-2.42559731e-01 -9.39742148e-01 -7.76500523e-01 5.98178685e-01
-1.20445621e+00 -6.99795067e-01 -4.50846553e-01 6.06701672e-01
1.96828306e-01 -3.04365993e-01 8.68731260e-01 2.52973646e-01
-4.56549734e-01 3.93852830e-01 9.66523290e-02 1.92041442e-01
7.03359008e-01 -1.23060846e+00 -8.34009200e-02 4.85015690e-01
4.42158341e-01 4.95782852e-01 8.66508305e-01 -1.71910629e-01
-6.44006789e-01 -6.11848176e-01 2.78142363e-01 -4.44585294e-01
6.40816092e-01 -4.58354115e-01 -1.18856323e+00 6.06512547e-01
3.27300340e-01 3.87136973e-02 5.18648148e-01 2.35826701e-01
-3.51502866e-01 -4.15803909e-01 -9.12184775e-01 4.39646006e-01
7.67129540e-01 -4.82760668e-01 -6.74189389e-01 1.08302139e-01
3.14080805e-01 2.58858085e-01 -7.72542715e-01 1.44251093e-01
5.58115482e-01 -1.27463627e+00 9.24451590e-01 -9.28939342e-01
3.88636321e-01 -2.10745279e-02 8.70100409e-02 -1.51047802e+00
-4.50922936e-01 -9.04408470e-02 1.86460301e-01 1.22818172e+00
4.83839035e-01 -5.19228339e-01 8.38143408e-01 3.31220716e-01
1.55514866e-01 -4.87963140e-01 -5.39707780e-01 -6.01708472e-01
1.77967265e-01 3.83933485e-02 2.21884981e-01 8.27141285e-01
-3.02107096e-01 5.45090735e-01 4.36698087e-02 -2.10409313e-01
2.73926228e-01 8.92672986e-02 7.62145817e-01 -1.65371943e+00
-2.36800611e-01 -6.34838700e-01 -5.29024363e-01 -2.20117927e-01
2.03448553e-02 -8.39470923e-01 -3.04337144e-01 -1.07171333e+00
5.96029572e-02 -2.73921579e-01 -5.37315428e-01 5.02906322e-01
1.30297914e-01 4.27208722e-01 1.96876392e-01 3.89593273e-01
-2.16998145e-01 3.67370993e-01 1.36205506e+00 -1.01897761e-01
-2.57424444e-01 2.22016685e-03 -4.95111227e-01 8.23077321e-01
1.04320538e+00 -5.74887335e-01 -2.52696216e-01 -3.11897457e-01
5.33119738e-01 -6.21351779e-01 5.62027991e-01 -1.24910784e+00
-7.12309182e-02 9.37258452e-02 7.81802356e-01 9.41960700e-03
2.76675969e-01 -9.30852354e-01 1.24398135e-02 7.80006588e-01
-6.74869239e-01 -2.61701673e-01 4.29139286e-01 2.83629566e-01
-2.58193702e-01 -4.35840607e-01 1.02562380e+00 -2.95863360e-01
-7.28493035e-01 -2.46790230e-01 -2.15713039e-01 2.53133625e-02
9.89032030e-01 -3.21484655e-01 -2.38516316e-01 -1.67171597e-01
-1.09445882e+00 -3.97070855e-01 4.49667960e-01 2.97129124e-01
1.13066740e-01 -1.15188074e+00 -4.44345057e-01 1.47394359e-01
1.02948770e-01 -2.59045780e-01 1.27732754e-03 6.36023879e-01
-5.17489612e-01 2.71629751e-01 -9.33999777e-01 -6.02066398e-01
-1.16420162e+00 8.34598839e-01 4.33697730e-01 -3.17774802e-01
-1.97208017e-01 6.19944572e-01 4.98799503e-01 -2.30178490e-01
1.96309701e-01 -4.21204120e-01 -4.35925514e-01 4.57697213e-01
2.22255439e-01 2.71835566e-01 1.92397714e-01 -3.93452823e-01
-1.45424381e-01 7.71453261e-01 -2.93727741e-02 -8.88909250e-02
1.59193730e+00 9.76652056e-02 -1.06021672e-01 6.05641246e-01
1.21645677e+00 -1.01091236e-01 -1.34931886e+00 -1.65565223e-01
2.44621366e-01 -1.83763117e-01 -1.99313581e-01 -9.35160518e-01
-1.27707911e+00 1.29246020e+00 8.53781283e-01 4.05430347e-01
1.19022477e+00 1.48226216e-01 3.75523530e-02 5.44571400e-01
-8.09860304e-02 -1.17030656e+00 5.91858864e-01 3.29652786e-01
1.08451068e+00 -1.06336248e+00 -1.73736081e-01 -7.84298405e-02
-3.67536813e-01 1.37916839e+00 4.55219954e-01 -4.72182840e-01
5.68843842e-01 -1.64187387e-01 8.20396841e-02 -4.15866464e-01
-3.78778219e-01 -6.67058900e-02 3.94019932e-01 4.75911587e-01
3.86385351e-01 -1.72540113e-01 1.17798135e-01 2.57591933e-01
-3.03729296e-01 1.05514385e-01 4.28590626e-01 4.87388879e-01
-4.36174303e-01 -1.05675924e+00 -3.48238349e-01 2.72087634e-01
-2.79359430e-01 5.41492067e-02 -7.70924926e-01 1.16956341e+00
3.95416260e-01 3.40008825e-01 1.16519816e-01 -2.58095860e-01
3.64989012e-01 3.30397964e-01 8.08792591e-01 -6.78768337e-01
-8.84822249e-01 -2.54281640e-01 -3.64622355e-01 -4.14011300e-01
-5.68370819e-01 -5.24120867e-01 -9.38744366e-01 2.30244219e-01
-1.76510796e-01 8.63966569e-02 8.32381129e-01 9.42638099e-01
1.82610974e-01 8.69661689e-01 3.97750050e-01 -9.46436226e-01
-4.08594787e-01 -1.23289967e+00 -4.52885360e-01 7.59929419e-01
1.95737347e-01 -6.51536405e-01 -5.25848091e-01 3.81433636e-01] | [9.761373519897461, 2.381601572036743] |
5ce24361-1282-42c4-999d-78934a5689af | towards-robust-face-recognition-with | 2208.13600 | null | https://arxiv.org/abs/2208.13600v2 | https://arxiv.org/pdf/2208.13600v2.pdf | Towards Robust Face Recognition with Comprehensive Search | Data cleaning, architecture, and loss function design are important factors contributing to high-performance face recognition. Previously, the research community tries to improve the performance of each single aspect but failed to present a unified solution on the joint search of the optimal designs for all three aspects. In this paper, we for the first time identify that these aspects are tightly coupled to each other. Optimizing the design of each aspect actually greatly limits the performance and biases the algorithmic design. Specifically, we find that the optimal model architecture or loss function is closely coupled with the data cleaning. To eliminate the bias of single-aspect research and provide an overall understanding of the face recognition model design, we first carefully design the search space for each aspect, then a comprehensive search method is introduced to jointly search optimal data cleaning, architecture, and loss function design. In our framework, we make the proposed comprehensive search as flexible as possible, by using an innovative reinforcement learning based approach. Extensive experiments on million-level face recognition benchmarks demonstrate the effectiveness of our newly-designed search space for each aspect and the comprehensive search. We outperform expert algorithms developed for each single research track by large margins. More importantly, we analyze the difference between our searched optimal design and the independent design of the single factors. We point out that strong models tend to optimize with more difficult training datasets and loss functions. Our empirical study can provide guidance in future research towards more robust face recognition systems. | ['Hongsheng Li', 'Yu Liu', 'Guanglu Song', 'Manyuan Zhang'] | 2022-08-29 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [-9.09459665e-02 -8.00799280e-02 -5.63962042e-01 -4.33854133e-01
-7.91707635e-01 -3.48389596e-01 2.46701270e-01 -4.09391075e-01
-1.74315840e-01 4.07691628e-01 9.07988399e-02 -2.46353358e-01
-5.87206423e-01 -4.89087820e-01 -7.12369919e-01 -8.06106806e-01
2.25060791e-01 2.44424626e-01 -3.16231936e-01 -2.25008521e-02
3.45755637e-01 6.12321258e-01 -1.77396333e+00 -3.45615029e-01
8.67210031e-01 1.08429611e+00 4.68062572e-02 1.07372031e-01
-7.28502346e-04 2.27577209e-01 -5.20289958e-01 -5.54148972e-01
5.01975238e-01 -2.45756596e-01 -4.48993444e-01 2.43070588e-01
6.77274942e-01 -7.31352940e-02 5.01390807e-02 9.39800560e-01
7.18494952e-01 -1.67316630e-01 3.64215761e-01 -1.47752368e+00
-6.99378908e-01 5.56428850e-01 -6.96347654e-01 -1.04525246e-01
-1.36253849e-01 4.33230042e-01 1.06569839e+00 -7.48621285e-01
2.55602747e-01 1.12636065e+00 8.64545941e-01 7.88034797e-01
-1.25763440e+00 -8.44542265e-01 6.39091671e-01 1.12016454e-01
-1.52353394e+00 -8.21801662e-01 9.03549314e-01 -3.70642513e-01
8.36843610e-01 3.23711097e-01 4.32253748e-01 1.03094888e+00
-2.47633783e-03 7.66321063e-01 9.78273094e-01 -5.03303289e-01
2.71488607e-01 2.67032236e-01 3.32852542e-01 9.20158207e-01
6.82301164e-01 3.22131723e-01 -4.21655297e-01 -2.53858238e-01
4.37656850e-01 -1.16512179e-01 -1.59533709e-01 -8.29459369e-01
-6.13353312e-01 7.73950994e-01 -6.05991064e-03 2.02155188e-01
-2.08683714e-01 2.38575652e-01 2.30625972e-01 3.23714733e-01
6.44692257e-02 9.01041329e-01 -5.98255515e-01 1.30941004e-01
-9.19751167e-01 1.74443409e-01 8.14634740e-01 9.18648541e-01
7.15501368e-01 9.24623758e-02 -3.57661307e-01 9.83922601e-01
4.00481969e-01 4.43421543e-01 1.34365216e-01 -1.07939410e+00
3.09864819e-01 7.63968170e-01 -5.62008424e-03 -1.06821311e+00
-3.91069651e-01 -8.12304795e-01 -6.97771847e-01 -2.37148888e-02
3.13592464e-01 -3.05316031e-01 -8.26797187e-01 2.09769773e+00
3.81265700e-01 1.34862764e-02 -1.94129229e-01 7.85899878e-01
4.91664767e-01 8.98688566e-03 7.02668950e-02 -3.32370073e-01
1.52434337e+00 -1.06256771e+00 -7.97738254e-01 -3.59389067e-01
5.70032477e-01 -6.41735375e-01 1.07037008e+00 5.04366338e-01
-8.98117065e-01 -4.08571810e-01 -1.13399220e+00 3.57619971e-01
-5.85005172e-02 4.88674521e-01 8.32490623e-01 1.10100555e+00
-8.50315571e-01 3.35600793e-01 -7.58365989e-01 -3.61639619e-01
6.18512213e-01 5.37524581e-01 -1.01520538e-01 -1.02437623e-02
-8.78267944e-01 8.60732615e-01 -7.07992241e-02 2.13987991e-01
-7.59388208e-01 -1.02586138e+00 -5.79854906e-01 1.97603703e-02
8.89167666e-01 -8.50734055e-01 8.79580677e-01 -7.16735005e-01
-1.38528657e+00 5.79208672e-01 -2.85350680e-01 -2.76423275e-01
2.59104431e-01 -1.48189560e-01 -3.79110217e-01 -2.51665890e-01
-1.76128924e-01 4.12161261e-01 1.13911021e+00 -1.44147837e+00
-6.03309333e-01 -5.96601784e-01 -1.32606804e-01 -8.21384490e-02
-7.99144089e-01 -2.73834784e-02 -8.92125368e-01 -6.61687911e-01
-2.36991316e-01 -8.72593403e-01 -2.80332059e-01 -1.15149841e-01
-2.36757666e-01 -1.58355519e-01 4.78176415e-01 -4.38816130e-01
1.68745041e+00 -2.15422463e+00 1.64573997e-01 2.48069867e-01
-8.23653713e-02 3.34298849e-01 -4.21958059e-01 5.11656329e-02
-1.86737794e-02 3.85821372e-01 -1.33573443e-01 -5.65249383e-01
1.86968908e-01 2.27860138e-01 -2.03709632e-01 4.51075405e-01
3.66202712e-01 8.87679100e-01 -4.73653138e-01 -2.95598060e-01
-1.45284161e-01 4.74586010e-01 -1.07282746e+00 1.71839118e-01
-6.64582923e-02 -5.39961681e-02 -5.67237079e-01 9.84842479e-01
8.12569022e-01 -1.47027910e-01 2.40281239e-01 -5.69240034e-01
8.81162956e-02 -7.91408941e-02 -1.51848149e+00 1.32901359e+00
-3.55259925e-01 -1.68425851e-02 4.67084557e-01 -1.08282125e+00
1.04283893e+00 1.53748887e-02 6.58590436e-01 -6.95445001e-01
-2.06899401e-02 1.12289838e-01 -5.91180325e-02 -5.61534345e-01
2.47464851e-01 -5.30317947e-02 6.95059896e-02 4.48601365e-01
-5.99167757e-02 2.33045161e-01 3.06655616e-02 -4.06102866e-01
8.99434686e-01 -1.67648628e-01 2.77404577e-01 -3.23401928e-01
5.48985064e-01 -3.10576260e-01 9.84124541e-01 9.02594507e-01
-4.55690384e-01 5.18602431e-01 5.74365139e-01 -4.05568361e-01
-7.84530818e-01 -5.19583344e-01 -3.50213498e-01 1.04166973e+00
3.17168795e-02 -4.10835415e-01 -8.23036909e-01 -8.64866138e-01
3.32447857e-01 5.23302615e-01 -7.33666718e-01 -2.98525304e-01
-6.49325311e-01 -1.09503984e+00 6.18183017e-01 3.90318543e-01
2.05497548e-01 -8.10800970e-01 -4.71968472e-01 -3.82201672e-02
2.13144779e-01 -9.69492018e-01 -6.21283710e-01 3.30730349e-01
-7.74562299e-01 -1.11564267e+00 -2.19016209e-01 -5.63646197e-01
7.50688791e-01 1.50993630e-01 1.07265913e+00 5.36428511e-01
-3.61524016e-01 4.53575730e-01 -1.63827345e-01 -3.06486636e-01
-1.38108477e-01 3.80675644e-01 1.84857875e-01 1.90508723e-01
4.51015294e-01 -3.57379377e-01 -5.36392987e-01 6.99807286e-01
-7.78185129e-01 -4.38875228e-01 7.18639195e-01 9.67537165e-01
5.35187602e-01 2.36171618e-01 6.42876446e-01 -7.11269379e-01
7.63770282e-01 -4.31939274e-01 -7.54155517e-01 7.23125637e-01
-1.33360016e+00 3.44504416e-01 4.23867971e-01 -3.10278803e-01
-8.76985788e-01 1.83388591e-01 2.56692115e-02 -5.35891235e-01
-1.58478960e-01 1.94608539e-01 -6.13093555e-01 -2.46213526e-01
4.09748137e-01 4.69556972e-02 1.47862300e-01 -6.68628633e-01
2.13252082e-01 5.45608938e-01 1.83792174e-01 -9.42235589e-01
7.58031905e-01 4.49324936e-01 -1.09327771e-01 -5.89207351e-01
-6.97613418e-01 -1.71609625e-01 -2.22862467e-01 1.33533433e-01
4.41644430e-01 -8.42157841e-01 -8.26760769e-01 3.56506407e-01
-8.00329685e-01 -1.83477197e-02 -1.16437688e-01 9.76661369e-02
-4.05952543e-01 3.67328733e-01 -1.54829949e-01 -9.05428588e-01
-4.16967660e-01 -1.57670414e+00 1.13338125e+00 2.35647827e-01
-2.28362828e-01 -7.11377144e-01 -6.26540408e-02 7.74523318e-01
4.85049367e-01 -2.35996246e-01 8.54471326e-01 -5.13950467e-01
-5.74542165e-01 2.81872880e-02 -1.17278442e-01 4.03619289e-01
2.63657033e-01 2.33317658e-01 -9.54271436e-01 -5.23278713e-01
1.77303284e-01 -9.58818123e-02 9.48569357e-01 4.82807785e-01
1.43728483e+00 -3.49628240e-01 -2.84406692e-01 9.73858595e-01
1.45781577e+00 3.17779392e-01 4.58561182e-01 5.64012229e-01
6.94726527e-01 7.52352655e-01 5.75025499e-01 4.55090404e-01
2.43465498e-01 9.14335251e-01 3.84802878e-01 -2.43291855e-02
9.40311234e-03 -1.33760005e-01 3.50532860e-01 3.91078144e-01
1.93663567e-01 6.89336881e-02 -8.07537138e-01 4.40393597e-01
-1.79899669e+00 -8.20484757e-01 5.06017268e-01 2.13642263e+00
7.96972156e-01 -3.52606103e-02 9.73909125e-02 -6.52469741e-03
5.81853390e-01 -4.62546982e-02 -7.65912533e-01 -3.48149627e-01
-1.52550176e-01 -3.54203433e-02 6.07367694e-01 5.04376829e-01
-1.07473457e+00 8.12663674e-01 6.91421747e+00 1.00777602e+00
-1.00859213e+00 -1.99471727e-01 6.50975049e-01 -4.60956693e-01
-5.11898994e-01 -2.73636915e-02 -1.36980748e+00 4.01060581e-01
6.31873310e-01 3.70335691e-02 8.05844665e-01 9.32021976e-01
2.54597217e-01 2.71118820e-01 -1.20163691e+00 1.09019220e+00
1.09688275e-01 -1.23310435e+00 1.98286489e-01 3.48977119e-01
6.97862744e-01 -3.67281288e-01 2.23076954e-01 2.99309790e-01
1.71967112e-02 -1.24369919e+00 6.32684469e-01 5.39708674e-01
3.25503796e-01 -7.92334795e-01 5.02044261e-01 5.57059869e-02
-9.54352319e-01 -6.02551997e-01 -7.60002434e-02 2.45830745e-01
-1.67513743e-01 6.87681317e-01 -2.97480434e-01 4.10918504e-01
6.97993875e-01 3.77527148e-01 -7.40701973e-01 9.66189086e-01
5.49428463e-02 5.29634237e-01 -2.44826630e-01 1.39067456e-01
-1.46351203e-01 -2.75366366e-01 5.01220942e-01 9.57177639e-01
2.45667860e-01 -1.59213349e-01 1.21481381e-01 9.03165936e-01
-3.89464423e-02 6.55277893e-02 -4.30946231e-01 -1.44668609e-01
7.51698017e-01 1.11529636e+00 -3.98212135e-01 2.23372966e-01
-3.65235448e-01 2.77779013e-01 4.54040319e-01 4.02547240e-01
-7.69547403e-01 -6.53129071e-02 1.19653797e+00 1.23425294e-02
6.30952716e-01 2.93137841e-02 -8.07552516e-01 -1.05550873e+00
4.21002120e-01 -1.38112640e+00 5.02396524e-01 -3.14029232e-02
-1.43949234e+00 4.93977308e-01 -1.11560725e-01 -8.24476957e-01
4.87757027e-02 -5.29298186e-01 -3.80483180e-01 6.57997847e-01
-1.72377431e+00 -8.86787772e-01 1.88357588e-02 4.64369327e-01
3.07993233e-01 -5.14916360e-01 6.10039234e-01 4.88372535e-01
-1.33698523e+00 1.15789282e+00 -7.55737200e-02 -1.16687350e-01
8.38468373e-01 -7.28986084e-01 1.32917851e-01 8.33265424e-01
-1.57181956e-02 1.04521811e+00 6.94892347e-01 -4.95503306e-01
-1.97713780e+00 -8.40397239e-01 5.54888964e-01 -5.11493862e-01
4.17978853e-01 -3.29660743e-01 -8.96678567e-01 2.24329188e-01
-3.68024632e-02 -2.44095966e-01 7.74755359e-01 4.95220065e-01
-3.91651362e-01 -7.39475727e-01 -1.24000728e+00 6.99957550e-01
1.23676515e+00 -2.43958190e-01 -2.72557676e-01 -7.79108480e-02
6.07467830e-01 -4.44773100e-02 -9.01514232e-01 7.83940256e-01
6.02536798e-01 -9.10803974e-01 1.08625507e+00 -6.34836853e-01
4.51159775e-02 -2.55093247e-01 -2.69761413e-01 -1.09778166e+00
-3.45630586e-01 -7.53677785e-01 -4.06707644e-01 1.53943825e+00
4.78344142e-01 -5.97676218e-01 1.03120279e+00 9.69667435e-01
-3.70742800e-03 -1.31439745e+00 -8.13058794e-01 -8.98721993e-01
1.87015459e-02 -4.76456016e-01 1.08051944e+00 6.86618268e-01
-5.11611283e-01 2.14574456e-01 -5.33087492e-01 2.71521598e-01
8.21926951e-01 3.31410497e-01 8.33321929e-01 -1.19296789e+00
-3.46383005e-01 -8.72570992e-01 2.05344446e-02 -8.13658714e-01
1.65780634e-01 -5.11859596e-01 -1.38450116e-01 -9.97000098e-01
3.00359488e-01 -8.12708259e-01 -4.64458704e-01 6.78972602e-01
-3.11293870e-01 -1.21715866e-01 1.46992981e-01 2.50854120e-02
-3.72223943e-01 7.55362988e-01 1.13143897e+00 -1.15034945e-01
-3.33580166e-01 -7.80669153e-02 -1.59017301e+00 4.58331972e-01
8.32380474e-01 -4.22151208e-01 -4.43216294e-01 -5.76897442e-01
7.95192048e-02 -3.71513039e-01 1.27715781e-01 -7.96025217e-01
2.90548801e-01 -5.21579623e-01 1.38981491e-01 -1.17239118e-01
9.04528126e-02 -1.12450910e+00 8.92514810e-02 3.46515030e-01
-2.22147509e-01 -1.78598836e-01 2.78544217e-01 5.76681614e-01
4.42761853e-02 -2.08606243e-01 8.32106769e-01 1.92707121e-01
-5.46894312e-01 3.89695615e-01 2.32476562e-01 2.93172952e-02
1.00680983e+00 -3.41180533e-01 -1.61725044e-01 -5.68386763e-02
-4.22690868e-01 5.70726871e-01 4.98300731e-01 7.82398462e-01
2.82392591e-01 -1.30201972e+00 -4.68384176e-01 5.84344506e-01
2.96375696e-02 -3.57859075e-01 9.01459381e-02 7.63043225e-01
1.56253040e-01 5.14101684e-01 -1.33320913e-01 -4.64185268e-01
-1.33540893e+00 6.36111200e-01 4.61168200e-01 -3.51873189e-01
-1.57847136e-01 6.58592999e-01 1.05916768e-01 -5.06998539e-01
8.07904184e-01 -1.26387239e-01 -1.46016210e-01 2.89611906e-01
4.77602541e-01 3.58696193e-01 2.36324877e-01 -3.03693116e-01
-5.56515992e-01 9.64448273e-01 -1.50792584e-01 2.85110235e-01
1.56286192e+00 -1.94938213e-01 -6.86782375e-02 -3.18149179e-01
1.11512721e+00 -2.19270792e-02 -1.31403911e+00 -8.06047544e-02
5.98996691e-02 -4.74239230e-01 2.14865997e-01 -7.79285908e-01
-1.58827138e+00 4.38109815e-01 6.12511635e-01 -2.57612672e-02
1.36911297e+00 -2.57479846e-01 4.85782236e-01 4.36376154e-01
3.37488890e-01 -1.21364379e+00 -3.78527381e-02 2.32132867e-01
8.19682837e-01 -1.30768073e+00 1.00674756e-01 -3.56761545e-01
-5.08658350e-01 8.21155488e-01 1.02239180e+00 1.74200073e-01
6.79242551e-01 5.38129628e-01 -4.56023812e-02 -3.49498689e-01
-9.32937503e-01 -4.37567569e-02 3.44366014e-01 5.80326557e-01
2.84423620e-01 -7.85745829e-02 -3.33566308e-01 9.27468479e-01
-5.55250272e-02 9.50563625e-02 -9.80106592e-02 7.89343178e-01
-3.16131920e-01 -1.56663871e+00 -6.04016066e-01 4.91340756e-01
-3.38285148e-01 9.88718420e-02 -3.39816004e-01 6.81652069e-01
3.16881239e-01 9.55236673e-01 -2.11305305e-01 -4.18601602e-01
7.06986427e-01 1.64541125e-01 4.82774734e-01 -2.70505667e-01
-6.95981026e-01 -1.82042196e-02 -5.12902178e-02 -7.07732141e-01
-1.85178041e-01 -5.83383918e-01 -7.38706648e-01 -2.94949830e-01
-3.45582575e-01 1.66719228e-01 7.07828462e-01 7.97396123e-01
8.12494457e-01 3.75795066e-01 7.73497343e-01 -2.46779084e-01
-1.01443672e+00 -4.03162539e-01 -4.24606264e-01 1.37713775e-01
3.31051409e-01 -9.15087700e-01 -4.10553366e-01 -3.45479637e-01] | [9.343235969543457, 3.205422878265381] |
a9eddadf-a638-4dc2-8749-8ce820db0c23 | improved-attention-models-for-memory | 1910.01189 | null | http://arxiv.org/abs/1910.01189v7 | http://arxiv.org/pdf/1910.01189v7.pdf | Improved Attention Models for Memory Augmented Neural Network Adaptive Controllers | We introduced a {\it working memory} augmented adaptive controller in our
recent work. The controller uses attention to read from and write to the
working memory. Attention allows the controller to read specific information
that is relevant and update its working memory with information based on its
relevance. The retrieved information is used to modify the final control input
computed by the controller. We showed that this modification speeds up
learning. In the above work, we used a soft-attention mechanism for the
adaptive controller. Controllers that use soft attention or hard attention
mechanisms are limited either because they can forget the information or fail
to shift attention when the information they are reading becomes less relevant.
We propose an attention mechanism that comprises of (i) a hard attention
mechanism and additionally (ii) an attention reallocation mechanism. The
attention reallocation enables the controller to reallocate attention to a
different location when the relevance of the location it is reading from
diminishes. The reallocation also ensures that the information stored in the
memory before the shift in attention is retained which can be lost in both soft
and hard attention mechanisms. We illustrate through detailed simulations of
various scenarios for two link robot and three link robot arm systems we
illustrate the effectiveness of the proposed attention mechanism. | [] | 2020-03-19 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 3.93711627e-01 4.30532634e-01 -2.69572139e-01 2.87981451e-01
-1.56798456e-02 -2.92429954e-01 1.89252958e-01 2.62615949e-01
-6.18667841e-01 9.60067630e-01 -3.73986699e-02 8.65887329e-02
-4.09433305e-01 -8.16394567e-01 -6.62399888e-01 -9.21207249e-01
3.92770112e-01 4.52107191e-01 9.23789978e-01 -2.71944791e-01
6.60982311e-01 4.98164326e-01 -1.81634617e+00 -1.93218559e-01
3.71951580e-01 6.69640720e-01 1.06014907e+00 7.21129656e-01
-5.42330137e-03 8.88316154e-01 -7.54008412e-01 8.28576624e-01
3.83363478e-02 -3.34952295e-01 -1.04430854e+00 -1.69477925e-01
-6.47428855e-02 1.26578324e-02 -2.30904724e-02 8.50981891e-01
4.54135120e-01 5.92017591e-01 4.17832464e-01 -1.22719944e+00
-5.68577588e-01 2.82702386e-01 -3.95439357e-01 7.28549123e-01
2.16215506e-01 8.46920311e-02 4.30413663e-01 -6.85368598e-01
4.85493720e-01 1.18272555e+00 1.13610737e-01 6.06837153e-01
-8.88865352e-01 -6.19643152e-01 5.29085755e-01 3.32142234e-01
-1.16653359e+00 -4.27341908e-01 5.07148564e-01 -4.11232144e-01
1.22869694e+00 1.38241798e-01 5.43945253e-01 2.66342968e-01
7.67621279e-01 3.30695540e-01 5.86502850e-01 -8.65979075e-01
3.08248311e-01 8.88329521e-02 4.36423451e-01 5.36764860e-01
4.55815673e-01 2.56183207e-01 -5.50437093e-01 1.91247240e-01
8.55165720e-01 -5.92392916e-03 -4.73692417e-01 -2.97498167e-01
-9.27423060e-01 5.50960362e-01 6.55015767e-01 4.41413879e-01
-7.42864370e-01 3.69336426e-01 1.05318166e-01 6.05146646e-01
-7.93505982e-02 5.54093421e-01 -3.28497142e-01 7.08997771e-02
-2.20011413e-01 -3.39965224e-01 5.90600610e-01 9.59711194e-01
7.93651760e-01 1.71944439e-01 -2.63576388e-01 5.19191384e-01
2.67031133e-01 5.37833452e-01 7.71183372e-01 -8.33798409e-01
2.88241148e-01 7.80789971e-01 3.54025722e-01 -9.55612957e-01
-3.69815469e-01 -1.90696642e-01 -5.63149929e-01 7.38002479e-01
2.54534502e-02 -2.59360015e-01 -7.55587399e-01 2.00639796e+00
1.76524058e-01 -1.48657858e-01 -1.71732400e-02 9.79587018e-01
2.07165822e-01 6.68050647e-01 7.30940551e-02 -4.64686900e-01
1.06875336e+00 -9.44994330e-01 -9.80464697e-01 -3.06061894e-01
1.83214471e-01 -7.09405243e-01 8.23050082e-01 2.00655580e-01
-1.27231538e+00 -9.26412761e-01 -1.41677439e+00 3.13701540e-01
-4.32111442e-01 -5.64610250e-02 -4.86367494e-02 2.97670439e-02
-1.08062530e+00 5.44134498e-01 -8.06304157e-01 -7.93347538e-01
-3.26245785e-01 7.33955860e-01 -1.43907651e-01 4.64419335e-01
-1.15019190e+00 1.26932085e+00 6.33169234e-01 -9.09389369e-03
-7.15816200e-01 -6.90333471e-02 -5.07982671e-01 2.99082875e-01
6.88018560e-01 -7.36842155e-01 1.09755266e+00 -1.22300744e+00
-1.66014552e+00 3.19506228e-01 -1.17832072e-01 -3.51185858e-01
-2.07919091e-01 -2.89722562e-01 -1.17484562e-01 2.10602116e-02
-9.01629403e-02 5.78623176e-01 1.24901116e+00 -9.76844311e-01
-8.24180126e-01 -2.79671103e-01 1.60236191e-02 2.78267562e-01
-2.74985969e-01 -3.75495970e-01 -2.70128727e-01 -5.77205360e-01
1.69051781e-01 -1.07435143e+00 3.67774129e-01 -9.77066383e-02
-1.53314695e-01 -3.28277424e-02 1.20014918e+00 4.62011211e-02
1.59190369e+00 -2.03340197e+00 5.94541609e-01 7.25355595e-02
-9.10448804e-02 5.91879070e-01 -1.03151008e-01 5.96094251e-01
-3.51883143e-01 -1.25405028e-01 2.66208142e-01 1.54229343e-01
-3.92775983e-01 3.71606976e-01 -3.26211393e-01 8.03344771e-02
1.90560952e-01 2.19645277e-01 -7.58514643e-01 -3.92246515e-01
1.77644789e-01 2.92882621e-01 -4.56371367e-01 3.32321554e-01
5.63341714e-02 3.29946250e-01 -5.26772141e-01 1.34493336e-01
2.53119498e-01 3.50418314e-02 4.63408679e-02 -4.62913364e-02
-4.67326403e-01 1.92960724e-01 -1.41090155e+00 7.79884160e-01
-3.32707047e-01 3.31749260e-01 4.45149362e-01 -6.87412083e-01
9.69908774e-01 3.98292303e-01 -9.23540890e-02 -7.98630476e-01
3.54946792e-01 -1.72996327e-01 1.60671040e-01 -4.44163889e-01
5.98668993e-01 6.25874400e-02 2.05109701e-01 7.64800787e-01
-1.59409389e-01 1.04923092e-01 3.43773626e-02 -1.49212569e-01
1.10381806e+00 7.87013322e-02 5.85156441e-01 -4.31924224e-01
1.18527269e+00 6.85474947e-02 5.20647585e-01 7.93153703e-01
-4.40521866e-01 4.22048867e-02 4.61540557e-02 -1.49251297e-01
-5.74582160e-01 -8.03585052e-01 5.70216835e-01 1.89208734e+00
3.85461658e-01 1.28446370e-01 -4.54218805e-01 -3.18246573e-01
5.22540286e-02 6.79962754e-01 -7.34989226e-01 -9.22146142e-01
-7.59598017e-01 -2.45909784e-02 -9.07665789e-02 6.84833169e-01
5.94438970e-01 -1.52329266e+00 -1.35772502e+00 1.87986493e-01
2.14140400e-01 -8.95617064e-03 -8.06424141e-01 7.29642987e-01
-8.71614695e-01 -1.09950948e+00 -3.85708243e-01 -1.10735357e+00
8.61729443e-01 4.38649625e-01 4.26775455e-01 4.24048364e-01
2.05013677e-01 6.00710034e-01 -2.27378815e-01 -4.99735981e-01
-3.01871091e-01 3.45655590e-01 1.26981258e-01 -2.53186136e-01
9.71584395e-02 -1.10187843e-01 -3.26127589e-01 5.10259807e-01
-7.46306181e-01 -1.88790873e-01 6.94392204e-01 8.52358758e-01
5.18673003e-01 1.52620256e-01 9.92465138e-01 -5.16836226e-01
7.08785236e-01 -4.56108272e-01 -7.13478863e-01 3.29289548e-02
-6.75657153e-01 6.14818156e-01 3.17754239e-01 -8.17543507e-01
-1.08143604e+00 5.93726449e-02 3.01822960e-01 -3.43431324e-01
2.59701043e-01 4.80436474e-01 2.14676186e-03 4.16933838e-03
1.91787049e-01 1.09658584e-01 2.92833328e-01 -2.29917303e-01
-1.46781206e-01 5.86956918e-01 3.59201968e-01 -2.92706549e-01
3.00892144e-01 -5.03449813e-02 4.58190888e-02 -4.58247632e-01
-2.38713622e-01 -3.49963039e-01 -7.40707040e-01 -3.34962159e-01
9.46442723e-01 -2.59636402e-01 -9.99417782e-01 3.19049090e-01
-8.96736145e-01 -4.26038265e-01 -4.31880206e-01 4.11109775e-01
-4.72875565e-01 -8.13440010e-02 -4.51489091e-01 -9.45506096e-01
-3.94250631e-01 -1.12179399e+00 6.86332524e-01 4.40906048e-01
-6.30663157e-01 -8.63508105e-01 1.56923220e-01 -5.68027377e-01
8.71532619e-01 -3.23195249e-01 1.20067239e+00 -5.29252529e-01
-5.07189274e-01 -1.05361640e-01 6.21569194e-02 3.13702039e-02
4.03932542e-01 -2.83733875e-01 -5.53512454e-01 -6.09433770e-01
1.70432359e-01 1.90718565e-02 8.75074685e-01 1.61923960e-01
4.27416265e-01 -2.19759256e-01 -7.64844418e-01 -2.25470245e-01
1.29064918e+00 8.15776765e-01 6.59487963e-01 7.09645748e-01
2.78167188e-01 2.98029363e-01 8.27798665e-01 1.87845856e-01
-2.75892373e-02 6.74008489e-01 5.67442954e-01 2.18641281e-01
-3.38166028e-01 1.79951265e-01 4.48007226e-01 8.96690786e-01
7.37369284e-02 -3.65311801e-01 -5.54056108e-01 7.23749280e-01
-2.00614691e+00 -9.61827219e-01 1.42800137e-01 2.71469855e+00
9.73946214e-01 5.65191448e-01 -3.17217141e-01 1.81399569e-01
1.03377724e+00 -3.04175258e-01 -6.80641949e-01 -8.31674516e-01
3.30012351e-01 2.03476071e-01 3.71319652e-01 8.81196499e-01
-6.81278944e-01 7.48157203e-01 6.00687313e+00 3.92798752e-01
-1.39849257e+00 6.56701848e-02 -2.00313166e-01 -1.80571258e-01
5.63511729e-01 -1.38240561e-01 -9.41639006e-01 4.60771590e-01
7.92976201e-01 -2.52383828e-01 2.96719790e-01 7.37439156e-01
-2.69668410e-03 -9.55729902e-01 -1.03986514e+00 5.21526754e-01
1.06385291e-01 -5.73944926e-01 1.01598762e-01 -9.92909819e-02
1.11606300e-01 -5.38214862e-01 1.88387543e-01 5.31087279e-01
4.51047868e-02 -4.88146663e-01 7.56711602e-01 1.23704410e+00
1.35875568e-01 -6.72417462e-01 8.10596764e-01 5.85868478e-01
-1.27623141e+00 -4.53705072e-01 -3.26774448e-01 -5.11041939e-01
-7.97395632e-02 -1.26589090e-01 -7.66184807e-01 6.36322945e-02
4.53183144e-01 7.82136619e-02 -4.46893483e-01 9.40323651e-01
-2.04477847e-01 1.68598950e-01 -2.07745939e-01 -1.41939759e-01
-6.80961013e-02 2.81015813e-01 7.93171108e-01 5.83844423e-01
2.42603153e-01 2.45192975e-01 9.18996856e-02 7.07486272e-01
3.65079522e-01 -3.62425596e-01 -4.07852471e-01 2.10509688e-01
6.00886583e-01 8.21323872e-01 -6.12310350e-01 -5.06297648e-01
-2.32389034e-03 8.57528210e-01 3.26780140e-01 2.68006593e-01
-6.12702549e-01 -8.66459727e-01 2.43550479e-01 1.48774788e-01
4.56589431e-01 -1.88020363e-01 5.81072420e-02 -1.46402866e-01
-5.65265596e-01 -2.03838781e-01 5.00890613e-01 -1.18278241e+00
-7.31084466e-01 6.86431408e-01 1.84055388e-01 -1.04398930e+00
-3.27230215e-01 -3.34399909e-01 -5.80272496e-01 1.17923379e+00
-1.19081068e+00 -4.14121985e-01 -3.35264981e-01 3.88246745e-01
7.54883289e-01 -2.64525503e-01 7.21552134e-01 -9.00308117e-02
-6.30042136e-01 3.38265568e-01 -3.57798994e-01 -4.55943763e-01
1.02276838e+00 -9.75024462e-01 -6.60185277e-01 5.56241393e-01
-9.21341419e-01 9.58154500e-01 6.61320508e-01 -9.41505909e-01
-1.39529872e+00 -1.01657581e+00 7.99562931e-01 -1.99214503e-01
2.98207641e-01 1.98704526e-01 -1.22627914e+00 8.95395696e-01
4.87570345e-01 -5.19830644e-01 1.98480397e-01 -4.36325014e-01
2.82872260e-01 -1.57557443e-01 -1.21567321e+00 4.93041068e-01
4.05594885e-01 -9.79444906e-02 -1.14583981e+00 -2.30041534e-01
6.86702549e-01 -1.77509964e-01 -4.00921226e-01 5.06504327e-02
5.68638682e-01 -5.59289098e-01 5.69600344e-01 -2.37536460e-01
-2.88497627e-01 -7.39424944e-01 -3.21531333e-02 -1.36139572e+00
-9.10928071e-01 -4.08180237e-01 -1.27666861e-01 1.08350480e+00
1.43470198e-01 -9.51380134e-01 1.66934773e-01 3.79969805e-01
-2.23346949e-01 -3.90026778e-01 -7.98010588e-01 -6.61502361e-01
-1.44987792e-01 2.01552317e-01 5.27411163e-01 4.23789740e-01
4.62646544e-01 6.89012468e-01 -3.38837095e-02 2.91974932e-01
2.97332034e-02 -1.06913276e-01 3.60227495e-01 -1.16995084e+00
-1.27294794e-01 -4.68191147e-01 -1.19331546e-01 -8.23842168e-01
-2.62727179e-02 -5.66770613e-01 4.01489317e-01 -1.42082715e+00
2.18192935e-01 -3.42029154e-01 -4.73044485e-01 9.08487141e-01
-1.05685778e-01 -3.44116911e-02 3.34787220e-01 4.19712305e-01
-6.47337317e-01 2.71105289e-01 1.10270321e+00 5.08800708e-02
-8.52657020e-01 5.91689087e-02 -3.58032227e-01 6.26463950e-01
8.79390001e-01 -4.40978199e-01 -5.22401035e-01 -3.69981170e-01
8.66998434e-02 1.23764426e-01 1.74046233e-01 -1.45233941e+00
7.34404385e-01 -2.08992973e-01 3.67600352e-01 -6.43137217e-01
3.35080475e-01 -1.25995064e+00 1.31003588e-01 1.03766084e+00
-5.43863595e-01 4.07644689e-01 5.74634373e-01 5.99632204e-01
1.08718067e-01 -5.98523438e-01 8.38631153e-01 1.37981237e-03
-7.73227096e-01 -4.81924474e-01 -1.01444924e+00 -3.43569279e-01
1.21089709e+00 -3.63770962e-01 -3.52789015e-01 -2.94243395e-01
-1.26898944e+00 4.60301787e-01 2.14089975e-01 5.61331749e-01
5.31013846e-01 -1.27672422e+00 8.53720307e-02 4.24941778e-01
-1.47819325e-01 -5.26885867e-01 -1.10211194e-01 9.44587469e-01
-3.85904610e-02 8.92059565e-01 -6.21816397e-01 -3.09248000e-01
-1.38362598e+00 1.14806354e+00 2.53404528e-01 6.48695324e-03
-2.13561103e-01 5.22852540e-01 8.95282850e-02 2.01440260e-01
3.81928861e-01 -6.64016366e-01 -6.22027934e-01 -8.81749857e-03
6.02597833e-01 7.42957950e-01 9.64808688e-02 -6.59036040e-01
-4.82039392e-01 8.84019554e-01 -3.73018049e-02 -1.09327801e-01
8.74258339e-01 -5.87611973e-01 -1.00007161e-01 8.30806255e-01
4.50450301e-01 -1.65757760e-01 -9.71693695e-01 -7.50093162e-02
-4.88649905e-02 -2.19401240e-01 2.26009652e-01 -9.49919879e-01
-9.31357741e-01 5.85310578e-01 8.67200673e-01 3.06692898e-01
1.47035015e+00 -2.39060760e-01 4.33732152e-01 5.41003883e-01
6.03805304e-01 -1.36756051e+00 4.20992672e-01 1.03675091e+00
1.23971379e+00 -5.96437156e-01 1.33574428e-02 -2.44552523e-01
-5.55513203e-01 1.14550924e+00 1.09731948e+00 -3.32935721e-01
7.78020024e-01 3.96435469e-01 -4.52406138e-01 -1.37459010e-01
-1.03405464e+00 -4.57986802e-01 4.91906218e-02 7.35309660e-01
1.28067717e-01 -5.59320033e-01 -3.84081483e-01 5.60793579e-01
2.00217843e-01 2.32404932e-01 6.02520704e-01 1.57781184e+00
-1.34863269e+00 -8.19769382e-01 -8.06673288e-01 3.12763490e-02
-5.89137226e-02 2.96463281e-01 -2.68595785e-01 1.03846765e+00
2.38077626e-01 9.47791696e-01 6.20669365e-01 -2.15906486e-01
4.97065008e-01 2.45400384e-01 5.42016804e-01 -8.36663187e-01
-5.88610649e-01 1.01051129e-01 -2.89212018e-01 -3.81386399e-01
-1.95863768e-01 -3.23629111e-01 -1.78186715e+00 9.47937816e-02
-6.50322318e-01 9.19785425e-02 2.84590960e-01 7.51118004e-01
5.71656883e-01 1.10742331e+00 9.09886658e-02 -8.14826965e-01
-3.75798821e-01 -1.11184573e+00 -2.98822880e-01 -1.84433609e-01
6.03536129e-01 -1.31942630e+00 -2.25272477e-01 3.24859880e-02] | [4.658388614654541, 0.9706303477287292] |
e0d7d820-d568-429e-ba7b-caeba249eb61 | prediction-of-chronic-kidney-disease-a | null | null | https://ieeexplore.ieee.org/abstract/document/9333572 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9333572 | Prediction of Chronic Kidney Disease - A Machine Learning Perspective | Chronic Kidney Disease is one of the most critical illness nowadays and proper diagnosis is
required as soon as possible. Machine learning technique has become reliable for medical treatment. With
the help of a machine learning classifier algorithms, the doctor can detect the disease on time. For this
perspective, Chronic Kidney Disease prediction has been discussed in this article. Chronic Kidney Disease
dataset has been taken from the UCI repository. Seven classifier algorithms have been applied in this research
such as artificial neural network, C5.0, Chi-square Automatic interaction detector, logistic regression, linear
support vector machine with penalty L1 & with penalty L2 and random tree. The important feature selection
technique was also applied to the dataset. For each classifier, the results have been computed based on
(i) full features, (ii) correlation-based feature selection, (iii) Wrapper method feature selection, (iv) Least
absolute shrinkage and selection operator regression, (v) synthetic minority over-sampling technique with
least absolute shrinkage and selection operator regression selected features, (vi) synthetic minority oversampling technique with full features. From the results, it is marked that LSVM with penalty L2 is giving
the highest accuracy of 98.86% in synthetic minority over-sampling technique with full features. Along
with accuracy, precision, recall, F-measure, area under the curve and GINI coefficient have been computed
and compared results of various algorithms have been shown in the graph. Least absolute shrinkage and
selection operator regression selected features with synthetic minority over-sampling technique gave the best
after synthetic minority over-sampling technique with full features. In the synthetic minority over-sampling
technique with least absolute shrinkage and selection operator selected features, again linear support vector
machine gave the highest accuracy of 98.46%. Along with machine learning models one deep neural network
has been applied on the same dataset and it has been noted that deep neural network achieved the highest
accuracy of 99.6%. | ['Vadim Bolshev', 'Elżbieta Jasińska', 'Radomir Gono', 'Łukasz Jasiński', 'Michał Jasiński', 'Zbigniew Leonowicz', 'Tulika Chakrabarti', 'Gaurav Kumawat', 'Prasun Chakrabarti', 'Sandeep Chaurasia', 'Pankaj Chittora'] | 2021-01-22 | null | null | null | journal-2021-1 | ['disease-prediction'] | ['medical'] | [ 1.08658053e-01 -1.66205782e-02 -2.54766464e-01 -4.86443877e-01
-2.53487349e-01 6.53637620e-03 3.58306020e-01 7.84049094e-01
-4.86452401e-01 1.04747379e+00 5.02204814e-04 -2.63774127e-01
-7.38218188e-01 -8.69460881e-01 2.34710542e-03 -7.05214560e-01
-5.11580944e-01 6.38096392e-01 -3.15964192e-01 1.92669883e-01
4.60584819e-01 6.14750028e-01 -1.44864416e+00 1.86051428e-01
9.57298577e-01 1.01154208e+00 -9.32868719e-02 7.21375942e-01
4.99603688e-04 8.10615599e-01 -3.31130117e-01 3.63671780e-01
5.45229554e-01 -3.96876991e-01 -4.77014244e-01 -4.19461459e-01
1.30465239e-01 -5.82379997e-02 6.69298992e-02 6.09820247e-01
7.46559203e-01 2.57634409e-02 1.17306721e+00 -1.23513258e+00
-4.00913268e-01 3.82656127e-01 -5.95744610e-01 2.54283339e-01
3.39540392e-01 3.68516590e-03 1.83013484e-01 -8.77106726e-01
5.08495688e-01 1.09905326e+00 8.02370846e-01 3.13052744e-01
-1.09189916e+00 -7.91571677e-01 -6.73409820e-01 9.16349217e-02
-1.36284542e+00 1.92051113e-01 4.51967031e-01 -6.10531509e-01
1.31350553e+00 5.37948906e-01 6.10245049e-01 9.74817723e-02
7.63041198e-01 3.55273932e-01 1.53387082e+00 -8.11958134e-01
1.51157618e-01 6.78074121e-01 7.22568512e-01 6.89986706e-01
4.95928556e-01 3.71628255e-01 9.42911357e-02 -3.60626310e-01
5.61918795e-01 5.75479448e-01 -4.65329811e-02 1.78028718e-01
-1.10101855e+00 1.02468288e+00 3.86930704e-01 5.66570878e-01
-8.00005734e-01 -1.01391040e-01 6.57050729e-01 6.76749110e-01
3.77265066e-02 3.44743013e-01 -1.04372311e+00 2.84847558e-01
-8.05384338e-01 3.04266572e-01 9.55966949e-01 6.39181495e-01
4.59901333e-01 6.11063913e-02 -3.42534035e-01 7.60323942e-01
1.84449747e-01 4.51734722e-01 7.19660938e-01 -3.86475593e-01
-3.06441467e-02 1.12374818e+00 -1.80948243e-01 -7.61085153e-01
-6.68950379e-01 -4.95835602e-01 -1.25741374e+00 6.22965217e-01
2.33934313e-01 -5.77299297e-01 -1.41619492e+00 1.13324809e+00
2.40540892e-01 -2.08638489e-01 4.31156456e-01 3.58717203e-01
1.00165391e+00 3.74556303e-01 2.65654534e-01 -4.66103315e-01
1.58657837e+00 -4.67440814e-01 -6.55584276e-01 5.58236420e-01
5.75877786e-01 -7.24857152e-01 2.66345829e-01 6.26106679e-01
-4.54631984e-01 -3.96255165e-01 -9.69130516e-01 3.94960314e-01
-6.18282676e-01 3.15002054e-01 5.98183215e-01 7.22652912e-01
-7.91316509e-01 7.29859650e-01 -6.39654815e-01 -7.57797241e-01
5.03776193e-01 7.65872359e-01 -6.56735659e-01 -8.05881768e-02
-1.15970969e+00 1.30207300e+00 4.04641122e-01 -2.04002738e-01
-1.33970484e-01 -7.33334482e-01 -5.55146873e-01 -1.12051256e-01
-3.32617253e-01 -8.40444446e-01 5.33504844e-01 -8.66633832e-01
-9.00785446e-01 6.87902093e-01 3.15662250e-02 -6.72278821e-01
5.90662479e-01 8.91636238e-02 -3.35727513e-01 -3.42744850e-02
4.65231501e-02 2.68427163e-01 2.46892095e-01 -6.81798279e-01
-8.98334086e-01 -8.12197506e-01 -7.65914679e-01 1.68820769e-01
1.86855912e-01 1.51365966e-01 6.57902956e-01 -4.41158235e-01
2.31280625e-01 -8.28873575e-01 -3.29914242e-01 -2.14589685e-01
-4.49953973e-01 -5.26973188e-01 1.11314738e+00 -7.63819575e-01
1.31206846e+00 -1.71287358e+00 -4.07285273e-01 6.23000622e-01
5.68771176e-02 2.40648299e-01 7.08231628e-01 4.82017517e-01
-4.25031602e-01 9.65759754e-02 -1.65071338e-01 4.96613503e-01
-3.75988454e-01 8.79767388e-02 4.63620394e-01 4.68373477e-01
8.15625265e-02 3.61160040e-01 -2.78041512e-01 -5.31631529e-01
5.59499085e-01 8.27123046e-01 -5.23555726e-02 -1.78722009e-01
3.67384434e-01 4.51882333e-02 -4.38659906e-01 9.63338614e-01
7.24908829e-01 1.84008740e-02 -1.07790172e-01 -2.50844181e-01
-1.05020911e-01 -4.44834977e-01 -1.25814593e+00 8.39072049e-01
-2.96599030e-01 5.78500867e-01 -3.25481564e-01 -1.35061789e+00
1.49789762e+00 7.42084086e-01 6.23895645e-01 -3.92450958e-01
2.96660990e-01 5.45990884e-01 2.56102949e-01 -8.81725550e-01
-3.99561971e-01 -2.89235890e-01 2.89338022e-01 2.36173645e-02
-2.16456681e-01 3.32088321e-01 1.27556935e-01 -1.11513659e-01
1.13369524e+00 -3.05412292e-01 1.06445134e+00 -4.51373547e-01
7.74679840e-01 2.09142253e-01 2.60939986e-01 5.98160505e-01
-4.71760422e-01 5.02529800e-01 1.45267561e-01 -8.55815053e-01
-8.86311889e-01 -4.62730557e-01 -8.05857122e-01 4.78056997e-01
-4.15618569e-01 4.66230273e-01 -3.17872167e-01 -6.62937045e-01
6.67288423e-01 5.45386553e-01 -9.50655580e-01 1.76158413e-01
-3.82426083e-01 -9.24562156e-01 4.58204597e-01 4.20427024e-01
6.15841150e-01 -1.30797172e+00 -8.72629523e-01 1.88310489e-01
7.38491714e-01 -1.61609545e-01 4.02666926e-01 6.62572384e-01
-1.23612034e+00 -1.35713410e+00 -6.90550804e-01 -9.25279975e-01
6.98721945e-01 -3.61362755e-01 7.22850442e-01 3.48950714e-01
-9.06268477e-01 -2.68458813e-01 -1.82585284e-01 -8.72706413e-01
-1.87844276e-01 -3.02406639e-01 -3.11005451e-02 -5.86682916e-01
1.06422675e+00 -6.03300154e-01 -8.21358860e-01 -3.71639431e-01
-4.08292264e-01 -4.14965391e-01 9.40625131e-01 1.01149452e+00
2.43062511e-01 2.54709989e-01 1.11148453e+00 -1.11657000e+00
5.79869866e-01 -6.36292875e-01 -1.41135454e-01 1.26491472e-01
-1.20186448e+00 -1.34566138e-02 4.84562308e-01 -1.24370262e-01
-5.00311017e-01 2.53198624e-01 2.17631385e-01 -1.45074457e-01
-3.63020599e-01 6.37141466e-01 3.96787167e-01 -1.94795690e-02
1.08979809e+00 1.42752826e-01 3.08430463e-01 -4.37940061e-01
-3.16020072e-01 1.04294884e+00 -9.96356234e-02 1.12912416e-01
2.01695710e-01 7.44604766e-02 3.88269454e-01 -6.56127632e-01
-3.95786799e-02 -4.99945343e-01 -4.99310285e-01 -5.23380563e-02
7.39886761e-01 -7.53284037e-01 -6.77066624e-01 5.37153125e-01
-5.95011473e-01 2.52253920e-01 1.24288071e-02 9.84209061e-01
-1.65641651e-01 -1.00742981e-01 -2.42254719e-01 -1.29306769e+00
-1.30469871e+00 -8.75509918e-01 1.64114892e-01 5.40851474e-01
-6.22020900e-01 -8.77574384e-01 -2.54469603e-01 9.19517651e-02
6.39685631e-01 1.03322160e+00 1.20324600e+00 -1.40702081e+00
1.99772343e-01 -9.07899320e-01 -3.13572794e-01 3.70541573e-01
5.42293429e-01 1.98993072e-01 -5.48833370e-01 -2.34036539e-02
-2.09881328e-02 7.51113966e-02 6.82299435e-01 8.00614595e-01
7.65810013e-01 -2.55229145e-01 -6.28383875e-01 2.71891534e-01
1.99456263e+00 8.81793857e-01 3.37774754e-01 5.76334715e-01
1.30746618e-01 3.27631831e-01 5.35423577e-01 5.48786700e-01
-4.77664312e-03 -2.72324570e-02 2.20275894e-02 -2.60798875e-02
2.82932483e-02 4.34401512e-01 -2.74053991e-01 1.79203883e-01
-2.30313197e-01 3.12359750e-01 -1.09896016e+00 5.08082986e-01
-1.49129868e+00 -9.31782782e-01 -7.77993858e-01 2.36099601e+00
5.92063725e-01 1.79853861e-03 1.03976831e-01 7.83044994e-01
8.70651484e-01 -7.09476531e-01 -4.07560587e-01 -7.56215692e-01
1.16011545e-01 5.68056464e-01 6.63294792e-01 4.83624309e-01
-1.15304291e+00 1.50956726e-02 5.00297022e+00 3.14253330e-01
-1.26939762e+00 -2.26405442e-01 7.14925587e-01 7.56322071e-02
2.95569569e-01 -7.40960240e-02 -5.71004391e-01 7.59281933e-01
6.86300933e-01 -2.87321001e-01 5.92026003e-02 1.01840222e+00
2.46111900e-01 -6.27138078e-01 -9.77544308e-01 7.14872539e-01
-2.61124253e-01 -8.33815515e-01 -1.56162098e-01 -2.89273392e-02
6.80189431e-01 1.59191445e-01 -4.46314812e-01 4.44824368e-01
-2.71968935e-02 -1.27723193e+00 -2.72991329e-01 7.03508735e-01
6.44297004e-01 -8.98508966e-01 1.28693485e+00 3.99555832e-01
-9.94857550e-01 -3.29053432e-01 4.29446669e-03 -2.59949118e-01
-3.17266315e-01 1.03782833e+00 -1.13063681e+00 5.36330700e-01
6.00703776e-01 3.28923732e-01 -1.77933246e-01 1.21130764e+00
3.38162899e-01 3.25722575e-01 -7.45779216e-01 -3.41577828e-01
-1.08472817e-01 -1.35260135e-01 4.17066902e-01 1.36341989e+00
5.13706923e-01 2.52349645e-01 8.47314894e-02 3.42723697e-01
5.65957606e-01 6.27827406e-01 -7.98185050e-01 3.78411651e-01
6.12039506e-01 8.96477520e-01 -6.92718327e-01 -6.83132052e-01
-1.86562136e-01 3.88467252e-01 -2.45702535e-01 1.00272205e-02
-2.53388762e-01 -9.90101576e-01 1.15135372e-01 2.94940978e-01
-1.27797887e-01 5.84054589e-01 -8.54011357e-01 -3.90904099e-01
-2.80113697e-01 -7.10665703e-01 7.37807512e-01 -3.36801410e-01
-1.18962467e+00 5.82536757e-01 -2.25348491e-02 -9.48015571e-01
-1.43028334e-01 -5.15944421e-01 -6.41812682e-01 1.41092885e+00
-1.02507710e+00 -9.04590845e-01 -4.38280493e-01 4.55732673e-01
5.34630001e-01 -5.24900854e-01 1.14763629e+00 5.08830249e-01
-4.82134998e-01 3.35075408e-01 3.36719543e-01 1.26530603e-01
5.42144656e-01 -1.40356076e+00 -5.27494550e-01 2.13747367e-01
-8.81816149e-01 7.14542210e-01 8.18140507e-01 -9.64492381e-01
-1.01098180e+00 -8.99034262e-01 1.17971909e+00 5.21720126e-02
-1.29323944e-01 5.05781054e-01 -8.70646894e-01 4.46314991e-01
2.09976569e-01 2.74555176e-01 7.16543555e-01 -2.32090741e-01
3.63251001e-01 -1.65959477e-01 -2.06628442e+00 -1.52431414e-01
7.88270831e-02 9.51502919e-02 -4.79474962e-01 5.87397575e-01
-8.76720101e-02 -2.80415535e-01 -1.28654957e+00 7.93999374e-01
1.00637949e+00 -9.62254822e-01 7.87906170e-01 -5.48734665e-01
4.44069654e-01 -3.49506408e-01 -4.26327810e-02 -8.37978542e-01
-2.78214633e-01 8.66362751e-02 3.91469151e-02 1.00937796e+00
6.80921316e-01 -9.46995080e-01 8.34123909e-01 9.34415519e-01
4.08823490e-01 -1.53347290e+00 -6.15284920e-01 -4.98607099e-01
1.81048026e-03 1.91235811e-01 8.95413682e-02 1.32542455e+00
2.76603159e-02 -1.26422876e-02 -7.44827241e-02 -2.74356157e-01
5.91856897e-01 -1.59177706e-01 6.01658821e-01 -1.67181396e+00
5.29255830e-02 -1.41594484e-01 -8.24115157e-01 1.92094088e-01
-6.51431382e-01 -8.27840388e-01 -4.58195597e-01 -1.95081258e+00
4.02767450e-01 -6.52912199e-01 -5.33622921e-01 6.65843844e-01
-2.51411796e-01 -3.73562761e-02 -1.73539683e-01 8.56012926e-02
6.42244101e-01 -3.60880435e-01 9.44638908e-01 1.70178622e-01
-6.01961493e-01 3.32446724e-01 -5.84491670e-01 5.39838314e-01
1.04592264e+00 -6.73280835e-01 -2.45409340e-01 3.27134669e-01
-1.61360607e-01 1.69220507e-01 -8.76684338e-02 -9.67112303e-01
1.50677100e-01 -1.69992909e-01 1.10079360e+00 -6.94365025e-01
-1.15577601e-01 -1.05378652e+00 5.12298286e-01 1.26173484e+00
-2.43225560e-01 7.74561465e-02 -3.11052315e-02 3.03194046e-01
-8.19447786e-02 -4.18104351e-01 9.02611434e-01 -2.53545672e-01
-4.34935033e-01 -3.00746597e-02 -2.80162275e-01 -4.11249876e-01
1.47736239e+00 -7.45257914e-01 9.93011426e-03 -8.01207721e-02
-1.09188557e+00 3.52326453e-01 -1.33280484e-02 -8.40540081e-02
4.32116717e-01 -1.18833697e+00 -9.57291424e-01 2.34603390e-01
-1.05464980e-02 -1.12860762e-01 8.00925791e-02 1.14968097e+00
-1.03773081e+00 4.33056056e-01 -3.74338537e-01 -5.02465308e-01
-1.91043723e+00 3.78818244e-01 3.35257053e-01 -2.54204392e-01
-7.43069828e-01 7.99708962e-01 -6.26906276e-01 -3.53744417e-01
1.19540349e-01 -4.63519275e-01 -6.04833245e-01 8.55381936e-02
3.12507540e-01 8.28262568e-01 1.26207009e-01 -3.17546517e-01
-6.58101082e-01 7.68781722e-01 -1.94571421e-01 2.53230512e-01
1.50830019e+00 3.29025477e-01 -3.34786832e-01 2.21270233e-01
1.25249326e+00 -1.78381488e-01 -2.03770533e-01 5.70796169e-02
1.34230584e-01 -3.59626085e-01 9.70835537e-02 -1.45462918e+00
-8.67366493e-01 5.12142420e-01 1.18936145e+00 3.87601316e-01
1.26996613e+00 -7.09760308e-01 4.55085307e-01 3.26788515e-01
9.73876752e-03 -1.02871692e+00 -8.65599811e-01 1.70446679e-01
7.99517453e-01 -1.59661543e+00 5.63461721e-01 -1.77783623e-01
-6.29600942e-01 1.42970681e+00 4.24103916e-01 -7.46135652e-01
1.31621408e+00 3.84628266e-01 2.08428502e-01 -3.09038252e-01
-8.84691060e-01 3.22992593e-01 -9.41267796e-03 6.06756091e-01
9.09732580e-01 3.57170373e-01 -1.21000385e+00 5.24051368e-01
-5.63864671e-02 6.57847285e-01 1.93116456e-01 1.01070762e+00
-6.46991730e-01 -7.36067533e-01 -4.75590080e-01 1.31174529e+00
-8.33926320e-01 1.01869330e-02 -2.79631406e-01 1.23001540e+00
4.05537277e-01 7.10941136e-01 -4.59849387e-02 -1.59603298e-01
3.12428385e-01 3.97999883e-01 3.33642870e-01 -3.72101307e-01
-1.18382668e+00 -3.26124787e-01 1.24957979e-01 6.88694194e-02
-2.54304290e-01 -6.48266912e-01 -1.71856046e+00 -2.25662336e-01
-5.22428155e-01 2.58660793e-01 1.17130518e+00 9.38352525e-01
1.63346961e-01 4.48392004e-01 4.32004988e-01 -3.85266632e-01
-5.88255703e-01 -1.53189683e+00 -5.13261795e-01 3.41377109e-01
4.04486477e-01 -4.99855757e-01 -4.50290173e-01 -1.82771906e-01] | [8.402148246765137, 4.829646110534668] |
fd0d88ba-2477-40b7-b3a3-1140123a80d1 | chip-channel-independence-based-pruning-for | 2110.13981 | null | https://arxiv.org/abs/2110.13981v3 | https://arxiv.org/pdf/2110.13981v3.pdf | CHIP: CHannel Independence-based Pruning for Compact Neural Networks | Filter pruning has been widely used for neural network compression because of its enabled practical acceleration. To date, most of the existing filter pruning works explore the importance of filters via using intra-channel information. In this paper, starting from an inter-channel perspective, we propose to perform efficient filter pruning using Channel Independence, a metric that measures the correlations among different feature maps. The less independent feature map is interpreted as containing less useful information$/$knowledge, and hence its corresponding filter can be pruned without affecting model capacity. We systematically investigate the quantification metric, measuring scheme and sensitiveness$/$reliability of channel independence in the context of filter pruning. Our evaluation results for different models on various datasets show the superior performance of our approach. Notably, on CIFAR-10 dataset our solution can bring $0.90\%$ and $0.94\%$ accuracy increase over baseline ResNet-56 and ResNet-110 models, respectively, and meanwhile the model size and FLOPs are reduced by $42.8\%$ and $47.4\%$ (for ResNet-56) and $48.3\%$ and $52.1\%$ (for ResNet-110), respectively. On ImageNet dataset, our approach can achieve $40.8\%$ and $44.8\%$ storage and computation reductions, respectively, with $0.15\%$ accuracy increase over the baseline ResNet-50 model. The code is available at https://github.com/Eclipsess/CHIP_NeurIPS2021. | ['Bo Yuan', 'Saman Zonouz', 'Huy Phan', 'Yi Xie', 'Miao Yin', 'Yang Sui'] | 2021-10-26 | null | http://proceedings.neurips.cc/paper/2021/hash/ce6babd060aa46c61a5777902cca78af-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/ce6babd060aa46c61a5777902cca78af-Paper.pdf | neurips-2021-12 | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 1.51530117e-01 -2.77076978e-02 3.00913379e-02 -1.98016807e-01
-2.41035059e-01 -1.56904891e-01 -6.91402555e-02 2.13869408e-01
-8.46649528e-01 1.02311623e+00 -2.65999436e-01 -3.20985883e-01
-2.61844993e-01 -1.01322937e+00 -8.62340510e-01 -4.69351679e-01
-3.01045239e-01 -2.58137375e-01 2.54572421e-01 -1.21479206e-01
3.29707675e-02 2.75483847e-01 -1.55041528e+00 1.15909263e-01
6.25793815e-01 1.57806003e+00 2.71619946e-01 2.81823426e-01
2.18665674e-01 4.82674927e-01 -6.81670368e-01 -5.77146471e-01
4.86616313e-01 -2.54701018e-01 -3.86939079e-01 -4.33306962e-01
3.32038194e-01 -3.80655795e-01 -7.11323023e-01 1.42985928e+00
5.07486343e-01 -4.95961867e-02 1.86030537e-01 -9.49872136e-01
-1.20119721e-01 1.01582515e+00 -4.85390306e-01 5.68266690e-01
-3.05499285e-01 -6.43645898e-02 1.00239909e+00 -5.94374299e-01
3.44892889e-01 8.75627220e-01 5.19120693e-01 4.50698286e-01
-1.42197883e+00 -1.41590536e+00 2.97359794e-01 2.76747674e-01
-1.66529214e+00 -5.55739820e-01 4.98860598e-01 -7.56685883e-02
1.05521703e+00 2.07780197e-01 8.03097606e-01 4.01882857e-01
-6.74590142e-03 7.13377416e-01 9.50389445e-01 -2.26521015e-01
2.20864460e-01 4.45972569e-02 4.25115734e-01 8.20480645e-01
4.40089971e-01 1.47581354e-01 -6.26787663e-01 1.18192688e-01
7.61701763e-01 3.09542306e-02 -4.88243878e-01 4.60489869e-01
-5.41815519e-01 6.39446020e-01 5.63129604e-01 5.22332862e-02
-6.45357892e-02 4.72340435e-01 4.65111136e-01 5.23218930e-01
2.11766750e-01 1.49837807e-01 -5.98743498e-01 -1.14970393e-01
-1.04344976e+00 3.03249300e-01 4.56850886e-01 1.11520123e+00
8.10384870e-01 2.22602636e-01 -2.33162921e-02 8.92078161e-01
2.10540861e-01 5.14408290e-01 2.63553977e-01 -7.93983757e-01
5.19359648e-01 6.35274649e-01 -3.57137442e-01 -9.88746464e-01
-4.55399394e-01 -8.75283062e-01 -1.24364877e+00 -1.87426116e-02
2.07501471e-01 -1.61807075e-01 -8.55029643e-01 2.06273603e+00
-3.13135982e-01 1.20245054e-01 -8.89762267e-02 4.30153280e-01
7.41703868e-01 4.65965807e-01 4.61188890e-02 -4.14279759e-01
1.31953120e+00 -5.25111556e-01 -5.74621439e-01 -3.11515361e-01
5.18313766e-01 -8.32898974e-01 7.46278107e-01 5.35463691e-01
-1.42596328e+00 -6.97034001e-01 -1.34854448e+00 2.72830844e-01
-1.90664247e-01 3.24583888e-01 5.95274270e-01 8.67545605e-01
-1.02812159e+00 7.17696130e-01 -7.94383943e-01 1.36477053e-01
8.87403131e-01 7.75677085e-01 7.79499486e-02 -1.24690041e-01
-1.11794400e+00 4.16221499e-01 4.23853964e-01 -2.44491130e-01
-6.71720743e-01 -9.77857471e-01 -4.18500185e-01 4.33965266e-01
1.70941830e-01 -1.88909903e-01 9.97862279e-01 -5.69424748e-01
-9.90609527e-01 2.46372953e-01 -1.78272173e-01 -1.05775893e+00
8.89377668e-02 -2.01257452e-01 -6.12395823e-01 2.00884894e-01
-3.59596848e-01 8.29680741e-01 5.03837943e-01 -6.74594045e-01
-9.41008985e-01 -3.23386401e-01 1.38089895e-01 -1.20490491e-01
-6.41202390e-01 -6.91345036e-02 -5.25094569e-01 -8.87787402e-01
2.45502114e-01 -8.01548004e-01 -2.29940295e-01 -3.67650688e-02
-3.38993609e-01 1.97039172e-01 4.90029067e-01 -5.77874720e-01
1.75208449e+00 -2.08034587e+00 -3.86687815e-01 4.83621299e-01
5.56279838e-01 7.09043622e-01 2.01436575e-03 -5.75360060e-02
4.98763844e-02 2.14339927e-01 -1.84199288e-01 5.11483476e-02
-3.04526925e-01 -1.32691041e-01 -2.19092324e-01 2.83020884e-01
-2.27317493e-02 5.68340898e-01 -2.60941774e-01 -9.75080803e-02
6.99725896e-02 5.66363573e-01 -7.75917768e-01 -3.58295441e-01
8.23630095e-02 -1.57931119e-01 -3.07978690e-01 4.58407104e-01
1.02067494e+00 -2.19794571e-01 2.75792807e-01 -4.20013905e-01
-2.42282394e-02 2.60191947e-01 -1.16403198e+00 1.52951288e+00
-3.83836001e-01 7.78261065e-01 1.44792780e-01 -1.05246770e+00
1.03271854e+00 1.85094029e-01 2.81104892e-01 -1.18618166e+00
4.70168799e-01 2.22666100e-01 4.64334249e-01 2.40749508e-01
3.27133060e-01 1.47547737e-01 9.33620036e-02 2.61469066e-01
1.54185504e-01 3.77375305e-01 4.87378657e-01 3.82245600e-01
1.52726829e+00 -5.68425894e-01 -2.88209673e-02 -6.15093291e-01
5.09106100e-01 -4.20280814e-01 6.54825091e-01 6.12238705e-01
-1.73517391e-01 1.86206147e-01 4.33187127e-01 -3.18451673e-01
-5.78213036e-01 -9.65846896e-01 -4.30068046e-01 7.00634062e-01
2.62835175e-01 -7.94771969e-01 -8.30022037e-01 -2.16454044e-01
2.32432261e-02 4.87645090e-01 -2.25795612e-01 -2.51939178e-01
-6.31402135e-01 -9.11922097e-01 7.63481259e-01 5.77785134e-01
1.05417275e+00 -8.05402815e-01 -9.89284992e-01 3.11021775e-01
1.61327973e-01 -9.61649597e-01 1.95025206e-02 4.07230854e-01
-1.10965228e+00 -9.45407510e-01 -4.38190788e-01 -4.94314671e-01
6.62234247e-01 2.12243110e-01 8.57816339e-01 4.08314139e-01
-4.73874718e-01 -2.40636885e-01 -2.92169392e-01 -4.17560577e-01
1.87278777e-01 5.71345650e-02 3.81195769e-02 -3.18918645e-01
3.53850722e-01 -8.26961815e-01 -9.22717571e-01 3.59400779e-01
-7.80777752e-01 6.03810418e-03 6.45030439e-01 7.71487892e-01
8.79962325e-01 4.05943483e-01 6.69683993e-01 -1.03919768e+00
5.54429471e-01 -1.44296527e-01 -7.71803617e-01 -6.49086460e-02
-9.50053990e-01 -5.23269596e-03 9.00817871e-01 -1.79181457e-01
-5.80121875e-01 -1.23265032e-02 -3.34655017e-01 -3.76052678e-01
1.22923076e-01 4.40569103e-01 -1.65992394e-01 -2.20627740e-01
6.33114517e-01 2.77809083e-01 -2.79333442e-01 -5.74319839e-01
-1.06961187e-02 2.59896010e-01 2.07857877e-01 -2.17747405e-01
5.88691890e-01 5.97404242e-01 1.27298580e-02 -7.28394568e-01
-2.68783510e-01 -7.97313824e-02 -9.46185067e-02 -4.04229835e-02
1.78727627e-01 -1.04676831e+00 -9.02216613e-01 2.90731579e-01
-8.15808773e-01 -2.30356485e-01 -1.22160375e-01 5.56646764e-01
-1.48732767e-01 9.05488059e-02 -5.66974938e-01 -5.35186410e-01
-7.48576880e-01 -1.30148923e+00 2.54676580e-01 3.39349270e-01
-1.06685355e-01 -3.99153531e-01 -6.47260725e-01 -1.04290955e-01
6.45900488e-01 -8.86792094e-02 1.02857065e+00 -4.02865440e-01
-6.62524462e-01 -2.59738654e-01 -7.87496090e-01 5.57708859e-01
-2.19617695e-01 -3.88193309e-01 -9.82446313e-01 -4.47812319e-01
-1.80163696e-01 -8.93148854e-02 1.32435334e+00 5.30912340e-01
1.60175776e+00 -3.12121809e-01 -4.65927213e-01 7.85189390e-01
1.64490807e+00 6.31937385e-01 7.54612744e-01 -9.03354511e-02
2.11906731e-01 2.20494531e-02 2.13923603e-01 6.90179229e-01
-1.03939958e-01 4.66685176e-01 4.10242677e-01 1.54926360e-01
-3.54858249e-01 -1.44824654e-01 1.88304088e-03 8.36413562e-01
-5.41489869e-02 -2.37380043e-01 -6.37402475e-01 3.36469829e-01
-1.52339292e+00 -7.12282538e-01 4.94122803e-02 2.14665627e+00
8.48852217e-01 6.67315722e-01 -1.07331537e-01 6.75206840e-01
5.65161169e-01 -2.23396048e-02 -5.94359994e-01 -2.11153895e-01
-1.26590684e-01 8.63564372e-01 8.54504704e-01 2.15306997e-01
-9.18198168e-01 7.54140198e-01 4.59881783e+00 1.33861172e+00
-1.18034112e+00 1.50234073e-01 9.20380294e-01 -6.15772903e-01
7.04366565e-02 -1.03658624e-01 -1.15986574e+00 6.42792642e-01
9.95547593e-01 -2.31985196e-01 5.18474996e-01 5.66765189e-01
5.34110842e-03 -2.96824753e-01 -9.30577576e-01 1.34467959e+00
-2.43226379e-01 -1.64796531e+00 -1.10406086e-01 7.35588819e-02
5.62244117e-01 1.26120403e-01 1.17986023e-01 1.14191458e-01
6.34834319e-02 -1.04094672e+00 9.14764106e-01 2.98738867e-01
9.99730825e-01 -1.18426991e+00 5.43969691e-01 1.10787161e-01
-1.39635515e+00 -1.82362273e-01 -6.45639241e-01 -2.56532222e-01
4.55507077e-02 9.86562014e-01 -3.52371007e-01 1.75019607e-01
1.23408902e+00 5.77759445e-01 -4.72801894e-01 1.05583799e+00
-1.48255318e-01 6.99711204e-01 -5.47283649e-01 -1.47161737e-01
-1.64239943e-01 9.27190110e-02 2.63988256e-01 1.18968070e+00
3.07729751e-01 3.22311997e-01 -1.84588581e-01 7.20897615e-01
-4.85597849e-01 1.94168650e-02 -1.19772173e-01 2.89758027e-01
7.96973348e-01 9.58890378e-01 -8.49172771e-01 -3.41793716e-01
-2.80383646e-01 4.29946244e-01 2.82396823e-01 2.37469822e-01
-9.91754949e-01 -7.82953858e-01 7.26502478e-01 1.59318626e-01
3.47632855e-01 -2.90220678e-02 -5.93390584e-01 -8.08423460e-01
1.25247926e-01 -6.55915976e-01 2.56022900e-01 -1.34110242e-01
-5.66071332e-01 9.51068282e-01 -1.43010080e-01 -1.05717671e+00
3.83443683e-01 -5.18359900e-01 -2.54972875e-01 7.82815278e-01
-1.42623973e+00 -4.02947932e-01 -1.51545793e-01 5.44544816e-01
3.18620771e-01 -2.77378947e-01 6.66650772e-01 7.78287709e-01
-7.20480740e-01 1.30151474e+00 1.61599368e-01 3.69637124e-02
1.82625055e-01 -4.15614128e-01 2.97553539e-01 9.37210321e-01
1.97462849e-02 8.60848546e-01 4.65136379e-01 -4.30052876e-01
-1.28426981e+00 -1.19141984e+00 7.29739666e-01 4.62130845e-01
3.28668773e-01 -3.74013901e-01 -7.41908967e-01 2.95536995e-01
9.58172698e-03 -8.11014511e-03 5.16410172e-01 -7.20677003e-02
-3.90320987e-01 -6.90661430e-01 -1.32688713e+00 6.41026199e-01
1.38783133e+00 -1.44271180e-01 2.30262831e-01 2.16429234e-02
6.02955282e-01 -2.04827234e-01 -9.75500286e-01 4.91782546e-01
6.18894041e-01 -1.08527410e+00 9.35653985e-01 3.89654301e-02
2.58543134e-01 -1.12426609e-01 -4.79659796e-01 -7.75726080e-01
-4.28826004e-01 -4.75631773e-01 -2.06606403e-01 1.04745317e+00
7.79600739e-01 -7.14777887e-01 9.00821626e-01 4.17692125e-01
-2.08865851e-01 -1.01355255e+00 -1.15786970e+00 -7.71721721e-01
3.37000079e-02 -7.71023571e-01 6.28364503e-01 3.09856296e-01
-1.36249244e-01 1.14409983e-01 -9.84417126e-02 -8.44337940e-02
5.34172714e-01 -1.51636809e-01 2.89820433e-01 -1.08567119e+00
-2.80895978e-01 -6.69312656e-01 -3.82477939e-01 -1.18059182e+00
-2.05578357e-01 -7.94836402e-01 -4.68829244e-01 -9.95594442e-01
-2.60917167e-03 -7.93544412e-01 -6.05152190e-01 6.11781895e-01
2.95556545e-01 5.78828573e-01 3.98130000e-01 6.03644699e-02
-3.51479381e-01 3.45764875e-01 8.55264187e-01 -1.06930584e-01
-1.62230536e-01 -1.03376815e-02 -7.83954382e-01 7.19065607e-01
1.29743481e+00 -4.44971949e-01 -5.49964547e-01 -6.34454668e-01
2.99361676e-01 -8.01023394e-02 3.40368778e-01 -1.55594289e+00
3.02286237e-01 3.74361783e-01 5.20298600e-01 -4.66950685e-01
5.15317976e-01 -7.49587119e-01 1.30843997e-01 1.00521398e+00
-2.86126465e-01 3.30024399e-02 6.71160519e-01 5.65695643e-01
-2.09227085e-01 -7.28131607e-02 1.08743930e+00 -7.63040930e-02
-7.15449631e-01 3.07346731e-01 -1.71004683e-01 -2.57369727e-02
8.26297700e-01 -1.69158950e-01 -4.44420129e-01 -7.38687664e-02
-6.06260836e-01 3.34198326e-02 1.77917182e-02 6.11543581e-02
8.53649914e-01 -1.00870264e+00 -4.55357313e-01 5.16190529e-01
-1.95854560e-01 -3.21147382e-01 4.71162647e-01 7.10863471e-01
-4.83156174e-01 7.30838597e-01 -3.10118914e-01 -2.35471934e-01
-1.29828024e+00 2.18621999e-01 8.05924758e-02 -1.57334194e-01
-3.01359415e-01 1.40233016e+00 1.42492503e-01 3.80557269e-01
3.97857010e-01 -5.22945642e-01 -1.15328524e-02 -1.15937397e-01
6.44786656e-01 7.94628859e-01 4.30463165e-01 -3.83555502e-01
-3.64834249e-01 3.82962435e-01 -4.15750295e-01 1.53200375e-02
1.26231813e+00 2.55593285e-02 -1.13607354e-01 -3.86311054e-01
1.28891659e+00 -2.46319741e-01 -1.12116742e+00 -3.67734611e-01
-4.02214408e-01 -2.89705336e-01 3.31294447e-01 -9.84309971e-01
-1.74467576e+00 9.81780350e-01 9.94636357e-01 -1.98192131e-02
1.69614530e+00 -2.07773000e-01 9.07387197e-01 1.67869031e-01
6.36233807e-01 -1.10052764e+00 -4.08177227e-01 5.03313184e-01
4.95865852e-01 -6.49283648e-01 2.04699174e-01 -5.14561832e-01
-1.07540771e-01 8.39290500e-01 6.78501070e-01 -2.39133313e-01
1.11666930e+00 4.05015945e-01 -3.68251026e-01 -2.24074155e-01
-7.56238222e-01 -1.07885540e-01 4.07810360e-02 2.75856733e-01
2.63153315e-01 2.64011502e-01 -6.99472427e-01 9.92477715e-01
-4.41640317e-01 7.80251622e-02 1.44663066e-01 9.03694749e-01
-5.71643054e-01 -1.09485626e+00 1.38771668e-01 1.07286036e+00
-6.57778978e-01 -5.24545729e-01 5.52906580e-02 5.70477605e-01
4.22296941e-01 1.06473398e+00 1.82753533e-01 -8.50319743e-01
2.45538875e-01 -2.29186267e-01 4.71457571e-01 -3.06622624e-01
-7.01851726e-01 2.13281825e-01 -3.60294357e-02 -6.20436907e-01
-1.13340303e-01 -2.89084107e-01 -1.54852831e+00 -6.37151718e-01
-3.15517753e-01 2.29826067e-02 6.07629359e-01 5.07658184e-01
6.85581923e-01 6.51374400e-01 2.35493600e-01 -2.46373773e-01
-1.79355875e-01 -6.01903796e-01 -7.29467452e-01 7.25087337e-03
-9.11332816e-02 -5.84820390e-01 -1.90369383e-01 -1.56765863e-01] | [8.545400619506836, 3.007338762283325] |
60666bc6-ace6-4b76-b23c-0f70f3c99fc8 | multi-task-learning-and-adapted-knowledge | 2106.09790 | null | https://arxiv.org/abs/2106.09790v1 | https://arxiv.org/pdf/2106.09790v1.pdf | Multi-Task Learning and Adapted Knowledge Models for Emotion-Cause Extraction | Detecting what emotions are expressed in text is a well-studied problem in natural language processing. However, research on finer grained emotion analysis such as what causes an emotion is still in its infancy. We present solutions that tackle both emotion recognition and emotion cause detection in a joint fashion. Considering that common-sense knowledge plays an important role in understanding implicitly expressed emotions and the reasons for those emotions, we propose novel methods that combine common-sense knowledge via adapted knowledge models with multi-task learning to perform joint emotion classification and emotion cause tagging. We show performance improvement on both tasks when including common-sense reasoning and a multitask framework. We provide a thorough analysis to gain insights into model performance. | ['Smaranda Muresan', 'Yaser Al-Onaizan', 'Kasturi Bhattacharjee', 'Rishita Anubhai', 'Shuai Wang', 'Elsbeth Turcan'] | 2021-06-17 | null | https://aclanthology.org/2021.findings-acl.348 | https://aclanthology.org/2021.findings-acl.348.pdf | findings-acl-2021-8 | ['emotion-cause-extraction'] | ['natural-language-processing'] | [ 2.37380430e-01 -9.12649557e-02 -3.67737293e-01 -7.03413188e-01
-8.47734928e-01 -6.39144003e-01 4.84579355e-01 6.49309278e-01
-5.01587868e-01 6.10521197e-01 6.00787163e-01 1.16607226e-01
-7.60536194e-02 -4.22570854e-01 -3.43549341e-01 -4.06846732e-01
-1.77117751e-03 1.68149486e-01 -2.49859512e-01 -2.38654330e-01
8.40685442e-02 1.31518826e-01 -1.64178157e+00 8.31170440e-01
7.96888530e-01 9.83359873e-01 -4.00579721e-01 7.83035576e-01
-5.79270065e-01 1.35541785e+00 -6.64777637e-01 -5.91322958e-01
-5.20626307e-01 -4.09917682e-01 -1.15093482e+00 -4.94412612e-03
-8.18951987e-03 4.27664071e-01 6.52703524e-01 8.51697862e-01
2.33957857e-01 2.85223126e-01 5.93230188e-01 -1.43230057e+00
-1.52345896e-01 6.13582075e-01 -4.01145339e-01 5.99154457e-02
6.59584820e-01 -4.35489953e-01 1.44616580e+00 -5.53534448e-01
4.33349013e-01 1.41134834e+00 6.26525462e-01 4.66174096e-01
-9.11710858e-01 -6.22893393e-01 4.50298578e-01 3.75242174e-01
-1.01891911e+00 -3.03356975e-01 9.37165618e-01 -3.65497470e-01
1.40506613e+00 2.49981850e-01 4.24119115e-01 9.73339379e-01
2.97275037e-01 1.10278034e+00 1.37662768e+00 -5.64241707e-01
2.82021165e-01 2.78688788e-01 5.58845818e-01 5.39768755e-01
-2.29534149e-01 -5.86768985e-01 -7.76920974e-01 -5.00351131e-01
3.00356120e-01 -9.40521806e-02 6.70058206e-02 2.84399502e-02
-9.47913826e-01 1.00337374e+00 8.57688934e-02 6.96622491e-01
-7.85265803e-01 2.51711607e-01 9.10557747e-01 5.25891244e-01
9.53732073e-01 6.94192410e-01 -1.12623405e+00 -5.39678514e-01
-7.05767751e-01 2.24271029e-01 1.22101021e+00 3.69403750e-01
9.88132179e-01 -3.34493905e-01 -3.18268016e-02 1.20802760e+00
2.21681017e-02 2.39548713e-01 3.35080475e-01 -7.20770776e-01
-2.01151684e-01 8.51317048e-01 1.64583594e-01 -1.19383955e+00
-6.99966252e-01 -3.30795765e-01 -5.20631254e-01 -2.67273694e-01
3.38638961e-01 -5.41409194e-01 -6.83773220e-01 1.93356586e+00
3.34643275e-01 3.61897945e-01 3.52385819e-01 8.57536614e-01
8.02303016e-01 5.90184867e-01 6.62408471e-01 -3.85829389e-01
2.08619380e+00 -7.66699791e-01 -1.17046833e+00 -4.87236023e-01
9.59594727e-01 -8.02206874e-01 8.84491801e-01 4.08248961e-01
-5.53619862e-01 -6.69616386e-02 -4.85476911e-01 -1.65499389e-01
-7.45182872e-01 2.04235792e-01 1.18291032e+00 4.73724484e-01
-5.81088364e-01 -1.31073520e-01 -3.63546550e-01 -6.37360096e-01
3.20962220e-02 9.44714919e-02 -1.41768724e-01 8.82474631e-02
-1.66910243e+00 1.09861100e+00 2.82371819e-01 -2.71850944e-01
-3.17948818e-01 -7.53661692e-01 -1.01645768e+00 -2.54996996e-02
8.42944801e-01 -5.09453654e-01 1.32724464e+00 -1.37288809e+00
-1.48437083e+00 1.05661464e+00 -5.87456942e-01 -2.00495392e-01
-3.83908302e-01 -6.91392660e-01 -6.72829151e-01 -3.32457689e-03
-3.74866948e-02 2.76878029e-01 6.89430475e-01 -1.07388222e+00
-6.70613408e-01 -3.60451072e-01 -1.36480844e-02 3.13316554e-01
-4.02295947e-01 7.97126234e-01 -2.06770599e-01 -5.65842330e-01
-3.34271401e-01 -7.37300992e-01 -3.31167012e-01 -5.11208773e-01
-1.50857657e-01 -8.62294316e-01 5.47867775e-01 -3.80816817e-01
1.18608618e+00 -2.01188087e+00 1.04840631e-02 1.71281964e-01
1.93267465e-01 -1.37417773e-02 -6.81895763e-02 3.32628846e-01
-2.06523567e-01 1.17406547e-01 1.39892608e-01 -3.22681069e-01
2.69475490e-01 4.34724510e-01 -5.70090294e-01 -1.67953148e-01
3.98421228e-01 9.58367944e-01 -9.42838728e-01 -3.23839545e-01
3.36079001e-02 4.20633972e-01 -3.45854223e-01 4.84320015e-01
-3.27727169e-01 1.09281354e-01 -7.61776865e-01 4.03521508e-01
1.79318979e-01 -2.83470154e-01 2.44601443e-01 -3.53261679e-01
8.99762511e-02 3.55614632e-01 -1.13086939e+00 1.48466074e+00
-9.29592669e-01 2.37917826e-01 3.25075477e-01 -9.64443326e-01
9.44853544e-01 7.00108528e-01 4.57025588e-01 -4.49626774e-01
2.62261868e-01 9.73826721e-02 -2.28945404e-01 -8.33544791e-01
5.09060681e-01 -9.07382667e-01 -7.77748823e-01 6.78045213e-01
1.00415193e-01 -1.40306547e-01 1.60194542e-02 2.47410178e-01
1.23321140e+00 -7.89906606e-02 8.57252598e-01 -2.84859352e-02
3.50546360e-01 2.23778918e-01 7.31828570e-01 5.72605431e-01
-2.64005333e-01 -1.78902507e-01 7.51840413e-01 -6.17931664e-01
-3.83837968e-01 -3.88707966e-01 2.28381515e-01 1.96109247e+00
-2.06399992e-01 -6.50563240e-01 -3.60772848e-01 -7.58902729e-01
-3.65512781e-02 7.24992156e-01 -8.55245769e-01 -1.68585647e-02
-8.40323716e-02 -7.48059630e-01 7.20593810e-01 4.73202527e-01
1.90041751e-01 -1.41231632e+00 -7.62020230e-01 1.95415184e-01
-3.84149581e-01 -1.42578447e+00 1.27791479e-01 7.80940235e-01
-3.60158771e-01 -1.00655472e+00 -3.01299989e-01 -5.39352179e-01
1.36218861e-01 -1.65223822e-01 1.48039651e+00 -1.45753294e-01
-2.67986685e-01 7.22989917e-01 -7.79946327e-01 -6.93145931e-01
-2.33869389e-01 -4.29723300e-02 -1.09635673e-01 1.42781556e-01
9.10979152e-01 -3.26324612e-01 4.49440889e-02 5.76141383e-03
-8.79189491e-01 -2.95692775e-02 5.58103323e-01 4.72547680e-01
2.26175830e-01 1.64601788e-01 7.72356927e-01 -1.03842068e+00
9.40295637e-01 -6.84565544e-01 2.05762178e-01 4.50362772e-01
-3.46913427e-01 1.81759924e-01 2.89944619e-01 -2.82637954e-01
-1.32945704e+00 2.51335632e-02 -2.84433633e-01 -2.20592961e-01
-8.45274568e-01 7.85656273e-01 -1.33403972e-01 2.28684291e-01
2.68118620e-01 -1.11774072e-01 -2.49133304e-01 -3.26488972e-01
7.84593940e-01 5.76882482e-01 2.22661734e-01 -8.54107797e-01
9.62385982e-02 3.03258896e-01 -3.19759160e-01 -7.47221410e-01
-1.55340278e+00 -1.00685477e+00 -3.71436208e-01 -2.01059625e-01
1.08102727e+00 -1.15484750e+00 -7.61717141e-01 2.77316988e-01
-1.36762035e+00 -3.08320880e-01 -5.00796027e-02 4.05125260e-01
-4.29969192e-01 -6.95003793e-02 -6.79327428e-01 -1.08968937e+00
-2.87652105e-01 -4.74586934e-01 1.23344862e+00 7.07809851e-02
-8.89255404e-01 -1.51508176e+00 1.87755585e-01 3.58711541e-01
2.95511186e-01 2.26162583e-01 9.93947923e-01 -1.08573723e+00
3.77163202e-01 -1.14197098e-01 -3.04534078e-01 -1.84126571e-02
2.66844720e-01 -2.29646623e-01 -1.04119241e+00 3.07461262e-01
4.31778841e-02 -8.90018761e-01 1.02847862e+00 2.31568694e-01
8.44234169e-01 -3.63029003e-01 -2.55367488e-01 -3.16186175e-02
1.27006865e+00 1.88006125e-02 3.96520048e-01 1.66603148e-01
7.36711144e-01 1.10183442e+00 8.68580759e-01 7.21562266e-01
8.31319511e-01 5.23757577e-01 2.77718231e-02 -3.23230118e-01
3.43653262e-01 2.70101100e-01 2.86388040e-01 6.11947417e-01
1.16826795e-01 -5.55292331e-02 -9.11083758e-01 6.96227312e-01
-2.04138541e+00 -9.79841173e-01 -3.70904863e-01 1.26910126e+00
1.24810183e+00 -3.04352283e-01 7.37800077e-02 7.23522231e-02
5.27873755e-01 2.87640572e-01 -2.60416180e-01 -1.05550897e+00
-2.24667057e-01 2.49847189e-01 -1.63805068e-01 4.40045595e-01
-1.40243244e+00 1.21707106e+00 6.29453754e+00 9.38556373e-01
-1.00125408e+00 3.14481348e-01 6.26382470e-01 1.16033666e-01
-4.53204840e-01 2.51130667e-02 -4.73699242e-01 -1.91213340e-01
6.59561992e-01 5.84318303e-02 6.19814992e-02 8.45261693e-01
1.34813070e-01 -3.15546185e-01 -8.16011369e-01 1.06922376e+00
3.77457321e-01 -8.99030507e-01 -2.39879966e-01 -3.42542112e-01
5.51331520e-01 -2.50874609e-01 -3.92523557e-01 6.01098061e-01
5.12263894e-01 -7.33593404e-01 3.94371241e-01 6.26068115e-01
2.47316331e-01 -8.41197312e-01 9.12984431e-01 2.38979071e-01
-1.04419875e+00 4.55189422e-02 -1.19843252e-01 -7.85545111e-01
1.79219410e-01 1.30221105e+00 -4.85082746e-01 5.48648238e-01
6.50160313e-01 9.75490153e-01 -2.40804702e-01 4.16488886e-01
-6.42529070e-01 7.91940868e-01 -2.08545923e-01 -1.85163796e-01
4.06504832e-02 2.71999836e-01 3.41612428e-01 1.69816637e+00
-5.04326671e-02 4.70135003e-01 5.94967544e-01 6.64781809e-01
-4.33289185e-02 4.87358093e-01 -4.27381963e-01 -1.90673754e-01
-3.22643183e-02 1.71758139e+00 -8.92158806e-01 -5.77451587e-01
-4.43585366e-01 1.00004947e+00 5.08819520e-01 3.32369387e-01
-5.61822951e-01 -2.59611338e-01 1.03760612e+00 -7.51701415e-01
2.30199873e-01 1.66681707e-02 -2.99519658e-01 -1.25975406e+00
-2.62901783e-01 -7.02810884e-01 7.94540286e-01 -8.85228634e-01
-1.68963051e+00 4.59880024e-01 -2.02507883e-01 -5.41332006e-01
-3.35508019e-01 -6.44769669e-01 -7.09376097e-01 4.74635541e-01
-1.75560594e+00 -1.02800250e+00 -1.35698438e-01 5.61002672e-01
3.87510896e-01 3.48048449e-01 1.36458910e+00 -3.49710323e-02
-5.04484296e-01 2.87995130e-01 -6.47075951e-01 1.13127179e-01
1.01817894e+00 -1.33713341e+00 -3.65396768e-01 5.15377164e-01
2.01086119e-01 6.35877252e-01 8.71762037e-01 -6.32646263e-01
-1.32406878e+00 -8.92286599e-01 1.30951881e+00 -4.45852250e-01
8.42241466e-01 -3.23880315e-01 -1.03975570e+00 4.25661564e-01
5.66080868e-01 -8.76232702e-03 1.43725193e+00 1.13216209e+00
-8.32935631e-01 1.39492884e-01 -9.31005776e-01 1.99617222e-01
5.29351473e-01 -7.37655461e-01 -8.34560931e-01 -6.16435409e-02
7.00828433e-01 3.62633215e-03 -1.02377188e+00 3.32014501e-01
5.53693712e-01 -7.26398170e-01 3.67202163e-01 -7.63121247e-01
4.81169790e-01 -7.15852808e-03 -1.15573041e-01 -1.46498048e+00
-7.78053254e-02 -6.64608121e-01 3.18044387e-02 1.52623439e+00
3.62729877e-01 -3.89887124e-01 2.82504171e-01 8.77879143e-01
2.06748992e-01 -6.77278697e-01 -5.39844751e-01 -3.65881085e-01
-7.46728331e-02 -1.05913520e+00 4.13470447e-01 1.42097545e+00
6.04975104e-01 8.61026764e-01 -5.52499712e-01 2.40135677e-02
1.99930564e-01 5.63147306e-01 5.23559153e-01 -1.35458148e+00
4.97024581e-02 -4.18270677e-01 -2.47449636e-01 -5.64564466e-01
8.06734622e-01 -7.08557129e-01 3.41966957e-01 -1.56710470e+00
4.40973312e-01 -2.81102359e-01 -6.48818731e-01 1.18893683e+00
-6.63223863e-01 1.84999511e-01 -4.64865677e-02 -2.48834416e-01
-1.34502864e+00 4.69076097e-01 7.38390326e-01 2.64384091e-01
-1.10720612e-01 -4.10577357e-01 -8.27960312e-01 9.23424423e-01
7.45601952e-01 -5.36208153e-01 -1.82035491e-01 -8.64226222e-02
9.52032924e-01 -1.20638452e-01 3.19141567e-01 -6.17802858e-01
1.33998513e-01 -3.97452384e-01 1.82439499e-02 -9.39084962e-02
1.56963482e-01 -7.63686419e-01 -4.32919323e-01 -2.11383887e-02
-6.59252107e-01 -9.61373374e-02 4.53000456e-01 4.02938753e-01
-4.14248705e-01 1.34494007e-01 3.97908777e-01 -1.21583343e-01
-1.24884558e+00 -2.01836571e-01 -7.44306624e-01 1.51345685e-01
8.44487965e-01 1.81826964e-01 -1.56973638e-02 -6.82651520e-01
-9.30783272e-01 2.17026860e-01 1.81830272e-01 7.30598211e-01
5.46967566e-01 -1.08949220e+00 -6.84842110e-01 -2.14896992e-01
3.74119610e-01 -5.55256128e-01 2.22031027e-01 9.27996397e-01
5.80880821e-01 3.10777336e-01 1.30287230e-01 -1.63778856e-01
-1.42780888e+00 5.36723673e-01 1.30137235e-01 -6.20024025e-01
-4.47024480e-02 9.68470216e-01 4.43364382e-02 -4.34995055e-01
7.23745003e-02 -2.21713185e-01 -6.52883053e-01 5.67361832e-01
5.66355050e-01 -8.27220082e-02 -5.67661710e-02 -6.41016424e-01
-7.43040264e-01 6.54934883e-01 7.01931864e-02 2.34846026e-02
1.35797095e+00 -2.89123595e-01 -6.66131258e-01 9.95107472e-01
1.07162154e+00 4.39977683e-02 -5.54531038e-01 -4.28218722e-01
4.00009394e-01 9.73249152e-02 1.16783813e-01 -1.31319976e+00
-6.74730003e-01 7.60184288e-01 1.53727412e-01 3.80976021e-01
1.38453889e+00 4.92096424e-01 6.65317535e-01 4.30237412e-01
2.76078373e-01 -1.05974782e+00 2.78048664e-01 1.00102556e+00
7.60734558e-01 -1.30682385e+00 -1.85079515e-01 -7.23234296e-01
-1.05548596e+00 1.08714867e+00 5.89064538e-01 -4.44768146e-02
5.56215942e-01 5.31522036e-01 6.43208981e-01 -6.48210287e-01
-1.24808955e+00 -7.93984115e-01 4.23511565e-01 1.08951673e-01
9.21596825e-01 2.22876459e-01 -2.85083026e-01 1.18063247e+00
1.14093930e-01 3.69541980e-02 1.32138461e-01 8.77144933e-01
-3.73975396e-01 -1.20448780e+00 -2.71506160e-01 3.81192356e-01
-9.05673027e-01 -1.58482283e-01 -8.17809105e-01 2.27500096e-01
3.19345474e-01 1.28526855e+00 -3.85913104e-02 -2.58139938e-01
4.95461151e-02 7.08975136e-01 2.40862221e-01 -8.52983952e-01
-9.67453420e-01 1.61767825e-01 4.38958377e-01 -7.11914539e-01
-1.02734983e+00 -4.75292355e-01 -1.59607196e+00 2.65231162e-01
-1.28855273e-01 5.24821341e-01 6.00421250e-01 1.62909758e+00
5.32229424e-01 7.47037947e-01 4.42354679e-01 -3.54957402e-01
1.53481320e-01 -9.32974339e-01 -5.65876961e-01 7.54792809e-01
1.75611272e-01 -6.15300298e-01 -3.07613522e-01 6.84006959e-02] | [12.687495231628418, 6.2456374168396] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.