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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
739049e4-b7de-4e3c-8c3f-3a041e703f7e | digiface-1m-1-million-digital-face-images-for | 2210.02579 | null | https://arxiv.org/abs/2210.02579v1 | https://arxiv.org/pdf/2210.02579v1.pdf | DigiFace-1M: 1 Million Digital Face Images for Face Recognition | State-of-the-art face recognition models show impressive accuracy, achieving over 99.8% on Labeled Faces in the Wild (LFW) dataset. Such models are trained on large-scale datasets that contain millions of real human face images collected from the internet. Web-crawled face images are severely biased (in terms of race, lighting, make-up, etc) and often contain label noise. More importantly, the face images are collected without explicit consent, raising ethical concerns. To avoid such problems, we introduce a large-scale synthetic dataset for face recognition, obtained by rendering digital faces using a computer graphics pipeline. We first demonstrate that aggressive data augmentation can significantly reduce the synthetic-to-real domain gap. Having full control over the rendering pipeline, we also study how each attribute (e.g., variation in facial pose, accessories and textures) affects the accuracy. Compared to SynFace, a recent method trained on GAN-generated synthetic faces, we reduce the error rate on LFW by 52.5% (accuracy from 91.93% to 96.17%). By fine-tuning the network on a smaller number of real face images that could reasonably be obtained with consent, we achieve accuracy that is comparable to the methods trained on millions of real face images. | ['Jingjing Shen', 'Roberto Cipolla', 'Julien Valentin', 'Dong Chen', 'Charlie Hewitt', 'Tadas Baltrusaitis', 'Martin de La Gorce', 'Gwangbin Bae'] | 2022-10-05 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 3.71377081e-01 5.13594627e-01 2.53195047e-01 -9.21372950e-01
-9.12509501e-01 -6.10554814e-01 6.41431034e-01 -8.91133368e-01
-2.85860658e-01 7.10282087e-01 -5.75192198e-02 5.55091538e-03
5.66213667e-01 -6.72365308e-01 -9.12750959e-01 -3.68304044e-01
2.13401735e-01 5.25091469e-01 -5.48785031e-01 5.69476038e-02
-1.25633627e-01 6.89688921e-01 -1.76000464e+00 4.12580043e-01
4.77255762e-01 1.03491175e+00 -6.00401878e-01 3.20984632e-01
1.51928291e-01 3.18654299e-01 -9.13071275e-01 -9.26834583e-01
8.15121293e-01 -4.69040960e-01 -3.46145958e-01 4.37752753e-01
1.31959212e+00 -9.93713140e-01 -1.23089701e-01 8.65329564e-01
6.29161835e-01 -3.93460155e-01 4.55828130e-01 -1.61998224e+00
-8.12376559e-01 1.26908854e-01 -7.77466536e-01 -4.82698739e-01
3.57241035e-01 5.94639659e-01 2.47932583e-01 -1.13624799e+00
8.58019829e-01 1.41561520e+00 7.45570600e-01 1.23102283e+00
-1.45535040e+00 -1.36982799e+00 -1.68168619e-01 -4.48823929e-01
-1.49983191e+00 -1.16839910e+00 5.80211639e-01 -5.16640604e-01
2.75096089e-01 2.07393855e-01 4.71358329e-01 1.71142125e+00
-4.62026864e-01 2.54320651e-01 1.46179891e+00 -3.24940443e-01
1.36364788e-01 2.07869619e-01 -4.64332104e-01 7.02506244e-01
3.27832520e-01 1.64556667e-01 -6.62028670e-01 -4.98633176e-01
6.91869557e-01 -2.47301191e-01 -2.20326975e-01 -2.36096546e-01
-8.28433692e-01 7.39798129e-01 6.05168119e-02 -3.94657373e-01
-1.29147127e-01 1.48496972e-02 2.33157367e-01 1.92628443e-01
6.19199336e-01 2.77814329e-01 -4.73621786e-01 5.64780384e-02
-8.63539040e-01 2.25862980e-01 7.72368848e-01 9.26905453e-01
7.77803719e-01 2.86053687e-01 -7.24453032e-02 8.99886847e-01
1.41278893e-01 7.75144577e-01 2.03051805e-01 -1.27871299e+00
2.58419454e-01 2.74812192e-01 2.95565367e-01 -9.34091091e-01
8.77100080e-02 3.05821020e-02 -7.70077646e-01 5.62838554e-01
7.66950548e-01 -2.58328050e-01 -1.18356228e+00 2.11327505e+00
4.28890496e-01 1.12890959e-01 -2.60846853e-01 6.47786856e-01
6.76478386e-01 1.37858048e-01 1.51366025e-01 -1.60390124e-01
1.30524027e+00 -6.18284404e-01 -5.34617007e-01 -2.86200315e-01
2.73588538e-01 -8.08515429e-01 1.37611580e+00 3.99208307e-01
-9.23557520e-01 -3.51795435e-01 -8.23049963e-01 1.83629319e-01
-2.43237004e-01 4.01825964e-01 5.82275510e-01 1.37078249e+00
-1.31644189e+00 4.03708845e-01 -4.71483737e-01 -4.66230512e-01
1.19174981e+00 5.29985309e-01 -8.54111195e-01 -4.22788680e-01
-6.84062958e-01 3.90317172e-01 -2.33554959e-01 3.83637659e-02
-1.08037698e+00 -9.53973532e-01 -7.91077018e-01 -2.49276131e-01
3.57490540e-01 -3.40740055e-01 1.16970491e+00 -1.37566841e+00
-1.41543829e+00 1.39942074e+00 -9.25518423e-02 2.27843504e-02
8.97334456e-01 -7.49173909e-02 -5.65667808e-01 2.71670055e-02
1.67214915e-01 1.01165175e+00 1.30552173e+00 -1.40823090e+00
4.53487709e-02 -6.47540689e-01 -3.40802640e-01 -3.96593034e-01
-5.23001134e-01 3.88130486e-01 -4.19384867e-01 -6.10876501e-01
-4.59687442e-01 -1.01574504e+00 1.93551183e-01 6.49107397e-01
-2.44065225e-01 1.77746922e-01 8.45869124e-01 -8.30609500e-01
3.73240024e-01 -2.21269774e+00 -5.94132066e-01 1.69463292e-01
1.28959462e-01 5.30437469e-01 -5.71617782e-01 -6.06692992e-02
-1.96099922e-01 4.50595737e-01 -1.82376727e-01 -6.05855227e-01
-9.22994912e-02 -1.52832363e-03 -2.26870477e-01 4.47112471e-01
4.89437789e-01 7.97375083e-01 -4.13731426e-01 -2.48451233e-01
-2.22480252e-01 7.65171349e-01 -7.90121138e-01 3.52913409e-01
-7.30771124e-02 5.99252641e-01 -7.23548755e-02 8.34995270e-01
1.11798704e+00 3.54317464e-02 2.69598246e-01 -1.32009819e-01
4.89722729e-01 -1.40151650e-01 -8.18972468e-01 1.52153742e+00
-3.69113803e-01 5.90623319e-01 4.16486830e-01 -2.62753457e-01
9.82870042e-01 3.53825718e-01 2.31732890e-01 -6.21448636e-01
1.70220122e-01 1.67022705e-01 -9.01983976e-02 -2.30363473e-01
-1.96096674e-03 -5.01884706e-02 3.33265960e-01 6.94981933e-01
5.73675483e-02 -3.77963409e-02 -7.04235509e-02 3.68214282e-03
9.58341777e-01 1.64514780e-01 -2.32487530e-01 -1.93781942e-01
7.32308030e-02 -3.29874903e-01 6.74394310e-01 4.56899464e-01
-2.30348468e-01 9.54488635e-01 5.70144594e-01 -6.33241117e-01
-1.27475417e+00 -8.73201728e-01 -5.62136769e-02 9.69421983e-01
-6.33134067e-01 -3.82005811e-01 -1.37214100e+00 -8.72171998e-01
1.43577397e-01 4.02342528e-01 -9.87993181e-01 -1.69768542e-01
-4.30909693e-01 -7.28927970e-01 8.93728077e-01 3.94650847e-01
6.13210082e-01 -9.54093158e-01 -2.87323087e-01 -2.26696685e-01
1.04562476e-01 -1.27205038e+00 -5.41150749e-01 -7.53941119e-01
-3.96242023e-01 -1.20901906e+00 -6.06350064e-01 -3.47099483e-01
1.05371070e+00 -1.21867269e-01 1.37379503e+00 3.12834561e-01
-6.94869936e-01 3.70969653e-01 -2.40118634e-02 -4.66822326e-01
-4.69732404e-01 -2.15384364e-01 2.29790136e-01 4.57654327e-01
4.00921762e-01 -4.57716107e-01 -6.43481970e-01 5.00838220e-01
-7.13693202e-01 5.22458169e-04 1.95827305e-01 8.22343647e-01
2.31799766e-01 -4.97208923e-01 6.61856115e-01 -1.25620377e+00
3.84182274e-01 -2.64277846e-01 -7.11478055e-01 2.36897647e-01
-5.10092497e-01 -2.75669664e-01 4.98628736e-01 -6.29180312e-01
-1.18670237e+00 9.57906991e-02 -1.16555482e-01 -5.44070065e-01
-4.94005769e-01 -4.26741034e-01 -4.73158270e-01 -3.42433751e-01
8.16690028e-01 -1.01244248e-01 3.64058644e-01 -4.26500142e-01
1.77952096e-01 8.30062926e-01 5.33568382e-01 -7.90222585e-01
8.94913852e-01 5.19994795e-01 -1.58461779e-01 -5.52474618e-01
-7.37211943e-01 5.10378361e-01 -4.28351372e-01 -2.80090153e-01
5.72297633e-01 -1.04815352e+00 -9.68750894e-01 7.05424666e-01
-9.01688516e-01 -5.51767588e-01 -2.51285136e-01 9.57939848e-02
-3.33801299e-01 -1.91730604e-01 -4.81262952e-01 -9.48838532e-01
-2.28793487e-01 -9.89943206e-01 1.39087820e+00 2.71725729e-02
-2.41283104e-01 -4.12299424e-01 -3.69654357e-01 6.94798708e-01
6.73408031e-01 6.21739805e-01 5.10545492e-01 -4.29561585e-01
-3.82973641e-01 -2.17529297e-01 -3.93833488e-01 4.50138271e-01
3.35115284e-01 2.25754529e-01 -1.45097864e+00 -4.73522246e-01
-1.34744301e-01 -1.00323188e+00 3.72286767e-01 -1.93311483e-01
1.41008949e+00 -5.14303982e-01 -2.62277778e-02 7.52057850e-01
1.09246540e+00 4.91736233e-02 7.44924605e-01 -1.63223967e-01
7.54697740e-01 8.45526040e-01 3.83081436e-01 5.35463274e-01
9.04914108e-04 5.34112036e-01 2.69650787e-01 -2.42817625e-01
-4.37842846e-01 -5.16162395e-01 2.32941717e-01 -8.33073631e-02
-4.29289276e-03 -4.64186892e-02 -9.18333113e-01 1.71517119e-01
-1.15085793e+00 -6.38070881e-01 2.37491667e-01 2.33364677e+00
9.38369215e-01 -2.53001243e-01 2.17188641e-01 -2.93687195e-01
7.99554408e-01 -1.03042617e-01 -7.89244652e-01 -2.31873587e-01
-1.78725064e-01 4.28470850e-01 3.99867177e-01 3.43934059e-01
-7.07098186e-01 8.79577219e-01 6.99408007e+00 6.07738495e-01
-1.20778751e+00 1.58781543e-01 1.41455328e+00 -4.87585366e-01
-2.62974590e-01 -4.23929334e-01 -6.52687013e-01 5.07265985e-01
9.76230621e-01 1.74145728e-01 9.82783914e-01 9.11717832e-01
-6.03944361e-02 1.91082031e-01 -1.29946303e+00 1.25827610e+00
5.44853806e-01 -1.07309914e+00 6.49422854e-02 4.41655874e-01
8.40060115e-01 -1.64407864e-01 4.43202227e-01 8.66540968e-02
5.56845605e-01 -1.51763070e+00 6.64805532e-01 2.85008460e-01
1.81258512e+00 -6.57331526e-01 3.60790282e-01 3.85309644e-02
-5.98377347e-01 1.29819661e-01 -2.59025395e-01 1.26634151e-01
-1.82470769e-01 5.07386506e-01 -8.28459680e-01 -1.13984428e-01
8.60192835e-01 2.06938729e-01 -9.05583262e-01 1.63270921e-01
-1.72333598e-01 6.84692979e-01 -5.11095583e-01 4.61840123e-01
-3.15330952e-01 -1.89997509e-01 -1.81681484e-01 7.18817055e-01
4.03458923e-01 1.40923247e-01 -2.83612818e-01 1.09286666e+00
-7.65744746e-01 -6.43113777e-02 -8.48570049e-01 -1.11001842e-01
5.25175810e-01 1.31590664e+00 -2.39895582e-01 -1.73913240e-01
-4.13475007e-01 9.28356409e-01 4.33112949e-01 4.49665546e-01
-6.92390561e-01 1.59027025e-01 9.68399942e-01 4.25487131e-01
1.30610885e-02 1.13667943e-01 -1.14908986e-01 -1.13603437e+00
2.36792505e-01 -1.36935067e+00 -3.22563276e-02 -8.26101243e-01
-1.40534329e+00 8.86981249e-01 -3.36654186e-01 -8.12169313e-01
-4.18858796e-01 -7.30162024e-01 -2.98697650e-01 9.30563509e-01
-1.05915141e+00 -1.43864810e+00 -5.79417288e-01 6.82269216e-01
2.12317526e-01 -5.24865508e-01 1.13241720e+00 3.82288665e-01
-7.15430379e-01 1.22364783e+00 -1.75361887e-01 4.34256375e-01
1.09905183e+00 -5.51792562e-01 9.00357425e-01 5.92713356e-01
-2.43522623e-03 6.03696525e-01 2.80375808e-01 -5.16151190e-01
-1.41990769e+00 -1.20853126e+00 3.36706698e-01 -7.52876461e-01
2.03751802e-01 -9.57280874e-01 -7.31603801e-01 8.43261898e-01
5.90231642e-02 5.01429319e-01 9.40373957e-01 4.43078876e-02
-9.68358219e-01 -2.61689276e-01 -1.88024437e+00 6.86053276e-01
1.39778304e+00 -6.18628263e-01 1.70720801e-01 4.08951223e-01
2.83049315e-01 -1.17140733e-01 -7.32402325e-01 3.85955513e-01
9.08306241e-01 -1.16080451e+00 7.44687080e-01 -8.74142826e-01
3.89370024e-01 2.16250517e-03 -2.03476116e-01 -1.17204642e+00
7.24546388e-02 -8.15636396e-01 2.15618610e-01 1.73020184e+00
2.54592031e-01 -8.96361351e-01 1.12130785e+00 1.37229550e+00
5.65544307e-01 -5.92550099e-01 -7.49335110e-01 -6.96955681e-01
4.72158082e-02 -2.09032089e-01 1.10408831e+00 1.09474301e+00
-6.97966218e-01 2.79703108e-03 -6.23641372e-01 -1.32284373e-01
9.21425641e-01 -1.94423035e-01 1.21630347e+00 -1.27176070e+00
-2.45714560e-03 -2.89202109e-02 -2.26607174e-01 -2.91089863e-01
6.70129240e-01 -5.77598572e-01 -1.72745556e-01 -7.20529497e-01
2.81772554e-01 -5.69800854e-01 3.11514884e-01 8.11526358e-01
-8.88896920e-03 1.01015878e+00 2.15056598e-01 -1.21517794e-03
4.73443866e-02 4.81969208e-01 1.12489247e+00 2.83501558e-02
3.94442827e-01 -4.83086824e-01 -8.87843609e-01 7.77335107e-01
9.09313202e-01 -4.51680750e-01 -2.67352581e-01 -6.07708335e-01
-1.40462413e-01 -2.12419957e-01 3.88759017e-01 -7.91110992e-01
-2.30244637e-01 -1.69775799e-01 8.31937850e-01 2.12111250e-01
5.98171711e-01 -7.70747960e-01 4.48045313e-01 1.84940740e-01
-2.95017779e-01 -1.50085360e-01 2.41311938e-01 1.88599125e-01
1.27898633e-01 2.18559593e-01 9.61552501e-01 -1.68498620e-01
-1.17835872e-01 6.20049894e-01 7.65645429e-02 2.92970777e-01
9.99050260e-01 -8.45756903e-02 -5.83645284e-01 -5.05333304e-01
-3.41941953e-01 -1.82290480e-01 1.04639387e+00 4.77606863e-01
4.09361005e-01 -1.49610364e+00 -1.00637579e+00 8.86598229e-01
1.95615798e-01 -2.01757371e-01 -4.76399735e-02 1.83600396e-01
-5.14743805e-01 -1.78061247e-01 -4.11063671e-01 -3.76877248e-01
-1.33466935e+00 4.36831146e-01 3.07305664e-01 3.33538890e-01
-2.28396788e-01 9.21641886e-01 2.93677807e-01 -5.90149760e-01
-1.36198085e-02 3.59172136e-01 5.35628736e-01 -7.25652948e-02
7.76146948e-01 1.81390420e-01 1.43172666e-01 -8.01354170e-01
-3.74111414e-01 5.92907906e-01 -1.04253553e-01 -3.29210103e-01
1.26065063e+00 3.34503293e-01 -1.32388070e-01 -2.24406585e-01
1.28398144e+00 2.07792655e-01 -1.79109561e+00 1.68895975e-01
-5.61919808e-01 -1.10989702e+00 -4.32547867e-01 -9.68472183e-01
-1.55863738e+00 7.55884767e-01 7.55576015e-01 -2.07723796e-01
1.10638928e+00 -1.48312688e-01 6.49901152e-01 9.02109742e-02
7.26681173e-01 -8.00460339e-01 9.33164954e-02 -2.89066844e-02
1.08786130e+00 -1.49800193e+00 -1.38542891e-01 -6.79150224e-01
-4.75992084e-01 7.54576504e-01 8.78127575e-01 1.47042990e-01
4.74451184e-01 5.90161324e-01 4.28493500e-01 3.15111987e-02
-6.41033411e-01 4.56388026e-01 -8.73469040e-02 1.03384042e+00
3.40129673e-01 4.55038100e-02 2.73277640e-01 4.08581555e-01
-5.23400903e-01 1.90382943e-01 4.11796093e-01 6.40346229e-01
2.13053674e-01 -1.22771466e+00 -6.20301902e-01 6.73929870e-01
-5.02742827e-01 1.00743927e-01 -7.07381070e-01 7.04755008e-01
2.17780158e-01 9.11631703e-01 2.38180682e-01 -2.86400974e-01
2.52107918e-01 2.64515102e-01 6.48162007e-01 -6.04014754e-01
-5.16743422e-01 -3.36814016e-01 2.10065871e-01 -8.89224172e-01
-1.85546905e-01 -6.68866932e-01 -5.40802300e-01 -5.87685585e-01
1.20913982e-01 -3.05355281e-01 7.54989445e-01 5.46571910e-01
7.14143991e-01 -7.29967281e-02 7.56052077e-01 -9.77336049e-01
-5.29663801e-01 -9.03552532e-01 -5.76373041e-01 8.78880203e-01
1.25542149e-01 -5.59399426e-01 -5.17020643e-01 2.69645154e-01] | [12.890278816223145, 0.621039867401123] |
aa07752b-7ca6-4c0d-b40e-90c8987a3213 | how-crucial-is-transformer-in-decision | 2211.14655 | null | https://arxiv.org/abs/2211.14655v1 | https://arxiv.org/pdf/2211.14655v1.pdf | How Crucial is Transformer in Decision Transformer? | Decision Transformer (DT) is a recently proposed architecture for Reinforcement Learning that frames the decision-making process as an auto-regressive sequence modeling problem and uses a Transformer model to predict the next action in a sequence of states, actions, and rewards. In this paper, we analyze how crucial the Transformer model is in the complete DT architecture on continuous control tasks. Namely, we replace the Transformer by an LSTM model while keeping the other parts unchanged to obtain what we call a Decision LSTM model. We compare it to DT on continuous control tasks, including pendulum swing-up and stabilization, in simulation and on physical hardware. Our experiments show that DT struggles with continuous control problems, such as inverted pendulum and Furuta pendulum stabilization. On the other hand, the proposed Decision LSTM is able to achieve expert-level performance on these tasks, in addition to learning a swing-up controller on the real system. These results suggest that the strength of the Decision Transformer for continuous control tasks may lie in the overall sequential modeling architecture and not in the Transformer per se. | ['Jan Peters', 'Junning Huang', 'Boris Belousov', 'Max Siebenborn'] | 2022-11-26 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-2.10869294e-02 2.81842440e-01 -2.19776079e-01 -2.16732454e-02
-8.65990967e-02 -4.84164923e-01 6.18751466e-01 -2.04014644e-01
-3.21275890e-01 9.85748053e-01 -2.92387873e-01 -7.73013771e-01
-8.15595910e-02 -5.80459714e-01 -7.78277755e-01 -6.84827447e-01
-2.33501002e-01 3.17096591e-01 3.34017694e-01 -6.24711812e-01
1.45015910e-01 2.53178895e-01 -1.44346023e+00 8.86141732e-02
6.25508189e-01 1.03741395e+00 3.15760493e-01 7.18638837e-01
2.35655636e-01 1.43128276e+00 -5.42913377e-01 1.36840031e-01
3.68710607e-02 -4.82536376e-01 -9.21369910e-01 -1.36099190e-01
-8.03190470e-02 -4.90724266e-01 -1.16295554e-01 7.22301245e-01
5.12059033e-01 1.46024659e-01 4.66020226e-01 -1.33300781e+00
-7.92011842e-02 7.36636519e-01 -2.86368042e-01 1.41447499e-01
1.94653839e-01 4.29177791e-01 8.51949990e-01 -4.43571448e-01
3.62787992e-01 1.22137451e+00 5.18022478e-01 6.12159252e-01
-1.19122875e+00 -4.97991323e-01 3.80240738e-01 3.98704499e-01
-7.45452464e-01 -2.70557433e-01 5.29871523e-01 -3.72754633e-01
1.39000165e+00 -2.83952028e-01 6.87297761e-01 1.13363111e+00
8.91244471e-01 9.62523162e-01 1.20728791e+00 -4.84723568e-01
5.50370336e-01 -2.15930626e-01 -5.62813953e-02 6.33749545e-01
-6.62147105e-02 6.70073211e-01 -2.08546624e-01 1.62125468e-01
8.15957069e-01 -1.66668415e-01 -2.42599413e-01 -6.10885501e-01
-1.12785900e+00 6.25913858e-01 4.08209622e-01 3.69443566e-01
-6.14177465e-01 5.45464396e-01 6.25895202e-01 8.66787016e-01
-7.14332163e-02 5.15167296e-01 -7.01922417e-01 -2.91139930e-01
-5.58804989e-01 3.33904803e-01 8.40483725e-01 6.16318166e-01
4.14206535e-01 8.34453046e-01 -2.16575399e-01 2.69967288e-01
7.38369823e-02 3.75029683e-01 7.87876487e-01 -1.02584028e+00
2.46786684e-01 2.45413825e-01 4.76241142e-01 -3.80555809e-01
-5.80453634e-01 -5.47811687e-01 -5.58825135e-01 8.54094207e-01
2.37465844e-01 -7.70687819e-01 -8.46170187e-01 1.86897516e+00
7.02004880e-02 2.06401944e-01 1.22012578e-01 6.53510809e-01
-1.20921142e-01 5.93105376e-01 -2.61734333e-02 -2.82410115e-01
1.06795490e+00 -9.00095820e-01 -8.05798054e-01 -2.43160725e-01
4.44384396e-01 -2.84483671e-01 9.06440377e-01 5.65359652e-01
-1.20652759e+00 -7.73719549e-01 -1.44650972e+00 4.72530693e-01
-8.76913741e-02 1.49943307e-01 2.26188451e-01 2.60152668e-01
-1.27260768e+00 1.27573752e+00 -1.18249559e+00 -1.71445385e-01
-3.63194466e-01 6.34560287e-01 2.88716644e-01 5.72987497e-01
-1.48477006e+00 1.41482091e+00 3.98729384e-01 2.58355737e-02
-1.15920830e+00 -3.21800798e-01 -8.07845294e-01 2.46567860e-01
6.76736414e-01 -7.61304677e-01 2.06119633e+00 -9.50592041e-01
-2.19840097e+00 1.58845216e-01 9.82624367e-02 -1.05046391e+00
5.36517262e-01 -2.14810163e-01 -3.93376470e-01 -4.84474748e-02
-2.91709870e-01 4.47416604e-01 1.25949371e+00 -9.45890665e-01
-1.00861526e+00 -1.48145512e-01 2.42183298e-01 1.27668604e-01
5.67326415e-03 -4.81701225e-01 4.55511034e-01 -3.83039176e-01
-2.92482853e-01 -1.27311075e+00 -2.55810916e-01 -4.39760864e-01
-5.55437505e-02 -2.36304313e-01 1.09699953e+00 -3.94605786e-01
1.15425301e+00 -1.85448730e+00 2.91611463e-01 -8.05367082e-02
6.88796118e-02 4.40942466e-01 3.26229967e-02 6.33607090e-01
-3.69775057e-01 -2.73108482e-01 7.53585100e-02 -5.28616570e-02
1.51088923e-01 2.99791813e-01 -5.38428068e-01 2.17844129e-01
3.98240596e-01 9.07288432e-01 -9.52862680e-01 -6.45512417e-02
2.83884197e-01 3.69792543e-02 -2.01939508e-01 3.16602215e-02
-6.54471099e-01 4.79301244e-01 -5.36715150e-01 1.44241795e-01
1.32123768e-01 -9.40883011e-02 6.01052105e-01 2.46443763e-01
-5.35283983e-01 5.55985332e-01 -9.37676430e-01 1.40107834e+00
-5.45848668e-01 6.33879483e-01 1.95377231e-01 -1.14555526e+00
8.68481159e-01 6.23921812e-01 6.15147352e-01 -1.20937061e+00
3.84472817e-01 2.18318641e-01 3.52472931e-01 -4.38409895e-01
4.21817124e-01 -4.02143002e-01 5.47171943e-03 5.64182460e-01
1.52795628e-01 -2.01796979e-01 2.81359971e-01 -3.05468231e-01
1.05321968e+00 6.86753571e-01 3.40727150e-01 -3.84174019e-01
5.03714800e-01 1.43093780e-01 5.13808787e-01 5.09315789e-01
-1.80411831e-01 -1.88512117e-01 6.60665512e-01 -4.13337916e-01
-9.46797907e-01 -8.66482675e-01 3.98953527e-01 1.02568936e+00
-2.01434314e-01 -2.07898170e-01 -6.04702294e-01 -3.87986779e-01
-8.88860524e-02 1.11732984e+00 -5.05532205e-01 -5.17589211e-01
-8.16556692e-01 -2.30359256e-01 2.69827396e-01 6.10510468e-01
4.37229723e-01 -1.36104476e+00 -1.33568501e+00 7.16066778e-01
9.43544358e-02 -8.03478837e-01 -2.76055872e-01 7.81362414e-01
-9.34690475e-01 -1.14695024e+00 -5.14990568e-01 -8.57068956e-01
4.31783758e-02 -1.65175721e-01 8.93142343e-01 -3.37843239e-01
1.87889233e-01 2.62288868e-01 -1.71184123e-01 -3.88828158e-01
-6.28401458e-01 -1.53471194e-02 3.25709432e-01 -3.47956955e-01
-1.78372294e-01 -7.14926600e-01 -3.79948199e-01 3.68241757e-01
-5.72650731e-01 8.52240622e-02 5.18784463e-01 1.03617191e+00
1.50149569e-01 1.72872022e-01 8.16905499e-01 -3.51647645e-01
9.98477876e-01 -1.41013250e-01 -9.14141357e-01 2.29642555e-01
-7.66588807e-01 4.54570889e-01 1.02320611e+00 -4.86955911e-01
-7.93019056e-01 1.64709002e-01 -1.30411685e-01 -5.78447044e-01
3.39334548e-01 5.21290183e-01 4.19463277e-01 1.01057641e-01
5.03629625e-01 4.93356943e-01 4.71007615e-01 -1.37350947e-01
1.54931545e-02 2.66547173e-01 1.99967667e-01 -4.14556086e-01
4.50532585e-01 -1.22310489e-01 1.60329148e-01 -4.28894103e-01
-4.15478170e-01 2.72337317e-01 -4.08470213e-01 -3.78850251e-01
5.99713504e-01 -6.49262011e-01 -1.41055596e+00 6.18147850e-01
-8.28927457e-01 -8.83727908e-01 -5.18516362e-01 3.86751086e-01
-1.17014682e+00 -4.18051109e-02 -9.13761616e-01 -1.04852223e+00
-2.23218784e-01 -1.15849125e+00 6.67783141e-01 1.43622279e-01
-2.50616223e-01 -9.62913752e-01 1.92320868e-01 -4.27764535e-01
5.30538261e-01 3.26352924e-01 1.04965234e+00 -2.16867566e-01
-2.65563697e-01 8.09226744e-03 4.85232502e-01 4.54586148e-01
1.20198697e-01 -8.04030746e-02 -6.82859957e-01 -6.64279044e-01
4.77129877e-01 -4.96460170e-01 6.38331413e-01 4.75184858e-01
5.40527105e-01 -1.57020465e-01 -1.44894198e-01 -1.13067225e-01
1.23558950e+00 8.43042374e-01 4.48615611e-01 2.50046611e-01
1.30785614e-01 2.88863778e-01 8.61435473e-01 4.35582191e-01
2.76874870e-01 6.84750915e-01 5.12465298e-01 4.81928736e-02
1.15701213e-01 -4.56529677e-01 1.00463855e+00 5.80550075e-01
1.09301083e-01 1.03743702e-01 -8.64848137e-01 3.30433637e-01
-2.09060407e+00 -1.00576878e+00 2.21735492e-01 2.33042955e+00
5.60302496e-01 5.90603888e-01 3.89243305e-01 4.62020457e-01
6.90111220e-01 -9.00152400e-02 -9.28904414e-01 -7.89019227e-01
3.46430540e-01 2.80240834e-01 3.55053961e-01 5.16577423e-01
-9.31259573e-01 9.34487045e-01 6.85755014e+00 5.49189806e-01
-1.76305056e+00 -1.67718753e-01 3.95926893e-01 -2.53158761e-03
3.21692616e-01 -3.44379693e-02 -6.23570979e-01 3.62865329e-01
1.20629561e+00 -4.43517774e-01 4.61933464e-01 8.21479857e-01
6.38537347e-01 -1.63463295e-01 -1.28972614e+00 4.45125043e-01
-5.52740753e-01 -1.10253990e+00 -3.32671106e-01 -5.95723540e-02
4.67814386e-01 8.59333575e-03 8.60451832e-02 9.08563673e-01
6.43240154e-01 -8.98390830e-01 9.49706495e-01 2.90513873e-01
4.90191579e-01 -7.44141281e-01 5.73246062e-01 7.83229172e-01
-1.15049970e+00 -5.89147627e-01 -3.54242115e-03 -4.24478501e-01
1.19199440e-01 -8.80338848e-02 -1.04793715e+00 4.47166800e-01
2.44382232e-01 6.56503320e-01 -1.96793914e-01 7.88685918e-01
-3.69865388e-01 5.66860795e-01 -1.84798494e-01 -4.32606220e-01
6.16177082e-01 -2.50547826e-02 4.43873554e-01 7.24087358e-01
2.74662942e-01 -1.88410819e-01 3.24248791e-01 6.82405829e-01
5.19732177e-01 -5.68809211e-01 -4.65245128e-01 6.65399134e-02
1.31755456e-01 7.37214088e-01 -3.58145803e-01 -3.25338840e-01
1.08357564e-01 6.34733737e-01 2.92700082e-01 3.86407197e-01
-1.03930569e+00 -2.30802432e-01 5.82318068e-01 -1.43922248e-03
6.74172580e-01 -4.00316060e-01 1.80558451e-02 -8.82060111e-01
-1.10307679e-01 -8.74466836e-01 1.45326555e-01 -8.06541979e-01
-6.46599054e-01 6.95074260e-01 -6.98192418e-02 -1.64792919e+00
-1.14382410e+00 -4.83018696e-01 -6.86686039e-01 5.99454820e-01
-1.66584945e+00 -4.87101644e-01 2.67566293e-01 7.34739304e-01
5.72782874e-01 4.18339334e-02 7.29126513e-01 2.10284963e-02
-6.11429572e-01 6.70597702e-02 1.43222228e-01 -1.94342792e-01
3.58236641e-01 -1.44148350e+00 4.61168259e-01 5.48976183e-01
-3.87179643e-01 3.15357149e-01 1.03470433e+00 -3.65933418e-01
-1.39265919e+00 -8.04761767e-01 6.12717867e-01 1.68731287e-01
8.39654624e-01 -4.66669574e-02 -6.01617813e-01 7.38204598e-01
5.33923984e-01 -3.82735193e-01 -1.32203728e-01 -1.97469801e-01
3.07694554e-01 -8.06313083e-02 -1.01843202e+00 7.31930494e-01
6.65539443e-01 -1.51779339e-01 -7.53271043e-01 -7.00980797e-02
7.39472270e-01 -5.07216752e-01 -6.30163848e-01 3.11716765e-01
5.04282594e-01 -8.22805464e-01 6.55920446e-01 -7.86126077e-01
5.39438009e-01 -3.73731673e-01 2.37663090e-01 -1.77867627e+00
-4.59150434e-01 -7.22270131e-01 -5.27199507e-01 5.61186016e-01
4.18910414e-01 -7.45012045e-01 5.94205379e-01 2.49212801e-01
-1.42950892e-01 -9.39444304e-01 -1.00049150e+00 -9.79860902e-01
2.09226772e-01 -1.36089161e-01 3.85348111e-01 7.36409068e-01
3.27970386e-01 7.79501557e-01 -5.89845061e-01 -9.53320637e-02
2.26689115e-01 2.97703981e-01 4.12127376e-01 -9.19770837e-01
-5.43308377e-01 -4.35815096e-01 -2.23142561e-02 -1.07106888e+00
2.43612915e-01 -5.12669921e-01 1.42922834e-01 -1.43693888e+00
-6.09910905e-01 -7.89692700e-02 -4.47362989e-01 7.09872305e-01
1.01303257e-01 -5.75190425e-01 4.80043292e-01 -2.44357251e-02
-3.38127792e-01 7.84601212e-01 1.35379446e+00 -1.46033600e-01
-4.58640218e-01 4.67243582e-01 -3.06917876e-01 6.31818295e-01
1.06818259e+00 -3.00959378e-01 -6.69005513e-01 -2.29726568e-01
-9.45708081e-02 9.14769292e-01 2.23731354e-01 -1.39956093e+00
3.99005651e-01 -2.24432856e-01 1.21684618e-01 -4.70962942e-01
3.79398018e-01 -8.85003328e-01 -6.73973337e-02 1.35001540e+00
-3.23859692e-01 4.71717834e-01 4.37431544e-01 4.17831838e-01
-2.16089264e-01 1.27213616e-02 8.02835524e-01 6.31321892e-02
-8.39436412e-01 -9.53195468e-02 -1.00705242e+00 -2.80000776e-01
1.10223246e+00 -1.46674737e-01 -5.71092479e-02 -6.07803702e-01
-1.21173739e+00 6.46171808e-01 7.92622194e-02 6.10831857e-01
3.84293437e-01 -1.03481674e+00 -4.33060378e-01 3.71330157e-02
-4.68016148e-01 -5.16833305e-01 4.07840647e-02 8.09993207e-01
-1.86732262e-01 7.57096291e-01 -6.66883707e-01 -4.70149428e-01
-9.63088036e-01 6.85464561e-01 7.66101956e-01 -7.40534484e-01
-6.20302737e-01 7.85700828e-02 -2.67659068e-01 -2.37208575e-01
2.95423537e-01 -6.91722929e-01 -2.77912587e-01 1.88141242e-02
2.39725515e-01 1.24542927e-02 7.64292255e-02 1.80595681e-01
-1.99557737e-01 3.18203121e-01 1.25268269e-02 -3.00522268e-01
1.25683320e+00 9.66492891e-02 2.96646953e-01 7.29361713e-01
5.14078081e-01 -7.00441122e-01 -1.48621178e+00 -9.31506827e-02
2.39443049e-01 2.01767102e-01 1.32652689e-02 -1.01460254e+00
-9.20365870e-01 7.67671525e-01 5.61939478e-01 3.95685822e-01
1.12687695e+00 -7.44805038e-01 6.95383012e-01 7.65281916e-01
7.07121849e-01 -1.27428555e+00 1.31316915e-01 9.65280652e-01
9.29144919e-01 -7.28381097e-01 -2.52551705e-01 8.58853683e-02
-8.01871061e-01 1.18390584e+00 7.44161367e-01 -5.70654690e-01
5.07156849e-01 5.38759768e-01 -2.18679205e-01 1.69010192e-01
-1.46383274e+00 -2.14357957e-01 -3.03069443e-01 4.48654145e-01
3.03583980e-01 1.47893261e-02 -2.46412918e-01 1.80813372e-01
-8.49618912e-02 5.66410303e-01 6.15701675e-01 1.31043792e+00
-5.60291409e-01 -1.25477421e+00 -2.52617240e-01 1.61523312e-01
-1.37332976e-01 2.02294379e-01 -2.57634640e-01 1.01619470e+00
-1.65033355e-01 8.38263631e-01 9.13448483e-02 -6.10073805e-01
5.19893885e-01 2.53199130e-01 6.29394054e-01 -3.54865491e-01
-9.53116179e-01 -1.08480394e-01 1.60389781e-01 -5.39083779e-01
-4.12162058e-02 -5.77598274e-01 -1.42324638e+00 -2.78735995e-01
-2.79080262e-03 9.36159268e-02 5.91076195e-01 1.06004417e+00
4.84619915e-01 1.11027575e+00 8.15885544e-01 -8.90583515e-01
-1.27090514e+00 -9.77704942e-01 -4.55178559e-01 -2.02901170e-01
6.96318924e-01 -8.24627221e-01 2.69524287e-02 -4.05042022e-01] | [4.2574591636657715, 1.9062401056289673] |
21f97f27-04d5-445c-bc72-bac5ae722d7b | automatic-microscopic-cell-counting-by-use-of | 1903.00388 | null | http://arxiv.org/abs/1903.00388v2 | http://arxiv.org/pdf/1903.00388v2.pdf | Automatic microscopic cell counting by use of unsupervised adversarial domain adaptation and supervised density regression | Accurate cell counting in microscopic images is important for medical
diagnoses and biological studies. However, manual cell counting is very
time-consuming, tedious, and prone to subjective errors. We propose a new
density regression-based method for automatic cell counting that reduces the
need to manually annotate experimental images. A supervised learning-based
density regression model (DRM) is trained with annotated synthetic images (the
source domain) and their corresponding ground truth density maps. A domain
adaptation model (DAM) is built to map experimental images (the target domain)
to the feature space of the source domain. By use of the unsupervised
learning-based DAM and supervised learning-based DRM, a cell density map of a
given target image can be estimated, from which the number of cells can be
counted. Results from experimental immunofluorescent microscopic images of
human embryonic stem cells demonstrate the promising performance of the
proposed counting method. | ['Lilianna Solnica-Krezel', 'Kyaw Thu Minn', 'Shenghua He', 'Mark Anastasio', 'Hua Li'] | 2019-03-01 | null | null | null | null | ['automatic-cell-counting'] | ['miscellaneous'] | [ 8.87836888e-02 -8.59979615e-02 3.55632440e-03 -2.37390757e-01
-6.35607719e-01 -1.62818432e-01 2.69320518e-01 5.26169479e-01
-8.00804734e-01 1.23272216e+00 -4.14744824e-01 -2.76863948e-02
3.97551328e-01 -1.03217709e+00 -4.60062593e-01 -9.78759825e-01
3.60716343e-01 1.01861787e+00 3.93710077e-01 4.76128370e-01
2.82397121e-01 5.60096562e-01 -1.14510286e+00 -3.50634664e-01
9.76736784e-01 7.12296307e-01 5.60764432e-01 8.36063087e-01
-2.72714764e-01 8.65953982e-01 -6.68870270e-01 1.52429223e-01
-1.17357783e-01 -5.71852863e-01 -6.40136898e-01 3.87787312e-01
-2.38198116e-01 -3.06159765e-01 1.54088631e-01 1.04602456e+00
3.08380872e-01 -5.21311983e-02 1.35071146e+00 -9.84080076e-01
-3.23986828e-01 -2.18644962e-01 -7.69448280e-01 2.08502203e-01
-4.05376554e-02 -1.40034303e-01 1.20986104e-01 -9.91496086e-01
7.59254336e-01 8.41395140e-01 3.72606039e-01 6.17974401e-01
-1.52020431e+00 -5.39401352e-01 -4.68022913e-01 -2.46975511e-01
-1.71611178e+00 -6.93467110e-02 3.55149448e-01 -7.30198205e-01
5.92030823e-01 -3.82862955e-01 8.19661319e-01 2.12476879e-01
3.60531926e-01 3.86272609e-01 1.28781724e+00 -7.47086048e-01
8.93472493e-01 5.57159066e-01 -1.60828903e-01 9.51324284e-01
5.03539681e-01 -2.27533519e-01 -1.47907376e-01 -5.68020083e-02
1.29556072e+00 5.67870727e-03 -8.66164863e-02 -4.14073259e-01
-9.73613441e-01 8.27867210e-01 2.18373701e-01 6.60771787e-01
-2.08015114e-01 -4.97306623e-02 2.14757368e-01 -3.37148726e-01
6.50062740e-01 2.60431021e-01 -2.35052988e-01 -1.82249565e-02
-1.18403387e+00 1.54894497e-02 7.15470672e-01 8.65873992e-01
1.05709100e+00 -9.82812643e-02 -6.33997545e-02 1.00089157e+00
3.08983922e-01 5.25599360e-01 4.94963855e-01 -9.10191536e-01
-2.97170162e-01 1.00437796e+00 2.90062338e-01 -8.84855747e-01
-3.28879029e-01 2.59062294e-02 -9.22273874e-01 3.55315655e-01
8.35559130e-01 1.24777332e-01 -8.76277208e-01 1.38941467e+00
5.27960122e-01 3.04886252e-01 1.39131490e-02 7.94453621e-01
3.43850166e-01 7.18613684e-01 2.89671957e-01 -5.80127597e-01
8.49809349e-01 -5.04650116e-01 -5.30607164e-01 4.13823202e-02
9.34843779e-01 -2.91579127e-01 9.53723371e-01 8.41098949e-02
-6.91910565e-01 -1.56087279e-01 -1.07610321e+00 -9.70735028e-03
-3.06228459e-01 4.72696513e-01 2.21751973e-01 4.41574335e-01
-7.43310571e-01 2.68612653e-01 -1.10335207e+00 -6.59174860e-01
7.64162898e-01 4.98887450e-01 -4.27975178e-01 -8.63373429e-02
-3.55387449e-01 8.24501216e-01 4.69125748e-01 -3.49989563e-01
-7.55366087e-01 -5.87929666e-01 -8.65462422e-01 -1.51328191e-01
-2.95521438e-01 -6.42724276e-01 9.03638303e-01 -5.79037607e-01
-1.50848091e+00 1.20950520e+00 -4.63794112e-01 -3.67978275e-01
1.44556865e-01 4.85425651e-01 1.39602721e-01 4.11264509e-01
2.70677984e-01 7.71291792e-01 4.68643606e-01 -1.44575310e+00
-7.31124640e-01 -3.15061152e-01 -5.31145275e-01 -8.02195165e-04
-2.29340434e-01 -2.14463741e-01 -2.56789863e-01 -4.38974887e-01
5.98700065e-03 -5.33129275e-01 -2.69734263e-01 2.39897221e-01
-2.27035895e-01 1.64041013e-01 7.97805548e-01 -5.21768808e-01
9.70873833e-01 -1.96677053e+00 9.61846337e-02 2.20829993e-01
4.21361327e-01 1.54074326e-01 1.67118698e-01 -1.56581566e-01
4.68630731e-01 -1.11015365e-01 -5.46612501e-01 -2.79954612e-01
-6.11460924e-01 1.87962994e-01 3.74427497e-01 6.60862207e-01
2.40275651e-01 5.19296527e-01 -1.03577185e+00 -1.14221013e+00
5.80080211e-01 6.88809752e-01 -3.67650777e-01 3.85066390e-01
-1.90908298e-01 1.03011966e+00 -3.14607620e-01 6.40340626e-01
7.16119826e-01 -5.15445828e-01 1.60729393e-01 -3.41017246e-02
1.05705415e-03 -4.15837616e-01 -6.73451006e-01 1.19925296e+00
-5.05475640e-01 3.79012942e-01 -2.26789504e-01 -9.05735195e-01
1.15577292e+00 -5.70949614e-02 6.38383687e-01 -4.48125154e-01
6.07900560e-01 4.85859841e-01 -3.26645970e-01 -1.69913709e-01
1.92636196e-02 -7.26167858e-01 -4.30482626e-02 3.56710494e-01
4.19582069e-01 -5.23195386e-01 4.04293627e-01 1.26143694e-01
8.71611297e-01 -1.01676211e-01 7.66903043e-01 -4.95388180e-01
8.19229782e-01 1.70612022e-01 6.06534839e-01 6.77953809e-02
-2.00881004e-01 6.18581653e-01 4.61398512e-01 -5.10940492e-01
-1.33919609e+00 -9.68060613e-01 -3.47830564e-01 3.87412220e-01
2.16794938e-01 1.33578002e-01 -1.11873806e+00 -4.27951574e-01
-7.14867711e-02 6.01827085e-01 -7.05826879e-01 1.18657947e-02
-1.40043691e-01 -8.77577424e-01 3.26070845e-01 4.63450491e-01
4.96153623e-01 -6.97093427e-01 -4.91821259e-01 3.16208124e-01
-2.70272732e-01 -1.20999932e+00 -1.75121725e-01 5.05402945e-02
-9.83548999e-01 -1.06397486e+00 -1.18018436e+00 -1.00608122e+00
1.25566375e+00 -7.14950785e-02 7.49765038e-01 1.92527264e-01
-5.06034076e-01 1.72326282e-01 -2.26605814e-02 -6.90746605e-01
-7.66443074e-01 -1.67573124e-01 1.78894714e-01 -9.51510370e-02
7.48527050e-01 -4.88650918e-01 -6.01295948e-01 4.27245885e-01
-7.36582279e-01 2.36387625e-02 2.98029006e-01 9.23206270e-01
1.26690352e+00 5.59134148e-02 8.10045362e-01 -1.06622279e+00
2.88096935e-01 -2.75498301e-01 -1.04806387e+00 1.31018952e-01
-5.65054476e-01 -4.39118184e-02 7.13914216e-01 -2.87710458e-01
-1.06075096e+00 3.72786880e-01 1.96855664e-01 -3.89180571e-01
-1.85877189e-01 3.34433228e-01 7.80580118e-02 -1.55308485e-01
7.99364686e-01 3.13898087e-01 4.33378756e-01 -4.47199270e-02
-1.87845215e-01 7.39796817e-01 6.70218587e-01 -3.56782854e-01
5.49698353e-01 6.65612578e-01 3.35918039e-01 -9.98406887e-01
-7.25017130e-01 -5.45862436e-01 -8.28963041e-01 -3.85923713e-01
9.34505999e-01 -8.29348028e-01 -4.85484391e-01 7.16660202e-01
-9.89893734e-01 -5.47972083e-01 -2.86349148e-01 7.37197816e-01
-7.95087814e-01 3.59818578e-01 -7.99973369e-01 -8.89006376e-01
-1.76450387e-01 -7.78729320e-01 1.13593698e+00 6.45793021e-01
-1.22091345e-01 -1.38826108e+00 4.98410404e-01 2.09963620e-01
1.96342528e-01 3.48650277e-01 1.23346543e+00 -2.97967017e-01
-4.22937810e-01 -6.37685776e-01 -4.61841822e-01 2.54210562e-01
3.28936219e-01 2.13790432e-01 -8.90908062e-01 -8.95837322e-02
-9.23282560e-03 -3.53475600e-01 5.85516274e-01 8.91122282e-01
9.76631224e-01 2.20876373e-02 -7.23518729e-01 4.68065053e-01
1.77999365e+00 1.81090266e-01 7.93006778e-01 6.82110712e-02
4.38696444e-01 3.15152526e-01 7.96424091e-01 4.66101795e-01
2.65957177e-01 3.95656615e-01 -5.73798120e-02 -3.07432652e-01
6.92366064e-02 -3.49529445e-01 -2.80258983e-01 8.26083481e-01
-2.57999331e-01 -1.65738910e-01 -8.24372232e-01 6.43237710e-01
-1.50783288e+00 -5.50717056e-01 6.79668859e-02 2.51455617e+00
9.29296136e-01 -7.11940452e-02 3.53531688e-01 2.98518091e-01
9.79088962e-01 -7.68112242e-01 -5.39911568e-01 -8.61314312e-02
1.18289135e-01 2.09288597e-01 3.96607876e-01 5.66517830e-01
-8.58030558e-01 7.33991265e-01 6.41591072e+00 8.10796499e-01
-1.05861330e+00 -6.57561272e-02 1.11984503e+00 1.46886572e-01
1.17259771e-01 -2.50460714e-01 -7.44314253e-01 5.34715474e-01
7.50873089e-01 -4.51636434e-01 1.67763546e-01 7.55796254e-01
2.47561246e-01 -9.12439227e-01 -9.83684421e-01 1.03405035e+00
-1.11383222e-01 -1.46002614e+00 9.28578749e-02 3.95690411e-01
7.04774439e-01 -4.48526442e-01 -4.13584232e-01 1.99501857e-01
2.78082728e-01 -1.00749660e+00 1.92157134e-01 6.50912523e-01
1.33605337e+00 -7.89073348e-01 1.19848728e+00 7.89577961e-01
-1.06817973e+00 3.25052857e-01 -7.60344028e-01 1.75633997e-01
1.37490943e-01 9.10050929e-01 -1.18984973e+00 -1.16076238e-01
4.09582496e-01 6.22860968e-01 -3.50398451e-01 1.02541828e+00
2.39695147e-01 4.38247919e-01 -4.61926281e-01 -2.27268234e-01
-2.76522458e-01 -4.52985883e-01 8.96785855e-02 1.19462013e+00
5.03863335e-01 2.33414918e-02 -1.32472426e-01 1.12144136e+00
-1.46409646e-01 2.32191116e-01 -7.21215606e-01 -2.14423209e-01
5.29995501e-01 1.55283773e+00 -1.51172757e+00 -2.94600815e-01
-1.87809706e-01 8.05491567e-01 5.95466018e-01 1.77305967e-01
-6.94774747e-01 -1.89131320e-01 -1.23565361e-01 4.16463792e-01
3.40421759e-02 -1.42352674e-02 -3.51786464e-01 -8.60409737e-01
-5.16269624e-01 -1.92633763e-01 1.84984058e-01 -6.27252936e-01
-1.26118541e+00 1.96069703e-01 3.91894206e-02 -1.45285749e+00
-9.02996510e-02 -6.24831676e-01 -5.08970320e-01 8.41208696e-01
-1.22099912e+00 -5.78403592e-01 -4.44743723e-01 2.79259950e-01
2.62943983e-01 -2.37465948e-01 1.11839724e+00 1.92185864e-01
-5.61285973e-01 2.54474550e-01 9.70960557e-02 2.25274369e-01
3.22683603e-01 -1.37236643e+00 -3.59452635e-01 3.93474132e-01
-1.77540392e-01 2.63888717e-01 5.83412230e-01 -6.07783556e-01
-5.44558823e-01 -1.31329095e+00 7.86522031e-01 -2.34305814e-01
2.53305525e-01 -1.44688442e-01 -9.28627312e-01 2.24031791e-01
-4.24188823e-01 3.66050333e-01 9.55332994e-01 -6.00074351e-01
1.63910076e-01 7.37095773e-02 -1.65616393e+00 4.33273137e-01
1.38698220e-01 -3.65334421e-01 1.24239825e-01 3.38614553e-01
6.77150115e-02 -2.23514989e-01 -1.07956982e+00 7.41989017e-02
4.82807100e-01 -8.30822408e-01 5.58156133e-01 2.39114091e-02
4.03881669e-01 -5.87982893e-01 2.66627278e-02 -1.13549471e+00
-1.47316813e-01 3.12882692e-01 1.26938820e-01 1.01328337e+00
3.68963301e-01 -4.29443598e-01 1.08748281e+00 4.14183140e-01
4.65449780e-01 -8.05627108e-01 -1.07627296e+00 -7.53127277e-01
1.53927773e-01 2.13099703e-01 1.37778983e-01 7.37864196e-01
3.48041624e-01 4.69906986e-01 2.62549132e-01 -1.26275524e-01
7.53096104e-01 -4.72817719e-01 7.85027683e-01 -1.43235874e+00
-2.27769576e-02 5.28373662e-03 -8.97718728e-01 -6.97004855e-01
3.87950279e-02 -6.53402746e-01 3.04949641e-01 -1.73499835e+00
5.99529505e-01 -3.87206614e-01 -6.95529655e-02 4.25547361e-02
-3.17164175e-02 4.96967256e-01 -3.21901590e-01 4.35193747e-01
-4.16104436e-01 3.06160390e-01 1.31985331e+00 -1.14513576e-01
-2.11810961e-01 -1.29477037e-02 -2.31598154e-01 8.53336751e-01
7.99200237e-01 -7.24613547e-01 -4.73715276e-01 1.09464280e-01
-9.83081535e-02 2.49045134e-01 3.26200008e-01 -1.27474034e+00
1.52603328e-01 -1.09153427e-01 6.32143855e-01 -4.13301140e-01
2.54986197e-01 -7.40551710e-01 -1.31591246e-01 4.51007545e-01
-1.49760157e-01 -5.58119833e-01 1.24196872e-01 7.93013871e-01
-2.79262632e-01 -4.30483013e-01 1.39011550e+00 -1.86002925e-01
-1.30868360e-01 1.70764491e-01 -7.14733183e-01 -9.26377103e-02
1.30718184e+00 -6.39004290e-01 -1.86234027e-01 -2.57976979e-01
-5.95184565e-01 -2.05122232e-01 8.92276943e-01 -6.40940845e-01
7.47629523e-01 -1.31733990e+00 -5.08761823e-01 6.61605150e-02
4.02829856e-01 4.79007751e-01 8.20050240e-02 9.39765453e-01
-1.09383583e+00 1.67548910e-01 -3.62233728e-01 -9.05443013e-01
-9.38476920e-01 4.63284194e-01 4.76135671e-01 -5.44988573e-01
-3.09240371e-01 7.54979432e-01 3.54343921e-01 -3.97924066e-01
-2.36831918e-01 -4.23778683e-01 -3.45204085e-01 -5.83860755e-01
5.67116201e-01 4.97292727e-01 -1.29460931e-01 -7.62671292e-01
-2.10067287e-01 6.58823669e-01 -9.06766281e-02 -1.34665370e-01
1.12925708e+00 -1.16149433e-01 -1.10010087e-01 6.25375450e-01
1.18304265e+00 -2.75288641e-01 -1.34670317e+00 -1.23908035e-01
-5.79617918e-02 -4.66042519e-01 1.19079567e-01 -4.74088311e-01
-9.60472584e-01 8.08773518e-01 5.59350193e-01 -1.46071181e-01
1.07795560e+00 -3.32385302e-02 4.35935557e-01 2.61550933e-01
8.02003264e-01 -1.33222008e+00 5.63199678e-03 1.58951238e-01
3.77593040e-01 -1.20879388e+00 1.94922060e-01 -6.87261283e-01
-3.63465220e-01 9.65360105e-01 6.42541945e-01 -2.79891759e-01
8.13647270e-01 6.65741384e-01 -5.90351447e-02 -1.59641802e-01
-4.99231219e-01 1.53900683e-01 -1.79709807e-01 1.09747565e+00
4.40434694e-01 -9.88480374e-02 -1.94125384e-01 4.04786080e-01
2.88805842e-01 7.51679659e-01 6.25291348e-01 8.78990531e-01
-6.76930845e-01 -8.66559803e-01 -3.75941783e-01 7.54339814e-01
-2.60792166e-01 2.94107854e-01 -2.40679502e-01 6.80635631e-01
1.75452065e-02 6.82570338e-01 4.14434910e-01 -1.56036258e-01
-1.03752390e-01 -1.02577262e-01 7.88377762e-01 -8.32370639e-01
1.75184384e-01 2.18572572e-01 -4.02050734e-01 1.05858013e-01
-7.00707495e-01 -3.62371385e-01 -1.72659421e+00 -2.47735575e-01
-6.11681223e-01 -1.11354247e-03 4.32683766e-01 1.01910913e+00
-3.33554484e-02 1.61738828e-01 4.78735834e-01 -8.15303087e-01
2.88182527e-01 -9.82780397e-01 -1.18629348e+00 5.08947857e-02
-1.55857978e-02 -8.17860425e-01 -3.83097738e-01 6.87183380e-01] | [14.667957305908203, -3.191068649291992] |
9d77590d-c1fc-4781-b448-3d0823b6984b | equivariant-energy-guided-sde-for-inverse | 2209.15408 | null | https://arxiv.org/abs/2209.15408v3 | https://arxiv.org/pdf/2209.15408v3.pdf | Equivariant Energy-Guided SDE for Inverse Molecular Design | Inverse molecular design is critical in material science and drug discovery, where the generated molecules should satisfy certain desirable properties. In this paper, we propose equivariant energy-guided stochastic differential equations (EEGSDE), a flexible framework for controllable 3D molecule generation under the guidance of an energy function in diffusion models. Formally, we show that EEGSDE naturally exploits the geometric symmetry in 3D molecular conformation, as long as the energy function is invariant to orthogonal transformations. Empirically, under the guidance of designed energy functions, EEGSDE significantly improves the baseline on QM9, in inverse molecular design targeted to quantum properties and molecular structures. Furthermore, EEGSDE is able to generate molecules with multiple target properties by combining the corresponding energy functions linearly. | ['Jun Zhu', 'Chongxuan Li', 'Peiyao Li', 'Zhongkai Hao', 'Min Zhao', 'Fan Bao'] | 2022-09-30 | null | null | null | null | ['3d-molecule-generation'] | ['medical'] | [ 2.09600896e-01 7.94712547e-03 -2.24879816e-01 2.68751085e-02
-3.18034858e-01 -8.91561270e-01 6.55270517e-01 -5.16515598e-02
-1.45372534e-02 1.15453041e+00 3.33211631e-01 -4.04238284e-01
-3.79497528e-01 -9.90180433e-01 -8.56361508e-01 -1.13481820e+00
-9.64164957e-02 2.04881262e-02 -2.67832398e-01 -5.76605737e-01
5.65274179e-01 7.90831625e-01 -6.88286126e-01 -1.78226680e-01
1.22204494e+00 3.85335237e-01 1.82678588e-02 5.19844174e-01
2.43639529e-01 2.87524432e-01 -5.51742315e-02 -2.66493022e-01
2.86305010e-01 -7.14545310e-01 -5.90256631e-01 -2.23841354e-01
2.61824399e-01 1.61963835e-01 -4.99060124e-01 9.69763160e-01
7.64403105e-01 4.85139400e-01 1.24427450e+00 -6.96996629e-01
-1.41003072e+00 2.95913488e-01 -6.93468824e-02 -7.17820004e-02
3.83830607e-01 6.68038964e-01 9.27772939e-01 -1.22763395e+00
1.05356061e+00 1.06730759e+00 2.71103621e-01 9.14978445e-01
-1.75015581e+00 -5.13913870e-01 -2.37763271e-01 -1.13276802e-01
-1.30174983e+00 -2.59890795e-01 6.90950751e-01 -5.73060095e-01
9.82234716e-01 2.35071689e-01 5.22269249e-01 1.18451965e+00
9.79164541e-01 2.55412132e-01 9.89000142e-01 -2.38692269e-01
7.68659174e-01 -2.79354788e-02 7.36159533e-02 6.83000267e-01
4.59135026e-01 4.71007138e-01 -6.92124665e-01 -4.02448475e-01
6.73464656e-01 -2.96375696e-02 -3.52237344e-01 -7.78570890e-01
-1.42487407e+00 9.73040581e-01 4.62432653e-01 1.57944709e-02
-7.03264654e-01 9.49920118e-02 -1.14338279e-01 2.05716360e-02
2.49641702e-01 9.64828908e-01 -2.83021748e-01 1.02738135e-01
-3.92756194e-01 4.68183041e-01 6.44336939e-01 9.50416267e-01
6.00468695e-01 2.20771506e-01 -3.37597400e-01 -6.95404559e-02
3.92939657e-01 8.49906266e-01 -2.56310385e-02 -1.04132867e+00
2.21876785e-01 4.27036345e-01 2.09872514e-01 -7.33293533e-01
-4.05146748e-01 -5.79362154e-01 -1.16013694e+00 4.02470753e-02
-1.16771050e-02 -2.08357722e-01 -6.83896601e-01 1.96558428e+00
5.23385644e-01 -1.77198082e-01 3.84245008e-01 6.30398035e-01
5.50699294e-01 7.36591280e-01 6.20843060e-02 -5.28539717e-01
6.54341638e-01 -4.53000784e-01 -7.19580829e-01 4.98941749e-01
7.50690997e-01 -4.20256585e-01 9.05570865e-01 1.30173102e-01
-1.07487595e+00 -2.91759055e-02 -9.00771916e-01 1.22991107e-01
-2.39165545e-01 -2.40827158e-01 8.46637011e-01 6.12694263e-01
-9.50839043e-01 1.11855245e+00 -5.74130476e-01 9.30705592e-02
4.41317648e-01 6.76899195e-01 -5.81077337e-01 2.83407241e-01
-1.18812072e+00 7.53612459e-01 1.50970355e-01 6.67742714e-02
-9.68987942e-01 -1.07408834e+00 -6.48877382e-01 -1.61896139e-01
-8.60643238e-02 -1.33754265e+00 5.61882734e-01 -2.02400118e-01
-2.15604615e+00 3.05707663e-01 -2.01332137e-01 -3.90591115e-01
4.16519463e-01 1.73192039e-01 -2.79046502e-02 -3.76438140e-03
-6.99681118e-02 6.89013362e-01 6.78736150e-01 -7.58261740e-01
3.57063442e-01 -5.05134404e-01 -1.27898484e-01 2.76155733e-02
-1.70701265e-01 -5.10697961e-01 4.59129035e-01 -5.70969045e-01
-7.49325156e-02 -1.30001783e+00 -8.79593313e-01 -2.33910754e-01
-7.34892786e-01 -1.18019104e-01 4.45985973e-01 -1.37008384e-01
1.01157248e+00 -1.73147178e+00 1.01764500e+00 5.14113188e-01
3.14344943e-01 4.71820608e-02 -2.51516968e-01 7.39169002e-01
-2.93581098e-01 4.57212538e-01 -3.78976941e-01 3.69052470e-01
7.37159476e-02 -2.15006381e-01 -4.20984179e-01 5.26800573e-01
2.07841858e-01 1.27623284e+00 -9.66977656e-01 2.13197879e-02
2.72947215e-02 7.84056008e-01 -1.11140406e+00 -1.23062208e-01
-3.83837342e-01 8.16348553e-01 -9.84625876e-01 3.94470632e-01
6.88395917e-01 -1.15457661e-01 5.45483343e-02 -2.17601016e-01
-1.54383719e-01 3.87315042e-02 -7.84167349e-01 1.72013366e+00
-1.88370287e-01 1.58598155e-01 -3.95777941e-01 -5.29301763e-01
7.66290069e-01 1.38775527e-01 6.69829428e-01 -4.67308611e-01
-1.26254633e-01 3.03048283e-01 2.39984155e-01 -1.43647864e-01
1.42733768e-01 -4.54382330e-01 6.69371560e-02 3.42467278e-01
5.39758801e-02 -5.53732693e-01 8.78019352e-03 3.33998561e-01
9.60143030e-01 1.24110892e-01 1.15133107e-01 -7.04648435e-01
6.01482451e-01 -2.01795936e-01 2.57779598e-01 6.45474851e-01
7.01806918e-02 3.22519034e-01 4.03054208e-01 -2.82448769e-01
-1.17055190e+00 -1.35819447e+00 -3.84444356e-01 3.08559328e-01
1.45294905e-01 -4.25807178e-01 -9.74073052e-01 -6.42710507e-01
-1.01369314e-01 7.21127868e-01 -6.17620111e-01 -7.57300735e-01
-4.42279965e-01 -1.03634667e+00 2.36062542e-01 -4.62236442e-02
9.97930691e-02 -8.42777014e-01 5.22779673e-02 3.55369002e-01
5.84867477e-01 -6.45097435e-01 -9.92296338e-01 8.11369866e-02
-8.96455407e-01 -8.10115099e-01 -9.15229380e-01 -2.56389558e-01
7.41904795e-01 3.08930546e-01 7.08388686e-01 -3.92216861e-01
-4.51202959e-01 2.73835868e-01 -1.09949827e-01 -1.65747195e-01
-4.87277985e-01 1.80923179e-01 4.93130594e-01 4.28759307e-02
-5.94384782e-02 -7.87051022e-01 -1.05815005e+00 9.51098651e-02
-9.08837080e-01 9.76567194e-02 4.57952678e-01 7.37593591e-01
8.52578461e-01 -2.09748447e-01 6.61963105e-01 -4.87600297e-01
7.83635259e-01 -3.94587487e-01 -5.01386285e-01 1.60380885e-01
-5.71313441e-01 8.10475767e-01 6.48392260e-01 -3.60363454e-01
-8.20139587e-01 2.21467748e-01 -3.43276352e-01 3.90504599e-02
1.59506321e-01 2.25608468e-01 -4.26185191e-01 -6.26355171e-01
8.54128301e-01 3.90270084e-01 -8.08752477e-02 -1.08675241e-01
7.14025497e-01 2.52197325e-01 -1.42649651e-01 -9.27600801e-01
8.43235195e-01 4.77918953e-01 7.74471641e-01 -1.00537395e+00
-5.61801970e-01 1.46433637e-01 -6.91068292e-01 1.60751998e-01
9.78465796e-01 -5.56777239e-01 -1.08119226e+00 9.95322764e-02
-1.12732685e+00 -2.16297805e-01 -5.06177664e-01 6.78653002e-01
-7.89162815e-01 1.72464773e-01 -3.48917961e-01 -6.78699553e-01
-5.90919316e-01 -1.49502361e+00 1.05750084e+00 1.24173887e-01
-7.02687502e-02 -1.14488339e+00 3.50724071e-01 -1.21753372e-01
3.88418227e-01 6.54518783e-01 1.12189496e+00 -1.37630478e-01
-8.84481788e-01 -9.00492892e-02 4.04620469e-01 9.59843472e-02
3.50621909e-01 1.57409415e-01 -3.98421288e-01 -1.68156236e-01
-1.44371822e-01 1.30211800e-01 8.28024268e-01 6.83995545e-01
9.51275051e-01 -4.29455251e-01 -3.17139834e-01 6.95640683e-01
1.09578919e+00 3.37310016e-01 5.17007232e-01 -8.78483206e-02
6.70339942e-01 4.01342660e-01 5.47117367e-02 4.64283288e-01
-6.23350851e-02 6.72447383e-01 1.81594461e-01 3.76980342e-02
1.75656155e-01 -4.59762067e-01 6.31082296e-01 7.45514989e-01
-3.12058359e-01 -2.48519644e-01 -4.29960877e-01 -1.08599767e-01
-1.46203697e+00 -1.10895371e+00 -4.72146302e-01 2.19414186e+00
1.01618028e+00 -2.02489808e-01 -5.64631410e-02 -4.03867394e-01
4.37888682e-01 -1.73034310e-01 -9.51495171e-01 -4.48415518e-01
-4.40403283e-01 6.00444198e-01 6.09944880e-01 8.06788623e-01
-5.66645920e-01 9.49180484e-01 7.25055170e+00 9.71099138e-01
-9.89854515e-01 -1.23440653e-01 5.22193193e-01 9.05450201e-04
-9.80728447e-01 2.46504828e-01 -9.00300980e-01 3.79327118e-01
8.29669058e-01 -5.14272749e-01 4.12524909e-01 4.53636974e-01
6.72100544e-01 2.78492033e-01 -9.87036765e-01 5.16107023e-01
-4.33951914e-01 -2.08344412e+00 5.60555518e-01 5.85246742e-01
1.20889342e+00 -4.24886256e-01 6.45909607e-01 -2.09182084e-01
4.54441309e-02 -1.21637464e+00 4.18460160e-01 7.60432243e-01
7.03755140e-01 -9.49541926e-01 1.89895332e-01 2.12422088e-01
-7.33893216e-01 4.08562630e-01 -2.97997922e-01 3.02949697e-01
2.98156679e-01 5.33384800e-01 -5.32350540e-01 4.50607389e-01
-1.93185017e-01 8.29240799e-01 -1.07280143e-01 5.83112717e-01
-1.23868100e-01 2.80848831e-01 -1.66328534e-01 -3.93812448e-01
2.13861257e-01 -1.10947061e+00 8.63578081e-01 8.76495898e-01
3.58639419e-01 3.52776468e-01 -2.46513262e-01 1.41827130e+00
-1.49908274e-01 2.11074039e-01 -9.01180029e-01 -3.43761355e-01
1.13318637e-01 8.46810997e-01 -3.56052220e-01 4.93589975e-02
2.52734840e-01 1.00909412e+00 -3.24791268e-04 7.49194622e-01
-8.85791600e-01 -5.59549570e-01 1.08703232e+00 4.86114733e-02
5.13414890e-02 -7.21213460e-01 -6.25177547e-02 -1.27886510e+00
-2.14411318e-01 -7.03143120e-01 -3.35631549e-01 -4.97075051e-01
-1.11605453e+00 4.62002367e-01 -3.91001791e-01 -8.52247536e-01
1.29084930e-01 -7.95243204e-01 -4.90099132e-01 7.62100458e-01
-1.16787922e+00 -7.04629421e-01 3.60049754e-01 4.79214281e-01
1.76763043e-01 -4.78177443e-02 9.57072318e-01 6.29377663e-02
-6.58068419e-01 3.71585906e-01 5.40830791e-01 -6.12444997e-01
4.92032826e-01 -1.31418574e+00 4.56381291e-01 4.85594839e-01
1.93271369e-01 1.09575963e+00 8.27906132e-01 -8.13613713e-01
-2.06529093e+00 -1.03935623e+00 4.01804507e-01 -8.17398190e-01
6.78583443e-01 -3.69865417e-01 -5.35920620e-01 1.75743967e-01
8.71658325e-02 4.15739082e-02 8.68416250e-01 -4.01584834e-01
-4.47541028e-02 2.32683331e-01 -1.18992651e+00 7.93987215e-01
1.46858335e+00 -5.80406010e-01 -8.74338523e-02 8.60304475e-01
1.00695455e+00 -3.29389185e-01 -1.28173792e+00 3.80321503e-01
3.64063174e-01 -6.19092166e-01 1.35235810e+00 -1.27609253e+00
4.13999230e-01 -3.63769174e-01 -2.60643661e-01 -1.39798892e+00
-2.91645914e-01 -1.30291545e+00 -3.17456514e-01 6.05282724e-01
7.41175830e-01 -6.59707963e-01 6.97147906e-01 7.86605299e-01
5.65888286e-02 -8.41162503e-01 -9.01884794e-01 -9.57413733e-01
6.80075169e-01 -3.73024307e-02 5.19630134e-01 5.46484590e-01
-1.71243306e-02 4.64882672e-01 -3.60841930e-01 7.98047855e-02
5.38995564e-01 5.66299744e-02 5.32895565e-01 -7.98540115e-01
-2.77891248e-01 -5.29104948e-01 -2.44305670e-01 -1.13477135e+00
3.77077788e-01 -1.13907063e+00 -4.15404022e-01 -1.30150557e+00
3.02614808e-01 -2.77522087e-01 -5.12834340e-02 -2.94619679e-01
-9.83879119e-02 -2.70471387e-02 1.06346607e-01 2.41234619e-03
-3.22250605e-01 1.35957015e+00 1.80777872e+00 -1.08214699e-01
-7.18224645e-01 1.51051790e-01 -7.20461011e-01 2.05556825e-01
6.92576945e-01 -4.17712480e-01 -3.31100523e-01 1.29779220e-01
5.85685134e-01 1.12905793e-01 8.25301707e-02 -5.15707076e-01
-5.21121994e-02 -5.49928546e-01 3.03471923e-01 -1.03871867e-01
2.96529859e-01 -3.99168968e-01 3.44044000e-01 7.86086023e-01
-4.19218898e-01 -2.10584939e-01 -1.09499432e-01 7.73686230e-01
1.61918893e-01 4.40586284e-02 7.87159324e-01 5.03040068e-02
6.39387295e-02 8.69317830e-01 -5.45199871e-01 1.21918237e-02
1.13643026e+00 -1.46921739e-01 -1.04444539e-02 -1.76112279e-02
-6.95239484e-01 -1.77384734e-01 5.66954613e-01 1.03704214e-01
7.81822443e-01 -1.42805171e+00 -5.86724997e-01 1.41007915e-01
-1.01447858e-01 -2.30684564e-01 2.62191117e-01 8.47105443e-01
-4.67414856e-01 6.77636802e-01 2.28811473e-01 -4.37831223e-01
-6.95378900e-01 7.02195406e-01 5.94032943e-01 -2.82897856e-02
-2.58982390e-01 6.97424948e-01 5.77966690e-01 -5.63852191e-01
-3.64644766e-01 -3.01428318e-01 2.92638004e-01 -2.24372461e-01
4.01037723e-01 1.80456877e-01 -3.81313562e-02 -4.88524824e-01
-3.86987865e-01 9.83017981e-01 1.34462938e-01 -2.70871539e-02
1.43572593e+00 1.82725698e-01 -5.06864935e-02 -2.65030444e-01
1.34541500e+00 2.86926001e-01 -1.28480172e+00 2.10980311e-01
-8.98519680e-02 -1.95348814e-01 1.10929348e-01 -3.51075590e-01
-6.24752343e-01 8.08083296e-01 4.31247145e-01 -2.98761398e-01
4.82384741e-01 -1.94425106e-01 6.83769524e-01 6.94763720e-01
5.26470244e-01 -6.11338377e-01 2.77426243e-01 4.86631930e-01
1.00640690e+00 -9.06679809e-01 -9.74412728e-03 -3.14470172e-01
-5.44189453e-01 1.26691806e+00 2.17044011e-01 -2.50317127e-01
7.39312768e-01 -1.58333614e-01 -7.60143936e-01 -2.04301745e-01
-6.32371485e-01 -5.22082672e-04 4.60766494e-01 6.71732485e-01
4.93617356e-01 1.48008034e-01 -4.09402579e-01 2.57152200e-01
-7.22703338e-02 -2.95114547e-01 5.14833570e-01 6.40467107e-01
-3.37385595e-01 -1.49768245e+00 7.58604193e-03 -1.14646100e-01
-8.55743438e-02 -4.24661189e-01 -7.25495279e-01 4.31549519e-01
-5.06673381e-02 6.08663380e-01 -4.58898872e-01 -7.45566189e-02
4.22091305e-01 -3.85101229e-01 7.82306492e-01 -5.69970012e-01
-2.13407338e-01 -2.17187747e-01 -3.81050408e-01 -5.68357050e-01
-1.80330381e-01 -4.31866050e-01 -1.30956590e+00 -4.86320347e-01
-5.97686172e-01 4.84160781e-01 5.32240510e-01 8.30892861e-01
1.06728327e+00 3.60116810e-01 9.75323677e-01 -7.60329366e-01
-6.24812007e-01 -3.06107789e-01 -4.57418412e-01 4.85137589e-02
4.77233589e-01 -5.11728287e-01 -2.60057747e-01 -8.79871771e-02] | [5.082918643951416, 5.684427261352539] |
aaa39187-adc6-4aef-9448-9ce457b4de4f | bplf-a-bi-parallel-linear-flow-model-for | 2106.07563 | null | https://arxiv.org/abs/2106.07563v1 | https://arxiv.org/pdf/2106.07563v1.pdf | BPLF: A Bi-Parallel Linear Flow Model for Facial Expression Generation from Emotion Set Images | The flow-based generative model is a deep learning generative model, which obtains the ability to generate data by explicitly learning the data distribution. Theoretically its ability to restore data is stronger than other generative models. However, its implementation has many limitations, including limited model design, too many model parameters and tedious calculation. In this paper, a bi-parallel linear flow model for facial emotion generation from emotion set images is constructed, and a series of improvements have been made in terms of the expression ability of the model and the convergence speed in training. The model is mainly composed of several coupling layers superimposed to form a multi-scale structure, in which each coupling layer contains 1*1 reversible convolution and linear operation modules. Furthermore, this paper sorted out the current public data set of facial emotion images, made a new emotion data, and verified the model through this data set. The experimental results show that, under the traditional convolutional neural network, the 3-layer 3*3 convolution kernel is more conducive to extracte the features of the face images. The introduction of principal component decomposition can improve the convergence speed of the model. | ['Kunxian Shu', 'Xiaoming Yao', 'Shimei Xu', 'Lijia Yang', 'Siwei Liu', 'Yuanpeng Long', 'Gao Xu'] | 2021-05-27 | null | null | null | null | ['facial-expression-generation'] | ['computer-vision'] | [-3.20651680e-01 -1.08273968e-01 3.61518800e-01 -5.39704084e-01
2.63675600e-01 -8.62327311e-03 3.81991148e-01 -1.13727534e+00
-3.87420692e-03 7.26963401e-01 2.59298474e-01 1.52767077e-01
6.96710125e-02 -1.25212324e+00 -3.95494133e-01 -1.05776465e+00
-3.34996171e-02 -3.26542296e-02 -3.56623411e-01 -4.63577867e-01
-2.92267129e-02 7.80507445e-01 -1.78051436e+00 4.59607154e-01
7.90733695e-01 9.88458395e-01 1.22464001e-01 6.00174665e-01
-4.22535509e-01 1.00054896e+00 -6.11432910e-01 -3.47766042e-01
-2.11297385e-02 -8.83612096e-01 -6.48249328e-01 2.90572941e-01
-2.01749682e-01 -4.23589379e-01 -4.83520716e-01 9.27415669e-01
9.51035678e-01 3.34251300e-02 6.09706819e-01 -1.74476218e+00
-1.17216790e+00 2.38258034e-01 -3.02219212e-01 6.58538640e-02
4.99566272e-02 -8.38424452e-03 4.28291142e-01 -1.02184021e+00
6.01286888e-01 1.60101223e+00 6.35710359e-01 1.06231749e+00
-9.38302100e-01 -1.07997179e+00 -1.79404840e-01 2.83092916e-01
-1.30271506e+00 -3.50951195e-01 1.06831908e+00 -2.50412881e-01
8.96180570e-01 4.71423268e-02 1.13795185e+00 1.16418576e+00
3.69339317e-01 7.95225978e-01 1.23764884e+00 -1.19985387e-01
-4.16663177e-02 1.65256530e-01 -2.51556307e-01 9.53559041e-01
-1.45954266e-01 2.00368196e-01 -5.39796114e-01 3.35727260e-02
1.09057152e+00 8.22351500e-02 -3.66232634e-01 9.69653018e-03
-4.38614815e-01 9.62529659e-01 6.07564688e-01 4.73699987e-01
-2.57091552e-01 1.21388242e-01 2.21302286e-01 2.29460523e-01
5.86358428e-01 -1.13698706e-01 -2.20789790e-01 -7.99377412e-02
-8.83656859e-01 -6.21981733e-02 7.69814134e-01 6.27610743e-01
9.92022634e-01 6.85289502e-01 2.05478873e-02 6.25582159e-01
4.93586838e-01 5.93104422e-01 8.76531959e-01 -6.77707374e-01
-1.37950152e-01 8.47227693e-01 -5.33427954e-01 -1.36846519e+00
-4.10260707e-01 -2.73903608e-01 -1.42859232e+00 3.75444204e-01
-2.91679204e-01 -5.28929830e-01 -9.90965784e-01 1.79533446e+00
1.72712177e-01 3.26640517e-01 3.94428074e-01 1.00448668e+00
1.45885146e+00 1.20256078e+00 1.66637227e-02 -4.47447181e-01
1.03573644e+00 -7.51631320e-01 -1.20548773e+00 3.32771569e-01
2.68459767e-01 -1.00314236e+00 9.08393979e-01 3.17943394e-01
-8.62221837e-01 -9.64835763e-01 -7.13620365e-01 -3.24916616e-02
-4.10227984e-01 4.57849562e-01 1.23942912e+00 7.24862516e-01
-1.43683398e+00 4.35114294e-01 -5.87407589e-01 -2.43789569e-01
5.93141258e-01 3.31081241e-01 -4.22226101e-01 1.39411747e-01
-1.34451509e+00 6.27859414e-01 3.24744374e-01 5.96797585e-01
-8.46407056e-01 -5.82305610e-01 -6.96735740e-01 3.08727950e-01
-3.61555576e-01 -7.48682141e-01 6.54621840e-01 -1.42426920e+00
-1.89501441e+00 5.29375851e-01 -2.65661210e-01 3.04424882e-01
2.83980608e-01 -4.06304337e-02 -7.05954850e-01 2.97775231e-02
-3.43713552e-01 8.44817460e-01 8.17900956e-01 -1.26117694e+00
-3.41480762e-01 -1.19787790e-01 -3.23399603e-01 6.75435588e-02
-5.17759860e-01 2.53668278e-02 -1.84703231e-01 -4.45616931e-01
8.09568986e-02 -6.78695977e-01 -1.10729836e-01 -9.82595086e-02
8.89939740e-02 -2.66332656e-01 1.36033571e+00 -4.68404979e-01
1.22743607e+00 -2.14000869e+00 7.67269954e-02 1.80731148e-01
1.36833012e-01 3.24823707e-01 -1.26176953e-01 3.74046266e-01
-3.89243484e-01 2.04540834e-01 -8.85235593e-02 -4.33813073e-02
-2.14824647e-01 4.99500632e-01 -3.94828856e-01 1.61318243e-01
4.29774135e-01 1.17400849e+00 -4.89492029e-01 -5.64396203e-01
-2.75336057e-02 1.04584587e+00 -8.38843167e-01 4.43486720e-01
2.43469089e-01 1.37885109e-01 -3.25864047e-01 5.11429429e-01
1.03872859e+00 -1.20002218e-01 -1.43775761e-01 -3.94900948e-01
9.88400914e-03 -4.77330565e-01 -1.21120548e+00 1.48660469e+00
-3.26631218e-01 6.86665297e-01 5.79436570e-02 -9.67608154e-01
1.49978566e+00 4.80138808e-01 6.24746203e-01 -5.13157606e-01
3.31825525e-01 -1.01414531e-01 -3.84540819e-02 -1.01067626e+00
2.75810003e-01 -3.84459138e-01 3.48193288e-01 3.85315001e-01
3.99497181e-01 -2.02463284e-01 2.10031364e-02 6.43762723e-02
4.96538848e-01 3.96591932e-01 -2.34698609e-01 -3.57064515e-01
6.14998221e-01 -2.36437574e-01 8.34378541e-01 -3.38740125e-02
1.74204424e-01 4.48236167e-01 6.33225083e-01 -7.21888840e-01
-8.45835328e-01 -7.37225831e-01 -6.70411661e-02 6.67677999e-01
8.16752166e-02 -2.69269198e-01 -8.22592795e-01 -3.64986390e-01
-3.33053023e-01 2.21230596e-01 -7.53415346e-01 -5.32162428e-01
-4.50966567e-01 -1.54897940e+00 6.24709129e-01 4.53461170e-01
1.30482090e+00 -1.61994219e+00 -3.17989826e-01 1.50590599e-01
-3.37623060e-01 -6.19138598e-01 -1.50179893e-01 -3.30497056e-01
-9.51469004e-01 -1.01790893e+00 -4.35734868e-01 -1.17216957e+00
1.07131743e+00 3.47172888e-03 8.51436257e-01 2.44717062e-01
-5.31526089e-01 2.17833832e-01 -3.53609204e-01 -4.00022596e-01
-3.28390956e-01 -2.29009405e-01 -1.47863805e-01 3.62136304e-01
4.26744521e-01 -6.70207262e-01 -3.68555039e-01 1.51412738e-02
-1.13006616e+00 1.53371021e-01 7.47876048e-01 1.15271461e+00
2.41811335e-01 3.68534863e-01 5.45997024e-01 -5.29829919e-01
1.12145519e+00 -2.51059741e-01 -1.96313977e-01 7.53063634e-02
-7.45228291e-01 1.07786292e-02 6.11173987e-01 -4.33134168e-01
-1.65509224e+00 1.57284699e-02 -3.90715092e-01 -4.02020484e-01
-1.02937020e-01 3.30201149e-01 -4.69742507e-01 -8.67136568e-02
3.29352051e-01 5.73425591e-01 6.72050536e-01 -2.98428118e-01
3.56071800e-01 4.61861491e-01 3.11964720e-01 -4.46108997e-01
6.13993824e-01 3.46883833e-01 1.25008926e-01 -7.25482345e-01
-2.55064428e-01 1.86400622e-01 -4.87737685e-01 -6.47047579e-01
8.83395016e-01 -8.01470637e-01 -9.97739792e-01 9.07948971e-01
-1.25111127e+00 -9.48164612e-02 -1.89401969e-01 5.32202244e-01
-4.37672228e-01 5.12605645e-02 -9.92047489e-01 -5.81695616e-01
-6.64601803e-01 -1.11082709e+00 6.22844756e-01 8.16021740e-01
3.11016649e-01 -1.07513571e+00 -5.05968481e-02 -1.27854615e-01
7.78441012e-01 2.08139434e-01 9.06757295e-01 8.03543106e-02
-2.89544731e-01 1.29149660e-01 -1.32309645e-01 6.66012049e-01
2.80456543e-01 6.43917382e-01 -1.14569497e+00 -1.53582111e-01
3.05801302e-01 -3.36204797e-01 7.32169092e-01 3.52181703e-01
1.35999346e+00 -4.94003743e-01 -1.07482642e-01 1.02719963e+00
1.49486256e+00 4.46584105e-01 1.24867821e+00 3.81694809e-02
5.35963356e-01 5.20242155e-01 9.03811604e-02 3.57071251e-01
1.41045719e-01 5.45362048e-02 4.34542358e-01 -7.63673306e-01
-3.66759330e-01 -1.83215618e-01 4.00154412e-01 1.07725942e+00
-6.20114505e-01 -2.28524171e-02 -2.80649036e-01 1.59007043e-01
-1.66438115e+00 -1.47450852e+00 -2.15093225e-01 1.41963971e+00
9.02976513e-01 -3.52658570e-01 -2.49629736e-01 -3.04227062e-02
5.17224431e-01 -6.04822375e-02 -2.62930512e-01 -5.58512688e-01
-3.70427221e-01 4.34522659e-01 -3.11971337e-01 3.14518631e-01
-7.74636030e-01 1.14775825e+00 7.16724348e+00 8.66533756e-01
-1.67920697e+00 -1.48509443e-01 7.93891430e-01 -1.07679712e-02
-3.07077765e-01 -2.15715423e-01 -8.25411916e-01 5.16964734e-01
5.98875821e-01 -2.12120742e-01 3.59169483e-01 8.90109360e-01
2.27587104e-01 2.21112207e-01 -5.21000624e-01 1.27536392e+00
3.48881423e-01 -1.26005352e+00 3.83246571e-01 1.34790182e-01
7.17030227e-01 -5.88122725e-01 1.47645786e-01 3.92175525e-01
1.86788440e-01 -1.33901179e+00 1.73618183e-01 9.39592063e-01
7.13448703e-01 -1.09186125e+00 9.49348450e-01 1.54554486e-01
-1.01097298e+00 1.38607129e-01 -7.11540818e-01 -2.69607842e-01
-4.45262939e-02 5.76361179e-01 -5.50413728e-01 5.52674294e-01
9.79878545e-01 7.47735083e-01 -4.52623457e-01 6.68289840e-01
-3.43485206e-01 6.56829357e-01 -1.08784631e-01 -2.81088084e-01
1.20996438e-01 -5.67596436e-01 -1.87183954e-02 1.28754985e+00
6.28376186e-01 3.63929421e-01 -2.41992205e-01 1.06661797e+00
-9.09000915e-03 2.48960793e-01 -5.41920662e-01 1.52940899e-01
5.15290685e-02 1.80846477e+00 -5.70380449e-01 -3.08014244e-01
-1.81666896e-01 8.18744481e-01 6.90173358e-02 5.09883046e-01
-1.01325905e+00 -5.66044927e-01 7.20724940e-01 -2.63771236e-01
1.36129603e-01 6.04896173e-02 -3.43162827e-02 -1.12781155e+00
-2.30534390e-01 -6.61496639e-01 4.01410200e-02 -1.11149180e+00
-1.14342690e+00 9.43530083e-01 -2.71677077e-01 -9.91474450e-01
-1.88971117e-01 -7.94416904e-01 -8.44326258e-01 1.07646036e+00
-1.48713958e+00 -1.24631143e+00 -7.82510817e-01 1.01682305e+00
2.18004495e-01 -5.35010695e-01 1.07821894e+00 3.76613021e-01
-6.82895482e-01 3.86877537e-01 -1.36300042e-01 2.97908485e-01
5.20465493e-01 -7.06396163e-01 -3.50212932e-01 8.37613821e-01
-1.56450033e-01 4.99962807e-01 4.11830932e-01 -4.07799840e-01
-1.27115047e+00 -1.03160787e+00 9.12589312e-01 1.73489794e-01
2.31427461e-01 -2.98449904e-01 -1.00229406e+00 4.68230844e-01
4.77055043e-01 1.23760626e-01 8.40597212e-01 -1.24961831e-01
4.45484407e-02 -5.29556632e-01 -1.32804966e+00 4.09079939e-01
9.60655391e-01 -1.36415005e-01 -3.98602068e-01 1.04723290e-01
4.62363243e-01 -1.64539054e-01 -8.85871530e-01 4.46183056e-01
4.98126209e-01 -1.09661114e+00 7.19664693e-01 -6.88809633e-01
9.05953526e-01 -2.40936354e-01 2.38842949e-01 -1.41801941e+00
-7.25674450e-01 -5.17016053e-01 1.27597556e-01 1.48862696e+00
1.40370443e-01 -7.22024322e-01 7.57413208e-01 4.18671310e-01
-9.18225199e-02 -9.16658640e-01 -7.15315461e-01 -2.94471532e-01
-8.87178406e-02 1.21828890e-03 9.60124612e-01 9.40306127e-01
-2.08526358e-01 5.29431462e-01 -5.16004622e-01 -1.31812900e-01
1.74635559e-01 3.19647670e-01 7.36150682e-01 -1.15088797e+00
1.35778785e-01 -5.69505990e-01 -4.07752961e-01 -8.33336174e-01
2.44665191e-01 -8.84082139e-01 -2.55192220e-01 -1.65638447e+00
1.85139090e-01 -3.69216144e-01 2.46475334e-03 7.44342625e-01
-1.89613372e-01 2.62891710e-01 -8.47519711e-02 2.11479828e-01
1.09065972e-01 1.04530048e+00 1.86071455e+00 1.49706215e-01
-1.83290392e-01 -1.14928328e-01 -7.39705503e-01 6.05923772e-01
8.75219285e-01 -1.71045154e-01 -6.66425526e-01 -3.13221633e-01
5.75939566e-02 -1.84893295e-01 3.08857799e-01 -8.23438823e-01
1.97412387e-01 -2.18256533e-01 1.01613653e+00 -4.00366932e-01
3.57103467e-01 -8.26673448e-01 4.69294995e-01 5.88091016e-01
5.06570563e-02 6.87195361e-02 2.93567389e-01 -9.43443850e-02
-5.92470706e-01 -6.54731318e-02 7.44728804e-01 -1.81493387e-01
-9.62158740e-01 6.45179868e-01 -2.50800401e-01 -3.90771002e-01
1.03519297e+00 -2.63754576e-01 -1.98239431e-01 -3.16307038e-01
-7.08136499e-01 -2.58049369e-01 -2.32410058e-02 4.26271021e-01
9.55861330e-01 -1.81507444e+00 -7.65475392e-01 7.82134175e-01
-5.07501662e-01 1.62895732e-02 4.89565551e-01 4.67847615e-01
-7.13610232e-01 2.75869910e-02 -8.49740803e-01 -2.89899766e-01
-1.24324322e+00 4.29919273e-01 5.68938375e-01 1.86914895e-02
-5.38024247e-01 8.25768590e-01 2.51872778e-01 -3.93358201e-01
-2.21208587e-01 4.96367142e-02 -5.97745299e-01 1.33772165e-01
4.06817883e-01 3.33968639e-01 -2.46470362e-01 -7.55215406e-01
-1.00782931e-01 6.56603277e-01 2.97396034e-01 1.17654666e-01
1.42863607e+00 1.07788488e-01 -7.01099336e-01 -1.33263662e-01
1.28755760e+00 -3.25790256e-01 -1.15396833e+00 1.78812191e-01
-9.60568190e-01 -3.23845655e-01 6.79587275e-02 -6.04932547e-01
-1.83786654e+00 1.00668120e+00 6.08946264e-01 2.10079342e-01
1.68176234e+00 -3.54949027e-01 6.09963179e-01 3.28734927e-02
3.16509865e-02 -1.13131440e+00 3.38933587e-01 4.90536571e-01
1.02615499e+00 -9.14071262e-01 -1.79808825e-01 -3.72665673e-01
-5.33699811e-01 1.53990519e+00 1.04544556e+00 -2.10432082e-01
9.33780909e-01 5.81963599e-01 1.55118018e-01 -3.37955564e-01
-7.60114193e-01 7.94858392e-03 1.13455728e-01 6.22276425e-01
5.63105822e-01 -1.66205481e-01 -5.19750595e-01 7.32871652e-01
-7.02009737e-01 3.50186557e-01 4.73147780e-01 6.72597051e-01
-3.72347534e-01 -8.54819715e-01 -3.48790705e-01 9.21684504e-02
-3.00115049e-01 -1.71752498e-02 3.21897352e-03 8.44142675e-01
5.57932913e-01 8.57106149e-01 2.40424678e-01 -6.14425123e-01
2.75472224e-01 2.15359211e-01 4.40070927e-01 -1.60332903e-01
-3.06802332e-01 1.83746576e-01 -5.24979115e-01 -4.99225885e-01
-7.76934326e-01 -1.84567556e-01 -1.31064212e+00 -6.78825855e-01
-1.34471700e-01 3.46523255e-01 6.99171782e-01 6.96562529e-01
6.23801291e-01 7.81968892e-01 8.40271711e-01 -6.39413953e-01
-1.28427640e-01 -1.11658549e+00 -7.19789267e-01 4.65900719e-01
-1.69936866e-01 -5.31119585e-01 -4.61484671e-01 6.79834664e-01] | [13.576674461364746, 1.7141032218933105] |
b0774883-623a-41e0-893f-5368fbaac01f | empirically-validating-conformal-prediction | 2307.01088 | null | https://arxiv.org/abs/2307.01088v1 | https://arxiv.org/pdf/2307.01088v1.pdf | Empirically Validating Conformal Prediction on Modern Vision Architectures Under Distribution Shift and Long-tailed Data | Conformal prediction has emerged as a rigorous means of providing deep learning models with reliable uncertainty estimates and safety guarantees. Yet, its performance is known to degrade under distribution shift and long-tailed class distributions, which are often present in real world applications. Here, we characterize the performance of several post-hoc and training-based conformal prediction methods under these settings, providing the first empirical evaluation on large-scale datasets and models. We show that across numerous conformal methods and neural network families, performance greatly degrades under distribution shifts violating safety guarantees. Similarly, we show that in long-tailed settings the guarantees are frequently violated on many classes. Understanding the limitations of these methods is necessary for deployment in real world and safety-critical applications. | ['Graham W. Taylor', 'Kevin Kasa'] | 2023-07-03 | null | null | null | null | ['conformal-prediction', 'conformal-prediction'] | ['computer-vision', 'reasoning'] | [-2.78575607e-02 1.30662113e-01 -3.15526664e-01 -6.85235798e-01
-9.81632233e-01 -7.03672767e-01 6.27383292e-01 3.11281413e-01
-1.68337762e-01 9.72286761e-01 5.11936378e-03 -4.80839819e-01
-5.24760067e-01 -7.04662323e-01 -1.03650331e+00 -8.10386479e-01
-6.01209462e-01 5.96279442e-01 2.77388334e-01 1.50223970e-01
9.14872438e-02 5.28319955e-01 -1.26720524e+00 -5.49796708e-02
5.59358835e-01 1.35917962e+00 -6.74451828e-01 3.32898527e-01
3.70298654e-01 5.33880234e-01 -3.31852823e-01 -6.73819423e-01
3.56429577e-01 2.22382650e-01 -3.74676377e-01 -7.89120793e-01
8.98915291e-01 -5.14521241e-01 -2.95714319e-01 9.58554149e-01
6.08524323e-01 1.12433173e-01 1.11653101e+00 -1.67988586e+00
-4.13961977e-01 5.47704995e-01 -3.68600070e-01 1.61997333e-01
-2.70812571e-01 -1.57546237e-01 8.10914874e-01 -4.76740658e-01
7.31574893e-02 9.05846655e-01 1.39184904e+00 7.23150373e-01
-1.18896568e+00 -1.01843143e+00 -4.49817628e-02 -3.40343500e-03
-1.26840055e+00 -4.22452390e-01 3.57827634e-01 -6.16382003e-01
6.20251238e-01 5.28944749e-03 1.42295465e-01 1.47782350e+00
1.08517587e+00 4.44603682e-01 5.39344490e-01 1.34729266e-01
6.28500581e-01 6.70924634e-02 2.06177384e-01 1.95771307e-01
4.77853119e-01 8.30634475e-01 -5.09469271e-01 -6.97035789e-01
2.55142570e-01 -7.87841678e-02 -7.94028193e-02 -7.16738701e-01
-5.13834417e-01 1.01222491e+00 1.33534655e-01 -3.53672415e-01
-2.16131173e-02 7.50340164e-01 5.48469126e-01 -3.16804759e-02
7.42039323e-01 1.49166986e-01 -5.88884115e-01 -1.36729345e-01
-8.64285529e-01 6.24507606e-01 7.70422459e-01 1.33915293e+00
7.91548863e-02 -2.88088508e-02 -4.09546942e-01 4.95623559e-01
1.19657800e-01 5.77157557e-01 -1.23282485e-01 -9.85115588e-01
2.63469130e-01 -3.36232930e-01 2.27642417e-01 -9.15449500e-01
-8.29177976e-01 -9.82218385e-01 -9.55014825e-01 2.90481031e-01
4.51726407e-01 -2.59608567e-01 -6.18934214e-01 2.13083005e+00
1.97174549e-01 2.78289080e-01 -1.71266049e-01 4.39288676e-01
2.76541889e-01 1.78370625e-01 4.81425315e-01 -1.13945521e-01
7.64750838e-01 -4.19187635e-01 -5.72095096e-01 1.03514850e-01
6.35831475e-01 -3.74391288e-01 9.01252925e-01 5.23742974e-01
-8.92303884e-01 -6.96801618e-02 -1.12080503e+00 1.99452534e-01
-1.69214550e-02 -6.21651590e-01 5.82638919e-01 8.08435023e-01
-8.48376215e-01 9.02286947e-01 -9.06901360e-01 -4.19410653e-02
9.62491632e-01 1.62303537e-01 -1.98893219e-01 3.39740552e-02
-1.27518713e+00 1.05698061e+00 3.27514522e-02 -2.35041797e-01
-1.30087817e+00 -1.49348414e+00 -6.72963142e-01 1.26262277e-01
-2.34703496e-02 -5.17312467e-01 1.51013327e+00 -3.69803995e-01
-8.23310256e-01 3.62367481e-01 5.01414061e-01 -8.67040038e-01
9.60254848e-01 -4.28991884e-01 -5.41251957e-01 -2.79757887e-01
-2.10944712e-01 4.51644033e-01 8.36125255e-01 -9.92658436e-01
-5.87823629e-01 -3.05572510e-01 -1.69135422e-01 -2.38252223e-01
-4.62917835e-01 -1.41809568e-01 2.56043673e-01 -4.53547806e-01
-4.68696117e-01 -9.55780208e-01 -1.50536224e-01 3.88769150e-01
-3.72454017e-01 -1.20051086e-01 5.27690530e-01 -2.56316483e-01
1.12306893e+00 -2.37879181e+00 -3.63508373e-01 3.66896182e-01
1.36750832e-01 -2.34962046e-01 1.30085707e-01 3.83334816e-01
-1.59808069e-01 -5.15792798e-03 -4.27056342e-01 -3.37236494e-01
3.22278053e-01 1.95008144e-01 -6.66554928e-01 8.58210385e-01
1.03794562e-03 5.49827397e-01 -6.11884296e-01 -4.33065265e-01
-1.46701768e-01 5.36317468e-01 -7.77490795e-01 -9.77672264e-02
-1.43431947e-01 2.70021617e-01 -2.33956337e-01 2.97381312e-01
7.82809496e-01 -3.50434855e-02 -2.22056001e-01 1.15935825e-01
4.70425487e-01 -1.71125636e-01 -5.68490505e-01 1.28524566e+00
-3.43657583e-01 5.88959694e-01 -2.28912249e-01 -8.45129609e-01
4.69891101e-01 1.57712787e-01 4.55245882e-01 -4.63318706e-01
4.25234079e-01 -6.98783323e-02 -1.48172557e-01 -7.33901374e-03
1.82285100e-01 -7.17454135e-01 -1.93354338e-01 2.61469185e-01
-2.33048517e-02 -1.43804342e-01 -4.84788597e-01 4.68560159e-02
1.32022667e+00 -1.77012935e-01 -1.53659984e-01 -9.35077071e-01
-1.65453553e-01 -2.55895197e-01 6.37936771e-01 8.56337488e-01
-3.64368320e-01 8.06561649e-01 7.34548271e-01 -5.52879393e-01
-1.09021258e+00 -1.50256312e+00 -1.02398491e+00 9.11498308e-01
-5.43622077e-02 -1.77154645e-01 -8.16055357e-01 -7.55805850e-01
5.30229092e-01 1.20631826e+00 -9.61309910e-01 -8.38040769e-01
-3.55248451e-02 -8.84649754e-01 7.79493153e-01 9.54962671e-01
-1.33102313e-01 -4.90649194e-01 -5.56133032e-01 -2.20695734e-02
4.16740328e-01 -9.99017537e-01 -4.70055610e-01 4.42835331e-01
-6.60442173e-01 -1.21326447e+00 -3.89032006e-01 -1.56371683e-01
3.77995789e-01 -4.78379071e-01 1.32313418e+00 -2.27640927e-01
-2.43470088e-01 5.45038819e-01 8.56787041e-02 -8.70549381e-01
-3.41033459e-01 -6.21499829e-02 5.35322726e-01 -3.48336279e-01
3.99760492e-02 -7.67274678e-01 -5.69200993e-01 4.20806974e-01
-9.71838772e-01 -4.03735280e-01 1.25195280e-01 7.36712396e-01
6.22799039e-01 2.92707831e-01 9.80930448e-01 -8.47587645e-01
6.78951263e-01 -7.91095197e-01 -9.16397929e-01 2.05274045e-01
-1.01631081e+00 1.12049080e-01 6.13109469e-01 -5.13199806e-01
-8.50828767e-01 -1.71715721e-01 -1.56859562e-01 -5.94512522e-01
-9.62598026e-02 3.12926382e-01 5.22245429e-02 5.08670919e-02
9.90662158e-01 -2.43367448e-01 -1.19217768e-01 -3.45084846e-01
-4.11718488e-02 5.00987470e-01 6.39077127e-01 -1.03279173e+00
8.24686408e-01 3.94762903e-01 4.34944451e-01 -1.60859674e-01
-1.25123930e+00 2.90945977e-01 -3.25050414e-01 -1.68546531e-02
8.42645764e-01 -8.58504891e-01 -4.97966975e-01 3.89962882e-01
-8.06275368e-01 -6.83657885e-01 -3.25213253e-01 3.33943844e-01
-7.76225567e-01 2.43451297e-01 -2.99826384e-01 -9.20682669e-01
-5.55358589e-01 -9.22119975e-01 1.01060224e+00 -5.97831309e-02
-2.10388035e-01 -1.10178852e+00 1.40057698e-01 -2.07615584e-01
6.75629973e-01 4.19759333e-01 1.30360448e+00 -9.97196794e-01
-1.00485864e-03 -4.06902730e-01 -1.41841676e-02 4.53686684e-01
-1.98298410e-01 -4.54421230e-02 -1.16733265e+00 -3.96319211e-01
-2.97339767e-01 -3.69312227e-01 5.41826069e-01 6.59163892e-01
1.80901599e+00 -1.37791395e-01 -2.63465017e-01 6.59307420e-01
1.06056380e+00 1.24336399e-01 3.73248756e-01 -3.15491781e-02
1.64032474e-01 6.09715939e-01 5.10365546e-01 5.36179245e-01
1.63977295e-01 4.13514614e-01 5.79273522e-01 4.05782044e-01
1.83267117e-01 -2.71422863e-01 4.20991518e-02 -3.50459591e-02
6.00516975e-01 -1.70638293e-01 -1.18565655e+00 4.36114013e-01
-1.89240408e+00 -8.33337963e-01 1.69484802e-02 2.46057272e+00
9.21440661e-01 5.85171461e-01 3.98459733e-02 -1.23846412e-01
4.28788662e-01 -3.49660546e-01 -1.00938976e+00 -3.96172822e-01
7.96088204e-02 1.62673056e-01 7.40481734e-01 2.77066201e-01
-1.20263469e+00 4.68643188e-01 7.90000725e+00 9.05780494e-01
-7.96990991e-01 2.54850954e-01 9.99078095e-01 -3.76905710e-01
-5.04623115e-01 -3.24952871e-01 -7.40151703e-01 5.35251677e-01
1.15304160e+00 -3.74342352e-01 1.09308511e-01 1.08448148e+00
-2.00854927e-01 6.45508468e-02 -1.59883893e+00 7.99609780e-01
-7.14768469e-02 -1.28250921e+00 -4.25620168e-01 1.70736447e-01
7.75791109e-01 3.31087202e-01 4.83501941e-01 4.24531460e-01
5.10723233e-01 -1.32079077e+00 8.53681087e-01 4.80270952e-01
1.07894516e+00 -1.26247358e+00 8.40062678e-01 3.88223708e-01
-4.20669258e-01 -2.46304899e-01 -4.23291951e-01 2.60710567e-01
2.45976336e-02 9.86021101e-01 -5.29816151e-02 1.57438904e-01
1.05827844e+00 4.61277783e-01 -3.99160385e-01 1.14196718e+00
2.41041079e-01 7.77046382e-01 -5.20128429e-01 2.14839116e-01
-2.65327375e-02 3.01773906e-01 3.26045513e-01 8.53446007e-01
3.95874888e-01 -7.70400837e-02 -3.11554402e-01 7.13756382e-01
3.28735635e-02 -4.34536248e-01 -8.11021328e-01 2.22368598e-01
5.11909485e-01 8.92593384e-01 -2.22105742e-01 7.39150941e-02
-1.32233799e-01 2.10324720e-01 3.02908152e-01 2.87142068e-01
-1.39991665e+00 -1.58658803e-01 9.16377604e-01 3.54916185e-01
-9.06564668e-02 -1.85824424e-01 -5.78446865e-01 -8.57732058e-01
-1.17146805e-01 -5.03899515e-01 7.07680166e-01 -5.51195502e-01
-1.80476940e+00 4.28004920e-01 4.89946067e-01 -1.14207649e+00
-1.51276812e-01 -7.41038740e-01 -6.87969804e-01 4.82173353e-01
-1.18183112e+00 -1.04106939e+00 -6.28325865e-02 5.32687962e-01
-5.96617833e-02 -3.72897983e-01 8.62452209e-01 4.08367008e-01
-4.24431324e-01 1.26295304e+00 3.51795107e-01 -2.19045043e-01
9.32952046e-01 -1.05879915e+00 1.03777111e-01 5.21551669e-01
-1.72451064e-01 3.54198873e-01 1.08474112e+00 -5.36505699e-01
-1.15229833e+00 -1.34426081e+00 1.80577919e-01 -7.58108735e-01
8.66622925e-01 -4.45707828e-01 -8.40695977e-01 8.39227080e-01
1.28830764e-02 2.90577322e-01 1.03089523e+00 4.94737357e-01
-6.67272389e-01 -3.40306371e-01 -1.48482633e+00 3.00590426e-01
1.18533397e+00 -4.59913552e-01 -2.37133384e-01 4.85685915e-01
7.62576044e-01 -5.59659958e-01 -1.10032010e+00 9.16414201e-01
8.15593004e-01 -9.42728639e-01 8.61068487e-01 -9.96084690e-01
6.08178020e-01 1.80901110e-01 -5.81821322e-01 -1.40163457e+00
-2.66023844e-01 -4.35305983e-01 -3.63281041e-01 1.09210813e+00
6.25446916e-01 -7.17114031e-01 6.70990407e-01 1.22368002e+00
-3.05805206e-01 -9.59350467e-01 -1.37955201e+00 -1.06188512e+00
1.00601041e+00 -9.80877042e-01 7.24357069e-01 7.43073106e-01
-9.32933167e-02 -2.35576272e-01 -3.46116692e-01 3.28510880e-01
1.10722208e+00 -2.10080042e-01 3.79769444e-01 -1.30641150e+00
-1.51332676e-01 -4.78235543e-01 -6.26946151e-01 -2.15453774e-01
4.92707640e-01 -7.13134646e-01 1.26011804e-01 -9.84778106e-01
2.65490711e-01 -9.66664314e-01 -6.39750004e-01 4.65821862e-01
2.02474952e-01 2.78043002e-02 -2.78267354e-01 -2.44967043e-01
-4.21338558e-01 8.99463356e-01 5.41931927e-01 -1.49450719e-01
4.16886240e-01 3.63546729e-01 -7.77558565e-01 6.94800735e-01
8.45527589e-01 -5.93796372e-01 -8.59568596e-01 -3.86916637e-01
5.91458499e-01 -4.63301502e-02 4.78716254e-01 -1.26529121e+00
1.99022532e-01 -2.75836051e-01 5.22489429e-01 -6.72048211e-01
4.12039645e-02 -1.19799542e+00 2.44034275e-01 2.72397965e-01
-6.53074861e-01 4.88392971e-02 6.74364448e-01 8.94534111e-01
2.96908975e-01 7.19753355e-02 1.16198277e+00 7.30536401e-01
2.34770086e-02 8.69966447e-01 -7.11803287e-02 4.50601071e-01
1.32856774e+00 3.68638843e-01 -4.09023762e-01 -6.53979301e-01
-5.22025883e-01 4.35190350e-01 2.28429660e-01 5.96729279e-01
4.90000546e-01 -1.60434544e+00 -6.24953151e-01 1.09631307e-01
4.35898662e-01 3.01308542e-01 4.67502803e-01 7.84524977e-01
-2.11537510e-01 6.56778589e-02 -3.63461263e-02 -7.35259831e-01
-7.27142215e-01 6.27763450e-01 7.15841353e-01 1.97830170e-01
-6.17523015e-01 8.57089400e-01 6.18644178e-01 -4.72517997e-01
7.74535716e-01 -2.88555354e-01 3.16899598e-01 -3.42079878e-01
5.36514461e-01 1.72459856e-01 2.46301010e-01 -1.55838117e-01
-3.96959633e-01 1.94693923e-01 8.01261291e-02 -1.54291047e-04
1.30034876e+00 2.05421984e-01 3.42318356e-01 2.67608166e-01
1.13568366e+00 -3.04429919e-01 -1.51991212e+00 1.47196338e-01
-3.66115272e-02 -2.64627218e-01 9.19726864e-02 -8.75409484e-01
-1.07020175e+00 1.07874751e+00 8.78166020e-01 -1.07054129e-01
8.35128784e-01 -4.57096733e-02 7.94706404e-01 1.85674161e-01
6.15341783e-01 -1.01678586e+00 -3.64106447e-01 3.16584468e-01
1.05935490e+00 -9.79458869e-01 -6.76783770e-02 2.73219799e-03
-4.54037577e-01 6.79111183e-01 7.77131259e-01 -6.67893440e-02
1.61785245e+00 9.04670894e-01 -2.07222864e-01 -4.79472987e-03
-9.67097402e-01 7.65410125e-01 4.23344076e-01 8.34429502e-01
1.76944379e-02 1.72013566e-01 6.06889613e-02 1.13511395e+00
-4.85008329e-01 -3.47035587e-01 3.65517259e-01 7.43429005e-01
-2.36667112e-01 -7.37889290e-01 2.99270405e-03 7.63264596e-01
-6.25365257e-01 -9.63304266e-02 6.09771386e-02 6.01502061e-01
-1.85047641e-01 6.35026991e-01 3.31720769e-01 -4.79502141e-01
2.81371474e-01 1.43810779e-01 4.65707630e-01 -1.85002089e-01
-2.28644475e-01 -3.34318578e-01 1.32785037e-01 -6.41260386e-01
2.31329679e-01 -6.34580255e-01 -1.37160504e+00 -6.18352413e-01
-1.89653799e-01 -7.69757852e-02 6.22010946e-01 9.94107425e-01
4.95356739e-01 4.16538805e-01 3.90254557e-01 -5.37130296e-01
-1.35593832e+00 -7.01554775e-01 -7.55768120e-01 2.11774930e-01
5.24142683e-01 -1.18636978e+00 -5.17324626e-01 -5.08571804e-01] | [7.8886260986328125, 4.0882182121276855] |
6866fe52-73dc-46c4-b365-b2618b827e50 | r2h-building-multimodal-navigation-helpers | 2305.14260 | null | https://arxiv.org/abs/2305.14260v1 | https://arxiv.org/pdf/2305.14260v1.pdf | R2H: Building Multimodal Navigation Helpers that Respond to Help | The ability to assist humans during a navigation task in a supportive role is crucial for intelligent agents. Such agents, equipped with environment knowledge and conversational abilities, can guide individuals through unfamiliar terrains by generating natural language responses to their inquiries, grounded in the visual information of their surroundings. However, these multimodal conversational navigation helpers are still underdeveloped. This paper proposes a new benchmark, Respond to Help (R2H), to build multimodal navigation helpers that can respond to help, based on existing dialog-based embodied datasets. R2H mainly includes two tasks: (1) Respond to Dialog History (RDH), which assesses the helper agent's ability to generate informative responses based on a given dialog history, and (2) Respond during Interaction (RdI), which evaluates the helper agent's ability to maintain effective and consistent cooperation with a task performer agent during navigation in real-time. Furthermore, we propose a novel task-oriented multimodal response generation model that can see and respond, named SeeRee, as the navigation helper to guide the task performer in embodied tasks. Through both automatic and human evaluations, we show that SeeRee produces more effective and informative responses than baseline methods in assisting the task performer with different navigation tasks. Project website: https://sites.google.com/view/respond2help/home. | ['Xin Eric Wang', 'Jing Gu', 'Kaizhi Zheng', 'Yue Fan'] | 2023-05-23 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [-2.44766504e-01 5.37905753e-01 5.89261174e-01 -5.62193096e-01
-5.94990551e-01 -5.93377590e-01 9.55520689e-01 -1.74352840e-01
-4.46375579e-01 1.03425133e+00 7.89906681e-01 1.94867015e-01
1.57588676e-01 -7.27911472e-01 1.37194633e-01 -5.67260623e-01
1.46162540e-01 9.16383922e-01 -6.11731037e-02 -1.00304544e+00
2.41270348e-01 1.77972252e-03 -1.41611636e+00 4.35775548e-01
1.15993953e+00 4.66983348e-01 5.59306920e-01 9.12351549e-01
-1.75107390e-01 1.25569987e+00 -8.56742263e-01 -2.54876137e-01
-2.47808442e-01 -6.92932129e-01 -1.30148351e+00 -6.58447817e-02
-4.22959417e-01 -7.77742863e-01 -2.24181116e-01 4.68008667e-01
9.63529229e-01 1.00182629e+00 6.86166883e-01 -1.56749010e+00
-8.25027287e-01 5.08332610e-01 4.17515486e-01 -2.96246380e-01
1.34073579e+00 6.25719309e-01 4.89180386e-01 -7.34269679e-01
8.19776595e-01 1.61573720e+00 2.23286420e-01 1.26785851e+00
-8.85826647e-01 -1.21311307e-01 1.89843908e-01 -1.31156780e-02
-8.63260567e-01 -6.14948213e-01 7.22617686e-01 -4.42483991e-01
9.53242064e-01 4.02099401e-01 4.81682062e-01 1.47900701e+00
-1.78108573e-01 8.70069623e-01 9.47509527e-01 -1.54949084e-01
3.65643710e-01 4.50646967e-01 1.71090961e-01 5.81132531e-01
-7.25360036e-01 1.16216980e-01 -1.09241140e+00 -7.87058324e-02
4.67263937e-01 -3.01858932e-01 -5.20547867e-01 3.84295210e-02
-1.49764955e+00 9.50168788e-01 6.20760977e-01 2.71671504e-01
-7.27287412e-01 -2.40700960e-01 2.61927515e-01 5.65665722e-01
1.97756574e-01 7.36084521e-01 8.05069432e-02 -5.39415658e-01
3.72186780e-01 5.58511496e-01 1.16178572e+00 1.07731283e+00
6.55928552e-01 -1.17940016e-01 -6.27681911e-01 1.18405342e+00
2.96591938e-01 6.16404772e-01 4.03608829e-01 -1.12655902e+00
2.89844304e-01 1.02909136e+00 6.19713485e-01 -8.75989795e-01
-7.76931763e-01 2.71733850e-01 -7.50460386e-01 3.25771302e-01
3.52417350e-01 -7.89515257e-01 -1.56196952e-01 2.00612473e+00
7.21136153e-01 -9.32641268e-01 7.18176961e-01 1.15171540e+00
1.68482006e+00 8.54109585e-01 1.30074292e-01 1.80080440e-03
1.42781997e+00 -1.29186320e+00 -9.85629916e-01 -4.32986319e-01
8.18350196e-01 -3.12358439e-01 1.44075620e+00 1.68361649e-01
-1.07680857e+00 -5.60262263e-01 -5.44369936e-01 -2.20426857e-01
-5.48736274e-01 4.44940217e-02 5.79593837e-01 2.66217530e-01
-1.51632965e+00 -2.12757692e-01 -5.87736107e-02 -9.60444212e-01
-4.66042995e-01 1.52897120e-01 -7.64325857e-01 2.50721216e-01
-1.53431463e+00 1.16682196e+00 9.66813713e-02 7.04425573e-02
-1.07201433e+00 -2.42044497e-03 -1.12254465e+00 -3.05751204e-01
3.36714089e-01 -1.09520376e+00 1.70783257e+00 -3.81768286e-01
-1.85806394e+00 8.64394724e-01 -1.73273012e-01 -9.62385833e-02
5.34991980e-01 -9.02610943e-02 -1.15012161e-01 1.58788532e-01
3.67645979e-01 1.25935292e+00 1.31818116e-01 -1.56222272e+00
-6.90324903e-01 -4.54290807e-01 5.35351396e-01 1.03536904e+00
-2.32486770e-01 -1.22260354e-01 -2.70000428e-01 -3.07230383e-01
-2.87270963e-01 -8.93955231e-01 -3.12921435e-01 -1.05073564e-01
-5.91043830e-01 -7.78633118e-01 6.35633588e-01 -5.90702832e-01
7.56226957e-01 -1.79044628e+00 5.54788709e-01 -2.24191844e-01
3.17675710e-01 -1.13315418e-01 -3.24395031e-01 1.10996950e+00
7.46334910e-01 -2.12666765e-01 -6.25193417e-02 -7.15590656e-01
4.18501675e-01 3.28674577e-02 2.91279014e-02 -1.67423695e-01
-2.50891745e-01 1.01100123e+00 -1.02487218e+00 -3.84783536e-01
2.44215071e-01 4.12433177e-01 -1.79361045e-01 1.08086419e+00
-7.14887008e-02 1.01247525e+00 -6.31040394e-01 5.14020741e-01
7.96944574e-02 -1.02076670e-02 3.01253162e-02 2.78893620e-01
-2.10497364e-01 2.05681339e-01 -5.84239483e-01 1.74756074e+00
-6.12295449e-01 4.89279091e-01 5.87584734e-01 -1.55059457e-01
1.28811252e+00 4.18079585e-01 -3.63356881e-02 -1.13906169e+00
4.34568673e-02 -1.33381292e-01 -2.33390182e-01 -8.36909652e-01
7.49005020e-01 1.36267915e-01 -6.40225589e-01 8.67573619e-01
-3.07632118e-01 -2.92291611e-01 9.26061422e-02 6.30980372e-01
1.19755566e+00 3.98953110e-02 3.76377106e-01 6.50422350e-02
6.79792523e-01 1.26175225e-01 -3.78254727e-02 9.28407729e-01
-4.04987991e-01 5.77661842e-02 3.05360943e-01 -3.37753862e-01
-4.65718597e-01 -8.19983304e-01 5.96486151e-01 1.65702629e+00
5.29843926e-01 -1.01900339e-01 -9.39393163e-01 -6.62785172e-01
-3.42409343e-01 9.90788817e-01 -6.41705036e-01 -1.38922647e-01
-4.71313953e-01 -1.25957906e-01 6.10992730e-01 4.82376903e-01
1.11226928e+00 -1.94180155e+00 -9.78533089e-01 1.07267894e-01
-9.53346252e-01 -8.27434480e-01 -4.69110876e-01 -1.76423877e-01
-1.59059763e-01 -7.64824092e-01 -9.51479435e-01 -9.60274339e-01
7.29748309e-01 4.22505349e-01 1.04546857e+00 3.76940161e-01
3.02679539e-01 1.03180933e+00 -8.50524664e-01 -3.34789045e-02
-5.47603190e-01 1.26114264e-02 5.60490228e-02 -4.22034651e-01
1.11697942e-01 -3.42763573e-01 -5.82292080e-01 9.47317541e-01
-4.07043755e-01 5.83788574e-01 2.22661138e-01 8.29432607e-01
-3.19294333e-01 -5.45397997e-01 8.43057752e-01 -4.61986184e-01
1.65186548e+00 -5.05803227e-01 8.56105909e-02 3.73334378e-01
-1.26387984e-01 -1.95133939e-01 3.95712435e-01 -3.59204322e-01
-1.61307907e+00 -1.56657293e-01 -2.98801005e-01 7.02794313e-01
-7.00923264e-01 3.61466348e-01 -3.50367367e-01 -4.18265574e-02
8.29824209e-01 3.97497654e-01 1.80029303e-01 -1.41224623e-01
5.28389037e-01 9.25791681e-01 8.42368007e-01 -7.88865447e-01
5.25602698e-01 1.14501275e-01 -5.77627361e-01 -8.13437939e-01
-7.62298927e-02 -3.42664957e-01 -3.71493965e-01 -9.15824831e-01
1.04422748e+00 -6.76459670e-01 -1.57559633e+00 5.97421885e-01
-1.39503372e+00 -1.08680904e+00 1.39029771e-01 1.03186734e-01
-7.27523744e-01 1.44424707e-01 -5.45607626e-01 -1.34073031e+00
-5.34925222e-01 -1.09170258e+00 1.08421659e+00 6.11261308e-01
-7.77171612e-01 -1.10769320e+00 1.43124372e-01 6.81044519e-01
7.64340758e-01 3.01407665e-01 6.83392763e-01 -9.08073246e-01
-2.73507237e-01 2.12444998e-02 3.27880047e-02 -3.72514635e-01
2.07014289e-02 -7.81095266e-01 -7.35057354e-01 -7.92319104e-02
-7.72987828e-02 -9.30193782e-01 1.72624528e-01 -1.13787845e-01
8.97388011e-02 -5.87462485e-01 -3.09441239e-01 -3.94379608e-02
2.67243743e-01 8.54775071e-01 6.29216969e-01 3.63852233e-01
2.21040651e-01 1.50012076e+00 1.14765429e+00 5.27329087e-01
1.15125823e+00 1.00465953e+00 2.77420163e-01 -9.92164090e-02
4.45513800e-02 -5.90597749e-01 6.51010334e-01 2.98896044e-01
-7.49913678e-02 -7.70103574e-01 -8.40223610e-01 3.76299173e-01
-2.25279593e+00 -8.81126225e-01 -1.10734023e-01 1.76024878e+00
5.89555740e-01 -5.22526562e-01 4.22359020e-01 -3.67223889e-01
5.56564510e-01 -1.79407820e-02 -6.18641675e-01 -2.56726682e-01
-1.11182682e-01 -7.46623516e-01 -6.18685842e-01 8.20935905e-01
-6.40898645e-01 1.19716036e+00 5.25360107e+00 2.05087051e-01
-3.54941040e-01 3.44670042e-02 4.71110225e-01 3.10214847e-01
-4.35064524e-01 -2.20796824e-01 -5.46306372e-01 2.43654296e-01
4.42220181e-01 -3.12833875e-01 6.07822418e-01 8.04457128e-01
4.35010344e-01 -5.03103197e-01 -1.18556237e+00 9.21325922e-01
1.96533769e-01 -8.60759854e-01 -1.11723252e-01 -1.45427167e-01
8.43951106e-02 -7.92551696e-01 1.33636668e-02 8.47778201e-01
7.30442405e-01 -1.03746486e+00 4.27654713e-01 8.16085696e-01
3.99149120e-01 -4.53717530e-01 9.42621946e-01 7.08762050e-01
-1.23706925e+00 3.37391975e-03 8.94733071e-02 -2.71392345e-01
9.06842113e-01 -5.44999719e-01 -1.28265989e+00 3.26650918e-01
6.32499099e-01 2.09978968e-01 -4.24538523e-01 4.87656355e-01
-6.08743727e-01 -2.47527078e-01 2.20814228e-01 -6.87101305e-01
1.51543394e-01 -3.31079632e-01 8.76745760e-01 8.92268181e-01
3.02290827e-01 6.97799742e-01 4.61862266e-01 8.17114651e-01
2.65344799e-01 1.99466482e-01 -9.14955437e-01 1.10729106e-01
8.85670066e-01 1.26804006e+00 -3.36148411e-01 -2.35343859e-01
2.12161362e-01 1.27155244e+00 3.14303607e-01 5.34119427e-01
-4.34339583e-01 -4.80603039e-01 4.02898610e-01 -2.22522914e-01
-6.75730884e-01 -3.54809202e-02 1.10664949e-01 -4.78880316e-01
-9.50915888e-02 -1.16774714e+00 4.89805013e-01 -1.57065809e+00
-1.19853938e+00 1.04229724e+00 6.23093545e-02 -9.62688863e-01
-8.62835348e-01 -2.91424841e-01 -8.67009819e-01 8.79599094e-01
-6.54458284e-01 -1.45308745e+00 -1.04569900e+00 7.54988849e-01
1.10886991e+00 -4.55184251e-01 1.08238947e+00 -1.87328711e-01
-4.06658560e-01 3.85984778e-01 -8.81308317e-01 -7.05001801e-02
9.12865579e-01 -1.38210261e+00 3.27485800e-01 1.63403884e-01
-8.00158858e-01 7.94251502e-01 7.51057863e-01 -8.20306659e-01
-1.31835699e+00 -4.83697057e-01 8.43512058e-01 -5.20703435e-01
1.66606948e-01 -3.15669715e-01 -7.59260535e-01 5.25335252e-01
7.07412899e-01 -1.06639445e+00 6.79475248e-01 5.72243296e-02
3.02860856e-01 4.69389141e-01 -1.45156288e+00 9.80229914e-01
1.45847106e+00 -4.77693945e-01 -7.78953254e-01 4.69244361e-01
1.03878498e+00 -4.38021332e-01 -5.60940385e-01 3.51501405e-02
4.32272822e-01 -1.20912516e+00 9.05265391e-01 -5.84485352e-01
2.75239080e-01 -2.99172223e-01 -1.12382323e-01 -1.67436874e+00
-5.37224449e-02 -9.70682681e-01 3.11365932e-01 1.41484237e+00
2.80999005e-01 -8.64612460e-01 3.34803015e-01 1.26296318e+00
-3.79966795e-01 -2.79287517e-01 -6.71390355e-01 -2.47733533e-01
-3.81139964e-01 5.04529886e-02 7.18336105e-01 7.30858982e-01
7.80426025e-01 7.71085024e-01 -7.36551821e-01 -1.49046266e-02
6.23740181e-02 -3.46639425e-01 1.57216406e+00 -1.24169838e+00
1.44599482e-01 -3.67646933e-01 1.51776791e-01 -1.41684222e+00
3.47700953e-01 -5.08102775e-01 5.64415872e-01 -2.27362895e+00
5.40475175e-02 -9.84202251e-02 7.19290018e-01 5.53475022e-01
-2.14305818e-01 -2.67056733e-01 1.72979444e-01 2.05353349e-01
-9.40883875e-01 1.07113719e+00 1.39437115e+00 -2.62431484e-02
-7.07178771e-01 1.49653569e-01 -9.20148849e-01 6.56872749e-01
6.97805583e-01 1.51360735e-01 -6.29959226e-01 -1.29249647e-01
1.09502584e-01 7.29683161e-01 6.55890822e-01 -8.06305707e-01
8.10503781e-01 -2.19914198e-01 6.90992251e-02 -5.94779670e-01
8.11117291e-01 -3.79068643e-01 9.53953266e-02 3.74562681e-01
-7.86721826e-01 4.44430649e-01 -4.68079709e-02 3.03620845e-01
-1.11260734e-01 -1.70048475e-01 1.74405679e-01 -2.55573690e-01
-1.03483284e+00 -1.89623401e-01 -1.01094174e+00 -1.11566842e-01
1.04852808e+00 -1.85031474e-01 -9.52453852e-01 -1.52303779e+00
-6.95487440e-01 1.09881222e+00 4.75234926e-01 5.43061912e-01
1.12221086e+00 -1.49693155e+00 -6.66143417e-01 -3.22293818e-01
4.52605188e-01 -1.58124194e-01 8.19740891e-01 5.86112976e-01
-1.92957386e-01 3.76731783e-01 -2.80906528e-01 -1.89637661e-01
-1.43497956e+00 1.54547030e-02 4.00656909e-01 -2.82731038e-02
-5.17781198e-01 1.05338752e+00 4.43534702e-01 -1.13803625e+00
5.75617731e-01 3.89031857e-01 -8.53321016e-01 1.60076290e-01
8.83887351e-01 5.59089005e-01 -5.38503230e-01 -7.78312325e-01
-2.02255741e-01 1.82449162e-01 1.53986126e-01 -8.45870495e-01
8.56344044e-01 -5.83165109e-01 -8.67856964e-02 2.70878851e-01
3.89649123e-01 -2.42456779e-01 -1.11119068e+00 -1.07519716e-01
-2.91312337e-01 -3.32337171e-01 -6.74406588e-01 -1.35714376e+00
-3.41708958e-01 5.88746607e-01 3.80089998e-01 4.43082482e-01
8.86071980e-01 1.63566515e-01 6.80034339e-01 1.18219745e+00
6.79395258e-01 -1.07002747e+00 7.73559988e-01 5.43910861e-01
1.72985518e+00 -1.46964979e+00 -6.62705898e-01 -2.03180224e-01
-1.59399021e+00 7.67048597e-01 1.38283074e+00 7.36664712e-01
-2.25768849e-01 -1.15157932e-01 6.11315846e-01 -3.54672104e-01
-1.02389717e+00 -4.62282568e-01 8.62670019e-02 1.35291803e+00
3.27481329e-01 3.86534482e-02 -1.00263827e-01 8.08905423e-01
-6.52087271e-01 -3.31441313e-01 4.08305705e-01 8.63706350e-01
-6.92495227e-01 -8.51707935e-01 -4.75850761e-01 -2.24912494e-01
8.35069120e-01 2.74316430e-01 -1.37570894e+00 7.86688209e-01
-3.38080257e-01 1.93560398e+00 -3.20715696e-01 -6.31729364e-01
7.81553745e-01 1.24783061e-01 -1.76220343e-01 -5.14133036e-01
-9.03919995e-01 -4.30391133e-01 7.90403247e-01 -5.08038402e-01
-2.80784845e-01 -2.81397700e-01 -1.46300042e+00 -4.29223359e-01
2.47795925e-01 4.38359916e-01 2.62138277e-01 8.08973253e-01
1.54829144e-01 3.69969934e-01 5.63343287e-01 -1.09645569e+00
-3.56969327e-01 -1.32997012e+00 -1.03828378e-01 4.92240459e-01
5.71733713e-02 -7.40694702e-01 -3.07991087e-01 -3.36430550e-01] | [12.939884185791016, 7.835784912109375] |
9867f4ba-0217-43e9-a957-fb5ac76c107f | high-quality-diversification-for-task | 2106.00891 | null | https://arxiv.org/abs/2106.00891v2 | https://arxiv.org/pdf/2106.00891v2.pdf | High-Quality Diversification for Task-Oriented Dialogue Systems | Many task-oriented dialogue systems use deep reinforcement learning (DRL) to learn policies that respond to the user appropriately and complete the tasks successfully. Training DRL agents with diverse dialogue trajectories prepare them well for rare user requests and unseen situations. One effective diversification method is to let the agent interact with a diverse set of learned user models. However, trajectories created by these artificial user models may contain generation errors, which can quickly propagate into the agent's policy. It is thus important to control the quality of the diversification and resist the noise. In this paper, we propose a novel dialogue diversification method for task-oriented dialogue systems trained in simulators. Our method, Intermittent Short Extension Ensemble (I-SEE), constrains the intensity to interact with an ensemble of diverse user models and effectively controls the quality of the diversification. Evaluations on the Multiwoz dataset show that I-SEE successfully boosts the performance of several state-of-the-art DRL dialogue agents. | ['Grace Hui Yang', 'Hrishikesh Kulkarni', 'Zhiwen Tang'] | 2021-06-02 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [-2.84239709e-01 3.33899528e-01 5.06063178e-02 -2.88374633e-01
-3.42045635e-01 -6.31008744e-01 7.57980585e-01 -1.94834128e-01
-5.62679946e-01 1.09463286e+00 9.59596410e-02 -2.31826141e-01
1.48368880e-01 -7.44501591e-01 -2.89710701e-01 -7.05734015e-01
1.06582522e-01 1.15509605e+00 3.19803447e-01 -1.00976777e+00
2.95861214e-01 1.90853834e-01 -1.33190060e+00 9.60493088e-02
1.43238902e+00 3.68613958e-01 3.50430638e-01 9.77769017e-01
-1.84459955e-01 1.07253206e+00 -1.26935256e+00 -7.05283806e-02
8.72229636e-02 -6.21012986e-01 -8.68307889e-01 1.00304402e-01
-2.25240842e-01 -5.64278424e-01 -2.73021400e-01 8.30348790e-01
7.85573006e-01 5.77917099e-01 4.51224983e-01 -1.17155862e+00
-4.91327606e-02 7.35026062e-01 -2.49371424e-01 1.45519435e-01
3.54196250e-01 7.06082106e-01 6.63275659e-01 -2.38777623e-01
6.89655781e-01 1.45786643e+00 3.37733835e-01 1.29926586e+00
-1.15370619e+00 -4.87821341e-01 2.37037212e-01 -4.21641082e-01
-5.38941503e-01 -4.85813320e-01 5.68107426e-01 -2.83576578e-01
9.93245304e-01 1.05213098e-01 5.71085095e-01 1.41245711e+00
2.89857626e-01 9.24257517e-01 9.38527286e-01 -2.94425845e-01
5.90117633e-01 4.19414669e-01 4.81436960e-03 7.04478800e-01
-2.99965054e-01 1.71488732e-01 -4.86458242e-01 -4.63617027e-01
6.09951079e-01 -4.78876740e-01 -1.77508116e-01 -1.61929294e-01
-8.19385111e-01 1.15927827e+00 -7.75460824e-02 2.52017170e-01
-5.18012524e-01 -1.57449976e-01 6.17238700e-01 5.24614453e-01
3.10287386e-01 1.13794744e+00 -5.27424157e-01 -6.96646154e-01
-1.68718621e-01 7.27058649e-01 1.32683504e+00 7.78433144e-01
3.36373985e-01 2.85906941e-01 -5.50621390e-01 1.24226284e+00
1.13761306e-01 3.44371140e-01 6.88625395e-01 -1.20018125e+00
4.35611606e-01 6.30836308e-01 3.97143751e-01 -3.19861889e-01
-6.06900632e-01 -2.16559857e-01 -5.94747961e-01 3.38893086e-01
6.49499536e-01 -1.01656842e+00 -5.85673690e-01 1.98577452e+00
4.48696524e-01 -3.18585098e-01 4.68144774e-01 7.61777103e-01
5.33440471e-01 8.59994471e-01 1.00589074e-01 -2.89029330e-01
9.17108417e-01 -1.02198684e+00 -8.05016100e-01 -2.97345281e-01
6.67753100e-01 -3.21604848e-01 1.20062101e+00 4.29430068e-01
-1.26631296e+00 -4.64739919e-01 -7.72451401e-01 5.92253566e-01
8.46678615e-02 -2.58156359e-01 5.22316813e-01 6.31467044e-01
-1.01761198e+00 5.69921792e-01 -5.87681174e-01 -1.82276368e-01
-3.01405098e-02 5.29468060e-01 2.01477125e-01 4.24672216e-01
-1.58107328e+00 1.16040874e+00 2.64362425e-01 -1.86456010e-01
-9.44216669e-01 -3.52402747e-01 -7.12624729e-01 3.34937759e-02
5.41897058e-01 -6.23452008e-01 1.96266806e+00 -8.88140142e-01
-2.34596014e+00 2.63930470e-01 2.86943257e-01 -5.54777443e-01
7.78689742e-01 -6.56246543e-02 6.12357445e-02 -1.24701701e-01
-1.76088259e-01 8.50667536e-01 7.52701759e-01 -1.43769193e+00
-7.94887483e-01 -4.52077091e-02 2.34260559e-01 6.63256407e-01
-1.79616407e-01 -3.97634178e-01 -1.38096169e-01 -2.37239763e-01
-6.25823379e-01 -1.06276679e+00 -5.68167329e-01 -7.81474888e-01
-2.67760664e-01 -7.07746089e-01 8.02055717e-01 -3.01775992e-01
1.08226228e+00 -1.71193755e+00 4.92984653e-01 3.34900059e-02
1.60921574e-01 5.63847303e-01 -2.83995628e-01 6.78979278e-01
5.42312682e-01 -1.20972104e-01 -2.91755535e-02 -4.08146173e-01
2.30980650e-01 4.22641605e-01 -2.45369017e-01 -9.45990384e-02
8.72438923e-02 6.00185812e-01 -1.15137815e+00 -7.17109069e-02
1.66237265e-01 9.21163056e-03 -8.58815968e-01 9.88444030e-01
-1.00737929e+00 9.04643834e-01 -7.00390458e-01 8.65318105e-02
3.01823437e-01 1.08324867e-02 3.51760834e-01 3.81276906e-01
-2.95631383e-02 3.07659149e-01 -5.51341534e-01 1.34192729e+00
-6.66261137e-01 3.10312778e-01 1.63445324e-01 -6.16025388e-01
1.12461507e+00 2.50987530e-01 2.90619910e-01 -8.87001216e-01
2.95704544e-01 -2.61670295e-02 5.09134471e-01 -5.97519457e-01
7.50740886e-01 8.46448541e-02 -2.61967719e-01 6.65135860e-01
5.69590814e-02 -5.00424027e-01 3.46985161e-01 1.71140194e-01
9.14739728e-01 -1.03524581e-01 1.53722525e-01 -7.45179802e-02
5.71063399e-01 -2.40604710e-02 7.24386513e-01 1.12339723e+00
-4.19152379e-01 -1.00355662e-01 8.51128459e-01 -3.54628354e-01
-1.05635810e+00 -7.32784867e-01 3.30077738e-01 1.39940846e+00
6.28673658e-02 -1.58179834e-01 -1.08566415e+00 -1.04016423e+00
3.32881100e-02 1.29550898e+00 -3.82135332e-01 -4.42786902e-01
-7.37453282e-01 -6.96994841e-01 5.59455156e-01 6.40085787e-02
8.17224085e-01 -1.47251260e+00 -7.45436549e-01 6.45060241e-01
-2.57689476e-01 -8.13642621e-01 -6.08133793e-01 1.12536304e-01
-5.25746107e-01 -8.52906406e-01 -4.92959470e-01 -5.47684669e-01
3.49720776e-01 -1.62577227e-01 1.15516090e+00 1.38410062e-01
6.99291527e-02 3.39166343e-01 -4.59404916e-01 -2.84029871e-01
-1.41610134e+00 4.50106561e-01 2.56845981e-01 -2.69392163e-01
1.53906196e-01 -1.87362880e-01 -3.21227461e-01 5.30030668e-01
-7.35722065e-01 4.39310521e-02 -2.21859440e-02 1.29919374e+00
-1.39782593e-01 -3.27626579e-02 1.18684113e+00 -1.00979447e+00
1.69084001e+00 -4.85747904e-01 -7.34353483e-01 3.42154771e-01
-5.49013257e-01 4.83982205e-01 9.84572887e-01 -7.09499240e-01
-1.42696869e+00 -2.91805536e-01 -3.68065655e-01 -2.74917893e-02
-5.08318618e-02 1.49179012e-01 1.14561534e-02 1.14483997e-01
8.16471994e-01 2.55651057e-01 4.24603909e-01 -2.17735358e-02
2.88141340e-01 9.18089986e-01 5.50417230e-02 -8.52100849e-01
2.57625282e-01 -2.20620513e-01 -5.93858898e-01 -9.26979303e-01
-3.82321358e-01 -1.44007010e-02 -9.36345384e-02 -4.64089781e-01
5.87581158e-01 -5.58812678e-01 -1.03329289e+00 8.09292436e-01
-1.03728747e+00 -1.13045037e+00 -1.81498036e-01 5.07869609e-02
-8.17518890e-01 9.38769653e-02 -5.99026680e-01 -1.20547211e+00
-4.19708878e-01 -1.52863538e+00 6.99323118e-01 7.84771144e-01
-5.54619074e-01 -1.06490290e+00 4.20214504e-01 2.04781979e-01
6.39366865e-01 -1.54231831e-01 8.21048617e-01 -1.05700183e+00
-2.67322004e-01 2.41104051e-01 5.13985395e-01 1.83380082e-01
2.08315611e-01 -2.36655101e-02 -6.56101823e-01 -4.10646349e-01
1.20743163e-01 -8.93864036e-01 4.02307540e-01 2.63563663e-01
7.10402966e-01 -5.44478834e-01 -1.20415531e-01 1.13624342e-01
4.76263493e-01 7.38533556e-01 3.91485959e-01 3.47319007e-01
1.07467815e-01 7.89188802e-01 8.56417239e-01 8.06672454e-01
4.65875655e-01 7.03203619e-01 2.82234401e-01 1.61456272e-01
5.21699488e-01 -1.69649184e-01 5.47458231e-01 5.53731918e-01
2.51593351e-01 -6.29935801e-01 -7.07988679e-01 1.23164497e-01
-2.10623145e+00 -9.63800371e-01 4.91911948e-01 1.95296180e+00
1.31160808e+00 3.40105295e-01 5.04266083e-01 -4.74294037e-01
6.03388667e-01 1.21677980e-01 -9.71924603e-01 -7.49202132e-01
1.57872379e-01 -1.34386584e-01 1.28607929e-01 8.41183126e-01
-7.28583038e-01 1.27391005e+00 5.86030293e+00 5.92958212e-01
-1.01280487e+00 -1.87056333e-01 7.79509127e-01 -4.27011326e-02
-2.85720348e-01 -4.36877668e-01 -9.58192050e-01 6.33805752e-01
1.00761151e+00 -2.46399194e-01 7.39761829e-01 7.25526392e-01
4.96702552e-01 -4.27575856e-01 -8.79292488e-01 4.60154980e-01
-2.50388235e-01 -1.07798231e+00 -2.77204216e-01 -1.11465059e-01
7.09193587e-01 -8.09782669e-02 9.83433202e-02 9.23008025e-01
1.08129561e+00 -7.94729292e-01 2.45648623e-01 4.41719145e-01
2.29658216e-01 -8.81382704e-01 5.95325053e-01 9.88353550e-01
-3.74404520e-01 -2.47937977e-01 -1.78846896e-01 2.25926712e-01
2.23807171e-01 -1.58022732e-01 -1.42792165e+00 -4.32441980e-02
2.37772316e-01 8.02862123e-02 -1.43141448e-01 5.55221438e-01
-1.81019872e-01 5.61652124e-01 -1.64468899e-01 -6.13956034e-01
5.76727808e-01 -3.64045262e-01 7.84811020e-01 9.23055768e-01
-1.83421507e-01 2.06280038e-01 3.61720562e-01 7.14584887e-01
2.88306903e-02 -1.67060897e-01 -6.38424873e-01 -2.18495414e-01
6.32140756e-01 9.74625885e-01 -1.89668208e-01 -2.18036503e-01
1.77697495e-01 9.74706888e-01 5.83663225e-01 3.30202848e-01
-7.29019105e-01 -2.04654232e-01 7.75797069e-01 -3.17093194e-01
-1.89480200e-01 -1.43638864e-01 -6.01119697e-02 -9.04592514e-01
-4.80010986e-01 -1.54821181e+00 2.07997754e-01 -3.13244522e-01
-1.25741673e+00 9.19993401e-01 -2.41774127e-01 -8.38978171e-01
-9.51009452e-01 -8.13765004e-02 -5.21445394e-01 8.24779987e-01
-1.17240441e+00 -3.96472514e-01 -8.55579749e-02 4.27728057e-01
1.12554395e+00 -7.96969771e-01 7.80981600e-01 -2.89867878e-01
-7.24495232e-01 6.36869550e-01 2.40048110e-01 -1.70829311e-01
7.54708648e-01 -1.49148989e+00 5.14050007e-01 7.99174234e-02
-6.10572577e-01 5.06850362e-01 9.89916563e-01 -8.05507064e-01
-1.22167528e+00 -7.51368344e-01 4.04449165e-01 -1.58725038e-01
3.98653448e-01 -2.97101229e-01 -1.02407348e+00 2.66006231e-01
5.25643349e-01 -7.90461779e-01 4.36060935e-01 1.76524799e-02
1.78770438e-01 1.55551597e-01 -1.33882046e+00 1.00669372e+00
5.95621765e-01 -3.52909863e-02 -5.35671949e-01 4.41963047e-01
7.41807520e-01 -7.66712546e-01 -5.72322965e-01 1.47129878e-01
2.07948104e-01 -1.05777311e+00 5.89559972e-01 -8.62542450e-01
1.03957698e-01 2.01128289e-01 3.31185967e-01 -2.07109475e+00
6.02530763e-02 -1.15064692e+00 -1.64795399e-01 9.14439976e-01
3.60375673e-01 -9.04491842e-01 7.05603242e-01 9.87089574e-01
-2.15837788e-02 -7.46196568e-01 -7.04409361e-01 -4.90285844e-01
2.43334204e-01 3.81075114e-01 5.61061263e-01 7.61349142e-01
3.91497284e-01 6.95069075e-01 -6.45926952e-01 -1.37237415e-01
3.28428894e-01 -2.64100760e-01 1.07729220e+00 -9.66904163e-01
-4.93178785e-01 -5.36850274e-01 5.04339814e-01 -1.35166883e+00
4.82960939e-01 -2.23746076e-01 3.99872750e-01 -1.18109202e+00
-3.06906134e-01 -8.22827876e-01 4.60981876e-01 2.40358680e-01
-4.50574934e-01 -9.04510200e-01 2.25825220e-01 3.49103883e-02
-6.34343207e-01 1.13905382e+00 1.47459733e+00 -4.52590361e-02
-9.17844057e-01 5.05748451e-01 -4.54556346e-01 5.13901055e-01
1.13244462e+00 -1.98577672e-01 -7.73647606e-01 -4.37939465e-02
6.02353364e-02 7.18349278e-01 -3.46682549e-01 -6.07707262e-01
2.79223889e-01 -4.71899480e-01 -2.99498104e-02 -3.01553309e-01
3.63225371e-01 -3.60887229e-01 -3.46551150e-01 7.12950170e-01
-7.32836545e-01 1.09409370e-01 2.75387615e-01 3.79484147e-01
1.39831424e-01 -4.35767353e-01 1.01443410e+00 -2.70006746e-01
-3.35744500e-01 5.77894747e-02 -1.00241911e+00 4.30961132e-01
9.96433854e-01 2.10360438e-01 -5.27377665e-01 -9.50186670e-01
-4.86878544e-01 1.07539487e+00 2.69265950e-01 5.55900812e-01
4.46837306e-01 -7.67306566e-01 -6.82161748e-01 2.02529386e-01
-2.15967312e-01 1.36311967e-02 8.35388154e-02 2.41719559e-01
-4.17411655e-01 3.03740174e-01 -2.03569174e-01 -4.87887979e-01
-1.06333661e+00 9.27346647e-02 9.21512306e-01 -6.37641966e-01
-4.09267068e-01 8.15088451e-01 -7.98359327e-03 -9.74851608e-01
5.58524907e-01 3.36621441e-02 -4.85071599e-01 1.26863390e-01
3.95920962e-01 2.25013867e-01 -2.44792864e-01 -1.43601462e-01
4.10609059e-02 -1.94462329e-01 -6.04894638e-01 -4.98274714e-01
1.03833413e+00 -5.42179681e-02 2.91773558e-01 3.59910011e-01
5.92713416e-01 -1.92831129e-01 -1.72390747e+00 -1.15447514e-01
-1.04187630e-01 -2.72814393e-01 -2.89790899e-01 -1.09475565e+00
-6.09669447e-01 5.63295186e-01 2.03788221e-01 7.12812841e-01
7.29251146e-01 -4.74881113e-01 8.23688984e-01 7.22645462e-01
4.66382831e-01 -1.45973957e+00 5.93953788e-01 1.13227153e+00
9.04904127e-01 -1.25169146e+00 -5.16996264e-01 2.02535823e-01
-1.45268106e+00 1.02393091e+00 1.30536139e+00 -2.06034929e-02
5.06813452e-02 2.76016086e-01 3.31492037e-01 -6.04499914e-02
-1.29460835e+00 -3.20166536e-02 -3.07554603e-01 6.03983462e-01
1.99897647e-01 -6.67644739e-02 -2.25493878e-01 2.89045304e-01
-1.20044537e-01 -2.20519945e-01 6.60210490e-01 8.44550371e-01
-7.27346957e-01 -1.41696227e+00 -1.87767237e-01 2.78934836e-01
8.69253203e-02 3.08183849e-01 -6.52587593e-01 7.12848365e-01
-4.14647043e-01 1.16295016e+00 -3.16408463e-02 -3.24992090e-01
5.20949066e-01 6.76182136e-02 2.79419720e-01 -5.62825084e-01
-1.26387727e+00 -1.09581100e-02 5.23784757e-01 -3.01313698e-01
8.32479075e-02 -6.14701092e-01 -1.43986404e+00 -3.29663336e-01
-3.43127847e-01 5.76835334e-01 3.00809026e-01 8.73218715e-01
2.78133243e-01 7.16504037e-01 1.15172637e+00 -5.78596652e-01
-1.40644896e+00 -1.15045357e+00 -3.33167881e-01 1.36830345e-01
3.95962119e-01 -7.70784318e-01 -3.16666037e-01 -5.99782407e-01] | [13.020994186401367, 8.062849998474121] |
015eade8-1e95-4644-9331-d36048d3e509 | stable-yaw-estimation-of-boats-from-the | 2306.14056 | null | https://arxiv.org/abs/2306.14056v1 | https://arxiv.org/pdf/2306.14056v1.pdf | Stable Yaw Estimation of Boats from the Viewpoint of UAVs and USVs | Yaw estimation of boats from the viewpoint of unmanned aerial vehicles (UAVs) and unmanned surface vehicles (USVs) or boats is a crucial task in various applications such as 3D scene rendering, trajectory prediction, and navigation. However, the lack of literature on yaw estimation of objects from the viewpoint of UAVs has motivated us to address this domain. In this paper, we propose a method based on HyperPosePDF for predicting the orientation of boats in the 6D space. For that, we use existing datasets, such as PASCAL3D+ and our own datasets, SeaDronesSee-3D and BOArienT, which we annotated manually. We extend HyperPosePDF to work in video-based scenarios, such that it yields robust orientation predictions across time. Naively applying HyperPosePDF on video data yields single-point predictions, resulting in far-off predictions and often incorrect symmetric orientations due to unseen or visually different data. To alleviate this issue, we propose aggregating the probability distributions of pose predictions, resulting in significantly improved performance, as shown in our experimental evaluation. Our proposed method could significantly benefit downstream tasks in marine robotics. | ['Andreas Zell', 'Timon Höfer', 'Benjamin Kiefer'] | 2023-06-24 | null | null | null | null | ['trajectory-prediction'] | ['computer-vision'] | [-1.79556563e-01 -3.70987505e-02 5.10660410e-01 -4.34773654e-01
-1.71863168e-01 -1.11608255e+00 4.99824584e-01 -2.61762410e-01
-4.53008980e-01 3.64804178e-01 -1.86582450e-02 -3.57903481e-01
1.15720909e-02 -6.77655399e-01 -9.91617858e-01 -5.29946983e-01
-3.23585898e-01 3.50075603e-01 7.55958617e-01 -3.62902254e-01
3.24668765e-01 6.60762191e-01 -1.60858738e+00 -1.63382694e-01
4.94708478e-01 8.65838587e-01 2.65558690e-01 9.62964475e-01
3.38254601e-01 3.25637251e-01 -4.93423551e-01 -3.30397934e-01
7.34804034e-01 1.28635183e-01 -4.74640340e-01 1.18840680e-01
8.28095257e-01 -5.87349832e-01 -9.40333679e-02 8.64559412e-01
2.62763351e-01 2.01274008e-01 6.34771764e-01 -1.28301978e+00
1.02064297e-01 1.48175564e-02 -3.81490529e-01 -1.91198793e-04
3.79113555e-01 1.81271896e-01 7.70412445e-01 -7.58046865e-01
5.48192322e-01 1.21191609e+00 9.27130640e-01 3.90294403e-01
-6.95504010e-01 -3.66574019e-01 4.59771842e-01 6.79065064e-02
-1.42048371e+00 -3.25396806e-01 3.27330709e-01 -6.28918946e-01
7.77338266e-01 3.08732450e-01 9.44986999e-01 7.31408656e-01
4.68522519e-01 7.31574833e-01 6.94331229e-01 6.47844896e-02
2.60867447e-01 -2.28850871e-01 -5.35351932e-01 7.32406616e-01
3.60394597e-01 9.33536366e-02 -3.50675285e-01 6.57914728e-02
8.15837204e-01 8.93872753e-02 -4.46817279e-01 -8.01149249e-01
-1.45171201e+00 5.60347795e-01 4.62007195e-01 -6.63225293e-01
-2.07701743e-01 3.93878445e-02 5.95326945e-02 1.03652760e-01
4.78756964e-01 5.20941496e-01 -8.93335283e-01 -2.44566157e-01
-5.43874979e-01 7.41869509e-01 8.58894229e-01 1.56871343e+00
5.27265668e-01 8.18852261e-02 4.06086922e-01 3.39336336e-01
5.97663343e-01 8.24913919e-01 1.57235432e-02 -1.01128590e+00
4.69219148e-01 4.52759653e-01 7.49027371e-01 -1.26434374e+00
-6.56726182e-01 -1.30658075e-01 -2.61408895e-01 3.81403923e-01
2.78568149e-01 -5.32893240e-01 -9.08607602e-01 1.28465784e+00
6.08243942e-01 -2.50491742e-02 2.08357841e-01 1.37698817e+00
6.89211071e-01 6.82023227e-01 -4.19655472e-01 1.43205300e-01
1.13217497e+00 -8.95062566e-01 -3.32199931e-01 -4.89300698e-01
8.22473764e-01 -8.45038295e-01 6.10362709e-01 4.73661929e-01
-4.89256978e-01 -3.22973818e-01 -9.11548615e-01 3.01783174e-01
-2.95473099e-01 2.63700634e-01 4.84062880e-01 2.52664089e-01
-1.04610276e+00 5.39624512e-01 -1.18734539e+00 -6.03088081e-01
9.39344913e-02 3.22516769e-01 -4.30455565e-01 1.47247896e-01
-6.85012639e-01 1.00902045e+00 2.69689888e-01 3.09815198e-01
-9.94227111e-01 -3.89031291e-01 -1.21730375e+00 -3.76143903e-01
4.87498134e-01 -5.14463007e-01 1.38913119e+00 -4.38034743e-01
-1.28574228e+00 4.33137715e-01 8.60187188e-02 -5.34247756e-01
6.55326188e-01 -7.24715352e-01 -1.21274613e-01 -6.27405420e-02
-4.87622172e-02 9.31642413e-01 7.94327557e-01 -1.20681334e+00
-1.01173711e+00 -5.75080872e-01 4.40586716e-01 6.61472321e-01
1.21223211e-01 -4.62174296e-01 -4.25474763e-01 -6.09928489e-01
5.75056911e-01 -1.49402642e+00 -4.51360375e-01 2.96957701e-01
-2.40207151e-01 1.53912023e-01 1.11460626e+00 -5.26441991e-01
6.51855826e-01 -2.04523134e+00 2.95579255e-01 -2.55001247e-01
-5.61377853e-02 1.40364602e-01 5.28155901e-02 3.94569784e-01
3.77331913e-01 -1.01087295e-01 1.10639967e-02 -1.29239351e-01
-2.31453136e-01 3.73605400e-01 -3.94328445e-01 7.04927325e-01
1.07687682e-01 4.34684753e-01 -9.92416501e-01 -2.04015285e-01
2.91049391e-01 2.59161860e-01 -8.63792479e-01 3.67141664e-01
-3.26528221e-01 5.05262971e-01 -5.72527707e-01 8.52466047e-01
8.96602631e-01 3.15955222e-01 -2.91467905e-02 -3.00247014e-01
-5.61170936e-01 4.10858244e-02 -9.79606271e-01 1.51184833e+00
-2.31949747e-01 6.79355443e-01 2.32602451e-02 -5.09288847e-01
1.05416358e+00 5.67970313e-02 3.19561332e-01 8.17064419e-02
7.31563335e-03 -1.95864271e-02 -1.42854348e-01 -4.61698472e-01
1.02787924e+00 8.90932456e-02 -8.94410908e-02 -3.01820368e-01
-9.58486553e-03 -7.69206643e-01 3.52456495e-02 -7.96272680e-02
9.23502982e-01 7.60520577e-01 3.07524711e-01 -2.45401695e-01
3.27546149e-02 5.22702456e-01 5.67483664e-01 4.40614164e-01
-2.07283005e-01 6.74748778e-01 3.30601245e-01 -7.85419583e-01
-9.08581376e-01 -8.57258976e-01 -1.01771846e-01 6.60722792e-01
6.09603882e-01 -5.82016766e-01 -4.89861190e-01 -7.91029096e-01
1.52357861e-01 4.36993241e-01 -3.33409786e-01 2.07908571e-01
-3.72313559e-01 -5.90509117e-01 1.33323342e-01 5.74417412e-01
1.92065939e-01 -4.17207092e-01 -1.02730739e+00 8.77396613e-02
6.78337142e-02 -1.49876833e+00 -1.75327331e-01 8.49388316e-02
-9.45552468e-01 -1.19018376e+00 -6.97343230e-01 -4.03722376e-01
7.28051782e-01 7.28992283e-01 7.49291360e-01 -3.73608112e-01
5.23476116e-02 5.61327338e-01 -7.40274131e-01 -7.46211708e-01
-1.84278004e-02 -2.81322509e-01 5.79296649e-01 -3.29091638e-01
7.38330185e-02 -2.82107025e-01 -7.09591448e-01 9.29528534e-01
-5.23112714e-01 1.00110218e-01 4.62333113e-01 4.57292199e-01
5.13364196e-01 -6.97850958e-02 -8.82744491e-02 -3.94005269e-01
-3.67909335e-02 -3.60912949e-01 -1.13400495e+00 -7.49326646e-02
-1.51421160e-01 -1.10485576e-01 7.12861121e-01 -2.19653800e-01
-5.45991123e-01 5.20321727e-01 1.06773369e-01 -7.36976743e-01
-3.40957195e-01 3.92031878e-01 7.29781017e-02 -4.76625949e-01
4.28884327e-01 5.97875230e-02 5.19888662e-02 -3.90781462e-01
2.58169062e-02 5.71677089e-01 3.11657965e-01 -1.79495409e-01
8.82370651e-01 4.54228729e-01 2.97626823e-01 -9.40455735e-01
-7.64857352e-01 -4.31338310e-01 -7.83761322e-01 -5.10983527e-01
8.57028723e-01 -1.29830396e+00 -6.75188065e-01 4.44133252e-01
-1.23172951e+00 -2.07551405e-01 3.38772774e-01 8.74387503e-01
-5.97405016e-01 3.90329540e-01 -5.67783676e-02 -8.57490659e-01
2.29736213e-02 -1.53018188e+00 1.40967941e+00 2.80121326e-01
-9.98758804e-03 -7.64568865e-01 5.26544340e-02 9.99166667e-02
1.51767833e-02 3.46641690e-01 1.39737979e-01 -5.43022990e-01
-8.64375114e-01 -2.85261810e-01 8.01094025e-02 9.13127884e-02
2.94232089e-03 1.93464220e-01 -6.44804001e-01 -4.01402116e-01
-3.72887433e-01 -2.08541766e-01 6.41576767e-01 3.11354816e-01
8.21774602e-01 -2.19954818e-01 -3.94857973e-01 7.87599683e-01
1.12396574e+00 9.97996256e-02 1.02999426e-01 4.26737010e-01
7.45163858e-01 7.20232904e-01 1.47973430e+00 8.62316787e-01
7.31850803e-01 9.49751794e-01 1.26332843e+00 2.50218272e-01
5.23060203e-01 -3.36549491e-01 5.15394449e-01 4.93387789e-01
-5.04978597e-01 -4.69436646e-01 -9.65275466e-01 4.36034262e-01
-1.86029541e+00 -4.78720784e-01 -3.54532450e-01 2.31031299e+00
8.25715624e-03 -3.80312428e-02 -3.58578973e-02 -3.12305629e-01
3.78236175e-01 1.18943565e-01 -4.86579388e-01 -1.03737444e-01
2.69938648e-01 -6.16194427e-01 1.16741765e+00 3.93396020e-01
-1.39618683e+00 1.14992893e+00 5.76588392e+00 3.17119211e-01
-1.17728853e+00 -6.10233307e-01 -9.45666507e-02 3.49053778e-02
-4.94224243e-02 1.99309513e-01 -1.24761891e+00 7.89379999e-02
4.17122066e-01 2.22357556e-01 2.92518109e-01 1.07787287e+00
1.34650767e-01 -2.11413831e-01 -9.78448510e-01 7.79748917e-01
1.64159641e-01 -9.50470030e-01 8.77759755e-02 1.02962345e-01
4.52375144e-01 2.30360955e-01 -1.55480072e-01 2.05465317e-01
2.61967361e-01 -4.66933042e-01 1.10126626e+00 3.81216913e-01
4.35975462e-01 -6.45557642e-01 9.64742422e-01 6.55427694e-01
-1.21422398e+00 6.09007105e-02 -6.85265601e-01 -2.53384382e-01
2.19752595e-01 1.33418038e-01 -1.50603664e+00 6.24661386e-01
9.97749984e-01 9.79104757e-01 -5.99005044e-01 1.38429797e+00
-1.17517278e-01 2.78947324e-01 -6.12097025e-01 -3.40007573e-01
3.95017266e-01 -2.68678516e-01 9.73669887e-01 7.21963525e-01
8.50658953e-01 1.50736675e-01 3.84092301e-01 1.96050718e-01
4.71679240e-01 -3.89208198e-02 -1.04122162e+00 7.54178315e-02
3.14892679e-01 1.31407928e+00 -6.59527719e-01 1.06253736e-01
-3.86475533e-01 7.01031089e-01 6.42232504e-03 1.94262251e-01
-9.68563318e-01 -2.36595154e-01 8.59401107e-01 2.29760259e-01
6.11038327e-01 -7.75432706e-01 2.83856332e-01 -1.29823685e+00
-3.26743051e-02 -3.80202919e-01 -6.57425914e-03 -9.56157684e-01
-7.99395859e-01 6.97086632e-01 2.58304864e-01 -2.19248056e+00
-1.89638555e-01 -1.01612270e+00 -3.07662129e-01 3.87470752e-01
-1.31894052e+00 -1.10224795e+00 -5.93012214e-01 9.35638323e-02
7.93485403e-01 -7.76840076e-02 6.74649417e-01 -2.38301218e-01
-1.72245085e-01 -4.16355915e-02 -2.35895097e-01 -1.19721621e-01
7.98977792e-01 -1.27560163e+00 6.95665896e-01 8.56227636e-01
7.16297850e-02 2.81852037e-01 1.11142516e+00 -7.15294242e-01
-1.84210157e+00 -1.35459495e+00 3.12865555e-01 -7.63321280e-01
6.61675453e-01 -2.39496484e-01 -4.19179350e-01 9.01538730e-01
-2.66536981e-01 1.83003694e-01 4.25104022e-01 -1.22794151e-01
1.95844203e-01 4.77443524e-02 -8.92074764e-01 7.90000260e-01
1.10486495e+00 3.06033432e-01 -4.30078626e-01 4.31419760e-01
6.77476645e-01 -1.08275998e+00 -8.74633431e-01 8.35723996e-01
7.07964361e-01 -1.02367973e+00 8.83412957e-01 -5.45366645e-01
2.42474347e-01 -7.71825910e-01 -3.71131927e-01 -1.59029853e+00
-9.86746475e-02 -4.49104875e-01 7.96278939e-02 6.13638163e-01
3.70689124e-01 -2.77591407e-01 7.53470063e-01 3.98597807e-01
-7.16814160e-01 -6.21876299e-01 -7.43819654e-01 -7.29122043e-01
-4.90255415e-01 -3.66427869e-01 4.15743500e-01 4.43725020e-01
-2.48882636e-01 -7.49613196e-02 -6.84003949e-01 1.04657185e+00
4.54014271e-01 2.35980317e-01 1.44069421e+00 -1.19438982e+00
-3.81370112e-02 1.32336780e-01 -9.69049692e-01 -1.72407126e+00
-2.38606762e-02 -2.31534973e-01 5.41478992e-01 -1.27710581e+00
-5.30856907e-01 -3.07572097e-01 5.02792180e-01 3.45111966e-01
5.83783388e-02 9.75042060e-02 1.10789396e-01 5.94179668e-02
-7.93007314e-01 7.02260733e-01 1.34969580e+00 1.68694437e-01
-7.21785277e-02 3.77221912e-01 -1.47366852e-01 1.20972574e+00
5.13940394e-01 -3.06679666e-01 -3.19971591e-01 -7.60814786e-01
2.50415444e-01 3.94195616e-01 5.40319204e-01 -1.18856311e+00
2.73641258e-01 -2.28644624e-01 4.92787749e-01 -9.68972623e-01
6.97783053e-01 -1.09223151e+00 -1.55542968e-02 2.69894868e-01
3.03024292e-01 3.19796920e-01 3.52651715e-01 8.40877235e-01
-1.39022082e-01 -1.97995931e-01 4.53924686e-01 -1.80355325e-01
-1.06317830e+00 4.15734857e-01 -5.40397584e-01 -1.20824419e-01
1.22930872e+00 -3.18237901e-01 -3.01622123e-01 -4.26871836e-01
-5.35962403e-01 4.40649986e-01 8.58818293e-01 6.05592668e-01
8.90706778e-01 -7.90971994e-01 -4.41275716e-01 3.75955909e-01
3.90499741e-01 6.68909252e-01 2.56471366e-01 9.53056514e-01
-1.09365535e+00 4.26128447e-01 -1.75322473e-01 -1.14429724e+00
-1.21261930e+00 2.58680940e-01 1.12055764e-01 5.53990245e-01
-4.62408960e-01 9.05162156e-01 3.50139290e-01 -8.12443137e-01
-8.40757880e-03 -6.94426179e-01 -3.19314271e-01 -2.13763744e-01
3.04969609e-01 2.08574802e-01 8.31183270e-02 -7.92648673e-01
-4.31575477e-01 8.66393268e-01 4.52815369e-02 1.20059319e-01
1.34799469e+00 -2.48479024e-01 3.69767636e-01 2.06544206e-01
6.37386322e-01 2.47728556e-01 -1.88770378e+00 3.19592625e-01
-3.29374582e-01 -7.49141932e-01 2.61079278e-02 -2.93684006e-01
-8.19703102e-01 7.16320157e-01 4.31267530e-01 1.99496686e-01
6.86678171e-01 -1.03295341e-01 5.79476655e-01 7.50411451e-01
8.97237539e-01 -5.60416639e-01 -4.28064317e-01 8.60675037e-01
8.38616073e-01 -1.41617227e+00 9.98602062e-02 -3.88378054e-01
-1.03146577e+00 1.35286796e+00 8.66283298e-01 -2.60418683e-01
4.08511221e-01 1.90155923e-01 2.59699732e-01 -1.87714137e-02
-7.42454410e-01 -1.12879559e-01 3.96112561e-01 6.55378997e-01
7.21882656e-02 4.56548594e-02 1.75348952e-01 4.19858605e-01
-5.19178391e-01 -3.73864144e-01 8.65238428e-01 1.24046981e+00
-6.51807368e-01 -5.80369174e-01 -5.26893318e-01 1.45300478e-01
-1.27274618e-01 -2.63234414e-02 -3.12793583e-01 9.40972149e-01
3.54485847e-02 7.24108517e-01 1.47935912e-01 -8.17525506e-01
6.38822913e-01 -5.22497654e-01 4.43581998e-01 -5.24213731e-01
3.06030717e-02 4.23069187e-02 2.60782093e-01 -5.48946559e-01
-5.29041767e-01 -7.49974012e-01 -1.15725219e+00 3.89920212e-02
-5.00101626e-01 1.33059472e-01 8.20385993e-01 9.04747903e-01
4.16394621e-01 9.38927010e-02 6.27726436e-01 -1.42861187e+00
-5.56428254e-01 -8.53578508e-01 -5.50371945e-01 -1.73901111e-01
4.30449843e-01 -9.99684155e-01 -4.26395804e-01 1.60220459e-01] | [7.551446437835693, -1.8543404340744019] |
fb2432e6-f140-464d-9ea4-027a03abc1c4 | deepvel-deep-learning-for-the-estimation-of | 1703.05128 | null | http://arxiv.org/abs/1703.05128v2 | http://arxiv.org/pdf/1703.05128v2.pdf | DeepVel: deep learning for the estimation of horizontal velocities at the solar surface | Many phenomena taking place in the solar photosphere are controlled by plasma
motions. Although the line-of-sight component of the velocity can be estimated
using the Doppler effect, we do not have direct spectroscopic access to the
components that are perpendicular to the line-of-sight. These components are
typically estimated using methods based on local correlation tracking. We have
designed DeepVel, an end-to-end deep neural network that produces an estimation
of the velocity at every single pixel and at every time step and at three
different heights in the atmosphere from just two consecutive continuum images.
We confront DeepVel with local correlation tracking, pointing out that they
give very similar results in the time- and spatially-averaged cases. We use the
network to study the evolution in height of the horizontal velocity field in
fragmenting granules, supporting the buoyancy-braking mechanism for the
formation of integranular lanes in these granules. We also show that DeepVel
can capture very small vortices, so that we can potentially expand the scaling
cascade of vortices to very small sizes and durations. | ['I. S. Requerey', 'A. Asensio Ramos', 'N. Vitas'] | 2017-03-15 | null | null | null | null | ['3d-depth-estimation'] | ['computer-vision'] | [-3.82317334e-01 -1.82091966e-01 2.80634791e-01 -6.46224916e-02
1.68193698e-01 -5.53138435e-01 7.22697258e-01 6.23399317e-02
-3.93421561e-01 6.30027294e-01 -2.04140976e-01 -2.98332334e-01
-7.93302134e-02 -1.02320814e+00 -3.03632349e-01 -1.07800901e+00
-3.28568667e-01 8.44915330e-01 2.05886021e-01 -5.46397269e-01
2.98066676e-01 7.49738812e-01 -1.47047877e+00 4.22563553e-02
6.21743679e-01 1.05369318e+00 -1.62142947e-01 1.08137274e+00
-1.48871064e-01 6.56983852e-01 -4.14396286e-01 3.52912515e-01
3.33256841e-01 -7.81580448e-01 -3.29303831e-01 3.26981470e-02
5.81502378e-01 -3.34194720e-01 -7.76655525e-02 9.01423693e-01
3.69037300e-01 2.43867949e-01 8.54306042e-01 -6.58458173e-01
-2.81399578e-01 7.44575113e-02 -9.79770958e-01 5.24312913e-01
4.90083024e-02 7.95456469e-02 9.28960443e-01 -5.31488478e-01
8.62218022e-01 8.97451520e-01 8.31249177e-01 1.80988371e-01
-1.14016140e+00 1.05814040e-01 -4.26646680e-01 4.86292653e-02
-9.78677452e-01 -3.79644275e-01 7.54605770e-01 -7.42350996e-01
1.06893551e+00 1.09034300e-01 8.95149410e-01 5.55463433e-01
3.11406195e-01 -2.88130976e-02 1.06714249e+00 -7.24382401e-01
3.23082358e-01 3.82690988e-02 1.17914513e-01 7.49099672e-01
2.84067541e-01 3.98107201e-01 -6.24444127e-01 -2.93973058e-01
1.00815582e+00 -3.73117983e-01 -5.48203588e-01 -5.73735535e-01
-1.01683629e+00 9.50576484e-01 5.75185478e-01 4.46124047e-01
-3.22157651e-01 -2.36114915e-02 1.63563296e-01 2.05145881e-01
2.62632966e-01 7.33721793e-01 -3.69884998e-01 -2.17508212e-01
-9.00370777e-01 3.08361471e-01 1.03167009e+00 1.91726327e-01
6.80050254e-01 6.80789724e-02 2.66190827e-01 2.37452418e-01
9.15954784e-02 8.35956872e-01 7.69394815e-01 -1.23142660e+00
-6.99831620e-02 3.59433979e-01 3.99042040e-01 -8.80925536e-01
-8.23879778e-01 -3.44988793e-01 -9.49123085e-01 8.00890744e-01
7.12716758e-01 -6.07091010e-01 -5.14532268e-01 1.30464602e+00
3.30698639e-01 -5.18526174e-02 3.53553444e-01 9.92362797e-01
3.74479204e-01 6.59315586e-01 -8.80698740e-01 -6.54125333e-01
1.12403023e+00 -6.97929800e-01 -4.02364999e-01 -3.45296301e-02
4.95146364e-01 -3.02385598e-01 6.98566258e-01 2.06864610e-01
-8.66459608e-01 -3.76476437e-01 -1.17815101e+00 6.11472912e-02
-4.39717025e-01 -2.75777817e-01 6.82636023e-01 1.56027779e-01
-9.15193081e-01 1.07346630e+00 -1.08531070e+00 -3.34173173e-01
-1.54449582e-01 -1.12936728e-01 -1.89726859e-01 9.50491428e-01
-6.92543209e-01 7.25349724e-01 -4.16987501e-02 3.25105548e-01
-1.89449713e-01 -4.14713293e-01 -4.43378240e-01 1.29552290e-01
-2.62381822e-01 -6.11079752e-01 1.51704192e+00 -9.72801030e-01
-1.38223982e+00 6.01863861e-01 -4.94805634e-01 -7.34902620e-01
5.97484767e-01 4.67208065e-02 -1.47255450e-01 3.43979597e-01
-1.95579559e-01 4.09996271e-01 6.61439538e-01 -1.02304304e+00
-5.16780853e-01 -4.85828638e-01 -4.18017954e-01 2.85453707e-01
-1.40742853e-01 -4.21861738e-01 1.06131449e-01 2.79657394e-01
1.93976492e-01 -8.89708042e-01 -4.61677387e-02 1.30871832e-01
-1.75232694e-01 -1.32906567e-02 8.38284850e-01 -3.81579310e-01
4.12013739e-01 -1.67566264e+00 5.71795218e-02 1.86065082e-02
2.12954387e-01 2.37473294e-01 2.59331137e-01 4.11515057e-01
7.93565959e-02 -1.60976306e-01 -3.79254460e-01 -1.77873194e-01
-3.25136274e-01 -2.61795446e-02 -5.06635904e-01 7.06146538e-01
-1.60816267e-01 7.24896967e-01 -8.07615638e-01 -6.32019341e-02
3.14747304e-01 6.05208278e-01 -2.37134889e-01 -7.08948821e-02
-3.86156350e-01 3.34784329e-01 -1.34428233e-01 9.34236646e-02
8.54762018e-01 -3.18426371e-01 2.38136724e-01 1.04954280e-01
-8.47961545e-01 1.95373908e-01 -7.34314680e-01 1.16981983e+00
-3.56858760e-01 1.44887352e+00 1.57526881e-01 -7.51845241e-01
9.19494331e-01 -3.43069173e-02 4.08079863e-01 -1.00154805e+00
1.35183349e-01 1.89205438e-01 3.91109675e-01 -4.01661515e-01
5.10637105e-01 -5.78023612e-01 6.37653172e-01 3.44528049e-01
-3.39320242e-01 -3.58947963e-01 2.28693157e-01 3.67794335e-02
8.34782243e-01 1.01716947e-02 1.77019879e-01 -4.82511431e-01
1.25664100e-01 7.55403042e-02 1.95121229e-01 6.69079900e-01
-2.17044950e-01 7.51645625e-01 9.93492067e-01 -5.89341462e-01
-1.19650102e+00 -6.76114798e-01 -6.30819321e-01 3.07276160e-01
1.90805212e-01 -3.64763945e-01 -3.28275949e-01 -1.31346434e-01
3.42414409e-01 3.80614042e-01 -8.48769367e-01 8.85856971e-02
-2.94442296e-01 -9.64409053e-01 2.37271823e-02 3.90307754e-01
5.04492462e-01 -1.29598415e+00 -1.27640283e+00 1.13861866e-01
2.13224724e-01 -9.76919889e-01 3.03068429e-01 4.03574169e-01
-7.57601559e-01 -1.10149431e+00 -4.08901066e-01 -2.88035482e-01
3.56057882e-01 1.86350808e-01 1.01304233e+00 -7.18470439e-02
-4.05151218e-01 -2.73426231e-02 -1.63849026e-01 -4.68991667e-01
-1.86796710e-01 -3.75502795e-01 2.27654263e-01 -2.20883071e-01
4.01611000e-01 -6.84249222e-01 -5.34546852e-01 9.01470482e-02
-2.00790867e-01 -1.28779680e-01 -2.54161566e-01 6.21695161e-01
3.17638725e-01 3.12904874e-03 -2.68771239e-02 -3.67699355e-01
5.32051146e-01 -1.61035970e-01 -1.39875519e+00 -1.54486895e-01
-2.96968669e-01 9.92057249e-02 6.88793004e-01 -5.22989705e-02
-9.33265388e-01 -4.57641602e-01 5.97933047e-02 -3.65141332e-01
5.32531291e-02 2.88263291e-01 8.37114751e-01 -2.32547417e-01
9.40714300e-01 1.24152377e-01 2.24368408e-01 -1.37753457e-01
2.58641899e-01 5.49083412e-01 5.82527101e-01 -3.00032794e-02
6.51205957e-01 1.16933405e+00 4.95275497e-01 -1.66687822e+00
-6.53682530e-01 -6.06308818e-01 -8.70698154e-01 -6.05838075e-02
8.95650327e-01 -7.99524486e-01 -1.17210448e+00 5.38072467e-01
-1.11947322e+00 -7.04627752e-01 -7.20344126e-01 4.68242139e-01
-4.65575248e-01 3.62642914e-01 -5.58382273e-01 -1.06971490e+00
-3.06264192e-01 -6.55317307e-01 9.14262414e-01 6.90790951e-01
-9.12840962e-02 -1.33124053e+00 7.30695844e-01 -2.05130965e-01
4.58129138e-01 3.19040328e-01 6.04761958e-01 5.09194331e-03
-2.81114668e-01 5.32503501e-02 -1.79190591e-01 7.53166005e-02
-6.11457974e-02 7.18729556e-01 -1.11644852e+00 -1.93206698e-01
4.73679304e-01 -3.12078297e-01 1.28386033e+00 1.03637528e+00
5.66850483e-01 2.06182361e-01 -4.42859262e-01 1.01169991e+00
1.38391995e+00 2.76283890e-01 3.68514091e-01 5.19869804e-01
5.94442189e-01 5.12079895e-01 1.34560600e-01 5.04979610e-01
-1.07895076e-01 6.68775320e-01 3.68468732e-01 -3.88106883e-01
1.66012257e-01 2.09562838e-01 1.66799068e-01 5.27592778e-01
-4.79474545e-01 -1.07386410e-01 -7.35358953e-01 3.81703496e-01
-1.83249378e+00 -1.32256830e+00 -5.90979755e-01 2.43032670e+00
2.16162041e-01 1.87066823e-01 5.52868296e-04 -3.53534281e-01
2.57581055e-01 3.59237522e-01 -7.13307381e-01 -6.04282796e-01
-3.41899753e-01 4.54225987e-02 5.68608820e-01 1.01590025e+00
-7.05611408e-01 5.44221342e-01 7.15053320e+00 -2.29471043e-01
-1.61965871e+00 -1.82436466e-01 3.73624921e-01 -3.08148712e-01
-2.43019581e-01 8.50147158e-02 -8.16777647e-01 3.45002651e-01
9.55423892e-01 -2.54695620e-02 4.40709054e-01 5.64977825e-01
5.05775988e-01 -4.69488442e-01 -5.44129670e-01 8.64694297e-01
-5.65047562e-02 -1.63091302e+00 -5.31666756e-01 3.51048172e-01
8.44918966e-01 7.00554430e-01 -3.19422215e-01 -6.47021756e-02
-9.60566849e-03 -6.50360823e-01 2.47073963e-01 6.92070186e-01
6.73419476e-01 -4.91701961e-01 5.48080564e-01 4.59920764e-01
-9.93433356e-01 5.13415486e-02 -9.10500407e-01 -4.84988153e-01
2.82588661e-01 8.61126184e-01 -8.88174355e-01 1.73951522e-01
6.37208939e-01 5.78562975e-01 -3.23690295e-01 9.87844467e-01
-2.00966746e-01 2.86740899e-01 -7.96959877e-01 -2.06761166e-01
2.12264001e-01 -1.00979960e+00 5.48297346e-01 7.61560798e-01
5.68704844e-01 -4.22878526e-02 -3.71854246e-01 9.06598866e-01
-6.30365917e-03 -2.53984749e-01 -8.54919553e-01 -8.33627060e-02
4.76332940e-02 1.61430573e+00 -8.19260478e-01 -3.71084660e-01
-3.43759149e-01 7.36012280e-01 3.97947669e-01 3.36699158e-01
-3.61632526e-01 -4.17688370e-01 8.14359844e-01 2.47029468e-01
7.39888608e-01 -5.01138985e-01 -7.00152963e-02 -1.21704471e+00
2.23560147e-02 -1.75102621e-01 -1.38861835e-01 -1.30250299e+00
-1.02672982e+00 5.98018050e-01 -3.86127353e-01 -6.61302865e-01
-3.74386668e-01 -1.08532751e+00 -1.17548108e+00 1.01834500e+00
-1.40659547e+00 -3.19721371e-01 -5.54348528e-01 1.82666808e-01
3.09002586e-02 -1.47979587e-01 7.40484178e-01 -6.96928740e-01
-3.25231940e-01 -4.13460582e-01 8.42929482e-01 -1.98668435e-01
5.26799202e-01 -1.56817544e+00 5.18669426e-01 8.35864842e-01
2.50909507e-01 4.62657034e-01 1.00226057e+00 -5.41961968e-01
-1.18243957e+00 -3.86671424e-01 6.60822451e-01 -4.40160483e-01
8.64292443e-01 -4.44465309e-01 -9.91066396e-01 4.87827033e-01
5.54221511e-01 4.35301691e-01 1.66292161e-01 1.91533789e-01
1.19095117e-01 -5.81585169e-02 -8.18931997e-01 2.24194869e-01
6.65783346e-01 -6.62483871e-01 -4.38774914e-01 5.56997359e-01
4.00802225e-01 -3.77086431e-01 -3.17391127e-01 2.62713403e-01
6.79271400e-01 -1.82485497e+00 5.16345978e-01 -4.14225101e-01
3.96869957e-01 -3.93314987e-01 2.36942872e-01 -1.36969769e+00
-1.25023231e-01 -7.25320697e-01 -1.71959028e-01 6.45064235e-01
3.74455780e-01 -9.17832077e-01 9.37294185e-01 -4.70407493e-02
2.06594110e-01 -5.10243773e-01 -9.55673277e-01 -5.41553199e-01
2.71057844e-01 -4.04613353e-02 5.50035685e-02 9.91123259e-01
1.61051661e-01 6.92843258e-01 8.30070227e-02 3.12227577e-01
5.58767974e-01 7.86674798e-01 4.76008028e-01 -1.50885403e+00
-4.20945287e-01 -2.86840230e-01 -3.11313458e-02 -8.80083799e-01
-1.03352986e-01 -4.37384248e-01 2.04599351e-01 -1.42234230e+00
-9.67348143e-02 -1.13471806e-01 1.74729124e-01 2.99344789e-02
3.25232774e-01 -7.73375407e-02 -1.11518037e-02 3.84372592e-01
7.82788079e-03 5.45186818e-01 1.09018421e+00 3.49194914e-01
-3.81303579e-01 -2.62554377e-01 -1.00566484e-01 8.98791552e-01
8.57590616e-01 5.52543029e-02 -2.66568065e-01 -1.37086064e-01
4.83456641e-01 1.33675247e-01 2.55592346e-01 -1.36731160e+00
5.88979721e-02 -5.11750802e-02 7.16153264e-01 -5.09656429e-01
5.25149226e-01 -3.04621667e-01 -4.39145386e-01 3.19067955e-01
-2.33243760e-02 7.18319565e-02 1.77693009e-01 2.62490600e-01
-1.13349326e-01 -3.93411189e-01 1.07326436e+00 -3.61144245e-01
-3.09022814e-01 5.67844398e-02 -3.80937934e-01 -1.44843891e-01
8.65752578e-01 6.00983277e-02 -8.73723567e-01 -4.96273458e-01
-3.40997875e-01 6.21870048e-02 9.62663472e-01 -1.56781435e-01
2.08338544e-01 -7.74326622e-01 -3.27258945e-01 5.69390714e-01
-2.21135825e-01 3.12331002e-02 1.15122169e-01 9.57449198e-01
-1.11406267e+00 4.42872375e-01 -2.07786560e-01 -8.34836900e-01
-1.05570960e+00 5.22535980e-01 1.12632060e+00 -2.45180260e-02
-7.57218838e-01 1.04089785e+00 1.75562441e-01 -3.19774479e-01
-3.73518109e-01 -3.23147148e-01 -3.04203838e-01 3.43131423e-01
6.61246777e-01 1.70432374e-01 5.11248223e-02 -5.72196126e-01
-1.12925187e-01 8.93019617e-01 2.11797029e-01 -2.46617287e-01
1.08213723e+00 -1.99980456e-02 -3.52990746e-01 8.36798787e-01
1.02574396e+00 2.07727998e-01 -1.57872093e+00 3.52666706e-01
-5.74588537e-01 -1.26421541e-01 3.14746261e-01 -3.81827384e-01
-8.53486121e-01 1.27648807e+00 6.86926544e-01 8.42660606e-01
6.93666875e-01 -1.88558325e-01 3.98846000e-01 6.42543375e-01
1.72878373e-02 -8.21795523e-01 -1.90338075e-01 7.45337546e-01
4.92586702e-01 -1.22042990e+00 -2.42133215e-02 2.63433997e-02
-3.06778193e-01 1.53947604e+00 5.15526712e-01 -1.34351492e-01
3.67833555e-01 3.97042692e-01 4.33469445e-01 -3.77624363e-01
-1.00763226e+00 -1.27319247e-01 -1.98056772e-01 5.43539703e-01
4.09138590e-01 3.74191180e-02 -1.83690917e-02 -8.12522113e-01
-3.83951634e-01 -3.22622538e-01 1.01220500e+00 6.52551413e-01
-8.95257473e-01 -7.36933231e-01 -5.05251765e-01 2.92684823e-01
2.74870634e-01 4.16887328e-02 -3.20897311e-01 7.05655873e-01
1.58051059e-01 5.79611123e-01 9.06445324e-01 1.87386200e-01
-1.39013175e-02 8.43462944e-02 6.36025786e-01 -1.65627584e-01
1.47295311e-01 1.91671941e-02 1.44920275e-01 -1.93960756e-01
-3.93015176e-01 -8.42502534e-01 -1.46262419e+00 -2.45072365e-01
-1.90305024e-01 4.84265566e-01 7.68546462e-01 7.78237581e-01
9.06936079e-02 3.53364319e-01 6.00905359e-01 -1.37007427e+00
-2.06196979e-01 -1.05692542e+00 -1.17479491e+00 2.30083987e-01
1.01958036e+00 -4.27510262e-01 -9.57456529e-01 -1.67322695e-01] | [6.690825462341309, 3.246318817138672] |
de45e7a5-d176-4415-a7a6-935d9a6bf010 | a-bayesian-evaluation-framework-for-ground | 2007.06711 | null | https://arxiv.org/abs/2007.06711v2 | https://arxiv.org/pdf/2007.06711v2.pdf | A Bayesian Evaluation Framework for Subjectively Annotated Visual Recognition Tasks | An interesting development in automatic visual recognition has been the emergence of tasks where it is not possible to assign objective labels to images, yet still feasible to collect annotations that reflect human judgements about them. Machine learning-based predictors for these tasks rely on supervised training that models the behavior of the annotators, i.e., what would the average person's judgement be for an image? A key open question for this type of work, especially for applications where inconsistency with human behavior can lead to ethical lapses, is how to evaluate the epistemic uncertainty of trained predictors, i.e., the uncertainty that comes from the predictor's model. We propose a Bayesian framework for evaluating black box predictors in this regime, agnostic to the predictor's internal structure. The framework specifies how to estimate the epistemic uncertainty that comes from the predictor with respect to human labels by approximating a conditional distribution and producing a credible interval for the predictions and their measures of performance. The framework is successfully applied to four image classification tasks that use subjective human judgements: facial beauty assessment, social attribute assignment, apparent age estimation, and ambiguous scene labeling. | ['Walter J. Scheirer', 'Mel McCurrie', 'Derek S. Prijatelj'] | 2020-06-20 | null | null | null | null | ['scene-labeling'] | ['computer-vision'] | [ 4.62055653e-01 8.23246062e-01 -5.80902584e-02 -7.90989995e-01
-4.65159774e-01 -3.62903446e-01 6.49526119e-01 2.82657713e-01
-5.77324867e-01 6.86604440e-01 -2.09703986e-02 -1.50917202e-01
-1.71777546e-01 -3.49642545e-01 -6.98537230e-01 -8.02747369e-01
5.30758381e-01 5.83107829e-01 -6.56237006e-02 2.59421527e-01
3.40025604e-01 1.00972436e-01 -1.74783218e+00 1.32725954e-01
7.29714215e-01 1.49527335e+00 -4.13269252e-01 4.55783755e-01
3.56323689e-01 1.07895315e+00 -5.04675627e-01 -1.27220726e+00
-3.47646810e-02 -2.30954453e-01 -8.61067832e-01 3.75162870e-01
3.35704565e-01 -2.06057116e-01 4.56334233e-01 1.40557754e+00
-1.26067281e-01 -3.53062600e-01 1.35556567e+00 -1.39505994e+00
-8.57354283e-01 5.61781466e-01 -4.85263050e-01 -2.60160685e-01
4.28154558e-01 3.48298438e-02 9.83827055e-01 -5.32104075e-01
6.17107153e-01 1.33029509e+00 7.13596821e-01 5.89426696e-01
-1.46520269e+00 -4.14038032e-01 1.12506114e-01 3.70518714e-01
-1.41470528e+00 -5.40047705e-01 4.78107125e-01 -1.10764849e+00
-1.36610307e-02 1.84394076e-01 2.80817509e-01 1.54079044e+00
1.93373695e-01 3.06698591e-01 1.70181584e+00 -6.72325015e-01
5.66274822e-01 8.75014544e-01 1.65096700e-01 4.96569574e-01
3.18224579e-01 2.97786474e-01 -6.69989169e-01 -3.13270807e-01
3.23202461e-01 -6.19762421e-01 1.49350855e-02 -5.11100590e-01
-8.87293160e-01 7.97211647e-01 1.56373039e-01 -5.81072606e-02
-5.76824129e-01 1.00384153e-01 3.66629899e-01 -1.49994213e-02
6.17495894e-01 3.99831295e-01 -3.01040530e-01 -1.68974604e-02
-6.85584843e-01 1.47261769e-01 7.78021812e-01 5.66188872e-01
5.38642704e-01 -3.37158442e-01 -2.65051365e-01 7.00694144e-01
4.90404308e-01 4.69328880e-01 2.63672203e-01 -1.34144890e+00
-2.60546416e-01 4.19189572e-01 5.48203528e-01 -1.18332267e+00
-8.12225342e-02 -1.41863301e-01 -4.93967772e-01 6.36335492e-01
9.68642473e-01 9.28548723e-03 -6.78224683e-01 1.93992889e+00
4.23331261e-01 -1.30566522e-01 -1.98583692e-01 9.48445559e-01
4.19914275e-01 9.15520787e-02 5.76230645e-01 -3.31953853e-01
1.53401709e+00 -2.90539473e-01 -4.55275714e-01 -3.05129737e-01
4.62945163e-01 -5.09061694e-01 9.34030712e-01 7.11586475e-01
-7.30562449e-01 -4.28644329e-01 -9.32166874e-01 2.17556685e-01
-2.36372277e-02 3.86560321e-01 5.80761731e-01 1.13972092e+00
-8.12591016e-01 6.64789200e-01 -4.33120847e-01 -3.08127165e-01
4.83907104e-01 3.09909105e-01 -4.75845665e-01 2.49404624e-01
-1.05345166e+00 1.40147007e+00 3.02410781e-01 3.50970775e-01
-9.96104956e-01 -1.52829394e-01 -5.27333915e-01 7.20862241e-04
3.55890334e-01 -4.99571145e-01 1.24985778e+00 -1.95324302e+00
-1.33307528e+00 1.56673360e+00 4.24927622e-02 -4.17669863e-01
8.98316920e-01 1.13540627e-01 -1.70750692e-01 -2.44678766e-01
2.08753034e-01 6.56263113e-01 1.35884619e+00 -1.47688985e+00
-4.77501363e-01 -7.27856636e-01 -1.63285702e-01 8.69051367e-02
-7.39888400e-02 2.86800414e-01 2.30094075e-01 -1.24780215e-01
-3.00132646e-03 -1.24703801e+00 -2.54673302e-01 2.30147213e-01
-4.48119581e-01 -4.57712352e-01 -7.20763288e-04 -6.26685798e-01
8.14988315e-01 -2.21575475e+00 -1.22103296e-01 3.62112731e-01
2.66160846e-01 -2.54506115e-02 4.08381402e-01 -2.49645114e-01
1.94617547e-02 1.59182221e-01 -1.80380225e-01 -3.97900790e-01
5.17075479e-01 1.23187691e-01 -2.97255158e-01 7.85511553e-01
5.95720820e-02 3.41179490e-01 -6.41441643e-01 -7.52847612e-01
-1.05345689e-01 3.47778708e-01 -1.98712111e-01 2.67388582e-01
-1.10431924e-01 5.07611811e-01 -2.16282010e-01 4.03338194e-01
4.85891491e-01 -2.14139596e-01 2.47520685e-01 -1.05068259e-01
8.04762840e-02 1.96924573e-03 -9.66316640e-01 9.63880837e-01
-6.67762607e-02 5.21281898e-01 3.82425636e-02 -9.24849451e-01
1.21625400e+00 2.74945229e-01 3.02461330e-02 -3.57853055e-01
4.11248207e-01 2.28199318e-01 -7.68381134e-02 -5.64271748e-01
2.03775391e-01 -4.24480647e-01 -2.25599959e-01 5.05688488e-01
-9.28110536e-03 6.51859641e-02 -2.50019878e-01 -1.15220800e-01
8.11039567e-01 3.57608706e-01 5.42523623e-01 -2.91736782e-01
6.06323481e-01 -3.09740394e-01 7.70327508e-01 7.39580989e-01
-6.62258625e-01 3.44541520e-01 1.08121157e+00 -6.43114567e-01
-1.21644127e+00 -1.01936352e+00 -3.73735726e-01 1.14452636e+00
-5.88018931e-02 -2.24291421e-02 -1.10604978e+00 -7.88141966e-01
-1.38010293e-01 9.62989390e-01 -1.11569536e+00 -3.01395476e-01
2.42941067e-01 -5.23328483e-01 3.34360927e-01 2.76345164e-01
1.50558893e-02 -1.00251007e+00 -9.09051597e-01 -2.16364920e-01
-1.84885845e-01 -9.87642467e-01 9.60940570e-02 4.94942628e-02
-3.50089669e-01 -9.03423965e-01 -4.39358771e-01 -7.52321407e-02
8.36303115e-01 -6.19664073e-01 1.25205040e+00 -6.96667610e-03
-1.39666665e-02 5.80406666e-01 -2.85874605e-01 -7.22845912e-01
-5.89301944e-01 -3.32011670e-01 1.98160991e-01 5.91766238e-01
6.60311222e-01 -3.40749085e-01 -4.90163267e-01 6.12200081e-01
-4.99023259e-01 -1.19425729e-01 5.04468322e-01 7.23653615e-01
3.66040945e-01 -2.24319085e-01 2.69060880e-01 -1.17701471e+00
4.46595758e-01 -3.04130495e-01 -5.64785242e-01 6.38812602e-01
-9.10704613e-01 5.50867505e-02 8.57055262e-02 -4.95868236e-01
-1.43500805e+00 1.52308673e-01 2.16800064e-01 -1.27881855e-01
-6.31212354e-01 1.50056913e-01 2.05882583e-02 6.78785145e-02
9.71620679e-01 -3.79772216e-01 -4.17106114e-02 -1.25864938e-01
2.63179302e-01 9.11483884e-01 6.93926573e-01 -9.01644051e-01
4.89822268e-01 3.61049861e-01 1.55044854e-01 -5.81431806e-01
-1.38095844e+00 -1.06062345e-01 -7.63870478e-01 -7.73890793e-01
1.07017684e+00 -6.87774897e-01 -1.29920483e+00 2.98644215e-01
-1.32827878e+00 -1.10468224e-01 -1.98897883e-01 5.29896319e-01
-9.14928436e-01 2.25909576e-01 -1.55198857e-01 -1.44996071e+00
-2.13742014e-02 -1.14900124e+00 8.83796036e-01 8.80023837e-02
-9.05237377e-01 -8.74970257e-01 -1.41034156e-01 8.06780279e-01
1.66065827e-01 3.93276244e-01 8.97500396e-01 -8.10710192e-01
-1.70982629e-01 -3.30037326e-01 -3.48098427e-01 6.57208979e-01
-4.67474222e-01 1.87247440e-01 -1.41316319e+00 1.66748285e-01
2.82880425e-01 -8.80985200e-01 4.35472369e-01 4.34334010e-01
1.08184552e+00 -4.95864600e-01 -9.61059704e-02 1.13787457e-01
1.12770772e+00 1.84199754e-02 7.42802739e-01 2.34799296e-01
2.21425086e-01 1.24349892e+00 6.46190405e-01 4.97917473e-01
2.79860646e-01 7.51839578e-01 5.49179137e-01 3.55038643e-01
4.33475941e-01 -2.89745808e-01 3.59514147e-01 4.70554084e-02
-3.55733752e-01 7.24845529e-02 -9.58102822e-01 1.83699265e-01
-1.89705789e+00 -7.65437901e-01 -1.42433703e-01 2.53036022e+00
7.06233799e-01 1.86749682e-01 -9.27263033e-03 1.50925517e-01
9.27400947e-01 -1.96827769e-01 -5.21224260e-01 -6.34064257e-01
2.14844823e-01 -3.83814782e-01 3.53397906e-01 3.80862832e-01
-1.15339124e+00 5.59266329e-01 6.19816399e+00 6.58385456e-01
-6.46704972e-01 2.55288929e-01 1.34046400e+00 4.36572820e-01
-1.24888361e-01 3.01459044e-01 -6.24934614e-01 5.09379745e-01
1.06356180e+00 -8.92485604e-02 2.44016021e-01 1.08703983e+00
-1.49255693e-02 -5.84352434e-01 -1.53530324e+00 8.12930286e-01
2.62120605e-01 -5.89690745e-01 -3.22671622e-01 1.46635041e-01
4.83856320e-01 -6.25945926e-01 3.13550264e-01 8.25938061e-02
4.58200723e-01 -1.35375333e+00 1.24062860e+00 9.69846487e-01
7.89027870e-01 -4.21113104e-01 9.54664350e-01 4.43596005e-01
-1.89087510e-01 -1.91174775e-01 -3.53000045e-01 -2.52933770e-01
-2.33826578e-01 5.74518204e-01 -1.02853191e+00 -2.50805408e-01
6.75838828e-01 1.85179204e-01 -7.43420243e-01 7.57197022e-01
-5.05319059e-01 6.03370845e-01 -9.49156880e-02 -5.15427329e-02
-7.79321641e-02 -2.23525986e-01 2.07578436e-01 8.75534594e-01
2.62906402e-01 -1.23955801e-01 -1.48242667e-01 1.05124605e+00
1.64346043e-02 2.25357428e-01 -4.72442150e-01 1.98971421e-01
3.12849075e-01 1.31520939e+00 -7.65783429e-01 3.20575796e-02
-1.00443043e-01 6.74914181e-01 5.07732689e-01 1.22575119e-01
-7.37092376e-01 2.96657681e-01 1.22285262e-01 7.19271079e-02
-2.44508296e-01 4.34538871e-01 -6.64575636e-01 -7.17254579e-01
8.92363638e-02 -8.97643566e-01 3.31165761e-01 -1.19279623e+00
-1.48426974e+00 5.32249808e-01 -3.16375233e-02 -9.05545413e-01
-3.03715557e-01 -7.50340044e-01 -2.77584314e-01 7.45059788e-01
-8.32403600e-01 -1.25681782e+00 -2.16276050e-01 3.03511798e-01
-2.55924404e-01 1.32745011e-02 9.16059077e-01 -2.09964156e-01
-3.53660017e-01 5.18931210e-01 -3.73841673e-01 9.00096297e-02
9.12400186e-01 -1.26715732e+00 -4.78053868e-01 4.55665082e-01
3.23863439e-02 2.70159036e-01 1.25223124e+00 -3.53489518e-01
-5.20652592e-01 -4.70136285e-01 1.03162527e+00 -9.31928337e-01
7.76204467e-01 -1.56979680e-01 -8.05874228e-01 7.75099337e-01
-3.22604738e-02 2.92263553e-02 8.22341919e-01 4.44164455e-01
-7.23020613e-01 -2.86410809e-01 -1.23097408e+00 3.25747222e-01
6.67835176e-01 -5.03211617e-01 -4.52156574e-01 3.08236063e-01
5.35369441e-02 1.19595136e-03 -8.84720802e-01 1.99970290e-01
8.11938047e-01 -1.36911476e+00 6.34092391e-01 -5.98934233e-01
6.38536453e-01 -1.32554412e-01 -2.50715971e-01 -1.11910439e+00
-2.37729475e-01 -2.10936770e-01 3.60775083e-01 1.37815452e+00
5.98555028e-01 -4.57667530e-01 7.24156618e-01 1.50602913e+00
3.87606084e-01 -4.89067703e-01 -1.29123259e+00 -4.21807140e-01
2.08879658e-03 -5.53003311e-01 1.98064879e-01 7.81536102e-01
1.37056177e-02 2.88766354e-01 -6.48068368e-01 1.30567461e-01
1.07984281e+00 -2.34271988e-01 6.74920022e-01 -1.61497796e+00
-1.93840832e-01 -3.53350520e-01 -7.64641106e-01 -1.88587993e-01
6.25247598e-01 -4.30560172e-01 4.08139437e-01 -8.03332746e-01
5.54686368e-01 -3.86655718e-01 -1.37844428e-01 2.30029717e-01
-3.37158963e-02 1.96549699e-01 9.13639367e-02 2.58714914e-01
-8.31388533e-01 3.16054195e-01 7.56011784e-01 -3.89063149e-03
4.25991148e-01 1.57447293e-01 -7.66865969e-01 1.26587105e+00
4.32403505e-01 -6.85868263e-01 -1.43124059e-01 -1.03312051e-02
7.73435175e-01 9.29217264e-02 7.74789512e-01 -8.42152357e-01
2.64076889e-01 -2.86257386e-01 5.13904870e-01 1.53024405e-01
3.35173547e-01 -9.27170932e-01 3.78091425e-01 1.27627805e-01
-9.72123802e-01 -3.67615819e-01 -5.24469435e-01 6.83547795e-01
-1.22954035e-02 -7.06167579e-01 9.58263814e-01 -1.85595945e-01
-3.50516230e-01 -9.68943164e-02 -3.94185007e-01 -9.34172720e-02
1.27669859e+00 -4.76832204e-02 -2.29936987e-01 -5.54424107e-01
-1.15613687e+00 -9.89924744e-02 6.19527996e-01 6.25365153e-02
3.43425393e-01 -1.26193511e+00 -8.87489617e-01 -2.75880486e-01
4.10380006e-01 -6.40409470e-01 1.13435484e-01 9.95791376e-01
-1.93858624e-01 1.01369798e-01 -2.25082070e-01 -6.62748039e-01
-1.39222431e+00 6.15894854e-01 3.84813875e-01 -1.11215577e-01
1.12020180e-01 8.42942476e-01 3.41699868e-01 -4.53546420e-02
2.97139525e-01 1.19817391e-01 -2.83481181e-01 2.06375659e-01
4.16481644e-01 4.01567727e-01 -2.80536175e-01 -9.87798393e-01
-3.34089458e-01 3.21177721e-01 7.31566995e-02 -2.39146605e-01
9.96304452e-01 -3.38362455e-01 -3.80061328e-01 8.24153960e-01
6.85228765e-01 -3.26361239e-01 -1.49420416e+00 -1.55960545e-01
4.49346840e-01 -6.44409835e-01 -9.14641544e-02 -1.05487096e+00
-5.58325171e-01 9.35821056e-01 6.96744442e-01 2.83700198e-01
8.55994105e-01 1.61865681e-01 -3.03290457e-01 2.48222575e-01
5.23002148e-01 -1.54702723e+00 -3.01067606e-02 -1.20157503e-01
9.77762401e-01 -1.56997573e+00 1.40035823e-01 -3.57973844e-01
-1.17515576e+00 9.50780213e-01 5.58107853e-01 2.03384504e-01
6.48910761e-01 -1.87487334e-01 2.17615932e-01 -1.11403853e-01
-7.92380691e-01 1.05321750e-01 4.41973031e-01 7.57104456e-01
3.17731440e-01 5.50066173e-01 -3.05519104e-01 8.07121336e-01
-1.54856727e-01 1.28393918e-01 6.74298704e-01 2.38419592e-01
-2.18952775e-01 -6.87464774e-01 -6.18486285e-01 4.46883202e-01
-6.31636441e-01 3.68977070e-01 -5.73923945e-01 2.92096347e-01
6.05577648e-01 1.04502451e+00 1.92905311e-03 -2.59059995e-01
9.30165946e-02 2.64686674e-01 3.72697532e-01 -3.90493155e-01
-2.47130886e-01 -3.56419325e-01 5.04579484e-01 -3.98400456e-01
-6.83075845e-01 -9.56611335e-01 -4.30414200e-01 -2.08304331e-01
-3.09025288e-01 2.28599176e-01 7.53342211e-01 1.14667737e+00
-2.27626681e-01 -2.28699073e-01 4.98903841e-01 -6.20947421e-01
-8.79132926e-01 -8.89767706e-01 -8.16794693e-01 7.94821680e-01
1.11906203e-02 -8.59018743e-01 -6.32612586e-01 2.59251773e-01] | [8.714601516723633, 5.110504150390625] |
69c6dd53-1939-4f40-9f88-1b94abd2854b | slk-ner-exploiting-second-order-lexicon | 2007.08416 | null | https://arxiv.org/abs/2007.08416v1 | https://arxiv.org/pdf/2007.08416v1.pdf | SLK-NER: Exploiting Second-order Lexicon Knowledge for Chinese NER | Although character-based models using lexicon have achieved promising results for Chinese named entity recognition (NER) task, some lexical words would introduce erroneous information due to wrongly matched words. Existing researches proposed many strategies to integrate lexicon knowledge. However, they performed with simple first-order lexicon knowledge, which provided insufficient word information and still faced the challenge of matched word boundary conflicts; or explored the lexicon knowledge with graph where higher-order information introducing negative words may disturb the identification. To alleviate the above limitations, we present new insight into second-order lexicon knowledge (SLK) of each character in the sentence to provide more lexical word information including semantic and word boundary features. Based on these, we propose a SLK-based model with a novel strategy to integrate the above lexicon knowledge. The proposed model can exploit more discernible lexical words information with the help of global context. Experimental results on three public datasets demonstrate the validity of SLK. The proposed model achieves more excellent performance than the state-of-the-art comparison methods. | ['Dou Hu', 'Lingwei Wei'] | 2020-07-16 | null | null | null | null | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-4.44903262e-02 -2.23880902e-01 -3.26385647e-01 -1.63225919e-01
-3.82853180e-01 -5.83936691e-01 2.46930495e-01 4.22389150e-01
-7.33223081e-01 8.53288591e-01 3.72518778e-01 -4.04681653e-01
1.64667256e-02 -9.52630162e-01 -6.51514754e-02 -4.02085781e-01
3.15103143e-01 1.08891778e-01 5.23823023e-01 -4.26944524e-01
7.86554754e-01 1.53303996e-01 -1.17786205e+00 3.49765956e-01
1.26064527e+00 6.02690160e-01 4.24993992e-01 -7.67441317e-02
-1.07706428e+00 3.60232651e-01 -7.30436027e-01 -4.38885450e-01
-4.09002341e-02 -3.17865133e-01 -6.41790867e-01 -2.84299701e-01
-2.52877057e-01 8.82691443e-02 2.30085909e-01 1.35341144e+00
6.08377457e-01 5.91070019e-02 4.33022767e-01 -7.06323147e-01
-8.95252585e-01 7.90819764e-01 -4.58378732e-01 3.90708447e-01
3.96506071e-01 -2.42747396e-01 8.58936965e-01 -1.03802967e+00
6.32001340e-01 1.18493605e+00 7.96682954e-01 4.61793572e-01
-4.69669759e-01 -9.53272998e-01 6.68936610e-01 4.06345397e-01
-1.73074293e+00 -1.25592798e-01 7.35576332e-01 -6.41297102e-02
1.29606414e+00 7.80771747e-02 3.43451530e-01 6.13773048e-01
6.47929311e-02 5.74374557e-01 1.17961419e+00 -7.64210999e-01
-5.83590232e-02 3.24395061e-01 6.71744406e-01 6.20117605e-01
6.51141763e-01 -3.67302060e-01 -3.14020216e-01 -1.53832123e-01
4.39824879e-01 -2.24371552e-01 -1.77828878e-01 1.74961016e-01
-1.01395202e+00 7.22116649e-01 -8.90331194e-02 8.70909750e-01
-2.08731130e-01 -6.04759753e-01 5.01662254e-01 -4.01279889e-02
1.78609818e-01 2.74670601e-01 -7.64267683e-01 4.08668257e-02
-6.62715614e-01 -3.02720845e-01 8.42749834e-01 1.06100476e+00
9.91215527e-01 5.71488887e-02 2.46035345e-02 7.24739850e-01
3.78470272e-01 4.87431258e-01 9.77962136e-01 -3.71280615e-03
6.00348234e-01 9.96539891e-01 -5.99740706e-02 -1.32238007e+00
-5.69789886e-01 -3.40529054e-01 -7.19817758e-01 -4.83930141e-01
2.26682961e-01 -2.77200788e-01 -9.69974041e-01 1.49870539e+00
2.90253490e-01 3.52235079e-01 4.34240580e-01 6.80619121e-01
1.10864723e+00 6.55909061e-01 4.13541108e-01 -4.33170974e-01
1.69042325e+00 -6.05204642e-01 -1.22273827e+00 -3.04101110e-01
8.30482602e-01 -1.10331714e+00 9.99584973e-01 1.67132407e-01
-5.55156827e-01 -5.04838407e-01 -1.13324988e+00 2.40852669e-01
-9.75224912e-01 2.82417566e-01 7.24820316e-01 1.18101966e+00
-7.09808052e-01 2.18028143e-01 -3.70069206e-01 -5.95196068e-01
1.22013642e-02 2.31137186e-01 -2.92610765e-01 1.17638230e-01
-1.83427131e+00 1.01077819e+00 1.08099306e+00 3.86826754e-01
-7.08991960e-02 -3.83622676e-01 -8.80219221e-01 -1.80114195e-01
7.56835341e-01 -1.28993317e-01 7.44166970e-01 -7.92058110e-01
-1.09883440e+00 7.41582036e-01 -2.87946552e-01 -2.39124731e-03
1.40172597e-02 -8.11453387e-02 -9.72229004e-01 -1.47180557e-01
2.69125998e-01 1.57309771e-01 1.07580155e-01 -1.00518572e+00
-6.53933346e-01 -3.13772976e-01 -2.33516753e-01 4.05542016e-01
-5.97329140e-01 2.37069651e-01 -5.96942842e-01 -6.53542638e-01
2.24818587e-01 -4.05332744e-01 -1.22087002e-01 -7.37218857e-01
-1.72507614e-01 -3.26016068e-01 5.02406776e-01 -7.59117603e-01
2.03450084e+00 -1.83136070e+00 -5.43762922e-01 2.84232110e-01
-2.03435645e-01 8.84769917e-01 -6.51005656e-02 6.60910785e-01
2.50341177e-01 7.23067284e-01 -1.27662763e-01 2.50175327e-01
-2.89139748e-01 3.51741999e-01 -6.27994910e-02 7.86219910e-03
2.46275827e-01 7.33761787e-01 -8.94541204e-01 -8.74736667e-01
1.15277700e-01 3.08368832e-01 -2.04641744e-01 -1.92822695e-01
-6.78264885e-04 2.65660584e-02 -7.74543583e-01 6.56610191e-01
9.77228701e-01 1.63223326e-01 5.60142457e-01 -2.59112895e-01
-2.50286430e-01 2.50914961e-01 -1.67227805e+00 1.37310767e+00
-3.36834967e-01 2.70277113e-02 -2.99234927e-01 -7.73702264e-01
1.19910359e+00 3.85656387e-01 -1.60644367e-01 -7.21714258e-01
2.80969262e-01 4.18722183e-01 -5.10458276e-02 -7.94270992e-01
5.29583573e-01 -2.51016706e-01 -1.45870492e-01 -4.29801308e-02
-1.28436666e-02 3.78653854e-01 1.19314462e-01 -4.89480682e-02
7.60327160e-01 -1.39306895e-02 9.51070964e-01 -4.31456625e-01
1.08530557e+00 2.22474441e-01 9.67768252e-01 6.75062418e-01
-3.75721723e-01 1.64289355e-01 2.52032727e-01 -2.51934141e-01
-7.22333550e-01 -7.16823459e-01 -1.24527052e-01 8.65566313e-01
4.83352005e-01 -5.22395670e-01 -7.82079041e-01 -7.66464770e-01
-3.86437982e-01 7.80369580e-01 -3.43808293e-01 -4.89157848e-02
-8.74007344e-01 -9.40530658e-01 1.02043259e+00 5.59085548e-01
8.21941853e-01 -1.21265197e+00 -1.31751135e-01 4.93343771e-01
-2.96169609e-01 -1.09684777e+00 -5.68581998e-01 -3.69355902e-02
-7.15243697e-01 -1.12789929e+00 -4.29233879e-01 -1.28634691e+00
6.81278229e-01 1.60302788e-01 8.12843978e-01 3.86257440e-01
-1.48769245e-01 -5.32558002e-02 -7.88283467e-01 -4.43742454e-01
-2.17749298e-01 2.94871598e-01 -4.30683456e-02 -1.27935067e-01
1.07795596e+00 -1.68563694e-01 -2.60498673e-01 3.32981050e-01
-1.01547325e+00 -2.67132699e-01 6.50578737e-01 8.21085453e-01
4.55857724e-01 2.36465856e-01 1.09684765e+00 -9.60563481e-01
8.74216557e-01 -5.40724993e-01 -3.51382703e-01 7.57712305e-01
-8.42127681e-01 3.29916209e-01 8.03482711e-01 -4.94723529e-01
-1.63172150e+00 -1.66686013e-01 -4.07821715e-01 3.99468273e-01
-3.08938146e-01 7.82566190e-01 -6.96531057e-01 9.44662616e-02
2.00871974e-01 4.92112458e-01 -3.47035050e-01 -7.89618134e-01
3.96917433e-01 8.94161642e-01 3.42867792e-01 -7.41784513e-01
3.25348258e-01 1.78759515e-01 -3.51846308e-01 -7.85073221e-01
-7.40609288e-01 -7.68310845e-01 -8.95209193e-01 2.14409307e-02
1.00155187e+00 -8.28089535e-01 -5.42369902e-01 4.94290352e-01
-1.27827728e+00 5.06958663e-01 3.55570912e-01 5.57945609e-01
2.60489762e-01 9.65208590e-01 -5.24843752e-01 -1.00603223e+00
-3.54411900e-01 -7.66170323e-01 4.92633611e-01 6.35967135e-01
5.33690155e-02 -1.17691302e+00 -9.27618220e-02 3.44429500e-02
3.34215850e-01 -1.24774113e-01 1.16392064e+00 -1.33138311e+00
-1.53259829e-01 -1.85562134e-01 -2.06459895e-01 1.60069048e-01
3.39567721e-01 -2.92134643e-01 -5.68979502e-01 2.17240840e-01
8.67621377e-02 3.66433859e-02 7.11159170e-01 -2.05894381e-01
5.82722068e-01 -1.55407742e-01 -4.01317358e-01 1.67635411e-01
1.70789969e+00 5.54017961e-01 6.67450070e-01 6.59280241e-01
6.30261719e-01 5.05354285e-01 8.87115419e-01 4.03030425e-01
4.32150543e-01 3.21203679e-01 1.78893264e-02 6.20953292e-02
6.17102385e-02 -3.65215033e-01 1.25820324e-01 1.44918835e+00
7.03406185e-02 -2.98419118e-01 -1.06905556e+00 7.40874708e-01
-1.74574757e+00 -6.17365897e-01 -2.99233496e-01 1.94100118e+00
1.08877897e+00 3.83073151e-01 -3.87554765e-01 1.15550332e-01
1.03892088e+00 1.14374638e-01 -2.26323873e-01 -4.20025736e-01
-6.77107513e-01 2.15665370e-01 6.10358059e-01 4.80298877e-01
-9.71300662e-01 1.52449214e+00 5.57560825e+00 1.45378017e+00
-9.42860425e-01 1.49247706e-01 1.38045162e-01 6.43374383e-01
-4.87458318e-01 2.41552830e-01 -1.47994888e+00 4.22030836e-01
5.09136558e-01 -3.49126339e-01 -8.27967152e-02 6.76706672e-01
2.84759346e-02 -2.03911871e-01 -2.99920410e-01 7.69667625e-01
4.12655741e-01 -1.00425613e+00 4.37066257e-01 -3.17553669e-01
5.77521741e-01 -3.82764906e-01 -4.54637140e-01 3.67077857e-01
1.99651401e-02 -7.54343092e-01 3.39069098e-01 5.81368208e-01
5.64070284e-01 -8.10202658e-01 1.20973408e+00 4.80068147e-01
-1.55299211e+00 2.86207527e-01 -6.42664313e-01 -2.83084989e-01
2.15332210e-01 4.16286796e-01 -7.56682336e-01 1.00595748e+00
3.06166410e-01 4.18678284e-01 -7.23970592e-01 9.31963027e-01
-3.77550602e-01 6.92163706e-01 -2.35012472e-01 -6.10721409e-01
4.54357654e-01 -2.12643698e-01 3.47368866e-01 1.64902151e+00
2.50101089e-01 4.39566344e-01 2.36137956e-01 3.80363047e-01
2.40283638e-01 1.08843791e+00 -5.15038729e-01 -4.60911989e-02
8.06501150e-01 1.09332848e+00 -1.24343693e+00 -4.73080844e-01
-7.51926899e-01 8.62957418e-01 2.61876285e-01 1.70371205e-01
-5.68313777e-01 -6.99568748e-01 4.13980007e-01 -1.82709292e-01
3.59451979e-01 -2.33248502e-01 -4.97425109e-01 -1.27115202e+00
1.65998906e-01 -6.89624310e-01 6.33798361e-01 -4.34711546e-01
-1.45335972e+00 6.46774173e-01 -6.57294616e-02 -9.19666827e-01
3.33514810e-01 -7.97956705e-01 -5.49550593e-01 9.43349957e-01
-1.69162381e+00 -1.01242507e+00 8.40213429e-03 4.45816159e-01
5.95204473e-01 -2.64796883e-01 9.12799656e-01 4.23736513e-01
-6.44343555e-01 8.16396952e-01 9.76475328e-02 4.60936815e-01
7.51131296e-01 -8.41099322e-01 1.44628942e-01 9.59470093e-01
-7.46999756e-02 1.05561805e+00 2.85076678e-01 -1.27549088e+00
-8.94105792e-01 -6.96146011e-01 1.56627595e+00 -2.02810973e-01
5.75288236e-01 -7.79077262e-02 -1.02731359e+00 4.38985765e-01
2.51856387e-01 -4.36191648e-01 1.06865728e+00 6.84361458e-02
-1.45674512e-01 1.13781221e-01 -9.50994074e-01 7.21806943e-01
1.02984512e+00 -2.95557737e-01 -1.18329847e+00 -1.83605328e-01
8.30016732e-01 -9.95551050e-02 -5.25444090e-01 5.40345788e-01
4.23099250e-01 -5.95716834e-01 7.09590137e-01 -5.56441784e-01
-3.02878767e-01 -6.94540918e-01 -1.81757271e-01 -9.90804613e-01
-2.03274921e-01 -1.54996529e-01 3.83760452e-01 1.74948990e+00
6.23167038e-01 -7.50075400e-01 4.34599549e-01 3.52243751e-01
-1.18836671e-01 -6.03677034e-01 -8.86423588e-01 -8.73566926e-01
1.78405046e-01 -4.72976089e-01 9.21070039e-01 1.14524698e+00
2.80436307e-01 5.08551240e-01 -3.30847919e-01 1.86058179e-01
7.30806738e-02 -2.74098605e-01 -2.40857769e-02 -1.13661051e+00
3.64360273e-01 -4.73552138e-01 -4.67371762e-01 -9.51808393e-01
3.36814106e-01 -8.43508601e-01 -3.23270224e-02 -1.52180302e+00
3.04266363e-01 -4.79739100e-01 -5.46269357e-01 4.31991458e-01
-7.66825140e-01 7.80452415e-02 1.08413711e-01 1.56535823e-02
-6.81016624e-01 3.43346089e-01 9.39413071e-01 1.94474220e-01
-2.90099263e-01 -4.59786087e-01 -8.82964790e-01 8.55082512e-01
8.48475635e-01 -4.68687236e-01 -1.89190149e-01 -5.31280458e-01
4.91721213e-01 -3.56921136e-01 -2.49085769e-01 -7.49365330e-01
6.08122468e-01 -2.89726943e-01 3.49336624e-01 -7.59474158e-01
-2.76718110e-01 -7.50320315e-01 -2.17594933e-02 3.47515672e-01
1.20974968e-04 2.71213204e-01 4.08402890e-01 4.49974805e-01
-3.11568767e-01 -7.44998753e-01 3.34301263e-01 -4.62545633e-01
-1.43192720e+00 -2.52027720e-01 -5.04519582e-01 3.76232952e-01
9.78328705e-01 -4.26407814e-01 -2.84819990e-01 9.94373634e-02
-4.99385089e-01 2.82015115e-01 2.08645701e-01 6.11565053e-01
5.63595414e-01 -1.22784197e+00 -5.17673969e-01 2.67933220e-01
2.53226966e-01 -4.25150335e-01 2.63555735e-01 3.80887508e-01
-6.55392706e-01 6.10139310e-01 -1.62925884e-01 -4.04452309e-02
-1.18902194e+00 7.63331294e-01 1.97829753e-01 -4.70970839e-01
-2.94014961e-01 6.92702413e-01 -1.92760285e-02 -6.41627908e-01
1.02079228e-01 -1.29870012e-01 -9.23165739e-01 3.66054088e-01
5.20810187e-01 2.13327944e-01 1.98878288e-01 -8.98316324e-01
-8.73904288e-01 9.76045191e-01 -1.32046580e-01 1.33540146e-02
6.08870804e-01 -5.05159259e-01 -1.48706466e-01 3.84374559e-01
7.84015477e-01 4.28617835e-01 -2.79073477e-01 -5.31823218e-01
5.81538260e-01 -2.58836836e-01 -2.17237100e-01 -9.82310772e-01
-6.27117515e-01 7.70662844e-01 3.68289441e-01 -1.44862086e-01
1.17629635e+00 -4.58901137e-01 1.00734794e+00 4.52584505e-01
5.57428837e-01 -1.44889438e+00 -3.48631263e-01 8.57773066e-01
1.15624137e-01 -1.25030839e+00 -1.28474936e-01 -7.69740224e-01
-7.97842741e-01 1.28679037e+00 7.86285639e-01 2.15033308e-01
9.28331017e-01 1.77605346e-01 4.78433222e-01 8.84854347e-02
-4.12407279e-01 -6.55579269e-01 1.61814868e-01 5.98477185e-01
4.79063481e-01 -8.15935805e-02 -1.42957544e+00 1.35043585e+00
5.90309203e-02 -1.12709366e-01 4.01267976e-01 1.06115270e+00
-7.56214261e-01 -1.54420567e+00 -3.94383758e-01 1.86664343e-01
-8.41432512e-01 -7.02520549e-01 -4.18234885e-01 8.40469599e-01
5.41238546e-01 1.27458179e+00 -2.51790434e-01 -2.88183838e-01
3.55883181e-01 4.74465549e-01 8.12705383e-02 -6.08067453e-01
-7.16022313e-01 9.28288326e-02 1.71378642e-01 -1.32635072e-01
-5.36733508e-01 -5.10617971e-01 -1.57799292e+00 -4.03052531e-02
-7.81454206e-01 6.01796329e-01 4.58954602e-01 1.12681186e+00
2.14000285e-01 2.92318553e-01 2.35536069e-01 5.56024956e-03
-2.52631813e-01 -1.00524759e+00 -5.67163944e-01 2.29492724e-01
-3.16926599e-01 -5.37026346e-01 -2.42694795e-01 -8.64275843e-02] | [9.78710651397705, 9.671896934509277] |
463c550a-9ae1-4ebf-b116-f0b0a470fbd4 | a-survey-of-vision-language-pre-training-from | 2306.07198 | null | https://arxiv.org/abs/2306.07198v1 | https://arxiv.org/pdf/2306.07198v1.pdf | A Survey of Vision-Language Pre-training from the Lens of Multimodal Machine Translation | Large language models such as BERT and the GPT series started a paradigm shift that calls for building general-purpose models via pre-training on large datasets, followed by fine-tuning on task-specific datasets. There is now a plethora of large pre-trained models for Natural Language Processing and Computer Vision. Recently, we have seen rapid developments in the joint Vision-Language space as well, where pre-trained models such as CLIP (Radford et al., 2021) have demonstrated improvements in downstream tasks like image captioning and visual question answering. However, surprisingly there is comparatively little work on exploring these models for the task of multimodal machine translation, where the goal is to leverage image/video modality in text-to-text translation. To fill this gap, this paper surveys the landscape of language-and-vision pre-training from the lens of multimodal machine translation. We summarize the common architectures, pre-training objectives, and datasets from literature and conjecture what further is needed to make progress on multimodal machine translation. | ['Kevin Duh', 'Jeremy Gwinnup'] | 2023-06-12 | null | null | null | null | ['visual-question-answering-1', 'image-captioning', 'multimodal-machine-translation'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [ 5.67101538e-01 1.99441165e-01 -2.56421864e-01 -5.11357367e-01
-1.46339405e+00 -7.54100323e-01 1.10606277e+00 -1.41063690e-01
-5.48103452e-01 5.26663244e-01 4.81308579e-01 -6.47659183e-01
5.01530111e-01 -2.00546995e-01 -9.39471185e-01 -2.78167456e-01
4.76913780e-01 8.20840061e-01 -3.41744721e-01 -1.89965576e-01
1.12688825e-01 1.83122028e-02 -9.77781594e-01 8.53685975e-01
4.66751158e-01 7.01599002e-01 2.74322242e-01 1.02087319e+00
-1.78030714e-01 6.76088631e-01 2.30693258e-03 -8.11843336e-01
7.14982972e-02 -7.04447448e-01 -1.07907343e+00 3.05551797e-01
7.07180023e-01 -2.53076494e-01 -2.64500380e-01 6.43678188e-01
6.45214438e-01 -1.97484359e-01 5.63615561e-01 -1.35245585e+00
-1.19142175e+00 3.03267747e-01 -5.77934682e-01 -7.65102333e-04
3.55160534e-01 4.82907623e-01 1.08498740e+00 -1.14857137e+00
8.66384566e-01 1.49220693e+00 3.30838889e-01 7.82134116e-01
-1.29750204e+00 -2.58070171e-01 -1.15355730e-01 2.82410920e-01
-8.18310440e-01 -7.46683538e-01 3.79784316e-01 -4.76548493e-01
1.10845542e+00 1.98011920e-01 2.07605213e-01 1.51061034e+00
-6.12750277e-02 1.03664005e+00 1.12400246e+00 -7.84863293e-01
-2.92351991e-01 9.31638330e-02 -3.35769206e-01 6.42758429e-01
-2.66078115e-01 1.26934201e-01 -6.57792985e-01 -3.70410420e-02
6.75406158e-01 -3.76122028e-01 -6.32727295e-02 -4.56089377e-01
-1.93371511e+00 9.91144598e-01 2.18731344e-01 1.30396768e-01
-1.15936242e-01 3.00256908e-01 5.57848632e-01 4.57955986e-01
4.10186380e-01 4.83086765e-01 -5.02159595e-01 -1.05182737e-01
-9.64250267e-01 6.28631637e-02 6.71956122e-01 9.59453404e-01
7.66935349e-01 -2.50569403e-01 -4.83070850e-01 8.20291936e-01
5.55995107e-01 7.39766836e-01 3.47249031e-01 -1.13663244e+00
1.04236317e+00 3.00964743e-01 -5.94375795e-03 -5.94582021e-01
-1.09075539e-01 4.52955328e-02 -6.67811334e-01 -7.22431317e-02
4.94503558e-01 -3.13091576e-01 -1.11314285e+00 1.70749998e+00
-4.77231033e-02 -1.89534605e-01 3.79518270e-01 1.15606475e+00
8.86139929e-01 7.69990385e-01 3.27082485e-01 2.17911415e-02
1.40528262e+00 -1.39822304e+00 -4.67389107e-01 -6.53435349e-01
6.27248824e-01 -1.28126550e+00 1.14520717e+00 -1.16981946e-01
-1.27699745e+00 -5.65806031e-01 -5.31628609e-01 -5.94383180e-01
-3.12243670e-01 2.70809919e-01 5.68889141e-01 3.01035404e-01
-1.41486526e+00 -7.89934024e-02 -6.63903534e-01 -1.01278090e+00
3.92458677e-01 2.89417595e-01 -5.96196175e-01 -3.39491695e-01
-1.12147963e+00 1.24705327e+00 3.54035050e-02 1.70717433e-01
-1.18655801e+00 -3.00016642e-01 -8.70542765e-01 -2.33321846e-01
4.24437448e-02 -1.36541748e+00 1.52416229e+00 -1.53549111e+00
-1.45815241e+00 1.55192363e+00 -4.13117558e-01 -5.33295214e-01
4.47576135e-01 -1.01904996e-01 -4.68158163e-02 3.00758809e-01
3.74878459e-02 1.58641064e+00 9.98613834e-01 -1.16393089e+00
-5.17584622e-01 -3.29867303e-01 2.45843589e-01 5.40851057e-01
-1.27057269e-01 4.63458389e-01 -5.93963325e-01 -2.29365468e-01
-3.81842643e-01 -1.15622556e+00 -2.53309488e-01 3.13315615e-02
-1.79451749e-01 -8.58296677e-02 4.37056780e-01 -7.66942382e-01
5.04624605e-01 -2.06317234e+00 4.58768278e-01 -4.64470059e-01
4.13657278e-02 7.27284178e-02 -7.53922522e-01 7.35207617e-01
-9.57170725e-02 2.52648294e-02 -1.71175286e-01 -6.53132737e-01
1.26136974e-01 2.08845407e-01 -6.23653650e-01 2.90772766e-01
5.01762450e-01 1.71084929e+00 -7.90432155e-01 -5.93880892e-01
2.05287099e-01 4.22318667e-01 -3.79217416e-01 3.89069200e-01
-5.36012232e-01 6.77174628e-01 -1.90655753e-01 6.87049687e-01
2.36871168e-01 -5.37701547e-01 -1.25595272e-01 -1.14684962e-01
-1.65441781e-01 8.75308216e-02 -1.57185182e-01 2.08758807e+00
-3.99757773e-01 1.25206542e+00 3.72434735e-01 -1.14986122e+00
6.04848862e-01 6.09171033e-01 2.74407029e-01 -7.72192061e-01
1.70471534e-01 2.31699526e-01 -1.15620516e-01 -8.53978992e-01
4.37069386e-01 -2.31027633e-01 3.76573019e-03 4.75750387e-01
2.13323146e-01 -3.54798824e-01 7.65299201e-02 2.63714999e-01
6.80107296e-01 3.94621044e-01 -8.13755766e-02 1.77728534e-01
3.66156846e-01 6.83534682e-01 -2.73944795e-01 6.07809544e-01
-3.36161047e-01 1.01436996e+00 1.84830114e-01 -2.29908913e-01
-1.32053828e+00 -8.89099658e-01 2.01069295e-01 1.42690134e+00
-2.13978454e-01 -2.27468714e-01 -8.93027723e-01 -4.71141368e-01
-2.01861531e-01 4.86572057e-01 -6.09158337e-01 1.86828505e-02
-5.22543490e-01 -7.02886403e-01 7.19864428e-01 2.92683572e-01
3.03299278e-01 -1.16902864e+00 -3.41919810e-01 -1.41828626e-01
-6.06821477e-01 -1.51197529e+00 -4.98653829e-01 -1.78730175e-01
-1.04232872e+00 -6.83890760e-01 -1.24859667e+00 -1.15771282e+00
6.62434757e-01 4.17097032e-01 1.36053622e+00 -3.77212673e-01
-1.34125456e-01 1.00492704e+00 -1.91380009e-01 -4.11147445e-01
-5.56195080e-01 1.53702036e-01 -3.72657090e-01 -1.12052858e-02
4.17603850e-01 -1.30333856e-01 -5.53187549e-01 2.25222006e-01
-8.25167298e-01 5.80302358e-01 1.01397908e+00 8.05872917e-01
5.05641162e-01 -1.31693149e+00 3.73411417e-01 -3.44582587e-01
7.84269929e-01 -4.49828923e-01 -2.34003380e-01 5.60723186e-01
-3.42347503e-01 -4.73899916e-02 1.42598435e-01 -4.69935775e-01
-8.87784839e-01 1.44402057e-01 1.03777960e-01 -5.38255394e-01
-2.87235379e-01 6.46286547e-01 1.59071103e-01 3.94328088e-02
5.45089483e-01 2.21527517e-01 1.33883536e-01 -2.89503783e-01
1.02255154e+00 6.56097710e-01 6.47312462e-01 -6.48569882e-01
6.55450165e-01 4.36178446e-01 -7.02528656e-02 -6.54521286e-01
-7.32031286e-01 -3.86197567e-01 -6.80464327e-01 -3.40341381e-03
1.38562095e+00 -1.18956399e+00 -2.97566563e-01 8.58286247e-02
-1.50245976e+00 -4.85728830e-01 -3.87048870e-02 4.80666935e-01
-9.32976663e-01 3.81682575e-01 -5.05018830e-01 -4.83279437e-01
-5.01986980e-01 -1.23595810e+00 1.51617634e+00 9.81691703e-02
-1.41841888e-01 -1.22554159e+00 3.81534129e-01 9.81199145e-01
4.63637114e-01 -9.36769769e-02 9.60920393e-01 -4.43042874e-01
-7.72523582e-01 3.70311700e-02 -5.81498742e-01 3.03357810e-01
-2.88352489e-01 -2.20284209e-01 -9.41415131e-01 -2.52072871e-01
-1.68391511e-01 -8.88355732e-01 8.26430500e-01 3.96883279e-01
5.57551920e-01 2.76735518e-02 -2.98753291e-01 5.07906735e-01
1.05509281e+00 -2.51855522e-01 5.37505984e-01 3.36637735e-01
6.96329236e-01 6.96224391e-01 4.34227854e-01 -3.43391716e-01
7.61652827e-01 4.78822261e-01 3.63643438e-01 -4.50868309e-01
-3.90431821e-01 -2.68020570e-01 5.60781837e-01 7.31711268e-01
-4.70585600e-02 -3.02540183e-01 -1.11877155e+00 7.47028828e-01
-2.02944231e+00 -7.93940008e-01 -1.01741195e-01 1.78937685e+00
8.17186713e-01 -4.00501609e-01 -8.97098333e-02 -8.58889937e-01
6.87754869e-01 5.54522797e-02 -3.59599292e-01 -5.48578441e-01
-2.79733539e-01 -1.56251535e-01 1.89838707e-01 5.38021922e-01
-9.72501755e-01 1.31518996e+00 6.93851995e+00 4.42598999e-01
-1.13861084e+00 4.57460672e-01 9.06971991e-01 2.77050398e-02
-2.90445983e-01 3.69323462e-01 -5.73212147e-01 -1.41063169e-01
1.21504259e+00 1.39449835e-01 4.64040011e-01 3.44733864e-01
3.45068753e-01 -7.42973834e-02 -1.45440531e+00 1.28421330e+00
5.67117453e-01 -1.35843146e+00 1.89241216e-01 5.23514524e-02
8.67453098e-01 6.39843822e-01 2.96657503e-01 4.88404840e-01
1.83936089e-01 -1.17424917e+00 5.47616720e-01 4.30746704e-01
8.37980270e-01 -3.29473913e-02 4.49229121e-01 2.34487444e-01
-5.30249715e-01 1.74273849e-01 -2.83069819e-01 8.62051360e-03
4.77623314e-01 4.84669618e-02 -9.12033319e-01 4.10577863e-01
3.00997853e-01 6.01108193e-01 -6.22427046e-01 8.34686399e-01
-5.93205467e-02 5.78214526e-01 -9.07868445e-02 2.40826800e-01
6.44364119e-01 -2.35194579e-01 3.64401579e-01 1.39115787e+00
2.37460166e-01 -2.65807837e-01 1.81242332e-01 6.79214120e-01
-3.46201628e-01 1.90367997e-01 -6.66468263e-01 -4.70383793e-01
-2.39432216e-01 1.31926119e+00 -3.31192315e-01 -4.38971996e-01
-9.83044207e-01 1.15906107e+00 2.83662766e-01 8.47002149e-01
-8.80197465e-01 3.18925083e-01 4.84617889e-01 1.71193685e-02
5.91723286e-02 -4.86196995e-01 -3.50317478e-01 -1.48134422e+00
-2.62093186e-01 -1.11787307e+00 2.66399652e-01 -1.34471369e+00
-1.31118441e+00 5.37340641e-01 -8.62322301e-02 -9.31937754e-01
-5.04394889e-01 -7.39176691e-01 -3.39002550e-01 9.71123159e-01
-1.61860824e+00 -1.83957231e+00 4.53588478e-02 6.66089594e-01
9.89370763e-01 -2.34200507e-01 8.68902326e-01 1.21905953e-01
-1.80419087e-01 4.14327532e-01 1.65119335e-01 2.07677409e-01
1.27803004e+00 -7.62344897e-01 5.00937819e-01 6.04157507e-01
5.20845354e-01 3.19410563e-01 6.11102939e-01 -4.58935142e-01
-1.84846425e+00 -8.46884191e-01 1.15773499e+00 -1.07835710e+00
8.39228630e-01 -4.46532190e-01 -4.06201571e-01 1.01613009e+00
1.01117027e+00 -2.17395082e-01 5.33561647e-01 -8.62305984e-02
-4.37986463e-01 2.09325984e-01 -6.77838624e-01 7.35360324e-01
7.71365643e-01 -9.98778105e-01 -6.01743102e-01 6.57759547e-01
6.43205643e-01 -2.34184206e-01 -5.58515310e-01 2.39969790e-01
4.46826994e-01 -5.66502154e-01 1.09421945e+00 -1.03036678e+00
8.61403167e-01 7.17384508e-03 -3.55409235e-01 -1.12704146e+00
-1.10185094e-01 -9.29857850e-01 1.41056493e-01 1.11495173e+00
8.13908577e-01 -1.69765651e-01 4.84470338e-01 6.57280326e-01
-6.15151152e-02 -4.90145594e-01 -7.50854492e-01 -3.15657705e-01
3.28581303e-01 -3.16289037e-01 -8.10931996e-02 8.86716187e-01
1.43446606e-02 9.74647462e-01 -4.22467679e-01 -2.23592043e-01
3.66446733e-01 2.79051453e-01 9.92907465e-01 -6.55741036e-01
-2.25873366e-01 -4.93646979e-01 2.08209455e-03 -1.15999401e+00
1.11596204e-01 -1.18955195e+00 9.82281417e-02 -1.95940578e+00
6.67687833e-01 2.72583872e-01 5.79569563e-02 5.23792207e-01
-1.31211832e-01 5.40296733e-01 3.89381230e-01 5.06065130e-01
-9.12236392e-01 4.29046780e-01 1.62246048e+00 -2.82444626e-01
1.25713006e-01 -3.77352327e-01 -7.61285901e-01 4.33063328e-01
6.88946068e-01 -1.31383434e-01 -2.91857004e-01 -1.36573756e+00
4.03881341e-01 1.50404945e-01 5.64971268e-01 -3.65383327e-01
5.02615198e-02 -2.72008657e-01 3.83403867e-01 -3.25153917e-01
4.52366889e-01 -6.15196049e-01 -2.44932219e-01 4.70407605e-02
-7.25067675e-01 4.97013927e-01 3.01559448e-01 3.59128535e-01
-4.16176796e-01 3.22521329e-02 6.46233201e-01 -3.41363966e-01
-6.20104313e-01 1.91995770e-01 -4.05047178e-01 2.20296726e-01
8.02437961e-01 1.09235592e-01 -5.48214495e-01 -7.48120308e-01
-8.61066818e-01 4.18798923e-01 4.64539826e-01 8.21547091e-01
5.23795605e-01 -1.28312635e+00 -1.08933711e+00 -1.92321941e-01
2.89658666e-01 -4.30510223e-01 -7.17120767e-02 1.24126172e+00
-2.52413273e-01 9.16136265e-01 -1.44299433e-01 -9.12980318e-01
-1.20725405e+00 6.73976541e-01 1.48606896e-01 -2.20165893e-01
-1.66185990e-01 7.88262010e-01 5.07839501e-01 -5.38030982e-01
-8.48660246e-03 -4.85877618e-02 3.22708599e-02 -7.93473795e-02
3.78029674e-01 -1.05089396e-01 -2.50346065e-01 -8.37698400e-01
-2.50256866e-01 6.42117858e-01 3.54772955e-02 -8.20235431e-01
1.06165576e+00 -5.73465407e-01 -3.40362489e-01 3.48226368e-01
1.31827223e+00 -5.38232923e-01 -1.17701280e+00 -2.20369190e-01
2.58849915e-02 -9.22741294e-02 -1.09016545e-01 -1.10456562e+00
-7.02841997e-01 1.50869298e+00 6.21433973e-01 -1.70924634e-01
9.67605650e-01 4.22356427e-01 7.64547527e-01 5.22016227e-01
4.50685583e-02 -9.96719301e-01 2.56466925e-01 8.76908600e-01
1.16122103e+00 -1.68570650e+00 -3.93375695e-01 1.03600629e-01
-9.70200539e-01 1.07274699e+00 4.69179064e-01 2.17237830e-01
1.04819536e-01 -1.28958911e-01 3.74029577e-01 -7.63426498e-02
-9.90829527e-01 -3.84072334e-01 5.35123706e-01 6.41572058e-01
6.85218632e-01 -3.77283618e-02 -1.23414174e-02 1.66325986e-01
-3.04492880e-02 1.77059636e-01 8.54400098e-02 7.48149693e-01
-3.83739382e-01 -1.26836491e+00 -4.82270747e-01 4.92589809e-02
-3.72254074e-01 -5.40586710e-01 -7.65080214e-01 5.43517053e-01
-1.51580110e-01 9.78098452e-01 -1.57239866e-02 2.80446280e-02
1.01358317e-01 4.80119228e-01 9.20029461e-01 -6.29730880e-01
-4.82938051e-01 3.16620648e-01 2.80975074e-01 -4.35719162e-01
-5.54112077e-01 -6.09921396e-01 -8.81802797e-01 -2.44121067e-03
4.12194468e-02 -3.29187177e-02 9.76793766e-01 1.03475237e+00
5.84837735e-01 3.50249074e-02 5.51258214e-02 -1.02087665e+00
-3.67689073e-01 -9.91934597e-01 4.43006366e-01 4.00190264e-01
4.30116743e-01 6.72295094e-02 -5.61641715e-02 7.39353478e-01] | [11.310126304626465, 1.4156560897827148] |
5815b073-4ffa-4460-8bb2-1a8fdde856c3 | word-sense-extension | 2306.05609 | null | https://arxiv.org/abs/2306.05609v1 | https://arxiv.org/pdf/2306.05609v1.pdf | Word sense extension | Humans often make creative use of words to express novel senses. A long-standing effort in natural language processing has been focusing on word sense disambiguation (WSD), but little has been explored about how the sense inventory of a word may be extended toward novel meanings. We present a paradigm of word sense extension (WSE) that enables words to spawn new senses toward novel context. We develop a framework that simulates novel word sense extension by first partitioning a polysemous word type into two pseudo-tokens that mark its different senses, and then inferring whether the meaning of a pseudo-token can be extended to convey the sense denoted by the token partitioned from the same word type. Our framework combines cognitive models of chaining with a learning scheme that transforms a language model embedding space to support various types of word sense extension. We evaluate our framework against several competitive baselines and show that it is superior in predicting plausible novel senses for over 7,500 English words. Furthermore, we show that our WSE framework improves performance over a range of transformer-based WSD models in predicting rare word senses with few or zero mentions in the training data. | ['Yang Xu', 'Lei Yu'] | 2023-06-09 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 5.76648176e-01 8.82119983e-02 -2.32116371e-01 -4.67511117e-01
-4.47356790e-01 -9.88756955e-01 7.90963352e-01 5.37379801e-01
-8.69932294e-01 8.22459519e-01 7.60552764e-01 -4.85446066e-01
2.04407841e-01 -9.44273055e-01 -3.28129560e-01 -2.92934686e-01
2.27053776e-01 2.54097998e-01 2.89811641e-01 -7.44681895e-01
3.03420126e-01 1.90368757e-01 -1.36174989e+00 3.75114977e-01
8.04056406e-01 5.96990049e-01 6.65152252e-01 2.40687340e-01
-8.40782344e-01 6.08341619e-02 -7.59374738e-01 -4.43401992e-01
4.52202447e-02 -1.21568650e-01 -1.08902788e+00 -2.65632451e-01
2.03488022e-01 4.67796355e-01 1.28918931e-01 1.02932227e+00
4.63572621e-01 4.78604943e-01 4.96196806e-01 -8.81091058e-01
-1.30494916e+00 1.16376233e+00 -1.51366919e-01 5.18881023e-01
6.80662692e-01 5.88008799e-02 1.40503991e+00 -1.13421404e+00
8.60909939e-01 1.48701322e+00 5.38654387e-01 6.77966177e-01
-1.51422012e+00 -8.02890122e-01 4.43181783e-01 6.29938990e-02
-1.33220208e+00 -1.92589268e-01 6.99215591e-01 -2.13255495e-01
1.59712672e+00 4.69667790e-03 9.72335935e-01 1.48968124e+00
3.17147613e-01 6.96143985e-01 1.31379974e+00 -5.86894333e-01
4.90495145e-01 7.41333282e-03 5.26320219e-01 2.07061306e-01
5.10599136e-01 -1.79112405e-01 -7.57238448e-01 -2.83318281e-01
2.37662643e-01 -1.69231221e-01 8.31647590e-02 1.34606063e-01
-1.37145865e+00 7.65730500e-01 3.73837680e-01 7.50191450e-01
-3.85144591e-01 5.95328398e-02 3.73295516e-01 3.36004525e-01
5.29199958e-01 1.36383510e+00 -7.72703052e-01 5.62455319e-03
-6.55010343e-01 4.93192941e-01 5.17566442e-01 9.34996843e-01
9.08989191e-01 1.73929557e-02 -2.14589715e-01 1.02480447e+00
4.29197513e-02 4.88806546e-01 1.13892090e+00 -5.51750660e-01
-9.81033742e-02 6.52763247e-01 6.71435297e-02 -8.01893890e-01
-2.05744714e-01 -5.35594583e-01 -3.84275258e-01 -2.26503670e-01
-2.72781044e-01 -3.82402018e-02 -7.46591508e-01 2.24896598e+00
8.98238048e-02 3.77924234e-01 3.69147509e-01 3.18752497e-01
6.10921025e-01 6.18194997e-01 4.82678384e-01 -1.54000416e-01
1.48091543e+00 -3.00923765e-01 -5.72698653e-01 -8.95567536e-01
5.73567450e-01 -6.78914189e-01 1.67171252e+00 1.60906971e-01
-9.07073796e-01 -5.00677049e-01 -1.08676016e+00 -1.80091366e-01
-9.37179565e-01 -6.57380641e-01 6.74823523e-01 4.31931257e-01
-1.09632051e+00 3.46953452e-01 -3.57364655e-01 -6.65927470e-01
3.40253748e-02 -1.23747671e-02 -2.25446761e-01 9.11040455e-02
-1.93324876e+00 1.11170912e+00 1.01531637e+00 -5.95948994e-01
-2.62274772e-01 -9.14511085e-01 -1.13041520e+00 6.03035614e-02
4.34524238e-01 -1.11081076e+00 1.07646167e+00 -7.37768233e-01
-7.55331039e-01 1.02604342e+00 -4.64639097e-01 -4.90348488e-01
-4.78813499e-01 -1.40803903e-01 -1.07291424e+00 -3.32972884e-01
6.61003888e-01 8.15549493e-01 5.73160231e-01 -1.23916280e+00
-6.30785763e-01 -1.50052398e-01 3.10636789e-01 1.74113452e-01
-5.16780972e-01 -1.18916705e-01 3.07296932e-01 -1.30082798e+00
-7.41663426e-02 -8.21078360e-01 -2.27601111e-01 -2.87749112e-01
-4.33601975e-01 -6.94692671e-01 3.45557183e-01 1.68789551e-01
1.51791096e+00 -2.11666107e+00 -1.51156738e-01 1.17721245e-01
2.62758344e-01 1.70759335e-01 -3.68004978e-01 4.67082202e-01
-3.34321111e-01 5.09780765e-01 -3.74479026e-01 -1.29407227e-01
3.48900825e-01 6.96553826e-01 -9.18673813e-01 -4.12991017e-01
2.56290287e-01 1.10344172e+00 -1.45518136e+00 -1.58141479e-01
-6.51433542e-02 1.66579306e-01 -6.05354548e-01 -1.39342964e-01
-4.48527366e-01 -2.86746413e-01 -3.51602823e-01 2.44359046e-01
3.96678686e-01 -1.70274869e-01 3.31178635e-01 -4.69351932e-02
1.14386149e-01 9.65764225e-01 -1.05520344e+00 1.83922231e+00
-7.78501630e-01 3.16635251e-01 -6.12378061e-01 -5.05927205e-01
9.63155746e-01 3.32010314e-02 -3.36802676e-02 -3.20335925e-01
-2.81361490e-01 2.04501420e-01 -1.02106407e-01 -3.85632634e-01
1.11911380e+00 -9.43170607e-01 -5.77067614e-01 6.25298560e-01
8.31905454e-02 -4.73565459e-01 1.63629845e-01 2.26854160e-01
1.13214087e+00 -2.24949807e-01 7.65198529e-01 -5.85496843e-01
2.30622590e-01 -4.05126922e-02 5.75417161e-01 8.71753693e-01
-1.12385508e-02 -1.35931700e-01 -1.29328854e-02 -3.17353368e-01
-5.53243160e-01 -1.78101707e+00 -1.36574179e-01 1.63735867e+00
2.38520727e-01 -8.04529965e-01 -2.40040511e-01 -5.09889245e-01
1.39006555e-01 1.50860429e+00 -5.39757311e-01 -4.74941373e-01
-2.66372979e-01 -4.31620270e-01 8.36377978e-01 6.68947875e-01
2.12686345e-01 -1.38914776e+00 -5.23092985e-01 4.87778902e-01
-2.05887631e-01 -9.44730282e-01 -6.45637393e-01 3.83609772e-01
-4.73503709e-01 -4.95884925e-01 -3.09274495e-01 -8.83283675e-01
1.41761422e-01 2.77469665e-01 1.51427770e+00 -2.52535254e-01
-5.51731363e-02 4.19026464e-01 -6.09336317e-01 -5.84382951e-01
-3.65158916e-01 2.88451612e-01 6.78397954e-01 -4.66876388e-01
9.56475735e-01 -6.68131590e-01 -1.71928838e-01 -2.00685695e-01
-1.48279309e+00 -3.33012998e-01 4.19749498e-01 7.43738770e-01
4.60042357e-01 -2.10375190e-01 8.31970930e-01 -9.48211908e-01
1.22158217e+00 -7.95832038e-01 1.60149798e-01 2.45630234e-01
-8.22362483e-01 6.10129476e-01 3.83522689e-01 -6.65356457e-01
-1.14718485e+00 -3.26795131e-01 -3.01099867e-01 7.95344412e-02
-2.21552998e-01 8.01742673e-01 -1.33609429e-01 6.30548418e-01
9.22020614e-01 4.47590470e-01 -5.03319263e-01 -2.74276167e-01
1.07228231e+00 4.63973731e-01 6.15742922e-01 -9.60910201e-01
7.04791188e-01 1.66001648e-01 -3.85128409e-01 -9.92420971e-01
-1.22512901e+00 -5.08265316e-01 -5.34200609e-01 3.11182112e-01
9.30135727e-01 -6.02159023e-01 -4.39412929e-02 7.44683743e-02
-1.23824573e+00 1.18749768e-01 -6.50928557e-01 1.95924938e-01
-2.40534753e-01 4.50370789e-01 -1.58946857e-01 -5.03611803e-01
-3.08692545e-01 -7.13486969e-01 1.05723786e+00 6.12215996e-02
-1.01142085e+00 -1.29499638e+00 1.48229077e-01 -2.35973388e-01
5.00653505e-01 2.93889120e-02 1.46469295e+00 -1.05972886e+00
6.09541833e-02 3.15297216e-01 2.75001228e-01 8.27639326e-02
4.80014682e-01 -5.91730475e-01 -8.52885365e-01 -1.73891798e-01
-2.82690763e-01 -4.30397570e-01 1.22918475e+00 -6.05337042e-03
9.43523169e-01 -4.61397678e-01 -5.75369596e-01 3.25109214e-01
1.22446918e+00 1.58988163e-01 3.20887387e-01 3.88156354e-01
2.13166997e-01 2.45724678e-01 4.10427421e-01 3.39767843e-01
4.32753980e-01 2.37337157e-01 -1.37132540e-01 1.79816246e-01
8.15983266e-02 -4.35049504e-01 4.00256902e-01 6.67093277e-01
4.67626512e-01 -4.31010813e-01 -9.08306777e-01 9.32262838e-01
-1.14748335e+00 -1.05750680e+00 3.81498277e-01 1.80792344e+00
1.42926073e+00 5.26813745e-01 -1.38938218e-01 -6.35174941e-03
7.16855168e-01 5.14373302e-01 -6.40591204e-01 -7.02815771e-01
-4.22940195e-01 8.96571279e-01 1.80341136e-02 7.18746185e-01
-5.82674861e-01 1.65395951e+00 6.65552711e+00 7.43415236e-01
-8.43969524e-01 2.66226120e-02 1.29668238e-02 6.56601265e-02
-1.12283015e+00 -4.51118462e-02 -6.34048283e-01 1.80107504e-01
4.10653681e-01 -6.89987242e-01 3.12521756e-01 6.48408055e-01
-2.22741604e-01 2.06027940e-01 -1.11517489e+00 8.27894986e-01
1.19583130e-01 -1.23067987e+00 8.11880648e-01 -3.18781823e-01
5.77406943e-01 2.06383853e-03 1.92629650e-01 4.24026310e-01
6.68074310e-01 -8.93717945e-01 6.54928923e-01 5.44203937e-01
9.43266988e-01 -6.83244824e-01 2.81287819e-01 8.48126113e-02
-1.31553757e+00 6.23079622e-03 -6.81201875e-01 -4.57802147e-01
9.71396044e-02 6.85412169e-01 -1.12614322e+00 1.42250344e-01
3.94730270e-01 7.17061639e-01 -6.17322266e-01 2.73896098e-01
-5.71057022e-01 6.85148001e-01 -1.40206814e-01 -3.41157913e-01
2.76453674e-01 3.14295709e-01 9.45608914e-01 1.54301083e+00
4.17847127e-01 3.55677187e-01 3.63060534e-01 1.17766070e+00
-1.49965256e-01 1.35493372e-02 -7.83669591e-01 -3.90287727e-01
1.11669075e+00 8.74073565e-01 -5.71879983e-01 -7.23118365e-01
-1.74244747e-01 1.14239109e+00 2.01256737e-01 3.95798266e-01
-2.07287028e-01 -5.76173246e-01 1.38486743e+00 3.98124382e-03
7.92550743e-02 -3.93364072e-01 -3.20725709e-01 -1.09790850e+00
-4.95472848e-02 -3.24756026e-01 5.36926687e-01 -1.16062522e+00
-1.92170167e+00 6.64390981e-01 1.78012580e-01 -6.07049763e-01
-4.29858148e-01 -7.95554698e-01 -6.23270571e-01 9.91326571e-01
-1.31647670e+00 -1.01943135e+00 2.84547865e-01 3.45668107e-01
7.89864719e-01 -2.59155899e-01 1.20044124e+00 -4.69700396e-01
2.15059146e-01 7.89431989e-01 -3.71261328e-01 -5.73193058e-02
8.10309052e-01 -1.52969253e+00 9.46822882e-01 8.29549313e-01
3.75373840e-01 1.31608367e+00 8.96928072e-01 -8.99302125e-01
-8.52308631e-01 -1.22005546e+00 1.54563808e+00 -5.97107828e-01
9.14213181e-01 -4.69481289e-01 -1.01972377e+00 8.75468791e-01
4.08629060e-01 -2.37505287e-01 1.19933414e+00 4.03503329e-01
-8.93511832e-01 1.93050906e-01 -9.40864086e-01 8.94583106e-01
1.50681460e+00 -7.77562737e-01 -1.85959208e+00 -1.09654255e-01
1.37710226e+00 2.46585786e-01 -5.13721228e-01 1.68951191e-02
4.51193631e-01 -4.85410154e-01 1.20445168e+00 -1.15221691e+00
3.71710896e-01 -3.05028796e-01 -5.03715396e-01 -2.02744627e+00
-7.40267038e-01 -5.16855001e-01 2.77491212e-01 1.17658305e+00
6.58050299e-01 -8.44984651e-01 2.38433089e-02 3.49643528e-01
-3.17507446e-01 -6.44545853e-01 -8.04270923e-01 -9.42797780e-01
5.50268650e-01 -9.22788978e-01 1.15263867e+00 1.31951547e+00
5.03522098e-01 8.74792218e-01 3.38541299e-01 -6.28525689e-02
5.93405180e-02 4.69991267e-02 -1.62668712e-02 -1.19016039e+00
-2.42222846e-01 -6.41312540e-01 -4.76795971e-01 -1.18795443e+00
7.97221243e-01 -1.60217202e+00 -4.91365977e-02 -1.54659438e+00
1.11459628e-01 -2.32316121e-01 -8.31147671e-01 8.92453551e-01
-6.43784225e-01 7.95194209e-02 1.72128052e-01 -3.31987262e-01
-4.57316965e-01 4.83972013e-01 9.43266630e-01 -3.62797499e-01
-2.05276653e-01 -5.32559514e-01 -1.27341700e+00 8.67844462e-01
6.44075572e-01 -4.65254277e-01 -6.25493288e-01 -4.57875669e-01
9.71907794e-01 -5.66350102e-01 2.06906348e-01 -6.34565294e-01
-3.39463279e-02 -5.33101737e-01 3.58328253e-01 -2.24227846e-01
1.88059166e-01 -5.21066010e-01 -3.39571714e-01 3.00296724e-01
-8.72900069e-01 6.02068484e-01 3.27064633e-01 4.89673257e-01
1.43620461e-01 -7.88757280e-02 5.65072238e-01 -3.55958909e-01
-1.34550726e+00 -2.02812955e-01 -7.58733571e-01 6.49080217e-01
7.06010103e-01 -2.90878236e-01 -4.07328159e-02 -8.75059143e-02
-8.77669096e-01 1.21798627e-02 4.95948344e-01 1.09107554e+00
8.38986874e-01 -1.59283030e+00 -6.67089641e-01 1.57831520e-01
7.82791376e-01 -4.64566469e-01 -2.45909885e-01 -2.26824880e-01
4.92410719e-01 4.84132767e-02 6.36106506e-02 -2.09002763e-01
-9.13435102e-01 8.49981785e-01 7.99973980e-02 1.81383807e-02
-4.74329114e-01 1.16254187e+00 2.48365894e-01 -5.72111785e-01
-4.74421412e-01 -8.91820192e-01 -1.03903241e-01 1.21166945e-01
6.17669523e-01 1.83854476e-01 -1.62425011e-01 -5.94743311e-01
-5.93822122e-01 2.91191161e-01 -1.01855822e-01 -3.72945428e-01
1.10201883e+00 -3.55145425e-01 -2.22175494e-01 1.03667223e+00
9.62543547e-01 1.75864324e-01 -3.74679953e-01 -6.11814916e-01
4.19123292e-01 -2.44508430e-01 -2.73550063e-01 -1.28038049e+00
-1.38037741e-01 3.51197928e-01 2.76964068e-01 -1.03213200e-02
1.06279862e+00 5.02550006e-01 1.27677631e+00 7.70756423e-01
5.59420884e-01 -9.96857703e-01 1.86081573e-01 1.21007192e+00
8.49861026e-01 -7.52536893e-01 -5.49225867e-01 -2.02604547e-01
-7.19516754e-01 8.35860431e-01 3.34548682e-01 -7.38982186e-02
6.25870347e-01 3.33692729e-01 1.26020849e-01 -1.91629067e-01
-7.88614869e-01 -4.14630771e-01 4.01256979e-01 7.46626914e-01
6.34962797e-01 3.70590866e-01 -4.56441313e-01 1.05314136e+00
-9.40042138e-01 -1.74038440e-01 5.18745005e-01 7.81132519e-01
-7.70867884e-01 -1.19968629e+00 -1.09499902e-01 5.47695160e-01
-3.25119287e-01 -8.09847236e-01 -7.23124027e-01 4.54733491e-01
5.74756026e-01 1.04706502e+00 1.00036636e-01 -5.39056361e-01
1.58963725e-01 7.08540499e-01 3.80350113e-01 -1.29355514e+00
-3.70258898e-01 -3.77030909e-01 9.46016517e-03 -3.51697147e-01
-2.82788068e-01 -4.56739873e-01 -1.71017051e+00 3.90970036e-02
5.73486537e-02 2.81545132e-01 5.55013828e-02 1.35650146e+00
3.65450889e-01 4.19545561e-01 2.07726464e-01 -1.65317178e-01
-5.62287867e-01 -9.07944500e-01 -8.12781751e-01 7.23698735e-01
2.97697246e-01 -8.25545609e-01 -7.93908536e-02 8.15756023e-02] | [10.398555755615234, 8.89050579071045] |
69ae2a3f-1799-40b6-b429-f3975b40254d | portable-speech-to-speech-translation-on-an | null | null | https://aclanthology.org/W18-1923 | https://aclanthology.org/W18-1923.pdf | Portable Speech-to-Speech Translation on an Android Smartphone: The MFLTS System | null | ['Sean Colbath', 'Ralf Meermeier', 'Martha Lillie'] | 2018-03-01 | null | null | null | ws-2018-3 | ['speech-to-speech-translation'] | ['speech'] | [-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] | [-7.3298139572143555, 3.642245054244995] |
44dd0a27-940b-47e9-9333-312787ce51d5 | large-margin-representation-learning-for | 2206.08537 | null | https://arxiv.org/abs/2206.08537v1 | https://arxiv.org/pdf/2206.08537v1.pdf | Large-Margin Representation Learning for Texture Classification | This paper presents a novel approach combining convolutional layers (CLs) and large-margin metric learning for training supervised models on small datasets for texture classification. The core of such an approach is a loss function that computes the distances between instances of interest and support vectors. The objective is to update the weights of CLs iteratively to learn a representation with a large margin between classes. Each iteration results in a large-margin discriminant model represented by support vectors based on such a representation. The advantage of the proposed approach w.r.t. convolutional neural networks (CNNs) is two-fold. First, it allows representation learning with a small amount of data due to the reduced number of parameters compared to an equivalent CNN. Second, it has a low training cost since the backpropagation considers only support vectors. The experimental results on texture and histopathologic image datasets have shown that the proposed approach achieves competitive accuracy with lower computational cost and faster convergence when compared to equivalent CNNs. | ['Alessandro Lameiras Koerich', 'Alceu de Souza Britto Junior', 'Luiz Eduardo Soares de Oliveira', 'Jonathan de Matos'] | 2022-06-17 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 2.50629067e-01 2.01678157e-01 -2.94359297e-01 -6.44040525e-01
-6.44435763e-01 4.56771851e-02 2.11887777e-01 5.06343484e-01
-7.82796144e-01 7.68855453e-01 -4.37293500e-01 -2.60762841e-01
-3.75473291e-01 -8.34266245e-01 -5.15908360e-01 -7.84789622e-01
-2.93819159e-01 3.16425115e-01 3.67015421e-01 9.33321640e-02
2.35696763e-01 7.68423200e-01 -1.52129281e+00 4.79258060e-01
7.60868728e-01 1.64795768e+00 1.94920197e-01 3.61446351e-01
-1.46242723e-01 9.79319394e-01 -2.86200523e-01 -1.36270061e-01
2.13978052e-01 -5.49742244e-02 -9.13995981e-01 -3.52885090e-02
5.24573088e-01 -3.54061760e-02 -1.55327097e-01 8.23627710e-01
5.60240149e-01 6.70753345e-02 8.85150611e-01 -9.55241799e-01
-3.37177306e-01 2.97276229e-01 -4.87248361e-01 3.09607089e-01
-2.21874252e-01 -4.65785265e-01 7.67368972e-01 -8.78110111e-01
5.05453467e-01 9.56758916e-01 1.14777231e+00 4.66399103e-01
-1.29521894e+00 -6.14535272e-01 -3.44384223e-01 2.75528699e-01
-1.56243587e+00 -2.18387116e-02 6.99039221e-01 -3.91691744e-01
7.53795266e-01 3.55911106e-01 5.19246817e-01 5.00693858e-01
4.59497422e-01 7.32203066e-01 1.05215287e+00 -5.16726434e-01
3.10072273e-01 3.89556020e-01 3.63846928e-01 1.18857884e+00
4.05790269e-01 -1.20995417e-01 -9.41644832e-02 -1.35530204e-01
7.36782908e-01 5.29454201e-02 -1.53887402e-02 -4.05668676e-01
-8.60983014e-01 9.81088579e-01 7.89810419e-01 5.51690519e-01
-2.01994434e-01 1.46381736e-01 7.42487967e-01 3.48526716e-01
5.20037651e-01 2.37926543e-01 -1.71348467e-01 3.84613574e-01
-9.99783039e-01 -7.13075474e-02 5.74005544e-01 4.46132004e-01
8.48624229e-01 -9.28573012e-02 -9.83856991e-02 1.02497792e+00
1.45301342e-01 1.42962724e-01 8.60903382e-01 -3.26421946e-01
3.03552151e-01 8.81365120e-01 -4.44031656e-01 -1.26479590e+00
-7.27633774e-01 -5.64822614e-01 -1.19306505e+00 5.57315588e-01
4.81217802e-01 2.72271573e-03 -7.80713022e-01 1.55082190e+00
1.47546202e-01 4.37183052e-01 5.68965748e-02 6.34835005e-01
8.91615987e-01 3.36501867e-01 -1.48102522e-01 6.66052178e-02
1.22217584e+00 -7.79736698e-01 -5.73493659e-01 3.55235226e-02
9.31816518e-01 -5.59892476e-01 6.63341284e-01 1.75239518e-01
-5.98545074e-01 -5.94273567e-01 -1.32940269e+00 -9.49123781e-03
-6.35714591e-01 6.50932372e-01 7.46130526e-01 5.93186021e-01
-9.77435946e-01 8.48859906e-01 -8.83910716e-01 -2.86072552e-01
7.36973822e-01 7.65193999e-01 -6.09228253e-01 1.13093860e-01
-1.10481429e+00 8.82854223e-01 6.41424775e-01 2.38169506e-01
-4.70901608e-01 -5.82074702e-01 -1.10928762e+00 7.84874484e-02
-3.87783736e-01 -1.41805783e-01 9.01964009e-01 -1.32068503e+00
-1.35987365e+00 9.61003840e-01 1.19155198e-01 -6.20458066e-01
5.94058454e-01 2.12388024e-01 -2.75759876e-01 2.84894079e-01
-1.34694532e-01 5.29328942e-01 6.74353600e-01 -7.94690251e-01
-6.99440181e-01 -2.44283155e-01 -3.23611796e-01 -1.20125107e-01
-7.19070971e-01 -2.38113537e-01 -9.59413946e-02 -6.03223443e-01
2.75753468e-01 -8.36028874e-01 -2.43582815e-01 4.54130441e-01
-3.03826839e-01 -3.61364037e-01 9.11164999e-01 -3.36155891e-01
1.06904137e+00 -2.03708100e+00 -1.32012025e-01 5.58430076e-01
1.74352348e-01 4.94174510e-01 -1.45153344e-01 1.50967523e-01
-2.96206743e-01 -1.82197139e-01 -2.27011085e-01 -2.94295847e-01
-4.28506315e-01 1.90969259e-01 1.85147449e-01 7.23829091e-01
4.11893100e-01 6.25067472e-01 -6.72946393e-01 -6.45286143e-01
2.26074159e-01 6.36023998e-01 -3.58770370e-01 3.51307020e-02
2.30388135e-01 -2.15485748e-02 -2.38098413e-01 5.02912164e-01
8.21823955e-01 -3.80028784e-01 1.88479230e-01 -3.36463302e-01
3.93157825e-02 -9.04638022e-02 -1.38615191e+00 1.50482190e+00
-5.15663683e-01 8.46276522e-01 -2.17464447e-01 -1.46962607e+00
1.32155740e+00 1.44724190e-01 3.55663657e-01 -4.75534678e-01
3.51407886e-01 6.73134983e-01 -2.25904901e-02 -5.64828396e-01
-9.99037642e-03 -2.00276285e-01 2.09205583e-01 2.75259912e-01
4.17996854e-01 3.99720252e-01 7.10957721e-02 -2.10487172e-01
7.69695759e-01 -1.64166570e-01 4.82722968e-01 -5.31001270e-01
1.03106761e+00 -2.75443286e-01 7.13108301e-01 4.45724815e-01
4.36580135e-03 3.50141913e-01 5.44795275e-01 -1.02340174e+00
-8.00857604e-01 -6.86669588e-01 -6.59929037e-01 5.15074015e-01
1.15454026e-01 -7.21703768e-02 -6.07392430e-01 -8.04762006e-01
2.41301984e-01 6.36191107e-03 -1.12760878e+00 -1.91389620e-01
-6.60139799e-01 -7.40783513e-01 6.08289182e-01 7.64369547e-01
4.07026857e-01 -9.40158963e-01 -6.27445400e-01 -6.36870712e-02
2.27160990e-01 -7.67079175e-01 -6.29895404e-02 3.75198454e-01
-1.34386241e+00 -1.33005774e+00 -7.37840176e-01 -1.47384286e+00
1.15127766e+00 -1.06612980e-01 6.49242282e-01 9.87552777e-02
-7.71847367e-01 -1.91576019e-01 -4.39565070e-02 -2.78249085e-01
-2.32099086e-01 1.03941225e-01 -2.38391012e-02 4.07168359e-01
4.39138651e-01 -3.44538838e-01 -6.89617157e-01 1.91657126e-01
-9.04464781e-01 -9.05487761e-02 9.23161030e-01 1.21307397e+00
8.59586895e-01 -1.97893325e-02 7.67725527e-01 -1.09443510e+00
5.02501190e-01 -5.16513765e-01 -4.62895393e-01 3.65821123e-01
-8.46398950e-01 1.23452775e-01 7.27010071e-01 -4.40090865e-01
-7.74427116e-01 2.96937060e-02 1.90399900e-01 -1.79637775e-01
-9.38268900e-02 5.39325833e-01 4.41853672e-01 -5.48100471e-01
7.50336945e-01 2.77481228e-01 5.47624350e-01 -5.64690948e-01
-3.60810645e-02 8.49812984e-01 2.52401650e-01 -4.78156507e-01
3.71912241e-01 5.29867232e-01 3.63129914e-01 -7.26111948e-01
-6.60977244e-01 -5.18967688e-01 -8.46217930e-01 -1.46725714e-01
6.64161742e-01 -6.28497183e-01 -6.43991947e-01 6.15532994e-01
-7.76168287e-01 -9.86555666e-02 -2.85245657e-01 8.90217483e-01
-5.13635695e-01 4.28980887e-01 -6.88879550e-01 -4.71666962e-01
-7.47024119e-01 -1.05559433e+00 6.11554205e-01 2.94476599e-01
-4.59781066e-02 -1.29478347e+00 4.40099835e-03 3.57321948e-02
4.77256268e-01 4.89198804e-01 1.08125448e+00 -8.72565329e-01
-6.17978610e-02 -8.59540582e-01 -4.45436776e-01 7.85952032e-01
3.17173153e-01 -1.46208301e-01 -9.83984232e-01 -6.19499743e-01
-1.01432011e-01 -6.65674865e-01 9.09855306e-01 3.04216892e-01
1.52829397e+00 -2.60825396e-01 -3.56180251e-01 9.06296074e-01
1.83252716e+00 -5.14762849e-03 5.35971165e-01 5.62288046e-01
3.22869301e-01 4.61196810e-01 5.43689251e-01 2.76503533e-01
-8.36038664e-02 4.29678738e-01 3.22421849e-01 -5.86975873e-01
-2.31693089e-01 2.09510326e-01 -1.72540709e-01 8.00830007e-01
5.10623828e-02 3.50778818e-01 -9.00412858e-01 5.63050568e-01
-1.86324406e+00 -7.12047875e-01 -2.60456260e-02 2.28105998e+00
7.49201238e-01 1.64503008e-01 -2.23034516e-01 4.51216251e-01
6.60181999e-01 -2.43017957e-01 -4.65371937e-01 -5.36282778e-01
-1.17382314e-02 5.23906350e-01 4.22727019e-01 3.49053651e-01
-1.23399675e+00 4.37044889e-01 5.81052351e+00 1.06244111e+00
-1.55956471e+00 -1.03965469e-01 8.24772000e-01 -1.29356682e-02
3.37423354e-01 -5.27399302e-01 -6.45403802e-01 2.69351959e-01
7.43229330e-01 5.84912039e-02 -1.57588005e-01 8.75047863e-01
-6.60487637e-02 3.36146960e-03 -1.18209183e+00 1.02063656e+00
2.86250710e-01 -1.58556426e+00 2.12287039e-01 -1.76886141e-01
6.91798151e-01 1.78136490e-02 1.63635358e-01 1.49118211e-02
-2.52900839e-01 -1.15531480e+00 3.77645254e-01 3.82469833e-01
9.75418508e-01 -1.14713609e+00 1.33143377e+00 9.86258313e-02
-1.11421359e+00 -1.71822101e-01 -7.89190233e-01 8.05799570e-03
-6.20573223e-01 5.88548720e-01 -1.00500453e+00 4.24076140e-01
4.89497095e-01 6.73455834e-01 -6.05668664e-01 1.28286600e+00
3.76993828e-02 2.79958576e-01 -2.21696839e-01 -2.68315762e-01
5.06322324e-01 -1.27339616e-01 -1.27879903e-02 1.58981788e+00
2.19529688e-01 -2.65881509e-01 2.31385797e-01 5.75621486e-01
-2.50489209e-02 6.26492202e-01 -5.25960624e-01 4.30249989e-01
1.97942331e-01 1.50683546e+00 -8.77827346e-01 -1.17068827e-01
-3.96496177e-01 7.28272915e-01 7.25955188e-01 -4.69342805e-02
-4.68959481e-01 -8.26951385e-01 4.55441803e-01 1.80260807e-01
3.81683230e-01 2.29916573e-01 -3.48531783e-01 -7.24828660e-01
8.86155367e-02 -4.89222199e-01 6.69829786e-01 -1.65182576e-01
-1.31051886e+00 9.61081564e-01 -3.82238775e-01 -1.52250433e+00
1.02479095e-02 -1.02606964e+00 -6.04746759e-01 9.29644465e-01
-1.85874414e+00 -1.24937999e+00 -3.91767204e-01 7.28455901e-01
2.20495060e-01 -4.38448191e-01 1.16610563e+00 2.11287349e-01
-5.38540542e-01 1.00954139e+00 5.33513844e-01 5.77259541e-01
5.29054761e-01 -1.24267137e+00 -3.57365549e-01 4.18174475e-01
-2.19541267e-01 4.65524614e-01 4.98785498e-03 -1.48031756e-01
-8.45751464e-01 -1.17869186e+00 1.05184472e+00 1.14536889e-01
5.34727037e-01 -3.08991551e-01 -8.09906542e-01 3.55641514e-01
-1.49138033e-01 6.88385069e-01 1.30343819e+00 2.84484625e-02
-4.37808692e-01 -6.14022672e-01 -1.53602731e+00 4.92381155e-02
2.12717727e-01 -5.07757485e-01 -3.70990813e-01 3.50913644e-01
5.05749322e-02 -2.71276623e-01 -1.08777249e+00 4.68590915e-01
7.12394416e-01 -6.88611925e-01 7.72196949e-01 -6.68681741e-01
4.36785132e-01 -2.52814293e-01 -5.63758984e-02 -9.66558814e-01
-3.47155601e-01 -1.77474350e-01 1.13612212e-01 7.76636541e-01
5.40473163e-01 -6.95475578e-01 8.37588489e-01 3.01496893e-01
4.02816683e-02 -1.39015269e+00 -1.18458998e+00 -8.53777766e-01
1.23771384e-01 -1.56011805e-01 1.78663298e-01 1.12749696e+00
2.11935535e-01 -1.61175802e-02 -1.45349890e-01 -2.52632331e-02
8.13540101e-01 3.10997039e-01 1.83940560e-01 -1.56831551e+00
-6.62894128e-03 -2.57875264e-01 -1.19231880e+00 -3.79685789e-01
1.85629234e-01 -1.44035816e+00 -1.90227330e-01 -1.34370792e+00
2.28927404e-01 -1.00322545e+00 -1.03212166e+00 7.89127111e-01
1.18977748e-01 5.49338996e-01 -2.05788791e-01 1.96579844e-01
-3.99810642e-01 2.58810431e-01 1.00745142e+00 -2.06124768e-01
-1.90871879e-02 1.83066189e-01 -4.15295064e-01 7.98728228e-01
9.40262377e-01 -6.08690262e-01 -4.64554489e-01 -2.51739770e-01
-1.55273408e-01 -2.58922696e-01 1.84069067e-01 -1.25631797e+00
3.14882934e-01 2.72039175e-01 7.79473722e-01 -3.55816424e-01
-1.05137065e-01 -8.19601536e-01 -1.65976033e-01 1.03006387e+00
-6.07230246e-01 -2.37276524e-01 4.16377932e-01 5.28549016e-01
-5.02427578e-01 -5.39000213e-01 1.27654564e+00 9.98090878e-02
-7.00993955e-01 2.91689932e-01 -2.00024977e-01 -2.93301046e-01
1.24446690e+00 -5.33792257e-01 1.54648712e-02 2.62204826e-01
-8.83152366e-01 -6.85998872e-02 1.22744851e-02 3.82283837e-01
8.57948184e-01 -1.69023407e+00 -6.98013246e-01 4.19295162e-01
2.57435173e-01 -1.75483972e-02 3.76084782e-02 8.81970346e-01
-1.02838552e+00 5.74802935e-01 -5.72023273e-01 -6.88923657e-01
-1.46473658e+00 2.77361244e-01 6.94042683e-01 -4.23650354e-01
-7.04058230e-01 9.93337154e-01 -2.17963122e-02 -6.06219471e-01
3.99543017e-01 -3.82163852e-01 -4.82169092e-01 1.25230864e-01
7.61306703e-01 4.03819621e-01 5.05104601e-01 -4.59041685e-01
-4.39087868e-01 5.75768173e-01 -3.25591147e-01 4.69742984e-01
1.65109384e+00 4.70845550e-01 -4.01292771e-01 3.50300729e-01
1.70380557e+00 -5.18578410e-01 -1.00478494e+00 -5.12022376e-01
5.90653755e-02 -3.90831649e-01 3.23774785e-01 -6.45510793e-01
-1.31788516e+00 8.43202770e-01 1.21821570e+00 -3.02076161e-01
9.67723966e-01 -2.15792090e-01 5.43112934e-01 7.49723375e-01
1.99896663e-01 -1.30090487e+00 1.62525803e-01 2.99529999e-01
7.91679978e-01 -1.26687372e+00 -1.10142753e-01 -2.45202571e-01
-1.27344385e-01 1.79548836e+00 6.53438926e-01 -4.95893240e-01
9.67860937e-01 2.73595929e-01 1.46809295e-01 -2.17895985e-01
-5.39618611e-01 9.55063328e-02 4.37318772e-01 6.25563025e-01
5.76197684e-01 6.53556213e-02 -4.85883534e-01 3.64515424e-01
2.44696945e-01 3.15159559e-01 2.85342693e-01 8.65696847e-01
-4.80170995e-01 -9.63679552e-01 -5.14616519e-02 6.73524380e-01
-6.45967543e-01 5.49925864e-02 -4.13013957e-02 7.64915526e-01
2.51165390e-01 5.75952590e-01 3.36677521e-01 -4.08210993e-01
1.73979416e-01 -1.34569192e-02 4.78162080e-01 -5.12217999e-01
-5.55213630e-01 -2.95197845e-01 -2.19046757e-01 -6.79304540e-01
-3.50297540e-01 -4.32445586e-01 -1.33289480e+00 -3.81557718e-02
-5.10009944e-01 2.14342758e-01 7.10970759e-01 6.70204759e-01
2.37525463e-01 4.91309345e-01 8.22960317e-01 -5.00151515e-01
-7.69017279e-01 -1.01613462e+00 -8.43682408e-01 4.44577247e-01
3.45658749e-01 -6.11177862e-01 -1.95901588e-01 -1.85171932e-01] | [15.088761329650879, -2.7370591163635254] |
4c7f3b3b-54dc-414e-ab6e-9fcd428305d3 | a-proxy-free-strategy-for-practically | 2306.08313 | null | https://arxiv.org/abs/2306.08313v1 | https://arxiv.org/pdf/2306.08313v1.pdf | A Proxy-Free Strategy for Practically Improving the Poisoning Efficiency in Backdoor Attacks | Poisoning efficiency is a crucial factor in poisoning-based backdoor attacks. Attackers prefer to use as few poisoned samples as possible to achieve the same level of attack strength, in order to remain undetected. Efficient triggers have significantly improved poisoning efficiency, but there is still room for improvement. Recently, selecting efficient samples has shown promise, but it requires a proxy backdoor injection task to find an efficient poisoned sample set, which can lead to performance degradation if the proxy attack settings are different from the actual settings used by the victims. In this paper, we propose a novel Proxy-Free Strategy (PFS) that selects efficient poisoned samples based on individual similarity and set diversity, effectively addressing this issue. We evaluate the proposed strategy on several datasets, triggers, poisoning ratios, architectures, and training hyperparameters. Our experimental results demonstrate that PFS achieves higher backdoor attack strength while x500 faster than previous proxy-based selection approaches. | ['Bin Li', 'Wei zhang', 'Xue Rui', 'Beihao Xia', 'Pengfei Xia', 'Hong Sun', 'Ziqiang Li'] | 2023-06-14 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [-2.39368886e-01 -7.93836534e-01 -4.02370214e-01 2.08138958e-01
-9.02058542e-01 -1.01115048e+00 4.87258017e-01 4.43484485e-01
-8.96290720e-01 5.92923582e-01 1.56381175e-01 -2.91204840e-01
1.20654434e-01 -9.96582687e-01 -3.07970852e-01 -8.54107976e-01
-2.88120687e-01 4.26494181e-01 5.65173686e-01 -1.70083478e-01
4.99959737e-01 7.03506947e-01 -8.73606741e-01 4.16066647e-02
5.35362840e-01 4.47862089e-01 -1.70441523e-01 4.38575506e-01
1.07903168e-01 5.39048433e-01 -1.30390024e+00 -4.80687737e-01
7.04542756e-01 -3.79665375e-01 -3.94087762e-01 -6.03028953e-01
3.96358781e-02 -6.85487330e-01 -4.80624706e-01 1.02501655e+00
7.67986894e-01 -2.40342468e-01 4.29512948e-01 -1.30856276e+00
-1.43522754e-01 9.92266715e-01 -7.49998689e-01 5.12476146e-01
1.51308268e-01 6.42048001e-01 9.79042709e-01 -4.49050784e-01
1.09625898e-01 1.29794550e+00 2.82155246e-01 7.68832564e-01
-1.23890471e+00 -1.40167093e+00 -3.78299272e-04 5.13820052e-02
-1.55585885e+00 -3.19875926e-01 6.00379288e-01 5.07410727e-02
6.48008943e-01 5.35938323e-01 5.47256291e-01 1.40499556e+00
1.58167899e-01 5.80457151e-01 8.82152498e-01 9.48815793e-03
6.73393726e-01 2.08153665e-01 1.55408651e-01 2.63690650e-01
9.06587303e-01 4.60495412e-01 -6.02834642e-01 -1.09405780e+00
5.42276442e-01 1.87298447e-01 -5.40955842e-01 -1.49496868e-01
-8.56594920e-01 1.28693378e+00 4.77472007e-01 1.92542017e-01
-2.95105785e-01 4.26794738e-01 4.40743983e-01 2.27406770e-01
-1.83420870e-02 1.07702756e+00 -1.71379536e-01 2.40795147e-02
-8.10661077e-01 6.59478843e-01 8.84175181e-01 3.75788897e-01
4.20722872e-01 -4.79529146e-03 -5.34397781e-01 3.57510209e-01
-1.37984687e-02 5.52110076e-01 1.95385993e-01 -5.48044264e-01
7.01756179e-01 3.55397075e-01 2.24201664e-01 -8.54482293e-01
-1.29458141e-02 -5.80708444e-01 -5.15161395e-01 1.84684485e-01
4.89619792e-01 -1.53725192e-01 -5.93603909e-01 1.91921449e+00
5.17583549e-01 8.14260244e-02 -1.41897663e-01 8.01875830e-01
5.99608533e-02 6.73639476e-01 3.26768994e-01 -1.87690571e-01
1.28366995e+00 -4.13891524e-01 -4.05762374e-01 8.63650069e-02
4.63883191e-01 -3.34635645e-01 1.35504258e+00 5.96556008e-01
-6.65702105e-01 3.07691365e-01 -1.25688303e+00 8.45849276e-01
-3.03424239e-01 -4.85547811e-01 6.00420654e-01 1.30463493e+00
-3.70772749e-01 5.52842081e-01 -6.79300725e-01 -1.75686210e-01
6.29539371e-01 6.57709897e-01 7.68150389e-02 1.04038514e-01
-1.19429743e+00 4.57720339e-01 5.15708864e-01 -7.04484165e-01
-1.54747987e+00 -9.53591943e-01 -1.97782144e-01 2.17687607e-01
6.18547499e-01 -6.02465272e-01 9.71768737e-01 1.35858059e-01
-1.34611595e+00 7.12300837e-02 4.24457133e-01 -9.00666177e-01
6.09010637e-01 -3.04507047e-01 3.45327593e-02 2.66184330e-01
-1.42562419e-01 4.04794574e-01 9.37080383e-01 -1.25995839e+00
-5.69323778e-01 -4.38695788e-01 3.03936034e-01 1.43710017e-01
-1.19887531e+00 2.70485073e-01 -3.30834538e-01 -6.68131173e-01
-4.68060553e-01 -8.49997878e-01 -6.19943082e-01 -3.24087292e-01
-7.33915448e-01 -8.33776146e-02 9.72512364e-01 1.28880665e-02
1.48358798e+00 -1.95291269e+00 -1.59295484e-01 4.18564498e-01
4.41905171e-01 5.11170089e-01 -3.04408014e-01 7.42236018e-01
2.29122967e-01 5.06939113e-01 -1.49696335e-01 -5.94595931e-02
-1.61621720e-01 -3.46072614e-02 -8.38063180e-01 6.32171810e-01
-2.46042088e-01 3.52867782e-01 -9.87810552e-01 -3.21356624e-01
1.09844998e-01 2.62491345e-01 -7.89082766e-01 3.57326061e-01
-3.02646339e-01 2.65860558e-01 -7.65393198e-01 7.22219527e-01
6.79280996e-01 5.58295548e-02 -5.54271303e-02 -1.81331143e-01
9.49598774e-02 2.13196993e-01 -1.04728639e+00 7.63751268e-01
-2.86434770e-01 -1.29525736e-01 -1.84510574e-01 -2.98005193e-01
8.47814620e-01 1.54987752e-01 5.85970700e-01 -5.59289813e-01
2.22624958e-01 3.00761253e-01 9.67124198e-03 2.17370749e-01
2.99645334e-01 -1.55089227e-02 -2.74406403e-01 6.08791947e-01
-5.12277365e-01 -6.53814375e-02 2.29254380e-01 3.97643149e-01
1.59616089e+00 -6.43302321e-01 3.16378385e-01 -1.63660094e-01
2.81716079e-01 1.12860128e-01 5.89439988e-01 1.21046078e+00
-2.27577031e-01 1.55105352e-01 5.80370665e-01 -4.37631071e-01
-8.04387331e-01 -1.13080287e+00 3.24833579e-02 7.61587441e-01
4.58298177e-01 -8.91270876e-01 -8.50573540e-01 -9.74467814e-01
-8.21314566e-03 8.03803146e-01 -4.40888911e-01 -4.40094411e-01
-6.96396530e-01 -1.23868859e+00 1.23361397e+00 9.55901891e-02
6.16044700e-01 -7.93080032e-01 -1.02528667e+00 1.77033857e-01
1.76173508e-01 -7.21267939e-01 -5.98653734e-01 4.06492770e-01
-7.38056660e-01 -1.13853312e+00 -3.93731296e-01 1.44503921e-01
3.04114878e-01 4.78724509e-01 6.76078022e-01 1.24411635e-01
-2.22667411e-01 -8.77044126e-02 -3.98152530e-01 -4.03819114e-01
-2.38768309e-01 2.50934839e-01 3.72879595e-01 -3.07783484e-01
1.57471567e-01 -5.56677699e-01 -7.61685133e-01 4.01270360e-01
-1.08203018e+00 -8.47760797e-01 5.66486537e-01 7.46445358e-01
4.72385734e-01 4.59253997e-01 4.63557392e-01 -8.65274072e-01
1.08506572e+00 -5.62042296e-01 -7.94051051e-01 2.98520196e-02
-4.78215158e-01 1.22081317e-01 1.13494670e+00 -8.02459061e-01
-3.88078660e-01 -7.01654032e-02 -8.70847180e-02 -7.77992427e-01
2.53168404e-01 -1.59518138e-01 -4.37367082e-01 -2.42064536e-01
1.07765198e+00 8.79177004e-02 -4.28612798e-01 -4.57986742e-01
2.64423162e-01 3.63421172e-01 2.26664647e-01 -7.81707466e-01
1.18751144e+00 4.89202142e-01 -1.27640711e-02 -7.63695061e-01
-4.12509441e-01 -3.98317188e-01 2.25356326e-01 4.45929587e-01
3.02494287e-01 -7.13738143e-01 -8.14808309e-01 2.03486875e-01
-1.01016593e+00 -1.46357059e-01 -3.27097774e-01 2.72226989e-01
6.82082698e-02 5.69341421e-01 -5.14165342e-01 -9.32876587e-01
-7.88470805e-01 -1.35179722e+00 7.15942562e-01 8.59606937e-02
-3.26592505e-01 -4.67108697e-01 2.19799280e-01 -1.12580806e-01
4.35145974e-01 3.63305360e-01 8.76760900e-01 -1.04155087e+00
-7.65022874e-01 -4.23215419e-01 2.95445591e-01 -1.15443110e-01
1.88277230e-01 -1.87820837e-01 -6.87262654e-01 -6.97832286e-01
2.22596154e-01 -3.90438855e-01 8.73487592e-01 1.03782393e-01
1.30126333e+00 -7.90217459e-01 -5.53560555e-01 7.85683990e-01
1.38558340e+00 3.19374233e-01 5.68499684e-01 6.22738421e-01
4.90079552e-01 1.21852629e-01 6.66092336e-01 8.73901188e-01
-2.85293549e-01 6.74915791e-01 6.99093461e-01 6.36126399e-02
2.48156622e-01 -5.69044232e-01 5.46858907e-01 -5.02141453e-02
6.39790416e-01 -5.96535385e-01 -5.77294886e-01 3.90772790e-01
-1.22106969e+00 -1.12972128e+00 2.16582358e-01 2.75082207e+00
9.49337423e-01 3.40930521e-01 7.97795057e-01 4.10875350e-01
5.67132175e-01 1.62235022e-01 -7.41564155e-01 -2.35804439e-01
4.85983975e-02 3.95722002e-01 1.06370771e+00 3.18730325e-01
-9.89392638e-01 1.13853621e+00 6.52839804e+00 1.48440003e+00
-1.15737534e+00 1.37733474e-01 6.87460244e-01 -4.71037209e-01
-3.32212150e-01 3.34770799e-01 -1.12702441e+00 6.30471885e-01
9.06539321e-01 -2.37966418e-01 4.13379222e-01 7.67486095e-01
3.14706415e-02 3.64926383e-02 -9.60220814e-01 9.81001496e-01
-1.90228224e-01 -1.18007505e+00 3.53011549e-01 4.37879115e-01
1.76408753e-01 -1.14963979e-01 4.32439774e-01 8.17309320e-02
6.40730143e-01 -9.08132672e-01 3.97232980e-01 -4.22630638e-01
5.57785690e-01 -1.25767696e+00 3.57510030e-01 3.86119246e-01
-1.09700727e+00 -5.31559646e-01 -4.81219262e-01 4.59074318e-01
-1.12495027e-01 5.62976122e-01 -1.12020612e+00 3.83723192e-02
8.00616741e-01 -1.80270135e-01 -5.75812876e-01 1.26143885e+00
-2.58727610e-01 8.95823121e-01 -8.67218316e-01 -4.38152671e-01
2.70203382e-01 5.02489582e-02 9.27839398e-01 9.75650072e-01
7.77749298e-03 4.82121073e-02 3.08550417e-01 9.16809082e-01
-1.93324894e-01 8.79450366e-02 -6.93243623e-01 -1.60850152e-01
1.23368716e+00 1.15910840e+00 -7.15370178e-01 -1.48341790e-01
2.55417019e-01 7.69791603e-01 7.45086074e-02 1.46756917e-01
-1.21340966e+00 -4.84557748e-01 9.12528813e-01 1.34269387e-01
2.28040144e-02 -2.08594471e-01 -2.15898350e-01 -8.40413511e-01
-4.15094346e-01 -1.32456124e+00 7.51388967e-01 3.92906845e-01
-1.30025494e+00 5.71998775e-01 2.26050392e-01 -1.09650135e+00
1.24409482e-01 -2.09669396e-01 -9.00483787e-01 5.01168251e-01
-1.01572382e+00 -6.70261860e-01 1.02842212e-01 6.76351428e-01
4.38405246e-01 -1.78661183e-01 5.90804100e-01 2.51608640e-01
-7.85996437e-01 1.13867760e+00 -1.38595253e-01 -1.82332873e-01
6.01253092e-01 -9.65752244e-01 3.70154291e-01 8.63611281e-01
2.38849685e-01 9.22268212e-01 8.82523239e-01 -8.04063976e-01
-1.65236568e+00 -1.00821853e+00 6.56207651e-02 -5.22666276e-01
3.77062619e-01 -4.09134150e-01 -6.37611985e-01 1.14716426e-03
-5.11586703e-02 -3.35466951e-01 8.81458044e-01 -1.87555924e-01
-7.42729187e-01 -2.01339796e-01 -1.63428223e+00 1.05747437e+00
8.38856399e-01 -2.02241272e-01 4.71618585e-02 3.62744838e-01
9.49357688e-01 5.93351312e-02 -5.93055129e-01 2.75653481e-01
2.03254744e-01 -1.00219762e+00 1.33144426e+00 -2.14942276e-01
-3.61281782e-01 -4.35739875e-01 -1.82236329e-01 -1.04245174e+00
-2.12012798e-01 -9.50810611e-01 -3.63231242e-01 1.03416359e+00
1.50592044e-01 -7.45267510e-01 1.04114068e+00 2.91201204e-01
4.51629639e-01 -8.94503951e-01 -8.72806787e-01 -1.11033177e+00
1.41339347e-01 -3.34422896e-03 1.08618557e+00 5.81751168e-01
-1.10137440e-01 3.30717683e-01 -5.29429853e-01 4.05209959e-01
1.08184278e+00 -3.55284423e-01 1.05053270e+00 -8.38852763e-01
-7.02926457e-01 -4.50119376e-01 -6.13440052e-02 -7.11728990e-01
-2.23003909e-01 -4.63773131e-01 -3.96102220e-01 -7.63775766e-01
2.96306074e-01 -6.61486626e-01 -3.16765934e-01 6.90614283e-01
-3.28527719e-01 3.56255621e-01 2.21119136e-01 4.79004681e-01
-4.75082606e-01 6.26796246e-01 7.51577973e-01 -2.51934916e-01
-6.90832913e-01 -5.94889447e-02 -9.01247859e-01 2.96126753e-01
1.17785072e+00 -9.01815832e-01 -6.45491183e-01 1.22967465e-02
-1.38717949e-01 -1.06461108e-01 9.52725634e-02 -1.26714909e+00
1.69082150e-01 -4.25417304e-01 4.18878913e-01 -6.34406626e-01
4.50645208e-01 -7.53638029e-01 1.48981616e-01 1.26202023e+00
-2.42334709e-01 2.86787122e-01 2.09403098e-01 6.64631963e-01
4.06017631e-01 -1.60967186e-01 1.02676284e+00 -1.18556228e-02
4.78977039e-02 6.33695185e-01 -2.82791227e-01 -3.27890553e-02
1.19085634e+00 -9.75829214e-02 -5.03486037e-01 -1.69842839e-01
2.50133842e-01 2.09655911e-01 7.22269177e-01 6.10584840e-02
8.32543075e-01 -9.71791863e-01 -6.16185427e-01 -8.33609551e-02
1.13473803e-01 -4.17787641e-01 -2.89263636e-01 3.42475981e-01
-4.53706056e-01 1.67121083e-01 2.25426685e-02 -3.29801202e-01
-1.36350465e+00 9.61431861e-01 2.18113720e-01 -5.69196403e-01
-5.26777208e-01 8.57581854e-01 1.41865283e-01 2.67565195e-02
4.37557220e-01 1.13111608e-01 1.97306108e-02 -3.74184787e-01
8.66436899e-01 4.78830218e-01 -1.51658937e-01 -1.22316323e-01
-4.22598690e-01 1.89817801e-01 -4.01886523e-01 -2.59184927e-01
9.76941228e-01 5.23399949e-01 1.27436519e-01 -2.89550155e-01
9.80830550e-01 4.03901517e-01 -8.84909511e-01 1.06931515e-01
5.33923805e-02 -1.01820505e+00 -7.18312487e-02 -5.58354020e-01
-1.05942452e+00 6.04846060e-01 6.76073492e-01 7.44231194e-02
9.87148702e-01 -2.76132435e-01 1.17710602e+00 7.07327962e-01
7.54635274e-01 -6.38548076e-01 5.55009425e-01 -5.80750499e-03
4.43830043e-01 -5.69382429e-01 2.87563115e-01 -3.89107853e-01
-4.19040352e-01 5.67498207e-01 8.23719740e-01 -2.82704413e-01
1.86387897e-01 3.65584075e-01 -4.08622980e-01 -2.36916080e-01
-6.85381770e-01 8.92320871e-02 -5.16378343e-01 6.22432172e-01
-1.17278025e-01 3.68250422e-02 -4.14810866e-01 2.78333634e-01
-1.34049639e-01 -6.52555346e-01 3.67032856e-01 8.54865551e-01
-5.57783842e-01 -1.72142422e+00 -9.07337427e-01 4.73603606e-01
-7.57916212e-01 -1.11457035e-01 -6.97145045e-01 7.26969719e-01
-3.09762396e-02 1.13175154e+00 -2.86578953e-01 -7.09508598e-01
2.63600767e-01 -4.22495544e-01 4.63652104e-01 -5.06955266e-01
-1.14888394e+00 3.17073278e-02 -1.12852119e-01 -5.51083028e-01
5.13165951e-01 -1.16799243e-01 -1.14253664e+00 -5.40382326e-01
-6.85397685e-01 6.64723635e-01 4.10513848e-01 3.57677042e-01
2.19613180e-01 -1.83029044e-02 1.14162612e+00 -4.86982882e-01
-1.28155375e+00 -5.01239598e-01 -4.98936743e-01 2.19444260e-01
6.54430538e-02 -5.96022964e-01 -5.23967564e-01 -7.65661120e-01] | [5.782889366149902, 7.593871116638184] |
5c3897e2-aacd-4a57-9332-f3022d2e06c4 | recurrent-convolutional-neural-networks-for | 1306.3584 | null | http://arxiv.org/abs/1306.3584v1 | http://arxiv.org/pdf/1306.3584v1.pdf | Recurrent Convolutional Neural Networks for Discourse Compositionality | The compositionality of meaning extends beyond the single sentence. Just as
words combine to form the meaning of sentences, so do sentences combine to form
the meaning of paragraphs, dialogues and general discourse. We introduce both a
sentence model and a discourse model corresponding to the two levels of
compositionality. The sentence model adopts convolution as the central
operation for composing semantic vectors and is based on a novel hierarchical
convolutional neural network. The discourse model extends the sentence model
and is based on a recurrent neural network that is conditioned in a novel way
both on the current sentence and on the current speaker. The discourse model is
able to capture both the sequentiality of sentences and the interaction between
different speakers. Without feature engineering or pretraining and with simple
greedy decoding, the discourse model coupled to the sentence model obtains
state of the art performance on a dialogue act classification experiment. | ['Nal Kalchbrenner', 'Phil Blunsom'] | 2013-06-15 | recurrent-convolutional-neural-networks-for-2 | https://aclanthology.org/W13-3214 | https://aclanthology.org/W13-3214.pdf | ws-2013-8 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 4.95638490e-01 7.36757100e-01 7.63976648e-02 -5.96435130e-01
-9.12156925e-02 -4.78596568e-01 1.17258167e+00 1.89262986e-01
-3.55782449e-01 3.87838572e-01 9.00076270e-01 -4.44200516e-01
3.14137697e-01 -8.57415617e-01 -3.34265947e-01 -5.24068236e-01
1.62291989e-01 3.20589781e-01 1.77051857e-01 -8.10045362e-01
1.51451126e-01 -4.21800092e-03 -1.33770645e+00 8.20216477e-01
3.29896182e-01 7.33412325e-01 3.50660145e-01 1.05612969e+00
-8.51690650e-01 1.33197534e+00 -8.37256014e-01 -3.42043459e-01
-9.72671956e-02 -7.36234188e-01 -1.21018219e+00 3.03323150e-01
3.30681324e-01 -2.01753765e-01 -3.27597201e-01 6.26495302e-01
1.39349595e-01 8.76945332e-02 4.87964123e-01 -6.62177205e-01
-5.21220088e-01 1.13095081e+00 2.63436288e-01 1.67538673e-02
4.82353598e-01 2.57522762e-02 1.30355561e+00 -6.34330630e-01
5.67192852e-01 1.44459033e+00 6.22701883e-01 8.52293313e-01
-1.36439252e+00 5.72953038e-02 3.34136039e-01 -8.79044756e-02
-6.00637078e-01 -5.74968278e-01 6.11795247e-01 -5.25421321e-01
1.37804830e+00 2.56238431e-01 7.59513974e-01 1.28017509e+00
1.28455088e-01 7.36007094e-01 6.37779713e-01 -4.56888616e-01
2.91798323e-01 2.65208423e-01 5.40889621e-01 4.97131050e-01
-3.49309057e-01 -4.59188104e-01 -4.05606240e-01 -1.02324992e-01
4.01490510e-01 -8.26172605e-02 -1.05138078e-01 -6.95887581e-02
-1.06785131e+00 1.01043952e+00 2.91719824e-01 6.09469414e-01
-3.02994430e-01 3.09240997e-01 7.29939997e-01 5.35946608e-01
3.42455089e-01 5.57859540e-01 -4.96711701e-01 -1.86862275e-01
-7.73484349e-01 3.18631679e-01 1.33285832e+00 4.39125270e-01
4.58390623e-01 3.99931520e-02 -2.37878859e-01 8.41851592e-01
3.40998918e-01 1.88588098e-01 7.32127786e-01 -8.58957767e-01
3.77384484e-01 7.49053597e-01 -1.51658326e-01 -6.59629345e-01
-6.84367776e-01 -4.08294320e-01 -6.85931087e-01 -5.24482168e-02
4.72021192e-01 -3.70073766e-01 -6.92597032e-01 1.95965946e+00
1.15576580e-01 2.09372491e-02 4.06611830e-01 6.25724852e-01
1.11382508e+00 7.18981266e-01 -8.78661051e-02 -5.74491285e-02
1.58316088e+00 -9.74920809e-01 -7.01795638e-01 -3.30138415e-01
8.00377727e-01 -4.87356365e-01 7.49714375e-01 -4.72420715e-02
-1.43066919e+00 -4.91106272e-01 -1.26409078e+00 -4.33995306e-01
-5.13466239e-01 -2.64603168e-01 5.16245186e-01 3.46417129e-01
-1.32414436e+00 5.94809890e-01 -4.41237867e-01 -4.00023609e-01
1.72805414e-01 3.54059547e-01 -4.23909903e-01 5.76475680e-01
-1.41390550e+00 1.25608635e+00 3.36700201e-01 -6.61586877e-03
-5.85174143e-01 -3.63920867e-01 -1.17125428e+00 4.92790341e-01
6.48661107e-02 -1.15646327e+00 1.53928483e+00 -1.22492790e+00
-2.05406976e+00 8.74016285e-01 -4.53117162e-01 -7.00857997e-01
1.62901878e-01 1.38927877e-01 1.71371832e-01 1.61434904e-01
-6.74095824e-02 5.38260996e-01 7.75896013e-01 -1.14660776e+00
-5.17118752e-01 -2.26540729e-01 4.55553383e-01 3.94867122e-01
-1.74323127e-01 -1.36476725e-01 -2.18949810e-01 -3.78093421e-01
-1.12485047e-03 -6.52958632e-01 -1.50658518e-01 -4.54759926e-01
-3.85754764e-01 -4.06891972e-01 6.48792803e-01 -6.27338767e-01
1.18825650e+00 -2.31951904e+00 6.51638091e-01 -1.77561224e-01
4.06457990e-01 1.33941457e-01 -1.89897373e-01 8.14273596e-01
-3.25028479e-01 7.17151016e-02 -5.09573758e-01 -8.62545967e-01
2.46253118e-01 4.56654459e-01 -3.65707695e-01 2.34069020e-01
4.81041342e-01 9.45467055e-01 -7.37129092e-01 9.67142209e-02
9.78288129e-02 3.06286931e-01 -5.17690122e-01 2.46647522e-01
-5.10392725e-01 2.56788999e-01 -1.79333359e-01 -3.54254484e-01
2.66505867e-01 -4.16716993e-01 6.45086884e-01 1.72634721e-01
-9.47531760e-02 1.29080248e+00 -8.27331185e-01 1.80653298e+00
-6.46656215e-01 6.33351922e-01 3.85221690e-01 -1.27517378e+00
6.14995599e-01 7.43947327e-01 5.02226911e-02 -5.09650052e-01
3.63968670e-01 -9.53426138e-02 2.84514278e-01 -5.32661200e-01
6.98069990e-01 -6.85291648e-01 -4.32916790e-01 7.91394174e-01
4.75704432e-01 -2.88743109e-01 2.94225544e-01 4.43465739e-01
1.14256847e+00 -1.65995285e-01 2.97907799e-01 -3.30157757e-01
7.94238150e-01 -2.11122662e-01 1.35139108e-01 4.63826776e-01
2.35831216e-01 4.27880853e-01 1.03130949e+00 -5.43141782e-01
-1.05004990e+00 -1.19384503e+00 -2.20829081e-02 1.28658009e+00
-1.19293191e-01 -4.87048149e-01 -9.07783449e-01 -5.61627746e-01
-1.70778167e-02 1.00887156e+00 -8.79915118e-01 -2.19314352e-01
-5.79666317e-01 -3.04280698e-01 3.86573762e-01 3.19973499e-01
3.52647871e-01 -1.20295370e+00 -5.37839472e-01 3.13917547e-01
-2.24484354e-01 -1.11721301e+00 -3.69653463e-01 3.96288425e-01
-4.30750698e-01 -8.89499843e-01 -1.36319533e-01 -8.00669789e-01
2.74304926e-01 -5.11093922e-02 1.17509627e+00 1.56917110e-01
1.85016394e-01 3.48483831e-01 -3.54402006e-01 -3.41110229e-01
-1.08191609e+00 2.97394812e-01 -2.38365039e-01 2.97998071e-01
2.06330553e-01 -7.00540006e-01 -1.26468331e-01 -2.87406921e-01
-9.65690196e-01 2.85488188e-01 2.74129510e-01 1.04727542e+00
-4.47143018e-01 -3.17999810e-01 4.69000131e-01 -9.33620334e-01
8.82497847e-01 -5.14241874e-01 9.04965121e-03 3.31635438e-02
5.87057993e-02 3.19243044e-01 6.14807487e-01 -1.56198308e-01
-1.10633397e+00 -1.57247528e-01 -4.83042300e-01 2.63508230e-01
-3.51856351e-01 3.61716509e-01 -3.88816297e-02 6.74796999e-01
4.07158107e-01 3.67142558e-01 4.60948110e-01 -4.33447957e-01
5.58705986e-01 7.81933010e-01 5.03752053e-01 -2.47818142e-01
4.68473822e-01 5.12044966e-01 -2.18154848e-01 -1.19983244e+00
-8.07750165e-01 -4.22936589e-01 -7.76552260e-01 3.78717147e-02
1.30096269e+00 -6.81322634e-01 -8.41988742e-01 5.40937901e-01
-1.62119293e+00 -3.39515507e-01 -5.93173385e-01 2.55950034e-01
-5.76526463e-01 3.85317177e-01 -9.04980123e-01 -7.97002256e-01
-5.10734618e-01 -8.79276037e-01 1.03452408e+00 6.99909702e-02
-7.22938776e-01 -1.45530033e+00 -1.76018357e-01 3.15603942e-01
5.32701075e-01 -5.05023375e-02 1.16887069e+00 -1.02269316e+00
-1.46649301e-01 -5.09098805e-02 6.77606687e-02 5.82684100e-01
1.96356520e-01 -2.67887026e-01 -1.01737320e+00 5.94037473e-02
6.36803329e-01 -1.51679501e-01 1.09106183e+00 1.98258430e-01
1.92131847e-01 -4.45076734e-01 8.06168243e-02 1.66932449e-01
9.11172152e-01 8.42256024e-02 5.56111753e-01 8.13162848e-02
5.96005917e-01 9.16968107e-01 -4.64367598e-01 2.37352267e-01
7.10771263e-01 4.87445951e-01 3.28231275e-01 -8.21681470e-02
-4.78650816e-02 -1.73816383e-02 5.77690542e-01 1.04897189e+00
3.99761349e-01 -1.21629722e-02 -7.52641082e-01 4.81612265e-01
-1.96075404e+00 -1.29878008e+00 -1.76990137e-01 1.90359962e+00
1.06177163e+00 3.17407578e-01 5.45385368e-02 -2.51425970e-02
3.62131298e-01 4.68638450e-01 -1.14199758e-01 -9.98689592e-01
-1.30303904e-01 2.42478624e-01 3.38418572e-03 9.65395272e-01
-1.11385417e+00 9.77858901e-01 6.79610491e+00 3.59589100e-01
-1.18841076e+00 2.52078444e-01 1.25628471e-01 -5.66886589e-02
-4.19337183e-01 -1.93543024e-02 -7.51107574e-01 3.71500820e-01
1.14773822e+00 -1.17403395e-01 4.76755023e-01 3.79807711e-01
2.12766975e-01 -2.03686669e-01 -1.25720835e+00 4.82504338e-01
2.06773251e-01 -1.62388313e+00 8.26474875e-02 -5.71483374e-02
3.39228034e-01 1.70024812e-01 -2.90431291e-01 3.04286540e-01
3.72173995e-01 -1.05333400e+00 8.35852265e-01 5.42415142e-01
2.18158901e-01 -2.77310193e-01 9.27111506e-01 5.94446719e-01
-9.64623451e-01 -1.85006618e-01 5.04077151e-02 -6.96823120e-01
2.41172105e-01 5.88019863e-02 -7.60866940e-01 4.30935949e-01
1.04430012e-01 7.35837460e-01 -3.68423730e-01 1.25252843e-01
-4.91879165e-01 4.19525027e-01 -1.83258250e-01 -1.98471874e-01
5.68249285e-01 -7.98188448e-02 6.41003489e-01 1.46079063e+00
-3.99504066e-01 -2.17095673e-01 1.83562100e-01 7.37144589e-01
-5.30693419e-02 -5.32420091e-02 -4.45800424e-01 -2.09336802e-01
1.46476194e-01 1.15728569e+00 -3.92799944e-01 -7.97465563e-01
-4.92224365e-01 9.48948205e-01 3.37988585e-01 3.13467145e-01
-3.95526379e-01 -2.63703734e-01 9.32530284e-01 -1.14860900e-01
4.13381308e-01 -3.56262505e-01 -5.26476562e-01 -1.22692037e+00
1.38737142e-01 -7.35021651e-01 2.81457920e-02 -4.22683984e-01
-1.26726830e+00 6.59259796e-01 -2.47278988e-01 -5.06702602e-01
-7.15910971e-01 -6.48883820e-01 -1.03959596e+00 1.13744497e+00
-1.17018688e+00 -1.07311904e+00 1.28966406e-01 2.46450499e-01
6.68420732e-01 -1.96830258e-01 1.32626820e+00 -2.04200238e-01
-3.54400814e-01 9.38640609e-02 -1.94891945e-01 3.79470080e-01
1.53324351e-01 -1.50300896e+00 7.70213008e-01 4.28165823e-01
-1.71973839e-01 8.36075604e-01 7.79183745e-01 -1.94302090e-02
-1.10462403e+00 -6.69128418e-01 1.40007114e+00 -4.09112930e-01
8.94602656e-01 -8.65064502e-01 -9.69603240e-01 6.50279820e-01
7.43054211e-01 -7.91785002e-01 8.57423186e-01 3.25265765e-01
-5.37465155e-01 1.48659557e-01 -6.41607165e-01 5.37408233e-01
6.62808418e-01 -9.63337839e-01 -1.40758085e+00 1.82208121e-01
1.11423600e+00 -3.01800549e-01 -6.70240283e-01 -3.69156674e-02
5.83557189e-01 -1.10287511e+00 6.49495423e-01 -7.66422570e-01
8.17544162e-01 8.05378333e-03 -2.19562769e-01 -1.23437798e+00
-2.00938180e-01 -4.82035935e-01 8.07466805e-02 1.01002133e+00
7.07596183e-01 -6.89680576e-01 5.52959561e-01 5.85547626e-01
-2.69936442e-01 -6.97373867e-01 -9.28544879e-01 -3.92428845e-01
3.54393244e-01 -4.68169898e-01 5.28607368e-01 9.52561557e-01
4.94102269e-01 1.38131976e+00 2.59494912e-02 -2.39988744e-01
9.40278023e-02 1.05440810e-01 7.41229832e-01 -1.20718265e+00
-7.30319679e-01 -1.00720417e+00 -2.88011014e-01 -1.32644391e+00
4.12612379e-01 -1.11855352e+00 8.79026670e-03 -1.93542373e+00
1.08326964e-01 4.75419074e-01 1.62045538e-01 3.74364585e-01
-1.53346434e-01 -2.73970783e-01 3.94576460e-01 -1.24026738e-01
-3.96886468e-01 7.03867674e-01 8.86498630e-01 -3.44255805e-01
-2.19963893e-01 -4.32972759e-02 -7.46713459e-01 7.27822602e-01
8.02052975e-01 -5.07462695e-02 -3.12171966e-01 -4.32065457e-01
3.09376895e-01 4.15373668e-02 1.68116987e-01 -5.58180630e-01
2.32902616e-01 2.96738565e-01 -2.20050085e-02 -1.94185719e-01
5.74531615e-01 -6.71584129e-01 -3.51638585e-01 5.68846941e-01
-7.57257402e-01 -2.16661289e-01 1.24918386e-01 3.31620783e-01
-4.33388501e-01 -4.18560058e-01 7.28384852e-01 -3.99621934e-01
-1.71285644e-01 -3.77142131e-01 -8.84573936e-01 -1.10659651e-01
5.65510035e-01 -1.40658125e-01 -1.98846176e-01 -6.84358835e-01
-1.17840457e+00 3.02521735e-01 2.18270794e-01 6.66442811e-01
2.59608001e-01 -9.27610159e-01 -7.56606817e-01 4.02670115e-01
-2.96772689e-01 -9.33127070e-04 4.39256243e-02 7.40768194e-01
-2.66530812e-01 5.62114716e-01 1.91781297e-01 -4.49453890e-01
-1.24177039e+00 3.38455409e-01 5.65958917e-01 -5.16936839e-01
-5.64135790e-01 8.31952572e-01 4.55474585e-01 -8.00312161e-01
2.04047486e-01 -4.37889129e-01 -5.23945451e-01 3.61510962e-01
6.95203364e-01 -7.13841766e-02 -1.81221053e-01 -7.93112576e-01
-7.34404996e-02 1.87421158e-01 -6.17119521e-02 -3.08148265e-01
1.39353859e+00 -2.02830046e-01 -6.36231780e-01 1.17677617e+00
1.33551002e+00 -2.45498389e-01 -8.09028566e-01 -2.65786141e-01
2.97371626e-01 4.11521941e-01 -2.05448091e-01 -5.65241396e-01
-3.92470241e-01 1.09746218e+00 -1.91433012e-01 8.59284937e-01
6.65220201e-01 8.15228894e-02 8.21290016e-01 3.66717875e-01
-1.36468738e-01 -1.03104496e+00 4.90441509e-02 1.39778292e+00
1.04978168e+00 -8.38770807e-01 -3.83537650e-01 -3.03353339e-01
-5.96261799e-01 1.30772173e+00 1.26100704e-01 -1.14497550e-01
7.08186388e-01 4.05896813e-01 1.37978613e-01 -3.46911490e-01
-1.20877075e+00 -2.18424171e-01 -3.85499522e-02 1.39042541e-01
6.45560741e-01 8.11090395e-02 -3.38222653e-01 6.65350497e-01
-4.93575513e-01 -3.32443863e-01 5.38902879e-01 7.09469795e-01
-5.02818763e-01 -1.21184683e+00 -7.12701306e-02 2.92402565e-01
-3.21681321e-01 -3.11848104e-01 -6.70946538e-01 4.71437365e-01
4.06338796e-02 1.13979840e+00 5.84994316e-01 -4.56013829e-01
2.74318129e-01 7.50719309e-01 2.03663588e-01 -1.13624096e+00
-1.21426618e+00 -1.10320374e-01 6.20678365e-01 -3.48050803e-01
-3.82154942e-01 -5.27067125e-01 -1.70674086e+00 -4.24240559e-01
4.45097731e-03 2.41583884e-01 8.49264741e-01 1.50036085e+00
2.82522470e-01 8.97127390e-01 6.43450260e-01 -8.85696352e-01
-7.45598793e-01 -1.36445868e+00 -5.39501131e-01 5.39386332e-01
5.61920345e-01 3.97080556e-02 -2.94491380e-01 6.43719658e-02] | [12.535086631774902, 7.560083866119385] |
a26d2350-92f9-4069-93d8-47ee7052923c | self-supervised-feature-learning-for-long | 2212.00122 | null | https://arxiv.org/abs/2212.00122v1 | https://arxiv.org/pdf/2212.00122v1.pdf | Self-Supervised Feature Learning for Long-Term Metric Visual Localization | Visual localization is the task of estimating camera pose in a known scene, which is an essential problem in robotics and computer vision. However, long-term visual localization is still a challenge due to the environmental appearance changes caused by lighting and seasons. While techniques exist to address appearance changes using neural networks, these methods typically require ground-truth pose information to generate accurate image correspondences or act as a supervisory signal during training. In this paper, we present a novel self-supervised feature learning framework for metric visual localization. We use a sequence-based image matching algorithm across different sequences of images (i.e., experiences) to generate image correspondences without ground-truth labels. We can then sample image pairs to train a deep neural network that learns sparse features with associated descriptors and scores without ground-truth pose supervision. The learned features can be used together with a classical pose estimator for visual stereo localization. We validate the learned features by integrating with an existing Visual Teach & Repeat pipeline to perform closed-loop localization experiments under different lighting conditions for a total of 22.4 km. | ['Timothy D. Barfoot', 'Yuxuan Chen'] | 2022-11-30 | null | null | null | null | ['visual-localization'] | ['computer-vision'] | [ 6.00419492e-02 -4.43134576e-01 -6.82821544e-03 -6.54711008e-01
-8.05482507e-01 -8.48438263e-01 6.44274950e-01 -2.11154353e-02
-6.52140558e-01 4.96916801e-01 -8.69594142e-02 1.88891411e-01
7.88755044e-02 -4.90318030e-01 -1.29210556e+00 -4.51418996e-01
1.06668755e-01 5.79278290e-01 3.01981986e-01 -3.71351056e-02
4.21385586e-01 7.45045006e-01 -1.83962643e+00 -4.53275219e-02
5.89176655e-01 1.05262423e+00 5.77048481e-01 5.88411629e-01
5.23185544e-02 6.43543780e-01 -4.11993086e-01 2.98624672e-02
6.13477468e-01 -2.35298261e-01 -5.92028856e-01 4.30701375e-01
1.08895588e+00 -4.61676717e-01 -3.59406978e-01 1.03238177e+00
3.71734619e-01 2.09350452e-01 4.90377843e-01 -1.27549088e+00
-3.18512559e-01 -1.64976820e-01 -4.60499227e-01 -1.65790096e-01
6.11247599e-01 3.79542500e-01 7.87119389e-01 -8.91409934e-01
7.60412693e-01 1.01981604e+00 8.91707659e-01 2.80474931e-01
-1.38443708e+00 -4.87290353e-01 -8.86438712e-02 4.07464206e-01
-1.44723964e+00 -5.12914717e-01 7.73401201e-01 -6.43154562e-01
7.46356130e-01 -1.58847541e-01 6.97444141e-01 1.16501796e+00
-8.58187079e-02 4.25317556e-01 9.57821727e-01 -4.15494382e-01
1.76814377e-01 1.49104789e-01 -4.82564032e-01 8.76271963e-01
-3.18853818e-02 3.35124791e-01 -7.76724994e-01 2.03663737e-01
7.71536410e-01 3.24494749e-01 -3.38996530e-01 -1.06805563e+00
-1.45662844e+00 7.82055318e-01 1.05552697e+00 -2.75821146e-03
-3.13562393e-01 4.22939837e-01 8.23231190e-02 2.00959131e-01
2.08908301e-02 5.67847192e-01 -3.27388614e-01 1.65413722e-01
-1.03105414e+00 1.18419975e-01 5.42846084e-01 1.00025415e+00
1.47857952e+00 -8.96332860e-02 1.70148667e-02 7.85310090e-01
2.94680566e-01 9.40375984e-01 4.23902154e-01 -1.23632658e+00
2.14354575e-01 4.06156689e-01 3.11633110e-01 -1.21881771e+00
-3.40117574e-01 -2.95629054e-01 -4.85033721e-01 3.97867590e-01
5.04387200e-01 1.07714988e-01 -8.49333346e-01 1.68321002e+00
3.24966878e-01 2.06732154e-01 -1.59979612e-01 1.20546746e+00
5.48046887e-01 4.72772419e-01 -3.85746419e-01 2.47901440e-01
8.73354197e-01 -1.12785709e+00 -2.02871189e-01 -8.02522302e-01
3.65106642e-01 -8.86350572e-01 8.15314114e-01 1.22823134e-01
-4.04750973e-01 -5.97871304e-01 -9.64262545e-01 -1.33822784e-01
-4.29005951e-01 4.84747678e-01 3.23797792e-01 8.79065692e-02
-1.26486945e+00 4.33495402e-01 -6.76653683e-01 -8.00734222e-01
-1.64959989e-02 3.25291753e-01 -7.97928572e-01 -8.52691755e-02
-8.41689348e-01 1.08851504e+00 2.87111700e-01 2.50260979e-01
-1.30352557e+00 -3.78559828e-01 -1.29238820e+00 -3.29641312e-01
1.38608277e-01 -5.45731664e-01 1.23183501e+00 -1.25819933e+00
-1.36941135e+00 1.10091007e+00 -2.50558168e-01 -4.08846676e-01
5.97256958e-01 -2.73608088e-01 1.26350880e-01 1.31915256e-01
4.51344460e-01 1.03169775e+00 9.57310855e-01 -1.59252930e+00
-6.67524397e-01 -2.76388764e-01 2.65364386e-02 3.48008126e-01
1.73275336e-03 -3.39962989e-01 -5.39551020e-01 -1.38408929e-01
2.41560549e-01 -1.15580916e+00 -1.95676535e-01 4.87781793e-01
-1.05628699e-01 2.34421983e-01 7.06534505e-01 -6.15312934e-01
3.05259347e-01 -1.93610013e+00 1.99306682e-01 1.92833513e-01
-1.37404785e-01 -6.48056120e-02 -3.27254057e-01 3.86337519e-01
9.95500684e-02 -6.74316525e-01 -8.08547363e-02 -5.79157472e-01
-2.11227104e-01 1.45349517e-01 -1.08887434e-01 8.29665303e-01
2.06771821e-01 8.07556450e-01 -1.05170107e+00 -2.86888659e-01
7.40903616e-01 4.45370495e-01 -4.52822357e-01 5.78435421e-01
-3.61937463e-01 8.20890665e-01 1.23543210e-01 5.52715182e-01
5.83707750e-01 -1.39376238e-01 -9.22834426e-02 -4.24654722e-01
-2.61054248e-01 -6.98339418e-02 -1.11281097e+00 2.38049436e+00
-8.97871256e-01 9.69800711e-01 -1.29727954e-02 -1.00649154e+00
1.19539404e+00 -2.52045840e-01 2.43148521e-01 -7.39089549e-01
2.15971246e-01 3.45697820e-01 -4.80926126e-01 -3.77521634e-01
4.68921036e-01 2.11369514e-01 -9.44964811e-02 8.46029744e-02
2.86553860e-01 -4.68536019e-01 -4.66777757e-02 -5.32154068e-02
9.90492165e-01 4.80167180e-01 9.59279090e-02 9.55613237e-03
5.66864073e-01 2.71816581e-01 4.43754345e-01 4.97766227e-01
-2.97399499e-02 9.35967386e-01 -1.54439747e-01 -7.36731350e-01
-1.31024146e+00 -1.14268827e+00 1.20282084e-01 9.01129007e-01
5.36264479e-01 -8.36640894e-02 -4.29160953e-01 -4.79886174e-01
1.22112557e-01 2.31173113e-01 -5.50472856e-01 -1.24023288e-01
-3.08923990e-01 6.91376105e-02 1.59755692e-01 4.31040436e-01
5.44077575e-01 -9.71199036e-01 -8.61550629e-01 -2.39373073e-02
-2.40941912e-01 -1.17411268e+00 -4.64771807e-01 3.82345706e-01
-3.56373906e-01 -1.17637873e+00 -6.70223475e-01 -9.41393793e-01
9.33605134e-01 5.81868768e-01 9.84725893e-01 -2.63381958e-01
-4.44847286e-01 6.24228418e-01 -1.95867673e-01 -1.18059725e-01
-2.24344537e-01 8.04612637e-02 2.38308087e-01 8.78733322e-02
1.95194587e-01 -4.79224235e-01 -7.92384565e-01 5.12566149e-01
-6.05722368e-01 -1.06437253e-02 6.94360077e-01 1.04314017e+00
7.08657563e-01 -5.83967149e-01 -2.22783834e-02 -2.94245303e-01
1.12846671e-02 -1.87841907e-01 -1.26035643e+00 2.57703424e-01
-2.57071704e-01 3.41870636e-01 5.79956353e-01 -2.59876698e-01
-7.10520148e-01 9.19008732e-01 1.91249296e-01 -9.14079428e-01
-3.09263408e-01 3.50388676e-01 -5.31334579e-02 -5.91842711e-01
8.72509837e-01 3.36955637e-01 2.52428472e-01 -1.67864263e-01
4.92474109e-01 6.60522401e-01 9.33260679e-01 -4.19525921e-01
1.19394767e+00 5.59601545e-01 6.20387048e-02 -7.19795883e-01
-1.09504998e+00 -8.32321703e-01 -9.43857431e-01 -4.67103958e-01
7.87912726e-01 -1.31670308e+00 -6.37194812e-01 4.06560868e-01
-1.28838909e+00 -4.44254905e-01 1.41791806e-01 5.13000369e-01
-8.08151126e-01 2.66343117e-01 -1.24109268e-01 -4.07812893e-01
-9.15236399e-02 -1.12285721e+00 1.57415450e+00 3.40813816e-01
-9.16107297e-02 -9.25943136e-01 2.90035367e-01 3.14999133e-01
3.08014154e-01 2.59623557e-01 -7.90124610e-02 -4.91280342e-03
-9.26478267e-01 -3.03199351e-01 -2.94217885e-01 3.25319320e-01
2.42992401e-01 -2.26989716e-01 -9.41076338e-01 -5.04692852e-01
-3.81198525e-01 -7.26545691e-01 6.77776754e-01 1.95054814e-01
1.01185417e+00 1.30635202e-02 -2.98897564e-01 1.06076431e+00
1.59056687e+00 -2.06059992e-01 2.76604414e-01 6.82481647e-01
7.57433414e-01 6.78419173e-01 7.56743252e-01 3.64723176e-01
4.71527636e-01 9.98560429e-01 6.37436926e-01 -1.87381208e-01
-1.29467091e-02 -5.33470571e-01 3.37889433e-01 3.40584159e-01
2.98956126e-01 2.87762761e-01 -1.11370599e+00 6.41964853e-01
-1.89166963e+00 -8.23225081e-01 1.34951994e-01 2.45523667e+00
7.43588746e-01 -3.34018260e-01 -1.99445218e-01 -2.65041202e-01
7.11211920e-01 3.31368186e-02 -5.68555593e-01 -1.25614251e-03
-2.73385481e-03 -2.49010295e-01 9.16401207e-01 5.97256243e-01
-1.27751493e+00 1.03751290e+00 5.37383652e+00 1.86792746e-01
-1.48179793e+00 -1.51019827e-01 1.68552503e-01 8.85173827e-02
3.76799777e-02 1.27696633e-01 -6.44377649e-01 2.92563468e-01
5.75230658e-01 2.32881159e-02 7.38140821e-01 1.08869112e+00
1.13358289e-01 -3.59385461e-01 -1.28136230e+00 1.44960678e+00
5.37333786e-01 -1.26191926e+00 -3.39301080e-01 -1.64892167e-01
8.53537798e-01 4.12475377e-01 -1.27845377e-01 -5.37410658e-03
3.94702762e-01 -8.95808101e-01 9.18326199e-01 5.54738998e-01
8.99536908e-01 -5.47736168e-01 6.20596945e-01 4.22446996e-01
-1.14920318e+00 -8.87231678e-02 -4.84352231e-01 -8.39079767e-02
1.18454747e-01 3.07974100e-01 -1.01512969e+00 5.13789892e-01
8.62015665e-01 1.11654758e+00 -9.26305652e-01 1.38755023e+00
-4.48218495e-01 -1.12541849e-02 -4.84617949e-01 5.98034002e-02
2.33206198e-01 -2.02702075e-01 3.14461142e-01 8.45248044e-01
4.56924856e-01 -7.39846647e-01 3.95849496e-01 7.88582981e-01
2.43935082e-02 -1.88766614e-01 -1.04945505e+00 3.79195213e-01
4.70728904e-01 1.44858754e+00 -5.28386772e-01 -5.26717864e-02
-4.10434455e-01 1.46008992e+00 5.65516591e-01 2.38303006e-01
-7.39067912e-01 -3.05100322e-01 6.21862173e-01 1.31937042e-01
2.61302501e-01 -5.45611918e-01 2.20058188e-01 -1.19281721e+00
1.36590987e-01 -5.69778025e-01 -1.51042596e-01 -1.01907885e+00
-1.18257344e+00 4.56411213e-01 -1.49387255e-01 -1.64291322e+00
-4.56235051e-01 -5.26046813e-01 -3.63184243e-01 8.91068041e-01
-1.58935833e+00 -1.36244094e+00 -1.08513653e+00 5.29415965e-01
3.87906462e-01 -1.46317616e-01 7.43346870e-01 2.78905660e-01
-1.33769676e-01 3.32932025e-01 2.44874313e-01 2.90091276e-01
1.09768593e+00 -1.27267444e+00 1.83144465e-01 8.81528914e-01
4.15237069e-01 4.31073725e-01 7.89817095e-01 -2.41263837e-01
-1.54261065e+00 -1.42009187e+00 6.22318566e-01 -4.81310487e-01
5.35322905e-01 -6.02846682e-01 -5.33273160e-01 6.72468901e-01
1.17423713e-01 4.46733445e-01 1.26572549e-01 -1.39477953e-01
-3.99647743e-01 -5.70802629e-01 -9.08599138e-01 3.38568807e-01
1.12741387e+00 -7.89593279e-01 -4.57219869e-01 4.26159173e-01
4.54023778e-01 -7.49266088e-01 -4.51114833e-01 1.53402448e-01
5.91514170e-01 -9.61792827e-01 1.03143156e+00 -1.08017307e-03
1.84812903e-01 -8.34609151e-01 -2.00142950e-01 -1.40351415e+00
-8.58177021e-02 -2.66942590e-01 5.72151899e-01 1.11711812e+00
1.76069960e-01 -4.45724845e-01 8.34941328e-01 2.33074933e-01
4.86836284e-02 5.88063449e-02 -9.07987475e-01 -9.47395205e-01
-3.82875383e-01 -1.79221034e-01 2.18581274e-01 7.14304209e-01
-4.62200373e-01 3.99244070e-01 -6.17615044e-01 4.08096731e-01
8.74442041e-01 3.80858183e-01 1.27881682e+00 -1.11078501e+00
-2.22051233e-01 -1.24823209e-02 -9.16040182e-01 -1.00637841e+00
6.28507257e-01 -9.07926619e-01 6.95315659e-01 -1.62633407e+00
8.27408507e-02 -2.52170146e-01 1.25476196e-01 4.77239668e-01
2.19356567e-01 4.54759002e-01 1.75895747e-02 1.73979446e-01
-8.38666260e-01 7.30821192e-01 7.08922327e-01 -2.53142089e-01
3.15185674e-02 -2.68763214e-01 3.08922934e-03 4.81549203e-01
7.25081146e-01 -2.66162127e-01 -3.49431783e-01 -7.43129015e-01
1.99987441e-01 -1.39093429e-01 8.09743226e-01 -1.37059343e+00
4.32760507e-01 -1.13271587e-01 6.74591362e-01 -5.22792101e-01
4.01692808e-01 -1.24051595e+00 2.69013822e-01 4.52336192e-01
-1.86295122e-01 1.13408580e-01 2.40808632e-02 6.24882996e-01
-4.18301702e-01 -1.53698981e-01 7.97945499e-01 -9.10018981e-02
-1.34193504e+00 3.51175517e-01 1.36481766e-02 -2.28731543e-01
1.11130273e+00 -2.27797151e-01 -1.22571118e-01 -5.70214510e-01
-3.09609622e-01 3.32592160e-01 1.10538948e+00 6.47881389e-01
7.40859687e-01 -1.51443005e+00 -3.96072060e-01 2.96321660e-01
7.53581703e-01 3.31895769e-01 1.65012367e-02 5.83528221e-01
-9.61959004e-01 2.23611265e-01 -3.77046585e-01 -1.26820731e+00
-1.15272224e+00 4.72355962e-01 5.97021043e-01 3.83995980e-01
-3.27992558e-01 6.91236973e-01 8.16431642e-02 -8.33583117e-01
2.56360590e-01 -6.88172206e-02 1.36996597e-01 -1.42264500e-01
3.00363958e-01 -1.09433740e-01 8.97891968e-02 -9.07997012e-01
-5.20923674e-01 9.81058180e-01 4.02904749e-01 -1.21290669e-01
1.30171001e+00 -2.88800508e-01 -5.92009723e-02 4.40394521e-01
1.47360301e+00 -1.75655738e-01 -1.62803423e+00 -4.27095473e-01
1.27775982e-01 -7.32513189e-01 -4.42875968e-03 -4.91368026e-01
-9.61066186e-01 8.18867803e-01 9.23104227e-01 -5.63867450e-01
8.38217974e-01 1.14936590e-01 3.49291354e-01 7.49551833e-01
8.03363085e-01 -1.07683468e+00 3.88301402e-01 6.15775347e-01
1.02352428e+00 -1.71290219e+00 -1.68179646e-01 -1.03838146e-02
-4.33898091e-01 1.20121813e+00 7.11191118e-01 -1.67820260e-01
2.78809041e-01 1.55715682e-02 4.91936654e-01 1.18356436e-01
-3.09544355e-01 -5.48512876e-01 2.70876735e-01 8.72619629e-01
2.10734680e-01 -2.16118246e-01 2.93283045e-01 -2.52640486e-01
-2.49285877e-01 -1.26513630e-01 2.64379710e-01 7.85915494e-01
-4.31382924e-01 -1.02227843e+00 -4.82596189e-01 4.20338847e-02
3.13995957e-01 1.23533122e-01 -4.22087073e-01 3.64590585e-01
2.39202529e-01 7.61596382e-01 8.89649093e-02 -5.16965568e-01
2.77522653e-01 -1.90949097e-01 6.34541988e-01 -6.23545527e-01
-1.21438041e-01 -2.25282639e-01 -2.09577426e-01 -9.53888774e-01
-6.47042751e-01 -8.56490612e-01 -8.42912316e-01 5.01881577e-02
-1.99100077e-01 -1.15742631e-01 1.03422344e+00 7.28627920e-01
3.23573619e-01 1.53578565e-01 8.97963345e-01 -1.44884634e+00
-4.31401789e-01 -8.68333876e-01 -2.80818790e-01 6.75908566e-01
5.98419666e-01 -8.45543802e-01 -4.04139489e-01 2.62021363e-01] | [7.714034080505371, -2.1335225105285645] |
bc60f30d-02ce-453d-8287-1a4954dbf3d8 | weakly-supervised-hoi-detection-via-prior | 2303.01313 | null | https://arxiv.org/abs/2303.01313v1 | https://arxiv.org/pdf/2303.01313v1.pdf | Weakly-supervised HOI Detection via Prior-guided Bi-level Representation Learning | Human object interaction (HOI) detection plays a crucial role in human-centric scene understanding and serves as a fundamental building-block for many vision tasks. One generalizable and scalable strategy for HOI detection is to use weak supervision, learning from image-level annotations only. This is inherently challenging due to ambiguous human-object associations, large search space of detecting HOIs and highly noisy training signal. A promising strategy to address those challenges is to exploit knowledge from large-scale pretrained models (e.g., CLIP), but a direct knowledge distillation strategy~\citep{liao2022gen} does not perform well on the weakly-supervised setting. In contrast, we develop a CLIP-guided HOI representation capable of incorporating the prior knowledge at both image level and HOI instance level, and adopt a self-taught mechanism to prune incorrect human-object associations. Experimental results on HICO-DET and V-COCO show that our method outperforms the previous works by a sizable margin, showing the efficacy of our HOI representation. | ['Xuming He', 'Tinne Tuytelaars', 'Desen Zhou', 'Yongfei Liu', 'Bo Wan'] | 2023-03-02 | null | null | null | null | ['human-object-interaction-detection'] | ['computer-vision'] | [ 2.69747764e-01 1.41017392e-01 -1.37377962e-01 -2.33400866e-01
-5.65163374e-01 -2.26896256e-01 4.12447602e-01 -3.77665609e-02
-4.70731914e-01 6.81882679e-01 3.79596613e-02 1.04494847e-01
1.12933457e-01 -3.80171090e-01 -9.22592580e-01 -6.37268007e-01
1.54366851e-01 4.98708814e-01 7.59466827e-01 -1.56067103e-01
-1.58267897e-02 2.90304631e-01 -1.79218090e+00 3.75656486e-01
8.13014090e-01 8.95753562e-01 4.60131347e-01 5.85137963e-01
4.49580580e-01 8.84770632e-01 -4.39299136e-01 -2.46410400e-01
2.55898118e-01 -4.49733377e-01 -7.62654960e-01 1.97166145e-01
7.40781665e-01 -2.23714933e-01 -4.24737692e-01 1.19557464e+00
5.03549755e-01 1.69852227e-02 5.79040945e-01 -1.28930414e+00
-2.97641367e-01 2.52406001e-01 -6.85051441e-01 4.30730373e-01
2.38876730e-01 5.75515926e-01 1.02202857e+00 -1.14788151e+00
8.19788873e-01 1.11388671e+00 4.88766819e-01 5.69292545e-01
-1.18100309e+00 -7.41656542e-01 1.07853770e-01 5.39255500e-01
-1.58861935e+00 -2.17266575e-01 7.49933720e-01 -5.17461836e-01
8.93461525e-01 2.51410395e-01 7.80036628e-01 1.16316283e+00
-1.49639025e-01 1.48993754e+00 1.15296674e+00 -4.51375902e-01
-8.69457275e-02 3.95352751e-01 2.39274636e-01 8.70558739e-01
3.81949484e-01 1.25919461e-01 -7.71319091e-01 7.71145746e-02
6.73279762e-01 4.16123448e-03 -4.78731453e-01 -5.68737090e-01
-1.46967936e+00 6.54244006e-01 7.74918020e-01 1.50354594e-01
-2.04940140e-01 -2.94641964e-02 2.91969925e-01 -8.22555870e-02
4.86915559e-03 5.97069323e-01 -2.59456098e-01 3.74316394e-01
-7.08844125e-01 3.98877174e-01 5.26326180e-01 1.12575984e+00
7.82415748e-01 -1.82563603e-01 -5.18383801e-01 7.59183943e-01
9.37178265e-03 5.42928874e-01 1.56766757e-01 -8.77872288e-01
4.97969478e-01 6.52410448e-01 8.95919949e-02 -8.82801592e-01
-3.12667340e-01 -5.91442049e-01 -9.61319268e-01 3.83537650e-01
4.95563447e-01 6.53493255e-02 -1.17489433e+00 1.55303895e+00
4.57783163e-01 2.64811546e-01 -2.54943609e-01 1.15768278e+00
9.26731944e-01 4.04757470e-01 1.86091557e-01 2.75953244e-02
1.29860258e+00 -1.30171323e+00 -4.39907223e-01 -4.78940606e-01
4.83348995e-01 -4.82697189e-01 1.01061523e+00 3.79596919e-01
-8.92123997e-01 -6.41545534e-01 -1.05287957e+00 -2.17231512e-01
-4.07803297e-01 3.63725960e-01 6.55639708e-01 1.93918407e-01
-7.35597074e-01 5.28659485e-02 -3.33164901e-01 -4.98705208e-01
7.15740204e-01 2.01575309e-01 -3.88656557e-01 -3.18367302e-01
-1.03498518e+00 7.51975894e-01 8.63491833e-01 2.99047768e-01
-1.24227464e+00 -4.83088553e-01 -8.72236788e-01 4.78046946e-02
8.75508487e-01 -8.71184468e-01 8.76337528e-01 -7.48501599e-01
-1.10851133e+00 1.05682123e+00 -4.36418355e-02 -6.99808598e-01
7.61221111e-01 -3.42189997e-01 4.88072298e-02 3.87746304e-01
1.24944545e-01 1.04344344e+00 1.08615148e+00 -1.44487727e+00
-9.65915203e-01 -2.12791860e-01 7.09582493e-02 3.70493531e-01
-2.06769094e-01 1.03475027e-01 -9.31917906e-01 -4.31743562e-01
-1.14002451e-01 -1.28664720e+00 -2.67205358e-01 3.28161754e-02
-6.38133287e-01 -4.38433200e-01 8.36110592e-01 -5.90543687e-01
1.06635308e+00 -2.06553435e+00 1.37458056e-01 1.34225115e-01
4.27815109e-01 6.66110396e-01 -7.22938851e-02 6.01429343e-02
1.37803152e-01 -1.89354122e-01 -3.45400035e-01 -1.41950145e-01
-2.86011845e-01 2.06520751e-01 -5.76577298e-02 4.46388870e-01
3.91778499e-01 1.29614985e+00 -1.06123555e+00 -7.21714616e-01
3.88549328e-01 3.07034552e-01 -6.88968062e-01 4.46720839e-01
-3.83267462e-01 6.73239529e-01 -2.42174432e-01 7.51658976e-01
3.08328062e-01 -5.71634650e-01 5.93073480e-03 -2.87941426e-01
7.40932822e-02 -5.78662008e-02 -1.13424122e+00 1.49499822e+00
-1.47552252e-01 6.15968883e-01 -6.36762455e-02 -1.10611331e+00
4.88413304e-01 2.91192621e-01 3.01174253e-01 -5.10633945e-01
5.40960915e-02 6.85120299e-02 1.88913524e-01 -5.43345511e-01
5.09359725e-02 3.47195804e-01 2.03218646e-02 -2.74768531e-01
3.22949857e-01 -4.82250983e-03 2.10623696e-01 3.82557362e-01
1.12237525e+00 -1.68934181e-01 5.40018618e-01 -5.20681180e-02
4.97747242e-01 7.53147155e-02 7.28750587e-01 1.26487625e+00
-5.95963418e-01 7.96827853e-01 2.86284536e-01 -3.51125479e-01
-9.94933307e-01 -7.08953738e-01 -7.18412921e-02 1.16030216e+00
3.97505194e-01 -2.88288683e-01 -6.23450279e-01 -9.58600819e-01
-5.38404807e-02 3.86340380e-01 -7.54310071e-01 -9.53152329e-02
-6.31934404e-01 -5.80780983e-01 2.02003822e-01 7.12984860e-01
7.21084416e-01 -1.38942027e+00 -6.95679307e-01 4.86272201e-02
-2.93673009e-01 -1.37242627e+00 -3.83273482e-01 3.41781408e-01
-2.91467905e-01 -1.14850247e+00 -7.61352837e-01 -8.67605150e-01
6.22387707e-01 6.39850676e-01 1.21694362e+00 2.10311174e-01
-8.73588443e-01 3.19834620e-01 -2.83158988e-01 -6.73100770e-01
8.29596519e-02 3.17212157e-02 -1.54387355e-01 1.08313926e-01
3.09746981e-01 -2.34724969e-01 -8.80049825e-01 5.50597429e-01
-6.03858948e-01 2.05787420e-01 1.01648736e+00 1.18117118e+00
5.66933036e-01 8.79787728e-02 3.04388911e-01 -1.12231922e+00
-1.54023156e-01 -2.52455831e-01 -5.02928138e-01 3.29868019e-01
-4.03138936e-01 -2.30101913e-01 3.52634460e-01 -4.67556894e-01
-1.24794948e+00 5.28374314e-01 7.47060180e-02 -6.74128115e-01
-5.32957315e-01 1.94081828e-01 -4.52838749e-01 -2.23883584e-01
7.38015354e-01 2.10245565e-01 -3.23800385e-01 -2.96460927e-01
3.87064010e-01 3.88048738e-01 1.04165173e+00 -4.36484456e-01
9.63962495e-01 6.35093749e-01 -2.55501300e-01 -1.00216758e+00
-1.53194416e+00 -1.19647491e+00 -6.61456048e-01 -1.63266689e-01
1.16017675e+00 -1.16780293e+00 -6.57911956e-01 2.89068699e-01
-1.21001923e+00 -2.67600417e-01 -1.27182573e-01 3.73279572e-01
-4.65557814e-01 2.90332884e-01 -4.04804230e-01 -7.85433590e-01
-2.13707909e-01 -1.03846133e+00 1.29386067e+00 2.86280513e-01
3.70691456e-02 -4.39583421e-01 -2.08258778e-01 8.46300364e-01
-4.42401543e-02 1.72400385e-01 5.47086239e-01 -5.78044713e-01
-9.65272367e-01 -2.60545820e-01 -7.02676892e-01 4.18910205e-01
-5.27546048e-01 -4.81873602e-01 -1.18639553e+00 -3.31764281e-01
-1.83012441e-01 -7.61374652e-01 1.12262404e+00 5.97676858e-02
1.38266993e+00 -1.34525582e-01 -4.11778331e-01 6.49280190e-01
1.24156606e+00 -2.32943848e-01 6.55609548e-01 2.09039882e-01
1.20355284e+00 6.01228893e-01 7.58498192e-01 2.95266539e-01
4.22218680e-01 9.03794229e-01 4.79592800e-01 -3.45730662e-01
-5.47374785e-01 -2.15547949e-01 1.20951861e-01 3.47916842e-01
-2.91463464e-01 -6.83969483e-02 -1.10444784e+00 7.95865834e-01
-2.23738384e+00 -1.08762348e+00 -3.28338265e-01 2.09825349e+00
7.37655461e-01 3.35245788e-01 2.58063555e-01 4.77123335e-02
6.10816836e-01 -3.65034454e-02 -4.44875449e-01 5.37837803e-01
-2.28580162e-01 -8.50735754e-02 5.81734300e-01 1.27733439e-01
-1.32007265e+00 1.21645725e+00 5.70192671e+00 7.42613792e-01
-8.08823764e-01 2.41296858e-01 3.08379352e-01 1.77423209e-02
4.07409608e-01 -9.57288742e-02 -1.12189591e+00 3.55892390e-01
3.50292414e-01 8.31387267e-02 2.50080675e-01 1.11381078e+00
-5.51059246e-02 -2.66026199e-01 -1.19632697e+00 1.27820599e+00
4.17313546e-01 -1.22400761e+00 -4.35612649e-02 1.05103113e-01
8.85643721e-01 2.96069860e-01 -2.93715913e-02 5.09713471e-01
1.59017667e-01 -9.11596775e-01 5.40083647e-01 3.38536501e-01
5.92111111e-01 -3.84223312e-01 9.11638737e-01 5.66633046e-01
-1.26446247e+00 -1.21800736e-01 -4.44129050e-01 1.37305960e-01
-8.93253535e-02 4.72967803e-01 -1.02048504e+00 3.65902811e-01
1.07947910e+00 7.15111613e-01 -8.23673308e-01 1.28788388e+00
-5.93888462e-01 7.11558700e-01 -5.14219582e-01 3.09597611e-01
1.94330383e-02 2.56389499e-01 6.00939453e-01 1.37833345e+00
-2.72225648e-01 3.07305813e-01 7.88550496e-01 6.14586234e-01
-1.13928467e-01 -5.35140820e-02 -6.98315024e-01 1.11862890e-01
1.87984750e-01 1.24760163e+00 -6.31643116e-01 -5.13909638e-01
-7.02175558e-01 1.34348917e+00 6.52774990e-01 4.23605889e-01
-8.60874116e-01 -3.41397762e-01 2.77521312e-01 1.73129514e-01
5.60797215e-01 -2.36116379e-01 -1.22301482e-01 -1.37070537e+00
1.20603509e-01 -9.54509914e-01 6.89158380e-01 -5.65161347e-01
-1.24460411e+00 3.68314236e-01 9.37114879e-02 -1.19071972e+00
-1.47046909e-01 -6.08790755e-01 -5.34064233e-01 5.77677190e-01
-1.79036295e+00 -1.53939795e+00 -7.86309063e-01 8.87229383e-01
7.68090248e-01 1.06222034e-01 4.36459243e-01 3.00266266e-01
-7.02344418e-01 7.32676864e-01 -3.71344090e-01 2.59544164e-01
7.08515465e-01 -1.37692821e+00 -5.60662784e-02 9.34441507e-01
5.14187932e-01 5.02029300e-01 7.99220562e-01 -5.98615050e-01
-1.27912104e+00 -1.25168705e+00 6.58018649e-01 -9.20407891e-01
5.13933063e-01 -6.15318775e-01 -1.12418056e+00 6.58099949e-01
1.26329765e-01 6.43466055e-01 3.49564672e-01 5.49441241e-02
-4.34985846e-01 -2.00163033e-02 -6.35153472e-01 4.55760121e-01
1.46475959e+00 -5.51708341e-01 -5.99249899e-01 6.20352745e-01
4.71756339e-01 -3.62824410e-01 -5.14976621e-01 7.54837811e-01
3.17352355e-01 -9.38043654e-01 1.13422620e+00 -4.33918387e-01
1.62927419e-01 -6.09818339e-01 1.22091316e-01 -8.51953745e-01
-4.31710064e-01 -5.56840718e-01 -3.66822213e-01 1.01702046e+00
2.54412621e-01 -1.92403033e-01 8.30315769e-01 3.81848663e-01
4.36834525e-03 -6.75755858e-01 -6.17310464e-01 -8.85656297e-01
-5.04308879e-01 -3.42179120e-01 -1.61259741e-01 7.46643841e-01
-1.84545115e-01 6.45171583e-01 -7.23894298e-01 6.06258750e-01
1.04177797e+00 5.05491346e-02 1.34522152e+00 -1.16675639e+00
-5.85888386e-01 -2.01725736e-01 -5.28358936e-01 -1.13072133e+00
9.39907059e-02 -7.80921996e-01 4.94254023e-01 -1.29638243e+00
7.82280862e-01 -3.33764732e-01 -5.72438478e-01 7.13843465e-01
-6.06228232e-01 6.90176368e-01 3.54740292e-01 5.25685847e-01
-1.24772918e+00 5.13935208e-01 1.12851131e+00 -2.12267369e-01
-9.50192437e-02 -2.36230358e-01 -3.51090670e-01 9.66979027e-01
4.72671211e-01 -3.34581941e-01 -4.33086663e-01 -1.82820722e-01
-5.72486147e-02 -3.00328344e-01 7.70519674e-01 -1.22270620e+00
5.37825704e-01 -1.20970391e-01 4.36169326e-01 -6.72489941e-01
2.03982428e-01 -7.40738273e-01 -1.43700838e-01 2.42666662e-01
-2.54186839e-01 -4.88519192e-01 -7.05850124e-02 8.72760713e-01
-3.14646602e-01 -2.98962463e-02 9.18605745e-01 -3.19989145e-01
-1.14148700e+00 4.15738255e-01 6.63114339e-02 1.81147113e-01
1.10622048e+00 -7.52698556e-02 -2.79362768e-01 -7.97000155e-02
-6.05597734e-01 5.05361438e-01 3.44460845e-01 4.27528232e-01
6.58110797e-01 -9.25137341e-01 -6.40610337e-01 1.74877569e-01
8.10130239e-01 2.55471826e-01 1.11509249e-01 1.06078422e+00
-1.09534703e-01 4.57537323e-01 -9.59330797e-02 -9.01602507e-01
-1.35949481e+00 6.66504741e-01 1.32575214e-01 -2.48779729e-01
-1.00911152e+00 1.00655437e+00 8.23370039e-01 -1.06989376e-01
6.47834778e-01 1.49371149e-02 -1.26086861e-01 -2.52929330e-01
6.74093723e-01 -2.87257656e-02 -1.01251431e-01 -5.60552955e-01
-4.07432735e-01 2.28453234e-01 -3.59689415e-01 1.79865792e-01
1.12509727e+00 -9.02868062e-02 2.00394884e-01 2.16579959e-01
9.62175012e-01 -3.76753032e-01 -1.33986700e+00 -4.47492421e-01
3.33793759e-02 -5.45712352e-01 -3.26284729e-02 -8.49220276e-01
-8.11545014e-01 9.54939544e-01 6.08179390e-01 -1.81619883e-01
8.88540387e-01 4.12445068e-01 6.88069224e-01 6.36864603e-01
6.49941266e-01 -1.09943998e+00 5.05863488e-01 3.81915003e-01
9.11240935e-01 -1.81279349e+00 -1.85751721e-01 -7.39113152e-01
-9.48863089e-01 4.86398399e-01 1.13209307e+00 1.28823519e-02
2.94144452e-01 -2.30094418e-01 -1.68829307e-01 -4.33727473e-01
-6.35698140e-01 -7.72271633e-01 5.07574379e-01 6.36948168e-01
3.57112773e-02 -1.54341206e-01 -1.26889586e-01 2.13727817e-01
2.48842135e-01 -4.23563160e-02 1.69245213e-01 8.91495526e-01
-5.31769633e-01 -5.91830850e-01 -2.95309782e-01 4.00663406e-01
-1.20709956e-01 1.31027149e-02 -3.57722044e-01 7.83941805e-01
3.72992784e-01 7.10938215e-01 -4.84241635e-01 -1.81520611e-01
3.66249174e-01 -6.27618060e-02 4.83791918e-01 -8.74582529e-01
-3.78705382e-01 5.80596551e-02 -3.08475061e-03 -9.14419413e-01
-3.93964201e-01 -5.34998178e-01 -9.81635749e-01 2.87276715e-01
-5.43965697e-01 -2.47382492e-01 3.11526418e-01 9.59064543e-01
3.29338521e-01 4.39616740e-01 2.93093204e-01 -1.03055978e+00
-2.85893530e-01 -6.79130554e-01 -4.37223822e-01 9.75703418e-01
2.59523779e-01 -9.66061115e-01 -2.80424029e-01 1.67462379e-01] | [9.608762741088867, 1.3919748067855835] |
dce7d873-74e1-49d4-b517-b2b02fd48b5e | optical-flow-based-motion-detection-for | 2203.11693 | null | https://arxiv.org/abs/2203.11693v1 | https://arxiv.org/pdf/2203.11693v1.pdf | Optical Flow Based Motion Detection for Autonomous Driving | Motion detection is a fundamental but challenging task for autonomous driving. In particular scenes like highway, remote objects have to be paid extra attention for better controlling decision. Aiming at distant vehicles, we train a neural network model to classify the motion status using optical flow field information as the input. The experiments result in high accuracy, showing that our idea is viable and promising. The trained model also achieves an acceptable performance for nearby vehicles. Our work is implemented in PyTorch. Open tools including nuScenes, FastFlowNet and RAFT are used. Visualization videos are available at https://www.youtube.com/playlist?list=PLVVrWgq4OrlBnRebmkGZO1iDHEksMHKGk . | ['Ka Man Lo'] | 2022-03-03 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [-5.60985923e-01 -2.00140461e-01 -5.91976583e-01 -3.40303421e-01
-6.55296892e-02 -2.91873515e-01 4.59776729e-01 -4.44834441e-01
-6.23374581e-01 8.72998118e-01 4.09798883e-02 -6.45087659e-01
-1.01282205e-02 -8.27752233e-01 -6.05394065e-01 -7.28766561e-01
-1.16675586e-01 -7.72538632e-02 7.91995168e-01 -2.80797809e-01
4.71204907e-01 6.03804886e-01 -1.77190244e+00 -1.09381109e-01
6.38594091e-01 8.87200356e-01 5.09948909e-01 9.38042521e-01
5.28603978e-03 1.17475307e+00 -2.86723405e-01 -1.20138757e-01
2.29674518e-01 1.40341595e-02 -6.94213152e-01 -4.33505028e-01
4.76756811e-01 -6.89099967e-01 -6.91848338e-01 9.86509025e-01
4.75789279e-01 4.23182577e-01 4.54799920e-01 -1.42182863e+00
6.51993603e-02 6.28662482e-02 -3.57422471e-01 8.07976902e-01
3.59930061e-02 5.54656446e-01 5.79802275e-01 -8.64020050e-01
6.35262251e-01 1.14009762e+00 1.39783442e-01 4.91835445e-01
-5.90768278e-01 -9.55479860e-01 1.09259963e-01 1.16500056e+00
-1.17902136e+00 -6.63462996e-01 5.37866235e-01 -4.72310424e-01
7.44477689e-01 3.56021494e-01 6.00983858e-01 1.04954565e+00
4.07404572e-01 8.71505022e-01 6.62499249e-01 2.47939065e-01
1.82759404e-01 1.65215470e-02 2.44724721e-01 6.38812840e-01
1.94145143e-01 7.95839950e-02 -3.98476481e-01 4.21340853e-01
8.31998944e-01 7.20145181e-02 -4.69137549e-01 -8.36313143e-02
-1.21206248e+00 9.59447443e-01 5.04252732e-01 1.32638812e-01
-1.85967892e-01 3.82804632e-01 3.05423260e-01 1.83201894e-01
2.57212073e-01 3.50462855e-04 -6.63390532e-02 -4.56644893e-01
-6.62056804e-01 3.87376726e-01 3.93644422e-01 8.94762754e-01
8.94495964e-01 2.04124242e-01 -1.76878661e-01 4.64433372e-01
3.34267110e-01 5.95100462e-01 2.12075084e-01 -1.79499805e+00
3.56253535e-01 3.26149285e-01 2.50787854e-01 -1.19672191e+00
-3.37096542e-01 -9.89132896e-02 -5.40725529e-01 8.03910971e-01
3.40141833e-01 -1.68914020e-01 -6.58960998e-01 1.11378467e+00
3.98904741e-01 4.67026263e-01 -5.64451329e-03 1.29744804e+00
1.31990755e+00 1.00377727e+00 1.84198901e-01 8.42459947e-02
9.97515321e-01 -1.21210480e+00 -9.72617865e-01 -6.60021082e-02
7.29414880e-01 -6.65828824e-01 7.80255795e-01 3.06476325e-01
-9.94866669e-01 -7.34310865e-01 -8.17303240e-01 -1.42461061e-01
-3.95021230e-01 8.78301263e-02 4.41238105e-01 2.96484768e-01
-1.12285841e+00 5.46985805e-01 -8.42859387e-01 -3.04972887e-01
5.73123991e-01 2.17322811e-01 -2.64401644e-01 -1.20989084e-01
-1.34456491e+00 7.73860335e-01 5.60697675e-01 3.67855132e-01
-1.10650122e+00 -4.12517220e-01 -8.23742092e-01 -3.07100385e-01
3.94609541e-01 -3.52526844e-01 1.37929726e+00 -4.66417998e-01
-1.51360285e+00 4.72942829e-01 -3.40143323e-01 -4.61737037e-01
8.89408767e-01 -2.23898143e-01 -5.98753810e-01 4.00973141e-01
1.52864441e-01 9.55897331e-01 5.27411044e-01 -8.95979762e-01
-1.01720989e+00 -8.95659105e-05 1.61068633e-01 2.38764197e-01
-1.33573219e-01 2.50526041e-01 -5.03814459e-01 -1.87664986e-01
-2.74149656e-01 -6.28588438e-01 -2.57118165e-01 2.42921710e-01
-2.23563552e-01 -3.21140766e-01 1.29726279e+00 -5.23904741e-01
1.26843143e+00 -1.90041816e+00 -4.16137487e-01 -6.18589483e-02
3.58521640e-01 7.08704352e-01 -2.80259941e-02 3.89952540e-01
2.41564453e-01 1.14457179e-02 3.26869547e-01 2.86304027e-01
-3.28092337e-01 1.44114405e-01 3.40802707e-02 5.48749447e-01
6.48837313e-02 1.01131368e+00 -9.25191939e-01 -6.11953616e-01
7.57595062e-01 3.37339282e-01 -3.13637257e-01 2.92763203e-01
-5.04264161e-02 6.73405707e-01 -7.23185122e-01 4.45626289e-01
7.39521086e-01 1.54256728e-02 -4.82244492e-01 -4.76099364e-02
-6.62572205e-01 8.47492740e-02 -1.18081653e+00 1.41744471e+00
-6.20501637e-02 1.12818372e+00 5.21583557e-02 -8.12171578e-01
8.44351649e-01 1.53232679e-01 4.06328857e-01 -7.40478098e-01
4.78929698e-01 7.82559812e-02 1.04271583e-02 -1.09436631e+00
6.30508840e-01 4.53303367e-01 2.54350692e-01 -7.82521516e-02
-3.46418530e-01 1.82326496e-01 4.90369260e-01 3.16480935e-01
1.02813137e+00 2.36245841e-01 7.04302639e-02 -1.26874864e-01
7.39329994e-01 2.33124897e-01 6.30883753e-01 5.14990926e-01
-7.35074222e-01 3.88414472e-01 3.42999488e-01 -5.07354319e-01
-8.09845388e-01 -7.21903741e-01 -2.54625499e-01 8.46568286e-01
6.57368362e-01 -2.28918791e-01 -5.81788480e-01 -4.39316899e-01
-1.90403059e-01 6.07817411e-01 -3.27920854e-01 -2.71565299e-02
-7.11616576e-01 -1.40258253e-01 2.50889599e-01 6.24907076e-01
6.26384437e-01 -1.41412330e+00 -9.28686798e-01 1.00475937e-01
-1.88181981e-01 -1.32797658e+00 -1.67603180e-01 -3.48034322e-01
-6.53702855e-01 -8.38820517e-01 -7.60960281e-01 -7.12239087e-01
3.57329190e-01 7.96542525e-01 7.65690744e-01 2.62162268e-01
-3.25792015e-01 1.95458271e-02 -2.78791577e-01 -2.57159650e-01
-1.68360144e-01 6.30048364e-02 -1.11305453e-02 -2.85067596e-02
5.47695458e-01 -2.14534685e-01 -1.04168653e+00 5.85772693e-01
-5.61754107e-01 2.94755816e-01 2.28382081e-01 3.40069205e-01
4.03477959e-02 -1.46145329e-01 2.93130517e-01 -3.14784169e-01
1.33292377e-01 -5.70160389e-01 -8.07302177e-01 -4.61733401e-01
-1.46905243e-01 -3.21254492e-01 5.17098248e-01 -1.66921347e-01
-9.45646405e-01 -8.52987617e-02 -5.44843912e-01 -6.61207616e-01
-6.81138575e-01 1.15010180e-01 -1.48893306e-02 -8.02677646e-02
4.13788140e-01 -1.18951574e-01 2.60106027e-01 -1.45498142e-01
1.82267994e-01 7.47125983e-01 3.59164685e-01 2.21361026e-01
5.22843897e-01 7.96974242e-01 -3.03242236e-01 -9.80199873e-01
-3.07114810e-01 -6.00161374e-01 -5.50670087e-01 -9.09475982e-01
1.07896388e+00 -1.08286262e+00 -1.03128088e+00 3.37248474e-01
-1.07684135e+00 -4.89124179e-01 1.38461560e-01 9.00681436e-01
-4.30312723e-01 2.81336606e-01 -5.61814249e-01 -6.11352742e-01
-1.66004732e-01 -1.04475689e+00 6.10179424e-01 5.94669342e-01
1.56505942e-01 -9.95284259e-01 3.97499092e-02 4.00779307e-01
5.98636568e-01 2.02706873e-01 5.16896546e-02 -4.09280926e-01
-1.00350916e+00 4.11838144e-02 -4.69747752e-01 1.16064779e-01
-2.97060069e-02 5.10821640e-01 -8.34603429e-01 -7.34616071e-02
-4.30262387e-01 -9.05934051e-02 1.09144866e+00 6.95914209e-01
1.12598848e+00 -2.79145122e-01 -4.22365189e-01 6.64695382e-01
1.23678231e+00 4.23139721e-01 8.24471951e-01 5.76699972e-01
8.53415191e-01 6.91708446e-01 8.58856201e-01 4.12632465e-01
5.38127661e-01 5.39364219e-01 8.24039340e-01 -1.59502760e-01
-2.41829574e-01 1.55113682e-01 5.29512048e-01 5.68019927e-01
-2.47070685e-01 -3.96901637e-01 -9.80440617e-01 5.37952483e-01
-2.04856157e+00 -1.26417732e+00 -7.52214611e-01 1.75464010e+00
1.62726507e-01 3.52936000e-01 3.61285433e-02 1.82115026e-02
6.77157223e-01 3.65791976e-01 -4.77756172e-01 -5.31715870e-01
9.88278314e-02 -1.89150885e-01 7.32831061e-01 5.52428424e-01
-1.08078468e+00 1.08534563e+00 5.15089893e+00 7.49314904e-01
-1.46075141e+00 5.48567176e-02 4.99959171e-01 -3.34639400e-01
3.63833681e-02 -2.17449088e-02 -9.85805511e-01 7.09371328e-01
1.13314152e+00 4.87436652e-02 7.18225092e-02 6.29337549e-01
1.08745217e+00 -5.49197555e-01 -3.30487520e-01 8.87451768e-01
-2.50731468e-01 -1.54773831e+00 -3.32998544e-01 -9.58118215e-03
3.76234174e-01 5.00366628e-01 -4.06273566e-02 2.03561738e-01
2.32704684e-01 -8.23985696e-01 6.34621799e-01 4.88524586e-01
5.20731330e-01 -6.60149872e-01 6.72446370e-01 4.35810953e-01
-1.34036231e+00 -5.90757467e-02 -5.39735794e-01 -2.31403321e-01
4.37136441e-01 3.40737432e-01 -8.48320246e-01 4.25538599e-01
1.00905001e+00 1.12697339e+00 -4.67954040e-01 1.52668333e+00
-3.44574660e-01 5.22695422e-01 -5.96500896e-02 -2.96206117e-01
4.14377689e-01 -2.32369468e-01 6.94975674e-01 1.11233914e+00
3.97977918e-01 8.25403035e-02 1.14301123e-01 6.13369107e-01
3.05388957e-01 1.18838638e-01 -7.30473340e-01 3.29916358e-01
3.01610351e-01 1.64468169e+00 -6.28803551e-01 -4.61684585e-01
-4.76793587e-01 4.65490609e-01 4.01116647e-02 4.87064451e-01
-1.18276823e+00 -5.03323734e-01 1.10841548e+00 3.31421852e-01
4.94254380e-01 -4.15547997e-01 2.98287541e-01 -1.07281256e+00
-1.57143742e-01 -3.34267110e-01 9.98114198e-02 -1.05697143e+00
-4.09204483e-01 6.04793489e-01 6.65196031e-02 -1.49503303e+00
-2.04710960e-01 -7.84580290e-01 -9.10342038e-01 5.83050728e-01
-1.94790256e+00 -5.99732339e-01 -7.26169586e-01 7.91069865e-01
7.13464856e-01 -4.57657948e-02 2.17980281e-01 5.72067797e-01
-8.50101650e-01 1.81139074e-02 9.44332406e-03 2.07921579e-01
5.98593414e-01 -8.46901476e-01 3.96757782e-01 9.12639737e-01
-4.36887890e-01 2.19724029e-02 7.53731012e-01 -5.56445301e-01
-1.17812884e+00 -1.36257482e+00 5.91909945e-01 -2.25470051e-01
7.37111032e-01 -1.17243223e-01 -1.02426302e+00 3.92318666e-01
5.51489353e-01 3.00399691e-01 1.53745413e-01 -5.97847641e-01
2.21629575e-01 -1.01125099e-01 -7.89168239e-01 6.63020611e-01
1.06080580e+00 -1.65516436e-02 -1.61396697e-01 9.87874568e-02
4.91936386e-01 -4.39961851e-01 -6.24320924e-01 3.97341311e-01
4.91375804e-01 -1.19648635e+00 6.56661570e-01 -4.69529003e-01
4.20663625e-01 -5.76243818e-01 2.12005273e-01 -9.46183622e-01
-2.46756464e-01 -4.86647666e-01 -1.25263408e-01 7.84767330e-01
2.42531627e-01 -6.61643207e-01 9.70843375e-01 3.77088517e-01
-1.07305288e-01 -5.50217271e-01 -8.46626222e-01 -6.75503612e-01
-1.74171284e-01 -6.31064951e-01 2.07376525e-01 7.52229452e-01
-2.34778658e-01 1.24292485e-01 -4.33552474e-01 2.23915443e-01
5.00504613e-01 -1.15249634e-01 9.76452112e-01 -1.05351293e+00
2.80240715e-01 -5.23320735e-01 -7.93596327e-01 -1.12425721e+00
7.71781728e-02 -6.16518438e-01 -1.56103477e-01 -1.69131947e+00
-1.14593059e-01 -2.46193334e-01 -1.83805421e-01 3.91036898e-01
1.73645411e-02 3.83725792e-01 2.24702224e-01 1.29398763e-01
-7.71725893e-01 5.30424237e-01 1.44253683e+00 -3.69737707e-02
-2.98603345e-02 1.87976897e-01 -2.80342758e-01 7.10437238e-01
1.43893337e+00 -2.97256321e-01 -3.27687860e-01 -4.55484092e-01
-6.60980940e-02 1.44904971e-01 7.48617530e-01 -1.16406775e+00
3.95455956e-01 -3.99353385e-01 2.50911772e-01 -8.59616518e-01
4.49609905e-01 -6.65376723e-01 5.26561290e-02 6.12813592e-01
-1.82534859e-01 2.71361738e-01 1.91922441e-01 3.16488296e-01
-2.02494919e-01 -7.56229311e-02 5.16046464e-01 -3.98218408e-02
-1.50198340e+00 5.19428849e-01 -7.97923326e-01 -9.13863555e-02
1.28289270e+00 -6.45619690e-01 -5.11400104e-01 -5.52303433e-01
-6.28676414e-01 7.89921343e-01 2.23128930e-01 8.48736942e-01
7.59388566e-01 -1.29534090e+00 -7.82676160e-01 1.54311642e-01
1.29441053e-01 -9.76075977e-02 5.37992716e-01 9.65938210e-01
-7.95868218e-01 5.62056184e-01 -5.46656609e-01 -9.06543016e-01
-1.34167683e+00 3.26500446e-01 3.29598129e-01 2.75562048e-01
-7.43109584e-01 7.21040130e-01 1.73725396e-01 -1.55426830e-01
1.77071124e-01 -4.34093088e-01 -8.61643791e-01 -4.59269062e-02
6.38603568e-01 8.15791428e-01 6.67851865e-02 -9.17205036e-01
-2.97518760e-01 4.88061607e-01 1.48051046e-02 1.50453910e-01
1.09915352e+00 -4.05039907e-01 2.16852725e-01 1.27704546e-01
1.42209125e+00 -1.96293190e-01 -1.64748251e+00 1.90407395e-01
-1.92744032e-01 -7.83465624e-01 1.71251640e-01 -2.23705411e-01
-1.23894751e+00 9.75506127e-01 9.09609318e-01 8.78718719e-02
8.53155792e-01 -6.23840056e-02 7.30808437e-01 1.96769461e-01
4.99767035e-01 -8.72866690e-01 -9.12773237e-02 5.83435178e-01
6.10549688e-01 -1.57516360e+00 -8.47659260e-02 -3.76845986e-01
-6.84676051e-01 1.10351372e+00 7.78793156e-01 -1.66459829e-01
6.44832432e-01 -1.14251804e-02 4.39340472e-01 -1.15981683e-01
-8.05154383e-01 -4.80753541e-01 6.57312497e-02 3.20994318e-01
1.45313278e-01 -1.03207760e-01 -4.09632772e-01 -1.96452677e-01
-1.09081917e-01 4.84011509e-03 7.62877762e-01 8.72163177e-01
-7.19464540e-01 -7.65859544e-01 -2.98160136e-01 3.47437292e-01
-4.31005329e-01 1.65949345e-01 2.98905671e-01 7.57525802e-01
9.87020284e-02 1.10726142e+00 1.99918464e-01 -2.76683092e-01
2.75138378e-01 -4.38954920e-01 -1.75911516e-01 -1.83790907e-01
-2.15267986e-01 8.76051709e-02 4.89468724e-02 -1.06972945e+00
-7.00208068e-01 -4.91567522e-01 -1.50501204e+00 -5.11855900e-01
-7.78325722e-02 1.25147894e-01 6.49626493e-01 4.94652152e-01
5.28213680e-01 4.94408786e-01 7.24087059e-01 -1.04965854e+00
2.37842381e-01 -7.00504243e-01 -3.55724663e-01 -4.30087931e-02
6.08992279e-01 -7.80172527e-01 -4.58535284e-01 -2.18075700e-02] | [8.097597122192383, -1.3552002906799316] |
4b893367-31ce-49bd-aa61-1f4150585afa | mandoline-model-evaluation-under-distribution | 2107.00643 | null | https://arxiv.org/abs/2107.00643v2 | https://arxiv.org/pdf/2107.00643v2.pdf | Mandoline: Model Evaluation under Distribution Shift | Machine learning models are often deployed in different settings than they were trained and validated on, posing a challenge to practitioners who wish to predict how well the deployed model will perform on a target distribution. If an unlabeled sample from the target distribution is available, along with a labeled sample from a possibly different source distribution, standard approaches such as importance weighting can be applied to estimate performance on the target. However, importance weighting struggles when the source and target distributions have non-overlapping support or are high-dimensional. Taking inspiration from fields such as epidemiology and polling, we develop Mandoline, a new evaluation framework that mitigates these issues. Our key insight is that practitioners may have prior knowledge about the ways in which the distribution shifts, which we can use to better guide the importance weighting procedure. Specifically, users write simple "slicing functions" - noisy, potentially correlated binary functions intended to capture possible axes of distribution shift - to compute reweighted performance estimates. We further describe a density ratio estimation framework for the slices and show how its estimation error scales with slice quality and dataset size. Empirical validation on NLP and vision tasks shows that Mandoline can estimate performance on the target distribution up to 3x more accurately compared to standard baselines. | ['Nimit S. Sohoni', 'Christopher Ré', 'Kayvon Fatahalian', 'Fait Poms', 'Karan Goel', 'Mayee Chen'] | 2021-07-01 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 2.27321431e-01 -9.27657783e-02 -5.98613679e-01 -6.14103019e-01
-1.12934232e+00 -8.29082191e-01 5.89256525e-01 5.09670854e-01
-4.81726527e-01 8.22835207e-01 4.17853177e-01 -5.42148113e-01
-2.69789994e-01 -5.89713931e-01 -6.40203774e-01 -7.38470614e-01
1.29189149e-01 7.02739179e-01 -9.57196951e-03 3.07003230e-01
3.53273749e-01 3.67814779e-01 -1.12749660e+00 7.58020580e-02
6.31784678e-01 8.19453597e-01 -1.80755943e-01 6.82987750e-01
-1.33439675e-01 5.82218289e-01 -1.13366699e+00 -4.12002921e-01
1.94963202e-01 -3.46590839e-02 -5.52897096e-01 -1.86308213e-02
6.14624619e-01 -4.07086968e-01 1.00867555e-01 9.78505909e-01
3.32591712e-01 -2.12645397e-01 1.25219297e+00 -1.39302397e+00
-4.08500642e-01 7.49201655e-01 -9.32186782e-01 4.52577591e-01
-7.26349503e-02 3.00076544e-01 8.93518627e-01 -4.80859548e-01
2.17439801e-01 1.07835340e+00 9.74646151e-01 2.05331117e-01
-1.51581836e+00 -6.32012129e-01 2.64302284e-01 9.32718664e-02
-1.27113080e+00 -3.35527480e-01 1.70951143e-01 -6.29979014e-01
4.92769748e-01 3.32110792e-01 3.02821755e-01 1.21058619e+00
7.50081763e-02 5.56032956e-01 9.97620821e-01 -1.34667218e-01
4.13047969e-01 4.88352835e-01 1.21412702e-01 1.68908581e-01
2.74530977e-01 1.17179945e-01 -5.19678652e-01 -6.28557086e-01
3.76947075e-01 -2.55432469e-03 -3.64330441e-01 -3.60528737e-01
-9.79441285e-01 8.99649441e-01 4.04900610e-01 -2.68758126e-02
-3.79344761e-01 3.30278397e-01 2.24474132e-01 -1.07252918e-01
7.12627053e-01 5.67855775e-01 -5.63674927e-01 -2.77286828e-01
-1.16632795e+00 5.04197538e-01 7.18071520e-01 7.41114736e-01
5.01920223e-01 -3.07541072e-01 -6.33426428e-01 8.29742968e-01
2.20679924e-01 6.95877135e-01 8.74929652e-02 -8.60927343e-01
4.59566236e-01 3.44237268e-01 4.77124125e-01 -6.59435451e-01
-2.72883713e-01 -4.86868292e-01 -5.75843096e-01 2.03858644e-01
9.65867341e-01 -3.29958826e-01 -1.02355850e+00 1.78810251e+00
4.48000193e-01 1.40642956e-01 -3.27099323e-01 5.83823204e-01
2.38124773e-01 5.24477363e-01 2.95198411e-01 -4.01234366e-02
1.36099041e+00 -4.40157950e-01 -2.97855109e-01 -2.21068844e-01
3.28029722e-01 -7.35864222e-01 1.13388014e+00 4.36662614e-01
-8.31186414e-01 -1.58544391e-01 -8.79524946e-01 4.13009912e-01
-1.76890358e-01 -1.74300700e-01 4.90383029e-01 8.39235723e-01
-8.31603467e-01 7.03133225e-01 -6.37677312e-01 -1.73043251e-01
8.09909225e-01 3.09558082e-02 8.55381340e-02 -1.55046731e-01
-7.96181202e-01 9.55442727e-01 6.99814558e-02 -3.75372857e-01
-8.75248849e-01 -1.37538099e+00 -6.05056643e-01 2.00097233e-01
2.97458261e-01 -5.74270785e-01 1.56149781e+00 -7.27948070e-01
-7.55311728e-01 4.56789643e-01 2.27969680e-02 -5.59828162e-01
6.18901551e-01 -1.64477289e-01 -1.67073011e-01 -2.02386364e-01
2.57091612e-01 4.92505968e-01 9.63930786e-01 -9.36636567e-01
-9.24119294e-01 -5.29110968e-01 -2.36991085e-02 2.39354089e-01
-2.71993369e-01 -1.05684258e-01 -4.05436575e-01 -5.72592676e-01
-4.75665182e-01 -6.30455971e-01 -2.40415096e-01 8.15749243e-02
-5.22169232e-01 5.49441986e-02 6.30028307e-01 -3.69643509e-01
1.45327914e+00 -1.87648273e+00 -3.27826738e-01 4.66969073e-01
5.02045631e-01 1.13443062e-01 1.01403352e-02 1.35412186e-01
8.90441164e-02 8.60669762e-02 -3.26942563e-01 3.46086696e-02
-2.73011215e-02 -4.80413251e-02 -3.39386344e-01 6.04179859e-01
1.71988338e-01 5.20390868e-01 -1.03501916e+00 -2.47438028e-01
1.32518440e-01 5.54241002e-01 -4.20817971e-01 8.92895013e-02
-2.98201054e-01 2.23041371e-01 -9.12123397e-02 6.30251348e-01
7.16568172e-01 -5.44909358e-01 9.52131525e-02 -2.66291499e-01
2.22774610e-01 1.69487730e-01 -1.03915071e+00 9.42038298e-01
-4.20663744e-01 6.99598730e-01 1.62510984e-02 -8.61121476e-01
6.16825700e-01 -3.00267041e-01 3.86365235e-01 -3.02972108e-01
1.25998957e-02 -9.33090895e-02 5.10877967e-02 -2.35507220e-01
3.72990012e-01 -2.51788139e-01 4.15783711e-02 5.73184967e-01
-2.29613230e-01 -6.77090138e-02 -4.51177210e-02 -6.35393038e-02
1.26753283e+00 -1.98294148e-01 3.66947234e-01 -2.87756115e-01
-1.74752206e-01 1.53111741e-01 1.98819697e-01 9.39921737e-01
-2.20831171e-01 6.62618876e-01 8.11754823e-01 -1.67067498e-01
-1.16092741e+00 -1.51966596e+00 -5.50254226e-01 1.18564796e+00
-4.64062095e-02 -2.04418927e-01 -6.16211772e-01 -1.01610982e+00
3.86167347e-01 8.98568094e-01 -7.90449023e-01 -3.99373844e-02
-3.08985244e-02 -1.21703494e+00 5.42539835e-01 6.96023464e-01
6.76080883e-02 -4.97757763e-01 -7.47194231e-01 1.98972613e-01
2.44455263e-02 -7.22780406e-01 -6.83241665e-01 2.63267815e-01
-6.32325470e-01 -1.34485674e+00 -1.00164640e+00 -6.43406883e-02
5.84745705e-01 7.56531879e-02 1.48203933e+00 -2.80447841e-01
-3.66492242e-01 5.46280324e-01 -1.04547113e-01 -9.42009628e-01
-4.01559025e-01 -2.90315654e-02 -1.33608088e-01 -2.07868725e-01
7.05746174e-01 -3.21001977e-01 -8.32721770e-01 3.25980753e-01
-7.81316817e-01 -3.04709554e-01 4.91052538e-01 6.06597722e-01
3.24893028e-01 2.28499267e-02 6.85130417e-01 -1.23597252e+00
9.37175155e-01 -9.88036275e-01 -5.78153193e-01 3.46326590e-01
-8.23015749e-01 2.52376854e-01 2.40216732e-01 -7.14079440e-01
-7.84775078e-01 -3.40006590e-01 3.06608193e-02 -4.99022335e-01
-1.17733754e-01 4.07483399e-01 1.71422482e-01 6.07370555e-01
1.19285834e+00 -1.86799511e-01 6.03146777e-02 -3.54420990e-01
3.69510829e-01 8.41357470e-01 5.60760319e-01 -4.97788638e-01
6.01482511e-01 3.31953108e-01 -2.04063445e-01 -5.37664473e-01
-1.03392482e+00 -4.98424739e-01 -1.49617046e-01 3.85503024e-02
3.71494889e-01 -7.70264626e-01 -4.57133710e-01 3.55246305e-01
-8.29786777e-01 -6.32955492e-01 -4.02139783e-01 4.76637393e-01
-1.51413992e-01 -3.72638926e-02 -1.20950475e-01 -9.96870518e-01
-3.26304495e-01 -1.01717067e+00 1.12642682e+00 1.97645560e-01
-5.58393240e-01 -1.17413425e+00 1.06590763e-01 7.20955729e-02
7.37508178e-01 1.42845482e-01 1.20072424e+00 -8.51536989e-01
-1.56367391e-01 -1.71152383e-01 -4.21885848e-01 3.20212573e-01
3.45456868e-01 1.90045252e-01 -1.05315089e+00 -2.04218701e-01
-2.05676347e-01 -1.11428760e-01 6.66920006e-01 9.19814646e-01
1.38873339e+00 -3.78775835e-01 -4.59351003e-01 3.00710410e-01
1.28972089e+00 -6.27805665e-02 3.47890407e-01 9.13902074e-02
4.84353989e-01 7.07812548e-01 5.99799097e-01 6.11215353e-01
4.85733300e-01 4.44008172e-01 2.42127523e-01 -4.87069860e-02
1.26815513e-01 -1.96694806e-01 6.50295690e-02 -1.64922848e-01
5.14076710e-01 -2.22120181e-01 -1.09861672e+00 6.60870016e-01
-1.57118535e+00 -8.96900177e-01 1.30937785e-01 2.64176774e+00
1.09048176e+00 3.90816867e-01 5.83403587e-01 -2.09840670e-01
7.54475176e-01 5.75478934e-02 -1.16362762e+00 -2.36400351e-01
3.22981805e-01 1.12441294e-01 9.16580915e-01 5.27174234e-01
-1.00364470e+00 3.70211303e-01 7.16728735e+00 1.07348824e+00
-1.21781540e+00 -1.22515716e-01 1.23780394e+00 -4.44715321e-01
-4.28076565e-01 -2.30965465e-01 -9.07016695e-01 8.46957564e-01
1.12371695e+00 -2.71411955e-01 1.40287593e-01 8.37714374e-01
6.41458258e-02 -4.53760505e-01 -1.40422428e+00 1.13077664e+00
-1.66521162e-01 -1.40132356e+00 -2.10344538e-01 2.95537323e-01
5.49426734e-01 2.54004478e-01 2.89149076e-01 4.30745870e-01
6.92993343e-01 -1.30060840e+00 5.20575345e-01 4.43885446e-01
9.96800959e-01 -7.40626931e-01 6.64014161e-01 4.56599414e-01
-7.58865654e-01 2.18043458e-02 -3.29313159e-01 2.67420113e-01
-9.32898149e-02 9.59021866e-01 -1.44469285e+00 -6.81615025e-02
7.21718371e-01 2.26521596e-01 -6.37238801e-01 1.19869196e+00
6.69629648e-02 8.39366257e-01 -5.96968234e-01 -1.64626744e-02
5.52993678e-02 1.01085588e-01 3.11468482e-01 1.21103895e+00
3.59962970e-01 -2.85001993e-01 -1.79211691e-01 9.43044305e-01
-1.24678195e-01 -1.66379094e-01 -5.52064478e-01 -3.07925716e-02
8.95367622e-01 1.17236960e+00 -6.65569365e-01 -4.26034033e-01
-2.92205065e-01 4.51867521e-01 2.45613560e-01 3.66258383e-01
-1.03958452e+00 -1.38929278e-01 8.09687018e-01 5.01695573e-01
2.15185523e-01 2.88857222e-01 -3.79272521e-01 -8.01615536e-01
-1.39178425e-01 -9.55141962e-01 4.44367766e-01 -5.47095299e-01
-1.69491863e+00 2.89710253e-01 5.62425613e-01 -1.02766621e+00
-4.18264627e-01 -4.48728800e-01 -5.40002525e-01 1.13232970e+00
-1.31466985e+00 -7.54578471e-01 -1.70897052e-01 2.74151802e-01
4.26510751e-01 -1.35289699e-01 5.53263485e-01 9.35973525e-02
-2.33324125e-01 7.83301115e-01 1.63324043e-01 -1.22704692e-01
8.47154975e-01 -1.58781230e+00 4.01042074e-01 4.74301130e-01
2.12975562e-01 3.56982917e-01 9.03395474e-01 -5.65784276e-01
-8.47719789e-01 -1.10586047e+00 3.75061005e-01 -8.39341521e-01
6.14428818e-01 -5.97942136e-02 -9.43302870e-01 5.20575762e-01
-8.72553792e-03 -1.22859702e-01 1.01738799e+00 4.13258821e-01
-6.16732538e-01 -1.85037568e-01 -1.46146321e+00 4.99986947e-01
6.12918079e-01 -2.77624786e-01 -3.23771894e-01 3.11992109e-01
2.39636093e-01 -3.94493759e-01 -7.36795962e-01 4.38522398e-01
6.33857310e-01 -1.00451767e+00 9.35967743e-01 -6.77982211e-01
4.46387738e-01 -1.07728228e-01 -3.30844373e-01 -1.63860965e+00
-1.34798750e-01 -1.86620995e-01 -2.03317389e-01 1.22774792e+00
5.81872582e-01 -4.91480559e-01 7.98506260e-01 1.04520726e+00
4.83508617e-01 -9.79775488e-01 -5.63447595e-01 -4.97349352e-01
1.98658377e-01 -6.92922711e-01 7.91677833e-01 7.98922539e-01
-3.68037820e-01 5.91257751e-01 -1.51060343e-01 1.73808768e-01
7.83535898e-01 -1.77035749e-01 8.03388596e-01 -1.25760853e+00
-4.32133943e-01 -6.56023860e-01 -6.46188185e-02 -9.30410385e-01
-2.66266018e-01 -5.16849041e-01 9.83126927e-03 -1.33108699e+00
3.12105775e-01 -7.34259307e-01 -2.68714398e-01 4.86388445e-01
-4.47765976e-01 1.60030529e-01 4.59810607e-02 9.49054211e-02
-4.23692793e-01 2.22402051e-01 7.31860936e-01 -8.88938159e-02
-1.60421386e-01 2.63659269e-01 -1.20588648e+00 8.07498813e-01
7.84117639e-01 -5.28932989e-01 -7.53103137e-01 -3.47692102e-01
1.47777006e-01 -9.91860181e-02 3.31874698e-01 -8.83603692e-01
-1.21025462e-02 -3.84819984e-01 7.75756299e-01 -6.99789107e-01
6.17486276e-02 -7.62888610e-01 6.66689947e-02 1.03081942e-01
-5.82443416e-01 7.51471892e-02 3.09354663e-01 6.55854881e-01
2.06445649e-01 -3.87205124e-01 9.12041843e-01 2.91296542e-02
-1.95492044e-01 3.79452974e-01 -1.97082952e-01 5.10084689e-01
9.46607113e-01 -8.01488198e-03 -4.94279861e-01 -5.37708521e-01
-3.92879814e-01 4.16220874e-01 4.05338973e-01 2.65366763e-01
3.50562781e-01 -1.27508807e+00 -9.05021608e-01 -1.35394372e-03
2.83483952e-01 2.27806042e-03 1.90167993e-01 8.17595780e-01
-2.93909699e-01 -5.06894030e-02 1.15907870e-01 -8.04906726e-01
-1.02222586e+00 3.09857726e-01 3.70656401e-01 -3.63535851e-01
-2.23461807e-01 8.92060757e-01 3.74966860e-01 -4.20009106e-01
5.13555169e-01 -3.83600563e-01 1.47747630e-02 4.10593390e-01
1.04190087e+00 3.94399822e-01 5.13041690e-02 -9.75552574e-02
-4.11555588e-01 1.63714796e-01 -2.28604570e-01 -1.84651107e-01
1.27543914e+00 -1.35146424e-01 3.67792398e-01 4.45561528e-01
1.14513969e+00 4.05931734e-02 -1.60192013e+00 -3.90707552e-01
4.78220843e-02 -6.86239064e-01 -2.91928444e-02 -1.15181696e+00
-8.18702340e-01 8.42938244e-01 7.21935987e-01 3.63873720e-01
9.52761352e-01 1.25099495e-01 3.74614596e-01 -4.02882807e-02
1.15838304e-01 -9.97823298e-01 -3.43605429e-02 4.21888418e-02
7.89462626e-01 -1.35476911e+00 1.25527278e-01 1.70792729e-01
-7.59820700e-01 8.33962977e-01 5.16501904e-01 1.14344835e-01
7.93949604e-01 5.53806722e-01 9.35573950e-02 -1.61535278e-01
-7.69505918e-01 -8.37033615e-03 3.83263886e-01 8.89038265e-01
2.21194044e-01 2.49422401e-01 2.93737948e-01 2.94271797e-01
-2.96524525e-01 -1.48851827e-01 4.15800720e-01 5.22541106e-01
-3.57694238e-01 -9.86585557e-01 -7.07886875e-01 1.16461694e+00
-6.46511793e-01 -2.25651532e-01 -1.71591595e-01 6.88835800e-01
3.46280597e-02 7.16859400e-01 5.42645752e-01 -1.80119768e-01
3.05878371e-01 1.31250937e-02 4.67621416e-01 -8.44277978e-01
-3.91691595e-01 2.08480999e-01 7.50061274e-02 -3.29362005e-01
-1.65459514e-01 -6.24910116e-01 -8.61887991e-01 -4.97386217e-01
-7.92426094e-02 -2.36796290e-02 6.56535447e-01 7.73259640e-01
2.37142667e-01 3.25528622e-01 4.10237670e-01 -6.54394448e-01
-9.88540113e-01 -1.03859460e+00 -5.75384915e-01 3.90540689e-01
6.10194802e-01 -6.38652325e-01 -4.52690631e-01 -1.97750419e-01] | [8.58965015411377, 4.547660827636719] |
323666c3-253b-4d9d-a877-d1081d358a00 | a-survey-on-deep-learning-approaches-for-data | 2306.11740 | null | https://arxiv.org/abs/2306.11740v1 | https://arxiv.org/pdf/2306.11740v1.pdf | A survey on deep learning approaches for data integration in autonomous driving system | The perception module of self-driving vehicles relies on a multi-sensor system to understand its environment. Recent advancements in deep learning have led to the rapid development of approaches that integrate multi-sensory measurements to enhance perception capabilities. This paper surveys the latest deep learning integration techniques applied to the perception module in autonomous driving systems, categorizing integration approaches based on "what, how, and when to integrate." A new taxonomy of integration is proposed, based on three dimensions: multi-view, multi-modality, and multi-frame. The integration operations and their pros and cons are summarized, providing new insights into the properties of an "ideal" data integration approach that can alleviate the limitations of existing methods. After reviewing hundreds of relevant papers, this survey concludes with a discussion of the key features of an optimal data integration approach. | ['Lei Chen', 'Yue Gong', 'Xiya Cao', 'Caifa Zhou', 'Likang Wang', 'Xi Zhu'] | 2023-06-17 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-1.64963201e-01 -1.96650252e-01 -4.99735266e-01 -5.76237321e-01
-5.31392992e-01 -3.82474869e-01 5.18771768e-01 1.19034043e-02
-4.05126572e-01 3.53377968e-01 -2.11371686e-02 -1.24790251e-01
-1.10175267e-01 -9.13284183e-01 -5.92254102e-01 -6.97727263e-01
4.30965871e-01 -8.92349612e-03 3.83672565e-01 -6.62742734e-01
1.48644045e-01 4.41924959e-01 -2.52230978e+00 3.39283139e-01
4.79158700e-01 1.31583309e+00 3.61663520e-01 5.95630467e-01
2.13304674e-03 6.44212842e-01 -4.53177869e-01 -5.03394604e-02
3.20370138e-01 8.54005218e-02 -1.64305031e-01 -1.78396665e-02
5.55520117e-01 -2.98187137e-01 -2.60492712e-01 1.02630758e+00
4.75106061e-01 1.38393501e-02 3.46977711e-01 -1.57367635e+00
-3.26992184e-01 -1.12837695e-01 -1.77079961e-01 3.45037818e-01
2.82726228e-01 2.24815756e-02 7.75140107e-01 -7.17386186e-01
3.91222924e-01 1.23989224e+00 7.83712626e-01 3.33383560e-01
-7.90873647e-01 -4.96320426e-01 1.65727541e-01 9.52474058e-01
-1.02443159e+00 -6.07485414e-01 9.10923004e-01 -5.79571247e-01
1.38822758e+00 7.36818835e-02 7.44008660e-01 1.07470739e+00
9.58452523e-01 7.26463258e-01 1.13367045e+00 -2.37129122e-01
4.18464601e-01 1.62063867e-01 2.04581782e-01 3.50042641e-01
3.95130098e-01 5.96834719e-01 -8.41462195e-01 2.85659164e-01
1.60725459e-01 -2.50510216e-01 6.50698125e-01 -6.65194094e-01
-8.83309662e-01 9.09505129e-01 3.35224718e-01 8.64826813e-02
-2.54467160e-01 5.53139001e-02 7.90218055e-01 3.33808333e-01
1.59948841e-01 3.52638885e-02 -3.23675781e-01 -6.31339848e-02
-4.45423603e-01 3.79397184e-01 3.92148912e-01 7.91621387e-01
1.08249414e+00 4.05831128e-01 5.21716118e-01 8.15400541e-01
5.84226251e-01 8.15903008e-01 4.40210849e-01 -1.47665584e+00
2.71370381e-01 4.13615197e-01 5.68930581e-02 -7.81000733e-01
-6.44841552e-01 -2.23305807e-01 -5.52477837e-01 9.67278957e-01
-3.72058868e-01 -2.37857983e-01 -5.84979892e-01 1.38391006e+00
2.49267697e-01 -1.78933978e-01 5.78846753e-01 7.30954111e-01
1.20242643e+00 1.43792272e-01 -3.39436419e-02 2.26232693e-01
1.30740106e+00 -1.03891242e+00 -1.11493051e+00 -6.95808709e-01
3.52016419e-01 -6.45079315e-01 5.53815901e-01 3.78894210e-01
-7.09096313e-01 -1.41096413e+00 -1.76256394e+00 -2.28604436e-01
-1.19370222e+00 -1.79859012e-01 5.59698522e-01 6.43472075e-01
-9.81456578e-01 3.37136276e-02 -6.65741920e-01 -4.72331524e-01
3.59763466e-02 2.37051576e-01 -3.48006010e-01 4.53097671e-02
-1.35298908e+00 1.54526281e+00 4.66172636e-01 -2.67894119e-02
-9.02559578e-01 -3.46812457e-01 -1.13866663e+00 -5.28585017e-01
2.66067326e-01 -7.20824063e-01 1.23801899e+00 -5.86063206e-01
-1.40321171e+00 4.66896385e-01 -4.45440680e-01 -5.77138424e-01
2.68748790e-01 -4.08284545e-01 -7.40639210e-01 -3.77072319e-02
3.09927613e-01 8.48771155e-01 4.84031826e-01 -1.47434318e+00
-1.28806973e+00 -4.26132441e-01 3.46139193e-01 5.07098436e-01
5.27937412e-02 -2.27299809e-01 -6.67511821e-02 2.24904343e-01
2.19432369e-01 -6.55974448e-01 -2.50558048e-01 -1.18824840e-01
1.10221863e-01 -3.41373235e-01 1.29622006e+00 -8.39755014e-02
6.01969838e-01 -2.26784348e+00 -1.19854100e-01 -3.06105047e-01
2.91732699e-01 1.76236600e-01 -8.30255374e-02 6.38203800e-01
1.61344767e-01 -5.23860097e-01 2.77456075e-01 -6.39069855e-01
1.28971338e-01 7.47209132e-01 -2.41015017e-01 3.93698335e-01
-1.79424688e-01 9.74503756e-01 -8.12129498e-01 -2.73348927e-01
1.05801797e+00 5.07410228e-01 2.49628443e-02 1.17141061e-01
1.05185501e-01 2.51208723e-01 -1.79180920e-01 6.92964137e-01
7.36926138e-01 3.73073220e-01 -5.01845367e-02 -5.86458564e-01
-6.56881630e-01 2.70669162e-01 -1.13603330e+00 1.48371911e+00
-4.08561766e-01 9.48667347e-01 2.49063909e-01 -1.16241789e+00
1.06923270e+00 2.84846544e-01 6.50829494e-01 -1.22864556e+00
2.91789562e-01 2.33659580e-01 -9.59622487e-02 -9.60519850e-01
6.80776000e-01 -1.46475494e-01 -2.33324677e-01 -1.30133107e-01
1.41015589e-01 -1.30566701e-01 2.51564771e-01 -2.91302651e-01
6.84615374e-01 2.29867578e-01 4.40662324e-01 -3.16487141e-02
5.27740657e-01 4.74598370e-02 6.86513364e-01 6.93628669e-01
-7.77453959e-01 1.75359607e-01 -1.00209422e-01 -6.57722414e-01
-7.90628672e-01 -1.10895348e+00 -1.43335909e-01 1.01971996e+00
6.41542494e-01 -1.89879388e-01 -4.16943222e-01 -2.65046626e-01
3.05935800e-01 5.13195932e-01 -6.26589596e-01 -2.04425856e-01
-1.74063131e-01 -4.19813335e-01 4.79381412e-01 1.11241627e+00
7.71716475e-01 -8.83594453e-01 -1.21411836e+00 1.58802822e-01
-4.00304645e-01 -1.56181979e+00 6.57204926e-01 3.52583319e-01
-6.25096560e-01 -9.03518379e-01 1.46067321e-01 -4.68858540e-01
-3.68390493e-02 8.27525437e-01 7.44727015e-01 -4.75732684e-01
5.17465509e-02 6.54518723e-01 -2.17910156e-01 -1.02825141e+00
-2.85224289e-01 -2.71694422e-01 3.89879882e-01 -1.76534787e-01
9.84916627e-01 -4.04612929e-01 -2.11607382e-01 1.52588055e-01
-5.66707313e-01 -1.12977147e-01 6.30110323e-01 2.89221615e-01
5.59063196e-01 2.84061711e-02 7.80413508e-01 -5.54016382e-02
4.62285012e-01 -4.57030416e-01 -6.04142547e-01 1.95595976e-02
-6.42733395e-01 -3.65552515e-01 2.97637105e-01 2.25688219e-02
-9.67190564e-01 1.63640343e-02 -3.86719584e-01 -3.85484159e-01
-8.19979906e-01 4.23450023e-01 -3.38341266e-01 -3.81028861e-01
5.81942558e-01 8.79157782e-02 3.50731671e-01 -2.57027209e-01
5.34897625e-01 6.74438834e-01 6.22226417e-01 -1.53809205e-01
2.54370153e-01 8.28903019e-01 7.42029399e-02 -1.16839230e+00
-7.32176840e-01 -7.57690787e-01 -9.55391169e-01 -7.84571111e-01
1.04231143e+00 -1.12467372e+00 -7.40529954e-01 6.16479456e-01
-9.98082340e-01 -3.06553785e-02 -4.18767422e-01 8.23114276e-01
-9.03211057e-01 3.42005759e-01 -2.02557176e-01 -6.82128727e-01
1.44302189e-01 -1.47947919e+00 1.04277968e+00 3.21808547e-01
3.84784229e-02 -1.06454945e+00 3.47502708e-01 1.01190758e+00
5.35177290e-01 2.03052491e-01 3.22816819e-01 -3.14246982e-01
-4.64718670e-01 -2.82352328e-01 -1.53016031e-01 5.93628109e-01
9.02942270e-02 -1.66550055e-01 -1.44650364e+00 -1.11109339e-01
1.66314676e-01 -2.96470225e-01 8.60116720e-01 4.61184025e-01
5.77767789e-01 4.96012151e-01 -6.07080996e-01 4.28414166e-01
1.60515118e+00 4.33027059e-01 3.62611234e-01 5.79983950e-01
5.13152003e-01 6.90915763e-01 8.45617354e-01 2.24013075e-01
1.09826970e+00 4.65291828e-01 9.36868906e-01 -7.41923079e-02
-3.14861178e-01 2.66381234e-01 4.11223501e-01 9.57719088e-01
-6.21135235e-02 -1.29343376e-01 -6.46554112e-01 5.42968035e-01
-2.08163166e+00 -8.78021777e-01 -7.91277550e-03 1.59710765e+00
-6.03579059e-02 3.47702086e-01 9.65848342e-02 1.97808906e-01
3.39303195e-01 3.60616297e-01 -9.11318183e-01 -7.31688559e-01
-3.30538988e-01 -2.74770677e-01 5.60454249e-01 7.08877683e-01
-1.24999249e+00 6.80876315e-01 8.19806767e+00 4.36392695e-01
-1.16076553e+00 2.73544550e-01 -2.80282438e-01 2.01084435e-01
-6.24266230e-02 -2.31038943e-01 -1.15315747e+00 1.77669406e-01
1.11640835e+00 2.55997758e-02 -5.68792261e-02 9.65587616e-01
3.33906382e-01 -5.95572114e-01 -9.55349863e-01 9.30632651e-01
3.72395724e-01 -1.42258394e+00 -3.32527220e-01 8.06497484e-02
4.37855780e-01 8.94938111e-01 1.09983042e-01 3.10276717e-01
5.48888892e-02 -5.19479096e-01 8.12068343e-01 5.71485043e-01
2.76991069e-01 -5.58791637e-01 9.33467209e-01 3.14638376e-01
-1.51108479e+00 -5.05244195e-01 -4.49238628e-01 -4.45770711e-01
2.28898898e-01 5.11111617e-01 -3.10141891e-01 9.79421675e-01
9.64147806e-01 8.41959059e-01 -5.22049606e-01 7.55507588e-01
1.10350981e-01 -1.16086014e-01 -3.06573242e-01 1.35895565e-01
1.48400381e-01 -1.81145385e-01 5.41310370e-01 1.09245694e+00
1.15809150e-01 -3.93216610e-01 3.78248692e-01 5.53359449e-01
7.24105239e-01 -4.55586970e-01 -1.15241647e+00 4.91213858e-01
5.50353646e-01 1.27686369e+00 -2.62493968e-01 -4.66735095e-01
-1.06456447e+00 8.01165923e-02 4.28929739e-02 3.72696340e-01
-7.65364766e-01 -2.83430398e-01 1.10024488e+00 -2.22725734e-01
2.62932062e-01 -7.39277840e-01 -6.38672769e-01 -6.91057801e-01
-5.76574206e-02 -6.28639519e-01 1.39335513e-01 -9.49848115e-01
-9.59308922e-01 3.90081257e-01 4.39004987e-01 -1.35881054e+00
-3.02112788e-01 -8.42892051e-01 -2.78203309e-01 5.58093965e-01
-1.80709529e+00 -1.26603246e+00 -6.65431142e-01 3.11472416e-01
6.91536546e-01 -5.32739758e-01 8.09917688e-01 3.99211764e-01
-3.45109642e-01 2.04702392e-01 9.84690338e-02 -2.74426252e-01
5.37667513e-01 -8.25733483e-01 1.50085658e-01 7.53911853e-01
-4.01203364e-01 -4.16545793e-02 7.73930073e-01 -3.46839458e-01
-1.41940522e+00 -8.09892774e-01 5.46619236e-01 -4.58341390e-01
5.04003644e-01 -3.06929350e-01 -5.17225266e-01 7.28991926e-01
7.26711273e-01 -1.37109190e-01 8.62881243e-01 1.86492205e-02
-3.49600881e-01 -6.15917206e-01 -1.10336971e+00 3.48449469e-01
4.98624504e-01 -6.34607077e-01 -7.72180259e-01 -7.88331181e-02
4.59219515e-01 -3.18767339e-01 -8.05929959e-01 8.63796115e-01
8.56833518e-01 -1.44034517e+00 1.04638183e+00 -1.40650898e-01
-4.83158380e-02 -5.04409492e-01 -7.51555145e-01 -1.11455572e+00
-4.32759494e-01 6.51421025e-02 -3.01882774e-01 4.65995729e-01
-1.22582942e-01 -1.09524786e+00 5.02456903e-01 1.29215971e-01
-6.54160738e-01 -5.77991128e-01 -1.16062987e+00 -8.13509464e-01
-7.86024630e-02 -8.81074190e-01 5.07711470e-01 4.51609552e-01
-1.62670165e-01 4.22456533e-01 1.18619561e-01 3.33570272e-01
8.55381608e-01 -2.83770869e-03 7.90438354e-01 -1.34319580e+00
3.10420066e-01 -4.45521146e-01 -8.78675878e-01 -9.65745389e-01
1.51374206e-01 -2.29484722e-01 1.67586431e-01 -1.83187008e+00
-1.89743638e-01 -7.44057670e-02 -4.18152958e-01 3.37799013e-01
3.49965841e-01 2.88141578e-01 6.77787885e-02 -4.16132547e-02
-6.82125032e-01 5.25648296e-01 1.00638366e+00 -1.58949956e-01
2.76080191e-01 -7.60804638e-02 -5.76182127e-01 9.51755643e-01
1.05384183e+00 1.84458792e-01 -6.82411551e-01 -4.96369690e-01
2.48557672e-01 -3.15769911e-01 4.54928547e-01 -1.70443106e+00
4.56332713e-01 -2.02140421e-01 1.46891639e-01 -1.25807250e+00
1.02204335e+00 -7.89406419e-01 -1.34049803e-01 3.44034970e-01
-4.64608558e-02 2.61919618e-01 5.08236289e-01 4.58964646e-01
-6.16553605e-01 -4.49637957e-02 8.64872396e-01 -5.16322181e-02
-1.66040385e+00 -2.13509202e-01 -1.01332116e+00 -5.67007065e-01
1.23357856e+00 -6.03512347e-01 -4.69673514e-01 -2.96054602e-01
-7.11680412e-01 2.29962260e-01 1.16200501e-03 1.09011245e+00
6.87564075e-01 -1.50119412e+00 -1.75037801e-01 5.61464608e-01
3.95356864e-01 -1.75488278e-01 3.82396042e-01 6.99755430e-01
-5.87761477e-02 8.45046520e-01 -6.37234926e-01 -8.64089668e-01
-1.15612185e+00 5.76232374e-01 6.76012397e-01 3.87157321e-01
-3.07618409e-01 4.05210406e-01 1.27301127e-01 -5.39204776e-01
1.93594888e-01 -3.26585174e-01 -5.95282376e-01 2.28730306e-01
3.26199532e-01 6.90407395e-01 4.66290146e-01 -1.09189248e+00
-4.00088012e-01 8.80037606e-01 3.30131203e-01 -2.05685824e-01
7.10934818e-01 -7.46167421e-01 8.26198012e-02 9.60416436e-01
1.14760435e+00 -5.96475482e-01 -1.13333035e+00 -9.87655520e-02
-4.00390297e-01 4.23333049e-02 2.75030404e-01 -5.87273717e-01
-7.71016002e-01 1.15420461e+00 1.17174506e+00 4.76018488e-01
9.49006975e-01 -7.00695962e-02 6.53661609e-01 5.45517087e-01
4.56167430e-01 -1.52295673e+00 1.18388124e-01 9.10059571e-01
7.76968062e-01 -1.57973564e+00 -5.00917770e-02 -1.19754635e-01
-5.22245586e-01 1.06132555e+00 1.01840425e+00 -6.38127094e-03
1.16528726e+00 5.65108776e-01 6.45803392e-01 -4.52505589e-01
-8.32963765e-01 -5.49745739e-01 1.90460846e-01 1.08312809e+00
4.97542173e-02 1.71050727e-01 7.30822086e-02 1.25299677e-01
-3.05351824e-01 -1.66109446e-02 2.22938746e-01 1.16599822e+00
-8.31689358e-01 -9.51906741e-01 -4.82918113e-01 1.13963306e-01
2.46830836e-01 7.09121585e-01 -7.54840076e-02 7.51488388e-01
8.94222081e-01 1.46802866e+00 1.86444446e-01 -9.04186308e-01
5.41212618e-01 -2.39752177e-02 3.35597187e-01 -2.15554878e-01
-4.96071689e-02 -7.17199966e-02 6.67683855e-02 -7.73481786e-01
-9.99664307e-01 -7.15800464e-01 -1.13378036e+00 -1.61796406e-01
-9.46822613e-02 -3.51757079e-01 1.14651000e+00 1.35554564e+00
5.79445779e-01 8.70226741e-01 4.12574351e-01 -1.07576752e+00
-1.11206248e-01 -7.61611044e-01 -1.57773748e-01 -2.60225028e-01
7.72091806e-01 -1.15999961e+00 -1.81770936e-01 -5.39627932e-02] | [5.690496921539307, 0.8801959156990051] |
4103bbff-dc8d-49e5-90b1-33a0a2432422 | protein-interface-prediction-using-graph | null | null | http://papers.nips.cc/paper/7231-protein-interface-prediction-using-graph-convolutional-networks | http://papers.nips.cc/paper/7231-protein-interface-prediction-using-graph-convolutional-networks.pdf | Protein Interface Prediction using Graph Convolutional Networks | We consider the prediction of interfaces between proteins, a challenging problem with important applications in drug discovery and design, and examine the performance of existing and newly proposed spatial graph convolution operators for this task. By performing convolution over a local neighborhood of a node of interest, we are able to stack multiple layers of convolution and learn effective latent representations that integrate information across the graph that represent the three dimensional structure of a protein of interest. An architecture that combines the learned features across pairs of proteins is then used to classify pairs of amino acid residues as part of an interface or not. In our experiments, several graph convolution operators yielded accuracy that is better than the state-of-the-art SVM method in this task. | ['Asa Ben-Hur', 'Jonathon Byrd', 'Alex Fout', 'Basir Shariat'] | 2017-12-01 | null | null | null | neurips-2017-12 | ['protein-interface-prediction'] | ['miscellaneous'] | [ 2.52486855e-01 1.34245947e-01 -1.46693110e-01 -3.02366674e-01
-2.49188870e-01 -4.47055399e-01 2.23298088e-01 7.16780186e-01
-4.69980806e-01 9.42390144e-01 -9.82419774e-02 -8.28449845e-01
-1.31519228e-01 -8.45546246e-01 -9.69623923e-01 -9.57377672e-01
-5.51189601e-01 6.27652109e-01 3.98718119e-01 -1.67750604e-02
8.98920596e-02 1.02541482e+00 -8.44122469e-01 8.22440803e-01
9.91490781e-01 9.84828651e-01 2.84242958e-01 5.67483902e-01
-3.44266802e-01 4.54369873e-01 -4.05140579e-01 -1.33612201e-01
-1.38345152e-01 -1.48535118e-01 -1.00312865e+00 -2.78667927e-01
7.21708179e-01 4.34680402e-01 -2.39709616e-01 1.05060089e+00
1.71490327e-01 -2.80297641e-03 8.54067326e-01 -6.26098752e-01
-8.33435476e-01 1.36242017e-01 -3.12548041e-01 3.88154447e-01
2.26569131e-01 1.97523683e-02 1.13211322e+00 -7.76907802e-01
1.01989710e+00 1.07260919e+00 8.85633528e-01 2.70675659e-01
-1.64180815e+00 -3.54084164e-01 2.73566902e-01 2.02506110e-01
-1.05160546e+00 1.55532315e-01 5.44054627e-01 -5.63290060e-01
1.60204041e+00 -1.02335187e-02 6.11367881e-01 8.23885977e-01
7.89275825e-01 3.34080875e-01 8.13923001e-01 -1.59613818e-01
2.72312403e-01 -1.71149239e-01 5.59632659e-01 1.07197309e+00
1.34861201e-01 -3.30394745e-01 -4.51343477e-01 -7.76022434e-01
7.82256961e-01 2.16396943e-01 -3.91810447e-01 -8.05887163e-01
-1.05935454e+00 1.04125118e+00 1.05820930e+00 2.77187467e-01
-6.24634981e-01 7.18137622e-02 8.39262977e-02 2.69186169e-01
7.17960596e-01 6.98698580e-01 -7.32578754e-01 2.44677156e-01
-5.24542391e-01 2.27424815e-01 8.71926785e-01 3.14700186e-01
7.61464000e-01 -5.39453983e-01 -1.01374770e-02 5.99902511e-01
1.91465691e-01 -3.00437301e-01 2.42769569e-02 -3.07314962e-01
2.76987642e-01 1.06693172e+00 -8.50805081e-03 -6.91373825e-01
-6.94469929e-01 -3.82502496e-01 -8.47395003e-01 5.80890536e-01
6.31086349e-01 -7.68925995e-02 -1.13556290e+00 1.59547174e+00
2.01857269e-01 3.85503262e-01 1.62126973e-01 5.79008818e-01
9.21517789e-01 5.74298382e-01 2.53471822e-01 -5.89285903e-02
1.29957736e+00 -9.49335754e-01 -4.82723296e-01 -2.15784043e-01
8.44544530e-01 -4.76735085e-01 5.59008658e-01 8.46996251e-03
-8.69976580e-01 -4.52276200e-01 -1.14222455e+00 -3.38195302e-02
-8.02342892e-01 -4.05454002e-02 1.01608276e+00 -6.52188137e-02
-9.74186659e-01 1.32907271e+00 -8.38148475e-01 -4.74905908e-01
7.22093940e-01 8.19194436e-01 -7.53014803e-01 4.00352478e-02
-1.24652183e+00 1.01865733e+00 1.36705309e-01 -7.65518099e-02
-3.87815446e-01 -8.91655087e-01 -9.07432675e-01 1.96370736e-01
-1.29601046e-01 -6.39116704e-01 5.84642410e-01 -8.91703844e-01
-8.85942996e-01 8.59297872e-01 -4.90968883e-01 -6.76708519e-01
1.65318474e-02 -1.18554853e-01 -1.79232523e-01 8.14903378e-02
-1.02877617e-01 4.45987761e-01 2.87095040e-01 -6.61692560e-01
-2.02955052e-01 -6.55124366e-01 3.97193581e-02 1.26922160e-01
-5.30580394e-02 8.89471918e-02 -1.21810883e-01 -3.56991768e-01
1.25928715e-01 -7.66109943e-01 -6.38294101e-01 4.51440126e-01
-3.08963120e-01 -3.38653237e-01 8.80337179e-01 -6.79061890e-01
7.45818555e-01 -1.83445895e+00 5.24658084e-01 3.59920651e-01
6.52907789e-01 5.63189209e-01 -1.78317353e-01 4.76808459e-01
-5.48481107e-01 -4.49263901e-02 -3.55661541e-01 1.65667102e-01
-5.68018198e-01 -7.12066144e-02 -5.63944951e-02 6.58471584e-01
4.97593313e-01 1.19658601e+00 -7.04863012e-01 3.75698619e-02
5.67774922e-02 8.81984234e-01 -3.83309543e-01 4.19783473e-01
-5.09234011e-01 2.90830910e-01 -5.26257634e-01 2.76265651e-01
6.70734107e-01 -9.74360526e-01 6.97306335e-01 -2.18299896e-01
2.26406097e-01 5.22163928e-01 -6.62099838e-01 1.58099651e+00
4.15708795e-02 5.69995582e-01 -3.99904400e-02 -1.18078041e+00
8.95228565e-01 2.70900697e-01 5.88309407e-01 -3.39274406e-01
-2.23879755e-01 1.16860338e-01 3.33010048e-01 -2.10734159e-01
-2.89602697e-01 6.64759055e-02 4.23462242e-01 2.10371614e-01
2.73992658e-01 6.59679115e-01 -2.38160789e-01 -4.83719166e-03
1.59819520e+00 1.05017692e-01 2.71990687e-01 -4.57516938e-01
3.88836056e-01 2.18482558e-02 2.32643589e-01 3.81556809e-01
-7.23224431e-02 3.30451518e-01 9.04986203e-01 -9.90076184e-01
-1.00086629e+00 -1.00138617e+00 -1.30660444e-01 8.45281065e-01
9.13323164e-02 -3.36870849e-01 -7.53489137e-01 -1.02544904e+00
5.76732337e-01 -1.08706303e-01 -9.27973688e-01 -2.19746888e-01
-6.88742816e-01 -8.82198691e-01 1.14097916e-01 6.67330325e-01
2.97860000e-02 -1.22718811e+00 -2.90496558e-01 3.97739291e-01
4.31363642e-01 -8.86511147e-01 -2.71444738e-01 1.03076708e+00
-9.89710867e-01 -1.46532881e+00 -7.45975196e-01 -1.22461367e+00
7.76440978e-01 5.84411360e-02 1.29247987e+00 9.64951143e-02
-5.25766909e-01 -5.90490878e-01 1.76447585e-01 -1.12916634e-03
-1.80330440e-01 7.23680630e-02 -1.38633132e-01 -9.55451876e-02
7.27900624e-01 -5.62180102e-01 -5.98196566e-01 6.55732378e-02
-4.43423271e-01 -1.86614618e-02 3.54230344e-01 1.05349708e+00
6.29211366e-01 -2.14910433e-01 2.32439771e-01 -1.28984857e+00
8.73571336e-01 -5.02199113e-01 -6.05233312e-01 4.54874724e-01
-2.11051866e-01 5.94889998e-01 7.04027534e-01 -5.76604307e-01
-5.64421475e-01 3.76717597e-01 -1.12978831e-01 -3.46589178e-01
-2.40389749e-01 6.33968234e-01 4.64454889e-02 -6.06436014e-01
5.49405754e-01 -8.22877735e-02 2.94918120e-01 -6.20528460e-01
2.64929801e-01 4.22107667e-01 9.24291313e-02 -3.38028491e-01
1.85102522e-01 2.78022885e-01 3.05717319e-01 -8.01707625e-01
-5.36032319e-01 -4.43768770e-01 -9.71821010e-01 3.19035798e-01
1.22278845e+00 -5.22123039e-01 -9.42845762e-01 4.62643087e-01
-1.40989363e+00 -4.59293306e-01 1.43737629e-01 3.02505761e-01
-4.49538022e-01 3.83353591e-01 -8.70201588e-01 -2.21381083e-01
-1.77116960e-01 -1.45696759e+00 9.65072215e-01 1.44102827e-01
-3.02701116e-01 -1.41684926e+00 2.17570722e-01 5.59996143e-02
3.50009650e-01 3.34297448e-01 1.62349761e+00 -9.60918128e-01
-6.34886742e-01 -1.01130620e-01 -3.51447374e-01 -1.33126602e-01
2.92166978e-01 -1.63088724e-01 -7.58678555e-01 -4.52952117e-01
-4.79329973e-01 -1.98496237e-01 1.17910540e+00 6.28801167e-01
1.12501788e+00 -2.13982746e-01 -1.00481832e+00 8.13219428e-01
1.35145223e+00 2.96535075e-01 5.44007301e-01 8.61429796e-02
7.40618587e-01 5.51437616e-01 2.02968925e-01 -1.51450306e-01
-2.64793217e-01 1.03476965e+00 4.56359863e-01 -8.02272081e-01
1.61796063e-01 3.34401987e-02 -5.35626560e-02 -8.70304257e-02
-9.42118093e-02 -2.42541090e-01 -8.35373700e-01 2.11220592e-01
-2.19810367e+00 -6.93395615e-01 -1.45018220e-01 1.97347379e+00
8.51766467e-01 1.40047878e-01 5.67729846e-02 -4.84544456e-01
6.98471427e-01 1.06810600e-01 -6.96382165e-01 -6.68668687e-01
-5.87346032e-02 7.80175328e-01 6.96426094e-01 6.40183628e-01
-1.35604203e+00 1.02164161e+00 7.20527935e+00 4.57745612e-01
-1.22780263e+00 -4.08142895e-01 6.94688618e-01 4.89930004e-01
-1.25638381e-01 -1.49308309e-01 -6.37273014e-01 2.18713388e-01
9.10388350e-01 3.03912312e-01 3.25160652e-01 5.47834277e-01
-1.28699780e-01 2.43770763e-01 -1.23628330e+00 4.81327862e-01
-2.51218200e-01 -1.97144592e+00 6.27786443e-02 4.15504873e-01
5.09194791e-01 1.39401212e-01 3.96127254e-02 -3.47505033e-01
4.53569263e-01 -1.68011057e+00 -2.15347365e-01 5.70150316e-01
5.28536379e-01 -6.15757704e-01 6.73753202e-01 7.79003426e-02
-1.28672183e+00 3.32789987e-01 -6.47033274e-01 -2.22105458e-01
-8.85018855e-02 4.42027360e-01 -1.04750276e+00 1.88052163e-01
5.14557838e-01 7.80373216e-01 -4.72404957e-01 1.10051346e+00
-1.59568876e-01 2.04598740e-01 -5.28359488e-02 -6.69055730e-02
4.90441591e-01 -3.69948208e-01 2.57872045e-01 9.96958315e-01
-1.45866215e-01 -1.15700951e-02 2.62179643e-01 1.08961380e+00
-2.74904221e-01 1.37692764e-01 -9.11225200e-01 -1.49127349e-01
7.69938082e-02 1.07910025e+00 -6.02586746e-01 -2.27518559e-01
-4.99218792e-01 1.28588438e+00 1.25479448e+00 6.66912973e-01
-6.45999312e-01 -5.86094737e-01 1.54782856e+00 1.26031935e-01
6.41092002e-01 -2.81217515e-01 -4.81738988e-03 -7.54474163e-01
-5.60089685e-02 -6.16632581e-01 3.28294039e-01 -5.38256764e-01
-1.23116291e+00 7.76615739e-01 -6.56174004e-01 -6.64698541e-01
2.26427749e-01 -1.39437699e+00 -8.15261126e-01 1.45611250e+00
-1.23909104e+00 -9.92425621e-01 3.91059071e-02 2.28076667e-01
6.38215095e-02 -3.09410095e-01 1.24957764e+00 -5.44243380e-02
-4.29647148e-01 4.30663288e-01 4.34562713e-01 6.62458390e-02
4.29180235e-01 -1.41181326e+00 1.05392683e+00 2.15629235e-01
2.64572531e-01 9.89927888e-01 4.94188875e-01 -8.56932938e-01
-1.06678140e+00 -1.17226899e+00 1.04868305e+00 -3.22700351e-01
5.82304835e-01 -5.93249023e-01 -1.51792073e+00 5.74889719e-01
9.57743078e-02 4.47674453e-01 8.10233295e-01 2.77132362e-01
-4.89266396e-01 4.81900066e-01 -1.12418091e+00 4.88482207e-01
1.14244437e+00 -7.11107314e-01 -3.37243617e-01 5.69246888e-01
8.23944747e-01 -3.71054649e-01 -9.60061729e-01 4.67495024e-01
5.58030248e-01 -8.04316878e-01 1.23963499e+00 -1.67106867e+00
1.71165183e-01 -1.36119440e-01 1.85390040e-01 -1.33072174e+00
-7.55602479e-01 -2.68891454e-01 -8.86914060e-02 1.72673404e-01
8.90280843e-01 -7.55372047e-01 1.23631549e+00 2.62883544e-01
4.63330932e-02 -1.39359272e+00 -8.14840615e-01 -4.32078093e-01
2.61441767e-01 3.54334861e-01 2.76679039e-01 7.84889400e-01
3.43095005e-01 6.83453679e-01 7.68564492e-02 3.48344237e-01
4.52793390e-01 3.03801417e-01 3.49550933e-01 -1.45498812e+00
-3.91990244e-01 -4.41518754e-01 -8.47980797e-01 -1.07422936e+00
5.62918246e-01 -1.16500878e+00 -4.65767860e-01 -1.81737638e+00
2.37193733e-01 -1.12790041e-01 -5.39464414e-01 5.83722293e-01
-3.62351418e-01 1.15059502e-01 -3.30623895e-01 -1.71358213e-02
-4.04460907e-01 1.67364255e-01 1.12832391e+00 -4.12535936e-01
-1.19842514e-01 -3.64595540e-02 -4.35886532e-01 4.96197343e-01
5.92781723e-01 -2.35797450e-01 -5.32262102e-02 6.07180409e-02
-1.96745083e-01 -3.42591517e-02 2.44999409e-01 -7.77219236e-01
-1.18451761e-02 -1.62572160e-01 6.59394562e-01 -3.84607226e-01
5.01503825e-01 -7.99817383e-01 2.01745719e-01 7.37657905e-01
-4.99679297e-01 1.08655788e-01 2.46051729e-01 6.10647619e-01
-1.09252840e-01 1.43147811e-01 9.05222595e-01 -1.77002981e-01
-4.17560518e-01 5.36846101e-01 -3.25025350e-01 -2.43516445e-01
1.17327082e+00 -1.65232867e-01 -4.10478860e-01 3.85564310e-03
-1.09341109e+00 -3.41622718e-02 8.34508300e-01 1.31207928e-01
7.29792535e-01 -1.08786488e+00 -2.67935425e-01 4.76554066e-01
1.69511721e-01 -2.79911041e-01 -2.44854853e-01 5.86560965e-01
-8.06628644e-01 6.96886420e-01 -4.35730368e-01 -5.48244953e-01
-1.84314561e+00 7.01828420e-01 8.25798869e-01 -5.57298124e-01
-6.80555582e-01 1.03058851e+00 5.93049347e-01 -3.13786298e-01
2.85695553e-01 -4.00312513e-01 -3.65837783e-01 -2.54626274e-01
2.69761741e-01 -1.64239734e-01 3.02949280e-01 -5.85342050e-01
-5.50215304e-01 6.39860153e-01 -3.14126730e-01 7.06549704e-01
1.50034058e+00 7.78912485e-01 -2.89368898e-01 1.50491312e-01
1.46823490e+00 -5.07157326e-01 -1.26626325e+00 -3.16170841e-01
1.79856449e-01 -1.35663405e-01 -2.10537180e-01 -7.30994225e-01
-8.40931177e-01 9.77600873e-01 7.87370920e-01 7.32920319e-02
4.69386667e-01 3.15647244e-01 6.33160472e-01 5.10106862e-01
1.48626462e-01 -3.89221936e-01 -2.06404645e-02 6.07600808e-01
4.98034030e-01 -1.17035735e+00 -7.46346489e-02 -7.51250923e-01
-2.61377215e-01 1.24264884e+00 5.68480670e-01 -3.03688198e-01
7.69573927e-01 3.96817178e-01 6.45953268e-02 -8.07614803e-01
-8.83032918e-01 -2.05942541e-01 6.01809800e-01 6.78627968e-01
9.81488526e-01 2.78602056e-02 -2.49414176e-01 1.13604516e-01
6.09282613e-01 -1.17457844e-01 -2.02727169e-01 9.41419899e-01
-6.58280015e-01 -1.32673204e+00 2.01373503e-01 6.64686382e-01
-4.25482780e-01 -3.46386522e-01 -8.95773351e-01 3.42040271e-01
-8.00233632e-02 3.07055503e-01 1.40944168e-01 -2.62031674e-01
2.52625108e-01 4.02647972e-01 5.39493084e-01 -7.77958751e-01
-7.91041434e-01 -1.29699096e-01 4.44520935e-02 -6.20092273e-01
-2.15702474e-01 -2.55921930e-01 -1.51165211e+00 1.02374386e-02
-3.92840981e-01 3.94740641e-01 3.02042037e-01 8.71817470e-01
6.85341001e-01 6.33055449e-01 1.88343883e-01 -6.82189226e-01
-2.70700812e-01 -7.68521726e-01 -6.45325005e-01 6.61406934e-01
4.22385842e-01 -7.77372599e-01 -2.72920411e-02 -2.55114943e-01] | [4.8799147605896, 5.673135280609131] |
bfdb6dde-7911-463a-97bb-50a1945b4ed3 | anomaly-detection-in-multiplex-dynamic | 2211.08378 | null | https://arxiv.org/abs/2211.08378v1 | https://arxiv.org/pdf/2211.08378v1.pdf | Anomaly Detection in Multiplex Dynamic Networks: from Blockchain Security to Brain Disease Prediction | The problem of identifying anomalies in dynamic networks is a fundamental task with a wide range of applications. However, it raises critical challenges due to the complex nature of anomalies, lack of ground truth knowledge, and complex and dynamic interactions in the network. Most existing approaches usually study networks with a single type of connection between vertices, while in many applications interactions between objects vary, yielding multiplex networks. We propose ANOMULY, a general, unsupervised edge anomaly detection framework for multiplex dynamic networks. In each relation type, ANOMULY sees node embeddings at different GNN layers as hierarchical node states and employs a GRU cell to capture temporal properties of the network and update node embeddings over time. We then add an attention mechanism that incorporates information across different types of relations. Our case study on brain networks shows how this approach could be employed as a new tool to understand abnormal brain activity that might reveal a brain disease or disorder. Extensive experiments on nine real-world datasets demonstrate that ANOMULY achieves state-of-the-art performance. | ['Margo Seltzer', 'Ali Behrouz'] | 2022-11-15 | null | null | null | null | ['disease-prediction'] | ['medical'] | [ 1.67790204e-01 1.59885913e-01 -4.02027508e-03 -1.29752100e-01
7.47416854e-01 -5.65973639e-01 7.14792728e-01 4.51398998e-01
-7.82120377e-02 4.46911067e-01 4.27311212e-02 -2.16505870e-01
-3.68908137e-01 -1.06577241e+00 -5.24607301e-01 -5.45742273e-01
-9.51989114e-01 5.54604530e-01 6.42822504e-01 -3.71840745e-01
-1.83639824e-02 7.57615387e-01 -9.79593217e-01 -4.92314398e-02
5.00900567e-01 8.25248837e-01 -5.56707323e-01 6.08419478e-01
-2.84305662e-01 6.42898321e-01 -4.85875845e-01 -6.19769812e-01
2.65407771e-01 -3.67445588e-01 -8.06010604e-01 -6.58354536e-02
1.88115045e-01 1.09282322e-01 -8.95997465e-01 1.41180003e+00
3.70476067e-01 -1.88909601e-02 6.67217970e-01 -1.97467208e+00
-3.46962094e-01 9.09025550e-01 -6.74706697e-01 1.20232236e+00
1.56464487e-01 1.17133081e-01 1.15459967e+00 -3.14625263e-01
7.39381671e-01 1.21653569e+00 6.60747230e-01 5.46190798e-01
-1.66696525e+00 -6.56144083e-01 6.60691440e-01 4.99609292e-01
-1.07470167e+00 -2.66015023e-01 8.40026379e-01 -5.97800255e-01
9.95293021e-01 3.30899209e-02 1.06312895e+00 1.52156901e+00
5.13965070e-01 4.23754245e-01 5.60010374e-01 2.73884684e-01
1.67324185e-01 -5.49406290e-01 5.07909715e-01 7.39226758e-01
6.26275182e-01 -2.66178608e-01 -7.02409565e-01 -5.14779687e-01
7.68530905e-01 3.18106622e-01 -2.18206137e-01 -4.57122684e-01
-1.26095653e+00 6.09506667e-01 6.77412689e-01 4.82696623e-01
-4.45864081e-01 4.65213984e-01 8.11990499e-01 6.91129565e-01
4.68116909e-01 4.38318044e-01 -4.40078765e-01 -1.01897590e-01
-4.43303198e-01 -6.17034659e-02 9.22323287e-01 5.20962954e-01
3.12149197e-01 1.80790171e-01 1.16430290e-01 6.49893761e-01
3.21821839e-01 -2.35657007e-01 3.07719499e-01 -4.57234800e-01
2.29337797e-01 1.02829134e+00 -7.21324027e-01 -1.47820234e+00
-7.85770953e-01 -6.39292777e-01 -1.30831218e+00 -6.58816658e-03
3.23530704e-01 2.75505092e-02 -6.96276605e-01 2.00535536e+00
3.15203279e-01 8.98631632e-01 -3.14870745e-01 3.06869209e-01
5.86207271e-01 2.10218981e-01 -5.53360470e-02 -9.00847167e-02
1.28321731e+00 -6.57908320e-01 -7.33981431e-01 -4.29674119e-01
6.38079882e-01 8.17100182e-02 4.36200291e-01 2.11760342e-01
-7.53981590e-01 1.75157651e-01 -9.45039272e-01 3.70538682e-01
-7.17638731e-01 -9.01971102e-01 8.33435357e-01 4.95002717e-01
-1.26218104e+00 6.12986028e-01 -1.15173042e+00 -7.19461679e-01
7.85877347e-01 5.66337407e-01 -5.17334461e-01 5.65484017e-02
-1.16522861e+00 5.93423605e-01 3.41802746e-01 2.93270856e-01
-7.73358107e-01 -6.71992838e-01 -8.48725200e-01 2.44592100e-01
4.46791112e-01 -5.55946529e-01 6.60648286e-01 -6.25604033e-01
-8.34283531e-01 6.35330498e-01 -1.11741260e-01 -5.99910736e-01
3.66486132e-01 2.24715963e-01 -1.01025224e+00 1.01138487e-01
-5.28921597e-02 9.83027294e-02 6.60884917e-01 -7.90971398e-01
-2.27465659e-01 -6.81745648e-01 -6.99094124e-03 -1.51796997e-01
-5.40219545e-01 -2.42699191e-01 -2.64431953e-01 -6.23347342e-01
5.18930495e-01 -8.85514438e-01 -2.74635196e-01 1.20611124e-01
-7.09718704e-01 -3.09667081e-01 1.03888249e+00 -1.36329100e-01
1.53920758e+00 -1.95123053e+00 3.52699101e-01 7.82070994e-01
1.08327448e+00 6.40076101e-02 -3.29260170e-01 4.33198601e-01
-4.40206438e-01 3.34436268e-01 -4.56612766e-01 -3.06824669e-02
-3.63238543e-01 4.40540195e-01 -3.14247608e-02 5.22079229e-01
2.88750827e-01 9.90316153e-01 -1.29247737e+00 4.61815409e-02
-2.54308730e-01 2.68647701e-01 -4.27619070e-01 -2.80331254e-01
-6.13000318e-02 2.57297993e-01 -4.47211266e-01 6.77034736e-01
4.33815420e-01 -4.09755766e-01 4.47015107e-01 -1.81334585e-01
4.00060862e-01 1.00413710e-02 -1.06463289e+00 1.33453465e+00
1.31631330e-01 8.51359010e-01 -1.01089619e-01 -1.31867993e+00
5.49298584e-01 2.51094341e-01 7.09916115e-01 -4.85180408e-01
1.82139605e-01 -7.55226165e-02 7.47814238e-01 -5.50711334e-01
2.29805559e-02 5.02086818e-01 1.87904999e-01 7.21796632e-01
1.93027884e-01 4.42988724e-01 4.63148594e-01 7.47183800e-01
2.05762887e+00 -6.56331539e-01 3.35255504e-01 -2.07986698e-01
3.81590098e-01 -4.75231767e-01 6.74707055e-01 6.70835435e-01
-6.36226416e-01 1.01257533e-01 1.48469043e+00 -5.46454787e-01
-7.57282853e-01 -1.31899846e+00 -2.22013444e-01 8.32298517e-01
-1.31402850e-01 -5.74352384e-01 -4.04642701e-01 -8.56250584e-01
1.89539716e-01 7.69514367e-02 -9.54629838e-01 -5.60903430e-01
-3.92202586e-01 -1.24136043e+00 6.73529267e-01 3.63886446e-01
3.19839925e-01 -1.08011842e+00 -1.51467383e-01 3.58726472e-01
1.81562249e-02 -1.28356063e+00 -1.99520245e-01 1.03005674e-02
-1.12281740e+00 -1.58774185e+00 8.08173791e-03 -6.29226327e-01
9.32432830e-01 -2.18814444e-02 1.20142484e+00 4.27166373e-01
-6.54570758e-01 5.82156003e-01 -1.87055543e-01 -5.55900335e-02
-2.32594013e-01 1.21373691e-01 5.22843719e-01 3.13774765e-01
4.25512761e-01 -1.24732542e+00 -6.34900272e-01 5.48360273e-02
-1.05147743e+00 -7.06726670e-01 3.54091108e-01 7.15190887e-01
1.75393060e-01 3.57003450e-01 7.42043316e-01 -1.32200861e+00
9.21425045e-01 -1.10140991e+00 -3.88251305e-01 -4.18172181e-02
-6.93036318e-01 5.24366647e-02 4.03041869e-01 -5.63422143e-01
-2.68820226e-01 -2.95976639e-01 9.18288380e-02 -2.26388022e-01
6.32305965e-02 5.05296052e-01 -7.95770809e-02 -1.06379241e-01
6.48968458e-01 3.40516642e-02 1.83941886e-01 -9.84300524e-02
1.64964199e-01 -6.62947670e-02 4.91712064e-01 -2.64214694e-01
6.82187378e-01 5.02354085e-01 3.88939768e-01 -8.23723078e-01
-4.44091797e-01 -3.63571256e-01 -8.76041651e-01 -3.05963039e-01
5.34213126e-01 -4.19979155e-01 -9.57652509e-01 6.69936776e-01
-9.61183965e-01 -3.08557212e-01 -9.51624140e-02 1.41002923e-01
-9.67110600e-03 3.34405690e-01 -8.43058467e-01 -4.26021695e-01
-1.78089082e-01 -8.55879605e-01 2.59263903e-01 -2.85451710e-02
-4.24276501e-01 -1.42124510e+00 2.06312567e-01 -3.10136586e-01
3.96384239e-01 5.88430524e-01 1.18799937e+00 -9.77824271e-01
-5.53725541e-01 -2.56613314e-01 -2.39922479e-01 -2.50780024e-02
3.42293650e-01 1.43338755e-01 -6.80043042e-01 -2.30952039e-01
-3.74518335e-01 2.31411606e-01 8.46192598e-01 1.01657137e-01
1.24402356e+00 -3.72252852e-01 -6.39264345e-01 4.79703069e-01
1.22846949e+00 4.38944958e-02 5.60185313e-01 1.23658560e-01
8.92892897e-01 4.73594517e-01 -4.09437209e-01 3.76157492e-01
3.93331587e-01 1.37887612e-01 8.47537458e-01 2.70919561e-01
1.47436336e-01 1.50508299e-01 3.30078542e-01 9.47202206e-01
1.08469248e-01 -3.89830917e-01 -1.20012379e+00 7.28173077e-01
-2.00997376e+00 -1.20374286e+00 -3.66406381e-01 1.84535551e+00
5.08265674e-01 5.71907341e-01 1.84407786e-01 1.73321918e-01
6.84316993e-01 3.29196274e-01 -9.75421369e-01 -3.33561033e-01
-7.38383755e-02 -7.28787929e-02 1.55882731e-01 1.71448916e-01
-7.07594395e-01 7.80197084e-01 7.03694105e+00 7.57258758e-02
-8.87741446e-01 1.11673534e-01 5.18134356e-01 -1.91721424e-01
-1.07470751e-01 -3.02007258e-01 -2.79923677e-01 5.99444509e-01
9.11396325e-01 -3.13520938e-01 5.39757490e-01 2.34676912e-01
-6.53615072e-02 6.51640492e-03 -1.39622104e+00 8.78754854e-01
-2.41274852e-02 -1.13987684e+00 5.32133467e-02 2.60904521e-01
7.58339584e-01 3.90792608e-01 -4.96761352e-02 -9.17359889e-02
5.53709447e-01 -9.04491484e-01 -2.58122385e-02 7.10186899e-01
3.53854358e-01 -6.17086112e-01 6.21549129e-01 4.50393148e-02
-1.39246583e+00 -1.95690095e-01 -1.25078514e-01 -9.22528729e-02
-2.03027446e-02 9.62718368e-01 -7.15980709e-01 1.77951977e-01
7.70571947e-01 1.16612339e+00 -7.88570464e-01 1.20841503e+00
-2.42078453e-01 6.30941153e-01 -3.94274831e-01 1.69816315e-01
8.74081329e-02 -2.93116987e-01 9.00191367e-01 1.08203530e+00
-3.14463652e-03 -1.11227229e-01 -8.04227367e-02 8.77848685e-01
-3.69404763e-01 -2.65552253e-01 -1.10707235e+00 -3.95489693e-01
6.07764304e-01 1.47278619e+00 -1.21982741e+00 -2.16216594e-01
-3.78714055e-01 1.10570669e+00 5.05118489e-01 4.78597343e-01
-4.34712529e-01 -2.09256381e-01 1.26026475e+00 1.41137838e-01
-9.25600901e-02 -3.34451973e-01 -7.33387247e-02 -1.25791621e+00
-7.41126575e-03 -6.38507128e-01 6.22245312e-01 -1.84237435e-01
-1.67971444e+00 6.60752594e-01 -2.13333681e-01 -9.41687822e-01
-1.33022889e-01 -5.64006686e-01 -1.14111674e+00 3.30127597e-01
-1.17987931e+00 -5.09764016e-01 -3.64785194e-01 7.57769346e-01
1.49879724e-01 -2.12146521e-01 6.89635634e-01 3.14346582e-01
-1.09586132e+00 4.76726741e-01 -2.76360065e-02 5.19145787e-01
3.51335078e-01 -1.17733550e+00 8.70982111e-01 1.08826268e+00
2.23321393e-01 7.06511974e-01 5.70883751e-01 -9.00825977e-01
-1.18960190e+00 -1.09167373e+00 5.41175187e-01 -4.83919710e-01
1.48067260e+00 -6.67624772e-01 -1.16496480e+00 1.01493728e+00
7.23158196e-02 6.20292842e-01 6.89628124e-01 3.65433186e-01
-4.71304834e-01 -7.82948136e-02 -1.11735237e+00 8.41867149e-01
1.78637183e+00 -6.97082460e-01 -2.00024784e-01 3.14393491e-01
5.69204807e-01 -8.78085643e-02 -9.48958278e-01 2.53114313e-01
3.88274848e-01 -7.29309261e-01 7.71269500e-01 -9.54570949e-01
2.26593316e-01 -2.19019130e-01 3.76261353e-01 -1.72648358e+00
-5.27871788e-01 -6.14559710e-01 -9.21769559e-01 1.10071790e+00
5.24941862e-01 -1.19738567e+00 7.29913294e-01 4.40081507e-01
1.25327632e-01 -8.74508262e-01 -9.07072067e-01 -6.71462774e-01
-3.13775659e-01 -4.66978818e-01 6.18896127e-01 1.41063130e+00
2.56756365e-01 3.68657559e-01 5.71196936e-02 3.48603159e-01
6.81069493e-01 -6.00234449e-01 3.48694593e-01 -1.86681187e+00
1.76964656e-01 -8.21363926e-01 -1.35783255e+00 -3.00635129e-01
3.32068235e-01 -1.18163300e+00 -4.79162872e-01 -1.45627356e+00
-6.83155796e-03 -3.31524730e-01 -6.32435858e-01 4.45730358e-01
2.69417744e-02 2.31030449e-01 -2.89313018e-01 -1.27091274e-01
-4.73832160e-01 3.15919459e-01 8.92681718e-01 -2.87217528e-01
-1.18896700e-01 -2.68112849e-02 -6.56880021e-01 7.73713231e-01
8.18594933e-01 -5.60518086e-01 -6.37923360e-01 -2.42824242e-01
6.42672241e-01 -4.17246580e-01 4.70633596e-01 -1.00268042e+00
6.28985405e-01 3.83733541e-01 3.82621944e-01 -2.36790434e-01
1.67168811e-01 -9.94383872e-01 7.76859522e-02 5.78532934e-01
-8.23871046e-02 6.41262412e-01 8.97194222e-02 1.15625679e+00
-4.53917608e-02 2.07400098e-01 4.54363942e-01 -5.22786826e-02
-7.71331608e-01 1.03160405e+00 -4.59411025e-01 2.61572301e-01
1.15521288e+00 -1.71236724e-01 -5.87598920e-01 -1.84106559e-01
-1.19374931e+00 5.98812759e-01 1.54799893e-01 6.47089779e-01
6.55301034e-01 -1.50937140e+00 -3.62031519e-01 3.77503335e-01
1.72107711e-01 -1.79602444e-01 2.09799409e-01 9.99294937e-01
-2.86114275e-01 -2.14486703e-01 -4.66018826e-01 -6.56467676e-01
-1.19423652e+00 2.91232377e-01 5.55786669e-01 -3.25749576e-01
-8.74460459e-01 8.56342494e-01 -1.76677443e-02 -3.02618325e-01
2.50996083e-01 -2.51676053e-01 -3.50188881e-01 6.06106222e-01
4.45351928e-01 4.65849549e-01 -5.52874617e-03 -5.86296558e-01
-4.49397504e-01 1.47087142e-01 -2.52157062e-01 1.84373617e-01
1.44131374e+00 -1.13486871e-01 -6.81800783e-01 7.74718940e-01
1.09759986e+00 -4.24632221e-01 -8.26135218e-01 -6.21840060e-01
3.91880572e-01 -2.53087252e-01 -1.77965268e-01 -4.47584033e-01
-1.63368142e+00 6.61967933e-01 5.28760910e-01 7.71692157e-01
9.22227859e-01 2.28671536e-01 5.99487960e-01 6.12174451e-01
4.83831875e-02 -7.84575343e-01 3.31524849e-01 7.46170998e-01
6.45057261e-01 -1.01459301e+00 -1.36317447e-01 -3.44715744e-01
-2.07078636e-01 1.21004713e+00 9.66839254e-01 -1.85124308e-01
1.15798402e+00 2.39059106e-01 -3.53500336e-01 -8.30564439e-01
-9.27551448e-01 -1.36512639e-02 6.79095760e-02 6.10468090e-01
2.13015065e-01 -6.24499284e-02 7.02472776e-02 1.50380462e-01
-3.93241039e-03 -4.98167753e-01 8.97658825e-01 7.70870030e-01
-1.40845776e-01 -1.00942433e+00 1.28932938e-01 1.25223172e+00
-4.75991756e-01 -1.87515557e-01 -5.38965464e-01 6.68973744e-01
-8.45405310e-02 6.37857139e-01 4.56899256e-01 -4.96433437e-01
3.80529761e-01 3.24541807e-01 3.16988856e-01 -6.39257848e-01
-4.69129175e-01 -5.14275134e-01 -1.01820663e-01 -8.43660951e-01
-1.55182868e-01 -8.19079459e-01 -1.29636443e+00 -6.21680439e-01
5.55567211e-03 -2.72273988e-01 2.35003129e-01 8.64333451e-01
6.65155709e-01 1.05035698e+00 4.35641646e-01 -3.24454159e-01
2.29381844e-01 -8.68893921e-01 -8.38617802e-01 4.91549850e-01
6.50726557e-01 -8.33625376e-01 -4.47658211e-01 -3.69642735e-01] | [7.055962562561035, 5.997568130493164] |
9c8d7e99-6e3b-45a3-845c-f669f0a80b10 | refineloc-iterative-refinement-for-weakly | 1904.00227 | null | https://arxiv.org/abs/1904.00227v3 | https://arxiv.org/pdf/1904.00227v3.pdf | RefineLoc: Iterative Refinement for Weakly-Supervised Action Localization | Video action detectors are usually trained using datasets with fully-supervised temporal annotations. Building such datasets is an expensive task. To alleviate this problem, recent methods have tried to leverage weak labeling, where videos are untrimmed and only a video-level label is available. In this paper, we propose RefineLoc, a novel weakly-supervised temporal action localization method. RefineLoc uses an iterative refinement approach by estimating and training on snippet-level pseudo ground truth at every iteration. We show the benefit of this iterative approach and present an extensive analysis of five different pseudo ground truth generators. We show the effectiveness of our model on two standard action datasets, ActivityNet v1.2 and THUMOS14. RefineLoc shows competitive results with the state-of-the-art in weakly-supervised temporal localization. Additionally, our iterative refinement process is able to significantly improve the performance of two state-of-the-art methods, setting a new state-of-the-art on THUMOS14. | ['Alejandro Pardo', 'Humam Alwassel', 'Fabian Caba Heilbron', 'Bernard Ghanem', 'Ali Thabet'] | 2019-03-30 | null | null | null | null | ['weakly-supervised-action-localization', 'weakly-supervised-temporal-action'] | ['computer-vision', 'computer-vision'] | [ 2.33863905e-01 7.52040073e-02 -6.97263598e-01 -1.25552505e-01
-9.47226942e-01 -4.68170613e-01 6.59717619e-01 -2.10638165e-01
-5.86040556e-01 7.71179438e-01 4.58185256e-01 2.55399287e-01
2.62565345e-01 -2.70938188e-01 -8.14906120e-01 -6.15346074e-01
-3.33440870e-01 3.35202545e-01 1.02347422e+00 1.95633754e-01
-2.58680731e-02 -8.91076773e-02 -1.40105605e+00 7.86745250e-01
3.33963126e-01 9.87433910e-01 -2.73319986e-02 6.62053466e-01
3.26478511e-01 1.55861640e+00 -2.48115838e-01 -1.09671749e-01
2.69478530e-01 -7.80244112e-01 -9.87134933e-01 2.63860017e-01
5.64528465e-01 -4.21707302e-01 -5.56200743e-01 8.18261564e-01
1.61016658e-01 1.17090223e-02 2.67816514e-01 -1.41047204e+00
-1.36457726e-01 8.82639468e-01 -5.27884126e-01 2.29281843e-01
4.37762052e-01 4.16811347e-01 9.75658715e-01 -7.33841300e-01
9.86494541e-01 9.70380604e-01 9.47847843e-01 5.82390606e-01
-1.28965950e+00 -4.80574310e-01 3.64154577e-01 4.15665865e-01
-1.39033663e+00 -5.73829889e-01 5.14950514e-01 -5.27806401e-01
1.05015075e+00 -2.03714713e-01 6.91796303e-01 1.58181417e+00
-2.48375565e-01 1.27626741e+00 1.17530215e+00 -3.84500235e-01
3.87437582e-01 -2.13137224e-01 -1.15875997e-01 1.03032255e+00
-3.56559679e-02 -1.83925126e-02 -9.02959883e-01 8.62472132e-02
7.69822717e-01 -1.16598018e-01 -1.98036641e-01 -6.44076884e-01
-1.57893229e+00 5.02193511e-01 2.25472167e-01 5.50031245e-01
-1.46464393e-01 7.05005407e-01 5.49155831e-01 8.54277238e-02
5.71857274e-01 2.11726815e-01 -5.26803970e-01 -7.51314402e-01
-1.31071794e+00 9.47725177e-02 7.09937692e-01 9.23429310e-01
5.46601295e-01 -2.08681211e-01 -4.80845749e-01 3.84905338e-01
7.74887949e-02 1.46618575e-01 5.94399631e-01 -1.35601258e+00
5.44509053e-01 5.22372901e-01 2.73215234e-01 -4.33574051e-01
-4.53452110e-01 -2.17515379e-01 -3.46576065e-01 1.44315243e-01
7.24321365e-01 -1.09375440e-01 -1.08680177e+00 1.78465784e+00
2.35572278e-01 9.19311106e-01 -5.44536300e-02 8.37909520e-01
3.08488995e-01 3.24006349e-01 2.54873127e-01 -2.38793090e-01
9.54447210e-01 -1.59329426e+00 -6.52328849e-01 -2.01023310e-01
1.09405494e+00 -2.41799474e-01 9.36995208e-01 4.88760501e-01
-8.98952007e-01 -5.38856447e-01 -9.72824931e-01 1.93912372e-01
-1.72863543e-01 6.64471567e-01 6.61721706e-01 4.43027467e-01
-9.53907609e-01 8.26465607e-01 -1.57753956e+00 -6.38115048e-01
6.83416009e-01 4.50836755e-02 -7.29214251e-01 -6.20101094e-02
-8.79203081e-01 7.22489119e-01 6.43192708e-01 -1.89032584e-01
-1.38705182e+00 -4.92200762e-01 -9.14192259e-01 -2.90900797e-01
8.99696767e-01 -4.84908402e-01 1.61505437e+00 -1.06686461e+00
-1.50095427e+00 8.80513251e-01 -1.02213755e-01 -9.88596857e-01
9.77372706e-01 -5.55812538e-01 -2.39394292e-01 4.84680384e-01
4.02672172e-01 8.22819710e-01 7.77190566e-01 -7.44985759e-01
-8.37166369e-01 1.66931637e-02 1.81430057e-01 -9.76647511e-02
-1.80112883e-01 -1.09532930e-01 -7.01306522e-01 -6.45792663e-01
-1.42964676e-01 -1.18622243e+00 -2.90131450e-01 2.28160061e-02
-2.02553198e-01 -2.92160153e-01 8.91414881e-01 -3.86308461e-01
1.27369142e+00 -2.14000344e+00 6.48617297e-02 -3.35997880e-01
3.93881537e-02 2.59931833e-01 -1.70028239e-01 3.07999134e-01
2.96491235e-02 4.74215532e-03 -2.53806740e-01 -7.04867899e-01
8.29264075e-02 2.52297699e-01 1.01227192e-02 7.08088398e-01
2.30265126e-01 9.47772801e-01 -1.43335080e+00 -6.92039609e-01
3.09444785e-01 1.93342447e-01 -6.05708897e-01 1.29832491e-01
-5.79772413e-01 6.46323442e-01 -4.08117741e-01 7.85583198e-01
3.40067632e-02 -4.17618096e-01 2.30232939e-01 -2.51745135e-01
-1.92807078e-01 2.07984924e-01 -1.13078630e+00 2.42979026e+00
-2.56769329e-01 7.14126706e-01 -3.32823753e-01 -8.67381871e-01
1.54352054e-01 4.46714818e-01 9.44357395e-01 -3.77912313e-01
1.85483824e-02 2.05527857e-01 -2.11222038e-01 -6.24110758e-01
3.06390196e-01 1.52557015e-01 -7.39244819e-02 5.39907634e-01
5.50443649e-01 3.25147033e-01 6.81374013e-01 3.34643096e-01
1.71367931e+00 1.15032828e+00 3.32946271e-01 2.38610040e-02
4.45353329e-01 1.04197182e-01 6.53518617e-01 9.03605044e-01
-5.53409696e-01 6.88017190e-01 6.95446193e-01 -3.75496805e-01
-8.40709865e-01 -7.72999823e-01 2.75599122e-01 1.19676542e+00
1.86307635e-02 -9.89497781e-01 -9.89607334e-01 -1.35523796e+00
-2.58729190e-01 4.82979745e-01 -9.23340559e-01 6.57950640e-02
-6.27062917e-01 -2.93415844e-01 7.88450956e-01 8.87412071e-01
7.55932331e-01 -1.11823463e+00 -8.09669018e-01 2.68655807e-01
-3.85878801e-01 -1.67531109e+00 -4.29340780e-01 1.36965349e-01
-7.77986765e-01 -1.29658663e+00 -5.70120454e-01 -2.85479218e-01
4.31736797e-01 9.72105563e-02 1.17795181e+00 -1.15681320e-01
-9.29336548e-02 4.25625235e-01 -7.19683111e-01 2.97637265e-02
-2.91622460e-01 3.44410002e-01 9.73646417e-02 2.73829281e-01
4.01648313e-01 -5.36975801e-01 -5.53236365e-01 5.48995674e-01
-7.83899069e-01 1.36908293e-01 5.81296980e-01 5.44614196e-01
6.39909387e-01 -8.48900750e-02 2.08995789e-01 -8.43977034e-01
-1.35248095e-01 -3.48590523e-01 -6.12877429e-01 1.47356406e-01
-3.15860033e-01 3.30723524e-01 2.87820220e-01 -4.95378584e-01
-9.91775274e-01 7.60610461e-01 8.15886706e-02 -7.79515207e-01
-2.65898287e-01 3.62014323e-01 1.49858519e-02 -6.77354857e-02
8.95357430e-01 7.47731403e-02 -4.45471048e-01 -6.07647836e-01
5.95798910e-01 2.45035157e-01 7.53715098e-01 -4.15282100e-01
5.57552576e-01 8.94177854e-01 -8.29102471e-02 -4.63127762e-01
-1.35278118e+00 -7.98574626e-01 -1.06311369e+00 -4.45109159e-01
1.08828914e+00 -1.10327816e+00 -4.29161638e-01 6.23555303e-01
-8.90380621e-01 -9.73718166e-01 -4.66821134e-01 6.09243333e-01
-9.68599796e-01 4.24120337e-01 -5.53395867e-01 -5.75679600e-01
2.39173636e-01 -9.74937797e-01 1.31424665e+00 -2.28083313e-01
-2.98147470e-01 -8.52166176e-01 4.79540855e-01 4.47956115e-01
4.78127450e-02 4.90152448e-01 -1.84297517e-01 -5.68199813e-01
-7.51299739e-01 -1.41937286e-01 -6.19066972e-03 2.91146964e-01
7.24333972e-02 -2.08806217e-01 -1.09421813e+00 -1.21163495e-01
-3.42222899e-01 -8.12428236e-01 1.35574400e+00 2.70065248e-01
1.12990463e+00 -1.73653468e-01 -5.69908917e-01 4.91809636e-01
1.21529448e+00 -3.11772257e-01 6.14467502e-01 3.39879781e-01
7.52546012e-01 2.32179776e-01 1.06885910e+00 4.97585326e-01
2.30405822e-01 1.01868558e+00 4.18106943e-01 1.07459657e-01
-3.70056510e-01 -6.09072745e-01 7.44111598e-01 2.27980942e-01
-3.87544215e-01 -2.58067518e-01 -7.89511681e-01 8.46685886e-01
-2.49016213e+00 -1.35588646e+00 -8.25751498e-02 1.88302541e+00
8.73754680e-01 3.88497025e-01 4.29162174e-01 1.31759912e-01
3.91065568e-01 4.83000457e-01 -3.41795534e-01 4.89767611e-01
-4.22280617e-02 1.04164504e-01 7.97316551e-01 2.43570670e-01
-1.70608139e+00 1.41830277e+00 5.72286463e+00 7.48490572e-01
-8.41429055e-01 5.64903498e-01 2.53828317e-01 -3.26587707e-01
4.69798088e-01 2.45867506e-01 -7.27122068e-01 4.65792626e-01
9.94705200e-01 3.25668991e-01 1.44281119e-01 1.06987965e+00
4.89555538e-01 -5.55699110e-01 -1.50055015e+00 1.11854911e+00
2.08298326e-01 -1.45087373e+00 -5.13454676e-01 -1.44421205e-01
1.03510857e+00 4.56614196e-01 -3.86850119e-01 5.00810981e-01
4.24332917e-01 -7.81801820e-01 1.02475381e+00 4.42226201e-01
7.67904878e-01 -2.29266137e-01 6.74376130e-01 3.45763654e-01
-1.50273979e+00 -7.57258683e-02 1.53500333e-01 -1.33240104e-01
5.49681604e-01 3.21390063e-01 -5.57398260e-01 2.92774588e-01
7.81083882e-01 1.40623641e+00 -6.92368805e-01 1.10935056e+00
-7.73591101e-01 9.25392985e-01 -3.32703680e-01 3.52018565e-01
6.05904222e-01 2.34772399e-01 4.04275656e-01 1.45206654e+00
7.44742304e-02 4.79413709e-03 6.74633026e-01 4.60044742e-01
-1.67359382e-01 -3.52983773e-01 -5.28901935e-01 -2.99941897e-01
-1.04029514e-02 1.20327950e+00 -9.01532412e-01 -5.40861249e-01
-4.06263620e-01 1.24753094e+00 3.44449639e-01 1.66029900e-01
-1.20383751e+00 1.47761822e-01 3.61111671e-01 3.21464479e-01
4.94226009e-01 -3.06429744e-01 2.06892595e-01 -1.41732955e+00
4.76504527e-02 -7.44940042e-01 6.38337374e-01 -8.49231124e-01
-7.56501734e-01 3.33004266e-01 2.15493664e-01 -1.53201795e+00
-4.02583152e-01 -4.29899514e-01 -1.98101401e-01 1.60532203e-02
-1.24537611e+00 -1.38123298e+00 -4.45293605e-01 6.20760143e-01
8.74889255e-01 4.12602946e-02 5.14739156e-01 3.03289264e-01
-4.43247795e-01 2.85067916e-01 -3.37972313e-01 3.24786305e-01
8.54650974e-01 -1.28971708e+00 2.80064315e-01 1.13486254e+00
6.21962667e-01 8.39780495e-02 5.41236520e-01 -6.06108427e-01
-1.01074612e+00 -1.42838085e+00 7.08848476e-01 -8.67536366e-01
1.12193191e+00 -5.10107338e-01 -5.62337041e-01 1.07647288e+00
1.56655163e-01 4.50616837e-01 3.33264530e-01 -1.68565586e-01
-3.33434165e-01 2.18854323e-02 -8.11805964e-01 4.90395457e-01
1.65305221e+00 -6.12951100e-01 -5.02376258e-01 6.55106187e-01
5.76531410e-01 -5.28976440e-01 -6.53974295e-01 4.13307339e-01
6.02430105e-01 -1.21318424e+00 6.06428325e-01 -5.28014064e-01
3.45489800e-01 -5.99343002e-01 6.06317334e-02 -1.01155496e+00
-1.18126214e-01 -9.62704778e-01 -5.83521128e-01 1.01795638e+00
3.57751250e-01 -9.40186977e-02 1.16313601e+00 1.55345336e-01
-1.89990744e-01 -4.76191521e-01 -9.09938455e-01 -1.04635561e+00
-7.41216719e-01 -7.82183528e-01 -4.97562326e-02 8.14461410e-01
3.07284057e-01 2.34957322e-01 -6.59595072e-01 -9.37614813e-02
5.19461930e-01 -3.08073848e-01 8.88854384e-01 -7.77360857e-01
-4.91824865e-01 -8.18827003e-02 -7.10342526e-01 -1.19669318e+00
3.67350280e-01 -5.31288028e-01 2.67546982e-01 -1.40446007e+00
3.03384244e-01 -6.52172565e-02 -4.44068491e-01 1.09030843e+00
1.60546750e-01 6.40849471e-01 2.03479931e-01 2.47895196e-01
-1.43420529e+00 4.16765749e-01 8.41460288e-01 -1.55420914e-01
-1.78547829e-01 -8.99763182e-02 -2.75732838e-02 1.03733313e+00
6.51648819e-01 -7.29708195e-01 -6.51524603e-01 -2.87054867e-01
5.14472164e-02 -1.45699844e-01 6.15216076e-01 -1.48956835e+00
2.47652486e-01 -4.18333039e-02 -1.30020864e-02 -5.69530487e-01
4.36046481e-01 -7.74578810e-01 4.73367609e-02 5.30281603e-01
-4.23775494e-01 -2.26015747e-01 -3.95694841e-03 8.58974934e-01
-3.63088995e-01 -8.79089013e-02 7.89933801e-01 -3.47091168e-01
-1.11839247e+00 2.88960516e-01 -3.16643327e-01 1.82041228e-01
1.32034540e+00 -1.85529947e-01 -2.14132786e-01 -3.22487354e-01
-8.68550956e-01 1.45852730e-01 5.03467321e-01 4.16637599e-01
2.57891417e-01 -1.24368489e+00 -5.96011400e-01 -2.74631977e-01
3.91334116e-01 -3.54283810e-01 -5.86563200e-02 1.42287970e+00
-4.18451905e-01 5.22627473e-01 3.11222710e-02 -8.54736388e-01
-1.25412238e+00 5.17570615e-01 3.08964103e-01 -6.09745741e-01
-5.70068598e-01 8.09579968e-01 4.59775515e-02 8.90239924e-02
3.62868488e-01 -7.37052202e-01 -5.00642806e-02 3.06485519e-02
4.96652931e-01 4.42565411e-01 -4.90307584e-02 -5.83016753e-01
-5.79862654e-01 4.06910479e-01 1.06641591e-01 -3.88769150e-01
1.24054241e+00 1.28353089e-01 1.58809051e-01 5.00433862e-01
9.88635123e-01 -1.15315795e-01 -1.85836542e+00 -1.73187658e-01
3.36582601e-01 -4.97444630e-01 -5.09684905e-02 -7.61374474e-01
-9.86873448e-01 5.47598064e-01 5.49895227e-01 -5.70011772e-02
1.01433623e+00 2.16125637e-01 6.19200230e-01 4.11524802e-01
6.84193671e-01 -1.34257400e+00 5.39192319e-01 4.52381462e-01
5.65701008e-01 -1.33278406e+00 7.82439038e-02 -3.60268056e-01
-6.98770046e-01 7.48986363e-01 7.04186022e-01 -6.98327199e-02
4.35359120e-01 1.93676680e-01 -1.41858816e-01 -1.56905830e-01
-8.05794716e-01 -6.06761336e-01 1.28688067e-01 4.72407967e-01
3.81126761e-01 -2.11984321e-01 -3.04715782e-01 2.95713872e-01
4.19075161e-01 6.07877016e-01 3.82746041e-01 1.29027104e+00
-2.85667479e-01 -1.18851709e+00 -2.87815724e-02 2.63375267e-02
-5.50056994e-01 -1.64645109e-02 -5.02905786e-01 9.43801463e-01
3.60680640e-01 9.44921732e-01 -2.53974795e-01 -3.88408482e-01
1.43844575e-01 5.97450621e-02 6.36358917e-01 -7.08493471e-01
-3.04136336e-01 6.06285743e-02 4.84315455e-01 -1.34370852e+00
-1.18183470e+00 -8.77963543e-01 -1.17669916e+00 2.98663139e-01
-1.49737194e-01 5.15008047e-02 3.02490354e-01 1.12928605e+00
2.32006803e-01 3.86424989e-01 2.28200808e-01 -1.04388177e+00
-2.87340492e-01 -1.14258313e+00 -3.83080006e-01 5.58712602e-01
2.39415392e-01 -8.86503220e-01 -2.24316671e-01 6.47049785e-01] | [8.405200958251953, 0.546775221824646] |
1068477c-465e-4926-a311-5d97ca29d746 | trading-off-quality-for-efficiency-of | 2209.14825 | null | https://arxiv.org/abs/2209.14825v1 | https://arxiv.org/pdf/2209.14825v1.pdf | Trading off Quality for Efficiency of Community Detection: An Inductive Method across Graphs | Many network applications can be formulated as NP-hard combinatorial optimization problems of community detection (CD). Due to the NP-hardness, to balance the CD quality and efficiency remains a challenge. Most existing CD methods are transductive, which are independently optimized only for the CD on a single graph. Some of these methods use advanced machine learning techniques to obtain high-quality CD results but usually have high complexity. Other approaches use fast heuristic approximation to ensure low runtime but may suffer from quality degradation. In contrast to these transductive methods, we propose an alternative inductive community detection (ICD) method across graphs of a system or scenario to alleviate the NP-hard challenge. ICD first conducts the offline training of an adversarial dual GNN on historical graphs to capture key properties of the system. The trained model is then directly generalized to new unseen graphs for online CD without additional optimization, where a better trade-off between quality and efficiency can be achieved. ICD can also capture the permutation invariant community labels in the offline training and tackle the online CD on new graphs with non-fixed number of nodes and communities. Experiments on a set of benchmarks demonstrate that ICD can achieve a significant trade-off between quality and efficiency over various baselines. | ['Dit-yan Yeung', 'Gong Zhang', 'Bo Bai', 'Chaorui Zhang', 'Meng Qin'] | 2022-09-29 | null | null | null | null | ['community-detection'] | ['graphs'] | [ 2.05548689e-01 1.47854254e-01 -9.30898413e-02 3.08473650e-02
-8.83728027e-01 -9.99501109e-01 3.71697485e-01 3.72384071e-01
-2.74374746e-02 4.42986488e-01 -4.73084003e-01 -4.32911843e-01
-2.08637536e-01 -1.04644871e+00 -8.00912142e-01 -7.96097815e-01
-6.53592825e-01 7.44937181e-01 5.83202660e-01 4.91077714e-02
-3.95902731e-02 3.58198911e-01 -1.00330567e+00 -5.43803722e-02
9.74891186e-01 6.47821486e-01 2.76335571e-02 7.54933417e-01
1.37399450e-01 5.43268621e-01 -3.95352900e-01 -5.02310157e-01
6.14999652e-01 -3.87918919e-01 -8.64132941e-01 3.39078099e-01
1.81927800e-01 1.05259530e-01 -4.98069763e-01 1.32973528e+00
5.42987347e-01 -3.52536976e-01 3.43573898e-01 -1.77498460e+00
-4.72680122e-01 8.67675722e-01 -7.84385324e-01 -1.94186158e-02
2.89763749e-01 4.44334410e-02 1.31148386e+00 -6.18825972e-01
7.24508703e-01 1.15798056e+00 7.17149198e-01 1.89834803e-01
-1.72823453e+00 -7.92285562e-01 3.39873016e-01 1.72941759e-01
-1.44118476e+00 -2.04872806e-03 8.43930840e-01 -4.78657931e-01
4.50338393e-01 2.24142909e-01 7.08143890e-01 8.64815116e-01
-3.19890589e-01 5.48279107e-01 1.02203548e+00 -3.00512552e-01
3.86746079e-01 3.78481522e-02 1.17477849e-01 7.31196702e-01
6.46674395e-01 -8.59271139e-02 -6.43523335e-02 -3.83679628e-01
5.77531993e-01 1.88521221e-02 -3.63851279e-01 -8.50463212e-01
-9.79661226e-01 9.14601862e-01 8.02138686e-01 9.22764465e-02
2.80982945e-02 2.97629893e-01 6.03575408e-01 5.66616774e-01
1.43312946e-01 3.42656791e-01 -1.67507812e-01 2.86960393e-01
-6.26489043e-01 -8.03550780e-02 1.14839864e+00 1.14218044e+00
8.89405131e-01 -2.46898219e-01 -1.73762310e-02 5.13598204e-01
2.20158547e-02 2.88829386e-01 -2.08321676e-01 -4.60292816e-01
6.51870668e-01 8.80719960e-01 -2.82064021e-01 -1.53123641e+00
-1.24404185e-01 -7.56542563e-01 -1.07373810e+00 -1.14337891e-01
3.64703268e-01 -5.51940175e-03 -7.58643985e-01 1.65820825e+00
5.12267470e-01 3.25635791e-01 -2.86130339e-01 7.54623413e-01
2.22013488e-01 7.46157229e-01 -3.51400793e-01 -4.95255679e-01
1.13453507e+00 -1.12606382e+00 -2.61969954e-01 -2.97332078e-01
6.33533001e-01 -3.99762422e-01 8.08801115e-01 3.04860681e-01
-7.05697238e-01 -1.30523862e-02 -1.11162817e+00 3.60142052e-01
-1.31171927e-01 -2.11113527e-01 5.57748795e-01 8.17901909e-01
-1.13973999e+00 5.06322682e-01 -8.79419267e-01 -3.18646342e-01
3.36625397e-01 6.24950111e-01 -4.09400523e-01 -5.11218071e-01
-9.38202620e-01 2.71053076e-01 4.87871259e-01 2.16211602e-01
-1.26381326e+00 -4.68134105e-01 -7.68356204e-01 1.74880475e-01
1.13521135e+00 -4.75384325e-01 7.13983119e-01 -1.07006466e+00
-1.17138135e+00 6.74433529e-01 4.48238015e-01 -3.91012132e-01
5.77717543e-01 4.61942673e-01 -2.29847461e-01 3.92306119e-01
1.71331152e-01 3.48239177e-04 7.67327785e-01 -1.43050158e+00
-4.45488632e-01 -1.19994931e-01 5.09157717e-01 -7.52861276e-02
-6.06186628e-01 -1.42649591e-01 -7.82484472e-01 -4.04207498e-01
2.45507881e-01 -1.32423139e+00 -3.99958223e-01 7.92567655e-02
-6.25314772e-01 1.23740966e-02 9.02904868e-01 -4.87477183e-01
1.07910502e+00 -1.83374441e+00 4.13368583e-01 6.08655751e-01
6.01275384e-01 3.38662922e-01 -3.98749709e-01 7.38786280e-01
-2.76071597e-02 2.64392346e-01 -3.08181107e-01 -3.47119331e-01
-9.22624171e-02 3.17631394e-01 3.05435546e-02 6.79712355e-01
3.04097861e-01 6.66806936e-01 -1.14427233e+00 -5.01320899e-01
-2.73846239e-01 -6.84226975e-02 -7.16359556e-01 2.09145859e-01
-9.84183773e-02 1.57834828e-01 -1.80869311e-01 7.81415641e-01
8.67023587e-01 -8.24576914e-01 9.26102340e-01 -3.83713655e-02
4.99424666e-01 -1.25002444e-01 -1.51328778e+00 1.14279950e+00
-3.04737091e-01 2.70872682e-01 5.23535967e-01 -1.57674205e+00
5.34774780e-01 1.46777406e-01 3.43583941e-01 -3.65710974e-01
-5.89678576e-03 1.66895404e-01 1.71030998e-01 -2.81239599e-01
-2.34920923e-02 3.44149210e-02 -1.51092738e-01 5.34856021e-01
-2.71815713e-02 1.41975895e-01 1.63102701e-01 5.01864076e-01
1.62031651e+00 -4.81679052e-01 9.59990099e-02 -2.16076151e-01
4.97653037e-01 6.40530065e-02 7.80508816e-01 7.51581788e-01
-9.15257856e-02 4.35076058e-01 9.69758689e-01 -2.85024140e-02
-1.06712091e+00 -1.05659521e+00 2.85244286e-01 8.66584361e-01
4.11740273e-01 -4.16567951e-01 -6.69032454e-01 -9.68116224e-01
-6.46073818e-02 1.21661155e-02 -4.64215040e-01 -2.53746331e-01
-4.30696607e-01 -8.69509220e-01 3.04688692e-01 3.52300644e-01
2.26162747e-01 -5.70433378e-01 2.86618441e-01 2.38099039e-01
-2.28641614e-01 -1.50565469e+00 -5.65124094e-01 1.78278223e-01
-7.90158451e-01 -1.26074076e+00 -3.28669518e-01 -1.21212101e+00
1.15624011e+00 5.76066911e-01 9.60506558e-01 5.56823015e-01
-3.64767253e-01 2.44101435e-01 -5.02709270e-01 6.44709691e-02
-4.94270176e-01 3.28088552e-01 -3.76545005e-02 2.46828616e-01
-1.99334234e-01 -8.56538951e-01 -5.18220544e-01 3.70385259e-01
-9.24281359e-01 -1.47493869e-01 5.86196303e-01 1.03879046e+00
6.20938420e-01 5.64905822e-01 5.36812365e-01 -1.04584229e+00
5.67537487e-01 -6.41321003e-01 -1.04097152e+00 3.78408015e-01
-8.46129537e-01 1.09471306e-01 9.03316617e-01 -8.27522993e-01
-3.85542035e-01 1.50511563e-01 4.69123125e-01 -4.80511576e-01
5.13200641e-01 7.33845353e-01 -3.30656856e-01 -3.99545431e-01
5.45025527e-01 1.01445630e-01 2.08999263e-03 -4.58453633e-02
1.90794826e-01 2.91925132e-01 3.50301355e-01 -7.03357816e-01
1.47267663e+00 6.09350502e-01 1.59994155e-01 -4.13809001e-01
-7.70438135e-01 -7.28131533e-01 -4.60983515e-01 -2.16758415e-01
2.53594309e-01 -8.72984946e-01 -6.81546628e-01 3.25709134e-01
-6.70338809e-01 -2.60949671e-01 1.39406577e-01 1.27505019e-01
-9.76200998e-02 8.91128540e-01 -6.64583385e-01 -8.28289747e-01
-4.06433910e-01 -1.03736627e+00 7.95360804e-01 -2.10180283e-01
5.21669984e-01 -1.00865972e+00 -1.42975394e-02 2.24055916e-01
-3.98154780e-02 6.26472831e-01 1.02653384e+00 -4.70044464e-01
-8.74785066e-01 -5.49064755e-01 -5.36742330e-01 3.03286701e-01
2.44223159e-02 4.31439504e-02 -6.61830664e-01 -9.88368571e-01
-3.33091021e-01 -3.07472020e-01 5.70016742e-01 -1.53022185e-01
1.15146756e+00 -6.94192708e-01 -5.26516795e-01 4.10575122e-01
1.90914476e+00 -2.14149535e-01 6.31852269e-01 1.23255029e-01
8.38398993e-01 3.15985799e-01 6.77750647e-01 1.87338963e-01
3.22211176e-01 4.49093372e-01 8.31148207e-01 -1.97953805e-01
2.33201578e-01 -2.15005532e-01 4.39802021e-01 1.13645685e+00
5.12577817e-02 -3.51504117e-01 -1.08971500e+00 8.03374887e-01
-2.05586815e+00 -9.06120777e-01 -2.71293491e-01 2.35086584e+00
8.66661549e-01 4.41947997e-01 4.62287664e-01 3.75856102e-01
1.18037653e+00 -2.71004528e-01 -4.14152831e-01 -5.71899191e-02
8.20654407e-02 7.45655522e-02 6.12378895e-01 2.71412134e-01
-8.31317842e-01 7.09652603e-01 5.59482098e+00 1.00807226e+00
-8.85990500e-01 2.42206126e-01 3.81806463e-01 1.28999919e-01
-1.77383810e-01 6.46779239e-01 -4.56272542e-01 2.90250093e-01
7.77156472e-01 -3.66289109e-01 8.60006154e-01 9.75326419e-01
-2.25919753e-01 1.52441114e-01 -1.26422381e+00 7.49866486e-01
-1.66228473e-01 -1.10293257e+00 -2.70040631e-01 4.21620339e-01
1.01823306e+00 -2.67517883e-02 -2.66682535e-01 4.23620343e-01
3.86535138e-01 -8.00351799e-01 4.07428503e-01 1.70906842e-01
7.82130897e-01 -7.69529700e-01 7.14436412e-01 6.34416997e-01
-1.46654832e+00 -3.27051908e-01 -4.12865371e-01 1.08067192e-01
-4.75835018e-02 8.36031795e-01 -1.13635123e+00 7.99078643e-01
6.65827572e-01 5.05877137e-01 -7.14458287e-01 1.20950556e+00
-2.27373689e-02 7.36677289e-01 -6.14134908e-01 -8.72975662e-02
1.11595832e-01 -3.38498652e-01 8.30688715e-01 1.04825711e+00
1.14804886e-01 -2.38957345e-01 6.24255240e-01 7.58400679e-01
-3.88859034e-01 4.36924212e-02 -5.79573452e-01 -4.25472260e-01
6.34324610e-01 1.55769026e+00 -9.76337016e-01 -1.34739354e-01
-2.67649591e-01 9.61013675e-01 6.21748269e-01 5.26185110e-02
-1.07934296e+00 -2.94319719e-01 2.20371351e-01 4.13632035e-01
4.91126269e-01 -2.42701992e-01 1.85756281e-01 -1.18163645e+00
2.90716410e-01 -1.11240983e+00 5.89588761e-01 -8.92181024e-02
-1.59680963e+00 4.69996750e-01 -1.79120079e-01 -1.32559180e+00
1.93701655e-01 -3.28774661e-01 -5.42508721e-01 2.79776037e-01
-1.28954196e+00 -1.28819215e+00 -4.57995534e-01 6.04042768e-01
2.51942631e-02 1.66432172e-01 5.30214608e-01 5.99880219e-01
-7.60209382e-01 7.88992465e-01 1.66099206e-01 1.86613336e-01
6.01562083e-01 -1.62300408e+00 2.45113477e-01 1.17700982e+00
-1.43180534e-01 3.52311194e-01 5.20714521e-01 -5.63718975e-01
-1.59997559e+00 -1.54868948e+00 4.35964316e-01 -1.55789688e-01
8.80363703e-01 -8.61248255e-01 -8.75307441e-01 5.23667753e-01
-2.15883836e-01 4.83070523e-01 4.08532411e-01 2.35825256e-01
-6.44559264e-01 -4.29172546e-01 -1.14303839e+00 5.56233883e-01
1.37021589e+00 -6.17603898e-01 -3.42084281e-02 7.07432508e-01
9.73786235e-01 -8.53296891e-02 -1.05839753e+00 2.92808115e-01
4.81064292e-03 -5.67773044e-01 1.05443597e+00 -4.01024908e-01
1.65985748e-01 -6.42527819e-01 1.53715521e-01 -1.05169213e+00
-3.89286101e-01 -9.06939209e-01 -3.08801562e-01 1.41956675e+00
3.14340830e-01 -8.01599920e-01 6.31148160e-01 1.70291379e-01
2.53217429e-01 -5.44948697e-01 -8.23322177e-01 -1.16525495e+00
-1.57013938e-01 -1.58620030e-01 4.85354751e-01 1.27999294e+00
-6.11770526e-02 6.80036962e-01 -3.28333467e-01 6.59265637e-01
9.53518569e-01 4.76141214e-01 9.21614766e-01 -1.43766713e+00
-5.75137615e-01 -2.00944126e-01 -6.09149337e-01 -4.88505989e-01
-2.98988279e-02 -1.16028750e+00 1.33690436e-03 -1.35247946e+00
6.63421154e-01 -7.77735651e-01 -2.04904587e-03 6.40982151e-01
-1.13151617e-01 1.59188360e-01 1.05781384e-01 2.45153919e-01
-9.42787647e-01 3.42389435e-01 1.21998572e+00 -2.85855263e-01
-2.55463302e-01 4.27776612e-02 -5.74490726e-01 2.83158988e-01
5.14291286e-01 -9.31481302e-01 -4.40186173e-01 -1.31339416e-01
5.77285469e-01 2.78178364e-01 4.58874553e-01 -9.79705691e-01
5.44536591e-01 -1.46462277e-01 -3.35864782e-01 -3.83148491e-01
-9.45335478e-02 -1.08674312e+00 3.47499877e-01 7.66124666e-01
1.46797851e-01 -6.65155202e-02 -5.13024367e-02 1.36542118e+00
-1.09535925e-01 -1.72041401e-01 7.69797564e-01 1.28821447e-01
-2.15691969e-01 7.34331846e-01 -1.06474586e-01 1.70161217e-01
1.33862031e+00 -2.17562258e-01 -3.45674217e-01 -4.35153902e-01
-7.60698378e-01 7.48506010e-01 5.09974837e-01 2.26659134e-01
2.29152963e-01 -1.27464628e+00 -6.94729328e-01 -2.63050318e-01
2.61470139e-01 2.77816623e-01 -1.53853493e-02 1.06440103e+00
-6.42484248e-01 -9.38134640e-02 1.26945153e-01 -7.77370870e-01
-1.26293945e+00 1.05966270e+00 3.56877118e-01 -7.31145799e-01
-5.24312139e-01 6.43158376e-01 1.54737592e-01 -6.65093958e-01
2.63745904e-01 1.38565436e-01 2.64108092e-01 2.44906880e-02
1.35426596e-02 2.42856666e-01 1.35406613e-01 -2.41116464e-01
-4.14547324e-01 3.22045177e-01 -5.54132275e-02 2.81706214e-01
1.48137665e+00 -1.21406406e-01 -4.30439740e-01 2.70420890e-02
1.34564817e+00 2.13397499e-02 -9.36004937e-01 -4.89950448e-01
-4.87738196e-03 -4.39019978e-01 -8.79175663e-02 -4.04380172e-01
-1.45425999e+00 6.00399494e-01 4.22113001e-01 6.73670948e-01
1.35934889e+00 -9.52271000e-02 6.53367817e-01 4.03834134e-01
5.69053173e-01 -1.08284438e+00 3.73562276e-01 1.74856037e-01
7.25101233e-01 -1.09045315e+00 1.09374091e-01 -1.04100192e+00
-2.69448847e-01 1.05582964e+00 5.66006482e-01 -2.73974806e-01
7.15267062e-01 2.08876058e-01 -6.83460832e-01 -3.10779721e-01
-6.70180619e-01 -9.38103795e-02 6.71783313e-02 6.21304035e-01
-3.65024000e-01 1.24787591e-01 -7.97477365e-02 5.58496296e-01
2.02963859e-01 -4.04853433e-01 6.25861228e-01 8.07333648e-01
-1.50857285e-01 -1.32323420e+00 -2.48542354e-01 3.57461154e-01
-2.20999435e-01 1.00611135e-01 -5.38792253e-01 7.58628249e-01
-2.08574794e-02 9.24446821e-01 -3.04398686e-01 -4.87864286e-01
1.34195536e-01 -5.07125974e-01 3.54868561e-01 -8.02198350e-01
-6.00135982e-01 7.86476955e-02 2.43475169e-01 -4.63761866e-01
-2.90097028e-01 -4.33851123e-01 -9.34269309e-01 -5.11037946e-01
-1.10653293e+00 1.46196947e-01 2.52255499e-01 6.88494444e-01
3.37639481e-01 4.01285201e-01 1.17410922e+00 -3.91421854e-01
-8.92296195e-01 -6.65104389e-01 -6.20057583e-01 3.40530932e-01
2.32271448e-01 -4.93505299e-01 -7.51999259e-01 -1.67671382e-01] | [7.190313816070557, 5.98133659362793] |
f3e27ace-c795-44a4-ba09-8be36c041faa | generative-adversarial-training-for-weakly | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Zou_Generative_Adversarial_Training_for_Weakly_Supervised_Cloud_Matting_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Zou_Generative_Adversarial_Training_for_Weakly_Supervised_Cloud_Matting_ICCV_2019_paper.pdf | Generative Adversarial Training for Weakly Supervised Cloud Matting | The detection and removal of cloud in remote sensing images are essential for earth observation applications. Most previous methods consider cloud detection as a pixel-wise semantic segmentation process (cloud v.s. background), which inevitably leads to a category-ambiguity problem when dealing with semi-transparent clouds. We re-examine the cloud detection under a totally different point of view, i.e. to formulate it as a mixed energy separation process between foreground and background images, which can be equivalently implemented under an image matting paradigm with a clear physical significance. We further propose a generative adversarial framework where the training of our model neither requires any pixel-wise ground truth reference nor any additional user interactions. Our model consists of three networks, a cloud generator G, a cloud discriminator D, and a cloud matting network F, where G and D aim to generate realistic and physically meaningful cloud images by adversarial training, and F learns to predict the cloud reflectance and attenuation. Experimental results on a global set of satellite images demonstrate that our method, without ever using any pixel-wise ground truth during training, achieves comparable and even higher accuracy over other fully supervised methods, including some recent popular cloud detectors and some well-known semantic segmentation frameworks.
| [' Jieping Ye', ' Zhenwei Shi', ' Tianyang Shi', ' Wenyuan Li', 'Zhengxia Zou'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['cloud-detection'] | ['computer-vision'] | [ 7.11043179e-01 -1.21525936e-01 5.50846338e-01 -2.80189902e-01
-5.72724283e-01 -9.19300914e-01 6.64646447e-01 -3.35746706e-01
-1.33418739e-01 5.20442724e-01 -7.85729170e-01 -5.19363403e-01
2.65602589e-01 -1.18663752e+00 -8.31018984e-01 -1.18784273e+00
2.88543075e-01 5.46145976e-01 2.59774297e-01 -3.09209935e-02
-2.72239111e-02 5.02292275e-01 -1.42088866e+00 -1.02596417e-01
1.58815527e+00 8.77717078e-01 3.81403625e-01 8.03903043e-01
-2.54121184e-01 5.42201221e-01 -5.81345975e-01 -2.96589792e-01
5.76449037e-01 -5.08213162e-01 -7.62842119e-01 7.11831510e-01
6.33868992e-01 -3.49533968e-02 8.23847950e-02 1.66148686e+00
2.16500878e-01 -1.69029273e-02 7.08303332e-01 -1.27608669e+00
-7.16781080e-01 1.88567102e-01 -6.17773831e-01 7.19865263e-02
-2.92143047e-01 3.43313724e-01 5.81403255e-01 -8.37132692e-01
2.76184380e-01 8.65408003e-01 6.47195220e-01 3.38661581e-01
-1.12079275e+00 -6.22496545e-01 3.85221839e-01 -2.53879100e-01
-1.56913447e+00 -7.97559023e-02 8.28124046e-01 -8.11805904e-01
3.36169124e-01 5.85410416e-01 8.66150618e-01 6.04679704e-01
-8.76807049e-02 5.20036578e-01 1.49110472e+00 -2.92683810e-01
3.96794587e-01 3.65256310e-01 -8.02203342e-02 2.73577929e-01
5.51468909e-01 -7.39220669e-03 2.22355083e-01 -2.13632397e-02
8.32344711e-01 1.18087120e-02 -5.49850464e-01 -2.94957459e-01
-6.80735826e-01 8.83832932e-01 6.38526976e-01 7.36497492e-02
-4.45092291e-01 1.27146348e-01 -3.54130685e-01 1.35225698e-01
9.46502507e-01 2.08863303e-01 -1.06880300e-01 8.11032057e-01
-1.16831613e+00 3.49420905e-01 6.15533710e-01 9.94859695e-01
8.51800978e-01 6.09574497e-01 1.11186840e-01 3.02423716e-01
4.16959137e-01 1.16313589e+00 -3.74089219e-02 -7.52012074e-01
1.61687478e-01 4.01846975e-01 5.75284123e-01 -8.33930612e-01
2.23988801e-01 -9.03747916e-01 -8.04000974e-01 5.42141676e-01
1.53161481e-01 -1.96330458e-01 -1.23063064e+00 1.36316609e+00
5.71343780e-01 6.68964863e-01 1.82035536e-01 1.18905699e+00
6.50202870e-01 6.97009623e-01 -1.95419062e-02 -6.73858076e-02
1.13811147e+00 -9.44972456e-01 -5.38387060e-01 -6.43659234e-01
-1.54947892e-01 -6.36182487e-01 7.76913285e-01 2.74856925e-01
-9.57787156e-01 -3.73620778e-01 -9.73237991e-01 3.41675192e-01
-6.80101395e-01 -5.91859221e-02 7.65231311e-01 8.42251718e-01
-1.10618603e+00 3.98791939e-01 -8.03359091e-01 -4.98882309e-02
3.24346185e-01 1.13684252e-01 1.83002606e-01 1.95119143e-01
-9.94071603e-01 8.08091164e-01 2.68198252e-01 5.59690416e-01
-1.21982908e+00 -4.45690721e-01 -5.94927073e-01 -1.77663099e-02
4.11116481e-01 -1.01618147e+00 1.06471765e+00 -1.75632811e+00
-1.25606430e+00 1.18150616e+00 -7.33568072e-02 -4.29105163e-01
7.64514983e-01 -1.84218004e-01 -3.22678775e-01 3.21956962e-01
9.40298885e-02 3.97732645e-01 1.41195393e+00 -2.07320309e+00
-5.23873925e-01 -1.80199623e-01 9.77710783e-02 3.67697835e-01
1.92608386e-01 -7.77551085e-02 -3.91726106e-01 -7.21131742e-01
3.27499002e-01 -1.04252136e+00 -3.69271219e-01 8.79960805e-02
-5.28367221e-01 4.96294260e-01 9.64812756e-01 -7.81605065e-01
5.17597258e-01 -1.67131138e+00 5.57876453e-02 2.43508741e-01
2.88972557e-01 4.30219352e-01 -5.70818298e-02 -9.21959952e-02
-2.22986624e-01 3.32581937e-01 -1.09125566e+00 -3.56291920e-01
-2.95252264e-01 3.53375226e-01 -5.56083620e-01 6.45408094e-01
5.90995133e-01 9.08514321e-01 -1.13629699e+00 -3.34909528e-01
3.54813665e-01 6.82235837e-01 -3.12984645e-01 3.93753827e-01
-5.74357510e-01 8.08973551e-01 -4.61565763e-01 7.82736242e-01
1.43783665e+00 -2.63544470e-01 4.07016911e-02 2.87184924e-01
-2.16212720e-01 -1.82204247e-01 -1.08340347e+00 1.08289814e+00
-2.97717988e-01 4.27129060e-01 5.29970706e-01 -9.54900205e-01
7.19321191e-01 9.49506238e-02 2.11716190e-01 -2.73994118e-01
1.56518281e-01 3.03891867e-01 -1.98305383e-01 -3.85522336e-01
4.86092597e-01 -3.94784540e-01 3.32805812e-01 1.60283893e-01
-3.89885128e-01 -8.82741213e-01 -4.66836751e-01 7.40907490e-02
4.26543236e-01 1.78721249e-01 -2.66999662e-01 -2.35098377e-01
4.55815911e-01 1.99883536e-01 4.75508332e-01 9.00887966e-01
1.28086224e-01 9.85171974e-01 -4.94657904e-02 -2.69788861e-01
-8.67969573e-01 -1.12686968e+00 -1.25843525e-01 5.32431424e-01
5.75461864e-01 4.30230111e-01 -1.06896448e+00 -4.85263318e-01
1.36694256e-02 6.45498335e-01 -5.42399287e-01 8.02890435e-02
-3.05580735e-01 -1.22488308e+00 2.71122575e-01 1.80834219e-01
7.92916894e-01 -8.57931972e-01 -4.39541221e-01 -4.90988344e-02
-3.50341141e-01 -1.44070685e+00 -2.35866070e-01 7.47912452e-02
-9.09769893e-01 -9.42941248e-01 -6.03226364e-01 -5.14748335e-01
9.51253772e-01 7.97599256e-01 1.34460735e+00 4.60006088e-01
-1.86088741e-01 2.77398050e-01 -3.30823213e-01 -7.55287111e-01
-4.84655559e-01 -2.43353069e-01 -3.38551521e-01 4.79660779e-01
4.49921191e-02 -6.15018129e-01 -7.04359233e-01 1.15358405e-01
-1.31749880e+00 1.82786614e-01 5.20326138e-01 5.22882164e-01
7.56891847e-01 3.60257149e-01 -3.99147905e-02 -1.22026312e+00
-3.07365041e-02 -5.36900103e-01 -9.71801579e-01 2.08942726e-01
-6.30853474e-01 -4.00903255e-01 1.68016195e-01 -1.19566530e-01
-1.16819143e+00 3.18903923e-01 1.10183515e-01 -7.21457124e-01
-1.33488655e-01 3.18007410e-01 -5.73607326e-01 -5.52867234e-01
3.96112412e-01 6.84904099e-01 -3.52255076e-01 -2.26072744e-01
5.22885084e-01 6.48471773e-01 7.61603415e-01 -4.44561422e-01
1.59244347e+00 1.01687467e+00 -2.00482592e-01 -9.62535918e-01
-8.16469014e-01 -5.24321914e-01 -7.58329570e-01 -4.13518310e-01
1.13910329e+00 -1.21236384e+00 -1.79673672e-01 8.69538069e-01
-1.24466550e+00 -4.20656741e-01 -1.33819446e-01 1.06642872e-01
-2.06764042e-01 4.26129103e-01 -1.71767816e-01 -1.18952131e+00
-5.43786407e-01 -9.53488469e-01 1.23359787e+00 2.75880128e-01
7.48561978e-01 -1.07427371e+00 -2.82002002e-01 5.08868992e-01
4.98823345e-01 8.36305261e-01 5.09594738e-01 -7.05615133e-02
-1.27316415e+00 1.18420027e-01 -4.65457737e-01 8.36857677e-01
2.08537221e-01 2.44832590e-01 -1.56836939e+00 -2.30446279e-01
4.34423357e-01 2.33947724e-01 1.05379784e+00 3.26315999e-01
1.14610064e+00 -5.21344960e-01 -2.24656135e-01 1.03843141e+00
1.92966473e+00 9.80841368e-02 6.76262677e-01 1.74191251e-01
1.08805990e+00 4.45430636e-01 3.39912206e-01 1.19883753e-01
2.38660038e-01 3.54871869e-01 1.18650436e+00 -6.00802898e-01
1.62345245e-02 2.22600251e-01 3.10007017e-02 5.44302642e-01
-3.95459145e-01 -4.93893057e-01 -9.39184010e-01 5.76875210e-01
-1.65918839e+00 -9.60896134e-01 -5.09948611e-01 2.28762078e+00
7.97952652e-01 1.03067197e-02 -3.19470525e-01 -1.33083016e-01
8.88899028e-01 2.24369496e-01 -6.70601666e-01 7.16523975e-02
-3.64241302e-01 4.35842961e-01 9.14825916e-01 7.22508252e-01
-1.46001136e+00 1.10129952e+00 6.01438761e+00 4.42038536e-01
-1.52088535e+00 5.30420661e-01 5.32965362e-01 4.01474446e-01
-6.35945857e-01 2.92942494e-01 -3.28357279e-01 5.13914049e-01
3.62641871e-01 1.03177831e-01 4.95596111e-01 7.80812025e-01
3.41111600e-01 -1.33542404e-01 -5.31392217e-01 8.07076991e-01
6.56815544e-02 -1.11405325e+00 3.05701166e-01 -1.04923323e-02
9.51594591e-01 1.43898442e-01 7.23783597e-02 -1.23409212e-01
4.15974617e-01 -1.01720560e+00 1.11411524e+00 7.61227071e-01
5.88006794e-01 -3.60439569e-01 5.85243464e-01 4.42783087e-01
-1.24264216e+00 3.61902326e-01 -3.12145293e-01 -6.15582988e-02
-1.27712339e-01 9.19216990e-01 -5.04813790e-01 9.15358603e-01
4.84962553e-01 3.68212521e-01 -4.94568795e-01 1.07413733e+00
-5.19922197e-01 7.14462340e-01 -2.74697721e-01 5.67150056e-01
2.32095912e-01 -7.47345328e-01 7.65934169e-01 1.20234025e+00
1.33736089e-01 2.75722086e-01 3.12463105e-01 1.34713709e+00
3.68210599e-02 -2.02572644e-01 -4.03026462e-01 6.52514324e-02
3.49831253e-01 1.35553944e+00 -1.07334077e+00 -4.47469622e-01
-2.58248180e-01 1.38010180e+00 -2.46880412e-01 5.15995860e-01
-1.15775657e+00 -1.76453248e-01 7.86200047e-01 2.15604037e-01
2.77276576e-01 -1.27663597e-01 -4.40540582e-01 -1.44056737e+00
6.23308495e-02 -7.06491351e-01 -1.59454480e-01 -1.16711652e+00
-1.30343866e+00 7.10636497e-01 -1.74786195e-01 -1.40831232e+00
2.30666772e-01 -5.94489217e-01 -1.03258991e+00 1.31157374e+00
-2.14936113e+00 -1.47110569e+00 -9.38517153e-01 5.20890892e-01
2.99988627e-01 4.38160181e-01 6.53610051e-01 2.54362404e-01
-4.72010195e-01 6.88345730e-02 1.94543049e-01 2.28673443e-01
1.62965044e-01 -1.58270741e+00 5.40846705e-01 1.49861729e+00
-2.73313969e-02 3.51832718e-01 9.58694994e-01 -6.60501182e-01
-1.22300923e+00 -1.82602084e+00 3.79653186e-01 -5.79502940e-01
3.91714543e-01 -5.12962699e-01 -1.12455893e+00 6.14062369e-01
1.84836328e-01 2.58774251e-01 1.63593531e-01 -7.65522718e-01
-9.55712497e-02 -6.04099482e-02 -1.25140190e+00 1.49594933e-01
8.49958658e-01 -5.88996410e-01 -4.15702730e-01 7.63447344e-01
8.23955059e-01 -6.39669716e-01 -3.36521894e-01 4.06688482e-01
5.42369857e-02 -9.01998699e-01 1.11151564e+00 -3.63584578e-01
3.40223759e-01 -7.65724540e-01 -7.32080871e-03 -1.19209099e+00
-7.68687427e-02 -3.90725017e-01 3.18757117e-01 1.16923749e+00
1.89089358e-01 -7.78365910e-01 6.10597312e-01 3.83995295e-01
-2.57368505e-01 -1.50738105e-01 -5.82450986e-01 -9.68504488e-01
2.52188683e-01 -4.02634829e-01 6.99705660e-01 1.42094016e+00
-1.14094186e+00 -6.14982098e-02 -2.17028677e-01 1.13538587e+00
9.21170473e-01 8.59522164e-01 7.53292918e-01 -1.35698450e+00
-3.35725874e-01 -3.89945447e-01 8.37587658e-03 -7.00347126e-01
2.62317117e-02 -8.74552488e-01 3.59141499e-01 -1.53976500e+00
1.71700016e-01 -7.09416926e-01 1.04931317e-01 1.96510836e-01
-4.95093137e-01 4.57878470e-01 8.36747214e-02 4.45668668e-01
-1.05961882e-01 5.43647587e-01 1.33147919e+00 -4.80777562e-01
-2.49060616e-02 2.85643488e-01 -5.79769373e-01 8.93827975e-01
7.77556002e-01 -7.32942760e-01 -3.14108223e-01 -8.36281896e-01
1.13361843e-01 -1.54435560e-01 9.44454312e-01 -9.00986671e-01
-5.44397235e-02 -5.73572397e-01 3.35517406e-01 -3.24779481e-01
2.27846652e-01 -8.60478878e-01 5.15781462e-01 3.98599356e-01
3.75351876e-01 -3.14033896e-01 7.60486498e-02 5.30003130e-01
-2.65135109e-01 -2.78225720e-01 1.00504410e+00 -5.52069902e-01
-5.70702851e-01 4.74406749e-01 -2.55735159e-01 -3.75674702e-02
1.00987256e+00 -2.72721767e-01 -2.26677641e-01 -2.65500516e-01
-8.73888910e-01 1.17399506e-01 7.26008594e-01 1.35608450e-01
4.38364446e-01 -8.56670320e-01 -7.94142246e-01 1.11233748e-01
-8.61069001e-03 5.17454505e-01 1.29823357e-01 5.27655840e-01
-8.06029916e-01 -1.34396151e-01 2.45875865e-01 -8.67372513e-01
-9.42808449e-01 4.22329724e-01 1.01397777e+00 5.31277694e-02
-4.91794765e-01 9.17032123e-01 5.26168764e-01 -5.25176764e-01
-2.69548357e-01 -4.83135074e-01 1.30275726e-01 -1.81809857e-01
1.40322596e-01 -6.56934455e-02 1.83506608e-01 -6.67943418e-01
-2.49187931e-01 5.54693580e-01 5.71571231e-01 1.47556374e-02
1.20143211e+00 -1.82875320e-01 -3.92154396e-01 2.71049529e-01
4.50642496e-01 1.88606754e-01 -1.26803601e+00 -2.74956226e-01
-3.55201870e-01 -6.94742799e-01 2.86491424e-01 -7.49984443e-01
-1.64391565e+00 9.92283642e-01 6.82111859e-01 4.11713213e-01
1.30058038e+00 -1.65891543e-01 5.36805332e-01 4.84804297e-03
3.69904280e-01 -7.05142617e-01 -2.71892965e-01 3.57441664e-01
7.69697249e-01 -1.44813228e+00 -4.89103273e-02 -8.61938953e-01
-4.14967805e-01 7.70955145e-01 5.69139242e-01 -2.42696419e-01
6.40042722e-01 2.41098687e-01 2.96355397e-01 -2.68825114e-01
-1.66795954e-01 -6.05501473e-01 1.63344845e-01 6.54413164e-01
-7.97750056e-02 4.50643033e-01 1.40017360e-01 4.49853987e-01
-2.12432407e-02 -2.71004319e-01 6.13982558e-01 8.74251008e-01
-5.46786785e-01 -8.45326185e-01 -7.48944521e-01 1.87202260e-01
-3.43475789e-01 -4.21989053e-01 -4.50637132e-01 6.59898520e-01
6.10411406e-01 9.18615937e-01 1.68813080e-01 -2.11342528e-01
-5.12807071e-02 -5.37433885e-02 2.22168952e-01 -8.09364974e-01
-5.61444461e-01 1.55076683e-01 -4.46089923e-01 -1.75516456e-01
-8.75391185e-01 -6.50694430e-01 -1.02794993e+00 -2.12794885e-01
-7.94092774e-01 1.69668227e-01 8.05829227e-01 8.46408606e-01
-4.92422208e-02 6.12126172e-01 9.03613150e-01 -1.07496703e+00
-2.66626388e-01 -8.56862426e-01 -7.39580154e-01 2.77861208e-01
5.15836656e-01 -5.34974873e-01 -6.92611396e-01 4.16485608e-01] | [9.817062377929688, -1.7739253044128418] |
c79cd6f7-0bcb-4499-b326-4ae45a168ed5 | point-cloud-augmentation-with-weighted-local-1 | 2110.05379 | null | https://arxiv.org/abs/2110.05379v1 | https://arxiv.org/pdf/2110.05379v1.pdf | Point Cloud Augmentation with Weighted Local Transformations | Despite the extensive usage of point clouds in 3D vision, relatively limited data are available for training deep neural networks. Although data augmentation is a standard approach to compensate for the scarcity of data, it has been less explored in the point cloud literature. In this paper, we propose a simple and effective augmentation method called PointWOLF for point cloud augmentation. The proposed method produces smoothly varying non-rigid deformations by locally weighted transformations centered at multiple anchor points. The smooth deformations allow diverse and realistic augmentations. Furthermore, in order to minimize the manual efforts to search the optimal hyperparameters for augmentation, we present AugTune, which generates augmented samples of desired difficulties producing targeted confidence scores. Our experiments show our framework consistently improves the performance for both shape classification and part segmentation tasks. Particularly, with PointNet++, PointWOLF achieves the state-of-the-art 89.7 accuracy on shape classification with the real-world ScanObjectNN dataset. | ['Hyunwoo J. Kim', 'Seong Jae Hwang', 'Jaewon Lee', 'Dasol Hwang', 'Sanghyeok Lee', 'Sihyeon Kim'] | 2021-10-11 | point-cloud-augmentation-with-weighted-local | http://openaccess.thecvf.com//content/ICCV2021/html/Kim_Point_Cloud_Augmentation_With_Weighted_Local_Transformations_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Kim_Point_Cloud_Augmentation_With_Weighted_Local_Transformations_ICCV_2021_paper.pdf | iccv-2021-1 | ['point-cloud-classification'] | ['computer-vision'] | [ 5.96995577e-02 2.37934798e-01 -2.30725557e-01 -5.07511199e-01
-7.68946409e-01 -4.51378673e-01 6.07341170e-01 -6.10666387e-02
-1.52390882e-01 2.21222281e-01 -2.70282984e-01 -2.26231158e-01
3.02466094e-01 -6.04102731e-01 -9.75976169e-01 -6.05729043e-01
3.25026691e-01 8.21762323e-01 2.64895886e-01 -3.95871466e-03
1.67103350e-01 1.14037740e+00 -1.32006991e+00 -2.24783823e-01
1.09507346e+00 1.00581539e+00 2.32405528e-01 2.07395479e-01
-4.78695810e-01 -2.69882649e-01 -2.93412864e-01 -5.20356357e-01
6.24958336e-01 2.64812797e-01 -5.35407007e-01 4.70638812e-01
8.49612594e-01 -4.06407714e-01 -1.52788848e-01 1.00981486e+00
5.10713637e-01 2.77057439e-02 5.11956394e-01 -1.34121597e+00
-7.70177960e-01 3.68814260e-01 -9.32851672e-01 -1.71035409e-01
-1.09821342e-01 1.78014800e-01 7.00069904e-01 -1.24681854e+00
6.02500081e-01 1.06012094e+00 8.88615489e-01 6.95058823e-01
-1.18724954e+00 -7.04398036e-01 2.21445337e-01 -2.67166734e-01
-1.16664827e+00 -2.30877414e-01 1.12622654e+00 -2.94770956e-01
9.46389616e-01 7.91020170e-02 8.69127035e-01 9.06696379e-01
-2.66792476e-01 8.14985514e-01 7.98185110e-01 -3.15223366e-01
1.20873466e-01 -2.51234680e-01 -3.18154879e-02 5.35840690e-01
2.69869387e-01 -6.35546669e-02 -1.14705317e-01 -1.99513733e-01
1.26948118e+00 2.27264568e-01 -1.83584586e-01 -8.39567482e-01
-1.20647943e+00 6.07273877e-01 8.24649930e-01 1.09927192e-01
-4.37554389e-01 2.65248388e-01 1.38311863e-01 -8.47248435e-02
8.33371520e-01 4.91237015e-01 -6.60816431e-01 3.55514921e-02
-6.37475848e-01 4.69227850e-01 1.67124972e-01 1.23853314e+00
5.62906504e-01 3.60459566e-01 8.95235687e-02 1.09765863e+00
4.70984697e-01 7.50650227e-01 1.31283820e-01 -9.73301232e-01
5.80914319e-01 9.89221811e-01 -2.30415575e-02 -7.69258738e-01
-3.61577213e-01 -5.30856073e-01 -7.71204650e-01 4.58003044e-01
4.01978970e-01 1.87558115e-01 -1.43610728e+00 1.49072421e+00
5.92646301e-01 5.09734452e-01 -2.36234352e-01 1.00700426e+00
8.82683098e-01 3.55888218e-01 6.00762926e-02 2.28888869e-01
9.76598322e-01 -9.05511558e-01 -3.50548744e-01 -3.01005363e-01
4.17821139e-01 -7.53874600e-01 1.38199687e+00 5.05211800e-02
-1.22080147e+00 -4.96179730e-01 -8.91551375e-01 -2.47208253e-02
-8.59814696e-03 3.00452616e-02 8.69913518e-01 3.98086041e-01
-8.33406150e-01 6.11310303e-01 -1.20832598e+00 -2.46862948e-01
9.57526505e-01 4.27187800e-01 -5.76698720e-01 -5.74981607e-02
-3.82590413e-01 6.38983548e-01 3.49189825e-02 1.19809192e-02
-5.21110356e-01 -1.13536954e+00 -7.56004930e-01 -1.22784875e-01
2.00995788e-01 -6.55886173e-01 1.31744528e+00 -4.40054476e-01
-1.44986653e+00 1.12093258e+00 1.65328711e-01 -4.77637649e-01
6.10234797e-01 -2.74234205e-01 1.38837576e-01 -4.67754602e-02
-9.08801258e-02 1.24337137e+00 9.18434799e-01 -1.54959941e+00
-1.65335014e-01 -6.66057348e-01 -2.06022739e-01 3.29587340e-01
-2.78205663e-01 -2.36181304e-01 -8.47484529e-01 -7.71369398e-01
7.50574589e-01 -1.16793144e+00 -3.96734536e-01 4.41260368e-01
-2.53554434e-01 -2.86302209e-01 1.06349897e+00 -3.07113677e-01
3.74404162e-01 -2.07227421e+00 -7.40841627e-02 2.83190191e-01
1.56987712e-01 4.88953769e-01 -3.01787317e-01 -1.34931460e-01
-1.14653774e-01 2.66103148e-01 -5.55844188e-01 -6.10711634e-01
-7.75152445e-02 4.25146222e-01 -3.69783759e-01 4.03640240e-01
5.79846323e-01 1.16829276e+00 -5.68425536e-01 -3.03774625e-01
5.28417766e-01 6.13219798e-01 -6.94914818e-01 1.05393574e-01
-4.43127185e-01 4.64870483e-01 -5.33358335e-01 1.05573666e+00
1.04465401e+00 -2.51426190e-01 -5.01983762e-01 -8.31558853e-02
1.63042143e-01 -3.20575461e-02 -1.03286946e+00 1.97551894e+00
-5.29053435e-02 3.04326802e-01 8.95522442e-03 -6.87300861e-01
1.25220072e+00 9.84535441e-02 6.32335365e-01 -9.19583067e-02
1.53668106e-01 3.20209324e-01 -1.03528779e-02 -1.13383077e-01
3.11132967e-01 -8.50473642e-02 3.65968704e-01 1.78993464e-01
-3.37090306e-02 -7.94852614e-01 -3.94332379e-01 -2.38962799e-01
8.25615108e-01 4.99108583e-01 -1.50738154e-02 7.86049813e-02
2.49383286e-01 1.85591862e-01 4.84192133e-01 3.40884507e-01
-1.79242834e-01 1.03030086e+00 -5.09828962e-02 -5.18792212e-01
-1.49113095e+00 -9.70318913e-01 -3.62473339e-01 6.06542647e-01
1.45115033e-01 1.09835513e-01 -7.18605876e-01 -7.29488552e-01
3.52475852e-01 5.23120761e-01 -3.49305272e-01 2.87741497e-02
-7.17990875e-01 -5.12009382e-01 4.24746722e-01 9.73976433e-01
7.47800529e-01 -1.26620281e+00 -1.99773833e-01 -8.47678483e-02
5.43341972e-02 -1.39319992e+00 -5.18287122e-01 -1.53734848e-01
-1.54906225e+00 -8.52677166e-01 -7.36113131e-01 -7.51702189e-01
1.09844673e+00 5.19840002e-01 1.03847754e+00 2.27519006e-01
-9.09366552e-03 3.31584185e-01 -2.71737933e-01 -7.19333768e-01
-3.02288443e-01 2.36945152e-01 -3.47195603e-02 -3.52454633e-01
3.97738665e-01 -5.74198842e-01 -4.61370319e-01 4.07882929e-01
-8.20522189e-01 5.34432419e-02 6.58658147e-01 6.94624841e-01
1.05080247e+00 -6.42143071e-01 4.00909573e-01 -6.28488421e-01
4.11105067e-01 1.31339878e-02 -7.59539366e-01 -1.51827171e-01
-5.54311872e-01 -1.19892873e-01 1.57181457e-01 -6.61126316e-01
-7.93057621e-01 3.96432459e-01 -4.74536091e-01 -1.12878537e+00
-4.39459383e-01 2.94454187e-01 -1.17827006e-01 -5.90353727e-01
6.77635849e-01 6.03816174e-02 2.43348822e-01 -6.52634740e-01
4.97150153e-01 3.90871584e-01 6.64179981e-01 -6.86666667e-01
1.19246078e+00 5.63712955e-01 1.00215092e-01 -7.70673394e-01
-5.36749959e-01 -3.08192760e-01 -8.30258548e-01 -1.15791529e-01
7.92439282e-01 -8.02004814e-01 -4.82724190e-01 5.18712819e-01
-1.19567811e+00 -2.54505485e-01 -6.34566903e-01 3.29666883e-01
-6.23148203e-01 3.59148264e-01 -2.89448410e-01 -6.98097825e-01
-6.47441745e-01 -1.16027069e+00 1.26669192e+00 1.52997106e-01
2.11866526e-03 -6.15171611e-01 -1.88778698e-01 3.40814650e-01
1.53435603e-01 4.40954804e-01 7.49850631e-01 -7.32061386e-01
-8.94892693e-01 -5.18458605e-01 -2.96506286e-01 4.10657465e-01
1.62773415e-01 1.49471685e-01 -9.31158900e-01 -2.46640205e-01
-1.28561944e-01 -2.30434895e-01 5.35020530e-01 4.43896234e-01
1.64816725e+00 1.10041142e-01 -3.80123168e-01 9.61107612e-01
1.06561852e+00 1.05676644e-01 6.18823111e-01 3.56724709e-01
8.77868176e-01 2.03255504e-01 6.67388976e-01 2.33546302e-01
6.10678568e-02 6.18041337e-01 9.33011532e-01 -2.57190436e-01
-1.85904473e-01 -2.78713614e-01 -3.76939327e-01 5.81314862e-01
-4.63880271e-01 2.10135221e-01 -1.17288530e+00 5.59817791e-01
-1.63124204e+00 -5.59055924e-01 -1.64119273e-01 2.11479282e+00
7.14710951e-01 3.06654543e-01 -5.18804230e-02 7.67418966e-02
6.20868742e-01 -5.88991418e-02 -8.32581222e-01 1.20201960e-01
7.27488846e-02 3.22823644e-01 4.33097333e-01 2.63989240e-01
-1.13991106e+00 1.20051014e+00 6.01721001e+00 6.06711805e-01
-1.17752147e+00 -1.31109953e-01 4.41482902e-01 1.38610587e-01
-3.31937045e-01 -3.31409812e-01 -6.45692110e-01 1.44262031e-01
1.61307633e-01 2.25510076e-01 1.29608124e-01 1.29434383e+00
1.06255636e-01 5.10635853e-01 -9.00680006e-01 1.04751575e+00
-1.99234173e-01 -1.35920870e+00 1.38600335e-01 1.65073216e-01
9.17747557e-01 3.76354784e-01 7.65526891e-02 1.25284582e-01
2.46726274e-02 -8.50481033e-01 4.50425446e-01 3.54060501e-01
9.11943555e-01 -7.68078804e-01 6.40172303e-01 3.38223279e-01
-8.25388849e-01 3.12364101e-01 -5.24435163e-01 2.91359812e-01
8.90917629e-02 4.46459800e-01 -1.16810191e+00 2.94918627e-01
6.27776682e-01 4.81113523e-01 -4.51201200e-01 1.26686609e+00
-1.94378197e-01 5.32012761e-01 -6.68621182e-01 1.78162768e-01
1.69974595e-01 -3.15757394e-01 6.98453128e-01 7.52845585e-01
4.82389808e-01 1.71532586e-01 2.74622142e-01 9.94278252e-01
-3.32826853e-01 -5.48416115e-02 -7.48706818e-01 -1.15057044e-01
5.96194148e-01 1.36253929e+00 -7.27682114e-01 -3.58164348e-02
-1.67085633e-01 5.84876835e-01 3.30452174e-01 2.34957173e-01
-6.76586092e-01 6.52060471e-03 8.97267342e-01 3.82640868e-01
2.06834882e-01 -6.48001134e-01 -9.02571857e-01 -9.64099884e-01
1.83247149e-01 -6.12096667e-01 -7.46148899e-02 -9.03311431e-01
-1.41777265e+00 5.20785809e-01 8.87625143e-02 -1.40379560e+00
3.48386429e-02 -5.97208261e-01 -6.81989133e-01 8.72002780e-01
-1.43589437e+00 -1.58458066e+00 -7.26061642e-01 3.25756907e-01
6.39678955e-01 -3.12381178e-01 7.56202161e-01 1.88705280e-01
-2.41226926e-01 5.44292331e-01 -4.33394581e-01 1.38545588e-01
4.81772274e-01 -1.25640094e+00 1.11441970e+00 6.30293489e-01
1.49672151e-01 5.70960999e-01 5.40318608e-01 -7.97459602e-01
-1.30251217e+00 -1.12234771e+00 3.49789560e-01 -3.87023926e-01
3.81383747e-01 -2.15866715e-01 -1.40919769e+00 6.64496720e-01
-2.21327707e-01 5.16541243e-01 3.18286091e-01 -1.17195867e-01
-3.19434673e-01 5.75326420e-02 -1.53385890e+00 7.14653552e-01
1.32798350e+00 -1.01517454e-01 -6.04701579e-01 3.30695510e-01
8.85535419e-01 -9.19348180e-01 -9.19830739e-01 9.63175535e-01
2.48432055e-01 -4.81122732e-01 1.27176714e+00 -6.91930413e-01
4.44802940e-01 -3.68509203e-01 -8.96413326e-02 -1.40554070e+00
-2.33523861e-01 -2.46446714e-01 -2.20907196e-01 1.21815228e+00
2.98381597e-01 -5.89623868e-01 1.45749581e+00 7.85309494e-01
-6.95100307e-01 -1.09501088e+00 -9.18657005e-01 -6.66805744e-01
1.87877774e-01 -6.32715940e-01 9.46607053e-01 1.05678892e+00
-6.99991643e-01 -1.35219917e-01 1.42939240e-01 2.56308496e-01
6.99948490e-01 1.11695230e-02 1.14270127e+00 -1.56645405e+00
9.30114836e-02 -4.84480619e-01 -4.91895288e-01 -1.10293639e+00
2.57054299e-01 -1.09375334e+00 5.73666010e-04 -1.63479090e+00
-1.48602158e-01 -8.85604858e-01 1.82590187e-01 7.75213242e-01
-1.97416037e-01 4.61297631e-01 2.60566622e-01 3.33499879e-01
1.75060764e-01 7.91754842e-01 1.72787023e+00 -1.60057172e-01
-3.92428100e-01 2.67128944e-01 -4.74901021e-01 1.00028646e+00
9.26271737e-01 -2.07544252e-01 -4.46581036e-01 -6.83514357e-01
-1.47981912e-01 -2.63885856e-01 4.31983948e-01 -9.25985992e-01
-1.29162058e-01 -2.98172116e-01 5.87645054e-01 -1.21893215e+00
6.05951428e-01 -1.10214758e+00 5.18111475e-02 1.75090790e-01
-2.21466161e-02 9.98580605e-02 4.74905312e-01 4.66412246e-01
4.81942892e-02 -2.73672819e-01 9.02619362e-01 -4.58803736e-02
-4.09945309e-01 8.91830027e-01 4.52924103e-01 -1.75893083e-01
1.03647292e+00 -4.57832783e-01 -2.57238388e-01 -3.63166910e-03
-7.56943643e-01 2.25330368e-01 7.09270895e-01 5.22098899e-01
1.03023458e+00 -1.62176478e+00 -5.88304341e-01 4.35545653e-01
1.84273273e-01 8.33909094e-01 1.50883552e-02 4.83842760e-01
-6.13918662e-01 9.00470465e-02 -2.61886090e-01 -1.09810245e+00
-1.25784779e+00 3.84524107e-01 3.34270924e-01 8.23154077e-02
-9.81262207e-01 9.47693288e-01 -8.46532881e-02 -9.34383512e-01
2.16150001e-01 -6.72614217e-01 2.20561940e-02 -4.31739211e-01
2.83391662e-02 2.05904454e-01 3.58878881e-01 -3.97186756e-01
-2.77827948e-01 8.16435337e-01 -2.23605540e-02 1.05078958e-01
1.52247262e+00 3.68715763e-01 4.10804860e-02 1.43960088e-01
8.69582832e-01 -2.53419578e-01 -1.54884362e+00 -2.18917862e-01
-1.35745645e-01 -7.44238794e-01 -7.34371366e-03 -5.53630412e-01
-1.17074633e+00 9.28739190e-01 6.22983456e-01 1.33860588e-01
7.85602152e-01 1.12549037e-01 8.42532933e-01 4.33253974e-01
4.06611621e-01 -6.04542613e-01 7.62431175e-02 5.15537143e-01
1.31021857e+00 -1.35906899e+00 -2.52614496e-03 -8.07007313e-01
-5.44967830e-01 8.84113550e-01 1.06690097e+00 -3.89135122e-01
4.86304849e-01 1.65415540e-01 2.11211845e-01 -1.78984329e-01
-3.23004484e-01 -9.12096500e-02 4.59973902e-01 9.03876126e-01
2.14863956e-01 -9.55637023e-02 -4.20118310e-02 1.97429359e-01
-2.81981468e-01 -1.61026105e-01 1.88390329e-01 7.66994178e-01
-3.90255153e-01 -1.07349885e+00 -5.94301581e-01 4.98295248e-01
-2.25597322e-01 2.81718671e-01 -4.51351553e-01 9.69094098e-01
-1.53683841e-01 2.51539320e-01 2.42463425e-01 -2.26996869e-01
6.19684398e-01 1.99592128e-01 6.37414873e-01 -6.44573271e-01
-4.26540345e-01 2.03111678e-01 -2.97911614e-01 -3.18426937e-01
-3.88205349e-01 -6.97851181e-01 -1.65616369e+00 -1.11498959e-01
-4.62076813e-01 -3.06453377e-01 1.09662771e+00 8.04022431e-01
5.20452797e-01 4.69955415e-01 4.70125020e-01 -1.35807478e+00
-8.11737716e-01 -1.11422718e+00 -1.52715817e-01 5.57685018e-01
3.51150371e-02 -9.05599356e-01 -1.18746147e-01 -1.48278564e-01] | [8.027142524719238, -3.381268262863159] |
818c5eec-6393-461e-8266-7dd896ea205c | question-driven-span-labeling-model-for | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/17523 | https://ojs.aaai.org/index.php/AAAI/article/view/17523/17330 | Question-Driven Span Labeling Model for Aspect–Opinion Pair Extraction | Aspect term extraction and opinion word extraction are two fundamental subtasks of aspect-based sentiment analysis. The internal relationship between aspect terms and opinion words is typically ignored, and information for the decision-making of buyers and sellers is insufficient. In this paper, we explore an aspect–opinion pair extraction (AOPE) task and propose a Question-Driven Span Labeling (QDSL) model to extract all the aspect–opinion pairs from user-generated reviews. Specifically, we divide the AOPE task into aspect term extraction (ATE) and aspect-specified opinion extraction (ASOE) subtasks; we first extract all the candidate aspect terms and then the corresponding opinion words given the aspect term. Unlike existing approaches that use the BIO-based tagging scheme for extraction, the QDSL model adopts a span-based tagging scheme and builds a question–answer-based machine-reading comprehension task for an effective aspect–opinion pair extraction. Extensive experiments conducted on three tasks (ATE, ASOE, and AOPE) on four benchmark datasets demonstrate that the proposed method significantly outperforms state-of-the-art approaches. | ['Jianxin Liao', 'Lei Zhang', 'Jingyu Wang', 'Tongcun Liu', 'Yulong Wang', 'Lei Gao'] | 2021-05-18 | null | null | null | aaai-2021-5 | ['term-extraction', 'aspect-based-sentiment-analysis', 'reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.89335212e-01 2.29306981e-01 -1.95573285e-01 -5.97648799e-01
-1.03588116e+00 -7.65423357e-01 5.83491147e-01 6.61393285e-01
-3.91042888e-01 3.11379284e-01 4.06097770e-01 -5.87150395e-01
3.19200665e-01 -8.44787538e-01 -1.96478903e-01 -4.62622970e-01
4.00655925e-01 3.08518320e-01 1.57335743e-01 -4.54216480e-01
5.28261244e-01 -2.72050112e-01 -1.34663856e+00 3.25588942e-01
1.06681621e+00 1.31015313e+00 -3.01707476e-01 5.68523288e-01
-8.80922735e-01 6.96576416e-01 -7.02339470e-01 -1.01788044e+00
6.74997568e-02 -4.03255969e-01 -8.08552027e-01 3.97906750e-01
-2.14362279e-01 1.38961792e-01 5.98532617e-01 9.97011065e-01
2.61795133e-01 -8.75494927e-02 6.79769814e-01 -1.18872452e+00
-8.68708313e-01 5.64456820e-01 -8.21615398e-01 -1.23553835e-01
4.83867943e-01 -1.77178442e-01 1.62689233e+00 -1.11327791e+00
3.09811831e-01 1.04607260e+00 4.67737556e-01 1.14359275e-01
-4.40122485e-01 -3.01578164e-01 7.27228820e-01 9.36088562e-02
-9.53245282e-01 -1.32171094e-01 9.10731018e-01 -2.37296879e-01
1.13125455e+00 2.24073604e-01 1.11644685e+00 4.07579511e-01
2.77554542e-01 1.12611151e+00 1.14362526e+00 -4.20788109e-01
3.19091797e-01 5.37012637e-01 8.28431547e-01 4.60085630e-01
4.03404981e-01 -5.07680595e-01 -7.10142434e-01 -3.50744039e-01
1.51904486e-02 -1.79154158e-01 -1.94748402e-01 -3.86639945e-02
-8.17077518e-01 8.42082620e-01 -8.12207311e-02 2.79646009e-01
-6.53920949e-01 -3.82812917e-01 4.30231661e-01 4.77457553e-01
8.47820818e-01 5.49074888e-01 -1.06885612e+00 -1.19180515e-01
-5.49986482e-01 2.06498981e-01 1.35667026e+00 1.12277591e+00
8.83711159e-01 -3.22937161e-01 -2.99547672e-01 8.13516557e-01
5.55959761e-01 7.08239436e-01 3.71610254e-01 -3.80936027e-01
3.23585153e-01 1.16184628e+00 1.63934216e-01 -8.92530978e-01
-1.52931929e-01 -6.74351335e-01 -3.98666114e-01 -2.99929857e-01
-2.57704794e-01 -3.70726287e-01 -9.31723654e-01 1.18550313e+00
7.38266289e-01 -5.17018557e-01 2.42513195e-01 5.18098652e-01
1.27953553e+00 5.45765579e-01 1.95694417e-01 -3.96190256e-01
2.25029016e+00 -1.38668489e+00 -1.05855513e+00 -6.94275618e-01
4.82840627e-01 -1.21715605e+00 9.36458051e-01 2.93171167e-01
-8.75267088e-01 -8.41983929e-02 -1.05179870e+00 -1.48273259e-01
-6.58927441e-01 1.30406007e-01 9.35001671e-01 5.32488883e-01
-7.32680500e-01 -2.60367930e-01 -4.86951947e-01 -7.46929571e-02
2.81516969e-01 1.62789419e-01 -2.57155865e-01 2.38106027e-02
-1.25521958e+00 6.95449293e-01 -1.13814019e-01 2.93060720e-01
-1.60518840e-01 -5.41723847e-01 -1.23595977e+00 -6.96530044e-02
5.41114211e-01 -9.49700415e-01 1.54538071e+00 -1.17327058e+00
-1.25897181e+00 7.92795002e-01 -7.22670794e-01 1.17457241e-01
-5.04944205e-01 -4.63806510e-01 -5.98160744e-01 2.39788480e-02
3.75333041e-01 2.82704830e-01 9.69015658e-01 -1.33968544e+00
-8.61198008e-01 -4.66856152e-01 5.58952868e-01 5.21028817e-01
-2.48230740e-01 3.11788738e-01 -6.56987250e-01 -5.91222107e-01
1.93809971e-01 -7.38658130e-01 -3.39958251e-01 -2.18628675e-01
-4.03775334e-01 -6.59513652e-01 5.34833074e-01 -6.59230530e-01
1.41769886e+00 -1.84599686e+00 -1.93555087e-01 1.96722358e-01
2.20705479e-01 2.37647876e-01 -2.80767351e-01 4.72396016e-01
1.44315764e-01 1.05081603e-01 -3.46193105e-01 -3.90695453e-01
1.81667060e-01 -5.02602719e-02 -2.66170025e-01 -3.55870038e-01
4.62693155e-01 1.18492937e+00 -9.01268065e-01 -5.20541728e-01
-4.48633760e-01 2.89192587e-01 -5.09025693e-01 4.66274530e-01
-4.25256819e-01 -8.25550482e-02 -7.37260282e-01 1.16754949e+00
6.33723974e-01 -1.73568860e-01 -7.40555972e-02 -4.15105939e-01
-4.02397886e-02 7.72995174e-01 -8.44342470e-01 1.25151873e+00
-8.37931156e-01 2.76907563e-01 9.17767733e-02 -4.69177514e-01
9.32788968e-01 4.11543697e-01 2.28088245e-01 -6.21272624e-01
2.95615762e-01 2.92081118e-01 -1.21826842e-01 -5.95089018e-01
8.58743548e-01 -4.37838793e-01 -4.07370985e-01 7.80690193e-01
1.06490463e-01 -4.57353085e-01 6.20032191e-01 3.03479224e-01
9.99226034e-01 7.08064139e-02 8.19177151e-01 -1.75002635e-01
8.70329320e-01 1.13051429e-01 6.52913690e-01 2.76957721e-01
-1.94515407e-01 5.07217109e-01 8.78654599e-01 -3.19762021e-01
-4.73108113e-01 -6.42528892e-01 2.12094024e-01 1.04338169e+00
1.61676809e-01 -9.34136808e-01 -6.71651483e-01 -1.14248955e+00
-3.61072510e-01 7.53882885e-01 -7.14276373e-01 1.28415570e-01
-2.56325632e-01 -8.92641783e-01 -2.32752904e-01 4.37520146e-01
4.47197706e-01 -1.13090312e+00 -1.08150467e-01 2.50837117e-01
-3.61378998e-01 -1.24492526e+00 -7.53787696e-01 1.17016785e-01
-4.22856212e-01 -1.04973269e+00 -4.05816972e-01 -9.80707526e-01
8.62172663e-01 3.16126078e-01 1.63185692e+00 1.36484891e-01
2.60114163e-01 2.99323410e-01 -8.45382988e-01 -9.12148297e-01
1.56716257e-01 2.36219704e-01 -4.68009740e-01 3.41156013e-02
1.15330112e+00 -5.79451799e-01 -8.26049387e-01 3.15521918e-02
-1.06983113e+00 -1.70431912e-01 9.79271293e-01 5.06356597e-01
7.31761694e-01 -8.30779150e-02 6.79892898e-01 -1.43267465e+00
1.07730937e+00 -2.96152890e-01 -3.23606402e-01 3.51672083e-01
-1.00242805e+00 1.66948531e-02 3.31422240e-01 -9.77902040e-02
-1.15232217e+00 -2.82086819e-01 -5.61143279e-01 5.72601616e-01
1.30963117e-01 8.84522974e-01 -5.03706157e-01 2.78172493e-01
-4.49690782e-02 3.20973486e-01 -4.80819881e-01 -2.07388744e-01
4.72695023e-01 9.76866424e-01 -9.08395573e-02 -2.07081199e-01
7.08965600e-01 2.84352809e-01 -4.49875295e-01 -6.46748304e-01
-1.77072442e+00 -8.88825297e-01 -3.68939549e-01 -2.60539306e-03
9.04862821e-01 -1.06136882e+00 -3.48640710e-01 5.65396011e-01
-1.29225719e+00 5.92806637e-01 -4.55245495e-01 1.26988411e-01
-6.93952572e-03 2.08435133e-01 -4.23187196e-01 -9.64216769e-01
-1.06243157e+00 -1.04073811e+00 1.48872042e+00 5.10134637e-01
-4.10979539e-01 -9.72062826e-01 2.52983004e-01 7.31112778e-01
5.03390133e-01 -2.38331288e-01 1.10314000e+00 -8.65721285e-01
-4.77564335e-01 -5.74847817e-01 -2.02160582e-01 5.19624889e-01
3.31485659e-01 -2.16665417e-01 -8.50808442e-01 2.94291288e-01
3.53010207e-01 -1.63401142e-01 7.73698211e-01 5.68755381e-02
4.09490198e-01 -4.11030352e-01 8.19877759e-02 3.23868655e-02
1.21907210e+00 1.35381043e-01 4.10044432e-01 3.78259599e-01
5.94174564e-01 6.88286722e-01 1.05563390e+00 2.57471800e-01
1.22561431e+00 1.49772540e-01 9.48087573e-02 -4.34338301e-02
1.32543519e-01 -1.60557643e-01 5.65815508e-01 1.71838820e+00
3.35674465e-01 -3.23452920e-01 -4.51504737e-01 9.45275724e-01
-1.61999118e+00 -3.52043390e-01 -2.90215790e-01 1.60051787e+00
9.92540121e-01 3.25240940e-01 -1.39256656e-01 1.51602462e-01
2.67689586e-01 5.96842408e-01 -3.83329868e-01 -7.37725079e-01
-5.05741984e-02 3.79711241e-01 -1.26164764e-01 5.89574337e-01
-1.11574697e+00 9.27159369e-01 4.96743584e+00 7.25481033e-01
-5.43956220e-01 2.87958026e-01 5.81855178e-01 4.20663297e-01
-9.94572520e-01 4.64321107e-01 -8.87131155e-01 -1.28488451e-01
5.34232914e-01 -2.20577687e-01 -1.52672678e-01 8.95001292e-01
1.47217624e-02 -4.47826087e-01 -7.45490849e-01 6.64204776e-01
4.73445386e-01 -6.27480567e-01 3.98823529e-01 -3.58191341e-01
8.39842796e-01 -3.80678236e-01 -1.56738654e-01 4.61617351e-01
-1.04972096e-02 -5.71043968e-01 4.73986000e-01 4.23630029e-01
1.93993539e-01 -8.27335536e-01 1.36516535e+00 4.15694043e-02
-1.74440563e+00 2.79535085e-01 -1.68065995e-01 -5.64418584e-02
5.51702797e-01 1.25982857e+00 -2.23278701e-01 7.19682574e-01
5.79521298e-01 6.43824339e-01 -5.26306212e-01 6.28954172e-01
-9.80248809e-01 6.58074379e-01 2.19936147e-02 -5.00091910e-01
2.48424605e-01 -4.58895385e-01 5.55599749e-01 1.02123380e+00
4.92145196e-02 1.91173032e-01 -7.86396489e-02 3.68338615e-01
-2.31168643e-01 6.50860369e-01 -1.95131451e-01 -3.79437476e-01
1.65949330e-01 1.88394940e+00 -7.73605227e-01 -4.53366458e-01
-7.24330127e-01 8.97140801e-01 2.30605438e-01 2.11261034e-01
-3.36217731e-01 -7.76546299e-01 6.81857705e-01 -2.05466062e-01
7.59717226e-01 5.72567061e-02 -4.86452430e-01 -1.44449043e+00
5.92461228e-01 -1.19734466e+00 2.68697262e-01 -7.51275837e-01
-1.46583652e+00 9.80529249e-01 -4.01841313e-01 -1.27793181e+00
-1.68304101e-01 -4.09657240e-01 -9.36801732e-01 7.29677975e-01
-2.02610540e+00 -1.37027967e+00 -8.34020674e-02 -3.46891396e-02
8.02672267e-01 1.79664418e-01 7.88918018e-01 1.34260818e-01
-2.25281790e-01 3.21983486e-01 -7.39508569e-01 1.00007787e-01
3.74545097e-01 -1.41043890e+00 5.63895285e-01 7.03349233e-01
1.51499018e-01 8.66886020e-01 6.29361510e-01 -5.61158180e-01
-1.39554214e+00 -9.74613786e-01 1.80645311e+00 -5.69918156e-01
7.50715017e-01 -2.72918284e-01 -7.60065913e-01 4.52480227e-01
7.79008210e-01 -4.07513171e-01 1.22146964e+00 3.63710791e-01
-5.20358205e-01 -2.65239865e-01 -7.50778496e-01 5.87840438e-01
6.73252523e-01 -5.75084388e-01 -9.66865063e-01 2.30328292e-01
1.07645166e+00 6.90930411e-02 -7.55517900e-01 5.01502454e-01
5.98765492e-01 -8.06703508e-01 4.21023458e-01 -3.56430262e-01
4.75359261e-01 -5.55065870e-01 -1.04801636e-02 -1.42126989e+00
1.75342247e-01 -4.63277549e-01 -2.36896411e-01 1.69793153e+00
1.09195471e+00 -5.96908569e-01 4.02785659e-01 6.16971612e-01
1.62292078e-01 -1.34864295e+00 -4.60833311e-01 -1.31726980e-01
-2.77634561e-01 -5.62938988e-01 8.68365824e-01 4.49021548e-01
1.94805965e-01 1.25645351e+00 3.88257429e-02 7.55936056e-02
2.88160950e-01 7.84890234e-01 4.32455242e-01 -8.58350992e-01
-2.79802144e-01 -2.96688169e-01 -4.54747938e-02 -1.17335963e+00
8.11372325e-02 -4.19868052e-01 2.32720450e-01 -2.00942612e+00
4.36014712e-01 -4.17389609e-02 -1.97731093e-01 3.19823563e-01
-9.43446279e-01 7.78982267e-02 -1.00276016e-01 -1.74038902e-01
-1.14446652e+00 9.13607478e-01 1.33369207e+00 -2.06552848e-01
-2.31399499e-02 4.02251363e-01 -1.46857440e+00 9.37685013e-01
5.64686596e-01 -5.34036279e-01 -4.70339209e-01 -3.33361626e-01
9.82461214e-01 -3.48260343e-01 -4.68932956e-01 -2.57197201e-01
1.06024370e-01 1.04704633e-01 -1.50343344e-01 -8.59498084e-01
1.37435362e-01 -7.72608042e-01 -5.97664595e-01 7.33225718e-02
-1.23271011e-01 5.37874460e-01 -1.25188708e-01 4.12327975e-01
-7.11681485e-01 -3.61365229e-01 1.28166348e-01 -2.88350433e-01
-3.89141470e-01 3.21239561e-01 -5.07424414e-01 4.03812051e-01
5.95745742e-01 3.52251470e-01 -2.80095130e-01 -6.74041390e-01
-2.64558643e-01 5.10500252e-01 1.44469425e-01 5.09128690e-01
7.04913020e-01 -1.12424827e+00 -5.72562814e-01 1.37981042e-01
5.56200624e-01 1.44779667e-01 -3.70130017e-02 7.95716047e-01
-5.54581136e-02 3.67962182e-01 6.99900866e-01 -8.10191929e-02
-1.25503325e+00 2.79263973e-01 2.15559695e-02 -8.72507453e-01
2.25612715e-01 9.65389788e-01 2.73502439e-01 -8.05004478e-01
-1.31918907e-01 -3.69792938e-01 -8.74051869e-01 4.33424085e-01
5.97958386e-01 -2.64800638e-01 2.29069799e-01 -6.34866655e-01
-4.21484053e-01 7.86104620e-01 -3.31903249e-01 -2.68561333e-01
1.25836730e+00 -3.71764660e-01 -7.72537768e-01 4.55738664e-01
1.04152536e+00 3.90521318e-01 -4.11164254e-01 -2.88091421e-01
2.96196967e-01 2.28875391e-02 2.91192159e-02 -8.26129556e-01
-1.07047224e+00 6.98350012e-01 -1.76929738e-02 4.34065133e-01
1.37149978e+00 1.69580996e-01 1.35865903e+00 3.14040184e-01
5.78851104e-02 -1.06963861e+00 -1.28442183e-01 7.70514011e-01
7.33063281e-01 -1.37708485e+00 1.33274525e-01 -7.90478170e-01
-8.38597476e-01 6.88253045e-01 7.50170112e-01 2.63094544e-01
1.09861386e+00 2.22155422e-01 6.70964241e-01 -5.63969374e-01
-1.21343195e+00 -6.75969362e-01 6.44889474e-01 2.43306965e-01
7.12943017e-01 -1.05211236e-01 -9.50077415e-01 1.25023067e+00
-3.21489692e-01 -2.85445511e-01 1.99142084e-01 1.37422109e+00
-4.85127687e-01 -1.48246443e+00 1.81652725e-01 7.19867587e-01
-7.15529442e-01 -6.96293890e-01 -7.57128894e-01 2.67908931e-01
6.76366389e-02 1.44560361e+00 -2.40406796e-01 -3.35419685e-01
5.52686453e-01 2.16882050e-01 -6.15373533e-03 -8.84845555e-01
-1.08330607e+00 5.65630943e-02 5.34221053e-01 -3.57838601e-01
-8.13609064e-01 -4.26373929e-01 -1.07645321e+00 2.10609913e-01
-7.52197385e-01 7.45819271e-01 8.58274698e-01 1.48162293e+00
5.14159441e-01 5.69736183e-01 8.59369636e-01 -5.13008721e-02
-1.28627241e-01 -1.04085469e+00 -4.31326956e-01 2.65964538e-01
2.85384744e-01 -2.14537919e-01 -4.35069323e-01 -1.10537127e-01] | [11.48515510559082, 6.623805999755859] |
87dda7e0-168e-485e-80fd-65bfdc1d79b2 | an-exploratory-study-into-automated-pr-ecis | null | null | https://aclanthology.org/2020.lrec-1.50 | https://aclanthology.org/2020.lrec-1.50.pdf | An Exploratory Study into Automated Pr\'ecis Grading | Automated writing evaluation is a popular research field, but the main focus has been on evaluating argumentative essays. In this paper, we consider a different genre, namely pr{\'e}cis texts. A pr{\'e}cis is a written text that provides a coherent summary of main points of a spoken or written text. We present a corpus of English pr{\'e}cis texts which all received a grade assigned by a highly-experienced English language teacher and were subsequently annotated following an exhaustive error typology. With this corpus we trained a machine learning model which relies on a number of linguistic, automatic summarization and AWE features. Our results reveal that this model is able to predict the grade of pr{\'e}cis texts with only a moderate error margin. | ['Senne Van Hoecke', 'Orphee De Clercq'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['automated-writing-evaluation'] | ['natural-language-processing'] | [ 1.51828468e-01 4.59225148e-01 -2.41791800e-01 -1.86517745e-01
-1.03202903e+00 -6.66905403e-01 8.37690115e-01 7.79035628e-01
-5.81042111e-01 1.01483655e+00 4.68329221e-01 -5.95491588e-01
-3.75617981e-01 -6.21372581e-01 -4.64355558e-01 -2.06257313e-01
6.34097934e-01 4.73403186e-01 6.17959425e-02 -4.14735019e-01
1.05727887e+00 5.04470542e-02 -1.26844072e+00 3.96534652e-01
1.34605515e+00 6.12744987e-01 1.57650188e-01 1.00267470e+00
-2.22552180e-01 1.15355337e+00 -1.25786209e+00 -1.09090102e+00
-5.69395304e-01 -6.56987429e-01 -1.28977859e+00 -1.32002413e-01
5.21347642e-01 -9.90423784e-02 -2.60665044e-02 1.08116174e+00
3.12536210e-01 -1.67308837e-01 9.82578218e-01 -7.88799345e-01
-3.53388637e-01 9.88644302e-01 -5.66708073e-02 3.93709511e-01
7.79896259e-01 -1.78642407e-01 1.31875288e+00 -8.44262898e-01
7.96633184e-01 8.28924239e-01 8.03422034e-01 3.83362442e-01
-1.01203704e+00 -2.71116346e-01 -1.22624494e-01 2.18341097e-01
-8.07652593e-01 -3.13459545e-01 8.06009293e-01 -7.44906247e-01
8.97376716e-01 2.53567129e-01 6.80044889e-01 1.12535572e+00
3.56831580e-01 9.44665611e-01 1.36232626e+00 -9.60244000e-01
3.17805022e-01 3.39563042e-01 8.38748753e-01 5.50297379e-01
1.93612427e-01 -6.44582748e-01 -5.93478441e-01 -1.04827926e-01
-1.60773501e-01 -6.08766794e-01 -2.95848459e-01 6.49767756e-01
-9.36104417e-01 9.61491764e-01 -3.52954149e-01 6.67117000e-01
-3.06764156e-01 -2.27981314e-01 5.57903290e-01 6.08393490e-01
5.71961403e-01 6.65802956e-01 -3.47605139e-01 -5.92564404e-01
-1.23325610e+00 3.52338791e-01 1.34123814e+00 5.33810318e-01
3.40067387e-01 -3.58932644e-01 -4.08449888e-01 1.02490103e+00
4.49983925e-02 4.28659260e-01 5.51419199e-01 -1.01919401e+00
8.14539254e-01 9.55121934e-01 1.31453246e-01 -1.02730203e+00
-2.65576124e-01 -2.64181256e-01 -5.52223921e-01 1.45026669e-01
4.69463080e-01 -2.47290775e-01 -1.37438253e-01 1.34126174e+00
-1.63449571e-01 -7.07447469e-01 2.25749478e-01 3.84817213e-01
1.30788708e+00 6.68637693e-01 8.35919380e-03 -7.10972488e-01
1.26830399e+00 -9.56929803e-01 -9.51930404e-01 1.95523486e-01
6.69164300e-01 -1.08473217e+00 1.22651088e+00 8.32886517e-01
-1.53832376e+00 -2.59974837e-01 -1.12271571e+00 -1.68655917e-01
-1.48082271e-01 6.48396194e-01 -9.46331620e-02 6.79341853e-01
-8.17262530e-01 7.65274346e-01 -4.11563158e-01 -1.98685870e-01
2.26335004e-01 -9.29190144e-02 -3.29950869e-01 2.74563968e-01
-1.02104819e+00 1.11788177e+00 2.72593379e-01 -1.75778344e-01
-3.52265120e-01 -6.72909617e-01 -6.32522047e-01 7.40656555e-02
2.93793708e-01 -3.05524945e-01 1.56605613e+00 -8.71500015e-01
-1.67938232e+00 1.32402670e+00 -1.35106996e-01 -4.16849136e-01
7.54878581e-01 -3.47333252e-01 -4.26849544e-01 4.30454940e-01
2.36427262e-01 -1.55473888e-01 5.31365633e-01 -8.46660614e-01
-5.87887883e-01 -3.40761751e-01 -7.40857497e-02 8.18302184e-02
-5.39257884e-01 6.13016963e-01 1.78075552e-01 -9.04236317e-01
-1.28461719e-01 -6.26332402e-01 3.84931713e-01 -5.83191454e-01
-6.49867117e-01 -9.09362257e-01 3.43162358e-01 -9.34335053e-01
1.90268362e+00 -1.43700516e+00 1.74527735e-01 -3.37495953e-02
4.05883193e-01 5.61697185e-01 3.17794889e-01 7.95525491e-01
2.96745330e-01 4.04010832e-01 -2.37320602e-01 -3.26917946e-01
3.44901204e-01 -8.35259184e-02 -3.67351264e-01 1.42737865e-01
5.31569570e-02 5.81716895e-01 -9.54106808e-01 -7.31145918e-01
-1.45356938e-01 -1.01517335e-01 -3.25754076e-01 2.98398942e-01
-4.20680791e-01 8.89590830e-02 -3.94336820e-01 3.42419952e-01
9.17355493e-02 -1.82038933e-01 -4.28673811e-02 3.07509005e-01
-6.13565803e-01 8.84667516e-01 -8.17394614e-01 1.25879359e+00
-3.07116687e-01 9.61511791e-01 -3.39184165e-01 -9.43981886e-01
1.03376055e+00 3.74058306e-01 -1.31845325e-02 -5.18466055e-01
3.51042897e-01 6.99538887e-01 1.76070314e-02 -6.61640584e-01
9.35829878e-01 2.17950687e-01 -2.83538967e-01 1.01781023e+00
7.76065141e-02 -4.13232446e-01 9.79403555e-01 4.37778860e-01
1.22726357e+00 -8.97600129e-02 5.89949906e-01 -4.49948937e-01
1.06856143e+00 3.13369840e-01 3.22816730e-01 8.80443096e-01
-1.01666443e-01 6.35535181e-01 8.97363186e-01 -2.72845268e-01
-1.27116716e+00 -5.29690027e-01 -3.43538970e-01 1.08391452e+00
-4.58706200e-01 -8.16190779e-01 -1.16453505e+00 -7.64999747e-01
-4.08669591e-01 1.08269155e+00 -4.99491125e-01 4.51897830e-02
-7.40341187e-01 -4.96884972e-01 7.73289680e-01 1.08771227e-01
4.94245738e-01 -1.29038095e+00 -9.15228724e-01 3.05256277e-01
-4.25815135e-01 -7.66350925e-01 -2.72025794e-01 1.74133703e-01
-5.07926047e-01 -1.24821019e+00 -5.13623178e-01 -6.36689484e-01
3.79021406e-01 -4.03477311e-01 1.42757869e+00 3.90653551e-01
6.78310990e-02 3.33982527e-01 -7.67158985e-01 -8.82249117e-01
-1.11396849e+00 4.29093540e-01 -2.12501168e-01 -2.90639669e-01
4.78791118e-01 -2.92464316e-01 -1.11558422e-01 -1.20975457e-01
-7.44341075e-01 1.29761770e-01 4.08939958e-01 9.61552739e-01
2.15874702e-01 -3.73791337e-01 6.32701218e-01 -1.15968907e+00
1.22025943e+00 -2.12404713e-01 -2.42679432e-01 4.03206289e-01
-7.79232264e-01 9.10629705e-02 7.36235321e-01 -3.53920251e-01
-1.07947242e+00 -7.92828918e-01 -4.98553932e-01 3.49641711e-01
-1.11334920e-01 7.70852864e-01 -2.37594470e-02 3.09462965e-01
8.03727388e-01 7.31527507e-02 -3.43903482e-01 -3.80183816e-01
-9.40469876e-02 9.52829361e-01 7.73792267e-01 -7.25294352e-01
3.50035369e-01 -3.33727449e-01 -2.39162967e-01 -9.74047124e-01
-1.29903328e+00 -2.79513717e-01 -6.34733677e-01 -4.05071050e-01
5.37323773e-01 -3.39307129e-01 -7.33439088e-01 3.79142016e-01
-1.46810591e+00 -3.41434628e-01 -2.61351407e-01 5.83075762e-01
-5.32710195e-01 5.07934868e-01 -7.15998769e-01 -8.15001905e-01
-6.88614070e-01 -7.57881582e-01 7.30683804e-01 3.99430543e-01
-1.02737284e+00 -1.00937939e+00 4.42985952e-01 7.35770822e-01
-3.27683836e-02 1.00089304e-01 1.11848116e+00 -1.05364132e+00
2.87773877e-01 -2.38221571e-01 3.18825059e-02 5.51280975e-01
-4.56194371e-01 4.53471452e-01 -7.05180049e-01 9.12480503e-02
1.53554082e-01 -6.85370624e-01 7.70269036e-01 1.12093747e-01
1.04099321e+00 -8.39317024e-01 1.79701909e-01 -7.38657936e-02
1.05886042e+00 -1.40090987e-01 5.20485461e-01 5.77958763e-01
1.87308311e-01 6.33566380e-01 5.55812478e-01 5.54418743e-01
5.25420606e-01 2.76420653e-01 1.93155631e-02 6.06401384e-01
-1.98795527e-01 -3.47387254e-01 4.59428430e-01 1.34714007e+00
-2.67366678e-01 -5.21702111e-01 -1.10557485e+00 7.33062685e-01
-1.73940980e+00 -1.19848931e+00 -6.61728323e-01 1.86795795e+00
1.22156215e+00 3.35850596e-01 -1.08324930e-01 8.03945184e-01
6.64894104e-01 1.28415957e-01 7.85222799e-02 -9.49851751e-01
-3.36887956e-01 5.02970994e-01 -1.43139437e-01 4.87677395e-01
-8.97246540e-01 6.80060804e-01 5.97736216e+00 8.80312860e-01
-7.65105009e-01 -4.03284803e-02 4.86831576e-01 3.47133160e-01
-4.17748958e-01 -1.13037705e-01 -8.79204690e-01 6.85211301e-01
1.05037606e+00 -4.59353924e-01 -2.74061561e-01 5.36725104e-01
2.86434352e-01 -5.21973312e-01 -9.14157033e-01 4.89593923e-01
5.60939372e-01 -1.36499023e+00 -1.35065138e-01 -1.06240585e-01
1.02100039e+00 -3.44608188e-01 -1.99724793e-01 3.65621239e-01
4.64187771e-01 -1.05522799e+00 1.07588613e+00 6.35195613e-01
7.59518385e-01 -5.96357167e-01 8.91750038e-01 9.28876460e-01
-3.02692562e-01 1.65039338e-02 -1.74343735e-01 -4.48980868e-01
-1.06990814e-01 6.24417305e-01 -5.63867569e-01 5.31490564e-01
5.83041549e-01 4.12421376e-01 -8.09595585e-01 8.37898374e-01
-8.04653585e-01 1.13474000e+00 1.04709128e-02 -7.45826662e-01
2.16293141e-01 -2.56646633e-01 8.00934970e-01 1.56937718e+00
3.88114274e-01 4.23192978e-01 -3.35423946e-02 7.75211215e-01
-4.21812654e-01 5.22346497e-01 -2.62472630e-01 -8.82459059e-02
4.36403990e-01 1.30287647e+00 -5.78556478e-01 -4.55283403e-01
-6.92189559e-02 7.59133995e-01 5.61751127e-01 -1.86380789e-01
-4.40592349e-01 -6.76376224e-01 -2.11339071e-01 2.40171095e-03
-1.43863750e-03 1.53869718e-01 -6.64390028e-01 -1.18847167e+00
1.63543344e-01 -9.35639560e-01 3.69728476e-01 -8.27636182e-01
-1.16438234e+00 5.99237382e-01 -9.24236700e-02 -8.61454666e-01
-2.92561293e-01 -6.45913541e-01 -9.14607525e-01 6.94461167e-01
-1.34361184e+00 -7.80033767e-01 -2.95958072e-01 1.49678826e-01
6.81698203e-01 -5.50048351e-01 8.74534011e-01 -2.96430290e-02
-6.79530442e-01 5.20124257e-01 1.35429963e-01 2.95191526e-01
7.16052413e-01 -1.67341554e+00 -3.16544294e-01 8.08486819e-01
1.05979301e-01 5.53770304e-01 1.07136679e+00 -5.48372209e-01
-8.94652307e-01 -8.98597777e-01 2.13197970e+00 -6.12898767e-01
7.52961874e-01 2.65134335e-01 -8.54668856e-01 4.18403774e-01
7.02810168e-01 -5.41050911e-01 8.48311186e-01 -1.04240976e-01
-1.84180275e-01 2.49663234e-01 -9.50537741e-01 5.45989335e-01
5.89311957e-01 -4.46900249e-01 -1.52414465e+00 3.24921876e-01
2.00291172e-01 -3.90944600e-01 -1.03457439e+00 1.86285526e-01
4.08715457e-01 -1.04920113e+00 2.81929940e-01 -5.68531275e-01
1.28827417e+00 6.98386226e-03 3.16410333e-01 -1.32844734e+00
1.45527914e-01 -6.79477274e-01 -1.53185353e-01 1.41302991e+00
5.02928972e-01 -2.76641905e-01 4.74520773e-01 4.49563861e-01
-4.99197274e-01 -5.66784799e-01 -7.08860874e-01 -3.06602627e-01
5.57580411e-01 -3.99497092e-01 2.51393691e-02 9.24138367e-01
5.77351689e-01 6.09466016e-01 1.25596315e-01 -5.84450364e-01
3.30704808e-01 1.95990339e-01 6.62167251e-01 -1.63530052e+00
8.59951526e-02 -1.06726706e+00 8.64409804e-02 -7.22272575e-01
6.40780747e-01 -9.11760211e-01 -3.37963812e-02 -1.62535214e+00
5.24437010e-01 -1.69583112e-01 2.61889368e-01 2.47944325e-01
-2.59562671e-01 2.96870649e-01 -6.41687438e-02 2.22590908e-01
-7.69999385e-01 3.41019064e-01 8.88305366e-01 -1.48348123e-01
-8.26068819e-02 1.78601831e-01 -6.52652800e-01 1.02750444e+00
7.89250374e-01 -4.50326145e-01 3.84837538e-02 -1.70990735e-01
7.51458108e-01 1.86366513e-01 2.41304457e-01 -5.93195915e-01
5.61803520e-01 -1.14060819e-01 1.97109222e-01 -6.74193025e-01
-3.09822887e-01 -8.89977440e-02 -2.67783523e-01 3.83445263e-01
-7.63535976e-01 2.35561445e-01 -1.79078907e-01 -3.54518257e-02
-5.04358411e-01 -1.02757549e+00 5.96950769e-01 -1.55881628e-01
-4.62750904e-02 -2.64758974e-01 -7.23587215e-01 4.93191779e-01
8.14645112e-01 -1.69710338e-01 -5.61744153e-01 -2.69566625e-01
-4.05703545e-01 4.07422148e-02 3.79611343e-01 1.02236822e-01
4.15451527e-01 -9.94603932e-01 -1.28028750e+00 -2.43959084e-01
6.79793656e-02 -2.51018077e-01 -1.18669525e-01 1.04355276e+00
-8.98633122e-01 5.61038315e-01 -3.47019988e-03 -2.16805130e-01
-1.46353340e+00 3.85637069e-03 -1.39951650e-02 -6.53769672e-01
-5.88941574e-01 5.47977567e-01 -6.31048381e-01 -3.49380851e-01
8.39253739e-02 -1.90584846e-02 -9.44738746e-01 4.72134441e-01
6.41097903e-01 8.14321697e-01 3.38079274e-01 -8.43527973e-01
1.49794430e-01 3.01752716e-01 -1.70744272e-04 -3.94347966e-01
1.27796555e+00 -8.58263671e-02 -4.50883031e-01 8.81217182e-01
7.44761288e-01 4.55054760e-01 -5.18298090e-01 -1.15208097e-01
4.90119517e-01 -1.27713442e-01 -1.42848408e-02 -1.04366648e+00
-1.27159253e-01 9.01454985e-01 -2.23042190e-01 4.12364751e-01
6.95265591e-01 -8.95493776e-02 5.95279157e-01 5.92383742e-01
-5.53473309e-02 -1.73032951e+00 2.75996476e-01 9.65558827e-01
1.07122397e+00 -1.20046926e+00 1.70641944e-01 -8.76296684e-02
-6.68074489e-01 1.55167544e+00 3.15589815e-01 -2.05349028e-02
1.70222044e-01 4.45316583e-02 -1.03902198e-01 -3.02249104e-01
-8.17828596e-01 8.36444199e-02 4.29440916e-01 -1.55410003e-02
8.55815291e-01 6.27899393e-02 -1.36984193e+00 1.13871706e+00
-8.04666519e-01 -1.60037845e-01 1.07178426e+00 6.84893906e-01
-6.27373815e-01 -1.09867966e+00 -3.64455372e-01 4.72094029e-01
-9.73640859e-01 5.05206324e-02 -6.71997428e-01 6.62054777e-01
-1.56028122e-01 1.20828676e+00 -1.75448909e-01 1.16658658e-01
2.89920300e-01 2.44235367e-01 5.56702256e-01 -6.27706707e-01
-1.16213500e+00 -2.09924400e-01 4.54909116e-01 1.90052003e-01
-5.99633873e-01 -9.25017536e-01 -1.09647465e+00 -5.96433938e-01
-2.12234899e-01 5.83242893e-01 5.82146823e-01 1.32514942e+00
-2.83517182e-01 4.61426258e-01 5.18128932e-01 -2.26560891e-01
-8.49993169e-01 -1.37722278e+00 -3.06982368e-01 3.49485606e-01
-9.01553966e-03 -2.06864089e-01 -5.01658738e-01 1.36022642e-01] | [11.797853469848633, 9.423715591430664] |
1909ce6c-4b50-418f-87ff-661d5ca23c15 | learning-spatio-temporal-features-with-two | 1905.02540 | null | https://arxiv.org/abs/1905.02540v2 | https://arxiv.org/pdf/1905.02540v2.pdf | Learning Spatio-Temporal Features with Two-Stream Deep 3D CNNs for Lipreading | We focus on the word-level visual lipreading, which requires recognizing the word being spoken, given only the video but not the audio. State-of-the-art methods explore the use of end-to-end neural networks, including a shallow (up to three layers) 3D convolutional neural network (CNN) + a deep 2D CNN (e.g., ResNet) as the front-end to extract visual features, and a recurrent neural network (e.g., bidirectional LSTM) as the back-end for classification. In this work, we propose to replace the shallow 3D CNNs + deep 2D CNNs front-end with recent successful deep 3D CNNs --- two-stream (i.e., grayscale video and optical flow streams) I3D. We evaluate different combinations of front-end and back-end modules with the grayscale video and optical flow inputs on the LRW dataset. The experiments show that, compared to the shallow 3D CNNs + deep 2D CNNs front-end, the deep 3D CNNs front-end with pre-training on the large-scale image and video datasets (e.g., ImageNet and Kinetics) can improve the classification accuracy. Also, we demonstrate that using the optical flow input alone can achieve comparable performance as using the grayscale video as input. Moreover, the two-stream network using both the grayscale video and optical flow inputs can further improve the performance. Overall, our two-stream I3D front-end with a Bi-LSTM back-end results in an absolute improvement of 5.3% over the previous art on the LRW dataset. | ['Kris Kitani', 'Xinshuo Weng'] | 2019-05-04 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [-8.39303955e-02 -2.42472872e-01 -3.11630875e-01 -1.67993605e-01
-6.76298022e-01 -2.62966782e-01 3.86423409e-01 -6.43097162e-01
-6.16475523e-01 1.96959347e-01 2.78750509e-01 -6.19733810e-01
6.58971667e-01 -3.93763751e-01 -7.73389816e-01 -3.94767940e-01
1.30678490e-01 -2.53794849e-01 2.53968775e-01 -1.14846090e-02
2.79872388e-01 4.99944717e-01 -1.95430446e+00 4.75071132e-01
2.83800244e-01 1.62735868e+00 1.01616301e-01 1.13571692e+00
-3.41462225e-01 1.07892430e+00 -5.55761874e-01 1.49425613e-02
3.06206793e-01 -3.93221706e-01 -6.23171031e-01 -8.95107587e-05
7.66763806e-01 -8.78865182e-01 -7.36254275e-01 7.34957814e-01
8.87425244e-01 1.56906366e-01 4.05065507e-01 -1.46959603e+00
-7.65780807e-01 -8.63465294e-02 -5.11861265e-01 6.54501542e-02
4.06658292e-01 7.02556133e-01 7.40678310e-01 -1.29496050e+00
4.17837203e-01 1.49737394e+00 6.39528096e-01 9.38433766e-01
-7.59495020e-01 -8.07113409e-01 1.10065676e-01 5.89906335e-01
-1.11408114e+00 -8.53281856e-01 9.84308422e-01 -2.93784976e-01
1.40173793e+00 -2.88738459e-01 7.20125318e-01 1.31058502e+00
4.45558056e-02 1.06983197e+00 9.53148305e-01 -2.17814431e-01
1.45643175e-01 -1.21631734e-01 -4.75388952e-02 8.17832708e-01
-3.37362081e-01 2.73286581e-01 -1.02496660e+00 3.35648060e-01
6.52750790e-01 2.47441921e-02 -4.29904014e-01 1.29274279e-01
-9.09606576e-01 6.14360750e-01 5.29242933e-01 2.05659177e-02
-2.42922336e-01 3.75239313e-01 6.35070741e-01 3.30368817e-01
7.00536489e-01 -1.85302556e-01 -5.62902808e-01 -3.86066407e-01
-1.26590323e+00 -2.23020837e-01 7.77037263e-01 7.68794596e-01
9.35226500e-01 3.72660846e-01 -2.05492079e-01 9.16488290e-01
5.57748795e-01 5.58381855e-01 7.60089278e-01 -9.63215947e-01
6.34704113e-01 3.72179747e-01 -3.16782951e-01 -6.43061161e-01
-3.74209911e-01 -1.60675824e-01 -8.62426758e-01 6.74590170e-01
2.56435990e-01 -2.74890274e-01 -1.31232023e+00 1.72589290e+00
1.52970538e-01 5.48163235e-01 4.28860903e-01 1.19509518e+00
1.24958980e+00 7.45158732e-01 -1.10022224e-01 2.81438772e-02
1.19315994e+00 -1.44978690e+00 -8.45606327e-01 -3.28361206e-02
5.21575511e-01 -8.99566650e-01 1.19186592e+00 1.50812507e-01
-1.35848343e+00 -9.95132983e-01 -1.02438474e+00 -5.24347842e-01
-6.47025287e-01 2.59301275e-01 6.46589100e-02 5.23727417e-01
-1.65444005e+00 3.30382079e-01 -6.29475653e-01 -2.37334594e-01
3.48553330e-01 3.98244917e-01 -5.69846570e-01 -1.16511598e-01
-1.20049441e+00 5.98086238e-01 -1.77997395e-01 3.89578283e-01
-1.09593940e+00 -7.53500581e-01 -1.05313516e+00 -1.02142589e-02
9.58924145e-02 -6.88450098e-01 1.27218664e+00 -1.11006653e+00
-2.29996777e+00 7.04259217e-01 -5.62747180e-01 -4.50776249e-01
6.11490130e-01 -1.40474379e-01 -1.00258673e-02 5.70312262e-01
-3.21456522e-01 1.12657142e+00 1.14286649e+00 -1.02263403e+00
-6.40962481e-01 -1.89249545e-01 6.15940355e-02 1.97035074e-01
-3.09158772e-01 -1.00906764e-03 -5.90261102e-01 -4.94315743e-01
-7.52120614e-02 -7.88305640e-01 3.83794338e-01 4.76878583e-01
-3.20818007e-01 -2.87511200e-01 1.39279175e+00 -8.07427049e-01
7.43156850e-01 -2.22171187e+00 -1.63193017e-01 -3.58366460e-01
1.92690358e-01 6.96888983e-01 -5.16459942e-01 4.67835665e-02
-6.86524510e-02 2.44731814e-01 1.14296570e-01 -9.79712486e-01
-6.18514642e-02 7.05276281e-02 -5.55455275e-02 5.82163215e-01
5.23717701e-01 1.11789620e+00 -6.28485560e-01 -3.04562151e-01
4.47616458e-01 9.10227835e-01 -6.04143798e-01 5.72380006e-01
-2.98747309e-02 1.99990049e-01 5.04086651e-02 5.98614514e-01
7.73792684e-01 -1.96747884e-01 -3.37587625e-01 -3.34863275e-01
-1.99425220e-01 5.13713360e-01 -8.31844985e-01 1.88686204e+00
-8.58559072e-01 1.09075379e+00 3.07291448e-01 -9.15129662e-01
9.56957936e-01 7.60403454e-01 3.48112047e-01 -8.94802213e-01
7.76657686e-02 3.18723977e-01 -2.33573854e-01 -7.85607457e-01
2.18471050e-01 -1.03189209e-02 4.28633273e-01 4.22482282e-01
5.38855255e-01 -3.70481946e-02 -3.01088423e-01 -9.00549367e-02
9.14221048e-01 2.86449969e-01 -3.37491393e-01 1.62162930e-01
8.56402755e-01 -5.33467114e-01 4.55950826e-01 3.59275043e-01
-4.54555809e-01 9.36301768e-01 4.62494344e-01 -2.91297466e-01
-9.98835742e-01 -6.93880558e-01 3.62956822e-01 9.88991737e-01
1.16807707e-01 -2.31106773e-01 -6.84462726e-01 -6.36893451e-01
-4.30064654e-04 6.85105696e-02 -5.85693777e-01 -3.15063596e-01
-5.46475530e-01 -4.47140932e-02 7.92976797e-01 6.40043974e-01
1.00117147e+00 -1.09256649e+00 -5.94007909e-01 4.15460691e-02
-2.13419005e-01 -1.52785099e+00 -8.13286245e-01 1.18001282e-01
-5.16846120e-01 -8.75638247e-01 -1.20540798e+00 -9.35257852e-01
2.90239453e-01 3.19737673e-01 5.67714393e-01 1.96456462e-02
-8.34232662e-03 5.20207107e-01 -4.08214867e-01 -1.32704169e-01
-2.76159048e-01 8.72939229e-02 -1.34180458e-02 3.50475937e-01
2.41669804e-01 -3.97202939e-01 -8.72415662e-01 2.35198647e-01
-7.21890867e-01 2.43321300e-01 4.40535873e-01 8.49217713e-01
3.20492089e-01 -6.30412817e-01 4.93306398e-01 -7.21346959e-02
4.06390309e-01 -1.47720814e-01 -3.85291398e-01 2.21190695e-02
-2.85283297e-01 -2.49024816e-02 6.04444921e-01 -6.78335369e-01
-9.26425874e-01 -6.34671981e-03 -6.57294989e-01 -1.25110126e+00
-2.02173874e-01 1.80027306e-01 -5.40447235e-03 -2.54935741e-01
1.46517903e-01 3.60210091e-01 4.47206616e-01 -7.14796185e-01
2.21024349e-01 1.12495673e+00 5.79603016e-01 -4.99536609e-03
4.71157372e-01 4.59469140e-01 2.26070806e-02 -9.63941276e-01
-5.05589187e-01 -5.90733051e-01 -5.02573133e-01 -5.22104383e-01
1.13673854e+00 -1.19643247e+00 -1.14057732e+00 1.14668751e+00
-1.55076885e+00 -7.47305095e-01 -9.90714580e-02 6.71804249e-01
-5.80221415e-01 2.26998389e-01 -9.24004674e-01 -9.32571232e-01
-4.96894449e-01 -1.44756186e+00 1.21166956e+00 3.11196476e-01
1.99116766e-01 -9.60006177e-01 -2.41550222e-01 4.35544729e-01
6.71960235e-01 -1.46594629e-01 6.12310231e-01 -5.23819029e-01
-3.96262348e-01 8.31000358e-02 -4.58054125e-01 8.88285995e-01
6.43060952e-02 3.91097516e-02 -1.60671771e+00 -2.01968819e-01
-1.37549927e-02 -6.22664750e-01 1.17336822e+00 5.07389367e-01
1.12294531e+00 -3.26711208e-01 1.75897494e-01 9.19904768e-01
1.14341128e+00 1.48305804e-01 6.58366263e-01 -1.82901040e-01
7.26620138e-01 5.68705738e-01 8.74179453e-02 1.23515643e-01
6.50944293e-01 6.59183204e-01 5.17094374e-01 -6.53112113e-01
-1.03089392e+00 -4.53509241e-01 8.52544785e-01 1.03020120e+00
-8.02298822e-03 -3.17763418e-01 -7.91392267e-01 4.21152532e-01
-1.68621862e+00 -7.82734811e-01 3.26117009e-01 1.96879470e+00
6.56494856e-01 5.18107675e-02 2.10336447e-01 3.40103477e-01
7.68343627e-01 4.08319235e-01 -7.65970051e-01 -5.18266320e-01
-1.64936960e-01 2.80913353e-01 2.35842824e-01 9.19084728e-01
-7.79062569e-01 1.12045348e+00 5.39919376e+00 7.29622066e-01
-1.99184060e+00 2.63403475e-01 6.07452393e-01 -1.86251163e-01
3.67194712e-02 -4.86626416e-01 -8.71707976e-01 3.84423345e-01
1.07045650e+00 5.08910716e-01 4.51680541e-01 4.80141550e-01
6.32312715e-01 5.18201888e-02 -9.27020192e-01 1.24455762e+00
3.71046931e-01 -1.32904124e+00 5.35475537e-02 1.41185746e-01
3.44799876e-01 3.99345577e-01 1.67539790e-01 2.96012223e-01
-1.60854816e-01 -1.00935817e+00 9.72286344e-01 4.58872557e-01
1.39062524e+00 -3.94856334e-01 6.75560534e-01 2.29746327e-01
-1.44098616e+00 -3.29478197e-02 -1.84509978e-02 -5.43893166e-02
4.51952741e-02 4.13275182e-01 -5.40775359e-01 2.95854270e-01
9.24781859e-01 1.02160954e+00 -4.99756038e-02 8.04702163e-01
-2.97541142e-01 6.68807089e-01 -3.34046245e-01 -8.82961899e-02
5.34627378e-01 2.55199611e-01 4.22533274e-01 1.43255401e+00
1.11965373e-01 -1.39675528e-01 -7.60417730e-02 8.72432053e-01
-4.27969307e-01 -2.44869977e-01 -5.30579090e-01 1.61497056e-01
-1.03140920e-02 1.14022100e+00 -7.58915842e-02 -4.68479276e-01
-7.86916256e-01 9.81348336e-01 1.78191617e-01 8.51478636e-01
-6.76903009e-01 -7.36873031e-01 1.16026568e+00 4.95403819e-03
4.56158310e-01 -3.61460149e-01 -1.30945975e-02 -1.21285307e+00
3.02165821e-02 -5.84987521e-01 -2.55982988e-02 -1.13830316e+00
-1.15196335e+00 7.88539290e-01 -6.18063748e-01 -1.11760759e+00
-4.76654232e-01 -9.70246077e-01 -6.39949322e-01 1.14948654e+00
-2.44665551e+00 -1.02993631e+00 -4.94712114e-01 1.10520554e+00
7.86516726e-01 -1.13264345e-01 5.42092919e-01 3.29794377e-01
-5.65278292e-01 7.53212333e-01 -2.52874702e-01 5.84991276e-01
7.19336331e-01 -8.90547454e-01 4.38959569e-01 6.72374129e-01
-4.05978821e-02 1.06123060e-01 -1.91938803e-02 -2.66003013e-01
-1.68209255e+00 -1.06948519e+00 9.92119789e-01 2.89013863e-01
4.80478138e-01 -7.48466551e-01 -8.20807815e-01 3.02235574e-01
4.44086969e-01 6.03769481e-01 2.88803339e-01 -4.41518754e-01
-5.67671239e-01 -2.92376727e-01 -1.20000255e+00 3.61937732e-01
1.22451138e+00 -1.10869431e+00 -2.76286602e-01 -1.58438191e-01
9.54745770e-01 -4.18605506e-01 -6.22718215e-01 4.81323242e-01
7.77602077e-01 -1.03553438e+00 9.44227576e-01 -4.94733274e-01
5.08076191e-01 -1.12492353e-01 -2.64597565e-01 -1.18478096e+00
3.53325188e-01 -9.59316015e-01 -2.41108298e-01 1.02813685e+00
4.37760890e-01 -8.10614049e-01 7.91758239e-01 1.48824409e-01
-2.64954239e-01 -9.13290799e-01 -1.18374157e+00 -6.65638566e-01
5.46331843e-03 -7.32724607e-01 2.08675995e-01 3.56272072e-01
-1.01017043e-01 3.57151330e-01 -4.49932069e-01 -8.78131464e-02
2.84420103e-01 -8.24505091e-02 6.94531679e-01 -8.78187537e-01
6.80434927e-02 -4.30786908e-01 -5.66965461e-01 -1.64154530e+00
6.61282539e-01 -7.92475581e-01 1.94812223e-01 -1.44488537e+00
-2.04340264e-01 4.62202542e-02 -2.75656462e-01 7.56458521e-01
2.35105008e-02 4.39716160e-01 5.63744128e-01 5.89945838e-02
-2.57396668e-01 7.03416944e-01 1.48290348e+00 -2.66819060e-01
-2.73508191e-01 -1.83992535e-01 -9.10703391e-02 5.25529981e-01
6.70681596e-01 -9.85560566e-02 -2.57040322e-01 -5.21507621e-01
-2.55326450e-01 2.08330438e-01 6.45420372e-01 -1.03054917e+00
6.44360185e-01 3.01562011e-01 3.83949369e-01 -5.24149597e-01
8.73100042e-01 -5.79128861e-01 -7.54013419e-01 4.57297176e-01
-3.55177552e-01 6.21226095e-02 4.30413008e-01 2.43574500e-01
-5.68599641e-01 1.47354767e-01 8.35205078e-01 -2.81626545e-02
-6.46063566e-01 5.25751770e-01 -4.76072818e-01 1.06849208e-01
5.13027072e-01 -3.85657638e-01 -5.81525087e-01 -7.20011592e-01
-6.48293257e-01 4.53604870e-02 4.20383364e-02 5.39052308e-01
9.84667540e-01 -1.14973903e+00 -6.09721065e-01 6.38145387e-01
-3.43697488e-01 -1.42172426e-01 2.03137413e-01 8.44609737e-01
-2.70795375e-01 5.26518226e-01 -3.85078877e-01 -8.88960302e-01
-1.25568771e+00 3.44317317e-01 7.24129260e-01 1.02059506e-01
-5.11860967e-01 9.29179251e-01 3.57483588e-02 -3.55271846e-01
7.36055911e-01 -4.68606621e-01 -7.61465505e-02 1.33683190e-01
4.78992343e-01 3.87697756e-01 7.11643472e-02 -6.63371027e-01
-4.80121791e-01 1.02453101e+00 2.98218369e-01 -4.07212019e-01
1.30101311e+00 -1.99505091e-01 2.61681229e-01 5.25472462e-01
1.98446763e+00 -3.89094025e-01 -1.57134080e+00 -2.00712189e-01
-5.43543398e-01 -2.25655124e-01 4.48395491e-01 -6.22880876e-01
-1.64824367e+00 1.69883466e+00 6.74858809e-01 -3.77045982e-02
1.36344826e+00 -1.21709920e-01 9.79150176e-01 7.04258084e-02
-8.29522163e-02 -9.42273915e-01 4.38176394e-01 6.64887667e-01
9.11865771e-01 -1.27377129e+00 -6.08730316e-01 -7.95421079e-02
-5.75415015e-01 1.42341530e+00 5.35031915e-01 6.14826493e-02
9.51236784e-01 4.17018324e-01 4.63563681e-01 2.16707543e-01
-1.10331619e+00 -5.76950967e-01 2.98847914e-01 5.25762379e-01
2.70548493e-01 -4.76431400e-01 3.81514281e-01 2.36567199e-01
5.15693724e-02 4.41577822e-01 3.44912171e-01 6.44978523e-01
-3.08741536e-02 -6.67786658e-01 -1.04842812e-01 1.11129276e-01
-3.46445322e-01 -2.91507572e-01 -2.43989840e-01 6.56713128e-01
2.25348413e-01 1.17944121e+00 3.82049263e-01 -6.73592091e-01
4.02639359e-01 2.97937244e-01 2.55667597e-01 -1.87423840e-01
-7.51972497e-01 1.06874563e-01 -1.49573341e-01 -9.02455270e-01
-5.84464669e-01 -2.87371546e-01 -1.10742176e+00 -4.80985224e-01
-1.58539087e-01 -3.42923790e-01 1.03139997e+00 8.41252983e-01
5.64824283e-01 2.94496208e-01 8.52408588e-01 -1.29109442e+00
-2.69210219e-01 -1.08313549e+00 -3.57108325e-01 1.81584045e-01
1.26485860e+00 -5.29807448e-01 -8.27436447e-01 9.82285663e-02] | [14.333357810974121, 5.034838676452637] |
2d99fa6a-83d3-47dd-96af-d8744a50636b | llms-for-knowledge-graph-construction-and | 2305.13168 | null | https://arxiv.org/abs/2305.13168v1 | https://arxiv.org/pdf/2305.13168v1.pdf | LLMs for Knowledge Graph Construction and Reasoning: Recent Capabilities and Future Opportunities | This paper presents an exhaustive quantitative and qualitative evaluation of Large Language Models (LLMs) for Knowledge Graph (KG) construction and reasoning. We employ eight distinct datasets that encompass aspects including entity, relation and event extraction, link prediction, and question answering. Empirically, our findings suggest that GPT-4 outperforms ChatGPT in the majority of tasks and even surpasses fine-tuned models in certain reasoning and question-answering datasets. Moreover, our investigation extends to the potential generalization ability of LLMs for information extraction, which culminates in the presentation of the Virtual Knowledge Extraction task and the development of the VINE dataset. Drawing on these empirical findings, we further propose AutoKG, a multi-agent-based approach employing LLMs for KG construction and reasoning, which aims to chart the future of this field and offer exciting opportunities for advancement. We anticipate that our research can provide invaluable insights for future undertakings of KG\footnote{Code and datasets will be available in https://github.com/zjunlp/AutoKG. | ['Ningyu Zhang', 'Huajun Chen', 'Shumin Deng', 'Yunzhi Yao', 'Yixin Ou', 'Shuofei Qiao', 'Jing Chen', 'Xiaohan Wang', 'Yuqi Zhu'] | 2023-05-22 | null | null | null | null | ['link-prediction', 'graph-construction', 'event-extraction'] | ['graphs', 'graphs', 'natural-language-processing'] | [-2.36020193e-01 8.31219196e-01 -4.93620843e-01 1.59189645e-02
-6.29736125e-01 -6.04691803e-01 8.71484518e-01 6.78012729e-01
-8.67323875e-02 1.04232132e+00 3.48153621e-01 -6.21239185e-01
-7.15201318e-01 -1.26116526e+00 -5.43016076e-01 -7.01354071e-03
-4.30203825e-01 7.27305412e-01 3.33617032e-01 -1.92263126e-01
1.49224386e-01 2.96669960e-01 -1.22943282e+00 3.04797977e-01
7.50795245e-01 7.89120853e-01 -2.07279891e-01 4.33409095e-01
-2.79683739e-01 1.77068186e+00 -3.12546492e-01 -9.54126120e-01
-1.07598968e-01 3.25723626e-02 -1.50247586e+00 -3.21110219e-01
1.99945554e-01 2.42833346e-02 -4.21552479e-01 6.87437952e-01
5.46208918e-01 1.29901126e-01 3.88460994e-01 -1.65764999e+00
-6.52293205e-01 1.11655581e+00 -7.96273202e-02 2.31553167e-01
8.50914955e-01 -4.18780968e-02 1.28509378e+00 -5.52411258e-01
9.64007795e-01 1.17801070e+00 7.22322404e-01 1.99399799e-01
-8.61616433e-01 -5.07331312e-01 1.73690394e-01 5.37427485e-01
-1.44477332e+00 -4.61153179e-01 5.35154819e-01 -2.64494240e-01
1.51906371e+00 2.22759470e-01 5.91253579e-01 7.81632304e-01
7.33728334e-02 1.00277531e+00 1.08368874e+00 -6.14432275e-01
1.25469893e-01 2.97791451e-01 3.02033067e-01 1.13183701e+00
4.31787193e-01 -2.60137375e-02 -8.30573618e-01 -2.51659721e-01
5.54553270e-01 -6.49129510e-01 -9.43846628e-02 -3.85897547e-01
-1.14491832e+00 7.44552016e-01 2.60018378e-01 2.05719858e-01
-5.29035628e-01 2.64430612e-01 2.22742707e-01 4.70410913e-01
4.81867343e-01 4.69418198e-01 -7.00420439e-01 -2.81023502e-01
-5.70393980e-01 4.90822077e-01 1.46787322e+00 1.13453984e+00
5.41711688e-01 -2.46396586e-01 -8.44603330e-02 6.10110998e-01
3.64905745e-01 1.06002860e-01 -6.27339631e-02 -1.04649687e+00
6.58732593e-01 1.00729823e+00 4.59006503e-02 -1.01696610e+00
-5.99246264e-01 -2.59542465e-01 -2.85267353e-01 -3.13912511e-01
3.62969816e-01 -4.01329994e-01 -3.79076719e-01 1.55973804e+00
5.23481965e-01 1.08465202e-01 4.73608404e-01 4.28584278e-01
1.45077312e+00 2.22171485e-01 5.46009004e-01 -9.53071862e-02
1.56094074e+00 -9.83403265e-01 -7.01884151e-01 -3.41118611e-02
1.03636861e+00 -5.07262886e-01 5.54849446e-01 1.87265873e-01
-1.34082687e+00 -4.34859619e-02 -6.60290241e-01 -1.31027564e-01
-8.74816895e-01 -4.63906489e-02 1.33517182e+00 5.10272205e-01
-1.22755158e+00 4.00715500e-01 -8.00750852e-01 -9.25838470e-01
5.26262641e-01 2.23937452e-01 -2.80787617e-01 -1.06124058e-01
-1.51493204e+00 1.21077299e+00 8.25899482e-01 -3.68779413e-02
-4.20452178e-01 -7.45642066e-01 -8.70503545e-01 -7.73524791e-02
9.47496831e-01 -1.12953198e+00 1.11451268e+00 -4.21642475e-02
-1.16371465e+00 6.79584920e-01 3.11050322e-02 -8.19083154e-01
3.03828895e-01 -8.94335005e-03 -6.80733502e-01 1.67259559e-01
1.97433382e-01 5.31297147e-01 4.75055389e-02 -9.74992037e-01
-6.73957765e-01 -1.48047850e-01 5.11652410e-01 1.37735918e-01
-1.61185592e-01 2.15364993e-01 -6.53944433e-01 -3.86629254e-01
-3.08636397e-01 -6.25167608e-01 -1.29937202e-01 -6.07668400e-01
-6.58748806e-01 -6.85236394e-01 4.85902607e-01 -6.28287852e-01
1.58819139e+00 -1.47356462e+00 9.67511162e-02 3.32073480e-01
3.09541643e-01 1.69913605e-01 -1.16068525e-02 1.31317973e+00
1.11948840e-01 1.96762443e-01 3.14996481e-01 6.18191762e-03
2.86964804e-01 2.79386282e-01 -2.05865607e-01 -3.46030667e-02
2.89605260e-01 1.59814024e+00 -9.06121492e-01 -8.81984472e-01
9.39998329e-02 1.49422586e-01 -2.09805965e-01 -3.11413050e-01
-5.94030857e-01 -4.21950035e-02 -8.63574505e-01 9.07991350e-01
1.87681764e-01 -6.69929385e-01 4.59833324e-01 -1.17668442e-01
-2.03038622e-02 4.86410528e-01 -1.00152862e+00 1.56140804e+00
-1.99541956e-01 3.25855196e-01 -1.65583596e-01 -8.69116068e-01
5.91769218e-01 5.31981885e-01 3.17291617e-01 -6.21025860e-01
-9.75122675e-02 -3.81956995e-02 -8.68593380e-02 -7.69841254e-01
5.07353365e-01 1.35197595e-01 -1.23659447e-01 6.31615937e-01
7.06446767e-02 -2.15005562e-01 6.56211197e-01 8.44027638e-01
1.36244690e+00 2.75851816e-01 5.01097620e-01 -1.80964053e-01
4.96363372e-01 3.52970898e-01 9.49091986e-02 7.47997344e-01
-4.78500091e-02 -2.43307143e-01 7.95929790e-01 -3.63206804e-01
-4.70879316e-01 -8.06646585e-01 -3.74852903e-02 8.21116626e-01
-1.86457440e-01 -9.46057677e-01 -5.30311286e-01 -8.59017491e-01
4.02208775e-01 1.01284969e+00 -5.16977549e-01 7.67261162e-02
-3.56436521e-01 -7.70878136e-01 8.88094604e-01 6.02387309e-01
5.88775873e-01 -1.35488737e+00 -1.77504271e-01 1.97623670e-01
-3.29820991e-01 -1.42197430e+00 4.70694244e-01 -1.46642655e-01
-7.21249223e-01 -1.48745716e+00 4.43830080e-02 -7.38852918e-01
3.93097341e-01 -2.86774278e-01 1.29411507e+00 2.96849310e-01
-8.67118835e-02 8.63286138e-01 -5.62741995e-01 -5.70692360e-01
-3.37873161e-01 4.39240634e-01 -1.69679016e-01 -6.72740400e-01
5.88603973e-01 -6.49802625e-01 -3.67588073e-01 5.32842055e-02
-6.69415414e-01 2.54368305e-01 5.35450161e-01 1.81567237e-01
1.91819414e-01 4.60403204e-01 1.07701087e+00 -1.13943040e+00
1.09670639e+00 -7.92930722e-01 -3.54499638e-01 5.91662943e-01
-8.85700285e-01 7.27024451e-02 1.58137918e-01 1.97137579e-01
-1.14867342e+00 -5.95639944e-01 -4.87558059e-02 2.07811415e-01
-4.20127958e-02 1.21381497e+00 1.13004461e-01 -1.58395961e-01
4.71748680e-01 -5.01386076e-02 -3.44086111e-01 -2.38333628e-01
6.46820903e-01 3.04239601e-01 2.90635973e-01 -9.83074129e-01
7.48051584e-01 3.02995265e-01 5.83620481e-02 -5.58912277e-01
-7.84372211e-01 -2.91564584e-01 -3.40696454e-01 -1.64907947e-01
5.92409670e-01 -9.23335671e-01 -9.71504450e-01 1.55452266e-01
-9.35775101e-01 -6.10627353e-01 -3.92839521e-01 1.90962598e-01
-4.85469431e-01 4.14004326e-01 -9.12846923e-01 -8.24029267e-01
-4.26545352e-01 -5.51600158e-01 7.68811703e-01 1.66509271e-01
-5.67187011e-01 -1.64614427e+00 -1.89678278e-02 1.02179253e+00
2.08908200e-01 2.57426977e-01 1.27026963e+00 -9.06009853e-01
-9.24744368e-01 -3.62944484e-01 -3.29013288e-01 -2.40060315e-01
-5.49196377e-02 -7.45393336e-02 -6.27610028e-01 -7.23640174e-02
-7.74029493e-01 -7.23362386e-01 6.30030572e-01 5.78214042e-02
8.26185346e-01 -2.42682382e-01 -6.49971068e-01 3.61613110e-02
1.39343679e+00 4.24298942e-02 6.10748708e-01 6.08644485e-01
5.76707363e-01 7.29526341e-01 6.68509245e-01 3.46276581e-01
1.26213872e+00 3.53224576e-01 2.24907652e-01 1.95081919e-01
-3.01997185e-01 -3.99005055e-01 -7.09292367e-02 8.15327823e-01
-6.02937281e-01 -5.34862578e-01 -1.20368969e+00 7.36785114e-01
-2.12521005e+00 -9.42302644e-01 -2.36214578e-01 1.73246360e+00
8.16356719e-01 2.55860537e-02 1.90517884e-02 6.82265088e-02
2.06300884e-01 1.21907786e-01 -2.94880897e-01 -2.95898646e-01
-1.96989864e-01 4.32724416e-01 2.58011937e-01 5.32513857e-01
-8.18913281e-01 1.41831768e+00 6.09153175e+00 8.35672617e-01
-2.39873946e-01 -5.09816706e-02 8.80187377e-02 1.62354544e-01
-2.65594900e-01 3.12046707e-01 -7.91389346e-01 -3.52313183e-02
1.19694233e+00 -5.37881196e-01 5.96631289e-01 3.60168308e-01
5.15158623e-02 -3.77137572e-01 -9.76445496e-01 3.13156426e-01
-2.44236842e-01 -1.59018707e+00 1.36194855e-01 1.00023244e-02
6.04784846e-01 1.07629649e-01 -4.92751271e-01 6.89314425e-01
9.05526519e-01 -7.48790741e-01 3.66205543e-01 6.49444163e-01
1.69538841e-01 -6.12428904e-01 6.89681113e-01 4.32293147e-01
-1.29459512e+00 -1.08995244e-01 4.93540242e-02 -1.93153292e-01
1.94811583e-01 3.00341278e-01 -1.10875797e+00 1.46459293e+00
5.63093901e-01 7.39622176e-01 -6.96261585e-01 9.07825887e-01
-4.70461011e-01 5.80083311e-01 -1.13847330e-01 -7.32211992e-02
-9.39291790e-02 4.69360910e-02 4.22295660e-01 1.22907770e+00
-4.03065868e-02 3.35837781e-01 2.83778936e-01 6.78148329e-01
-9.00403112e-02 2.12366045e-01 -5.04208207e-01 -6.29484236e-01
6.19149327e-01 1.30893493e+00 -7.66660631e-01 -5.25326669e-01
-6.22250021e-01 3.82030547e-01 6.55088186e-01 6.48402154e-01
-6.39237940e-01 -4.05912250e-01 1.47512287e-01 2.98304819e-02
2.25887328e-01 -4.40664664e-02 -1.05086267e-01 -1.20635045e+00
-6.08368814e-02 -7.36306906e-01 8.85051906e-01 -1.07324243e+00
-1.25285065e+00 4.37916100e-01 4.18993473e-01 -5.00594318e-01
-4.70128685e-01 -3.09014678e-01 -4.38106388e-01 6.76068425e-01
-1.69004405e+00 -1.60139585e+00 -2.08533019e-01 5.11983275e-01
2.35359538e-02 -2.38452740e-02 1.07497025e+00 3.85857403e-01
-5.85237801e-01 3.01402241e-01 -4.91459191e-01 2.92927325e-01
3.98515224e-01 -1.18966675e+00 4.63191569e-01 5.39785683e-01
2.69246876e-01 6.33359253e-01 5.60644865e-01 -1.06291854e+00
-1.55847383e+00 -1.18406105e+00 1.26125741e+00 -8.16217065e-01
1.24922514e+00 8.89332127e-03 -6.29394293e-01 1.29796743e+00
4.51877326e-01 -3.70261818e-01 8.01830471e-01 4.52339977e-01
-2.88299508e-02 2.07500532e-01 -1.03650391e+00 5.82663178e-01
1.24942541e+00 -4.87943947e-01 -5.84862828e-01 6.00857735e-01
6.72188401e-01 -4.03613091e-01 -1.52748311e+00 5.14843047e-01
4.22719926e-01 -7.48979568e-01 1.04152799e+00 -8.39824617e-01
2.86144316e-01 -1.56065062e-01 -2.53054705e-02 -1.03999448e+00
-2.50482917e-01 -5.91112375e-01 -7.70970643e-01 1.51989949e+00
7.69197106e-01 -9.48775887e-01 9.07194257e-01 7.90537477e-01
1.08840570e-01 -9.77000833e-01 -6.06037915e-01 -6.17320716e-01
-2.96130832e-02 -8.01905990e-01 6.17356837e-01 1.14979303e+00
5.68937004e-01 3.35822374e-01 3.67283449e-02 4.30415332e-01
6.86128676e-01 3.39842618e-01 7.17922807e-01 -1.30204082e+00
-1.64486989e-01 -1.66105479e-01 -3.40174824e-01 -5.35747230e-01
4.12605673e-01 -1.11406004e+00 -7.84900308e-01 -2.32658243e+00
1.43225446e-01 -4.81280088e-01 -1.93306893e-01 1.02057624e+00
4.20500524e-02 7.71604804e-03 -2.06991974e-02 -2.34368388e-02
-9.32450056e-01 3.83227319e-01 1.26418042e+00 3.85977738e-02
-1.49159402e-01 1.66868716e-02 -8.16954195e-01 5.82976997e-01
9.32617188e-01 -1.86266944e-01 -6.89468265e-01 -1.31071791e-01
6.11390352e-01 2.90788233e-01 5.41567981e-01 -5.91621280e-01
5.67487121e-01 -2.07323059e-01 5.07648773e-02 -4.97476488e-01
3.48503709e-01 -5.11991680e-01 2.65991092e-01 2.01500505e-01
-2.40913779e-01 -6.85130134e-02 5.32603264e-01 4.30201620e-01
-1.59195915e-01 -3.40262800e-02 -1.74335942e-01 -3.36961746e-01
-1.02463210e+00 2.07011521e-01 -2.86804825e-01 2.30158448e-01
1.01641691e+00 -9.01254416e-02 -8.44942153e-01 -4.43099439e-01
-8.94279420e-01 8.41317654e-01 1.25222385e-01 4.19088751e-01
3.70877028e-01 -1.14995623e+00 -6.77982509e-01 -3.39251429e-01
1.68383777e-01 -1.27783537e-01 1.38854578e-01 1.09411991e+00
-3.55566412e-01 9.04421687e-01 1.42761916e-01 1.82977140e-01
-1.20550060e+00 5.20323038e-01 1.30496636e-01 -6.06612802e-01
-5.07660985e-01 6.47145391e-01 -4.31626230e-01 -4.77635384e-01
1.06714047e-01 -3.01301241e-01 -4.25021440e-01 1.95947867e-02
8.56314376e-02 4.87494290e-01 1.61960237e-02 -2.25498542e-01
-4.08719301e-01 1.05964489e-01 -2.07108751e-01 9.99675244e-02
1.45364702e+00 -1.57439590e-01 -4.04041380e-01 1.41637042e-01
4.11993414e-01 1.49604604e-01 -4.94725972e-01 -3.48687261e-01
4.82331097e-01 1.57510906e-01 -8.62809718e-02 -1.32459247e+00
-7.44649470e-01 2.44430467e-01 -2.77590543e-01 6.49237752e-01
9.22623336e-01 4.17604655e-01 6.02560639e-01 5.61149001e-01
7.05359519e-01 -9.17154849e-01 -2.66739279e-01 3.08498859e-01
7.73563623e-01 -1.04678738e+00 4.09029514e-01 -7.42812634e-01
-6.29354656e-01 8.07750702e-01 5.33960223e-01 3.22158426e-01
5.72979510e-01 2.76573002e-01 -2.25228041e-01 -9.07657444e-01
-1.29150259e+00 -3.84485394e-01 1.68219656e-01 5.60970426e-01
3.72479528e-01 4.60880585e-02 -2.91207820e-01 4.90807384e-01
-3.65263969e-01 4.95395005e-01 2.72547781e-01 1.19231832e+00
-1.15680583e-01 -1.43502903e+00 1.27033070e-01 4.88376737e-01
-3.32095891e-01 -3.59122783e-01 -7.33034790e-01 1.27774763e+00
1.21172993e-02 1.14907193e+00 -4.56513435e-01 -2.28914127e-01
4.14681435e-01 3.50550890e-01 6.79650664e-01 -5.47427893e-01
-5.79748094e-01 -4.80760843e-01 1.13529778e+00 -4.42126036e-01
-6.99825287e-01 -5.85440814e-01 -1.34819865e+00 -5.98977089e-01
-3.03836793e-01 4.48791385e-01 1.55092180e-01 8.97956908e-01
6.10449791e-01 5.27480721e-01 -3.62088829e-01 -3.96308899e-02
2.09926032e-02 -8.90115619e-01 -5.65408707e-01 -1.32214442e-01
-2.91078269e-01 -6.52739465e-01 -1.96672045e-04 -6.03538528e-02] | [9.407784461975098, 8.15185832977295] |
9964db82-5b34-42ff-96e1-8328ffbb0496 | conditional-sequential-modulation-for-1 | 2009.10390 | null | https://arxiv.org/abs/2009.10390v1 | https://arxiv.org/pdf/2009.10390v1.pdf | Conditional Sequential Modulation for Efficient Global Image Retouching | Photo retouching aims at enhancing the aesthetic visual quality of images that suffer from photographic defects such as over/under exposure, poor contrast, inharmonious saturation. Practically, photo retouching can be accomplished by a series of image processing operations. In this paper, we investigate some commonly-used retouching operations and mathematically find that these pixel-independent operations can be approximated or formulated by multi-layer perceptrons (MLPs). Based on this analysis, we propose an extremely light-weight framework - Conditional Sequential Retouching Network (CSRNet) - for efficient global image retouching. CSRNet consists of a base network and a condition network. The base network acts like an MLP that processes each pixel independently and the condition network extracts the global features of the input image to generate a condition vector. To realize retouching operations, we modulate the intermediate features using Global Feature Modulation (GFM), of which the parameters are transformed by condition vector. Benefiting from the utilization of $1\times1$ convolution, CSRNet only contains less than 37k trainable parameters, which is orders of magnitude smaller than existing learning-based methods. Extensive experiments show that our method achieves state-of-the-art performance on the benchmark MIT-Adobe FiveK dataset quantitively and qualitatively. Code is available at https://github.com/hejingwenhejingwen/CSRNet. | ['Chao Dong', 'Yihao Liu', 'Yu Qiao', 'Jingwen He'] | 2020-09-22 | conditional-sequential-modulation-for | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2014_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123580664.pdf | eccv-2020-8 | ['photo-retouching', 'image-retouching'] | ['computer-vision', 'computer-vision'] | [ 6.31332576e-01 -1.16116159e-01 -4.19470333e-02 4.51393100e-03
-3.20966452e-01 -9.12036672e-02 3.64473879e-01 -3.17452133e-01
-6.02187097e-01 7.57275760e-01 -1.58929303e-01 -3.58361095e-01
1.28038555e-01 -6.29386842e-01 -9.80032325e-01 -9.50410485e-01
3.26798826e-01 -7.00631440e-01 2.55320877e-01 -1.22974731e-01
4.64897603e-01 5.05689085e-01 -1.63401186e+00 3.98922920e-01
7.75830984e-01 1.10807729e+00 2.99738795e-01 9.13345695e-01
1.51070312e-01 7.90089846e-01 -5.11905730e-01 -3.75719011e-01
3.07122499e-01 -3.76226217e-01 -6.93232596e-01 -1.72340740e-02
4.79736805e-01 -4.63663578e-01 -6.49457335e-01 1.19140792e+00
6.51872575e-01 1.96231544e-01 4.64379579e-01 -9.95300174e-01
-1.27639973e+00 2.53148198e-01 -7.86172450e-01 4.37285691e-01
1.23144731e-01 4.27654028e-01 7.94575095e-01 -9.50311780e-01
3.11799079e-01 9.93630469e-01 5.17499149e-01 5.63735366e-01
-1.20225859e+00 -7.71003723e-01 -1.21712372e-01 5.66403925e-01
-1.47425163e+00 -3.74096036e-01 8.87734711e-01 -2.35478476e-01
1.04836249e+00 4.96355504e-01 6.66225731e-01 7.29257643e-01
7.35509336e-01 6.20280266e-01 1.20153606e+00 -6.08364999e-01
-7.96265975e-02 1.51462913e-01 -8.15943182e-02 7.78305948e-01
9.36354026e-02 3.41702402e-01 -4.72047985e-01 3.71624380e-01
1.00026667e+00 2.04767540e-01 -5.56403160e-01 1.76728472e-01
-1.02896965e+00 3.50613505e-01 9.21459317e-01 1.65862620e-01
-1.72812298e-01 4.74763423e-01 3.43977101e-02 3.86147887e-01
2.87445277e-01 2.82267898e-01 -2.58451402e-01 1.82793319e-01
-6.61980867e-01 -2.89557904e-01 1.95932448e-01 6.59701824e-01
1.23980546e+00 1.16515830e-01 -1.57753974e-01 8.99444878e-01
-2.31616590e-02 3.86659175e-01 7.59167850e-01 -7.66383529e-01
2.66107202e-01 6.13321662e-01 -5.75353280e-02 -9.41505492e-01
-2.94345170e-01 -2.66699880e-01 -1.18400204e+00 6.84003055e-01
2.96263229e-02 -2.05687523e-01 -1.07306337e+00 1.29646051e+00
1.30095050e-01 2.88371980e-01 -4.08375934e-02 8.62471998e-01
7.98374116e-01 7.86137938e-01 1.34891391e-01 -1.71313569e-01
1.24955964e+00 -1.15137303e+00 -6.12348557e-01 -1.76005080e-01
3.70519400e-01 -8.22082341e-01 1.27195334e+00 3.58336508e-01
-1.17671502e+00 -8.55843306e-01 -1.46491349e+00 -3.52302343e-01
-5.46543479e-01 5.04814506e-01 3.83056343e-01 5.80625474e-01
-1.20372832e+00 6.51868761e-01 -3.74287009e-01 -1.19853280e-02
5.09799719e-01 4.85750258e-01 -3.59274536e-01 1.14468243e-02
-9.29717064e-01 9.07913804e-01 4.55533743e-01 4.18874860e-01
-6.12033725e-01 -7.99712121e-01 -7.03636110e-01 8.17200989e-02
1.90047622e-01 -5.62327504e-01 1.10148942e+00 -1.20686555e+00
-1.87472403e+00 8.03845286e-01 -1.35680540e-02 -3.03690165e-01
4.35523927e-01 -8.75805765e-02 -7.48898506e-01 1.85670897e-01
-5.06981015e-01 7.28738129e-01 1.20928419e+00 -1.16954780e+00
-7.32959807e-01 1.05759859e-01 2.63061132e-02 1.90784127e-01
-5.11983871e-01 -1.40750468e-01 -4.68940139e-01 -5.99889338e-01
-1.39274672e-01 -7.74848700e-01 -9.61031020e-02 2.48693049e-01
-4.97707993e-01 7.46722966e-02 8.58481646e-01 -6.62755728e-01
1.29283166e+00 -2.55086660e+00 -1.68202236e-01 -1.72847912e-01
3.95969093e-01 6.64724827e-01 -2.96330750e-01 2.62752593e-01
-4.44220901e-01 2.04227298e-01 -2.29369074e-01 -1.39160872e-01
-3.07227761e-01 -8.73050615e-02 -1.61925301e-01 6.08269095e-01
2.38038927e-01 1.23184550e+00 -6.42304420e-01 -3.77911657e-01
6.06969357e-01 5.68039179e-01 -3.22393000e-01 -4.47183102e-03
7.21920803e-02 1.66512001e-02 -8.36217701e-02 5.68087339e-01
8.97654951e-01 -1.39965281e-01 -2.47747794e-01 -5.58057249e-01
-3.75508696e-01 -2.22266182e-01 -8.82997453e-01 1.49738574e+00
-5.65076649e-01 9.17800546e-01 -4.34458405e-01 -7.48731911e-01
9.10439134e-01 1.56115377e-02 2.82477252e-02 -1.04949987e+00
5.07463276e-01 1.53208375e-01 -2.17424154e-01 -6.45572662e-01
5.54961979e-01 -1.13500118e-01 1.69164196e-01 2.82620907e-01
-4.40202206e-02 5.57938516e-02 -3.14925686e-02 -1.18419804e-01
9.55617011e-01 -7.17949718e-02 3.33345354e-01 3.06498744e-02
6.53331995e-01 -3.55258286e-01 3.71716082e-01 4.35404897e-01
-2.25401640e-01 6.69317782e-01 1.96121708e-01 -5.93468964e-01
-1.03227389e+00 -1.06526840e+00 4.06458275e-03 8.21332872e-01
5.07048428e-01 -4.52252328e-02 -5.25607586e-01 -3.45960736e-01
-2.37960741e-01 6.14832759e-01 -6.18186653e-01 -5.37841976e-01
-4.71037358e-01 -8.59971285e-01 4.23452526e-01 3.40409040e-01
1.08314764e+00 -1.20701253e+00 -6.32942975e-01 7.87261687e-03
2.24491373e-01 -9.26934540e-01 -6.36065125e-01 2.45830901e-02
-6.14209950e-01 -9.49378133e-01 -7.06284642e-01 -9.36710715e-01
7.98400164e-01 6.42072260e-01 5.70778906e-01 1.94536358e-01
-6.89963341e-01 5.80626987e-02 -2.13080153e-01 -2.11878106e-01
-3.33120435e-01 -9.70229357e-02 -2.91852981e-01 2.21653655e-01
1.70741901e-01 -8.02454710e-01 -9.81922567e-01 1.35280266e-01
-1.17848992e+00 4.36153412e-01 1.11872792e+00 7.12007046e-01
5.65349162e-01 3.08962882e-01 3.13702792e-01 -6.65791690e-01
5.00618219e-01 3.84941176e-02 -5.55084705e-01 3.47469658e-01
-5.63086689e-01 -9.28422287e-02 6.45443499e-01 -6.68504179e-01
-1.23566711e+00 -7.19366595e-02 1.62201703e-01 -4.83112961e-01
6.06144662e-04 2.14447483e-01 -2.51687989e-02 -6.46994114e-01
7.44385719e-01 4.19702560e-01 -1.13716714e-01 -3.49102378e-01
5.91704786e-01 8.51049602e-01 7.75337934e-01 8.54030810e-03
9.14311528e-01 5.01289666e-01 -8.90253112e-02 -8.35255921e-01
-3.44237089e-01 -1.13107152e-01 -4.04494256e-01 -5.02251029e-01
7.18595147e-01 -7.94755042e-01 -9.48546112e-01 8.91351402e-01
-1.03649032e+00 -5.22188127e-01 -2.48050466e-01 3.06300670e-01
-1.65514424e-01 2.93870747e-01 -7.55845845e-01 -6.03301167e-01
-5.05967438e-01 -9.76804852e-01 5.38105547e-01 7.41230011e-01
2.96920359e-01 -7.01135218e-01 -1.99423656e-01 2.01907337e-01
3.93712014e-01 4.26446348e-02 8.96400094e-01 2.17059076e-01
-7.52018452e-01 -2.23905340e-01 -6.50456369e-01 8.28490198e-01
2.23788992e-01 1.25665143e-01 -1.12441170e+00 -2.66635686e-01
-1.82311833e-01 -8.09171349e-02 1.10580790e+00 3.09265435e-01
1.48608553e+00 -3.79918993e-01 -1.48765400e-01 8.45184505e-01
1.81075227e+00 4.37726438e-01 1.18141353e+00 2.84184575e-01
7.68191159e-01 1.61266521e-01 1.23572133e-01 2.62748420e-01
2.14257360e-01 2.73277789e-01 4.05217469e-01 -5.28978586e-01
-6.80308282e-01 -1.47438467e-01 3.47033411e-01 7.32550502e-01
-7.47381970e-02 -2.39074662e-01 -4.84090298e-01 4.43851411e-01
-1.48735583e+00 -7.22501934e-01 -1.53780798e-03 2.13904643e+00
9.19431090e-01 8.02304894e-02 -4.55320925e-01 1.82772353e-01
9.43905473e-01 2.72152007e-01 -6.09371245e-01 -6.24356866e-01
-2.78991640e-01 2.93760628e-01 9.31354523e-01 4.22927946e-01
-8.70798767e-01 9.40000594e-01 5.72567511e+00 1.10477507e+00
-1.47806919e+00 4.08816747e-02 7.93343186e-01 -1.51703119e-01
-2.61528969e-01 -1.08038582e-01 -5.09474695e-01 5.30894458e-01
5.89976668e-01 -4.10703830e-02 6.85188353e-01 3.79703611e-01
2.44172126e-01 -3.41851085e-01 -7.22369492e-01 1.30833733e+00
1.79234385e-01 -1.48911476e+00 9.31903347e-02 -1.30825058e-01
9.37659025e-01 -7.42932633e-02 5.06231546e-01 -2.63447315e-03
-4.28925902e-02 -9.85056400e-01 5.22511244e-01 6.31930649e-01
1.24399543e+00 -5.67751527e-01 4.41664100e-01 -3.15061510e-02
-1.05516291e+00 -2.62287170e-01 -4.29863304e-01 -1.34263635e-01
-1.64087728e-01 6.66873634e-01 -4.62267786e-01 4.92175221e-01
8.22499573e-01 6.59813523e-01 -6.51644588e-01 1.18894053e+00
-4.63108629e-01 3.62944216e-01 -6.40498623e-02 -3.98027264e-02
8.31115767e-02 -4.22619507e-02 3.21060687e-01 1.04609430e+00
2.44149536e-01 2.21825764e-01 -4.56118494e-01 6.92205667e-01
-3.75101537e-01 2.06376649e-02 -1.65335745e-01 7.53473565e-02
2.49024525e-01 1.45386934e+00 -5.44477344e-01 -2.26737842e-01
-3.96533012e-01 1.33686101e+00 2.04428434e-01 7.26966202e-01
-7.48749018e-01 -7.96826124e-01 4.12603259e-01 -7.65210837e-02
3.25555444e-01 -1.15639850e-01 -3.66157502e-01 -9.38387632e-01
1.52616203e-01 -5.17916441e-01 1.06798373e-01 -1.23140121e+00
-1.04629302e+00 5.89407623e-01 -3.26690167e-01 -1.18368733e+00
5.53715110e-01 -7.12799489e-01 -7.87477434e-01 9.29283142e-01
-1.96445167e+00 -1.08625543e+00 -6.48556232e-01 5.84794462e-01
4.35487509e-01 1.71274338e-02 4.30623621e-01 3.69256407e-01
-9.80596781e-01 6.27814233e-01 1.26966257e-02 -9.02198404e-02
7.22124040e-01 -9.33927894e-01 2.77241617e-01 1.05018127e+00
-1.19379781e-01 1.52841821e-01 5.68122447e-01 -3.32745016e-01
-1.23453391e+00 -1.15054011e+00 6.36039555e-01 1.13605589e-01
5.65581560e-01 -1.11179598e-01 -8.78877938e-01 4.84468073e-01
4.49122816e-01 1.79826260e-01 4.60677058e-01 -5.95905304e-01
-2.75783420e-01 -4.93697852e-01 -1.01488185e+00 8.71105731e-01
9.52007711e-01 -5.29596329e-01 -2.83613175e-01 2.85701565e-02
8.81789088e-01 -3.43568921e-01 -5.70742846e-01 1.58535153e-01
5.99444985e-01 -9.00901496e-01 9.21267152e-01 -1.31915852e-01
6.65985763e-01 -3.54553908e-01 1.69911403e-02 -1.29167247e+00
-5.47967434e-01 -8.19959700e-01 1.02661967e-01 1.06481743e+00
3.72623593e-01 -9.75645125e-01 4.98560429e-01 4.70210373e-01
-1.83459491e-01 -7.89035857e-01 -9.62873042e-01 -5.62143862e-01
-1.63756490e-01 -1.76305860e-01 5.54337502e-01 5.99858403e-01
-1.97830766e-01 1.26870945e-01 -5.32771170e-01 2.19153598e-01
3.44300658e-01 -2.25815237e-01 4.52873051e-01 -5.86283207e-01
-1.28379226e-01 -5.87607861e-01 -4.40129608e-01 -1.01515841e+00
-8.30270723e-02 -6.15102947e-01 8.72711930e-03 -1.30085242e+00
1.10279746e-01 -2.57852852e-01 -6.09244585e-01 6.75933540e-01
-3.14953327e-01 8.40544760e-01 1.88342229e-01 2.05348149e-01
-3.28695029e-01 5.98794937e-01 1.59959674e+00 -2.91827142e-01
-3.88787389e-01 -2.53877044e-01 -8.69234383e-01 5.00877321e-01
1.03588843e+00 -1.37273043e-01 -3.91727060e-01 -5.93712389e-01
3.27540219e-01 -3.57068539e-01 5.98282039e-01 -1.21596479e+00
4.63948697e-01 -1.81291655e-01 5.78771710e-01 -2.95448035e-01
4.32170302e-01 -5.69215417e-01 9.22352597e-02 5.07314146e-01
-2.11989060e-01 -1.00510135e-01 4.09313798e-01 5.00981927e-01
-9.13329329e-03 -2.51329899e-01 8.71306360e-01 9.74762216e-02
-1.02911270e+00 2.64800042e-01 -3.22280616e-01 -4.56710339e-01
1.13096738e+00 -4.19663697e-01 -7.35739708e-01 1.91972200e-02
-2.93824852e-01 -2.17724010e-01 2.53946722e-01 2.14604512e-01
9.39879537e-01 -1.21906126e+00 -5.14010966e-01 3.98487777e-01
-8.28181878e-02 -2.88557976e-01 8.20882499e-01 8.30758691e-01
-8.03107500e-01 1.15970053e-01 -3.60658377e-01 -2.12789252e-01
-1.28705263e+00 6.56073153e-01 3.65405262e-01 1.83794871e-01
-7.04806983e-01 1.01545727e+00 2.77147979e-01 2.31717885e-01
-1.25832483e-01 -3.78757179e-01 -1.47549465e-01 -1.95302248e-01
6.47876382e-01 4.88001436e-01 8.54041576e-02 -4.39074099e-01
-6.00837618e-02 6.79333389e-01 -2.54622579e-01 1.25488147e-01
1.17431045e+00 -2.53183663e-01 -1.62512451e-01 3.04595649e-01
1.47647727e+00 -2.57909715e-01 -1.47021842e+00 -2.89929837e-01
-6.43646955e-01 -5.81616044e-01 3.53246510e-01 -8.25894177e-01
-1.18730915e+00 8.49860609e-01 8.45372319e-01 9.66673046e-02
1.72154140e+00 -3.02564472e-01 8.79339337e-01 2.53000915e-01
-1.16460547e-02 -1.15781450e+00 3.02619934e-01 -4.56267819e-02
9.31794763e-01 -9.45388556e-01 5.19094504e-02 -2.97873288e-01
-4.66194063e-01 1.08272135e+00 5.98173738e-01 -3.53170246e-01
5.30228317e-01 7.36746043e-02 1.10529914e-01 -1.77106578e-02
-6.69919968e-01 -4.85753529e-02 2.94453382e-01 4.34147358e-01
-1.11997895e-01 1.10318258e-01 -2.03973770e-01 2.22364143e-01
-1.57825097e-01 1.66624770e-01 7.34763384e-01 7.30400324e-01
-4.56631184e-01 -7.87851453e-01 -3.29905748e-01 5.18475652e-01
-2.39911914e-01 -3.17906499e-01 -1.12176025e-02 6.25299633e-01
3.99433911e-01 8.47793937e-01 7.38970414e-02 -6.95036829e-01
1.99146196e-01 -3.37891608e-01 5.03989160e-01 -2.27360383e-01
-3.96400839e-01 -8.65708813e-02 -3.41895193e-01 -4.78862315e-01
-3.94052178e-01 -1.90833390e-01 -9.74143982e-01 -5.06851435e-01
-3.87495786e-01 -3.82583827e-01 6.96945786e-01 8.06270123e-01
2.40620941e-01 9.29681659e-01 6.71981215e-01 -8.14860582e-01
-1.05245598e-01 -9.26993847e-01 -6.60385489e-01 4.42891866e-02
5.43620229e-01 -3.92244399e-01 -4.38894123e-01 3.19383681e-01] | [11.006936073303223, -2.2379114627838135] |
83fd0d4f-04bb-4bb4-9605-df4ba69ac2bb | constructing-highly-inductive-contexts-for | 2212.01810 | null | https://arxiv.org/abs/2212.01810v1 | https://arxiv.org/pdf/2212.01810v1.pdf | Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation | Large pretrained language models can easily produce toxic or biased content, which is prohibitive for practical use. In order to detect such toxic generations, existing methods rely on templates, real-world data extraction, crowdsourcing workers, or automatic generation to construct adversarial contexts that are likely to induce toxic generations. However, what type of context is more likely to induce unsafe responses is still under-explored. In this paper, we identify that context toxicity and context category (e.g., \textit{profanity}, \textit{insult}, \textit{drugs}, etc.) are two important factors to cause safety issues in response generation. Hence, we propose a method called \emph{reverse generation} to construct adversarial contexts conditioned on a given response, with the flexibility to control category, toxicity level, and inductivity of the generated contexts. Via reverse generation, we augment the existing BAD dataset and construct a new dataset BAD+ which contains more than 120K diverse and highly inductive contexts in 12 categories. We test three popular pretrained dialogue models (Blender, DialoGPT, and Plato2) and find that BAD+ can largely expose their safety problems. Furthermore, we show that BAD+ can greatly enhance the safety of generation and reveal the key factors of safety improvement. Our code and dataset is available at \url{https://github.com/thu-coai/Reverse_Generation}. | ['Minlie Huang', 'Lifeng Shang', 'Yasheng Wang', 'Fei Mi', 'Jiawen Deng', 'Hao Sun', 'Jiale Cheng', 'Zhexin Zhang'] | 2022-12-04 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 2.70821571e-01 -3.67588066e-02 -3.05943012e-01 -1.55725509e-01
-7.84582794e-01 -1.22748268e+00 7.81443596e-01 2.39730150e-01
-2.02252999e-01 1.08928740e+00 5.12149394e-01 -4.08137351e-01
2.04801977e-01 -1.01348758e+00 -6.16867781e-01 -6.77940845e-01
3.17186862e-01 2.66222119e-01 -4.87331441e-03 -6.67865157e-01
1.02660477e-01 1.37722194e-01 -8.31788957e-01 6.01191759e-01
1.19693947e+00 5.28653085e-01 -2.94152051e-01 5.62413394e-01
1.19420597e-02 9.38044667e-01 -8.19638371e-01 -7.42094934e-01
1.63565889e-01 -6.88642621e-01 -8.16596210e-01 -5.33255577e-01
-1.33687913e-01 -5.95856845e-01 -2.76063681e-01 9.28053677e-01
1.03226805e+00 -3.13586160e-03 6.41757786e-01 -1.22081876e+00
-7.92377353e-01 9.26984966e-01 -2.06888050e-01 6.61098436e-02
6.50247991e-01 7.11293757e-01 7.55917609e-01 -4.56356615e-01
5.82868159e-01 1.33335078e+00 4.38873708e-01 1.26939118e+00
-1.13357043e+00 -1.09642076e+00 1.94592357e-01 -1.86714619e-01
-1.14237499e+00 -3.83354902e-01 7.80510426e-01 -5.27520657e-01
5.25410652e-01 6.89318299e-01 2.94470161e-01 2.05280256e+00
1.86773717e-01 5.94811976e-01 1.49616754e+00 3.26500051e-02
4.77376670e-01 2.90870845e-01 -7.38212913e-02 3.66792560e-01
-8.98909271e-02 4.49769735e-01 -4.63095874e-01 -7.05173612e-01
2.32366070e-01 -4.49948153e-03 -3.72664332e-01 5.54777205e-01
-9.18001890e-01 1.14296997e+00 3.47241372e-01 2.46327981e-01
-9.60076377e-02 -8.00009631e-03 4.91226286e-01 4.91058379e-02
5.41017532e-01 7.23209083e-01 -4.27247167e-01 -7.34432787e-02
-7.92089030e-02 6.15326941e-01 8.15038741e-01 8.49797249e-01
4.04902756e-01 -1.47795929e-02 -6.92538738e-01 8.86249483e-01
1.72056202e-02 4.54381943e-01 4.09745604e-01 -6.34725094e-01
5.20702243e-01 6.49465561e-01 1.52538195e-01 -8.97200525e-01
-3.84010792e-01 1.07893934e-02 -7.16915131e-01 -3.47362071e-01
3.51641059e-01 -8.93665135e-01 -5.18215060e-01 2.11159825e+00
5.01599848e-01 -6.78603426e-02 9.24238935e-02 7.31366932e-01
9.88166630e-01 6.19432032e-01 6.35638177e-01 -1.97604597e-01
1.19462299e+00 -4.72311974e-01 -6.47433102e-01 -1.14693701e-01
8.10404420e-01 -5.63267887e-01 1.28352070e+00 4.30701911e-01
-6.98165059e-01 -3.46289761e-02 -7.99204051e-01 2.26529375e-01
-6.63092732e-01 -4.74539638e-01 7.67081559e-01 8.27586293e-01
-5.48327386e-01 5.36852956e-01 -2.03191079e-02 -3.85052450e-02
2.58716464e-01 3.80645335e-01 -1.02070458e-01 4.00571898e-02
-2.13710713e+00 7.55116880e-01 3.59666824e-01 -1.73849016e-01
-1.31432402e+00 -7.47066557e-01 -8.14024687e-01 -5.63834488e-01
5.48151553e-01 -4.51317787e-01 1.10908163e+00 -6.83454573e-01
-1.47567296e+00 4.74693507e-01 4.02182460e-01 -1.67270735e-01
5.60137093e-01 -2.15460941e-01 -4.42534804e-01 -9.40110162e-02
-1.27250880e-01 6.95578456e-01 4.98120338e-01 -1.16943753e+00
-1.46164060e-01 -2.74564564e-01 4.56690192e-01 5.24425618e-02
-5.13965070e-01 4.87647802e-01 3.62098701e-02 -9.03316081e-01
-9.68653977e-01 -1.08797514e+00 -4.17380720e-01 -4.35119212e-01
-8.78032327e-01 -4.59872007e-01 2.90579945e-01 -5.96302092e-01
1.40736318e+00 -1.82055950e+00 -5.11785410e-02 -7.47517869e-02
2.43921652e-01 4.56146955e-01 -2.67738253e-01 7.07704782e-01
-3.85148712e-02 7.05046773e-01 -2.70554423e-01 4.33657616e-01
1.24304414e-01 2.31210459e-02 -6.40602291e-01 1.80905014e-01
3.07168275e-01 8.59580457e-01 -1.22626543e+00 -3.60943824e-01
1.77619010e-02 3.91642243e-01 -7.62643456e-01 3.82755339e-01
-6.78601027e-01 7.57284820e-01 -8.61671686e-01 6.35434985e-01
5.26846647e-01 2.61644989e-01 1.59649491e-01 8.03350210e-02
5.77402376e-02 2.26453111e-01 -6.95949018e-01 1.03870726e+00
-2.36367732e-01 -7.94837102e-02 -3.56276810e-01 -2.38427043e-01
9.12219107e-01 3.99938762e-01 2.84198821e-01 -6.85770273e-01
4.15648997e-01 2.06225857e-01 5.73275087e-04 -6.23630762e-01
2.95739800e-01 -3.86527002e-01 -5.56424499e-01 4.99721706e-01
-2.85699099e-01 -3.10726553e-01 -8.75233859e-03 2.03467011e-01
1.24440408e+00 -1.78719223e-01 1.09251112e-01 -5.86310513e-02
3.86603385e-01 -4.72918265e-02 8.57726038e-01 6.25158846e-01
-4.31336045e-01 4.63029563e-01 9.39615607e-01 -2.53650934e-01
-7.23584235e-01 -7.29740262e-01 2.50549195e-03 1.35884023e+00
2.96251886e-02 -4.69685078e-01 -1.04338837e+00 -9.89277899e-01
-2.31819302e-01 9.77639496e-01 -7.22769916e-01 -4.90743309e-01
-5.35207868e-01 -1.07344604e+00 1.13393521e+00 2.67305434e-01
4.29726243e-01 -1.17486095e+00 3.03972065e-02 2.21747741e-01
-3.31178874e-01 -8.20624173e-01 -8.03413868e-01 -3.83447520e-02
-2.01870769e-01 -1.10221100e+00 -4.39972430e-01 -2.39612937e-01
4.12951350e-01 -2.76795506e-01 8.71289611e-01 1.22572362e-01
-3.34660858e-02 -6.23779930e-02 -4.59323853e-01 -7.42632926e-01
-7.80342758e-01 -1.53856799e-01 1.42277792e-01 -1.68226868e-01
3.10148418e-01 -3.99542838e-01 -7.18267798e-01 5.11479855e-01
-1.28985703e+00 -1.93796262e-01 1.12932950e-01 6.72497988e-01
2.85284132e-01 -2.86353141e-01 1.03802490e+00 -1.35284984e+00
1.30334330e+00 -8.88906121e-01 -2.77606487e-01 2.35036686e-01
-4.22800630e-01 -1.84889108e-01 9.48876083e-01 -8.70474219e-01
-1.01619494e+00 3.58632319e-02 -4.96422976e-01 -1.12336546e-01
-3.20697039e-01 2.69159168e-01 -5.71747065e-01 6.26486167e-02
1.10443211e+00 1.41637325e-01 -4.36007023e-01 -1.65465266e-01
4.77489889e-01 8.58680546e-01 2.04691757e-02 -9.14719105e-01
6.92815244e-01 7.42275175e-03 -5.25719464e-01 -3.75884980e-01
-7.35028267e-01 1.08488627e-01 2.70288140e-02 -1.29850000e-01
9.88716185e-01 -8.75047505e-01 -6.61260247e-01 6.12841904e-01
-1.29592574e+00 -6.63750648e-01 1.51309922e-01 9.48138535e-02
-2.01446995e-01 3.73437107e-01 -8.30904245e-01 -8.03547919e-01
-7.58789897e-01 -1.16081166e+00 8.01232219e-01 1.85496584e-01
-5.68150580e-01 -7.92694151e-01 4.15954918e-01 6.16824925e-01
3.45319331e-01 7.20896482e-01 1.09825170e+00 -1.20850420e+00
-1.53445646e-01 -1.41283989e-01 2.50742435e-01 2.41527840e-01
1.52758539e-01 -1.20855104e-02 -1.00828457e+00 1.10271379e-01
6.84666410e-02 -7.88038790e-01 3.37753236e-01 -1.78575650e-01
1.27140915e+00 -1.04905748e+00 -1.42871276e-01 2.54405886e-01
9.43653643e-01 5.94147444e-01 7.68264592e-01 -4.64598611e-02
6.24152899e-01 8.16998839e-01 5.67098737e-01 6.13382518e-01
2.11307198e-01 6.42485321e-01 2.16797888e-01 1.00067340e-01
3.50692511e-01 -5.75186670e-01 7.45057523e-01 3.51177841e-01
2.31043890e-01 -6.00180566e-01 -8.17550063e-01 4.58282709e-01
-1.48440778e+00 -9.51536596e-01 -1.03805400e-01 1.91806817e+00
1.77220786e+00 1.82879977e-02 2.39398777e-01 -2.00665310e-01
7.67100334e-01 1.06516488e-01 -5.59364498e-01 -7.82559454e-01
-1.22194089e-01 1.83598354e-01 2.29155079e-01 4.93836164e-01
-9.53823268e-01 1.04570854e+00 5.92702389e+00 1.16912937e+00
-1.04001415e+00 1.57993272e-01 1.05344009e+00 -1.07137114e-01
-8.72558177e-01 -1.13469131e-01 -7.64585078e-01 9.61994588e-01
8.56672704e-01 -2.68503219e-01 4.92992520e-01 6.01214409e-01
3.58489692e-01 2.46844560e-01 -1.15962005e+00 3.92291099e-01
-5.20673096e-02 -1.02875388e+00 2.04642385e-01 3.77170667e-02
7.31602728e-01 -3.69463116e-01 8.77509713e-02 6.00206196e-01
8.38215530e-01 -1.27796352e+00 4.59744394e-01 2.09821194e-01
7.40869999e-01 -7.26952732e-01 5.14147401e-01 5.27342379e-01
-5.86615145e-01 -6.92481250e-02 -1.62901372e-01 1.39323890e-01
-8.67906734e-02 5.76534331e-01 -9.60009158e-01 3.36967081e-01
3.64178479e-01 5.34044541e-02 -5.13265789e-01 4.69430923e-01
-5.19789755e-01 7.96746135e-01 -5.77706285e-02 -4.93526638e-01
1.82269588e-01 9.87114199e-03 5.01547813e-01 1.20663357e+00
-3.30588594e-02 6.16659582e-01 2.20784977e-01 9.58712757e-01
-3.08057457e-01 2.20524594e-01 -9.51265275e-01 -1.87784895e-01
6.37362361e-01 1.13462663e+00 -1.53048202e-01 -1.48041427e-01
5.98528534e-02 6.81471050e-01 8.78306851e-02 3.08159888e-01
-1.12318397e+00 -3.31782758e-01 6.19976401e-01 -1.00373350e-01
-4.85758036e-01 2.69077182e-01 -3.51792514e-01 -7.90926635e-01
-2.56895185e-01 -1.57848990e+00 5.31307399e-01 -5.04807174e-01
-1.44381809e+00 6.01526618e-01 5.01472242e-02 -1.04815769e+00
-1.59484982e-01 -3.41085076e-01 -8.60424936e-01 9.01180744e-01
-8.70596290e-01 -1.02403152e+00 -3.79828066e-02 7.20948160e-01
6.67317450e-01 -6.67936578e-02 7.68338144e-01 2.51326710e-01
-9.24015701e-01 8.98087978e-01 -3.52998197e-01 2.31142163e-01
9.33785856e-01 -1.04204559e+00 2.06263810e-01 4.99684960e-01
-5.30046940e-01 9.20545340e-01 7.38265395e-01 -9.56758499e-01
-1.20328248e+00 -1.42176318e+00 8.14774036e-01 -9.50623989e-01
7.58642912e-01 -7.52259016e-01 -7.37235785e-01 3.29693824e-01
3.24293286e-01 -4.63377237e-01 1.00346804e+00 -2.40929961e-01
-4.52995300e-01 8.46159533e-02 -1.40915048e+00 1.18046474e+00
1.04138839e+00 -5.63119650e-01 -2.47112960e-01 8.99420142e-01
1.28426790e+00 -3.52763385e-01 -8.89843464e-01 3.65587294e-01
1.56389505e-01 -7.60979772e-01 7.90821731e-01 -9.84696031e-01
7.83080578e-01 -2.09448598e-02 4.68471795e-02 -1.35681200e+00
-7.24725872e-02 -1.08193016e+00 1.05116561e-01 1.64296520e+00
7.68470705e-01 -7.60393977e-01 3.67434889e-01 1.17308342e+00
-2.74929288e-03 -9.46013689e-01 -6.50037706e-01 -4.10212457e-01
5.77076495e-01 -2.41830960e-01 9.52067971e-01 1.24846494e+00
2.16863975e-01 4.19037223e-01 -7.28370130e-01 -1.36970773e-01
1.91451147e-01 -3.06189179e-01 7.17873096e-01 -5.56719840e-01
-2.66853333e-01 -3.09836328e-01 2.55487859e-01 -4.47819501e-01
2.54506499e-01 -7.62417138e-01 1.65385404e-03 -9.14676309e-01
3.78837228e-01 -5.50944567e-01 -9.94553119e-02 7.99300969e-01
-6.06864989e-01 -4.72655967e-02 4.65961248e-02 -1.59129307e-01
-4.20756102e-01 6.20766878e-01 1.28694558e+00 -2.57172495e-01
-2.90382594e-01 -1.22310236e-01 -1.39177239e+00 5.25152683e-01
1.15778434e+00 -5.24750590e-01 -4.79854345e-01 -1.75019622e-01
3.05528909e-01 3.34531628e-02 2.04396859e-01 -3.59560698e-01
-8.68267044e-02 -7.87305057e-01 2.52154678e-01 -8.84441808e-02
2.52136532e-02 -3.25882196e-01 2.53986746e-01 5.75239599e-01
-7.82391071e-01 7.11962872e-04 2.15736091e-01 4.76865768e-01
2.11742938e-01 -2.83718526e-01 5.35162151e-01 -3.16537559e-01
-1.97781369e-01 4.08903152e-01 -4.15860593e-01 5.15717924e-01
1.04707491e+00 3.65411073e-01 -8.99635077e-01 -3.63024324e-01
-2.65087396e-01 4.47291285e-01 2.66828537e-01 6.06226444e-01
5.10810077e-01 -1.30304921e+00 -8.67603242e-01 -3.15224111e-01
2.00525180e-01 -2.08516613e-01 4.57154632e-01 5.65586090e-01
-2.53116339e-01 1.42713189e-01 1.24415062e-01 -3.12649347e-02
-1.19249117e+00 8.22121918e-01 3.89568210e-01 -3.51565331e-01
8.36815014e-02 7.83600628e-01 6.41685784e-01 -7.34956145e-01
8.09659660e-02 4.56347764e-02 -2.98482180e-01 -4.22880352e-02
6.21240199e-01 2.47700930e-01 -1.58746496e-01 -2.72610337e-01
-3.35966051e-01 1.07862368e-01 -1.85799107e-01 5.53570688e-02
9.12225485e-01 4.32189286e-01 -1.15694098e-01 -6.06747940e-02
9.91393030e-01 2.99856097e-01 -9.56131101e-01 9.17244479e-02
-1.53420657e-01 -2.60104537e-01 -5.89038193e-01 -1.25522411e+00
-8.39651048e-01 6.13924623e-01 4.02601391e-01 8.43566060e-02
9.85432208e-01 -1.16233788e-01 9.43572283e-01 3.05458307e-01
3.37704837e-01 -1.14974332e+00 1.80845067e-01 4.75476146e-01
1.31648183e+00 -1.02235639e+00 -2.28193909e-01 -4.31976467e-01
-1.06533515e+00 4.69216973e-01 1.05951154e+00 2.62443960e-01
2.31961027e-01 1.75639555e-01 1.45753607e-01 -8.95248726e-02
-9.77423489e-01 1.38120160e-01 -6.77903369e-02 7.05129981e-01
4.74551380e-01 2.92179823e-01 -7.56312609e-01 1.12743604e+00
-2.67739832e-01 -2.24146992e-01 4.91772324e-01 7.53267527e-01
-9.81210470e-02 -1.47456706e+00 -4.28148180e-01 3.26894492e-01
-7.50296056e-01 -2.28290007e-01 -1.36427343e+00 3.78602356e-01
4.45685744e-01 1.38149273e+00 -8.10414791e-01 -8.74018610e-01
5.54912865e-01 4.21200022e-02 -1.10867564e-02 -6.38556957e-01
-1.36278462e+00 -7.38710910e-02 5.29924095e-01 -4.22417879e-01
9.36477110e-02 -2.61478156e-01 -1.12477064e+00 -4.92096752e-01
-2.46318564e-01 1.53389484e-01 4.91269045e-02 7.36595392e-01
2.76349902e-01 1.86447695e-01 1.02982843e+00 -1.88912153e-01
-6.82392240e-01 -9.24643397e-01 -1.97514623e-01 6.67882860e-01
-3.03850621e-02 -4.10396278e-01 -3.83198470e-01 -1.80970088e-01] | [6.1325860023498535, 8.158246994018555] |
8e7a0330-46a5-4c6e-9486-4798860e447f | towards-the-extraction-of-robust-sign | 2306.17558 | null | https://arxiv.org/abs/2306.17558v1 | https://arxiv.org/pdf/2306.17558v1.pdf | Towards the extraction of robust sign embeddings for low resource sign language recognition | Isolated Sign Language Recognition (SLR) has mostly been applied on relatively large datasets containing signs executed slowly and clearly by a limited group of signers. In real-world scenarios, however, we are met with challenging visual conditions, coarticulated signing, small datasets, and the need for signer independent models. To tackle this difficult problem, we require a robust feature extractor to process the sign language videos. One could expect human pose estimators to be ideal candidates. However, due to a domain mismatch with their training sets and challenging poses in sign language, they lack robustness on sign language data and image based models often still outperform keypoint based models. Furthermore, whereas the common practice of transfer learning with image based models yields even higher accuracy, keypoint based models are typically trained from scratch on every SLR dataset. These factors limit their usefulness for SLR. From the existing literature, it is also not clear which, if any, pose estimator performs best for SLR. We compare the three most popular pose estimators for SLR: OpenPose, MMPose and MediaPipe. We show that through keypoint normalization, missing keypoint imputation, and learning a pose embedding, we can obtain significantly better results and enable transfer learning. We show that keypoint-based embeddings contain cross-lingual features: they can transfer between sign languages and achieve competitive performance even when fine-tuning only the classifier layer of an SLR model on a target sign language. We furthermore achieve better performance using fine-tuned transferred embeddings than models trained only on the target sign language. The application of these embeddings could prove particularly useful for low resource sign languages in the future. | ['Joni Dambre', 'Anthony Ventresque', 'Ruth Holmes', 'Ellen Rushe', 'Mathieu De Coster'] | 2023-06-30 | null | null | null | null | ['sign-language-recognition', 'imputation', 'transfer-learning', 'imputation', 'imputation'] | ['computer-vision', 'computer-vision', 'miscellaneous', 'miscellaneous', 'time-series'] | [ 2.76269037e-02 -3.51850331e-01 -2.91581362e-01 -1.68376699e-01
-1.08571053e+00 -7.73103893e-01 5.14677286e-01 -8.63981962e-01
-7.99859166e-01 5.06708980e-01 3.51330698e-01 6.25574738e-02
6.90651610e-02 -1.48050085e-01 -8.52721870e-01 -6.53118908e-01
4.25329758e-03 6.48980260e-01 2.83544600e-01 -2.54565924e-01
6.62505999e-02 7.56022215e-01 -1.77519023e+00 6.02744073e-02
6.59516811e-01 6.90182149e-01 1.50917560e-01 9.04590011e-01
1.33858889e-01 6.63121581e-01 -5.03169119e-01 -3.01837534e-01
5.64514518e-01 -3.56842726e-01 -3.46054256e-01 -6.21515885e-02
1.20966518e+00 -5.12807250e-01 -5.96087039e-01 7.46630967e-01
8.26628625e-01 -1.51734546e-01 7.01037526e-01 -1.34620821e+00
-5.04753649e-01 8.55936110e-03 -3.62145662e-01 -4.32561815e-01
3.55919272e-01 5.89922369e-01 1.03942227e+00 -9.08924222e-01
1.18036020e+00 1.05856204e+00 9.05947089e-01 8.61976922e-01
-1.07908142e+00 -7.33945727e-01 1.35325529e-02 2.43075132e-01
-1.23919547e+00 -4.01514322e-01 7.27195561e-01 -3.65890354e-01
7.94427216e-01 7.86457397e-03 7.91578352e-01 1.40480983e+00
-3.28496486e-01 1.05014431e+00 1.44537687e+00 -4.26726460e-01
-1.42083302e-01 3.08904480e-02 -1.97679862e-01 6.17036462e-01
1.80157900e-01 2.21276283e-01 -8.60299468e-01 4.25609462e-02
8.14043760e-01 -2.21695900e-01 -6.19527042e-01 -8.60617995e-01
-1.30931270e+00 5.79747379e-01 3.48355383e-01 2.20436990e-01
-2.59880900e-01 3.93619925e-01 2.75463521e-01 5.24685621e-01
-1.67766690e-01 3.49656254e-01 -6.76931560e-01 -4.86897588e-01
-8.83888543e-01 2.55126327e-01 8.26526880e-01 8.88334811e-01
4.22501534e-01 -6.30091922e-03 1.96279451e-01 6.49465561e-01
3.81688863e-01 9.04011011e-01 5.37644446e-01 -6.86158240e-01
5.79027891e-01 3.13286364e-01 2.30272729e-02 -4.24883306e-01
-2.35418841e-01 -1.47025123e-01 -3.88213396e-01 8.01386476e-01
1.10004067e+00 -1.25680700e-01 -1.43771803e+00 1.70557821e+00
-4.93172091e-03 2.77777258e-02 7.34815150e-02 1.26559794e+00
4.04039472e-01 1.89502463e-01 4.04934287e-02 3.33550900e-01
1.05961263e+00 -7.40383685e-01 -3.67760152e-01 -2.59775221e-01
7.35681355e-01 -9.79404211e-01 1.31212878e+00 6.40626013e-01
-6.74291611e-01 -2.63675839e-01 -8.97380769e-01 -2.36697391e-01
-4.21225756e-01 3.91752332e-01 5.68230331e-01 6.61873579e-01
-8.45996261e-01 3.37125808e-01 -9.69708145e-01 -6.71959937e-01
2.76782602e-01 4.85915273e-01 -8.91669691e-01 -4.06101048e-01
-7.35863030e-01 1.37185609e+00 -1.04224250e-01 3.62009495e-01
-6.19460166e-01 -5.94487190e-01 -8.83111477e-01 -5.27106583e-01
1.94948658e-01 -4.59099352e-01 1.10006487e+00 -1.09027934e+00
-1.64286375e+00 1.09683514e+00 -5.51292077e-02 -2.77636677e-01
1.19300270e+00 -5.21644831e-01 -1.32020414e-01 1.11521915e-01
-3.64700556e-01 7.22678006e-01 1.09408998e+00 -1.32775128e+00
-4.15552676e-01 -3.49772155e-01 -2.07988083e-01 2.25522950e-01
-5.49787134e-02 1.44259185e-01 -4.60586369e-01 -5.18487096e-01
9.14225727e-02 -1.15161026e+00 2.64179241e-03 6.35944426e-01
1.08592473e-01 6.11987449e-02 9.57252264e-01 -1.06484830e+00
5.12502372e-01 -2.06463504e+00 3.52570474e-01 2.67348975e-01
-8.12025517e-02 5.14967442e-01 -5.07690847e-01 2.72748619e-01
6.23546615e-02 -1.60308063e-01 -1.54451489e-01 -8.30757394e-02
1.70145020e-01 4.10409600e-01 -2.05076531e-01 7.50719965e-01
2.00266525e-01 1.09954166e+00 -7.40031481e-01 -4.35331941e-01
3.73919040e-01 9.03115273e-01 -6.12280369e-01 1.42908618e-01
-1.88953176e-01 5.88796973e-01 -4.29234654e-02 8.50705624e-01
5.19300222e-01 1.39664873e-01 1.25783339e-01 -5.08542836e-01
7.58156553e-02 -3.68302539e-02 -1.39765227e+00 1.63825834e+00
-7.69766629e-01 8.69449556e-01 2.72349566e-01 -7.08981991e-01
7.57253051e-01 2.56409705e-01 3.90355110e-01 -5.42352319e-01
6.83449581e-02 8.40168774e-01 3.31757247e-01 -5.74996829e-01
2.06517503e-01 -2.67188698e-01 -6.39727712e-02 2.52089500e-01
3.20386022e-01 -2.89791763e-01 7.01786652e-02 -2.64326602e-01
1.12823796e+00 6.38488710e-01 2.89740264e-02 3.37908983e-01
2.27416530e-01 -3.13221626e-02 2.89872259e-01 4.10826057e-01
-3.50420326e-01 1.03166735e+00 2.17837662e-01 -2.57701248e-01
-1.01277590e+00 -1.29361713e+00 -1.37388304e-01 8.82383823e-01
-1.16176881e-01 -5.70895448e-02 -2.34407380e-01 -9.12068367e-01
2.82891542e-01 3.65872651e-01 -4.36663955e-01 2.40532577e-01
-1.05662823e+00 -3.09527546e-01 7.89071321e-01 7.47139454e-01
3.38348061e-01 -9.56998348e-01 -6.39206231e-01 -4.41899039e-02
1.07117653e-01 -1.13820541e+00 -6.82347238e-01 1.06150195e-01
-5.55074871e-01 -1.15586889e+00 -1.34088576e+00 -9.65925992e-01
7.15463102e-01 -1.95431769e-01 5.90739250e-01 -1.91124961e-01
-4.08368021e-01 7.56108165e-01 -5.05887866e-01 -1.78724170e-01
-4.62964177e-01 -9.74658355e-02 1.43144950e-01 2.52551232e-02
3.75566036e-01 -5.90500832e-01 -4.87402618e-01 4.54732239e-01
-7.00537086e-01 -3.08265835e-01 9.59511578e-01 1.02247286e+00
2.89741635e-01 -7.94536173e-01 6.26199618e-02 -2.34948963e-01
2.31027588e-01 2.70249426e-01 -5.57690322e-01 3.82634580e-01
-2.70964682e-01 3.52701128e-01 4.13779825e-01 -7.87438393e-01
-7.61323810e-01 4.59583342e-01 -1.67175695e-01 -6.25048637e-01
-2.60684967e-01 9.62355435e-02 -2.17276230e-01 -4.92957741e-01
7.47758210e-01 3.77129048e-01 4.28329915e-01 -6.29141569e-01
5.62315524e-01 7.66128123e-01 5.77255428e-01 -4.45568085e-01
1.12700140e+00 5.28325438e-01 -5.60681634e-02 -1.03488255e+00
-3.51554066e-01 -5.48899531e-01 -9.28926826e-01 -3.36879492e-01
7.33681440e-01 -8.32549930e-01 -5.96213341e-01 6.92040324e-01
-1.06379735e+00 -4.74595159e-01 -3.37585002e-01 8.48071635e-01
-7.63439536e-01 4.09209937e-01 -2.40157261e-01 -5.33207119e-01
3.69105004e-02 -1.17890441e+00 1.25576913e+00 -2.88983583e-02
-3.71016622e-01 -6.94793344e-01 1.97809041e-01 4.60665435e-01
3.42022032e-01 2.03014180e-01 4.23576325e-01 -4.73312527e-01
-7.79698253e-01 -5.49956083e-01 -1.91686541e-01 6.76207423e-01
2.01415792e-01 -6.15725145e-02 -9.96555507e-01 -3.33915561e-01
-6.90637052e-01 -7.63481498e-01 1.00350833e+00 1.41657367e-01
5.37299514e-01 -3.18874046e-02 -6.24089949e-02 7.51239955e-01
1.37225533e+00 -2.73408383e-01 5.73788702e-01 2.16952875e-01
7.94917166e-01 5.83145618e-01 4.60514218e-01 -4.20673713e-02
2.84807086e-01 9.45290625e-01 1.13273868e-02 -3.55008841e-02
-7.19376504e-01 -5.94071746e-01 7.99956977e-01 7.93970346e-01
-6.38454318e-01 1.41677782e-01 -9.93966281e-01 6.11934483e-01
-1.62940621e+00 -9.22220230e-01 1.44221559e-01 2.34209776e+00
8.31858695e-01 -2.10619211e-01 1.53078273e-01 1.66929439e-01
1.78069890e-01 4.33605127e-02 -5.43392718e-01 -1.73332781e-01
-4.03078258e-01 3.95916283e-01 9.07055140e-01 4.61952001e-01
-1.02012515e+00 1.04828584e+00 5.76729488e+00 3.76967669e-01
-1.60073674e+00 2.46531535e-02 -3.36433738e-01 -1.98189825e-01
1.09633110e-01 9.87112988e-03 -8.14307928e-01 1.70051187e-01
6.00278258e-01 1.81960449e-01 3.18086982e-01 9.46894765e-01
8.06470066e-02 9.86618847e-02 -1.22167230e+00 1.28779900e+00
4.91412848e-01 -8.16249728e-01 -8.24328884e-02 1.88861579e-01
6.69172347e-01 3.60215813e-01 3.35073732e-02 3.51140529e-01
2.15198100e-01 -1.05716825e+00 5.64534485e-01 4.81236428e-01
9.58841741e-01 -3.36202323e-01 6.17849767e-01 1.19314916e-01
-9.85970914e-01 7.01167211e-02 2.02969816e-02 1.86503038e-01
3.51769000e-01 -1.92614794e-01 -7.96858072e-01 1.82964310e-01
4.08745706e-01 5.91292858e-01 -5.61556160e-01 1.34781325e+00
-5.35006106e-01 5.02784252e-01 -6.74581110e-01 -1.33147731e-01
7.63384029e-02 5.01904376e-02 5.96266448e-01 1.15660989e+00
3.41145635e-01 -5.13431132e-01 2.79436074e-02 2.72158325e-01
1.20034583e-01 2.62386322e-01 -7.77831852e-01 -2.61595082e-02
-7.72866011e-02 9.58344579e-01 -2.91661322e-01 -5.59637621e-02
-5.16866028e-01 1.31953049e+00 1.35874122e-01 3.91081661e-01
-5.93905389e-01 -2.77857363e-01 8.95365238e-01 1.81937829e-01
5.59299409e-01 -6.22644186e-01 -1.24721542e-01 -1.54961562e+00
4.74501789e-01 -1.26274228e+00 1.53038710e-01 -6.19029880e-01
-1.25227976e+00 2.79353350e-01 -1.19522594e-01 -1.57122874e+00
-6.26577735e-01 -1.31301355e+00 -3.06135505e-01 7.36623347e-01
-1.72271967e+00 -1.74107826e+00 -2.05923483e-01 6.10125661e-01
2.49214560e-01 -6.69253021e-02 8.79958212e-01 4.66634274e-01
1.77183345e-01 7.95209825e-01 8.57928023e-02 5.73388755e-01
1.16271222e+00 -1.17389715e+00 9.62910578e-02 7.60835350e-01
6.54675186e-01 4.97301608e-01 6.56284511e-01 -4.85463023e-01
-1.77131569e+00 -7.18817711e-01 8.46196532e-01 -7.80500650e-01
6.67408347e-01 -5.14767349e-01 -6.87850058e-01 7.10677803e-01
-1.70905381e-01 3.64511907e-01 3.92436117e-01 8.41003805e-02
-8.05006862e-01 -7.81354457e-02 -8.91026378e-01 7.66046703e-01
1.29908729e+00 -6.96879566e-01 -7.76730835e-01 3.44834417e-01
5.45974076e-03 -5.88514388e-01 -7.01797247e-01 2.84835279e-01
1.25181592e+00 -5.77887774e-01 8.72997046e-01 -6.63336217e-01
1.23615377e-01 -4.17809546e-01 -2.51311243e-01 -1.26478219e+00
4.10837233e-01 -4.33879972e-01 5.31527847e-02 9.62822497e-01
4.61533308e-01 -6.72814190e-01 1.01642132e+00 5.62404811e-01
2.77544975e-01 -3.09613556e-01 -1.13403058e+00 -1.27283859e+00
1.93537742e-01 -6.30397677e-01 4.04703282e-02 5.96047103e-01
-3.00209969e-01 -1.01186067e-01 -6.20225668e-01 1.31815150e-01
8.57655644e-01 1.62683338e-01 1.31685245e+00 -1.03936183e+00
-4.55348819e-01 -3.77962828e-01 -8.96919429e-01 -1.05858529e+00
2.86404371e-01 -9.36625540e-01 2.23014057e-01 -1.40600765e+00
-2.69049674e-01 -3.31755638e-01 1.64059419e-02 7.27617562e-01
1.90486938e-01 5.07636845e-01 5.00017047e-01 7.38386437e-02
-1.77491903e-01 4.22943115e-01 1.42156196e+00 -1.42089620e-01
-2.45258480e-01 1.73161272e-02 -6.22164309e-02 8.87472332e-01
5.98125577e-01 -2.43815958e-01 -4.11988460e-02 -3.69514525e-01
-3.42350528e-02 -3.64110827e-01 7.71397293e-01 -1.03864491e+00
3.29175532e-01 3.43462229e-02 2.53060609e-01 -3.47130090e-01
6.03004277e-01 -9.16253269e-01 -1.24326356e-01 4.97343421e-01
-1.80253424e-02 -3.40894789e-01 8.27146471e-02 2.21748695e-01
-3.26385975e-01 2.96156015e-02 8.78968000e-01 -7.31454045e-02
-1.09878051e+00 2.97351629e-01 -1.34411484e-01 2.07481772e-01
7.50142217e-01 -4.97458011e-01 9.69365686e-02 -5.55748343e-01
-6.24447227e-01 -6.68134764e-02 5.63467920e-01 7.09680498e-01
6.49672329e-01 -1.32493484e+00 -8.24692249e-01 4.70183015e-01
4.07677770e-01 -2.21439421e-01 2.98075043e-02 8.97215366e-01
-7.37117171e-01 2.56954521e-01 -3.32862377e-01 -7.01245487e-01
-1.68039954e+00 1.61239915e-02 3.79821837e-01 2.25001555e-02
-7.38851309e-01 6.91236198e-01 -2.75371373e-01 -7.81055510e-01
3.08673054e-01 -4.29508358e-01 3.59294593e-01 1.09250799e-01
3.34339291e-01 5.23343943e-02 -9.14100632e-02 -1.07503641e+00
-4.25386608e-01 1.30962646e+00 1.54138908e-01 -3.92210722e-01
1.37310767e+00 4.60786134e-01 4.04829204e-01 2.51787931e-01
1.46679270e+00 3.12829763e-01 -1.46706247e+00 -1.57639980e-01
-1.40134826e-01 -6.15221262e-01 -1.50055319e-01 -8.69644046e-01
-8.31676066e-01 1.05713177e+00 7.53163517e-01 -7.92253613e-01
9.36003268e-01 9.94411781e-02 6.25879943e-01 4.77899432e-01
5.34426630e-01 -1.13980198e+00 -5.22127934e-02 4.65239465e-01
1.26532125e+00 -1.42609274e+00 -2.05006525e-01 4.64806110e-02
-6.21247232e-01 1.16861641e+00 4.33981538e-01 -2.71208316e-01
6.14226997e-01 3.71451318e-01 8.06891143e-01 1.47973925e-01
-3.08914989e-01 -5.93062460e-01 5.36812007e-01 1.02311909e+00
2.95232505e-01 -6.23035012e-03 -2.69318074e-01 1.56197593e-01
-4.52806264e-01 1.51105419e-01 2.75047690e-01 1.00654137e+00
-2.00639423e-02 -1.52828455e+00 -5.11210501e-01 2.45988190e-01
-2.72939876e-02 1.44989967e-01 -4.04657245e-01 1.27559936e+00
2.33373687e-01 3.05244476e-01 -3.21507782e-01 -2.28237420e-01
6.71657443e-01 3.68320376e-01 1.13855386e+00 -3.39073271e-01
-4.36659187e-01 -1.35536849e-01 8.80957320e-02 -6.42558873e-01
-4.15788889e-01 -9.49785531e-01 -1.00641191e+00 1.31590590e-01
-1.88670620e-01 -5.44043362e-01 8.74600470e-01 8.96359742e-01
-2.09194086e-02 -8.95392746e-02 1.06744610e-01 -1.04325199e+00
-1.00367379e+00 -9.66711581e-01 -5.52840889e-01 6.82429790e-01
4.87084329e-01 -7.20962822e-01 -5.40363014e-01 1.12200424e-01] | [9.144371032714844, -6.462841987609863] |
baf2c959-da03-4314-a7b4-148391921ecc | 0-1-phase-transitions-in-sparse-spiked-matrix | 1911.05030 | null | https://arxiv.org/abs/1911.05030v1 | https://arxiv.org/pdf/1911.05030v1.pdf | 0-1 phase transitions in sparse spiked matrix estimation | We consider statistical models of estimation of a rank-one matrix (the spike) corrupted by an additive gaussian noise matrix in the sparse limit. In this limit the underlying hidden vector (that constructs the rank-one matrix) has a number of non-zero components that scales sub-linearly with the total dimension of the vector, and the signal strength tends to infinity at an appropriate speed. We prove explicit low-dimensional variational formulas for the asymptotic mutual information between the spike and the observed noisy matrix in suitable sparse limits. For Bernoulli and Bernoulli-Rademacher distributed vectors, and when the sparsity and signal strength satisfy an appropriate scaling relation, these formulas imply sharp 0-1 phase transitions for the asymptotic minimum mean-square-error. A similar phase transition was analyzed recently in the context of sparse high-dimensional linear regression (compressive sensing). | ['Jean Barbier', 'Nicolas Macris'] | 2019-11-12 | null | null | null | null | ['object-detection-in-aerial-images', 'video-inpainting'] | ['computer-vision', 'computer-vision'] | [ 4.97844607e-01 1.86354920e-01 -4.89857979e-03 2.58611385e-02
-7.94588745e-01 -2.91727632e-01 4.92919028e-01 -3.34208786e-01
-4.45399791e-01 7.70748854e-01 2.15915501e-01 1.56399652e-01
-4.10276383e-01 -2.11528420e-01 -8.04631054e-01 -1.21579957e+00
-7.02618539e-01 4.60389107e-01 -1.97073385e-01 3.35503630e-02
5.57334051e-02 -3.72350216e-04 -1.25858307e+00 -2.57390440e-01
5.51976621e-01 1.14911139e+00 7.29220659e-02 9.65219378e-01
2.89013892e-01 7.57304966e-01 -3.61087680e-01 1.86821539e-02
5.21136105e-01 -7.52759159e-01 -1.55096024e-01 9.83508229e-02
3.35142426e-02 3.12059939e-01 -7.99698412e-01 1.42939401e+00
1.69569477e-01 -2.24526078e-01 9.28399384e-01 -1.08630610e+00
-4.69586939e-01 8.42089534e-01 -8.59393656e-01 3.17625850e-01
1.36370867e-01 -2.52188742e-01 7.43684292e-01 -9.03356671e-01
6.08604670e-01 8.19348335e-01 8.92534673e-01 2.67342687e-01
-1.62304842e+00 -6.47389352e-01 -2.96526879e-01 -2.03096718e-01
-1.66835701e+00 -8.36546004e-01 6.19110942e-01 -7.17157364e-01
3.92966330e-01 1.96191251e-01 5.64521194e-01 1.10846257e+00
4.24963564e-01 3.93811345e-01 8.77035797e-01 -4.40607280e-01
5.81688166e-01 -1.95189148e-01 1.44786507e-01 4.49478507e-01
9.47370827e-01 1.20360509e-01 -7.83760428e-01 -3.54457974e-01
8.82790506e-01 -5.62995933e-02 -3.45342040e-01 -6.25932634e-01
-1.40764320e+00 7.86244571e-01 -1.44014686e-01 7.80183136e-01
-7.51989484e-01 4.45641816e-01 -4.41276766e-02 5.48388779e-01
2.66339362e-01 1.97162077e-01 -6.57645836e-02 -2.48772241e-02
-1.30744243e+00 7.11389109e-02 1.12938309e+00 9.24441874e-01
6.87616408e-01 5.90039432e-01 2.37934887e-01 4.14299011e-01
3.33103061e-01 1.34355605e+00 2.33605325e-01 -1.10914433e+00
5.46687305e-01 -3.17644536e-01 2.95394510e-01 -9.21030283e-01
-4.39576536e-01 -1.09484220e+00 -1.50801194e+00 -1.34558082e-01
5.77411950e-01 -3.66487235e-01 -3.90199870e-01 1.94067466e+00
-2.23496199e-01 4.70276773e-01 3.23824733e-01 9.07715440e-01
1.00942321e-01 5.25019407e-01 -4.99157578e-01 -9.27590132e-01
6.55456543e-01 1.10665321e-01 -9.69161928e-01 -5.21159172e-01
2.25153640e-01 -6.13626420e-01 3.30136269e-01 4.63206768e-01
-1.38237369e+00 1.88306093e-01 -1.20548356e+00 6.36487484e-01
4.42735285e-01 -2.33688444e-01 1.37877688e-01 3.66129011e-01
-8.50120187e-01 6.56648695e-01 -8.08812857e-01 2.25790646e-02
-5.54166101e-02 1.06218807e-01 -4.05883968e-01 -2.06303492e-01
-8.18209827e-01 3.82777989e-01 -2.06364304e-01 3.40935469e-01
-7.71431506e-01 -5.42565286e-01 -6.31869137e-01 -1.55009404e-01
-5.57338707e-02 -5.66366971e-01 8.97324741e-01 -8.24784458e-01
-1.04265821e+00 6.35067463e-01 -3.34949285e-01 -5.46831191e-01
2.61668533e-01 -1.84315339e-01 -4.18935567e-01 2.18944132e-01
4.23594445e-01 -4.00769383e-01 1.16316032e+00 -1.21505594e+00
1.79323405e-01 -6.21362209e-01 -7.77466536e-01 -8.63150731e-02
2.03824062e-02 -4.23160195e-01 3.78198065e-02 -6.46239340e-01
9.27677095e-01 -9.17615473e-01 -5.62275708e-01 -4.03358638e-01
-1.07586220e-01 6.75735176e-01 3.30841810e-01 -4.96014208e-01
1.13051903e+00 -2.44911695e+00 6.83668792e-01 5.60895503e-01
3.09221506e-01 -5.16786635e-01 -1.75865233e-01 5.75944841e-01
-1.51241437e-01 -2.53732800e-01 -5.37643433e-01 -1.56402901e-01
-1.99929744e-01 1.41093671e-01 -2.30226785e-01 1.24632680e+00
-4.85229045e-02 4.34058964e-01 -8.27492833e-01 -1.79711357e-01
-1.38673320e-01 4.00151193e-01 -5.93246400e-01 -1.50825940e-02
2.49652103e-01 4.83655006e-01 -5.48286438e-01 1.81696042e-01
7.30017006e-01 -7.99264073e-01 1.90436497e-01 -4.09316979e-02
9.51680243e-02 -4.14139092e-01 -1.87186503e+00 1.37054253e+00
-3.33676398e-01 7.44892538e-01 9.30000365e-01 -1.17985618e+00
5.28764606e-01 5.34627676e-01 7.32143521e-01 -3.14033002e-01
2.46138945e-01 3.98902059e-01 9.51341391e-02 -3.36796373e-01
1.12738542e-01 -5.31335950e-01 -2.18016624e-01 1.90356925e-01
1.64943151e-02 1.14619784e-01 1.63903072e-01 6.04424477e-01
1.59669280e+00 -6.39016628e-01 4.12937820e-01 -5.28105617e-01
2.52979845e-01 -3.96700203e-01 5.91510832e-01 9.12701607e-01
2.19286546e-01 5.82241833e-01 6.64336562e-01 2.34359920e-01
-1.27359474e+00 -1.21976602e+00 -3.59162062e-01 4.57365602e-01
1.44022092e-01 -2.82579094e-01 -5.91565311e-01 5.68643332e-01
3.56762204e-03 5.80204427e-01 -7.32123017e-01 -1.24208838e-01
-3.48522663e-01 -9.66864586e-01 2.92454690e-01 -6.93995804e-02
3.42569590e-01 -4.37268645e-01 -1.56266317e-01 2.53770769e-01
-4.67250384e-02 -1.39020360e+00 -3.55413616e-01 3.11399579e-01
-9.77945805e-01 -9.75939214e-01 -8.44560146e-01 -3.20094049e-01
7.06407309e-01 1.96394235e-01 7.18403339e-01 -4.98949856e-01
-6.12244941e-02 7.30408430e-01 6.59390390e-02 8.18092525e-02
-3.04523349e-01 -5.09504914e-01 4.74948853e-01 4.00070548e-01
-2.60051303e-02 -7.83381522e-01 -4.48377341e-01 2.24823087e-01
-7.98973858e-01 -3.12477380e-01 4.75796998e-01 7.72939980e-01
5.14870822e-01 -1.34029329e-01 3.85152549e-01 -3.01855475e-01
4.96762037e-01 -8.61545980e-01 -7.36346304e-01 -2.39221945e-01
-2.75202185e-01 4.63449180e-01 4.91333812e-01 -7.01612473e-01
-4.53920543e-01 1.49581045e-01 5.40308297e-01 -4.43723708e-01
5.17682552e-01 4.78429765e-01 1.26911014e-01 -1.57826245e-01
6.97124541e-01 4.57408726e-01 -1.83688905e-02 -4.33374137e-01
3.12217236e-01 5.90214789e-01 6.87562704e-01 -3.38267624e-01
1.04076600e+00 4.69610691e-01 5.26594818e-01 -1.48399055e+00
-7.70864785e-01 -3.24331641e-01 -4.35362488e-01 -5.45729361e-02
5.61510324e-01 -1.09398842e+00 -7.19490528e-01 3.83871108e-01
-8.56242716e-01 -1.81256846e-01 -7.42415607e-01 1.01546264e+00
-7.96608567e-01 2.86103189e-01 -6.74404263e-01 -1.39784098e+00
8.31467137e-02 -5.26386440e-01 7.94125319e-01 -1.24884464e-01
-1.77620873e-01 -6.69409573e-01 3.37803364e-01 -2.57731438e-01
6.35901272e-01 1.57771960e-01 5.52693725e-01 -4.10866320e-01
-5.98259866e-01 -4.88037556e-01 5.99500537e-02 2.60835171e-01
-5.32014191e-01 -3.60494256e-01 -4.70382512e-01 -1.35539323e-01
6.63471282e-01 1.15990087e-01 7.43093789e-01 9.94861841e-01
4.54126388e-01 -4.60240126e-01 -2.94376999e-01 6.22729480e-01
1.48604369e+00 -3.62319291e-01 1.57876432e-01 -1.44739687e-01
3.56215894e-01 1.98029056e-01 3.35543230e-02 8.29590380e-01
-1.80295482e-01 2.14639902e-01 2.10848957e-01 3.86004776e-01
2.64700204e-01 7.63102025e-02 3.91525090e-01 1.33504248e+00
3.19474638e-02 4.19172496e-02 -5.72455406e-01 5.68968832e-01
-1.72485924e+00 -1.27418399e+00 -4.15042967e-01 2.74600363e+00
6.07001543e-01 2.02261791e-01 -9.95336697e-02 2.54368842e-01
7.02108800e-01 2.33481392e-01 -7.15578198e-01 2.84124136e-01
-8.69326413e-01 1.04935758e-01 1.13142729e+00 8.88831913e-01
-5.39814591e-01 2.24555120e-01 7.62212229e+00 3.88535082e-01
-7.09653795e-01 3.41418833e-01 4.06666361e-02 -4.05129284e-01
-5.27396739e-01 -1.24011353e-01 -3.92688721e-01 5.56031585e-01
1.19357800e+00 -5.01030564e-01 6.16925776e-01 4.37097907e-01
1.81409284e-01 -2.49266982e-01 -9.82568085e-01 1.30921173e+00
-7.01469481e-02 -1.36314785e+00 -6.29956543e-01 5.80158293e-01
8.44159424e-01 4.91364449e-01 1.28771111e-01 -2.05894962e-01
1.89382359e-01 -7.55450964e-01 5.31791449e-01 8.99128556e-01
8.90803754e-01 -4.76304531e-01 3.84445101e-01 5.77473998e-01
-1.01584232e+00 -2.57721126e-01 -3.65308046e-01 -1.43896908e-01
4.77442503e-01 1.35350502e+00 9.30504128e-02 -4.79164682e-02
3.40055764e-01 7.68483877e-01 1.23054802e-01 1.10271633e+00
3.43433380e-01 7.51941681e-01 -7.91097403e-01 8.96962658e-02
3.51438075e-02 -7.96543181e-01 1.09011626e+00 8.33617985e-01
7.22356856e-01 4.13792223e-01 -2.16384083e-01 6.51152909e-01
3.61971706e-02 -1.47439033e-01 -7.36280143e-01 -2.85226673e-01
5.16297758e-01 8.68733048e-01 -5.26346862e-01 -1.95739076e-01
-3.27976972e-01 8.23491573e-01 -2.35188589e-01 7.56595254e-01
-2.01232553e-01 -2.09281921e-01 7.08916426e-01 4.74689931e-01
6.57411635e-01 -7.66406596e-01 -3.99618626e-01 -1.41707015e+00
1.50629103e-01 -6.81894779e-01 -1.04197592e-01 -5.73179841e-01
-1.38988900e+00 5.56325018e-01 -1.44822478e-01 -1.15137196e+00
-5.19680977e-01 -2.99650192e-01 -2.71883994e-01 7.14062452e-01
-6.50146902e-01 -3.19401681e-01 3.38863552e-01 5.89188755e-01
-2.85895467e-01 -4.43589270e-01 7.08379805e-01 1.05104543e-01
-4.77318615e-01 1.22169040e-01 1.01240039e+00 -7.84186348e-02
2.08821725e-02 -8.34520102e-01 6.61987141e-02 8.25007260e-01
1.41999066e-01 6.60618067e-01 1.44264042e+00 -7.01747119e-01
-1.85494959e+00 -5.40037811e-01 6.58987105e-01 -1.58345327e-01
1.21205020e+00 -4.03073281e-01 -7.10964620e-01 6.41948998e-01
7.92919472e-02 2.94620633e-01 7.24739254e-01 -2.57181209e-02
-2.38096386e-01 1.51340708e-01 -9.98917162e-01 2.18557253e-01
9.58867788e-01 -5.30313969e-01 -4.02882516e-01 5.55436432e-01
2.75720567e-01 2.03339476e-02 -8.05898905e-01 2.92259216e-01
4.90257800e-01 -7.72232771e-01 7.42250860e-01 -4.36544538e-01
7.10008293e-02 -1.94583520e-01 -8.57317269e-01 -9.11511064e-01
-5.54210305e-01 -1.18046844e+00 -4.23690349e-01 3.92905831e-01
2.45480910e-01 -5.10707915e-01 8.84070039e-01 9.23376828e-02
4.29385841e-01 -4.78305608e-01 -1.27780962e+00 -9.39881802e-01
-1.44378662e-01 -4.64787692e-01 -2.88856208e-01 5.75633526e-01
1.87667713e-01 6.15203321e-01 -6.47557318e-01 4.15394634e-01
1.42547023e+00 -2.78274655e-01 2.80933112e-01 -1.46229887e+00
-5.91250837e-01 -2.17584386e-01 -6.58646226e-01 -1.34619832e+00
-1.40600707e-02 -9.29079652e-01 -5.70925213e-02 -1.07186782e+00
2.87347674e-01 -3.52596730e-01 -1.28092527e-01 -4.80416417e-01
4.73856717e-01 1.34724036e-01 1.23819172e-01 4.47510064e-01
-5.79622209e-01 6.08818114e-01 6.28128052e-01 3.23953599e-01
-1.72533542e-02 3.79386723e-01 -6.21263266e-01 8.53288174e-01
3.84062916e-01 -8.15195382e-01 -7.67571181e-02 -8.68514329e-02
7.51848280e-01 7.20043659e-01 1.51297033e-01 -1.14857662e+00
3.73888761e-01 2.49311030e-02 1.76010013e-01 -5.61069846e-01
6.52404845e-01 -7.38567352e-01 4.69099879e-01 5.11096895e-01
-4.46251482e-01 -1.00651719e-01 -4.24782187e-01 1.11354172e+00
-1.06191069e-01 -4.33222920e-01 8.87327671e-01 1.99106082e-01
2.20841512e-01 1.44359857e-01 -8.20712447e-01 4.84855026e-01
9.17167306e-01 -6.56274334e-02 1.34456396e-01 -9.96738613e-01
-1.03738153e+00 -8.68482068e-02 3.15474391e-01 -2.68170238e-01
5.02057612e-01 -1.53364933e+00 -9.92831051e-01 5.30303836e-01
-2.19954580e-01 -5.04269540e-01 3.64376366e-01 1.03223789e+00
7.31798634e-02 2.22528815e-01 2.08336890e-01 -7.56793201e-01
-7.65343606e-01 3.81246179e-01 4.66092348e-01 -1.81310643e-02
-3.31732810e-01 8.18997562e-01 4.19134870e-02 1.10197514e-01
2.43699387e-01 3.84075977e-02 3.30245733e-01 -2.09332835e-02
6.96799517e-01 5.39519787e-01 -2.82785445e-01 -8.84052873e-01
-2.24055439e-01 8.22331190e-01 3.26068044e-01 -6.66938424e-01
1.37154889e+00 -1.85273066e-01 -3.64455193e-01 1.01773798e+00
1.51242185e+00 3.16477209e-01 -1.12701499e+00 -6.70555532e-01
4.69053239e-02 -2.40471780e-01 3.66983153e-02 -1.51974326e-02
-9.26646411e-01 5.75569987e-01 5.49685478e-01 5.60166895e-01
6.27821863e-01 4.76596922e-01 3.37722480e-01 3.35570484e-01
4.63909417e-01 -6.96949601e-01 -5.51518835e-02 5.50227106e-01
9.67024684e-01 -8.62055957e-01 -4.96787354e-02 -1.63430542e-01
-3.21552068e-01 7.98291385e-01 -3.45438391e-01 -6.66945815e-01
1.11839867e+00 6.76290929e-01 -4.86360162e-01 -1.43019006e-01
-6.55685186e-01 1.16290845e-01 2.03961745e-01 7.08205581e-01
4.25531745e-01 2.09150180e-01 -2.29785904e-01 5.31378865e-01
-2.40610033e-01 -1.46333780e-02 7.96591103e-01 5.08575380e-01
-8.18591297e-01 -6.89631045e-01 -6.03683293e-01 7.72120833e-01
-3.43349993e-01 -2.07342818e-01 8.91120285e-02 2.67970741e-01
-4.76460129e-01 9.35430646e-01 1.15419522e-01 -1.63442418e-01
1.31864011e-01 -1.45039726e-02 7.69992650e-01 -4.33869332e-01
2.72404701e-01 4.29203331e-01 -1.96303010e-01 -6.77045584e-01
-3.48605990e-01 -1.13817239e+00 -9.44458544e-01 -3.34037513e-01
-3.37151326e-02 4.36940789e-01 6.89918578e-01 1.05261838e+00
1.30450949e-01 1.59752488e-01 5.30812263e-01 -7.10910201e-01
-1.04244471e+00 -9.48240221e-01 -1.17447221e+00 -6.60756305e-02
7.76230931e-01 -2.22500563e-01 -1.12759984e+00 -2.49610886e-01] | [6.96699333190918, 4.434881210327148] |
e5e7ea1d-e692-45d2-bedc-3838563bbb6c | anomaly-detection-in-emails-using-machine | 2203.10408 | null | https://arxiv.org/abs/2203.10408v1 | https://arxiv.org/pdf/2203.10408v1.pdf | Anomaly Detection in Emails using Machine Learning and Header Information | Anomalies in emails such as phishing and spam present major security risks such as the loss of privacy, money, and brand reputation to both individuals and organizations. Previous studies on email anomaly detection relied on a single type of anomaly and the analysis of the email body and subject content. A drawback of this approach is that it takes into account the written language of the email content. To overcome this deficit, this study conducted feature extraction and selection on email header datasets and leveraged both multi and one-class anomaly detection approaches. Experimental analysis results obtained demonstrate that email header information only is enough to reliably detect spam and phishing emails. Supervised learning algorithms such as Random Forest, SVM, MLP, KNN, and their stacked ensembles were found to be very successful, achieving high accuracy scores of 97% for phishing and 99% for spam emails. One-class classification with One-Class SVM achieved accuracy scores of 87% and 89% with spam and phishing emails, respectively. Real-world email filtering applications will benefit from the use of only the header information in terms of resources utilization and efficiency. | ['Haruna Isah', 'Craig Beaman'] | 2022-03-19 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 9.08590481e-02 -1.65923640e-01 1.59679249e-01 -3.73173356e-01
-1.81872621e-01 -4.99764800e-01 6.39599383e-01 6.60349905e-01
-2.08646610e-01 5.29107690e-01 -2.59991497e-01 -6.81573510e-01
-9.46454480e-02 -7.73098588e-01 -5.30960113e-02 -5.54480255e-01
1.92234173e-01 3.79939109e-01 2.86693960e-01 -8.15165564e-02
1.08413136e+00 5.74883163e-01 -1.66000235e+00 4.11199570e-01
9.01810467e-01 1.25281191e+00 -6.11136854e-01 9.51014876e-01
-7.90806711e-01 4.82370555e-01 -8.68210495e-01 -7.25658357e-01
3.15066010e-01 -2.50321358e-01 -7.11350918e-01 -2.14945167e-01
4.74709481e-01 -4.25410062e-01 -3.20514113e-01 1.00347674e+00
3.38564903e-01 -3.86381708e-02 6.71607852e-01 -1.37920642e+00
-4.55224097e-01 1.70355201e-01 -5.78895688e-01 4.22620237e-01
5.60322821e-01 1.02374524e-01 9.35278356e-01 -5.68118632e-01
1.76535785e-01 1.11835253e+00 8.13597679e-01 3.64082992e-01
-1.23848581e+00 -7.76562870e-01 -1.17740467e-01 1.08000472e-01
-6.96037412e-01 -2.19498038e-01 5.79258442e-01 -6.15051687e-01
9.75574672e-01 3.83026063e-01 2.88835049e-01 9.81355250e-01
5.27431726e-01 4.08529907e-01 1.30575573e+00 -5.72943211e-01
2.06425115e-01 5.68299294e-01 1.05835569e+00 6.13370657e-01
5.83455443e-01 8.98964033e-02 -3.21419030e-01 -1.04022932e+00
1.25129238e-01 3.16690654e-01 1.67509884e-01 2.34587733e-02
-7.33825684e-01 1.06373668e+00 -1.37619274e-02 3.56328398e-01
-4.47759002e-01 -3.48024070e-01 8.41185331e-01 7.01361477e-01
5.11322737e-01 5.04420638e-01 -5.71216643e-01 -2.81831980e-01
-6.82637393e-01 -2.77338903e-02 1.24688351e+00 6.42225564e-01
7.79561877e-01 -1.75039861e-02 1.67483017e-01 7.58070111e-01
4.06431109e-01 3.85103464e-01 7.09647417e-01 -4.78318930e-01
2.87922442e-01 1.07146716e+00 -5.63397557e-02 -1.31085944e+00
-2.77270854e-01 -1.77133009e-01 -6.07542396e-01 2.36870408e-01
6.76213324e-01 4.62182276e-02 -9.95285690e-01 1.04221797e+00
2.82923460e-01 1.71645638e-02 -1.70955256e-01 2.95816898e-01
5.89075446e-01 2.47034326e-01 2.46331304e-01 -8.40315223e-02
1.23545158e+00 -4.45233017e-01 -7.27577746e-01 -1.47679001e-01
7.56831944e-01 -1.01830709e+00 1.01102507e+00 3.04951489e-01
-6.75852060e-01 -1.81299612e-01 -8.81061077e-01 6.71343207e-01
-7.78352439e-01 -4.84834343e-01 6.82344675e-01 1.38277566e+00
-5.04671335e-01 8.52864981e-01 -4.27391797e-01 -5.13392150e-01
5.27489841e-01 5.48134446e-01 -3.07412118e-01 1.87307060e-01
-9.80704725e-01 9.96559560e-01 1.65185392e-01 -3.02286118e-01
-2.07823329e-02 -5.91738522e-01 -4.03711885e-01 8.58010724e-02
-3.02257249e-04 -1.80877045e-01 1.14765835e+00 -9.03276801e-01
-1.03353083e+00 8.24340701e-01 -2.03744233e-01 -4.71022934e-01
4.22728568e-01 -1.86063319e-01 -6.56844199e-01 1.44296750e-01
1.89034864e-02 -3.78266454e-01 1.01058149e+00 -1.03206909e+00
-7.90136456e-01 -7.75916100e-01 -5.73068976e-01 -4.69275415e-01
-5.60118496e-01 1.60299137e-01 5.35737634e-01 -3.59148264e-01
5.08975565e-01 -7.96764016e-01 2.06192844e-02 -7.31588721e-01
-2.82629848e-01 -1.94672033e-01 1.36000514e+00 -1.20647717e+00
1.45506895e+00 -1.81508243e+00 -7.87689090e-01 7.88890958e-01
2.69317985e-01 5.29141128e-01 3.14279616e-01 5.65583825e-01
-3.89901139e-02 3.29025298e-01 -3.31206471e-02 2.07727805e-01
-1.55969590e-01 9.79353189e-02 -5.02414525e-01 4.77092773e-01
-1.03223316e-01 7.21711278e-01 -5.95695734e-01 -4.26404029e-01
3.30375403e-01 5.62508143e-02 -3.86764973e-01 2.66384661e-01
2.62428194e-01 6.07125610e-02 -4.52938557e-01 1.00743246e+00
6.13913715e-01 -4.72813062e-02 -2.19591539e-02 1.94225639e-01
3.36678714e-01 3.90682042e-01 -1.07234156e+00 5.71096718e-01
-4.26070631e-01 5.28916419e-01 5.81247211e-02 -8.84082973e-01
1.16273069e+00 2.51610994e-01 4.00213420e-01 -4.76912439e-01
2.81552404e-01 5.42120457e-01 1.21544756e-01 -6.21971726e-01
3.63636881e-01 2.88910195e-02 5.16536348e-02 9.64406610e-01
-2.81450637e-02 5.98195434e-01 -2.51727123e-02 4.94461447e-01
1.34778285e+00 -4.34113204e-01 4.69231129e-01 -2.19454214e-01
8.66115212e-01 -2.69605994e-01 1.56309485e-01 1.02830923e+00
-5.70797324e-01 3.03941995e-01 6.57454908e-01 -4.45437789e-01
-1.07389164e+00 -8.76262486e-01 -2.23197609e-01 1.33049572e+00
-2.02965885e-01 -1.05030611e-01 -7.25327075e-01 -1.35445368e+00
5.76759994e-01 8.73035967e-01 -2.32505128e-01 -2.21478835e-01
-7.55601704e-01 -9.87075269e-01 5.39607227e-01 2.15735689e-01
2.33535066e-01 -1.21915400e+00 -1.90503836e-01 2.46302858e-01
-6.47395924e-02 -7.41424620e-01 -8.75190571e-02 1.90761536e-01
-1.12070334e+00 -1.45560431e+00 -9.55041274e-02 -4.71113920e-01
7.35899210e-01 1.49149105e-01 6.25318050e-01 4.20916170e-01
-6.56768858e-01 5.64091861e-01 -3.28342319e-01 -4.31968689e-01
-8.30649316e-01 9.97594669e-02 1.22992083e-01 1.55591682e-01
9.18842852e-01 -6.02021992e-01 -4.15377557e-01 4.42067832e-01
-6.09012783e-01 -1.21082771e+00 6.05241001e-01 1.03445911e+00
-4.74867314e-01 3.39187542e-03 1.21184456e+00 -1.33029544e+00
9.00187790e-01 -6.77812815e-01 -1.46500856e-01 8.84433091e-03
-1.43815303e+00 -2.23993406e-01 4.66135919e-01 -2.63228267e-01
-1.22975707e+00 -3.11348706e-01 -1.57362789e-01 3.34213465e-01
-6.79214001e-01 3.90519947e-02 2.83887498e-02 -4.10645068e-01
8.68917942e-01 3.09614956e-01 5.57529032e-01 -6.41445816e-01
-1.86374396e-01 1.50594532e+00 1.71211794e-01 -2.37964794e-01
7.00039029e-01 2.90644348e-01 -1.30057156e-01 -1.01096535e+00
-5.00937164e-01 -1.11764193e+00 -8.50577116e-01 -1.43201724e-01
1.96718052e-01 -2.02208340e-01 -1.05109811e+00 8.60477448e-01
-1.00300801e+00 6.10203505e-01 2.49989897e-01 1.19746849e-01
2.44982000e-02 1.21711087e+00 -8.11376512e-01 -1.27291763e+00
-6.50104105e-01 -6.15328312e-01 4.24201220e-01 8.03193375e-02
-7.32230723e-01 -1.11259484e+00 -2.70283788e-01 6.65834010e-01
7.29947746e-01 6.16566725e-02 1.23310924e+00 -1.83751571e+00
-1.64790973e-01 -1.11179638e+00 -2.81867236e-01 5.40655792e-01
2.56883174e-01 -3.86628173e-02 -1.08016062e+00 -5.14746130e-01
2.33639762e-01 5.28675728e-02 6.00841165e-01 -1.08115599e-02
9.07613814e-01 -4.90128756e-01 -4.11277384e-01 -5.59860794e-03
1.11261547e+00 4.53406423e-01 6.11074328e-01 7.41361856e-01
4.11791623e-01 9.06849921e-01 5.10234237e-01 4.02201265e-01
-2.29340851e-01 2.22339407e-01 6.11786008e-01 4.90989864e-01
4.19784755e-01 8.37929174e-02 2.85279125e-01 4.03608233e-01
3.29692870e-01 1.50019735e-01 -9.23681319e-01 3.11656624e-01
-1.73893869e+00 -1.04143667e+00 -7.50100255e-01 2.38629460e+00
3.08140457e-01 3.81919384e-01 3.17856431e-01 6.93765342e-01
1.12165821e+00 -1.38363138e-01 -2.12187231e-01 -8.06177020e-01
2.20928863e-01 9.86687317e-02 4.93042022e-01 3.48001838e-01
-1.25704813e+00 5.19342005e-01 6.60453510e+00 4.20417398e-01
-7.93139160e-01 -2.55968958e-01 4.79233891e-01 6.10295758e-02
2.12699607e-01 7.07983300e-02 -1.05322409e+00 9.27735150e-01
1.32505405e+00 1.07776292e-01 1.37757450e-01 1.02567112e+00
-5.50734177e-02 -2.07334384e-01 -6.35752380e-01 6.60367727e-01
1.74957030e-02 -9.56590712e-01 4.39992845e-02 2.61678666e-01
4.76604849e-01 -3.57709676e-01 -2.06395194e-01 2.04630539e-01
7.41958767e-02 -9.63963032e-01 -1.23622812e-01 2.57092685e-01
-4.90673818e-03 -8.40733230e-01 1.34618878e+00 3.06922078e-01
-4.24276888e-01 -3.60347748e-01 -5.46780266e-02 -2.49649972e-01
-2.73086857e-02 7.60400653e-01 -1.17657173e+00 3.44915152e-01
6.88857257e-01 2.52408922e-01 -7.54027426e-01 9.56103623e-01
2.72458702e-01 8.64086330e-01 -3.11147749e-01 -2.58352578e-01
1.67775556e-01 -1.74681962e-01 6.22779191e-01 1.43395233e+00
1.70180053e-02 -2.92996794e-01 6.18754178e-02 2.22741902e-01
4.38267022e-01 3.50014925e-01 -5.95713794e-01 -7.17308372e-02
5.03387392e-01 1.32752407e+00 -7.75492311e-01 -2.22259283e-01
-4.69154239e-01 6.44233942e-01 -4.41226177e-02 1.42122433e-01
-3.47150713e-01 -8.64564598e-01 7.42706478e-01 4.80043530e-01
6.98654503e-02 5.16600646e-02 -1.04015481e+00 -8.25422525e-01
1.37588270e-02 -9.15532410e-01 7.07994282e-01 1.86954439e-01
-1.75179040e+00 2.16334224e-01 -4.66360062e-01 -7.99686491e-01
-2.10337192e-01 -9.36442852e-01 -9.02525783e-01 9.84071136e-01
-1.04101598e+00 -1.00784338e+00 -4.14015859e-01 3.06384027e-01
2.28966892e-01 -8.75781417e-01 8.02916944e-01 1.60467342e-01
-4.51609731e-01 6.89695239e-01 2.20490858e-01 2.43769646e-01
8.00961971e-01 -1.32855546e+00 4.94044662e-01 4.71833736e-01
-3.22987676e-01 9.50615644e-01 6.17489040e-01 -9.20808733e-01
-9.78975952e-01 -6.73255503e-01 1.08944488e+00 -7.68272400e-01
7.56276190e-01 -1.43937869e-02 -1.36486340e+00 5.71845531e-01
-5.55770509e-02 -4.61660832e-01 9.89481151e-01 4.21943545e-01
-4.00617421e-01 1.92718968e-01 -1.63667119e+00 3.75372440e-01
5.17060995e-01 -4.42146331e-01 -6.98200703e-01 1.52483881e-01
-1.36130983e-02 8.32440257e-02 -8.73281240e-01 2.32757971e-01
6.89681351e-01 -1.30049539e+00 9.28611636e-01 -1.18018198e+00
1.31423488e-01 1.73755243e-01 -2.83449013e-02 -8.29503953e-01
-1.86331362e-01 -4.35045063e-01 -5.15290558e-01 1.42986691e+00
2.76032835e-01 -1.26500177e+00 1.00596535e+00 7.73031056e-01
2.73821115e-01 -6.19670689e-01 -7.18196750e-01 -6.97209716e-01
-2.83744801e-02 -2.72906721e-01 2.35670820e-01 9.37158942e-01
2.03373253e-01 3.26166183e-01 -1.92348242e-01 2.40208604e-03
8.21050346e-01 -1.48371747e-02 8.76509190e-01 -1.76594150e+00
1.27977699e-01 -4.76049453e-01 -7.40478635e-01 -3.79647225e-01
2.86937833e-01 -8.29211414e-01 -5.04677534e-01 -1.09897864e+00
5.78833185e-02 -4.36602682e-01 -1.90831542e-01 2.46856585e-01
-3.40625763e-01 2.10615933e-01 -2.01899260e-01 4.58921194e-01
-3.55082192e-02 -5.50585389e-02 4.67868328e-01 7.65010118e-02
-2.02509254e-01 7.35968888e-01 -7.01894999e-01 8.84261012e-01
1.03722644e+00 -4.83030975e-01 1.70003742e-01 4.59189326e-01
-1.89424679e-01 -4.04703051e-01 4.59231853e-01 -6.55069470e-01
2.25783095e-01 -8.84159058e-02 5.13000906e-01 -5.41947842e-01
7.36406371e-02 -7.31707394e-01 -4.46034282e-01 8.45727503e-01
-2.58131027e-01 5.42829931e-02 -7.38636553e-02 9.30365205e-01
-6.39199615e-02 -6.59531295e-01 1.13395917e+00 -2.42324457e-01
-5.09130239e-01 -2.98300236e-01 -6.33757293e-01 -2.04252109e-01
1.32301724e+00 -4.70337391e-01 -5.45997322e-01 -5.51744580e-01
-6.10104859e-01 -1.42174810e-02 2.73536205e-01 5.56496561e-01
5.14544010e-01 -8.95583987e-01 -4.51577693e-01 5.67903936e-01
-2.04226095e-02 -8.11745346e-01 -3.68153839e-03 7.50021577e-01
-6.06379747e-01 5.92119038e-01 -3.46042603e-01 -5.22568941e-01
-1.77473712e+00 4.20794845e-01 1.93056837e-02 -1.21177696e-01
-4.80688810e-01 4.33902144e-01 -4.84345943e-01 -7.45975435e-01
3.38412702e-01 4.78295773e-01 -7.76848048e-02 2.85062909e-01
5.14871836e-01 1.13315058e+00 4.44564462e-01 -4.33546126e-01
-5.17213643e-01 1.89689934e-01 -7.15358257e-01 3.30311537e-01
7.73060024e-01 -1.34003952e-01 -4.64643180e-01 2.42262349e-01
1.29066539e+00 5.93748204e-02 -2.49068022e-01 -2.21267387e-01
7.51058519e-01 -9.36178088e-01 -2.30731040e-01 -8.35668981e-01
-4.66423959e-01 7.35040367e-01 5.42574108e-01 8.71981382e-01
5.82123995e-01 -3.87161642e-01 9.78676856e-01 5.62990010e-01
1.14856020e-01 -1.18485153e+00 2.94265926e-01 5.19430339e-01
2.51220882e-01 -1.58451307e+00 4.07648571e-02 -4.00682479e-01
-5.29245436e-01 1.34636807e+00 8.61095846e-01 -9.18675289e-02
8.67901385e-01 -4.95178737e-02 2.27690294e-01 -1.03253022e-01
-5.21763563e-01 1.77809015e-01 1.44562945e-01 6.60147309e-01
6.28375471e-01 -1.89653128e-01 -6.40583932e-01 3.66590321e-01
-1.38528332e-01 -5.50493896e-01 5.88881016e-01 1.22964501e+00
-9.47522104e-01 -1.06656420e+00 -6.05219245e-01 1.29486215e+00
-1.03854918e+00 6.54164627e-02 -8.01883340e-01 5.81575871e-01
-4.88641739e-01 1.13186693e+00 -2.83475500e-02 -4.68857080e-01
2.69734204e-01 9.20429707e-01 -1.04019068e-01 -5.33754110e-01
-1.14900076e+00 -4.07007247e-01 1.48547128e-01 -2.52391547e-01
3.39334518e-01 -6.89534783e-01 -9.68295932e-01 -9.36484873e-01
-6.37979209e-01 3.15224081e-01 8.60773981e-01 9.95938480e-01
4.42752242e-01 1.72245160e-01 7.91854084e-01 -3.39748770e-01
-1.22258902e+00 -1.18668759e+00 -6.95765972e-01 7.48743176e-01
3.10316473e-01 -4.88361478e-01 -9.14703250e-01 -2.74547040e-01] | [7.826504707336426, 10.008919715881348] |
83095bdd-3913-4986-9bb7-90bbf02e489c | estimation-of-speaker-age-and-height-from | 2203.11774 | null | https://arxiv.org/abs/2203.11774v1 | https://arxiv.org/pdf/2203.11774v1.pdf | Estimation of speaker age and height from speech signal using bi-encoder transformer mixture model | The estimation of speaker characteristics such as age and height is a challenging task, having numerous applications in voice forensic analysis. In this work, we propose a bi-encoder transformer mixture model for speaker age and height estimation. Considering the wide differences in male and female voice characteristics such as differences in formant and fundamental frequencies, we propose the use of two separate transformer encoders for the extraction of specific voice features in the male and female gender, using wav2vec 2.0 as a common-level feature extractor. This architecture reduces the interference effects during backpropagation and improves the generalizability of the model. We perform our experiments on the TIMIT dataset and significantly outperform the current state-of-the-art results on age estimation. Specifically, we achieve root mean squared error (RMSE) of 5.54 years and 6.49 years for male and female age estimation, respectively. Further experiment to evaluate the relative importance of different phonetic types for our task demonstrate that vowel sounds are the most distinguishing for age estimation. | ['Chng Eng Siong', 'Tran The Anh', 'Duc-Tuan Truong', 'Tarun Gupta'] | 2022-03-22 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-2.75580645e-01 1.10350531e-02 3.37039419e-02 -4.85879570e-01
-8.38526845e-01 -2.10039988e-01 5.08611381e-01 1.73684895e-01
-5.12990892e-01 4.65248287e-01 4.54018682e-01 -1.20207593e-01
2.24647522e-01 -5.77904522e-01 -3.49929303e-01 -6.69106960e-01
-1.50585487e-01 2.29068935e-01 -3.64425659e-01 1.76846862e-01
2.21213885e-02 3.03864419e-01 -1.92543495e+00 -2.75152326e-01
9.54848349e-01 1.28547657e+00 -1.79033816e-01 8.35656106e-01
-1.20237032e-02 4.90779012e-01 -8.83922160e-01 -9.80570972e-01
-2.63898104e-01 1.38181290e-02 -5.40229559e-01 -7.98107460e-02
7.82670915e-01 -5.15674531e-01 -4.47387338e-01 8.49145174e-01
1.10852671e+00 1.20137371e-01 8.80431235e-01 -1.12375808e+00
-6.87525451e-01 1.13675296e+00 -4.98550713e-01 1.48559660e-01
2.74524271e-01 -2.38954395e-01 9.00357306e-01 -8.67191672e-01
-2.31226347e-02 1.53758490e+00 8.03772390e-01 7.21656144e-01
-7.21023798e-01 -1.13171554e+00 -2.89135659e-03 4.80989486e-01
-1.72790420e+00 -8.55113864e-01 8.61437023e-01 -5.06676555e-01
7.06581771e-01 2.73566782e-01 4.84229535e-01 1.26926076e+00
2.63613407e-02 6.59220994e-01 9.21779096e-01 -2.59015918e-01
1.19627811e-01 1.94662035e-01 -4.11121212e-02 5.92066228e-01
-3.49652693e-02 2.60955509e-04 -9.07056153e-01 -1.73332244e-01
5.40059865e-01 -4.53202218e-01 -1.31777423e-02 1.97966754e-01
-8.67023528e-01 8.40732694e-01 -7.81233003e-03 1.36094689e-01
-2.45351359e-01 1.94106594e-01 4.02600616e-01 -9.56576169e-02
7.89441943e-01 1.69273898e-01 -2.38526538e-01 -4.76299465e-01
-1.17079139e+00 3.28935653e-01 6.92245781e-01 5.67356765e-01
1.43311784e-01 3.80221605e-01 -1.53668508e-01 1.26352119e+00
4.52855438e-01 7.52316236e-01 5.82400203e-01 -1.03632188e+00
2.75525093e-01 4.22557518e-02 -2.67705351e-01 -6.34081185e-01
-2.41543338e-01 -6.74447954e-01 -7.25853920e-01 -2.07545996e-01
5.43490410e-01 -2.68077582e-01 -8.57298315e-01 1.94433606e+00
3.32835674e-01 3.41321766e-01 -9.69761610e-02 6.06252968e-01
9.79719579e-01 3.30692023e-01 2.19156280e-01 -2.41357118e-01
1.47897542e+00 -7.02298462e-01 -7.97385573e-01 -5.43522537e-01
-1.59564987e-03 -7.86285281e-01 8.39685261e-01 3.47664177e-01
-1.15457869e+00 -6.98535383e-01 -9.00830388e-01 -5.13192341e-02
-1.69830337e-01 3.68686318e-01 6.96290314e-01 1.05917120e+00
-6.80304646e-01 7.61275828e-01 -8.70801449e-01 -2.61644535e-02
2.11657286e-01 4.68549252e-01 -2.54507244e-01 4.56455797e-01
-1.20520496e+00 6.43436491e-01 -1.04512602e-01 9.98534858e-02
-6.59746587e-01 -1.05773497e+00 -1.09073710e+00 2.59408891e-01
-2.36953408e-01 -4.74437296e-01 1.43564165e+00 -3.00003260e-01
-1.71792257e+00 7.74122834e-01 -1.30050272e-01 -4.66224730e-01
3.70298058e-01 -5.47599554e-01 -6.76172197e-01 -7.24412277e-02
-1.21503666e-01 7.26420999e-01 1.01938426e+00 -6.68820441e-01
-6.18516445e-01 -7.02511966e-01 -5.87495148e-01 8.30885619e-02
-9.38069165e-01 2.13884369e-01 -3.94172221e-01 -5.61733484e-01
-4.01928313e-02 -6.80519462e-01 2.00838804e-01 -3.29453558e-01
-2.67589867e-01 -2.36344010e-01 2.83905149e-01 -1.40355158e+00
1.67366290e+00 -2.21045637e+00 1.12172008e-01 -7.52088651e-02
1.77165046e-01 -1.08311497e-01 2.76246935e-01 -5.80937192e-02
1.36551499e-01 -6.72647804e-02 -3.55052366e-03 -7.64290988e-01
2.34854534e-01 -1.65063575e-01 4.15261909e-02 4.19899046e-01
5.99044971e-02 3.75023812e-01 -4.60559398e-01 -6.13622487e-01
1.02428079e-01 9.17249560e-01 -6.39604867e-01 5.07023871e-01
5.89797854e-01 3.05730492e-01 2.63954718e-02 6.42372787e-01
7.43082702e-01 5.75177014e-01 -1.06689639e-01 -2.29488224e-01
-4.14993986e-02 6.37781620e-01 -1.09343481e+00 1.34222794e+00
-9.16092277e-01 7.13413894e-01 2.88104236e-01 -3.78814697e-01
1.13953865e+00 4.12385523e-01 2.89014399e-01 -3.45621973e-01
4.04023707e-01 2.88511485e-01 3.76289517e-01 -2.57136047e-01
5.85868418e-01 -1.55731872e-01 -1.88837141e-01 -8.90993848e-02
4.67926532e-01 -1.52691647e-01 1.09247766e-01 -2.29971617e-01
5.47586322e-01 -3.43100548e-01 8.22222326e-04 -1.54735118e-01
6.58809125e-01 -1.18744683e+00 6.47004783e-01 2.80520707e-01
-4.16522622e-01 6.75246298e-01 2.04181701e-01 1.01584651e-01
-8.33493710e-01 -1.30025744e+00 -3.65110159e-01 1.22998762e+00
-5.61574578e-01 -6.27315581e-01 -9.61528063e-01 -4.02656198e-01
2.22743079e-01 9.01891828e-01 -5.73691249e-01 -2.95051217e-01
-5.03597617e-01 -5.52851737e-01 9.39945698e-01 8.57031941e-01
2.07975179e-01 -6.21734619e-01 -2.42478624e-01 7.46876076e-02
-1.42438173e-01 -1.38072324e+00 -6.23945355e-01 1.30901441e-01
-5.24305403e-01 -5.21282852e-01 -9.33880091e-01 -5.48468292e-01
1.42037362e-01 -6.29562795e-01 9.54650044e-01 -4.86639678e-01
-6.84637204e-02 3.61932397e-01 1.41487392e-02 -7.02868104e-01
-3.56959403e-01 4.38824415e-01 5.80862701e-01 7.21547529e-02
4.48655218e-01 -8.29683721e-01 -5.64156294e-01 -7.32955784e-02
-1.02576867e-01 -3.47673535e-01 2.45262668e-01 6.76650584e-01
7.80469030e-02 -6.78202510e-02 6.30149424e-01 -5.03954887e-01
7.36743391e-01 -2.20660657e-01 -2.21208259e-01 -1.72173828e-01
-7.10278869e-01 -4.30705808e-02 4.06317711e-01 -6.66747928e-01
-1.00947678e+00 -1.41725570e-01 -7.44201601e-01 -4.15860862e-01
-1.62919611e-01 1.38518021e-01 -2.67339230e-01 3.23066920e-01
2.62305677e-01 1.56526253e-01 -1.66449267e-02 -7.17506945e-01
1.46760225e-01 1.22185099e+00 8.77708673e-01 -5.06057322e-01
5.53657115e-01 -1.10674515e-01 -3.29220831e-01 -1.05858517e+00
-5.33944011e-01 -3.38939369e-01 -4.24604952e-01 -2.21452534e-01
8.00890446e-01 -1.23480093e+00 -9.64967728e-01 1.00651002e+00
-9.12628055e-01 6.83914358e-03 -1.16194040e-02 6.74604297e-01
-4.02605444e-01 4.16403830e-01 -9.93878841e-01 -1.24705005e+00
-9.29815710e-01 -1.18378735e+00 1.11438096e+00 3.74261916e-01
-4.16664273e-01 -8.48856509e-01 -2.21255898e-01 5.74587643e-01
3.55857134e-01 -1.64922044e-01 6.01050079e-01 -5.67886591e-01
4.40580755e-01 -8.70464817e-02 1.98610663e-01 5.61618686e-01
1.03061587e-01 1.34696990e-01 -1.58217704e+00 -2.62639284e-01
-2.87109882e-01 -7.10787922e-02 7.63124645e-01 5.51709592e-01
1.40243614e+00 -1.68902576e-01 6.03808388e-02 6.73565447e-01
5.76759517e-01 -2.26564240e-02 4.50331360e-01 -2.15452969e-01
8.68142009e-01 6.77524924e-01 3.30042601e-01 7.25288391e-01
5.99417865e-01 6.40565276e-01 1.65522844e-01 2.46944726e-01
-2.50613421e-01 -3.50964487e-01 6.57125175e-01 1.37134337e+00
-1.09241836e-01 1.58249274e-01 -9.18095946e-01 6.98427975e-01
-1.20135903e+00 -7.45643258e-01 2.03520030e-01 2.44042373e+00
1.13495171e+00 1.94455788e-01 4.00234997e-01 5.07466197e-01
7.87506163e-01 3.13183576e-01 -3.35740983e-01 -7.47788727e-01
3.50793511e-01 4.96891052e-01 5.19400656e-01 5.42303562e-01
-1.13792157e+00 7.98354387e-01 6.51323223e+00 9.26365852e-01
-1.32593262e+00 -6.15672246e-02 7.03808665e-01 -1.78363577e-01
-6.67234436e-02 -6.14062130e-01 -1.07605338e+00 5.87591231e-01
1.40734386e+00 -2.10269064e-01 4.59022254e-01 9.11724865e-01
-1.51360899e-01 2.87653744e-01 -1.13184488e+00 1.30518639e+00
1.59993663e-01 -5.99633276e-01 -3.51332724e-01 1.19575150e-01
1.74154967e-01 -3.61995220e-01 5.63590288e-01 5.84747076e-01
-2.37965375e-01 -1.12859046e+00 9.58778918e-01 2.02630281e-01
1.04181838e+00 -1.19916940e+00 6.74805582e-01 9.68357027e-02
-1.18994427e+00 -2.28642672e-01 -2.91600734e-01 -1.40287384e-01
1.23984758e-02 8.54668915e-01 -9.42488730e-01 1.72576830e-01
7.27222025e-01 2.93985873e-01 -6.62362993e-01 7.72878230e-01
-1.57126933e-01 9.08412695e-01 -3.45105290e-01 1.60071328e-01
-4.69150096e-01 1.66032657e-01 2.91979164e-01 1.18422627e+00
7.65011966e-01 -4.28748012e-01 -3.27349246e-01 5.55872858e-01
-1.79286778e-01 2.44769126e-01 -9.65430960e-02 -2.92222112e-01
9.60553348e-01 1.27844131e+00 -1.13144048e-01 -1.65020213e-01
-2.06444904e-01 8.56510639e-01 3.38264793e-01 -2.69442480e-02
-9.40929055e-01 -4.93633837e-01 1.20930707e+00 1.09411977e-01
4.26205248e-01 -2.09828064e-01 -2.43247464e-01 -9.38076496e-01
-2.47764677e-01 -9.56027746e-01 1.64470077e-01 -3.02232385e-01
-1.19048953e+00 4.99969423e-01 -1.62632614e-01 -7.99829900e-01
-6.52839065e-01 -5.85005403e-01 -8.19408000e-01 9.85520601e-01
-1.22965229e+00 -1.17477036e+00 -1.11256145e-01 1.61066428e-01
6.06420219e-01 -2.60167480e-01 8.94993603e-01 7.59622872e-01
-7.88584411e-01 1.38161314e+00 -1.95578054e-01 2.03119040e-01
9.67390358e-01 -1.43849432e+00 6.95337415e-01 6.66236818e-01
6.87057059e-03 5.92961371e-01 8.51641953e-01 -2.69307047e-01
-1.14278018e+00 -8.36795270e-01 1.31381965e+00 -3.24565679e-01
5.22265911e-01 -6.32673562e-01 -6.85208857e-01 3.66024882e-01
1.77445382e-01 -3.87158215e-01 1.03946149e+00 7.39630461e-01
-5.17008901e-01 -2.46019378e-01 -1.04600847e+00 5.02814531e-01
8.19127917e-01 -7.28166759e-01 -3.35692436e-01 -3.18472028e-01
5.85177004e-01 -4.74673599e-01 -1.41860306e+00 3.96400005e-01
9.84360933e-01 -9.67783213e-01 1.19066703e+00 -1.60618454e-01
4.88006532e-01 3.38463932e-01 -1.29165947e-01 -1.41400564e+00
-3.85069698e-01 -2.73499429e-01 -5.76445937e-01 2.01237798e+00
2.85289973e-01 -3.41202140e-01 6.16025925e-01 5.02561152e-01
-7.12179989e-02 -8.46275866e-01 -9.52523470e-01 -5.49845695e-01
3.73413205e-01 -5.75619757e-01 9.07324195e-01 5.37461758e-01
-2.92752367e-02 5.42616665e-01 -4.72869307e-01 2.12077856e-01
6.30925000e-01 -1.97472900e-01 5.23860335e-01 -1.39667463e+00
-3.30252349e-01 -5.39178550e-01 -5.22337794e-01 -6.36642575e-01
5.02534330e-01 -5.63619435e-01 -7.31851459e-02 -1.01508939e+00
-3.85738574e-02 -1.33318761e-02 -4.57815975e-01 4.23728712e-02
-5.30159950e-01 2.65469164e-01 1.36415482e-01 -4.88285094e-01
8.58600661e-02 7.34703422e-01 6.96541369e-01 -2.57492810e-01
-3.40612829e-02 3.05391788e-01 -6.74487650e-01 9.38527763e-01
8.56172264e-01 -1.02112532e-01 -2.07152843e-01 -2.44590089e-01
-2.24867731e-01 6.40639365e-02 -1.92200258e-01 -1.28260434e+00
-8.76754671e-02 1.63271666e-01 6.57635748e-01 -4.23861414e-01
7.98534214e-01 -3.08984667e-01 -2.90276319e-01 4.00197297e-01
-3.65287602e-01 -1.00939116e-02 2.04068899e-01 5.45431525e-02
-3.27897459e-01 -1.81089535e-01 7.28486657e-01 3.74605834e-01
-3.83484095e-01 3.80628169e-01 -4.05625343e-01 -3.18255462e-02
4.22316700e-01 8.05137753e-02 -6.85163587e-02 -6.96507335e-01
-6.85191035e-01 5.81707880e-02 6.68758303e-02 5.77816546e-01
5.08343756e-01 -1.49022079e+00 -9.16630208e-01 3.29477638e-01
-9.50019248e-03 -5.81686914e-01 4.35021788e-01 7.46132731e-01
-1.00809090e-01 7.51235709e-02 -1.98883355e-01 -3.06619793e-01
-1.72087049e+00 3.83116633e-01 3.06696743e-01 -4.72451821e-02
2.48497296e-02 1.29380906e+00 3.10375765e-02 -5.70898950e-01
6.31048262e-01 -1.62145689e-01 -5.30077755e-01 4.26702261e-01
6.21271253e-01 8.44124675e-01 9.62994695e-02 -8.54038298e-01
-5.43870509e-01 4.21512812e-01 -7.62628615e-02 -2.96856493e-01
1.21124637e+00 -2.45171785e-01 6.03271239e-02 5.77524304e-01
1.21040058e+00 6.18445814e-01 -8.74832153e-01 6.63627014e-02
-2.50714839e-01 -3.95621628e-01 2.78444022e-01 -4.92223114e-01
-1.17570317e+00 1.31698406e+00 1.01703417e+00 -1.71535481e-02
1.05540848e+00 -6.01042733e-02 9.51297820e-01 -2.41159007e-01
-1.68178469e-01 -1.28156793e+00 -3.33742946e-01 4.72592056e-01
6.89642429e-01 -9.94581878e-01 -8.30464959e-02 -5.37716329e-01
-5.92974782e-01 9.12356615e-01 8.05635035e-01 2.96405047e-01
4.72837359e-01 3.14068258e-01 3.04692358e-01 3.68472427e-01
-6.53283775e-01 -3.05572867e-01 5.83568215e-01 6.52901947e-01
1.12909067e+00 3.68111372e-01 -5.06767809e-01 1.05683172e+00
-1.14338064e+00 -4.86651301e-01 2.56449550e-01 1.51126146e-01
-2.32748613e-01 -1.13267660e+00 -4.50557142e-01 5.56376338e-01
-9.61273253e-01 -1.08052552e-01 -2.79970795e-01 2.24986508e-01
2.92936414e-01 1.09335101e+00 4.61387843e-01 -7.70197570e-01
9.03619528e-02 4.86242265e-01 7.70609021e-01 -4.08481121e-01
-7.06116021e-01 9.00525898e-02 3.76852363e-01 -1.69495210e-01
1.28230840e-01 -8.24633479e-01 -1.07675850e+00 -4.51989949e-01
-3.29287410e-01 2.02166215e-01 9.55071092e-01 6.19345129e-01
1.90446034e-01 7.58986831e-01 7.12818921e-01 -7.75968850e-01
-8.34877610e-01 -1.36837780e+00 -8.71029258e-01 3.16390932e-01
3.00299168e-01 -7.09640622e-01 -3.92017394e-01 -1.26094460e-01] | [14.167863845825195, 6.043668270111084] |
0a1e306b-3a8f-41b6-a678-7ae9cb411778 | the-effectiveness-of-morphology-aware | 2103.11189 | null | https://arxiv.org/abs/2103.11189v1 | https://arxiv.org/pdf/2103.11189v1.pdf | The Effectiveness of Morphology-aware Segmentation in Low-Resource Neural Machine Translation | This paper evaluates the performance of several modern subword segmentation methods in a low-resource neural machine translation setting. We compare segmentations produced by applying BPE at the token or sentence level with morphologically-based segmentations from LMVR and MORSEL. We evaluate translation tasks between English and each of Nepali, Sinhala, and Kazakh, and predict that using morphologically-based segmentation methods would lead to better performance in this setting. However, comparing to BPE, we find that no consistent and reliable differences emerge between the segmentation methods. While morphologically-based methods outperform BPE in a few cases, what performs best tends to vary across tasks, and the performance of segmentation methods is often statistically indistinguishable. | ['Constantine Lignos', 'Jonne Sälevä'] | 2021-03-20 | null | https://aclanthology.org/2021.eacl-srw.22 | https://aclanthology.org/2021.eacl-srw.22.pdf | eacl-2021-2 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.08341435e-01 -7.61559159e-02 -1.75162822e-01 -3.20273638e-01
-1.38259554e+00 -9.79786396e-01 6.31153345e-01 1.80133685e-01
-8.82696450e-01 1.09528196e+00 4.13575292e-01 -9.33468103e-01
1.90936759e-01 -6.09977663e-01 -6.07180357e-01 -4.72889125e-01
5.36977053e-01 9.27421629e-01 -1.30252585e-01 -2.88380444e-01
2.37749666e-01 2.72797376e-01 -5.12027621e-01 6.64931089e-02
1.13238990e+00 3.30290794e-01 1.17786437e-01 4.06738639e-01
-3.54475915e-01 9.51285139e-02 -7.39935756e-01 -6.40053272e-01
6.00330174e-01 -8.12485278e-01 -1.21044517e+00 -2.86791533e-01
7.08162367e-01 1.64290085e-01 -1.09486610e-01 1.01097918e+00
3.83963525e-01 3.62350307e-02 8.14442277e-01 -3.06213796e-01
-1.10830963e+00 1.31681287e+00 -3.98961455e-01 6.52627587e-01
8.41554999e-02 1.69621810e-01 1.24818277e+00 -7.78497338e-01
1.03831053e+00 1.32441199e+00 8.10241818e-01 4.79166478e-01
-1.58133245e+00 -3.64810437e-01 -1.68617800e-01 -2.27601469e-01
-1.40476787e+00 -5.52980065e-01 3.99656892e-01 -2.57777303e-01
1.10728693e+00 1.96606383e-01 4.38065320e-01 9.89516497e-01
3.57180059e-01 7.41751671e-01 1.67314279e+00 -4.88476574e-01
-3.90247889e-02 -1.49695083e-01 1.96188778e-01 5.02140045e-01
2.68385053e-01 -9.98774320e-02 -4.36762124e-01 6.00145161e-02
6.96785092e-01 -7.79394925e-01 -1.38567358e-01 6.99533105e-01
-1.60867000e+00 8.78744006e-01 7.51100332e-02 6.61399245e-01
-4.59522486e-01 -8.53976905e-02 5.98237574e-01 6.87116861e-01
8.10828090e-01 7.71934748e-01 -7.45596945e-01 -3.25217724e-01
-1.34337091e+00 1.55107215e-01 6.42001748e-01 7.44075835e-01
5.91761410e-01 4.60848689e-01 -2.73968309e-01 1.12220991e+00
-1.43699229e-01 5.29497921e-01 6.38422132e-01 -7.65815079e-01
9.29219842e-01 1.48799978e-02 -1.21620953e-01 -6.12653911e-01
-3.83801490e-01 -2.26048574e-01 -3.99128735e-01 -2.02366292e-01
9.24916029e-01 -4.53629702e-01 -1.13055515e+00 1.78653693e+00
-1.29295975e-01 -8.45221043e-01 1.23172984e-01 7.80677259e-01
6.95536792e-01 9.11930978e-01 2.18233228e-01 -3.44614923e-01
1.22775078e+00 -9.41600919e-01 -4.98666972e-01 -5.35015583e-01
7.46626794e-01 -9.70057070e-01 1.31514728e+00 1.59045011e-01
-1.38471723e+00 -6.96352422e-01 -9.45452273e-01 -2.74382979e-01
-2.72727847e-01 1.56425342e-01 3.59634101e-01 7.72060871e-01
-1.20075381e+00 7.94620514e-01 -7.41524100e-01 -4.80606705e-01
1.18821569e-01 2.70921737e-01 -2.92836040e-01 1.91611096e-01
-1.07147980e+00 1.35568643e+00 6.67114913e-01 -7.84353465e-02
-1.25214681e-01 -2.84425288e-01 -8.34096789e-01 -2.18356341e-01
-4.65222523e-02 -1.51729211e-01 1.34917939e+00 -1.15613997e+00
-1.40023935e+00 1.19056022e+00 -2.22223744e-01 -4.90143657e-01
4.50726777e-01 -2.16171760e-02 -4.27918211e-02 -7.31965527e-02
1.77841291e-01 1.03412032e+00 3.10915112e-01 -8.88387322e-01
-5.81686080e-01 -3.11630189e-01 -3.11429143e-01 4.45507824e-01
6.15487471e-02 7.92404830e-01 -1.94914594e-01 -8.46701920e-01
2.29475483e-01 -8.39181483e-01 -1.19066134e-01 -7.58115232e-01
-2.69147187e-01 -5.34308910e-01 1.48090020e-01 -1.38673282e+00
1.01431227e+00 -1.93468356e+00 2.53170341e-01 4.05399241e-02
-1.42176077e-01 7.87701383e-02 -3.40321481e-01 1.43082097e-01
3.36410701e-01 6.05306029e-01 -5.26926100e-01 -2.83870429e-01
9.52239241e-03 5.69457352e-01 -3.69477086e-02 4.45917070e-01
5.97962677e-01 1.48553908e+00 -5.63325524e-01 -6.15263760e-01
-1.90942585e-01 4.61301543e-02 -1.32336482e-01 -4.61397052e-01
-5.02562299e-02 5.75561762e-01 -1.92458648e-02 8.44255388e-01
3.86011869e-01 4.13721323e-01 2.45828927e-01 6.35651469e-01
-3.62944096e-01 1.04213929e+00 -4.60795134e-01 1.55988526e+00
-4.39990759e-01 8.69684994e-01 6.99564144e-02 -1.03253484e+00
8.87725294e-01 2.79702127e-01 -1.54794112e-01 -1.00672317e+00
1.93193153e-01 6.57245278e-01 7.06563175e-01 -2.04349682e-02
8.93287599e-01 -3.96089762e-01 -3.70533645e-01 5.16820967e-01
1.69562638e-01 -4.94460016e-01 3.57182145e-01 -4.15672868e-01
8.60826433e-01 2.91685879e-01 3.00597489e-01 -6.23153985e-01
4.24828120e-02 6.17485821e-01 8.58835459e-01 8.46939743e-01
-1.74680203e-01 8.75597000e-01 3.03057581e-01 -1.70135468e-01
-1.20174348e+00 -1.17064333e+00 -3.46815377e-01 1.31975079e+00
-2.74775177e-01 -7.16914609e-02 -1.08666563e+00 -7.26948619e-01
-4.08110350e-01 8.72843683e-01 -3.98579866e-01 3.55561763e-01
-1.13061690e+00 -1.01480281e+00 9.27174985e-01 5.59137225e-01
1.96732834e-01 -1.46827137e+00 -3.04784864e-01 5.06728411e-01
-3.16730171e-01 -1.22255659e+00 -6.13011420e-01 2.92979062e-01
-1.02320492e+00 -2.77203172e-01 -9.27875817e-01 -1.10314488e+00
1.79348826e-01 -1.27998576e-01 1.33121943e+00 -5.88023402e-02
1.80451050e-01 -2.01162055e-01 -3.83542299e-01 -5.04073799e-01
-8.66055965e-01 5.67736089e-01 -5.74099459e-02 -6.60583794e-01
5.08618832e-01 -1.30008072e-01 1.03963781e-02 -3.95431146e-02
-4.90155935e-01 -3.46305072e-02 6.73500538e-01 6.06665969e-01
7.26153553e-01 -5.48811197e-01 5.97006679e-01 -1.03514481e+00
8.22566330e-01 -3.99198174e-01 -3.57194364e-01 1.69988364e-01
-4.89426613e-01 -1.23022854e-01 7.88412809e-01 -4.57394451e-01
-8.45618963e-01 -1.50668129e-01 -3.09250087e-01 2.75576442e-01
-3.08743060e-01 7.36470401e-01 5.88631593e-02 1.54374003e-01
7.25846946e-01 2.12312654e-01 -3.06530118e-01 -6.09340847e-01
4.03016925e-01 7.84573555e-01 6.27015948e-01 -9.06402588e-01
6.34292483e-01 -1.68515086e-01 -5.37823617e-01 -8.41861606e-01
-4.25004750e-01 -1.36420969e-03 -9.10658658e-01 2.16260999e-01
1.11103380e+00 -6.14868999e-01 2.33810931e-01 3.11624408e-01
-1.42597651e+00 -6.45064473e-01 -4.55354035e-01 5.94340384e-01
-6.91897333e-01 1.30051225e-01 -1.12124026e+00 -2.85795063e-01
-5.38394809e-01 -1.39059317e+00 8.30503702e-01 1.07113160e-01
-6.08338058e-01 -1.05680943e+00 1.76170930e-01 4.22756463e-01
3.59574139e-01 2.14297809e-02 1.33485723e+00 -1.07835078e+00
-4.42170054e-02 3.40743773e-02 -3.02669555e-01 4.26213324e-01
1.76826477e-01 -1.09089985e-01 -4.19784486e-01 -3.39486711e-02
-1.26092181e-01 -3.32799196e-01 7.77202606e-01 4.59581524e-01
3.78823280e-01 -3.14187229e-01 1.73040450e-01 4.21589494e-01
1.29908013e+00 3.57432127e-01 3.67309839e-01 2.62664258e-01
5.87968469e-01 7.35759974e-01 3.50202382e-01 -4.60165471e-01
4.77046669e-01 4.87839490e-01 -5.51427066e-01 -1.94421887e-01
-4.08521235e-01 2.77837124e-02 6.78687215e-01 1.37690580e+00
-1.90703437e-01 -8.66722986e-02 -1.30013847e+00 8.25174630e-01
-1.64111638e+00 -5.66034019e-01 -2.54454106e-01 1.83080554e+00
1.36920738e+00 2.41681382e-01 2.61389226e-01 -3.20820153e-01
7.10896015e-01 1.13624468e-01 -7.13095814e-02 -1.21299314e+00
-5.22577941e-01 9.96159017e-01 6.29657924e-01 5.44259965e-01
-9.37068939e-01 1.73154831e+00 7.81380129e+00 9.26386356e-01
-1.04082847e+00 4.66397941e-01 7.65970170e-01 1.46527156e-01
-2.28189513e-01 -7.45428130e-02 -6.65443242e-01 3.24179232e-01
1.41647387e+00 -5.93031794e-02 6.63610816e-01 2.90239185e-01
2.90176421e-01 -9.59961042e-02 -1.14289808e+00 6.43799961e-01
-2.32042074e-02 -1.17824650e+00 1.91970821e-02 6.75869407e-03
9.43089545e-01 5.85100472e-01 -1.05467238e-01 3.55409145e-01
4.89434332e-01 -1.11962843e+00 1.01428831e+00 3.61733995e-02
6.98975980e-01 -6.56997442e-01 6.92516267e-01 2.78943986e-01
-6.63544416e-01 4.46555614e-01 -5.30225456e-01 -2.93758899e-01
8.76998976e-02 8.01648349e-02 -6.27978921e-01 4.23951924e-01
3.29051435e-01 1.14713751e-01 -7.64846087e-01 8.27539086e-01
-4.69137162e-01 1.39635277e+00 -2.69974321e-01 -1.17098421e-01
7.55087197e-01 -7.58776844e-01 4.45169747e-01 1.78570509e+00
7.05945641e-02 -8.60352516e-02 1.84831411e-01 8.32294464e-01
-3.02022636e-01 7.99292922e-01 -5.46359599e-01 -3.28580767e-01
2.83266842e-01 9.71382320e-01 -1.26906300e+00 -4.38332826e-01
-6.77239954e-01 1.23384690e+00 4.53970343e-01 2.96031386e-01
-5.46630323e-01 -3.75301659e-01 5.36724985e-01 -1.06156170e-01
2.40462124e-01 -7.05868959e-01 -8.26515496e-01 -8.87320340e-01
1.07918963e-01 -1.07202649e+00 8.71861577e-02 -3.79588842e-01
-1.27684021e+00 6.75261319e-01 -4.21125032e-02 -3.93473238e-01
-3.47511679e-01 -7.84675181e-01 -6.19050026e-01 1.28406608e+00
-1.11615729e+00 -9.76483703e-01 7.24655509e-01 7.01634884e-02
8.68707597e-01 -1.69331536e-01 7.60118723e-01 -2.69181877e-02
-4.75446016e-01 7.16867626e-01 3.92846167e-01 7.05225468e-01
5.69065571e-01 -1.39115143e+00 1.18650258e+00 1.07702827e+00
8.12740147e-01 4.55309808e-01 4.19046819e-01 -6.96856618e-01
-8.17375779e-01 -9.17040408e-01 1.25213289e+00 -4.97605413e-01
5.87854147e-01 -3.00362408e-01 -8.76719177e-01 8.05416584e-01
6.35881066e-01 -5.51301837e-01 5.26282012e-01 3.64307731e-01
-1.99426532e-01 5.39170742e-01 -1.03944659e+00 6.89385951e-01
1.06861722e+00 -6.40016794e-01 -1.17558634e+00 4.10036922e-01
6.83996260e-01 -2.21048295e-01 -8.65345180e-01 6.78102672e-02
3.55311811e-01 -4.46363151e-01 5.18122733e-01 -9.97485340e-01
5.37070096e-01 1.23914286e-01 -2.73978829e-01 -1.70805228e+00
-4.88908112e-01 -4.48994249e-01 7.43213058e-01 1.33836162e+00
1.09472072e+00 -6.81848764e-01 5.60982764e-01 2.28749394e-01
-3.65061760e-01 -6.25214338e-01 -1.10854495e+00 -1.03485560e+00
1.13929343e+00 -3.85448784e-01 2.00351596e-01 1.04394197e+00
-1.78101912e-01 6.16958618e-01 2.82795191e-01 -3.57750714e-01
1.95366696e-01 1.31845906e-01 2.06134409e-01 -1.05909503e+00
-4.09532487e-01 -9.37605441e-01 -2.46159121e-01 -6.84612751e-01
5.12029052e-01 -1.36262381e+00 3.81309032e-01 -1.87333584e+00
5.07258549e-02 -3.60158503e-01 -1.60505578e-01 8.07939410e-01
-2.45151654e-01 8.16178083e-01 4.51842010e-01 2.39315271e-01
-8.61668363e-02 -2.60037724e-02 1.13686192e+00 -1.46698177e-01
-5.67287564e-01 -1.28621206e-01 -8.25299799e-01 6.99697852e-01
1.18172634e+00 -5.70181668e-01 1.33111998e-01 -9.58020210e-01
1.32531645e-02 1.69487391e-02 -3.16559792e-01 -6.39759004e-01
-2.05662295e-01 -3.01305324e-01 4.45870489e-01 -2.31742546e-01
-1.13020271e-01 9.67652425e-02 -2.70602196e-01 2.42535397e-01
-2.28369713e-01 5.08120120e-01 3.30641478e-01 -2.34762549e-01
-1.27716795e-01 -4.77722526e-01 7.53517091e-01 -5.06404042e-01
-3.43921721e-01 -4.39706892e-02 -9.26580787e-01 6.39440894e-01
3.48634422e-01 -5.29943824e-01 -5.78833148e-02 4.65452448e-02
-5.33312440e-01 -1.41586304e-01 5.07910550e-01 4.56338078e-01
1.71239868e-01 -1.10596967e+00 -1.24553704e+00 -1.33904263e-01
-1.84565261e-01 -1.92057908e-01 -5.60779989e-01 9.18429673e-01
-7.03481972e-01 4.21572894e-01 -2.75782079e-01 -1.53189406e-01
-9.57503140e-01 3.32980528e-02 5.42504676e-02 -2.89354026e-01
-4.05233026e-01 8.87360752e-01 -2.18689173e-01 -8.31185937e-01
-1.71378255e-01 -3.40069771e-01 1.29195929e-01 2.56758541e-01
5.37835294e-03 2.08810508e-01 1.90301016e-01 -1.06046331e+00
-2.31668770e-01 4.27315086e-01 -2.85042554e-01 -6.38213813e-01
1.11508310e+00 -2.48393118e-02 -2.68471718e-01 6.89121842e-01
7.88458169e-01 3.29111308e-01 -6.74544692e-01 -2.31566146e-01
3.54242176e-01 2.83515081e-02 -4.53093275e-03 -9.05339718e-01
-6.36543512e-01 8.41563404e-01 7.68490881e-02 -7.33690113e-02
8.03057253e-01 5.36693558e-02 1.08348370e+00 3.23496580e-01
2.91681290e-01 -1.57251620e+00 -8.20349991e-01 9.98697400e-01
5.07472575e-01 -1.08357084e+00 -3.22249889e-01 -3.97642069e-02
-7.70949781e-01 9.08367932e-01 2.93609470e-01 -2.51031369e-01
1.76181898e-01 1.58912659e-01 2.55263895e-01 1.77121684e-01
-2.88705379e-01 -3.33447963e-01 2.51489580e-01 5.01252592e-01
7.27900922e-01 3.55590284e-01 -1.22137868e+00 5.24332583e-01
-8.51627231e-01 -6.75785124e-01 4.76501465e-01 7.36125886e-01
-3.07936609e-01 -1.24891126e+00 -3.98766071e-01 6.14019692e-01
-9.77169693e-01 -6.71446681e-01 -8.91237378e-01 1.02693450e+00
2.10406423e-01 7.84992933e-01 4.46134895e-01 -1.88409418e-01
-1.17219891e-03 6.92822456e-01 7.32402802e-01 -9.70839620e-01
-1.07509148e+00 2.03450456e-01 3.90130520e-01 -1.06792592e-01
-2.76409596e-01 -9.96785104e-01 -1.34874415e+00 -2.62915552e-01
-2.36953974e-01 1.43034026e-01 6.42623186e-01 1.22857785e+00
-1.45991772e-01 -8.80304072e-03 1.76122054e-01 -7.22548485e-01
-5.54386258e-01 -1.15678632e+00 -3.61211151e-01 2.47831672e-01
-9.45354067e-03 -4.45065610e-02 1.04328003e-02 -1.03392720e-01] | [11.3802490234375, 10.30298900604248] |
b15a4d17-3f63-40d1-a36b-b8acbd5d4623 | towards-understanding-ecg-rhythm | null | null | http://proceedings.mlr.press/v85/goodfellow18a.html | http://proceedings.mlr.press/v85/goodfellow18a/goodfellow18a.pdf | Towards understanding ECG rhythm classification using convolutional neural networks and attention mappings | Access to electronic health record (EHR) data has motivated computational advances in medical research. However, various concerns, particularly over privacy, can limit access to and collaborative use of EHR data. Sharing synthetic EHR data could mitigate risk. In this paper, we propose a new approach, medical Generative Adversarial Network (medGAN), to generate realistic synthetic patient records. Based on input real patient records, medGAN can generate high-dimensional discrete variables (e.g., binary and count features) via a combination of an autoencoder and generative adversarial networks. We also propose minibatch averaging to efficiently avoid mode collapse, and increase the learning efficiency with batch normalization and shortcut connections. To demonstrate feasibility, we showed that medGAN generates synthetic patient records that achieve comparable performance to real data on many experiments including distribution statistics, predictive modeling tasks and a medical expert review. We also empirically observe a limited privacy risk in both identity and attribute disclosure using medGAN. | ['Peter C. Laussen', 'Robert Greer', 'Sebastian D. Goodfellow', 'Mjaye Mazwi', 'Danny Eytan', 'Andrew Goodwin'] | 2018-08-17 | null | null | null | proceedings-of-the-3rd-machine-learning-for | ['arrhythmia-detection', 'ecg-classification', 'electrocardiography-ecg'] | ['medical', 'medical', 'methodology'] | [ 1.60482481e-01 6.25702918e-01 2.46817112e-01 -6.60849571e-01
-1.08920252e+00 -5.40082037e-01 3.41578960e-01 2.43824735e-01
-3.53339404e-01 1.18960214e+00 3.79001081e-01 -4.45666999e-01
1.02335975e-01 -1.04675889e+00 -7.00089455e-01 -4.78820920e-01
-4.92877848e-02 4.86258864e-01 -9.55537975e-01 1.50705829e-01
-4.20659304e-01 2.79560447e-01 -7.65202940e-01 5.03270447e-01
8.23355019e-01 6.97424293e-01 -7.19432831e-01 6.85514629e-01
1.85625672e-01 8.69748950e-01 -9.12628412e-01 -9.23789442e-01
6.06798410e-01 -5.55447280e-01 -4.21371847e-01 -2.29690373e-01
5.39944321e-02 -6.17375970e-01 -5.40798128e-01 1.05082273e+00
1.00979531e+00 -1.59417436e-01 5.82458079e-01 -1.42996955e+00
-9.83857930e-01 1.10340917e+00 -2.24542782e-01 -4.47611094e-01
3.32644999e-01 4.37780142e-01 6.03761256e-01 -4.30536747e-01
6.76153898e-01 1.05509305e+00 1.10571480e+00 8.76155138e-01
-1.51611483e+00 -9.41670120e-01 -4.84971881e-01 -5.59892952e-01
-1.58518469e+00 -3.58350545e-01 3.64713073e-01 -3.15117836e-01
3.90751362e-01 5.98329127e-01 6.04114056e-01 1.42396867e+00
4.08354938e-01 6.74801469e-01 1.04613495e+00 -4.15220251e-03
4.43995893e-01 3.89841080e-01 -1.79847553e-01 4.69429225e-01
5.09882390e-01 9.05216411e-02 6.58235624e-02 -1.08848310e+00
7.78024852e-01 4.59456384e-01 -2.02345431e-01 -8.47892538e-02
-1.17742109e+00 9.38824594e-01 2.00430408e-01 -2.13679597e-01
-6.11590862e-01 1.38746008e-01 4.13163811e-01 2.62179554e-01
2.62488037e-01 5.39877236e-01 -2.95733809e-01 1.93300888e-01
-6.41623259e-01 4.69853669e-01 1.16415644e+00 1.21579874e+00
3.97798032e-01 9.57721025e-02 -5.82224190e-01 4.93486911e-01
3.50694694e-02 7.20921934e-01 5.60265779e-01 -8.83705556e-01
4.05909061e-01 2.44911253e-01 1.55947939e-01 -8.38521183e-01
-1.91969678e-01 -7.48340040e-02 -1.52634203e+00 -1.12256721e-01
3.73609960e-01 -8.03341269e-01 -1.03962374e+00 1.89425302e+00
2.54711062e-01 9.65353176e-02 5.09773672e-01 4.36736077e-01
8.48040462e-01 4.22415078e-01 4.30380553e-01 -2.88350917e-02
1.35357404e+00 -1.37513921e-01 -8.84149492e-01 4.47330803e-01
6.63703263e-01 -4.57043231e-01 7.98459172e-01 7.99111500e-02
-1.20669842e+00 -1.35835662e-01 -7.99790740e-01 2.89967190e-02
-3.02286237e-01 -2.50007480e-01 8.23633254e-01 1.00011873e+00
-8.20490122e-01 6.05576634e-01 -8.57301772e-01 -7.14068413e-02
9.71441567e-01 4.63969380e-01 -4.35145766e-01 -6.96420074e-02
-1.47237563e+00 2.30742037e-01 2.68114328e-01 -3.29203606e-01
-6.96989775e-01 -1.14302528e+00 -9.95029211e-01 3.15043390e-01
4.56331708e-02 -1.37741923e+00 1.11301911e+00 -7.49574184e-01
-1.23779023e+00 7.01406598e-01 4.72828954e-01 -9.34366107e-01
8.89381289e-01 -3.72676402e-02 -4.99574333e-01 -2.29792222e-01
-2.21985176e-01 5.95612824e-01 5.31392336e-01 -1.12918377e+00
-6.99105486e-02 -4.47054863e-01 -3.74867827e-01 -1.39769182e-01
-2.67973572e-01 -2.97742903e-01 6.86798170e-02 -1.06673062e+00
-5.02107739e-01 -9.50797975e-01 -6.83568537e-01 -1.77605614e-01
-9.82185543e-01 3.33919674e-01 2.56616116e-01 -8.71140718e-01
1.10500860e+00 -2.20311379e+00 -4.06776637e-01 4.64426845e-01
4.30668175e-01 2.27032974e-01 1.95058100e-02 3.39819640e-01
-7.42804855e-02 5.09084523e-01 -5.90405643e-01 -4.06727016e-01
-6.37079552e-02 1.39239147e-01 -3.73434216e-01 3.19914937e-01
2.33692657e-02 1.17351210e+00 -7.03218222e-01 -3.77891332e-01
-8.87786746e-02 7.39303827e-01 -8.69870484e-01 4.30611163e-01
-1.65619448e-01 4.29053843e-01 -4.63048220e-01 6.35518670e-01
8.30937386e-01 -4.22487229e-01 3.17424297e-01 3.94836999e-02
6.93614781e-01 -2.33588368e-02 -9.31148291e-01 1.40717041e+00
-1.66662112e-01 9.72766504e-02 -1.68180078e-01 -3.86289150e-01
7.93583930e-01 4.50694084e-01 7.44639874e-01 -1.68656528e-01
2.45319977e-01 -2.35361859e-01 -8.45403820e-02 -2.57849127e-01
3.42702538e-01 -2.32147157e-01 -4.61257458e-01 6.13501191e-01
-2.05400556e-01 2.64043175e-02 -5.76926768e-01 2.52050340e-01
1.22586274e+00 -2.74923772e-01 5.68730772e-01 1.68953389e-02
-5.09117730e-02 -1.14899285e-01 5.93659163e-01 1.14057434e+00
-9.29448083e-02 8.86055112e-01 5.40714383e-01 -5.99411488e-01
-1.19541943e+00 -1.21227276e+00 -1.28884643e-01 4.34693336e-01
-4.28233564e-01 -4.17692512e-01 -1.01129711e+00 -7.41621912e-01
4.38325644e-01 9.35278714e-01 -7.32112527e-01 -3.15390408e-01
-3.75473738e-01 -1.22658551e+00 1.26829922e+00 5.20221770e-01
2.48236790e-01 -1.06838083e+00 -4.99079078e-01 3.19365174e-01
-5.37707247e-02 -7.74090290e-01 -7.74096429e-01 -7.66798630e-02
-6.97999477e-01 -9.66071963e-01 -7.55236328e-01 -3.05137485e-01
7.30916440e-01 -6.30702734e-01 1.20122933e+00 -3.83605301e-01
-6.18171990e-01 2.90191442e-01 5.41351028e-02 -8.23211610e-01
-8.67976546e-01 1.21400550e-01 4.57049021e-03 5.74467964e-02
5.33355057e-01 -4.73951608e-01 -8.45244169e-01 -1.22738540e-01
-1.20009875e+00 -1.14471197e-01 4.28845108e-01 1.06781662e+00
7.11510599e-01 -2.80299932e-01 7.89396942e-01 -1.59381914e+00
1.02961123e+00 -7.84389317e-01 -3.83760959e-01 2.09955662e-01
-8.48993063e-01 4.65883948e-02 8.62298667e-01 -4.01640952e-01
-9.35662508e-01 1.11809768e-01 -3.50048482e-01 -4.88086790e-01
-2.51816899e-01 2.69806534e-01 -3.96461159e-01 3.57140452e-01
7.91216433e-01 9.78952423e-02 2.74530530e-01 -1.95466965e-01
4.27060336e-01 9.18852031e-01 5.01905680e-01 -2.65781343e-01
3.69285852e-01 4.34657246e-01 -1.60047963e-01 -3.25167507e-01
-4.86518234e-01 2.14387611e-01 -1.45522915e-02 6.18617177e-01
9.05865192e-01 -1.11673093e+00 -1.07524145e+00 4.49586928e-01
-8.54593277e-01 -2.03129381e-01 -7.25528121e-01 4.98380452e-01
-5.46947122e-01 1.41081229e-01 -8.73100221e-01 -7.15264678e-01
-8.68265629e-01 -6.88056290e-01 8.11413646e-01 4.47060168e-02
-5.55117309e-01 -7.89124310e-01 7.36827031e-02 1.69325173e-01
4.66941476e-01 9.26578641e-01 8.53636146e-01 -1.18360519e+00
-6.01929247e-01 -4.74814087e-01 6.66843206e-02 3.08623791e-01
3.55149806e-01 -2.57555723e-01 -1.00409639e+00 -4.37168062e-01
4.26422022e-02 -3.15266192e-01 5.12204230e-01 5.29752135e-01
1.71834075e+00 -9.93470430e-01 -3.65194559e-01 1.17920566e+00
1.28606427e+00 3.64616036e-01 9.23870683e-01 -2.53158152e-01
8.05213332e-01 4.04432803e-01 1.54160932e-01 9.86280620e-01
4.15694863e-01 1.12858266e-01 -5.58040058e-03 -4.86259490e-01
3.70309204e-01 -5.83883166e-01 -1.24210827e-01 3.47726375e-01
3.18659306e-01 -4.80731606e-01 -7.92557001e-01 3.89035076e-01
-1.58429456e+00 -9.66435850e-01 3.38770777e-01 2.36412597e+00
1.21851552e+00 -2.84426957e-01 1.46393687e-01 -3.13746572e-01
6.57148361e-01 -2.88445204e-01 -9.26889718e-01 -1.99689910e-01
-1.93216145e-01 4.98122752e-01 9.20145690e-01 1.89144433e-01
-9.84490812e-01 5.22242486e-01 6.87134361e+00 3.31286043e-01
-8.80129635e-01 -1.51022396e-03 1.41310155e+00 -2.39495367e-01
-6.78589821e-01 -6.08086824e-01 -4.81458455e-01 7.00700521e-01
1.05504000e+00 -5.05137026e-01 3.34028900e-01 9.40973043e-01
-3.07949334e-01 6.05054438e-01 -1.28904116e+00 1.07644320e+00
-1.03010550e-01 -1.43036735e+00 2.21101001e-01 2.87281573e-01
8.80710304e-01 -1.45437136e-01 3.76033694e-01 2.13537037e-01
1.11645067e+00 -1.58663106e+00 7.74328085e-03 6.72521651e-01
1.22946513e+00 -1.08524978e+00 9.68641996e-01 1.56193793e-01
-3.87904704e-01 7.17823878e-02 -2.40235239e-01 5.18980861e-01
1.68888703e-01 5.91394722e-01 -1.24527967e+00 4.48168129e-01
5.33994734e-01 1.93285123e-01 -2.00023279e-01 7.02386498e-01
2.26946726e-01 5.33156037e-01 -3.59526873e-01 1.79070070e-01
-2.10137859e-01 6.37801131e-03 2.01956347e-01 1.18089235e+00
5.16358078e-01 3.10006082e-01 2.89568827e-02 1.05675054e+00
-6.03955328e-01 9.74152759e-02 -8.35408211e-01 -2.65906602e-01
6.25572026e-01 9.76848722e-01 8.13375972e-03 -3.96925718e-01
-4.06586193e-02 8.97995412e-01 -7.98073485e-02 3.37679893e-01
-8.06286931e-01 -5.62974870e-01 9.34648097e-01 2.00470075e-01
5.91425272e-03 4.68989968e-01 -5.25274813e-01 -1.23128450e+00
-1.02677651e-01 -1.26523387e+00 7.32603014e-01 -3.59302998e-01
-1.62107491e+00 7.77802706e-01 -2.60589033e-01 -1.07491124e+00
-7.20367610e-01 1.24387443e-02 -2.48858407e-01 1.13268101e+00
-9.10439730e-01 -1.01839900e+00 -2.10260674e-01 7.98246324e-01
-2.53871471e-01 -3.21933866e-01 1.37315881e+00 3.23825389e-01
-2.93792307e-01 1.46236324e+00 3.25282931e-01 5.38558304e-01
7.52447069e-01 -1.27920222e+00 8.09058964e-01 3.03661883e-01
-1.67747021e-01 9.15177941e-01 4.74115938e-01 -8.80103350e-01
-1.29965854e+00 -1.62316096e+00 5.56368411e-01 -4.91178215e-01
2.25468781e-02 -3.91460150e-01 -1.03054047e+00 1.03782094e+00
1.84210524e-01 5.23182303e-02 1.43593502e+00 -3.17054957e-01
-2.83636272e-01 -7.54429176e-02 -2.03875828e+00 7.32000351e-01
7.20035017e-01 -5.33638358e-01 -3.12259823e-01 3.04899633e-01
8.65216315e-01 -5.63224375e-01 -1.51074088e+00 4.19090658e-01
6.91172421e-01 -6.77593827e-01 1.02643836e+00 -1.10083902e+00
4.69572514e-01 1.44911977e-03 -5.34350723e-02 -1.40886533e+00
-2.03843579e-01 -1.14760888e+00 -1.84403583e-01 1.15524352e+00
4.19154942e-01 -8.87967050e-01 9.29409683e-01 1.29428589e+00
3.97639662e-01 -6.63933098e-01 -5.65727651e-01 -3.37583482e-01
1.83422893e-01 2.23111473e-02 1.38982570e+00 1.26591671e+00
-1.70446321e-01 -1.21011004e-01 -8.03293526e-01 1.29943073e-01
7.73929775e-01 -1.72417074e-01 9.48007286e-01 -9.23715115e-01
-6.21706188e-01 1.10903420e-01 -3.27247947e-01 -3.46209913e-01
-5.66047952e-02 -9.74132657e-01 -3.33576620e-01 -1.08221233e+00
1.55730322e-01 -5.21751940e-01 -4.30055588e-01 5.47398031e-01
-4.35648769e-01 1.72463581e-01 5.72223775e-02 -5.27707748e-02
-4.94252294e-02 3.93246144e-01 1.09992015e+00 -1.14467591e-01
-2.96448886e-01 1.77385822e-01 -1.16710246e+00 3.66670907e-01
9.73279119e-01 -6.63733721e-01 -3.63042265e-01 -5.00371493e-02
1.80735644e-02 5.82721829e-01 2.43669361e-01 -6.70700312e-01
1.41920345e-02 -5.56636676e-02 8.12852442e-01 -3.93278688e-01
7.60255307e-02 -6.78288877e-01 8.31657708e-01 5.64647496e-01
-7.42287755e-01 5.32513782e-02 1.73602954e-01 5.79632461e-01
-1.17823379e-02 3.41288865e-01 6.31570101e-01 -2.95478165e-01
3.49357873e-01 6.53038621e-01 -2.42013961e-01 1.29313171e-01
1.09417617e+00 1.64835572e-01 -2.20787525e-01 -6.38220668e-01
-9.60732877e-01 3.12763572e-01 5.32591343e-01 1.34717882e-01
7.23162353e-01 -1.40179503e+00 -9.43616331e-01 5.70037365e-01
1.27602741e-01 1.40203863e-01 4.09423172e-01 1.47241384e-01
-5.23318648e-01 1.22890003e-01 -2.48726398e-01 -1.60102427e-01
-1.16796172e+00 8.55276406e-01 3.38569373e-01 -4.11068380e-01
-6.35590196e-01 5.73743820e-01 2.24111393e-01 -6.90549314e-01
2.67830223e-01 -2.74333745e-01 3.57328475e-01 -2.78989077e-01
6.73155010e-01 2.63261557e-01 -1.75433233e-01 -1.31825507e-02
-1.71892688e-01 -1.85848996e-01 -1.95065022e-01 -1.06486000e-01
1.21424651e+00 1.70425728e-01 1.25176057e-01 7.10902587e-02
1.19799852e+00 1.29013970e-01 -1.11109293e+00 -1.03261977e-01
-4.03354198e-01 -4.64320838e-01 -5.36650896e-01 -1.01075602e+00
-1.13338017e+00 4.31446344e-01 6.17285073e-01 2.86403567e-01
1.07327688e+00 -2.45913267e-01 1.03474760e+00 4.91485685e-01
1.42304540e-01 -5.63822091e-01 -6.17967188e-01 -1.19961746e-01
6.03242218e-01 -1.22782910e+00 -1.51552409e-01 -2.49587148e-01
-9.26888347e-01 4.95931447e-01 3.88414443e-01 3.17572504e-02
7.70521104e-01 6.31850421e-01 3.49542230e-01 3.12238894e-02
-7.24477291e-01 5.11448503e-01 -5.35286381e-05 7.61611521e-01
4.14098561e-01 5.41227758e-01 -5.96333668e-02 9.79733646e-01
-4.63192880e-01 3.08299810e-01 5.58266521e-01 6.76605225e-01
5.48145294e-01 -1.19572222e+00 -3.68328154e-01 9.58692729e-01
-1.20724142e+00 -3.91730428e-01 -2.49522954e-01 6.21603131e-01
-2.41063442e-02 4.80960995e-01 2.15547651e-01 -3.44881386e-01
1.89703763e-01 1.92082837e-01 1.93225354e-01 -4.33468223e-01
-1.13057840e+00 -3.22042912e-01 -3.31305154e-02 -3.62671643e-01
2.64415175e-01 -6.00513160e-01 -1.05903566e+00 -5.39018989e-01
1.19303152e-01 7.40892291e-02 5.23510456e-01 2.95332879e-01
9.13437307e-01 5.96689820e-01 5.67318022e-01 1.69014677e-01
-1.15759146e+00 -8.71686399e-01 -6.93946958e-01 8.79286706e-01
3.03510725e-01 3.78178626e-01 -7.00905770e-02 2.14145586e-01] | [6.263184070587158, 6.7803568840026855] |
da37803f-6b55-4321-98fa-13811151dff9 | deep-reinforcement-learning-from-self-play-in | 1603.01121 | null | http://arxiv.org/abs/1603.01121v2 | http://arxiv.org/pdf/1603.01121v2.pdf | Deep Reinforcement Learning from Self-Play in Imperfect-Information Games | Many real-world applications can be described as large-scale games of
imperfect information. To deal with these challenging domains, prior work has
focused on computing Nash equilibria in a handcrafted abstraction of the
domain. In this paper we introduce the first scalable end-to-end approach to
learning approximate Nash equilibria without prior domain knowledge. Our method
combines fictitious self-play with deep reinforcement learning. When applied to
Leduc poker, Neural Fictitious Self-Play (NFSP) approached a Nash equilibrium,
whereas common reinforcement learning methods diverged. In Limit Texas Holdem,
a poker game of real-world scale, NFSP learnt a strategy that approached the
performance of state-of-the-art, superhuman algorithms based on significant
domain expertise. | ['David Silver', 'Johannes Heinrich'] | 2016-03-03 | null | null | null | null | ['game-of-poker', 'card-games'] | ['playing-games', 'playing-games'] | [-6.45977736e-01 1.57519326e-01 -9.04219598e-02 3.07695359e-01
-7.32122958e-01 -7.06125557e-01 2.50687122e-01 -5.06039858e-01
-7.40078568e-01 1.31798565e+00 -1.33315429e-01 -8.51956680e-02
-1.16329148e-01 -6.20942831e-01 -9.90803778e-01 -2.10815862e-01
-4.18896347e-01 9.76043820e-01 2.64709651e-01 -9.37156975e-01
3.17994922e-01 -1.26789272e-01 -1.23444295e+00 2.94694871e-01
9.40476239e-01 8.07968915e-01 3.88994031e-02 1.01688409e+00
1.70120150e-01 1.56066930e+00 -7.66838074e-01 -7.47112393e-01
9.04974639e-01 -5.44684291e-01 -1.11536551e+00 -7.04183042e-01
2.51767516e-01 -8.98315251e-01 -6.11715019e-01 1.04334438e+00
5.98942459e-01 2.14403033e-01 5.01886904e-01 -1.67275310e+00
-1.91039696e-01 1.11739433e+00 -3.87196988e-01 1.46203339e-01
3.60100478e-01 7.79318631e-01 1.35308862e+00 -5.85761946e-03
7.15018868e-01 1.18489766e+00 9.94777441e-01 7.94222474e-01
-1.13867676e+00 -9.44021702e-01 -3.39096367e-01 2.83853024e-01
-1.18169272e+00 -1.66531026e-01 5.58454573e-01 -1.39802262e-01
1.18221021e+00 -3.44677657e-01 1.14975059e+00 1.04664123e+00
2.12726682e-01 1.20276403e+00 1.08083487e+00 7.45486021e-02
6.22750998e-01 -4.96250480e-01 -8.18737447e-01 7.55719602e-01
2.92218309e-02 8.48248541e-01 -5.68560719e-01 -4.50322121e-01
1.04755330e+00 -5.78009844e-01 3.04217637e-01 -5.59703887e-01
-7.45021462e-01 8.55283320e-01 6.46172702e-01 1.17465287e-01
-5.83669603e-01 8.27426374e-01 4.08749640e-01 7.03077078e-01
-1.78871602e-02 1.23965538e+00 -6.50933444e-01 -8.77191305e-01
-1.09771228e+00 1.00290751e+00 1.52663600e+00 7.18839169e-01
7.20407128e-01 3.48065883e-01 3.78608406e-01 3.71410221e-01
-2.47566298e-01 2.08379433e-01 4.73888397e-01 -1.95065928e+00
1.44730851e-01 1.18937455e-01 6.67857766e-01 -5.43097079e-01
-5.32071769e-01 -2.86410332e-01 -3.60296994e-01 7.64556587e-01
8.03537488e-01 -8.00857425e-01 -4.74350542e-01 1.72190547e+00
1.78541377e-01 7.53090322e-01 2.92720854e-01 1.19764638e+00
2.72474855e-01 6.07534230e-01 5.51569611e-02 2.12582067e-01
7.58782983e-01 -9.22675908e-01 -1.09731052e-02 -3.96243513e-01
3.32553506e-01 -2.49636903e-01 8.94515812e-01 7.55365849e-01
-1.51564169e+00 -1.37605429e-01 -1.02142024e+00 2.18420640e-01
-1.91528331e-02 -7.05513120e-01 8.48332644e-01 5.15871644e-01
-1.29936624e+00 1.05885375e+00 -8.25909376e-01 -1.53841078e-01
6.74634159e-01 5.75944483e-01 -1.89572796e-01 1.95120245e-01
-1.42071259e+00 1.18749511e+00 6.24438584e-01 -3.90238076e-01
-1.56319666e+00 -8.23006034e-01 -4.31083113e-01 2.92071909e-01
9.05961812e-01 -6.98244095e-01 2.01768637e+00 -1.57526934e+00
-2.01432467e+00 7.50645518e-01 7.98339188e-01 -9.11870956e-01
8.91994834e-01 -1.65226519e-01 1.46003351e-01 2.61550039e-01
1.36675507e-01 7.58641124e-01 7.45516360e-01 -1.12862122e+00
-7.07168102e-01 1.28223568e-01 6.63426518e-01 5.18427670e-01
9.61465389e-02 -2.22754464e-01 2.36840338e-01 -3.36605132e-01
-6.84680283e-01 -8.65885317e-01 -4.49757487e-01 -6.86327070e-02
2.75190443e-01 -3.36412489e-01 2.07603022e-01 -3.91651779e-01
7.05212057e-01 -1.67053151e+00 3.41735125e-01 5.85431680e-02
5.78720570e-01 2.22094953e-01 -4.31296438e-01 4.70728517e-01
3.35698634e-01 1.52477384e-01 1.11038163e-01 3.32239985e-01
4.78453517e-01 1.67670935e-01 1.29196793e-01 3.60733837e-01
-1.24876514e-01 1.18851888e+00 -1.61202419e+00 -4.91241306e-01
-2.20283002e-01 -5.33018768e-01 -1.25329316e+00 3.81231040e-01
-6.12662315e-01 -1.70591623e-01 -4.38881487e-01 4.54928428e-01
4.26698357e-01 -5.42995706e-02 5.12361765e-01 4.01964754e-01
6.06885515e-02 -2.96938475e-02 -1.23350739e+00 1.92524505e+00
-4.92989272e-01 4.26017314e-01 3.39978665e-01 -9.88630295e-01
3.41744512e-01 2.80415207e-01 7.19558239e-01 -7.25092709e-01
2.80212790e-01 3.97973031e-01 4.56250697e-01 -1.29657820e-01
6.89324439e-01 -6.13832176e-01 -4.77731824e-01 5.83171546e-01
4.29113567e-01 -7.84430563e-01 2.03271598e-01 2.82914281e-01
1.50994945e+00 4.30353373e-01 6.85823441e-01 -3.27833474e-01
-1.80505171e-01 6.88525140e-01 5.47769845e-01 1.30560279e+00
-7.90475488e-01 2.59249330e-01 8.32136333e-01 -6.20727837e-01
-1.40863061e+00 -1.32071757e+00 6.56922758e-01 1.28100705e+00
3.35778952e-01 -4.22170430e-01 -9.28421378e-01 -5.40505171e-01
1.01424277e-01 4.79278952e-01 -5.27016163e-01 -3.01460356e-01
-4.69505489e-01 -1.32994130e-01 1.02920663e+00 3.99519324e-01
9.16121066e-01 -1.38211894e+00 -9.09869611e-01 6.44044578e-01
-2.29420319e-01 -9.25052285e-01 -4.42136645e-01 1.20399194e-03
-4.88581985e-01 -1.21495652e+00 -6.45652175e-01 -5.51420271e-01
-3.70631546e-01 -2.67502666e-01 1.64358294e+00 2.29027227e-01
-2.65350342e-01 4.39328969e-01 -2.36742571e-01 -3.19174498e-01
-5.95786989e-01 2.14853644e-01 3.03877205e-01 -8.57412696e-01
4.10605550e-01 -8.22032511e-01 -5.80010831e-01 2.82883763e-01
-6.63877249e-01 -1.86172321e-01 6.56098723e-01 1.28968632e+00
-6.36995658e-02 2.14217901e-01 5.69168746e-01 -5.70223331e-01
9.80925739e-01 -6.16185725e-01 -9.69379842e-01 3.07408534e-02
-1.77554905e-01 2.90869862e-01 1.11819565e+00 -7.14950383e-01
-9.75427747e-01 4.99587767e-02 7.88599849e-02 -5.20262599e-01
6.25799149e-02 3.64343792e-01 3.65975678e-01 -2.44857877e-01
1.40771496e+00 2.23154098e-01 1.68725103e-01 2.38200724e-01
3.76836270e-01 4.94479418e-01 4.67135608e-01 -1.25548482e+00
7.70288527e-01 2.09840946e-03 -1.50246650e-01 -5.24436355e-01
-5.25676370e-01 -1.20356731e-01 -2.62104213e-01 -3.37853193e-01
5.26983738e-01 -1.01300561e+00 -1.59940326e+00 7.73287714e-01
-7.03219533e-01 -1.23062980e+00 -5.51028371e-01 2.10254788e-01
-1.41735804e+00 5.69471903e-02 -8.15425038e-01 -6.92651272e-01
-5.87900691e-02 -6.76117301e-01 4.87120807e-01 4.83715504e-01
-2.31830329e-01 -5.81304252e-01 5.83198786e-01 3.91984910e-01
6.18197739e-01 7.55895227e-02 2.68345892e-01 -4.93753165e-01
-5.54248691e-01 4.25870642e-02 -6.82108328e-02 1.96502075e-01
-5.44656217e-01 -3.86778355e-01 -5.25789022e-01 -2.00261906e-01
-5.33924401e-01 -1.33654404e+00 1.92704186e-01 2.89215177e-01
5.82853854e-01 -3.81174564e-01 2.64402211e-01 4.97828662e-01
1.34376526e+00 2.40920484e-02 7.72744060e-01 5.76449215e-01
1.37531549e-01 -6.48899525e-02 4.43066269e-01 9.44230855e-01
4.88275468e-01 2.88967311e-01 5.96136332e-01 3.07615161e-01
3.52836758e-01 -7.60464430e-01 3.29405993e-01 -1.91793516e-01
-3.32566768e-01 4.33290266e-02 -1.12146401e+00 5.73228896e-01
-2.12651610e+00 -1.55915964e+00 7.60201573e-01 1.62603045e+00
1.15274358e+00 4.19962913e-01 8.25469077e-01 -3.66613418e-01
3.30863684e-01 -6.22656494e-02 -1.16900659e+00 -4.73637164e-01
1.00694008e-01 4.29355472e-01 5.90935886e-01 3.14800888e-01
-8.97727668e-01 1.69894230e+00 6.93654490e+00 1.12700176e+00
-7.11790979e-01 -4.22091894e-02 4.33054924e-01 -4.21445429e-01
6.15212843e-02 -5.86653464e-02 2.91108638e-02 2.35416129e-01
9.94567215e-01 -5.49991786e-01 1.57885277e+00 1.10964370e+00
2.48564575e-02 -2.99664319e-01 -9.36074257e-01 8.89993787e-01
-3.96873385e-01 -1.45978117e+00 -4.96830553e-01 3.71642858e-02
9.39775705e-01 2.76151121e-01 -1.79125052e-02 1.07956707e+00
1.75159311e+00 -1.28029239e+00 8.40513229e-01 2.07553819e-01
5.66818655e-01 -9.86724496e-01 6.66385889e-01 7.77443230e-01
-9.38878477e-01 -5.05377591e-01 -4.94366527e-01 -7.93393850e-01
-1.92646965e-01 -3.82713169e-01 -8.22730362e-01 1.41726077e-01
8.01243365e-01 4.34731096e-01 8.44026431e-02 1.07440126e+00
-1.10936962e-01 6.24211252e-01 -7.52675235e-01 -6.11725986e-01
7.90293455e-01 7.17793405e-03 3.79555523e-01 5.01464248e-01
-1.32498011e-01 6.99083865e-01 3.94382805e-01 1.01187074e+00
-4.67506945e-01 -1.50209010e-01 -5.56501448e-01 -3.59836131e-01
3.65847439e-01 9.95261490e-01 -4.73453611e-01 -1.38944522e-01
-1.12447001e-01 1.10976672e+00 8.25293422e-01 1.46771908e-01
-8.71796250e-01 -3.69982541e-01 1.11298013e+00 1.50895014e-01
2.98769057e-01 -1.26380175e-01 -1.09171189e-01 -1.40879691e+00
-5.13649046e-01 -1.28846061e+00 4.02278751e-01 -7.87093043e-01
-1.49906278e+00 2.72461563e-01 -4.37617488e-02 -1.09154999e+00
-7.81196058e-01 -5.54453492e-01 -7.49631584e-01 4.15839851e-01
-1.07838595e+00 -8.95161092e-01 5.15395068e-02 7.05337584e-01
2.49542907e-01 -5.96315980e-01 5.54479420e-01 -1.97508976e-01
7.71645978e-02 7.19272256e-01 2.81133860e-01 3.05394888e-01
3.11209559e-01 -1.44907200e+00 4.70236868e-01 4.53230470e-01
-1.17070585e-01 1.80458128e-02 1.03071368e+00 -5.45020759e-01
-1.34825969e+00 -2.61807323e-01 -9.23206508e-02 -2.21358120e-01
1.19949853e+00 -1.66676015e-01 -3.32120001e-01 4.75049287e-01
3.39914382e-01 -2.70687360e-02 2.05527872e-01 2.44028702e-01
-3.03898275e-01 -4.07337323e-02 -1.51717210e+00 1.00444305e+00
1.22477174e+00 -4.77549881e-01 -7.47487247e-01 6.90064859e-03
3.05955231e-01 -7.75859058e-01 -5.68926752e-01 -3.04949313e-01
8.56626093e-01 -1.04256082e+00 9.53453422e-01 -1.01053166e+00
8.33861649e-01 -1.24561647e-02 7.85265416e-02 -1.79219794e+00
-4.17641252e-01 -1.14762497e+00 -5.92553839e-02 2.70401955e-01
7.67817795e-02 -2.38870665e-01 1.44571197e+00 6.78281009e-01
1.40704572e-01 -3.80120814e-01 -1.13793719e+00 -9.66613591e-01
8.01622987e-01 -1.75545827e-01 5.31135678e-01 6.33360386e-01
6.61227465e-01 -2.95895133e-02 -8.38310063e-01 -1.92658707e-01
8.91651034e-01 -4.75607701e-02 9.39575195e-01 -1.11703539e+00
-8.08002532e-01 -5.95921099e-01 -5.40469706e-01 -9.88738060e-01
6.68921649e-01 -5.80675066e-01 3.65112811e-01 -1.01797152e+00
7.61127472e-02 -1.18898466e-01 -4.85200286e-02 5.25420725e-01
1.11216582e-01 2.45645329e-01 5.31527817e-01 -7.33815059e-02
-1.16231465e+00 5.80741525e-01 1.13166940e+00 -2.47660339e-01
-1.99448034e-01 -8.91285017e-02 -7.58303225e-01 7.00890064e-01
1.03356326e+00 -5.35623252e-01 -3.45537215e-01 -2.43514776e-01
5.11731505e-01 5.52653432e-01 3.03061575e-01 -1.30196106e+00
5.68637609e-01 -7.84249306e-01 1.98991105e-01 2.63564110e-01
2.34097883e-01 -7.45302498e-01 -4.05642726e-02 8.62893224e-01
-1.95922881e-01 -1.31034508e-01 3.81527990e-01 4.31584120e-01
-2.04074313e-03 -2.96479970e-01 9.68075514e-01 -7.90436745e-01
-9.72515464e-01 1.80951461e-01 -8.19577932e-01 9.64878142e-01
1.32298303e+00 -1.97960660e-01 -1.35944858e-01 -1.05774355e+00
-5.65831184e-01 6.94981515e-01 6.34212136e-01 -7.74730518e-02
5.16014874e-01 -9.72600639e-01 -8.23950291e-01 -3.06262404e-01
-3.46235514e-01 -3.41987133e-01 4.42829072e-01 -3.49479578e-02
-1.15081024e+00 -2.70436972e-01 -1.01411498e+00 2.92321667e-02
-6.95419073e-01 3.15268248e-01 1.13648140e+00 -6.64572418e-01
-2.14503706e-01 7.59925187e-01 -2.98187345e-01 -7.75382102e-01
-8.63180868e-03 1.23226590e-01 3.56279701e-01 -4.10495847e-01
2.40751743e-01 4.77557659e-01 -4.42381799e-01 -9.99793177e-04
-1.71039239e-01 -1.17007919e-01 9.02029872e-02 -4.69943762e-01
1.55509973e+00 3.92796397e-01 3.55036587e-01 -9.71080661e-02
6.18760526e-01 -6.18905187e-01 -1.81516337e+00 -1.25097260e-01
-8.89807492e-02 -2.71512151e-01 8.58849585e-02 -1.13422751e+00
-8.44898880e-01 4.90860850e-01 3.35472345e-01 -2.70593408e-02
1.05837119e+00 -2.02145383e-01 9.73535836e-01 1.25664937e+00
1.03478479e+00 -1.56749737e+00 4.08395171e-01 8.96187305e-01
4.40136433e-01 -1.33693242e+00 -6.57141730e-02 6.34303868e-01
-1.02879059e+00 9.56651330e-01 1.15709448e+00 -9.04680490e-01
5.98150849e-01 4.97016907e-01 -1.43942744e-01 4.48378660e-02
-9.97950852e-01 -2.57839739e-01 -3.34100872e-01 7.69558787e-01
-2.74521202e-01 6.43359423e-02 6.11829087e-02 9.41031158e-01
-6.84876561e-01 6.65961087e-01 8.07172000e-01 1.01287210e+00
-6.02018237e-01 -7.12451518e-01 -1.03401765e-01 4.18623328e-01
-5.27251065e-01 -1.54804572e-01 -4.49271530e-01 9.21607494e-01
-2.85983719e-02 6.57763839e-01 -1.25253767e-01 -4.24449980e-01
3.01984489e-01 -3.10025692e-01 7.83427835e-01 -2.78044581e-01
-1.10420501e+00 -4.09385502e-01 2.87194438e-02 -9.70117688e-01
5.49187250e-02 -4.78146136e-01 -1.33973110e+00 -1.15969634e+00
2.04576030e-01 3.28980267e-01 1.30236521e-01 8.97026837e-01
4.79433760e-02 -2.03746520e-02 3.11915129e-01 -1.22708273e+00
-1.04712653e+00 -5.28264403e-01 -1.07534957e+00 4.92091030e-01
1.53166294e-01 -4.72457886e-01 -1.99819714e-01 -4.34136510e-01] | [3.5965468883514404, 1.562039852142334] |
91cedda0-320f-4cff-8cb1-e2828d49efdb | video-based-computer-aided-arthroscopy-for | 1807.09627 | null | http://arxiv.org/abs/1807.09627v1 | http://arxiv.org/pdf/1807.09627v1.pdf | Video-based computer aided arthroscopy for patient specific reconstruction of the Anterior Cruciate Ligament | The Anterior Cruciate Ligament (ACL) tear is a common medical condition that
is treated using arthroscopy by pulling a tissue graft through a tunnel opened
with a drill. The correct anatomical position and orientation of this tunnel is
crucial for knee stability, and drilling an adequate bone tunnel is the most
technically challenging part of the procedure. This paper presents, for the
first time, a guidance system based solely on intra-operative video for guiding
the drilling of the tunnel. Our solution uses small, easily recognizable visual
markers that are attached to the bone and tools for estimating their relative
pose. A recent registration algorithm is employed for aligning a pre-operative
image of the patient's anatomy with a set of contours reconstructed by touching
the bone surface with an instrumented tool. Experimental validation using
ex-vivo data shows that the method enables the accurate registration of the
pre-operative model with the bone, providing useful information for guiding the
surgeon during the medical procedure. | ['Joao P. Barreto', 'Joao Oliveira', 'Cristovao Sousa', 'Pedro Marques', 'Luis Ribeiro', 'Carolina Raposo', 'Rui Melo', 'Fernando Fonseca'] | 2018-07-25 | null | null | null | null | ['medical-procedure'] | ['medical'] | [-2.68407494e-01 2.11874377e-02 -2.16809079e-01 4.48188692e-01
-2.62380868e-01 -3.63186717e-01 -2.29458287e-02 2.76744902e-01
-7.27917075e-01 5.01341522e-01 2.47292668e-01 -3.39310348e-01
8.62846300e-02 -2.09174603e-01 -3.34996939e-01 -2.45310843e-01
-2.97560662e-01 9.57152069e-01 7.16227233e-01 -1.71003014e-01
4.71789718e-01 6.17056012e-01 -1.08037257e+00 -2.36623734e-01
6.05622292e-01 6.34780049e-01 5.03012180e-01 2.02744156e-01
-7.04627857e-02 -2.12504603e-02 -2.65237749e-01 -4.78649944e-01
5.11809647e-01 -3.30873698e-01 -6.62428975e-01 4.21389461e-01
-7.58106783e-02 -1.76540792e-01 3.17565918e-01 8.87213945e-01
1.25869304e-01 -3.15952986e-01 4.61167634e-01 -5.58598042e-01
2.05682531e-01 1.65352911e-01 -9.41049099e-01 -3.05508465e-01
6.10290408e-01 2.42459904e-02 3.12696606e-01 -1.17811048e+00
1.23885977e+00 5.15178442e-01 8.71256292e-01 4.72184211e-01
-1.42754114e+00 -6.47894740e-01 -5.83704650e-01 -1.24292776e-01
-1.26899910e+00 -1.17488228e-01 6.98215842e-01 -1.06155717e+00
3.19301009e-01 1.39046341e-01 1.70161760e+00 2.87666947e-01
1.27840972e+00 1.11165270e-01 1.35807288e+00 -8.98449242e-01
3.59209776e-01 -1.24165108e-02 -1.17686696e-01 6.34443223e-01
5.08599877e-01 4.59724188e-01 -3.15432608e-01 -2.48768210e-01
1.51604629e+00 2.17369869e-02 -5.88398099e-01 -1.35299516e+00
-1.00522339e+00 4.56506491e-01 2.13030130e-01 3.53710234e-01
-7.12212205e-01 -1.04002040e-02 4.90367323e-01 -1.68015823e-01
-5.50532788e-02 2.26058409e-01 1.76642701e-01 -4.48649645e-01
-8.24745893e-01 -2.79544115e-01 8.02018523e-01 2.47097626e-01
5.13479933e-02 -3.31712455e-01 5.93077123e-01 7.68526852e-01
6.42721295e-01 1.98374212e-01 5.43610632e-01 -7.67143309e-01
-4.17770892e-02 7.02760994e-01 1.98342204e-02 -6.37729168e-01
-2.63721675e-01 -1.80756912e-01 -4.98147339e-01 8.05734158e-01
6.79351509e-01 9.79548916e-02 -1.05933762e+00 9.06970322e-01
4.31532979e-01 -2.96861559e-01 -4.39673156e-01 1.00509775e+00
-1.07329287e-01 -1.54243961e-01 9.53731835e-02 -4.52088624e-01
1.79107535e+00 -6.66766286e-01 -7.32204974e-01 -4.75928009e-01
5.44733465e-01 -1.06973410e+00 1.06688190e+00 6.70017600e-01
-1.06481242e+00 -1.60562649e-01 -9.52046216e-01 3.56260419e-01
1.68950200e-01 2.57657886e-01 5.11986613e-01 6.61377370e-01
-5.70486665e-01 3.29919428e-01 -1.14316738e+00 -4.57410812e-01
-2.23836258e-01 5.31455278e-01 -1.15313482e+00 3.26526731e-01
-5.12267232e-01 1.23522913e+00 1.43546611e-01 2.86916822e-01
-1.80942789e-01 -3.24031025e-01 -7.31104314e-01 -4.47250456e-01
4.32317793e-01 -8.36757243e-01 9.17042792e-01 -4.77626354e-01
-1.87874854e+00 1.55405259e+00 1.35828227e-01 -3.44377339e-01
7.46332347e-01 -4.75067407e-01 -4.41736840e-02 4.44456607e-01
1.60991281e-01 -7.17749521e-02 6.05343223e-01 -1.25707340e+00
-1.68304279e-01 -9.08785939e-01 -4.21579123e-01 -1.18534341e-01
1.78331092e-01 -9.34362933e-02 -5.66372275e-01 -8.04707050e-01
8.20272565e-01 -1.16429102e+00 -4.56371456e-01 5.39800704e-01
-2.78449208e-01 3.30679864e-01 6.33421242e-01 -9.88627732e-01
1.20590663e+00 -2.11573768e+00 6.44817874e-02 4.11892295e-01
1.47609323e-01 -7.11575476e-03 6.47308409e-01 5.04815459e-01
-5.93313798e-02 -2.48643532e-01 -1.09712362e-01 2.23202556e-01
-5.73200583e-01 9.31390598e-02 3.39666516e-01 8.13611925e-01
-8.96578670e-01 4.48362380e-01 -5.23745358e-01 -7.89400220e-01
-7.58209005e-02 1.61400527e-01 -3.30534130e-01 2.68951505e-01
4.42737043e-01 6.00172341e-01 -2.85767466e-01 6.56796813e-01
5.15784085e-01 1.51558623e-01 5.11748850e-01 -2.01694906e-01
4.20282595e-02 9.32484716e-02 -1.38993466e+00 2.06623483e+00
-6.04541123e-01 -1.46680728e-01 7.91155279e-01 -3.42186213e-01
1.34902108e+00 4.58637267e-01 6.59570336e-01 -5.39668441e-01
2.00695604e-01 9.60305095e-01 -4.08219881e-02 -4.09162700e-01
5.49820811e-02 -7.99928010e-01 1.61197349e-01 4.64645803e-01
-7.11576119e-02 -4.72689360e-01 -5.56867644e-02 -3.52769941e-02
3.83482903e-01 2.67847180e-01 7.70145893e-01 -5.26181579e-01
4.39211071e-01 1.49822563e-01 3.22100759e-01 6.30270466e-02
1.79905519e-01 5.99211097e-01 2.46004522e-01 -7.46108472e-01
-8.24611545e-01 -1.10725009e+00 -2.46052623e-01 -2.60713339e-01
2.27957189e-01 -1.28237188e-01 -7.77233601e-01 -4.15056527e-01
3.02768290e-01 1.31566763e-01 -6.25766993e-01 -2.05519378e-01
-5.04247904e-01 2.85254478e-01 -1.82723179e-01 2.93761224e-01
-1.36255115e-01 -5.11229217e-01 -1.18745661e+00 6.42089784e-01
8.33979994e-02 -7.22898483e-01 -1.66766062e-01 1.10252008e-01
-1.54112458e+00 -1.28536260e+00 -8.99210930e-01 -6.93771303e-01
6.49950802e-01 -1.48774266e-01 8.60239506e-01 3.39269787e-01
-4.69355196e-01 4.05141234e-01 -1.44789919e-01 2.29953546e-02
-6.21093988e-01 -2.91170746e-01 -3.66429612e-02 -1.55117422e-01
3.74878533e-02 -4.94211525e-01 -6.77372754e-01 4.74665821e-01
-5.37701845e-01 2.56704241e-01 6.25311315e-01 6.56169653e-01
7.53345668e-01 -6.76232755e-01 -8.65609422e-02 -5.31127930e-01
6.27105653e-01 -2.00870587e-03 -4.62913394e-01 -2.51920998e-01
-4.82122421e-01 -3.48638475e-01 6.00792319e-02 -1.93346113e-01
-8.87004972e-01 1.11191705e-01 -6.84392080e-02 -2.73731560e-01
-1.38893187e-01 5.25955617e-01 2.60579377e-01 -2.71208823e-01
4.96128649e-01 -2.00251669e-01 9.97703373e-01 -7.45950103e-01
-1.84932530e-01 4.21736091e-01 8.85728955e-01 -4.76521730e-01
4.28163230e-01 7.22685695e-01 3.33855093e-01 -4.96627182e-01
-1.44318506e-01 -6.56916142e-01 -1.04154110e+00 -6.66007459e-01
5.95644534e-01 -1.85153902e-01 -7.93235242e-01 1.22117475e-01
-9.69173789e-01 -4.87816893e-03 -4.67488855e-01 1.28371775e+00
-7.12016582e-01 6.55798972e-01 -8.92909765e-01 -6.07431471e-01
-5.79788148e-01 -1.36655962e+00 8.42075527e-01 -1.47115692e-01
-3.53984356e-01 -1.04449332e+00 5.67805052e-01 2.88279533e-01
1.20717585e-01 5.14111459e-01 7.33019888e-01 -1.50264978e-01
-1.80604964e-01 -9.02361751e-01 6.05954707e-01 1.74149424e-01
-2.39566411e-03 -7.80032249e-03 -2.64036715e-01 -7.80789331e-02
4.71207768e-01 8.01792964e-02 3.76786888e-01 4.88158703e-01
5.30245185e-01 3.07469398e-01 -5.65147698e-01 4.79961544e-01
1.73491645e+00 2.15028301e-01 8.87022853e-01 7.45479345e-01
2.22122386e-01 6.27801299e-01 1.09215057e+00 1.56628758e-01
-1.64252594e-01 1.08400035e+00 3.98932576e-01 -8.95838141e-02
-2.22146943e-01 -1.74855858e-01 8.81159827e-02 9.47545230e-01
-4.76842493e-01 6.58444226e-01 -1.15093458e+00 3.10848415e-01
-1.12566268e+00 -2.31311411e-01 -4.49157119e-01 2.90335155e+00
9.15362656e-01 2.60379463e-01 -7.46816248e-02 1.76961929e-01
4.05420005e-01 -5.95779955e-01 6.81689829e-02 -5.39742589e-01
7.63939321e-01 2.54782528e-01 4.85495031e-01 7.41256058e-01
-3.68194193e-01 5.80754399e-01 6.30948257e+00 6.17370248e-01
-1.60383725e+00 1.96176190e-02 -2.22010061e-01 3.64446938e-01
-3.23320664e-02 3.36331576e-01 -4.48708445e-01 3.33495766e-01
4.07094151e-01 4.61678952e-02 -3.82029206e-01 5.85274637e-01
3.80163342e-01 -1.13178253e+00 -9.02254164e-01 5.51323175e-01
2.16385182e-02 -1.48265398e+00 -1.56694442e-01 3.21210593e-01
5.30559011e-02 -6.80821657e-01 -3.95965308e-01 -2.70853102e-01
-5.95522821e-01 -5.02409458e-01 6.53293550e-01 7.91578293e-01
9.45546806e-01 -2.56681830e-01 7.87787795e-01 1.76303759e-01
-1.10254645e+00 2.32868150e-01 2.04694681e-02 -1.11998923e-01
6.89617753e-01 5.38007557e-01 -9.94708896e-01 3.00434053e-01
4.07693803e-01 -6.43521696e-02 -1.31584302e-01 1.57297337e+00
-6.68337718e-02 1.72747448e-01 -3.46982658e-01 4.88204479e-01
-1.00925937e-01 -7.07383692e-01 8.77010167e-01 5.73104143e-01
4.16406959e-01 -9.53069478e-02 1.22125946e-01 4.02178794e-01
4.73285854e-01 7.25397348e-01 -4.85240906e-01 4.72316653e-01
1.31864935e-01 9.65689719e-01 -1.06774592e+00 1.65453747e-01
-1.15864486e-01 7.34330177e-01 -2.78781235e-01 3.59086622e-03
-5.34494333e-02 5.54395095e-02 1.28845675e-02 8.83939564e-01
-1.83370873e-01 -3.34014118e-01 -4.89200741e-01 -7.07425654e-01
5.32099307e-01 -4.67125535e-01 2.16990054e-01 -1.04127634e+00
-3.41347158e-01 3.85190606e-01 -2.39217907e-01 -1.58749378e+00
-3.25478286e-01 -7.64737010e-01 -3.29522759e-01 9.63767946e-01
-6.72970414e-01 -1.02590883e+00 -1.85370818e-01 2.17029989e-01
2.57891208e-01 3.00469160e-01 9.60830212e-01 -3.76248479e-01
1.36611357e-01 -2.03389794e-01 -3.85884911e-01 4.37298492e-02
7.91892588e-01 -1.11662734e+00 -4.11088794e-01 5.84004462e-01
-4.61696267e-01 7.60967016e-01 9.44842696e-01 -1.13856125e+00
-1.37187552e+00 1.59396548e-02 6.76387429e-01 6.45605177e-02
7.95905411e-01 -1.02119818e-01 -9.35985804e-01 7.66583741e-01
1.64560795e-01 -1.29933268e-01 9.59310532e-01 -1.32887229e-01
4.78786975e-02 -6.75304905e-02 -1.18208385e+00 4.53538120e-01
4.83276218e-01 -1.41885906e-01 -9.67161000e-01 -2.13912353e-02
-2.95338064e-01 -1.03592479e+00 -1.32961547e+00 2.85334557e-01
1.22798467e+00 -1.07557201e+00 9.72471356e-01 -4.39221710e-02
1.98814094e-01 -2.24347398e-01 6.26803875e-01 -1.07147801e+00
2.26710632e-01 -5.89766383e-01 2.75261402e-01 5.38051367e-01
8.62201601e-02 -7.28825390e-01 1.16444933e+00 3.90215248e-01
-9.28163975e-02 -1.01971018e+00 -1.18794167e+00 -5.98358452e-01
-2.24026993e-01 2.87666116e-02 -2.60340095e-01 4.64907140e-01
5.80084324e-01 -2.15652153e-01 1.63349599e-01 -5.06664850e-02
5.94400167e-01 -3.69498902e-03 8.55864704e-01 -1.67071688e+00
-3.08168352e-01 -2.84889430e-01 -8.31239700e-01 -4.56382275e-01
-3.87900561e-01 -6.68999016e-01 -1.17368333e-01 -1.57646322e+00
4.86121774e-02 -5.23578584e-01 5.20050287e-01 3.23338620e-02
2.25831762e-01 1.48369133e-01 1.26726121e-01 5.32681525e-01
4.61709172e-01 1.33477628e-01 1.80527651e+00 7.08602071e-01
-2.79112309e-01 1.97084770e-01 6.38348237e-02 1.01903093e+00
5.36656916e-01 -5.59265435e-01 3.42644081e-02 2.15016873e-04
7.08355904e-02 5.30346870e-01 2.85693944e-01 -9.35285032e-01
4.54989552e-01 2.29035094e-01 8.65679458e-02 -5.27353823e-01
4.34267610e-01 -1.54447746e+00 7.41346955e-01 1.25957119e+00
2.14709282e-01 2.13771179e-01 1.36946246e-01 3.41097236e-01
-4.05627102e-01 -5.56648672e-01 9.72486317e-01 -4.11894143e-01
-2.81025529e-01 -2.19174623e-01 -5.83167732e-01 -1.32147908e-01
1.26550782e+00 -9.92101789e-01 4.81915742e-01 -2.52699051e-02
-1.56209230e+00 -3.75173867e-01 9.96014893e-01 -1.63352937e-01
9.12159920e-01 -9.06165779e-01 -4.63348657e-01 4.09845561e-01
-1.56904031e-02 -1.11688718e-01 1.88905060e-01 1.67999470e+00
-1.29599786e+00 1.64257288e-01 -7.13712692e-01 -1.06401479e+00
-1.30257213e+00 4.88002181e-01 1.01545799e+00 -3.54538292e-01
-8.21751833e-01 5.61036050e-01 -7.99598470e-02 -1.37115896e-01
-1.35162389e-02 -1.73960313e-01 -1.15346909e-01 -1.73836827e-01
2.47911662e-01 2.22381547e-01 2.73852468e-01 -4.87159133e-01
-2.79923946e-01 1.10816836e+00 1.97273027e-02 -2.86576450e-01
1.13468993e+00 -3.12239885e-01 -4.50626582e-01 6.01812482e-01
5.58317780e-01 6.95138872e-01 -8.53399217e-01 -8.03277642e-02
-5.15933894e-02 -8.96275222e-01 7.42467046e-02 -7.92170823e-01
-1.09096396e+00 8.17483127e-01 6.61409497e-01 -1.87009647e-01
9.63256299e-01 2.04748306e-02 4.77161974e-01 -5.38865745e-01
1.17070365e+00 -9.49862659e-01 -1.39744446e-01 -3.03784519e-01
1.21647108e+00 -3.89564395e-01 -2.89003085e-02 -1.00314283e+00
-4.01178449e-01 1.48113775e+00 3.03933144e-01 -2.79385388e-01
9.43757832e-01 5.74673712e-01 5.38216949e-01 -5.05199313e-01
4.30195928e-02 2.66794860e-01 1.32613659e-01 5.44325590e-01
5.68208992e-01 2.18408421e-01 -1.33032787e+00 1.03411041e-01
-7.04320073e-02 4.62094516e-01 5.81225097e-01 1.43898427e+00
-6.30910814e-01 -1.13005519e+00 -7.84893632e-01 3.14491063e-01
-8.80940735e-01 3.63111764e-01 8.93692523e-02 1.16681290e+00
-2.31089503e-01 3.25251698e-01 -7.04732761e-02 2.93659687e-01
6.22376144e-01 -5.39298505e-02 6.32241488e-01 -5.66644311e-01
-7.70695329e-01 8.07474017e-01 1.52319809e-02 -6.00240171e-01
-1.43092692e-01 -7.49449611e-01 -1.32714498e+00 3.09811532e-01
-2.42658257e-01 5.08165121e-01 1.11288905e+00 7.54275441e-01
-1.10043332e-01 1.38113812e-01 2.47512147e-01 -8.31532001e-01
-5.59400320e-01 -5.72806358e-01 -1.21974623e+00 3.13913524e-01
-2.05308404e-02 -1.27807927e+00 -1.72879532e-01 -7.48084933e-02] | [13.757312774658203, -2.9336304664611816] |
05f13ec2-8a3b-42a3-86dc-96891ccc0dc7 | improving-mesh-based-motion-compensation-by | 2301.04836 | null | https://arxiv.org/abs/2301.04836v1 | https://arxiv.org/pdf/2301.04836v1.pdf | Improving mesh-based motion compensation by using edge adaptive graph-based compensated wavelet lifting for medical data sets | Medical applications like Computed Tomography (CT) or Magnetic Resonance Tomography (MRT) often require an efficient scalable representation of their huge output volumes in the further processing chain of medical routine. A downscaled version of such a signal can be obtained by using image and video coders based on wavelet transforms. The visual quality of the resulting lowpass band, which shall be used as a representative, can be improved by applying motion compensation methods during the transform. This paper presents a new approach of using the distorted edge lengths of a mesh-based compensated grid instead of the approximated intensity values of the underlying frame to perform a motion compensation. We will show that an edge adaptive graph-based compensation and its usage for compensated wavelet lifting improves the visual quality of the lowpass band by approximately 2.5 dB compared to the traditional mesh-based compensation, while the additional filesize required for coding the motion information doesn't change. | ['André Kaup', 'Daniela Lanz'] | 2023-01-12 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 6.37824953e-01 9.70509723e-02 2.70034015e-01 -9.12450403e-02
-4.62179899e-01 8.30499977e-02 1.25447467e-01 3.43212485e-01
-6.63663626e-01 8.27220142e-01 1.71642363e-01 -8.60745832e-02
-4.74362262e-03 -9.61416543e-01 -3.35618228e-01 -6.31156564e-01
-2.25243270e-01 2.82441467e-01 6.28151238e-01 -9.00760517e-02
1.94467589e-01 6.82065248e-01 -1.40898979e+00 6.02267385e-01
5.91167569e-01 1.08345783e+00 2.96043843e-01 9.61733580e-01
-2.23021001e-01 6.09456956e-01 -5.37386239e-01 4.80918363e-02
2.07542688e-01 -7.00270712e-01 -7.18149245e-01 3.60015452e-01
1.65305004e-01 -4.24429446e-01 -2.56802272e-02 9.65535283e-01
5.33418119e-01 3.37397531e-02 4.24517125e-01 -5.92306614e-01
3.25859278e-01 3.03935945e-01 -8.02094400e-01 1.77774444e-01
3.84034187e-01 -1.35925844e-01 2.71729559e-01 -9.82110322e-01
1.18525040e+00 1.01915371e+00 9.47193801e-01 1.42067328e-01
-1.63437140e+00 -2.30048969e-01 -6.78836644e-01 4.96342003e-01
-1.41506743e+00 -1.38051137e-01 9.36456382e-01 -2.56901324e-01
7.80947149e-01 5.11541069e-01 1.19593048e+00 7.82029107e-02
4.69813496e-01 -1.14957333e-01 1.10713911e+00 -6.33371532e-01
1.20222867e-01 -1.70711786e-01 -1.13361627e-01 7.74093211e-01
2.70675898e-01 -1.72001436e-01 -2.95286626e-01 -2.57212043e-01
1.01859128e+00 -6.12127557e-02 -7.52027988e-01 -3.09471339e-01
-1.33387518e+00 6.54247522e-01 3.81555170e-01 7.96252787e-01
-5.81048548e-01 4.60671633e-01 5.55161178e-01 4.35862601e-01
7.80600369e-01 -6.75979778e-02 -6.65918440e-02 6.19596541e-02
-1.51695621e+00 -5.93649130e-03 6.01420701e-01 5.71303427e-01
8.69583070e-01 1.50005758e-01 -3.54031809e-02 6.00838423e-01
-7.55619854e-02 2.32032955e-01 4.93860990e-01 -1.23064804e+00
-3.36787179e-02 4.16197866e-01 -2.84125190e-02 -9.86920416e-01
-5.27628005e-01 -1.93060279e-01 -1.06124687e+00 4.96700376e-01
5.58150232e-01 1.46764189e-01 -7.44501650e-01 1.05917239e+00
6.54762149e-01 1.97107673e-01 -2.43443191e-01 6.36666477e-01
5.92866719e-01 8.72402191e-01 -2.64013678e-01 -8.07114244e-01
1.39096069e+00 -5.77380896e-01 -9.83577907e-01 4.62937891e-01
6.55288815e-01 -1.19419074e+00 7.18172133e-01 7.29932487e-01
-1.57311225e+00 -6.85666323e-01 -1.12094164e+00 -1.58452004e-01
1.57789409e-01 -1.41017109e-01 -3.14536542e-02 4.87631440e-01
-1.14183593e+00 1.04866564e+00 -9.43414152e-01 -1.05383238e-02
1.94338500e-01 2.13730142e-01 -3.32286865e-01 -1.22243457e-01
-7.01485276e-01 8.13449979e-01 1.01207733e-01 -2.30607063e-01
-8.89598057e-02 -1.00467455e+00 -4.98350978e-01 2.25882336e-01
-1.05945216e-02 -6.17368221e-01 7.56405413e-01 -1.05273938e+00
-1.23946416e+00 8.87337923e-01 -7.77566507e-02 -3.62036556e-01
8.67723703e-01 3.14200252e-01 -1.23643428e-01 9.56808388e-01
-1.84749186e-01 2.99441934e-01 1.34967768e+00 -1.07507610e+00
-4.30080861e-01 -4.11969833e-02 -5.49060583e-01 -4.88250516e-02
-9.66993794e-02 -2.10186690e-01 -2.26781592e-01 -1.05041790e+00
3.87917757e-01 -6.47707701e-01 -4.92483020e-01 5.46795726e-01
9.70826447e-02 5.28200686e-01 7.86371410e-01 -8.91235769e-01
1.19402051e+00 -2.31528521e+00 8.79206806e-02 4.65073615e-01
2.12319911e-01 4.09238134e-03 1.40128613e-01 4.05852169e-01
-3.35410774e-01 -3.32833737e-01 -5.04990041e-01 -1.36901081e-01
-6.84075654e-01 2.61224866e-01 -1.40601257e-02 8.23690295e-01
-1.46860585e-01 2.54875600e-01 -7.28300214e-01 -7.70296812e-01
3.57278109e-01 9.75264370e-01 -6.91865563e-01 -1.45109788e-01
3.45760256e-01 6.53750181e-01 -8.65707397e-02 7.95406923e-02
9.38810825e-01 -2.37849075e-02 2.33203441e-01 -7.13631570e-01
-2.15149835e-01 -9.34771672e-02 -1.21911204e+00 1.62562203e+00
-5.70087969e-01 6.44888639e-01 3.97126973e-01 -1.00383389e+00
7.87517130e-01 6.87752903e-01 9.97905672e-01 -5.30152082e-01
2.56521143e-02 4.97155845e-01 -1.10866010e-01 -3.52273464e-01
3.29192489e-01 -5.05645573e-01 4.48314071e-01 2.18819782e-01
-4.21126902e-01 -4.79966283e-01 2.44855866e-01 -1.91431735e-02
1.12014389e+00 -1.08256832e-01 4.44531947e-01 -6.08288944e-01
8.72716486e-01 3.29293936e-01 3.21464777e-01 7.05892444e-02
1.97137743e-01 7.71929085e-01 2.84951657e-01 -7.50863969e-01
-1.61034977e+00 -6.61048174e-01 -4.62720662e-01 3.61608773e-01
-1.95941001e-01 -3.30536991e-01 -9.69217181e-01 -4.95064557e-02
-3.71417224e-01 4.52609628e-01 -2.51644075e-01 1.45996243e-01
-1.06926513e+00 -7.29703426e-01 3.72702144e-02 2.36975387e-01
4.65119451e-01 -7.06345737e-01 -1.22733307e+00 6.11898124e-01
-3.48876774e-01 -9.95320559e-01 -4.15665209e-01 -5.17200446e-03
-1.56283450e+00 -9.14209008e-01 -1.11824560e+00 -6.25051260e-01
9.03906167e-01 3.12486947e-01 1.06887722e+00 5.96270740e-01
-4.89359289e-01 2.72445291e-01 -3.92823845e-01 2.00866796e-02
-7.40671337e-01 -4.48196590e-01 -3.08247119e-01 5.73614165e-02
-5.46244860e-01 -9.24362779e-01 -8.95578921e-01 1.26949787e-01
-1.14006698e+00 4.40911204e-01 2.19082251e-01 8.61581326e-01
8.21166813e-01 9.35360864e-02 1.41991109e-01 -8.11052322e-01
3.65145475e-01 1.40133187e-01 -6.09455526e-01 -1.98191255e-01
-4.54415888e-01 2.86788214e-02 7.20851123e-01 -4.12547350e-01
-7.71741688e-01 1.76701382e-01 -2.90429860e-01 -4.72603798e-01
3.48503441e-01 2.17394888e-01 4.90142077e-01 -5.35160124e-01
5.71555376e-01 1.63705632e-01 3.83076742e-02 -4.70891595e-01
8.98528248e-02 2.43452579e-01 5.37647367e-01 2.27530710e-02
5.72945118e-01 9.32472825e-01 6.75658286e-01 -1.20046496e+00
5.87508008e-02 -4.52730089e-01 -4.85592663e-01 -5.33540070e-01
8.62367392e-01 -4.34299231e-01 -4.88354445e-01 1.42630905e-01
-1.45582902e+00 -1.94067582e-01 -8.52171183e-01 7.67335653e-01
-8.05376232e-01 6.72792494e-01 -7.49330997e-01 -5.60955584e-01
-4.07982528e-01 -1.11289752e+00 8.08142722e-01 -5.00472188e-01
-2.41424561e-01 -8.85661185e-01 3.47133912e-02 7.04409881e-03
6.00276887e-01 5.69920838e-01 1.14712322e+00 2.77072281e-01
-4.21005577e-01 -2.17448547e-01 1.83549039e-02 2.31714472e-01
-9.39045697e-02 -1.47265539e-01 -7.45049417e-01 -3.60957265e-01
6.49865806e-01 2.62820929e-01 6.76048636e-01 7.46726394e-01
1.00503457e+00 -1.99690476e-01 -1.86505958e-01 7.23705709e-01
1.72310638e+00 -5.07460609e-02 8.36998522e-01 2.55567525e-02
2.95423001e-01 6.14353299e-01 4.68204558e-01 6.02176547e-01
-3.55198175e-01 8.88608456e-01 1.96293905e-01 -3.40233564e-01
-6.53856993e-01 2.81834453e-01 2.76595820e-02 1.12704945e+00
-5.70659220e-01 2.24130973e-01 -6.31006181e-01 4.10626978e-01
-1.41773677e+00 -9.97432411e-01 -6.23060167e-01 2.44389439e+00
7.31084704e-01 -2.49980614e-02 -4.36525494e-02 1.04174507e+00
7.33992815e-01 -1.00568078e-01 -8.13853890e-02 -6.40797198e-01
1.93458840e-01 5.63869834e-01 7.48705804e-01 7.68465340e-01
-5.44902384e-01 -3.67285646e-02 6.71203804e+00 1.08710313e+00
-1.12431777e+00 3.28944594e-01 4.19454694e-01 9.41959172e-02
-2.14129642e-01 -1.21787801e-01 -1.90472603e-02 4.15591419e-01
1.07180500e+00 -2.90030807e-01 1.76409826e-01 6.03575945e-01
5.18786192e-01 -5.74776590e-01 -6.16420627e-01 1.11072052e+00
-2.20307410e-01 -1.45816827e+00 -6.56248108e-02 2.61912704e-01
4.35876191e-01 -4.99577016e-01 -2.74317771e-01 -5.05475760e-01
-5.33973217e-01 -7.25036800e-01 6.92987978e-01 5.15242159e-01
1.06206000e+00 -9.18429971e-01 5.17649055e-01 2.17007965e-01
-1.39471471e+00 3.46393883e-01 -4.60615993e-01 6.77841827e-02
5.82668543e-01 1.08125591e+00 -7.05912650e-01 5.51835060e-01
4.83507633e-01 3.46292168e-01 -1.02487423e-01 1.00860858e+00
3.35494787e-01 3.84487838e-01 -3.77368450e-01 5.66320002e-01
4.65196334e-02 -4.43837374e-01 7.22288728e-01 1.17584610e+00
8.51100206e-01 3.86504799e-01 -8.88952687e-02 3.21876585e-01
1.54556975e-01 4.24585789e-01 -3.03782403e-01 5.27553082e-01
-1.33860245e-01 1.10769975e+00 -1.28148937e+00 -8.09605360e-01
-5.91884315e-01 1.24258864e+00 -4.71317023e-01 1.04398489e-01
-6.28872037e-01 -3.85875404e-01 3.35740000e-02 7.96455860e-01
4.98993784e-01 -1.94994912e-01 -1.84033826e-01 -9.33741331e-01
-1.23234086e-01 -5.19071937e-01 1.99256897e-01 -8.46443117e-01
-6.69808328e-01 8.40186357e-01 1.28861576e-01 -1.60199440e+00
-2.59009689e-01 -3.15482199e-01 -2.95465946e-01 8.22194040e-01
-1.21382129e+00 -6.33039296e-01 -3.70025456e-01 7.91093647e-01
5.42968333e-01 5.02005637e-01 7.22360134e-01 6.16099238e-01
3.90559554e-01 2.67309360e-02 4.14905250e-02 -2.55567163e-01
3.33296001e-01 -8.96393836e-01 -4.63987179e-02 7.07000256e-01
-3.24714780e-01 -1.40495285e-01 1.07333958e+00 -6.24361575e-01
-1.04236841e+00 -6.56285405e-01 1.00287783e+00 4.25049990e-01
3.08817059e-01 2.97236502e-01 -1.11301041e+00 2.79562294e-01
3.35124254e-01 3.58038157e-01 4.08008426e-01 -1.07836747e+00
1.61629036e-01 -2.20113173e-01 -1.68570924e+00 2.61444718e-01
6.25489593e-01 -9.56623256e-02 -4.39895868e-01 2.17006475e-01
2.89277077e-01 -4.02050644e-01 -1.21719348e+00 2.31922090e-01
5.73978007e-01 -1.40828192e+00 1.27079105e+00 1.18473031e-01
4.15548086e-01 -3.01263481e-01 9.43902228e-03 -1.07937396e+00
-1.84064150e-01 -5.89321792e-01 1.77337065e-01 5.08399248e-01
-1.96635544e-01 -3.58352482e-01 7.59981573e-01 1.97564200e-01
-9.33442414e-02 -6.10803366e-01 -1.39485025e+00 -5.36514640e-01
-3.45393509e-01 -3.84911746e-01 4.89672683e-02 8.53528738e-01
1.72230124e-01 -1.63931862e-01 -2.84957975e-01 -2.50208646e-01
9.47305858e-01 -3.63144488e-03 2.88607776e-01 -1.13662505e+00
-4.31564152e-01 -1.48977995e-01 -8.59073758e-01 -7.48787522e-01
-6.53196990e-01 -8.03815186e-01 -2.24487096e-01 -1.44480503e+00
-2.22629815e-01 -1.76988244e-01 2.56215513e-01 -6.69999868e-02
3.23702127e-01 8.68818521e-01 3.44652504e-01 2.62549877e-01
3.88847977e-01 -8.32171459e-03 1.68226075e+00 8.54764283e-02
-8.44320580e-02 -1.60539240e-01 3.52022916e-01 9.60691810e-01
3.84766877e-01 -5.94305754e-01 -3.54881406e-01 -1.55563682e-01
2.58605778e-01 6.87881410e-01 2.09325716e-01 -1.28002620e+00
8.59522074e-02 2.47902349e-01 4.51098293e-01 -5.96620798e-01
5.35225093e-01 -1.25293517e+00 8.45358551e-01 1.26886654e+00
-1.26932099e-01 3.12456936e-01 1.04381397e-01 3.35527807e-01
-3.15467954e-01 -5.90714514e-01 1.26216972e+00 -3.45227540e-01
-2.27442607e-01 -2.49883644e-02 -8.46216798e-01 -4.37013298e-01
9.20010626e-01 -6.90230966e-01 4.73380297e-01 -4.44988072e-01
-1.10692859e+00 -5.83133817e-01 6.14043772e-01 -7.25090086e-01
9.30251420e-01 -1.35899603e+00 -6.99806690e-01 3.85175228e-01
-5.14261365e-01 -1.21365421e-01 4.55344379e-01 1.35608435e+00
-1.44083679e+00 -1.34097785e-01 -4.83535200e-01 -5.79431832e-01
-1.62924051e+00 5.02890825e-01 1.84768394e-01 -5.91500938e-01
-1.17468572e+00 3.38523835e-01 -1.69996440e-01 6.78379297e-01
-1.43146366e-01 -6.34211540e-01 -5.60141988e-02 1.28340706e-01
7.33453691e-01 7.90714085e-01 3.47564459e-01 -6.90575361e-01
-1.56479359e-01 9.98079360e-01 5.08936524e-01 -4.98501062e-01
1.47361124e+00 -2.69154191e-01 -3.23166519e-01 2.25499123e-01
1.30349112e+00 2.59674728e-01 -9.78934467e-01 1.14522755e-01
-2.15900779e-01 -6.11010969e-01 4.38191921e-01 -3.04609351e-02
-1.34145057e+00 9.69988525e-01 6.92369461e-01 4.70045507e-01
1.71724892e+00 -4.71895367e-01 1.04091716e+00 -1.88669860e-01
4.55767423e-01 -9.81108904e-01 -2.03055963e-01 -1.87253296e-01
9.52603042e-01 -3.98600519e-01 5.74992776e-01 -8.39768171e-01
-2.07652487e-02 1.68745708e+00 -5.10941446e-01 -4.32239473e-01
7.40446031e-01 5.34442067e-01 -1.76750779e-01 -1.40799237e-02
-4.96329844e-01 -2.27849446e-02 1.74810320e-01 3.96522492e-01
5.75954616e-01 -3.23811620e-01 -1.03896010e+00 -2.82992929e-01
8.48003551e-02 3.19003671e-01 7.64063835e-01 1.00731444e+00
-7.43262887e-01 -1.11506629e+00 -9.51679289e-01 3.47656459e-01
-6.33434057e-01 -6.50199130e-02 3.58772784e-01 5.85697651e-01
1.00108102e-01 4.53128785e-01 1.96026966e-01 1.10722370e-01
4.28879976e-01 -8.46351963e-03 8.75430048e-01 -1.67060390e-01
-6.40004933e-01 6.90030396e-01 4.28563133e-02 -8.39947104e-01
-8.42888772e-01 -3.55842292e-01 -1.40776742e+00 -5.63732445e-01
8.68466347e-02 2.15211753e-02 6.27349615e-01 3.95292878e-01
-1.50034636e-01 7.08006144e-01 4.73732620e-01 -1.30456638e+00
-3.95478234e-02 -5.76029003e-01 -7.56484628e-01 6.92001164e-01
4.50059563e-01 -3.40481699e-01 -3.58569920e-01 5.58121979e-01] | [11.511945724487305, -2.299452066421509] |
878b1631-70ea-482f-ab5c-333c58d18a42 | a-unified-dro-view-of-multi-class-loss | 2112.14869 | null | https://arxiv.org/abs/2112.14869v4 | https://arxiv.org/pdf/2112.14869v4.pdf | Label Distributionally Robust Losses for Multi-class Classification: Consistency, Robustness and Adaptivity | We study a family of loss functions named label-distributionally robust (LDR) losses for multi-class classification that are formulated from distributionally robust optimization (DRO) perspective, where the uncertainty in the given label information are modeled and captured by taking the worse case of distributional weights. The benefits of this perspective are several fold: (i) it provides a unified framework to explain the classical cross-entropy (CE) loss and SVM loss and their variants, (ii) it includes a special family corresponding to the temperature-scaled CE loss, which is widely adopted but poorly understood; (iii) it allows us to achieve adaptivity to the uncertainty degree of label information at an instance level. Our contributions include: (1) we study both consistency and robustness by establishing top-$k$ ($\forall k\geq 1$) consistency of LDR losses for multi-class classification, and a negative result that a top-$1$ consistent and symmetric robust loss cannot achieve top-$k$ consistency simultaneously for all $k\geq 2$; (2) we propose a new adaptive LDR loss that automatically adapts the individualized temperature parameter to the noise degree of class label of each instance; (3) we demonstrate stable and competitive performance for the proposed adaptive LDR loss on 7 benchmark datasets under 6 noisy label and 1 clean settings against 13 loss functions, and on one real-world noisy dataset. The code is open-sourced at \url{https://github.com/Optimization-AI/ICML2023_LDR}. | ['Tianbao Yang', 'Yiming Ying', 'Dixian Zhu'] | 2021-12-30 | null | null | null | null | ['image-classification-shift-consistency'] | ['computer-vision'] | [-4.40261364e-02 -2.14473739e-01 -2.40110949e-01 -7.60815918e-01
-1.42869127e+00 -5.84084809e-01 5.05098850e-02 4.14582431e-01
-4.93361354e-01 9.57047403e-01 -2.93174952e-01 -1.35094717e-01
-6.03865147e-01 -6.80594742e-01 -6.15786195e-01 -9.89534438e-01
-1.70783296e-01 3.23106885e-01 -1.15393914e-01 1.65917631e-02
-1.51697304e-02 5.22513926e-01 -1.61133790e+00 -3.05733345e-02
9.86928880e-01 1.38195384e+00 -2.07553223e-01 2.79374808e-01
1.74550429e-01 4.18980241e-01 -4.56370473e-01 -5.03767550e-01
3.85116875e-01 -2.23423183e-01 -8.63295972e-01 -2.24398375e-01
2.24380851e-01 3.98226023e-01 9.30964872e-02 1.35991514e+00
7.70535707e-01 2.72811502e-01 1.00521302e+00 -1.51259005e+00
-5.70055664e-01 5.24418771e-01 -7.37362802e-01 -4.06102501e-02
3.60604450e-02 7.96020962e-03 1.25508249e+00 -8.88649344e-01
4.29958135e-01 1.11766756e+00 7.68054664e-01 5.13805211e-01
-1.20708239e+00 -9.40349936e-01 2.53211379e-01 6.39532134e-02
-1.67450929e+00 -7.78851286e-02 7.15235174e-01 -4.10404980e-01
5.46971738e-01 4.16160047e-01 -6.38676658e-02 8.62803936e-01
1.44967705e-01 7.44519591e-01 1.42043734e+00 -2.83647418e-01
3.63677919e-01 3.00600439e-01 4.46197450e-01 5.94471633e-01
5.45434318e-02 1.97718758e-02 -1.92052752e-01 -3.44905972e-01
-3.67884263e-02 -1.89321131e-01 -3.38805258e-01 -2.50367165e-01
-7.69927800e-01 9.13545370e-01 3.16893429e-01 2.25552410e-01
5.11984080e-02 3.98392193e-02 6.17472947e-01 4.52288061e-01
8.86151612e-01 2.65874922e-01 -7.26422608e-01 4.80128489e-02
-7.72556067e-01 2.59862602e-01 6.22783482e-01 1.02623761e+00
8.10129046e-01 -5.96676692e-02 -2.76544660e-01 1.24600494e+00
3.86151165e-01 6.90866351e-01 3.75339001e-01 -7.42454290e-01
5.15901744e-01 4.46396828e-01 4.97640632e-02 -7.59336054e-01
-6.48765802e-01 -6.54616952e-01 -9.33413088e-01 1.30779460e-01
4.64650363e-01 -8.55523348e-02 -6.61252856e-01 2.19857359e+00
3.69773000e-01 -4.52403948e-02 -1.85381934e-01 6.11447692e-01
7.39957631e-01 3.27681690e-01 2.34177664e-01 -4.03004587e-01
1.23497808e+00 -5.37420273e-01 -6.14833534e-01 1.02721721e-01
5.70966780e-01 -4.91036564e-01 1.32343102e+00 4.68035877e-01
-9.30029094e-01 -3.46255392e-01 -1.05263937e+00 1.34590447e-01
-6.03325486e-01 8.49154368e-02 2.69595832e-01 7.73001909e-01
-7.69973516e-01 5.53911984e-01 -4.07834142e-01 -1.47236753e-02
4.07179505e-01 3.39892566e-01 -3.41676623e-01 -1.74766839e-01
-1.41699207e+00 7.30632663e-01 3.46854120e-01 8.37932825e-02
-7.34106600e-01 -9.30128932e-01 -9.67709005e-01 -1.88105091e-01
3.83996129e-01 -4.21956867e-01 7.79506385e-01 -7.11673796e-01
-1.33032453e+00 1.09000528e+00 -4.30647470e-02 -2.03671739e-01
7.35634208e-01 -1.00545108e-03 -4.48584080e-01 -1.25707045e-01
2.10040569e-01 2.64282167e-01 5.89454651e-01 -1.32426679e+00
-3.89222562e-01 -6.21048391e-01 -2.46820956e-01 1.39219716e-01
-2.64294893e-01 1.30097821e-01 1.50805876e-01 -8.70855272e-01
8.95200893e-02 -7.51384020e-01 -4.09263000e-02 -1.43472508e-01
-5.73527753e-01 -3.56883675e-01 5.95929801e-01 -4.07485783e-01
1.31048119e+00 -2.20252061e+00 -2.71846596e-02 3.72022659e-01
2.58092433e-02 2.61816114e-01 -1.04033776e-01 3.05083334e-01
-3.90817016e-01 4.42766398e-01 -6.93965256e-01 -3.69004041e-01
4.31504190e-01 -1.27117798e-01 -2.97128588e-01 7.58082330e-01
2.92391181e-02 8.07424664e-01 -9.25542533e-01 -3.46353114e-01
1.24091983e-01 4.88216072e-01 -2.66778886e-01 3.27497013e-02
-1.42351955e-01 2.84379005e-01 -3.75269234e-01 8.53505492e-01
9.85192120e-01 -2.39304110e-01 -1.17395721e-01 -2.69228578e-01
1.65620536e-01 -2.43834272e-01 -1.56807721e+00 1.39410496e+00
-2.86142379e-01 -1.34345159e-01 2.16507122e-01 -1.14143825e+00
1.03203619e+00 1.66896358e-01 6.50788784e-01 -3.90622228e-01
1.65064707e-01 4.18443233e-01 -4.62761611e-01 -2.12070197e-01
2.44070441e-01 -6.26101613e-01 -4.96687025e-01 4.32656527e-01
2.16839835e-01 -2.12153882e-01 3.77337970e-02 -1.63052306e-01
9.08523679e-01 -1.07653618e-01 3.36206764e-01 -6.82302594e-01
6.81291759e-01 -5.42950213e-01 9.04200256e-01 6.41842365e-01
-6.55681968e-01 5.76006174e-01 4.64440644e-01 -2.17786178e-01
-6.35610938e-01 -1.31610250e+00 -6.91187084e-01 1.10596240e+00
2.29857415e-01 -7.89881721e-02 -5.13518810e-01 -8.85687828e-01
2.40344658e-01 8.53406727e-01 -5.38945675e-01 -2.01262981e-01
-1.77844748e-01 -1.49412060e+00 7.25360751e-01 3.94160867e-01
6.01908803e-01 -7.84961522e-01 -6.57187251e-04 -4.68012132e-02
-2.75391728e-01 -6.74051404e-01 -6.50954604e-01 5.92434466e-01
-4.91923630e-01 -1.15860450e+00 -4.62250501e-01 -5.68239868e-01
5.10177314e-01 -3.52924645e-01 1.02275300e+00 -2.91804522e-01
-3.34789723e-01 4.55323964e-01 -2.50260562e-01 -4.82908547e-01
-1.43636912e-01 -2.30507622e-03 2.37526432e-01 1.41056344e-01
3.20127994e-01 -4.99928951e-01 -4.89682913e-01 5.40476263e-01
-1.00021386e+00 -7.47309268e-01 1.08419627e-01 9.17149007e-01
1.14853120e+00 3.76840115e-01 1.07960057e+00 -1.03608155e+00
8.24505150e-01 -7.27828860e-01 -5.50032139e-01 5.12163699e-01
-9.59571242e-01 -2.61824336e-02 8.30925286e-01 -1.54210940e-01
-9.20811653e-01 -1.81110293e-01 -4.15992945e-01 -2.85315186e-01
-1.91490278e-01 2.87200600e-01 -5.48542500e-01 -8.25445727e-03
6.15618050e-01 1.13258407e-01 -1.94557056e-01 -5.46729624e-01
4.98764753e-01 7.38509595e-01 2.13089481e-01 -1.18692064e+00
5.82657933e-01 4.47897524e-01 1.57041565e-01 -5.40660203e-01
-9.64783072e-01 -5.30838072e-01 -5.46953142e-01 -6.62346184e-02
6.06017232e-01 -7.25873113e-01 -8.69420171e-01 7.85168946e-01
-5.46684563e-01 -1.93294436e-01 -5.41681886e-01 3.73274207e-01
-6.77807808e-01 5.24929821e-01 -6.14677370e-01 -8.74061823e-01
-5.14194727e-01 -1.24464524e+00 9.84084368e-01 9.80615914e-02
1.44359186e-01 -1.18718433e+00 7.05563603e-03 2.96810150e-01
3.17192823e-01 6.36227846e-01 1.06068695e+00 -9.13814247e-01
2.08422933e-02 -2.27549866e-01 -2.42880747e-01 9.25406277e-01
3.83065864e-02 -9.94081497e-02 -1.08577192e+00 -5.74702442e-01
1.77687705e-01 -6.84411228e-01 8.15837204e-01 5.05673826e-01
1.44369900e+00 -1.39117420e-01 -1.75549343e-01 7.38005161e-01
1.69839501e+00 1.07205108e-01 3.38687778e-01 1.16375580e-01
4.26536113e-01 5.11513948e-01 7.48752534e-01 4.96112108e-01
4.19275284e-01 6.32906199e-01 3.84485722e-01 4.19729128e-02
1.42890885e-01 -9.81878042e-02 1.75916016e-01 8.90481412e-01
2.90665090e-01 -3.30504179e-01 -6.31611586e-01 4.89156723e-01
-1.74365199e+00 -7.51966536e-01 5.52756973e-02 2.47511840e+00
1.00972736e+00 -9.72131118e-02 2.09590688e-01 1.25751063e-01
9.35547054e-01 2.18103662e-01 -8.21057320e-01 -4.49446887e-01
-3.79420668e-01 3.96002769e-01 5.22299767e-01 6.48299396e-01
-1.37061095e+00 5.34066439e-01 5.72277641e+00 1.41921616e+00
-8.42532694e-01 2.65979648e-01 9.98097837e-01 -1.00172132e-01
-4.41979885e-01 -4.30159360e-01 -8.56240988e-01 7.10507512e-01
7.98606157e-01 -2.78060198e-01 1.37867719e-01 9.22991395e-01
5.75175788e-03 1.41494542e-01 -1.01893961e+00 9.45569992e-01
-5.58970235e-02 -6.97762132e-01 -1.85979828e-01 -6.71742531e-03
6.54038548e-01 -5.71087562e-03 4.19000894e-01 3.03196073e-01
4.25642371e-01 -1.17353892e+00 7.15832233e-01 5.15153527e-01
1.04251575e+00 -1.04290342e+00 8.71234596e-01 2.59121150e-01
-1.32194829e+00 -1.69643208e-01 -3.36861730e-01 3.51385206e-01
6.03156574e-02 1.06876373e+00 -2.61019558e-01 9.76585805e-01
8.21826160e-01 6.30568743e-01 -4.25209433e-01 9.03277516e-01
-1.48448318e-01 5.43185890e-01 -3.60285491e-01 2.27424487e-01
3.43582481e-02 -2.31510714e-01 6.18759036e-01 1.27617192e+00
1.77953884e-01 -7.15181530e-02 1.87466249e-01 8.87960136e-01
-2.74882048e-01 3.69288713e-01 -4.05286074e-01 4.67089266e-01
7.16001332e-01 1.17584753e+00 -6.46118820e-01 9.23301578e-02
-5.46937529e-03 6.83574378e-01 3.27067852e-01 2.99760431e-01
-9.93806124e-01 -7.07753241e-01 8.86245668e-01 -2.08101735e-01
1.34101540e-01 3.18426937e-01 -3.70581210e-01 -1.11095512e+00
2.84801155e-01 -7.36820936e-01 9.43210006e-01 -3.64159375e-01
-2.03771663e+00 5.56335747e-01 1.12208672e-01 -1.25071514e+00
1.32365435e-01 -6.54649854e-01 -2.22669154e-01 9.51492846e-01
-1.70407748e+00 -9.49269295e-01 6.13039285e-02 6.53619409e-01
1.98436528e-02 4.56549115e-02 9.20978427e-01 6.22793436e-01
-6.43439174e-01 1.18369055e+00 5.01531363e-01 -7.79152801e-03
8.02802444e-01 -1.50146043e+00 -3.70150328e-01 6.10512435e-01
-1.11365452e-01 4.65651006e-01 5.38048744e-01 -2.92731166e-01
-7.96819866e-01 -1.46313250e+00 7.54888237e-01 -5.63075721e-01
5.29172897e-01 -4.21443850e-01 -8.67784917e-01 5.03840685e-01
-2.76030540e-01 4.18828279e-01 1.03423059e+00 -1.12640664e-01
-7.05175042e-01 -5.33261359e-01 -1.97234869e+00 7.83001184e-02
9.10067260e-01 -5.19358873e-01 -3.06059062e-01 4.67156231e-01
7.47958601e-01 -3.29861611e-01 -1.46174073e+00 7.87725627e-01
3.63691092e-01 -8.71040702e-01 1.01629269e+00 -4.56049711e-01
3.27201784e-02 -4.53542560e-01 -5.34205019e-01 -1.26042795e+00
-1.52076304e-01 -3.65968496e-01 1.90280631e-01 1.43594933e+00
5.27369320e-01 -1.03847873e+00 2.56995589e-01 6.70730829e-01
-1.34346306e-01 -1.20533812e+00 -1.25810111e+00 -9.76493180e-01
6.63523912e-01 -7.25928187e-01 6.94332659e-01 1.01457953e+00
-6.51552305e-02 -1.86032578e-01 -2.74036914e-01 1.22843720e-01
8.72227311e-01 7.60490000e-02 3.19874063e-02 -1.11537218e+00
-2.10254937e-01 -4.64777142e-01 -3.65824431e-01 -5.52976072e-01
4.40344989e-01 -1.28952491e+00 -5.41189639e-03 -1.04684520e+00
2.24974856e-01 -7.80880332e-01 -8.46144617e-01 5.75069785e-01
-1.76704377e-01 3.27701747e-01 1.15626283e-01 6.39820844e-02
-8.02834868e-01 7.09465981e-01 8.99427772e-01 -1.26351297e-01
6.08663708e-02 2.17783585e-01 -9.16533589e-01 6.08820736e-01
7.21514225e-01 -5.86466670e-01 -4.11542594e-01 1.19197872e-02
1.57950118e-01 -3.05270791e-01 2.59199560e-01 -7.98073053e-01
-1.36020899e-01 -9.70157012e-02 1.28340675e-02 -3.39916468e-01
1.13485165e-01 -7.58787930e-01 5.70234433e-02 2.02420279e-01
-5.35562098e-01 -3.75540480e-02 1.75840661e-01 7.64342129e-01
-2.66946316e-01 -2.13397115e-01 1.31820881e+00 -2.43600849e-02
-4.54276681e-01 5.26218653e-01 2.18992874e-01 7.30976582e-01
1.25416970e+00 1.09496623e-01 -4.21330571e-01 -1.12525657e-01
-7.19579637e-01 3.60475361e-01 4.31302875e-01 2.76001960e-01
3.12252283e-01 -1.51039493e+00 -8.09693694e-01 5.05123213e-02
3.36560905e-01 9.65636224e-02 3.34798932e-01 7.83258021e-01
-1.06432833e-01 2.45495513e-01 2.97438681e-01 -5.38209558e-01
-1.00021243e+00 5.34022272e-01 6.42195463e-01 -4.10807878e-01
-2.83850789e-01 1.18157673e+00 1.64789036e-01 -9.30444360e-01
4.21058357e-01 -2.30138734e-01 2.18207613e-01 3.50306481e-02
4.23852623e-01 6.60225093e-01 2.71193862e-01 -7.77594864e-01
-6.61965072e-01 8.04514885e-01 1.29387990e-01 1.50851935e-01
1.33657134e+00 -9.25428867e-02 -3.09487909e-01 5.78469813e-01
1.55838585e+00 -7.81541169e-02 -1.15497601e+00 -3.29216063e-01
1.95158029e-03 -3.20068419e-01 -2.94736177e-01 -1.18745291e+00
-1.20138407e+00 7.47698426e-01 6.96585357e-01 1.17100418e-01
1.26854885e+00 5.22675551e-02 5.86235464e-01 -1.19548187e-01
4.63925511e-01 -1.29542387e+00 -2.15583757e-01 3.73884976e-01
6.91612542e-01 -1.17453837e+00 -1.06255703e-01 -3.05207610e-01
-6.42234802e-01 8.31209540e-01 4.43626136e-01 1.11464433e-01
1.20512259e+00 2.17535019e-01 6.25014082e-02 -1.81935281e-01
-4.56202775e-01 -6.19840845e-02 3.06994736e-01 5.49320519e-01
3.98930579e-01 2.35797331e-01 -6.93006992e-01 1.07596517e+00
-1.67860053e-02 -2.39969760e-01 3.72234993e-02 4.47979540e-01
-1.17896810e-01 -1.08716083e+00 -1.20663367e-01 4.77170825e-01
-7.07823932e-01 5.86636290e-02 -1.00780241e-02 6.45321071e-01
4.74998087e-01 1.03587103e+00 -4.45604801e-01 -2.95395404e-01
3.47831249e-01 4.09010947e-01 2.54616737e-01 -4.64149863e-01
-4.56916094e-01 -3.31386000e-01 -2.10843652e-01 -5.79752088e-01
-3.74431044e-01 -6.84913814e-01 -1.21353483e+00 -1.49977818e-01
-3.24864298e-01 3.65180939e-01 5.20685971e-01 7.22651362e-01
4.22699720e-01 1.71685532e-01 8.72257292e-01 -3.37107062e-01
-1.07432961e+00 -8.68589222e-01 -1.21547425e+00 6.83783710e-01
1.66970655e-01 -8.51609290e-01 -9.15315390e-01 -3.13715667e-01] | [9.222163200378418, 3.914842367172241] |
3fef91ff-70f0-40b0-8dfa-f410303b72ab | interpretable-deep-learning-methods-for | 2302.07930 | null | https://arxiv.org/abs/2302.07930v1 | https://arxiv.org/pdf/2302.07930v1.pdf | Interpretable Deep Learning Methods for Multiview Learning | Technological advances have enabled the generation of unique and complementary types of data or views (e.g. genomics, proteomics, metabolomics) and opened up a new era in multiview learning research with the potential to lead to new biomedical discoveries. We propose iDeepViewLearn (Interpretable Deep Learning Method for Multiview Learning) for learning nonlinear relationships in data from multiple views while achieving feature selection. iDeepViewLearn combines deep learning flexibility with the statistical benefits of data and knowledge-driven feature selection, giving interpretable results. Deep neural networks are used to learn view-independent low-dimensional embedding through an optimization problem that minimizes the difference between observed and reconstructed data, while imposing a regularization penalty on the reconstructed data. The normalized Laplacian of a graph is used to model bilateral relationships between variables in each view, therefore, encouraging selection of related variables. iDeepViewLearn is tested on simulated and two real-world data, including breast cancer-related gene expression and methylation data. iDeepViewLearn had competitive classification results and identified genes and CpG sites that differentiated between individuals who died from breast cancer and those who did not. The results of our real data application and simulations with small to moderate sample sizes suggest that iDeepViewLearn may be a useful method for small-sample-size problems compared to other deep learning methods for multiview learning. | ['Sandra E Safo', 'Ju Sun', 'Han Lu', 'Hengkang Wang'] | 2023-02-15 | null | null | null | null | ['multiview-learning'] | ['computer-vision'] | [-2.30588347e-01 3.18371169e-02 -4.09603924e-01 -6.89825654e-01
-4.98670310e-01 -1.82677701e-01 4.11342978e-01 3.25536609e-01
-1.34822447e-02 7.17662275e-01 3.83121550e-01 2.41525874e-01
-2.19635054e-01 -7.86890864e-01 -6.92225575e-01 -9.11063731e-01
-4.96703207e-01 3.17625821e-01 -5.98666191e-01 1.32808149e-01
-5.73495291e-02 3.64766449e-01 -1.41981709e+00 4.87316012e-01
6.86713934e-01 7.84046650e-01 3.98894995e-02 4.23700899e-01
-8.56498182e-02 1.59491986e-01 -2.08450437e-01 -3.48484427e-01
8.21094364e-02 -3.86957586e-01 -2.76231587e-01 -1.39992490e-01
5.55217683e-01 1.78810686e-01 -1.17108613e-01 8.32569480e-01
8.70007277e-01 -2.60672897e-01 7.37405658e-01 -1.40656781e+00
-7.99328625e-01 4.20932055e-01 -8.82481217e-01 -1.74310952e-02
1.27492264e-01 1.58149615e-01 1.07668877e+00 -1.30393040e+00
8.80985856e-01 1.53482819e+00 6.29746616e-01 4.86957222e-01
-1.65953279e+00 -5.93923271e-01 1.41115189e-01 3.92437398e-01
-1.20593333e+00 -2.70280868e-01 6.32701218e-01 -5.78091621e-01
7.01883137e-01 4.38693523e-01 9.70380545e-01 1.45482934e+00
1.13964963e+00 6.73978508e-01 1.15878308e+00 -6.00389168e-02
2.89144605e-01 7.45649859e-02 1.30966932e-01 1.00276339e+00
2.25332528e-01 5.43338209e-02 -9.62522388e-01 -3.37723136e-01
2.62348920e-01 4.62231606e-01 -1.57620162e-01 -7.87666082e-01
-1.35264933e+00 9.42506433e-01 3.88588041e-01 3.90995555e-02
-3.04685116e-01 -1.71358109e-01 6.29728377e-01 4.78401095e-01
7.23771393e-01 2.55524963e-01 -5.80171287e-01 6.14542484e-01
-3.52311581e-01 -1.24790512e-01 5.98377824e-01 5.10791361e-01
8.77269983e-01 1.47177562e-01 -1.41560286e-02 8.31234574e-01
4.33779180e-01 4.72966313e-01 5.93034923e-01 -7.17351854e-01
2.08509266e-01 1.07702255e+00 -4.75473017e-01 -1.27812171e+00
-9.55003202e-01 -4.95658666e-01 -1.40883207e+00 3.76676142e-01
1.34436578e-01 -3.30959499e-01 -6.80815816e-01 1.62913549e+00
3.70963693e-01 -1.03829086e-01 8.34589750e-02 7.19094932e-01
1.06153798e+00 6.37454152e-01 -8.14608634e-02 -2.85231829e-01
1.30676544e+00 -4.96588677e-01 -7.20210850e-01 1.45550475e-01
7.43804157e-01 7.61251077e-02 9.07448292e-01 5.76799154e-01
-6.92528009e-01 -4.10749435e-01 -1.03014088e+00 9.82237235e-02
-5.98895907e-01 8.63144994e-02 6.38241291e-01 2.37031385e-01
-8.20912182e-01 4.92618740e-01 -8.47708344e-01 -1.40918225e-01
5.97176075e-01 4.38300282e-01 -9.36440289e-01 1.81972478e-02
-1.29038668e+00 5.36391199e-01 3.36795926e-01 1.77091599e-01
-9.31418002e-01 -1.02923155e+00 -9.21024740e-01 1.64806485e-01
2.35842451e-01 -1.07637846e+00 2.60747075e-01 -1.01301754e+00
-1.12398171e+00 8.49892437e-01 -1.97425425e-01 -5.01339883e-02
3.80845100e-01 -1.15536749e-01 -4.87787575e-01 -1.84398636e-01
1.48225486e-01 4.28058863e-01 7.02616215e-01 -1.01703322e+00
-1.78331241e-01 -1.03774190e+00 -5.19182682e-01 1.34650677e-01
-6.14447653e-01 -6.13650680e-01 -3.18078786e-01 -3.98914367e-01
1.94754168e-01 -7.57164657e-01 -1.44916847e-01 3.75735521e-01
-4.88127649e-01 -2.88696349e-01 9.11555767e-01 -5.80251992e-01
5.60627222e-01 -2.11302423e+00 8.40862691e-01 2.34339744e-01
5.93848407e-01 -7.06690252e-02 -1.23639390e-01 5.63021481e-01
-3.82129699e-01 2.21207440e-01 3.28192040e-02 -2.92664953e-02
-3.37609529e-01 -3.39858457e-02 4.06018734e-01 8.31567407e-01
-7.21862093e-02 8.64248574e-01 -8.25408041e-01 -3.21261346e-01
1.62440494e-01 5.71796894e-01 -2.82434404e-01 1.00926407e-01
-5.43426126e-02 3.63979876e-01 -3.06596190e-01 8.54167402e-01
5.24343133e-01 -5.91599762e-01 5.72699308e-01 -5.09860337e-01
9.14571285e-02 -6.30794942e-01 -1.03932798e+00 1.58338904e+00
-3.51619035e-01 8.76183331e-01 2.04236329e-01 -1.15179205e+00
9.99842405e-01 2.90748388e-01 5.18648446e-01 -7.21840441e-01
-3.12049519e-02 -1.84355721e-01 -2.08665773e-01 -9.40234959e-01
-3.16924036e-01 -1.94074214e-01 1.41381219e-01 1.60468802e-01
3.46554697e-01 4.89354372e-01 -2.92295247e-01 1.40448421e-01
7.14785814e-01 -7.82186389e-02 6.12766206e-01 -4.27812696e-01
5.64138174e-01 -4.86377388e-01 1.02111709e+00 2.95856148e-01
5.20997606e-02 1.42469659e-01 8.54347944e-01 -7.88356602e-01
-8.10940087e-01 -1.11236346e+00 -1.60305187e-01 1.02853322e+00
-1.19019328e-02 -5.19370794e-01 -8.26280415e-02 -7.85373509e-01
3.44994634e-01 3.13221812e-01 -9.37851906e-01 -2.33338207e-01
-2.70509124e-01 -1.16729510e+00 2.00591147e-01 4.47401017e-01
5.47591932e-02 -5.55996418e-01 -2.08359852e-01 -2.03953698e-01
-4.93079051e-03 -4.59057897e-01 -5.87822236e-02 3.94396544e-01
-1.00467241e+00 -1.24474585e+00 -5.19638717e-01 -6.29850268e-01
8.30333471e-01 8.97699147e-02 9.03999805e-01 -3.37876320e-01
-5.88313222e-01 3.34117293e-01 -5.87699190e-02 -5.77266097e-01
-1.11119471e-01 -2.06350520e-01 2.71090209e-01 1.58175960e-01
1.51976839e-01 -5.55957258e-01 -6.49690092e-01 1.99437097e-01
-7.35045195e-01 2.14815304e-01 5.23635805e-01 1.29098427e+00
7.90770829e-01 -3.27640027e-01 9.81991649e-01 -1.06740212e+00
3.45615506e-01 -6.65517271e-01 -7.01403618e-01 4.29280847e-01
-7.75388718e-01 2.76808560e-01 7.41769850e-01 -2.02272609e-01
-8.57418418e-01 7.10519850e-02 1.87999904e-01 -3.52636337e-01
1.38440114e-02 9.78197575e-01 -5.16509712e-01 1.92289546e-01
6.87899947e-01 1.24071851e-01 4.56383497e-01 -3.72305959e-01
4.23980802e-01 3.78661007e-01 6.50816932e-02 1.27701953e-01
1.39084086e-01 5.97323895e-01 5.41758239e-01 -9.37279820e-01
-4.78664339e-01 -1.79575086e-01 -8.63550305e-01 -4.32304174e-01
9.66814280e-01 -1.16500974e+00 -1.10392761e+00 3.52441758e-01
-5.98001242e-01 2.43060123e-02 1.74073845e-01 6.72062099e-01
-3.99850309e-01 1.99794367e-01 -3.27196002e-01 -2.93698281e-01
-5.71617901e-01 -9.51163888e-01 8.32803428e-01 9.09289494e-02
-1.96502522e-01 -1.35573030e+00 2.76119500e-01 2.26994306e-01
-6.00317121e-02 8.14019978e-01 1.54955661e+00 -8.58111918e-01
-3.28533739e-01 -2.41916195e-01 1.45595372e-01 9.04359743e-02
4.06897843e-01 2.98872162e-02 -1.03922069e+00 -5.36956668e-01
-2.49138460e-01 -5.20352066e-01 1.03724849e+00 5.06836832e-01
1.36048234e+00 -3.03094000e-01 -8.07190597e-01 1.00097382e+00
1.45281315e+00 9.30636132e-04 1.99510187e-01 9.26744640e-02
9.05215621e-01 8.27824652e-01 2.67618418e-01 6.31324470e-01
2.77554125e-01 5.15016556e-01 7.16880798e-01 -2.75274426e-01
3.34215045e-01 1.18296426e-02 2.90452868e-01 7.24084437e-01
1.91563189e-01 -2.24210203e-01 -8.64704609e-01 2.19662845e-01
-1.86138940e+00 -9.81673717e-01 -1.70533225e-01 2.14732933e+00
4.78414595e-01 -3.62561435e-01 7.60908192e-03 -3.94551545e-01
6.08884811e-01 2.54859895e-01 -1.05758989e+00 -3.29831868e-01
-3.38862032e-01 -3.11818153e-01 1.78638101e-01 2.45867640e-01
-1.09063125e+00 2.10135639e-01 6.17522573e+00 3.03211063e-01
-1.54332912e+00 -1.60270467e-01 8.48556697e-01 -2.86977082e-01
-5.07684350e-01 -4.95929509e-01 -6.57425761e-01 1.22457050e-01
9.24891889e-01 -2.69330055e-01 -6.60654008e-02 6.53903961e-01
3.47360879e-01 1.06575966e-01 -1.43637431e+00 1.10879970e+00
3.04241925e-01 -1.62922275e+00 8.68035778e-02 9.96275917e-02
6.01234794e-01 2.24567801e-01 7.07113892e-02 -3.23538221e-02
1.41171739e-01 -9.92391229e-01 3.67137156e-02 9.11177337e-01
6.08467758e-01 -8.63974631e-01 6.73426151e-01 4.32753026e-01
-8.83946717e-01 -2.10941851e-01 -3.67237121e-01 4.24491465e-01
-7.69515708e-02 8.92325997e-01 -9.88165617e-01 1.00516665e+00
6.69000804e-01 1.44462192e+00 -6.98387861e-01 7.89419889e-01
2.85732865e-01 2.29954302e-01 2.06248656e-01 1.32846003e-02
-3.34026694e-01 -2.99801826e-01 5.77434778e-01 9.90573049e-01
3.51811528e-01 -3.03701758e-01 2.67751352e-03 8.40348303e-01
-7.57502988e-02 3.24995846e-01 -1.00359547e+00 3.27509455e-03
1.67832538e-01 1.51578462e+00 -5.56656837e-01 -2.28769466e-01
-6.77486300e-01 8.24877679e-01 4.37561363e-01 3.37020993e-01
-3.68634671e-01 -5.39968722e-02 6.10099792e-01 -1.36635065e-01
1.69332877e-01 2.01830581e-01 -3.03074569e-01 -1.35468578e+00
-3.29238892e-01 -1.02548957e+00 6.90124750e-01 -4.16634530e-01
-1.58845663e+00 1.41496554e-01 1.74655784e-02 -1.24314392e+00
-1.60245746e-01 -7.54308820e-01 -6.14247739e-01 8.12013388e-01
-1.02533281e+00 -1.02074063e+00 -2.30590850e-01 4.49806720e-01
3.83160859e-01 -5.52239776e-01 1.13457835e+00 9.13903117e-02
-8.32075357e-01 5.06514311e-01 6.86002195e-01 -4.59355488e-02
9.73363280e-01 -1.24818730e+00 1.67523846e-02 5.92932254e-02
-4.58203331e-02 5.88990033e-01 5.20946264e-01 -6.93134904e-01
-1.96321738e+00 -1.17437494e+00 4.14046794e-01 -1.73370644e-01
2.61791646e-01 -6.76404119e-01 -7.76644647e-01 6.52665138e-01
3.49935919e-01 2.71245718e-01 1.48839915e+00 4.16497082e-01
-3.99631530e-01 -5.11720717e-01 -9.72747147e-01 5.60238719e-01
5.80536962e-01 -3.42086822e-01 -1.63880050e-01 2.92504519e-01
3.40615064e-01 -4.98647951e-02 -1.15766513e+00 4.41389799e-01
1.08365321e+00 -9.76244450e-01 1.01020348e+00 -8.49620223e-01
1.90666467e-01 2.18031229e-03 -1.46286294e-01 -1.89706540e+00
-4.61319864e-01 -4.85603549e-02 -1.03626177e-01 9.34793174e-01
4.62875247e-01 -7.21458495e-01 6.21503353e-01 3.18390191e-01
2.06082426e-02 -1.00299239e+00 -8.90215397e-01 -5.17255962e-01
-2.60866266e-02 2.28690684e-01 1.84698030e-01 1.18939853e+00
1.82060212e-01 5.19221961e-01 -5.92382729e-01 1.82767920e-02
7.63875723e-01 1.94623411e-01 7.86625624e-01 -1.62688828e+00
-2.46384576e-01 -9.54827890e-02 -5.97792208e-01 -1.21438071e-01
4.85845238e-01 -1.37429249e+00 -5.44893563e-01 -1.54928935e+00
5.10808527e-01 1.14075825e-01 -3.80017996e-01 4.81304705e-01
-1.03293598e-01 -1.75805643e-01 -4.43103947e-02 -2.22891420e-01
-2.81468302e-01 7.45835304e-01 1.26823282e+00 -6.36733472e-01
-2.80355483e-01 -1.59019336e-01 -7.70985544e-01 6.95497692e-01
6.16565943e-01 -2.96236336e-01 -3.73496860e-01 -8.97728314e-04
5.07832468e-01 3.37862998e-01 3.63689691e-01 -5.64552665e-01
9.89744160e-03 -1.33627355e-01 1.14604604e+00 -6.38377726e-01
3.84304553e-01 -7.32954025e-01 4.46773887e-01 7.20545411e-01
-4.55351382e-01 2.30328858e-01 1.24362960e-01 9.55268681e-01
-1.73848309e-02 2.31666714e-01 5.16111732e-01 -2.45005414e-01
-6.52775407e-01 3.17782521e-01 -4.86625522e-01 -2.47554541e-01
1.12182641e+00 -1.18613886e-02 -3.85361999e-01 -3.62978905e-01
-1.15239370e+00 6.08531356e-01 1.82300210e-01 4.83079493e-01
9.80798900e-01 -1.53696501e+00 -8.54258597e-01 6.33113563e-01
3.21784168e-01 -2.91154116e-01 4.00107682e-01 9.51685369e-01
-9.97848883e-02 1.92741528e-01 -5.58127403e-01 -8.99798691e-01
-1.83491635e+00 6.03737354e-01 4.00610238e-01 -4.61180955e-02
-7.04240859e-01 5.78269124e-01 4.48592097e-01 -6.78583741e-01
2.02274427e-01 1.32681131e-01 -4.62435752e-01 7.50255644e-01
5.66863596e-01 7.41434991e-01 -5.44089787e-02 -4.22860831e-01
-4.64798182e-01 3.99968028e-01 -1.55591205e-01 2.58538872e-01
1.78638756e+00 -1.26487538e-01 -4.97717559e-01 1.05118489e+00
1.51640415e+00 -3.16933692e-01 -8.95985067e-01 2.13036968e-04
-1.37519062e-01 -2.82223940e-01 2.11297814e-02 -7.65636146e-01
-1.17599308e+00 9.78684187e-01 1.07171810e+00 -8.93535987e-02
8.99897099e-01 -2.40021553e-02 6.26325095e-03 6.37706041e-01
6.04565553e-02 -1.00874388e+00 4.06829894e-01 1.84428662e-01
1.03353095e+00 -1.56154156e+00 3.04685920e-01 -1.46187348e-02
-5.95823824e-01 1.50928295e+00 7.17977524e-01 2.89395541e-01
8.75444889e-01 7.73848146e-02 1.75140500e-01 -6.47784591e-01
-1.10211957e+00 5.19147158e-01 3.93116385e-01 6.06524885e-01
7.27855027e-01 1.94465965e-01 -2.22380459e-01 3.80863696e-01
4.01217610e-01 -1.15909703e-01 2.71669984e-01 5.34668684e-01
-1.87837481e-01 -1.03048766e+00 -3.20588291e-01 9.46841061e-01
-4.53210771e-01 2.59232372e-01 -6.00192010e-01 7.72559762e-01
2.02077534e-02 4.61414576e-01 5.14651984e-02 -4.77012992e-01
1.06409207e-01 3.27050924e-01 2.02709660e-01 -4.17968929e-01
-3.81843537e-01 3.10180873e-01 -3.66275422e-02 -6.55528963e-01
-3.86383235e-01 -8.58647645e-01 -1.23509729e+00 -4.11414541e-02
-2.90665895e-01 -1.38379127e-01 4.30272996e-01 7.80163825e-01
7.21459150e-01 5.82700670e-01 8.12189043e-01 -7.03914225e-01
-2.81619430e-01 -6.54270232e-01 -7.94519424e-01 3.14391792e-01
5.73671877e-01 -5.28941095e-01 -2.18547314e-01 -3.80754247e-02] | [6.465691566467285, 5.506194591522217] |
e9fdb63a-994b-4fb4-b7b0-8343ccebc566 | globalflownet-video-stabilization-using-deep | 2210.13769 | null | https://arxiv.org/abs/2210.13769v3 | https://arxiv.org/pdf/2210.13769v3.pdf | GlobalFlowNet: Video Stabilization using Deep Distilled Global Motion Estimates | Videos shot by laymen using hand-held cameras contain undesirable shaky motion. Estimating the global motion between successive frames, in a manner not influenced by moving objects, is central to many video stabilization techniques, but poses significant challenges. A large body of work uses 2D affine transformations or homography for the global motion. However, in this work, we introduce a more general representation scheme, which adapts any existing optical flow network to ignore the moving objects and obtain a spatially smooth approximation of the global motion between video frames. We achieve this by a knowledge distillation approach, where we first introduce a low pass filter module into the optical flow network to constrain the predicted optical flow to be spatially smooth. This becomes our student network, named as \textsc{GlobalFlowNet}. Then, using the original optical flow network as the teacher network, we train the student network using a robust loss function. Given a trained \textsc{GlobalFlowNet}, we stabilize videos using a two stage process. In the first stage, we correct the instability in affine parameters using a quadratic programming approach constrained by a user-specified cropping limit to control loss of field of view. In the second stage, we stabilize the video further by smoothing global motion parameters, expressed using a small number of discrete cosine transform coefficients. In extensive experiments on a variety of different videos, our technique outperforms state of the art techniques in terms of subjective quality and different quantitative measures of video stability. The source code is publicly available at \href{https://github.com/GlobalFlowNet/GlobalFlowNet}{https://github.com/GlobalFlowNet/GlobalFlowNet} | ['Ajit Rajwade', 'Devansh Jain', 'Jerin Geo James'] | 2022-10-25 | null | null | null | null | ['video-stabilization'] | ['computer-vision'] | [-1.48849398e-01 -2.98857868e-01 -2.51514167e-01 -1.20982625e-01
-2.87526309e-01 -6.56773210e-01 1.80551916e-01 -4.20546412e-01
-3.51231605e-01 6.99932337e-01 5.06612323e-02 -6.19358942e-02
3.38570923e-02 -4.88403231e-01 -9.08811331e-01 -8.35565090e-01
-7.72418827e-03 -2.97351211e-01 6.06367409e-01 -1.72734186e-01
4.37201649e-01 3.88403952e-01 -1.06864893e+00 -1.33685797e-01
7.53268301e-01 8.67934763e-01 8.40124637e-02 8.91403317e-01
4.45479691e-01 1.11204600e+00 -2.70498037e-01 -3.39342475e-01
5.61542451e-01 -6.26782238e-01 -9.25693691e-01 2.07834616e-01
8.30755889e-01 -6.47664905e-01 -6.29748702e-01 1.28675747e+00
4.30468917e-01 5.77274203e-01 6.49710745e-02 -1.09562922e+00
-4.11135554e-01 1.26566231e-01 -5.67510009e-01 5.42551875e-01
3.65809649e-01 4.86134231e-01 6.91619933e-01 -7.61595786e-01
8.20345998e-01 1.20233226e+00 5.48815131e-01 6.27099454e-01
-1.29943299e+00 -6.89810336e-01 2.12449670e-01 2.09753588e-01
-1.32762218e+00 -6.23514593e-01 8.65442276e-01 -5.34708023e-01
4.41893280e-01 1.47236109e-01 6.32584095e-01 7.24500775e-01
2.56309003e-01 4.49552685e-01 5.36139369e-01 -2.90394992e-01
-7.50909671e-02 -1.44222021e-01 -7.99566507e-02 1.00206101e+00
-9.18118656e-03 5.69374636e-02 -5.18083811e-01 1.29947737e-01
1.37841427e+00 -3.62840220e-02 -8.47906053e-01 -4.41075683e-01
-1.30465019e+00 6.78075790e-01 3.23241353e-01 9.54506770e-02
-1.38446704e-01 3.16762149e-01 2.75231928e-01 3.26664388e-01
4.90522921e-01 3.41730505e-01 -2.27489829e-01 -1.31832317e-01
-1.05681968e+00 1.80684566e-01 7.87276447e-01 8.00941288e-01
1.01485968e+00 1.40981108e-01 -7.60740563e-02 7.18498588e-01
3.65069509e-01 2.30999798e-01 4.59601969e-01 -1.75339305e+00
4.84775186e-01 2.10805878e-01 2.38897622e-01 -1.41480172e+00
6.90399949e-03 -4.70730923e-02 -8.44982564e-01 4.78584796e-01
5.96170068e-01 -4.25198853e-01 -5.84165335e-01 1.81222045e+00
6.26694322e-01 6.66882396e-01 -1.89088464e-01 1.25530684e+00
4.26129431e-01 7.52633154e-01 -3.76611978e-01 -3.84069830e-01
9.11111951e-01 -1.43436503e+00 -6.49469435e-01 3.65697481e-02
3.39811206e-01 -9.94936109e-01 7.62189329e-01 3.77344638e-01
-1.44259822e+00 -6.80052578e-01 -1.01897836e+00 -2.26151973e-01
1.82400137e-01 7.72805139e-02 9.44268927e-02 3.15393984e-01
-1.27165020e+00 8.87393117e-01 -1.20457351e+00 -2.63055921e-01
1.81116432e-01 3.78393382e-01 -3.55142385e-01 2.26964112e-02
-1.13253319e+00 7.13753045e-01 3.31932545e-01 2.84497738e-01
-9.11791265e-01 -7.87922621e-01 -1.13937426e+00 -2.11943965e-02
5.20677626e-01 -8.19313467e-01 1.08433425e+00 -1.22706389e+00
-1.98852360e+00 6.18032157e-01 -2.49285266e-01 -1.95778310e-01
7.94198632e-01 -4.70328957e-01 -1.35535121e-01 4.89587218e-01
-8.28192830e-02 6.94495440e-01 1.11089587e+00 -1.03839195e+00
-5.92309833e-01 2.31657848e-01 2.23700613e-01 2.12939128e-01
-2.60113865e-01 1.09871440e-01 -8.86681080e-01 -8.13160419e-01
-1.37805253e-01 -1.16174018e+00 -2.01693639e-01 4.24580902e-01
-2.03862578e-01 1.24531254e-01 9.99426782e-01 -7.33921349e-01
1.45693958e+00 -2.09185839e+00 2.53738582e-01 1.41351491e-01
2.39244595e-01 4.94509488e-01 -1.56688839e-01 2.07699090e-02
-2.11538672e-01 -1.38416275e-01 -1.60466567e-01 -3.57060820e-01
-5.26626885e-01 2.05366444e-02 -1.80930912e-01 7.15606689e-01
1.64196864e-01 5.86049676e-01 -1.04490221e+00 -4.64357227e-01
4.85784382e-01 7.11816072e-01 -8.81469667e-01 2.28916332e-01
2.31044460e-02 6.66806698e-01 -4.42002535e-01 2.16015533e-01
7.12278128e-01 -2.66178668e-01 3.80360372e-02 -2.31596723e-01
-3.26039076e-01 -5.84539622e-02 -1.46441889e+00 1.77304780e+00
-6.21325932e-02 7.76520908e-01 2.99942046e-01 -1.04984069e+00
7.13810146e-01 3.55169326e-01 8.25013638e-01 -8.71174112e-02
3.00278783e-01 -1.54637825e-02 -1.85903981e-01 -5.99352777e-01
3.83819371e-01 1.50786653e-01 5.72255611e-01 1.36278138e-01
1.71929330e-01 -1.18357405e-01 4.99814868e-01 1.41605362e-01
9.06722009e-01 5.38862050e-01 4.85231578e-02 -2.09170625e-01
9.70160544e-01 -1.29262313e-01 8.89982820e-01 3.30696166e-01
-5.43017805e-01 1.06447828e+00 6.09772325e-01 -7.04255879e-01
-1.06546390e+00 -7.64939189e-01 2.04349458e-01 6.43807471e-01
4.54667151e-01 -5.56899011e-01 -8.98383737e-01 -4.65441465e-01
-3.23226154e-01 7.04475865e-02 -3.72032881e-01 -1.57334492e-01
-9.54981625e-01 -3.75001490e-01 2.80988425e-01 3.28933448e-01
6.84100807e-01 -9.41432416e-01 -6.40328348e-01 1.83193520e-01
-4.84415233e-01 -1.26251113e+00 -1.18341208e+00 -4.71597612e-01
-8.58455658e-01 -1.19594908e+00 -9.55272615e-01 -8.03036094e-01
7.16399968e-01 5.29185593e-01 8.56217980e-01 4.05937225e-01
-2.55315974e-02 2.31706366e-01 -1.00609310e-01 2.72769392e-01
-2.23560393e-01 -9.07846987e-02 1.45024911e-01 2.77575552e-01
-1.37569189e-01 -5.09217083e-01 -9.71216857e-01 5.49014390e-01
-9.73919570e-01 8.94329231e-03 -4.73369248e-02 6.95102572e-01
4.65884715e-01 -8.47332366e-03 -2.12141126e-01 -4.21158403e-01
3.82256120e-01 -1.73391327e-01 -9.72476840e-01 1.43571505e-02
-2.30169788e-01 -1.24448694e-01 7.98420668e-01 -6.39659643e-01
-1.00582540e+00 1.48273274e-01 3.63048986e-02 -1.06564796e+00
1.29622236e-01 1.91127300e-01 9.85414386e-02 -3.82703424e-01
4.51792002e-01 7.41705894e-02 3.63714635e-01 -4.33220752e-02
2.20761195e-01 6.46525621e-02 7.23737299e-01 -4.01194781e-01
1.02675092e+00 5.74910998e-01 9.75144189e-03 -8.15563083e-01
-8.51385593e-01 -5.73444426e-01 -7.80961275e-01 -4.24584240e-01
1.10815895e+00 -9.08621192e-01 -9.62788880e-01 7.21684813e-01
-1.17124665e+00 -5.56960464e-01 -5.48233241e-02 8.02663445e-01
-6.21167004e-01 6.50967717e-01 -6.73715413e-01 -3.02769214e-01
-2.69514263e-01 -1.28828835e+00 7.14901805e-01 4.54942793e-01
-1.03903323e-01 -1.28711069e+00 3.24111402e-01 3.03229570e-01
2.56635994e-01 3.69306803e-01 2.37399995e-01 1.26655564e-01
-8.65284681e-01 9.44272894e-03 9.68207605e-03 8.61409128e-01
2.80684799e-01 4.15088058e-01 -5.17903924e-01 -5.66051185e-01
1.37507170e-01 -8.33460242e-02 6.94437385e-01 6.49810553e-01
1.06814885e+00 -3.85739744e-01 -7.00933933e-02 1.07789576e+00
1.42726779e+00 1.53024912e-01 8.05544972e-01 4.16149378e-01
8.74866068e-01 3.92558753e-01 4.71814662e-01 3.06198090e-01
2.60084480e-01 5.96208572e-01 3.46515954e-01 2.07008682e-02
-1.86388910e-01 -8.91036242e-02 6.39584661e-01 7.63722122e-01
-5.00923395e-01 -1.69056952e-01 -6.14768326e-01 4.04164851e-01
-2.00391889e+00 -1.21108341e+00 3.79450596e-03 2.30952454e+00
7.99052298e-01 -4.39959653e-02 3.80340181e-02 -9.29718316e-02
9.40359116e-01 4.61627007e-01 -4.04527783e-01 -1.47950858e-01
1.32701635e-01 -1.34231716e-01 5.00058711e-01 9.34677064e-01
-1.25246704e+00 1.10975301e+00 5.15095758e+00 4.39970762e-01
-1.57033658e+00 -8.96036401e-02 5.97500384e-01 -1.85592845e-01
2.77089477e-01 2.05302060e-01 -6.78910673e-01 6.70359313e-01
5.62702417e-01 -1.11450680e-01 5.32617152e-01 4.78808254e-01
7.31439590e-01 -6.75858930e-02 -8.57617199e-01 8.83349061e-01
1.90633446e-01 -1.40812683e+00 -2.00390786e-01 -1.74643412e-01
8.82023633e-01 -2.12271869e-01 -4.34061065e-02 -2.54710823e-01
1.34153023e-01 -5.63426077e-01 7.29831934e-01 6.51062846e-01
6.92333698e-01 -5.10625660e-01 4.53506351e-01 1.33090615e-01
-1.26909733e+00 1.82295188e-01 -3.39451343e-01 7.65924081e-02
3.98776203e-01 3.10561299e-01 -4.10212763e-02 4.43850219e-01
8.94524336e-01 1.04492283e+00 -2.19555244e-01 1.16230845e+00
-3.03139240e-01 5.50194979e-01 -2.00474262e-01 5.08324504e-01
2.60536164e-01 -5.31970799e-01 7.45637119e-01 1.00839639e+00
2.88670808e-01 2.87666231e-01 2.59145975e-01 6.47094369e-01
-1.43208474e-01 -4.93108816e-02 -2.51231134e-01 4.51332659e-01
5.28743342e-02 1.33900177e+00 -5.74347138e-01 -3.27003181e-01
-4.76920187e-01 1.09898937e+00 2.16802984e-01 6.63383961e-01
-8.93520176e-01 -4.73804742e-01 7.70742059e-01 9.53822583e-02
2.96170235e-01 -4.06061321e-01 2.75126457e-01 -1.72424495e+00
1.86254114e-01 -8.08720410e-01 2.87501633e-01 -8.59225690e-01
-7.74578035e-01 4.28807944e-01 -9.22785774e-02 -1.55156815e+00
-2.36426339e-01 -3.86940598e-01 -8.85033667e-01 7.82156527e-01
-1.50973797e+00 -7.35336900e-01 -6.15846872e-01 9.45690453e-01
6.12648487e-01 4.45564762e-02 1.96198389e-01 4.26100016e-01
-6.57128632e-01 4.14583504e-01 6.07804619e-02 3.64568561e-01
1.17496049e+00 -9.32271063e-01 1.22980021e-01 1.26158106e+00
-2.81290501e-01 6.39146030e-01 7.27504432e-01 -4.92233485e-01
-1.06768298e+00 -1.19941580e+00 7.03308702e-01 -2.91173279e-01
7.57603765e-01 9.19313170e-03 -1.09950531e+00 8.93963099e-01
3.79110724e-01 6.12711072e-01 1.33977890e-01 -7.28317022e-01
9.43368971e-02 -2.87823498e-01 -8.71298969e-01 5.64497471e-01
7.69463778e-01 -2.06620589e-01 -2.17248604e-01 1.88463360e-01
7.12637126e-01 -7.29380369e-01 -8.85764301e-01 1.47054777e-01
5.57969213e-01 -9.39332664e-01 9.08028483e-01 -3.41834277e-01
7.13633060e-01 -6.85149908e-01 2.24890575e-01 -1.28156793e+00
-3.66695404e-01 -1.33030176e+00 -2.65266359e-01 1.10263801e+00
3.28537002e-02 -5.37145793e-01 8.99480402e-01 6.22713327e-01
-2.11546812e-02 -5.67771375e-01 -6.99942410e-01 -7.17880666e-01
2.13051513e-02 -1.20252362e-02 3.23847472e-03 1.04988039e+00
-1.31420091e-01 -3.66164707e-02 -7.06378520e-01 2.96420991e-01
6.80666387e-01 -3.18353087e-01 8.57858837e-01 -6.96329117e-01
-9.10552293e-02 -3.27217758e-01 -4.21306312e-01 -1.41790700e+00
1.78962991e-01 -4.49957550e-01 7.22773150e-02 -1.07485950e+00
-6.69211224e-02 1.70793403e-02 -1.47092566e-01 2.52354592e-01
-3.22185099e-01 3.27016979e-01 4.48276520e-01 4.18478608e-01
-5.47942162e-01 4.62725818e-01 1.58774912e+00 1.46820813e-01
-4.74934578e-01 9.30106118e-02 -3.21871638e-01 8.62179816e-01
7.57451355e-01 -3.29794079e-01 -4.24080580e-01 -6.96717441e-01
-1.11213043e-01 2.34347269e-01 5.15190661e-01 -1.12227178e+00
5.68074703e-01 -3.44363421e-01 3.10542583e-01 -2.04855636e-01
2.26649895e-01 -6.76808000e-01 -6.42213132e-03 5.51682770e-01
-1.91058040e-01 2.82545328e-01 -2.30985135e-02 3.22824895e-01
-4.70793307e-01 -1.78686440e-01 1.14417243e+00 -2.93940343e-02
-6.01968050e-01 6.04073286e-01 -2.98636705e-01 1.74637754e-02
1.07732046e+00 -2.73182124e-01 -2.77343005e-01 -4.67231363e-01
-7.19194949e-01 3.71383786e-01 6.03826165e-01 4.65614647e-01
6.36375487e-01 -1.24091613e+00 -4.38038439e-01 2.46399239e-01
-5.03142416e-01 2.04017013e-01 2.31243521e-01 1.11548436e+00
-1.12969029e+00 8.76126513e-02 -2.09161907e-01 -6.25082552e-01
-1.25486302e+00 5.61057150e-01 6.20607257e-01 -4.10426175e-03
-5.79379916e-01 7.02665031e-01 2.34439746e-01 -3.76786776e-02
2.12454274e-01 -3.64773870e-01 -1.45260260e-01 -2.03333884e-01
5.83669722e-01 6.01827800e-01 -3.12668830e-01 -7.99100578e-01
-1.72582746e-01 9.90281880e-01 1.02809794e-01 1.11606993e-01
1.22129238e+00 -5.47519207e-01 -2.52642900e-01 -5.42105641e-03
1.42698920e+00 -3.45901586e-02 -1.94238389e+00 -3.04536521e-02
-4.45088536e-01 -7.47599006e-01 3.18223275e-02 -4.93515246e-02
-1.65357542e+00 6.83951199e-01 5.37014842e-01 -2.51008626e-02
1.16418087e+00 -4.01061714e-01 8.24682772e-01 1.17327683e-01
-1.34960964e-01 -9.05417621e-01 2.83721507e-01 7.23602235e-01
6.38229191e-01 -1.38990879e+00 4.64801751e-02 -3.99466783e-01
-4.69953984e-01 1.35091949e+00 8.42107892e-01 -5.54712474e-01
7.40874231e-01 1.17613394e-02 3.56722802e-01 1.93623707e-01
-5.93919635e-01 1.99770674e-01 2.01130852e-01 1.95277825e-01
3.89500618e-01 -5.78836977e-01 -2.52209604e-01 -1.36686102e-01
-5.48071824e-02 3.24690908e-01 8.09567809e-01 8.60779583e-01
-3.47492158e-01 -9.29634273e-01 -4.62655187e-01 -2.47657582e-01
-6.67111099e-01 2.20749378e-02 2.12580159e-01 6.68428659e-01
8.05546194e-02 8.92655611e-01 6.36756644e-02 -1.27437711e-01
3.24779212e-01 -3.51688743e-01 3.06096911e-01 -2.75751740e-01
-4.15065944e-01 2.82465935e-01 -3.39584798e-01 -9.28249657e-01
-8.34437191e-01 -6.15934551e-01 -1.06125438e+00 -6.35854900e-01
-1.95094168e-01 7.56367994e-03 1.10469431e-01 6.24995768e-01
1.34461671e-01 3.08099866e-01 7.31768548e-01 -1.15024042e+00
-2.47546270e-01 -5.27038336e-01 -2.45542184e-01 5.31329513e-01
7.93756485e-01 -4.54378784e-01 -5.24349153e-01 7.98687816e-01] | [10.610638618469238, -1.3948496580123901] |
d5b363aa-4652-4d05-99da-1296a287d024 | an-application-of-topological-graph | 1408.5634 | null | http://arxiv.org/abs/1408.5634v1 | http://arxiv.org/pdf/1408.5634v1.pdf | An application of topological graph clustering to protein function prediction | We use a semisupervised learning algorithm based on a topological data
analysis approach to assign functional categories to yeast proteins using
similarity graphs. This new approach to analyzing biological networks yields
results that are as good as or better than state of the art existing
approaches. | ["Danielle O'Donnol", 'R. Sean Bowman', 'Jesse Johnson', 'Douglas Heisterkamp'] | 2014-08-24 | null | null | null | null | ['protein-function-prediction'] | ['medical'] | [ 2.35633537e-01 5.48135638e-02 -2.06230700e-01 -4.89063352e-01
3.29578787e-01 -7.18054235e-01 4.43798184e-01 6.78919494e-01
-1.29226089e-01 1.15370572e+00 -2.02242866e-01 -6.70992553e-01
-7.44362772e-01 -8.04784417e-01 -3.23387474e-01 -6.11387312e-01
-6.64899051e-01 6.89217985e-01 7.03593075e-01 -2.48240486e-01
4.93991464e-01 7.62328327e-01 -1.43676043e+00 2.08100706e-01
7.89015770e-01 4.03088510e-01 -1.53321043e-01 5.41700542e-01
-3.83935004e-01 4.30224895e-01 -4.64736551e-01 2.59315055e-02
-9.91210938e-02 -6.87912822e-01 -1.06135118e+00 -1.64768845e-02
-1.51456267e-01 7.72773802e-01 -2.07062755e-02 1.10701048e+00
-3.40592526e-02 -1.33335125e-02 1.01065433e+00 -1.39459682e+00
-5.59866905e-01 6.41633749e-01 -3.78553629e-01 4.30325001e-01
7.46881247e-01 -2.64333427e-01 9.51288700e-01 -7.48081267e-01
1.19240844e+00 1.20383883e+00 6.71311498e-01 -7.00298697e-03
-1.95661390e+00 -3.61396074e-01 -5.32766059e-02 4.40997660e-01
-1.34072626e+00 -2.26592600e-01 8.12532187e-01 -5.20533025e-01
1.00812101e+00 1.40280455e-01 6.68526590e-01 2.20269278e-01
6.62154198e-01 2.95027799e-04 1.44515991e+00 -6.14737272e-01
5.12625277e-01 -1.45430848e-01 5.59733868e-01 1.02818501e+00
5.78288317e-01 -1.98082924e-01 -4.84535068e-01 -5.33747911e-01
4.95572716e-01 -8.92660692e-02 -5.17756566e-02 -1.12306750e+00
-1.43546844e+00 1.04296780e+00 3.94537836e-01 8.37676823e-01
-1.27860561e-01 -3.36729348e-01 7.22664535e-01 6.58501863e-01
6.82829678e-01 7.28917539e-01 -7.13169038e-01 2.84886330e-01
-6.67182028e-01 -2.94382274e-01 1.06410384e+00 5.61545074e-01
7.08457172e-01 -3.00435364e-01 6.50314093e-01 7.98938394e-01
6.26834571e-01 -1.54056221e-01 4.46699053e-01 -7.89216757e-01
-5.05423725e-01 1.19884562e+00 -3.67521167e-01 -1.17760432e+00
-5.01789272e-01 1.56885594e-01 -7.04703391e-01 3.77009958e-01
6.54793978e-01 2.94319242e-01 -6.84670866e-01 1.36874688e+00
3.31304878e-01 -1.87114358e-01 2.11535528e-01 4.87293690e-01
7.95119286e-01 5.60750484e-01 1.42268121e-01 -5.97999215e-01
1.11798692e+00 -5.77980816e-01 -8.09357226e-01 7.29092300e-01
7.55036533e-01 -7.91029215e-01 8.61042500e-01 5.95724940e-01
-5.46983242e-01 -2.12842122e-01 -1.11867821e+00 3.61693144e-01
-1.25069904e+00 -2.59357635e-02 1.22184038e+00 5.20366848e-01
-1.19495702e+00 1.07556617e+00 -6.01372123e-01 -1.01933920e+00
2.45851815e-01 5.97325146e-01 -1.07156444e+00 2.22191717e-02
-9.24432695e-01 7.32963562e-01 8.25288594e-01 -5.40870786e-01
-1.90568626e-01 -3.18693280e-01 -7.73804605e-01 -1.09570794e-01
1.04879960e-01 -3.06427747e-01 4.09393877e-01 -4.50238466e-01
-1.41445971e+00 1.01218319e+00 -1.57011852e-01 -1.75158426e-01
-1.20642178e-01 6.80765390e-01 -4.24687028e-01 5.61661959e-01
2.03879789e-01 4.45959061e-01 -2.99021672e-03 -1.03278553e+00
-1.65435717e-01 -4.05480593e-01 -3.74387622e-01 -3.20485204e-01
-4.79020238e-01 8.22938308e-02 2.33200803e-01 -4.74362016e-01
5.34307599e-01 -8.31525624e-01 -8.50357711e-02 1.04877830e-01
-5.99873424e-01 -7.67698765e-01 1.03284073e+00 -3.73107374e-01
9.73182678e-01 -1.56333494e+00 6.15260124e-01 6.25031412e-01
4.08772260e-01 3.94757837e-02 9.10259411e-02 6.14969432e-01
-7.43187428e-01 2.55720526e-01 -3.39576036e-01 4.44075853e-01
-1.73790187e-01 2.99169838e-01 3.01890254e-01 7.09465504e-01
-1.80947129e-02 3.38811159e-01 -1.07609653e+00 -6.76756680e-01
2.49928564e-01 -3.64666828e-03 5.11310995e-02 8.31933841e-02
2.45433562e-02 1.87341467e-01 -1.86384857e-01 9.56636310e-01
3.87927532e-01 -4.77385581e-01 1.07571054e+00 -3.08309287e-01
-1.05556250e-02 5.16009033e-02 -7.85599828e-01 1.52057779e+00
4.19679374e-01 8.20923924e-01 -1.65016964e-01 -1.62822330e+00
1.06134558e+00 2.56986380e-01 7.43160844e-01 3.02665174e-01
-4.13805107e-03 8.63951519e-02 1.65663958e-01 -2.33432472e-01
-3.40904623e-01 -1.33233946e-02 2.71783292e-01 5.87588906e-01
6.82329834e-01 8.89186114e-02 6.04296207e-01 4.55545843e-01
1.48301530e+00 3.55619155e-02 8.20017040e-01 -1.17516780e+00
9.90926981e-01 3.03961545e-01 5.16071558e-01 -2.31379345e-02
-1.75094008e-01 -2.59161383e-01 1.11534107e+00 -5.92958570e-01
-1.08368134e+00 -1.10997558e+00 -2.65341699e-01 8.83236289e-01
-4.77323905e-02 -5.38811982e-01 -8.76389205e-01 -8.60986650e-01
3.21916997e-01 -4.88609187e-02 -5.98386526e-01 -1.50768578e-01
7.87820201e-03 -7.64629900e-01 3.93441826e-01 1.00740351e-01
6.88744336e-02 -8.73076439e-01 1.56401515e-01 1.43851489e-02
2.71114349e-01 -6.75769627e-01 -1.22669023e-02 7.32818902e-01
-1.12006366e+00 -1.75460470e+00 -2.76955426e-01 -1.13469684e+00
1.10420048e+00 2.54751325e-01 9.13782299e-01 6.07038029e-02
-5.74956059e-01 -7.64983296e-02 -2.76604176e-01 -1.52886420e-01
-5.46604276e-01 1.35596946e-01 5.26301205e-01 -4.19037849e-01
6.42488718e-01 -8.99457693e-01 -2.39703115e-02 6.41695738e-01
-8.71133268e-01 -4.53145385e-01 4.58079308e-01 8.31267715e-01
5.18582225e-01 5.74365914e-01 8.97497296e-01 -9.75340962e-01
7.98944235e-01 -5.91897368e-01 -7.05397964e-01 2.55156159e-01
-1.23891294e+00 3.02686960e-01 8.36671948e-01 -4.93853390e-01
-4.57797289e-01 5.46017706e-01 4.67978746e-01 4.84729856e-02
-4.63178962e-01 3.17921817e-01 -3.36050361e-01 -7.25322545e-01
5.95702112e-01 7.03395009e-02 5.15294373e-01 -4.31534499e-01
1.91689581e-01 5.52174866e-01 2.76329309e-01 -3.14100802e-01
6.20632529e-01 3.54252428e-01 5.44406593e-01 -7.52488673e-01
6.40824586e-02 -7.36817777e-01 -1.27924275e+00 -4.15908962e-01
6.75569177e-01 -1.22196734e-01 -1.00616813e+00 1.52777418e-01
-7.27820754e-01 2.21498638e-01 1.17073841e-01 4.74145621e-01
-6.58942997e-01 6.95515811e-01 -8.10782194e-01 -3.19816589e-01
1.59910962e-01 -7.16373503e-01 5.48407793e-01 5.89841269e-02
-4.02048141e-01 -1.40439332e+00 4.86242443e-01 1.91532329e-01
1.13446459e-01 4.86441374e-01 1.30790484e+00 -1.15903354e+00
8.20856318e-02 5.69149777e-02 -2.34689265e-01 -2.89103210e-01
4.89291251e-01 3.81727934e-01 -3.56787145e-01 -1.51325092e-01
-6.24790549e-01 -8.03700238e-02 9.08519804e-01 1.37614578e-01
8.63193154e-01 5.40406890e-02 -1.02542400e+00 -2.27234438e-02
1.36828864e+00 3.78418684e-01 3.45893592e-01 2.31120601e-01
2.82428712e-01 1.00868738e+00 7.01168060e-01 -1.34668410e-01
-6.78029656e-02 4.34327036e-01 1.28488317e-01 -2.00743720e-01
3.18195254e-01 -8.85726884e-02 3.53527889e-02 6.13986433e-01
1.18712276e-01 -3.94211896e-02 -1.16060889e+00 4.54075098e-01
-2.15501499e+00 -7.11598516e-01 -2.93870717e-01 1.74203205e+00
8.34263086e-01 2.52957612e-01 4.57887352e-01 5.63616395e-01
9.88835037e-01 -3.29865545e-01 -3.45596790e-01 -6.14992261e-01
-1.21528774e-01 5.98010467e-03 2.96570897e-01 8.55835825e-02
-1.13777614e+00 1.08479953e+00 8.76113987e+00 5.91108263e-01
-7.83403277e-01 -3.01776201e-01 3.75807434e-01 6.28600717e-01
1.62096113e-01 1.38071716e-01 -3.87039706e-02 1.80115253e-01
1.17681336e+00 -2.17051268e-01 2.39085406e-01 8.48606110e-01
1.22659028e-01 -2.47000471e-01 -9.80603755e-01 5.81743360e-01
7.74713010e-02 -1.43153453e+00 -1.49265826e-01 4.06819373e-01
3.87077302e-01 -3.91473204e-01 -5.47031820e-01 -5.31035185e-01
6.95504069e-01 -8.26896310e-01 -2.45823294e-01 8.66320133e-01
4.33357924e-01 -8.17960858e-01 8.57339323e-01 -2.84475852e-02
-1.39953685e+00 9.33741704e-02 -6.42404437e-01 -2.81272799e-01
-1.92711830e-01 1.09877539e+00 -1.02116132e+00 5.22737622e-01
6.91477299e-01 1.06871521e+00 -7.54737556e-01 9.20988560e-01
-2.42317274e-01 6.18867695e-01 -2.52498180e-01 -4.10556197e-01
-2.61256725e-01 -2.62053311e-01 4.36981082e-01 1.28376174e+00
-1.29290819e-01 -5.26581593e-02 1.70610622e-01 6.89414084e-01
1.50748610e-01 5.27313590e-01 -1.40749252e+00 -5.16080797e-01
3.02350223e-01 1.46222472e+00 -1.78052402e+00 -5.93066096e-01
-1.42860934e-01 7.98709035e-01 3.88652474e-01 -2.39626303e-01
-9.95708555e-02 -9.68757093e-01 6.46247506e-01 -2.99407225e-02
4.70715053e-02 -4.46166992e-01 1.25468615e-02 -9.51183677e-01
-5.31335294e-01 -6.12296581e-01 7.39783168e-01 -6.98881090e-01
-1.67398429e+00 1.72573835e-01 1.73570633e-01 -1.06178045e+00
-2.03086093e-01 -1.05032003e+00 -4.96898651e-01 1.84534907e-01
-9.39134240e-01 -8.25247347e-01 2.54803926e-01 3.36947948e-01
1.00582466e-01 -6.98637247e-01 1.23932278e+00 -3.48847136e-02
-4.41043556e-01 -1.40089035e-01 2.84848332e-01 -1.36997595e-01
8.56376469e-01 -1.56115830e+00 -5.86664304e-02 5.97942948e-01
1.39214456e-01 7.87404060e-01 5.61620176e-01 -6.37675345e-01
-1.36613679e+00 -6.54803455e-01 1.07640278e+00 -1.94597468e-01
1.00852835e+00 -3.42514485e-01 -1.00359416e+00 2.41180941e-01
3.94138098e-01 1.74656928e-01 1.37966585e+00 2.76302338e-01
-2.27812320e-01 8.12059492e-02 -1.40617478e+00 1.83206022e-01
1.17293644e+00 -2.64332354e-01 -7.53596783e-01 6.52994514e-01
4.30266052e-01 6.08174801e-01 -1.37778163e+00 3.76091450e-01
3.64490658e-01 -6.58527911e-01 8.74593377e-01 -1.10723400e+00
5.13268001e-02 -8.20107937e-01 9.52666998e-02 -1.33405066e+00
-7.14635074e-01 -4.41860497e-01 1.33822381e-01 1.04343951e+00
5.60395956e-01 -7.25704253e-01 7.76234925e-01 -3.35717916e-01
3.53187174e-01 -3.67848963e-01 -8.50669861e-01 -8.75142276e-01
-1.35705262e-01 4.25537854e-01 2.33607128e-01 1.71148646e+00
1.30285299e+00 6.71423137e-01 3.14156055e-01 -3.56918931e-01
8.70277405e-01 1.85818821e-01 2.55557626e-01 -1.96245706e+00
2.51979351e-01 -5.03642261e-01 -1.28129303e+00 1.47963360e-01
8.62470090e-01 -1.02141118e+00 -7.98783079e-02 -1.45830786e+00
2.88793743e-01 2.74455454e-02 -6.44243717e-01 8.12036932e-01
1.91994756e-01 5.95119298e-01 -5.73384523e-01 1.74804524e-01
-8.76834631e-01 -1.96791232e-01 5.57907641e-01 -7.72972628e-02
-8.10271651e-02 -4.17030096e-01 -5.53541362e-01 7.83928096e-01
9.68729556e-01 -4.17137563e-01 -3.19361269e-01 6.61350250e-01
-4.77273129e-02 -1.54194072e-01 2.65851822e-02 -1.10386884e+00
3.25329810e-01 -1.50668889e-01 5.99317908e-01 -5.10057747e-01
-1.24698095e-01 -7.82090306e-01 2.78951257e-01 1.02178359e+00
-4.04863834e-01 1.07546598e-01 1.15093097e-01 7.58381665e-01
-2.50037760e-01 -2.08108276e-02 8.67235839e-01 -1.49973705e-01
-6.27833486e-01 -4.51739103e-01 -1.01606810e+00 -7.68426955e-01
1.24465704e+00 -1.96081370e-01 -2.56863564e-01 -1.05950773e-01
-1.19501638e+00 3.96173354e-03 7.72474766e-01 1.39170960e-01
5.55415988e-01 -1.27444458e+00 -8.00597221e-02 -4.37046401e-02
3.57801974e-01 -1.00717270e+00 -5.20459890e-01 6.50506616e-01
-7.89504826e-01 8.58203828e-01 -8.35462153e-01 -6.13529146e-01
-1.97132969e+00 9.66387451e-01 -1.18683456e-02 -3.77955854e-01
-2.35441327e-01 2.19517693e-01 -5.13902783e-01 -1.03131735e+00
-2.63818920e-01 -5.84459677e-03 -6.47015452e-01 8.87804329e-02
7.03129172e-02 6.40277386e-01 -7.00630620e-02 -7.22398221e-01
-8.20366263e-01 6.01016641e-01 3.57332915e-01 2.53442079e-01
1.28281891e+00 1.00839278e-02 -9.82447922e-01 6.83601320e-01
1.14097881e+00 -2.61282653e-01 -2.33621135e-01 -9.65195522e-02
5.68886757e-01 -3.25496107e-01 -2.76551634e-01 -8.84680271e-01
-6.07593596e-01 3.87318820e-01 7.31800139e-01 7.05601275e-01
1.00484705e+00 2.38811493e-01 -8.51373076e-02 7.64441550e-01
3.12684625e-01 -1.08759713e+00 4.45989110e-02 2.49364987e-01
3.11650038e-01 -1.17613685e+00 3.49412024e-01 -1.10770452e+00
3.90349887e-03 1.59724414e+00 2.15483204e-01 -3.08139324e-01
9.43969727e-01 3.73063415e-01 -2.06151396e-01 -3.88681680e-01
-8.94847095e-01 -1.38028502e-01 3.00799102e-01 1.05112100e+00
8.26055765e-01 1.05987832e-01 -1.14008987e+00 1.22027677e-02
1.39436960e-01 -9.78525728e-02 3.35385740e-01 1.14730465e+00
-8.29977214e-01 -1.49127030e+00 -4.15860653e-01 5.06669462e-01
-3.65346491e-01 3.40650231e-01 -1.30635202e+00 8.07691932e-01
-1.83666989e-01 1.05031335e+00 -2.97411114e-01 -5.28471053e-01
3.11511476e-02 4.71370667e-01 2.13008508e-01 -5.06872714e-01
-5.15669733e-02 5.63895851e-02 2.93677151e-01 -5.11128306e-01
-7.84994602e-01 -4.95813519e-01 -1.54768372e+00 -4.58638608e-01
-5.97881913e-01 8.61610889e-01 7.07575083e-01 7.45470762e-01
5.43479681e-01 5.99312067e-01 6.78409278e-01 -7.35494673e-01
3.55837584e-01 -9.04138803e-01 -8.50073874e-01 2.80691445e-01
-3.91791314e-01 -1.03913271e+00 -3.85358572e-01 4.01756406e-01] | [6.573915958404541, 5.500241279602051] |
49356300-a4ff-4c68-972f-c23bef137899 | lightglue-local-feature-matching-at-light | 2306.13643 | null | https://arxiv.org/abs/2306.13643v1 | https://arxiv.org/pdf/2306.13643v1.pdf | LightGlue: Local Feature Matching at Light Speed | We introduce LightGlue, a deep neural network that learns to match local features across images. We revisit multiple design decisions of SuperGlue, the state of the art in sparse matching, and derive simple but effective improvements. Cumulatively, they make LightGlue more efficient - in terms of both memory and computation, more accurate, and much easier to train. One key property is that LightGlue is adaptive to the difficulty of the problem: the inference is much faster on image pairs that are intuitively easy to match, for example because of a larger visual overlap or limited appearance change. This opens up exciting prospects for deploying deep matchers in latency-sensitive applications like 3D reconstruction. The code and trained models are publicly available at https://github.com/cvg/LightGlue. | ['Marc Pollefeys', 'Paul-Edouard Sarlin', 'Philipp Lindenberger'] | 2023-06-23 | null | null | null | null | ['3d-reconstruction'] | ['computer-vision'] | [-2.33680472e-01 -2.16492340e-01 -2.84489691e-01 -5.10831952e-01
-6.60953701e-01 -4.17141676e-01 4.45165187e-01 1.83756799e-02
-1.82329834e-01 2.89496839e-01 2.21869692e-01 -1.46399334e-01
-1.42752808e-02 -8.39495778e-01 -1.11183679e+00 -2.87488371e-01
-1.38060823e-01 3.94359440e-01 2.10179061e-01 -9.68367234e-02
1.19434789e-01 7.52012968e-01 -1.62580597e+00 4.09821659e-01
3.29910010e-01 1.12980306e+00 3.41942579e-01 6.20120168e-01
2.84878463e-01 7.09498823e-01 -8.16709846e-02 -4.12736356e-01
6.56613231e-01 -4.80870530e-02 -8.64244819e-01 -2.10911348e-01
1.28680038e+00 -6.81828439e-01 -8.00402880e-01 9.47913826e-01
6.22096717e-01 1.73012272e-01 2.17014283e-01 -1.22030890e+00
-8.12214613e-01 1.27641723e-01 -5.47538459e-01 4.09136355e-01
3.24750185e-01 2.17633680e-01 1.15618205e+00 -1.22562790e+00
8.10120165e-01 1.13899684e+00 1.07925069e+00 3.18680495e-01
-1.12561238e+00 -6.40408576e-01 1.21871427e-01 2.76740879e-01
-1.43278372e+00 -8.60995531e-01 2.56046474e-01 -1.64117917e-01
1.14764440e+00 1.68413788e-01 6.84244275e-01 8.14358473e-01
3.33218127e-01 8.86671066e-01 8.38952303e-01 -3.12514216e-01
2.70214817e-03 -3.09967935e-01 -1.09794162e-01 9.39993918e-01
1.94419682e-01 4.81827557e-01 -7.94922054e-01 -1.53392524e-01
1.21827054e+00 3.42979848e-01 -2.19968885e-01 -6.46171331e-01
-1.29004526e+00 6.25490785e-01 8.75357866e-01 2.53066748e-01
-3.76797348e-01 6.42986059e-01 2.85077721e-01 5.41465938e-01
4.82098252e-01 4.02168632e-01 -2.76600540e-01 3.13094780e-02
-1.02662456e+00 4.85536754e-01 6.66479588e-01 1.09296417e+00
1.01594460e+00 -1.63521156e-01 5.45235015e-02 7.34860718e-01
1.02341786e-01 3.94509733e-01 9.88214910e-02 -1.43172038e+00
2.41667926e-01 2.64813781e-01 -5.47516905e-02 -1.16006529e+00
-3.63652706e-01 -5.40086567e-01 -7.91208267e-01 4.45522487e-01
1.86471194e-01 2.31652915e-01 -8.35657477e-01 1.77890015e+00
1.43424153e-01 4.34141606e-01 -4.21061933e-01 1.10369694e+00
1.06514776e+00 4.89401013e-01 -1.63382784e-01 5.15315831e-01
1.30078363e+00 -1.18940377e+00 -2.41462201e-01 -5.12282312e-01
4.89633799e-01 -1.05736351e+00 8.21396232e-01 6.42495379e-02
-1.37027597e+00 -5.62920213e-01 -1.04708612e+00 -4.25399065e-01
-1.65229678e-01 -7.07910433e-02 1.07329500e+00 -6.82957768e-02
-1.55578363e+00 8.21746826e-01 -8.35944593e-01 -5.18280864e-01
5.75210869e-01 5.29188335e-01 -6.24103069e-01 -4.48575974e-01
-9.85180438e-01 7.87111938e-01 -1.28598571e-01 -3.74963097e-02
-8.29073250e-01 -8.61643016e-01 -8.31773758e-01 1.12680413e-01
1.58639416e-01 -1.16919422e+00 1.59285069e+00 -9.76974130e-01
-1.03683817e+00 1.36585581e+00 -3.29180032e-01 -4.25343573e-01
4.09845352e-01 -2.68914789e-01 -4.09277603e-02 1.30543008e-01
2.09717393e-01 9.22968209e-01 7.23334730e-01 -9.20747459e-01
-5.51980674e-01 -2.27891013e-01 5.79987951e-02 2.30968952e-01
7.83912316e-02 6.97876746e-03 -6.74737692e-01 -6.00759685e-01
7.08222687e-02 -9.07561004e-01 -3.67427677e-01 7.73400962e-01
-3.33636813e-02 -1.60343677e-01 4.15685654e-01 -3.81689608e-01
7.35268295e-01 -2.18170047e+00 -1.43910766e-01 2.55751014e-01
7.16243088e-01 2.90495634e-01 -5.62641144e-01 6.03935361e-01
-2.43050501e-01 -2.62428343e-01 5.41889444e-02 -4.75615650e-01
-2.12867558e-02 4.17258292e-02 -2.55701035e-01 6.02258384e-01
-7.53940493e-02 1.29460073e+00 -8.30171108e-01 -2.35098198e-01
2.93479025e-01 3.82455468e-01 -5.54509044e-01 2.79913366e-01
6.87865820e-03 6.49168864e-02 -2.90481508e-01 7.81490147e-01
9.22897875e-01 -7.25545883e-01 -5.70350811e-02 -4.25361365e-01
-9.31971967e-02 4.41291541e-01 -1.07763362e+00 2.08414149e+00
-5.84038079e-01 8.54041159e-01 2.79976696e-01 -9.70250428e-01
7.62988329e-01 -4.14584298e-03 4.91316885e-01 -9.35968935e-01
-3.84241343e-02 2.90494114e-01 -4.21079040e-01 -1.69813707e-01
4.86398369e-01 1.28552943e-01 3.31556946e-01 7.00717390e-01
2.16583550e-01 -1.67666376e-01 -7.72396401e-02 3.49352986e-01
1.23638237e+00 8.02010372e-02 4.02135134e-01 -2.60429800e-01
-1.86202168e-01 -1.51141778e-01 4.40094352e-01 9.93774414e-01
-1.41636714e-01 7.31778622e-01 -1.17573693e-01 -9.80855644e-01
-1.23477781e+00 -1.29359055e+00 -7.72343650e-02 1.02304244e+00
4.02831346e-01 -4.94001538e-01 -4.16090429e-01 -2.58767754e-01
3.24705958e-01 -1.27588078e-01 -5.12928188e-01 -3.50842886e-02
-6.59882069e-01 -4.82702479e-02 2.86122859e-01 7.49348640e-01
4.25895840e-01 -9.23850894e-01 -4.23521757e-01 7.72772729e-02
1.71877574e-02 -1.09640121e+00 -6.76230669e-01 -2.58796643e-02
-9.84722316e-01 -9.62305903e-01 -6.72885239e-01 -9.77422714e-01
6.54278815e-01 8.85643899e-01 1.75093484e+00 6.57089412e-01
-5.95257878e-01 4.03669536e-01 -1.67778566e-01 -9.17544365e-02
-4.46369760e-02 8.40268582e-02 -3.79486866e-02 -4.83246386e-01
4.00489807e-01 -8.27624798e-01 -9.80290651e-01 3.66160452e-01
-6.80490613e-01 1.19023107e-01 6.85940504e-01 7.34435141e-01
7.80125856e-01 -4.82275635e-01 3.17872576e-02 -6.08284175e-01
4.17710990e-01 -4.69948351e-01 -7.46420801e-01 1.53939337e-01
-4.53978449e-01 -4.82912175e-02 3.36030930e-01 -2.47230694e-01
-4.93494540e-01 7.09759966e-02 -2.49838322e-01 -7.05359399e-01
-9.63571519e-02 2.72099853e-01 5.05148768e-01 -5.75777948e-01
6.18358612e-01 8.27771053e-02 1.24040857e-01 -4.64754432e-01
4.25636709e-01 2.08386868e-01 7.36374557e-01 -7.17003167e-01
7.15353847e-01 6.68871462e-01 1.19697034e-01 -3.97224516e-01
-7.65087843e-01 -5.31071186e-01 -3.41290981e-01 5.74438907e-02
4.58571196e-01 -1.19029653e+00 -9.40977395e-01 5.17098904e-01
-1.17025781e+00 -6.09088004e-01 -2.25709274e-01 3.55738372e-01
-6.93653464e-01 2.84614623e-01 -8.32550228e-01 -1.15791209e-01
-6.14010215e-01 -9.87652898e-01 1.22059727e+00 2.98550248e-01
-3.23648989e-01 -9.98684943e-01 2.16039658e-01 1.21471658e-01
6.35633707e-01 6.75472617e-02 5.73033571e-01 -1.94853902e-01
-1.13596272e+00 -2.60818660e-01 -6.95623577e-01 1.74759731e-01
3.38935740e-02 -1.53821846e-02 -8.33751857e-01 -6.10183895e-01
-3.83331746e-01 -5.91655076e-01 9.38791275e-01 6.09182477e-01
1.27645814e+00 -2.60690331e-01 -5.32548070e-01 1.20951176e+00
1.54064906e+00 -3.14961344e-01 5.72015643e-01 3.34613293e-01
4.85416144e-01 3.60267848e-01 4.71410215e-01 4.14179206e-01
4.42329764e-01 9.51530993e-01 5.17975032e-01 -4.96875316e-01
-4.58530873e-01 -2.84614652e-01 3.72575112e-02 7.87914276e-01
-2.21702941e-02 -4.19828892e-02 -7.80065000e-01 6.22024894e-01
-2.08256531e+00 -1.05129361e+00 1.97498143e-01 2.09217691e+00
6.50477827e-01 -2.34413415e-01 -3.88501249e-02 -5.69358528e-01
6.35776818e-01 3.15243125e-01 -4.94177967e-01 -3.71745437e-01
-2.30183244e-01 6.52002633e-01 5.92681348e-01 5.69386721e-01
-9.93909180e-01 8.57127130e-01 7.02915239e+00 8.48425567e-01
-1.15664291e+00 1.81987762e-01 6.86914146e-01 -3.61277580e-01
-3.41117501e-01 -1.83221288e-02 -5.78438699e-01 2.83842325e-01
4.57770407e-01 1.04923114e-01 6.89206004e-01 8.20162654e-01
-1.28064573e-01 -1.35422111e-01 -1.43851614e+00 1.15775549e+00
-1.21215127e-01 -2.01612973e+00 -1.13377228e-01 -1.56228272e-02
8.75816226e-01 7.24687576e-01 2.20624954e-01 -1.54612318e-01
4.68033850e-01 -1.18201053e+00 7.22606957e-01 4.96850818e-01
8.86540532e-01 -4.93180633e-01 6.49742007e-01 7.03349486e-02
-1.33561230e+00 2.61580586e-01 -6.73027873e-01 -2.06204295e-01
-8.19171034e-03 6.17690504e-01 -3.97981972e-01 2.33737767e-01
9.83051956e-01 8.82082701e-01 -4.67508852e-01 1.25120521e+00
-3.91111188e-02 1.42847255e-01 -5.37277758e-01 1.89883649e-01
2.71853626e-01 -5.84474243e-02 3.41252357e-01 1.17551982e+00
4.80855495e-01 -1.35867251e-02 1.80641651e-01 9.90201950e-01
-3.43563408e-01 -1.19364195e-01 -8.88656616e-01 2.84837276e-01
8.78526568e-01 1.19578671e+00 -4.58156228e-01 -2.41172612e-01
-6.35780275e-01 1.18022215e+00 7.61460483e-01 2.56326258e-01
-6.31183207e-01 -2.58641422e-01 1.00715911e+00 1.48725301e-01
4.37250853e-01 -2.83581078e-01 -2.95119107e-01 -1.15340781e+00
2.12642923e-01 -9.45478320e-01 3.75213385e-01 -9.44465458e-01
-1.44919658e+00 5.76942444e-01 -2.11443856e-01 -1.17145324e+00
-1.34634852e-01 -7.14508176e-01 -7.97381639e-01 6.89318419e-01
-1.53614700e+00 -1.18012297e+00 -5.90935409e-01 8.23209405e-01
3.04325044e-01 2.64942534e-02 8.00983608e-01 5.51694036e-01
-1.51062697e-01 9.01303887e-01 3.72127630e-02 4.93256412e-02
8.73963833e-01 -8.72341633e-01 1.13118315e+00 8.21522832e-01
4.33787674e-01 8.53122592e-01 5.04568756e-01 -2.93268323e-01
-1.53623486e+00 -7.99520373e-01 9.82551277e-01 -4.08688903e-01
6.44332528e-01 -4.17996258e-01 -6.32541180e-01 8.01394343e-01
2.39760265e-01 5.19207120e-01 4.33980078e-01 3.18860412e-01
-7.50292838e-01 -3.19707453e-01 -9.84482110e-01 5.63324928e-01
1.43918204e+00 -8.49890947e-01 -3.21720749e-01 4.40133035e-01
5.33814073e-01 -7.48723626e-01 -8.94406974e-01 3.21745157e-01
7.72169650e-01 -1.45859122e+00 1.36244166e+00 -4.09386903e-01
4.34325367e-01 -1.17414489e-01 -3.72697294e-01 -1.06859899e+00
-7.88576663e-01 -7.45240092e-01 -1.77783996e-01 6.24603152e-01
2.18144923e-01 -7.32973576e-01 1.02217281e+00 4.39169288e-01
-3.21177691e-01 -1.07715249e+00 -9.28087711e-01 -8.43830824e-01
-1.92102641e-01 -1.72643378e-01 7.65952706e-01 9.73380566e-01
-2.60804892e-01 6.80614216e-03 -5.20692706e-01 1.21188127e-01
6.59974158e-01 5.55703044e-01 8.98800075e-01 -9.84413683e-01
-5.20785928e-01 -4.21378821e-01 -5.19065380e-01 -1.59175837e+00
4.73379903e-02 -8.88761401e-01 -1.25221848e-01 -1.46274710e+00
3.49159509e-01 -7.59660244e-01 -1.84509933e-01 7.83138990e-01
-1.13061117e-02 7.15445936e-01 3.15455377e-01 3.87395054e-01
-7.09133446e-01 2.29372978e-01 1.35340250e+00 -9.19587091e-02
2.14235008e-01 1.92721537e-03 -6.07515991e-01 4.02559698e-01
7.61750817e-01 -3.89104515e-01 -1.25067353e-01 -9.81361866e-01
4.03182089e-01 -7.37386346e-02 6.43403172e-01 -9.28506017e-01
4.10345167e-01 3.97838540e-02 4.85162467e-01 -4.52212870e-01
4.72076595e-01 -6.85929239e-01 3.03985149e-01 4.46737736e-01
-2.37397105e-01 5.45543373e-01 3.28582764e-01 1.44657999e-01
-2.16285050e-01 -9.61599573e-02 7.77508855e-01 -3.45534980e-01
-1.09301460e+00 9.24141109e-01 -4.39531431e-02 9.38455239e-02
7.14985728e-01 -3.68627042e-01 -4.90039319e-01 -6.06680930e-01
-3.63780826e-01 -1.89022142e-02 8.17222416e-01 3.37161690e-01
6.91915691e-01 -1.69834244e+00 -6.98556423e-01 3.12674463e-01
1.97996542e-01 -1.42571032e-02 2.70992666e-01 7.77177691e-01
-7.39395022e-01 3.57823104e-01 -4.20810431e-01 -6.50771260e-01
-1.22936881e+00 4.90116030e-01 4.83518958e-01 -7.83905461e-02
-7.92100549e-01 1.29921615e+00 3.72718871e-01 -3.49139273e-01
2.82299817e-01 -2.87922323e-01 6.72837079e-01 -4.81099844e-01
5.52056968e-01 2.72679865e-01 2.21807137e-01 -2.93394059e-01
-4.94698584e-01 8.28286052e-01 -1.26066253e-01 3.88428777e-01
1.39602947e+00 -1.23918943e-01 -3.36095959e-01 -2.12699190e-01
1.40835416e+00 -1.47907898e-01 -1.40517676e+00 -6.11808240e-01
-4.73004192e-01 -8.00153852e-01 2.84233451e-01 -5.15380144e-01
-1.35470378e+00 7.22166538e-01 6.84873700e-01 -1.17087252e-01
1.02548039e+00 2.47551411e-01 1.10687494e+00 4.80575144e-01
4.73752797e-01 -7.83175349e-01 1.64002925e-01 4.60384130e-01
8.63685966e-01 -1.33037627e+00 2.04060286e-01 -2.49269992e-01
-4.98191193e-02 1.08779120e+00 6.44595802e-01 -5.45816898e-01
6.68750107e-01 5.51019073e-01 1.70270130e-01 -5.01093388e-01
-7.89923370e-01 -7.12785870e-02 3.60766143e-01 7.26576030e-01
4.68367994e-01 -1.61835879e-01 2.36351147e-01 -1.25078753e-01
-1.44009544e-02 6.10882044e-02 -3.04492395e-02 7.95364559e-01
-3.19333911e-01 -1.14049196e+00 -1.30712330e-01 4.26594973e-01
-1.57629892e-01 -3.48044604e-01 -2.37259507e-01 6.24616265e-01
2.06749979e-02 4.79519904e-01 3.79654676e-01 -3.13505024e-01
1.60917372e-01 -6.19447052e-01 8.23553562e-01 -5.14785230e-01
-4.91635919e-01 -3.63331378e-01 6.34527067e-03 -1.26049066e+00
-4.82410848e-01 -3.67928356e-01 -9.26393509e-01 -1.01324868e+00
-7.46710524e-02 -2.28281468e-01 3.39600712e-01 7.25617409e-01
8.26252878e-01 -6.04906939e-02 6.06240392e-01 -1.21494687e+00
-4.11311924e-01 -3.85157913e-01 -2.64846474e-01 4.04278517e-01
5.20519137e-01 -4.92663443e-01 -7.68318772e-02 -2.52394915e-01] | [8.188416481018066, -1.999177098274231] |
7929b2e0-0760-41f3-9445-6872db177cb4 | identifying-therapist-conversational-actions | null | null | https://aclanthology.org/W19-3002 | https://aclanthology.org/W19-3002.pdf | Identifying therapist conversational actions across diverse psychotherapeutic approaches | While conversation in therapy sessions can vary widely in both topic and style, an understanding of the underlying techniques used by therapists can provide valuable insights into how therapists best help clients of different types. Dialogue act classification aims to identify the conversational {``}action{''} each speaker takes at each utterance, such as sympathizing, problem-solving or assumption checking. We propose to apply dialogue act classification to therapy transcripts, using a therapy-specific labeling scheme, in order to gain a high-level understanding of the flow of conversation in therapy sessions. We present a novel annotation scheme that spans multiple psychotherapeutic approaches, apply it to a large and diverse corpus of psychotherapy transcripts, and present and discuss classification results obtained using both SVM and neural network-based models. The results indicate that identifying the structure and flow of therapeutic actions is an obtainable goal, opening up the opportunity in the future to provide therapeutic recommendations tailored to specific client situations. | ['Fei-Tzin Lee', 'Jacob Levine', 'Kathy McKeown', 'Derrick Hull', 'Bonnie Ray'] | 2019-06-01 | null | null | null | ws-2019-6 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 6.87706649e-01 6.18438900e-01 -6.48386061e-01 -7.84166038e-01
-5.20396352e-01 -6.31780446e-01 5.41750431e-01 3.64104033e-01
-7.87231773e-02 5.70249498e-01 1.15304613e+00 -2.14116558e-01
-4.10257548e-01 -1.88966289e-01 5.15297413e-01 -7.48186052e-01
2.41593212e-01 8.79954576e-01 -4.26649481e-01 -2.85713494e-01
5.98973691e-01 7.79798090e-01 -8.54710102e-01 1.17247450e+00
4.32446629e-01 3.32395196e-01 -3.17047350e-02 4.29316163e-01
-5.08094549e-01 1.55763423e+00 -6.08160138e-01 -5.39810181e-01
-3.81890863e-01 -1.04778588e+00 -1.72701406e+00 4.89001125e-01
5.38952707e-04 -6.04860127e-01 -4.30323809e-01 5.40856898e-01
2.83226490e-01 1.74431711e-01 7.13257790e-01 -1.02063060e+00
-2.60757282e-02 1.05378151e+00 1.81935970e-02 5.21987796e-01
7.01572359e-01 -2.58907247e-02 8.27301979e-01 7.62732700e-02
8.23726892e-01 1.32714188e+00 5.29280722e-01 1.22379982e+00
-1.28402758e+00 -3.70438457e-01 -3.44900414e-02 4.16875154e-01
-3.22351396e-01 -9.22703743e-01 7.57989705e-01 -6.70997083e-01
9.46475029e-01 4.02835935e-01 7.87156403e-01 1.54338801e+00
-1.89715162e-01 9.28747475e-01 1.16757655e+00 -3.55105281e-01
1.48879111e-01 3.21648836e-01 3.37246209e-01 2.37131953e-01
-7.33990073e-01 -5.81616640e-01 -5.09798348e-01 -5.90023577e-01
6.50658369e-01 -2.03194126e-01 -4.53720808e-01 9.44723785e-02
-7.38527894e-01 1.03750205e+00 6.54728040e-02 8.30661952e-01
-5.39759576e-01 -2.67856389e-01 1.07266915e+00 3.87423843e-01
5.51581085e-01 5.92436314e-01 -1.93831235e-01 -1.07083404e+00
-5.20687580e-01 2.04616919e-01 1.27603126e+00 3.76138806e-01
1.81960333e-02 -2.15946093e-01 -4.12856460e-01 1.44791651e+00
-6.28631935e-02 -4.83522981e-01 4.13368225e-01 -1.45623040e+00
3.70792091e-01 5.80884516e-01 -1.03033535e-01 -5.68407357e-01
-6.22641861e-01 8.74207616e-02 -2.67233133e-01 -1.04023644e-03
6.75481677e-01 -2.37690717e-01 -2.33671993e-01 1.51983893e+00
3.13689381e-01 -2.01446950e-01 1.32454466e-02 4.68032628e-01
7.26469338e-01 5.96246161e-02 3.81463170e-01 -8.65218341e-01
1.27476704e+00 -1.01455891e+00 -1.13638258e+00 -2.13280231e-01
1.24254155e+00 -6.88404679e-01 4.87243891e-01 2.60328680e-01
-1.32192481e+00 2.35505909e-01 -2.04969436e-01 -1.09774388e-01
4.18680727e-01 -1.78313524e-01 7.31090128e-01 7.21709073e-01
-6.99102283e-01 8.58281791e-01 -8.47605050e-01 -8.51203144e-01
6.34594917e-01 3.64777803e-01 -5.90304017e-01 1.86333299e-01
-7.01424181e-01 1.41026533e+00 1.79886207e-01 6.43070787e-02
-2.50463098e-01 -6.50470197e-01 -5.52012503e-01 2.46472403e-01
2.00960889e-01 -1.62333116e-01 1.80350769e+00 -1.15547025e+00
-1.64721954e+00 1.24070334e+00 -4.34420675e-01 -1.80555731e-01
4.91316378e-01 2.41198838e-01 4.17194702e-02 4.64896917e-01
-1.37042403e-01 2.06978768e-01 3.84181410e-01 -9.53331590e-01
-6.40283406e-01 -5.64510465e-01 2.14212865e-01 6.79607749e-01
-4.07610506e-01 7.22478092e-01 1.78219527e-01 -2.20703408e-01
2.67580658e-01 -8.58264208e-01 6.43898621e-02 9.16350540e-03
-2.34084725e-01 -6.18265092e-01 5.14527261e-01 -7.64787257e-01
1.23975277e+00 -2.00983763e+00 6.72717988e-01 -2.37684995e-01
4.74939525e-01 1.54597700e-01 4.48421448e-01 1.28616953e+00
-3.09339106e-01 -3.78057733e-02 -1.05778299e-01 -3.97178531e-01
-4.31033112e-02 4.80032086e-01 -1.65178031e-01 5.40179849e-01
-2.39236087e-01 5.36385298e-01 -8.58804762e-01 -5.47069550e-01
4.33222175e-01 1.79381162e-01 -4.17921126e-01 5.20818055e-01
7.04019889e-02 8.27401102e-01 -6.96587503e-01 1.94851622e-01
6.32092804e-02 -4.02129814e-02 8.73896360e-01 6.95883930e-02
2.10094407e-01 7.71260202e-01 -3.02585691e-01 1.24675679e+00
-3.65775466e-01 8.76700342e-01 4.94841009e-01 -1.33419347e+00
5.56759119e-01 8.22391033e-01 9.74136531e-01 -1.12797223e-01
4.30357844e-01 -4.06422354e-02 5.66740692e-01 -1.07863867e+00
3.88189517e-02 -6.07947767e-01 9.02314186e-02 9.31288302e-01
-1.95407141e-02 -5.61319515e-02 -5.32115810e-02 2.05468863e-01
1.37057161e+00 -3.79870832e-01 3.58671308e-01 5.19096702e-02
3.71979773e-01 1.36136547e-01 2.95105606e-01 6.27737820e-01
-6.62768245e-01 7.77645409e-02 1.05156744e+00 -3.42103451e-01
-9.12070096e-01 -3.41506273e-01 -3.59200031e-01 1.28011656e+00
-4.41618502e-01 -1.21288538e-01 -1.01137924e+00 -5.36936700e-01
-1.58348367e-01 1.06528223e+00 -7.66176343e-01 -1.55451030e-01
-7.75339723e-01 -3.53335023e-01 5.52915096e-01 4.64445293e-01
1.99521482e-01 -1.62270081e+00 -2.91754991e-01 4.99327242e-01
-5.40117741e-01 -1.24175847e+00 -7.92247728e-02 2.26585194e-01
-1.03805363e+00 -1.12292218e+00 -1.85648978e-01 -7.78026342e-01
3.79105985e-01 8.43646452e-02 5.25368094e-01 1.15166210e-01
-8.10428336e-02 5.43028355e-01 -6.95171416e-01 -2.98983343e-02
-1.05071867e+00 2.45343037e-02 -1.98935062e-01 -1.72610879e-01
5.77360392e-01 -8.19198847e-01 -2.05151379e-01 2.39466876e-01
-5.28053343e-01 2.68566579e-01 -2.11782619e-01 8.67804229e-01
-6.07314885e-01 -2.79994935e-01 5.23965776e-01 -1.26397789e+00
1.18842685e+00 -5.79810321e-01 5.48621714e-01 1.61975965e-01
-1.82169303e-01 -4.37885612e-01 3.19386095e-01 -4.10529971e-01
-1.39856577e+00 -3.59693587e-01 -4.79232639e-01 -6.32219091e-02
-4.38793719e-01 3.00432444e-01 2.93222159e-01 -5.41988797e-02
8.10404420e-01 -1.89257320e-02 2.60997444e-01 -5.45191884e-01
5.04222885e-02 1.08858418e+00 3.44767213e-01 -5.24215698e-01
8.12783837e-02 4.31941897e-01 -4.32705730e-01 -7.59720743e-01
-9.60237503e-01 -6.51844800e-01 -6.17915392e-01 -5.51655471e-01
8.22322726e-01 -2.70223916e-01 -1.51669252e+00 2.80738741e-01
-1.06678104e+00 -5.64776123e-01 -7.33646154e-02 5.17878354e-01
-1.12337422e+00 3.32079589e-01 -1.16203332e+00 -1.16454923e+00
-3.58379036e-01 -1.09182632e+00 7.32388496e-01 2.55237848e-01
-1.15705335e+00 -1.41319239e+00 9.67616960e-02 1.23576319e+00
1.30352318e-01 2.81650983e-02 1.16858816e+00 -1.32147110e+00
2.89114058e-01 -1.31495565e-01 6.83028549e-02 9.94361937e-02
6.00440085e-01 -2.00789884e-01 -9.80744779e-01 -4.81976308e-02
6.50266945e-01 -7.86872447e-01 1.34921432e-01 2.19938874e-01
1.23447919e+00 -5.00730813e-01 -3.27036738e-01 6.52950257e-02
6.26688778e-01 5.79087198e-01 5.63003302e-01 1.90099508e-01
4.26010340e-01 1.28818643e+00 4.79175746e-01 4.79094416e-01
1.20740362e-01 9.24979150e-01 7.99402446e-02 3.27816188e-01
1.63991213e-01 2.18441593e-03 1.26158774e-01 4.02112365e-01
-1.04470119e-01 -1.51629746e-01 -8.83545697e-01 3.64077836e-01
-1.89596522e+00 -1.53479612e+00 -1.12190723e-01 1.46515727e+00
1.13547099e+00 4.55762260e-02 3.67304385e-01 2.34485045e-01
5.56053102e-01 1.79081857e-01 -4.40594018e-01 -1.02085328e+00
5.11072636e-01 2.77617782e-01 -9.30881873e-02 5.19761503e-01
-6.31215930e-01 9.89119232e-01 7.09464741e+00 6.68927789e-01
-9.70640421e-01 6.88809678e-02 4.34762895e-01 -1.50032654e-01
6.29071444e-02 -1.68143585e-01 -2.94143438e-01 2.70847261e-01
9.81665432e-01 -3.39182347e-01 6.81234241e-01 7.25765586e-01
7.72977471e-01 -1.64867789e-01 -1.38268590e+00 7.53994703e-01
-2.74112094e-02 -1.42353523e+00 -6.06648207e-01 5.96751869e-02
5.23333028e-02 -3.46426070e-01 -5.77929497e-01 5.84332235e-02
3.75722587e-01 -1.02891505e+00 3.12554534e-03 1.06089905e-01
1.78233802e-01 -1.80243045e-01 6.02110088e-01 5.76837361e-01
-3.94876361e-01 -3.67335290e-01 1.98531687e-01 -2.43325591e-01
5.05916476e-01 -4.32336450e-01 -1.42238724e+00 1.51543349e-01
3.56286645e-01 7.25412607e-01 3.22898626e-01 7.15115368e-01
-1.87820256e-01 7.63616621e-01 2.30498165e-02 2.24935580e-02
1.41607478e-01 -2.12961391e-01 5.32401741e-01 1.11460447e+00
-3.89310032e-01 9.24822330e-01 1.39172822e-01 3.32447499e-01
1.17020540e-01 2.58001089e-01 -3.37549090e-01 -4.83442813e-01
5.59365809e-01 1.34724450e+00 -7.26816237e-01 -2.44814321e-01
-1.21729439e-02 9.41680908e-01 5.77930033e-01 1.84868965e-02
-3.35542440e-01 4.65652466e-01 8.75738502e-01 4.83170301e-02
-2.81492889e-01 2.40705505e-01 -5.42414069e-01 -9.05624568e-01
-3.15089315e-01 -1.23223531e+00 4.57876146e-01 -8.51424634e-01
-1.25798869e+00 3.71944100e-01 1.95265323e-01 -7.73315251e-01
-5.89772105e-01 -4.16284412e-01 -9.04293060e-01 8.05654466e-01
-4.30218130e-01 -1.00545180e+00 -2.45905459e-01 2.92000294e-01
9.97160673e-01 -7.43073896e-02 1.09834301e+00 2.63992976e-02
-7.46930957e-01 3.45956564e-01 6.88821152e-02 2.42140964e-01
6.30164325e-01 -9.60065603e-01 -3.18072408e-01 -3.45542639e-01
-2.63895571e-01 5.26910126e-01 1.14969945e+00 -4.80822831e-01
-9.96200383e-01 -1.80332959e-01 1.00321198e+00 -4.24271494e-01
9.90243793e-01 3.25780623e-02 -1.25572944e+00 9.42015231e-01
3.65336627e-01 -9.31142747e-01 1.25493634e+00 4.31483924e-01
7.43651167e-02 3.09756577e-01 -1.46028590e+00 6.18419170e-01
1.22133994e+00 -8.08807075e-01 -1.00525546e+00 1.03745127e+00
2.08770007e-01 -6.30210340e-01 -1.25496113e+00 -8.04737285e-02
6.18744671e-01 -9.56746817e-01 5.06593943e-01 -1.35263395e+00
8.61947238e-01 8.67266297e-01 2.67890722e-01 -1.19588566e+00
-2.41100371e-01 -7.44114399e-01 3.71438384e-01 1.18170512e+00
-4.03609872e-03 -5.12599766e-01 1.03588665e+00 1.37251782e+00
-3.11128139e-01 -8.22709620e-01 -8.53700459e-01 3.99719849e-02
1.89912692e-01 -3.62378627e-01 1.67983204e-01 1.40176535e+00
1.36512256e+00 3.17593902e-01 -4.24254775e-01 -5.46013892e-01
1.65168434e-01 -2.34680772e-01 4.66415167e-01 -1.07456374e+00
-2.50926763e-01 -7.13368177e-01 -3.58962625e-01 -7.35409379e-01
5.86572766e-01 -8.83563042e-01 -1.74954578e-01 -1.59395611e+00
3.30036163e-01 -2.59425849e-01 4.19689089e-01 8.82268310e-01
1.49577528e-01 -2.73378909e-01 1.41299516e-01 4.12132293e-01
1.86045781e-01 5.16420603e-01 1.12870896e+00 1.03703596e-01
-5.01190603e-01 3.55342835e-01 -8.90555322e-01 8.71499240e-01
7.39711404e-01 -5.00388205e-01 -4.78329629e-01 -2.70356238e-02
-4.69279021e-01 8.37957799e-01 -1.05576821e-01 -1.70732275e-01
2.75540650e-01 -7.06565440e-01 -1.22799978e-01 -3.68811190e-02
6.97484910e-01 -6.38044655e-01 3.67355458e-02 4.12854820e-01
-1.07839787e+00 -5.88094890e-01 2.48405002e-02 1.28976077e-01
9.92368907e-02 -8.07579100e-01 8.52381349e-01 -1.57641143e-01
-9.12430137e-02 -2.72485256e-01 -9.39430892e-01 -2.39441991e-01
1.15987539e+00 -4.16349471e-01 -3.82083446e-01 -9.96846855e-01
-1.44239557e+00 2.12693080e-01 2.59391069e-01 3.64550561e-01
1.19358733e-01 -9.77184057e-01 -5.25085092e-01 -4.92449641e-01
3.84975858e-02 -6.46571279e-01 5.03755271e-01 1.01669288e+00
-4.96638149e-01 3.59819114e-01 -4.86203253e-01 -3.17943305e-01
-1.86192501e+00 -1.94010185e-03 6.95110559e-01 -1.55925706e-01
-8.60178709e-01 6.66337013e-01 2.08406020e-02 -1.76383927e-01
3.64263564e-01 3.96171242e-01 -8.11699808e-01 3.65020245e-01
6.00361705e-01 3.93489897e-01 -1.63522616e-01 -7.83866525e-01
-2.19952017e-02 -1.02031969e-01 -5.41368783e-01 -1.78320065e-01
1.41248536e+00 -1.31409690e-01 -3.70379329e-01 4.78187770e-01
1.02604175e+00 -5.42455196e-01 -6.86749697e-01 -2.19948128e-01
1.87303275e-01 -6.26884639e-01 -1.43943414e-01 -7.26254761e-01
-6.31171405e-01 7.99003243e-01 8.25705379e-02 6.55289948e-01
9.73392606e-01 3.28838646e-01 3.99737567e-01 2.79167444e-01
1.61379144e-01 -9.52992857e-01 5.63610420e-02 2.78449178e-01
1.04370868e+00 -9.61663663e-01 -5.56067340e-02 -8.01531076e-01
-1.00523460e+00 1.37237144e+00 5.89205861e-01 3.17286521e-01
3.17280263e-01 -2.27887314e-02 1.01013012e-01 -5.02980232e-01
-9.76115942e-01 1.27338573e-01 -2.42083848e-01 4.34027344e-01
7.95773983e-01 -1.19273923e-01 -8.48666906e-01 8.44408423e-02
-1.77161857e-01 -2.83877067e-02 6.57340288e-01 9.91529107e-01
-4.34528977e-01 -1.56999266e+00 -1.78079501e-01 8.88281226e-01
-4.73628700e-01 2.89039403e-01 -1.10432708e+00 4.35503453e-01
-3.20490927e-01 1.28420913e+00 -4.97943722e-02 -1.32412732e-01
2.92517990e-01 5.22293150e-01 6.38000786e-01 -1.00897992e+00
-1.32437336e+00 1.59731016e-01 9.52673137e-01 -4.02278274e-01
-6.73111677e-01 -1.15505898e+00 -1.19885945e+00 -7.27897584e-01
-1.04582429e-01 2.71595001e-01 4.12997663e-01 1.57000387e+00
-2.64887393e-01 3.63496065e-01 5.59337199e-01 -6.12342656e-01
-6.13354802e-01 -1.37728643e+00 -4.66662794e-01 7.59108484e-01
-3.09282057e-02 -3.92637432e-01 -3.19956809e-01 -2.07085256e-02] | [12.798222541809082, 8.009910583496094] |
18d67173-c4ed-410c-af22-333cc505aed7 | stglow-a-flow-based-generative-framework-with | 2211.11220 | null | https://arxiv.org/abs/2211.11220v2 | https://arxiv.org/pdf/2211.11220v2.pdf | STGlow: A Flow-based Generative Framework with Dual Graphormer for Pedestrian Trajectory Prediction | Pedestrian trajectory prediction task is an essential component of intelligent systems, and its applications include but are not limited to autonomous driving, robot navigation, and anomaly detection of monitoring systems. Due to the diversity of motion behaviors and the complex social interactions among pedestrians, accurately forecasting the future trajectory of pedestrians is challenging. Existing approaches commonly adopt GANs or CVAEs to generate diverse trajectories. However, GAN-based methods do not directly model data in a latent space, which makes them fail to have full support over the underlying data distribution; CVAE-based methods optimize a lower bound on the log-likelihood of observations, causing the learned distribution to deviate from the underlying distribution. The above limitations make existing approaches often generate highly biased or unnatural trajectories. In this paper, we propose a novel generative flow based framework with dual graphormer for pedestrian trajectory prediction (STGlow). Different from previous approaches, our method can more accurately model the underlying data distribution by optimizing the exact log-likelihood of motion behaviors. Besides, our method has clear physical meanings to simulate the evolution of human motion behaviors, where the forward process of the flow gradually degrades the complex motion behavior into a simple behavior, while its reverse process represents the evolution of a simple behavior to the complex motion behavior. Further, we introduce a dual graphormer combining with the graph structure to more adequately model the temporal dependencies and the mutual spatial interactions. Experimental results on several benchmarks demonstrate that our method achieves much better performance compared to previous state-of-the-art approaches. | ['Xia Li', 'Jiantao Zhou', 'Yuanman Li', 'Rongqin Liang'] | 2022-11-21 | null | null | null | null | ['robot-navigation'] | ['robots'] | [-3.58044684e-01 -2.62687117e-01 -1.68824032e-01 -2.89692342e-01
4.47472930e-02 -3.05944562e-01 6.46360040e-01 -4.17663813e-01
-3.71165201e-02 8.12286019e-01 3.02055240e-01 -3.59546959e-01
2.71786064e-01 -1.16650784e+00 -6.71334743e-01 -9.19978738e-01
6.43226802e-02 3.99584442e-01 5.63453972e-01 -3.43795896e-01
-2.48668537e-01 2.51555175e-01 -1.38558841e+00 -8.77616554e-02
1.11195886e+00 5.42168200e-01 1.20042682e-01 7.38831878e-01
-1.58684522e-01 8.56827438e-01 -4.49413985e-01 -4.86006796e-01
-9.81218740e-03 -7.35025048e-01 -2.44948372e-01 1.55991614e-01
-6.90438449e-02 -4.16636258e-01 -9.71680284e-01 9.81850982e-01
1.86984807e-01 3.73087287e-01 8.39057863e-01 -1.85340345e+00
-7.76715994e-01 2.58618712e-01 -4.73925471e-01 1.42315060e-01
3.66028875e-01 6.54911339e-01 4.50569868e-01 -4.02808577e-01
2.58607000e-01 1.34590375e+00 7.03600347e-01 7.73546457e-01
-1.00640845e+00 -6.71768367e-01 6.02662325e-01 5.26882410e-01
-1.26203763e+00 -9.79331583e-02 7.93740988e-01 -5.27989089e-01
4.65758443e-01 2.35424697e-01 1.18913019e+00 1.42990971e+00
2.75276572e-01 1.09303606e+00 7.04584658e-01 2.78026134e-01
2.51702249e-01 -1.65624842e-01 4.99186432e-03 6.44262969e-01
2.67899126e-01 2.70045489e-01 -1.51088670e-01 -1.16877899e-01
7.08517849e-01 3.36265981e-01 -2.05154762e-01 -3.30300301e-01
-1.05377519e+00 6.94001317e-01 4.60289091e-01 -8.59652311e-02
-3.19001377e-01 3.22429448e-01 6.49041384e-02 -6.30792752e-02
1.15963466e-01 -3.98577720e-01 8.99944678e-02 -5.38694501e-01
-7.88392305e-01 7.00630903e-01 6.84253216e-01 1.26541829e+00
7.45529711e-01 3.08592319e-01 -3.72479975e-01 4.28151488e-01
4.78250623e-01 6.96064413e-01 4.13495123e-01 -8.30177307e-01
5.16637623e-01 6.06804490e-01 2.97354013e-01 -1.14765418e+00
-1.84041843e-01 -2.27756917e-01 -1.28169131e+00 -1.82598289e-02
6.63778722e-01 -4.10123259e-01 -9.29948568e-01 1.75905442e+00
3.34385812e-01 5.84022701e-01 -6.47443831e-02 9.63195503e-01
6.51435614e-01 9.57178950e-01 1.15358487e-01 1.91896316e-02
7.56624401e-01 -1.04757464e+00 -7.09755778e-01 -1.85920447e-01
4.32812721e-01 -3.32191527e-01 1.06183076e+00 -2.45006252e-02
-7.43087947e-01 -6.90964818e-01 -7.46455908e-01 3.29857558e-01
-1.11068413e-01 -7.15280250e-02 5.13999403e-01 7.58843422e-01
-8.75460625e-01 4.43310380e-01 -1.12464058e+00 -2.01445624e-01
5.25625706e-01 -2.48314794e-02 1.92020200e-02 -2.24329501e-01
-1.08616281e+00 5.33196092e-01 2.02997893e-01 2.41356567e-01
-6.84623003e-01 -6.32189870e-01 -8.34016562e-01 -5.63680790e-02
1.49804384e-01 -1.00836551e+00 1.06112266e+00 -5.06060898e-01
-1.52684367e+00 4.55600582e-02 -5.61342895e-01 -3.97955537e-01
9.56205070e-01 -1.00838251e-01 -6.09732926e-01 -3.80027175e-01
9.48488861e-02 6.17182612e-01 6.49908602e-01 -1.16429317e+00
-7.59112954e-01 7.29022129e-03 -1.14365853e-01 1.20178409e-01
-1.77365139e-01 -6.33094907e-01 -6.11348629e-01 -6.51277781e-01
-2.48943120e-01 -1.14686608e+00 -5.54757714e-01 -1.43723786e-01
-6.02068961e-01 -1.69738829e-01 1.28191447e+00 -3.51446450e-01
1.43985212e+00 -2.10184908e+00 5.39216399e-02 1.13972843e-01
2.72736728e-01 1.92543432e-01 1.25043347e-01 4.39672619e-01
4.42664087e-01 -1.29718669e-02 -3.05134773e-01 -2.44151115e-01
1.94934845e-01 5.40029585e-01 -4.36424822e-01 3.85294557e-01
-8.07233248e-03 1.19839728e+00 -1.25083542e+00 -3.11483771e-01
3.75328094e-01 4.63152796e-01 -5.60373008e-01 2.16012985e-01
-1.44413874e-01 8.74362230e-01 -6.53518856e-01 3.22713941e-01
8.17765474e-01 -2.12006614e-01 -2.05662727e-01 7.40039945e-02
-7.64891058e-02 -1.26087576e-01 -1.06163228e+00 1.21552551e+00
-1.06983095e-01 6.21199310e-01 -4.64070022e-01 -7.75648594e-01
8.98805022e-01 4.26060893e-02 6.47679627e-01 -5.38954020e-01
-3.90537493e-02 -1.35628074e-01 1.13378592e-01 -5.53502083e-01
6.33710384e-01 1.33844063e-01 -1.31239533e-01 2.69299328e-01
-4.18240368e-01 1.39577776e-01 2.37833872e-01 1.76948905e-01
1.02565622e+00 2.36529022e-01 -1.34969413e-01 4.43718769e-02
4.87405092e-01 8.95022377e-02 9.09261227e-01 6.45413339e-01
-3.60525280e-01 5.70423722e-01 4.67359751e-01 -6.06474817e-01
-1.05734706e+00 -1.29459858e+00 2.06611678e-01 3.54440778e-01
5.82817674e-01 -4.16312665e-01 -8.05507720e-01 -8.10455322e-01
-4.30412889e-02 9.35005128e-01 -4.39206213e-01 -4.80486006e-01
-7.78837562e-01 -1.12673247e+00 5.35786092e-01 5.55123091e-01
9.28014576e-01 -9.37938988e-01 -4.40651625e-01 3.18672091e-01
-4.25462693e-01 -1.18609071e+00 -6.61510825e-01 -8.25756848e-01
-5.59455276e-01 -9.66113269e-01 -9.39567327e-01 -4.58468556e-01
6.56083405e-01 3.60206038e-01 8.35259736e-01 2.23282259e-02
2.31217463e-02 4.17513877e-01 -3.28228921e-01 -2.19644353e-01
-4.94965404e-01 3.48936319e-02 -4.71637100e-02 3.41305286e-01
4.67326015e-01 -6.47161067e-01 -9.62234318e-01 6.20254695e-01
-7.08936572e-01 2.53568351e-01 2.79066294e-01 7.54963636e-01
5.02520621e-01 3.23785186e-01 4.99074876e-01 -6.57202423e-01
5.76695561e-01 -8.53993356e-01 -3.39580208e-01 4.66467068e-02
-4.70025837e-01 -1.13634013e-01 9.34040010e-01 -7.26380348e-01
-1.10526979e+00 9.42808613e-02 -2.52780110e-01 -5.52996695e-01
-4.79946733e-01 1.42513752e-01 -4.49678153e-01 3.71744603e-01
2.47569770e-01 5.77296078e-01 -1.06899247e-01 -7.45676085e-02
4.75232452e-01 2.29472652e-01 6.83825970e-01 -4.52733248e-01
1.00468576e+00 6.15992367e-01 2.19299361e-01 -8.98356199e-01
-3.36579144e-01 -2.37117559e-01 -5.87149143e-01 -6.79088473e-01
8.90684724e-01 -6.52534127e-01 -7.32285023e-01 9.05711412e-01
-1.02021694e+00 -6.30815864e-01 -3.74872923e-01 4.10219222e-01
-6.45633459e-01 4.71904993e-01 -4.96951610e-01 -8.79043043e-01
1.95640311e-01 -1.26214278e+00 9.31044400e-01 5.17651439e-01
-1.33046851e-01 -1.18596947e+00 1.37885645e-01 -3.81459892e-02
4.09668714e-01 4.68440950e-01 8.31485808e-01 -2.45995168e-02
-8.08541477e-01 -2.03505218e-01 -1.86727326e-02 -8.91790241e-02
3.19662124e-01 1.59719050e-01 -4.52733934e-01 -1.35874838e-01
-3.29843014e-01 3.09153795e-01 7.66868174e-01 5.93420744e-01
1.11769748e+00 -4.49001253e-01 -6.92008436e-01 7.66295671e-01
9.80904520e-01 2.73167521e-01 9.53153610e-01 9.77986455e-02
1.08226585e+00 5.34084976e-01 5.23436248e-01 4.04253095e-01
8.88632715e-01 6.90330505e-01 5.03049433e-01 1.27853468e-01
-1.58086762e-01 -8.06501150e-01 6.04201019e-01 7.11179554e-01
-1.65408731e-01 -6.60540879e-01 -8.32914591e-01 5.00665724e-01
-2.26503682e+00 -1.37905550e+00 -7.07258761e-01 2.14030743e+00
1.61325872e-01 9.45396274e-02 5.79349577e-01 -9.70724672e-02
6.92138851e-01 1.04611248e-01 -8.87933195e-01 1.50040002e-03
-9.24561638e-04 -5.53357363e-01 4.96873677e-01 3.76036525e-01
-9.49667692e-01 9.70395088e-01 6.42727137e+00 7.65102029e-01
-9.70605135e-01 -7.34982342e-02 6.84496045e-01 2.66727954e-02
-4.93354529e-01 -3.59136090e-02 -1.04481256e+00 1.06039894e+00
7.66116560e-01 -3.39620084e-01 3.26161176e-01 6.57866120e-01
4.39991117e-01 1.00981779e-01 -8.99663746e-01 1.08338797e+00
-2.83840328e-01 -1.19279122e+00 1.94799200e-01 2.92598248e-01
8.03556263e-01 -1.22744679e-01 1.40777037e-01 4.13999617e-01
7.06733823e-01 -9.88690376e-01 8.78896534e-01 9.75802243e-01
4.76280451e-01 -8.47916484e-01 5.16849518e-01 8.99755597e-01
-1.62342834e+00 -1.21158361e-02 -3.81678075e-01 -2.41302326e-02
8.38732600e-01 4.88095820e-01 -4.94295508e-01 5.20812631e-01
6.83483422e-01 1.01480389e+00 -4.69050705e-01 1.13518953e+00
-2.68197775e-01 8.83587956e-01 -2.42851406e-01 -2.67595887e-01
2.50731587e-01 -6.55936599e-01 1.00711322e+00 1.14172804e+00
6.01540744e-01 2.88304258e-02 3.72704923e-01 1.11388338e+00
3.43549222e-01 -1.98587507e-01 -8.58964384e-01 1.32845104e-01
3.73805702e-01 8.43044519e-01 -6.54835999e-01 -3.07018131e-01
-5.61101854e-01 9.72249687e-01 1.36334270e-01 7.48501718e-01
-1.12719119e+00 2.95019113e-02 8.35204303e-01 3.66279781e-01
3.21509689e-01 -4.97195274e-01 -5.19642122e-02 -1.24409354e+00
1.51778162e-01 -4.63360071e-01 1.41178191e-01 -4.46664363e-01
-1.45525444e+00 5.95511258e-01 1.28432378e-01 -1.51999450e+00
-5.85041344e-01 -1.78690165e-01 -1.06239760e+00 7.64403284e-01
-1.22401989e+00 -1.17522275e+00 -5.35291612e-01 7.12818563e-01
5.48267007e-01 -2.28671461e-01 3.92953783e-01 3.36870015e-01
-7.41020799e-01 5.58472574e-01 4.83736433e-02 2.03853264e-01
2.21433416e-01 -1.12892735e+00 8.33498359e-01 1.08559191e+00
-1.14443779e-01 2.57502109e-01 7.56893277e-01 -9.99412537e-01
-1.11760128e+00 -1.58134639e+00 4.16962773e-01 -4.26690906e-01
4.56090480e-01 -1.49199769e-01 -9.36613679e-01 7.30462909e-01
-1.12482915e-02 1.67322066e-02 4.97791588e-01 -3.68967533e-01
1.81122154e-01 6.37607649e-02 -8.66968393e-01 9.82441664e-01
1.32973015e+00 1.73231475e-02 -1.23257987e-01 -1.49403391e-02
4.84090209e-01 -3.53555501e-01 -4.32697743e-01 3.14642400e-01
4.94302601e-01 -1.06456506e+00 8.87229502e-01 -6.24452055e-01
3.16007555e-01 -5.90397120e-01 1.06944311e-02 -1.51332808e+00
-5.20913064e-01 -6.55554652e-01 -6.00650787e-01 1.25757396e+00
5.37031107e-02 -7.45608270e-01 1.08179522e+00 7.46622086e-01
-1.08084962e-01 -8.39526176e-01 -8.50328505e-01 -8.09615552e-01
9.12331343e-02 -5.31471372e-01 1.12656510e+00 6.07954621e-01
-4.08137858e-01 8.01260248e-02 -5.75048685e-01 2.12896198e-01
8.09423983e-01 7.05339834e-02 1.27921510e+00 -8.96527112e-01
-2.61615485e-01 -6.72039807e-01 -7.47239172e-01 -1.78776944e+00
1.55849740e-01 -7.09904075e-01 1.81957055e-02 -1.64735162e+00
4.24983762e-02 -5.61859787e-01 2.41469845e-01 -1.50974348e-01
-4.64407146e-01 -1.85198054e-01 1.25513524e-01 2.57324576e-01
-4.64685708e-01 1.10221851e+00 1.45860064e+00 -2.49009326e-01
-3.52444440e-01 5.76638281e-01 -3.69565636e-01 8.14552605e-01
7.52709568e-01 -2.82609552e-01 -7.84734190e-01 -2.47075632e-01
3.64609025e-02 6.82886690e-02 7.40780771e-01 -1.16677284e+00
3.97981614e-01 -3.89887124e-01 3.51529062e-01 -9.16462183e-01
2.41348043e-01 -6.54731810e-01 4.60483223e-01 5.28547704e-01
1.23832166e-01 3.15186977e-01 3.33997011e-02 9.94103670e-01
-1.72967494e-01 1.96449339e-01 4.41518873e-01 -2.44174786e-02
-8.28626454e-01 9.80072737e-01 -8.33687305e-01 1.15962736e-01
1.17138600e+00 -3.61379862e-01 -2.00069070e-01 -8.80309582e-01
-7.12961376e-01 5.71092129e-01 6.88193142e-01 6.08247340e-01
7.43734062e-01 -1.78253770e+00 -6.20485306e-01 2.37395987e-01
-1.51608303e-01 2.59669214e-01 5.14773905e-01 6.26370907e-01
-4.09140766e-01 1.38683364e-01 -1.41536236e-01 -8.48995328e-01
-7.73087502e-01 4.71454829e-01 3.91653597e-01 -1.70301348e-01
-9.48857307e-01 3.56293857e-01 4.41006452e-01 -4.21116889e-01
-1.33232459e-01 -2.24820167e-01 -2.85649091e-01 -4.77921367e-01
4.67980504e-01 8.01845014e-01 -5.32700658e-01 -1.05736101e+00
-1.15651943e-04 4.10642385e-01 2.29568824e-01 2.48844139e-02
9.74674404e-01 -3.74716997e-01 3.30402195e-01 5.18133879e-01
8.37665558e-01 -1.29162923e-01 -1.73422515e+00 -1.30675007e-02
-3.82895440e-01 -5.41952372e-01 -3.83461505e-01 -1.98311374e-01
-1.21872163e+00 8.53100717e-01 2.91280508e-01 3.47262651e-01
8.99127781e-01 -1.65036455e-01 1.39865935e+00 4.15376537e-02
5.16818464e-01 -7.85355866e-01 -4.81299683e-02 4.30695474e-01
5.33620417e-01 -1.02548182e+00 -3.34555745e-01 -5.79874158e-01
-8.65891159e-01 8.87149870e-01 9.82470393e-01 -2.49283656e-01
8.17250311e-01 1.37009382e-01 -1.63661599e-01 9.12601426e-02
-6.32641613e-01 -1.01200707e-01 2.91317165e-01 9.54629421e-01
-1.09129719e-01 3.11590284e-01 8.98403823e-02 5.95001042e-01
-4.99132127e-01 -1.26482457e-01 4.42380816e-01 5.39848030e-01
-7.99679086e-02 -1.16003501e+00 -2.96412110e-01 4.53584552e-01
4.85832766e-02 3.90962571e-01 -1.52953165e-02 7.81582594e-01
1.95712015e-01 1.00083303e+00 2.14361951e-01 -5.70073903e-01
3.88698429e-01 -2.11559862e-01 2.19146162e-01 -1.53977290e-01
1.70576110e-01 -1.24114968e-01 -2.31146693e-01 -5.42651355e-01
-2.90592641e-01 -1.03785706e+00 -1.28607273e+00 -7.65937746e-01
1.31685302e-01 1.69299487e-02 8.84975344e-02 1.02679348e+00
3.88579965e-01 6.47830904e-01 5.11946321e-01 -7.76890099e-01
-1.75823376e-01 -6.68361306e-01 -3.51435751e-01 6.54036105e-01
3.39909583e-01 -8.67393672e-01 -8.08397681e-02 1.76035628e-01] | [6.416214466094971, 1.0386052131652832] |
3995b755-f472-4109-810f-8b911eae4ccf | cross-domain-few-shot-learning-by-1 | 2010.06498 | null | https://arxiv.org/abs/2010.06498v2 | https://arxiv.org/pdf/2010.06498v2.pdf | Cross-Domain Few-Shot Learning by Representation Fusion | In order to quickly adapt to new data, few-shot learning aims at learning from few examples, often by using already acquired knowledge. The new data often differs from the previously seen data due to a domain shift, that is, a change of the input-target distribution. While several methods perform well on small domain shifts like new target classes with similar inputs, larger domain shifts are still challenging. Large domain shifts may result in high-level concepts that are not shared between the original and the new domain, whereas low-level concepts like edges in images might still be shared and useful. For cross-domain few-shot learning, we suggest representation fusion to unify different abstraction levels of a deep neural network into one representation. We propose Cross-domain Hebbian Ensemble Few-shot learning (CHEF), which achieves representation fusion by an ensemble of Hebbian learners acting on different layers of a deep neural network. Ablation studies show that representation fusion is a decisive factor to boost cross-domain few-shot learning. On the few-shot datasets miniImagenet and tieredImagenet with small domain shifts, CHEF is competitive with state-of-the-art methods. On cross-domain few-shot benchmark challenges with larger domain shifts, CHEF establishes novel state-of-the-art results in all categories. We further apply CHEF on a real-world cross-domain application in drug discovery. We consider a domain shift from bioactive molecules to environmental chemicals and drugs with twelve associated toxicity prediction tasks. On these tasks, that are highly relevant for computational drug discovery, CHEF significantly outperforms all its competitors. Github: https://github.com/ml-jku/chef | ['Sepp Hochreiter', 'Günter Klambauer', 'Michael Kopp', 'David Kreil', 'Andreas Mayr', 'Michael Widrich', 'Johannes Brandstetter', 'Thomas Adler'] | 2020-10-13 | cross-domain-few-shot-learning-by | https://openreview.net/forum?id=w5bNwUzj33 | https://openreview.net/pdf?id=w5bNwUzj33 | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 2.79670954e-01 -1.84634522e-01 -2.21471757e-01 -1.38177678e-01
-8.20141554e-01 -3.48291785e-01 5.81797659e-01 4.24788862e-01
-3.94571155e-01 9.59341466e-01 1.94702577e-02 1.08695202e-01
-2.20230147e-01 -9.46897328e-01 -9.34227586e-01 -7.61866927e-01
4.67107892e-02 6.13312662e-01 5.14555037e-01 -4.53920275e-01
-1.33545429e-01 2.70759553e-01 -1.32014942e+00 7.60116339e-01
6.41063988e-01 6.54760599e-01 3.75204176e-01 2.43249223e-01
-2.23182052e-01 5.20327449e-01 -5.75894058e-01 -1.89650878e-01
2.33691618e-01 -5.21944404e-01 -8.35405052e-01 -2.00246170e-01
2.14802727e-01 -6.48074271e-03 -3.29238027e-01 1.23273456e+00
8.10116589e-01 5.07305503e-01 9.38552141e-01 -9.95114207e-01
-9.83237326e-01 4.56242502e-01 -6.66356683e-01 3.92538160e-01
1.84394270e-01 4.30646896e-01 7.84348607e-01 -1.19299209e+00
1.08101904e+00 1.10062945e+00 6.25233471e-01 9.52354491e-01
-1.53048205e+00 -8.41512203e-01 1.47369623e-01 4.88680273e-01
-1.20896351e+00 -2.00601339e-01 6.44189060e-01 -6.40984476e-01
1.07737136e+00 -1.80165514e-01 3.34019929e-01 1.53948903e+00
2.88859516e-01 6.10121310e-01 9.12644267e-01 -2.67799925e-02
8.23810637e-01 -2.06728317e-02 2.79976487e-01 2.99213976e-01
1.79034486e-01 1.66254565e-01 -5.27118802e-01 -2.20965266e-01
2.90146351e-01 6.47357285e-01 -4.57055628e-01 -6.07830107e-01
-9.08331394e-01 9.58546877e-01 7.56271660e-01 5.89734912e-01
-4.90414083e-01 -9.12712216e-02 5.81078351e-01 4.93578136e-01
5.77519417e-01 7.47762024e-01 -7.28986979e-01 7.13183060e-02
-5.66989720e-01 1.65931970e-01 5.66524565e-01 7.35267222e-01
9.51976001e-01 4.02208567e-02 -3.93000931e-01 9.53431249e-01
-3.54645699e-01 6.97293505e-02 8.54587853e-01 -2.89361179e-01
3.41926515e-02 5.45651197e-01 -2.76970357e-01 -4.25470650e-01
-3.23695481e-01 -4.64101195e-01 -1.02120006e+00 3.63716722e-01
2.53983319e-01 -1.72644794e-01 -1.41379511e+00 1.76666272e+00
2.62984961e-01 6.21314347e-01 1.90091431e-01 6.50785089e-01
1.23455799e+00 7.40099847e-01 2.89230943e-01 -1.89009249e-01
1.37200165e+00 -7.24097848e-01 -3.64500880e-01 -1.55098602e-01
6.60030544e-01 -3.29851717e-01 1.03412855e+00 3.17764521e-01
-6.73710763e-01 -6.16279364e-01 -1.04684329e+00 2.58864671e-01
-1.02271247e+00 -8.32734704e-01 3.33118528e-01 1.64498672e-01
-4.32811767e-01 1.02015567e+00 -3.88009876e-01 -5.46242535e-01
8.74594867e-01 1.99593529e-01 -5.37652075e-01 -6.21093035e-01
-1.65498090e+00 9.62356329e-01 6.54449403e-01 -7.30096281e-01
-1.27088094e+00 -1.45571005e+00 -9.48033690e-01 3.09841037e-01
5.11261702e-01 -7.83131242e-01 1.21606290e+00 -8.33617389e-01
-1.22189426e+00 6.61699831e-01 1.39229506e-01 -6.15242302e-01
2.17704117e-01 3.50953490e-02 -4.80227649e-01 3.59250829e-02
1.15738966e-01 5.64320505e-01 7.92218089e-01 -1.03850901e+00
-5.66094398e-01 -4.44169104e-01 -2.32081398e-01 -3.90420221e-02
-2.02034757e-01 -3.77855480e-01 -6.60775304e-02 -5.53016067e-01
-4.66809601e-01 -6.04214668e-01 -2.86541581e-01 -1.19348295e-01
6.10236451e-02 -2.91174650e-01 9.98258173e-01 -7.85618052e-02
9.64548945e-01 -2.09279680e+00 2.32124910e-01 -1.69198513e-01
4.47640389e-01 7.02100456e-01 -4.99204189e-01 4.24667954e-01
-6.68081939e-01 -1.77467361e-01 -3.54099035e-01 1.95587724e-01
-2.97940969e-01 1.80099890e-01 -3.15618366e-01 2.96555042e-01
2.90324837e-01 1.16789472e+00 -1.29316509e+00 1.31483391e-01
2.29512140e-01 4.38222766e-01 -3.16033691e-01 -9.15581584e-02
-3.87980133e-01 3.36780161e-01 -1.80272803e-01 6.50363326e-01
7.42924809e-01 -3.97661746e-01 -2.96879467e-02 -6.34914562e-02
2.78316170e-01 -2.54187167e-01 -7.48972237e-01 2.09298587e+00
-2.97400922e-01 2.98507899e-01 -5.73399663e-01 -1.31654322e+00
8.54046524e-01 3.42511415e-01 5.55530190e-01 -8.78569067e-01
5.80119602e-02 1.90809295e-01 2.87239760e-01 -3.41273099e-01
-2.16490418e-01 -6.17779315e-01 -8.76082703e-02 1.97027609e-01
8.05348098e-01 -6.20820671e-02 1.44222856e-01 2.09180132e-01
1.46861947e+00 -1.80549785e-01 8.69603515e-01 -1.81338400e-01
4.74166758e-02 -6.34856075e-02 7.97562540e-01 6.62421823e-01
-4.33052868e-01 6.73176706e-01 4.59987670e-01 -7.71157026e-01
-8.78241301e-01 -1.08993447e+00 -2.36185715e-01 1.37129366e+00
1.35095149e-01 -2.84048051e-01 -4.53396261e-01 -8.67125928e-01
2.77563274e-01 6.91517115e-01 -1.14254165e+00 -7.31192470e-01
7.88022652e-02 -8.92521977e-01 2.26441815e-01 4.61465597e-01
1.04992725e-01 -1.19390190e+00 -6.39920354e-01 6.17543638e-01
4.64307964e-01 -7.94227362e-01 -2.63205171e-01 7.40864396e-01
-7.10469484e-01 -1.29925287e+00 -1.35647297e+00 -6.41030610e-01
3.69344175e-01 4.00456101e-01 1.12041163e+00 -4.75469857e-01
-7.17933774e-01 5.42118363e-02 -4.70531106e-01 -7.48130441e-01
-1.80776194e-01 -7.75472224e-02 1.28389955e-01 -1.29623160e-01
7.78098941e-01 -7.64353812e-01 -6.12331808e-01 2.13865101e-01
-1.03841126e+00 -3.49483013e-01 5.93191445e-01 1.43622625e+00
6.32889390e-01 -5.88440299e-02 9.25392926e-01 -1.25976861e+00
6.76480711e-01 -9.56203878e-01 -2.44228184e-01 2.05630466e-01
-3.84469748e-01 9.15623084e-02 8.05286527e-01 -8.16240370e-01
-8.47151816e-01 1.01128094e-01 4.72666463e-03 -8.36608112e-01
-3.90265405e-01 5.05392075e-01 -1.32821739e-01 9.11547095e-02
1.41744053e+00 2.39841431e-01 -6.39884025e-02 -6.18851125e-01
4.71475542e-01 4.30466175e-01 2.06522346e-01 -4.21091795e-01
5.21222413e-01 4.39873755e-01 -1.64547548e-01 -7.69915044e-01
-7.25141823e-01 -6.65163100e-01 -6.39263868e-01 1.90841138e-01
6.95616305e-01 -8.98535311e-01 -3.19542438e-01 4.33746934e-01
-1.02736783e+00 -3.23666692e-01 -7.64500082e-01 3.26017588e-01
-5.24934590e-01 1.28690317e-01 -3.91208559e-01 -8.20159391e-02
-4.48313862e-01 -1.14280701e+00 8.81647289e-01 2.55605549e-01
-2.80509979e-01 -9.96455789e-01 5.81697583e-01 -2.41236418e-01
2.27048472e-01 4.24606711e-01 1.12541652e+00 -1.21543813e+00
7.42393583e-02 -8.14964026e-02 -1.75566882e-01 1.09927639e-01
3.55271727e-01 -4.05056894e-01 -1.21486890e+00 -4.05472189e-01
-2.01515257e-01 -5.35787880e-01 1.47189832e+00 4.80121225e-01
1.10645556e+00 5.45506291e-02 -4.65476394e-01 6.31207824e-01
1.44324934e+00 5.52691400e-01 7.06020653e-01 1.22863792e-01
4.20880139e-01 4.17757392e-01 5.29567420e-01 6.73970342e-01
-2.33528182e-01 5.88221729e-01 3.77454937e-01 -1.85099676e-01
-2.05109164e-01 -9.09743384e-02 2.26659402e-01 1.60394073e-01
4.50834744e-02 8.36479515e-02 -1.03755927e+00 6.46387517e-01
-1.96889365e+00 -1.32619727e+00 4.86784011e-01 2.00405335e+00
1.00530303e+00 8.29963908e-02 1.59917161e-01 -2.55445898e-01
8.00023615e-01 1.33086696e-01 -1.23751724e+00 -2.73974389e-01
-6.71099406e-04 7.88895369e-01 2.38178834e-01 1.70655414e-01
-1.06768501e+00 1.01620662e+00 4.99660063e+00 1.30930793e+00
-1.19760323e+00 4.28329170e-01 6.96409225e-01 -3.52699727e-01
-1.51309371e-01 -2.51095861e-01 -6.42946839e-01 4.98489261e-01
8.79864573e-01 -6.56555116e-01 8.08504894e-02 1.08044338e+00
-2.93789506e-01 1.50149465e-01 -1.31053507e+00 1.03681028e+00
-3.34161171e-03 -1.76477242e+00 1.48835570e-01 5.90123832e-02
9.50687230e-01 4.95804846e-01 1.69169039e-01 7.68640876e-01
5.35745919e-01 -1.08486032e+00 -1.29806384e-01 4.08505082e-01
1.07549918e+00 -7.54586518e-01 7.14973688e-01 2.68428832e-01
-1.06316066e+00 -1.74513623e-01 -8.87472391e-01 9.41339657e-02
-2.13405207e-01 6.18773699e-01 -9.31847990e-01 4.64449376e-01
6.21991217e-01 1.16825402e+00 -2.80179620e-01 1.09968948e+00
6.51088580e-02 9.31729898e-02 5.43443076e-02 6.25467747e-02
3.35944772e-01 2.25891545e-01 5.05293846e-01 1.07296026e+00
3.45181882e-01 3.05038720e-01 6.05554655e-02 8.51981401e-01
-3.81253153e-01 1.91316083e-02 -1.15638089e+00 3.87899056e-02
3.11803550e-01 1.12711835e+00 -4.40669060e-01 -7.16851056e-01
-4.73214865e-01 1.02995288e+00 3.14204603e-01 5.40478468e-01
-7.23624051e-01 -6.48752391e-01 1.11090076e+00 1.38668582e-01
5.72203875e-01 5.06615281e-01 7.95821622e-02 -1.26458073e+00
-5.34547031e-01 -9.45813537e-01 6.75743401e-01 -4.18130219e-01
-1.67607152e+00 6.10685766e-01 -1.20316990e-01 -1.52796805e+00
2.53638234e-02 -7.43162572e-01 -7.96320379e-01 7.79470861e-01
-1.51105487e+00 -8.81656051e-01 -1.54414386e-01 8.34396482e-01
1.04792142e+00 -6.13830090e-01 1.17419958e+00 3.21325153e-01
-3.84564877e-01 4.40711737e-01 5.00735044e-01 -7.30143785e-02
1.10309529e+00 -1.08220172e+00 5.35533011e-01 4.43007410e-01
1.47013396e-01 5.52788675e-01 5.97547114e-01 -7.22843587e-01
-1.05883741e+00 -1.42549014e+00 1.99243218e-01 -3.62428814e-01
7.34264493e-01 -3.82621258e-01 -1.41586328e+00 3.34178120e-01
1.71748340e-01 5.48979044e-01 1.09454703e+00 1.21922009e-01
-7.59581447e-01 -6.37132674e-02 -1.25188482e+00 3.35862964e-01
9.99740362e-01 -4.87238914e-01 -8.44514430e-01 3.45491171e-01
7.47650146e-01 4.24805433e-02 -8.29168856e-01 4.82521653e-01
5.20443141e-01 -7.37777770e-01 1.03712535e+00 -1.26420045e+00
4.99277323e-01 -1.33060366e-01 -2.89828449e-01 -1.96942508e+00
-7.95540929e-01 -3.33617836e-01 -3.57837766e-01 6.17103934e-01
3.41888309e-01 -5.23248196e-01 6.39309585e-01 2.87105948e-01
-1.18057817e-01 -9.59424555e-01 -1.01804435e+00 -1.20132780e+00
5.07928908e-01 -9.58096012e-02 6.20475829e-01 1.15496135e+00
3.61056387e-01 7.87523985e-01 -4.45074797e-01 -3.59175831e-01
5.51023543e-01 3.16136852e-02 5.68424761e-01 -1.61324263e+00
-4.21438366e-01 -5.55912018e-01 -5.89863837e-01 -3.85779321e-01
1.33758783e-01 -1.06349528e+00 -8.64275917e-02 -1.45452523e+00
6.31819487e-01 3.20576191e-01 -8.79786193e-01 7.11086810e-01
-1.48554698e-01 1.01196945e-01 2.73548931e-01 3.45796533e-02
-7.20721185e-01 7.68769622e-01 1.20379984e+00 -7.16654897e-01
-3.35812092e-01 -8.41452330e-02 -8.57739925e-01 6.67970538e-01
5.93945444e-01 -5.86443365e-01 -4.98197883e-01 4.31010574e-02
-2.16191933e-01 -2.29592308e-01 1.36099920e-01 -1.11683297e+00
1.83542728e-01 -2.51381189e-01 6.74478590e-01 -2.00870216e-01
3.35368067e-01 -7.53622472e-01 1.36251852e-01 9.22663689e-01
-1.74995467e-01 -5.75237393e-01 3.90501291e-01 8.67392540e-01
-2.56271005e-01 -1.62423983e-01 1.33697510e+00 -4.11544591e-01
-1.34117520e+00 8.36145818e-01 -1.72573313e-01 2.60270327e-01
1.43576765e+00 -2.37136781e-01 -5.34097254e-01 -4.62046564e-02
-1.12183571e+00 2.56694527e-03 3.48143667e-01 5.48115075e-01
9.70969200e-01 -1.41128230e+00 -6.44570351e-01 2.24724382e-01
8.02873492e-01 -1.28489301e-01 7.13348448e-01 5.88905394e-01
3.78532298e-02 1.52931184e-01 -6.44991755e-01 -3.83742452e-01
-1.20911241e+00 9.67004001e-01 3.89682382e-01 -2.56541222e-01
-6.58225060e-01 1.05821443e+00 9.09917176e-01 -3.33534360e-01
6.97934702e-02 -1.88129440e-01 -1.49651617e-01 5.28594434e-01
9.33708370e-01 1.58254594e-01 8.40596557e-02 -1.76963225e-01
-3.68380606e-01 5.82323015e-01 -5.26508927e-01 3.40510756e-01
1.71115673e+00 4.34668809e-01 4.72886562e-01 5.01979887e-01
1.39133954e+00 -7.89900124e-01 -1.33104777e+00 -6.98564827e-01
-4.80613336e-02 -4.09544498e-01 -1.93105280e-01 -1.00667512e+00
-8.74569297e-01 1.31565988e+00 8.33124518e-01 -3.21725458e-01
9.90064740e-01 2.08158880e-01 8.29549909e-01 5.30939460e-01
3.63456070e-01 -1.18388224e+00 4.23766136e-01 7.25315452e-01
8.42469573e-01 -1.54717064e+00 -7.57838860e-02 2.14981422e-01
-8.76502752e-01 1.01976895e+00 6.80801809e-01 -9.43820365e-03
8.10021579e-01 -7.34614804e-02 -3.81629050e-01 -4.16663617e-01
-1.05001616e+00 -4.97991830e-01 3.33939224e-01 1.00048816e+00
1.87813431e-01 1.45984078e-02 9.57952216e-02 9.09017444e-01
5.57590783e-01 3.01460773e-01 3.60031217e-01 7.71949708e-01
-7.25958943e-01 -1.01922846e+00 -9.80405416e-03 5.19292414e-01
-2.03321666e-01 -2.15806887e-01 -4.21982408e-01 6.28242433e-01
2.73393720e-01 3.44515681e-01 -8.17441344e-02 -4.70693469e-01
5.56649745e-01 3.10778886e-01 4.57433969e-01 -1.25806260e+00
-3.76131892e-01 3.70733142e-02 -4.01775569e-01 -5.09144366e-01
4.11081186e-04 -2.93419629e-01 -1.29603195e+00 -2.11720973e-01
2.23752093e-02 1.05752148e-01 1.94398671e-01 7.55449355e-01
6.99961185e-01 7.87805080e-01 4.28952456e-01 -7.94768155e-01
-5.08755267e-01 -1.01358831e+00 -9.02091444e-01 7.80995429e-01
3.15807104e-01 -9.93354142e-01 -1.28380641e-01 -7.18353987e-02] | [9.996828079223633, 2.9291234016418457] |
e7f579ef-ebdc-4aa5-adc8-be8eb4fb421b | adversarial-attacks-and-defences-for-skin | 2212.06822 | null | https://arxiv.org/abs/2212.06822v1 | https://arxiv.org/pdf/2212.06822v1.pdf | Adversarial Attacks and Defences for Skin Cancer Classification | There has been a concurrent significant improvement in the medical images used to facilitate diagnosis and the performance of machine learning techniques to perform tasks such as classification, detection, and segmentation in recent years. As a result, a rapid increase in the usage of such systems can be observed in the healthcare industry, for instance in the form of medical image classification systems, where these models have achieved diagnostic parity with human physicians. One such application where this can be observed is in computer vision tasks such as the classification of skin lesions in dermatoscopic images. However, as stakeholders in the healthcare industry, such as insurance companies, continue to invest extensively in machine learning infrastructure, it becomes increasingly important to understand the vulnerabilities in such systems. Due to the highly critical nature of the tasks being carried out by these machine learning models, it is necessary to analyze techniques that could be used to take advantage of these vulnerabilities and methods to defend against them. This paper explores common adversarial attack techniques. The Fast Sign Gradient Method and Projected Descent Gradient are used against a Convolutional Neural Network trained to classify dermatoscopic images of skin lesions. Following that, it also discusses one of the most popular adversarial defense techniques, adversarial training. The performance of the model that has been trained on adversarial examples is then tested against the previously mentioned attacks, and recommendations to improve neural networks robustness are thus provided based on the results of the experiment. | ['Shraddha Surtkar', 'Samina Attari', 'Ishaan Shivhare', 'Joy Purohit', 'Vinay Jogani'] | 2022-12-13 | null | null | null | null | ['adversarial-defense', 'skin-cancer-classification'] | ['adversarial', 'medical'] | [ 6.01817548e-01 3.26874286e-01 1.29798859e-01 -2.12120146e-01
-1.55933604e-01 -7.45550454e-01 5.13778210e-01 1.41481698e-01
-6.27552390e-01 3.96561712e-01 -2.25227773e-01 -7.28028893e-01
-7.07151964e-02 -8.10475230e-01 -5.12239873e-01 -6.34328663e-01
-1.46320775e-01 4.34015319e-02 1.56318799e-01 -3.59941393e-01
3.31581563e-01 9.76499617e-01 -1.23187184e+00 3.83648664e-01
8.92930090e-01 9.20931399e-01 -3.60556662e-01 9.65574324e-01
3.97302955e-02 6.25878751e-01 -6.24205887e-01 -6.50872529e-01
4.44560528e-01 -2.92731345e-01 -8.41387272e-01 -1.47689536e-01
2.78128952e-01 -2.79044569e-01 -2.91087747e-01 1.17769241e+00
5.98162651e-01 -1.42964870e-01 4.19924229e-01 -7.24120319e-01
-9.59233940e-02 2.21354216e-01 -4.14142877e-01 3.35957766e-01
2.21834704e-01 2.62685239e-01 3.11214149e-01 -2.82044187e-02
5.12285650e-01 9.42741573e-01 4.05985087e-01 7.14576840e-01
-9.66411352e-01 -5.99101603e-01 -2.07522437e-01 6.78163841e-02
-8.55379462e-01 6.27419800e-02 5.25842071e-01 -3.47472072e-01
5.93830347e-01 5.57100058e-01 4.89229441e-01 9.95226502e-01
4.63119328e-01 5.32327473e-01 1.30116475e+00 -6.11663699e-01
1.00268602e-01 6.19587243e-01 -6.43969849e-02 6.30941570e-01
2.21192226e-01 2.52289474e-01 1.41526550e-01 -8.44735503e-02
6.03280902e-01 -2.52897531e-01 -1.91311911e-01 4.85663209e-03
-4.26641822e-01 1.01695788e+00 6.59711659e-01 5.64772904e-01
-4.26551491e-01 -1.35736749e-01 5.48151731e-01 1.31377026e-01
3.88996392e-01 7.33253181e-01 -1.01040853e-02 -1.07356928e-01
-7.03550816e-01 1.15467317e-01 7.81622231e-01 6.94881529e-02
-1.00902148e-01 1.16837457e-01 1.92741811e-01 5.80164969e-01
-6.39867038e-02 2.23615721e-01 5.33313036e-01 -5.61551213e-01
2.68980622e-01 7.53867567e-01 -2.03347206e-01 -9.91229832e-01
-2.05546349e-01 -3.23915005e-01 -7.39567578e-01 8.68046701e-01
5.35365701e-01 -2.95545220e-01 -1.17611563e+00 1.21396327e+00
4.30291057e-01 5.00646941e-02 2.01921672e-01 7.37333238e-01
3.18764120e-01 2.46701851e-01 1.52165636e-01 3.71002465e-01
1.16271842e+00 -4.16046828e-01 -3.11377108e-01 -5.23504317e-02
3.70391250e-01 -9.64184880e-01 4.56450105e-01 6.54682100e-01
-9.29012597e-01 -4.05360013e-01 -1.08900726e+00 4.75467831e-01
-6.57919168e-01 -2.36725181e-01 4.31704253e-01 1.25742614e+00
-5.56154072e-01 8.61614287e-01 -1.08405292e+00 -3.89562726e-01
6.32910669e-01 6.81681037e-01 -3.00633967e-01 -6.01663925e-02
-1.22315586e+00 1.16969287e+00 2.95089334e-01 1.47880539e-01
-4.50051248e-01 -4.89123106e-01 -5.79536557e-01 -1.55262366e-01
9.18561816e-02 -2.62581587e-01 8.21939707e-01 -1.35060322e+00
-1.07115471e+00 1.17788720e+00 3.79435807e-01 -9.05420423e-01
8.20353746e-01 -1.23318627e-01 -6.28246725e-01 2.88593978e-01
-4.43594813e-01 3.30951214e-01 8.93562257e-01 -9.20646727e-01
-8.61313820e-01 -4.41719562e-01 4.16132033e-01 -1.09344237e-01
-3.94511938e-01 2.66482949e-01 2.15825588e-02 -3.96509469e-01
-4.59506840e-01 -1.29378283e+00 -3.88283610e-01 4.34238836e-02
-6.26729608e-01 2.75216967e-01 9.68812108e-01 -8.94668162e-01
9.33300734e-01 -2.16145205e+00 -6.50582612e-02 7.20430851e-01
8.85251537e-02 1.10113680e+00 2.74818778e-01 3.65611881e-01
-3.29394549e-01 2.87092745e-01 -2.10375339e-01 1.84476703e-01
-5.23258924e-01 1.67832166e-01 -3.53716135e-01 3.87554258e-01
1.96393758e-01 5.16041040e-01 -5.22761166e-01 -2.71436691e-01
6.19225502e-01 5.95820487e-01 -1.63004428e-01 2.30895892e-01
1.47333622e-01 6.36186719e-01 -5.72730005e-01 4.72126156e-01
4.33760285e-01 -3.60442773e-02 2.04963028e-01 -1.43491864e-01
1.08621970e-01 -2.72047073e-01 -8.66276860e-01 9.03936505e-01
-4.87530440e-01 6.46674335e-01 6.65152520e-02 -1.05256045e+00
4.78483230e-01 4.81545866e-01 4.38565642e-01 -3.54876131e-01
3.38379622e-01 2.36740008e-01 5.66258192e-01 -8.66612732e-01
1.69046119e-01 -6.14734017e-04 3.91118646e-01 1.09990343e-01
-4.71647203e-01 -7.71356300e-02 1.06395766e-01 2.13098779e-01
1.09540415e+00 -3.24041694e-02 2.14036569e-01 2.17260867e-01
8.94926727e-01 2.54455507e-01 -3.30122709e-02 4.07281280e-01
-2.95093894e-01 3.79746765e-01 2.50160933e-01 -4.87174720e-01
-1.02124310e+00 -7.75413632e-01 -3.08780491e-01 3.28483522e-01
2.26223245e-02 3.16617399e-01 -1.01089013e+00 -7.93209255e-01
6.64078742e-02 4.50200588e-01 -7.59378135e-01 -3.99397731e-01
-5.22421420e-01 -6.32937729e-01 9.50538933e-01 4.60093439e-01
5.71837485e-01 -1.19026434e+00 -9.89251673e-01 9.51001272e-02
4.32501495e-01 -1.06215930e+00 1.39884695e-01 -3.04869916e-02
-7.73273647e-01 -1.42289114e+00 -5.67534328e-01 -5.88250220e-01
7.79052317e-01 -1.25865415e-01 5.12733161e-01 2.19754115e-01
-1.01142538e+00 5.16301453e-01 -4.62977499e-01 -8.13961089e-01
-9.17927682e-01 -9.01671275e-02 -8.57397839e-02 1.71244711e-01
3.51510257e-01 -3.88454795e-01 -7.21960247e-01 2.18462385e-02
-1.35318851e+00 -3.34468007e-01 7.07465827e-01 8.85141373e-01
2.16949433e-01 3.09618562e-01 2.52764612e-01 -1.35936272e+00
8.18241119e-01 -4.91319567e-01 -4.85554755e-01 2.57933170e-01
-4.85210598e-01 -1.27908066e-01 8.06445539e-01 -5.31650364e-01
-8.64249587e-01 -6.17195401e-05 -5.09490073e-01 -3.38779390e-01
-4.70136255e-01 5.04598320e-01 3.48894358e-01 -7.94080436e-01
9.99740183e-01 -1.37454659e-01 3.67826045e-01 -1.20988779e-01
4.96657975e-02 8.62464309e-01 4.38793689e-01 -7.59801641e-02
9.08365607e-01 5.13435721e-01 2.33472541e-01 -1.02145278e+00
-4.21207726e-01 -2.42245302e-01 -1.72704294e-01 -3.70645016e-01
9.76637721e-01 -1.40798330e-01 -6.59087896e-01 6.74137592e-01
-7.87167370e-01 3.67739499e-02 -1.03125483e-01 4.71207976e-01
-1.73121393e-01 4.57309902e-01 -6.00704193e-01 -9.51779366e-01
-3.88052791e-01 -1.25556302e+00 3.84417981e-01 6.19769156e-01
-2.68710703e-01 -1.33562565e+00 -1.49329141e-01 6.41557515e-01
7.00703859e-01 8.33445728e-01 1.18585336e+00 -8.89364004e-01
-2.94305772e-01 -9.09042895e-01 1.57769471e-01 9.31935132e-01
2.72433817e-01 2.70753622e-01 -1.10044885e+00 -4.24078584e-01
1.01777963e-01 -3.01189780e-01 3.89798105e-01 3.66739452e-01
1.30077195e+00 -8.20897594e-02 -3.23190093e-01 5.51250756e-01
1.41695642e+00 5.86723387e-01 8.86301458e-01 3.37252587e-01
4.64178950e-01 1.04346120e+00 6.64162517e-01 -9.91099626e-02
-4.24149603e-01 4.15966898e-01 8.94505262e-01 -4.69930530e-01
1.16371371e-01 7.05965087e-02 4.75155609e-03 4.58834283e-02
-4.02382731e-01 1.36967316e-01 -8.16211641e-01 2.50691473e-01
-1.39904475e+00 -9.95246530e-01 1.37592942e-01 2.21499944e+00
4.92768824e-01 4.72796112e-01 5.84929846e-02 3.99286270e-01
8.19431365e-01 -8.85164067e-02 -7.76884317e-01 -9.94158328e-01
2.01734886e-01 7.63477325e-01 7.49307334e-01 2.00102508e-01
-1.19231653e+00 7.27378368e-01 5.66669893e+00 6.51563883e-01
-1.70705581e+00 -4.44509655e-01 7.97875166e-01 2.38105997e-01
2.35223040e-01 -4.55460101e-01 -2.86396593e-01 5.35740376e-01
8.71056557e-01 4.58215475e-02 3.86393040e-01 8.05774808e-01
1.40187770e-01 -2.40841925e-01 -6.71296000e-01 5.80342889e-01
-1.37036324e-01 -1.29814839e+00 -1.24047257e-01 3.12594473e-01
6.15543187e-01 -2.30856806e-01 6.18282318e-01 -1.77668750e-01
2.43542671e-01 -1.29852414e+00 -3.95765249e-03 1.68119416e-01
7.10717380e-01 -9.86416161e-01 9.24161494e-01 2.60452062e-01
-5.87893665e-01 -2.02003345e-01 -1.17445057e-02 2.27584213e-01
5.84275499e-02 2.28405654e-01 -9.57102597e-01 5.18214822e-01
5.86675584e-01 -1.52690532e-02 -3.70927006e-01 1.10444236e+00
-3.33636940e-01 8.18788111e-01 -1.57667398e-01 -3.84893566e-02
4.13987905e-01 -9.25382823e-02 5.82646787e-01 9.17237222e-01
-4.40200344e-02 -7.97538757e-02 -2.38550097e-01 4.79676843e-01
2.10102141e-01 1.06339559e-01 -7.30639160e-01 -3.19334567e-01
2.12784916e-01 1.39700055e+00 -7.55058885e-01 1.08230803e-02
-2.62989581e-01 1.01235175e+00 -2.05439970e-01 1.47022352e-01
-8.32979620e-01 -6.10745311e-01 8.14783871e-01 3.52632374e-01
1.70717519e-02 5.64710908e-02 -2.82195866e-01 -5.78530848e-01
-4.33524214e-02 -1.33511257e+00 2.90214449e-01 -3.20638955e-01
-1.08443546e+00 7.39049256e-01 -2.07895026e-01 -1.05314410e+00
-5.60340405e-01 -9.68229234e-01 -7.93949485e-01 1.08934486e+00
-1.31435895e+00 -1.19465292e+00 -1.16385728e-01 7.01677084e-01
2.28679433e-01 -4.23436612e-01 9.25623834e-01 1.43105492e-01
-3.89763236e-01 6.52826130e-01 5.39079830e-02 3.95562053e-01
5.61897397e-01 -1.15482509e+00 3.80466163e-01 1.05492818e+00
1.44932061e-01 6.80525661e-01 6.40539944e-01 -6.33631051e-01
-9.56755936e-01 -8.35136056e-01 2.64324903e-01 -1.09669842e-01
5.98080218e-01 9.84589979e-02 -7.69116163e-01 3.93516332e-01
1.02678500e-01 -1.46711677e-01 7.22765207e-01 -2.96276808e-01
-1.20524645e-01 -7.87991751e-03 -1.66925132e+00 7.86418438e-01
3.15565109e-01 -4.96522099e-01 -3.61882240e-01 4.11633611e-01
1.18423156e-01 -6.48689210e-01 -8.15077424e-01 3.79260898e-01
6.16791964e-01 -9.19589818e-01 9.85496879e-01 -9.46886241e-01
6.17680490e-01 5.39946333e-02 3.63176525e-01 -1.30787301e+00
1.77064419e-01 -4.66300786e-01 1.75433204e-01 7.75860667e-01
3.43561858e-01 -9.56738353e-01 1.18081963e+00 8.95398259e-01
2.73657352e-01 -1.13698924e+00 -7.52178729e-01 -3.70309204e-01
2.05744728e-02 -8.86533186e-02 6.86746538e-02 8.43574643e-01
-2.86929309e-01 -2.74592698e-01 -2.75962502e-01 3.01782906e-01
3.64111036e-01 -3.23083282e-01 6.79361284e-01 -9.50899184e-01
-4.13416624e-01 -2.66094506e-01 -1.05928862e+00 -2.74324477e-01
-3.27267289e-01 -5.59804618e-01 -5.39802849e-01 -1.06593943e+00
-3.26437145e-01 -5.08167565e-01 -2.56391883e-01 4.11223799e-01
-3.61209691e-01 5.01058340e-01 2.41927803e-01 4.12813500e-02
3.59200358e-01 -4.99908775e-01 1.02478063e+00 -1.55651972e-01
-4.68659103e-02 5.01742184e-01 -6.40759230e-01 8.09981763e-01
1.00536036e+00 -2.87915796e-01 -2.54204571e-01 4.27306779e-02
-5.62051684e-02 -1.20865747e-01 2.61531264e-01 -1.20032203e+00
2.18967304e-01 1.18545122e-01 5.13918757e-01 -2.67158691e-02
3.11598748e-01 -1.12624741e+00 1.75635070e-02 1.15332210e+00
-3.37406456e-01 -1.01409569e-01 2.11061671e-01 1.72512904e-01
-4.02525991e-01 -3.47187370e-01 1.06189644e+00 -1.88362479e-01
-6.87325656e-01 5.42257205e-02 -3.72897625e-01 -2.81495512e-01
1.55514407e+00 -5.49009740e-01 -2.59881288e-01 -2.94763863e-01
-8.94971192e-01 4.09723446e-02 3.17343354e-01 3.36418033e-01
5.55471897e-01 -6.49143457e-01 -5.35373449e-01 1.97687984e-01
-1.78614363e-01 -2.98857331e-01 4.37630534e-01 7.80136883e-01
-1.12101698e+00 2.65551299e-01 -6.28865123e-01 -2.63987273e-01
-1.82529891e+00 5.95726490e-01 6.53339386e-01 -4.76599395e-01
-3.89095992e-01 7.63905346e-01 -1.72486812e-01 1.32326365e-01
2.56108314e-01 1.70795545e-01 -2.99415857e-01 -3.32192659e-01
5.04558980e-01 4.25429612e-01 2.23299697e-01 -2.82559663e-01
-2.03155041e-01 2.95233071e-01 -4.99819756e-01 7.82688931e-02
1.13009787e+00 3.85173380e-01 4.50980291e-02 -8.14186633e-02
8.29368949e-01 1.79740131e-01 -8.99488509e-01 3.26567709e-01
-1.43176615e-01 -4.52936441e-01 -2.49074981e-01 -1.15791667e+00
-1.21432281e+00 1.13047469e+00 1.03422797e+00 6.83677077e-01
1.31705475e+00 -6.05540335e-01 7.60830104e-01 3.69990505e-02
2.72664994e-01 -1.01770091e+00 -1.42706528e-01 -8.19972157e-02
4.02263612e-01 -1.17175889e+00 -6.40320554e-02 -5.19103825e-01
-6.28010690e-01 1.24702024e+00 3.04327220e-01 -4.74221051e-01
3.98095816e-01 3.05423707e-01 5.63158631e-01 -5.61814308e-02
4.56823371e-02 7.45695531e-02 2.16190040e-01 7.21936643e-01
3.09387475e-01 5.43830805e-02 -5.21598816e-01 -2.56296873e-01
3.88352983e-02 -4.01855856e-02 4.60379899e-01 1.06184435e+00
-8.07618946e-02 -1.56684589e+00 -5.14566183e-01 6.88061059e-01
-1.29352760e+00 1.45882145e-01 -6.69657409e-01 9.81063843e-01
1.83263630e-01 8.21375847e-01 -4.20644432e-01 -4.66727048e-01
3.45445156e-01 -1.05092093e-01 4.06423450e-01 -4.38402623e-01
-1.20528269e+00 -4.49527889e-01 2.08792657e-01 -5.68226576e-01
-2.42337346e-01 -2.16490820e-01 -9.42558229e-01 -4.33555394e-01
-1.87754348e-01 1.20408144e-02 1.08890367e+00 8.33847582e-01
-5.70271611e-02 6.74269617e-01 6.40271962e-01 -4.61096466e-01
-9.24648464e-01 -5.63221514e-01 -5.63369215e-01 6.81074500e-01
1.08556077e-01 -1.87482029e-01 -2.84963787e-01 -1.98840082e-01] | [5.591829776763916, 7.808200836181641] |
03ae97cc-62c2-498d-98a1-4c387c16d00d | labels-are-not-perfect-improving | 2008.04168 | null | https://arxiv.org/abs/2008.04168v1 | https://arxiv.org/pdf/2008.04168v1.pdf | Labels Are Not Perfect: Improving Probabilistic Object Detection via Label Uncertainty | Reliable uncertainty estimation is crucial for robust object detection in autonomous driving. However, previous works on probabilistic object detection either learn predictive probability for bounding box regression in an un-supervised manner, or use simple heuristics to do uncertainty regularization. This leads to unstable training or suboptimal detection performance. In this work, we leverage our previously proposed method for estimating uncertainty inherent in ground truth bounding box parameters (which we call label uncertainty) to improve the detection accuracy of a probabilistic LiDAR-based object detector. Experimental results on the KITTI dataset show that our method surpasses both the baseline model and the models based on simple heuristics by up to 3.6% in terms of Average Precision. | ['Klaus Dietmayer', 'Lars Rosenbaum', 'Di Feng', 'Fabian Timm'] | 2020-08-10 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [-8.97399150e-03 2.56783247e-01 -9.98333469e-02 -4.86656874e-01
-1.12053812e+00 -5.16662538e-01 4.97083783e-01 1.43970191e-01
-6.65536940e-01 8.96762729e-01 -5.26489079e-01 -3.94143075e-01
8.96502882e-02 -5.56143582e-01 -8.20056140e-01 -6.73585951e-01
4.16779034e-02 6.02142572e-01 9.59159374e-01 2.66450346e-01
3.21199507e-01 6.20803654e-01 -1.74283957e+00 -3.09831321e-01
8.34463358e-01 1.10669863e+00 6.10381030e-02 5.95166683e-01
-5.99586479e-02 5.23935318e-01 -4.28505719e-01 -1.37494311e-01
3.37992728e-01 1.97283655e-01 -9.75123048e-02 -1.29487142e-01
6.53828561e-01 -4.72058922e-01 -2.61625618e-01 1.02561617e+00
1.49127513e-01 1.98018074e-01 9.46035743e-01 -1.31810558e+00
6.43654019e-02 6.39862478e-01 -8.21918905e-01 2.67438918e-01
-1.45355761e-01 8.26805905e-02 7.20781028e-01 -1.04138362e+00
2.45133013e-01 1.23049438e+00 8.56880546e-01 5.52196324e-01
-1.33231056e+00 -9.44777191e-01 2.40328923e-01 8.23503062e-02
-1.91432667e+00 -5.18053412e-01 5.18145800e-01 -6.90781295e-01
7.10592031e-01 -1.69826135e-01 2.14388758e-01 6.88579679e-01
3.71166259e-01 6.23722315e-01 1.21306705e+00 -1.66554734e-01
4.16425854e-01 3.95333022e-01 3.04112554e-01 7.68617272e-01
9.05979514e-01 7.15569317e-01 -4.30215567e-01 -1.24936163e-01
4.37611490e-01 -4.45017457e-01 2.97847241e-01 -7.50120938e-01
-8.19111228e-01 1.00013196e+00 4.92231369e-01 -3.79663944e-01
5.31665003e-03 4.75133181e-01 1.39882907e-01 -2.50947505e-01
4.65929985e-01 2.78184265e-01 -1.26481444e-01 8.28010030e-03
-9.91363049e-01 4.02765214e-01 6.88549876e-01 1.25344992e+00
9.35888290e-01 9.66991186e-02 -2.60662764e-01 4.46213365e-01
7.12473810e-01 8.25623155e-01 -1.57460660e-01 -1.03725743e+00
2.72584707e-01 3.11293513e-01 7.15722263e-01 -7.54727721e-01
-3.72947246e-01 -1.14804268e-01 -2.25414976e-01 7.40478098e-01
4.57425386e-01 -1.72503814e-01 -1.37843132e+00 1.34275150e+00
4.69737828e-01 1.98639452e-01 6.86296970e-02 8.08899224e-01
7.11607993e-01 3.45264941e-01 2.30617315e-01 2.65783668e-02
9.76004303e-01 -7.21388400e-01 -6.29483938e-01 -5.87357461e-01
4.02341276e-01 -6.35032952e-01 5.72569132e-01 3.83100748e-01
-4.72522408e-01 -4.67318028e-01 -1.45727885e+00 1.54412135e-01
-2.38613695e-01 2.27808952e-01 4.22299981e-01 9.20672417e-01
-5.70295990e-01 3.81267786e-01 -1.08782160e+00 -1.66057453e-01
5.66888452e-01 2.81553775e-01 -5.76573573e-02 -2.09272861e-01
-6.26568854e-01 1.32997227e+00 8.40013504e-01 1.55948894e-02
-1.05505753e+00 -5.36939800e-01 -9.78380680e-01 -4.53092873e-01
8.51570547e-01 -3.01421523e-01 1.33038557e+00 1.34442478e-01
-1.43981397e+00 5.33752799e-01 -8.45459551e-02 -9.29728031e-01
8.35210800e-01 -6.28670216e-01 7.29024261e-02 -1.50604635e-01
2.91240271e-02 1.09376323e+00 1.10293806e+00 -1.60986197e+00
-9.93207574e-01 -1.30574927e-01 -1.07775576e-01 -1.23335809e-01
4.79874432e-01 -1.08289286e-01 -3.85940045e-01 -2.11645126e-01
5.36441207e-01 -1.16400552e+00 -4.68274832e-01 4.77474928e-02
-3.50492924e-01 -2.66002715e-01 1.05185246e+00 -2.29113117e-01
8.69625866e-01 -1.96413064e+00 -3.00729871e-01 2.95476586e-01
2.76981741e-01 3.10190842e-02 3.93192142e-01 -7.27643967e-02
5.62678039e-01 8.15493464e-02 -5.58405399e-01 -4.30075705e-01
3.37950792e-03 5.95172405e-01 -4.59670693e-01 5.95282972e-01
4.96627808e-01 5.60161710e-01 -1.02010620e+00 -8.56606066e-01
5.88352978e-01 3.20264399e-01 -3.23457986e-01 1.71702594e-01
-2.30010644e-01 3.19754779e-01 -4.75382984e-01 7.98937321e-01
8.38046014e-01 3.17874014e-01 -2.41803452e-01 1.74859855e-02
-2.83144712e-01 2.36561358e-01 -1.40136313e+00 1.12965107e+00
-2.13317454e-01 6.55766547e-01 -6.75719231e-02 -4.32578951e-01
1.07801437e+00 -1.90448776e-01 2.89177150e-01 1.82623744e-01
1.94990605e-01 3.30843896e-01 -6.76491708e-02 8.80121961e-02
7.63553202e-01 -1.63585573e-01 -2.46083826e-01 4.99545299e-02
-1.65912747e-01 -8.32031786e-01 1.75419077e-01 1.25149816e-01
8.52303505e-01 4.47001070e-01 3.46448392e-01 -3.86943370e-01
3.01808983e-01 3.33228499e-01 6.80815697e-01 1.18921137e+00
-6.26525044e-01 5.73276639e-01 1.07738085e-01 -2.17725933e-01
-8.62587690e-01 -1.24925387e+00 -7.13999450e-01 7.04988360e-01
4.92375106e-01 -3.40798765e-01 -6.96819901e-01 -8.17438066e-01
4.31050181e-01 9.83693719e-01 -4.64210570e-01 -2.57923067e-01
-3.50027800e-01 -7.15279043e-01 4.82349277e-01 8.83094013e-01
2.94242233e-01 -6.30389631e-01 -8.18893433e-01 2.51566499e-01
1.98687553e-01 -1.46923542e+00 8.82109329e-02 4.21584874e-01
-8.41860890e-01 -1.00643516e+00 -6.61091208e-02 1.05025701e-01
5.92695236e-01 3.36917102e-01 8.75459909e-01 -1.05955571e-01
-3.80635321e-01 2.35231504e-01 -2.31936783e-01 -1.07297289e+00
-4.32551503e-01 -1.58851594e-01 4.21571881e-01 -3.71665955e-01
5.93205571e-01 -2.29260903e-02 -2.32294679e-01 5.54451883e-01
-5.10309756e-01 -3.66169780e-01 6.94531918e-01 6.95004404e-01
8.44707072e-01 5.66234142e-02 2.98287332e-01 -6.44052327e-01
1.45775065e-01 -3.59582841e-01 -1.37483060e+00 -2.09795728e-01
-9.75840867e-01 3.55119377e-01 -2.31060058e-01 -3.33457053e-01
-1.02833271e+00 6.88092530e-01 2.21553743e-01 -6.07106686e-01
-1.40151694e-01 1.25422865e-01 2.25873634e-01 -5.49164295e-01
8.81575525e-01 -3.11618686e-01 -2.45427966e-01 -1.87657833e-01
4.55519617e-01 4.79292154e-01 6.60823226e-01 -7.88297534e-01
1.16937184e+00 7.37183094e-01 2.11807176e-01 -6.65250778e-01
-1.22786486e+00 -6.88536942e-01 -8.17196608e-01 -2.13785812e-01
6.97005451e-01 -1.07851505e+00 -4.22117233e-01 6.99669570e-02
-1.07184017e+00 -6.34107366e-02 -2.71573633e-01 7.89920330e-01
-5.54489136e-01 3.19099307e-01 -1.58527911e-01 -1.48536229e+00
7.51297176e-02 -1.12116098e+00 1.45606685e+00 1.34655535e-01
-4.45451178e-02 -2.79473662e-01 -8.99446458e-02 2.70085543e-01
1.52601421e-01 1.60375014e-01 7.68117830e-02 -3.18615347e-01
-1.01595771e+00 -4.88415539e-01 -6.53274596e-01 3.15373838e-01
-2.98019737e-01 3.64777297e-01 -1.20743859e+00 -7.38001615e-02
-2.02833876e-01 -2.25712851e-01 1.39478958e+00 3.28667521e-01
9.37314928e-01 3.65340501e-01 -7.01481104e-01 2.36397132e-01
1.35454392e+00 -8.83529484e-02 3.77285540e-01 2.58626074e-01
6.02496564e-01 5.53341568e-01 1.30412221e+00 4.94913340e-01
3.22771698e-01 5.97397387e-01 6.48796797e-01 4.86383706e-01
-5.58972964e-03 -3.26107025e-01 2.88758516e-01 1.76613906e-03
1.63292494e-02 7.21638650e-02 -1.16473770e+00 5.99247873e-01
-2.10172939e+00 -5.99403262e-01 -2.41152704e-01 2.39230037e+00
5.89758098e-01 6.93413436e-01 3.69570479e-02 1.86003372e-02
6.77239835e-01 -1.58782125e-01 -5.87211013e-01 -1.66056946e-01
3.30168009e-01 -2.40669757e-01 1.30137467e+00 7.05537796e-01
-1.43208981e+00 1.31663489e+00 7.30680513e+00 7.80481875e-01
-6.78498805e-01 2.10039854e-01 4.51126099e-02 -1.79673776e-01
2.11510301e-01 5.86838946e-02 -1.70807946e+00 1.98293626e-01
8.75657797e-01 -3.57016921e-02 -3.73824835e-02 1.29582989e+00
8.34391713e-02 -6.61190927e-01 -1.01001072e+00 8.58385801e-01
-1.38992727e-01 -8.19340110e-01 -3.82656097e-01 -1.78806018e-02
7.65213907e-01 3.80476683e-01 -1.01543620e-01 5.96428394e-01
7.55280554e-01 -8.26880157e-01 1.03789926e+00 4.72681046e-01
4.36572731e-01 -8.15028369e-01 1.03813243e+00 5.16814351e-01
-1.12567377e+00 1.24751348e-02 -7.33238995e-01 9.17569324e-02
1.49181485e-01 8.69992733e-01 -1.45036530e+00 1.38258576e-01
7.94607043e-01 1.93153888e-01 -5.57452679e-01 1.30592656e+00
-6.43776774e-01 7.04445422e-01 -9.84711826e-01 -3.99729908e-02
3.12764883e-01 -7.93057159e-02 9.04316247e-01 1.14053130e+00
2.11843014e-01 2.77929213e-02 4.86643404e-01 1.13992667e+00
1.13760918e-01 -2.49630034e-01 -8.33614111e-01 2.53313690e-01
6.36035919e-01 1.17105031e+00 -7.53126860e-01 -2.30085373e-01
2.20356062e-02 2.75704265e-01 3.12255591e-01 1.02542751e-01
-1.07850158e+00 -2.21115530e-01 5.90692222e-01 6.19875677e-02
5.62787294e-01 -6.73401594e-01 -5.10280788e-01 -7.08720684e-01
1.25642553e-01 4.42668907e-02 2.20911220e-01 -6.40534997e-01
-1.03012514e+00 4.74530548e-01 6.40744865e-01 -1.20776069e+00
-3.88278872e-01 -8.74730527e-01 -3.38832974e-01 6.29854858e-01
-1.73822391e+00 -1.04905367e+00 -3.42840195e-01 -4.57942784e-02
3.50745916e-01 1.49078086e-01 4.08087075e-01 -1.02999374e-01
-5.93617857e-01 3.21263313e-01 -2.17545778e-02 -2.92299390e-01
7.08012402e-01 -1.37200940e+00 3.06915402e-01 1.02854764e+00
3.90551388e-02 3.01125705e-01 1.17098927e+00 -9.21600461e-01
-1.01945734e+00 -1.23687315e+00 4.78301823e-01 -9.81713474e-01
7.83845544e-01 -3.05657893e-01 -8.22701275e-01 6.57175958e-01
-3.73873472e-01 3.48960221e-01 1.71423405e-01 1.43070593e-01
-3.85422945e-01 -6.90922439e-02 -1.41446054e+00 4.51422870e-01
8.48277330e-01 -2.26229444e-01 -8.28863621e-01 2.42305920e-01
6.65459692e-01 -6.15858734e-01 -5.99570513e-01 9.10037577e-01
5.76213121e-01 -7.16103554e-01 8.64538014e-01 -2.99921423e-01
-2.89825678e-01 -8.97763371e-01 -3.51546109e-01 -9.25840616e-01
-2.81494260e-02 -3.84267837e-01 -3.99463981e-01 9.11131382e-01
4.80885744e-01 -5.70831776e-01 9.67997134e-01 7.37961590e-01
-2.00972080e-01 -3.74082804e-01 -1.19206643e+00 -1.20164323e+00
-1.39418304e-01 -1.04196644e+00 1.97321221e-01 1.57859743e-01
-4.27665025e-01 9.16685313e-02 -1.62277639e-01 7.13750601e-01
1.21955764e+00 -8.49228576e-02 9.10814643e-01 -1.44872034e+00
1.76131219e-01 -3.51340950e-01 -7.44486630e-01 -8.77718031e-01
2.23016649e-01 -3.09782445e-01 1.07469606e+00 -1.13706660e+00
3.35404165e-02 -8.13003004e-01 -3.38348836e-01 2.77949899e-01
-3.01027834e-01 4.21658337e-01 -6.36460036e-02 1.21787265e-01
-7.29294598e-01 4.72145915e-01 5.88604569e-01 -1.60084888e-01
-1.67062417e-01 1.88043237e-01 -3.00112456e-01 1.02876151e+00
9.01697695e-01 -9.15655255e-01 -1.72638059e-01 -4.83585745e-02
-2.31508128e-02 -3.74370068e-01 5.15889049e-01 -1.27990937e+00
2.73284286e-01 -2.46067271e-01 1.78739756e-01 -1.23345196e+00
5.33822834e-01 -9.01133358e-01 -3.78697723e-01 2.44630367e-01
8.50069895e-02 -4.68253732e-01 6.92136288e-01 9.95191097e-01
-1.73775628e-02 -6.14576280e-01 9.98706162e-01 2.40992710e-01
-9.35533881e-01 1.52065009e-01 -4.79136646e-01 -1.72541738e-01
1.20603430e+00 -2.37019002e-01 -2.03551009e-01 -1.19295880e-01
-5.37496090e-01 3.56712967e-01 3.76279712e-01 4.03728455e-01
6.54590249e-01 -1.09588611e+00 -7.44669855e-01 -5.03869429e-02
5.88979721e-01 3.38494033e-01 -3.52778554e-01 7.63916016e-01
-3.54122609e-01 5.35767734e-01 1.05765887e-01 -1.16612756e+00
-1.05142927e+00 4.92323160e-01 2.89159715e-01 8.52806121e-02
-4.76775795e-01 9.92326379e-01 -3.12890718e-03 -3.64937574e-01
3.89015049e-01 -6.43434525e-01 3.56705524e-02 -2.68231004e-01
3.66702735e-01 4.46770370e-01 -6.36510476e-02 -6.85396075e-01
-4.50088441e-01 5.87090313e-01 -2.05782205e-01 -1.48823589e-01
6.57668710e-01 -1.46987423e-01 3.38979840e-01 6.71356559e-01
5.05011261e-01 4.49167229e-02 -1.64894533e+00 -2.14696959e-01
2.89301515e-01 -6.38521910e-01 6.10303402e-01 -5.06083846e-01
-4.27775979e-01 7.43554711e-01 7.60633111e-01 -1.53159171e-01
2.94559151e-01 2.31757447e-01 3.21493894e-01 8.87013614e-01
7.45583415e-01 -1.08959556e+00 -3.82384628e-01 4.61224556e-01
8.40105414e-01 -1.96059370e+00 4.79655445e-01 -7.67835379e-01
-5.40810406e-01 9.08922911e-01 8.33086908e-01 -1.87187314e-01
6.66938663e-01 5.36161304e-01 5.50759695e-02 -6.95130005e-02
-5.74340045e-01 -6.28151178e-01 5.04687786e-01 6.19378030e-01
-8.47355202e-02 2.33309433e-01 -9.57322940e-02 3.22015435e-01
-1.08319648e-01 -1.35402426e-01 3.76299083e-01 1.07307637e+00
-1.06790137e+00 -7.29845524e-01 -6.72633588e-01 4.51285303e-01
-1.38958037e-01 1.10268176e-01 -3.10524881e-01 9.60630298e-01
1.16570882e-01 1.20558155e+00 4.59699333e-03 -3.76916200e-01
3.43764126e-01 5.22223935e-02 4.57469583e-01 -8.03550839e-01
1.80937648e-01 -1.19313478e-01 1.84134483e-01 -4.89701152e-01
-3.53222847e-01 -8.62128973e-01 -1.41339600e+00 1.53114155e-01
-8.69492412e-01 4.98575345e-03 9.77436543e-01 9.01125729e-01
4.29824740e-02 3.52531791e-01 3.29937309e-01 -1.07380772e+00
-8.95298481e-01 -1.12485766e+00 -4.51806664e-01 -3.96049887e-01
3.33394945e-01 -1.38062143e+00 -6.36061490e-01 -2.71445602e-01] | [7.580136775970459, -1.1072161197662354] |
74f04822-3ff6-4da3-9862-f334a6e64282 | semantic-guided-context-modeling-for-indoor | 2305.12661 | null | https://arxiv.org/abs/2305.12661v1 | https://arxiv.org/pdf/2305.12661v1.pdf | Semantic-guided context modeling for indoor scene recognition | Exploring the semantic context in scene images is essential for indoor scene recognition. However, due to the diverse intra-class spatial layouts and the coexisting inter-class objects, modeling contextual relationships to adapt various image characteristics is a great challenge. Existing contextual modeling methods for indoor scene recognition exhibit two limitations: 1) During training, space-independent information, such as color, may hinder optimizing the network's capacity to represent the spatial context. 2) These methods often overlook the differences in coexisting objects across different scenes, suppressing the performance of scene recognition. To address these limitations, we propose SpaCoNet, a novel approach that simultaneously models the Spatial relation and Co-occurrence of objects based on semantic segmentation. Firstly, the semantic spatial relation module (SSRM) is designed to explore the spatial relations among objects within a scene. With the help of semantic segmentation, this module decouples the spatial information from the image, effectively avoiding the influence of irrelevant features. Secondly, both spatial context features from SSRM and deep features from RGB feature extractor are used to distinguish the coexisting object across different scenes. Finally, utilizing the discriminative features mentioned above, we employ the self-attention mechanism to explore the long-range co-occurrence relationships among objects, and further generate a semantic-guided feature representation for indoor scene recognition. Experimental results on three publicly available datasets demonstrate the effectiveness and generality of the proposed method. The code will be made publicly available after the blind-review process is completed. | ['Yibin Li', 'Xin Ma', 'Hanbo Wu', 'Chuanxin Song'] | 2023-05-22 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 3.76069307e-01 -7.51946747e-01 5.76608926e-02 -6.48011029e-01
-1.57682315e-01 -4.58140165e-01 3.25398803e-01 -1.43507384e-02
-2.49692693e-01 3.14604312e-01 2.85808474e-01 -1.30093560e-01
-6.10675633e-01 -6.97715461e-01 -4.91127312e-01 -7.60374725e-01
3.23937476e-01 -3.44633549e-01 3.16807747e-01 6.56806231e-02
4.74563956e-01 6.01464748e-01 -1.71075928e+00 3.53479624e-01
1.19382000e+00 1.25046396e+00 8.48646164e-01 2.68047571e-01
-4.80028063e-01 4.73518282e-01 -5.80176830e-01 2.83092111e-01
9.24631655e-02 -4.02910829e-01 -5.16263962e-01 3.70905370e-01
3.55914861e-01 -9.58900899e-02 -4.56532240e-01 1.18713140e+00
3.30976516e-01 4.86323535e-01 3.34382474e-01 -8.79107475e-01
-8.50094318e-01 -1.32004661e-03 -2.16457874e-01 3.89055699e-01
3.37935269e-01 1.69661462e-01 7.79714227e-01 -9.82700825e-01
6.50618523e-02 1.04744744e+00 2.38452256e-01 2.24645332e-01
-9.05306816e-01 -6.10461891e-01 8.13411593e-01 5.52236795e-01
-1.74331951e+00 -3.20960969e-01 1.23963118e+00 -2.20067650e-01
6.59738779e-01 6.29104912e-01 7.66982496e-01 7.57988572e-01
-2.95068175e-01 8.76963377e-01 1.25210774e+00 -2.06213862e-01
1.97682425e-01 1.38133675e-01 1.12185679e-01 3.49337727e-01
1.75149050e-02 -1.27561972e-01 -4.68671769e-01 3.15547168e-01
6.87830746e-01 5.04567683e-01 -4.64119762e-01 -5.12152851e-01
-1.14035332e+00 2.73095310e-01 1.01115608e+00 5.79934299e-01
-1.86161697e-01 -1.96310222e-01 -1.11299589e-01 -2.66489089e-01
9.02430415e-02 2.46006623e-01 -3.21542293e-01 1.31651923e-01
-7.87563324e-01 -1.49636433e-01 8.53973404e-02 1.12467647e+00
1.04389405e+00 -2.88627625e-01 -2.00012773e-01 1.09405875e+00
4.14239109e-01 4.05371338e-01 5.47027588e-01 -5.87539673e-01
6.39958680e-01 9.69246507e-01 -4.81262505e-02 -1.42273319e+00
-4.26159322e-01 -6.26069248e-01 -8.62073779e-01 -4.17743236e-01
8.38967487e-02 3.60946715e-01 -1.02371585e+00 1.42755449e+00
3.71083468e-01 3.64078730e-01 -1.67183354e-01 1.16593182e+00
9.73457873e-01 5.23712695e-01 2.11081758e-01 -2.00453661e-02
1.11395895e+00 -9.69908237e-01 -4.39888090e-01 -5.90232253e-01
1.43697530e-01 -7.37328470e-01 1.24347997e+00 -5.80878593e-02
-4.12468225e-01 -7.99321711e-01 -1.01927912e+00 -3.26810330e-02
-7.48918176e-01 2.00771049e-01 8.25075567e-01 4.15802956e-01
-6.52611196e-01 1.35077134e-01 -4.46034431e-01 -3.89329582e-01
5.32930493e-01 3.94888610e-01 -1.46258578e-01 -5.27469158e-01
-1.00363898e+00 2.14886069e-01 5.65441310e-01 6.45773649e-01
-5.42486489e-01 -3.56590301e-01 -8.06772292e-01 4.63176668e-02
4.33973581e-01 -4.81965035e-01 7.06506550e-01 -1.19651544e+00
-1.08508372e+00 4.39358026e-01 -3.99892837e-01 3.98118883e-01
1.81501418e-01 -1.33515015e-01 -6.37905598e-01 -9.09008309e-02
3.57278228e-01 4.32834268e-01 6.03670955e-01 -1.48212755e+00
-8.01577449e-01 -4.71145630e-01 2.00732082e-01 6.85849309e-01
-3.57238322e-01 -2.44652301e-01 -1.06300128e+00 -7.01988220e-01
7.76183963e-01 -7.67748117e-01 -2.62193233e-01 -2.71522164e-01
-7.06195414e-01 6.13109395e-02 9.03805315e-01 -3.25753570e-01
1.17519557e+00 -2.44161344e+00 7.19946483e-03 3.50353509e-01
-2.87193507e-01 2.16001838e-01 -1.88245714e-01 2.10161824e-02
-4.45179306e-02 -5.13009056e-02 -3.88020486e-01 1.74387284e-02
-2.01531500e-01 3.77228588e-01 -1.61727771e-01 2.95928240e-01
7.00576007e-02 8.37233424e-01 -9.80773926e-01 -4.16534901e-01
6.98372245e-01 3.85378212e-01 -4.53471184e-01 2.58247465e-01
-5.04820198e-02 8.46742034e-01 -8.04606974e-01 7.07034051e-01
8.18368733e-01 -1.56094968e-01 4.01339978e-02 -3.90667468e-01
-3.24783102e-02 2.77225912e-01 -1.51402843e+00 1.80762529e+00
-5.90371370e-01 3.91466409e-01 -1.89060539e-01 -1.00295556e+00
8.44616711e-01 -1.93564340e-01 2.93478042e-01 -1.12779498e+00
-4.30342481e-02 1.50892839e-01 -1.06360748e-01 -6.36397421e-01
3.35750401e-01 2.60512531e-01 -7.39922598e-02 -5.02545387e-02
-2.19368115e-01 1.32394657e-01 -1.31333292e-01 -8.96885172e-02
6.01470649e-01 -2.45052408e-02 5.16920500e-02 -1.59253478e-01
7.92331636e-01 -2.28059515e-01 7.01370537e-01 6.88205183e-01
-1.88003391e-01 6.45049334e-01 -1.76052436e-01 -4.51653987e-01
-4.79946554e-01 -1.06185651e+00 -3.13266844e-01 9.33399439e-01
1.03301048e+00 -3.54981303e-01 -4.87357289e-01 -5.55419803e-01
-1.16599172e-01 4.73900706e-01 -5.41249216e-01 -2.57944375e-01
-3.84285897e-01 -8.19390595e-01 -1.84634089e-01 6.31234407e-01
9.75449443e-01 -1.01445711e+00 -5.40335953e-01 -5.73975639e-03
-3.36995304e-01 -1.20608342e+00 -4.89015698e-01 7.47177005e-02
-5.66245198e-01 -1.17022383e+00 -4.38030511e-01 -9.52304661e-01
1.00464189e+00 9.73516047e-01 5.95541298e-01 3.89532566e-01
-2.83685565e-01 4.86345381e-01 -2.67114252e-01 -2.94607952e-02
4.95035946e-01 -7.42134601e-02 -1.16473518e-01 4.41475689e-01
4.04539526e-01 -5.80616117e-01 -9.71182048e-01 5.15553057e-01
-9.14777339e-01 4.13244873e-01 5.30901670e-01 7.55373001e-01
8.34131360e-01 6.00075603e-01 1.60666272e-01 -6.28152549e-01
2.09471002e-01 -4.03401375e-01 -3.46312284e-01 3.61986846e-01
-1.98122159e-01 -2.97990113e-01 6.52285039e-01 -3.29465657e-01
-1.18832290e+00 1.10847704e-01 8.36848170e-02 -2.29507193e-01
-5.26581943e-01 4.14115191e-01 -1.03554642e+00 -1.09795071e-01
1.18425086e-01 6.56821370e-01 -6.67486072e-01 -4.35225934e-01
3.48400772e-01 6.73304856e-01 5.22804260e-01 -5.51834106e-01
7.45866776e-01 4.29395080e-01 -2.22298071e-01 -8.57806206e-01
-1.03334761e+00 -7.59167910e-01 -8.16342056e-01 -8.92119035e-02
8.86658609e-01 -1.00074399e+00 -3.76576930e-01 5.64197779e-01
-8.59666228e-01 -2.04812363e-01 -5.85531816e-02 4.18573588e-01
-5.38654961e-02 3.14803123e-01 -1.40716493e-01 -6.98759258e-01
2.47709021e-01 -1.21638596e+00 9.72323716e-01 7.56836116e-01
1.39625043e-01 -7.89314806e-01 -4.70282197e-01 5.70705056e-01
3.25653523e-01 -1.54319346e-01 9.59235609e-01 -2.90903777e-01
-1.14420819e+00 -8.75483602e-02 -6.80016756e-01 1.88190937e-01
7.38680959e-01 -2.02594638e-01 -1.03263676e+00 3.36227790e-02
3.00257839e-02 1.73057958e-01 7.89654970e-01 2.46882215e-01
1.80075598e+00 -3.16371560e-01 -3.81449670e-01 9.22666430e-01
1.41184163e+00 5.17863035e-01 5.54293692e-01 5.56146979e-01
1.25423980e+00 8.04847240e-01 7.74516582e-01 2.35203296e-01
5.10342896e-01 7.41534650e-01 4.70556438e-01 -4.98126969e-02
-1.69355646e-01 -4.27841485e-01 -1.09552689e-01 7.59065986e-01
1.53094679e-01 6.75264969e-02 -8.08652520e-01 5.42338073e-01
-1.81591451e+00 -6.99274123e-01 3.14284960e-04 2.06867099e+00
4.84997869e-01 -2.71975137e-02 -4.09693450e-01 -4.68576103e-02
6.69080794e-01 3.39003444e-01 -7.25369275e-01 1.81503326e-01
-3.24717700e-01 -2.11043358e-01 3.47164363e-01 2.86469340e-01
-1.33987701e+00 9.61350679e-01 4.66705656e+00 1.01810563e+00
-1.22034407e+00 -1.33563384e-01 6.49811924e-01 -2.63486244e-02
-4.10563529e-01 8.22448656e-02 -5.62194228e-01 7.73059845e-01
1.12739056e-01 2.93384433e-01 6.25831902e-01 8.33975494e-01
8.59706774e-02 -2.11608246e-01 -7.61904120e-01 1.19764340e+00
3.83560322e-02 -1.00208521e+00 6.53711930e-02 -1.76826224e-01
8.50628138e-01 -1.41657174e-01 2.33062953e-01 1.44296184e-01
-8.76061469e-02 -9.07333910e-01 7.65423000e-01 5.76253891e-01
5.88762522e-01 -7.23157525e-01 4.55891609e-01 3.47879231e-01
-1.69159305e+00 -3.90209973e-01 -4.31383759e-01 6.94621801e-02
-1.95516452e-01 4.96982634e-01 -2.94092804e-01 6.87409997e-01
9.03374434e-01 9.03785706e-01 -1.05993438e+00 1.17959535e+00
-4.56336558e-01 1.73346996e-01 -2.14963987e-01 7.13077262e-02
2.43112788e-01 -2.78411508e-01 3.79599601e-01 1.10830855e+00
2.02950761e-01 1.34902745e-01 3.50813389e-01 8.89816105e-01
2.22735956e-01 3.89760025e-02 -3.79585624e-01 3.31417173e-01
5.69392145e-01 1.08879614e+00 -8.79851699e-01 -2.06313029e-01
-4.22380865e-01 1.10911727e+00 1.47102401e-01 9.36316848e-01
-6.67631924e-01 -3.41981918e-01 6.66160166e-01 -1.78287968e-01
4.42923039e-01 -4.46240038e-01 -5.06180942e-01 -1.20236051e+00
3.79003167e-01 -6.91863537e-01 2.22619906e-01 -8.76347840e-01
-1.17410052e+00 4.61683869e-01 -1.28403366e-01 -1.32180512e+00
5.06350040e-01 -5.23711860e-01 -6.45670176e-01 9.59088326e-01
-1.58254623e+00 -1.25463760e+00 -7.24469841e-01 8.20256829e-01
6.82943583e-01 5.49927093e-02 5.37440240e-01 3.44695091e-01
-9.13691401e-01 4.50497180e-01 9.33585763e-02 1.99930221e-01
3.09536308e-01 -1.03299344e+00 -5.56490757e-02 1.11723006e+00
3.38894844e-01 9.75160956e-01 1.83405176e-01 -5.25042415e-01
-1.24222946e+00 -1.44844568e+00 5.34616232e-01 -2.52289206e-01
3.17986131e-01 -4.75529075e-01 -1.07398248e+00 2.93114096e-01
-3.21929663e-01 2.95568734e-01 8.54332507e-01 2.14320838e-01
-4.19924170e-01 -5.22811890e-01 -6.76570714e-01 7.99515903e-01
1.37760890e+00 -9.72534060e-01 -4.42475140e-01 3.07372451e-01
6.48990095e-01 -3.36925060e-01 -3.74474347e-01 6.43413186e-01
2.31085360e-01 -1.03215408e+00 1.25480556e+00 -1.00196816e-01
1.68530926e-01 -8.42073560e-01 -5.07909358e-01 -1.06056499e+00
-4.43103224e-01 1.02725677e-01 2.13603467e-01 1.41637981e+00
-1.33757722e-02 -6.89465344e-01 6.33804381e-01 8.12182307e-01
-2.25176856e-01 -6.88051343e-01 -9.48976040e-01 -6.72462881e-01
-4.67540383e-01 -6.55785799e-01 1.00439250e+00 9.38101172e-01
-4.93073374e-01 1.89241052e-01 -2.95527484e-02 5.46032608e-01
3.80638868e-01 6.93835258e-01 6.33898199e-01 -7.22443998e-01
-1.42622381e-01 -5.23358583e-01 -5.10394514e-01 -1.40895748e+00
6.36804476e-02 -8.21155965e-01 2.32481077e-01 -1.70563543e+00
2.87855297e-01 -9.79305625e-01 -8.65263283e-01 3.21750224e-01
-5.74611843e-01 1.83485210e-01 2.40158334e-01 4.39480901e-01
-9.96127903e-01 8.71003389e-01 1.34788775e+00 -4.09073085e-01
-2.46532589e-01 -3.22676972e-02 -7.97151446e-01 6.87209547e-01
6.46325231e-01 -4.26761769e-02 -6.45603716e-01 -6.41607642e-01
-2.18463182e-01 -4.17156994e-01 8.20313931e-01 -1.34487987e+00
4.24537718e-01 -4.97274995e-01 8.47817481e-01 -6.26269221e-01
3.03820044e-01 -1.20287395e+00 1.03237994e-01 6.63387030e-02
-3.94359715e-02 -3.65428180e-01 2.46052042e-01 7.81326890e-01
-4.04216856e-01 3.92971821e-02 4.88515168e-01 -9.91398916e-02
-1.30485916e+00 4.53539729e-01 -1.01406001e-01 -1.66270062e-01
9.18129683e-01 -5.24520695e-01 -1.57245472e-01 -8.15484747e-02
-4.69539195e-01 2.96281189e-01 6.46981001e-01 7.29935288e-01
8.17046940e-01 -1.41877401e+00 4.16459367e-02 5.66976309e-01
3.78392756e-01 4.89841372e-01 9.17395830e-01 5.65135539e-01
-2.62458354e-01 5.00537097e-01 4.61516082e-02 -8.05621922e-01
-1.09219682e+00 6.69348478e-01 4.56482083e-01 3.47146302e-01
-4.20332223e-01 9.52894270e-01 9.14808989e-01 -4.98300165e-01
1.27858311e-01 -3.38591188e-01 -3.80361110e-01 -1.53277159e-01
3.76975507e-01 -5.43670319e-02 -1.30564019e-01 -9.16072607e-01
-6.00715339e-01 9.69998181e-01 7.11262226e-02 4.22973067e-01
1.02648354e+00 -6.77898943e-01 -2.02855498e-01 5.16251028e-01
1.25510097e+00 5.04411059e-03 -1.46788430e+00 -4.86452699e-01
-2.96535399e-02 -9.51350510e-01 -1.09981466e-02 -8.03741157e-01
-1.04525566e+00 9.29403365e-01 7.34656870e-01 -5.17969243e-02
1.62803495e+00 -7.01032281e-02 5.59503198e-01 1.36157304e-01
4.30440217e-01 -1.13286555e+00 1.54823244e-01 4.35843766e-01
6.76760852e-01 -1.33824289e+00 3.31485495e-02 -6.72056198e-01
-4.66485351e-01 8.77079546e-01 9.12173331e-01 1.83674797e-01
6.07307017e-01 -3.82137299e-01 9.48257968e-02 -4.34006900e-02
3.10173631e-02 -2.85676330e-01 7.62672305e-01 5.94962537e-01
8.57555047e-02 2.52442569e-01 2.80275077e-01 7.22753525e-01
-8.76355544e-02 -4.55155343e-01 -1.57435775e-01 8.41767907e-01
-3.49965334e-01 -9.83488977e-01 -3.31561595e-01 1.89664111e-01
7.23762363e-02 -2.08348840e-01 -3.32546651e-01 4.25164878e-01
6.64562285e-01 1.16322362e+00 1.13258995e-01 -5.04858315e-01
4.49185580e-01 -3.18900436e-01 1.93669707e-01 -6.66510224e-01
-1.16054766e-01 8.69946703e-02 -4.44495499e-01 -6.56251729e-01
-5.24262071e-01 -4.85985518e-01 -1.18704569e+00 1.55724913e-01
-1.92284420e-01 6.89403620e-03 6.32930338e-01 1.03544796e+00
4.66456234e-01 7.62760341e-01 8.66877437e-01 -7.20973790e-01
3.02263707e-01 -5.52772343e-01 -5.82756221e-01 5.08453846e-01
3.98357272e-01 -8.11384797e-01 -5.99546432e-02 -1.15569457e-01] | [9.487125396728516, -0.7156233787536621] |
da8a569e-7ba1-4e09-8096-4b330f5f522d | elf-an-extensive-lightweight-and-flexible | 1707.01067 | null | http://arxiv.org/abs/1707.01067v2 | http://arxiv.org/pdf/1707.01067v2.pdf | ELF: An Extensive, Lightweight and Flexible Research Platform for Real-time Strategy Games | In this paper, we propose ELF, an Extensive, Lightweight and Flexible
platform for fundamental reinforcement learning research. Using ELF, we
implement a highly customizable real-time strategy (RTS) engine with three game
environments (Mini-RTS, Capture the Flag and Tower Defense). Mini-RTS, as a
miniature version of StarCraft, captures key game dynamics and runs at 40K
frame-per-second (FPS) per core on a Macbook Pro notebook. When coupled with
modern reinforcement learning methods, the system can train a full-game bot
against built-in AIs end-to-end in one day with 6 CPUs and 1 GPU. In addition,
our platform is flexible in terms of environment-agent communication
topologies, choices of RL methods, changes in game parameters, and can host
existing C/C++-based game environments like Arcade Learning Environment. Using
ELF, we thoroughly explore training parameters and show that a network with
Leaky ReLU and Batch Normalization coupled with long-horizon training and
progressive curriculum beats the rule-based built-in AI more than $70\%$ of the
time in the full game of Mini-RTS. Strong performance is also achieved on the
other two games. In game replays, we show our agents learn interesting
strategies. ELF, along with its RL platform, is open-sourced at
https://github.com/facebookresearch/ELF. | ['Yuandong Tian', 'Yuxin Wu', 'Wenling Shang', 'Qucheng Gong', 'C. Lawrence Zitnick'] | 2017-07-04 | elf-an-extensive-lightweight-and-flexible-1 | http://papers.nips.cc/paper/6859-elf-an-extensive-lightweight-and-flexible-research-platform-for-real-time-strategy-games | http://papers.nips.cc/paper/6859-elf-an-extensive-lightweight-and-flexible-research-platform-for-real-time-strategy-games.pdf | neurips-2017-12 | ['real-time-strategy-games'] | ['playing-games'] | [-5.48693776e-01 -2.61461467e-01 -1.13907252e-02 1.84363306e-01
-4.38387632e-01 -9.00305867e-01 3.16542149e-01 -3.55564773e-01
-1.12177682e+00 7.08239675e-01 -3.12200397e-01 -7.81323433e-01
-2.77239770e-01 -8.95709813e-01 -6.33406162e-01 -4.94906723e-01
-5.71488857e-01 7.08945096e-01 6.31984293e-01 -1.15641308e+00
6.60641789e-02 2.99941301e-01 -1.49305749e+00 -1.09899648e-01
3.93083215e-01 6.39295220e-01 3.79837245e-01 1.44896269e+00
3.86124521e-01 1.50292408e+00 -7.70432055e-01 -2.48473272e-01
7.76701748e-01 -6.39965758e-02 -6.02076888e-01 -5.01193047e-01
-2.14713693e-01 -6.35224938e-01 -9.17216420e-01 6.27776086e-01
9.83034730e-01 3.70213896e-01 -2.22802907e-01 -1.48935807e+00
2.95237839e-01 7.21663177e-01 -4.89510298e-01 5.11946619e-01
1.66423187e-01 8.87972236e-01 6.31883085e-01 -9.07319114e-02
4.58147883e-01 9.82684791e-01 6.42534733e-01 6.64877117e-01
-8.72380614e-01 -1.02817249e+00 1.08425476e-01 6.14636391e-02
-1.11424172e+00 -2.99152046e-01 2.99289852e-01 1.41935088e-02
1.26990449e+00 1.49225384e-01 9.20668125e-01 1.14259815e+00
3.96292210e-01 8.98051679e-01 9.37661231e-01 -1.08625451e-02
5.92570961e-01 -5.41001797e-01 -2.67557472e-01 9.44777191e-01
-2.91396528e-01 7.02809095e-01 -3.42381716e-01 -1.12610891e-01
1.25115442e+00 -1.65549695e-01 1.93407480e-02 -7.91266039e-02
-9.22876537e-01 9.57577050e-01 4.14435297e-01 -2.92784244e-01
-3.12020063e-01 8.89540672e-01 6.38963699e-01 7.07803071e-01
-1.50065180e-02 6.30559146e-01 -4.89118427e-01 -1.07186449e+00
-4.31685269e-01 7.97073543e-01 9.46455598e-01 9.89872694e-01
5.06988645e-01 5.74506938e-01 2.71267623e-01 6.29155278e-01
-3.87746692e-02 3.25376034e-01 4.54454362e-01 -1.73455191e+00
1.91306785e-01 1.75053671e-01 -1.65683348e-02 -5.80851376e-01
-8.68562341e-01 -7.46914625e-01 -3.13776165e-01 9.20350671e-01
5.07442653e-01 -8.10820818e-01 -4.48152810e-01 1.80137014e+00
3.60647649e-01 5.34874022e-01 2.40909830e-01 9.33485627e-01
5.09068847e-01 5.54726243e-01 -2.09219173e-01 1.47312671e-01
1.25683033e+00 -1.21085095e+00 2.03808211e-02 -5.45908213e-01
7.77007163e-01 -4.02073413e-01 1.16521943e+00 6.75103784e-01
-1.29620564e+00 -2.16021389e-01 -9.54237640e-01 3.79791170e-01
-1.92723885e-01 -4.94451433e-01 1.10123324e+00 5.13189495e-01
-1.29454887e+00 7.14716196e-01 -1.17987633e+00 -2.69488096e-01
1.78084895e-01 6.36163652e-01 -5.86858839e-02 1.50483906e-01
-9.79473829e-01 7.21877456e-01 6.08158529e-01 -3.81837547e-01
-1.31046414e+00 -6.00059688e-01 -5.93855023e-01 -4.09302860e-03
1.00090098e+00 -7.20280468e-01 1.82796025e+00 -6.12318218e-01
-2.21668482e+00 4.89754438e-01 8.91577423e-01 -7.13197827e-01
5.37724674e-01 -2.81910766e-02 -1.42273098e-01 8.08149576e-03
-1.02209568e-01 7.39746928e-01 4.65180486e-01 -7.45680630e-01
-9.76571977e-01 -7.58031234e-02 8.24970722e-01 6.14906847e-01
1.31710261e-01 2.58392811e-01 -3.59236419e-01 -3.88678819e-01
-6.75781071e-01 -1.10910463e+00 -6.02213800e-01 -3.77548695e-01
3.37145120e-01 -4.09757309e-02 6.92991078e-01 -1.09886765e-01
8.09965611e-01 -2.04000163e+00 2.70308200e-02 7.33223334e-02
2.32811153e-01 1.91849500e-01 -2.30918288e-01 4.10527289e-01
2.60917038e-01 -3.39289844e-01 1.85963199e-01 7.28220046e-02
3.38573903e-01 7.91441917e-01 -2.40584090e-01 2.69447893e-01
-3.74010086e-01 8.56171191e-01 -1.14477801e+00 -9.08244476e-02
2.38000348e-01 2.18665204e-03 -1.10411763e+00 2.23271608e-01
-2.08850056e-01 2.29360268e-01 -3.69731337e-01 3.54367763e-01
2.32964933e-01 1.36542320e-01 2.25235552e-01 6.74568594e-01
-4.23745424e-01 2.59559602e-01 -1.24779475e+00 2.10200739e+00
-4.39704180e-01 4.66147512e-01 4.82111037e-01 -7.25780606e-01
5.73633611e-01 -6.96019754e-02 4.58391696e-01 -9.29062605e-01
4.94173199e-01 8.95283818e-02 2.20283970e-01 -7.93948025e-02
8.14471483e-01 3.12843323e-01 -4.59835559e-01 5.44585228e-01
3.10104311e-01 -4.41285223e-01 7.16101527e-01 4.56106037e-01
1.83335841e+00 3.71254802e-01 7.54251853e-02 -2.32706279e-01
-2.23355219e-01 3.31708252e-01 6.25013351e-01 1.14472175e+00
-3.61489713e-01 -1.85811460e-01 6.31151795e-01 -4.43344116e-01
-8.91496181e-01 -1.12944663e+00 4.81160641e-01 1.98059416e+00
1.84482858e-01 -7.82114148e-01 -6.67468667e-01 -3.20521742e-01
-1.25044715e-02 7.13520885e-01 -3.25896025e-01 -1.25069380e-01
-6.62879527e-01 -6.11622155e-01 1.09366012e+00 4.41698551e-01
7.59822965e-01 -1.27808034e+00 -1.19629955e+00 5.16364992e-01
1.47454754e-01 -8.59970927e-01 -4.48551238e-01 5.34269154e-01
-4.50853258e-01 -1.07133687e+00 -9.95574743e-02 -6.10887885e-01
-1.65758654e-01 5.09547055e-01 1.15098929e+00 1.57553121e-01
-3.26453954e-01 4.86722589e-01 -4.35362577e-01 -3.64137173e-01
-2.05493435e-01 8.46658275e-02 1.55661434e-01 -9.32829797e-01
-1.76087633e-01 -8.86879981e-01 -5.41505992e-01 3.34672272e-01
-6.49501801e-01 1.47528544e-01 2.04103157e-01 8.36307526e-01
1.67818502e-01 2.68157691e-01 2.32754752e-01 -4.23259765e-01
8.90187860e-01 -5.41858673e-01 -1.09737420e+00 -3.40080738e-01
-2.85287380e-01 -7.98476338e-02 7.48099029e-01 -4.53490227e-01
-6.17867112e-01 -4.33819652e-01 -3.39763701e-01 -3.77637148e-01
-2.53113639e-02 3.81101817e-01 4.81124312e-01 -3.82152438e-01
1.02830589e+00 2.67448634e-01 2.55826026e-01 1.59581572e-01
2.72521347e-01 3.46952647e-01 7.98286259e-01 -9.69869852e-01
6.96513176e-01 9.47314501e-02 -3.62663358e-01 -5.15891790e-01
-2.22591922e-01 -2.09790781e-01 3.40333998e-01 -4.87913698e-01
4.48073059e-01 -1.14905357e+00 -1.63894999e+00 7.50542402e-01
-4.45169032e-01 -1.55312824e+00 -4.84092206e-01 4.31398571e-01
-9.88628507e-01 -1.55003503e-01 -1.00752366e+00 -6.16720021e-01
-3.12430799e-01 -1.25059760e+00 6.26447439e-01 7.14925289e-01
1.61688656e-01 -7.79142022e-01 4.02481914e-01 3.01519454e-01
7.89929688e-01 3.20135467e-02 1.52211249e-01 -4.67915207e-01
-3.16503227e-01 1.17700875e-01 1.61822781e-01 -1.00107692e-01
-4.73242074e-01 -4.98316400e-02 -5.04378676e-01 -7.25421727e-01
-2.71511734e-01 -7.28311956e-01 4.27512914e-01 3.18252087e-01
9.20769155e-01 -2.52070963e-01 1.86276168e-01 9.59798217e-01
1.32343018e+00 4.95283037e-01 6.55955791e-01 8.67486775e-01
2.99510300e-01 -3.66278030e-02 6.81869805e-01 8.28238785e-01
6.02144301e-01 6.70955360e-01 8.79280686e-01 8.54349881e-02
3.76509517e-01 -8.31715241e-02 8.30483496e-01 6.03377640e-01
-4.18928057e-01 -2.50050575e-01 -9.58042741e-01 5.92656201e-03
-2.04308558e+00 -1.34247160e+00 3.12302500e-01 1.94381595e+00
7.37533867e-01 6.24241650e-01 8.79527867e-01 -2.85704076e-01
1.15669042e-01 2.62112543e-03 -7.70468354e-01 -6.65291131e-01
-2.29140129e-02 6.23899817e-01 9.49701667e-01 6.43470764e-01
-8.47837925e-01 1.56345284e+00 6.00520849e+00 1.24454904e+00
-1.18985486e+00 1.56329975e-01 7.83438161e-02 -7.13770390e-01
2.14503959e-01 -8.54643136e-02 -4.50722009e-01 3.50338370e-01
1.18050635e+00 -5.21328092e-01 1.39877963e+00 1.15162611e+00
1.86978027e-01 -2.58333623e-01 -4.10450459e-01 8.50865364e-01
-3.96068931e-01 -1.51067603e+00 -7.54180729e-01 3.54214132e-01
3.17305535e-01 7.89410710e-01 -4.62716855e-02 1.28645575e+00
1.64582503e+00 -9.85337317e-01 7.58000672e-01 -1.10075347e-01
6.72141016e-01 -1.35056698e+00 6.14561141e-01 6.42044008e-01
-1.09087360e+00 -4.51568544e-01 -4.31998730e-01 -7.58746982e-01
-1.34957939e-01 -4.07255441e-01 -7.52952754e-01 4.16004270e-01
9.41397369e-01 3.46082568e-01 -2.90631264e-01 8.51100266e-01
-2.08777964e-01 6.46071434e-01 -6.79191053e-01 -1.24970965e-01
6.94053948e-01 -1.38555318e-01 5.50675869e-01 1.01037180e+00
6.14033788e-02 4.42429930e-01 6.35835707e-01 3.81847948e-01
1.46058902e-01 -2.43604481e-01 -4.68018711e-01 2.09604189e-01
4.54296768e-01 1.52977300e+00 -7.10587502e-01 -4.32995185e-02
-3.32184811e-03 7.20712721e-01 5.13551891e-01 7.64135048e-02
-1.33323169e+00 -5.40338874e-01 1.16570199e+00 -1.92401946e-01
1.34350345e-01 -5.93610406e-01 2.17419267e-01 -1.07014048e+00
-5.82871914e-01 -1.22113204e+00 4.79034573e-01 -8.34386349e-01
-7.91450679e-01 5.65958977e-01 -1.47407547e-01 -9.01568353e-01
-5.74814498e-01 -7.28950083e-01 -8.58382940e-01 2.23852485e-01
-1.11858916e+00 -6.65269017e-01 -3.95221293e-01 1.07981205e+00
5.77244401e-01 -5.48095882e-01 8.70264411e-01 2.11910918e-01
-8.48816454e-01 6.85992658e-01 1.83065668e-01 1.13035843e-01
3.69889438e-01 -1.45704424e+00 7.39275038e-01 7.43417919e-01
-2.54974663e-01 2.15681314e-01 8.34129035e-01 -3.43635559e-01
-1.74032569e+00 -8.17523539e-01 -5.35756886e-01 -6.63611144e-02
1.12162673e+00 -3.38371187e-01 -3.31804037e-01 8.32753420e-01
3.37618113e-01 -9.66142416e-02 4.93058443e-01 9.02896449e-02
-1.46952763e-01 -1.34794876e-01 -9.42283034e-01 1.22482479e+00
1.15524173e+00 -1.01819724e-01 -4.55414057e-02 3.50600898e-01
7.51753807e-01 -1.10470200e+00 -6.69618368e-01 -1.06103532e-02
5.43981791e-01 -1.08214152e+00 8.85870337e-01 -6.52758002e-01
9.13962498e-02 -2.36846045e-01 -2.76903987e-01 -1.65753639e+00
-5.85191131e-01 -1.20190620e+00 8.67570788e-02 5.02362847e-01
-4.38223928e-02 -8.37806284e-01 1.08352017e+00 1.38794601e-01
-4.51200217e-01 -7.54132152e-01 -8.80003870e-01 -9.62618232e-01
2.19468355e-01 -8.03811669e-01 3.25125694e-01 6.04855597e-01
3.72891307e-01 1.98128968e-01 -4.63161260e-01 4.86113243e-02
5.75625300e-01 -2.46995881e-01 1.33237875e+00 -7.04142332e-01
-1.08217728e+00 -7.01750278e-01 -3.73653919e-01 -1.03074086e+00
8.59020576e-02 -7.62670040e-01 -3.52394469e-02 -9.76259708e-01
-2.30348945e-01 -5.96360266e-01 -1.35762066e-01 7.94948399e-01
2.89632231e-01 1.13729902e-01 6.12696290e-01 -9.49638188e-02
-9.68983650e-01 4.41958129e-01 9.37281013e-01 1.57938823e-01
-2.62751102e-01 -1.92589350e-02 -6.81296706e-01 8.52127254e-01
1.30498922e+00 -3.97058457e-01 -4.43630427e-01 -3.76234531e-01
2.52214372e-01 3.92154992e-01 3.24641913e-01 -1.39084315e+00
4.62985277e-01 -5.91817379e-01 -1.27601400e-01 4.38379534e-02
4.55647588e-01 -5.57378173e-01 5.55731282e-02 7.69201458e-01
-8.71888399e-02 5.79506040e-01 5.92456639e-01 1.99285462e-01
3.99787396e-01 -6.42632786e-03 6.85186207e-01 -3.56463611e-01
-9.54013288e-01 2.79221624e-01 -9.16966081e-01 5.39039910e-01
1.29383731e+00 -1.33992687e-01 -8.83710206e-01 -6.85831904e-01
-3.95933360e-01 8.47144544e-01 5.26453793e-01 2.77552903e-01
3.48981380e-01 -9.77692962e-01 -7.31877267e-01 1.25438245e-02
-2.66586721e-01 -9.52409804e-02 4.63129878e-01 6.11923397e-01
-1.21678364e+00 -1.93002135e-01 -7.39953637e-01 -2.90091813e-01
-1.17591178e+00 4.21277791e-01 6.64410472e-01 -6.18086517e-01
-6.48778975e-01 9.36091661e-01 -7.88921565e-02 -8.56739163e-01
2.36366242e-01 1.73770562e-01 1.50858566e-01 -5.64548552e-01
6.81602478e-01 4.36673403e-01 -2.66072571e-01 8.16242397e-02
-3.23972285e-01 -5.21233119e-02 -3.26527357e-02 -4.41695958e-01
1.45131648e+00 3.00545722e-01 3.05302143e-01 6.68400377e-02
4.40989017e-01 -6.92139864e-02 -1.79120553e+00 -1.65067371e-02
-4.69703615e-01 -2.08114803e-01 2.29061827e-01 -8.17881942e-01
-1.13765919e+00 3.53790343e-01 4.95674759e-01 2.30267849e-02
1.09288108e+00 -4.73788559e-01 6.57198548e-01 6.29486084e-01
1.06328714e+00 -1.28376269e+00 2.99161226e-01 1.18688226e+00
5.91819525e-01 -8.43927562e-01 -2.80848682e-01 2.07839683e-01
-8.04585099e-01 1.03664601e+00 1.01320004e+00 -7.27553308e-01
1.82820275e-01 9.74684000e-01 8.59667137e-02 -1.93114340e-01
-1.30748081e+00 -3.85404170e-01 -8.62841725e-01 7.49903023e-01
-2.54471153e-01 2.16389090e-01 1.99125111e-01 4.65780318e-01
-9.25289035e-01 -1.66576710e-02 9.54624057e-01 1.35389566e+00
-6.91425860e-01 -8.59120846e-01 -2.11055353e-01 9.61142480e-02
-3.56265008e-01 1.66220851e-02 9.28422436e-02 9.81733918e-01
-8.96501541e-02 9.62184727e-01 2.06490949e-01 -7.74600387e-01
4.25668627e-01 -4.81520534e-01 6.17269039e-01 -3.61271530e-01
-1.24599338e+00 -9.12419036e-02 1.96398482e-01 -1.08913922e+00
3.72436613e-01 -3.33013892e-01 -1.48698354e+00 -1.41046107e+00
7.51954690e-02 1.63195089e-01 5.05738258e-01 6.05882525e-01
3.21360201e-01 8.05099368e-01 5.06030619e-01 -1.21621287e+00
-6.12841189e-01 -5.35769165e-01 -6.87411845e-01 -2.53987014e-01
-1.78405836e-01 -5.84783733e-01 -1.53093398e-01 -7.96493411e-01] | [3.700322151184082, 1.4646884202957153] |
27b407e6-39b1-4348-badd-b7e6cc3290a5 | commonsense-knowledge-base-construction-in | 2105.01925 | null | https://arxiv.org/abs/2105.01925v1 | https://arxiv.org/pdf/2105.01925v1.pdf | Commonsense Knowledge Base Construction in the Age of Big Data | Compiling commonsense knowledge is traditionally an AI topic approached by manual labor. Recent advances in web data processing have enabled automated approaches. In this demonstration we will showcase three systems for automated commonsense knowledge base construction, highlighting each time one aspect of specific interest to the data management community. (i) We use Quasimodo to illustrate knowledge extraction systems engineering, (ii) Dice to illustrate the role that schema constraints play in cleaning fuzzy commonsense knowledge, and (iii) Ascent to illustrate the relevance of conceptual modelling. The demos are available online at https://quasimodo.r2.enst.fr, https://dice.mpi-inf.mpg.de and ascent.mpi-inf.mpg.de. | ['Simon Razniewski'] | 2021-05-05 | null | null | null | null | ['commonsense-knowledge-base-construction'] | ['knowledge-base'] | [-1.79732274e-02 2.74712950e-01 5.09572737e-02 -3.35388035e-01
-3.27665687e-01 -8.80088866e-01 8.02255690e-01 4.18144286e-01
-2.69405186e-01 9.72031891e-01 1.36145532e-01 -5.09658158e-01
-6.84192598e-01 -1.03333414e+00 -7.72330642e-01 -2.41181716e-01
-1.21563174e-01 5.53279996e-01 -1.30572960e-01 -6.09580815e-01
4.70886797e-01 4.25766468e-01 -1.81150305e+00 5.62293112e-01
1.14155459e+00 8.16634834e-01 1.65633053e-01 2.29845569e-01
-2.33613014e-01 1.03009355e+00 -6.22737646e-01 -6.58985913e-01
9.57138166e-02 -3.69267881e-01 -1.07678628e+00 -3.57224762e-01
-1.51129782e-01 5.63466847e-01 -1.06322765e-01 1.56201339e+00
6.49645105e-02 2.56140143e-01 7.92165697e-01 -1.76503515e+00
-8.45486343e-01 1.08100152e+00 1.29056722e-01 1.47644818e-01
7.78621852e-01 -5.98180145e-02 9.73293781e-01 -5.85754335e-01
1.06950808e+00 1.32122457e+00 3.51865202e-01 3.42318743e-01
-9.90953445e-01 -5.32031059e-01 -2.39854723e-01 8.53300095e-01
-1.35146618e+00 -3.51270437e-01 6.59052610e-01 -4.99629468e-01
8.18956077e-01 7.06284165e-01 7.35838950e-01 7.59952784e-01
-1.12111755e-01 5.92631102e-01 1.48970556e+00 -6.16327584e-01
4.39588249e-01 5.52207410e-01 4.65293467e-01 6.09215021e-01
4.66949046e-01 -7.18320981e-02 -4.33223695e-01 -1.84171587e-01
6.40334308e-01 -2.07240790e-01 -2.26526171e-01 -1.99995875e-01
-9.79853868e-01 9.66812491e-01 3.16731453e-01 6.25773430e-01
-3.16153556e-01 1.06202014e-01 4.06830162e-01 5.33961952e-01
8.05266872e-02 7.98102379e-01 -3.05242419e-01 -2.35773072e-01
-8.72753620e-01 5.02106428e-01 1.10620117e+00 1.37113941e+00
5.08462310e-01 -2.81443447e-01 5.53799927e-01 8.46660495e-01
6.87135160e-02 1.33967102e-01 2.86161304e-01 -1.23384166e+00
-2.55067758e-02 6.73204243e-01 2.33209163e-01 -9.87469733e-01
-3.02697092e-01 -8.18255022e-02 -5.68253636e-01 1.34396762e-01
2.67835021e-01 1.03516001e-02 -5.63279569e-01 1.51462770e+00
1.17237799e-01 -4.33239877e-01 2.67917424e-01 5.99507391e-01
8.69024217e-01 3.87303233e-01 1.21322125e-01 -3.89927328e-01
1.53800988e+00 -1.85180321e-01 -9.00233805e-01 1.01108275e-01
5.61297417e-01 -5.34914494e-01 1.12859380e+00 6.64771020e-01
-1.09343362e+00 -1.16901040e-01 -8.72215152e-01 -2.97080934e-01
-1.28678536e+00 -3.18008184e-01 8.08932722e-01 2.78407127e-01
-7.98073173e-01 6.88084722e-01 -4.48901832e-01 -6.17275000e-01
6.95878327e-01 -1.12182833e-01 -1.35058716e-01 -1.58463970e-01
-1.79881465e+00 1.29797435e+00 1.05443323e+00 -2.31329918e-01
-3.14386994e-01 -6.96902633e-01 -8.73384655e-01 -1.05642125e-01
7.97972798e-01 -6.84455276e-01 9.82028723e-01 -5.54061413e-01
-8.83951068e-01 1.07759285e+00 4.41622511e-02 -6.01950645e-01
6.60893977e-01 -1.73650712e-01 -8.38574708e-01 1.95796490e-01
1.86106175e-01 5.14478922e-01 8.76816958e-02 -1.35025740e+00
-6.56597793e-01 -2.14603499e-01 4.02402371e-01 -1.29819140e-01
6.71813414e-02 2.78345853e-01 4.73801158e-02 -7.70149052e-01
-6.86521903e-02 -5.98309815e-01 2.62948364e-01 7.56521076e-02
-4.64729488e-01 -4.39940631e-01 4.83593434e-01 -5.38665116e-01
1.49516225e+00 -1.66791534e+00 1.48895293e-01 4.11511689e-01
7.76046095e-03 -2.90589780e-02 4.55258697e-01 8.98053646e-01
-4.84948367e-01 5.52050531e-01 -6.98482037e-01 4.93626237e-01
4.85882849e-01 3.08684349e-01 -2.78839886e-01 -1.12073697e-01
9.82261598e-02 7.89239109e-01 -1.10146999e+00 -7.92516112e-01
3.29097718e-01 9.91701558e-02 -2.31565654e-01 -1.96821079e-01
-3.87112588e-01 -2.73927808e-01 -1.95082217e-01 7.34926701e-01
3.56927693e-01 -4.70348708e-02 2.48534635e-01 -3.63309741e-01
-2.75787413e-01 2.87677735e-01 -1.35883260e+00 1.51007259e+00
-2.14959055e-01 5.66568255e-01 -6.53265696e-03 -7.75937259e-01
8.45253408e-01 1.22948602e-01 3.78881246e-01 -1.42592311e-01
1.94242388e-01 2.03672111e-01 3.34793441e-02 -6.34630859e-01
4.89242733e-01 -6.95157588e-01 -1.74404100e-01 3.45644563e-01
5.88926747e-02 -7.96086848e-01 9.19902384e-01 4.22137171e-01
8.95959437e-01 2.18298137e-01 7.07126319e-01 -6.78245068e-01
3.61489117e-01 6.32337391e-01 5.30819058e-01 2.15340361e-01
-8.75198655e-03 5.84096983e-02 7.23992884e-01 -2.80015349e-01
-1.02962768e+00 -9.10005748e-01 -5.51844418e-01 8.90769601e-01
9.32511017e-02 -6.62155926e-01 -6.61034048e-01 -4.16213125e-01
8.96641389e-02 1.85540140e+00 -6.75546706e-01 2.51148343e-01
-3.69507335e-02 -3.97440612e-01 7.70617366e-01 2.51258612e-01
2.11425513e-01 -1.21071720e+00 -8.23046207e-01 4.97827455e-02
-5.61335802e-01 -8.60901475e-01 2.25047290e-01 3.13720644e-01
-5.52571297e-01 -1.46183157e+00 1.29714623e-01 -3.08628172e-01
3.03910583e-01 -1.60153851e-01 1.40104723e+00 6.81450069e-02
-5.09536147e-01 3.16248655e-01 -6.59852445e-01 -8.24401855e-01
-7.35878766e-01 -3.97073597e-01 9.11659449e-02 -6.33294761e-01
7.49105513e-01 -7.25713730e-01 3.50470170e-02 1.49610803e-01
-1.34562540e+00 1.21004887e-01 -6.12433217e-02 5.38026214e-01
5.97654402e-01 5.44590950e-01 4.78489727e-01 -1.13986826e+00
9.15440083e-01 -7.89963365e-01 -3.41175169e-01 4.60832030e-01
-6.35420203e-01 -1.27698824e-01 4.99553859e-01 9.58916992e-02
-8.45076442e-01 -3.53275985e-01 1.39521360e-01 -3.14023226e-01
-4.52489167e-01 9.35111105e-01 -4.65223968e-01 5.94553411e-01
9.62543428e-01 -5.00535499e-03 -2.02800602e-01 -3.19683731e-01
7.05114663e-01 7.07396209e-01 5.24318397e-01 -9.44604814e-01
6.43153369e-01 3.36566359e-01 1.10268388e-02 -8.67922246e-01
-4.57443178e-01 3.97860911e-03 -8.15874219e-01 -2.38888264e-01
6.21182501e-01 -4.53990817e-01 -5.04672825e-01 -1.95734918e-01
-9.24105644e-01 -2.56957471e-01 -6.06476426e-01 1.55740947e-01
-8.59751165e-01 1.41061798e-01 -2.29494587e-01 -7.74938405e-01
-1.97170839e-01 -4.84270692e-01 2.69919008e-01 1.61293373e-01
-6.75840259e-01 -9.36526537e-01 -2.86408186e-01 4.12401885e-01
-7.73951784e-02 6.98449492e-01 1.12584305e+00 -7.61835456e-01
-8.59529972e-02 -1.23876616e-01 -2.97939274e-02 3.50198418e-01
1.06463738e-01 4.48279917e-01 -6.44184351e-01 3.05206984e-01
-9.41972360e-02 -2.22405121e-01 2.63804346e-01 3.28884423e-02
1.24223745e+00 -4.89268094e-01 -2.77575612e-01 1.83360279e-01
1.59649062e+00 1.00080475e-01 8.33996654e-01 5.78142762e-01
1.04081497e-01 9.18780267e-01 1.00587034e+00 3.88532996e-01
4.57206488e-01 3.19398403e-01 2.10779250e-01 5.74475646e-01
-7.99761061e-03 5.79803716e-03 -2.48892810e-02 5.96137941e-01
-3.76008540e-01 8.45799819e-02 -1.46277285e+00 8.78283203e-01
-2.00979829e+00 -1.35718536e+00 -3.07788551e-01 1.66633499e+00
1.26481116e+00 2.47865766e-01 2.35372767e-01 3.07754546e-01
7.47132659e-01 -3.36466074e-01 -1.33553922e-01 -6.04050636e-01
-3.58082324e-01 2.24091634e-01 -1.16600180e-02 7.16001570e-01
-7.28584051e-01 9.68198359e-01 4.39179087e+00 1.07339334e+00
-1.82413474e-01 4.79154773e-02 -7.63943493e-02 -2.90516317e-01
-4.36834037e-01 2.84803420e-01 -4.34305854e-02 6.10488534e-01
7.33059943e-01 -8.97259116e-01 7.34820426e-01 6.72295690e-01
1.15503684e-01 -3.50203037e-01 -1.15441811e+00 7.58233905e-01
-1.18409283e-01 -1.25185263e+00 1.47972539e-01 -3.48840505e-02
2.87384033e-01 -2.40565509e-01 -5.61321974e-01 2.50129163e-01
4.25636202e-01 -1.05998588e+00 8.40157390e-01 9.30778265e-01
6.09895527e-01 -1.00434601e+00 4.42049026e-01 3.40416461e-01
-8.20629597e-01 -3.47977392e-02 -2.60722965e-01 -1.47799209e-01
1.33197591e-01 9.96439993e-01 -4.49031621e-01 1.02688277e+00
7.54119277e-01 5.19380987e-01 -5.59588909e-01 7.93104529e-01
-5.94024777e-01 3.00534576e-01 -4.18473840e-01 -2.54312336e-01
-8.62442479e-02 -3.36506814e-01 7.32910037e-01 1.55365312e+00
1.20008722e-01 4.99759108e-01 -2.80642033e-01 1.43029070e+00
1.83438715e-02 -1.85388550e-01 -5.53191304e-01 -3.65183383e-01
8.74614954e-01 9.80930746e-01 -6.29694641e-01 -5.41737139e-01
1.79713994e-01 5.79188645e-01 1.33052304e-01 1.04761355e-01
-8.59021485e-01 -1.09062612e+00 7.90692985e-01 1.53447270e-01
-2.05025733e-01 -1.44247353e-01 -4.17260796e-01 -1.14347863e+00
-1.61950648e-01 -9.56262767e-01 6.23653352e-01 -1.35733271e+00
-1.48178959e+00 3.14251065e-01 6.77004755e-01 -9.81246829e-01
-4.35314447e-01 -7.26425886e-01 -4.13319916e-01 6.67889476e-01
-9.59483981e-01 -8.81949544e-01 -3.15270543e-01 5.32591641e-01
2.37480015e-01 2.27975309e-01 9.59903359e-01 -6.43730462e-02
-2.28553385e-01 -3.22085060e-03 -8.53217468e-02 1.70286428e-02
5.06017685e-01 -1.56954050e+00 2.52889115e-02 6.80045068e-01
5.83579838e-02 9.40318584e-01 1.11770535e+00 -8.45481277e-01
-1.09743619e+00 -8.99669886e-01 1.03218699e+00 -7.08962858e-01
1.04114163e+00 -1.10863581e-01 -1.08502519e+00 8.24707508e-01
2.90042669e-01 -3.80658120e-01 9.37610924e-01 1.71902671e-01
-4.58656222e-01 1.53786212e-01 -1.51540923e+00 7.33391523e-01
1.29737878e+00 -5.55244029e-01 -1.33680892e+00 6.77149475e-01
4.32049066e-01 -2.01700941e-01 -1.34086573e+00 1.04145028e-01
4.68371212e-01 -8.52240324e-01 6.92832530e-01 -8.30744445e-01
6.90199316e-01 -6.31725311e-01 -4.59149808e-01 -1.26128018e+00
-2.07473397e-01 -6.33433402e-01 -1.47054464e-01 1.28995717e+00
3.29938442e-01 -5.96855342e-01 7.36059994e-02 8.36506784e-01
3.68708521e-02 -3.78363758e-01 -9.07608628e-01 -9.66617823e-01
3.30721140e-01 -9.91354704e-01 6.23198032e-01 1.51595378e+00
7.82767415e-01 2.13586047e-01 2.07007527e-01 -9.98719260e-02
7.91084111e-01 3.19027245e-01 3.45411569e-01 -1.46351647e+00
1.14177778e-01 -7.18861878e-01 -3.68440926e-01 1.16132218e-02
-6.97314367e-03 -1.23807180e+00 -2.42273480e-01 -1.90803957e+00
-1.31114319e-01 -3.01021606e-01 -2.01223820e-01 7.81777561e-01
8.02450031e-02 -4.70013097e-02 3.74454349e-01 1.01595126e-01
-3.04453552e-01 5.46562560e-02 8.43874395e-01 1.31390169e-01
4.86343633e-03 -5.19166172e-01 -9.53449011e-01 8.68055046e-01
1.46150005e+00 -7.00126290e-01 -3.04332525e-01 3.89991179e-02
6.37147903e-01 -3.57644767e-01 8.22191894e-01 -7.96015322e-01
3.16601366e-01 -5.61040699e-01 5.86061068e-02 -3.40890527e-01
1.99706137e-01 -9.65191364e-01 4.49619591e-01 2.87546843e-01
-2.92729706e-01 -8.13736394e-03 3.90050888e-01 1.37185037e-01
-1.41049638e-01 -7.45145142e-01 7.09332526e-01 -4.25254524e-01
-7.82717049e-01 -5.00559568e-01 -2.52031475e-01 2.46169135e-01
8.63937259e-01 -2.08327189e-01 -4.44934398e-01 -1.68853238e-01
-9.94306564e-01 2.73577034e-01 6.32143199e-01 3.43017757e-01
3.76436919e-01 -1.25594175e+00 -6.08984888e-01 -5.25826514e-01
3.29597414e-01 -2.42439628e-01 1.22775838e-01 6.12849116e-01
-6.11206710e-01 4.35437769e-01 -4.19580430e-01 1.47738636e-01
-1.21663547e+00 8.44701588e-01 2.24501029e-01 2.53629804e-01
-4.25297469e-01 6.40908182e-01 -4.82945055e-01 -4.01355684e-01
-5.66928685e-02 -3.16535234e-01 -1.08152196e-01 2.50123799e-01
4.37992394e-01 8.11357200e-01 2.80763069e-03 -2.96733469e-01
-6.96287930e-01 2.59884000e-01 3.79507333e-01 -2.07409069e-01
1.55627215e+00 -6.87186196e-02 -7.08134770e-01 7.85551608e-01
6.90288782e-01 -1.45824447e-01 -2.85279423e-01 1.77494198e-01
3.81275922e-01 -2.54337013e-01 -5.78066260e-02 -1.58971155e+00
-2.24127859e-01 5.35094678e-01 1.99527010e-01 8.81653607e-01
1.03101349e+00 1.90658420e-01 6.02407642e-02 6.43759191e-01
5.73027134e-01 -1.53323567e+00 -6.93485022e-01 3.60852838e-01
1.49476516e+00 -9.98370767e-01 3.80993724e-01 -6.15387321e-01
-7.10900664e-01 1.28387117e+00 1.96481720e-01 -6.71921298e-03
8.20510030e-01 2.46871233e-01 -2.39346460e-01 -7.19781280e-01
-6.78502619e-01 -4.95190412e-01 1.33019313e-01 7.02960312e-01
4.24288541e-01 3.53878140e-01 -7.62773216e-01 8.50108027e-01
-8.10880005e-01 5.31393230e-01 8.55254769e-01 1.21994638e+00
-4.63534266e-01 -9.19209242e-01 -5.16987264e-01 5.01900017e-01
-3.89794707e-02 -4.31159288e-01 -8.88042212e-01 1.21057105e+00
4.46423620e-01 1.12591946e+00 -2.97758818e-01 -2.29478627e-01
5.13828635e-01 5.58871388e-01 6.09847724e-01 -5.97771406e-01
-5.37346244e-01 -3.99558365e-01 6.15703046e-01 -3.60397130e-01
-5.87838590e-01 -8.32564652e-01 -1.47593427e+00 -6.48005366e-01
-1.55438766e-01 5.12263477e-01 6.74693167e-01 6.64754212e-01
3.11875701e-01 5.96703410e-01 -2.51998395e-01 -2.21483633e-01
-1.51674882e-01 -1.09827709e+00 -7.38419354e-01 3.30719173e-01
-3.23302984e-01 -7.30531037e-01 -3.07901144e-01 4.37997103e-01] | [9.798919677734375, 8.190047264099121] |
4fa9bfb3-7edb-4764-abe6-2ef49d7dc8ef | inverted-pyramid-multi-task-transformer-for | 2203.07997 | null | https://arxiv.org/abs/2203.07997v3 | https://arxiv.org/pdf/2203.07997v3.pdf | InvPT: Inverted Pyramid Multi-task Transformer for Dense Scene Understanding | Multi-task dense scene understanding is a thriving research domain that requires simultaneous perception and reasoning on a series of correlated tasks with pixel-wise prediction. Most existing works encounter a severe limitation of modeling in the locality due to heavy utilization of convolution operations, while learning interactions and inference in a global spatial-position and multi-task context is critical for this problem. In this paper, we propose a novel end-to-end Inverted Pyramid multi-task Transformer (InvPT) to perform simultaneous modeling of spatial positions and multiple tasks in a unified framework. To the best of our knowledge, this is the first work that explores designing a transformer structure for multi-task dense prediction for scene understanding. Besides, it is widely demonstrated that a higher spatial resolution is remarkably beneficial for dense predictions, while it is very challenging for existing transformers to go deeper with higher resolutions due to huge complexity to large spatial size. InvPT presents an efficient UP-Transformer block to learn multi-task feature interaction at gradually increased resolutions, which also incorporates effective self-attention message passing and multi-scale feature aggregation to produce task-specific prediction at a high resolution. Our method achieves superior multi-task performance on NYUD-v2 and PASCAL-Context datasets respectively, and significantly outperforms previous state-of-the-arts. The code is available at https://github.com/prismformore/InvPT | ['Dan Xu', 'Hanrong Ye'] | 2022-03-15 | null | null | null | null | ['surface-normals-estimation', 'human-parsing', 'boundary-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.81709468e-01 -3.79243881e-01 2.54217833e-01 -5.29106975e-01
-9.61418211e-01 -1.37274697e-01 4.15371090e-01 9.54967830e-03
-2.94389218e-01 5.33933580e-01 1.45908371e-01 -4.75673079e-02
-1.32405937e-01 -7.43672967e-01 -1.05360854e+00 -5.69111347e-01
1.91763401e-01 3.41684759e-01 8.01345050e-01 -2.19782442e-01
3.60511363e-01 1.44701391e-01 -1.85027277e+00 9.39554155e-01
1.00769246e+00 1.09343469e+00 1.15782392e+00 8.58017027e-01
-3.45038064e-02 1.01561701e+00 -1.69695199e-01 -1.93808705e-01
8.33119750e-02 1.12398088e-01 -9.39642489e-01 -1.67736664e-01
8.04518402e-01 -3.10483843e-01 -1.15297109e-01 8.29884350e-01
5.90475976e-01 2.31778964e-01 2.13690311e-01 -1.12005329e+00
-6.17515206e-01 9.95460376e-02 -8.36322069e-01 4.05264497e-01
5.55055961e-02 1.76598787e-01 1.00984883e+00 -9.99467075e-01
9.02463123e-02 1.36313283e+00 5.21261811e-01 2.15591937e-02
-9.99929488e-01 -7.77916908e-01 4.80073094e-01 7.21330047e-01
-1.34973419e+00 -1.64055005e-01 7.21806526e-01 -3.36893111e-01
1.25602674e+00 2.85508126e-01 4.51109558e-01 9.17974591e-01
4.03762668e-01 7.88053513e-01 1.24366271e+00 8.50911159e-03
1.01599835e-01 -1.75065368e-01 6.01903163e-02 7.26408005e-01
-2.38140792e-01 -2.21302047e-01 -8.94091129e-01 1.39655992e-01
1.00644600e+00 3.55851680e-01 -7.11045936e-02 -6.61065504e-02
-1.30776012e+00 5.47984004e-01 6.60543442e-01 2.33648941e-01
-5.79874814e-01 4.24262106e-01 4.52314883e-01 1.29838109e-01
5.37587106e-01 -5.03081195e-02 -8.68333817e-01 -1.76933691e-01
-8.10794711e-01 3.20584953e-01 4.35937971e-01 8.22588265e-01
9.98152792e-01 -1.16125636e-01 -2.13425010e-01 9.90148127e-01
1.82418153e-01 5.15697598e-01 2.92920649e-01 -1.04735923e+00
6.21868968e-01 4.85445917e-01 2.57648025e-02 -9.39592242e-01
-6.18996024e-01 -6.67276978e-01 -9.50655997e-01 2.66614914e-01
2.17829704e-01 2.06466205e-02 -7.76549041e-01 1.57415271e+00
5.40009856e-01 7.69359469e-01 4.71472442e-02 1.07342958e+00
7.73187876e-01 9.39712346e-01 2.90301710e-01 1.95390806e-01
1.66118515e+00 -1.62207699e+00 -4.10258591e-01 -5.51113367e-01
6.19499624e-01 -9.08774674e-01 1.36459017e+00 5.10757625e-01
-9.41871107e-01 -1.04323125e+00 -7.90489674e-01 -7.12292612e-01
-3.88652831e-01 1.21416099e-01 8.67457032e-01 7.08312392e-02
-9.79945838e-01 2.69711316e-01 -7.91642070e-01 -2.86636770e-01
7.09193468e-01 3.01806659e-01 -2.12637424e-01 -4.52844530e-01
-9.48299527e-01 8.63310277e-01 2.36087754e-01 8.60727429e-02
-9.31392610e-01 -1.15863192e+00 -7.79333830e-01 6.08510785e-02
4.60403025e-01 -8.13744664e-01 1.29207647e+00 -6.89025402e-01
-1.23697925e+00 4.96338755e-01 -5.98213673e-01 -3.65641713e-01
3.00110966e-01 -6.92580462e-01 -4.07007709e-02 -1.36128575e-01
1.86752066e-01 7.37133861e-01 7.52796054e-01 -1.21719348e+00
-8.84388208e-01 -6.67671561e-01 3.25584620e-01 8.10210586e-01
-2.62925744e-01 -1.68813273e-01 -5.89367628e-01 -3.90461147e-01
1.16601050e-01 -5.93044221e-01 -3.87911230e-01 7.64926597e-02
-1.59629747e-01 -1.88162610e-01 1.10509634e+00 -5.78310370e-01
7.88288355e-01 -2.12887311e+00 1.37371704e-01 -5.08362353e-01
3.87769520e-01 1.44520149e-01 -5.60451932e-02 2.82200575e-01
1.63606972e-01 -2.23118305e-01 2.25476571e-04 -7.41130888e-01
-6.65734932e-02 2.72493780e-01 -3.55160385e-01 2.21865758e-01
-1.70383021e-01 9.65482116e-01 -6.80264413e-01 -5.22203743e-01
5.98640442e-01 7.45042741e-01 -7.59252846e-01 1.03796534e-01
-3.95169526e-01 5.07221878e-01 -5.55615127e-01 5.51366270e-01
1.06967807e+00 -5.92172801e-01 -2.05021262e-01 -6.39440775e-01
-3.51780683e-01 2.11407885e-01 -1.07090521e+00 2.18603134e+00
-9.57708657e-01 4.23005670e-01 1.50359839e-01 -1.00061810e+00
6.07866585e-01 1.02060519e-01 2.10941598e-01 -1.07225633e+00
-1.14689976e-01 -3.99710871e-02 -2.55635440e-01 -4.08094615e-01
7.05201924e-01 6.69300407e-02 -3.48329730e-02 6.26155585e-02
-1.65296227e-01 -1.20920472e-01 -1.87874079e-01 4.31593545e-02
1.03256917e+00 2.81353682e-01 1.39822260e-01 -2.28401884e-01
5.64164042e-01 6.03187941e-02 4.92947310e-01 7.03436911e-01
-6.78585619e-02 4.96185511e-01 9.49617699e-02 -6.17388487e-01
-9.58628297e-01 -1.07677960e+00 -5.71455061e-02 1.57273960e+00
3.87142539e-01 -3.86398375e-01 -3.37043196e-01 -2.27254584e-01
3.41233760e-02 5.91179192e-01 -4.56402242e-01 3.61502051e-01
-5.72398782e-01 -7.33447313e-01 2.29662836e-01 6.99693859e-01
1.04093671e+00 -1.01438177e+00 -7.66582310e-01 1.97821707e-01
-2.92780578e-01 -1.47359419e+00 -3.29954267e-01 2.05931604e-01
-7.91570961e-01 -8.88030469e-01 -4.34675068e-01 -8.81538451e-01
3.89524639e-01 6.04683816e-01 9.93091583e-01 -1.36013985e-01
-3.37234169e-01 6.31224439e-02 -2.22180188e-01 -4.36347932e-01
3.58037859e-01 -2.46080407e-03 -3.40532899e-01 -8.93453881e-02
1.65607095e-01 -8.97680163e-01 -8.22501481e-01 4.21762675e-01
-6.50651991e-01 9.01717961e-01 6.95009053e-01 7.20895171e-01
9.94276702e-01 1.30306184e-01 3.15233707e-01 -9.43409979e-01
1.40956789e-01 -5.01466811e-01 -5.32273173e-01 2.09764525e-01
-6.71268627e-02 -1.16746098e-01 8.38472247e-01 -1.96493998e-01
-1.39767957e+00 1.45208895e-01 -3.05075437e-01 -5.96619308e-01
-3.37199867e-01 2.24158555e-01 9.12348181e-03 -1.33701012e-01
4.14591849e-01 4.80739027e-01 -5.03274202e-01 -5.18085480e-01
4.13578957e-01 3.46107572e-01 4.70117688e-01 -7.04133034e-01
4.64428514e-01 7.04118490e-01 3.27072330e-02 -7.63450861e-01
-1.13580883e+00 -5.80288053e-01 -6.51131332e-01 -8.22431669e-02
1.05598950e+00 -1.40909576e+00 -8.92934620e-01 7.25196838e-01
-1.36667633e+00 -7.49197781e-01 1.55482337e-01 3.08976769e-01
-5.75950682e-01 2.18306214e-01 -5.54793477e-01 -5.40156245e-01
-4.84537184e-01 -1.25153863e+00 1.44388556e+00 1.73344105e-01
2.18367904e-01 -8.40671420e-01 -1.17293291e-01 8.68637025e-01
5.96665025e-01 -7.29801953e-02 6.49616420e-01 -6.18854836e-02
-1.06230521e+00 4.23195601e-01 -8.03951085e-01 5.68759218e-02
-1.29602611e-01 -5.71600556e-01 -1.23356616e+00 -1.46339551e-01
3.96659039e-02 -5.19065857e-01 9.68303978e-01 4.51474786e-01
1.77164268e+00 6.77644238e-02 -3.61318052e-01 9.43247199e-01
1.67409205e+00 4.10301685e-02 6.10317469e-01 1.60503671e-01
1.15040016e+00 3.27770352e-01 7.70382226e-01 5.18990099e-01
9.45852339e-01 7.64388680e-01 6.40112519e-01 -3.14281851e-01
-2.16033086e-01 -6.23259991e-02 -1.68472845e-02 6.67141855e-01
5.73516674e-02 -2.23519072e-01 -9.83465433e-01 5.76291263e-01
-2.14189672e+00 -9.00183558e-01 -3.18237633e-01 1.82285178e+00
5.04750609e-01 1.10973038e-01 -3.63620013e-01 -1.51526511e-01
4.18020815e-01 4.64160085e-01 -6.76274359e-01 -3.10929000e-01
4.68530692e-02 2.78559893e-01 4.70624179e-01 7.14548707e-01
-1.15290296e+00 1.14520276e+00 4.93889093e+00 1.15853345e+00
-1.08841634e+00 6.18415058e-01 7.20756412e-01 -4.23155457e-01
-1.11630805e-01 -9.26841125e-02 -9.70026791e-01 3.72275472e-01
5.94453573e-01 1.53650984e-01 3.18081468e-01 8.80800724e-01
2.70463586e-01 -4.79757994e-01 -7.79620826e-01 1.25538635e+00
2.48024631e-02 -1.35044312e+00 -1.31757483e-02 -1.30316451e-01
6.96621656e-01 4.39613700e-01 1.29443318e-01 2.94703126e-01
2.25800663e-01 -9.79963243e-01 6.00813925e-01 4.43694800e-01
5.80350220e-01 -6.20282769e-01 5.04567981e-01 5.27122617e-01
-1.69628525e+00 -2.12889165e-01 -5.17436922e-01 -4.85831112e-01
2.66796857e-01 7.38662481e-01 -3.58776301e-01 5.13200760e-01
1.03336596e+00 8.52974474e-01 -5.11713386e-01 8.49998355e-01
1.42612174e-01 3.84819597e-01 -3.98261696e-01 1.08994529e-01
3.10427338e-01 2.34985143e-01 1.50978580e-01 1.24320185e+00
3.45692366e-01 1.40382066e-01 5.27318120e-01 6.01471961e-01
1.19681850e-01 4.11666837e-03 -4.05045420e-01 6.06388509e-01
4.22531754e-01 1.33837950e+00 -4.20631021e-01 -4.09872532e-01
-7.42675900e-01 1.25499642e+00 6.74173534e-01 3.10101599e-01
-1.08043075e+00 1.36151500e-02 8.25369775e-01 1.53111145e-01
5.09689629e-01 -5.18201590e-01 -7.31152415e-01 -1.05334711e+00
8.62843767e-02 -4.98886228e-01 2.84523427e-01 -1.10787952e+00
-1.08836925e+00 6.63118958e-01 -1.15633003e-01 -8.79547119e-01
3.64199758e-01 -4.06711161e-01 -6.51834667e-01 1.04257774e+00
-1.94612861e+00 -1.63734543e+00 -8.53230596e-01 1.01237071e+00
1.05368781e+00 2.44433135e-01 7.19371676e-01 4.05429751e-01
-4.75552291e-01 2.97221005e-01 -1.17437400e-01 -4.90019768e-01
6.30545557e-01 -1.17598677e+00 4.17925656e-01 6.71889424e-01
-3.99018563e-02 8.47070813e-02 5.03253281e-01 -5.06441176e-01
-1.42479634e+00 -1.26872945e+00 6.80621088e-01 -3.22214067e-01
5.26781559e-01 -5.66497922e-01 -9.20970500e-01 7.91029751e-01
2.57324755e-01 2.99026966e-01 3.67300868e-01 2.07357839e-01
-3.01403761e-01 -4.02950943e-01 -7.85851896e-01 5.32773316e-01
1.26947773e+00 -6.03853583e-01 -2.11818188e-01 6.57696962e-01
9.62704301e-01 -7.61508882e-01 -7.92604506e-01 4.35902357e-01
3.69699478e-01 -1.28908277e+00 1.14947116e+00 1.25943169e-01
6.30247891e-01 -4.47240978e-01 -4.49166179e-01 -9.30655003e-01
-5.65513432e-01 -1.41948149e-01 3.60617973e-02 8.34164381e-01
1.77867234e-01 -6.22091115e-01 8.16839695e-01 2.66339034e-01
-4.95727122e-01 -1.15433550e+00 -9.25323606e-01 -2.58643836e-01
-1.40851960e-01 -7.63870955e-01 3.93962473e-01 6.44281745e-01
-4.48281854e-01 4.29698288e-01 -5.37932336e-01 7.28993475e-01
7.75723338e-01 4.02547568e-01 9.61748421e-01 -9.35852706e-01
-4.69786942e-01 -1.20218083e-01 -2.38416389e-01 -1.56288457e+00
-1.08514108e-01 -5.81119120e-01 1.05060195e-03 -1.76304150e+00
3.15648049e-01 -5.63708484e-01 -2.85528094e-01 5.62788546e-01
-2.01665819e-01 2.86333740e-01 3.06576759e-01 1.48072377e-01
-9.82146680e-01 6.65158570e-01 1.54006326e+00 2.12594196e-01
-9.43307113e-03 -2.75179595e-01 -6.33823633e-01 7.67726183e-01
8.19020927e-01 -2.66359299e-01 -7.76490390e-01 -9.44835424e-01
7.53684193e-02 1.71743065e-01 7.39826739e-01 -1.36329854e+00
4.52827901e-01 -2.61968464e-01 5.19677818e-01 -9.77185905e-01
9.70624864e-01 -6.57808006e-01 5.20914048e-02 1.80344969e-01
-4.69485372e-02 1.64792210e-01 6.20989799e-01 5.13618112e-01
-2.46340230e-01 3.56503010e-01 7.23657370e-01 -2.95742601e-01
-1.41893816e+00 4.92507786e-01 1.84291322e-03 -6.46540057e-03
1.09160101e+00 -6.55603111e-02 -5.91267169e-01 -1.27568871e-01
-5.42307794e-01 5.21962404e-01 2.13574976e-01 3.71732354e-01
7.27914035e-01 -1.04677880e+00 -6.79020345e-01 1.76081613e-01
-1.23793207e-01 4.93869424e-01 1.10616779e+00 8.96836638e-01
-4.80440706e-01 3.42017025e-01 -2.54444987e-01 -8.38027477e-01
-1.42655885e+00 3.55991751e-01 1.47677749e-01 -4.77926582e-01
-7.40468800e-01 1.12799442e+00 9.03987706e-01 -4.17964607e-01
5.51166534e-02 -2.77664006e-01 2.11437289e-02 -3.02865267e-01
6.07350945e-01 1.26827285e-01 -9.08258557e-02 -5.60788751e-01
-3.14339101e-01 8.82149100e-01 -2.51357466e-01 2.15712011e-01
1.34774935e+00 -3.39679331e-01 -5.47896549e-02 3.58189195e-01
1.16782153e+00 -3.35398197e-01 -1.68033600e+00 -4.77018535e-01
-5.70329130e-01 -6.20157242e-01 2.95178413e-01 -8.40581894e-01
-1.00350273e+00 1.12706709e+00 5.23866534e-01 -2.04031259e-01
1.36224878e+00 4.58070859e-02 1.12175548e+00 3.34390044e-01
4.42116082e-01 -8.29927921e-01 3.56211245e-01 7.65975654e-01
1.09254467e+00 -1.36341476e+00 -3.18938829e-02 -7.25990236e-01
-7.44604588e-01 6.88107014e-01 1.23899281e+00 -1.48807570e-01
8.31751227e-01 4.77732688e-01 -2.08107099e-01 -2.66533047e-01
-1.01017678e+00 -2.76596248e-01 1.84729472e-01 4.57136929e-01
3.06524038e-01 4.21017669e-02 1.53488398e-01 4.98029619e-01
-1.94471449e-01 1.45561278e-01 1.56106036e-02 7.70462394e-01
-5.72268844e-01 -7.78564215e-01 -2.93677986e-01 5.46290815e-01
-2.16656864e-01 -4.04597163e-01 4.88210618e-01 5.06723583e-01
5.34335554e-01 7.60381043e-01 2.18392983e-01 -3.92844468e-01
2.22179875e-01 -1.31207079e-01 4.12261099e-01 -5.59584260e-01
-5.28740227e-01 -1.07099429e-01 -7.63614178e-02 -9.65233028e-01
-2.73856878e-01 -5.09496570e-01 -1.36100769e+00 -4.14586216e-01
1.32608801e-01 -3.30283403e-01 6.11295998e-01 9.85218704e-01
5.97198188e-01 8.77505779e-01 1.58139080e-01 -1.11320424e+00
-6.25615641e-02 -9.40125585e-01 -4.35781419e-01 2.89034784e-01
3.02355856e-01 -7.65661418e-01 6.95577636e-02 -1.28646955e-01] | [9.692279815673828, 1.2163903713226318] |
077cfce5-6e49-4bbe-a7f9-9a1ff351d18a | open-vocabulary-point-cloud-object-detection | 2304.00788 | null | https://arxiv.org/abs/2304.00788v2 | https://arxiv.org/pdf/2304.00788v2.pdf | Open-Vocabulary Point-Cloud Object Detection without 3D Annotation | The goal of open-vocabulary detection is to identify novel objects based on arbitrary textual descriptions. In this paper, we address open-vocabulary 3D point-cloud detection by a dividing-and-conquering strategy, which involves: 1) developing a point-cloud detector that can learn a general representation for localizing various objects, and 2) connecting textual and point-cloud representations to enable the detector to classify novel object categories based on text prompting. Specifically, we resort to rich image pre-trained models, by which the point-cloud detector learns localizing objects under the supervision of predicted 2D bounding boxes from 2D pre-trained detectors. Moreover, we propose a novel de-biased triplet cross-modal contrastive learning to connect the modalities of image, point-cloud and text, thereby enabling the point-cloud detector to benefit from vision-language pre-trained models,i.e.,CLIP. The novel use of image and vision-language pre-trained models for point-cloud detectors allows for open-vocabulary 3D object detection without the need for 3D annotations. Experiments demonstrate that the proposed method improves at least 3.03 points and 7.47 points over a wide range of baselines on the ScanNet and SUN RGB-D datasets, respectively. Furthermore, we provide a comprehensive analysis to explain why our approach works. | ['Shanghang Zhang', 'Kurt Keutzer', 'Masayoshi Tomizuka', 'Xiaodong Xie', 'Xiaobao Wei', 'Chenfeng Xu', 'Yuheng Lu'] | 2023-04-03 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lu_Open-Vocabulary_Point-Cloud_Object_Detection_Without_3D_Annotation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lu_Open-Vocabulary_Point-Cloud_Object_Detection_Without_3D_Annotation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['cloud-detection'] | ['computer-vision'] | [ 4.16399390e-02 -4.34771329e-02 -2.81141460e-04 -2.86682963e-01
-1.22798789e+00 -7.70363271e-01 8.01149011e-01 2.32637197e-01
-3.19578350e-01 -2.37707466e-01 -2.76807904e-01 -1.88381732e-01
4.54228818e-01 -5.29544353e-01 -1.08491123e+00 -6.00140035e-01
1.58090919e-01 5.28088927e-01 5.73020399e-01 -1.04504071e-01
3.16833198e-01 8.26914430e-01 -1.78218615e+00 1.94257736e-01
5.52117527e-01 1.27367783e+00 5.14350235e-01 3.78919333e-01
-4.84991908e-01 2.18527243e-01 -2.59851843e-01 -1.42675772e-01
5.96552372e-01 3.99404228e-01 -3.00748467e-01 4.14331198e-01
9.19163704e-01 -3.73615503e-01 -2.02189177e-01 1.03983653e+00
2.80769289e-01 -2.58465298e-02 9.79369998e-01 -1.14897895e+00
-7.77521670e-01 -6.81435019e-02 -8.75782311e-01 7.62443244e-02
3.63547266e-01 4.47955698e-01 9.54933882e-01 -1.41573548e+00
4.68570769e-01 1.35191071e+00 5.73262274e-01 5.00653625e-01
-1.11991644e+00 -7.94903815e-01 4.66151118e-01 -3.48469312e-03
-1.84405327e+00 -2.02623516e-01 8.77813458e-01 -7.69795060e-01
1.10519028e+00 2.99521573e-02 5.08549690e-01 9.28308964e-01
-1.33295015e-01 8.77196729e-01 6.44372880e-01 -3.60433847e-01
2.71024704e-02 4.12836850e-01 1.19858412e-02 7.16578126e-01
3.10869515e-01 1.09563425e-01 -2.57244647e-01 -1.23059303e-01
7.73521841e-01 3.22038621e-01 1.12221964e-01 -9.82404351e-01
-1.20830822e+00 6.64958239e-01 7.73585498e-01 5.04384264e-02
-3.61658394e-01 8.71540383e-02 4.04265016e-01 -6.61452040e-02
8.06394458e-01 4.16653864e-02 -3.30147237e-01 5.55176437e-01
-7.21843600e-01 1.27588600e-01 1.53293416e-01 1.52288687e+00
9.59815979e-01 -2.19324201e-01 2.33091190e-02 5.53526640e-01
5.91483593e-01 1.14318061e+00 -1.91976801e-02 -4.09294695e-01
6.83796287e-01 9.40114737e-01 4.38921265e-02 -9.54586089e-01
-2.16288596e-01 -2.70664006e-01 -5.95268309e-01 2.61558056e-01
1.20915323e-02 4.77975577e-01 -1.16891801e+00 1.15999603e+00
6.58501863e-01 1.80135220e-01 -3.28197256e-02 1.04556978e+00
9.28141654e-01 6.44137084e-01 -1.93988346e-02 2.20882267e-01
1.53256726e+00 -6.31553948e-01 -7.00789019e-02 -3.05616230e-01
6.12960994e-01 -6.85984135e-01 1.20828617e+00 -9.29800514e-03
-8.11476469e-01 -7.09922433e-01 -9.38260734e-01 -3.56206954e-01
-6.46450579e-01 1.73799768e-01 1.42908990e-01 2.48962209e-01
-7.63202488e-01 -1.38034284e-01 -7.70231187e-01 -6.05718136e-01
5.98737657e-01 2.00374573e-01 -3.54084760e-01 -3.09311226e-02
-4.54084963e-01 9.38535154e-01 6.96055233e-01 -3.20087910e-01
-1.18523967e+00 -6.28651321e-01 -9.78509963e-01 -4.21178788e-02
5.08212328e-01 -9.52073753e-01 1.13386846e+00 -6.50569797e-01
-9.37611103e-01 1.40572798e+00 -5.17216995e-02 -2.48869404e-01
4.68259275e-01 -1.57744646e-01 -1.94101319e-01 4.71849144e-01
4.23352331e-01 9.62908149e-01 1.26308000e+00 -1.55317557e+00
-1.15140653e+00 -6.95824444e-01 8.41119885e-02 3.17256123e-01
-1.57865882e-01 -1.35456994e-01 -8.37709069e-01 -4.02435780e-01
5.63555598e-01 -1.01579058e+00 -2.81525552e-02 3.83327931e-01
-4.50977683e-01 -6.86330616e-01 1.01159906e+00 -1.58643231e-01
4.89267468e-01 -2.32858324e+00 -1.31306365e-01 8.51663426e-02
3.83560777e-01 -3.78781632e-02 -1.73431039e-01 1.11915909e-01
-3.28616761e-02 2.22383509e-03 1.07480384e-01 -4.14444625e-01
5.47551252e-02 6.04338907e-02 -5.99290013e-01 6.67242110e-01
6.62067652e-01 8.99977148e-01 -8.10592234e-01 -6.25319660e-01
7.51358092e-01 3.07973713e-01 -4.08420146e-01 5.23027666e-02
-5.91796339e-01 4.10240263e-01 -5.59993923e-01 9.75523114e-01
8.63438725e-01 -4.09397393e-01 -5.67409098e-01 -3.05279553e-01
-3.88680995e-01 1.21191867e-01 -1.08792031e+00 1.88579762e+00
-4.14999068e-01 3.60905439e-01 -1.52018502e-01 -7.70606041e-01
1.06861401e+00 -7.69569213e-03 2.71778792e-01 -4.17787105e-01
1.40411437e-01 2.82892436e-01 -6.84341908e-01 -4.69273567e-01
2.77705759e-01 -4.18118164e-02 -1.82782874e-01 1.96788996e-01
1.44944489e-01 -5.21712661e-01 -2.13078648e-01 3.34629744e-01
8.70319247e-01 -1.30146118e-02 1.64037794e-01 1.72347222e-02
7.13874042e-01 3.07700634e-01 2.59611029e-02 8.94252062e-01
-6.90051764e-02 6.74758017e-01 -1.10510752e-01 -4.63510692e-01
-1.16472352e+00 -1.30957794e+00 -2.37754911e-01 1.06813085e+00
5.87675333e-01 -2.41893470e-01 -7.45075718e-02 -9.41033840e-01
3.43846351e-01 6.32949352e-01 -5.34574330e-01 -5.72832301e-04
-3.38153094e-01 1.34411409e-01 3.36302429e-01 4.63194788e-01
2.58686870e-01 -4.73011822e-01 -6.28063142e-01 -3.33004564e-01
-9.09839123e-02 -1.29146135e+00 -3.58105808e-01 3.03035736e-01
-8.68598878e-01 -9.08644617e-01 -7.35893428e-01 -1.06254005e+00
7.73176134e-01 1.02135408e+00 8.07026148e-01 -1.14435948e-01
-1.56129152e-01 7.81027734e-01 -4.38579977e-01 -6.53564453e-01
-2.07877204e-01 9.59622636e-02 3.08517277e-01 6.06607459e-03
7.46344984e-01 -2.37991616e-01 -4.87195879e-01 3.52390975e-01
-6.35867357e-01 1.44712850e-01 8.02414119e-01 4.92646366e-01
1.05290616e+00 -3.40603262e-01 -1.38992723e-02 -2.06261396e-01
-1.94868639e-01 -3.55386853e-01 -8.19610059e-01 1.75145462e-01
-3.00775200e-01 -3.09370831e-03 1.79161176e-01 -6.46767914e-01
-7.29072154e-01 5.84205985e-01 2.10783526e-01 -1.24248397e+00
-3.87296021e-01 3.13015166e-03 -1.04385138e-01 -3.51682663e-01
7.57240355e-01 5.44851959e-01 -1.68313980e-01 -5.16464055e-01
8.63600791e-01 7.97615826e-01 6.47227764e-01 -4.71320480e-01
1.52644491e+00 9.44075465e-01 -1.68782294e-01 -9.81472194e-01
-8.77952933e-01 -1.15085256e+00 -8.68669629e-01 -2.24618539e-01
1.06526148e+00 -1.48851728e+00 -5.50508261e-01 5.86001649e-02
-1.40794849e+00 3.19634736e-01 -2.80151188e-01 3.05902421e-01
-5.01931250e-01 3.76104355e-01 -9.40991193e-02 -7.81325221e-01
-3.34791481e-01 -9.41256464e-01 1.82166934e+00 3.52407992e-02
3.93001109e-01 -4.65346158e-01 -2.23536208e-01 3.46234202e-01
-1.47560909e-01 1.83527246e-01 7.03588068e-01 -7.47178555e-01
-1.02957571e+00 -5.31630099e-01 -6.30770743e-01 3.04831594e-01
-7.09876344e-02 -2.11678848e-01 -1.14048219e+00 -3.46959502e-01
4.20689583e-02 -3.13640296e-01 7.95362771e-01 4.99557666e-02
1.06314135e+00 1.44889176e-01 -6.87218130e-01 7.19962180e-01
1.27133524e+00 -1.29781276e-01 1.52730495e-01 3.26021463e-01
9.36549187e-01 4.46133137e-01 7.37440586e-01 4.15958405e-01
4.99455899e-01 6.09697938e-01 7.37477779e-01 -1.22761510e-01
2.50718910e-02 -5.47498286e-01 1.92154348e-01 3.69300783e-01
1.26321018e-01 1.48214191e-01 -1.22183311e+00 6.05001688e-01
-1.68521190e+00 -5.92072427e-01 -2.52568901e-01 1.85755122e+00
4.28121835e-01 3.33994508e-01 1.09941393e-01 -3.06404471e-01
9.03789520e-01 2.23568715e-02 -8.62384260e-01 3.40832263e-01
-1.02063641e-02 -8.85965973e-02 7.60205507e-01 2.84826458e-02
-1.32401502e+00 1.15115118e+00 4.96363974e+00 7.95303404e-01
-1.14253592e+00 2.13266581e-01 -5.88968396e-02 -2.11374611e-01
-1.39659764e-02 5.23018241e-02 -1.08190012e+00 1.61958650e-01
2.60566771e-01 -1.65584177e-01 -3.01102456e-02 1.19108617e+00
2.90986244e-02 2.29188159e-01 -1.38777339e+00 1.51350629e+00
3.20877075e-01 -1.30720103e+00 5.05399704e-01 -5.59261665e-02
5.41229248e-01 5.02928734e-01 8.73517543e-02 4.06261683e-01
-4.47900183e-02 -3.79044056e-01 1.16382480e+00 3.31569463e-01
8.72018039e-01 -4.26038206e-01 3.82096142e-01 6.58224881e-01
-1.40122771e+00 -7.65178353e-02 -5.59902370e-01 1.28546998e-01
-9.86797363e-02 3.08547765e-01 -1.10818839e+00 5.78222454e-01
9.88716841e-01 8.75428677e-01 -7.61189103e-01 9.53499556e-01
-3.38800669e-01 1.01385973e-01 -6.12517297e-01 -4.70501184e-02
3.90335470e-01 1.17191382e-01 8.50725412e-01 9.49098110e-01
2.52192259e-01 1.53540701e-01 4.17071968e-01 1.17901838e+00
9.63491127e-02 5.66299725e-03 -8.10737729e-01 2.65548497e-01
5.17735064e-01 1.07771015e+00 -6.10163450e-01 -3.63056481e-01
-7.25706756e-01 7.08553910e-01 2.76393890e-01 1.62425786e-01
-7.71164119e-01 -2.90299475e-01 5.78959942e-01 2.79163748e-01
8.20842862e-01 -4.20069993e-01 -2.69234717e-01 -1.29227412e+00
1.33229405e-01 -3.52872550e-01 3.26274067e-01 -1.10454416e+00
-1.57455349e+00 3.21864516e-01 2.19843417e-01 -1.62214339e+00
7.99024701e-02 -7.86673367e-01 -3.34522635e-01 8.10667098e-01
-1.52899432e+00 -1.68916035e+00 -4.90660906e-01 8.36731315e-01
7.16427803e-01 -5.55004887e-02 3.08267981e-01 2.14174375e-01
-1.53224871e-01 3.40753704e-01 4.40157438e-03 1.34870768e-01
6.25682831e-01 -1.05221117e+00 6.49351358e-01 7.30650961e-01
4.84429151e-01 4.36273366e-01 4.94249374e-01 -6.40507817e-01
-1.57243538e+00 -1.48453355e+00 6.07451618e-01 -9.67770576e-01
6.32036328e-01 -8.79727542e-01 -8.61032963e-01 6.93260968e-01
-4.55930680e-01 2.54280448e-01 1.96701512e-01 -1.46319941e-02
-7.69473195e-01 -1.08134449e-01 -9.18238163e-01 4.75193202e-01
1.09381413e+00 -8.57697666e-01 -9.43499207e-01 7.41581380e-01
9.95790601e-01 -6.39668643e-01 -4.56549644e-01 4.09081191e-01
1.25776991e-01 -4.98678207e-01 1.41909337e+00 -3.37637305e-01
1.07880898e-01 -6.62376940e-01 -4.12913263e-01 -7.84011364e-01
-3.19531471e-01 4.13066074e-02 -6.34441450e-02 9.65318501e-01
1.12962253e-01 -4.63347077e-01 5.61317921e-01 2.15527192e-01
-4.12571877e-01 -3.25203985e-01 -1.04694629e+00 -1.03677130e+00
6.94980994e-02 -8.93676758e-01 5.04949510e-01 8.11837316e-01
-4.33562219e-01 5.46080351e-01 2.15840503e-03 7.61406720e-01
7.89586067e-01 3.79305422e-01 1.25862360e+00 -1.25363684e+00
1.89564526e-01 -3.53158832e-01 -7.58851290e-01 -1.55198669e+00
1.12105638e-01 -1.25641310e+00 8.53945389e-02 -1.21871912e+00
1.97351962e-01 -4.77516413e-01 -1.71024382e-01 4.10442352e-01
3.68333310e-02 3.24911356e-01 4.60259855e-01 5.96026063e-01
-9.65212405e-01 6.55073166e-01 9.71561730e-01 -6.18928909e-01
-1.65355071e-01 -2.77912840e-02 -4.47380841e-01 7.85005450e-01
3.60250086e-01 -6.49200261e-01 -1.68837562e-01 -6.38076603e-01
3.32310423e-02 -3.08720827e-01 1.02698636e+00 -9.63564694e-01
3.86662722e-01 8.10360163e-02 4.45144445e-01 -1.46097338e+00
3.91133964e-01 -1.13097215e+00 -2.77249545e-01 3.07144701e-01
-1.22033887e-01 -3.97476435e-01 3.78492534e-01 1.06795585e+00
1.66512847e-01 5.92427514e-03 6.79983199e-01 -1.48749143e-01
-1.03711057e+00 4.83889371e-01 -8.48635286e-03 -1.56875134e-01
1.28023255e+00 -3.60554993e-01 -1.59001514e-01 1.18791007e-01
-7.22810626e-01 5.08990169e-01 6.99247360e-01 8.62046301e-01
8.18160892e-01 -1.30544055e+00 -6.06371224e-01 3.73430192e-01
1.09407365e+00 5.81270635e-01 2.72862077e-01 6.14718437e-01
-3.84152383e-01 5.52255988e-01 2.63903350e-01 -1.55530453e+00
-1.17817211e+00 9.24290597e-01 2.58359194e-01 4.11737829e-01
-9.59087014e-01 8.73033106e-01 6.67449176e-01 -5.24871111e-01
4.21750218e-01 -6.84561312e-01 2.13582426e-01 -1.45513222e-01
4.16775525e-01 -6.09996319e-02 4.29962352e-02 -7.92440772e-01
-7.60670960e-01 8.84902596e-01 -1.47274524e-01 2.45983467e-01
1.08990049e+00 -4.25187171e-01 1.88806072e-01 7.41542220e-01
1.12261355e+00 -5.43223679e-01 -1.19169104e+00 -8.16783011e-01
-1.17500275e-01 -6.43727541e-01 1.64183676e-01 -5.88556111e-01
-5.51408231e-01 8.84739101e-01 9.76084888e-01 -4.58773226e-02
7.24893689e-01 7.29740679e-01 4.63703990e-01 6.47006154e-01
6.91273332e-01 -6.78837776e-01 2.65318543e-01 6.52416408e-01
8.94395530e-01 -1.70370853e+00 -2.56701168e-02 -3.54817510e-01
-4.26243067e-01 1.07082140e+00 6.24248147e-01 -2.28355035e-01
5.40241241e-01 -3.09411794e-01 -1.72857083e-02 -5.56118846e-01
-5.47374368e-01 -5.89759827e-01 5.60279429e-01 7.98490286e-01
-4.28877443e-01 3.15738246e-02 5.00419676e-01 3.82067919e-01
4.63442914e-02 -3.29488695e-01 1.42526282e-02 8.05722237e-01
-6.40642762e-01 -4.87008661e-01 -7.42092431e-01 3.26169461e-01
3.01402867e-01 -2.80055273e-02 -4.60522264e-01 9.47530866e-01
3.80431265e-01 7.10708022e-01 3.72314662e-01 -4.25318152e-01
5.81503034e-01 5.39056258e-03 4.75440770e-01 -1.01512921e+00
-2.06747383e-01 4.50509936e-02 -4.06826824e-01 -3.82955343e-01
-3.58035654e-01 -7.29781270e-01 -1.15198374e+00 6.54669777e-02
-7.73991704e-01 -2.78801948e-01 9.12309647e-01 8.97732139e-01
4.95270997e-01 1.59451202e-01 6.59473956e-01 -1.31471860e+00
-6.30212069e-01 -7.25919306e-01 -4.31049794e-01 4.81463522e-01
5.79739630e-01 -9.68348920e-01 -5.01149297e-01 5.73944449e-02] | [7.869972229003906, -2.9402194023132324] |
d01ca06d-f0e4-4e64-8b3f-f7bd8c8b4250 | deeper-text-understanding-for-ir-with | 1905.09217 | null | https://arxiv.org/abs/1905.09217v1 | https://arxiv.org/pdf/1905.09217v1.pdf | Deeper Text Understanding for IR with Contextual Neural Language Modeling | Neural networks provide new possibilities to automatically learn complex language patterns and query-document relations. Neural IR models have achieved promising results in learning query-document relevance patterns, but few explorations have been done on understanding the text content of a query or a document. This paper studies leveraging a recently-proposed contextual neural language model, BERT, to provide deeper text understanding for IR. Experimental results demonstrate that the contextual text representations from BERT are more effective than traditional word embeddings. Compared to bag-of-words retrieval models, the contextual language model can better leverage language structures, bringing large improvements on queries written in natural languages. Combining the text understanding ability with search knowledge leads to an enhanced pre-trained BERT model that can benefit related search tasks where training data are limited. | ['Zhuyun Dai', 'Jamie Callan'] | 2019-05-22 | null | null | null | null | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [ 1.56286463e-01 1.19765371e-01 -1.02529001e+00 -5.29360950e-01
-8.20178628e-01 -3.66004884e-01 1.15931892e+00 4.83205438e-01
-7.13034511e-01 2.63611674e-01 9.67101455e-01 -6.35221720e-01
-3.97309512e-01 -8.02831769e-01 -3.71640056e-01 -5.66182807e-02
6.91363961e-02 6.14871740e-01 9.25913006e-02 -5.00375807e-01
5.20038724e-01 3.86466384e-01 -1.16010213e+00 6.14851415e-01
4.22175527e-01 9.04367447e-01 2.11352065e-01 5.04447460e-01
-6.10739827e-01 9.15581644e-01 -4.68866140e-01 -2.68157497e-02
-1.28810436e-01 -5.80059290e-02 -1.12672246e+00 -7.32581496e-01
6.19837463e-01 -4.91314769e-01 -6.72582626e-01 5.59544504e-01
3.33481610e-01 4.93930757e-01 6.98465049e-01 -5.60689867e-01
-1.63345063e+00 7.73371935e-01 -3.27651322e-01 5.16309857e-01
2.25243479e-01 -4.36693370e-01 1.86095142e+00 -1.03034616e+00
5.00333309e-01 1.58447886e+00 3.40421021e-01 6.07670844e-01
-1.23333704e+00 -3.27794611e-01 3.10525715e-01 3.31541210e-01
-1.06735539e+00 -1.77590832e-01 6.34924889e-01 -3.38586904e-02
1.65714705e+00 1.99751228e-01 4.32837844e-01 1.01967514e+00
2.57656991e-01 1.21786344e+00 4.14010674e-01 -1.03951490e+00
-1.56215772e-01 2.70494491e-01 6.76013947e-01 3.80530030e-01
3.20026934e-01 2.03537956e-01 -5.51250815e-01 -2.94507325e-01
4.66433823e-01 3.37513000e-01 -8.76400918e-02 -1.87684044e-01
-6.89994872e-01 1.37595022e+00 8.03669572e-01 8.29429150e-01
-2.84375817e-01 5.86695015e-01 4.32014257e-01 2.15501815e-01
7.30574787e-01 1.13584006e+00 -5.63106537e-01 1.06877461e-01
-7.35479772e-01 2.40562856e-01 7.93152511e-01 8.79973292e-01
8.84487987e-01 -1.53845996e-01 -6.25469267e-01 1.03831685e+00
5.58454633e-01 4.10729945e-01 7.45357394e-01 -4.11590040e-01
3.02691907e-01 6.71213806e-01 -1.54407874e-01 -1.35088754e+00
-2.64362723e-01 -3.72893333e-01 -2.32600421e-01 -4.79553759e-01
-3.09172004e-01 2.33576491e-01 -9.15124178e-01 1.48994505e+00
-3.72834414e-01 -4.29690033e-01 3.44301283e-01 6.14151955e-01
8.48452270e-01 8.84930670e-01 3.68012488e-01 2.85636604e-01
1.41378403e+00 -1.12073779e+00 -8.43355238e-01 -8.13004434e-01
9.84246612e-01 -7.46189237e-01 1.35828555e+00 -1.45823555e-02
-7.96098888e-01 -5.55236340e-01 -8.49331439e-01 -6.64262354e-01
-1.02762783e+00 9.63834599e-02 1.05531931e+00 4.29152399e-01
-1.17090559e+00 6.50776848e-02 -6.25775397e-01 -7.01934814e-01
4.41011369e-01 3.05131048e-01 6.34636655e-02 -4.87299919e-01
-1.42990887e+00 1.01054752e+00 5.33034563e-01 -6.18618131e-02
-6.61535084e-01 -6.35785401e-01 -8.79095972e-01 3.48893017e-01
5.51836431e-01 -6.14203870e-01 1.39112580e+00 -5.72209716e-01
-9.46979463e-01 5.25461257e-01 -4.83135611e-01 -4.73831236e-01
-6.47185206e-01 -6.29067123e-01 -5.07930160e-01 1.55561149e-01
-6.12393655e-02 9.60205555e-01 4.03002262e-01 -1.08565903e+00
-2.75843054e-01 -4.07498002e-01 2.18873903e-01 3.65959525e-01
-9.30935919e-01 2.69239068e-01 -6.74812436e-01 -7.76156843e-01
-3.61210406e-01 -5.96734881e-01 -1.18153408e-01 -1.01210512e-01
-1.03326999e-01 -1.02794790e+00 9.03350234e-01 -3.41806352e-01
1.55991685e+00 -1.84105420e+00 -2.55742073e-01 1.39142096e-01
-6.42441660e-02 3.93874764e-01 -5.62260091e-01 1.04362965e+00
7.67835453e-02 5.72460592e-01 5.49031138e-01 -7.61285424e-02
1.25875205e-01 4.48911697e-01 -9.34104800e-01 -3.81839782e-01
2.78169125e-01 1.65920687e+00 -8.33984792e-01 -3.44455957e-01
-3.81308347e-02 4.59347814e-01 -7.74517536e-01 1.03352226e-01
-7.81756520e-01 -4.89454150e-01 -9.86969769e-01 5.21391511e-01
-8.28121081e-02 -5.07625937e-01 8.42880234e-02 4.30374481e-02
3.49421322e-01 7.36244678e-01 -1.99011967e-01 1.81883693e+00
-8.18334281e-01 1.02363789e+00 -4.57016379e-01 -1.00646472e+00
9.21344101e-01 2.90935546e-01 8.39785114e-02 -1.17712247e+00
-1.06693886e-01 -1.38571769e-01 -3.34094197e-01 -6.79574430e-01
1.03401566e+00 -9.15793702e-02 -1.52417883e-01 7.61244118e-01
2.15065815e-02 -1.69175684e-01 2.15726376e-01 6.02787137e-01
9.91717577e-01 -1.49805903e-01 9.54596102e-02 -3.36144358e-01
2.16568187e-01 2.57083982e-01 -1.80195838e-01 1.07470798e+00
2.93549687e-01 5.42615987e-02 1.01992168e-01 -4.58105028e-01
-6.15713418e-01 -6.54009938e-01 4.40682583e-02 1.83338892e+00
-2.61093602e-02 -6.99627817e-01 -8.80813375e-02 -6.63060546e-01
1.98181897e-01 8.96755099e-01 -7.89199412e-01 -5.99811316e-01
-4.57563460e-01 -3.51340860e-01 5.74251294e-01 9.58463192e-01
-4.83235084e-02 -1.22401285e+00 -2.04068676e-01 8.35619271e-02
9.86499861e-02 -8.62520278e-01 -6.75949395e-01 3.79207373e-01
-1.05102992e+00 -7.34221876e-01 -4.43709493e-01 -1.10415566e+00
3.80428731e-01 5.16229570e-01 1.48616290e+00 3.07885051e-01
-6.05106711e-01 9.01853204e-01 -5.69888353e-01 -5.99641085e-01
-8.74904171e-02 2.66113549e-01 -1.02347068e-01 -6.89130127e-01
1.34257293e+00 -1.30403414e-01 -7.72998393e-01 1.17431812e-01
-1.37786186e+00 -4.95145917e-01 7.65653133e-01 1.13218904e+00
2.77731299e-01 -2.48907864e-01 7.60627449e-01 -6.35862589e-01
1.46310318e+00 -4.89774495e-01 -2.82181472e-01 6.88225269e-01
-1.29624736e+00 5.68421602e-01 2.04927146e-01 -5.84551036e-01
-1.21960640e+00 -7.08581984e-01 6.69728220e-02 -3.40706483e-02
-6.00237846e-02 1.20755124e+00 4.27608043e-01 3.32178533e-01
1.03414035e+00 -2.89794691e-02 -2.31688365e-01 -5.64900398e-01
9.77900147e-01 6.71830356e-01 1.00667097e-01 -8.41779053e-01
3.79469693e-01 2.18620449e-01 -4.28985685e-01 -8.24163854e-01
-1.09891188e+00 -1.11910713e+00 -4.05484587e-01 4.53098446e-01
9.92595911e-01 -9.00076509e-01 -1.89897269e-01 -5.14653385e-01
-1.18814409e+00 -6.95174485e-02 -4.47515368e-01 5.21844208e-01
-1.19766966e-01 3.32505673e-01 -6.87660992e-01 -6.80489600e-01
-4.18601602e-01 -6.06880248e-01 1.20377672e+00 2.15948671e-01
-4.32059348e-01 -1.55986762e+00 4.04460549e-01 3.23528022e-01
7.98035443e-01 -6.78322256e-01 1.44585168e+00 -1.11148190e+00
-5.93928039e-01 -5.13357759e-01 -4.45410550e-01 1.51718348e-01
1.92091435e-01 -4.58818167e-01 -1.01887190e+00 -1.91079900e-01
-2.08454385e-01 -8.41490924e-01 1.19660854e+00 2.63677120e-01
1.13767445e+00 -5.28892636e-01 -5.68811476e-01 2.85091773e-02
1.41449106e+00 4.14665937e-01 3.98120403e-01 3.61157805e-01
3.88097882e-01 7.22173810e-01 4.72367465e-01 1.43320309e-02
2.70811915e-01 6.66248560e-01 1.72105238e-01 -1.33104131e-01
-7.19539449e-02 -6.63556457e-01 1.76267549e-01 5.48457265e-01
4.18395698e-01 -5.96433282e-01 -1.02968752e+00 6.22936964e-01
-1.70405853e+00 -8.67210090e-01 4.84764785e-01 1.61842930e+00
1.07810462e+00 -1.76045954e-01 -4.78818089e-01 -5.63401759e-01
1.22940384e-01 4.87060398e-01 -5.42737603e-01 -6.05208337e-01
3.92054692e-02 4.59568322e-01 1.84415847e-01 6.98706508e-01
-8.10696363e-01 1.33625591e+00 7.20178699e+00 8.08834136e-01
-8.31013560e-01 -1.91275463e-01 3.10686320e-01 -7.18851313e-02
-7.67298043e-01 7.50168636e-02 -1.10079932e+00 -4.10534650e-01
9.05004621e-01 -3.54401290e-01 2.14591652e-01 1.00269961e+00
-7.07507059e-02 9.22149643e-02 -1.49313200e+00 7.32536435e-01
5.68731010e-01 -1.50690067e+00 7.25890100e-01 1.75899386e-01
4.69715476e-01 1.67605773e-01 2.51211017e-01 8.13561261e-01
5.82441270e-01 -1.26070774e+00 -1.33629188e-01 6.30725443e-01
4.80147809e-01 -4.62462574e-01 7.15115607e-01 2.26848707e-01
-9.24581051e-01 -2.42769271e-01 -7.58756340e-01 -2.38523353e-03
-8.98564011e-02 2.53410131e-01 -1.11408520e+00 3.29235733e-01
5.24068773e-01 8.68819714e-01 -9.18172359e-01 5.45702100e-01
-1.53569773e-01 4.74674881e-01 -8.97257179e-02 -6.29253626e-01
6.45037055e-01 1.87725559e-01 9.11628604e-02 1.41513324e+00
-4.06917036e-02 -4.82068881e-02 2.94969231e-01 1.02799082e+00
-2.34248519e-01 4.44753468e-01 -9.84533846e-01 -9.07789350e-01
2.61877567e-01 9.61690426e-01 -2.34275118e-01 -4.25900549e-01
-5.87523282e-01 6.75072014e-01 3.60990882e-01 8.24053526e-01
4.35748175e-02 -4.43078786e-01 6.07825816e-01 -1.81446776e-01
2.61932075e-01 -3.25996250e-01 -7.63167515e-02 -1.01059639e+00
-3.08993235e-02 -8.08229148e-01 6.10836208e-01 -9.10476089e-01
-1.52693450e+00 5.08062780e-01 2.69135058e-01 -5.17286181e-01
-7.11253285e-01 -9.06597674e-01 -4.63049948e-01 1.08127630e+00
-1.72756052e+00 -1.25990748e+00 2.28827149e-01 3.96810889e-01
8.02732825e-01 -4.62914705e-01 1.24369395e+00 -8.65362734e-02
-9.54871625e-02 5.15561581e-01 2.44819239e-01 1.66686147e-01
7.34465241e-01 -1.06155777e+00 2.96984643e-01 3.96191031e-01
7.96157539e-01 1.44590223e+00 7.28742555e-02 -5.97708881e-01
-1.56925583e+00 -6.76508546e-01 1.22632098e+00 -8.35629225e-01
8.04588974e-01 -3.17915976e-01 -9.70231712e-01 7.73098528e-01
7.59790897e-01 -1.79395869e-01 1.16702795e+00 9.59272802e-01
-8.16139996e-01 -6.90120980e-02 -4.92340922e-01 6.57411814e-01
6.96102083e-01 -1.19444239e+00 -1.29945505e+00 3.39162618e-01
1.30958354e+00 3.99279356e-01 -6.94045603e-01 2.64529467e-01
4.63632762e-01 -2.93932080e-01 1.36504102e+00 -1.21588826e+00
4.40849006e-01 2.45044813e-01 -2.99414873e-01 -1.24848676e+00
-3.33680153e-01 -7.99230412e-02 -3.01700979e-01 8.79427016e-01
7.01021791e-01 -4.62710291e-01 7.14258492e-01 8.94437492e-01
3.40394787e-02 -8.24778199e-01 -4.12079811e-01 -7.06860006e-01
5.31506300e-01 -5.57743609e-01 4.66798127e-01 8.78648460e-01
4.20451611e-01 9.20456707e-01 4.75204214e-02 -2.56574452e-01
-1.16422899e-01 1.19494490e-01 5.03302038e-01 -1.28916669e+00
-2.96474099e-01 -7.02100575e-01 -7.85885379e-02 -1.82607353e+00
4.38344687e-01 -1.23732114e+00 -7.26030767e-02 -1.93569505e+00
2.90083647e-01 -4.34068620e-01 -5.53019881e-01 6.81862950e-01
-2.81739026e-01 -1.37089178e-01 -1.15131356e-01 2.18298897e-01
-6.07782781e-01 7.21257091e-01 1.03215718e+00 -5.45669854e-01
-1.43707976e-01 -3.45391572e-01 -1.12368762e+00 2.94278979e-01
6.90726936e-01 -2.99316704e-01 -9.22194839e-01 -1.10635877e+00
6.81324720e-01 -2.94086665e-01 1.20370038e-01 -2.76240498e-01
5.29264867e-01 -8.18299130e-02 3.21891963e-01 -7.21844554e-01
3.72245431e-01 -8.81714344e-01 -9.81713414e-01 9.59157187e-04
-1.26197112e+00 1.98985785e-01 5.26925504e-01 9.46353674e-01
-5.79171836e-01 -4.64480460e-01 -1.01802386e-01 -1.87609464e-01
-8.33477080e-01 2.42412299e-01 -5.39740741e-01 2.93018579e-01
1.72336996e-01 6.31740093e-02 -5.05935848e-01 -6.58952713e-01
-4.81854528e-01 4.62206841e-01 -4.41252440e-02 1.13200355e+00
7.63095796e-01 -1.31603611e+00 -2.60368019e-01 2.23733950e-02
7.02499568e-01 -1.73885152e-01 -2.95431644e-01 -2.13568714e-02
9.52294469e-03 1.45489943e+00 4.27361935e-01 -3.88623238e-01
-1.01936817e+00 8.97361875e-01 1.01235405e-01 -6.49546325e-01
-3.48897725e-01 9.85613644e-01 3.98456573e-01 -4.21291411e-01
5.85567594e-01 -5.77600062e-01 -5.73330522e-01 1.66664332e-01
7.75947928e-01 -1.86291724e-01 -6.75125048e-02 -4.40896824e-02
-2.48864712e-03 4.84592378e-01 -6.59881115e-01 -3.86321843e-01
1.41950905e+00 -1.01063795e-01 -2.56539941e-01 4.25718129e-01
1.48555589e+00 -1.10031165e-01 -3.42367649e-01 -8.80283892e-01
7.53414750e-01 -4.19681787e-01 5.13426423e-01 -1.11422205e+00
-5.08450091e-01 1.29837894e+00 4.91798788e-01 2.03240052e-01
8.71622860e-01 4.61439908e-01 6.35525763e-01 1.32273173e+00
6.33221194e-02 -1.17319119e+00 6.59420133e-01 9.04699326e-01
1.08004844e+00 -1.30170858e+00 5.29417992e-02 1.65011093e-01
-3.56930584e-01 1.36122882e+00 6.29456460e-01 1.92702457e-01
7.91633129e-01 -1.54359967e-01 1.09792784e-01 -6.78531587e-01
-1.04691100e+00 -3.10778022e-01 9.74988639e-01 4.86847520e-01
8.31449449e-01 -3.67226899e-01 -1.98705494e-01 2.77653277e-01
1.20520353e-01 -8.28865767e-02 -2.45502200e-02 1.06129253e+00
-5.71123719e-01 -1.45992088e+00 5.95328361e-02 4.94922280e-01
-5.00950277e-01 -8.58666778e-01 -7.90014923e-01 5.91530979e-01
-5.55253446e-01 1.06734228e+00 2.51798421e-01 -1.73496813e-01
1.24904707e-01 5.37668288e-01 1.45606369e-01 -1.23763871e+00
-4.10964310e-01 -6.23584818e-03 3.52360457e-01 -4.71022159e-01
-3.54691505e-01 -5.14115281e-02 -1.01362109e+00 3.59228641e-01
-6.29685342e-01 5.99313855e-01 6.35972857e-01 8.06032717e-01
5.81913531e-01 2.99945474e-01 2.66285211e-01 -3.62472743e-01
-5.80519974e-01 -1.20270205e+00 -1.85138658e-01 1.20345004e-01
3.58319581e-01 -2.39351049e-01 -1.66391820e-01 -1.48180097e-01] | [11.112361907958984, 8.064037322998047] |
ee469afc-8a5f-4abc-938c-b6cff3b715b5 | adversarial-deep-learning-in-eeg-biometrics | 1903.11673 | null | http://arxiv.org/abs/1903.11673v1 | http://arxiv.org/pdf/1903.11673v1.pdf | Adversarial Deep Learning in EEG Biometrics | Deep learning methods for person identification based on
electroencephalographic (EEG) brain activity encounters the problem of
exploiting the temporally correlated structures or recording session specific
variability within EEG. Furthermore, recent methods have mostly trained and
evaluated based on single session EEG data. We address this problem from an
invariant representation learning perspective. We propose an adversarial
inference approach to extend such deep learning models to learn
session-invariant person-discriminative representations that can provide
robustness in terms of longitudinal usability. Using adversarial learning
within a deep convolutional network, we empirically assess and show
improvements with our approach based on longitudinally collected EEG data for
person identification from half-second EEG epochs. | ['Toshiaki Koike-Akino', 'Ozan Ozdenizci', 'Deniz Erdogmus', 'Ye Wang'] | 2019-03-27 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 3.70484084e-01 -2.96312541e-01 4.13683772e-01 -7.73704410e-01
-7.90548205e-01 -6.25527501e-01 4.57554072e-01 -1.43191054e-01
-6.49340987e-01 1.06768012e+00 3.58612835e-01 1.58953443e-01
-4.52373087e-01 -2.62137473e-01 -8.34699869e-01 -5.42087615e-01
-7.69527912e-01 2.49677673e-01 -5.59253871e-01 -3.72633711e-02
-1.06452189e-01 5.76091051e-01 -9.56444860e-01 4.55950320e-01
6.87419116e-01 9.44888055e-01 -6.26584053e-01 5.56998193e-01
6.36685073e-01 4.05943930e-01 -1.17989433e+00 -2.91718602e-01
2.82989800e-01 -3.52698505e-01 -7.13234723e-01 -2.28029862e-01
5.04319012e-01 -3.02918017e-01 -7.98387945e-01 9.79781210e-01
9.63446915e-01 6.73412234e-02 7.93547392e-01 -1.41458082e+00
-7.27789342e-01 5.25916874e-01 -1.36116952e-01 8.05246949e-01
6.13096654e-01 3.15404236e-01 3.01373422e-01 -3.34375381e-01
4.65288982e-02 8.79639983e-01 1.01514900e+00 8.05023074e-01
-1.62295628e+00 -1.16623795e+00 1.62079722e-01 5.05599380e-01
-1.53087342e+00 -4.15025353e-01 8.83975327e-01 -5.79577208e-01
9.98238266e-01 1.77897751e-01 6.30975306e-01 2.47138548e+00
7.63751507e-01 4.37694490e-01 1.48418903e+00 2.36604452e-01
3.80193114e-01 -1.54403716e-01 5.18520474e-01 1.78654473e-02
4.55246791e-02 5.08241057e-01 -7.72688627e-01 -2.15013579e-01
7.11673915e-01 1.34316668e-01 -4.55806494e-01 2.36021457e-04
-1.08575010e+00 4.36580747e-01 5.40223539e-01 3.72810125e-01
-5.75911880e-01 2.42646217e-01 9.21501338e-01 7.28725970e-01
4.30005401e-01 7.66169727e-01 -5.72881579e-01 -2.19481274e-01
-1.05951703e+00 2.77434528e-01 6.60530150e-01 6.65224791e-01
2.74829417e-01 2.47363999e-01 -6.71048641e-01 6.29077673e-01
-4.40274686e-01 4.00198579e-01 7.80637443e-01 -3.58178228e-01
2.30449736e-01 2.88875252e-01 -9.74277258e-02 -7.66654909e-01
-8.82881463e-01 -6.00559711e-01 -1.11932909e+00 9.31952968e-02
2.21166387e-01 -3.44796479e-01 -8.91304195e-01 2.00893235e+00
-5.33209801e-01 7.08142459e-01 -1.53659768e-02 8.10011744e-01
6.06176913e-01 -4.89595979e-02 3.16022545e-01 -1.63551778e-01
1.24263513e+00 -3.66740614e-01 -7.97858179e-01 -1.82946384e-01
-8.51164535e-02 9.94261727e-02 8.36396396e-01 5.19970834e-01
-8.95306408e-01 -6.54208899e-01 -1.16400719e+00 3.37704629e-01
-5.92778981e-01 -3.26740563e-01 6.26670241e-01 8.96228194e-01
-1.03029406e+00 7.87365198e-01 -9.37806010e-01 -2.03108877e-01
7.12279320e-01 9.10734534e-01 -6.31117940e-01 2.69544333e-01
-1.58775437e+00 7.11083949e-01 2.28835270e-01 9.21186730e-02
-1.23060560e+00 -9.37725842e-01 -7.18732774e-01 2.70449165e-02
-2.75431067e-01 -8.09577048e-01 7.65530229e-01 -1.21182442e+00
-1.51192987e+00 6.97823763e-01 3.82661931e-02 -8.57000053e-01
5.04052341e-01 -2.92935044e-01 -7.87681043e-01 2.88571864e-01
-8.38062391e-02 2.58782089e-01 9.25546527e-01 -8.90938997e-01
1.53697699e-01 -7.76813567e-01 -2.28559691e-02 -2.20916390e-01
-5.20623922e-01 2.08012745e-01 2.08687663e-01 -1.01556599e+00
-4.01339829e-01 -9.55264091e-01 9.55510661e-02 -4.44833636e-01
-3.70007396e-01 -6.66326508e-02 4.37853545e-01 -9.96395051e-01
7.13709176e-01 -1.94415450e+00 2.21168339e-01 3.05122852e-01
2.07747072e-01 -3.50688957e-03 -1.76193699e-01 6.13904856e-02
-6.23576105e-01 -8.23489949e-02 -3.17777872e-01 -4.40859646e-01
6.72051534e-02 -1.44193113e-01 -2.86980033e-01 6.75237119e-01
1.78017750e-01 1.19066787e+00 -8.59402061e-01 2.39081040e-01
9.17799622e-02 6.64186120e-01 -3.17072839e-01 4.58327830e-01
7.03684747e-01 1.09228384e+00 -2.58231498e-02 6.34439647e-01
5.81515014e-01 3.26196641e-01 -8.09492096e-02 -2.79294103e-01
4.62206155e-01 1.41949251e-01 -5.54899573e-01 2.05079651e+00
-4.04399782e-01 8.50160003e-01 -2.91290373e-01 -1.39118755e+00
7.12447524e-01 7.00969994e-01 8.04828644e-01 -8.32421124e-01
2.71564394e-01 -2.71155030e-01 8.02135766e-02 -3.22942615e-01
-1.28756061e-01 -7.69931749e-02 -4.26379234e-01 3.60590547e-01
5.79481602e-01 3.54077399e-01 -4.88229334e-01 -1.08917072e-01
1.74952304e+00 -6.01684526e-02 2.55804583e-02 -4.81979936e-01
3.89528096e-01 -7.67616391e-01 4.53021556e-01 1.01347268e+00
-6.90554857e-01 5.70840836e-01 2.76405156e-01 -6.06494486e-01
-8.32267940e-01 -1.49235260e+00 -6.68586433e-01 6.33131564e-01
-9.31522250e-02 -1.81966662e-01 -9.45035815e-01 -8.64718914e-01
1.18611626e-01 5.57435215e-01 -1.24895668e+00 -8.87622714e-01
-4.19657201e-01 -8.89683902e-01 1.08649421e+00 1.13541996e+00
4.64036405e-01 -1.15321624e+00 -3.47601056e-01 3.36668462e-01
7.69889727e-02 -1.14375055e+00 -2.55767643e-01 5.44452667e-01
-5.64666033e-01 -1.04132533e+00 -9.51719165e-01 -4.34284300e-01
4.64927554e-01 -5.38398683e-01 1.09183836e+00 -5.67149222e-01
-4.74124759e-01 8.52647603e-01 -1.63148701e-01 -6.61493182e-01
1.21881820e-01 -6.17884919e-02 9.77331460e-01 2.21509948e-01
8.37172627e-01 -1.27084339e+00 -8.27374279e-01 3.01131129e-01
-5.67466259e-01 -5.91927826e-01 2.32868373e-01 1.02303040e+00
2.52804071e-01 -3.34458947e-02 1.13577890e+00 -5.14432669e-01
9.23574030e-01 -5.34799874e-01 -3.79941761e-02 2.27855980e-01
-3.68847340e-01 -6.60343021e-02 7.02089787e-01 -7.03176916e-01
-6.94355309e-01 -1.43958718e-01 -9.64892432e-02 -6.95931613e-01
-4.81806964e-01 1.60779074e-01 -1.42428353e-01 -3.07485342e-01
8.28932524e-01 4.67123508e-01 -4.83660996e-01 -1.16740964e-01
-1.20893642e-01 6.16658032e-01 1.04442692e+00 -6.01418436e-01
7.43796587e-01 4.36263412e-01 -1.45623311e-01 -3.90424967e-01
-4.68103558e-01 -2.71659225e-01 -1.04218268e+00 -1.78711042e-01
9.91203487e-01 -1.15808582e+00 -1.01292861e+00 5.68227947e-01
-1.03770816e+00 -4.24245834e-01 -1.61065027e-01 6.67134821e-01
-8.56924474e-01 8.98763761e-02 -6.13476872e-01 -6.79939806e-01
-6.82697713e-01 -7.59908855e-01 1.03920889e+00 1.66543797e-02
-5.86269081e-01 -1.03444302e+00 4.02162939e-01 -3.23104039e-02
3.43903840e-01 5.55052936e-01 4.46598321e-01 -9.48624134e-01
4.20869924e-02 -5.85885227e-01 4.55744416e-02 4.43732172e-01
2.98064619e-01 -9.73760784e-01 -1.43137145e+00 -7.73904085e-01
1.67582348e-01 -3.93549681e-01 6.71724796e-01 4.26715463e-01
1.76349723e+00 -2.45531306e-01 -2.43062958e-01 1.19384193e+00
1.09944785e+00 1.73957482e-01 8.92348349e-01 2.22734958e-01
6.89412951e-01 3.44128042e-01 -4.03393626e-01 3.71736467e-01
-5.28216287e-02 7.11196184e-01 -4.73044328e-02 1.05491355e-01
1.99003056e-01 1.87803134e-02 3.91667426e-01 2.74655610e-01
-3.24185699e-01 -3.70569117e-02 -7.64942706e-01 4.50870186e-01
-1.64565372e+00 -1.21010077e+00 6.30705714e-01 2.12075710e+00
6.36220634e-01 -7.15526147e-03 2.22455785e-01 2.89045990e-01
4.02405977e-01 -2.94759959e-01 -8.40337336e-01 -3.38246554e-01
-1.74777359e-01 7.39697039e-01 4.19731438e-01 -2.19975531e-01
-1.22205174e+00 2.79887170e-01 7.41382599e+00 1.33918419e-01
-1.08812451e+00 3.52393270e-01 2.59843767e-01 -4.43285853e-01
2.11039543e-01 -7.56338179e-01 -3.19806933e-01 8.49070370e-01
1.42271972e+00 -3.78993839e-01 6.92126572e-01 3.35490912e-01
2.10074350e-01 4.22317386e-01 -1.86642826e+00 1.59274375e+00
3.03524703e-01 -7.56750286e-01 -1.61728308e-01 1.67771772e-01
6.39835000e-01 -7.18863979e-02 3.67526054e-01 6.77864254e-01
1.72314714e-04 -1.66387606e+00 2.58935153e-01 9.60086107e-01
8.39117587e-01 -9.91884470e-01 8.05844247e-01 -6.36866465e-02
-9.58906889e-01 -2.51758903e-01 -8.48514587e-02 -9.85906422e-02
-1.65796563e-01 8.00569728e-02 -5.30923963e-01 6.26988828e-01
1.03496373e+00 9.60948825e-01 -7.67055333e-01 8.81656051e-01
-6.51903078e-02 7.08229601e-01 -6.46267459e-03 5.02973616e-01
-2.58793738e-02 1.05372794e-01 5.92075527e-01 1.35970330e+00
1.64612293e-01 3.39863002e-02 -3.46610993e-02 1.10403085e+00
-2.80970722e-01 -3.66171420e-01 -9.72588718e-01 1.57824099e-01
2.51702219e-01 8.79745126e-01 -3.22953969e-01 -1.25711262e-01
-2.99762398e-01 1.65649486e+00 4.20705140e-01 5.75046241e-01
-8.10973525e-01 -3.85507911e-01 9.84536588e-01 -2.12323606e-01
-1.72968850e-01 -1.24128357e-01 -4.80098605e-01 -1.39270496e+00
1.25540689e-01 -9.38673854e-01 3.14159840e-01 -6.22585297e-01
-1.95378387e+00 8.97419870e-01 6.98784664e-02 -1.05472493e+00
-4.81391430e-01 -5.87251067e-01 -9.64255631e-01 1.11952841e+00
-1.03050101e+00 -1.11877179e+00 -3.51166368e-01 1.25052559e+00
3.92680734e-01 -5.15995800e-01 1.36193085e+00 4.10563976e-01
-7.71188319e-01 1.35347140e+00 1.37272239e-01 4.52593297e-01
8.64010572e-01 -1.43984926e+00 4.37905401e-01 7.19263077e-01
1.18468806e-01 1.00673175e+00 5.43487787e-01 -3.34468305e-01
-1.06677222e+00 -1.22442305e+00 2.44520888e-01 -1.05245554e+00
4.85872656e-01 -8.37685764e-01 -1.06626952e+00 9.54871833e-01
4.10550326e-01 5.27306497e-02 1.10094738e+00 5.47894061e-01
-4.47967499e-01 -3.09137434e-01 -1.36698508e+00 3.69118869e-01
1.23420238e+00 -1.21239126e+00 -9.72627044e-01 3.59526157e-01
2.32676238e-01 -1.07362524e-01 -1.24962044e+00 4.68853623e-01
7.08349347e-01 -5.02418101e-01 1.12190115e+00 -8.61476839e-01
-1.07292652e-01 1.45517334e-01 1.16094068e-01 -1.77257895e+00
-4.43023562e-01 -7.20640779e-01 -1.04428694e-01 9.48986650e-01
-6.99326349e-03 -6.01347506e-01 5.24515927e-01 8.07533681e-01
-4.22482826e-02 -3.23049247e-01 -1.13573921e+00 -1.14187455e+00
2.79218525e-01 -3.83005738e-01 7.79831052e-01 9.32672322e-01
2.85849452e-01 2.71867692e-01 -6.18682861e-01 3.88035357e-01
7.47978389e-01 -3.92830014e-01 4.53279376e-01 -1.26272821e+00
-1.82743341e-01 -1.41858295e-01 -1.15000188e+00 -2.63031334e-01
8.91849279e-01 -9.94769692e-01 -8.23016092e-03 -9.31404054e-01
5.10588765e-01 1.88057601e-01 -1.33128846e+00 4.67456788e-01
-1.91570103e-01 4.19204116e-01 -2.34337032e-01 -9.68272984e-02
-4.28714544e-01 5.78803122e-01 4.12317097e-01 -5.48278093e-01
-7.10863397e-02 -1.59023621e-03 -4.96881127e-01 4.47988927e-01
8.59802604e-01 -3.08346957e-01 -3.59371692e-01 -1.99554712e-01
-3.27567160e-01 -2.69223332e-01 9.45888519e-01 -1.58121371e+00
1.84918657e-01 4.89049762e-01 1.23212588e+00 -6.71150684e-02
4.25818086e-01 -8.08903515e-01 9.07561556e-02 3.06836724e-01
-5.37476659e-01 1.55082390e-01 6.25795245e-01 7.63239920e-01
-3.47385444e-02 4.30301368e-01 5.74162960e-01 3.80154327e-02
-5.33561409e-01 6.47363007e-01 -4.69392121e-01 -1.06561176e-01
9.70318973e-01 -2.58730620e-01 1.62231084e-02 -3.26725751e-01
-1.09058869e+00 3.28186229e-02 9.36832875e-02 6.44375980e-01
6.20942473e-01 -1.44952917e+00 -7.09139585e-01 5.61647177e-01
3.08230370e-01 -9.16688502e-01 4.38750535e-01 7.12832808e-01
3.24966729e-01 5.01706719e-01 -1.06689477e+00 -5.05654752e-01
-9.88620222e-01 7.21677244e-01 7.62056768e-01 -1.51550015e-02
-7.55631387e-01 7.96743572e-01 4.31890726e-01 -2.45185733e-01
3.97999495e-01 -1.39997259e-01 -2.20647737e-01 -2.74963193e-02
7.16439128e-01 1.56152919e-01 3.79910320e-01 -3.70002419e-01
-7.65645683e-01 1.24285027e-01 -5.45713678e-02 -1.44291818e-01
1.37364900e+00 1.36372671e-01 2.87136704e-01 5.38976848e-01
1.46294343e+00 -5.23677528e-01 -1.37657452e+00 6.39943555e-02
-1.67119622e-01 -1.40985653e-01 1.16238363e-01 -1.19225931e+00
-9.70254719e-01 7.40527332e-01 1.31785953e+00 -1.62032649e-01
1.14743471e+00 -3.54478598e-01 5.51243544e-01 5.32226920e-01
6.71323776e-01 -9.45141673e-01 1.12776995e-01 1.46362871e-01
1.10008597e+00 -1.17076385e+00 -2.01497361e-01 3.88992310e-01
-4.06195819e-01 1.00560474e+00 5.78540981e-01 -5.34451723e-01
7.85184860e-01 2.50106961e-01 -7.58028254e-02 -1.68202892e-01
-3.85319650e-01 1.45135537e-01 5.78812242e-01 1.23349071e+00
3.11671227e-01 1.89365238e-01 -1.64680332e-02 1.31879473e+00
-3.13632935e-01 5.72074801e-02 1.29896149e-01 5.39271176e-01
5.11212528e-01 -9.28396940e-01 -1.11590154e-01 5.82628608e-01
-7.17645466e-01 -5.90878315e-02 -4.18278873e-01 5.79571664e-01
6.71738386e-02 7.48319626e-01 2.54817873e-01 -7.14739859e-01
5.29400647e-01 3.89653057e-01 8.11635554e-01 -4.36976403e-01
-9.93440747e-01 -6.30917490e-01 -2.94598430e-01 -7.89159477e-01
-3.93875539e-01 -6.72225118e-01 -6.04415894e-01 -1.50874317e-01
2.85239100e-01 -6.53168112e-02 4.87223476e-01 1.02999771e+00
4.70401138e-01 1.13997781e+00 7.61031747e-01 -9.57517505e-01
-5.55783153e-01 -1.23479724e+00 -9.19650972e-01 7.78768778e-01
5.27392924e-01 -7.33312249e-01 -2.69647241e-01 -2.38523558e-02] | [13.22960376739502, 3.4742350578308105] |
33922071-78e0-4724-b9ca-658c2866f57a | disentangling-semantic-features-of | 2106.14192 | null | https://arxiv.org/abs/2106.14192v1 | https://arxiv.org/pdf/2106.14192v1.pdf | Disentangling semantic features of macromolecules in Cryo-Electron Tomography | Cryo-electron tomography (Cryo-ET) is a 3D imaging technique that enables the systemic study of shape, abundance, and distribution of macromolecular structures in single cells in near-atomic resolution. However, the systematic and efficient $\textit{de novo}$ recognition and recovery of macromolecular structures captured by Cryo-ET are very challenging due to the structural complexity and imaging limits. Even macromolecules with identical structures have various appearances due to different orientations and imaging limits, such as noise and the missing wedge effect. Explicitly disentangling the semantic features of macromolecules is crucial for performing several downstream analyses on the macromolecules. This paper has addressed the problem by proposing a 3D Spatial Variational Autoencoder that explicitly disentangle the structure, orientation, and shift of macromolecules. Extensive experiments on both synthesized and real cryo-ET datasets and cross-domain evaluations demonstrate the efficacy of our method. | ['Min Xu', 'Xiangrui Zeng', 'Yungeng Zhang', 'Jianye Pang', 'Kai Yi'] | 2021-06-27 | null | null | null | null | ['electron-tomography'] | ['medical'] | [ 3.90866809e-02 -4.96860951e-01 3.62658232e-01 -3.91129971e-01
-4.92135137e-01 -6.75676525e-01 5.28515160e-01 1.16781645e-01
-4.69595701e-01 1.02164817e+00 2.09566802e-01 -5.45426831e-02
-6.08488806e-02 -4.45927113e-01 -8.66437733e-01 -1.35396314e+00
-3.76393506e-03 8.77610624e-01 -1.74792111e-01 1.28805041e-01
2.04847962e-01 8.17975581e-01 -1.24673200e+00 1.69432938e-01
5.95614314e-01 7.01566935e-01 6.74282312e-01 5.43404996e-01
-3.57390106e-01 5.57214499e-01 -1.81177229e-01 -1.53989598e-01
-3.13009173e-02 -2.68108964e-01 -4.85961735e-01 1.57121271e-02
3.91948313e-01 -7.95045123e-02 -4.45220947e-01 1.06197178e+00
5.34395993e-01 2.43557692e-01 1.00542998e+00 -4.73825485e-01
-7.75572538e-01 -1.55055998e-02 -3.86782229e-01 3.72009307e-01
1.73900053e-01 1.96708262e-01 8.53309214e-01 -9.54642057e-01
1.12923992e+00 9.48820591e-01 2.30271474e-01 4.17360842e-01
-1.46406591e+00 -4.80734199e-01 -3.63887139e-02 1.52338475e-01
-1.03917170e+00 -3.96977425e-01 8.26257825e-01 -8.28215301e-01
1.01460910e+00 2.63238531e-02 4.04144734e-01 1.29610157e+00
5.46711683e-01 1.41259208e-01 1.30188167e+00 -3.39703053e-01
2.65432149e-01 1.93335190e-02 2.43510991e-01 7.36299098e-01
1.65668145e-01 1.03276297e-01 -3.27861249e-01 -1.71606511e-01
9.34632778e-01 2.34994203e-01 -3.51224393e-01 -6.96302116e-01
-1.25679708e+00 5.11909783e-01 2.72345036e-01 2.85826445e-01
-4.56734866e-01 -5.51179796e-02 3.45622510e-01 -7.34210983e-02
2.53798425e-01 6.68087244e-01 -4.84555095e-01 4.63066585e-02
-6.28222346e-01 1.03826918e-01 5.67636907e-01 6.65735185e-01
8.48782897e-01 -7.75342435e-02 3.74602586e-01 4.79593635e-01
2.36631379e-01 3.70249569e-01 3.78979445e-01 -1.11250424e+00
-3.51939052e-02 4.41582620e-01 2.63126075e-01 -6.16658151e-01
-3.98210466e-01 -2.00656950e-02 -9.71914768e-01 3.19258213e-01
5.20286560e-01 2.26621237e-02 -1.01542068e+00 1.65449023e+00
6.07947230e-01 -5.10591008e-02 -1.42965496e-01 1.10640073e+00
9.95244205e-01 6.29395962e-01 1.24512471e-01 -3.54711592e-01
1.27691877e+00 -5.31617999e-01 -6.63888693e-01 8.29790160e-02
-3.91118601e-03 -5.49550653e-01 5.10424376e-01 -5.58563648e-03
-8.22936952e-01 -4.51046020e-01 -9.95041549e-01 -4.42984700e-01
-4.16331798e-01 -1.94587708e-01 7.79595494e-01 3.98036152e-01
-3.68006229e-01 8.08116972e-01 -1.09726274e+00 -1.00008965e-01
2.53914207e-01 4.09991533e-01 -7.53476977e-01 -4.39725816e-02
-6.48466945e-01 7.49775231e-01 1.01481207e-01 2.42450293e-02
-8.03207040e-01 -8.80849123e-01 -6.60808146e-01 2.03791797e-01
5.01172207e-02 -7.14637756e-01 5.69486141e-01 -4.05850500e-01
-1.38385963e+00 9.00154591e-01 -5.83701193e-01 1.08309858e-01
2.18447953e-01 2.24143729e-01 -5.67341223e-02 2.62041271e-01
-1.54582575e-01 3.36774260e-01 6.47812903e-01 -1.38018858e+00
1.28273025e-01 -8.01296294e-01 -3.15738440e-01 -6.08735345e-02
1.84266701e-01 -1.55089438e-01 -3.95519240e-03 -1.52969882e-01
5.08949697e-01 -5.56387424e-01 -1.15385935e-01 -1.40472010e-01
-2.63908088e-01 2.78684080e-01 6.55839920e-01 -7.39158213e-01
3.23335141e-01 -1.92762923e+00 8.16806138e-01 -1.27213653e-02
6.23903453e-01 -1.08375140e-01 1.12132430e-01 5.46263158e-01
-2.27600187e-01 -6.92319050e-02 -8.96888599e-02 -2.49503642e-01
-4.28317897e-02 8.72180983e-02 -1.09990679e-01 7.65423894e-01
-1.58817507e-02 8.62488151e-01 -6.35475218e-01 -4.40826088e-01
4.39384282e-01 8.70096326e-01 -2.06732020e-01 1.79267019e-01
-3.17128479e-01 7.68173575e-01 -3.82311553e-01 7.98149705e-01
8.09776366e-01 -6.47262096e-01 8.44936728e-01 -6.03203297e-01
-3.18169385e-01 4.71688062e-02 -7.59286165e-01 1.40848064e+00
1.29664555e-01 5.31419098e-01 5.89013636e-01 -1.00859344e+00
5.88127375e-01 1.34510562e-01 6.54650092e-01 -7.78475404e-01
1.62864760e-01 -4.74095754e-02 2.36232852e-04 -5.94659388e-01
3.25219810e-01 -7.27148592e-01 2.19220892e-01 4.04157519e-01
1.82507694e-01 1.66427970e-01 -1.68454349e-01 1.90024711e-02
6.59151435e-01 4.09812331e-01 1.17779426e-01 -5.19481897e-01
3.01847696e-01 -2.73726046e-01 5.36447525e-01 4.81718868e-01
-4.60118145e-01 6.11970842e-01 3.11492324e-01 -7.94884145e-01
-1.76072705e+00 -1.20899069e+00 -3.57096076e-01 7.45296121e-01
2.10233137e-01 2.23626625e-02 -5.18569112e-01 -2.38898680e-01
1.44250005e-01 -8.78970474e-02 -7.07066894e-01 1.61143571e-01
-5.92154860e-01 -1.18654227e+00 6.28478155e-02 3.66383225e-01
-1.54962152e-01 -8.85718822e-01 -4.28822100e-01 2.78935730e-01
-9.74352285e-02 -1.30630827e+00 -2.33974919e-01 3.25351894e-01
-7.77175128e-01 -1.13765812e+00 -5.84800422e-01 -5.98997831e-01
7.46798396e-01 2.30074748e-01 8.19020510e-01 -2.08264530e-01
-6.52809381e-01 7.71273747e-02 1.22329608e-01 1.04789762e-02
-3.48439842e-01 -2.14897558e-01 3.64976346e-01 -7.07281232e-02
2.45360747e-01 -9.68097150e-01 -7.37147808e-01 2.74140649e-02
-6.35394156e-01 -1.39067262e-01 3.47723186e-01 1.04702139e+00
1.06765532e+00 -2.59978510e-02 1.89246878e-01 -7.37978935e-01
3.72443050e-01 -2.22186208e-01 -7.18924105e-01 2.43575647e-01
-3.45441043e-01 4.08537239e-01 6.62299573e-01 -1.88965321e-01
-1.13552368e+00 1.47797894e-02 -9.13655236e-02 -4.99074608e-01
-4.18483287e-01 2.67866611e-01 -3.56238157e-01 -7.89583772e-02
2.18085691e-01 6.33426130e-01 1.95804536e-01 -5.61787426e-01
-4.01343703e-02 2.19316706e-01 4.88603532e-01 -9.60193574e-01
2.55943477e-01 8.52442086e-01 3.04469258e-01 -1.19985890e+00
-3.46183270e-01 -3.31544310e-01 -8.34388673e-01 -2.98958551e-03
8.80547643e-01 -8.67006242e-01 -1.27706492e+00 3.95549685e-01
-9.55635250e-01 4.47527505e-02 1.35607019e-01 5.43349028e-01
-3.79476368e-01 9.50367689e-01 -9.30640399e-01 -7.65604019e-01
-2.78600574e-01 -1.42536116e+00 1.02768373e+00 1.38766885e-01
4.40201759e-02 -9.16599154e-01 2.41681904e-01 5.42127907e-01
1.52641818e-01 3.26796472e-01 1.38524485e+00 -1.10582761e-01
-1.06283450e+00 9.77357700e-02 -4.15527105e-01 -1.74544498e-01
1.81882665e-01 1.38976589e-01 -1.06245065e+00 -3.90123218e-01
1.08282946e-01 -1.78032085e-01 9.21644628e-01 7.46956527e-01
9.07074332e-01 1.84717402e-01 -4.43165869e-01 8.62950444e-01
1.50794005e+00 2.79550880e-01 5.21906734e-01 1.67455301e-01
8.37669969e-01 7.88766742e-01 5.82918860e-02 4.32622641e-01
-3.61768343e-02 3.87046397e-01 4.76824969e-01 2.96008110e-01
1.69019014e-01 -8.20424929e-02 1.51927080e-02 1.00856471e+00
-4.38920379e-01 -2.73559481e-01 -6.24006391e-01 2.33261734e-01
-1.41430056e+00 -1.22512805e+00 7.47966841e-02 1.81937706e+00
5.54655552e-01 -2.71417171e-01 -1.21888421e-01 -3.29718590e-01
7.02346861e-01 2.05366477e-01 -9.66524541e-01 -5.88540994e-02
-3.59550059e-01 2.70038366e-01 2.55565763e-01 5.83264053e-01
-8.54741693e-01 5.19744098e-01 6.97370386e+00 3.78134429e-01
-1.10849750e+00 -7.69104436e-02 4.01309520e-01 -5.09397276e-02
-4.18964207e-01 -6.21817447e-02 -8.93224537e-01 6.84948981e-01
7.46720791e-01 9.78716984e-02 4.89513934e-01 5.59220791e-01
4.94203456e-02 2.33338084e-02 -9.83186662e-01 8.78689766e-01
-1.98338374e-01 -1.77276480e+00 -2.49608979e-02 2.02942148e-01
3.92761648e-01 3.51338416e-01 -5.97355925e-02 -1.08711809e-01
9.71185341e-02 -9.86138999e-01 4.90151465e-01 6.36010587e-01
8.69716585e-01 -5.05234540e-01 5.81782937e-01 3.95040482e-01
-9.19908643e-01 1.92737162e-01 -5.49388707e-01 -4.27936912e-02
1.30600631e-01 7.57157445e-01 -7.01052070e-01 3.89844626e-01
7.16813922e-01 4.64150637e-01 8.36843103e-02 3.31784904e-01
3.00932050e-01 1.19435027e-01 -2.67335832e-01 -1.79567844e-01
-2.19995230e-01 -9.70898151e-01 4.62412357e-01 1.04385817e+00
1.61411092e-01 3.46477419e-01 -5.99183477e-02 1.26805031e+00
-2.68172890e-01 -3.73251766e-01 -2.75161624e-01 -5.85971057e-01
5.14811218e-01 1.06054640e+00 -6.76412165e-01 -7.57354647e-02
-2.35896230e-01 8.68803084e-01 6.11911535e-01 5.88250816e-01
-5.50108612e-01 -1.01909362e-01 1.04640210e+00 1.57894179e-01
5.83618641e-01 -6.90563202e-01 -1.28254294e-01 -1.45226181e+00
-3.65927704e-02 -4.68122214e-01 -5.00551388e-02 -7.65542507e-01
-1.49424505e+00 3.30046535e-01 -4.30132091e-01 -3.45380574e-01
1.99397266e-01 -1.03967869e+00 -1.66128039e-01 9.68761086e-01
-1.30726695e+00 -9.22234178e-01 -1.27079621e-01 1.73253641e-01
1.44415870e-01 -6.54358370e-03 9.93540108e-01 3.17596048e-01
-8.40472102e-01 9.87217650e-02 9.40136731e-01 -1.28992587e-01
5.12307286e-01 -1.30057919e+00 2.12746352e-01 3.14186722e-01
-2.59266526e-01 9.56519306e-01 9.49152112e-01 -7.01337814e-01
-1.90696728e+00 -6.40060425e-01 6.14186466e-01 -7.17130065e-01
4.60215062e-01 -6.90330684e-01 -1.01055634e+00 6.63862944e-01
4.88256738e-02 3.44938561e-02 8.85644257e-01 8.38819817e-02
-4.24220026e-01 3.06138426e-01 -1.27964115e+00 2.49992073e-01
9.14131105e-01 -1.02446055e+00 -5.09300232e-01 3.97372246e-01
5.46758115e-01 -3.18800926e-01 -1.30287075e+00 1.52230322e-01
9.90796447e-01 -1.19064844e+00 1.38801312e+00 -8.86380553e-01
5.72770298e-01 -2.24440739e-01 -5.77204287e-01 -9.30305898e-01
-6.51832163e-01 -3.03122163e-01 -1.10824876e-01 8.08840394e-01
1.62295271e-02 -5.45836508e-01 1.04439855e+00 8.23767006e-01
-3.68478149e-02 -5.59410572e-01 -1.10487854e+00 -4.34089750e-01
1.17368571e-01 1.52068794e-01 6.88365579e-01 1.12131095e+00
-1.50807351e-01 3.68449032e-01 -2.88559288e-01 3.06528479e-01
9.31620181e-01 7.48693347e-01 3.42843205e-01 -1.20342183e+00
-4.02254492e-01 6.58198353e-03 -3.37632895e-01 -8.72069836e-01
3.32150489e-01 -7.04191923e-01 -1.19444229e-01 -9.80836153e-01
6.99478209e-01 -6.43044338e-02 -2.64652908e-01 -3.18139642e-01
9.66452062e-02 -2.10075885e-01 -9.25838426e-02 3.06678057e-01
-3.81859124e-01 8.32167506e-01 1.48138094e+00 -1.26845017e-01
2.78869033e-01 -6.20024562e-01 -3.67755175e-01 3.75874788e-01
3.91823143e-01 -2.66542196e-01 -5.40352762e-02 -4.46329981e-01
1.75102174e-01 3.68696988e-01 3.70436192e-01 -6.47964180e-01
1.24052487e-01 -3.17494333e-01 7.84315407e-01 -7.69469857e-01
7.68865585e-01 -7.49654412e-01 3.33370328e-01 1.40088513e-01
-3.73408385e-02 -3.22300382e-02 1.58549100e-02 9.27190185e-01
3.32264253e-03 -6.44035414e-02 9.23692822e-01 -6.73746526e-01
-4.03906435e-01 4.84356374e-01 -5.68323612e-01 -1.44095391e-01
6.48717046e-01 -2.62498945e-01 -4.30563241e-01 -5.17880060e-02
-8.29272032e-01 -1.05239846e-01 9.92040932e-01 -3.56733263e-01
7.27506518e-01 -9.36605990e-01 -1.89509094e-01 5.16804934e-01
-1.97678238e-01 7.21571669e-02 7.07149804e-01 4.84149635e-01
-7.15668380e-01 5.75899661e-01 -4.94372010e-01 -4.76735055e-01
-1.14129186e+00 7.35858798e-01 5.44949710e-01 -1.31931156e-01
-4.78787929e-01 7.03394473e-01 4.83857781e-01 -7.40071535e-01
-2.04006851e-01 -7.87998363e-02 8.67854133e-02 -7.61194155e-02
4.11995620e-01 3.53999853e-01 -8.67771596e-05 -6.44604206e-01
-1.93301961e-01 6.68106258e-01 -3.54947925e-01 3.07005048e-01
1.69194734e+00 -2.71266490e-01 -3.60698789e-01 2.06287503e-01
1.21179485e+00 -2.28357837e-01 -1.42409909e+00 -1.96260601e-01
-2.67587721e-01 -2.96712816e-01 -4.13418934e-02 -3.40031624e-01
-7.82540619e-01 1.11495900e+00 3.85011941e-01 -2.14031264e-01
5.06923437e-01 7.03224018e-02 8.15292716e-01 6.03958905e-01
3.62778842e-01 -8.25381756e-01 -1.91699684e-01 3.67739469e-01
4.12961572e-01 -1.32578743e+00 2.50696599e-01 -2.76760042e-01
-5.50861936e-03 1.07313406e+00 3.60583454e-01 7.48105794e-02
5.31085432e-01 4.63596731e-01 -1.73827410e-01 -5.74212492e-01
-6.68566942e-01 3.32555383e-01 -2.54997820e-01 7.03413904e-01
4.65295821e-01 8.29758868e-02 2.72835106e-01 3.90918523e-01
2.12490961e-01 -3.07836652e-01 4.69120830e-01 1.02443624e+00
-3.68395716e-01 -1.02310801e+00 -2.57464558e-01 1.65140480e-01
-5.07043302e-01 1.08965531e-01 -4.28838253e-01 5.72602987e-01
-8.91012475e-02 2.56407648e-01 7.86224231e-02 -1.01745971e-01
1.56935547e-02 3.94817054e-01 1.01922250e+00 -2.28194743e-01
-1.76783893e-02 1.95858970e-01 -2.36286104e-01 -3.01117122e-01
-6.00244701e-01 -7.61658490e-01 -1.48552465e+00 -4.88406092e-01
-3.84557068e-01 2.72492677e-01 8.84821653e-01 8.17027986e-01
7.41085708e-01 4.07072663e-01 4.60829586e-01 -1.09056461e+00
-5.49606144e-01 -8.05282414e-01 -9.97568727e-01 6.10833347e-01
5.98432541e-01 -7.38304734e-01 -3.37420702e-01 1.48058146e-01] | [13.297715187072754, -3.0710034370422363] |
80a7869d-fa80-4d4f-a135-7d8ddbe1d70d | luke-graph-a-transformer-based-approach-with | 2303.06675 | null | https://arxiv.org/abs/2303.06675v1 | https://arxiv.org/pdf/2303.06675v1.pdf | LUKE-Graph: A Transformer-based Approach with Gated Relational Graph Attention for Cloze-style Reading Comprehension | Incorporating prior knowledge can improve existing pre-training models in cloze-style machine reading and has become a new trend in recent studies. Notably, most of the existing models have integrated external knowledge graphs (KG) and transformer-based models, such as BERT into a unified data structure. However, selecting the most relevant ambiguous entities in KG and extracting the best subgraph remains a challenge. In this paper, we propose the LUKE-Graph, a model that builds a heterogeneous graph based on the intuitive relationships between entities in a document without using any external KG. We then use a Relational Graph Attention (RGAT) network to fuse the graph's reasoning information and the contextual representation encoded by the pre-trained LUKE model. In this way, we can take advantage of LUKE, to derive an entity-aware representation; and a graph model - to exploit relation-aware representation. Moreover, we propose Gated-RGAT by augmenting RGAT with a gating mechanism that regulates the question information for the graph convolution operation. This is very similar to human reasoning processing because they always choose the best entity candidate based on the question information. Experimental results demonstrate that the LUKE-Graph achieves state-of-the-art performance on the ReCoRD dataset with commonsense reasoning. | ['Kourosh Kiani', 'Shima Foolad'] | 2023-03-12 | null | null | null | null | ['reading-comprehension'] | ['natural-language-processing'] | [-1.25478348e-03 7.51200557e-01 -1.96151704e-01 -4.37804908e-01
-3.88703436e-01 -4.55498666e-01 4.41720486e-01 4.30041790e-01
-3.77500772e-01 3.35548550e-01 5.38585782e-01 -4.41621959e-01
-3.17524642e-01 -1.36485147e+00 -8.89326930e-01 -2.85749733e-01
3.09724122e-01 5.25257409e-01 4.59797055e-01 -4.80613828e-01
3.39450002e-01 -2.65607387e-02 -1.13310504e+00 5.27254820e-01
1.36024976e+00 9.56837296e-01 2.60066241e-01 3.65199059e-01
-5.03359318e-01 1.36991322e+00 -4.26972717e-01 -9.31239247e-01
-2.62238383e-01 -6.07387245e-01 -1.26384652e+00 -3.46188486e-01
2.08042949e-01 -2.93745399e-01 -5.31132579e-01 1.12339711e+00
4.16717649e-01 4.11246091e-01 4.95572120e-01 -8.67874980e-01
-1.56618214e+00 1.36157942e+00 -2.03746602e-01 3.89620155e-01
4.13287103e-01 -2.79195961e-02 1.57629538e+00 -7.49472260e-01
7.09845185e-01 1.40394604e+00 3.64172459e-01 3.81684095e-01
-9.01200056e-01 -3.37465823e-01 6.03096664e-01 1.05301619e+00
-1.26347721e+00 1.59713887e-02 9.09910679e-01 -8.27963799e-02
1.42154956e+00 1.02160834e-01 7.72595823e-01 7.45350540e-01
-1.37792900e-01 9.19034719e-01 8.36820722e-01 -6.30555391e-01
-2.46372782e-02 -1.62248909e-01 7.19922721e-01 1.00503230e+00
2.85552561e-01 -4.44412559e-01 -4.58261192e-01 1.15354098e-01
6.06876493e-01 -1.04343966e-01 -7.97815561e-01 -2.98853546e-01
-9.60921049e-01 8.12821507e-01 1.12262225e+00 4.57635015e-01
-4.65909153e-01 8.86320472e-02 2.09766522e-01 3.32744032e-01
6.26867265e-02 6.90922916e-01 -3.92597347e-01 3.89410377e-01
-3.95191878e-01 1.15477219e-01 8.25338244e-01 9.99794304e-01
5.98526061e-01 -3.55076760e-01 -5.83584547e-01 6.94805145e-01
5.76656342e-01 1.91286013e-01 5.24151683e-01 -4.74289060e-01
8.99795294e-01 1.28988469e+00 -4.45347279e-01 -1.24169636e+00
-4.31914568e-01 -6.49178267e-01 -6.34985030e-01 -4.71374571e-01
1.13785580e-01 2.34838963e-01 -9.09663677e-01 1.84580493e+00
3.07997018e-01 1.04848176e-01 3.30673456e-01 7.75629342e-01
1.45261097e+00 4.26776767e-01 2.43362933e-01 4.01042700e-02
1.59997773e+00 -1.15127563e+00 -1.05035281e+00 -2.84555316e-01
6.68457270e-01 -4.02197279e-02 1.32554007e+00 9.52347070e-02
-9.49609876e-01 -3.80660176e-01 -1.07423937e+00 -7.28139222e-01
-8.32408190e-01 5.84987253e-02 5.18953562e-01 1.96022898e-01
-1.07129920e+00 4.69788194e-01 -4.39773798e-01 -4.80386764e-01
5.81016660e-01 9.31340531e-02 -1.29436895e-01 -3.04682493e-01
-1.81349826e+00 1.15369153e+00 8.78458679e-01 2.74298310e-01
-4.43651557e-01 -3.88440132e-01 -1.06758165e+00 5.78478694e-01
8.91124547e-01 -1.07599461e+00 9.18038130e-01 -4.19932008e-01
-1.21836543e+00 7.99496651e-01 -2.07563579e-01 -4.99795496e-01
-6.71285242e-02 -4.03432101e-01 -4.88490492e-01 2.56424069e-01
8.42920840e-02 2.10763782e-01 4.47792888e-01 -1.05117238e+00
-1.61711216e-01 -5.85722029e-01 5.64396024e-01 4.55301583e-01
-1.31934285e-01 -1.64764047e-01 -7.62311101e-01 -5.94320476e-01
1.71858549e-01 -4.83157575e-01 1.48243845e-01 -5.18991470e-01
-6.42621458e-01 -7.52500296e-01 4.48561788e-01 -1.08999944e+00
1.76139164e+00 -1.83734417e+00 5.29509544e-01 2.33943462e-01
6.63447618e-01 3.57618570e-01 -7.33923465e-02 5.29173851e-01
8.96593034e-02 3.20463419e-01 -3.95760499e-02 1.22927368e-01
1.68232054e-01 2.05605805e-01 -4.34795260e-01 -2.20732287e-01
2.72662222e-01 1.52790010e+00 -1.18619335e+00 -4.46364820e-01
-1.39768913e-01 9.67144817e-02 -6.28186345e-01 3.02003652e-01
-5.26936173e-01 -3.82450894e-02 -6.61331415e-01 4.67425615e-01
4.91590858e-01 -7.92029560e-01 3.33305180e-01 -4.80203569e-01
5.55678308e-01 7.07545996e-01 -7.90051997e-01 1.67739570e+00
-2.54316866e-01 3.16421300e-01 -3.67801368e-01 -1.01901913e+00
8.84968877e-01 1.10646568e-01 -3.52342159e-01 -8.10459435e-01
3.11639816e-01 5.94802536e-02 1.71647951e-01 -7.83651412e-01
4.81147438e-01 -1.98130980e-02 3.39918137e-02 3.03034246e-01
4.36647385e-01 7.19995424e-03 3.02197963e-01 7.44534492e-01
1.13783550e+00 2.08703548e-01 4.57992226e-01 5.46367839e-04
6.25231564e-01 -1.66339129e-01 2.61020184e-01 6.12411201e-01
1.43214419e-01 1.93191171e-01 7.08213449e-01 -4.64641377e-02
-2.53116041e-01 -1.02027273e+00 3.23718399e-01 1.01324785e+00
3.60389769e-01 -8.29391241e-01 -7.18744040e-01 -9.33313847e-01
-1.23724952e-01 1.15992832e+00 -7.78902888e-01 -6.21980846e-01
-5.75309396e-01 -4.80467170e-01 4.31019664e-01 8.15544367e-01
9.81039166e-01 -1.29339910e+00 -2.16347411e-01 2.97718227e-01
-4.55378294e-01 -1.26319861e+00 -1.18683964e-01 1.70375347e-01
-5.23609340e-01 -1.32243526e+00 -7.30432719e-02 -7.89680839e-01
6.46100581e-01 3.00541800e-02 1.53782010e+00 5.25153458e-01
3.33877087e-01 3.30117702e-01 -7.03152299e-01 -2.50865877e-01
-2.24330239e-02 2.32913777e-01 -6.46063268e-01 -1.51893511e-01
7.54405618e-01 -5.81347406e-01 -3.80692095e-01 -5.74955083e-02
-8.08710515e-01 2.62436777e-01 4.70745265e-01 8.02944839e-01
6.74856901e-01 1.00965239e-01 8.44981313e-01 -1.16987908e+00
1.00500703e+00 -6.30990505e-01 -2.72769421e-01 8.11549306e-01
-6.01911783e-01 4.69977707e-01 7.80102789e-01 -5.61439507e-02
-1.24230158e+00 -6.54514492e-01 -5.16562946e-02 -4.27266538e-01
4.17895794e-01 1.04338670e+00 -6.66187882e-01 3.29125881e-01
5.68079472e-01 1.86271697e-01 -5.41675210e-01 -3.33858758e-01
9.99612451e-01 5.05574346e-01 6.66831970e-01 -7.14439869e-01
6.46972895e-01 9.25463066e-02 -2.41339833e-01 -1.90586925e-01
-1.44983339e+00 -4.26539660e-01 -5.70103645e-01 4.83999699e-02
1.23860109e+00 -8.31269920e-01 -6.68295622e-01 3.23976099e-01
-1.28549182e+00 -3.46869886e-01 -2.91690111e-01 2.53002286e-01
-2.87592441e-01 2.59775817e-01 -5.69425881e-01 -4.58338022e-01
-4.29639995e-01 -8.35705459e-01 9.37175870e-01 4.65494961e-01
-8.38009734e-03 -1.09707212e+00 -2.28254393e-01 5.82532883e-01
2.85770595e-01 7.77576268e-02 1.52432847e+00 -9.58051145e-01
-9.00180697e-01 1.65328592e-01 -6.18758321e-01 1.08893096e-01
-1.03518907e-02 -4.74734128e-01 -7.92667627e-01 1.64320558e-01
-1.49995908e-01 -4.14373457e-01 1.15257525e+00 4.50324342e-02
1.20501196e+00 -3.50285769e-01 -3.78941357e-01 5.58499932e-01
1.41690326e+00 4.33110110e-02 8.10158968e-01 3.35437298e-01
1.25884736e+00 3.99401307e-01 3.76968801e-01 -3.25205550e-02
1.06742048e+00 4.13437247e-01 4.99673814e-01 3.44823807e-01
-3.12371075e-01 -8.45280528e-01 8.52185488e-02 9.94113743e-01
-2.61499792e-01 -4.33712155e-01 -8.91427755e-01 5.66789567e-01
-2.01837754e+00 -1.03594112e+00 -1.12491570e-01 1.75247788e+00
9.76353109e-01 1.33545864e-02 -4.81670678e-01 4.24439237e-02
7.75695443e-01 1.53574988e-01 -6.30990267e-01 -3.59280735e-01
-2.05076352e-01 3.98713827e-01 1.64673209e-01 4.71050769e-01
-8.74511302e-01 1.36863923e+00 4.65898371e+00 7.29914546e-01
-7.03840256e-01 3.57767418e-02 1.57006383e-01 4.01536137e-01
-7.00327635e-01 1.01445973e-01 -7.54012406e-01 8.54525864e-02
6.62670970e-01 -3.84244710e-01 6.04644418e-01 6.47548676e-01
-3.30094039e-01 1.14217348e-01 -1.00492513e+00 8.87395203e-01
4.73095149e-01 -1.30274498e+00 5.75610876e-01 -2.29553789e-01
3.90479892e-01 -8.15145820e-02 -4.36418265e-01 9.27037477e-01
5.29966474e-01 -1.08566892e+00 6.65730834e-01 6.73245847e-01
2.54498899e-01 -4.50059891e-01 6.92583144e-01 3.98592085e-01
-1.48965824e+00 -1.40659928e-01 -4.30444181e-01 4.74609695e-02
1.04217969e-01 3.99192512e-01 -6.26208067e-01 1.11413670e+00
7.04002202e-01 7.66452432e-01 -9.90553200e-01 7.83578277e-01
-1.26440346e+00 6.60851777e-01 -5.16185444e-03 -2.73450017e-01
1.57338068e-01 -1.21719308e-01 3.92937005e-01 9.38758433e-01
7.16753453e-02 4.60938632e-01 3.58272195e-02 1.23655641e+00
-6.06171608e-01 2.36497760e-01 -3.87035340e-01 -2.48949111e-01
5.24069607e-01 1.14012170e+00 -5.69251835e-01 -6.41934872e-01
-4.34717029e-01 1.02502465e+00 1.11806560e+00 4.48551774e-01
-8.36756706e-01 -6.43815815e-01 1.62252560e-02 -5.95452264e-02
7.09266663e-01 8.41230527e-02 -4.39651497e-02 -1.51102328e+00
2.62366116e-01 -7.20596313e-01 9.22523737e-01 -1.16756320e+00
-1.36002946e+00 6.76288605e-01 2.77699791e-02 -6.14015579e-01
-1.05530202e-01 -6.10355377e-01 -6.58075213e-01 9.31652308e-01
-1.77788723e+00 -1.33954751e+00 -4.89477932e-01 7.11046219e-01
2.37907752e-01 1.43337086e-01 6.62077427e-01 -6.51288554e-02
-4.47852105e-01 4.71935540e-01 -4.85184252e-01 4.82668787e-01
3.09559703e-01 -1.64643228e+00 4.71887827e-01 9.37050819e-01
3.82212579e-01 1.13321936e+00 9.21920389e-02 -8.03848505e-01
-1.29546237e+00 -1.19264042e+00 1.10004807e+00 -5.57836771e-01
7.46697187e-01 -7.49020055e-02 -1.47761774e+00 1.08172202e+00
5.07008374e-01 1.66831035e-02 6.56791091e-01 3.97209942e-01
-7.46211290e-01 2.08485782e-01 -8.80494714e-01 6.96862698e-01
1.46084702e+00 -7.57977366e-01 -1.39937818e+00 3.10339592e-02
1.13181067e+00 -5.06246328e-01 -8.06111455e-01 2.13358611e-01
-1.62199195e-02 -6.51435852e-01 7.55396307e-01 -8.63400519e-01
4.35416341e-01 -5.15247822e-01 -4.79928814e-02 -1.53424108e+00
-6.27050042e-01 -3.05389822e-01 -6.05965137e-01 1.41543901e+00
4.68733847e-01 -6.32855773e-01 2.18129396e-01 5.57418346e-01
-3.15488935e-01 -9.54864264e-01 -5.86784005e-01 -5.60896695e-01
-1.16858125e-01 -2.18671173e-01 9.18033242e-01 1.07609010e+00
3.88275892e-01 1.07872558e+00 1.43383473e-01 4.14311260e-01
3.12441647e-01 3.91380131e-01 4.43134636e-01 -1.18308544e+00
-4.65747327e-01 -4.06473160e-01 -3.33572865e-01 -1.26925588e+00
4.18648064e-01 -1.51821542e+00 -2.98996419e-01 -2.43055987e+00
2.53006160e-01 -1.87799688e-02 -5.08474231e-01 7.28872657e-01
-9.10531044e-01 -4.57038730e-01 2.45110780e-01 -1.83168322e-01
-8.45198631e-01 6.24705315e-01 1.55607927e+00 -2.52144694e-01
-1.03029564e-01 -5.98517776e-01 -1.22593713e+00 5.92776597e-01
6.68278456e-01 -1.19758233e-01 -8.47688317e-01 -6.10964656e-01
7.48877466e-01 -1.77574262e-01 5.10174036e-01 -6.37201607e-01
3.71304989e-01 5.11258692e-02 2.62970448e-01 -4.30501640e-01
-5.75061813e-02 -6.35155678e-01 -3.54030728e-01 -2.10484434e-02
-4.94448841e-01 1.74869858e-02 2.54348963e-02 6.84977472e-01
-2.87606299e-01 -2.50785500e-01 3.41652989e-01 -2.11128294e-01
-1.07900977e+00 1.81778222e-01 2.51324296e-01 5.13550937e-01
5.59005618e-01 1.00958914e-01 -8.12732518e-01 -2.76915193e-01
-8.41175497e-01 5.03812551e-01 1.02411292e-01 4.89293188e-01
6.53463304e-01 -1.33232927e+00 -4.67404902e-01 -9.17601436e-02
3.93957675e-01 2.74131000e-01 2.75566459e-01 6.42999053e-01
-1.54702693e-01 3.46603990e-01 6.52303174e-02 -1.20065987e-01
-9.66309905e-01 9.07619417e-01 3.41268033e-01 -7.67829776e-01
-6.25750780e-01 9.57557917e-01 2.38632992e-01 -4.22429621e-01
-1.85834572e-01 -7.51787066e-01 -8.73440206e-01 1.27643468e-02
5.07752717e-01 9.62001644e-03 1.99262142e-01 -5.11641026e-01
-2.58726180e-01 4.87474144e-01 -1.81256369e-01 2.92312384e-01
1.22668958e+00 -1.38603628e-01 -4.36698794e-01 2.56752074e-01
8.32573056e-01 7.41086006e-02 -4.52928066e-01 -5.20682454e-01
1.51763722e-01 -2.10439116e-01 1.36125475e-01 -1.03139782e+00
-9.91022527e-01 8.24303269e-01 -1.95066869e-01 2.67973155e-01
1.17854273e+00 3.56656313e-01 7.41444230e-01 6.28305793e-01
3.16304833e-01 -7.93681204e-01 4.94798198e-02 6.68831348e-01
1.18431282e+00 -9.13254678e-01 -1.40531212e-01 -6.95670247e-01
-6.45517290e-01 9.16447103e-01 9.56634998e-01 1.05167683e-02
4.95380133e-01 -3.72956455e-01 -2.37652898e-01 -7.74894476e-01
-7.20444918e-01 -7.02827394e-01 6.35723472e-01 5.66563845e-01
3.72375011e-01 1.08304389e-01 -4.06800449e-01 1.26252222e+00
-3.14338207e-01 7.29161650e-02 2.65805840e-01 7.72760332e-01
-4.12050754e-01 -8.77455354e-01 1.64308846e-01 6.33127928e-01
-6.57209978e-02 -6.71367407e-01 -6.32210851e-01 7.21808553e-01
1.16078295e-01 8.81888151e-01 -3.23446304e-01 -3.64404082e-01
6.59698069e-01 4.25468653e-01 6.28687382e-01 -8.75967979e-01
-6.79876089e-01 -5.70647657e-01 2.82878309e-01 -5.81214845e-01
-2.89219320e-01 -8.63150358e-02 -1.74251449e+00 -1.24485027e-02
-6.65507495e-01 2.13321060e-01 3.16407271e-02 1.16000056e+00
5.59022725e-01 9.04341757e-01 1.76004916e-02 -5.87316714e-02
-3.04105818e-01 -1.00369811e+00 -4.95096624e-01 6.07711673e-01
-2.15389095e-02 -7.46191919e-01 -1.64067611e-01 -4.75238562e-02] | [10.365232467651367, 7.942054748535156] |
2b0a95e9-eb20-4c94-83f2-405e14483036 | poissonian-blurred-image-deconvolution-by | 2206.05283 | null | https://arxiv.org/abs/2206.05283v1 | https://arxiv.org/pdf/2206.05283v1.pdf | Poissonian Blurred Image Deconvolution by Framelet based Local Minimal Prior | Image production tools do not always create a clear image, noisy and blurry images are sometimes created. Among these cases, Poissonian noise is one of the most famous noises that appear in medical images and images taken in astronomy. Blurred image with Poissonian noise obscures important details that are of great importance in medicine or astronomy. Therefore, studying and increasing the quality of images that are affected by this type of noise is always considered by researchers. In this paper, in the first step, based on framelet transform, a local minimal prior is introduced, and in the next step, this tool together with fractional calculation is used for Poissonian blurred image deconvolution. In the following, the model is generalized to the blind case. To evaluate the performance of the presented model, several images such as real images have been investigated. | ['Reza Parvaz'] | 2022-06-10 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 2.76711136e-01 -4.00013089e-01 4.67994362e-01 1.08192386e-02
2.73465961e-01 -1.93848997e-01 3.18795711e-01 -2.11397007e-01
-6.58044159e-01 1.27737367e+00 2.26407856e-01 8.28930289e-02
-3.15261006e-01 -5.61171949e-01 -2.83944219e-01 -9.39936221e-01
3.67178380e-01 4.82942276e-02 1.07827179e-01 2.05098093e-01
2.44503543e-01 4.73102033e-01 -1.49858642e+00 -1.34199157e-01
1.09125590e+00 6.68723583e-01 7.99893498e-01 3.83650690e-01
-5.64398050e-01 7.22913802e-01 -7.05944836e-01 -2.98686683e-01
1.38251439e-01 -7.30844557e-01 -4.77286607e-01 4.70296979e-01
-2.64871001e-01 -5.18773675e-01 -2.41897833e-02 1.47191763e+00
3.46598238e-01 1.52344048e-01 6.81194365e-01 -5.12343287e-01
-7.63925612e-01 3.45225722e-01 -6.15072787e-01 4.12851065e-01
1.04541130e-01 2.89054692e-01 2.31972951e-02 -7.48006880e-01
4.09305125e-01 1.38495922e+00 3.08845401e-01 2.06176296e-01
-1.13857079e+00 -3.33468884e-01 -3.87616277e-01 4.05806482e-01
-1.13742554e+00 -3.15567642e-01 8.40590775e-01 -3.74741793e-01
-1.42758563e-01 2.62554228e-01 6.02012575e-01 8.15527022e-01
5.03816724e-01 3.19293678e-01 1.65719843e+00 -3.90479535e-01
1.90380409e-01 1.97465733e-01 2.39076942e-01 1.56526715e-01
8.28458428e-01 1.33850956e-02 8.68386850e-02 1.15542546e-01
8.09452057e-01 1.04354791e-01 -8.04695904e-01 -5.77352494e-02
-9.13525343e-01 3.95599455e-01 2.53487617e-01 9.69966054e-01
-4.98596638e-01 -1.91439390e-01 -1.27186319e-02 -4.48016450e-02
5.09946465e-01 4.51457411e-01 2.12782577e-01 -3.07588279e-01
-9.20517504e-01 1.08180679e-01 6.81026280e-01 3.42312694e-01
6.21853173e-01 1.20729404e-02 -1.92173168e-01 8.38285327e-01
2.91188830e-03 6.55726731e-01 5.76419711e-01 -7.51213968e-01
-2.45334744e-01 2.55544692e-01 4.77858186e-01 -9.78627741e-01
-1.56175360e-01 -7.52580106e-01 -1.18544292e+00 4.52524245e-01
5.05918026e-01 1.07695423e-01 -8.92543018e-01 1.12764585e+00
2.78506041e-01 4.82765436e-01 1.80209994e-01 1.23270965e+00
7.57928014e-01 5.92470109e-01 -1.23046644e-01 -7.60912657e-01
1.44757581e+00 -4.88824159e-01 -1.28416836e+00 1.33463830e-01
-4.41537917e-01 -1.42587841e+00 7.10105419e-01 5.82722425e-01
-1.06986594e+00 -5.92557192e-01 -6.07837737e-01 -3.96200642e-02
-7.72621855e-02 2.26116460e-02 1.42019942e-01 6.05761230e-01
-7.26044774e-01 6.25920653e-01 -4.26790297e-01 -3.64336669e-01
2.32960910e-01 -2.77123123e-01 -1.20885424e-01 -5.13976574e-01
-9.57717776e-01 1.16102827e+00 2.73018777e-01 3.86045456e-01
-4.12687391e-01 -4.24540728e-01 -3.85580599e-01 4.35149707e-02
1.95084915e-01 -7.45120347e-01 9.83921230e-01 -9.05284941e-01
-1.24303555e+00 5.39861023e-01 -3.97861689e-01 -4.61605132e-01
7.32604980e-01 -2.59053349e-01 -2.94829398e-01 2.94868022e-01
-2.73704622e-02 -1.11285910e-01 1.22354496e+00 -1.49110818e+00
-2.21472710e-01 -2.51519501e-01 -1.41306922e-01 1.89706698e-01
3.24870884e-01 -1.99215058e-02 -2.25028619e-01 -7.15195656e-01
-7.17619201e-03 -4.28438246e-01 6.67172577e-03 -1.42060876e-01
-2.44601682e-01 2.11036026e-01 7.45087326e-01 -9.44148242e-01
1.05641890e+00 -2.38653731e+00 6.44180700e-02 6.25632852e-02
1.75126866e-01 4.84741151e-01 3.03682297e-01 7.56649971e-02
8.72889385e-02 -1.69919834e-01 -7.04851031e-01 -2.70351559e-01
-4.65883732e-01 3.66005808e-01 1.12654328e-01 4.87637490e-01
-5.98390475e-02 3.99938971e-01 -8.21860850e-01 -5.69973528e-01
5.12250960e-01 6.55534685e-01 -3.19263163e-05 1.81277111e-01
2.47931108e-02 1.05103385e+00 -2.63519704e-01 2.95844197e-01
1.23449063e+00 -3.12477704e-02 -3.28828335e-01 -4.42337304e-01
-3.93905789e-01 -6.42633021e-01 -1.14040923e+00 9.95274544e-01
-3.68969232e-01 4.00838643e-01 3.82961303e-01 -8.88839006e-01
8.33355188e-01 3.14331889e-01 3.81046832e-01 -6.23269737e-01
3.55213106e-01 3.79378349e-01 1.08650163e-01 -9.84517455e-01
3.93036246e-01 -3.48308027e-01 6.94726706e-01 1.17878705e-01
-4.72876638e-01 -2.96377420e-01 3.70651066e-01 -3.91607583e-01
7.36363411e-01 -1.53523564e-01 4.18662786e-01 -1.64083183e-01
8.95406365e-01 -2.54362822e-01 4.38550353e-01 5.08053720e-01
-1.15439571e-01 9.79570925e-01 2.46037096e-01 -2.88122177e-01
-1.16687465e+00 -8.64675641e-01 -4.29522216e-01 -1.25635907e-01
5.08609235e-01 4.71116036e-01 -8.26367855e-01 3.51057686e-02
-1.37113571e-01 7.21619368e-01 -3.53235126e-01 7.96932075e-03
-2.82638967e-01 -8.79731476e-01 1.20466314e-01 -3.18007052e-01
1.15661061e+00 -1.14751625e+00 -7.07332313e-01 2.21064314e-01
-4.03512478e-01 -1.25730097e+00 -3.29034954e-01 -4.29738849e-01
-8.48984778e-01 -1.30728531e+00 -1.53796995e+00 -5.68079948e-01
8.29841912e-01 6.03162289e-01 7.38044500e-01 8.28041360e-02
-5.11684835e-01 2.06573054e-01 -4.80623633e-01 -2.43399069e-01
-7.26061761e-01 -5.88712752e-01 -2.11158201e-01 6.32022440e-01
1.10262923e-01 -5.30425370e-01 -6.81004345e-01 2.31538936e-01
-1.27215457e+00 2.20714193e-02 1.03724062e+00 7.85431802e-01
2.14099407e-01 4.88296419e-01 2.04916418e-01 -7.39321887e-01
8.78754199e-01 -1.38222784e-01 -7.17882276e-01 8.69875774e-02
-3.17311764e-01 6.45424426e-02 6.99671388e-01 -3.90980601e-01
-1.53559780e+00 -3.66150528e-01 7.62537047e-02 -4.77376640e-01
-4.80397016e-01 4.06308711e-01 6.29364103e-02 -8.12115222e-02
5.95760763e-01 5.18902898e-01 4.56853807e-01 -8.78413737e-01
1.21577084e-02 6.94574594e-01 5.64757943e-01 3.97828221e-02
8.63906980e-01 5.23598015e-01 1.47530586e-01 -1.30793226e+00
-5.07864118e-01 -5.42170346e-01 -1.13584116e-01 -2.00566098e-01
1.06782126e+00 -3.80006313e-01 -6.17806137e-01 8.35012555e-01
-1.36131501e+00 2.06714064e-01 -2.33356223e-01 8.51357996e-01
-1.93305582e-01 7.92264760e-01 -5.87089777e-01 -1.18898857e+00
-2.07086921e-01 -1.32072973e+00 4.09225732e-01 7.85097837e-01
1.28719077e-01 -7.71910548e-01 -2.41776273e-01 2.94244409e-01
8.34300697e-01 5.40602803e-02 6.74426317e-01 1.34894932e-02
-6.79062188e-01 2.03367427e-01 -4.88691628e-01 6.95353627e-01
6.31733835e-01 -3.95450830e-01 -9.42860961e-01 1.15896784e-01
9.22443330e-01 5.13556778e-01 6.93160355e-01 7.76811719e-01
9.34613824e-01 -2.92305619e-01 7.54090538e-03 4.57505465e-01
1.51338780e+00 5.37482917e-01 1.03441429e+00 1.75328091e-01
3.56328368e-01 4.36186820e-01 5.76221228e-01 2.27735236e-01
-3.64741117e-01 4.82715815e-01 3.75861853e-01 -2.53131151e-01
-5.20424187e-01 3.46476376e-01 -5.57721630e-02 5.83870232e-01
-4.51254487e-01 -5.36095910e-02 -2.92285383e-01 4.99297827e-01
-1.42040646e+00 -1.11516416e+00 -4.38663036e-01 2.28764701e+00
8.21830511e-01 -1.05004199e-01 -3.68658870e-01 3.00505787e-01
1.11027551e+00 -2.57766824e-02 -2.00274006e-01 -8.44079405e-02
-4.32797104e-01 3.08156759e-01 4.67233062e-01 8.43247950e-01
-8.58251214e-01 4.24240857e-01 5.19855881e+00 8.29522610e-01
-1.20356250e+00 2.16601133e-01 4.70527709e-01 3.96382391e-01
-1.99843168e-01 2.34695654e-02 -1.97925687e-01 1.06829381e+00
2.64096498e-01 -2.46866062e-01 7.16538429e-01 3.46628278e-01
7.80186176e-01 -9.46951687e-01 -1.41538978e-01 1.44350612e+00
2.76393685e-02 -7.16609776e-01 -4.05417755e-02 8.42616186e-02
6.43624008e-01 -6.49944544e-01 6.44582063e-02 -4.63466257e-01
-2.36407116e-01 -1.04542220e+00 1.66580707e-01 1.11609614e+00
4.59825486e-01 -6.21128678e-01 1.33802736e+00 2.68943697e-01
-4.27357286e-01 1.06351428e-01 -5.40394008e-01 -1.65272243e-02
4.64515001e-01 1.44314003e+00 -6.50141478e-01 7.74044812e-01
5.32178164e-01 5.20502090e-01 -4.56704736e-01 1.84436691e+00
-2.23954543e-01 4.39858913e-01 -1.48661107e-01 -3.23073082e-02
-5.11006117e-02 -8.35062981e-01 1.03276086e+00 8.32013726e-01
6.86334491e-01 2.95388401e-01 -4.93901432e-01 1.25737333e+00
2.08810449e-01 3.83054465e-02 -7.07414985e-01 2.38738835e-01
8.24989527e-02 1.10539770e+00 -7.21505225e-01 -2.93779343e-01
-1.80594772e-01 1.14921236e+00 -6.59288824e-01 6.47489548e-01
-6.94832921e-01 -1.50886014e-01 3.70730132e-01 1.28636047e-01
2.01495830e-02 -1.11694083e-01 -2.10713506e-01 -1.03920162e+00
-1.69309497e-01 -6.69311166e-01 -2.91266888e-01 -9.63838518e-01
-1.16393733e+00 4.40302223e-01 1.03524372e-01 -1.26671326e+00
1.76606894e-01 -4.22501236e-01 -4.62108314e-01 1.29628992e+00
-1.51040125e+00 -5.45558512e-01 -6.34123743e-01 3.93101305e-01
6.35072768e-01 9.04745534e-02 3.97281855e-01 4.55790401e-01
-3.43126118e-01 -2.18544781e-01 4.14954096e-01 -2.13300854e-01
6.78189933e-01 -9.80921030e-01 -1.94022909e-01 1.04509771e+00
-2.57860035e-01 5.12992263e-01 1.27967894e+00 -7.03513145e-01
-7.06332266e-01 -6.17484927e-01 6.48642242e-01 2.58068234e-01
1.56455740e-01 2.94006616e-01 -1.05835783e+00 4.79910262e-02
3.86546433e-01 -7.50900060e-02 -8.09214860e-02 -5.96997082e-01
4.53217298e-01 -1.92271382e-01 -1.40772367e+00 4.77159977e-01
4.21416700e-01 -1.31753117e-01 -7.93139994e-01 1.10255964e-01
3.61330628e-01 -3.93518418e-01 -5.70616424e-01 1.89877957e-01
1.42809853e-01 -1.32427347e+00 8.22935522e-01 2.25736424e-01
2.88777888e-01 -6.59015715e-01 4.01292622e-01 -1.52629960e+00
-1.67492434e-01 -5.71780980e-01 2.99705178e-01 1.14820492e+00
-8.79182443e-02 -8.23628664e-01 3.21402133e-01 9.51105580e-02
5.49615361e-03 -7.90454522e-02 -7.60798931e-01 -6.89080477e-01
-5.11649609e-01 6.14575148e-02 3.00545067e-01 7.47428119e-01
-5.34299195e-01 1.04912333e-01 -5.50062537e-01 6.95940107e-02
1.08210194e+00 -1.06028251e-01 5.98597288e-01 -1.22937322e+00
-2.81989872e-01 -3.51303428e-01 -3.55907977e-01 -7.69275486e-01
-5.58310509e-01 -1.94406003e-01 7.23013729e-02 -1.67512488e+00
1.75833553e-01 -2.04404637e-01 9.13078114e-02 -3.76730442e-01
-3.29835624e-01 2.51973331e-01 2.25794077e-01 4.17600989e-01
1.12439252e-01 3.82023245e-01 1.72583485e+00 -8.17910284e-02
3.39520574e-02 2.75994480e-01 -3.89081806e-01 6.81159496e-01
6.54987037e-01 -1.83602661e-01 -2.29218125e-01 -1.71636939e-01
-3.18822533e-01 -3.35279405e-02 6.07066691e-01 -1.16999710e+00
1.22795470e-01 -1.35462610e-02 3.53797972e-01 -3.71424466e-01
4.21055049e-01 -1.11819386e+00 4.76725131e-01 5.75515866e-01
2.47301430e-01 -3.45784038e-01 -8.21691453e-02 5.15471876e-01
-4.71919984e-01 -8.17863047e-01 1.15818667e+00 -4.98584539e-01
-4.39318001e-01 -1.61376446e-01 -2.86447614e-01 -2.91936815e-01
1.00152218e+00 -3.69895637e-01 -3.39690208e-01 -5.72460175e-01
-7.43369937e-01 -2.65848547e-01 4.65802848e-01 -1.47811836e-02
6.38679445e-01 -7.37028360e-01 -5.75758219e-01 -7.39170313e-02
-3.24369043e-01 6.93347976e-02 7.11271822e-01 1.26158607e+00
-1.10682547e+00 2.30702952e-01 -2.59813607e-01 -4.50550199e-01
-1.31593585e+00 8.20183396e-01 3.02322805e-01 -1.09422747e-02
-5.64024627e-01 2.73708224e-01 2.86365628e-01 4.29287612e-01
-5.17392755e-02 -5.09658277e-01 -5.45449972e-01 -1.69561565e-01
6.90891087e-01 5.98457634e-01 -2.05069393e-01 -7.14469135e-01
9.49769616e-02 8.13744724e-01 2.29674742e-01 -1.59298941e-01
9.75865364e-01 -4.70837861e-01 -5.12285590e-01 5.78629315e-01
7.85955667e-01 4.17681724e-01 -1.09504092e+00 -2.65638120e-02
-1.49938226e-01 -7.32879043e-01 9.26681794e-03 -6.43402040e-01
-9.37913239e-01 7.83859134e-01 7.08358228e-01 4.58827645e-01
1.31210542e+00 -3.33818406e-01 6.94401443e-01 -3.09370011e-01
2.49945536e-01 -7.34393895e-01 -1.03721254e-01 2.37159610e-01
8.80242825e-01 -9.73978400e-01 -5.42398542e-03 -6.99266076e-01
-3.87548029e-01 1.11860871e+00 1.27392232e-01 -3.96833085e-02
4.73816752e-01 1.83644846e-01 8.27762708e-02 1.86233610e-01
1.93061650e-01 -5.72606385e-01 -4.17177230e-02 5.23539841e-01
3.14575970e-01 -2.45411307e-01 -1.11814868e+00 2.94901520e-01
1.08306699e-01 5.30793667e-01 8.96945417e-01 6.36536181e-01
-6.49970114e-01 -8.43114972e-01 -1.31992769e+00 2.55355418e-01
-8.79310310e-01 -6.33987831e-04 1.98821932e-01 4.89777148e-01
3.68850678e-01 1.12337291e+00 -6.36742041e-02 2.67391801e-01
1.85897619e-01 -9.46442410e-02 4.56009179e-01 -2.08394900e-01
-5.39816320e-02 2.11876169e-01 -2.42656186e-01 -8.64843354e-02
-7.21419036e-01 -3.77701312e-01 -9.84149218e-01 -3.43140781e-01
-2.56576091e-01 3.63933414e-01 7.51013041e-01 9.46098566e-01
-9.69166309e-02 6.54410541e-01 3.42990249e-01 -6.44529700e-01
-3.22951972e-01 -1.23491311e+00 -9.17071700e-01 7.46804476e-01
5.03971159e-01 -7.51891911e-01 -7.20144808e-01 1.47166342e-01] | [11.626350402832031, -2.665297031402588] |
e2a8b6d4-a536-4af6-b57c-ed463e2ebb19 | free-hyperbolic-neural-networks-with-limited | 2107.11472 | null | https://arxiv.org/abs/2107.11472v4 | https://arxiv.org/pdf/2107.11472v4.pdf | Clipped Hyperbolic Classifiers Are Super-Hyperbolic Classifiers | Hyperbolic space can naturally embed hierarchies, unlike Euclidean space. Hyperbolic Neural Networks (HNNs) exploit such representational power by lifting Euclidean features into hyperbolic space for classification, outperforming Euclidean neural networks (ENNs) on datasets with known semantic hierarchies. However, HNNs underperform ENNs on standard benchmarks without clear hierarchies, greatly restricting HNNs' applicability in practice. Our key insight is that HNNs' poorer general classification performance results from vanishing gradients during backpropagation, caused by their hybrid architecture connecting Euclidean features to a hyperbolic classifier. We propose an effective solution by simply clipping the Euclidean feature magnitude while training HNNs. Our experiments demonstrate that clipped HNNs become super-hyperbolic classifiers: They are not only consistently better than HNNs which already outperform ENNs on hierarchical data, but also on-par with ENNs on MNIST, CIFAR10, CIFAR100 and ImageNet benchmarks, with better adversarial robustness and out-of-distribution detection. | ['Stella X. Yu', 'Yubei Chen', 'Xudong Wang', 'Yunhui Guo'] | 2021-07-23 | free-hyperbolic-neural-networks-with-limited-1 | http://openaccess.thecvf.com//content/CVPR2022/html/Guo_Clipped_Hyperbolic_Classifiers_Are_Super-Hyperbolic_Classifiers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Guo_Clipped_Hyperbolic_Classifiers_Are_Super-Hyperbolic_Classifiers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['classification'] | ['methodology'] | [-2.32076988e-01 6.49622679e-01 9.84986499e-02 -3.23026836e-01
-5.68156183e-01 -6.66882336e-01 5.79992175e-01 -1.18854769e-01
-7.33771682e-01 5.83525836e-01 3.94060999e-01 -2.24797040e-01
-4.57879454e-01 -9.70725954e-01 -7.21971095e-01 -7.07692623e-01
-4.18760866e-01 1.22739412e-01 4.61814225e-01 -5.21028876e-01
1.92765549e-01 5.03544033e-01 -1.15689051e+00 3.55236024e-01
6.38775170e-01 1.13237083e+00 -5.94986677e-01 7.71800995e-01
5.43949246e-01 1.18663883e+00 -4.07225400e-01 -6.46535516e-01
3.67310286e-01 -3.05761784e-01 -8.21030676e-01 -4.71858889e-01
6.75082147e-01 -4.66960430e-01 -1.12439179e+00 1.05444229e+00
3.76309037e-01 3.09157878e-01 7.42213964e-01 -1.59267235e+00
-1.55047810e+00 1.23550832e+00 -1.48427069e-01 1.59257680e-01
-1.89203352e-01 -2.05334321e-01 1.46711171e+00 -1.07617831e+00
3.68356556e-01 1.18278122e+00 1.54493022e+00 5.28063953e-01
-1.33067560e+00 -6.36658788e-01 -4.27057408e-02 4.37239736e-01
-1.63285112e+00 -4.85980362e-02 6.25659585e-01 -8.65922198e-02
8.64212811e-01 2.85281718e-01 6.06334686e-01 1.34210587e+00
9.86724198e-02 8.12314987e-01 8.69975448e-01 9.05817747e-02
3.60801190e-01 8.28530118e-02 3.51410866e-01 7.37718582e-01
5.46059720e-02 9.20763910e-02 -3.06395888e-01 -9.40186307e-02
7.65708506e-01 9.28849056e-02 -4.21404362e-01 -4.15377259e-01
-9.65260029e-01 1.28688669e+00 1.51803982e+00 3.31640065e-01
-8.52602124e-02 4.58894968e-01 5.56203723e-01 3.54847789e-01
1.26462817e-01 9.15227115e-01 -3.15130651e-01 1.67760998e-01
-5.77839851e-01 3.60416979e-01 5.86747408e-01 1.10955667e+00
4.70555127e-01 7.60457572e-03 1.89196616e-01 8.13581109e-01
-9.46717113e-02 1.82329983e-01 6.51148140e-01 -1.20776916e+00
1.66806623e-01 5.39309680e-01 -4.15636420e-01 -1.17201877e+00
-9.84858096e-01 -8.38050485e-01 -1.18512547e+00 1.15647517e-01
4.95929629e-01 1.18075125e-01 -5.68834484e-01 1.84762967e+00
-1.99214429e-01 -7.88883045e-02 4.95497495e-01 8.44501376e-01
9.21082020e-01 6.95945621e-01 -2.49767855e-01 5.30710936e-01
1.23701000e+00 -1.03363872e+00 -2.38692686e-01 1.51044711e-01
1.11468518e+00 -7.95306917e-03 1.15625358e+00 4.47371960e-01
-9.38398302e-01 -4.01736796e-01 -1.43750751e+00 -5.61480880e-01
-8.85471046e-01 -3.53292465e-01 7.46783137e-01 5.52863419e-01
-1.23911047e+00 1.23788404e+00 -5.56268573e-01 -2.04147384e-01
6.81901217e-01 4.96320665e-01 -4.49191451e-01 -5.64058602e-04
-1.60345125e+00 7.55272031e-01 6.28716290e-01 -2.39307452e-02
-5.81583679e-01 -1.19623744e+00 -1.08554566e+00 1.59158185e-01
-2.45427415e-01 -3.10975790e-01 1.24284983e+00 -4.56075847e-01
-1.22903728e+00 5.28359532e-01 9.03072894e-01 -1.09499276e+00
7.49697030e-01 -1.63995832e-01 -2.45612204e-01 2.39096150e-01
-2.03326136e-01 1.35248613e+00 4.61253405e-01 -1.12339616e+00
-6.31476045e-01 -3.83443356e-01 3.44252050e-01 1.17731594e-01
-8.12886238e-01 -5.59926987e-01 3.47615242e-01 -5.37541628e-01
4.99709696e-01 -1.25015700e+00 -2.01717597e-02 7.23411143e-02
-8.32396269e-01 -5.57164431e-01 1.12274039e+00 -3.75621080e-01
9.47634637e-01 -2.28098297e+00 9.51207057e-02 -1.82053205e-02
6.53292775e-01 -5.05134091e-02 -9.54762101e-02 2.71551013e-01
3.57095972e-02 5.71878441e-02 -2.99732983e-01 -2.38318086e-01
6.80914283e-01 3.38852525e-01 -6.17787719e-01 8.86132956e-01
4.61697802e-02 8.17778051e-01 -1.00638962e+00 -2.19520330e-01
6.41445443e-03 8.10681760e-01 -9.54950571e-01 -3.09168160e-01
2.42174566e-01 -2.26018041e-01 2.00008258e-01 2.35722691e-01
6.94674373e-01 -1.80811793e-01 -3.08326930e-01 -1.70373246e-01
6.51620096e-03 1.17016487e-01 -8.07290256e-01 1.29184544e+00
-3.93123895e-01 1.18147159e+00 -4.86187547e-01 -1.08837938e+00
8.71293426e-01 8.21098909e-02 2.13447928e-01 -8.18725169e-01
1.78657584e-02 4.19070512e-01 2.12011755e-01 8.97914767e-02
5.95162630e-01 5.57438247e-02 -2.66423583e-01 -4.41475306e-03
2.20685512e-01 7.94772655e-02 -1.02745458e-01 3.07801068e-01
1.26407075e+00 -3.24940771e-01 -2.49103844e-01 -5.33382297e-01
4.05428171e-01 -1.50414407e-01 4.04346257e-01 9.95093703e-01
-3.95864993e-01 8.66730213e-01 7.88957596e-01 -5.53484499e-01
-1.43542862e+00 -1.50500953e+00 -6.79144382e-01 1.50048864e+00
1.88003302e-01 -3.15666676e-01 -8.20971072e-01 -9.34726179e-01
1.14825480e-01 9.35676217e-01 -1.08036327e+00 -7.09958136e-01
-6.18772209e-01 -8.10120821e-01 1.24618244e+00 9.96590257e-01
8.98506522e-01 -8.23347211e-01 -5.08763969e-01 -5.39917126e-02
3.30555618e-01 -1.19364262e+00 -3.62939537e-01 6.35251582e-01
-6.97325230e-01 -9.00822699e-01 -8.72418523e-01 -7.53794014e-01
3.60740781e-01 -2.67006084e-02 8.67927730e-01 -2.30300859e-01
-2.54205644e-01 1.46409899e-01 -3.19889009e-01 8.50431062e-03
-3.62665892e-01 6.72011912e-01 2.14056119e-01 -5.18019557e-01
1.45004019e-01 -6.80259109e-01 -8.89051080e-01 5.51809430e-01
-8.39831233e-01 -9.47156250e-02 4.33555454e-01 1.02709901e+00
-8.89293775e-02 4.05127585e-01 8.13601673e-01 -8.59313011e-01
3.09159577e-01 -5.39600790e-01 -4.02968347e-01 -2.28371173e-01
-5.99704742e-01 1.58034749e-02 1.30739236e+00 -3.66051763e-01
-6.01192355e-01 -3.63661736e-01 -1.90435991e-01 -4.73955840e-01
1.72626451e-02 -2.06687152e-02 2.39005059e-01 -1.81778744e-01
1.25149763e+00 4.76739593e-02 -3.23326379e-01 -1.36798620e-01
6.04318917e-01 3.36515337e-01 8.67648065e-01 -2.50909001e-01
8.60590756e-01 5.71342170e-01 -9.60171074e-02 -8.72071087e-01
-1.15218723e+00 -2.42413327e-01 -5.92275977e-01 3.81401658e-01
1.08297741e+00 -6.65333927e-01 -6.85898602e-01 2.77312696e-01
-1.11032093e+00 -3.85654777e-01 -5.07080138e-01 4.66648430e-01
-5.57286322e-01 -2.13257343e-01 -1.27007675e+00 -3.49397659e-01
-6.44583479e-02 -8.78444672e-01 9.45112467e-01 -2.02632863e-02
-1.63506314e-01 -1.45489097e+00 -9.60948169e-02 1.36629954e-01
4.12464559e-01 1.61626533e-01 1.21783841e+00 -9.84294295e-01
-2.36535102e-01 -2.20039994e-01 -6.84511483e-01 4.52661991e-01
-2.55166590e-01 -1.32958695e-01 -1.33464170e+00 -3.30270082e-01
-1.64657086e-01 -5.74018002e-01 1.03598416e+00 2.24616006e-01
1.14225709e+00 -7.20677137e-01 1.72429048e-02 1.13485897e+00
1.48289919e+00 -1.70580775e-01 6.39267504e-01 7.17821956e-01
8.93548250e-01 5.51249683e-01 -1.00721762e-01 1.72646880e-01
4.55208838e-01 5.16719162e-01 5.79639852e-01 -3.28911245e-01
-1.26966342e-01 -2.40030706e-01 2.27660656e-01 6.01266742e-01
1.46180987e-01 1.63238019e-01 -1.04342258e+00 1.04637541e-01
-1.78526115e+00 -9.32253480e-01 -1.81805804e-01 1.77244210e+00
6.77165866e-01 3.78064632e-01 3.36273730e-01 5.84328532e-01
4.90679741e-01 1.03162602e-01 -6.46909416e-01 -3.83574724e-01
-3.71313989e-01 -1.89595580e-01 9.59138155e-01 3.75238508e-01
-1.47206199e+00 8.02691519e-01 6.09811211e+00 5.63271105e-01
-8.49045515e-01 6.12431802e-02 5.37711740e-01 6.47581816e-02
-7.23980889e-02 -4.60989296e-01 -8.41254890e-01 6.47665039e-02
1.10200512e+00 -1.02632768e-01 3.21336955e-01 1.27366078e+00
-6.02573752e-01 7.74586141e-01 -1.50869966e+00 9.63882685e-01
-2.76751593e-02 -1.49148917e+00 -7.32087642e-02 2.11147949e-01
8.86256695e-01 4.20554370e-01 7.69786596e-01 9.35252190e-01
6.36143923e-01 -1.40746963e+00 9.40504551e-01 1.79317258e-02
4.21746790e-01 -1.00742459e+00 9.34283912e-01 1.57749265e-01
-9.24237311e-01 -6.54565990e-01 -1.00041127e+00 1.46389529e-01
-5.41768372e-01 1.22315489e-01 -8.01365316e-01 -4.45620203e-03
8.31793129e-01 7.36503422e-01 -9.45451796e-01 8.07870388e-01
-2.10877005e-02 2.59656340e-01 -2.38091439e-01 5.37968464e-02
1.03319943e+00 2.43886169e-02 1.43393174e-01 1.38685298e+00
1.21933244e-01 1.58118531e-01 -7.52163455e-02 9.47959840e-01
-4.15799171e-01 5.65508986e-03 -8.44969690e-01 1.05999596e-01
4.50089782e-01 1.16205549e+00 -8.92738223e-01 -1.29994929e-01
-3.10599059e-01 9.67486441e-01 6.39204741e-01 3.95552725e-01
-1.37050581e+00 -1.00984681e+00 4.28526461e-01 2.81227157e-02
2.47834072e-01 -1.74020126e-01 -2.28777289e-01 -7.44901896e-01
-3.13682556e-01 -5.58531761e-01 5.74061751e-01 -6.56014144e-01
-1.19223607e+00 9.22933280e-01 -2.18649402e-01 -1.04069459e+00
-1.69837172e-03 -1.12808788e+00 -2.91270792e-01 -6.04779273e-02
-1.13448954e+00 -1.01413369e+00 -3.41973186e-01 6.34932935e-01
2.31543556e-01 -1.05395608e-01 7.30008185e-01 3.44519973e-01
-5.32226324e-01 1.11214328e+00 4.85581160e-01 7.63126254e-01
2.44722903e-01 -1.83792603e+00 3.53710145e-01 2.98577935e-01
2.59413213e-01 5.03185511e-01 6.67818725e-01 1.98103234e-01
-1.05896556e+00 -1.34785080e+00 4.02821988e-01 -5.17304301e-01
1.11665428e+00 -7.45541871e-01 -1.09147179e+00 8.54470432e-01
-3.48718688e-02 3.50287795e-01 6.36936069e-01 2.19067857e-01
-1.16341054e+00 -9.46223140e-02 -1.09149432e+00 8.15257370e-01
1.53724849e+00 -7.70596027e-01 -6.84838235e-01 7.91150510e-01
1.21243000e+00 -2.55827188e-01 -1.13453770e+00 3.82266819e-01
5.15139639e-01 -1.38645720e+00 9.80493724e-01 -7.84807622e-01
7.10245013e-01 1.75958917e-01 -7.78307498e-01 -1.38684082e+00
-4.35473979e-01 -4.76090372e-01 7.15918913e-02 7.57938504e-01
5.01165330e-01 -8.51261735e-01 1.00170112e+00 1.84356064e-01
-5.16931951e-01 -8.39942992e-01 -9.08225954e-01 -1.33739066e+00
7.60893822e-01 -2.41754889e-01 3.00354958e-01 1.15865958e+00
1.78306371e-01 3.48931104e-01 1.12425104e-01 2.27764547e-01
8.39119613e-01 -4.04417992e-01 2.09155455e-01 -1.17799580e+00
-1.09998016e-02 -9.15539801e-01 -1.16148341e+00 -7.29061007e-01
6.01687491e-01 -1.21167040e+00 1.43807726e-02 -1.12451792e+00
8.81448686e-02 -5.97978175e-01 -5.01848221e-01 4.63404715e-01
3.79046440e-01 9.35785055e-01 6.31463677e-02 4.81631272e-02
-8.66705239e-01 7.90467739e-01 9.32745814e-01 -1.30314246e-01
-1.60539344e-01 -2.40917072e-01 -7.66081691e-01 1.02049422e+00
6.53797746e-01 -3.73899728e-01 -7.19968379e-01 -4.63389874e-01
1.75882041e-01 -4.48562205e-01 6.99488819e-01 -1.54902041e+00
4.72057134e-01 5.80858946e-01 6.80358589e-01 -2.04521298e-01
4.58611190e-01 -6.77761137e-01 -2.78553694e-01 6.82083189e-01
-7.66175866e-01 2.16421485e-01 -2.41869837e-02 3.82005125e-01
-8.48661736e-02 -3.83010991e-02 1.17681909e+00 2.82162517e-01
-4.14939523e-01 3.35815758e-01 -1.03988767e-01 4.29351032e-01
1.00964260e+00 -3.28749627e-01 -8.62853706e-01 -4.00271684e-01
-8.70344520e-01 6.33741915e-03 3.95570874e-01 3.87002647e-01
5.93093932e-01 -1.59870923e+00 -5.97755611e-01 1.43881798e-01
2.27011531e-03 8.78911745e-03 1.92892343e-01 8.75205457e-01
-7.99535692e-01 6.73749328e-01 -1.20448537e-01 -6.96397543e-01
-7.16706395e-01 6.36536479e-01 7.24615037e-01 2.13192284e-01
-1.14074659e+00 1.22530317e+00 6.99816763e-01 -8.62790525e-01
7.03512192e-01 -4.45540994e-01 3.43277939e-02 -6.31024614e-02
2.48477563e-01 6.65235639e-01 5.81902228e-02 -4.21158940e-01
-2.87615806e-01 2.70614147e-01 -2.70531863e-01 -1.07447229e-01
1.36100733e+00 2.48372793e-01 1.22903392e-01 4.76220220e-01
1.90295279e+00 -4.66299742e-01 -1.41090131e+00 -2.32674971e-01
3.01849455e-01 1.68801080e-02 2.55797915e-02 -2.56538033e-01
-9.31033254e-01 9.31962371e-01 5.31948984e-01 6.89508319e-01
7.31002092e-01 2.23408431e-01 8.80157173e-01 9.66912746e-01
5.88072762e-02 -1.07180870e+00 3.32501441e-01 7.27918565e-01
1.12992394e+00 -9.58625257e-01 -1.96941048e-01 -1.30671144e-01
-4.18348730e-01 1.22771811e+00 6.45888090e-01 -6.36617064e-01
6.81083262e-01 6.10269839e-03 -8.51263553e-02 -8.20993260e-02
-6.93490386e-01 2.81487107e-01 1.19667709e-01 5.59608638e-01
2.49003470e-01 -1.89742103e-01 4.74580497e-01 6.10475123e-01
-1.14851797e+00 -7.32251823e-01 5.28987706e-01 3.83208752e-01
-5.45827687e-01 -1.58123914e-02 -1.41926289e-01 5.64134754e-02
-1.61506996e-01 -3.13772798e-01 -3.69882941e-01 1.36665821e+00
6.67727739e-02 5.60974061e-01 3.92322272e-01 -5.55129826e-01
5.50124586e-01 -1.73255235e-01 4.12216157e-01 -1.88831851e-01
-3.16664636e-01 -6.09502256e-01 -1.07588075e-01 -6.51469827e-01
5.45492113e-01 -4.33518350e-01 -1.67591798e+00 -6.89310551e-01
-1.75738871e-01 4.18085486e-01 5.90164840e-01 6.55101895e-01
-7.80851915e-02 5.70142686e-01 7.76900589e-01 -7.17721581e-01
-1.37710404e+00 -8.50729764e-01 -9.52766299e-01 5.69118679e-01
6.28374875e-01 -5.55986702e-01 -8.76577675e-01 -3.85476798e-01] | [9.185223579406738, 3.1686761379241943] |
e8e8fe29-368d-4446-ab48-7d8ba4c5112a | healthedge-a-machine-learning-based-smart | 2301.10450 | null | https://arxiv.org/abs/2301.10450v1 | https://arxiv.org/pdf/2301.10450v1.pdf | HealthEdge: A Machine Learning-Based Smart Healthcare Framework for Prediction of Type 2 Diabetes in an Integrated IoT, Edge, and Cloud Computing System | Diabetes Mellitus has no permanent cure to date and is one of the leading causes of death globally. The alarming increase in diabetes calls for the need to take precautionary measures to avoid/predict the occurrence of diabetes. This paper proposes HealthEdge, a machine learning-based smart healthcare framework for type 2 diabetes prediction in an integrated IoT-edge-cloud computing system. Numerical experiments and comparative analysis were carried out between the two most used machine learning algorithms in the literature, Random Forest (RF) and Logistic Regression (LR), using two real-life diabetes datasets. The results show that RF predicts diabetes with 6% more accuracy on average compared to LR. | ['Leila Ismail', 'Huned Materwala', 'Alain Hennebelle'] | 2023-01-25 | null | null | null | null | ['diabetes-prediction'] | ['medical'] | [ 7.09629655e-02 -1.38333887e-01 -5.70166647e-01 -6.27099276e-01
-1.20173924e-01 2.66349524e-01 1.31611332e-01 7.38601089e-01
-8.31808820e-02 8.89871597e-01 1.23749092e-01 -5.43366075e-01
-3.82592618e-01 -9.66436803e-01 -2.88371816e-02 -7.19682515e-01
-1.49388149e-01 8.86630595e-01 -4.05172884e-01 2.89891511e-01
8.06877762e-02 2.94641256e-01 -1.35365891e+00 3.00396293e-01
1.11794674e+00 1.41152775e+00 -3.08804482e-01 4.92096066e-01
-1.69569794e-02 8.55553985e-01 -1.58428997e-01 -3.47313911e-01
2.99498647e-01 -4.78917360e-01 -2.10273802e-01 -3.04446101e-01
-3.26848507e-01 -2.55026463e-02 3.11867326e-01 6.33397520e-01
9.16707814e-01 -6.41657054e-01 5.40032446e-01 -1.36717319e+00
-3.09200972e-01 1.60593778e-01 -2.77290314e-01 2.37811640e-01
1.99381933e-01 -1.42128691e-01 1.05871230e-01 -2.81449229e-01
4.30480778e-01 8.83754015e-01 8.87296736e-01 4.76718605e-01
-9.61794734e-01 -7.68928587e-01 -3.69606256e-01 6.95158780e-01
-1.34616077e+00 -1.53735951e-01 3.73390883e-01 -3.97552311e-01
8.63214612e-01 4.49740261e-01 1.07433522e+00 7.56675601e-01
8.41608763e-01 2.34968305e-01 1.52309108e+00 -6.56430483e-01
5.44085264e-01 1.80298403e-01 1.06655648e-02 4.10749257e-01
7.72090912e-01 3.31887424e-01 -1.12536348e-01 -3.37217599e-01
5.13030112e-01 4.53299105e-01 2.53862023e-01 -1.97365507e-01
-9.09810901e-01 6.35284424e-01 1.92051560e-01 1.20888807e-01
-7.20020354e-01 -3.03780466e-01 3.81773531e-01 2.96299160e-01
6.40280902e-01 -1.66166842e-01 -1.00806069e+00 -1.29767686e-01
-4.08050299e-01 1.74787305e-02 8.38802636e-01 5.15430570e-01
9.23623443e-02 -1.57726660e-01 1.17427617e-01 3.93930376e-01
4.40774918e-01 4.12106246e-01 4.57752198e-01 -5.45722306e-01
-3.99151258e-02 1.03866577e+00 -8.70132297e-02 -9.12415564e-01
-5.78952670e-01 -7.41037488e-01 -1.31081212e+00 1.03227817e-01
1.99210659e-01 -3.70105892e-01 -6.20454073e-01 7.22232163e-01
4.95084107e-01 4.19055551e-01 3.60190451e-01 7.25523293e-01
6.31215453e-01 2.27300495e-01 5.27759731e-01 -5.93698978e-01
1.26654828e+00 -5.10148048e-01 -7.21092820e-01 2.75386244e-01
6.46454990e-01 -5.81396878e-01 3.79372150e-01 6.89952016e-01
-5.86201966e-01 -4.64476198e-02 -5.37771881e-01 3.87684107e-01
-3.76831949e-01 -6.06181212e-02 8.52097750e-01 9.18829143e-01
-5.13551831e-01 1.89064816e-01 -8.30399156e-01 -6.80388808e-01
4.83981520e-01 2.68582970e-01 1.07176034e-02 -5.57697594e-01
-8.69107902e-01 1.03108144e+00 -5.50928414e-02 -2.33110458e-01
-2.37880796e-01 -8.43257844e-01 -1.96049944e-01 -2.40273654e-01
-7.84279481e-02 -1.21716654e+00 7.92182565e-01 -8.27952385e-01
-1.09004688e+00 7.51002669e-01 -3.07244509e-02 -7.17831612e-01
7.25688040e-01 -1.87501386e-01 -9.34491456e-01 -1.80570617e-01
-1.69627890e-01 -6.80857897e-03 1.64464444e-01 -7.48843193e-01
-8.73634815e-01 -9.13329124e-01 -6.02336466e-01 -1.28576547e-01
1.67108536e-01 1.04106806e-01 4.33352172e-01 -5.47438085e-01
1.59303285e-02 -8.34058821e-01 -6.09531820e-01 -1.44823596e-01
-1.53404072e-01 -2.50200808e-01 5.87039411e-01 -8.05701017e-01
1.18497205e+00 -1.43457460e+00 -2.87461460e-01 1.82488203e-01
-4.04423662e-02 6.62067160e-02 4.23541933e-01 2.78484672e-01
-3.67136411e-02 -3.54972631e-02 4.00873512e-01 4.45473075e-01
-5.39441407e-01 5.14474273e-01 2.74222523e-01 3.66963983e-01
-1.41944602e-01 5.60616016e-01 -6.50318623e-01 -3.71909708e-01
4.98507082e-01 8.97971213e-01 -3.69191080e-01 -7.29405284e-02
-9.52465534e-02 8.04821312e-01 -4.26641792e-01 1.00007534e+00
4.88783211e-01 -2.73612797e-01 6.40246391e-01 -2.21921191e-01
-2.60419190e-01 5.44132590e-02 -1.04668081e+00 1.02653027e+00
-3.77009332e-01 -6.93382993e-02 -2.72850960e-01 -1.16968179e+00
1.02618468e+00 5.41528642e-01 9.00468647e-01 -1.03872907e+00
2.30851620e-01 2.81830400e-01 1.11742532e-02 -1.02997398e+00
-7.67365754e-01 -1.11827470e-01 1.89421788e-01 -1.25313401e-01
-5.50525069e-01 8.17310512e-01 -1.29216805e-01 -3.34069252e-01
1.03552592e+00 -1.38457179e-01 8.23814631e-01 -2.67494917e-01
3.18859816e-01 2.94570595e-01 9.68355536e-01 1.86876550e-01
-2.48638555e-01 -2.33488232e-02 1.39862418e-01 -1.24123514e+00
-5.19534886e-01 -6.42838180e-01 -4.00652051e-01 4.58422482e-01
3.65962200e-02 -8.41292664e-02 -3.81094664e-01 -6.94929063e-01
3.75172049e-01 6.17318392e-01 -2.13605121e-01 1.90484211e-01
-2.62122899e-01 -1.22710335e+00 6.13708887e-03 3.62817615e-01
7.45738566e-01 -5.39299130e-01 -7.98149407e-01 3.37952346e-01
-6.49701431e-02 -8.19839120e-01 6.38217032e-01 1.20891646e-01
-1.26699495e+00 -1.31990480e+00 -3.41742605e-01 -6.51032865e-01
5.80937266e-01 -1.04347236e-01 1.21567023e+00 1.67641863e-01
-7.29210496e-01 8.41640681e-03 -2.81363815e-01 -7.05816686e-01
-3.80522609e-01 -2.50410438e-01 2.07920268e-01 -1.32321686e-01
8.79740655e-01 -6.60252452e-01 -1.02810574e+00 2.48828381e-01
-8.88175666e-02 1.83346197e-01 9.61396992e-01 2.38078192e-01
8.30428958e-01 6.11583054e-01 9.49642956e-01 -9.09501851e-01
-6.88990578e-02 -9.55547035e-01 -5.65567374e-01 8.41079578e-02
-1.60172141e+00 -2.53712654e-01 1.20427258e-01 -6.07707463e-02
-4.94129807e-01 2.00919718e-01 -2.35531852e-02 6.73757121e-02
-6.81035042e-01 4.91546512e-01 -2.71499932e-01 1.00227274e-01
2.67552376e-01 8.57333168e-02 6.59113079e-02 -9.15273547e-01
-2.80532569e-01 8.80232990e-01 -7.95033723e-02 2.18729749e-01
-6.19666055e-02 2.93480098e-01 3.98670286e-01 -4.59396809e-01
-5.23936331e-01 -4.39840794e-01 -1.67666033e-01 -4.37215835e-01
8.98616612e-01 -1.06861460e+00 -9.73637283e-01 3.44720691e-01
-4.41557586e-01 1.82128756e-03 2.66343027e-01 6.64772391e-01
-2.93202609e-01 -3.14294308e-01 -2.96109617e-01 -8.44454944e-01
-7.05718100e-01 -4.85592902e-01 3.18522185e-01 1.33686081e-01
-2.55095065e-01 -9.00115848e-01 2.28571936e-01 6.52143121e-01
7.71963000e-01 7.83229768e-01 1.27625835e+00 -6.13957703e-01
-2.99301147e-01 -4.59521204e-01 -2.73745298e-01 1.63246527e-01
3.64817649e-01 -5.27582578e-02 -3.97289306e-01 -9.05192494e-02
6.12509437e-02 5.65739095e-01 3.09976220e-01 5.86073577e-01
8.31978858e-01 -8.24847400e-01 -6.89499021e-01 3.08212966e-01
1.90008116e+00 6.39448464e-01 7.05384374e-01 3.28428298e-01
2.55437940e-01 1.33704692e-01 5.70642412e-01 6.67671442e-01
8.39327931e-01 5.77188909e-01 6.67256236e-01 -1.44853190e-01
-1.00832187e-01 1.22311123e-01 -1.38345286e-01 8.30429792e-01
-5.52635670e-01 2.14319695e-02 -1.20888972e+00 2.79940248e-01
-1.73505020e+00 -8.94318044e-01 -5.66615045e-01 2.35516047e+00
4.38154757e-01 -2.32551172e-02 4.24235225e-01 4.78671342e-01
5.20865262e-01 -7.80230105e-01 -5.02843440e-01 -4.54399049e-01
1.34548023e-01 2.28595451e-01 6.68152511e-01 -1.71639279e-01
-9.23182309e-01 4.61120456e-02 6.01581812e+00 -5.66748157e-02
-1.37546432e+00 2.84782797e-01 9.86853957e-01 -6.47009164e-02
3.32317978e-01 -1.71552688e-01 -3.72760594e-01 8.58465195e-01
1.25071740e+00 -2.38100216e-02 2.05217361e-01 1.04340971e+00
6.13923967e-01 -2.06097677e-01 -5.56400597e-01 8.62007141e-01
-2.67262042e-01 -1.18804669e+00 -8.23974833e-02 2.04777159e-02
6.32583797e-01 1.48406610e-01 -3.67807448e-01 6.16341829e-02
2.59478599e-01 -8.35398674e-01 6.41082451e-02 1.05403233e+00
6.83709979e-01 -9.25493598e-01 1.04742110e+00 3.85038286e-01
-7.42129266e-01 -4.54706788e-01 8.69403221e-03 -5.29602826e-01
-3.92163470e-02 1.23045194e+00 -8.50524247e-01 5.16124547e-01
9.44646299e-01 6.34126842e-01 -2.79451311e-01 1.40017581e+00
1.66608244e-01 7.15885639e-01 -1.34824842e-01 2.84806360e-02
-5.57661593e-01 -2.14469597e-01 1.51411369e-01 8.94179940e-01
6.19431198e-01 3.15857261e-01 1.89406320e-01 -1.33645013e-01
3.11683327e-01 3.10287774e-01 -3.38022262e-01 3.14501733e-01
2.49501899e-01 1.07202303e+00 -5.72603583e-01 -3.73197198e-01
-6.11590385e-01 4.61237252e-01 -5.23074687e-01 -1.82694286e-01
-7.92816043e-01 1.58635437e-01 7.74761856e-01 8.46524954e-01
-5.43213822e-02 1.10963620e-01 -8.03146958e-01 -8.49473000e-01
-4.53673303e-01 -6.90680563e-01 6.78958833e-01 -3.71477097e-01
-1.35947835e+00 2.75554746e-01 -4.60655838e-01 -1.19221330e+00
-1.78957447e-01 -2.41916254e-01 -3.05366844e-01 5.05722642e-01
-1.72175002e+00 -9.55663443e-01 -6.35564983e-01 4.94699210e-01
2.60347724e-01 -1.47808909e-01 1.41374290e+00 7.78647602e-01
-6.78418517e-01 8.05625394e-02 1.75390586e-01 -1.58799157e-01
5.15252054e-01 -8.83459628e-01 -4.15517956e-01 2.16268022e-02
-6.17512524e-01 -9.52376276e-02 5.75413585e-01 -9.20410395e-01
-1.41612446e+00 -1.33580172e+00 1.37920094e+00 1.05543606e-01
3.10733858e-02 2.00416267e-01 -2.74065435e-01 4.19633955e-01
1.25903636e-01 2.31462926e-01 1.16237354e+00 -2.87486110e-02
-6.09638728e-03 -7.39223719e-01 -1.73769343e+00 2.85292715e-02
5.36322176e-01 3.98221880e-01 4.32016095e-03 5.16470015e-01
1.42414659e-01 -3.62996734e-03 -1.25539899e+00 8.92869413e-01
9.89184141e-01 -1.04939449e+00 8.93295825e-01 -6.52237952e-01
2.99516648e-01 -1.61793783e-01 -2.03640178e-01 -9.47315633e-01
-1.68993965e-01 -1.65665731e-01 -3.69217873e-01 1.00318456e+00
1.09932356e-01 -9.28991735e-01 8.58526409e-01 6.33966744e-01
3.86298120e-01 -1.10637224e+00 -1.11440766e+00 -6.37992978e-01
-2.96678275e-01 -3.12962949e-01 6.73324704e-01 1.34423757e+00
-1.69024482e-01 2.35860884e-01 -1.14255838e-01 1.23170286e-01
8.39636147e-01 1.63357660e-01 4.96162862e-01 -1.71874499e+00
6.66051283e-02 2.16988325e-02 -9.87579942e-01 1.51179150e-01
-7.73597240e-01 -6.02967918e-01 -8.74776125e-01 -1.91310418e+00
2.12339759e-01 -9.49042499e-01 -7.58237064e-01 6.71703219e-01
2.90921181e-01 2.74081737e-01 -1.62175849e-01 1.72286779e-01
-2.14429587e-01 -1.67794332e-01 8.05418491e-01 -4.84113395e-02
-2.40930408e-01 6.10745430e-01 -5.02585351e-01 6.65689647e-01
1.18180728e+00 -5.82985580e-01 -4.45673801e-02 -2.60764025e-02
2.58984059e-01 2.87404031e-01 4.07400399e-01 -1.26081502e+00
5.12317903e-02 -4.57958966e-01 7.34306812e-01 -4.59952086e-01
-2.84122467e-01 -1.31140769e+00 9.77028906e-01 1.21850836e+00
-6.64057508e-02 1.86478749e-01 -2.53497720e-01 3.72038305e-01
1.96700960e-01 4.79484856e-01 6.98676825e-01 1.54930159e-01
-4.47140932e-01 -8.46365467e-02 -5.41989684e-01 -5.81915021e-01
1.43696010e+00 1.20789185e-01 -3.99182767e-01 4.62650470e-02
-1.11173582e+00 7.05484524e-02 2.45722130e-01 4.14150089e-01
5.26716232e-01 -1.21889257e+00 -8.29512775e-01 3.21489871e-01
1.98572893e-02 -4.62179959e-01 1.77312031e-01 1.33155096e+00
-6.01873517e-01 5.14920294e-01 -3.89464736e-01 -4.53055769e-01
-1.61421657e+00 9.56007540e-01 1.48958132e-01 -3.06679189e-01
-6.28376842e-01 8.47344622e-02 -9.41002369e-01 -9.74448621e-02
3.23314041e-01 -3.52989256e-01 -2.67127126e-01 -1.13697566e-01
5.41364193e-01 1.04616034e+00 3.25363487e-01 -2.82962590e-01
-7.33362556e-01 4.62656498e-01 1.76457241e-01 7.55246282e-01
1.46303940e+00 -1.73872530e-01 -2.77250707e-01 2.32131168e-01
7.10027933e-01 -1.47156507e-01 -4.26953942e-01 1.04280911e-01
2.92166203e-01 -3.59026998e-01 2.65647143e-01 -1.65480435e+00
-1.28053010e+00 3.54388803e-01 1.54857707e+00 3.83886009e-01
1.58863461e+00 -2.14799508e-01 7.66669154e-01 -9.50888172e-02
6.21398628e-01 -6.67683899e-01 -7.12061882e-01 -2.92981148e-01
3.82708102e-01 -1.22808290e+00 1.97642684e-01 -6.08108997e-01
-4.65098679e-01 1.02311432e+00 1.45878002e-01 -8.98369923e-02
1.05924964e+00 4.17586714e-01 4.92491305e-01 1.34062365e-01
-9.50203657e-01 -4.41275276e-02 -3.05653870e-01 7.27509201e-01
3.67326558e-01 4.80104238e-01 -5.40224731e-01 7.06050217e-01
3.16213638e-01 9.85324442e-01 -9.29957107e-02 1.06344056e+00
-3.23744178e-01 -1.22407591e+00 -2.96124578e-01 8.40305984e-01
-6.85065687e-01 2.91403495e-02 -1.17136136e-01 6.19985044e-01
7.76757121e-01 8.97302389e-01 1.73962876e-01 -2.65351892e-01
1.60178632e-01 4.32950646e-01 3.99399012e-01 1.98758524e-02
-4.33126062e-01 -1.22590706e-01 9.57775116e-02 -4.69536692e-01
-6.73016787e-01 -6.04640186e-01 -1.26413536e+00 -4.78590548e-01
-4.28901799e-02 -2.61085391e-01 1.24160290e+00 8.06604207e-01
1.01319253e+00 5.42098343e-01 7.40588367e-01 -1.04739688e-01
-1.93813384e-01 -7.28157043e-01 -3.46387476e-01 1.97832938e-02
2.71891449e-02 -5.34478486e-01 -1.25369117e-01 1.03242636e-01] | [8.40965747833252, 4.972101211547852] |
8332ab17-73ec-4b26-a602-ab962011ac74 | virtual-multi-view-fusion-for-3d-semantic | 2007.13138 | null | https://arxiv.org/abs/2007.13138v1 | https://arxiv.org/pdf/2007.13138v1.pdf | Virtual Multi-view Fusion for 3D Semantic Segmentation | Semantic segmentation of 3D meshes is an important problem for 3D scene understanding. In this paper we revisit the classic multiview representation of 3D meshes and study several techniques that make them effective for 3D semantic segmentation of meshes. Given a 3D mesh reconstructed from RGBD sensors, our method effectively chooses different virtual views of the 3D mesh and renders multiple 2D channels for training an effective 2D semantic segmentation model. Features from multiple per view predictions are finally fused on 3D mesh vertices to predict mesh semantic segmentation labels. Using the large scale indoor 3D semantic segmentation benchmark of ScanNet, we show that our virtual views enable more effective training of 2D semantic segmentation networks than previous multiview approaches. When the 2D per pixel predictions are aggregated on 3D surfaces, our virtual multiview fusion method is able to achieve significantly better 3D semantic segmentation results compared to all prior multiview approaches and competitive with recent 3D convolution approaches. | ['Xiaoqi Yin', 'Thomas Funkhouser', 'David Ross', 'Caroline Pantofaru', 'Abhijit Kundu', 'Brian Brewington', 'Alireza Fathi'] | 2020-07-26 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4670_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123690511.pdf | eccv-2020-8 | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 4.05499786e-01 4.64386702e-01 -1.29597187e-01 -7.62655914e-01
-8.62106383e-01 -6.73986495e-01 5.60028851e-01 2.68283993e-01
7.35658929e-02 -8.73311684e-02 -1.62169084e-01 -1.73803568e-01
1.51490510e-01 -1.16037667e+00 -1.27105629e+00 -1.81005895e-02
2.80469626e-01 1.05783522e+00 6.73024893e-01 -1.43327966e-01
2.55486697e-01 8.68295193e-01 -2.10439396e+00 3.87599319e-01
3.89245838e-01 1.44634771e+00 4.98823643e-01 6.71195507e-01
-8.45133781e-01 5.81016392e-02 -1.02800369e-01 5.74329682e-02
5.84857404e-01 2.38290742e-01 -9.92293179e-01 6.62699640e-01
1.10497427e+00 -2.24614635e-01 2.97977567e-01 1.13449252e+00
1.60119757e-01 4.35649976e-03 8.48378479e-01 -1.13005769e+00
-1.89279720e-01 3.33083749e-01 -6.12557828e-01 -2.43092477e-01
6.11321390e-01 -2.10821748e-01 9.12034750e-01 -1.11885035e+00
9.07087505e-01 1.78209245e+00 8.74965191e-01 3.80182117e-01
-1.33680630e+00 -2.81238288e-01 5.47678530e-01 -3.30858678e-01
-1.00116873e+00 -8.92035142e-02 1.14886665e+00 -5.09125888e-01
1.10031128e+00 4.10240650e-01 1.05145943e+00 8.03449333e-01
1.79879114e-01 7.59479821e-01 1.25423181e+00 -9.32100713e-02
4.02976871e-01 -5.74249662e-02 1.01130605e-01 9.72911060e-01
-2.84480155e-01 -4.28044558e-01 -4.25541162e-01 -1.51612058e-01
1.06543291e+00 3.02048326e-01 1.77157462e-01 -8.98043156e-01
-1.13054931e+00 7.00495720e-01 4.78174925e-01 -1.55201241e-01
-4.71888065e-01 5.74119687e-01 1.05847232e-01 1.42102703e-01
1.22246742e+00 -3.07610426e-02 -8.11806858e-01 4.69061941e-01
-7.82406330e-01 3.15858573e-01 5.10163009e-01 1.05762625e+00
1.33352172e+00 -9.97264013e-02 5.22917151e-01 9.20521736e-01
6.93120003e-01 7.81293392e-01 -4.49148566e-01 -1.89749444e+00
3.71026397e-01 1.00390089e+00 -2.62451679e-01 -7.74966896e-01
-5.47628582e-01 2.94578642e-01 -3.31635803e-01 4.95125324e-01
-3.07764113e-02 4.81242657e-01 -1.46870339e+00 9.27504241e-01
7.96353340e-01 3.73668969e-02 -2.22923517e-01 6.46552801e-01
1.19673169e+00 3.07495981e-01 2.70090997e-02 5.22460222e-01
1.14425468e+00 -6.96310103e-01 -7.20685646e-02 -4.23341542e-01
2.75978953e-01 -6.19935572e-01 7.17814922e-01 2.18138769e-01
-1.36165810e+00 -6.00556433e-01 -9.61373806e-01 -2.30528414e-01
-5.36245525e-01 -7.32524693e-01 7.19880998e-01 5.36976516e-01
-1.31626618e+00 6.98508263e-01 -8.77075851e-01 -4.08967048e-01
7.84939110e-01 4.43565965e-01 -3.95915359e-01 -2.58903861e-01
-6.40373051e-01 5.14131308e-01 2.42535010e-01 -4.91289228e-01
-9.58615839e-01 -1.03906405e+00 -1.23529112e+00 -5.16643822e-01
2.14602157e-01 -1.18848240e+00 1.25996387e+00 -7.41400182e-01
-1.04435098e+00 1.57616103e+00 -1.03533685e-01 -5.55860139e-02
2.83816248e-01 4.95218337e-02 2.89946645e-01 5.14972508e-01
4.77345586e-01 1.16295648e+00 8.71333420e-01 -2.07212782e+00
-6.16464794e-01 -8.73671830e-01 2.76040524e-01 5.89572310e-01
6.20950520e-01 -6.49135888e-01 -6.44346237e-01 -9.41885114e-02
9.73771393e-01 -5.86141646e-01 -5.77464283e-01 1.91848919e-01
-5.53048790e-01 -2.57957429e-01 9.00684655e-01 -4.12002653e-01
-2.08848402e-01 -1.82683933e+00 3.50444645e-01 3.50855082e-01
5.09197235e-01 -4.66019243e-01 1.55546278e-01 -8.14498439e-02
1.79841697e-01 3.98788065e-01 -3.98125559e-01 -8.17951202e-01
7.51247481e-02 5.67391455e-01 7.79065639e-02 4.65934366e-01
-2.34334156e-01 9.72561657e-01 -6.62477911e-01 -6.74275696e-01
1.00457036e+00 5.70278347e-01 -5.74669182e-01 6.11830354e-02
-9.52869415e-01 5.70709884e-01 -8.73468697e-01 1.02836215e+00
8.74186814e-01 -6.40350819e-01 -8.79590288e-02 -3.23978871e-01
6.02512620e-02 1.48440795e-02 -9.40891206e-01 2.44695663e+00
-5.79862177e-01 2.02743828e-01 4.07982111e-01 -1.03634608e+00
8.96594524e-01 2.26365045e-01 9.82479095e-01 -7.65262902e-01
2.58144826e-01 9.42839980e-02 -1.35947740e+00 -2.19103739e-01
3.55879307e-01 -2.30210423e-01 -1.59900174e-01 1.69020995e-01
2.94168949e-01 -1.25237763e+00 -7.64571011e-01 1.70112357e-01
7.16351986e-01 4.58200574e-01 -1.64170161e-01 -2.05985501e-01
1.82282209e-01 3.62285256e-01 1.58352271e-01 6.05749607e-01
2.34329924e-01 8.61414731e-01 1.65255681e-01 -5.68871260e-01
-1.24231339e+00 -1.58724630e+00 -7.66815469e-02 7.02803135e-01
9.31534946e-01 -4.24847715e-02 -6.19836628e-01 -8.00600350e-01
4.07021731e-01 5.79736292e-01 -4.93342161e-01 6.04409993e-01
-3.76780480e-01 -1.93290055e-01 9.29295421e-02 4.95594829e-01
5.17263353e-01 -3.77916843e-01 -6.73795700e-01 -1.10330202e-01
-2.03373075e-01 -1.44288063e+00 1.86309457e-01 3.90073776e-01
-1.21732223e+00 -1.22040498e+00 -5.03807068e-01 -8.28691244e-01
5.41331112e-01 5.48590362e-01 1.53246439e+00 2.07713936e-02
-2.67086953e-01 1.35758126e+00 -2.49784231e-01 -4.68755245e-01
-4.04105037e-01 -1.65436611e-01 -2.75717974e-01 -4.04448897e-01
-6.05902821e-03 -9.08392489e-01 -4.62716669e-01 1.94300488e-01
-5.43573856e-01 4.84271914e-01 -1.13733530e-01 2.61229519e-02
1.41487885e+00 -3.48944157e-01 -6.86737299e-02 -1.11006141e+00
-2.86836147e-01 -4.70181704e-01 -6.24374866e-01 -3.19543853e-03
-3.41786385e-01 -2.91328132e-01 9.51521769e-02 1.20655082e-01
-9.23525810e-01 3.13059479e-01 -4.64466661e-01 -1.03852403e+00
-6.82672262e-01 -1.61315426e-01 -1.37301415e-01 -3.01674515e-01
3.40635777e-01 -1.52227014e-01 1.65488590e-02 -7.10533798e-01
6.86153948e-01 2.51677305e-01 4.62570488e-02 -5.36063135e-01
5.09958029e-01 8.72874200e-01 3.07411194e-01 -9.57778752e-01
-7.99895763e-01 -6.81159914e-01 -1.08967149e+00 -6.14723146e-01
1.57965994e+00 -1.20344019e+00 -5.98607361e-01 3.12712729e-01
-1.37201917e+00 -6.39127910e-01 -3.51044267e-01 5.21643013e-02
-1.17967570e+00 2.29270339e-01 -4.58690166e-01 -6.30960763e-01
-1.50507204e-02 -1.44949305e+00 2.07625604e+00 -3.18653196e-01
-5.46692610e-02 -1.24803984e+00 -3.56348604e-01 7.44191468e-01
-8.33095983e-02 7.54443824e-01 1.17658317e+00 -2.54774839e-01
-1.01456559e+00 3.05793643e-01 -2.38867760e-01 3.48868847e-01
-1.67858019e-01 -2.00107351e-01 -1.30040359e+00 3.81167740e-01
-4.36870046e-02 -3.49681139e-01 8.13343346e-01 6.47298396e-01
1.31055105e+00 3.34445000e-01 -6.10527277e-01 8.62848818e-01
1.74223745e+00 -1.90846041e-01 1.27020523e-01 6.28650263e-02
1.35896683e+00 7.03798473e-01 4.81633097e-01 1.51552662e-01
9.13092792e-01 3.54481786e-01 1.20294929e+00 -2.96683401e-01
-2.10445330e-01 -1.73266962e-01 -3.13510209e-01 6.83925390e-01
-5.27716465e-02 -2.13475123e-01 -1.06484711e+00 3.76594931e-01
-1.53274381e+00 -2.57729053e-01 -5.72431386e-01 1.84260654e+00
1.07884027e-01 -1.09945482e-03 -1.51807770e-01 -4.95996960e-02
5.41425645e-01 5.91459751e-01 -7.29717255e-01 -4.70798522e-01
-1.14833176e-01 4.68707085e-01 9.60378647e-01 7.83710659e-01
-1.26661444e+00 9.95704949e-01 6.35492468e+00 6.21424377e-01
-5.85340977e-01 3.78323913e-01 7.26181686e-01 1.99949190e-01
-9.08744633e-01 -1.96748331e-01 -5.97984374e-01 -1.36647567e-01
4.93770123e-01 6.80818081e-01 2.18376398e-01 9.15482342e-01
-2.64226556e-01 -1.59486160e-01 -1.07904172e+00 1.32426786e+00
7.39475116e-02 -1.73485959e+00 3.72902989e-01 2.07344487e-01
9.91225898e-01 6.09933794e-01 -1.83757350e-01 -2.67932266e-01
7.04701602e-01 -1.01747596e+00 8.96138251e-01 6.38802290e-01
6.50612652e-01 -7.44722068e-01 2.28029564e-01 5.19330740e-01
-1.39774466e+00 3.66375923e-01 -2.29973614e-01 2.34841287e-01
4.26380605e-01 6.59287512e-01 -6.52148306e-01 8.44841719e-01
9.00165737e-01 1.10587418e+00 -4.08646435e-01 4.15145725e-01
1.70059189e-01 -1.21691637e-02 -5.06492198e-01 5.04213095e-01
2.99834937e-01 -2.10738614e-01 6.52055323e-01 4.45108265e-01
1.58871591e-01 -6.71297759e-02 5.50006807e-01 1.17553341e+00
-1.82060078e-01 -1.86804738e-02 -1.10415423e+00 4.01772171e-01
2.50085980e-01 9.08441722e-01 -1.31791401e+00 -5.52601993e-01
-4.54925984e-01 1.19755316e+00 1.39619261e-01 2.83199191e-01
-5.83364964e-01 4.35131401e-01 7.73830414e-01 6.30880669e-02
4.34742808e-01 -4.25250679e-01 -8.02363217e-01 -8.11780274e-01
-2.16040686e-01 -2.46706247e-01 -1.33634016e-01 -1.02194381e+00
-1.28369331e+00 1.16539955e-01 1.89130709e-01 -8.91513288e-01
1.51405856e-01 -6.27334118e-01 -7.46030435e-02 7.84766376e-01
-1.20206499e+00 -1.54627001e+00 -3.49406958e-01 4.52356339e-01
9.44700718e-01 4.65392917e-01 8.29043686e-01 7.07484111e-02
3.52194279e-01 -5.51976740e-01 -3.80802423e-01 -3.53183180e-01
5.80828302e-02 -1.46324527e+00 9.28010941e-01 1.86091557e-01
8.96515548e-02 -1.44887894e-01 3.65598768e-01 -9.76066172e-01
-1.61700380e+00 -1.26687896e+00 3.59409899e-01 -1.06755865e+00
6.27373978e-02 -4.44041729e-01 -5.83578050e-01 8.88115883e-01
-1.96513370e-01 1.32328868e-02 4.08126742e-01 -1.77220777e-01
-2.73482651e-01 3.62265140e-01 -1.34494054e+00 1.87081143e-01
1.64504111e+00 -6.24674678e-01 -5.26573181e-01 4.57731217e-01
1.26126850e+00 -7.15978920e-01 -1.38387299e+00 8.96028757e-01
4.18892413e-01 -1.17620885e+00 1.84117317e+00 -1.30521610e-01
4.49444920e-01 -3.44652194e-03 -1.04273164e+00 -1.10969830e+00
2.79336840e-01 2.88236648e-01 3.21913809e-01 7.32621253e-01
1.76266238e-01 -4.11921352e-01 9.78254318e-01 4.91886616e-01
-5.41710854e-01 -6.61608636e-01 -9.79252636e-01 -3.42575461e-01
5.02948940e-04 -1.13930309e+00 6.36809409e-01 9.15968418e-01
-9.84856367e-01 1.48933232e-01 2.61037469e-01 4.84133571e-01
1.24954844e+00 5.49359739e-01 8.17912519e-01 -1.80887377e+00
1.79306909e-01 -4.27781224e-01 -6.43508852e-01 -1.38207376e+00
4.32396203e-01 -1.49967039e+00 -2.09979251e-01 -2.16151786e+00
-3.01719815e-01 -5.78985572e-01 2.57253468e-01 2.21138418e-01
4.52207893e-01 5.86816311e-01 1.13257043e-01 -1.56483397e-01
-8.31720054e-01 4.36243534e-01 1.53648150e+00 -4.55354959e-01
-1.15168154e-01 -2.21145973e-01 -1.81185797e-01 1.12991202e+00
5.64855635e-01 -1.75085604e-01 -5.16025543e-01 -6.52375221e-01
1.83373496e-01 3.89977783e-01 8.39681327e-01 -7.50592232e-01
-7.27703124e-02 -1.33547947e-01 6.50259018e-01 -1.32625902e+00
1.04742658e+00 -9.32736933e-01 2.70690471e-01 -6.94997236e-02
4.29853611e-02 -1.85168937e-01 6.17371015e-02 8.58546436e-01
1.59378916e-01 2.95191765e-01 5.95970333e-01 -1.04095066e+00
-1.06159270e+00 5.29519975e-01 -2.91738838e-01 -2.18346808e-02
8.83811653e-01 -6.08520448e-01 2.20782951e-01 4.73107584e-02
-1.16776609e+00 3.29854578e-01 1.10589993e+00 4.59725142e-01
1.03329313e+00 -1.23668969e+00 -5.30201010e-02 2.81076699e-01
-2.30209995e-02 8.42673600e-01 6.25172675e-01 4.46266025e-01
-6.33584976e-01 7.08200037e-02 -7.50744864e-02 -1.41213572e+00
-1.38398325e+00 2.95604140e-01 5.11183679e-01 5.17337024e-01
-9.73542631e-01 1.05304670e+00 5.10378957e-01 -1.26080930e+00
2.07978800e-01 -6.59054160e-01 7.09162429e-02 4.85008098e-02
-1.50496721e-01 3.64480108e-01 1.94749281e-01 -8.50457132e-01
-4.17774588e-01 1.22277236e+00 6.75798237e-01 -1.85289785e-01
1.48661530e+00 -5.23912013e-01 -3.21479708e-01 1.10293090e+00
1.36926234e+00 -4.46272939e-01 -1.21753907e+00 -1.05772093e-02
-4.00461525e-01 -3.59400421e-01 1.91912413e-01 -4.06407028e-01
-1.12521553e+00 1.03561008e+00 5.82792699e-01 2.88094252e-01
5.63022375e-01 7.75877237e-01 7.46535599e-01 1.26366630e-01
9.92329657e-01 -9.63205278e-01 -1.49742095e-02 4.68900323e-01
4.68235016e-01 -1.39355433e+00 3.21513824e-02 -9.67289627e-01
-1.99397668e-01 1.01649308e+00 3.86577249e-01 -3.30012232e-01
1.04994857e+00 3.24162352e-03 4.03330028e-02 -1.03446662e+00
-1.83598682e-01 -1.81790337e-01 3.46666902e-01 6.63153589e-01
-1.61666095e-01 3.42019290e-01 6.00982130e-01 1.00289127e-02
-2.95020510e-02 -5.69824040e-01 8.52835551e-02 8.62525761e-01
-6.46853030e-01 -1.01336515e+00 -4.51640755e-01 6.06068313e-01
-1.13035321e-01 2.12993816e-01 -4.06074435e-01 8.13503206e-01
3.82261992e-01 7.32179880e-01 5.42523026e-01 -5.65941095e-01
4.10040855e-01 2.59831876e-01 8.79506886e-01 -9.10108805e-01
-3.33447486e-01 2.23458752e-01 6.06492609e-02 -1.11207879e+00
-9.53914165e-01 -6.12819612e-01 -1.55127537e+00 -6.51767179e-02
2.58108824e-01 -4.67823118e-01 1.21261299e+00 1.04504025e+00
2.48918131e-01 4.90619212e-01 6.13607705e-01 -1.71772337e+00
3.68004203e-01 -1.96476653e-01 -6.93884909e-01 4.05973166e-01
3.30181241e-01 -8.90135109e-01 -3.71756911e-01 8.73589441e-02] | [8.341943740844727, -2.9837286472320557] |
eabdd415-e1c0-4d37-a281-a5df34f121ef | conffusion-confidence-intervals-for-diffusion | 2211.09795 | null | https://arxiv.org/abs/2211.09795v1 | https://arxiv.org/pdf/2211.09795v1.pdf | Conffusion: Confidence Intervals for Diffusion Models | Diffusion models have become the go-to method for many generative tasks, particularly for image-to-image generation tasks such as super-resolution and inpainting. Current diffusion-based methods do not provide statistical guarantees regarding the generated results, often preventing their use in high-stakes situations. To bridge this gap, we construct a confidence interval around each generated pixel such that the true value of the pixel is guaranteed to fall within the interval with a probability set by the user. Since diffusion models parametrize the data distribution, a straightforward way of constructing such intervals is by drawing multiple samples and calculating their bounds. However, this method has several drawbacks: i) slow sampling speeds ii) suboptimal bounds iii) requires training a diffusion model per task. To mitigate these shortcomings we propose Conffusion, wherein we fine-tune a pre-trained diffusion model to predict interval bounds in a single forward pass. We show that Conffusion outperforms the baseline method while being three orders of magnitude faster. | ['Yedid Hoshen', 'Eliahu Horwitz'] | 2022-11-17 | null | null | null | null | ['facial-inpainting', 'image-inpainting', 'prediction-intervals'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 2.53501803e-01 2.48100698e-01 -2.85817355e-01 -1.77744582e-01
-1.22602677e+00 -7.36284435e-01 8.61234725e-01 5.82580678e-02
-3.96295398e-01 1.05439901e+00 3.02869171e-01 -2.95305401e-01
2.57736742e-01 -8.61449361e-01 -7.56485581e-01 -6.33194566e-01
1.69057399e-01 5.62659204e-01 2.61433899e-01 3.90613019e-01
2.67845005e-01 3.43569756e-01 -1.18543756e+00 3.08398575e-01
1.04293227e+00 9.64213967e-01 2.07008108e-01 7.90542603e-01
-1.09292001e-01 7.97977507e-01 -9.27632213e-01 -5.30113876e-01
3.52303624e-01 -9.68957126e-01 -5.35275102e-01 1.89804420e-01
6.84093982e-02 -7.35105932e-01 -3.85496095e-02 1.11369252e+00
2.97877043e-01 2.20300853e-01 9.60338116e-01 -1.00004160e+00
-7.93980479e-01 5.97960413e-01 -9.68968272e-01 2.08995417e-01
3.35659593e-01 9.75687057e-02 7.77511001e-01 -6.68706357e-01
1.02086806e+00 1.10632789e+00 6.18849039e-01 5.83090246e-01
-1.44654989e+00 -4.40328121e-01 -1.39428049e-01 -2.17885226e-01
-1.35941017e+00 -5.17293096e-01 4.33033109e-01 -4.20043468e-01
5.42008281e-01 2.84238458e-01 5.81589997e-01 1.07775986e+00
6.41037300e-02 7.72679210e-01 1.26443374e+00 -3.99382085e-01
5.14811158e-01 8.56582820e-02 -5.23791850e-01 4.41536874e-01
2.52096099e-03 3.31314392e-02 -4.92830038e-01 -4.33713466e-01
1.37552500e+00 -2.88339108e-01 -2.31499970e-01 -5.65079786e-02
-1.21301901e+00 9.79185343e-01 2.87424147e-01 1.84167787e-01
-5.22739291e-01 4.61665541e-01 1.39656112e-01 5.66037595e-02
8.64687204e-01 2.20889971e-01 8.73835385e-02 -3.09622049e-01
-1.53438318e+00 5.86475968e-01 5.92010736e-01 8.27765882e-01
5.84480345e-01 -1.85545415e-01 -5.00075877e-01 8.36838603e-01
2.13485897e-01 1.80325270e-01 3.03596795e-01 -1.24441457e+00
3.55426759e-01 -9.55263227e-02 5.32468498e-01 -8.53079796e-01
2.95684069e-01 -1.56629190e-01 -6.54514670e-01 5.18834412e-01
7.93044448e-01 -3.03051323e-01 -1.09382927e+00 1.58937800e+00
5.32762706e-01 1.72475174e-01 -2.76375860e-01 1.03253031e+00
5.49749210e-02 1.02160048e+00 -9.04123590e-04 -3.50254595e-01
1.10968113e+00 -8.40026855e-01 -6.57847881e-01 -6.86031133e-02
3.10949147e-01 -8.16984415e-01 1.11673832e+00 5.34725904e-01
-1.42431962e+00 -1.90898091e-01 -9.75075543e-01 -1.94216147e-01
9.79198888e-02 -1.77170336e-01 5.75110018e-01 6.04767323e-01
-1.00651431e+00 7.45914161e-01 -8.82349312e-01 6.65242076e-02
7.99640357e-01 2.42159478e-02 -1.95990518e-01 -5.38120866e-02
-8.31621408e-01 7.59577811e-01 1.01704754e-01 -3.21133763e-01
-9.75632012e-01 -7.84402430e-01 -7.03758180e-01 7.45363533e-02
2.58147806e-01 -8.21108758e-01 1.25781024e+00 -9.75347340e-01
-1.44876194e+00 8.85644615e-01 -3.38594258e-01 -7.23245084e-01
1.17173517e+00 -1.63628265e-01 9.60205719e-02 1.68699846e-01
3.28482181e-01 8.19549084e-01 1.21321177e+00 -1.24236286e+00
-5.98865628e-01 -7.88209140e-02 -9.93775055e-02 1.77068159e-01
-6.19911999e-02 4.64974158e-02 -7.18042910e-01 -8.07711065e-01
-9.76273194e-02 -7.23471880e-01 -5.02670467e-01 4.04067069e-01
-4.35912460e-01 -1.76342562e-01 5.16747296e-01 -6.28975213e-01
1.24559689e+00 -1.96107411e+00 -8.37619975e-02 3.92131358e-02
3.35855395e-01 9.53877941e-02 1.55363798e-01 1.98526114e-01
3.69989216e-01 2.27589354e-01 -3.74854237e-01 -7.06517875e-01
-6.91318065e-02 1.29562216e-02 -5.40550172e-01 5.01695216e-01
6.95618540e-02 7.94469655e-01 -9.75537300e-01 -7.20345259e-01
1.46033138e-01 6.98673069e-01 -5.50221145e-01 2.09507599e-01
-5.50398767e-01 5.69130421e-01 -2.89975762e-01 2.37035036e-01
8.06222022e-01 -3.07528228e-01 -1.65022537e-01 1.79824218e-01
9.95408818e-02 1.46360070e-01 -1.08400857e+00 1.79251254e+00
-5.19101322e-01 6.63307726e-01 -1.21580034e-01 -5.02888680e-01
8.07587624e-01 2.49006450e-01 4.22215432e-01 -4.13995653e-01
-7.91614503e-02 5.94824255e-02 -3.76052707e-01 -5.32214716e-02
5.35078585e-01 -4.94551390e-01 2.15625748e-01 7.40060985e-01
-2.73590267e-01 -4.20009196e-01 2.23210946e-01 2.49396667e-01
1.01396883e+00 2.84005761e-01 8.19900855e-02 1.11946790e-03
5.45556732e-02 -5.11694923e-02 4.71178234e-01 8.30800474e-01
-9.19337422e-02 1.27420855e+00 8.40380907e-01 -3.23763371e-01
-1.46817446e+00 -1.16537178e+00 -2.99974158e-02 5.04417479e-01
-1.11677960e-01 -2.18973085e-01 -1.17048836e+00 -6.73214972e-01
-1.67921543e-01 8.63614619e-01 -8.04214478e-01 3.06901693e-01
-1.87179610e-01 -5.36918044e-01 3.30327660e-01 4.80769008e-01
4.25399363e-01 -9.15380895e-01 -7.55716503e-01 3.28027219e-01
-1.84447840e-01 -1.00521100e+00 -8.01892817e-01 -2.37694457e-01
-8.50136161e-01 -6.78369224e-01 -1.27463210e+00 -2.96845824e-01
8.72940361e-01 -4.58054291e-03 1.16671765e+00 -1.77628800e-01
-7.66045898e-02 -1.67813838e-01 -8.77784640e-02 -2.62441158e-01
-5.58135033e-01 -6.71323016e-02 -4.51744616e-01 -6.14508055e-02
-6.09580465e-02 -5.33628523e-01 -9.07546043e-01 1.03314199e-01
-1.02600026e+00 2.73890615e-01 3.99115503e-01 7.46650994e-01
9.22175229e-01 1.55905262e-01 6.20925367e-01 -9.92310703e-01
9.52203751e-01 -4.32523787e-01 -7.77144909e-01 5.61568961e-02
-5.30885637e-01 1.27902016e-01 5.31853795e-01 -5.64727724e-01
-1.09523392e+00 9.59172286e-03 -5.43336868e-02 -5.42668998e-01
-4.04278673e-02 4.01491135e-01 2.62910128e-01 4.35814023e-01
7.53278315e-01 1.71358988e-01 1.68084636e-01 -1.94318905e-01
6.30966783e-01 4.28864330e-01 6.86229825e-01 -6.46251857e-01
5.35546601e-01 7.64181972e-01 -2.30234414e-01 -5.40450215e-01
-9.08231497e-01 -1.08239338e-01 -2.33123451e-01 -2.08790407e-01
9.04475272e-01 -8.42714787e-01 -3.75189811e-01 1.52819321e-01
-1.25219929e+00 -6.52985215e-01 -5.40381551e-01 2.71012038e-01
-8.44190657e-01 2.47147486e-01 -5.74627936e-01 -1.05662847e+00
-2.47501567e-01 -1.00682604e+00 1.08546674e+00 2.02576146e-01
-4.68671471e-01 -9.18525159e-01 8.79969001e-02 3.18463385e-01
4.99492079e-01 5.85619807e-01 5.65260351e-01 -6.19691946e-02
-4.95714277e-01 -2.76679963e-01 -3.24048907e-01 3.25221777e-01
5.26096225e-02 8.20659027e-02 -8.75925720e-01 8.90571345e-03
3.33873671e-03 -3.15404534e-01 7.62305856e-01 8.34120452e-01
1.37809432e+00 -4.05916840e-01 -3.43174726e-01 5.71158588e-01
1.25687957e+00 8.90870467e-02 9.04135406e-01 1.78525284e-01
3.53679419e-01 3.69524091e-01 4.98428822e-01 6.55672252e-01
3.00631464e-01 4.83281761e-01 5.74044883e-02 -6.78868443e-02
-3.10644507e-01 -4.91077960e-01 1.16136283e-01 4.45753597e-02
-3.23554203e-02 -4.70295608e-01 -5.62556803e-01 8.12844992e-01
-1.79261148e+00 -1.07457709e+00 1.09412521e-01 2.41059089e+00
1.20455623e+00 3.20787102e-01 2.43076086e-01 4.87947352e-02
6.19735539e-01 2.42371485e-01 -6.02092445e-01 -3.12844247e-01
2.53852140e-02 4.19615433e-02 2.65898556e-01 6.99426353e-01
-7.85875022e-01 7.99824953e-01 6.72967052e+00 1.04706371e+00
-8.43844473e-01 1.39047906e-01 1.26407063e+00 -2.11872697e-01
-6.86989188e-01 9.35739428e-02 -5.07447839e-01 6.92608178e-01
8.84234786e-01 -1.74177557e-01 4.32035029e-01 6.89861238e-01
3.61313701e-01 -5.60890794e-01 -8.85776997e-01 9.96926486e-01
-9.62319300e-02 -1.56841743e+00 -9.33504403e-02 2.79177785e-01
1.02649224e+00 -2.46540606e-01 2.27346703e-01 -2.13474944e-01
5.56920767e-01 -1.24850404e+00 9.81781781e-01 4.42251533e-01
9.43542778e-01 -9.17572796e-01 1.95715740e-01 4.63011026e-01
-6.28594875e-01 4.24202204e-01 -5.27029157e-01 4.76523787e-02
3.96320999e-01 1.18417680e+00 -8.90867651e-01 -1.73541252e-02
4.58938181e-01 2.22162977e-01 -5.97939044e-02 9.52953160e-01
-2.95316815e-01 6.83451474e-01 -4.61279213e-01 3.10099274e-01
1.51912227e-01 -3.83031905e-01 3.48170131e-01 1.08474314e+00
6.46684885e-01 2.71427464e-02 -2.66737312e-01 1.43182683e+00
-2.05131158e-01 -2.69134551e-01 -6.57335222e-01 -1.39326379e-01
6.10916138e-01 1.04816341e+00 -1.06866491e+00 -5.06424427e-01
-3.35735440e-01 1.32154214e+00 3.33618164e-01 3.82997841e-01
-1.03763783e+00 -2.05564409e-01 4.20296192e-01 2.58526742e-01
4.47851330e-01 -3.19940746e-01 -6.79970145e-01 -9.74657834e-01
4.21746038e-02 -7.54272282e-01 1.77717850e-01 -9.63148952e-01
-1.11293674e+00 6.27510369e-01 -8.02693516e-02 -1.01394331e+00
-5.85568607e-01 -8.75271112e-02 -5.41610897e-01 1.04054904e+00
-1.26879144e+00 -9.13409472e-01 -1.24828599e-01 3.67016256e-01
5.01240492e-01 3.05936724e-01 6.60660148e-01 7.47085037e-03
-2.51015335e-01 4.50370818e-01 7.42349178e-02 -5.46053834e-02
7.17673540e-01 -1.32680285e+00 5.00269353e-01 8.98058653e-01
2.47039452e-01 3.40511978e-01 1.03148484e+00 -6.52899981e-01
-1.00142431e+00 -6.88220561e-01 8.22256386e-01 -4.07377303e-01
4.87937689e-01 -3.29587191e-01 -8.33803296e-01 5.07427275e-01
1.22277446e-01 2.17987105e-01 4.81438160e-01 -1.60373867e-01
-2.88094878e-01 1.91588849e-01 -1.33918715e+00 7.18610823e-01
7.39289403e-01 -3.85194421e-01 -6.77512512e-02 4.06180531e-01
3.83923918e-01 -6.37539327e-01 -8.25776696e-01 -7.44286329e-02
4.68718916e-01 -1.29059935e+00 9.09415960e-01 -7.05301389e-02
1.00492287e+00 -4.02281404e-01 8.83186385e-02 -1.32640827e+00
4.54425029e-02 -1.00355196e+00 -3.87619436e-01 1.33566070e+00
4.71370131e-01 -3.73445153e-01 1.00715995e+00 9.96306241e-01
2.58244306e-01 -8.93194199e-01 -9.64381516e-01 -6.57005668e-01
2.04453319e-01 -5.07544219e-01 6.74525023e-01 6.02399588e-01
-1.91585958e-01 5.61960302e-02 -4.13984269e-01 -7.61140883e-02
7.72613049e-01 1.19303428e-01 8.12604725e-01 -6.78696513e-01
-4.68317747e-01 -4.99460310e-01 -9.12409499e-02 -1.33291936e+00
-2.16185451e-01 -3.34685057e-01 1.58659354e-01 -1.62401092e+00
2.55266696e-01 -6.85612738e-01 1.22210830e-01 1.51033148e-01
-3.21396440e-01 4.20617163e-01 4.82296720e-02 3.62085909e-01
-3.84248525e-01 3.88268292e-01 1.43685579e+00 1.29360259e-01
-9.45664644e-02 -8.42058957e-02 -8.48778009e-01 5.83346725e-01
5.49997926e-01 -5.15897810e-01 -6.25068188e-01 -5.08458495e-01
2.55615920e-01 6.08770430e-01 2.84483820e-01 -8.94247949e-01
1.05664492e-01 -3.03407967e-01 6.28613889e-01 -5.46179652e-01
2.69123226e-01 -5.15226781e-01 2.45322630e-01 1.71632588e-01
-4.44506794e-01 -1.64292715e-02 -1.89505203e-03 6.62026286e-01
-1.61794916e-01 -3.02403420e-01 8.94009054e-01 -8.97744521e-02
-9.52335522e-02 2.44757682e-01 -2.60035396e-01 1.03779376e-01
1.08302259e+00 -2.86281765e-01 -4.40928824e-02 -8.47173512e-01
-5.12726665e-01 -3.11688092e-02 8.00575495e-01 -2.29002256e-02
5.50834835e-01 -1.21194041e+00 -7.86834359e-01 8.03057775e-02
-2.89733738e-01 5.23532212e-01 1.36910409e-01 5.56797802e-01
-6.29466116e-01 -2.87242681e-02 2.03614280e-01 -4.43066984e-01
-8.53750765e-01 5.15802860e-01 2.04337269e-01 -3.76104176e-01
-7.77678370e-01 1.02144563e+00 3.07002217e-01 2.14011908e-01
8.26899037e-02 -3.15117031e-01 2.68375576e-01 -5.59129119e-02
8.20208788e-01 2.55082071e-01 -2.84689844e-01 -2.61863053e-01
1.88437719e-02 1.88181370e-01 -1.35137454e-01 -9.04524148e-01
1.27228975e+00 -1.25787854e-01 7.95503855e-02 2.82101840e-01
9.42249000e-01 9.20395926e-02 -2.08474135e+00 -7.70748630e-02
-2.16747329e-01 -7.57068515e-01 1.58673435e-01 -8.12580943e-01
-9.91551042e-01 7.14531422e-01 1.29282475e-01 3.01339716e-01
1.04785013e+00 -1.87891610e-02 1.01733458e+00 -4.01773661e-01
1.87301382e-01 -1.09052551e+00 -3.36248390e-02 -6.67137429e-02
8.77409637e-01 -1.09290338e+00 1.71122730e-01 -2.47543082e-01
-8.97176325e-01 9.04380858e-01 1.48525760e-01 -2.75419980e-01
5.11643589e-01 4.34740156e-01 5.20618968e-02 4.07232605e-02
-7.01812446e-01 2.10321620e-01 2.63550162e-01 6.95953369e-01
5.08312106e-01 -4.52161357e-02 -3.53488117e-01 2.05840796e-01
-2.67453015e-01 3.25515836e-01 5.19347668e-01 7.80044377e-01
-2.42159590e-01 -9.69071388e-01 -5.46970487e-01 4.31127608e-01
-7.66767085e-01 -1.21897772e-01 -3.24943773e-02 4.01046753e-01
-1.65482879e-01 8.82310092e-01 1.58706576e-01 1.10782169e-01
-6.11326806e-02 -1.91795409e-01 6.71878219e-01 -4.21798855e-01
-9.42605212e-02 3.81390125e-01 -6.03658007e-03 -4.01682347e-01
-1.65910840e-01 -7.43556499e-01 -1.22708952e+00 -4.33635533e-01
-3.61561567e-01 2.45925542e-02 7.00176179e-01 8.84242535e-01
4.27432120e-01 1.42088279e-01 4.84598309e-01 -9.41174924e-01
-6.37808979e-01 -8.97098958e-01 -5.46252728e-01 4.95588124e-01
2.40300193e-01 -3.97150278e-01 -2.64598697e-01 3.00823659e-01] | [11.19501781463623, -0.33948659896850586] |
27c0aabf-fad8-473f-a053-676f5231f970 | image-dehazing-transformer-with-transmission | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Guo_Image_Dehazing_Transformer_With_Transmission-Aware_3D_Position_Embedding_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Guo_Image_Dehazing_Transformer_With_Transmission-Aware_3D_Position_Embedding_CVPR_2022_paper.pdf | Image Dehazing Transformer With Transmission-Aware 3D Position Embedding | Despite single image dehazing has been made promising progress with Convolutional Neural Networks (CNNs), the inherent equivariance and locality of convolution still bottleneck dehazing performance. Though Transformer has occupied various computer vision tasks, directly leveraging Transformer for image dehazing is challenging: 1) it tends to result in ambiguous and coarse details that are undesired for image reconstruction; 2) previous position embedding of Transformer is provided in logic or spatial position order that neglects the variational haze densities, which results in the sub-optimal dehazing performance. The key insight of this study is to investigate how to combine CNN and Transformer for image dehazing. To solve the feature inconsistency issue between Transformer and CNN, we propose to modulate CNN features via learning modulation matrices (i.e., coefficient matrix and bias matrix) conditioned on Transformer features instead of simple feature addition or concatenation. The feature modulation naturally inherits the global context modeling capability of Transformer and the local representation capability of CNN. We bring a haze density-related prior into Transformer via a novel transmission-aware 3D position embedding module, which not only provides the relative position but also suggests the haze density of different spatial regions. Extensive experiments demonstrate that our method, DeHamer, attains state-of-the-art performance on several image dehazing benchmarks. | ['Chongyi Li', 'Wenqi Ren', 'Runmin Cong', 'Saeed Anwar', 'Qixin Yan', 'Chun-Le Guo'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['image-dehazing'] | ['computer-vision'] | [ 4.60358769e-01 -6.26676679e-02 1.52807951e-01 -8.49340260e-02
-3.69558126e-01 -2.54668951e-01 7.04579413e-01 -2.38800049e-01
-1.88913032e-01 4.13805395e-01 1.99299112e-01 -2.83547074e-01
-1.25280246e-01 -9.20649052e-01 -9.01047111e-01 -1.24341583e+00
1.98949501e-01 -5.19141257e-01 4.63089168e-01 -5.37106514e-01
1.09060593e-01 2.97061414e-01 -1.49554992e+00 3.37107092e-01
1.02224922e+00 1.29900169e+00 2.40280375e-01 4.53123152e-01
1.52590916e-01 1.01442611e+00 -5.98731220e-01 -4.73799467e-01
5.81404924e-01 -4.04773891e-01 -3.37478161e-01 3.02862488e-02
8.13376665e-01 -6.18684769e-01 -8.06109369e-01 1.35687912e+00
2.31830686e-01 1.37555357e-02 6.97504282e-01 -1.20804429e+00
-1.19702423e+00 3.65420371e-01 -5.90807080e-01 3.59199017e-01
-4.66631413e-01 4.32587385e-01 8.47371995e-01 -9.10609424e-01
2.55654044e-02 1.15499210e+00 6.84158504e-01 3.76392722e-01
-1.31232929e+00 -8.91227007e-01 1.95091784e-01 4.12849069e-01
-1.83142793e+00 -3.53781968e-01 9.69041705e-01 -3.06044340e-01
7.85140991e-01 1.61068544e-01 6.14793718e-01 7.90095985e-01
6.01802230e-01 5.45134723e-01 1.03209841e+00 -1.51040807e-01
4.68956269e-02 2.05924928e-01 -3.30546677e-01 8.00557554e-01
2.91471273e-01 2.87487835e-01 -6.10704184e-01 3.43560278e-01
9.91790771e-01 2.14242384e-01 -4.99640405e-01 -2.00956270e-01
-1.06571782e+00 8.45993042e-01 9.01915610e-01 1.26569107e-01
-2.62665480e-01 4.49701250e-01 -9.64970887e-03 3.75010669e-01
5.35888553e-01 4.31219250e-01 -5.67886867e-02 4.30686355e-01
-9.99255657e-01 2.85574824e-01 1.48377776e-01 9.98907447e-01
1.10626292e+00 3.83417785e-01 -4.03523475e-01 7.08951354e-01
6.64515644e-02 3.17890406e-01 1.55603826e-01 -7.02388167e-01
3.77229333e-01 4.65255916e-01 -7.34769031e-02 -1.31090581e+00
1.05645239e-01 -5.20796418e-01 -1.24789953e+00 4.06559885e-01
1.78010046e-01 2.17994094e-01 -1.11629534e+00 1.51379538e+00
9.39617604e-02 5.82409561e-01 -1.75410107e-01 8.86234462e-01
6.44969523e-01 9.62945461e-01 -2.21136749e-01 9.17913765e-02
1.30420589e+00 -1.04199934e+00 -8.85884106e-01 -8.95396173e-02
2.58458048e-01 -7.15602696e-01 8.65511477e-01 3.70557219e-01
-1.07066393e+00 -4.73138809e-01 -1.49684453e+00 -5.29676795e-01
-4.81143177e-01 -3.30936670e-01 3.21187407e-01 6.90416455e-01
-1.25449693e+00 5.49413443e-01 -8.13157916e-01 1.45926625e-01
7.59867251e-01 2.28802800e-01 -1.66407704e-01 -3.50927472e-01
-1.31372118e+00 7.80983984e-01 2.69090146e-01 3.57397169e-01
-1.01453567e+00 -1.16766453e+00 -1.03590250e+00 1.65280908e-01
2.98469394e-01 -6.51375949e-01 6.54256821e-01 -8.72127473e-01
-1.58565795e+00 3.33735734e-01 7.62175303e-03 -5.38423359e-01
3.99577916e-01 -1.71912357e-01 -3.69658291e-01 2.64234722e-01
-1.44627705e-01 7.89970756e-01 1.48860455e+00 -1.26270819e+00
-6.09984577e-01 -1.63178787e-01 1.35553956e-01 1.60085648e-01
-6.87860131e-01 -2.74868399e-01 -4.37650979e-01 -1.24316764e+00
2.48872191e-02 -5.57129085e-01 -1.54849172e-01 5.17528296e-01
-4.18682873e-01 1.80929378e-01 1.13447714e+00 -7.40497231e-01
1.29354513e+00 -2.41030550e+00 2.62164950e-01 9.91252884e-02
5.09028912e-01 2.81591326e-01 -9.93771553e-02 2.24707454e-01
-9.42182839e-02 2.64662236e-01 -4.70160097e-01 -2.71035314e-01
-6.93397522e-02 1.46891907e-01 -5.93417346e-01 6.79584444e-01
7.53137052e-01 9.83081341e-01 -7.92353332e-01 -3.77006322e-01
5.51816404e-01 1.09946537e+00 -9.42788541e-01 1.60398841e-01
-4.28814180e-02 3.42182428e-01 -2.24482179e-01 6.35260046e-01
1.00252783e+00 -1.87244862e-01 -3.98116320e-01 -7.03672886e-01
-2.80976564e-01 2.76434124e-01 -7.88643122e-01 1.49122894e+00
-4.53674048e-01 6.99730933e-01 6.31355792e-02 -1.00549829e+00
5.49032986e-01 1.13500372e-01 1.32430434e-01 -8.24487507e-01
5.59437759e-02 3.89628522e-02 -1.72518179e-01 -1.96219325e-01
5.53980231e-01 -3.01780730e-01 2.00436831e-01 -1.49932161e-01
5.81889488e-02 -3.63785446e-01 -4.54023063e-01 9.15443078e-02
9.72087979e-01 -1.08929217e-01 -4.66348492e-02 -5.58013797e-01
3.11165124e-01 -3.65774930e-01 4.56182033e-01 6.26757443e-01
-7.11254627e-02 9.21337962e-01 3.52155238e-01 -3.78046155e-01
-1.00706899e+00 -1.20460117e+00 -3.69886070e-01 6.42972708e-01
4.80433524e-01 -3.98955762e-01 -6.94361091e-01 -3.51451427e-01
-2.58912206e-01 6.07064962e-01 -8.32262516e-01 -7.56254315e-01
-6.15493655e-01 -6.38873935e-01 4.81492758e-01 1.58855379e-01
1.09447038e+00 -3.36083651e-01 -2.93604285e-01 3.76587771e-02
-1.67136222e-01 -1.12647164e+00 -8.20210934e-01 1.60889909e-01
-5.80524981e-01 -5.91184855e-01 -6.93820477e-01 -7.60048330e-01
5.32781482e-01 7.92533398e-01 7.73093462e-01 1.79872572e-01
-1.96234167e-01 -5.70944585e-02 -3.30303699e-01 -3.94048572e-01
-1.05207250e-01 -8.31630379e-02 -2.45862067e-01 3.36661637e-01
-6.04139417e-02 -1.02556634e+00 -1.09467208e+00 2.96489686e-01
-1.50794518e+00 1.74763605e-01 7.94613183e-01 1.09652817e+00
3.82807106e-01 6.77069247e-01 1.23443156e-01 -6.67917788e-01
1.19787313e-01 -5.21723270e-01 -5.42398691e-01 -1.33069716e-02
-7.07570910e-01 8.35623816e-02 5.70374489e-01 -4.49225485e-01
-8.78489912e-01 -3.45079422e-01 -5.34820631e-02 -8.98109555e-01
2.93487608e-01 3.48029196e-01 -4.24814403e-01 -3.90898645e-01
4.37940717e-01 5.48873782e-01 1.59480095e-01 -2.82918513e-01
3.12413096e-01 2.24560231e-01 6.38441026e-01 -3.29524666e-01
1.35710430e+00 6.95992172e-01 -1.21079981e-02 -1.03177106e+00
-6.37234271e-01 -9.59309414e-02 -2.89223194e-01 -2.24878900e-02
1.02076781e+00 -1.18978703e+00 -3.54940861e-01 7.93923557e-01
-1.12076974e+00 -2.42157593e-01 -9.38745961e-02 3.56205069e-02
-8.07458907e-03 4.53540236e-01 -7.16158569e-01 -4.68912661e-01
-8.25669616e-02 -1.34542871e+00 1.10652065e+00 -4.67593968e-02
3.57400298e-01 -9.22138929e-01 -5.61402977e-01 2.05090150e-01
7.11355031e-01 2.30693430e-01 1.19502056e+00 1.55140266e-01
-1.25171196e+00 1.37394279e-01 -5.46526074e-01 8.05518627e-01
3.05825949e-01 -2.10193411e-01 -1.12332439e+00 -3.56591940e-01
1.95459589e-01 9.26219448e-02 1.31480277e+00 2.87233055e-01
1.45633042e+00 -6.23221636e-01 -1.61256343e-02 1.10936069e+00
1.55931151e+00 -4.76635844e-02 1.14443374e+00 2.80311078e-01
1.21325064e+00 5.34682870e-01 2.46575713e-01 2.48496339e-01
2.94852257e-01 7.84395397e-01 7.99500287e-01 -1.82191312e-01
-4.18379277e-01 -3.57412159e-01 4.77630377e-01 7.38764524e-01
-5.61045203e-03 -2.23459840e-01 -5.20831943e-01 5.35433769e-01
-1.48909092e+00 -7.52571106e-01 2.69782841e-01 2.09110618e+00
9.42419052e-01 3.72623354e-02 -2.47051045e-01 1.75755247e-01
5.78974545e-01 6.52443349e-01 -3.63159090e-01 2.32901629e-02
-2.16785565e-01 3.59864473e-01 8.81148934e-01 5.52963555e-01
-1.12629783e+00 9.44129705e-01 5.64356279e+00 1.21243954e+00
-1.33712721e+00 2.10168079e-01 6.00186825e-01 -1.54588118e-01
-5.58044970e-01 -1.36608809e-01 -6.07417941e-01 7.41148889e-01
6.01440907e-01 1.77025914e-01 6.45514548e-01 3.07625771e-01
-7.93209299e-03 2.16069907e-01 -9.21456993e-01 9.98195410e-01
1.19935401e-01 -1.69305134e+00 7.27917492e-01 2.12730825e-01
8.41138363e-01 -2.28302702e-01 7.24947989e-01 1.72869191e-01
-6.96274713e-02 -1.29960918e+00 1.24355316e+00 2.08007529e-01
9.26675141e-01 -9.53189313e-01 5.78399122e-01 5.08426651e-02
-1.19555438e+00 -1.21821404e-01 -4.68300819e-01 4.50263657e-02
-2.22023129e-01 7.45501995e-01 -4.85658318e-01 6.47359967e-01
1.03453255e+00 1.05792856e+00 -5.16875684e-01 6.56281710e-01
-7.04715028e-02 5.93111932e-01 -8.09762254e-02 3.34082931e-01
3.91356409e-01 -2.28480250e-01 5.99235177e-01 1.05518115e+00
3.73457640e-01 1.36306927e-01 -2.23357067e-01 1.23286474e+00
-2.26220246e-02 -3.95104736e-01 -6.89474761e-01 8.46902877e-02
4.93233800e-01 9.15260255e-01 -4.24469709e-01 -3.96341011e-02
-4.78589892e-01 1.12958813e+00 7.86698312e-02 5.77117682e-01
-9.13514614e-01 -4.88441139e-01 1.14590359e+00 3.28010619e-01
7.42007673e-01 -2.92119473e-01 -2.60541350e-01 -1.13156998e+00
2.21759111e-01 -7.77542114e-01 -1.17940508e-01 -6.28808618e-01
-1.30241966e+00 5.31012058e-01 2.18839850e-03 -1.38837564e+00
3.86602402e-01 -7.68143594e-01 -5.77448666e-01 8.90881419e-01
-2.11824608e+00 -1.34651411e+00 -3.98143530e-01 8.60610843e-01
5.50031900e-01 9.07519534e-02 3.71845961e-01 6.88443959e-01
-6.78732038e-01 9.59259689e-01 -1.92477126e-02 1.05401583e-01
6.63000524e-01 -1.01950991e+00 1.67212844e-01 1.00804114e+00
-2.01509461e-01 6.19615555e-01 8.19072604e-01 -2.87944078e-01
-1.49872756e+00 -1.62653244e+00 3.40789080e-01 -5.09905696e-01
7.12366104e-01 -4.70212936e-01 -1.13018978e+00 5.94820142e-01
3.91885519e-01 2.25922778e-01 2.50846148e-01 -5.74000776e-01
-8.12051117e-01 -3.94285142e-01 -9.59824979e-01 8.53617251e-01
9.98173892e-01 -8.22143018e-01 -1.99803233e-01 -5.69085181e-02
1.19415951e+00 -4.02703285e-01 -6.69681847e-01 4.02559966e-01
3.11714739e-01 -1.12006330e+00 1.26001215e+00 -4.87396494e-02
8.13583910e-01 -6.29486918e-01 -3.63690853e-01 -1.29290438e+00
-4.32409912e-01 -8.81502032e-01 -1.11488827e-01 1.21877110e+00
2.00575083e-01 -7.00758517e-01 4.85205770e-01 2.62595415e-01
-3.88016343e-01 -7.94881225e-01 -9.55771029e-01 -7.05219924e-01
3.22363794e-01 -2.40581349e-01 7.24698067e-01 1.08097887e+00
-5.81310928e-01 6.23661578e-02 -7.58083999e-01 6.15981579e-01
7.67029345e-01 -3.83696407e-01 3.45237911e-01 -6.75611019e-01
-1.68002397e-01 -5.22439957e-01 -6.73888922e-01 -1.22931504e+00
-4.04917002e-02 -7.08445549e-01 1.57229617e-01 -1.04591548e+00
-2.59149652e-02 -3.36087674e-01 -5.31513691e-01 4.25629169e-01
-3.32273602e-01 5.08521497e-01 6.15882576e-02 1.73996016e-01
-1.58400208e-01 9.15863216e-01 1.48345363e+00 -5.84700048e-01
6.08462505e-02 -3.80903542e-01 -8.76383066e-01 3.64728093e-01
5.34805655e-01 -3.50035429e-01 -6.59522355e-01 -7.38908708e-01
1.98579907e-01 -3.34472448e-01 8.97383392e-01 -1.00251460e+00
3.24764252e-01 -9.24700424e-02 3.39331061e-01 -2.22442314e-01
4.40585464e-01 -8.97599399e-01 -8.39047879e-03 2.05714360e-01
-7.37897381e-02 -8.02716389e-02 2.56569177e-01 9.04224873e-01
-4.97076422e-01 2.14526236e-01 1.01644611e+00 1.17782421e-01
-6.99496388e-01 5.71554840e-01 -3.63738328e-01 -1.59669489e-01
6.20115817e-01 -4.47808146e-01 -3.84875178e-01 -3.07216197e-01
-1.26416739e-02 -1.26782492e-01 5.92006207e-01 4.15540963e-01
8.33375216e-01 -1.43769383e+00 -6.71324551e-01 3.80550057e-01
1.30038977e-01 3.83092195e-01 4.57745701e-01 8.89179111e-01
-6.03867352e-01 6.81793317e-02 -1.64035067e-01 -6.44400954e-01
-8.32229078e-01 5.92612207e-01 5.23957193e-01 8.46465155e-02
-8.85481894e-01 1.06939507e+00 9.01769161e-01 1.20802119e-01
1.00663945e-01 -4.40268904e-01 7.21997544e-02 -3.38526666e-01
7.37697959e-01 4.72150184e-02 1.42737553e-01 -4.64939624e-01
-7.01043159e-02 4.78996277e-01 -3.07693571e-01 1.12607181e-01
1.18746948e+00 -2.80359149e-01 -3.93127084e-01 1.74466684e-01
1.44537163e+00 -9.00152326e-02 -1.72970140e+00 -4.60130006e-01
-4.72285330e-01 -8.90319228e-01 7.04722106e-01 -3.15463901e-01
-1.44704294e+00 1.13766062e+00 6.18914425e-01 9.82725322e-02
1.26158011e+00 -2.33665034e-01 9.68793094e-01 1.55817524e-01
3.00323755e-01 -7.19728947e-01 2.93833643e-01 3.27632338e-01
9.28873062e-01 -1.09279716e+00 1.14562422e-01 -6.39763892e-01
-2.57171124e-01 6.98427022e-01 5.29677689e-01 -2.43615091e-01
9.45799351e-01 2.39853576e-01 -9.00118649e-02 -2.15824008e-01
-5.63829780e-01 2.11445644e-01 4.51874375e-01 5.17292678e-01
2.85105053e-02 -1.80242896e-01 3.99788857e-01 3.75282526e-01
-1.33015156e-01 -3.64472628e-01 3.41471016e-01 8.39092612e-01
-2.71467537e-01 -6.21608198e-01 -2.17998222e-01 2.14601368e-01
-4.50280637e-01 -5.85615575e-01 2.17433364e-04 7.52615333e-01
3.38420182e-01 8.41043711e-01 1.29443869e-01 -6.59391224e-01
1.78848743e-01 -4.52477902e-01 5.97102523e-01 -4.12865162e-01
-3.70217830e-01 1.25510216e-01 -3.75728279e-01 -5.51204562e-01
-2.93549478e-01 -2.25665823e-01 -8.28061044e-01 -5.84933221e-01
-3.35377067e-01 -1.00959174e-01 2.64956027e-01 9.17995453e-01
3.85096133e-01 8.37657094e-01 7.10361183e-01 -1.05613041e+00
-4.54963177e-01 -7.48295486e-01 -8.11686039e-01 1.12258591e-01
9.48375165e-01 -8.38846684e-01 -6.21785283e-01 1.50938928e-01] | [10.944387435913086, -3.0819332599639893] |
03795bf3-a087-4067-b6cd-b0a6f2b26953 | boosting-offline-reinforcement-learning-via | 2210.09241 | null | https://arxiv.org/abs/2210.09241v1 | https://arxiv.org/pdf/2210.09241v1.pdf | Boosting Offline Reinforcement Learning via Data Rebalancing | Offline reinforcement learning (RL) is challenged by the distributional shift between learning policies and datasets. To address this problem, existing works mainly focus on designing sophisticated algorithms to explicitly or implicitly constrain the learned policy to be close to the behavior policy. The constraint applies not only to well-performing actions but also to inferior ones, which limits the performance upper bound of the learned policy. Instead of aligning the densities of two distributions, aligning the supports gives a relaxed constraint while still being able to avoid out-of-distribution actions. Therefore, we propose a simple yet effective method to boost offline RL algorithms based on the observation that resampling a dataset keeps the distribution support unchanged. More specifically, we construct a better behavior policy by resampling each transition in an old dataset according to its episodic return. We dub our method ReD (Return-based Data Rebalance), which can be implemented with less than 10 lines of code change and adds negligible running time. Extensive experiments demonstrate that ReD is effective at boosting offline RL performance and orthogonal to decoupling strategies in long-tailed classification. New state-of-the-arts are achieved on the D4RL benchmark. | ['Shuicheng Yan', 'Gao Huang', 'Zhongwen Xu', 'Xiao Ma', 'Bingyi Kang', 'Yang Yue'] | 2022-10-17 | null | null | null | null | ['d4rl'] | ['robots'] | [-7.51574263e-02 9.89802480e-02 -7.46959686e-01 -4.35538799e-01
-6.87854290e-01 -6.98437989e-01 4.78570938e-01 2.12614432e-01
-8.85879338e-01 9.16337430e-01 1.73328251e-01 -4.48186755e-01
-4.21312004e-02 -7.23647177e-01 -8.96744490e-01 -8.27246666e-01
2.44108066e-02 5.01045346e-01 3.13069493e-01 -5.21470197e-02
1.31502211e-01 4.41735774e-01 -1.51319063e+00 6.61000013e-02
7.71506786e-01 9.61293638e-01 1.17151178e-01 4.09242898e-01
-2.16976672e-01 9.87790763e-01 -6.12137318e-01 -1.67024732e-01
4.33887869e-01 -6.04593456e-01 -5.61814129e-01 8.96890834e-03
3.41112494e-01 -7.03586042e-01 -3.53437036e-01 9.90089357e-01
5.24230778e-01 3.12082112e-01 4.66802180e-01 -1.18667114e+00
-4.15275812e-01 9.46131110e-01 -7.37681150e-01 1.77896574e-01
-4.58834283e-02 2.62718797e-01 1.06194282e+00 -7.46472254e-02
3.39779913e-01 1.16764402e+00 4.25276101e-01 7.18843281e-01
-1.47146618e+00 -6.17405653e-01 7.70352483e-01 1.28745213e-01
-8.41069043e-01 -3.40030372e-01 5.23229420e-01 -1.25948593e-01
6.13207817e-01 5.66694029e-02 5.12535393e-01 1.18566799e+00
-3.10930237e-02 1.02575684e+00 1.34459960e+00 -3.17110062e-01
7.06972420e-01 7.81960487e-02 1.03753351e-01 3.78022522e-01
2.46963501e-01 2.28210852e-01 -5.45584798e-01 -2.50344455e-01
4.84460384e-01 -1.83790438e-02 -1.78538144e-01 -7.20528483e-01
-8.81600201e-01 8.46012473e-01 2.85429984e-01 8.54607522e-02
-3.94967347e-01 4.15142655e-01 5.60278237e-01 4.27452415e-01
4.15762633e-01 3.40768963e-01 -6.15709960e-01 -5.43464780e-01
-6.49314106e-01 5.46922207e-01 6.49257421e-01 6.16259396e-01
8.78260255e-01 2.43751053e-02 -5.85555315e-01 7.65672624e-01
-7.92490169e-02 4.48641419e-01 7.01261580e-01 -9.14438128e-01
3.90977532e-01 2.67035365e-01 2.85079718e-01 -5.00588417e-01
-3.14215750e-01 -6.04012549e-01 -5.39758325e-01 2.04433277e-01
8.06686342e-01 -3.38052243e-01 -6.60290241e-01 2.15270233e+00
4.49412912e-01 1.03464544e-01 -1.50995970e-01 7.72874892e-01
-1.66089535e-01 3.77713561e-01 1.12941891e-01 -2.98435926e-01
1.00974786e+00 -8.57932985e-01 -5.68411767e-01 -3.15127790e-01
6.56855106e-01 -2.35973746e-01 1.48948169e+00 4.67408985e-01
-8.12468112e-01 -3.40364277e-01 -9.08004522e-01 2.08282664e-01
5.83725423e-03 5.09316996e-02 5.17869294e-01 5.85786462e-01
-6.87060297e-01 8.53308141e-01 -1.08485794e+00 -1.52675886e-04
5.09290695e-01 2.40468428e-01 -2.06866506e-02 1.04458719e-01
-9.46582258e-01 7.95396447e-01 4.50095087e-01 -2.64084667e-01
-8.78304243e-01 -9.51781452e-01 -5.22830009e-01 1.72729194e-01
8.55160534e-01 -3.60788703e-01 1.63531232e+00 -1.25838196e+00
-2.05838776e+00 4.40281421e-01 9.47763175e-02 -9.25545871e-01
8.98896992e-01 -4.05144781e-01 1.92463342e-02 -1.39515549e-01
-1.70587912e-01 4.29960936e-01 1.00695586e+00 -1.08134246e+00
-9.10049617e-01 -4.16546226e-01 1.58979312e-01 2.79433668e-01
-6.63120389e-01 -3.42892557e-01 -3.60594332e-01 -6.02480114e-01
-4.34137881e-01 -1.04163563e+00 -1.98318094e-01 -1.94878250e-01
-1.44527769e-02 -2.66773134e-01 6.67713404e-01 -3.86695594e-01
1.35353076e+00 -2.16102481e+00 -1.68945193e-02 8.38394165e-02
-4.01957445e-02 2.23609924e-01 -1.71864599e-01 3.01391721e-01
1.61279306e-01 -2.40883678e-01 -2.79644102e-01 -1.91141337e-01
2.55352139e-01 5.24383068e-01 -7.32647002e-01 7.30031133e-01
-1.80170372e-01 5.90864003e-01 -1.04632604e+00 3.82821709e-02
-1.81040149e-02 -2.19264589e-02 -9.99017537e-01 3.32357407e-01
-6.41107857e-01 5.60568392e-01 -3.64509553e-01 1.25591189e-01
6.98807299e-01 -5.39145917e-02 5.90568781e-01 1.17469326e-01
-2.51815412e-02 3.92152846e-01 -1.11892521e+00 1.54210186e+00
-6.00607157e-01 9.45571885e-02 -1.23945847e-01 -9.94561076e-01
9.00691628e-01 -2.33571693e-01 5.03987014e-01 -9.47276771e-01
-6.97273463e-02 5.72041944e-02 7.19513297e-02 -2.24466026e-01
3.49912226e-01 -2.33133122e-01 -2.94358246e-02 6.91045582e-01
-2.12255359e-01 1.87018886e-01 1.57229409e-01 -1.52607203e-01
1.03767288e+00 5.35055101e-01 3.23510170e-01 -2.67809451e-01
2.46055841e-01 -3.59692365e-01 6.79238915e-01 9.54128742e-01
-3.29211622e-01 -3.93364951e-02 9.19788957e-01 -2.12819308e-01
-7.94547915e-01 -9.78855491e-01 9.28243473e-02 1.63047421e+00
3.28262597e-02 -2.38642320e-01 -7.12512255e-01 -1.32118654e+00
4.88289833e-01 1.03833222e+00 -6.50319636e-01 -4.23913062e-01
-7.43801117e-01 -6.80792689e-01 5.52955866e-01 5.64513028e-01
4.92500782e-01 -8.80449533e-01 -9.15543795e-01 1.99952170e-01
2.88337786e-02 -9.13695872e-01 -6.44821107e-01 4.08890486e-01
-7.85738587e-01 -9.58673358e-01 -6.78574443e-01 -2.70419180e-01
5.37511110e-01 8.88966247e-02 9.31491077e-01 -2.23703533e-01
2.39429757e-01 3.15678120e-01 -3.78733546e-01 -2.32111037e-01
-4.35434699e-01 3.34909260e-01 1.62268803e-01 6.64164647e-02
1.66555598e-01 -5.59232056e-01 -5.44632077e-01 2.68370062e-01
-9.53644872e-01 -2.49466822e-01 4.66373742e-01 9.06646490e-01
5.89690030e-01 6.74410863e-03 8.21625292e-01 -9.20578063e-01
6.22727692e-01 -4.58480388e-01 -8.51812005e-01 2.88576543e-01
-7.72622049e-01 6.40929103e-01 1.18231642e+00 -7.36750603e-01
-1.07426608e+00 4.30963859e-02 -1.38233110e-01 -4.75806087e-01
-4.18032408e-02 1.01963744e-01 -5.64149432e-02 3.47558618e-01
5.15144467e-01 3.06798995e-01 2.76655406e-01 -5.80416203e-01
6.17291868e-01 6.01655841e-01 4.68611598e-01 -1.06027794e+00
5.48560083e-01 5.14353693e-01 -1.47250652e-01 -4.86645132e-01
-1.06482661e+00 -1.39068037e-01 -3.09978575e-01 -1.46391675e-01
3.46048176e-01 -7.10145533e-01 -1.02887487e+00 3.64092529e-01
-4.09516454e-01 -9.97578681e-01 -6.60222769e-01 4.23522949e-01
-8.29651296e-01 3.31167728e-01 -3.63070369e-01 -8.76244307e-01
-9.59101990e-02 -9.86475289e-01 7.10926771e-01 7.24620596e-02
5.08941263e-02 -8.16443980e-01 3.64997387e-01 -9.93637592e-02
4.65202957e-01 -7.80714974e-02 9.82716024e-01 -7.12229192e-01
-7.47904256e-02 2.24921539e-01 -3.59039046e-02 5.72518408e-01
2.42678642e-01 -3.26072395e-01 -8.67155135e-01 -7.06577420e-01
-5.88354953e-02 -6.01391435e-01 1.01424217e+00 2.13569015e-01
1.59643769e+00 -4.64100182e-01 5.94322383e-03 4.94318396e-01
1.27994680e+00 2.06550166e-01 4.36943352e-01 5.16966581e-01
4.75912869e-01 5.48061073e-01 8.37994754e-01 9.52053070e-01
2.75470048e-01 8.23845506e-01 4.48782623e-01 1.82303667e-01
4.69398238e-02 -7.01287866e-01 7.51817584e-01 2.41571322e-01
3.40124726e-01 -9.24723372e-02 -5.56000173e-01 3.03283155e-01
-1.99944890e+00 -9.85545576e-01 5.53743541e-01 2.59234238e+00
1.24877262e+00 1.86451271e-01 6.30317628e-01 -7.37173334e-02
3.86898458e-01 2.13944182e-01 -1.07155252e+00 -3.43680471e-01
2.48633116e-01 2.62389332e-01 8.74696791e-01 4.57739383e-01
-9.19546485e-01 9.78245020e-01 5.88677740e+00 1.16220522e+00
-1.33574808e+00 4.26894799e-02 6.34945273e-01 -4.52341229e-01
-3.60880792e-01 4.81463037e-02 -9.36334431e-01 7.73110688e-01
9.83526647e-01 -2.39092603e-01 8.06597114e-01 1.04275489e+00
2.36051857e-01 -1.99215561e-01 -1.23353267e+00 8.10118496e-01
-1.36124492e-01 -1.02515090e+00 -3.60789075e-02 8.06953311e-02
5.49425840e-01 -5.94910458e-02 2.02385008e-01 8.11934054e-01
5.07458806e-01 -6.42943382e-01 9.62565958e-01 3.52426618e-01
6.07708037e-01 -1.12054217e+00 3.58086705e-01 6.34756625e-01
-7.35174358e-01 -3.83187294e-01 -4.45123523e-01 -1.98621787e-02
-2.36753315e-01 5.09698212e-01 -7.34241843e-01 3.81215960e-01
6.17854714e-01 6.23274982e-01 -4.30658698e-01 7.32469797e-01
-2.35888138e-01 6.73629105e-01 -2.67026395e-01 -1.11807950e-01
2.32164279e-01 -1.79386675e-01 2.32754573e-01 1.01293600e+00
1.28975555e-01 -3.43747437e-01 5.06798863e-01 6.13411963e-01
-1.93310082e-01 3.88648584e-02 -3.98290992e-01 -1.70874894e-02
4.82397616e-01 9.19084370e-01 -4.41903710e-01 -2.93832272e-01
-1.90027520e-01 8.30142438e-01 8.77775133e-01 1.70659378e-01
-1.12898695e+00 -1.53890764e-02 7.97447503e-01 8.86299089e-02
6.29600704e-01 -1.86282352e-01 -2.06067134e-02 -1.10206366e+00
-2.37943307e-02 -1.23903310e+00 5.63520968e-01 -1.52641693e-02
-1.41494751e+00 2.05921829e-01 1.07056715e-01 -1.11890686e+00
-2.50701278e-01 -3.33976746e-01 -1.28441900e-01 4.27436173e-01
-1.77172971e+00 -4.87264484e-01 2.70059071e-02 5.31029522e-01
4.07477409e-01 6.72968896e-03 4.60552305e-01 2.12993056e-01
-6.21117115e-01 9.06057060e-01 4.03153658e-01 -3.08970660e-01
9.55069363e-01 -1.41737115e+00 -5.16639240e-02 5.31931639e-01
9.81755089e-03 3.93372208e-01 8.17204952e-01 -4.32889760e-01
-1.51663566e+00 -1.09979880e+00 1.48646906e-01 2.25014463e-02
7.20792890e-01 -2.75193393e-01 -9.15226579e-01 6.01350963e-01
5.37866578e-02 2.16885526e-02 4.83721823e-01 1.33712575e-01
-4.94764209e-01 -5.12450337e-01 -1.10460985e+00 7.50743985e-01
8.47737849e-01 -2.14643985e-01 -3.01976413e-01 2.39865318e-01
5.26719391e-01 -4.80503380e-01 -6.33928835e-01 3.23983133e-01
5.45051038e-01 -1.01731300e+00 6.55706763e-01 -8.17414284e-01
1.47976980e-01 -1.93777129e-01 -1.78923428e-01 -1.45046711e+00
-6.87195957e-02 -7.37960637e-01 -5.04011512e-01 1.03110373e+00
5.95781766e-02 -8.37141454e-01 8.20853055e-01 2.91287005e-01
1.60543621e-01 -1.02330124e+00 -8.62441361e-01 -1.15487146e+00
2.93514609e-01 -3.01376015e-01 6.94440722e-01 6.82905495e-01
-5.43222018e-02 8.64881352e-02 -5.98844051e-01 -2.57756636e-02
4.50537622e-01 3.24053705e-01 9.72456098e-01 -7.54519641e-01
-9.58413661e-01 -5.32654524e-01 1.93435177e-01 -1.63165259e+00
5.26872516e-01 -8.34761798e-01 2.02723950e-01 -9.72348034e-01
1.74738854e-01 -6.87163234e-01 -5.01289666e-01 7.95430303e-01
-1.11829929e-01 -1.95409626e-01 3.22514474e-01 8.62866789e-02
-6.68317258e-01 9.90657449e-01 1.12764549e+00 1.93378106e-01
-4.02962863e-01 1.51868671e-01 -7.20801532e-01 5.37299335e-01
9.32747960e-01 -5.58229744e-01 -5.93916118e-01 -1.96487576e-01
2.06861258e-01 3.60267051e-03 8.37071463e-02 -7.85773873e-01
-9.78862196e-02 -4.41585809e-01 6.97529539e-02 -3.79554510e-01
-5.92673011e-02 -7.78826892e-01 -3.00940871e-01 6.93647802e-01
-7.64982343e-01 -2.02383194e-02 2.98810869e-01 7.91700661e-01
1.43025547e-01 -2.40556866e-01 1.10356021e+00 9.29532200e-02
-3.63929749e-01 3.04357111e-01 -2.71197647e-01 3.37751508e-01
1.13347709e+00 3.12183291e-01 -4.45100754e-01 -3.58675063e-01
-3.19475263e-01 3.61587048e-01 4.79839236e-01 3.78167987e-01
1.88675523e-01 -1.09695721e+00 -4.49479252e-01 1.49722844e-01
-3.00895665e-02 -2.27683559e-01 1.78383797e-01 7.43728220e-01
-1.52793229e-01 2.44950846e-01 -2.33404338e-01 -3.36212397e-01
-1.03456759e+00 6.71034217e-01 4.64124888e-01 -5.95734835e-01
-7.68454313e-01 4.57058638e-01 7.61040300e-02 -4.60330933e-01
6.09046817e-01 -4.75233376e-01 -3.31343524e-02 1.35604635e-01
5.21450520e-01 4.02705640e-01 -3.97159085e-02 1.02832988e-01
-2.75766432e-01 1.39186725e-01 -3.77117723e-01 -1.44593596e-01
1.24003911e+00 8.38498473e-02 2.76852369e-01 5.57242513e-01
9.29193079e-01 2.51701295e-01 -1.82290292e+00 -3.18930775e-01
4.99258041e-02 -5.71256697e-01 7.64534809e-03 -9.33790207e-01
-1.00985730e+00 4.65870410e-01 5.85070312e-01 2.40532160e-01
1.10014832e+00 -1.80120930e-01 7.00014710e-01 4.90023315e-01
4.56194192e-01 -1.34109902e+00 2.67616123e-01 4.94779974e-01
6.35827601e-01 -9.77892458e-01 1.56139135e-01 3.06074142e-01
-9.15858388e-01 9.68079984e-01 7.57061064e-01 -2.98717588e-01
3.40713978e-01 2.39506170e-01 -1.86608925e-01 3.33471090e-01
-8.69923592e-01 -2.58997560e-01 -1.67324960e-01 3.26693892e-01
1.85211197e-01 1.66991338e-01 -3.90220165e-01 4.21011746e-01
-9.10120234e-02 2.13604630e-03 3.67966741e-01 9.31165040e-01
-4.54449058e-01 -1.41041505e+00 -1.88064337e-01 4.08735484e-01
-4.04050380e-01 1.76821008e-01 -1.29223550e-02 8.19580615e-01
-8.73582736e-02 6.54764354e-01 2.98296422e-01 -1.38493747e-01
2.85389036e-01 -3.79355028e-02 6.61837280e-01 -3.21720600e-01
-5.17277837e-01 3.05585787e-02 -6.24719402e-03 -7.07582057e-01
-1.39208019e-01 -6.98238134e-01 -1.31409895e+00 -3.37256491e-01
-1.20500706e-01 9.93407518e-02 4.10070628e-01 9.12943900e-01
4.16084439e-01 5.99272907e-01 8.73800099e-01 -5.06032348e-01
-1.59424245e+00 -7.34761000e-01 -6.31526411e-01 3.11336458e-01
3.87911648e-01 -7.94939697e-01 -3.51010621e-01 -4.53704566e-01] | [4.069770812988281, 2.2325141429901123] |
3070f70c-0921-4f87-b006-1536110fdc03 | coherent-parametric-contours-for-interactive | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Lu_Coherent_Parametric_Contours_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Lu_Coherent_Parametric_Contours_CVPR_2016_paper.pdf | Coherent Parametric Contours for Interactive Video Object Segmentation | Interactive video segmentation systems aim at producing sub-pixel-level object boundaries for visual effect applications. Recent approaches mainly focus on using sparse user input (i.e. scribbles) for efficient segmentation; however, the quality of the final object boundaries is not satisfactory for the following reasons: (1) the boundary on each frame is often not accurate; (2) boundaries across adjacent frames wiggle around inconsistently, causing temporal flickering; and (3) there is a lack of direct user control for fine tuning. We propose Coherent Parametric Contours, a novel video segmentation propagation framework that addresses all the above issues. Our approach directly models the object boundary using a set of parametric curves, providing direct user controls for manual adjustment. A spatio-temporal optimization algorithm is employed to produce object boundaries that are spatially accurate and temporally stable. We show that existing evaluation datasets are limited and demonstrate a new set to cover the common cases in professional rotoscoping. A new metric for evaluating temporal consistency is proposed. Results show that our approach generates higher quality, more coherent segmentation results than previous methods. | ['Jue Wang', 'Linda Shapiro', 'Yao Lu', 'Xue Bai'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['interactive-video-object-segmentation'] | ['computer-vision'] | [ 2.30780914e-01 -3.27089339e-01 -1.72455803e-01 -2.35359326e-01
-6.24715686e-01 -7.40496039e-01 1.81926087e-01 1.88442260e-01
-4.37432140e-01 6.12222016e-01 -1.44252300e-01 -9.78899151e-02
-4.08619130e-03 -3.27376455e-01 -6.76896751e-01 -4.75910723e-01
5.93603067e-02 1.61245003e-01 1.26661491e+00 -1.57905340e-01
4.01048273e-01 5.02216160e-01 -1.54803586e+00 1.39973685e-01
1.31848633e+00 6.64977908e-01 3.48414153e-01 1.01251936e+00
-1.26241595e-01 3.40868622e-01 -8.34218562e-01 -6.57793181e-03
3.40645581e-01 -7.60798335e-01 -9.12018776e-01 4.40909863e-01
5.81099749e-01 -4.46185708e-01 1.20996289e-01 1.16152513e+00
1.33772865e-01 4.72909570e-01 6.19338214e-01 -1.19736779e+00
-1.97712943e-01 2.19893649e-01 -9.37467992e-01 3.10532659e-01
4.14102674e-01 3.61469865e-01 5.55295467e-01 -3.05658191e-01
9.01491463e-01 1.11434090e+00 6.56032622e-01 4.12087560e-01
-1.38056850e+00 -3.37287039e-01 5.30458689e-01 -2.18099542e-02
-1.37604320e+00 -1.08725801e-01 5.02310455e-01 -7.36597657e-01
5.43520391e-01 5.52185059e-01 9.82554972e-01 6.09966278e-01
4.84799564e-01 6.61380529e-01 1.01346207e+00 -3.88952136e-01
1.81402013e-01 1.27043918e-01 1.83060423e-01 5.86063087e-01
-2.14390475e-02 7.61907101e-02 -3.71256769e-01 8.51750970e-02
1.26549888e+00 -2.44278073e-01 -6.54304326e-01 -4.27979320e-01
-1.25140905e+00 4.23785359e-01 3.81193548e-01 4.58869338e-01
-2.13060364e-01 2.57903129e-01 2.29645088e-01 3.07432236e-03
2.68025786e-01 4.56730068e-01 -1.25126049e-01 -3.07371467e-01
-1.45404458e+00 4.97952700e-01 6.88505769e-01 1.08703673e+00
6.05190754e-01 -6.07307591e-02 -4.73910213e-01 7.21920609e-01
2.81584144e-01 9.12197828e-02 3.86275083e-01 -1.51299179e+00
1.59908924e-02 2.64516979e-01 5.72856188e-01 -1.08493638e+00
-3.23292583e-01 -5.40677495e-02 -3.10084760e-01 7.63987541e-01
8.93947423e-01 2.65740305e-02 -1.25868523e+00 1.38168240e+00
5.80872297e-01 2.19969943e-01 -5.28812945e-01 1.27311957e+00
7.38236666e-01 7.66429663e-01 2.06400231e-01 -2.79069930e-01
1.20546925e+00 -1.22939026e+00 -1.21602273e+00 -9.19735581e-02
1.81621537e-01 -9.69156325e-01 1.28656900e+00 6.05563343e-01
-1.50879204e+00 -7.14460254e-01 -1.09704566e+00 -6.26992956e-02
-2.53754348e-01 -1.01543859e-01 3.74817312e-01 6.44096732e-01
-1.21405065e+00 6.74791574e-01 -9.58256662e-01 -2.04308733e-01
1.43331394e-01 3.13517720e-01 6.07935712e-02 1.63433224e-01
-7.69209146e-01 6.40697479e-01 4.17609960e-01 5.93467383e-03
-4.85461622e-01 -8.27385247e-01 -8.23923111e-01 -3.64197433e-01
4.70928371e-01 -6.71599209e-01 1.43780935e+00 -1.36452150e+00
-1.70682526e+00 6.85112715e-01 -1.50506631e-01 -1.86195031e-01
9.41488147e-01 -2.75681853e-01 2.78864335e-02 3.71521622e-01
-3.69336223e-03 1.09639645e+00 9.14904654e-01 -1.63234150e+00
-8.37324262e-01 3.38927358e-02 -5.81402890e-02 4.20983195e-01
4.10329029e-02 2.34507415e-02 -1.17869377e+00 -8.83625269e-01
1.44481584e-01 -9.53727961e-01 -4.82970923e-01 1.73594475e-01
-3.88527215e-01 9.59675945e-03 1.11945832e+00 -8.26901376e-01
1.84126174e+00 -1.99878705e+00 9.93056297e-02 2.24995241e-01
2.04243794e-01 3.53161335e-01 2.18956575e-01 -2.36760840e-01
1.36068717e-01 2.34328881e-01 -5.11560440e-01 -2.16164321e-01
-4.23657864e-01 9.49112698e-02 -3.72080654e-02 3.85839641e-01
-2.44886443e-01 7.67699182e-01 -1.00849855e+00 -9.08268094e-01
5.50767481e-01 3.05941910e-01 -5.83573520e-01 1.40898317e-01
-4.05082554e-01 6.79762423e-01 -2.93323994e-01 5.93837142e-01
5.95740080e-01 -8.22048411e-02 -2.48753235e-01 -1.91162005e-01
-4.27960485e-01 -1.27342835e-01 -1.51069582e+00 1.84632289e+00
-3.61593813e-02 8.01943958e-01 1.41451046e-01 -2.70328492e-01
4.91507471e-01 2.17455834e-01 6.03562713e-01 -4.04071659e-01
1.77350771e-02 1.47182345e-01 -8.42737854e-02 -6.24065220e-01
8.65119755e-01 2.71899998e-01 2.12541759e-01 2.39935935e-01
-4.27425206e-01 -7.44448543e-01 5.57291687e-01 1.49734572e-01
6.67183757e-01 6.35529816e-01 1.06286280e-01 -3.60279202e-01
1.91937611e-01 3.64941806e-01 7.18890548e-01 7.20501363e-01
-5.80237746e-01 1.28024638e+00 4.58039165e-01 -1.38952315e-01
-9.61430073e-01 -9.47651744e-01 -1.82078972e-01 6.79878950e-01
8.77581000e-01 -2.15239510e-01 -1.10908377e+00 -4.45043206e-01
-3.05278838e-01 5.77139437e-01 -3.86747330e-01 3.25863719e-01
-7.01602101e-01 -2.03659251e-01 1.56476140e-01 5.70753813e-01
3.47935140e-01 -1.03628123e+00 -1.09210706e+00 4.66082186e-01
-3.87450904e-01 -1.06006825e+00 -1.12687397e+00 -1.99073046e-01
-1.20229244e+00 -1.01680195e+00 -1.05238223e+00 -6.94336891e-01
8.13958228e-01 3.26186985e-01 1.13577545e+00 2.53893912e-01
-4.04508889e-01 3.80687535e-01 -2.63913989e-01 -1.10431358e-01
-4.87124622e-01 -3.61714333e-01 -3.50596994e-01 -2.08359390e-01
-2.06565186e-01 -1.54196143e-01 -8.09928238e-01 6.40086055e-01
-1.11388826e+00 3.73495191e-01 3.80551852e-02 5.44394851e-01
9.14490283e-01 9.57500935e-02 1.38228074e-01 -8.06128204e-01
5.60654402e-01 9.62238535e-02 -8.62420976e-01 1.50959060e-01
-3.70241612e-01 -1.53792977e-01 2.24978715e-01 -6.67426527e-01
-1.19813943e+00 8.64404514e-02 1.68196112e-01 -6.08781099e-01
-3.49517345e-01 -2.81975474e-02 3.09957415e-01 -1.80618748e-01
7.38967121e-01 -2.07577154e-01 4.40602237e-03 -1.12464577e-01
3.61422062e-01 2.43328989e-01 8.13035905e-01 -4.07320946e-01
3.87104154e-01 4.66518641e-01 -4.35731560e-01 -8.68523180e-01
-3.87952775e-01 -7.07792938e-01 -7.66273260e-01 -7.58656502e-01
1.11464059e+00 -3.48264456e-01 -4.51875478e-01 4.89138514e-01
-1.26264906e+00 -7.81414270e-01 -2.80096412e-01 5.69806933e-01
-7.23693132e-01 5.99109650e-01 -8.07252407e-01 -7.41817474e-01
-6.13157079e-02 -1.55915403e+00 9.57270145e-01 6.99226856e-01
-5.79535007e-01 -9.54431951e-01 -7.44497031e-02 2.47912899e-01
1.98747560e-01 5.64919353e-01 4.69550461e-01 1.89674243e-01
-8.16285908e-01 1.71618313e-01 -8.51926580e-02 1.45906180e-01
1.75796047e-01 7.08009481e-01 -6.41291678e-01 -2.79471674e-03
-3.30633730e-01 1.02668576e-01 5.34471035e-01 1.08252478e+00
1.30706811e+00 -2.91706529e-02 -5.58677316e-01 5.60476422e-01
1.17537904e+00 5.44167161e-01 7.55992830e-01 2.88344383e-01
6.16698384e-01 7.68543243e-01 1.07344186e+00 9.24789980e-02
1.07109435e-01 9.80218053e-01 6.69542179e-02 -4.53105420e-01
-4.11066383e-01 -3.84623855e-02 8.82158354e-02 2.79360592e-01
-2.26512447e-01 -3.47437918e-01 -8.45573545e-01 8.08039188e-01
-1.97695637e+00 -6.65580809e-01 -6.17076457e-01 2.27216101e+00
9.92010415e-01 2.31864750e-01 3.26627374e-01 1.53072085e-02
7.91232169e-01 -1.56375900e-01 -3.69620711e-01 -3.33543092e-01
2.60966301e-01 -7.45938346e-02 5.37852108e-01 7.96540260e-01
-1.14035475e+00 1.01641691e+00 6.56013823e+00 7.76876330e-01
-1.09025848e+00 -2.33653579e-02 8.09655786e-01 -1.55598512e-02
-3.30761284e-01 -1.01389429e-02 -6.66618168e-01 4.85090852e-01
2.56562054e-01 -5.95831871e-03 1.23102956e-01 4.36701715e-01
6.39854491e-01 -7.51049757e-01 -9.05647814e-01 1.04114544e+00
-1.38172120e-01 -1.16099167e+00 -3.14080119e-01 -1.76733196e-01
1.08714139e+00 -4.90178406e-01 8.63008723e-02 -3.55667830e-01
2.36759141e-01 -8.53613675e-01 1.13227475e+00 6.26104653e-01
7.25237668e-01 -6.54677570e-01 1.73510656e-01 9.39269513e-02
-1.08076525e+00 4.19453055e-01 7.43307024e-02 4.46420372e-01
5.46432972e-01 4.66170132e-01 -7.06903636e-01 2.25955769e-01
9.54777658e-01 4.04211283e-01 -4.29973006e-01 1.61995471e+00
-1.99600473e-01 4.39664304e-01 -4.57467526e-01 1.10674620e-01
3.66737664e-01 -4.05698776e-01 7.28188932e-01 1.50869524e+00
1.86610594e-01 2.46888846e-01 1.99363992e-01 1.13362193e+00
4.86469895e-01 1.54498860e-01 -3.73323530e-01 2.55936414e-01
2.88558125e-01 1.11095130e+00 -1.41544473e+00 -3.25630814e-01
-2.43283302e-01 1.28710747e+00 -2.19660193e-01 7.49170065e-01
-9.52533722e-01 -3.90644133e-01 4.15927798e-01 3.74317020e-01
1.58821076e-01 -5.22788346e-01 -5.63891113e-01 -6.74096406e-01
-1.25803187e-01 -7.34986365e-01 3.61614764e-01 -9.36026990e-01
-7.05859005e-01 4.70481962e-01 2.47967139e-01 -1.29263389e+00
-3.69174182e-01 -1.54019624e-01 -5.24841905e-01 7.12721467e-01
-1.22288001e+00 -6.24433815e-01 -5.70526898e-01 3.86438698e-01
9.31576014e-01 5.94359398e-01 2.25433797e-01 3.02026421e-01
-3.49200368e-01 3.33607316e-01 -2.94745058e-01 -2.76653856e-01
8.28985453e-01 -1.44085991e+00 1.50785759e-01 1.14533401e+00
6.78307237e-03 4.17799741e-01 9.58148599e-01 -8.35452676e-01
-6.61993682e-01 -7.43401945e-01 4.37318265e-01 -3.96299183e-01
1.64817914e-01 -6.19228333e-02 -1.24218976e+00 4.73815501e-01
2.35611320e-01 -5.61704934e-02 2.45276138e-01 -4.52347010e-01
3.60335469e-01 1.87707692e-01 -1.15986073e+00 9.01488960e-01
7.53589571e-01 1.61206722e-01 -2.10642308e-01 5.77193834e-02
7.91028202e-01 -1.09412205e+00 -6.63674891e-01 2.46372238e-01
5.27330816e-01 -1.29227519e+00 9.16984439e-01 -6.41420111e-02
2.90737927e-01 -7.75392294e-01 5.83236635e-01 -1.13874066e+00
-6.55535012e-02 -8.44398320e-01 4.76684142e-03 8.92055869e-01
2.91970313e-01 -1.85683787e-01 6.54646635e-01 1.01820230e+00
-2.53423423e-01 -6.44840360e-01 -6.86178207e-01 -5.77276051e-01
-4.25410062e-01 -4.42246646e-01 8.60835686e-02 8.69561851e-01
-2.15648457e-01 -2.80452162e-01 -5.91822788e-02 1.50006711e-01
5.66829979e-01 1.01565808e-01 6.04376197e-01 -7.64919341e-01
-2.55447596e-01 -7.94407248e-01 -1.81640625e-01 -1.46504319e+00
-3.48774552e-01 -1.25081852e-01 5.13557434e-01 -1.60615420e+00
-1.04163669e-01 -7.56619811e-01 2.22435951e-01 2.13658392e-01
-4.07006472e-01 4.10077095e-01 1.74228445e-01 8.55720863e-02
-4.79990512e-01 1.68637373e-02 1.82448471e+00 1.29051432e-01
-8.46696377e-01 1.32478237e-01 -1.06170312e-01 8.37407172e-01
4.57433164e-01 -1.55899957e-01 -5.52041650e-01 -2.26834565e-01
-2.59282827e-01 2.49560535e-01 3.22890520e-01 -1.03961825e+00
3.58001053e-01 -4.55896080e-01 3.15059721e-01 -7.51517177e-01
1.43263161e-01 -7.81418741e-01 3.42550635e-01 5.11388004e-01
-1.68800324e-01 -1.20954446e-01 4.55580086e-01 5.05879819e-01
-2.84811974e-01 -4.94377643e-01 1.17595494e+00 -9.40666199e-02
-8.75359714e-01 2.06772149e-01 -4.55012351e-01 -8.24127626e-03
1.33699322e+00 -7.80588686e-01 2.94658273e-01 -4.53138769e-01
-8.64942610e-01 4.38695550e-01 7.46693909e-01 4.31953937e-01
6.36806548e-01 -1.02604091e+00 -4.25855815e-01 8.57852772e-02
-2.31226578e-01 3.37166697e-01 2.94340521e-01 9.30122197e-01
-1.06565225e+00 5.75920716e-02 6.35866672e-02 -1.01393461e+00
-1.49924958e+00 2.61122495e-01 6.74071848e-01 1.26959290e-03
-9.41755414e-01 1.00301015e+00 3.86496007e-01 2.35386103e-01
3.90781313e-01 -7.26334393e-01 -1.56913206e-01 -2.21866299e-03
3.80431980e-01 4.18862611e-01 -9.44016948e-02 -4.18499768e-01
-5.20643033e-02 7.73281038e-01 1.00335754e-01 -4.89439547e-01
7.56183863e-01 -3.97128284e-01 1.87243611e-01 5.08599937e-01
6.68218613e-01 -2.60034017e-02 -1.81572771e+00 7.51575157e-02
-2.11604401e-01 -8.31543267e-01 1.20168678e-01 -6.63351357e-01
-9.07536030e-01 6.23654783e-01 6.69399679e-01 2.89925963e-01
1.17537928e+00 -1.66035816e-01 9.01143372e-01 -4.34448749e-01
2.15225011e-01 -1.44724989e+00 2.29546383e-01 1.98138073e-01
8.42985868e-01 -9.75032628e-01 -4.15496901e-02 -8.65960181e-01
-7.91208148e-01 1.15148818e+00 8.02682757e-01 1.63412705e-01
2.82008171e-01 4.78277296e-01 2.39119157e-01 1.41437864e-02
-1.93037897e-01 -1.73199266e-01 5.86019516e-01 5.57428062e-01
5.33927560e-01 -2.17833370e-01 -6.49099886e-01 1.02825515e-01
1.76449753e-02 3.71293463e-02 6.04656696e-01 9.73259449e-01
-4.78532255e-01 -9.16977525e-01 -7.47630656e-01 1.53723046e-01
-5.29054821e-01 2.79633611e-01 2.05149800e-02 8.39951158e-01
4.47818451e-02 9.39009726e-01 1.92482963e-01 1.50009051e-01
4.39286798e-01 -2.68585533e-01 6.16322160e-01 -4.75271106e-01
-6.45240843e-01 5.66857636e-01 -1.15536444e-01 -9.03806269e-01
-4.76778865e-01 -5.57194173e-01 -1.65259981e+00 -1.64687466e-02
-3.12884569e-01 -3.26154977e-02 4.80365664e-01 5.34050822e-01
5.04059065e-03 8.46219897e-01 2.13890269e-01 -1.06725800e+00
1.05080895e-01 -5.88109553e-01 -4.70614612e-01 6.32076085e-01
3.19886655e-01 -5.80402851e-01 -1.52900293e-01 7.03218997e-01] | [9.28069019317627, -0.4160038232803345] |
6a04ac50-88fd-41da-8c78-9fbc2ecdd90a | promptmix-text-to-image-diffusion-models | 2301.12914 | null | https://arxiv.org/abs/2301.12914v2 | https://arxiv.org/pdf/2301.12914v2.pdf | PromptMix: Text-to-image diffusion models enhance the performance of lightweight networks | Many deep learning tasks require annotations that are too time consuming for human operators, resulting in small dataset sizes. This is especially true for dense regression problems such as crowd counting which requires the location of every person in the image to be annotated. Techniques such as data augmentation and synthetic data generation based on simulations can help in such cases. In this paper, we introduce PromptMix, a method for artificially boosting the size of existing datasets, that can be used to improve the performance of lightweight networks. First, synthetic images are generated in an end-to-end data-driven manner, where text prompts are extracted from existing datasets via an image captioning deep network, and subsequently introduced to text-to-image diffusion models. The generated images are then annotated using one or more high-performing deep networks, and mixed with the real dataset for training the lightweight network. By extensive experiments on five datasets and two tasks, we show that PromptMix can significantly increase the performance of lightweight networks by up to 26%. | ['Alexandros Iosifidis', 'Qi Zhang', 'Arian Bakhtiarnia'] | 2023-01-30 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 3.21972728e-01 3.46908748e-01 3.23188931e-01 -4.83928293e-01
-3.88475835e-01 -4.43006516e-01 7.95836926e-01 8.39508623e-02
-9.04315412e-01 7.87930131e-01 1.03557870e-01 -2.16921851e-01
5.99927247e-01 -7.14890718e-01 -7.86210299e-01 -4.65208828e-01
-9.06940997e-02 8.98276210e-01 5.11153638e-01 -6.53303191e-02
-2.59669218e-02 2.80523896e-01 -1.28515506e+00 2.94195741e-01
7.58077264e-01 8.23411524e-01 2.60504305e-01 9.45556343e-01
-2.52705306e-01 1.11184919e+00 -8.95331979e-01 -5.83476305e-01
3.73929262e-01 -3.97622913e-01 -5.81045330e-01 2.33495906e-01
4.51154143e-01 -8.45093966e-01 -4.40074235e-01 8.54687929e-01
6.67249203e-01 1.87438101e-01 5.12557805e-01 -1.37571406e+00
-4.78106886e-01 7.37171769e-01 -8.07400286e-01 3.06729198e-01
1.03425346e-01 5.53134918e-01 3.05895865e-01 -8.71428311e-01
6.49891675e-01 1.47424865e+00 5.73546231e-01 8.89406085e-01
-1.11285973e+00 -7.64823377e-01 2.01377824e-01 -2.90448427e-01
-9.34296429e-01 -5.05064070e-01 6.59818828e-01 -5.88439703e-01
6.42149210e-01 -1.11401014e-01 5.08498371e-01 1.42251372e+00
-5.27094126e-01 8.82635355e-01 9.87948716e-01 -2.49849007e-01
2.56824315e-01 1.86089799e-01 -1.89878326e-02 7.77042985e-01
4.88107800e-01 -6.77425116e-02 -4.30630475e-01 -1.20641172e-01
7.73285747e-01 -3.94713171e-02 -8.73089507e-02 -9.44693610e-02
-1.41351950e+00 1.04188049e+00 7.15552688e-01 -8.72756937e-04
-6.64191842e-01 2.66169190e-01 4.76447821e-01 1.71474293e-02
8.71609569e-01 4.33497816e-01 -1.34020016e-01 5.04683405e-02
-1.14903331e+00 4.65461254e-01 8.04751873e-01 7.37575591e-01
6.55185938e-01 -6.83500618e-03 -4.90326554e-01 7.45253623e-01
1.71656400e-01 4.97200161e-01 3.83939892e-01 -1.16027927e+00
8.37390065e-01 6.32011175e-01 4.06373560e-01 -8.82154226e-01
-4.91180658e-01 -2.48256534e-01 -9.54977572e-01 2.29613528e-01
7.89132535e-01 -6.35718465e-01 -1.02098715e+00 1.64747286e+00
3.54315788e-01 1.85820431e-01 6.07710071e-02 9.53027606e-01
7.05023229e-01 7.32248187e-01 1.67552680e-01 1.26731679e-01
1.19666839e+00 -1.36516929e+00 -4.85050321e-01 -5.39210975e-01
6.74323976e-01 -3.89106601e-01 1.18585563e+00 9.98912528e-02
-1.11757994e+00 -4.78558451e-01 -7.81945765e-01 7.60708153e-02
-4.03609931e-01 1.11473642e-01 5.15548050e-01 3.97219628e-01
-1.14079368e+00 2.73388892e-01 -7.81591952e-01 -2.95512825e-01
9.98626113e-01 3.42058063e-01 -1.82564855e-01 -1.95953667e-01
-1.01387930e+00 6.04538500e-01 4.83657509e-01 9.46661830e-02
-1.01729369e+00 -5.32100320e-01 -8.63315940e-01 -1.64942458e-01
1.80664644e-01 -7.37871826e-01 1.28232861e+00 -1.32854176e+00
-1.06536400e+00 7.93231368e-01 1.04052229e-02 -8.77025545e-01
1.18326747e+00 -8.99928212e-02 1.58705667e-01 3.05570930e-01
4.05529112e-01 1.39329100e+00 9.14797783e-01 -1.44422555e+00
-5.67130923e-01 -2.53410071e-01 1.03422165e-01 -4.36540358e-02
-5.11950672e-01 2.14174494e-01 -6.68934822e-01 -6.35660231e-01
-6.77944243e-01 -1.04017866e+00 -6.20873332e-01 2.07397044e-01
-5.59338093e-01 -8.74530748e-02 9.79514599e-01 -7.57915556e-01
6.16822362e-01 -1.94515061e+00 -2.49753729e-01 6.66213930e-02
5.33738673e-01 5.54009974e-01 -5.13717890e-01 -4.31718305e-03
1.96283415e-01 8.26862529e-02 -3.65758866e-01 -7.46417701e-01
-3.06351244e-01 1.82954714e-01 -1.10132836e-01 2.46455267e-01
5.05396903e-01 1.11346376e+00 -1.07561326e+00 -5.62808871e-01
4.66142371e-02 5.62301457e-01 -4.02515024e-01 3.63206744e-01
-5.41523278e-01 5.77129245e-01 -2.95011342e-01 1.79893121e-01
5.66416621e-01 -5.89783490e-01 -1.80005863e-01 1.74534366e-01
2.82208771e-01 -3.56032215e-02 -8.66202414e-01 1.41830873e+00
-5.16603291e-01 9.66267705e-01 -2.82899775e-02 -6.38841093e-01
7.36566007e-01 8.52048695e-02 3.13364118e-01 -6.93774819e-01
1.12374172e-01 8.47742148e-03 -3.77163552e-02 -5.26124895e-01
5.11236489e-01 2.45337799e-01 1.48690909e-01 8.39055955e-01
-2.11529598e-01 2.03819036e-01 6.30935967e-01 4.64349896e-01
1.27755833e+00 -1.83942348e-01 -4.22981590e-01 1.76468402e-01
3.00558567e-01 1.72611251e-01 1.39425814e-01 9.40263391e-01
-4.33089249e-02 7.50498176e-01 5.33915341e-01 -7.96044171e-01
-1.59390962e+00 -6.31793201e-01 2.99946278e-01 1.10679197e+00
-9.97294411e-02 -2.34625246e-02 -1.28012180e+00 -1.08097839e+00
-3.61556709e-02 6.06364489e-01 -7.78153062e-01 4.84246500e-02
-5.74991107e-01 -8.14163387e-01 7.92408764e-01 7.70822942e-01
9.00804996e-01 -1.35760009e+00 -7.53452599e-01 3.57510209e-01
-3.48421335e-01 -1.44874346e+00 -3.04744869e-01 -7.29429871e-02
-6.58700228e-01 -9.56480682e-01 -1.23801148e+00 -5.51662505e-01
1.22713566e+00 3.72573435e-01 1.07012403e+00 3.73571128e-01
-1.56406522e-01 5.28652370e-02 -2.70818323e-01 -5.89654744e-01
-6.51880682e-01 3.12113822e-01 1.11395738e-03 1.94653884e-01
3.32143307e-01 -1.98786497e-01 -7.98701882e-01 1.31056294e-01
-1.13750494e+00 4.05373812e-01 5.31820357e-01 7.72930741e-01
1.00454219e-01 -1.68028101e-01 6.81807697e-01 -1.10855353e+00
9.98418152e-01 -4.09913898e-01 -6.85779631e-01 -1.21208727e-01
-2.08016232e-01 1.09247111e-01 6.65587723e-01 -7.72652566e-01
-9.93629813e-01 3.49580377e-01 -7.95262083e-02 -4.79417473e-01
-1.10934034e-01 2.43618131e-01 3.14265609e-01 4.51836474e-02
1.08016074e+00 1.02501892e-01 2.19503179e-01 -1.34612277e-01
3.78548563e-01 8.29249978e-01 5.31862080e-01 -2.82180458e-01
8.90036345e-01 6.08015001e-01 -3.35953653e-01 -6.82088971e-01
-1.10566211e+00 -3.79887782e-02 -6.31148696e-01 -2.64509678e-01
9.50957656e-01 -1.02157140e+00 -4.27806169e-01 7.20011711e-01
-1.40626919e+00 -9.82784629e-01 -2.66604662e-01 2.76714653e-01
-2.80054092e-01 7.14894533e-02 -6.55671418e-01 -7.34777272e-01
-4.75415915e-01 -1.06828082e+00 1.17205739e+00 3.29422385e-01
-1.71176389e-01 -1.03147411e+00 -2.06541106e-01 6.46374285e-01
6.55134141e-01 3.45802337e-01 4.74655539e-01 -8.26100469e-01
-5.27945936e-01 -4.24047142e-01 -5.98843634e-01 4.73729759e-01
-2.01112255e-01 -1.72504693e-01 -1.08328772e+00 -2.72139132e-01
-3.85622323e-01 -7.39216685e-01 9.53514338e-01 1.62768498e-01
1.41264009e+00 -5.00177443e-01 -4.73591298e-01 3.03281546e-01
9.88515735e-01 -2.28471547e-01 4.76513773e-01 2.21784547e-01
8.89049530e-01 7.95814812e-01 3.53604525e-01 3.65352541e-01
6.12083912e-01 3.60318840e-01 4.37185526e-01 -5.44099033e-01
-3.28238994e-01 -2.94896245e-01 1.86166316e-01 3.46373767e-01
3.24123050e-03 -4.64517772e-01 -1.06985855e+00 6.66852117e-01
-1.99961102e+00 -8.78322959e-01 -1.61519125e-01 2.01670146e+00
7.93529332e-01 2.36567914e-01 4.07989055e-01 -1.02201313e-01
8.37704360e-01 7.31118917e-02 -7.17633069e-01 9.07172188e-02
8.33128765e-02 -2.39655778e-01 6.93215370e-01 4.35977906e-01
-9.96315062e-01 1.11845279e+00 5.83280611e+00 3.90606165e-01
-1.15366209e+00 2.63537109e-01 1.10465813e+00 -2.43902415e-01
-2.59984750e-02 -2.59003103e-01 -6.16138875e-01 6.98098898e-01
9.40740407e-01 4.92750518e-02 3.38343263e-01 8.55795562e-01
3.66638720e-01 -1.91934273e-01 -9.02480423e-01 1.01510584e+00
1.91502810e-01 -1.39636421e+00 2.79199053e-02 2.30890945e-01
9.44487154e-01 4.17223752e-01 -1.09680079e-01 2.20662177e-01
6.77537382e-01 -9.32464302e-01 4.50822055e-01 3.26401740e-01
6.92885578e-01 -6.25758767e-01 8.65820169e-01 6.93954706e-01
-6.65616632e-01 -4.00347486e-02 -4.09922212e-01 1.51831424e-02
2.87241757e-01 7.45397925e-01 -1.06473529e+00 -1.75063819e-01
5.80066025e-01 3.83850247e-01 -9.04968917e-01 1.01975656e+00
-1.91342413e-01 6.42428398e-01 -5.50750732e-01 -2.22140923e-01
4.13170159e-01 1.76728010e-01 8.21137354e-02 1.33219242e+00
1.45172000e-01 -2.31397510e-01 2.13516951e-01 1.04904044e+00
-6.55452788e-01 -1.89630926e-01 -9.53855336e-01 -1.87303685e-02
3.43840688e-01 1.51740158e+00 -9.18734074e-01 -7.95928478e-01
-3.92769545e-01 1.27488494e+00 4.82699513e-01 5.23322701e-01
-8.76313806e-01 -2.52002329e-01 1.96206674e-01 4.09426659e-01
8.56399089e-02 -1.27429947e-01 -4.73115072e-02 -1.06802022e+00
6.09458648e-02 -8.79659534e-01 -1.92741316e-03 -9.05125201e-01
-1.38416517e+00 7.93159187e-01 -1.10653229e-01 -7.85499871e-01
-3.99761438e-01 -4.28980291e-01 -6.71851516e-01 8.44191432e-01
-1.46211934e+00 -1.08700716e+00 -8.89565110e-01 6.56415999e-01
6.32812738e-01 -1.60017088e-01 4.23503727e-01 3.96434397e-01
-5.83124936e-01 4.24659044e-01 -3.51913542e-01 6.46989167e-01
5.76735318e-01 -1.24293935e+00 1.05570054e+00 8.73632491e-01
1.19932003e-01 1.97329611e-01 5.99802434e-01 -7.41378248e-01
-7.58715749e-01 -1.57436860e+00 6.45606101e-01 -5.74107945e-01
5.66484034e-01 -6.77217960e-01 -8.66180360e-01 5.62134206e-01
2.41018221e-01 4.22235310e-01 2.68395394e-01 -3.94042224e-01
-1.30742610e-01 1.22250855e-01 -1.05068576e+00 8.24819982e-01
1.03193676e+00 -2.48427674e-01 -1.28826231e-01 7.74878085e-01
7.23782778e-01 -4.10155088e-01 -4.27890956e-01 2.04611346e-02
8.65958408e-02 -7.71770537e-01 7.72651911e-01 -5.39797187e-01
6.54249310e-01 -1.55541852e-01 3.75731856e-01 -1.48491299e+00
1.20957069e-01 -6.08672261e-01 -1.61719531e-01 1.15516889e+00
6.62632942e-01 -5.04628003e-01 9.92169261e-01 8.29093456e-01
4.39491361e-01 -4.31357592e-01 -5.70162952e-01 -6.41352177e-01
-1.08566113e-01 -2.79186785e-01 6.11627758e-01 9.28959250e-01
-5.36622167e-01 5.68089366e-01 -3.65562886e-01 -1.76243320e-01
6.97130263e-01 -4.92267936e-01 1.22960639e+00 -1.20378602e+00
1.35298690e-03 -2.57124096e-01 -1.93910643e-01 -1.15725493e+00
2.51830131e-01 -6.06212497e-01 7.29985908e-02 -1.79840922e+00
2.87378490e-01 -6.15028441e-01 2.67814755e-01 5.39039254e-01
-4.62148994e-01 5.86707413e-01 2.90315479e-01 2.45144635e-01
-7.48369575e-01 4.22858119e-01 1.32800198e+00 -2.79822469e-01
-1.10444762e-01 -7.39430683e-03 -5.80554724e-01 8.59717190e-01
8.95829976e-01 -5.63054919e-01 -5.09425461e-01 -6.77984774e-01
3.18919599e-01 -6.26169071e-02 6.09043479e-01 -1.19502854e+00
2.11472362e-01 2.25011215e-01 6.17473960e-01 -3.63651425e-01
2.66192764e-01 -7.65988827e-01 -4.81882840e-01 4.64735150e-01
-5.04907966e-01 2.02010393e-01 8.33220482e-02 5.12241840e-01
-2.07031053e-02 -2.12845460e-01 7.48201728e-01 -3.07393700e-01
-3.13840091e-01 4.50785249e-01 -1.88715622e-01 2.85440654e-01
1.16672647e+00 -6.54840693e-02 -6.45830989e-01 -7.26210773e-01
-4.73841280e-01 4.23882663e-01 4.08732682e-01 4.69696879e-01
4.90357280e-01 -1.21695161e+00 -9.23359871e-01 7.80376047e-02
9.42761377e-02 5.76765716e-01 -5.07431861e-04 7.26866424e-01
-7.68490970e-01 3.73787768e-02 -5.94233498e-02 -6.21898353e-01
-9.93567884e-01 6.94819808e-01 3.96089256e-01 -3.42889756e-01
-6.92705035e-01 7.43964851e-01 2.19141364e-01 -4.52834040e-01
2.98657089e-01 -2.00619757e-01 -5.44163249e-02 -3.22630517e-02
9.02261317e-01 5.82329631e-02 -2.45979652e-01 -5.91627896e-01
-8.36761147e-02 8.18902776e-02 -3.58317912e-01 -4.18217897e-01
1.31340837e+00 1.34887889e-01 1.26022294e-01 5.41212335e-02
1.06558454e+00 -2.74830729e-01 -1.92692864e+00 -3.31221104e-01
-1.87975302e-01 -3.41942072e-01 -1.08010426e-01 -7.82743335e-01
-1.41345561e+00 9.87381339e-01 5.04329026e-01 3.14980388e-01
8.50583255e-01 3.21944095e-02 9.75076556e-01 5.94329298e-01
2.54654288e-01 -1.07404721e+00 5.64852715e-01 3.69670957e-01
7.40106642e-01 -1.68251908e+00 -4.05744076e-01 2.31643803e-02
-9.25156832e-01 8.45459461e-01 9.10035253e-01 -1.46678135e-01
2.56750524e-01 4.43069607e-01 2.50319630e-01 -2.11661205e-01
-7.25748122e-01 -1.14193521e-01 -2.36498475e-01 8.38784993e-01
2.22397417e-01 -1.86667502e-01 2.02856302e-01 1.00701377e-01
-1.69739768e-01 2.37669438e-01 6.15191758e-01 7.31128752e-01
-3.83022547e-01 -9.00868118e-01 -3.72645169e-01 6.70683265e-01
-4.43003476e-01 -1.35205880e-01 -6.25612497e-01 6.28438175e-01
1.69494823e-02 9.33022201e-01 2.75017112e-01 -1.02448441e-01
1.12109967e-01 -1.37228087e-01 1.65179700e-01 -5.27733624e-01
-6.04738832e-01 -3.91898423e-01 2.05183417e-01 -2.76349485e-01
-4.25936669e-01 -3.87784719e-01 -1.30461299e+00 -3.09593052e-01
-1.83297202e-01 -1.02317668e-01 7.30674744e-01 9.61553514e-01
3.83331060e-01 5.05455852e-01 4.43643838e-01 -1.18851829e+00
-4.03057694e-01 -1.30062580e+00 6.96088597e-02 7.51423180e-01
3.62447977e-01 -4.56433684e-01 -2.41872936e-01 2.54907578e-01] | [9.960013389587402, 1.1280572414398193] |
74f8f98e-7cd9-41c1-9c45-4e36c5d47ece | mace-an-efficient-model-agnostic-framework | 2205.15540 | null | https://arxiv.org/abs/2205.15540v1 | https://arxiv.org/pdf/2205.15540v1.pdf | MACE: An Efficient Model-Agnostic Framework for Counterfactual Explanation | Counterfactual explanation is an important Explainable AI technique to explain machine learning predictions. Despite being studied actively, existing optimization-based methods often assume that the underlying machine-learning model is differentiable and treat categorical attributes as continuous ones, which restricts their real-world applications when categorical attributes have many different values or the model is non-differentiable. To make counterfactual explanation suitable for real-world applications, we propose a novel framework of Model-Agnostic Counterfactual Explanation (MACE), which adopts a newly designed pipeline that can efficiently handle non-differentiable machine-learning models on a large number of feature values. in our MACE approach, we propose a novel RL-based method for finding good counterfactual examples and a gradient-less descent method for improving proximity. Experiments on public datasets validate the effectiveness with better validity, sparsity and proximity. | ['Steven C. H. Hoi', 'Caiming Xiong', 'Jia Li', 'Wenzhuo Yang'] | 2022-05-31 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 3.00345540e-01 5.76670587e-01 -9.43877637e-01 -5.66074073e-01
-6.39033079e-01 -2.15379387e-01 7.14963078e-01 -7.56208077e-02
-2.49678567e-02 1.16043568e+00 5.37618577e-01 -8.42126608e-01
-3.03317875e-01 -7.15554416e-01 -1.12263978e+00 -2.34596819e-01
-2.30139241e-01 5.66307664e-01 -5.42302251e-01 -2.35776082e-02
3.99248838e-01 -5.10822749e-04 -1.43779778e+00 4.21592087e-01
1.33610320e+00 7.54886746e-01 -2.46697024e-01 1.85946584e-01
-1.36352301e-01 7.50829160e-01 -3.46792936e-01 -7.63707101e-01
5.05061686e-01 -5.62805772e-01 -7.79672205e-01 -2.81192183e-01
2.80443698e-01 -1.82091773e-01 -4.96511348e-02 7.56141663e-01
2.27202937e-01 -3.62907499e-02 6.73476577e-01 -1.80568945e+00
-1.07916331e+00 1.03588355e+00 -4.81129795e-01 -1.08311303e-01
1.72700107e-01 3.17124814e-01 1.21934140e+00 -5.92733979e-01
2.05800042e-01 1.45886648e+00 7.55112231e-01 7.49336302e-01
-1.46556950e+00 -7.98880935e-01 6.43535674e-01 4.79813635e-01
-7.21483767e-01 -2.53456756e-02 1.09473515e+00 -1.83713455e-02
6.75137460e-01 6.48474753e-01 7.42104530e-01 1.20118928e+00
1.89596981e-01 8.90925288e-01 1.15126300e+00 -3.26045096e-01
5.88606536e-01 9.56864059e-02 -2.21550822e-01 3.75787109e-01
5.56635559e-01 4.69287902e-01 -4.93302226e-01 -5.43446958e-01
5.89449048e-01 4.63996172e-01 -2.17768610e-01 -8.13293815e-01
-1.52742827e+00 1.32131279e+00 6.09810650e-01 -2.69852728e-01
-7.48412192e-01 4.32234585e-01 3.38683039e-01 4.21673357e-01
6.10946596e-01 9.88450825e-01 -1.01068950e+00 -5.43283038e-02
-4.91590202e-01 6.47544801e-01 7.85066605e-01 6.52187824e-01
5.91100693e-01 -1.32180257e-02 -2.25476936e-01 5.47447979e-01
2.07909644e-01 3.19242686e-01 7.35252440e-01 -1.00888252e+00
5.25000632e-01 7.96276212e-01 3.25445503e-01 -7.23122776e-01
-3.28533709e-01 -5.76098502e-01 -8.15329909e-01 1.12356305e-01
4.59069014e-01 -1.89207494e-01 -5.47182500e-01 1.86411965e+00
5.04804075e-01 6.75522149e-01 1.85536265e-01 1.14922464e+00
2.89830714e-01 3.37346524e-01 2.89114386e-01 -4.47389454e-01
1.07914448e+00 -1.00558484e+00 -5.93171895e-01 -4.37569886e-01
7.86999047e-01 -1.63365707e-01 1.53494227e+00 2.85796791e-01
-6.98451519e-01 -1.00650407e-01 -8.41649711e-01 5.33485413e-02
-3.16830397e-01 -2.89647281e-01 1.61457455e+00 5.71603656e-01
-2.34428972e-01 7.95582235e-01 -3.27912390e-01 3.01066160e-01
8.34189236e-01 4.56259102e-01 -1.64986223e-01 -5.43153547e-02
-1.20771039e+00 8.18183362e-01 4.56074059e-01 -4.44331430e-02
-7.01297283e-01 -1.27013004e+00 -7.87941396e-01 3.76630545e-01
7.46742666e-01 -1.11476505e+00 1.14912271e+00 -1.35305369e+00
-1.55369079e+00 3.49878699e-01 -2.31774434e-01 -1.04262447e+00
7.04599321e-01 -2.30494842e-01 -5.15067637e-01 -5.69661498e-01
1.25590473e-01 5.58654726e-01 8.03129792e-01 -1.13386333e+00
-5.75726628e-01 -3.26494545e-01 2.10677922e-01 2.27173552e-01
-9.51569453e-02 -5.65139055e-01 5.20346522e-01 -7.60422289e-01
-2.00041514e-02 -7.88394332e-01 -6.79903209e-01 -8.92746672e-02
-5.73422134e-01 -2.63269931e-01 6.10987782e-01 -1.56692520e-01
1.03227377e+00 -1.74446714e+00 -3.31817120e-01 7.10868984e-02
2.14022979e-01 -3.29478756e-02 -2.38936350e-01 1.48156375e-01
-4.34791178e-01 3.95948261e-01 -5.22750556e-01 -1.17391117e-01
2.83114761e-01 1.43225402e-01 -8.16766679e-01 3.71219158e-01
1.73743844e-01 1.18375778e+00 -1.02490163e+00 -2.74153262e-01
2.22539023e-01 1.15991987e-01 -9.15476143e-01 2.13879481e-01
-6.17751837e-01 3.19680482e-01 -5.92176676e-01 3.28220814e-01
7.76827037e-01 -2.85901278e-01 2.38206103e-01 2.20736250e-01
-8.46024323e-03 5.95179379e-01 -1.21076500e+00 1.42901051e+00
-7.80508459e-01 3.00787568e-01 -7.08167970e-01 -1.15375042e+00
6.99410021e-01 1.73467919e-01 1.87273592e-01 -5.00072300e-01
-6.42542765e-02 3.28637391e-01 2.76207149e-01 -2.52913475e-01
1.74308896e-01 -3.82760316e-01 -7.46561438e-02 5.87441564e-01
-6.19277537e-01 1.35585785e-01 -3.29159528e-01 -1.18076168e-01
8.22694421e-01 8.11545551e-03 7.63263226e-01 -2.63750851e-01
2.96319366e-01 2.78520823e-01 9.72575843e-01 9.08874333e-01
1.58002172e-02 5.53907275e-01 6.03675485e-01 -9.73317623e-01
-9.05220211e-01 -7.64360011e-01 1.24237472e-02 8.89626086e-01
1.37957796e-01 -3.27024944e-02 -5.47189891e-01 -1.30651045e+00
4.77091551e-01 1.38475561e+00 -9.47693944e-01 -4.87221986e-01
-4.17559266e-01 -8.84533405e-01 2.24148277e-02 4.98822719e-01
4.23193127e-01 -1.05931866e+00 -6.00399792e-01 2.15406552e-01
-1.14548735e-01 -4.10931110e-01 -5.35414577e-01 -1.50407270e-01
-9.75761354e-01 -1.09168708e+00 -3.77747953e-01 -5.74454181e-02
5.31820357e-01 2.70965308e-01 1.17682374e+00 9.46942568e-02
1.10724270e-01 -1.23372257e-01 -3.03100273e-02 -9.54772234e-01
-4.81017120e-02 -1.66985929e-01 1.15550876e-01 1.64921299e-01
5.81641793e-01 -6.11255407e-01 -8.61081421e-01 -1.22588286e-02
-5.86269617e-01 2.57602036e-01 7.99295008e-01 1.15602052e+00
6.25842154e-01 -3.64278167e-01 1.09111440e+00 -1.33696902e+00
7.55880952e-01 -8.97903621e-01 -5.42027056e-01 3.18366587e-01
-1.18773592e+00 4.74527806e-01 1.07823801e+00 -7.88945317e-01
-1.06715631e+00 1.81115902e-05 2.39086464e-01 -3.83680463e-01
-2.00813770e-01 5.21810949e-01 -4.20883268e-01 4.27630723e-01
6.08677804e-01 1.18351512e-01 -4.09540758e-02 -5.13998628e-01
7.86058247e-01 4.45945263e-01 3.77034932e-01 -3.20222080e-01
8.05103183e-01 6.15804374e-01 1.10737592e-01 -1.01170674e-01
-1.09115195e+00 4.37017046e-02 -1.83532044e-01 3.67835909e-01
3.81713301e-01 -7.28324413e-01 -1.17667449e+00 -2.47035757e-01
-1.07611167e+00 -3.71507496e-01 -6.19567394e-01 8.45852554e-01
-7.43775010e-01 1.25202779e-02 2.54523516e-01 -9.55202639e-01
-3.70266259e-01 -7.74064422e-01 8.51453125e-01 1.21883854e-01
-4.96448696e-01 -1.01021659e+00 9.58150700e-02 2.78300852e-01
4.03666228e-01 3.94602895e-01 1.12542760e+00 -9.90611255e-01
-6.38268411e-01 1.84094962e-02 -1.85655490e-01 -2.32281208e-01
4.13445989e-03 -5.04145801e-01 -7.47096539e-01 4.37325127e-02
-2.12526038e-01 -6.59318119e-02 8.42849791e-01 6.56117737e-01
1.76668096e+00 -1.14417326e+00 -4.14961100e-01 7.24213600e-01
1.24975216e+00 -8.04691855e-03 4.20333654e-01 6.17501259e-01
5.21255434e-01 6.19231403e-01 8.97980809e-01 6.25224411e-01
6.41607642e-01 5.47980845e-01 9.35731530e-01 -3.79926234e-01
3.91957104e-01 -6.86124086e-01 1.00813977e-01 -1.93533406e-01
9.10954773e-02 9.17157829e-02 -6.06202841e-01 7.87837207e-01
-2.17508340e+00 -1.25830793e+00 -2.62734652e-01 2.25827003e+00
7.35311389e-01 2.28324626e-02 2.12674677e-01 3.45906243e-02
5.23131251e-01 1.06117111e-02 -1.13735294e+00 -7.17270851e-01
-5.50834648e-02 -4.38208319e-02 5.60269952e-01 3.09570909e-01
-8.88784647e-01 7.30272293e-01 6.34734058e+00 5.77274799e-01
-1.13147581e+00 3.36364396e-02 8.92402232e-01 -2.94724405e-01
-1.05809438e+00 3.99039388e-01 -1.74486026e-01 6.58764124e-01
8.15637529e-01 -5.90170920e-01 5.59586644e-01 1.29340208e+00
5.01158178e-01 3.15331846e-01 -1.45671618e+00 8.30420017e-01
-4.30618554e-01 -1.84416211e+00 2.55575627e-01 2.03517735e-01
7.84588516e-01 -2.57601649e-01 3.17187726e-01 3.88665289e-01
5.77497303e-01 -1.22091663e+00 7.68765748e-01 4.50211287e-01
6.20722115e-01 -9.59567785e-01 6.82560802e-01 5.47538459e-01
-5.29829860e-01 -6.01225674e-01 -5.09801626e-01 -6.46935582e-01
-6.28546551e-02 5.85016906e-01 -1.01325512e+00 4.30404276e-01
3.17471504e-01 6.32073939e-01 -2.57845372e-01 8.62269282e-01
-3.67845416e-01 8.54267716e-01 -2.51113418e-02 -1.95911452e-01
1.42117664e-01 -5.41064255e-02 5.18989861e-01 8.34633827e-01
2.06165969e-01 2.06505880e-01 -1.26821563e-01 1.27435303e+00
-1.07318394e-01 1.28113270e-01 -8.35614264e-01 2.33309448e-01
6.97676301e-01 8.38875353e-01 -2.74099618e-01 -3.38754326e-01
-3.35949391e-01 7.74089396e-01 3.16658854e-01 3.66577983e-01
-1.12468183e+00 -9.77050066e-02 1.04769349e+00 -9.75321140e-03
1.72222242e-01 3.71291459e-01 -8.96618366e-01 -1.43788159e+00
7.94005245e-02 -1.18022275e+00 6.86324298e-01 -4.63755369e-01
-1.50417316e+00 1.76440418e-01 -1.12195179e-01 -1.19437683e+00
-6.01921022e-01 -2.57656127e-01 -1.15034997e+00 6.19970381e-01
-1.70950496e+00 -1.25472736e+00 2.15825140e-01 4.16730940e-01
5.01698136e-01 -1.83160141e-01 7.43937016e-01 -3.43192130e-01
-5.10737121e-01 6.26741529e-01 1.95789486e-01 -4.04253334e-01
3.67335439e-01 -1.44544041e+00 5.34100592e-01 6.73430681e-01
3.60499769e-01 8.65202427e-01 9.73089993e-01 -4.83460009e-01
-1.55814648e+00 -1.22559690e+00 1.05833817e+00 -4.80343729e-01
3.63890231e-01 -2.72413194e-01 -8.90600801e-01 8.31795692e-01
8.77493918e-02 -1.70869753e-02 7.54714787e-01 5.72973013e-01
-3.19869131e-01 -1.98102251e-01 -1.36736035e+00 8.76938224e-01
1.18417811e+00 -1.05015472e-01 -8.43094647e-01 3.45856011e-01
9.71828938e-01 -1.79276481e-01 -5.43983936e-01 2.73434073e-01
5.62587857e-01 -8.51674736e-01 1.04914474e+00 -1.21536362e+00
6.23507559e-01 -2.27824777e-01 6.64631873e-02 -1.58748126e+00
-1.39641881e-01 -8.43353212e-01 -3.23001951e-01 9.82379317e-01
6.94972277e-01 -1.15063560e+00 8.17759037e-01 1.01199377e+00
1.25875354e-01 -1.10013342e+00 -1.05641627e+00 -8.29398572e-01
1.45070637e-02 -5.39415002e-01 1.66453004e+00 1.33001328e+00
2.61399329e-01 2.81878054e-01 -6.93424463e-01 6.59601837e-02
7.39026368e-01 8.27731788e-01 1.01856291e+00 -1.27373004e+00
-3.55124801e-01 -4.62290704e-01 -7.48200528e-03 -8.76164258e-01
5.04151165e-01 -8.02184522e-01 -3.34301800e-01 -1.24315691e+00
3.46058756e-01 -4.55178708e-01 -2.81789750e-01 6.95765018e-01
-4.71823603e-01 -2.02031329e-01 6.53810427e-02 1.23431571e-01
-3.90558928e-01 8.81230652e-01 9.46082652e-01 -1.34235686e-02
-2.42472380e-01 3.09944540e-01 -1.29903841e+00 7.46693850e-01
7.27670670e-01 -7.70478785e-01 -5.43236196e-01 -2.01481178e-01
1.35079980e-01 -1.10967420e-01 7.32266963e-01 -3.21677297e-01
-2.75083005e-01 -7.78003573e-01 3.32767308e-01 -1.87702447e-01
1.97677910e-02 -8.63440752e-01 2.75593132e-01 6.39234066e-01
-7.89305389e-01 3.44966464e-02 -6.73818886e-02 7.88378477e-01
-1.17231784e-02 1.34189293e-01 3.77121329e-01 -4.58365791e-02
-5.02130985e-01 4.81931388e-01 1.82767317e-01 5.05264588e-02
1.11754858e+00 -1.15398228e-01 -3.36532801e-01 -5.19170463e-01
-2.57600784e-01 4.05152470e-01 3.95252109e-01 4.32135046e-01
6.63344979e-01 -1.52010298e+00 -9.34758127e-01 -7.25129899e-03
2.16981009e-01 -1.89631343e-01 9.75549445e-02 6.34168088e-01
1.78673387e-01 5.97090125e-01 -6.09785393e-02 -1.45685822e-01
-7.74406672e-01 9.69585001e-01 2.98379034e-01 -4.13911313e-01
-5.99973023e-01 4.50786024e-01 5.23721814e-01 -7.15696335e-01
-1.69587657e-01 -3.81290227e-01 -1.16023093e-01 -5.06823123e-01
4.17431444e-01 4.44864988e-01 -3.81886840e-01 -4.06183377e-02
-2.32165724e-01 -9.77478027e-02 4.68229167e-02 5.20868860e-02
1.67480576e+00 -1.87083893e-02 1.91220537e-01 3.49199504e-01
9.40074086e-01 -1.30346030e-01 -1.48622346e+00 -8.21642950e-02
3.09887320e-01 -8.50561559e-01 -1.37371421e-01 -9.66719270e-01
-7.11586893e-01 7.28645682e-01 2.75768101e-01 2.95750737e-01
1.03171694e+00 -7.51441121e-02 5.52419186e-01 3.23449373e-01
2.07318559e-01 -8.35520804e-01 -3.98697168e-01 -4.00741845e-02
1.16800570e+00 -1.49130797e+00 -3.07689011e-02 -3.29401225e-01
-1.03346324e+00 6.98637486e-01 5.79860091e-01 -8.84342715e-02
2.93918610e-01 -2.85808951e-01 -2.79917549e-02 -1.08585134e-03
-1.14425337e+00 8.20787922e-02 4.22986865e-01 6.22653306e-01
4.83015656e-01 2.85490066e-01 -6.28971994e-01 1.02595270e+00
-5.12416542e-01 -5.37696816e-02 3.83734196e-01 2.31654897e-01
7.13343993e-02 -1.02663672e+00 -1.93186700e-01 7.85031557e-01
-5.39494455e-01 -2.99700677e-01 -3.83289307e-01 1.06562662e+00
4.36533913e-02 8.78620028e-01 1.12524489e-02 2.54229177e-02
4.10218239e-01 -5.60322739e-02 -1.23432325e-02 -4.43110108e-01
-4.55051243e-01 -4.54598308e-01 -2.02624896e-03 -9.12316918e-01
-1.99835479e-01 -7.90093243e-01 -1.28320956e+00 -4.26858962e-01
-4.24478292e-01 4.35725868e-01 5.33448339e-01 1.12022448e+00
7.99384654e-01 2.72696614e-01 9.02777195e-01 -4.34948057e-01
-1.04658818e+00 -7.75161028e-01 -2.99922556e-01 6.54767990e-01
5.61244905e-01 -7.43214250e-01 -2.87698954e-01 -2.95620412e-01] | [8.699493408203125, 5.6237406730651855] |
2162c58d-1576-4691-ac60-c627b32b2ab9 | rf-based-3d-skeletons | null | null | https://doi.org/10.1145/3230543.3230579 | https://people.csail.mit.edu/mingmin/papers/rfpose3d-sigcomm-zhao.pdf | RF-based 3D skeletons | This paper introduces RF-Pose3D, the first system that infers 3D human skeletons from RF signals. It requires no sensors on the body, and works with multiple people and across walls and occlusions. Further, it generates dynamic skeletons that follow the people as they move, walk or sit. As such, RF-Pose3D provides a significant leap in RF-based sensing and enables new applications in gaming, healthcare, and smart homes.
RF-Pose3D is based on a novel convolutional neural network (CNN) architecture that performs high-dimensional convolutions by decomposing them into low-dimensional operations. This property allows the network to efficiently condense the spatio-temporal information in RF signals. The network first zooms in on the individuals in the scene, and crops the RF signals reflected off each person. For each individual, it localizes and tracks their body parts - head, shoulders, arms, wrists, hip, knees, and feet. Our evaluation results show that RF-Pose3D tracks each keypoint on the human body with an average error of 4.2 cm, 4.0 cm, and 4.9 cm along the X, Y, and Z axes respectively. It maintains this accuracy even in the presence of multiple people, and in new environments that it has not seen in the training set. Demo videos are available at our website: http://rfpose3d.csail.mit.edu. | ['Ming-Min Zhao', 'Yonglong Tian', 'Tianhong Li', 'Hang Zhao', 'Antonio Torralba', 'Zachary Kabelac', 'Rumen Hristov', 'Dina Katabi', 'Mohammad Abu Alsheikh'] | 2018-08-20 | null | null | null | sigcomm-18-proceedings-of-the-2018-conference | ['rf-based-pose-estimation'] | ['computer-vision'] | [ 7.97594041e-02 1.23767525e-01 1.05657540e-01 -4.65468429e-02
3.69789228e-02 -2.60648370e-01 1.58869445e-01 -4.16726798e-01
-2.09753245e-01 5.13023734e-01 2.94886291e-01 4.13286202e-02
-6.77404478e-02 -9.66455042e-01 -5.94181180e-01 -4.50015455e-01
-5.89821398e-01 4.35052037e-01 2.15378612e-01 -5.90791442e-02
-5.96149445e-01 6.19074523e-01 -1.20485580e+00 -6.26415387e-02
1.93499252e-01 1.15865803e+00 -2.21263424e-01 9.03503597e-01
5.96341372e-01 4.45846707e-01 -7.49243855e-01 -1.50249019e-01
3.12362164e-01 5.95735833e-02 -2.93130845e-01 -3.47119063e-01
3.94523770e-01 -6.38255894e-01 -6.14214182e-01 3.38874131e-01
9.80727851e-01 -9.56395566e-02 4.02919143e-01 -9.89517093e-01
-3.02690983e-01 4.98277485e-01 -8.69150519e-01 -1.95339158e-01
8.64285886e-01 -6.23134449e-02 3.13287795e-01 -6.38476431e-01
3.11723351e-01 1.34577048e+00 1.22433794e+00 7.07735658e-01
-8.19792032e-01 -8.27332437e-01 -6.58800900e-02 -5.13772011e-01
-1.52890754e+00 -2.79567122e-01 5.61707675e-01 -3.95293593e-01
5.85129082e-01 2.33906269e-01 1.02976990e+00 1.45628572e+00
3.59251946e-01 6.05937421e-01 4.10491437e-01 -1.62474468e-01
9.07196626e-02 -5.77732503e-01 -1.99931398e-01 8.26584160e-01
4.16671962e-01 1.14011198e-01 -9.41718578e-01 -1.67043313e-01
1.18590474e+00 3.20441455e-01 -3.08137983e-01 -2.25083098e-01
-1.48853695e+00 2.26497680e-01 5.73998749e-01 6.29435945e-03
-4.58738327e-01 6.55898333e-01 -1.75067276e-01 -2.22078443e-01
3.06490690e-01 -1.53232619e-01 -5.63497841e-01 -3.32263671e-02
-8.91777813e-01 3.63662302e-01 5.99721134e-01 8.96726191e-01
2.18707055e-01 4.81727943e-02 6.48357393e-03 5.19122124e-01
6.53874338e-01 1.34531903e+00 5.06085418e-02 -9.93905187e-01
5.50751865e-01 3.55782270e-01 2.84660995e-01 -1.12906611e+00
-1.13036549e+00 -7.11363435e-01 -1.25328457e+00 -2.48318594e-02
1.08642273e-01 -8.47703338e-01 -1.02965760e+00 1.84835672e+00
6.09903216e-01 2.30105728e-01 -3.28412801e-01 1.06166840e+00
9.95989263e-01 2.28839472e-01 -1.83681056e-01 2.60513961e-01
1.35619581e+00 -4.58675057e-01 -3.72516513e-01 -4.24426556e-01
-7.57830814e-02 -3.86943907e-01 6.28898144e-01 3.40151131e-01
-1.01220596e+00 -8.65170538e-01 -1.22261298e+00 3.07311356e-01
7.47459382e-02 3.58077794e-01 6.41493022e-01 8.07113290e-01
-8.79537225e-01 4.12785321e-01 -1.16010106e+00 -4.55296487e-01
1.57077104e-01 5.54060280e-01 -1.32943764e-01 2.08299920e-01
-1.33321404e+00 3.32896709e-01 -3.51426810e-01 4.44166571e-01
-6.60935521e-01 -5.06296456e-01 -8.03212762e-01 -1.94186360e-01
-6.46060193e-03 -1.27457297e+00 1.30064988e+00 -2.84700096e-01
-1.63421369e+00 4.40705091e-01 2.90633161e-02 -2.74123251e-01
5.64447701e-01 -8.92194510e-01 -8.24455738e-01 1.08523592e-01
2.35874295e-01 5.72430253e-01 7.15346098e-01 -7.98117697e-01
-4.29393023e-01 -7.47394979e-01 -1.26824498e-01 1.73101686e-02
3.29781584e-02 -3.57709318e-01 -6.13439620e-01 -6.35516763e-01
5.17917633e-01 -1.16243613e+00 -1.62600309e-01 2.12438494e-01
-7.44514287e-01 2.76723146e-01 5.74240029e-01 -4.64320511e-01
9.84851539e-01 -2.05136108e+00 -4.95674722e-02 5.24861991e-01
3.06633294e-01 -3.77212279e-02 4.17748570e-01 2.42365018e-01
2.64767885e-01 -2.02601939e-01 -1.25443621e-03 -2.98177540e-01
-4.20535319e-02 -1.75995886e-01 2.41724074e-01 6.63126230e-01
-3.61849904e-01 7.29278207e-01 -6.60025299e-01 -2.65269905e-01
1.32717490e-01 1.01371312e+00 -2.87248880e-01 -2.49062285e-01
3.62743795e-01 6.20443583e-01 -6.27649546e-01 9.30345356e-01
7.17770219e-01 -3.23466927e-01 1.98158011e-01 -2.70542979e-01
-2.56540123e-02 1.02862090e-01 -1.49176025e+00 1.83923626e+00
-3.17302227e-01 4.27388310e-01 2.98718780e-01 -5.11863470e-01
8.26911509e-01 4.05737102e-01 9.19836402e-01 -7.24075019e-01
1.89287722e-01 -7.88799450e-02 -4.45271462e-01 -5.06062686e-01
1.35721371e-01 4.16255966e-02 -6.12056732e-01 5.18986583e-01
-1.03204183e-01 4.60283935e-01 -1.29040554e-01 -4.29203967e-03
1.33664513e+00 2.02338457e-01 1.38319790e-01 7.10150599e-02
2.36924589e-01 -4.86718088e-01 5.79696178e-01 8.06288540e-01
1.31196931e-01 6.29619122e-01 -3.87297124e-01 -7.62534440e-01
-4.20130491e-01 -1.74202502e+00 1.13363050e-01 9.63096142e-01
3.24780464e-01 -2.96786398e-01 -5.32893419e-01 -2.61279136e-01
2.86380768e-01 -1.27310514e-01 -5.54218113e-01 -5.70575185e-02
-7.72061288e-01 -5.03185987e-01 8.99786532e-01 7.24217594e-01
9.23867941e-01 -8.76047373e-01 -1.19526196e+00 2.92317003e-01
-4.64247853e-01 -9.23875809e-01 -2.77369767e-01 -2.81262435e-02
-7.26429343e-01 -1.09293270e+00 -1.09970558e+00 -4.20036167e-01
3.90806586e-01 1.41954809e-01 1.11949456e+00 -1.41536742e-01
-5.02204001e-01 6.30241275e-01 -7.46511966e-02 -5.53979695e-01
5.02743781e-01 2.20572561e-01 5.37452638e-01 -8.12993199e-02
1.44481435e-01 -9.72441614e-01 -9.45067525e-01 5.30956805e-01
4.27557006e-02 -4.41127680e-02 6.21483028e-01 1.27651677e-01
2.91142255e-01 1.93147194e-02 1.36830255e-01 -2.88170606e-01
2.83177465e-01 -3.94927859e-01 -2.63321936e-01 -1.86487567e-02
1.39665827e-01 -2.75158942e-01 1.55272365e-01 -3.82138610e-01
-7.77721405e-01 4.93724078e-01 -2.77261157e-02 -1.60959706e-01
-3.83988649e-01 6.96170107e-02 -1.69330627e-01 3.21676612e-01
8.38761270e-01 -1.21884577e-01 -2.91754305e-01 -5.87100685e-01
2.12764993e-01 6.78002536e-01 1.01512396e+00 -4.99974549e-01
1.00710022e+00 8.37132156e-01 1.87098771e-01 -1.15070057e+00
-8.47690701e-01 -4.36554760e-01 -7.63516128e-01 -6.67210639e-01
9.20661151e-01 -1.19377339e+00 -1.21646953e+00 5.91531992e-01
-9.72425044e-01 -4.62204397e-01 -1.13689110e-01 7.02634275e-01
-3.05223972e-01 -1.48069998e-02 -6.46125197e-01 -8.38838339e-01
-6.72919691e-01 -4.15088773e-01 1.41743839e+00 4.37318772e-01
-6.52160227e-01 -5.65852940e-01 1.56699151e-01 2.02582940e-01
2.80270308e-01 6.23013139e-01 1.90377012e-01 2.93221593e-01
-3.35802555e-01 -2.28404880e-01 3.24439317e-01 -2.48961851e-01
8.71671364e-02 -2.93407530e-01 -8.49195242e-01 -2.58679003e-01
-5.26283801e-01 2.28106473e-02 6.64044023e-01 7.80279279e-01
7.88392425e-01 -1.41395360e-01 -9.06758070e-01 8.67018521e-01
8.66400898e-01 -1.90328155e-02 4.25182521e-01 1.21148571e-01
6.01167262e-01 3.07646334e-01 3.73073041e-01 6.52378678e-01
4.01789278e-01 7.43614256e-01 4.37388748e-01 -3.22404414e-01
-2.80815631e-01 -2.95809954e-01 5.12352645e-01 3.36115509e-01
-5.66125572e-01 -4.55513358e-01 -8.49795699e-01 1.64226726e-01
-1.68336260e+00 -9.28598166e-01 -2.92408735e-01 2.08664536e+00
2.78473139e-01 1.23802476e-01 4.36735034e-01 2.98890680e-01
7.87342608e-01 -2.14079209e-02 -6.27588511e-01 1.75575003e-01
1.25005916e-01 4.60638076e-01 7.73632526e-01 2.20410630e-01
-1.37934780e+00 3.92645925e-01 6.44307804e+00 -1.34158626e-01
-1.20543289e+00 -1.62018076e-01 -3.99952978e-02 -4.67786938e-01
3.09984356e-01 -7.29607463e-01 -8.43444765e-01 4.01796877e-01
6.43437982e-01 4.78920728e-01 2.01888934e-01 7.51637816e-01
2.10362241e-01 -5.25215305e-02 -8.96409392e-01 1.05440700e+00
-3.96496765e-02 -8.31618547e-01 -4.52662379e-01 1.37292281e-01
3.56598437e-01 2.46766984e-01 1.24977969e-01 -2.76826136e-02
4.13286686e-01 -1.10647023e+00 1.08644497e+00 6.13861561e-01
1.02328801e+00 -7.62462854e-01 6.19211733e-01 4.77377474e-01
-1.61483598e+00 -8.43692645e-02 -7.20824599e-02 -1.82126850e-01
6.27376020e-01 9.42585945e-01 -7.28038251e-01 3.33249956e-01
1.19938707e+00 4.58447546e-01 -1.80027917e-01 7.80351460e-01
-6.03712857e-01 3.88020039e-01 -7.54754007e-01 8.46255571e-02
-4.00511593e-01 3.69066298e-01 4.85287994e-01 1.04313290e+00
7.98555434e-01 1.51273534e-01 2.40838960e-01 4.93774325e-01
-3.25189568e-02 -4.83933955e-01 -4.42072690e-01 5.72552860e-01
6.29680276e-01 1.09185970e+00 -6.25742733e-01 -4.82195355e-02
-1.16581377e-02 1.04352927e+00 -4.13049132e-01 4.62209433e-01
-9.58769739e-01 -5.59297204e-01 7.23628998e-01 6.67059720e-01
2.96667665e-01 -6.09885514e-01 2.63703298e-02 -8.49915147e-01
1.54723331e-01 -3.72714251e-01 2.38547042e-01 -8.02011430e-01
-8.64149988e-01 3.07541370e-01 -1.20365053e-01 -1.20239449e+00
-3.67800951e-01 -4.91207540e-01 -3.53866637e-01 5.56170762e-01
-6.84726000e-01 -1.09091294e+00 -7.94289231e-01 9.76868868e-01
9.80070755e-02 1.26287058e-01 8.07575762e-01 6.04894102e-01
-3.40103298e-01 5.54096937e-01 -2.98652589e-01 5.44432282e-01
5.07369518e-01 -9.03396904e-01 8.62437665e-01 5.48782825e-01
2.40728501e-02 7.97225714e-01 5.18218577e-01 -8.64220619e-01
-1.29464197e+00 -1.04894173e+00 6.48107409e-01 -3.10579866e-01
-4.42985334e-02 -6.15946472e-01 3.00004911e-02 7.73777425e-01
-3.55409741e-01 1.67644247e-01 6.67302847e-01 3.60201508e-01
-5.05728424e-02 -3.25178385e-01 -1.18229222e+00 5.32321811e-01
1.57130086e+00 -1.98093951e-01 -3.62636745e-01 2.75363009e-02
4.79910940e-01 -7.45687723e-01 -7.59111345e-01 4.93224442e-01
1.33477306e+00 -8.96633744e-01 1.61179054e+00 8.76560807e-02
-1.03470393e-01 -4.93393362e-01 -1.20515220e-01 -1.01005375e+00
-5.31887054e-01 -4.98362243e-01 -5.86309433e-01 6.81240380e-01
2.87043840e-01 -4.60383803e-01 1.10066462e+00 6.62163720e-02
4.76764664e-02 -3.68114054e-01 -9.24927175e-01 -6.34308696e-01
-6.70206785e-01 -6.08440220e-01 7.73873866e-01 4.71107423e-01
-2.44121179e-01 5.04457474e-01 -7.36191273e-01 6.06811345e-01
1.00042343e+00 2.94349082e-02 1.10860515e+00 -1.66973984e+00
-4.40816313e-01 1.86635017e-01 -3.95779938e-01 -1.57184577e+00
-5.78941822e-01 -3.60546768e-01 5.02945073e-02 -1.69930935e+00
-2.90547878e-01 -4.07364964e-01 3.42851654e-02 6.30321503e-01
4.30132329e-01 4.20443982e-01 -2.56177559e-02 -7.71466568e-02
-5.09118259e-01 1.69791788e-01 1.09799421e+00 -2.79000372e-01
-3.71187389e-01 6.30113780e-01 -4.33977306e-01 9.98849452e-01
1.05836952e+00 -3.02675545e-01 -2.79283404e-01 -7.04895318e-01
3.79360765e-01 1.43822640e-01 7.19674230e-01 -1.85224259e+00
4.16826427e-01 1.68135911e-01 1.35768259e+00 -9.80980694e-01
7.13723958e-01 -8.42994690e-01 6.73016906e-01 8.07656825e-01
2.15027541e-01 -3.33302282e-02 -2.51107439e-02 4.22435164e-01
4.81639028e-01 5.92229843e-01 4.85407531e-01 -3.70219767e-01
-1.69032246e-01 4.26467478e-01 -5.02532125e-01 -1.92786962e-01
5.71032286e-01 -4.11571741e-01 -6.78624064e-02 -5.48214734e-01
-1.00388646e+00 1.38715699e-01 1.46600738e-01 4.21775699e-01
6.15640163e-01 -1.50705898e+00 -3.55533719e-01 5.03708005e-01
-3.46458077e-01 3.03622246e-01 3.16500694e-01 6.38467133e-01
-4.38436180e-01 3.77682000e-01 -1.37343779e-01 -8.65996480e-01
-1.27614307e+00 -5.87470569e-02 4.54045683e-01 4.08105031e-02
-8.81968319e-01 8.98003578e-01 -8.73661265e-02 -5.91555893e-01
4.94279563e-01 -4.16783661e-01 -7.86382407e-02 -2.48204768e-01
5.94311118e-01 6.92348719e-01 -1.97849616e-01 -7.10825562e-01
-7.61398733e-01 1.26298106e+00 6.94866180e-01 -3.05715889e-01
1.39526999e+00 -2.12964997e-01 3.95180404e-01 2.17487693e-01
6.40836477e-01 3.80665451e-01 -1.20302534e+00 -5.02091534e-02
-3.23057175e-01 -1.11026853e-01 -2.44482428e-01 -8.34454894e-01
-1.21890056e+00 7.61008382e-01 7.14092314e-01 2.39815898e-02
1.07368040e+00 1.69746906e-01 1.02732134e+00 4.18136358e-01
8.11064422e-01 -8.98914278e-01 9.95559096e-02 4.92510676e-01
6.64966822e-01 -3.56567472e-01 2.50111639e-01 -3.37126404e-01
8.29421282e-02 9.65851545e-01 4.37856108e-01 1.12454370e-01
7.69793928e-01 5.98925650e-01 1.09370351e-01 -4.56085294e-01
-1.65573820e-01 -1.64603889e-01 9.47007686e-02 1.15625155e+00
3.27288181e-01 4.42864865e-01 3.91862690e-01 8.63169909e-01
-6.91640377e-01 9.79583040e-02 -4.03067097e-02 1.08470774e+00
-5.96775055e-01 -6.07240617e-01 -6.72966182e-01 2.15608388e-01
-4.23978746e-01 4.22614425e-01 -3.75510246e-01 8.71865153e-01
6.86178982e-01 8.93121362e-01 -6.70976192e-02 -7.77843714e-01
6.24709904e-01 -2.58713573e-01 5.09828150e-01 -3.55163068e-01
-2.89326340e-01 1.15421370e-01 9.57817212e-02 -9.02486324e-01
-4.91537839e-01 -5.80717564e-01 -1.48035848e+00 -3.94927859e-01
7.35676065e-02 -1.40135795e-01 6.73510015e-01 4.57512647e-01
4.39824700e-01 7.59069383e-01 4.23217207e-01 -1.21622956e+00
-1.00352809e-01 -8.57336819e-01 -6.47168994e-01 -1.77058145e-01
4.23701882e-01 -8.47856939e-01 -1.82970166e-01 -7.00017512e-02] | [6.869505405426025, 0.2985776960849762] |
a911ab1e-19bd-408a-bc9d-e83ad9a41e62 | speakers-account-for-asymmetries-in-visual | 1807.09000 | null | https://arxiv.org/abs/1807.09000v4 | https://arxiv.org/pdf/1807.09000v4.pdf | The division of labor in communication: Speakers help listeners account for asymmetries in visual perspective | Recent debates over adults' theory of mind use have been fueled by surprising failures of perspective-taking in communication, suggesting that perspective-taking can be relatively effortful. How, then, should speakers and listeners allocate their resources to achieve successful communication? We begin with the observation that this shared goal induces a natural division of labor: the resources one agent chooses to allocate toward perspective-taking should depend on their expectations about the other's allocation. We formalize this idea in a resource-rational model augmenting recent probabilistic weighting accounts with a mechanism for (costly) control over the degree of perspective-taking. In a series of simulations, we first derive an intermediate degree of perspective weighting as an optimal tradeoff between expected costs and benefits of perspective-taking. We then present two behavioral experiments testing novel predictions of our model. In Experiment 1, we manipulated the presence or absence of occlusions in a director-matcher task and found that speakers spontaneously produced more informative descriptions to account for "known unknowns" in their partner's private view. In Experiment 2, we compared the scripted utterances used by confederates in prior work with those produced in interactions with unscripted directors. We found that confederates were systematically less informative than listeners would initially expect given the presence of occlusions, but listeners used violations to adaptively make fewer errors over time. Taken together, our work suggests that people are not simply "mindblind"; they use contextually appropriate expectations to navigate the division of labor with their partner. We discuss how a resource rational framework may provide a more deeply explanatory foundation for understanding flexible perspective-taking under processing constraints. | ['Noah D. Goodman', 'Robert D. Hawkins', 'Hyowon Gweon'] | 2018-07-24 | null | null | null | null | ['known-unknowns'] | ['miscellaneous'] | [ 2.75186867e-01 5.56655824e-01 8.35307837e-02 -9.63039756e-01
-5.47491312e-01 -7.15975225e-01 6.19644582e-01 3.24897885e-01
-8.88523102e-01 2.82211185e-01 1.01848698e+00 -2.36569971e-01
-6.93402737e-02 -5.96011758e-01 -1.46984458e-01 -4.04305965e-01
2.61268258e-01 4.02708471e-01 -2.77667850e-01 -5.80594316e-02
6.84475720e-01 2.48419851e-01 -1.33098304e+00 3.05069014e-02
5.76742053e-01 1.45016178e-01 2.32884139e-01 4.22362357e-01
2.19694048e-01 6.49954915e-01 -4.89791274e-01 -7.30744779e-01
1.32171750e-01 -4.38334316e-01 -5.43112218e-01 1.62517235e-01
3.87401551e-01 -7.18490481e-01 -1.27835765e-01 9.41249311e-01
4.70496684e-01 4.13342476e-01 5.65408468e-01 -8.21272016e-01
-7.91081667e-01 1.09260893e+00 -7.91170895e-02 3.05759788e-01
7.77584672e-01 4.80083346e-01 1.14173365e+00 -6.27150834e-01
5.64894199e-01 1.64258409e+00 2.16538519e-01 7.46469259e-01
-1.41688514e+00 -5.17908156e-01 7.26430714e-01 -3.22213262e-01
-1.04987764e+00 -1.12229013e+00 8.15332830e-01 -5.19257247e-01
5.99056959e-01 4.33129400e-01 1.10722542e+00 1.25644779e+00
1.45030124e-02 3.77539456e-01 1.44286919e+00 -2.02640250e-01
3.82025391e-01 2.74365693e-01 3.30705464e-01 4.43128258e-01
5.48341930e-01 2.68886060e-01 -1.24364913e+00 -5.46676576e-01
6.56647682e-01 -2.71094173e-01 -3.49994838e-01 -6.41336292e-02
-1.27439928e+00 8.13214242e-01 -3.28233503e-02 4.29043382e-01
-3.71021301e-01 -1.27126530e-01 2.02456415e-02 3.82564873e-01
3.65526170e-01 8.22684169e-01 -1.02462471e-01 -4.12769347e-01
-4.34336752e-01 6.28229082e-01 8.82052243e-01 7.43902802e-01
8.53279978e-02 -5.65564334e-02 -2.38493569e-02 8.63151133e-01
6.87877595e-01 3.28687161e-01 2.86622435e-01 -1.48149908e+00
4.01026845e-01 8.23963135e-02 4.96809006e-01 -9.24231827e-01
-3.94870132e-01 -4.26762164e-01 1.36296198e-01 1.83215216e-01
7.34243810e-01 -4.43544209e-01 -7.93873519e-02 2.57070088e+00
1.57578886e-01 -8.67293894e-01 -3.48115414e-02 8.76912296e-01
-2.57471949e-01 2.31238365e-01 4.01309490e-01 -6.68643177e-01
1.40588105e+00 -1.66024804e-01 -5.41344106e-01 -6.73609376e-01
3.95786732e-01 -7.83581555e-01 1.13569450e+00 3.17323834e-01
-1.45702314e+00 1.52327716e-02 -1.00339103e+00 -1.01355277e-01
4.34078038e-01 -6.36578381e-01 1.01176679e+00 8.42039824e-01
-1.11015320e+00 6.13305807e-01 -8.40911567e-01 -6.23239756e-01
1.50713593e-01 -5.85574284e-03 -2.53784180e-01 2.71331102e-01
-7.23538339e-01 9.07297075e-01 -2.35006541e-01 9.48661864e-02
-8.19176793e-01 -3.25635582e-01 -8.31625819e-01 3.13963562e-01
2.08384067e-01 -9.51863647e-01 1.67075264e+00 -1.21970761e+00
-1.57375872e+00 1.12746882e+00 -1.90235108e-01 -5.23141436e-02
2.88717479e-01 -1.11788221e-01 -4.31342795e-02 9.06879231e-02
3.36068690e-01 6.22388780e-01 7.12956846e-01 -1.01376987e+00
-1.74807191e-01 -8.98306727e-01 4.06637639e-01 7.74222136e-01
1.10110007e-01 3.97111803e-01 4.73375976e-01 -6.89669132e-01
6.98651433e-01 -9.13580179e-01 8.77137110e-02 2.30603993e-01
-2.74406463e-01 -2.60824710e-01 -2.80246317e-01 -1.18678614e-01
8.91381264e-01 -2.22913599e+00 -1.55775517e-01 -6.00050204e-02
4.14731890e-01 -5.67567647e-01 8.42326581e-02 6.50912464e-01
-1.50216356e-01 4.02165085e-01 3.23912427e-02 -5.65374911e-01
4.85540360e-01 -2.14504644e-01 -2.35803112e-01 7.18634963e-01
-5.51698327e-01 3.78348023e-01 -5.19409955e-01 -3.53521317e-01
-3.40837687e-01 2.79065192e-01 -7.54641116e-01 2.69005269e-01
-4.49613445e-02 2.92745024e-01 -1.59680679e-01 6.01074457e-01
4.11911905e-01 6.30680174e-02 6.11328781e-01 4.09064114e-01
-4.92529601e-01 8.91794145e-01 -8.30617785e-01 1.17794359e+00
-9.17745084e-02 6.42566383e-01 6.24387503e-01 -2.99520463e-01
3.59290868e-01 1.71197653e-01 -3.25403452e-01 -3.56705070e-01
4.62872773e-01 1.20675847e-01 7.96648681e-01 -4.23017591e-01
3.15016925e-01 -8.51253986e-01 -1.36545002e-01 1.08407509e+00
-4.87785637e-01 -3.68851036e-01 -3.32723767e-01 2.36610591e-01
5.42570174e-01 -1.44148082e-01 3.15435886e-01 -5.25044560e-01
1.17313534e-01 -3.74457330e-01 9.85331535e-01 9.19810951e-01
-5.45632064e-01 2.63383836e-01 6.49411380e-01 -3.46444398e-01
-8.73154819e-01 -1.01380336e+00 -1.53029831e-02 1.42807853e+00
-1.98499635e-01 7.45220408e-02 -1.00734198e+00 -7.07966536e-02
-2.88467765e-01 1.49279475e+00 -3.52641433e-01 -1.15120828e-01
-5.94814658e-01 -3.44484031e-01 3.93937737e-01 3.61470580e-01
2.62864500e-01 -7.94148266e-01 -1.19968498e+00 -6.09293878e-02
-1.46436885e-01 -6.55130148e-01 -9.55814838e-01 -4.49525923e-01
-8.61860454e-01 -6.48191988e-01 -4.03291613e-01 -8.56052116e-02
7.20021665e-01 6.52972937e-01 7.04347491e-01 2.42398500e-01
3.42100114e-01 9.39206004e-01 1.03334516e-01 -6.19304419e-01
-3.20026726e-01 -4.07746941e-01 2.70030141e-01 -2.27964029e-01
3.24394852e-01 -1.01216018e+00 -7.08775103e-01 2.68162996e-01
-4.83137429e-01 1.64688915e-01 2.24989682e-01 4.80229437e-01
-2.45208040e-01 -5.24430931e-01 2.09586263e-01 -9.34441209e-01
1.03341639e+00 -4.44962323e-01 -6.53545260e-01 -4.15750854e-02
-2.61976451e-01 1.21443784e-02 2.13281363e-01 -3.86067301e-01
-1.62882388e+00 -5.49807429e-01 2.57565856e-01 3.57382089e-01
-3.50083224e-02 2.15798467e-01 -2.36892283e-01 1.64859340e-01
5.45523763e-01 -4.59177382e-02 1.90248162e-01 -3.14731061e-01
9.65279713e-02 3.67804110e-01 1.02614440e-01 -1.27208114e+00
5.16177535e-01 3.43971759e-01 -2.98496276e-01 -9.73472536e-01
-9.48995590e-01 4.26427305e-01 -1.21970654e-01 -1.40746161e-01
7.85696447e-01 -1.06202030e+00 -1.23732662e+00 3.52925122e-01
-1.02291620e+00 -4.08371806e-01 -1.32284537e-01 9.19132411e-01
-8.30458164e-01 1.99004576e-01 -2.96961039e-01 -1.14856565e+00
-3.36426981e-02 -1.11530411e+00 5.39843440e-01 2.42212847e-01
-9.65958357e-01 -7.76699722e-01 -2.64762908e-01 6.72846138e-01
6.00522578e-01 -2.13499978e-01 1.02165377e+00 -1.13693011e+00
-7.10979402e-01 1.42634436e-01 3.41418803e-01 -2.88453430e-01
-2.07070261e-01 -5.62233962e-02 -1.04025781e+00 7.26429094e-03
9.88105178e-01 -1.76087856e-01 3.88868928e-01 3.67749721e-01
6.04682803e-01 -5.84040105e-01 -8.86951238e-02 4.30425584e-01
8.87006879e-01 4.89929289e-01 -4.90009002e-02 -1.94192022e-01
6.94035962e-02 1.22174215e+00 5.12948751e-01 4.60495919e-01
9.29018319e-01 6.43612742e-01 -7.25352243e-02 9.23702359e-01
3.73899609e-01 -6.35295749e-01 4.05375272e-01 3.10159743e-01
8.90990645e-02 -2.86594748e-01 -7.20137000e-01 4.97601926e-01
-1.52340674e+00 -1.06964540e+00 3.91843885e-01 2.62792706e+00
6.62905514e-01 3.80549729e-01 8.54572430e-02 -2.38921359e-01
7.14129269e-01 5.86284816e-01 -5.32287598e-01 -6.18170381e-01
3.00560236e-01 -3.99148583e-01 9.80082378e-02 1.07831502e+00
-3.55292052e-01 7.35197186e-01 6.86689234e+00 -1.94992349e-01
-6.76676810e-01 2.59849846e-01 7.54107237e-01 -4.29937392e-01
-1.01194942e+00 2.53549278e-01 -5.61208904e-01 2.56439149e-01
9.62878108e-01 -5.48224807e-01 4.94892806e-01 7.26026714e-01
3.26810509e-01 -6.78448617e-01 -1.74764538e+00 5.53791642e-01
5.72262108e-02 -7.21717834e-01 -5.70164740e-01 1.72603458e-01
-9.75225307e-03 -5.24474621e-01 1.59637719e-01 -1.00746453e-01
5.79630375e-01 -8.13919008e-01 1.27344227e+00 3.69582474e-01
1.79673791e-01 -2.05201358e-01 1.67872474e-01 7.85777986e-01
-2.50980705e-01 -1.85921162e-01 -1.81074470e-01 -9.30938423e-01
3.82936567e-01 1.68744266e-01 -7.60283113e-01 -7.18789697e-01
-5.44306822e-02 -2.62278706e-01 7.68882036e-02 3.01857471e-01
-4.06826138e-01 1.82757571e-01 -4.03166145e-01 -1.29192770e-01
-1.29431367e-01 -2.13919088e-01 9.42523539e-01 6.14769399e-01
8.92765727e-03 6.44920945e-01 -2.83002764e-01 1.08952200e+00
1.83811635e-01 5.51296696e-02 -7.02461481e-01 -1.44882753e-01
1.15026772e+00 8.26037943e-01 -6.76310837e-01 -2.36601204e-01
-1.93604276e-01 5.45869470e-01 4.47449148e-01 2.90701181e-01
-4.86953318e-01 2.93779761e-01 1.07590842e+00 3.02808315e-01
-3.11431020e-01 -5.46752632e-01 -5.48676908e-01 -1.43565404e+00
-2.57134866e-02 -6.46975100e-01 8.35305452e-02 -7.59285867e-01
-9.98437405e-01 8.70986953e-02 4.88182366e-01 -3.52981478e-01
-2.12509871e-01 -1.53611064e-01 -4.72148210e-01 1.02580523e+00
-5.92575431e-01 -8.02826047e-01 2.21013814e-01 -7.45736361e-02
9.67707038e-01 3.01195681e-01 6.61487818e-01 -4.06406760e-01
-5.26286662e-01 4.16841686e-01 -9.61118579e-01 -3.37323368e-01
7.12392628e-01 -1.00527942e+00 3.99647504e-01 7.79475689e-01
-1.34124488e-01 1.53561282e+00 1.01184690e+00 -6.09019756e-01
-1.32704794e+00 1.10777646e-01 1.18207359e+00 -4.47056830e-01
5.21398306e-01 -5.15829921e-01 -6.49071813e-01 1.10697281e+00
4.29182351e-01 -5.09063959e-01 9.79762673e-01 3.79324704e-01
-5.08936107e-01 1.04268454e-01 -1.43834889e+00 9.69858646e-01
1.71777534e+00 -5.90743124e-01 -1.18354833e+00 2.58785963e-01
9.62649107e-01 -1.99070051e-01 -5.03112376e-01 -2.17264384e-01
9.25459087e-01 -1.44226229e+00 6.80395544e-01 -3.90271783e-01
1.48887768e-01 1.88176110e-01 -4.97845024e-01 -1.21174610e+00
-4.48388278e-01 -8.28599751e-01 7.57518530e-01 1.40154839e+00
1.72206938e-01 -1.03964186e+00 2.61843175e-01 1.54497898e+00
-6.58262894e-02 -4.92945045e-01 -9.66775835e-01 -3.02083418e-02
1.05062678e-01 -4.08800662e-01 5.28395653e-01 7.20137000e-01
3.56932700e-01 3.55906963e-01 1.43508106e-01 2.26995870e-01
7.84206390e-01 -1.17232978e-01 3.61437947e-01 -1.25553524e+00
-4.87957418e-01 -5.21470964e-01 2.37677351e-01 -8.96760166e-01
2.44703367e-01 -2.43173271e-01 -1.63478758e-02 -8.58053207e-01
2.52245307e-01 -4.05061573e-01 4.57250863e-01 1.66179195e-01
-9.32206362e-02 -6.38105810e-01 5.94497800e-01 3.31939638e-01
-1.05968304e-01 3.40395421e-01 9.05549884e-01 5.54290593e-01
-3.51690799e-01 2.93690503e-01 -1.78727221e+00 1.36589634e+00
5.89199901e-01 -2.35076204e-01 -7.53462493e-01 -6.06284261e-01
6.30817294e-01 6.97804630e-01 2.13249326e-01 -5.39193034e-01
2.14571133e-01 -8.04813564e-01 1.37168154e-01 -3.01602464e-02
6.39300287e-01 -6.37079179e-01 1.80743396e-01 2.75570035e-01
-7.95637429e-01 3.06667268e-01 -1.74286842e-01 3.18262428e-01
4.57180262e-01 -2.70708412e-01 8.97138059e-01 -4.72341299e-01
9.21681151e-02 -4.43606108e-01 -9.02615905e-01 1.27340332e-01
7.80586660e-01 -6.40485734e-02 -4.39818144e-01 -7.75123835e-01
-8.55047047e-01 3.30114737e-02 8.46039474e-01 1.11890472e-01
5.41821897e-01 -6.94890797e-01 -3.97713065e-01 9.56290066e-02
-1.38735190e-01 -3.04060787e-01 1.83807299e-01 8.28341424e-01
-3.76442939e-01 4.08558667e-01 6.65841177e-02 1.29068762e-01
-1.19399714e+00 3.86960059e-01 6.03117347e-01 4.68317896e-01
-2.74950117e-01 1.40109944e+00 6.98567569e-01 -1.51096582e-01
-3.49626057e-02 -3.10705215e-01 1.10065296e-01 3.99747342e-01
7.30674684e-01 3.29228640e-01 -6.61537051e-01 -5.97935736e-01
-3.81940842e-01 2.98847795e-01 -3.82509887e-01 -9.20196235e-01
1.17139637e+00 -5.95179081e-01 -2.00761929e-01 7.91279614e-01
4.16732192e-01 6.14329219e-01 -1.32728040e+00 -3.95985357e-02
-3.75959873e-01 -1.02199662e+00 -5.23467541e-01 -6.79947495e-01
-2.87725955e-01 6.17013574e-01 -2.33307391e-01 1.47299081e-01
5.75788379e-01 1.46609083e-01 -4.62222807e-02 3.47410023e-01
7.23451793e-01 -9.11203802e-01 -6.26761913e-02 6.62349463e-02
1.12390578e+00 -8.70115399e-01 1.79032072e-01 -5.53286910e-01
-8.34244072e-01 6.26149476e-01 7.49886394e-01 1.00897051e-01
2.65847355e-01 -1.62479162e-01 -8.99848565e-02 -3.62634152e-01
-1.35485649e+00 2.56291956e-01 -5.25892675e-01 5.75338840e-01
8.15559924e-01 3.66314560e-01 -8.97431791e-01 7.46997654e-01
-8.23617697e-01 -5.03456950e-01 9.72554862e-01 8.46826255e-01
-8.40837479e-01 -9.61600244e-01 -6.22494936e-01 5.00299297e-02
-5.02672732e-01 -7.96471536e-02 -5.73053837e-01 6.63641572e-01
-7.91312233e-02 1.00076985e+00 3.00511628e-01 2.01331973e-01
2.24247888e-01 5.12794912e-01 5.06241262e-01 -1.10556877e+00
-4.68824863e-01 2.04997376e-01 5.30137658e-01 -4.60927904e-01
-1.58244178e-01 -1.43069267e+00 -7.97097445e-01 -3.98994893e-01
3.10294311e-02 -2.87811796e-04 4.86700654e-01 9.24001098e-01
2.79615134e-01 -1.15284398e-01 4.17707413e-01 -4.23746318e-01
-8.93455803e-01 -9.43648517e-01 -5.67740440e-01 -4.15513739e-02
3.54189992e-01 -5.01371920e-01 -7.18805671e-01 -3.77986312e-01] | [10.324118614196777, 8.55572509765625] |
8fece3c3-5164-40bd-846a-fcde358fe000 | unsupervised-domain-adaptation-for-automated | 2212.07023 | null | https://arxiv.org/abs/2212.07023v1 | https://arxiv.org/pdf/2212.07023v1.pdf | Unsupervised Domain Adaptation for Automated Knee Osteoarthritis Phenotype Classification | Purpose: The aim of this study was to demonstrate the utility of unsupervised domain adaptation (UDA) in automated knee osteoarthritis (OA) phenotype classification using a small dataset (n=50). Materials and Methods: For this retrospective study, we collected 3,166 three-dimensional (3D) double-echo steady-state magnetic resonance (MR) images from the Osteoarthritis Initiative dataset and 50 3D turbo/fast spin-echo MR images from our institute (in 2020 and 2021) as the source and target datasets, respectively. For each patient, the degree of knee OA was initially graded according to the MRI Osteoarthritis Knee Score (MOAKS) before being converted to binary OA phenotype labels. The proposed UDA pipeline included (a) pre-processing, which involved automatic segmentation and region-of-interest cropping; (b) source classifier training, which involved pre-training phenotype classifiers on the source dataset; (c) target encoder adaptation, which involved unsupervised adaption of the source encoder to the target encoder and (d) target classifier validation, which involved statistical analysis of the target classification performance evaluated by the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity and accuracy. Additionally, a classifier was trained without UDA for comparison. Results: The target classifier trained with UDA achieved improved AUROC, sensitivity, specificity and accuracy for both knee OA phenotypes compared with the classifier trained without UDA. Conclusion: The proposed UDA approach improves the performance of automated knee OA phenotype classification for small target datasets by utilising a large, high-quality source dataset for training. The results successfully demonstrated the advantages of the UDA approach in classification on small datasets. | ['Weitian Chen', 'James F. Griffith', 'Michael Tim-Yun Ong', 'Kevin Ki-Wai Ho', 'Jack Lee', 'Siyue Li', 'Fan Xiao', 'Donal G. Cahill', 'Yongcheng Yao', 'Junru Zhong'] | 2022-12-14 | null | null | null | null | ['phenotype-classification'] | ['medical'] | [ 3.21439505e-01 -2.46061221e-01 -2.59962291e-01 -2.26642236e-01
-1.29373574e+00 -1.77495807e-01 2.07066938e-01 3.26061219e-01
-7.36714125e-01 8.15334380e-01 3.99661690e-01 -5.39733991e-02
-5.08186102e-01 -4.37838048e-01 -2.27061808e-01 -6.65896952e-01
-6.98502004e-01 9.88219321e-01 4.77346510e-01 2.25819319e-01
1.31058291e-01 3.68642449e-01 -1.42892587e+00 5.00432432e-01
9.16299045e-01 1.17194653e+00 4.36975747e-01 7.26405084e-01
4.09765065e-01 6.14098251e-01 -5.75830042e-01 1.54340044e-01
1.38439357e-01 -2.76011586e-01 -6.99012578e-01 1.17433079e-01
5.37383705e-02 -8.74639452e-02 -2.95411609e-02 4.98912036e-01
9.65605795e-01 -9.67931449e-02 9.84501362e-01 -8.53541493e-01
-4.89488840e-01 2.51262616e-02 -1.71338230e-01 5.12382269e-01
-3.57680954e-02 4.76111621e-01 5.62846601e-01 -8.55220497e-01
9.36259151e-01 6.00591838e-01 7.33218551e-01 2.93307751e-01
-1.34674728e+00 -5.27733743e-01 -7.91014552e-01 1.07248910e-01
-1.29673624e+00 -1.75615326e-01 4.64552581e-01 -8.71835113e-01
6.82704985e-01 1.19147003e-01 1.28276098e+00 8.66568148e-01
7.78762758e-01 4.95489925e-01 1.67294753e+00 -2.80192643e-01
5.89067638e-01 -2.70628542e-01 6.92014173e-02 5.31221867e-01
7.64687434e-02 2.28322506e-01 -2.35093385e-01 -5.87418377e-01
7.89709330e-01 -4.10516798e-01 -3.05967610e-02 -8.26213300e-01
-1.54101455e+00 7.49372184e-01 2.96003968e-01 1.04117498e-01
-6.38862014e-01 -2.49962702e-01 7.35285997e-01 3.95630062e-01
1.75380692e-01 6.37307227e-01 -4.61129963e-01 -1.15484357e-01
-9.69204009e-01 1.55635282e-01 4.67193067e-01 4.24938500e-01
2.18576137e-02 8.43784492e-03 -1.16754957e-02 1.46388757e+00
3.98724377e-01 5.42498291e-01 1.38704145e+00 -8.57029438e-01
-2.20264286e-01 5.16260803e-01 -1.86569527e-01 -4.60616529e-01
-4.61388439e-01 -5.94194710e-01 -5.81368029e-01 2.54473776e-01
4.45841193e-01 3.17387879e-02 -1.49884927e+00 1.33656085e+00
1.72672421e-01 -2.74399608e-01 1.19806111e-01 1.25842869e+00
2.86107898e-01 -5.97571780e-04 3.47398013e-01 -1.60346717e-01
1.73777461e+00 -8.39387298e-01 -3.36959183e-01 -1.89254042e-02
1.02298868e+00 -8.16813886e-01 1.20319593e+00 4.93614763e-01
-7.08676696e-01 -4.18441862e-01 -1.10166800e+00 2.77499974e-01
-2.06323996e-01 4.78644073e-01 6.76199138e-01 8.01593721e-01
-7.29796827e-01 1.06749013e-01 -9.73953187e-01 -3.40101540e-01
4.83250499e-01 8.32413137e-01 -8.56973588e-01 -1.60226405e-01
-1.31498480e+00 9.77890074e-01 2.94430107e-01 -1.00728199e-01
-1.13639843e+00 -7.92850971e-01 -7.36889958e-01 -5.09542286e-01
3.53459045e-02 -8.92442107e-01 9.22406673e-01 -9.99170303e-01
-1.44998896e+00 1.04935408e+00 2.11166069e-01 -5.56908667e-01
3.04933906e-01 -4.22452331e-01 -6.14507377e-01 3.36343557e-01
5.03114164e-01 5.17049849e-01 6.72644436e-01 -9.36010003e-01
-3.80245268e-01 -6.98550999e-01 -6.41008973e-01 1.32116422e-01
1.04169443e-01 -7.10588992e-02 -4.55992110e-02 -8.66954267e-01
5.11000119e-03 -1.08306527e+00 -3.78897250e-01 -2.82642365e-01
5.78854932e-03 9.31700617e-02 5.14773250e-01 -1.13790321e+00
1.21684003e+00 -2.16388035e+00 8.45675245e-02 4.35287744e-01
2.83713400e-01 1.68810412e-01 9.28530991e-02 4.12287973e-02
-4.96228665e-01 -1.47619411e-01 -5.16637444e-01 4.98139888e-01
-5.24655342e-01 9.81930718e-02 3.89670759e-01 4.98604536e-01
1.76641256e-01 6.46743417e-01 -8.59588265e-01 -7.21474111e-01
1.85523435e-01 1.27832338e-01 -6.25575185e-01 3.78188282e-01
1.94730386e-01 6.48685753e-01 -3.17634970e-01 7.38184810e-01
5.81749864e-02 -1.48790240e-01 2.82839179e-01 -1.72923267e-01
2.23419785e-01 -3.28370221e-02 -8.75823915e-01 1.77054012e+00
-4.15788561e-01 4.07774091e-01 -1.82416543e-01 -9.05366123e-01
1.08931994e+00 4.20330733e-01 9.56230819e-01 -6.09898329e-01
-9.61319078e-03 9.37393129e-01 5.75759888e-01 -7.02439964e-01
-2.36620590e-01 -3.98606509e-01 -2.25185245e-01 2.21280321e-01
2.61317223e-01 -8.43173563e-02 1.07785039e-01 -5.77903092e-02
1.16188347e+00 1.63869247e-01 4.51286793e-01 -4.12452132e-01
6.92607343e-01 4.96368855e-01 3.77038598e-01 6.04303002e-01
-4.20077860e-01 7.79626846e-01 4.31311131e-01 -4.85718042e-01
-1.35025311e+00 -1.52483439e+00 -3.42507482e-01 2.67587483e-01
-5.20558953e-01 -1.70969874e-01 -5.67024410e-01 -5.58463454e-01
1.02036618e-01 1.93316191e-01 -6.90042257e-01 -4.74038273e-01
-5.42583644e-01 -9.22180533e-01 6.34671509e-01 4.64924872e-01
3.87878686e-01 -6.89154625e-01 -7.96270788e-01 1.66804865e-01
1.25552518e-02 -4.92534310e-01 -4.43604589e-01 4.60020989e-01
-1.35052812e+00 -1.26297760e+00 -1.33298135e+00 -1.06491601e+00
3.70235056e-01 -6.12125933e-01 5.03082931e-01 -4.01095033e-01
-6.10311210e-01 5.20764232e-01 -2.13078372e-02 -1.43987596e-01
-3.23800147e-01 2.36088887e-01 3.09790373e-01 -8.24449584e-02
3.21191937e-01 -4.64935392e-01 -7.82255650e-01 2.52006620e-01
-7.76440740e-01 -2.36008272e-01 1.45403695e+00 1.10325873e+00
8.87643933e-01 -2.56710380e-01 7.51836121e-01 -6.59079015e-01
7.20976710e-01 -6.57199621e-01 2.93957759e-02 1.24834225e-01
-9.17906106e-01 1.05242588e-01 1.47324368e-01 -5.93693435e-01
-8.19307983e-01 2.17192873e-01 -7.06812143e-02 -1.81615770e-01
-1.90079272e-01 8.11319470e-01 1.22136839e-01 2.50578076e-01
6.96579576e-01 2.95547277e-01 8.39076877e-01 -5.98820448e-01
-5.82952127e-02 1.20612192e+00 6.01098478e-01 -6.50827050e-01
9.59030166e-02 2.29910105e-01 1.60566658e-01 -8.05530369e-01
-3.63797009e-01 -7.11844921e-01 -9.60563779e-01 -2.76429951e-01
9.70101178e-01 -8.84890556e-01 7.70409079e-03 6.59130096e-01
-1.53827652e-01 -4.70949590e-01 -3.43025029e-01 1.12475801e+00
-8.47084820e-01 3.31522763e-01 -6.07285857e-01 -3.64920646e-01
-4.76879448e-01 -1.61616969e+00 1.00043297e+00 -1.59127370e-01
-5.49987912e-01 -7.34469473e-01 2.88970500e-01 4.20833796e-01
4.93332177e-01 1.69300199e-01 1.44762290e+00 -1.03033817e+00
2.22699836e-01 -3.89865458e-01 3.76619473e-02 4.93940741e-01
4.03616130e-01 -5.95062494e-01 -2.27399766e-01 7.29231611e-02
7.70055428e-02 -3.76085699e-01 4.39014822e-01 5.00345767e-01
7.42932796e-01 2.17837840e-01 -1.14558168e-01 8.72487277e-02
1.28773081e+00 4.62466180e-01 8.40329528e-01 8.03944528e-01
1.75675780e-01 3.79602700e-01 8.01423728e-01 -1.40639665e-02
1.15408599e-01 9.63980675e-01 -2.09474221e-01 1.57637998e-01
-6.21212840e-01 1.32223636e-01 2.73099661e-01 9.28786516e-01
-3.20431679e-01 6.40718937e-01 -1.28110790e+00 8.65434706e-01
-1.31918871e+00 -2.87537783e-01 -3.37933421e-01 2.30706954e+00
9.95684505e-01 1.38991669e-01 3.37304950e-01 2.66680002e-01
6.02626801e-01 -3.65370512e-01 -2.99659461e-01 -3.57949495e-01
4.97788563e-03 7.19365239e-01 5.33086419e-01 1.38639659e-01
-9.24645662e-01 4.56949592e-01 6.40025854e+00 7.40051985e-01
-1.28771341e+00 3.98784488e-01 4.74315763e-01 -1.61904782e-01
3.16398442e-01 -4.91957515e-02 6.79566264e-02 6.34448290e-01
1.39291823e+00 4.35157940e-02 2.02251635e-02 9.10064697e-01
1.88649133e-01 -5.81855357e-01 -6.63819790e-01 5.00522435e-01
-8.95927474e-02 -1.08383453e+00 1.50004357e-01 2.53414989e-01
3.40001553e-01 1.05761224e-03 -2.35751420e-01 3.37790519e-01
-1.70566477e-02 -7.72691607e-01 3.07543367e-01 6.78750277e-01
1.10137820e+00 -4.62752670e-01 1.21632612e+00 -3.87464017e-01
-7.59375870e-01 1.38924206e-02 1.76956639e-01 1.97290391e-01
1.23817876e-01 6.38931513e-01 -1.03019178e+00 5.60144186e-01
6.67735219e-01 3.75609338e-01 -7.24119008e-01 1.30270636e+00
1.04885422e-01 8.27186465e-01 -2.00509831e-01 2.79331535e-01
3.70378308e-02 -2.95473449e-02 6.22946918e-01 8.58391941e-01
1.96158335e-01 -1.13634087e-01 6.05988875e-02 1.54823452e-01
4.66555595e-01 2.93960184e-01 -1.93288162e-01 -2.03142762e-01
1.78040311e-01 7.32991874e-01 -7.57991314e-01 -3.49104941e-01
-2.60557711e-01 7.28404164e-01 -2.05723226e-01 9.89010558e-02
-4.35960770e-01 -4.54517841e-01 2.50048131e-01 5.36565483e-01
4.09038179e-02 -2.49518901e-01 -4.69711006e-01 -7.47841895e-01
-2.58282661e-01 -1.06163323e+00 8.24152827e-01 -9.47244585e-01
-1.21541989e+00 3.73831332e-01 9.34484825e-02 -1.30016506e+00
-1.82640821e-01 -7.52992511e-01 1.88656583e-01 8.58631253e-01
-7.38974333e-01 -9.91313398e-01 6.18860312e-03 2.59702355e-01
5.24030447e-01 -5.75000286e-01 9.07127500e-01 4.18650627e-01
-1.06881440e-01 4.77880016e-02 1.34833708e-01 3.80576462e-01
1.01290441e+00 -1.22529554e+00 -6.49906695e-01 2.04699785e-01
-4.91396725e-01 7.78409839e-01 3.73402834e-01 -1.19756615e+00
-1.27704227e+00 -8.64779830e-01 7.08741128e-01 -2.72620171e-01
9.29298341e-01 1.58853069e-01 -5.01114905e-01 4.94592398e-01
-3.16948116e-01 7.09383376e-03 1.11106598e+00 -2.85872668e-01
9.45741311e-02 1.01118043e-01 -1.14184237e+00 3.43434840e-01
5.12141109e-01 -3.87178749e-01 -1.07766974e+00 1.30354032e-01
3.23963873e-02 -3.44278425e-01 -1.90746319e+00 6.24718308e-01
1.20974052e+00 -4.71014649e-01 1.00782239e+00 -8.41902792e-01
6.79470658e-01 -2.28380442e-01 -2.87038624e-01 -1.04843283e+00
-9.60719585e-02 1.72830641e-01 7.53957480e-02 5.34555256e-01
2.88673729e-01 -5.71139514e-01 5.88556826e-01 2.68719018e-01
-5.49417257e-01 -1.14321983e+00 -1.19430900e+00 -7.43196547e-01
4.46231812e-02 -2.26215765e-01 6.10151403e-02 6.67043746e-01
-1.94783524e-01 1.17132008e-01 -1.70996398e-01 1.31118158e-02
4.60957140e-01 -2.41376311e-01 4.78469759e-01 -1.21108246e+00
-2.59713382e-01 -1.92191005e-01 -9.91307378e-01 -1.35613024e-01
-4.31569427e-01 -1.12337244e+00 -2.51263171e-01 -1.50292790e+00
2.48318940e-01 -6.80819690e-01 -4.52829450e-01 8.89628083e-02
8.79499782e-03 5.68984330e-01 -2.20344648e-01 7.77095795e-01
-6.38380796e-02 2.87492484e-01 1.15179586e+00 -8.13625082e-02
-2.95074910e-01 -2.99309231e-02 -1.09224245e-01 2.13907927e-01
5.75433195e-01 -7.68283486e-01 -3.31924886e-01 6.28623888e-02
-2.92801350e-01 4.78027351e-02 6.12822592e-01 -1.29781234e+00
-3.38599950e-01 1.53314903e-01 7.54023254e-01 -2.03338489e-01
1.66186064e-01 -5.33892453e-01 4.85583931e-01 1.17504883e+00
-2.14497879e-01 -7.61594921e-02 2.57145651e-02 4.72723275e-01
-3.07936728e-01 -2.24181190e-01 7.49815404e-01 -2.90495127e-01
-9.09047067e-01 -1.99497506e-01 -8.66379201e-01 1.86275840e-02
1.15828633e+00 -5.86417735e-01 1.26595289e-01 -3.28847324e-03
-1.52609515e+00 -2.33121008e-01 4.14124697e-01 3.53183270e-01
5.86238146e-01 -1.34513617e+00 -5.21687627e-01 2.28381827e-01
3.03375959e-01 -2.68341064e-01 2.84471840e-01 1.65701520e+00
-7.29847193e-01 5.73319793e-01 -7.74315536e-01 -8.76938462e-01
-1.02972877e+00 3.57345581e-01 5.67764819e-01 -5.65260231e-01
-7.61911750e-01 3.86087298e-01 -2.84545779e-01 -4.59913284e-01
-2.54467338e-01 5.61914816e-02 -1.44551665e-01 2.23668385e-02
1.31762877e-01 2.64272511e-01 3.36397409e-01 -7.24428296e-01
-4.42029864e-01 3.88259530e-01 -1.18533216e-01 -3.27493370e-01
1.37461793e+00 4.19596471e-02 -9.62731689e-02 7.41907597e-01
1.12012446e+00 1.26438150e-02 -6.44554973e-01 9.73440260e-02
3.59302640e-01 -2.53386647e-01 2.36823842e-01 -1.19601738e+00
-9.30673182e-01 5.51644504e-01 1.41257679e+00 -4.70847011e-01
1.09265316e+00 1.36983946e-01 7.72804677e-01 -1.78355947e-01
4.93214309e-01 -1.07080793e+00 2.32526343e-02 7.31088892e-02
8.72234344e-01 -8.11177671e-01 8.81947279e-02 -2.23391175e-01
-9.24148321e-01 1.15791619e+00 2.93680251e-01 -3.98779988e-01
6.63587093e-01 4.26106080e-02 2.54826009e-01 -5.18434823e-01
-3.21257085e-01 9.33477730e-02 4.26017582e-01 8.96584332e-01
4.30799782e-01 1.63447335e-01 -1.03733528e+00 9.83106136e-01
-1.46585166e-01 7.45978653e-01 2.45113432e-01 1.19207847e+00
-1.13041572e-01 -1.19534516e+00 -3.52408499e-01 1.18145943e+00
-5.06522357e-01 1.71622485e-02 -2.57741839e-01 1.07400727e+00
1.53898343e-01 3.69022697e-01 1.77870002e-02 -3.45935792e-01
1.57254189e-01 4.26000595e-01 3.84731084e-01 -6.94436073e-01
-5.01317441e-01 1.43277675e-01 4.60576028e-01 -3.05057824e-01
-3.76988411e-01 -8.28692853e-01 -1.30182076e+00 7.76654243e-01
-8.15793648e-02 3.23509842e-01 5.72015464e-01 9.64552522e-01
4.93960291e-01 6.94492936e-01 2.11325973e-01 -3.64618629e-01
-4.43005562e-01 -1.28723943e+00 -7.01537549e-01 2.63663232e-01
-6.86672181e-02 -1.15656960e+00 -2.21520379e-01 2.62654364e-01] | [14.532155990600586, -1.7705219984054565] |
bcaeff3c-230f-4ec5-9443-86444b89ea4c | an-aerial-image-recognition-framework-using | 1410.2188 | null | http://arxiv.org/abs/1410.2188v1 | http://arxiv.org/pdf/1410.2188v1.pdf | An Aerial Image Recognition Framework using Discrimination and Redundancy Quality Measure | Aerial image categorization plays an indispensable role in remote sensing and
artificial intelligence. In this paper, we propose a new aerial image
categorization framework, focusing on organizing the local patches of each
aerial image into multiple discriminative subgraphs. The subgraphs reflect both
the geometric property and the color distribution of an aerial image. First,
each aerial image is decomposed into a collection of regions in terms of their
color intensities. Thereby region connected graph (RCG), which models the
connection between the spatial neighboring regions, is constructed to encode
the spatial context of an aerial image. Second, a subgraph mining technique is
adopted to discover the frequent structures in the RCGs constructed from the
training aerial images. Thereafter, a set of refined structures are selected
among the frequent ones toward being highly discriminative and low redundant.
Lastly, given a new aerial image, its sub-RCGs corresponding to the refined
structures are extracted. They are further quantized into a discriminative
vector for SVM classification. Thorough experimental results validate the
effectiveness of the proposed method. In addition, the visualized mined
subgraphs show that the discriminative topologies of each aerial image are
discovered. | ['Luming Zhang', 'Yuxin Hu'] | 2014-10-06 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 3.42653036e-01 -1.96304813e-01 -2.18475133e-01 -3.02438319e-01
3.07182241e-02 -7.38497853e-01 1.16603389e-01 4.31844264e-01
2.39477530e-01 4.08930361e-01 9.22715068e-02 -8.30154568e-02
-6.17236137e-01 -1.19097984e+00 -1.69349551e-01 -8.58111084e-01
-2.50606835e-01 -2.75322109e-01 3.01173449e-01 -2.17713583e-02
1.31065086e-01 6.14378631e-01 -1.58611310e+00 1.02042079e-01
8.78945589e-01 1.03730130e+00 4.45108593e-01 1.33319408e-01
2.93418411e-02 5.53590536e-01 -4.17422861e-01 3.18435520e-01
2.73535222e-01 -6.75454974e-01 -6.66297436e-01 9.09747243e-01
1.52962625e-01 -1.15658239e-01 -2.99294591e-01 1.22647107e+00
-1.98516682e-01 2.81059712e-01 5.05681813e-01 -9.81541038e-01
-4.45058912e-01 4.34775740e-01 -8.92299712e-01 3.45936209e-01
-1.98793190e-04 -2.06444845e-01 1.16551089e+00 -7.31955171e-01
4.68166888e-01 9.07243848e-01 2.33416781e-02 -2.73306221e-01
-8.57345939e-01 -4.96957242e-01 7.85682797e-01 2.40563393e-01
-1.63825107e+00 1.58434570e-01 9.48639691e-01 -3.14161628e-01
3.21737140e-01 6.55587435e-01 1.09564447e+00 -2.96046920e-02
1.60204262e-01 5.25386512e-01 9.77107644e-01 -1.53434783e-01
2.31621757e-01 7.15143234e-02 2.87792087e-01 1.02985930e+00
3.13574791e-01 -5.29149473e-01 -1.23115806e-02 -1.09919161e-01
4.71153110e-01 4.91872638e-01 -5.56611836e-01 -4.68318969e-01
-8.71654510e-01 7.47301340e-01 9.90231574e-01 4.41115737e-01
-4.70396906e-01 -3.68726999e-01 1.92115232e-01 8.45368952e-02
4.10343140e-01 8.87900218e-02 -1.90014988e-01 8.99050653e-01
-8.78817737e-01 -2.12532744e-01 1.62935928e-01 7.99515843e-01
1.39164245e+00 -1.66374296e-01 1.61581278e-01 7.10366905e-01
4.93473083e-01 1.97935045e-01 1.29097328e-01 -3.72353345e-01
5.57388783e-01 1.64794922e+00 -1.89856619e-01 -2.02682948e+00
-3.78847212e-01 -4.98212576e-01 -1.05838215e+00 -3.49350572e-01
-3.56921017e-01 1.04068249e-01 -6.63672805e-01 1.12769067e+00
6.63883209e-01 1.33212984e-01 -1.54975662e-02 1.12364161e+00
7.54768848e-01 9.86493945e-01 9.97950435e-02 -2.84630388e-01
1.20298755e+00 -5.32932162e-01 -1.34104893e-01 -1.86546534e-01
1.56974822e-01 -3.25824648e-01 7.61272073e-01 2.07061172e-01
-3.82496119e-01 -6.03852689e-01 -1.21559989e+00 5.16377389e-01
-2.23830789e-01 7.27014661e-01 7.08673537e-01 7.05580413e-02
-7.18581796e-01 2.37006038e-01 -4.20401096e-01 -4.20734018e-01
2.88694829e-01 1.61669686e-01 -4.08818215e-01 -2.79910922e-01
-8.84508729e-01 8.71856362e-02 7.01177597e-01 3.57982367e-01
-6.67582810e-01 -1.13663867e-01 -9.30106223e-01 8.37501809e-02
4.02084112e-01 -3.92998382e-02 2.53892779e-01 -1.23958862e+00
-7.20733762e-01 6.71991467e-01 -5.47455214e-02 -7.06881061e-02
-2.59252399e-01 4.92994547e-01 -6.39064372e-01 5.94161928e-01
1.67589888e-01 2.52641767e-01 9.86506224e-01 -1.31980717e+00
-1.12381411e+00 -5.63974380e-01 3.16564560e-01 4.55164999e-01
-5.59919238e-01 -4.24345702e-01 -4.13499624e-01 -6.50071144e-01
7.88769126e-01 -7.46604025e-01 -2.24183455e-01 -4.57224309e-01
-6.37206137e-01 -1.74428239e-01 1.14132726e+00 -5.75661123e-01
1.52890992e+00 -2.38674974e+00 4.90653604e-01 9.91944611e-01
3.82953078e-01 -1.77811265e-01 -1.45828858e-01 3.81711572e-01
-1.55365601e-01 1.30994037e-01 -5.75345457e-01 7.24036992e-01
-5.41678905e-01 1.14230521e-01 -3.10755581e-01 5.23270547e-01
3.13808411e-01 3.07310760e-01 -8.49766254e-01 -6.83267117e-01
2.29829848e-01 -1.40639096e-01 -6.61332607e-02 4.21318300e-02
-6.36158586e-02 4.90480810e-01 -1.14188910e+00 7.31571794e-01
7.15587139e-01 -2.04554543e-01 5.86908877e-01 -4.68307376e-01
-3.79838347e-01 -2.66050994e-01 -1.08051264e+00 1.10258925e+00
-1.14229918e-01 1.79947346e-01 -1.73038971e-02 -1.28263342e+00
1.29335892e+00 -2.44553342e-01 6.90305889e-01 -2.32024983e-01
3.90218459e-02 -1.95707038e-01 -2.65030354e-01 -4.03649122e-01
5.00650048e-01 1.66810110e-01 -2.56493390e-01 2.82133013e-01
-2.38492295e-01 7.60306641e-02 3.06260794e-01 2.63655096e-01
6.05125487e-01 -8.37695301e-02 4.56503630e-01 -6.16704643e-01
8.58716726e-01 5.08823276e-01 5.46476066e-01 4.32043523e-02
1.36658445e-01 8.98861960e-02 2.90839374e-01 -7.60729730e-01
-6.00877047e-01 -5.74771345e-01 -1.10888205e-01 7.64386475e-01
7.09202290e-01 -4.08714652e-01 -5.71296155e-01 -7.20874667e-01
4.07986678e-02 1.22948892e-01 -6.42429709e-01 -1.97203904e-01
-1.65971458e-01 -8.15994501e-01 4.80692387e-02 1.29764110e-01
8.56557786e-01 -6.93790853e-01 -7.42643178e-01 1.41338110e-01
-3.31499070e-01 -6.74166143e-01 -2.64060289e-01 -1.42878115e-01
-1.00279975e+00 -1.37767220e+00 -2.44573265e-01 -1.04127336e+00
1.12510252e+00 1.18164694e+00 5.20315707e-01 3.93537492e-01
-3.84747565e-01 2.76370078e-01 -7.50302553e-01 1.38205066e-01
9.84241217e-02 -8.02596807e-02 -1.14862069e-01 6.21793151e-01
1.61021858e-01 -7.03921676e-01 -7.39058137e-01 4.60109025e-01
-8.99254918e-01 -3.39809842e-02 8.06311250e-01 5.73679507e-01
1.10359895e+00 9.84811187e-01 -3.20526958e-02 -4.75628227e-01
2.22610921e-01 -8.33536685e-01 -7.31866002e-01 6.13570392e-01
-9.60528031e-02 -2.13293627e-01 8.49959254e-01 -7.66966045e-02
-8.14393580e-01 3.78638029e-01 7.61021554e-01 -2.52714276e-01
-2.71889925e-01 9.94953394e-01 -2.87705064e-01 -5.98397590e-02
2.21784353e-01 6.18462801e-01 -3.08966637e-01 -2.22561941e-01
2.18684435e-01 6.63099289e-01 1.37759089e-01 -2.33682618e-01
8.68058980e-01 4.33245867e-01 1.34218782e-01 -1.29126322e+00
-6.61965430e-01 -5.47472537e-01 -6.99784935e-01 -4.40905839e-01
8.90480340e-01 -9.67606366e-01 -3.14022630e-01 7.51975784e-03
-6.28285468e-01 2.22284913e-01 4.70150411e-02 4.35156912e-01
8.68382156e-02 6.86372399e-01 -1.43789187e-01 -6.62859440e-01
-1.01509355e-01 -8.71252060e-01 8.22501421e-01 5.66700220e-01
3.57496813e-02 -4.46265340e-01 -9.01803672e-02 3.35052870e-02
-3.17292035e-01 4.18459356e-01 1.29927683e+00 -3.01100999e-01
-6.55543208e-01 -9.50609520e-02 -3.76064271e-01 7.33910203e-02
6.47791386e-01 3.91275823e-01 -3.14305454e-01 -4.15867597e-01
-1.97354257e-01 -7.26575032e-02 8.19084048e-01 3.78133148e-01
1.24679327e+00 -5.77465236e-01 -6.47801280e-01 5.16068220e-01
1.69690275e+00 4.96744126e-01 1.87987208e-01 1.83778703e-01
8.11855495e-01 7.03203321e-01 9.88770843e-01 5.40689349e-01
1.41467661e-01 2.30457082e-01 6.17389023e-01 -3.13736171e-01
5.28155625e-01 -3.43535900e-01 1.05020024e-01 7.42884815e-01
-3.37518692e-01 -1.02673300e-01 -6.32740259e-01 3.16707879e-01
-1.64907885e+00 -9.60883796e-01 -2.91878611e-01 1.91879427e+00
2.44984284e-01 -2.32988164e-01 8.80241022e-02 2.39303589e-01
9.60009694e-01 4.83382344e-01 -5.86179852e-01 7.65657201e-02
-1.64342850e-01 -4.52868193e-02 2.58037925e-01 1.43267855e-01
-1.33974457e+00 8.59762788e-01 4.85321474e+00 5.40490806e-01
-1.04065239e+00 -5.63313961e-01 7.19072223e-01 3.92674834e-01
-3.63984942e-01 2.31859669e-01 -5.24209261e-01 3.21617156e-01
1.52731240e-01 -1.27622798e-01 2.84718335e-01 1.11878073e+00
1.43062472e-01 -4.10030901e-01 -2.07328513e-01 6.61506355e-01
-2.20798887e-02 -9.56773520e-01 6.34677887e-01 1.54661313e-01
7.39107966e-01 -3.63907278e-01 -6.55235052e-02 -2.15864733e-01
2.34717667e-01 -6.28993213e-01 5.22813499e-01 3.97400320e-01
8.56496453e-01 -1.05387902e+00 5.72240710e-01 3.00005049e-01
-2.18401384e+00 -3.71858418e-01 -8.19456756e-01 2.13409081e-01
-5.86751103e-01 5.44185996e-01 -4.37289685e-01 1.16421115e+00
8.08720469e-01 1.22791505e+00 -6.99440897e-01 6.69003487e-01
-1.63123742e-01 4.96372938e-01 -9.94996876e-02 -1.50386438e-01
6.12303793e-01 -9.88043249e-01 3.71206611e-01 7.88801610e-01
4.24442649e-01 7.20053613e-01 7.44500220e-01 5.47593653e-01
2.36245722e-01 4.46552098e-01 -9.16300356e-01 -5.04362211e-02
4.92146015e-01 1.59569788e+00 -1.20433187e+00 -1.61387816e-01
-2.98315853e-01 9.45816934e-01 1.12672649e-01 3.17154139e-01
-6.26066148e-01 -4.84223932e-01 4.25848275e-01 1.09306211e-02
3.98607254e-01 -2.50449806e-01 3.91291678e-01 -9.80354190e-01
2.39276495e-02 -7.63629913e-01 6.78430796e-01 -5.58831334e-01
-9.32349920e-01 8.12542617e-01 1.54840410e-01 -1.73061585e+00
4.57989275e-01 -2.58073479e-01 -6.86983407e-01 6.51695371e-01
-1.32765782e+00 -1.17793643e+00 -8.72804284e-01 7.85368264e-01
4.72834885e-01 -8.96911770e-02 8.52600932e-01 -3.06249619e-01
-8.35972905e-01 -1.57489091e-01 4.30069901e-02 1.89365178e-01
-1.44642562e-01 -8.48779380e-01 -1.66844070e-01 1.13973272e+00
2.60500640e-01 5.99483430e-01 4.98443395e-02 -6.86010778e-01
-1.30959702e+00 -1.54396772e+00 4.63025659e-01 3.65508139e-01
5.43239415e-01 1.56562626e-01 -7.11560190e-01 5.03726065e-01
-1.25066534e-01 -8.36019218e-02 7.64844060e-01 -3.40731114e-01
-1.52117655e-01 -4.63303506e-01 -9.61445868e-01 5.36618054e-01
9.19931650e-01 -2.11052462e-01 -4.05877590e-01 3.65443200e-01
4.10463154e-01 6.36325628e-02 -8.11794043e-01 2.50641733e-01
1.98186025e-01 -8.22409928e-01 7.76852965e-01 -5.52453160e-01
5.17451644e-01 -8.19747031e-01 -2.70724773e-01 -1.33859396e+00
-8.57242823e-01 1.17783673e-01 7.15219021e-01 1.09149146e+00
1.49551630e-01 -3.74578536e-01 4.95471209e-01 -2.31926724e-01
-1.69659685e-02 -6.27746165e-01 -5.84665358e-01 -4.24941570e-01
-6.24331534e-01 2.15781614e-01 8.43827426e-01 9.66016233e-01
9.49254632e-02 3.84133637e-01 -9.42771062e-02 6.22041285e-01
5.71056783e-01 9.98364568e-01 3.61828655e-01 -1.45037282e+00
-3.93324308e-02 -1.85400695e-01 -6.38135791e-01 -9.25339401e-01
2.62500029e-02 -1.07568944e+00 -6.80053383e-02 -1.64660645e+00
3.30711693e-01 -6.13296866e-01 -3.25864762e-01 6.38734162e-01
-1.90729216e-01 2.83373803e-01 -1.27463758e-01 5.19855857e-01
-5.34852326e-01 5.21643400e-01 1.02588046e+00 -5.57677448e-01
-4.80721951e-01 2.34600324e-02 -6.75740063e-01 5.77126026e-01
1.00776386e+00 -1.82262093e-01 -6.41989708e-01 -1.80668890e-01
-1.62377939e-01 5.91666810e-02 3.60521317e-01 -1.04890597e+00
-8.55399519e-02 -4.99524057e-01 5.42259753e-01 -6.91713750e-01
-1.84147507e-01 -1.07509971e+00 3.77451837e-01 7.19835103e-01
-1.79735497e-02 3.20051834e-02 1.42165169e-01 7.56536067e-01
-4.78101969e-01 1.73706058e-02 6.73831940e-01 -3.03808242e-01
-1.09881377e+00 4.73313600e-01 -3.67384911e-01 -3.49099010e-01
1.43503070e+00 -3.41680914e-01 6.94745220e-04 1.10013057e-02
-5.99376798e-01 2.73269296e-01 3.59897017e-01 1.41474307e-01
1.00189054e+00 -1.18941081e+00 -6.35017157e-01 3.92078429e-01
5.96913636e-01 5.90162426e-02 5.74248910e-01 5.59923947e-01
-5.27045906e-01 2.80794889e-01 -3.62210244e-01 -6.48291171e-01
-1.68983984e+00 5.12069762e-01 9.28723812e-02 -1.30922467e-01
-8.06358159e-01 7.28986442e-01 2.16563806e-01 1.97436914e-01
-1.50418639e-01 -5.43102741e-01 -7.83596277e-01 3.44795048e-01
2.43347377e-01 1.57170475e-01 -9.48433354e-02 -9.60853875e-01
-5.27613044e-01 9.51389968e-01 2.31792629e-01 3.86491448e-01
1.37221527e+00 -2.15986729e-01 -6.78335965e-01 5.99000305e-02
1.03289700e+00 1.43745348e-01 -9.65004027e-01 -1.43364191e-01
-1.16462842e-01 -5.01610398e-01 8.01634267e-02 -2.32104003e-01
-1.36562121e+00 5.43131828e-01 3.69957924e-01 3.80605459e-01
1.61089098e+00 5.03664687e-02 3.35099280e-01 3.59512478e-01
6.07558966e-01 -5.91879785e-01 -9.82012823e-02 7.49025792e-02
7.81527877e-01 -7.51515269e-01 3.29512268e-01 -6.94078982e-01
-6.98650956e-01 1.19634843e+00 4.34530675e-01 -2.02529714e-01
7.02767789e-01 -3.66219580e-01 -3.12901944e-01 -4.53116477e-01
-6.70018643e-02 -4.17920262e-01 4.57406580e-01 3.98572892e-01
-1.23146109e-01 2.66244143e-01 -3.36318970e-01 4.34337169e-01
-9.36104506e-02 -4.52831745e-01 1.66713014e-01 6.88759029e-01
-8.33810091e-01 -6.22619271e-01 -2.95332253e-01 5.49467444e-01
2.16134876e-01 1.84129238e-01 -7.43562579e-01 7.42482305e-01
4.14632678e-01 1.10062492e+00 1.00826763e-01 -1.04195809e+00
4.12896574e-01 -4.26911592e-01 1.72424659e-01 -7.47718036e-01
-2.58884281e-01 -1.31722257e-01 -3.72543961e-01 -3.64535332e-01
-4.98287886e-01 -2.69385487e-01 -1.32438445e+00 7.80716166e-03
-2.98078626e-01 6.16365194e-01 1.58146974e-02 6.65696740e-01
2.27568269e-01 3.99429381e-01 1.38903809e+00 -5.92058599e-01
7.56244361e-03 -4.96811241e-01 -1.10458446e+00 1.57523587e-01
5.20353988e-02 -5.42516530e-01 -3.71098965e-01 1.22765087e-01] | [10.024248123168945, -1.5069433450698853] |
3ed6c174-ecc2-41b1-9a88-fa21c70e84b2 | actor-mimic-deep-multitask-and-transfer | 1511.06342 | null | http://arxiv.org/abs/1511.06342v4 | http://arxiv.org/pdf/1511.06342v4.pdf | Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning | The ability to act in multiple environments and transfer previous knowledge
to new situations can be considered a critical aspect of any intelligent agent.
Towards this goal, we define a novel method of multitask and transfer learning
that enables an autonomous agent to learn how to behave in multiple tasks
simultaneously, and then generalize its knowledge to new domains. This method,
termed "Actor-Mimic", exploits the use of deep reinforcement learning and model
compression techniques to train a single policy network that learns how to act
in a set of distinct tasks by using the guidance of several expert teachers. We
then show that the representations learnt by the deep policy network are
capable of generalizing to new tasks with no prior expert guidance, speeding up
learning in novel environments. Although our method can in general be applied
to a wide range of problems, we use Atari games as a testing environment to
demonstrate these methods. | ['Ruslan Salakhutdinov', 'Jimmy Lei Ba', 'Emilio Parisotto'] | 2015-11-19 | null | null | null | null | ['transfer-reinforcement-learning'] | ['methodology'] | [ 4.96740907e-01 3.00493568e-01 4.29843329e-02 -2.25882366e-01
-4.46062386e-01 -8.71655047e-01 7.52662599e-01 -9.70146507e-02
-5.54056346e-01 1.01383674e+00 -2.66676039e-01 -2.95483232e-01
-4.44025069e-01 -7.60597348e-01 -8.01474214e-01 -7.68000722e-01
-3.36601660e-02 9.07922626e-01 4.66892898e-01 -5.56341290e-01
1.18033543e-01 5.06980956e-01 -1.65432262e+00 1.79041624e-01
9.61359322e-01 2.84573972e-01 4.92032498e-01 8.29398215e-01
1.93581581e-01 1.04063141e+00 -8.29620302e-01 -6.67888969e-02
3.40232879e-01 -3.32413167e-01 -1.18044722e+00 2.65216291e-01
-6.36804895e-03 -5.48975348e-01 -2.29080454e-01 6.30224347e-01
3.55029076e-01 8.16078484e-01 6.65625989e-01 -1.06912029e+00
-6.17757320e-01 5.06975770e-01 -2.03062624e-01 2.89839655e-01
3.66460323e-01 3.47312599e-01 5.26067615e-01 -2.26991534e-01
3.16807330e-01 1.34955692e+00 2.52924562e-01 9.76445436e-01
-1.28575933e+00 -4.97702897e-01 2.65011460e-01 1.68136075e-01
-6.46761417e-01 -1.67153418e-01 5.65187871e-01 -2.31691688e-01
9.96690214e-01 -1.43708393e-01 6.05786800e-01 1.24941409e+00
1.99176028e-01 9.42484260e-01 1.49342167e+00 -6.65212929e-01
3.56483221e-01 1.69728249e-01 -3.43992710e-01 7.00791121e-01
1.47342324e-01 3.05991471e-01 -4.09600973e-01 -1.19427480e-01
6.88528538e-01 3.78700681e-02 4.99366000e-02 -5.71100116e-01
-9.22278106e-01 7.06328988e-01 9.86233950e-02 3.25744033e-01
-4.64202195e-01 4.54489857e-01 3.29218149e-01 1.04862106e+00
2.88162529e-01 8.79550338e-01 -6.45907640e-01 -2.55957007e-01
-1.26092061e-01 4.04961020e-01 8.96966457e-01 6.52183950e-01
9.15760696e-01 4.23443735e-01 8.81804749e-02 6.17553174e-01
-9.59282666e-02 5.91843367e-01 7.97753870e-01 -1.45017803e+00
2.70165324e-01 3.78417015e-01 1.56251788e-01 -2.40984604e-01
-2.50726402e-01 -1.96509123e-01 -2.33616278e-01 6.85530007e-01
7.36400113e-02 -5.67816019e-01 -9.52878356e-01 1.76995540e+00
5.10995150e-01 3.08749586e-01 4.48996931e-01 3.63673657e-01
7.81099424e-02 6.49699211e-01 1.79076046e-01 -5.03982827e-02
7.70626605e-01 -9.25567687e-01 -2.21401066e-01 -5.74179769e-01
6.36750102e-01 -1.20545536e-01 8.50805759e-01 6.61732495e-01
-1.37041128e+00 -6.57912374e-01 -8.77858222e-01 3.45798701e-01
-4.27340835e-01 -6.32744670e-01 6.09664977e-01 4.63481575e-01
-1.23379934e+00 6.27898932e-01 -7.88049579e-01 -2.98639208e-01
3.39403450e-01 6.08242273e-01 -1.98190853e-01 -2.03434423e-01
-1.12631464e+00 1.32507718e+00 6.72881305e-01 -6.87667012e-01
-1.71160114e+00 -3.82233262e-01 -7.53867269e-01 3.00706118e-01
8.74060154e-01 -9.06003535e-01 1.94469106e+00 -1.33983362e+00
-2.06371117e+00 4.62004423e-01 2.63718218e-01 -4.56241071e-01
2.47198895e-01 -8.96570086e-02 1.24040274e-02 1.92581519e-01
4.44503464e-02 6.22066021e-01 9.99108076e-01 -1.28776777e+00
-1.08020985e+00 -3.54010224e-01 5.50934911e-01 7.60430694e-01
-4.67360944e-01 -6.64542392e-02 1.39760986e-01 -4.49326932e-01
-4.30875927e-01 -1.04248738e+00 -4.65828121e-01 -6.11693501e-01
4.33699995e-01 -7.28418052e-01 7.63050079e-01 -2.34250911e-02
4.76919353e-01 -1.91693282e+00 5.96125543e-01 2.79071540e-01
2.72619903e-01 4.91421551e-01 -4.08179492e-01 4.47167099e-01
5.56585938e-02 -2.90652663e-02 9.75772813e-02 -4.60178368e-02
-1.50287421e-02 6.62789345e-01 -6.72061294e-02 -6.65587559e-03
1.05786338e-01 9.67982411e-01 -1.18423223e+00 2.67774779e-02
8.26779082e-02 2.22548516e-03 -5.68692148e-01 5.26401758e-01
-5.96993566e-01 8.01497877e-01 -8.74259055e-01 9.62470993e-02
-9.28950310e-02 -2.52785772e-01 5.10875642e-01 9.15753543e-01
3.03366631e-01 1.89978063e-01 -1.14432883e+00 1.72829878e+00
-8.11795056e-01 2.64193296e-01 2.75205284e-01 -1.44816983e+00
8.13250721e-01 6.21236503e-01 4.46998805e-01 -7.46399939e-01
2.27563232e-02 4.63087708e-02 3.42734218e-01 -6.86192036e-01
2.09439863e-02 -2.85896122e-01 -1.76417232e-01 1.05081439e+00
3.55834603e-01 -3.82774860e-01 1.93843558e-01 -3.14825145e-03
1.37782848e+00 1.93245605e-01 4.01125371e-01 -1.20107658e-01
3.00971806e-01 9.67792124e-02 4.62538898e-01 1.22690797e+00
-7.58922100e-02 -4.25529122e-01 2.12338358e-01 -8.00470293e-01
-8.86097372e-01 -1.09052372e+00 5.76091349e-01 1.98055172e+00
-1.48419499e-01 9.22059640e-02 -4.92263377e-01 -1.00468516e+00
1.35900944e-01 8.14544141e-01 -6.84666514e-01 -4.74664539e-01
-7.70724356e-01 -1.79274842e-01 3.31288129e-01 4.65817153e-01
5.81402898e-01 -1.43277168e+00 -1.26887310e+00 4.05933619e-01
4.06218141e-01 -8.36494923e-01 -9.29276347e-02 5.86066425e-01
-8.03953588e-01 -1.09014153e+00 -6.19230986e-01 -7.99674094e-01
5.00074446e-01 3.50574344e-01 1.05854380e+00 9.39134508e-02
5.74787483e-02 1.06456745e+00 -3.39614064e-01 -6.41749680e-01
-9.64411736e-01 2.95673877e-01 3.78080726e-01 -3.13481182e-01
2.07578883e-01 -9.72573817e-01 -1.04691520e-01 2.27356017e-01
-1.07721555e+00 1.31056666e-01 8.28267753e-01 1.02535582e+00
-4.22777012e-02 3.37578297e-01 7.83746779e-01 -1.03630078e+00
1.10608232e+00 -6.22759640e-01 -7.16320813e-01 5.53819120e-01
-7.14367807e-01 4.47821379e-01 9.80630577e-01 -7.43468285e-01
-1.34718454e+00 9.44749340e-02 3.10529768e-01 -2.60630786e-01
-5.24726629e-01 3.64285171e-01 6.19890578e-02 -2.89197683e-01
8.69490087e-01 4.37223285e-01 2.18077406e-01 -3.40323955e-01
4.92303222e-01 4.61188674e-01 3.86701494e-01 -1.24987817e+00
8.11316967e-01 8.56416449e-02 2.20768303e-02 -3.07470232e-01
-9.06151712e-01 -1.27555609e-01 -4.88995194e-01 -1.12771787e-01
6.44933641e-01 -6.62204802e-01 -1.04656637e+00 3.17128390e-01
-8.15846920e-01 -9.92106616e-01 -4.97269869e-01 2.41243288e-01
-1.00857770e+00 2.66271904e-02 -2.51599908e-01 -6.92149162e-01
1.22046478e-01 -1.30109549e+00 3.84415150e-01 5.38297236e-01
2.07315236e-01 -1.33108723e+00 3.35605741e-01 3.01454514e-01
5.32374620e-01 -2.41485640e-01 9.77958560e-01 -1.30174816e+00
-3.07568580e-01 2.97283888e-01 4.26878810e-01 3.16353828e-01
3.01578641e-01 -6.56917930e-01 -8.40227485e-01 -6.13636792e-01
2.27217987e-01 -1.10864007e+00 6.07275426e-01 -7.62586817e-02
1.21455014e+00 -4.32286710e-01 -3.19947213e-01 2.35252187e-01
1.02748203e+00 7.06640959e-01 2.77545959e-01 6.32074237e-01
3.09574932e-01 4.89252567e-01 2.78683722e-01 2.58621812e-01
4.45363224e-01 6.32892668e-01 3.05151463e-01 1.81220755e-01
1.34982631e-01 -2.22238488e-02 4.61524755e-01 3.51114690e-01
-3.05740476e-01 -1.18208587e-01 -1.12390959e+00 5.51909089e-01
-1.88158047e+00 -9.85281825e-01 1.04373121e+00 1.76298666e+00
9.70217347e-01 9.59447846e-02 3.30423474e-01 -3.00202340e-01
4.66410935e-01 -3.06631178e-01 -1.00245619e+00 -7.17517316e-01
4.31787968e-01 7.85166085e-01 2.82625228e-01 4.96087879e-01
-6.97355926e-01 1.19903386e+00 7.00530291e+00 5.45671523e-01
-9.40504253e-01 2.32397214e-01 3.61835718e-01 1.64968997e-01
-1.75831437e-01 -8.25487003e-02 -5.96089482e-01 -1.06534557e-02
1.18504739e+00 -4.00476933e-01 1.03197265e+00 8.68107677e-01
-1.51226642e-02 -1.49589637e-02 -1.12478328e+00 4.11139965e-01
5.35593182e-02 -1.04491723e+00 2.23916639e-02 2.56245993e-02
1.01703644e+00 -8.12644735e-02 3.16636890e-01 8.24972451e-01
1.43236506e+00 -1.11337423e+00 1.51551679e-01 1.86977580e-01
5.36394596e-01 -8.76477182e-01 1.64083689e-01 9.30443645e-01
-6.01816952e-01 -7.29896784e-01 -1.56070769e-01 -4.49252963e-01
-3.79397988e-01 -4.52701628e-01 -1.52334714e+00 4.64656472e-01
5.13247669e-01 3.56212825e-01 -4.38931555e-01 7.26345003e-01
-5.31617522e-01 5.77183187e-01 -2.60829955e-01 -1.92241430e-01
4.59185749e-01 6.86286539e-02 3.11039150e-01 6.05059505e-01
2.28451535e-01 4.79677349e-01 6.42435193e-01 6.14263237e-01
1.15070017e-02 -2.67444402e-01 -1.09071553e+00 9.78781506e-02
3.87749970e-01 1.00398397e+00 -3.26176971e-01 -4.04831201e-01
-4.62784082e-01 8.83456230e-01 8.12500358e-01 5.03461599e-01
-4.79557306e-01 -2.65173852e-01 4.71031159e-01 -1.70613825e-01
4.62000966e-01 -2.99540818e-01 2.78566331e-01 -1.02164805e+00
-4.94556457e-01 -1.46483552e+00 4.43370849e-01 -8.15694690e-01
-1.08222008e+00 2.92798966e-01 4.48601812e-01 -7.30667114e-01
-8.22859883e-01 -6.35232866e-01 -7.78614223e-01 4.98670340e-01
-1.42823267e+00 -9.03926611e-01 -5.21568619e-02 8.53042245e-01
7.78676629e-01 -8.12037230e-01 9.51274097e-01 -4.29710567e-01
-3.18180591e-01 3.47807080e-01 2.26664230e-01 -1.29367843e-01
5.78240097e-01 -1.45412207e+00 3.48524928e-01 4.71663475e-01
2.57729232e-01 3.51781726e-01 5.61413527e-01 -4.12068129e-01
-1.22487020e+00 -8.23356509e-01 3.34768631e-02 -4.75953668e-01
5.56630909e-01 -2.21783876e-01 -9.00662601e-01 1.18072486e+00
5.94790339e-01 -2.95789659e-01 4.88364786e-01 1.59063399e-01
-1.43550664e-01 -1.56070769e-01 -1.12007403e+00 6.92978919e-01
8.48721683e-01 -1.65648296e-01 -1.06419325e+00 4.65582132e-01
6.98171139e-01 -5.04708648e-01 -7.67810225e-01 1.56118035e-01
2.08712518e-01 -7.21565962e-01 8.15757990e-01 -1.35594976e+00
1.40028402e-01 1.57257393e-01 2.27720320e-01 -1.99746668e+00
-4.33609605e-01 -9.30768728e-01 -1.97458282e-01 6.54552102e-01
3.18101734e-01 -9.70031142e-01 7.48994648e-01 6.01962268e-01
-2.66206354e-01 -4.92158800e-01 -8.80042613e-01 -8.72448504e-01
5.36861897e-01 -7.06754550e-02 7.87539542e-01 9.07245815e-01
-3.32936738e-03 4.34101462e-01 -3.28222543e-01 2.49774903e-02
3.84198070e-01 -3.45024690e-02 9.75458682e-01 -1.34998679e+00
-8.87615681e-01 -2.11684257e-01 -2.83346251e-02 -1.23192859e+00
6.71250999e-01 -7.76580811e-01 1.20462894e-01 -1.22254920e+00
7.46444017e-02 -6.41511023e-01 -4.82400328e-01 8.84872854e-01
1.25887161e-02 -4.15798128e-01 2.32307717e-01 -2.29967143e-02
-1.03811288e+00 4.33241367e-01 1.64368415e+00 -1.89382866e-01
-3.53387594e-01 5.06469131e-01 -8.26595783e-01 8.67026985e-01
9.88837302e-01 -6.85394704e-01 -8.86127710e-01 -6.46965921e-01
3.03531736e-01 1.09414741e-01 2.37054527e-02 -1.09082913e+00
4.95884448e-01 -7.85985589e-01 5.33924520e-01 3.81587058e-01
3.11647028e-01 -9.52515185e-01 -2.57030010e-01 6.96358204e-01
-5.72432339e-01 4.13101196e-01 3.40580553e-01 6.54831290e-01
2.00123459e-01 -5.31480849e-01 6.64658964e-01 -7.00966299e-01
-1.00260651e+00 5.99561483e-02 -8.65526557e-01 2.06018075e-01
1.52320611e+00 -5.02359793e-02 -2.86995649e-01 -6.85451806e-01
-9.74727929e-01 5.96029103e-01 3.72017115e-01 2.35378981e-01
6.37963831e-01 -1.01278996e+00 -7.17109025e-01 3.12376291e-01
-8.51563141e-02 2.50803292e-01 -1.07470758e-01 1.89466655e-01
-1.78127721e-01 2.56740153e-01 -5.35632908e-01 -4.14798766e-01
-1.09500122e+00 6.72700286e-01 6.08199477e-01 -5.82607210e-01
-4.15098667e-01 5.82970977e-01 2.21378505e-01 -5.33395708e-01
5.85692599e-02 5.51686250e-02 -2.37794310e-01 -5.20899653e-01
4.39074814e-01 1.25765175e-01 -1.62281290e-01 7.16580972e-02
2.43745744e-01 2.70941883e-01 -4.41750228e-01 -4.09135818e-01
1.46905971e+00 3.15039232e-02 2.73549944e-01 2.85231352e-01
4.59845692e-01 -3.42200667e-01 -1.49649060e+00 -3.83005857e-01
1.19603455e-01 -3.04993480e-01 -3.23899746e-01 -1.08499217e+00
-5.79635441e-01 6.75243616e-01 3.93530756e-01 4.17850077e-01
1.18556297e+00 1.13820937e-02 2.85877228e-01 1.21416342e+00
4.45923358e-01 -1.31745458e+00 8.90137434e-01 8.78683329e-01
5.29259861e-01 -1.21645761e+00 -1.78146079e-01 2.92809695e-01
-7.98485458e-01 1.20430529e+00 1.08259320e+00 -1.30239606e-01
3.28446895e-01 1.84500441e-01 -1.23611189e-01 -1.00718349e-01
-1.19536602e+00 -2.02728599e-01 2.89504454e-02 1.05153728e+00
-1.68555796e-01 -1.68654814e-01 4.04697210e-01 7.97083452e-02
1.23618111e-01 3.33881825e-02 7.43768215e-01 1.30795956e+00
-8.57736170e-01 -1.51751590e+00 -2.28236526e-01 3.02207291e-01
-3.12248230e-01 4.20020252e-01 -3.66202921e-01 7.74939179e-01
1.42530888e-01 8.86138856e-01 -1.27192587e-01 -1.83270767e-01
2.22675949e-01 3.37420762e-01 9.02516425e-01 -1.22505105e+00
-8.22577000e-01 -2.94023424e-01 -7.80036524e-02 -3.65075141e-01
-5.10772943e-01 -5.17928064e-01 -1.09468269e+00 -1.54177159e-01
1.73736885e-01 2.65777409e-01 3.41822594e-01 1.28131568e+00
2.29180157e-01 7.00150907e-01 7.79623091e-01 -5.37395179e-01
-1.17528117e+00 -8.32612693e-01 -3.70927036e-01 3.93386662e-01
3.43871266e-01 -7.75205433e-01 -2.07075011e-02 -6.87080063e-03] | [4.051812171936035, 1.5673893690109253] |
9ddaa99b-5e0c-4602-ba3e-b7f027434856 | can-deepfakes-be-created-by-novice-users | 2304.14576 | null | https://arxiv.org/abs/2304.14576v1 | https://arxiv.org/pdf/2304.14576v1.pdf | Can deepfakes be created by novice users? | Recent advancements in machine learning and computer vision have led to the proliferation of Deepfakes. As technology democratizes over time, there is an increasing fear that novice users can create Deepfakes, to discredit others and undermine public discourse. In this paper, we conduct user studies to understand whether participants with advanced computer skills and varying levels of computer science expertise can create Deepfakes of a person saying a target statement using limited media files. We conduct two studies; in the first study (n = 39) participants try creating a target Deepfake in a constrained time frame using any tool they desire. In the second study (n = 29) participants use pre-specified deep learning-based tools to create the same Deepfake. We find that for the first study, 23.1% of the participants successfully created complete Deepfakes with audio and video, whereas, for the second user study, 58.6% of the participants were successful in stitching target speech to the target video. We further use Deepfake detection software tools as well as human examiner-based analysis, to classify the successfully generated Deepfake outputs as fake, suspicious, or real. The software detector classified 80% of the Deepfakes as fake, whereas the human examiners classified 100% of the videos as fake. We conclude that creating Deepfakes is a simple enough task for a novice user given adequate tools and time; however, the resulting Deepfakes are not sufficiently real-looking and are unable to completely fool detection software as well as human examiners | ['Brendan Dolan-Gavitt', 'Rachel Greenstadt', 'Siddharth Garg', 'Progga Deb', 'Brian Timmerman', 'Kevin Gallagher', 'Gauri Jagatap', 'Pulak Mehta'] | 2023-04-28 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 4.58588824e-02 2.53856152e-01 1.88276526e-02 1.37250468e-01
-6.92535281e-01 -1.02137101e+00 5.15506864e-01 -9.25335214e-02
-3.72895837e-01 1.09689146e-01 -3.59009691e-02 -8.87002945e-01
3.71864051e-01 -3.64030987e-01 -6.34134412e-01 -1.94218054e-01
5.68275034e-01 -6.69217557e-02 1.51933223e-01 -1.64316036e-02
6.15783691e-01 2.31907114e-01 -1.39965546e+00 6.98915064e-01
7.73758352e-01 5.39532363e-01 -8.88055414e-02 9.92165864e-01
2.26487249e-01 1.06021786e+00 -1.36870384e+00 -7.75421321e-01
2.84177929e-01 -4.53042507e-01 -6.83962882e-01 -9.31019112e-02
1.10830474e+00 -8.49894404e-01 -5.14242768e-01 1.26393390e+00
5.67253649e-01 -3.39938700e-01 1.28016502e-01 -1.55089104e+00
-7.06274092e-01 4.00742233e-01 -5.02849400e-01 1.89580202e-01
8.03076208e-01 8.14012170e-01 4.07503843e-01 -5.81506312e-01
3.71218354e-01 1.44214761e+00 7.95252979e-01 5.26535630e-01
-1.28888690e+00 -1.44596219e+00 -4.83417243e-01 -1.71391591e-02
-1.29445267e+00 -6.94808781e-01 6.15775585e-01 -1.06689537e+00
7.02891469e-01 3.17462057e-01 9.58461404e-01 1.58204818e+00
2.31339768e-01 3.62957746e-01 1.33737516e+00 -4.34244573e-01
-5.32169044e-02 6.42346382e-01 5.69733679e-02 7.15941370e-01
4.41746622e-01 3.75693887e-01 -4.51757818e-01 -4.87693012e-01
7.31408954e-01 -3.23445290e-01 -1.98604569e-01 3.48760873e-01
-1.23423290e+00 9.92521226e-01 6.06340282e-02 3.94015998e-01
-1.26051277e-01 -1.76626164e-02 4.22813028e-01 8.06783319e-01
3.52194995e-01 9.83697295e-01 1.36536762e-01 -5.88457048e-01
-9.96629536e-01 2.75944948e-01 9.71381783e-01 3.82341892e-01
3.28041464e-01 1.78485110e-01 3.83640267e-02 4.33479697e-01
1.27594084e-01 5.04606307e-01 6.90152824e-01 -9.83024657e-01
4.12237704e-01 3.19644898e-01 4.24687177e-01 -1.77167118e+00
1.33702591e-01 -2.36693174e-01 -4.47013453e-02 5.09310126e-01
8.83822918e-01 -4.22980130e-01 -6.51536345e-01 1.19724286e+00
-1.27461376e-02 1.75078139e-01 -3.92975777e-01 8.92577469e-01
6.35280669e-01 5.14313698e-01 -1.21413693e-01 1.16987757e-01
1.40104687e+00 -4.87312794e-01 -7.02660441e-01 -4.92848843e-01
9.72455680e-01 -1.17538297e+00 1.61760032e+00 1.02100754e+00
-8.90542805e-01 -7.72004128e-01 -1.35824966e+00 2.67045826e-01
-5.64131215e-02 -1.99853349e-02 2.66529620e-01 1.56761682e+00
-9.78910327e-01 4.61113781e-01 -5.12053072e-01 -1.61704928e-01
4.03074592e-01 6.84487000e-02 -3.72848392e-01 -1.88457388e-02
-1.10917532e+00 1.04598820e+00 -1.18900826e-02 -1.90703556e-01
-9.70851541e-01 -6.40014470e-01 -6.42688155e-01 -1.70967922e-01
1.26704812e-01 -4.70445976e-02 1.62673759e+00 -1.58539379e+00
-1.04276264e+00 1.29119968e+00 2.64825284e-01 -1.70800835e-01
7.94679582e-01 -3.94798577e-01 -5.76253057e-01 2.47860014e-01
1.23616323e-01 2.19876498e-01 1.39519453e+00 -1.05990136e+00
-3.87092829e-01 -1.62658110e-01 2.84330130e-01 -2.26525962e-01
-7.92162597e-01 2.22956598e-01 3.77135187e-01 -7.16745615e-01
-3.11787069e-01 -8.85903895e-01 5.16649604e-01 -3.69176976e-02
-5.85443258e-01 -4.28348258e-02 1.26308322e+00 -9.76531625e-01
1.41675615e+00 -2.48749185e+00 -4.43113506e-01 1.60768032e-01
8.52448881e-01 7.73546994e-01 -1.02162868e-01 2.84135699e-01
-2.22183719e-01 6.61530197e-01 6.10516131e-01 -1.31394133e-01
-1.55474752e-01 -3.89624536e-01 -1.53092667e-01 8.37342680e-01
-1.22207895e-01 5.24923146e-01 -1.05598927e+00 -1.24982528e-01
7.17581436e-02 2.70953745e-01 -6.41513646e-01 9.00970176e-02
3.20916563e-01 7.22121522e-02 -1.19709469e-01 5.46757936e-01
6.22829139e-01 -2.14061085e-02 -1.18789345e-01 -8.30100477e-02
-2.70593822e-01 3.32408875e-01 -7.97851205e-01 8.33186984e-01
-4.22559232e-01 1.53175950e+00 2.46010154e-01 -8.08915734e-01
5.90392768e-01 3.79596055e-01 -1.61203220e-01 -4.43050086e-01
3.17698449e-01 3.81689161e-01 4.09596115e-01 -9.75179613e-01
4.79084879e-01 -1.38182044e-01 -1.06593035e-01 8.01454604e-01
-2.76727468e-01 -2.54738003e-01 -4.18615490e-01 3.33789080e-01
1.47866452e+00 -5.95358551e-01 -6.38379902e-02 9.63847041e-02
-4.58987169e-02 6.13637567e-02 -2.68618524e-01 1.19241655e+00
-6.62180424e-01 3.01054031e-01 6.11647367e-01 -3.63334239e-01
-1.08073854e+00 -7.94020355e-01 4.89072174e-01 1.06318963e+00
-1.99439093e-01 -2.57218629e-01 -1.15670025e+00 -6.91252232e-01
1.16513083e-02 8.83998692e-01 -8.53687972e-02 -6.24700904e-01
-3.03556859e-01 3.11484277e-01 1.23858452e+00 -1.61907062e-01
7.06496179e-01 -9.48905349e-01 -8.40403140e-01 1.34553472e-02
-4.79906380e-01 -1.29203176e+00 -7.75068820e-01 -6.53905272e-01
-2.32434258e-01 -1.05187047e+00 -4.89530593e-01 -7.34388530e-01
4.45544362e-01 7.97082722e-01 7.04504430e-01 7.10522294e-01
-2.73366094e-01 3.69948208e-01 -3.94573241e-01 -3.79174709e-01
-1.00326121e+00 -4.23618704e-01 4.84699421e-02 -5.98911680e-02
7.11198270e-01 -8.96264538e-02 -2.07124233e-01 3.98267925e-01
-7.82238722e-01 -7.85439312e-02 3.16750795e-01 4.12229210e-01
-5.42943001e-01 3.25157225e-01 2.86159873e-01 -6.82153463e-01
9.94195700e-01 -3.34598482e-01 -3.22204500e-01 -2.93049574e-01
-3.39388072e-01 -6.43739820e-01 3.12079549e-01 -1.17030728e+00
-4.38941509e-01 -3.05050641e-01 1.32150486e-01 -5.54214239e-01
-1.64586246e-01 4.26496536e-01 1.62405282e-01 -4.28627223e-01
1.29579461e+00 -1.11597791e-01 2.48274639e-01 -1.41826078e-01
-1.21134989e-01 1.22900331e+00 4.21706319e-01 -9.11514536e-02
1.04300702e+00 1.37401596e-01 -9.95307028e-01 -1.19657123e+00
-6.05378568e-01 -1.77018955e-01 7.73763657e-02 -6.83098733e-01
7.12473571e-01 -8.33008289e-01 -9.58125234e-01 9.68931973e-01
-1.40280092e+00 -4.29564208e-01 3.86646152e-01 5.63213706e-01
-9.09718648e-02 6.01494551e-01 -5.00461936e-01 -8.30131710e-01
2.00259328e-01 -1.22393179e+00 6.07527733e-01 -7.87689015e-02
-9.50175643e-01 -5.18028319e-01 -2.96439916e-01 9.36137080e-01
3.22112143e-01 2.96241224e-01 5.94785333e-01 -8.18196058e-01
-3.02345425e-01 -8.15332234e-01 -1.61518231e-01 4.84053016e-01
2.63555497e-01 2.14607939e-01 -1.10184848e+00 -4.27750707e-01
2.16980964e-01 -4.94198859e-01 2.07507193e-01 1.18129663e-01
8.49448025e-01 -6.74902320e-01 -1.75524190e-01 5.85584082e-02
7.54082084e-01 2.64146417e-01 7.28531480e-01 3.04873317e-01
7.51234770e-01 6.58630192e-01 3.80367935e-01 2.41916478e-01
2.17692833e-02 5.36996305e-01 2.75815398e-01 1.84124336e-01
6.26009926e-02 -4.74070758e-01 9.22511578e-01 1.64935231e-01
1.96252480e-01 -3.41757029e-01 -9.45576429e-01 4.26495254e-01
-1.19471312e+00 -1.33247232e+00 -3.63862455e-01 2.22335291e+00
5.11251211e-01 6.39640689e-01 4.25322741e-01 4.93231535e-01
1.09499669e+00 4.08380292e-02 -2.70027965e-01 -7.51190841e-01
4.15994585e-01 1.80408105e-01 4.05203104e-01 4.98151481e-01
-9.84356523e-01 8.63259614e-01 6.07012606e+00 6.35203898e-01
-1.41516113e+00 2.50184119e-01 4.80128676e-01 -2.07952335e-01
-1.97940931e-01 -1.64501011e-01 -4.63671774e-01 7.47749329e-01
1.28184855e+00 1.21822786e-02 4.86836284e-01 7.14833915e-01
7.28250206e-01 -4.57094572e-02 -9.06942666e-01 1.16139174e+00
1.57642290e-01 -1.24915397e+00 -4.56149608e-01 2.64357358e-01
2.46251032e-01 -3.93127620e-01 3.61854196e-01 2.90728092e-01
3.30890954e-01 -1.33945251e+00 1.12646937e+00 -5.91223761e-02
9.03319836e-01 -4.40342516e-01 5.81940353e-01 7.09448159e-01
-3.92004490e-01 -1.83307484e-01 6.60836697e-02 -6.36795342e-01
-1.18880108e-01 4.92900044e-01 -1.03354967e+00 -5.43632388e-01
7.29474664e-01 1.87331393e-01 -5.21424711e-01 7.28641033e-01
-3.49509597e-01 9.86222088e-01 -2.01052234e-01 -1.17502511e-01
-3.76628004e-02 5.03168702e-01 5.61390340e-01 1.12389779e+00
4.05161619e-01 -1.13961168e-01 -2.27895360e-02 7.50943363e-01
3.41491168e-03 -3.80819052e-01 -8.93592954e-01 -9.34246004e-01
8.18245053e-01 1.03497517e+00 -4.84458953e-01 -2.57701367e-01
-3.35662931e-01 1.03423977e+00 -1.65363729e-01 1.70861289e-01
-8.80639493e-01 -5.99730492e-01 7.22604394e-01 7.27593064e-01
-1.65214598e-01 -3.42204958e-01 -2.87491798e-01 -1.13800073e+00
-3.36896017e-04 -1.63573885e+00 4.49726693e-02 -1.03858507e+00
-1.04237413e+00 2.96981812e-01 -3.03942353e-01 -1.08277631e+00
-2.99359024e-01 -4.25958097e-01 -5.78976512e-01 8.25025082e-01
-7.10231185e-01 -8.27450097e-01 -6.27084434e-01 4.01954025e-01
4.05124873e-01 -1.80551559e-01 4.20638710e-01 3.64234805e-01
-2.34594479e-01 7.92460561e-01 -6.27197623e-01 4.65269357e-01
9.29715037e-01 -6.18604600e-01 5.18326819e-01 8.17753136e-01
3.07672899e-02 9.11088586e-01 9.13610041e-01 -8.64514470e-01
-1.15375209e+00 -4.77902949e-01 8.95301759e-01 -6.43650711e-01
9.27118301e-01 -3.60976785e-01 -7.66498625e-01 6.44051909e-01
1.50858790e-01 -2.58027136e-01 7.35530138e-01 -3.13250571e-01
-6.55193090e-01 5.56541026e-01 -1.54567909e+00 5.88512838e-01
7.06570387e-01 -1.02725494e+00 -5.17643452e-01 2.38487497e-01
6.31529450e-01 -3.89769703e-01 -3.76860708e-01 -3.71148169e-01
1.02310431e+00 -1.23328984e+00 8.61943901e-01 -6.03148341e-01
6.84401989e-01 5.86655885e-02 2.26168483e-01 -1.26757455e+00
-8.39859247e-02 -9.98015165e-01 2.10240725e-02 1.09806395e+00
1.39227927e-01 -5.90659678e-01 7.78869033e-01 7.89618909e-01
-5.56609705e-02 5.02625667e-02 -7.24330544e-01 -7.52437294e-01
-4.50340174e-02 -6.42796874e-01 9.90413129e-02 1.41561508e+00
1.02114543e-01 -6.41334727e-02 -5.26561141e-01 2.05732286e-01
3.42100590e-01 -7.31290460e-01 1.09643006e+00 -1.01295984e+00
-3.79144490e-01 -4.51668650e-01 -6.21624708e-01 -7.60707378e-01
-6.22324347e-02 -4.88354981e-01 -2.22986579e-01 -7.43728518e-01
-3.06921825e-02 1.18726596e-01 3.59716982e-01 5.57603359e-01
1.20097995e-01 4.29315120e-01 3.15254986e-01 3.81812721e-01
2.20525578e-01 -3.56011391e-01 1.03806758e+00 -1.28379822e-01
-3.33450168e-01 4.82772589e-02 -1.11635983e+00 9.44490671e-01
5.81150532e-01 -5.71736217e-01 -3.46215665e-01 -3.14535558e-01
4.77964550e-01 -8.10367689e-02 1.14918959e+00 -1.06037390e+00
-2.00235069e-01 -1.19580507e-01 3.83108765e-01 5.88539876e-02
1.09806843e-01 -7.31035411e-01 -9.04230103e-02 9.67920899e-01
-2.64030695e-01 1.39902726e-01 3.60715359e-01 2.05376238e-01
1.11339472e-01 -4.39770818e-01 9.18598533e-01 -8.63007680e-02
-3.75731826e-01 -5.12185216e-01 -1.15600383e+00 7.96172321e-02
1.01251626e+00 -7.67093122e-01 -5.96561253e-01 -9.97561574e-01
-5.42804003e-01 -4.17683363e-01 5.05221307e-01 4.29375976e-01
7.64164567e-01 -1.01442242e+00 -4.33680624e-01 3.77705455e-01
-1.87636718e-01 -7.00417519e-01 5.35965301e-02 5.88753462e-01
-7.90782511e-01 8.75910819e-02 -2.00086400e-01 -3.17572653e-01
-1.60607815e+00 2.95779914e-01 2.29076892e-01 4.61379945e-01
-4.32791620e-01 1.02362561e+00 3.95190492e-02 -8.92554373e-02
1.04656301e-01 -1.11418746e-01 2.17946067e-01 1.85524821e-01
7.85725474e-01 3.69596392e-01 -3.83296050e-02 -5.30310214e-01
-2.31554478e-01 -4.81994338e-02 -2.56281167e-01 -2.54003316e-01
7.36412883e-01 1.39494881e-01 3.25544447e-01 1.28974453e-01
1.21477258e+00 4.06493664e-01 -8.34216118e-01 1.24308944e-01
-2.84148544e-01 -1.03002787e+00 2.43152350e-01 -7.23109066e-01
-7.55442142e-01 9.52947676e-01 4.89108533e-01 7.80984879e-01
6.90174043e-01 -8.52725208e-02 9.18966293e-01 1.14260249e-01
2.06395134e-01 -7.76400864e-01 8.56993675e-01 2.74976969e-01
7.46297002e-01 -9.95988309e-01 -2.05382422e-01 -2.32993677e-01
-5.98620892e-01 9.53112662e-01 7.86860168e-01 -1.72246620e-01
3.22315454e-01 9.02016982e-02 1.46370709e-01 -3.13583642e-01
-3.29779655e-01 4.95423198e-01 -1.31809441e-02 8.44283998e-01
4.57322568e-01 7.66067356e-02 -7.11404830e-02 6.52891755e-01
-8.14142883e-01 1.35909200e-01 1.06399941e+00 8.82211745e-01
-6.77942753e-01 -4.48885798e-01 -7.90815890e-01 5.04983008e-01
-5.95398784e-01 6.26467094e-02 -1.02410138e+00 8.50026906e-01
1.97109908e-01 1.39925563e+00 7.60627314e-02 -9.93285596e-01
2.20537990e-01 -8.67214352e-02 3.57577622e-01 -7.62731373e-01
-8.53693306e-01 -1.44139528e-01 3.83501470e-01 -4.42005545e-01
6.66669309e-02 -6.91220105e-01 -6.34926081e-01 -1.03436005e+00
-3.94881040e-01 -1.75346076e-01 5.96193016e-01 1.14108396e+00
2.41447195e-01 -1.11875236e-01 5.03798306e-01 -7.47081876e-01
-7.38373697e-01 -1.04961538e+00 -1.06867976e-01 2.37232313e-01
7.14858651e-01 -5.67176163e-01 -9.07516420e-01 6.34112135e-02] | [12.527132034301758, 1.1952327489852905] |
f2bbb6cd-bdad-4961-a6c1-fcfdc23a9bac | corola-a-sequential-solution-to-moving-object | 1505.03566 | null | http://arxiv.org/abs/1505.03566v2 | http://arxiv.org/pdf/1505.03566v2.pdf | COROLA: A Sequential Solution to Moving Object Detection Using Low-rank Approximation | Extracting moving objects from a video sequence and estimating the background
of each individual image are fundamental issues in many practical applications
such as visual surveillance, intelligent vehicle navigation, and traffic
monitoring. Recently, some methods have been proposed to detect moving objects
in a video via low-rank approximation and sparse outliers where the background
is modeled with the computed low-rank component of the video and the foreground
objects are detected as the sparse outliers in the low-rank approximation. All
of these existing methods work in a batch manner, preventing them from being
applied in real time and long duration tasks. In this paper, we present an
online sequential framework, namely contiguous outliers representation via
online low-rank approximation (COROLA), to detect moving objects and learn the
background model at the same time. We also show that our model can detect
moving objects with a moving camera. Our experimental evaluation uses simulated
data and real public datasets and demonstrates the superior performance of
COROLA in terms of both accuracy and execution time. | ['Hong Zhang', 'Moein Shakeri'] | 2015-05-13 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 1.31164432e-01 -6.40471160e-01 -1.25956208e-01 -1.62521765e-01
-7.32570410e-01 -2.74281800e-01 3.18111360e-01 -2.83973157e-01
-3.88224632e-01 3.65913749e-01 1.04159266e-01 -1.50053818e-02
-1.28193749e-02 -1.55220479e-01 -8.26829016e-01 -8.42221200e-01
-3.84873748e-01 1.62113667e-01 9.27765071e-01 3.63346398e-01
1.43714055e-01 4.74528521e-01 -1.69602251e+00 4.20691550e-01
7.97170401e-01 1.09079123e+00 2.22410217e-01 8.88193250e-01
1.40358871e-02 1.17881477e+00 -4.59045589e-01 -1.06690809e-01
6.10463023e-01 -1.68085113e-01 -1.23444321e-02 4.74958599e-01
8.41430724e-01 -6.66909873e-01 -7.75436938e-01 1.23557770e+00
1.74561664e-01 3.61782700e-01 2.34580949e-01 -1.42614317e+00
1.33221269e-01 6.73917085e-02 -1.07897437e+00 6.52937353e-01
2.89589435e-01 2.43445039e-02 3.63183796e-01 -1.17318225e+00
3.82524788e-01 1.32442319e+00 6.01072073e-01 2.84465075e-01
-8.48174572e-01 -7.02281535e-01 6.37140751e-01 5.64928651e-01
-1.37265861e+00 -6.66777194e-01 6.75645232e-01 -5.34365714e-01
3.88260573e-01 5.13972580e-01 5.11708081e-01 5.94216049e-01
-5.09617908e-04 1.17246604e+00 6.62857413e-01 9.67332497e-02
1.92147151e-01 -3.47911268e-01 2.75420129e-01 6.78834558e-01
7.66768694e-01 3.25314663e-02 -5.45981109e-01 -6.30294502e-01
5.90101600e-01 7.02840865e-01 -4.07322228e-01 -3.39650959e-01
-1.15698516e+00 4.40838724e-01 3.98946479e-02 3.77073176e-02
-5.28033853e-01 2.65049011e-01 3.94317210e-01 1.20286748e-01
5.41358590e-01 -5.37662625e-01 -2.41188705e-01 -2.90787444e-02
-1.26287520e+00 1.83488742e-01 6.53731287e-01 1.15092134e+00
6.20405614e-01 2.49343336e-01 -2.17821658e-01 5.50572097e-01
3.03250104e-01 8.38311076e-01 4.24275756e-01 -1.05412376e+00
5.73590040e-01 2.59965688e-01 3.90811831e-01 -1.45921695e+00
-1.23365141e-01 -1.02788530e-01 -9.99397993e-01 3.91373746e-02
3.54425430e-01 8.09381250e-03 -8.18660080e-01 1.24220908e+00
5.46141744e-01 1.30761731e+00 -5.51762283e-02 9.82772887e-01
6.86401427e-01 9.37088549e-01 -2.16591850e-01 -6.48533404e-01
1.02612245e+00 -9.46142912e-01 -7.15560555e-01 -2.24048570e-01
3.47014517e-01 -7.99819767e-01 4.85081941e-01 6.63970590e-01
-9.30190027e-01 -5.41711509e-01 -6.19778454e-01 2.52077401e-01
2.45944589e-01 3.55562389e-01 4.60448712e-01 4.62292641e-01
-6.25586510e-01 3.21431279e-01 -1.16931558e+00 -1.25668392e-01
5.09087384e-01 2.79466450e-01 -2.42449358e-01 -4.79414433e-01
-5.80959678e-01 2.17996389e-01 2.05779567e-01 4.74038839e-01
-1.05611396e+00 -2.85817593e-01 -6.63234055e-01 -1.38515279e-01
7.39358068e-01 -3.85910809e-01 8.88874829e-01 -1.23325050e+00
-1.08401942e+00 3.37260991e-01 -6.69821739e-01 -4.85362470e-01
7.19527721e-01 -6.80149078e-01 -5.74374080e-01 2.82119751e-01
9.57107767e-02 5.51911481e-02 1.31315827e+00 -1.08656538e+00
-1.14699745e+00 -3.20930153e-01 -3.92447650e-01 -1.11955211e-01
-1.24464467e-01 1.07070394e-01 -9.06515956e-01 -7.91844964e-01
4.96820748e-01 -1.06950009e+00 -5.79345405e-01 -1.46244811e-02
3.74819487e-02 2.58162078e-02 1.28882968e+00 -9.10873950e-01
1.25737190e+00 -2.48208904e+00 -6.29177392e-02 3.43373030e-01
3.27563167e-01 3.77695531e-01 -1.33051157e-01 -4.92775999e-02
9.76606086e-02 -3.95855039e-01 -3.40311751e-02 -1.33168221e-01
-3.84044051e-01 1.45286933e-01 -3.61070663e-01 9.67973053e-01
-1.34931237e-03 4.29547548e-01 -1.02638388e+00 -4.20805097e-01
2.08558097e-01 2.88792908e-01 -5.86613715e-01 2.40959674e-01
1.21881522e-01 3.09295356e-01 -5.09502232e-01 5.77598751e-01
1.02023923e+00 -6.44561201e-02 -1.17075346e-01 -2.83221871e-01
-7.05451071e-02 -4.72836643e-01 -1.71176207e+00 1.20892763e+00
1.20313972e-01 7.81191766e-01 3.46512735e-01 -7.40000367e-01
3.70812535e-01 1.86838225e-01 6.55920744e-01 -4.29315597e-01
-1.78319082e-01 1.87191486e-01 -1.66889369e-01 -7.94655800e-01
6.67759955e-01 4.00334388e-01 4.27277088e-01 -1.12573951e-01
-4.36974347e-01 5.60713947e-01 5.42763829e-01 4.62461263e-01
1.19831049e+00 -2.85977535e-02 1.66297540e-01 9.88164917e-02
6.87629342e-01 -1.01028033e-01 9.80739534e-01 9.68109071e-01
-1.99645415e-01 5.36098957e-01 1.40984371e-01 -6.69939816e-01
-6.83420539e-01 -8.45575094e-01 3.47145885e-01 9.45538700e-01
2.78095871e-01 -3.17802399e-01 -4.83533144e-01 -5.80650926e-01
-1.97135825e-02 1.39574498e-01 -1.47224218e-01 1.25836194e-01
-8.72775674e-01 -6.07869148e-01 -5.01149567e-03 3.49131316e-01
2.84879088e-01 -4.61212069e-01 -6.39069438e-01 1.38289109e-01
-3.36644143e-01 -1.53415644e+00 -7.38025665e-01 -4.52749223e-01
-1.06295991e+00 -1.25233316e+00 -6.90228462e-01 -7.01226711e-01
9.66232657e-01 1.17600894e+00 7.13085592e-01 2.64535338e-01
-3.49893570e-01 2.83502311e-01 -3.33852082e-01 -1.50378466e-01
3.41907591e-02 -5.77608049e-01 2.83728242e-01 6.67732358e-01
2.24591747e-01 -2.00584531e-01 -6.68713272e-01 4.61377978e-01
-1.06123209e+00 -1.45825922e-01 7.52488732e-01 5.65330982e-01
6.45449877e-01 3.53695631e-01 1.39567703e-01 -9.29134130e-01
7.35052209e-03 -6.18824124e-01 -1.12071633e+00 -7.57424086e-02
1.10920303e-01 -4.26112056e-01 5.00047743e-01 -6.59302890e-01
-8.53768587e-01 4.28966701e-01 1.95407614e-01 -8.80894840e-01
-7.49259666e-02 2.96255291e-01 -2.12465510e-01 -9.25598592e-02
2.36113876e-01 3.52642387e-01 -2.12658167e-01 -4.15107310e-01
1.44820571e-01 5.09396136e-01 6.77270353e-01 -2.22137839e-01
1.14238799e+00 9.92280424e-01 5.69166392e-02 -1.31828296e+00
-7.59374142e-01 -1.18569803e+00 -3.50940764e-01 -2.78074831e-01
4.55957919e-01 -1.40262997e+00 -5.57210326e-01 5.27167916e-01
-1.03859520e+00 1.06840588e-01 6.12432463e-03 8.26601982e-01
-2.65247136e-01 8.73065948e-01 -5.68899274e-01 -1.17181253e+00
5.43067418e-02 -1.00995839e+00 1.11496305e+00 4.35545631e-02
2.74459958e-01 -3.77420634e-01 -2.08659858e-01 3.00209671e-01
1.80557653e-01 2.98067600e-01 2.83790290e-01 -4.57532495e-01
-1.16554737e+00 -4.88053828e-01 -2.67071992e-01 2.12036416e-01
8.21439028e-02 1.95566490e-02 -7.79342592e-01 -6.55228913e-01
1.89775422e-01 2.91134119e-01 1.03365993e+00 6.44096911e-01
1.08006549e+00 -5.97689867e-01 -4.49038893e-01 7.02669084e-01
1.27903092e+00 2.18490362e-01 5.51196635e-01 -6.90179411e-03
9.21859980e-01 4.34277803e-01 8.84072542e-01 6.72065556e-01
-1.34024769e-02 3.91288936e-01 3.56266379e-01 9.89667252e-02
1.70778066e-01 1.49780557e-01 8.50515306e-01 1.04071259e+00
-9.95380580e-02 2.15778165e-02 -7.22997427e-01 6.48929119e-01
-2.32145953e+00 -1.43841302e+00 -6.32784188e-01 2.47458220e+00
1.49135187e-01 2.78319240e-01 2.99463063e-01 -8.42087865e-02
9.50005829e-01 1.29219487e-01 -5.02782881e-01 4.90149677e-01
-7.50569627e-02 -5.23104131e-01 6.47201359e-01 4.08841193e-01
-1.22717822e+00 7.38053977e-01 5.82108212e+00 6.97175443e-01
-8.91036272e-01 1.01663627e-01 6.34946883e-01 -3.85741264e-01
3.09892297e-01 5.45388609e-02 -8.50418329e-01 6.42897427e-01
8.08210433e-01 1.68778445e-03 2.51394600e-01 8.47700775e-01
6.79737210e-01 -2.77387500e-01 -8.68504882e-01 1.38853776e+00
3.25818896e-01 -1.11261117e+00 1.72781840e-01 -5.62037341e-02
9.56036091e-01 1.16406128e-01 -1.79166570e-01 7.14227334e-02
-7.90061150e-03 -4.91644233e-01 6.04661345e-01 6.61501050e-01
2.08560854e-01 -8.09375763e-01 7.87759840e-01 5.62968791e-01
-1.33342946e+00 -2.75711417e-01 -5.89126766e-01 -2.21899319e-02
1.84648633e-01 9.17263448e-01 -2.67021000e-01 3.94756407e-01
9.48756397e-01 8.02520156e-01 -4.19232726e-01 1.55318522e+00
3.54986966e-01 9.35027778e-01 -4.74396884e-01 2.78314382e-01
1.89370349e-01 -5.71497321e-01 8.83210778e-01 1.27281344e+00
4.27063763e-01 3.61532688e-01 8.95874619e-01 1.45711616e-01
1.29469812e-01 5.90820909e-02 -4.72009093e-01 1.09129429e-01
1.36936590e-01 1.02151227e+00 -9.07582402e-01 -5.34391999e-01
-8.27498853e-01 8.84375691e-01 -2.04629660e-01 6.11718416e-01
-1.07290900e+00 -1.85887545e-01 7.16452956e-01 2.53488034e-01
7.21673965e-01 -4.33037370e-01 4.45759118e-01 -1.49901009e+00
4.30179775e-01 -9.51355994e-01 3.74086887e-01 -3.58587861e-01
-1.15017688e+00 3.23569089e-01 -1.39486715e-01 -1.62106538e+00
1.44931704e-01 -4.90772069e-01 -4.42395240e-01 3.08834374e-01
-1.45068562e+00 -6.47545040e-01 -5.96764803e-01 7.89413929e-01
7.60640860e-01 -1.93495423e-01 -2.98249424e-02 5.38307786e-01
-7.59935856e-01 7.82480165e-02 6.83965325e-01 2.90613294e-01
5.35959899e-01 -6.93176329e-01 1.83796763e-01 1.38353670e+00
1.95368960e-01 3.78567755e-01 7.59690940e-01 -8.94465208e-01
-1.80018735e+00 -1.41446555e+00 3.31866026e-01 -1.56823099e-01
6.12449586e-01 -2.50169039e-01 -9.52479243e-01 4.89506304e-01
-1.96889505e-01 7.25741923e-01 4.29847747e-01 -4.55391586e-01
-9.27248131e-03 -4.13238704e-01 -8.20134044e-01 6.48308277e-01
1.04176521e+00 -8.82157832e-02 -2.50487089e-01 5.30643940e-01
5.14193535e-01 -4.80225176e-01 -2.01520845e-01 2.86065429e-01
4.09466028e-01 -9.10668373e-01 8.33647609e-01 -5.89128077e-01
-1.55275896e-01 -1.01337397e+00 -2.62204677e-01 -7.20280349e-01
-3.72796267e-01 -9.52914298e-01 -6.89253211e-01 1.00210488e+00
-2.30995834e-01 -3.99743080e-01 9.30195749e-01 4.08847451e-01
1.07988931e-01 -2.85339892e-01 -1.03696251e+00 -8.95267844e-01
-9.57258224e-01 -5.18718958e-01 1.46104395e-01 6.16385043e-01
-7.40321100e-01 -6.90801665e-02 -8.95040214e-01 8.40634406e-01
9.51585114e-01 2.53338665e-01 1.21993518e+00 -1.16688132e+00
-2.13419721e-01 6.27489388e-02 -9.26825106e-01 -1.43807316e+00
1.29491106e-01 -3.50885153e-01 3.87408108e-01 -1.17465878e+00
5.30079246e-01 4.03898321e-02 -4.11054462e-01 -1.41616911e-01
-4.34982210e-01 1.07503727e-01 2.00319707e-01 4.57534760e-01
-1.16562498e+00 3.56780350e-01 4.73054826e-01 -1.92285180e-01
-8.89162272e-02 2.80311644e-01 -1.98853850e-01 1.22727871e+00
2.04209417e-01 -8.34705770e-01 -2.79446959e-01 -5.36383808e-01
-1.24888696e-01 1.38664499e-01 3.11466575e-01 -1.24467683e+00
5.29099524e-01 -2.86970198e-01 5.87917745e-01 -9.75701272e-01
1.93805367e-01 -1.33257329e+00 1.19801998e-01 6.15418315e-01
1.41065270e-01 3.19461226e-01 1.46718159e-01 1.27754331e+00
-3.48097056e-01 9.57041886e-03 6.97895169e-01 1.08970731e-01
-1.05389917e+00 6.37790918e-01 -5.55346012e-01 1.03592895e-01
1.25090551e+00 -4.13517356e-01 -6.24475852e-02 -5.75998008e-01
-5.41787326e-01 3.23148787e-01 3.69645715e-01 3.12011361e-01
9.21308041e-01 -1.25088465e+00 -9.00768220e-01 2.66327828e-01
1.50828408e-02 -6.51819706e-02 3.85649204e-01 1.01398230e+00
-7.15414464e-01 1.18784405e-01 2.35707968e-01 -9.25682724e-01
-1.64240217e+00 7.87378609e-01 -1.00424208e-01 3.25960480e-02
-6.50051296e-01 5.34776807e-01 3.99577171e-01 4.16191518e-01
4.75044340e-01 -4.37101603e-01 8.89537036e-02 -1.65896192e-01
1.02953339e+00 7.72269368e-01 -1.27011359e-01 -1.09187031e+00
-3.88939470e-01 6.08691812e-01 -1.50255308e-01 2.68346399e-01
1.26237261e+00 -2.68414021e-01 -2.32981801e-01 5.07545948e-01
1.04675508e+00 4.94569540e-01 -1.60270977e+00 -4.15174842e-01
3.44922692e-01 -1.07781613e+00 -8.10944214e-02 1.86739177e-01
-1.36148787e+00 4.08132166e-01 8.06359172e-01 -6.49259165e-02
1.19513130e+00 -3.46244723e-01 8.83293211e-01 6.99350595e-01
4.27694887e-01 -9.76913512e-01 8.93369690e-02 4.76679921e-01
4.96539444e-01 -1.34084821e+00 2.87442148e-01 -5.89334846e-01
-5.43870866e-01 8.94006550e-01 5.32266855e-01 -4.10553843e-01
6.09580278e-01 2.65384376e-01 -4.52663191e-02 2.07467869e-01
-7.69557595e-01 -3.03323179e-01 3.18244159e-01 2.94674903e-01
5.74628972e-02 -2.04524815e-01 -1.86925326e-02 2.21661329e-01
5.82689464e-01 -3.16470228e-02 5.55125594e-01 1.10185754e+00
-7.29456842e-01 -5.50003350e-01 -8.38520944e-01 6.72610581e-01
-6.84620023e-01 4.10408601e-02 -4.39638346e-02 2.80729473e-01
2.33846363e-02 1.06999183e+00 -1.55947115e-02 2.76137665e-02
3.01273823e-01 -2.38698229e-01 9.44167003e-02 -3.96657020e-01
-1.56717867e-01 4.62827176e-01 -2.53151298e-01 -9.30611730e-01
-5.51575065e-01 -7.54728973e-01 -9.41618800e-01 -1.43554688e-01
-4.58099902e-01 3.37925971e-01 2.99068749e-01 6.70237303e-01
4.19178456e-01 2.01091439e-01 7.53773153e-01 -1.03939188e+00
-3.87099415e-01 -5.74205220e-01 -5.58718741e-01 6.45936787e-01
7.67814338e-01 -4.21250790e-01 -3.18092704e-01 3.18910152e-01] | [8.98952865600586, -0.8048866987228394] |
b9bbec46-1590-4756-8ab4-23467149701d | fourier-contour-embedding-for-arbitrary | 2104.10442 | null | https://arxiv.org/abs/2104.10442v2 | https://arxiv.org/pdf/2104.10442v2.pdf | Fourier Contour Embedding for Arbitrary-Shaped Text Detection | One of the main challenges for arbitrary-shaped text detection is to design a good text instance representation that allows networks to learn diverse text geometry variances. Most of existing methods model text instances in image spatial domain via masks or contour point sequences in the Cartesian or the polar coordinate system. However, the mask representation might lead to expensive post-processing, while the point sequence one may have limited capability to model texts with highly-curved shapes. To tackle these problems, we model text instances in the Fourier domain and propose one novel Fourier Contour Embedding (FCE) method to represent arbitrary shaped text contours as compact signatures. We further construct FCENet with a backbone, feature pyramid networks (FPN) and a simple post-processing with the Inverse Fourier Transformation (IFT) and Non-Maximum Suppression (NMS). Different from previous methods, FCENet first predicts compact Fourier signatures of text instances, and then reconstructs text contours via IFT and NMS during test. Extensive experiments demonstrate that FCE is accurate and robust to fit contours of scene texts even with highly-curved shapes, and also validate the effectiveness and the good generalization of FCENet for arbitrary-shaped text detection. Furthermore, experimental results show that our FCENet is superior to the state-of-the-art (SOTA) methods on CTW1500 and Total-Text, especially on challenging highly-curved text subset. | ['Zhanghui Kuang', 'Wayne Zhang', 'Lianwen Jin', 'Lingyu Liang', 'Jianyong Chen', 'Yiqin Zhu'] | 2021-04-21 | fourier-contour-embedding-for-arbitrary-1 | https://arxiv.org/abs/2104.10442 | https://arxiv.org/pdf/2104.10442.pdf | cvpr-2021-2021-4 | ['scene-text-detection'] | ['computer-vision'] | [ 4.19606358e-01 -2.27212116e-01 2.04036906e-01 -1.48853853e-01
-2.61668682e-01 -6.26269162e-01 7.45169878e-01 -4.08188939e-01
-4.66791019e-02 -3.01755704e-02 4.47084643e-02 -2.32683450e-01
-2.37346023e-01 -8.71539176e-01 -5.23307323e-01 -5.75212896e-01
1.55447811e-01 6.52276695e-01 5.41030645e-01 -2.11785093e-01
2.81189084e-01 7.59323895e-01 -1.21705747e+00 3.29892039e-01
1.02902091e+00 1.00887656e+00 1.91750914e-01 6.76076114e-01
-5.44178426e-01 2.30097458e-01 -4.84063357e-01 -7.92461514e-01
4.34105366e-01 5.03075868e-02 -2.32348502e-01 4.42615718e-01
6.28629565e-01 -2.91070789e-01 -5.77452540e-01 1.08933771e+00
5.26557863e-01 -3.49560119e-02 1.17636633e+00 -1.04628408e+00
-9.63034630e-01 4.11078155e-01 -1.01530635e+00 -2.44343653e-01
-4.51439992e-03 -9.68173817e-02 8.19052517e-01 -1.53007901e+00
5.40935755e-01 1.36830294e+00 1.23745537e+00 1.68139532e-01
-8.60112548e-01 -6.17420614e-01 1.86309472e-01 -1.52121827e-01
-1.51380920e+00 2.65398901e-02 1.03515196e+00 -3.63932282e-01
5.51435292e-01 3.33230883e-01 5.16095638e-01 7.13191390e-01
1.94085687e-01 1.21401370e+00 7.30436921e-01 -3.45758140e-01
-1.52298108e-01 6.28998205e-02 1.01803586e-01 9.43656206e-01
2.81657428e-01 -1.54022008e-01 -3.61797184e-01 7.52553269e-02
1.20165801e+00 -6.25440478e-02 -6.83121502e-01 -3.98973256e-01
-1.29838145e+00 9.08291161e-01 4.17999804e-01 3.60358983e-01
-1.28868669e-01 -1.86542764e-01 3.74487370e-01 3.17523591e-02
5.56416810e-01 7.36570954e-02 -2.35637352e-01 1.66905582e-01
-1.27879548e+00 1.09736025e-01 6.96787000e-01 1.17125118e+00
3.71137112e-01 2.94762850e-01 -2.24622786e-01 1.05522871e+00
2.29323804e-01 8.12301457e-01 5.76358736e-01 -9.65513960e-02
5.89863241e-01 7.50286520e-01 -2.05993056e-01 -1.33532441e+00
-4.96036261e-01 -2.85281658e-01 -1.18851471e+00 3.83298248e-02
4.53736693e-01 -6.94116652e-02 -9.60784793e-01 7.72276938e-01
2.03430012e-01 1.55567586e-01 -1.58539355e-01 9.29453611e-01
8.77701879e-01 8.69665086e-01 -6.32352352e-01 4.86106379e-03
1.36631203e+00 -8.51788223e-01 -7.99892247e-01 -3.52714472e-02
4.76955384e-01 -1.15282047e+00 1.22921383e+00 6.67631090e-01
-8.51311207e-01 -6.22728944e-01 -1.05803442e+00 -1.74141869e-01
-4.09659564e-01 7.41251707e-01 4.62257087e-01 9.27797973e-01
-9.31567371e-01 4.23058867e-01 -5.27825117e-01 -8.70682448e-02
5.20255923e-01 3.10140520e-01 -1.03965439e-01 -1.66485161e-01
-8.06745648e-01 3.80828887e-01 3.33103985e-01 2.64552891e-01
7.42694885e-02 -6.09527946e-01 -8.94116044e-01 1.85274437e-01
3.64680409e-01 -2.65067399e-01 7.61236608e-01 -8.13326895e-01
-1.45215106e+00 6.05471373e-01 -1.75867900e-01 -2.23195031e-01
9.31899130e-01 -1.19515136e-01 -4.74359125e-01 4.43471938e-01
-1.25588939e-01 6.11656249e-01 1.41083550e+00 -1.29351175e+00
-2.80932248e-01 -2.66693234e-01 -5.38837552e-01 2.52063155e-01
-5.07872224e-01 -7.41102314e-03 -7.66559422e-01 -1.01782489e+00
5.54497838e-01 -5.31091154e-01 2.77526617e-01 5.75347245e-01
-6.48132145e-01 -4.05470580e-01 1.67068732e+00 -7.73535669e-01
1.18266821e+00 -2.04884052e+00 -1.88623741e-01 3.75589579e-01
6.13822453e-02 3.73812497e-01 -5.83825223e-02 2.87703991e-01
-6.36993796e-02 2.56241739e-01 -3.39348078e-01 -4.19843256e-01
2.30959177e-01 7.26743340e-02 -5.20770907e-01 5.67304909e-01
3.96532118e-01 1.11786473e+00 -3.08195353e-01 -9.11775291e-01
6.06007278e-01 5.96411109e-01 -2.94994891e-01 -2.30981976e-01
-2.97562867e-01 -1.11036353e-01 -6.08806074e-01 7.33234227e-01
1.26632893e+00 -3.08451295e-01 -9.34045911e-02 -6.30568206e-01
-2.11361229e-01 -4.53157216e-01 -1.56095147e+00 1.09456766e+00
-1.95818990e-02 7.60064781e-01 2.08457679e-01 -9.93978083e-01
1.21819282e+00 2.08228633e-01 3.83030802e-01 -5.59356272e-01
3.95151466e-01 2.09278837e-01 -3.40962559e-01 -5.39207101e-01
6.69166625e-01 1.91233158e-02 3.01366180e-01 3.74178499e-01
-2.67717540e-01 -6.31388247e-01 7.08708614e-02 -2.23417319e-02
4.44064528e-01 1.31537497e-01 -1.06917433e-01 -3.42710555e-01
8.27080727e-01 -2.26117477e-01 2.44785324e-01 5.16667068e-01
-2.06295983e-03 1.09399700e+00 4.67418253e-01 -5.78514993e-01
-9.58760142e-01 -9.09562409e-01 -6.97519481e-01 9.34069633e-01
2.69099116e-01 -1.70726925e-01 -8.08481991e-01 -7.31962800e-01
4.33081388e-02 4.91261572e-01 -4.20433939e-01 2.81596452e-01
-8.42694819e-01 -7.97358453e-01 7.02941179e-01 6.19212270e-01
1.03113055e+00 -9.71310794e-01 -2.73424983e-01 -7.31972679e-02
-7.97274411e-02 -1.30643785e+00 -1.12841856e+00 -1.49098635e-01
-7.65291750e-01 -9.87790585e-01 -1.10175908e+00 -1.20662022e+00
7.19124675e-01 4.52747554e-01 7.49013901e-01 1.53075367e-01
-4.53764379e-01 3.10235918e-01 -4.31287795e-01 -2.69501477e-01
-1.72131017e-01 -3.67416114e-01 -2.69678265e-01 3.48885506e-01
1.42989233e-01 -1.03346244e-01 -5.13992667e-01 6.02380574e-01
-1.18599677e+00 1.60397500e-01 6.00997031e-01 9.35593247e-01
5.60175061e-01 6.34675682e-01 2.55646229e-01 -6.61933839e-01
7.29936779e-01 1.33211300e-01 -5.75524926e-01 4.01301652e-01
-3.83279681e-01 -1.72516063e-01 8.42371881e-01 -7.51978099e-01
-1.03094172e+00 3.38144861e-02 -1.65883765e-01 -7.20899820e-01
-5.20275868e-02 3.30757797e-01 -2.14632843e-02 -3.80926341e-01
5.77194095e-01 8.40612113e-01 -1.11434370e-01 -2.23931819e-01
2.74502993e-01 8.40073466e-01 4.30143833e-01 -7.22225666e-01
1.13835680e+00 7.27761686e-01 5.19674718e-02 -1.51451826e+00
-3.67168546e-01 -5.45188606e-01 -9.98227298e-01 -2.70091206e-01
7.50360966e-01 -5.63479006e-01 -6.70328856e-01 8.35443795e-01
-1.08136463e+00 -2.41491541e-01 -1.93996988e-02 1.66438162e-01
-5.09263754e-01 1.09161997e+00 -6.68160439e-01 -9.08582211e-01
-6.46008790e-01 -9.79307711e-01 1.51909101e+00 2.43947193e-01
3.19211930e-01 -1.07957637e+00 -4.34163630e-01 2.44261459e-01
1.98901534e-01 1.66208789e-01 1.02590191e+00 -4.21405315e-01
-5.35974324e-01 -2.01982871e-01 -7.50357866e-01 1.96317971e-01
-3.97323817e-02 2.76246876e-01 -7.99883842e-01 -2.75686532e-01
-7.76800588e-02 -2.72623271e-01 9.01694417e-01 6.26392603e-01
1.22794712e+00 -9.47837383e-02 -3.03869545e-01 8.88680995e-01
1.35175729e+00 3.22012715e-02 6.03889883e-01 -5.71260154e-02
7.63767302e-01 4.09001946e-01 4.92665768e-01 3.49253327e-01
1.26311377e-01 5.66439033e-01 9.60491151e-02 -2.94544488e-01
-1.84481025e-01 -1.98605433e-01 2.40333900e-01 7.29745090e-01
1.10040441e-01 -4.61285025e-01 -7.66486585e-01 3.70424122e-01
-1.63947523e+00 -8.18544328e-01 -6.02712452e-01 1.78277040e+00
5.89452922e-01 2.63095677e-01 -9.34684202e-02 5.01804352e-01
9.24472511e-01 2.26853535e-01 -5.46305060e-01 -1.43057808e-01
-4.81247604e-01 4.07178670e-01 3.82497549e-01 2.00644463e-01
-1.24587858e+00 1.15137446e+00 5.75233841e+00 1.30965793e+00
-1.35988784e+00 -2.43270174e-01 4.34955955e-01 5.96399546e-01
-9.55008939e-02 -5.99225700e-01 -6.99763477e-01 2.09322244e-01
1.34350628e-01 3.41516882e-02 1.78317502e-01 6.55980349e-01
4.95596044e-02 2.59709597e-01 -6.39380217e-01 1.07166255e+00
3.18005472e-01 -1.33430946e+00 3.07946265e-01 -2.29664698e-01
6.43103659e-01 -1.26108184e-01 1.95677668e-01 2.19935492e-01
-9.83905047e-02 -1.05619001e+00 8.41340601e-01 3.91099155e-01
1.29285264e+00 -5.32357395e-01 5.27705669e-01 5.16654789e-01
-1.77728140e+00 9.84258577e-02 -7.16032922e-01 5.29065669e-01
-1.34928808e-01 5.26572585e-01 -9.05593812e-01 6.91495895e-01
4.70463544e-01 8.45691562e-01 -5.54683626e-01 9.91858125e-01
9.30200368e-02 5.56306958e-01 -6.86983228e-01 -4.12634164e-01
3.48772615e-01 -5.34617364e-01 4.81567979e-01 1.48498058e+00
3.75584334e-01 7.46660903e-02 2.98143893e-01 1.10433650e+00
-3.35199721e-02 4.63129312e-01 -4.03669178e-01 -1.63744956e-01
1.94057286e-01 1.34471285e+00 -1.23473454e+00 -1.91198587e-01
-4.47639495e-01 1.11928642e+00 -1.27617747e-01 5.06098390e-01
-6.88822091e-01 -5.78909338e-01 -1.15203805e-01 1.82399437e-01
6.27913475e-01 -2.64059544e-01 -4.64580655e-01 -1.16010892e+00
3.62024188e-01 -6.54218376e-01 1.46935761e-01 -1.00099695e+00
-1.34074688e+00 4.25314337e-01 -3.64424169e-01 -1.29048383e+00
4.34830755e-01 -1.25925374e+00 -1.05390084e+00 8.83391023e-01
-1.66546774e+00 -1.72581851e+00 -3.08628172e-01 9.10954654e-01
1.01261008e+00 -1.11957289e-01 5.04410565e-01 1.40814763e-02
-4.22190040e-01 7.53841639e-01 4.09783840e-01 5.49529731e-01
5.03153563e-01 -1.29217160e+00 5.76909959e-01 7.36005425e-01
2.53235072e-01 4.07130927e-01 3.43164325e-01 -8.78632843e-01
-1.39799166e+00 -1.24728811e+00 3.55756104e-01 -9.05427635e-02
4.89221185e-01 -6.38110399e-01 -1.23588252e+00 5.21977842e-01
-6.08794764e-02 1.38226777e-01 9.28167254e-02 -2.94173986e-01
-2.84612983e-01 3.14236730e-01 -1.19106495e+00 7.32261539e-01
9.55917180e-01 -3.00543427e-01 -6.50375545e-01 5.16750038e-01
5.40751457e-01 -5.12417138e-01 -7.98964500e-01 4.75776434e-01
5.01086831e-01 -1.03663170e+00 1.11434305e+00 -2.88062245e-02
2.73399353e-01 -3.06146413e-01 -3.13326297e-03 -9.01036501e-01
-2.49130681e-01 -6.26894891e-01 8.36630464e-02 9.55851436e-01
1.65439229e-02 -6.00323737e-01 1.16425693e+00 -8.58840346e-02
-4.07778919e-01 -9.45102811e-01 -1.02002215e+00 -5.67877710e-01
4.82829422e-01 -5.74243486e-01 4.21591401e-01 1.03713465e+00
-1.96043879e-01 -1.30022049e-01 -2.42640659e-01 2.68639445e-01
4.92515266e-01 2.90395647e-01 5.94986618e-01 -1.57151186e+00
9.21360925e-02 -7.27643311e-01 -2.82605886e-01 -1.63159561e+00
1.42659675e-02 -8.62378359e-01 -2.25408413e-02 -1.45157230e+00
-1.22619964e-01 -4.20642316e-01 4.67961133e-01 4.01142314e-02
-1.78344518e-01 9.25961509e-02 1.62560627e-01 1.02166794e-01
-2.13298976e-01 8.40645313e-01 1.79298115e+00 -2.63310432e-01
-6.86510652e-02 1.21880129e-01 -2.38258094e-01 9.30032909e-01
4.92543548e-01 3.67937982e-02 -3.41497362e-01 -2.12820753e-01
1.56839430e-01 -3.79827581e-02 3.41924191e-01 -9.24605608e-01
4.78037357e-01 -1.14020444e-02 8.30710292e-01 -1.48917758e+00
3.60234797e-01 -1.00909781e+00 -5.18606782e-01 1.16492316e-01
2.21188188e-01 -2.04198793e-01 3.45441580e-01 6.05198681e-01
-2.35337913e-02 -4.31165963e-01 8.88823986e-01 1.26958534e-01
-3.60612899e-01 4.22705024e-01 -1.87546864e-01 8.81418139e-02
7.44180381e-01 -7.81724274e-01 -3.05747241e-01 -2.53986180e-01
-3.49171221e-01 2.79128492e-01 3.77828658e-01 3.82223547e-01
1.11632085e+00 -1.17584074e+00 -6.12658978e-01 7.35165060e-01
-1.24122836e-01 4.02072251e-01 2.64689296e-01 7.24854946e-01
-7.44769573e-01 5.65995038e-01 2.42268875e-01 -8.93001735e-01
-1.14118719e+00 3.32396656e-01 6.25846803e-01 -2.56858110e-01
-1.18297827e+00 6.91217840e-01 5.32297850e-01 -7.82819450e-01
2.46144503e-01 -5.08271098e-01 -2.66318560e-01 -1.35825962e-01
4.07138765e-01 1.39192969e-01 8.32406431e-02 -7.57517815e-01
1.78945675e-01 1.50267553e+00 -1.28024936e-01 1.22863732e-01
9.88040209e-01 1.11445621e-01 2.58105367e-01 3.13889571e-02
9.75695610e-01 1.84274554e-01 -1.30225074e+00 -4.47132647e-01
-2.46903837e-01 -4.42692518e-01 8.60370472e-02 -5.48999965e-01
-1.04707682e+00 1.12244976e+00 5.00637949e-01 2.68820822e-01
1.08809161e+00 -3.80920678e-01 8.91661286e-01 6.18821859e-01
4.00403664e-02 -1.17355335e+00 5.06572187e-01 7.00701058e-01
1.17701828e+00 -9.32034254e-01 1.29137626e-02 -9.62688029e-01
-6.43404424e-01 1.81457901e+00 4.28637624e-01 -1.20441608e-01
9.00991559e-01 4.39705014e-01 -2.91643720e-02 -4.28124130e-01
-1.17380671e-01 8.98600277e-03 5.73769569e-01 5.52850723e-01
3.33390146e-01 -1.33352801e-01 5.00293933e-02 5.70343792e-01
-3.34031761e-01 -3.22443575e-01 4.77637529e-01 5.93822658e-01
-4.88795877e-01 -7.66108811e-01 -8.30250204e-01 7.42190599e-01
-2.15093166e-01 -1.29771188e-01 -5.13063550e-01 1.17696941e+00
3.82440053e-02 5.93244076e-01 2.30099946e-01 -2.21205369e-01
4.34751481e-01 2.77357697e-01 3.40133518e-01 -3.42171550e-01
-4.33205694e-01 5.23649514e-01 -4.39490020e-01 -1.04424721e-02
-7.15818405e-02 -6.78706646e-01 -1.56920230e+00 -2.14037821e-01
-8.71451259e-01 -2.86003411e-01 4.34292287e-01 6.79269910e-01
7.26741133e-03 4.21728045e-01 6.95108414e-01 -1.00972497e+00
-5.71857095e-01 -1.07675028e+00 -8.22831810e-01 5.01580298e-01
7.51383901e-02 -6.85030878e-01 -4.38325047e-01 1.89335763e-01] | [12.088459014892578, 2.2674155235290527] |
9f9dc97d-2116-4a85-8524-868606a6b44d | robust-integrated-sensing-and-communication | 2303.07652 | null | https://arxiv.org/abs/2303.07652v1 | https://arxiv.org/pdf/2303.07652v1.pdf | Robust Integrated Sensing and Communication Beamforming for Dual-functional Radar and Communications: Method and Insights | This work presents a novel robust beamforming design dedicated for dual-functional radar and communication (DFRC) base stations (BSs) in the context of integrated sensing and communications (ISAC). The architecture is intended for circumstances with imperfect channel state information (CSI). Our suggested approach demonstrates several tradeoffs for joint radar-communication deployment. Due to the DFRC nature of the design, the beamformer can simultaneously point towards an intended target, while optimizing communication quality of service. We unveil several insights regarding closed form expressions, as well as optimality of the proposed beamformer. Lastly, simulation results demonstrate the effectiveness of the proposed ISAC beamformer. | ['Marwa Chafii', 'Ahmad Bazzi'] | 2023-03-14 | null | null | null | null | ['joint-radar-communication'] | ['robots'] | [ 3.29077214e-01 1.23243161e-01 5.73203862e-02 -6.38643652e-02
-7.27746248e-01 -8.49846423e-01 3.84335339e-01 -5.75977623e-01
1.31495940e-02 7.23179162e-01 2.72031695e-01 -7.32390523e-01
-8.67536724e-01 -5.51630139e-01 5.44173550e-03 -1.20695436e+00
-5.45598447e-01 -1.09637856e-01 -5.99137843e-01 1.33333923e-02
-1.71408325e-01 1.20882261e+00 -7.30786562e-01 -7.65187621e-01
6.29482627e-01 1.36695457e+00 3.08943510e-01 8.02329004e-01
4.91375118e-01 8.94844532e-02 -8.65678549e-01 -3.60629261e-01
6.02420807e-01 -1.23994112e-01 5.89405715e-01 6.49635419e-02
1.43912405e-01 -7.72268549e-02 -4.92439896e-01 9.48408425e-01
5.37088156e-01 -5.26230097e-01 7.40197062e-01 -1.08154607e+00
-2.97932804e-01 5.60745835e-01 -5.46440542e-01 2.29793832e-01
9.45642404e-03 -1.41626164e-01 8.85590732e-01 -8.31484199e-01
-5.16514294e-02 1.00188720e+00 3.87304306e-01 1.36486858e-01
-5.85771501e-01 -1.09656429e+00 1.38275176e-02 -3.24301243e-01
-1.59670794e+00 -8.78178477e-01 5.20513415e-01 -1.05948508e-01
1.56694785e-01 6.35138869e-01 7.84382999e-01 4.55592871e-01
6.41550004e-01 3.54802728e-01 8.14282894e-01 -4.45272416e-01
2.48270646e-01 -2.45947987e-01 1.88763067e-01 3.48232657e-01
1.37580192e+00 7.65481710e-01 -2.77203768e-01 -3.78468156e-01
7.58310020e-01 -2.14048326e-01 -6.63600385e-01 -3.85196954e-01
-1.12734866e+00 7.15893090e-01 2.22503409e-01 3.64668071e-01
-6.50524855e-01 4.84398514e-01 -6.94965899e-01 5.22868574e-01
1.89411446e-01 4.86335248e-01 7.78925344e-02 4.55663472e-01
-8.41552854e-01 -8.11472908e-03 9.59487438e-01 1.33897400e+00
1.05659671e-01 7.25440145e-01 -3.64491701e-01 2.50708401e-01
1.02083564e+00 1.87351537e+00 -6.70940280e-01 -7.27802515e-01
3.65100920e-01 -3.83106321e-01 5.76581597e-01 -9.57317531e-01
-5.40090680e-01 -2.06445694e+00 -9.12738383e-01 -8.98079202e-02
-9.74592566e-02 -1.03786433e+00 -5.80145657e-01 1.54751992e+00
-4.57573421e-02 1.08932346e-01 8.23376775e-01 6.61419392e-01
4.24803585e-01 4.14313674e-01 -5.63794792e-01 -8.13602984e-01
1.18964171e+00 -4.77171876e-02 -1.11861849e+00 -7.34901130e-01
-3.70612890e-02 -7.39202976e-01 -7.52347350e-01 4.66667384e-01
-9.90487933e-01 -4.77188751e-02 -1.53770578e+00 1.19795287e+00
2.94608712e-01 1.90005988e-01 5.04965425e-01 1.50829351e+00
-9.76733685e-01 -8.12810898e-01 -3.66745800e-01 -3.99485007e-02
1.14004292e-01 5.22802919e-02 3.53652090e-01 -4.29160357e-01
-8.14815760e-01 8.01371157e-01 -2.83778578e-01 6.20821834e-01
-7.34190524e-01 -6.98297024e-01 -6.17436230e-01 8.33964646e-02
8.11662972e-02 -6.23706818e-01 1.06616557e+00 -3.22304778e-02
-1.02117467e+00 -1.16315804e-01 -2.32759714e-01 -7.45160580e-01
-1.97258011e-01 -2.21120209e-01 -8.31850290e-01 1.30553603e-01
6.89366013e-02 -2.42584385e-02 5.28054118e-01 -1.53378963e+00
-8.79669130e-01 -4.74191248e-01 -2.88960844e-01 -1.62172560e-02
2.52688587e-01 -2.19076097e-01 1.54822871e-01 -9.84414339e-01
5.93460441e-01 -8.37091744e-01 -6.10724449e-01 -3.47419173e-01
-3.40729207e-01 4.17092234e-01 6.71498299e-01 -8.45631063e-02
1.35126746e+00 -2.25754356e+00 -1.48726283e-02 7.27101028e-01
-6.56236932e-02 6.43198639e-02 -8.57881904e-02 6.36156440e-01
1.34797812e-01 -4.99692768e-01 1.08225226e-01 -3.25902156e-03
-7.54135028e-02 -1.14654943e-01 -4.66139078e-01 9.66271400e-01
-3.24694067e-01 6.82287753e-01 -4.88496929e-01 6.23641610e-02
-1.50009215e-01 1.06784120e-01 -8.12803879e-02 1.36151537e-01
4.23605204e-01 3.56211036e-01 -9.11435008e-01 1.19791687e+00
1.12280273e+00 3.48903984e-01 2.13177130e-01 -4.53591764e-01
-5.45755625e-01 -4.11597908e-01 -1.05725944e+00 1.08441615e+00
-3.90161842e-01 4.80178326e-01 1.09110641e+00 -8.96432459e-01
1.14003241e+00 4.70207036e-01 4.93183345e-01 -6.39356434e-01
2.66623408e-01 2.96024114e-01 3.36853772e-01 1.07350610e-01
3.87902632e-02 -4.65973020e-01 -4.07368481e-01 3.95539582e-01
-1.19909875e-01 -6.29228875e-02 -2.25027502e-01 2.14831620e-01
1.41763031e+00 -8.85007262e-01 7.96553791e-01 -5.44318974e-01
6.09688461e-01 -2.21047342e-01 4.33904946e-01 1.15554368e+00
-9.49098915e-02 -2.62676328e-01 -4.68304098e-01 4.00668055e-01
-1.66109592e-01 -1.24368012e+00 -1.95089072e-01 1.96058288e-01
4.98214722e-01 3.17523152e-01 1.91513404e-01 1.54566050e-01
4.70655233e-01 9.86985743e-01 -1.67651698e-01 3.32979970e-02
-3.27620775e-01 -4.28239346e-01 5.93428791e-01 9.29916576e-02
3.08518767e-01 2.64507502e-01 -9.10720706e-01 4.35095996e-01
3.89423698e-01 -1.10592449e+00 -4.23399620e-02 2.52630651e-01
-5.66035211e-01 -8.78002644e-01 -6.27430141e-01 -3.48574787e-01
3.52425545e-01 1.36741579e+00 5.01071215e-01 -3.19650054e-01
7.51932561e-02 1.22876251e+00 -3.65423828e-01 -5.03917158e-01
3.26653272e-01 -6.12743735e-01 6.71522170e-02 2.74561763e-01
-2.10671112e-01 -5.13616621e-01 -6.02914631e-01 3.91455710e-01
-1.79980233e-01 -3.80328685e-01 1.23807573e+00 3.65830719e-01
-1.47603571e-01 2.61198401e-01 9.09445107e-01 -9.05987993e-02
6.30870104e-01 -5.96668303e-01 -8.66614103e-01 3.80348563e-01
-2.96952099e-01 -2.58942246e-01 -6.50436878e-02 2.61500716e-01
-1.40395463e+00 -2.62042042e-02 5.66436201e-02 -1.99567750e-02
1.60606667e-01 7.10489392e-01 -5.22550881e-01 -8.00680041e-01
1.21550135e-01 3.23607475e-01 -3.21640894e-02 -1.50860384e-01
4.58314240e-01 7.55073845e-01 6.21028960e-01 -1.61648154e-01
1.51311266e+00 6.47214592e-01 4.57608700e-01 -1.14474928e+00
-1.05558395e+00 -6.36899769e-01 -9.86836404e-02 -5.47713220e-01
2.55092859e-01 -1.43610346e+00 -5.10280430e-01 -1.47710204e-01
-1.10530305e+00 3.53055865e-01 1.62026003e-01 1.07179689e+00
-1.51759416e-01 9.55917165e-02 3.11207920e-01 -1.74390018e+00
-4.12339628e-01 -6.23203158e-01 7.80288398e-01 1.20893806e-01
1.93843320e-01 -9.65990543e-01 1.96709316e-02 1.21651746e-01
8.90816510e-01 4.37606543e-01 2.47833312e-01 -2.05018803e-01
-9.36241210e-01 -5.54803133e-01 -1.13265894e-01 -3.46819848e-01
1.68523952e-01 -9.09608126e-01 -4.13969040e-01 -8.06716025e-01
1.39243737e-01 4.25360590e-01 4.63354826e-01 9.41322029e-01
1.59417287e-01 -3.39750677e-01 -9.74522412e-01 7.01250911e-01
1.56700170e+00 8.47021461e-01 4.41174001e-01 -2.17992499e-01
-2.69828409e-01 1.16687402e-01 1.10947776e+00 7.12562501e-01
-2.24869978e-02 4.42932516e-01 8.35046470e-01 2.68152744e-01
1.06197156e-01 5.96945405e-01 2.88698614e-01 2.77474791e-01
3.79068345e-01 -7.94573426e-01 -2.18853369e-01 4.43542033e-01
-1.42677486e+00 -8.40535879e-01 -2.15940848e-01 2.06005621e+00
-6.93820864e-02 -1.31200179e-01 -5.15911520e-01 -1.20032147e-01
5.06102264e-01 1.90404609e-01 -3.91443908e-01 2.09285796e-01
-3.16974491e-01 1.69140518e-01 1.37674594e+00 6.82570577e-01
-8.62831950e-01 -1.26079069e-02 6.72329044e+00 7.41168380e-01
-7.11730957e-01 3.20576072e-01 -2.62059927e-01 -4.06666361e-02
-6.48989141e-01 -4.69475500e-02 -8.74225497e-01 -4.89250617e-03
7.52849758e-01 -4.13442910e-01 -2.07209155e-01 2.71033525e-01
3.64148200e-01 -3.91154200e-01 -5.71649790e-01 1.09724212e+00
2.27044106e-01 -1.35537517e+00 -4.53803509e-01 5.81682682e-01
3.88030022e-01 -2.21833363e-01 1.73898283e-02 1.25427157e-01
5.51594615e-01 -6.35731697e-01 8.32966328e-01 8.92667353e-01
4.95937943e-01 -7.48303473e-01 7.66875982e-01 2.73194313e-01
-1.35879970e+00 -7.38472164e-01 -1.70842055e-02 -1.75603613e-01
5.99105835e-01 1.23712254e+00 -4.23458010e-01 1.13946736e+00
6.47546500e-02 3.16185564e-01 -1.18146196e-01 1.61248875e+00
-2.22042277e-01 6.05547130e-01 -4.70463961e-01 -3.19120407e-01
3.65323365e-01 -2.21207693e-01 1.19529665e+00 1.17895341e+00
1.24126744e+00 6.55957580e-01 8.80449861e-02 4.95910883e-01
5.10468721e-01 -5.44080198e-01 -9.29434776e-01 1.40931472e-01
1.34945118e+00 1.26289070e+00 -1.85064763e-01 2.24386081e-01
-4.36372846e-01 -7.70405084e-02 -5.87661982e-01 6.57591641e-01
-6.16124749e-01 -4.99851435e-01 8.85845304e-01 -2.95856059e-01
6.86869025e-01 -8.33778679e-01 -4.86457884e-01 -4.64074522e-01
-3.53185922e-01 -2.95443296e-01 -4.34414260e-02 -3.98044616e-01
-1.01113367e+00 1.57995015e-01 2.65778881e-02 -1.64473569e+00
5.41316196e-02 -2.45371372e-01 -7.29469240e-01 8.42067599e-01
-1.63529658e+00 -1.26065791e+00 -1.15100868e-01 2.23651513e-01
4.46727574e-02 -6.16049409e-01 4.66499746e-01 1.14682391e-01
-3.37461054e-01 3.27386498e-01 6.00134611e-01 -2.68953681e-01
9.31134596e-02 -6.05183899e-01 -3.79893363e-01 1.30073059e+00
-3.19414139e-01 7.94353485e-01 1.14052474e+00 -6.54698372e-01
-2.22658873e+00 -1.17970431e+00 3.15528035e-01 2.85995871e-01
7.68409312e-01 -2.56022245e-01 4.58151907e-01 3.75612229e-01
5.61518371e-01 -5.68767726e-01 1.23419309e+00 -2.72977371e-02
1.07079171e-01 -3.44885975e-01 -1.02403343e+00 3.67659897e-01
9.04882133e-01 3.88867199e-01 -3.35404426e-01 2.79023468e-01
3.65271538e-01 -7.47862384e-02 -7.14585483e-01 6.06703937e-01
8.28557730e-01 -4.74497706e-01 1.01856303e+00 3.04627657e-01
-7.99296439e-01 -4.66619968e-01 -9.41504121e-01 -1.38086188e+00
-9.44943070e-01 -8.37168157e-01 -2.36178771e-01 1.07903540e+00
4.43781644e-01 -7.93356776e-01 4.36244369e-01 -3.26801002e-01
-2.38657981e-01 -3.50506842e-01 -1.28839660e+00 -1.17958903e+00
-3.88527542e-01 -6.21461809e-01 3.13087374e-01 6.24944195e-02
-1.62558913e-01 4.75266129e-01 -5.44370055e-01 1.23659635e+00
1.35717154e+00 2.05012590e-01 5.68933249e-01 -1.53855193e+00
-3.07471484e-01 -1.23469256e-01 -1.26150504e-01 -1.54529548e+00
-1.25381947e-01 -4.86889422e-01 8.03107917e-02 -1.71231627e+00
-5.21407723e-01 -8.02803576e-01 -3.23699936e-02 -7.06656799e-02
5.23860812e-01 1.54408187e-01 2.38164797e-01 5.50531447e-02
-3.47666651e-01 6.54929638e-01 1.06349850e+00 -1.84199452e-01
3.15826058e-01 9.38018024e-01 -1.38262177e+00 6.72703907e-02
6.44253194e-01 -1.36106148e-01 -3.80119026e-01 -4.28143889e-01
-4.36511151e-02 8.75412762e-01 4.64221746e-01 -1.27073205e+00
6.65709794e-01 -2.69669235e-01 2.34549209e-01 -1.32476091e+00
6.27666056e-01 -1.48137057e+00 6.65844977e-01 9.98148084e-01
2.36680746e-01 -6.61827862e-01 -1.41753659e-01 1.30592060e+00
1.45668522e-01 -2.43901640e-01 7.98804581e-01 4.86794502e-01
-1.64478302e-01 -7.12708011e-02 -8.14054012e-01 -6.85536265e-01
1.36548650e+00 -2.50330061e-01 -3.73334855e-01 -9.31147397e-01
-4.90618885e-01 7.13734508e-01 -4.68290865e-01 -1.13857612e-01
7.12477148e-01 -1.16419315e+00 -7.89751410e-01 1.75433263e-01
-1.94348380e-01 -8.18552315e-01 -1.10608235e-01 1.01028454e+00
1.66238189e-01 1.45166981e+00 2.23446965e-01 -3.77094537e-01
-1.10005641e+00 -6.88980669e-02 3.27867717e-01 2.04458222e-01
-1.27817243e-01 6.92510366e-01 6.10722490e-02 5.19456938e-02
3.06545675e-01 1.60484433e-01 -1.72517881e-01 -6.26870990e-02
5.73417723e-01 1.64423600e-01 3.64250839e-02 -4.14664090e-01
-6.05330646e-01 6.27124548e-01 4.04800802e-01 -3.35326940e-01
1.14278507e+00 -6.44684613e-01 -1.40425535e-02 -4.20754522e-01
3.37583542e-01 7.30199397e-01 -8.62260640e-01 -3.94485295e-01
-1.54534951e-01 -5.78512430e-01 5.05696118e-01 -7.99783230e-01
-1.12815130e+00 7.69240484e-02 7.71894097e-01 2.78766334e-01
9.33214545e-01 -1.69320591e-02 2.64226854e-01 7.91538954e-01
9.79248762e-01 -2.99747854e-01 -1.98159441e-01 4.32417542e-01
8.32099974e-01 -3.19560111e-01 3.28069389e-01 -5.36211491e-01
1.91758871e-02 9.17374074e-01 2.10753810e-02 -8.16481858e-02
1.10526669e+00 8.62263441e-01 -2.39522960e-02 -5.37230313e-01
-7.98270643e-01 -7.35572696e-01 4.59612273e-02 1.04381728e+00
3.00271928e-01 3.64332855e-01 -4.80660826e-01 7.27230191e-01
-2.11587325e-01 -6.85332239e-01 5.87754011e-01 1.16523087e+00
-1.28469920e+00 -7.55728006e-01 -1.02970648e+00 5.75485349e-01
-4.15861420e-03 3.35049666e-02 -2.09667563e-01 6.71949863e-01
-2.25444391e-01 1.73855686e+00 -4.77886461e-02 2.07693246e-03
5.49907625e-01 -6.16271675e-01 4.33009475e-01 -3.11775774e-01
2.39093259e-01 6.72209799e-01 4.36565876e-01 -5.26072271e-02
-4.47940886e-01 -7.88189173e-01 -5.31626105e-01 3.63893569e-01
-6.90264046e-01 7.94488549e-01 5.90250969e-01 9.52941477e-01
3.81740242e-01 7.40271688e-01 1.09687972e+00 -4.50987756e-01
-5.70473909e-01 -7.35782385e-01 -9.68770623e-01 -1.04632723e+00
4.92378533e-01 -9.78538573e-01 -3.92816663e-01 -6.63253188e-01] | [6.281074523925781, 1.2495671510696411] |
6cb4672c-b25e-4d34-9a17-a4cb5188f07d | inductive-and-unsupervised-representation | null | null | https://openreview.net/forum?id=rkem91rtDB | https://openreview.net/pdf?id=rkem91rtDB | Inductive and Unsupervised Representation Learning on Graph Structured Objects | Inductive and unsupervised graph learning is a critical technique for predictive or information retrieval tasks where label information is difficult to obtain. It is also challenging to make graph learning inductive and unsupervised at the same time, as learning processes guided by reconstruction error based loss functions inevitably demand graph similarity evaluation that is usually computationally intractable. In this paper, we propose a general framework SEED (Sampling, Encoding, and Embedding Distributions) for inductive and unsupervised representation learning on graph structured objects. Instead of directly dealing with the computational challenges raised by graph similarity evaluation, given an input graph, the SEED framework samples a number of subgraphs whose reconstruction errors could be efficiently evaluated, encodes the subgraph samples into a collection of subgraph vectors, and employs the embedding of the subgraph vector distribution as the output vector representation for the input graph. By theoretical analysis, we demonstrate the close connection between SEED and graph isomorphism. Using public benchmark datasets, our empirical study suggests the proposed SEED framework is able to achieve up to 10% improvement, compared with competitive baseline methods. | ['Wei Cheng', 'Qianqian Ma', 'Dongjin Song', 'Yanchi Liu', 'Wenchao Yu', 'Lichen Wang', 'Bo Zong', 'Yun Fu', 'Jingchao Ni', 'Haifeng Chen'] | 2020-05-01 | null | null | null | iclr-2020-1 | ['graph-similarity'] | ['graphs'] | [ 4.95089412e-01 5.63183784e-01 -4.84813571e-01 -2.26455316e-01
-7.28924930e-01 -5.42818606e-01 5.32691777e-01 8.11458766e-01
5.68164978e-03 4.98536319e-01 1.58936635e-01 -2.55941570e-01
-3.88452947e-01 -1.10385442e+00 -7.42261589e-01 -6.91606760e-01
-3.89905035e-01 7.65707374e-01 -1.48248628e-01 2.87564635e-01
2.21980765e-01 2.96297759e-01 -1.45785320e+00 -1.35380551e-01
9.39976931e-01 9.10617173e-01 1.03133336e-01 4.88021761e-01
-3.87440175e-01 9.69057143e-01 -6.21130802e-02 -4.37207431e-01
1.74869165e-01 -4.56185669e-01 -1.03368294e+00 3.25962722e-01
3.59167248e-01 -5.16850278e-02 -4.58500564e-01 1.37447143e+00
1.97446153e-01 1.74477533e-01 9.62309897e-01 -1.41129100e+00
-6.05682373e-01 7.58226514e-01 -2.45593980e-01 -1.55214310e-01
5.15051067e-01 -1.34879813e-01 1.58212662e+00 -7.55447626e-01
8.75440896e-01 9.97594893e-01 5.49706221e-01 6.74751699e-02
-1.56678808e+00 -2.94981420e-01 2.04147115e-01 5.60562685e-02
-1.43380523e+00 -1.29659981e-01 1.11395407e+00 -6.06783867e-01
5.37074387e-01 2.26878449e-01 6.24109685e-01 7.90444374e-01
-2.22824663e-01 7.89768815e-01 7.44392991e-01 -4.08403337e-01
4.20736462e-01 3.44201565e-01 3.87943625e-01 1.04578745e+00
4.40958977e-01 8.90691206e-02 -2.05887556e-01 -3.15861940e-01
4.17141259e-01 2.26258427e-01 -3.82305622e-01 -1.06787395e+00
-1.00370181e+00 9.62288916e-01 8.61332357e-01 6.27484471e-02
-3.15980494e-01 2.36601025e-01 3.76759499e-01 5.76727927e-01
5.64575493e-01 4.72029388e-01 4.03659940e-02 3.89096588e-01
-7.69304633e-01 -7.53879249e-02 8.72297168e-01 1.20619190e+00
1.30115688e+00 -9.48014781e-02 -1.67650789e-01 5.27029455e-01
4.47620928e-01 1.93565860e-01 3.77075858e-02 -3.06565553e-01
4.22355980e-01 1.06467044e+00 -2.79945731e-01 -1.37187207e+00
-2.02551916e-01 -4.77775693e-01 -9.88157690e-01 -3.83690119e-01
1.77519113e-01 2.44002759e-01 -7.00520158e-01 1.57228470e+00
3.93232703e-01 2.87848473e-01 -1.48889542e-01 6.87287033e-01
8.37665915e-01 5.36229372e-01 9.94270369e-02 -4.69670922e-01
9.15317059e-01 -7.99738348e-01 -4.01076883e-01 -3.68870273e-02
8.92228603e-01 -2.77286500e-01 1.08776045e+00 2.30782963e-02
-7.87243903e-01 -2.99259126e-01 -9.44338977e-01 6.35223016e-02
-1.58723563e-01 -6.22645989e-02 8.33192766e-01 2.59005696e-01
-1.11207724e+00 6.06260896e-01 -5.64497590e-01 -5.06970823e-01
4.29533005e-01 3.60983104e-01 -5.30230641e-01 -3.09337288e-01
-8.07223260e-01 2.95521438e-01 5.45964539e-01 -2.36824721e-01
-1.00153685e+00 -5.63887119e-01 -1.04447246e+00 2.89620012e-01
4.13319916e-01 -5.84896505e-01 6.42344594e-01 -9.66883421e-01
-9.71101999e-01 8.05998147e-01 2.44840584e-03 -5.99901021e-01
1.90597758e-01 1.12046920e-01 7.59052634e-02 2.93381840e-01
9.52158049e-02 1.95047140e-01 8.90067458e-01 -1.29056442e+00
-1.96618736e-01 -2.78306186e-01 1.25021040e-01 1.32524163e-01
-6.83376074e-01 -3.29561919e-01 -4.31782544e-01 -4.46359575e-01
3.12094480e-01 -7.26796746e-01 -4.38028067e-01 -1.48143113e-01
-6.95154309e-01 -4.60793078e-01 4.71868992e-01 -2.70960540e-01
1.33977938e+00 -2.20433664e+00 3.03575158e-01 7.53978372e-01
7.29831398e-01 -1.51834160e-01 -1.89291850e-01 6.80579484e-01
-1.38407454e-01 1.97684109e-01 -4.55521315e-01 -7.30432570e-02
1.00452349e-01 8.23238716e-02 -3.26515198e-01 6.60841882e-01
2.18536794e-01 9.28201795e-01 -1.36660480e+00 -5.45230389e-01
1.21440634e-01 1.70137733e-01 -7.13399112e-01 6.38130367e-01
-3.34978104e-01 2.04336956e-01 -6.24518275e-01 5.26553571e-01
3.56439769e-01 -7.88372993e-01 6.49651051e-01 -2.22906709e-01
5.55738091e-01 3.67207348e-01 -1.18776429e+00 1.51674581e+00
-2.21880823e-01 3.70317668e-01 -4.57530677e-01 -1.55761015e+00
1.05537164e+00 1.28617153e-01 4.52454686e-01 -2.75446683e-01
6.00350164e-02 -5.37745748e-03 -3.35442692e-01 -2.99095452e-01
2.37046048e-01 4.44297083e-02 -1.18972823e-01 6.11356199e-01
4.80688483e-01 6.76359981e-02 2.63843775e-01 6.52301073e-01
1.46361625e+00 -1.79594234e-01 4.90816325e-01 -3.63689959e-01
5.44365764e-01 -1.60618499e-01 3.62589598e-01 6.27906024e-01
1.71606719e-01 3.90222073e-01 9.33334410e-01 -5.18599391e-01
-9.77478683e-01 -1.32946897e+00 1.79227382e-01 9.97624338e-01
2.19874546e-01 -8.17682147e-01 -6.23463511e-01 -1.07274187e+00
6.91251010e-02 3.78514886e-01 -5.08662462e-01 -3.66987377e-01
-2.39287376e-01 -3.45537633e-01 1.55121952e-01 2.26325661e-01
4.78025377e-02 -7.68032074e-01 2.14070529e-01 1.87225491e-01
1.60951465e-02 -1.02610314e+00 -4.64568436e-01 1.49825752e-01
-1.03970444e+00 -1.38709402e+00 -3.04215044e-01 -1.26376224e+00
1.25929523e+00 2.70826221e-01 1.33577347e+00 4.73022908e-01
-2.83787131e-01 5.08871675e-01 -4.12519008e-01 1.72504067e-01
-4.18711990e-01 2.71157742e-01 -1.78816870e-01 2.76234418e-01
1.26945391e-01 -8.06774795e-01 -4.35408622e-01 -1.49135396e-01
-8.88774335e-01 6.22423217e-02 4.89902258e-01 9.65736806e-01
7.71133780e-01 -3.20843868e-02 3.71608138e-01 -1.57339370e+00
7.15552807e-01 -8.01821411e-01 -7.22852111e-01 5.52282929e-01
-1.05095124e+00 5.18688738e-01 6.78248703e-01 -1.66205108e-01
-4.69788820e-01 1.13947473e-01 3.20229232e-01 -4.44547892e-01
8.31769779e-02 8.85070801e-01 -2.13360265e-01 -1.31030470e-01
6.86438322e-01 2.98972517e-01 4.52094711e-02 -3.12818468e-01
6.26090050e-01 3.68457198e-01 1.67948171e-01 -5.90520620e-01
1.05295384e+00 1.89074710e-01 1.60795644e-01 -8.21812034e-01
-7.15851486e-01 -6.62153721e-01 -5.59291840e-01 -2.44524330e-01
4.88897592e-01 -7.35093892e-01 -4.85276073e-01 -1.82237849e-01
-7.40539372e-01 -9.36247483e-02 -3.57947171e-01 4.16483194e-01
-5.49861848e-01 6.71807349e-01 -3.38470370e-01 -7.86780238e-01
-3.23297381e-01 -9.74105000e-01 9.42225695e-01 -2.40508601e-01
-8.73111337e-02 -1.26055527e+00 1.14520647e-01 -2.29781903e-02
8.59084651e-02 5.21877170e-01 1.40051138e+00 -8.97435486e-01
-9.08128500e-01 -4.56715852e-01 -2.85485893e-01 1.08712561e-01
1.93429336e-01 -2.19790399e-01 -5.98299980e-01 -6.50170326e-01
-3.06775421e-01 -5.90755165e-01 7.31162131e-01 5.50521463e-02
1.32149434e+00 -5.33895910e-01 -4.26095605e-01 6.71489120e-01
1.65161705e+00 -3.66356850e-01 5.89099169e-01 -8.47149715e-02
9.40249979e-01 5.13157010e-01 4.77196395e-01 3.09303790e-01
3.86859804e-01 2.90122658e-01 2.73471028e-01 1.01386495e-01
5.85097335e-02 -7.92641997e-01 1.63838938e-01 1.05058396e+00
-7.58354664e-02 -2.68748373e-01 -7.31005073e-01 5.57244420e-01
-1.74731147e+00 -7.58619249e-01 9.66157094e-02 2.45056272e+00
7.25154459e-01 6.95536211e-02 1.51507333e-01 1.55006200e-01
8.04862738e-01 1.87503338e-01 -3.95672321e-01 -6.18042275e-02
2.66407877e-01 3.80945534e-01 3.84135991e-01 6.10722899e-01
-9.66397405e-01 9.68878627e-01 5.84617949e+00 5.44729233e-01
-9.42811728e-01 -1.34460390e-01 5.79592943e-01 3.40432853e-01
-8.16626251e-01 4.12673414e-01 -2.95423478e-01 2.74691999e-01
9.17946458e-01 -4.80130613e-01 6.29674554e-01 8.60458314e-01
-3.31572652e-01 3.75270247e-01 -1.52930391e+00 1.00139499e+00
3.63663654e-03 -1.29143572e+00 2.93175817e-01 1.68393657e-01
7.86323369e-01 -1.43553287e-01 -9.37925279e-02 4.19922739e-01
4.50356334e-01 -1.05654609e+00 2.49871492e-01 4.98860151e-01
7.76532590e-01 -7.30806172e-01 5.55927634e-01 3.20467114e-01
-1.36403143e+00 1.42024711e-01 -5.88829637e-01 8.54190588e-02
-2.25403160e-01 7.97279298e-01 -1.13880730e+00 7.67471075e-01
1.67422742e-01 1.01344979e+00 -5.70273042e-01 9.45850194e-01
-4.58765924e-01 9.45021391e-01 -6.74895421e-02 -2.37823069e-01
1.80561572e-01 -4.55707371e-01 5.85645437e-01 1.10172117e+00
1.03997417e-01 -2.05631763e-01 6.45024598e-01 7.93285012e-01
-5.78875303e-01 5.01663387e-01 -1.22851038e+00 -5.69459498e-01
5.25903523e-01 1.22897470e+00 -7.74185836e-01 -3.98938358e-01
-2.33447611e-01 8.66054237e-01 1.06197464e+00 4.53207284e-01
-5.42629898e-01 -3.13710868e-01 2.74845004e-01 2.45270953e-01
1.60195500e-01 -1.21764399e-01 1.71405926e-01 -1.23556685e+00
7.40106702e-02 -7.79398739e-01 5.73299408e-01 -2.68373191e-01
-1.42265975e+00 5.27676046e-01 -2.91630719e-02 -1.21146607e+00
-3.39341462e-01 -4.68368709e-01 -5.47020972e-01 5.67402482e-01
-1.35609984e+00 -9.72551167e-01 -4.26683992e-01 5.23150742e-01
4.34619077e-02 -1.90428391e-01 9.12716568e-01 1.04450531e-01
-2.40178049e-01 7.64148831e-01 1.58005878e-01 2.50648499e-01
2.63190359e-01 -1.50924337e+00 2.63308614e-01 6.60084307e-01
6.48645520e-01 6.40766799e-01 5.24974346e-01 -5.45616329e-01
-1.78260159e+00 -1.42659926e+00 7.72388220e-01 -1.45411193e-01
7.93814719e-01 -5.97871304e-01 -9.55722809e-01 7.17388272e-01
-9.39047858e-02 4.20208007e-01 7.13893592e-01 3.35001469e-01
-5.06639540e-01 -1.45847514e-01 -1.09888518e+00 4.75317270e-01
1.25299668e+00 -8.20188284e-01 -2.74957478e-01 7.10956693e-01
8.03324997e-01 4.34123464e-02 -1.04721379e+00 3.83738697e-01
9.08248648e-02 -6.49709582e-01 8.97285581e-01 -1.05205667e+00
2.97412932e-01 -2.20252946e-01 -2.18918249e-01 -1.18592775e+00
-4.43687260e-01 -5.89489281e-01 -2.43316457e-01 1.28595340e+00
4.25977647e-01 -6.40973806e-01 8.68301630e-01 2.40163699e-01
2.54636347e-01 -7.48387098e-01 -5.05772591e-01 -7.02613175e-01
-2.88111150e-01 -3.00984144e-01 5.96366942e-01 1.01719975e+00
1.83226228e-01 5.91102123e-01 -2.55087525e-01 3.06155980e-01
1.02012670e+00 4.81274813e-01 9.81021583e-01 -1.39760673e+00
-3.62793714e-01 -2.43782997e-01 -9.96003866e-01 -8.35829496e-01
4.56397504e-01 -1.71735907e+00 2.66100131e-02 -1.69957376e+00
2.81983078e-01 -5.11984408e-01 -4.61888999e-01 2.17121288e-01
-2.15501666e-01 -7.77082890e-02 -9.69173163e-02 2.15542823e-01
-7.56029427e-01 6.80784762e-01 1.01319098e+00 -4.20459837e-01
3.50637101e-02 -1.94174424e-01 -7.31105983e-01 6.20110750e-01
5.97508609e-01 -5.22116959e-01 -7.90030718e-01 -1.16146326e-01
3.02490324e-01 7.42696747e-02 4.56494659e-01 -7.40344465e-01
2.61606544e-01 -6.07410744e-02 -6.16082288e-02 -2.13704005e-01
-1.80163413e-01 -7.86408663e-01 1.96502611e-01 4.08302635e-01
-6.87899053e-01 -2.30324849e-01 -3.88343841e-01 1.03072977e+00
-2.34700978e-01 -2.50357300e-01 5.57915449e-01 -6.74690455e-02
-5.00578940e-01 7.60575354e-01 1.90136313e-01 3.70932847e-01
9.09425318e-01 4.93609346e-02 7.83369131e-03 -3.45383286e-01
-7.72062719e-01 1.68095395e-01 5.29044032e-01 5.94837517e-02
8.27456474e-01 -1.55869460e+00 -5.56375206e-01 3.05376023e-01
5.68853736e-01 -5.02342731e-03 -2.46706456e-01 6.36662483e-01
-4.53724086e-01 9.53853875e-02 3.02477896e-01 -6.28074884e-01
-1.01816428e+00 8.27160299e-01 -8.66625607e-02 -5.65952539e-01
-7.07185328e-01 8.45714211e-01 3.57760459e-01 -6.01614714e-01
3.66106093e-01 -1.30573988e-01 -1.09267354e-01 -1.16314657e-01
1.77165136e-01 1.41854674e-01 -3.54955532e-02 -3.52049261e-01
-1.05519742e-01 2.96957016e-01 -1.15685932e-01 4.04488266e-01
1.40365005e+00 1.00695118e-01 -3.23756963e-01 3.49057257e-01
1.61798036e+00 -2.25879714e-01 -7.95678973e-01 -6.04000270e-01
4.80007559e-01 -5.53798914e-01 -6.35311604e-02 -9.24106613e-02
-9.63110328e-01 6.77127957e-01 2.88592666e-01 5.33472896e-01
9.62490261e-01 3.50346744e-01 5.24987221e-01 6.51756346e-01
5.06562471e-01 -6.96334183e-01 1.64225265e-01 1.70426205e-01
7.42528975e-01 -1.23968518e+00 5.00363782e-02 -7.75627196e-01
-2.83219755e-01 9.72833216e-01 2.85441816e-01 -6.05881929e-01
7.41449475e-01 -2.29620785e-01 -6.28678381e-01 -6.36986256e-01
-6.80810928e-01 -3.88761580e-01 6.08248174e-01 5.51145375e-01
4.88151997e-01 2.58492678e-01 -2.23762289e-01 2.82608718e-01
-1.01856202e-01 -3.44681114e-01 1.20199554e-01 6.57404363e-01
-5.20532012e-01 -1.25026596e+00 2.25890979e-01 8.34511459e-01
-7.32018277e-02 -1.80766001e-01 -4.18282270e-01 6.68956757e-01
-3.78953099e-01 6.07693434e-01 -9.02621746e-02 -5.22299647e-01
8.12405422e-02 -5.79693876e-02 4.12476122e-01 -9.33008313e-01
-1.22708447e-01 -1.85573444e-01 1.00490712e-01 -5.39064169e-01
-2.62513697e-01 -4.15279180e-01 -1.16902030e+00 -6.95746019e-02
-3.31486136e-01 5.15364945e-01 2.40912706e-01 8.44566822e-01
3.09784561e-01 2.64637947e-01 1.01390481e+00 -4.22364235e-01
-6.86950505e-01 -7.63871372e-01 -7.43036509e-01 7.84705043e-01
7.26044327e-02 -7.12491632e-01 -6.08046591e-01 7.85244033e-02] | [7.200651168823242, 6.22418212890625] |
f78f8686-002b-4d3d-91a8-6b30f1ca19f7 | forensic-video-steganalysis-in-spatial-domain | 2305.18070 | null | https://arxiv.org/abs/2305.18070v1 | https://arxiv.org/pdf/2305.18070v1.pdf | Forensic Video Steganalysis in Spatial Domain by Noise Residual Convolutional Neural Network | This research evaluates a convolutional neural network (CNN) based approach to forensic video steganalysis. A video steganography dataset is created to train a CNN to conduct forensic steganalysis in the spatial domain. We use a noise residual convolutional neural network to detect embedded secrets since a steganographic embedding process will always result in the modification of pixel values in video frames. Experimental results show that the CNN-based approach can be an effective method for forensic video steganalysis and can reach a detection rate of 99.96%. Keywords: Forensic, Steganalysis, Deep Steganography, MSU StegoVideo, Convolutional Neural Networks | ['Meike Kombrink', 'Zeno Geradts', 'Mart Keizer'] | 2023-05-29 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 4.82039720e-01 -5.95950149e-02 3.87134910e-01 2.71850765e-01
8.23837370e-02 -2.24883214e-01 3.01684707e-01 -6.38258815e-01
-2.31465191e-01 4.40969944e-01 -2.76661903e-01 -9.30724144e-01
4.60257560e-01 -1.07329774e+00 -5.96438110e-01 -7.08764195e-01
-6.55695379e-01 -2.94926256e-01 3.51363093e-01 -5.21552444e-01
6.37967646e-01 6.16642237e-01 -1.06711888e+00 5.91892421e-01
1.95784807e-01 8.55755448e-01 -2.70587325e-01 1.17858315e+00
2.33826697e-01 1.31883812e+00 -1.00748348e+00 -3.39991480e-01
7.04246223e-01 -6.84809029e-01 -4.79959846e-01 -1.63849127e-02
9.21350569e-02 -7.77322829e-01 -1.27387166e+00 1.51092398e+00
4.94114608e-01 -2.48231083e-01 3.37483138e-01 -1.04761624e+00
-7.59355485e-01 7.51305640e-01 -3.72739732e-01 9.05999959e-01
1.50817409e-01 5.95721185e-01 -1.61675274e-01 -4.43281829e-01
5.64130127e-01 1.33985198e+00 8.91832232e-01 5.12135267e-01
-3.70093882e-01 -1.31421149e+00 -1.05912602e+00 5.45350611e-01
-1.21236420e+00 -3.78003448e-01 1.02861381e+00 -1.39814898e-01
9.01936531e-01 -1.00112446e-02 7.15402484e-01 8.05938840e-01
8.09580266e-01 3.04342568e-01 9.67520356e-01 -5.80774903e-01
-1.19547248e-01 -2.48201340e-01 -2.47272089e-01 9.15251851e-01
8.12681794e-01 6.07579291e-01 -1.93191051e-01 8.84807408e-02
1.14281476e+00 -2.13058433e-03 -1.81957632e-01 4.32755798e-01
-8.15825462e-01 1.27805007e+00 1.67076916e-01 4.52464759e-01
-3.45024288e-01 7.75031388e-01 1.13820326e+00 1.01115060e+00
6.61545470e-02 2.74878412e-01 2.97653973e-01 4.06946614e-02
-1.25713432e+00 3.71044800e-02 6.68060064e-01 5.13745368e-01
3.45021516e-01 9.84784424e-01 3.68713021e-01 -7.22041056e-02
4.63045657e-01 7.33185291e-01 5.60023904e-01 -1.17851853e+00
4.08370614e-01 1.12807333e-01 -6.05492353e-01 -1.58231378e+00
3.85457993e-01 3.62459272e-01 -6.72018707e-01 7.95534313e-01
8.27946812e-02 -1.77255929e-01 -1.02205169e+00 7.15116322e-01
-1.64910659e-01 6.73029542e-01 3.11195493e-01 3.93842041e-01
9.49303150e-01 5.27162373e-01 -7.57459179e-02 2.94980705e-01
1.13833427e+00 -4.99501318e-01 -8.68028164e-01 1.18656494e-01
5.64593315e-01 -7.38452792e-01 -2.97968537e-01 5.09957075e-01
-1.13045776e+00 -3.99887145e-01 -1.43909788e+00 4.34862226e-01
-5.73789418e-01 -6.42446339e-01 5.40439844e-01 1.61585426e+00
-1.17121255e+00 7.65046239e-01 -5.01361728e-01 1.77282587e-01
8.36276531e-01 6.79141104e-01 -6.59762800e-01 -3.21310997e-01
-1.63010705e+00 6.72879577e-01 1.27929831e+00 1.28905237e-01
-1.44073343e+00 2.62487400e-02 -1.10162866e+00 2.26975650e-01
-7.47043565e-02 3.22188407e-01 5.47641158e-01 -1.06667125e+00
-9.29359019e-01 9.61894274e-01 5.78295708e-01 -8.49453568e-01
5.04140079e-01 7.78732955e-01 -1.12332845e+00 9.13777769e-01
-2.12646231e-01 2.91320592e-01 1.24160182e+00 -6.70543134e-01
-4.49881762e-01 1.20592117e-01 -1.74468949e-01 -6.74637437e-01
-2.59653717e-01 5.89122534e-01 6.54741526e-02 -6.17275357e-01
-5.64981103e-02 -6.47695839e-01 1.56291425e-01 -2.65558153e-01
-2.32494161e-01 3.56261611e-01 1.75787377e+00 -1.21081197e+00
1.14839411e+00 -2.03696394e+00 -8.19974661e-01 5.55025995e-01
5.16487002e-01 1.22521555e+00 -8.33197776e-03 3.18699509e-01
-6.72724187e-01 4.03161645e-01 -1.64868921e-01 4.10420299e-01
-5.97208083e-01 -1.01018064e-01 6.87325224e-02 1.04640043e+00
-8.70615318e-02 1.15597296e+00 -8.49758625e-01 -6.50420606e-01
6.11028492e-01 7.31199205e-01 -1.76516965e-01 -2.25709558e-01
4.71743494e-01 2.89749444e-01 -2.74013668e-01 7.19511628e-01
1.18274450e+00 3.36718820e-02 4.18878138e-01 1.38110802e-01
2.45528370e-01 -7.51447529e-02 -7.91975975e-01 8.79761457e-01
1.36123329e-01 1.69152606e+00 -5.38024716e-02 -1.19422746e+00
1.19622838e+00 7.79177666e-01 3.29308957e-01 -5.19972086e-01
1.00386369e+00 4.10911381e-01 2.92370319e-01 -1.13235700e+00
4.28534329e-01 -1.52652919e-01 2.08560944e-01 7.45259225e-01
1.66338205e-01 4.07851756e-01 -3.48189287e-03 -4.83857580e-02
1.59232414e+00 -2.55325675e-01 1.14301063e-01 -2.12559383e-02
9.42963779e-01 -2.60171920e-01 2.70506889e-01 7.73474514e-01
-5.95297992e-01 3.38770211e-01 6.91203117e-01 -8.59710574e-01
-1.68749034e+00 -3.06476444e-01 4.16747719e-01 2.35088140e-01
-2.22420812e-01 1.95105240e-01 -1.13474500e+00 -7.19089270e-01
-3.14172864e-01 1.87090531e-01 -6.22233748e-01 -4.27882731e-01
-1.20010126e+00 -1.60923988e-01 1.58913612e+00 2.63089448e-01
1.50815809e+00 -1.64801419e+00 -8.70284975e-01 3.95619720e-01
1.07982732e-01 -1.37767613e+00 -6.89970776e-02 -2.47960523e-01
-9.63279486e-01 -1.81198001e+00 -4.27151978e-01 -9.47060347e-01
5.08632779e-01 5.54617822e-01 5.02976954e-01 7.63971984e-01
-3.86501342e-01 -1.47409782e-01 -5.46594262e-01 -2.90902525e-01
-1.03312457e+00 -5.00264704e-01 -3.84081125e-01 -1.74977541e-01
8.36182654e-01 -4.98233676e-01 -4.45463002e-01 -1.58153158e-02
-1.40087497e+00 -6.42949224e-01 1.85522601e-01 6.19626760e-01
-4.19661671e-01 7.52984524e-01 -6.35011494e-02 -8.57499778e-01
8.21515977e-01 -6.93505168e-01 -4.84677792e-01 -8.39929804e-02
-4.40737128e-01 -3.80422503e-01 4.98677909e-01 -4.18374598e-01
-4.41145271e-01 -6.83693290e-01 4.04895581e-02 -1.18990803e+00
-8.40710923e-02 4.47191030e-01 8.98262039e-02 -9.71837223e-01
6.28733337e-01 4.48575288e-01 5.70554316e-01 -6.62367493e-02
-3.52079868e-01 7.53654361e-01 7.15629816e-01 3.13916028e-01
1.11145234e+00 7.16336250e-01 3.09031963e-01 -8.58017564e-01
5.58793068e-01 -2.06361517e-01 -4.07441467e-01 -6.34892106e-01
9.52081263e-01 -7.56171584e-01 -1.00359416e+00 8.99181485e-01
-1.40738654e+00 1.75578687e-02 2.03059375e-01 4.04420853e-01
-1.54754698e-01 1.02371502e+00 -8.31184745e-01 -6.76853359e-01
-4.60402310e-01 -1.26660693e+00 2.84893602e-01 -1.18276983e-01
1.76333666e-01 -1.52200198e+00 -1.38734475e-01 3.09875280e-01
5.55897713e-01 1.14416385e+00 3.36280465e-01 -1.00916386e+00
-5.19996464e-01 -7.85465658e-01 -3.65077972e-01 8.71599436e-01
-1.25998408e-01 -6.58615828e-02 -8.23355258e-01 -4.32407796e-01
4.92152542e-01 1.09636061e-01 8.68175387e-01 3.59919935e-01
1.18490124e+00 -4.03232366e-01 3.82572524e-02 1.13217139e+00
1.65965819e+00 7.12755501e-01 1.84717584e+00 9.76584554e-01
7.20997155e-01 2.74136811e-01 -2.41802838e-02 2.23120928e-01
-2.88039535e-01 -2.62241930e-01 8.55298817e-01 1.87124595e-01
-2.18590245e-01 1.64194047e-01 4.96344030e-01 4.50043947e-01
-5.17483771e-01 -8.57216835e-01 -9.71082032e-01 3.14526469e-01
-1.20133591e+00 -1.60000932e+00 -5.91374457e-01 1.54569840e+00
-2.31696472e-01 1.80446774e-01 -3.05125952e-01 8.81803811e-01
1.51348186e+00 3.96846384e-01 5.93999885e-02 -8.62011969e-01
-2.67087251e-01 7.46640623e-01 1.27292919e+00 1.29314080e-01
-1.13507032e+00 1.11829674e+00 6.88988686e+00 8.84384930e-01
-1.48681700e+00 5.30596554e-01 3.04602295e-01 4.15926456e-01
1.23523131e-01 -2.46603712e-01 -1.30607128e-01 8.71165514e-01
1.40026414e+00 2.30061173e-01 2.67259359e-01 4.99197245e-01
3.31973553e-01 1.89673394e-01 -9.98255685e-02 1.15101147e+00
5.05557299e-01 -1.71354568e+00 -3.82157415e-02 6.47990763e-01
6.18674457e-01 -3.06003034e-01 1.94312304e-01 -1.41100422e-01
2.33668000e-01 -1.41829824e+00 3.85531813e-01 -4.67489250e-02
1.20984852e+00 -1.20215333e+00 1.20295548e+00 -2.81299710e-01
-1.06432962e+00 -1.67048112e-01 -7.02252805e-01 1.26170248e-01
1.58145487e-01 -1.03051990e-01 -9.30317521e-01 1.85916528e-01
7.48163640e-01 5.88944316e-01 -1.77265197e-01 9.53291655e-01
-3.33840728e-01 1.08791721e+00 1.28257751e-01 2.52424240e-01
8.29910517e-01 1.61269426e-01 7.58390963e-01 1.38901472e+00
5.81462920e-01 6.72275126e-02 -6.12989247e-01 6.89707518e-01
-1.34924427e-01 -4.57757860e-01 -1.13311100e+00 -3.00764918e-01
3.14859092e-01 7.44089425e-01 -1.06709111e+00 -6.11695588e-01
-3.40866297e-02 7.41555691e-01 -7.78733015e-01 1.12868130e-01
-7.97733903e-01 -9.04099464e-01 3.23624820e-01 2.77338773e-01
1.06505239e+00 -1.60258591e-01 -5.67165799e-02 -9.85005558e-01
-5.52712202e-01 -1.18795800e+00 2.07202420e-01 -6.21200919e-01
-4.04388905e-01 4.10493851e-01 -5.79184331e-02 -1.39806151e+00
-3.02819014e-01 -9.94800568e-01 -1.18085408e+00 6.36045814e-01
-1.47194493e+00 -1.24852562e+00 -2.96510637e-01 9.66164470e-01
3.52998048e-01 -1.18093085e+00 3.93646061e-01 4.47157443e-01
-1.95129007e-01 3.74581188e-01 1.84361711e-01 1.02129817e+00
2.52021164e-01 -4.41520602e-01 7.68859088e-01 1.25044167e+00
-8.36747169e-01 4.95052755e-01 9.10142839e-01 -1.12753820e+00
-1.21554494e+00 -1.05853367e+00 6.39674187e-01 2.55117193e-02
4.89282608e-01 2.09728360e-01 -8.40794802e-01 9.28776503e-01
5.62841356e-01 2.06636831e-01 5.86733282e-01 -1.32390964e+00
-5.02890468e-01 5.21103561e-01 -1.89707279e+00 8.28458145e-02
3.61944675e-01 -5.09700298e-01 -3.74753803e-01 -1.87285747e-02
2.54307270e-01 -1.82788476e-01 -6.59126759e-01 -6.19680360e-02
5.65908492e-01 -1.22460842e+00 1.14783752e+00 -5.23551404e-01
7.77899861e-01 -1.25397161e-01 -7.12650418e-02 -4.77594107e-01
-1.42205777e-02 -9.73313987e-01 -5.03723264e-01 5.30556738e-01
-3.30637485e-01 -7.35706687e-01 9.04299200e-01 -2.88115203e-01
1.35246411e-01 1.33453906e-01 -1.01459253e+00 -9.61086094e-01
-3.35554825e-03 -4.84685004e-01 9.03694093e-01 1.28898275e+00
-3.86374831e-01 -7.33659804e-01 -9.77600098e-01 4.03828686e-04
1.13117146e+00 -1.08690417e+00 4.59710717e-01 -1.00076663e+00
2.63710439e-01 -2.03589752e-01 -1.52438176e+00 -9.39555168e-02
1.66202456e-01 -4.42868382e-01 -4.21586275e-01 -8.65752816e-01
-3.12995404e-01 8.48644748e-02 -2.59121925e-01 1.59200370e-01
5.55207789e-01 1.03008676e+00 1.81033656e-01 2.58289516e-01
-3.03304195e-01 -4.11339909e-01 1.16462314e+00 -3.56360942e-01
3.40907276e-01 -4.47096914e-01 -3.67122233e-01 6.47276580e-01
1.01557767e+00 -1.02259636e+00 -3.40090878e-02 1.19930506e-03
1.71552837e-01 2.14859352e-01 5.94822645e-01 -1.32249224e+00
1.80407435e-01 7.98600465e-02 4.72679645e-01 -3.96221429e-01
6.03195205e-02 -7.87239790e-01 3.80385190e-01 1.37869632e+00
7.23137632e-02 6.74535707e-02 4.14978452e-02 4.18729126e-01
-3.11751842e-01 -7.01260626e-01 8.58733535e-01 -7.29470015e-01
-1.09833717e+00 2.81224310e-01 -1.14612663e+00 -5.53346097e-01
1.20638835e+00 -1.23119569e+00 -1.90507054e-01 -4.88404006e-01
-5.00488162e-01 -4.15606827e-01 4.31451410e-01 2.93648094e-02
1.32095587e+00 -1.28092659e+00 -7.00759292e-01 4.44185436e-01
-3.85934323e-01 -8.92872274e-01 1.38054937e-01 4.69803900e-01
-1.90063906e+00 4.59385365e-01 -1.05980933e+00 -1.12764731e-01
-1.67682946e+00 5.35385311e-01 5.81267297e-01 -1.89419240e-01
-7.04073310e-01 5.60727119e-01 -5.98239064e-01 1.22999050e-01
-2.36892864e-01 3.04943442e-01 -7.02908099e-01 -3.74212027e-01
1.04195225e+00 9.37287390e-01 -2.20640138e-01 -1.02766490e+00
2.60436162e-02 1.43190786e-01 2.14497074e-01 1.52574018e-01
1.40293145e+00 8.80580395e-03 -3.68514955e-01 -6.01579964e-01
1.82667494e+00 -5.33230722e-01 -7.01577008e-01 2.39851147e-01
-3.11035395e-01 -8.27353656e-01 3.98384839e-01 1.07055247e-01
-1.72642660e+00 1.08737648e+00 8.20376873e-01 4.74162191e-01
7.72493780e-01 -7.23026156e-01 1.20437098e+00 5.14743686e-01
1.87979117e-01 -1.01866913e+00 1.92919672e-01 5.79806387e-01
1.26442626e-01 -1.21537948e+00 -8.33448619e-02 -1.49117857e-01
-2.43665874e-01 2.06145620e+00 1.54499084e-01 -9.72784460e-01
8.39560151e-01 2.63621122e-01 -3.80408652e-02 -7.99656868e-01
7.53547475e-02 3.11347038e-01 -3.32829386e-01 1.05113935e+00
4.90607619e-02 -1.98400378e-01 -1.37209117e-01 -7.40885913e-01
-1.04420014e-01 1.11367300e-01 1.05504739e+00 1.31882381e+00
-8.47270131e-01 -8.86785567e-01 -1.15645111e+00 1.22939110e-01
-1.30611646e+00 -9.74415094e-02 -6.85539246e-02 1.09171939e+00
3.88942212e-01 9.50482309e-01 2.04308510e-01 -7.25787759e-01
-3.60449016e-01 -2.07590997e-01 3.62805456e-01 -1.62006274e-01
-8.27660799e-01 -1.43168822e-01 -5.29413261e-02 -5.48879981e-01
-5.49545050e-01 -3.91758859e-01 -9.30285454e-01 -1.41894424e+00
-2.50067621e-01 -4.18279395e-02 9.85818684e-01 9.75394070e-01
-3.43875736e-01 7.75483906e-01 6.09685063e-01 -6.77190006e-01
3.58200669e-02 -8.72249126e-01 -7.34681010e-01 2.07207531e-01
5.38767517e-01 -2.24483684e-01 -5.86810946e-01 3.37018043e-01] | [4.275097846984863, 8.09187126159668] |
8f1f7598-cbf4-4c6c-9820-e0429d7841b5 | unovost-unsupervised-offline-video-object | 2001.05425 | null | https://arxiv.org/abs/2001.05425v1 | https://arxiv.org/pdf/2001.05425v1.pdf | UnOVOST: Unsupervised Offline Video Object Segmentation and Tracking | We address Unsupervised Video Object Segmentation (UVOS), the task of automatically generating accurate pixel masks for salient objects in a video sequence and of tracking these objects consistently through time, without any input about which objects should be tracked. Towards solving this task, we present UnOVOST (Unsupervised Offline Video Object Segmentation and Tracking) as a simple and generic algorithm which is able to track and segment a large variety of objects. This algorithm builds up tracks in a number stages, first grouping segments into short tracklets that are spatio-temporally consistent, before merging these tracklets into long-term consistent object tracks based on their visual similarity. In order to achieve this we introduce a novel tracklet-based Forest Path Cutting data association algorithm which builds up a decision forest of track hypotheses before cutting this forest into paths that form long-term consistent object tracks. When evaluating our approach on the DAVIS 2017 Unsupervised dataset we obtain state-of-the-art performance with a mean J &F score of 67.9% on the val, 58% on the test-dev and 56.4% on the test-challenge benchmarks, obtaining first place in the DAVIS 2019 Unsupervised Video Object Segmentation Challenge. UnOVOST even performs competitively with many semi-supervised video object segmentation algorithms even though it is not given any input as to which objects should be tracked and segmented. | ['Idil Esen Zulfikar', 'Jonathon Luiten', 'Bastian Leibe'] | 2020-01-15 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 3.43977600e-01 -1.45964399e-01 -3.88087690e-01 -2.05618441e-01
-7.52582908e-01 -8.89704823e-01 5.34269333e-01 3.54189873e-01
-6.19742215e-01 4.05253381e-01 -1.49871916e-01 -1.00337468e-01
-3.59811671e-02 -5.08384705e-01 -1.00245953e+00 -5.20079672e-01
-3.88638228e-01 8.18676054e-01 1.09543121e+00 2.51931429e-01
8.92465785e-02 3.46039444e-01 -1.74074340e+00 3.63183945e-01
7.56426692e-01 1.20732009e+00 1.99977919e-01 1.01525033e+00
-1.53692156e-01 9.79160070e-01 -2.28805169e-01 -3.00673485e-01
4.86440420e-01 -2.64624834e-01 -1.31519580e+00 6.51459873e-01
9.53776717e-01 -2.27619246e-01 -7.85516277e-02 1.07950985e+00
9.69760993e-04 1.45087525e-01 6.81057990e-01 -1.32249475e+00
-5.92057221e-02 8.92271519e-01 -6.63154662e-01 6.47525311e-01
1.43934280e-01 2.32921153e-01 9.93545949e-01 -8.44814837e-01
1.12066925e+00 8.76960456e-01 7.08980560e-01 3.58350933e-01
-1.28738987e+00 -4.13710177e-01 3.95523697e-01 2.92242825e-01
-1.26228476e+00 -2.63940990e-01 3.18772495e-01 -8.75603616e-01
7.67789543e-01 3.45557630e-01 7.95318544e-01 6.10798478e-01
-1.09974734e-01 1.23978794e+00 8.63336384e-01 -2.01778814e-01
2.41156280e-01 -7.84087852e-02 6.52303815e-01 5.73896289e-01
1.04216218e-01 -5.10475263e-02 -4.63084698e-01 2.07967088e-01
5.12406707e-01 -1.86172605e-01 -5.12390248e-02 -5.90528250e-01
-1.34764433e+00 4.16022062e-01 5.09511650e-01 2.80344486e-01
-5.76755762e-01 2.29652688e-01 5.37825823e-01 -6.67815134e-02
2.88195163e-01 1.27795011e-01 -4.82394606e-01 -6.88674003e-02
-1.60206223e+00 4.35883254e-01 4.50319797e-01 1.20345807e+00
6.17123425e-01 -7.01187178e-02 -5.92404902e-01 4.15710837e-01
1.84754387e-01 2.63731569e-01 1.19483136e-01 -1.05741560e+00
3.95283669e-01 5.16880691e-01 1.63721830e-01 -6.62890315e-01
-3.14897656e-01 -5.30577779e-01 -3.82769793e-01 2.31261015e-01
6.62811041e-01 -5.41269369e-02 -1.48292673e+00 1.50327218e+00
6.26203179e-01 5.08071721e-01 -5.05187884e-02 9.44527566e-01
8.12251925e-01 7.95789123e-01 3.62717479e-01 -3.59398186e-01
1.43996561e+00 -1.17606318e+00 -5.02905011e-01 -1.12995945e-01
3.39578778e-01 -6.61274970e-01 5.19323111e-01 4.81200993e-01
-1.26335454e+00 -8.78149688e-01 -8.31057668e-01 1.15312219e-01
-2.20742434e-01 3.59165296e-02 5.27395308e-01 5.50973356e-01
-9.67146337e-01 6.47672057e-01 -1.03261530e+00 -2.86278099e-01
8.77585948e-01 4.25768286e-01 -1.81596443e-01 1.28653422e-01
-5.97003698e-01 4.75764692e-01 9.32294369e-01 2.15644985e-02
-1.18303144e+00 -8.12545717e-01 -7.29266047e-01 -2.10611254e-01
8.09793174e-01 -5.64903259e-01 1.15086305e+00 -1.28909206e+00
-6.96535110e-01 1.03457820e+00 -2.00069904e-01 -1.17688131e+00
9.38297272e-01 -3.56880993e-01 -2.75999755e-01 4.54808503e-01
3.42158854e-01 1.27442348e+00 8.85318160e-01 -1.37634706e+00
-1.34556067e+00 -1.30316705e-01 -2.55973399e-01 1.11205593e-01
1.55658513e-01 2.23392427e-01 -1.16599500e+00 -7.70098209e-01
1.23839714e-01 -1.05999446e+00 -3.15625310e-01 -1.41592920e-01
-6.39404178e-01 -3.74940813e-01 1.35497046e+00 -6.58488452e-01
1.16778469e+00 -1.81749928e+00 1.58222303e-01 2.41913185e-01
2.36784041e-01 4.01275158e-01 5.26475385e-02 -1.90148041e-01
-2.18587089e-02 7.20071327e-03 -4.92989630e-01 -3.50613564e-01
-2.12316185e-01 -3.55954915e-02 -2.80662686e-01 4.40988332e-01
2.98232794e-01 9.23673689e-01 -1.01906013e+00 -7.81845927e-01
2.39341602e-01 1.38883904e-01 -5.10392606e-01 5.39392754e-02
-7.49544024e-01 3.52653027e-01 -2.78736055e-01 7.43712604e-01
6.93226039e-01 -1.85396522e-01 -1.82973400e-01 -1.81599155e-01
-2.44826511e-01 -7.78950527e-02 -1.41807377e+00 1.60488391e+00
4.94321227e-01 8.36777806e-01 -2.28542209e-01 -7.84688413e-01
5.53767323e-01 1.35263890e-01 9.94902432e-01 -3.16378087e-01
1.58354878e-01 -5.03743347e-03 -2.49241844e-01 -4.43568885e-01
7.02203393e-01 1.90630570e-01 6.68926537e-02 2.28094593e-01
2.65751660e-01 1.90326780e-01 9.02043045e-01 3.82559776e-01
9.99346077e-01 4.62932676e-01 -2.01786786e-01 -4.68989998e-01
4.71316785e-01 5.72356939e-01 5.11727035e-01 7.45202720e-01
-3.28471482e-01 8.89261127e-01 2.63988018e-01 -4.28214312e-01
-9.11634982e-01 -1.29878926e+00 -1.06130570e-01 9.65922058e-01
3.85218978e-01 -4.68968600e-01 -1.06288362e+00 -1.01011240e+00
-1.04347229e-01 4.76326227e-01 -7.40807414e-01 3.14612776e-01
-4.87108618e-01 -4.66738462e-01 2.90006816e-01 5.99719763e-01
4.16092157e-01 -1.23244750e+00 -7.59778917e-01 3.09592545e-01
-1.34889707e-01 -1.42604613e+00 -7.02461898e-01 4.47347790e-01
-6.62864685e-01 -1.26708758e+00 -7.25472152e-01 -1.05464995e+00
7.19016314e-01 2.86317110e-01 1.25129497e+00 3.92724155e-03
-3.68475288e-01 3.34804505e-01 -4.51880991e-01 -3.59960467e-01
-2.42919251e-01 -5.74326768e-05 -3.39750424e-02 3.05316031e-01
1.58736423e-01 -1.90716311e-02 -5.60308218e-01 5.53014636e-01
-8.54365885e-01 1.38893455e-01 3.52606893e-01 1.47744700e-01
1.26145482e+00 2.44310960e-01 1.34176478e-01 -1.00439382e+00
-3.34881902e-01 -4.19181019e-01 -8.03912997e-01 2.36194730e-01
-2.79953927e-01 -2.34298915e-01 2.62500018e-01 -3.13869715e-01
-6.36941791e-01 5.35788715e-01 2.96616182e-02 -6.98518991e-01
-3.18751842e-01 1.09429426e-01 -7.69434050e-02 5.43473735e-02
4.95121658e-01 2.00216725e-01 -3.33645523e-01 -3.27675402e-01
4.82786804e-01 3.26734543e-01 1.05652118e+00 -3.08141619e-01
9.35697913e-01 5.40091813e-01 -2.19671577e-01 -7.24348128e-01
-9.34718966e-01 -1.00850725e+00 -1.07866395e+00 -6.41444743e-01
1.51571429e+00 -9.74163115e-01 -1.43408120e-01 4.59044695e-01
-7.95173705e-01 -6.78476393e-01 -4.75656360e-01 2.46294856e-01
-7.51680255e-01 2.96708673e-01 -3.78465325e-01 -6.03271961e-01
-2.50924677e-01 -1.10411847e+00 9.90978181e-01 4.62089300e-01
-3.48993748e-01 -6.31312907e-01 -1.54435903e-01 5.44576466e-01
-2.22800329e-01 5.61873555e-01 2.13451013e-01 -6.11752450e-01
-1.05389047e+00 -7.18212575e-02 -1.19849332e-01 2.91776419e-01
-1.35212466e-01 4.09426004e-01 -7.61541367e-01 -2.62906432e-01
-5.19367754e-01 -3.87162678e-02 1.25533712e+00 6.33284152e-01
1.11899197e+00 -6.63199648e-02 -6.05003357e-01 6.20035052e-01
1.41392827e+00 2.80733168e-01 5.28931677e-01 3.57471079e-01
1.05440366e+00 5.37365556e-01 1.01177776e+00 9.23136771e-02
2.95488328e-01 8.56518984e-01 3.79687935e-01 -8.70654806e-02
-2.40823820e-01 -1.83793396e-01 4.19820845e-01 1.54531419e-01
-8.21697265e-02 -2.84709930e-01 -9.18105900e-01 8.77095997e-01
-2.01944089e+00 -1.12170029e+00 -7.75463939e-01 2.02667308e+00
6.37738228e-01 6.18242621e-01 9.16900635e-01 2.00117454e-01
8.74074161e-01 -3.24560553e-02 -3.80026639e-01 -3.62579105e-03
-1.78889543e-01 1.21307008e-01 8.02301645e-01 2.10082859e-01
-1.77135134e+00 1.30697179e+00 5.46836996e+00 8.88780415e-01
-9.14098918e-01 4.35397588e-02 7.91002333e-01 -4.46852781e-02
1.56481981e-01 -8.58823583e-03 -1.02636242e+00 4.83619779e-01
7.06581652e-01 1.79372013e-01 4.12699357e-02 7.75987089e-01
2.56385446e-01 -4.61583972e-01 -1.15154874e+00 7.48702049e-01
-1.79632053e-01 -1.55196571e+00 6.96296394e-02 -2.51000077e-01
1.16953599e+00 1.94398761e-01 -1.37646198e-01 1.45719266e-02
2.48086900e-01 -8.38200271e-01 1.48311937e+00 3.40781361e-01
2.82261372e-01 -7.84248412e-01 3.88025194e-01 2.45459586e-01
-1.44346130e+00 1.02343075e-01 -1.53396595e-02 4.12237704e-01
3.27779502e-01 3.80950660e-01 -6.13599539e-01 5.17584980e-01
9.95168149e-01 8.43855321e-01 -7.47365117e-01 1.75723600e+00
-3.64257284e-02 9.02874529e-01 -4.04153556e-01 3.68229449e-01
5.75001478e-01 -1.05115332e-01 6.39376044e-01 1.60596728e+00
-7.63346180e-02 -6.99941590e-02 6.48528397e-01 7.00516343e-01
-2.78533380e-02 -4.82179448e-02 -8.61078203e-02 1.83009401e-01
7.93248937e-02 1.31510711e+00 -1.69159091e+00 -6.75467670e-01
-1.04957223e-02 8.78469467e-01 -6.25697523e-02 2.29257345e-01
-1.14677954e+00 -9.86171514e-03 3.62954110e-01 3.37912381e-01
9.63900745e-01 -2.98060060e-01 -2.50915408e-01 -7.20351696e-01
3.47531997e-02 -6.71047926e-01 6.25269532e-01 -6.79509640e-01
-7.76824832e-01 6.74528003e-01 8.57798830e-02 -1.37760317e+00
1.16939163e-02 -3.22306633e-01 -5.09160101e-01 3.14641505e-01
-1.26970255e+00 -1.11474204e+00 -3.44759911e-01 5.81434250e-01
8.75221610e-01 1.63313568e-01 6.36815801e-02 2.63642907e-01
-5.42076707e-01 2.88824588e-01 -1.00353062e-01 3.95041883e-01
3.07645947e-01 -1.45239711e+00 5.58769047e-01 1.28654552e+00
7.56047964e-01 1.73908055e-01 8.40247095e-01 -9.25999999e-01
-1.01519394e+00 -1.64030695e+00 5.63262820e-01 -8.04457128e-01
5.85388303e-01 -3.53572249e-01 -1.02594137e+00 6.99225545e-01
1.61701486e-01 3.08809161e-01 1.97619632e-01 -3.94766301e-01
-5.27823158e-02 1.11111887e-01 -9.30703759e-01 4.07779187e-01
1.18349922e+00 -1.49236977e-01 -4.56753731e-01 5.13424277e-01
6.67645454e-01 -5.55895984e-01 -6.44600570e-01 5.16792178e-01
1.90214857e-01 -9.11909282e-01 1.01015794e+00 -6.59772575e-01
3.84682000e-01 -9.33798850e-01 1.52607739e-01 -6.10276043e-01
-1.98863462e-01 -8.28859091e-01 -1.44631192e-01 1.30483007e+00
4.56448495e-01 1.78759202e-01 1.16160679e+00 2.10124031e-01
-3.24376047e-01 -7.03729630e-01 -6.95459187e-01 -9.33963478e-01
-4.18065488e-01 -8.03022921e-01 -4.17201954e-04 7.01460302e-01
-5.55041552e-01 -2.87182555e-02 -1.23828299e-01 2.84085512e-01
8.86452377e-01 2.22900048e-01 8.61642003e-01 -1.09076726e+00
-1.10824011e-01 -6.09415889e-01 -4.95135367e-01 -1.06843197e+00
-3.04385424e-02 -9.83787000e-01 3.12882125e-01 -1.57603753e+00
3.93182933e-01 -4.33317751e-01 -1.90862238e-01 3.92827481e-01
-3.52987260e-01 6.42187297e-01 4.06739742e-01 2.91188329e-01
-1.38464153e+00 4.69258940e-03 9.13247943e-01 -3.63466769e-01
-4.31768745e-01 2.30236918e-01 -2.24957168e-01 8.07287216e-01
3.85267287e-01 -5.89418530e-01 -2.15422824e-01 -1.59074992e-01
-3.05887252e-01 -1.32249430e-01 5.11938453e-01 -1.39146066e+00
3.58555645e-01 5.89537015e-03 4.74063128e-01 -1.11908448e+00
-7.00617135e-02 -8.81502509e-01 3.85406256e-01 5.35956979e-01
-1.89171284e-01 8.27917177e-03 4.44750190e-01 5.97156048e-01
-1.17320776e-01 -1.72271639e-01 9.15665209e-01 9.39722434e-02
-1.43926477e+00 6.01347089e-01 -5.03063977e-01 2.27059469e-01
1.53083050e+00 -6.40806735e-01 9.74214152e-02 1.07757561e-01
-1.08057916e+00 6.78984821e-01 4.90007401e-01 6.35701716e-01
3.31833214e-01 -1.05946815e+00 -6.69643164e-01 -4.37012576e-02
1.45679742e-01 2.72074819e-01 2.96099812e-01 8.44683111e-01
-9.13504660e-01 2.98630476e-01 -1.57950521e-01 -1.20569515e+00
-1.68427336e+00 6.98499858e-01 2.30208710e-01 -1.38909459e-01
-8.28262687e-01 1.07406986e+00 1.33707812e-02 3.31265450e-01
4.26559776e-01 -5.23061454e-01 -3.18848372e-01 4.13935333e-01
4.45126295e-01 2.59406894e-01 -5.11635328e-03 -1.00321710e+00
-4.74665165e-01 8.61140847e-01 -3.73939909e-02 9.06991120e-03
1.14898884e+00 4.91584726e-02 1.97632000e-01 3.23271036e-01
1.03595769e+00 -1.02930427e-01 -1.77842176e+00 -1.84282854e-01
4.63785112e-01 -3.57714713e-01 -5.78585528e-02 -7.00514019e-01
-1.25576162e+00 3.61148030e-01 6.75367177e-01 3.80891442e-01
1.08223128e+00 2.56480634e-01 8.08245480e-01 -9.94406492e-02
3.17953646e-01 -1.15674174e+00 -3.81879471e-02 5.73035777e-01
4.96062517e-01 -1.18263233e+00 6.00651689e-02 -7.00229943e-01
-7.72405922e-01 8.72606337e-01 5.00895619e-01 -2.73920745e-01
3.54823738e-01 2.52884775e-01 2.88931001e-02 -3.17424744e-01
-5.13804436e-01 -7.06900120e-01 7.06719458e-01 5.30594170e-01
-2.63663288e-02 -5.41715324e-02 -5.41505553e-02 2.57272243e-01
-1.41344935e-01 -1.25697449e-01 2.27246836e-01 8.81782353e-01
-5.98198831e-01 -6.98348999e-01 -3.41588467e-01 5.14160514e-01
-5.81542850e-01 2.44026631e-01 -4.12033349e-01 7.59794950e-01
5.52177668e-01 7.49612093e-01 2.50965238e-01 -3.76758367e-01
2.91838974e-01 -5.89619800e-02 3.51350009e-01 -5.69154322e-01
-7.87402034e-01 2.99203992e-01 5.85914925e-02 -5.99360466e-01
-7.29374409e-01 -1.18352628e+00 -1.57239914e+00 2.19624802e-01
-2.94222623e-01 1.81740314e-01 4.90923077e-01 1.04872012e+00
-5.17981723e-02 6.89187527e-01 2.94973344e-01 -1.10907149e+00
1.94012865e-01 -4.72579509e-01 -2.10693806e-01 5.77006996e-01
3.22317928e-01 -4.95752901e-01 1.16836421e-01 6.99641287e-01] | [9.158447265625, -0.23414117097854614] |
dfed0638-585b-40a7-ba44-abc93bdc86d9 | kgtn-ens-few-shot-image-classification-with | 2211.03199 | null | https://arxiv.org/abs/2211.03199v1 | https://arxiv.org/pdf/2211.03199v1.pdf | KGTN-ens: Few-Shot Image Classification with Knowledge Graph Ensembles | We propose KGTN-ens, a framework extending the recent Knowledge Graph Transfer Network (KGTN) in order to incorporate multiple knowledge graph embeddings at a small cost. We evaluate it with different combinations of embeddings in a few-shot image classification task. We also construct a new knowledge source - Wikidata embeddings - and evaluate it with KGTN and KGTN-ens. Our approach outperforms KGTN in terms of the top-5 accuracy on the ImageNet-FS dataset for the majority of tested settings. | ['Agata Filipowska', 'Anna Fensel', 'Dominik Filipiak'] | 2022-11-06 | null | null | null | null | ['few-shot-image-classification', 'knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['computer-vision', 'graphs', 'methodology'] | [-1.25842780e-01 2.43622988e-01 -4.95605946e-01 -6.75999969e-02
-2.43706152e-01 -2.20391497e-01 5.84074438e-01 8.91538113e-02
-7.36650467e-01 6.08593822e-01 3.50808918e-01 -7.09018260e-02
-3.67515624e-01 -1.07266438e+00 -7.52671540e-01 -3.02648813e-01
-4.48173061e-02 3.34723175e-01 5.78012586e-01 -8.16913173e-02
-3.52775715e-02 -8.82068742e-03 -1.13949895e+00 -6.28654435e-02
5.50130308e-01 9.99370039e-01 4.21848670e-02 5.36459506e-01
-1.28634319e-01 1.14445543e+00 -8.74138996e-02 -8.81827652e-01
1.81146070e-01 -1.20543949e-01 -9.97810662e-01 -1.87266812e-01
6.41007960e-01 -2.36000285e-01 -7.93965697e-01 1.03265691e+00
4.29562598e-01 2.33572885e-01 6.89429522e-01 -1.46813285e+00
-1.48331070e+00 6.80254042e-01 -1.74426720e-01 1.68147892e-01
1.36540070e-01 9.99698266e-02 1.11410856e+00 -7.36566424e-01
1.10915232e+00 1.13843155e+00 9.85273600e-01 3.26998085e-01
-1.25076354e+00 -4.03411835e-01 -7.33860806e-02 8.13570619e-01
-1.67968416e+00 7.11714104e-02 3.95113915e-01 -4.12982196e-01
1.40669000e+00 -6.39430404e-01 6.99719191e-01 1.09186697e+00
-5.23180934e-04 6.14947379e-01 7.69258380e-01 -5.42113245e-01
1.82270050e-01 6.85452372e-02 3.12712848e-01 1.07165384e+00
3.61667842e-01 -7.18141720e-02 -7.12378800e-01 4.73947264e-02
6.40961349e-01 -1.07778303e-01 -2.77530462e-01 -6.99180722e-01
-1.25757360e+00 1.04873204e+00 8.34979892e-01 3.02919000e-01
-1.61882252e-01 7.55323648e-01 6.12390757e-01 5.64602613e-01
7.22712517e-01 4.24763352e-01 -5.06119728e-01 -9.92945954e-02
-3.30466330e-01 -1.14214651e-01 1.16360307e+00 1.03786767e+00
9.93732333e-01 5.42505980e-02 -3.42234552e-01 7.39389837e-01
-1.09044261e-01 2.77311325e-01 4.06087935e-01 -6.19733870e-01
1.18193679e-01 6.64811492e-01 -1.57745466e-01 -8.33773971e-01
-1.75341830e-01 -2.99479216e-01 -4.07989800e-01 -8.18871334e-02
-1.88649431e-01 9.17666480e-02 -1.21598577e+00 1.53681004e+00
1.21555656e-01 8.68620038e-01 3.49960119e-01 6.19135201e-01
1.21917927e+00 4.73843634e-01 3.89354855e-01 2.14720756e-01
1.11137152e+00 -1.50980616e+00 -6.92539811e-01 8.89773667e-02
7.74028122e-01 -6.62104264e-02 1.08744073e+00 1.21630274e-01
-4.03660834e-01 -3.52717102e-01 -1.05448914e+00 -1.37293264e-01
-1.10471714e+00 -6.26842678e-02 8.88616562e-01 4.68814552e-01
-1.55505431e+00 7.62362540e-01 -5.55393338e-01 -7.43283153e-01
8.21874857e-01 1.49685010e-01 -7.74416864e-01 -6.28116965e-01
-1.39502013e+00 1.30636275e+00 9.51594889e-01 -3.47297460e-01
-9.91298318e-01 -1.07020330e+00 -1.20546556e+00 1.12625666e-01
6.72651887e-01 -9.75844800e-01 7.40927398e-01 -5.88120997e-01
-1.45245564e+00 8.42248082e-01 4.08130646e-01 -6.02719665e-01
2.25450411e-01 -2.43535742e-01 -5.43073177e-01 3.91851127e-01
-1.10090077e-02 1.04067004e+00 7.34722137e-01 -8.55768085e-01
-3.79511751e-02 -1.15702767e-02 2.64676720e-01 1.71882734e-01
-8.50887001e-01 -3.08235973e-01 -5.51558793e-01 -4.67044324e-01
-7.48078346e-01 -7.97046363e-01 -1.91839904e-01 2.01698884e-01
-1.41817197e-01 -4.89571422e-01 9.23822999e-01 -4.86412466e-01
9.04582977e-01 -2.13692498e+00 3.77394497e-01 -1.39769495e-01
4.11126792e-01 6.96857870e-01 -6.52893066e-01 5.02948344e-01
-2.70248447e-02 1.35821000e-01 -1.90908179e-01 -3.39544237e-01
1.04601108e-01 6.55237436e-01 1.14347167e-01 1.24277860e-01
3.28171521e-01 1.69809198e+00 -1.34659982e+00 -5.16118526e-01
3.84074360e-01 6.86036468e-01 -4.21165615e-01 -3.76751795e-02
-3.26821655e-01 -4.66978192e-01 -2.46912748e-01 4.64851737e-01
4.25308019e-01 -5.33493996e-01 1.02548553e-02 -3.93402457e-01
3.27421606e-01 -3.96543086e-01 -7.37401068e-01 2.02162457e+00
-4.21665043e-01 5.78648210e-01 -6.18904352e-01 -9.27707434e-01
5.89955866e-01 2.04267398e-01 2.83189237e-01 -3.08106869e-01
1.28547609e-01 -1.29688725e-01 -2.85684705e-01 -3.82529616e-01
4.25123155e-01 -1.39786094e-01 -1.54068768e-01 2.23345265e-01
1.18210173e+00 1.07276529e-01 8.32366198e-02 7.89054275e-01
1.58841324e+00 2.46490836e-02 5.51331401e-01 -2.92810261e-01
8.08628462e-03 -6.91304952e-02 7.67138675e-02 7.66943634e-01
-4.02570188e-01 3.25835973e-01 5.33752382e-01 -4.00716037e-01
-5.43600202e-01 -1.41657686e+00 1.75744906e-01 1.05508602e+00
1.37585372e-01 -7.77771115e-01 -4.85555500e-01 -1.23138785e+00
3.83454859e-01 7.82072365e-01 -1.07403851e+00 -4.82243359e-01
2.34740317e-01 -2.88534671e-01 6.13651395e-01 6.46612287e-01
5.88454723e-01 -1.25161731e+00 -1.99481413e-01 8.45429376e-02
1.62379116e-01 -1.57475305e+00 -4.02959526e-01 1.41549796e-01
-5.53209245e-01 -1.32700264e+00 -5.12793362e-01 -8.68871987e-01
4.52325553e-01 1.92328274e-01 1.12789357e+00 -1.21627778e-01
-5.92486024e-01 1.02547443e+00 -8.58693659e-01 -2.84916073e-01
-7.01037515e-03 1.93853483e-01 -5.84860966e-02 -7.82593414e-02
5.64359248e-01 -7.00456321e-01 -3.08993548e-01 -1.62044182e-01
-9.74891365e-01 -1.87570974e-01 4.51881915e-01 8.07956278e-01
4.09236252e-01 1.16985030e-01 9.19589758e-01 -9.49708879e-01
6.15575671e-01 -6.39811873e-01 -3.92887741e-01 4.55030501e-01
-8.09045374e-01 6.97136372e-02 4.57746238e-01 -5.05847037e-01
-6.64704919e-01 -2.04497427e-01 1.79546610e-01 -1.14018559e+00
1.19144782e-01 8.75668824e-01 1.22051984e-01 -6.56996667e-01
6.41191602e-01 1.73139483e-01 -1.60039455e-01 -2.98578560e-01
1.29669785e+00 2.11897627e-01 4.78700608e-01 -2.04385892e-01
7.92073190e-01 3.38747501e-01 -2.73049146e-01 -8.35511506e-01
-1.09711075e+00 -6.78513765e-01 -6.67264521e-01 -1.36805683e-01
1.07396364e+00 -1.05945504e+00 -4.16554809e-01 4.60723579e-01
-9.91484344e-01 -4.47771609e-01 -5.70869446e-01 4.65416014e-01
-5.05104423e-01 3.61740559e-01 -6.29046381e-01 -4.92034703e-02
-3.62847030e-01 -6.77231371e-01 6.90971434e-01 9.47504044e-02
2.22146973e-01 -1.52466071e+00 3.06135118e-01 5.07290699e-02
6.14314497e-01 3.16988200e-01 8.90172005e-01 -7.05123007e-01
-4.80076194e-01 -8.88817236e-02 -7.27737427e-01 6.66752875e-01
9.89337340e-02 -3.28357726e-01 -8.56338441e-01 -1.60813034e-01
-7.52370298e-01 -8.07531953e-01 1.50637233e+00 8.10581371e-02
1.01067531e+00 -1.31850347e-01 -3.74443501e-01 5.92593610e-01
2.10600495e+00 -3.72353196e-01 7.92428553e-01 2.03876749e-01
1.03852224e+00 -9.27293207e-03 4.73065637e-02 3.99822667e-02
6.64794922e-01 4.99957234e-01 4.35052216e-01 1.46359026e-01
-6.55839503e-01 -4.48098630e-01 4.53424156e-01 1.16898024e+00
-2.69024745e-02 -3.01139027e-01 -8.15533698e-01 1.20153522e+00
-2.05657673e+00 -7.46600032e-01 3.63080710e-01 1.55738974e+00
8.31417978e-01 -1.45095795e-01 -1.57116637e-01 -2.87628204e-01
5.72184443e-01 3.12269062e-01 -5.58072031e-01 -3.72720093e-01
2.24119928e-02 7.58594394e-01 6.85124636e-01 4.66056973e-01
-1.17695832e+00 1.60939729e+00 6.76594114e+00 9.08121705e-01
-7.17642903e-01 6.65705264e-01 -2.71010458e-01 1.78804606e-01
-1.01778515e-01 1.05508678e-01 -5.46855509e-01 2.16305628e-01
1.09558666e+00 -5.11946738e-01 4.28395301e-01 1.07971823e+00
-8.52068484e-01 -1.42408714e-01 -9.72768843e-01 1.07897186e+00
5.96219480e-01 -1.65213013e+00 4.48238432e-01 -2.17428222e-01
8.69423568e-01 5.81913471e-01 -1.92830637e-01 7.70015001e-01
7.59691000e-01 -1.05159557e+00 2.03634366e-01 4.76675153e-01
8.59877527e-01 -5.63068867e-01 8.15214455e-01 -1.44553840e-01
-1.32804811e+00 1.59880087e-01 -5.51657796e-01 3.84205021e-02
1.88842446e-01 6.82073534e-01 -9.50246096e-01 1.01897264e+00
6.05031848e-01 1.38122046e+00 -7.45957017e-01 1.07488906e+00
-6.68666720e-01 8.06209803e-01 -3.42562571e-02 1.68948382e-01
3.90861303e-01 1.61457166e-01 2.77761489e-01 1.33915257e+00
1.10775284e-01 -2.09521316e-02 4.01249558e-01 8.78910363e-01
-6.72738135e-01 -7.88674429e-02 -8.89160514e-01 -3.37310493e-01
3.02728117e-01 1.33397925e+00 -6.49303734e-01 -5.08877158e-01
-4.43063617e-01 1.40995383e+00 9.13389087e-01 3.75290126e-01
-7.83272386e-01 -7.53075242e-01 5.17522275e-01 -3.91906530e-01
8.74921679e-01 -1.35591909e-01 6.37000620e-01 -1.20575988e+00
-3.35377246e-01 2.50961110e-02 6.19878650e-01 -1.00594032e+00
-1.82130551e+00 5.74813604e-01 2.17994481e-01 -7.31303751e-01
2.06577227e-01 -8.48059297e-01 -3.58675212e-01 2.91852355e-01
-2.09077454e+00 -1.56803906e+00 -4.61663991e-01 8.45327258e-01
3.88315506e-02 -1.48415938e-01 1.14976549e+00 2.68219292e-01
-3.25270504e-01 6.12514913e-01 -2.95890402e-02 4.38802421e-01
6.97792053e-01 -1.34733820e+00 4.35713172e-01 4.31632847e-01
3.20159912e-01 7.60197192e-02 1.03457011e-01 -6.26823425e-01
-1.25629497e+00 -1.66328728e+00 7.19987571e-01 -7.38610268e-01
1.14954662e+00 -3.29525441e-01 -8.53607893e-01 1.07342851e+00
5.78758955e-01 7.16970384e-01 7.30901718e-01 2.83354193e-01
-1.00179839e+00 1.83813944e-01 -1.08290780e+00 2.43615642e-01
1.47814238e+00 -7.43883491e-01 -8.30550432e-01 3.10583115e-01
1.41464317e+00 1.18625090e-01 -1.45934999e+00 2.91360945e-01
2.68772185e-01 -3.91500801e-01 9.94020760e-01 -7.23876297e-01
4.02774513e-01 -1.88121423e-01 -1.69754207e-01 -1.93785942e+00
-5.02952337e-01 -2.43240803e-01 -3.65524232e-01 9.81579244e-01
2.30973646e-01 -8.95144343e-01 5.38739145e-01 -1.09773546e-01
-1.14336133e-01 -5.43480456e-01 -1.05802226e+00 -1.38371861e+00
-7.93593973e-02 -3.46867621e-01 1.81430355e-01 1.32448280e+00
1.54085129e-01 5.08829594e-01 -4.31241691e-01 -2.06590727e-01
8.00185263e-01 -3.25081855e-01 5.08475244e-01 -1.06978679e+00
-7.15121478e-02 -1.28359899e-01 -1.30214334e+00 -1.55481100e-01
6.08369529e-01 -1.48771799e+00 -2.69829392e-01 -2.01122069e+00
4.76588726e-01 -1.54217944e-01 -8.93954337e-01 1.08138871e+00
-1.69423923e-01 4.45015222e-01 3.22720081e-01 -2.71327168e-01
-1.27306557e+00 8.84710073e-01 9.07399833e-01 -2.84990609e-01
2.32095137e-01 -1.00687003e+00 -5.35218298e-01 4.52904373e-01
6.11127734e-01 -5.55090070e-01 -7.35029876e-01 -4.77493316e-01
2.19157562e-01 -4.83171642e-01 5.61122298e-01 -1.16685438e+00
2.42224470e-01 1.66904271e-01 -2.02230252e-02 -1.25075251e-01
2.10431725e-01 -6.18359864e-01 1.19256586e-01 2.48241976e-01
-1.38086841e-01 -3.30849916e-01 3.52305263e-01 1.08507109e+00
-2.57070243e-01 8.87591764e-02 5.65127015e-01 2.08125003e-02
-1.43662310e+00 6.34602129e-01 2.68687028e-02 3.02000970e-01
1.28668535e+00 -1.56576425e-01 -8.38382125e-01 -1.60632238e-01
-8.04204583e-01 2.64619589e-01 3.21481049e-01 7.45744407e-01
8.19271207e-01 -1.71932995e+00 -4.87587631e-01 -1.35406107e-01
8.84912789e-01 -6.11838579e-01 2.25968271e-01 8.24992239e-01
-3.99498820e-01 3.15960586e-01 -3.65747243e-01 -3.36319566e-01
-8.17358613e-01 6.89675987e-01 2.12037891e-01 -4.71969008e-01
-7.76601911e-01 9.94250953e-01 -5.07456996e-02 -6.20682657e-01
7.90727139e-02 -2.62412250e-01 -2.18612343e-01 6.49785548e-02
4.57028538e-01 2.87213206e-01 3.54748145e-02 -2.99801052e-01
-4.16502893e-01 6.07893109e-01 -1.18230410e-01 3.04002643e-01
1.48759401e+00 2.83832699e-01 -1.97911888e-01 3.67245674e-01
1.55567729e+00 -6.55926585e-01 -8.41992915e-01 -6.03398323e-01
2.44066678e-02 -4.08995122e-01 3.27327222e-01 -1.06587780e+00
-1.18816197e+00 5.35256863e-01 5.84065676e-01 -3.33156079e-01
7.92797208e-01 3.26998770e-01 7.40004480e-01 6.51952744e-01
7.16144025e-01 -9.00444567e-01 4.59116340e-01 6.92629397e-01
6.37184024e-01 -1.17509222e+00 1.16011299e-01 -2.45834664e-01
-6.68001592e-01 8.70982230e-01 6.37687683e-01 -4.05314714e-01
1.24961853e+00 -9.97759476e-02 -1.60338461e-01 -5.99295616e-01
-8.92227709e-01 -7.53524542e-01 2.90709168e-01 1.02058303e+00
-1.27197757e-01 2.00726613e-01 -7.08663538e-02 4.47861880e-01
2.27406949e-01 6.67081714e-01 4.90387410e-01 8.20321441e-01
-2.92151630e-01 -1.00100315e+00 5.60753942e-01 5.98390102e-01
-1.66881569e-02 -4.23307866e-01 -4.27972734e-01 1.02513540e+00
9.36648995e-02 6.06759846e-01 -2.85105586e-01 -7.01347053e-01
3.84386033e-01 1.73074394e-01 7.87928045e-01 -9.45564687e-01
-4.57252055e-01 -8.23216379e-01 1.98474735e-01 -7.45193243e-01
-7.30949342e-01 6.41547516e-02 -8.56331825e-01 -2.28681430e-01
-5.73166728e-01 3.36704925e-02 3.76205802e-01 7.69124091e-01
6.09397709e-01 6.66884720e-01 -6.84990138e-02 -6.27803564e-01
-2.29917631e-01 -8.86472821e-01 -8.53955686e-01 4.52845633e-01
-1.25772581e-01 -9.78398383e-01 -2.66772538e-01 1.50585435e-02] | [8.733317375183105, 7.798191070556641] |
e0b6cb47-0608-44b7-b9e3-d5a469db0956 | cam-cad-point-cloud-part-segmentation-via-few | 2207.01218 | null | https://arxiv.org/abs/2207.01218v2 | https://arxiv.org/pdf/2207.01218v2.pdf | CAM/CAD Point Cloud Part Segmentation via Few-Shot Learning | 3D part segmentation is an essential step in advanced CAM/CAD workflow. Precise 3D segmentation contributes to lower defective rate of work-pieces produced by the manufacturing equipment (such as computer controlled CNCs), thereby improving work efficiency and attaining the attendant economic benefits. A large class of existing works on 3D model segmentation are mostly based on fully-supervised learning, which trains the AI models with large, annotated datasets. However, the disadvantage is that the resulting models from the fully-supervised learning methodology are highly reliant on the completeness of the available dataset, and its generalization ability is relatively poor to new unknown segmentation types (i.e. further additional novel classes). In this work, we propose and develop a noteworthy few-shot learning-based approach for effective part segmentation in CAM/CAD; and this is designed to significantly enhance its generalization ability and flexibly adapt to new segmentation tasks by using only relatively rather few samples. As a result, it not only reduces the requirements for the usually unattainable and exhaustive completeness of supervision datasets, but also improves the flexibility for real-world applications. As further improvement and innovation, we additionally adopt the transform net and the center loss block in the network. These characteristics serve to improve the comprehension for 3D features of the various possible instances of the whole work-piece and ensure the close distribution of the same class in feature space. | ['Tong Heng Lee', 'Vadakkepat Prahlad', 'Abdullah Al Mamun', 'Haoren Guo', 'Haiyue Zhu', 'Jiahui Wang'] | 2022-07-04 | null | null | null | null | ['3d-part-segmentation'] | ['computer-vision'] | [ 2.64496505e-01 6.28577396e-02 -3.24497849e-01 -1.84670538e-01
-1.82214588e-01 -2.09521919e-01 1.77253619e-01 3.87999900e-02
-6.90511540e-02 4.85728383e-01 -4.75718290e-01 4.07517441e-02
-5.65694749e-01 -8.96229208e-01 -4.36980635e-01 -7.27097690e-01
2.37327382e-01 7.56368935e-01 4.31150764e-01 -3.14509600e-01
2.07198903e-01 8.91189933e-01 -1.78376603e+00 -1.10890366e-01
1.07526195e+00 1.14854014e+00 5.40734828e-01 1.48163941e-02
-3.90043229e-01 4.33749408e-01 -5.01656473e-01 -2.74032950e-01
5.34108937e-01 -2.74280846e-01 -6.26118422e-01 6.21477604e-01
-1.61301985e-01 2.11676601e-02 -8.44234154e-02 8.88510823e-01
2.72780389e-01 3.51407200e-01 6.18006885e-01 -1.20850885e+00
-3.85945588e-01 4.12658483e-01 -5.29885232e-01 -2.63558209e-01
4.86465394e-02 1.39277309e-01 6.36760712e-01 -5.40252507e-01
5.56401312e-01 9.82590675e-01 5.64295709e-01 5.73037326e-01
-6.95027471e-01 -4.70255524e-01 2.13694736e-01 1.61959410e-01
-1.18065810e+00 -1.70171052e-01 1.20385790e+00 -3.90204638e-01
5.95841944e-01 1.36838198e-01 9.87020850e-01 5.68645895e-01
1.15812749e-01 9.53987479e-01 7.30612814e-01 -3.84594977e-01
4.30046707e-01 2.68792331e-01 1.52878225e-01 5.20486295e-01
2.31939167e-01 -1.22305527e-01 2.55255811e-02 3.52226377e-01
9.47919726e-01 5.15872717e-01 -1.96953624e-01 -8.77132773e-01
-9.43167806e-01 5.65779924e-01 4.25982386e-01 4.81940508e-01
-2.58759558e-01 -4.32796776e-01 4.50392365e-01 1.52223408e-01
4.25622284e-01 6.62809253e-01 -6.95541084e-01 -1.64302796e-01
-8.53762448e-01 6.79085404e-02 7.27514625e-01 1.28456151e+00
9.49206173e-01 4.92410623e-02 1.29384845e-01 9.59849238e-01
5.41480444e-03 1.79099470e-01 3.83934021e-01 -6.74522102e-01
4.35377508e-01 1.22910941e+00 -1.44991443e-01 -8.00414801e-01
-4.87069190e-01 -5.08819163e-01 -9.18986022e-01 1.89086840e-01
2.03543887e-01 1.61782596e-02 -1.24344683e+00 1.14009106e+00
5.61550736e-01 -1.92292750e-01 -2.70686537e-01 6.76000416e-01
3.57798696e-01 4.87620860e-01 -1.14531986e-01 -4.30872530e-01
9.77623165e-01 -9.41649973e-01 -8.09712470e-01 -1.73581004e-01
5.86758554e-01 -7.35569179e-01 1.02778530e+00 4.79261905e-01
-8.18198562e-01 -9.48709488e-01 -1.16164708e+00 1.40638500e-01
-5.53606093e-01 1.78868696e-01 1.00373983e+00 5.11360347e-01
-4.00874645e-01 8.07728231e-01 -8.12885463e-01 -1.61192551e-01
6.99264050e-01 5.88689148e-01 -2.14460149e-01 -3.78017843e-01
-9.11446333e-01 9.33687270e-01 6.68956935e-01 5.67859769e-01
-5.35368443e-01 -6.38361514e-01 -8.50415170e-01 -3.79462610e-03
8.56182992e-01 -3.21798295e-01 9.56371546e-01 -8.94560039e-01
-1.47797084e+00 5.19226849e-01 3.17426205e-01 -1.54290572e-01
5.60404778e-01 -1.41040832e-01 -3.00020754e-01 2.58395690e-02
-1.06906749e-01 5.49337268e-01 8.64086926e-01 -1.44885302e+00
-6.30592942e-01 -5.34017742e-01 6.30798042e-02 1.20028898e-01
-2.98632324e-01 -5.20911753e-01 -4.41656113e-01 -5.05063474e-01
4.18596148e-01 -7.64541924e-01 -5.18482506e-01 -1.63795259e-02
-2.85198450e-01 -3.36792052e-01 1.12279022e+00 -4.06603664e-01
9.99549508e-01 -2.17390919e+00 9.50642005e-02 2.34910816e-01
-7.53146932e-02 4.42963004e-01 1.37349695e-01 3.91031384e-01
3.15269046e-02 -1.61347881e-01 -5.04531503e-01 -3.13174576e-02
-1.35340154e-01 3.89698118e-01 2.74267793e-01 2.59795904e-01
4.03286129e-01 7.57083118e-01 -6.52793050e-01 -5.97424567e-01
6.21156991e-01 5.84937166e-03 -3.21246654e-01 -4.74280156e-02
-2.89150685e-01 4.40640271e-01 -7.04542458e-01 9.52812552e-01
7.12174773e-01 1.07126553e-02 -1.95783034e-01 -1.68838099e-01
-7.38452841e-03 -4.10266846e-01 -1.39487135e+00 1.97108448e+00
-5.87235987e-01 3.85391293e-03 2.14746632e-02 -1.45101213e+00
1.45167482e+00 1.54320002e-01 8.23238373e-01 -4.16842043e-01
3.40950578e-01 2.15133667e-01 2.70877387e-02 -6.68946922e-01
3.82581949e-01 -1.70790002e-01 -1.93955824e-01 6.17912635e-02
1.61041453e-01 -6.08566165e-01 3.49381864e-01 -2.28706822e-01
5.46846211e-01 2.69023657e-01 2.45229065e-01 -2.13023111e-01
6.96565092e-01 1.42829821e-01 8.20739627e-01 4.23483402e-01
-1.72779873e-01 4.26146418e-01 3.45129043e-01 -4.21213448e-01
-1.00593936e+00 -7.20224142e-01 -2.23155707e-01 3.22055072e-01
5.54779947e-01 2.71347966e-02 -6.44143522e-01 -8.04167807e-01
4.06064726e-02 6.88557923e-01 -3.16485912e-01 -4.54438865e-01
-5.28764129e-01 -3.94286215e-01 -9.04010758e-02 6.26684725e-01
5.67966342e-01 -1.07777727e+00 -5.94886243e-01 2.67054051e-01
3.55628043e-01 -8.80907536e-01 -6.97762072e-02 3.81308436e-01
-1.37300026e+00 -1.28459370e+00 -7.75732756e-01 -9.68132079e-01
8.72305572e-01 2.43540183e-01 6.75496995e-01 1.90851286e-01
-5.32087624e-01 1.29452646e-01 -4.95299309e-01 -4.09915835e-01
-3.20079684e-01 2.09449843e-01 -8.92415792e-02 -7.09690750e-02
3.52922857e-01 -4.03147966e-01 -2.72623807e-01 4.54307348e-01
-7.96612918e-01 2.91815624e-02 8.18271697e-01 9.84727979e-01
8.18686068e-01 7.27941394e-01 7.88508892e-01 -1.01746213e+00
2.73609906e-01 -1.61072686e-01 -6.89998209e-01 3.05215925e-01
-7.70433009e-01 -2.07326040e-01 7.30836451e-01 -4.47297692e-01
-1.31386638e+00 3.73205125e-01 -1.17708966e-01 -6.10960722e-01
-3.51901323e-01 2.92134374e-01 -5.94413340e-01 1.44721987e-02
1.63692832e-01 1.27385020e-01 1.85057729e-01 -6.80816948e-01
2.65064865e-01 6.52260542e-01 2.15996474e-01 -3.92684788e-01
8.58802080e-01 1.45354047e-01 3.42020780e-01 -8.30044031e-01
-7.20678866e-01 -7.13992298e-01 -1.12914360e+00 -2.89673716e-01
7.04486251e-01 -4.83893424e-01 -5.26344776e-01 5.56387603e-01
-7.37737298e-01 -1.69028286e-02 -7.09473729e-01 5.74598730e-01
-5.65436125e-01 4.64554310e-01 -4.25143123e-01 -9.52040195e-01
-1.03457391e-01 -1.26728654e+00 9.49020147e-01 4.23380196e-01
4.77169193e-02 -7.56213903e-01 -4.55240935e-01 5.79806387e-01
9.21380520e-02 2.75185287e-01 1.20879042e+00 -6.46188498e-01
-7.15901494e-01 -7.30679035e-01 3.33425105e-02 7.11953461e-01
4.96453196e-01 6.91983849e-02 -7.05345213e-01 -1.01763770e-01
4.23238397e-01 -2.15757504e-01 6.52382731e-01 3.89786690e-01
1.40235734e+00 1.88997000e-01 -4.55967188e-01 1.54949814e-01
1.28235257e+00 6.40623569e-01 4.84348983e-01 6.38082027e-02
6.91809654e-01 8.88223648e-01 1.34579194e+00 3.63340765e-01
-2.98207164e-01 3.76434505e-01 3.97174984e-01 -2.16051951e-01
4.07201946e-02 -2.01285541e-01 -1.89389884e-01 1.01963294e+00
-2.25343660e-01 -1.61194354e-01 -7.27976024e-01 5.13957202e-01
-1.83798063e+00 -7.28703439e-01 -1.10676832e-01 2.03448844e+00
5.92438161e-01 4.23760056e-01 -1.60331637e-01 5.86409092e-01
8.52146924e-01 -1.54465437e-01 -7.62267649e-01 -1.50303647e-01
1.66094363e-01 3.79272401e-01 3.48464578e-01 -9.15175956e-03
-1.07713795e+00 8.14431489e-01 5.13946676e+00 1.23366439e+00
-7.32162535e-01 -2.14822680e-01 4.23215568e-01 2.07823217e-02
5.21126427e-02 -5.09938113e-02 -7.57745087e-01 3.17138195e-01
3.56534541e-01 9.68858376e-02 2.85801232e-01 1.22117937e+00
1.91803366e-01 -2.40456119e-01 -1.12912309e+00 9.26771224e-01
6.76384121e-02 -1.15632510e+00 -5.78430407e-02 5.33871679e-03
8.20512831e-01 -5.46237588e-01 -2.89708287e-01 3.36562634e-01
-3.65759999e-01 -6.90676630e-01 5.31916976e-01 5.74847519e-01
5.85506260e-01 -1.11655140e+00 1.21248877e+00 6.58490241e-01
-1.02542877e+00 -3.61216396e-01 -6.20034158e-01 1.13130240e-02
2.95090318e-01 7.45267212e-01 -7.08436251e-01 8.42653811e-01
5.05117893e-01 8.60608399e-01 -4.48813081e-01 1.17867196e+00
-1.11504532e-01 1.69490591e-01 -2.49673158e-01 -1.26375586e-01
1.27870217e-01 -4.32364911e-01 3.13519388e-01 5.13606846e-01
2.97401577e-01 -1.92532837e-02 4.52542394e-01 8.32632005e-01
7.48380795e-02 1.33085385e-01 -6.48140371e-01 -1.14135094e-01
3.70525062e-01 1.19337165e+00 -1.11900580e+00 -1.52069867e-01
-3.71021032e-01 7.27890134e-01 -9.95631218e-02 9.44250450e-02
-5.52599728e-01 -7.09607065e-01 3.40566844e-01 -6.10841997e-03
4.15172398e-01 -1.62179500e-01 -5.36635697e-01 -7.90480375e-01
1.18913017e-01 -6.35448575e-01 1.72689989e-01 -3.93270403e-01
-1.42426777e+00 2.35754624e-01 4.95955311e-02 -1.54680002e+00
-3.12590152e-02 -7.21254289e-01 -3.92772645e-01 5.26234567e-01
-1.24292409e+00 -1.21664107e+00 -2.91081756e-01 4.47453082e-01
1.02582622e+00 -2.20162228e-01 5.78417718e-01 4.18372065e-01
-7.08429992e-01 2.80291885e-01 -3.43923783e-03 -9.14373472e-02
3.98831129e-01 -9.10063207e-01 9.21246037e-02 6.14843905e-01
-8.01083744e-02 4.21978563e-01 3.68742049e-01 -7.10691571e-01
-1.47575951e+00 -1.00739801e+00 3.74543637e-01 -2.25091264e-01
3.19607288e-01 -4.37630564e-01 -8.95833254e-01 2.77622461e-01
-3.52963209e-01 -2.10931987e-01 4.24718410e-01 1.39908195e-01
4.66018617e-01 -3.08380276e-01 -1.31689763e+00 3.98096114e-01
1.09994686e+00 -1.83920160e-01 -7.81504869e-01 3.64062041e-01
8.01404238e-01 -3.22240233e-01 -1.11800241e+00 6.70253575e-01
3.05506110e-01 -8.07009161e-01 8.20782542e-01 -4.94055957e-01
4.58106726e-01 -3.07192862e-01 2.06638217e-01 -1.00239325e+00
-2.37569094e-01 -3.28127116e-01 -1.67222351e-01 1.46623051e+00
2.71028459e-01 -4.93984252e-01 1.05136001e+00 6.53675079e-01
-6.20485902e-01 -1.11325121e+00 -8.79436553e-01 -1.00311971e+00
-1.58447281e-01 -3.58973145e-01 7.81147301e-01 8.88777554e-01
-1.41712368e-01 1.01879537e-01 -2.08424360e-01 -1.67641431e-01
4.24253225e-01 3.11656296e-01 7.09853828e-01 -1.56860888e+00
-1.18398555e-01 -2.64990449e-01 -6.21729612e-01 -1.00433671e+00
-3.72706987e-02 -7.17171848e-01 2.66702414e-01 -1.46490395e+00
-6.29971102e-02 -8.66377354e-01 -2.95980960e-01 4.62128341e-01
9.14835483e-02 -1.18321031e-01 7.64648691e-02 2.48138189e-01
-2.88458169e-01 7.72461176e-01 1.95376050e+00 -2.25584477e-01
-5.81507266e-01 6.72899604e-01 -4.73654360e-01 8.05896342e-01
7.45238125e-01 -3.65218103e-01 -6.93416297e-01 -9.36820880e-02
-2.00935140e-01 -2.07557827e-02 5.58816455e-02 -1.01408112e+00
1.25035718e-01 -2.70980388e-01 4.94478166e-01 -9.16505754e-01
3.70422840e-01 -1.26663315e+00 7.45161846e-02 4.55030739e-01
-2.46687029e-02 -5.11833131e-01 7.58591369e-02 6.90285623e-01
-2.85961002e-01 -6.59881055e-01 6.55153632e-01 -4.74277049e-01
-9.79535043e-01 4.63989645e-01 -5.06542176e-02 -2.93045789e-01
1.40875530e+00 -8.07285488e-01 1.79168776e-01 2.53398806e-01
-5.96846104e-01 3.31956953e-01 5.87196052e-01 5.57826996e-01
5.21535575e-01 -1.11618137e+00 -9.70661417e-02 5.22566438e-01
8.42685848e-02 6.50110364e-01 5.50440252e-01 7.47433245e-01
-4.64663774e-01 3.26988757e-01 -3.62443537e-01 -5.79248726e-01
-7.43311346e-01 8.48400772e-01 6.43227696e-02 -8.79500806e-02
-7.63637662e-01 6.47400796e-01 -1.17992602e-01 -4.60202426e-01
2.66240567e-01 -5.52353382e-01 -1.81716785e-01 3.91092263e-02
4.19450328e-02 6.55953050e-01 1.71603695e-01 -2.55808949e-01
-1.64462090e-01 7.17055619e-01 -2.76453956e-03 5.80001891e-01
1.48078501e+00 -6.14104457e-02 5.10134064e-02 6.45460725e-01
9.17003632e-01 -2.40473345e-01 -1.42990434e+00 7.07477778e-02
2.73966696e-02 -5.11222005e-01 6.92992210e-02 -7.29407609e-01
-1.27177131e+00 9.33704913e-01 5.25663197e-01 1.81631207e-01
1.19985008e+00 7.58874193e-02 9.61217165e-01 4.99606133e-01
6.23633981e-01 -1.40652168e+00 -4.86176051e-02 1.74130350e-01
6.43448532e-01 -1.14402199e+00 2.85906911e-01 -7.14705348e-01
-6.50783122e-01 1.32935894e+00 7.62314796e-01 -3.33620980e-03
5.75554311e-01 1.66733041e-02 -1.44701317e-01 -3.63406122e-01
-1.83599725e-01 -1.40039846e-01 1.94196686e-01 7.51608610e-01
4.42250036e-02 -7.02255666e-02 -2.62314618e-01 7.85098612e-01
-7.44758593e-03 8.02460164e-02 1.89427763e-01 1.08558512e+00
-4.64865923e-01 -1.14120424e+00 -4.80195992e-02 6.54233575e-01
1.24919061e-02 5.16512930e-01 -2.29248941e-01 1.19212139e+00
5.00532329e-01 8.03147316e-01 -6.32236972e-02 -2.12906465e-01
5.58293283e-01 -2.14036815e-02 5.40100992e-01 -7.96142876e-01
-3.03009599e-01 7.33139068e-02 -1.82860866e-01 -3.29753190e-01
-4.82363224e-01 -6.22508168e-01 -1.38502359e+00 1.78484678e-01
-9.39350545e-01 2.01005116e-02 6.20224118e-01 1.08609271e+00
1.84376478e-01 6.98615134e-01 8.34793389e-01 -9.19048429e-01
-7.13744640e-01 -8.94551039e-01 -9.26142931e-01 3.09962064e-01
-1.67118877e-01 -9.98235404e-01 -1.56833217e-01 -2.93824021e-02] | [7.347378253936768, 1.984575867652893] |
c63e7a0b-a30c-4726-857c-80583e3ef64e | bazinga-a-dataset-for-multi-party-dialogues | null | null | https://aclanthology.org/2022.lrec-1.367 | https://aclanthology.org/2022.lrec-1.367.pdf | Bazinga! A Dataset for Multi-Party Dialogues Structuring | We introduce a dataset built around a large collection of TV (and movie) series. Those are filled with challenging multi-party dialogues. Moreover, TV series come with a very active fan base that allows the collection of metadata and accelerates annotation. With 16 TV and movie series, Bazinga! amounts to 400+ hours of speech and 8M+ tokens, including 500K+ tokens annotated with the speaker, addressee, and entity linking information. Along with the dataset, we also provide a baseline for speaker diarization, punctuation restoration, and person entity recognition. The results demonstrate the difficulty of the tasks and of transfer learning from models trained on mono-speaker audio or written text, which is more widely available. This work is a step towards better multi-party dialogue structuring and understanding. Bazinga! is available at hf.co/bazinga. Because (a large) part of Bazinga! is only partially annotated, we also expect this dataset to foster research towards self- or weakly-supervised learning methods. | ['Claude Barras', 'Ruiqing Yin', 'Léo Galmant', 'Aman Berhe', 'Martin Bouteiller', 'Sharleyne Lefevre', 'Benjamin Maurice', 'Hervé Bredin', 'Camille Guinaudeau', 'Juliette Bergoënd', 'Paul Lerner'] | null | null | null | null | lrec-2022-6 | ['punctuation-restoration'] | ['natural-language-processing'] | [ 2.82769836e-02 4.27930027e-01 -5.16439714e-02 -6.76868558e-01
-1.45391977e+00 -8.74049246e-01 7.01077163e-01 2.61734903e-01
-5.86147964e-01 8.49490941e-01 8.12068760e-01 7.57992938e-02
3.85379285e-01 -3.08891058e-01 -5.17842174e-01 -2.04182088e-01
-9.83899981e-02 9.53349650e-01 -5.43579347e-02 -4.83001232e-01
-2.76429236e-01 -3.14596653e-01 -1.28170192e+00 6.45970166e-01
5.63097775e-01 8.33809137e-01 2.20396116e-01 8.87098551e-01
-1.50103226e-01 7.42478371e-01 -8.81337762e-01 -8.48497033e-01
-2.25534756e-02 -3.84494841e-01 -1.39122796e+00 1.71860978e-01
8.97191584e-01 -3.66804659e-01 -3.30543816e-01 6.32199049e-01
8.23721826e-01 2.44050249e-01 1.81524783e-01 -1.20626903e+00
-3.10529888e-01 1.09865415e+00 -1.29755348e-01 3.07389319e-01
7.19340742e-01 4.26274538e-02 1.22701240e+00 -7.01608777e-01
8.34725976e-01 1.19927359e+00 8.94178212e-01 7.99077690e-01
-8.94142032e-01 -5.22434771e-01 -1.44131050e-01 3.54293957e-02
-1.18030870e+00 -1.16834259e+00 5.43888092e-01 -2.35534042e-01
8.75341117e-01 6.27917707e-01 4.55348670e-01 1.37538850e+00
-7.90048242e-01 1.11790061e+00 8.29043686e-01 -6.28021657e-01
-2.09468365e-01 4.99999434e-01 3.30414474e-01 6.25250161e-01
-5.99774361e-01 -6.40393317e-01 -9.89685118e-01 -3.82535547e-01
3.01973760e-01 -5.27269721e-01 -4.64160085e-01 2.77684420e-01
-1.23939419e+00 5.93921185e-01 -2.85915703e-01 2.32898965e-01
-9.57635343e-02 -2.70722836e-01 7.78758168e-01 4.66725320e-01
5.65100670e-01 5.26206553e-01 -5.85660100e-01 -9.66787338e-01
-9.14715827e-01 4.65932161e-01 1.20478368e+00 1.13628983e+00
5.09028614e-01 -1.08608514e-01 -1.19184539e-01 1.36428452e+00
6.72614425e-02 5.36985874e-01 4.07074839e-01 -1.33860385e+00
1.00856626e+00 3.22952807e-01 2.49653220e-01 -5.84710121e-01
-4.94340807e-01 2.75282264e-02 -5.96370935e-01 -6.21609390e-01
9.80439782e-01 -6.24290228e-01 -1.91590890e-01 1.86174595e+00
4.58621293e-01 -1.67036995e-01 -4.73138615e-02 4.42490339e-01
1.01867414e+00 8.00706625e-01 -1.81653380e-01 -1.57730177e-01
1.59195650e+00 -1.02869940e+00 -1.21543086e+00 -6.32891431e-02
8.90118182e-01 -9.58601713e-01 1.17411995e+00 3.36900443e-01
-1.56104314e+00 -4.87923622e-01 -5.07247448e-01 -3.04012805e-01
-2.94169724e-01 1.45630211e-01 6.39383495e-01 7.42395103e-01
-1.10505021e+00 3.71544063e-01 -5.85189819e-01 -5.21500289e-01
2.45741263e-01 1.55601710e-01 -4.98169869e-01 3.11102849e-02
-1.40717149e+00 7.96877801e-01 -6.75939545e-02 -9.23967883e-02
-6.76085114e-01 -6.14364266e-01 -1.08769572e+00 -1.64758310e-01
3.18545818e-01 -1.56275541e-01 1.67035460e+00 -6.20453835e-01
-1.52275002e+00 1.13014412e+00 -1.63011983e-01 -6.09876812e-01
5.69976032e-01 -3.12661409e-01 -5.13721228e-01 2.36564964e-01
2.35268354e-01 8.24155986e-01 5.51473320e-01 -6.36838078e-01
-7.80683160e-01 -4.36792403e-01 4.57797796e-02 3.71774882e-01
-8.22490692e-01 5.60701966e-01 -5.03909409e-01 -5.99872112e-01
-4.61456120e-01 -9.30452168e-01 3.54096174e-01 -3.77629966e-01
-6.17479801e-01 -5.11680901e-01 7.99752653e-01 -1.28742874e+00
1.31686258e+00 -2.24061012e+00 -3.50679606e-02 -2.52029479e-01
9.28249508e-02 1.35433659e-01 -1.37791922e-02 7.78554440e-01
1.33957639e-01 9.43035036e-02 -4.67154123e-02 -9.77262139e-01
2.30129004e-01 -3.73227820e-02 -3.21526825e-01 3.64499956e-01
-1.36206001e-02 6.30111217e-01 -8.44910383e-01 -5.14918804e-01
-1.50337741e-01 2.74710834e-01 -4.57194477e-01 2.62803555e-01
-4.06364463e-02 5.86291730e-01 -8.79175812e-02 5.53134680e-01
1.77459434e-01 -1.72877029e-01 1.69004023e-01 -6.75059110e-02
-2.73087561e-01 9.24300194e-01 -1.27540612e+00 1.96742630e+00
-4.06668067e-01 9.86335278e-01 7.13602960e-01 -5.81309438e-01
6.05042636e-01 8.16536784e-01 4.76425320e-01 -3.05327863e-01
7.23050609e-02 4.08780351e-02 -3.01861405e-01 -5.69321275e-01
1.11059403e+00 3.45122546e-01 -3.23725730e-01 6.75923824e-01
4.38553870e-01 -2.57110089e-01 4.02903736e-01 5.05737960e-01
1.09112954e+00 -3.17576200e-01 -7.50044435e-02 -1.36303334e-02
3.73497784e-01 2.03936715e-02 4.83208746e-01 6.80008948e-01
-4.47041750e-01 5.21588326e-01 2.52575904e-01 -1.34840846e-01
-1.20412421e+00 -6.14550352e-01 -2.90153831e-01 1.55644381e+00
-4.59953159e-01 -8.12098801e-01 -1.08246207e+00 -5.10773718e-01
-2.90561259e-01 5.33675432e-01 -3.10019165e-01 3.12733322e-01
-7.78096974e-01 -5.53306401e-01 1.11532152e+00 3.70640218e-01
5.89491248e-01 -8.65910232e-01 2.58277506e-01 1.76891088e-01
-8.65894437e-01 -1.22227859e+00 -9.60143328e-01 -1.16984442e-01
-3.52141827e-01 -9.07535195e-01 -8.13987613e-01 -1.00405157e+00
2.20125541e-01 2.30755396e-02 1.22496474e+00 -2.41620451e-01
-2.27180608e-02 6.97306573e-01 -4.01343763e-01 -2.21088931e-01
-7.67539442e-01 2.54460841e-01 7.42310360e-02 -4.42360491e-02
2.70266503e-01 -4.10535991e-01 -1.14960991e-01 3.20240825e-01
-4.36184317e-01 -4.32930589e-02 -4.01282944e-02 7.81658351e-01
1.11874953e-01 -2.46688157e-01 6.92209899e-01 -1.08148956e+00
5.88775337e-01 -3.50576520e-01 -2.30415791e-01 1.09369986e-01
1.26826810e-02 -4.71221030e-01 2.68131375e-01 -3.49125475e-01
-1.22115922e+00 -7.79061094e-02 -4.54392791e-01 1.64355874e-01
-4.01760340e-01 2.70767748e-01 -4.17523384e-01 4.62989211e-01
6.51782513e-01 4.96243834e-02 -1.55611113e-02 -8.23147655e-01
5.13810456e-01 1.28230381e+00 8.93103659e-01 -5.42795002e-01
4.81714308e-01 3.85698915e-01 -9.64189112e-01 -1.24530411e+00
-7.71962702e-01 -6.86999679e-01 -5.31589031e-01 -2.82637000e-01
7.90398479e-01 -1.14627767e+00 -7.52222359e-01 6.79888487e-01
-1.19939995e+00 -6.12525642e-01 -4.28003997e-01 3.64804834e-01
-3.41891021e-01 4.10421371e-01 -1.08871484e+00 -9.95292544e-01
-3.45295548e-01 -7.05062091e-01 1.03184938e+00 5.91303632e-02
-7.58036852e-01 -9.18358147e-01 1.30851135e-01 1.16726172e+00
2.30682015e-01 -2.79768795e-01 4.03504670e-01 -1.14589596e+00
-4.32901233e-02 -4.60766315e-01 2.78922558e-01 3.02392781e-01
3.75531055e-02 -2.43308127e-01 -1.23468351e+00 -3.57178241e-01
-7.04812929e-02 -8.96539092e-01 4.48372483e-01 1.64485127e-01
8.64916444e-01 -6.23542011e-01 -2.51011900e-03 1.51729375e-01
3.38018417e-01 -6.94393367e-02 2.82569736e-01 2.16026992e-01
6.24265850e-01 1.00177264e+00 4.78198379e-01 7.05742359e-01
1.00581419e+00 8.86741757e-01 -1.36811733e-01 1.41161308e-01
-2.87632257e-01 -2.38122106e-01 5.68270981e-01 1.38055480e+00
3.32006142e-02 -2.87723064e-01 -1.05738807e+00 7.08752394e-01
-1.59624267e+00 -1.26967871e+00 -2.51427352e-01 2.07546902e+00
1.55410326e+00 -2.70575322e-02 6.44279242e-01 9.09577832e-02
9.98353362e-01 1.63522288e-01 -4.32294980e-02 -7.41093680e-02
-3.05321127e-01 -1.45131186e-01 1.74816102e-01 6.52045608e-01
-1.31294382e+00 8.28060329e-01 6.58850384e+00 8.66263926e-01
-7.21334994e-01 3.51907402e-01 5.67503214e-01 -2.35525444e-01
-1.30803987e-01 -3.18447769e-01 -1.33676362e+00 6.23520374e-01
1.37598228e+00 -4.49648267e-03 4.60633218e-01 6.30185962e-01
1.99527621e-01 1.39380977e-01 -1.22357082e+00 1.03362036e+00
2.83306867e-01 -1.40438485e+00 -4.23618287e-01 1.12083629e-02
5.72925389e-01 3.34986240e-01 -1.83099702e-01 5.78385234e-01
4.68604922e-01 -5.93693078e-01 7.75021255e-01 1.26532644e-01
7.30819643e-01 -4.89460766e-01 6.06944025e-01 4.81296748e-01
-8.20586741e-01 1.44760162e-01 6.55731559e-02 6.74121752e-02
4.29302365e-01 2.83055365e-01 -9.92924809e-01 2.57609636e-01
8.75807285e-01 7.27731764e-01 -5.97429574e-01 8.19739282e-01
-5.25792800e-02 8.15173447e-01 -4.92302477e-01 9.84140486e-02
-4.23890054e-02 1.41048223e-01 7.08532035e-01 1.50326407e+00
-7.25300089e-02 -1.00230202e-01 4.24349815e-01 3.66764575e-01
-5.34494460e-01 2.34281793e-01 -4.31945145e-01 -2.83510149e-01
9.52326596e-01 1.41603398e+00 -4.26085562e-01 -5.29988885e-01
-4.66626018e-01 9.87627625e-01 3.12136650e-01 5.98898605e-02
-5.38478494e-01 -3.74265462e-01 4.87346172e-01 1.23973683e-01
2.13669688e-02 -2.14459017e-01 7.83093795e-02 -1.26889884e+00
-5.81806675e-02 -1.26601124e+00 6.25612020e-01 -5.26936650e-01
-1.41856289e+00 7.54290700e-01 -9.26396847e-02 -7.77089000e-01
-6.13656044e-01 -1.43170729e-01 -3.77703518e-01 5.33717453e-01
-1.21147668e+00 -1.27660680e+00 -1.00819170e-01 7.26300299e-01
8.58360171e-01 -2.74061024e-01 1.19435024e+00 7.69096434e-01
-7.35810518e-01 9.61717427e-01 1.54069886e-01 7.10083604e-01
1.36212850e+00 -1.38842463e+00 4.11718100e-01 3.68140429e-01
3.29594523e-01 3.81580114e-01 7.28676260e-01 -4.59051788e-01
-1.22602499e+00 -9.07253623e-01 1.28038979e+00 -1.13838136e+00
8.76475990e-01 -7.99647212e-01 -8.81163001e-01 8.85656357e-01
6.18372798e-01 -3.52358609e-01 1.11048615e+00 6.43987536e-01
-1.86612383e-01 -2.76099220e-02 -1.00848222e+00 2.82409787e-01
8.49060059e-01 -9.21175659e-01 -7.58513987e-01 7.51882613e-01
6.29161537e-01 -7.46446967e-01 -1.13718796e+00 -1.79760039e-01
2.19452962e-01 -5.19342065e-01 7.42089927e-01 -4.86420542e-01
1.78360686e-01 2.21768379e-01 -1.48450822e-01 -1.16884673e+00
3.01885843e-01 -1.23987305e+00 -1.36646077e-01 2.17275119e+00
6.20970964e-01 -1.81961119e-01 6.50351644e-01 1.02121162e+00
-5.14804959e-01 -9.61333439e-02 -1.10567760e+00 -6.86956942e-01
-1.30758628e-01 -5.22224188e-01 3.99238020e-01 1.42423105e+00
4.41049457e-01 6.90317094e-01 -6.01329803e-01 -1.25878692e-01
3.71119261e-01 -2.13812307e-01 1.02618194e+00 -1.04757464e+00
-3.83024186e-01 -1.89081967e-01 -5.46737760e-02 -1.20188224e+00
1.90123394e-01 -7.97176659e-01 6.10247068e-02 -1.10248160e+00
3.48184071e-02 -5.90269506e-01 3.37628394e-01 7.58042037e-01
-7.05217794e-02 3.14587295e-01 2.77091227e-02 3.48943561e-01
-1.06564951e+00 4.52101797e-01 7.96838343e-01 -1.93336025e-01
-3.83830339e-01 1.19044989e-01 -5.06579280e-01 8.06013703e-01
6.93112075e-01 -2.46006876e-01 3.51598598e-02 -6.52082622e-01
-2.85456870e-02 1.85647219e-01 9.53981355e-02 -7.34642506e-01
3.53625596e-01 4.27345932e-01 -2.17140149e-02 -6.09072268e-01
7.99932480e-01 -2.85239965e-01 -3.19557160e-01 -1.90492854e-01
-6.98899031e-01 3.65730152e-02 2.03669593e-01 3.36271465e-01
-3.63259435e-01 -2.01125771e-01 3.87030333e-01 -2.26073340e-01
-4.21711296e-01 7.57352710e-02 -5.12919068e-01 7.17982471e-01
4.82316494e-01 9.52650830e-02 -4.85107929e-01 -8.20334852e-01
-9.16776180e-01 4.45032001e-01 2.46025607e-01 4.24288213e-01
-2.23734621e-02 -1.16123974e+00 -8.54878604e-01 -9.27286074e-02
-4.26474288e-02 -1.19166691e-02 3.61632735e-01 7.65676439e-01
-6.15549572e-02 3.31762314e-01 4.59596738e-02 -3.50305647e-01
-1.59703445e+00 -2.17298567e-02 -6.40546754e-02 -8.73005837e-02
-5.93190312e-01 1.07009232e+00 -4.42279309e-01 -6.72507465e-01
7.60995567e-01 8.58148858e-02 -3.83678406e-01 5.70845187e-01
9.31817710e-01 5.60524344e-01 -7.89365694e-02 -8.73895407e-01
-2.24166349e-01 -3.40461761e-01 -1.83612078e-01 -6.10291243e-01
1.60526991e+00 -5.55808783e-01 -1.13568045e-01 7.15643942e-01
1.05403578e+00 5.26175797e-01 -1.15415049e+00 -4.84976053e-01
5.16226590e-02 -1.99773997e-01 -1.27809823e-01 -7.03948140e-01
-5.96847773e-01 8.09053719e-01 2.73020178e-01 5.72488964e-01
5.76350212e-01 3.35113168e-01 1.18410265e+00 5.94297945e-01
1.53540239e-01 -1.50672305e+00 1.02053590e-01 7.64518082e-01
8.59113693e-01 -1.56011415e+00 -3.36453438e-01 -2.95534253e-01
-7.86852777e-01 7.08441436e-01 6.31211340e-01 5.58485568e-01
3.58528733e-01 1.78742796e-01 2.64605373e-01 -7.26762637e-02
-6.12456799e-01 8.87591857e-03 -9.61951315e-02 7.02002466e-01
6.42000198e-01 -1.08447589e-01 1.63324594e-01 9.78533030e-01
-7.43795812e-01 -4.83119577e-01 5.47019243e-01 8.19598913e-01
-3.12133998e-01 -1.27774692e+00 -4.03552324e-01 1.66780710e-01
-6.91252172e-01 -3.68585795e-01 -4.91415888e-01 5.83829999e-01
-1.65222108e-01 1.33068419e+00 1.80460494e-02 -8.52844715e-02
2.83664107e-01 5.19283891e-01 2.67409056e-01 -8.74576628e-01
-9.62046146e-01 5.21333739e-02 9.48690414e-01 -2.43537441e-01
-4.01415050e-01 -1.05204797e+00 -9.25247252e-01 -5.10038376e-01
-4.75592613e-01 6.47081673e-01 7.05481470e-01 8.85839820e-01
3.60131204e-01 2.87899524e-01 5.54585099e-01 -7.51294076e-01
-3.80225092e-01 -1.27229333e+00 -5.32754123e-01 4.73259002e-01
1.55497432e-01 -1.65322974e-01 -2.28635982e-01 3.85933042e-01] | [13.713031768798828, 7.149754524230957] |
aee8e49e-6277-4152-b7de-0dcc549b3573 | sensor-data-driven-analysis-for | 2301.06300 | null | https://arxiv.org/abs/2301.06300v1 | https://arxiv.org/pdf/2301.06300v1.pdf | Sensor data-driven analysis for identification of causal relationships between exposure to air pollution and respiratory rate in asthmatics | According to the Lancet report on the global burden of disease published in October 2020, air pollution is among the five highest risk factors for global health, reducing life expectancy on average by 20 months. This paper describes a data-driven method for establishing causal relationships within and between two multivariate time series data streams derived from wearable sensors: personal exposure to airborne particulate matter of aerodynamic sizes less than 2.5um (PM2.5) gathered from the Airspeck monitor worn on the person and continuous respiratory rate (breaths per minute) measured by the Respeck monitor worn as a plaster on the chest. Results are presented for a cohort of 113 asthmatic adolescents using the PCMCI+ algorithm to learn the short-term causal relationships between lags of \pm exposure and respiratory rate. We consider causal effects up to a maximum delay of 8 hours, using data at both a 1 minute and 15 minute resolution in different experiments. For the first time a personalised exposure-response relationship between PM2.5 exposure and respiratory rate has been demonstrated to exist for short-term effects in asthmatic adolescents during their everyday lives. Our results lead to recommendations for work on specific open problems in causal discovery, to increase the feasibility of this approach for similar epidemiology studies in the future. | ['S Maiya', 'D K Arvind'] | 2023-01-16 | null | null | null | null | ['causal-discovery', 'epidemiology'] | ['knowledge-base', 'medical'] | [ 5.84627151e-01 -6.83751553e-02 -1.32007182e-01 -2.11023092e-01
-4.97916996e-01 -5.18978477e-01 5.51939070e-01 6.02271318e-01
-3.55691820e-01 9.81803119e-01 1.31126165e-01 -4.83647346e-01
-7.98125267e-01 -1.17075479e+00 -7.28909135e-01 -5.91460407e-01
-3.18765849e-01 3.22058111e-01 1.52260259e-01 5.94461322e-01
-4.79331285e-01 3.22782844e-01 -1.40216041e+00 6.03900813e-02
7.31042445e-01 5.80965400e-01 4.02191542e-02 8.94389331e-01
3.45913261e-01 -1.32280663e-01 -7.70410419e-01 7.74630159e-02
-8.42681155e-03 -2.56849289e-01 -2.06964865e-01 -4.02160972e-01
4.72887099e-01 1.83327012e-02 8.91097784e-02 3.00997436e-01
5.42171180e-01 6.02801740e-02 3.91056091e-01 -8.65806758e-01
-4.14386205e-02 2.78730869e-01 -1.46090880e-01 6.07672513e-01
5.08098423e-01 1.58142328e-01 4.52401221e-01 -6.04047580e-03
-6.03402033e-02 1.19542873e+00 8.30888987e-01 4.93414015e-01
-1.44457400e+00 -8.08151603e-01 9.82947648e-02 -2.79158533e-01
-1.11861062e+00 -1.33850396e-01 2.17907235e-01 -6.58381760e-01
9.95672405e-01 9.50132966e-01 1.07623100e+00 1.25279891e+00
9.18674231e-01 -5.40063262e-01 1.44391859e+00 6.06742576e-02
3.43183070e-01 6.04927577e-02 -1.24967292e-01 -1.64904490e-01
5.98562300e-01 6.22152984e-01 -3.43973845e-01 -5.53120315e-01
4.15133953e-01 3.31611991e-01 1.04080349e-01 7.76107609e-01
-7.45228112e-01 5.06318092e-01 7.70647004e-02 3.84340554e-01
-4.59812105e-01 5.12508452e-01 -6.25315830e-02 2.59083420e-01
1.00776851e+00 1.97939381e-01 -7.33388543e-01 -1.38728738e-01
-7.60817051e-01 5.16166866e-01 6.50098443e-01 2.97653973e-01
5.06088734e-02 -4.43812788e-01 -2.30835646e-01 4.63899642e-01
7.59849310e-01 1.19763982e+00 1.50520671e-02 -5.75609684e-01
4.19696569e-01 1.75441086e-01 3.22855443e-01 -8.19793761e-01
-5.84856093e-01 -2.07385212e-01 -5.53601444e-01 -4.00192499e-01
3.12313288e-01 -5.57556689e-01 -3.78901809e-01 1.30706573e+00
7.98659921e-01 5.51860988e-01 -5.57315707e-01 4.17619109e-01
6.17404401e-01 6.22875154e-01 3.94644320e-01 -7.36941516e-01
1.50183082e+00 8.61962140e-02 -7.87415981e-01 9.93531868e-02
5.42400852e-02 -4.87777650e-01 5.52294970e-01 4.60301071e-01
-9.28500354e-01 -7.22979844e-01 -6.84943974e-01 6.47665560e-01
-5.66741347e-01 -5.55800438e-01 2.76730686e-01 1.32999122e+00
-4.87058818e-01 9.29399252e-01 -1.06621647e+00 -3.42758477e-01
1.29698724e-01 4.06152457e-01 2.00559065e-01 1.70663759e-01
-1.52980602e+00 4.32847738e-01 -2.61781543e-01 -9.56243500e-02
-7.79194772e-01 -1.59106219e+00 -2.73028731e-01 -4.15753692e-01
-3.05180728e-01 -9.08985913e-01 1.01969969e+00 4.69235629e-02
-1.19974351e+00 4.04452443e-01 -6.11603111e-02 -5.32943487e-01
4.68191355e-01 -5.95535159e-01 -1.18461382e+00 3.64664122e-02
1.89704463e-01 -4.55427617e-02 8.18286896e-01 -5.33306360e-01
-4.32439029e-01 -6.98635399e-01 -4.78713661e-01 -3.42592239e-01
-1.95987657e-01 2.80674309e-01 5.02924263e-01 -3.75492722e-01
-3.25801015e-01 -1.00651705e+00 -2.89847493e-01 -4.37901586e-01
-2.34131724e-01 -3.66468966e-01 5.54907262e-01 -6.79549694e-01
1.37910330e+00 -1.60162270e+00 -7.67708719e-01 1.27967447e-01
-1.19683528e-02 2.41603732e-01 2.60149330e-01 6.67851269e-01
-3.63577336e-01 2.27995455e-01 -1.24633729e-01 -4.32137847e-02
-2.16604456e-01 1.48684427e-01 -2.04442248e-01 1.00720716e+00
1.51474595e-01 7.58493721e-01 -1.24752426e+00 8.28673616e-02
3.62225920e-01 7.99372256e-01 2.85243355e-02 2.37732321e-01
-2.19992757e-01 8.40423584e-01 -2.33135432e-01 1.06029361e-01
4.82595176e-01 2.29828686e-01 -7.71940202e-02 4.66335267e-01
-6.64667428e-01 7.44706273e-01 -1.06140673e+00 1.02967143e+00
-4.77558792e-01 3.78853381e-01 -1.18815623e-01 -2.74279773e-01
7.27548063e-01 1.03590596e+00 7.82494128e-01 -8.01688731e-01
-9.85679328e-02 -1.93843439e-01 -5.85892051e-02 -8.58202934e-01
-4.02037144e-01 -7.18058169e-01 2.24648342e-01 4.56717014e-01
-7.10939407e-01 -1.88182384e-01 -7.86902234e-02 -3.53194654e-01
1.36259365e+00 -3.89084905e-01 -2.48602945e-02 -3.76859426e-01
2.57668853e-01 -3.72278243e-01 2.06811100e-01 6.23712242e-01
1.32607773e-01 3.09844792e-01 -5.23564890e-02 -2.78498143e-01
-5.68906069e-01 -1.43689907e+00 -8.21680963e-01 6.11673951e-01
-5.97406328e-01 -2.72431701e-01 -5.41987777e-01 -1.88580737e-01
3.38148624e-01 8.96518052e-01 -6.61598563e-01 1.19207419e-01
-4.43842441e-01 -1.14234900e+00 5.00870168e-01 4.33847070e-01
4.90488745e-02 -6.60404205e-01 -5.92319787e-01 4.87754256e-01
2.72266746e-01 -7.79129326e-01 -6.39894009e-02 -2.22262577e-03
-1.13649249e+00 -1.16690552e+00 -2.54514724e-01 2.37411708e-01
3.10017884e-01 6.27912953e-02 9.67178822e-01 -1.36767194e-01
-6.14422441e-01 9.06096995e-01 1.52753433e-02 -1.28503108e+00
-1.29993916e-01 -4.06917244e-01 4.86630499e-01 -1.02651112e-01
2.79910803e-01 -1.02741396e+00 -1.02319026e+00 1.73640609e-01
-6.95178866e-01 -6.13812923e-01 -3.85220312e-02 -1.45971909e-01
6.34309769e-01 4.96348172e-01 8.35487068e-01 -8.40809524e-01
4.04459655e-01 -1.03153443e+00 -6.99837148e-01 -3.75874370e-01
-1.01508844e+00 -6.22009099e-01 3.42892736e-01 -5.48912287e-01
-9.36387658e-01 -2.45411262e-01 -1.90339088e-01 8.05240199e-02
-5.99268436e-01 -1.42908141e-01 -1.38549015e-01 6.40604496e-01
7.59604037e-01 -3.09511960e-01 -2.66643018e-01 -7.79678583e-01
1.89474761e-01 5.03083348e-01 3.67588013e-01 -4.71880198e-01
7.02683091e-01 5.84147811e-01 4.64249015e-01 -1.28432751e+00
-7.26183593e-01 -7.04349160e-01 -4.87766296e-01 -3.02788168e-01
1.28731203e+00 -1.05525315e+00 -9.33669388e-01 -2.99296863e-02
-9.62191343e-01 -3.71497571e-01 -4.26892132e-01 7.25798488e-01
-2.15037033e-01 -3.75745326e-01 -3.04099202e-01 -1.26808298e+00
-4.67140734e-01 -1.01209342e-01 9.66089785e-01 4.45755459e-02
-6.96576595e-01 -1.56800914e+00 8.05789173e-01 6.15866840e-01
2.64936447e-01 9.04118001e-01 6.19562328e-01 -3.00444126e-01
-6.16530180e-02 -3.83148998e-01 2.99937576e-01 1.37676314e-01
4.74893183e-01 3.76247317e-02 -1.19343710e+00 -3.61204088e-01
4.81918514e-01 5.52257240e-01 5.84788740e-01 5.67380250e-01
1.09074676e+00 -5.68906426e-01 -7.70633340e-01 5.99075370e-02
1.22779059e+00 3.80609453e-01 4.12699789e-01 -2.18208730e-01
6.49142802e-01 7.50252545e-01 3.89345974e-01 2.96566159e-01
4.32142578e-02 3.60118717e-01 4.16876882e-01 7.42457509e-02
-1.04955427e-01 -3.32463831e-01 4.04652625e-01 5.16729057e-01
-1.96569562e-01 -3.97618473e-01 -6.83941185e-01 5.99867463e-01
-1.05460441e+00 -1.06141412e+00 -1.19790792e+00 2.68204284e+00
7.66084731e-01 4.63283397e-02 5.32050788e-01 3.32696915e-01
5.71147144e-01 7.68627971e-03 -2.48779416e-01 -5.08181572e-01
5.84028482e-01 6.58872306e-01 6.71777129e-01 6.15038812e-01
-8.11545789e-01 -3.08721602e-01 7.17990875e+00 2.17663780e-01
-8.41755867e-01 5.66952288e-01 4.72606331e-01 -5.88265359e-01
-4.76299316e-01 -9.91478562e-02 -9.78410423e-01 9.63836551e-01
2.07151794e+00 -9.76318121e-02 8.74287933e-02 1.94621384e-02
1.06568420e+00 -2.74848044e-01 -1.18463159e+00 2.95197934e-01
-6.84498250e-01 -6.38755023e-01 -7.21324444e-01 4.20100570e-01
8.33846867e-01 8.51535648e-02 1.11944288e-01 -3.67199928e-01
-7.76272938e-02 -1.13620448e+00 9.28272754e-02 1.02018607e+00
8.42127740e-01 -6.82215512e-01 1.28324464e-01 5.86501122e-01
-1.09049571e+00 9.61734727e-02 -2.08856434e-01 -5.96675634e-01
1.84171543e-01 1.60391498e+00 -1.21557641e+00 4.02820528e-01
6.35104299e-01 4.05983150e-01 -2.47316256e-01 8.01163912e-01
-3.21475029e-01 1.40692663e+00 -7.70244956e-01 -1.72411546e-01
-2.93611050e-01 3.24637890e-02 7.26070344e-01 1.30477405e+00
4.24288660e-01 3.77315223e-01 -6.06091738e-01 9.61171091e-01
5.16658783e-01 -1.92295268e-01 -7.54880428e-01 -3.74791138e-02
5.23747563e-01 1.04037237e+00 -3.70604455e-01 5.57601079e-02
-4.85295326e-01 1.42533153e-01 -6.69785619e-01 8.89489353e-02
-8.66476178e-01 9.87766162e-02 7.05155551e-01 1.25560999e+00
-6.67265756e-03 -8.24265555e-03 -2.49833122e-01 -6.27167001e-02
-1.81394190e-01 -3.51688117e-01 5.53193986e-01 -3.53190452e-01
-1.32752550e+00 -3.04972500e-01 5.43082833e-01 -7.71296024e-01
1.52271967e-02 -3.64168316e-01 -1.10058832e+00 1.05531871e+00
-8.60122800e-01 -4.13041919e-01 7.12856203e-02 3.03161860e-01
1.98914960e-01 7.20829368e-01 9.46325004e-01 4.33530927e-01
-5.21598279e-01 -2.90236939e-02 7.20588714e-02 -5.69102287e-01
3.91165614e-01 -1.36629534e+00 7.03205526e-01 2.98013628e-01
-2.45388389e-01 7.14377820e-01 9.14168656e-01 -1.34521246e+00
-1.34479976e+00 -1.57830691e+00 1.31814623e+00 -1.14218450e+00
6.81279957e-01 -3.67020190e-01 -5.76940417e-01 2.22631395e-01
1.27763733e-01 -2.31202349e-01 1.11472821e+00 1.39617339e-01
1.70757383e-01 -5.27184308e-01 -1.28307307e+00 -1.12368353e-01
7.39482284e-01 -2.77284950e-01 -5.11596620e-01 8.59219551e-01
9.73521352e-01 4.09027515e-03 -1.54540694e+00 4.12576735e-01
6.21640921e-01 -6.89598382e-01 1.14552200e+00 -4.39981639e-01
3.80803011e-02 -1.48249894e-01 2.30572879e-01 -1.13687444e+00
-9.25490037e-02 -8.18835557e-01 -1.36044785e-01 1.32034028e+00
1.57319784e-01 -9.33662415e-01 3.15456659e-01 6.16114557e-01
3.47598910e-01 -6.32993162e-01 -1.21048939e+00 -9.96574938e-01
2.77819961e-01 -8.91085744e-01 7.51940489e-01 4.74995881e-01
-3.48093241e-01 4.48980629e-01 -3.97053957e-02 5.70658028e-01
6.76010728e-01 -3.02354395e-01 3.05805475e-01 -1.64812768e+00
-2.25347981e-01 2.22714648e-01 7.96964765e-02 -3.87955308e-01
-4.55582798e-01 -4.74217445e-01 -2.78047472e-02 -1.51259410e+00
-6.68570725e-03 -3.24153751e-01 -2.99361616e-01 -7.69151077e-02
-2.17925385e-01 -2.09086537e-01 -2.48815790e-01 -5.59941113e-01
2.09674202e-02 -6.97318017e-02 1.15737128e+00 -5.10696359e-02
-3.92196655e-01 7.25051820e-01 -3.33099842e-01 5.23001254e-01
8.82116020e-01 -1.08252883e+00 -7.10384190e-01 1.42198289e-03
6.21327102e-01 2.70813882e-01 7.24789083e-01 -1.22043276e+00
-9.29444432e-02 -3.70674551e-01 6.89660072e-01 -6.76833212e-01
2.61990845e-01 -1.01545513e+00 9.96996522e-01 8.23902786e-01
-1.55853420e-01 1.75986409e-01 3.17924112e-01 9.38402951e-01
6.01875603e-01 4.73570637e-02 4.80427027e-01 -9.48545262e-02
6.53314412e-01 4.32875335e-01 -6.21135056e-01 -7.31435372e-03
9.49679375e-01 1.87466159e-01 -3.41179013e-01 -5.61109222e-02
-1.00119781e+00 1.17532477e-01 -1.97876960e-01 4.80074376e-01
2.04833746e-01 -1.31591225e+00 -6.59045100e-01 2.50349849e-01
-3.79332870e-01 -2.10309580e-01 3.48991811e-01 9.90655541e-01
1.85302988e-01 8.27632070e-01 5.32133877e-01 -6.21636569e-01
-1.34684360e+00 6.53525412e-01 4.13002789e-01 -2.24909961e-01
-4.10189420e-01 8.34819257e-01 -1.54231802e-01 8.41189921e-02
-7.30445758e-02 -5.16973555e-01 4.23875488e-02 2.61229724e-01
8.58845651e-01 1.34301364e+00 2.32282043e-01 -1.05124325e-01
-4.26427811e-01 4.10829544e-01 6.03482008e-01 -1.48324326e-01
1.23138809e+00 -1.70742020e-01 -3.42312872e-01 1.04085732e+00
1.03617346e+00 3.83052200e-01 -9.27223623e-01 3.67125958e-01
-3.85872781e-01 -2.00185806e-01 -2.66603857e-01 -9.02287960e-01
-4.48651940e-01 7.21867502e-01 1.13077366e+00 8.23193908e-01
9.41595674e-01 6.55969419e-03 8.47177327e-01 -1.47019148e-01
-6.20683171e-02 -8.92269433e-01 -2.07295582e-01 -1.95978239e-01
7.92461574e-01 -5.94613135e-01 4.66367900e-01 -3.53124440e-01
4.99887615e-01 6.15754008e-01 7.50068352e-02 -8.92792270e-02
1.14748299e+00 2.69282520e-01 -9.65274423e-02 -4.72101092e-01
-1.00999534e+00 8.20447505e-02 4.64613706e-01 8.20925236e-01
6.43819749e-01 8.71128321e-01 -6.39472723e-01 5.70370674e-01
-4.24838036e-01 -2.93840915e-01 -1.40815467e-01 5.55138290e-01
-5.44394732e-01 -1.18651819e+00 -8.31019402e-01 7.95755506e-01
-8.20839703e-01 1.63652793e-01 -3.51170063e-01 6.75569892e-01
8.34925652e-01 1.78031588e+00 3.11009914e-01 1.61766578e-02
7.92888105e-01 1.50283575e-01 5.34465253e-01 -8.09870064e-01
-7.38783956e-01 -4.93422821e-02 2.14962035e-01 -4.57997650e-01
-5.49942732e-01 -1.18426335e+00 -1.06032646e+00 -2.62093544e-01
-9.53747332e-02 5.34094498e-02 6.81576967e-01 9.37258720e-01
1.94896515e-02 7.84349799e-01 7.39561200e-01 -2.26120010e-01
-1.73418239e-01 -9.52919006e-01 -6.49981737e-01 -7.51415687e-03
6.11779749e-01 -3.17753524e-01 -4.87398297e-01 -8.01752135e-02] | [6.190474033355713, 2.605403423309326] |
d13cf24f-e992-4a0d-968c-866b6f1b35e0 | large-scale-multi-modal-person-identification | 1912.12134 | null | https://arxiv.org/abs/1912.12134v1 | https://arxiv.org/pdf/1912.12134v1.pdf | Large-scale Multi-modal Person Identification in Real Unconstrained Environments | Person identification (P-ID) under real unconstrained noisy environments is a huge challenge. In multiple-feature learning with Deep Convolutional Neural Networks (DCNNs) or Machine Learning method for large-scale person identification in the wild, the key is to design an appropriate strategy for decision layer fusion or feature layer fusion which can enhance discriminative power. It is necessary to extract different types of valid features and establish a reasonable framework to fuse different types of information. In traditional methods, different persons are identified based on single modal features to identify, such as face feature, audio feature, and head feature. These traditional methods cannot realize a highly accurate level of person identification in real unconstrained environments. The study aims to propose a fusion module to fuse multi-modal features for person identification in real unconstrained environments. | ['Jiajie Ye', 'Yisheng Guan', 'Xinghong Huang', 'Hong Zhang', 'Junfa Liu'] | 2019-12-17 | null | null | null | null | ['person-identification', 'multi-modal-person-identification'] | ['computer-vision', 'miscellaneous'] | [ 1.43716354e-02 -6.74242914e-01 3.26263100e-01 -5.92117310e-01
-7.90640295e-01 -3.33438218e-01 3.60115469e-01 -7.44167566e-02
-6.83137119e-01 7.01578140e-01 4.83382314e-01 3.81484032e-01
-4.05189753e-01 -6.30181551e-01 -1.79902911e-01 -7.91404009e-01
8.22676048e-02 2.20052510e-01 -2.83029705e-01 -8.54741037e-02
-5.57768829e-02 6.07154310e-01 -1.92801499e+00 2.39621162e-01
5.48581421e-01 9.96160805e-01 -1.08147793e-01 6.25755250e-01
-7.17793405e-02 -3.22360992e-02 -1.01033580e+00 -4.45844829e-01
3.22505027e-01 -1.45402327e-01 -5.34938812e-01 -1.75946265e-01
3.52180809e-01 -5.57382762e-01 -1.97537750e-01 1.11419404e+00
1.36823869e+00 1.66084077e-02 6.89963162e-01 -1.40227258e+00
-3.36898655e-01 6.31097555e-01 -2.96832412e-01 1.07601799e-01
7.03584492e-01 1.04193695e-01 3.68448406e-01 -6.98440254e-01
-1.17994249e-01 1.72904515e+00 9.94507313e-01 6.36031747e-01
-8.76247108e-01 -8.43506336e-01 -8.24267883e-03 3.59511614e-01
-1.68040299e+00 -6.46806359e-01 5.87250888e-01 -3.58542562e-01
4.70884115e-01 3.71308178e-01 4.96453822e-01 1.28558278e+00
-1.65326551e-01 6.51754558e-01 9.90959287e-01 -3.51230264e-01
-1.29095286e-01 3.58083576e-01 3.75712365e-01 2.72749901e-01
4.08883810e-01 2.42554665e-01 -6.56169474e-01 -1.94058612e-01
5.53603947e-01 2.96215415e-01 -1.46581262e-01 5.90097368e-01
-1.15931880e+00 6.29609704e-01 2.37070054e-01 6.57954931e-01
-4.22075242e-01 -1.29293099e-01 5.47979474e-01 2.44275212e-01
-8.64599645e-02 1.80281162e-01 -3.81220967e-01 -1.72840282e-01
-7.20635056e-01 4.18082505e-01 7.34345555e-01 5.01148939e-01
6.99714541e-01 9.01950002e-02 -4.25068170e-01 7.98649013e-01
2.98684090e-01 7.48261273e-01 6.92117870e-01 -6.92480385e-01
1.87683940e-01 5.72809398e-01 1.54264912e-01 -1.10284638e+00
-5.33154786e-01 -3.94541144e-01 -1.12203491e+00 7.19574168e-02
3.37355405e-01 -5.63577414e-01 -5.33657134e-01 1.61891401e+00
2.06767872e-01 1.00727946e-01 1.94949120e-01 8.87940228e-01
1.31932998e+00 3.26503783e-01 1.93931729e-01 -1.26272142e-01
1.59022582e+00 -2.63812006e-01 -9.89692807e-01 2.34009251e-02
1.99464709e-01 -7.29657769e-01 4.71743613e-01 3.49126101e-01
-5.96642852e-01 -1.10847449e+00 -9.54105020e-01 3.46758962e-01
-5.90360940e-01 4.47930425e-01 4.66656655e-01 1.07968640e+00
-8.65268826e-01 2.64979005e-01 -8.40489343e-02 -3.31121802e-01
1.45440742e-01 9.65200484e-01 -8.18258405e-01 4.66264039e-02
-1.52269053e+00 6.55428827e-01 4.52676892e-01 5.21798313e-01
-6.39343262e-01 -2.01474652e-01 -8.55178535e-01 -3.10171470e-02
-6.64370582e-02 -7.14327276e-01 9.54575539e-01 -8.59221041e-01
-1.29316103e+00 4.70677882e-01 -3.11313450e-01 -1.87389463e-01
2.30106622e-01 -2.83777535e-01 -7.52826929e-01 -9.12588462e-02
2.40898132e-02 4.14833993e-01 9.52116847e-01 -1.26090991e+00
-6.88307345e-01 -7.97540665e-01 -1.55786917e-01 3.28546166e-01
-8.96989822e-01 5.92681885e-01 -7.50110224e-02 -3.93535137e-01
-9.21882316e-02 -4.86013740e-01 -1.05392314e-01 -4.91789997e-01
-5.01603544e-01 -4.95880604e-01 1.01544905e+00 -9.93653893e-01
1.12327778e+00 -2.12685251e+00 -1.28536388e-01 3.18488985e-01
1.77547261e-01 5.17098784e-01 1.25540733e-01 8.06820095e-02
1.35448188e-01 -4.43453193e-02 2.78777599e-01 -5.53539753e-01
2.04935014e-01 -7.61092529e-02 3.02127212e-01 2.74290949e-01
3.46100405e-02 6.54424429e-01 -4.32508737e-01 -6.44960821e-01
5.48201680e-01 8.10113013e-01 -6.15877174e-02 1.79104239e-01
8.07720661e-01 5.49851120e-01 -5.58466077e-01 9.32785988e-01
8.18282127e-01 4.39249396e-01 -3.01766723e-01 -5.93276680e-01
-8.08947831e-02 -4.98093039e-01 -1.92569089e+00 1.15315938e+00
-1.73261732e-01 3.04638267e-01 3.13503325e-01 -9.74710047e-01
1.17560351e+00 7.15204418e-01 5.29617429e-01 -2.41997123e-01
5.53570807e-01 2.56062269e-01 -1.52327567e-02 -6.91524863e-01
4.70627636e-01 8.17579329e-02 -4.77618247e-01 8.31565708e-02
2.28517562e-01 6.59401000e-01 -1.30703986e-01 -2.74600565e-01
6.97758615e-01 -4.09042776e-01 2.15448961e-01 -1.12649187e-01
9.86067295e-01 -6.44876838e-01 7.99211144e-01 8.04207981e-01
-6.24213457e-01 4.63770896e-01 -1.69110909e-01 -6.69900835e-01
-7.21795022e-01 -7.89536893e-01 -2.76880652e-01 1.01071334e+00
1.88363399e-02 -1.39482901e-01 -7.87385762e-01 -3.70481819e-01
1.08098231e-01 -1.72225699e-01 -4.53462511e-01 -1.89677432e-01
-5.12929022e-01 -9.49376881e-01 1.05305159e+00 5.87956607e-01
1.01002014e+00 -9.21845734e-01 -1.24889255e-01 2.64961421e-01
-2.99194127e-01 -1.07910490e+00 -4.10135716e-01 8.21846798e-02
-9.58484486e-02 -9.11512256e-01 -7.99975097e-01 -9.64605629e-01
4.47400123e-01 1.99184328e-01 5.83354712e-01 -5.99951185e-02
-2.69681334e-01 2.75835484e-01 -2.79139549e-01 -7.16318071e-01
-9.14727822e-02 -7.57461339e-02 7.63832569e-01 4.05249685e-01
8.45718980e-01 -3.61070871e-01 -5.34046531e-01 3.17070484e-01
-6.19304180e-01 -5.51139235e-01 4.76343393e-01 8.33813727e-01
3.71967405e-02 5.34050822e-01 9.19081986e-01 1.13900468e-01
9.77853835e-01 -1.62713692e-01 -1.43443152e-01 3.89071494e-01
1.46057932e-02 -2.66615182e-01 6.02679849e-01 -5.02916276e-01
-1.04574847e+00 2.26747930e-01 -3.72323185e-01 -1.37754500e-01
-7.37552404e-01 2.68352330e-01 -7.37852871e-01 -2.72531539e-01
4.93700236e-01 3.95127028e-01 -1.21260829e-01 -6.20217204e-01
-8.73823315e-02 1.34044242e+00 7.58175015e-01 -5.55326760e-01
8.60032380e-01 1.24923021e-01 -2.41797462e-01 -7.60482669e-01
-4.05766129e-01 -6.55684769e-01 -1.07482910e+00 -4.45007741e-01
1.04380083e+00 -1.19880855e+00 -1.35624051e+00 1.06205475e+00
-1.18796933e+00 5.02812684e-01 1.09912552e-01 8.12519968e-01
1.82604730e-01 4.15759295e-01 -4.12386686e-01 -1.21270096e+00
-7.01375782e-01 -1.17148602e+00 1.12350488e+00 9.06060278e-01
-6.60818964e-02 -6.23547435e-01 -3.05649966e-01 4.35749799e-01
5.56908548e-01 1.42135546e-01 1.50475830e-01 -9.64835703e-01
1.09582826e-01 -6.04955912e-01 -3.16020846e-01 5.04769683e-01
4.24762994e-01 -1.10696733e-01 -1.26138973e+00 -3.22863966e-01
-8.61935541e-02 -3.54856819e-01 7.44136989e-01 4.87947673e-01
1.08368504e+00 -1.70569330e-01 -3.83009851e-01 6.04646385e-01
1.09818029e+00 3.79031390e-01 3.41341168e-01 8.51994902e-02
8.05836201e-01 6.37255251e-01 1.40437096e-01 6.75816178e-01
5.22997260e-01 6.70525074e-01 -1.10834301e-01 2.92971935e-02
9.72043648e-02 3.84080112e-02 3.92355740e-01 6.40401244e-01
-2.63255835e-01 -7.33889937e-02 -6.67690277e-01 4.56249267e-01
-1.75313795e+00 -1.34308922e+00 1.86110929e-01 2.16787505e+00
6.23302042e-01 -2.40520135e-01 4.38778102e-01 6.80893779e-01
1.28780115e+00 -2.84560591e-01 -2.64508843e-01 -1.02253564e-01
-4.55028832e-01 -1.22271828e-01 1.58749491e-01 2.97981203e-01
-1.51833355e+00 3.76390368e-01 6.42537642e+00 9.04185474e-01
-1.05537975e+00 -1.31266974e-02 3.91633302e-01 1.88017577e-01
1.76972061e-01 -5.20538509e-01 -1.48176003e+00 6.64706171e-01
8.62707555e-01 -4.50401381e-02 1.79844379e-01 6.43292964e-01
1.52022272e-01 7.06798881e-02 -8.82271826e-01 1.58022380e+00
2.08371937e-01 -6.77119851e-01 -1.64354369e-01 1.40799895e-01
3.77635419e-01 -6.36143446e-01 1.34718865e-01 4.02504146e-01
1.51231945e-01 -1.29693854e+00 2.55004048e-01 7.92886972e-01
6.14471257e-01 -8.33388984e-01 1.29784143e+00 2.64915466e-01
-1.52385378e+00 -4.81143683e-01 -2.87154526e-01 -1.28075600e-01
3.34145427e-01 4.57492560e-01 -5.48915684e-01 8.27722907e-01
9.00230765e-01 3.84744525e-01 -6.08343899e-01 1.27884817e+00
3.76862198e-01 2.80437201e-01 -5.70333540e-01 -3.94377932e-02
-1.91435784e-01 1.17439725e-01 4.34967220e-01 1.21055138e+00
6.04330182e-01 -6.10231906e-02 3.90428364e-01 4.38680828e-01
1.66356504e-01 4.99676578e-02 -6.06233060e-01 2.15765640e-01
7.68558025e-01 1.37610924e+00 -2.17144564e-01 -2.69533128e-01
-3.30356091e-01 8.77953887e-01 -2.36697599e-01 2.22518250e-01
-5.40762484e-01 -3.97778541e-01 6.82643592e-01 -2.48839319e-01
-8.28505978e-02 -1.78699307e-02 -2.25119412e-01 -9.99445617e-01
-1.14056058e-02 -9.64423895e-01 5.44891953e-01 -2.53478825e-01
-1.62384534e+00 6.79524064e-01 2.25817058e-02 -1.28292632e+00
-3.08949977e-01 -6.38804436e-01 -7.72880793e-01 1.20738697e+00
-1.14863956e+00 -1.42729354e+00 -6.48668706e-01 1.12537682e+00
2.38616854e-01 -7.19606042e-01 1.09087408e+00 7.07544744e-01
-8.93963695e-01 1.04590738e+00 1.02964543e-01 6.17554724e-01
7.67005801e-01 -9.59936738e-01 -1.39931992e-01 8.10604572e-01
-2.53007323e-01 7.28314757e-01 5.73570132e-01 -5.47731042e-01
-1.34220076e+00 -7.96245754e-01 8.97034824e-01 -2.85606384e-01
-1.38513774e-01 -1.15985833e-01 -5.54579616e-01 2.60794044e-01
4.06366512e-02 -1.15628485e-02 1.01772904e+00 2.54401326e-01
-1.49126574e-01 -5.60492277e-01 -1.54590857e+00 1.99067906e-01
7.23431945e-01 -5.44903040e-01 -5.41664004e-01 4.63275909e-02
4.35107350e-01 9.62077901e-02 -1.10540211e+00 5.99388957e-01
9.66975749e-01 -7.48482466e-01 1.25360823e+00 -2.38730431e-01
-3.77667546e-01 -5.14478385e-01 -3.90132964e-01 -1.18806982e+00
-3.45004976e-01 -3.74205917e-01 2.19054744e-01 1.65540850e+00
-1.09990172e-01 -7.25896657e-01 3.95494998e-01 7.92964518e-01
7.86463618e-02 -1.71709955e-01 -1.12266409e+00 -6.99433625e-01
-3.44260514e-01 -2.14644611e-01 1.06556928e+00 7.16520071e-01
-1.21395014e-01 1.86540887e-01 -7.36426055e-01 3.17630529e-01
8.49410176e-01 -3.85846645e-01 8.38795960e-01 -1.74615705e+00
4.83123623e-02 -3.92483413e-01 -8.41804445e-01 -2.95983225e-01
2.07591459e-01 -3.42828631e-01 -1.51335761e-01 -1.43072879e+00
3.85481864e-01 -2.31765077e-01 -5.99784613e-01 4.89505976e-01
-3.32851917e-01 3.07786763e-01 2.72101182e-02 4.83744256e-02
-4.93088931e-01 6.60002708e-01 8.20053935e-01 -2.88901865e-01
-1.17065385e-01 2.49230206e-01 -8.76515806e-01 6.72995687e-01
7.49894798e-01 3.73406038e-02 -1.07808979e-02 -2.20872551e-01
-3.63123953e-01 -6.53121248e-02 6.40062511e-01 -1.38037884e+00
4.64525282e-01 6.29839525e-02 1.12438023e+00 -6.36270285e-01
6.20850861e-01 -8.04851294e-01 2.12828919e-01 4.49223161e-01
-1.08342040e-02 1.15174139e-02 8.88106320e-03 1.35935277e-01
-5.62722266e-01 -1.92010596e-01 4.70826983e-01 -2.20095024e-01
-9.53996778e-01 3.48099917e-01 -1.99011639e-01 -6.34688258e-01
8.57751906e-01 -5.31241715e-01 -3.60296011e-01 -4.41770166e-01
-8.69968712e-01 1.96584746e-01 -1.09483331e-01 5.64422131e-01
7.95896530e-01 -1.68294477e+00 -9.75295365e-01 4.88993466e-01
-5.47822267e-02 -4.49693203e-01 6.05841279e-01 5.41585267e-01
2.49885302e-02 4.71061289e-01 -5.69590032e-01 -3.76356870e-01
-1.72407210e+00 1.78729370e-01 5.33141732e-01 1.12786680e-03
-7.52753764e-02 9.42470729e-01 -1.32946104e-01 -6.34067535e-01
4.90830749e-01 3.42889637e-01 -8.74612093e-01 3.96726698e-01
1.02863491e+00 3.42251509e-01 -1.68857928e-02 -1.23200488e+00
-5.78784704e-01 6.46454036e-01 9.41060558e-02 -2.30279360e-02
1.05222404e+00 -2.52126932e-01 -8.89271796e-02 9.91561934e-02
1.18569219e+00 -2.10264906e-01 -7.73932517e-01 -4.23862666e-01
-3.44484806e-01 -5.15346885e-01 4.25597392e-02 -5.46391785e-01
-8.26736331e-01 7.30062723e-01 1.18062258e+00 1.96336433e-01
1.07781577e+00 -3.24035883e-01 8.32746387e-01 4.29992825e-01
4.91597354e-01 -1.25179470e+00 -7.06064850e-02 2.62411296e-01
8.16926658e-01 -1.51085782e+00 -2.92530596e-01 2.88589392e-03
-3.32504839e-01 1.25251913e+00 8.33416462e-01 3.37363034e-01
8.63953412e-01 4.79344219e-01 7.40954205e-02 6.43176809e-02
-7.54284337e-02 -5.73697865e-01 5.03698051e-01 1.02316904e+00
2.80373275e-01 2.22603306e-01 -2.59009749e-01 1.30689073e+00
-3.73001188e-01 -1.13021463e-01 -2.84696855e-02 5.97246349e-01
-5.90619624e-01 -1.14794552e+00 -1.10568810e+00 5.83457589e-01
-5.53362310e-01 2.10725293e-01 -2.11504191e-01 3.40435088e-01
7.60957897e-01 1.37230098e+00 -1.14428192e-01 -9.35692966e-01
2.26301327e-01 2.48215124e-01 3.46992254e-01 -7.10380152e-02
-8.61029446e-01 -7.10401237e-02 -1.10613573e-02 -1.11764021e-01
-7.06223905e-01 -8.62942219e-01 -7.81819105e-01 -5.38806438e-01
-3.32085818e-01 3.16997096e-02 6.00375414e-01 1.16687477e+00
-8.79331201e-04 4.19965446e-01 7.28791654e-01 -9.99188423e-01
-5.12006938e-01 -1.22936738e+00 -6.71761155e-01 4.41323251e-01
2.65236050e-01 -6.59367979e-01 -2.74337739e-01 -3.65223512e-02] | [14.568195343017578, 1.021847128868103] |
7a55559d-846a-4c09-97e3-e773f8c1f049 | embracing-consistency-a-one-stage-approach | 2209.13306 | null | https://arxiv.org/abs/2209.13306v2 | https://arxiv.org/pdf/2209.13306v2.pdf | Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Grounding | Spatio-Temporal video grounding (STVG) focuses on retrieving the spatio-temporal tube of a specific object depicted by a free-form textual expression. Existing approaches mainly treat this complicated task as a parallel frame-grounding problem and thus suffer from two types of inconsistency drawbacks: feature alignment inconsistency and prediction inconsistency. In this paper, we present an end-to-end one-stage framework, termed Spatio-Temporal Consistency-Aware Transformer (STCAT), to alleviate these issues. Specially, we introduce a novel multi-modal template as the global objective to address this task, which explicitly constricts the grounding region and associates the predictions among all video frames. Moreover, to generate the above template under sufficient video-textual perception, an encoder-decoder architecture is proposed for effective global context modeling. Thanks to these critical designs, STCAT enjoys more consistent cross-modal feature alignment and tube prediction without reliance on any pre-trained object detectors. Extensive experiments show that our method outperforms previous state-of-the-arts with clear margins on two challenging video benchmarks (VidSTG and HC-STVG), illustrating the superiority of the proposed framework to better understanding the association between vision and natural language. Code is publicly available at https://github.com/jy0205/STCAT. | ['Yadong Mu', 'Zehuan Yuan', 'Yongzhi Li', 'Yang Jin'] | 2022-09-27 | null | null | null | null | ['video-grounding', 'spatio-temporal-video-grounding'] | ['computer-vision', 'computer-vision'] | [ 1.14627652e-01 -1.35233566e-01 -2.73190767e-01 -3.71584505e-01
-9.23722625e-01 -3.69442374e-01 6.56132698e-01 -3.84816796e-01
-1.22587174e-01 3.83256525e-01 3.06381106e-01 -1.90935239e-01
1.88530475e-01 -4.19292271e-01 -9.49018121e-01 -4.76150006e-01
3.44333231e-01 -4.92783934e-02 4.85552549e-01 -9.36901197e-02
1.88549072e-01 -9.83718857e-02 -1.48369741e+00 5.53353548e-01
6.94810212e-01 1.33721411e+00 6.64614677e-01 3.31499934e-01
-6.55917525e-02 9.43804860e-01 -7.93800596e-03 -5.93752325e-01
1.40117049e-01 -4.70010877e-01 -6.90734029e-01 4.56833780e-01
7.63518155e-01 -4.51388568e-01 -3.88185829e-01 1.03008008e+00
3.19736660e-01 -2.44769119e-02 3.31231296e-01 -1.35767508e+00
-5.92168808e-01 3.27996403e-01 -7.50261426e-01 2.17178971e-01
5.87614775e-01 3.67577225e-01 1.23469329e+00 -1.21292341e+00
7.10632205e-01 1.15050197e+00 4.81350005e-01 3.02484006e-01
-9.09330785e-01 -5.09633303e-01 5.61935246e-01 4.83998775e-01
-1.57077456e+00 -5.93297660e-01 8.40735674e-01 -5.41920245e-01
9.11003768e-01 2.46061981e-01 7.91641235e-01 1.17596114e+00
1.84702933e-01 9.59388494e-01 8.90897751e-01 -2.64490128e-01
5.85406944e-02 -2.08689466e-01 -2.42107242e-01 8.71827722e-01
-1.03925362e-01 -1.98202450e-02 -9.04701114e-01 1.16177022e-01
8.22979331e-01 -3.62975746e-02 -4.98865962e-01 -4.88991708e-01
-1.59190547e+00 4.25566137e-01 2.60535240e-01 3.92135948e-01
-3.96781951e-01 2.69867957e-01 4.75311428e-01 -4.22006920e-02
4.15147424e-01 -1.64139554e-01 -3.46392721e-01 -4.80844565e-02
-9.24585521e-01 2.28146985e-01 2.31616527e-01 1.38986778e+00
7.31760144e-01 -3.54751162e-02 -4.90375698e-01 6.08156741e-01
5.97509027e-01 3.62336159e-01 3.38784426e-01 -9.14134264e-01
9.18521762e-01 5.02030492e-01 2.06863821e-01 -1.19397116e+00
-1.77688852e-01 -5.55399895e-01 -7.23487735e-01 -2.53571242e-01
1.85649753e-01 1.98205203e-01 -6.78921700e-01 1.73235965e+00
5.76403081e-01 7.61539698e-01 -9.12363157e-02 1.18476188e+00
9.41477835e-01 7.77319372e-01 2.52737463e-01 -2.92539835e-01
1.59945047e+00 -1.24930179e+00 -8.45096529e-01 -4.38416153e-01
4.78196681e-01 -8.95858526e-01 1.23399818e+00 1.86582789e-01
-1.11101639e+00 -6.84155643e-01 -9.25770223e-01 -3.11903954e-01
1.80422030e-02 3.02996635e-01 3.70262444e-01 7.55600110e-02
-1.09007978e+00 6.37096837e-02 -8.13507676e-01 -3.47878128e-01
4.01117593e-01 5.07579632e-02 -3.34815741e-01 -6.43567294e-02
-1.01018250e+00 6.18696451e-01 3.97344321e-01 3.92275900e-01
-6.78982198e-01 -4.80141699e-01 -1.04451668e+00 -2.26527780e-01
6.63863778e-01 -1.04983425e+00 1.32681608e+00 -9.42202091e-01
-1.43269503e+00 8.93927872e-01 -6.08815551e-01 -2.41648749e-01
6.55533612e-01 -3.50095659e-01 -3.82618189e-01 2.77343363e-01
3.33670974e-01 6.48564696e-01 1.03069139e+00 -1.14072311e+00
-8.38560164e-01 -1.36656150e-01 9.61245894e-02 3.60945642e-01
-1.88711241e-01 1.28909692e-01 -1.13602769e+00 -8.64630520e-01
3.65478903e-01 -7.79327631e-01 4.36388329e-02 1.81075901e-01
-4.93758470e-01 -2.79719740e-01 8.34751725e-01 -6.03052855e-01
1.32509685e+00 -2.10761046e+00 2.78639346e-01 -2.14694187e-01
1.61715895e-01 1.08253099e-01 -7.92488232e-02 3.66810739e-01
5.23870848e-02 -2.06662878e-01 -1.89121678e-01 -5.48499346e-01
-2.31114291e-02 8.58132094e-02 -4.33402717e-01 5.11235237e-01
3.07237357e-01 1.09375703e+00 -8.96538615e-01 -8.52725565e-01
4.94064271e-01 5.44062316e-01 -5.89885592e-01 2.79940486e-01
-4.42244232e-01 6.53603673e-01 -6.65551960e-01 7.97859073e-01
5.78913689e-01 -4.72899854e-01 9.18061137e-02 -7.35778153e-01
-2.35505998e-01 1.63524330e-01 -1.04443157e+00 2.25739408e+00
-2.61586368e-01 5.10317385e-01 -3.16789895e-01 -8.20102632e-01
6.54106498e-01 3.41013908e-01 5.33879042e-01 -9.04112816e-01
2.56382942e-01 1.15629509e-01 -4.97954309e-01 -8.02696943e-01
5.62889576e-01 2.89328843e-01 3.33325453e-02 -3.57520841e-02
9.98366177e-02 1.14740290e-01 5.32512553e-02 2.19226122e-01
6.70665026e-01 6.68667018e-01 3.50773603e-01 -4.66156863e-02
7.19096422e-01 -5.76145723e-02 7.61704385e-01 3.93597871e-01
-3.65929633e-01 7.80867279e-01 3.01276207e-01 -3.32141638e-01
-9.85624611e-01 -9.23647404e-01 6.37190193e-02 9.65930998e-01
6.05252504e-01 -8.45847487e-01 -5.87615192e-01 -4.57651973e-01
-4.41182137e-01 5.46397626e-01 -4.47837085e-01 2.47132748e-01
-5.90540588e-01 -2.62684077e-01 2.63536125e-01 5.52397907e-01
8.22869897e-01 -8.25011313e-01 -7.28786767e-01 5.96716143e-02
-8.90642643e-01 -1.62295377e+00 -8.49765956e-01 -3.39422882e-01
-6.33460641e-01 -1.01342690e+00 -6.84885263e-01 -8.81570578e-01
4.64385331e-01 5.79000354e-01 1.11783063e+00 1.61516041e-01
-1.72383320e-02 3.93316060e-01 -4.42333221e-01 1.45434424e-01
-9.60491505e-03 -1.72355682e-01 -2.70428509e-01 3.72148544e-01
1.64959759e-01 -4.54740405e-01 -9.17767763e-01 5.17096698e-01
-8.31450284e-01 7.84453273e-01 4.51186419e-01 7.68880844e-01
9.52620685e-01 -2.04927400e-01 2.18990460e-01 -4.44124728e-01
3.97124216e-02 -3.45523089e-01 -6.62925005e-01 3.78480077e-01
-2.11693272e-01 -1.92071840e-01 4.04024869e-01 -2.09218666e-01
-1.10736585e+00 2.79320955e-01 -1.62408613e-02 -8.21645141e-01
3.74532901e-02 4.45825547e-01 -4.28006053e-01 2.34675810e-01
1.48766309e-01 5.17978847e-01 -3.35480899e-01 -2.78939873e-01
5.06703019e-01 4.02868092e-01 8.50323975e-01 -5.99673033e-01
7.60994077e-01 5.35907328e-01 -1.02156073e-01 -5.75088739e-01
-9.70862329e-01 -6.13785565e-01 -6.28905475e-01 -4.55204874e-01
1.15349066e+00 -1.30098855e+00 -6.43929601e-01 3.71465147e-01
-1.29854620e+00 -3.10500592e-01 1.73944175e-01 4.34912086e-01
-8.04041207e-01 6.04875028e-01 -4.63258266e-01 -6.24087632e-01
-2.90211231e-01 -1.22402430e+00 1.48964620e+00 -7.60208070e-02
-1.73210562e-03 -6.60314322e-01 -2.68271834e-01 4.54333991e-01
1.12217508e-01 2.34243020e-01 5.11972010e-01 4.13646698e-02
-1.07080281e+00 2.02978328e-01 -3.67773771e-01 1.98335331e-02
-8.63484107e-03 -4.28745989e-03 -1.03040528e+00 -2.19450265e-01
9.28564090e-03 -1.85877129e-01 7.67735898e-01 3.23211879e-01
1.21595919e+00 -3.90167981e-01 -1.90008983e-01 8.19983304e-01
1.55222952e+00 1.13398746e-01 6.99714780e-01 5.13822436e-01
1.03121626e+00 5.30051589e-01 9.94958878e-01 5.07173538e-01
8.18258286e-01 1.15160775e+00 6.15855873e-01 -3.89725901e-02
-1.61553249e-01 -4.65389639e-01 5.02988994e-01 7.95896530e-01
-8.38438720e-02 -4.28734124e-01 -8.30598295e-01 4.69892234e-01
-2.29619408e+00 -1.06966960e+00 -2.11862713e-01 1.83757794e+00
6.17372751e-01 -7.50604346e-02 1.00363612e-01 -1.73875000e-02
8.09552729e-01 4.35474724e-01 -4.42334414e-01 2.36882508e-01
-2.55737990e-01 -3.81695837e-01 1.44910008e-01 3.73595893e-01
-1.27343082e+00 1.11607039e+00 4.73629999e+00 9.47224975e-01
-1.09903932e+00 2.27123976e-01 7.65773058e-01 4.64206301e-02
-2.17289448e-01 1.40485644e-01 -6.60733461e-01 4.87155110e-01
4.51177925e-01 4.16641794e-02 1.91199228e-01 6.26350224e-01
4.90328550e-01 -1.15625776e-01 -1.14146674e+00 1.31525564e+00
9.21360552e-02 -1.42433035e+00 1.20525591e-01 -1.34523734e-01
5.86838663e-01 -4.46494296e-02 1.43044069e-01 -7.34912306e-02
-3.14368606e-01 -5.72934270e-01 1.47508264e+00 4.85503763e-01
9.15938258e-01 -4.23857152e-01 4.27778661e-01 5.06232157e-02
-1.76919830e+00 7.07454830e-02 -2.14836910e-01 2.18249053e-01
3.79452884e-01 3.59139979e-01 -4.94876564e-01 8.64734709e-01
7.45295346e-01 1.10476184e+00 -5.97408473e-01 1.02273929e+00
-2.31173277e-01 4.94611830e-01 -1.45273954e-01 4.11323041e-01
3.57132822e-01 -1.95902213e-02 5.73777139e-01 1.17571449e+00
4.62466449e-01 1.80135399e-01 3.28763664e-01 9.32435989e-01
2.07642198e-01 1.48315117e-01 -4.13939506e-01 2.04790458e-01
5.21050990e-01 1.11792374e+00 -6.95633233e-01 -3.39312077e-01
-8.38425756e-01 1.10633695e+00 2.15562761e-01 4.39438909e-01
-1.27387404e+00 2.14675114e-01 4.72618669e-01 1.52936026e-01
6.22002780e-01 -2.71720260e-01 -9.73474905e-02 -1.50301373e+00
4.55587953e-01 -7.73625493e-01 3.97787154e-01 -1.12868178e+00
-1.12352633e+00 7.86647558e-01 -6.82401750e-03 -1.84496439e+00
-2.71151364e-01 -3.86371493e-01 -3.36620659e-01 4.93348122e-01
-1.58876181e+00 -1.61308920e+00 -5.42975485e-01 9.97035801e-01
7.86892653e-01 1.46148488e-01 3.33099484e-01 4.46070701e-01
-6.11282825e-01 5.54827511e-01 -3.65077108e-01 9.39182378e-03
6.67876542e-01 -7.79221892e-01 3.28827918e-01 1.17137468e+00
2.21227124e-01 4.36302304e-01 7.24859416e-01 -6.19231164e-01
-1.67400753e+00 -1.46954370e+00 9.18148577e-01 -3.52504492e-01
6.54123783e-01 -3.23363364e-01 -7.64434040e-01 7.80195653e-01
3.22192580e-01 2.62078583e-01 2.14070931e-01 -3.85701001e-01
-4.82719660e-01 -7.85593465e-02 -6.18512392e-01 7.55183160e-01
1.44830501e+00 -7.13515937e-01 -4.30812210e-01 2.85554290e-01
9.18258309e-01 -7.59172678e-01 -6.21833622e-01 4.09160554e-01
5.60157001e-01 -1.24948442e+00 1.00935316e+00 -1.26367912e-01
7.18013585e-01 -7.51629531e-01 -5.15987694e-01 -6.61678910e-01
-2.80555397e-01 -7.66646445e-01 -1.62405163e-01 1.35089707e+00
1.16993256e-01 -3.23113531e-01 5.15608907e-01 4.08551961e-01
-3.71390104e-01 -9.38863337e-01 -1.02664137e+00 -7.25067914e-01
-5.28368950e-01 -7.81913996e-01 5.05235434e-01 9.21458602e-01
-1.09388992e-01 3.68472338e-01 -6.60800099e-01 3.08519334e-01
5.00411510e-01 3.09090406e-01 8.77201676e-01 -5.41821003e-01
-3.39141399e-01 -2.16032669e-01 -5.85080206e-01 -1.41803408e+00
-2.66637380e-05 -6.53928816e-01 9.89347026e-02 -1.45531702e+00
2.20949024e-01 -1.15052328e-01 -2.60176122e-01 4.70131844e-01
-2.14222461e-01 3.69382143e-01 4.50640261e-01 4.00397629e-01
-1.14021552e+00 7.65888035e-01 1.39477837e+00 -1.09569274e-01
-2.27306858e-02 -3.52661639e-01 -5.01273215e-01 7.30653644e-01
4.73959774e-01 -2.15071887e-01 -6.41032815e-01 -7.37947226e-01
1.65985703e-01 3.33670139e-01 8.12308609e-01 -9.63836014e-01
3.26496392e-01 -2.60907471e-01 1.61181524e-01 -9.42998528e-01
5.25308609e-01 -8.79472733e-01 3.54401112e-01 9.32512507e-02
-1.82670131e-01 3.33027095e-01 1.20356604e-01 7.97391534e-01
-3.96854341e-01 2.69192696e-01 4.77465987e-01 1.16339087e-01
-1.14603114e+00 5.72383344e-01 -6.78400770e-02 8.30086842e-02
1.04535604e+00 -2.49147266e-01 -4.42176729e-01 -2.89536327e-01
-4.69145328e-01 2.05321148e-01 4.52235162e-01 5.17721772e-01
8.44511807e-01 -1.45525324e+00 -6.00287139e-01 1.44851310e-02
4.76148099e-01 1.01344667e-01 5.93310833e-01 1.16277575e+00
-3.74041200e-01 5.22058725e-01 3.10156457e-02 -1.10787797e+00
-1.19281054e+00 6.11054718e-01 3.22036684e-01 -1.71269197e-02
-9.82635498e-01 8.42201710e-01 7.05995858e-01 2.65326887e-01
2.32720345e-01 -6.36538267e-01 -2.02219207e-02 -1.66267559e-01
4.74204540e-01 -1.46744624e-01 -3.71499509e-02 -1.09638405e+00
-4.03660685e-01 8.55626822e-01 6.54363185e-02 -5.78647591e-02
1.04533076e+00 -6.33023858e-01 5.51689342e-02 3.36727202e-01
1.16974425e+00 -2.28389874e-01 -1.60009921e+00 -5.19585550e-01
-1.00258686e-01 -7.61187315e-01 8.25002138e-03 -4.30673748e-01
-1.16835093e+00 6.09355867e-01 4.45416570e-01 -2.26996019e-01
1.29923952e+00 1.29208580e-01 8.75980020e-01 2.01974049e-01
5.17519832e-01 -7.70758092e-01 2.75035799e-01 3.29496831e-01
1.11706412e+00 -1.21167648e+00 -2.28757709e-02 -7.87368655e-01
-7.71330535e-01 8.93772781e-01 7.52332211e-01 5.55396006e-02
4.29327101e-01 -7.47545660e-02 -7.66992718e-02 -2.17513219e-01
-9.44152236e-01 -3.02841574e-01 5.07082701e-01 4.19648200e-01
4.16023195e-01 -1.48214877e-01 -1.17554143e-01 4.97359037e-01
-9.11339670e-02 1.50694504e-01 5.43530658e-02 8.23425412e-01
-1.66501582e-01 -8.35730016e-01 -2.62004316e-01 8.98041949e-03
-3.61103803e-01 -2.64046550e-01 -6.09827898e-02 6.93426669e-01
2.99075454e-01 1.04220033e+00 -4.96172719e-02 -4.97629493e-01
2.17449889e-01 -2.13878661e-01 5.20768166e-01 -3.42959046e-01
-2.95654744e-01 6.18716955e-01 1.27313346e-01 -1.09548283e+00
-8.98353457e-01 -6.47788525e-01 -1.07756472e+00 -2.54000407e-02
-3.64803046e-01 -2.75681317e-01 2.93813020e-01 9.87777472e-01
4.62737828e-01 5.54971695e-01 4.06259000e-01 -9.51932073e-01
-1.23352036e-01 -6.62580132e-01 -2.43023351e-01 4.08689409e-01
3.79773259e-01 -8.47120345e-01 8.96578748e-03 4.88039166e-01] | [9.714398384094238, 0.6525529623031616] |
e21c4099-811d-4605-baff-096149654743 | learning-weakly-supervised-audio-visual | 2305.18797 | null | https://arxiv.org/abs/2305.18797v2 | https://arxiv.org/pdf/2305.18797v2.pdf | Learning Weakly Supervised Audio-Visual Violence Detection in Hyperbolic Space | In recent years, the task of weakly supervised audio-visual violence detection has gained considerable attention. The goal of this task is to identify violent segments within multimodal data based on video-level labels. Despite advances in this field, traditional Euclidean neural networks, which have been used in prior research, encounter difficulties in capturing highly discriminative representations due to limitations of the feature space. To overcome this, we propose HyperVD, a novel framework that learns snippet embeddings in hyperbolic space to improve model discrimination. Our framework comprises a detour fusion module for multimodal fusion, effectively alleviating modality inconsistency between audio and visual signals. Additionally, we contribute two branches of fully hyperbolic graph convolutional networks that excavate feature similarities and temporal relationships among snippets in hyperbolic space. By learning snippet representations in this space, the framework effectively learns semantic discrepancies between violent and normal events. Extensive experiments on the XD-Violence benchmark demonstrate that our method outperforms state-of-the-art methods by a sizable margin. | ['Zizhao Wu', 'Yigang Wang', 'Keyang Yu', 'Xiao Zhou', 'Yikai Luo', 'Hao Wen', 'Xiaogang Peng'] | 2023-05-30 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 2.81506896e-01 -1.19144998e-01 -1.49842510e-02 -1.75614968e-01
-1.12388945e+00 -4.63165671e-01 6.33607686e-01 2.64875740e-01
-4.64613646e-01 1.97810933e-01 5.62964499e-01 1.36341527e-01
-3.25107694e-01 -3.56523573e-01 -2.96723396e-01 -6.99341893e-01
-2.59761900e-01 -4.45514545e-02 -1.22941341e-02 -2.02397168e-01
1.03529200e-01 3.57783318e-01 -1.38416278e+00 5.17723620e-01
4.08012152e-01 9.92625952e-01 -5.79391539e-01 6.26447141e-01
2.48984143e-01 9.87876713e-01 -3.22483718e-01 -6.36953056e-01
9.29350555e-02 -4.70899522e-01 -7.80125141e-01 -3.88165824e-02
7.44808614e-01 -4.07464296e-01 -1.18550420e+00 8.56527328e-01
7.42204070e-01 1.91993445e-01 8.52428198e-01 -1.83984160e+00
-5.10041475e-01 6.03862047e-01 -7.30879247e-01 4.74073946e-01
6.85049236e-01 -1.44138190e-04 1.13755381e+00 -1.06260622e+00
3.92704159e-01 1.38101292e+00 1.05786061e+00 4.24395621e-01
-1.21364498e+00 -7.83832133e-01 2.03181015e-04 7.42520630e-01
-1.83616257e+00 -5.27730584e-01 1.07364523e+00 -5.98058522e-01
8.71594071e-01 3.11974436e-01 6.15808487e-01 1.70850611e+00
-1.08678930e-01 9.29607034e-01 5.57721674e-01 -9.43145156e-03
-1.29581645e-01 -3.14155072e-01 -1.69718489e-01 7.87959337e-01
-2.80979186e-01 -5.69028631e-02 -1.09722424e+00 -3.75702202e-01
5.78220546e-01 2.54208326e-01 -3.23187739e-01 -4.75683749e-01
-1.06217909e+00 1.08143294e+00 3.28769654e-01 2.57274359e-01
-1.25809968e-01 2.88459331e-01 7.79108644e-01 5.19889891e-02
4.92990106e-01 2.88671285e-01 1.95574090e-01 -5.25608659e-01
-1.00285637e+00 3.28264326e-01 3.19878370e-01 4.45224524e-01
3.25180054e-01 -5.38784266e-02 -4.01248127e-01 8.96626711e-01
9.32254568e-02 3.05291772e-01 4.00141105e-02 -1.10038352e+00
7.12874889e-01 5.65328121e-01 -5.18393815e-01 -1.57670486e+00
-6.94199204e-01 -1.54782802e-01 -9.30672050e-01 -7.36775473e-02
3.32804084e-01 -7.91945383e-02 -5.91990769e-01 1.92015016e+00
2.69730061e-01 7.07493842e-01 -7.69201368e-02 9.92811203e-01
9.95610237e-01 4.38461453e-01 -4.20542434e-02 -9.13372487e-02
1.02114761e+00 -6.32868409e-01 -6.52496576e-01 1.89966083e-01
5.47671318e-01 -4.16587174e-01 8.19045484e-01 3.89360934e-01
-1.14342046e+00 -2.28473783e-01 -7.91033924e-01 -9.37943980e-02
-2.13668257e-01 -2.33949453e-01 4.47652936e-01 3.50813270e-01
-8.05275321e-01 5.45502901e-01 -6.40436709e-01 -3.52591097e-01
6.64540648e-01 9.35572237e-02 -7.10790098e-01 5.62000424e-02
-1.31806600e+00 4.83328104e-01 2.61658669e-01 1.74242675e-01
-8.92141759e-01 -7.53225029e-01 -1.21469533e+00 1.15102991e-01
4.96179640e-01 -4.72921163e-01 1.05617034e+00 -7.17807651e-01
-9.68179405e-01 7.06678510e-01 2.15447083e-01 -3.01889867e-01
6.21052742e-01 -6.77662864e-02 -4.68698055e-01 8.29898775e-01
2.13242192e-02 4.63436127e-01 1.22888744e+00 -1.11216319e+00
-5.85789502e-01 -5.43692231e-01 -9.47899185e-04 3.27404976e-01
-7.34548867e-01 1.78199351e-01 -4.61939961e-01 -8.55949640e-01
7.94634521e-02 -8.90200913e-01 2.00443417e-01 1.72609404e-01
-4.41399872e-01 -4.52396542e-01 1.11119127e+00 -6.69701457e-01
1.48181427e+00 -2.33654523e+00 5.18754780e-01 2.63603866e-01
8.05297792e-01 1.28448814e-01 -1.77241340e-01 6.97538972e-01
-1.68722436e-01 -6.70034140e-02 -3.15595329e-01 -7.05685616e-01
2.30504513e-01 -4.83879298e-02 -3.63087386e-01 8.25618625e-01
3.49637926e-01 6.17458403e-01 -1.01111519e+00 -7.27630794e-01
1.35659888e-01 7.22030997e-01 -7.17595339e-01 2.28583500e-01
4.48313117e-01 3.20805877e-01 -4.36146930e-02 6.56800807e-01
5.20046055e-01 1.27187446e-01 -1.16456747e-01 -1.97103739e-01
2.40080446e-01 -1.73959672e-01 -1.03088701e+00 2.00672626e+00
-5.57131357e-02 1.02526700e+00 -8.41694549e-02 -1.39312708e+00
6.31635785e-01 5.60918808e-01 9.48688030e-01 -6.15129471e-01
1.98656783e-01 -2.26650819e-01 -3.36038500e-01 -9.80598807e-01
5.83341181e-01 -3.63555886e-02 -3.68849277e-01 1.05627552e-01
2.32384369e-01 9.35121775e-02 -3.72662134e-02 5.59163451e-01
1.42735875e+00 -2.31804371e-01 3.96781974e-02 5.16449213e-01
3.52073759e-01 -3.49120051e-01 6.26991868e-01 5.94396234e-01
-5.67108393e-01 1.04151309e+00 9.09366429e-01 -2.56252736e-01
-7.70231485e-01 -1.49141395e+00 -1.28158867e-01 1.14693201e+00
1.07032284e-01 -8.83128524e-01 -5.64300835e-01 -9.43730891e-01
3.64554673e-02 2.13180333e-01 -8.58006239e-01 -5.55688262e-01
-5.17154098e-01 -3.48227054e-01 1.20249379e+00 7.53186643e-01
2.78265744e-01 -7.61651993e-01 -4.03787225e-01 -2.40642596e-02
-4.61112291e-01 -1.40609252e+00 -3.54762793e-01 -1.16044775e-01
-2.96269506e-01 -1.35393333e+00 -5.30518353e-01 -5.43006778e-01
2.53538102e-01 4.22811389e-01 8.94434094e-01 -9.95585248e-02
-6.17482483e-01 9.51331913e-01 -5.83086491e-01 -4.47470583e-02
-6.57053962e-02 -1.04260512e-01 1.05431207e-01 2.40894124e-01
3.90570760e-01 -6.19140387e-01 -5.97341478e-01 9.31534916e-02
-1.05419040e+00 -1.07692644e-01 3.68308537e-02 7.58808553e-01
1.86242163e-01 -1.87714994e-02 6.01620793e-01 -2.79458255e-01
7.25267708e-01 -8.40067327e-01 1.34517411e-02 3.04613970e-02
-6.19019605e-02 -2.63668865e-01 5.20586312e-01 -4.26988423e-01
-6.26205146e-01 -1.05166785e-01 -2.09985431e-02 -9.66817319e-01
-1.14779651e-01 5.21524191e-01 8.59031901e-02 1.37536034e-01
5.80002189e-01 -1.35498000e-02 -1.12552173e-01 -2.65012115e-01
4.82672930e-01 7.06826448e-01 9.00200427e-01 -2.75768936e-01
6.95745111e-01 6.19760692e-01 1.29919976e-01 -9.92516100e-01
-6.51649117e-01 -6.61079943e-01 -6.56935692e-01 -5.48714459e-01
1.16369128e+00 -9.25634027e-01 -9.56451178e-01 3.37691963e-01
-1.08582854e+00 1.25305131e-01 -1.22208416e-01 6.43463612e-01
-6.64403021e-01 6.49928629e-01 -6.72015548e-01 -9.67879772e-01
6.05735779e-02 -1.02290845e+00 1.37282813e+00 -8.59237164e-02
-3.58841091e-01 -9.38763440e-01 3.34444642e-01 5.72358906e-01
-1.59289822e-01 6.65410340e-01 8.36208522e-01 -5.60513914e-01
-3.62097286e-02 -2.00681031e-01 -3.89901876e-01 1.93405628e-01
-1.59929365e-01 1.28947392e-01 -1.01935935e+00 -3.38060737e-01
-4.50294763e-01 -6.51337981e-01 1.28005254e+00 1.39765009e-01
1.40442872e+00 -2.27091342e-01 -1.56387284e-01 7.67513931e-01
8.76124620e-01 -2.18488976e-01 3.18236470e-01 9.36506540e-02
9.55112636e-01 6.52567208e-01 2.55178273e-01 9.40666795e-01
5.82250237e-01 9.06281650e-01 8.90579998e-01 -2.11229876e-01
-1.81364715e-01 -4.39737618e-01 3.72778416e-01 5.18085182e-01
5.95987998e-02 -1.32267937e-01 -1.19280159e+00 6.86168551e-01
-2.16346359e+00 -1.42318547e+00 2.07084715e-02 1.77332878e+00
6.10088289e-01 -1.90014154e-01 5.97257495e-01 6.82949483e-01
6.02677405e-01 5.04834890e-01 -2.33523563e-01 -3.34049255e-01
-1.08313814e-01 -1.29196033e-01 1.84345618e-02 1.81748688e-01
-1.45387161e+00 7.88753629e-01 5.85064840e+00 9.87061024e-01
-1.01424551e+00 1.78474896e-02 1.03854321e-01 -4.63598490e-01
-7.17561692e-02 -3.78290981e-01 -2.65924931e-01 3.81485969e-01
6.64788961e-01 1.72138274e-01 4.38197970e-01 3.92608106e-01
3.10608625e-01 2.59392738e-01 -1.30910027e+00 1.67155743e+00
6.73743129e-01 -1.18153358e+00 -6.73439205e-02 -1.40446648e-01
5.22146046e-01 -2.30669960e-01 4.80253249e-01 3.90644550e-01
-1.71867847e-01 -1.32263303e+00 7.06998765e-01 5.50085545e-01
8.00111055e-01 -1.08888292e+00 4.83950675e-01 9.32806954e-02
-1.42786157e+00 -4.52868849e-01 -1.60908267e-01 4.78316881e-02
1.17755145e-01 1.93321750e-01 -7.87711859e-01 5.68727612e-01
8.94239068e-01 1.04601169e+00 -6.61546588e-01 9.98769641e-01
-3.46244164e-02 5.57454705e-01 4.53356840e-02 5.53957164e-01
4.83663529e-01 3.84414941e-02 7.51055896e-01 1.43984473e+00
2.19331130e-01 7.22634187e-03 2.81274319e-01 6.74402058e-01
-2.37666398e-01 -3.39208394e-02 -1.01882756e+00 -1.78491443e-01
3.41376424e-01 1.19306374e+00 -3.10046285e-01 1.07399613e-01
-5.38210452e-01 9.99408960e-01 5.17715752e-01 4.60005134e-01
-1.16286945e+00 -4.16923672e-01 8.59942496e-01 -3.36897075e-02
4.84733954e-02 -3.31974864e-01 -4.36336920e-03 -1.04114997e+00
-1.99134164e-02 -8.38836968e-01 7.71541417e-01 -5.76721132e-01
-1.31588697e+00 5.17815530e-01 1.70557082e-01 -1.38076138e+00
-3.80699962e-01 -2.40645230e-01 -4.90419149e-01 1.12042286e-01
-1.05620527e+00 -1.25511444e+00 -4.61224735e-01 1.15018797e+00
4.21047688e-01 -3.30048591e-01 6.89754665e-01 4.70757335e-01
-7.02066720e-01 8.99407089e-01 1.27230152e-01 5.11589825e-01
7.51381457e-01 -1.00425720e+00 -1.11016795e-01 7.05470562e-01
4.99927104e-01 5.24486452e-02 5.64207613e-01 -3.59484166e-01
-1.29428339e+00 -1.12181807e+00 6.95801020e-01 -4.53392833e-01
8.74977291e-01 -5.83822191e-01 -8.33310544e-01 4.78511781e-01
9.95537415e-02 2.28455946e-01 1.06644106e+00 1.56810522e-01
-1.00148666e+00 -2.94093750e-02 -9.38210964e-01 7.27737129e-01
1.36361933e+00 -1.11558795e+00 -6.32771850e-01 2.07861185e-01
4.65114921e-01 -2.27445126e-01 -7.66773462e-01 4.18576121e-01
5.93939960e-01 -1.09988642e+00 1.08529425e+00 -8.64763021e-01
6.35369182e-01 5.72764240e-02 -3.01257759e-01 -1.30993891e+00
-2.05914065e-01 -7.33300388e-01 -3.47535282e-01 1.30383575e+00
-8.22543725e-02 -4.89672609e-02 6.49830520e-01 2.20081061e-01
-1.37258932e-01 -6.54829741e-01 -1.44829345e+00 -7.47382641e-01
3.13285515e-02 -8.25118244e-01 2.85544097e-01 1.21348155e+00
6.35322630e-01 4.48148519e-01 -6.99566603e-01 1.44405708e-01
6.77517056e-01 -2.49220133e-01 6.01239145e-01 -1.25951540e+00
-3.75871360e-02 -6.45155191e-01 -1.03680670e+00 -5.68336964e-01
6.17817700e-01 -1.14885032e+00 -2.03684554e-01 -1.30821681e+00
5.43431997e-01 5.32522500e-02 -3.70108962e-01 5.74145079e-01
-1.17866658e-01 7.47978806e-01 4.03797328e-01 -7.66807236e-04
-1.01497352e+00 8.24533165e-01 6.91174626e-01 -3.96906286e-01
-4.68788631e-02 -4.20772702e-01 -4.72031623e-01 8.60379219e-01
5.50348163e-01 -4.59369779e-01 -4.53375608e-01 -5.68795741e-01
2.35872701e-01 2.72898436e-01 8.55612755e-01 -1.04825020e+00
1.72529131e-01 5.58775552e-02 1.94971487e-01 -5.00244379e-01
8.42730522e-01 -9.29048002e-01 -1.63970649e-01 6.86413646e-02
-6.52939141e-01 9.33487490e-02 2.22878181e-03 7.24005818e-01
-4.58585113e-01 2.08897933e-01 5.65487027e-01 4.93305057e-01
-5.13671577e-01 5.24524212e-01 -4.31657434e-01 4.17542785e-01
1.07813001e+00 -1.63111791e-01 -2.26010427e-01 -6.62111163e-01
-7.53131986e-01 1.59779772e-01 1.35833338e-01 7.73252428e-01
1.03590608e+00 -1.82051802e+00 -8.90704930e-01 3.11788917e-02
4.73421901e-01 -4.71768290e-01 5.66346169e-01 1.16475129e+00
-8.53681639e-02 1.05864741e-01 -4.38728854e-02 -8.19504380e-01
-1.65037799e+00 4.15899336e-01 1.35194227e-01 4.70989905e-02
-8.19127142e-01 8.64995778e-01 1.93455040e-01 -1.11677654e-01
6.37864232e-01 7.73940235e-02 -2.90225238e-01 3.88361156e-01
5.11174560e-01 7.17259467e-01 -1.12076618e-01 -1.20583034e+00
-6.00621521e-01 5.48051000e-01 1.83692098e-01 -2.43599698e-01
1.34700739e+00 -8.54707882e-02 2.53229946e-01 3.84643853e-01
1.71188116e+00 -1.48470908e-01 -1.28362429e+00 -3.18001479e-01
-9.13403183e-02 -7.15225399e-01 1.68775126e-01 -3.77967566e-01
-1.05967391e+00 1.20542669e+00 6.12588286e-01 2.80213803e-01
1.20779824e+00 2.61830688e-01 7.95467436e-01 1.75131813e-01
-1.88451678e-01 -1.22002733e+00 6.27932906e-01 4.87727642e-01
1.04253757e+00 -1.28459299e+00 -3.25692743e-01 -1.80603325e-01
-7.98947155e-01 1.19330025e+00 3.52297634e-01 -6.34367168e-02
3.38539422e-01 5.29993363e-02 -1.27767786e-01 -3.83258551e-01
-5.78164101e-01 -2.69118875e-01 4.71436411e-01 8.51074517e-01
2.28934735e-01 -1.97751269e-01 -9.23754424e-02 7.38527417e-01
5.14085451e-03 -4.18735623e-01 3.17179412e-01 7.93070555e-01
-1.84720054e-01 -5.83170950e-01 -5.06961226e-01 1.16720162e-01
-6.39731705e-01 3.73288840e-02 -6.19123101e-01 7.80887723e-01
1.67636260e-01 1.28887665e+00 1.85782030e-01 -7.97025502e-01
4.42490667e-01 -1.79555163e-01 5.21039009e-01 -2.28595927e-01
-5.92551291e-01 1.04405724e-01 4.50508855e-02 -9.20708060e-01
-2.27491871e-01 -7.09693372e-01 -1.25354552e+00 -3.66645128e-01
1.17731377e-01 1.22133829e-01 3.17499936e-01 8.43389332e-01
1.82645708e-01 4.87284154e-01 8.01327705e-01 -9.27641451e-01
-5.88196099e-01 -5.78458488e-01 -7.99771249e-01 8.21953952e-01
5.77435315e-01 -9.26424682e-01 -3.41044903e-01 -1.70883462e-01] | [13.526847839355469, 4.668430328369141] |
c86133b8-deb6-4fde-808d-95e64d9307b5 | approximate-bayes-optimal-pseudo-label | 2302.08883 | null | https://arxiv.org/abs/2302.08883v5 | https://arxiv.org/pdf/2302.08883v5.pdf | Approximately Bayes-Optimal Pseudo Label Selection | Semi-supervised learning by self-training heavily relies on pseudo-label selection (PLS). The selection often depends on the initial model fit on labeled data. Early overfitting might thus be propagated to the final model by selecting instances with overconfident but erroneous predictions, often referred to as confirmation bias. This paper introduces BPLS, a Bayesian framework for PLS that aims to mitigate this issue. At its core lies a criterion for selecting instances to label: an analytical approximation of the posterior predictive of pseudo-samples. We derive this selection criterion by proving Bayes optimality of the posterior predictive of pseudo-samples. We further overcome computational hurdles by approximating the criterion analytically. Its relation to the marginal likelihood allows us to come up with an approximation based on Laplace's method and the Gaussian integral. We empirically assess BPLS for parametric generalized linear and non-parametric generalized additive models on simulated and real-world data. When faced with high-dimensional data prone to overfitting, BPLS outperforms traditional PLS methods. | ['Thomas Augustin', 'Thomas Nagler', 'Emilio Dorigatti', 'Jann Goschenhofer', 'Julian Rodemann'] | 2023-02-17 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 3.95433724e-01 5.72943449e-01 -1.81565434e-01 -6.90469623e-01
-9.17638898e-01 -3.80919993e-01 3.37766349e-01 2.26737037e-01
-3.54230016e-01 1.30766201e+00 -4.10868168e-01 -2.28754401e-01
-3.53880942e-01 -4.83848661e-01 -6.71929002e-01 -7.76176989e-01
2.73481179e-02 7.47729897e-01 9.76479650e-02 5.50434589e-01
3.12477529e-01 3.75274777e-01 -1.38129067e+00 9.44834426e-02
1.13882756e+00 7.17287540e-01 4.85126413e-02 3.41165960e-01
-1.52357012e-01 5.56879044e-01 -4.40406889e-01 -3.90880316e-01
2.17297170e-02 -5.34679055e-01 -5.50608516e-01 2.35566333e-01
1.08445354e-01 -1.34745449e-01 5.88969946e-01 1.24678266e+00
3.18014562e-01 -1.49353161e-01 1.13539672e+00 -1.58598268e+00
-3.64943683e-01 7.81337976e-01 -3.74687105e-01 -3.76659781e-01
2.81654358e-01 -3.05700690e-01 7.16621459e-01 -9.30559039e-01
2.90915281e-01 1.23969388e+00 1.27449393e+00 2.72268564e-01
-1.81634152e+00 -6.79364979e-01 -2.38971755e-01 -1.05627358e-01
-1.58524013e+00 -3.75327617e-01 6.98073447e-01 -6.17075384e-01
1.22786149e-01 1.45528555e-01 3.26576531e-01 1.22120702e+00
1.12192519e-01 7.78044760e-01 1.64427614e+00 -6.40117705e-01
7.34114110e-01 9.60273087e-01 3.03558975e-01 4.15711194e-01
4.60769504e-01 3.19513887e-01 -4.19010431e-01 -6.67291105e-01
5.25973141e-01 -3.56633812e-01 6.39561005e-03 -6.41000628e-01
-5.31556249e-01 9.29928958e-01 -3.14398736e-01 -5.72523735e-02
-4.84674871e-01 -2.51519769e-01 5.76001443e-02 7.14556649e-02
7.46789455e-01 3.24685574e-01 -6.82201624e-01 1.20883688e-01
-1.25332987e+00 1.00492693e-01 1.01915145e+00 9.60466862e-01
7.25051522e-01 -5.60735725e-02 -7.16074556e-02 1.00857735e+00
7.60993719e-01 4.78130430e-01 1.58150256e-01 -1.04812348e+00
-1.83202103e-01 3.73210013e-01 5.52545130e-01 -5.41808248e-01
-4.07695919e-01 -6.65636718e-01 -4.24407631e-01 2.83280253e-01
8.40893924e-01 -1.92452267e-01 -4.91112947e-01 1.80435145e+00
4.63032603e-01 1.63980089e-02 3.55093856e-03 4.78722364e-01
1.26939803e-01 3.83198828e-01 4.53432620e-01 -1.03343248e+00
8.18293929e-01 -4.69686508e-01 -7.67946899e-01 -2.24942546e-02
5.95954895e-01 -6.37631714e-01 9.96232927e-01 8.13631415e-01
-8.63335788e-01 -4.80385810e-01 -9.37840164e-01 4.84158605e-01
-4.43938747e-02 3.00673932e-01 3.85970056e-01 9.76025999e-01
-7.39606082e-01 1.11194956e+00 -5.24893939e-01 -3.15446526e-01
3.37370753e-01 4.36370164e-01 -1.56378686e-01 2.24467292e-01
-9.06284094e-01 7.09037125e-01 4.79536682e-01 -1.10450752e-01
-4.14493650e-01 -9.23112810e-01 -4.95151669e-01 -1.79670408e-01
2.26513803e-01 -3.30289453e-01 1.32407784e+00 -1.04866946e+00
-1.59392357e+00 6.76968277e-01 -7.09146410e-02 -4.60167468e-01
9.69778180e-01 -1.28027603e-01 -4.57566887e-01 -2.07890347e-02
1.87486157e-01 3.40130150e-01 1.01204491e+00 -1.54897499e+00
-3.68393630e-01 -4.10873890e-01 -7.06127644e-01 -1.25767514e-01
-6.94937678e-03 -1.69065088e-01 2.16915354e-01 -3.69175255e-01
4.39784735e-01 -9.94686663e-01 -2.24118277e-01 -2.38402635e-01
-6.77183092e-01 -3.45207900e-01 4.32809860e-01 -4.08942193e-01
9.74219441e-01 -2.12566161e+00 -6.01527810e-01 5.25124371e-01
-6.61569610e-02 1.60714537e-01 1.45554736e-01 4.12478060e-01
-3.11132729e-01 4.50986263e-04 -3.48561317e-01 -3.43988836e-01
3.40748250e-01 -3.04226447e-02 -3.47349882e-01 6.17412865e-01
1.88380674e-01 3.82033408e-01 -9.38526392e-01 -8.24592888e-01
-9.34987292e-02 2.85176337e-01 -1.76934823e-01 1.90026954e-01
-7.09207952e-02 2.41061896e-01 -2.66088873e-01 4.51708853e-01
9.32060421e-01 -2.99783528e-01 2.68049568e-01 -1.28391683e-01
-6.62862957e-02 1.22353574e-02 -1.34640348e+00 1.00711966e+00
2.19519600e-01 1.06449910e-01 -1.09945036e-01 -9.76572990e-01
1.29417241e+00 2.10216537e-01 6.02206945e-01 -9.44283605e-02
6.56822175e-02 4.51832414e-01 -4.48238760e-01 -3.09473902e-01
1.72752172e-01 -5.06404042e-01 1.89486206e-01 5.74011087e-01
2.19890639e-01 1.09618688e-02 2.97892299e-02 -3.43787484e-02
5.82038105e-01 7.42922306e-01 5.26422977e-01 -6.33267879e-01
3.35725933e-01 4.58170585e-02 6.06812060e-01 8.68548989e-01
-1.74072921e-01 4.10027117e-01 6.82516813e-01 -5.71889877e-02
-9.28437114e-01 -1.32070506e+00 -6.30323291e-01 8.29761505e-01
-3.86629075e-01 -1.84591666e-01 -8.04297030e-01 -7.76204109e-01
2.41754085e-01 1.40185893e+00 -5.01165926e-01 -1.45499364e-01
2.24855915e-01 -9.91091847e-01 3.59565020e-01 2.84581661e-01
-9.11944956e-02 -5.95433474e-01 -2.14429066e-01 5.18503666e-01
2.97761559e-01 -8.93762112e-01 2.02539712e-01 7.53781021e-01
-1.07493067e+00 -1.13968599e+00 -7.49778509e-01 -1.56069100e-01
8.42314899e-01 -2.80766845e-01 8.65502656e-01 -4.88511711e-01
3.36196385e-02 4.00036693e-01 -2.25818902e-01 -5.60229301e-01
-9.71636891e-01 -3.24477315e-01 2.97525108e-01 5.20998165e-02
7.59359598e-01 -6.06414258e-01 -1.33364365e-01 5.35278857e-01
-6.30761325e-01 -2.18090013e-01 5.15015543e-01 8.02210629e-01
8.34066331e-01 1.14495672e-01 8.73906672e-01 -1.32794702e+00
5.20760000e-01 -4.30952042e-01 -9.13797855e-01 5.58605790e-01
-1.23815370e+00 2.46022061e-01 4.60992843e-01 -6.77473307e-01
-1.17792737e+00 3.12298924e-01 2.58729964e-01 -3.02572638e-01
-3.89984310e-01 3.67506146e-01 1.48111722e-02 4.90730330e-02
9.35402274e-01 7.13404082e-03 1.28764451e-01 -5.89398921e-01
1.05855919e-01 7.96484053e-01 3.49947363e-01 -8.08736622e-01
5.80690324e-01 1.00900650e-01 3.97188216e-01 -7.19887257e-01
-1.12754130e+00 -1.69654652e-01 -8.86483729e-01 -2.92197764e-01
3.71456712e-01 -6.27279043e-01 -7.45107710e-01 5.38883731e-02
-7.36878037e-01 -4.68111664e-01 -4.58063811e-01 8.05426598e-01
-8.93379211e-01 6.20202899e-01 -3.17903250e-01 -1.53776062e+00
1.40271522e-02 -7.30333090e-01 7.58780897e-01 -4.48072627e-02
-5.06507099e-01 -1.08867514e+00 1.47411525e-01 1.34093478e-01
-9.49391350e-02 1.65379241e-01 7.95449734e-01 -1.21625030e+00
1.85855180e-01 -3.92797291e-01 -1.85883731e-01 5.18760026e-01
-1.56264663e-01 2.32113481e-01 -1.24169099e+00 -1.33299321e-01
1.22883350e-01 -3.56175631e-01 5.20711601e-01 5.73596597e-01
8.76105964e-01 -1.34176195e-01 -3.33969355e-01 1.71945572e-01
1.33996522e+00 1.54303655e-01 3.55122983e-01 4.07751091e-02
2.30174646e-01 9.63523865e-01 9.09243584e-01 7.84589708e-01
7.41619468e-02 3.52411240e-01 -2.38073450e-02 2.98102558e-01
2.81284422e-01 -5.72402179e-01 3.31545353e-01 4.46212232e-01
3.96821499e-01 -2.08699354e-03 -8.12962651e-01 -5.91084966e-03
-1.80656970e+00 -7.91525126e-01 -6.95307434e-01 2.66371942e+00
9.73577261e-01 2.63493747e-01 2.84906298e-01 4.34168130e-01
7.61611462e-01 -7.31949747e-01 -4.99409914e-01 -2.27430388e-01
-2.19803303e-02 -7.42937848e-02 6.12106264e-01 5.91586232e-01
-1.08668661e+00 5.28550744e-01 7.23482323e+00 9.25998449e-01
-6.42673552e-01 8.34340081e-02 7.42057979e-01 3.72004837e-01
-2.10100099e-01 1.67878941e-01 -1.11282039e+00 6.26635730e-01
1.09743524e+00 -1.81419745e-01 1.23737015e-01 1.00648654e+00
3.95959586e-01 -5.43245494e-01 -1.31885362e+00 6.78614736e-01
-1.48854017e-01 -4.71616864e-01 -4.20894027e-01 1.91209495e-01
7.14883983e-01 -4.88365561e-01 -1.07044160e-01 2.82283008e-01
3.99613947e-01 -7.78229356e-01 6.21916354e-01 7.70879984e-01
7.52846837e-01 -6.30535066e-01 6.07610047e-01 5.54677129e-01
-4.62339878e-01 -3.86641584e-02 -6.28470778e-01 1.41317949e-01
-1.44466935e-02 1.14932477e+00 -1.18982291e+00 2.87441581e-01
1.24406949e-01 3.43952954e-01 -5.10178566e-01 1.06068075e+00
-2.20366240e-01 8.21326315e-01 -5.21368682e-01 -1.76225789e-02
-9.66945142e-02 -5.44578910e-01 3.26243460e-01 1.09804523e+00
6.32104814e-01 -8.63607079e-02 -3.63446139e-02 1.35866320e+00
7.68005669e-01 4.06872332e-01 -3.60924900e-01 -8.23376700e-02
5.29379368e-01 9.42624867e-01 -7.94398010e-01 -2.65814215e-01
4.82315896e-03 6.29979372e-01 1.26801357e-01 4.68640059e-01
-6.64311409e-01 6.22195378e-03 -1.92132279e-01 1.05145171e-01
2.68154144e-02 1.89448401e-01 -3.74392569e-01 -6.61475480e-01
-4.18827444e-01 -5.88253677e-01 3.14136207e-01 -8.55718970e-01
-1.51324213e+00 2.52504766e-01 3.77193153e-01 -1.27992797e+00
-3.32711279e-01 -6.40480638e-01 -3.00915688e-01 8.96112621e-01
-1.00689662e+00 -7.92726398e-01 7.37944916e-02 2.65200943e-01
6.68169484e-02 -5.15436269e-02 9.62995946e-01 -5.79517186e-02
-4.76361454e-01 5.86908460e-01 2.85322309e-01 -5.86234808e-01
1.14787006e+00 -1.43848491e+00 -4.97380733e-01 4.20319468e-01
-2.54862309e-01 4.62631822e-01 1.21355808e+00 -8.56383801e-01
-7.12082863e-01 -9.15551126e-01 1.09453917e+00 -1.93947986e-01
6.44039810e-01 -9.99669582e-02 -1.01588774e+00 4.41796929e-01
-4.94273990e-01 -1.49117574e-01 1.10722280e+00 1.75181597e-01
-1.52385324e-01 -2.21491948e-01 -1.38698041e+00 3.33619744e-01
5.94913483e-01 -2.85000801e-01 -4.36733156e-01 5.36421120e-01
2.23429218e-01 1.71245307e-01 -1.02252257e+00 4.17246580e-01
6.44865692e-01 -1.22402537e+00 6.38067782e-01 -5.11899710e-01
1.35541603e-01 -1.07268959e-01 -1.58961728e-01 -1.14715350e+00
-4.25678551e-01 -6.84419572e-01 -1.20169837e-02 1.43164873e+00
5.36465228e-01 -5.91841519e-01 9.47691679e-01 9.82166231e-01
1.72346726e-01 -5.66557109e-01 -6.85389936e-01 -1.05646837e+00
-2.93757971e-02 -6.21816099e-01 3.12573016e-01 9.97105002e-01
1.25721872e-01 1.44267440e-01 -4.00479853e-01 6.62354603e-02
1.44011378e+00 -1.83500931e-01 5.89257121e-01 -1.83888793e+00
-3.64863902e-01 -1.70133069e-01 -7.34767765e-02 -6.19277477e-01
2.63352662e-01 -5.00487745e-01 1.58980951e-01 -9.72321332e-01
1.88448027e-01 -5.73555768e-01 -3.25251311e-01 3.29677045e-01
-6.74750730e-02 7.71328434e-02 -2.26454824e-01 1.99583262e-01
-4.21617687e-01 2.63014644e-01 8.51184249e-01 3.23323876e-01
-3.64687681e-01 4.72635835e-01 -4.32982266e-01 9.02480006e-01
9.35267866e-01 -9.34501290e-01 -6.01729691e-01 5.34916997e-01
4.06266153e-01 2.66956151e-01 3.50606352e-01 -9.00955856e-01
8.61340314e-02 -1.26184195e-01 6.02343559e-01 -8.21497023e-01
2.30138823e-01 -9.24924791e-01 6.18358731e-01 4.04165894e-01
-7.40507066e-01 -5.16061366e-01 -7.69993216e-02 6.90593779e-01
2.47726634e-01 -8.86260211e-01 8.80451679e-01 1.05949804e-01
-1.06249876e-01 -9.88762453e-02 -4.20515031e-01 -2.49107555e-01
1.03551078e+00 -2.70360649e-01 1.11741371e-01 -3.81606162e-01
-1.22378302e+00 -6.43452257e-02 4.18413758e-01 -2.99963385e-01
3.35536063e-01 -1.26790607e+00 -5.85082531e-01 3.42421919e-01
3.34117301e-02 -5.31123757e-01 -1.33448094e-01 1.14450872e+00
-1.62902191e-01 3.20874333e-01 7.93625787e-02 -6.14040434e-01
-1.07279396e+00 6.87984526e-01 1.09185316e-01 -1.25567345e-02
6.23304723e-03 5.45107305e-01 -6.84515610e-02 -3.15387189e-01
3.47527742e-01 2.91803293e-03 -4.58475016e-02 1.75769597e-01
3.24180633e-01 7.32513368e-01 -2.92046279e-01 -3.82489890e-01
-1.60505563e-01 2.48514980e-01 2.20813975e-01 -2.97401339e-01
1.20455587e+00 -1.83721706e-01 -1.45071670e-01 1.05086017e+00
1.01839149e+00 -1.10419452e-01 -1.60591722e+00 -3.83077353e-01
3.48402411e-01 -3.65219712e-01 -8.73813033e-02 -9.67638671e-01
-2.57753193e-01 6.05008662e-01 6.56915009e-01 3.91334236e-01
6.48845136e-01 -2.26327032e-01 -1.93706080e-02 3.89978260e-01
4.10517931e-01 -1.38092208e+00 -1.37818947e-01 -6.86601400e-02
6.05382860e-01 -1.30312538e+00 4.20093209e-01 -5.79135120e-01
-7.24333584e-01 1.38067114e+00 2.37737164e-01 1.38635756e-02
7.81701326e-01 1.99269086e-01 9.69336480e-02 2.05523536e-01
-5.60598016e-01 1.97422698e-01 3.28667611e-01 6.94139421e-01
4.13897574e-01 1.77974358e-01 -5.77763438e-01 8.57643902e-01
-5.95186255e-04 3.71459991e-01 3.17099929e-01 5.39323866e-01
-5.92330933e-01 -1.13136661e+00 -6.43277049e-01 4.52380866e-01
-4.89761680e-01 2.15406254e-01 -3.94908905e-01 7.08646119e-01
9.49294865e-02 1.04273272e+00 -1.81877479e-01 -1.23218171e-01
-8.39190409e-02 6.38907015e-01 5.13491511e-01 -3.64682168e-01
2.83394847e-02 5.75034320e-01 1.17431089e-01 -2.14181378e-01
-5.11940241e-01 -9.85011220e-01 -9.51366365e-01 -8.05459470e-02
-5.96579313e-01 5.61291993e-01 8.46407413e-01 8.68109584e-01
7.19845220e-02 1.50488988e-02 7.17360675e-01 -6.06612682e-01
-1.39503992e+00 -9.84464288e-01 -1.14276958e+00 2.59606153e-01
-5.56624234e-02 -6.97777629e-01 -6.14162922e-01 7.39618763e-02] | [7.913333892822266, 4.296858310699463] |
88601cb5-af10-4c33-82d3-50d22ebd01d9 | fire-detection-in-a-still-image-using-colour | 1803.03828 | null | http://arxiv.org/abs/1803.03828v1 | http://arxiv.org/pdf/1803.03828v1.pdf | Fire detection in a still image using colour information | Colour analysis is a crucial step in image-based fire detection algorithms.
Many of the proposed fire detection algorithms in a still image are prone to
false alarms caused by objects with a colour similar to fire. To design a
colour-based system with a better false alarm rate, a new
colour-differentiating conversion matrix, efficient on images of high colour
complexity, is proposed. The elements of this conversion matrix are obtained by
performing K-medoids clustering and Particle Swarm Optimisation procedures on a
fire sample image with a background of high fire-colour similarity. The
proposed conversion matrix is then used to construct two new fire colour
detection frameworks. The first detection method is a two-stage non-linear
image transformation framework, while the second is a direct transformation of
an image with the proposed conversion matrix. A performance comparison of the
proposed methods with alternate methods in the literature was carried out.
Experimental results indicate that the linear image transformation method
outperforms other methods regarding false alarm rate while the non-linear
two-stage image transformation method has the best performance on the F-score
metric and provides a better trade-off between missed detection and false alarm
rate. | ['Abdsamad Benkrid', 'Oluwarotimi Giwa'] | 2018-03-10 | null | null | null | null | ['fire-detection'] | ['time-series'] | [ 7.27200389e-01 -5.93932152e-01 6.27586246e-01 2.58156449e-01
-1.99268788e-01 -4.99524146e-01 7.79408574e-01 1.63671419e-01
-7.53363431e-01 5.68048358e-01 -4.42368001e-01 -1.26190633e-01
-7.04442620e-01 -9.55657423e-01 -2.35521290e-02 -1.23352480e+00
1.74112618e-01 3.92270654e-01 5.64276993e-01 2.96376012e-02
4.71143514e-01 7.82750964e-01 -1.83150268e+00 1.62002668e-01
4.95366514e-01 1.00966585e+00 3.14536005e-01 1.15531027e+00
1.07489660e-01 5.13586640e-01 -8.84612083e-01 1.41225204e-01
7.01152265e-01 -8.53842914e-01 -5.31710327e-01 4.01536793e-01
-4.21850175e-01 1.68949008e-01 5.17954767e-01 9.26421583e-01
5.47954202e-01 3.25331122e-01 9.67617273e-01 -1.42978728e+00
-1.93499010e-02 -1.43215761e-01 -6.03733242e-01 6.09665811e-01
2.10079774e-01 3.20361167e-01 2.01562718e-01 -7.92038560e-01
1.15671359e-01 1.19652724e+00 4.37007755e-01 8.99681225e-02
-1.22177541e+00 -5.88056922e-01 -4.67611045e-01 2.89803654e-01
-1.64301336e+00 -8.69881455e-03 6.45944953e-01 -4.65268046e-01
8.22935224e-01 8.60562146e-01 8.55612218e-01 4.18240190e-01
2.70880520e-01 -7.63971135e-02 1.54850340e+00 -7.81033337e-01
4.15551901e-01 2.32407928e-01 -4.02985096e-01 5.32693446e-01
5.51859558e-01 4.41774547e-01 3.93575095e-02 -1.37486219e-01
6.39801562e-01 1.44570231e-01 -1.79461852e-01 -1.87572896e-01
-9.96010184e-01 8.00819039e-01 2.32182324e-01 7.94909596e-01
-6.51579142e-01 -2.07325026e-01 1.15445510e-01 1.34235963e-01
7.88900554e-02 5.44687152e-01 7.55442902e-02 1.76384319e-02
-1.05941570e+00 -4.26163115e-02 8.95041466e-01 1.15867153e-01
3.71957481e-01 8.86284113e-02 -2.97802329e-01 5.76054931e-01
4.42776471e-01 6.79183304e-01 2.09215388e-01 -4.54173297e-01
-1.73176363e-01 9.40554261e-01 2.23224863e-01 -1.17057621e+00
-4.41527188e-01 -5.23783505e-01 -8.86637568e-01 8.01279128e-01
4.17993039e-01 -1.33451357e-01 -9.28565621e-01 9.32797492e-01
5.45169234e-01 1.09775342e-01 1.98246151e-01 7.69113839e-01
2.43297100e-01 8.60636175e-01 -2.97130905e-02 -5.96880198e-01
1.40133917e+00 -4.29667205e-01 -4.89226967e-01 6.66933879e-02
-1.11882657e-01 -1.14326644e+00 4.43192005e-01 6.66821837e-01
-6.21881306e-01 -5.44728220e-01 -1.12835073e+00 1.14931536e+00
-4.16631013e-01 1.40817240e-01 1.14647567e-01 9.93848145e-01
-9.10637796e-01 2.00930879e-01 -4.52159792e-01 -9.18392777e-01
-5.98999951e-03 2.57042229e-01 -7.40440935e-03 6.32150099e-02
-7.06246555e-01 1.06928515e+00 7.07206964e-01 2.70194918e-01
-6.80383563e-01 -1.29623652e-01 -3.15611064e-01 -2.14592382e-01
3.42825472e-01 -3.24263930e-01 6.90159142e-01 -9.48518217e-01
-1.23325384e+00 6.63527906e-01 1.60105169e-01 -3.21823269e-01
7.53213584e-01 1.88529626e-01 -6.75190687e-01 3.98281306e-01
2.65588854e-02 1.49504945e-01 1.19861901e+00 -1.68255508e+00
-1.17469442e+00 -1.36367217e-01 -3.35566580e-01 2.29885012e-01
1.29318684e-01 1.92909747e-01 1.40900448e-01 -3.98438364e-01
2.05921121e-02 -7.59391904e-01 -2.64219671e-01 -2.58571655e-01
-1.68132544e-01 -8.03996157e-03 1.10770369e+00 -1.88712433e-01
1.24127960e+00 -1.83598387e+00 -2.59443462e-01 9.26696360e-01
-1.57371119e-01 5.91566443e-01 2.03632504e-01 4.72040594e-01
-2.89447546e-01 -1.44025832e-01 -6.08528197e-01 4.70731795e-01
-5.29917657e-01 1.30516499e-01 5.24451196e-01 5.09057820e-01
5.64759552e-01 -1.17609084e-01 -9.42267597e-01 -4.91196007e-01
6.23006046e-01 6.40236080e-01 2.78549036e-04 1.91440806e-01
3.56663972e-01 4.98292685e-01 -2.38254368e-01 5.73995173e-01
7.29656219e-01 1.89711422e-01 -8.46067891e-02 -3.63388330e-01
-4.01170760e-01 -7.80947626e-01 -1.67974091e+00 5.68859279e-01
-1.10389747e-01 2.22090930e-01 1.04549527e-02 -9.82929528e-01
1.42246819e+00 5.60901642e-01 7.36696243e-01 -5.72152853e-01
5.45022190e-01 1.10826008e-02 1.69789955e-01 -4.96802539e-01
3.60720962e-01 -4.56581831e-01 3.41120452e-01 4.61426318e-01
-3.74842376e-01 -1.53752163e-01 6.05744243e-01 -2.46516839e-01
1.15574932e+00 -1.98398143e-01 3.58703077e-01 -3.68067205e-01
9.97918189e-01 9.19884816e-02 2.60816991e-01 8.01253021e-01
-2.63435513e-01 4.10790175e-01 5.43742478e-02 -4.43112880e-01
-8.49791944e-01 -9.67566431e-01 -5.51710017e-02 4.51126099e-01
2.45961294e-01 1.63103878e-01 -6.46299720e-01 -1.90200165e-01
-2.76796311e-01 6.68381214e-01 -4.62595940e-01 -3.87471765e-01
-2.35628903e-01 -1.31685030e+00 4.38575298e-01 -1.40810460e-01
7.51100421e-01 -1.13475490e+00 -1.52805090e+00 3.36972356e-01
-1.18386015e-01 -6.00490570e-01 2.89547354e-01 4.97083694e-01
-6.96966708e-01 -1.39226055e+00 -7.32193947e-01 -6.13088667e-01
8.15029204e-01 4.67831522e-01 7.84820259e-01 4.03723776e-01
-8.29036176e-01 4.85299587e-01 -1.04157257e+00 -5.82173109e-01
-7.63508201e-01 -6.11228585e-01 -2.33487099e-01 5.72588861e-01
5.46589613e-01 -1.52417585e-01 -7.29225278e-01 3.90450031e-01
-1.27743304e+00 -3.57951730e-01 8.48588049e-01 7.34497070e-01
3.62503827e-01 1.05769813e+00 5.70518970e-02 -3.13560724e-01
7.87075818e-01 -2.63667494e-01 -6.57248080e-01 3.27742100e-01
-7.38003373e-01 1.43128723e-01 3.78637165e-01 -3.43458802e-01
-1.04696071e+00 4.04132873e-01 2.93703318e-01 2.08778586e-02
-3.88327181e-01 4.48947512e-02 1.21234894e-01 -2.69406855e-01
8.86516154e-01 3.14221770e-01 6.01950288e-02 -2.17367664e-01
-8.53270739e-02 5.65007210e-01 6.37446880e-01 -1.15437344e-01
1.13015985e+00 5.41839361e-01 4.14954960e-01 -9.49168861e-01
-1.41105786e-01 -9.34809804e-01 -6.87982619e-01 -8.58660936e-01
1.00624156e+00 -2.74858207e-01 -5.02840817e-01 6.67006910e-01
-1.02016878e+00 2.63659298e-01 -1.71597674e-01 5.98640323e-01
-2.66658217e-01 3.14354718e-01 1.55514702e-01 -1.55815089e+00
-4.60702121e-01 -9.27764714e-01 6.31722152e-01 3.03038538e-01
2.30554938e-02 -5.16116500e-01 1.81656092e-01 -1.36692539e-01
3.25377136e-01 7.27408469e-01 5.67626536e-01 -3.94447237e-01
-1.61508322e-01 -4.74919617e-01 -2.23380089e-01 3.67017508e-01
4.69633847e-01 2.94656515e-01 -5.95251024e-01 -2.27763236e-01
4.14143711e-01 4.30919945e-01 6.64900780e-01 3.71550798e-01
2.53844708e-01 -7.63007849e-02 -2.79550403e-01 3.41587484e-01
2.03203201e+00 1.06340003e+00 7.88424850e-01 8.51375163e-01
2.74485588e-01 3.81036848e-01 9.45583045e-01 9.35519993e-01
-3.14091593e-01 3.49332780e-01 5.95366895e-01 -5.08307040e-01
5.44033907e-02 5.10993719e-01 1.37239814e-01 2.05113903e-01
-5.27412355e-01 -3.74395728e-01 -8.32457423e-01 3.79042566e-01
-1.60280287e+00 -1.30135477e+00 -4.08852875e-01 2.46827626e+00
1.56446904e-01 2.58552104e-01 4.78535920e-01 1.12608933e+00
1.13734710e+00 -3.32317740e-01 -4.32411022e-02 -5.27575910e-01
-1.27096534e-01 4.50218767e-01 7.51686156e-01 3.02617222e-01
-1.31485128e+00 2.62708992e-01 5.89671659e+00 6.27538383e-01
-9.93729234e-01 -7.14296699e-02 2.31740132e-01 2.16583073e-01
3.94652158e-01 8.34583193e-02 -3.48210752e-01 4.74838287e-01
7.03122258e-01 3.95094752e-02 4.25179839e-01 2.37834051e-01
4.73449171e-01 -9.29979622e-01 -5.04095495e-01 1.03434598e+00
1.85930192e-01 -5.12058079e-01 8.11910927e-02 3.23855244e-02
4.52362686e-01 -6.37786269e-01 -3.50902677e-01 -4.74036455e-01
1.54213384e-01 -7.82326043e-01 6.64002955e-01 8.20127964e-01
3.47586304e-01 -1.11271453e+00 9.75776553e-01 1.87636465e-01
-1.38867640e+00 -3.09078008e-01 -1.85396612e-01 -1.07503571e-01
2.13063180e-01 4.75526154e-01 -9.22793746e-01 5.96556783e-01
7.42738307e-01 9.77719650e-02 -8.19414437e-01 1.62974668e+00
1.75650075e-01 3.70445967e-01 -6.85123622e-01 -1.43616259e-01
2.91841924e-01 -4.70609754e-01 9.07300532e-01 1.32986104e+00
5.84218681e-01 6.74072560e-03 -6.11683279e-02 8.17431271e-01
9.46289897e-01 1.27855703e-01 -5.39555728e-01 1.22755887e-02
4.50763047e-01 1.33035731e+00 -1.67676187e+00 -3.51774603e-01
-8.20575804e-02 1.10809386e+00 -7.65260518e-01 4.94426116e-02
-8.35711360e-01 -4.18901384e-01 6.09156638e-02 1.51515007e-01
2.49204025e-01 5.44975400e-02 -6.62983861e-03 -1.76991746e-01
-3.58718067e-01 -6.41078353e-01 5.94002247e-01 -6.89657271e-01
-9.88566577e-01 8.39094520e-01 3.23539436e-01 -1.52132714e+00
-1.91875100e-01 -5.34414709e-01 -5.72851896e-01 8.87201786e-01
-8.21064532e-01 -8.85611713e-01 -5.41977644e-01 7.87101328e-01
5.03824711e-01 -2.62657881e-01 9.14030492e-01 4.28943112e-02
-4.42363173e-01 1.74989909e-01 1.65616974e-01 -3.05102050e-01
1.94726646e-01 -1.06870639e+00 -2.77493536e-01 1.47443295e+00
-2.44018823e-01 -4.64428216e-02 9.85063016e-01 -7.74306655e-01
-8.63230944e-01 -8.69436681e-01 3.60898525e-01 -1.38125747e-01
1.19866557e-01 1.21175624e-01 -4.51974094e-01 -3.64923954e-01
8.88805613e-02 -3.78963351e-01 3.67726654e-01 -7.70044386e-01
2.99916595e-01 -2.60840744e-01 -1.65464640e+00 4.39253934e-02
1.72263414e-01 4.65415828e-02 -4.95543092e-01 5.36627844e-02
-2.62074172e-01 3.02655458e-01 -7.75966406e-01 1.44389763e-01
6.00962698e-01 -1.23042977e+00 1.03650570e+00 2.00303093e-01
-2.94527769e-01 -9.74595010e-01 2.16345266e-01 -1.03164804e+00
-7.22860754e-01 -3.68759483e-01 5.78168571e-01 1.19304252e+00
1.01821423e-01 -5.05839944e-01 4.03657675e-01 1.36509016e-01
4.05137330e-01 -1.02679655e-01 -8.08985114e-01 -9.22328711e-01
-6.75765812e-01 -2.31451709e-02 1.93823934e-01 5.82326412e-01
-3.85788232e-01 3.68276536e-02 -1.99438125e-01 2.06666350e-01
8.08488429e-01 -4.32951391e-01 4.77039903e-01 -1.60661709e+00
-1.19751722e-01 -4.47184145e-01 -6.99500084e-01 4.30887520e-01
-7.98135757e-01 -4.08326477e-01 3.54540825e-01 -1.60754144e+00
1.07002974e-01 -1.82347715e-01 -4.52439308e-01 2.22163945e-01
-2.12982163e-01 5.85230708e-01 3.17056894e-01 2.55055696e-01
-2.29970124e-02 -5.28736301e-02 8.83127213e-01 -1.62494164e-02
-3.94168019e-01 2.40301371e-01 -2.93926895e-01 7.23102748e-01
8.66577148e-01 -8.19330692e-01 -3.93322557e-01 1.45395741e-01
-7.69691616e-02 -2.61933386e-01 4.73033011e-01 -1.66802943e+00
1.35675788e-01 -2.40743443e-01 6.32122040e-01 -6.98916018e-01
1.92584574e-01 -1.37921119e+00 7.17590332e-01 1.00961518e+00
2.75648415e-01 4.98363376e-01 1.00458845e-01 5.39759636e-01
-1.51167754e-02 -5.29913962e-01 1.18514705e+00 -4.39973861e-01
-7.89832294e-01 -3.97433639e-01 -9.99866128e-01 -4.87256497e-01
1.47124958e+00 -1.09541345e+00 3.66583876e-02 -1.71902508e-01
-5.68244815e-01 -4.34500545e-01 3.02934796e-01 1.96472868e-01
6.14584744e-01 -1.09338844e+00 -7.24119484e-01 1.99562281e-01
-7.21550407e-03 -3.33903790e-01 -1.19472973e-01 1.08236706e+00
-9.24060047e-01 8.43071863e-02 -7.39223242e-01 -6.85092449e-01
-1.68661380e+00 5.33235431e-01 3.39251220e-01 -2.63719946e-01
-3.50679129e-01 6.30998254e-01 -1.99832797e-01 2.31634229e-01
-7.47588351e-02 -1.99787185e-01 -6.20229900e-01 1.02921873e-01
5.04938424e-01 8.20574105e-01 1.94822073e-01 -1.06287503e+00
-5.84965527e-01 7.79076576e-01 5.48322380e-01 -3.89671892e-01
1.30638051e+00 1.89019945e-02 -2.06465170e-01 3.01252931e-01
6.44858897e-01 -3.92417222e-01 -9.90362227e-01 2.64803588e-01
1.30499795e-01 -7.52463877e-01 2.42144912e-01 -1.10512364e+00
-8.68370891e-01 3.30109686e-01 1.47243738e+00 6.54728413e-01
1.78408885e+00 -5.37613809e-01 1.02187589e-01 1.62635565e-01
8.07932019e-03 -1.25588810e+00 -6.43439591e-02 2.57526070e-01
8.36270630e-01 -8.94376636e-01 1.27416626e-01 -3.34684789e-01
-5.10059893e-01 1.22077286e+00 2.75599301e-01 -2.47172594e-01
5.77911496e-01 4.13352072e-01 2.58591361e-02 -3.79763365e-01
-2.97238797e-01 -7.05722988e-01 1.42212182e-01 1.01202774e+00
-5.38754044e-03 6.91274181e-02 -7.13419795e-01 1.66544281e-02
1.21487819e-01 8.43136683e-02 4.07184750e-01 1.15588546e+00
-8.07699859e-01 -9.86035228e-01 -1.41067529e+00 6.32220149e-01
-3.39237899e-01 1.16894126e-01 -5.70379317e-01 9.23599720e-01
5.24216592e-01 1.53694510e+00 8.82432461e-02 -6.80389106e-01
5.48000693e-01 -2.22100783e-02 5.30731618e-01 -1.20834611e-01
-8.11671734e-01 3.46034586e-01 -1.77471906e-01 -1.67935476e-01
-8.44775617e-01 -6.13866568e-01 -9.62363303e-01 -1.40150413e-01
-6.13586605e-01 2.78252929e-01 9.28263009e-01 9.67873871e-01
-3.08530718e-01 5.64504504e-01 8.48982096e-01 -7.86696792e-01
8.58629402e-03 -8.53676677e-01 -7.66072929e-01 5.13402820e-01
1.08619139e-01 -9.10189807e-01 -6.04896843e-01 1.83607221e-01] | [9.245498657226562, -1.2770963907241821] |
18a9b814-7bf8-4230-8e1c-5710228b10eb | subtype-former-a-deep-learning-approach-for | 2207.14639 | null | https://arxiv.org/abs/2207.14639v1 | https://arxiv.org/pdf/2207.14639v1.pdf | Subtype-Former: a deep learning approach for cancer subtype discovery with multi-omics data | Motivation: Cancer is heterogeneous, affecting the precise approach to personalized treatment. Accurate subtyping can lead to better survival rates for cancer patients. High-throughput technologies provide multiple omics data for cancer subtyping. However, precise cancer subtyping remains challenging due to the large amount and high dimensionality of omics data. Results: This study proposed Subtype-Former, a deep learning method based on MLP and Transformer Block, to extract the low-dimensional representation of the multi-omics data. K-means and Consensus Clustering are also used to achieve accurate subtyping results. We compared Subtype-Former with the other state-of-the-art subtyping methods across the TCGA 10 cancer types. We found that Subtype-Former can perform better on the benchmark datasets of more than 5000 tumors based on the survival analysis. In addition, Subtype-Former also achieved outstanding results in pan-cancer subtyping, which can help analyze the commonalities and differences across various cancer types at the molecular level. Finally, we applied Subtype-Former to the TCGA 10 types of cancers. We identified 50 essential biomarkers, which can be used to study targeted cancer drugs and promote the development of cancer treatments in the era of precision medicine. | ['Zhe Wang', 'Jing Zhang', 'Dongdong Li', 'Xiaoyang Fang', 'Yi Jiang', 'Yuhang Sheng', 'Hai Yang'] | 2022-07-28 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [-1.24703772e-01 -6.17020428e-01 -9.28927004e-01 -1.06527485e-01
-7.86528587e-01 -3.89309853e-01 1.35793716e-01 4.86738473e-01
8.44342858e-02 9.39356327e-01 3.91803384e-01 -5.06951511e-01
-4.99531478e-01 -8.29429150e-01 -2.59810328e-01 -1.23593974e+00
1.08526938e-01 5.92527986e-01 -4.90223318e-01 -1.71483904e-02
8.21603695e-04 6.71582341e-01 -1.27829027e+00 8.61859679e-01
1.03762460e+00 9.84404862e-01 9.94022489e-02 4.95367527e-01
-4.84203637e-01 3.87288183e-02 -3.61663550e-01 -2.40837373e-02
-2.54054964e-01 -3.36380482e-01 -4.68807638e-01 -5.31327307e-01
1.26747757e-01 2.33747467e-01 -2.31722593e-01 8.91731918e-01
6.82073116e-01 -6.08848155e-01 8.93863857e-01 -9.82067108e-01
-6.75831437e-01 3.27127695e-01 -1.16586529e-01 -3.62229049e-01
3.24404269e-01 7.24696219e-02 7.80479670e-01 -5.33130050e-01
4.80517566e-01 9.74499285e-01 6.59011602e-01 7.77945578e-01
-1.06571615e+00 -7.73245573e-01 -3.20327193e-01 3.24140698e-01
-1.27660608e+00 -2.53661454e-01 2.55967319e-01 -7.28905857e-01
8.82182062e-01 7.74256408e-01 9.73911405e-01 1.12603867e+00
1.09571207e+00 5.80478311e-01 1.03714585e+00 1.20167695e-01
2.64463097e-01 -1.59979105e-01 3.57047081e-01 7.99440801e-01
4.34129387e-01 2.75164068e-01 -4.51303601e-01 -5.19248009e-01
5.26764467e-02 8.15754592e-01 -3.19147736e-01 1.18755586e-01
-1.35684681e+00 5.56231916e-01 4.55381840e-01 4.23734635e-01
-1.64544493e-01 -1.21680953e-01 5.22957504e-01 2.05324143e-01
2.64311314e-01 6.91512287e-01 -6.49111927e-01 2.82166839e-01
-5.78216076e-01 -1.47573993e-01 6.52145028e-01 7.43761539e-01
4.03278440e-01 -3.72790217e-01 -4.55190539e-01 8.02664042e-01
2.42098391e-01 5.22205472e-01 1.05970573e+00 -1.93870649e-01
-1.06088430e-01 1.03163242e+00 -1.24847442e-01 -6.55583799e-01
-1.18584275e+00 -6.44265294e-01 -1.30666018e+00 -5.33404410e-01
3.04719061e-01 4.02557924e-02 -9.76470768e-01 1.33134186e+00
3.75857532e-01 3.87519240e-01 2.53468573e-01 6.62539363e-01
1.21690106e+00 2.17492238e-01 2.49988720e-01 -3.53325963e-01
1.74002087e+00 -4.28080380e-01 -5.94972849e-01 4.65417653e-01
1.27296090e+00 -3.78382862e-01 8.86197388e-01 3.48130494e-01
-1.54791445e-01 2.71609556e-02 -8.61256063e-01 8.13028291e-02
-7.99757779e-01 3.28447372e-01 1.07293653e+00 7.81496346e-01
-5.30853510e-01 5.44459879e-01 -7.71664441e-01 -5.00158906e-01
6.42248988e-01 5.04600763e-01 -4.67256218e-01 -3.29143316e-01
-1.14427078e+00 5.81258655e-01 3.14159513e-01 -1.59324482e-01
-9.77470756e-01 -1.38495743e+00 -5.71017742e-01 8.09969977e-02
-2.56842136e-01 -1.29913795e+00 6.17399812e-01 -3.23470145e-01
-1.65049839e+00 7.35873282e-01 -5.06118655e-01 7.52341151e-02
-1.73259825e-01 1.31325468e-01 -6.15051031e-01 -2.46580467e-01
7.42593140e-04 5.31592257e-02 -4.40922938e-02 -4.47434723e-01
-7.86889672e-01 -7.81083882e-01 -6.81743801e-01 1.52593911e-01
-4.57010776e-01 -3.09497982e-01 -2.51715668e-02 -1.15341522e-01
2.60045141e-01 -1.06547630e+00 -2.59957969e-01 -2.61564761e-01
-6.32513762e-01 -1.95336625e-01 6.23393893e-01 -5.26103079e-01
9.50287402e-01 -1.94860959e+00 5.64501524e-01 -1.58079833e-01
5.94425023e-01 1.27610698e-01 -1.36034071e-01 5.68056703e-01
-1.56300545e-01 3.35833848e-01 1.27604902e-01 6.95193931e-02
-7.54292905e-02 -1.26542792e-01 -8.26986581e-02 9.81540263e-01
-1.16858475e-01 1.21227181e+00 -1.00053489e+00 -3.04433286e-01
9.89135504e-02 3.44703108e-01 -1.07243121e-01 1.42672956e-01
-4.97040600e-01 3.64474088e-01 -6.76126897e-01 1.44326115e+00
5.35117269e-01 -3.16818237e-01 1.61962658e-01 -2.59317100e-01
1.78114936e-01 4.97974567e-02 -1.65296152e-01 1.76068699e+00
-3.02832991e-01 3.93005610e-01 -3.69026601e-01 -8.07711482e-01
7.28155613e-01 3.05927098e-01 9.27147806e-01 -6.41427219e-01
1.34917438e-01 5.28162360e-01 2.82216877e-01 -7.99651265e-01
-2.61865050e-01 -3.52701992e-01 -2.37159669e-01 -4.05732505e-02
-3.43618810e-01 2.26390421e-01 -6.72432929e-02 -2.51673728e-01
1.14164114e+00 -4.64023203e-01 3.49040031e-01 -5.89744151e-01
6.20471656e-01 2.69907236e-01 9.18712556e-01 2.46190727e-01
-1.36747599e-01 2.13106140e-01 6.25241935e-01 -6.40194535e-01
-6.21258974e-01 -7.51734018e-01 -7.43642926e-01 7.20295846e-01
1.22053787e-01 -2.90535659e-01 -1.03519075e-01 -5.30457973e-01
3.98590773e-01 2.67238170e-01 -7.97687948e-01 -4.13707882e-01
5.48106395e-02 -1.82408977e+00 8.57902884e-01 1.79124534e-01
3.44898164e-01 -1.84255421e-01 2.22339332e-01 1.44150779e-01
-7.36917257e-02 -7.10644245e-01 -1.52577341e-01 2.32125223e-01
-8.37971628e-01 -1.40046704e+00 -8.46420109e-01 -5.70396364e-01
5.81707835e-01 7.08516166e-02 6.67411208e-01 7.35436380e-02
-5.69029033e-01 -4.02462065e-01 -2.07123235e-01 -7.74753273e-01
-4.43483025e-01 3.87332201e-01 2.41197735e-01 -1.12615809e-01
7.92571425e-01 -2.72812188e-01 -6.49937510e-01 4.27015871e-01
-4.86909151e-01 1.44350916e-01 9.77331400e-01 1.08901846e+00
9.21186745e-01 8.70847106e-02 9.03220713e-01 -9.19957340e-01
5.36272407e-01 -8.89337003e-01 -3.23433518e-01 4.17769015e-01
-6.91312134e-01 -3.51638570e-02 9.40933943e-01 -4.33915406e-01
-4.99876052e-01 3.33347358e-02 -3.32807690e-01 -3.39498110e-02
-2.27451190e-01 7.69560039e-01 -5.35988331e-01 -2.34109581e-01
5.58403790e-01 4.83037353e-01 7.19807344e-03 -3.00536394e-01
-2.52002001e-01 1.08441627e+00 -2.70075543e-04 -3.20479751e-01
-8.20595399e-02 3.35931718e-01 7.54631937e-01 -8.15444708e-01
-7.67730713e-01 -6.60050035e-01 -2.12033942e-01 3.12763780e-01
8.88841867e-01 -9.00127649e-01 -1.17147219e+00 6.62514627e-01
-7.01818407e-01 -2.79060125e-01 4.37930971e-01 7.38013506e-01
-3.45549524e-01 2.32734442e-01 -9.30702627e-01 -8.16048831e-02
-7.40727544e-01 -1.42539215e+00 1.25076890e+00 1.64440721e-01
-1.23916708e-01 -1.00261724e+00 4.71912235e-01 1.64876923e-01
2.28317603e-01 5.14252663e-01 1.49752176e+00 -7.16768563e-01
-2.69783050e-01 -4.09510791e-01 -3.46770376e-01 -4.88817006e-01
6.37376249e-01 -5.08545488e-02 -1.06810808e+00 -1.44495338e-01
-2.75874585e-01 1.49247080e-01 8.37843776e-01 4.56454456e-01
1.62424707e+00 -1.20314993e-01 -1.32999909e+00 1.38235116e+00
1.45324564e+00 3.80346954e-01 5.71643949e-01 1.39533117e-01
9.77920830e-01 2.24824160e-01 5.36612511e-01 4.20503058e-02
2.52680689e-01 5.61031878e-01 4.04069513e-01 -1.36279792e-01
4.54458445e-02 4.05624509e-02 -9.04828236e-02 6.37078941e-01
2.60189027e-01 -2.72358537e-01 -9.98479724e-01 3.08593452e-01
-1.64253509e+00 -6.71635032e-01 -5.32392919e-01 1.96750462e+00
9.73249197e-01 -5.82788467e-01 3.03347893e-02 -2.47517660e-01
3.23473573e-01 -5.21166027e-01 -9.89890873e-01 -1.86532706e-01
-3.99170399e-01 -3.11079115e-01 6.41084075e-01 1.69585556e-01
-8.49866867e-01 5.80054820e-01 6.99612570e+00 8.28652442e-01
-1.44576871e+00 -1.45155594e-01 7.55076468e-01 -2.30523646e-01
-2.54898906e-01 -2.79416800e-01 -1.16807699e+00 8.66820276e-01
1.09489024e+00 -2.10582644e-01 9.66070965e-02 5.54751337e-01
1.95380270e-01 1.21699445e-01 -1.41791105e+00 1.07277822e+00
-3.07554245e-01 -1.86337101e+00 -2.75683645e-02 5.91242909e-01
7.37937629e-01 2.93949932e-01 1.77097891e-03 5.58350720e-02
9.01966766e-02 -1.39275002e+00 -2.25919649e-01 8.71311724e-01
1.17152941e+00 -6.17426932e-01 1.01253200e+00 4.30405021e-01
-9.53072250e-01 -2.81886429e-01 -6.91541910e-01 2.82684684e-01
-5.09115577e-01 1.15836680e+00 -1.10233426e+00 8.55186641e-01
3.80577892e-01 1.10642970e+00 -4.47510570e-01 1.16056728e+00
1.56982198e-01 1.91795155e-01 -1.43486321e-01 -5.89701891e-01
-3.38371605e-01 1.05270697e-02 6.59059361e-02 1.35609460e+00
5.42565823e-01 8.98629799e-02 2.63376385e-02 5.46744764e-01
4.29471955e-02 3.12374264e-01 -2.89386213e-01 -3.34006578e-01
4.72514957e-01 1.40291464e+00 -4.01488602e-01 -3.80093694e-01
-7.76581764e-02 4.99054939e-01 1.09464295e-01 2.45853484e-01
-6.22002840e-01 -3.90781790e-01 1.17179942e+00 -8.53828490e-02
-4.51066315e-01 1.53054923e-01 -5.01398623e-01 -1.38099432e+00
-7.49996185e-01 -1.08888435e+00 7.55647123e-01 -2.51166880e-01
-1.56689799e+00 3.12299907e-01 -3.26151609e-01 -1.11489296e+00
2.58298039e-01 -1.13547266e+00 -6.03842199e-01 1.08208263e+00
-1.44364262e+00 -1.10423303e+00 -6.12871766e-01 2.55303860e-01
1.64434761e-01 -3.16255867e-01 1.32182014e+00 2.72122264e-01
-1.15607858e+00 6.43155992e-01 5.99290133e-01 -2.34450534e-01
9.22662377e-01 -1.06024539e+00 -2.77816951e-01 9.07648653e-02
-6.82497084e-01 4.67972338e-01 4.24357653e-01 -8.89454007e-01
-2.08973980e+00 -1.29472172e+00 5.50986767e-01 -4.83142138e-01
5.46345472e-01 -2.90890336e-01 -6.35573745e-01 3.34678143e-01
-3.80909562e-01 -6.24549724e-02 1.67510056e+00 2.37733930e-01
-1.45413190e-01 -3.08313489e-01 -1.30335128e+00 5.73640049e-01
7.72815347e-01 -4.45408493e-01 3.23488489e-02 7.31796980e-01
7.70241141e-01 -4.67468321e-01 -1.51903069e+00 4.53372270e-01
1.01853526e+00 -5.41448653e-01 8.59283328e-01 -9.85538185e-01
4.18856382e-01 -3.36635172e-01 -3.46980631e-01 -1.55325103e+00
-6.82786167e-01 -1.34891108e-01 -1.50555909e-01 6.69642746e-01
5.00200331e-01 -1.06787109e+00 1.06769669e+00 5.18878579e-01
-3.97330731e-01 -1.24066448e+00 -7.91214287e-01 -5.97319126e-01
4.37676996e-01 1.66711450e-01 1.33333838e+00 1.04038358e+00
6.56082273e-01 1.14491202e-01 3.54423150e-02 2.48483452e-03
4.97880608e-01 3.62237871e-01 7.32539892e-01 -1.40148056e+00
-3.40279549e-01 -7.48043716e-01 -5.83181322e-01 -3.06937039e-01
2.03645736e-01 -1.25056410e+00 -3.98083866e-01 -1.53900933e+00
5.24139524e-01 -4.97646928e-01 -4.92775112e-01 4.93809909e-01
-2.21549898e-01 -1.65120780e-01 -4.38962519e-01 2.49912992e-01
-8.86699930e-03 3.04089099e-01 1.34041512e+00 -5.72339058e-01
-2.37665892e-01 -8.72701705e-02 -1.09353662e+00 2.51777530e-01
9.80039239e-01 -4.70770955e-01 -8.08093920e-02 1.50278080e-02
3.83083016e-01 1.89233318e-01 8.08396488e-02 -8.28663111e-01
1.64220393e-01 -8.19243312e-01 8.51570308e-01 -5.82113087e-01
2.35196084e-01 -5.40541410e-01 8.08859110e-01 1.07420790e+00
-1.28510520e-01 -3.90138000e-01 4.39160556e-01 4.21660960e-01
1.18772820e-01 1.21477269e-01 6.24778152e-01 -2.70153140e-03
-3.49562496e-01 7.30303049e-01 -4.11280811e-01 -5.99498928e-01
1.14265871e+00 -6.59301281e-02 -9.29902077e-01 4.05007303e-01
-4.62553561e-01 1.96528226e-01 4.77718651e-01 -3.69494632e-02
4.65883046e-01 -1.35079193e+00 -8.01379621e-01 3.11865509e-01
4.53957856e-01 -1.49003357e-01 4.59938884e-01 1.23728883e+00
-6.13703251e-01 8.67238224e-01 -2.03802556e-01 -8.76519144e-01
-1.27615964e+00 8.40022981e-01 8.74064803e-01 -2.07491413e-01
-3.10099155e-01 4.81391609e-01 6.03175938e-01 -4.00696278e-01
-2.16832515e-02 -3.39119345e-01 -1.92795128e-01 -1.74446832e-02
6.81951165e-01 3.57480109e-01 1.96905956e-01 -2.29428381e-01
-6.67375088e-01 6.56410456e-01 -1.18579321e-01 7.19827592e-01
1.17781019e+00 1.50239259e-01 -4.42068756e-01 3.36924165e-01
1.45486462e+00 -2.46264994e-01 -3.11900049e-01 2.72476315e-01
-1.95248619e-01 -3.11896622e-01 7.31691644e-02 -1.05791306e+00
-8.67730439e-01 6.37658834e-01 6.18203223e-01 -4.66529652e-02
1.24725020e+00 -5.93832061e-02 8.88071537e-01 4.73937750e-01
1.42742932e-01 -5.00033796e-01 -4.76167113e-01 2.51324594e-01
4.95934069e-01 -1.01745284e+00 2.43170813e-01 -4.36943203e-01
1.00851767e-01 1.26889443e+00 4.46461409e-01 3.38134527e-01
7.99423993e-01 3.39045495e-01 2.10757814e-02 -3.57912242e-01
-9.26055908e-01 2.41938084e-01 3.95172387e-01 5.84020138e-01
5.65317452e-01 6.05235636e-01 -4.47865337e-01 1.04627502e+00
-1.78170159e-01 2.78001338e-01 1.72645569e-01 4.22718853e-01
-3.63233417e-01 -1.22295797e+00 -3.76819283e-01 9.76541758e-01
-5.53970277e-01 -8.77244584e-03 -7.07051575e-01 4.41302389e-01
1.26033217e-01 6.55641377e-01 -3.05499077e-01 -7.35545576e-01
-1.66941211e-02 4.01887223e-02 2.50259608e-01 -4.69696134e-01
-4.06706005e-01 6.85978532e-02 1.06674090e-01 -5.41578531e-01
1.30441159e-01 -4.17944670e-01 -1.19983399e+00 -3.65425617e-01
-3.55302930e-01 1.73147768e-02 8.69805932e-01 8.78184438e-01
8.96848679e-01 7.34489083e-01 5.53371549e-01 -4.84554470e-01
-2.72887141e-01 -7.14472234e-01 -7.56195247e-01 4.08990541e-03
3.90262306e-01 -5.39207399e-01 -2.67184526e-01 -4.09604549e-01] | [5.951557636260986, 5.694282054901123] |
7165f6c0-d995-4266-9ecf-ed8398d0b787 | physics-informed-machine-learning-of | 2208.00880 | null | https://arxiv.org/abs/2208.00880v1 | https://arxiv.org/pdf/2208.00880v1.pdf | Physics-informed Machine Learning of Parameterized Fundamental Diagrams | Fundamental diagrams describe the relationship between speed, flow, and density for some roadway (or set of roadway) configuration(s). These diagrams typically do not reflect, however, information on how speed-flow relationships change as a function of exogenous variables such as curb configuration, weather or other exogenous, contextual information. In this paper we present a machine learning methodology that respects known engineering constraints and physical laws of roadway flux - those that are captured in fundamental diagrams - and show how this can be used to introduce contextual information into the generation of these diagrams. The modeling task is formulated as a probe vehicle trajectory reconstruction problem with Neural Ordinary Differential Equations (Neural ODEs). With the presented methodology, we extend the fundamental diagram to non-idealized roadway segments with potentially obstructed traffic data. For simulated data, we generalize this relationship by introducing contextual information at the learning stage, i.e. vehicle composition, driver behavior, curb zoning configuration, etc, and show how the speed-flow relationship changes as a function of these exogenous factors independent of roadway design. | ['Chase Dowling', 'Andisheh Ranjbari', 'Vinay Amatya', 'Thomas Maxner', 'James Koch'] | 2022-08-01 | null | null | null | null | ['physics-informed-machine-learning', 'machine-learning', 'machine-learning'] | ['graphs', 'methodology', 'miscellaneous'] | [ 1.48216158e-01 -2.26343218e-02 -1.85427785e-01 -4.02075261e-01
5.63210621e-02 -5.04149795e-01 7.77096212e-01 1.17460288e-01
-3.51859003e-01 8.86768401e-01 -1.82868645e-01 -8.30316722e-01
-7.15954721e-01 -9.96333957e-01 -9.71786439e-01 -6.64991736e-01
-6.53582141e-02 2.99578547e-01 1.94301426e-01 -5.74510813e-01
3.83919239e-01 1.00685775e+00 -1.90980101e+00 -4.11058664e-01
8.41550767e-01 4.54996467e-01 2.75023729e-02 8.03083777e-01
-1.13040879e-01 4.69790518e-01 -1.81998476e-01 -3.92227136e-02
2.72088289e-01 -1.73552364e-01 -4.93821174e-01 6.35120571e-02
2.18713716e-01 -1.88455448e-01 -6.12924278e-01 6.67240441e-01
2.18282938e-02 4.29651737e-01 1.18642628e+00 -1.38434005e+00
2.29669996e-02 2.65668660e-01 1.71284035e-01 2.62846559e-01
-1.08036146e-01 4.19558793e-01 5.54901600e-01 -3.00400764e-01
4.70634937e-01 1.22607422e+00 6.72636569e-01 2.19731301e-01
-1.39390206e+00 -3.89858395e-01 3.26404870e-01 2.63170272e-01
-1.24823141e+00 -1.90912113e-01 8.40650201e-01 -8.68819356e-01
7.44798720e-01 3.03322196e-01 7.83701003e-01 7.89426625e-01
5.88851511e-01 3.49560201e-01 7.76376665e-01 -3.85964751e-01
4.11300451e-01 5.57054698e-01 4.22307670e-01 4.33898777e-01
2.90466279e-01 6.33109450e-01 1.13819882e-01 3.73874843e-01
6.37651205e-01 -2.47043580e-01 -1.02322519e-01 -6.17870808e-01
-6.48490965e-01 7.59364903e-01 2.95242876e-01 1.79478660e-01
-2.41665691e-01 2.00027123e-01 2.82484800e-01 4.99059439e-01
2.81849466e-02 4.42509919e-01 -5.50012410e-01 -1.31945133e-01
-5.80110610e-01 7.99789488e-01 8.20346236e-01 9.34858441e-01
1.34409440e+00 2.95515865e-01 2.52766550e-01 5.82871616e-01
1.62302464e-01 6.75929308e-01 6.30412102e-02 -9.89353955e-01
3.82365763e-01 3.23105961e-01 4.17201519e-01 -1.09075320e+00
-7.27509916e-01 -3.06540608e-01 -5.26046336e-01 4.06131238e-01
6.02513671e-01 -5.01344740e-01 -1.00650179e+00 1.71189988e+00
3.06067944e-01 8.64832327e-02 -1.80031821e-01 6.77177429e-01
3.01073670e-01 7.58906543e-01 5.42670786e-02 -2.21914321e-01
7.84096718e-01 -4.14723486e-01 -7.11273015e-01 -2.10832786e-02
8.38366687e-01 -1.76033422e-01 8.43904793e-01 2.13718444e-01
-9.49691355e-01 -6.64601564e-01 -9.78544176e-01 2.51455575e-01
-1.13854253e+00 -2.12884918e-02 3.66027474e-01 7.98622012e-01
-8.23766768e-01 8.85478079e-01 -7.77401268e-01 -4.22516793e-01
-1.34765118e-01 4.04944569e-01 -1.41425073e-01 2.50437796e-01
-1.48321164e+00 1.40012240e+00 7.30900988e-02 3.93683970e-01
-8.58982444e-01 -1.09744537e+00 -8.10425520e-01 -1.28466502e-01
2.99987793e-01 -6.54709697e-01 1.17737460e+00 -5.28472841e-01
-1.54398167e+00 3.20840597e-01 -4.81941476e-02 -2.56648421e-01
7.88938224e-01 7.06072524e-02 -6.33290887e-01 -2.36868292e-01
-1.92767322e-01 4.22222286e-01 6.17075562e-01 -1.52531540e+00
-6.23313963e-01 -5.80364503e-02 2.75015473e-01 -1.07320957e-02
3.58564973e-01 -6.74246371e-01 -5.60210124e-02 -8.26062076e-03
-3.23571295e-01 -1.08424878e+00 -5.26099801e-01 -8.05756822e-02
-5.33224940e-01 5.06373048e-02 8.25207114e-01 -4.03778851e-01
1.16068304e+00 -1.84224963e+00 1.07409552e-01 7.72398949e-01
-1.82406217e-01 3.23395431e-01 -1.59036759e-02 6.72565103e-01
-2.76845872e-01 1.59878191e-03 -5.29906929e-01 1.04541615e-01
7.63315111e-02 4.65911835e-01 -9.94823128e-02 4.80759859e-01
3.20410222e-01 7.23016858e-01 -8.22832882e-01 -8.67507383e-02
7.89240062e-01 7.03497469e-01 -4.82638538e-01 -8.08559880e-02
-1.00322947e-01 4.41109478e-01 -6.30319893e-01 -7.53256306e-02
6.92818522e-01 7.01809943e-01 -2.07577303e-01 -1.26478821e-02
-7.42305577e-01 1.40142158e-01 -1.30139911e+00 9.08636630e-01
-9.68790948e-01 9.65804160e-01 -3.87348682e-02 -1.43296409e+00
1.28767395e+00 -1.13380320e-01 4.49951619e-01 -8.75094593e-01
3.72228712e-01 -4.77209799e-02 1.80079848e-01 -8.65759552e-01
4.69235152e-01 -5.31874001e-02 -2.02556867e-02 9.38493907e-02
-2.51607180e-01 -3.00300866e-01 3.17789257e-01 -1.43798649e-01
7.33513415e-01 -6.28012642e-02 5.32933101e-02 -4.52509314e-01
6.88057005e-01 2.27090970e-01 2.36774117e-01 6.36232436e-01
1.59366652e-01 1.15059264e-01 9.34565961e-01 -4.26630735e-01
-1.49944925e+00 -1.00331330e+00 -5.98391473e-01 3.03429753e-01
2.02621013e-01 3.69812245e-03 -8.33450198e-01 1.33486077e-01
3.76990229e-01 1.02035236e+00 -9.17937815e-01 -5.37373126e-01
-7.04767048e-01 -7.43004441e-01 2.51036972e-01 2.35662252e-01
2.38296062e-01 -6.13975406e-01 -7.55064368e-01 3.79559070e-01
4.02255595e-01 -9.08559084e-01 1.05019443e-01 7.84635171e-02
-6.97203636e-01 -1.22733986e+00 -1.73631057e-01 -3.61631185e-01
6.45732284e-01 -1.10384099e-01 7.78310299e-01 1.05924346e-01
-4.71312791e-01 4.59989518e-01 3.77837196e-02 -4.29020345e-01
-5.80988228e-01 9.99087617e-02 2.13904381e-02 2.80038387e-01
5.57971234e-03 -6.73019826e-01 -4.45331067e-01 5.27260900e-01
-8.26180398e-01 -1.05978446e-02 2.42371097e-01 3.83657724e-01
4.20293987e-01 5.02448678e-01 7.62461424e-01 -1.03728962e+00
6.14560902e-01 -6.90645814e-01 -8.08979809e-01 -4.28656414e-02
-7.00966299e-01 2.32022166e-01 6.93231940e-01 -4.55225855e-01
-1.09683454e+00 1.90906659e-01 -2.78068602e-01 -8.74797851e-02
-6.54721916e-01 2.95632213e-01 -6.96922660e-01 -4.83841598e-02
5.36376476e-01 4.64751199e-02 5.69810392e-03 -4.54314798e-01
5.23412645e-01 4.66014653e-01 5.12248278e-01 -6.81448221e-01
1.03772569e+00 3.72566432e-01 5.39597094e-01 -9.73618507e-01
-1.03192799e-01 -1.54047981e-01 -1.09074724e+00 -6.88763440e-01
5.39242148e-01 -4.43702549e-01 -9.22327399e-01 4.00458932e-01
-8.15087736e-01 -7.11098969e-01 -2.50413805e-01 5.66938877e-01
-8.96314025e-01 -2.89622277e-01 4.51114252e-02 -1.26966369e+00
6.08498871e-01 -1.11829841e+00 4.64375317e-01 2.62672335e-01
-1.08173497e-01 -1.48390758e+00 2.59385072e-02 -2.24872246e-01
5.39035261e-01 5.42996109e-01 1.35364115e+00 -1.42848849e-01
-6.80991650e-01 -2.61197895e-01 2.67270003e-02 2.83414543e-01
8.74776170e-02 4.28953797e-01 -5.78247368e-01 1.67769939e-01
-3.44185919e-01 5.50779641e-01 5.51685154e-01 6.41890705e-01
9.42231357e-01 -3.48785877e-01 -6.56241059e-01 5.88460624e-01
1.50705981e+00 5.70147872e-01 5.94241619e-01 3.69593978e-01
6.50602520e-01 1.05839133e+00 4.11437720e-01 9.28512365e-02
3.66834670e-01 7.05593586e-01 3.95534575e-01 -2.40802899e-01
1.59365565e-01 -3.66099745e-01 7.65534192e-02 3.18445683e-01
-8.90571102e-02 -4.57450468e-03 -9.90674496e-01 7.38415301e-01
-1.70570815e+00 -9.19470012e-01 -6.94695473e-01 2.22712517e+00
2.35799447e-01 3.70811224e-01 4.43473399e-01 2.44274437e-01
6.74308360e-01 -2.19452858e-01 -7.57950425e-01 -1.01407802e+00
7.15044215e-02 -2.03457385e-01 1.09991682e+00 1.02165306e+00
-5.58337152e-01 3.90651077e-01 6.80340195e+00 3.35234910e-01
-1.17196715e+00 -2.38912299e-01 3.57060373e-01 2.04854891e-01
-6.70342863e-01 6.85762614e-02 -7.99129426e-01 4.21967685e-01
1.38087666e+00 -3.94814789e-01 7.85410225e-01 4.12582010e-01
8.19001794e-01 -2.86547333e-01 -1.10591745e+00 4.25161213e-01
-3.54759276e-01 -1.13954806e+00 -3.76166441e-02 2.13790447e-01
5.21085978e-01 -3.33500087e-01 9.65469852e-02 3.09909642e-01
1.99031204e-01 -8.61598909e-01 5.25278032e-01 9.95605350e-01
5.33838391e-01 -9.59066272e-01 5.22880256e-01 5.41288972e-01
-1.05803597e+00 -2.10876882e-01 1.05204312e-02 -2.06791893e-01
4.95483637e-01 4.22540694e-01 -6.90608203e-01 5.21008432e-01
1.41461328e-01 6.31749332e-01 -3.22689056e-01 1.07367158e+00
-3.69538255e-02 6.04099333e-01 -4.47976083e-01 -2.00926930e-01
4.65252399e-01 -3.46820056e-01 6.28696620e-01 1.26433885e+00
2.43102893e-01 -3.01503073e-02 -3.85352701e-01 1.18333292e+00
6.62795603e-01 -1.89925134e-02 -9.45966363e-01 4.31154162e-01
3.38926047e-01 9.37851250e-01 -5.01784801e-01 9.05263871e-02
-3.18902612e-01 -1.74863413e-01 -1.75680086e-01 9.29456592e-01
-8.67979527e-01 -4.71343726e-01 1.20290661e+00 7.33597636e-01
1.77464947e-01 -2.92064488e-01 -1.38266057e-01 -4.91658449e-01
-2.40342870e-01 -1.75267264e-01 -2.75030464e-01 -7.02523649e-01
-8.63827348e-01 1.25912443e-01 8.07311058e-01 -1.02003443e+00
-4.23066437e-01 -8.72838914e-01 -8.28887224e-01 9.75852013e-01
-1.69107580e+00 -6.42348647e-01 -2.75951046e-02 4.23406363e-01
2.36860693e-01 -9.56029296e-02 1.84223428e-01 3.87247592e-01
-7.82130480e-01 3.27502638e-01 3.47763956e-01 -1.56711057e-01
-1.14120372e-01 -1.11743009e+00 5.68226457e-01 5.38267910e-01
-5.46365261e-01 5.00449240e-01 1.19325829e+00 -4.96483028e-01
-1.31189930e+00 -1.23116291e+00 6.26120567e-01 -5.29047191e-01
7.94262052e-01 -4.51145023e-01 -9.84066784e-01 5.93295157e-01
-2.88575083e-01 -2.19758987e-01 1.22124068e-01 -4.36146222e-02
1.43473908e-01 -4.28844929e-01 -9.42407370e-01 5.79057157e-01
9.37898457e-01 -4.85544026e-01 -3.52219194e-01 -1.59227531e-02
4.76924390e-01 -1.66045114e-01 -5.27885139e-01 4.09772217e-01
5.95113039e-01 -7.07259834e-01 1.07804084e+00 -8.86945903e-01
1.22054569e-01 -3.85131896e-01 1.34886220e-01 -1.66887462e+00
-2.21914053e-01 -6.45461977e-01 1.79645568e-01 9.20471251e-01
6.46245778e-01 -9.46400583e-01 4.66836959e-01 8.77403438e-01
-2.94686556e-01 -8.36181760e-01 -1.05891585e+00 -7.81336129e-01
6.41165257e-01 -8.12948823e-01 7.56074190e-01 5.27979374e-01
-1.87290505e-01 4.71559614e-02 -2.22387180e-01 3.60120088e-01
3.99569064e-01 -3.56711775e-01 9.56649482e-01 -1.31030333e+00
1.46017790e-01 -5.91515541e-01 -5.39202273e-01 -9.48153853e-01
4.11622614e-01 -5.87765038e-01 1.12205282e-01 -1.38387311e+00
-5.19488275e-01 -8.66121769e-01 1.05034458e-02 -1.83745489e-01
2.35991985e-01 -4.67196792e-01 -7.95697495e-02 -3.34019899e-01
4.07130569e-01 3.31915945e-01 1.30610216e+00 3.19401138e-02
-4.00677025e-01 2.24815518e-01 -2.29894012e-01 5.91600955e-01
8.72136950e-01 -3.59418958e-01 -8.61429930e-01 -7.42874593e-02
2.18679667e-01 2.73288339e-01 4.99024957e-01 -9.81363893e-01
2.29916453e-01 -5.46845853e-01 -1.34009168e-01 -5.81810892e-01
2.49221504e-01 -1.02689946e+00 4.32340801e-01 3.65833521e-01
-4.33541656e-01 2.76093721e-01 5.36083937e-01 5.84686100e-01
1.24919645e-01 -2.60491133e-01 5.73909819e-01 1.45691916e-01
-5.13470769e-01 1.08875707e-01 -1.00502777e+00 -2.43414730e-01
1.09644473e+00 -6.57546163e-01 -1.20611012e-01 -4.31821585e-01
-8.36568713e-01 4.28764999e-01 3.24113399e-01 5.04518390e-01
1.71879649e-01 -9.73966002e-01 -5.41749716e-01 4.37274367e-01
-9.47010741e-02 -1.51204675e-01 4.14406240e-01 7.15317249e-01
-4.66310591e-01 7.83666968e-01 -4.69713658e-01 -4.58511710e-01
-6.56892896e-01 6.21051311e-01 9.56532359e-01 2.34091654e-01
-5.30181766e-01 1.51254296e-01 6.89860284e-02 -5.78224957e-01
-6.02522381e-02 -7.48110175e-01 -3.34218115e-01 1.12892702e-01
1.12881280e-01 5.23692071e-01 5.77425733e-02 -9.42945778e-01
-1.38653377e-02 9.51265097e-01 4.52519089e-01 -3.94597650e-02
1.07576597e+00 -5.33620119e-01 6.49007976e-01 9.74512458e-01
1.10769391e+00 -4.24116701e-01 -1.43366349e+00 2.25169718e-01
-2.51191914e-01 -3.45918864e-01 1.80600911e-01 -5.60842276e-01
-9.68230009e-01 1.03990686e+00 4.71055388e-01 4.97119993e-01
8.28758538e-01 -2.67704368e-01 5.95569015e-01 3.95001531e-01
1.42025128e-01 -1.14878142e+00 -6.44178092e-01 4.33757365e-01
6.23957455e-01 -6.71655416e-01 -2.89784133e-01 -3.83002967e-01
-1.04917318e-01 1.10278833e+00 4.83969420e-01 -1.69580817e-01
1.05044425e+00 4.00792181e-01 -6.44184873e-02 4.90740165e-02
-7.49812543e-01 -3.06782961e-01 1.64567664e-01 7.90738881e-01
-3.60461324e-02 2.29863077e-01 -2.70355582e-01 6.35012835e-02
-4.42014247e-01 -1.55125529e-01 7.52388299e-01 5.27112603e-01
-4.46018070e-01 -9.03391182e-01 -2.51587182e-01 4.09720540e-01
3.87161404e-01 3.55534911e-01 5.09745255e-02 1.45363736e+00
4.47769016e-01 7.49891341e-01 5.73861122e-01 -2.81218976e-01
9.72645104e-01 1.22108690e-01 2.63228655e-01 -2.60644734e-01
-2.04535335e-01 -5.84459782e-01 3.56582642e-01 -2.95187026e-01
-1.95552692e-01 -1.06065643e+00 -1.23643827e+00 -6.50494158e-01
-1.79782614e-01 1.32196650e-01 9.45204496e-01 1.27451670e+00
-3.04104984e-02 7.01014161e-01 6.33460462e-01 -8.89744580e-01
3.63841280e-02 -5.17182946e-01 -5.42372346e-01 2.41914093e-01
8.87528896e-01 -1.19893408e+00 -6.13771737e-01 8.30765292e-02] | [5.628015518188477, 1.2593824863433838] |
998ad58c-8832-44ea-b9eb-f9d16d16bce2 | neural-code-summarization-how-far-are-we | 2107.07112 | null | https://arxiv.org/abs/2107.07112v2 | https://arxiv.org/pdf/2107.07112v2.pdf | On the Evaluation of Neural Code Summarization | Source code summaries are important for program comprehension and maintenance. However, there are plenty of programs with missing, outdated, or mismatched summaries. Recently, deep learning techniques have been exploited to automatically generate summaries for given code snippets. To achieve a profound understanding of how far we are from solving this problem and provide suggestions to future research, in this paper, we conduct a systematic and in-depth analysis of 5 state-of-the-art neural code summarization models on 6 widely used BLEU variants, 4 pre-processing operations and their combinations, and 3 widely used datasets. The evaluation results show that some important factors have a great influence on the model evaluation, especially on the performance of models and the ranking among the models. However, these factors might be easily overlooked. Specifically, (1) the BLEU metric widely used in existing work of evaluating code summarization models has many variants. Ignoring the differences among these variants could greatly affect the validity of the claimed results. Furthermore, we conduct human evaluations and find that the metric BLEU-DC is most correlated to human perception; (2) code pre-processing choices can have a large (from -18\% to +25\%) impact on the summarization performance and should not be neglected. We also explore the aggregation of pre-processing combinations and boost the performance of models; (3) some important characteristics of datasets (corpus sizes, data splitting methods, and duplication ratios) have a significant impact on model evaluation. Based on the experimental results, we give actionable suggestions for evaluating code summarization and choosing the best method in different scenarios. We also build a shared code summarization toolbox to facilitate future research. | ['Hongbin Sun', 'Dongmei Zhang', 'Hongyu Zhang', 'Shi Han', 'Junjie Chen', 'Lun Du', 'Yanlin Wang', 'Ensheng Shi'] | 2021-07-15 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 2.46233176e-02 -2.47969061e-01 -2.73721606e-01 -2.84165472e-01
-6.60533369e-01 -4.10824805e-01 2.14426905e-01 3.94468009e-01
-2.49982044e-01 3.70195150e-01 5.88645637e-01 -4.85307425e-01
-5.88323660e-02 -5.09975195e-01 -6.34064615e-01 -3.05940449e-01
1.04123158e-02 -1.42671123e-01 1.47120342e-01 -2.68950641e-01
7.35415161e-01 -1.11112177e-01 -1.63880706e+00 4.01076913e-01
1.46772587e+00 3.59448463e-01 4.46008205e-01 5.87281287e-01
-3.99838924e-01 9.56169486e-01 -1.10865629e+00 -5.23195982e-01
-1.03992097e-01 -6.27268374e-01 -8.33770812e-01 -1.37582093e-01
2.37806588e-01 -2.93432891e-01 -3.09399591e-04 1.19343531e+00
5.77603459e-01 -1.82485640e-01 6.29178762e-01 -1.02175820e+00
-8.67881000e-01 1.06254923e+00 -7.05418944e-01 3.22012246e-01
4.00498271e-01 3.33826035e-01 9.04485524e-01 -4.39736009e-01
3.51857156e-01 9.78929162e-01 7.40338147e-01 3.15183014e-01
-8.60428333e-01 -5.65005839e-01 1.65589392e-01 2.53192037e-01
-1.10702431e+00 -4.26662534e-01 7.05140233e-01 -6.39162958e-01
1.15217090e+00 3.31582397e-01 3.77311021e-01 9.88971770e-01
4.37312335e-01 6.38025463e-01 6.20232105e-01 -4.97390866e-01
1.78148881e-01 1.07373334e-01 3.72548074e-01 5.86629748e-01
6.51239872e-01 -4.32799041e-01 -2.63156265e-01 -2.35300541e-01
2.02913299e-01 -1.89923137e-01 -4.06585872e-01 -1.61303893e-01
-1.22729611e+00 8.01296592e-01 2.84965932e-01 4.70966727e-01
-1.49128258e-01 1.52076140e-01 9.51056004e-01 2.77574092e-01
1.63091198e-01 9.23505962e-01 -3.01300138e-01 -5.12530029e-01
-1.02323544e+00 3.02649379e-01 7.35341728e-01 1.05013347e+00
5.60028017e-01 1.26163393e-01 -4.91267115e-01 9.32253480e-01
1.16336085e-01 1.64578885e-01 8.71477187e-01 -7.70248890e-01
8.14088583e-01 8.20730209e-01 1.18448868e-01 -1.14046013e+00
-2.97413141e-01 -5.34380496e-01 -6.31275892e-01 -6.67582452e-02
2.52074897e-01 -1.72414765e-01 -4.69682962e-01 1.53660059e+00
-4.56639916e-01 -4.62299645e-01 -2.65548564e-02 6.36707664e-01
1.12374198e+00 4.80481058e-01 4.76063043e-02 -1.79411858e-01
1.22595060e+00 -1.15313375e+00 -6.94061756e-01 -4.34515387e-01
8.67434144e-01 -1.00919962e+00 1.42291474e+00 1.42876774e-01
-1.06605136e+00 -7.19047606e-01 -1.21613908e+00 -1.72707681e-02
-4.81170155e-02 4.74870056e-01 5.86301148e-01 6.85289323e-01
-1.01930082e+00 6.86869919e-01 -8.91266584e-01 -5.37462950e-01
1.07210331e-01 -9.46373027e-03 1.30941257e-01 3.10964659e-02
-8.15408111e-01 8.70818377e-01 4.45991665e-01 -1.38391867e-01
-6.53794646e-01 -4.56164151e-01 -7.61245430e-01 4.47507113e-01
3.85938913e-01 -7.26150692e-01 1.35744619e+00 -9.36660349e-01
-1.05590916e+00 4.46625113e-01 -2.36306459e-01 -3.05227876e-01
3.80089879e-01 -2.49229193e-01 -2.75379956e-01 -3.72533381e-01
1.46751046e-01 3.34035367e-01 2.59153336e-01 -1.05855179e+00
-4.84909445e-01 -1.17252536e-01 1.94848105e-01 1.48681059e-01
-5.46456754e-01 1.53881952e-01 -4.42966461e-01 -7.79335439e-01
-2.15627298e-01 -8.08532894e-01 -9.19583663e-02 -5.44770539e-01
-4.13777053e-01 -2.45534375e-01 2.00167477e-01 -8.18526804e-01
1.94020677e+00 -2.15518475e+00 1.10881343e-01 -4.05570120e-01
1.05661750e-01 3.76334816e-01 -3.91845673e-01 7.78726399e-01
9.78280753e-02 4.53775197e-01 -2.84290314e-01 -2.52357960e-01
-3.36347260e-02 -2.67325550e-01 -2.81646609e-01 4.52120639e-02
3.44420075e-02 7.53763556e-01 -8.24471295e-01 -4.46349651e-01
-1.78585798e-01 1.58640265e-01 -6.94362760e-01 2.63788640e-01
-2.87824929e-01 1.94042809e-02 -4.66124594e-01 4.76119071e-01
5.05325317e-01 -2.69095182e-01 -6.51179627e-02 -8.31707492e-02
-1.82474464e-01 6.50371671e-01 -8.78852189e-01 1.65935731e+00
-5.40767491e-01 8.58032942e-01 -3.48919690e-01 -7.18581915e-01
9.16813850e-01 4.02153693e-02 7.31709674e-02 -6.17748737e-01
5.63553460e-02 1.24512136e-01 4.17731851e-01 -9.52438414e-01
1.00752437e+00 4.28303808e-01 -1.45621644e-03 6.09149814e-01
-3.02460104e-01 7.30412826e-02 6.74196541e-01 2.09892556e-01
1.25257623e+00 -3.59910689e-02 4.63154525e-01 -3.63512695e-01
2.76452959e-01 3.47765423e-02 6.33877933e-01 8.77879262e-01
-3.31611961e-01 7.23580956e-01 9.65866864e-01 -1.95009530e-01
-8.34090769e-01 -4.49125439e-01 1.35464087e-01 1.10547912e+00
1.19351141e-01 -8.42253745e-01 -8.99923205e-01 -6.93937480e-01
-3.44793975e-01 1.14884031e+00 -4.35157120e-01 -2.83422232e-01
-4.62174863e-01 -9.46102262e-01 7.30237186e-01 6.46612883e-01
4.65375096e-01 -1.06481111e+00 -8.10862839e-01 6.60453364e-02
-5.96582234e-01 -6.66520178e-01 -5.97107291e-01 -2.30764970e-02
-9.67837989e-01 -1.04177940e+00 -5.22064984e-01 -6.64241612e-01
6.73867822e-01 6.72002196e-01 1.27232611e+00 5.50579607e-01
2.87642032e-01 1.02569357e-01 -6.55891299e-01 -5.66807151e-01
-9.95464444e-01 4.18548584e-01 -4.39105392e-01 -7.38802016e-01
4.62531149e-01 -3.27523947e-01 -5.15052676e-01 2.50674963e-01
-9.09380853e-01 1.41615599e-01 8.07947159e-01 5.83087921e-01
-1.52024813e-03 1.87119991e-01 7.32353866e-01 -1.07595706e+00
1.17155933e+00 -5.30079365e-01 -3.26134622e-01 5.16746521e-01
-6.88201010e-01 2.43795201e-01 7.13499308e-01 -3.31397086e-01
-1.05055332e+00 -5.18458545e-01 -1.37407646e-01 5.81279881e-02
-5.55050075e-02 9.55579340e-01 -1.02434352e-01 3.40640157e-01
9.87993598e-01 2.62970269e-01 -2.21839622e-02 -5.02804935e-01
-2.91947182e-02 8.38968039e-01 1.41214788e-01 -5.12878835e-01
3.60308141e-01 -3.35495509e-02 -7.38216519e-01 -6.19413972e-01
-4.52947646e-01 -1.67701527e-01 -3.12527984e-01 1.57053888e-01
5.52583873e-01 -8.87975097e-01 -7.39617497e-02 5.11906087e-01
-1.24260426e+00 -2.20739841e-01 2.33523160e-01 3.17627043e-01
-1.77589163e-01 8.29004884e-01 -6.99983656e-01 -4.53641504e-01
-4.71446484e-01 -1.72708189e+00 6.43723190e-01 3.95980179e-01
-7.08592176e-01 -6.89550340e-01 1.90535150e-02 4.89025176e-01
7.58435488e-01 6.99706888e-03 1.36434710e+00 -6.24768078e-01
-3.51749420e-01 3.34967151e-02 -2.06596568e-01 3.11264753e-01
1.34702042e-01 5.23508966e-01 -7.70922482e-01 -3.78055125e-01
6.55332655e-02 -1.15408145e-01 9.08087969e-01 3.62876743e-01
1.27944314e+00 -3.21407825e-01 -2.17311963e-01 2.98257142e-01
1.28065109e+00 3.88907254e-01 7.82709599e-01 3.97046477e-01
6.61022663e-01 5.48825741e-01 5.40788651e-01 4.62828100e-01
5.95104694e-01 4.53570455e-01 5.04888713e-01 2.32148692e-01
-1.29883915e-01 -8.71282890e-02 6.53886497e-01 1.35925257e+00
1.07951816e-02 -3.58906209e-01 -1.02004373e+00 6.75114393e-01
-1.63180697e+00 -8.52388859e-01 -2.50368029e-01 2.26125264e+00
9.17444229e-01 3.44759881e-01 1.05023138e-01 -4.02275324e-02
6.67986453e-01 2.48532042e-01 -4.73257124e-01 -6.90066457e-01
1.40535399e-01 -3.07480931e-01 1.50991738e-01 6.74658641e-02
-7.56215155e-01 5.48755527e-01 5.94613743e+00 8.22281718e-01
-1.15126121e+00 -9.32798535e-02 5.57883620e-01 6.76469803e-02
-5.92551947e-01 1.84152797e-01 -5.61928570e-01 7.98545599e-01
9.04844582e-01 -6.47476614e-01 3.70885491e-01 9.92989123e-01
1.30140603e-01 -2.84554183e-01 -1.29390049e+00 6.98071718e-01
2.69822747e-01 -1.09276462e+00 7.22943619e-02 -8.48472342e-02
9.33410108e-01 1.45149469e-01 -1.91809654e-01 7.13713765e-01
1.16180331e-01 -8.00169826e-01 7.38151133e-01 2.21224621e-01
3.99319887e-01 -6.33651078e-01 1.06131661e+00 4.85966325e-01
-8.97734463e-01 -1.96097821e-01 -5.80738485e-01 -1.70299411e-01
-2.35694781e-01 6.18319333e-01 -6.56764567e-01 5.34599066e-01
5.40378332e-01 6.34395659e-01 -1.16162539e+00 1.43522716e+00
-2.14898199e-01 7.76858151e-01 2.40004823e-01 -3.43198419e-01
-4.12246659e-02 8.28529075e-02 4.19197321e-01 1.41766548e+00
5.97800672e-01 -2.67103016e-01 -1.90303639e-01 1.20274043e+00
-9.55432430e-02 1.85779333e-01 -5.26321530e-01 -2.49591589e-01
6.89703941e-01 9.84260619e-01 -8.31911504e-01 -2.20647991e-01
-7.20779896e-01 6.12667918e-01 2.77684063e-01 4.10958827e-01
-1.03278804e+00 -7.27884412e-01 5.22050381e-01 4.23575565e-02
2.27333084e-01 -3.33665311e-02 -6.17246211e-01 -1.23306668e+00
2.52089351e-01 -1.20452952e+00 2.76407331e-01 -6.67334318e-01
-8.66621017e-01 6.46563828e-01 1.23511903e-01 -1.33370471e+00
-1.61121994e-01 -1.05516762e-01 -1.04599917e+00 5.97446561e-01
-1.25180447e+00 -5.37468135e-01 -5.32767296e-01 -2.61351287e-01
8.78024042e-01 -2.46957883e-01 5.41061819e-01 3.63481969e-01
-8.88269722e-01 8.06310952e-01 1.51702791e-01 1.32743150e-01
8.00190210e-01 -1.03887355e+00 7.00615644e-01 1.15679979e+00
-8.25756267e-02 1.07840252e+00 9.27506864e-01 -6.64514959e-01
-1.17006373e+00 -1.01319528e+00 8.97207975e-01 -5.11750758e-01
3.13440770e-01 -6.60858229e-02 -1.23984039e+00 5.09778380e-01
4.74311709e-01 -8.62950265e-01 6.79301918e-01 3.51703912e-01
-1.89364851e-01 5.80111444e-02 -7.61583865e-01 7.84488022e-01
7.86055088e-01 -2.48678610e-01 -6.64258718e-01 -2.34602224e-02
8.43480289e-01 -3.65143120e-01 -6.86241150e-01 3.81558090e-01
4.26708996e-01 -1.36234570e+00 4.50736761e-01 -2.34566927e-01
1.06046188e+00 -1.91351280e-01 1.20776691e-01 -1.52881420e+00
-3.07617396e-01 -1.61715150e-01 9.23148915e-02 1.60186720e+00
4.22009021e-01 -4.60500002e-01 3.71596307e-01 7.67263770e-01
-3.79082590e-01 -8.80904853e-01 -3.53316993e-01 -5.45638978e-01
2.43924916e-01 -2.36019805e-01 8.48545849e-01 8.35109890e-01
2.41921216e-01 3.84659916e-01 -1.81641296e-01 -2.02613935e-01
1.04211949e-01 2.18409747e-01 9.75229800e-01 -8.54575157e-01
-3.80395055e-01 -8.02931309e-01 -2.22336277e-02 -1.03528607e+00
7.89094865e-02 -7.07447350e-01 -5.07734977e-02 -1.66037762e+00
7.02403784e-01 -1.98545799e-01 -2.15200126e-01 3.78312528e-01
-6.24011636e-01 -4.77485865e-01 2.39466414e-01 4.27507401e-01
-6.69577122e-01 5.30136347e-01 1.02211714e+00 -1.38020888e-01
-2.63609886e-01 1.73460588e-01 -1.17879546e+00 6.41970277e-01
7.96053231e-01 -4.84701008e-01 -4.91110533e-01 -9.34111118e-01
4.49983984e-01 1.19662486e-01 -1.53101370e-01 -1.11771524e+00
9.21840742e-02 -1.40108496e-01 -4.42298763e-02 -4.32372063e-01
-4.60509986e-01 -3.47863913e-01 2.71075941e-03 5.70268035e-01
-5.17507732e-01 4.38683331e-01 4.19651061e-01 2.42996365e-01
-2.91717529e-01 -9.05843735e-01 5.39236426e-01 -2.53006458e-01
-6.56786382e-01 -3.60699862e-01 -4.65968817e-01 1.53789729e-01
6.39951825e-01 -2.60560632e-01 -6.36477947e-01 -4.40904081e-01
1.37766317e-01 2.89682925e-01 8.71113479e-01 7.52299190e-01
2.11778373e-01 -9.52617526e-01 -7.55305886e-01 7.73700848e-02
3.54784667e-01 -1.76044896e-01 3.15525711e-01 6.24580383e-01
-6.85823619e-01 5.57641149e-01 -2.71644443e-01 -3.68049264e-01
-1.21953535e+00 3.67863804e-01 4.36629206e-02 -2.94203550e-01
-2.55490065e-01 5.10647297e-01 9.76598561e-02 -2.80315995e-01
3.95252734e-01 -7.32186973e-01 -3.88938397e-01 6.80142492e-02
4.84625697e-01 6.24150634e-01 2.80234247e-01 -3.27266872e-01
-1.39437363e-01 3.12456429e-01 -3.52554232e-01 4.59260434e-01
1.19723439e+00 -7.08248317e-02 -2.79519796e-01 4.83905613e-01
1.01680338e+00 2.22242102e-01 -9.19129550e-01 2.01625889e-03
3.28277387e-02 -4.43793774e-01 -2.93510556e-01 -7.92048991e-01
-9.83706117e-01 9.17038620e-01 3.16203892e-01 3.57436776e-01
1.13233256e+00 -2.27861047e-01 6.50962591e-01 3.68689239e-01
3.23474705e-01 -9.46620941e-01 5.35709709e-02 5.58272719e-01
9.08331037e-01 -1.30145419e+00 1.84350669e-01 -1.23312347e-01
-7.86914587e-01 1.11382806e+00 9.56896782e-01 3.07720453e-01
-6.81052804e-02 1.11514460e-02 3.25458348e-02 -7.66353980e-02
-9.08249497e-01 2.46771857e-01 1.90751031e-01 2.75484324e-01
1.16595531e+00 -3.82817201e-02 -7.92706549e-01 7.66533971e-01
-3.27729166e-01 -1.10770270e-01 1.02366078e+00 8.38007033e-01
-4.76340592e-01 -1.06497788e+00 -2.57477105e-01 7.31834412e-01
-5.08924305e-01 -1.07760623e-01 -3.31348926e-01 7.18186259e-01
-1.46949574e-01 1.04390883e+00 -2.02541426e-01 -4.98168677e-01
3.71387899e-01 -6.53877780e-02 1.64239809e-01 -8.25183988e-01
-8.63454103e-01 -2.47474954e-01 6.32612407e-02 -2.97538102e-01
-1.27101496e-01 -6.35897875e-01 -1.06877303e+00 -4.45765316e-01
-5.94581187e-01 2.19900385e-01 5.95764875e-01 6.58400059e-01
6.78345621e-01 8.06534410e-01 2.74805933e-01 -5.10573804e-01
-8.23807061e-01 -1.36965561e+00 -7.83210993e-02 4.13922220e-01
6.52159303e-02 -4.61439490e-01 -4.67851013e-01 -1.50739364e-02] | [7.671576976776123, 7.938797473907471] |
c502ed40-5771-4947-b560-a14b3aabe76d | an-abstract-specification-of-voxml-as-an | 2305.13076 | null | https://arxiv.org/abs/2305.13076v1 | https://arxiv.org/pdf/2305.13076v1.pdf | An Abstract Specification of VoxML as an Annotation Language | VoxML is a modeling language used to map natural language expressions into real-time visualizations using commonsense semantic knowledge of objects and events. Its utility has been demonstrated in embodied simulation environments and in agent-object interactions in situated multimodal human-agent collaboration and communication. It introduces the notion of object affordance (both Gibsonian and Telic) from HRI and robotics, as well as the concept of habitat (an object's context of use) for interactions between a rational agent and an object. This paper aims to specify VoxML as an annotation language in general abstract terms. It then shows how it works on annotating linguistic data that express visually perceptible human-object interactions. The annotation structures thus generated will be interpreted against the enriched minimal model created by VoxML as a modeling language while supporting the modeling purposes of VoxML linguistically. | ['James Pustejovsky', 'Nikhil Krishnaswamy', 'Kiyong Lee'] | 2023-05-22 | null | null | null | null | ['human-object-interaction-detection'] | ['computer-vision'] | [-8.44840035e-02 7.92152762e-01 1.82115704e-01 -3.70980650e-01
1.28617093e-01 -5.13928592e-01 1.54430234e+00 4.40701872e-01
-2.45652646e-01 6.18170261e-01 6.35753214e-01 -3.92932177e-01
-1.75347686e-01 -8.75434875e-01 -2.71066695e-01 -2.51629800e-01
-4.13254946e-01 5.34521997e-01 1.87617064e-01 -5.68957508e-01
1.02145769e-01 7.17911243e-01 -1.92853463e+00 2.72127450e-01
4.06485796e-01 5.08678973e-01 6.38877869e-01 8.32976341e-01
-4.01053727e-01 1.62058210e+00 -6.43452466e-01 -4.11035717e-02
-2.37860739e-01 -6.07857823e-01 -1.14030993e+00 2.44487002e-01
-3.10455203e-01 -3.44233476e-02 1.00063205e-01 8.45477462e-01
8.48568305e-02 4.59399641e-01 4.79233712e-01 -1.78360534e+00
-8.17225277e-01 7.25761354e-01 2.58205026e-01 -4.78669047e-01
1.25717914e+00 1.90590009e-01 8.87165427e-01 -5.31142950e-01
1.40738285e+00 1.77098691e+00 3.32469009e-02 7.82950461e-01
-1.03708410e+00 3.00446689e-01 5.14923409e-02 2.25305825e-01
-1.33747244e+00 -2.77225584e-01 6.47588670e-01 -6.61880672e-01
1.41059685e+00 9.14979994e-01 1.14115512e+00 5.49845159e-01
-4.01338376e-02 7.73397028e-01 9.29210424e-01 -1.01920283e+00
3.64736915e-01 3.87124836e-01 2.53694326e-01 8.66194248e-01
-1.29301578e-01 1.95856825e-01 -4.13667798e-01 -3.92190367e-01
1.10176277e+00 -3.84622365e-01 2.71172971e-02 -5.49268842e-01
-1.66336095e+00 6.10460877e-01 4.29571897e-01 7.58218110e-01
-4.56967652e-01 6.38795614e-01 4.15964514e-01 6.11334890e-02
7.62425363e-02 5.62877417e-01 -1.01966068e-01 -3.14973563e-01
2.46274233e-01 6.04878187e-01 1.13542676e+00 1.24227738e+00
3.68666887e-01 -4.83217798e-02 1.96426943e-01 5.28304994e-01
1.07567930e+00 2.67029554e-01 2.41030678e-01 -1.32000875e+00
-3.87515277e-01 1.02335131e+00 6.46131456e-01 -8.84158850e-01
-6.34378195e-01 2.77423888e-01 1.09864078e-01 8.99644196e-01
2.13473380e-01 1.97935969e-01 -5.66307485e-01 1.62885237e+00
7.05785334e-01 -1.61770821e-01 7.06216872e-01 1.00511062e+00
1.50150287e+00 6.61998808e-01 8.22266340e-01 -4.62472811e-02
1.82702887e+00 -7.16958404e-01 -1.00692940e+00 -1.04250841e-01
1.19892418e+00 -1.92577407e-01 1.30658543e+00 -3.63433622e-02
-1.06821430e+00 -7.38694072e-02 -1.08228457e+00 -5.87080061e-01
-8.67532849e-01 -7.80149922e-02 1.01545835e+00 2.48185053e-01
-1.02363980e+00 7.83146266e-03 -7.84027457e-01 -1.06131852e+00
-8.81100744e-02 4.96838912e-02 -2.09031507e-01 6.11029744e-01
-9.79308009e-01 1.46192729e+00 6.96493268e-01 -6.86108097e-02
-9.24214005e-01 -3.84782478e-02 -1.43688715e+00 -1.79747716e-01
3.99929434e-01 -5.99518716e-01 1.45687819e+00 -8.68506491e-01
-1.56326842e+00 1.24553943e+00 1.95999682e-01 -4.76833075e-01
6.11362040e-01 3.63539681e-02 -7.74804950e-01 -1.37035087e-01
2.05902740e-01 6.76999629e-01 -1.57531545e-01 -1.74896920e+00
-7.24663317e-01 -3.06483060e-01 8.42931867e-01 5.15765727e-01
6.11270964e-01 4.77869660e-01 -1.59544319e-01 -3.50574940e-01
2.38963544e-01 -4.79281843e-01 -2.22454026e-01 2.27205321e-01
-3.28572571e-01 -6.92198575e-01 7.29214311e-01 -3.06818843e-01
7.88791060e-01 -2.14332151e+00 1.79625273e-01 1.54187396e-01
3.39138329e-01 -2.64652818e-01 4.18976955e-02 9.82070923e-01
1.48326457e-01 4.82165255e-02 5.08286357e-02 6.98566362e-02
8.96218956e-01 4.78574097e-01 -7.01254681e-02 4.12001222e-01
-4.11360264e-01 8.58819008e-01 -1.21561730e+00 -8.53874564e-01
9.42419291e-01 5.66380143e-01 -1.07974231e-01 4.06636000e-01
-6.29743814e-01 4.22476023e-01 -6.34362936e-01 4.85308707e-01
-1.52781948e-01 1.35971025e-01 5.91575444e-01 1.92995921e-01
-5.98019063e-01 4.40748513e-01 -1.23576021e+00 1.84404218e+00
-6.60524786e-01 6.72424912e-01 5.90107590e-02 -1.69979334e-01
9.58651543e-01 7.10779727e-01 6.97506294e-02 -4.14515287e-01
1.87172607e-01 1.19408384e-01 -1.27608761e-01 -1.35514307e+00
4.14921910e-01 -2.71219432e-01 -3.59280974e-01 5.86980522e-01
-1.73591763e-01 -6.74476862e-01 2.12804392e-01 3.28810811e-01
4.84672576e-01 9.35929298e-01 1.00832736e+00 -4.40947920e-01
5.41790187e-01 3.21399420e-01 9.66910087e-03 3.69761050e-01
-1.71351545e-02 -2.63963670e-01 4.00630504e-01 -6.72981262e-01
-8.96018982e-01 -1.06429231e+00 4.60115187e-02 1.44291639e+00
6.15794420e-01 -6.27354443e-01 -7.64992476e-01 8.72915760e-02
-5.42392969e-01 1.65413296e+00 -6.22502148e-01 1.86537027e-01
-4.64759111e-01 -1.79177478e-01 4.76264417e-01 2.24724963e-01
6.73224777e-02 -1.73111880e+00 -2.01547647e+00 3.14215243e-01
4.49241474e-02 -8.69491935e-01 4.25989360e-01 -1.24958813e-01
-3.34815949e-01 -9.91197288e-01 1.74593955e-01 -7.62399912e-01
4.88268018e-01 -3.36797535e-01 1.26754248e+00 4.71001029e-01
-4.05400217e-01 6.62269831e-01 -5.64615786e-01 -8.41652691e-01
-8.15193415e-01 -8.23111773e-01 -2.07220882e-01 -3.78683388e-01
2.76583999e-01 -3.94840688e-01 -4.47888114e-02 8.76123086e-02
-9.84612226e-01 6.87448919e-01 -5.65372527e-01 2.36404791e-01
-5.53143863e-03 -3.21022362e-01 4.33945516e-03 -3.89953971e-01
8.14166784e-01 -1.83971599e-01 -4.72745150e-01 4.28697467e-01
8.82694945e-02 -6.83279485e-02 -2.37788215e-01 -3.40430111e-01
-1.26915455e+00 -2.33471677e-01 2.36707017e-01 4.84343320e-01
-6.61490142e-01 6.64850116e-01 -7.12105691e-01 1.94481790e-01
5.99760830e-01 -3.05615067e-01 -7.69540220e-02 -3.57030094e-01
9.76406872e-01 5.75507820e-01 6.84691429e-01 -9.88758683e-01
-2.74581518e-02 5.83559394e-01 7.91277960e-02 -1.03534091e+00
1.73498213e-01 -7.54199997e-02 -6.73854232e-01 -6.84162319e-01
1.13201261e+00 -4.78986472e-01 -1.33471274e+00 -9.42761078e-02
-1.44075966e+00 -6.03711605e-01 -9.49730873e-01 5.03932178e-01
-1.24801755e+00 -6.61920533e-02 -3.44141185e-01 -1.48819530e+00
3.41207385e-01 -1.11488259e+00 1.10563731e+00 -1.98184788e-01
-9.37970638e-01 -1.10710180e+00 1.64811108e-02 -1.60265058e-01
-3.01017106e-01 8.33194137e-01 1.06083417e+00 -6.68939352e-01
-3.16664457e-01 -1.17825843e-01 5.02400510e-02 -5.19852579e-01
3.13008092e-02 1.34771660e-01 -7.18121469e-01 3.70338231e-01
-6.73276335e-02 -1.58651769e-01 -4.82489109e-01 8.34603906e-02
1.68594584e-01 -5.04950225e-01 -5.70809662e-01 -4.31133993e-02
1.62422431e+00 9.66724634e-01 8.60522687e-01 6.13988280e-01
5.69248080e-01 1.26378787e+00 7.90358245e-01 3.01789463e-01
5.49218833e-01 9.25124884e-01 5.20787597e-01 -1.25140265e-01
4.28498723e-03 -2.68156707e-01 5.77883013e-02 1.62197918e-01
-6.47919416e-01 -3.27111661e-01 -1.19817650e+00 3.77271444e-01
-2.26492810e+00 -1.00609255e+00 -6.38797998e-01 1.75444758e+00
4.29629564e-01 -4.66946721e-01 1.74232900e-01 -1.77050814e-01
4.65450257e-01 1.12246171e-01 -2.70421826e-03 -7.22783625e-01
-1.70080155e-01 -4.23119128e-01 -2.39340410e-01 1.07994509e+00
-5.92811227e-01 1.22694755e+00 6.23154306e+00 3.43923092e-01
-4.73578691e-01 5.17694533e-01 -2.70114928e-01 1.68330774e-01
-3.93648267e-01 9.49746370e-02 1.14268571e-01 -2.19608545e-01
6.47549093e-01 -2.64900088e-01 5.49113393e-01 6.74715638e-01
5.45319259e-01 -4.18900609e-01 -1.45715070e+00 7.94764578e-01
1.71742728e-03 -1.50565338e+00 3.75312716e-02 -8.24575499e-03
3.63047682e-02 -4.88888979e-01 -6.78676426e-01 -2.02069283e-01
8.68860543e-01 -1.09192479e+00 1.74156034e+00 7.08437204e-01
4.73351866e-01 -3.60033691e-01 3.58561546e-01 3.61741960e-01
-1.24417925e+00 1.37736142e-01 1.24471553e-01 -4.72000331e-01
6.32555783e-01 -4.75276291e-01 -8.03575218e-01 4.24368203e-01
4.16342705e-01 2.23035589e-01 -4.52334136e-02 7.08046257e-01
-3.93504202e-01 -1.91559196e-01 -2.94499725e-01 -5.81582606e-01
2.79631644e-01 -5.03109872e-01 1.01073301e+00 1.31773484e+00
9.94331483e-03 5.28979421e-01 2.61627555e-01 1.09510434e+00
6.80331588e-01 3.43564719e-01 -1.03523588e+00 8.61738771e-02
3.35666537e-01 7.15374231e-01 -9.48381007e-01 -7.42846668e-01
-2.15860188e-01 7.40939319e-01 -1.35530800e-01 5.74008346e-01
-7.36069441e-01 -3.39777827e-01 6.14849269e-01 8.53555575e-02
-6.32710576e-01 -3.42499435e-01 -5.46936728e-02 -6.63310289e-01
-3.54231715e-01 -4.62055475e-01 8.12291577e-02 -1.66380095e+00
-6.12881660e-01 7.18981385e-01 5.42484283e-01 -9.32176292e-01
-5.14038980e-01 -6.39376938e-01 -2.88147897e-01 6.04483426e-01
-1.65356740e-01 -1.78135788e+00 -1.26501992e-01 2.40115970e-01
3.91527534e-01 1.99988127e-01 1.38443434e+00 -3.54067594e-01
2.90307045e-01 -5.99921048e-01 -6.39151275e-01 -1.50601417e-01
-3.38413417e-01 -1.32439637e+00 1.42107323e-01 4.71454889e-01
2.41122261e-01 6.49938881e-01 1.42328501e+00 -6.47536516e-01
-1.31204236e+00 -3.62718165e-01 1.02506900e+00 -7.88196146e-01
8.51509988e-01 -4.43259567e-01 -6.04870558e-01 1.29056263e+00
6.61886811e-01 -1.74650043e-01 4.93833423e-01 -2.61090219e-01
-1.03430405e-01 6.52324080e-01 -1.43042445e+00 1.25127804e+00
1.44541264e+00 -7.40955710e-01 -1.12198389e+00 3.73942167e-01
8.55787694e-01 -5.39937377e-01 -8.32726777e-01 1.30178049e-01
5.78303039e-01 -9.37951803e-01 1.10228789e+00 -8.68828893e-01
-2.44483590e-01 -7.85156071e-01 -4.63441014e-01 -9.03329372e-01
3.49248154e-03 -6.05588317e-01 1.13735147e-01 9.58857954e-01
2.72687644e-01 -6.26499116e-01 -2.71792188e-02 9.44717467e-01
-2.08281457e-01 -7.03450888e-02 -9.32732165e-01 -5.52900612e-01
-5.76612234e-01 -9.43196118e-01 1.16089416e+00 1.11448348e+00
1.08320057e+00 2.56530315e-01 1.35419339e-01 3.81227247e-02
2.84638554e-01 -1.13491036e-01 7.79428959e-01 -1.52502215e+00
6.41050637e-02 -4.93315488e-01 -8.20397735e-01 -5.69432259e-01
2.59484947e-01 -8.09626579e-01 3.56837034e-01 -2.19477677e+00
-2.00476751e-01 -4.10397232e-01 5.44184625e-01 5.15516400e-01
6.48027122e-01 -1.50048435e-01 5.23215294e-01 1.78352207e-01
-6.10138655e-01 3.68398190e-01 1.32853043e+00 2.73211394e-02
-4.89118189e-01 -7.51209676e-01 -2.32581183e-01 1.17197955e+00
5.55465937e-01 -1.30274311e-01 -4.56294030e-01 -2.17493147e-01
5.25677145e-01 2.61969805e-01 1.05036974e+00 -6.90016210e-01
-2.20448477e-03 -7.36246526e-01 -3.80807996e-01 -1.02571040e-01
5.52563667e-01 -1.39935505e+00 9.78588164e-01 5.92552781e-01
-6.17474318e-01 -2.53714994e-02 1.51147291e-01 1.81163892e-01
1.32387266e-01 -4.11577106e-01 1.05892524e-01 -3.19803417e-01
-1.24045455e+00 -7.04024851e-01 -8.39063048e-01 -5.06146252e-01
1.38303339e+00 -4.16011453e-01 -2.52960503e-01 -2.88585693e-01
-1.42143118e+00 5.61802238e-02 8.59076917e-01 4.78026688e-01
5.51373124e-01 -1.23488569e+00 -3.72807026e-01 -1.27926975e-01
4.38610345e-01 -1.91820368e-01 -1.36697277e-01 2.89593905e-01
-1.15783775e+00 1.99016795e-01 -3.82297218e-01 -3.58542413e-01
-1.25773621e+00 7.07312405e-01 4.15797263e-01 4.94209498e-01
-9.87903833e-01 5.89667499e-01 3.47377121e-01 -7.09040642e-01
2.61005670e-01 -6.13358140e-01 -5.46486497e-01 -3.89260590e-01
5.89117825e-01 5.74093044e-01 -6.52867317e-01 -1.45591664e+00
-5.07791102e-01 3.17697197e-01 1.02524483e+00 -9.47262943e-01
1.15370166e+00 -4.75556076e-01 -7.67326176e-01 1.14628029e+00
6.51313901e-01 -1.37134224e-01 -8.45935583e-01 2.16510192e-01
4.52460438e-01 -4.75075394e-02 -4.88949090e-01 -9.39905226e-01
-7.11717308e-02 5.64795852e-01 6.53406620e-01 9.46818709e-01
4.26355243e-01 7.27486074e-01 -1.57805532e-01 3.35320055e-01
9.84350681e-01 -9.06714857e-01 -3.78749281e-01 8.61975327e-02
1.65138173e+00 -6.99185014e-01 3.06350216e-02 -8.03268611e-01
-8.41304004e-01 1.07730365e+00 3.45938206e-01 2.37237692e-01
4.85223323e-01 5.37597299e-01 4.36827332e-01 -1.09333742e+00
-6.05891287e-01 -4.36082453e-01 5.54449894e-02 1.05831206e+00
4.42967713e-01 7.50977278e-01 -5.00104129e-01 2.39125609e-01
-3.52886558e-01 2.51544684e-01 5.06445527e-01 1.47904980e+00
-4.79655951e-01 -6.86384976e-01 -5.27305186e-01 -2.51614332e-01
7.30299875e-02 2.55006403e-01 -6.15240693e-01 1.39566672e+00
5.01263022e-01 1.26616013e+00 4.35562104e-01 2.63336040e-02
5.37369132e-01 -1.01804331e-01 6.36996031e-01 -7.71670759e-01
-6.32927358e-01 -3.46789844e-02 7.45291054e-01 -6.61452889e-01
-1.20063174e+00 -5.28834760e-01 -2.28186893e+00 -1.19015105e-01
-1.43748727e-02 4.61043388e-01 8.52220237e-01 1.06314516e+00
-2.48754591e-01 6.05964482e-01 -4.90173697e-01 -1.04174042e+00
3.44898015e-01 -8.39475036e-01 -6.73576117e-01 3.24250638e-01
3.34460586e-01 -6.68490946e-01 -5.34430444e-02 4.57546055e-01] | [5.091229438781738, 0.5443795919418335] |
a2acb812-754f-45da-bf9a-75e45affa0a6 | unsupervised-abstractive-summarization-of | 2102.04490 | null | https://arxiv.org/abs/2102.04490v2 | https://arxiv.org/pdf/2102.04490v2.pdf | Unsupervised Abstractive Summarization of Bengali Text Documents | Abstractive summarization systems generally rely on large collections of document-summary pairs. However, the performance of abstractive systems remains a challenge due to the unavailability of parallel data for low-resource languages like Bengali. To overcome this problem, we propose a graph-based unsupervised abstractive summarization system in the single-document setting for Bengali text documents, which requires only a Part-Of-Speech (POS) tagger and a pre-trained language model trained on Bengali texts. We also provide a human-annotated dataset with document-summary pairs to evaluate our abstractive model and to support the comparison of future abstractive summarization systems of the Bengali Language. We conduct experiments on this dataset and compare our system with several well-established unsupervised extractive summarization systems. Our unsupervised abstractive summarization model outperforms the baselines without being exposed to any human-annotated reference summaries. | ['Taufiqul Jannat', 'Md. Saifur Rahman Chowdhury', 'Tahsin Tasnim Mim', 'Mir Tafseer Nayeem', 'Radia Rayan Chowdhury'] | 2021-01-26 | null | https://aclanthology.org/2021.eacl-main.224 | https://aclanthology.org/2021.eacl-main.224.pdf | eacl-2021-2 | ['unsupervised-extractive-summarization'] | ['natural-language-processing'] | [ 5.13283670e-01 4.24610734e-01 4.09783283e-03 -3.19040239e-01
-1.40495765e+00 -8.10615480e-01 8.77385795e-01 8.60359728e-01
-4.68900621e-01 7.46460319e-01 9.87272203e-01 -2.55618185e-01
1.54895335e-01 -5.03806353e-01 -4.57402736e-01 -2.74416894e-01
5.94834909e-02 9.06215787e-01 3.91189575e-01 -4.46653098e-01
6.97425783e-01 2.38959476e-01 -9.39635336e-01 4.62611794e-01
1.41361070e+00 6.90123811e-02 1.99178651e-01 1.28988731e+00
-1.54977709e-01 6.02804363e-01 -1.23835170e+00 -5.64281344e-01
-3.43196094e-02 -9.43580806e-01 -1.05351448e+00 1.21475324e-01
9.82517302e-01 -1.89734653e-01 -4.12498266e-01 9.26386654e-01
7.27641344e-01 3.27613443e-01 8.27133775e-01 -7.50567079e-01
-5.77926397e-01 1.06457603e+00 -4.41356182e-01 4.01472420e-01
6.08811140e-01 -2.03229755e-01 1.38152099e+00 -4.94325966e-01
8.36808681e-01 1.02232087e+00 3.43047887e-01 4.64075446e-01
-9.39211607e-01 -4.55474705e-02 3.65158841e-02 -5.79304732e-02
-8.07435393e-01 -7.90638506e-01 7.54796863e-01 1.20802738e-01
1.44017100e+00 7.54740536e-01 6.07506633e-01 6.87163830e-01
2.34486267e-01 1.16746974e+00 3.66412759e-01 -6.58253074e-01
1.90028220e-01 -3.46868724e-01 4.69154000e-01 6.44581914e-01
5.89567006e-01 -1.05462825e+00 -4.31823552e-01 -4.34134722e-01
5.26905358e-02 -3.83801877e-01 -3.52025956e-01 1.00256473e-01
-1.20124876e+00 7.40764380e-01 -4.47936095e-02 3.05468291e-01
-5.42675197e-01 -1.11895137e-01 9.33738172e-01 2.66458511e-01
9.41897094e-01 8.89730573e-01 -2.97693163e-01 -1.06394373e-01
-1.65973425e+00 3.63810122e-01 1.48049450e+00 1.22799289e+00
3.89355481e-01 1.09278396e-01 -5.64551353e-01 9.38229620e-01
-1.08409129e-01 5.34806013e-01 5.42303383e-01 -6.55597925e-01
1.09449792e+00 7.01473773e-01 -1.38830557e-01 -9.09469187e-01
-1.75740436e-01 -3.19552124e-01 -8.76122713e-01 -8.14318538e-01
-1.16619602e-01 -2.11232036e-01 -8.14681768e-01 9.04508352e-01
-3.82799320e-02 -3.87766331e-01 4.85149980e-01 4.43456441e-01
1.24745440e+00 1.07756245e+00 -3.32206100e-01 -7.76624084e-01
1.17068470e+00 -1.40482605e+00 -7.09170699e-01 -3.27727199e-01
6.55218303e-01 -9.05638814e-01 8.13632965e-01 6.08810931e-02
-1.37397802e+00 -1.36889622e-01 -1.08177507e+00 -3.40169907e-01
-4.84190546e-02 2.20646590e-01 2.15630025e-01 4.53892380e-01
-1.16411185e+00 4.76551563e-01 -1.02757621e+00 -9.03858304e-01
1.64190561e-01 2.31198281e-01 -4.24803704e-01 -2.34057046e-02
-7.00455844e-01 7.71795809e-01 7.46003747e-01 -1.53213814e-02
-4.65037495e-01 -2.94265836e-01 -1.10203493e+00 1.49500579e-01
4.28377628e-01 -9.66474473e-01 1.58543074e+00 -6.40454054e-01
-1.47470903e+00 6.60657227e-01 -3.86794150e-01 -6.62791729e-01
4.83842134e-01 -4.33196604e-01 -1.96075752e-01 7.02146113e-01
3.63875031e-01 1.70846552e-01 4.24839914e-01 -1.09238911e+00
-5.84870219e-01 -2.33481735e-01 -1.55376181e-01 5.10073423e-01
-2.92298287e-01 2.79633403e-01 -6.98852181e-01 -8.79902422e-01
-8.39335024e-02 -7.38878071e-01 -3.99405658e-01 -1.18058169e+00
-1.01784718e+00 -3.60176206e-01 8.41263831e-01 -1.18986285e+00
1.55106568e+00 -1.56473291e+00 2.03223929e-01 -1.34537265e-01
1.14479341e-01 4.73867953e-01 -4.38611716e-01 1.18602085e+00
4.03955281e-01 3.27651352e-01 -5.56589544e-01 -4.40706640e-01
-5.50839417e-02 1.36701927e-01 -3.98057193e-01 2.23890916e-01
1.51693910e-01 9.41662848e-01 -1.32760417e+00 -9.34597135e-01
-2.58161485e-01 -2.15982795e-01 -4.01059717e-01 3.48047942e-01
-4.10316348e-01 4.29024063e-02 -6.53794289e-01 4.18737590e-01
3.65541369e-01 2.44545966e-01 3.79607916e-01 -7.93537870e-02
-1.88333094e-01 8.52192044e-01 -5.41789591e-01 1.87507939e+00
-1.32339314e-01 7.14090884e-01 -2.11843133e-01 -1.09057999e+00
8.34038377e-01 2.49688685e-01 1.06559627e-01 -4.13558543e-01
-1.21820934e-01 3.81900340e-01 -1.44819170e-01 -4.23824757e-01
1.46959460e+00 2.25457892e-01 -5.38397253e-01 7.84229040e-01
2.24393442e-01 -7.52511442e-01 9.63976622e-01 1.03675628e+00
1.50080931e+00 -1.34611011e-01 5.50910771e-01 -3.66474718e-01
3.96994025e-01 4.56561536e-01 3.26283962e-01 1.13897717e+00
2.47944400e-01 9.17924702e-01 6.69797301e-01 4.86797132e-02
-1.14808989e+00 -7.16016591e-01 4.21797186e-01 8.95856380e-01
-1.91233128e-01 -1.04179478e+00 -9.20416355e-01 -1.05979609e+00
-4.44444805e-01 1.27956581e+00 -2.02926382e-01 6.33200034e-02
-1.04502475e+00 -5.20618379e-01 8.97040963e-01 4.59753007e-01
5.94000876e-01 -9.92589235e-01 -3.64299476e-01 2.70791590e-01
-4.23976660e-01 -1.01742399e+00 -8.24271202e-01 -5.75514287e-02
-9.45302784e-01 -8.60589027e-01 -7.80607820e-01 -8.51193070e-01
5.17639935e-01 3.08644384e-01 1.30868387e+00 4.63974252e-02
2.39574865e-01 6.41880751e-01 -5.62974513e-01 -5.81722140e-01
-1.07684076e+00 8.34198177e-01 -1.53838068e-01 -6.17978632e-01
4.84056175e-02 -3.38611186e-01 -3.70456874e-01 -3.11827302e-01
-1.04558504e+00 1.34812057e-01 6.98035479e-01 4.98145431e-01
3.19302648e-01 -1.81391746e-01 8.06256652e-01 -1.36206686e+00
1.22111356e+00 -1.70325726e-01 -2.09728882e-01 7.21703589e-01
-2.06545860e-01 1.50322333e-01 7.24872172e-01 -1.11573068e-02
-1.17017734e+00 -3.17732126e-01 -1.28329203e-01 4.11958039e-01
1.86266929e-01 9.82062578e-01 -1.49576768e-01 6.93544149e-01
6.12373471e-01 5.22444248e-01 -3.31792831e-01 -5.00872970e-01
6.10948443e-01 1.03542507e+00 8.64297569e-01 -4.01994377e-01
6.59463525e-01 -2.80688107e-02 -2.82168299e-01 -1.56230175e+00
-9.76313770e-01 -8.07754576e-01 -8.10036480e-01 2.13207021e-01
6.08271062e-01 -8.31394851e-01 1.13738529e-01 3.00029367e-01
-1.46376359e+00 -1.77999794e-01 -4.10853207e-01 1.96370721e-01
-5.19486189e-01 1.07787240e+00 -6.59625649e-01 -5.04650533e-01
-1.22646558e+00 -5.85618198e-01 1.15233505e+00 1.87670320e-01
-5.55159628e-01 -1.00878024e+00 4.13093448e-01 5.00393748e-01
7.19946399e-02 1.72385901e-01 9.17831659e-01 -1.48521495e+00
-8.85480121e-02 -5.98683894e-01 -7.41987452e-02 5.13234615e-01
3.30959946e-01 1.79443777e-01 -3.43224078e-01 -2.91281044e-01
-1.05988041e-01 -2.95235455e-01 1.29975462e+00 3.29024225e-01
5.37108481e-01 -1.03645432e+00 -2.32865766e-01 -9.54356641e-02
9.29045618e-01 -3.60981166e-01 4.51230019e-01 1.52104914e-01
7.64160752e-01 4.74419832e-01 5.46164453e-01 2.04838723e-01
7.01503217e-01 1.06599651e-01 -2.49466434e-01 1.51043847e-01
-2.45930940e-01 -4.32116836e-01 3.09184790e-01 1.79995251e+00
-2.28049576e-01 -1.07723129e+00 -8.43019366e-01 8.92358065e-01
-2.03441882e+00 -1.08876503e+00 -4.12654907e-01 1.91908622e+00
1.05555570e+00 1.35073019e-03 1.45757720e-01 -1.33160308e-01
6.49596035e-01 5.27532995e-01 -1.61424369e-01 -7.90229499e-01
-1.39579058e-01 1.53473586e-01 3.38828206e-01 4.28828150e-01
-1.01245940e+00 1.18760002e+00 6.12191153e+00 7.31511652e-01
-8.50939631e-01 -1.73006713e-01 1.31830454e-01 5.18665425e-02
-3.71779233e-01 1.63431928e-01 -6.25281692e-01 1.02186203e-01
1.21919405e+00 -7.01724350e-01 -1.33819848e-01 5.33173978e-01
1.79493055e-01 -3.28577399e-01 -1.06969428e+00 4.35258061e-01
7.00059772e-01 -1.50052083e+00 5.68626463e-01 -2.03256354e-01
9.20419335e-01 3.45370680e-01 -7.54970133e-01 3.75171036e-01
4.52638209e-01 -5.79591572e-01 5.43349266e-01 2.70906478e-01
5.52037001e-01 -7.21140504e-01 8.14596474e-01 5.29171944e-01
-8.52209330e-01 5.75239003e-01 -5.61300516e-01 1.56278640e-01
2.41543204e-01 3.84809345e-01 -1.16813672e+00 1.22720265e+00
8.84007439e-02 7.79880822e-01 -8.91944826e-01 1.05694628e+00
-6.30159080e-01 9.87050474e-01 -1.99021205e-01 -4.10790265e-01
3.82976890e-01 -1.42935365e-01 1.06690228e+00 1.81696451e+00
1.77206725e-01 3.15854341e-01 4.51885581e-01 1.05507508e-01
-4.13165599e-01 4.88926947e-01 -4.18445885e-01 -6.10527575e-01
1.93643644e-01 1.33893728e+00 -9.39546883e-01 -9.11931276e-01
-1.74827307e-01 1.32540846e+00 2.39644483e-01 3.72009635e-01
-2.02059314e-01 -9.56761539e-01 -2.83061594e-01 2.70777033e-03
2.17988834e-01 -5.62035739e-01 -4.15875800e-02 -1.63866198e+00
1.18165351e-02 -1.07798970e+00 5.93054771e-01 -6.16427183e-01
-9.99825478e-01 7.68154263e-01 2.25695416e-01 -7.79715300e-01
-5.62776744e-01 1.86488047e-01 -1.04241514e+00 6.17610633e-01
-1.22384417e+00 -1.37931323e+00 5.38647249e-02 6.07646182e-02
1.18550503e+00 -1.88724697e-01 8.26255262e-01 -2.59353310e-01
-6.63178205e-01 3.53744596e-01 1.87691689e-01 1.97759882e-01
8.61326814e-01 -1.53629947e+00 9.89733577e-01 1.28978944e+00
2.73681730e-01 6.18602812e-01 1.11489606e+00 -1.07399440e+00
-1.51263976e+00 -1.24713206e+00 1.36349249e+00 -5.51887214e-01
7.04926670e-01 -3.79957467e-01 -9.98970151e-01 8.47481787e-01
9.31054413e-01 -6.52130961e-01 7.43649185e-01 1.54742479e-01
-4.39585336e-02 2.20687389e-01 -5.93711078e-01 7.12554753e-01
9.51568425e-01 -2.77420610e-01 -1.41089463e+00 6.06251538e-01
9.11676168e-01 -3.57295394e-01 -5.98904967e-01 1.95134297e-01
2.27730617e-01 -4.60226893e-01 4.03981388e-01 -6.73428953e-01
6.64030313e-01 -9.04439464e-02 1.35922030e-01 -1.73060167e+00
3.24197114e-02 -9.01896775e-01 -1.66324228e-01 1.64866745e+00
5.59846640e-01 -4.87207979e-01 4.49730814e-01 3.21100563e-01
-6.14112735e-01 -2.22833589e-01 -4.53289837e-01 -7.40716577e-01
5.80716762e-04 1.30343407e-01 1.33312479e-01 6.25071943e-01
4.60557759e-01 1.11618221e+00 -1.19222859e-02 -3.48780043e-02
4.50165778e-01 3.27643573e-01 1.11102223e+00 -9.80705559e-01
-3.61805521e-02 -3.86277109e-01 -1.70340404e-01 -9.97121632e-01
4.54405427e-01 -1.06379068e+00 3.74019980e-01 -2.56300902e+00
5.95533192e-01 1.99971110e-01 3.34113121e-01 4.01075989e-01
-4.73678350e-01 1.34259118e-02 -1.57237276e-02 4.30130988e-01
-1.22569811e+00 6.69400215e-01 8.66426229e-01 -4.56363857e-01
-5.11896968e-01 1.44738341e-02 -8.27590466e-01 6.70342743e-01
7.63507068e-01 -6.40681922e-01 -4.03937936e-01 -5.14186323e-01
7.36625940e-02 1.96410328e-01 -2.01443419e-01 -6.48825586e-01
6.13765061e-01 -9.61167440e-02 -1.16591249e-02 -1.09074628e+00
-2.29162440e-01 2.14048684e-01 -5.04821837e-01 2.85664409e-01
-4.12406474e-01 2.69168973e-01 1.86003223e-01 5.40607452e-01
-3.49610269e-01 -4.21803325e-01 3.10285300e-01 -2.31373206e-01
-2.52768397e-01 -4.03214954e-02 -6.11609697e-01 5.52645862e-01
3.64977151e-01 -3.72142233e-02 -6.08515561e-01 -4.99183536e-01
-8.46741255e-03 4.19325083e-01 6.18209004e-01 1.97511822e-01
5.14144063e-01 -5.68071127e-01 -1.34579468e+00 -3.43382925e-01
1.07045181e-01 1.69462427e-01 -4.64038737e-02 5.88146329e-01
-9.47575271e-01 5.28669655e-01 5.44465668e-02 -2.90976822e-01
-1.38446772e+00 2.49098495e-01 -3.79403681e-01 -6.63438678e-01
-8.79893959e-01 3.66944969e-01 -1.89549178e-01 -3.72281879e-01
-1.43041044e-01 -2.76813000e-01 -2.24794701e-01 2.27610618e-01
5.94560623e-01 5.02568841e-01 3.53632420e-01 -5.86111903e-01
-2.51605541e-01 1.89011032e-03 -4.67136294e-01 -3.38914573e-01
1.54404163e+00 -1.45270750e-01 -5.57266951e-01 4.06631231e-01
9.38309610e-01 6.13817811e-01 -5.19864798e-01 -8.04422498e-02
4.11574602e-01 1.19944580e-01 -1.77242175e-01 -5.38095891e-01
-3.58782470e-01 4.75686729e-01 -6.50698125e-01 5.00170171e-01
8.61375153e-01 8.92160982e-02 1.01788831e+00 1.11422777e+00
-1.41645938e-01 -1.14907980e+00 3.09699699e-02 8.09627950e-01
1.20010448e+00 -9.89546180e-01 6.83512986e-01 -2.54860789e-01
-7.55787313e-01 1.13495588e+00 1.36970744e-01 -2.41851926e-01
-2.06152331e-02 -1.44205004e-01 -8.68721083e-02 -2.96452135e-01
-7.74012864e-01 -9.63800475e-02 6.20087028e-01 3.49200428e-01
5.81543505e-01 -9.77352113e-02 -7.96039999e-01 5.91476262e-01
-6.20332360e-01 -5.30553758e-01 1.02401936e+00 1.20459270e+00
-5.87294400e-01 -1.02431512e+00 9.49023664e-02 6.94828153e-01
-6.92685664e-01 -3.27723831e-01 -9.75390434e-01 7.80422211e-01
-9.84762311e-01 1.24688983e+00 -2.07386352e-02 2.34403297e-01
5.55582106e-01 -3.59016494e-03 6.86121762e-01 -1.29633892e+00
-8.63670468e-01 2.78608531e-01 7.75525212e-01 9.43687558e-02
-4.73524511e-01 -7.41155863e-01 -1.24259186e+00 -2.33785555e-01
-4.28626269e-01 7.12185979e-01 6.02100551e-01 8.88484716e-01
3.65165830e-01 4.11277860e-01 3.08456361e-01 -8.85648131e-01
-5.19216120e-01 -1.40198946e+00 -4.17177439e-01 1.39249131e-01
2.34033436e-01 5.21565914e-01 -1.63329780e-01 2.28672653e-01] | [12.516236305236816, 9.519414901733398] |
10a1fdbf-f26b-4088-aeae-459f9be0e5a1 | vision-transformer-based-model-for-describing | 2210.02762 | null | https://arxiv.org/abs/2210.02762v2 | https://arxiv.org/pdf/2210.02762v2.pdf | Vision Transformer Based Model for Describing a Set of Images as a Story | Visual Story-Telling is the process of forming a multi-sentence story from a set of images. Appropriately including visual variation and contextual information captured inside the input images is one of the most challenging aspects of visual storytelling. Consequently, stories developed from a set of images often lack cohesiveness, relevance, and semantic relationship. In this paper, we propose a novel Vision Transformer Based Model for describing a set of images as a story. The proposed method extracts the distinct features of the input images using a Vision Transformer (ViT). Firstly, input images are divided into 16X16 patches and bundled into a linear projection of flattened patches. The transformation from a single image to multiple image patches captures the visual variety of the input visual patterns. These features are used as input to a Bidirectional-LSTM which is part of the sequence encoder. This captures the past and future image context of all image patches. Then, an attention mechanism is implemented and used to increase the discriminatory capacity of the data fed into the language model, i.e. a Mogrifier-LSTM. The performance of our proposed model is evaluated using the Visual Story-Telling dataset (VIST), and the results show that our model outperforms the current state of the art models. | ['Ajmal Mian', 'Ghulam Mubashar Hassan', 'Zainy M. Malakan'] | 2022-10-06 | null | null | null | null | ['visual-storytelling'] | ['natural-language-processing'] | [ 5.42759001e-01 -3.37257594e-01 -2.78928448e-02 -3.29449832e-01
-4.11676675e-01 -4.08318222e-01 9.06335175e-01 -2.35641405e-01
-9.13242577e-04 5.40424347e-01 5.15874267e-01 1.23994380e-01
2.20535100e-01 -7.16909170e-01 -9.85062957e-01 -6.81020260e-01
4.90431905e-01 1.23756807e-02 3.18120480e-01 -1.64328173e-01
3.61623436e-01 8.38166699e-02 -1.65350485e+00 1.06581426e+00
4.08628672e-01 1.03935492e+00 9.21852171e-01 6.77696764e-01
-4.36636478e-01 1.22322702e+00 -5.34153104e-01 -8.46982598e-02
-1.46572247e-01 -7.91278839e-01 -6.04089141e-01 7.17036426e-01
5.16781151e-01 -4.21054840e-01 -6.00283623e-01 8.69058013e-01
1.26375288e-01 2.70162165e-01 5.24457633e-01 -1.23499990e+00
-1.15938401e+00 6.73311889e-01 -4.71253008e-01 1.00216374e-01
5.68199098e-01 2.44610250e-01 8.73441458e-01 -1.12253296e+00
9.62187648e-01 1.32344234e+00 -4.16119061e-02 3.97676438e-01
-1.12549138e+00 -3.11209857e-01 1.40792415e-01 6.99864388e-01
-1.24558139e+00 -2.88029015e-01 1.06071210e+00 -6.54682636e-01
8.46592605e-01 3.03101242e-01 1.24348974e+00 1.28411639e+00
3.53479087e-01 1.03564239e+00 1.02545440e+00 -2.27872625e-01
-5.06821722e-02 3.18893015e-01 -1.61701694e-01 4.18642312e-01
-3.29029620e-01 -9.44987014e-02 -9.12104666e-01 3.45416963e-01
9.64931726e-01 4.61734056e-01 -3.42541665e-01 -3.65094960e-01
-1.39608586e+00 6.11022592e-01 5.73479831e-01 5.41130185e-01
-5.83952785e-01 9.76343080e-02 9.31123495e-02 1.18023977e-01
2.46062130e-01 -3.54606025e-02 4.35039401e-01 1.27189368e-01
-9.57297325e-01 1.26830608e-01 1.17270172e-01 9.22824085e-01
6.94813848e-01 7.41290674e-02 -4.80398685e-01 7.60292470e-01
2.55162090e-01 4.21815842e-01 4.04798627e-01 -6.32234454e-01
6.56237185e-01 7.06102133e-01 -1.61055043e-01 -1.19703746e+00
5.12815565e-02 -1.93923473e-01 -8.25395882e-01 -1.16145723e-01
-3.83018441e-02 1.56552777e-01 -8.97789121e-01 1.48102844e+00
7.04957843e-02 1.45228207e-01 1.82244688e-01 1.07905877e+00
1.10074949e+00 1.25683045e+00 -2.39217788e-01 -3.10279578e-01
1.35738587e+00 -1.14535189e+00 -8.09212625e-01 -6.05507672e-01
-1.24683365e-01 -7.77519703e-01 1.14011431e+00 1.00024760e-01
-1.22952211e+00 -8.64768326e-01 -1.14944661e+00 -3.66543978e-01
-2.65468597e-01 1.22534707e-01 1.00063749e-01 -1.73286676e-01
-1.05293465e+00 4.54023443e-02 -1.52087763e-01 -4.76948887e-01
3.41359735e-01 -2.82648981e-01 -3.91145498e-01 -3.84084284e-01
-9.66744781e-01 7.35691071e-01 5.25460422e-01 3.92324589e-02
-1.19485474e+00 -4.01714772e-01 -9.77704942e-01 8.30284208e-02
1.57889023e-01 -7.68603563e-01 8.30203593e-01 -1.31572306e+00
-1.05025256e+00 6.95797682e-01 -3.81057113e-01 -3.51494938e-01
3.53017330e-01 1.75783992e-01 -2.99832851e-01 3.03293377e-01
1.02477707e-01 8.60683501e-01 1.15893042e+00 -1.58536935e+00
-6.16150081e-01 -1.54721946e-01 5.71545176e-02 5.22228837e-01
-2.67144859e-01 1.02165733e-02 -6.71043158e-01 -6.78732097e-01
9.64614097e-03 -5.69596827e-01 1.63074300e-01 -2.47362003e-01
-4.36902583e-01 1.82999924e-01 1.15281785e+00 -8.54904890e-01
1.05213320e+00 -2.30203819e+00 2.96240836e-01 -1.39298409e-01
1.40600577e-01 -4.69472557e-02 -2.10394725e-01 7.02348351e-01
1.89133193e-02 -2.29731590e-01 -3.06855977e-01 -3.25956941e-01
-3.90995800e-01 2.19878852e-01 -5.32450020e-01 1.10754691e-01
2.80328482e-01 1.16422951e+00 -8.42534363e-01 -7.18267858e-01
4.35644329e-01 5.84704995e-01 4.22179094e-03 4.78428423e-01
-3.67013991e-01 5.93854249e-01 -2.02432081e-01 2.84330815e-01
3.89917612e-01 -2.95319140e-01 -5.47569357e-02 -4.66742963e-02
-1.24435142e-01 1.11459993e-01 -7.67260253e-01 1.86525297e+00
-2.83632785e-01 1.04505312e+00 -5.98524690e-01 -9.26439762e-01
1.22844231e+00 2.62149364e-01 4.98923361e-02 -9.33474541e-01
8.06793422e-02 -3.28612477e-01 -1.30798042e-01 -9.27180529e-01
6.36507034e-01 -1.43222064e-01 -7.50129446e-02 6.38580799e-01
6.93265125e-02 -2.30742410e-01 1.33137345e-01 2.94460714e-01
6.38938308e-01 1.45040646e-01 2.14196458e-01 3.18857610e-01
5.59621930e-01 1.73175916e-01 3.90868902e-01 4.26173806e-01
1.60039619e-01 9.26187575e-01 3.84834498e-01 -5.94812572e-01
-1.28804600e+00 -1.06975758e+00 2.08501533e-01 7.85972893e-01
3.71775329e-01 -2.19641522e-01 -5.24797022e-01 -3.44725192e-01
-2.00811833e-01 9.75964069e-01 -8.85999858e-01 -2.14080185e-01
-3.81542414e-01 -2.63467908e-01 -2.03543082e-02 4.87001687e-01
7.67190754e-01 -1.45288134e+00 -8.36697757e-01 1.15871146e-01
-4.74711120e-01 -1.25611639e+00 -7.36530125e-01 -2.70875096e-01
-4.89067584e-01 -7.34673381e-01 -7.61005044e-01 -1.20133471e+00
7.03461409e-01 8.16491187e-01 7.97402799e-01 -2.40640044e-01
-6.12076148e-02 2.36835077e-01 -4.63846326e-01 -1.14672035e-01
-4.00576442e-01 -5.81877410e-01 -4.24236953e-01 7.06431270e-01
-4.58270162e-02 -6.14179909e-01 -6.00111365e-01 -6.76471367e-03
-1.07362735e+00 9.73389745e-01 7.34462202e-01 1.02259004e+00
8.06169689e-01 -1.34073421e-01 1.85683265e-01 -5.46212852e-01
5.60312688e-01 -4.95694429e-01 2.23729033e-02 3.90170306e-01
-1.88724231e-03 -1.21854953e-01 6.58048570e-01 -7.25389421e-01
-1.32948756e+00 1.06829859e-01 4.68353152e-01 -7.49877453e-01
-9.85378213e-03 4.69783247e-01 -2.57495284e-01 3.02217215e-01
2.60529667e-01 9.19332147e-01 -4.85304780e-02 -7.76663348e-02
5.40893912e-01 5.72031498e-01 7.07608581e-01 -2.81321574e-02
5.38964212e-01 5.40449083e-01 -4.26477879e-01 -9.20750499e-01
-5.03984988e-01 -1.71562731e-01 -5.39505064e-01 -7.73733437e-01
1.04501402e+00 -7.59376884e-01 -1.92439601e-01 4.24707174e-01
-1.45031857e+00 -1.15803674e-01 -3.11606020e-01 1.77390635e-01
-7.57239878e-01 1.65271416e-01 -4.57499146e-01 -6.72107160e-01
-2.81437695e-01 -1.08408105e+00 8.31738472e-01 4.62070197e-01
-1.47051245e-01 -7.33175576e-01 -1.67992249e-01 6.01120412e-01
1.33458704e-01 4.16299880e-01 9.75762665e-01 -2.59204805e-01
-7.21309483e-01 6.47373125e-02 -4.00170475e-01 1.42640293e-01
6.73515946e-02 6.74466342e-02 -8.84255528e-01 -5.06149754e-02
1.31515741e-01 -3.24685246e-01 1.03953695e+00 3.97133321e-01
1.00482106e+00 -4.62087631e-01 -2.44557202e-01 4.19323355e-01
1.37001276e+00 6.70747399e-01 7.20637381e-01 2.96733588e-01
9.51961696e-01 7.50847518e-01 6.25065148e-01 3.53107512e-01
7.10207999e-01 5.22585571e-01 5.38661778e-01 -2.37014174e-01
-3.31430942e-01 -7.46671557e-01 5.14842629e-01 9.37501550e-01
1.15807712e-01 -2.51232147e-01 -8.37813616e-01 7.80217588e-01
-2.00198793e+00 -1.37863660e+00 -1.64046455e-02 1.89214897e+00
5.81516564e-01 -9.36246570e-03 -1.29421845e-01 -2.44861208e-02
8.65338087e-01 5.50063670e-01 -7.36903548e-01 -5.26348591e-01
-5.25390148e-01 -3.69448334e-01 -7.89391175e-02 1.85528889e-01
-6.49077415e-01 9.59202111e-01 5.53878355e+00 7.06103981e-01
-1.28830874e+00 2.39005964e-02 7.15940654e-01 -2.52493858e-01
-5.68092704e-01 -2.69026477e-02 -1.90171942e-01 4.65476125e-01
2.71137863e-01 -3.12700361e-01 4.70991284e-01 5.27135670e-01
3.43949616e-01 -2.40751281e-01 -9.76006210e-01 1.19419193e+00
6.64240658e-01 -1.59886730e+00 5.27600944e-01 -1.45889014e-01
8.71031821e-01 -3.83087486e-01 4.61606354e-01 -9.65924785e-02
-1.24403629e-02 -1.09695041e+00 1.24056923e+00 8.75823140e-01
8.33156288e-01 -7.63587832e-01 4.71569687e-01 4.44963962e-01
-1.17984021e+00 -1.42453969e-01 -3.32502902e-01 -1.31154522e-01
3.50347757e-01 1.92746878e-01 -8.80384624e-01 4.32233840e-01
7.47415900e-01 1.16684115e+00 -4.96030003e-01 6.96461678e-01
-2.65816420e-01 2.95924634e-01 3.40243667e-01 -2.05992565e-01
3.02136689e-01 -1.83369875e-01 5.79121768e-01 1.03185380e+00
4.63144094e-01 2.06160367e-01 -1.36919171e-01 1.22920299e+00
6.07472509e-02 -5.41943200e-02 -8.42184544e-01 -2.59884924e-01
3.39845628e-01 1.11814713e+00 -5.53821921e-01 -5.84835708e-01
-4.57294822e-01 1.07163107e+00 2.76002139e-01 4.54235703e-01
-6.99917734e-01 -1.26543231e-02 3.15100878e-01 4.49831523e-02
4.92212594e-01 -1.41367108e-01 -3.53094280e-01 -9.94598508e-01
2.15465814e-01 -7.61515141e-01 2.67414093e-01 -1.56426942e+00
-9.26935136e-01 7.56834030e-01 7.06865191e-02 -1.44192600e+00
-2.71569848e-01 -3.54821123e-02 -8.70103955e-01 8.15405011e-01
-1.13924384e+00 -1.49007761e+00 -5.80635369e-01 8.23074996e-01
9.99282062e-01 -1.32893100e-01 4.37385172e-01 -2.04839528e-01
-3.24579686e-01 1.96252629e-01 -1.26587711e-02 -2.79463138e-02
4.85092193e-01 -8.52024198e-01 4.54616368e-01 1.12126732e+00
4.86536026e-01 2.15074748e-01 6.44243717e-01 -6.43407643e-01
-1.47397280e+00 -1.17426074e+00 9.97313023e-01 -8.92236456e-02
4.90452707e-01 -4.59655613e-01 -9.73178267e-01 6.80609405e-01
6.86674595e-01 -3.73646110e-01 5.05860209e-01 -7.07818210e-01
-3.28763396e-01 -8.46916884e-02 -8.13515425e-01 7.11208522e-01
9.09906328e-01 -5.92804432e-01 -8.74626040e-01 -3.24291475e-02
8.94026160e-01 -3.23959321e-01 -4.06587392e-01 -1.73510402e-01
6.55175507e-01 -1.04377401e+00 8.98096979e-01 -2.74852961e-01
1.02501452e+00 -3.07776242e-01 -2.50435114e-01 -1.28351438e+00
-3.61442387e-01 -3.44178975e-01 2.67825127e-01 1.35775149e+00
2.06888422e-01 -3.02701861e-01 2.79485464e-01 1.76964253e-01
-3.01426742e-02 -6.04086518e-01 -7.42181003e-01 -2.28795514e-01
-3.41028601e-01 -3.15749675e-01 6.97552025e-01 1.01568604e+00
2.01511100e-01 5.81981957e-01 -6.42467260e-01 -1.47134781e-01
4.12020296e-01 5.20540237e-01 6.15273297e-01 -4.13638979e-01
-2.82935888e-01 -2.84392357e-01 -4.67061102e-01 -9.57656085e-01
-1.68658569e-01 -8.23729992e-01 1.69923045e-02 -1.92156398e+00
6.72257125e-01 1.58570006e-01 -1.18475288e-01 4.16935563e-01
-1.50048897e-01 2.04446882e-01 6.73263907e-01 2.82308578e-01
-4.66528356e-01 7.49713778e-01 1.69044757e+00 -5.54837465e-01
-2.02624187e-01 -4.65909302e-01 -4.74433810e-01 4.20829266e-01
7.50114083e-01 -1.12472333e-01 -5.84023237e-01 -6.83523893e-01
-1.21558644e-02 6.11860454e-01 6.55917108e-01 -9.64590609e-01
1.87934443e-01 -4.39977586e-01 7.04143286e-01 -1.01284242e+00
7.32590377e-01 -6.25176191e-01 5.31951308e-01 2.86475182e-01
-4.73747432e-01 2.59459257e-01 -1.01990011e-02 4.51341689e-01
-4.27251697e-01 1.59414366e-01 4.68963206e-01 -1.82359964e-01
-9.44859862e-01 2.64647305e-01 -2.36790240e-01 -1.32879436e-01
1.20943308e+00 -5.41391850e-01 -4.97926205e-01 -5.73782086e-01
-4.94703263e-01 2.41839170e-01 5.15020311e-01 8.78394067e-01
1.44370699e+00 -1.72681499e+00 -8.37014794e-01 2.74168879e-01
4.25132066e-01 -6.96452931e-02 7.69165099e-01 5.12497246e-01
-3.08818549e-01 3.45147312e-01 -7.24898040e-01 -7.24441051e-01
-1.48329782e+00 7.10292161e-01 1.27246916e-01 1.41788140e-01
-8.37820411e-01 7.60136247e-01 6.33823395e-01 3.87049735e-01
-7.95863494e-02 -1.68204144e-01 -6.59260452e-01 2.08241656e-01
7.57495344e-01 -8.19770545e-02 -5.54927468e-01 -1.21029150e+00
-1.29326448e-01 5.83777249e-01 -5.00548035e-02 -4.82919574e-01
1.33374774e+00 -3.12809736e-01 -1.23786762e-01 9.36021507e-01
1.13220370e+00 -5.67967117e-01 -1.41609693e+00 -5.08313835e-01
-4.40764695e-01 -5.66589952e-01 -9.82350931e-02 -5.44607282e-01
-1.08129203e+00 1.23979986e+00 2.50617415e-01 3.98154818e-02
1.28821576e+00 2.16437340e-01 8.66205752e-01 -2.15130150e-01
1.16126634e-01 -7.75064945e-01 5.33626318e-01 2.88811892e-01
1.41735649e+00 -1.00991082e+00 -3.50987256e-01 3.46097251e-04
-1.22892833e+00 1.06072950e+00 5.26385903e-01 -1.12306595e-01
4.28959541e-02 -8.55452791e-02 3.69584411e-02 -1.25438705e-01
-9.57561910e-01 -3.10536325e-01 4.41303909e-01 6.38540387e-01
-4.26865593e-02 9.39189419e-02 -1.37513712e-01 6.27518833e-01
-3.19499910e-01 -1.69411838e-01 5.94622374e-01 5.72610080e-01
-4.50810075e-01 -5.79646766e-01 -5.86853743e-01 1.75498143e-01
1.32302746e-01 -4.54801321e-02 -6.45276189e-01 4.70948666e-01
2.86583155e-01 1.07742977e+00 2.80090988e-01 -7.85597742e-01
1.95078880e-01 1.41984951e-02 4.50065106e-01 -5.32754898e-01
-4.00758237e-01 2.41032131e-02 -1.05602309e-01 -3.00072521e-01
-5.25368810e-01 -7.22595692e-01 -1.14139557e+00 -9.85852405e-02
2.78843611e-01 -2.38862485e-01 5.21774948e-01 1.04150462e+00
1.54413208e-01 8.07684302e-01 6.95550501e-01 -8.71891320e-01
1.85456648e-01 -7.85770953e-01 -3.87943596e-01 5.27197361e-01
4.84475911e-01 -5.16633034e-01 -2.74879951e-02 5.71077108e-01] | [11.088898658752441, 0.6752956509590149] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.