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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
b07e52a5-7ced-4aef-8c29-4e0276fd67f3
|
improving-the-transferability-of-time-series
|
2307.00066
| null |
https://arxiv.org/abs/2307.00066v1
|
https://arxiv.org/pdf/2307.00066v1.pdf
|
Improving the Transferability of Time Series Forecasting with Decomposition Adaptation
|
Due to effective pattern mining and feature representation, neural forecasting models based on deep learning have achieved great progress. The premise of effective learning is to collect sufficient data. However, in time series forecasting, it is difficult to obtain enough data, which limits the performance of neural forecasting models. To alleviate the data scarcity limitation, we design Sequence Decomposition Adaptation Network (SeDAN) which is a novel transfer architecture to improve forecasting performance on the target domain by aligning transferable knowledge from cross-domain datasets. Rethinking the transferability of features in time series data, we propose Implicit Contrastive Decomposition to decompose the original features into components including seasonal and trend features, which are easier to transfer. Then we design the corresponding adaptation methods for decomposed features in different domains. Specifically, for seasonal features, we perform joint distribution adaptation and for trend features, we design an Optimal Local Adaptation. We conduct extensive experiments on five benchmark datasets for multivariate time series forecasting. The results demonstrate the effectiveness of our SeDAN. It can provide more efficient and stable knowledge transfer.
|
['Qiang Wang', 'Yan Wang', 'Yan Gao']
|
2023-06-30
| null | null | null | null |
['transfer-learning', 'time-series-forecasting', 'multivariate-time-series-forecasting']
|
['miscellaneous', 'time-series', 'time-series']
|
[-1.08843446e-01 -7.34710097e-01 -1.48680508e-01 -5.66962123e-01
-2.60499984e-01 -5.42816699e-01 3.07655007e-01 -5.27815163e-01
-7.55483061e-02 6.28055871e-01 3.39458972e-01 -3.86914611e-01
-2.11936221e-01 -7.92001784e-01 -5.00625610e-01 -9.89170849e-01
-1.42829910e-01 1.10677443e-01 -4.12743799e-02 -2.94339389e-01
-5.55036217e-03 2.67155647e-01 -1.24614227e+00 2.32873976e-01
1.11282742e+00 1.27183533e+00 2.83377379e-01 6.58379272e-02
-4.14022982e-01 7.12315857e-01 -6.19687021e-01 1.36718992e-02
2.36792251e-01 -5.21793067e-01 -4.42945123e-01 1.06639944e-01
-1.65498912e-01 -5.20702481e-01 -3.48882735e-01 8.26943815e-01
4.25144523e-01 2.55392998e-01 7.59394169e-01 -1.50960302e+00
-1.09625399e+00 5.07072389e-01 -6.46067441e-01 4.82135415e-01
-1.91684976e-01 8.98809358e-03 6.20617032e-01 -6.65876567e-01
1.26215279e-01 1.10904992e+00 8.37355852e-01 4.64205801e-01
-9.28452909e-01 -1.04128611e+00 4.95275468e-01 4.04174477e-01
-1.03447580e+00 -2.34932125e-01 1.09114754e+00 -4.76736814e-01
6.77579701e-01 -2.29177251e-03 3.64569932e-01 1.07336855e+00
2.25450486e-01 8.33425105e-01 9.14676368e-01 -9.64695364e-02
7.88557529e-02 -2.67651789e-02 3.10504772e-02 9.91609320e-02
-1.73652753e-01 7.96957687e-02 -3.91821653e-01 5.25873154e-03
7.34916389e-01 6.38816953e-01 -3.21604908e-01 -2.01890357e-02
-1.30787063e+00 7.12309539e-01 4.09735143e-01 3.28918755e-01
-5.48474610e-01 -2.36027360e-01 8.18333626e-01 8.85398209e-01
9.11432624e-01 -1.64542645e-01 -9.96563733e-01 -1.67008460e-01
-5.62345862e-01 2.67234165e-02 6.59116209e-01 7.52016604e-01
6.54975832e-01 4.68206644e-01 -1.01408571e-01 8.39528084e-01
-5.13932742e-02 7.26932764e-01 1.03855240e+00 -4.18284088e-01
7.22633243e-01 5.76644778e-01 3.69163826e-02 -1.37204909e+00
-4.83646035e-01 -5.71247041e-01 -1.40950274e+00 -4.35226649e-01
8.55199713e-03 -6.56228185e-01 -8.10014725e-01 1.85128856e+00
2.43653327e-01 5.04313350e-01 2.92247385e-01 1.01071787e+00
6.47364080e-01 1.39053464e+00 9.88378450e-02 -6.06644273e-01
1.01587486e+00 -9.81702387e-01 -8.09975207e-01 9.17922780e-02
5.84659815e-01 -6.35719597e-01 9.81022596e-01 2.84796864e-01
-3.73463005e-01 -7.96661794e-01 -8.50656271e-01 2.34722942e-01
-3.15353334e-01 2.52648234e-01 6.47574723e-01 1.83553338e-01
-5.44293404e-01 5.20369053e-01 -8.14347744e-01 -1.20137505e-01
2.64360964e-01 2.69608974e-01 -2.58354962e-01 1.23392399e-02
-1.59076011e+00 5.41999936e-01 6.15028381e-01 2.94846684e-01
-6.67943597e-01 -9.28580701e-01 -5.58515012e-01 2.42149130e-01
-3.63221951e-02 -4.62987214e-01 1.22523010e+00 -1.42442298e+00
-1.58187723e+00 1.82137147e-01 -9.73857269e-02 -2.41264269e-01
1.62424311e-01 -7.65908808e-02 -1.07117188e+00 -2.03695133e-01
-2.04433147e-02 2.06996486e-01 1.00484788e+00 -6.21828616e-01
-8.34309101e-01 -2.46120781e-01 -4.83817697e-01 1.45065725e-01
-1.15944970e+00 2.66118422e-02 1.59700632e-01 -9.93883967e-01
-6.13651648e-02 -5.75459242e-01 -1.71612725e-01 -4.77584541e-01
1.66962236e-01 -4.69631076e-01 1.06423366e+00 -1.13909090e+00
1.43321848e+00 -2.41078138e+00 1.40087515e-01 1.76466793e-01
1.85916852e-02 1.03567578e-01 -4.43309933e-01 5.35427451e-01
-1.92622155e-01 -3.53771478e-01 -3.38348538e-01 6.64803535e-02
-1.79582648e-02 4.16753918e-01 -1.08966327e+00 1.17006250e-01
1.96053371e-01 7.83939660e-01 -7.00662494e-01 -2.19910070e-02
-1.36511400e-01 2.39549771e-01 -1.77191019e-01 3.78161907e-01
-2.28547156e-01 6.15393877e-01 -9.11889911e-01 3.70994955e-01
1.01688147e+00 -4.42420453e-01 -8.66274908e-02 -3.84657979e-02
-1.97221786e-01 2.24317208e-01 -8.81871521e-01 1.54283631e+00
-5.66860855e-01 3.38336825e-01 -2.62289733e-01 -1.55733132e+00
1.48086596e+00 3.47583830e-01 7.09784925e-01 -7.52157867e-01
-6.85427785e-02 2.97870010e-01 7.02333823e-02 -6.30840361e-01
1.99755475e-01 -3.81628484e-01 -3.16672176e-02 4.80836421e-01
-1.61936623e-03 2.63942361e-01 -1.84827283e-01 -3.66046697e-01
7.90156901e-01 1.55865535e-01 -2.98957918e-02 -3.27460140e-01
6.90494657e-01 3.58025059e-02 9.78914082e-01 -1.27038226e-01
3.38520817e-02 3.12405497e-01 3.52701813e-01 -1.05426383e+00
-8.35303307e-01 -6.77874327e-01 -1.23204775e-01 1.38598347e+00
-9.66199413e-02 -1.27544791e-01 -3.88409555e-01 -5.89001596e-01
8.19364265e-02 3.39329898e-01 -6.41214490e-01 -4.26208794e-01
-5.49581766e-01 -1.06982291e+00 2.30930805e-01 8.44850063e-01
6.96777761e-01 -1.21974158e+00 -1.78731978e-01 5.34358561e-01
-3.54914069e-01 -6.05458200e-01 -7.02347696e-01 -1.27243716e-02
-1.09066200e+00 -4.97994959e-01 -1.12046230e+00 -9.72665429e-01
5.43856382e-01 4.50300097e-01 8.99001658e-01 -1.53122202e-01
2.79124081e-01 1.03394575e-01 -5.03347754e-01 -4.65079814e-01
1.03083169e-02 3.76172155e-01 4.09560800e-01 3.53444695e-01
6.25404596e-01 -1.04881501e+00 -6.18537962e-01 5.05737066e-01
-9.58232164e-01 -1.59973353e-02 4.71621692e-01 9.95577693e-01
3.87706459e-01 2.71412790e-01 1.21436155e+00 -3.69707882e-01
9.97854352e-01 -9.45807159e-01 -5.81887305e-01 4.19743061e-01
-6.34113252e-01 -1.02607675e-01 1.23565555e+00 -7.86189735e-01
-1.19616604e+00 -2.63026357e-01 2.21353009e-01 -7.10721791e-01
1.50265712e-02 1.05309987e+00 -7.06258193e-02 2.09278464e-01
4.61681843e-01 8.77288699e-01 1.58105910e-01 -6.51087582e-01
6.74965084e-02 8.49787116e-01 3.50362718e-01 -7.64777184e-01
7.10937083e-01 1.18278056e-01 -2.12763280e-01 -4.58330721e-01
-8.24498713e-01 -2.06145644e-01 -4.51992810e-01 1.12924576e-01
5.70277750e-01 -9.73886490e-01 -6.17019951e-01 7.48082340e-01
-1.08342266e+00 -2.95357019e-01 2.33950019e-02 6.87929809e-01
-2.79549867e-01 2.95780480e-01 -5.71366131e-01 -5.52635789e-01
-4.49370235e-01 -5.72553515e-01 6.29482746e-01 2.74544209e-01
1.88704714e-01 -1.23171663e+00 3.54020953e-01 -2.53689349e-01
7.23185360e-01 5.39056472e-02 9.22344804e-01 -7.29714990e-01
-1.50628492e-01 3.02692000e-02 -2.63101876e-01 5.74674964e-01
6.23763621e-01 -6.69874474e-02 -7.94635594e-01 -3.10008645e-01
2.33156875e-01 -1.82696730e-01 8.94879460e-01 3.45354795e-01
1.52010322e+00 -5.24424076e-01 -3.07984412e-01 7.80413091e-01
9.69343185e-01 5.29514194e-01 3.53847384e-01 4.21550781e-01
6.32536173e-01 7.93649435e-01 5.17226219e-01 5.79372525e-01
6.00746214e-01 2.56452441e-01 -1.13801360e-01 5.08128591e-02
4.04288590e-01 -3.66837353e-01 4.62075084e-01 1.57748771e+00
-3.08090359e-01 4.15581986e-02 -9.62989390e-01 5.84324300e-01
-2.11555958e+00 -9.71174240e-01 2.46549681e-01 1.86701584e+00
9.18875575e-01 -2.13589266e-01 1.39012635e-01 -5.95257431e-02
6.67383254e-01 9.10845250e-02 -9.57502842e-01 -3.82316299e-02
-2.05519050e-01 -2.16566801e-01 1.67109445e-02 -9.11126807e-02
-1.21733057e+00 6.54954553e-01 5.98480749e+00 1.13019216e+00
-1.57098067e+00 6.19948432e-02 8.32446158e-01 1.30180895e-01
-5.88151693e-01 -2.85891145e-01 -6.51043415e-01 9.90858257e-01
9.54741359e-01 -5.75992644e-01 4.52802360e-01 9.57581162e-01
1.03801727e-01 7.56935894e-01 -9.64108586e-01 1.07944167e+00
-1.91455126e-01 -1.25630176e+00 1.62681177e-01 -1.50232732e-01
9.29784238e-01 -4.45167981e-02 2.20761165e-01 7.54995823e-01
1.72962204e-01 -9.39264655e-01 1.73913315e-01 6.30683124e-01
5.44649005e-01 -9.05315518e-01 7.83804834e-01 5.65757513e-01
-1.46622169e+00 -2.85583973e-01 -6.99946165e-01 -3.48312557e-01
2.71482971e-02 6.90200925e-01 -2.00763702e-01 9.41102386e-01
7.36912608e-01 1.33956170e+00 -4.34842482e-02 6.99658871e-01
2.15687066e-01 8.91302288e-01 -2.74849951e-01 7.91722909e-02
2.24509433e-01 -6.61680818e-01 1.13927439e-01 8.93122256e-01
7.51482189e-01 3.13722163e-01 5.67067444e-01 4.14431393e-01
-3.41478139e-02 3.21705341e-01 -4.15140867e-01 -8.20894092e-02
5.27766287e-01 1.08808267e+00 -3.50728244e-01 -1.88973933e-01
-6.34707689e-01 8.57146919e-01 3.74188513e-01 5.93719423e-01
-6.94457293e-01 -5.29086292e-01 6.29041374e-01 -3.64625216e-01
4.10514534e-01 -2.07622081e-01 -1.59951240e-01 -1.42294729e+00
3.86449277e-01 -9.13484097e-01 7.43945360e-01 -5.56186259e-01
-1.90006280e+00 8.37211072e-01 -1.00014985e-01 -1.74849296e+00
-3.05900514e-01 -5.72204947e-01 -9.03650045e-01 1.09273517e+00
-1.73262846e+00 -1.21582901e+00 -2.12281182e-01 8.92948270e-01
4.99443889e-01 -6.93054318e-01 6.81536734e-01 4.95457709e-01
-5.83774328e-01 5.60910761e-01 6.61562085e-01 1.90323055e-01
7.92852998e-01 -7.81963110e-01 3.78304154e-01 6.73246980e-01
-2.93006778e-01 6.09396160e-01 1.80805027e-01 -4.68019336e-01
-1.30464768e+00 -1.33711720e+00 5.47484636e-01 -6.06325045e-02
8.33076835e-01 -4.06321846e-02 -1.40812349e+00 7.77962983e-01
1.56600401e-01 1.01990925e-04 8.58194828e-01 2.83438951e-01
-4.67434913e-01 -7.69411206e-01 -6.68515861e-01 3.14465255e-01
8.35680723e-01 -3.54299843e-01 -8.30874920e-01 3.73919815e-01
1.02514791e+00 -9.82432961e-02 -9.21179473e-01 5.02641916e-01
4.40496564e-01 -4.82135147e-01 6.29463553e-01 -8.93942237e-01
6.74510002e-01 -3.15704286e-01 -4.68429625e-02 -1.71532214e+00
-6.27693951e-01 -5.87094963e-01 -2.32419357e-01 1.38493013e+00
1.91253752e-01 -1.15657890e+00 3.44388723e-01 4.70476896e-01
-2.21452370e-01 -5.48221290e-01 -8.02490592e-01 -1.14212477e+00
2.74766177e-01 -6.72938302e-02 1.22802651e+00 1.53866124e+00
-4.03293082e-03 3.82176399e-01 -6.42952383e-01 8.09511617e-02
1.42317683e-01 7.90141761e-01 6.64359510e-01 -1.44064903e+00
-2.14550108e-01 -4.15497363e-01 -2.69272439e-02 -1.30567920e+00
4.38299209e-01 -7.42731452e-01 -2.54012108e-01 -1.12270892e+00
1.26461506e-01 -6.05915844e-01 -8.03992689e-01 7.17170298e-01
-2.26519108e-01 -3.10017467e-01 -1.87031880e-01 5.91926217e-01
-2.39006490e-01 1.13008499e+00 1.46994877e+00 -8.81821290e-02
-1.84727043e-01 5.98834008e-02 -6.10711217e-01 4.05804068e-01
9.96081531e-01 -1.86132446e-01 -6.15691245e-01 -7.30683386e-01
9.52328444e-02 1.29850090e-01 -5.36019988e-02 -7.08660603e-01
1.81505978e-01 -5.30039251e-01 6.46422446e-01 -6.54523194e-01
-1.39899090e-01 -1.00664902e+00 1.72911137e-01 4.26102310e-01
-1.30115449e-01 1.81976631e-01 3.61747772e-01 5.44652760e-01
-6.28032148e-01 2.44622856e-01 8.29549432e-02 2.09989414e-01
-1.04532921e+00 7.82322705e-01 -1.59994364e-01 -1.79666281e-01
8.54292154e-01 3.34020257e-02 -2.32828632e-01 -4.07929152e-01
-4.13583726e-01 5.98572552e-01 -1.00739107e-01 7.43313730e-01
4.27097201e-01 -1.77499104e+00 -9.68429148e-01 4.46111053e-01
1.07722329e-02 -5.50268181e-02 6.47124350e-01 7.94829071e-01
-1.06657065e-01 3.81639838e-01 -5.98720908e-01 -4.23114240e-01
-6.76789343e-01 9.60494578e-01 2.43149742e-01 -1.28500968e-01
-5.52897215e-01 4.66952384e-01 5.25751948e-01 -8.05989563e-01
-3.29838507e-02 -4.60664809e-01 -4.14136916e-01 1.46319672e-01
7.44412661e-01 3.17861199e-01 -6.12365529e-02 -9.95583460e-02
-2.38752306e-01 7.38340020e-01 -2.19601765e-01 2.57690161e-01
1.60450435e+00 -1.74511895e-01 -3.71191114e-01 4.77809370e-01
1.19052052e+00 -3.17212790e-01 -1.40598881e+00 -6.32294357e-01
-1.56722263e-01 -2.55279183e-01 -9.93760228e-02 -5.07950366e-01
-1.30137300e+00 9.61894393e-01 4.38743442e-01 5.99037170e-01
1.67967951e+00 -4.82226491e-01 1.14061546e+00 5.77921212e-01
1.05758332e-01 -8.53111804e-01 -2.24726066e-01 7.45707572e-01
9.39478397e-01 -1.32341146e+00 -3.28345299e-01 -3.48537862e-02
-6.70683384e-01 1.38222051e+00 7.14923561e-01 -1.79787844e-01
1.02829850e+00 1.69881601e-02 2.32294142e-01 1.94050342e-01
-9.15689528e-01 2.35118926e-01 3.74351025e-01 5.32517195e-01
3.71902585e-01 1.32825404e-01 -2.80943215e-01 1.24382663e+00
-1.13413282e-01 2.17868373e-01 -8.44554007e-02 4.86042053e-01
-2.47202083e-01 -1.03380692e+00 -3.03851396e-01 4.62532163e-01
-2.54324138e-01 -2.58841068e-02 1.23290837e-01 4.51724708e-01
-1.91561669e-01 7.72192180e-01 3.19902062e-01 -6.00538909e-01
2.20031217e-01 3.62714417e-02 -3.59880328e-02 -1.68545187e-01
-3.16474259e-01 -4.06660400e-02 -3.64712149e-01 -1.30412923e-02
-5.23851573e-01 -5.28949201e-01 -9.76390779e-01 -3.94776851e-01
6.40932694e-02 4.57059145e-01 2.10749775e-01 1.02814472e+00
7.62582541e-01 6.27843201e-01 1.27646470e+00 -5.94319224e-01
-9.35656607e-01 -1.14304543e+00 -6.77206218e-01 2.64338493e-01
4.85296458e-01 -5.91273129e-01 -1.74483091e-01 4.40166950e-01]
|
[6.924373626708984, 2.93058443069458]
|
888d49e2-bd45-4ff4-8c77-85b131a6baae
|
lwsis-lidar-guided-weakly-supervised-instance
|
2212.03504
| null |
https://arxiv.org/abs/2212.03504v2
|
https://arxiv.org/pdf/2212.03504v2.pdf
|
LWSIS: LiDAR-guided Weakly Supervised Instance Segmentation for Autonomous Driving
|
Image instance segmentation is a fundamental research topic in autonomous driving, which is crucial for scene understanding and road safety. Advanced learning-based approaches often rely on the costly 2D mask annotations for training. In this paper, we present a more artful framework, LiDAR-guided Weakly Supervised Instance Segmentation (LWSIS), which leverages the off-the-shelf 3D data, i.e., Point Cloud, together with the 3D boxes, as natural weak supervisions for training the 2D image instance segmentation models. Our LWSIS not only exploits the complementary information in multimodal data during training, but also significantly reduces the annotation cost of the dense 2D masks. In detail, LWSIS consists of two crucial modules, Point Label Assignment (PLA) and Graph-based Consistency Regularization (GCR). The former module aims to automatically assign the 3D point cloud as 2D point-wise labels, while the latter further refines the predictions by enforcing geometry and appearance consistency of the multimodal data. Moreover, we conduct a secondary instance segmentation annotation on the nuScenes, named nuInsSeg, to encourage further research on multimodal perception tasks. Extensive experiments on the nuInsSeg, as well as the large-scale Waymo, show that LWSIS can substantially improve existing weakly supervised segmentation models by only involving 3D data during training. Additionally, LWSIS can also be incorporated into 3D object detectors like PointPainting to boost the 3D detection performance for free. The code and dataset are available at https://github.com/Serenos/LWSIS.
|
['Jianbing Shen', 'Ruigang Yang', 'Yikang Li', 'Botian Shi', 'Junbo Yin', 'Xiang Li']
|
2022-12-07
| null | null | null | null |
['weakly-supervised-instance-segmentation']
|
['computer-vision']
|
[ 9.87017304e-02 4.20500904e-01 -4.40899640e-01 -5.57503879e-01
-8.65223408e-01 -5.43734670e-01 4.40590113e-01 8.84241760e-02
-3.65748525e-01 3.13078195e-01 -4.36615735e-01 -4.74185616e-01
3.25203300e-01 -8.28671396e-01 -1.15331960e+00 -6.51528597e-01
3.28967780e-01 6.28108025e-01 6.74758673e-01 -3.99406642e-01
2.29240417e-01 5.58340847e-01 -1.83038759e+00 4.00918536e-03
1.27737677e+00 1.10336661e+00 4.77481335e-01 2.86783189e-01
-3.03344101e-01 1.92351580e-01 6.83147600e-03 -5.04716873e-01
4.31439131e-01 8.85021314e-03 -4.57435817e-01 3.60403150e-01
6.76706910e-01 -2.28059888e-01 -5.73380999e-02 1.17514765e+00
1.96802303e-01 2.72902459e-01 4.98295367e-01 -1.54210639e+00
-1.21600747e-01 5.51734604e-02 -9.38021898e-01 -4.00604159e-01
-1.71293944e-01 3.30430686e-01 8.76671731e-01 -1.09454691e+00
3.83764178e-01 1.12511563e+00 7.01783717e-01 3.98849726e-01
-9.78776753e-01 -7.57188141e-01 3.87581646e-01 1.02063715e-01
-1.42543411e+00 -2.75255382e-01 1.22363067e+00 -4.69393760e-01
6.19166851e-01 3.02112788e-01 5.37125051e-01 7.86870718e-01
-3.29667568e-01 9.47414994e-01 9.98108089e-01 -2.47736037e-01
1.56989783e-01 1.97886020e-01 2.60529429e-01 1.03518248e+00
6.65469915e-02 1.57643735e-01 -3.91705930e-01 1.64612919e-01
7.76752114e-01 -1.35857999e-01 8.07236508e-02 -6.69658661e-01
-1.07998216e+00 8.30495000e-01 6.70373976e-01 -3.01057607e-01
4.52038422e-02 3.25553231e-02 1.47540390e-01 -2.06245497e-01
7.30958462e-01 1.17254769e-02 -5.09767056e-01 1.19000688e-01
-7.63499618e-01 1.44648105e-01 3.87745231e-01 1.33876836e+00
1.52019751e+00 -3.78597200e-01 7.07493424e-02 9.77822006e-01
4.91831481e-01 6.27639055e-01 -1.91218719e-01 -1.35010004e+00
7.80871809e-01 9.24811482e-01 -4.51414362e-02 -7.99239039e-01
-3.17356557e-01 -2.64740139e-01 -5.59512675e-01 3.79448146e-01
3.63356352e-01 3.14836390e-03 -1.23756337e+00 1.47787249e+00
7.46963620e-01 3.18403512e-01 -1.86010540e-01 1.05134678e+00
1.05452812e+00 5.36703289e-01 -3.16635221e-02 3.69515210e-01
1.13940740e+00 -1.20003653e+00 -2.74379671e-01 -6.00597858e-01
8.16945732e-01 -6.15696490e-01 1.21406281e+00 6.25272840e-02
-9.55528140e-01 -8.08611691e-01 -9.29339409e-01 -5.30069351e-01
-4.35990334e-01 4.72016841e-01 6.96181715e-01 4.37725991e-01
-7.49739528e-01 1.35960162e-01 -1.04842782e+00 -2.13144764e-01
8.08918476e-01 1.59639448e-01 -4.04424369e-01 -1.73103884e-01
-9.23683465e-01 7.40848422e-01 3.83707792e-01 2.71648407e-01
-6.30364776e-01 -6.61210716e-01 -1.38031375e+00 -2.88343400e-01
7.60892272e-01 -6.08698010e-01 1.09310675e+00 -5.12861729e-01
-1.23120141e+00 1.34417629e+00 -3.62665772e-01 -2.35245317e-01
6.09325588e-01 -1.70549512e-01 1.11719266e-01 1.30563736e-01
3.82097989e-01 1.35149789e+00 8.69590759e-01 -1.81014550e+00
-8.06400359e-01 -4.97950792e-01 9.86088812e-03 3.18287581e-01
1.74107105e-01 -6.54400170e-01 -1.03975046e+00 -1.99329406e-01
4.83985543e-01 -9.24299002e-01 -4.19418246e-01 1.48209885e-01
-7.31168032e-01 -4.13415968e-01 9.62881744e-01 -4.29827064e-01
5.49026251e-01 -2.28009319e+00 6.73950315e-02 3.10211271e-01
5.76978996e-02 1.62059925e-02 -2.19547138e-01 6.99063689e-02
1.31693631e-01 -1.64540876e-02 -8.36673677e-01 -8.40597928e-01
2.15466946e-01 6.43698096e-01 -2.30684504e-01 3.73617500e-01
7.13306606e-01 1.22554541e+00 -8.14421356e-01 -6.37792945e-01
6.61191881e-01 2.29514435e-01 -5.44039607e-01 1.55245170e-01
-5.28412819e-01 7.15504169e-01 -4.90340561e-01 7.85004497e-01
1.14633322e+00 8.55875388e-03 -3.83442730e-01 -4.36430097e-01
-4.17221695e-01 1.03065208e-01 -9.70917225e-01 2.04916859e+00
-2.43827283e-01 3.78625304e-01 2.64007866e-01 -1.13196945e+00
9.77764487e-01 -2.36835733e-01 3.89451534e-01 -6.72507763e-01
1.07297726e-01 2.87585646e-01 -4.45946664e-01 -6.35382116e-01
5.02634525e-01 1.08446814e-01 -6.34626821e-02 -1.40229896e-01
-1.55945063e-01 -8.21561933e-01 1.27186954e-01 3.12844008e-01
4.82534200e-01 5.95712900e-01 -2.46046230e-01 1.59657258e-03
5.28504014e-01 4.93857384e-01 6.50779009e-01 5.49180865e-01
-5.38490638e-02 7.43352234e-01 3.99084985e-01 1.51193477e-02
-9.45984483e-01 -9.52537298e-01 -3.87395829e-01 7.21135736e-01
8.17257822e-01 -1.51926324e-01 -7.56255627e-01 -9.13331211e-01
2.72111744e-01 7.84447253e-01 -4.49692100e-01 -5.84197268e-02
-4.59251106e-01 -4.46863592e-01 4.09605145e-01 6.88814580e-01
7.36879587e-01 -6.14639282e-01 -2.91632921e-01 -3.26638758e-01
-2.55949795e-01 -1.44203138e+00 -4.20695245e-01 2.69643635e-01
-8.71352792e-01 -1.08667421e+00 -3.94219786e-01 -7.73210526e-01
9.16454673e-01 5.66627741e-01 1.00928807e+00 3.19771022e-01
-1.06871448e-01 3.71677071e-01 -2.52331585e-01 -4.63696837e-01
-6.88905418e-02 1.65607437e-01 -2.78499544e-01 -2.54447460e-02
4.39312249e-01 -4.17281806e-01 -4.06898320e-01 6.43629313e-01
-6.77425623e-01 4.85537171e-01 7.34843612e-01 5.77962279e-01
1.14075136e+00 -3.60594168e-02 2.21456826e-01 -9.90743995e-01
-3.11198264e-01 -2.50993580e-01 -7.69380987e-01 -1.98726267e-01
-2.35333368e-01 -2.10702673e-01 -4.98082973e-02 7.70216016e-03
-1.15262508e+00 4.82271552e-01 -5.60229063e-01 -5.39267361e-01
-6.28427207e-01 3.98233503e-01 -5.18622041e-01 -3.40162545e-01
3.74642283e-01 2.24000905e-02 9.85176861e-02 -5.93580246e-01
5.95494151e-01 5.98749280e-01 6.34336114e-01 -6.80468857e-01
1.11866748e+00 8.10574293e-01 1.30631298e-01 -9.61892724e-01
-1.22221100e+00 -8.21135342e-01 -1.04801178e+00 -2.23139435e-01
9.83971953e-01 -1.18878031e+00 -4.96149957e-01 4.33848947e-01
-1.07126153e+00 -5.95870972e-01 -2.49713063e-01 3.63867640e-01
-6.37752533e-01 4.51966643e-01 -3.21189642e-01 -8.18987131e-01
2.09939733e-01 -1.23279691e+00 1.74298751e+00 1.47727430e-01
2.71479398e-01 -8.14474583e-01 -3.08041632e-01 1.01764941e+00
-2.86732823e-01 2.26161525e-01 7.89350510e-01 -1.86087310e-01
-9.37387943e-01 -1.72643691e-01 -5.46479046e-01 5.43076277e-01
-1.45131201e-01 -5.42016216e-02 -1.21039188e+00 1.05481647e-01
-2.78389663e-01 -4.85161692e-01 1.17105651e+00 3.55518281e-01
1.32470679e+00 3.16069812e-01 -4.54220802e-01 8.39388371e-01
1.03730834e+00 -2.16239393e-01 5.66169858e-01 1.38260350e-01
1.33091903e+00 9.53077316e-01 1.15657902e+00 3.80409360e-02
8.58744621e-01 6.24084651e-01 9.27308142e-01 -5.17231286e-01
-2.76943386e-01 -5.72898746e-01 7.74136484e-02 7.07147062e-01
-1.82461455e-01 8.43433887e-02 -9.48675096e-01 3.49239230e-01
-2.06329560e+00 -4.92328793e-01 -6.87901437e-01 2.02034044e+00
6.15207613e-01 2.96774775e-01 -1.65900607e-02 4.58226539e-02
5.93299150e-01 9.43178311e-02 -7.11372197e-01 1.40549794e-01
-4.71270718e-02 -8.61071609e-03 7.89235413e-01 7.61483908e-01
-1.33233023e+00 1.30190277e+00 4.39146471e+00 1.04327130e+00
-7.60778189e-01 1.50724009e-01 6.16797924e-01 2.13995397e-01
-3.02186370e-01 7.90727213e-02 -1.00936615e+00 3.23778331e-01
3.03649545e-01 5.24585783e-01 7.03182966e-02 9.08800840e-01
4.70369279e-01 -3.04747343e-01 -9.36829865e-01 1.01674914e+00
-1.08543634e-01 -1.26690149e+00 -1.54996580e-02 1.38053522e-01
7.66422629e-01 3.35667104e-01 -2.34730691e-02 3.00367594e-01
-6.76182099e-03 -7.41748929e-01 8.58850241e-01 3.35072517e-01
7.94154048e-01 -8.27320218e-01 7.54215837e-01 7.91009486e-01
-1.26584971e+00 1.57171786e-01 -4.00528103e-01 5.48540726e-02
2.29421139e-01 7.20836818e-01 -5.03081858e-01 8.53625417e-01
7.24349201e-01 1.09925640e+00 -7.02372193e-01 1.00876307e+00
-5.97401738e-01 5.29057026e-01 -5.47221303e-01 4.56376642e-01
4.72249776e-01 -5.94260216e-01 6.20651484e-01 9.68576789e-01
8.39425176e-02 3.95816937e-02 5.06472230e-01 1.11701596e+00
-2.87580397e-02 -1.82200819e-01 -4.96158600e-01 3.52045119e-01
3.65327567e-01 1.48172379e+00 -7.83877611e-01 -1.77129507e-01
-4.26823050e-01 7.52837002e-01 3.56284142e-01 4.77426171e-01
-8.81553173e-01 -1.41184613e-01 6.85541093e-01 2.28254050e-01
3.32486600e-01 -5.91555178e-01 -7.36850917e-01 -1.04823029e+00
2.12584794e-01 -3.80985647e-01 7.14839175e-02 -7.92668045e-01
-1.32080984e+00 2.02273130e-01 1.06155559e-01 -1.33031762e+00
3.12178671e-01 -6.34077251e-01 -5.88665545e-01 9.20470536e-01
-2.07211065e+00 -1.40989637e+00 -7.08839655e-01 4.70416933e-01
5.98630726e-01 3.81373316e-01 3.13140243e-01 3.25624734e-01
-8.31706166e-01 2.64078379e-01 -4.56949621e-01 3.66895795e-02
5.70446193e-01 -1.23162270e+00 4.01368976e-01 7.11450219e-01
1.13036752e-01 2.47164652e-01 3.13674301e-01 -7.02401340e-01
-1.14908218e+00 -1.45191026e+00 5.15218794e-01 -7.02302933e-01
4.70898509e-01 -6.04521453e-01 -1.08391786e+00 5.22845209e-01
-2.48043329e-01 8.13754201e-02 2.77246565e-01 -1.28615275e-01
-1.48253888e-01 -5.98274283e-02 -9.55833316e-01 5.56239247e-01
1.38445771e+00 -4.45187539e-01 -4.03279215e-01 5.54126918e-01
9.31293666e-01 -9.01807010e-01 -5.37467003e-01 7.19395995e-01
1.61297925e-04 -9.34542537e-01 1.10953033e+00 -2.35483482e-01
4.44030017e-01 -6.72211289e-01 -1.62333176e-01 -8.73539329e-01
9.75813493e-02 -2.36171171e-01 -4.00609821e-02 1.17414880e+00
5.11650860e-01 -5.52392721e-01 1.09525490e+00 6.74516022e-01
-8.37961674e-01 -8.27792168e-01 -1.05353677e+00 -5.88192940e-01
1.87035482e-02 -1.07116079e+00 4.70486134e-01 8.01337004e-01
-5.42698562e-01 3.66685331e-01 2.15258151e-02 6.69416964e-01
8.46285880e-01 2.10308492e-01 1.15229857e+00 -1.25539958e+00
5.71233630e-02 -3.68047565e-01 -4.19062376e-01 -1.60987782e+00
3.24114680e-01 -1.07426667e+00 4.14929807e-01 -1.58522487e+00
-1.60340846e-01 -8.05647552e-01 2.24960074e-01 7.14107335e-01
-2.88882017e-01 5.28372228e-01 -6.71835914e-02 2.38860235e-01
-5.80263853e-01 8.52270246e-01 1.49217594e+00 -2.86554873e-01
-2.70880938e-01 2.62751430e-01 -4.51126516e-01 1.05444920e+00
7.95172811e-01 -2.89959699e-01 -4.91307348e-01 -4.70631689e-01
-3.55680916e-03 -3.83234084e-01 8.61462772e-01 -7.24099159e-01
2.04321802e-01 -2.56072193e-01 1.63766786e-01 -1.28248751e+00
6.60340846e-01 -9.25751984e-01 -3.57423067e-01 -4.22629416e-02
6.08489616e-04 -4.48110163e-01 3.99841815e-01 6.62828743e-01
-2.10064307e-01 -2.71278709e-01 7.36532509e-01 -6.66900575e-02
-9.64881897e-01 6.78836644e-01 2.00047165e-01 -1.38752600e-02
1.09112191e+00 -4.31210577e-01 -2.28632823e-01 1.42967822e-02
-5.69165349e-01 7.72429526e-01 6.55111194e-01 4.02588755e-01
6.93968296e-01 -1.13595831e+00 -5.10192096e-01 4.82117593e-01
5.02674997e-01 1.07981980e+00 3.65713716e-01 1.07618928e+00
-3.62341404e-01 1.62109554e-01 2.70973146e-01 -1.18211544e+00
-1.14785647e+00 2.42890522e-01 2.34777734e-01 3.19932461e-01
-8.18380535e-01 1.05501509e+00 3.99173051e-01 -8.66206408e-01
2.76331663e-01 -4.27877605e-01 -3.44135761e-02 -7.17916340e-02
3.21696326e-02 1.49214879e-01 1.40145913e-01 -8.20745230e-01
-3.20456445e-01 9.63081956e-01 1.25855058e-01 -1.62875541e-02
1.12799442e+00 -3.59052479e-01 -1.63694158e-01 5.38470805e-01
9.41713512e-01 -6.80819973e-02 -1.74394226e+00 -1.55198559e-01
-5.81748299e-02 -5.26854038e-01 2.05250621e-01 -5.16835332e-01
-1.15829337e+00 1.11057854e+00 2.81807214e-01 -7.21888617e-04
8.04172933e-01 3.36789966e-01 9.26067829e-01 3.22764337e-01
3.87126356e-01 -1.06794024e+00 -1.07011221e-01 4.63476330e-01
7.12916493e-01 -1.75036716e+00 -1.42261267e-01 -1.12498569e+00
-7.33900785e-01 8.04014027e-01 8.27484608e-01 -8.36356170e-03
5.66829920e-01 -4.82418016e-02 2.32465431e-01 -3.09425861e-01
-2.83496052e-01 -6.27904177e-01 5.93086839e-01 7.16873109e-01
-3.72510813e-02 -1.86014045e-02 3.66927832e-02 5.14305353e-01
-1.94790632e-01 -3.38411540e-01 2.44897250e-02 6.97889388e-01
-5.36517560e-01 -1.12832761e+00 -4.48958278e-01 3.48958254e-01
4.75837737e-01 -2.11665295e-02 -4.94535923e-01 8.24290156e-01
7.22488999e-01 9.11730528e-01 1.09903693e-01 -4.52275604e-01
5.59570372e-01 -1.19067580e-01 3.14681530e-01 -8.54023755e-01
1.45575963e-02 1.43863782e-02 1.69804748e-02 -6.92944765e-01
-5.16005158e-01 -6.28941298e-01 -1.65813816e+00 2.22754143e-02
-3.86440605e-01 5.98792806e-02 8.16068113e-01 1.04785979e+00
3.71067613e-01 3.48978460e-01 5.33676326e-01 -1.41313910e+00
-8.37034211e-02 -6.62339807e-01 -3.78814995e-01 2.86414504e-01
3.20231140e-01 -1.01005852e+00 -4.83399004e-01 2.85757966e-02]
|
[8.065104484558105, -2.9501547813415527]
|
e34938a1-dd63-4e58-91f5-a0bfc944edb0
|
a-fully-spiking-hybrid-neural-network-for
|
2104.10719
| null |
https://arxiv.org/abs/2104.10719v2
|
https://arxiv.org/pdf/2104.10719v2.pdf
|
A Fully Spiking Hybrid Neural Network for Energy-Efficient Object Detection
|
This paper proposes a Fully Spiking Hybrid Neural Network (FSHNN) for energy-efficient and robust object detection in resource-constrained platforms. The network architecture is based on Convolutional SNN using leaky-integrate-fire neuron models. The model combines unsupervised Spike Time-Dependent Plasticity (STDP) learning with back-propagation (STBP) learning methods and also uses Monte Carlo Dropout to get an estimate of the uncertainty error. FSHNN provides better accuracy compared to DNN based object detectors while being 150X energy-efficient. It also outperforms these object detectors, when subjected to noisy input data and less labeled training data with a lower uncertainty error.
|
['Saibal Mukhopadhyay', 'Xueyuan She', 'Biswadeep Chakraborty']
|
2021-04-21
| null | null | null | null |
['robust-object-detection']
|
['computer-vision']
|
[ 2.27000669e-01 -4.45053518e-01 2.13208437e-01 -1.44182533e-01
-3.74582082e-01 -1.63823813e-01 5.56439996e-01 1.74806431e-01
-1.02961481e+00 1.16978049e+00 -3.89333278e-01 1.94915086e-01
2.34891176e-02 -8.74581277e-01 -9.89548326e-01 -9.46981668e-01
1.66225851e-01 3.52534354e-01 1.02826321e+00 1.59863830e-01
4.53763485e-01 6.67483151e-01 -1.52882731e+00 1.13310270e-01
3.87525767e-01 1.22545207e+00 4.64646190e-01 5.78687489e-01
1.87375888e-01 6.49126470e-01 -5.27963042e-01 8.86374936e-02
3.29287052e-01 -3.50476086e-01 -7.00507825e-03 -8.44120681e-01
-1.67710669e-02 2.08667833e-02 -6.03962362e-01 9.75891411e-01
7.85771310e-01 1.49857681e-02 7.12548077e-01 -1.28208125e+00
-4.17119563e-01 5.69609046e-01 1.26661435e-01 7.97463596e-01
-3.35689217e-01 6.45338356e-01 1.33868545e-01 -9.62860227e-01
6.05544209e-01 8.43468010e-01 8.56279910e-01 5.17063856e-01
-1.44005895e+00 -9.27910149e-01 -6.12764955e-01 -9.71204042e-02
-1.49812520e+00 -4.16160583e-01 2.16990471e-01 4.28203642e-02
1.93273461e+00 -1.99926123e-01 9.80823457e-01 1.12312210e+00
8.32715094e-01 6.54469371e-01 1.28510427e+00 -8.92194510e-02
1.14642489e+00 -2.73646057e-01 2.73864686e-01 4.29459304e-01
8.78464341e-01 4.71034646e-01 -1.03596115e+00 -9.37755927e-02
8.29225004e-01 1.70101777e-01 1.41892746e-01 2.48108253e-01
-8.88623774e-01 3.46865147e-01 7.77889252e-01 3.78221035e-01
-4.60272938e-01 1.12949932e+00 2.21964180e-01 -1.33769691e-01
-2.14782938e-01 2.45042965e-01 -3.61614317e-01 -1.13778077e-01
-1.32292366e+00 3.02343994e-01 7.16139495e-01 8.43753755e-01
5.20173788e-01 6.94014788e-01 -2.73792118e-01 3.29998970e-01
8.38985503e-01 7.97215283e-01 5.64584851e-01 -9.98654544e-01
-1.52334407e-01 6.63135111e-01 -1.07371237e-03 -2.22715870e-01
-3.96281630e-01 -3.33697796e-01 -6.50879025e-01 1.11495531e+00
2.26677120e-01 -2.24280730e-02 -1.58655655e+00 1.22358370e+00
-5.52725971e-01 1.95603877e-01 1.41769692e-01 6.01343513e-01
6.82589173e-01 7.30310857e-01 1.94388911e-01 7.03179166e-02
1.14518452e+00 -5.60757816e-01 -7.56756365e-01 -5.33131778e-01
2.62142662e-02 1.38441985e-02 9.78644565e-02 3.35429281e-01
-1.30204523e+00 -1.20280810e-01 -1.60825372e+00 -1.13232143e-01
-8.09161544e-01 -5.36143631e-02 2.31548712e-01 9.33648348e-01
-1.44575441e+00 9.39633071e-01 -1.37674630e+00 -3.14185679e-01
1.01108932e+00 9.73371029e-01 1.44522220e-01 2.67995358e-01
-6.43103898e-01 1.20979726e+00 7.75017858e-01 -2.74551332e-01
-1.16507745e+00 -4.55740988e-01 -3.95991504e-01 1.10248521e-01
-3.51267457e-01 -5.89431047e-01 1.08474147e+00 -6.05193198e-01
-1.77599847e+00 5.06931663e-01 2.10763160e-02 -1.17430615e+00
1.01993047e-01 2.88623959e-01 -1.37535244e-01 1.42109871e-01
-5.61824918e-01 1.15889895e+00 4.30984676e-01 -8.35029423e-01
-2.41508335e-01 -4.04550910e-01 -8.24492633e-01 -2.64456362e-01
2.61039853e-01 2.28511170e-01 -2.83403173e-02 -4.55365419e-01
4.59022284e-01 -7.50383794e-01 -1.42619610e-01 5.11441171e-01
-1.03053235e-01 -3.21582481e-02 9.29446042e-01 -2.63317734e-01
5.73099494e-01 -1.95203424e+00 -2.43879735e-01 1.13332361e-01
-4.37753946e-02 3.42093170e-01 1.52119547e-01 1.06519595e-01
4.82484400e-01 -5.65204583e-02 -7.45350122e-01 -3.49137723e-01
-7.16879219e-02 4.72299784e-01 3.30800563e-02 4.39185530e-01
5.95476091e-01 1.14612186e+00 -6.17063582e-01 -2.63290554e-01
8.67810398e-02 8.01143289e-01 -8.71799812e-02 -1.75716385e-01
-2.21252382e-01 7.94137716e-02 -9.36841220e-02 1.05137265e+00
6.91673636e-01 5.96318431e-02 -3.34044009e-01 -7.35546136e-03
-5.97707868e-01 1.23971840e-02 -1.16744554e+00 1.69079518e+00
2.22360611e-01 8.38724971e-01 -2.06754088e-01 -6.64845347e-01
1.32364106e+00 5.30060418e-02 1.32487401e-01 -1.00935698e+00
6.69449329e-01 9.38627124e-01 -6.12601638e-02 5.10070324e-02
2.01703221e-01 -3.64095643e-02 3.10450971e-01 9.33400020e-02
9.19322252e-01 -2.85051391e-02 9.97766387e-03 -5.08493595e-02
1.73016357e+00 3.36277515e-01 -1.26008093e-01 -5.90530992e-01
-3.25313240e-01 -2.63560921e-01 6.41667247e-01 7.99035907e-01
-3.77443314e-01 6.77241206e-01 8.56379233e-03 -2.33641222e-01
-1.14013720e+00 -1.44879496e+00 -5.08842051e-01 3.81674260e-01
2.78617859e-01 2.17087612e-01 -5.81029177e-01 1.72177598e-01
-1.98828112e-02 6.91058397e-01 -4.59352255e-01 -4.20960754e-01
-3.54918271e-01 -1.20535123e+00 1.06082296e+00 9.18850243e-01
7.69401848e-01 -1.45224285e+00 -1.45029712e+00 6.70952380e-01
6.31382346e-01 -7.23280311e-01 4.01206315e-01 1.55188358e+00
-1.11047685e+00 -5.92282295e-01 -7.50874460e-01 -6.03154719e-01
5.57927430e-01 -5.69427669e-01 8.38236928e-01 -2.37994641e-01
-7.63982654e-01 -6.87424242e-02 -6.34310171e-02 -7.42496490e-01
1.01937503e-01 -5.31971037e-01 5.15105277e-02 -5.24290919e-01
7.51320004e-01 -6.25190556e-01 -7.06105351e-01 -1.00672916e-01
-9.24175203e-01 -4.46475625e-01 7.01426744e-01 6.95906997e-01
1.22500932e+00 -7.36709908e-02 4.59567070e-01 -2.09147424e-01
2.19928041e-01 -4.62231934e-01 -9.93559957e-01 -1.07723445e-01
-8.39287698e-01 3.89280587e-01 3.40833127e-01 -3.90115589e-01
-8.56559992e-01 4.71796393e-01 -9.18307528e-02 -2.52482504e-01
4.02531354e-03 -1.36791885e-01 3.26113969e-01 -8.38419795e-01
9.68230307e-01 5.56708395e-01 -4.11406338e-01 -2.47662634e-01
-4.10891026e-01 7.51423761e-02 6.50469184e-01 3.96875927e-04
1.63434818e-01 7.18217731e-01 1.85025305e-01 -3.76757145e-01
2.16415241e-01 -9.41299498e-02 -2.29066744e-01 -2.38171637e-01
8.29309106e-01 -8.79025996e-01 -7.54897952e-01 1.14660966e+00
-1.07458282e+00 -6.20503187e-01 -4.50553745e-01 7.20889747e-01
-7.99842894e-01 -4.23022121e-01 -8.39329600e-01 -1.20400047e+00
-6.20644987e-01 -9.54599917e-01 4.68506306e-01 7.98933387e-01
2.92636037e-01 -5.00013053e-01 2.56677359e-01 -6.05581224e-01
1.01603889e+00 2.65725493e-01 1.60458907e-01 -7.76859760e-01
-8.92942607e-01 -3.50350946e-01 -2.78444469e-01 2.82670051e-01
-6.33526623e-01 3.01790893e-01 -1.22549391e+00 -1.04211025e-01
-1.75231509e-02 -4.22342032e-01 1.49741471e+00 1.02624094e+00
8.56067955e-01 -4.17751037e-02 -6.17284715e-01 6.22805297e-01
2.28000522e+00 2.78839707e-01 9.83121932e-01 4.31680679e-01
-9.84123051e-02 -1.96146458e-01 -2.78679430e-01 4.59771842e-01
-5.18382005e-02 1.45449311e-01 8.02171052e-01 4.38474596e-01
-2.44506434e-01 6.91946894e-02 3.32271785e-01 3.71151924e-01
-3.05793267e-02 -5.58575690e-01 -1.09979033e+00 5.24981022e-01
-1.71948051e+00 -9.21956301e-01 -2.74191648e-01 2.22462606e+00
7.62298584e-01 6.92914426e-01 -4.89645116e-02 2.53320664e-01
6.33707047e-01 -5.35627067e-01 -9.17235494e-01 -3.82786810e-01
-8.37814450e-01 7.64938951e-01 1.28400040e+00 -1.61526948e-01
-6.12741053e-01 4.41074312e-01 6.95087862e+00 5.00124514e-01
-9.10987318e-01 3.31389815e-01 4.04251039e-01 -3.95816416e-01
6.89973682e-02 -1.29170939e-01 -1.14273047e+00 9.76098537e-01
1.71826470e+00 -1.20199487e-01 4.49135363e-01 4.51332569e-01
-2.06478208e-01 -7.86929488e-01 -7.70800769e-01 1.04644728e+00
-2.74228603e-01 -1.72980928e+00 -5.03652334e-01 -1.80272549e-01
5.50106585e-01 1.07354093e+00 -5.25250146e-03 2.66152501e-01
4.74769920e-01 -9.79847193e-01 9.52681124e-01 9.80053246e-01
2.94271737e-01 -6.97643220e-01 8.67753446e-01 3.29404056e-01
-9.75951195e-01 -2.86559284e-01 -7.95944035e-01 -7.33363777e-02
-4.23884355e-02 6.68866456e-01 -1.39878318e-01 -3.03472489e-01
1.20997393e+00 2.21259922e-01 -3.35173637e-01 2.08538127e+00
9.79197472e-02 4.37258780e-01 -1.17400849e+00 -6.39488518e-01
2.37361446e-01 1.84208959e-01 3.80178481e-01 1.38982403e+00
7.05680549e-01 1.89963996e-01 -3.17053884e-01 1.48463774e+00
-4.48126197e-01 -4.88158256e-01 -4.14874941e-01 -8.41065198e-02
8.47708166e-01 1.07725394e+00 -1.41015875e+00 -2.16702878e-01
9.19388160e-02 7.97647893e-01 -2.71296874e-02 2.47098505e-01
-4.80587631e-01 -6.38028204e-01 6.00676909e-02 4.39203568e-02
5.91589868e-01 -1.95791721e-01 -7.71877170e-01 -4.19721872e-01
-2.15408981e-01 3.12361360e-01 -6.96086511e-02 -1.08046627e+00
-8.87686014e-01 5.63625038e-01 -3.92464250e-01 -8.83758605e-01
2.28548855e-01 -1.02448559e+00 -8.60169709e-01 9.17773366e-01
-1.53003979e+00 -5.97453117e-01 -8.40035081e-02 2.98462272e-01
1.49526238e-01 -8.10599700e-02 8.96714330e-01 1.33237973e-01
-4.14907962e-01 2.92755425e-01 5.45723140e-01 6.61175400e-02
1.68479070e-01 -1.15207744e+00 2.37482294e-01 9.05861735e-01
-2.22329170e-01 2.06480622e-01 7.38235176e-01 -7.10841000e-01
-1.40410256e+00 -1.30725050e+00 4.90711868e-01 -1.11603037e-01
3.18670660e-01 -2.99803942e-01 -1.09922850e+00 3.28945428e-01
4.85563278e-01 6.21822059e-01 4.52878296e-01 -1.01804709e+00
-2.67586142e-01 -3.75969298e-02 -1.90679097e+00 8.38470235e-02
6.64049268e-01 -2.62847513e-01 -5.42533934e-01 3.52680162e-02
1.96786687e-01 -1.24197125e-01 -5.34111381e-01 4.35777038e-01
3.94150794e-01 -8.04376662e-01 5.06146729e-01 1.75019845e-01
-2.44884372e-01 -5.33221185e-01 -1.69960424e-01 -9.68730927e-01
-4.71194267e-01 -3.52559507e-01 -5.03165424e-01 8.15468967e-01
5.93399286e-01 -6.05953634e-01 8.57945919e-01 4.98445004e-01
-3.59004676e-01 -6.67902231e-01 -1.57162547e+00 -1.33486390e+00
-1.84035450e-01 -2.51989961e-01 3.19771818e-03 2.80099101e-02
-1.65367618e-01 -4.25210238e-01 3.17667067e-01 6.54768795e-02
9.72063184e-01 -6.45501316e-01 -4.89745617e-01 -1.16272271e+00
2.03637660e-01 -7.26977289e-01 -1.03392804e+00 -1.15744047e-01
-3.77868533e-01 -6.72595024e-01 6.47381783e-01 -1.68252325e+00
3.14442627e-02 -1.78435698e-01 -1.00631213e+00 6.97139800e-01
4.65516478e-01 8.36779892e-01 -1.37501925e-01 1.81549102e-01
-7.08419919e-01 3.03377867e-01 1.18148960e-01 -8.05739835e-02
1.27947584e-01 -1.92392305e-01 2.05888942e-01 4.88191605e-01
1.05889368e+00 -1.28712380e+00 1.21976547e-01 -3.07299554e-01
3.09882730e-01 -4.57859188e-01 6.53783143e-01 -1.98511350e+00
1.04954684e+00 3.42716277e-01 8.47813129e-01 -7.21441150e-01
4.78261143e-01 -7.01865613e-01 3.94125998e-01 1.18472588e+00
-3.45979631e-02 -6.02841265e-02 5.92841506e-01 7.41484761e-01
9.01648253e-02 -5.66698015e-01 1.30533969e+00 -2.71912843e-01
-7.05331326e-01 2.27716416e-01 -9.73388016e-01 -2.65818745e-01
1.07801545e+00 -9.96631622e-01 -6.00174010e-01 4.89579707e-01
-5.62465489e-01 -2.22147584e-01 4.92824644e-01 -1.57620579e-01
9.89579082e-01 -1.24050152e+00 -5.86196065e-01 2.07481697e-01
-2.09081899e-02 -2.79972494e-01 1.04855304e-03 7.43994892e-01
-9.09671783e-01 4.31360364e-01 -9.43857789e-01 -4.96243894e-01
-4.58747238e-01 2.18161598e-01 6.51000619e-01 2.47978210e-01
-3.08210522e-01 1.40963948e+00 -8.62386584e-01 3.75686884e-02
4.55368906e-01 -4.75358337e-01 1.90238819e-01 -2.45024234e-01
5.58637738e-01 6.00334287e-01 3.10481340e-01 1.84464455e-02
-6.83406234e-01 4.65897508e-02 3.34009439e-01 -3.59874576e-01
1.54906130e+00 4.26563531e-01 1.95310311e-03 8.44995379e-01
9.19628143e-01 -9.53487158e-01 -1.69724941e+00 -1.55638024e-01
3.87824148e-01 1.32779062e-01 6.69174194e-01 -1.31471455e+00
-9.75222766e-01 8.54598641e-01 1.27196610e+00 -1.67962372e-01
1.08030951e+00 -2.26037353e-01 5.40768743e-01 7.41936982e-01
4.49401438e-01 -1.55237818e+00 6.13882355e-02 5.51011026e-01
4.72149730e-01 -1.03477728e+00 -1.56122953e-01 3.96433145e-01
-7.52369240e-02 1.32554913e+00 6.84854984e-01 -8.31469834e-01
8.44932854e-01 1.24944162e+00 -5.32847404e-01 -1.01382442e-01
-1.04037523e+00 -1.49184495e-01 -2.76481152e-01 6.79512858e-01
-4.81311046e-02 -9.71273780e-02 -1.50316522e-01 6.18929207e-01
6.42211556e-01 6.12449586e-01 4.01567757e-01 1.36395895e+00
-1.09802961e+00 -4.41319466e-01 -1.07979871e-01 7.74916649e-01
-5.36319435e-01 -2.68843710e-01 -7.97314197e-02 3.04413766e-01
2.41207942e-01 4.36600089e-01 3.11462581e-01 -1.22169554e-01
9.30723250e-02 1.81318998e-01 7.76699007e-01 -4.58336949e-01
-1.00893176e+00 -5.69384545e-02 -4.63016599e-01 -4.92577344e-01
-3.61935854e-01 -4.07622606e-01 -2.06734324e+00 -4.31455262e-02
-5.69296062e-01 -4.30650622e-01 1.45031083e+00 8.37524593e-01
4.00731444e-01 7.14542866e-01 1.20010033e-01 -1.09179556e+00
-2.60093004e-01 -7.94495583e-01 -9.07418013e-01 -4.16558295e-01
3.53796929e-02 -6.95903301e-01 -4.35681939e-01 -8.45211223e-02]
|
[8.203936576843262, 2.4740843772888184]
|
01f5fcc8-96c3-42a2-b8db-34e5b8a4a665
|
multi-agent-collaborative-inference-via-dnn
|
2205.11854
| null |
https://arxiv.org/abs/2205.11854v2
|
https://arxiv.org/pdf/2205.11854v2.pdf
|
Multi-Agent Collaborative Inference via DNN Decoupling: Intermediate Feature Compression and Edge Learning
|
Recently, deploying deep neural network (DNN) models via collaborative inference, which splits a pre-trained model into two parts and executes them on user equipment (UE) and edge server respectively, becomes attractive. However, the large intermediate feature of DNN impedes flexible decoupling, and existing approaches either focus on the single UE scenario or simply define tasks considering the required CPU cycles, but ignore the indivisibility of a single DNN layer. In this paper, we study the multi-agent collaborative inference scenario, where a single edge server coordinates the inference of multiple UEs. Our goal is to achieve fast and energy-efficient inference for all UEs. To achieve this goal, we first design a lightweight autoencoder-based method to compress the large intermediate feature. Then we define tasks according to the inference overhead of DNNs and formulate the problem as a Markov decision process (MDP). Finally, we propose a multi-agent hybrid proximal policy optimization (MAHPPO) algorithm to solve the optimization problem with a hybrid action space. We conduct extensive experiments with different types of networks, and the results show that our method can reduce up to 56\% of inference latency and save up to 72\% of energy consumption.
|
['Shiwen Mao', 'Jianping An', 'Han Hu', 'Yong Luo', 'Guanyu Xu', 'Zhiwei Hao']
|
2022-05-24
| null | null | null | null |
['feature-compression']
|
['computer-vision']
|
[-2.10438296e-01 1.60491914e-01 -2.49105260e-01 -2.71329969e-01
-4.84369308e-01 -5.03586948e-01 3.04977447e-01 -4.94810641e-01
-5.70082724e-01 9.16100860e-01 -6.41714483e-02 -5.96663177e-01
-2.55092770e-01 -8.28442633e-01 -8.33636999e-01 -9.69289660e-01
2.08923072e-01 5.82380295e-01 9.94354784e-02 3.60644639e-01
-3.63141149e-01 3.71502846e-01 -1.18020475e+00 -2.06853926e-01
9.76968527e-01 1.31884277e+00 6.55614614e-01 7.98218966e-01
1.57616600e-01 1.37669861e+00 -6.47299469e-01 -4.02035147e-01
1.73075303e-01 -1.88634351e-01 -7.67604351e-01 -1.65715456e-01
-6.77093118e-02 -1.04012179e+00 -7.41515815e-01 1.23288488e+00
8.46604586e-01 4.28686321e-01 3.30497742e-01 -1.62007558e+00
-5.05921721e-01 6.46441221e-01 -9.43620354e-02 1.93490937e-01
-3.04037899e-01 -1.15647480e-01 8.22387934e-01 -3.70167673e-01
4.28690255e-01 9.78705168e-01 4.09775794e-01 8.17526042e-01
-7.61253417e-01 -7.48690546e-01 5.75409889e-01 5.63046038e-01
-1.19871509e+00 -8.46349716e-01 7.19584703e-01 1.74597085e-01
1.01684654e+00 1.02367513e-01 5.48748672e-01 1.27573490e+00
-4.92978990e-02 1.24285722e+00 6.19952261e-01 -1.14836968e-01
6.44613922e-01 -2.05384642e-01 -1.50000051e-01 8.11296642e-01
1.76340833e-01 -3.14174086e-01 -1.71852142e-01 -2.33462974e-01
7.02787638e-01 4.95102882e-01 -1.02557354e-01 1.36267126e-01
-1.06491542e+00 5.33698678e-01 1.88306212e-01 -1.45298138e-01
-8.94952953e-01 5.78862190e-01 3.74720544e-01 -2.65458673e-02
4.01871026e-01 -4.06340897e-01 -6.13778293e-01 -3.56766194e-01
-8.07807207e-01 -4.51631732e-02 1.16559947e+00 1.16876209e+00
5.32494068e-01 2.16663703e-01 -2.61959672e-01 5.83852947e-01
5.69689989e-01 4.97909725e-01 2.91977853e-01 -1.57239377e+00
4.89015609e-01 1.81750536e-01 1.54083282e-01 -5.67004204e-01
-2.71913469e-01 -3.02737534e-01 -1.15102327e+00 -5.82280792e-02
3.10399234e-02 -1.17926192e+00 -5.72929978e-01 1.92892277e+00
6.16887450e-01 6.08383358e-01 2.64963478e-01 8.62417817e-01
4.03089941e-01 9.11226749e-01 9.54485014e-02 -3.72056276e-01
1.24336970e+00 -1.59218812e+00 -8.69369328e-01 -9.28272977e-02
4.48054194e-01 -2.19376683e-01 4.41034228e-01 1.72269776e-01
-1.29361796e+00 -3.09638649e-01 -9.43392277e-01 -8.72194245e-02
-9.16403458e-02 4.80856419e-01 6.37556791e-01 3.57234895e-01
-1.29564738e+00 4.41624373e-01 -1.27155542e+00 9.49274451e-02
3.68901342e-01 7.67432153e-01 3.11415344e-01 -1.32100090e-01
-1.10965705e+00 5.72057009e-01 3.59871298e-01 3.97370875e-01
-1.24941790e+00 -4.65365231e-01 -3.29577863e-01 4.30807501e-01
7.27776766e-01 -1.14360642e+00 1.48547697e+00 -7.71187007e-01
-2.03596497e+00 -1.73664689e-02 -2.03661337e-01 -3.82761240e-01
3.92186284e-01 -9.55009684e-02 -5.01659989e-01 1.25076875e-01
-2.70580143e-01 4.86204475e-01 6.29334092e-01 -1.01513946e+00
-9.70597088e-01 -2.60747671e-01 7.43670702e-01 4.28430200e-01
-6.15372360e-01 -2.60110646e-02 -7.34693170e-01 -4.30806816e-01
-3.41916025e-01 -1.15322971e+00 -1.76481798e-01 -1.28105339e-02
-4.53077406e-01 -5.30557930e-01 1.10306668e+00 -8.45770597e-01
1.24236333e+00 -1.91057897e+00 2.56244004e-01 -9.66148823e-02
6.43833220e-01 1.98257089e-01 1.94859281e-01 1.36953620e-02
6.53375626e-01 -9.43553373e-02 1.83939144e-01 -8.54735732e-01
5.18559039e-01 7.15737939e-01 3.14308852e-02 2.52153166e-02
-5.08003533e-01 7.68742681e-01 -8.78669858e-01 -7.14162171e-01
-5.83355650e-02 5.61768234e-01 -6.63364649e-01 3.29252303e-01
-3.51124913e-01 4.31984738e-02 -8.68341863e-01 7.04399407e-01
5.69714308e-01 -5.18831313e-01 6.96359694e-01 -3.98647785e-01
1.49456263e-01 1.86566502e-01 -1.22381961e+00 1.53057301e+00
-7.08682597e-01 5.83513141e-01 6.71956897e-01 -1.02034831e+00
2.38860562e-01 6.22532189e-01 6.22615278e-01 -4.65622813e-01
1.86585605e-01 4.37723994e-02 -1.60838515e-01 -4.20613199e-01
1.71329618e-01 3.13279361e-01 2.19321594e-01 6.82893813e-01
1.97545752e-01 8.61791611e-01 9.29793417e-02 1.31811649e-01
1.36196303e+00 2.79338937e-02 -5.51304854e-02 -3.53090875e-02
9.25750062e-02 -4.92632538e-01 1.06536281e+00 6.65430009e-01
-4.53980595e-01 -1.11459918e-01 4.38960224e-01 -3.32615167e-01
-6.79555058e-01 -9.15876567e-01 4.79765952e-01 1.32041001e+00
2.71100879e-01 -2.46258125e-01 -1.08414900e+00 -8.56528878e-01
-3.70260417e-01 7.56804705e-01 -2.70539552e-01 -2.16283664e-01
-5.06575286e-01 -9.17220891e-01 5.82033694e-01 7.69939065e-01
1.09407198e+00 -8.57103705e-01 -6.12419069e-01 2.72803128e-01
-5.13216019e-01 -1.42003345e+00 -4.71341759e-01 1.77594155e-01
-8.52377534e-01 -5.73324084e-01 -5.06558537e-01 -7.51490355e-01
6.51627481e-01 5.75847775e-02 8.82052839e-01 -1.38334796e-01
3.64870667e-01 2.32804805e-01 8.77506286e-03 -3.05666327e-01
-4.26319912e-02 3.08280677e-01 3.42825592e-01 6.84831366e-02
1.22329265e-01 -8.08845162e-01 -7.48746097e-01 1.00337796e-01
-6.63160443e-01 2.63316214e-01 8.28305840e-01 6.12141073e-01
6.65086448e-01 3.15483302e-01 4.96821195e-01 -5.12254477e-01
6.12836242e-01 -7.14956224e-01 -6.36936903e-01 6.09357476e-01
-5.17357230e-01 1.38347819e-01 9.72114205e-01 -6.10056102e-01
-1.38119745e+00 -6.53060675e-02 -7.97968060e-02 -8.41031611e-01
5.50949238e-02 2.98820853e-01 -5.51715851e-01 1.54207930e-01
-9.68810916e-02 3.00902933e-01 -2.96626538e-01 -3.91709596e-01
2.38111764e-02 6.74097776e-01 3.07966501e-01 -7.77595282e-01
2.94226587e-01 4.05405730e-01 -1.56852994e-02 -3.78696829e-01
-7.94357538e-01 2.89775252e-01 5.60631324e-03 -3.69155586e-01
1.00416040e+00 -1.10973442e+00 -1.53123236e+00 3.30524296e-01
-1.32358646e+00 -8.04790318e-01 6.52889535e-02 7.12918460e-01
-4.05611247e-01 5.73632419e-02 -9.41843569e-01 -8.45405221e-01
-6.63691521e-01 -1.33693659e+00 8.50570023e-01 4.97229099e-01
4.08119857e-01 -1.12965834e+00 -4.06843647e-02 4.62720037e-01
4.64505821e-01 -9.89520475e-02 6.46649420e-01 -5.23787320e-01
-9.93524313e-01 9.31016505e-02 -2.86624312e-01 3.61374617e-01
-1.27455652e-01 -1.13251373e-01 -9.64060009e-01 -2.91340113e-01
6.42695278e-02 -1.13150351e-01 3.91145945e-01 5.02283275e-01
1.73005331e+00 -6.58153653e-01 -5.24635553e-01 7.99121261e-01
1.51763773e+00 4.95276153e-01 2.95113444e-01 -1.15486691e-02
8.93239081e-01 -5.65805286e-02 1.83305904e-01 6.03931904e-01
9.17099535e-01 5.74429333e-01 6.87013030e-01 2.93100402e-02
1.27923176e-01 -5.09992056e-02 6.66569710e-01 1.23392677e+00
-5.57051837e-01 -7.63319254e-01 -5.20089030e-01 2.03462243e-01
-2.28242850e+00 -7.73314118e-01 5.06379426e-01 1.62632120e+00
5.72807133e-01 1.71347875e-02 -6.71878979e-02 -2.79098034e-01
6.09099627e-01 8.22708309e-02 -1.07450187e+00 -1.27335399e-01
3.13797444e-01 -2.98343450e-01 6.78756714e-01 2.68268049e-01
-8.66949320e-01 6.28290296e-01 5.64940643e+00 1.01814008e+00
-9.11429465e-01 6.84551001e-01 5.81872940e-01 -5.41123450e-01
8.15146193e-02 -2.13324979e-01 -9.40787375e-01 9.92844224e-01
1.38851476e+00 1.79330632e-02 1.00657928e+00 1.10447443e+00
2.78239280e-01 1.99466780e-01 -1.00547385e+00 1.12807214e+00
-2.58211344e-01 -1.28924191e+00 -1.87359706e-01 3.23996037e-01
7.68629968e-01 5.34226120e-01 -3.31903428e-01 5.46365619e-01
8.62240613e-01 -4.46499884e-01 6.62840068e-01 8.09465230e-01
2.74430692e-01 -9.69238639e-01 6.47959769e-01 7.14273274e-01
-1.24162340e+00 -4.11760777e-01 -4.80204076e-01 -1.82138309e-02
3.25031430e-01 6.08347833e-01 -2.96380132e-01 5.69796443e-01
8.43211412e-01 3.17262918e-01 1.21699564e-01 5.19280016e-01
2.10379604e-02 6.03212416e-01 -5.80357432e-01 -2.45532066e-01
2.52969384e-01 -5.02648592e-01 2.57188767e-01 7.75211036e-01
3.65868092e-01 3.25201809e-01 2.86729366e-01 7.01146960e-01
-6.68204963e-01 -4.39907134e-01 -1.53092831e-01 1.00230746e-01
8.83468866e-01 1.37082851e+00 -6.04699314e-01 -7.43396819e-01
-5.71890712e-01 1.44553626e+00 6.40851080e-01 6.70993865e-01
-1.50229168e+00 -3.49228650e-01 9.64991868e-01 -7.01719582e-01
2.93667018e-01 -2.19946355e-01 1.55973092e-01 -1.27577353e+00
3.32210451e-01 -4.69332397e-01 8.97664130e-02 -9.00190055e-01
-7.40078926e-01 6.05124950e-01 -2.88937598e-01 -6.42949104e-01
-2.43052781e-01 -2.76408881e-01 -6.83905065e-01 4.94394869e-01
-1.50328887e+00 -1.07362080e+00 -3.80973071e-01 6.74316645e-01
5.38657367e-01 -2.15970203e-02 8.10740709e-01 9.17303801e-01
-1.42257845e+00 5.92966557e-01 6.62151217e-01 2.47247890e-01
3.21796834e-01 -1.20514107e+00 4.95047867e-02 7.78952539e-01
-2.59030789e-01 2.11584747e-01 2.26877317e-01 -3.69109541e-01
-1.80620432e+00 -1.33942842e+00 7.18232393e-01 -4.65693064e-02
3.74523461e-01 -3.43447030e-01 -2.74646133e-01 1.08981621e+00
3.49925190e-01 2.88172007e-01 5.10107100e-01 -1.87897637e-01
3.98034543e-01 -4.77845430e-01 -1.00740778e+00 8.43430519e-01
1.36037052e+00 -5.57496786e-01 1.11077294e-01 4.73881871e-01
9.95584071e-01 -3.61782730e-01 -1.13994217e+00 5.10263033e-02
4.18026268e-01 -4.95819062e-01 9.70127523e-01 -5.91836989e-01
1.34660363e-01 -2.94995934e-01 -3.63352597e-01 -1.26844740e+00
-2.82050759e-01 -7.10004449e-01 -1.21109283e+00 1.26882148e+00
2.04345211e-01 -6.13000631e-01 9.35000658e-01 9.80108440e-01
-3.21025372e-01 -9.58708346e-01 -8.38261604e-01 -5.58224618e-01
-5.91214597e-01 -2.09165871e-01 7.59547234e-01 7.86456704e-01
-4.07878995e-01 4.42696512e-01 -4.45461512e-01 4.98144329e-01
7.49080360e-01 3.29931416e-02 3.96967560e-01 -1.00523496e+00
-6.33142412e-01 -2.37836823e-01 2.56113112e-01 -1.54842770e+00
4.29448903e-01 -6.70055985e-01 -8.44983011e-03 -1.69619083e+00
1.50449589e-01 -4.80764061e-01 -4.13224757e-01 5.30999005e-01
3.83395492e-03 -2.94910342e-01 2.06729442e-01 2.34094188e-01
-1.27466714e+00 7.79995382e-01 1.02811658e+00 -2.29743905e-02
3.59929278e-02 1.39195874e-01 -6.51263297e-01 8.97081435e-01
7.60475755e-01 -4.02629882e-01 -7.79618680e-01 -1.13110840e+00
1.33054420e-01 3.41463983e-01 3.30442965e-01 -1.08516467e+00
8.05044770e-01 -6.71241879e-02 3.49156350e-01 -4.39282954e-01
5.82531691e-01 -1.19369936e+00 2.22539842e-01 3.56869638e-01
4.39040773e-02 1.47526801e-01 -1.86227977e-01 7.44974494e-01
1.07723251e-01 -9.36616585e-02 3.69007319e-01 1.43692508e-01
-7.18664408e-01 7.06296146e-01 -5.46729445e-01 -2.46809676e-01
1.15446365e+00 2.61593044e-01 -2.68875659e-01 -3.44732285e-01
-7.98368156e-01 5.67330658e-01 -2.90191844e-02 -5.46658337e-02
2.03745678e-01 -1.39477527e+00 -2.94102162e-01 -2.14314744e-01
-7.08924592e-01 3.53733152e-01 5.60950756e-01 9.65026617e-01
-3.54450315e-01 2.64327466e-01 -4.42426233e-03 -2.35088706e-01
-1.00309265e+00 4.88292605e-01 5.22701085e-01 -4.32401747e-01
-2.61535227e-01 6.88881397e-01 -1.30741775e-01 -3.80649269e-01
5.42532325e-01 -3.06806713e-01 5.45920469e-02 -8.41503888e-02
2.94278324e-01 9.29691255e-01 -2.70159412e-02 -1.17315715e-02
-2.48517215e-01 -3.39617860e-03 -2.54706070e-02 1.08634174e-01
1.32159793e+00 -4.25758123e-01 -5.27123064e-02 -1.09274415e-02
1.28200459e+00 -4.70748872e-01 -1.63137126e+00 -3.18819284e-01
-4.51745987e-01 3.17957491e-01 7.39614427e-01 -7.06217527e-01
-1.56152737e+00 5.95417142e-01 4.61722851e-01 7.18327835e-02
1.39673424e+00 -1.41698554e-01 1.33525050e+00 9.49103713e-01
2.41614997e-01 -1.58324468e+00 -2.81896979e-01 3.98998350e-01
1.00845881e-01 -1.08746195e+00 -2.90366113e-01 5.90536511e-03
-4.11715806e-01 8.99816871e-01 7.56817877e-01 1.40890241e-01
9.71680641e-01 4.38819140e-01 -4.26442921e-01 2.25245487e-02
-1.18576837e+00 7.39452392e-02 -4.29181874e-01 1.68013871e-01
-2.00496614e-01 3.15781772e-01 1.86326921e-01 9.89159882e-01
3.13310117e-01 3.88623238e-01 1.25395939e-01 1.05895674e+00
-1.74277321e-01 -8.42128813e-01 3.35332565e-02 6.65925026e-01
-5.33525586e-01 -1.83634888e-02 1.50240973e-01 4.07644868e-01
2.61551321e-01 8.13145518e-01 2.18238428e-01 -5.70260704e-01
-1.61673464e-02 2.20753495e-02 2.28888810e-01 -1.40587124e-03
-2.16484785e-01 7.63842389e-02 2.19907060e-01 -5.36436915e-01
-3.56128275e-01 -4.43239152e-01 -1.26422310e+00 -8.90157163e-01
-2.41026938e-01 -9.87186003e-03 8.69088471e-01 1.18179989e+00
8.23421180e-01 1.03339899e+00 6.26866758e-01 -9.73441184e-01
-6.74718499e-01 -8.41278672e-01 -5.33484280e-01 -3.66009533e-01
2.70870119e-01 -5.36597133e-01 -1.40057117e-01 1.56596452e-02]
|
[8.115602493286133, 2.8405816555023193]
|
c6b5d272-9dd8-47c0-a304-f50bdd467231
|
xinggan-for-person-image-generation
|
2007.09278
| null |
https://arxiv.org/abs/2007.09278v1
|
https://arxiv.org/pdf/2007.09278v1.pdf
|
XingGAN for Person Image Generation
|
We propose a novel Generative Adversarial Network (XingGAN or CrossingGAN) for person image generation tasks, i.e., translating the pose of a given person to a desired one. The proposed Xing generator consists of two generation branches that model the person's appearance and shape information, respectively. Moreover, we propose two novel blocks to effectively transfer and update the person's shape and appearance embeddings in a crossing way to mutually improve each other, which has not been considered by any other existing GAN-based image generation work. Extensive experiments on two challenging datasets, i.e., Market-1501 and DeepFashion, demonstrate that the proposed XingGAN advances the state-of-the-art performance both in terms of objective quantitative scores and subjective visual realness. The source code and trained models are available at https://github.com/Ha0Tang/XingGAN.
|
['Philip H. S. Torr', 'Li Zhang', 'Song Bai', 'Hao Tang', 'Nicu Sebe']
|
2020-07-17
| null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5192_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123700715.pdf
|
eccv-2020-8
|
['pose-transfer']
|
['computer-vision']
|
[-2.01650374e-02 7.13420063e-02 3.54696006e-01 -2.83923447e-01
-3.80653858e-01 -4.38323021e-01 8.22100699e-01 -5.72268546e-01
-1.59323037e-01 8.92740488e-01 6.28398880e-02 1.57032371e-01
4.81016248e-01 -1.03724432e+00 -7.35966742e-01 -7.65175760e-01
6.09057367e-01 3.82871956e-01 -1.19381376e-01 -2.16759488e-01
-2.71671563e-01 -5.10947593e-03 -1.10106277e+00 -2.43422300e-01
1.03736365e+00 7.88722456e-01 -1.35437161e-01 4.93613422e-01
2.99821347e-01 4.82170701e-01 -7.57061601e-01 -8.70358944e-01
4.28028882e-01 -8.22272360e-01 -4.78304923e-01 3.88258755e-01
4.72976893e-01 -5.73566735e-01 -6.32404208e-01 9.79351342e-01
7.83601522e-01 1.40024787e-02 5.96091270e-01 -1.43372059e+00
-1.16879284e+00 2.54854739e-01 -8.09098244e-01 -1.45330265e-01
4.31161702e-01 6.78458750e-01 3.98439497e-01 -7.38211215e-01
4.91589755e-01 1.44658422e+00 4.53038305e-01 9.53546762e-01
-1.00018787e+00 -9.60774124e-01 8.59449655e-02 -6.20787218e-02
-1.32449341e+00 -3.15812528e-01 8.79296064e-01 -4.81567472e-01
-5.41049056e-03 1.66732177e-01 8.18332016e-01 1.45614731e+00
-1.13222390e-01 8.23651910e-01 1.19337022e+00 -2.44963676e-01
-1.06553338e-01 1.27490968e-01 -4.24040228e-01 7.55691469e-01
3.83144289e-01 3.05352747e-01 -2.43506461e-01 2.19485879e-01
1.14011991e+00 3.15175354e-02 -3.17344695e-01 -2.51392752e-01
-1.22160280e+00 7.11093724e-01 6.26283228e-01 -4.72409911e-02
-3.83190185e-01 2.77011782e-01 7.71276131e-02 1.98009666e-02
4.87876326e-01 1.87718496e-01 1.01142801e-01 1.93930596e-01
-5.00554681e-01 5.54312587e-01 2.36633226e-01 1.14029562e+00
6.66543365e-01 2.36071631e-01 -6.65295541e-01 7.04881668e-01
2.58632004e-01 8.48899305e-01 4.27599937e-01 -6.00161076e-01
4.23186570e-01 7.08047330e-01 4.30281460e-01 -1.01840389e+00
3.58109362e-02 -5.44388235e-01 -1.11522806e+00 1.59374058e-01
4.09613192e-01 -6.04376853e-01 -1.19934857e+00 1.94801068e+00
5.16613781e-01 1.94590658e-01 -5.19109629e-02 1.06391549e+00
1.17621315e+00 7.50075638e-01 7.61820525e-02 2.61657387e-01
1.41258478e+00 -1.31420755e+00 -6.61872268e-01 -5.04189074e-01
-6.93442971e-02 -8.22167218e-01 1.10491145e+00 1.76532328e-01
-1.33487117e+00 -1.04560244e+00 -9.39604342e-01 -2.22163200e-02
-1.59937069e-01 6.39441729e-01 5.97800314e-01 6.33132041e-01
-8.95238757e-01 2.31093630e-01 -6.49995923e-01 -2.48824999e-01
6.29212379e-01 8.35830122e-02 -2.85015434e-01 -6.70374483e-02
-1.16553736e+00 4.48804259e-01 3.44988465e-01 2.69149363e-01
-1.03556490e+00 -4.75671798e-01 -8.56507897e-01 -1.96512759e-01
3.05259377e-01 -1.20816720e+00 1.16953409e+00 -1.08102560e+00
-1.47269452e+00 9.49170589e-01 8.72257631e-03 -1.53278351e-01
1.12847054e+00 -3.55453134e-01 -3.83072674e-01 4.27431799e-02
1.24531753e-01 9.87597823e-01 1.05164540e+00 -1.43334877e+00
-2.94607073e-01 -2.79039234e-01 -1.04565471e-02 2.36875609e-01
-3.49838853e-01 -1.62311047e-01 -8.61746967e-01 -1.07013535e+00
-4.32349026e-01 -1.04790759e+00 -1.92476109e-01 9.70446542e-02
-8.48505020e-01 -8.14856067e-02 6.74203277e-01 -9.57314253e-01
8.25828493e-01 -1.89503193e+00 1.86843798e-01 7.71318749e-02
3.29717785e-01 5.49385726e-01 -3.01009238e-01 3.65121305e-01
-1.93849429e-01 1.91441745e-01 -1.52335659e-01 -7.70237625e-01
6.56184480e-02 -1.77530915e-01 -1.85349565e-02 3.31559032e-01
2.40243450e-01 1.39752972e+00 -8.43443871e-01 -2.78993398e-01
3.86676580e-01 7.31738150e-01 -3.26705813e-01 6.17157876e-01
-2.11363044e-02 8.92443538e-01 -3.11206460e-01 4.57468003e-01
9.79038119e-01 -1.38648614e-01 -1.42326653e-01 -1.38263106e-01
2.91796178e-01 -1.62258565e-01 -9.39366221e-01 1.58219326e+00
-2.56739974e-01 2.26278439e-01 -3.77240121e-01 -4.10700560e-01
1.07765889e+00 3.59083712e-01 7.18808025e-02 -6.07327521e-01
3.31223577e-01 -1.17179409e-01 -1.24862812e-01 -1.21495225e-01
2.84577161e-01 -6.08373545e-02 -1.05977051e-01 2.40359351e-01
-1.36259906e-02 1.32805318e-01 2.67858684e-01 1.24170862e-01
5.88410437e-01 2.30811670e-01 8.18989575e-02 -6.85141906e-02
7.42208838e-01 -4.46328819e-01 6.11810207e-01 5.01662135e-01
-5.93684055e-02 1.03150630e+00 3.19950789e-01 -3.40758801e-01
-1.18769789e+00 -1.32343757e+00 4.99737471e-01 4.27968204e-01
3.50888669e-01 -2.93653429e-01 -1.10817635e+00 -7.44383991e-01
-1.31449789e-01 4.54766214e-01 -8.90723467e-01 -2.67996222e-01
-4.69334006e-01 -5.92263877e-01 4.45386827e-01 6.11455083e-01
1.04026973e+00 -1.30044663e+00 -2.48506427e-01 -8.78187492e-02
-3.41026962e-01 -1.14643383e+00 -1.01832592e+00 -9.37442660e-01
-3.05550009e-01 -9.13591504e-01 -1.24201381e+00 -8.37821186e-01
1.13949633e+00 4.33685109e-02 1.02399492e+00 2.08502874e-01
-2.15844572e-01 2.20866472e-01 -4.46506828e-01 -4.61813122e-01
-2.47624934e-01 3.16035859e-02 -9.89199281e-02 3.96661788e-01
1.12267688e-01 -5.05532086e-01 -1.31473446e+00 5.04764855e-01
-8.56736898e-01 5.24980545e-01 7.05790639e-01 8.08668017e-01
5.12228370e-01 1.21769011e-01 5.69024682e-01 -8.78533363e-01
6.24152839e-01 -2.03004315e-01 -4.12808836e-01 3.87238786e-02
-4.58752751e-01 -3.19378376e-01 7.00680315e-01 -4.62558031e-01
-1.28284180e+00 -9.93352756e-02 -2.74281979e-01 -4.96298045e-01
-3.44421059e-01 -2.13096160e-02 -4.39830899e-01 3.20751704e-02
3.28814387e-01 4.87227201e-01 9.39292088e-02 -3.82410020e-01
4.11078811e-01 4.05729204e-01 8.47811818e-01 -5.75261891e-01
1.45647132e+00 5.02652586e-01 -3.57844710e-01 -2.98574299e-01
-6.03790700e-01 -8.27583298e-02 -3.63334626e-01 -2.05204859e-01
1.06862187e+00 -1.12293327e+00 -6.89540684e-01 1.08980846e+00
-1.18788373e+00 -3.11772048e-01 -3.26494992e-01 6.98042437e-02
-4.41089422e-01 1.89861774e-01 -6.16756976e-01 -5.43598592e-01
-7.01681376e-01 -9.67477024e-01 1.12745988e+00 7.68923640e-01
4.18859022e-03 -8.88467252e-01 8.50745589e-02 6.97724700e-01
3.23501259e-01 7.47962058e-01 3.06133479e-01 -1.24633327e-01
-5.91755092e-01 -2.79309064e-01 -3.00822347e-01 4.95134085e-01
2.96587706e-01 -3.15211833e-01 -7.14879155e-01 -7.26615727e-01
-3.65960866e-01 -2.77587235e-01 7.01475739e-01 1.50660947e-01
1.02810740e+00 -4.84679520e-01 -2.51664966e-01 7.23564088e-01
1.34396100e+00 3.30134809e-01 1.03486669e+00 1.66397411e-02
9.22424316e-01 3.26169103e-01 4.52775240e-01 4.47163522e-01
2.97288239e-01 7.75674284e-01 2.51872182e-01 -6.05750144e-01
-5.37765384e-01 -8.06980789e-01 2.14644134e-01 4.50345397e-01
-4.22401994e-01 -4.84377414e-01 -6.20836735e-01 4.86334920e-01
-1.82273722e+00 -9.03173804e-01 6.06794208e-02 1.99561834e+00
7.23542809e-01 1.89463422e-02 3.97581756e-01 -2.44982108e-01
9.44854319e-01 1.71394482e-01 -6.21834219e-01 1.51104882e-01
2.12980490e-02 2.53264934e-01 3.34466755e-01 1.91384822e-01
-1.08481574e+00 1.00647688e+00 4.93367577e+00 7.64614224e-01
-9.27927732e-01 1.84931144e-01 1.04359102e+00 -3.23953666e-02
-2.48983905e-01 -2.37238586e-01 -7.33768821e-01 8.40649426e-01
3.19447696e-01 -2.69762158e-01 4.72792268e-01 5.31924963e-01
2.14359328e-01 3.31685275e-01 -7.37884223e-01 1.17263186e+00
2.61014730e-01 -1.09164929e+00 2.75653869e-01 1.27941117e-01
9.57034647e-01 -6.99514687e-01 3.76559734e-01 3.55048656e-01
3.84978920e-01 -1.13181329e+00 7.53528655e-01 7.16736555e-01
9.74239171e-01 -9.60324228e-01 8.69319499e-01 1.16120530e-02
-1.18857765e+00 3.48086029e-01 -3.15859281e-02 1.92867756e-01
4.06430691e-01 3.51329893e-01 -4.67844397e-01 7.68348515e-01
6.45195723e-01 5.73366880e-01 -8.52810681e-01 8.88569534e-01
-7.50051260e-01 5.79284847e-01 9.84925702e-02 3.62489164e-01
4.81309295e-02 -3.83932501e-01 6.13539040e-01 8.25243652e-01
4.00106817e-01 -1.35623356e-02 1.97334051e-01 1.26830447e+00
-2.51095027e-01 -7.78239220e-02 -3.26935172e-01 1.02476865e-01
4.20716792e-01 1.35529494e+00 -3.49360019e-01 -2.67576426e-01
-1.91116408e-01 1.39312267e+00 1.35536157e-02 4.21765953e-01
-1.20027745e+00 -5.16820490e-01 5.66576779e-01 4.69345659e-01
3.06699842e-01 8.14870968e-02 -1.14297662e-02 -1.28106332e+00
3.49176377e-01 -1.04396904e+00 1.55080572e-01 -9.63741601e-01
-1.41341615e+00 8.83507550e-01 -8.60334188e-02 -1.32367432e+00
-2.88365275e-01 -3.09570849e-01 -9.68087792e-01 1.13143241e+00
-1.26768267e+00 -1.77878964e+00 -8.40031505e-01 6.73471570e-01
2.66309112e-01 -3.65787387e-01 4.83372658e-01 3.35519612e-01
-7.94470668e-01 1.02276230e+00 -3.31120610e-01 5.91847658e-01
7.74371564e-01 -1.18555069e+00 8.81654918e-01 1.22199500e+00
-2.36342505e-01 5.90780377e-01 5.76356649e-01 -6.93783641e-01
-1.00583756e+00 -1.44417977e+00 4.03820306e-01 -2.55493462e-01
9.62498114e-02 -5.60596108e-01 -4.82464910e-01 7.74918258e-01
6.58204794e-01 -1.31474324e-02 3.33037019e-01 -2.88387567e-01
-1.72634646e-01 -1.46034941e-01 -1.26386845e+00 8.20242941e-01
1.17246103e+00 -3.96997295e-02 -1.46375120e-01 2.35258281e-01
6.06086373e-01 -6.39885366e-01 -6.52045429e-01 2.74806023e-01
4.68119502e-01 -1.05534148e+00 1.14840508e+00 -3.65570992e-01
6.80897117e-01 -5.90757549e-01 3.49300176e-01 -1.57952082e+00
-4.31690812e-01 -8.89305174e-01 -2.91068479e-02 1.57231045e+00
-6.13580495e-02 -8.93293381e-01 8.63930643e-01 3.64682853e-01
1.09784879e-01 -7.72104859e-01 -4.37975883e-01 -8.51284802e-01
6.24184757e-02 2.19469085e-01 1.02126122e+00 6.52523875e-01
-6.95271075e-01 4.05321658e-01 -9.07083452e-01 8.12399462e-02
8.50336015e-01 5.31710982e-02 1.28742409e+00 -7.38891184e-01
-4.41605449e-01 -2.40524739e-01 -5.91874361e-01 -1.06940484e+00
-1.21058427e-01 -5.89953303e-01 -1.15065336e-01 -1.68485332e+00
3.86541575e-01 -3.50403905e-01 -4.71949130e-02 5.97390413e-01
-5.73270261e-01 6.30345166e-01 4.48736012e-01 -1.71287768e-02
-2.23797172e-01 9.42988515e-01 1.73624778e+00 -3.26007187e-01
2.38834489e-02 2.93777920e-02 -1.10830796e+00 5.17334461e-01
9.76792514e-01 -1.22686073e-01 -5.75509250e-01 -4.51918870e-01
-1.46565422e-01 -1.41374856e-01 8.25398088e-01 -1.29449773e+00
-1.06974512e-01 -2.05746219e-02 8.30842674e-01 -3.86425138e-01
4.57269847e-01 -3.89036804e-01 4.36498374e-01 4.99470413e-01
-1.34390563e-01 1.59623221e-01 1.65052652e-01 5.23641467e-01
-1.17144413e-01 2.87896395e-01 9.70501482e-01 -1.29952356e-01
-5.05816102e-01 7.13576496e-01 3.05793405e-01 1.19774438e-01
1.24980569e+00 8.66866950e-03 -5.63704550e-01 -8.15857589e-01
-5.28502882e-01 3.25978786e-01 5.59984446e-01 6.06146097e-01
7.31160223e-01 -1.80987978e+00 -1.16065395e+00 2.36584887e-01
1.36604384e-01 7.67073259e-02 5.81331253e-01 5.40947616e-01
-4.03100342e-01 9.16674361e-02 -4.99248236e-01 -2.24462271e-01
-1.18904245e+00 5.62167943e-01 4.68597502e-01 -4.23299372e-01
-5.56491077e-01 8.32713068e-01 7.43703663e-01 -4.64602888e-01
-2.01155812e-01 1.36782572e-01 -8.87273462e-04 -4.52110678e-01
5.47734976e-01 1.71955809e-01 -3.57312888e-01 -7.72003472e-01
-2.29162663e-01 4.87233162e-01 -2.20643178e-01 1.60728414e-02
1.01424420e+00 -9.20174047e-02 1.24461807e-01 -1.16328448e-01
8.91506076e-01 6.11978993e-02 -1.48014891e+00 -1.03245124e-01
-9.55736339e-01 -7.68039167e-01 -3.90102655e-01 -1.01981282e+00
-1.59729588e+00 6.54393077e-01 8.27474833e-01 -2.46508747e-01
1.23948431e+00 -6.66348934e-02 1.21786284e+00 -3.31860662e-01
2.67294288e-01 -6.12580121e-01 3.35374802e-01 -3.05565782e-02
1.15228009e+00 -1.14549839e+00 -9.48127955e-02 -5.68489611e-01
-6.82526290e-01 5.63247442e-01 1.07814169e+00 -2.34425351e-01
2.59360254e-01 -2.49626726e-01 1.87069401e-01 2.41048615e-02
-3.39268088e-01 -2.60818332e-01 5.10195851e-01 7.80783057e-01
2.47236356e-01 2.91863173e-01 -2.83393234e-01 7.77460039e-01
-4.94928449e-01 5.59571013e-02 3.02237749e-01 5.07239163e-01
2.51241148e-01 -1.25220084e+00 -3.92532349e-01 3.18756253e-01
-2.12136865e-01 2.81918682e-02 -4.34577465e-01 7.58107781e-01
5.69579363e-01 8.93768132e-01 1.78703514e-03 -6.38674498e-01
3.98202598e-01 -4.27375317e-01 4.88328725e-01 -4.60723788e-01
-4.22536343e-01 2.08479725e-03 -5.95108047e-02 -2.41766304e-01
-1.91905335e-01 -5.38616002e-01 -7.72919834e-01 -5.10527909e-01
6.13881983e-02 4.47503999e-02 1.20462239e-01 6.22898400e-01
4.17881668e-01 6.93126321e-01 6.37828648e-01 -7.11015582e-01
-2.50838757e-01 -1.10705054e+00 -3.35662335e-01 8.07380140e-01
-2.45397463e-02 -6.20166183e-01 -6.13604300e-02 2.29647219e-01]
|
[12.013676643371582, -0.8299381732940674]
|
5e1dc150-f35e-4c52-b8c7-e17618d50978
|
efficient-tms-based-motor-cortex-mapping
| null | null |
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9515996
|
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9515996
|
Efficient TMS-Based Motor Cortex Mapping Using Gaussian Process Active Learning
|
Transcranial Magnetic Stimulation (TMS) can be used to map cortical motor topography by spatially sampling the sensorimotor cortex while recording Motor Evoked Potentials (MEP) with surface electromyography (EMG). Traditional sampling strategies are time-consuming and inefficient, as they ignore the fact that responsive sites are typically sparse and highly spatially correlated. An alternative approach, commonly employed when TMS mapping is used for presurgical planning, is to leverage the expertise of the coil operator to use MEPs elicited by previous stimuli as feedback to decide which loci to stimulate next. In this paper, we propose to automatically infer optimal future stimulus loci using active learning Gaussian Process-based sampling in place of user expertise. We first compare the user-guided (USRG) method to the traditional grid selection method and randomized sampling to verify that the USRG approach has superior performance. We then compare several novel active Gaussian Process (GP) strategies with the USRG approach. Experimental results using real data show that, as expected, the USRG method is superior to the grid and random approach in both time efficiency and MEP map accuracy. We also found that an active warped GP entropy and a GP random-based strategy performed equally as well as, or even better than, the USRG method. These methods were completely automatic, and succeeded in efficiently sampling the regions in which the MEP response variations are largely confined. This work provides the foundation for highly efficient, fully automatized TMS mapping, especially when considered in the context of advances in robotic coil operation.
|
['D Erdoğmuş', 'D Brooks', 'E Tunik', 'T Imbiriba', 'M Yarossi', 'R Faghihpirayesh']
|
2021-08-18
| null | null | null |
ieee-transactions-on-neural-systems-and-1
|
['electromyography-emg']
|
['medical']
|
[ 4.21824098e-01 1.53493345e-01 -2.37794593e-01 1.85101017e-01
-1.31584597e+00 -4.80769187e-01 4.19082224e-01 -1.75980493e-01
-7.35698342e-01 1.02726114e+00 3.59503448e-01 -1.19830489e-01
-4.20525879e-01 -3.52264196e-01 -5.38780212e-01 -1.03512549e+00
-2.23855883e-01 7.52567947e-01 3.08649689e-01 1.91845357e-01
8.33154023e-01 5.54064035e-01 -1.56419683e+00 -1.22961458e-02
1.30683362e+00 7.66727805e-01 1.03626788e+00 4.26723845e-02
8.16698596e-02 4.13596064e-01 -3.01832706e-01 4.27137315e-01
1.66854277e-01 -3.07324082e-01 -8.54247034e-01 -4.75738585e-01
-2.35761434e-01 4.57290299e-02 8.14570785e-02 9.46690023e-01
8.51981044e-01 3.54004145e-01 8.19732964e-01 -5.19765794e-01
8.27284809e-03 5.94041288e-01 -4.31944996e-01 1.58357531e-01
3.82012516e-01 3.31006736e-01 2.65597880e-01 -9.38941777e-01
9.91854131e-01 6.94235981e-01 4.42503363e-01 4.72836286e-01
-1.55772161e+00 -6.39917731e-01 -5.92678413e-02 4.74152178e-01
-1.32864022e+00 -4.75167692e-01 6.60663903e-01 -6.95508361e-01
9.57645655e-01 2.04143241e-01 7.61033356e-01 1.23298717e+00
5.54036021e-01 6.70011282e-01 1.54831755e+00 -5.83174348e-01
6.84218884e-01 4.75814342e-02 -3.34159851e-01 1.13152914e-01
-2.09847555e-01 5.91742635e-01 -6.43316746e-01 -4.29050326e-01
9.78277564e-01 -3.32889467e-01 -6.55097663e-01 -6.71495020e-01
-1.42499745e+00 8.23511720e-01 1.67128935e-01 6.05604589e-01
-1.15991914e+00 -5.14059886e-02 2.32201576e-01 -1.41897142e-01
3.67052823e-01 9.89461303e-01 -6.75512776e-02 -5.56909680e-01
-1.43500495e+00 9.37565044e-02 6.35572970e-01 4.65124011e-01
5.04250050e-01 -3.39550704e-01 -4.09262747e-01 8.80062461e-01
-2.08632722e-01 3.03944260e-01 6.44664526e-01 -1.09705663e+00
1.30379677e-01 1.99451596e-01 2.13510752e-01 -6.95903838e-01
-3.91485751e-01 -1.66060701e-01 -4.14014876e-01 7.64216065e-01
9.88043472e-02 -2.83892244e-01 -1.00645518e+00 1.37927270e+00
-1.06473707e-01 -1.91109568e-01 -5.73265791e-01 9.63197529e-01
-3.28030586e-02 1.45524591e-01 2.12606415e-01 -4.86808032e-01
8.64622831e-01 -3.36900562e-01 -6.92203522e-01 -2.24959597e-01
4.41278785e-01 -6.86256051e-01 1.06399751e+00 7.27210343e-01
-1.07235301e+00 -1.75788313e-01 -8.28834355e-01 6.48829103e-01
-1.94068123e-02 6.95338622e-02 5.99886954e-01 2.19344079e-01
-1.20641041e+00 1.02674675e+00 -1.28949463e+00 -3.00216794e-01
5.37499309e-01 5.65017879e-01 -4.07319844e-01 3.76546323e-01
-1.13811731e+00 1.60652673e+00 3.37729007e-01 -4.68684211e-02
-6.55809641e-01 -6.54268563e-01 -4.27628964e-01 -1.64905384e-01
2.23041058e-01 -2.60532528e-01 1.05047107e+00 -6.17751598e-01
-1.90312707e+00 6.88103974e-01 -2.91196048e-01 -4.84036177e-01
3.73969764e-01 -9.15628113e-03 1.70864210e-01 4.02932793e-01
1.06703855e-01 7.58969545e-01 6.76674724e-01 -9.82638299e-01
-2.20305443e-01 -4.31988239e-01 -5.48570335e-01 3.72649848e-01
1.55139998e-01 6.02875650e-02 1.79308504e-01 -4.35611606e-01
2.10518360e-01 -9.97399926e-01 -5.88153601e-01 -2.95022219e-01
-1.41509771e-01 -1.77632168e-01 1.99349120e-01 -8.69845569e-01
1.02830493e+00 -1.86977160e+00 4.53280509e-01 5.81472635e-01
-8.26844573e-02 7.83775747e-02 2.04108566e-01 6.15432560e-01
-1.65054545e-01 -2.77579904e-01 -4.46346194e-01 1.91458866e-01
-2.05752075e-01 -1.37290150e-01 -2.26249874e-01 5.74811161e-01
-7.83286467e-02 8.79408062e-01 -9.89927053e-01 -5.57727337e-01
4.35248166e-01 4.34329569e-01 -3.22616428e-01 -7.15907812e-02
2.14129746e-01 8.55439663e-01 -1.03860363e-01 4.44566727e-01
2.21692756e-01 1.48291677e-01 1.29405677e-01 -2.91303962e-01
-3.72121483e-01 3.35105926e-01 -8.46062183e-01 2.00039244e+00
-5.69979012e-01 6.55752301e-01 2.28381425e-01 -1.09255838e+00
9.53672707e-01 5.89502931e-01 8.73444498e-01 -7.55749524e-01
1.08870298e-01 6.96198404e-01 1.45212159e-01 -6.26929462e-01
1.89113110e-01 -3.96184266e-01 2.49304265e-01 3.85020643e-01
2.94501543e-01 -2.69650578e-01 1.31919356e-02 -1.24699272e-01
1.09172308e+00 5.89614511e-01 4.87206101e-01 -7.98052609e-01
1.85448155e-02 2.76548684e-01 2.41670266e-01 6.65833950e-01
-6.84949011e-02 6.00321174e-01 3.28897744e-01 4.21680182e-01
-6.27429903e-01 -1.19527185e+00 -3.72540295e-01 4.13488865e-01
1.19873613e-01 9.72387940e-02 -6.95087552e-01 -3.48466516e-01
-1.57954887e-01 1.02820957e+00 -6.53878391e-01 -7.07004741e-02
-6.88996851e-01 -4.29986447e-01 -6.33472428e-02 5.32002807e-01
-1.52695179e-01 -1.47025931e+00 -8.55911374e-01 5.37016094e-01
-3.52036983e-01 -3.84971708e-01 -1.62859052e-01 5.85934877e-01
-1.04176307e+00 -7.57473052e-01 -1.25255620e+00 -5.59671879e-01
5.11735022e-01 -4.06685084e-01 5.35703778e-01 -7.81680703e-01
-3.19707394e-01 3.01981479e-01 -4.52376604e-01 -3.86961013e-01
-1.17324650e-01 7.13298768e-02 7.07342476e-02 -4.76734728e-01
3.81782800e-01 -1.02458954e+00 -8.13275039e-01 2.02215642e-01
-4.13060248e-01 -4.56719063e-02 1.05057657e+00 1.00493085e+00
7.47327209e-01 -4.02729005e-01 7.17060268e-01 -5.45529783e-01
1.05333507e+00 -4.11363423e-01 -2.46698797e-01 2.14553550e-02
-6.75900519e-01 6.50335252e-02 2.54755944e-01 -7.77207136e-01
-8.78900588e-01 2.36947820e-01 2.85099894e-02 -4.29908901e-01
-2.23738670e-01 5.52257538e-01 2.92977363e-01 -3.95925254e-01
1.13284159e+00 3.67382616e-01 5.54288924e-01 -2.29334250e-01
-4.80493754e-02 7.54736304e-01 4.40224439e-01 -3.45982283e-01
1.97251946e-01 4.39868361e-01 -2.42677703e-01 -4.99765217e-01
1.48209319e-01 -4.45172101e-01 -6.93055570e-01 -4.95554954e-01
6.66608036e-01 -3.15558821e-01 -5.62279761e-01 2.06029356e-01
-1.02540898e+00 -6.17896199e-01 -2.28802964e-01 1.18247616e+00
-1.40607584e+00 2.82501340e-01 -3.26611400e-01 -1.00980401e+00
-5.92950702e-01 -1.46218467e+00 9.71155226e-01 -1.26995683e-01
-9.58709717e-01 -9.06471908e-01 1.65530115e-01 -9.59715396e-02
7.59591699e-01 1.96399570e-01 7.16388106e-01 -4.97461379e-01
-2.99984992e-01 -4.03862417e-01 3.43361974e-01 3.96848731e-02
2.64364272e-01 -5.00138164e-01 -7.01978564e-01 -2.39651918e-01
3.65871698e-01 -6.60963506e-02 4.91803616e-01 8.51383865e-01
1.07044566e+00 2.33520836e-01 -7.35403419e-01 2.85569280e-01
1.48910785e+00 5.27116001e-01 1.05243766e+00 2.73476183e-01
-9.79865342e-02 7.58853078e-01 7.87745893e-01 -8.24008584e-02
-5.22139549e-01 8.86233628e-01 3.24529633e-02 -7.00012036e-03
1.34394169e-01 -2.83722281e-02 2.80773997e-01 5.34305513e-01
-5.56544244e-01 4.77277339e-01 -1.04432642e+00 5.46986639e-01
-1.65184081e+00 -1.11003458e+00 3.51659268e-01 2.54585934e+00
9.61956501e-01 1.67632252e-01 -1.97364260e-02 1.30679309e-01
7.41605818e-01 -4.76081401e-01 -5.14183164e-01 -2.99383610e-01
1.84310377e-01 1.05136752e+00 7.79780388e-01 4.55221504e-01
-6.60763443e-01 5.50824285e-01 6.49583483e+00 1.26696754e+00
-1.36455393e+00 3.66151005e-01 -7.06633404e-02 -3.72419238e-01
6.53674304e-02 -6.80139987e-03 -3.46992731e-01 7.20012009e-01
6.49838805e-01 -3.56703848e-01 5.36642432e-01 6.29154623e-01
6.58353686e-01 -8.40533137e-01 -9.04805481e-01 9.34111118e-01
-2.02360898e-01 -1.47065306e+00 -5.28324127e-01 1.97589800e-01
4.07168865e-01 2.33911172e-01 -2.32399106e-01 2.28479989e-02
-1.00303307e-01 -1.05070543e+00 5.95486045e-01 9.57982361e-01
1.00183606e+00 -4.33421999e-01 5.65961003e-01 6.24206722e-01
-7.83763170e-01 4.36419472e-02 -6.07566945e-02 1.78125754e-01
5.42001665e-01 3.35309684e-01 -9.89050448e-01 3.71107399e-01
5.53178072e-01 1.49834946e-01 5.54009750e-02 1.45287836e+00
-2.15372071e-01 5.11544406e-01 -4.47345048e-01 -2.81274974e-01
9.61931702e-03 -1.49641767e-01 1.04141402e+00 8.65668118e-01
5.23219764e-01 -2.10044637e-01 -3.39480191e-01 1.23563290e+00
9.63068485e-01 2.05917567e-01 -5.07941544e-01 5.28477989e-02
6.25947356e-01 1.04497230e+00 -8.65240037e-01 1.11251481e-01
2.24965334e-01 9.09919798e-01 2.84504503e-01 4.82475340e-01
-3.52077246e-01 -5.96161485e-01 2.26983223e-02 2.83128619e-01
1.39663830e-01 -2.38412991e-01 -5.96862257e-01 -6.48748159e-01
9.85587388e-02 -6.29937351e-01 -1.06424809e-01 -8.64057958e-01
-1.07651222e+00 5.72188914e-01 3.42894346e-01 -1.44577348e+00
-8.57996643e-01 -6.07054591e-01 -7.26766348e-01 1.53185654e+00
-6.84029460e-01 -3.56100768e-01 1.48545519e-01 4.54811901e-01
4.52158749e-01 4.69603725e-02 1.11734307e+00 -5.26308641e-02
5.76663986e-02 1.13853186e-01 -5.07741328e-03 -3.68476272e-01
5.64184964e-01 -1.12811732e+00 -1.96516961e-01 6.18030071e-01
-2.65597582e-01 8.96415651e-01 8.55914772e-01 -8.69153738e-01
-1.02455258e+00 -2.43084535e-01 4.09678429e-01 9.99730080e-02
6.06906414e-01 -2.55611151e-01 -7.33018339e-01 1.72495142e-01
8.09842348e-02 -6.91492081e-01 4.42059815e-01 1.17423482e-01
6.91510320e-01 2.52737343e-01 -1.15877306e+00 7.28326023e-01
8.28475654e-01 -5.16605079e-01 -9.78578806e-01 2.73493141e-01
-1.08107030e-01 -2.05079913e-01 -9.41286206e-01 7.79015601e-01
5.81529379e-01 -7.08115757e-01 6.01471484e-01 4.64419201e-02
1.14642330e-01 2.08348513e-01 2.49329075e-01 -1.80183113e+00
-4.70582217e-01 -7.05994964e-01 1.73599780e-01 5.46744108e-01
5.66119671e-01 -9.80310500e-01 9.65977132e-01 4.42983717e-01
-3.92155051e-01 -1.27754581e+00 -1.19305491e+00 -7.41865158e-01
2.86907881e-01 -1.87538639e-01 2.79916488e-02 5.07306635e-01
8.89264226e-01 -1.89597204e-01 1.00437485e-01 -4.68715847e-01
6.73535824e-01 3.51979099e-02 1.04014806e-01 -1.04765379e+00
-3.17153275e-01 -6.99857116e-01 -5.62175035e-01 -7.04010010e-01
-1.62550826e-02 -1.03670681e+00 5.86458921e-01 -1.84655797e+00
-5.52257970e-02 -4.55423534e-01 -2.03816742e-01 1.45697936e-01
-1.79931998e-01 -8.07789415e-02 -2.48190969e-01 4.56740469e-01
4.01163429e-01 3.65827501e-01 1.33318138e+00 3.65437418e-01
-4.89167631e-01 2.02555031e-01 -2.92031944e-01 5.23009181e-01
9.73131478e-01 -4.15445298e-01 -3.82074445e-01 1.60939336e-01
-1.56965524e-01 3.32867742e-01 4.18376237e-01 -1.06571925e+00
6.40399933e-01 -9.61449444e-02 3.38748664e-01 -7.60839522e-01
5.13308465e-01 -5.33564389e-01 4.02353168e-01 5.06031394e-01
-3.15246671e-01 -4.17343944e-01 1.50060609e-01 3.11756283e-01
-2.49696285e-01 -2.80072153e-01 7.45911896e-01 -2.72865653e-01
-6.74782693e-01 7.33348876e-02 -8.35626245e-01 -2.90744454e-01
1.03076005e+00 -8.58152866e-01 1.90432698e-01 -2.27796838e-01
-1.22875214e+00 -5.46354204e-02 4.31672961e-01 7.05693513e-02
6.20480835e-01 -1.07723093e+00 -5.12529254e-01 1.51150554e-01
-1.12139799e-01 -4.00547266e-01 3.59185129e-01 1.87298787e+00
-3.24190795e-01 6.47911549e-01 -4.54663068e-01 -8.74580920e-01
-8.76320958e-01 3.00243765e-01 4.67113256e-01 -9.04327184e-02
-7.10729837e-01 7.98583865e-01 -1.27203554e-01 -2.31525794e-01
2.05994368e-01 7.51498193e-02 -2.09854543e-01 8.61181915e-02
1.77535072e-01 4.08732325e-01 4.04051989e-01 -1.94376811e-01
-2.30575532e-01 3.61058176e-01 3.10869545e-01 -6.12160563e-01
1.18056202e+00 5.43856211e-02 -2.33954079e-02 6.15244508e-01
8.35072696e-01 -7.81035870e-02 -9.56379890e-01 1.04022384e-01
1.47711918e-01 -2.95927137e-01 3.19038391e-01 -1.01240134e+00
-6.80850506e-01 7.75894046e-01 7.53375351e-01 -1.41031027e-01
1.11731005e+00 6.95499182e-02 2.88367510e-01 -1.50275514e-01
1.22675288e+00 -1.35160279e+00 -5.44367373e-01 -6.71710894e-02
1.18899167e+00 -5.73161125e-01 -1.09375931e-01 -4.07636046e-01
-7.47352958e-01 9.88136768e-01 2.36975208e-01 -3.52785259e-01
6.89915121e-01 5.35863578e-01 -2.11810805e-02 -2.02790350e-01
-4.72979814e-01 -2.62465149e-01 2.45213464e-01 7.23059535e-01
3.84342700e-01 2.91560024e-01 -9.00040507e-01 4.10267174e-01
-3.27744782e-01 6.51493907e-01 5.46723558e-03 1.07656550e+00
-4.98546183e-01 -1.02817273e+00 -4.13346499e-01 9.45876777e-01
-1.59067050e-01 -1.69406667e-01 -6.73648566e-02 7.92948961e-01
-4.49718572e-02 7.27982283e-01 3.01986188e-02 -4.88215476e-01
1.73365027e-01 2.10928053e-01 9.95587766e-01 -7.10607886e-01
-5.08523047e-01 2.30168179e-01 5.03235385e-02 -8.91582370e-01
-3.34655166e-01 -9.41871881e-01 -1.23833585e+00 3.73999894e-01
-4.65993822e-01 3.02239656e-01 9.24543560e-01 9.11949098e-01
2.65884042e-01 1.44231662e-01 3.22346479e-01 -1.41994286e+00
-7.95375109e-01 -1.36736274e+00 -8.56937826e-01 7.06307963e-02
-2.52615631e-01 -1.21309435e+00 -7.50698864e-01 -3.93887252e-01]
|
[12.880169868469238, 3.367546319961548]
|
82e65b6f-dcb3-47e4-9069-310834fe8a14
|
towards-a-real-time-demand-response-framework
|
2303.00186
| null |
https://arxiv.org/abs/2303.00186v2
|
https://arxiv.org/pdf/2303.00186v2.pdf
|
Targeted demand response for flexible energy communities using clustering techniques
|
The present study explores the use of clustering techniques for the design and implementation of a demand response (DR) program for commercial and residential prosumers. The goal of the program is to alter the consumption behavior of the prosumers pertaining to a distributed energy community in Italy. This aggregation aims to: a) minimize the reverse power flow at the primary substation, that occurs when generation from solar panels in the local grid exceeds consumption, and b) shave the system wide peak demand, that typically occurs during the hours of late afternoon. Regarding the clustering stage, three popular machine learning algorithms for electrical load clustering are employed -namely k-means, k-medoids and an agglomerative hierarchical clustering- alongside two different distance measures -namely euclidean and constrained dynamic time warping (DTW). We evaluate the methods using multiple validation metrics including a novel metric -namely peak performance score (PPS)- that we propose in the context of this study. The best model is employed to divide daily prosumer load profiles into clusters and each cluster is analyzed in terms of load shape, mean entropy, and load type distribution. These characteristics are then used to distinguish the clusters that have the potential to serve the optimization objectives by matching them to appropriate DR schemes including time of use (TOU), critical peak pricing (CPP), and real-time pricing (RTP). The results of this study can be useful for network operators, utilities, and aggregators that aim to develop targeted DR programs for groups of prosumers within flexible energy communities.
|
['Dimitris Askounis', 'Mohammad Ghoreishi', 'Francesca Santori', 'Spiros Mouzakitis', 'Evangelos Karakolis', 'Angelos Pipergias', 'Sotiris Pelekis']
|
2023-03-01
| null | null | null | null |
['dynamic-time-warping']
|
['time-series']
|
[-3.62452388e-01 -1.19072311e-01 -9.73866582e-02 -1.70246020e-01
-4.35164779e-01 -9.54903185e-01 4.58253264e-01 5.15410662e-01
2.31053069e-01 4.65575784e-01 3.31068397e-01 -2.78004944e-01
-9.01616991e-01 -1.06947899e+00 3.56153250e-01 -1.25568354e+00
-4.24049973e-01 8.27492297e-01 -3.72887403e-01 -2.20080629e-01
1.23573892e-01 1.04064202e+00 -1.38412428e+00 1.87178761e-01
1.22703838e+00 8.51815999e-01 1.95348814e-01 9.28827934e-03
5.35051227e-01 3.63878638e-01 -7.00411975e-01 3.18806082e-01
2.82788157e-01 -5.93931854e-01 -1.90588057e-01 -2.34027267e-01
-1.06032491e+00 -6.64949045e-02 1.91077992e-01 8.71785402e-01
8.01064968e-01 3.24537694e-01 8.94199550e-01 -1.50243187e+00
-1.20818973e-01 9.70563591e-01 -3.52628201e-01 2.82842666e-01
2.66082913e-01 3.81939695e-04 1.06279790e+00 -1.42318979e-02
8.19498375e-02 6.51668966e-01 2.94262081e-01 -7.51460865e-02
-1.61277318e+00 -3.88014048e-01 -3.27160329e-01 6.12958133e-01
-1.64364493e+00 1.98621359e-02 1.14153433e+00 -4.47194755e-01
1.20909631e+00 7.14999437e-01 9.18679357e-01 1.66963965e-01
6.50364682e-02 4.57809836e-01 1.21579063e+00 -3.74967307e-01
1.01005077e+00 3.20541970e-02 -1.99117333e-01 -6.81947291e-01
1.23436585e-01 -1.10722207e-01 3.46538782e-01 -3.17130536e-01
-2.72988647e-01 -5.10231793e-01 -3.43813211e-01 -5.25781989e-01
-6.80001736e-01 1.06029499e+00 2.34824836e-01 9.56412733e-01
-5.55248976e-01 -5.94524622e-01 4.90673900e-01 1.25230998e-01
4.19923633e-01 4.48007613e-01 -2.98811972e-01 8.20327550e-03
-1.36762726e+00 4.17509004e-02 1.06427908e+00 5.77783525e-01
3.88908893e-01 2.85109907e-01 -4.65293266e-02 6.44036949e-01
3.95717949e-01 4.91904557e-01 6.81894124e-01 -7.64625013e-01
5.23073301e-02 5.93156159e-01 2.39894971e-01 -6.62581503e-01
-7.05642343e-01 -4.43562955e-01 -7.80149400e-01 2.35459983e-01
8.91888738e-02 -4.34568107e-01 8.96552056e-02 1.15394700e+00
1.54634058e-01 -5.14967144e-01 1.25640184e-01 5.37600279e-01
2.18688510e-02 9.31864560e-01 6.26710653e-02 -9.26352382e-01
9.69239354e-01 -3.34703624e-01 -7.32165575e-01 5.85707486e-01
8.13528359e-01 -5.08852959e-01 2.45110899e-01 3.29035014e-01
-1.14814746e+00 -1.82749763e-01 -1.02418780e+00 6.98626101e-01
-5.49095690e-01 9.75665748e-02 -1.55434415e-01 9.33630824e-01
-1.07951903e+00 6.80606604e-01 -7.59437740e-01 -4.13694620e-01
1.30963130e-02 3.41282874e-01 4.63243276e-01 5.54370105e-01
-9.79081213e-01 1.17476439e+00 4.38964009e-01 1.15774907e-01
-2.57477939e-01 -8.12424064e-01 -4.91063356e-01 5.55586815e-01
-2.55440146e-01 -3.03577989e-01 8.84050727e-01 -8.47494304e-01
-1.28816378e+00 2.89220452e-01 3.24147552e-01 -6.40792429e-01
4.66530204e-01 6.29978538e-01 -8.17521453e-01 7.10553825e-02
-6.08778037e-02 -8.23940709e-02 1.71297416e-01 -1.36972237e+00
-6.84869230e-01 -5.42287648e-01 -5.85595667e-01 4.24644589e-01
2.81485287e-03 -1.14739381e-01 6.50306463e-01 -5.17819226e-01
-2.52792507e-01 -8.01750779e-01 -2.07149521e-01 -1.07401979e+00
-5.18764019e-01 -6.78308249e-01 1.21394813e+00 -8.91918719e-01
1.46350634e+00 -1.94404233e+00 3.27002555e-01 1.03528893e+00
-3.86695445e-01 2.51541138e-02 5.69830835e-01 1.04225540e+00
-5.80839336e-01 -1.07205376e-01 -2.53970385e-01 2.08943233e-01
5.38832366e-01 2.67957538e-01 -1.21470660e-01 5.30944586e-01
-4.25367445e-01 4.71607983e-01 -8.75479102e-01 1.88779101e-01
8.34226966e-01 1.97941527e-01 -7.44606405e-02 -1.02566533e-01
-2.53380030e-01 2.20599353e-01 -2.88280368e-01 2.45921806e-01
6.15091443e-01 3.16444546e-01 6.45369887e-01 -5.09895086e-01
-4.15165126e-01 7.39031062e-02 -1.20089269e+00 8.60241413e-01
-2.92312026e-01 4.31941599e-01 -6.04731143e-02 -1.31301963e+00
9.97237742e-01 1.35973632e-01 1.40350449e+00 -8.18951368e-01
3.29313017e-02 1.48018032e-01 2.41246372e-01 -3.25160325e-01
1.03592344e-01 2.49864578e-01 -3.31441537e-02 7.36313760e-01
-4.04280454e-01 -1.19199738e-01 5.46970963e-01 -1.08640388e-01
6.40103757e-01 -2.95799792e-01 3.98278564e-01 -1.02467847e+00
6.35722578e-01 -1.96934119e-01 4.41526026e-01 -3.31470370e-01
8.57251929e-04 2.11040638e-02 6.99818492e-01 1.66722372e-01
-1.11828613e+00 -1.16446638e+00 -3.09251547e-01 5.10350168e-01
-5.75176589e-02 1.39788181e-01 -8.75345945e-01 -2.49271914e-01
1.50865689e-01 2.00275254e+00 -1.14264488e-01 -7.84860626e-02
-4.14225072e-01 -1.28577280e+00 -2.29676366e-01 1.72425047e-01
1.13904908e-01 -7.14670122e-01 -8.04854453e-01 1.58182278e-01
7.88950361e-03 -5.16140878e-01 -3.85100067e-01 5.55461287e-01
-6.24372303e-01 -8.75356197e-01 -4.17217523e-01 -4.23310101e-01
6.70708537e-01 -2.67945617e-01 9.83796656e-01 -3.17020863e-01
-3.46824527e-02 5.26723146e-01 -5.40473223e-01 -9.88906622e-02
-5.31801939e-01 1.56694680e-01 7.79896975e-03 -6.28677234e-02
5.96754253e-01 -1.11912370e+00 -7.86103129e-01 6.56380296e-01
-7.20224380e-01 -3.72888833e-01 1.10121459e-01 1.88685656e-01
4.22254562e-01 8.20489943e-01 9.66701448e-01 -2.58796811e-01
8.23960483e-01 -8.32493961e-01 -9.39074039e-01 3.25030088e-01
-1.17636120e+00 -4.35981005e-01 7.97359943e-01 4.46015075e-02
-7.90547013e-01 -9.23741162e-02 2.78517287e-02 7.84487277e-02
-1.34929270e-01 3.66640389e-01 -6.12085998e-01 1.05804250e-01
2.75984138e-01 2.81891823e-01 -1.97731227e-01 -4.35668498e-01
3.64272475e-01 7.75931060e-01 1.25726104e-01 -2.83837914e-01
9.64041889e-01 2.54232377e-01 2.01704670e-02 -7.83236206e-01
1.61972955e-01 -4.81820256e-01 -6.39154494e-01 -3.41439962e-01
8.40733171e-01 -6.95461214e-01 -9.81966138e-01 2.87796110e-01
-4.81569916e-01 -3.60538363e-01 -6.95862651e-01 3.01112473e-01
-6.27606153e-01 4.27909791e-01 -5.42357638e-02 -1.01532698e+00
-4.78247792e-01 -1.08806300e+00 2.35005528e-01 4.40926611e-01
-3.50047737e-01 -1.26105714e+00 1.37185037e-01 2.78921217e-01
4.99299228e-01 6.17447853e-01 1.28250325e+00 -1.04242504e+00
-1.10501312e-01 1.03805356e-01 1.73279002e-01 6.66727364e-01
3.14091355e-01 1.81870356e-01 -6.81862295e-01 -7.68796384e-01
9.05708447e-02 3.92928392e-01 -1.38312191e-01 6.61746264e-01
9.12861466e-01 -3.48042190e-01 -3.17567766e-01 1.36278912e-01
1.84553051e+00 1.12619400e+00 7.78625369e-01 2.14983612e-01
1.51404496e-02 7.16833055e-01 3.39222848e-01 8.35881054e-01
6.86011076e-01 8.62919748e-01 4.26565409e-01 -1.11295991e-02
4.48094726e-01 2.35893473e-01 2.97021359e-01 8.35809588e-01
1.80306789e-02 -3.93906325e-01 -7.31132686e-01 7.09795713e-01
-1.74778223e+00 -1.20421767e+00 -7.10847154e-02 2.22399426e+00
1.81078091e-01 -1.28570318e-01 6.37536764e-01 4.71922696e-01
7.59507716e-01 -4.37488928e-02 -3.43147755e-01 -8.61347258e-01
-2.13372618e-01 -2.34546699e-02 6.22260392e-01 2.98204869e-01
-4.92811799e-01 -3.15361142e-01 5.07189894e+00 9.15018976e-01
-8.32450807e-01 -3.52522843e-02 9.27273393e-01 -1.50617734e-01
-5.41903377e-01 -1.12069227e-01 -3.58407021e-01 1.12885809e+00
1.26846087e+00 -1.01255858e+00 7.27769852e-01 7.78120935e-01
1.26649022e+00 -4.63895351e-01 -1.15929472e+00 6.00636721e-01
-1.85236722e-01 -8.51185441e-01 -4.35697556e-01 2.98076212e-01
1.03730237e+00 1.35081664e-01 -3.50694478e-01 -4.49381694e-02
3.74953240e-01 -6.39549971e-01 4.81428206e-01 5.97239792e-01
-1.83688607e-02 -1.32280755e+00 6.61122382e-01 2.72765636e-01
-1.28622472e+00 -6.20637834e-01 1.29886642e-01 3.50501657e-01
6.04023099e-01 9.28781807e-01 -4.05394763e-01 1.02481174e+00
7.20723331e-01 2.19844252e-01 -1.10888146e-01 1.10215509e+00
2.38301292e-01 4.83658582e-01 -6.63311481e-01 -1.30666554e-01
8.22211653e-02 -9.59812164e-01 4.80463982e-01 8.76903594e-01
6.71140492e-01 1.24978013e-01 2.58272551e-02 8.95534337e-01
3.19940239e-01 1.63190380e-01 -2.88945585e-01 1.81695923e-01
9.68580663e-01 1.42387581e+00 -8.28954935e-01 -3.87938619e-02
-6.43500686e-02 5.35150707e-01 -6.14270687e-01 4.54749852e-01
-6.61213517e-01 -3.59627813e-01 5.49704731e-01 2.33132109e-01
3.93049985e-01 5.50748431e-04 -6.12451077e-01 -5.75652719e-01
-1.35108247e-01 -5.99982440e-01 6.16997480e-01 -8.14325571e-01
-1.44029105e+00 5.05994111e-02 5.10992467e-01 -1.21345830e+00
-7.72120833e-01 -1.65978670e-02 -1.22419488e+00 1.06538069e+00
-1.44177306e+00 -7.56753862e-01 1.07721470e-01 5.28532922e-01
3.28428984e-01 -2.70283371e-01 6.18583798e-01 5.21023095e-01
-6.56852782e-01 1.90592408e-01 1.11782277e+00 -1.71586320e-01
-2.80829191e-01 -1.60372996e+00 -3.35611463e-01 5.35883725e-01
-3.77350837e-01 -2.09714457e-01 8.18448186e-01 -4.27351594e-01
-1.04340863e+00 -8.24938178e-01 8.07997644e-01 2.69401282e-01
8.22812617e-01 -1.21951406e-03 -4.31148648e-01 2.25829810e-01
8.45760047e-01 -8.21120799e-01 1.10019219e+00 -3.77649784e-01
4.63160127e-01 -6.30842745e-01 -1.67134309e+00 3.19967568e-01
-5.71944863e-02 -3.24654251e-01 -2.72812515e-01 5.05326867e-01
-1.75451413e-01 2.89236486e-01 -1.60155630e+00 1.74421430e-01
1.81042254e-01 -1.07735538e+00 7.85910606e-01 2.80894786e-01
-4.04645741e-01 -5.73529065e-01 -1.18288629e-01 -1.98782301e+00
-4.29458469e-01 -8.81655574e-01 8.36798325e-02 1.47522438e+00
4.40075070e-01 -9.08707678e-01 4.94110644e-01 6.47294700e-01
-1.00830495e-01 -6.29518628e-01 -1.23455036e+00 -7.29333282e-01
2.41736799e-01 2.06639934e-02 9.20775771e-01 1.21855712e+00
4.62384492e-01 -3.43803056e-02 4.12478894e-01 3.23389113e-01
9.43341255e-01 2.94876993e-01 8.88145864e-02 -8.79290819e-01
5.26512798e-04 -9.21013772e-01 -1.93611994e-01 -2.16566920e-02
-1.07144587e-01 -8.93130958e-01 -4.38397586e-01 -1.72998440e+00
-3.33359092e-01 -4.43375021e-01 -2.34134570e-01 2.28099450e-01
6.24409497e-01 -3.31233203e-01 4.60582703e-01 2.27720335e-01
2.46513486e-01 6.15780294e-01 3.96710366e-01 -9.08173770e-02
-4.14851278e-01 4.30835038e-01 -2.35955626e-01 4.11039442e-01
1.10678601e+00 -6.31621033e-02 -7.01358974e-01 2.02888072e-01
8.50289389e-02 1.76938072e-01 -1.14349961e-01 -8.50603878e-01
2.95342207e-01 -8.34495723e-02 2.10122108e-01 -1.09129894e+00
-3.41337502e-01 -1.47859001e+00 1.01832616e+00 6.91636086e-01
6.77536577e-02 3.60693634e-01 -6.04117140e-02 1.31417498e-01
1.14133343e-01 -2.79803246e-01 8.62932384e-01 3.49016309e-01
-3.54333311e-01 -3.28783602e-01 -7.75574028e-01 -4.84236360e-01
1.86167109e+00 -2.47035995e-01 -2.77195871e-01 -4.91264135e-01
-8.70861709e-01 8.00058007e-01 4.04056013e-01 1.94189385e-01
-1.85675859e-01 -1.37607694e+00 -8.19694638e-01 -2.35647529e-01
-3.10619265e-01 -5.04772723e-01 3.71549755e-01 9.05411720e-01
-5.23939312e-01 3.10206383e-01 -9.97624621e-02 -4.91745234e-01
-9.33087647e-01 5.16556323e-01 7.21578598e-01 -4.42319334e-01
-3.00580859e-01 -2.27403611e-01 -4.12754864e-01 -6.14734180e-02
-1.51593059e-01 -2.44432881e-01 -5.06391644e-01 9.78990853e-01
7.06200674e-02 1.13867521e+00 3.66724968e-01 -6.83792949e-01
-2.72403032e-01 2.92980790e-01 5.56730211e-01 3.83278430e-01
1.39217210e+00 -3.67047101e-01 -3.36188138e-01 2.73928702e-01
1.08355582e+00 -2.82061875e-01 -7.93535054e-01 4.56647485e-01
2.86485136e-01 4.59617190e-02 2.20423549e-01 -1.07843029e+00
-1.28044128e+00 1.58951640e-01 8.94021332e-01 1.40464079e+00
1.64702380e+00 -1.85392052e-01 4.65521574e-01 -1.40022963e-01
1.37242138e-01 -1.73607028e+00 -8.34351838e-01 -2.48245940e-01
6.75516486e-01 -3.34613830e-01 1.51965454e-01 5.21981530e-02
-4.57126021e-01 1.10659063e+00 -2.21801490e-01 -6.25917464e-02
9.04522121e-01 4.36033875e-01 -3.60173553e-01 4.32882011e-02
-4.16833848e-01 -8.72298777e-02 3.65645885e-02 7.93166220e-01
-5.38375974e-02 5.69336534e-01 -8.48373830e-01 4.30540323e-01
-2.52529889e-01 -3.11967105e-01 4.94043171e-01 5.91633260e-01
-2.44974196e-01 -9.80933845e-01 -3.57224762e-01 7.21556783e-01
6.26170859e-02 3.10061127e-01 -6.08224720e-02 7.04260707e-01
4.31474715e-01 1.24126422e+00 3.66299659e-01 -3.66629511e-02
5.98131120e-01 -5.27339987e-02 -3.32751647e-02 1.10428212e-02
-8.41033757e-01 4.24707294e-01 1.79810300e-01 -7.88717046e-02
-4.13422614e-01 -1.30674851e+00 -1.10677075e+00 -6.35073721e-01
-3.60628307e-01 7.30267882e-01 9.19250071e-01 1.01702106e+00
4.40321397e-03 4.84090596e-01 1.61271870e+00 -9.35975671e-01
-5.66207528e-01 -7.86190927e-01 -1.16842103e+00 1.40223667e-01
-3.51705343e-01 -8.86230245e-02 -7.18339205e-01 -1.51320502e-01]
|
[5.728696823120117, 2.575523614883423]
|
2b33745c-c64d-48ed-b7ef-d991bd552fa3
|
uct-learning-unified-convolutional-networks
|
1711.04661
| null |
http://arxiv.org/abs/1711.04661v1
|
http://arxiv.org/pdf/1711.04661v1.pdf
|
UCT: Learning Unified Convolutional Networks for Real-time Visual Tracking
|
Convolutional neural networks (CNN) based tracking approaches have shown
favorable performance in recent benchmarks. Nonetheless, the chosen CNN
features are always pre-trained in different task and individual components in
tracking systems are learned separately, thus the achieved tracking performance
may be suboptimal. Besides, most of these trackers are not designed towards
real-time applications because of their time-consuming feature extraction and
complex optimization details.In this paper, we propose an end-to-end framework
to learn the convolutional features and perform the tracking process
simultaneously, namely, a unified convolutional tracker (UCT). Specifically,
The UCT treats feature extractor and tracking process both as convolution
operation and trains them jointly, enabling learned CNN features are tightly
coupled to tracking process. In online tracking, an efficient updating method
is proposed by introducing peak-versus-noise ratio (PNR) criterion, and scale
changes are handled efficiently by incorporating a scale branch into network.
The proposed approach results in superior tracking performance, while
maintaining real-time speed. The standard UCT and UCT-Lite can track generic
objects at 41 FPS and 154 FPS without further optimization, respectively.
Experiments are performed on four challenging benchmark tracking datasets:
OTB2013, OTB2015, VOT2014 and VOT2015, and our method achieves state-of-the-art
results on these benchmarks compared with other real-time trackers.
|
['Chang Huang', 'Wei Zou', 'Guan Huang', 'Dalong Du', 'Zheng Zhu']
|
2017-11-10
| null | null | null | null |
['real-time-visual-tracking']
|
['computer-vision']
|
[-3.77777249e-01 -6.68014646e-01 -2.49680743e-01 -1.12462685e-01
-4.81424391e-01 -6.26764715e-01 4.49338078e-01 -9.61963907e-02
-8.25616896e-01 5.17583370e-01 -2.20160261e-01 -3.56381275e-02
-8.60571209e-03 -4.23713356e-01 -8.22069407e-01 -6.20438457e-01
-1.06341459e-01 1.09721050e-01 6.32313669e-01 1.54191419e-01
-3.46628129e-01 5.27308464e-01 -1.30519521e+00 -4.59758967e-01
7.75749564e-01 1.47575903e+00 2.15803683e-01 5.80969512e-01
-2.32448820e-02 5.87891161e-01 -6.06564403e-01 -4.85265881e-01
4.21882302e-01 9.62721109e-02 -1.56565174e-01 -1.62717417e-01
7.84766793e-01 -2.83218443e-01 -7.58117378e-01 1.26163042e+00
6.09656334e-01 -1.10449225e-01 2.38522992e-01 -1.38112581e+00
-5.31349182e-01 3.38274986e-01 -5.54320335e-01 2.70976752e-01
-1.36806175e-01 5.04809618e-01 6.31059170e-01 -7.31941164e-01
2.85378009e-01 1.22324586e+00 1.15139043e+00 5.94869018e-01
-9.61651146e-01 -1.13169360e+00 2.85487324e-01 -4.20344882e-02
-1.28672111e+00 -2.61245072e-01 4.20020998e-01 -4.93141651e-01
6.31878734e-01 4.86035235e-02 8.51237297e-01 8.51102769e-01
4.13080990e-01 9.11852598e-01 5.56016624e-01 1.66185305e-01
-2.42532730e-01 -1.17764473e-01 -4.20359261e-02 6.74382329e-01
6.16724312e-01 5.52149415e-01 -3.82563442e-01 1.65847957e-01
8.74512434e-01 2.72057384e-01 -3.29548538e-01 -4.49567407e-01
-1.55388594e+00 5.16108513e-01 9.14312303e-01 8.40441808e-02
-2.49251768e-01 6.32676721e-01 7.56210089e-01 1.81248724e-01
1.31397352e-01 -1.05832793e-01 -5.28705835e-01 -1.78376690e-01
-8.79679441e-01 3.92241597e-01 2.31878668e-01 1.32111216e+00
3.43113631e-01 2.67592460e-01 -7.30544090e-01 3.89181674e-01
6.11210823e-01 9.89057481e-01 4.65535730e-01 -3.88219625e-01
4.05092239e-01 4.96955097e-01 5.06148577e-01 -9.07100856e-01
-6.00952268e-01 -8.74436378e-01 -6.66496158e-01 6.06257170e-02
5.66522777e-01 -3.03724289e-01 -1.02027452e+00 1.68536913e+00
6.01570904e-01 4.80791867e-01 -5.87379672e-02 1.21075284e+00
1.03653538e+00 4.18108553e-01 4.43108797e-01 -6.56344742e-02
1.62386262e+00 -1.15848172e+00 -9.69720006e-01 -1.36223864e-02
4.47909921e-01 -1.05983949e+00 5.66099703e-01 1.10463701e-01
-8.94572973e-01 -1.03723800e+00 -1.08491576e+00 1.19429506e-01
-2.26832315e-01 7.62129784e-01 6.51736438e-01 5.67802131e-01
-6.59903109e-01 5.35918057e-01 -1.18935657e+00 -2.56564230e-01
5.97753048e-01 5.74497461e-01 -1.85181405e-02 3.11540991e-01
-1.03970706e+00 7.79638886e-01 5.02732873e-01 6.19172573e-01
-9.17418242e-01 -7.25744188e-01 -6.72614753e-01 1.38795033e-01
5.33862114e-01 -6.08975351e-01 1.44546294e+00 -4.53115344e-01
-1.43710661e+00 3.27638447e-01 1.62225068e-01 -6.05494499e-01
7.81991661e-01 -6.33377373e-01 -6.63899422e-01 -4.09598708e-01
-1.48396967e-02 6.49050891e-01 7.87184238e-01 -7.55167067e-01
-9.94101584e-01 -8.90733153e-02 -1.30289108e-01 -2.32911348e-01
-4.56539929e-01 2.69248843e-01 -9.44283426e-01 -8.27801645e-01
-1.47424892e-01 -9.49619830e-01 -6.32814169e-02 7.04303622e-01
-2.27037892e-01 -3.47591192e-01 1.39298093e+00 -2.77808994e-01
1.14770770e+00 -2.10735583e+00 -2.64745981e-01 -2.48699844e-01
3.48012447e-01 8.24120641e-01 -9.18460358e-03 -1.26743540e-01
2.95300692e-01 -5.18919706e-01 3.46591204e-01 -3.93662572e-01
3.34137142e-01 1.37741177e-03 -1.70068055e-01 8.10270786e-01
1.42216459e-01 1.19761324e+00 -9.40078199e-01 -7.85588741e-01
5.78809142e-01 7.33096659e-01 -3.15751582e-01 1.33863613e-01
-2.87120104e-01 5.89529216e-01 -6.53323174e-01 8.57460439e-01
8.86197448e-01 -3.19385260e-01 -2.00947717e-01 -5.35347760e-01
-4.81633008e-01 -8.35756063e-02 -1.19281721e+00 1.49009383e+00
-3.57159644e-01 7.54065931e-01 -3.72744314e-02 -4.75727201e-01
9.71895933e-01 2.65677214e-01 4.97414619e-01 -5.17219305e-01
5.81197143e-01 1.82725012e-01 1.58892006e-01 -1.81024238e-01
4.60046053e-01 2.72312552e-01 -1.47266081e-02 -1.52470514e-01
2.91195899e-01 6.16724968e-01 1.82916746e-01 -1.58889949e-01
8.86013329e-01 5.11885524e-01 1.02667011e-01 -2.40272284e-01
6.63700521e-01 8.26024860e-02 1.05981898e+00 6.10241354e-01
-5.99390090e-01 1.77635178e-01 -4.95436564e-02 -6.05307043e-01
-7.55166650e-01 -9.73091424e-01 -1.58960268e-01 8.36413264e-01
6.26094460e-01 -4.75886971e-01 -3.64689916e-01 -6.78387463e-01
2.92553753e-01 -5.27935103e-02 -5.60964763e-01 -1.49256319e-01
-8.17933261e-01 -4.08337086e-01 6.36210442e-01 9.50767219e-01
7.26760566e-01 -9.88795161e-01 -9.53142464e-01 5.85227549e-01
2.71737665e-01 -1.45931685e+00 -1.08865178e+00 8.68204050e-03
-8.01887631e-01 -1.05486643e+00 -8.52690518e-01 -7.43054688e-01
4.77590412e-01 3.77987355e-01 6.62561238e-01 1.65703341e-01
-4.13634002e-01 5.79848699e-02 -1.48005515e-01 -4.97282088e-01
2.02385366e-01 4.65128645e-02 1.38764307e-01 1.76208876e-02
2.95823991e-01 -4.95338216e-02 -8.39696586e-01 6.49017513e-01
-4.30986494e-01 -1.29955992e-01 6.89251721e-01 9.01906848e-01
5.88211060e-01 -2.31009617e-01 2.38842726e-01 -2.60491520e-01
1.12821676e-01 8.35101977e-02 -1.47357988e+00 3.13462913e-01
-3.36171240e-01 -2.06646889e-01 8.18685889e-01 -9.17138338e-01
-8.50647688e-01 3.93137455e-01 6.37452379e-02 -1.06553209e+00
3.43465149e-01 6.39509633e-02 -1.77642912e-01 -6.72283649e-01
3.29167515e-01 3.34176749e-01 -1.97190061e-01 -5.29610157e-01
1.11934252e-01 2.90824264e-01 8.53123426e-01 -3.75228345e-01
1.43079770e+00 3.48336041e-01 -7.69791231e-02 -3.71148914e-01
-8.98313403e-01 -4.83983278e-01 -5.47035873e-01 -6.33928120e-01
8.35596502e-01 -1.19205165e+00 -1.33097970e+00 6.74529254e-01
-1.01795602e+00 -6.77066818e-02 -1.36880323e-01 9.94331658e-01
-2.45094702e-01 7.45815411e-02 -4.74179447e-01 -6.77281857e-01
-6.75487339e-01 -1.35532331e+00 1.21454155e+00 9.15293038e-01
3.83102030e-01 -7.94323206e-01 -9.78022814e-02 -1.75650835e-01
8.62043917e-01 4.85396236e-01 -5.93789220e-02 -4.77427304e-01
-9.34372723e-01 -4.45206493e-01 -6.46445692e-01 -1.06219091e-02
9.21300054e-02 1.15720719e-01 -7.89241672e-01 -8.34445536e-01
-5.01344740e-01 -2.05575004e-01 8.12319160e-01 4.73037273e-01
1.06408775e+00 -1.73136741e-02 -6.77304566e-01 1.07562661e+00
1.52863157e+00 2.01937839e-01 1.23857565e-01 3.93078059e-01
8.80955040e-01 -2.78855175e-01 9.66080129e-01 3.54620248e-01
1.36893034e-01 1.05379379e+00 4.56574291e-01 2.11108513e-02
-3.63330454e-01 -2.07772553e-01 3.97283226e-01 6.46387100e-01
1.41999824e-02 -3.07359006e-02 -6.15004420e-01 5.08216083e-01
-2.07866621e+00 -7.51714408e-01 -1.87439844e-01 2.09005594e+00
5.33099174e-01 4.74160343e-01 3.24121177e-01 -3.79653513e-01
8.13756168e-01 1.62136734e-01 -7.94618905e-01 4.40637320e-01
1.45343155e-01 7.65984505e-02 1.06317425e+00 -3.45980823e-02
-1.53117216e+00 1.03672397e+00 5.18895435e+00 8.57664406e-01
-1.49207497e+00 3.04074019e-01 -2.10011229e-01 -2.24987224e-01
5.37781179e-01 -1.58836037e-01 -1.24730313e+00 6.65515661e-01
6.69933736e-01 -2.81800151e-01 3.33079062e-02 1.08171678e+00
3.82660590e-02 3.95393759e-01 -7.43684292e-01 1.09904325e+00
-4.78878140e-01 -1.25981891e+00 -5.08500278e-01 -4.52976711e-02
3.71902525e-01 2.21320212e-01 7.24087059e-02 7.37515092e-01
2.93227434e-01 -6.69515312e-01 9.97078061e-01 5.09413719e-01
7.18865633e-01 -6.78794563e-01 7.58448422e-01 -2.24719383e-02
-2.15031242e+00 -1.25698531e-02 -6.13670111e-01 4.58195448e-01
2.06901327e-01 2.48342901e-01 -2.52045661e-01 7.29430854e-01
1.05400479e+00 8.73059094e-01 -6.36340320e-01 1.81484139e+00
2.05151066e-02 4.14505154e-01 -4.20801759e-01 -2.58741021e-01
4.48463380e-01 3.49763066e-01 6.00186288e-01 1.25166869e+00
2.95467496e-01 -2.87705898e-01 5.71024418e-01 6.73033953e-01
-1.96778342e-01 -1.27547920e-01 -2.82986183e-02 3.66405547e-02
5.20601273e-01 1.61603844e+00 -7.54008293e-01 -2.30606467e-01
-5.86332858e-01 4.26713228e-01 1.53865740e-01 3.40363905e-02
-1.64573634e+00 -4.53593165e-01 8.73466790e-01 -2.26864204e-01
9.06587124e-01 -1.28288925e-01 2.05735490e-01 -1.28262258e+00
1.06235586e-01 -5.03766179e-01 2.16438398e-01 -3.97276491e-01
-1.20903444e+00 6.57866001e-01 -2.52486169e-01 -1.85912693e+00
3.01053435e-01 -7.12733686e-01 -6.48481429e-01 7.41890669e-01
-1.68526924e+00 -1.37414145e+00 -7.56695688e-01 5.45103908e-01
3.78643811e-01 -1.35289297e-01 1.80257335e-01 8.84171486e-01
-8.08617294e-01 1.08994997e+00 1.79760501e-01 5.65886140e-01
8.85871172e-01 -1.04678619e+00 6.22462988e-01 8.36860001e-01
-8.79499838e-02 6.15859091e-01 5.68710089e-01 -6.27116323e-01
-1.74134064e+00 -1.64975297e+00 4.48069543e-01 -1.70010239e-01
7.44774640e-01 -3.26995969e-01 -8.01708519e-01 7.02052057e-01
1.71478271e-01 8.60421836e-01 1.27831921e-01 -1.28439844e-01
-2.13796720e-01 -2.78036833e-01 -8.17526877e-01 3.41480404e-01
1.12616277e+00 -2.51052845e-02 -3.09009463e-01 2.50770718e-01
7.96848238e-01 -1.07238877e+00 -1.07861376e+00 5.55736244e-01
8.59917760e-01 -4.53190804e-01 1.04723990e+00 -5.17664254e-01
-4.00620967e-01 -1.06161106e+00 2.14253873e-01 -8.23622465e-01
-3.80826205e-01 -7.82668769e-01 -4.61374223e-01 1.16080892e+00
3.71347517e-02 -5.87725341e-01 8.15244973e-01 1.45404711e-01
-2.32709169e-01 -8.63905907e-01 -1.03071642e+00 -1.33245790e+00
-3.50921124e-01 -1.26966938e-01 7.21419632e-01 6.29244149e-01
-7.58796036e-01 -1.99526381e-02 -7.42683887e-01 2.82116979e-01
1.02042317e+00 2.90344775e-01 1.10763967e+00 -1.32363939e+00
-1.03481777e-01 -4.63734716e-01 -7.01817632e-01 -1.31403399e+00
-1.75936848e-01 -6.12646461e-01 3.21687758e-01 -1.11764038e+00
-1.24719786e-02 -5.41730523e-01 -4.26178515e-01 5.38178921e-01
-3.28187048e-01 1.69973686e-01 6.54786468e-01 3.03105921e-01
-1.00035286e+00 8.78108323e-01 1.38909018e+00 -2.09416226e-01
8.07249770e-02 1.35856763e-01 -2.12428614e-01 6.37974560e-01
3.48912925e-01 -7.06784546e-01 -1.47293666e-02 -5.44233620e-01
-4.90257382e-01 -3.99874058e-03 5.90796888e-01 -1.54407740e+00
8.33160758e-01 1.45296782e-01 8.43479753e-01 -1.11815321e+00
2.59861410e-01 -1.11247098e+00 3.17482263e-01 8.70875418e-01
6.80320011e-03 1.21683121e-01 5.65301776e-01 6.76261663e-01
-1.70803919e-01 1.08112998e-01 9.43452477e-01 4.81596470e-01
-7.03056753e-01 9.55572069e-01 2.43022650e-01 -1.76633336e-02
1.11769521e+00 -2.14296505e-01 -3.75788301e-01 1.85280561e-01
-5.17743587e-01 6.75640404e-01 2.77242303e-01 6.92342043e-01
2.77534366e-01 -1.84961414e+00 -4.87056851e-01 -1.46116808e-01
1.53115451e-01 -1.09089920e-02 1.71447977e-01 1.14205849e+00
-3.65581065e-01 6.73068225e-01 -2.90628940e-01 -9.87676561e-01
-1.24180186e+00 6.11398101e-01 5.21012425e-01 -3.75896066e-01
-8.90252471e-01 7.15351403e-01 1.59352094e-01 -2.44251251e-01
7.19268978e-01 -4.97213304e-01 -1.76200122e-02 -1.55255079e-01
6.05500877e-01 -2.07596701e-02 -1.67471915e-01 -7.80814230e-01
-5.87563872e-01 8.76531899e-01 -2.42827609e-01 5.80289900e-01
1.09542477e+00 1.60851642e-01 5.36198378e-01 -4.01190482e-02
9.91085708e-01 -3.59245688e-01 -1.77564108e+00 -5.45050621e-01
7.88958445e-02 -7.12036431e-01 7.14734495e-02 -5.27314961e-01
-1.69502592e+00 5.51859915e-01 1.00574994e+00 -1.44186735e-01
9.86607552e-01 -3.33993524e-01 1.24778831e+00 3.67868572e-01
3.40610653e-01 -8.59803438e-01 -8.56809020e-02 4.87006992e-01
4.11541849e-01 -1.34938371e+00 1.34390146e-01 -2.07533255e-01
-3.08014512e-01 1.11505675e+00 1.14871609e+00 -3.15818995e-01
5.84785819e-01 4.46847260e-01 4.84520495e-02 -1.00807481e-01
-6.61891639e-01 -4.17840362e-01 6.09084487e-01 2.73657113e-01
3.63899976e-01 -1.32358000e-01 -2.89736152e-01 5.70373237e-01
9.25026461e-02 2.21515208e-01 -2.15869308e-01 9.78498638e-01
-3.45025957e-01 -7.89480567e-01 -4.89044398e-01 2.76456505e-01
-4.73776907e-01 3.63801330e-01 1.04926802e-01 1.27533829e+00
2.70494014e-01 4.24203753e-01 -6.90932274e-02 -4.55314249e-01
6.34802282e-01 -5.03573477e-01 4.78142053e-01 -2.55650003e-02
-9.60660458e-01 4.14299816e-01 -2.12760657e-01 -7.34867752e-01
-5.81524432e-01 -7.01525271e-01 -1.16648471e+00 -3.32194030e-01
-8.45354378e-01 7.88540989e-02 5.41651905e-01 8.46986234e-01
2.23579451e-01 9.65263367e-01 3.60830426e-01 -8.95096719e-01
-6.23328745e-01 -9.27184939e-01 -1.74605936e-01 1.43494144e-01
4.42312181e-01 -1.27221966e+00 1.57634303e-01 -1.38320237e-01]
|
[6.309696197509766, -2.1380774974823]
|
0a547910-2330-420f-aa0a-8eeb4b3b6639
|
model-agnostic-vs-model-intrinsic
|
2108.05317
| null |
https://arxiv.org/abs/2108.05317v2
|
https://arxiv.org/pdf/2108.05317v2.pdf
|
Model-agnostic vs. Model-intrinsic Interpretability for Explainable Product Search
|
Product retrieval systems have served as the main entry for customers to discover and purchase products online. With increasing concerns on the transparency and accountability of AI systems, studies on explainable information retrieval has received more and more attention in the research community. Interestingly, in the domain of e-commerce, despite the extensive studies on explainable product recommendation, the studies of explainable product search is still in an early stage. In this paper, we study how to construct effective explainable product search by comparing model-agnostic explanation paradigms with model-intrinsic paradigms and analyzing the important factors that determine the performance of product search explanations. We propose an explainable product search model with model-intrinsic interpretability and conduct crowdsourcing to compare it with the state-of-the-art explainable product search model with model-agnostic interpretability. We observe that both paradigms have their own advantages and the effectiveness of search explanations on different properties are affected by different factors. For example, explanation fidelity is more important for user's overall satisfaction on the system while explanation novelty may be more useful in attracting user purchases. These findings could have important implications for the future studies and design of explainable product search engines.
|
['Lakshmi Narayanan Ramasamy', 'Qingyao Ai']
|
2021-08-11
| null | null | null | null |
['product-recommendation']
|
['miscellaneous']
|
[-7.58555755e-02 5.71234167e-01 -8.37936223e-01 -5.61427295e-01
-2.01014549e-01 -5.90477943e-01 6.44523859e-01 3.15891594e-01
2.46161800e-02 1.92316756e-01 2.62795120e-01 -4.68730897e-01
-5.86541891e-01 -3.57027143e-01 -4.90496904e-01 -1.45792454e-01
4.59375888e-01 7.42248893e-01 -1.85086265e-01 -3.76942217e-01
4.81278092e-01 1.21520787e-01 -1.57652104e+00 4.14707750e-01
1.29897892e+00 8.25655282e-01 1.54386520e-01 4.00098979e-01
-1.39928281e-01 5.77536941e-01 -1.44363552e-01 -9.36055422e-01
3.49460870e-01 -4.27433908e-01 -8.51171494e-01 9.35102720e-03
2.59007961e-01 -3.68721843e-01 -1.47694752e-01 1.04166734e+00
6.96153417e-02 -1.69755220e-02 3.94715458e-01 -1.65547192e+00
-2.05768180e+00 7.54317462e-01 -1.56807043e-02 1.81913488e-02
3.17158222e-01 2.26723269e-01 1.48681331e+00 -8.90610814e-01
4.35874254e-01 1.29455066e+00 2.70172924e-01 6.41145289e-01
-1.21334720e+00 -3.00952524e-01 4.49134737e-01 3.72818381e-01
-1.18928838e+00 -1.87174603e-01 3.94801646e-01 -4.12804574e-01
8.51754487e-01 7.32403040e-01 8.33252668e-01 6.98255420e-01
2.43279546e-01 8.04898500e-01 7.81255364e-01 -3.96320194e-01
4.17237163e-01 6.63829863e-01 6.82452798e-01 3.37538034e-01
9.12439883e-01 1.74448952e-01 -5.58029890e-01 -2.70624995e-01
7.66254425e-01 5.75700760e-01 -2.52099097e-01 -1.31600678e-01
-6.49619162e-01 1.22848940e+00 5.18355787e-01 3.28235298e-01
-6.84854329e-01 2.22970873e-01 -2.58278310e-01 3.20435464e-01
7.21131980e-01 1.04299831e+00 -4.73289877e-01 -1.12005517e-01
-5.24913788e-01 5.03706217e-01 8.97539794e-01 1.08060479e+00
6.60149217e-01 -1.50327668e-01 -3.65722239e-01 3.28240871e-01
8.71218503e-01 4.98168081e-01 5.45092344e-01 -8.31201434e-01
-4.83221821e-02 1.11392999e+00 3.64538938e-01 -1.04314113e+00
-1.18070908e-01 -4.88072902e-01 -1.49153396e-01 -1.13440678e-01
6.68567047e-02 4.07263935e-01 -6.63342178e-01 1.23681843e+00
2.64545437e-02 -3.25707763e-01 -2.08522260e-01 1.64940202e+00
7.84128249e-01 3.02052110e-01 3.63296002e-01 4.92221750e-02
1.46641648e+00 -1.32315850e+00 -1.00510573e+00 -5.18435538e-01
6.98117316e-01 -9.21555936e-01 1.49995911e+00 1.78361639e-01
-1.08307624e+00 -4.78229910e-01 -8.41308236e-01 -4.26343262e-01
-3.78305763e-01 -8.98817778e-02 1.26385570e+00 6.52208984e-01
-6.51788771e-01 5.14548600e-01 -6.02372169e-01 -4.77584898e-01
1.28630832e-01 3.22521210e-01 1.90785788e-02 -8.15897286e-02
-1.16780531e+00 1.17004514e+00 -8.72502774e-02 1.85752615e-01
-4.88214791e-01 -6.00299895e-01 -4.42126900e-01 4.97136116e-01
3.00164104e-01 -9.67920840e-01 1.46445727e+00 -1.34866381e+00
-1.11934638e+00 3.73600692e-01 -3.64139766e-01 -4.33851242e-01
1.46447882e-01 -4.36022311e-01 -4.70703393e-01 -2.97985256e-01
1.88965753e-01 4.17396933e-01 6.08718216e-01 -1.25105202e+00
-4.88114357e-01 -5.15346110e-01 5.17080367e-01 2.37818852e-01
-2.04356357e-01 -8.03414509e-02 -3.64600837e-01 -3.79356652e-01
1.11556366e-01 -1.13580763e+00 -2.59560406e-01 1.58224583e-01
-1.62855387e-01 -5.18167675e-01 2.38159001e-01 -4.67295200e-01
1.44492340e+00 -2.00256395e+00 -6.33274717e-03 4.73128408e-02
6.89438641e-01 2.31156304e-01 -1.94823563e-01 4.87593263e-01
2.36786902e-01 7.93971121e-01 5.41971087e-01 -3.35302025e-01
5.25863826e-01 1.17595911e-01 -3.02108437e-01 9.31600481e-02
-4.80686501e-02 1.41040397e+00 -8.55737925e-01 -1.36795580e-01
-6.66191205e-02 3.75665575e-01 -7.13795006e-01 3.11217271e-02
-3.60745341e-01 1.72647491e-01 -1.09170532e+00 6.98793352e-01
3.54842454e-01 -7.29993761e-01 -1.47262663e-01 6.73026964e-02
9.01522115e-02 3.15146893e-01 -7.34922945e-01 1.19888330e+00
-4.16504413e-01 7.30738461e-01 -2.60498226e-01 -1.99577048e-01
6.93651080e-01 2.72120446e-01 5.49652800e-02 -1.07713783e+00
-1.49939815e-02 4.62890744e-01 3.52368385e-01 -6.93911850e-01
1.01797509e+00 1.87050067e-02 4.74544972e-01 8.00198317e-01
-6.31492317e-01 -4.71906625e-02 -1.62483349e-01 2.15984166e-01
6.81912601e-01 -1.33191913e-01 2.76296735e-01 -2.94968277e-01
1.96270213e-01 3.36657703e-01 2.07286760e-01 1.00024140e+00
-2.54251450e-01 4.76611823e-01 -1.99419186e-02 -7.95432687e-01
-8.94923031e-01 -3.82261097e-01 -1.68892033e-02 1.09433198e+00
8.90559375e-01 -4.98259097e-01 -8.24449539e-01 -5.03510296e-01
3.21122944e-01 1.25377882e+00 -6.65739238e-01 -4.54600990e-01
1.75530910e-02 -6.91384003e-02 4.30865679e-03 5.31200767e-01
2.63432115e-01 -1.05653119e+00 -4.92944270e-01 -9.13970321e-02
-1.31425112e-01 -7.61362672e-01 -6.94225311e-01 -2.37272173e-01
-9.16504443e-01 -8.88510406e-01 -3.90485495e-01 -3.08765978e-01
8.79942477e-01 1.02365625e+00 1.31748247e+00 9.35745895e-01
-2.16754247e-02 6.77409589e-01 -6.29592717e-01 -7.91001141e-01
-2.54370689e-01 2.92762309e-01 9.43988189e-02 -1.08408548e-01
9.92488086e-01 3.92989777e-02 -1.08508658e+00 8.28402340e-01
-1.14867806e+00 6.26674965e-02 4.30406958e-01 4.51818943e-01
4.36874926e-01 -1.56731993e-01 4.84718323e-01 -8.29165518e-01
1.21975172e+00 -6.77982688e-01 -3.54986042e-01 5.39661825e-01
-1.61153650e+00 2.20938683e-01 2.57040054e-01 -4.87823308e-01
-1.03418493e+00 -5.12313306e-01 3.97104204e-01 -4.96298634e-02
1.91269711e-01 6.71503186e-01 2.26002023e-01 -6.42286167e-02
9.67279077e-01 -2.64431387e-01 -4.58308059e-04 -5.94073653e-01
6.35467947e-01 7.40013480e-01 -3.80512863e-01 -8.41181278e-02
6.98141217e-01 4.10591453e-01 -5.43371916e-01 -3.80941302e-01
-8.59224975e-01 -3.98346871e-01 2.52212957e-02 -6.55806661e-02
8.24436367e-01 -5.72872877e-01 -8.51449847e-01 -4.81200665e-01
-1.35731387e+00 8.84360000e-02 -5.09315252e-01 5.59451222e-01
-1.25588357e-01 9.06133577e-02 -4.23896909e-01 -1.02931774e+00
-5.14898956e-01 -1.31253099e+00 1.23155916e+00 4.38414514e-01
-8.31831217e-01 -9.83009875e-01 -4.34943102e-02 7.42253959e-01
8.78867924e-01 -6.62942231e-01 1.00293791e+00 -1.25536406e+00
-1.09559453e+00 -5.34604847e-01 -2.09704727e-01 -1.54823691e-01
2.04149242e-02 -2.37688035e-01 -8.16454172e-01 1.73324779e-01
8.41137096e-02 1.07684016e-01 3.88871223e-01 4.59733635e-01
6.03411734e-01 -7.08350837e-01 -4.12384361e-01 -4.86803465e-02
1.15086424e+00 1.40708506e-01 6.28628075e-01 5.60473025e-01
3.99802089e-01 7.89494693e-01 9.79985535e-01 1.26489565e-01
5.37013590e-01 1.00798774e+00 5.63760400e-01 -9.24945846e-02
5.91586903e-02 -5.78604519e-01 8.52670968e-02 3.32058132e-01
-2.55098552e-01 -3.37328464e-01 -4.23464119e-01 3.37022841e-01
-2.39281607e+00 -7.70328522e-01 -4.32119936e-01 2.12768269e+00
2.92579204e-01 -2.88027108e-01 -2.02950567e-01 -3.03366929e-01
5.16018093e-01 -2.64243096e-01 -7.78794706e-01 -6.97018921e-01
8.16309005e-02 -4.60913002e-01 2.86036938e-01 7.62290359e-01
-2.35485688e-01 9.67351794e-01 6.49441719e+00 5.27170837e-01
-5.96069098e-01 2.06494719e-01 5.89374125e-01 5.76963015e-02
-1.28343451e+00 4.93811607e-01 -6.28860235e-01 3.80597025e-01
6.51003480e-01 -4.39417154e-01 6.93475068e-01 1.31385696e+00
5.54525316e-01 5.22165447e-02 -1.51907325e+00 8.02174807e-01
-9.96132717e-02 -1.26702142e+00 3.39990258e-01 4.65962797e-01
7.21105218e-01 -5.10743856e-01 2.50539690e-01 2.19468534e-01
1.53906733e-01 -1.21927392e+00 9.44174647e-01 5.50574839e-01
-2.91320048e-02 1.31174484e-02 9.39930201e-01 4.49274659e-01
-8.01315308e-01 -1.26765683e-01 -5.37137568e-01 -5.17704904e-01
3.60752225e-01 3.91236871e-01 -7.41039395e-01 5.56324907e-02
3.99278551e-01 3.58353019e-01 -5.57430208e-01 8.99919748e-01
-2.58094639e-01 5.07800817e-01 1.04575343e-01 -5.01968026e-01
2.89006028e-02 -3.36576343e-01 4.03073043e-01 5.00872731e-01
2.71796376e-01 3.67177904e-01 -1.85857713e-01 1.52007699e+00
1.65056095e-01 2.41283923e-01 -5.71915209e-01 -4.42537397e-01
5.89857459e-01 1.21610558e+00 -5.84222138e-01 -2.66890913e-01
-3.73235196e-01 9.88559604e-01 -1.31810438e-02 5.24734855e-01
-8.79295945e-01 3.29408556e-01 7.60005891e-01 5.60084224e-01
3.86733264e-02 4.20512967e-02 -6.74347281e-01 -1.03865743e+00
1.15261920e-01 -1.04182577e+00 -1.20492265e-01 -1.12853944e+00
-1.22299993e+00 6.80757880e-01 -1.84991896e-01 -1.01308882e+00
-2.74504721e-01 -2.89754182e-01 -3.05695117e-01 9.30488467e-01
-1.50554585e+00 -1.12223446e+00 -4.98581916e-01 8.94895718e-02
1.02235830e+00 1.78002827e-02 7.58438766e-01 9.76656154e-02
-2.79153705e-01 3.37915361e-01 -1.56858727e-01 -6.87391996e-01
4.27681655e-01 -8.67196500e-01 6.64081752e-01 3.42912287e-01
4.17207986e-01 1.38791394e+00 9.52358782e-01 -9.08538401e-01
-1.72821248e+00 -5.26442468e-01 1.42055845e+00 -9.99941409e-01
4.43376899e-01 -5.88484779e-02 -8.89834702e-01 8.23330581e-01
3.20657283e-01 -4.26650733e-01 8.22498441e-01 4.26961273e-01
-1.70423508e-01 1.46916896e-01 -1.05076182e+00 8.24218929e-01
1.13584220e+00 -4.95397300e-01 -3.95467818e-01 4.23651814e-01
1.07873702e+00 3.90176214e-02 -6.35774076e-01 -6.04024939e-02
7.75831461e-01 -9.90332305e-01 6.21506631e-01 -8.48654628e-01
4.01817799e-01 -2.88463593e-01 2.87578627e-02 -1.10079157e+00
-7.79176116e-01 -6.67358041e-01 4.64243889e-02 1.09111524e+00
1.01103663e+00 -9.92165506e-01 7.88925827e-01 2.02213550e+00
7.79949203e-02 -6.29355252e-01 -3.93070042e-01 -5.98843575e-01
-3.14301372e-01 -3.25044930e-01 1.07936180e+00 5.26618421e-01
1.69009775e-01 3.43592823e-01 -2.82502234e-01 1.31797895e-01
2.51686960e-01 2.62011349e-01 7.62603462e-01 -1.37492943e+00
-5.23167670e-01 -3.75702292e-01 -1.66624457e-01 -1.42898560e+00
-4.30095196e-02 -7.30022073e-01 -2.14442939e-01 -1.67443240e+00
5.81869364e-01 -4.13200766e-01 1.60912015e-02 3.19282442e-01
-3.51451784e-01 -9.27755982e-02 4.91657943e-01 9.08857107e-01
-8.40647399e-01 5.08293450e-01 1.29566002e+00 -9.74589437e-02
-3.21060926e-01 2.74794530e-02 -1.44710231e+00 5.90974152e-01
4.52644229e-01 -4.77968991e-01 -7.81850040e-01 -7.92169452e-01
8.22997153e-01 -1.71778262e-01 3.71351391e-01 -1.35709196e-01
2.67115444e-01 -3.30909729e-01 -2.21102953e-01 1.00299977e-01
1.79911748e-01 -1.28653288e+00 5.94204843e-01 3.39892447e-01
-8.48141789e-01 3.64095509e-01 -1.29459143e-01 7.63215184e-01
-1.45250022e-01 -5.95713019e-01 -2.23975271e-01 -8.74398742e-03
-5.08945227e-01 1.72809675e-01 -2.69725293e-01 -4.30595398e-01
7.74978101e-01 -7.04101145e-01 -4.95303124e-01 -8.44242454e-01
-4.11995709e-01 2.04695612e-01 4.65635240e-01 8.84349585e-01
4.47379738e-01 -1.30272138e+00 -3.71215850e-01 -1.72877181e-02
2.97727525e-01 -5.47461152e-01 2.59796251e-02 5.73494077e-01
-1.49656579e-01 1.09367585e+00 2.97376007e-01 -4.20946747e-01
-8.93946052e-01 5.83019614e-01 1.35260895e-01 -5.97948954e-02
-2.09877826e-02 5.03995001e-01 4.21719909e-01 -1.01563402e-01
3.32779326e-02 -4.10445780e-01 -9.84330326e-02 -4.94918555e-01
4.60613638e-01 6.14442170e-01 -1.77324355e-01 -4.24938589e-01
-1.54988572e-01 4.36741769e-01 -2.64245450e-01 4.74041663e-02
9.82714295e-01 -4.95607644e-01 1.42726853e-01 1.57834008e-01
6.83711469e-01 -1.59529790e-01 -8.56813490e-01 3.80994678e-02
7.85135031e-02 -8.44114542e-01 3.62555198e-02 -9.35509920e-01
-1.02547669e+00 6.08215511e-01 6.73560262e-01 9.88038361e-01
5.79532385e-01 3.39688092e-01 6.34057760e-01 1.98527575e-01
5.28299987e-01 -8.65601838e-01 -3.32471877e-02 -1.85922787e-01
1.06316328e+00 -1.62820005e+00 1.51985243e-01 -7.33594239e-01
-1.06600749e+00 7.95892179e-01 5.64067662e-01 3.10951114e-01
5.57561040e-01 -5.75656116e-01 6.21087551e-02 -7.09432364e-01
-6.44156039e-01 -1.44931465e-01 8.17471325e-01 3.55673850e-01
8.28069329e-01 1.55427799e-01 -8.01837564e-01 1.21965301e+00
-4.49541658e-02 2.85159528e-01 7.02202246e-02 4.85017508e-01
-5.22885919e-01 -1.08817542e+00 -6.89326152e-02 2.83530354e-01
-2.62653261e-01 -4.15067345e-01 -8.15088034e-01 7.90550470e-01
-2.36562788e-01 1.47920299e+00 -2.67249703e-01 -2.79532075e-01
4.47009921e-01 -1.18907765e-01 4.51711519e-03 -6.18303955e-01
-8.40161264e-01 -3.24738771e-01 1.80638731e-01 -7.33363569e-01
-3.11051518e-01 -2.94084668e-01 -1.22729945e+00 -4.79316264e-01
-1.04373372e+00 4.77078825e-01 9.80390549e-01 1.09423900e+00
1.14460135e+00 -8.25281218e-02 2.52836019e-01 -5.10942817e-01
-7.31673419e-01 -8.18773508e-01 -5.37618876e-01 8.12237024e-01
-9.46029648e-03 -6.24703407e-01 -5.62592268e-01 -9.91611406e-02]
|
[9.409615516662598, 5.98390007019043]
|
6991e46e-03de-43e0-bae4-f495a0a39a36
|
sampling-of-bayesian-posteriors-with-a-non
|
1910.12717
| null |
https://arxiv.org/abs/1910.12717v1
|
https://arxiv.org/pdf/1910.12717v1.pdf
|
Sampling of Bayesian posteriors with a non-Gaussian probabilistic learning on manifolds from a small dataset
|
This paper tackles the challenge presented by small-data to the task of Bayesian inference. A novel methodology, based on manifold learning and manifold sampling, is proposed for solving this computational statistics problem under the following assumptions: 1) neither the prior model nor the likelihood function are Gaussian and neither can be approximated by a Gaussian measure; 2) the number of functional input (system parameters) and functional output (quantity of interest) can be large; 3) the number of available realizations of the prior model is small, leading to the small-data challenge typically associated with expensive numerical simulations; the number of experimental realizations is also small; 4) the number of the posterior realizations required for decision is much larger than the available initial dataset. The method and its mathematical aspects are detailed. Three applications are presented for validation: The first two involve mathematical constructions aimed to develop intuition around the method and to explore its performance. The third example aims to demonstrate the operational value of the method using a more complex application related to the statistical inverse identification of the non-Gaussian matrix-valued random elasticity field of a damaged biological tissue (osteoporosis in a cortical bone) using ultrasonic waves.
|
['Christian Soize', 'Roger Ghanem']
|
2019-10-28
| null | null | null | null |
['small-data']
|
['computer-vision']
|
[ 3.62437159e-01 5.22302449e-01 1.42872423e-01 2.16634795e-01
-7.06407607e-01 -6.03503198e-04 4.69657570e-01 1.33162215e-01
-4.59499091e-01 1.02348602e+00 -2.87090689e-01 -1.20498940e-01
-7.09590018e-01 -4.34532106e-01 -6.65119946e-01 -1.23893189e+00
-5.16813993e-01 7.16495395e-01 1.39457658e-01 -1.06701568e-01
2.48317108e-01 5.21217167e-01 -1.67852664e+00 -4.95425731e-01
7.38483429e-01 1.16932905e+00 3.66144091e-01 6.13374352e-01
4.26511884e-01 2.55471677e-01 -4.20234114e-01 2.09245861e-01
4.51407768e-02 -3.62671077e-01 -6.40714824e-01 3.66543382e-01
-1.47051781e-01 -7.53220916e-03 2.47376859e-01 1.20895076e+00
6.33217871e-01 2.69517452e-01 1.32514012e+00 -8.36069286e-01
-1.54170364e-01 2.44558185e-01 -1.49003819e-01 4.88070510e-02
3.31264250e-02 -9.84399486e-03 5.78511596e-01 -9.87287879e-01
5.61328411e-01 1.12426090e+00 5.59663236e-01 2.15141043e-01
-1.61672103e+00 1.65727288e-01 -5.43961108e-01 9.56720114e-03
-1.70000219e+00 -3.77478421e-01 5.46532154e-01 -7.64611185e-01
2.05996379e-01 3.05197448e-01 5.42561114e-01 7.09943712e-01
3.19383293e-01 1.30356222e-01 1.26051199e+00 -6.72548115e-01
8.39073956e-01 2.83529073e-01 5.21498397e-02 5.51812649e-01
3.34529996e-01 2.34014645e-01 -3.78437564e-02 -4.26725179e-01
9.13676143e-01 -3.96724254e-01 -1.71994939e-01 -5.07652164e-01
-1.12315261e+00 9.53482151e-01 -1.78308949e-01 6.83387697e-01
-5.46878636e-01 -9.93619487e-02 1.71978578e-01 2.76306748e-01
4.92541999e-01 3.74854594e-01 -2.75700182e-01 1.05894193e-01
-8.24481964e-01 2.55310446e-01 1.03264356e+00 7.04847276e-01
8.16996634e-01 1.58121243e-01 2.39051357e-01 5.45733452e-01
4.28057760e-01 9.18834269e-01 2.70181596e-01 -1.04609227e+00
1.08448662e-01 2.18230672e-02 3.45875800e-01 -9.02562320e-01
-6.34190619e-01 -2.60148644e-01 -9.99862432e-01 4.60019857e-01
9.36017156e-01 -2.05141261e-01 -6.25299633e-01 1.54699075e+00
4.99332964e-01 -4.92864624e-02 -4.63074967e-02 7.37791896e-01
4.10734147e-01 5.30355930e-01 -3.07890475e-01 -8.27404141e-01
1.47182620e+00 1.28438249e-01 -7.08352268e-01 2.37840027e-01
4.30356979e-01 -7.44991720e-01 7.86655664e-01 4.81765538e-01
-1.21856976e+00 -3.86574507e-01 -9.73215878e-01 4.39302236e-01
-6.47119656e-02 1.66907340e-01 1.88768357e-01 5.92127562e-01
-7.09553540e-01 8.91123593e-01 -9.80865359e-01 -2.15755790e-01
1.21513672e-01 3.30414683e-01 -2.66390353e-01 -3.62440310e-02
-9.64188278e-01 1.02912915e+00 2.81435668e-01 3.45074475e-01
-7.48426974e-01 -7.20391035e-01 -5.86118579e-01 6.51255623e-02
3.21549594e-01 -5.92315614e-01 7.30724812e-01 -4.97415721e-01
-1.70255959e+00 4.59504515e-01 2.19099075e-02 1.02351561e-01
7.74891913e-01 8.52238238e-02 -3.04254331e-02 3.85163695e-01
2.77960915e-02 -5.56951314e-02 1.11230946e+00 -1.16766465e+00
1.55396849e-01 -4.56246793e-01 -4.83951360e-01 -3.05248588e-01
5.68847433e-02 -2.55542666e-01 1.45370588e-01 -6.29923880e-01
4.16343987e-01 -8.59005570e-01 -5.14419734e-01 -2.28879958e-01
-6.03749037e-01 -1.66494459e-01 2.77920932e-01 -5.56201875e-01
9.56400931e-01 -2.15397811e+00 4.90052462e-01 7.15667009e-01
4.44334634e-02 -1.51420981e-01 1.58718437e-01 7.90480494e-01
-2.09499881e-01 -6.34908751e-02 -7.13641584e-01 -1.43828206e-02
-1.37949258e-01 3.02832089e-02 8.10270831e-02 1.04359722e+00
1.75745368e-01 2.73019791e-01 -7.02387214e-01 -7.83914149e-01
1.60777524e-01 3.65793914e-01 -2.66092807e-01 2.10204855e-01
-4.61694449e-02 7.00846851e-01 -6.01982117e-01 3.14135015e-01
6.34706080e-01 -2.66071677e-01 -5.87018728e-02 -4.32927310e-01
-1.72058731e-01 -3.25503439e-01 -1.95218289e+00 1.12248540e+00
-4.06205624e-01 2.17391551e-01 4.06428009e-01 -1.53969920e+00
8.34707618e-01 5.22546411e-01 8.37730944e-01 -3.19375806e-02
3.52953464e-01 5.00195682e-01 1.82500973e-01 -8.35873008e-01
-2.98128724e-01 -6.24704838e-01 5.65193594e-02 6.31126285e-01
2.38573700e-01 -6.22045577e-01 2.51440287e-01 -6.56701550e-02
8.89729381e-01 -1.00090265e-01 5.68449676e-01 -9.66074586e-01
6.34177744e-01 -3.77126604e-01 4.16642189e-01 5.01325130e-01
2.46001616e-01 3.53837579e-01 6.87766790e-01 1.44297052e-02
-1.14465034e+00 -1.16664863e+00 -8.03199410e-01 2.87851691e-01
-4.72761132e-02 1.33047430e-02 -8.66425991e-01 4.23583947e-02
3.32777016e-02 5.77949941e-01 -6.81970537e-01 1.73254237e-01
-4.24753100e-01 -1.02775574e+00 -4.98070195e-02 8.39525908e-02
-6.79744035e-02 -6.86998308e-01 -6.68565869e-01 1.85539931e-01
-1.79224312e-01 -9.50122774e-01 3.10808569e-02 2.00546622e-01
-1.18139362e+00 -1.23426867e+00 -8.65981519e-01 -6.19023681e-01
8.36426258e-01 -5.13228595e-01 8.40276122e-01 -1.44226179e-01
-5.63483000e-01 7.21946299e-01 3.38934697e-02 -4.35288459e-01
-8.46187174e-01 -4.26722944e-01 2.78748751e-01 1.96314245e-01
-2.12246239e-01 -7.32777536e-01 -3.80853206e-01 4.88498479e-01
-9.68156219e-01 -4.84409988e-01 6.53536916e-01 1.15216088e+00
6.42066836e-01 3.86021644e-01 7.88358510e-01 -5.83557248e-01
4.55797315e-01 -5.14559150e-01 -7.96400726e-01 -9.24786702e-02
-5.66073775e-01 2.38075927e-01 4.96879399e-01 -5.79602063e-01
-7.20853567e-01 7.88604375e-04 1.71898201e-01 -4.11973372e-02
-2.57112414e-01 6.80910468e-01 -2.83745468e-01 7.67638348e-03
7.48830974e-01 1.22101024e-01 6.54869020e-01 -7.29328692e-01
2.42364675e-01 4.45331693e-01 5.37842035e-01 -7.08489716e-01
6.35180354e-01 5.59074163e-01 8.30984592e-01 -1.55186367e+00
-2.68072903e-01 -3.72876257e-01 -9.65010524e-01 -3.24745387e-01
7.40123928e-01 -4.54205871e-01 -1.00589776e+00 4.06571954e-01
-8.28605115e-01 -2.31858552e-01 -5.86478651e-01 9.23926651e-01
-1.25397408e+00 5.86341858e-01 -5.43422759e-01 -1.39318466e+00
1.33057265e-02 -1.17868781e+00 9.57144260e-01 -1.67567357e-01
-1.65084243e-01 -1.23523104e+00 -6.59482256e-02 3.64959091e-02
4.35059309e-01 4.48382556e-01 1.17810237e+00 -4.43443418e-01
-3.94417852e-01 -3.28419358e-01 1.48220986e-01 4.98176306e-01
1.40548140e-01 -2.82811671e-02 -6.72244310e-01 -3.94797146e-01
8.05705130e-01 -1.73252374e-01 3.88592422e-01 7.44605362e-01
8.66039455e-01 -2.06968993e-01 -3.60146493e-01 -1.00092180e-01
1.32205200e+00 -3.45743388e-01 4.89160657e-01 -2.20866144e-01
4.43182826e-01 1.06706405e+00 7.38658726e-01 5.96418738e-01
-3.37419689e-01 7.97812283e-01 3.99444193e-01 1.65769279e-01
1.52692735e-01 4.40952569e-01 1.53737172e-01 8.79983783e-01
-3.59023809e-01 1.94885224e-01 -6.59719944e-01 3.99990410e-01
-1.60967875e+00 -9.37796772e-01 -3.02384883e-01 2.67646360e+00
7.54119456e-01 -9.43190884e-03 2.10091889e-01 4.87523079e-01
7.83271849e-01 -4.99685228e-01 -3.63929510e-01 9.41828042e-02
1.16844960e-02 3.44370455e-01 3.96830797e-01 6.65954173e-01
-8.74218106e-01 3.55309667e-03 6.53444767e+00 1.01096439e+00
-8.45315576e-01 -2.10386161e-02 5.12479186e-01 4.07295674e-01
-9.80045274e-03 -9.80886538e-03 -5.11489987e-01 6.15070343e-01
1.17411196e+00 -8.89492221e-03 2.21068606e-01 4.90382552e-01
4.03962493e-01 -6.27284408e-01 -1.12877119e+00 8.61579776e-01
-1.89592361e-01 -1.05525005e+00 -3.68804812e-01 3.56272280e-01
3.72683287e-01 -2.01166123e-01 -2.68412549e-02 -3.12621713e-01
-2.39636704e-01 -9.93467689e-01 5.96273482e-01 9.79370058e-01
7.65026867e-01 -5.41558206e-01 7.92994261e-01 6.24770463e-01
-7.72406220e-01 -6.34559914e-02 -4.71224010e-01 1.13186948e-01
4.95109558e-01 1.18083656e+00 -7.96999037e-01 5.51371932e-01
3.38166982e-01 3.56159002e-01 -2.01371580e-01 1.25078225e+00
2.16553226e-01 8.07975948e-01 -7.62324631e-01 -1.87131733e-01
-2.34485582e-01 -6.43988252e-01 9.93132353e-01 8.88369739e-01
6.54791355e-01 -8.45915601e-02 4.37090062e-02 1.06893122e+00
6.28884792e-01 3.23705614e-01 -5.13926387e-01 3.90108190e-02
2.43483499e-01 1.04605258e+00 -8.52456689e-01 3.48885618e-02
-3.59120779e-02 2.72664696e-01 -2.10613504e-01 4.21995431e-01
-3.09979975e-01 -1.96362168e-01 1.21031530e-01 3.98784876e-01
4.27522957e-01 -2.06855074e-01 -1.86514601e-01 -7.29338467e-01
5.39333671e-02 -4.89500910e-01 1.96688592e-01 -3.75547826e-01
-1.35832012e+00 2.98541076e-02 6.19323492e-01 -1.28797531e+00
-6.50472283e-01 -9.04583395e-01 -4.76149231e-01 1.00845599e+00
-7.74436057e-01 -7.44202554e-01 2.44869053e-01 4.75887805e-01
-2.43868250e-02 -2.48913318e-01 1.03780723e+00 3.16973031e-01
-3.27159643e-01 -3.30765471e-02 5.38978279e-01 -1.95380941e-01
2.39050180e-01 -1.38368130e+00 -3.46770644e-01 4.96855915e-01
-4.43011791e-01 5.05631864e-01 1.34342742e+00 -5.49587548e-01
-1.48875880e+00 -6.23938739e-01 5.82444489e-01 -3.42001528e-01
8.66600156e-01 -3.89902413e-01 -1.01075256e+00 6.16842173e-02
-3.99986714e-01 3.19136888e-01 5.79099715e-01 -1.36400595e-01
3.36748213e-01 6.51578382e-02 -1.28175235e+00 2.68155992e-01
2.23015621e-01 -1.76989421e-01 -4.55052584e-01 6.21742129e-01
7.31858239e-02 -4.85528223e-02 -1.45434344e+00 4.12833512e-01
4.20530915e-01 -6.66238427e-01 9.80996907e-01 -4.44175988e-01
-1.33284628e-02 -3.23057562e-01 -3.17182541e-01 -1.07215571e+00
-1.01370066e-01 -7.60855317e-01 -2.54297048e-01 9.86367047e-01
2.34887183e-01 -8.17816138e-01 4.83139306e-01 2.86416829e-01
1.13928534e-01 -7.82597482e-01 -1.60221910e+00 -1.03233767e+00
3.55508238e-01 -3.06189746e-01 -6.12152293e-02 5.80181122e-01
6.68834597e-02 3.29434097e-01 -1.72762483e-01 2.02507049e-01
9.78169978e-01 -1.45309806e-01 6.22292399e-01 -1.65242982e+00
-5.91852546e-01 -2.19735444e-01 -7.29242265e-01 -6.76087081e-01
-3.07660867e-02 -5.29032111e-01 3.27186257e-01 -1.06130469e+00
6.33698404e-02 -5.51757455e-01 8.75503272e-02 -3.60724866e-01
4.71403114e-02 -1.56117380e-01 -2.76889473e-01 2.79754668e-01
1.01017542e-01 5.10476291e-01 1.36050284e+00 4.88097891e-02
3.28290313e-02 5.65929294e-01 -2.54332244e-01 6.62943244e-01
3.40708643e-01 -4.13769096e-01 -4.08670634e-01 6.16215765e-01
1.74751058e-01 6.24360204e-01 5.57978690e-01 -9.75485682e-01
-5.61087914e-02 -1.13141753e-01 7.55054206e-02 -3.79059553e-01
3.83803457e-01 -8.08892012e-01 2.70873427e-01 5.01484692e-01
-4.35980670e-02 -3.19940239e-01 -2.32027918e-01 8.28695834e-01
-5.16380370e-02 -8.73581588e-01 1.07365191e+00 5.76408058e-02
-1.18509429e-02 -2.22537983e-02 -6.87717199e-01 -1.75523218e-02
8.75085711e-01 1.24209419e-01 1.48395330e-01 -5.07772624e-01
-1.27182698e+00 -1.06581509e-01 1.89413115e-01 -3.49559605e-01
3.65357429e-01 -1.14364517e+00 -9.23718274e-01 1.71690360e-01
-7.46148080e-02 5.18328361e-02 2.38563225e-01 1.51817536e+00
-2.78528512e-01 1.52204484e-01 2.50700563e-01 -9.85675573e-01
-9.27252948e-01 7.59558737e-01 3.56889635e-01 -5.78627512e-02
-5.74378371e-01 4.17901039e-01 7.31949508e-02 -1.69978827e-01
-5.40739819e-02 -3.96141768e-01 -2.35856831e-01 3.79874930e-02
3.86920959e-01 9.54171300e-01 1.57181352e-01 -7.23940372e-01
-5.05005047e-02 5.95449805e-01 5.51887989e-01 -3.84281546e-01
1.42679441e+00 -2.89273590e-01 -3.55269670e-01 9.44535732e-01
1.24589694e+00 -7.41180703e-02 -1.12694550e+00 -2.10483029e-01
1.03074819e-01 -9.62090343e-02 -2.98027825e-02 -2.85768718e-01
-6.63965702e-01 8.12469065e-01 7.42855191e-01 5.49777091e-01
7.87265897e-01 2.99965739e-01 5.51103055e-02 4.77957457e-01
2.71547139e-01 -1.38264251e+00 6.01614453e-02 1.89905748e-01
1.17081535e+00 -1.04999685e+00 3.68299991e-01 -6.68795109e-01
-5.82787357e-02 1.46269608e+00 -1.79360166e-01 -2.28183255e-01
1.24222493e+00 2.54614174e-01 -1.35978431e-01 -3.29710066e-01
-1.88974679e-01 -1.50008470e-01 4.75353569e-01 6.15087569e-01
2.45742500e-01 -4.79391962e-02 -5.60352564e-01 3.54023099e-01
-1.45195156e-01 -2.07091346e-01 6.60576105e-01 8.67671549e-01
-4.21318769e-01 -7.23994434e-01 -7.27031767e-01 3.97272646e-01
-4.84974951e-01 2.39980966e-01 2.79415786e-01 8.88145387e-01
-8.41989964e-02 9.59059536e-01 -1.38704956e-01 3.09451193e-01
1.35845870e-01 -6.76595867e-02 5.85455954e-01 -6.39711440e-01
4.41122055e-01 5.19609332e-01 -2.09212787e-02 -4.54498649e-01
-5.52240312e-01 -1.18817830e+00 -1.13379240e+00 2.02306330e-01
-7.14531839e-01 5.02936244e-01 8.01039875e-01 1.26379800e+00
-1.09852739e-01 3.22897017e-01 6.04066789e-01 -1.13828230e+00
-1.08067048e+00 -1.35859895e+00 -1.26215470e+00 3.13326180e-01
3.61168951e-01 -1.04199719e+00 -7.11373150e-01 1.93817601e-01]
|
[6.919426918029785, 4.014157295227051]
|
3e019778-6535-4357-a6a7-6b97e72bbff1
|
curiosity-killed-the-cat-and-the
|
2006.03357
| null |
https://arxiv.org/abs/2006.03357v2
|
https://arxiv.org/pdf/2006.03357v2.pdf
|
Curiosity Killed or Incapacitated the Cat and the Asymptotically Optimal Agent
|
Reinforcement learners are agents that learn to pick actions that lead to high reward. Ideally, the value of a reinforcement learner's policy approaches optimality--where the optimal informed policy is the one which maximizes reward. Unfortunately, we show that if an agent is guaranteed to be "asymptotically optimal" in any (stochastically computable) environment, then subject to an assumption about the true environment, this agent will be either "destroyed" or "incapacitated" with probability 1. Much work in reinforcement learning uses an ergodicity assumption to avoid this problem. Often, doing theoretical research under simplifying assumptions prepares us to provide practical solutions even in the absence of those assumptions, but the ergodicity assumption in reinforcement learning may have led us entirely astray in preparing safe and effective exploration strategies for agents in dangerous environments. Rather than assuming away the problem, we present an agent, Mentee, with the modest guarantee of approaching the performance of a mentor, doing safe exploration instead of reckless exploration. Critically, Mentee's exploration probability depends on the expected information gain from exploring. In a simple non-ergodic environment with a weak mentor, we find Mentee outperforms existing asymptotically optimal agents and its mentor.
|
['Marcus Hutter', 'Elliot Catt', 'Michael K. Cohen']
|
2020-06-05
| null | null | null | null |
['safe-exploration']
|
['robots']
|
[-2.34143883e-01 3.99095178e-01 -1.99540406e-01 8.52713659e-02
-6.91511273e-01 -6.53552532e-01 4.02913809e-01 3.42571437e-02
-9.69356239e-01 1.27790773e+00 1.17395535e-01 -7.25007296e-01
-3.33433896e-01 -8.05318773e-01 -8.24420154e-01 -1.02388632e+00
-5.49687266e-01 6.09211802e-01 -8.97614956e-02 -3.47891062e-01
5.81463635e-01 3.18544298e-01 -1.31249607e+00 -6.14928186e-01
8.89286458e-01 2.09443852e-01 3.46972942e-01 1.09970152e+00
1.77917704e-01 1.09402657e+00 -7.67532349e-01 3.04962462e-03
4.25041616e-01 -7.62487531e-01 -1.10856295e+00 -5.38690612e-02
-4.38307196e-01 -8.43246579e-01 -4.18925345e-01 1.31050360e+00
1.84398785e-01 4.38654959e-01 4.64271069e-01 -1.18974197e+00
-4.32800114e-01 1.16590238e+00 -4.28710341e-01 9.95637774e-02
2.85330951e-01 5.28807104e-01 9.40177679e-01 1.94857150e-01
4.39585358e-01 1.23093188e+00 1.51366621e-01 6.87326908e-01
-1.06235325e+00 -5.34424961e-01 2.96567947e-01 5.90951256e-02
-9.12270069e-01 -2.85393745e-01 3.75953645e-01 7.04836920e-02
7.18957901e-01 1.62051022e-01 9.38565016e-01 1.10018587e+00
4.90660638e-01 9.49106276e-01 1.35666108e+00 -3.54825318e-01
8.15013528e-01 9.93533731e-02 -4.10520166e-01 4.13670123e-01
5.91051221e-01 8.67115498e-01 -3.90103698e-01 -2.89922148e-01
8.55711281e-01 -6.25414252e-02 -7.30512068e-02 -3.96951437e-01
-1.17247677e+00 7.47789085e-01 1.35103688e-01 8.90299901e-02
-7.19620824e-01 6.09826684e-01 1.57178909e-01 8.82433474e-01
-2.60792434e-01 8.71898174e-01 -2.23927468e-01 -7.82232046e-01
-4.66225773e-01 5.85944414e-01 1.06278694e+00 6.60119534e-01
5.21704733e-01 3.71587306e-01 3.36420119e-01 2.24454924e-02
2.26356015e-01 6.21251047e-01 4.73650306e-01 -1.46952391e+00
1.37180910e-01 2.35492680e-02 9.47862089e-01 -3.05499583e-01
-2.77576298e-01 -4.99400079e-01 2.00518034e-02 9.03241634e-01
4.86132443e-01 -7.41324365e-01 -4.97044057e-01 1.89286268e+00
1.53347775e-01 -9.06495675e-02 5.73944628e-01 8.44255149e-01
-3.84664863e-01 6.03522420e-01 -3.41434106e-02 -4.92962539e-01
8.57440114e-01 -5.11416972e-01 -5.34905016e-01 -1.59298837e-01
6.93401575e-01 -3.31252366e-01 1.15435469e+00 5.84249556e-01
-1.18321419e+00 2.16735572e-01 -1.16894877e+00 8.23812068e-01
1.51983276e-01 -9.12573040e-01 5.94980657e-01 7.21625865e-01
-1.06313205e+00 9.42192197e-01 -1.06987154e+00 -3.36573839e-01
1.75974876e-01 3.08884352e-01 -2.66772136e-02 3.29042822e-01
-9.47911203e-01 1.11313152e+00 4.25075263e-01 -2.82616198e-01
-1.77133965e+00 7.81145543e-02 -4.03370440e-01 1.13471702e-01
8.97830963e-01 -2.69551426e-01 1.59568572e+00 -1.01246476e+00
-1.65654027e+00 6.39559701e-02 1.44867986e-01 -7.27606595e-01
8.12428355e-01 -2.96461642e-01 5.84322400e-02 1.62996680e-01
6.14649989e-02 4.52836514e-01 6.26136720e-01 -1.31913364e+00
-7.91983843e-01 -2.34518304e-01 5.30804813e-01 7.93044031e-01
-3.99080127e-01 -1.21044099e-01 5.79566419e-01 2.64174677e-02
-2.39570841e-01 -1.06783378e+00 -6.83725297e-01 -5.22436082e-01
-1.62477359e-01 -2.88452357e-01 5.47506511e-01 6.51435629e-02
5.89684010e-01 -2.00543714e+00 -2.77955234e-01 2.12667167e-01
1.06010027e-01 -1.40017629e-01 -2.11448953e-01 6.18513465e-01
2.28589058e-01 2.08450884e-01 1.18378974e-01 7.67275840e-02
2.06686303e-01 4.23172623e-01 -3.19144040e-01 7.19375312e-01
-3.67609680e-01 5.18186569e-01 -1.41277301e+00 -2.90157139e-01
1.36336503e-05 -5.32474890e-02 -7.09400058e-01 4.14949805e-01
-1.55958131e-01 3.10045689e-01 -8.24057758e-01 4.55703110e-01
2.14501828e-01 3.57308649e-02 3.99963796e-01 1.11795557e+00
-3.31314206e-01 4.47597682e-01 -1.28750694e+00 1.00130785e+00
-1.79341331e-01 5.24226546e-01 2.02642456e-01 -9.49589908e-01
7.38518775e-01 4.87485975e-01 5.94668627e-01 -5.30021369e-01
2.56287187e-01 3.79469335e-01 3.59883279e-01 -4.86466318e-01
6.28325582e-01 -4.87775594e-01 6.85783103e-02 1.13226295e+00
-3.70145351e-01 -6.57403562e-03 -1.26270847e-02 3.69012624e-01
1.45812118e+00 2.38225132e-01 6.18098602e-02 -5.37801683e-01
-1.67245269e-01 1.53600916e-01 5.74951291e-01 1.37179637e+00
-6.47336721e-01 -3.26797664e-01 7.17853010e-01 -3.21584582e-01
-1.03762877e+00 -1.15541887e+00 7.90867433e-02 1.08530736e+00
4.50945050e-01 -2.34303594e-01 -8.46582174e-01 -6.70310140e-01
-7.00433878e-03 1.03689814e+00 -7.27517128e-01 -1.76486015e-01
-4.11109149e-01 -5.28590798e-01 4.76774484e-01 3.07182729e-01
5.39370954e-01 -1.38940763e+00 -1.39885473e+00 3.68495673e-01
8.88647437e-02 -2.27775499e-01 -4.17730898e-01 6.87280476e-01
-8.43496919e-01 -1.05398774e+00 -3.80193144e-01 -2.34907061e-01
7.08787560e-01 3.51308882e-01 7.45652795e-01 2.80864298e-01
1.38220638e-01 6.94794178e-01 -3.03756952e-01 -5.59707165e-01
-6.29455686e-01 -1.96899101e-01 5.41896403e-01 -7.10907876e-01
4.02278781e-01 -6.01665258e-01 -6.81570888e-01 4.91579100e-02
-7.18994498e-01 -4.21862781e-01 5.03322542e-01 8.11421156e-01
-3.56027260e-02 5.59419513e-01 8.37923527e-01 -5.73091626e-01
8.99433017e-01 -5.57976425e-01 -8.31466258e-01 -1.24887684e-02
-8.10750902e-01 5.07964194e-01 8.13418925e-01 -5.19831896e-01
-8.79130721e-01 -3.20861131e-01 2.55635768e-01 -6.09179363e-02
-1.23188533e-01 5.30506261e-02 6.14001676e-02 8.17296728e-02
8.63803566e-01 4.68685865e-01 2.27148160e-01 -2.79424489e-01
1.04248077e-01 4.45057958e-01 2.81108081e-01 -1.09594238e+00
8.91787291e-01 1.26573801e-01 8.64950046e-02 -5.89537978e-01
-4.28393781e-01 1.52504772e-01 3.58782113e-02 -3.35560262e-01
2.68579423e-01 -6.86867535e-01 -1.34740567e+00 5.33908196e-02
-5.43656707e-01 -6.90163851e-01 -7.41195142e-01 7.07452953e-01
-1.12398934e+00 2.11524948e-01 -3.16946477e-01 -1.46970117e+00
1.53661057e-01 -1.30038667e+00 3.66130993e-02 6.00105762e-01
-1.50208488e-01 -9.65010345e-01 2.26210490e-01 -9.01791751e-02
6.36767685e-01 1.08396690e-02 4.78894532e-01 -7.73842394e-01
-7.51879990e-01 5.80003336e-02 5.21705210e-01 6.24621883e-02
5.73843531e-02 -8.50080773e-02 -6.07079089e-01 -6.19036019e-01
9.54794437e-02 -7.71743715e-01 3.76279056e-01 3.43207031e-01
7.37705529e-01 -8.36084127e-01 -9.71614346e-02 3.30265701e-01
1.28971946e+00 7.66756952e-01 3.39769125e-01 9.87670779e-01
-5.04641384e-02 5.29212177e-01 8.27787876e-01 9.83482718e-01
4.18593973e-01 -5.61305061e-02 9.34898853e-01 4.37631667e-01
7.07679689e-01 -5.25332451e-01 7.82027960e-01 1.43605292e-01
-3.21773708e-01 -7.99657851e-02 -6.39854193e-01 5.84826767e-01
-1.79803395e+00 -1.31184757e+00 5.37545919e-01 2.65396571e+00
9.54382062e-01 3.78927231e-01 5.61843753e-01 -2.54790306e-01
4.27312374e-01 -1.31622657e-01 -1.02973032e+00 -8.49775732e-01
1.24564916e-01 -2.24934861e-01 8.79190743e-01 7.45452583e-01
-6.01199865e-01 7.32037127e-01 7.53380919e+00 3.07840705e-01
-7.09360301e-01 -4.94427793e-02 4.29814845e-01 -5.42043269e-01
-5.32984495e-01 4.84658629e-01 -6.34478331e-01 3.69530499e-01
1.26916337e+00 -6.13267064e-01 9.86819685e-01 1.15145540e+00
4.05352414e-01 -5.29725492e-01 -1.15258515e+00 3.02021652e-01
-6.21387959e-01 -8.85416806e-01 -4.39154238e-01 5.69487631e-01
6.52169108e-01 -3.86299081e-02 1.08195096e-01 5.71433663e-01
1.58597064e+00 -1.20474720e+00 8.58454347e-01 1.08525291e-01
1.91775665e-01 -1.34621644e+00 5.38498342e-01 9.62152600e-01
-5.86882591e-01 -4.42211956e-01 -4.86821771e-01 -4.21111494e-01
-2.24038452e-01 -1.92209259e-01 -1.03855169e+00 -1.39057348e-02
4.07292575e-01 -3.21303755e-02 1.02816261e-01 9.35809135e-01
-2.85635769e-01 7.04784930e-01 -4.14343268e-01 -5.87567925e-01
8.61648083e-01 -3.50921065e-01 7.46232867e-01 7.80352533e-01
3.35315406e-01 2.80951321e-01 3.20358515e-01 7.01618671e-01
1.45129427e-01 -6.63309023e-02 -6.67649925e-01 -2.18164906e-01
7.98959315e-01 8.61038029e-01 -6.76003218e-01 -3.81330043e-01
1.75417244e-01 5.38733006e-01 3.05950195e-01 3.62254441e-01
-6.38124287e-01 -4.92795765e-01 7.65868664e-01 -3.91341969e-02
-5.64855859e-02 -2.96272576e-01 -9.80495885e-02 -7.95761347e-01
-2.71698475e-01 -1.04791307e+00 2.54417717e-01 -3.69440883e-01
-6.26212180e-01 2.79904425e-01 -3.09438426e-02 -8.96008909e-01
-5.92291236e-01 -2.28867859e-01 -7.53303826e-01 6.46937370e-01
-1.35725451e+00 -1.13003045e-01 4.58037704e-01 4.41540837e-01
5.13850033e-01 -2.51938373e-01 6.38562620e-01 -6.43078327e-01
-3.05583090e-01 3.67915809e-01 4.59950358e-01 -2.62680918e-01
3.40964466e-01 -1.71198320e+00 -6.37471601e-02 7.19315350e-01
-2.10306317e-01 7.28995681e-01 1.25362873e+00 -6.14774346e-01
-1.80780625e+00 -5.46001673e-01 3.31211835e-01 -3.61121446e-01
8.42684269e-01 1.89581826e-01 -6.81388021e-01 8.76804888e-01
3.48445028e-01 -4.71768290e-01 3.01998198e-01 3.08029860e-01
3.64906713e-02 1.83210433e-01 -1.17255354e+00 1.14304972e+00
8.86321068e-01 -1.88241422e-01 -6.81240141e-01 3.81489336e-01
5.50076604e-01 -1.99377000e-01 -5.38511515e-01 -1.73891366e-01
4.90015656e-01 -9.59086597e-01 4.96780962e-01 -9.01487291e-01
1.65033475e-01 -1.45533219e-01 1.76839298e-03 -1.65034676e+00
-2.70483971e-01 -1.29969382e+00 2.31973771e-02 4.96787935e-01
6.96522444e-02 -7.32725024e-01 1.03346384e+00 7.27878392e-01
1.67880625e-01 -5.53533494e-01 -9.38962579e-01 -1.27308106e+00
5.95760226e-01 -1.60407051e-01 5.75809598e-01 6.42822802e-01
6.68756247e-01 -2.29280159e-01 -5.62345028e-01 1.38344407e-01
1.04201758e+00 -1.06982633e-01 6.97012007e-01 -7.02039421e-01
-5.17587781e-01 -4.92768914e-01 1.13767989e-01 -1.03809071e+00
6.95547611e-02 -3.91874373e-01 5.18043697e-01 -1.22601283e+00
3.55785191e-01 -6.56628072e-01 -6.04698002e-01 4.05568600e-01
-5.76338172e-02 -5.32580554e-01 1.79064170e-01 2.43275344e-01
-8.49877179e-01 5.27814806e-01 1.41226411e+00 2.17931360e-01
-4.30208981e-01 1.34632379e-01 -1.09109509e+00 7.96508789e-01
1.05644631e+00 -6.77381992e-01 -6.79159582e-01 3.45943216e-03
2.40629882e-01 5.84028900e-01 1.46737322e-01 -7.54705191e-01
2.00344130e-01 -9.66634333e-01 1.38354167e-01 -1.22180939e-01
5.73134720e-02 -6.92756772e-01 4.69539985e-02 1.07585371e+00
-8.49273443e-01 2.07109541e-01 -2.70086139e-01 6.09922349e-01
5.05202830e-01 -7.89165556e-01 8.90104711e-01 -3.86619419e-01
-3.02209347e-01 6.26585335e-02 -9.69264209e-01 6.72036484e-02
1.22238421e+00 -5.16377687e-02 -3.89318287e-01 -9.34598386e-01
-5.11985123e-01 5.98232448e-01 8.67856026e-01 -3.87333445e-02
5.11768758e-01 -8.83382976e-01 -6.42868698e-01 -1.02269232e-01
-3.02659333e-01 -3.95902485e-01 -8.62873495e-02 4.48645025e-01
-2.46105641e-01 1.80485785e-01 -3.36919308e-01 -1.73380207e-02
-8.19281518e-01 6.60401523e-01 5.91759980e-01 -1.45131394e-01
-8.80908191e-01 5.16734242e-01 2.28959266e-02 -5.12195602e-02
4.22687888e-01 1.19621314e-01 6.97955713e-02 -4.25868064e-01
7.54610896e-01 3.95547092e-01 -5.29604375e-01 3.54637317e-02
-4.79011834e-02 -2.35505760e-01 -3.47013950e-01 -7.30219007e-01
1.28676724e+00 -1.54992327e-01 3.36497664e-01 3.25246692e-01
5.57501972e-01 2.09352802e-02 -1.92475140e+00 -1.08484566e-01
-4.20993529e-02 -7.62762845e-01 -2.20971908e-02 -5.89010298e-01
-4.57087398e-01 6.35061204e-01 3.86763394e-01 5.44746518e-01
6.25807345e-01 -3.30597162e-01 3.51431847e-01 1.21062481e+00
8.64125013e-01 -1.47851110e+00 3.04348737e-01 4.77426708e-01
4.39491808e-01 -1.09911931e+00 -6.07039686e-03 7.63088524e-01
-9.45139229e-01 9.77096438e-01 5.87213039e-01 -3.56662840e-01
2.33904839e-01 4.74466562e-01 -1.27549335e-01 6.95602736e-03
-1.23188925e+00 -2.02231482e-01 -8.85090113e-01 7.63566375e-01
5.90335727e-02 4.34322476e-01 -5.84089346e-02 -3.48075777e-02
-5.45028448e-01 -1.28012300e-01 1.23201418e+00 1.24981141e+00
-1.47767651e+00 -1.01380002e+00 -6.80735171e-01 2.82333255e-01
-4.64677215e-01 1.66711748e-01 -6.92829490e-03 8.90181661e-01
-3.77644330e-01 1.05712593e+00 1.02907531e-01 -9.61069837e-02
-1.26170740e-01 -1.49421245e-01 4.75939184e-01 -2.69485086e-01
-3.37643623e-01 7.81380162e-02 -2.59863827e-02 -4.88054574e-01
-6.07829243e-02 -9.66089129e-01 -1.58949721e+00 -9.31097031e-01
-4.31928337e-02 6.60891831e-01 5.67200601e-01 1.01407838e+00
-2.35764369e-01 2.44150400e-01 1.04323661e+00 -3.16769600e-01
-1.55790854e+00 -5.51470220e-01 -9.50986385e-01 3.44257466e-02
5.86317778e-01 -5.87029755e-01 -5.48882484e-01 -5.00688553e-01]
|
[4.1863508224487305, 2.310831308364868]
|
7c917a82-9838-40ac-abc5-4ec25edc71b6
|
bridging-the-gap-between-events-and-frames
|
2109.02618
| null |
https://arxiv.org/abs/2109.02618v2
|
https://arxiv.org/pdf/2109.02618v2.pdf
|
Bridging the Gap between Events and Frames through Unsupervised Domain Adaptation
|
Reliable perception during fast motion maneuvers or in high dynamic range environments is crucial for robotic systems. Since event cameras are robust to these challenging conditions, they have great potential to increase the reliability of robot vision. However, event-based vision has been held back by the shortage of labeled datasets due to the novelty of event cameras. To overcome this drawback, we propose a task transfer method to train models directly with labeled images and unlabeled event data. Compared to previous approaches, (i) our method transfers from single images to events instead of high frame rate videos, and (ii) does not rely on paired sensor data. To achieve this, we leverage the generative event model to split event features into content and motion features. This split enables efficient matching between latent spaces for events and images, which is crucial for successful task transfer. Thus, our approach unlocks the vast amount of existing image datasets for the training of event-based neural networks. Our task transfer method consistently outperforms methods targeting Unsupervised Domain Adaptation for object detection by 0.26 mAP (increase by 93%) and classification by 2.7% accuracy.
|
['Davide Scaramuzza', 'Mathias Gehrig', 'Daniel Gehrig', 'Nico Messikommer']
|
2021-09-06
| null | null | null | null |
['event-based-vision']
|
['computer-vision']
|
[ 6.06379330e-01 -7.85188675e-02 -2.00576931e-01 -2.72380888e-01
-8.23032975e-01 -5.41027844e-01 9.56384838e-01 -2.35612512e-01
-6.58037484e-01 6.34738028e-01 1.43999949e-01 1.03216274e-02
4.28793468e-02 -5.59705615e-01 -1.07888138e+00 -7.29207695e-01
2.52447277e-01 1.66988954e-01 6.94645584e-01 1.31283030e-01
2.08334312e-01 4.16314095e-01 -1.87595451e+00 3.11192513e-01
5.93046069e-01 1.04061818e+00 7.38241613e-01 5.72332561e-01
1.18133567e-01 1.06738210e+00 -3.37369651e-01 6.01713993e-02
1.82969302e-01 -4.26659614e-01 -4.81264263e-01 3.64826649e-01
4.91475791e-01 -6.72370374e-01 -6.52427733e-01 8.62505734e-01
2.63957232e-01 3.43081564e-01 6.55738711e-01 -1.45310080e+00
-4.42871988e-01 8.39828774e-02 -4.84986156e-01 1.78740069e-01
3.86538029e-01 3.12112480e-01 6.41788661e-01 -9.14730728e-01
8.75087440e-01 1.03178251e+00 4.69304979e-01 6.96199536e-01
-1.03404200e+00 -5.54595947e-01 6.26836866e-02 6.47388935e-01
-1.00543845e+00 -7.17518806e-01 7.23538816e-01 -6.63731217e-01
1.07781744e+00 -3.65062714e-01 6.95467174e-01 1.64656746e+00
2.35779554e-01 7.94414818e-01 7.68342912e-01 -2.86119252e-01
3.14124763e-01 -7.18345270e-02 -2.87456661e-01 4.22143310e-01
1.59622908e-01 4.64044303e-01 -8.28531563e-01 4.09200668e-01
7.86811352e-01 2.84218758e-01 -2.18703285e-01 -6.83711588e-01
-1.67376626e+00 7.05730319e-01 4.22768384e-01 2.55145114e-02
-5.81259370e-01 3.85251850e-01 3.70322317e-01 1.53709993e-01
1.48103461e-01 1.72326297e-01 -2.33544186e-01 -3.30048054e-01
-6.14220083e-01 -3.09084263e-02 6.42060816e-01 9.48365629e-01
6.62493408e-01 1.87697783e-01 5.51479794e-02 5.33763707e-01
2.77976424e-01 7.94105232e-01 4.57946628e-01 -1.18063772e+00
4.67734277e-01 4.44225132e-01 3.12747002e-01 -8.44471455e-01
-2.97955364e-01 1.30201742e-01 -4.55016404e-01 3.46750677e-01
5.55877328e-01 7.81039372e-02 -1.01066971e+00 1.71431291e+00
2.26560965e-01 1.58588141e-01 2.29230598e-01 9.06776905e-01
4.55474138e-01 7.72671402e-01 1.57795459e-01 -1.94474488e-01
1.16892302e+00 -7.72187769e-01 -6.60661101e-01 -6.09496593e-01
2.32871532e-01 -6.97439611e-01 9.46521759e-01 2.78059989e-01
-7.87798584e-01 -6.35291517e-01 -1.09160149e+00 4.18183766e-02
-2.70120442e-01 1.11926626e-02 4.94231075e-01 1.74579963e-01
-7.95602679e-01 1.66120455e-01 -1.19624686e+00 -8.29678893e-01
3.71763408e-01 3.40503126e-01 -6.47720516e-01 -3.18805039e-01
-8.18945229e-01 1.02820623e+00 7.03684688e-01 -1.67849094e-01
-1.13077044e+00 -3.90608639e-01 -1.13456750e+00 -1.62832588e-01
4.79801118e-01 -5.93347967e-01 1.34084702e+00 -9.63251889e-01
-1.61414647e+00 6.98358238e-01 -1.14906825e-01 -6.92441285e-01
4.21797961e-01 -3.32672626e-01 -4.06773277e-02 4.53972012e-01
1.48051873e-01 1.13392794e+00 1.00684178e+00 -1.16314542e+00
-9.76648688e-01 -3.93244438e-02 -1.19373910e-01 2.64134198e-01
-4.48371351e-01 -2.91194469e-02 -4.61739510e-01 -2.65318930e-01
-1.06045175e-02 -9.83108342e-01 -3.51878628e-02 1.11044444e-01
2.12440357e-01 -9.86840203e-02 1.12439501e+00 -5.31055272e-01
3.43701571e-01 -2.30290389e+00 1.21108837e-01 -3.68965209e-01
-2.85574608e-02 1.74976319e-01 -7.52301738e-02 4.39743370e-01
2.20450610e-01 -6.82539880e-01 -1.34675682e-01 -3.16752881e-01
-1.32570654e-01 4.48713779e-01 -5.48480093e-01 5.23429513e-01
4.64298725e-01 9.92779255e-01 -1.07498348e+00 -6.18600726e-01
8.58429074e-01 4.72297728e-01 -3.23126078e-01 3.54904503e-01
-3.24455202e-01 8.93406689e-01 -2.81930536e-01 5.95888138e-01
2.52451926e-01 -3.32807213e-01 6.01971224e-02 -3.32522720e-01
-2.69642234e-01 1.52007714e-01 -9.03192163e-01 1.95626533e+00
-3.20725709e-01 8.42279553e-01 -1.18701465e-01 -1.13162494e+00
8.67432415e-01 4.64894444e-01 7.27823317e-01 -8.17897737e-01
1.96273401e-01 4.80470583e-02 -3.50482911e-01 -7.20180392e-01
5.34261644e-01 -4.44831662e-02 -1.99387938e-01 1.20064102e-01
3.20198864e-01 -1.72646120e-01 2.28023186e-01 6.99358732e-02
1.16463721e+00 6.10611558e-01 3.04755986e-01 3.33330154e-01
5.86798191e-02 2.41954342e-01 6.81586921e-01 7.14618802e-01
-4.69136685e-01 7.23443031e-01 -1.13791995e-01 -3.71824771e-01
-1.19326234e+00 -1.33736670e+00 9.30275843e-02 9.85853314e-01
5.04512429e-01 5.98361269e-02 -2.59905338e-01 -5.59336722e-01
-1.31213233e-01 6.21216118e-01 -4.33100164e-01 -4.30688649e-01
-6.24560058e-01 -4.12237018e-01 2.81643689e-01 9.17608142e-01
6.43065572e-01 -1.34004664e+00 -1.08040214e+00 3.19662243e-01
-5.30782878e-01 -1.66499913e+00 -2.85847127e-01 2.28740245e-01
-6.44670784e-01 -1.11125922e+00 -4.79848325e-01 -7.72244036e-01
4.87455845e-01 7.03880727e-01 8.74074817e-01 -7.55344570e-01
-3.43518257e-01 7.85927117e-01 -6.05063498e-01 -4.97977674e-01
-3.20769370e-01 -2.16872171e-01 1.58779621e-01 2.18402863e-01
3.87012482e-01 -4.89731163e-01 -5.76031804e-01 3.75782549e-01
-1.03655243e+00 1.73235774e-01 8.40322077e-01 9.01094139e-01
4.74091351e-01 -3.18860799e-01 5.81560433e-01 -1.29971787e-01
-1.38321355e-01 -5.65053642e-01 -7.93211639e-01 7.46826008e-02
-3.72344077e-01 -1.56342536e-01 3.08609635e-01 -7.38883138e-01
-1.37134016e+00 6.97915435e-01 3.83857161e-01 -6.63148582e-01
-3.50069910e-01 2.60929555e-01 6.50792643e-02 -1.00379754e-02
7.18953788e-01 2.02507243e-01 1.38634399e-01 5.57600800e-03
5.02700031e-01 6.05242550e-01 9.43291187e-01 -2.86892474e-01
6.81449413e-01 8.31118703e-01 -5.53083159e-02 -8.08017552e-01
-6.76342309e-01 -6.82768345e-01 -6.07278407e-01 -5.01801610e-01
1.21985710e+00 -1.27771866e+00 -5.17069638e-01 4.66043264e-01
-1.18656635e+00 -4.25534874e-01 -3.95608455e-01 1.03823352e+00
-8.00534785e-01 3.35392952e-01 -5.26535094e-01 -9.26904321e-01
1.17385402e-01 -9.15820420e-01 1.25076854e+00 2.30622292e-01
-5.27694039e-02 -6.72712803e-01 -1.20974243e-01 3.98394614e-01
1.93284675e-01 3.99365366e-01 2.71496683e-01 -3.26244980e-01
-8.72432709e-01 -3.11601222e-01 -4.33484048e-01 3.19621235e-01
1.93982974e-01 -2.58176714e-01 -9.31673408e-01 -3.01755935e-01
1.61491662e-01 -6.82071090e-01 9.48545456e-01 3.74943674e-01
6.48071587e-01 2.65666455e-01 -4.17917013e-01 3.96886557e-01
1.17371011e+00 4.37110633e-01 6.61023557e-01 4.19370979e-01
6.61713898e-01 6.60704255e-01 8.77325356e-01 4.69863713e-01
5.50331891e-01 8.07823598e-01 6.20161593e-01 1.07837610e-01
-3.51639688e-01 -2.98706263e-01 8.97733152e-01 5.88619351e-01
1.10955596e-01 -2.25883588e-01 -9.23295856e-01 8.94721150e-01
-2.17329979e+00 -1.04857874e+00 1.40262067e-01 2.20265532e+00
5.84704697e-01 6.11445121e-02 -3.93683314e-02 -8.44299942e-02
7.37272918e-01 -7.11892353e-05 -7.40741909e-01 3.69029671e-01
-9.06170085e-02 -3.42857391e-01 5.13879299e-01 1.51703000e-01
-1.21137512e+00 8.91146362e-01 5.77776051e+00 3.82552207e-01
-1.12929142e+00 2.76653934e-02 -6.42310232e-02 -1.93347827e-01
4.15324867e-01 -7.23507553e-02 -8.50822568e-01 4.31411594e-01
1.02469170e+00 3.79001871e-02 2.65158087e-01 8.85625839e-01
1.87893704e-01 -2.69791633e-01 -1.32133925e+00 1.10371506e+00
3.21901768e-01 -1.05404198e+00 -9.50543210e-02 2.55621076e-02
7.38248825e-01 2.33265907e-01 -6.29185438e-02 2.90439188e-01
3.69646013e-01 -6.87487602e-01 8.91783118e-01 4.92893696e-01
6.52060807e-01 -4.05476958e-01 5.60190797e-01 3.63108695e-01
-1.11688721e+00 -2.03124732e-01 -3.18062216e-01 -5.55651262e-02
4.33931082e-01 3.74789178e-01 -1.11125433e+00 4.63025123e-01
8.88384759e-01 1.04078555e+00 -2.39038423e-01 9.60306764e-01
-1.21170804e-01 5.17991424e-01 -4.61767614e-01 2.95563996e-01
3.94326150e-02 1.51270973e-02 5.48438966e-01 9.93474603e-01
4.19906527e-01 -2.73472935e-01 4.16441739e-01 6.11869156e-01
1.04705811e-01 -4.83682901e-01 -1.02057326e+00 -7.87608605e-03
4.63825166e-01 1.07721293e+00 -6.70698285e-01 -2.54522175e-01
-6.59113705e-01 1.24034441e+00 1.95342883e-01 3.43604982e-01
-9.45642531e-01 -1.72461703e-01 2.66202599e-01 -2.49197453e-01
6.24376118e-01 -7.20483243e-01 2.88701952e-02 -1.20496368e+00
1.49394006e-01 -4.51411963e-01 2.83270329e-01 -9.60840762e-01
-1.23908067e+00 3.28841835e-01 1.68022975e-01 -1.56793809e+00
-6.06008947e-01 -7.35106468e-01 -2.36179501e-01 3.19189072e-01
-1.57249069e+00 -1.52373993e+00 -7.21110344e-01 8.25417519e-01
8.63743782e-01 -1.02264114e-01 5.84084332e-01 2.72737801e-01
-2.79115409e-01 9.88739580e-02 1.08778849e-01 -1.49164433e-02
1.10623014e+00 -1.02751946e+00 2.14253768e-01 1.00186539e+00
1.58759221e-01 2.25931615e-01 5.89202225e-01 -6.13320887e-01
-1.63270104e+00 -1.31139183e+00 6.44608080e-01 -7.24249899e-01
6.54185236e-01 -3.82496536e-01 -7.22517550e-01 7.58606493e-01
1.20090961e-01 1.37924820e-01 4.00423795e-01 -2.76036322e-01
-3.80881727e-01 -2.81961113e-01 -7.93571353e-01 5.41624486e-01
9.83501554e-01 -4.88319665e-01 -8.47387195e-01 2.37864017e-01
7.18610227e-01 -2.23418102e-01 -6.36031806e-01 6.03102267e-01
5.37332356e-01 -7.68740296e-01 1.01220119e+00 -2.30166063e-01
3.43175858e-01 -4.86286402e-01 -4.11271840e-01 -9.79651153e-01
-1.91540867e-01 -3.48680973e-01 -4.08775598e-01 1.14210963e+00
5.84580339e-02 -5.26779771e-01 6.31840885e-01 3.89465958e-01
-2.92744309e-01 2.09699012e-02 -9.88105237e-01 -1.06944633e+00
-4.13493454e-01 -4.75120276e-01 1.18774725e-02 7.24484146e-01
-1.89487472e-01 3.75063509e-01 -6.19456649e-01 1.51438475e-01
7.32159913e-01 1.51141420e-01 8.36995661e-01 -1.07348752e+00
-2.70999461e-01 -1.45093009e-01 -5.51989853e-01 -1.05211425e+00
1.77480429e-01 -5.65279782e-01 5.81169009e-01 -1.53861701e+00
2.73696393e-01 -7.08801895e-02 -2.76174486e-01 5.65008759e-01
-9.67111588e-02 4.72031534e-01 1.78703815e-01 5.56252480e-01
-9.58913624e-01 7.88716495e-01 8.53837967e-01 1.12940133e-01
-1.05591245e-01 -4.04928237e-01 -1.57140762e-01 6.66567922e-01
7.25308776e-01 -2.91147441e-01 -5.08577168e-01 -5.05558789e-01
5.15388604e-03 1.54979318e-01 8.06084096e-01 -1.32663977e+00
4.53639805e-01 -3.14724237e-01 4.32102948e-01 -5.54025769e-01
6.42669141e-01 -1.05375838e+00 8.91144872e-02 3.98656905e-01
-5.18667735e-02 -2.18188182e-01 1.57196864e-01 1.05325472e+00
-2.83824414e-01 1.22326598e-01 7.44284213e-01 1.91143788e-02
-1.35482299e+00 9.73005295e-02 -6.01504683e-01 -1.60549149e-01
1.26759779e+00 -4.99277651e-01 -3.42979580e-01 -4.33377534e-01
-3.53044599e-01 1.19434290e-01 4.98141289e-01 6.59643471e-01
6.73517644e-01 -1.28446198e+00 -4.82592046e-01 1.51804850e-01
5.61293960e-01 1.40955463e-01 1.13697127e-01 9.07089114e-01
-1.54248312e-01 4.35396731e-01 -4.70622122e-01 -1.02545655e+00
-1.04238653e+00 5.68227172e-01 -8.76834393e-02 1.64416894e-01
-8.31912160e-01 4.92048800e-01 4.01257664e-01 -2.15437919e-01
2.46849775e-01 -8.91873240e-02 -4.22692224e-02 -2.03235541e-02
5.17610908e-01 1.80095360e-01 -1.44513682e-01 -5.43358564e-01
-2.91493952e-01 4.64010298e-01 -2.82537136e-02 -3.62908453e-01
1.18512702e+00 -3.20979774e-01 3.25167984e-01 6.83153629e-01
9.76044118e-01 -3.74377966e-01 -2.13441467e+00 -3.96525681e-01
-2.02559829e-02 -3.07716221e-01 -1.11657545e-01 -6.37942731e-01
-6.50362313e-01 8.18666875e-01 6.54574454e-01 -8.99650231e-02
1.17166352e+00 1.41536027e-01 8.15744936e-01 5.76822639e-01
6.81428075e-01 -9.98734355e-01 6.11822128e-01 7.12209105e-01
6.84853077e-01 -1.64746702e+00 -2.92229414e-01 -2.36985341e-01
-7.25161076e-01 8.80591869e-01 5.98939538e-01 -1.40466362e-01
1.81670398e-01 8.13697651e-02 2.19620895e-02 4.71183583e-02
-8.32093716e-01 -4.97088164e-01 1.18731931e-01 9.17877078e-01
-9.78655964e-02 -2.83858597e-01 4.71508168e-02 1.61151141e-01
2.54391879e-01 2.20219404e-01 3.36566120e-01 1.22202742e+00
-6.20712161e-01 -8.50353360e-01 -3.49525303e-01 1.40891254e-01
-1.06511965e-01 2.31943533e-01 -1.10211186e-01 7.75660396e-01
4.55186106e-02 1.17578816e+00 1.47648692e-01 -4.96087939e-01
2.67017812e-01 1.47527963e-01 6.83548510e-01 -5.20338595e-01
-1.81411821e-02 -2.66618934e-02 -6.89246207e-02 -7.39658058e-01
-7.01917768e-01 -9.46002424e-01 -1.15781760e+00 1.76394418e-01
-1.72650129e-01 -3.38874876e-01 8.83779466e-01 1.00391018e+00
5.64806819e-01 5.31307340e-01 4.89250392e-01 -1.28068483e+00
-5.00330150e-01 -8.91662478e-01 -2.54958421e-01 6.48824036e-01
3.00430864e-01 -1.02417278e+00 -2.63480723e-01 8.21355999e-01]
|
[8.436281204223633, -0.8754507899284363]
|
cb522fbe-98ca-497b-a3a0-9ee553976ba4
|
extending-monocular-visual-odometry-to-stereo
|
1905.12723
| null |
https://arxiv.org/abs/1905.12723v3
|
https://arxiv.org/pdf/1905.12723v3.pdf
|
Extending Monocular Visual Odometry to Stereo Camera Systems by Scale Optimization
|
This paper proposes a novel approach for extending monocular visual odometry to a stereo camera system. The proposed method uses an additional camera to accurately estimate and optimize the scale of the monocular visual odometry, rather than triangulating 3D points from stereo matching. Specifically, the 3D points generated by the monocular visual odometry are projected onto the other camera of the stereo pair, and the scale is recovered and optimized by directly minimizing the photometric error. It is computationally efficient, adding minimal overhead to the stereo vision system compared to straightforward stereo matching, and is robust to repetitive texture. Additionally, direct scale optimization enables stereo visual odometry to be purely based on the direct method. Extensive evaluation on public datasets (e.g., KITTI), and outdoor environments (both terrestrial and underwater) demonstrates the accuracy and efficiency of a stereo visual odometry approach extended by scale optimization, and its robustness in environments with challenging textures.
|
['Junaed Sattar', 'Jiawei Mo']
|
2019-05-29
| null | null | null | null |
['stereo-matching', 'monocular-visual-odometry']
|
['computer-vision', 'robots']
|
[ 2.45367941e-02 7.62954578e-02 2.72256106e-01 -3.49171281e-01
-2.36559808e-01 -7.22069621e-01 4.44703609e-01 -2.75203109e-01
-3.45839292e-01 4.64795172e-01 1.45124933e-02 3.22692841e-02
3.51460636e-01 -6.88169599e-01 -7.74917603e-01 -5.34719646e-01
2.43309081e-01 9.09672499e-01 5.90333819e-01 -8.61396194e-02
6.57245040e-01 6.01025522e-01 -1.69071925e+00 -7.21015334e-01
8.09775174e-01 9.61736143e-01 7.36484587e-01 8.03660214e-01
-7.31726410e-03 4.08515453e-01 -2.21930355e-01 -2.33671933e-01
7.79807091e-01 2.87210885e-02 -6.16920710e-01 6.59170568e-01
1.08627307e+00 -7.33144641e-01 -3.15638512e-01 1.28967702e+00
5.69480062e-01 -5.62616624e-02 2.69913077e-01 -7.06156611e-01
-3.55458617e-01 -5.08397102e-01 -5.66630483e-01 -4.04321074e-01
9.04245436e-01 -4.60643396e-02 7.80316532e-01 -9.03830409e-01
9.03409362e-01 1.38356471e+00 9.11674559e-01 9.71786454e-02
-1.36363029e+00 -4.08571839e-01 -3.20020795e-01 -7.92073160e-02
-1.48247945e+00 -5.50429285e-01 4.39737111e-01 -6.29253030e-01
7.98284471e-01 -1.29985675e-01 1.00615096e+00 2.80054241e-01
2.64714748e-01 6.02939986e-02 1.28840458e+00 -5.41607499e-01
2.89745450e-01 1.47622256e-02 -2.88165957e-01 8.31879616e-01
6.38903975e-01 3.84785235e-01 -7.12874174e-01 -2.15215981e-01
1.03674376e+00 2.48942986e-01 -2.16002479e-01 -1.17998302e+00
-9.39025581e-01 5.58638275e-01 3.87522548e-01 -5.84245086e-01
-7.36508369e-02 2.18874626e-02 4.50582616e-02 3.90620828e-01
2.65845776e-01 3.14561665e-01 -7.84967989e-02 -1.93590447e-01
-6.68673158e-01 3.02482080e-02 9.20431256e-01 1.33270621e+00
1.51087987e+00 1.61766902e-01 8.23180199e-01 8.84310782e-01
7.09723473e-01 1.21913004e+00 2.61542410e-01 -1.34755313e+00
4.67360973e-01 6.61385417e-01 2.26626217e-01 -1.00940037e+00
-3.29575509e-01 2.96852589e-01 -3.92450690e-01 6.63668513e-01
2.01336801e-01 -9.50617194e-02 -9.92533445e-01 9.78488743e-01
6.40851259e-01 4.61760312e-02 2.65315861e-01 1.06515145e+00
7.54374385e-01 1.83502272e-01 -8.64129066e-01 2.22852156e-01
1.12911081e+00 -6.28871977e-01 -4.69633192e-01 -7.12552190e-01
3.71829003e-01 -1.16896164e+00 5.71172535e-01 1.08730718e-01
-1.00606561e+00 -3.88529301e-01 -1.53642178e+00 -4.07515138e-01
1.27878115e-01 -2.61465251e-01 2.15913251e-01 5.91202557e-01
-1.41425371e+00 3.99262160e-01 -7.78183460e-01 -7.20191717e-01
-4.36274737e-01 5.13427973e-01 -3.59339118e-01 -5.36381640e-02
-8.46331358e-01 1.17227077e+00 2.41725713e-01 -1.38115302e-01
-7.18057692e-01 -1.61296114e-01 -1.31124806e+00 -4.44026053e-01
-6.52617514e-02 -7.37395585e-01 1.11078060e+00 -6.80517912e-01
-1.87756181e+00 1.35540831e+00 -6.00886405e-01 -3.02915573e-01
5.72307408e-01 -2.55466670e-01 5.19506773e-03 3.61602753e-01
3.78350914e-01 7.79834151e-01 7.02202797e-01 -1.38930118e+00
-7.00033307e-01 -5.78908622e-01 1.77164078e-01 8.10194314e-01
1.91107288e-01 -4.32094425e-01 -7.33810425e-01 2.48644818e-02
1.19485998e+00 -1.32226825e+00 -1.07954577e-01 4.01381075e-01
-9.86041650e-02 7.98079729e-01 7.22739279e-01 -5.96368492e-01
5.35318315e-01 -2.16172552e+00 6.37917295e-02 1.96211874e-01
4.58665825e-02 -1.52123883e-01 3.31548482e-01 3.83974522e-01
4.12432313e-01 -6.51168466e-01 1.32777318e-01 -6.10522270e-01
-1.70023412e-01 6.36525512e-01 -1.46434575e-01 9.59246874e-01
-3.85843873e-01 2.65771151e-01 -7.33576357e-01 -4.85136330e-01
5.89917064e-01 2.00920269e-01 -5.33060491e-01 1.92652717e-01
3.72364014e-01 5.53602040e-01 5.18648587e-02 8.84424806e-01
1.10823846e+00 4.31355760e-02 3.71469736e-01 -8.38963762e-02
-6.00396395e-01 5.37346184e-01 -1.75891435e+00 1.53079534e+00
-4.87978429e-01 6.34246886e-01 2.58754015e-01 -2.44989321e-01
1.58996272e+00 6.17051870e-03 2.06124291e-01 -7.50451207e-01
-8.56920183e-02 6.10037684e-01 -6.38119578e-01 -3.32871020e-01
9.89140153e-01 -3.33752692e-01 1.32309631e-01 1.26841739e-01
-9.03243646e-02 -1.00879192e+00 3.64225581e-02 -1.43113211e-01
4.10918057e-01 4.54826266e-01 6.95501864e-01 -4.77595717e-01
5.00086844e-01 2.77716339e-01 5.49743414e-01 4.30792779e-01
-7.87326396e-02 7.73112297e-01 -1.78650588e-01 -6.27948821e-01
-1.46848333e+00 -1.19802618e+00 -4.44594920e-01 2.72177845e-01
1.15528572e+00 8.53419900e-02 -3.44124585e-01 1.90547049e-01
4.59856242e-01 -3.27695608e-01 -2.35899657e-01 2.67687589e-01
-4.52474713e-01 -2.31622905e-01 3.18112105e-01 3.50979000e-01
7.46589482e-01 -2.54088730e-01 -5.18434465e-01 2.35160589e-01
-1.60821453e-01 -1.17529058e+00 -3.82531017e-01 5.26574403e-02
-1.41485429e+00 -1.04601908e+00 -6.57519042e-01 -9.56855953e-01
8.01017582e-01 1.00560582e+00 9.26556706e-01 -1.16750292e-01
2.17098482e-02 1.92007795e-01 -8.85106772e-02 -2.41456762e-01
-1.61090344e-01 -3.05791438e-01 3.50535214e-01 -1.44400373e-01
4.54657614e-01 -5.34237742e-01 -5.58833838e-01 8.53632808e-01
-2.91430831e-01 -6.17977278e-03 1.18768044e-01 8.41297388e-01
6.33632004e-01 -7.16936290e-01 -4.85112101e-01 -2.51974523e-01
-2.40690127e-01 -5.00736907e-02 -1.33815503e+00 -3.36211741e-01
-5.56918263e-01 1.12728618e-01 1.04534850e-01 5.07941060e-02
-1.07069278e+00 5.50566196e-01 1.81222185e-01 -3.52397263e-01
1.52231589e-01 1.59287483e-01 1.64318860e-01 -7.98051894e-01
7.65220881e-01 2.90743887e-01 5.08886218e-01 -5.13893306e-01
-1.43312424e-01 8.24487746e-01 6.45124912e-01 -2.61901021e-01
1.26011646e+00 1.40766013e+00 3.71609867e-01 -1.08745623e+00
-5.39165437e-01 -1.00862157e+00 -1.18388605e+00 -2.14268118e-01
6.91284478e-01 -1.62639153e+00 -5.70881844e-01 6.51030421e-01
-1.08230650e+00 2.43382994e-02 7.17958137e-02 7.91400075e-01
-6.87525094e-01 9.56942797e-01 -3.74259144e-01 -6.15162492e-01
-8.01350251e-02 -1.35300100e+00 1.53137732e+00 2.52710521e-01
-2.59988755e-01 -1.32281435e+00 5.68959892e-01 3.35864961e-01
-1.11891948e-01 -2.15107277e-02 -3.01755723e-02 4.10052359e-01
-9.94111180e-01 -2.14005619e-01 -3.92677546e-01 -1.71759110e-02
2.12074444e-01 2.49271444e-03 -1.06157553e+00 -5.45262575e-01
-3.09127904e-02 -1.29582956e-01 3.15541625e-01 3.10362339e-01
-2.53566712e-01 1.24586783e-01 -4.23481725e-02 1.21121335e+00
1.79172397e+00 1.64862692e-01 7.66004384e-01 9.94254112e-01
7.32283473e-01 5.68517387e-01 9.40402389e-01 5.76038182e-01
7.92591155e-01 1.17212856e+00 6.71089113e-01 -2.30892286e-01
9.73420218e-02 -2.81393379e-01 4.88219291e-01 8.52572680e-01
-2.45054632e-01 6.96119666e-01 -7.65173078e-01 5.72304368e-01
-1.52525675e+00 -7.87188649e-01 -4.43617076e-01 2.55732036e+00
5.49283803e-01 -2.66549528e-01 -1.95197567e-01 -6.89850450e-02
8.35096717e-01 2.20795706e-01 -3.89791280e-01 -6.26245081e-01
-2.67480314e-01 -2.85510033e-01 1.25247717e+00 1.03108001e+00
-8.52538109e-01 9.29891944e-01 6.80586863e+00 -2.08805010e-01
-1.20879555e+00 -1.34210572e-01 -5.19930661e-01 1.64471775e-01
-1.22984827e-01 4.42641467e-01 -1.31034923e+00 2.65731484e-01
2.77691156e-01 -5.37656024e-02 6.32196128e-01 1.02434242e+00
2.54920095e-01 -3.40394735e-01 -7.44706333e-01 1.15992796e+00
2.80457139e-01 -1.03173482e+00 -1.92137599e-01 3.87787759e-01
1.09208012e+00 4.72948402e-01 -2.94885933e-01 -3.33902538e-01
6.44524276e-01 -3.04265916e-01 1.08228433e+00 1.69565782e-01
8.90586674e-01 -2.72077769e-01 7.41373599e-01 2.24318773e-01
-1.57094073e+00 7.93835893e-02 -9.02037084e-01 -5.88795364e-01
3.38134170e-01 4.43321645e-01 -1.06589174e+00 4.49779749e-01
1.10614431e+00 9.24562871e-01 -3.67435783e-01 1.13642383e+00
-2.84409165e-01 -3.65597636e-01 -4.51156020e-01 2.32390791e-01
1.86615765e-01 -5.91817379e-01 6.82083786e-01 8.11768472e-01
6.70665026e-01 -3.69295329e-01 2.51406491e-01 2.74122983e-01
2.58504570e-01 1.33107370e-02 -1.05138385e+00 7.11073697e-01
6.43429756e-01 8.76254380e-01 -1.58644333e-01 -4.88695800e-01
-5.25364459e-01 9.86647069e-01 1.17767520e-01 6.31562471e-02
-3.41643631e-01 -3.41803193e-01 9.84493494e-01 8.33113492e-02
3.17793906e-01 -5.52804589e-01 -6.86033666e-01 -1.26820314e+00
1.91174582e-01 -4.61846620e-01 -4.72271677e-06 -1.00862348e+00
-8.14265072e-01 2.70661652e-01 -2.52679467e-01 -2.06233907e+00
-2.73261160e-01 -7.04908729e-01 -3.03449035e-01 1.10371947e+00
-1.69111407e+00 -6.49400890e-01 -8.97205770e-01 5.26857615e-01
2.85321087e-01 2.10751906e-01 5.43575883e-01 1.63299099e-01
2.17098534e-01 1.52924443e-02 5.29450595e-01 -2.66455948e-01
9.62536931e-01 -1.42079890e+00 6.23438060e-01 8.47900391e-01
-4.22266036e-01 6.17940843e-01 8.68774176e-01 -7.94341207e-01
-1.59078133e+00 -7.41035938e-01 1.09843409e+00 -4.31192070e-01
7.25246847e-01 -1.31227851e-01 -4.83585268e-01 7.23526001e-01
-8.71807560e-02 -3.37948561e-01 5.61715588e-02 -2.47701034e-01
-1.96683615e-01 -2.05333784e-01 -1.21613979e+00 6.66551054e-01
1.12530124e+00 -7.88098991e-01 -7.51417220e-01 1.84639581e-02
2.58574843e-01 -1.19035745e+00 -9.10739243e-01 1.99318051e-01
1.04419577e+00 -1.16255629e+00 1.19364738e+00 4.54971462e-01
6.15850054e-02 -8.06712449e-01 -4.99584049e-01 -1.13281965e+00
-1.04011312e-01 -6.15937829e-01 3.89760196e-01 7.81599998e-01
-8.80482942e-02 -9.89814937e-01 8.73194754e-01 1.89710408e-01
-3.38520050e-01 2.68983513e-01 -1.07150769e+00 -9.76698875e-01
-6.39033675e-01 5.42457849e-02 5.57729840e-01 7.98328817e-01
6.28848048e-03 3.47373337e-02 -6.84373677e-01 8.91544402e-01
1.04222631e+00 2.58550048e-01 1.45910156e+00 -1.52317226e+00
-3.69625896e-01 3.18811953e-01 -1.06031656e+00 -1.83520806e+00
-2.28077546e-01 -4.97589588e-01 3.19146574e-01 -1.35369372e+00
9.12392396e-04 -1.85007378e-01 6.69843853e-01 -2.84879673e-02
1.93883285e-01 6.28478289e-01 2.85626836e-02 7.74821579e-01
-2.81385891e-02 2.10415453e-01 1.43486774e+00 1.12925425e-01
-2.27097467e-01 -2.48243675e-01 -7.04542100e-02 1.04418910e+00
3.71291757e-01 -4.43146050e-01 -7.52842352e-02 -8.36260021e-01
3.06913674e-01 3.19252282e-01 3.75144064e-01 -1.09028578e+00
2.93458909e-01 -2.36525878e-01 8.71912688e-02 -6.73308909e-01
6.11863256e-01 -1.00345826e+00 4.28318769e-01 7.10381210e-01
4.55967933e-01 1.67509183e-01 -1.23716429e-01 4.35275048e-01
-3.78987730e-01 -4.68540847e-01 1.17289913e+00 -1.22486256e-01
-1.07281792e+00 1.91909950e-02 -2.27773547e-01 -2.78559387e-01
7.07990587e-01 -1.14717245e+00 -3.52671206e-01 -3.88953120e-01
-5.97540855e-01 1.05267607e-01 1.56291747e+00 1.29048109e-01
6.52689576e-01 -1.33702517e+00 -1.79047793e-01 6.32184565e-01
4.81358588e-01 1.85823768e-01 -4.78770118e-03 8.30386639e-01
-1.54879344e+00 3.35872024e-01 -4.28852409e-01 -1.29644382e+00
-1.41486537e+00 8.04663897e-02 5.94945788e-01 2.82622546e-01
-7.69751608e-01 4.18547660e-01 1.96702227e-01 -8.91131520e-01
-4.05440247e-03 -2.81114280e-01 1.23226494e-01 -2.39964277e-01
3.46450001e-01 6.94277346e-01 1.09718263e-01 -9.94896650e-01
-3.09872210e-01 1.45998371e+00 3.89978260e-01 -4.33603585e-01
9.20083880e-01 -1.03488910e+00 -7.73959979e-02 2.71616697e-01
1.14991677e+00 9.92047712e-02 -1.65278792e+00 -3.45957160e-01
-2.06376195e-01 -9.47884560e-01 -1.19323947e-01 9.06140953e-02
-6.50613666e-01 8.36078048e-01 6.61054432e-01 -1.06036022e-01
6.69092119e-01 -3.00662458e-01 5.92324317e-01 6.11080885e-01
8.97841334e-01 -1.19059741e+00 -1.26607195e-01 1.02089322e+00
5.83129585e-01 -1.30781603e+00 2.92499214e-01 -5.76523781e-01
-5.29727459e-01 1.34002626e+00 3.45429122e-01 -2.55301565e-01
4.24257398e-01 3.38093966e-01 6.91181481e-01 -1.61810949e-01
-1.75063193e-01 -2.09415674e-01 -9.78654325e-02 6.63420141e-01
1.32973045e-02 -2.76475310e-01 6.42617121e-02 -9.57665622e-01
-2.54048884e-01 -8.51370394e-02 8.26386571e-01 1.24796641e+00
-1.01807773e+00 -8.59250307e-01 -9.82265770e-01 -2.39512876e-01
2.89754774e-02 3.30107808e-02 -3.16424161e-01 5.53013146e-01
-1.68299079e-01 7.61011839e-01 2.56508857e-01 -3.96861821e-01
4.30254996e-01 -4.28558946e-01 3.55416566e-01 -6.03436887e-01
-5.03725000e-02 1.21289439e-01 2.58776695e-01 -5.93662679e-01
-5.72215140e-01 -8.18524480e-01 -1.19623256e+00 -4.72135514e-01
-3.19156349e-01 -8.57511461e-02 1.00445211e+00 6.98741198e-01
1.49465710e-01 -5.19892991e-01 1.02015865e+00 -1.44420815e+00
-5.74213862e-01 -6.01045489e-01 -8.61574233e-01 3.48013371e-01
4.97129679e-01 -9.74742651e-01 -7.13967085e-01 2.22716630e-01]
|
[7.688779354095459, -2.241863489151001]
|
5df53652-6242-4723-ba0a-7e87d4e3f5e7
|
fast-model-based-policy-search-for-universal
|
2202.05843
| null |
https://arxiv.org/abs/2202.05843v1
|
https://arxiv.org/pdf/2202.05843v1.pdf
|
Fast Model-based Policy Search for Universal Policy Networks
|
Adapting an agent's behaviour to new environments has been one of the primary focus areas of physics based reinforcement learning. Although recent approaches such as universal policy networks partially address this issue by enabling the storage of multiple policies trained in simulation on a wide range of dynamic/latent factors, efficiently identifying the most appropriate policy for a given environment remains a challenge. In this work, we propose a Gaussian Process-based prior learned in simulation, that captures the likely performance of a policy when transferred to a previously unseen environment. We integrate this prior with a Bayesian Optimisation-based policy search process to improve the efficiency of identifying the most appropriate policy from the universal policy network. We empirically evaluate our approach in a range of continuous and discrete control environments, and show that it outperforms other competing baselines.
|
['Svetha Venkatesh', 'Santu Rana', 'Thommen George Karimpanal', 'Buddhika Laknath Semage']
|
2022-02-11
| null | null | null | null |
['bayesian-optimisation']
|
['methodology']
|
[-5.00832163e-02 -3.79282832e-01 -2.34842569e-01 8.12206045e-03
-5.75954854e-01 -5.79749763e-01 9.08244252e-01 2.58345634e-01
-8.43556941e-01 1.05550873e+00 1.70881823e-01 -3.12023669e-01
-3.64377737e-01 -6.00516915e-01 -9.78914857e-01 -7.41223514e-01
-1.27224192e-01 7.93767989e-01 5.12148201e-01 -9.42883920e-03
3.15186709e-01 7.17144549e-01 -1.40201473e+00 -4.43481356e-01
6.64975941e-01 5.79869151e-01 5.26612282e-01 7.78968334e-01
2.12928772e-01 4.62979615e-01 -7.10596085e-01 3.91900003e-01
3.04861754e-01 -2.22917646e-01 -4.11414623e-01 -3.42124343e-01
-5.16792722e-02 -5.33614397e-01 -3.67917657e-01 9.65099275e-01
5.15182793e-01 8.64902616e-01 7.13113248e-01 -9.68554080e-01
-1.29151434e-01 2.27337241e-01 -4.99276072e-02 4.46036994e-01
2.03124613e-01 7.28377938e-01 3.55260611e-01 -4.61818464e-02
5.31714916e-01 1.57009494e+00 2.77155280e-01 5.39432287e-01
-1.38169348e+00 -5.80597341e-01 4.42063689e-01 -1.17969021e-01
-1.13476586e+00 -2.44374141e-01 5.41850030e-01 -3.15962911e-01
1.01691735e+00 -1.92076519e-01 8.28152597e-01 1.45458448e+00
6.42276645e-01 5.92220426e-01 1.53198934e+00 -2.01126471e-01
9.36216116e-01 -1.59684256e-01 -5.02182841e-01 5.07035851e-01
1.25219032e-01 7.83363342e-01 -4.97845083e-01 -5.54880261e-01
1.17980552e+00 -2.07844555e-01 -1.10795699e-01 -5.62840402e-01
-9.90784645e-01 6.36332333e-01 1.70421451e-01 -9.89412814e-02
-5.82100034e-01 7.47821510e-01 8.33725035e-02 1.84115693e-02
1.74575031e-01 8.34502280e-01 -6.81835413e-01 -6.23878837e-01
-6.01976275e-01 8.58054161e-01 9.74982917e-01 6.06705070e-01
5.38866341e-01 1.62525579e-01 -3.06934536e-01 4.32766736e-01
5.29768825e-01 7.36162603e-01 3.82028848e-01 -1.31453133e+00
-2.02618279e-02 6.13602996e-02 6.40887141e-01 -5.38196206e-01
-1.97651550e-01 -3.67595136e-01 -1.02832682e-01 7.70561874e-01
3.55793297e-01 -4.46024299e-01 -1.25272834e+00 1.83017802e+00
5.27858615e-01 5.56565762e-01 -4.75766556e-03 7.42448807e-01
-1.43913612e-01 7.96797335e-01 4.80001628e-01 -1.04432121e-01
6.62509203e-01 -7.23381042e-01 -4.70321268e-01 -5.72971821e-01
-2.34800264e-01 -5.79961002e-01 7.88230598e-01 1.87741339e-01
-9.18588638e-01 -3.69581491e-01 -8.65615070e-01 6.41239583e-01
-2.10083976e-01 -5.06945729e-01 5.19908667e-01 1.91364512e-01
-1.02521336e+00 1.06081128e+00 -1.36414075e+00 -4.87777084e-01
-3.81062441e-02 4.13368702e-01 7.18424022e-02 2.05489576e-01
-1.13934100e+00 1.44993150e+00 7.00544357e-01 -4.32508916e-01
-1.62606275e+00 -5.52762568e-01 -5.61977446e-01 -7.20229745e-02
7.90354013e-01 -8.20046365e-01 1.81001997e+00 -4.59167153e-01
-2.36794853e+00 -2.11346686e-01 -6.72547668e-02 -6.64636910e-01
3.56093884e-01 -4.16404575e-01 -2.03415081e-01 -6.54254407e-02
-1.01748094e-01 6.54176652e-01 9.65804100e-01 -1.35224116e+00
-6.17098212e-01 -9.53978077e-02 1.24611117e-01 6.17003024e-01
7.78049976e-02 5.10444865e-02 -4.12530482e-01 -4.71113890e-01
-4.04210091e-01 -1.11070096e+00 -8.17049563e-01 -2.46455356e-01
1.56335041e-01 -3.29027414e-01 8.05808663e-01 -3.51006567e-01
8.43908012e-01 -1.72025251e+00 4.34077293e-01 2.48021036e-01
-3.02551568e-01 2.48884514e-01 1.49765804e-01 4.69742268e-01
4.50486481e-01 -9.04251412e-02 3.95917483e-02 -1.32225826e-01
2.68653810e-01 6.84543252e-01 -4.66810882e-01 3.28935474e-01
9.99426022e-02 7.19203591e-01 -1.16257060e+00 -4.57861394e-01
4.73725051e-01 4.15382087e-01 -3.74068737e-01 3.93885493e-01
-8.48582983e-01 1.02541208e+00 -9.24993992e-01 1.54996097e-01
1.06072120e-01 -9.69981775e-02 2.64083624e-01 3.26174647e-01
-2.10109130e-01 2.76145309e-01 -1.13608372e+00 1.62532604e+00
-4.24694389e-01 -4.77990657e-02 2.66462285e-02 -5.28902411e-01
8.01754117e-01 3.33289832e-01 5.51487505e-01 -6.28414094e-01
3.15663546e-01 -6.09934852e-02 1.25474259e-01 -3.26147437e-01
4.27671999e-01 -1.73716128e-01 -1.63434252e-01 4.84740824e-01
1.68407202e-01 -4.93072033e-01 8.75776857e-02 -2.10747629e-01
1.31438196e+00 8.69734645e-01 2.98166722e-01 -3.11816037e-01
3.10161319e-02 -5.06753139e-02 7.68737912e-01 1.20278096e+00
-5.31496406e-01 1.70276407e-02 1.19224243e-01 -2.82246172e-01
-1.05470812e+00 -1.11738515e+00 -3.39705609e-02 1.06098592e+00
1.79617912e-01 -2.74359316e-01 -4.74475235e-01 -4.52914953e-01
2.46166512e-01 8.48878682e-01 -5.14350355e-01 -1.70131803e-01
-6.84297442e-01 -6.12935007e-01 3.55557688e-02 4.65479702e-01
3.14810961e-01 -1.14546537e+00 -1.13445556e+00 6.41035080e-01
3.87450874e-01 -9.33198690e-01 -9.21544209e-02 4.04001117e-01
-7.65367806e-01 -7.60565937e-01 -4.81572211e-01 -9.44639817e-02
2.20381439e-01 -2.99582809e-01 9.86949682e-01 -3.43484432e-01
1.20264463e-01 6.93664134e-01 -2.64692791e-02 -4.29120600e-01
-7.08283722e-01 -9.87532139e-02 4.99896675e-01 -5.53711772e-01
-6.26115575e-02 -6.35610759e-01 -5.65481961e-01 1.53018177e-01
-7.45308340e-01 -6.67818859e-02 4.86282349e-01 6.27049148e-01
6.84680700e-01 1.49551556e-01 1.63900450e-01 -2.15936884e-01
9.46069062e-01 -4.33445305e-01 -1.14098537e+00 1.53917924e-01
-5.27628839e-01 6.29472017e-01 6.11242950e-01 -7.94328034e-01
-1.06809688e+00 3.63125876e-02 7.07571059e-02 -5.39342344e-01
-4.76562053e-01 3.36550564e-01 5.50344214e-02 -2.16329210e-02
5.42444408e-01 2.00389195e-02 -8.96515325e-02 -5.22282004e-01
3.98372263e-01 -8.95116292e-03 4.33281422e-01 -1.34631944e+00
6.99481726e-01 2.23906189e-01 2.36075029e-01 -4.14646864e-01
-6.81923270e-01 -2.58982807e-01 -2.90627182e-01 -2.99601406e-01
7.80479789e-01 -5.60665309e-01 -8.47706199e-01 2.84376383e-01
-8.25258076e-01 -8.57531369e-01 -4.00653154e-01 6.51028335e-01
-9.13790941e-01 1.21596567e-01 -3.37507278e-01 -9.39475358e-01
-5.82488291e-02 -1.45938826e+00 8.65453184e-01 6.84314191e-01
-2.91664451e-01 -1.08821166e+00 7.62434602e-01 -4.46937382e-01
6.05766714e-01 2.94525564e-01 7.17495024e-01 -4.16267961e-01
-6.72946036e-01 1.06847301e-01 3.53012770e-01 6.40248284e-02
8.53960589e-03 -1.78857166e-02 -6.07423306e-01 -5.36978126e-01
-9.12211929e-03 -1.49709553e-01 7.09561110e-01 5.53390145e-01
8.43713582e-01 -1.53670698e-01 -6.10822737e-01 3.60005289e-01
1.31666505e+00 5.34860253e-01 2.20473379e-01 6.51136160e-01
2.69315124e-01 4.03985307e-02 5.83316624e-01 5.18662810e-01
1.13577880e-01 6.41538501e-01 4.50019598e-01 4.34613109e-01
1.50626272e-01 -4.78650391e-01 5.34055471e-01 3.41223419e-01
5.67672476e-02 -2.36072659e-01 -9.36657488e-01 3.70831370e-01
-2.02313018e+00 -7.95663416e-01 8.50089073e-01 2.28221607e+00
7.38015652e-01 3.79853666e-01 1.04137093e-01 -5.89116275e-01
5.71783721e-01 -3.83500233e-02 -1.12864757e+00 -2.81584233e-01
2.76550144e-01 4.61225927e-01 7.79937685e-01 5.27937770e-01
-1.19942093e+00 1.07431197e+00 7.22464800e+00 6.33042634e-01
-9.87184525e-01 -4.69826274e-02 2.67382741e-01 -8.04931074e-02
4.10982668e-02 2.86189646e-01 -8.81377995e-01 5.18756449e-01
1.37205219e+00 -4.01951551e-01 6.35593116e-01 8.05540085e-01
6.09279275e-01 -4.74269629e-01 -9.36172187e-01 3.42551500e-01
-4.47123945e-01 -1.12989080e+00 -2.22862005e-01 2.05758438e-01
8.02203715e-01 3.80176216e-01 2.20567435e-01 6.07859671e-01
1.21694469e+00 -9.40546036e-01 6.64571404e-01 7.63766050e-01
2.94680148e-01 -7.34440684e-01 2.72347420e-01 5.51298618e-01
-9.31301475e-01 -1.22410282e-01 -2.45003030e-01 -1.96586385e-01
2.40222976e-01 -1.11844666e-01 -1.08827770e+00 2.56742150e-01
7.01536417e-01 3.54885042e-01 -3.58509690e-01 1.41184628e+00
-4.45787549e-01 8.54654610e-01 -8.21198106e-01 -2.16077641e-01
4.65381056e-01 -4.95587550e-02 9.29855466e-01 7.00976908e-01
2.41391018e-01 -8.12050924e-02 7.02006638e-01 9.23460782e-01
3.44806671e-01 -3.93969566e-01 -6.35177732e-01 -7.02944398e-02
5.00334322e-01 9.27818298e-01 -7.86946952e-01 -2.28991300e-01
1.31970882e-01 5.20644724e-01 3.96711588e-01 5.92481494e-01
-8.06428432e-01 2.99852699e-01 8.58352482e-01 -3.58613372e-01
4.16334599e-01 -7.30192065e-01 3.87131035e-01 -7.65438974e-01
-4.00852919e-01 -7.65108526e-01 7.63215795e-02 -6.61159217e-01
-1.05146790e+00 1.62421644e-01 5.26029944e-01 -9.11234617e-01
-6.16011858e-01 -5.31694114e-01 -4.81159896e-01 9.37070429e-01
-1.37706411e+00 -6.18185699e-01 2.07997382e-01 2.53507882e-01
4.65998501e-01 -1.60242543e-01 9.27122295e-01 -2.76322246e-01
-5.16845822e-01 -6.06258363e-02 6.53305411e-01 -3.47250730e-01
7.18945503e-01 -1.42020774e+00 4.06070650e-01 8.18745375e-01
-1.70841917e-01 6.51193857e-01 1.30875373e+00 -9.86294150e-01
-1.44669366e+00 -9.54913378e-01 -2.28119478e-01 -4.29499507e-01
8.07606995e-01 3.75711843e-02 -1.03287148e+00 5.87420046e-01
2.39045933e-01 -2.07432881e-01 1.42461419e-01 -7.30251968e-02
1.08999953e-01 2.32075170e-01 -8.73508096e-01 8.26250076e-01
6.92892313e-01 -2.01323539e-01 -4.70751613e-01 1.58954456e-01
7.14983523e-01 -7.51714110e-01 -9.14635003e-01 4.99889761e-01
3.91922176e-01 -5.74914932e-01 8.97866189e-01 -7.76613057e-01
-7.08745494e-02 -5.13369858e-01 1.45642180e-02 -1.75947940e+00
-3.17060053e-01 -9.47907090e-01 -5.08031607e-01 7.11122036e-01
-4.29350138e-02 -5.58604240e-01 5.93579650e-01 6.59879863e-01
5.85191585e-02 -6.63680911e-01 -1.03884673e+00 -9.14556980e-01
3.26581448e-01 -3.65557253e-01 7.53089905e-01 2.99619883e-01
-4.11677688e-01 1.54342413e-01 -3.68068933e-01 4.32417274e-01
7.07110465e-01 -4.76480760e-02 8.69344890e-01 -8.85076106e-01
-7.63461351e-01 -4.85006392e-01 -1.86909452e-01 -1.08097029e+00
5.15312612e-01 -5.10705471e-01 5.04498065e-01 -1.59463787e+00
9.91332829e-02 -4.29860950e-01 -4.67118025e-01 3.17915142e-01
-2.43314281e-01 -6.59942150e-01 3.24948400e-01 1.81253985e-01
-7.34643161e-01 9.49526727e-01 1.40226424e+00 1.29422218e-01
-2.97392845e-01 7.28659183e-02 -1.45573884e-01 7.90979564e-01
1.00181508e+00 -5.36805570e-01 -3.76166552e-01 -3.52550834e-01
-3.35482024e-02 1.86346501e-01 4.07172143e-01 -1.27035642e+00
3.58014107e-01 -8.60636294e-01 4.23085660e-01 -3.38088244e-01
5.39653003e-01 -7.20500588e-01 6.61019459e-02 4.65885788e-01
-3.59108746e-01 2.65890062e-01 6.13111317e-01 1.16788435e+00
3.11713099e-01 -1.53112069e-01 6.93956614e-01 -2.31054217e-01
-9.11672175e-01 3.38214278e-01 -6.39261425e-01 -3.88446939e-03
9.35999990e-01 2.37735078e-01 -1.96366504e-01 -3.48112494e-01
-7.02292740e-01 4.21633840e-01 6.59891129e-01 5.09689927e-01
5.23666203e-01 -9.17279303e-01 -2.43457839e-01 -1.18267834e-01
-3.00464571e-01 -6.08699173e-02 -1.71347946e-01 1.98087379e-01
-1.59424394e-01 3.20101470e-01 -3.84009838e-01 -5.36900699e-01
-6.59761667e-01 5.01668096e-01 6.10313296e-01 -5.62128961e-01
-7.18047917e-01 3.96670312e-01 -1.24800220e-01 -3.96170467e-01
8.38257745e-02 -2.50741929e-01 -4.39652838e-02 -8.66834402e-01
3.04506391e-01 2.55485415e-01 -3.61060500e-01 -4.85931307e-01
-4.55922410e-02 3.16530108e-01 9.98745784e-02 -6.63287282e-01
1.14757144e+00 2.35346809e-01 1.85522869e-01 5.95356584e-01
5.61205685e-01 -4.55581188e-01 -2.11523438e+00 -2.90229797e-01
1.32283121e-01 -4.49709266e-01 2.32782051e-01 -9.11885798e-01
-3.99520397e-01 4.15504843e-01 7.63897240e-01 -1.68739393e-01
5.54436684e-01 -1.70180291e-01 5.15971243e-01 5.16755164e-01
7.56334126e-01 -1.23664904e+00 1.49167672e-01 8.85498226e-01
6.45281732e-01 -9.89244998e-01 1.65872216e-01 4.16756779e-01
-3.84164870e-01 1.00303090e+00 7.20767856e-01 -3.18962812e-01
6.67017221e-01 4.23342347e-01 -1.73165187e-01 -5.93096279e-02
-9.55785036e-01 -2.62872607e-01 1.83603317e-01 6.47825897e-01
-1.53485030e-01 1.63216531e-01 1.58700272e-01 -8.93675014e-02
-1.16967931e-02 1.41713232e-01 2.27000758e-01 1.57969832e+00
-7.24787533e-01 -1.30914807e+00 -4.60670531e-01 2.77559102e-01
-2.65891641e-01 3.45853716e-01 1.13595285e-01 7.06327438e-01
-1.42462507e-01 7.47513413e-01 -9.97513384e-02 2.33009923e-02
2.64528304e-01 2.40028530e-01 7.98851132e-01 -6.89929128e-01
-4.49095845e-01 2.94440061e-01 -1.97514638e-01 -6.98270023e-01
-3.48559886e-01 -8.90785873e-01 -1.31158578e+00 -4.45268378e-02
-4.00675200e-02 1.59824997e-01 7.48472095e-01 1.14801145e+00
4.11185145e-01 7.25015879e-01 8.38362649e-02 -1.12016761e+00
-1.05254507e+00 -8.46204042e-01 -1.58740431e-01 2.32362628e-01
2.74340242e-01 -1.27104235e+00 -1.27153873e-01 -2.58310497e-01]
|
[4.283660888671875, 1.8487696647644043]
|
00f02bd0-985d-4cfc-a1cf-32271de2a669
|
semantically-grounded-object-matching-for
|
2111.07975
| null |
https://arxiv.org/abs/2111.07975v1
|
https://arxiv.org/pdf/2111.07975v1.pdf
|
Semantically Grounded Object Matching for Robust Robotic Scene Rearrangement
|
Object rearrangement has recently emerged as a key competency in robot manipulation, with practical solutions generally involving object detection, recognition, grasping and high-level planning. Goal-images describing a desired scene configuration are a promising and increasingly used mode of instruction. A key outstanding challenge is the accurate inference of matches between objects in front of a robot, and those seen in a provided goal image, where recent works have struggled in the absence of object-specific training data. In this work, we explore the deterioration of existing methods' ability to infer matches between objects as the visual shift between observed and goal scenes increases. We find that a fundamental limitation of the current setting is that source and target images must contain the same $\textit{instance}$ of every object, which restricts practical deployment. We present a novel approach to object matching that uses a large pre-trained vision-language model to match objects in a cross-instance setting by leveraging semantics together with visual features as a more robust, and much more general, measure of similarity. We demonstrate that this provides considerably improved matching performance in cross-instance settings, and can be used to guide multi-object rearrangement with a robot manipulator from an image that shares no object $\textit{instances}$ with the robot's scene.
|
['Ingmar Posner', 'Ioannis Havoutis', 'Sagar Vaze', 'Walter Goodwin']
|
2021-11-15
| null | null | null | null |
['robot-manipulation']
|
['robots']
|
[ 4.51576620e-01 -3.12448829e-01 -1.60786986e-01 -4.24932241e-01
-6.78365827e-01 -6.29884779e-01 3.64336610e-01 4.74568844e-01
-4.17001039e-01 1.79336250e-01 -2.81140894e-01 -1.51136816e-01
-3.82148862e-01 -4.68012691e-01 -1.20687950e+00 -2.62145996e-01
-5.59647754e-02 7.61183500e-01 5.55119336e-01 -5.83718836e-01
6.20161772e-01 8.04874003e-01 -2.11983919e+00 3.99663597e-01
3.06470573e-01 8.96201849e-01 9.95888591e-01 7.09425747e-01
-4.87746000e-02 5.85461974e-01 -5.32522440e-01 -9.98093709e-02
5.04696429e-01 -9.85087305e-02 -1.04786801e+00 4.11597133e-01
8.66994798e-01 -3.12763333e-01 -3.32289994e-01 1.25071061e+00
2.46296749e-01 6.21721685e-01 5.02486527e-01 -1.56169355e+00
-4.58145797e-01 4.45594490e-01 -4.52818274e-01 3.06168884e-01
7.74985790e-01 3.97274226e-01 8.57725978e-01 -8.14542651e-01
8.59310210e-01 1.51594210e+00 2.60982186e-01 3.68011177e-01
-1.04583919e+00 -4.28009331e-01 3.35760564e-01 5.49370408e-01
-1.14148974e+00 -4.37507093e-01 6.01580560e-01 -4.53108728e-01
1.19988680e+00 1.36425510e-01 5.46423972e-01 6.08780742e-01
3.59445512e-01 7.20200300e-01 7.70231366e-01 -5.61834097e-01
-1.45569108e-02 3.57784852e-02 -3.44324932e-02 9.07420456e-01
8.20493102e-02 6.73658401e-02 -6.48743033e-01 1.94833398e-01
1.03266418e+00 3.11670810e-01 -2.21381202e-01 -1.08524823e+00
-1.47691464e+00 4.82590765e-01 6.49987340e-01 3.34749579e-01
-2.48497590e-01 3.81853133e-01 9.97271314e-02 4.73482221e-01
-2.74303854e-01 7.76513457e-01 -2.25625083e-01 -1.58103436e-01
-3.40012312e-01 4.97863650e-01 6.77746773e-01 1.81510389e+00
9.41099823e-01 -2.73037136e-01 7.35016093e-02 5.57618022e-01
1.02505885e-01 3.63896132e-01 1.44037575e-01 -1.32481301e+00
7.01199114e-01 5.79912901e-01 2.03444079e-01 -1.04676354e+00
-3.54244679e-01 -1.63487822e-01 -3.60648148e-02 6.37232959e-01
5.65339446e-01 6.21512294e-01 -8.33231449e-01 1.64140403e+00
5.43905556e-01 -2.03943133e-01 9.23333466e-02 1.01114261e+00
5.01492798e-01 2.74247646e-01 -1.12116955e-01 1.50592759e-01
1.39977467e+00 -1.11142445e+00 -2.73322552e-01 -6.35872006e-01
5.19426227e-01 -9.60846245e-01 1.10778952e+00 3.52295041e-01
-1.30119753e+00 -6.67989910e-01 -9.80221152e-01 -2.84147620e-01
-4.11868125e-01 -2.99492121e-01 6.85413301e-01 -6.56524347e-03
-9.10703480e-01 7.78663814e-01 -7.62005746e-01 -7.35919535e-01
2.91040897e-01 4.94435877e-01 -7.15862036e-01 -6.26220047e-01
-3.57741624e-01 1.37250733e+00 6.67581975e-01 -2.62446180e-02
-1.02210975e+00 -5.16080856e-01 -1.14544666e+00 -4.29857001e-02
9.12642360e-01 -5.34474552e-01 1.56386936e+00 -6.17190719e-01
-1.17738223e+00 1.03257287e+00 -8.30532387e-02 -1.15928277e-02
2.59631336e-01 -1.51758581e-01 6.62164912e-02 2.90914208e-01
2.82128662e-01 7.85250604e-01 9.51757133e-01 -1.58319306e+00
-9.71300900e-01 -7.23441958e-01 3.99335206e-01 4.99682218e-01
2.68514484e-01 4.05882485e-03 -6.71329975e-01 -1.01988882e-01
5.83233535e-01 -9.17471707e-01 -2.00735509e-01 2.83286095e-01
-1.79352105e-01 -3.41697782e-01 7.87668288e-01 -2.64957279e-01
4.82856750e-01 -2.02988768e+00 3.79724354e-01 1.51207820e-01
-6.49395138e-02 -7.50684142e-02 -3.05594623e-01 5.81070602e-01
1.69422161e-02 -4.29995507e-01 -1.54481939e-04 -2.11866528e-01
1.38021171e-01 3.60642552e-01 -3.80116850e-01 4.47488278e-01
2.60441989e-01 9.60912406e-01 -1.33140755e+00 -4.69164133e-01
3.33985448e-01 1.76236080e-03 -5.43715477e-01 4.13731217e-01
-3.98076057e-01 4.92864460e-01 -4.04927939e-01 7.33578801e-01
3.26779991e-01 -1.88319683e-01 1.28223568e-01 -2.17342645e-01
-5.14273569e-02 8.69634971e-02 -1.22720444e+00 2.22264457e+00
-2.89560795e-01 4.52725410e-01 1.69691890e-01 -1.18865955e+00
8.37621212e-01 -1.27892673e-01 4.23560113e-01 -7.08256543e-01
7.16044232e-02 1.53007343e-01 1.44501373e-01 -8.67661715e-01
7.53610611e-01 2.72910371e-02 -2.31689438e-01 2.93311179e-01
5.45792431e-02 -1.01141071e+00 3.12125683e-01 1.25931770e-01
1.11109412e+00 3.97429705e-01 3.72243613e-01 -7.52569810e-02
2.09424924e-02 6.70134068e-01 7.58379102e-02 1.15647948e+00
-2.75324762e-01 5.37207365e-01 6.88494146e-02 -2.28201941e-01
-1.13321280e+00 -1.09256220e+00 5.94459474e-02 1.40281975e+00
8.60420346e-01 2.24925708e-02 -3.86542857e-01 -4.93940204e-01
2.64461994e-01 5.96583724e-01 -2.31050298e-01 -2.29412526e-01
-9.83357370e-01 7.95678645e-02 -1.29470408e-01 6.94673300e-01
-2.11031381e-02 -1.36394060e+00 -1.21637452e+00 2.44466692e-01
4.15199064e-02 -1.29074264e+00 -4.90568608e-01 4.27845031e-01
-8.56909692e-01 -1.27433610e+00 -3.19878727e-01 -1.19818771e+00
1.00771010e+00 7.12436616e-01 1.25588500e+00 1.63264811e-01
-6.42926991e-01 9.52395082e-01 -3.51879954e-01 -3.52335215e-01
-5.93556702e-01 -2.83551544e-01 -2.76350882e-02 -6.35949790e-01
1.56505719e-01 -2.53746033e-01 -3.80054027e-01 3.56023192e-01
-8.76054764e-01 -2.16643289e-02 7.68964469e-01 7.46597469e-01
4.87014949e-01 -4.90019768e-02 1.66824371e-01 -4.45119560e-01
2.99740434e-01 -2.74585128e-01 -6.71350598e-01 3.28212649e-01
-2.88758665e-01 1.07003063e-01 2.90422350e-01 -5.01851261e-01
-5.77884495e-01 1.70784548e-01 3.22242737e-01 -7.64634132e-01
-2.78000653e-01 4.48831946e-01 -9.96662751e-02 -1.62916273e-01
4.46260244e-01 1.33165747e-01 1.21981665e-01 -3.16406876e-01
4.13103312e-01 3.56717467e-01 9.45233881e-01 -7.61367083e-01
7.37905204e-01 4.35171604e-01 2.04568684e-01 -3.49291116e-01
-6.95678294e-01 -9.30373430e-01 -9.34706211e-01 -1.77249819e-01
5.82696795e-01 -7.25891829e-01 -1.12404358e+00 1.29243582e-01
-1.25612485e+00 -5.54691732e-01 -3.50464821e-01 5.62103987e-01
-1.10964775e+00 3.14541698e-01 -2.92978108e-01 -5.56765676e-01
2.12013945e-01 -1.40803885e+00 1.33574855e+00 8.91450197e-02
-1.10313594e-01 -5.13168216e-01 -5.53568900e-01 3.97664815e-01
1.59174487e-01 -1.75696332e-03 1.06528342e+00 -5.51180601e-01
-1.00608563e+00 -2.25203261e-01 -2.68129349e-01 -7.48878047e-02
2.27543727e-01 -3.82759154e-01 -3.93455029e-01 -6.11692011e-01
1.58456676e-02 -5.48674166e-01 4.43484485e-01 -8.11849684e-02
1.04086185e+00 -2.40667500e-02 -5.36494195e-01 1.69568285e-01
1.41083825e+00 3.52698028e-01 3.82140577e-01 4.21278387e-01
6.64001942e-01 9.01842952e-01 1.10427666e+00 2.40459919e-01
4.00273412e-01 8.74942362e-01 8.87990952e-01 2.63268679e-01
-1.54612541e-01 -7.05952048e-02 2.32045874e-01 4.36026573e-01
1.29860222e-01 -1.66431770e-01 -1.01339018e+00 7.52758563e-01
-2.02242517e+00 -1.02786660e+00 2.05907553e-01 2.14431572e+00
6.98637068e-01 1.10917591e-01 -1.30807176e-01 -1.52052134e-01
7.37340987e-01 -2.44037107e-01 -7.80483186e-01 -2.44835392e-01
3.03887039e-01 1.58711508e-01 3.77972603e-01 4.12159860e-01
-9.50231671e-01 1.00565565e+00 6.23178530e+00 5.24978399e-01
-8.96111727e-01 -6.98994249e-02 -1.17841415e-01 -1.55221242e-02
2.34928019e-02 -8.66331384e-02 -8.25348854e-01 2.08777580e-02
2.32777491e-01 -1.15034562e-02 5.68204939e-01 1.04761004e+00
-1.80645451e-01 -4.76055562e-01 -1.94560039e+00 1.16640294e+00
4.77587163e-01 -1.09334588e+00 -6.49350658e-02 -2.42378607e-01
4.13797826e-01 -1.47932544e-01 1.20645471e-01 3.83573502e-01
4.65539008e-01 -9.98554409e-01 1.08820271e+00 3.48268777e-01
5.89773178e-01 -3.19277287e-01 2.42810890e-01 5.64970016e-01
-1.18681085e+00 -3.79255325e-01 -5.18893838e-01 -4.68334816e-02
4.61474359e-02 -2.28156611e-01 -1.18297780e+00 4.80108589e-01
9.12627041e-01 6.10857308e-01 -2.70736635e-01 1.02978158e+00
-4.23937961e-02 -3.96727234e-01 -2.77416497e-01 -4.04971056e-02
1.25256032e-01 3.91460136e-02 7.42308378e-01 8.48190784e-01
3.20310920e-01 9.02730376e-02 7.20935583e-01 8.98819923e-01
1.17605999e-01 -2.43268386e-01 -9.01786745e-01 1.81441814e-01
5.48681378e-01 9.41178083e-01 -7.69175351e-01 -3.36098611e-01
-3.65207911e-01 9.12443697e-01 6.10477090e-01 2.45638534e-01
-6.25453353e-01 -2.43329585e-01 6.97144270e-01 2.03137454e-02
5.61585903e-01 -6.23002887e-01 3.59308347e-02 -8.68614912e-01
2.90501505e-01 -9.89896119e-01 2.50482589e-01 -1.15225124e+00
-9.95455086e-01 2.76301920e-01 4.71955985e-01 -1.36164737e+00
-3.77879053e-01 -8.90986621e-01 -1.68007776e-01 6.64214313e-01
-1.54486465e+00 -1.08217514e+00 -6.17542803e-01 6.09770954e-01
8.61956179e-01 3.02692074e-02 5.92123449e-01 -7.81137496e-02
-1.51651204e-01 2.50489354e-01 -3.53660956e-02 -1.20637409e-01
5.84676981e-01 -1.06281388e+00 9.26365927e-02 6.76803410e-01
1.49548933e-01 6.65848076e-01 9.84914482e-01 -5.89205801e-01
-2.04419231e+00 -8.32809269e-01 3.84060800e-01 -8.36914420e-01
5.28067648e-01 -3.43027502e-01 -8.34796369e-01 1.02249241e+00
-7.48395026e-02 8.56255889e-02 -3.70395463e-03 -2.70437330e-01
-3.59106809e-01 1.82268675e-02 -1.12252426e+00 7.84413815e-01
1.53551388e+00 -5.21216214e-01 -9.12763953e-01 6.21590853e-01
7.39464104e-01 -9.97878075e-01 -9.23487961e-01 5.08703828e-01
5.24441063e-01 -7.36835778e-01 1.11187768e+00 -8.57691228e-01
3.75481308e-01 -4.13413644e-01 -4.94188339e-01 -1.06963313e+00
-2.63101190e-01 -3.93506676e-01 5.25847897e-02 7.70431817e-01
1.72436275e-02 -3.68698716e-01 6.79336250e-01 7.89056122e-01
-6.09503031e-01 -5.39623797e-01 -8.41239810e-01 -1.09906423e+00
-2.22074926e-01 -5.10811508e-01 5.44388831e-01 6.73603833e-01
2.93250620e-01 2.99739949e-02 1.09343417e-01 3.72640759e-01
4.74627703e-01 5.76299787e-01 1.12916827e+00 -9.46655273e-01
-1.68931857e-01 -4.71587092e-01 -8.86648357e-01 -1.39136910e+00
2.39258587e-01 -9.43663716e-01 8.14606249e-01 -1.50758386e+00
3.36029440e-01 -6.77447677e-01 -9.94394720e-02 5.80273867e-01
8.03410634e-03 1.15454476e-02 5.78521371e-01 3.35561514e-01
-7.66245306e-01 2.67265886e-01 1.56594610e+00 -3.47428530e-01
-3.16919796e-02 -2.01039001e-01 -3.55565995e-01 5.37501097e-01
4.89421487e-01 -4.29897279e-01 -2.89158046e-01 -7.62485325e-01
-1.02773458e-02 1.96449772e-01 5.85794508e-01 -1.00240266e+00
5.55431545e-01 -5.22627711e-01 1.60618588e-01 -5.18023312e-01
6.25217855e-01 -1.22143531e+00 7.71043524e-02 5.14311433e-01
-4.77980345e-01 5.26832998e-01 3.93539578e-01 7.00144470e-01
-4.34410535e-02 -6.49891675e-01 5.56274474e-01 -4.53557730e-01
-1.38868427e+00 1.41867265e-01 -4.35680114e-02 -1.56381518e-01
1.41062081e+00 -6.47053301e-01 -4.89038140e-01 -7.19285235e-02
-6.49342000e-01 3.61862272e-01 8.07026803e-01 7.95654893e-01
9.02802229e-01 -1.06589043e+00 -4.28000629e-01 2.19536364e-01
5.15381098e-01 4.25609052e-01 3.06335539e-02 6.81568861e-01
-4.28191185e-01 3.13255817e-01 -4.23575819e-01 -1.05831695e+00
-1.31320143e+00 8.35475564e-01 1.03356518e-01 9.64951292e-02
-7.93710113e-01 8.41653883e-01 2.88314193e-01 -4.03063476e-01
3.98903072e-01 -5.04090667e-01 7.53268376e-02 -5.02792478e-01
2.41413534e-01 1.45798206e-01 1.07299857e-01 -8.11926782e-01
-3.79553229e-01 6.34500265e-01 -1.68917120e-01 3.89261432e-02
1.13214934e+00 -2.36386314e-01 -1.69008195e-01 5.84058762e-01
1.03852654e+00 -3.87884587e-01 -1.35672462e+00 -4.46182042e-01
1.30429730e-01 -8.01340520e-01 -4.33658898e-01 -4.71659541e-01
-6.37988985e-01 7.25813687e-01 5.09731591e-01 -8.01976677e-03
6.77070200e-01 5.39641142e-01 5.19980907e-01 9.01772618e-01
1.04436553e+00 -9.45713222e-01 6.99530602e-01 6.27422094e-01
1.22662115e+00 -1.58382678e+00 2.91472003e-02 -5.24176896e-01
-4.44273680e-01 1.14857543e+00 9.94204223e-01 4.16515255e-03
1.62459970e-01 9.35149342e-02 -8.96571428e-02 -4.84729916e-01
-5.62130332e-01 -3.68811876e-01 3.11256349e-01 7.82158256e-01
-1.42268524e-01 -1.60579383e-01 4.90902185e-01 -2.24551260e-02
-1.77363411e-01 -3.51592004e-01 4.62201267e-01 1.67911780e+00
-8.26267600e-01 -8.15691769e-01 -4.03381258e-01 3.81713867e-01
4.93404903e-02 1.36366978e-01 -7.07287714e-02 9.40583527e-01
2.12110728e-02 9.73260045e-01 3.19928557e-01 -7.78763518e-02
6.84753060e-01 -4.38966453e-02 1.11374366e+00 -1.22032142e+00
-5.08944094e-01 -3.56585205e-01 -2.85148859e-01 -8.88317406e-01
-5.07733524e-01 -7.83098936e-01 -1.53812957e+00 -1.79324318e-02
-3.76374543e-01 -1.34450734e-01 8.59141648e-01 1.06789899e+00
3.13074410e-01 4.50007200e-01 3.13884079e-01 -1.20975447e+00
-7.54003406e-01 -6.03851438e-01 -3.60696167e-01 8.46437454e-01
3.98270488e-01 -9.63880658e-01 7.90303852e-03 1.92632973e-01]
|
[4.959063529968262, 0.20806580781936646]
|
2632cd88-3162-47b0-94d4-50bb14eccce3
|
how-to-ask-better-questions-a-large-scale
|
1911.09247
| null |
https://arxiv.org/abs/1911.09247v1
|
https://arxiv.org/pdf/1911.09247v1.pdf
|
How to Ask Better Questions? A Large-Scale Multi-Domain Dataset for Rewriting Ill-Formed Questions
|
We present a large-scale dataset for the task of rewriting an ill-formed natural language question to a well-formed one. Our multi-domain question rewriting MQR dataset is constructed from human contributed Stack Exchange question edit histories. The dataset contains 427,719 question pairs which come from 303 domains. We provide human annotations for a subset of the dataset as a quality estimate. When moving from ill-formed to well-formed questions, the question quality improves by an average of 45 points across three aspects. We train sequence-to-sequence neural models on the constructed dataset and obtain an improvement of 13.2% in BLEU-4 over baseline methods built from other data resources. We release the MQR dataset to encourage research on the problem of question rewriting.
|
['Manaal Faruqui', 'Xiance Si', 'Zewei Chu', 'Mingda Chen', 'Kevin Gimpel', 'Miaosen Wang', 'Jing Chen']
|
2019-11-21
| null | null | null | null |
['question-rewriting']
|
['natural-language-processing']
|
[ 3.91657829e-01 4.27998543e-01 2.34550506e-01 -5.88311255e-01
-1.65117788e+00 -1.00498116e+00 5.11890948e-01 8.18310529e-02
-5.66446781e-01 7.81224847e-01 5.25095940e-01 -5.14083743e-01
2.28565380e-01 -6.18920982e-01 -1.04823983e+00 2.54367381e-01
6.58471704e-01 4.74366397e-01 6.73719227e-01 -7.25444555e-01
4.15873617e-01 -2.39259586e-01 -1.19101131e+00 7.29262650e-01
1.27190912e+00 3.70093465e-01 1.72769949e-01 1.21555555e+00
-4.10175771e-01 9.58850443e-01 -9.36677217e-01 -1.19452155e+00
1.79783881e-01 -6.25145316e-01 -1.43149090e+00 -3.69574368e-01
1.02456760e+00 -1.83184564e-01 -2.38003731e-01 9.61480796e-01
3.49891156e-01 2.38850445e-01 2.67394155e-01 -9.41547751e-01
-1.38715959e+00 5.51542461e-01 2.62407307e-02 2.65586019e-01
9.12919760e-01 3.67714226e-01 1.60527456e+00 -7.09454179e-01
1.11464715e+00 1.09570074e+00 5.41250229e-01 8.97349954e-01
-1.02771127e+00 -2.17193022e-01 -2.11520448e-01 3.00994873e-01
-5.54190695e-01 -4.86738652e-01 1.13403901e-01 -1.83672071e-01
1.59493017e+00 5.37719309e-01 7.71362707e-02 8.40496302e-01
1.97061554e-01 4.19233143e-01 1.07732713e+00 -4.80821162e-01
-2.10193336e-01 -2.14051709e-01 6.95612729e-01 8.18443537e-01
1.24952763e-01 -2.68869519e-01 -3.91993493e-01 -3.69635284e-01
3.47366631e-02 -5.43223858e-01 -1.67050555e-01 2.91822165e-01
-9.81741369e-01 6.87044680e-01 2.43809253e-01 2.34218717e-01
-9.41239856e-03 9.66532305e-02 3.75050068e-01 9.52694356e-01
1.09671183e-01 1.28509331e+00 -7.41063476e-01 -5.83554685e-01
-2.60608047e-01 8.52595806e-01 1.29796636e+00 1.25357771e+00
7.92226613e-01 -6.00184321e-01 -4.10414815e-01 9.44878340e-01
1.85490549e-01 4.90266293e-01 3.63095641e-01 -1.75748551e+00
7.69294798e-01 9.92939770e-01 3.65646511e-01 -4.60835457e-01
-7.14860708e-02 2.74016410e-01 -4.94997650e-02 -2.87215710e-01
7.40989089e-01 -1.18172746e-02 -7.19610035e-01 1.72572100e+00
2.93588489e-01 -5.01098335e-01 -4.17889208e-02 5.64042091e-01
1.35686362e+00 7.16881990e-01 -2.30695590e-01 1.87382698e-01
1.80244124e+00 -1.34262729e+00 -7.64450192e-01 -1.37795463e-01
7.97938704e-01 -8.68460298e-01 1.64764488e+00 1.80384606e-01
-1.37757158e+00 -4.49333906e-01 -7.99386144e-01 -8.35443974e-01
-2.74719834e-01 -2.81842619e-01 -8.12472850e-02 4.43374574e-01
-1.38733625e+00 3.49511743e-01 -1.92861259e-01 -4.80796278e-01
-1.29769996e-01 -1.42667338e-01 -4.10303593e-01 -1.63967490e-01
-1.30708206e+00 1.25929034e+00 4.11339700e-02 -3.11134219e-01
-7.25198209e-01 -1.02876008e+00 -7.30886161e-01 -1.23899497e-01
5.42345583e-01 -7.55040765e-01 2.15498686e+00 -5.15150011e-01
-1.43330300e+00 1.18504846e+00 -5.21194577e-01 -3.15789104e-01
3.12106222e-01 -4.07716483e-01 -5.65304279e-01 1.21231094e-01
2.90347517e-01 4.05075461e-01 3.88656944e-01 -8.00447166e-01
-6.10998094e-01 -1.06078096e-01 7.08595395e-01 -1.25987157e-01
2.19267800e-01 4.67036933e-01 -4.73607093e-01 -4.64790285e-01
-4.23596859e-01 -1.02307832e+00 -5.82656488e-02 -1.27558082e-01
-1.29762217e-01 -5.73965788e-01 3.06293815e-01 -1.34914553e+00
1.30149102e+00 -1.51852298e+00 -1.33039560e-02 -5.49996853e-01
4.60915625e-01 3.80304039e-01 -6.64380610e-01 6.99087918e-01
-8.59294906e-02 3.57848585e-01 -6.86896384e-01 -1.87225506e-01
1.94098949e-01 1.05191007e-01 -3.14699024e-01 -1.46539971e-01
5.44598103e-01 1.37207651e+00 -1.03464031e+00 -3.39821845e-01
-5.79643607e-01 -2.63109326e-01 -7.82828629e-01 8.28295887e-01
-8.70555520e-01 2.62433082e-01 -1.98652565e-01 4.02769446e-01
4.15084451e-01 -4.75995153e-01 -1.58740841e-02 2.19255343e-01
2.43546695e-01 1.05277359e+00 -3.03729028e-01 1.93545902e+00
-5.04938364e-01 7.16907680e-01 -2.32724130e-01 -5.62321171e-02
9.84112799e-01 3.57263833e-01 -1.42961472e-01 -1.06805313e+00
-3.26599896e-01 2.51268297e-01 3.08496095e-02 -1.00456619e+00
1.21138084e+00 -1.06130965e-01 -5.06599307e-01 8.02166283e-01
1.40753806e-01 -3.69502068e-01 6.51064873e-01 4.45383251e-01
1.73214841e+00 -2.15090793e-02 2.50809193e-01 -5.87427057e-02
6.70832515e-01 1.80175632e-01 3.94010425e-01 8.53138328e-01
-4.00298297e-01 6.64501488e-01 8.26816261e-01 -3.65313977e-01
-1.26306653e+00 -1.10267985e+00 1.85675442e-01 1.42266071e+00
-1.58554077e-01 -4.82561439e-01 -9.48102593e-01 -1.08604479e+00
1.38718113e-02 1.13509727e+00 -6.32398367e-01 -1.63777739e-01
-1.24734330e+00 -1.16707161e-01 1.04187202e+00 2.49517381e-01
3.16750884e-01 -1.33205259e+00 -2.35049978e-01 2.21541941e-01
-7.68732846e-01 -1.23964798e+00 -9.64552879e-01 -4.63767678e-01
-4.82369691e-01 -1.35335994e+00 -4.42225456e-01 -7.51627147e-01
2.98485667e-01 9.51106995e-02 1.96720159e+00 5.23431659e-01
1.02812588e-01 3.73926222e-01 -6.09832168e-01 -7.77918398e-02
-1.10372794e+00 4.59787071e-01 -6.54289186e-01 -5.92290401e-01
5.17436624e-01 -1.99164689e-01 -2.18101501e-01 3.72515082e-01
-1.00965750e+00 -3.61653626e-01 4.06202525e-02 6.16387069e-01
1.94096699e-01 -1.21409214e+00 9.26431477e-01 -1.21003056e+00
1.15963686e+00 -4.42357779e-01 -6.99073076e-01 9.24618840e-01
-4.96821642e-01 3.80848527e-01 6.11475170e-01 -6.26221374e-02
-1.48627102e+00 -5.87972999e-01 -4.54548746e-01 3.77651632e-01
-7.74830626e-03 2.00487658e-01 3.35602998e-03 1.99180335e-01
9.84115303e-01 -9.03923716e-03 -1.65465623e-01 -4.77959365e-01
9.30911005e-01 7.30097950e-01 8.21243942e-01 -8.44130695e-01
6.37015581e-01 -2.51632869e-01 -5.86246789e-01 -4.67466861e-01
-9.09250855e-01 -3.28909308e-01 -2.45552480e-01 -2.65717693e-02
1.05340397e+00 -5.50029278e-01 -8.22137654e-01 3.28073293e-01
-1.63150311e+00 -4.68619466e-01 -3.41478705e-01 -3.43290865e-01
-4.93597627e-01 5.97719669e-01 -8.88056636e-01 -4.21978027e-01
-7.21741736e-01 -1.01485348e+00 9.62419271e-01 3.41681182e-01
-6.47759497e-01 -7.52396524e-01 7.36085892e-01 1.01671958e+00
5.04042089e-01 5.44288643e-02 1.21902037e+00 -9.69804883e-01
-7.96395540e-01 1.18231736e-02 -1.57228798e-01 3.64104241e-01
9.62095559e-02 6.92708716e-02 -5.09277225e-01 -2.65055094e-02
1.18246555e-01 -7.39359975e-01 6.04476810e-01 -4.33527946e-01
7.38229454e-01 -3.28239381e-01 3.21984470e-01 9.93354544e-02
1.10403001e+00 4.52002324e-02 7.21835315e-01 3.86760473e-01
4.49254453e-01 6.38428152e-01 5.37312329e-01 -3.42430174e-02
1.08849812e+00 3.65335315e-01 2.77919322e-01 6.17377043e-01
-3.01795185e-01 -3.99269104e-01 4.44979817e-01 1.23284662e+00
4.45724487e-01 -3.46557736e-01 -1.03240907e+00 7.92478740e-01
-1.53012395e+00 -7.40506709e-01 -6.62564814e-01 1.73340619e+00
1.24489987e+00 -9.37475711e-02 1.03293449e-01 -5.58547080e-01
7.13381410e-01 2.12942094e-01 -5.91987193e-01 -8.90695274e-01
-6.28990903e-02 4.39668775e-01 3.60803902e-02 8.16256940e-01
-5.00903666e-01 8.99179459e-01 6.82295847e+00 1.95517480e-01
-3.86855215e-01 3.40129435e-01 1.36168242e-01 -6.24268390e-02
-1.02897263e+00 3.34513366e-01 -8.46739173e-01 4.55038458e-01
1.38054216e+00 -4.19550210e-01 6.25747859e-01 4.13107306e-01
-2.45508820e-01 -8.43250975e-02 -1.19570470e+00 3.77177536e-01
2.39434972e-01 -1.23198199e+00 4.29761708e-02 -3.21899116e-01
8.50138068e-01 2.38200337e-01 -3.00882012e-01 7.39044785e-01
9.33913589e-01 -9.84380424e-01 5.86811423e-01 8.63779426e-01
7.04981148e-01 -3.75866532e-01 4.10573006e-01 4.01807606e-01
-8.40748966e-01 6.58196881e-02 -3.60666543e-01 -1.03072394e-02
5.56663096e-01 3.85851830e-01 -7.35024393e-01 6.44215405e-01
7.46469736e-01 3.37802291e-01 -1.12558782e+00 7.66911805e-01
-4.60904270e-01 6.51073694e-01 -7.36877769e-02 -4.14322823e-01
8.28979015e-02 -2.53209412e-01 5.05870402e-01 1.03907084e+00
3.20694372e-02 2.86158442e-01 -4.51627702e-01 1.05930436e+00
-7.53129661e-01 -1.30613312e-01 -3.55327398e-01 -2.02370644e-01
5.55750191e-01 1.24207318e+00 1.91153228e-01 -5.24902463e-01
-5.69685996e-01 9.75253940e-01 8.17997217e-01 2.51428723e-01
-8.72371495e-01 -8.12723517e-01 7.14775980e-01 -2.10598841e-01
2.45326340e-01 -1.28798991e-01 -2.65186489e-01 -1.33938360e+00
6.63767695e-01 -1.75685465e+00 4.76443261e-01 -9.91191328e-01
-1.52001894e+00 8.39692712e-01 -3.28805327e-01 -6.27256930e-01
-3.81428599e-01 -3.33058506e-01 -6.24463022e-01 1.06022954e+00
-1.51511276e+00 -8.89874697e-01 -3.04297149e-01 1.49486646e-01
5.74878931e-01 4.88646068e-02 7.56582439e-01 4.77296919e-01
-2.85182774e-01 9.62464750e-01 -1.20070599e-01 1.90217540e-01
1.12156522e+00 -1.51737714e+00 1.53002584e+00 9.10481811e-01
-4.61537428e-02 9.49814737e-01 5.98452985e-01 -7.42624104e-01
-1.32003009e+00 -1.17457533e+00 1.68664551e+00 -1.39754701e+00
7.73637593e-01 -3.68642092e-01 -1.52226269e+00 7.85759330e-01
8.31325293e-01 -3.39061230e-01 6.01192415e-01 -7.94699863e-02
-7.92730629e-01 1.76259860e-01 -1.15684283e+00 6.99700236e-01
1.27379143e+00 -1.10304534e+00 -1.45999110e+00 2.13652447e-01
1.59747553e+00 -5.74699819e-01 -1.29875350e+00 2.15377733e-01
2.46704027e-01 -8.50286961e-01 5.15444160e-01 -1.17352164e+00
8.77314091e-01 -3.24731767e-01 -2.01336056e-01 -1.01044071e+00
-8.09067637e-02 -8.85539055e-01 -2.50722975e-01 1.32367122e+00
8.39938045e-01 -7.52472103e-01 4.54258859e-01 9.83872831e-01
-2.28351712e-01 -7.04199731e-01 -8.92034352e-01 -6.69335067e-01
6.66010261e-01 -1.67484775e-01 8.48908901e-01 6.37897551e-01
-3.09751723e-02 6.56065702e-01 2.20475048e-02 -2.28728235e-01
2.59255290e-01 1.75964132e-01 8.98488700e-01 -7.58858383e-01
-3.56981039e-01 -1.50252149e-01 3.80637705e-01 -1.45229948e+00
4.06257004e-01 -1.07326412e+00 2.62855232e-01 -1.72323978e+00
9.70230922e-02 6.50677681e-02 3.00753176e-01 1.92301899e-01
-8.13940465e-01 -2.08927006e-01 3.24200183e-01 -9.19202119e-02
-9.99053359e-01 5.73784709e-01 1.34324288e+00 4.93054232e-03
2.77238190e-01 -3.85152400e-01 -9.62715805e-01 2.25412890e-01
6.95061862e-01 -6.94624722e-01 -8.29942152e-02 -1.04095554e+00
6.54182971e-01 3.69923502e-01 -3.09039075e-02 -4.49100375e-01
1.62633017e-01 -4.63126339e-02 -4.73153949e-01 -4.89291340e-01
7.84140304e-02 -2.22060144e-01 -1.56659678e-01 3.92293841e-01
-1.11487269e+00 7.71468937e-01 -9.65541154e-02 3.96163702e-01
-5.48386611e-02 -6.99019849e-01 6.42357409e-01 -4.59690064e-01
-4.82307673e-01 -8.12004209e-02 -3.94392312e-01 1.25311196e+00
6.72833979e-01 1.56872869e-01 -9.86495316e-01 -6.16917968e-01
-2.77888119e-01 6.05890393e-01 7.60631502e-01 9.29299831e-01
2.59148687e-01 -1.13521326e+00 -9.96663570e-01 -3.05270910e-01
4.92884874e-01 -2.24866882e-01 -4.92027961e-02 9.66229439e-02
-6.81144297e-01 5.21141171e-01 1.43585904e-02 -1.96606889e-01
-1.11925542e+00 3.66359502e-01 3.76505136e-01 -6.23259962e-01
-1.58594757e-01 8.01894486e-01 -3.83004248e-01 -1.41005909e+00
-2.26505905e-01 -5.55698574e-01 4.80848877e-03 -2.65632391e-01
6.55893385e-01 7.09073424e-01 2.31162056e-01 -3.21152031e-01
-2.69635588e-01 4.02024627e-01 -2.79458821e-01 -2.79406607e-01
1.12378550e+00 -2.03814685e-01 -7.08550453e-01 2.06603810e-01
1.59672225e+00 -8.47490430e-02 -6.92222834e-01 -5.50795257e-01
6.70676529e-01 -2.74931163e-01 -1.03655481e+00 -9.44800317e-01
-4.70565498e-01 6.07342184e-01 5.57904579e-02 2.03289732e-01
7.50360906e-01 1.93171024e-01 1.35850275e+00 1.13489175e+00
2.22386599e-01 -8.62702012e-01 4.84737605e-01 1.37720799e+00
1.20963371e+00 -1.24073565e+00 -5.14602542e-01 -1.42825648e-01
-6.69282436e-01 8.07774842e-01 9.98880088e-01 -2.92918742e-01
9.21965465e-02 -1.24508411e-01 2.79203534e-01 -2.57570922e-01
-1.20832515e+00 5.71443848e-02 1.74532682e-01 4.67591792e-01
5.33157527e-01 -2.48922125e-01 -4.48929906e-01 6.79367661e-01
-5.75982869e-01 1.09590799e-01 8.85010540e-01 9.61732507e-01
-4.86875594e-01 -1.41721964e+00 -9.48320478e-02 6.22230470e-01
-6.27435863e-01 -2.29179114e-01 -7.11609423e-01 3.90343308e-01
-4.25255746e-01 1.35891116e+00 -1.54945776e-01 -1.62906528e-01
9.03665364e-01 3.20343107e-01 4.54694122e-01 -8.39441895e-01
-1.30073416e+00 -1.02444065e+00 9.56945896e-01 -6.49416327e-01
-3.10258325e-02 -5.03582597e-01 -1.31070805e+00 -4.21230048e-01
-1.53110191e-01 1.18825585e-01 2.63320893e-01 1.01261771e+00
6.24955714e-01 4.06706125e-01 3.57001662e-01 3.53365719e-01
-1.01902616e+00 -9.77242529e-01 2.38485247e-01 7.76175618e-01
3.99423569e-01 1.65240586e-01 -1.74352586e-01 1.75375879e-01]
|
[11.409646034240723, 8.072209358215332]
|
abe7bd98-d5b7-4707-be35-97e3e81fed6a
|
finbert-mrc-financial-named-entity
|
2205.15485
| null |
https://arxiv.org/abs/2205.15485v1
|
https://arxiv.org/pdf/2205.15485v1.pdf
|
FinBERT-MRC: financial named entity recognition using BERT under the machine reading comprehension paradigm
|
Financial named entity recognition (FinNER) from literature is a challenging task in the field of financial text information extraction, which aims to extract a large amount of financial knowledge from unstructured texts. It is widely accepted to use sequence tagging frameworks to implement FinNER tasks. However, such sequence tagging models cannot fully take advantage of the semantic information in the texts. Instead, we formulate the FinNER task as a machine reading comprehension (MRC) problem and propose a new model termed FinBERT-MRC. This formulation introduces significant prior information by utilizing well-designed queries, and extracts start index and end index of target entities without decoding modules such as conditional random fields (CRF). We conduct experiments on a publicly available Chinese financial dataset ChFinAnn and a real-word bussiness dataset AdminPunish. FinBERT-MRC model achieves average F1 scores of 92.78% and 96.80% on the two datasets, respectively, with average F1 gains +3.94% and +0.89% over some sequence tagging models including BiLSTM-CRF, BERT-Tagger, and BERT-CRF. The source code is available at https://github.com/zyz0000/FinBERT-MRC.
|
['Hong Zhang', 'Yuzhe Zhang']
|
2022-05-31
| null | null | null | null |
['machine-reading-comprehension']
|
['natural-language-processing']
|
[-2.47780696e-01 1.35301873e-01 -2.03166142e-01 -5.94138205e-01
-1.03728843e+00 -8.57188880e-01 4.12311822e-01 1.66062012e-01
-7.85441577e-01 9.28209960e-01 3.00344139e-01 -5.53929150e-01
4.80203003e-01 -6.49469733e-01 -7.08620846e-01 -2.42342681e-01
2.20803827e-01 4.36649501e-01 2.23515764e-01 9.38909352e-02
5.67744792e-01 4.97924462e-02 -6.51594043e-01 3.13019991e-01
1.04935920e+00 8.11418712e-01 5.01491904e-01 8.02190065e-01
-5.70055902e-01 1.34224880e+00 -6.94667995e-01 -1.07872498e+00
-2.02375218e-01 -3.40633839e-01 -9.69891846e-01 -3.55496019e-01
-1.51961058e-01 -1.93408176e-01 -1.74024016e-01 1.20842898e+00
3.00099224e-01 -1.22674011e-01 5.79963863e-01 -7.17441976e-01
-9.28813279e-01 9.21783984e-01 -4.44647819e-01 3.81528348e-01
9.27357972e-02 -1.32654205e-01 1.17852294e+00 -9.89903867e-01
6.44286394e-01 7.22878039e-01 4.88416702e-01 6.00929856e-01
-4.47629690e-01 -7.62938380e-01 3.04389223e-02 2.21444845e-01
-1.19877350e+00 -3.17848146e-01 2.67085731e-01 -5.72769821e-01
1.31215000e+00 5.94059229e-02 2.15668917e-01 9.05752957e-01
2.05384418e-01 1.12116838e+00 9.09375250e-01 -4.31979924e-01
6.87915161e-02 3.99347395e-02 4.13307279e-01 5.38201392e-01
2.59727180e-01 -2.38081679e-01 -6.16603494e-01 4.57570478e-02
6.80439770e-01 4.13320251e-02 -1.93496153e-01 5.12876689e-01
-1.14648020e+00 9.19292986e-01 2.19200924e-02 5.88390887e-01
-3.52593780e-01 -1.09025739e-01 4.06883538e-01 -2.38907095e-02
5.86039424e-01 4.30057138e-01 -1.09256470e+00 -5.01491070e-01
-8.00894856e-01 -3.64711545e-02 1.15520084e+00 1.30224562e+00
3.73935670e-01 -5.99576980e-02 -9.37092677e-02 6.29792392e-01
1.68608621e-01 7.43175209e-01 7.18930185e-01 -5.42454839e-01
8.60544622e-01 3.53200853e-01 1.79556668e-01 -7.23219335e-01
-2.33688816e-01 -4.53313559e-01 -6.02747858e-01 -5.76464832e-01
4.43625987e-01 -5.25116563e-01 -7.62863696e-01 1.52826643e+00
1.21048354e-01 1.84501410e-01 4.25310314e-01 6.79133654e-01
8.64640892e-01 8.49485695e-01 4.04665679e-01 -2.20436364e-01
1.54174972e+00 -1.15013242e+00 -7.20064938e-01 -3.23281020e-01
7.36648679e-01 -1.03619421e+00 5.00260830e-01 2.08258212e-01
-1.04357612e+00 -4.00023490e-01 -7.14576125e-01 -1.46473199e-01
-2.85750926e-01 2.83970684e-01 5.57443500e-01 5.49695671e-01
-5.17925024e-01 2.97778100e-01 -1.00373447e+00 4.79935147e-02
2.28212833e-01 1.00975208e-01 -1.68108717e-01 -1.03056729e-01
-1.31822288e+00 7.06569493e-01 7.09487379e-01 8.84727910e-02
-4.42710012e-01 -5.71487188e-01 -7.20724940e-01 8.90661478e-02
3.64161938e-01 -4.03836071e-01 1.76436436e+00 -8.06558788e-01
-1.48236620e+00 7.95631707e-01 -3.75247568e-01 -4.91513550e-01
2.90140331e-01 -7.39620447e-01 -5.81320345e-01 1.22215979e-01
1.75356597e-01 2.34983996e-01 1.76163599e-01 -5.17506421e-01
-6.47474527e-01 -2.26658851e-01 -3.39837939e-01 -2.17196159e-02
-2.08017021e-01 6.87029541e-01 -4.47671801e-01 -9.30929959e-01
-3.88758421e-01 -5.09967089e-01 -1.99592993e-01 -8.30299199e-01
-5.70299506e-01 -3.31558794e-01 2.09248394e-01 -1.51540482e+00
1.38917303e+00 -1.90772486e+00 -4.35951650e-01 -2.07377136e-01
2.00764984e-01 6.65944338e-01 -2.78726518e-02 4.81023967e-01
-1.46230124e-03 3.30817014e-01 -3.03517848e-01 -3.57360989e-02
3.03287208e-02 -4.42898646e-02 -2.30980501e-01 -8.84189084e-02
3.65542531e-01 1.23176801e+00 -9.23197627e-01 -5.18694162e-01
-1.74925074e-01 3.34467113e-01 -2.69240290e-01 5.27637720e-01
-3.89474779e-01 4.73174065e-01 -9.07487929e-01 7.26498723e-01
7.23848522e-01 -4.05490130e-01 3.38544637e-01 2.67288864e-01
-2.54808873e-01 5.99589229e-01 -8.45939696e-01 1.51023901e+00
-1.78763151e-01 1.78624749e-01 -3.58806282e-01 -1.16055393e+00
1.10692894e+00 5.57211161e-01 1.85420498e-01 -8.98652852e-01
1.11049414e-01 4.78021979e-01 -2.66858190e-01 -7.26692379e-01
6.04831278e-01 -2.46421143e-01 -4.19267386e-01 2.21750900e-01
1.85873419e-01 5.38014412e-01 2.59784460e-01 1.94789141e-01
1.29574049e+00 2.82053292e-01 6.17404044e-01 3.95805389e-02
4.14388239e-01 1.65174842e-01 1.02858019e+00 5.50936282e-01
-1.27974069e-02 4.65268105e-01 3.82672548e-01 -1.85170352e-01
-9.86167133e-01 -8.36004436e-01 1.60802770e-02 8.34141672e-01
-2.58316994e-01 -3.59604686e-01 -8.00068617e-01 -9.23843086e-01
-3.02349865e-01 9.90197778e-01 -1.18395582e-01 3.14926118e-01
-8.38408053e-01 -8.75205398e-01 7.32071817e-01 6.84854269e-01
8.28415453e-01 -1.22995389e+00 -4.73723143e-01 5.95023215e-01
-5.56564212e-01 -1.40013742e+00 -6.73519671e-01 1.17690749e-01
-6.90774143e-01 -1.07094514e+00 -1.00184894e+00 -8.91704917e-01
2.56094545e-01 -3.80583048e-01 1.41561699e+00 -1.46100923e-01
7.38530904e-02 -3.03419948e-01 -6.62043512e-01 -4.00804341e-01
-2.33331650e-01 4.01761830e-01 -5.10621846e-01 -2.76334345e-01
8.30153704e-01 -1.45868465e-01 -3.59968543e-01 4.66947816e-02
-6.34436250e-01 1.63366079e-01 9.30556834e-01 8.17250550e-01
5.28440058e-01 -2.07259431e-01 8.73255312e-01 -1.23648632e+00
2.82994092e-01 -6.12230957e-01 -6.25647902e-01 6.67939186e-01
-5.19110620e-01 6.50847629e-02 7.34350502e-01 -2.66532481e-01
-1.38411093e+00 -3.33696231e-02 -4.97610122e-01 -3.41804302e-03
-2.20957592e-01 9.48352695e-01 -3.41324836e-01 6.33977473e-01
6.01484925e-02 3.27976614e-01 -6.72926247e-01 -9.25031006e-01
4.44963649e-02 1.05262816e+00 7.40150988e-01 -4.41092521e-01
3.80400330e-01 -3.00120473e-01 -5.64770997e-01 -5.14566779e-01
-1.41904461e+00 -6.75415516e-01 -6.74946070e-01 1.89841941e-01
1.19169414e+00 -9.55856204e-01 -6.54841185e-01 8.43688011e-01
-1.31195927e+00 -2.75588393e-01 9.55290496e-02 7.14593649e-01
-2.39610627e-01 3.95188302e-01 -1.09433663e+00 -8.73015821e-01
-6.48809969e-01 -6.95755661e-01 6.66655779e-01 5.06384909e-01
-4.69375625e-02 -1.09770429e+00 3.61142159e-02 5.67804694e-01
2.29841873e-01 2.47910157e-01 7.68067360e-01 -1.32036364e+00
-5.31954050e-01 -2.63686955e-01 -2.57039815e-01 3.97588104e-01
-1.37746587e-01 -2.66748130e-01 -7.39806473e-01 1.00801237e-01
-4.17000242e-02 -3.27775925e-01 9.12006617e-01 1.17495164e-01
9.19454813e-01 -5.74698567e-01 -7.70866126e-02 2.35140756e-01
1.44492090e+00 6.57353759e-01 6.09425426e-01 3.78125966e-01
6.78251088e-01 3.57284665e-01 6.55873299e-01 5.20120144e-01
7.69701660e-01 2.35850587e-01 -1.93366244e-01 1.87427551e-01
2.27087006e-01 -5.32432854e-01 5.24807453e-01 1.48557270e+00
9.50311571e-02 -5.06815374e-01 -1.29031026e+00 5.62739670e-01
-1.79840887e+00 -8.50426912e-01 -2.62135029e-01 1.84799516e+00
1.25713527e+00 2.48548552e-01 -1.37858063e-01 -1.77578956e-01
9.39928532e-01 -1.36886209e-01 -5.38018286e-01 -1.54457629e-01
-1.81153908e-01 4.53546941e-01 8.08225214e-01 3.13603401e-01
-1.18764246e+00 1.11368191e+00 4.78660250e+00 8.53426456e-01
-7.72737980e-01 2.04600990e-01 8.69150400e-01 4.54256594e-01
-4.24831025e-02 1.96185619e-01 -1.33447778e+00 7.46660829e-01
1.57455587e+00 -2.01437846e-01 -5.75014874e-02 7.74994671e-01
8.99935812e-02 3.45520824e-02 -7.10740745e-01 6.24730945e-01
-2.91026700e-02 -1.16097784e+00 -2.36337841e-01 -1.42865539e-01
5.47219992e-01 1.38848394e-01 -3.17160338e-01 4.98649687e-01
7.89656103e-01 -1.05882490e+00 8.70292127e-01 9.58629429e-01
7.13478744e-01 -8.15402210e-01 1.14174569e+00 6.16568387e-01
-1.19692266e+00 1.79907680e-01 -3.08442384e-01 1.78393070e-02
4.30895567e-01 8.51646304e-01 -7.52727687e-01 8.97921860e-01
6.23420238e-01 8.66754413e-01 -4.30863023e-01 1.11667418e+00
-5.12087405e-01 1.21317899e+00 -3.76575589e-02 -3.74593675e-01
1.06577821e-01 -8.23800731e-03 9.27518159e-02 1.49160016e+00
3.64195555e-01 4.44383532e-01 2.74484128e-01 5.47254264e-01
-3.80040079e-01 2.78827339e-01 2.21931249e-01 -5.84319770e-01
4.64986205e-01 1.02563429e+00 -5.99159658e-01 -6.49032474e-01
-6.80458009e-01 1.03718495e+00 4.87656295e-01 1.98760316e-01
-8.13669145e-01 -8.00611258e-01 1.58590212e-01 -2.36154661e-01
6.16777241e-01 -3.13622653e-01 -2.53660560e-01 -1.61881065e+00
7.04393908e-02 -7.30800748e-01 5.80608666e-01 -8.46694469e-01
-1.42639911e+00 5.68141460e-01 -2.33986676e-01 -8.77163947e-01
-2.80169725e-01 -5.32202244e-01 -4.41929996e-01 1.08337367e+00
-1.69494069e+00 -1.01827586e+00 8.85014459e-02 2.58862346e-01
5.09340227e-01 -1.68396756e-01 6.81811988e-01 3.80975127e-01
-8.98936093e-01 4.83680665e-01 4.01679784e-01 8.85655046e-01
7.10150361e-01 -1.36945701e+00 7.05132663e-01 1.00976622e+00
8.34226832e-02 7.05266118e-01 3.20737332e-01 -8.60661447e-01
-8.76726031e-01 -1.26513946e+00 1.83319235e+00 -3.75094146e-01
7.18519747e-01 -2.77500421e-01 -1.12574506e+00 8.53624642e-01
3.06524366e-01 -2.89010555e-01 8.50435495e-01 -1.34414628e-01
-4.73134667e-01 4.09291446e-01 -9.42524135e-01 1.37030765e-01
8.81465673e-01 -3.15448731e-01 -8.85616839e-01 2.46550038e-01
9.23508823e-01 -3.92636478e-01 -1.15620363e+00 2.36892760e-01
1.92839593e-01 -6.07030511e-01 6.60866797e-01 -5.82400501e-01
7.07043111e-01 7.31192008e-02 -6.38203323e-02 -8.25842381e-01
-1.52892306e-01 -4.38719600e-01 -8.37257579e-02 1.68713951e+00
7.49291122e-01 -6.86141133e-01 5.50239563e-01 6.54735982e-01
-1.89614221e-01 -5.86916804e-01 -7.67117202e-01 -8.20726573e-01
4.65030700e-01 -3.51048678e-01 6.05629742e-01 1.14788008e+00
1.63909644e-01 5.26335597e-01 -2.61321932e-01 8.73223096e-02
3.36127669e-01 1.09833449e-01 2.86640823e-01 -9.59933698e-01
-4.68035519e-01 -1.40295792e-02 4.29053651e-03 -1.35824072e+00
3.22552115e-01 -8.27203274e-01 6.45664195e-03 -1.60930645e+00
5.05555212e-01 -3.60274047e-01 -4.29152101e-01 7.56192625e-01
-4.47967947e-01 -2.78898180e-01 7.78348520e-02 4.60144311e-01
-8.94237399e-01 4.95867014e-01 1.02927172e+00 1.61853641e-01
1.00102141e-01 -4.58198180e-03 -6.19377553e-01 5.90772212e-01
1.12419868e+00 -5.73542237e-01 1.11637905e-01 -5.28525889e-01
2.04946741e-01 5.48744023e-01 7.54130259e-02 -5.43676913e-01
1.30991742e-01 -3.75310928e-01 5.11802018e-01 -6.68178201e-01
-2.53814191e-01 -1.95242599e-01 -1.30224153e-01 4.28642690e-01
-3.10722560e-01 3.36118251e-01 -6.26664162e-02 3.62029731e-01
-4.58830506e-01 -6.24568105e-01 4.75891292e-01 -4.85148132e-01
-8.91141117e-01 6.45658374e-02 -3.52478117e-01 5.85903525e-01
7.56064832e-01 2.53816932e-01 -4.84899968e-01 -3.10120508e-02
-5.61290264e-01 2.46541634e-01 -2.17358377e-02 3.39532822e-01
4.33049440e-01 -8.55379939e-01 -8.60241413e-01 -6.44237697e-02
-5.53588662e-03 4.63363640e-02 1.91387132e-01 6.29160047e-01
-5.20588815e-01 8.90979469e-01 1.47555232e-01 6.05782792e-02
-1.05491138e+00 4.46775019e-01 5.48561253e-02 -8.39688599e-01
-4.40510750e-01 9.31890011e-01 1.16900906e-01 -4.98692751e-01
-1.85180724e-01 -2.78549373e-01 -3.89113367e-01 -2.72270262e-01
5.90235114e-01 1.45403072e-01 1.79531023e-01 -5.69106936e-01
-3.02642584e-01 2.29894936e-01 -3.65404189e-01 -4.64190692e-02
1.38736820e+00 -1.24547206e-01 6.02635890e-02 2.56535828e-01
1.01233923e+00 1.31020248e-01 -1.01018858e+00 -4.04059291e-01
8.60867321e-01 -1.29750431e-01 -2.72702157e-01 -1.13955569e+00
-9.65841472e-01 8.24059725e-01 -1.05164640e-01 -1.65986866e-01
1.00702488e+00 5.74207455e-02 1.27025437e+00 3.36261094e-01
2.90270627e-01 -1.00709045e+00 -1.67093992e-01 9.46078837e-01
4.22945708e-01 -9.78235245e-01 -4.54951316e-01 -4.88899589e-01
-8.95181656e-01 1.03608418e+00 4.71496373e-01 1.18573226e-01
5.98698378e-01 3.48772407e-01 1.55964777e-01 -7.62250368e-03
-6.88319683e-01 -1.09249331e-01 -1.71593670e-02 1.95811540e-01
7.86183357e-01 5.90664223e-02 -5.12337565e-01 1.38231099e+00
-4.14287329e-01 2.11536154e-01 4.52059627e-01 1.19982350e+00
-5.91092765e-01 -1.29767597e+00 -1.71723664e-02 4.65636849e-01
-1.36506748e+00 -5.51467896e-01 -1.41426325e-01 5.31751454e-01
-1.31592676e-01 1.00914395e+00 -3.93999554e-02 -2.34048396e-01
2.66680777e-01 4.55778629e-01 -5.07068001e-02 -7.10627317e-01
-7.27845192e-01 2.31042176e-01 3.35804641e-01 -3.88837047e-02
-4.97850716e-01 -9.44420159e-01 -1.61717319e+00 -1.49473533e-01
-3.89305115e-01 6.47326887e-01 4.23291981e-01 1.07410419e+00
3.43156725e-01 4.08185571e-01 4.68012869e-01 8.01763013e-02
-5.97452998e-01 -1.13526654e+00 -5.64661324e-01 1.71336368e-01
-6.18910417e-02 -2.64778346e-01 -5.11821695e-02 6.31457746e-01]
|
[9.82951545715332, 9.649425506591797]
|
02d31dbe-0e3a-4c07-926c-8e9a6c2c3133
|
concept-extraction-using-pointer-generator
|
2008.11295
| null |
https://arxiv.org/abs/2008.11295v1
|
https://arxiv.org/pdf/2008.11295v1.pdf
|
Concept Extraction Using Pointer-Generator Networks
|
Concept extraction is crucial for a number of downstream applications. However, surprisingly enough, straightforward single token/nominal chunk-concept alignment or dictionary lookup techniques such as DBpedia Spotlight still prevail. We propose a generic open-domain OOV-oriented extractive model that is based on distant supervision of a pointer-generator network leveraging bidirectional LSTMs and a copy mechanism. The model has been trained on a large annotated corpus compiled specifically for this task from 250K Wikipedia pages, and tested on regular pages, where the pointers to other pages are considered as ground truth concepts. The outcome of the experiments shows that our model significantly outperforms standard techniques and, when used on top of DBpedia Spotlight, further improves its performance. The experiments furthermore show that the model can be readily ported to other datasets on which it equally achieves a state-of-the-art performance.
|
['Alexander Shvets', 'Leo Wanner']
|
2020-08-25
| null | null | null | null |
['concept-alignment']
|
['computer-vision']
|
[-2.51874894e-01 6.76825702e-01 -4.83650565e-01 2.77805813e-02
-8.95043850e-01 -7.12677240e-01 9.93064165e-01 6.28304064e-01
-7.57676780e-01 1.09640527e+00 3.96361321e-01 -5.89275897e-01
4.01160233e-02 -1.11837661e+00 -1.10188520e+00 -1.89645439e-01
-5.77021837e-01 8.12514067e-01 4.65863615e-01 -4.80978161e-01
2.98135370e-01 -1.08483359e-01 -1.54592824e+00 5.82400620e-01
6.91647530e-01 9.48731303e-01 1.00947946e-01 3.26092452e-01
-6.75281405e-01 8.66858721e-01 -5.30617476e-01 -8.10979247e-01
5.52643776e-01 3.16824585e-01 -1.24569261e+00 -7.51897752e-01
9.41292822e-01 -2.00295702e-01 -1.61631793e-01 8.42461407e-01
4.46654677e-01 -1.08176591e-02 3.70086879e-01 -8.48238468e-01
-8.65295649e-01 1.21859515e+00 -1.38929591e-01 1.92040801e-01
3.32125336e-01 -2.76349574e-01 1.53886366e+00 -1.10817862e+00
9.77864444e-01 7.14860678e-01 9.00991380e-01 3.71611744e-01
-1.12824535e+00 -5.38680553e-01 2.24417280e-02 2.40703151e-01
-1.20555437e+00 -4.68964130e-01 6.57039136e-02 -3.59425545e-01
1.69202924e+00 -1.16062120e-01 5.12730956e-01 1.11534810e+00
2.04212479e-02 7.21623898e-01 6.75228179e-01 -7.54292965e-01
9.10109505e-02 3.08619261e-01 1.91567972e-01 8.18217695e-01
6.53848648e-01 -8.83149635e-03 -6.23557389e-01 -2.45332792e-02
2.55481064e-01 -4.78890717e-01 -3.36733729e-01 -6.23116672e-01
-1.47337413e+00 6.55790329e-01 4.70837027e-01 2.80641735e-01
-2.34412819e-01 4.07515615e-02 8.30095768e-01 3.46241564e-01
4.48788941e-01 7.91456699e-01 -8.62219989e-01 -4.63495776e-02
-1.26794088e+00 4.90240544e-01 1.27513278e+00 1.46425068e+00
7.17582762e-01 -4.44305509e-01 -2.41273671e-01 6.99070156e-01
6.48577437e-02 1.75811827e-01 6.75070226e-01 -8.19152057e-01
1.01564312e+00 6.49579406e-01 2.97961012e-02 -7.20972836e-01
-2.95037270e-01 -2.96571374e-01 -2.07768857e-01 -1.64498985e-01
5.05603969e-01 -1.27601624e-01 -9.36138511e-01 1.43799150e+00
9.47718993e-02 4.95417677e-02 2.98476964e-01 6.54312372e-01
1.13341808e+00 5.19046068e-01 2.30502442e-01 4.96823378e-02
1.52228105e+00 -9.67954934e-01 -3.91409785e-01 -3.06644648e-01
8.35717857e-01 -3.35380435e-01 8.91524732e-01 4.26988930e-01
-1.08000064e+00 -3.94674182e-01 -1.22213757e+00 -4.19064552e-01
-1.10468340e+00 -2.35139981e-01 7.29038894e-01 3.99530739e-01
-1.13254893e+00 6.41818285e-01 -5.87123215e-01 -6.43987060e-01
4.02144969e-01 2.22830951e-01 -5.33200204e-01 -1.24742530e-01
-1.57130551e+00 1.16171634e+00 1.16209948e+00 -1.02128536e-01
-6.36577487e-01 -7.96734214e-01 -8.58671129e-01 4.59740721e-02
5.96993029e-01 -8.91219139e-01 1.18628550e+00 -6.14685774e-01
-1.07950664e+00 1.05198193e+00 2.33758435e-01 -1.08540702e+00
2.43224546e-01 -5.44039726e-01 -4.24955726e-01 9.44840536e-02
4.41505700e-01 8.57704103e-01 3.51419985e-01 -1.05631948e+00
-1.03134358e+00 -1.33147329e-01 3.42475206e-01 1.15580395e-01
-6.61649883e-01 -9.41610634e-02 -6.41331196e-01 -5.40906608e-01
-2.37718880e-01 -5.61273217e-01 -5.52421762e-03 -3.90487105e-01
-7.78220356e-01 -1.51376382e-01 2.21772224e-01 -9.23052073e-01
1.40806663e+00 -1.62819886e+00 1.69535562e-01 2.44235709e-01
2.20763385e-01 2.69359499e-01 -7.85593614e-02 6.68339968e-01
-2.67588720e-02 1.13786377e-01 -1.75314233e-01 6.38401508e-02
1.08291350e-01 1.17422670e-01 -5.55491865e-01 -7.48877972e-02
2.10624009e-01 1.03202677e+00 -1.09671283e+00 -5.93790114e-01
-3.51511240e-01 -4.73144092e-02 -5.79039216e-01 3.53514329e-02
-7.10463464e-01 -3.83869141e-01 -1.36722609e-01 6.24880314e-01
2.90018260e-01 -2.42791504e-01 4.94043291e-01 -3.47989388e-02
-1.62337840e-01 8.23251784e-01 -9.80700314e-01 2.07194233e+00
-4.75750744e-01 7.20963955e-01 -2.30538294e-01 -8.92053306e-01
6.80804670e-01 3.37071508e-01 1.91685304e-01 -8.36317420e-01
-3.41380417e-01 3.92236054e-01 -1.38265178e-01 -5.48976183e-01
1.15978003e+00 2.36849234e-01 -1.03960045e-01 2.41726354e-01
5.67582130e-01 5.19988179e-01 6.69284701e-01 5.82461953e-01
1.26481092e+00 6.85931921e-01 5.25585175e-01 -4.35160935e-01
5.00752807e-01 4.32931870e-01 2.01601088e-01 7.76456058e-01
3.45493525e-01 1.86922282e-01 4.47317094e-01 -4.50586557e-01
-1.16771746e+00 -9.78943408e-01 -2.18982711e-01 1.38752913e+00
-6.36803582e-02 -8.23395550e-01 -6.07004344e-01 -1.04629672e+00
3.57961148e-01 8.31814706e-01 -5.42919278e-01 1.60920188e-01
-6.74594045e-01 -5.37381649e-01 8.35191965e-01 6.76461518e-01
3.41830134e-01 -1.18959594e+00 -4.11208302e-01 5.56379020e-01
-1.46838680e-01 -1.30297840e+00 1.97452635e-01 4.67205346e-01
-7.17269480e-01 -1.05241776e+00 -5.59399724e-01 -9.36254859e-01
3.03955168e-01 -1.07447252e-01 1.73320043e+00 -9.15069655e-02
-7.83059001e-03 1.48349673e-01 -4.80291128e-01 -4.56549495e-01
-2.51708955e-01 7.44291604e-01 -7.64650777e-02 -7.91494131e-01
8.99776936e-01 -4.13209200e-01 -2.66103774e-01 -1.91636071e-01
-6.30615056e-01 -1.89627290e-01 5.47664702e-01 9.97790754e-01
4.80896771e-01 -1.97594821e-01 7.72035897e-01 -1.33961189e+00
5.51424742e-01 -6.70920372e-01 -6.51677430e-01 2.96023548e-01
-9.43981111e-01 3.59577388e-01 6.33414686e-01 3.16608176e-02
-9.39761639e-01 -1.65590495e-01 -3.19014415e-02 8.45786408e-02
-1.62000209e-01 9.62506235e-01 -1.79883331e-01 3.39362383e-01
9.65807736e-01 1.18419111e-01 -2.80027598e-01 -6.23952746e-01
9.02150512e-01 7.49021113e-01 8.45873713e-01 -8.22835088e-01
7.76724041e-01 2.53752887e-01 -4.32866305e-01 -6.07527375e-01
-9.74995971e-01 -4.26271588e-01 -7.45730340e-01 3.92908543e-01
5.80895603e-01 -1.25806451e+00 -3.42107892e-01 -5.57552204e-02
-1.04631734e+00 -4.46710080e-01 -2.56444633e-01 9.96225029e-02
-3.64469677e-01 2.34058201e-01 -8.03633630e-01 -2.19013110e-01
-7.31888652e-01 -4.70168561e-01 7.92500973e-01 1.28995003e-02
-6.45744875e-02 -1.10124564e+00 1.44617170e-01 1.85971275e-01
2.18946695e-01 -1.08166382e-01 1.11378300e+00 -1.27339447e+00
-7.87275016e-01 -1.53583303e-01 -2.64183581e-01 1.56100333e-01
-2.51461506e-01 -4.96139564e-02 -8.49445999e-01 -9.59374085e-02
-8.63554537e-01 -5.38279295e-01 1.15839529e+00 -9.21126157e-02
8.17018807e-01 -6.44704998e-01 -6.58410072e-01 4.71075088e-01
1.61691236e+00 -3.19128603e-01 6.18323147e-01 1.16158235e+00
6.86264455e-01 6.84876621e-01 6.37287974e-01 1.07582673e-01
8.50775123e-01 6.05509102e-01 4.30401146e-01 3.02722663e-01
-1.24060795e-01 -4.76582557e-01 3.22426021e-01 8.30031633e-01
-8.39595720e-02 -1.68435529e-01 -1.15467119e+00 9.50564623e-01
-1.68345654e+00 -1.27046788e+00 -1.14348292e-01 2.39718699e+00
1.28840160e+00 4.91889656e-01 1.21106952e-01 -1.89340055e-01
4.17383432e-01 -3.66834886e-02 -9.32234004e-02 -4.71721828e-01
-8.26996751e-03 5.05833209e-01 9.00984704e-01 2.37505496e-01
-1.21151102e+00 1.24619055e+00 6.55784750e+00 6.59073293e-01
-8.92926693e-01 1.34199530e-01 1.47005508e-03 -1.91827416e-01
-1.92526951e-01 1.23561561e-01 -1.24589312e+00 5.41064858e-01
1.13853049e+00 -2.80307710e-01 3.37479830e-01 1.01145434e+00
-4.86092776e-01 3.47183272e-02 -1.39527416e+00 4.01857615e-01
2.62656752e-02 -1.64461255e+00 9.58870426e-02 1.64510384e-01
3.96373779e-01 4.36008245e-01 -3.58372778e-01 8.61516178e-01
6.54869139e-01 -1.00068808e+00 7.91035354e-01 2.66740263e-01
8.53590727e-01 -5.84763587e-01 8.94477069e-01 2.16416270e-01
-1.04033124e+00 -3.03433910e-02 -5.13691366e-01 8.85864627e-03
4.24605906e-02 5.12193680e-01 -1.23395085e+00 8.54113281e-01
7.43671954e-01 7.46357322e-01 -6.50078416e-01 1.05699301e+00
-4.71163303e-01 3.95130843e-01 -1.73477039e-01 -4.25847508e-02
2.85347968e-01 3.55081201e-01 3.74902099e-01 1.72037327e+00
1.85319126e-01 -3.64109904e-01 -3.21699679e-02 7.29492128e-01
-4.50885773e-01 3.23081315e-01 -7.50290632e-01 -1.47610784e-01
9.08243775e-01 1.35121202e+00 -3.26862305e-01 -6.03063881e-01
-7.93634653e-01 6.00036561e-01 8.57281566e-01 2.57346302e-01
-6.86592817e-01 -8.27722549e-01 4.97932017e-01 -6.36209995e-02
8.60338628e-01 -1.53897703e-02 -1.16794202e-02 -1.23542094e+00
2.11920306e-01 -7.81462193e-01 6.95432484e-01 -5.61545551e-01
-1.39588368e+00 5.86548626e-01 8.17684233e-02 -9.66177583e-01
-5.92893004e-01 -8.81025612e-01 -3.89924347e-01 7.43281007e-01
-1.78997636e+00 -1.09604299e+00 5.00712022e-02 3.59369010e-01
2.20577449e-01 -4.08081412e-01 1.02590358e+00 5.18547416e-01
-4.48126644e-01 6.02393329e-01 1.40440807e-01 6.69302940e-01
8.64088416e-01 -1.64979863e+00 7.74192750e-01 9.06100094e-01
4.20028895e-01 1.09191871e+00 5.60596883e-01 -6.47445977e-01
-1.16613066e+00 -1.18477857e+00 1.28141499e+00 -6.24208629e-01
9.80224073e-01 -5.83306909e-01 -9.50093329e-01 1.09231281e+00
5.37637591e-01 2.69940197e-02 7.47042894e-01 6.28587544e-01
-7.37630785e-01 -8.37906003e-02 -6.76345646e-01 3.40741456e-01
1.14218390e+00 -3.84117633e-01 -1.32868314e+00 3.62592638e-01
9.39816773e-01 -6.26120210e-01 -1.25053942e+00 1.40639305e-01
6.10940158e-01 -7.08670914e-01 1.12305164e+00 -9.66271400e-01
8.78859758e-01 -1.15873955e-01 -1.37758553e-01 -1.34157145e+00
-1.14208885e-01 -2.63016611e-01 -8.03912580e-01 1.44952655e+00
9.54356670e-01 -4.84666675e-01 8.43014300e-01 3.92093271e-01
-1.91551551e-01 -5.57591677e-01 -8.40407550e-01 -1.04015005e+00
3.50305229e-01 -3.14850897e-01 5.35617590e-01 1.08561695e+00
6.23291910e-01 7.12228298e-01 -3.53722982e-02 5.29262833e-02
5.67467690e-01 1.93076447e-01 6.99216187e-01 -1.45354474e+00
-3.59978586e-01 -3.79987985e-01 -3.32122326e-01 -9.30134118e-01
3.26324552e-01 -1.35286915e+00 2.46480592e-02 -1.80361319e+00
2.34154146e-02 -6.45228684e-01 -3.41605961e-01 8.78446341e-01
-1.38537824e-01 8.86949375e-02 5.37787341e-02 8.64246022e-03
-7.46859252e-01 1.96916297e-01 6.09474719e-01 -2.74546474e-01
-1.35219723e-01 -5.45754194e-01 -9.53916490e-01 3.44314843e-01
5.42954624e-01 -5.10923564e-01 -5.60965389e-02 -5.88291228e-01
6.73151195e-01 -2.63798356e-01 1.63512200e-01 -9.89893615e-01
5.16463518e-01 3.46722156e-01 2.63783008e-01 -2.88736105e-01
5.60276583e-02 -6.59524441e-01 -4.90467578e-01 1.20605394e-01
-5.23778439e-01 -1.96984503e-02 3.71506393e-01 5.83781719e-01
-2.31420770e-01 -4.57788616e-01 1.10140853e-01 -4.21466112e-01
-1.12789083e+00 8.97760093e-02 -2.21786462e-02 7.74337411e-01
7.52136290e-01 -7.83672556e-02 -7.59333849e-01 1.39016673e-01
-3.79831046e-01 3.17911685e-01 3.90616864e-01 4.11478609e-01
2.42242083e-01 -1.23496628e+00 -6.58275068e-01 -9.39404145e-02
5.81847727e-01 4.60560396e-02 -4.84680355e-01 5.84469199e-01
-4.79355633e-01 9.66834247e-01 -2.52689779e-01 -1.82164937e-01
-8.26496422e-01 8.35823476e-01 9.53627080e-02 -4.24833417e-01
-8.08734596e-01 7.70493388e-01 -3.46179545e-01 -6.71007872e-01
3.68073493e-01 -3.67585301e-01 -3.71576637e-01 3.08742315e-01
5.34330010e-01 1.38351008e-01 6.15238547e-01 -3.62984121e-01
-3.78170699e-01 -1.86534785e-02 -3.22488099e-01 -2.99390638e-03
1.50156748e+00 -9.69158672e-03 -2.65481234e-01 2.28408337e-01
9.05887544e-01 2.45817080e-01 -7.88468599e-01 -4.24522370e-01
6.98127687e-01 -9.80678052e-02 -1.56029955e-01 -1.11284149e+00
-8.50775838e-01 6.41051888e-01 9.13665667e-02 2.49635801e-01
6.13423526e-01 -1.53053254e-01 7.43367374e-01 9.46053863e-01
7.10418880e-01 -1.19187975e+00 -3.28854561e-01 7.74839759e-01
4.49685007e-01 -1.09825361e+00 8.45929906e-02 -1.25313014e-01
-4.24256563e-01 1.21799254e+00 7.00154960e-01 -7.26874545e-02
2.94835687e-01 3.42915535e-01 -1.28447175e-01 -1.74282938e-01
-9.52556670e-01 -3.90255362e-01 1.56896576e-01 7.28064060e-01
6.94903791e-01 -2.34136939e-01 -3.06427658e-01 5.94443500e-01
-3.20444614e-01 1.60184905e-01 5.82676232e-01 8.94104540e-01
-4.74782676e-01 -1.24524713e+00 2.48786900e-02 5.85531890e-01
-6.74104512e-01 -9.41933393e-01 -9.66925919e-02 1.23507667e+00
2.48201430e-01 3.75063956e-01 1.76867872e-01 -1.88043162e-01
2.70675838e-01 4.93662119e-01 4.09610480e-01 -9.82730508e-01
-8.44263554e-01 -6.49731219e-01 7.27932990e-01 -6.27714574e-01
-3.74218404e-01 -4.80183333e-01 -1.31181300e+00 -1.95929617e-01
-2.77583271e-01 2.88955599e-01 5.55544913e-01 7.39038110e-01
6.05123520e-01 2.62303293e-01 -2.80934155e-01 -6.09011590e-01
-5.06612301e-01 -1.01123524e+00 -4.40923423e-01 2.39797831e-01
8.08410347e-02 -5.41450858e-01 4.00526971e-02 1.68288559e-01]
|
[9.502586364746094, 8.572388648986816]
|
94914d7c-0888-43f8-a2c8-6487a0999e4d
|
ticket-bert-labeling-incident-management
|
2307.00108
| null |
https://arxiv.org/abs/2307.00108v1
|
https://arxiv.org/pdf/2307.00108v1.pdf
|
Ticket-BERT: Labeling Incident Management Tickets with Language Models
|
An essential aspect of prioritizing incident tickets for resolution is efficiently labeling tickets with fine-grained categories. However, ticket data is often complex and poses several unique challenges for modern machine learning methods: (1) tickets are created and updated either by machines with pre-defined algorithms or by engineers with domain expertise that share different protocols, (2) tickets receive frequent revisions that update ticket status by modifying all or parts of ticket descriptions, and (3) ticket labeling is time-sensitive and requires knowledge updates and new labels per the rapid software and hardware improvement lifecycle. To handle these issues, we introduce Ticket- BERT which trains a simple yet robust language model for labeling tickets using our proposed ticket datasets. Experiments demonstrate the superiority of Ticket-BERT over baselines and state-of-the-art text classifiers on Azure Cognitive Services. We further encapsulate Ticket-BERT with an active learning cycle and deploy it on the Microsoft IcM system, which enables the model to quickly finetune on newly-collected tickets with a few annotations.
|
['Siduo Jiang', 'Cris Benge', 'Zhexiong Liu']
|
2023-06-30
| null | null | null | null |
['active-learning', 'management', 'active-learning']
|
['methodology', 'miscellaneous', 'natural-language-processing']
|
[ 1.99910015e-01 -1.14413515e-01 -6.41346753e-01 -8.96189570e-01
-8.68202090e-01 -3.72369081e-01 4.78228509e-01 3.93264979e-01
-5.89709997e-01 9.02911007e-01 8.53457768e-03 -2.96988279e-01
-5.32287478e-01 -6.33713365e-01 -4.78425294e-01 -2.72461712e-01
-2.27894157e-01 1.53560460e+00 4.28937107e-01 -2.75744945e-01
1.95997506e-01 1.22844473e-01 -1.56313562e+00 5.80524266e-01
5.30849278e-01 1.34717643e+00 -3.13927352e-01 3.28492463e-01
-2.84622937e-01 9.70926762e-01 -9.54288065e-01 -2.62705505e-01
4.53761756e-01 7.41172075e-01 -1.08870828e+00 -1.47480443e-01
5.00708282e-01 -4.18608099e-01 5.36925066e-03 3.06622416e-01
1.36140734e-01 9.06758979e-02 6.00555241e-01 -1.69717419e+00
5.67033468e-03 9.81298208e-01 -5.28590083e-01 5.52321315e-01
-2.54962556e-02 5.43904155e-02 8.45457137e-01 -5.85047424e-01
2.52283633e-01 9.71081257e-01 7.86775589e-01 3.12575579e-01
-1.24827373e+00 -1.32173693e+00 6.18989944e-01 6.93575025e-01
-1.66681707e+00 -8.51582289e-01 1.56690150e-01 -4.56488848e-01
1.32485020e+00 5.28204203e-01 -4.28673439e-02 1.21991086e+00
-6.43757507e-02 5.28147995e-01 8.49849880e-01 -6.14732325e-01
3.27234238e-01 2.92383641e-01 8.99018407e-01 5.84023416e-01
1.87733978e-01 1.46995917e-01 -8.16424429e-01 -3.95658970e-01
1.12286489e-02 2.79760122e-01 1.78014413e-02 6.50964081e-02
-9.50173438e-01 7.43975699e-01 6.33084178e-02 7.31255263e-02
-4.61027890e-01 7.48820603e-02 5.05285978e-01 4.50312197e-01
6.08702540e-01 3.83254379e-01 -7.47872531e-01 -2.34467044e-01
-1.09305370e+00 -9.03039426e-02 8.82573724e-01 9.97475088e-01
8.46942723e-01 -1.78376779e-01 -3.69664878e-01 9.43195224e-01
1.19437091e-01 3.26829761e-01 -7.00491369e-02 -4.86441195e-01
5.03206491e-01 5.31983435e-01 2.33577102e-01 -6.12861753e-01
-4.32241231e-01 -6.33871555e-01 -6.44769967e-01 -9.08352807e-02
6.72752932e-02 3.58825140e-02 -1.01909852e+00 1.44826984e+00
1.32729247e-01 5.53528249e-01 -7.89268732e-01 5.96866310e-01
3.56240153e-01 3.78290772e-01 3.09701264e-01 -2.63116241e-01
1.46393383e+00 -7.89478242e-01 -6.74370050e-01 -5.94005346e-01
3.91504526e-01 -6.26973450e-01 9.71597850e-01 6.36636436e-01
-5.37634254e-01 -1.56476766e-01 -1.02379858e+00 5.08291185e-01
-4.76037472e-01 -3.50747287e-01 7.32435882e-01 8.48331988e-01
-8.46302092e-01 -1.09955929e-01 -8.22898328e-01 -3.40885073e-01
5.97665846e-01 5.21097302e-01 2.91352898e-01 -1.19439848e-01
-1.23297799e+00 1.08862758e+00 2.30105549e-01 -9.16796848e-02
-1.02314544e+00 -9.24815953e-01 -3.68615329e-01 1.09471224e-01
9.75710332e-01 -5.58730245e-01 1.84089947e+00 -3.81424576e-01
-1.08536506e+00 6.33589625e-01 -1.53135046e-01 -4.60468620e-01
3.46502274e-01 -1.46592245e-01 -8.90663028e-01 -1.87106848e-01
4.32805687e-01 2.00324506e-01 1.11401761e+00 -8.86054575e-01
-1.25423169e+00 -2.65499502e-01 -5.33073805e-02 -1.44616768e-01
-5.30697107e-01 -3.19844414e-03 -4.68363255e-01 -3.05186182e-01
-3.41254890e-01 -7.05611527e-01 -1.30153745e-01 -6.50367618e-01
-2.80262381e-01 -3.58217180e-01 1.03960526e+00 -2.74470687e-01
1.58868742e+00 -1.78093219e+00 -6.40280485e-01 5.82125008e-01
5.25032938e-01 6.52679801e-02 2.63599962e-01 1.51172563e-01
9.59547237e-02 1.20703124e-01 3.48190904e-01 -2.54480332e-01
2.86329687e-01 3.23506922e-01 -7.57701814e-01 -4.82204258e-02
-2.69737095e-01 4.42178726e-01 -6.53845668e-01 -4.93964046e-01
1.68049976e-01 2.66736355e-02 -4.58052754e-01 -1.08647324e-01
-3.04496378e-01 2.53614902e-01 -3.34574699e-01 8.19175780e-01
4.04934764e-01 -5.92658639e-01 2.24797428e-01 -1.80908546e-01
-2.73615152e-01 7.92252362e-01 -1.02137339e+00 9.70539868e-01
-6.58501685e-01 2.82511473e-01 1.70200780e-01 -1.23614132e+00
7.22659111e-01 2.66838819e-01 6.44257009e-01 -7.72906542e-01
4.24387418e-02 2.63185613e-02 -8.33550930e-01 -2.51820385e-01
4.66218799e-01 6.91770837e-02 -5.55119216e-01 1.25887883e+00
-1.50537163e-01 4.11324739e-01 1.67414024e-01 1.70330301e-01
1.47785366e+00 -6.92802668e-01 1.45560741e-01 1.65598437e-01
-6.71407357e-02 4.25060868e-01 6.31923199e-01 1.30506134e+00
-3.58708560e-01 -2.94575781e-01 -1.33451656e-01 -8.51905167e-01
-6.31964147e-01 -8.78665626e-01 -3.73124689e-01 1.93167186e+00
-1.56258404e-01 -4.33084875e-01 -3.52870107e-01 -1.03987384e+00
-3.49034220e-02 9.23220873e-01 -2.28298858e-01 -1.03371412e-01
-4.82626230e-01 -8.15766752e-01 6.45451128e-01 4.73951250e-01
7.40545273e-01 -6.52343214e-01 -5.47258317e-01 5.29210210e-01
-7.59488344e-01 -1.11497724e+00 -6.43382728e-01 7.66741335e-01
-2.75882930e-01 -9.20600533e-01 1.11588754e-01 -6.48058534e-01
3.73073220e-01 2.92282581e-01 1.38487411e+00 2.43628144e-01
-4.23982352e-01 4.74824905e-01 -4.50800121e-01 -8.02726567e-01
4.98789884e-02 4.58871305e-01 3.14215362e-01 -7.40495650e-03
1.03400481e+00 -2.08940759e-01 -2.99926579e-01 9.42291141e-01
-6.26367152e-01 -1.51775867e-01 4.06232178e-01 6.96066082e-01
3.83927315e-01 7.02992797e-01 7.02487230e-01 -1.15656352e+00
5.77433467e-01 -5.54556251e-01 -7.68367290e-01 6.40017450e-01
-1.17492342e+00 -5.64777777e-02 -1.97479007e-04 -4.30944920e-01
-9.51846421e-01 -1.47556603e-01 2.48774245e-01 1.72655553e-01
-2.47279063e-01 5.97154140e-01 2.33400285e-01 2.19879746e-01
9.71977472e-01 -3.31374645e-01 -4.06085789e-01 -3.56910795e-01
7.85960779e-02 1.24955904e+00 5.65724134e-01 -7.49026358e-01
8.95302832e-01 5.24183750e-01 -7.46007383e-01 -2.74015039e-01
-1.31572342e+00 -7.95817554e-01 -4.42622632e-01 -4.42791641e-01
4.68074262e-01 -5.97959459e-01 -1.01191032e+00 3.03787172e-01
-8.63611460e-01 -5.90781927e-01 2.05481425e-01 3.40324223e-01
-1.76604435e-01 -1.66064963e-01 -5.81544340e-01 -6.20836675e-01
-6.23508990e-01 -7.70596504e-01 9.28001285e-01 -1.99814308e-02
-6.80414796e-01 -7.18379021e-01 5.41509967e-03 9.70053852e-01
8.16350698e-01 -2.80290395e-01 1.21220922e+00 -8.26542974e-01
-8.42131138e-01 -5.40631652e-01 -1.39085144e-01 -2.39099309e-01
6.34583533e-02 -4.70977485e-01 -1.04002190e+00 -4.62476730e-01
-2.57641822e-01 -4.25701857e-01 5.55133998e-01 2.33015940e-01
1.20842552e+00 -4.93196756e-01 -8.75462532e-01 3.45403969e-01
8.04048896e-01 2.87961394e-01 3.40838209e-02 7.76402354e-01
2.78711170e-01 4.76647288e-01 9.03340161e-01 6.58235610e-01
6.00304425e-01 1.03084159e+00 2.32885420e-01 5.10521196e-02
3.59417289e-01 2.27167204e-01 1.16349071e-01 5.15716672e-01
1.53712913e-01 -1.10489585e-01 -1.48027086e+00 2.01533675e-01
-2.03260827e+00 -1.19081068e+00 1.82293564e-01 2.32117820e+00
1.02891040e+00 8.45272779e-01 7.79931992e-02 3.92092139e-01
9.54588056e-01 -2.04747155e-01 -6.75204456e-01 -2.66544163e-01
4.66448605e-01 5.51103234e-01 8.75376701e-01 3.55128258e-01
-1.26313329e+00 1.02627873e+00 6.59315538e+00 7.64889359e-01
-1.26478827e+00 4.24642026e-01 4.80740547e-01 -4.69008088e-01
2.81874239e-01 -9.09083039e-02 -1.16896021e+00 3.85797352e-01
1.10633361e+00 -1.12325333e-01 7.86687315e-01 9.98969257e-01
1.80062782e-02 2.23335370e-01 -1.31393826e+00 1.08954358e+00
-5.35917655e-02 -1.74567914e+00 -2.75101095e-01 7.90405348e-02
2.71318287e-01 2.81637579e-01 2.85991076e-02 1.15485024e+00
7.40989268e-01 -9.44698274e-01 9.61719334e-01 4.63719726e-01
8.99494231e-01 -8.57605934e-01 6.47659361e-01 4.86681312e-01
-1.19131577e+00 -4.49709415e-01 4.79017422e-02 1.64482281e-01
3.51678580e-01 7.17524588e-01 -1.75037980e+00 2.31155381e-01
1.19283795e+00 1.59575805e-01 -5.23501277e-01 1.11314368e+00
7.35218674e-02 1.09955668e+00 -4.74594593e-01 1.90345868e-01
-1.18275709e-01 4.46067840e-01 2.57818341e-01 1.20428133e+00
-9.68408734e-02 -5.03423214e-02 7.33810127e-01 4.73607421e-01
1.08689405e-01 -5.30005395e-01 2.52996664e-02 1.19291708e-01
1.33435702e+00 1.21446633e+00 -6.66107476e-01 -5.02031684e-01
-3.45698416e-01 3.74661446e-01 1.43106043e-01 4.90755975e-01
-9.12460864e-01 -4.04737741e-01 5.56895733e-01 6.17447078e-01
2.29421571e-01 -1.28549546e-01 -3.51610869e-01 -6.96255624e-01
-3.73328656e-01 -9.12589014e-01 9.35796380e-01 -5.48574209e-01
-1.59358156e+00 7.24578798e-01 3.01986039e-01 -9.98583972e-01
-3.79107416e-01 -3.39974642e-01 -2.67545789e-01 2.82143295e-01
-1.42174578e+00 -1.12335944e+00 -6.38068020e-01 9.85329390e-01
4.33360964e-01 -5.07046759e-01 1.03640366e+00 1.06543338e+00
-7.42335737e-01 9.10280645e-01 -4.42366481e-01 2.01377615e-01
1.11009002e+00 -8.73246670e-01 3.49874586e-01 4.76683140e-01
1.15046769e-01 6.88209116e-01 6.39993012e-01 -6.56157076e-01
-9.45113599e-01 -1.39137900e+00 9.72365856e-01 -9.15441275e-01
6.01139247e-01 -7.26329982e-01 -7.09671378e-01 1.16533661e+00
-2.63848037e-01 -2.45859325e-01 9.39921975e-01 7.39837229e-01
-5.51843584e-01 -5.76208949e-01 -9.63020205e-01 3.40433985e-01
8.40988576e-01 -6.77456439e-01 -5.50236940e-01 5.83714187e-01
6.20365798e-01 -1.35211140e-01 -3.70229125e-01 3.69264007e-01
2.71607041e-01 -3.71654809e-01 8.58705938e-01 -5.09346426e-01
-6.98810935e-01 -3.01323861e-01 -5.30615225e-02 -1.04891980e+00
-6.21574223e-01 -7.48605907e-01 6.38848096e-02 1.17619562e+00
5.54991126e-01 -7.27125049e-01 8.22775304e-01 9.14247096e-01
-3.21402848e-01 -1.08464040e-01 -1.04331374e+00 -6.59351587e-01
-7.40999103e-01 -1.02322018e+00 1.21844292e+00 1.27872157e+00
1.83791950e-01 5.60204327e-01 -6.38580739e-01 2.15278134e-01
8.31922293e-01 -1.31790310e-01 8.46834183e-01 -1.72818887e+00
-9.98036340e-02 -3.35883677e-01 -1.31760479e-03 -8.06803346e-01
1.24904461e-01 -8.86093497e-01 2.25986466e-01 -1.18694937e+00
5.56420581e-03 -1.17473423e+00 -6.01603985e-01 1.31539094e+00
3.78830105e-01 7.23917410e-02 -4.70433384e-01 4.89594430e-01
-1.04852533e+00 -7.67727345e-02 1.25618130e-01 -5.46489060e-01
-3.04568142e-01 2.79407352e-01 -7.43942142e-01 5.20035446e-01
5.88701487e-01 -7.20764995e-01 -3.74868214e-01 -4.74914551e-01
2.88926244e-01 -1.66666448e-01 2.82500803e-01 -7.72214353e-01
7.13647425e-01 -4.98982102e-01 6.82940781e-02 -7.09822595e-01
2.47077182e-01 -1.09542954e+00 1.27955973e-01 1.03905149e-01
-5.43573201e-01 -1.31805167e-01 -2.60855705e-02 5.37370563e-01
2.13947311e-01 1.60514116e-01 5.04446387e-01 2.19724104e-01
-1.07504034e+00 5.92821896e-01 -6.99790716e-01 7.44378194e-02
9.28899586e-01 -1.03985347e-01 -9.36225951e-01 -2.80127764e-01
-4.53667819e-01 5.00911415e-01 -9.53893960e-02 6.50489509e-01
3.08452010e-01 -1.11332309e+00 -5.47480166e-01 2.75941044e-01
6.00628555e-01 -1.85297668e-01 2.53346823e-02 9.70525563e-01
-1.05011612e-01 5.94117284e-01 -1.47947341e-01 -8.10254633e-01
-1.14346099e+00 3.37373406e-01 1.95659846e-01 -3.18033457e-01
-6.01100504e-01 6.87953293e-01 -3.25774401e-01 -4.27993238e-01
8.47434163e-01 -1.52773440e-01 -1.74394041e-01 1.93991795e-01
7.23004282e-01 2.48102024e-01 8.07327092e-01 -1.84888110e-01
-6.94234312e-01 -2.15048790e-02 -7.26486027e-01 -1.67930603e-01
1.37776196e+00 -3.18918318e-01 1.53518453e-01 2.25003928e-01
4.94932175e-01 -2.54304409e-01 -9.17952001e-01 -7.80406713e-01
6.10506177e-01 -2.64577985e-01 5.60084105e-01 -1.27412724e+00
-7.54926205e-01 1.22032247e-01 7.39717364e-01 1.17343850e-01
9.54611719e-01 7.47450814e-02 1.00245512e+00 5.43258846e-01
9.73062873e-01 -1.52268827e+00 2.78559774e-01 6.34060562e-01
4.35126841e-01 -1.08229816e+00 -1.22487232e-01 -1.40994877e-01
-6.03587925e-01 6.48945928e-01 5.91366351e-01 6.73198760e-01
8.54209065e-01 6.18897080e-01 4.01266307e-01 -5.00195205e-01
-1.41367030e+00 7.88870528e-02 1.19081572e-01 5.73812842e-01
2.75388863e-02 3.53384584e-01 2.17402562e-01 6.77675962e-01
-3.45956981e-01 9.62279812e-02 -3.22823599e-02 1.08474207e+00
-6.67218268e-01 -1.15943277e+00 -6.64492488e-01 7.85533547e-01
-8.35059136e-02 -1.29505739e-01 -1.87260300e-01 5.67782879e-01
2.03846321e-01 1.50395632e+00 2.59638697e-01 -6.81569397e-01
3.20228070e-01 2.79000551e-01 9.11400095e-02 -9.99663293e-01
-7.25271463e-01 -6.01630099e-02 4.70324725e-01 -3.38163018e-01
1.29386485e-01 -4.75347042e-01 -1.19656575e+00 -5.38132131e-01
-3.36158037e-01 3.94385189e-01 6.29079700e-01 1.23350680e+00
3.50330621e-01 5.67398965e-01 5.44086874e-01 -4.09256518e-01
-6.69361532e-01 -1.02722418e+00 -3.96016955e-01 1.74685493e-01
1.63150921e-01 -1.11553156e+00 -3.11168849e-01 6.33977056e-02]
|
[9.527668952941895, 4.457243919372559]
|
80462f27-e593-4cbd-9f32-2806e98d119a
|
language-driven-region-pointer-advancement
|
2011.14901
| null |
https://arxiv.org/abs/2011.14901v1
|
https://arxiv.org/pdf/2011.14901v1.pdf
|
Language-Driven Region Pointer Advancement for Controllable Image Captioning
|
Controllable Image Captioning is a recent sub-field in the multi-modal task of Image Captioning wherein constraints are placed on which regions in an image should be described in the generated natural language caption. This puts a stronger focus on producing more detailed descriptions, and opens the door for more end-user control over results. A vital component of the Controllable Image Captioning architecture is the mechanism that decides the timing of attending to each region through the advancement of a region pointer. In this paper, we propose a novel method for predicting the timing of region pointer advancement by treating the advancement step as a natural part of the language structure via a NEXT-token, motivated by a strong correlation to the sentence structure in the training data. We find that our timing agrees with the ground-truth timing in the Flickr30k Entities test data with a precision of 86.55% and a recall of 97.92%. Our model implementing this technique improves the state-of-the-art on standard captioning metrics while additionally demonstrating a considerably larger effective vocabulary size.
|
['John D. Kelleher', 'Robert J. Ross', 'Annika Lindh']
|
2020-11-30
| null |
https://aclanthology.org/2020.coling-main.174
|
https://aclanthology.org/2020.coling-main.174.pdf
|
coling-2020-8
|
['controllable-image-captioning']
|
['computer-vision']
|
[ 4.70458299e-01 4.31138158e-01 -4.80214745e-01 -6.76385880e-01
-1.05355406e+00 -9.18262184e-01 8.72184992e-01 2.38128260e-01
-4.71808791e-01 5.25549889e-01 7.27669239e-01 -3.28817874e-01
3.69608641e-01 -5.32902896e-01 -1.22881377e+00 -4.82771903e-01
1.71175793e-01 4.86176938e-01 4.75372039e-02 -2.58673638e-01
2.59177715e-01 9.63689834e-02 -1.25735867e+00 6.52016878e-01
2.64721036e-01 7.71998286e-01 2.11081475e-01 8.02883029e-01
3.73904146e-02 8.66707265e-01 -5.73979080e-01 -5.15502393e-01
8.82244669e-03 -2.76300281e-01 -9.38556731e-01 2.40815505e-01
9.60540891e-01 -3.58574599e-01 -3.80664438e-01 6.82181776e-01
2.30366588e-01 -6.29817545e-02 6.87894881e-01 -1.23095834e+00
-1.01328492e+00 7.57671535e-01 -7.32120395e-01 2.32993290e-01
4.71275896e-01 1.55695155e-01 1.16756773e+00 -4.86779064e-01
9.96330023e-01 9.34508562e-01 1.21002793e-01 7.64699042e-01
-1.14979303e+00 -4.04353172e-01 3.39376897e-01 6.09000921e-02
-1.33690608e+00 -6.32176459e-01 5.22692800e-01 -6.13891900e-01
8.12476456e-01 4.15226012e-01 1.83316872e-01 1.04595566e+00
-7.27126971e-02 7.74669766e-01 7.49659419e-01 -5.88068902e-01
5.42368069e-02 1.65035844e-01 -1.11127369e-01 6.05687439e-01
-1.31905422e-01 -9.31020454e-02 -5.52834868e-01 1.21261127e-01
7.34472215e-01 -3.64457667e-01 -3.42682034e-01 -3.46986622e-01
-1.41774142e+00 7.19863892e-01 5.67400932e-01 2.37805009e-01
-1.31539211e-01 5.88653505e-01 3.75495344e-01 -2.22941473e-01
4.19615507e-01 7.36672163e-01 -4.81527835e-01 -1.35775149e-01
-1.02569306e+00 4.15912777e-01 5.12639880e-01 1.17230570e+00
6.06142104e-01 -4.68293995e-01 -8.13942671e-01 2.44166836e-01
1.96803838e-01 5.44000626e-01 -1.01172356e-02 -1.09610975e+00
8.05511117e-01 4.88424838e-01 6.00036740e-01 -7.65618861e-01
8.41551274e-03 -2.72340924e-01 -3.35286409e-01 -1.43049017e-01
4.76545900e-01 3.74638736e-02 -1.14778709e+00 2.03698397e+00
1.62676319e-01 1.52749538e-01 -1.37335151e-01 9.51955557e-01
4.98765707e-01 1.05386090e+00 4.39294785e-01 2.85693314e-02
1.74012649e+00 -9.32979047e-01 -4.93661344e-01 -4.75459039e-01
6.38867855e-01 -7.47778118e-01 1.28107917e+00 -2.58784425e-02
-8.45598161e-01 -3.89922321e-01 -7.39632130e-01 -2.66044825e-01
-1.91703513e-01 1.19625278e-01 5.80146611e-01 2.46905878e-01
-1.28177154e+00 1.30991980e-01 -4.98299897e-01 -2.62377948e-01
2.22711131e-01 1.20449148e-01 -3.53780627e-01 -2.23580196e-01
-1.20859456e+00 9.66559708e-01 2.15134650e-01 -1.22644216e-01
-8.13786626e-01 -9.38473761e-01 -1.02588212e+00 5.61350547e-02
2.86283135e-01 -8.04294050e-01 1.63503957e+00 -9.98940706e-01
-9.53845263e-01 1.21207690e+00 -4.57351059e-01 -5.79473019e-01
4.71460104e-01 -1.39896914e-01 -1.70783684e-01 3.03474486e-01
2.46600121e-01 1.53106380e+00 8.08529615e-01 -1.43671691e+00
-5.09596646e-01 -2.05925584e-01 4.35802490e-01 4.42153364e-01
-1.61253795e-01 1.33428037e-01 -9.45621729e-01 -4.07303780e-01
-5.58004022e-01 -1.17347956e+00 -3.87650877e-01 2.12941661e-01
-4.64773089e-01 -2.19947338e-01 4.47569668e-01 -5.36105156e-01
1.36632156e+00 -2.16076398e+00 7.88331851e-02 -2.76654303e-01
1.63715854e-01 -6.51396886e-02 -2.52988636e-01 4.19312209e-01
-8.24348405e-02 5.13214409e-01 -8.33400413e-02 -5.25818706e-01
2.69522965e-02 8.89659822e-02 -6.34207487e-01 2.53264427e-01
3.34348351e-01 1.15168798e+00 -1.16881168e+00 -7.03057528e-01
2.86884964e-01 4.98556376e-01 -4.66261983e-01 2.33485177e-01
-6.63366079e-01 1.79964259e-01 -5.17472029e-01 4.02203053e-01
2.42477387e-01 -6.26297951e-01 -9.66467410e-02 -2.97495365e-01
-1.91717058e-01 1.19675681e-01 -5.62767744e-01 1.90925193e+00
-6.18515968e-01 9.41217422e-01 -3.18611085e-01 -3.20734978e-01
5.16936779e-01 4.73081142e-01 3.44061732e-01 -6.63509667e-01
-1.10093020e-01 -2.55983114e-01 -3.18921268e-01 -5.71816564e-01
7.93964922e-01 2.56972432e-01 -4.42872435e-01 3.18566352e-01
-3.19105387e-01 -5.53585179e-02 4.14674044e-01 5.19456804e-01
9.44731832e-01 2.35388339e-01 2.06863299e-01 -1.29866108e-01
2.22549468e-01 2.84948915e-01 -7.97661096e-02 8.82625759e-01
-1.29764423e-01 9.78819430e-01 6.01781309e-01 -5.47558010e-01
-1.53351629e+00 -7.78087795e-01 1.58351269e-02 1.10251224e+00
1.86073020e-01 -4.64556634e-01 -9.25224423e-01 -6.43735766e-01
-3.44955504e-01 9.79217172e-01 -1.11698067e+00 9.32064578e-02
-6.15545273e-01 -2.29523838e-01 3.72067690e-01 5.14187992e-01
2.33526096e-01 -1.30258167e+00 -6.19494021e-01 -4.17699367e-02
-6.54286385e-01 -1.67567623e+00 -1.04811084e+00 -2.35767499e-01
-3.94724011e-01 -7.46213853e-01 -8.05353582e-01 -9.45304930e-01
9.35124874e-01 2.28560418e-01 1.48872995e+00 -2.04592459e-02
3.51585390e-04 3.80040169e-01 -3.75165254e-01 -2.92949766e-01
-4.46270019e-01 2.31907904e-01 -4.24348801e-01 -5.48988543e-02
1.49363294e-01 -2.89830882e-02 -7.53080249e-01 2.20355362e-01
-1.08174443e+00 6.72203422e-01 5.71447790e-01 4.08296049e-01
7.37031996e-01 -4.80436444e-01 -5.69003075e-02 -7.14865327e-01
3.69263113e-01 -3.96915406e-01 -5.56758046e-01 4.67712283e-01
-4.21951890e-01 4.58487779e-01 4.35900003e-01 -5.77271342e-01
-7.71162570e-01 5.06090701e-01 2.14682207e-01 -4.11046803e-01
-2.26486072e-01 2.30817765e-01 1.92566551e-02 3.43784094e-01
5.37492633e-01 3.21056426e-01 -2.08483770e-01 -2.22079493e-02
7.77121842e-01 4.82976168e-01 6.57914400e-01 -5.78898251e-01
7.04134166e-01 5.08950949e-01 -1.30013794e-01 -5.24399638e-01
-1.24584413e+00 -6.32677436e-01 -3.74987125e-01 -3.12677443e-01
1.23587382e+00 -1.11732709e+00 -6.51891291e-01 6.44712076e-02
-1.48545253e+00 -4.05007392e-01 -9.97835547e-02 -1.81237925e-02
-6.06962860e-01 1.27325386e-01 -3.75300348e-01 -4.60120231e-01
-4.64610249e-01 -1.29310071e+00 1.73669183e+00 2.07463220e-01
-3.80902767e-01 -7.59631932e-01 -1.34459749e-01 6.62278593e-01
2.98777729e-01 3.51473421e-01 8.10780585e-01 -2.80197501e-01
-8.43561172e-01 -2.39452288e-01 -4.28137690e-01 -6.48234114e-02
-1.57871276e-01 4.96099293e-02 -8.17935348e-01 -6.09061532e-02
-5.11540174e-01 -2.85271168e-01 6.07366979e-01 4.48605508e-01
1.27934182e+00 -3.81419748e-01 -4.20723587e-01 2.88669884e-01
1.63474393e+00 -1.21032260e-01 8.95665228e-01 4.87993181e-01
7.19417453e-01 5.09979129e-01 6.25743508e-01 2.59653747e-01
6.37597978e-01 9.96075332e-01 7.17778921e-01 -2.40750387e-01
-2.19100490e-01 -6.34403169e-01 2.66629755e-01 1.67314902e-01
3.00864130e-01 -8.42187285e-01 -9.19521153e-01 8.18626165e-01
-1.78672409e+00 -1.02630472e+00 -2.52012432e-01 2.12950277e+00
9.84056950e-01 1.66191816e-01 -1.16429940e-01 -3.89045447e-01
8.43662918e-01 4.97513443e-01 -3.33061635e-01 -4.34526265e-01
1.02885082e-01 -1.60835981e-01 9.41219985e-01 7.58243680e-01
-1.13666701e+00 1.09096348e+00 6.46013641e+00 5.97353756e-01
-1.03244996e+00 -1.66637406e-01 9.69470263e-01 3.83349620e-02
-4.72240180e-01 1.06405243e-01 -1.04336238e+00 5.40077269e-01
1.07353556e+00 -5.34088425e-02 3.95084769e-01 6.11257732e-01
4.57155406e-01 -1.75851718e-01 -1.38815057e+00 9.06085372e-01
3.59409720e-01 -1.54481435e+00 4.53670681e-01 1.96983546e-01
6.67414248e-01 -2.23576762e-02 2.08048373e-01 5.11793904e-02
4.35008258e-02 -1.09797716e+00 9.99314904e-01 3.70489150e-01
1.17553413e+00 -3.39747936e-01 4.03142303e-01 1.71652555e-01
-9.90683734e-01 2.84845471e-01 -7.38509325e-03 5.78634106e-02
4.95090842e-01 3.26373935e-01 -1.30421281e+00 1.93619996e-01
3.23222578e-01 3.89630139e-01 -5.94540775e-01 7.99143732e-01
-4.77197915e-01 5.47188640e-01 -1.14556804e-01 -1.96043074e-01
4.69488442e-01 2.48063952e-01 3.47131282e-01 1.29799318e+00
4.44684029e-02 7.43732601e-02 -5.71506433e-02 7.28838742e-01
-3.23636115e-01 5.54430187e-02 -5.34038723e-01 -2.51935333e-01
4.70043391e-01 1.05377984e+00 -6.28092885e-01 -3.91342789e-01
-2.35765040e-01 1.07465625e+00 2.37300351e-01 2.58151025e-01
-1.30350101e+00 -3.52678522e-02 5.51385462e-01 3.39291215e-01
4.65005577e-01 -2.21189886e-01 -2.76207924e-01 -1.00763631e+00
1.29922867e-01 -7.55209684e-01 9.51212496e-02 -1.48883426e+00
-7.49136209e-01 7.71273673e-01 1.50855705e-01 -1.25322902e+00
-3.58946592e-01 -5.42494118e-01 -2.69169986e-01 7.34951615e-01
-1.64674449e+00 -1.25180936e+00 -4.19474334e-01 1.83168918e-01
7.14991868e-01 5.05936921e-01 9.79543567e-01 2.16124326e-01
-2.61486411e-01 5.03463686e-01 -2.99868971e-01 2.93821186e-01
8.36190164e-01 -1.23749852e+00 7.48929799e-01 9.34301972e-01
3.80040020e-01 4.77881551e-01 1.17389953e+00 -3.88471842e-01
-1.18985176e+00 -1.11155248e+00 1.35104680e+00 -1.01737010e+00
6.74048007e-01 -5.78837276e-01 -6.23657644e-01 6.44993961e-01
4.55874383e-01 -2.51447950e-02 4.17983323e-01 -6.98193610e-02
-6.46492839e-01 4.95386273e-02 -7.87371874e-01 7.39953697e-01
8.22167695e-01 -5.54428041e-01 -4.67943102e-01 4.66925651e-01
1.21471465e+00 -6.77888513e-01 -4.96169478e-01 1.89867932e-02
4.58037287e-01 -4.61751819e-01 8.74612629e-01 -5.46641588e-01
1.11405003e+00 -4.77483809e-01 -4.59521152e-02 -9.62726474e-01
-3.48590165e-01 -5.16429663e-01 -3.23708542e-02 1.20497990e+00
8.07948172e-01 2.54245013e-01 1.01978445e+00 9.92934048e-01
-1.05664961e-01 -7.61682630e-01 -6.00760698e-01 -3.62569928e-01
-1.56239167e-01 -2.98313320e-01 4.53668863e-01 7.11231172e-01
-2.33104765e-01 4.55235690e-01 -5.09743392e-01 2.03277960e-01
5.46733916e-01 6.55421168e-02 6.53895080e-01 -6.02302074e-01
-5.81084713e-02 -1.61085099e-01 -3.73458236e-01 -1.35869789e+00
1.15580596e-01 -7.11167932e-01 1.51748359e-01 -1.75837052e+00
4.91199732e-01 -1.53452188e-01 -8.47159028e-02 6.56823933e-01
-1.72844470e-01 3.91617745e-01 2.80068904e-01 1.36219054e-01
-1.04727876e+00 1.94007456e-01 1.45042980e+00 -2.38060161e-01
4.32594540e-03 -3.31010848e-01 -9.72363591e-01 2.20498428e-01
6.24327958e-01 -4.80203539e-01 -5.28384089e-01 -1.02384531e+00
3.02504569e-01 1.96070701e-01 3.24773818e-01 -7.04870880e-01
2.25777552e-01 -3.26581806e-01 2.40337104e-01 -3.68961751e-01
4.41599220e-01 -7.24917054e-01 1.27254259e-02 1.64504722e-01
-9.25152123e-01 3.41132253e-01 2.26133958e-01 5.00430167e-01
-4.12162207e-02 -1.95890348e-02 5.26070714e-01 -1.45793721e-01
-9.29579616e-01 4.12151009e-01 -6.64801747e-02 2.36190081e-01
1.15242302e+00 -2.00490043e-01 -4.91224021e-01 -6.18074536e-01
-4.66211438e-01 4.47080046e-01 7.01467574e-01 8.95600677e-01
4.85046208e-01 -1.06074786e+00 -7.70539105e-01 -2.40722135e-01
5.95728159e-01 -1.43423434e-02 1.45734549e-01 2.96512574e-01
-4.48911607e-01 6.80293679e-01 5.07018529e-02 -6.30543411e-01
-1.26394439e+00 6.51430488e-01 2.42680088e-01 -3.44459057e-01
-3.52840394e-01 9.16375041e-01 2.54979134e-01 2.67674476e-01
3.00850481e-01 -3.85503441e-01 -1.41548380e-01 -1.61183313e-01
6.63448811e-01 -4.24357414e-01 -1.25885591e-01 -7.95907438e-01
-2.99010277e-01 4.60630596e-01 -2.41077453e-01 -2.84042090e-01
1.14386785e+00 -2.24486604e-01 4.74943183e-02 2.91583270e-01
1.29323328e+00 -5.53209856e-02 -1.43802059e+00 -3.93170165e-03
-1.64465040e-01 -3.69717509e-01 1.14584550e-01 -1.13754511e+00
-7.13115036e-01 5.44643700e-01 4.80488241e-01 -4.88897823e-02
8.12151134e-01 4.68053341e-01 6.95365131e-01 1.21261649e-01
3.49191427e-01 -6.30198359e-01 -1.97147150e-02 2.47818172e-01
9.29924428e-01 -1.43034434e+00 -9.60346535e-02 -3.41677457e-01
-9.04039204e-01 8.36619973e-01 7.53545165e-01 9.53687206e-02
1.30683437e-01 7.02372193e-02 1.83811218e-01 -2.09425539e-01
-8.91111910e-01 7.69358128e-02 4.00960028e-01 4.52846557e-01
5.63739419e-01 9.31465775e-02 -1.28229231e-01 6.35804832e-02
-1.05842851e-01 -2.77178511e-02 7.44237125e-01 6.23604774e-01
-3.11505139e-01 -9.46229994e-01 -1.40312374e-01 1.93212107e-01
-5.78240454e-01 -4.92585272e-01 -3.48972350e-01 6.78986728e-01
-1.30530139e-02 8.59453201e-01 2.72251934e-01 -4.32385691e-02
3.16792279e-01 -1.06282130e-01 5.30032158e-01 -8.09944689e-01
-3.07018608e-01 -3.75563115e-01 1.56937972e-01 -5.38632512e-01
-4.90472555e-01 -5.89504361e-01 -1.15726972e+00 5.58599830e-02
-1.47327885e-01 2.94012398e-01 9.21950400e-01 9.02291536e-01
5.28079629e-01 3.77143919e-01 3.91873062e-01 -6.98097885e-01
-3.60957563e-01 -7.28450656e-01 4.17182520e-02 7.19014704e-01
4.58046973e-01 -1.82124823e-01 -3.29904526e-01 6.33257449e-01]
|
[10.979716300964355, 0.9775854349136353]
|
672f7b71-0cb7-4882-a702-bce821e6b398
|
a-deep-learning-model-for-forecasting-global
|
2202.09967
| null |
https://arxiv.org/abs/2202.09967v1
|
https://arxiv.org/pdf/2202.09967v1.pdf
|
A Deep Learning Model for Forecasting Global Monthly Mean Sea Surface Temperature Anomalies
|
Sea surface temperature (SST) variability plays a key role in the global weather and climate system, with phenomena such as El Ni\~{n}o-Southern Oscillation regarded as a major source of interannual climate variability at the global scale. The ability to be able to make long-range forecasts of sea surface temperature anomalies, especially those associated with extreme marine heatwave events, has potentially significant economic and societal benefits. We have developed a deep learning time series prediction model (Unet-LSTM) based on more than 70 years (1950-2021) of ECMWF ERA5 monthly mean sea surface temperature and 2-metre air temperature data. The Unet-LSTM model is able to learn the underlying physics driving the temporal evolution of the 2-dimensional global sea surface temperatures. The model accurately predicts sea surface temperatures over a 24 month period with a root mean square error remaining below 0.75$^\circ$C for all predicted months. We have also investigated the ability of the model to predict sea surface temperature anomalies in the Ni\~{n}o3.4 region, as well as a number of marine heatwave hot spots over the past decade. Model predictions of the Ni\~{n}o3.4 index allow us to capture the strong 2010-11 La Ni\~{n}a, 2009-10 El Nino and the 2015-16 extreme El Ni\~{n}o up to 24 months in advance. It also shows long lead prediction skills for the northeast Pacific marine heatwave, the Blob. However, the prediction of the marine heatwaves in the southeast Indian Ocean, the Ningaloo Ni\~{n}o, shows limited skill. These results indicate the significant potential of data driven methods to yield long-range predictions of sea surface temperature anomalies.
|
['Ming Feng', 'John Taylor']
|
2022-02-21
| null | null | null | null |
['time-series-prediction']
|
['time-series']
|
[-2.48149112e-01 -4.84279156e-01 3.04166079e-01 -2.98644155e-01
-5.30563593e-01 -4.72685635e-01 7.39111841e-01 -1.30838454e-01
-3.62618506e-01 1.08661997e+00 -3.01760924e-03 -1.09278524e+00
-2.65703350e-02 -1.09451973e+00 -4.30005193e-01 -1.09588301e+00
-7.58762896e-01 9.44928378e-02 -4.13155228e-01 -8.42181504e-01
-6.38359692e-03 5.84716856e-01 -1.39806056e+00 -3.36039007e-01
7.94514716e-01 1.08683276e+00 -3.44189703e-02 7.21970320e-01
-1.90821677e-01 -4.71372455e-02 -2.07968548e-01 5.96371293e-01
3.41798723e-01 -3.77926826e-01 -5.22059441e-01 -5.40417492e-01
2.93703854e-01 -5.60761206e-02 3.05668145e-01 7.31076479e-01
3.34697753e-01 3.32088709e-01 3.86344820e-01 -8.16159487e-01
2.46631607e-01 1.55915648e-01 -4.77613658e-01 4.30116385e-01
-3.13077867e-01 2.23330230e-01 7.63979495e-01 -7.50707865e-01
3.26694429e-01 9.10707533e-01 8.76361191e-01 2.55722016e-01
-1.13820767e+00 -9.29310739e-01 5.57692721e-02 -4.70958740e-01
-1.13809001e+00 -1.46810740e-01 4.45926309e-01 -6.66858971e-01
1.26140106e+00 4.49672401e-01 8.60901356e-01 6.84625328e-01
9.68364120e-01 -2.52048578e-02 1.63428164e+00 -3.14966083e-01
2.49787658e-01 -3.82616162e-01 -2.42375284e-02 2.64172792e-01
2.21885610e-02 7.56690025e-01 -2.58103281e-01 -2.57060707e-01
4.49004501e-01 9.67806950e-02 -7.70240054e-02 7.88582385e-01
-7.86466539e-01 7.89982080e-01 6.08436048e-01 4.23479408e-01
-4.47112143e-01 1.96568802e-01 1.33885860e-01 6.31386757e-01
1.16985989e+00 5.02525151e-01 -1.45569551e+00 -2.14624584e-01
-1.20527899e+00 5.80414653e-01 5.24741948e-01 2.61901557e-01
8.38925838e-01 6.96272254e-01 5.89501500e-01 5.43496192e-01
4.96338427e-01 1.38470292e+00 1.31674290e-01 -6.89630508e-01
1.38281032e-01 1.53170198e-01 4.65814024e-01 -7.85405636e-01
-9.06475544e-01 -5.67207158e-01 -1.19996190e+00 6.21181488e-01
4.05077517e-01 -1.03176260e+00 -1.05688071e+00 1.69670689e+00
1.23713776e-01 -6.02582619e-02 1.79153010e-01 7.57915914e-01
1.43642485e-01 1.46476674e+00 2.53896385e-01 -5.03466010e-01
1.18722343e+00 -3.07242543e-01 -6.26089156e-01 -5.21716774e-01
8.47623467e-01 -4.93954062e-01 4.08253253e-01 -2.22293511e-01
-5.60379982e-01 -4.05127913e-01 -5.28899610e-01 6.69688344e-01
-7.98534036e-01 -3.68833154e-01 4.76823539e-01 1.72717616e-01
-1.33666646e+00 8.33810806e-01 -1.15618241e+00 -3.86837542e-01
-2.41299778e-01 2.10268319e-01 -7.89208412e-02 6.35420382e-01
-1.80068302e+00 8.99797380e-01 8.83104056e-02 8.76763344e-01
-6.39692426e-01 -9.42879438e-01 -9.51491058e-01 1.98855966e-01
-5.62050998e-01 -2.50493884e-01 9.70545232e-01 -1.02095044e+00
-1.17276776e+00 3.47310185e-01 -5.71950793e-01 -3.47289056e-01
-1.19277565e-02 -2.74762809e-01 -9.63763714e-01 -5.47338665e-01
5.48120178e-02 3.58281106e-01 2.12455973e-01 -9.75644052e-01
-9.06956077e-01 -4.78528857e-01 -8.11407268e-01 7.69317672e-02
3.41405943e-02 3.18752527e-01 6.38053298e-01 -7.38641262e-01
3.24929357e-01 -1.18226528e+00 -7.45203733e-01 -5.28904617e-01
1.41487941e-01 -1.30193397e-01 8.64506006e-01 -9.89425600e-01
1.05219340e+00 -1.89869535e+00 -4.39777792e-01 4.40182567e-01
-3.49206537e-01 1.11791559e-01 9.24580991e-02 6.12494588e-01
-1.46900639e-01 5.02734601e-01 -6.36686027e-01 -7.39179403e-02
-3.75386596e-01 5.46744347e-01 -7.86167741e-01 5.26080191e-01
1.47139505e-01 5.48333049e-01 -6.68430448e-01 3.83948356e-01
1.68779910e-01 3.72471690e-01 1.06451310e-01 1.11012913e-01
-3.95772159e-01 8.17055464e-01 -2.78401494e-01 3.97904515e-01
9.48666036e-01 1.56690300e-01 -2.05946609e-01 8.47377181e-01
-1.05449462e+00 2.55893558e-01 -6.18575335e-01 1.14592826e+00
-3.37290794e-01 9.06154931e-01 3.19492489e-01 -4.38692063e-01
1.28649962e+00 6.23003125e-01 3.44746143e-01 -1.05767417e+00
-1.00536242e-01 6.01775050e-01 2.40053385e-01 -3.69878024e-01
3.99291664e-01 -7.20133543e-01 -2.44232584e-02 5.55480719e-01
-4.82352853e-01 -2.86956251e-01 -3.28216344e-01 -4.23423886e-01
2.33358011e-01 3.00063848e-01 -1.82423890e-01 -1.08058894e+00
2.03848287e-01 1.36880755e-01 8.54292870e-01 5.36031723e-01
9.98490024e-03 3.16471130e-01 3.58490914e-01 -1.21272349e+00
-1.14957154e+00 -5.59244573e-01 -6.91294730e-01 1.17103076e+00
-5.37233055e-01 9.33390930e-02 -1.02877982e-01 -8.77284736e-04
-8.04748461e-02 9.06829000e-01 -8.40875626e-01 2.33738661e-01
-6.22227788e-01 -1.37716138e+00 4.96895641e-01 2.81830013e-01
4.06379998e-01 -1.18781102e+00 -6.14054143e-01 3.91164243e-01
-6.96684793e-02 -6.83639109e-01 3.80546272e-01 5.17191768e-01
-1.45580626e+00 -3.08965594e-01 -6.77859306e-01 -4.54426080e-01
4.52485919e-01 -4.34936434e-01 1.25761938e+00 -3.95573825e-01
3.15576524e-01 -2.51982421e-01 -7.80845955e-02 -9.09417927e-01
-2.56921619e-01 1.90454796e-02 3.40449870e-01 -3.19174498e-01
4.59297150e-01 -7.62520552e-01 -6.97569013e-01 3.24593276e-01
-5.58453500e-01 2.05159821e-02 5.51171089e-03 5.05268037e-01
3.88648689e-01 -5.88358305e-02 7.44329751e-01 -3.63841206e-01
2.14545559e-02 -6.77590132e-01 -1.11307752e+00 -2.74112672e-01
-9.44450200e-01 -7.55073922e-03 4.73690897e-01 2.66065270e-01
-1.08181250e+00 -1.03422351e-01 -4.21102524e-01 1.25457078e-01
-4.14184660e-01 1.09902501e+00 5.23686707e-01 2.92021453e-01
4.03187633e-01 5.42694807e-01 -2.35445112e-01 -8.79851162e-01
-2.58639216e-01 4.42137599e-01 3.42620283e-01 -3.96120369e-01
9.23000753e-01 3.74637574e-01 2.05567226e-01 -1.22762489e+00
-4.35956091e-01 -2.26608872e-01 -5.64673960e-01 -2.75371999e-01
9.35025871e-01 -1.18359101e+00 -2.62892812e-01 1.01139224e+00
-7.92040110e-01 -6.81536794e-01 2.17253938e-01 6.06172204e-01
3.92407514e-02 -3.85099530e-01 -4.46834952e-01 -1.14018071e+00
-7.48312891e-01 -6.89530671e-01 6.39286578e-01 4.09885257e-01
-2.04106048e-01 -1.45546079e+00 7.07235754e-01 -4.54305232e-01
1.08043420e+00 9.02189374e-01 9.60022271e-01 -3.17153215e-01
9.84799676e-03 -5.73994741e-02 5.50596826e-02 2.83455640e-01
3.84195954e-01 3.23089033e-01 -1.04842734e+00 -6.18498504e-01
9.45827365e-02 1.55287847e-01 9.35785294e-01 7.48247743e-01
5.10314703e-01 -3.53672385e-01 -4.27136943e-02 8.32951367e-01
1.45907223e+00 5.57076752e-01 2.82192707e-01 7.26050377e-01
1.70644969e-01 4.91077006e-01 2.23005980e-01 3.69688064e-01
1.56236783e-01 -9.02872160e-03 5.77252328e-01 -5.47487736e-01
6.81573868e-01 3.49907994e-01 4.66003656e-01 6.60511851e-01
-3.05527896e-01 2.85413831e-01 -1.69147801e+00 8.80539536e-01
-1.19890416e+00 -7.26271510e-01 -6.12380981e-01 2.11205459e+00
6.18491530e-01 4.19421494e-02 -4.55962718e-01 -1.66040763e-01
1.47171095e-01 7.76133716e-01 -2.85306454e-01 -1.09817588e+00
-1.20434232e-01 6.01743758e-01 9.09086287e-01 6.48790061e-01
-1.04197991e+00 8.86005402e-01 6.18851900e+00 -1.50328159e-01
-1.82118440e+00 -1.85354054e-01 7.49444008e-01 9.17438194e-02
-3.48798037e-01 1.33574188e-01 -8.19150031e-01 2.79994726e-01
1.78923678e+00 -1.23550639e-01 5.70367835e-02 4.87604767e-01
1.12850118e+00 -2.80646920e-01 -3.60675663e-01 6.95332214e-02
-7.70767808e-01 -1.30277109e+00 -2.12283298e-01 3.59037459e-01
1.18936396e+00 7.90392697e-01 7.82719702e-02 2.73189187e-01
3.27096760e-01 -1.17790079e+00 2.11755410e-01 7.35198081e-01
1.10030341e+00 -9.09391284e-01 9.04330671e-01 3.94575626e-01
-1.26462948e+00 1.03156269e-01 -2.72903085e-01 -8.38730991e-01
1.53355598e-01 8.19073439e-01 -2.56373078e-01 4.59881067e-01
1.32848179e+00 8.41620028e-01 2.13352200e-02 5.54139793e-01
-6.56225830e-02 1.33669126e+00 -7.65106380e-01 2.91914731e-01
7.99859941e-01 -2.89583474e-01 6.06605649e-01 9.83144760e-01
7.23160863e-01 5.62006474e-01 -1.63661912e-01 4.58125472e-01
1.53545424e-01 -4.27283309e-02 -7.18396664e-01 4.13469940e-01
5.94217852e-02 7.98581779e-01 -1.44632712e-01 -1.17167301e-01
-2.68108338e-01 2.24837884e-01 -4.50295597e-01 9.19668078e-01
-3.54730487e-01 -3.75287950e-01 1.49562478e+00 1.71897758e-03
7.82978833e-02 -6.49683118e-01 -5.43530285e-01 -9.66878772e-01
-3.16339821e-01 -6.20323896e-01 1.49489924e-01 -7.06070662e-01
-6.87533855e-01 6.77731872e-01 -1.77636892e-01 -1.21392953e+00
-5.07853985e-01 -6.16713583e-01 -1.22729313e+00 1.72588003e+00
-2.02844357e+00 -6.83745921e-01 1.64082736e-01 2.11772516e-01
9.55749229e-02 1.81006059e-01 1.35233259e+00 -2.57863879e-01
-6.00290000e-01 -1.14059925e-01 1.18217146e+00 2.84319401e-01
4.60706919e-01 -1.26147783e+00 1.08026469e+00 7.87041187e-01
-5.96851528e-01 5.18225014e-01 9.49184656e-01 -6.93432868e-01
-1.18061495e+00 -1.40206897e+00 1.50276995e+00 5.49771003e-02
8.25025082e-01 -2.16850117e-01 -1.16880476e+00 9.15882945e-01
2.77081788e-01 2.45616734e-01 6.23042643e-01 1.92945629e-01
7.18965232e-02 -4.89054143e-01 -9.05295968e-01 3.43497127e-01
-1.02974758e-01 -4.49783564e-01 -4.26622897e-01 1.01328060e-01
5.07144213e-01 -1.13407336e-01 -1.20151806e+00 8.30315173e-01
7.77160168e-01 -6.10113025e-01 2.11283922e-01 -7.42051959e-01
4.30740505e-01 -5.72130047e-02 -1.81725889e-01 -1.80493522e+00
-3.33470404e-01 -4.85099465e-01 4.03325260e-01 8.74298394e-01
8.19093049e-01 -1.09528148e+00 4.96310920e-01 7.36784160e-01
-3.57307047e-02 -5.26418388e-01 -1.49050534e+00 -6.49478912e-01
1.00877607e+00 -6.24696851e-01 5.33926249e-01 1.08148706e+00
-4.56829935e-01 -4.00004745e-01 -4.95434433e-01 7.65555501e-01
5.26026547e-01 3.93284351e-01 4.00757313e-01 -1.55037582e+00
4.99243766e-01 -4.65540111e-01 2.62413263e-01 -5.05061388e-01
-1.29232347e-01 -2.44397938e-01 5.82103767e-02 -1.33564663e+00
-3.69095981e-01 -4.61673260e-01 -8.41446996e-01 8.22989821e-01
-4.16920744e-02 1.25241429e-01 -2.20320091e-01 1.16830222e-01
5.99516809e-01 6.44307315e-01 9.37367141e-01 2.29655430e-01
-3.82328659e-01 8.65127444e-02 1.36775792e-01 6.87142789e-01
9.86851275e-01 -5.88284731e-01 4.21172112e-01 -5.52822948e-01
6.28961802e-01 4.07812446e-01 -9.97491106e-02 -1.10843527e+00
-2.57917076e-01 -7.04703271e-01 6.05068505e-01 -5.77779174e-01
-1.51942074e-01 -7.40467489e-01 2.66926855e-01 6.77315116e-01
6.38460219e-02 4.97842878e-01 9.55510914e-01 6.89840689e-02
-4.55782473e-01 5.71135938e-01 7.72965372e-01 -1.52528688e-01
-6.29639089e-01 3.69282842e-01 -1.25565827e+00 -2.96283841e-01
6.23054802e-01 2.19787225e-01 -1.43960208e-01 -4.16370988e-01
-6.71489775e-01 8.84561956e-01 3.93576056e-01 6.16295040e-01
1.05502032e-01 -9.14698660e-01 -1.19943738e+00 7.07630754e-01
-1.27529755e-01 -7.23551586e-02 3.00094754e-01 5.65546453e-01
-6.76691234e-01 4.40107465e-01 -2.46668339e-01 -6.15357280e-01
-7.45184600e-01 -1.24030903e-01 1.07809186e+00 -1.86592668e-01
-4.71478045e-01 8.28748643e-01 -4.03536893e-02 -7.36743450e-01
-4.21458125e-01 -5.94439507e-01 -2.63546407e-01 3.17472160e-01
5.57519197e-01 1.23256564e-01 -8.40032250e-02 -8.39697540e-01
-4.52171475e-01 7.81679869e-01 4.39472795e-01 -2.15065822e-01
1.59931755e+00 -3.08544636e-01 -5.20334005e-01 1.11471915e+00
1.28933096e+00 -3.92619610e-01 -1.25182605e+00 -4.26552966e-02
1.33465886e-01 1.00779854e-01 4.41970050e-01 -1.14076781e+00
-1.24112749e+00 1.01245618e+00 9.56109703e-01 1.94336637e-03
1.15819025e+00 -4.85369772e-01 1.08417976e+00 2.90546268e-01
-6.01722253e-03 -6.56816721e-01 -1.22899806e+00 1.14990306e+00
7.11314619e-01 -1.23911548e+00 -7.99041316e-02 8.57388079e-01
-1.28578410e-01 1.35632336e+00 2.59867489e-01 9.57469046e-02
1.20297289e+00 2.85837114e-01 7.63336122e-01 -1.70473173e-01
-8.84569705e-01 9.29520950e-02 8.58846456e-02 -2.27933630e-01
6.08795941e-01 5.46492696e-01 -2.73294955e-01 -3.39576751e-02
-3.52004081e-01 -1.71204448e-01 3.88890952e-01 8.99377584e-01
-5.10830820e-01 -7.57362068e-01 -7.57877290e-01 4.52332556e-01
-7.81927824e-01 -3.98546994e-01 2.70394117e-01 5.30911148e-01
-7.20681772e-02 8.18155885e-01 4.69319969e-01 -1.80293322e-01
-2.07936823e-01 6.99167073e-01 -6.74230456e-01 3.19285169e-02
-7.82798588e-01 1.40647173e-01 -8.84794146e-02 -4.84497368e-01
-3.09955508e-01 -9.50199366e-01 -1.34724772e+00 -6.07852936e-01
1.50222639e-02 7.20019162e-01 1.00457597e+00 1.16710210e+00
1.03000253e-01 1.80288270e-01 1.11376989e+00 -1.20352542e+00
-1.33115143e-01 -1.40440154e+00 -9.08189714e-01 -1.76891640e-01
1.03531456e+00 -1.23112779e-02 -8.46706450e-01 -3.26350659e-01]
|
[6.521899700164795, 2.9653947353363037]
|
57afddf5-9317-47e0-a503-aa3049a360a0
|
interactive-generative-adversarial-networks
|
1801.09092
| null |
http://arxiv.org/abs/1801.09092v2
|
http://arxiv.org/pdf/1801.09092v2.pdf
|
Interactive Generative Adversarial Networks for Facial Expression Generation in Dyadic Interactions
|
A social interaction is a social exchange between two or more
individuals,where individuals modify and adjust their behaviors in response to
their interaction partners. Our social interactions are one of most fundamental
aspects of our lives and can profoundly affect our mood, both positively and
negatively. With growing interest in virtual reality and avatar-mediated
interactions,it is desirable to make these interactions natural and human like
to promote positive effect in the interactions and applications such as
intelligent tutoring systems, automated interview systems and e-learning. In
this paper, we propose a method to generate facial behaviors for an agent.
These behaviors include facial expressions and head pose and they are generated
considering the users affective state. Our models learn semantically meaningful
representations of the face and generate appropriate and temporally smooth
facial behaviors in dyadic interactions.
|
['Yuchi Huang', 'Behnaz Nojavanasghari', 'Saad Khan']
|
2018-01-27
| null | null | null | null |
['facial-expression-generation']
|
['computer-vision']
|
[-1.22271866e-01 6.07579052e-01 1.97617501e-01 -8.88101339e-01
4.19274390e-01 -5.13825893e-01 7.35015213e-01 -2.88811564e-01
2.71966904e-02 9.60358024e-01 2.68964231e-01 5.43260872e-01
2.83477426e-01 -8.01117480e-01 -1.92773208e-01 -4.59830523e-01
-6.96124062e-02 3.54644209e-01 -1.13692023e-02 -6.54461741e-01
4.59179208e-02 7.61063397e-01 -1.98628485e+00 1.81841135e-01
4.49965179e-01 4.22066987e-01 -2.66117841e-01 8.90841722e-01
-2.39494011e-01 8.64128470e-01 -5.56957781e-01 -5.35896719e-01
-2.20053658e-01 -5.21748126e-01 -5.94532013e-01 2.55814344e-01
-1.42813981e-01 -3.19025844e-01 -3.65272909e-02 1.05350637e+00
4.20474172e-01 6.38183475e-01 6.08188808e-01 -1.87390566e+00
-6.14374995e-01 4.10657048e-01 -8.77736330e-01 -4.29578751e-01
1.21562755e+00 7.09276125e-02 2.28875071e-01 -3.03575873e-01
7.20054984e-01 1.72405601e+00 3.39137584e-01 1.32873869e+00
-8.98359656e-01 -1.00340676e+00 3.45472515e-01 6.89913929e-02
-9.58934724e-01 -6.80981040e-01 7.75744557e-01 -3.63055021e-01
5.96753001e-01 3.95617992e-01 9.83323812e-01 1.19742048e+00
2.07455039e-01 6.86790287e-01 7.00942695e-01 -4.11663830e-01
-1.41290829e-01 4.89185959e-01 -3.47845219e-02 7.85825789e-01
-4.91546243e-01 -4.91362996e-02 -4.83256668e-01 -3.32667619e-01
8.64872515e-01 5.77184521e-02 -1.27478819e-02 -1.92587957e-01
-8.63118231e-01 5.82312763e-01 2.34172896e-01 2.26307929e-01
-6.57325327e-01 2.31509194e-01 2.18018875e-01 2.09729165e-01
8.94102156e-02 2.44467437e-01 -2.19756011e-02 -6.39551878e-01
7.71637857e-02 4.90090430e-01 7.03869820e-01 7.71656454e-01
5.82585037e-01 -4.53489013e-02 -6.80120215e-02 1.06211197e+00
4.82394010e-01 4.54154581e-01 4.71303135e-01 -1.09353793e+00
-6.40115499e-01 8.75212073e-01 2.65760243e-01 -1.23117530e+00
-6.79714620e-01 4.54822779e-01 -4.61879253e-01 3.72908205e-01
1.02973700e-01 -6.20262802e-01 -4.85902131e-01 2.23614478e+00
1.07237673e+00 3.19438159e-01 1.65212974e-01 6.98399544e-01
1.53299546e+00 6.87680125e-01 3.23439628e-01 -3.98109436e-01
1.30906940e+00 -7.19548166e-01 -1.46202993e+00 8.70733485e-02
5.35799563e-01 -8.16404104e-01 8.86162877e-01 1.46266580e-01
-1.29530430e+00 -3.16277176e-01 -5.41862190e-01 2.00931385e-01
-2.20736012e-01 -1.65128425e-01 1.00813425e+00 5.99621356e-01
-1.08394635e+00 4.90105122e-01 -5.78293025e-01 -7.47806549e-01
-1.44032435e-02 7.59318650e-01 -5.71317434e-01 7.99894392e-01
-1.16345537e+00 8.71888578e-01 -4.49785352e-01 1.30606771e-01
-3.66310000e-01 -1.11185893e-01 -9.87533569e-01 -2.83295900e-01
3.26415561e-02 -5.20522594e-01 1.66343784e+00 -1.45915556e+00
-2.08272767e+00 1.32827580e+00 -2.48277396e-01 9.81930122e-02
3.71114582e-01 1.29192621e-01 -4.69878137e-01 -1.98544592e-01
-4.06573772e-01 1.18131518e+00 5.02747238e-01 -1.29880500e+00
-3.52221757e-01 -7.27508008e-01 2.16335967e-01 6.82313323e-01
-1.55945256e-01 4.42769974e-01 -1.56693459e-01 -1.87527359e-01
-2.92703539e-01 -1.32251799e+00 -3.29501003e-01 2.13689640e-01
-2.08999321e-01 -3.53070647e-01 1.41890538e+00 -3.38912964e-01
5.16192496e-01 -2.09597850e+00 3.02995831e-01 1.68724030e-01
2.66615838e-01 1.61528990e-01 1.19785042e-02 2.96028972e-01
-1.04672700e-01 -1.99848950e-01 3.54909688e-01 -3.45008433e-01
5.96605316e-02 6.30542114e-02 1.20547175e-01 3.92775089e-01
-4.04802889e-01 8.68625283e-01 -8.89491200e-01 -5.35043955e-01
3.90681267e-01 8.14884067e-01 -3.82576436e-01 8.14015150e-01
-3.71877514e-02 9.88010466e-01 -5.85713625e-01 2.54476875e-01
3.76898944e-01 2.91513115e-01 2.56285936e-01 1.68852359e-01
1.49856761e-01 -1.70812517e-01 -8.65508437e-01 1.16640282e+00
-3.94897312e-01 4.44379479e-01 4.12674904e-01 -2.95887291e-01
1.26942670e+00 5.71132123e-01 6.98913276e-01 -4.60580409e-01
7.39052176e-01 -4.22647744e-01 -1.27915010e-01 -7.21694887e-01
4.08043295e-01 -1.37743086e-01 -9.32149887e-02 8.77275109e-01
-4.09871161e-01 -4.93704647e-01 -2.41659090e-01 2.04403207e-01
5.65134943e-01 2.26777241e-01 7.10721254e-01 1.29264459e-01
5.78551352e-01 -6.40211999e-01 4.52808201e-01 -2.32471272e-01
-6.71364963e-01 1.12678669e-01 4.53715354e-01 -4.46458250e-01
-3.50251466e-01 -8.70586038e-01 2.98061311e-01 1.22861457e+00
2.99683928e-01 1.93832874e-01 -1.06646717e+00 -6.01554692e-01
-1.89491093e-01 5.97716570e-01 -6.60388410e-01 -5.05324125e-01
-3.49759191e-01 -2.07929224e-01 3.90012413e-01 2.14576691e-01
4.00303662e-01 -1.83781493e+00 -4.29081291e-01 -8.82015005e-02
-4.59922142e-02 -1.00011551e+00 -4.70345706e-01 -4.24891591e-01
-4.65766817e-01 -8.01862121e-01 -4.57525820e-01 -1.07913089e+00
9.33425725e-01 2.94193149e-01 8.39611948e-01 2.12747023e-01
-1.91039696e-01 7.35109448e-01 -3.67244810e-01 -6.55566931e-01
-3.83066505e-01 -4.13907677e-01 4.64837849e-01 3.42670202e-01
4.88864750e-01 -9.80039775e-01 -2.91881651e-01 3.81351233e-01
-5.65783799e-01 3.48332405e-01 -2.69282132e-01 2.29799241e-01
-8.36415812e-02 -6.41874611e-01 8.02314281e-01 -1.07023597e+00
9.98000503e-01 -3.72718960e-01 -1.94187433e-01 1.50683835e-01
1.74994349e-01 -3.71033162e-01 4.17957872e-01 -6.39563620e-01
-1.50247848e+00 4.06002581e-01 9.48234126e-02 -1.23819545e-01
-5.02037346e-01 -1.22449227e-01 -2.92623788e-01 -4.10956651e-01
5.84936142e-01 -3.33245397e-01 3.23392987e-01 2.71293104e-01
2.88393974e-01 1.02208447e+00 4.54078227e-01 -5.84596395e-01
6.21113539e-01 4.19809669e-01 -1.66120883e-02 -1.22358871e+00
-2.70678431e-01 1.71890147e-02 -6.11065805e-01 -1.07849145e+00
1.02575314e+00 -5.07373452e-01 -1.67626047e+00 6.61807656e-01
-1.27355707e+00 -2.39326388e-01 1.78550240e-02 2.87274957e-01
-6.22554779e-01 6.57023266e-02 -5.12228847e-01 -1.15446997e+00
-1.72762781e-01 -1.11046529e+00 9.54428375e-01 1.05575287e+00
-1.06164789e+00 -1.05104268e+00 1.74029127e-01 4.43763018e-01
2.28165686e-01 4.89825755e-01 4.24678504e-01 -3.71478677e-01
-8.23489353e-02 7.96114653e-02 1.09128639e-01 -2.97285110e-01
6.65695608e-01 5.40452957e-01 -1.00718725e+00 1.02983244e-01
-3.68468404e-01 -6.86756492e-01 -2.34803170e-01 3.99498671e-01
9.68365252e-01 -1.77270710e-01 -3.22696686e-01 3.09935033e-01
2.73902714e-01 1.08452153e+00 7.83463299e-01 -2.96419680e-01
5.15534759e-01 1.33008087e+00 6.49914026e-01 6.72601044e-01
5.07266939e-01 7.09062874e-01 3.57591808e-01 -1.85641065e-01
3.35904658e-01 -1.63099214e-01 5.18744648e-01 4.52529132e-01
-3.74817818e-01 -1.23897336e-01 -5.29468119e-01 1.91208541e-01
-1.99834144e+00 -1.01106882e+00 -2.71472782e-01 1.95953643e+00
7.30279088e-01 -5.69970787e-01 2.26114154e-01 -3.45694959e-01
1.01263654e+00 3.42762806e-02 -6.66899264e-01 -9.86262918e-01
2.94572145e-01 -5.41823730e-02 -3.16039264e-01 6.80219889e-01
-7.60067880e-01 1.21561635e+00 6.32470798e+00 -1.37314305e-01
-1.15832782e+00 -1.78939268e-01 9.03363764e-01 -1.66198775e-01
-3.41300875e-01 -4.85987127e-01 -5.01598895e-01 -3.00666131e-02
5.59001327e-01 -4.63783562e-01 7.53133476e-01 5.67711949e-01
5.85057974e-01 -8.67368206e-02 -1.17064846e+00 1.30054903e+00
1.45783961e-01 -7.11710393e-01 -3.70964378e-01 -2.22697973e-01
7.75730848e-01 -9.56390083e-01 2.16484889e-01 3.04263115e-01
6.84834003e-01 -1.28540909e+00 1.59982711e-01 6.29551768e-01
5.06603420e-01 -9.65653539e-01 3.34319204e-01 1.70033798e-01
-1.18250823e+00 3.17155063e-01 3.42267245e-01 -3.82499218e-01
4.79392231e-01 -2.26020619e-01 -9.21581507e-01 -4.26586688e-01
5.66206455e-01 4.18472379e-01 -2.44374573e-02 4.82114732e-01
-1.16287231e-01 -8.73505101e-02 -8.53244290e-02 -6.80800974e-01
-1.73069701e-01 -6.71024203e-01 5.77749431e-01 6.89861655e-01
1.36971418e-02 7.08723903e-01 5.34606865e-03 6.48265958e-01
-1.93306208e-01 3.28867406e-01 -1.05406380e+00 -2.57690251e-01
2.03321680e-01 1.43379998e+00 -6.35577798e-01 -8.05983171e-02
-1.51799589e-01 1.14187503e+00 2.70499498e-01 3.17206502e-01
-1.04324293e+00 -1.88137084e-01 1.28328192e+00 1.34099022e-01
-8.10388803e-01 4.80719283e-02 4.28863429e-02 -9.02292371e-01
-1.93951130e-01 -9.51712310e-01 2.62267012e-02 -1.02223182e+00
-9.67919052e-01 7.02901602e-01 -1.91749692e-01 -9.48404312e-01
-7.36501873e-01 -2.95621082e-02 -7.37906933e-01 4.82103944e-01
-5.57702065e-01 -1.14380181e+00 -6.98094726e-01 6.09941006e-01
3.52849156e-01 -2.20806360e-01 9.78020787e-01 7.12039694e-02
-5.35170734e-01 5.40556788e-01 -7.37620771e-01 -8.12799856e-02
8.94118011e-01 -8.37973416e-01 1.20451830e-01 4.36948873e-02
-2.76068181e-01 6.05749667e-01 9.08583462e-01 -5.85154593e-01
-1.22850418e+00 -6.55529141e-01 7.29230106e-01 -1.91197470e-01
2.07106143e-01 -2.23889887e-01 -5.80687821e-01 9.38442349e-01
3.16489458e-01 -1.51784629e-01 7.89242506e-01 7.66867176e-02
2.66890675e-01 1.08863451e-02 -1.74230361e+00 1.20347297e+00
1.22828090e+00 -4.42996025e-01 -2.74534106e-01 3.37129414e-01
5.70206046e-01 -5.14232218e-01 -4.08753008e-01 2.47137725e-01
8.31684768e-01 -1.21884179e+00 7.19985068e-01 -8.13970149e-01
2.39858598e-01 -8.36562365e-02 2.32238188e-01 -1.49848914e+00
2.90526561e-02 -1.07949650e+00 3.83170485e-01 1.32596076e+00
8.66822973e-02 -7.47201741e-01 1.18038309e+00 1.37499976e+00
2.00342715e-01 -4.05550331e-01 -3.66994619e-01 3.66256721e-02
-5.37729502e-01 7.95470402e-02 9.62133288e-01 1.25774813e+00
7.14881182e-01 2.96137840e-01 -5.37588358e-01 -2.06294656e-02
2.07855925e-01 -2.08282948e-01 1.34067595e+00 -1.29781973e+00
-2.28597838e-02 -2.23504156e-01 -5.39537847e-01 -7.41831362e-01
7.27317452e-01 -4.42133307e-01 3.76598388e-02 -1.20690954e+00
2.10267901e-01 -3.25451553e-01 4.24688995e-01 4.25626695e-01
5.92635609e-02 1.77561969e-01 -1.78078096e-02 -4.15148139e-01
-4.61610228e-01 8.97369683e-01 1.60283780e+00 2.70242870e-01
-7.81444371e-01 2.41764933e-01 -5.64144552e-01 1.15463686e+00
6.61659837e-01 4.87759672e-02 -6.43040597e-01 2.66942799e-01
1.78652927e-01 7.39060342e-01 -4.09193896e-02 -5.43836057e-01
9.50464830e-02 -7.38049150e-01 2.24135056e-01 -2.10945657e-03
7.87182093e-01 -8.95762205e-01 4.42714989e-01 3.15393418e-01
-4.06289488e-01 8.13112706e-02 3.00704082e-03 1.39434217e-02
-6.89915493e-02 -1.36253424e-03 1.07002854e+00 -9.37106684e-02
-3.84699792e-01 3.12210768e-01 -8.05254817e-01 -5.23152053e-01
1.84031856e+00 -2.43480995e-01 -8.70909393e-02 -1.57857811e+00
-1.02629054e+00 5.08729935e-01 5.23408532e-01 7.95500815e-01
6.83120966e-01 -1.48697925e+00 -3.51251155e-01 2.99031764e-01
-9.29647014e-02 -2.18752533e-01 4.60918814e-01 3.46984833e-01
-5.26801288e-01 -1.11676008e-01 -6.30586386e-01 -2.72923142e-01
-2.24988151e+00 -1.47091284e-01 5.44149399e-01 1.88121870e-01
-1.22096352e-01 9.54148114e-01 6.57270432e-01 -9.72530007e-01
4.88485277e-01 9.59229767e-02 -1.00827801e+00 5.04657580e-03
8.06143224e-01 4.77563798e-01 -4.78697389e-01 -1.14767301e+00
-8.88097137e-02 4.95443225e-01 1.36726722e-01 -2.42719188e-01
1.27825618e+00 -8.50587785e-02 -3.37900400e-01 5.40990889e-01
8.07799935e-01 1.31840371e-02 -9.95505214e-01 3.31055552e-01
-5.75926065e-01 -5.43441474e-01 -3.23977172e-01 -4.44638789e-01
-1.32064891e+00 4.98408794e-01 4.81284380e-01 2.17818424e-01
9.25985813e-01 -7.19195977e-02 9.66419876e-01 4.04275417e-01
5.81463099e-01 -1.03102243e+00 3.97616267e-01 4.48941499e-01
1.03584099e+00 -1.10949957e+00 -1.96351737e-01 -6.83721900e-01
-1.16096628e+00 9.23503816e-01 1.19912899e+00 1.04968853e-01
7.39040792e-01 5.66675425e-01 5.57228446e-01 -2.69497246e-01
-8.59523356e-01 1.50612056e-01 1.78929605e-02 1.02483082e+00
7.68754184e-01 3.70763749e-01 -2.25869939e-01 6.09437287e-01
-4.43193465e-01 -4.41017821e-02 6.47065043e-01 7.80527711e-01
-4.90056306e-01 -1.23263264e+00 -6.24067545e-01 1.53013229e-01
-2.34649032e-01 7.61544049e-01 -1.12756455e+00 3.77665997e-01
-7.31775686e-02 1.21586740e+00 1.81327477e-01 -5.22189558e-01
4.59417611e-01 1.10540755e-01 5.49440980e-01 -6.57044351e-01
-9.07489657e-01 -4.11336958e-01 3.82029861e-01 -6.94914281e-01
-5.67455947e-01 -8.09923351e-01 -2.06994820e+00 -7.84974158e-01
-4.52857725e-02 4.93162870e-02 6.03364468e-01 7.70082653e-01
2.17205286e-01 2.99618125e-01 9.35473680e-01 -1.29749596e+00
1.21491157e-01 -8.44792783e-01 -6.05627120e-01 7.34805703e-01
-9.43028182e-03 -8.12126637e-01 -2.23295763e-01 4.51098792e-02]
|
[13.145583152770996, -0.34196510910987854]
|
0c1ebf6c-3fa5-4ee5-a0d0-905d00f26b85
|
the-neural-architecture-of-language
| null | null |
https://www.pnas.org/content/118/45/e2105646118
|
https://www.pnas.org/doi/epdf/10.1073/pnas.2105646118
|
The neural architecture of language: Integrative modeling converges on predictive processing
|
The neuroscience of perception has recently been revolutionized with an integrative modeling approach in which computation, brain function, and behavior are linked across many datasets and many computational models. By revealing trends across models, this approach yields novel insights into cognitive and neural mechanisms in the target domain. We here present a systematic study taking this approach to higher-level cognition: human language processing, our species' signature cognitive skill. We find that the most powerful "transformer" models predict nearly 100% of explainable variance in neural responses to sentences and generalize across different datasets and imaging modalities (functional MRI and electrocorticography). Models' neural fits ("brain score") and fits to behavioral responses are both strongly correlated with model accuracy on the next-word prediction task (but not other language tasks). Model architecture appears to substantially contribute to neural fit. These results provide computationally explicit evidence that predictive processing fundamentally shapes the language comprehension mechanisms in the human brain.
|
['Evelina Fedorenko', 'Joshua Tenenbaum', 'Nancy Kanwisher', 'Eghbal Hosseini', 'Carina Kauf', 'Greta Tuckute', 'Idan Blank', 'Martin Schrimpf']
|
2021-11-04
| null | null | null |
proceedings-of-the-national-academy-of-1
|
['probing-language-models']
|
['natural-language-processing']
|
[ 3.36427540e-01 -1.16328493e-01 2.22161204e-01 -4.49861079e-01
-2.68062145e-01 -5.89323044e-01 9.57824647e-01 1.97374523e-01
-7.39790797e-01 1.42851800e-01 5.33396304e-01 -4.05665338e-01
-3.89464855e-01 -4.89667624e-01 -3.62690359e-01 -2.24695474e-01
-4.29956876e-02 1.31603628e-01 -1.51643157e-01 -1.59117803e-01
7.32992530e-01 2.63317436e-01 -1.30744171e+00 5.13160288e-01
1.02211380e+00 9.94113326e-01 6.44800663e-01 5.60684144e-01
2.19712749e-01 6.57756209e-01 9.50018466e-02 -5.91424108e-01
-9.01177377e-02 -5.33937097e-01 -6.22846186e-01 -4.54428732e-01
3.98001581e-01 3.60278748e-02 -1.18067630e-01 9.62123513e-01
6.33458436e-01 6.52138889e-02 7.81249583e-01 -3.99243444e-01
-1.06624591e+00 6.39911592e-01 -2.15347126e-01 7.53437102e-01
1.87189162e-01 6.36976004e-01 8.91863227e-01 -1.05340731e+00
5.50267935e-01 1.56963789e+00 7.14562178e-01 4.44849998e-01
-1.70435619e+00 -5.18970430e-01 1.07672781e-01 2.45959476e-01
-1.04800153e+00 -6.74027383e-01 2.73277134e-01 -9.91370916e-01
1.39546049e+00 1.10723246e-02 1.14880693e+00 9.67174649e-01
1.10341787e+00 -8.38175416e-03 1.71233404e+00 -2.49715909e-01
1.93174198e-01 -1.08231548e-02 5.41757524e-01 6.41103268e-01
3.28029245e-01 5.84267020e-01 -9.65132236e-01 -1.71342447e-01
4.62799639e-01 -2.07907662e-01 -1.77514866e-01 3.90998393e-01
-1.42404366e+00 6.43058777e-01 3.24471056e-01 3.74250799e-01
-6.94087803e-01 2.47649670e-01 4.70067143e-01 1.26191586e-01
4.18398887e-01 7.80594826e-01 -6.27449334e-01 1.27130330e-01
-9.93716419e-01 1.62443340e-01 3.57376814e-01 1.46830633e-01
3.62892300e-01 3.87613147e-01 -2.35192508e-01 1.22624457e+00
4.99095082e-01 6.57201409e-01 8.71962190e-01 -9.33445334e-01
7.75530711e-02 1.58796266e-01 -4.62916970e-01 -1.20157504e+00
-1.05556107e+00 -7.02468693e-01 -5.27535260e-01 -3.44549469e-03
4.69035916e-02 1.02811009e-01 -5.79711020e-01 2.12837458e+00
-4.21266705e-01 -7.18848765e-01 -3.47606301e-01 8.08565378e-01
5.68433046e-01 3.82269770e-01 9.36377764e-01 -3.24439079e-01
1.56703198e+00 -4.20934737e-01 -2.06990913e-01 -9.84222353e-01
5.35110295e-01 -4.29931313e-01 9.97117877e-01 4.08313513e-01
-1.58596647e+00 -8.60313356e-01 -8.02498579e-01 -3.60648066e-01
-2.00703174e-01 -1.85027033e-01 1.00649071e+00 6.64003372e-01
-1.45477331e+00 3.69262964e-01 -6.97969437e-01 -4.11050588e-01
5.82639098e-01 1.89218298e-01 -3.44611585e-01 -4.88516502e-02
-9.59285140e-01 1.60380054e+00 3.88390869e-01 -1.73044100e-01
-8.46866012e-01 -8.48990023e-01 -4.31405485e-01 3.07908177e-01
-1.89384460e-01 -1.17996001e+00 1.09535933e+00 -1.19506133e+00
-9.79425371e-01 1.41474819e+00 -5.74524879e-01 -3.27403635e-01
-3.13057393e-01 9.38653946e-02 -4.21779305e-01 7.28091300e-02
4.09855656e-02 9.42274153e-01 4.29513186e-01 -7.85424948e-01
2.24149287e-01 -7.96133041e-01 -5.15302777e-01 2.54640430e-01
-9.88123659e-03 2.92821854e-01 2.76695430e-01 -6.93570077e-01
5.03488183e-01 -7.54738092e-01 -9.56234410e-02 2.19328105e-01
2.20712706e-01 -1.01786897e-01 -7.28085816e-01 -1.16019011e+00
1.02383745e+00 -1.93006909e+00 2.17881754e-01 2.18454942e-01
6.00367844e-01 -3.49559247e-01 -2.92368203e-01 2.72144675e-01
-1.97125927e-01 4.56777900e-01 -4.69700783e-01 4.99268286e-02
2.65229791e-01 -3.08887482e-01 -2.57171839e-01 5.32798409e-01
6.98851421e-02 1.58749294e+00 -5.17250478e-01 -1.89870045e-01
-1.56573981e-01 1.93475664e-01 -1.04949367e+00 -1.68286785e-01
-1.76825389e-01 1.83771878e-01 1.86603051e-02 4.24696475e-01
6.48353517e-01 -2.86871612e-01 4.28393930e-01 8.12508017e-02
-2.72781849e-01 5.59462547e-01 -2.76151877e-02 1.76160955e+00
-2.22926691e-01 8.00841331e-01 -1.06087886e-02 -1.04057419e+00
7.55315065e-01 1.35563314e-01 -2.82113016e-01 -1.45059431e+00
2.76214927e-01 7.92620331e-02 1.38638639e+00 -4.30498064e-01
-1.28094122e-01 -6.16228700e-01 7.26747513e-02 5.65822363e-01
4.11107987e-01 -3.28975767e-01 -7.32331350e-02 -1.09398961e-01
7.06854820e-01 3.32165249e-02 5.42370737e-01 -9.36008632e-01
3.47226113e-01 -1.18907794e-01 3.91234756e-01 8.74659181e-01
-2.34642163e-01 1.00316159e-01 4.72371072e-01 -2.42400840e-01
-1.03892589e+00 -1.25755405e+00 -3.47001314e-01 1.51778591e+00
-6.31035268e-01 -1.96966633e-01 -7.49187171e-01 5.31907797e-01
-2.31284618e-01 1.16504455e+00 -7.62888849e-01 -7.44046509e-01
-2.23392144e-01 -1.08211064e+00 5.13013899e-01 6.00614846e-01
1.00795366e-01 -1.17367470e+00 -8.70572329e-01 2.41164565e-01
4.58750576e-02 -8.93808782e-01 -1.34914190e-01 -1.18640266e-01
-1.14643645e+00 -6.59666419e-01 -1.93995059e-01 -5.86252987e-01
3.40246737e-01 2.63016243e-02 1.30109322e+00 2.34712183e-01
-6.63814425e-01 4.72798467e-01 3.52602959e-01 -4.53839272e-01
-3.86180341e-01 -1.14609249e-01 1.96624935e-01 -4.15663630e-01
1.92730054e-01 -6.95277154e-01 -8.48874152e-01 -2.50286371e-01
-6.07419550e-01 3.24677348e-01 7.45783985e-01 6.31313860e-01
4.15449440e-01 -8.93366575e-01 8.57192338e-01 -3.92548412e-01
1.21492481e+00 -7.53020585e-01 -3.62160981e-01 2.65203029e-01
-7.99827278e-01 -4.79538254e-02 3.73717666e-01 -3.21224302e-01
-1.24310887e+00 -4.60149676e-01 -2.83699241e-02 4.42983836e-01
-2.74283975e-01 8.85679543e-01 3.74893427e-01 -1.05736762e-01
7.19865203e-01 6.78606391e-01 1.92694888e-01 -3.01920027e-01
2.13039443e-01 -1.22477598e-01 3.69420558e-01 -7.18340755e-01
1.91204771e-01 1.54981881e-01 -8.92453082e-03 -9.96793389e-01
-5.88709891e-01 5.10530829e-01 -6.86392844e-01 -1.87259749e-01
1.11857760e+00 -1.06491590e+00 -1.12233627e+00 1.88050926e-01
-1.27002847e+00 -2.08443388e-01 3.08278024e-01 9.57235932e-01
-6.59598231e-01 1.05704680e-01 -5.47902644e-01 -8.10184836e-01
-5.45049548e-01 -1.02283394e+00 6.66668713e-01 2.90962141e-02
-7.91977227e-01 -1.07566857e+00 1.04522929e-01 3.53665322e-01
7.19278097e-01 -1.38012588e-01 1.53614497e+00 -7.40565181e-01
-3.12023431e-01 1.46548763e-01 -3.25613439e-01 1.00290999e-01
-6.88942611e-01 -3.75498891e-01 -1.01657033e+00 2.43519336e-01
5.15432835e-01 -5.01063049e-01 1.14942372e+00 6.88023925e-01
1.32676375e+00 8.39824453e-02 -1.47003338e-01 5.08850515e-01
1.24649239e+00 3.04380804e-01 5.39673030e-01 -3.51228654e-01
8.02176595e-02 1.04051292e+00 -2.72154808e-01 -1.59479961e-01
5.10758221e-01 2.48100236e-01 -2.98951328e-01 4.70663458e-01
-2.65355054e-02 -7.72593766e-02 5.98616004e-01 1.08522689e+00
-1.34991348e-01 8.90654176e-02 -1.30746686e+00 2.17805222e-01
-1.11682487e+00 -1.45982504e+00 -1.95108712e-01 2.03091311e+00
7.38908291e-01 1.96067125e-01 -3.65152866e-01 -4.44366068e-01
2.03879029e-01 2.59987488e-02 -7.97577083e-01 -8.49069536e-01
-4.43155438e-01 4.45958436e-01 -6.73848018e-02 4.68202889e-01
-2.85968065e-01 9.99003828e-01 8.38715076e+00 5.12027383e-01
-1.00155079e+00 4.51953053e-01 8.92491579e-01 -3.16547930e-01
-4.78003144e-01 -2.88792878e-01 -1.40533045e-01 1.71218872e-01
1.57215452e+00 -5.06716549e-01 9.77211714e-01 3.84213150e-01
5.20269156e-01 -5.13012588e-01 -1.16131508e+00 6.72434032e-01
3.93595457e-01 -1.49831057e+00 -2.63815839e-02 1.15784316e-03
3.89609516e-01 2.64000297e-01 4.04309720e-01 1.27922133e-01
-2.21549392e-01 -1.48159051e+00 9.63095725e-01 1.26748228e+00
8.58122706e-01 -2.01519817e-01 -6.24945089e-02 5.28818369e-01
-6.18671358e-01 -1.74940497e-01 -5.42957008e-01 -6.33775771e-01
1.69872984e-01 3.33374292e-01 -3.05863023e-01 -6.10655963e-01
4.78142142e-01 3.67835224e-01 -7.12724566e-01 8.17135155e-01
2.11837888e-01 7.27423906e-01 6.45982474e-02 -2.21376121e-01
-3.05646360e-01 1.50528029e-01 2.28703991e-01 1.28614402e+00
2.12028101e-01 5.93531311e-01 -3.58132124e-01 1.65705395e+00
2.66192377e-01 3.03844184e-01 -4.76346791e-01 -3.09249163e-01
4.62347478e-01 1.04299605e+00 -8.97724986e-01 -3.03585887e-01
-3.50980550e-01 6.64931178e-01 4.67207164e-01 2.67374754e-01
-4.77844000e-01 2.71351576e-01 6.81928158e-01 -1.76855505e-01
-2.40237951e-01 -6.34908319e-01 -8.79849553e-01 -9.98124361e-01
-1.63962960e-01 -6.31686687e-01 -1.89639404e-01 -1.30173635e+00
-1.35957289e+00 4.05536503e-01 -1.68095361e-02 -2.07280423e-02
-1.16565533e-01 -1.04118991e+00 -4.26926851e-01 1.53268421e+00
-9.58330035e-01 -7.42369950e-01 1.07350960e-01 4.52662736e-01
4.03715521e-01 -7.23817050e-02 1.18259704e+00 -1.01014204e-01
-3.08241874e-01 2.78166413e-01 -1.30363524e-01 -3.24488759e-01
3.73711914e-01 -6.19235337e-01 5.10829568e-01 5.88229001e-01
-3.65317501e-02 1.13360357e+00 2.82492816e-01 -7.05005705e-01
-1.28061855e+00 -3.20364743e-01 9.45215642e-01 -6.96719348e-01
6.32520735e-01 -6.56020939e-01 -8.21338713e-01 5.86317778e-01
2.66620815e-01 -4.82646734e-01 1.20213354e+00 8.04622173e-02
-5.63986957e-01 2.92089432e-01 -9.05964434e-01 6.06327891e-01
1.56630409e+00 -1.07206511e+00 -9.70750034e-01 3.95272911e-01
6.47463918e-01 3.98224145e-01 -8.82184505e-01 7.67030045e-02
9.60329890e-01 -1.00637579e+00 1.15233529e+00 -9.49978054e-01
8.37535977e-01 3.62952501e-01 -2.42664039e-01 -1.47651756e+00
-9.62393761e-01 9.89853889e-02 1.82372361e-01 5.41892231e-01
5.87037086e-01 -9.73871946e-01 -1.73087910e-01 9.51121390e-01
-2.81039804e-01 -5.89419127e-01 -9.12345171e-01 -1.90460965e-01
6.87304616e-01 -8.20674360e-01 8.69498029e-02 6.84897721e-01
2.18706921e-01 3.41899425e-01 2.52279758e-01 -1.82702437e-01
4.57518041e-01 -2.00523332e-01 -3.03508312e-01 -1.59324312e+00
-2.70265549e-01 -1.02881610e+00 -1.45029366e-01 -4.41101730e-01
4.28929865e-01 -1.56053758e+00 -2.20009446e-01 -1.48169625e+00
6.88612938e-01 1.04746930e-01 -1.02208361e-01 2.61187911e-01
-6.31833225e-02 8.91411230e-02 4.17868406e-01 8.95367265e-02
9.71877947e-03 3.98484647e-01 1.06802046e+00 3.35423499e-01
3.38073522e-01 -6.41313314e-01 -1.27135479e+00 9.60269153e-01
9.36963379e-01 -1.05902761e-01 -4.11821544e-01 -7.23471224e-01
5.71987092e-01 2.14689791e-01 7.72507846e-01 -1.00995886e+00
3.38017680e-02 -3.50608259e-01 1.05521834e+00 2.93023456e-02
3.87200058e-01 -3.14755708e-01 -1.16175257e-01 9.53638196e-01
-7.81667948e-01 6.81819856e-01 5.54413259e-01 1.65843159e-01
4.51246649e-01 -2.48473193e-02 9.50566173e-01 -3.88675332e-01
-6.92581952e-01 -1.83172524e-01 -9.63487148e-01 3.69657665e-01
4.92251188e-01 -2.29792222e-01 -6.19601011e-01 -1.97063431e-01
-1.03450739e+00 -3.52515161e-01 -4.94303294e-02 4.12029713e-01
6.87300742e-01 -9.12196517e-01 -8.46260369e-01 1.40447542e-01
-3.06804508e-01 -1.24509537e+00 4.90505487e-01 1.21636987e+00
-2.31941789e-01 8.96389663e-01 -4.53725398e-01 -1.48030475e-01
-5.28248847e-01 5.32720923e-01 5.74888825e-01 3.43327641e-01
-1.26524508e-01 1.01576865e+00 5.84584117e-01 -2.43740305e-01
-4.27180916e-01 -3.39334995e-01 -1.45081356e-01 3.27807307e-01
7.00883865e-01 3.00991178e-01 -2.78825015e-01 -6.93683684e-01
-3.81742626e-01 5.00817776e-01 1.05216578e-01 -3.46579552e-01
1.35338581e+00 -1.59368530e-01 -7.36680865e-01 8.36479664e-01
7.72710860e-01 -7.12140575e-02 -6.83818460e-01 -3.36050242e-02
-5.72605617e-02 4.98240031e-02 9.74958483e-03 -1.49620271e+00
-7.34960556e-01 1.29168856e+00 6.85313582e-01 -2.71861106e-01
8.88494968e-01 6.08660914e-02 5.68701103e-02 3.77346039e-01
3.27079117e-01 -1.23122561e+00 -9.53611508e-02 6.21716559e-01
1.23504603e+00 -9.79144156e-01 8.39878768e-02 1.23558585e-02
-7.11861789e-01 8.40215921e-01 7.83807278e-01 -2.79558808e-01
8.40583086e-01 -1.28754824e-01 -4.61575001e-01 -3.29059958e-01
-1.30400097e+00 -4.35601734e-02 7.33971715e-01 4.25002486e-01
1.04841447e+00 2.74788678e-01 -9.61211860e-01 1.23320735e+00
-6.80663168e-01 -2.61936873e-01 1.12143911e-01 3.27729046e-01
-8.48971009e-01 -5.19340217e-01 -4.49910223e-01 9.02439952e-01
-2.61973262e-01 -8.79585207e-01 -3.80728513e-01 4.76711541e-01
2.37519518e-01 6.30968273e-01 3.88121635e-01 -6.32037818e-02
2.37354562e-02 8.98090780e-01 6.86074615e-01 -6.75140083e-01
-8.19362760e-01 -3.73159349e-02 9.09403414e-02 -6.33074641e-01
-2.18172669e-01 -9.98528421e-01 -1.36737251e+00 -3.69697452e-01
2.26108745e-01 -5.37794709e-01 5.52940845e-01 1.01208210e+00
6.02736950e-01 4.26926553e-01 -3.79611701e-01 -7.84066439e-01
-4.76123452e-01 -1.08476651e+00 -3.83519709e-01 -2.05265563e-02
-1.40136436e-01 -5.80124319e-01 -7.11862594e-02 1.25482753e-01]
|
[10.219429969787598, 8.331756591796875]
|
da9fab5b-0ca7-44a9-8921-cc0bf757eba5
|
abhe-all-attention-based-homography
|
2212.03029
| null |
https://arxiv.org/abs/2212.03029v3
|
https://arxiv.org/pdf/2212.03029v3.pdf
|
AbHE: All Attention-based Homography Estimation
|
Homography estimation is a basic computer vision task, which aims to obtain the transformation from multi-view images for image alignment. Unsupervised learning homography estimation trains a convolution neural network for feature extraction and transformation matrix regression. While the state-of-theart homography method is based on convolution neural networks, few work focuses on transformer which shows superiority in highlevel vision tasks. In this paper, we propose a strong-baseline model based on the Swin Transformer, which combines convolution neural network for local features and transformer module for global features. Moreover, a cross non-local layer is introduced to search the matched features within the feature maps coarsely. In the homography regression stage, we adopt an attention layer for the channels of correlation volume, which can drop out some weak correlation feature points. The experiment shows that in 8 Degree-of-Freedoms(DOFs) homography estimation our method overperforms the state-of-the-art method.
|
['Xianqiang Yang', 'Xinyang Ren', 'Zhihao Zhang', 'Mingxiao Huo']
|
2022-12-06
| null | null | null | null |
['homography-estimation']
|
['computer-vision']
|
[-5.47211953e-02 -2.57807404e-01 4.71415259e-02 -3.14160258e-01
-4.25742567e-01 -3.00544035e-02 6.93342149e-01 -6.65329695e-01
-2.93720961e-01 2.15946555e-01 2.70915180e-01 2.77502626e-01
-1.05599694e-01 -8.64565015e-01 -8.62393916e-01 -7.74253368e-01
3.89142215e-01 5.62452137e-01 2.47102603e-01 -2.89685398e-01
4.33400184e-01 5.05574584e-01 -1.22816586e+00 5.60592152e-02
6.96019292e-01 9.24832165e-01 1.12401798e-01 4.43957895e-01
3.69883627e-01 6.90820992e-01 -4.55517769e-02 -4.48904842e-01
4.77802128e-01 -3.89349967e-01 -4.67973113e-01 1.33256204e-02
1.05107665e+00 -6.34060204e-01 -9.24868524e-01 1.31768155e+00
5.64034343e-01 -3.20419446e-02 5.50079644e-01 -1.23819411e+00
-7.54419148e-01 1.71826735e-01 -6.76820397e-01 -7.99405575e-02
2.13021100e-01 4.31002006e-02 1.02844119e+00 -1.09449172e+00
9.20828223e-01 1.12362266e+00 6.78869426e-01 4.38712239e-02
-9.29330170e-01 -7.46271431e-01 -3.49461615e-01 3.86734545e-01
-1.22653353e+00 -2.90235937e-01 1.10145831e+00 -4.65000242e-01
1.33188307e+00 3.79937366e-02 6.53613329e-01 8.40947986e-01
7.01080322e-01 6.45243108e-01 9.85400617e-01 -1.95868477e-01
-4.24338192e-01 -2.33732790e-01 -2.19548672e-01 9.15097237e-01
-4.57029343e-02 5.82702696e-01 -6.05529010e-01 2.29432702e-01
1.41397619e+00 2.80904889e-01 -3.92748475e-01 -9.07299995e-01
-1.59208274e+00 7.93933928e-01 8.07913244e-01 1.75403908e-01
-2.79977590e-01 1.23817950e-01 2.23562166e-01 2.59392411e-01
1.53063580e-01 4.66044486e-01 -1.58541858e-01 -1.46303009e-02
-9.13106561e-01 2.83781551e-02 5.94673753e-01 1.29939365e+00
1.06685042e+00 6.23427406e-02 1.12180086e-02 6.43067181e-01
1.85831085e-01 5.71607709e-01 3.81131470e-01 -8.53112817e-01
6.57332361e-01 7.67375529e-01 -3.85303617e-01 -1.41312826e+00
-5.23834467e-01 -6.06574714e-01 -1.34071052e+00 6.78852946e-02
2.91207321e-02 8.12884718e-02 -9.12658274e-01 1.19196773e+00
1.00667387e-01 1.23127736e-01 -3.12124759e-01 1.29398465e+00
8.50149214e-01 3.50329101e-01 -7.74215519e-01 2.95175090e-02
1.16930020e+00 -1.41360819e+00 -5.72036028e-01 -1.57863289e-01
2.10236803e-01 -1.23473442e+00 7.44954348e-01 1.83163628e-01
-1.01794112e+00 -6.81900263e-01 -1.25352216e+00 -7.20083654e-01
-3.83483618e-01 3.26134890e-01 7.39277899e-01 4.81553413e-02
-9.21758056e-01 6.99142933e-01 -8.45103979e-01 -3.49396408e-01
1.27265096e-01 5.64061642e-01 -8.45723748e-01 1.26132265e-01
-9.66384113e-01 9.13950384e-01 3.27705234e-01 2.05755919e-01
-4.52508211e-01 -5.91389477e-01 -9.54728425e-01 2.37287283e-01
2.44511023e-01 -1.18938816e+00 7.64605224e-01 -4.89051968e-01
-1.61593127e+00 9.17649031e-01 4.75732051e-02 -1.97207898e-01
7.09445000e-01 -3.41372967e-01 -1.93534091e-01 1.36785701e-01
9.92863625e-02 6.22773409e-01 1.07141733e+00 -9.84136283e-01
-5.38572013e-01 -4.76638943e-01 -2.51184553e-01 3.69957060e-01
-1.52089447e-01 -6.66033626e-02 -8.76510024e-01 -5.11970103e-01
8.43699217e-01 -9.45669115e-01 -1.11798286e-01 -1.38651013e-01
-4.77059811e-01 1.79893434e-01 8.52157295e-01 -8.34544301e-01
1.08488059e+00 -1.92383742e+00 3.63618851e-01 4.08887118e-01
5.39599776e-01 -1.74922422e-01 1.09062985e-01 3.89445782e-01
-2.35558048e-01 -5.28482080e-01 1.72119632e-01 -2.40614429e-01
1.23062842e-02 -1.52193114e-01 -2.10111693e-01 8.65249336e-01
-3.82804684e-02 1.06366301e+00 -6.29046977e-01 -3.12664300e-01
7.26087749e-01 6.42470360e-01 -6.72482014e-01 3.93399775e-01
3.91153574e-01 4.38561976e-01 -3.09235781e-01 5.87959111e-01
9.24651504e-01 -3.75563085e-01 -1.24203861e-01 -9.46277201e-01
-3.89262497e-01 7.16970712e-02 -9.74619329e-01 2.01052618e+00
-4.95439261e-01 6.54394865e-01 -2.23932713e-01 -8.00759435e-01
1.11543489e+00 -7.04655424e-02 8.29897881e-01 -6.88565612e-01
6.74228728e-01 1.79962456e-01 9.95676070e-02 -2.35841632e-01
3.68703574e-01 4.43926126e-01 1.38980359e-01 5.50648905e-02
3.83469671e-01 -3.01299840e-01 -1.84393391e-01 -1.12678722e-01
7.93111861e-01 2.37904593e-01 4.55272198e-01 -1.74171776e-01
6.25756741e-01 -2.13792205e-01 3.08115393e-01 4.43973750e-01
1.58787426e-02 1.00852430e+00 2.51377970e-01 -8.19732487e-01
-1.32991958e+00 -8.57939243e-01 2.03451179e-02 4.80047584e-01
4.14352715e-01 -5.17846167e-01 -6.97821140e-01 -3.49853218e-01
2.46367278e-03 -2.80570537e-02 -5.70919633e-01 -1.47989541e-01
-7.94617653e-01 -4.73983616e-01 7.84316659e-02 5.94810188e-01
1.29651117e+00 -7.41811931e-01 -7.12064803e-01 5.85909337e-02
-2.51209158e-02 -1.38273311e+00 -8.55162799e-01 2.36201659e-02
-7.37021327e-01 -1.17947459e+00 -6.65496767e-01 -8.48553181e-01
8.41535449e-01 3.80630523e-01 9.19284523e-01 -2.65924931e-01
-1.14137769e-01 1.35776952e-01 6.67151511e-02 1.40285775e-01
2.24063754e-01 2.76803881e-01 -7.78730139e-02 -6.92265555e-02
6.62832081e-01 -9.59443867e-01 -7.74220526e-01 6.41006529e-01
-3.70617181e-01 4.78777289e-01 8.30474615e-01 1.21176624e+00
6.61755860e-01 -3.59498084e-01 -3.97312462e-01 -7.04669893e-01
3.60498071e-01 1.26993582e-01 -1.07205606e+00 3.87781173e-01
-6.73169434e-01 1.94705576e-01 7.36393690e-01 -2.52829939e-01
-6.99477851e-01 5.91848671e-01 1.29336596e-01 -1.06286216e+00
8.07981193e-02 2.66540885e-01 -1.87314793e-01 -6.66143477e-01
4.46317255e-01 3.03395718e-01 2.66411211e-02 -4.40009147e-01
2.93868095e-01 3.76857042e-01 8.67578924e-01 -6.13166392e-02
1.26008964e+00 5.40057719e-01 3.78254831e-01 -5.47905803e-01
-5.24548769e-01 -6.91515326e-01 -1.08882070e+00 -7.96606317e-02
1.11282110e+00 -1.04300773e+00 -9.78475749e-01 4.88897830e-01
-1.27441156e+00 2.06708640e-01 3.82677883e-01 9.00019825e-01
-7.63365209e-01 4.61931258e-01 -4.95099992e-01 -7.65927583e-02
-4.59598064e-01 -1.55902553e+00 1.36207986e+00 9.87475440e-02
1.20782033e-01 -7.25664437e-01 1.00695007e-01 3.40103865e-01
3.66093069e-01 5.50841987e-02 5.87868154e-01 -3.18934053e-01
-1.16850019e+00 -3.15780520e-01 -5.00121295e-01 2.13497937e-01
-1.57698572e-01 -1.23664476e-01 -9.10325110e-01 -2.94222146e-01
1.64491341e-01 -1.38838083e-01 9.71612990e-01 5.15750587e-01
9.88532484e-01 -1.27597541e-01 -8.62627625e-02 1.58927834e+00
1.40467584e+00 1.03928037e-01 9.62834716e-01 6.50796592e-01
1.36775255e+00 3.14489901e-01 3.80655259e-01 3.26159328e-01
4.27284837e-01 8.98773789e-01 4.88011926e-01 -4.73987281e-01
-5.12145534e-02 -4.42492038e-01 6.28378987e-02 1.19899428e+00
-3.66128296e-01 3.44773322e-01 -6.88613713e-01 1.86987430e-01
-2.14492369e+00 -7.75160313e-01 1.42519083e-03 2.23926377e+00
3.48763376e-01 1.88630726e-02 -2.00740114e-01 -2.28979886e-01
5.14034986e-01 3.37416053e-01 -4.47405607e-01 -1.89329684e-02
-1.00817271e-01 -8.52631032e-02 8.23628247e-01 5.58019817e-01
-1.41186523e+00 1.20323527e+00 5.59665966e+00 6.40291750e-01
-1.36949599e+00 -1.96428537e-01 5.71340183e-03 2.65288830e-01
6.04390129e-02 2.76519418e-01 -7.59980798e-01 1.07588425e-01
1.64468028e-02 -1.24891689e-02 6.90416455e-01 9.42944884e-01
-1.81935191e-01 1.48184285e-01 -1.20998681e+00 1.67067862e+00
5.25841236e-01 -1.42469752e+00 1.07644834e-01 1.62983492e-01
9.10742640e-01 3.91218722e-01 7.08192587e-02 5.17652463e-03
2.54120864e-03 -9.36309040e-01 1.85356095e-01 5.89319885e-01
7.43032634e-01 -6.46060348e-01 9.39197481e-01 1.13659389e-01
-1.34627211e+00 2.34462500e-01 -7.02654004e-01 1.77588761e-01
1.43780887e-01 3.44879240e-01 -5.42496324e-01 8.35804284e-01
6.94829643e-01 1.00336289e+00 -6.59305632e-01 1.08079815e+00
-8.37244615e-02 -1.21368818e-01 -3.39436442e-01 3.27997595e-01
3.95859689e-01 -6.71860039e-01 3.01790237e-01 8.49457085e-01
4.56057340e-01 -1.22362703e-01 3.36128622e-01 9.26026583e-01
-9.37582105e-02 1.18298501e-01 -8.24293792e-01 2.63795376e-01
7.75850043e-02 1.38980877e+00 -5.43495119e-01 -2.66466737e-01
-6.52018785e-01 1.18268263e+00 2.42923781e-01 3.48741800e-01
-6.42732441e-01 -5.52947462e-01 2.77794600e-01 -4.56285998e-02
2.98489362e-01 -2.28272364e-01 -4.11778510e-01 -1.64310968e+00
3.02374810e-02 -8.53835642e-01 9.82570872e-02 -8.86155725e-01
-1.05556095e+00 6.56248271e-01 -1.08475372e-01 -1.60319173e+00
-4.67318475e-01 -8.70076537e-01 -7.06888497e-01 8.82308006e-01
-1.32671595e+00 -1.56375027e+00 -8.08099329e-01 8.43930304e-01
4.45751965e-01 -6.05830848e-01 5.69690943e-01 3.06299865e-01
-2.07267404e-01 5.63162982e-01 9.51917246e-02 3.61050546e-01
9.77319002e-01 -1.06888497e+00 6.21307373e-01 8.39041173e-01
9.56433937e-02 8.59555483e-01 5.43919563e-01 -6.70169652e-01
-1.66182029e+00 -6.28608108e-01 9.58469927e-01 -2.78023511e-01
6.36083782e-01 -4.11724389e-01 -5.83545029e-01 8.21493983e-01
3.81327778e-01 3.01828414e-01 -7.09204283e-03 -7.51141757e-02
-4.54161316e-01 -2.03325078e-01 -6.80772185e-01 6.13080382e-01
1.14299107e+00 -8.42079699e-01 -2.33365357e-01 2.22574547e-01
3.07965040e-01 -8.89678359e-01 -9.71334517e-01 5.57530701e-01
9.72216964e-01 -1.39381719e+00 1.05865800e+00 -2.84670234e-01
6.79880738e-01 -4.19762433e-01 -8.46410394e-02 -1.20864201e+00
-6.09562635e-01 -7.87038386e-01 -3.25401761e-02 6.89348280e-01
2.71549402e-03 -5.63015521e-01 9.30728376e-01 1.89889088e-01
-2.28208274e-01 -7.37025440e-01 -8.31693172e-01 -5.46082795e-01
-2.68100202e-01 1.74614102e-01 5.87764800e-01 1.06826138e+00
-2.27546304e-01 8.39472771e-01 -7.11271942e-01 2.81556457e-01
7.53062665e-01 4.06093299e-01 1.20406401e+00 -1.00023365e+00
-3.39070141e-01 -5.24322271e-01 -8.76922309e-01 -1.27610373e+00
1.44982412e-01 -7.67993987e-01 -1.69938609e-01 -1.15519571e+00
3.03047091e-01 2.79415607e-01 -3.05984598e-02 1.59627989e-01
1.01895213e-01 1.55022770e-01 2.57820249e-01 2.59336203e-01
-2.10215181e-01 7.23752677e-01 1.66764021e+00 -1.85855821e-01
-8.22858214e-02 -1.23884924e-01 -1.00698024e-01 7.44491279e-01
5.39756715e-01 -3.10335066e-02 -4.37906384e-01 -5.98161757e-01
4.38664332e-02 5.50314635e-02 4.21966523e-01 -1.17811739e+00
8.04469347e-01 -5.32538593e-02 6.92803502e-01 -1.17316127e+00
4.13084060e-01 -1.10688257e+00 3.13724041e-01 2.92922556e-01
-1.44406827e-02 5.73153496e-01 -2.34830782e-01 -3.38169932e-03
-5.31017721e-01 2.36020133e-01 6.24615252e-01 -2.64213264e-01
-5.60058355e-01 7.93043196e-01 3.90748471e-01 -1.68664277e-01
6.60791457e-01 -4.77049083e-01 -5.53737700e-01 -5.18550277e-01
-3.93317074e-01 9.24282297e-02 6.23437166e-01 4.20950353e-01
9.76713538e-01 -1.45283878e+00 -4.56144243e-01 7.86454797e-01
1.27653241e-01 4.82863840e-03 1.49237797e-01 1.15152740e+00
-8.27041209e-01 4.67558026e-01 -7.24032640e-01 -6.99568689e-01
-1.23117375e+00 5.60224116e-01 4.51958627e-01 -2.58840650e-01
-7.79998481e-01 5.60026765e-01 6.07276976e-01 -4.74135101e-01
2.98404604e-01 -1.72664464e-01 -2.79028028e-01 -3.28898191e-01
2.10219979e-01 5.63437462e-01 -2.86328308e-02 -1.03850365e+00
-2.64986813e-01 1.29553330e+00 -2.50313729e-01 2.60805478e-03
1.32473087e+00 6.27692193e-02 -3.96966457e-01 -8.51672068e-02
1.63070166e+00 -1.62178844e-01 -1.27691019e+00 -1.45086825e-01
-2.99262345e-01 -6.69967890e-01 2.53492057e-01 -3.11633140e-01
-1.32876539e+00 1.21542728e+00 6.54426992e-01 -3.44043344e-01
9.91315901e-01 -4.42366451e-01 7.72575974e-01 5.87695956e-01
2.55913079e-01 -1.14883816e+00 1.80918314e-02 7.89261103e-01
1.17244709e+00 -1.43142796e+00 3.17780644e-01 -7.58516610e-01
-5.28915465e-01 1.49032903e+00 1.06327927e+00 -6.64244056e-01
7.74866700e-01 1.28150105e-01 8.13645720e-02 -5.98553002e-01
-3.56887579e-01 -1.49240434e-01 6.55113816e-01 4.85826850e-01
2.68784940e-01 -2.62086302e-01 -1.18436784e-01 8.23948309e-02
-6.15893602e-01 -2.25803018e-01 1.38171092e-01 4.76372063e-01
-1.93862021e-01 -9.88030910e-01 -2.53940165e-01 2.74840176e-01
9.80007462e-04 -3.20241958e-01 -5.97093642e-01 1.04833210e+00
1.34126535e-02 1.67562231e-01 8.13249275e-02 -8.43449235e-01
4.96389598e-01 -4.13906723e-01 7.22378850e-01 -9.18391794e-02
-7.20467567e-01 6.10973656e-01 -1.80195600e-01 -1.01870835e+00
-4.82381791e-01 -2.99692065e-01 -7.68973768e-01 -4.97952044e-01
-3.14317465e-01 -3.87056381e-01 4.25897360e-01 8.94496977e-01
1.70995161e-01 3.53680193e-01 7.56833375e-01 -1.10419476e+00
-5.51032603e-01 -9.22844410e-01 -4.32345301e-01 4.52248931e-01
4.34189707e-01 -8.70111823e-01 -1.15104668e-01 -8.06719065e-02]
|
[8.64319133758545, -2.2339982986450195]
|
a294adb6-f6b4-4f2f-a639-4679caaf64c8
|
stationarity-analysis-of-the-stock-market
|
2112.12459
| null |
https://arxiv.org/abs/2112.12459v3
|
https://arxiv.org/pdf/2112.12459v3.pdf
|
Stationarity analysis of the stock market data and its application to mechanical trading
|
This study proposes a scheme for stationarity analysis of stock price fluctuations based on KM$_2$O-Langevin theory. Using this scheme, we classify the time-series data of stock price fluctuations into three periods: stationary, non-stationary, and intermediate. We then suggest an example of a low-risk stock trading strategy to demonstrate the usefulness of our scheme by using actual stock index data. Our strategy uses a trend-based indicator, moving averages, for stationary periods and an oscillator-based indicator, psychological lines, for non-stationary periods to make trading decisions. Finally, we confirm that our strategy is a safe trading strategy with small maximum drawdown by back testing on the Nikkei Stock Average. Our study, the first to apply the stationarity analysis of KM$_2$O-Langevin theory to actual mechanical trading, opens up new avenues for stock price prediction.
|
['Norikazu Todoroki', 'Kazuki Kanehira']
|
2021-12-23
| null | null | null | null |
['stock-price-prediction']
|
['time-series']
|
[-5.93334317e-01 -4.71804023e-01 -1.19309552e-01 3.00547779e-01
-1.92721635e-01 -8.62600684e-01 5.48565447e-01 -2.77262747e-01
-4.14180070e-01 1.01140690e+00 -2.37899661e-01 -4.37128216e-01
-1.58495262e-01 -1.06520557e+00 -4.48354572e-01 -8.76924217e-01
-6.32981420e-01 1.98177949e-01 5.66038430e-01 -4.41399246e-01
7.24465430e-01 3.43054861e-01 -1.07783139e+00 -5.89009583e-01
2.16052815e-01 1.30208933e+00 -4.33313668e-01 6.35841787e-01
-1.12465762e-01 6.81232870e-01 -4.43635881e-01 -2.49542922e-01
8.82428110e-01 -6.14363015e-01 -2.26167277e-01 -5.71107030e-01
-4.44851905e-01 -3.47971886e-01 -1.44042224e-01 1.12752771e+00
1.91789821e-01 -7.69673958e-02 5.25392771e-01 -9.77074146e-01
-5.12012005e-01 8.99587810e-01 -4.23400760e-01 6.88065052e-01
-1.42099887e-01 2.05152333e-01 1.07602084e+00 -3.97836626e-01
3.88735652e-01 5.87299645e-01 9.85430777e-01 2.03760609e-01
-1.40729213e+00 -9.81044054e-01 -3.50159705e-01 -3.98496121e-01
-1.04919529e+00 -8.57574120e-02 1.14364028e+00 -4.89146680e-01
7.64862537e-01 2.16763780e-01 1.20725536e+00 5.20775139e-01
9.78656352e-01 1.95962459e-01 1.43920803e+00 -2.08710387e-01
5.00465870e-01 -1.42384872e-01 3.21923316e-01 2.23153666e-01
8.49234998e-01 7.92532086e-01 -4.95944828e-01 -4.67255026e-01
1.23318553e+00 -1.97936431e-01 7.72576332e-02 2.92872488e-02
-1.10201848e+00 8.15944076e-01 -2.59537339e-01 6.89271450e-01
-4.23035413e-01 6.79070354e-01 1.11514360e-01 8.24251592e-01
7.37144768e-01 6.73623502e-01 -7.33389080e-01 -4.12600815e-01
-1.28045285e+00 6.08997285e-01 1.03795230e+00 4.70293015e-01
5.19515157e-01 4.76144969e-01 1.75400734e-01 -1.00439176e-01
5.76676071e-01 9.57881153e-01 7.23916709e-01 -1.15672100e+00
-1.32300064e-01 -1.40785992e-01 5.98414063e-01 -5.95548928e-01
-4.43786919e-01 -4.55449909e-01 -5.92321515e-01 5.52200735e-01
6.25989377e-01 -4.55284119e-01 -2.27231719e-02 1.49662709e+00
-1.70881405e-01 4.61940825e-01 2.16002077e-01 1.45897791e-01
-1.58073053e-01 5.86511195e-01 -5.04054308e-01 -9.01620150e-01
1.25611305e+00 -2.84508258e-01 -8.30489516e-01 6.82354033e-01
3.37793857e-01 -6.73806548e-01 4.41346884e-01 3.56500387e-01
-1.25347829e+00 -2.26753980e-01 -1.02555060e+00 9.11093891e-01
-5.68165146e-02 -6.86486959e-01 5.79268575e-01 8.46086204e-01
-9.92357671e-01 1.18308198e+00 -9.72771823e-01 4.63668853e-01
-4.13410872e-01 1.95675120e-01 4.38362241e-01 1.47487402e+00
-1.48327434e+00 5.80939770e-01 2.28228420e-01 4.71373685e-02
-2.46906504e-01 -6.24069810e-01 -3.10855687e-01 -1.76426083e-01
-4.78437878e-02 -2.92981893e-01 1.29259920e+00 -3.64351779e-01
-1.95567286e+00 4.23632294e-01 1.60458028e-01 -9.85056579e-01
7.54741371e-01 2.41316393e-01 -5.07801533e-01 -2.25741486e-03
-2.61095464e-02 -2.32683420e-01 8.31426740e-01 -6.78875685e-01
-3.29502136e-01 7.46992603e-02 -4.35656846e-01 -5.16126871e-01
3.21521074e-01 -2.22087409e-02 6.38625264e-01 -1.19005609e+00
2.70069331e-01 -9.92719829e-01 -1.15814909e-01 -6.88775539e-01
-4.47509959e-02 -2.25333840e-01 3.73119384e-01 -3.23220968e-01
1.34054482e+00 -1.93397415e+00 -6.37255371e-01 5.97235084e-01
-5.96415401e-02 -3.64195973e-01 3.02039027e-01 6.03467345e-01
-1.30683452e-01 4.36665922e-01 -1.91352054e-01 1.93088323e-01
1.75857127e-01 -1.70188114e-01 -9.51903641e-01 3.73780221e-01
7.45684206e-02 1.04981077e+00 -6.97557449e-01 -6.66405186e-02
-1.50952682e-01 -2.44364664e-01 -4.51669753e-01 -2.26270139e-01
-2.78146476e-01 5.54596364e-01 -4.50485557e-01 3.26261818e-01
6.27093911e-01 -9.91566554e-02 -2.72764057e-01 2.47515485e-01
-9.73007798e-01 2.51299173e-01 -1.26284456e+00 6.83490813e-01
-7.68068340e-03 6.05481148e-01 -3.86981577e-01 -6.10494316e-01
1.20322204e+00 3.33958566e-01 8.83537769e-01 -6.27707720e-01
1.36972770e-01 7.02770114e-01 1.06631324e-01 -1.02994092e-01
5.51183462e-01 -7.82325029e-01 -3.43512833e-01 1.11610389e+00
-3.88963133e-01 -3.43455613e-01 1.96281523e-01 -4.40600991e-01
8.55273306e-01 2.47978151e-01 2.31253535e-01 -8.46196234e-01
2.10880965e-01 -1.33091822e-01 5.12693524e-01 7.04233825e-01
-4.44585025e-01 1.02532431e-01 9.18417513e-01 -5.82029939e-01
-1.01626098e+00 -1.04330528e+00 -3.16142380e-01 3.25172603e-01
1.58245161e-01 -1.59466937e-01 -5.83066821e-01 1.93925202e-01
4.65369552e-01 7.86024749e-01 -6.28423929e-01 2.87204329e-02
-5.80951154e-01 -1.10030532e+00 4.80777919e-01 2.03424007e-01
7.07274377e-01 -1.15656722e+00 -9.52898860e-01 6.13161325e-01
3.89598399e-01 -4.99185532e-01 -4.36402440e-01 2.79976338e-01
-1.15162253e+00 -7.66949356e-01 -7.39061117e-01 -3.03706855e-01
-6.33377582e-03 -1.22120269e-01 9.98737633e-01 -7.57588148e-02
2.74371743e-01 4.39635634e-01 2.50074118e-02 -6.46122277e-01
-4.70488071e-01 -3.20641696e-01 4.05780554e-01 -3.46611328e-02
3.22041690e-01 -6.89323246e-01 -7.99198985e-01 4.94894266e-01
-7.73234785e-01 -6.52554154e-01 1.15456313e-01 5.03498077e-01
6.31597638e-01 4.09805745e-01 6.11270070e-01 -3.06786656e-01
8.32310915e-01 -3.34440500e-01 -1.51422572e+00 -1.25280730e-02
-9.74605680e-01 5.20419836e-01 3.57938707e-01 -4.20184493e-01
-5.32516897e-01 -3.30783188e-01 1.15190372e-01 -8.65092203e-02
7.45926678e-01 5.10372579e-01 7.51741350e-01 -1.45423383e-01
1.56617686e-01 5.80062807e-01 2.66821593e-01 -4.04309541e-01
-7.97218978e-02 1.56372651e-01 3.69655311e-01 -4.67639446e-01
1.11305964e+00 5.55013537e-01 2.44676188e-01 -8.04334104e-01
-2.47266263e-01 -1.52469337e-01 -6.03962719e-01 -2.42588475e-01
8.03790152e-01 -5.64338863e-01 -1.30407727e+00 1.02481735e+00
-7.93504238e-01 -4.52640712e-01 -6.93688273e-01 6.99817419e-01
-8.36713135e-01 3.14315349e-01 -1.01636147e+00 -1.44338322e+00
-2.18534768e-01 -6.02550805e-01 5.18054962e-01 1.59518972e-01
-2.51169622e-01 -1.37362826e+00 9.13315594e-01 -2.84772396e-01
6.47902429e-01 3.82486701e-01 5.63014984e-01 -8.48905861e-01
-8.76715839e-01 -2.86907017e-01 4.99770701e-01 2.21078098e-01
6.68724552e-02 4.50088263e-01 -6.01640403e-01 -1.31403953e-01
6.85610354e-01 1.69039339e-01 8.43644619e-01 6.47122860e-01
2.62733132e-01 -2.88782060e-01 1.96898729e-01 3.59842271e-01
1.35038269e+00 6.44803464e-01 6.02063119e-01 6.86617136e-01
-1.87821627e-01 2.35233873e-01 2.96475053e-01 6.07850432e-01
2.95566395e-02 2.84710377e-01 8.18892475e-03 4.64498311e-01
6.88283145e-01 -2.58856177e-01 8.71543825e-01 1.29085898e+00
-5.60766518e-01 3.46431673e-01 -5.38032413e-01 8.09461027e-02
-1.55899835e+00 -1.56631505e+00 -2.52755433e-01 2.36083579e+00
8.03556859e-01 7.46349216e-01 6.79517925e-01 6.21926077e-02
4.98751879e-01 3.42500567e-01 -5.31592548e-01 -2.14892715e-01
-1.78729087e-01 4.83889043e-01 1.10213530e+00 4.43567991e-01
-9.67375636e-01 4.98297989e-01 8.08085346e+00 5.69311023e-01
-1.29991579e+00 -5.34729660e-02 3.98782790e-01 1.45492673e-01
-6.49323702e-01 1.92967415e-01 -8.73429179e-01 9.59628224e-01
1.33783054e+00 -8.24690878e-01 3.14386964e-01 4.47238743e-01
5.14924705e-01 -1.79627970e-01 -7.03996778e-01 6.86787784e-01
-6.77905917e-01 -1.55935216e+00 -2.88459539e-01 4.12265718e-01
5.22212446e-01 -1.20369419e-01 1.74834639e-01 1.19443871e-01
3.89442146e-01 -4.46924776e-01 1.15133429e+00 9.75756526e-01
2.75284737e-01 -5.90208292e-01 8.60967278e-01 2.83286035e-01
-1.50957918e+00 3.50228369e-01 -3.41961205e-01 -4.07188714e-01
3.92100781e-01 6.31342173e-01 -2.53262334e-02 3.05529445e-01
3.03578973e-01 5.88617146e-01 -5.66479713e-02 7.95057893e-01
1.69352815e-01 7.98477530e-01 -6.68221653e-01 -4.68039602e-01
1.49229467e-01 -8.96694005e-01 6.08677924e-01 5.73880672e-01
7.34773397e-01 2.17344105e-01 -3.71332526e-01 1.24637890e+00
4.25798088e-01 -3.31806272e-01 -6.28325641e-01 -2.02978536e-01
3.42846930e-01 6.00545943e-01 -1.10990584e+00 -1.78608254e-01
-3.26744020e-01 5.48078179e-01 -7.80429780e-01 1.36477083e-01
-5.55501103e-01 -4.45556819e-01 6.86557055e-01 1.38538778e-01
4.56597745e-01 -7.44986951e-01 -3.88871282e-01 -1.35305738e+00
-1.12900257e-01 -6.17944077e-02 1.38295472e-01 -4.22491193e-01
-1.25254810e+00 1.85279578e-01 2.34835327e-01 -1.63846982e+00
-6.86340868e-01 -7.41017640e-01 -1.01276803e+00 7.66748369e-01
-1.42045689e+00 -1.00490630e-01 7.45760918e-01 3.35582763e-01
3.92945111e-02 -4.74526525e-01 4.98922199e-01 -2.44779125e-01
-4.13578928e-01 1.63739458e-01 6.71447694e-01 3.43685448e-01
2.21811198e-02 -1.38870418e+00 8.57129395e-01 6.89357221e-01
1.94274157e-01 6.31055295e-01 1.09888625e+00 -9.52830017e-01
-1.15711749e+00 -5.26398778e-01 7.11194038e-01 -5.44081569e-01
1.59030414e+00 -7.52666146e-02 -8.55790257e-01 5.40887237e-01
2.25268334e-01 -2.21114531e-01 6.66016281e-01 -4.92593080e-01
1.81173868e-02 -2.25417569e-01 -1.06924701e+00 4.65006560e-01
5.27202189e-01 -3.59666586e-01 -9.35734928e-01 -3.85904424e-02
6.12703443e-01 1.26701429e-01 -1.16234910e+00 3.46903168e-02
1.02799129e+00 -1.24265087e+00 5.66806853e-01 -1.32517427e-01
-8.19451287e-02 -2.65190512e-01 6.67685121e-02 -1.12325037e+00
-2.21124738e-01 -1.41221642e+00 1.57618463e-01 7.96386123e-01
5.32919884e-01 -1.46487308e+00 5.74753284e-01 5.54577589e-01
4.80281234e-01 -2.58308530e-01 -1.33132529e+00 -1.40823328e+00
6.16448879e-01 -5.21862268e-01 5.59320986e-01 7.78173745e-01
9.73731801e-02 -2.54686207e-01 -3.03185314e-01 -2.01850265e-01
9.62393522e-01 4.53391522e-01 4.18366551e-01 -1.49579489e+00
-6.43379033e-01 -8.19714963e-01 -2.69501418e-01 -8.55661094e-01
-1.24541894e-02 -4.59442466e-01 -1.73755705e-01 -4.19268340e-01
-1.20101482e-01 -3.29447806e-01 -6.32299006e-01 -2.05681771e-01
3.71746987e-01 3.05489391e-01 2.14889303e-01 7.94879079e-01
-5.97461164e-02 3.86828572e-01 1.02936780e+00 1.72772408e-01
-3.70583802e-01 5.11928737e-01 -1.46337211e-01 6.39539242e-01
8.19474101e-01 -5.12371600e-01 -8.08613449e-02 5.70987880e-01
6.97736144e-01 3.73657137e-01 1.72938526e-01 -8.23072612e-01
-3.23189162e-02 -3.42992634e-01 -8.23523179e-02 -8.76885891e-01
-1.41395196e-01 -5.34637630e-01 6.36554420e-01 1.24318051e+00
-8.41156468e-02 5.89967012e-01 5.17027602e-02 5.07736802e-01
-2.62451023e-01 -4.46490586e-01 6.27767205e-01 -3.48812073e-01
-4.01242435e-01 1.69372782e-01 -6.69074476e-01 1.79885030e-01
1.11205816e+00 -2.13478029e-01 -2.58943826e-01 -5.76105654e-01
-7.55430400e-01 1.19635649e-01 4.54099655e-01 -1.86382517e-01
1.48099467e-01 -1.51375663e+00 -5.51925004e-01 4.99468029e-01
-5.17914176e-01 -6.93171501e-01 -1.08523257e-01 8.68887484e-01
-9.35278058e-01 5.62818110e-01 -1.90067068e-01 -3.81655663e-01
-4.43980783e-01 3.29091758e-01 5.74423313e-01 -4.48450714e-01
-5.34811735e-01 3.39208037e-01 -3.37054908e-01 3.32254589e-01
-3.07849973e-01 -9.96952355e-01 3.82796675e-02 4.14121777e-01
5.30437589e-01 4.49813187e-01 -2.84798622e-01 -3.50646317e-01
-1.33270949e-01 1.16410315e+00 6.19543552e-01 -6.73906326e-01
1.21303916e+00 -7.82548860e-02 -3.80818218e-01 1.39732659e+00
8.33777189e-01 4.11635518e-01 -1.42044950e+00 1.88749433e-01
6.31537199e-01 1.98551379e-02 -3.02014738e-01 1.21812243e-02
-8.95084918e-01 3.22086811e-01 3.44546884e-01 1.26573956e+00
5.72275758e-01 -2.39678830e-01 1.01279473e+00 4.76719677e-01
8.66328359e-01 -1.10859382e+00 -2.67508715e-01 5.96221685e-01
7.46067047e-01 -8.22250426e-01 -2.64585525e-01 1.43264771e-01
-3.33144188e-01 1.38303602e+00 -3.84359092e-01 -7.99598634e-01
1.62406886e+00 5.40866733e-01 1.43782616e-01 -2.60962769e-02
-8.92589688e-01 -6.25443980e-02 -9.94333699e-02 -1.43630609e-01
2.72607893e-01 1.36483297e-01 -5.22185087e-01 5.30669391e-01
-6.94692910e-01 2.10506916e-01 8.83229554e-01 9.20922875e-01
-6.28267705e-01 -9.79824722e-01 -4.42870587e-01 3.43463123e-01
-7.69034386e-01 -3.75524461e-02 -1.59304932e-01 1.18555760e+00
-4.72116530e-01 4.42497998e-01 6.62126780e-01 -3.66849720e-01
2.28736624e-01 4.67346609e-01 2.83004373e-01 -3.97258885e-02
-6.36663437e-01 6.11211538e-01 -3.51188481e-01 -2.09247798e-01
-7.71288097e-01 -9.66243565e-01 -1.22174120e+00 -7.96836317e-01
-2.53562182e-01 5.31933606e-01 2.72770107e-01 8.22442889e-01
-1.69307604e-01 3.58577110e-02 1.17741072e+00 -8.52954149e-01
-1.22194052e+00 -7.10026324e-01 -1.49012637e+00 5.70286512e-02
6.69592142e-01 -6.34954095e-01 -9.54536319e-01 -2.10950673e-01]
|
[4.793015480041504, 4.106166839599609]
|
b99e89a8-6f62-4cd4-a04e-45af486149e4
|
face-recognition-by-fusion-of-local-and
|
1002.00382
| null |
http://arxiv.org/abs/1002.0382v1
|
http://arxiv.org/pdf/1002.0382v1.pdf
|
Face Recognition by Fusion of Local and Global Matching Scores using DS Theory: An Evaluation with Uni-classifier and Multi-classifier Paradigm
|
Faces are highly deformable objects which may easily change their appearance
over time. Not all face areas are subject to the same variability. Therefore
decoupling the information from independent areas of the face is of paramount
importance to improve the robustness of any face recognition technique. This
paper presents a robust face recognition technique based on the extraction and
matching of SIFT features related to independent face areas. Both a global and
local (as recognition from parts) matching strategy is proposed. The local
strategy is based on matching individual salient facial SIFT features as
connected to facial landmarks such as the eyes and the mouth. As for the global
matching strategy, all SIFT features are combined together to form a single
feature. In order to reduce the identification errors, the Dempster-Shafer
decision theory is applied to fuse the two matching techniques. The proposed
algorithms are evaluated with the ORL and the IITK face databases. The
experimental results demonstrate the effectiveness and potential of the
proposed face recognition techniques also in the case of partially occluded
faces or with missing information.
|
['Jamuna Kanta Sing', 'Phalguni Gupta', 'Massimo Tistarelli', 'Dakshina Ranjan Kisku']
|
2010-02-02
| null | null | null | null |
['robust-face-recognition']
|
['computer-vision']
|
[ 1.53425530e-01 -1.56688914e-01 -7.74287106e-03 -6.02236271e-01
-2.28809580e-01 -3.91500294e-01 6.18412852e-01 -1.72521710e-01
-2.13302732e-01 4.19798434e-01 1.53353736e-02 4.84816819e-01
-3.78924042e-01 -6.31084681e-01 -2.70185083e-01 -1.06369615e+00
7.82886147e-02 9.57146287e-02 1.77273616e-01 -4.56456952e-02
4.04972076e-01 1.22787642e+00 -2.04870677e+00 1.89938709e-01
3.38386625e-01 1.12765837e+00 -5.78945018e-02 1.94776371e-01
-9.50131342e-02 2.60785133e-01 -2.81576574e-01 -2.36964390e-01
6.72269940e-01 -5.21958396e-02 -4.75772738e-01 4.27242428e-01
6.74545527e-01 -4.35612589e-01 -1.03315048e-01 1.17898726e+00
4.18041497e-01 1.13838524e-01 5.53967774e-01 -1.28437209e+00
-2.70951241e-01 -2.66068727e-01 -7.63198376e-01 6.11984953e-02
6.48072541e-01 -2.19456449e-01 4.34577525e-01 -1.25989544e+00
6.62242115e-01 1.48923385e+00 4.49585885e-01 1.87623337e-01
-7.85496950e-01 -6.57006502e-01 -3.71317342e-02 4.72418547e-01
-1.70723140e+00 -9.84853268e-01 9.89574373e-01 -3.03722560e-01
5.12759626e-01 1.30515322e-01 3.49324405e-01 1.43298507e-01
2.66762346e-01 3.28849196e-01 1.05043221e+00 -6.94421053e-01
-5.62438220e-02 2.85933435e-01 1.43190927e-03 9.31545913e-01
5.58051527e-01 1.01002112e-01 -4.98585939e-01 -1.86227143e-01
5.66927731e-01 1.58797801e-01 6.45031705e-02 -5.69015324e-01
-6.01487517e-01 5.97340822e-01 2.75490642e-01 4.81963426e-01
-5.86566567e-01 -2.15561181e-01 1.41950294e-01 2.72619963e-01
2.27870822e-01 -5.06733716e-01 -2.01019436e-01 4.11054879e-01
-1.08247328e+00 -6.30496442e-02 5.19164979e-01 7.60115623e-01
1.01944780e+00 7.02324659e-02 1.40607029e-01 5.43701768e-01
8.09433937e-01 6.19204342e-01 2.95624882e-01 -4.62513536e-01
1.13633819e-01 9.23474848e-01 1.30471021e-01 -1.50300145e+00
-1.78046465e-01 7.44203925e-02 -4.59261686e-01 5.00499427e-01
2.68791020e-01 8.03517476e-02 -9.70278859e-01 1.31870377e+00
8.89917016e-01 4.93722446e-02 6.67142719e-02 5.47061145e-01
1.02417099e+00 3.94659609e-01 -2.09253188e-02 -2.63123870e-01
1.47884202e+00 -3.98769677e-01 -9.70226347e-01 5.62032349e-02
6.21664673e-02 -1.23636460e+00 6.71576858e-02 8.99148658e-02
-8.10074449e-01 -7.71959066e-01 -1.02988338e+00 1.87876090e-01
-6.09920263e-01 5.33820570e-01 2.08592057e-01 1.00779533e+00
-1.10517311e+00 2.89767563e-01 -7.34804630e-01 -4.68732059e-01
4.80646461e-01 8.59155476e-01 -8.97058606e-01 -1.76573992e-01
-7.11806297e-01 9.01655436e-01 1.69310346e-01 4.75549072e-01
-3.51141840e-01 -1.57133177e-01 -7.91667998e-01 -2.70261951e-02
-9.80004296e-02 -4.94691841e-02 5.74499607e-01 -1.17513382e+00
-1.40820098e+00 8.55638266e-01 -5.27571499e-01 2.32875906e-02
3.08565408e-01 1.68509915e-01 -5.78670204e-01 5.47997415e-01
-4.29977141e-02 3.27223599e-01 1.32767379e+00 -8.60527217e-01
-5.11377752e-01 -8.75937760e-01 -5.21519840e-01 1.86683953e-01
-2.26478308e-01 3.72636765e-01 -1.72737494e-01 -4.31438595e-01
3.36218178e-01 -8.03243458e-01 4.30816889e-01 3.12085837e-01
9.28788483e-02 -2.63013482e-01 1.57620335e+00 -8.54116857e-01
7.59536922e-01 -2.34004283e+00 -1.17376402e-01 5.87898314e-01
-3.07518989e-01 4.54922706e-01 -9.61140022e-02 3.76509190e-01
-2.08841830e-01 -3.45261961e-01 2.50674188e-01 -1.47388235e-01
-3.38157564e-01 -4.21660915e-02 2.80450851e-01 1.01978910e+00
3.31241637e-01 2.77913570e-01 -2.34559223e-01 -6.11883640e-01
5.10637343e-01 7.46977866e-01 -1.26526743e-01 -1.35860533e-01
5.03658652e-01 1.50169760e-01 -5.72938442e-01 1.02840710e+00
1.22636223e+00 4.19390887e-01 7.02238008e-02 -5.46433330e-01
-1.20841064e-01 -2.82351315e-01 -1.73716545e+00 1.12428021e+00
1.82336960e-02 3.17033768e-01 5.13426185e-01 -1.00845408e+00
1.17762291e+00 7.16913998e-01 7.14220405e-01 -4.44583118e-01
3.20454866e-01 1.84044465e-01 -1.03639163e-01 -4.20481592e-01
1.33223802e-01 -3.13082822e-02 3.60790938e-01 1.74091265e-01
3.06975931e-01 2.36068293e-01 2.71171448e-03 -2.51076460e-01
4.77187514e-01 4.96963337e-02 7.45473087e-01 -5.73412299e-01
1.09930241e+00 -3.96763235e-01 5.12072742e-01 -4.77948040e-02
-3.72996867e-01 1.73211709e-01 -2.04573020e-01 -5.32918692e-01
-4.27087814e-01 -8.01301897e-01 -4.17319566e-01 4.49096054e-01
2.43786231e-01 6.96561038e-02 -6.71398103e-01 -6.15710437e-01
1.64336592e-01 -1.02259256e-01 -6.36298358e-01 -8.36553946e-02
-2.94634908e-01 -3.98648977e-01 1.70443475e-01 1.39771223e-01
7.58625984e-01 -8.68640661e-01 -6.65482938e-01 1.25745729e-01
3.51118207e-01 -8.96243095e-01 -2.78768659e-01 -3.83309543e-01
-8.26510608e-01 -1.26864660e+00 -5.10853887e-01 -9.40845430e-01
1.03698087e+00 6.41596496e-01 1.80425808e-01 3.36522222e-01
-6.46764278e-01 6.13785446e-01 -2.66983539e-01 -3.03415895e-01
-2.14282051e-01 -4.48101133e-01 2.94771194e-01 6.63831651e-01
6.65189326e-01 -2.04986274e-01 -5.51541269e-01 5.88586330e-01
-7.08646119e-01 -6.07711971e-01 4.56061989e-01 6.14954531e-01
4.51315552e-01 4.37987298e-01 4.95380491e-01 -2.87907034e-01
2.28439167e-01 -9.70698223e-02 -7.86679208e-01 5.08695126e-01
-3.39940846e-01 -1.34547517e-01 2.10302472e-01 -2.55578250e-01
-1.21961260e+00 7.44445264e-01 2.18583018e-01 -2.65046954e-01
-4.09203321e-01 5.27064912e-02 -6.59208715e-01 -9.35445309e-01
1.63741052e-01 2.08854809e-01 4.02276427e-01 -4.09943968e-01
9.84304622e-02 7.68854499e-01 2.01355413e-01 -2.21851155e-01
1.08230495e+00 7.52137423e-01 3.78867507e-01 -1.31453991e+00
2.07605101e-02 -7.17534363e-01 -9.89098549e-01 -4.62409884e-01
6.74577475e-01 -8.62906635e-01 -6.08357668e-01 6.97778583e-01
-9.05127823e-01 7.42848575e-01 5.41922078e-02 5.83850384e-01
-2.63002634e-01 6.16484225e-01 -1.90988004e-01 -8.70426536e-01
-4.93339300e-01 -1.30331194e+00 9.74828541e-01 6.58436358e-01
1.83450282e-01 -8.22635591e-01 -1.56395003e-01 1.24871947e-01
3.79561931e-01 1.62350371e-01 4.47100103e-01 -6.03018701e-01
-5.61012208e-01 -8.33537996e-01 -5.73064424e-02 2.63109177e-01
9.33370411e-01 5.00683725e-01 -1.04825914e+00 -4.11045939e-01
2.30984554e-01 1.76984787e-01 4.85607237e-01 3.33090097e-01
3.67510706e-01 -2.92778432e-01 -3.87962878e-01 2.57076383e-01
1.53548026e+00 6.55747235e-01 7.55317867e-01 6.28827438e-02
3.29851031e-01 8.06586444e-01 7.87336171e-01 4.11454767e-01
-1.04010485e-01 7.20445454e-01 3.33418310e-01 -8.61645937e-02
-1.51810110e-01 1.82542577e-01 4.25618201e-01 4.28153604e-01
-1.55032799e-01 1.59239531e-01 -4.23377603e-01 4.99678731e-01
-1.40680385e+00 -1.33648837e+00 7.98160359e-02 2.46475077e+00
2.06113160e-01 -5.09141445e-01 -9.09578204e-02 4.42821205e-01
1.05097592e+00 -9.22042131e-02 -1.56385139e-01 -3.24456006e-01
1.40637252e-02 3.44083190e-01 2.96140283e-01 5.17936409e-01
-1.22520316e+00 7.12625623e-01 5.67701149e+00 6.86443686e-01
-1.25328791e+00 -1.41716018e-01 2.37507284e-01 2.74571866e-01
2.58738488e-01 3.95335034e-02 -1.10258257e+00 3.04605871e-01
3.88936758e-01 -1.24863923e-01 1.29017234e-01 6.96938634e-01
1.16885595e-01 -3.15492034e-01 -5.75700521e-01 1.03734887e+00
2.51564115e-01 -7.31549382e-01 1.15220442e-01 2.00683311e-01
5.20011425e-01 -5.18573523e-01 1.91045329e-01 -3.97321582e-01
-3.22307199e-01 -7.35342383e-01 5.26872218e-01 6.12192512e-01
5.42149484e-01 -7.46365428e-01 6.67269170e-01 -5.74184433e-02
-1.69409633e+00 1.08656585e-01 -3.78327012e-01 1.99445128e-01
-9.65824872e-02 2.03047320e-01 -7.63321400e-01 6.07257605e-01
5.15162408e-01 5.56904137e-01 -6.08805120e-01 1.11519575e+00
7.01936334e-02 -7.65628070e-02 -6.23433888e-01 2.12315977e-01
-1.39873579e-01 -3.59077841e-01 5.23426712e-01 8.34348917e-01
5.78467846e-01 1.87890038e-01 6.30476652e-03 4.32498664e-01
9.37055424e-02 5.68389297e-01 -8.34318399e-01 4.74245884e-02
5.07373571e-01 1.44316399e+00 -8.35168183e-01 -1.62900209e-01
-7.52124727e-01 8.92301500e-01 -3.11983079e-01 1.18910901e-01
-3.08955282e-01 -4.40690845e-01 6.02885246e-01 1.59226701e-01
5.71089745e-01 -1.57436565e-01 2.63866693e-01 -8.12717140e-01
1.38840795e-01 -6.55799270e-01 3.50978017e-01 -3.00000817e-01
-9.38731372e-01 5.96045434e-01 1.21752419e-01 -1.38729930e+00
-1.20269567e-01 -6.58273757e-01 -5.80604792e-01 9.37150121e-01
-1.43693137e+00 -1.41311991e+00 -3.55906427e-01 1.07151735e+00
3.73078108e-01 -4.76830423e-01 7.29840219e-01 3.19252014e-01
-4.50286448e-01 5.96950173e-01 7.45756030e-02 4.98109348e-02
7.02189744e-01 -4.34545338e-01 -1.78235561e-01 9.10910964e-01
-3.52702625e-02 7.22042739e-01 4.66490924e-01 -7.91472793e-01
-1.51156449e+00 -6.74581468e-01 8.35052669e-01 1.49239421e-01
6.14484586e-02 2.17256285e-02 -7.40325153e-01 4.04743969e-01
3.44891213e-02 3.12215090e-01 5.57973981e-01 -4.02565837e-01
-8.26023072e-02 -5.50458670e-01 -1.81136274e+00 8.14833418e-02
4.48698431e-01 -5.04084349e-01 -5.16390800e-01 1.06599621e-01
-3.49964708e-01 5.19616380e-02 -9.42659855e-01 4.45702940e-01
7.96172917e-01 -9.60840106e-01 9.67279971e-01 -1.76272303e-01
-5.19838512e-01 -6.35652483e-01 -3.45041692e-01 -6.88416064e-01
-2.66190559e-01 -2.23348573e-01 3.60490978e-01 1.50257540e+00
-1.22347414e-01 -7.12171197e-01 7.41634905e-01 6.28031194e-01
4.57631230e-01 -4.23297584e-02 -1.34649992e+00 -7.03785598e-01
-6.79311574e-01 4.28074330e-01 5.69160581e-01 6.22511744e-01
-1.95288286e-01 -1.45037740e-01 -1.39354825e-01 3.29129487e-01
7.90981650e-01 1.30798727e-01 6.36307597e-01 -1.58747530e+00
3.79564613e-01 -8.30265284e-02 -1.08231688e+00 -1.94070280e-01
1.53691188e-01 -5.31546652e-01 -1.36583254e-01 -1.12924862e+00
6.18609674e-02 -1.80865929e-01 -2.98398137e-01 4.79303956e-01
2.53361225e-01 2.46771902e-01 4.87661324e-02 1.13695964e-01
4.64999638e-02 4.42964226e-01 9.56538498e-01 -5.82996197e-02
-1.50481597e-01 1.38061777e-01 -2.24659711e-01 7.02546835e-01
7.13391244e-01 -3.29560995e-01 -3.72479916e-01 -9.53301787e-02
-5.75580120e-01 -4.04905863e-02 1.77462339e-01 -1.14744174e+00
3.64237159e-01 -6.50219694e-02 7.18555689e-01 -4.84739602e-01
6.05284214e-01 -1.37632585e+00 4.53919619e-01 5.75265288e-01
3.68601322e-01 1.78784207e-01 3.19732964e-01 4.53377604e-01
-4.27578539e-01 -2.88358659e-01 1.17909634e+00 -6.59739748e-02
-7.91074097e-01 3.01966399e-01 -2.71607131e-01 -9.05771375e-01
1.50321233e+00 -7.41406918e-01 -1.12695508e-02 -2.10331917e-01
-7.59015262e-01 -3.99238288e-01 4.60477442e-01 4.37820673e-01
7.74552226e-01 -1.30736387e+00 -5.29829979e-01 8.05853307e-01
-7.22060278e-02 -5.91732204e-01 3.58131647e-01 7.70316839e-01
-4.95708317e-01 4.74765539e-01 -8.27453315e-01 -3.14221472e-01
-2.13121009e+00 6.30730271e-01 3.36610705e-01 3.64210308e-01
-6.54121861e-02 5.16928673e-01 -7.74638057e-02 1.02864191e-01
1.47473902e-01 4.18994837e-02 -5.00658512e-01 3.58853817e-01
6.50478721e-01 3.63958240e-01 3.77049029e-01 -1.47475028e+00
-8.31276417e-01 1.16474485e+00 -2.32617378e-01 1.30318999e-01
1.15419257e+00 -2.62730092e-01 -3.67787153e-01 -2.33707130e-01
1.31054974e+00 2.74304718e-01 -8.24239850e-01 -2.77842194e-01
-6.15260713e-02 -8.55441809e-01 6.99689612e-02 -3.07274282e-01
-1.30257225e+00 7.67222524e-01 1.14005852e+00 -1.03892431e-01
1.33942008e+00 -3.69657785e-01 7.63972327e-02 1.24866053e-01
5.34681559e-01 -1.02865362e+00 -3.14393759e-01 -4.25156988e-02
8.83217752e-01 -1.21386147e+00 3.56375962e-01 -6.49433970e-01
-1.90198332e-01 1.45153272e+00 2.62106657e-01 -1.57835141e-01
1.02066922e+00 9.57267955e-02 -4.88730781e-02 -2.05180362e-01
-2.66276091e-01 -2.47671142e-01 5.85440636e-01 5.53959489e-01
2.65508920e-01 -1.93243101e-01 -5.47804177e-01 1.65694952e-02
3.93043637e-01 1.27860263e-01 1.63601592e-01 1.18452322e+00
-6.09335303e-01 -1.19843614e+00 -7.73282051e-01 3.07142079e-01
-5.58596313e-01 4.80562985e-01 -3.61125380e-01 8.89576316e-01
2.33118623e-01 1.10655820e+00 -1.67057708e-01 -2.39490137e-01
1.58101544e-01 1.96178421e-01 6.98358655e-01 -2.56651998e-01
-3.33190829e-01 2.13554442e-01 -3.05498481e-01 -5.10683239e-01
-8.94848228e-01 -9.55004513e-01 -1.04356682e+00 -2.08301768e-01
-4.88185644e-01 6.29586428e-02 1.01877785e+00 7.79280841e-01
2.96307176e-01 -2.36394614e-01 8.40931177e-01 -9.10227180e-01
-4.11814064e-01 -6.82383835e-01 -7.89854169e-01 3.30711514e-01
3.63442898e-01 -1.00925541e+00 -3.98717463e-01 2.24635929e-01]
|
[13.110581398010254, 0.6071850657463074]
|
40fd0181-bd71-413f-b8f5-6f46c4477763
|
opera-attention-regularized-transformers-for
|
2103.03873
| null |
https://arxiv.org/abs/2103.03873v1
|
https://arxiv.org/pdf/2103.03873v1.pdf
|
OperA: Attention-Regularized Transformers for Surgical Phase Recognition
|
In this paper we introduce OperA, a transformer-based model that accurately predicts surgical phases from long video sequences. A novel attention regularization loss encourages the model to focus on high-quality frames during training. Moreover, the attention weights are utilized to identify characteristic high attention frames for each surgical phase, which could further be used for surgery summarization. OperA is thoroughly evaluated on two datasets of laparoscopic cholecystectomy videos, outperforming various state-of-the-art temporal refinement approaches.
|
['Nassir Navab', 'Benjamin Busam', 'Seong Tae Kim', 'Daniel Ostler', 'Magdalini Paschali', 'Tobias Czempiel']
|
2021-03-05
| null | null | null | null |
['surgical-phase-recognition']
|
['computer-vision']
|
[ 1.80946797e-01 4.24809158e-01 -6.30382359e-01 -7.76316524e-02
-6.95261478e-01 4.63016145e-02 4.33646232e-01 2.02583313e-01
-3.64238560e-01 4.00457710e-01 7.14377999e-01 -1.07305288e-01
-2.63873160e-01 -2.24356875e-01 -6.11059010e-01 -7.98618734e-01
-3.63582134e-01 2.22749442e-01 7.80251250e-02 -5.28012179e-02
1.06343046e-01 2.19235674e-01 -1.12174344e+00 5.72353065e-01
7.05301583e-01 9.48217511e-01 4.90319759e-01 5.42751014e-01
4.74379659e-02 1.20550704e+00 -1.95276380e-01 -2.58269638e-01
5.07409163e-02 -5.52022517e-01 -7.58903503e-01 -3.19697559e-02
4.32068497e-01 -1.79098979e-01 -8.47598612e-01 1.02498937e+00
3.39788079e-01 2.73619920e-01 3.16971749e-01 -4.98507261e-01
-4.59684014e-01 7.99696028e-01 -3.13712925e-01 8.06364179e-01
2.56978899e-01 1.98346063e-01 7.98718333e-01 -8.42227519e-01
1.04153085e+00 8.46414268e-01 5.74488759e-01 7.14888275e-01
-9.38562870e-01 -5.72537661e-01 3.84143710e-01 3.73749018e-01
-8.85704875e-01 -2.90008724e-01 8.03053439e-01 -2.94745594e-01
8.78883541e-01 1.33158699e-01 1.19337118e+00 1.29787803e+00
7.94309616e-01 1.21144176e+00 5.00307977e-01 -2.49740481e-02
-3.45011503e-01 -5.74206591e-01 -1.61801595e-02 1.10246837e+00
7.80247003e-02 2.78094411e-01 -7.58295834e-01 3.68448108e-01
9.04743671e-01 3.56641203e-01 -7.65745044e-01 -3.04721057e-01
-1.73566771e+00 6.86499298e-01 7.64788926e-01 4.39379245e-01
-9.06852186e-01 2.34630823e-01 6.23881817e-01 -1.16089121e-01
3.48440856e-01 8.25364709e-01 -1.56934232e-01 -2.67397046e-01
-8.50401342e-01 -1.17297918e-01 2.85762608e-01 1.02739966e+00
2.86827177e-01 8.93632248e-02 -7.35237181e-01 5.32359421e-01
2.44186670e-02 -1.56402171e-01 8.09553802e-01 -1.06903803e+00
1.74821407e-01 4.33973014e-01 -2.63734519e-01 -6.40223086e-01
-5.27736366e-01 -6.65678561e-01 -1.04790699e+00 -3.26496214e-01
-1.89689457e-01 1.85066342e-01 -1.26787448e+00 1.25832748e+00
-2.44439375e-02 7.20622957e-01 6.56876191e-02 1.00353968e+00
1.15442431e+00 5.28174341e-01 2.26749569e-01 -5.99819243e-01
1.19607735e+00 -1.48784828e+00 -1.14262199e+00 -2.19618097e-01
4.68974590e-01 -4.94189739e-01 6.87500358e-01 5.33126891e-01
-1.51928067e+00 -5.77094853e-01 -8.51965368e-01 -2.86923587e-01
2.70076513e-01 7.41394013e-02 8.55659723e-01 -1.82040423e-01
-8.25149536e-01 7.15649784e-01 -1.43661058e+00 -1.08501159e-01
6.82594001e-01 4.35736597e-01 -3.02041709e-01 -2.38599762e-01
-9.15497899e-01 9.32238340e-01 4.83302802e-01 4.35983688e-01
-1.27009439e+00 -1.11413431e+00 -1.29181004e+00 2.11405337e-01
2.38253310e-01 -9.38069820e-01 1.46891189e+00 -9.15595710e-01
-1.48604572e+00 9.13612962e-01 -3.04724574e-01 -8.97389710e-01
3.08563083e-01 -4.97692645e-01 -1.86401650e-01 4.87716287e-01
-1.62446648e-01 7.53292918e-01 7.96588778e-01 -1.02251947e+00
-6.09976590e-01 -5.01830466e-02 1.39647856e-01 2.17526674e-01
-2.76613593e-01 -2.39424691e-01 -9.40647602e-01 -1.14128470e+00
1.01098880e-01 -1.00552249e+00 -7.19191909e-01 -2.73457710e-02
-3.16922337e-01 -4.01187763e-02 5.52620530e-01 -6.86871827e-01
1.63200605e+00 -2.18332171e+00 8.88050258e-01 -1.98561549e-01
4.18005347e-01 2.97316611e-01 -6.01572320e-02 -8.16638246e-02
-1.54119357e-01 -5.97404428e-02 8.09221491e-02 -4.92768943e-01
-5.76738060e-01 4.86221790e-01 -9.91430208e-02 5.86265266e-01
1.38614863e-01 9.81160343e-01 -1.43083572e+00 -6.67618573e-01
6.17382944e-01 4.24838752e-01 -8.42149854e-01 3.98598224e-01
6.70123994e-02 8.15757751e-01 -5.12243807e-01 6.18926823e-01
7.38022178e-02 -4.85234350e-01 2.97092140e-01 -5.90816796e-01
-1.84855924e-03 3.11444223e-01 -1.00536041e-01 2.52961922e+00
-6.16816163e-01 7.87346900e-01 -7.20177069e-02 -1.04452980e+00
3.80361706e-01 4.48704720e-01 1.05643630e+00 -6.76815450e-01
3.81962478e-01 -1.09589607e-01 1.42767644e-02 -7.93869019e-01
5.46238899e-01 4.01919819e-02 6.30901232e-02 -2.29453534e-01
4.59638506e-01 4.79466617e-02 4.44170982e-01 1.92817375e-01
9.32705879e-01 2.10181803e-01 4.46416676e-01 -3.08883160e-01
5.98907590e-01 7.36801103e-02 9.50696051e-01 6.74375594e-01
-3.49034637e-01 7.26238370e-01 3.71931970e-01 -8.12572122e-01
-7.05632210e-01 -8.17196250e-01 -3.45421620e-02 6.98640168e-01
5.07127762e-01 -6.57613873e-01 -2.63030052e-01 -8.78519833e-01
-3.24336112e-01 4.68246102e-01 -1.12929499e+00 -4.45311874e-01
-9.62327957e-01 -3.52035880e-01 -1.74023777e-01 6.41790509e-01
-6.98987022e-02 -1.25106299e+00 -8.25465143e-01 5.19665062e-01
-3.90979081e-01 -1.00533366e+00 -7.30122089e-01 1.89924613e-01
-1.17570472e+00 -1.30716956e+00 -1.01513839e+00 -1.10174942e+00
7.58579075e-01 3.64164174e-01 1.28479600e+00 3.18681002e-01
-4.96239871e-01 2.40776896e-01 -5.55287242e-01 -3.85917425e-01
-2.58590609e-01 1.78630635e-01 -2.75268763e-01 -2.33725652e-01
1.50672738e-02 -3.11331153e-01 -9.19108331e-01 -4.38730083e-02
-5.38031399e-01 6.19831920e-01 7.21769214e-01 1.22346067e+00
8.85278404e-01 -4.88719523e-01 2.84556840e-02 -1.03817451e+00
1.50463581e-01 -5.60239494e-01 -4.53116357e-01 2.05916658e-01
-3.69347394e-01 2.56911308e-01 4.51292187e-01 -5.59818208e-01
-9.41274643e-01 2.36641281e-04 -2.22520400e-02 -1.15445518e+00
4.25739795e-01 7.46272862e-01 7.27610171e-01 -3.27529795e-02
1.86634466e-01 3.22039604e-01 -1.01664234e-02 -2.18341097e-01
-1.00732848e-01 -2.17145517e-01 8.27379167e-01 -9.65333059e-02
3.40456516e-01 3.73865068e-01 -3.39767896e-02 -5.64218640e-01
-1.27796662e+00 -6.15900755e-01 -3.33510339e-01 -4.22511756e-01
9.42830145e-01 -8.48718822e-01 -7.32821584e-01 1.51363596e-01
-9.57450092e-01 -4.03233081e-01 -4.43871975e-01 9.61028755e-01
-7.18098879e-01 1.46018565e-01 -9.69790041e-01 -3.06432515e-01
-6.01762533e-01 -1.43633330e+00 1.09118426e+00 5.09549618e-01
-1.86316445e-01 -1.02251840e+00 1.34949729e-01 1.35076135e-01
1.98827833e-01 2.55075395e-01 6.11046314e-01 -3.64980280e-01
-8.42382789e-01 1.08571544e-01 7.11147487e-02 -7.17748627e-02
8.25647935e-02 7.53726857e-03 -4.42118019e-01 -4.96791035e-01
-1.81620508e-01 -1.50653526e-01 1.22389698e+00 1.03967953e+00
1.81989110e+00 -1.42640010e-01 -5.97076118e-01 1.41314328e+00
1.05780566e+00 2.08210766e-01 6.97998643e-01 4.06200677e-01
6.21676743e-01 1.77123055e-01 8.47772598e-01 3.47141296e-01
1.51122913e-01 5.12798250e-01 7.31960177e-01 -2.77454287e-01
-2.87695497e-01 -4.44974229e-02 7.24131167e-02 1.00481296e+00
-3.23112428e-01 -1.95202470e-01 -8.69328201e-01 7.80288517e-01
-1.80833387e+00 -1.10028672e+00 1.44096136e-01 1.87968218e+00
6.53359175e-01 7.24104568e-02 -4.23207790e-01 -3.97605598e-01
3.99753034e-01 5.64065039e-01 -3.82660091e-01 -3.52913551e-02
1.65803924e-01 4.46329594e-01 5.20246685e-01 2.79921144e-01
-1.37145102e+00 7.12337017e-01 6.93216324e+00 5.42471826e-01
-1.21852255e+00 1.02287292e-01 4.02642131e-01 -5.35685062e-01
-1.75240949e-01 -3.06775481e-01 -3.01134974e-01 3.49896580e-01
7.08834410e-01 -5.72096169e-01 1.19880907e-01 6.17522717e-01
2.16698661e-01 1.87188648e-02 -1.00330424e+00 9.63316619e-01
1.98794737e-01 -1.96220696e+00 2.05238268e-01 -1.97663784e-01
9.63992059e-01 1.66765332e-01 6.58316016e-02 5.28643548e-01
1.20926850e-01 -9.55841541e-01 3.67970854e-01 7.92197108e-01
7.16074705e-01 -6.15099370e-01 9.32935178e-01 -1.09077550e-01
-1.30401731e+00 -2.61202961e-01 -7.91762844e-02 5.79866827e-01
2.82203704e-01 7.78476149e-03 -5.46729982e-01 8.94689262e-01
8.31261277e-01 1.56819451e+00 -3.76684010e-01 1.44726896e+00
-1.87261298e-01 3.67311329e-01 1.47131994e-01 4.47033763e-01
7.57371366e-01 1.73337162e-02 6.59944475e-01 1.30869281e+00
5.73501348e-01 3.83106112e-01 4.06197488e-01 2.49431536e-01
-2.24776357e-01 -9.27989110e-02 -4.20334399e-01 -1.35680273e-01
-1.72342751e-02 1.29812264e+00 -4.32583958e-01 -4.66684192e-01
-4.66726512e-01 9.65117157e-01 2.50878394e-01 2.31358975e-01
-8.70975554e-01 -3.85745727e-02 5.70653498e-01 2.39371378e-02
4.28649127e-01 9.97583792e-02 2.22237498e-01 -1.34980142e+00
-4.10839498e-01 -5.92263043e-01 6.60398841e-01 -8.98187041e-01
-7.16255724e-01 7.41824925e-01 -1.94663301e-01 -1.71909773e+00
-1.90622881e-01 -3.26168031e-01 -7.21377671e-01 4.64495003e-01
-1.81601691e+00 -1.12935197e+00 -7.55590796e-01 5.34965277e-01
1.20336151e+00 -4.67229411e-02 7.04986751e-01 2.49984741e-01
-6.83140635e-01 3.05893838e-01 -9.07789990e-02 5.04621342e-02
7.28764832e-01 -1.12253845e+00 -2.29226369e-02 8.18157792e-01
-1.32669106e-01 6.55540168e-01 6.89700067e-01 -4.81315374e-01
-1.38553607e+00 -1.31801641e+00 6.72923028e-01 2.88075842e-02
5.61405361e-01 2.94921070e-01 -9.58589733e-01 9.90478337e-01
4.65953290e-01 4.04918492e-01 5.29263496e-01 -2.12375559e-02
2.59235024e-01 2.29703691e-02 -5.49008906e-01 4.70815182e-01
1.05440533e+00 -1.37854069e-01 -7.97143936e-01 4.49251503e-01
6.83982134e-01 -1.12353778e+00 -1.13419414e+00 6.18793190e-01
6.21952355e-01 -9.26193416e-01 1.11493528e+00 -5.85984230e-01
8.77921283e-01 2.19307989e-01 5.97869277e-01 -1.29359853e+00
-6.50693417e-01 -9.69980001e-01 -6.29370987e-01 3.25766355e-01
2.77043711e-02 6.21220134e-02 8.64924729e-01 -2.24555098e-02
-8.69138062e-01 -9.94688094e-01 -7.82973707e-01 -2.78543323e-01
-2.90125132e-01 4.04320061e-02 -4.39225510e-02 8.23619366e-01
1.56022638e-01 -1.53985068e-01 -5.24090409e-01 -1.97389517e-02
5.06201684e-01 1.18811056e-01 3.10217232e-01 -9.95912969e-01
6.85730651e-02 -6.51241601e-01 -4.31176782e-01 -1.00715947e+00
2.13821918e-01 -9.22912419e-01 1.13896392e-01 -1.66616082e+00
3.84292722e-01 -4.77065854e-02 -7.55192220e-01 3.72325003e-01
-3.46318871e-01 2.90522903e-01 1.99692890e-01 2.99803466e-01
-8.48851204e-01 7.98115611e-01 1.66579711e+00 -3.60323191e-01
-2.53100455e-01 1.95456650e-02 -3.37311625e-01 7.91022420e-01
3.66940081e-01 -4.25715715e-01 -3.21471900e-01 -5.62691689e-01
-1.90669268e-01 5.37261367e-01 1.04466081e-01 -1.12543261e+00
4.72724527e-01 -8.10958669e-02 3.03201228e-01 -8.04941475e-01
2.22915605e-01 -7.27033854e-01 1.31349772e-01 9.49225068e-01
-6.36192858e-01 4.92394380e-02 2.85951555e-01 6.40015006e-01
-6.12273991e-01 9.27576646e-02 8.33906412e-01 -2.21735612e-01
-9.22892511e-01 8.75061810e-01 -3.29143494e-01 7.94241950e-02
1.11991107e+00 -4.83414941e-02 -3.30518670e-02 -7.50850439e-02
-1.05522645e+00 4.88448411e-01 3.06998670e-01 5.94643891e-01
8.68255913e-01 -1.28639221e+00 -7.50450075e-01 3.06727260e-01
2.70603061e-01 3.01505357e-01 7.65815258e-01 1.42648280e+00
-7.68406034e-01 6.04329526e-01 -3.98618340e-01 -7.61232615e-01
-1.27693689e+00 8.23451042e-01 4.08725381e-01 -6.75861955e-01
-1.37831128e+00 9.62660909e-01 5.69234192e-01 4.11745876e-01
4.57380861e-01 -6.87432289e-01 -5.63088715e-01 -1.37696236e-01
6.52227461e-01 -2.19525889e-01 -1.76353067e-01 -3.97149980e-01
-3.06471169e-01 5.00640094e-01 -3.95196021e-01 5.06666422e-01
1.47705293e+00 -9.93945003e-02 2.51227945e-01 1.85988873e-01
1.02991641e+00 -4.21947598e-01 -1.47039914e+00 -3.58326912e-01
-2.23122567e-01 -4.95330423e-01 5.34182012e-01 -5.12381732e-01
-1.55010831e+00 6.29168749e-01 4.09412414e-01 -3.44954967e-01
1.36136270e+00 -5.73305748e-02 9.14768338e-01 1.54261678e-01
2.42536142e-01 -7.21108735e-01 6.09025918e-02 4.39897239e-01
1.03609872e+00 -1.29564762e+00 1.55912906e-01 -1.46733612e-01
-8.33554029e-01 1.32132149e+00 6.11210823e-01 -2.71800935e-01
3.91958207e-01 2.95537025e-01 -2.70766646e-01 -3.96833241e-01
-9.13650334e-01 -3.49175814e-03 6.52781785e-01 2.98241049e-01
6.19942665e-01 -3.11221182e-01 -3.00798774e-01 4.63702887e-01
2.24820837e-01 3.19459110e-01 5.13332248e-01 8.46906006e-01
-1.48912966e-01 -5.79151571e-01 1.85639560e-01 5.27749598e-01
-7.28598475e-01 -2.37156063e-01 4.26254570e-01 8.55430245e-01
-1.11581199e-01 2.91738212e-01 1.90501466e-01 -6.49099573e-02
3.48874569e-01 -2.85636544e-01 7.08847642e-01 -6.47325456e-01
-1.04107368e+00 3.64123255e-01 -3.57621610e-02 -1.10284805e+00
-7.38798022e-01 -6.71222329e-01 -1.20987129e+00 -7.70641789e-02
-4.82226685e-02 1.50981560e-01 9.49748829e-02 5.81068337e-01
3.19565475e-01 1.39578712e+00 3.02046001e-01 -1.29639983e+00
3.04747019e-02 -7.55285859e-01 -2.24559397e-01 4.82816190e-01
7.20591307e-01 -8.05095494e-01 -4.05089445e-02 2.29646355e-01]
|
[14.136795043945312, -3.32171368598938]
|
15ca92e1-40a5-4ab3-9d16-99e7e1281a03
|
beyond-greedy-search-tracking-by-multi-agent
|
2205.09676
| null |
https://arxiv.org/abs/2205.09676v3
|
https://arxiv.org/pdf/2205.09676v3.pdf
|
Beyond Greedy Search: Tracking by Multi-Agent Reinforcement Learning-based Beam Search
|
To track the target in a video, current visual trackers usually adopt greedy search for target object localization in each frame, that is, the candidate region with the maximum response score will be selected as the tracking result of each frame. However, we found that this may be not an optimal choice, especially when encountering challenging tracking scenarios such as heavy occlusion and fast motion. To address this issue, we propose to maintain multiple tracking trajectories and apply beam search strategy for visual tracking, so that the trajectory with fewer accumulated errors can be identified. Accordingly, this paper introduces a novel multi-agent reinforcement learning based beam search tracking strategy, termed BeamTracking. It is mainly inspired by the image captioning task, which takes an image as input and generates diverse descriptions using beam search algorithm. Accordingly, we formulate the tracking as a sample selection problem fulfilled by multiple parallel decision-making processes, each of which aims at picking out one sample as their tracking result in each frame. Each maintained trajectory is associated with an agent to perform the decision-making and determine what actions should be taken to update related information. When all the frames are processed, we select the trajectory with the maximum accumulated score as the tracking result. Extensive experiments on seven popular tracking benchmark datasets validated the effectiveness of the proposed algorithm.
|
['Bo Jiang', 'DaCheng Tao', 'Bin Luo', 'Jin Tang', 'Zhe Chen', 'Xiao Wang']
|
2022-05-19
| null | null | null | null |
['visual-tracking']
|
['computer-vision']
|
[ 1.30679831e-01 -5.11637986e-01 -4.54884946e-01 -1.00582510e-01
-5.08031666e-01 -5.75756192e-01 2.32353702e-01 2.57446259e-01
-6.37194276e-01 8.41973364e-01 -2.58084774e-01 -5.81594296e-02
-1.44672751e-01 -5.74826658e-01 -4.96744365e-01 -1.09357071e+00
1.29905224e-01 3.59632492e-01 7.08663702e-01 4.04322654e-01
3.79803449e-01 4.77449983e-01 -1.19701350e+00 -1.92480281e-01
7.44458079e-01 1.01755941e+00 7.89550543e-01 4.16858971e-01
-9.45937261e-02 9.38613713e-01 -6.61619246e-01 -6.87823892e-02
8.80169719e-02 -5.53100884e-01 -3.93255293e-01 4.01313663e-01
1.77697182e-01 -4.82314885e-01 -1.03101499e-01 1.20116770e+00
4.82543051e-01 3.83901775e-01 2.33222485e-01 -1.50138915e+00
-1.98155478e-01 4.90668774e-01 -7.32673228e-01 7.46165872e-01
3.05533826e-01 3.79300267e-01 7.05949903e-01 -6.88602626e-01
7.09430754e-01 1.19905555e+00 3.50614101e-01 7.64616191e-01
-9.08683538e-01 -8.21221411e-01 6.95140183e-01 3.75260383e-01
-1.33575487e+00 -4.36539352e-01 7.57897019e-01 -3.36136132e-01
1.32822067e-01 1.03461392e-01 8.42200398e-01 6.70732975e-01
2.36554638e-01 1.07873070e+00 8.15257490e-01 -1.18928708e-01
4.01881695e-01 5.21282619e-03 -1.38992935e-01 6.33017540e-01
4.44006741e-01 2.51392186e-01 -4.32223886e-01 -2.20958546e-01
6.78722203e-01 6.38820529e-02 -4.50440228e-01 -4.72109765e-01
-1.48126531e+00 7.64477015e-01 3.84078056e-01 1.82980075e-01
-7.38389850e-01 2.58340538e-01 2.78884530e-01 5.57228401e-02
5.16990833e-02 -2.51591891e-01 -1.09354831e-01 1.03946202e-01
-8.29964519e-01 4.94533509e-01 2.21146807e-01 8.17435145e-01
6.17062449e-01 1.79337449e-02 -5.50286293e-01 3.95805568e-01
7.26620018e-01 6.99646890e-01 2.53589422e-01 -9.89710033e-01
3.66407514e-01 4.37246561e-01 5.96605778e-01 -1.02099609e+00
-1.72605976e-01 -5.45864105e-01 -6.03782296e-01 4.59127098e-01
3.64594102e-01 -3.78568202e-01 -6.09938681e-01 1.67892575e+00
9.64348793e-01 4.99064475e-01 1.84504241e-01 1.28499472e+00
6.17392600e-01 7.79030263e-01 4.87204820e-01 -8.86135936e-01
1.25348115e+00 -7.63997614e-01 -8.02716434e-01 -1.77097380e-01
2.33890206e-01 -8.87823343e-01 2.15065092e-01 1.95177332e-01
-1.00716460e+00 -7.10929811e-01 -8.65691662e-01 1.07401276e+00
2.37265423e-01 2.38665819e-01 1.80494681e-01 2.90498614e-01
-7.23871410e-01 2.79761732e-01 -7.59707689e-01 -3.47803235e-01
3.95249903e-01 1.25778690e-01 4.53760624e-02 -3.80383730e-02
-8.83798778e-01 5.89598835e-01 7.60501862e-01 1.36693731e-01
-1.12836123e+00 -4.80105668e-01 -4.68436062e-01 -2.67744392e-01
6.91823900e-01 -7.29731977e-01 1.32408226e+00 -9.74156559e-01
-1.21373379e+00 2.33163550e-01 -3.24455261e-01 -5.46370387e-01
6.08196557e-01 2.40267888e-02 -3.27008396e-01 -2.29319856e-02
3.14571321e-01 6.30319297e-01 1.07346797e+00 -1.32253647e+00
-1.33602118e+00 -1.45240366e-01 -3.39459069e-02 6.02861226e-01
-1.65992007e-01 3.42296928e-01 -6.91041350e-01 -5.09756505e-01
4.41696979e-02 -1.03123426e+00 -2.57335603e-01 -1.18414196e-03
6.40156027e-03 -4.77875203e-01 9.89379764e-01 -2.88485616e-01
1.43580139e+00 -1.94891632e+00 1.47051677e-01 7.42033124e-02
9.82683003e-02 3.29228610e-01 6.98130727e-02 1.54869109e-01
3.57948780e-01 -3.72870326e-01 2.43605584e-01 -7.79114440e-02
-3.23651165e-01 -4.83516157e-02 -1.16591178e-01 5.45447469e-01
-1.14450984e-01 5.69156706e-01 -1.18569183e+00 -1.08736408e+00
3.51806313e-01 6.16969988e-02 -1.76585630e-01 4.18233931e-01
-5.31364858e-01 7.87903070e-01 -1.14206648e+00 6.66437805e-01
8.21469426e-01 -5.52422345e-01 7.71424547e-02 -4.03912336e-01
-4.36926484e-01 -5.69817603e-01 -1.62345445e+00 1.29638004e+00
4.56804829e-03 3.59148532e-01 9.26377326e-02 -5.51846623e-01
8.50057721e-01 2.61134386e-01 7.74892926e-01 -6.44451499e-01
7.72058219e-02 1.46167260e-02 1.37959421e-01 -4.63230073e-01
3.26046437e-01 1.16257191e-01 3.06606770e-01 4.14110601e-01
-5.77620625e-01 5.57433844e-01 3.65267426e-01 6.55807555e-02
9.95139658e-01 3.87935162e-01 2.54832834e-01 2.17170164e-01
8.78238976e-01 4.02743757e-01 1.16167998e+00 8.52636039e-01
-6.26760662e-01 2.71400660e-02 -2.83132732e-01 -5.72141886e-01
-6.76560402e-01 -8.93231034e-01 2.25921065e-01 1.01317692e+00
7.09812343e-01 -1.01146296e-01 -5.26883364e-01 -6.85213923e-01
-1.82210639e-01 3.39125574e-01 -2.83680350e-01 -3.24985050e-02
-8.87597501e-01 -5.41925490e-01 3.63638736e-02 1.75965369e-01
6.85531020e-01 -1.54231608e+00 -1.29147029e+00 4.78632480e-01
-1.72334671e-01 -8.96364033e-01 -6.87739015e-01 -3.11596960e-01
-6.84255779e-01 -1.18837810e+00 -8.83298278e-01 -8.66341650e-01
7.64063656e-01 5.48218727e-01 6.11891866e-01 3.20855409e-01
1.12748131e-01 1.71683669e-01 -5.59826016e-01 -1.65944338e-01
-4.69290435e-01 -1.97630927e-01 3.54039371e-02 3.54119569e-01
1.10851206e-01 1.89684123e-01 -9.05335605e-01 5.39253891e-01
-7.36161590e-01 6.82495758e-02 6.53611302e-01 6.17511630e-01
8.48939180e-01 3.86297673e-01 5.63712656e-01 -3.29956591e-01
5.67786396e-01 -4.26534653e-01 -9.74644482e-01 4.46994871e-01
-3.49945903e-01 -1.77795589e-01 6.81846917e-01 -7.27337122e-01
-1.03884912e+00 4.00183856e-01 2.78359503e-01 -8.52898538e-01
-1.48689542e-02 4.39507842e-01 -6.24175258e-02 -2.57472724e-01
1.96711570e-01 5.90240359e-01 -2.00416204e-02 -1.47305354e-01
5.30156977e-02 3.39775234e-01 4.00482446e-01 -2.44555905e-01
9.36919034e-01 3.00956786e-01 -3.83150354e-02 -3.74456346e-01
-6.29430056e-01 -4.83293802e-01 -2.30079278e-01 -9.32103157e-01
1.11692321e+00 -7.97086477e-01 -1.17854059e+00 3.61209214e-01
-1.15191114e+00 1.23788565e-01 2.40489677e-01 7.42627382e-01
-5.28223813e-01 3.76030892e-01 -1.54824749e-01 -1.15371299e+00
-4.48483855e-01 -1.30707633e+00 8.21640909e-01 8.98089290e-01
-3.02158669e-03 -7.78316081e-01 1.77737564e-01 -1.48890736e-02
2.63374627e-01 3.83605093e-01 4.52570796e-01 -4.26035672e-01
-1.18206322e+00 -4.22843024e-02 -2.03601848e-02 -2.81977743e-01
2.05705032e-01 -6.59599900e-02 -3.87495339e-01 -6.53012037e-01
-2.05051929e-01 -2.03958806e-02 4.60938960e-01 6.91405654e-01
8.03855300e-01 -1.11533619e-01 -7.83660412e-01 2.92103440e-01
1.50391591e+00 9.64942932e-01 1.99314505e-01 6.74366176e-01
4.89257276e-01 7.51935765e-02 1.54603171e+00 6.02763772e-01
1.95228219e-01 9.44380760e-01 7.18294680e-01 1.69546217e-01
3.86190116e-02 -2.34847993e-01 4.86610055e-01 3.81751984e-01
2.55315661e-01 -3.92601281e-01 -5.68742275e-01 3.33254755e-01
-2.23532796e+00 -1.19482660e+00 1.30057231e-01 2.32120013e+00
5.48192680e-01 2.87885636e-01 3.81587654e-01 -2.30978638e-01
1.07863104e+00 2.45375752e-01 -9.59931433e-01 3.77594292e-01
7.99115673e-02 -5.61389506e-01 4.39152211e-01 2.82675326e-01
-1.17968810e+00 1.01232672e+00 4.78262091e+00 9.25175428e-01
-1.07690895e+00 7.74161704e-03 3.75518113e-01 -4.76632155e-02
1.42224774e-01 1.26861662e-01 -1.09129274e+00 8.05517614e-01
5.31666696e-01 -3.15655142e-01 3.06604594e-01 5.75657248e-01
6.71166420e-01 -3.69294047e-01 -7.00199425e-01 1.00726974e+00
-1.68946773e-01 -1.29274845e+00 5.42209893e-02 -2.03681022e-01
4.93194252e-01 -2.17858478e-01 -1.17238685e-01 -1.72373373e-02
4.70239937e-01 -3.67530882e-01 8.85892868e-01 6.91549718e-01
1.56460315e-01 -6.83174312e-01 3.37458313e-01 5.76157629e-01
-1.67557049e+00 -2.34484091e-01 -3.73614967e-01 4.77924198e-01
3.94667327e-01 1.24510035e-01 -7.95876086e-01 6.54779494e-01
7.09880888e-01 7.47863531e-01 -3.22692335e-01 1.83229458e+00
7.18895867e-02 3.47515136e-01 -1.50479451e-01 -3.72794479e-01
2.55894363e-01 -2.90116012e-01 9.46517289e-01 7.59437501e-01
4.16154116e-01 1.48821354e-01 8.37825775e-01 5.03799081e-01
3.42729896e-01 1.48126587e-01 -3.33800405e-01 2.69213796e-01
8.99158478e-01 1.29385781e+00 -1.05660713e+00 -3.93614620e-01
-3.07669699e-01 7.30181694e-01 2.64427513e-01 3.22109848e-01
-1.18301773e+00 -7.78261721e-02 2.58271515e-01 -1.99209020e-01
6.04946375e-01 3.26985978e-02 4.26952928e-01 -7.09210575e-01
1.06631042e-02 -7.06833482e-01 7.54342377e-01 -8.75327528e-01
-7.75374651e-01 7.30632603e-01 1.27841160e-02 -1.72659051e+00
-2.26285279e-01 6.54537007e-02 -7.81093717e-01 7.14697421e-01
-1.40830231e+00 -8.07545722e-01 -6.00447536e-01 6.02282226e-01
8.01699400e-01 -7.69787729e-02 1.85272336e-01 3.56985092e-01
-7.02706993e-01 2.33065277e-01 -2.39072535e-02 -2.39029881e-02
4.98586893e-01 -8.79013181e-01 -1.96422949e-01 8.20304334e-01
-1.77787751e-01 1.87452614e-01 8.55660260e-01 -1.03382051e+00
-1.45436454e+00 -1.26684165e+00 5.20144343e-01 7.68091902e-02
6.01047575e-01 3.11321914e-01 -7.82486558e-01 4.70357656e-01
1.92634001e-01 9.89729092e-02 1.26850918e-01 -7.58343577e-01
5.38197577e-01 -2.36448362e-01 -1.01115882e+00 6.11482799e-01
9.89639938e-01 3.17433387e-01 -3.02758753e-01 4.07675624e-01
6.13725603e-01 -5.76908529e-01 -7.85499156e-01 4.47029233e-01
2.83429772e-01 -6.40298724e-01 8.38884771e-01 -3.19050103e-01
-3.12073022e-01 -1.07000315e+00 1.24667764e-01 -1.29150307e+00
-6.03755414e-01 -6.14050567e-01 -2.83687830e-01 1.32181871e+00
6.91901818e-02 -2.43168861e-01 9.55608606e-01 3.14148247e-01
2.25220442e-01 -8.01037192e-01 -9.50666606e-01 -6.93048894e-01
-6.36323690e-01 1.45451218e-01 5.23501873e-01 4.25545216e-01
-5.67877233e-01 1.03782192e-01 -5.13798952e-01 4.23417330e-01
1.03399706e+00 6.32220745e-01 8.38234544e-01 -1.06713438e+00
-4.90198769e-02 -2.28746876e-01 -2.52098024e-01 -1.28838086e+00
1.44160450e-01 -6.64755225e-01 3.79156768e-01 -1.53061628e+00
3.32523614e-01 -6.75153196e-01 -3.46897453e-01 1.95496485e-01
-4.65081692e-01 -1.70863956e-01 5.78003883e-01 5.72224557e-01
-1.24992156e+00 5.25780737e-01 1.37393773e+00 -2.06235856e-01
-3.12628448e-01 5.47608495e-01 -2.75706261e-01 5.79589963e-01
6.06371641e-01 -6.20631814e-01 -4.19624835e-01 -3.15974146e-01
-2.48423785e-01 5.83091080e-01 2.79762417e-01 -1.16392815e+00
7.62537837e-01 -6.06844366e-01 6.64056182e-01 -1.02070177e+00
1.53495699e-01 -1.23215425e+00 5.12181520e-01 9.78059769e-01
-2.83911437e-01 4.33441460e-01 1.26521140e-01 8.62290859e-01
-3.24575393e-03 -4.99340683e-01 7.84358680e-01 -8.42353180e-02
-1.21591198e+00 6.39638007e-01 -4.54157025e-01 -1.90002233e-01
1.58474576e+00 -2.87536323e-01 -1.90724626e-01 -1.57297909e-01
-7.84342051e-01 8.03560555e-01 3.36162150e-01 4.55033213e-01
7.60163188e-01 -1.57263374e+00 -7.57736802e-01 -9.16888192e-02
8.12612027e-02 -1.60733357e-01 2.74323106e-01 7.90374637e-01
-1.31025955e-01 2.40557134e-01 -8.66980758e-03 -8.49653006e-01
-1.64493906e+00 7.85733223e-01 3.74244511e-01 -2.34614551e-01
-6.33389473e-01 4.70376432e-01 8.14609975e-03 4.12079543e-01
3.73940259e-01 4.02295887e-02 -6.79910719e-01 -1.61845777e-02
6.98349416e-01 2.88041770e-01 -3.25932056e-01 -8.18145633e-01
-5.71124375e-01 8.33836913e-01 -1.44855469e-01 3.96026187e-02
9.14027810e-01 -4.42441225e-01 1.59759462e-01 1.72632441e-01
7.14109182e-01 -1.87003642e-01 -1.63425231e+00 -4.13640767e-01
1.62070664e-03 -8.11342061e-01 2.58047041e-02 -6.69083357e-01
-1.34755373e+00 -5.82748353e-02 9.28659141e-01 1.25269890e-01
1.24655354e+00 -1.03322230e-01 7.51328588e-01 1.31041273e-01
5.78101933e-01 -9.00946081e-01 4.12259735e-02 1.69517487e-01
5.39624214e-01 -1.16627967e+00 -8.32971651e-04 -2.46448606e-01
-6.70606136e-01 8.98351490e-01 1.06786263e+00 1.69085473e-01
4.24897075e-01 -1.18168416e-02 -8.17670859e-03 -4.11457159e-02
-8.68438661e-01 -3.80700052e-01 -5.95050193e-02 4.73454744e-01
-1.37465224e-01 -2.42106676e-01 -3.58199894e-01 -1.37188435e-01
4.27316427e-01 3.76820594e-01 1.29117593e-01 1.25374770e+00
-8.81980121e-01 -9.41536784e-01 -7.28768885e-01 3.59300256e-01
-2.00657770e-01 3.94250393e-01 4.55346070e-02 5.28035820e-01
4.50185798e-02 1.10777199e+00 6.38820902e-02 -2.72783130e-01
2.49860838e-01 -4.05727267e-01 4.07958806e-01 -2.09772661e-01
-4.79401916e-01 2.95888901e-01 -1.99891776e-01 -3.42110604e-01
-7.76882052e-01 -1.01183856e+00 -1.53000963e+00 -1.85059950e-01
-4.15865064e-01 4.50587302e-01 3.89100373e-01 7.96833515e-01
2.51860559e-01 5.82379222e-01 8.49894524e-01 -7.48945832e-01
-5.88830769e-01 -5.92775941e-01 -1.37243748e-01 3.36824626e-01
4.14928526e-01 -9.65275705e-01 -3.10030654e-02 1.27671063e-01]
|
[6.385588645935059, -2.099343776702881]
|
0a543024-8c85-4b1f-b390-a693fa70d24f
|
meta3d-single-view-3d-object-reconstruction
|
2003.03711
| null |
https://arxiv.org/abs/2003.03711v3
|
https://arxiv.org/pdf/2003.03711v3.pdf
|
Single-View 3D Object Reconstruction from Shape Priors in Memory
|
Existing methods for single-view 3D object reconstruction directly learn to transform image features into 3D representations. However, these methods are vulnerable to images containing noisy backgrounds and heavy occlusions because the extracted image features do not contain enough information to reconstruct high-quality 3D shapes. Humans routinely use incomplete or noisy visual cues from an image to retrieve similar 3D shapes from their memory and reconstruct the 3D shape of an object. Inspired by this, we propose a novel method, named Mem3D, that explicitly constructs shape priors to supplement the missing information in the image. Specifically, the shape priors are in the forms of "image-voxel" pairs in the memory network, which is stored by a well-designed writing strategy during training. We also propose a voxel triplet loss function that helps to retrieve the precise 3D shapes that are highly related to the input image from shape priors. The LSTM-based shape encoder is introduced to extract information from the retrieved 3D shapes, which are useful in recovering the 3D shape of an object that is heavily occluded or in complex environments. Experimental results demonstrate that Mem3D significantly improves reconstruction quality and performs favorably against state-of-the-art methods on the ShapeNet and Pix3D datasets.
|
['Jiahao Xia', 'Stuart Perry', 'Haozhe Xie', 'Min Xu', 'Shuo Yang']
|
2020-03-08
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Yang_Single-View_3D_Object_Reconstruction_From_Shape_Priors_in_Memory_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Yang_Single-View_3D_Object_Reconstruction_From_Shape_Priors_in_Memory_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['single-view-3d-reconstruction', '3d-object-reconstruction']
|
['computer-vision', 'computer-vision']
|
[-0.09800956 -0.21140818 -0.03886572 -0.45397347 -0.5575351 -0.32716322
0.3906545 -0.3277357 0.02688159 0.31849223 0.25715947 0.091216
0.1458595 -1.0181725 -1.2022129 -0.6970538 0.58048135 0.52941924
0.26801237 0.01351733 0.26658985 0.89097 -1.6670485 0.49770024
0.44737503 1.3577267 0.7948068 0.01339569 -0.555476 0.3282257
-0.26757917 -0.08295921 0.46908477 -0.05814155 -0.47577155 0.45784953
0.68717206 -0.8107353 -0.6065187 0.8361949 0.5628569 -0.07113281
0.74200743 -0.6814422 -0.9881792 -0.09801971 -0.5868768 -0.20536675
0.32820523 0.23150524 0.5015106 -1.5210515 0.89435494 1.6056556
0.42166504 0.45673633 -1.1821716 -0.47167698 0.22310768 0.03648836
-1.3491505 -0.5103083 1.1821356 -0.30823243 1.0878897 0.02170682
0.8619704 0.9733083 0.14480625 1.0525737 1.0427796 -0.07443909
-0.11663974 0.03482532 -0.27011833 0.9399018 0.07894713 0.37334988
-0.7715385 -0.06699237 1.3981299 0.349479 -0.4189938 -0.79689014
-1.1887196 0.3897985 0.7659594 0.0625122 -0.706273 0.17672811
-0.04168329 0.08731811 0.6148699 -0.01689957 -0.3425502 0.3385421
-0.6079384 0.09313464 0.34050235 1.141135 0.9471532 0.26967663
-0.18812022 0.7827831 0.5938939 1.0218047 0.0375424 -0.99043
0.47744626 0.93020153 0.20283087 -1.0531049 -0.09525058 -0.4220624
-0.7445483 0.25252405 -0.03528697 0.66978985 -1.1798027 1.480025
0.53373134 0.1307935 -0.19922891 1.2223607 1.5111364 0.71431243
-0.32165977 0.06591471 1.1137112 -0.64806294 -0.39754814 -0.2753913
-0.13651489 -0.67408264 1.0091064 0.05501895 -1.5062782 -0.7478295
-1.0091794 -0.444388 -0.0779975 0.10108873 0.32190993 0.21082509
-0.8924974 0.42657658 -0.769935 0.11509304 0.8800382 0.13778976
-0.526095 -0.5742756 -0.58138907 0.8966249 0.23933522 0.15155365
-1.1589301 -0.85819364 -1.1006227 -0.00595325 0.3091872 -1.1890568
0.77167505 -0.51231265 -1.2028005 1.267349 -0.4031461 0.0639628
0.22141248 -0.03677579 0.19156691 0.3703226 -0.0172294 0.8478146
1.2343514 -1.8606249 -0.05857671 -0.7594032 -0.07326698 0.32771984
0.17473994 -0.4636271 -0.75590616 -0.707877 0.74345475 -0.5011153
0.01667996 0.6186328 -0.2635534 -0.10098164 0.8376553 -0.7099079
0.40047446 -2.1335108 0.23552062 0.08127722 0.11138274 0.27109465
-0.41149342 0.17544022 0.28237903 -0.11933727 -0.2867176 -0.62916726
-0.06710748 0.5579631 -0.6070261 0.42634493 0.31629154 1.272342
-0.7912726 -0.3187928 0.41864845 0.947757 -0.4729609 0.4997697
-0.39774203 0.7493008 -0.6613205 0.86513793 1.0096722 -0.40051466
-0.3800352 -0.5880441 0.09947063 0.28345504 -0.85111654 2.349952
-0.43152314 0.07993835 0.2026884 -0.90679073 1.2039735 0.2104427
0.4204105 -0.8668799 0.22634944 0.33819422 -0.6734472 -0.52563393
0.20825128 -0.18405977 0.40358618 0.4538398 0.11335699 -0.7234789
-0.36594072 -0.11136241 0.5613355 0.55930024 -0.02629795 0.1401926
0.48180762 -0.3744765 0.48370633 0.49287966 0.2787572 1.0225401
-0.05602008 -0.7106781 -1.1770984 -1.4914639 -0.15685366 0.42286417
0.4039516 0.06602854 -0.17667171 -0.50710076 0.20098633 0.5040927
-0.3882805 -0.26958987 -0.6456558 -0.12810169 -0.07186934 0.5712812
0.6748538 -1.2979368 -0.65413857 0.19131841 -0.36072978 -1.1683347
-0.6654451 -0.08268233 -1.3754146 -1.0038998 -0.8360089 -1.0269294
1.218607 0.693931 1.2580688 0.2875432 -0.27119824 0.561616
-0.165151 -0.1524625 -0.2025151 -0.52908945 -0.20334275 0.08704074
0.13231502 -0.99537116 -0.71406454 0.29400858 -0.997614 0.33647963
0.92864954 0.9488368 1.304954 -0.29738578 0.4090687 -0.59178275
0.01618218 -0.21948911 -0.4499054 0.17511077 -0.22310095 0.18698345
0.36656266 -0.45143777 -1.1342543 0.3956623 -0.27402827 -1.033915
-0.19500539 0.26078296 -0.42848855 -0.08365289 0.28426465 0.89166677
0.14293255 -0.9465411 0.37064365 0.24221729 0.53355813 -0.7089626
0.73682374 0.8643756 0.03000202 -0.76246864 -1.0384403 -0.30328208
-0.6540453 -0.09396601 0.6065926 -1.1194273 -0.47980043 0.6384473
-1.3854957 -0.11286672 -0.38236168 0.31471953 -0.8978209 0.5046694
-0.55760396 -0.5048304 -0.6008441 -1.1882647 1.5200014 0.0567646
0.45760053 -0.55911493 -0.44091657 0.49471745 0.31623828 0.24441803
1.2846986 0.21741351 -1.1737056 0.02972745 -0.5213571 0.40426672
0.14509791 -0.61108804 -1.0248814 -0.35185117 0.31956235 -0.42317277
1.0014722 0.59235406 1.3177037 -0.21429114 -0.30799437 0.8955744
1.3537778 -0.02465348 0.6108365 -0.2258402 0.7783318 0.4666467
0.26952702 0.47082996 0.49095848 0.59376985 0.67257255 -0.06175154
-0.61617 -0.8140425 0.14642373 0.9940886 0.05425542 0.04802701
-0.40976572 0.47783396 -1.5141724 -0.55371904 0.02523652 2.1548233
0.9539331 0.03785039 -0.4435482 -0.16897342 0.3774532 0.2950594
-0.905146 0.21607094 -0.36791474 0.24963087 0.04984337 0.35624936
-0.6475953 0.8306165 5.5964704 0.7258529 -1.088009 -0.01965677
0.4318816 -0.01063467 -0.65596545 -0.06284087 -0.61288124 0.2072484
0.01943741 0.23088633 0.39833957 0.5565487 0.08327828 -0.17096528
-1.1571863 1.4742576 0.3078272 -1.3923081 0.51509845 0.07670063
0.8241902 0.01831887 0.24970222 -0.08131522 -0.14646193 -0.97560465
1.0326823 1.0159966 0.9028881 -0.40765008 0.3186555 0.5353872
-1.1349514 0.22837411 -0.86079293 0.29749548 0.32879668 0.8072761
-0.6791302 0.54017305 0.82566357 1.0076475 -0.3710446 1.219128
-0.28359398 0.09492966 -0.31457242 0.2300563 0.10350128 -0.17749071
0.7762932 0.4104371 0.45853734 0.38860047 0.09887962 1.6085849
-0.21630502 -0.08209748 -0.96312237 0.15550464 0.46411344 0.9242454
-0.5780351 -0.25628275 -0.33136597 0.91998833 0.38392192 0.5097495
-0.3287435 0.26217288 0.35626447 0.34873337 0.7250477 -0.49766064
-0.3270033 -1.0009345 0.39312017 -0.45199156 -0.10160718 -1.363381
-1.4825379 0.48363358 -0.08904405 -1.3374652 0.18302737 -0.6251867
-0.26568094 0.90458024 -1.7776072 -1.3155797 -0.40793324 0.6814948
0.53562546 -0.07486775 0.69268334 0.18401107 0.17739545 0.11266921
-0.24879462 -0.07974774 0.43725616 -0.7385587 0.4191651 0.3098622
0.30453825 0.4239161 0.11877357 -0.80580354 -1.9551425 -1.0155495
0.5509583 -0.49265045 -0.27155513 -0.41845337 -1.0291845 0.37064323
-0.19804011 0.38799343 0.2341025 -0.5665282 -0.5686529 -0.01710699
-1.180687 0.34952807 1.5187242 -0.6193699 -0.76874113 0.06506611
0.81753975 -0.673477 -0.7296151 0.45563883 0.46298584 -1.0462854
1.4802239 -0.44960785 0.38487113 -0.34001055 -0.57843834 -1.199416
-0.12448333 -0.05571303 -0.33109388 0.8144925 0.0581092 -0.39149886
0.7959087 0.3946691 -0.38782588 -0.95718557 -1.0196083 -0.6579419
-0.0413594 -0.32182342 0.7670285 0.49940503 -0.83209676 0.10542824
-0.3357304 0.0922937 0.8727144 0.8904858 0.69384277 -1.2255459
0.01229214 -0.3265031 -0.37362987 -1.6917834 0.31232136 -1.1768777
0.11921496 -1.8234897 0.23631418 -0.6217743 -0.01245835 0.5141815
0.08684704 0.38954556 0.23187833 0.19088966 -0.36836186 1.2220414
2.066492 -0.4216899 -0.14953189 0.01057902 -0.53618586 0.62004024
0.19425453 -0.3973286 -0.40093625 -0.8852261 0.08480028 0.29060617
0.73155165 -0.55172294 0.11929915 -0.02342628 0.88393766 -1.341338
0.884493 -1.0668181 0.18813024 0.21392626 0.05410201 -0.20423117
0.09694059 0.678266 -0.14739454 -0.05704729 0.57627225 -0.5711201
-0.59240305 0.80621684 0.32793558 0.04315044 0.5009321 -0.38497758
0.02717878 -0.28598434 -0.7072023 0.10360129 0.61944544 0.5980776
1.5687679 -1.8260022 -0.7655404 0.7444614 0.06567514 0.5251201
0.39370146 0.64060193 -0.36853114 0.30685088 -0.49793258 -0.8948755
-0.80245984 0.7436691 0.34179628 0.2837347 -1.1772116 0.76749456
0.56637007 -0.6110496 0.3846172 -0.49611947 0.16594951 -0.2986787
0.41701487 -0.13671571 0.12073844 -0.8171859 -0.16817333 1.048514
0.23905794 0.1238435 1.7463318 -0.20358361 -0.18832976 0.37531438
1.2881342 -0.27643383 -1.7194724 -0.8504391 -0.5175523 -0.9231883
0.17721313 -0.63383 -1.5300529 1.0880718 0.5623067 -0.45624703
1.0882215 0.2528642 1.0993491 0.3929097 0.68526644 -0.6397473
0.42791024 0.4619275 1.4414897 -1.166604 0.08574624 -0.41650462
-0.32951272 1.0017171 0.58878654 -0.24349843 0.913894 0.05318854
-0.1890204 -0.53739136 -0.49510592 -0.13929154 0.6367893 0.800543
-0.1879565 -0.15247065 0.2529402 0.69732153 0.1692674 -0.25792086
0.1043852 0.752478 -0.44411486 -0.8416797 -0.37236112 0.4112066
0.11785283 0.03486422 -0.35897747 0.32597205 0.04018283 0.4387722
0.07079142 -0.18836533 0.53479093 0.00596015 1.0076553 -0.61356866
-0.10480253 0.23843719 -0.4409755 -0.7755239 -0.6199682 -0.30405986
-1.3520261 0.02173954 -0.0957964 -0.50950474 0.60477865 0.7892779
0.52133185 0.20395552 0.64781827 -1.4069715 -0.35439786 -0.596153
-0.56317836 0.5273468 0.33964226 -1.1634088 -0.00682554 -0.03708532]
|
[8.44268798828125, -3.362149477005005]
|
c2dfb984-4907-4a9d-94ff-87c23fa021d2
|
conformer-and-blind-noisy-students-for
|
2204.12819
| null |
https://arxiv.org/abs/2204.12819v1
|
https://arxiv.org/pdf/2204.12819v1.pdf
|
Conformer and Blind Noisy Students for Improved Image Quality Assessment
|
Generative models for image restoration, enhancement, and generation have significantly improved the quality of the generated images. Surprisingly, these models produce more pleasant images to the human eye than other methods, yet, they may get a lower perceptual quality score using traditional perceptual quality metrics such as PSNR or SSIM. Therefore, it is necessary to develop a quantitative metric to reflect the performance of new algorithms, which should be well-aligned with the person's mean opinion score (MOS). Learning-based approaches for perceptual image quality assessment (IQA) usually require both the distorted and reference image for measuring the perceptual quality accurately. However, commonly only the distorted or generated image is available. In this work, we explore the performance of transformer-based full-reference IQA models. We also propose a method for IQA based on semi-supervised knowledge distillation from full-reference teacher models into blind student models using noisy pseudo-labeled data. Our approaches achieved competitive results on the NTIRE 2022 Perceptual Image Quality Assessment Challenge: our full-reference model was ranked 4th, and our blind noisy student was ranked 3rd among 70 participants, each in their respective track.
|
['Radu Timofte', 'Maxime Burchi', 'Marcos V. Conde']
|
2022-04-27
| null | null | null | null |
['blind-image-quality-assessment', 'no-reference-image-quality-assessment']
|
['computer-vision', 'computer-vision']
|
[ 9.29147601e-02 -3.09756815e-01 1.32569402e-01 -4.60555673e-01
-1.33335996e+00 -3.85828137e-01 4.27209914e-01 -3.48871686e-02
-3.19811255e-01 5.19695699e-01 4.20488387e-01 -1.99926600e-01
-5.37040047e-02 -5.43792665e-01 -6.61280990e-01 -5.17152965e-01
3.22550654e-01 -1.06950618e-01 4.21336293e-02 -1.18491121e-01
3.49446088e-01 -4.27832119e-02 -1.70972979e+00 3.51669878e-01
1.53119457e+00 1.03890121e+00 2.70367563e-01 9.01666880e-01
1.84958562e-01 9.61109281e-01 -1.02915204e+00 -8.14786077e-01
3.28777641e-01 -6.46850824e-01 -7.12220907e-01 2.17747271e-01
8.10250461e-01 -5.66634893e-01 -3.12844932e-01 1.53684652e+00
9.47016120e-01 4.85400185e-02 6.34283364e-01 -1.25684655e+00
-1.29769647e+00 4.82416153e-01 -3.01717818e-01 2.80287474e-01
6.41326368e-01 4.21618015e-01 8.59583974e-01 -9.21697855e-01
2.35036790e-01 1.34007752e+00 3.33760560e-01 4.61964458e-01
-1.18514764e+00 -6.92920029e-01 3.91605757e-02 9.04801369e-01
-1.22445536e+00 -6.65856361e-01 5.40388465e-01 -3.28253835e-01
5.41287839e-01 1.48681834e-01 6.70758069e-01 9.55830932e-01
-1.01450242e-01 7.50262141e-01 1.56848872e+00 -3.84398192e-01
2.91492015e-01 1.89199060e-01 -3.51733714e-01 4.25630420e-01
-1.04785904e-01 4.80539024e-01 -5.58126748e-01 2.99082518e-01
6.14162624e-01 -2.47741759e-01 -7.06692040e-01 -3.33853632e-01
-1.18222880e+00 5.02424121e-01 7.15855479e-01 4.81206663e-02
-2.94480354e-01 -1.07922852e-01 -5.26005588e-02 5.67370474e-01
2.61839479e-01 5.39587617e-01 -2.41715740e-02 -1.90935925e-01
-1.09121990e+00 6.83365166e-02 4.74616766e-01 7.52378404e-01
3.41230869e-01 9.97618511e-02 -5.70872843e-01 1.06562424e+00
4.53522533e-01 7.45309055e-01 5.04008889e-01 -1.39245367e+00
2.15374500e-01 3.05331141e-01 4.82139707e-01 -8.90430450e-01
3.26698452e-01 -6.96623087e-01 -8.29091191e-01 7.05795169e-01
1.97876200e-01 2.35208258e-01 -1.19806063e+00 1.63801515e+00
-1.34310825e-02 -9.79261007e-03 2.71416038e-01 1.20322943e+00
9.91569877e-01 7.80019164e-01 -9.78342444e-02 -3.36479247e-01
1.18380487e+00 -1.19440341e+00 -7.78276443e-01 -6.16160221e-02
-2.06689745e-01 -1.15054464e+00 1.35082912e+00 1.00121641e+00
-1.47780371e+00 -1.10800266e+00 -1.29433310e+00 7.64661506e-02
3.11918631e-02 1.58280164e-01 -6.62737191e-02 1.02907550e+00
-1.51202714e+00 5.69236219e-01 -3.44742566e-01 -8.16191360e-02
3.75409335e-01 -6.26447275e-02 -3.32001410e-02 -5.39798379e-01
-1.04943621e+00 1.11146140e+00 -7.41493255e-02 -6.65341541e-02
-1.42928541e+00 -5.93356133e-01 -6.56884491e-01 -1.23606451e-01
-1.29551649e-01 -6.97342336e-01 1.59063959e+00 -8.81967247e-01
-1.63779843e+00 7.17121661e-01 -2.32795045e-01 -3.85261983e-01
6.44734740e-01 -2.83894777e-01 -8.76662612e-01 2.81489521e-01
8.73846635e-02 7.59230435e-01 1.24199545e+00 -1.73650372e+00
-6.73486710e-01 -1.50374889e-01 1.88033864e-01 4.06813532e-01
-3.29989225e-01 -5.64320711e-03 -4.94224429e-01 -7.24602997e-01
1.80393904e-01 -4.96725023e-01 -1.16071045e-01 1.10943742e-01
-2.60387003e-01 -7.72145987e-02 4.99737859e-01 -1.03087997e+00
1.00761104e+00 -2.23544931e+00 -1.99626699e-01 6.36545150e-03
3.03489447e-01 5.31997621e-01 -3.75448346e-01 -2.93091908e-02
-3.90966795e-02 -1.47735411e-02 4.61178413e-03 -2.90034473e-01
1.47495314e-01 -1.48313075e-01 -1.76872596e-01 2.54709601e-01
5.44622466e-02 6.26935840e-01 -1.07761335e+00 -5.83323419e-01
2.96021670e-01 6.41583204e-01 -4.44804937e-01 5.85595727e-01
1.68233544e-01 4.71798390e-01 1.00497015e-01 5.72948575e-01
8.04785788e-01 -2.42547318e-01 -2.84875542e-01 -5.90035081e-01
1.89840510e-01 1.78454965e-01 -1.12076294e+00 1.69158340e+00
-6.63013458e-01 6.43516898e-01 -7.91808441e-02 -4.56700712e-01
9.46939051e-01 6.35212719e-01 -3.24515486e-03 -1.26953518e+00
-1.86508805e-01 2.48433486e-01 -2.62398086e-02 -4.70687002e-01
5.34877300e-01 -2.46658288e-02 4.30903643e-01 3.45570415e-01
1.21913683e-02 -5.16506970e-01 3.08726817e-01 2.54895419e-01
8.87866855e-01 -1.01498440e-01 -1.14879347e-01 2.34967172e-01
5.11455178e-01 -3.56249005e-01 3.97004932e-01 8.08044851e-01
-6.61561906e-01 9.97236609e-01 4.34853956e-02 1.83287282e-02
-1.12468576e+00 -1.70602369e+00 -1.61910150e-02 8.28287840e-01
4.86702710e-01 -2.69634396e-01 -9.17890549e-01 -4.60645884e-01
-5.59383154e-01 6.59283340e-01 -1.94506258e-01 -3.32191765e-01
-9.86813102e-03 -5.01791358e-01 3.71720791e-01 3.74435097e-01
8.35821152e-01 -9.67330515e-01 -2.14363873e-01 3.39571908e-02
-7.71567643e-01 -1.04669762e+00 -7.39657760e-01 -4.22410607e-01
-5.67157090e-01 -9.12194312e-01 -1.18735373e+00 -9.04569447e-01
8.60256135e-01 3.66898328e-01 1.27238286e+00 -4.03621569e-02
-7.53281638e-02 3.40048343e-01 -4.98199850e-01 -1.85271993e-01
-6.68963194e-01 -6.20090723e-01 1.64762616e-01 -1.25439480e-01
-4.53879349e-02 -5.00369549e-01 -1.05715144e+00 5.34477592e-01
-8.34365547e-01 8.15631473e-04 8.81285489e-01 7.40043461e-01
6.06542230e-01 4.70127136e-01 6.73229992e-01 -5.89717329e-02
9.69988763e-01 2.56908923e-01 -4.46897119e-01 4.00471479e-01
-1.04470420e+00 -9.19223800e-02 5.40758193e-01 -5.56537747e-01
-1.28299069e+00 -3.80798906e-01 -1.64771423e-01 -4.21663076e-01
-2.49025486e-02 2.50320137e-01 -3.68604153e-01 -1.63534686e-01
9.24860656e-01 3.46535236e-01 -2.69046843e-01 -4.08138663e-01
5.47767401e-01 9.43383276e-01 1.12738383e+00 -2.98129290e-01
1.00669372e+00 3.90074439e-02 -5.14037311e-01 -3.49284202e-01
-9.27213132e-01 -3.67300838e-01 -1.05576433e-01 -4.14481461e-01
7.40825891e-01 -1.26994801e+00 -5.71738243e-01 7.41719425e-01
-1.16223300e+00 -1.37539193e-01 -3.05528671e-01 6.24802232e-01
-4.82625693e-01 5.28776467e-01 -5.14304996e-01 -7.75238276e-01
-3.78709763e-01 -1.48221731e+00 6.73132479e-01 4.08777595e-01
2.39381101e-02 -5.22990823e-01 9.67365410e-03 8.50015104e-01
6.25227511e-01 -2.34611049e-01 7.01188803e-01 1.56537548e-01
-4.81170386e-01 -6.45475313e-02 -5.93662858e-01 1.08545899e+00
2.82161325e-01 -3.78901213e-01 -1.19511688e+00 -4.84567612e-01
7.18989000e-02 -4.40652192e-01 7.49256194e-01 3.75278234e-01
1.25050378e+00 -3.16042930e-01 3.52925837e-01 4.82742250e-01
1.30110407e+00 2.91122288e-01 1.08826077e+00 1.91806212e-01
4.13616389e-01 3.72498810e-01 5.51630795e-01 1.46634728e-01
6.16456270e-01 5.60552955e-01 4.28679496e-01 -2.40010783e-01
-7.34619260e-01 -4.86480594e-01 7.59730935e-01 9.74371016e-01
-1.29116684e-01 -2.96036065e-01 -6.08582854e-01 6.76970184e-01
-1.27574778e+00 -9.88538444e-01 4.07453105e-02 2.28302884e+00
1.19469690e+00 1.00541681e-01 -5.96045190e-03 5.33725619e-01
8.13265860e-01 2.98853982e-02 -6.11765802e-01 -1.98447853e-01
-1.98765144e-01 1.86611325e-01 1.43596023e-01 4.52353925e-01
-9.14871633e-01 5.23970008e-01 6.81928205e+00 9.03936744e-01
-7.76807487e-01 1.63079232e-01 7.31367707e-01 -5.13026044e-02
-2.41471782e-01 -1.78334057e-01 -2.01447263e-01 5.32235026e-01
1.00663912e+00 -3.76295328e-01 6.29105330e-01 5.74912131e-01
4.04494315e-01 -7.15266019e-02 -9.85501707e-01 1.39555514e+00
3.82698864e-01 -8.43038857e-01 2.21625909e-01 -6.30005524e-02
1.15264344e+00 -1.90923125e-01 6.69893980e-01 1.74284577e-01
5.48916161e-01 -1.33026850e+00 7.89846838e-01 7.71964192e-01
9.25208688e-01 -6.73137367e-01 7.42720127e-01 1.09748393e-01
-9.33047175e-01 1.10175356e-03 -2.63139784e-01 2.66899675e-01
1.74438611e-01 6.64277375e-01 -5.90259075e-01 5.47210455e-01
1.06896305e+00 5.12840033e-01 -9.64298785e-01 1.62459540e+00
-5.64946711e-01 5.91683030e-01 3.22233975e-01 3.66957664e-01
-2.43982002e-01 4.93577495e-02 4.67313170e-01 7.98412025e-01
5.60470104e-01 8.63226969e-03 -7.09710568e-02 8.07259202e-01
-2.78010070e-01 -2.87817251e-02 9.12178308e-03 8.99534151e-02
3.12214017e-01 1.00748503e+00 -1.59226626e-01 -3.77617210e-01
-2.32921228e-01 1.10053551e+00 -3.77732635e-01 4.92384046e-01
-6.14708185e-01 -4.81182724e-01 4.45641071e-01 -5.65944724e-02
1.44337162e-01 1.44704357e-01 -9.73625705e-02 -1.03234708e+00
2.45731696e-01 -1.47693992e+00 5.55141419e-02 -1.44372582e+00
-1.58812904e+00 8.92143130e-01 -4.62717593e-01 -1.76790690e+00
-1.30614534e-01 -3.19108188e-01 -4.04853791e-01 1.06980729e+00
-1.70542920e+00 -8.20930660e-01 -7.07149386e-01 7.37200677e-01
4.84562665e-01 -1.39183700e-01 6.90243185e-01 5.06018341e-01
-1.49535730e-01 8.10994744e-01 3.71339694e-02 5.29208556e-02
1.07149577e+00 -1.41663647e+00 2.93038666e-01 1.27180612e+00
2.12315187e-01 2.23951653e-01 9.54278588e-01 -3.54941517e-01
-9.68871057e-01 -1.03620481e+00 6.33392334e-01 -3.81337762e-01
2.21574485e-01 3.23065728e-01 -9.08386350e-01 3.29633662e-03
3.77951831e-01 -1.91002376e-02 4.80305791e-01 -1.55615419e-01
-5.57951570e-01 -3.85986418e-01 -1.36164212e+00 4.48416442e-01
9.41930890e-01 -8.30290020e-01 -5.75397789e-01 6.21877462e-02
7.94951081e-01 -3.28440756e-01 -8.80474985e-01 3.77650380e-01
4.00387645e-01 -1.27806723e+00 1.12927496e+00 1.77226309e-02
6.91886127e-01 -7.49980450e-01 -2.82251418e-01 -1.73775077e+00
-1.88267171e-01 -3.95723104e-01 -1.86065912e-01 1.27224326e+00
3.70252967e-01 -1.69847205e-01 3.59475017e-01 2.10212082e-01
-8.97398219e-02 -3.23069662e-01 -6.04546130e-01 -9.22072649e-01
-1.26594231e-01 -5.38594306e-01 7.44551897e-01 5.51263392e-01
-2.84472883e-01 1.10793665e-01 -4.95168060e-01 4.19335902e-01
9.47900891e-01 -1.99937131e-02 6.22328758e-01 -9.88638520e-01
-3.58505964e-01 -4.82301027e-01 -5.78415096e-01 -1.09566474e+00
-3.51134002e-01 -5.43744087e-01 3.21906090e-01 -2.06418848e+00
2.61710048e-01 -2.31582299e-02 -5.47779024e-01 8.13399777e-02
-3.58864307e-01 5.70179760e-01 1.51027814e-01 2.49172151e-01
-8.97316039e-01 7.14731395e-01 1.76459420e+00 -5.19655704e-01
-3.58715802e-02 1.26734659e-01 -9.52370346e-01 5.16456246e-01
5.97273290e-01 -1.58049405e-01 -5.76546371e-01 -6.78771079e-01
2.19132900e-01 1.14643246e-01 6.62124813e-01 -1.34273887e+00
3.99647951e-01 8.50725621e-02 6.31950855e-01 -4.91168141e-01
3.19877177e-01 -5.09949923e-01 -8.79491940e-02 3.08691353e-01
-4.20097589e-01 -3.70428897e-02 -2.07686990e-01 4.82103348e-01
-5.92510223e-01 -8.51113349e-02 1.06846583e+00 -2.25619748e-02
-7.62822986e-01 1.31808758e-01 -4.30649631e-02 -1.64636932e-02
5.74159861e-01 -1.18024461e-01 -4.78572369e-01 -9.51008081e-01
-7.67860830e-01 5.18838279e-02 4.55847234e-01 5.52700758e-01
1.20033669e+00 -1.40952635e+00 -1.06062591e+00 -4.25433964e-02
2.90003866e-01 -3.96342456e-01 5.16108453e-01 4.99562740e-01
-2.87643433e-01 -2.45884676e-02 -2.70841330e-01 -5.46513617e-01
-1.20654535e+00 5.37793577e-01 3.90696675e-01 -1.37949064e-01
-2.67104447e-01 8.59523058e-01 -2.24389751e-02 -2.64571875e-01
5.39488196e-01 -9.15909186e-02 -2.91192710e-01 -2.44626731e-01
9.84630823e-01 5.05183935e-01 1.31801546e-01 -7.09768236e-01
5.68882143e-03 4.03451443e-01 5.16113117e-02 -4.70256597e-01
1.01614523e+00 -5.15042543e-01 2.56577283e-02 1.26528621e-01
8.79312575e-01 -4.31284569e-02 -1.30243075e+00 -4.19883728e-01
-3.82808030e-01 -7.94221461e-01 4.15723085e-01 -1.51157463e+00
-1.24273038e+00 9.54870522e-01 1.45150125e+00 1.18891507e-01
1.71803021e+00 -1.01841770e-01 8.58291805e-01 9.96640921e-02
3.96061838e-01 -1.02005279e+00 6.57112300e-01 7.49117583e-02
1.00850666e+00 -1.47015250e+00 -1.87497124e-01 -1.17621124e-01
-6.59918368e-01 5.64286888e-01 6.61641121e-01 2.62283921e-01
2.25529552e-01 -2.05416650e-01 6.46602511e-01 2.35471472e-01
-5.31342983e-01 -3.36497009e-01 7.05738187e-01 1.16201770e+00
2.77918339e-01 9.97764170e-02 1.52755529e-01 5.18828392e-01
-6.91209257e-01 -2.21178159e-02 5.51663518e-01 3.99455547e-01
-4.31850672e-01 -1.10665083e+00 -6.46350086e-01 2.09067568e-01
-3.95287037e-01 -2.41738170e-01 -1.87988449e-02 1.93960629e-02
2.51973808e-01 1.72420633e+00 -2.86920369e-01 -5.70244193e-01
5.55996180e-01 -4.01981473e-01 6.63744807e-01 -3.42444450e-01
-2.39109337e-01 -5.58823608e-02 -2.49973997e-01 -5.13330936e-01
-6.50037348e-01 -3.39721888e-01 -7.56060243e-01 -3.97450089e-01
-1.47431955e-01 1.59181863e-01 7.02812314e-01 7.14668095e-01
4.18832861e-02 6.98971570e-01 8.40257466e-01 -4.90038604e-01
-5.39994478e-01 -1.03166890e+00 -5.42906582e-01 4.58163977e-01
5.01721919e-01 -3.40533078e-01 -4.22218353e-01 4.04019743e-01]
|
[11.906563758850098, -1.8558233976364136]
|
8745a33f-e355-449a-a255-0ee2a465df3a
|
noisetrans-point-cloud-denoising-with
|
2304.11812
| null |
https://arxiv.org/abs/2304.11812v1
|
https://arxiv.org/pdf/2304.11812v1.pdf
|
NoiseTrans: Point Cloud Denoising with Transformers
|
Point clouds obtained from capture devices or 3D reconstruction techniques are often noisy and interfere with downstream tasks. The paper aims to recover the underlying surface of noisy point clouds. We design a novel model, NoiseTrans, which uses transformer encoder architecture for point cloud denoising. Specifically, we obtain structural similarity of point-based point clouds with the assistance of the transformer's core self-attention mechanism. By expressing the noisy point cloud as a set of unordered vectors, we convert point clouds into point embeddings and employ Transformer to generate clean point clouds. To make the Transformer preserve details when sensing the point cloud, we design the Local Point Attention to prevent the point cloud from being over-smooth. In addition, we also propose sparse encoding, which enables the Transformer to better perceive the structural relationships of the point cloud and improve the denoising performance. Experiments show that our model outperforms state-of-the-art methods in various datasets and noise environments.
|
['Zhonghan Zhang', 'Jie Yan', 'Yanhua Liang', 'Minghui Sun', 'Guihe Qin', 'Guangzhe Hou']
|
2023-04-24
| null | null | null | null |
['3d-reconstruction']
|
['computer-vision']
|
[-5.07780723e-02 -9.99956653e-02 4.69699353e-01 -2.37896487e-01
-7.90162325e-01 -4.82712954e-01 3.61581922e-01 -2.25188881e-01
1.46792322e-01 5.49776703e-02 3.93840820e-01 2.23408207e-01
1.47690147e-01 -8.36656511e-01 -1.31212842e+00 -6.88680410e-01
2.68846214e-01 6.01237565e-02 -1.06275722e-01 -1.66312754e-01
1.99622184e-01 8.15966547e-01 -1.53342915e+00 2.89723158e-01
7.38192737e-01 1.08985949e+00 4.67744708e-01 2.97898769e-01
-3.88596952e-01 4.41431791e-01 -5.87053418e-01 -3.42696041e-01
5.52951872e-01 3.45426291e-01 -1.31486729e-01 2.81077981e-01
5.64249575e-01 -4.59815532e-01 -8.59658778e-01 1.49143124e+00
3.34803194e-01 -2.42247246e-02 4.11671311e-01 -9.96223152e-01
-1.24773467e+00 5.04345149e-02 -5.33061504e-01 -1.31340459e-01
1.18961394e-01 3.40603232e-01 6.90841377e-01 -1.21021366e+00
2.82991260e-01 1.60984004e+00 1.07431066e+00 4.26932067e-01
-1.10319197e+00 -7.87630737e-01 3.74569207e-01 -3.84967253e-02
-1.51791739e+00 -4.80803043e-01 1.31223178e+00 -2.68956274e-01
8.18615437e-01 2.99153864e-01 6.44901514e-01 1.15258849e+00
-1.44701481e-01 5.96948683e-01 4.81955796e-01 2.45728433e-01
1.77966163e-01 -2.52550155e-01 -2.17121914e-01 3.12668741e-01
2.26599723e-01 1.12509020e-01 -3.01386058e-01 -3.36095870e-01
1.09694970e+00 8.65414441e-01 -4.89623934e-01 -2.95483589e-01
-1.18490255e+00 5.21728575e-01 8.96773398e-01 1.34614468e-01
-7.03191102e-01 5.22019267e-01 1.97047487e-01 3.46842438e-01
5.79170346e-01 6.12185150e-02 -1.64610147e-01 7.80311003e-02
-6.94545150e-01 1.87768042e-01 2.57012427e-01 1.50293553e+00
7.86126614e-01 2.19176725e-01 4.73151691e-02 5.81138015e-01
6.24032259e-01 8.22682619e-01 1.13738447e-01 -1.30406547e+00
6.81081474e-01 6.56143486e-01 5.92079833e-02 -1.39105940e+00
3.16781461e-01 -5.14095128e-01 -1.40556586e+00 3.36405039e-01
-4.15511757e-01 3.02469820e-01 -9.50017691e-01 1.35715353e+00
7.65274018e-02 1.04542661e+00 -5.66853303e-03 1.05522358e+00
8.63194704e-01 9.27323759e-01 -3.14405203e-01 1.40188783e-01
8.75525415e-01 -4.11572605e-01 -5.99599659e-01 -2.11582035e-02
1.06208071e-01 -6.22900605e-01 9.16458666e-01 3.23336661e-01
-1.17686462e+00 -7.47240365e-01 -9.84597504e-01 -6.03492498e-01
-3.52413505e-02 -1.55025452e-01 1.22534230e-01 -1.04222432e-01
-8.29633534e-01 8.88757765e-01 -1.27377212e+00 7.95344636e-02
7.27313817e-01 1.58484727e-01 -3.84913862e-01 -4.93572652e-01
-5.88844478e-01 2.44576469e-01 -4.53531653e-01 5.00619769e-01
-9.04546320e-01 -1.00729334e+00 -1.05659246e+00 2.98255593e-01
-1.04651049e-01 -1.00714147e+00 7.82160342e-01 -5.51353753e-01
-1.03344631e+00 5.97599149e-01 -3.77362907e-01 -1.83533356e-01
2.93959916e-01 -3.74317676e-01 -2.12222353e-01 7.00433925e-03
3.50655705e-01 4.25481111e-01 1.17497790e+00 -1.83963621e+00
-3.22537661e-01 -6.55282557e-01 -2.20420793e-01 1.82137057e-01
6.35429844e-02 -5.36249161e-01 -7.89268613e-01 -8.16218615e-01
7.05879390e-01 -5.50899923e-01 -3.76991838e-01 4.05022323e-01
-4.71584260e-01 6.31544665e-02 1.17024887e+00 -5.46955109e-01
7.05424666e-01 -2.72507095e+00 2.66999424e-01 3.16123992e-01
5.97808301e-01 -3.60358059e-02 -3.80012482e-01 1.34470627e-01
-2.79364765e-01 1.89981312e-01 -3.02184016e-01 -9.47263360e-01
-2.42615547e-02 6.61533713e-01 -7.34026253e-01 5.78952610e-01
4.89218950e-01 8.83375943e-01 -8.38946462e-01 1.38639256e-01
5.78626335e-01 9.47616041e-01 -5.67792356e-01 2.56913126e-01
-9.89283994e-03 4.08885449e-01 -7.24452615e-01 8.07019770e-01
1.28076994e+00 -1.26360491e-01 -7.68950284e-01 -7.21904218e-01
-1.04891425e-02 2.08152846e-01 -1.10338044e+00 2.19164419e+00
-4.71880406e-01 2.40234450e-01 5.00506461e-01 -7.34254062e-01
1.10560441e+00 1.71201795e-01 5.80914795e-01 -4.31050569e-01
-5.01584774e-03 -3.45207402e-03 -5.06494701e-01 -3.34564298e-01
3.14149618e-01 -6.47162274e-02 2.18654573e-01 -3.57443273e-01
-1.60892293e-01 -5.56372106e-01 -8.59895229e-01 1.64180472e-01
1.32093465e+00 -6.02340512e-02 -4.62366372e-01 2.08518758e-01
4.27517682e-01 -2.56960839e-01 7.17642307e-01 5.97703993e-01
6.90573752e-02 1.25546825e+00 9.90545303e-02 -3.39372993e-01
-1.21639216e+00 -1.37949479e+00 9.24696475e-02 2.05144867e-01
3.84971023e-01 -2.95397311e-01 -3.45095515e-01 -3.87865573e-01
3.86477351e-01 5.20000935e-01 -4.92581725e-01 -4.07973707e-01
-5.60505033e-01 -7.38686621e-02 1.93897232e-01 6.16476953e-01
3.21196049e-01 -5.14971077e-01 1.50204837e-01 1.68821022e-01
-1.44863561e-01 -1.00353289e+00 -6.19086802e-01 3.39016281e-02
-1.19611835e+00 -8.19662273e-01 -5.00608563e-01 -7.92247951e-01
8.03490877e-01 8.25427830e-01 1.21151292e+00 2.64690697e-01
3.06330919e-01 3.65866840e-01 -4.14910138e-01 -4.49958831e-01
4.20934968e-02 -3.84616822e-01 1.46015152e-01 3.11933141e-02
5.50141692e-01 -1.27449775e+00 -5.87244034e-01 5.84108531e-02
-1.04303300e+00 -2.86742270e-01 5.23420811e-01 8.08744252e-01
1.21913946e+00 1.36589214e-01 -1.11154892e-01 -4.60240275e-01
3.83692831e-01 -7.03813612e-01 -4.68312502e-01 -3.50559771e-01
-5.25818439e-03 5.46463877e-02 6.01294935e-01 -2.04455256e-01
-4.91551757e-01 2.56321341e-01 -4.85107750e-01 -1.68175638e+00
-1.18865617e-01 1.61757171e-01 -4.98760641e-01 -2.61262625e-01
2.75600642e-01 4.30680722e-01 8.16011801e-02 -1.08807707e+00
2.95627505e-01 5.41218519e-01 9.33997393e-01 -4.47461575e-01
1.28421462e+00 9.66668963e-01 -3.32060531e-02 -9.19188738e-01
-6.91749990e-01 -6.22770190e-01 -3.71555984e-01 1.48293808e-01
7.02552080e-01 -1.44807136e+00 -6.63495362e-01 2.43870124e-01
-1.62967229e+00 2.39806741e-01 -4.86853898e-01 8.94745588e-02
-3.87419403e-01 4.88123238e-01 -6.32577658e-01 -6.02697432e-01
-3.55074674e-01 -1.23218036e+00 1.87764812e+00 -1.47145346e-01
3.45636040e-01 -4.65790838e-01 -1.52825519e-01 -1.16323590e-01
8.66245255e-02 2.24857315e-01 4.67741221e-01 4.19426076e-02
-1.19552696e+00 -1.76569760e-01 -2.99395204e-01 8.13744426e-01
2.42016628e-01 -1.73541948e-01 -1.06801927e+00 -2.19911531e-01
6.28983617e-01 2.81043798e-01 8.18109512e-01 4.02235955e-01
1.66540301e+00 -2.64759362e-01 -3.08053523e-01 1.45348763e+00
1.47353733e+00 -2.57870883e-01 7.72711277e-01 6.72535375e-02
1.29882324e+00 2.15803131e-01 2.51557261e-01 3.48784924e-01
2.43779019e-01 2.79716849e-01 1.10423303e+00 2.61745248e-02
-4.15005498e-02 -6.58069611e-01 2.39239305e-01 1.26803851e+00
3.12695317e-02 9.47539136e-02 -5.95465899e-01 6.90769553e-01
-1.68892157e+00 -8.28541577e-01 -2.07068995e-01 1.80867434e+00
2.44726554e-01 -3.65200005e-02 -6.71844542e-01 8.88486058e-02
5.87622285e-01 2.44297788e-01 -7.53122628e-01 1.71741188e-01
-1.97129279e-01 2.58961558e-01 5.69669783e-01 4.37901258e-01
-8.84342968e-01 5.10935605e-01 5.50820923e+00 5.95015526e-01
-1.05997515e+00 6.70841038e-02 2.28231132e-01 -2.60750353e-01
-7.18841732e-01 -3.51574838e-01 -3.16106766e-01 8.11345577e-01
4.31196898e-01 -5.83767258e-02 4.57544088e-01 8.57019007e-01
2.72725523e-01 8.09743404e-01 -1.08810699e+00 1.57217455e+00
-1.36927888e-01 -1.50777233e+00 2.62084246e-01 8.28801692e-02
5.36493957e-01 5.24269164e-01 1.46009058e-01 1.42688394e-01
1.83316424e-01 -8.63110244e-01 7.26326585e-01 9.41940367e-01
7.19209373e-01 -7.27368951e-01 6.96525931e-01 4.48651582e-01
-1.13489926e+00 1.03753902e-01 -9.31280136e-01 9.14132446e-02
1.74135074e-01 9.44505215e-01 -4.84791398e-02 6.14567935e-01
1.18330276e+00 1.46072471e+00 -2.85650939e-01 9.99395788e-01
1.12700919e-02 4.35132504e-01 -6.97106123e-01 5.78704536e-01
3.12737048e-01 -5.69348931e-01 9.31658566e-01 6.99333489e-01
8.94083738e-01 3.14931214e-01 1.65182091e-02 1.27513635e+00
-2.37788573e-01 -5.34211338e-01 -1.04114652e+00 2.07155928e-01
7.21038282e-01 8.54331017e-01 2.23459721e-01 -1.54221252e-01
-5.59602559e-01 1.22299802e+00 1.68295830e-01 4.83511060e-01
-4.08628762e-01 -3.17351967e-01 1.58810735e+00 1.14654176e-01
5.80289543e-01 -3.40602010e-01 -6.48719370e-01 -1.30248809e+00
5.04893959e-01 -5.88448107e-01 -3.73018146e-01 -1.30527616e+00
-1.82266653e+00 6.37666345e-01 -5.14982462e-01 -1.73049557e+00
3.66732150e-01 -3.82020861e-01 -7.54555047e-01 1.12930262e+00
-1.56085718e+00 -1.07291532e+00 -5.93835413e-01 6.30171180e-01
3.93584847e-01 9.48419720e-02 5.32994986e-01 4.92367148e-01
-2.32869834e-01 7.85962716e-02 4.39176977e-01 7.78081566e-02
2.80323595e-01 -1.05256236e+00 1.04695117e+00 8.66792142e-01
-8.13888758e-02 9.33312297e-01 5.43062568e-01 -7.45019019e-01
-1.81754076e+00 -1.60789561e+00 3.22407484e-01 -5.69251299e-01
4.62128490e-01 -5.74521422e-01 -1.36400306e+00 7.27166891e-01
-2.94085503e-01 3.86190534e-01 8.40172321e-02 -3.05169374e-01
-4.63827640e-01 -3.00290316e-01 -1.24946487e+00 4.37501848e-01
1.43194389e+00 -7.68979430e-01 -8.32485616e-01 3.48196834e-01
1.30157936e+00 -5.97770512e-01 -8.62625062e-01 3.61907065e-01
-8.18759501e-02 -7.84941792e-01 1.46605003e+00 -2.92779237e-01
5.29864490e-01 -5.98452926e-01 -5.11836648e-01 -1.54023254e+00
-6.93628073e-01 -4.87792552e-01 -3.81343782e-01 1.24128449e+00
-1.44814000e-01 -3.33729148e-01 9.97475922e-01 3.01428139e-01
-6.93681657e-01 -5.45776367e-01 -9.94180679e-01 -5.70412099e-01
3.54236737e-02 -6.60047174e-01 1.12147629e+00 9.36412752e-01
-8.37737381e-01 9.76773426e-02 -2.61675119e-01 9.40743744e-01
9.77078736e-01 1.41049745e-02 7.13651478e-01 -1.18038619e+00
8.22651088e-02 -2.08155036e-01 -8.57164919e-01 -1.54180324e+00
2.66051650e-01 -6.90295994e-01 2.27441296e-01 -1.38580096e+00
-3.87512922e-01 -5.22168279e-01 -1.47145554e-01 1.16962537e-01
-6.71763048e-02 1.75285414e-01 2.31970370e-01 3.59617531e-01
-1.53719708e-01 9.94772613e-01 1.22706127e+00 -4.96749640e-01
7.89716914e-02 3.43876029e-03 -9.10681367e-01 7.42981613e-01
3.42186987e-01 -5.36533594e-01 -2.38530129e-01 -1.40693593e+00
3.18038575e-02 -2.01090246e-01 6.56973720e-01 -1.14000070e+00
2.61122584e-01 7.25130886e-02 5.24374962e-01 -9.54275966e-01
6.92505121e-01 -1.54607689e+00 3.58349800e-01 -3.31945643e-02
9.39734504e-02 2.52077341e-01 9.63094980e-02 9.95242953e-01
-4.24966156e-01 1.68444708e-01 5.23017049e-01 -2.42288545e-01
-4.20432776e-01 9.21018720e-01 2.68247575e-01 -3.63487780e-01
5.93238354e-01 -2.19441324e-01 5.56958057e-02 -2.31259346e-01
-5.99668086e-01 3.78429353e-01 8.69570911e-01 4.91176516e-01
1.10854161e+00 -1.69906723e+00 -5.42497933e-01 7.79754460e-01
2.05395371e-01 9.75995243e-01 3.57833833e-01 3.82167876e-01
-6.27424896e-01 -3.76345180e-02 1.45047531e-01 -9.23273504e-01
-8.70369971e-01 6.88922048e-01 2.94967860e-01 4.50111181e-01
-1.22181213e+00 9.75010455e-01 2.74362445e-01 -6.42771959e-01
3.69211107e-01 -9.04158831e-01 3.87801230e-01 -6.19747579e-01
5.97209334e-01 1.86558679e-01 2.58857757e-01 -7.32929826e-01
-2.40459904e-01 8.82936537e-01 2.38895923e-01 3.50697279e-01
1.66814399e+00 -3.03851336e-01 -2.90517300e-01 3.40372443e-01
1.63920128e+00 5.90642355e-02 -1.45126438e+00 -3.78737509e-01
-5.05268812e-01 -8.88475299e-01 5.31872869e-01 -6.88866675e-02
-1.36322117e+00 9.76220608e-01 5.97932339e-01 3.14050913e-02
1.12720764e+00 -9.91139859e-02 1.06140876e+00 4.71333772e-01
3.81938875e-01 -3.90918940e-01 -4.16409701e-01 5.25310218e-01
1.17051816e+00 -1.22207403e+00 -2.18546018e-01 -5.98213971e-01
-2.84482211e-01 7.00822055e-01 3.65177423e-01 -7.89661229e-01
8.53884101e-01 4.40808892e-01 -1.10478373e-02 -4.26016510e-01
-6.80341601e-01 -4.98240627e-02 1.32781029e-01 9.91428494e-01
-2.28117868e-01 -8.73169452e-02 6.15804970e-01 7.79506445e-01
-2.88538992e-01 5.24646826e-02 1.30382672e-01 7.70600319e-01
-3.84474486e-01 -7.94098556e-01 -6.10107958e-01 5.23945987e-01
-6.76884204e-02 -2.10073158e-01 -2.98584968e-01 1.88362762e-01
2.92141050e-01 6.12892568e-01 3.91684383e-01 -7.15215623e-01
9.33080912e-01 -3.36513221e-01 1.25685439e-01 -7.30421364e-01
-3.70186985e-01 9.01396647e-02 -6.45008981e-01 -9.14347172e-01
-1.69226363e-01 -6.10014439e-01 -9.61598456e-01 -5.77328742e-01
6.23028316e-02 1.97791547e-01 6.13818049e-01 5.27465403e-01
8.22547555e-01 5.41009426e-01 8.17325473e-01 -1.33545232e+00
-5.21332979e-01 -7.74059236e-01 -7.29173362e-01 7.68263221e-01
9.91303146e-01 -5.49506307e-01 -6.42623305e-01 -4.90747578e-02]
|
[8.215133666992188, -3.59460186958313]
|
d6e00d47-9854-4d07-90f7-c0132557284a
|
endnet-sparse-autoencoder-network-for
|
1708.01894
| null |
http://arxiv.org/abs/1708.01894v4
|
http://arxiv.org/pdf/1708.01894v4.pdf
|
EndNet: Sparse AutoEncoder Network for Endmember Extraction and Hyperspectral Unmixing
|
Data acquired from multi-channel sensors is a highly valuable asset to
interpret the environment for a variety of remote sensing applications.
However, low spatial resolution is a critical limitation for previous sensors
and the constituent materials of a scene can be mixed in different fractions
due to their spatial interactions. Spectral unmixing is a technique that allows
us to obtain the material spectral signatures and their fractions from
hyperspectral data. In this paper, we propose a novel endmember extraction and
hyperspectral unmixing scheme, so called \textit{EndNet}, that is based on a
two-staged autoencoder network. This well-known structure is completely
enhanced and restructured by introducing additional layers and a projection
metric (i.e., spectral angle distance (SAD) instead of inner product) to
achieve an optimum solution. Moreover, we present a novel loss function that is
composed of a Kullback-Leibler divergence term with SAD similarity and
additional penalty terms to improve the sparsity of the estimates. These
modifications enable us to set the common properties of endmembers such as
non-linearity and sparsity for autoencoder networks. Lastly, due to the
stochastic-gradient based approach, the method is scalable for large-scale data
and it can be accelerated on Graphical Processing Units (GPUs). To demonstrate
the superiority of our proposed method, we conduct extensive experiments on
several well-known datasets. The results confirm that the proposed method
considerably improves the performance compared to the state-of-the-art
techniques in literature.
|
['Gozde Bozdagi Akar', 'Savas Ozkan', 'Berk Kaya']
|
2017-08-06
| null | null | null | null |
['hyperspectral-unmixing']
|
['computer-vision']
|
[ 4.91149992e-01 -7.37830162e-01 1.92152545e-01 -2.66156346e-01
-2.89682359e-01 -3.08840930e-01 3.74737263e-01 -9.06997360e-03
-5.22103131e-01 7.98609376e-01 -1.48547180e-02 -1.29039556e-01
-4.47983682e-01 -8.37836742e-01 -6.49414539e-01 -1.21735060e+00
2.38866284e-02 -1.54705480e-01 -8.77849758e-02 -1.78294957e-01
1.72025096e-02 6.11629069e-01 -1.82687700e+00 1.54445972e-02
1.33458853e+00 1.12335742e+00 5.67573547e-01 1.35929227e-01
-4.41881716e-02 5.27560294e-01 -1.60064578e-01 1.35656387e-01
5.57177126e-01 -1.98235735e-01 -2.27428883e-01 3.59915495e-01
3.38470012e-01 -4.41603035e-01 -1.84738860e-01 1.43827891e+00
4.47783530e-01 4.75684613e-01 6.11325324e-01 -8.14163506e-01
-3.86918247e-01 4.76854205e-01 -8.76558602e-01 -9.67838094e-02
-2.97530681e-01 -1.32303104e-01 8.39636683e-01 -9.01594877e-01
-2.43067779e-02 9.56398606e-01 7.46068418e-01 -1.17521830e-01
-1.00974643e+00 -6.24993861e-01 9.02877748e-02 2.94659644e-01
-1.63559783e+00 -2.15808302e-01 9.55210388e-01 -4.59150106e-01
4.60440338e-01 4.59863096e-01 6.97924316e-01 4.56193119e-01
-1.42749682e-01 4.43022788e-01 1.43715096e+00 -3.64651412e-01
1.59494177e-01 1.26250029e-01 1.19631641e-01 4.95669186e-01
5.24838507e-01 1.03648573e-01 -1.17625423e-01 -7.60820787e-03
7.51515627e-01 5.44191539e-01 -6.94400847e-01 -2.68376112e-01
-1.05929005e+00 7.91620851e-01 7.84041047e-01 2.86061198e-01
-6.53918087e-01 -3.54014128e-01 2.21269876e-02 -1.34114474e-01
4.40972179e-01 6.63535297e-02 -2.47808576e-01 4.56332713e-01
-9.58550215e-01 -1.93946213e-02 6.08288765e-01 4.32219058e-01
1.23196876e+00 3.37925196e-01 1.73021466e-01 1.16043365e+00
5.16660511e-01 7.90986001e-01 6.78512633e-01 -4.11368877e-01
3.53566408e-01 7.09906936e-01 1.41915843e-01 -1.18954110e+00
-5.06607056e-01 -7.45472074e-01 -1.31557441e+00 2.63080835e-01
-2.09573116e-02 -2.16152146e-01 -8.27059090e-01 1.47104979e+00
3.99091065e-01 3.65659982e-01 2.98734337e-01 1.26548386e+00
8.60768318e-01 1.06402290e+00 1.88869312e-02 -4.18474853e-01
1.07894683e+00 -8.87555778e-01 -7.87364304e-01 -3.31694901e-01
6.69070631e-02 -6.52841985e-01 8.36178362e-01 3.98131967e-01
-4.91679937e-01 -5.87212801e-01 -1.42188334e+00 2.98863709e-01
-4.50530410e-01 4.99586225e-01 7.62471735e-01 5.39214790e-01
-5.34880400e-01 7.59933770e-01 -8.65899205e-01 -7.98867866e-02
1.92690298e-01 1.18525639e-01 -3.06391269e-01 1.72089599e-02
-1.05207133e+00 5.31171143e-01 9.54487979e-01 5.24510741e-01
-6.31124616e-01 -4.50109899e-01 -7.24925339e-01 1.15515612e-01
2.95299083e-01 -2.15402126e-01 5.82401037e-01 -1.21141744e+00
-1.56041253e+00 2.63118804e-01 1.36287492e-02 -2.13437602e-02
1.52611643e-01 -2.14798912e-01 -7.42375731e-01 1.35458842e-01
-2.39037052e-01 8.93420279e-02 9.95230019e-01 -1.37152195e+00
-5.75784326e-01 -5.19372463e-01 -1.46111295e-01 4.97176319e-01
-8.97600055e-01 -2.34007850e-01 -5.67664988e-02 -6.91689730e-01
6.36692762e-01 -8.28583539e-01 -2.65427619e-01 -1.75456524e-01
-3.42748225e-01 3.88485640e-01 7.87898719e-01 -9.48118508e-01
1.12356544e+00 -2.29021525e+00 1.15294963e-01 4.12432671e-01
3.87411378e-02 5.37415564e-01 1.20617170e-02 4.17546451e-01
-2.58550823e-01 -2.08417818e-01 -8.83960843e-01 -1.67030171e-02
-2.76712000e-01 5.17818006e-03 -6.28935844e-02 6.21424615e-01
-9.47084557e-03 2.68792003e-01 -7.53711343e-01 -1.42905116e-01
5.01766741e-01 7.71929502e-01 -6.21253848e-02 1.84236422e-01
3.83756086e-02 4.16481704e-01 -2.34641686e-01 6.00647151e-01
1.28022516e+00 -2.23956168e-01 1.26543954e-01 -6.19456589e-01
-4.85012591e-01 -1.94805115e-01 -1.75554156e+00 1.48400486e+00
-3.64679128e-01 1.70782387e-01 5.01610279e-01 -1.19763625e+00
1.00818586e+00 1.02319956e-01 3.58350426e-01 -2.20672011e-01
1.84845522e-01 4.51931059e-01 -1.76873520e-01 -5.17464101e-01
5.44335783e-01 -2.93249935e-01 6.31499648e-01 2.19027907e-01
-2.68751264e-01 8.85484219e-02 8.63583460e-02 -4.55109239e-01
3.11362535e-01 -6.62026703e-02 4.04735148e-01 -4.27916974e-01
8.63321483e-01 -1.12364434e-01 6.27331257e-01 4.69852805e-01
1.43042952e-01 3.38561773e-01 -3.68457526e-01 -3.51771146e-01
-7.85782993e-01 -8.46030712e-01 -4.23651963e-01 7.47966588e-01
3.29424500e-01 2.08026484e-01 -4.99868035e-01 -2.16548428e-01
-5.12001440e-02 4.02731061e-01 -1.82390839e-01 6.05920926e-02
-2.28363454e-01 -1.50422549e+00 2.11266115e-01 3.78491551e-01
1.02395272e+00 -6.50711000e-01 -4.08906281e-01 2.16501638e-01
-1.70921728e-01 -9.34317827e-01 -5.04330173e-02 1.42610431e-01
-1.19447112e+00 -9.28926229e-01 -6.65460885e-01 -7.06723928e-01
6.64826214e-01 7.53128290e-01 4.84072208e-01 -2.64290333e-01
-6.40561879e-02 -2.26122960e-01 -4.80368763e-01 -4.87152308e-01
-1.42771468e-01 -6.25997633e-02 7.83946291e-02 6.41669214e-01
2.72257060e-01 -9.47075784e-01 -6.89507842e-01 1.75120812e-02
-1.49531639e+00 1.50085300e-01 9.74651456e-01 9.18503582e-01
6.80987775e-01 6.71416163e-01 2.83284187e-01 -6.58475876e-01
4.36003536e-01 -5.22794425e-01 -8.27714026e-01 1.93558872e-01
-5.94927728e-01 -1.05362877e-01 8.09479892e-01 -3.42184484e-01
-1.32060564e+00 2.63268173e-01 1.37766367e-02 -3.54676604e-01
-2.29566693e-01 9.70779598e-01 -2.94285953e-01 -3.66209537e-01
5.42566061e-01 5.88383734e-01 7.77441487e-02 -6.87274396e-01
1.65694013e-01 1.02826881e+00 5.07405877e-01 -2.65530258e-01
9.31112409e-01 7.70715415e-01 9.39834863e-02 -1.21349394e+00
-7.18552351e-01 -6.21818542e-01 -3.76801878e-01 -8.33872855e-02
5.53415835e-01 -1.19838393e+00 -5.59936821e-01 8.18217695e-01
-8.73787224e-01 7.64862224e-02 1.36859655e-01 9.50396657e-01
-2.34967154e-02 6.56473637e-01 -4.82673436e-01 -1.01646876e+00
-5.96404016e-01 -9.53635991e-01 5.83031178e-01 4.76472050e-01
7.98435330e-01 -8.40964973e-01 -9.26010013e-02 2.19780236e-01
4.40538198e-01 3.29712301e-01 6.98018432e-01 -3.50492835e-01
-4.55606639e-01 -7.93355610e-03 -4.94347811e-01 9.77067947e-01
4.12816972e-01 -7.49627650e-02 -9.84565020e-01 -4.19495463e-01
3.48891407e-01 -4.16648239e-02 1.11265326e+00 4.22731459e-01
1.28804362e+00 -3.86225045e-01 -3.97100523e-02 8.81357372e-01
2.06297445e+00 2.48422652e-01 5.79298377e-01 3.65536898e-01
9.62356269e-01 5.22383273e-01 3.88401389e-01 7.22792208e-01
-1.60396492e-04 1.85494885e-01 7.00814962e-01 -2.48051181e-01
3.80444080e-01 1.20028459e-01 1.65595755e-01 1.08213687e+00
-3.94811809e-01 -9.75175053e-02 -6.41995907e-01 3.83372515e-01
-1.89746094e+00 -9.84586418e-01 -3.66336703e-01 2.37128282e+00
8.01016748e-01 -3.75278980e-01 -2.86684543e-01 4.95413750e-01
9.00516570e-01 5.14419436e-01 -4.91021693e-01 1.56513646e-01
-4.28138942e-01 1.53965726e-01 7.87530243e-01 4.44648445e-01
-1.42160392e+00 5.41509032e-01 5.27574492e+00 9.31116700e-01
-1.43531239e+00 1.19267151e-01 2.30667859e-01 3.65287185e-01
-1.93050325e-01 -4.52488437e-02 -5.77444494e-01 5.05987704e-01
4.79668856e-01 2.90172637e-01 7.65613317e-01 6.82730913e-01
3.31268013e-01 -1.43143535e-01 -4.70573217e-01 1.15230179e+00
1.71117216e-01 -9.94885921e-01 -5.55401202e-03 -1.90865800e-01
9.13514376e-01 9.41528603e-02 -2.99849827e-02 -1.65784210e-01
-3.78627405e-02 -7.00844228e-01 5.18221736e-01 6.01371586e-01
5.70769191e-01 -7.48911858e-01 7.71751404e-01 4.49338973e-01
-1.30236661e+00 -3.42719674e-01 -5.60312271e-01 -3.44993353e-01
-2.93174442e-02 1.24867809e+00 -4.53123122e-01 1.01457858e+00
5.60213804e-01 7.84976244e-01 -1.74252912e-01 1.14000535e+00
-2.32699946e-01 5.47304928e-01 -5.90158403e-01 -4.72465046e-02
2.77347714e-01 -8.49330962e-01 6.29060149e-01 1.04128408e+00
6.21304750e-01 2.64165103e-01 2.82740980e-01 8.30205917e-01
1.25436157e-01 4.54542965e-01 -2.79659063e-01 -1.92001164e-01
5.30760825e-01 1.54826796e+00 -3.46886486e-01 -2.33677849e-01
-4.88477141e-01 7.16814637e-01 -1.19421318e-01 4.76466626e-01
-6.28400922e-01 -5.07422924e-01 5.34524202e-01 -1.47567153e-01
2.99152732e-01 -2.66675740e-01 -1.08905740e-01 -1.32898080e+00
2.73939759e-01 -9.59802091e-01 2.04114884e-01 -6.71920776e-01
-1.30342329e+00 5.45748532e-01 -1.39280602e-01 -1.55571008e+00
3.42201084e-01 -8.22174489e-01 -3.82866502e-01 1.02867126e+00
-2.01110244e+00 -1.13734317e+00 -9.92784798e-01 6.31954551e-01
1.35564372e-01 -4.45131995e-02 7.59058893e-01 4.74101633e-01
-8.13508332e-01 2.35906169e-02 8.34137976e-01 -4.04178128e-02
4.42627162e-01 -9.47019696e-01 -4.89429265e-01 1.14684618e+00
-1.16507508e-01 4.43141490e-01 5.66176593e-01 -4.89166349e-01
-1.31237030e+00 -1.38480747e+00 2.33841792e-01 5.55103719e-01
5.82159460e-01 1.84153691e-01 -9.75152135e-01 4.81758654e-01
2.50657275e-02 -5.47270216e-02 8.61957133e-01 -1.60052642e-01
-1.74933910e-01 -6.96859896e-01 -9.99940574e-01 2.99952328e-01
5.65317810e-01 -3.65136027e-01 -3.97676826e-01 4.03003186e-01
4.23275381e-01 -2.79567063e-01 -8.27010632e-01 7.88221538e-01
4.99144703e-01 -1.08241117e+00 9.21484292e-01 6.97186077e-03
3.87601912e-01 -7.76108205e-01 -3.79171014e-01 -1.39721084e+00
-5.09107113e-01 -2.48180941e-01 -4.67486568e-02 1.03017914e+00
2.05257043e-01 -9.22523499e-01 4.97141153e-01 1.46220163e-01
-2.56284535e-01 -4.72680032e-01 -5.99572301e-01 -8.17754567e-01
-4.80511844e-01 -1.09595209e-01 7.73199081e-01 1.03562784e+00
-2.58372903e-01 5.90021461e-02 -5.83653510e-01 6.92011356e-01
8.12117636e-01 3.12617600e-01 4.12421882e-01 -1.46989810e+00
-4.16517675e-01 -3.08425009e-01 -8.21714923e-02 -9.98760760e-01
-7.60525987e-02 -7.48682678e-01 1.92704752e-01 -1.38993824e+00
2.19137460e-01 -4.38526690e-01 -6.75077438e-01 3.82354975e-01
-3.87727410e-01 1.32193550e-01 -7.16529191e-02 4.87254113e-01
1.15902148e-01 8.72880220e-01 1.05662870e+00 -3.15846920e-01
-5.40666759e-01 -6.13626018e-02 -6.38090491e-01 8.88946533e-01
8.56438398e-01 -2.39711836e-01 -3.55336726e-01 -7.05970883e-01
1.48321345e-01 -2.09131643e-01 3.74895930e-01 -1.38385701e+00
1.44044101e-01 -3.79783332e-01 4.09985840e-01 -6.25937700e-01
4.48466212e-01 -1.09823108e+00 4.00746107e-01 2.26613075e-01
1.76968098e-01 -4.09827083e-01 1.95111692e-01 4.74825412e-01
-4.58253920e-01 -4.31817502e-01 8.22714508e-01 -1.54809430e-01
-7.98198819e-01 4.47324097e-01 -6.98698312e-02 -5.94631612e-01
7.88569987e-01 -2.21515581e-01 -2.20613211e-01 -1.17794409e-01
-2.98969477e-01 -5.36876395e-02 2.57329911e-01 -3.33429389e-02
6.05008006e-01 -1.33811939e+00 -8.60499799e-01 2.61546642e-01
3.10411025e-02 7.44669586e-02 6.72184289e-01 8.91357243e-01
-7.83755004e-01 1.19660802e-01 -2.73706257e-01 -4.79550540e-01
-9.96749759e-01 3.13884377e-01 3.83253872e-01 -1.34122774e-01
-3.27121198e-01 6.07895494e-01 1.44119933e-01 -4.93392438e-01
-5.07027842e-02 -1.36990115e-01 -3.62675220e-01 4.02553529e-02
6.99438095e-01 5.98999858e-01 2.15194106e-01 -7.34949410e-01
-1.76078573e-01 5.77305257e-01 3.69355083e-01 -3.90644837e-03
1.60399103e+00 -5.17087728e-02 -6.63515687e-01 3.98058593e-01
1.15725279e+00 3.66380513e-02 -1.25314963e+00 -4.16928083e-01
-6.00334585e-01 -4.57778722e-01 4.97323662e-01 -5.90938449e-01
-1.16443598e+00 8.93263161e-01 9.10634995e-01 2.14157701e-01
1.56201613e+00 -6.83347404e-01 7.11753964e-01 6.18151486e-01
4.38995007e-03 -1.23410869e+00 -3.90993536e-01 3.52250397e-01
6.98434412e-01 -1.29530537e+00 2.67242968e-01 -5.16557992e-01
-1.87667951e-01 1.16735709e+00 4.14881200e-01 -6.06232276e-03
8.08181882e-01 -6.84437603e-02 1.14989847e-01 1.08196609e-01
2.15252906e-01 -3.09753507e-01 3.20112407e-01 2.55118519e-01
2.28904828e-01 3.10803354e-01 -4.98819619e-01 4.94356215e-01
7.15518892e-02 -1.30545795e-01 2.64023721e-01 7.47056067e-01
-7.58821368e-01 -8.05488706e-01 -7.85270393e-01 4.96455193e-01
-2.67569453e-01 -2.67244935e-01 2.17928827e-01 3.98692757e-01
3.40041518e-01 1.07422209e+00 -2.39020020e-01 -4.94235843e-01
7.96717703e-02 -8.00130740e-02 9.77791101e-02 -3.40482235e-01
-2.92463060e-02 2.69583791e-01 -3.38325024e-01 -1.76017135e-01
-9.41373944e-01 -4.75431263e-01 -1.00246322e+00 -1.45581931e-01
-6.48296356e-01 1.38177857e-01 1.01044381e+00 9.56024945e-01
1.47419706e-01 4.19721186e-01 8.67984414e-01 -1.04329538e+00
-7.23711073e-01 -1.22872090e+00 -1.08822834e+00 3.98244143e-01
3.30699027e-01 -6.13042474e-01 -5.13418019e-01 1.36009619e-01]
|
[10.142683982849121, -2.035994291305542]
|
d77fccae-50d1-4af4-b92a-49e99171d6ce
|
co-evolving-graph-reasoning-network-for
|
2306.04340
| null |
https://arxiv.org/abs/2306.04340v1
|
https://arxiv.org/pdf/2306.04340v1.pdf
|
Co-evolving Graph Reasoning Network for Emotion-Cause Pair Extraction
|
Emotion-Cause Pair Extraction (ECPE) aims to extract all emotion clauses and their corresponding cause clauses from a document. Existing approaches tackle this task through multi-task learning (MTL) framework in which the two subtasks provide indicative clues for ECPE. However, the previous MTL framework considers only one round of multi-task reasoning and ignores the reverse feedbacks from ECPE to the subtasks. Besides, its multi-task reasoning only relies on semantics-level interactions, which cannot capture the explicit dependencies, and both the encoder sharing and multi-task hidden states concatenations can hardly capture the causalities. To solve these issues, we first put forward a new MTL framework based on Co-evolving Reasoning. It (1) models the bidirectional feedbacks between ECPE and its subtasks; (2) allows the three tasks to evolve together and prompt each other recurrently; (3) integrates prediction-level interactions to capture explicit dependencies. Then we propose a novel multi-task relational graph (MRG) to sufficiently exploit the causal relations. Finally, we propose a Co-evolving Graph Reasoning Network (CGR-Net) that implements our MTL framework and conducts Co-evolving Reasoning on MRG. Experimental results show that our model achieves new state-of-the-art performance, and further analysis confirms the advantages of our method.
|
['Ivor W. Tsang', 'Bowen Xing']
|
2023-06-07
| null | null | null | null |
['emotion-cause-pair-extraction']
|
['natural-language-processing']
|
[-1.38716185e-02 3.85649592e-01 -3.17146689e-01 -3.49436611e-01
-8.44248533e-01 -3.09212476e-01 6.17023230e-01 -1.11366279e-01
3.87170091e-02 5.72353005e-01 5.14160991e-01 -1.53516874e-01
-2.23201007e-01 -6.06838822e-01 -7.18200684e-01 -3.95835310e-01
-1.51251614e-01 4.03314471e-01 3.72278631e-01 -4.27557170e-01
-1.75630450e-01 -3.12247396e-01 -1.15996873e+00 7.03823686e-01
1.02719808e+00 8.18307817e-01 1.53484613e-01 5.74002922e-01
-3.59725714e-01 1.77708769e+00 -4.12726074e-01 -8.04789543e-01
-4.23761666e-01 -5.31541169e-01 -1.12095785e+00 -2.55219400e-01
-4.00183499e-01 -1.52906537e-01 -2.63190538e-01 8.92131329e-01
2.74093449e-01 5.54837892e-03 3.75712216e-01 -1.57679319e+00
-8.98139894e-01 1.20717156e+00 -5.98155499e-01 1.12062935e-02
3.55701774e-01 -1.40038997e-01 1.62606430e+00 -7.14028239e-01
6.56227291e-01 1.31539333e+00 4.81295586e-01 5.36996782e-01
-8.09858859e-01 -8.06743383e-01 8.00876737e-01 5.20711303e-01
-1.08559585e+00 -1.17457487e-01 1.18933523e+00 -1.56087801e-01
1.46044743e+00 2.18336228e-02 7.41546988e-01 1.36634529e+00
3.72342259e-01 1.30951345e+00 9.69982386e-01 -1.93033040e-01
-9.77673978e-02 -1.96593419e-01 3.88110787e-01 8.72970760e-01
-2.80269533e-01 -2.29824334e-01 -7.15588093e-01 4.48266305e-02
3.78921360e-01 -1.55502215e-01 -1.98659271e-01 8.53086486e-02
-1.13491392e+00 6.68650210e-01 3.68228406e-01 6.06961370e-01
-3.00092995e-01 2.98363984e-01 6.49973512e-01 4.28757340e-01
5.30559361e-01 2.44464606e-01 -7.38810182e-01 -1.21604495e-01
-4.97559041e-01 4.65126745e-02 7.75309205e-01 1.12734771e+00
6.81791425e-01 -1.28460780e-01 -5.86477578e-01 8.99598122e-01
3.42754096e-01 4.02720645e-02 4.39227879e-01 -6.64719641e-01
7.11602509e-01 9.52508032e-01 -2.70083308e-01 -1.03889179e+00
-4.78490055e-01 -4.23669726e-01 -8.35960984e-01 -5.90887427e-01
-2.35185474e-01 -3.48011434e-01 -5.04079759e-01 1.95999682e+00
2.89757967e-01 2.06525236e-01 3.33024323e-01 6.81381404e-01
1.16065407e+00 7.56426454e-01 7.90883005e-02 -4.64206308e-01
1.39990163e+00 -1.44548941e+00 -1.20012057e+00 -4.60598737e-01
8.25500786e-01 -3.59336019e-01 9.78890896e-01 4.86076206e-01
-1.00464356e+00 -3.58577698e-01 -9.47201192e-01 -3.58761460e-01
-1.32164910e-01 6.93301708e-02 7.40228295e-01 -1.50155559e-01
-9.05740440e-01 2.28214696e-01 -6.72466874e-01 2.14763403e-01
2.64693111e-01 2.16569453e-01 2.65703369e-02 -5.47048263e-02
-1.86630881e+00 8.97999704e-01 5.71397603e-01 5.37447810e-01
-7.47270942e-01 -4.21246260e-01 -9.78551745e-01 2.19238892e-01
1.05167389e+00 -7.41827130e-01 1.21389711e+00 -8.51427376e-01
-1.48049819e+00 4.71250355e-01 -4.46102023e-01 4.13180608e-03
2.69138157e-01 -4.82798725e-01 -6.92042708e-01 -1.87356800e-01
-5.76578639e-02 3.00818533e-01 6.79142833e-01 -1.19033480e+00
-6.79627120e-01 -1.77474201e-01 2.92557865e-01 3.50301176e-01
-3.73482108e-01 2.81409860e-01 -8.71402621e-01 -5.68412423e-01
-2.10776404e-01 -6.06705487e-01 -1.54712245e-01 -6.43651664e-01
-8.53222132e-01 -8.11492324e-01 8.41573238e-01 -5.87373614e-01
1.60689712e+00 -1.93916714e+00 7.51267433e-01 -1.10856466e-01
6.34708226e-01 3.69004868e-02 -2.69420207e-01 5.55344999e-01
-1.50250211e-01 1.59748763e-01 5.85407652e-02 -3.73809874e-01
1.74978733e-01 5.09967685e-01 -3.32615227e-01 -1.95610315e-01
5.17405570e-01 1.50885463e+00 -1.06451106e+00 -9.03178096e-01
-1.84014738e-01 3.47083986e-01 -4.00872231e-01 4.03384775e-01
-5.88540316e-01 1.60202429e-01 -7.56421924e-01 5.59412897e-01
2.66871512e-01 -6.39239907e-01 5.37487447e-01 -4.62586939e-01
8.71025175e-02 4.74446326e-01 -8.71745884e-01 1.68638027e+00
-5.19441545e-01 2.23757684e-01 -2.43565366e-01 -9.98455584e-01
9.89092231e-01 4.55660105e-01 3.70653301e-01 -7.86401510e-01
-3.31647918e-02 2.04791538e-02 -1.17183840e-02 -8.99032235e-01
1.46628529e-01 -2.00662583e-01 -3.83108974e-01 4.58726615e-01
2.88304627e-01 1.80536076e-01 1.74256831e-01 5.48311114e-01
1.22087967e+00 4.82396990e-01 4.03852016e-01 1.37574062e-01
6.80516541e-01 -2.78971583e-01 1.04015350e+00 2.79909074e-01
-9.89704207e-02 -7.99684748e-02 1.14742637e+00 -3.67654175e-01
-4.35482532e-01 -5.11782587e-01 4.98886228e-01 1.21496761e+00
1.96315482e-01 -9.23639536e-01 -2.44570240e-01 -9.90254104e-01
-2.39035264e-01 8.23017836e-01 -8.85503769e-01 -3.76102328e-01
-6.68533146e-01 -8.96768153e-01 6.58091724e-01 6.38708711e-01
5.20303249e-01 -1.27362931e+00 -2.92349011e-01 3.56927812e-01
-8.36747587e-01 -1.33296239e+00 -1.94234058e-01 3.94263834e-01
-4.44195420e-01 -1.18509364e+00 -2.66193122e-01 -6.00460827e-01
2.38654599e-01 -1.83063015e-01 1.31607640e+00 1.74968749e-01
1.57242090e-01 8.73630792e-02 -6.79720998e-01 -1.91487536e-01
-2.82136500e-01 1.52620912e-01 -4.70988095e-01 4.88045774e-02
4.39343989e-01 -7.48296440e-01 -1.19982600e-01 1.07703574e-01
-6.15906537e-01 5.74637949e-01 7.67122924e-01 9.27948892e-01
5.82058609e-01 2.00576797e-01 7.43265867e-01 -1.26983488e+00
8.04209232e-01 -4.87753481e-01 -2.06192151e-01 9.31016147e-01
-7.04391301e-01 3.63729268e-01 7.47040033e-01 -3.13745230e-01
-1.47109389e+00 -2.95839399e-01 9.79400948e-02 -7.17944205e-01
2.76253283e-01 9.24616754e-01 -3.37026983e-01 5.27038574e-01
7.33348057e-02 1.04393780e-01 -6.77629292e-01 -6.55199885e-02
5.24023533e-01 3.32646370e-01 5.71822405e-01 -8.23229074e-01
6.89841568e-01 -3.48911062e-02 -8.34602863e-02 -1.67690948e-01
-1.39017725e+00 -1.13827221e-01 -5.89955926e-01 -4.74017709e-01
9.85281289e-01 -9.65192139e-01 -9.65055704e-01 2.65785396e-01
-1.47958946e+00 -5.11812508e-01 -2.70956811e-02 1.55633062e-01
-3.59408110e-01 2.00337395e-01 -1.01683402e+00 -9.79191542e-01
-4.68244255e-01 -1.00354958e+00 1.10641539e+00 -1.21325664e-02
-1.52455330e-01 -1.13282347e+00 9.10903960e-02 4.34807867e-01
-1.10068200e-02 3.27785611e-01 1.38386691e+00 -7.24252880e-01
-4.70308602e-01 3.01595211e-01 -3.26494277e-01 -7.14021772e-02
1.31458655e-01 1.74648687e-01 -9.05860662e-01 1.44061223e-01
1.20864259e-02 -8.34206641e-01 9.04628277e-01 6.89022169e-02
1.11542654e+00 -4.50631261e-01 -3.56639594e-01 3.19116086e-01
1.17892814e+00 2.40949720e-01 5.37570298e-01 1.57927275e-01
1.14972830e+00 8.52314889e-01 6.49121881e-01 1.86591491e-01
1.15973270e+00 6.61785722e-01 4.79960203e-01 -2.29549050e-01
2.44415086e-02 -4.43792433e-01 6.45157158e-01 1.60083377e+00
-2.36478835e-01 -3.00878972e-01 -8.63440275e-01 4.15925771e-01
-2.42567778e+00 -8.67056251e-01 -3.98256630e-01 1.30705011e+00
1.04575658e+00 2.02195868e-01 -1.57558337e-01 2.16440093e-02
5.30549645e-01 4.71520483e-01 -7.23203480e-01 -4.94087070e-01
-1.82335287e-01 -1.08603179e-01 -2.35660344e-01 5.05326509e-01
-8.95169556e-01 1.16172624e+00 5.40120363e+00 7.76295781e-01
-7.33701348e-01 3.18587273e-01 4.85248566e-01 -7.52183273e-02
-7.45042801e-01 7.32179210e-02 -5.86915016e-01 1.61974967e-01
6.96277797e-01 -3.43243688e-01 4.23564315e-01 5.26059151e-01
-1.34811789e-01 2.37324908e-01 -1.15047634e+00 7.18092918e-01
2.46407241e-01 -1.02187896e+00 7.74634555e-02 -2.53209859e-01
5.18180668e-01 -8.17242563e-02 -3.45942914e-01 9.23246205e-01
8.77481043e-01 -7.62727797e-01 7.87799776e-01 5.26300788e-01
6.03593051e-01 -8.67909551e-01 8.46742809e-01 5.15369415e-01
-1.63697612e+00 -2.49641776e-01 -4.02762331e-02 -2.68807523e-02
3.62943828e-01 7.47458696e-01 -3.48142982e-01 1.26810312e+00
6.72149599e-01 1.14778018e+00 -6.03374302e-01 8.43073949e-02
-9.21483457e-01 4.58104998e-01 2.51353562e-01 -1.93519145e-01
3.12113851e-01 -1.48416668e-01 3.63710105e-01 1.26819158e+00
-6.54295366e-03 2.48386234e-01 2.74843901e-01 1.10582507e+00
-2.23359704e-01 -3.43561545e-02 -3.99499983e-01 -8.37715119e-02
3.28392565e-01 1.40329778e+00 -4.53104109e-01 -3.79483342e-01
-6.07836783e-01 1.07155263e+00 9.68585432e-01 4.98312116e-01
-1.07728863e+00 -3.08367610e-01 2.98768878e-01 -7.08787799e-01
3.18537563e-01 1.66874573e-01 8.81539807e-02 -1.52020049e+00
2.44380742e-01 -7.97325134e-01 7.37849712e-01 -1.10155451e+00
-1.35509455e+00 8.78732979e-01 -2.81448360e-03 -7.11204529e-01
-4.77222323e-01 -1.39874175e-01 -6.44974470e-01 6.09551013e-01
-1.86452973e+00 -1.56687808e+00 -2.08428670e-02 7.49855042e-01
6.08312130e-01 1.88433290e-01 7.99037457e-01 2.24142745e-01
-1.05580246e+00 4.07136649e-01 -7.02062011e-01 2.33185306e-01
5.48409879e-01 -1.41006351e+00 2.02773601e-01 9.83945131e-01
1.95280001e-01 4.22670364e-01 2.66819865e-01 -8.69100273e-01
-1.47820342e+00 -1.18337417e+00 1.39565611e+00 -4.23624247e-01
9.05299067e-01 -5.52754343e-01 -1.08436263e+00 1.15824592e+00
2.89457500e-01 -1.54954880e-01 6.28727436e-01 6.62671268e-01
-6.82512164e-01 -6.50806874e-02 -2.91254520e-01 5.08706748e-01
1.19752634e+00 -7.62375355e-01 -9.06522512e-01 3.14703405e-01
1.23473525e+00 -4.46604401e-01 -8.85114074e-01 5.55685461e-01
3.15833807e-01 -8.81280422e-01 5.71151376e-01 -8.13451886e-01
1.07623780e+00 -6.55204803e-02 1.61973983e-02 -1.25610793e+00
-5.55158198e-01 -7.44263470e-01 -6.55924141e-01 1.68208182e+00
6.89613104e-01 -4.14427429e-01 1.03325114e-01 6.72941923e-01
-2.81994790e-01 -1.26243556e+00 -5.89634299e-01 -5.53444862e-01
-2.73202956e-01 -7.35980511e-01 5.78868628e-01 1.24477696e+00
4.16942984e-01 9.23629165e-01 -6.88399851e-01 1.12925649e-01
1.71753913e-01 6.23586476e-01 4.16670978e-01 -1.22753561e+00
-5.80687940e-01 -4.18869168e-01 5.02628565e-01 -8.34468603e-01
6.41582727e-01 -1.14301777e+00 1.26706734e-01 -1.84367669e+00
5.59746385e-01 -2.17212528e-01 -5.43096006e-01 1.13197124e+00
-7.03335524e-01 -5.12967944e-01 1.81220084e-01 2.19263330e-01
-1.18733883e+00 9.14291024e-01 1.45420361e+00 -2.88545322e-02
-1.68541119e-01 -4.29400682e-01 -9.74569917e-01 6.72118247e-01
3.54550958e-01 -5.34291923e-01 -7.12271154e-01 -5.70069551e-01
8.86535883e-01 3.18654537e-01 2.51062065e-01 -3.47821563e-01
5.15756667e-01 -1.99446425e-01 9.44744721e-02 -4.39465404e-01
7.76340663e-02 -6.54379725e-01 5.34975342e-02 1.44559070e-01
-5.39354324e-01 1.58228800e-01 -7.38550257e-03 5.88601708e-01
-4.24885541e-01 2.14476913e-01 2.34847777e-02 1.36136927e-03
-7.43570387e-01 2.70152211e-01 -1.12240620e-01 3.16203654e-01
8.89262795e-01 4.38165545e-01 -5.87091625e-01 -3.87952656e-01
-5.63272059e-01 8.40659201e-01 -1.78289115e-01 6.46852672e-01
5.89193940e-01 -1.46953273e+00 -7.59177744e-01 -1.52741313e-01
2.81714082e-01 2.73136407e-01 2.70935714e-01 9.90240872e-01
4.16337430e-01 3.61752182e-01 1.92887262e-01 -2.04529837e-01
-1.26204967e+00 7.16539979e-01 2.67333180e-01 -1.08758450e+00
-6.26470447e-01 9.20346081e-01 2.27626264e-01 -3.63391519e-01
6.91346452e-02 -4.60052758e-01 -5.83593130e-01 2.68198550e-01
2.04308853e-01 -1.11329751e-02 -4.15062010e-02 -3.81745934e-01
-4.24103111e-01 4.22488779e-01 -2.10701749e-01 1.91522650e-02
1.53954840e+00 -7.98158199e-02 -5.47186732e-01 1.01335132e+00
9.82088804e-01 -2.43577704e-01 -1.10171366e+00 -4.86008197e-01
2.35468373e-01 1.98139995e-01 3.10518835e-02 -9.83792424e-01
-1.27551579e+00 9.14219260e-01 -6.70003057e-01 1.36046380e-01
1.44297206e+00 2.76930481e-01 8.73920560e-01 4.28786397e-01
1.88419625e-01 -1.05958247e+00 4.94579673e-01 6.91489995e-01
9.96163368e-01 -8.45823407e-01 -1.35536253e-01 -6.91557288e-01
-1.09756279e+00 1.03434539e+00 9.93130147e-01 2.99512923e-01
4.18677419e-01 5.70285618e-01 -1.23590738e-01 -6.79745317e-01
-1.68946099e+00 -2.23989785e-01 4.34058577e-01 2.15147883e-01
5.80002129e-01 -8.38850290e-02 -2.07473010e-01 1.27538395e+00
1.85376760e-02 2.01754645e-01 2.30967086e-02 7.31607497e-01
1.14769623e-01 -1.22126341e+00 2.82718062e-01 3.21737319e-01
-2.31384441e-01 -2.80782700e-01 -7.34147370e-01 8.05854440e-01
2.32059300e-01 9.47762966e-01 -3.16233784e-01 -7.48505414e-01
3.62980157e-01 3.03549141e-01 1.63196400e-01 -3.56729895e-01
-8.90310049e-01 2.64002234e-01 4.17936355e-01 -8.46203744e-01
-7.10087299e-01 -3.66975814e-01 -1.44587231e+00 1.22715324e-01
-3.47534716e-01 8.99009109e-02 3.60594504e-02 1.35947704e+00
3.53204995e-01 1.37834287e+00 6.31454170e-01 -2.01232105e-01
-1.38967797e-01 -8.21011782e-01 -3.93969715e-01 3.12919080e-01
1.86802357e-01 -6.19579971e-01 -1.71308845e-01 -3.88122983e-02]
|
[12.581441879272461, 6.2674713134765625]
|
43eb10b0-7df4-4d6c-8de0-37b245ac04d5
|
intelligent-video-editing-incorporating
|
2110.08580
| null |
https://arxiv.org/abs/2110.08580v1
|
https://arxiv.org/pdf/2110.08580v1.pdf
|
Intelligent Video Editing: Incorporating Modern Talking Face Generation Algorithms in a Video Editor
|
This paper proposes a video editor based on OpenShot with several state-of-the-art facial video editing algorithms as added functionalities. Our editor provides an easy-to-use interface to apply modern lip-syncing algorithms interactively. Apart from lip-syncing, the editor also uses audio and facial re-enactment to generate expressive talking faces. The manual control improves the overall experience of video editing without missing out on the benefits of modern synthetic video generation algorithms. This control enables us to lip-sync complex dubbed movie scenes, interviews, television shows, and other visual content. Furthermore, our editor provides features that automatically translate lectures from spoken content, lip-sync of the professor, and background content like slides. While doing so, we also tackle the critical aspect of synchronizing background content with the translated speech. We qualitatively evaluate the usefulness of the proposed editor by conducting human evaluations. Our evaluations show a clear improvement in the efficiency of using human editors and an improved video generation quality. We attach demo videos with the supplementary material clearly explaining the tool and also showcasing multiple results.
|
['C. V. Jawahar', 'Vinay P. Namboodiri', 'Rudrabha Mukhopadhyay', 'Faizan Farooq Khan', 'Anchit Gupta']
|
2021-10-16
| null | null | null | null |
['talking-face-generation']
|
['computer-vision']
|
[ 1.90273792e-01 4.48027283e-01 1.76477388e-01 -1.64627805e-01
-7.62401819e-01 -5.56270003e-01 7.12190032e-01 -2.76364893e-01
-1.69625044e-01 6.45295143e-01 3.63659084e-01 4.88525778e-02
2.64771223e-01 -3.51719409e-01 -6.33150041e-01 -2.63833702e-01
3.21037769e-01 -4.00468633e-02 2.55416989e-01 -2.73083568e-01
2.41793215e-01 4.21542972e-01 -1.94323134e+00 5.23292899e-01
6.64880395e-01 6.22621953e-01 1.98290020e-01 8.96004081e-01
-4.40186620e-01 6.61275446e-01 -8.27853739e-01 -6.71544373e-01
-1.56410504e-02 -5.20042777e-01 -3.64393294e-01 3.39170754e-01
5.75911403e-01 -5.69044292e-01 -1.90243065e-01 8.87155056e-01
7.41039693e-01 -4.06739637e-02 3.05550158e-01 -1.54166377e+00
-2.84776211e-01 4.75330770e-01 -4.26898241e-01 -4.44566935e-01
1.28413963e+00 2.37499505e-01 4.65190917e-01 -9.36341166e-01
1.07883847e+00 1.19828665e+00 3.99817556e-01 7.71925926e-01
-8.35109770e-01 -7.73248076e-01 -1.47865906e-01 1.01256676e-01
-1.40652442e+00 -1.05370033e+00 8.03301513e-01 -3.79783094e-01
5.21184623e-01 5.82449734e-01 8.07865202e-01 1.17448604e+00
-3.28000903e-01 6.30030274e-01 6.19725883e-01 -8.36724997e-01
-1.11968167e-01 7.07181513e-01 -3.78626972e-01 8.43207121e-01
-5.71087718e-01 -1.78571612e-01 -7.84078419e-01 6.27471730e-02
6.92515850e-01 -3.42352301e-01 -5.60466111e-01 -1.77571386e-01
-1.21465063e+00 2.45720893e-01 -3.60221624e-01 1.36764467e-01
-2.66148467e-02 1.78317614e-02 4.93339479e-01 2.14915395e-01
3.14894468e-01 -1.24737576e-01 1.90159857e-01 -4.68625039e-01
-1.30426800e+00 2.31198996e-01 8.00193369e-01 1.28495920e+00
4.04057384e-01 6.31850511e-02 -3.18195313e-01 7.93789625e-01
1.89226776e-01 5.32837272e-01 3.03423762e-01 -1.28822303e+00
2.88183838e-01 3.30451638e-01 2.31961936e-01 -8.34065676e-01
1.08701877e-01 3.51696014e-01 -2.21167341e-01 4.63745803e-01
3.16722870e-01 -3.25264037e-01 -3.47603530e-01 1.52146983e+00
3.69569749e-01 7.28422496e-03 -1.70385867e-01 6.60562515e-01
1.14988744e+00 6.67675197e-01 -1.18345179e-01 -5.39121926e-01
1.38053012e+00 -1.00648057e+00 -1.44054389e+00 4.35359299e-01
3.40730608e-01 -1.18588769e+00 1.27618146e+00 3.83650988e-01
-1.39813149e+00 -6.04575098e-01 -8.06113482e-01 -9.89427790e-02
-1.70364648e-01 5.46441793e-01 4.99313399e-02 8.11474264e-01
-1.22046602e+00 5.66256404e-01 -4.07015800e-01 -4.58139449e-01
7.68321753e-02 1.90104127e-01 -7.48943448e-01 3.86964709e-01
-8.23446870e-01 6.06657624e-01 -1.14230841e-01 -3.13327432e-01
-3.60233873e-01 -8.75178218e-01 -8.79555404e-01 3.83756347e-02
4.21886742e-01 -3.84777904e-01 1.46120107e+00 -1.39481235e+00
-2.34528399e+00 1.00500989e+00 -3.67736816e-01 1.01332843e-01
1.12546146e+00 -3.32060456e-01 -5.37253678e-01 7.34376192e-01
-1.34814262e-01 9.81417477e-01 1.08248329e+00 -1.12038684e+00
-4.94719952e-01 2.10151583e-01 5.97178638e-02 1.75465658e-01
-3.96615624e-01 3.74531895e-01 -7.80810356e-01 -8.33545387e-01
-5.66320360e-01 -6.61695898e-01 4.71875191e-01 6.58757985e-01
-3.41343254e-01 6.75783008e-02 1.25977671e+00 -7.32414663e-01
1.42417407e+00 -2.45005369e+00 -9.73904505e-02 -5.10340789e-03
1.39427595e-02 2.15063095e-01 1.13500643e-03 6.27098024e-01
-2.03975201e-01 9.12492499e-02 1.58732459e-01 -5.90868652e-01
9.14562345e-02 -2.87053972e-01 -1.37001142e-01 2.44283393e-01
-1.43565983e-01 4.14104104e-01 -6.07635379e-01 -8.05551589e-01
4.90524322e-01 8.01421821e-01 -5.00790179e-01 3.26715678e-01
-1.14258066e-01 1.60821140e-01 2.12289065e-01 5.63800514e-01
6.08941257e-01 2.74177760e-01 2.70388901e-01 -2.38740578e-01
-4.79518801e-01 -7.82686919e-02 -1.34589958e+00 1.75736713e+00
-5.88311255e-01 1.03419876e+00 5.28535545e-01 -3.04283947e-01
8.03378880e-01 7.21561134e-01 2.46651888e-01 -2.89997309e-01
-1.55074839e-02 -1.66318677e-02 -5.74070334e-01 -1.01425374e+00
6.49493754e-01 1.26684427e-01 2.25624129e-01 6.39262259e-01
2.30458602e-01 -3.85660768e-01 2.96634436e-01 5.03559589e-01
4.62004006e-01 6.44171536e-01 3.66567314e-01 3.35039459e-02
6.43065333e-01 -5.53451955e-01 -7.45862946e-02 1.54835805e-01
-2.78313570e-02 7.54676163e-01 6.30348682e-01 1.57072656e-02
-8.42481196e-01 -9.08012271e-01 2.63613075e-01 1.07096064e+00
-2.07573861e-01 -9.58756447e-01 -1.30556357e+00 -4.45527285e-01
-4.73745435e-01 6.41169727e-01 -3.46168607e-01 3.71104509e-01
-3.68564963e-01 3.36433411e-01 6.20277762e-01 9.99517515e-02
2.54898667e-01 -1.20428407e+00 -7.48068452e-01 -1.74692929e-01
-5.42650282e-01 -1.23237276e+00 -8.62591505e-01 -7.20867038e-01
-3.84978175e-01 -8.33065391e-01 -1.11535156e+00 -8.08560610e-01
6.85669065e-01 1.24480411e-01 8.11768055e-01 1.24015637e-01
-4.34489638e-01 7.47584581e-01 -3.50519598e-01 -3.64494383e-01
-8.53156328e-01 -1.97530895e-01 7.51772523e-02 1.45548224e-01
-2.44657233e-01 -5.60913146e-01 -2.96143860e-01 3.62349778e-01
-8.11141253e-01 5.37476778e-01 -1.25939220e-01 2.09665880e-01
1.55260548e-01 -5.27658522e-01 3.34735215e-01 -6.47712350e-01
7.07765281e-01 5.93047179e-02 -5.75636744e-01 3.74342293e-01
1.19963661e-02 -1.74582720e-01 5.72809637e-01 -4.41681087e-01
-1.13780415e+00 1.48659602e-01 -2.84562379e-01 -5.61561763e-01
-2.76929498e-01 -2.47103393e-01 -1.89415082e-01 -5.40234670e-02
3.51380199e-01 -9.48677957e-02 3.46524000e-01 -1.73589244e-01
5.07378340e-01 1.00610340e+00 7.93465793e-01 -2.67773747e-01
7.04874694e-01 4.17968422e-01 -4.03353125e-01 -1.21928120e+00
6.72454527e-03 -4.15577218e-02 -5.03447235e-01 -1.04046571e+00
6.22747898e-01 -7.48094380e-01 -1.20891476e+00 4.82046247e-01
-1.23187470e+00 -2.21183881e-01 -2.73770571e-01 2.43924811e-01
-7.21018672e-01 6.07447743e-01 -4.60820556e-01 -9.30574954e-01
-8.38561282e-02 -1.15057755e+00 1.04300630e+00 3.99252683e-01
-5.37243068e-01 -7.40729094e-01 -5.22049926e-02 4.03678924e-01
2.69544572e-01 3.46145213e-01 4.46808189e-01 -5.10225482e-02
-3.74286890e-01 -1.08969025e-01 8.41032490e-02 1.53040722e-01
2.13483691e-01 8.77130270e-01 -1.31575894e+00 1.67476475e-01
-6.03669345e-01 -1.56944275e-01 2.84633875e-01 1.71575040e-01
9.28792655e-01 -4.14487481e-01 -2.87475109e-01 4.25119221e-01
1.01583266e+00 1.98781833e-01 7.36074746e-01 -9.44204777e-02
3.09957057e-01 1.05828834e+00 5.79993784e-01 6.09862447e-01
6.83268234e-02 1.04026473e+00 7.19562545e-02 -1.35656059e-01
-2.04533696e-01 -4.18016464e-01 6.67575121e-01 7.22857594e-01
-1.39621407e-01 -2.41793200e-01 -5.11786401e-01 3.03791076e-01
-1.50180900e+00 -1.21257603e+00 -6.94749206e-02 2.06715536e+00
8.29595625e-01 -9.99263749e-02 3.87508392e-01 3.40612382e-01
9.40602958e-01 -9.35510173e-02 1.52341887e-01 -6.05479240e-01
2.15962842e-01 2.10420355e-01 -5.04684038e-02 9.32090044e-01
-7.63591349e-01 8.99269462e-01 6.41482306e+00 8.12383890e-01
-1.37067187e+00 -9.17162821e-02 1.38046011e-01 -4.57284808e-01
-3.54957521e-01 -3.55604917e-01 -4.70722020e-01 4.91670042e-01
6.77842021e-01 -3.01530451e-01 4.37283069e-01 6.68106914e-01
7.75073051e-01 -3.28386962e-01 -1.09192050e+00 1.22543204e+00
4.34759289e-01 -1.54690588e+00 1.92468509e-01 -1.81831121e-01
3.04717958e-01 -8.14077973e-01 -6.37442172e-02 2.05651764e-02
-4.41670299e-01 -7.30391145e-01 1.20998669e+00 6.17362618e-01
1.51079261e+00 -6.74658656e-01 3.03076744e-01 -3.08325700e-02
-1.13735104e+00 2.38870770e-01 1.41918942e-01 1.55500874e-01
4.81690526e-01 1.63653940e-01 -8.23584199e-01 4.22305763e-01
5.02687335e-01 2.92043090e-01 -3.50382686e-01 8.14883888e-01
-3.10574502e-01 5.70916794e-02 -2.43178606e-01 -7.13532371e-03
-3.58777612e-01 -1.12169199e-01 5.96973896e-01 1.50273645e+00
5.02679110e-01 6.74679652e-02 -3.49448144e-01 6.40506685e-01
-1.40915290e-01 3.89885724e-01 -7.49649167e-01 -1.41933084e-01
3.61762673e-01 1.35266745e+00 -6.64724886e-01 -2.67380804e-01
-2.48374432e-01 1.18101776e+00 -3.14159751e-01 1.82861537e-01
-1.04654527e+00 -8.02861989e-01 6.13974571e-01 6.81791306e-01
2.10090484e-02 6.89545199e-02 2.81174392e-01 -9.01476026e-01
1.27120286e-01 -9.58646119e-01 -1.79719284e-01 -1.41067314e+00
-4.22220528e-01 6.77047789e-01 9.47871134e-02 -1.41891813e+00
-4.43401605e-01 -2.56976098e-01 -8.28982532e-01 4.69651341e-01
-9.93426144e-01 -1.05061615e+00 -5.73442459e-01 6.99725807e-01
8.40683877e-01 -1.25109568e-01 8.04023445e-01 4.45596665e-01
-3.36421371e-01 7.95267940e-01 -3.33149135e-01 -9.79214441e-03
1.23089945e+00 -7.79295385e-01 7.10041523e-02 6.96319997e-01
-6.56328052e-02 3.06617498e-01 7.84407854e-01 -3.13522249e-01
-9.22630668e-01 -4.70597774e-01 9.36853588e-01 -1.69075176e-01
3.02064031e-01 -6.18955076e-01 -5.07051349e-01 3.45700920e-01
7.06943631e-01 -2.83150136e-01 6.50036454e-01 -5.18804312e-01
-2.99597718e-02 -2.52897531e-01 -1.14519346e+00 8.95676315e-01
9.72039461e-01 -6.08373463e-01 -3.96199733e-01 1.19550280e-01
5.57001352e-01 -5.14211237e-01 -3.49251479e-01 -1.30077943e-01
1.02012157e+00 -1.34687185e+00 6.61173880e-01 4.85788994e-02
6.68154001e-01 -3.11791927e-01 2.97730595e-01 -1.00879121e+00
3.83681506e-01 -1.43859351e+00 3.09680462e-01 1.70280874e+00
2.97429234e-01 -1.57026142e-01 5.09034336e-01 4.95237172e-01
4.56071272e-02 -2.55396366e-01 -6.57738209e-01 -2.46481538e-01
-6.62317991e-01 -3.99414212e-01 3.73466492e-01 7.87674606e-01
7.26585090e-01 -5.72797880e-02 -4.33421314e-01 -1.47253931e-01
3.97917747e-01 -3.59210789e-01 1.15508127e+00 -9.20341432e-01
-5.96795492e-02 -3.41230005e-01 -1.52892470e-01 -6.24746561e-01
1.64474458e-01 -5.79888999e-01 -1.91420063e-01 -1.22205341e+00
-1.24647304e-01 2.66352594e-01 6.37074769e-01 2.24373400e-01
1.35775268e-01 3.34664553e-01 7.10322499e-01 -8.95029455e-02
-4.26707298e-01 2.88307518e-01 1.18496406e+00 2.25595400e-01
-3.67423207e-01 -1.21413127e-01 -2.98670679e-01 7.55367339e-01
4.96212542e-01 -3.16628456e-01 -4.38787401e-01 -3.92668396e-02
1.42785132e-01 4.26579982e-01 3.69071752e-01 -1.14655805e+00
1.32646546e-01 1.23177126e-01 1.29262462e-01 -2.63509303e-01
4.39866722e-01 -7.37436593e-01 3.73109967e-01 1.90796331e-01
-4.27980900e-01 1.44825757e-01 3.19545239e-01 4.95555028e-02
-2.53369451e-01 -2.73816735e-01 8.38344514e-01 -1.58896863e-01
-2.51866937e-01 -1.91247821e-01 -8.46716285e-01 -1.42260656e-01
1.24568641e+00 -6.57303274e-01 -7.16308728e-02 -1.18867779e+00
-9.53500390e-01 -1.03861973e-01 6.68205619e-01 4.09669399e-01
6.49565160e-01 -1.09326816e+00 -3.90403807e-01 6.12759888e-01
-1.27197325e-01 -5.20112813e-01 4.19219196e-01 6.19409621e-01
-9.67782497e-01 1.07492767e-01 -4.46063012e-01 -5.03338039e-01
-2.01673007e+00 3.24398875e-01 1.84869409e-01 2.85153508e-01
-6.12578154e-01 5.36978960e-01 8.88815373e-02 -4.14945791e-03
5.89736342e-01 -3.54779899e-01 -1.95399985e-01 3.29928696e-01
7.84485817e-01 6.19735837e-01 1.30261050e-03 -3.34739625e-01
-2.80007005e-01 5.30649722e-01 3.74747545e-01 -6.11527324e-01
9.78191316e-01 -4.47510242e-01 1.34176373e-01 3.45582545e-01
8.83037269e-01 8.95482898e-01 -1.17627144e+00 3.91868502e-01
-6.16457403e-01 -4.64798659e-01 -4.38184172e-01 -6.92832589e-01
-8.69682133e-01 9.23713326e-01 4.63563353e-01 7.73990750e-02
1.07144046e+00 -3.30561727e-01 4.80701953e-01 1.24350460e-02
1.40291139e-01 -1.25551987e+00 1.76345110e-01 2.61451211e-02
1.11971617e+00 -7.65210569e-01 -7.47301728e-02 -7.04230964e-01
-9.22630668e-01 1.47153223e+00 4.64407384e-01 3.17332625e-01
6.06320083e-01 8.35127056e-01 3.80024463e-01 1.97130188e-01
-7.90426850e-01 1.37439504e-01 -3.73988263e-02 7.19623923e-01
5.18410265e-01 -8.07578266e-02 -2.82786936e-01 4.18156147e-01
-4.11062211e-01 4.28101927e-01 9.81057525e-01 6.31433547e-01
-1.30077526e-01 -1.03146780e+00 -5.61007023e-01 -2.57240117e-01
-6.87064290e-01 4.20158431e-02 -5.66229582e-01 8.86228740e-01
7.86802247e-02 9.59940076e-01 9.65420231e-02 -1.40518010e-01
2.90133893e-01 3.60222638e-01 5.38658738e-01 -3.84679735e-01
-7.28316188e-01 2.23718986e-01 1.79548606e-01 -6.26467228e-01
-5.29183447e-01 -3.36808443e-01 -8.88367355e-01 -4.77483958e-01
-1.05621554e-01 1.47509202e-01 7.92221367e-01 5.92837036e-01
4.28262293e-01 3.71159047e-01 4.44799036e-01 -1.15723455e+00
7.16519505e-02 -9.46545899e-01 -2.80342281e-01 3.41920286e-01
3.20507348e-01 -4.38218981e-01 -3.44475120e-01 6.44551218e-01]
|
[13.2421875, -0.4506858289241791]
|
73bcb3f7-c7cc-44c2-8e25-3b6684a6e5d6
|
designing-and-evaluating-speech-emotion
|
2304.00860
| null |
https://arxiv.org/abs/2304.00860v1
|
https://arxiv.org/pdf/2304.00860v1.pdf
|
Designing and Evaluating Speech Emotion Recognition Systems: A reality check case study with IEMOCAP
|
There is an imminent need for guidelines and standard test sets to allow direct and fair comparisons of speech emotion recognition (SER). While resources, such as the Interactive Emotional Dyadic Motion Capture (IEMOCAP) database, have emerged as widely-adopted reference corpora for researchers to develop and test models for SER, published work reveals a wide range of assumptions and variety in its use that challenge reproducibility and generalization. Based on a critical review of the latest advances in SER using IEMOCAP as the use case, our work aims at two contributions: First, using an analysis of the recent literature, including assumptions made and metrics used therein, we provide a set of SER evaluation guidelines. Second, using recent publications with open-sourced implementations, we focus on reproducibility assessment in SER.
|
['Shrikanth Narayanan', 'Theodoros Giannakopoulos', 'Athanasios Katsamanis', 'Nikolaos Antoniou']
|
2023-04-03
| null | null | null | null |
['speech-emotion-recognition']
|
['speech']
|
[-1.38378918e-01 -2.56111145e-01 -9.63389352e-02 -3.50932389e-01
-8.08377862e-01 -6.07487321e-01 5.84641933e-01 -1.41935110e-01
-6.14051998e-01 4.52131331e-01 4.19001669e-01 -1.03084426e-02
-9.56425071e-02 1.87555149e-01 -3.45531285e-01 -2.46991187e-01
-2.67778426e-01 -1.20688997e-01 1.06557339e-01 -2.85411328e-01
-4.09000851e-02 6.09773755e-01 -2.13625789e+00 4.10175264e-01
2.60170668e-01 1.24281836e+00 -6.03136234e-02 7.98864186e-01
8.55320618e-02 1.01836383e+00 -9.54128623e-01 -7.87122905e-01
-3.55936378e-01 -5.42823434e-01 -8.88379276e-01 -1.11717083e-01
1.02951221e-01 -9.14499909e-02 -3.18517059e-01 6.71738625e-01
1.11515808e+00 5.44965029e-01 4.11739409e-01 -1.48413718e+00
-5.81536651e-01 1.23360522e-01 3.37549359e-01 6.92891002e-01
6.79424703e-01 -1.64639559e-02 9.06055391e-01 -1.05116200e+00
1.00542343e+00 8.86773407e-01 8.33129764e-01 9.25948322e-01
-6.46771669e-01 -5.52621484e-01 -7.13963807e-02 6.76737905e-01
-1.32402778e+00 -1.04475760e+00 8.07975113e-01 -1.43032074e-01
1.36535084e+00 6.71982825e-01 8.44604790e-01 1.88086677e+00
-1.12035573e-01 8.69080484e-01 1.12027729e+00 -3.49087477e-01
4.09375161e-01 5.10434926e-01 -2.68251114e-02 2.70674713e-02
-2.32903585e-01 2.48648927e-01 -8.65901053e-01 -2.92443991e-01
3.07460487e-01 -8.06463242e-01 -3.23622048e-01 -2.08701462e-01
-7.79694617e-01 3.89830351e-01 -4.31682736e-01 7.52397776e-01
-4.74342972e-01 -3.13172936e-01 9.21037197e-01 4.17345405e-01
5.69786966e-01 1.18305264e-02 -2.34824240e-01 -1.33514988e+00
-1.02781153e+00 1.15927562e-01 8.83216560e-01 5.83846807e-01
5.84620936e-03 2.28011653e-01 2.08561763e-01 1.22969091e+00
4.00835931e-01 1.45432368e-01 6.10899329e-01 -9.50264454e-01
1.43371314e-01 -5.16879298e-02 5.87675124e-02 -1.17473006e+00
-1.75725132e-01 -3.51544134e-02 -3.37290883e-01 2.70288080e-01
-4.38622646e-02 -4.10458058e-01 -2.59117693e-01 1.69263661e+00
1.45797819e-01 4.19385225e-01 3.43516201e-01 9.85064328e-01
1.09888685e+00 2.47021839e-01 5.21795332e-01 -4.61628586e-01
1.16707838e+00 -5.45640409e-01 -1.06523550e+00 6.84640408e-02
4.80061054e-01 -6.91564798e-01 9.58944023e-01 4.49889421e-01
-1.26814854e+00 -2.56144583e-01 -1.13159156e+00 2.50995681e-02
-5.92245460e-01 -2.91638970e-01 4.37123924e-01 1.04589641e+00
-1.34135091e+00 1.75561160e-01 -1.01340616e+00 -8.45821023e-01
6.29214421e-02 -5.08863367e-02 -6.72436595e-01 4.08451617e-01
-1.35869026e+00 1.25210285e+00 -5.12528718e-02 1.42637417e-01
-4.58718985e-01 -5.20151913e-01 -6.87437296e-01 -2.79259741e-01
2.41761710e-02 -2.92493612e-01 1.49894249e+00 -1.18723691e+00
-1.97364366e+00 9.95210946e-01 -9.21887159e-02 -4.09352064e-01
4.53367025e-01 -4.13019024e-02 -1.27036643e+00 5.33533454e-01
-5.77023506e-01 5.21808624e-01 4.87436980e-01 -9.51725900e-01
-4.70481694e-01 -6.86396435e-02 -1.59837872e-01 3.55865866e-01
-4.11933124e-01 1.03958356e+00 -4.46343184e-01 -5.33001304e-01
-4.99484301e-01 -7.30360925e-01 1.75569877e-01 -1.65577099e-01
3.02035779e-01 -6.87475801e-02 9.70152676e-01 -7.75869489e-01
1.62622297e+00 -2.43569732e+00 1.80078167e-02 -1.32235900e-01
1.08119130e-01 4.06540543e-01 1.86061785e-01 5.93657017e-01
-2.66625255e-01 2.36169532e-01 -1.01750001e-01 -7.07072675e-01
3.40287983e-01 1.55820146e-01 -1.77317247e-01 3.90843123e-01
1.11257084e-01 7.99435437e-01 -7.45920479e-01 -5.61832786e-01
6.80640697e-01 1.00783646e+00 -2.71976411e-01 2.28825152e-01
4.34857905e-01 1.84739128e-01 2.96965633e-02 4.86235410e-01
3.81371617e-01 4.72490229e-02 5.85338511e-02 1.63079649e-01
-2.16273680e-01 3.64737064e-01 -1.12466013e+00 1.48029256e+00
-4.48538005e-01 1.12670648e+00 3.29801619e-01 -7.03789651e-01
8.23457956e-01 1.02800250e+00 6.66167140e-01 -4.97129321e-01
3.72290403e-01 1.91861272e-01 -1.09035976e-01 -6.93446577e-01
6.12124383e-01 -1.44144118e-01 -1.52043663e-02 2.14616552e-01
3.32172573e-01 -3.98375131e-02 -8.40875953e-02 -3.55500467e-02
1.11368990e+00 -2.90128998e-02 3.08318138e-01 1.85554959e-02
6.99598312e-01 -1.14823692e-01 4.80594695e-01 3.20592850e-01
-9.69457328e-01 7.10687339e-01 4.11727339e-01 -2.48303622e-01
-8.77024591e-01 -1.00274479e+00 -4.28913683e-01 7.10971951e-01
-2.23763451e-01 -7.39960492e-01 -8.72268081e-01 -4.99275684e-01
-6.34966969e-01 6.98790848e-01 -5.59953392e-01 1.03034720e-01
-1.87328964e-01 -7.36282647e-01 9.33926165e-01 4.88910347e-01
4.09297466e-01 -1.42331207e+00 -1.06491363e+00 1.29369259e-01
-3.33090454e-01 -1.49590480e+00 -5.34337685e-02 -9.48153883e-02
-4.14348215e-01 -8.91313016e-01 -6.19843781e-01 -4.58251178e-01
-4.23892364e-02 1.39822826e-01 1.21259081e+00 -1.33543998e-01
1.04199932e-03 1.22675502e+00 -7.02722073e-01 -4.54327643e-01
-4.60814953e-01 -2.77335197e-01 1.57369673e-01 -1.15459837e-01
6.88580573e-01 -4.45914596e-01 -4.37287003e-01 4.43294466e-01
-9.77429092e-01 -3.89351398e-01 2.56601751e-01 4.55237657e-01
3.31309438e-01 -2.16318637e-01 7.01611996e-01 -1.61457062e-01
1.01651120e+00 -8.18382561e-01 -1.37952685e-01 6.84892908e-02
-5.82569957e-01 -5.80096066e-01 1.30910009e-01 -3.70720357e-01
-1.17410219e+00 -5.03678262e-01 -6.38780951e-01 -8.35631251e-01
-5.18762827e-01 5.55167198e-01 -6.26566485e-02 -1.34957835e-01
5.44368505e-01 9.29865390e-02 1.22987159e-01 -3.58254343e-01
1.86906829e-02 9.51134980e-01 5.78612328e-01 -3.03417265e-01
1.44083407e-02 2.96652704e-01 -5.68561196e-01 -1.25400364e+00
-4.47809920e-02 -4.11628813e-01 -3.94475996e-01 -7.54338682e-01
9.42748666e-01 -7.94275105e-01 -5.97614765e-01 4.05243874e-01
-8.88266087e-01 -3.41436446e-01 -2.51934409e-01 7.09082842e-01
-6.66117966e-01 5.24079680e-01 -4.22584713e-01 -1.25890017e+00
-3.17271173e-01 -9.57488656e-01 8.77890587e-01 2.38177739e-02
-7.91158915e-01 -1.24867630e+00 3.93261790e-01 3.34499985e-01
4.95159298e-01 2.63118714e-01 9.00250971e-02 -8.30873489e-01
1.48856774e-01 -1.33548766e-01 3.98118645e-01 6.61740482e-01
-2.35205486e-01 3.64331514e-01 -1.28935015e+00 6.05075099e-02
4.06710878e-02 -4.02590752e-01 9.58777890e-02 1.28909916e-01
7.83336759e-01 1.56742901e-01 1.03766114e-01 3.21126312e-01
8.06958556e-01 4.43078101e-01 8.88545156e-01 4.45188075e-01
-9.13033932e-02 7.08231986e-01 6.14354551e-01 4.37517375e-01
6.23214543e-01 8.69421482e-01 -2.85237897e-02 3.03493321e-01
-8.20207000e-02 8.99630710e-02 7.89306939e-01 1.25345659e+00
-1.30806372e-01 -4.11379814e-01 -7.90649056e-01 5.48325062e-01
-1.65914690e+00 -1.27589500e+00 2.46138468e-01 2.11788130e+00
2.72600770e-01 -2.32959241e-01 4.82038081e-01 2.39509851e-01
5.07321298e-01 4.82793301e-01 3.56043503e-02 -8.01463902e-01
-4.10032541e-01 2.27328524e-01 -2.52494901e-01 3.33106369e-01
-7.54002452e-01 7.04100370e-01 8.12218666e+00 5.32417357e-01
-1.40609562e+00 3.03549141e-01 4.75742877e-01 -4.24508482e-01
6.13129549e-02 -3.85739923e-01 -3.48365545e-01 4.52912629e-01
1.74313414e+00 -2.79387355e-01 4.54943269e-01 6.63362622e-01
4.60564345e-01 1.04888782e-01 -7.85372317e-01 1.36702859e+00
3.08838159e-01 -1.00046110e+00 -6.84769869e-01 -1.76982567e-01
2.48086065e-01 4.01692003e-01 1.87966138e-01 3.82443011e-01
-1.67007092e-02 -7.13839948e-01 9.62688029e-01 4.24672335e-01
7.59460628e-01 -6.34112418e-01 7.98357606e-01 -2.36687183e-01
-1.11895072e+00 1.46206662e-01 -3.75656523e-02 -4.11699861e-02
3.40516359e-01 3.95515040e-02 -4.93833184e-01 5.66928625e-01
9.83886123e-01 5.68269730e-01 -1.11087725e-01 8.54478300e-01
8.36094841e-02 7.64128685e-01 -1.29707366e-01 -7.43384361e-02
-1.94513630e-02 -7.87252095e-04 7.87023485e-01 1.65474033e+00
2.97853380e-01 2.18308866e-01 -4.84361082e-01 4.27477837e-01
2.07367375e-01 4.26121205e-01 -6.22323215e-01 -4.13866460e-01
5.37941813e-01 1.41274393e+00 -3.16486269e-01 -1.49274960e-01
-7.83594608e-01 1.00506556e+00 2.98677415e-01 3.79541069e-01
-9.66745138e-01 -2.73354292e-01 1.32372212e+00 -3.03287685e-01
1.35768071e-01 -4.29355860e-01 -2.58650514e-03 -1.07495153e+00
1.45188615e-01 -1.04967999e+00 6.55951679e-01 -1.13046038e+00
-9.92692292e-01 9.52806294e-01 2.32354417e-01 -1.22924042e+00
-6.20859742e-01 -5.09739101e-01 -5.64777315e-01 7.24529326e-01
-1.15518260e+00 -7.37552583e-01 -2.00748742e-01 2.77847290e-01
5.05303383e-01 -1.73294127e-01 9.74336386e-01 5.53924263e-01
-7.37706065e-01 7.27309763e-01 -2.81774193e-01 -2.99591631e-01
6.51032865e-01 -8.20417583e-01 2.32258573e-01 6.21779442e-01
1.96536466e-01 4.63362962e-01 9.44283605e-01 -3.14694673e-01
-1.38195312e+00 -4.25758988e-01 9.16324735e-01 -8.44424665e-01
8.60509157e-01 -4.35994059e-01 -8.66273344e-01 7.24370360e-01
4.77899909e-01 8.61771032e-02 9.36045408e-01 -6.54002577e-02
-1.90728065e-02 2.17942759e-01 -1.27766395e+00 6.08211398e-01
1.09150374e+00 -8.86543393e-01 -5.38768530e-01 -3.46173257e-01
2.89969444e-01 -4.27402049e-01 -1.50558388e+00 3.00495595e-01
9.11572397e-01 -1.28852844e+00 7.36629128e-01 -4.10730720e-01
6.39150068e-02 1.52884930e-01 -2.94945776e-01 -1.27291179e+00
4.98655885e-02 -9.63519335e-01 -4.64426130e-02 1.32256949e+00
2.71768838e-01 -4.84084606e-01 5.83036184e-01 8.98945212e-01
-3.21742177e-01 -7.53545821e-01 -1.37320006e+00 -5.74815214e-01
-8.67085531e-02 -1.28229916e+00 4.10074919e-01 9.45090592e-01
5.67365646e-01 1.27619749e-03 -2.85322458e-01 5.99826612e-02
-1.72017068e-02 -8.16878200e-01 7.49839008e-01 -1.02762580e+00
2.99376715e-02 -6.25876904e-01 -8.26595187e-01 -5.37891269e-01
2.83937782e-01 -4.27717090e-01 -2.37311497e-01 -1.13662088e+00
-8.09257925e-02 9.87694971e-03 -5.99791646e-01 6.89163953e-02
2.47170031e-01 3.35948795e-01 1.64448515e-01 -1.60529967e-02
-8.45751226e-01 6.06919110e-01 6.72168076e-01 4.95350718e-01
-1.76901251e-01 -3.36178690e-01 -5.79173267e-01 4.95635927e-01
4.77095306e-01 -1.41502142e-01 -4.21931148e-01 2.89355163e-02
1.14371963e-01 1.11814931e-01 2.66781807e-01 -1.17549515e+00
1.13840587e-01 -1.79967843e-02 6.41351193e-02 -3.62131953e-01
8.72734427e-01 -8.97719741e-01 4.94275481e-01 -1.77624673e-01
-1.45650312e-01 4.33371156e-01 7.43327737e-01 2.31536746e-01
-5.10700643e-01 -9.71744359e-02 6.16535723e-01 2.52665788e-01
-1.09318578e+00 7.86229745e-02 -8.42294991e-01 1.65807456e-01
9.24131930e-01 -3.29424173e-01 -1.88511446e-01 -8.96848738e-01
-8.89003396e-01 -8.56623128e-02 4.48948056e-01 9.56566751e-01
5.58341980e-01 -1.23584056e+00 -4.61586386e-01 -1.02814831e-01
2.94072747e-01 -8.37313473e-01 5.18924952e-01 1.16506553e+00
-5.94172850e-02 5.16016781e-01 -3.48243266e-01 -3.28984261e-01
-1.60281432e+00 3.02229106e-01 5.42266846e-01 1.30237773e-01
-5.83730757e-01 4.62482512e-01 -2.60446638e-01 -3.31254840e-01
1.84501424e-01 -1.13860868e-01 -2.62417257e-01 1.04642563e-01
6.75719321e-01 5.58149815e-01 3.44508290e-01 -1.19726372e+00
-5.84669411e-01 -3.28591131e-02 3.86782140e-01 -8.20616782e-01
1.27042747e+00 -4.42607880e-01 2.21533045e-01 9.45267677e-01
1.05870986e+00 2.11475849e-01 -9.94969308e-01 3.59223306e-01
1.53998941e-01 -2.69913435e-01 2.27358565e-01 -8.37029159e-01
-6.43373907e-01 6.70880377e-01 8.73262942e-01 2.85663277e-01
1.23951948e+00 -1.27379835e-01 5.00441134e-01 -3.65866721e-02
2.83579439e-01 -1.33886719e+00 -1.96092024e-01 5.49418032e-01
9.39102530e-01 -9.12731290e-01 -4.90204930e-01 -3.29922587e-01
-1.14051735e+00 8.21492434e-01 3.47307891e-01 4.04042482e-01
6.64140463e-01 4.04328555e-01 4.78400886e-01 -2.52488136e-01
-1.05552185e+00 -9.07559041e-03 2.23376602e-01 7.46259570e-01
6.89089954e-01 -1.50025338e-01 -5.35929918e-01 8.86716902e-01
-2.65633464e-01 2.12597564e-01 3.21566224e-01 9.76113319e-01
1.48437191e-02 -1.14314425e+00 -2.93526769e-01 5.87760331e-03
-7.60418892e-01 1.77837014e-01 -6.54564738e-01 1.04237306e+00
-2.60254443e-01 1.20452416e+00 2.30860993e-01 -6.30839765e-01
4.41315860e-01 6.01362586e-01 4.09930348e-01 -6.39124773e-04
-8.70253146e-01 -3.01515132e-01 7.16530502e-01 -6.02277994e-01
-6.18907213e-01 -1.20007753e+00 -9.21124995e-01 -2.47380599e-01
-5.06811514e-02 2.95376360e-01 1.00142038e+00 8.93275082e-01
8.69298220e-01 3.71568143e-01 2.49391824e-01 -6.12366915e-01
-8.08206722e-02 -1.11394358e+00 -5.88336527e-01 5.66829562e-01
8.37556645e-02 -7.83284903e-01 -6.35444283e-01 1.90096740e-02]
|
[13.328653335571289, 5.559164047241211]
|
bf466837-c7d1-4a5a-b7a9-83e5bcf8c516
|
perturb-predict-paraphrase-semi-supervised
| null | null |
https://www.ijcai.org/proceedings/2021/105
|
https://www.ijcai.org/proceedings/2021/0105.pdf
|
Perturb, Predict & Paraphrase: Semi-Supervised Learning using Noisy Student for Image Captioning
|
Recent semi-supervised learning (SSL) methods are predominantly focused on multi-class classification tasks. Classification tasks allow for easy mixing of class labels during augmentation which does not trivially extend to structured outputs such as word sequences that appear in tasks like image captioning. Noisy Student Training is a recent SSL paradigm proposed for image classification that is an extension of self-training and teacher-student learning. In this work, we provide an in-depth analysis of the noisy student SSL framework for the task of image captioning and derive state-of-the-art results. The original algorithm relies on computationally expensive data augmentation steps that involve perturbing the raw images and computing features for each perturbed image. We show that, even in the absence of raw image augmentation, the use of simple model and feature perturbations to the input images for the student model are beneficial to SSL training. We also show how a paraphrase generator could be effectively used for label augmentation to improve the quality of pseudo labels and significantly improve performance. Our final results in the limited labeled data setting (1% of the MS-COCO labeled data) outperform previous state-of-the-art approaches by 2.5 on BLEU4 and 11.5 on CIDEr scores.
|
['Maneesh Singh', 'Deepak Mittal', 'Preethi Jyothi', 'Pranay Reddy Samala', 'Arjit Jain']
|
2021-08-19
| null | null | null |
ijcai-2021-8
|
['semi-supervised-learning-for-image-captioning', 'image-augmentation']
|
['computer-vision', 'computer-vision']
|
[ 1.01178312e+00 4.30200279e-01 -4.29615438e-01 -4.68189836e-01
-1.17334175e+00 -9.62576568e-01 8.80520344e-01 3.08944464e-01
-6.35359108e-01 8.24654222e-01 5.64944185e-02 -5.36201715e-01
4.14726853e-01 -3.45381647e-01 -1.26368749e+00 -6.37681544e-01
3.85852993e-01 6.19383156e-01 5.90544343e-02 -1.78364187e-01
4.05685697e-03 2.17253625e-01 -1.67418337e+00 5.81749797e-01
7.84208417e-01 8.49655867e-01 1.31754950e-01 8.08425188e-01
-3.31334561e-01 8.12085509e-01 -8.83835793e-01 -4.27843511e-01
8.57822150e-02 -7.53153563e-01 -1.06874144e+00 3.93241912e-01
7.52612531e-01 7.03440979e-04 -7.74218738e-02 9.59279418e-01
5.87497234e-01 2.21178755e-02 7.27027893e-01 -1.53040898e+00
-8.42818081e-01 6.64044738e-01 -3.16418946e-01 -1.28304765e-01
2.90332496e-01 1.36418581e-01 8.30593467e-01 -8.85771990e-01
6.63150907e-01 1.06717789e+00 3.98152143e-01 9.99048173e-01
-1.51152182e+00 -6.32865489e-01 -2.55719516e-02 1.95448011e-01
-9.57839012e-01 -3.23144555e-01 5.76742172e-01 -2.98131645e-01
4.56792951e-01 3.98964882e-01 1.83341026e-01 1.42497027e+00
-6.97717667e-01 1.18402934e+00 1.49168968e+00 -9.47849691e-01
1.46825314e-01 5.57252228e-01 1.08652011e-01 4.97924566e-01
8.66471231e-02 -6.54213428e-02 -3.69971961e-01 -1.09037429e-01
3.18009406e-01 -3.11087370e-01 -2.69247115e-01 -5.09160876e-01
-1.48011756e+00 9.65790749e-01 4.39102829e-01 1.39803877e-02
8.38644058e-02 2.35295579e-01 5.45043886e-01 3.89673889e-01
5.89389384e-01 7.95675993e-01 -4.49042022e-01 1.37805082e-02
-9.47280049e-01 9.71557945e-02 5.88120997e-01 1.04114175e+00
7.05760121e-01 2.39824317e-02 -5.45040250e-01 8.05375516e-01
-7.63815120e-02 5.99723577e-01 7.14818537e-01 -9.90879416e-01
3.80331069e-01 3.36086094e-01 1.35329932e-01 -1.54421955e-01
5.20605780e-03 -4.95892406e-01 -7.22182155e-01 6.99141920e-02
5.26298583e-01 -1.24972142e-01 -1.38262463e+00 1.90342987e+00
1.51718616e-01 4.05660808e-01 2.54188955e-01 4.15960848e-01
1.01074326e+00 6.18494272e-01 1.91906169e-01 -1.18248940e-01
1.21748757e+00 -1.43960655e+00 -7.22729146e-01 -4.01923090e-01
1.11794412e+00 -7.62792408e-01 1.45800817e+00 1.63155168e-01
-8.70366454e-01 -6.75432205e-01 -1.00544977e+00 7.47865736e-02
-4.38199669e-01 1.69050872e-01 3.55555654e-01 8.43287408e-01
-1.04233634e+00 4.11650926e-01 -4.18231189e-01 -2.77924895e-01
6.48804486e-01 3.23148996e-01 -4.58183825e-01 -3.67981255e-01
-1.11104274e+00 9.19898748e-01 4.47316289e-01 -4.26357061e-01
-9.67479467e-01 -6.24596715e-01 -1.07029271e+00 -9.87005830e-02
4.51659530e-01 -3.53737712e-01 1.58228564e+00 -1.26592255e+00
-1.32002616e+00 1.14293802e+00 -7.62011111e-02 -5.56347668e-01
5.46623468e-01 -6.94476739e-02 -2.09901817e-02 2.15841323e-01
1.18898809e-01 1.34613359e+00 8.78965378e-01 -1.56684458e+00
-2.86909401e-01 1.47560477e-01 5.55971544e-03 3.04971933e-01
-4.84869301e-01 -2.77238712e-02 -2.51620978e-01 -8.38287234e-01
-2.13275462e-01 -1.28628814e+00 -3.23213607e-01 2.62654573e-02
-3.78143609e-01 -1.98361456e-01 6.99204385e-01 -2.82419592e-01
7.12855816e-01 -2.11287642e+00 -5.54390326e-02 -4.54614572e-02
-2.01065511e-01 6.45924509e-01 -6.58298194e-01 2.73770630e-01
-5.14726162e-01 3.26628715e-01 -4.37470734e-01 -7.45194316e-01
-1.43393725e-01 2.53273845e-01 -4.78710085e-01 3.45127463e-01
4.25304949e-01 1.09912252e+00 -1.05102730e+00 -4.46912944e-01
1.44892752e-01 2.08309054e-01 -3.38180155e-01 4.42540109e-01
-4.91250843e-01 5.77089846e-01 1.52806595e-01 3.28801185e-01
4.75247502e-01 -4.95501757e-01 -1.83882058e-01 -6.84923902e-02
3.47996056e-01 1.83647156e-01 -1.03507054e+00 1.86441386e+00
-5.83480597e-01 6.41017139e-01 -3.24483395e-01 -1.17778099e+00
6.54830217e-01 5.19027054e-01 1.58140719e-01 -4.61077034e-01
-5.01763970e-02 2.77357817e-01 -1.85864776e-01 -3.89317214e-01
1.92453653e-01 -3.20402756e-02 -2.16519877e-01 6.79989815e-01
4.87822980e-01 -4.99222130e-01 3.64409119e-01 3.95315409e-01
7.32235432e-01 1.82072669e-01 1.55970454e-01 5.11605889e-02
5.54736257e-01 3.88372727e-02 2.48695686e-02 1.05423570e+00
-8.54634941e-02 9.72499371e-01 2.84150243e-01 -1.27860501e-01
-1.25955880e+00 -7.83449769e-01 -4.10778224e-02 1.21240258e+00
-5.78665100e-02 -2.02414095e-01 -9.18675601e-01 -1.04660773e+00
-2.77424932e-01 7.64857769e-01 -6.91738844e-01 -2.96367824e-01
-3.00823748e-01 -7.32676625e-01 8.38384509e-01 5.03483772e-01
3.62429947e-01 -1.28658199e+00 -1.73452571e-01 -3.38998586e-02
-2.60629565e-01 -1.43358874e+00 -5.35263062e-01 4.89976555e-01
-6.28626406e-01 -8.28627408e-01 -9.33210194e-01 -1.15305996e+00
9.76842225e-01 2.12317213e-01 1.05597806e+00 7.21541643e-02
-1.28533617e-01 3.02524567e-01 -5.64751625e-01 -4.80043083e-01
-9.50802088e-01 1.73833698e-01 -8.10088068e-02 1.09834233e-02
3.57717350e-02 -3.45660388e-01 -3.21986586e-01 2.14453429e-01
-1.27735662e+00 4.33041602e-01 6.48827612e-01 1.11985159e+00
4.74974930e-01 -5.44743061e-01 6.81921601e-01 -1.26540864e+00
4.08722013e-01 -3.63960713e-01 -3.36676151e-01 4.11027551e-01
-5.37752807e-01 4.10191566e-01 6.86737657e-01 -1.01329648e+00
-7.00946808e-01 4.48679745e-01 -1.90813228e-01 -4.07791972e-01
-4.33032721e-01 2.53639460e-01 -1.83409154e-02 -1.55174658e-01
8.95565808e-01 4.69315112e-01 5.56849986e-02 -3.65471780e-01
8.53938580e-01 9.31323647e-01 7.66851306e-01 -5.03021300e-01
8.98275673e-01 2.99803317e-01 -1.03675582e-01 -4.94148493e-01
-1.32408857e+00 -4.69151914e-01 -5.12195945e-01 3.17210183e-02
7.51410007e-01 -1.06447232e+00 -2.94024110e-01 4.86826718e-01
-1.07018650e+00 -5.27412951e-01 -6.72049046e-01 2.19732195e-01
-7.35784233e-01 3.48158509e-01 -5.56305230e-01 -4.95284528e-01
-2.39991948e-01 -1.35924327e+00 1.10671949e+00 9.47641134e-02
-1.80381790e-01 -7.78693914e-01 -5.94258569e-02 7.35327005e-01
2.89620161e-01 1.52129848e-02 8.06032062e-01 -1.14996338e+00
-3.42241555e-01 -1.25369698e-01 -1.67190447e-01 7.81474650e-01
1.23439230e-01 -5.63576937e-01 -1.20830560e+00 -4.16866750e-01
-1.99668735e-01 -1.13523674e+00 8.43495548e-01 -8.78932625e-02
1.35285568e+00 -3.27845931e-01 -7.05389306e-02 4.50649798e-01
1.22168946e+00 -1.38398752e-01 4.76327777e-01 3.83843839e-01
8.23767602e-01 6.46035135e-01 5.76072037e-01 -2.21783239e-02
9.85753909e-02 6.35681748e-01 4.37177628e-01 -3.71014416e-01
-5.40022731e-01 -3.44931841e-01 2.82014281e-01 4.73177373e-01
6.32716954e-01 -4.87257510e-01 -7.66605556e-01 7.93671489e-01
-1.65063560e+00 -6.97670102e-01 -1.30731821e-01 2.11477494e+00
1.44138896e+00 1.29664853e-01 4.45851423e-02 3.26473027e-01
8.19598317e-01 6.19814778e-03 -4.83294159e-01 -3.18579167e-01
-1.79942384e-01 4.08748806e-01 6.96310759e-01 4.32925612e-01
-1.29230905e+00 1.01783752e+00 6.08592129e+00 9.91507769e-01
-8.86001348e-01 2.59380639e-01 8.93352628e-01 9.62552149e-04
-2.46990412e-01 -1.46083161e-01 -7.26023853e-01 4.86842453e-01
1.17212701e+00 2.04605818e-01 2.32634932e-01 8.45005155e-01
-1.42767012e-01 -9.12491977e-02 -1.28412700e+00 1.07352614e+00
3.92589748e-01 -1.28244877e+00 2.23082006e-01 -1.42662644e-01
1.22812176e+00 6.01490028e-03 2.91938215e-01 2.69542545e-01
3.52945566e-01 -1.12097073e+00 6.69355035e-01 -9.81246978e-02
1.11741519e+00 -6.18410289e-01 7.74897814e-01 4.25724953e-01
-4.22342449e-01 8.04097205e-02 4.36137244e-02 1.21832646e-01
1.23513430e-01 4.05864060e-01 -1.08057523e+00 1.43614665e-01
3.22212815e-01 3.20271522e-01 -9.81505752e-01 1.08248210e+00
-5.11845767e-01 9.85083699e-01 -2.53399253e-01 3.24327163e-02
4.74689037e-01 2.49305576e-01 1.69130191e-01 1.35947275e+00
7.17543885e-02 -2.70730019e-01 3.42852116e-01 3.74136627e-01
-5.46669602e-01 6.01065382e-02 -6.17651820e-01 -5.93638867e-02
4.20922428e-01 1.03471518e+00 -6.89496696e-01 -6.88220322e-01
-3.68222803e-01 1.26721668e+00 2.70137310e-01 2.11675599e-01
-7.06489325e-01 -3.97571266e-01 2.90132105e-01 -4.59363796e-02
3.46862793e-01 6.01763576e-02 -3.61953497e-01 -1.08577132e+00
-7.22272992e-02 -1.19282138e+00 2.00841919e-01 -8.51955652e-01
-1.14065933e+00 6.28247023e-01 -3.08474395e-02 -1.28123379e+00
-4.84742641e-01 -3.67314786e-01 -3.47318947e-01 6.81603074e-01
-1.63976979e+00 -1.29441202e+00 -2.71226108e-01 4.60263729e-01
7.45544910e-01 -6.22810647e-02 1.07043529e+00 1.48783654e-01
-2.81412601e-01 9.50571001e-01 3.72503519e-01 2.36096025e-01
1.10161805e+00 -1.44512177e+00 5.61186790e-01 8.33265722e-01
5.32942772e-01 1.19153723e-01 8.99202108e-01 -4.68861103e-01
-7.32158303e-01 -1.33460569e+00 9.21218157e-01 -6.08319044e-01
5.81965983e-01 -6.64702892e-01 -8.76148701e-01 6.56210065e-01
4.40752149e-01 4.14445877e-01 7.53585041e-01 -3.94088715e-01
-5.55991650e-01 2.35390842e-01 -1.17328501e+00 6.22620881e-01
8.67695153e-01 -6.52149677e-01 -5.02700567e-01 7.88941443e-01
1.04578781e+00 -4.06857550e-01 -3.72690827e-01 3.39416862e-01
5.84775843e-02 -3.01415145e-01 9.69876051e-01 -8.43072176e-01
4.86830115e-01 -2.74203539e-01 -7.83205032e-03 -1.46156549e+00
1.08916141e-01 -7.08918333e-01 9.33714025e-03 1.28943515e+00
6.05772734e-01 -9.17323381e-02 9.68475342e-01 3.63138497e-01
-5.62307946e-02 -5.79287946e-01 -6.76769495e-01 -9.04382050e-01
3.23923081e-02 -3.87255520e-01 2.33982772e-01 9.77170706e-01
-1.92039341e-01 6.35360956e-01 -4.87848818e-01 -1.66811664e-02
7.98672795e-01 -6.37285709e-02 7.29168773e-01 -9.69142139e-01
-2.84741312e-01 -3.30779776e-02 -1.73135206e-01 -1.00777912e+00
5.14751613e-01 -1.11766791e+00 3.08745146e-01 -1.34003711e+00
2.46404052e-01 -5.23715675e-01 -1.64590910e-01 7.28829503e-01
-5.09287834e-01 8.48148227e-01 3.16864014e-01 9.48435068e-02
-7.98686862e-01 4.47165579e-01 1.17642260e+00 -2.76553869e-01
3.02576777e-02 6.63806871e-02 -6.94985032e-01 4.04590011e-01
7.87913084e-01 -7.94101000e-01 -6.71566069e-01 -3.86162758e-01
-3.59872244e-02 -2.58231729e-01 2.33161956e-01 -9.15199041e-01
5.78988232e-02 5.72869629e-02 8.12393576e-02 -2.85514593e-01
2.96792686e-01 -7.39041269e-01 -3.00721169e-01 6.07453167e-01
-9.78793919e-01 -1.06961206e-01 2.03848034e-01 5.17464101e-01
-1.55605167e-01 -7.25543857e-01 1.02272618e+00 -1.89678282e-01
-4.57608074e-01 4.15932275e-02 -3.23458135e-01 2.52658129e-01
1.07782686e+00 2.41167974e-02 -3.23795617e-01 -6.65168405e-01
-7.41960466e-01 2.29715884e-01 4.71963704e-01 4.47705448e-01
3.58216137e-01 -1.39364219e+00 -8.45703781e-01 8.12371001e-02
4.15056527e-01 8.89286175e-02 -1.74437165e-01 3.16232383e-01
-3.34285527e-01 3.63190889e-01 -7.71047175e-02 -6.79734826e-01
-1.46088994e+00 7.37492502e-01 -9.96190356e-04 -4.34488237e-01
-1.61440283e-01 9.99039292e-01 1.19696900e-01 -6.45720601e-01
5.31636953e-01 -9.49321985e-02 -9.96206254e-02 2.07416505e-01
5.66522956e-01 -9.92503986e-02 1.88029051e-01 -5.08824944e-01
-1.02172650e-01 2.03849614e-01 -2.74596632e-01 -3.80511820e-01
1.33243561e+00 -9.57990624e-03 2.47196287e-01 4.39627707e-01
1.46509171e+00 -3.42678726e-01 -1.16317046e+00 -4.60498154e-01
6.10323809e-02 -1.56832859e-01 -1.97664350e-01 -1.04433787e+00
-6.73955262e-01 9.19672966e-01 7.48527765e-01 2.26584747e-02
9.70367849e-01 2.50820339e-01 7.18697786e-01 5.68044245e-01
2.22590882e-02 -7.92156100e-01 5.04848719e-01 3.94189894e-01
6.16554856e-01 -1.60578847e+00 -2.57726789e-01 -4.84311342e-01
-8.28553796e-01 6.48553848e-01 6.06793821e-01 1.04457080e-01
1.85550600e-01 1.21834613e-01 2.66398489e-01 3.26998174e-01
-6.56291723e-01 -3.64146471e-01 2.52544045e-01 7.23563254e-01
3.64070863e-01 -2.84155071e-01 -4.30298112e-02 1.31043524e-01
-1.24150097e-01 -1.31653786e-01 8.02083433e-01 1.05124140e+00
-3.11427534e-01 -1.53388512e+00 -3.28914881e-01 3.65115464e-01
-5.22048056e-01 -3.60499859e-01 -3.51856112e-01 4.85885620e-01
6.02001511e-02 8.74186635e-01 1.09523144e-02 -1.04101107e-01
1.89271316e-01 4.99296665e-01 5.39398789e-01 -1.03175068e+00
-5.69948196e-01 -2.57784128e-01 1.05847783e-01 -1.47446334e-01
-6.45162821e-01 -5.50398409e-01 -9.18794155e-01 2.82662779e-01
-5.00359118e-01 3.03720385e-01 8.99784982e-01 9.62271988e-01
1.76421270e-01 2.68059999e-01 7.31916547e-01 -7.55676508e-01
-7.35378027e-01 -1.10061955e+00 -7.75989145e-02 8.33402634e-01
5.57998180e-01 -3.73621553e-01 -4.77594078e-01 4.47147012e-01]
|
[10.954536437988281, 0.6306782960891724]
|
4ed9a996-06a0-4887-a783-34526184b94d
|
learning-typographic-style
|
1603.04000
| null |
http://arxiv.org/abs/1603.04000v1
|
http://arxiv.org/pdf/1603.04000v1.pdf
|
Learning Typographic Style
|
Typography is a ubiquitous art form that affects our understanding,
perception, and trust in what we read. Thousands of different font-faces have
been created with enormous variations in the characters. In this paper, we
learn the style of a font by analyzing a small subset of only four letters.
From these four letters, we learn two tasks. The first is a discrimination
task: given the four letters and a new candidate letter, does the new letter
belong to the same font? Second, given the four basis letters, can we generate
all of the other letters with the same characteristics as those in the basis
set? We use deep neural networks to address both tasks, quantitatively and
qualitatively measure the results in a variety of novel manners, and present a
thorough investigation of the weaknesses and strengths of the approach.
|
['Shumeet Baluja']
|
2016-03-13
| null | null | null | null |
['font-recognition']
|
['computer-vision']
|
[ 1.97066367e-01 -2.92534649e-01 1.13430262e-01 -3.68169338e-01
2.92783901e-02 -1.11224079e+00 5.91126800e-01 -1.44531325e-01
-1.69089988e-01 8.45541418e-01 2.14845687e-01 -3.38637561e-01
1.00785188e-01 -7.68911064e-01 -9.17915642e-01 -3.97723913e-01
6.22826278e-01 4.53858942e-01 1.85150683e-01 -4.19237733e-01
7.08937228e-01 6.02141798e-01 -1.75816250e+00 8.61246347e-01
7.54645944e-01 9.57698107e-01 -4.04530466e-02 5.84883630e-01
-2.17273191e-01 4.31020677e-01 -9.84134257e-01 -7.24104822e-01
5.28523088e-01 -4.58914757e-01 -6.10740066e-01 1.93659931e-01
8.68766904e-01 -3.49844575e-01 -2.31819481e-01 1.05913067e+00
2.36213163e-01 -9.38050300e-02 8.23509991e-01 -6.16516769e-01
-1.44581735e+00 5.10840356e-01 -2.64275551e-01 -1.13232709e-01
5.72816253e-01 3.17814112e-01 8.14694524e-01 -1.06026936e+00
7.34821141e-01 1.27661347e+00 4.48590785e-01 6.07315600e-01
-1.32804441e+00 -5.16423643e-01 7.27042602e-03 -1.80190429e-01
-9.90215421e-01 -1.65056825e-01 6.77225649e-01 -7.52743781e-01
5.93559265e-01 3.20732862e-01 8.37727666e-01 1.51971316e+00
2.81362087e-01 3.78459185e-01 1.79859984e+00 -7.98740268e-01
5.09315059e-02 5.14946282e-01 2.43174464e-01 3.66752446e-01
4.34979379e-01 1.45830318e-01 -6.74743116e-01 1.11334108e-01
7.85156131e-01 -2.47320771e-01 -4.00391407e-02 -2.80210655e-02
-1.31325316e+00 5.04435599e-01 2.68997401e-01 1.99023098e-01
2.17265487e-02 -1.28616422e-01 -9.46930200e-02 6.13493264e-01
7.32202679e-02 1.10649657e+00 -2.71533459e-01 -3.44532579e-02
-5.29205263e-01 2.28355497e-01 1.08542073e+00 1.01139843e+00
7.30678916e-01 -4.86176573e-02 -3.14318627e-01 8.67188156e-01
-2.51201212e-01 8.87513697e-01 1.37400523e-01 -7.03572273e-01
3.94134462e-01 5.03500462e-01 2.31099352e-01 -9.09056962e-01
-5.60117066e-02 -2.87777573e-01 -6.31618321e-01 7.94155717e-01
7.38646567e-01 -1.89651147e-01 -9.81624067e-01 1.50077510e+00
-2.73540229e-01 -5.93809724e-01 -3.94707084e-01 6.91841424e-01
5.83283722e-01 4.13948387e-01 -4.21230078e-01 1.66088969e-01
1.21127605e+00 -7.62672424e-01 -3.96881759e-01 -2.50897318e-01
1.39378324e-01 -1.21580100e+00 1.51292539e+00 7.28131115e-01
-1.07753313e+00 -8.54049861e-01 -1.32628810e+00 -9.01888013e-02
-7.45121598e-01 4.55025613e-01 4.18745488e-01 7.64910460e-01
-1.04260516e+00 7.77413428e-01 2.11527914e-01 -2.05508575e-01
4.06281412e-01 1.57796219e-01 -2.18280151e-01 8.40078071e-02
-1.04062867e+00 1.05259597e+00 1.58035278e-01 -1.65793762e-01
-3.82253766e-01 -3.90818417e-01 -1.64385930e-01 7.31394812e-02
7.37430155e-02 -6.18252993e-01 1.01212060e+00 -1.45263922e+00
-1.62896073e+00 1.22582126e+00 -2.21055210e-01 6.33261278e-02
8.66907001e-01 -1.54111311e-01 -7.66167164e-01 -4.04263914e-01
-9.79505945e-03 5.99977791e-01 1.20013332e+00 -1.35349739e+00
-5.13246059e-01 -1.51465431e-01 -3.16525698e-02 -1.17700733e-01
-4.05564338e-01 5.44868633e-02 -1.88378915e-01 -8.98639619e-01
-1.13125481e-02 -1.01341331e+00 5.41780531e-01 2.65308976e-01
-5.63454032e-01 2.32051928e-02 2.48538643e-01 -4.75667417e-01
1.16034377e+00 -2.29750204e+00 -9.65038091e-02 3.88111264e-01
2.49760807e-01 2.12256372e-01 -2.05382094e-01 3.96581680e-01
6.82209954e-02 3.42313141e-01 2.91012317e-01 1.77602857e-01
3.01423073e-02 -1.83015808e-01 -6.56341493e-01 -1.28481090e-01
4.09659952e-01 6.85692668e-01 -6.99655712e-01 1.88967690e-01
-2.43867263e-01 1.47600293e-01 -3.02954108e-01 1.06030054e-01
-3.78764451e-01 2.71118999e-01 -3.04591693e-02 4.82184857e-01
6.64174378e-01 -2.40156874e-01 1.50081038e-01 1.13329124e-02
-2.15075016e-01 2.65520900e-01 -1.06436491e+00 9.38291550e-01
-2.40854517e-01 1.12221372e+00 -6.96690440e-01 -1.31384015e-01
1.35288000e+00 -2.11435571e-01 -4.94024009e-01 -7.95399785e-01
1.44028753e-01 5.19498348e-01 3.83991241e-01 -4.17228371e-01
6.32077813e-01 -6.53056428e-02 7.11226165e-02 7.89542377e-01
-3.96661937e-01 -2.46308550e-01 3.01172197e-01 -2.58053780e-01
8.83554935e-01 -7.58182630e-02 9.57130194e-02 -1.82945505e-01
2.20558047e-01 -2.67059058e-01 1.99074760e-01 1.18528104e+00
3.62714194e-03 8.39538872e-01 7.56039262e-01 -9.31895971e-01
-1.50421798e+00 -1.36350942e+00 -1.58374861e-01 9.52791214e-01
-2.28598285e-02 -1.10197723e-01 -5.57138562e-01 -3.86827856e-01
3.54668558e-01 9.75889444e-01 -1.03775358e+00 -5.80879375e-02
-4.37986910e-01 -2.11583108e-01 3.90260607e-01 6.24695182e-01
2.78644443e-01 -1.35678029e+00 -5.16193926e-01 -1.33457914e-01
3.20871532e-01 -8.41309905e-01 -5.87815642e-01 1.60546362e-01
-6.00986838e-01 -9.06951904e-01 -5.73312759e-01 -8.67888272e-01
9.04325783e-01 7.11072013e-02 1.33426595e+00 2.42582485e-01
1.45610310e-02 -1.20601416e-01 -2.84082025e-01 -7.12819576e-01
-7.64526784e-01 -1.38305917e-01 1.11224376e-01 3.64224762e-02
5.98062575e-01 -2.69759625e-01 -3.72869551e-01 3.84485304e-01
-6.81250393e-01 2.45505646e-01 7.66753852e-01 8.09160590e-01
2.92404741e-01 -1.53930470e-01 3.49683881e-01 -1.10567927e+00
9.31114852e-01 -7.56498650e-02 -3.97843122e-01 4.56943035e-01
-4.86414135e-01 1.34388253e-01 9.94354248e-01 -7.15869844e-01
-6.55329466e-01 -1.69814304e-01 3.13876033e-01 6.87830476e-03
-1.71930596e-01 1.46267131e-01 -1.10356763e-01 2.28451014e-01
9.26431715e-01 1.28527880e-01 -6.63683340e-02 -5.31040549e-01
3.99200499e-01 6.81819677e-01 4.98953044e-01 -7.76093483e-01
9.17552173e-01 3.32866848e-01 -1.20566711e-01 -5.48300624e-01
-7.85683453e-01 4.70401525e-01 -9.80357111e-01 -2.72992373e-01
6.92850113e-01 -4.83819664e-01 -6.77339435e-01 6.20345712e-01
-1.41095400e+00 -1.46492735e-01 -4.97759879e-01 8.32224488e-02
-4.54043269e-01 1.07952200e-01 -1.48126706e-01 -5.58660209e-01
4.97737713e-02 -1.14049840e+00 6.84782803e-01 3.10556561e-01
-6.82388604e-01 -6.43484175e-01 -1.44759431e-01 -1.75564408e-01
2.38929704e-01 1.78248867e-01 1.59566021e+00 -6.03771091e-01
-6.91447675e-01 -3.63752753e-01 -3.91876519e-01 5.87465167e-01
4.94034439e-01 4.51767027e-01 -1.05687535e+00 -2.52677530e-01
-2.75992036e-01 -3.25224608e-01 7.93085456e-01 2.84522194e-02
1.38832057e+00 -1.57453299e-01 3.47105786e-02 7.88709521e-01
1.10218060e+00 6.26747608e-01 6.16154253e-01 5.62145650e-01
6.54149413e-01 6.12485170e-01 2.52683666e-02 2.94918686e-01
-2.20561787e-01 4.26221907e-01 -1.85764775e-01 1.95241034e-01
-3.19052935e-01 -4.21422124e-01 2.59021759e-01 8.03877473e-01
-1.29608124e-01 -4.85625744e-01 -7.72486389e-01 1.69476405e-01
-1.32292998e+00 -8.55713367e-01 6.99523762e-02 2.42509055e+00
1.04376495e+00 5.74831724e-01 3.77071351e-02 1.88810006e-02
9.32194114e-01 1.69280142e-01 -7.73737192e-01 -1.01718080e+00
-7.76727021e-01 3.37862968e-01 3.73037495e-02 1.13569051e-01
-8.41694117e-01 8.20317090e-01 7.62118196e+00 3.13246459e-01
-1.20344198e+00 -6.68199241e-01 9.53151762e-01 1.79967321e-02
-6.15413904e-01 1.00075088e-01 -8.49716365e-01 6.50153816e-01
5.58830082e-01 -9.15295407e-02 1.00945783e+00 4.87669945e-01
1.82012860e-02 -1.24907140e-02 -1.72465670e+00 8.54146719e-01
3.15142751e-01 -1.45805395e+00 6.50802910e-01 -9.36876610e-02
1.18842041e+00 -5.81574440e-01 6.34477198e-01 -8.79492611e-02
2.45819017e-01 -1.39437509e+00 1.09981108e+00 1.01303303e+00
1.32662570e+00 -3.94728035e-01 2.31344193e-01 -1.80510730e-02
-3.78087401e-01 -1.71483710e-01 -4.23531950e-01 -4.86317635e-01
-5.39355338e-01 4.82635677e-01 -5.98489523e-01 -7.21897632e-02
7.54560351e-01 4.57930803e-01 -1.05475831e+00 8.96993935e-01
-4.50548053e-01 3.33141059e-01 4.24278975e-02 -4.56872553e-01
-7.96290264e-02 -2.87361711e-01 4.80107307e-01 9.79569614e-01
4.74637121e-01 -2.86913276e-01 -3.25135797e-01 1.35539532e+00
-5.43060660e-01 -6.37957081e-02 -5.92577696e-01 -2.70541072e-01
6.52546525e-01 8.04432988e-01 -5.90546429e-01 -3.92288476e-01
-4.96876746e-01 9.73866224e-01 4.65371013e-01 3.68651301e-01
-2.80592233e-01 -5.82118750e-01 5.68443894e-01 2.59977549e-01
3.28090727e-01 -4.69773263e-02 -1.06703079e+00 -9.31140482e-01
4.18471634e-01 -1.40798330e+00 -4.70155664e-02 -1.20994961e+00
-1.85453737e+00 7.33817577e-01 -3.84528846e-01 -1.48385429e+00
1.45360738e-01 -1.14073074e+00 -9.30548191e-01 1.26190460e+00
-9.20709372e-01 -8.32078516e-01 -4.02278274e-01 -4.31390293e-03
3.85147154e-01 -6.24266028e-01 6.89754605e-01 -3.25484276e-01
-2.30851725e-01 8.30613852e-01 6.82842851e-01 2.94427246e-01
1.12904716e+00 -1.34779561e+00 9.04156089e-01 5.66884935e-01
2.17087448e-01 9.15074170e-01 5.75560749e-01 -7.51899958e-01
-1.07316566e+00 -7.25842774e-01 9.99814987e-01 -1.03101134e+00
6.87940121e-01 -5.37422776e-01 -9.08449054e-01 4.31184649e-01
2.29123533e-01 -4.53286618e-01 5.27847826e-01 6.03979416e-02
-9.41485107e-01 1.07523920e-02 -7.37700105e-01 9.02551711e-01
9.88406062e-01 -6.82581544e-01 -8.04451764e-01 1.84664607e-01
5.11399806e-01 -2.61386305e-01 -4.28497404e-01 -1.35075107e-01
1.14881110e+00 -1.23852408e+00 6.08628631e-01 -8.91705751e-01
1.10265040e+00 -2.81497985e-01 -7.34362975e-02 -1.69181955e+00
-6.97982430e-01 -5.19230902e-01 2.57603645e-01 1.05648184e+00
6.38028562e-01 -7.54728079e-01 6.07255816e-01 4.93851423e-01
1.02142937e-01 -8.61322641e-01 -4.79910284e-01 -1.01884437e+00
6.53239489e-01 7.86460117e-02 1.11923921e+00 6.91509962e-01
-2.98566878e-01 2.82632530e-01 -3.56396675e-01 -2.65986055e-01
2.18932614e-01 3.96918774e-01 7.46092618e-01 -1.58154392e+00
-3.85469973e-01 -6.73509657e-01 -7.30918795e-02 -9.32618439e-01
-2.57846545e-02 -8.63412559e-01 -1.50099248e-01 -1.12241840e+00
2.16868937e-01 -1.27423033e-01 -1.60165399e-01 2.12023467e-01
-3.37107718e-01 1.52967945e-01 5.50839126e-01 3.12436700e-01
-4.94446270e-02 1.74769819e-01 1.47398579e+00 -3.77772897e-01
-1.13173388e-01 3.74174044e-02 -8.76455247e-01 6.76752925e-01
7.02775180e-01 -1.74867809e-01 -7.79695287e-02 -8.03386152e-01
8.41070950e-01 -4.99834329e-01 2.36937344e-01 -1.03313065e+00
-4.01674360e-02 -2.04326183e-01 1.22639728e+00 -4.54343438e-01
1.48191378e-01 -6.99426889e-01 -2.29683816e-02 2.74563819e-01
-7.11418509e-01 5.58841944e-01 1.56033039e-01 4.88317788e-01
2.09949449e-01 -4.56332684e-01 8.68543863e-01 -7.62050748e-02
-5.21763802e-01 -2.29678348e-01 -4.59533691e-01 2.19579130e-01
6.68978214e-01 -6.78837836e-01 -8.04116428e-01 -3.06370884e-01
-5.93799651e-01 -3.96598011e-01 8.92093778e-01 6.47251248e-01
5.29992938e-01 -1.41643119e+00 -6.96607053e-01 5.89650095e-01
3.79943639e-01 -5.46846151e-01 -3.66974860e-01 1.54506490e-01
-5.75404525e-01 2.22210184e-01 -7.45993435e-01 -1.28337368e-01
-8.84074628e-01 5.86606622e-01 1.85947254e-01 3.29682380e-01
-3.22869092e-01 7.90087819e-01 2.19816849e-01 -9.60208550e-02
1.50589412e-02 -3.74639124e-01 -1.39209226e-01 2.46056821e-02
4.00571644e-01 3.40737611e-01 -7.83009455e-02 -3.74514401e-01
8.68618023e-03 7.79412150e-01 -3.39840770e-01 -8.20031166e-02
1.07740581e+00 1.88923776e-01 -2.19814166e-01 9.51729834e-01
9.00716126e-01 2.91559070e-01 -1.28663814e+00 -2.75150895e-01
-1.01009943e-02 -9.57412422e-01 -6.47295594e-01 -1.34424329e+00
-5.31238019e-01 8.36479366e-01 5.73867202e-01 2.13860452e-01
8.26416016e-01 -3.27247858e-01 5.56219697e-01 6.95327640e-01
2.92382002e-01 -1.39933670e+00 3.05805713e-01 8.34453464e-01
1.10963237e+00 -1.19570899e+00 6.57259375e-02 -2.40286842e-01
-7.26819515e-01 1.61832964e+00 8.45163584e-01 -1.54093489e-01
4.03176099e-01 -1.79546289e-02 3.76694918e-01 -2.55311392e-02
-7.63194025e-01 1.68462053e-01 5.27174771e-01 6.48691773e-01
6.45708919e-01 1.87505990e-01 -1.39067456e-01 3.30064625e-01
-6.83314264e-01 -1.33813575e-01 9.11812305e-01 6.17822886e-01
-4.64578897e-01 -1.08705997e+00 -4.97136623e-01 8.61895919e-01
1.61434442e-01 -2.28733063e-01 -1.15610933e+00 4.90055144e-01
2.88118899e-01 6.46303177e-01 1.97852761e-01 -7.58110344e-01
4.49573666e-01 2.89589643e-01 5.69082677e-01 -6.21365786e-01
-6.72063172e-01 -4.52484876e-01 -6.92349747e-02 1.03163682e-01
2.77219087e-01 -6.72395349e-01 -7.12454319e-01 -6.09772205e-01
-2.24222131e-02 -3.37208927e-01 4.98154074e-01 7.45855212e-01
2.74611086e-01 3.63487571e-01 8.81621957e-01 -4.86642480e-01
-8.30758810e-01 -7.02498257e-01 -6.16672575e-01 6.76255286e-01
3.12258005e-01 -5.92494726e-01 -4.26415324e-01 1.18048184e-01]
|
[11.702119827270508, -0.11591404676437378]
|
6403ac25-c4a5-445f-9bc2-4ad57c61fd93
|
self-attention-between-datapoints-going
|
2106.02584
| null |
https://arxiv.org/abs/2106.02584v2
|
https://arxiv.org/pdf/2106.02584v2.pdf
|
Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning
|
We challenge a common assumption underlying most supervised deep learning: that a model makes a prediction depending only on its parameters and the features of a single input. To this end, we introduce a general-purpose deep learning architecture that takes as input the entire dataset instead of processing one datapoint at a time. Our approach uses self-attention to reason about relationships between datapoints explicitly, which can be seen as realizing non-parametric models using parametric attention mechanisms. However, unlike conventional non-parametric models, we let the model learn end-to-end from the data how to make use of other datapoints for prediction. Empirically, our models solve cross-datapoint lookup and complex reasoning tasks unsolvable by traditional deep learning models. We show highly competitive results on tabular data, early results on CIFAR-10, and give insight into how the model makes use of the interactions between points.
|
['Yarin Gal', 'Tom Rainforth', 'Aidan N. Gomez', 'Clare Lyle', 'Neil Band', 'Jannik Kossen']
|
2021-06-04
| null |
http://proceedings.neurips.cc/paper/2021/hash/f1507aba9fc82ffa7cc7373c58f8a613-Abstract.html
|
http://proceedings.neurips.cc/paper/2021/file/f1507aba9fc82ffa7cc7373c58f8a613-Paper.pdf
|
neurips-2021-12
|
['3d-part-segmentation']
|
['computer-vision']
|
[-2.70672202e-01 4.63633120e-01 -5.20128965e-01 -9.43902552e-01
-6.87030733e-01 -5.99923968e-01 6.20853484e-01 3.29383820e-01
-4.30927008e-01 4.83169556e-01 2.82671511e-01 -5.01495540e-01
-2.46090636e-01 -9.18129444e-01 -1.34611726e+00 -2.10989550e-01
-4.93854806e-02 1.29777682e+00 4.39954847e-02 -9.23142508e-02
3.40627402e-01 4.91251051e-01 -1.22880101e+00 8.28239143e-01
4.54738140e-01 1.00914490e+00 9.75115132e-03 6.69902027e-01
-3.11416388e-01 1.11339259e+00 -4.74043459e-01 -6.53428614e-01
2.72749532e-02 1.70045406e-01 -1.09064066e+00 -2.86230147e-01
6.27833307e-01 -5.52003741e-01 -3.41692775e-01 6.93451405e-01
-6.54414743e-02 2.30782166e-01 9.59143937e-01 -1.47175121e+00
-9.89689708e-01 8.36618900e-01 -3.60183388e-01 2.56372094e-01
5.17041907e-02 2.64342636e-01 1.40995276e+00 -7.88613319e-01
3.23168486e-01 1.52489328e+00 7.82924235e-01 3.99350673e-01
-1.53951669e+00 -5.18809140e-01 4.53591436e-01 3.84822786e-01
-9.20472562e-01 -4.12337720e-01 6.38409853e-01 -4.96628851e-01
1.22784984e+00 -7.08627626e-02 2.95031875e-01 9.52273190e-01
1.99392915e-01 9.54179108e-01 4.66481000e-01 -2.70424306e-01
2.43290439e-01 -2.34028324e-01 5.88785529e-01 6.79405391e-01
-5.08974604e-02 2.97011193e-02 -4.66958731e-01 -1.91198587e-01
7.46445358e-01 1.38375074e-01 -3.57196964e-02 -5.31073570e-01
-1.18800831e+00 9.03432071e-01 7.13230610e-01 -5.01972064e-02
-3.44582498e-01 4.92408782e-01 3.71132374e-01 4.35867608e-01
7.66161680e-02 7.69183755e-01 -1.00047576e+00 -1.11903615e-01
-7.44003892e-01 5.02072215e-01 8.50655258e-01 8.39866519e-01
8.44384968e-01 -2.32886955e-01 -2.03390002e-01 5.88039279e-01
2.80845821e-01 1.77879930e-01 4.84685838e-01 -1.10363901e+00
5.49441576e-01 5.01184940e-01 2.40608931e-01 -6.18147433e-01
-6.44867182e-01 -2.51134872e-01 -4.94679749e-01 2.79030323e-01
6.55836821e-01 -9.59896669e-02 -9.55315590e-01 1.79761040e+00
-8.69512632e-02 1.90084837e-02 1.04300238e-01 8.14927638e-01
5.37356913e-01 8.82634103e-01 3.42665583e-01 3.96957308e-01
9.71290350e-01 -1.06572831e+00 -3.68045956e-01 -3.61653954e-01
8.98227692e-01 -3.10623407e-01 1.34561694e+00 5.40711999e-01
-1.29822707e+00 -7.35180676e-01 -9.45312679e-01 -8.63990545e-01
-5.56506515e-01 6.06026053e-02 7.57260919e-01 -2.07496006e-02
-9.94648457e-01 6.94652498e-01 -9.72621322e-01 -1.45360410e-01
5.96583307e-01 6.31115675e-01 -2.00012520e-01 -4.26917039e-02
-1.05282688e+00 1.00166273e+00 4.94899184e-01 2.21599340e-02
-8.15464914e-01 -1.19312978e+00 -6.87390804e-01 5.16565263e-01
3.70347381e-01 -9.04215515e-01 1.69739819e+00 -1.12684655e+00
-1.34971642e+00 7.34257817e-01 -3.22183132e-01 -8.86220515e-01
3.26823622e-01 -4.51972634e-01 -9.36737880e-02 -2.10802659e-01
-8.70362073e-02 9.75820780e-01 6.86374068e-01 -1.14017820e+00
-3.99358183e-01 -3.21735650e-01 4.39294726e-01 4.67043594e-02
2.44750813e-01 -1.30534381e-01 -4.03322846e-01 -3.66894394e-01
-1.40902489e-01 -7.63425887e-01 -1.41710520e-01 2.27715135e-01
-5.84665120e-01 -5.60086906e-01 7.05342054e-01 -4.16802973e-01
6.20809376e-01 -2.19651270e+00 1.60419494e-01 2.39700988e-01
2.26050660e-01 1.14404730e-01 -3.53983998e-01 4.07663852e-01
-4.79790032e-01 1.92336440e-01 -3.77346091e-02 -3.58848751e-01
3.54539305e-01 3.83722752e-01 -8.06563318e-01 2.53620118e-01
4.58452761e-01 1.34982455e+00 -6.61770642e-01 -1.32314920e-01
1.23040676e-01 2.09106892e-01 -8.97113085e-01 1.73138574e-01
-8.96533787e-01 -1.66213721e-01 -4.20302749e-01 2.45413825e-01
3.82876515e-01 -4.71713781e-01 3.00741438e-02 -2.01103210e-01
1.91583365e-01 5.76179564e-01 -8.51773620e-01 1.67962313e+00
-5.51245749e-01 6.56611502e-01 -3.44949722e-01 -1.10106480e+00
6.87382638e-01 -5.59838079e-02 -4.93763434e-03 -5.84515929e-01
-9.23482776e-02 -9.01855752e-02 1.37013763e-01 -3.39311182e-01
2.98816353e-01 -1.44701153e-01 -6.03088662e-02 6.41221464e-01
1.09705195e-01 1.44406587e-01 -1.12083688e-01 2.53801674e-01
8.88918221e-01 1.78610772e-01 1.41345203e-01 -2.32201248e-01
1.09990589e-01 2.24745229e-01 2.71304756e-01 9.36231434e-01
3.84451687e-01 5.17044246e-01 1.07639098e+00 -1.06507897e+00
-1.10408330e+00 -1.10883248e+00 9.24846604e-02 1.51229692e+00
-1.47535160e-01 -5.06168485e-01 -5.13683498e-01 -7.16128290e-01
3.16859096e-01 1.10262275e+00 -1.10208058e+00 -2.86525756e-01
-4.85173017e-01 -3.76576364e-01 3.51138741e-01 9.79226887e-01
2.06601977e-01 -1.19185174e+00 -5.16763449e-01 1.30181849e-01
6.95124045e-02 -7.46160328e-01 -2.17444822e-01 5.82784891e-01
-1.00732923e+00 -1.14203858e+00 -6.19376171e-03 -6.60122871e-01
6.53845966e-01 -1.52388245e-01 1.52019703e+00 1.74798042e-01
2.91814599e-02 9.92306471e-02 8.98051187e-02 -4.67381477e-01
-1.83842584e-01 4.28852826e-01 -2.30454519e-01 -2.71608293e-01
5.94757020e-01 -4.04576689e-01 -2.46582493e-01 1.29321039e-01
-7.09692001e-01 1.05415113e-01 7.44779229e-01 9.67058718e-01
6.78389251e-01 -1.21418431e-01 5.82822621e-01 -1.38684046e+00
4.57129508e-01 -8.10819685e-01 -5.89450538e-01 4.21180338e-01
-2.55049050e-01 6.47809148e-01 8.73303831e-01 -3.90653253e-01
-8.03435504e-01 4.25087400e-02 1.16564535e-01 -7.45128334e-01
-3.08640629e-01 6.02461338e-01 -2.65611857e-01 6.19829953e-01
5.94434083e-01 1.00413356e-02 6.77313581e-02 -6.04940653e-01
6.28910422e-01 2.12937325e-01 7.42884159e-01 -1.00977230e+00
6.34767890e-01 2.91602224e-01 -3.79093997e-02 -2.16136381e-01
-1.16845107e+00 2.30508000e-02 -8.34621012e-01 4.74663585e-01
9.19329762e-01 -9.07481670e-01 -1.06554198e+00 1.65982455e-01
-1.37808514e+00 -9.56144571e-01 -4.50165302e-01 1.40865341e-01
-7.67684579e-01 -3.35439771e-01 -4.88477588e-01 -3.69548470e-01
-8.20156559e-02 -9.93321061e-01 8.91765714e-01 -2.62510508e-01
-4.60416496e-01 -1.32815874e+00 -1.17280930e-01 2.87760735e-01
3.18640351e-01 -1.73280835e-01 1.65562010e+00 -1.12271547e+00
-7.32442796e-01 -1.08828709e-01 -3.74017447e-01 -3.64493532e-03
-3.18942785e-01 2.86458656e-02 -8.76406252e-01 2.44609509e-02
-3.25130135e-01 -6.13621354e-01 8.85269105e-01 3.07463795e-01
2.07709670e+00 -4.74547863e-01 -3.47835571e-01 7.97649384e-01
1.19387817e+00 -1.00944191e-03 6.07460856e-01 2.81319469e-01
7.45113671e-01 5.03279686e-01 3.37985307e-01 2.14670613e-01
7.19775319e-01 4.42386240e-01 6.50126636e-01 3.32035534e-02
1.54851422e-01 -5.47206879e-01 -4.11503110e-03 -4.94510382e-02
1.87449664e-01 -1.63006321e-01 -1.13926387e+00 5.06345272e-01
-2.01225924e+00 -9.95304525e-01 7.52073247e-03 2.04586744e+00
7.43294477e-01 3.35305780e-01 7.62718096e-02 -1.90935656e-01
3.78688157e-01 2.20984463e-02 -1.01539516e+00 -6.06446981e-01
2.53217518e-01 2.52413720e-01 4.73266840e-01 6.73082292e-01
-1.09797919e+00 9.57231104e-01 7.09216309e+00 4.31423038e-01
-1.21230531e+00 -2.41925240e-01 9.70464528e-01 -3.99098784e-01
-4.70575571e-01 -1.16772249e-01 -6.78100944e-01 2.53665566e-01
1.14484787e+00 6.33517653e-02 6.53286397e-01 8.46823335e-01
-1.17316417e-01 7.28439018e-02 -2.03848362e+00 7.70793080e-01
-1.41723111e-01 -1.56610310e+00 3.74739081e-01 7.98535496e-02
3.36652696e-01 2.28401899e-01 2.98748940e-01 6.48714423e-01
8.18916380e-01 -1.53397942e+00 7.41757572e-01 7.55538881e-01
4.65206683e-01 -8.38312626e-01 4.33956534e-01 5.20385146e-01
-3.91278923e-01 -3.63680124e-01 -5.61212897e-01 -3.06875408e-01
-2.53281325e-01 3.03513497e-01 -8.57284188e-01 1.03604756e-01
5.51161766e-01 7.91744530e-01 -5.55777848e-01 8.10057759e-01
-2.66686559e-01 5.24520814e-01 -2.50542521e-01 2.41752923e-01
4.08669502e-01 2.26271734e-01 -1.14822343e-01 8.72854352e-01
1.15291424e-01 1.33371845e-01 -2.10928749e-02 1.11152518e+00
-3.92805815e-01 -2.56502599e-01 -5.38460314e-01 1.24082446e-01
3.85834664e-01 7.64623225e-01 -1.42721191e-01 -5.02227306e-01
-5.55512130e-01 6.05748951e-01 1.10658658e+00 6.02029860e-01
-7.92286038e-01 -9.58220363e-02 7.46423900e-01 3.04649383e-01
5.72484493e-01 -9.26158726e-02 -5.77001750e-01 -1.15548170e+00
-1.23210378e-01 -7.14114189e-01 5.59195757e-01 -1.24386978e+00
-1.51393139e+00 2.39131987e-01 1.54022664e-01 -6.57288611e-01
-4.56399143e-01 -8.73920441e-01 -8.32817376e-01 9.79665697e-01
-1.56022310e+00 -1.08637500e+00 1.09247722e-01 6.04430437e-01
4.89805192e-01 -2.63160076e-02 9.23343539e-01 -2.80568272e-01
-3.56755465e-01 6.84561610e-01 -7.01364782e-03 3.71154040e-01
6.35521591e-01 -1.44232678e+00 8.34864497e-01 2.75515139e-01
4.02113736e-01 9.30804074e-01 5.79837561e-01 -2.70780683e-01
-1.38829803e+00 -1.05646086e+00 1.09246004e+00 -7.44856536e-01
8.32263768e-01 -4.24443543e-01 -1.27280223e+00 1.44248605e+00
1.71847120e-01 3.49091828e-01 7.29527652e-01 7.52673745e-01
-9.08959568e-01 -1.49192065e-01 -9.45681214e-01 4.65563029e-01
8.00513983e-01 -6.32458389e-01 -1.05700767e+00 2.49948785e-01
7.66273975e-01 -3.60932916e-01 -6.70350075e-01 9.96196270e-02
6.05944574e-01 -8.51960063e-01 1.07508266e+00 -1.39181650e+00
8.08127284e-01 8.94031301e-02 -1.47501573e-01 -1.52472949e+00
-5.67621648e-01 -3.47725600e-01 -3.79383922e-01 7.02812552e-01
7.05123484e-01 -5.16912282e-01 9.00188327e-01 1.22600591e+00
-1.83681041e-01 -8.96592438e-01 -6.32654786e-01 -3.48598510e-01
6.02759540e-01 -4.93003726e-01 8.98019552e-01 9.68772650e-01
-8.23324621e-02 5.35556078e-01 -3.69910500e-03 3.62167358e-01
3.70708317e-01 5.18021286e-01 7.93164253e-01 -1.40542126e+00
-5.23886859e-01 -4.74519640e-01 -1.91616222e-01 -1.31113207e+00
6.13261402e-01 -8.91321838e-01 -8.79783407e-02 -1.48049760e+00
1.37734637e-01 -5.83874106e-01 -3.66491735e-01 1.00945175e+00
-1.10877633e-01 -9.09307897e-02 2.42831126e-01 2.29125887e-01
-7.47086167e-01 3.27865303e-01 1.03202081e+00 -1.19582266e-01
-3.66395339e-02 -1.19918525e-01 -9.29192841e-01 7.13079810e-01
8.36185873e-01 -4.02961224e-01 -7.72048533e-01 -1.04329360e+00
5.07189155e-01 2.31085211e-01 4.57201421e-01 -8.18029404e-01
3.48421305e-01 -3.36995333e-01 7.65106201e-01 -7.12704301e-01
4.73724484e-01 -9.57566798e-01 -1.60118327e-01 1.65637091e-01
-9.93474543e-01 2.29370713e-01 3.54985744e-01 4.78128433e-01
-1.41976878e-01 -7.76443556e-02 5.11139214e-01 -9.46855471e-02
-6.96638346e-01 4.59044188e-01 -1.39810577e-01 1.87121585e-01
8.97104502e-01 1.79797247e-01 -6.22179568e-01 -4.30605292e-01
-9.67916250e-01 4.31945801e-01 5.08707643e-01 3.89534622e-01
4.21512455e-01 -1.13869953e+00 -3.82506728e-01 4.21770722e-01
1.00766271e-01 3.92721027e-01 4.72586602e-02 4.22117949e-01
-3.46565783e-01 5.91635406e-01 -3.47454906e-01 -3.69518846e-01
-6.23255372e-01 1.07710803e+00 4.95935380e-01 -2.15136632e-01
-3.97739083e-01 7.31877029e-01 5.41474223e-01 -5.93848884e-01
2.88229793e-01 -8.42419624e-01 9.16293636e-02 -1.66532367e-01
4.35596198e-01 -1.19049117e-01 3.33560295e-02 -1.78777784e-01
-1.97864935e-01 2.48422071e-01 -4.89620984e-01 -2.61667762e-02
1.37165892e+00 3.09911728e-01 -5.93691990e-02 8.00364792e-01
1.38621771e+00 -4.21638966e-01 -1.64101756e+00 -3.41350943e-01
-1.75835360e-02 -2.36586288e-01 -7.80016035e-02 -1.21373582e+00
-9.20207858e-01 1.27440619e+00 -2.76012897e-01 -5.12497984e-02
7.61004269e-01 2.16982260e-01 5.14970005e-01 9.93822575e-01
1.14666717e-02 -6.45181000e-01 -1.51521405e-02 6.84436262e-01
9.24092472e-01 -1.28038120e+00 -1.65405333e-01 -2.48538069e-02
-6.36896908e-01 1.29568565e+00 7.91015148e-01 -3.81435841e-01
7.18514264e-01 2.40019321e-01 -5.89151233e-02 -2.23662823e-01
-1.53295195e+00 2.75571108e-01 2.98297584e-01 5.18612146e-01
2.74769425e-01 -2.37392616e-02 5.46986818e-01 7.69699752e-01
-3.90366703e-01 2.32512042e-01 2.73816645e-01 5.98219573e-01
-3.30890894e-01 -1.00958931e+00 -9.40794572e-02 7.27299452e-01
-3.83712143e-01 -1.86555520e-01 -3.02617669e-01 9.89227712e-01
-1.63257346e-01 3.91936511e-01 7.16564119e-01 -1.35141894e-01
2.17285499e-01 4.03762341e-01 3.15869540e-01 -7.14717984e-01
-5.27734578e-01 -4.84271377e-01 -5.73637113e-02 -6.61613762e-01
5.73816411e-02 -6.31116390e-01 -1.49152744e+00 -4.11823422e-01
4.56400782e-01 1.43853739e-01 6.02744102e-01 1.01118433e+00
5.26414931e-01 4.40744519e-01 2.28666365e-01 -6.09483600e-01
-8.31954181e-01 -7.84732282e-01 -1.77295446e-01 3.80614430e-01
7.27361858e-01 -4.30106908e-01 -9.68098640e-02 1.36045426e-01]
|
[9.609108924865723, 7.069088935852051]
|
db86d175-eed7-4500-9d99-7d63994bf250
|
robot-motion-planning-as-video-prediction-a
|
2208.11287
| null |
https://arxiv.org/abs/2208.11287v1
|
https://arxiv.org/pdf/2208.11287v1.pdf
|
Robot Motion Planning as Video Prediction: A Spatio-Temporal Neural Network-based Motion Planner
|
Neural network (NN)-based methods have emerged as an attractive approach for robot motion planning due to strong learning capabilities of NN models and their inherently high parallelism. Despite the current development in this direction, the efficient capture and processing of important sequential and spatial information, in a direct and simultaneous way, is still relatively under-explored. To overcome the challenge and unlock the potentials of neural networks for motion planning tasks, in this paper, we propose STP-Net, an end-to-end learning framework that can fully extract and leverage important spatio-temporal information to form an efficient neural motion planner. By interpreting the movement of the robot as a video clip, robot motion planning is transformed to a video prediction task that can be performed by STP-Net in both spatially and temporally efficient ways. Empirical evaluations across different seen and unseen environments show that, with nearly 100% accuracy (aka, success rate), STP-Net demonstrates very promising performance with respect to both planning speed and path cost. Compared with existing NN-based motion planners, STP-Net achieves at least 5x, 2.6x and 1.8x faster speed with lower path cost on 2D Random Forest, 2D Maze and 3D Random Forest environments, respectively. Furthermore, STP-Net can quickly and simultaneously compute multiple near-optimal paths in multi-robot motion planning tasks
|
['Bo Yuan', 'Saman Zonouz', 'Jingjin Yu', 'Lingyi Huang', 'Miao Yin', 'Xiao Zang']
|
2022-08-24
| null | null | null | null |
['video-prediction']
|
['computer-vision']
|
[ 3.55304420e-01 -1.23943590e-01 -4.53833908e-01 3.95314917e-02
-4.41187859e-01 -2.97751933e-01 4.83113468e-01 -9.72704738e-02
-8.13069642e-01 5.80305755e-01 8.35453570e-02 -3.61244380e-01
-3.71223003e-01 -8.93640161e-01 -5.95345497e-01 -5.33599377e-01
-4.01383966e-01 7.11395919e-01 5.62110960e-01 -1.76712275e-01
2.78802842e-01 7.92922378e-01 -1.32187104e+00 -5.05516976e-02
4.58803236e-01 7.68132627e-01 8.09667289e-01 8.76367092e-01
7.72486255e-02 7.57174909e-01 1.07408516e-01 4.17055309e-01
3.36037725e-01 2.72190161e-02 -1.07680595e+00 -2.05355600e-01
8.30846280e-03 -3.59845519e-01 -7.07564950e-01 5.65549612e-01
4.33214515e-01 5.68483531e-01 5.85012734e-01 -1.22729182e+00
8.63041803e-02 2.50979275e-01 -6.81169748e-01 -1.82965305e-03
2.75554985e-01 4.94962633e-01 7.59800971e-01 -7.60394394e-01
9.29610491e-01 1.26551056e+00 5.86359203e-01 5.07300735e-01
-9.11600888e-01 -3.94156307e-01 3.64806876e-02 4.69845384e-01
-1.10309196e+00 -2.68699139e-01 4.69655216e-01 -1.29327755e-02
1.52284193e+00 -1.84508756e-01 6.83334172e-01 1.07429516e+00
6.82284772e-01 1.03646767e+00 4.79838848e-01 -2.90303946e-01
3.21525276e-01 -8.47144961e-01 -2.66113222e-01 9.13366556e-01
-9.78288502e-02 4.91685748e-01 -4.08686787e-01 2.66872853e-01
1.10728395e+00 2.55451016e-02 -8.08511749e-02 -4.74489540e-01
-1.61555505e+00 6.19586587e-01 7.76948094e-01 7.11231828e-02
-6.93377614e-01 7.17387676e-01 3.91361028e-01 -4.09747437e-02
-2.13735759e-01 3.90546292e-01 -4.00462806e-01 -5.30145705e-01
-7.83772647e-01 4.80590433e-01 5.95514536e-01 1.10104728e+00
6.62895262e-01 1.86906725e-01 1.36043951e-01 6.60827935e-01
2.42212534e-01 5.30808270e-01 4.16202307e-01 -1.46014428e+00
6.46392584e-01 3.92538428e-01 1.74821481e-01 -1.34638715e+00
-1.08140743e+00 5.06318547e-02 -1.07893288e+00 4.33613151e-01
3.63821030e-01 -1.15414262e-01 -1.02978563e+00 1.39783287e+00
3.26495320e-01 1.27468079e-01 2.87127584e-01 8.92199099e-01
3.69684041e-01 1.02921796e+00 6.46988824e-02 2.30461303e-02
1.10335171e+00 -1.68252182e+00 -4.13256943e-01 -7.41841614e-01
6.15258276e-01 -3.44145477e-01 6.53880954e-01 3.20823699e-01
-1.12487483e+00 -5.06249011e-01 -1.00449204e+00 -9.53291804e-02
-4.91966382e-02 -5.44246808e-02 6.90275669e-01 -6.45944998e-02
-1.10016608e+00 7.94289589e-01 -1.31690192e+00 -5.78862607e-01
5.86062372e-01 6.42506540e-01 -5.44218898e-01 -6.62174761e-01
-6.58314586e-01 9.54389989e-01 8.57647777e-01 2.34276518e-01
-1.07210159e+00 -8.81837085e-02 -9.04653072e-01 -3.79275940e-02
6.47797525e-01 -7.66067445e-01 1.36391604e+00 -2.88928598e-01
-1.70247126e+00 1.71210170e-02 -2.17497453e-01 -6.07905447e-01
3.37600470e-01 -3.38537037e-01 -3.78414318e-02 3.11866283e-01
9.97162536e-02 1.35631430e+00 3.85339767e-01 -8.80786955e-01
-1.08285129e+00 -1.79284424e-01 -9.82918367e-02 6.29062772e-01
5.44918468e-03 -2.25956768e-01 -7.49072254e-01 -3.00028145e-01
4.27974820e-01 -1.27774227e+00 -1.04538870e+00 2.44808212e-01
-2.23754495e-01 -2.06709534e-01 9.87701535e-01 -2.75068939e-01
7.24888504e-01 -1.71759117e+00 3.20755959e-01 -1.29984403e-02
6.19726628e-02 3.05453539e-01 -3.90176237e-01 5.76194704e-01
5.14184356e-01 -2.47201502e-01 -4.09320951e-01 -1.93722934e-01
-1.20093130e-01 6.77755117e-01 -5.02072535e-02 3.33088100e-01
5.07197231e-02 1.11835003e+00 -1.26474655e+00 -4.64618683e-01
4.57848817e-01 2.16808513e-01 -3.70921165e-01 -8.24185684e-02
-3.97564232e-01 3.24184328e-01 -4.86297011e-01 6.47369146e-01
2.64345437e-01 -7.62828887e-02 3.03873926e-01 3.37828040e-01
-2.89105773e-01 -5.42708822e-02 -1.00664723e+00 2.10790443e+00
-4.76816297e-01 9.37034607e-01 6.95901364e-02 -9.03737128e-01
8.11725557e-01 2.35202804e-01 6.26769781e-01 -9.73430574e-01
-1.31326258e-01 3.28884184e-01 -1.63485050e-01 -4.95949686e-01
9.32816803e-01 1.94324419e-01 -1.87057287e-01 3.56061280e-01
-1.87554330e-01 -7.08021522e-02 9.97395217e-02 -2.83829778e-01
1.62447643e+00 5.53072631e-01 4.82738227e-01 1.69733644e-01
2.18359515e-01 7.13436067e-01 5.53684533e-01 7.45989740e-01
-3.87192547e-01 4.74969834e-01 -1.91841628e-02 -7.75239408e-01
-1.05737615e+00 -8.63766193e-01 5.21613836e-01 9.92059112e-01
5.28442085e-01 -1.19232669e-01 -2.56157905e-01 -3.36555660e-01
-3.69105935e-01 4.15161937e-01 -6.34016022e-02 5.54521456e-02
-1.40591121e+00 -5.15919507e-01 5.78050375e-01 9.37393785e-01
7.55162597e-01 -1.66622293e+00 -1.36558664e+00 6.91204011e-01
-2.75093317e-01 -1.41890693e+00 -2.02337146e-01 3.29922616e-01
-1.13668156e+00 -8.77094746e-01 -5.50135612e-01 -1.12313628e+00
4.81506616e-01 7.54827797e-01 7.64582932e-01 -1.56780660e-01
-2.42621154e-01 9.66455564e-02 -2.78437436e-01 -8.62709433e-03
-1.39317930e-01 2.75663733e-01 -5.22053763e-02 -6.76955998e-01
-1.03057995e-01 -7.94597983e-01 -5.81282735e-01 3.84185731e-01
-5.26693642e-01 3.55064243e-01 1.08282542e+00 7.70507514e-01
7.39528477e-01 3.48869771e-01 2.55618274e-01 -2.12318316e-01
2.81964600e-01 -5.79547405e-01 -6.31312907e-01 -1.83832318e-01
-5.89517593e-01 9.52119082e-02 6.28014922e-01 -3.82934511e-01
-8.97461832e-01 7.01454818e-01 -9.44506153e-02 -3.95897925e-01
-3.57292086e-01 7.20914781e-01 1.48134440e-01 -4.08915877e-02
4.58447546e-01 2.78220385e-01 -1.16501264e-01 -9.46646705e-02
4.18448329e-01 7.38582239e-02 9.49318886e-01 -3.64907920e-01
7.11812675e-01 6.19853616e-01 5.16548336e-01 -9.01633561e-01
-5.57602979e-02 -6.94254398e-01 -9.08100009e-01 -2.34370977e-01
9.40255940e-01 -6.84288919e-01 -7.99768627e-01 5.75939596e-01
-1.40636206e+00 -8.05873573e-01 3.22937891e-02 5.33905327e-01
-1.02546918e+00 3.89093131e-01 -5.24643183e-01 -7.47414708e-01
-3.21175367e-01 -1.12839413e+00 8.73291790e-01 2.04725996e-01
-3.44477385e-01 -9.04730856e-01 7.68327713e-02 4.76422533e-02
3.54919136e-01 5.01116872e-01 8.08100581e-01 -2.89498210e-01
-1.06610060e+00 -2.69606352e-01 -3.66303742e-01 -5.32660663e-01
-2.13889644e-01 -2.15500161e-01 -4.38058645e-01 -1.09590173e-01
-3.57746422e-01 -2.06908986e-01 8.15795004e-01 6.81072950e-01
7.83444881e-01 -3.66023451e-01 -8.27759683e-01 6.03818476e-01
1.64133334e+00 4.91837591e-01 6.20821536e-01 7.56294250e-01
7.38602757e-01 6.57388628e-01 9.21133578e-01 3.10304493e-01
4.28133160e-01 9.03986812e-01 1.08157134e+00 2.25982934e-01
5.59633933e-02 -2.77379632e-01 4.61221814e-01 4.97431755e-01
-1.58984050e-01 -5.10080993e-01 -1.26306677e+00 7.77762771e-01
-2.30826306e+00 -1.09792280e+00 -2.21691430e-01 1.81848145e+00
1.63371399e-01 2.34214708e-01 7.69957453e-02 1.24517448e-01
4.76482779e-01 4.19217318e-01 -5.62129855e-01 -3.20988744e-01
1.21486701e-01 -8.66054818e-02 8.24079692e-01 4.41143632e-01
-1.08859110e+00 1.14476740e+00 6.32428312e+00 8.63170147e-01
-9.92262661e-01 -2.03355208e-01 6.00464828e-02 -1.67157829e-01
2.37828866e-01 -4.40235361e-02 -6.32232904e-01 -7.93694332e-03
9.63064194e-01 1.98433682e-01 4.50936496e-01 1.11953378e+00
3.86559606e-01 -3.83398533e-01 -8.66844535e-01 8.26463044e-01
-3.90937716e-01 -1.55387425e+00 -2.80248672e-01 3.27612281e-01
6.29393458e-01 5.41981220e-01 -1.17667824e-01 2.70645201e-01
5.77344894e-01 -1.33703673e+00 6.08082592e-01 1.12627402e-01
4.74648744e-01 -9.91413593e-01 7.61258900e-01 8.77007961e-01
-1.61325824e+00 -4.32961106e-01 -4.58025187e-01 -4.01255578e-01
7.32246935e-01 1.79348364e-01 -1.16114879e+00 7.56834805e-01
6.30567551e-01 7.93493152e-01 -8.35018419e-03 1.20169270e+00
-2.49391690e-01 1.17015034e-01 -5.03652215e-01 -2.23970532e-01
9.81659830e-01 2.43579168e-02 5.96818149e-01 1.12809837e+00
4.43346947e-01 1.41090453e-01 5.34258306e-01 4.08104241e-01
2.72993535e-01 -2.98989624e-01 -9.09905493e-01 1.43992066e-01
5.11036694e-01 1.15461779e+00 -1.12093592e+00 6.10478781e-02
-1.78938076e-01 1.01605940e+00 4.64479208e-01 2.69123197e-01
-7.73699045e-01 -3.45705003e-01 4.62827861e-01 -2.87376851e-01
4.98353630e-01 -9.01574135e-01 -3.64036888e-01 -3.09314758e-01
7.16330037e-02 -3.79353642e-01 -1.94697957e-02 -6.87384784e-01
-4.77964580e-01 6.31679833e-01 1.89918280e-02 -1.27524722e+00
-6.19200706e-01 -8.77092123e-01 -6.53540909e-01 4.93064135e-01
-1.58171308e+00 -1.10572028e+00 -4.67587292e-01 3.74250412e-01
1.00629139e+00 -1.74624726e-01 8.89602304e-01 -1.09607302e-01
-4.82233554e-01 -9.90919676e-03 2.04242721e-01 -3.03607304e-02
1.58234090e-01 -9.32545066e-01 8.05259168e-01 7.84954667e-01
-4.32290323e-02 2.67192554e-02 4.96244937e-01 -6.37843668e-01
-1.60101080e+00 -1.25933480e+00 7.43192434e-01 -3.04236233e-01
3.23059112e-01 2.02061698e-01 -5.32369375e-01 4.77070212e-01
3.63735706e-02 -2.21543126e-02 3.40489894e-02 -3.40756088e-01
1.64247945e-01 2.54026085e-01 -8.94160688e-01 9.87654090e-01
1.45994163e+00 1.32687703e-01 -2.28861257e-01 2.80314118e-01
7.78619707e-01 -6.43413603e-01 -5.62251866e-01 5.12763202e-01
7.88891673e-01 -8.55723441e-01 1.17164671e+00 -2.69495636e-01
6.73634350e-01 -1.67830899e-01 -2.78970897e-01 -9.76867676e-01
-5.31018019e-01 -7.10094154e-01 -2.27215722e-01 4.62544113e-01
3.10546577e-01 -1.73386365e-01 1.10241902e+00 2.63866335e-01
-4.49703723e-01 -1.01679766e+00 -1.13298535e+00 -9.08287704e-01
-2.09274724e-01 -8.36424351e-01 2.68634796e-01 3.27723861e-01
3.36789563e-02 2.64273316e-01 -5.79484165e-01 2.68833846e-01
3.08795929e-01 1.75572298e-02 8.24975610e-01 -1.01149666e+00
5.19437082e-02 -6.10475779e-01 -4.19059306e-01 -1.57860041e+00
2.69789904e-01 -6.72877908e-01 7.09347844e-01 -2.00854135e+00
-2.13499442e-01 -4.66029495e-01 7.65586123e-02 5.76288521e-01
2.83540606e-01 -1.24745734e-01 2.73563981e-01 4.39629614e-01
-5.82499385e-01 5.89510560e-01 1.30768538e+00 -6.56259730e-02
-4.11182761e-01 5.54121211e-02 -7.92840309e-03 9.18725073e-01
8.16495538e-01 -4.41874862e-01 -5.03744304e-01 -7.69732475e-01
-3.37038934e-02 6.27690136e-01 2.57900387e-01 -1.51320612e+00
7.83577561e-01 -7.22806513e-01 2.01054558e-01 -1.02872252e+00
7.50547826e-01 -9.51887131e-01 1.12688495e-02 8.85494053e-01
-6.54161423e-02 5.75059533e-01 3.55815202e-01 9.86450851e-01
-4.80754673e-02 -3.04860383e-01 4.38730806e-01 -3.72476965e-01
-1.48207092e+00 5.07083833e-01 -7.76832879e-01 -3.70867848e-01
1.15538919e+00 -6.20874941e-01 -1.78750366e-01 -2.56172270e-01
-3.78800809e-01 5.34356177e-01 3.58246177e-01 4.17362303e-01
1.07387972e+00 -1.29719818e+00 -3.85029107e-01 -1.67663336e-01
-2.03822836e-01 6.33152246e-01 3.09378743e-01 7.51792312e-01
-1.00143874e+00 8.27453196e-01 -5.63905716e-01 -7.31812835e-01
-1.03128159e+00 5.51371992e-01 -2.70301849e-02 -6.63556755e-01
-9.69640851e-01 5.91974020e-01 -4.22780290e-02 -4.47425336e-01
2.40730807e-01 -3.34182352e-01 -2.72110224e-01 -3.96402538e-01
3.45216751e-01 8.48102510e-01 -3.35675657e-01 -4.76875752e-01
-3.75584930e-01 5.69325864e-01 2.51854390e-01 -4.50752765e-01
1.46653974e+00 -1.28788069e-01 1.55039191e-01 6.23742566e-02
9.05986309e-01 -5.23624837e-01 -1.62952983e+00 -5.96094355e-02
3.13765377e-01 -1.58219665e-01 -1.17292441e-01 -5.98691463e-01
-8.57689381e-01 8.81832600e-01 2.24435598e-01 -3.48686635e-01
9.67907667e-01 -3.86043996e-01 1.22984147e+00 8.26939404e-01
7.98443615e-01 -1.08488381e+00 3.51370215e-01 1.13028014e+00
6.76980913e-01 -1.13241899e+00 4.08935212e-02 -2.67775476e-01
-5.60348451e-01 1.37320435e+00 8.01453948e-01 -1.62997365e-01
2.10431054e-01 1.73550740e-01 -2.23335341e-01 -4.95281667e-02
-8.90453517e-01 -1.45305758e-02 -9.56403688e-02 1.00987697e+00
-2.87742198e-01 -1.13095000e-01 1.95457742e-01 1.12157486e-01
-3.74646881e-03 4.82291691e-02 5.19138694e-01 1.29706419e+00
-8.53971839e-01 -8.53128374e-01 -1.44782439e-01 -6.97411690e-03
1.44504458e-01 2.88513839e-01 1.23904854e-01 1.06859493e+00
-2.32844770e-01 8.17845464e-01 -3.27310376e-02 -5.30425251e-01
2.29440600e-01 -2.20732585e-01 2.89932191e-01 -5.74523687e-01
-1.78666189e-01 -1.52454153e-01 2.60491371e-01 -1.01238632e+00
-4.75594193e-01 -5.31277061e-01 -1.84943128e+00 -2.85652429e-01
6.73992634e-02 -3.92545462e-01 6.76073849e-01 1.14658570e+00
4.46748078e-01 5.20823658e-01 2.45210528e-01 -1.62505281e+00
-3.63210976e-01 -6.08899415e-01 -1.01890136e-02 -3.93281221e-01
2.82313168e-01 -5.59208155e-01 2.23482773e-01 -1.62854299e-01]
|
[4.784753799438477, 0.9893090128898621]
|
0c1faeac-f992-4d61-8db5-4b474b7d4949
|
joint-architecture-and-knowledge-distillation
|
1912.07806
| null |
https://arxiv.org/abs/1912.07806v3
|
https://arxiv.org/pdf/1912.07806v3.pdf
|
Joint Architecture and Knowledge Distillation in CNN for Chinese Text Recognition
|
The technique of distillation helps transform cumbersome neural network into compact network so that the model can be deployed on alternative hardware devices. The main advantages of distillation based approaches include simple training process, supported by most off-the-shelf deep learning softwares and no special requirement of hardwares. In this paper, we propose a guideline to distill the architecture and knowledge of pre-trained standard CNNs simultaneously. We first make a quantitative analysis of the baseline network, including computational cost and storage overhead in different components. And then, according to the analysis results, optional strategies can be adopted to the compression of fully-connected layers. For vanilla convolution layers, the proposed parsimonious convolution (ParConv) block only consisting of depthwise separable convolution and pointwise convolution is used as a direct replacement without other adjustments such as the widths and depths in the network. Finally, the knowledge distillation with multiple losses is adopted to improve performance of the compact CNN. The proposed algorithm is first verified on offline handwritten Chinese text recognition (HCTR) where the CNNs are characterized by tens of thousands of output nodes and trained by hundreds of millions of training samples. Compared with the CNN in the state-of-the-art system, our proposed joint architecture and knowledge distillation can reduce the computational cost by >10x and model size by >8x with negligible accuracy loss. And then, by conducting experiments on one of the most popular data sets: MNIST, we demonstrate the proposed approach can also be successfully applied on mainstream backbone networks.
|
['Zi-Rui Wang', 'Jun Du']
|
2019-12-17
| null | null | null | null |
['handwritten-chinese-text-recognition', 'handwritten-chinese-text-recognition']
|
['computer-vision', 'natural-language-processing']
|
[ 5.15772998e-02 1.92324072e-01 7.23632611e-03 -3.70453447e-01
1.80701584e-01 -2.86345184e-01 2.11487651e-01 -2.65496224e-01
-9.86118734e-01 7.33834684e-01 -4.98854101e-01 -6.98945165e-01
-1.17146738e-01 -8.67644668e-01 -8.58444095e-01 -7.44644523e-01
1.72123119e-01 8.91079679e-02 3.95202935e-01 -5.72921708e-02
9.20309722e-02 7.76340306e-01 -1.28781688e+00 1.00905806e-01
8.02648783e-01 1.44449091e+00 6.18756771e-01 4.34136540e-01
-7.56313428e-02 9.59572732e-01 -6.38575256e-01 -8.88568640e-01
5.15942514e-01 -6.25497177e-02 -7.36972630e-01 -1.33676305e-01
2.73211867e-01 -9.19566393e-01 -8.85202348e-01 9.75454628e-01
4.98924106e-01 5.43081760e-02 1.81942940e-01 -7.92844236e-01
-5.51493585e-01 1.09925270e+00 -2.58904010e-01 1.01817697e-01
-7.26602018e-01 7.77500272e-02 6.05556965e-01 -8.13072562e-01
1.82642400e-01 9.75080788e-01 6.95985436e-01 5.02829671e-01
-6.45620942e-01 -1.01045108e+00 1.95022047e-01 3.40824157e-01
-1.50418937e+00 -4.43287551e-01 6.41240656e-01 9.76170599e-02
1.31513417e+00 5.77137545e-02 6.27071440e-01 6.38945103e-01
-1.65698543e-01 7.18619466e-01 5.81265330e-01 -4.30474997e-01
1.43293232e-01 2.50375897e-01 2.84205705e-01 9.88585293e-01
7.42740870e-01 -2.60125905e-01 -1.74086064e-01 2.12102741e-01
8.20917845e-01 4.08447325e-01 -2.80278504e-01 2.30123416e-01
-9.07059491e-01 5.04815757e-01 7.71422863e-01 2.58755028e-01
-2.04164878e-01 4.72938478e-01 5.51865280e-01 2.14000866e-01
1.09688438e-01 1.83236703e-01 -7.57688582e-01 -1.15672961e-01
-1.38517594e+00 -1.39671966e-01 6.63347781e-01 1.32688892e+00
8.18794668e-01 4.30785835e-01 2.18363896e-01 8.95675778e-01
4.22657542e-02 2.42049888e-01 8.40609550e-01 -6.11577570e-01
7.60260940e-01 8.03104639e-01 -4.84362155e-01 -7.33845890e-01
-2.29342282e-01 -6.85152233e-01 -1.43210304e+00 -8.80797673e-03
2.17895076e-01 -3.75180095e-01 -9.64768589e-01 1.17581201e+00
-1.95000723e-01 2.88276404e-01 1.35030717e-01 6.74121499e-01
6.30867064e-01 5.30139208e-01 -2.73521423e-01 2.55893618e-02
1.38154495e+00 -1.30733001e+00 -5.48443675e-01 -1.64164945e-01
8.34231019e-01 -6.26055241e-01 9.11038756e-01 5.38277388e-01
-1.17451739e+00 -7.51981735e-01 -1.51006711e+00 -3.52950126e-01
-4.49010611e-01 7.99108624e-01 6.21998131e-01 6.28700674e-01
-8.89908910e-01 1.04184175e+00 -9.86431837e-01 5.13525419e-02
8.38507712e-01 6.57616019e-01 -2.32859626e-01 -2.05298603e-01
-9.85230446e-01 7.60929763e-01 1.09423280e+00 5.49728274e-01
-7.70078242e-01 -6.70614719e-01 -5.05830646e-01 5.51390469e-01
2.42071912e-01 -4.43104297e-01 1.17628980e+00 -9.73068595e-01
-1.71306515e+00 3.44553202e-01 2.39593938e-01 -8.36771846e-01
5.14568567e-01 -3.33545834e-01 -2.03246430e-01 2.31480628e-01
-8.37372422e-01 6.21520638e-01 5.88129461e-01 -6.33812785e-01
-6.93922579e-01 -1.07587511e-02 2.43928626e-01 -3.85933928e-02
-9.22031164e-01 -1.09395996e-01 -6.97640181e-01 -7.52602279e-01
1.41592249e-01 -6.37504518e-01 -2.14548215e-01 3.78361166e-01
-4.19740677e-01 1.16694123e-01 1.13202441e+00 -8.01266968e-01
1.29199111e+00 -2.20638943e+00 -1.88292116e-01 9.92010832e-02
2.29660630e-01 7.97341228e-01 9.72099826e-02 -8.78281891e-02
9.45914760e-02 1.29730448e-01 -1.21308744e-01 -5.59825957e-01
-1.60518736e-01 3.80510271e-01 -3.03820789e-01 3.30114812e-01
3.02703977e-01 9.15878177e-01 -3.90669674e-01 -4.18651462e-01
5.35224117e-02 5.88041663e-01 -6.56375051e-01 3.82225141e-02
-1.04738414e-01 -2.68519372e-01 -2.25136667e-01 7.33316183e-01
1.03063834e+00 -3.96438181e-01 2.70358711e-01 -5.18684566e-01
-4.34295870e-02 3.46011072e-01 -1.09569561e+00 1.51807690e+00
-3.62858504e-01 8.11400712e-01 1.55916270e-02 -1.24827445e+00
1.25911760e+00 9.64283347e-02 -2.43868411e-01 -4.33295429e-01
5.28499544e-01 5.10609806e-01 2.49082357e-01 -1.68867558e-01
5.32744944e-01 2.08251998e-01 4.33002502e-01 3.71499330e-01
3.62422198e-01 2.54577219e-01 -1.72200631e-02 -3.35785858e-02
8.27449560e-01 -2.22906724e-01 2.48006210e-02 -2.21320465e-01
4.70722616e-01 -1.39500052e-01 4.61615324e-01 5.33985436e-01
9.30704474e-02 3.23151052e-01 3.52700263e-01 -5.58482111e-01
-1.37309432e+00 -4.88316000e-01 -2.18411714e-01 7.59626269e-01
-1.19115725e-01 -2.08109245e-01 -8.92782927e-01 -3.88973296e-01
-3.21402878e-01 3.30905199e-01 -3.43595594e-01 -1.90572694e-01
-8.46684396e-01 -7.36896217e-01 1.10748136e+00 9.09136295e-01
1.53099132e+00 -7.31413722e-01 -8.10077429e-01 4.94465567e-02
3.96288693e-01 -1.18070054e+00 -1.46871954e-01 4.21844006e-01
-1.42209435e+00 -8.44826341e-01 -9.04208601e-01 -1.09009135e+00
9.04391110e-01 2.15218663e-01 5.92013597e-01 5.99070966e-01
-2.03834727e-01 -3.77385885e-01 -3.94436568e-01 -2.09484413e-01
4.26110812e-03 4.13507134e-01 -1.48679474e-02 -2.27307394e-01
1.62645876e-01 -6.79988086e-01 -7.87942648e-01 7.65439495e-02
-9.69207764e-01 4.80110317e-01 9.25452471e-01 1.00104225e+00
5.13361633e-01 4.73228276e-01 1.85689956e-01 -8.50545228e-01
3.10332209e-01 -8.90545547e-02 -7.32062280e-01 4.12066489e-01
-7.54381120e-01 9.73361656e-02 1.08412755e+00 -7.40010560e-01
-1.04685235e+00 3.65851521e-02 -6.70427782e-03 -5.94189346e-01
2.68382996e-01 5.82513034e-01 -1.00597583e-01 -3.44221205e-01
3.40373129e-01 4.70082134e-01 -5.20807058e-02 -7.91593790e-01
2.13953868e-01 7.91504502e-01 6.26782835e-01 -3.93200606e-01
5.42591691e-01 3.22689354e-01 2.82961018e-02 -6.89345002e-01
-3.79198849e-01 3.02494735e-01 -7.08926439e-01 2.92790979e-01
3.97958666e-01 -1.12172019e+00 -9.55243587e-01 8.66657972e-01
-1.28367794e+00 -4.84983146e-01 2.01988425e-02 5.80407500e-01
8.93855542e-02 4.08666641e-01 -1.02026653e+00 -5.43444872e-01
-8.15342188e-01 -1.11706579e+00 3.24651808e-01 4.31939840e-01
5.24029016e-01 -8.29121530e-01 -8.59433472e-01 1.13326870e-01
7.17276037e-01 -2.54341245e-01 9.45740104e-01 -6.70183182e-01
-8.33996475e-01 -2.72579551e-01 -6.38109028e-01 1.04681420e+00
-1.39477432e-01 8.04811120e-02 -1.04421210e+00 -4.66697544e-01
-4.01012897e-02 -3.74234796e-01 1.21902144e+00 1.70489877e-01
1.69492364e+00 -6.10863149e-01 -2.13598773e-01 1.10325420e+00
1.63869143e+00 2.89710850e-01 9.01813984e-01 3.70520860e-01
7.18638778e-01 1.56079799e-01 1.97363704e-01 5.21376967e-01
-3.10840961e-02 1.58364445e-01 5.22786915e-01 -9.80229303e-02
-9.35590044e-02 -1.72002226e-01 1.52382284e-01 1.34486938e+00
-3.65771413e-01 -6.29480481e-02 -7.59452343e-01 2.49418631e-01
-1.53650975e+00 -6.40868664e-01 2.05200911e-01 1.97011542e+00
1.04652393e+00 3.92470986e-01 -5.47756553e-01 2.99042970e-01
6.26395822e-01 -5.12817688e-02 -6.44714892e-01 -3.57574046e-01
-2.59710789e-01 5.85056782e-01 9.04385328e-01 2.61695713e-01
-8.23194206e-01 8.61701846e-01 5.69550943e+00 1.12904167e+00
-1.42996347e+00 1.38451517e-01 8.74300003e-01 -3.48693907e-01
1.00149408e-01 -3.91126163e-02 -1.10641003e+00 3.68154407e-01
9.68476892e-01 4.29449379e-02 5.29602826e-01 1.01792324e+00
-1.57735795e-01 1.06572807e-01 -1.09008574e+00 1.03531849e+00
-2.46396691e-01 -1.70775807e+00 1.32975072e-01 -1.56670019e-01
6.33599579e-01 4.46850583e-02 -6.96260761e-03 3.38665426e-01
-9.82959047e-02 -1.03157091e+00 7.53111660e-01 3.56769055e-01
1.10134375e+00 -9.53813255e-01 1.10448492e+00 4.80275273e-01
-1.14539135e+00 -2.49883786e-01 -9.95419204e-01 -3.05899292e-01
-2.09544197e-01 6.68157637e-01 -6.12151802e-01 4.50717658e-01
8.23245823e-01 6.04292691e-01 -6.04809523e-01 9.34846520e-01
-2.12483481e-02 6.61562383e-01 -3.79425079e-01 -2.01647818e-01
2.93034285e-01 5.63262664e-02 -3.76927555e-01 1.32656193e+00
4.25720751e-01 2.54703313e-01 -5.00143528e-01 8.20411265e-01
-5.33813417e-01 -1.49851531e-01 -1.56996027e-01 -1.45246446e-01
9.02290225e-01 1.11192083e+00 -6.75893962e-01 -6.20011330e-01
-3.93336087e-01 1.04945695e+00 5.10285378e-01 3.15669626e-01
-7.95797169e-01 -9.31757212e-01 3.08767498e-01 -1.98662385e-01
6.84877276e-01 -2.07093537e-01 -4.67180371e-01 -1.15986490e+00
2.59995759e-01 -7.40949690e-01 -6.54918700e-02 -5.78537822e-01
-7.19292402e-01 8.76473844e-01 -2.01702282e-01 -1.10569429e+00
2.94816941e-01 -1.09548092e+00 -5.72415352e-01 1.07786787e+00
-1.74289477e+00 -9.72964108e-01 -5.02782941e-01 6.55322552e-01
4.58367199e-01 -4.51206565e-01 7.19714284e-01 6.39618516e-01
-9.91820157e-01 1.13528025e+00 3.12693805e-01 5.32466948e-01
2.49026388e-01 -5.33622026e-01 3.84261191e-01 9.09432828e-01
-3.92792434e-01 8.89924228e-01 1.34567901e-01 -2.92239338e-01
-1.35519779e+00 -1.29074144e+00 6.86341763e-01 4.97585237e-01
4.83842373e-01 -4.53530163e-01 -1.00300801e+00 7.12587595e-01
5.49966358e-02 2.91998893e-01 3.12998712e-01 -3.37747753e-01
-4.00838703e-01 -4.55718517e-01 -1.15135598e+00 5.93251348e-01
8.75635862e-01 -3.15879524e-01 -2.04575524e-01 2.04986423e-01
1.02854228e+00 -4.77878392e-01 -7.76255190e-01 2.60429949e-01
5.82237363e-01 -7.73552060e-01 7.66092479e-01 -4.17941213e-01
7.15390086e-01 -1.00598738e-01 -2.54079938e-01 -6.91787004e-01
-1.81597486e-01 -4.57784027e-01 -4.24170822e-01 1.15069997e+00
4.37990487e-01 -8.42542231e-01 9.15666401e-01 6.62924707e-01
-3.49226505e-01 -1.11675930e+00 -9.23493207e-01 -7.88624465e-01
7.32470304e-02 -3.17913890e-01 7.29318678e-01 8.68316650e-01
-2.04556808e-01 -7.75103271e-02 -4.40052152e-01 8.75932500e-02
3.49746615e-01 -2.32934028e-01 4.37543958e-01 -9.94110763e-01
-6.67341530e-01 -5.24610817e-01 -4.15789425e-01 -1.61667383e+00
-1.56750947e-01 -7.22328305e-01 -3.12132090e-01 -1.06199539e+00
1.59794956e-01 -7.08223820e-01 -2.58150637e-01 8.69004250e-01
2.90474355e-01 1.30543262e-01 3.99980605e-01 2.92130232e-01
-1.40139982e-01 5.56631982e-01 1.26039898e+00 -2.71849334e-01
5.17018773e-02 -2.60574132e-01 -5.73624849e-01 6.31695271e-01
8.68502080e-01 -2.98364669e-01 -4.91853088e-01 -9.44724143e-01
1.35515660e-01 -3.47105980e-01 3.27536136e-01 -1.16325998e+00
6.03844941e-01 1.93210304e-01 4.12835181e-01 -5.61051071e-01
3.21334511e-01 -1.04137945e+00 3.54491435e-02 8.36830735e-01
-1.51571676e-01 6.78599104e-02 5.43415666e-01 1.51941463e-01
-2.71021485e-01 -7.43299961e-01 8.60927582e-01 -8.68479311e-02
-5.65315127e-01 2.61100054e-01 1.24066852e-01 -4.05586779e-01
8.80784690e-01 -4.07072484e-01 -7.35225081e-01 9.60384756e-02
-2.92821586e-01 -5.26465997e-02 2.17759963e-02 -8.86704624e-02
8.29102397e-01 -1.19267690e+00 -4.30974960e-01 3.28651667e-01
-5.86934745e-01 4.47803199e-01 4.08577502e-01 8.08241189e-01
-1.25624788e+00 7.73738801e-01 -3.27779174e-01 -1.33783743e-01
-1.16219747e+00 5.04536569e-01 4.76995975e-01 -2.12115854e-01
-7.65344560e-01 9.67891574e-01 -1.81966767e-01 -1.38564900e-01
6.56762779e-01 -8.08528125e-01 1.61659673e-01 -4.35629010e-01
6.69194102e-01 5.03703058e-01 1.80436075e-01 -3.68908718e-02
-1.35235623e-01 3.36750031e-01 -4.46554303e-01 4.33490783e-01
1.46724355e+00 1.97251663e-01 -2.43743941e-01 -2.82500796e-02
1.32991993e+00 -6.17960393e-01 -1.49776411e+00 -3.75115365e-01
-5.13132989e-01 -1.35872081e-01 3.09875309e-01 -5.99238753e-01
-1.67794776e+00 1.23036444e+00 5.02557397e-01 -2.09315419e-01
1.56583369e+00 -4.16365951e-01 9.03154492e-01 1.06654024e+00
3.74485373e-01 -1.08177650e+00 -5.35884425e-02 6.37665153e-01
5.55802763e-01 -8.83538008e-01 2.47449532e-01 -3.24703276e-01
-2.64752358e-01 1.54584014e+00 7.58749306e-01 -2.57303536e-01
8.51807594e-01 5.22281766e-01 -3.37400675e-01 9.67692956e-02
-8.34316254e-01 3.93492430e-01 -9.42360908e-02 1.99221283e-01
2.10995421e-01 -3.60966139e-02 -2.34193861e-01 9.45415020e-01
-2.87651539e-01 2.10191473e-01 5.54000199e-01 9.66547906e-01
-4.77649033e-01 -7.65809119e-01 8.11104476e-02 6.09010220e-01
-4.54075664e-01 -5.04624665e-01 1.98916569e-01 8.33786309e-01
3.76245558e-01 4.44127798e-01 2.54880041e-01 -7.06550837e-01
2.16849729e-01 -5.95025793e-02 3.48403871e-01 -2.05719337e-01
-6.44554734e-01 -1.85446799e-01 -1.94019347e-01 -2.23725379e-01
-2.17426762e-01 2.67738272e-02 -1.42193735e+00 -8.25106084e-01
-7.59141088e-01 -3.09697121e-01 7.43953645e-01 8.62250924e-01
3.80353212e-01 6.63917303e-01 2.59744227e-01 -7.73750842e-01
-8.81182671e-01 -1.13457239e+00 -6.41498327e-01 -1.16146110e-01
3.80442403e-02 -2.84936786e-01 -2.30747923e-01 2.29082525e-01]
|
[8.582704544067383, 2.977672815322876]
|
7ac4e8b8-6f3c-4f9a-9cca-ca55c1f009fa
|
multi-class-zero-shot-learning-for-artistic
|
2010.13850
| null |
https://arxiv.org/abs/2010.13850v1
|
https://arxiv.org/pdf/2010.13850v1.pdf
|
Multi-Class Zero-Shot Learning for Artistic Material Recognition
|
Zero-Shot Learning (ZSL) is an extreme form of transfer learning, where no labelled examples of the data to be classified are provided during the training stage. Instead, ZSL uses additional information learned about the domain, and relies upon transfer learning algorithms to infer knowledge about the missing instances. ZSL approaches are an attractive solution for sparse datasets. Here we outline a model to identify the materials with which a work of art was created, by learning the relationship between English descriptions of the subject of a piece and its composite materials. After experimenting with a range of hyper-parameters, we produce a model which is capable of correctly identifying the materials used on pieces from an entirely distinct museum dataset. This model returned a classification accuracy of 48.42% on 5,000 artworks taken from the Tate collection, which is distinct from the Rijksmuseum network used to create and train our model.
|
['Tom Bock', 'Andreea Cucu', 'Alexander W Olson']
|
2020-10-26
| null | null | null | null |
['material-recognition']
|
['computer-vision']
|
[ 3.58203322e-01 1.90133765e-01 -1.34119824e-01 -5.07256389e-01
-8.79852772e-01 -5.83138347e-01 8.65390539e-01 1.59148127e-02
-4.66444969e-01 7.80286610e-01 2.39367113e-01 1.18920490e-01
-3.45533907e-01 -1.05927885e+00 -9.20702755e-01 -3.82315308e-01
-2.52621695e-02 1.07986784e+00 2.57031530e-01 -1.31641343e-01
4.49723989e-01 3.01255435e-02 -1.71577382e+00 6.63129568e-01
3.63912165e-01 7.94118285e-01 4.06547576e-01 7.00563014e-01
-4.01300997e-01 7.50164032e-01 -6.76113367e-01 -5.90563774e-01
2.35066205e-01 -5.03491580e-01 -1.13752151e+00 1.29557863e-01
7.29586244e-01 -2.61457205e-01 -4.03007537e-01 6.05376184e-01
3.95778894e-01 1.85015202e-01 1.05290723e+00 -1.06837881e+00
-5.77202976e-01 7.94078171e-01 5.20840921e-02 1.78125098e-01
4.22748148e-01 1.97551146e-01 1.09333491e+00 -8.52791011e-01
1.17195439e+00 9.25249517e-01 1.06356311e+00 7.17198908e-01
-1.24300957e+00 -6.91812277e-01 -5.33569396e-01 4.91939098e-01
-1.16533434e+00 -5.21149099e-01 7.75357783e-01 -6.21704280e-01
6.85273111e-01 -7.92104602e-02 8.84513676e-01 1.20770013e+00
-3.65620494e-01 7.46789157e-01 1.03696942e+00 -6.65684521e-01
5.63182831e-01 3.80936950e-01 1.85589418e-02 6.83365524e-01
2.62174718e-02 -2.08931714e-01 -1.04271984e+00 -4.38381940e-01
8.51695538e-01 -2.05363289e-01 -2.53420863e-02 -3.54790926e-01
-1.04971004e+00 6.71406388e-01 4.36730266e-01 3.21186960e-01
-1.77770019e-01 5.81682399e-02 3.45356703e-01 3.11250031e-01
3.76162082e-01 6.46710992e-01 -4.86637533e-01 -2.78817266e-01
-1.03553617e+00 5.90220913e-02 1.14607334e+00 1.02125728e+00
1.08575511e+00 -1.61822319e-01 2.16095969e-01 1.13839900e+00
6.34090304e-02 1.80811971e-01 6.43549860e-01 -9.80240881e-01
3.61664593e-01 6.98026478e-01 -1.07625961e-01 -4.26770657e-01
-4.35385145e-02 -2.82545891e-02 -3.27139676e-01 1.08440883e-01
4.87155706e-01 -4.35711369e-02 -1.06311679e+00 1.43138754e+00
5.72943464e-02 3.21518570e-01 8.47226977e-02 6.10281169e-01
1.05294740e+00 5.58004379e-01 2.78743776e-03 3.21131855e-01
9.97531235e-01 -6.13193333e-01 -3.20044070e-01 -4.70532417e-01
4.82574165e-01 -3.77583891e-01 1.19590282e+00 2.62315124e-01
-9.82991993e-01 -4.47429568e-01 -1.04384720e+00 -1.09571107e-01
-5.93798280e-01 -1.27502054e-01 8.65070820e-01 4.11256969e-01
-9.95794296e-01 8.06784928e-01 -5.23518741e-01 -6.93521500e-01
7.06763208e-01 3.70276749e-01 -5.66715062e-01 -3.68491054e-01
-9.52655315e-01 8.76060009e-01 4.53733742e-01 -2.61227518e-01
-1.09761333e+00 -9.02740479e-01 -9.02924418e-01 2.91121062e-02
1.52047530e-01 -4.52189773e-01 1.14469492e+00 -8.84161234e-01
-1.03616297e+00 1.40397143e+00 2.00385466e-01 -3.40254337e-01
4.21134531e-01 3.75860184e-02 -3.40103447e-01 1.63842455e-01
1.90252885e-01 5.76315582e-01 7.58869588e-01 -1.41908443e+00
-4.82955754e-01 -2.98698276e-01 1.33669376e-03 -1.63186699e-01
-3.55888754e-01 -3.33664179e-01 -3.87658685e-01 -3.54437113e-01
-1.09567354e-02 -9.09491837e-01 2.24350601e-01 2.62666315e-01
3.59744616e-02 -2.01840863e-01 4.71681207e-01 -7.03880489e-01
6.73653543e-01 -2.22428107e+00 1.63878858e-01 2.28189752e-01
-2.53722370e-01 -3.08210701e-02 -1.92095444e-01 7.73741007e-01
2.20281124e-01 4.66502644e-02 -6.25808716e-01 -1.23037823e-01
-4.52094972e-02 4.17042911e-01 -2.18070567e-01 3.69281530e-01
6.88513368e-02 6.14272296e-01 -1.23231518e+00 -6.14678323e-01
1.58992022e-01 2.76850939e-01 -4.76649672e-01 2.44164586e-01
-3.08231324e-01 2.37768695e-01 -1.89367205e-01 6.01186812e-01
2.07881436e-01 -1.32385924e-01 1.46455139e-01 -6.12971326e-03
1.72010317e-01 1.56443924e-01 -9.51373816e-01 2.24449730e+00
-6.13655746e-01 8.32508266e-01 -3.02094575e-02 -5.99908888e-01
9.67697382e-01 2.91695744e-01 4.37346756e-01 -2.97204435e-01
9.09660980e-02 1.78329155e-01 -1.73550695e-01 -8.03085148e-01
2.60522932e-01 -5.64015329e-01 -1.67535990e-01 6.88068032e-01
7.81593978e-01 -4.63631809e-01 9.80949625e-02 2.54945725e-01
1.36298287e+00 2.93267190e-01 8.87784511e-02 -1.78598031e-01
1.31306767e-01 2.24133953e-01 4.28760439e-01 6.99539006e-01
-9.26951244e-02 4.92589206e-01 8.39219242e-03 -8.14001322e-01
-1.36342585e+00 -1.15044153e+00 -1.63691565e-01 1.15461087e+00
-1.13483980e-01 -4.65959966e-01 -5.02691329e-01 -4.53777015e-01
5.67186289e-02 8.69863272e-01 -9.52323735e-01 -3.09071720e-01
-1.96106657e-01 -2.78045148e-01 6.53024077e-01 3.21256101e-01
4.02943820e-01 -1.34258640e+00 -5.78398228e-01 -2.26377919e-02
-1.16348736e-01 -7.48489022e-01 1.10645987e-01 3.60008419e-01
-9.27520514e-01 -1.32030582e+00 -5.00078857e-01 -1.08261228e+00
5.98574281e-01 -1.84180483e-01 1.22014141e+00 8.90724733e-02
-6.60753667e-01 5.03085554e-01 -3.04022402e-01 -4.52247292e-01
-4.33396548e-01 1.75416678e-01 -3.07972897e-02 -1.18760087e-01
6.60361111e-01 -8.01221907e-01 -2.10123599e-01 1.29142448e-01
-9.05721724e-01 -1.48516774e-01 6.31577134e-01 9.25570428e-01
1.78401694e-01 -2.43822648e-03 5.53701580e-01 -1.17340541e+00
3.58367503e-01 -8.14923167e-01 4.08084430e-02 4.19634432e-01
-1.76372781e-01 1.65467381e-01 3.95626813e-01 -5.24918258e-01
-1.17390370e+00 3.84912401e-01 1.35033783e-02 -3.85984629e-01
-4.11417753e-01 1.23770319e-01 9.30313300e-03 -4.74803224e-02
8.63513768e-01 2.73652852e-01 -1.36625409e-01 -6.09768331e-01
2.70059139e-01 9.86471951e-01 7.39803493e-01 -8.39033365e-01
6.63119495e-01 5.01065850e-01 -2.93228060e-01 -1.02599132e+00
-9.58830774e-01 -5.12431324e-01 -1.09684277e+00 -4.08813149e-01
5.87168097e-01 -7.39734828e-01 -4.09643739e-01 2.94773817e-01
-7.22686589e-01 -6.37833416e-01 -5.85622668e-01 5.00951648e-01
-8.71679485e-01 -1.10586889e-01 -5.58885872e-01 -8.60707879e-01
-1.48531184e-01 -4.72546965e-01 9.31327164e-01 -7.86778927e-02
-5.17622709e-01 -8.81935775e-01 2.72693396e-01 4.28828448e-01
2.19794482e-01 1.75973147e-01 1.26110744e+00 -7.31139719e-01
-2.32600957e-01 -2.22117007e-01 2.09715888e-01 9.47171077e-02
2.30040640e-01 -1.53964713e-01 -1.20413744e+00 -2.17921093e-01
1.43425204e-02 -7.82476008e-01 7.00459301e-01 -8.92383754e-02
9.55093861e-01 -1.36502221e-01 -1.59015313e-01 4.96872693e-01
1.54102886e+00 7.76687786e-02 6.52009547e-01 6.20024979e-01
3.38402003e-01 8.70788574e-01 2.33553246e-01 5.15841663e-01
2.59036362e-01 1.60429627e-01 2.30479494e-01 3.96603405e-01
-2.68687665e-01 -6.05659068e-01 1.48397222e-01 6.77584827e-01
-1.73539713e-01 2.39984229e-01 -1.23281252e+00 9.36785638e-01
-1.53772473e+00 -1.00075710e+00 1.09536357e-01 2.21501136e+00
9.42428350e-01 1.53361246e-01 8.45632255e-02 7.52793476e-02
7.63508320e-01 -6.62528872e-02 -7.29079962e-01 -1.26600489e-01
1.74661100e-01 3.93317610e-01 4.21372384e-01 2.55637556e-01
-9.44664359e-01 1.10431504e+00 7.83728552e+00 6.53447509e-01
-3.90825152e-01 -2.28231940e-02 1.62347227e-01 -3.06752741e-01
-2.20626727e-01 -7.84340501e-02 -5.15394807e-01 5.09478271e-01
1.21058941e+00 -7.48862103e-02 6.45464122e-01 8.56301367e-01
-3.74923587e-01 -9.12247226e-02 -1.52565897e+00 1.02674401e+00
6.15693152e-01 -1.39290416e+00 7.26871341e-02 -2.39297729e-02
6.75733984e-01 4.01770025e-02 -4.07623827e-01 6.23894751e-01
5.31201959e-01 -8.36097360e-01 6.45961106e-01 6.88973904e-01
8.08651805e-01 -5.79242826e-01 5.56089580e-01 4.94248390e-01
-6.98179305e-01 -1.97671890e-01 -6.42637610e-01 -3.08304280e-01
-1.27014726e-01 2.08730809e-02 -1.08201885e+00 1.31064117e-01
8.60555053e-01 7.86323786e-01 -7.05782831e-01 1.16063678e+00
-3.60448420e-01 6.64342940e-01 -2.17760652e-01 -7.18352497e-02
1.05211422e-01 -1.63352874e-03 2.35571399e-01 1.06706131e+00
2.73855150e-01 1.24053419e-01 -4.54167910e-02 9.49229062e-01
-4.11319494e-01 -8.45074654e-02 -9.23054338e-01 -6.26278669e-02
5.11204004e-01 1.09360409e+00 -6.24577165e-01 -4.30113912e-01
-2.98714548e-01 9.61761653e-01 5.97551405e-01 -3.87659520e-02
-2.86387861e-01 -2.73419201e-01 2.27674469e-01 2.44889587e-01
3.09138179e-01 -1.36185125e-01 -1.90305516e-01 -9.88940954e-01
-1.35905072e-01 -4.73407447e-01 5.19010246e-01 -1.07894909e+00
-1.76384437e+00 4.51691002e-01 -7.40178153e-02 -1.30705166e+00
-1.39481217e-01 -5.49294889e-01 -5.92428565e-01 6.32412136e-01
-9.68835890e-01 -1.33680832e+00 -3.90748978e-01 5.22656739e-01
7.04077542e-01 -3.92511368e-01 1.14590633e+00 1.14397347e-01
1.38785476e-02 1.75571710e-01 2.19440550e-01 4.27630872e-01
9.37636495e-01 -1.21745920e+00 1.92992702e-01 1.46253869e-01
3.63740057e-01 4.74453300e-01 7.07995713e-01 -7.92406440e-01
-1.43342388e+00 -8.41639817e-01 6.41451955e-01 -7.53738582e-01
9.36960816e-01 -4.00865465e-01 -1.04120028e+00 8.94393265e-01
3.91701236e-02 -2.46701449e-01 1.29961669e+00 2.98830807e-01
-5.32303929e-01 1.56186149e-01 -1.36182535e+00 1.28292993e-01
1.10272408e+00 -8.10192585e-01 -1.32488418e+00 4.97180372e-01
1.92748562e-01 2.33535431e-02 -1.19436550e+00 -1.40606239e-01
6.56608105e-01 -5.92687130e-01 7.87404954e-01 -8.88742030e-01
8.60207260e-01 1.35813564e-01 -3.12208474e-01 -1.45004201e+00
-2.33146563e-01 -1.62234083e-01 -1.16509251e-01 1.30532336e+00
4.34235483e-01 3.80439498e-02 1.11709547e+00 1.03190565e+00
-8.36443231e-02 -2.59691715e-01 -8.63860607e-01 -9.02748585e-01
-2.44162083e-02 -3.58074814e-01 4.79632676e-01 1.24305022e+00
9.90549549e-02 5.80569804e-01 -4.00804967e-01 -4.28115189e-01
8.94613683e-01 6.91472962e-02 7.65523314e-01 -1.63860703e+00
-3.32333952e-01 -1.24168664e-01 -6.28170431e-01 -2.56889462e-01
4.04357672e-01 -1.20147598e+00 2.89286435e-01 -1.82280660e+00
3.79759133e-01 -5.35036206e-01 -1.06324174e-01 8.13116312e-01
2.71889478e-01 3.62774074e-01 2.86713298e-02 4.98168528e-01
-4.11421597e-01 4.84199196e-01 7.58601964e-01 -3.91974747e-01
-4.31290790e-02 -1.29874155e-01 -4.00673538e-01 8.85094762e-01
7.07646549e-01 -5.57998359e-01 -3.06318521e-01 -6.16771340e-01
1.63554668e-01 -1.41458705e-01 2.33921036e-01 -1.26043117e+00
4.92661297e-01 1.18193768e-01 5.97122192e-01 -4.06453758e-01
6.56530857e-01 -1.06625044e+00 2.40522489e-01 4.10824001e-01
-6.46574080e-01 -5.30134022e-01 1.29775535e-02 7.39507616e-01
-1.39879892e-02 -7.91076660e-01 6.60615146e-01 -6.18420482e-01
-1.23378789e+00 1.28604591e-01 -2.56865352e-01 6.07812963e-02
1.01602030e+00 -2.34722137e-01 -3.00543964e-01 -1.24299943e-01
-9.33069825e-01 2.82499194e-02 8.47034216e-01 3.23824108e-01
5.71063519e-01 -1.30495632e+00 -6.80730641e-01 3.40543300e-01
5.10102689e-01 -5.82310073e-02 1.98912784e-01 8.36276263e-03
-5.87127626e-01 -6.36020303e-02 -7.00048685e-01 -2.92037100e-01
-9.13432240e-01 4.90169406e-01 -3.38376537e-02 2.97317505e-01
-9.09656346e-01 8.61373484e-01 -4.84447896e-01 -6.85802519e-01
1.48690075e-01 3.15399677e-01 -2.34657396e-02 1.49909407e-01
5.88631988e-01 4.22564209e-01 1.11086018e-01 -5.61849236e-01
-2.90398031e-01 4.14733231e-01 6.15564957e-02 -3.16477388e-01
2.01880312e+00 4.11452621e-01 -4.78243195e-02 1.20839787e+00
1.18725288e+00 -3.94065350e-01 -1.24345255e+00 -5.41694283e-01
2.84854621e-01 -6.10789239e-01 -5.20997606e-02 -9.88753200e-01
-6.89953327e-01 8.94084573e-01 3.09052676e-01 -9.53358337e-02
6.05464339e-01 3.75500560e-01 6.25264347e-01 7.44007945e-01
7.59086967e-01 -1.30191302e+00 9.12203565e-02 4.46294874e-01
7.54353464e-01 -1.42070460e+00 4.93578538e-02 -6.84482455e-02
-5.09839475e-01 1.20061219e+00 6.45367980e-01 -4.33862507e-01
5.65377533e-01 2.06557170e-01 -3.28988820e-01 -4.80876833e-01
-7.70125568e-01 -1.07320122e-01 6.13316149e-02 7.65696645e-01
2.55636185e-01 -1.87772542e-01 9.17713046e-02 5.55637538e-01
-4.50737208e-01 1.52637318e-01 5.43615103e-01 1.34531188e+00
-6.86869919e-01 -8.84811878e-01 -2.79147148e-01 7.43865967e-01
-5.75619563e-02 -1.43694514e-02 -1.00565350e+00 7.26340652e-01
2.22133636e-01 6.25602782e-01 2.58645177e-01 -4.96463597e-01
2.08641008e-01 5.27527750e-01 6.56723261e-01 -9.80226457e-01
-6.41431868e-01 -4.40174311e-01 1.21862628e-01 -1.23740882e-01
-6.26496017e-01 -5.12268007e-01 -1.07928777e+00 -2.63792485e-01
7.03917909e-03 4.00192678e-01 8.19337606e-01 8.55855048e-01
-4.11485806e-02 3.30359370e-01 3.55533183e-01 -1.09493220e+00
-1.26576394e-01 -1.20182633e+00 -9.41279233e-01 7.23858535e-01
-1.06629923e-01 -6.79992974e-01 -5.25189459e-01 4.33454931e-01]
|
[9.98342227935791, 2.990692138671875]
|
c4087a30-683f-4621-bc64-f0441bc520a7
|
unsupervised-learning-for-intrinsic-image
|
1911.09930
| null |
https://arxiv.org/abs/1911.09930v2
|
https://arxiv.org/pdf/1911.09930v2.pdf
|
Unsupervised Learning for Intrinsic Image Decomposition from a Single Image
|
Intrinsic image decomposition, which is an essential task in computer vision, aims to infer the reflectance and shading of the scene. It is challenging since it needs to separate one image into two components. To tackle this, conventional methods introduce various priors to constrain the solution, yet with limited performance. Meanwhile, the problem is typically solved by supervised learning methods, which is actually not an ideal solution since obtaining ground truth reflectance and shading for massive general natural scenes is challenging and even impossible. In this paper, we propose a novel unsupervised intrinsic image decomposition framework, which relies on neither labeled training data nor hand-crafted priors. Instead, it directly learns the latent feature of reflectance and shading from unsupervised and uncorrelated data. To enable this, we explore the independence between reflectance and shading, the domain invariant content constraint and the physical constraint. Extensive experiments on both synthetic and real image datasets demonstrate consistently superior performance of the proposed method.
|
['ShaoDi You', 'Yunfei Liu', 'Feng Lu', 'Yu Li']
|
2019-11-22
|
unsupervised-learning-for-intrinsic-image-1
|
http://openaccess.thecvf.com/content_CVPR_2020/html/Liu_Unsupervised_Learning_for_Intrinsic_Image_Decomposition_From_a_Single_Image_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Liu_Unsupervised_Learning_for_Intrinsic_Image_Decomposition_From_a_Single_Image_CVPR_2020_paper.pdf
|
cvpr-2020-6
|
['intrinsic-image-decomposition']
|
['computer-vision']
|
[ 8.69005620e-01 -1.09188393e-01 1.76651940e-01 -2.86676794e-01
-5.11862457e-01 -3.17513704e-01 5.26927590e-01 -2.23559067e-01
-2.03193471e-01 6.70931995e-01 -2.28889603e-02 -1.15904910e-02
-5.94912618e-02 -6.62533700e-01 -3.91696751e-01 -1.15833557e+00
6.46440089e-01 1.51619911e-01 3.16604264e-02 5.51349595e-02
1.28344581e-01 1.71401247e-01 -1.53737509e+00 -9.95773003e-02
1.11298633e+00 9.47273612e-01 5.97592831e-01 2.69343585e-01
-1.16003223e-01 5.36687493e-01 -1.62618041e-01 -5.78861870e-02
4.81435150e-01 -4.33977455e-01 -3.56653333e-01 7.03652442e-01
1.79386720e-01 -3.78394514e-01 -4.57179546e-02 1.29349482e+00
1.23699315e-01 4.76968922e-02 6.90753281e-01 -8.12453806e-01
-2.44284049e-01 -4.74306904e-02 -9.28676248e-01 -2.45677859e-01
9.29828808e-02 1.25645638e-01 1.06167340e+00 -7.97413349e-01
1.42557904e-01 8.55094016e-01 3.51610482e-01 3.14603984e-01
-1.52612531e+00 -3.11411798e-01 1.31089866e-01 -2.02538237e-01
-1.37068319e+00 -5.43317318e-01 1.27884114e+00 -5.56534648e-01
1.43082276e-01 1.19155273e-01 3.05775464e-01 8.77078235e-01
-2.83605874e-01 5.80031455e-01 1.50986636e+00 -5.79318166e-01
3.13537031e-01 2.59069890e-01 1.53765446e-02 4.19916987e-01
3.87060493e-01 1.03039192e-02 -4.38201785e-01 -9.52444598e-02
8.59038293e-01 9.35516357e-02 -7.53462851e-01 -5.63221157e-01
-1.06638420e+00 4.74314809e-01 1.43175542e-01 -1.18878886e-01
-4.59870785e-01 -2.40558192e-01 -7.87110627e-02 -1.23381138e-01
3.61530751e-01 1.01576030e-01 -3.10766429e-01 2.77953327e-01
-8.95420551e-01 -1.72418401e-01 6.08476520e-01 6.89304590e-01
1.16141307e+00 7.43657574e-02 3.21469009e-01 9.82385993e-01
5.40034711e-01 6.92980111e-01 1.77592024e-01 -9.18742299e-01
2.95125514e-01 5.44513404e-01 5.43173254e-01 -1.30231285e+00
-7.94744417e-02 -5.34626007e-01 -9.05049682e-01 3.08940709e-01
5.82719147e-01 -5.51918782e-02 -9.39496815e-01 1.72467637e+00
5.29839337e-01 2.08875939e-01 1.75809592e-01 1.17933547e+00
5.97812176e-01 4.83043909e-01 -2.71411330e-01 -5.13053715e-01
1.11179042e+00 -7.52154291e-01 -5.85251451e-01 -4.60437447e-01
-5.78730889e-02 -9.70192313e-01 1.20208108e+00 5.74739456e-01
-9.57241237e-01 -5.29109657e-01 -1.12259126e+00 -5.97341433e-02
2.27241188e-01 5.34698427e-01 6.75215006e-01 6.65470898e-01
-6.47200108e-01 1.74235404e-01 -8.35892379e-01 -2.40030870e-01
-4.23554406e-02 2.89354712e-01 -1.50218576e-01 -3.29039752e-01
-6.01759851e-01 3.83442581e-01 2.64723539e-01 4.55249876e-01
-8.05366099e-01 -2.29487538e-01 -7.15439081e-01 -1.91228371e-02
6.07283533e-01 -5.65612137e-01 7.40821481e-01 -1.11300278e+00
-1.77665174e+00 7.45902598e-01 -3.27644736e-01 -2.17749029e-02
3.61861259e-01 -1.03687719e-01 9.16972104e-03 2.09248006e-01
-1.63950726e-01 9.81483981e-02 1.39512503e+00 -1.90215230e+00
-2.55811721e-01 -3.93048257e-01 2.21052527e-01 2.79246300e-01
-3.94918442e-01 -3.94247383e-01 -5.57840586e-01 -4.15245354e-01
6.40948474e-01 -5.92721403e-01 -2.07209960e-01 -1.25898853e-01
-3.76804858e-01 2.98905551e-01 5.49809635e-01 -5.88351429e-01
6.30002797e-01 -2.28131151e+00 6.39457479e-02 1.36101022e-01
1.99173033e-01 4.23237309e-02 -1.00824006e-01 2.12113634e-01
3.81887443e-02 -3.63343954e-01 -5.39884388e-01 -3.45410317e-01
-2.09725380e-01 3.84914488e-01 -4.63525206e-01 6.68896735e-01
9.98608544e-02 2.67504096e-01 -8.35214376e-01 -4.16736394e-01
3.60184997e-01 7.09562063e-01 -4.85394359e-01 4.29780632e-01
-2.89945483e-01 1.01363945e+00 -6.18616462e-01 4.51596260e-01
9.55409408e-01 -3.40455085e-01 4.71034914e-01 -3.88535291e-01
-1.50183052e-01 1.52385145e-01 -1.27014911e+00 1.77394032e+00
-5.25755405e-01 3.17465276e-01 4.24928576e-01 -1.28217661e+00
1.15184021e+00 2.74808317e-01 6.44418240e-01 -4.80158359e-01
2.37983063e-01 1.31436184e-01 -2.80347258e-01 -5.53606451e-01
1.60895154e-01 -4.84977186e-01 4.75178272e-01 3.80950689e-01
-3.63808542e-01 -1.08248889e-01 -3.96127589e-02 -1.50445327e-01
8.82618606e-01 4.71084833e-01 2.51232922e-01 -2.82800823e-01
9.72851276e-01 -1.62311316e-01 1.03586745e+00 4.67601120e-01
2.33611226e-01 8.54839921e-01 3.66964757e-01 -2.18900934e-01
-9.16620195e-01 -1.18875575e+00 -1.09397553e-01 5.04472792e-01
4.53406125e-01 -1.25215918e-01 -7.16471195e-01 -2.23915592e-01
-3.01665962e-01 3.59858304e-01 -3.59375924e-01 1.79877713e-01
-3.21320027e-01 -9.85221505e-01 -1.85746893e-01 8.91223550e-02
7.29656875e-01 -4.72174555e-01 -7.74132192e-01 2.44851988e-02
-4.23674285e-01 -1.45524681e+00 -1.44879147e-01 -1.07237071e-01
-8.67049336e-01 -1.09432554e+00 -5.25445759e-01 -5.55501580e-01
1.21337521e+00 7.96097219e-01 9.36048269e-01 3.06389809e-01
-3.94724607e-01 4.21529770e-01 -3.51755023e-01 -9.89699289e-02
6.55772462e-02 -1.84707150e-01 -1.03893012e-01 6.98027909e-01
7.94838816e-02 -9.14724827e-01 -7.16201782e-01 4.62905228e-01
-1.02574277e+00 4.14312899e-01 7.08646715e-01 9.98762906e-01
6.69201255e-01 4.56287175e-01 5.26816957e-02 -9.27689195e-01
1.50502935e-01 -1.08171932e-01 -9.51035380e-01 2.55340338e-01
-5.23568928e-01 -8.01613107e-02 7.70037115e-01 -8.35741684e-02
-1.58550632e+00 5.13439953e-01 1.86692968e-01 -2.65743017e-01
-4.10194397e-01 5.61682701e-01 -6.28102779e-01 -8.73703212e-02
3.61905068e-01 5.63323617e-01 -3.60368676e-02 -7.52997518e-01
1.84891239e-01 5.51072299e-01 6.78017914e-01 -8.11460435e-01
1.11301005e+00 1.02176654e+00 7.88661465e-02 -1.27714360e+00
-9.92508769e-01 -6.03187919e-01 -7.90717840e-01 -1.37760907e-01
8.50048006e-01 -9.18684423e-01 -6.19928122e-01 2.97520131e-01
-1.06644475e+00 -3.56846094e-01 -1.55216614e-02 5.22964120e-01
-4.88583863e-01 9.27414715e-01 -1.31512061e-01 -1.04899085e+00
5.94112650e-02 -1.08920574e+00 1.01213372e+00 1.94675326e-01
3.94220531e-01 -9.58198965e-01 2.54123621e-02 6.92168832e-01
2.12964922e-01 4.04179364e-01 8.27992380e-01 2.71903068e-01
-8.70878279e-01 1.13518015e-02 -6.53409302e-01 4.96864080e-01
5.50311983e-01 -5.37680238e-02 -1.10254288e+00 -3.01984072e-01
5.63347876e-01 -2.78261572e-01 8.20657372e-01 2.64565915e-01
1.03523123e+00 -7.68376365e-02 1.27812754e-02 8.00676405e-01
1.86309612e+00 -9.18235332e-02 5.83811402e-01 9.58584920e-02
6.77024662e-01 9.78013933e-01 5.80188155e-01 6.52211487e-01
3.52745056e-01 5.34361303e-01 5.55615664e-01 -2.96598583e-01
-4.44233716e-02 -1.04556352e-01 1.40421331e-01 9.52067256e-01
-2.51607716e-01 -2.07184300e-01 -7.89468706e-01 3.68672758e-01
-1.78892994e+00 -6.17888510e-01 -3.27755183e-01 2.36782217e+00
7.05620825e-01 -5.68029992e-02 -3.09730530e-01 3.49795312e-01
5.05755067e-01 1.87333629e-01 -5.70007622e-01 3.71638864e-01
-1.21817641e-01 2.87943352e-02 3.06464285e-01 6.39413476e-01
-8.24030161e-01 7.15686798e-01 5.79364920e+00 5.22555709e-01
-1.20303464e+00 3.41188349e-02 4.17377919e-01 2.53935307e-01
-4.68170941e-01 3.88493150e-01 -4.53575701e-01 2.69018561e-01
2.24510953e-01 3.01405728e-01 7.00708210e-01 4.33497846e-01
5.05341411e-01 -4.69153851e-01 -9.29307163e-01 1.22255158e+00
6.82764500e-02 -5.51022232e-01 -1.70476183e-01 7.70946965e-02
6.72856987e-01 -3.51889402e-01 3.93535122e-02 -2.57973105e-01
9.33147222e-02 -7.84220219e-01 3.33909005e-01 5.86478233e-01
5.38784862e-01 -2.81253487e-01 4.19485927e-01 5.85463107e-01
-1.02620494e+00 1.66943908e-01 -6.33746088e-01 -3.40208232e-01
2.59745121e-02 1.05099773e+00 -3.34794700e-01 6.17713153e-01
4.89818752e-01 8.06455433e-01 -2.32480586e-01 7.89592385e-01
-6.04737997e-01 6.07501745e-01 -3.69256347e-01 6.29342616e-01
-9.11797211e-02 -8.56458187e-01 4.08217639e-01 6.99073613e-01
3.80780175e-02 3.50499064e-01 4.39840347e-01 1.03549290e+00
2.27330998e-01 1.49857521e-01 -4.28415507e-01 1.05648100e-01
-1.16610050e-01 1.52202177e+00 -8.83717835e-01 8.35800394e-02
-6.67858362e-01 9.84679997e-01 1.71180159e-01 8.37416291e-01
-6.79606676e-01 2.35182978e-02 4.98419970e-01 9.39738229e-02
2.33942240e-01 -6.46833599e-01 -5.26032507e-01 -1.53244281e+00
2.65444845e-01 -7.42920280e-01 -2.26350781e-02 -6.88699722e-01
-1.20463192e+00 3.33637029e-01 -2.37491697e-01 -1.37830472e+00
1.93020865e-01 -6.98482871e-01 -5.64745724e-01 9.14909780e-01
-2.06519294e+00 -1.13403904e+00 -7.73434937e-01 7.59417653e-01
4.27806646e-01 2.22768828e-01 5.51339328e-01 2.26598904e-01
-7.78122723e-01 8.16122219e-02 4.27518457e-01 -7.34363571e-02
6.64701462e-01 -1.12794089e+00 -3.82104695e-01 9.86227274e-01
-3.53527963e-02 5.86644232e-01 9.19994533e-01 -4.03005779e-01
-1.58690095e+00 -7.90990233e-01 2.39353597e-01 3.38654853e-02
4.69737887e-01 -2.59286702e-01 -8.53358567e-01 3.15341890e-01
1.55707315e-01 1.11733124e-01 7.47255325e-01 -6.29031733e-02
-2.17164084e-01 -3.89102280e-01 -8.25852752e-01 4.33305651e-01
7.62427807e-01 -5.18959701e-01 -2.44491622e-01 3.45890969e-01
3.42188418e-01 -1.29345298e-01 -6.62900627e-01 4.61966902e-01
4.58134532e-01 -1.32493794e+00 1.03371847e+00 -8.82803276e-02
4.24077451e-01 -6.15554750e-01 -4.76829708e-01 -9.31423664e-01
-9.67308506e-02 -4.31201220e-01 2.40984634e-01 1.43050289e+00
1.54969707e-01 -7.31272817e-01 9.41972792e-01 6.63283765e-01
6.63050562e-02 -4.85340446e-01 -4.52686548e-01 -7.24861979e-01
-3.84546697e-01 -2.74302095e-01 4.11188781e-01 9.72756743e-01
-4.78147805e-01 3.42179805e-01 -6.11306369e-01 6.43253088e-01
1.21650445e+00 5.55822849e-01 1.00375545e+00 -1.31247413e+00
-5.91626883e-01 -1.41140103e-01 -1.13527037e-01 -1.30279076e+00
1.30108744e-01 -5.15936375e-01 4.94653791e-01 -1.48984694e+00
3.22678655e-01 -5.82106888e-01 -2.00624183e-01 1.81520000e-01
-2.50795662e-01 2.33404681e-01 -2.02725247e-01 5.48605561e-01
-2.50101805e-01 8.87244344e-01 1.12019014e+00 -1.49586663e-01
-1.65393844e-01 -5.69545738e-02 -5.90497375e-01 9.94434297e-01
8.23675931e-01 -2.33526111e-01 -8.72210622e-01 -8.31647635e-01
2.05496699e-01 8.17993656e-02 4.52320963e-01 -7.84386039e-01
1.04391158e-01 -5.57832956e-01 1.37093425e-01 -5.23795128e-01
4.99125451e-01 -1.09116232e+00 1.30031228e-01 3.74868768e-03
1.17561050e-01 -5.56998610e-01 -2.59764642e-01 8.00992370e-01
-2.98445404e-01 -3.35555375e-01 8.50016057e-01 -2.85763621e-01
-4.29116964e-01 1.95917487e-01 8.27253424e-03 -1.32155776e-01
6.60423040e-01 -2.38691717e-01 3.07168942e-02 -2.42340475e-01
-4.19824094e-01 4.07619998e-02 8.02138865e-01 -1.28184229e-01
6.60377502e-01 -9.73094165e-01 -5.30093014e-01 2.05689639e-01
2.20506210e-02 3.02872539e-01 1.66265547e-01 9.73342240e-01
-4.78648186e-01 2.00212106e-01 2.14606449e-02 -8.27538669e-01
-1.06308186e+00 5.89740694e-01 1.02128364e-01 -9.44730267e-02
-6.47957861e-01 3.94346952e-01 7.70407736e-01 -3.67271513e-01
6.05734810e-02 -1.15157709e-01 -4.72009070e-02 -3.04663301e-01
3.02670896e-01 6.58359975e-02 -1.57234073e-01 -6.89580441e-01
-6.64680004e-02 8.60158026e-01 1.81630552e-01 4.29648310e-02
1.30247867e+00 -6.16949320e-01 -3.44932586e-01 3.38816375e-01
9.26524639e-01 1.05089933e-01 -1.50413740e+00 -4.80351478e-01
-7.01661631e-02 -9.05440807e-01 3.48919243e-01 -2.28785262e-01
-1.20356524e+00 1.07131803e+00 1.86966434e-01 1.76579937e-01
1.56664670e+00 -5.11968374e-01 7.43216693e-01 2.90245503e-01
5.00582218e-01 -9.35954690e-01 1.83495432e-01 1.12110510e-01
5.62977910e-01 -1.49023390e+00 3.45113546e-01 -8.96566987e-01
-2.07689345e-01 1.10192335e+00 5.40522814e-01 1.72762126e-02
6.28980219e-01 -4.35574763e-02 2.02082276e-01 -8.24491680e-02
-4.03203964e-01 -4.02371794e-01 2.20509544e-01 4.58127677e-01
4.27873701e-01 -1.72675088e-01 -3.20036709e-01 3.64194572e-01
2.29411215e-01 -1.88487694e-01 4.32592601e-01 7.02217996e-01
-3.85879427e-01 -1.21222293e+00 -5.53764999e-01 1.30603224e-01
-3.30177963e-01 5.41935526e-02 -1.84969947e-01 3.34436417e-01
8.38717893e-02 1.10150588e+00 -4.67650652e-01 -7.67943338e-02
1.63780469e-02 -2.68233806e-01 4.66432750e-01 -6.62996531e-01
2.10636884e-01 6.07115924e-01 -2.72527635e-01 -2.47581273e-01
-8.94391418e-01 -7.32106686e-01 -9.86871958e-01 1.60108373e-01
-4.98390526e-01 -2.89137252e-02 6.70609534e-01 9.86034214e-01
-1.28852144e-01 2.69961894e-01 9.10223246e-01 -7.76829422e-01
-3.66985977e-01 -5.65088570e-01 -8.41185868e-01 4.06085193e-01
4.73628372e-01 -8.74219954e-01 -5.24650633e-01 3.53847951e-01]
|
[10.00757122039795, -2.7585465908050537]
|
c4b1aa51-05bb-4e57-af94-127be4992462
|
enabling-factorized-piano-music-modeling-and
|
1810.12247
| null |
http://arxiv.org/abs/1810.12247v5
|
http://arxiv.org/pdf/1810.12247v5.pdf
|
Enabling Factorized Piano Music Modeling and Generation with the MAESTRO Dataset
|
Generating musical audio directly with neural networks is notoriously
difficult because it requires coherently modeling structure at many different
timescales. Fortunately, most music is also highly structured and can be
represented as discrete note events played on musical instruments. Herein, we
show that by using notes as an intermediate representation, we can train a
suite of models capable of transcribing, composing, and synthesizing audio
waveforms with coherent musical structure on timescales spanning six orders of
magnitude (~0.1 ms to ~100 s), a process we call Wave2Midi2Wave. This large
advance in the state of the art is enabled by our release of the new MAESTRO
(MIDI and Audio Edited for Synchronous TRacks and Organization) dataset,
composed of over 172 hours of virtuosic piano performances captured with fine
alignment (~3 ms) between note labels and audio waveforms. The networks and the
dataset together present a promising approach toward creating new expressive
and interpretable neural models of music.
|
['Cheng-Zhi Anna Huang', 'Adam Roberts', 'Sander Dieleman', 'Jesse Engel', 'Douglas Eck', 'Ian Simon', 'Erich Elsen', 'Curtis Hawthorne', 'Andriy Stasyuk']
|
2018-10-29
|
enabling-factorized-piano-music-modeling-and-1
|
https://openreview.net/forum?id=r1lYRjC9F7
|
https://openreview.net/pdf?id=r1lYRjC9F7
|
iclr-2019-5
|
['music-modeling', 'piano-music-modeling']
|
['music', 'music']
|
[ 4.47343171e-01 7.12233782e-03 1.47376493e-01 -8.47291499e-02
-9.68671560e-01 -1.31162202e+00 3.94721210e-01 -1.48726150e-01
1.38426825e-01 4.79140431e-01 5.72203398e-01 1.85956419e-01
-4.39201206e-01 -5.05263448e-01 -7.06983805e-01 -3.13357592e-01
-4.46727097e-01 5.47045469e-01 -2.72571683e-01 -3.26703578e-01
1.91964760e-01 3.61751944e-01 -1.72801375e+00 7.69341588e-01
1.56610355e-01 9.39096391e-01 -9.81244072e-02 1.23726201e+00
-6.29915744e-02 9.86204147e-01 -1.11480725e+00 -1.78267017e-01
2.72745699e-01 -6.89907789e-01 -6.21563017e-01 -3.72512937e-01
8.80689263e-01 -1.37167528e-01 -2.15953603e-01 6.15693271e-01
7.26872087e-01 4.35679942e-01 4.61542994e-01 -1.05672336e+00
-3.85297626e-01 1.51992726e+00 1.34170223e-02 -2.87158079e-02
3.11895400e-01 -4.51278798e-02 1.46182740e+00 -5.26711524e-01
6.96585059e-01 8.45689714e-01 1.15464497e+00 4.10979629e-01
-1.33249331e+00 -9.50768411e-01 -3.57421637e-01 4.70430702e-02
-1.27946997e+00 -8.21395874e-01 9.67864215e-01 -5.54203331e-01
8.40021968e-01 6.00327373e-01 1.00137770e+00 1.30316484e+00
-3.17339033e-01 7.87683010e-01 5.64983726e-01 -3.54134977e-01
3.92360799e-02 -6.55176401e-01 -2.20913127e-01 2.65007406e-01
-5.98170519e-01 3.58172119e-01 -1.21138549e+00 -1.56307995e-01
1.02207708e+00 -4.65597153e-01 -4.45828424e-04 4.21470284e-01
-1.54619825e+00 3.06339920e-01 7.70568177e-02 3.40492219e-01
-2.79799849e-01 7.70206809e-01 7.26527274e-01 5.03998816e-01
2.31517795e-02 1.17788148e+00 -3.60455751e-01 -9.66374636e-01
-1.42197216e+00 6.65240347e-01 8.02055359e-01 7.82283545e-01
6.62983656e-02 8.93050134e-01 -2.62869000e-01 8.17106605e-01
-3.52886975e-01 3.47819567e-01 5.02706826e-01 -1.62668455e+00
2.63648033e-01 -1.88357279e-01 1.19377352e-01 -8.90759289e-01
-3.65803361e-01 -6.98479176e-01 -8.76198113e-01 3.21544558e-01
5.92961252e-01 -2.81881571e-01 -5.86169541e-01 2.00205922e+00
-3.60315323e-01 6.89502418e-01 -2.84471869e-01 5.76217413e-01
8.81668925e-01 8.15597534e-01 -3.09458584e-01 -1.65833198e-02
1.17445827e+00 -8.05283070e-01 -7.89182484e-01 3.34280550e-01
-9.83133540e-03 -1.00271368e+00 1.32696855e+00 1.01274955e+00
-1.63013422e+00 -1.03545356e+00 -1.18198514e+00 -1.58852756e-01
2.25030243e-01 2.44723871e-01 7.46887684e-01 1.48711950e-01
-9.48576212e-01 1.20220220e+00 -6.03172541e-01 2.17797145e-01
1.66266143e-01 2.93594390e-01 -9.24939737e-02 9.17722642e-01
-1.09772992e+00 1.91081703e-01 4.38961387e-01 2.41950974e-02
-1.17580104e+00 -1.27694464e+00 -3.70294482e-01 2.57523268e-01
4.26752083e-02 -5.18794835e-01 1.80415988e+00 -9.33790147e-01
-1.92233562e+00 6.76369727e-01 2.03929752e-01 -6.79054439e-01
2.08609909e-01 -4.40330505e-01 -5.03773391e-01 5.19872233e-02
-2.87763476e-01 7.14571059e-01 8.50090027e-01 -8.21363986e-01
-6.36574805e-01 1.99307665e-01 6.89352825e-02 4.87649962e-02
-2.83165663e-01 2.14834362e-01 -1.23718299e-01 -1.29645526e+00
-1.68013617e-01 -1.13898003e+00 2.72032917e-01 -2.86946625e-01
-6.98413670e-01 1.46864715e-03 2.51235545e-01 -6.63690388e-01
1.46822500e+00 -2.31495762e+00 5.28564036e-01 3.90706286e-02
1.56540915e-01 -1.37155041e-01 -4.25485194e-01 4.93471950e-01
-1.86738893e-01 1.75363440e-02 -1.00315168e-01 -4.56927419e-01
3.66353273e-01 -1.12405919e-01 -1.07795429e+00 -1.04021817e-01
-1.54933751e-01 9.07631993e-01 -7.94118583e-01 -3.70483659e-02
-2.62871474e-01 4.56185848e-01 -5.30566394e-01 2.02049300e-01
-5.19525170e-01 5.94437420e-01 1.63195476e-01 6.04672551e-01
-1.60031855e-01 4.24282849e-02 4.87032607e-02 -3.26419979e-01
-3.76014024e-01 7.64643908e-01 -1.23257375e+00 2.35181570e+00
-3.81794214e-01 1.16098034e+00 -1.41131416e-01 -3.47566307e-01
1.03801501e+00 8.28781664e-01 6.99091315e-01 -1.89780012e-01
6.28965423e-02 1.23441868e-01 8.95559788e-02 -8.08040425e-02
7.66480029e-01 -4.28580850e-01 -4.15379256e-01 8.88335586e-01
2.66657650e-01 -5.46062410e-01 4.02131736e-01 -2.59235948e-01
1.16126597e+00 3.85204077e-01 -4.87481169e-02 1.61739156e-01
-2.22233772e-01 -1.53126970e-01 4.82132792e-01 8.17109168e-01
2.11802423e-01 9.91312861e-01 3.86248022e-01 -6.15194023e-01
-1.19427073e+00 -1.39093733e+00 1.95204318e-01 1.49100208e+00
-6.63919747e-01 -8.43021691e-01 -6.55882120e-01 2.39642680e-01
-1.88680917e-01 5.71418285e-01 -5.92593431e-01 -4.71558534e-02
-9.45460498e-01 -2.20694765e-01 1.57101941e+00 7.75345325e-01
-8.09910521e-02 -1.75603962e+00 -4.43693697e-01 6.44204080e-01
-4.08066034e-01 -6.60606027e-01 -6.91159844e-01 4.04279679e-01
-9.07828093e-01 -6.39501512e-01 -4.96378630e-01 -6.67982638e-01
-2.51279235e-01 -2.71301121e-01 1.74083257e+00 -1.91949651e-01
-3.66384804e-01 -1.40278816e-01 -1.98383942e-01 -8.32612157e-01
-5.16600728e-01 3.96813899e-01 3.72558981e-01 -2.03542858e-01
-2.31810987e-01 -1.40836275e+00 -2.53983706e-01 -1.86285481e-01
-8.66468668e-01 2.77217060e-01 3.46565962e-01 5.13863504e-01
7.33495355e-01 1.11279227e-02 7.12444246e-01 -6.75381482e-01
8.55752587e-01 -1.87189296e-01 -3.49960029e-01 -1.78665385e-01
3.74812353e-03 -7.17241764e-02 8.32410872e-01 -8.77410352e-01
-7.43797839e-01 -1.16518876e-02 -1.75492272e-01 -5.38590074e-01
-2.56526709e-01 5.24565279e-01 2.84454584e-01 3.98866564e-01
1.00835443e+00 1.08981486e-02 -3.59982640e-01 -7.18213022e-01
6.64602280e-01 3.41228575e-01 1.45806241e+00 -9.82449353e-01
9.49792325e-01 3.56042325e-01 -1.93331502e-02 -5.34773767e-01
-1.08880591e+00 2.95180790e-02 -6.91088080e-01 -3.84826750e-01
4.96718824e-01 -8.12300503e-01 -1.01819098e+00 3.92994225e-01
-1.07486415e+00 -5.86866140e-01 -8.17531228e-01 3.47955793e-01
-1.00980997e+00 -2.67003626e-01 -1.11842871e+00 -6.16089702e-01
-6.20261669e-01 -5.11780500e-01 8.81596744e-01 6.29619434e-02
-1.22093165e+00 -6.59641087e-01 4.88709688e-01 -1.04374625e-02
3.32055241e-01 6.62995934e-01 1.04112792e+00 -2.56266475e-01
-4.36384588e-01 -1.06003629e-02 3.05871695e-01 1.75323829e-01
1.12633884e-01 4.50120568e-01 -1.28045380e+00 -2.42872313e-02
-1.95366770e-01 -5.93258679e-01 4.68860716e-01 4.62723017e-01
1.22657037e+00 -2.99237847e-01 4.36106116e-01 9.01327491e-01
8.34526539e-01 3.27004552e-01 5.33240199e-01 1.55262187e-01
5.96577346e-01 4.03233945e-01 2.53515184e-01 6.81307375e-01
-6.57710657e-02 8.21933448e-01 1.90396085e-01 1.15400985e-01
-5.00571132e-01 -5.25197089e-01 4.85767961e-01 1.55223119e+00
-6.08984828e-01 7.87517950e-02 -6.62163258e-01 6.10446811e-01
-1.52170229e+00 -1.60086012e+00 8.07709470e-02 2.04877543e+00
1.41213715e+00 2.78143048e-01 4.01446313e-01 6.09017730e-01
3.99552107e-01 2.39789486e-01 -6.06356263e-01 -4.59247231e-01
-3.10844660e-01 7.84814775e-01 -2.66970396e-01 2.79191852e-01
-9.40618753e-01 6.91145182e-01 7.59759331e+00 8.56771231e-01
-1.24819660e+00 -2.80476034e-01 1.23184994e-01 -9.01995063e-01
-4.03240174e-01 -1.68954104e-01 -3.65197301e-01 3.80627513e-01
1.32570279e+00 -2.82366097e-01 1.07803023e+00 4.73458439e-01
1.91711202e-01 6.96314096e-01 -1.45969343e+00 1.19497669e+00
-6.12692945e-02 -1.89455271e+00 1.31193489e-01 -2.87720084e-01
8.39369714e-01 -6.95311278e-02 4.59785908e-01 3.98787618e-01
4.06959623e-01 -1.43067014e+00 1.45692599e+00 8.45417082e-01
1.25689566e+00 -8.15082252e-01 -4.16368954e-02 1.02694169e-01
-1.53208065e+00 8.95212218e-03 4.07699682e-02 -3.69419396e-01
3.48954618e-01 2.37130925e-01 -6.80371881e-01 2.81869590e-01
6.51162565e-01 1.01983309e+00 -2.17781246e-01 1.00525594e+00
-1.07759885e-01 1.08940923e+00 -2.90840924e-01 3.35686922e-01
2.32011266e-02 -4.47589532e-02 7.07484186e-01 1.18306768e+00
7.24681258e-01 -4.08467092e-02 -1.04493618e-01 9.59199369e-01
-3.12797308e-01 -2.40594327e-01 -3.62749070e-01 -5.22311449e-01
8.17593098e-01 1.18801928e+00 -5.36525190e-01 -3.40547025e-01
2.02939078e-01 6.92916572e-01 1.83420867e-01 1.42529100e-01
-8.32651675e-01 -6.34649873e-01 7.00250804e-01 -9.90057662e-02
1.31617442e-01 -3.70670021e-01 -4.72736210e-01 -8.96800637e-01
-1.83689877e-01 -1.28388166e+00 1.11575276e-01 -1.21232283e+00
-1.20624638e+00 8.05233240e-01 -3.58241946e-01 -1.56294191e+00
-8.87249470e-01 -2.00055122e-01 -8.04396987e-01 6.17093444e-01
-6.61229789e-01 -9.56056118e-01 -1.27381265e-01 2.77332813e-01
5.16157329e-01 -3.30008000e-01 1.33980894e+00 4.81105655e-01
2.68757455e-02 5.10487616e-01 -1.62486453e-02 2.26558775e-01
9.04474497e-01 -1.47581470e+00 7.84294367e-01 1.29819930e-01
1.22889566e+00 5.68296432e-01 1.02804005e+00 -1.68808758e-01
-1.02329683e+00 -9.30013359e-01 8.00575316e-01 -7.01667428e-01
9.12307024e-01 -4.94756311e-01 -7.80226588e-01 7.74317026e-01
2.76568174e-01 -6.20691895e-01 1.13983405e+00 4.12948698e-01
-6.19861960e-01 -2.15576366e-01 -3.00153166e-01 8.14461648e-01
1.27400100e+00 -1.06125546e+00 -8.49886477e-01 1.67620450e-01
8.47089767e-01 -6.63388669e-01 -1.12575376e+00 3.27757537e-01
1.10979497e+00 -8.58316958e-01 1.03958082e+00 -7.87129819e-01
6.80164397e-01 -4.08147544e-01 1.58753321e-02 -1.32321203e+00
-4.14425462e-01 -1.42834270e+00 -3.99408311e-01 1.41254044e+00
3.53717446e-01 4.01694864e-01 7.56992280e-01 -1.74629599e-01
-4.82762754e-01 -2.80324012e-01 -6.56985164e-01 -7.51393795e-01
1.01844314e-02 -8.73381674e-01 8.78460705e-01 9.53131855e-01
-8.78880173e-03 1.77126870e-01 -6.88651562e-01 -2.79325843e-01
4.14613277e-01 5.09642839e-01 8.84293675e-01 -1.64502108e+00
-8.68082523e-01 -7.35088885e-01 -7.84711987e-02 -8.55468750e-01
1.53786436e-01 -1.06073177e+00 2.14306846e-01 -9.44313824e-01
-1.92217439e-01 -3.45360637e-01 -6.02463126e-01 5.63304663e-01
3.54015172e-01 7.32710123e-01 5.29977441e-01 4.12959903e-01
-3.43497843e-01 3.46278012e-01 9.62819040e-01 -2.42244929e-01
-2.89645553e-01 3.23778428e-02 -5.41516721e-01 9.25305545e-01
7.54232347e-01 -5.92176616e-01 -4.13414598e-01 -6.93439722e-01
8.02311718e-01 4.99117762e-01 2.49284536e-01 -1.52987790e+00
2.17301756e-01 -1.17637338e-02 2.55221188e-01 -6.99299037e-01
8.17931533e-01 -1.20862044e-01 7.77905107e-01 2.16086190e-02
-1.05872869e+00 2.17272535e-01 4.77095425e-01 1.94257319e-01
-5.81408083e-01 -1.60189196e-01 2.84053952e-01 -1.20112173e-01
-3.94592524e-01 2.20594183e-01 -4.56553996e-01 2.83688247e-01
1.56514868e-01 2.15666294e-02 -9.83694196e-02 -6.88621044e-01
-1.11912954e+00 -6.20609224e-01 -1.17428936e-02 5.47995627e-01
2.22999573e-01 -1.78221774e+00 -7.45786726e-01 3.21752690e-02
-1.09902564e-02 -2.07148284e-01 2.99575895e-01 2.36261368e-01
-4.70736802e-01 2.78572619e-01 -5.56601107e-01 -3.30318689e-01
-1.26302230e+00 -1.13335975e-01 1.77048758e-01 -1.74649149e-01
-7.15389431e-01 1.02188098e+00 -1.87546238e-01 -4.30245042e-01
4.54984903e-01 -5.11614859e-01 -9.06665996e-02 2.13355660e-01
6.68956935e-01 3.39999229e-01 -2.38200188e-01 -4.16161299e-01
2.50842452e-01 5.68547130e-01 4.47373658e-01 -6.60953522e-01
1.44674742e+00 4.22219634e-01 -1.50665894e-01 1.23140717e+00
6.82849109e-01 5.87690115e-01 -1.38251734e+00 1.10958882e-01
2.39204597e-02 1.40237108e-01 -3.51207227e-01 -9.13859725e-01
-7.37939477e-01 1.00415218e+00 7.35229477e-02 3.52763861e-01
1.00445318e+00 -3.07779938e-01 1.11619294e+00 6.65968955e-01
3.47958684e-01 -1.00484037e+00 3.66107404e-01 8.35623145e-01
1.21010041e+00 -3.25915515e-01 -4.67704743e-01 2.74394542e-01
-4.79960412e-01 1.30227506e+00 1.95845038e-01 -4.30384904e-01
4.03045774e-01 7.14504302e-01 3.58623952e-01 -8.98501948e-02
-1.01445293e+00 3.28812689e-01 6.08189702e-01 2.98702180e-01
7.52569973e-01 2.20029607e-01 3.96371424e-01 1.11966920e+00
-1.31368876e+00 1.12863295e-01 5.45419991e-01 4.84050483e-01
-1.99828908e-01 -1.09289956e+00 -3.50907862e-01 1.66019216e-01
-6.59822404e-01 -2.73079604e-01 -4.38811302e-01 3.40179443e-01
2.59758502e-01 6.64374232e-01 3.46144408e-01 -6.87729955e-01
2.92773336e-01 3.31506044e-01 5.87656736e-01 -5.72998405e-01
-1.13617182e+00 1.52698860e-01 2.92705506e-01 -5.23884892e-01
-3.00965935e-01 -4.78538960e-01 -1.37633526e+00 -2.03887761e-01
3.53615493e-01 2.24638209e-01 6.04334354e-01 5.52380979e-01
4.03565586e-01 1.12580967e+00 4.11812425e-01 -1.37492979e+00
-4.59235996e-01 -1.14259505e+00 -7.74768054e-01 5.40019512e-01
3.01385492e-01 -1.83630392e-01 -2.80031472e-01 7.53596663e-01]
|
[15.95609188079834, 5.500547409057617]
|
ce8cb608-b6d9-46f7-b84c-c0b4913b1432
|
movie-plot-analysis-via-turning-point
|
1908.10328
| null |
https://arxiv.org/abs/1908.10328v2
|
https://arxiv.org/pdf/1908.10328v2.pdf
|
Movie Plot Analysis via Turning Point Identification
|
According to screenwriting theory, turning points (e.g., change of plans, major setback, climax) are crucial narrative moments within a screenplay: they define the plot structure, determine its progression and segment the screenplay into thematic units (e.g., setup, complications, aftermath). We propose the task of turning point identification in movies as a means of analyzing their narrative structure. We argue that turning points and the segmentation they provide can facilitate processing long, complex narratives, such as screenplays, for summarization and question answering. We introduce a dataset consisting of screenplays and plot synopses annotated with turning points and present an end-to-end neural network model that identifies turning points in plot synopses and projects them onto scenes in screenplays. Our model outperforms strong baselines based on state-of-the-art sentence representations and the expected position of turning points.
|
['Frank Keller', 'Pinelopi Papalampidi', 'Mirella Lapata']
|
2019-08-27
|
movie-plot-analysis-via-turning-point-1
|
https://aclanthology.org/D19-1180
|
https://aclanthology.org/D19-1180.pdf
|
ijcnlp-2019-11
|
['turning-point-identification']
|
['natural-language-processing']
|
[ 4.12390858e-01 6.82472512e-02 -4.01612788e-01 -5.22434831e-01
-8.62063289e-01 -1.32952321e+00 1.01758361e+00 6.89710915e-01
3.32287923e-02 2.38150924e-01 1.20047402e+00 -2.98975140e-01
1.31550118e-01 -7.63461292e-01 -7.64701426e-01 3.00362229e-01
1.68292865e-01 3.26368481e-01 2.63064593e-01 -4.96731013e-01
7.21175492e-01 2.61649311e-01 -1.03202045e+00 1.36587763e+00
3.91425133e-01 7.88079083e-01 6.13041222e-03 1.19673908e+00
-3.83522838e-01 1.32984221e+00 -1.11788726e+00 -7.82082081e-01
-4.66827929e-01 -6.91163898e-01 -1.28509045e+00 3.31483215e-01
4.90657091e-01 -5.06796896e-01 -5.56739390e-01 5.51597059e-01
1.47719249e-01 1.45970389e-01 8.09490979e-01 -7.94736028e-01
-3.53352606e-01 1.24310362e+00 -6.82361484e-01 7.46406019e-01
9.15888846e-01 2.89008636e-02 1.34211099e+00 -7.19871759e-01
1.30218565e+00 8.87966037e-01 8.55730057e-01 -8.23831651e-03
-1.02035761e+00 -1.11156113e-01 2.21875980e-01 1.69027314e-01
-7.64301181e-01 -6.35816693e-01 9.97206688e-01 -7.65473545e-01
1.23527181e+00 6.44133508e-01 9.37525928e-01 1.23506343e+00
2.07559973e-01 1.28156972e+00 2.40676746e-01 -1.55633122e-01
1.41448108e-02 -4.16485161e-01 3.69367510e-01 5.64981878e-01
-6.33166373e-01 -9.02963877e-01 -7.19997704e-01 2.40968943e-01
7.45917201e-01 -2.62006670e-01 -1.24213107e-01 2.29637787e-01
-1.20155466e+00 5.76793134e-01 3.25884283e-01 1.21808231e-01
-3.02791327e-01 2.89687663e-02 8.94881070e-01 -6.36534765e-02
4.36951876e-01 9.10011351e-01 -8.38901699e-02 -9.78849173e-01
-1.39949858e+00 6.91733539e-01 8.30269277e-01 8.07453215e-01
8.90774950e-02 -2.95538992e-01 -7.16585815e-01 8.47963810e-01
-4.05544013e-01 -2.79637069e-01 8.57202932e-02 -9.82174098e-01
1.30631781e+00 8.38907123e-01 1.67599186e-01 -1.37351358e+00
-7.86968887e-01 -1.09452575e-01 -3.48221391e-01 -4.26576883e-01
3.15534621e-01 -3.48332644e-01 -5.58784246e-01 1.35636175e+00
-1.39768533e-02 -1.35919392e-01 -2.78340518e-01 6.70902669e-01
1.49874377e+00 1.17183018e+00 -3.16249043e-01 -1.84435487e-01
1.63872826e+00 -9.39078093e-01 -8.50027382e-01 -2.51377076e-01
4.50794935e-01 -8.10941935e-01 1.51879466e+00 4.32352930e-01
-1.52340245e+00 -5.29405057e-01 -1.08112442e+00 -5.69623113e-01
2.66212989e-02 4.10385698e-01 3.30208093e-01 -4.47603613e-02
-3.80368769e-01 9.05203819e-01 -9.57942545e-01 -3.44757169e-01
5.36004186e-01 -3.44037980e-01 -1.00012913e-01 6.83025658e-01
-7.92922258e-01 4.18301880e-01 3.04666102e-01 5.98756503e-03
-5.95465541e-01 -1.01592827e+00 -9.10556376e-01 1.65711641e-01
6.44246459e-01 -1.05771638e-01 1.76780438e+00 -7.44302213e-01
-1.58002496e+00 1.23412991e+00 -2.95779333e-02 -4.59485829e-01
5.76818466e-01 -6.22257888e-01 -1.85015455e-01 3.99416953e-01
1.91349313e-01 3.73853981e-01 5.64416766e-01 -8.86711597e-01
-6.18144631e-01 -3.80418487e-02 4.35112804e-01 5.01341760e-01
-7.54357409e-03 6.15119159e-01 -8.83946538e-01 -5.55127978e-01
1.32213861e-01 -4.94295865e-01 2.22012773e-01 -6.28640890e-01
-1.31112480e+00 -1.19414859e-01 4.81122464e-01 -1.18114710e+00
2.08444428e+00 -1.87602186e+00 5.11892319e-01 -3.30643415e-01
4.90053833e-01 -2.80961782e-01 2.39897758e-01 9.70319629e-01
-4.44038697e-02 2.18452334e-01 9.44089983e-03 -2.36968398e-01
1.22872904e-01 -3.90317976e-01 -6.00179315e-01 1.03742316e-01
2.59901702e-01 1.09583521e+00 -8.10970902e-01 -4.96544540e-01
-9.78787690e-02 -4.95080352e-02 -5.36061347e-01 9.51300561e-02
-5.99637747e-01 5.28250076e-02 -2.65860856e-01 4.59984541e-01
8.13603923e-02 -4.57761109e-01 9.95572209e-02 -3.18225294e-01
-4.52068895e-01 8.77410948e-01 -7.50887752e-01 1.72771752e+00
-5.74570969e-02 1.47505939e+00 -4.23036635e-01 -4.38622326e-01
6.15526259e-01 -5.63043654e-02 1.25756472e-01 -5.79247773e-01
2.40443096e-01 -5.67954779e-01 -2.92904586e-01 -8.95278156e-01
1.37769353e+00 2.83259034e-01 -1.07876647e+00 5.25827885e-01
-1.32152915e-01 -5.21506727e-01 5.30941665e-01 6.09057963e-01
1.28198302e+00 1.42850429e-01 4.74592358e-01 3.63139510e-01
-6.56238245e-03 4.47042763e-01 2.49824569e-01 6.17848754e-01
8.36370140e-02 1.06609046e+00 1.69109309e+00 -5.64265609e-01
-1.31108809e+00 -9.84368503e-01 4.30807263e-01 1.24737465e+00
-4.30765636e-02 -1.29653573e+00 -1.01053452e+00 -5.65524995e-01
-5.01053512e-01 9.45327461e-01 -7.75155306e-01 4.56463210e-02
-1.06157422e+00 -2.29834780e-01 8.76374185e-01 6.11107767e-01
3.88877869e-01 -1.10418785e+00 -1.07463658e+00 3.06022257e-01
-6.24168575e-01 -1.09742808e+00 -6.25489056e-01 6.77798167e-02
-3.51181239e-01 -1.14682734e+00 -2.65462190e-01 -4.68174994e-01
2.63113916e-01 -3.18164118e-02 1.63508451e+00 -4.16744590e-01
7.24061951e-02 2.77185708e-01 -4.82024491e-01 -2.79229343e-01
-5.13363004e-01 2.40072593e-01 -6.25749588e-01 -2.13836342e-01
-1.32063970e-01 -4.45555687e-01 -3.81455958e-01 -5.37170358e-02
-8.01436186e-01 7.06291139e-01 -5.25774471e-02 2.31755674e-01
6.16112471e-01 -7.43821934e-02 -5.17774783e-02 -1.39066982e+00
1.19968212e+00 -4.17613626e-01 -8.71005431e-02 2.48524830e-01
4.50947464e-01 -3.06061596e-01 7.31953263e-01 -3.45024675e-01
-1.05812681e+00 -3.16548645e-01 -1.38655007e-01 -1.12721294e-01
-1.08838640e-01 8.07443738e-01 -1.20896026e-01 9.54567194e-01
7.47235894e-01 -4.20718752e-02 -6.41162038e-01 -2.71124035e-01
6.95379615e-01 5.44910371e-01 1.08122075e+00 -3.30972880e-01
4.84448731e-01 4.40891147e-01 -2.75224268e-01 -9.13060129e-01
-1.05949461e+00 -3.36987615e-01 -8.42519403e-01 -8.27055752e-01
9.71256375e-01 -6.71586156e-01 -6.91076279e-01 4.11168396e-01
-1.48294151e+00 -4.84357417e-01 -4.80113953e-01 -1.48595229e-01
-7.07237959e-01 1.41649127e-01 -8.68097246e-01 -3.51677060e-01
-1.66822448e-01 -6.98795021e-01 1.09456027e+00 5.84099531e-01
-1.21072221e+00 -6.60325825e-01 1.57474905e-01 5.65565109e-01
-5.02147853e-01 1.00190651e+00 1.06272519e+00 -5.63190401e-01
-1.25861213e-01 -2.30729237e-01 4.31879312e-02 -3.28618646e-01
9.26769746e-04 7.42247224e-01 -4.03515339e-01 4.34465677e-01
-3.26289296e-01 -5.73478520e-01 8.00693095e-01 5.39432704e-01
1.20747662e+00 -8.22330177e-01 -2.63963252e-01 7.47726619e-01
9.10694182e-01 2.76347280e-01 6.69243515e-01 4.16592419e-01
6.95099533e-01 5.59561133e-01 5.51431000e-01 8.80642235e-01
5.16023397e-01 4.40474987e-01 2.54915476e-01 -5.81767457e-03
-1.07529938e-01 -6.18788302e-01 4.19287056e-01 7.20734775e-01
7.45941699e-02 -5.31087935e-01 -1.06501341e+00 4.91882861e-01
-1.71472692e+00 -1.32677531e+00 -5.73575616e-01 1.52485800e+00
8.99575651e-01 6.79219365e-01 4.78250772e-01 6.87165111e-02
5.81223786e-01 9.86459434e-01 -3.98918062e-01 -9.40832019e-01
-2.37946928e-01 -1.33543074e-01 -7.99261183e-02 3.31009686e-01
-1.39488351e+00 1.18217230e+00 6.19210148e+00 6.69071972e-01
-9.87638295e-01 -1.98134229e-01 9.04800236e-01 -5.30775487e-01
-2.03649402e-01 -1.67809620e-01 -6.15811944e-01 3.36704612e-01
7.02199042e-01 1.05204388e-01 2.65144765e-01 4.99544561e-01
5.89128673e-01 -2.21490636e-01 -1.34562445e+00 8.79071474e-01
4.50591803e-01 -2.06287456e+00 -3.04776896e-02 -5.09743810e-01
7.60816932e-01 -3.23256880e-01 -7.52476603e-02 2.22742841e-01
1.16064660e-01 -1.02206457e+00 1.28239405e+00 6.52862787e-01
9.37440336e-01 -6.76272094e-01 1.31412640e-01 1.65233970e-01
-1.00593185e+00 1.25659421e-01 1.96339980e-01 -3.65621984e-01
4.93259311e-01 -7.99099952e-02 -1.07584512e+00 4.07463223e-01
3.06334704e-01 9.12465811e-01 -8.30711305e-01 6.95530474e-01
-5.88272393e-01 9.88138318e-01 -1.56525716e-01 -3.89909029e-01
5.01632631e-01 -7.89611414e-02 7.82160401e-01 1.46757293e+00
-1.27977245e-02 3.45020860e-01 -1.51826873e-01 9.75887239e-01
-3.26729536e-01 -6.28402689e-03 -2.53822803e-01 -6.99544191e-01
3.93955469e-01 1.21934962e+00 -1.51952386e+00 -2.19192505e-01
8.74688029e-02 1.14453828e+00 2.27497905e-01 2.97266454e-01
-1.22715569e+00 -5.29084384e-01 3.28495383e-01 4.02568161e-01
3.97105902e-01 -1.49825603e-01 -6.90505683e-01 -9.87095475e-01
1.78314090e-01 -8.70974600e-01 4.45051014e-01 -1.18063998e+00
-6.55510128e-01 6.43502712e-01 -3.46097560e-03 -1.12640834e+00
-2.20424011e-01 -9.74668190e-02 -1.32436895e+00 1.48953825e-01
-5.60714185e-01 -1.05688226e+00 -2.98317701e-01 9.54367071e-02
1.19246459e+00 2.36612946e-01 3.20562452e-01 -2.51817733e-01
-6.35863245e-01 3.31330806e-01 1.54399276e-01 7.82437265e-01
3.52218866e-01 -1.32087576e+00 9.00065124e-01 9.37247813e-01
5.26102960e-01 2.82374948e-01 9.86810863e-01 -8.09013546e-01
-1.03709352e+00 -7.36985028e-01 7.61555493e-01 -8.56413364e-01
9.20509100e-01 -7.36527443e-01 -5.61793625e-01 7.64548838e-01
3.54394704e-01 -9.51055408e-01 8.90975058e-01 5.74386060e-01
-2.98736155e-01 1.72122017e-01 -2.96759486e-01 9.86253917e-01
8.92395020e-01 -6.39761508e-01 -1.06722951e+00 5.16826093e-01
7.41224766e-01 -1.02668989e+00 -4.80062932e-01 -2.93058723e-01
5.56810379e-01 -9.18686569e-01 8.44161510e-01 -1.10118556e+00
1.79189456e+00 2.59265751e-01 3.02729517e-01 -1.17618263e+00
-2.00942829e-01 -1.21778393e+00 -1.25189215e-01 1.63259208e+00
4.16833788e-01 7.25639760e-01 8.34312201e-01 2.58523852e-01
-4.25616950e-01 -7.90844262e-01 -6.15040243e-01 4.26253304e-02
-1.24871083e-01 -7.28747725e-01 4.81360614e-01 7.93483675e-01
6.14276946e-01 8.99166167e-01 -2.29807287e-01 -3.53847593e-01
-8.29670727e-02 2.78671294e-01 8.56683850e-01 -1.00652623e+00
-1.46502495e-01 -9.38796699e-01 -1.06567210e-02 -1.18589699e+00
-1.04387090e-01 -7.47110307e-01 7.00929062e-03 -2.06664300e+00
1.72010362e-01 2.72633761e-01 3.23558778e-01 2.06447676e-01
-7.56044388e-02 -5.83235137e-02 5.16845345e-01 1.77697331e-01
-1.23058653e+00 1.20944418e-01 1.08671522e+00 -3.79278004e-01
-7.55209684e-01 9.93823353e-03 -7.26483285e-01 1.04816079e+00
5.30483186e-01 -2.05340460e-01 -3.43324333e-01 -4.04591143e-01
1.03676593e+00 6.73674345e-01 5.46877310e-02 -8.89697611e-01
2.85536438e-01 -2.28908822e-01 5.41087270e-01 -1.11694074e+00
5.29907823e-01 1.25603527e-01 -1.39975682e-01 -1.39425620e-01
-9.71415043e-01 2.56182373e-01 8.56274143e-02 2.34165460e-01
-1.95532128e-01 -2.14537323e-01 2.50134349e-01 -1.92198738e-01
-4.72032160e-01 -2.03400970e-01 -8.72349620e-01 4.62109655e-01
8.84788215e-01 -6.81306005e-01 -4.97416198e-01 -6.95893645e-01
-7.68441319e-01 2.76269764e-01 2.70962000e-01 5.45576394e-01
7.25734055e-01 -1.06887484e+00 -7.11170554e-01 -4.85244364e-01
-7.51405433e-02 2.12657973e-01 3.58696878e-01 5.38793445e-01
-1.10941577e+00 2.82696672e-02 -2.79289603e-01 -3.49198431e-01
-1.56668127e+00 9.91247445e-02 -5.69587247e-03 -5.32549083e-01
-9.50046420e-01 1.13362277e+00 2.90315673e-02 7.34236166e-02
1.68565288e-01 -9.66102839e-01 -9.27434504e-01 9.23757792e-01
7.06301391e-01 2.64179140e-01 -2.63052613e-01 -6.33169472e-01
-1.65473342e-01 3.14485878e-01 -1.68683931e-01 -3.89095575e-01
1.57664251e+00 -8.78215209e-02 -3.50952409e-02 1.06104040e+00
1.00968575e+00 2.90871233e-01 -1.51737607e+00 2.46282995e-01
1.21755451e-01 -9.52795520e-02 -2.86694974e-01 -8.76625478e-01
-2.64966249e-01 9.89467919e-01 -6.20202422e-01 6.83190465e-01
8.62282872e-01 4.02502090e-01 8.41228187e-01 3.85122925e-01
-3.99033070e-01 -1.46766233e+00 4.34570670e-01 8.87887955e-01
1.27404392e+00 -7.59683073e-01 2.12844566e-01 -2.37903178e-01
-1.27344477e+00 1.47079682e+00 5.15555859e-01 -1.71743497e-01
-4.39830981e-02 3.47070158e-01 -3.56681913e-01 -6.33126915e-01
-8.91294777e-01 2.59265125e-01 4.22331572e-01 1.56137794e-01
3.98389786e-01 1.60353050e-01 -1.14778519e-01 1.32600534e+00
-9.14604127e-01 -4.63366419e-01 1.17537785e+00 8.15904796e-01
-3.59331727e-01 -1.66496798e-01 -3.69900018e-01 6.27017558e-01
-6.67747915e-01 7.49110803e-02 -1.29502082e+00 7.48337269e-01
6.56805420e-03 8.48915160e-01 6.21016741e-01 -5.14370620e-01
6.45365775e-01 -8.64494145e-02 5.34723759e-01 -7.20470846e-01
-1.20680618e+00 1.29419371e-01 7.22335815e-01 -5.43429554e-01
2.11582091e-02 -9.44488347e-01 -1.45852888e+00 -2.88215131e-01
3.21221083e-01 8.33861232e-02 2.89632499e-01 9.10909534e-01
1.91761017e-01 1.18849027e+00 3.64208102e-01 -6.83194280e-01
8.96371454e-02 -1.06265342e+00 -1.63196295e-01 4.93853748e-01
2.12797552e-01 -1.71819419e-01 3.60302806e-01 4.97506469e-01]
|
[12.405367851257324, 9.441787719726562]
|
94ead6d8-c006-419d-8582-2cef2fa8b088
|
storm-a-diffusion-based-stochastic
|
2212.11851
| null |
https://arxiv.org/abs/2212.11851v1
|
https://arxiv.org/pdf/2212.11851v1.pdf
|
StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberation
|
Diffusion models have shown a great ability at bridging the performance gap between predictive and generative approaches for speech enhancement. We have shown that they may even outperform their predictive counterparts for non-additive corruption types or when they are evaluated on mismatched conditions. However, diffusion models suffer from a high computational burden, mainly as they require to run a neural network for each reverse diffusion step, whereas predictive approaches only require one pass. As diffusion models are generative approaches they may also produce vocalizing and breathing artifacts in adverse conditions. In comparison, in such difficult scenarios, predictive models typically do not produce such artifacts but tend to distort the target speech instead, thereby degrading the speech quality. In this work, we present a stochastic regeneration approach where an estimate given by a predictive model is provided as a guide for further diffusion. We show that the proposed approach uses the predictive model to remove the vocalizing and breathing artifacts while producing very high quality samples thanks to the diffusion model, even in adverse conditions. We further show that this approach enables to use lighter sampling schemes with fewer diffusion steps without sacrificing quality, thus lifting the computational burden by an order of magnitude. Source code and audio examples are available online (https://uhh.de/inf-sp-storm).
|
['Timo Gerkmann', 'Simon Welker', 'Julius Richter', 'Jean-Marie Lemercier']
|
2022-12-22
| null | null | null | null |
['speech-dereverberation']
|
['speech']
|
[ 1.90225407e-01 1.96713969e-01 2.29414597e-01 1.85735136e-01
-1.00246322e+00 -4.41657186e-01 5.35904050e-01 -3.45486253e-02
-1.88895434e-01 6.55218780e-01 3.48578542e-01 -1.84379280e-01
-9.44150537e-02 -6.36435807e-01 -4.60338205e-01 -1.03359032e+00
6.25605136e-02 -2.87895426e-02 1.36893466e-01 -2.51095623e-01
-3.95093001e-02 2.28679180e-01 -1.77257097e+00 1.54886112e-01
9.81387019e-01 9.06530797e-01 2.89223641e-01 9.35285807e-01
1.69081256e-01 5.16191661e-01 -9.53045309e-01 -2.54330665e-01
1.42297521e-01 -7.05965281e-01 -2.54051596e-01 -1.28909582e-02
2.76255757e-01 -3.16812277e-01 -3.12352568e-01 1.08416200e+00
1.09879422e+00 1.87510401e-01 5.12183964e-01 -9.05952275e-01
-5.66373885e-01 5.37782907e-01 -2.49481216e-01 4.50905785e-02
4.71849531e-01 3.98145825e-01 8.29953015e-01 -9.04972613e-01
3.55894983e-01 9.75982189e-01 9.83183622e-01 7.26931572e-01
-1.50852382e+00 -5.98731816e-01 -2.45703101e-01 2.05911204e-01
-1.13811612e+00 -9.24418390e-01 9.22579765e-01 -2.03215525e-01
8.01990211e-01 4.87470329e-01 7.66271651e-01 1.29398584e+00
-1.49758831e-01 7.10360050e-01 1.17951906e+00 -4.74046975e-01
2.66298741e-01 1.66667342e-01 -1.39314085e-01 2.90981948e-01
-1.43434420e-01 5.19600987e-01 -6.14366531e-01 -1.66760683e-01
3.55324984e-01 -3.57134849e-01 -7.51357615e-01 1.44804716e-01
-7.93035090e-01 6.11301839e-01 1.14677019e-01 5.37367225e-01
-7.10584283e-01 2.30196223e-01 2.94751912e-01 4.64287281e-01
7.07975984e-01 3.03961307e-01 -4.16281298e-02 -3.78446370e-01
-1.37584794e+00 2.76155263e-01 8.00064206e-01 5.59307933e-01
4.62917089e-01 5.38161516e-01 -2.24795669e-01 1.11306190e+00
1.95332929e-01 6.48993194e-01 4.81921524e-01 -9.27745640e-01
3.63395870e-01 -2.79374212e-01 1.12090550e-01 -9.50357974e-01
-3.21467549e-01 -7.38904178e-01 -1.10425925e+00 4.13339764e-01
3.12611848e-01 -3.25281948e-01 -7.78965235e-01 1.71459162e+00
2.00910345e-01 5.30609302e-02 1.18105598e-01 8.34179759e-01
5.03764033e-01 8.87691438e-01 -2.34310150e-01 -5.99374831e-01
8.76454413e-01 -7.79241860e-01 -1.11423445e+00 6.42962977e-02
2.85124779e-01 -1.15557790e+00 1.15804553e+00 7.52307534e-01
-1.47998512e+00 -5.22290587e-01 -9.43145275e-01 2.75872260e-01
1.42094254e-01 1.23094089e-01 4.27724048e-02 1.12553024e+00
-1.44170892e+00 8.88326347e-01 -6.09287381e-01 -1.26634911e-01
9.08272490e-02 1.33672208e-01 1.05633169e-01 2.77635664e-01
-1.06363225e+00 6.77382708e-01 -3.35594654e-01 1.77861273e-01
-9.20397639e-01 -8.53580296e-01 -6.24822795e-01 1.62223935e-01
2.22626910e-01 -3.54770750e-01 1.35900283e+00 -9.05707657e-01
-2.00790477e+00 1.96540475e-01 -1.16795279e-01 -5.78286529e-01
7.96769619e-01 -4.82077360e-01 -5.44559777e-01 1.83176234e-01
-2.79643923e-01 3.11152905e-01 1.24442303e+00 -1.30167210e+00
-1.55819789e-01 1.60173297e-01 -1.15110926e-01 -1.03886180e-01
-5.16152918e-01 -2.08453774e-01 -1.66790187e-01 -1.00906515e+00
1.01257846e-01 -9.68358457e-01 -2.62638062e-01 2.54897550e-02
-4.07662898e-01 3.19336891e-01 9.28602159e-01 -7.81112790e-01
1.23116481e+00 -2.18450379e+00 7.17088431e-02 4.85182330e-02
6.09496757e-02 6.91980779e-01 -1.01865493e-01 7.24948287e-01
-7.28392042e-03 4.15553413e-02 -4.76300508e-01 -6.96281374e-01
-8.15934017e-02 1.07394643e-01 -5.25893927e-01 5.11693537e-01
1.03697933e-01 3.60432506e-01 -7.14404583e-01 -8.02584812e-02
3.48572403e-01 9.19973195e-01 -8.31624269e-01 2.90957838e-01
-2.11168686e-03 5.50268829e-01 1.80501953e-01 2.55267352e-01
6.79132819e-01 3.36749524e-01 -6.06212616e-02 -1.87572062e-01
-1.37548029e-01 3.75400215e-01 -1.34070039e+00 1.38386858e+00
-1.02870047e+00 9.09396946e-01 5.01396120e-01 -9.42261517e-01
1.04199767e+00 8.04858565e-01 2.86064327e-01 -6.21668577e-01
-7.38774985e-02 4.45677876e-01 1.04639612e-01 -4.53638226e-01
3.73371989e-01 -5.83838522e-01 4.48536485e-01 4.61255640e-01
6.19899295e-02 -4.81672168e-01 -1.16939088e-02 -1.27429096e-02
1.00626504e+00 -4.57419679e-02 -1.64583046e-02 -1.31136551e-01
5.95509887e-01 -5.28924704e-01 4.68165398e-01 5.18508852e-01
-1.15635298e-01 7.52962232e-01 3.64602298e-01 7.21930712e-02
-1.11379790e+00 -1.09306943e+00 -2.18330827e-02 6.01360023e-01
-6.76198974e-02 -3.86457294e-01 -9.22252595e-01 -1.29664794e-01
-3.50513667e-01 9.31161463e-01 -3.38487029e-01 -2.17538148e-01
-6.01916552e-01 -7.20512271e-01 7.41551220e-01 1.98028103e-01
3.00130278e-01 -8.70101213e-01 -3.80269200e-01 4.21742052e-01
-2.33136281e-01 -7.86482453e-01 -4.02819097e-01 2.44617164e-01
-7.98945725e-01 -5.42619228e-01 -1.16001928e+00 -2.96839267e-01
3.91213298e-01 1.24903642e-01 7.85050035e-01 3.36970761e-02
-2.05711164e-02 3.02430451e-01 -3.99733603e-01 -1.42799377e-01
-9.75054860e-01 -2.55817324e-01 2.17246920e-01 1.60635978e-01
-4.03042406e-01 -1.08227563e+00 -6.80771947e-01 3.25385928e-01
-8.77339065e-01 -1.32634178e-01 4.05636251e-01 9.97054696e-01
3.36448014e-01 1.52126491e-01 7.63372123e-01 -5.40714920e-01
9.72163737e-01 -3.44316036e-01 -5.17756522e-01 -4.24325228e-01
-7.04720318e-01 -4.61756773e-02 1.04697096e+00 -5.57488978e-01
-1.13062096e+00 -2.77795017e-01 -8.68257642e-01 -4.90113407e-01
-5.47013320e-02 2.24579006e-01 1.85818225e-02 2.73283944e-02
7.99026549e-01 4.40620333e-01 1.29680689e-02 -8.30947995e-01
2.72197276e-01 7.51109898e-01 3.88215303e-01 -2.40825087e-01
9.72384691e-01 2.80542761e-01 -8.22544619e-02 -1.16329324e+00
-4.39930201e-01 -3.10265929e-01 -1.46445453e-01 -3.99459958e-01
4.04245734e-01 -6.24860048e-01 -3.88969153e-01 5.10210872e-01
-1.09102941e+00 -3.59284163e-01 -6.03519738e-01 4.99972016e-01
-7.63338327e-01 4.52573478e-01 -7.17508852e-01 -1.17835486e+00
-4.58857149e-01 -1.10189164e+00 7.42562473e-01 3.55905108e-02
-2.33892769e-01 -8.68265331e-01 1.19044863e-01 8.39437023e-02
8.37222099e-01 -2.58642603e-02 6.84977293e-01 -2.56458014e-01
-2.00906828e-01 -6.55398890e-02 3.02795827e-01 6.52023077e-01
1.69600695e-01 2.68353280e-02 -1.37160242e+00 -2.70769805e-01
3.75530154e-01 5.08399047e-02 7.64902651e-01 4.83429283e-01
8.46516132e-01 -6.28278553e-01 1.43537179e-01 5.85550189e-01
1.37719345e+00 1.08866356e-01 7.43398905e-01 -2.85240002e-02
1.86626107e-01 5.81816077e-01 3.92964423e-01 5.24769068e-01
-4.00222659e-01 8.57546091e-01 3.17193478e-01 -3.27119380e-01
-8.08483481e-01 -1.20195709e-01 6.53876066e-01 1.15535605e+00
-2.50273198e-01 -4.43782896e-01 -6.77341759e-01 6.51276946e-01
-1.36524570e+00 -1.21726239e+00 -3.82194757e-01 2.30043101e+00
9.60048378e-01 1.72974560e-02 2.21227571e-01 8.85697842e-01
5.82076371e-01 3.03048700e-01 -2.37265483e-01 -6.27235651e-01
-8.67451802e-02 3.93768877e-01 2.94314772e-01 7.91811347e-01
-6.56489193e-01 4.90227580e-01 6.67067146e+00 9.22534883e-01
-1.36712646e+00 4.88949716e-01 3.15430373e-01 -3.41210485e-01
-3.67135912e-01 -3.30716163e-01 -3.77908349e-01 5.20273089e-01
1.19384360e+00 -1.36616349e-01 4.34627593e-01 6.42616808e-01
6.25942945e-01 -1.91593334e-01 -5.80505490e-01 8.59942794e-01
-2.12970003e-01 -1.08832407e+00 -2.28496268e-01 7.71024749e-02
6.27605915e-01 -1.58656448e-01 3.70438576e-01 -9.27344486e-02
-5.81086390e-02 -8.65593016e-01 9.55660582e-01 4.93107945e-01
7.53406048e-01 -8.40993285e-01 5.77872574e-01 5.35271525e-01
-8.77413154e-01 -8.85541290e-02 -2.11387247e-01 -6.87558996e-03
4.40497577e-01 1.15236199e+00 -7.01790273e-01 4.45150644e-01
5.22473335e-01 2.67145306e-01 5.96110113e-02 1.24196780e+00
-5.04600942e-01 1.10933518e+00 -3.21151137e-01 1.25140309e-01
1.13064677e-01 -1.05736546e-01 1.16852760e+00 1.31844950e+00
7.91143417e-01 -2.09604666e-01 -3.61859888e-01 9.02209044e-01
2.79677898e-01 1.12216979e-01 -5.96980751e-01 1.52650252e-01
1.97573870e-01 9.10815358e-01 -3.41651678e-01 -2.39208281e-01
-3.56167443e-02 1.07638526e+00 -2.39856318e-01 4.30465132e-01
-9.55544114e-01 -4.43623662e-01 6.76590741e-01 2.70853668e-01
3.43886733e-01 -2.98153788e-01 -1.69719443e-01 -7.53059208e-01
6.23186911e-03 -8.94422293e-01 -2.36525923e-01 -6.80455387e-01
-8.32371712e-01 8.23909581e-01 -3.84505332e-01 -1.38657844e+00
-4.90346044e-01 -1.10063590e-01 -5.57723224e-01 1.06569207e+00
-1.42041159e+00 -6.33345425e-01 -1.18048504e-01 4.59481001e-01
7.45804548e-01 1.56692743e-01 7.92964339e-01 6.38516605e-01
-3.44821930e-01 5.98932564e-01 3.48217010e-01 -4.42408472e-01
6.07714415e-01 -1.12275648e+00 1.59681007e-01 9.20993865e-01
1.80627152e-01 4.05831695e-01 1.35655701e+00 -3.53237629e-01
-1.24534118e+00 -9.06599641e-01 8.29213560e-01 2.30607614e-01
4.88777846e-01 -3.45976949e-01 -1.07295024e+00 -3.47568206e-02
4.31258023e-01 -1.72409832e-01 6.00540280e-01 6.38800934e-02
-1.42171308e-02 -1.09545983e-01 -1.14375246e+00 4.92001265e-01
8.89158309e-01 -5.70550859e-01 -2.31865704e-01 2.93142557e-01
5.99932790e-01 -2.17289791e-01 -7.57677913e-01 8.38317163e-03
3.42438102e-01 -1.43024611e+00 8.37129414e-01 2.06205085e-01
3.50179762e-01 -3.66737664e-01 -7.99272209e-02 -1.64605570e+00
4.64368872e-02 -1.29998958e+00 -2.80243933e-01 1.52793777e+00
5.52897990e-01 -7.17252672e-01 4.75093395e-01 6.16314374e-02
-1.62716404e-01 -6.07905865e-01 -1.13263130e+00 -1.20338237e+00
-2.16630977e-02 -8.27868700e-01 2.56023794e-01 4.96911854e-01
-5.11052422e-02 8.55301917e-02 -7.65970349e-01 2.56914169e-01
6.22165859e-01 -1.98299587e-01 5.45726240e-01 -6.75354958e-01
-6.55187905e-01 -5.31896532e-01 -1.98720008e-01 -1.14004827e+00
-2.21601978e-01 -5.83077967e-01 3.81883025e-01 -1.31422055e+00
-4.07070309e-01 -7.07754433e-01 2.42443196e-02 1.15452647e-01
-8.38183761e-02 5.49329638e-01 2.27125406e-01 2.82881051e-01
2.31708273e-01 7.62907863e-01 1.17196441e+00 -8.52335989e-02
-3.94211322e-01 4.82596695e-01 -4.00197268e-01 7.29722440e-01
9.16149080e-01 -6.05339110e-01 -5.25883079e-01 -2.64687806e-01
5.08012809e-02 2.12625787e-01 3.95735055e-01 -1.50550187e+00
9.43005383e-02 3.29477400e-01 -5.78895994e-02 -2.91082561e-01
7.21218526e-01 -6.92267776e-01 3.99375916e-01 7.44301140e-01
-3.05954516e-01 -3.22235107e-01 3.59657496e-01 6.14188254e-01
-4.75914359e-01 -2.91473657e-01 1.07754707e+00 3.68283331e-01
-5.54122310e-03 -5.67349158e-02 -8.37264061e-01 -3.30961555e-01
6.41949832e-01 -1.93353042e-01 1.04403965e-01 -9.35115516e-01
-1.03869033e+00 -4.62859154e-01 3.14634055e-01 8.76683649e-03
4.85983312e-01 -1.08770108e+00 -7.00733483e-01 1.12890907e-01
-5.40545702e-01 -4.39055502e-01 5.18178880e-01 1.21468699e+00
-2.27400050e-01 1.48647040e-01 1.59635335e-01 -5.30571938e-01
-1.40615296e+00 5.68648696e-01 4.21405762e-01 -1.61710307e-01
-8.42559695e-01 9.26619411e-01 8.99796858e-02 -1.76586453e-02
2.99772561e-01 -2.38391131e-01 -7.17664137e-02 1.82758451e-01
5.31555653e-01 4.51231211e-01 3.38274390e-01 -5.13134122e-01
-2.39172578e-02 4.87042725e-01 4.50902373e-01 -5.43925941e-01
1.37124300e+00 -2.52255619e-01 3.07995826e-01 3.61182868e-01
1.11352873e+00 7.07715273e-01 -1.23753977e+00 -9.65854377e-02
-2.84178436e-01 -5.08299530e-01 2.77656913e-01 -7.43232191e-01
-1.22122049e+00 1.24084842e+00 7.37375319e-01 5.78949809e-01
1.69491160e+00 -2.38498941e-01 9.99174416e-01 -1.76218212e-01
1.05028450e-01 -1.00332499e+00 -1.29029285e-02 2.88116634e-01
1.16554201e+00 -5.19877851e-01 -1.91244498e-01 -4.93647426e-01
-3.68607342e-01 1.07763124e+00 -1.06015198e-01 -6.84351549e-02
4.70083892e-01 6.60081506e-01 2.16848597e-01 2.27220401e-01
-7.20185220e-01 -2.81425834e-01 6.05907850e-02 9.90520716e-01
4.94079083e-01 -4.53026518e-02 -3.67940784e-01 3.09899271e-01
-4.13654387e-01 -2.32148632e-01 5.89323163e-01 5.95679462e-01
-4.12194490e-01 -1.34969473e+00 -6.91358209e-01 1.01683475e-01
-6.23077571e-01 -4.39830929e-01 -1.44784868e-01 4.28810924e-01
-4.31157127e-02 1.27281129e+00 -2.53421396e-01 -4.24578995e-01
4.02763367e-01 1.22436136e-02 4.38284010e-01 -1.99041530e-01
-8.39573920e-01 4.92044479e-01 3.69827151e-01 -5.63879073e-01
-3.45901847e-01 -7.69705832e-01 -8.53196323e-01 -3.94840449e-01
-5.61469674e-01 1.36331871e-01 7.56591141e-01 6.20185673e-01
3.12110752e-01 7.56333947e-01 7.52712667e-01 -9.23012674e-01
-6.32033110e-01 -1.10855544e+00 -5.81838906e-01 1.02115132e-01
6.38839960e-01 -4.70755041e-01 -6.36761665e-01 1.33521765e-01]
|
[15.171443939208984, 5.937561511993408]
|
e990a128-b862-46c7-9410-d0af8ace7c9f
|
similarity-based-android-malware-detection
|
1908.05759
| null |
https://arxiv.org/abs/1908.05759v2
|
https://arxiv.org/pdf/1908.05759v2.pdf
|
Similarity-based Android Malware Detection Using Hamming Distance of Static Binary Features
|
In this paper, we develop four malware detection methods using Hamming distance to find similarity between samples which are first nearest neighbors (FNN), all nearest neighbors (ANN), weighted all nearest neighbors (WANN), and k-medoid based nearest neighbors (KMNN). In our proposed methods, we can trigger the alarm if we detect an Android app is malicious. Hence, our solutions help us to avoid the spread of detected malware on a broader scale. We provide a detailed description of the proposed detection methods and related algorithms. We include an extensive analysis to asses the suitability of our proposed similarity-based detection methods. In this way, we perform our experiments on three datasets, including benign and malware Android apps like Drebin, Contagio, and Genome. Thus, to corroborate the actual effectiveness of our classifier, we carry out performance comparisons with some state-of-the-art classification and malware detection algorithms, namely Mixed and Separated solutions, the program dissimilarity measure based on entropy (PDME) and the FalDroid algorithms. We test our experiments in a different type of features: API, intent, and permission features on these three datasets. The results confirm that accuracy rates of proposed algorithms are more than 90% and in some cases (i.e., considering API features) are more than 99%, and are comparable with existing state-of-the-art solutions.
|
['Mauro Conti', 'Zahra Pooranian', 'Rahim Taheri', 'Meysam Ghahramani', 'Mohammad Shojafar', 'Reza Javidan']
|
2019-08-13
| null | null | null | null |
['android-malware-detection']
|
['miscellaneous']
|
[ 1.85670227e-01 -5.36583006e-01 -5.62011957e-01 -8.11854098e-03
-1.85001254e-01 -6.25500381e-01 7.99269021e-01 3.43147755e-01
-2.52467364e-01 5.28773785e-01 -2.13598281e-01 -4.89209145e-01
-3.10850739e-01 -8.66319299e-01 -1.73344478e-01 -3.89577478e-01
-4.54082400e-01 -1.47949919e-01 6.43387556e-01 -1.16513327e-01
9.33404207e-01 4.31496829e-01 -1.95701301e+00 2.21987173e-01
9.58431542e-01 1.15057051e+00 -3.03227752e-01 5.62337697e-01
1.64742887e-01 5.64618289e-01 -6.87561214e-01 -5.95189333e-01
2.42690057e-01 -1.77472726e-01 -4.84894007e-01 -5.46814919e-01
-1.76323548e-01 -3.99617851e-01 7.85823241e-02 1.22248733e+00
1.21533148e-01 -9.93637741e-02 8.30905616e-01 -1.45474541e+00
-6.13883436e-01 2.73732007e-01 -8.59760582e-01 5.23122668e-01
7.96383321e-01 9.43403170e-02 6.53579473e-01 -5.44581711e-01
4.59354758e-01 9.87893701e-01 1.01435041e+00 2.31948003e-01
-6.85024858e-01 -7.97291219e-01 -2.04529971e-01 4.73951221e-01
-1.50030577e+00 -2.88242161e-01 6.32916510e-01 -6.15798235e-01
7.81330526e-01 5.70832908e-01 6.67843342e-01 1.21473920e+00
6.59361482e-01 2.51568645e-01 1.29740381e+00 -1.66410819e-01
3.18807960e-01 4.01875257e-01 4.69409823e-01 7.22855389e-01
6.87904358e-01 2.47646086e-02 -1.04176909e-01 -9.50080514e-01
-1.40197545e-01 5.97537339e-01 -1.21105812e-01 1.70455705e-02
-7.97047496e-01 1.05496645e+00 8.35130829e-03 5.28396726e-01
-2.79126674e-01 -5.43476403e-01 8.17376733e-01 -9.32318345e-02
3.17092687e-01 7.90049210e-02 -1.33830115e-01 -3.79118472e-01
-9.98241127e-01 -1.07229419e-01 8.81817162e-01 5.21169007e-01
5.42590797e-01 -2.54846245e-01 -2.77254935e-02 5.64152420e-01
5.24337590e-01 2.19539866e-01 1.11409128e+00 -4.58211839e-01
7.43819699e-02 6.37322664e-01 -2.41524622e-01 -1.54806852e+00
1.54082522e-01 7.01990053e-02 -6.19847238e-01 1.32554621e-02
-1.33600071e-01 -8.22168868e-03 -2.91865408e-01 1.20524073e+00
3.84221822e-01 6.80606902e-01 2.49657407e-02 9.60483477e-02
5.64513743e-01 6.28004432e-01 8.92275944e-03 -5.37903249e-01
1.36546779e+00 -5.46346188e-01 -5.77313960e-01 2.27735654e-01
5.24950445e-01 -9.60180759e-01 9.50021207e-01 4.41433787e-01
-3.31093431e-01 -5.36992788e-01 -1.24317777e+00 7.12874711e-01
-7.64346957e-01 -4.40802313e-02 4.04221922e-01 1.21900415e+00
-7.64368534e-01 7.17861652e-01 -7.56617844e-01 -7.61134684e-01
3.25460702e-01 3.17166597e-01 -2.25191250e-01 2.51767069e-01
-8.16839814e-01 4.14463073e-01 4.70775604e-01 -3.72492015e-01
-8.70838344e-01 -2.47185364e-01 -5.62431693e-01 -2.14489445e-01
1.39204174e-01 1.25085101e-01 8.19832325e-01 -5.09834707e-01
-1.21348393e+00 6.19723976e-01 -1.16638020e-01 -4.44663674e-01
1.21033220e-02 7.02792406e-02 -9.56567585e-01 -5.07391281e-02
-2.57487018e-02 4.34863269e-02 6.91126764e-01 -1.00605893e+00
-6.36130393e-01 -5.30778825e-01 1.49684604e-02 -4.14263785e-01
-7.89970100e-01 2.83503473e-01 7.68159404e-02 -4.87311006e-01
-3.66256237e-01 -1.00688756e+00 3.75625312e-01 -4.47047234e-01
-5.29702306e-01 -3.38664681e-01 1.53381836e+00 -8.27939689e-01
1.76390433e+00 -2.48054838e+00 -4.31993723e-01 4.78521198e-01
2.27543429e-01 7.53974557e-01 2.52847940e-01 4.75714535e-01
1.67394817e-01 3.80252302e-01 -2.37790182e-01 7.22363517e-02
-2.37323046e-01 1.32799834e-01 -9.64869484e-02 6.74442947e-01
-1.59947857e-01 2.75162429e-01 -7.73061752e-01 -6.06323004e-01
9.52350795e-02 6.23817801e-01 -3.24213892e-01 1.66991547e-01
1.80642962e-01 -5.74370399e-02 -5.05032003e-01 1.03082740e+00
7.36316979e-01 1.81376010e-01 1.33843452e-01 -2.06742778e-01
-1.66133523e-01 1.09033726e-01 -1.07061768e+00 6.65493965e-01
-3.94576490e-01 4.94184047e-01 -2.53570348e-01 -8.00890565e-01
1.10197663e+00 3.98205630e-02 4.15368021e-01 -3.22149634e-01
3.52517337e-01 4.31592882e-01 4.04136002e-01 -6.92866802e-01
3.35034728e-01 6.72654390e-01 2.12710500e-01 5.58440030e-01
-2.65040785e-01 5.55390775e-01 1.42492652e-01 9.21171680e-02
1.29390287e+00 -4.03519005e-01 9.18757915e-01 -4.03670937e-01
1.04356349e+00 -2.35708401e-01 3.92140597e-01 3.52517813e-01
-6.86372280e-01 -1.15164451e-01 7.13563085e-01 -4.51774374e-02
-6.61337733e-01 -8.10772240e-01 -1.36932313e-01 8.35659862e-01
2.00022057e-01 -7.49946058e-01 -1.05885231e+00 -1.20762229e+00
3.06881815e-02 4.97434586e-01 -7.12910831e-01 -2.65023261e-01
-4.34501231e-01 -8.44151974e-01 8.07365656e-01 6.98160976e-02
7.84129441e-01 -8.22450340e-01 -6.51353776e-01 -2.83083022e-01
2.26088867e-01 -8.48104000e-01 -3.84143174e-01 -2.23276421e-01
-7.30111659e-01 -1.58173037e+00 -2.76906490e-01 -5.74869752e-01
3.55740815e-01 3.69190961e-01 5.82831740e-01 1.96300521e-01
-3.20444047e-01 1.65916577e-01 -7.04380810e-01 -8.56222957e-02
-6.27547383e-01 -1.79432780e-02 2.66078591e-01 1.30643278e-01
6.93026006e-01 -5.02813637e-01 -6.14732683e-01 5.04575670e-01
-7.56340265e-01 -1.03028715e+00 5.90823233e-01 2.86214620e-01
4.45897996e-01 3.99257839e-01 4.86248255e-01 -8.31327736e-01
1.10360384e+00 -1.14198756e+00 -3.83524060e-01 1.30182877e-01
-8.10684085e-01 -2.63082296e-01 8.06524873e-01 -8.30435991e-01
-6.22150004e-01 -7.50519857e-02 -3.28307241e-01 -2.14512974e-01
-1.70894563e-01 6.64883777e-02 -1.08970240e-01 -4.79090542e-01
7.12978780e-01 3.64298195e-01 8.27814937e-02 -2.39808410e-01
-1.28962640e-02 1.50802767e+00 5.81867099e-02 -7.77347758e-02
4.21809614e-01 3.48216385e-01 -2.34418094e-01 -7.88968921e-01
2.90018786e-02 -5.06401777e-01 -2.11197898e-01 -2.87164785e-02
9.28417325e-01 -4.23110962e-01 -1.05579066e+00 4.98138845e-01
-8.98093343e-01 4.44579959e-01 3.28932554e-01 3.39695692e-01
-5.87643944e-02 8.75591815e-01 -5.33920527e-01 -1.06664145e+00
-5.80086052e-01 -1.45430064e+00 7.98526287e-01 1.52070343e-01
-4.84587878e-01 -8.87573123e-01 1.73233092e-01 1.53450280e-01
4.03057456e-01 5.62533140e-01 8.67164910e-01 -1.34099948e+00
1.00760870e-01 -2.89809853e-01 -8.71530771e-02 3.56081516e-01
5.97025096e-01 7.00002253e-01 -7.85093367e-01 -2.90979117e-01
2.94351935e-01 2.41950616e-01 5.45017898e-01 -2.03787073e-01
1.31112933e+00 -7.24703074e-01 -8.19662750e-01 2.40395322e-01
1.39653981e+00 8.20424199e-01 5.29101074e-01 2.13773504e-01
4.14796859e-01 4.79269892e-01 8.11237574e-01 5.88445008e-01
2.04323828e-01 5.96681654e-01 7.47904003e-01 5.32117009e-01
1.53762281e-01 -7.64183104e-02 6.83157861e-01 8.83888125e-01
2.97657419e-02 -2.11292550e-01 -8.23714614e-01 3.80561978e-01
-1.47744429e+00 -1.06753767e+00 -3.00705731e-01 2.47188497e+00
5.74974835e-01 2.13710845e-01 6.50146246e-01 5.80254257e-01
1.01917958e+00 1.45699516e-01 -3.79357696e-01 -1.04849827e+00
4.43602622e-01 9.40704122e-02 3.84238273e-01 2.36049339e-01
-1.20184064e+00 4.09963846e-01 6.03214836e+00 1.25376344e+00
-1.13913572e+00 4.00132358e-01 6.30076528e-01 2.76380658e-01
1.59799278e-01 -2.66688198e-01 -7.84718752e-01 1.17508304e+00
1.16607320e+00 4.47781943e-02 5.64821661e-01 1.19064939e+00
-4.54205312e-02 -1.65173039e-01 -8.40416610e-01 1.06701732e+00
4.53661144e-01 -9.34564769e-01 -1.61413670e-01 3.69722217e-01
6.20686769e-01 -1.67470753e-01 1.25139043e-01 2.26797640e-01
3.18752788e-02 -7.33076036e-01 1.15580350e-01 1.74413979e-01
4.42422807e-01 -9.23754752e-01 8.33532691e-01 3.89773667e-01
-1.39591086e+00 -5.31526983e-01 -1.55169189e-01 9.48519930e-02
-3.39884669e-01 6.34037375e-01 -7.93567240e-01 4.42773312e-01
7.92449117e-01 7.04744339e-01 -8.05100799e-01 1.03564990e+00
2.57889926e-01 5.93980193e-01 -9.58640203e-02 -5.10103106e-01
6.94995522e-02 -5.28488867e-02 3.03350747e-01 1.13393044e+00
6.81759596e-01 -1.65670887e-01 -8.46130401e-02 6.20703459e-01
2.35008433e-01 3.91612470e-01 -9.14824486e-01 -1.28481984e-01
8.35187018e-01 1.35100806e+00 -7.26493657e-01 -4.99828249e-01
-4.29773927e-01 7.08149672e-01 -1.12928651e-01 -2.32406020e-01
-1.07522404e+00 -8.39042902e-01 8.43713939e-01 2.01053530e-01
2.37891451e-01 -1.11950152e-02 8.33615884e-02 -7.96722651e-01
5.43640144e-02 -1.19388592e+00 4.09317940e-01 -1.80170223e-01
-1.30044627e+00 7.62490392e-01 1.57073945e-01 -1.48229253e+00
-3.92878503e-01 -6.81959510e-01 -1.00459385e+00 2.87477732e-01
-9.19051945e-01 -5.66945732e-01 -2.88367957e-01 6.27436042e-01
1.78548157e-01 -3.91863793e-01 5.65306664e-01 5.17447233e-01
-6.17403924e-01 7.22367287e-01 2.94807971e-01 3.25677847e-03
4.59921867e-01 -6.60755336e-01 1.20052375e-01 7.05659270e-01
-1.63449138e-01 1.06840503e+00 3.58825773e-01 -1.04609060e+00
-1.36862683e+00 -1.05263305e+00 4.65502560e-01 -1.99641123e-01
7.17273116e-01 -2.66064316e-01 -7.49663293e-01 4.28833544e-01
2.32356235e-01 -5.08820266e-02 9.20626879e-01 -3.42858672e-01
-4.80783015e-01 -1.52080253e-01 -1.83146715e+00 3.56811047e-01
8.13549995e-01 -3.12236756e-01 -2.56936818e-01 3.06552351e-01
4.62706029e-01 2.16685519e-01 -9.83121872e-01 6.33239865e-01
8.07023525e-01 -1.57049727e+00 8.64654124e-01 -5.83977699e-02
2.97826771e-02 -3.56341809e-01 -4.70786661e-01 -7.41636634e-01
1.28132194e-01 -3.51567209e-01 -4.80693728e-01 1.47427380e+00
4.48362306e-02 -9.65321958e-01 4.42985386e-01 -2.31377289e-01
2.87602246e-01 -1.08939660e+00 -8.28982115e-01 -1.00050032e+00
-3.80607665e-01 -3.04607779e-01 8.56215179e-01 1.12542665e+00
1.78979516e-01 -1.51704088e-01 -1.91731676e-01 7.36491978e-02
4.67689812e-01 -3.34204972e-01 5.16256154e-01 -1.25332606e+00
-3.30973357e-01 -5.29406846e-01 -9.93217170e-01 -1.87837645e-01
1.32162079e-01 -5.66777289e-01 -4.72961634e-01 -7.57660329e-01
5.65251648e-01 -2.89136827e-01 -1.60004273e-01 3.51612717e-01
-7.75757506e-02 5.10253847e-01 -6.21860893e-03 3.80134791e-01
-4.00161892e-01 9.75931808e-02 5.25421321e-01 -1.48558132e-02
-2.80764580e-01 1.45107567e-01 -6.60108387e-01 9.14047062e-01
1.06305742e+00 -4.28252727e-01 -4.45592493e-01 2.46999264e-01
-2.38106012e-01 -4.82792169e-01 2.10808143e-01 -1.28709650e+00
-1.08957328e-01 -1.68363098e-02 1.81927904e-01 -4.07877803e-01
2.25895956e-01 -7.99797654e-01 1.12832189e-01 1.05175757e+00
8.13425705e-02 3.27068657e-01 -8.79676566e-02 4.60783154e-01
-2.67750048e-03 -3.97009671e-01 7.86782801e-01 2.17930645e-01
-5.60230613e-01 9.15256515e-03 -4.92261857e-01 -2.99153864e-01
1.66335130e+00 -5.20415902e-01 -6.49923146e-01 -7.93357193e-02
-1.74319655e-01 -4.08080667e-01 6.62733078e-01 3.75493050e-01
6.40699327e-01 -1.30418408e+00 -2.12164029e-01 3.19435954e-01
2.74307728e-01 -1.16461885e+00 -1.58391476e-01 9.60242987e-01
-5.93330920e-01 3.05584669e-01 -4.06295657e-01 -5.48959374e-01
-1.75636244e+00 8.66903841e-01 -3.40090133e-02 -1.34623989e-01
8.25399160e-02 2.60662943e-01 -4.69712317e-01 -3.63169402e-01
1.62162483e-01 -2.04926789e-01 -5.91439009e-01 -1.16196722e-01
6.84510946e-01 1.06161308e+00 2.84228269e-02 -9.62610960e-01
-9.20289934e-01 7.49556899e-01 -9.21355635e-02 3.68208647e-01
6.97026610e-01 3.34296733e-01 -4.78918135e-01 1.60729006e-01
1.46431816e+00 6.26061916e-01 -4.11311686e-01 4.50196385e-01
1.63607933e-02 -7.15031981e-01 -4.23306376e-01 -4.33204949e-01
-8.09335172e-01 6.80256367e-01 1.26923907e+00 8.35659921e-01
1.07452989e+00 -1.42379701e-01 9.14771378e-01 1.43423170e-01
2.71819800e-01 -4.99461293e-01 1.85531005e-02 1.86584100e-01
2.16440916e-01 -1.18367195e+00 9.97273847e-02 -5.77608049e-01
-3.19491655e-01 9.30938423e-01 6.86971188e-01 -2.15890482e-01
9.54206944e-01 2.02478424e-01 -3.99745464e-01 -9.40564275e-03
-3.57430816e-01 1.31562948e-02 2.53849328e-01 8.03056121e-01
5.46708167e-01 5.68502843e-02 -7.87653863e-01 3.92021596e-01
-1.59040570e-01 -3.08799803e-01 3.89663041e-01 9.40370798e-01
-5.74696302e-01 -1.17139971e+00 -4.97353226e-01 7.26800203e-01
-6.48869097e-01 7.95761645e-02 -8.20660532e-01 8.73478889e-01
6.10693276e-01 1.34049726e+00 -8.21692720e-02 -1.14366579e+00
-5.74762374e-02 -1.82666674e-01 5.79514503e-02 -3.47025573e-01
-6.42996192e-01 -4.40214366e-01 -5.14213294e-02 -5.49085736e-01
-2.87851334e-01 -5.54864109e-01 -9.93087590e-01 -6.70351565e-01
-4.58644837e-01 1.66233391e-01 7.44793892e-01 7.96104670e-01
6.56102598e-01 2.10652083e-01 8.76048923e-01 -5.21958888e-01
-4.83474612e-01 -9.00652945e-01 -2.72561133e-01 2.66018450e-01
1.03749439e-01 -7.36163080e-01 -5.61952055e-01 -3.21512163e-01]
|
[14.415809631347656, 9.673895835876465]
|
c272578b-b73f-47a4-adfe-6383648e84b2
|
cn-lbp-complex-networks-based-local-binary
|
2105.06652
| null |
https://arxiv.org/abs/2105.06652v3
|
https://arxiv.org/pdf/2105.06652v3.pdf
|
CN-LBP: Complex Networks-based Local Binary Patterns for Texture Classification
|
To overcome the limitations of original local binary patterns (LBP), this article proposes a new texture descriptor aided by complex networks (CN) and LBP, named CN-LBP. Specifically, we first abstract a texture image (TI) as directed graphs over different bands with the help of pixel distance, intensity, and gradient (magnitude and angle). Second, several CN-based feature measurements (including clustering coefficient, in-degree centrality, out-degree centrality, and eigenvector centrality) are selected to further decipher the texture features, which generates four feature images that can retain the image information as much as possible. Third, given the original TIs, gradient images (GI), and generated feature images, we can obtain the discriminative representation of texture images based on uniform LBP (ULBP). Finally, the feature vector is obtained by jointly calculating and concatenating the spatial histograms. In contrast to original LBP, the proposed texture descriptor contains more detailed image information, and shows resistance to imaging and noise. Experiment results on four datasets demonstrate that the proposed texture descriptor can significantly improve the classification accuracies compared with the state-of-the-art LBP-based variants and deep learning-based methods.
|
['Zhengrui Huang']
|
2021-05-14
| null | null | null | null |
['texture-classification']
|
['computer-vision']
|
[ 2.04324886e-01 -5.86432993e-01 -3.32855552e-01 -2.04743832e-01
-2.44384676e-01 5.30603761e-03 2.64266551e-01 7.03992918e-02
7.19141588e-02 6.17636323e-01 -2.53181085e-02 1.04798339e-01
-7.03043938e-01 -1.17367637e+00 -1.29545376e-01 -1.29061592e+00
-4.83494490e-01 -1.28163800e-01 4.19125736e-01 -1.77527685e-02
7.00879097e-01 8.37823689e-01 -1.68232131e+00 4.11534071e-01
7.58119583e-01 1.95396805e+00 -2.41771117e-02 -7.39323348e-02
-9.67166200e-02 8.20081890e-01 -5.28356254e-01 2.35159963e-01
-1.07543744e-01 -2.11736992e-01 -6.03897274e-01 2.66938984e-01
2.83191651e-01 -7.65299574e-02 -4.47340131e-01 1.00053060e+00
4.02159870e-01 3.43083680e-01 6.89140975e-01 -8.22388291e-01
-1.00400460e+00 -1.20971844e-01 -7.87598431e-01 1.84739143e-01
3.28378141e-01 -2.44596064e-01 5.67908764e-01 -7.64684021e-01
5.46797216e-01 1.34554279e+00 7.76328623e-01 -3.91784966e-01
-1.15563631e+00 -4.03422773e-01 -4.01982278e-01 8.58307898e-01
-1.68095553e+00 -3.07868719e-02 1.12049639e+00 -9.89852026e-02
5.86476624e-01 4.33223128e-01 6.24731481e-01 7.72202730e-01
6.91798747e-01 4.44457740e-01 1.76964188e+00 -5.33939064e-01
6.42815307e-02 -5.41046143e-01 1.00229405e-01 1.19870925e+00
1.87752053e-01 -1.38057053e-01 -4.95979965e-01 -1.50952995e-01
7.62370467e-01 1.97652549e-01 -3.05735260e-01 -3.41318190e-01
-1.05358779e+00 5.97465992e-01 6.22683346e-01 5.98165452e-01
-4.59720999e-01 -9.53666344e-02 4.49431300e-01 1.52606815e-01
5.58733523e-01 -2.85142332e-01 1.71831250e-01 2.28472985e-02
-5.88333905e-01 -1.90669328e-01 5.76073945e-01 3.01849872e-01
1.31613445e+00 -1.22263700e-01 -3.61985981e-01 1.16727638e+00
7.64834583e-02 7.79510260e-01 5.85060596e-01 -7.08174348e-01
-8.35287943e-03 5.73463619e-01 -3.09684098e-01 -2.38990021e+00
-2.36021191e-01 -2.19120577e-01 -1.41631234e+00 -1.74446851e-01
6.91745728e-02 5.46738625e-01 -6.33315444e-01 1.04893327e+00
1.47840932e-01 3.36064473e-02 -5.74362278e-01 7.57185936e-01
7.41306305e-01 6.27441525e-01 -2.00753540e-01 -1.44240797e-01
1.20729840e+00 -8.10733736e-01 -7.87458122e-01 1.85728222e-01
1.26379086e-02 -6.92502081e-01 6.61606133e-01 4.51410472e-01
-2.67560512e-01 -9.10408437e-01 -9.64073122e-01 1.71488777e-01
-5.38035810e-01 3.95827591e-01 5.39190471e-01 4.88644838e-01
-9.28039670e-01 7.32179582e-01 -7.42930591e-01 -4.27317977e-01
2.40132973e-01 2.96248466e-01 -6.93925798e-01 -3.34806055e-01
-9.19878781e-01 5.31955898e-01 2.41079420e-01 4.75716859e-01
-1.12853423e-02 2.17871845e-01 -9.47539985e-01 1.50380254e-01
3.62500064e-02 2.02333942e-01 1.49469137e-01 -8.00315440e-01
-1.54568624e+00 3.78231525e-01 -3.88921797e-01 3.05060416e-01
-1.58689678e-01 4.98335898e-01 -5.76715052e-01 6.80629194e-01
8.05539861e-02 2.86664367e-01 8.94398987e-01 -1.05479658e+00
-4.83972013e-01 -3.93676728e-01 -5.66282988e-01 -6.95906356e-02
-3.91116738e-01 -2.85510600e-01 -3.45802128e-01 -5.44075549e-01
7.24621594e-01 -5.46843648e-01 2.84873545e-01 1.30497543e-02
-3.71467412e-01 -2.84165949e-01 1.34600651e+00 -9.62755382e-01
1.20349729e+00 -2.44635296e+00 -7.91610181e-02 9.79401290e-01
1.09899558e-01 2.97652483e-02 -2.67543286e-01 3.69108051e-01
7.49081969e-02 3.07781436e-02 -7.99505273e-04 2.88259208e-01
-8.63604844e-02 4.99998093e-01 1.85694039e-01 7.41320252e-01
1.20178536e-01 5.79667330e-01 -6.01373255e-01 -9.31598783e-01
4.33032870e-01 4.86564666e-01 -1.73355907e-01 -3.17604899e-01
4.89862084e-01 3.95416230e-01 -7.44567513e-01 7.73297548e-01
1.10878301e+00 -1.09190926e-01 5.12197733e-01 -7.21111536e-01
6.15662150e-02 -1.27697900e-01 -9.88274574e-01 8.45255196e-01
-2.46947154e-01 6.74510062e-01 -3.08809489e-01 -1.63050973e+00
1.63994420e+00 -2.41288543e-01 5.91875970e-01 -1.17303360e+00
1.28801882e-01 2.60398865e-01 -2.46344447e-01 -6.74114943e-01
9.26762000e-02 4.11866665e-01 -6.48637414e-02 2.50203699e-01
4.57552411e-02 3.35507959e-01 2.79157192e-01 -5.47922671e-01
8.97036791e-01 3.79688479e-03 1.57787696e-01 -4.76699382e-01
9.97370720e-01 -3.53801131e-01 4.81551856e-01 6.35198891e-01
-2.92351931e-01 2.49127284e-01 8.02815497e-01 -7.44200528e-01
-6.10839128e-01 -8.22122812e-01 -4.84916568e-01 8.27674210e-01
5.74919641e-01 -3.48977633e-02 -5.43927193e-01 -4.97178078e-01
3.49614859e-01 -3.04449946e-01 -8.14121366e-01 -1.12689257e-01
-6.04323387e-01 -6.86140001e-01 3.69915694e-01 1.22856684e-01
1.21289980e+00 -1.07951045e+00 -1.88260511e-01 -6.94900230e-02
-6.02828562e-01 -6.77205741e-01 -3.13894778e-01 -2.07120217e-02
-6.63111210e-01 -1.06358314e+00 -3.67901266e-01 -1.24039435e+00
7.33878314e-01 3.27500671e-01 4.26015377e-01 1.03283070e-01
-4.36726481e-01 2.30263144e-01 -5.23932815e-01 6.03010476e-01
2.13220567e-01 -1.76882282e-01 -1.59453258e-01 5.77263713e-01
3.80233824e-01 -5.51882267e-01 -7.06233919e-01 5.09453773e-01
-6.88472092e-01 -2.93750644e-01 8.28996658e-01 1.32131338e+00
1.00821030e+00 6.63218856e-01 2.67390698e-01 -2.13997558e-01
7.10180581e-01 -2.19444215e-01 -1.30449876e-01 4.79833096e-01
-2.79380590e-01 -4.09930088e-02 5.92934310e-01 -3.19538295e-01
-1.08800077e+00 -4.56272721e-01 2.18823016e-01 -3.26437354e-02
-1.69279829e-01 8.81964505e-01 2.02360749e-01 -9.37107801e-01
4.02345985e-01 8.32690477e-01 4.53475833e-01 -5.55514276e-01
-8.54262188e-02 9.63167608e-01 4.01157081e-01 -8.31091464e-01
1.13960363e-01 5.48071742e-01 4.77220893e-01 -1.03636599e+00
-3.78306627e-01 -3.35259914e-01 -7.00692713e-01 -2.49710381e-01
5.08088529e-01 -2.44143337e-01 -1.15132093e+00 8.16046476e-01
-7.24332809e-01 4.78296131e-02 1.42720848e-01 3.20973545e-01
-3.94803077e-01 9.96351480e-01 -8.25989008e-01 -4.76211041e-01
-1.42656833e-01 -1.11205518e+00 9.99917626e-01 1.76590502e-01
2.21357062e-01 -1.01142788e+00 -2.70554870e-01 5.76850288e-02
8.07072580e-01 6.71034813e-01 1.18014646e+00 -1.63025055e-02
-3.68315935e-01 -2.76997983e-01 -7.86397696e-01 5.45628786e-01
6.10292435e-01 -2.05696449e-02 -3.06019366e-01 -2.25977361e-01
-1.45951491e-02 -2.19678178e-01 8.22427213e-01 2.83755898e-01
1.79604352e+00 -5.25469959e-01 -4.59706515e-01 6.25484586e-01
1.46742332e+00 3.07842940e-01 7.41814375e-01 5.10565698e-01
3.63816023e-01 4.35483307e-01 4.40319479e-01 3.56029928e-01
2.08941698e-01 5.55849791e-01 3.65984328e-02 -2.12359279e-01
-2.06423670e-01 7.24095153e-04 -9.70643833e-02 1.12495017e+00
-3.61357480e-01 9.85849947e-02 -7.08151460e-01 2.71697551e-01
-1.72724128e+00 -1.11438775e+00 8.41039978e-03 1.66767204e+00
6.13182008e-01 -2.04415873e-01 -3.33645731e-01 3.66104841e-01
1.06078529e+00 5.06593525e-01 -1.72951713e-01 -4.46387768e-01
-3.71246219e-01 6.90339863e-01 5.07935107e-01 9.64579061e-02
-1.24225855e+00 5.91722071e-01 6.00687742e+00 1.60151088e+00
-1.66508448e+00 -1.01055123e-01 8.84952486e-01 8.07600796e-01
2.16725051e-01 -2.52877593e-01 -8.43000188e-02 6.24664068e-01
2.40549773e-01 1.22930616e-01 7.30543077e-01 7.89404690e-01
2.24645231e-02 -5.05075514e-01 -1.68745503e-01 1.19012189e+00
2.16457233e-01 -9.28145528e-01 1.22596897e-01 3.23700085e-02
5.55093229e-01 -3.51755470e-01 3.23813498e-01 1.42074168e-01
-3.59236926e-01 -8.61750901e-01 3.34577680e-01 8.47215414e-01
1.19077909e+00 -7.10711300e-01 1.15459263e+00 -1.24081984e-01
-1.38479412e+00 -2.43959099e-01 -7.83700883e-01 -8.39128122e-02
-5.47392607e-01 8.34964514e-01 -1.93688244e-01 9.89224017e-01
8.97904098e-01 9.76477981e-01 -7.21286178e-01 8.85219932e-01
-4.62333485e-02 4.50889260e-01 -3.30506802e-01 -3.36997718e-01
2.04015493e-01 -5.17540812e-01 -4.71221805e-02 1.23659968e+00
5.06218731e-01 -6.60753399e-02 3.13084364e-01 5.58467388e-01
3.27833921e-01 3.67995948e-01 -3.59763891e-01 3.76133583e-02
3.97622615e-01 1.41460216e+00 -9.77090597e-01 -2.64367908e-01
-1.53814897e-01 9.17273343e-01 2.92336613e-01 5.66692293e-01
-3.45045835e-01 -1.07315862e+00 1.68845281e-01 -2.68541992e-01
5.57933867e-01 -3.81217778e-01 -1.65124536e-01 -7.77794003e-01
9.99968946e-02 -8.23145926e-01 5.75081892e-02 -6.54964685e-01
-1.40491188e+00 5.00593722e-01 -2.06630856e-01 -1.09920692e+00
2.93000102e-01 -8.79213631e-01 -2.93236643e-01 9.38284397e-01
-1.27565169e+00 -1.24963462e+00 -6.18897378e-01 6.49381101e-01
-2.05862865e-01 -1.17844671e-01 7.34156132e-01 2.93576419e-01
-4.77042347e-01 3.93596679e-01 6.72639370e-01 3.00961077e-01
5.91597021e-01 -6.76657557e-01 -3.39408100e-01 3.19438964e-01
-4.58518058e-01 5.28326035e-01 -9.45606753e-02 -5.85106075e-01
-1.31462860e+00 -8.31991732e-01 7.66722798e-01 3.54930341e-01
5.07472277e-01 -1.40014365e-02 -7.08149791e-01 8.90064090e-02
-1.98011994e-01 3.66620392e-01 4.35647845e-01 -1.13107875e-01
-1.72979414e-01 -6.87883556e-01 -1.26618290e+00 2.35484406e-01
8.34519565e-01 -7.32617378e-01 -1.70965686e-01 1.19191445e-01
-8.66753981e-02 -2.08804816e-01 -1.00619221e+00 6.32979691e-01
9.33682203e-01 -1.29879451e+00 8.41160595e-01 4.59901653e-02
2.45982677e-01 -3.39703113e-01 -2.76764661e-01 -8.88223827e-01
-9.10949171e-01 1.44264743e-01 3.08510303e-01 8.15939724e-01
-3.43578726e-01 -9.75274384e-01 4.51841503e-01 -4.64287788e-01
1.09198526e-01 -9.94645059e-01 -1.22282791e+00 -8.40737879e-01
-5.94220519e-01 9.06718448e-02 6.06848121e-01 9.80006635e-01
2.18811542e-01 -4.32220995e-01 -3.78735602e-01 -9.18871239e-02
6.04816079e-01 3.70253235e-01 2.87700176e-01 -1.14587092e+00
1.37081772e-01 -5.60749114e-01 -1.03664589e+00 -8.56719971e-01
1.59249470e-01 -7.55608439e-01 2.30036397e-02 -1.44109917e+00
2.57042378e-01 -6.58062398e-01 -7.13489592e-01 8.14183116e-01
1.80507571e-01 6.47629857e-01 -3.00929934e-01 4.36788023e-01
-5.25129914e-01 5.89113474e-01 1.63944602e+00 -3.95269871e-01
2.09041446e-01 -4.87474710e-01 -2.85447985e-01 3.55903566e-01
5.65764666e-01 -1.98078364e-01 1.51861921e-01 3.93837802e-02
-4.71582443e-01 4.79749404e-02 4.16020602e-01 -1.19901192e+00
1.81430608e-01 -2.73002207e-01 6.89471185e-01 -3.52595299e-01
1.92930028e-01 -5.16849101e-01 8.13139752e-02 6.10524714e-01
1.57168299e-01 -1.08246677e-01 -3.71828340e-02 5.70315063e-01
-8.60193908e-01 1.78344920e-01 5.99342525e-01 2.12451726e-01
-9.93830323e-01 1.91665024e-01 -4.33098406e-01 -7.78052807e-01
6.09261870e-01 -4.93393898e-01 -8.13366830e-01 -1.83999479e-01
-6.58223629e-01 -3.83351177e-01 2.60548145e-01 1.08283050e-01
8.22278380e-01 -1.66216803e+00 -2.84720778e-01 6.64333940e-01
-9.87113193e-02 -5.25977075e-01 5.99605560e-01 1.19086397e+00
-1.15459239e+00 3.88032615e-01 -8.67374420e-01 -8.17013502e-01
-9.89925802e-01 4.15188462e-01 2.53918797e-01 -2.80658662e-01
-3.64798099e-01 3.47863764e-01 -4.55195121e-02 -2.54237831e-01
-9.07809362e-02 -2.14859813e-01 -3.72482032e-01 -7.98883289e-02
1.27736315e-01 6.11629188e-01 1.90874264e-01 -9.32062268e-01
-4.25808132e-01 1.16518080e+00 3.40738058e-01 2.81307548e-01
1.14540553e+00 -1.95110291e-01 -1.02148724e+00 -1.11366019e-01
1.49367881e+00 9.50749293e-02 -6.97560608e-01 -2.75409788e-01
-7.83126429e-02 -8.21488321e-01 1.23025328e-01 -5.48692048e-01
-1.04061699e+00 7.85586119e-01 9.26504314e-01 3.96000892e-01
1.41760170e+00 -4.53416705e-01 7.06122816e-01 4.71492618e-01
5.69691420e-01 -9.42387879e-01 3.17627251e-01 5.90532184e-01
8.30699921e-01 -7.60025442e-01 6.00842126e-02 -5.37409961e-01
-1.82754308e-01 1.51175320e+00 3.31316233e-01 -3.43361199e-01
8.93672049e-01 -1.48071647e-01 -1.15567230e-01 -1.15595013e-01
-3.02367769e-02 -8.64076838e-02 5.79615235e-01 7.76887894e-01
1.72199562e-01 1.25824273e-01 -6.55298471e-01 1.39892861e-01
1.53212354e-01 -6.40288442e-02 -9.97282416e-02 9.08659637e-01
-7.25338161e-01 -1.01002932e+00 -5.39075732e-01 5.65315306e-01
-2.98170656e-01 2.44792506e-01 5.69085451e-03 5.92729151e-01
4.54123050e-01 8.98051858e-01 2.47422487e-01 -7.69408405e-01
-7.29703829e-02 -1.54997528e-01 5.49935818e-01 3.14573757e-02
2.69368947e-01 -2.15624757e-02 -6.69458508e-02 -5.13352036e-01
-6.29297078e-01 -1.26334548e-01 -6.88917100e-01 -6.39061809e-01
-2.16928363e-01 1.61793411e-01 6.05875969e-01 8.84933472e-01
5.22235453e-01 3.11791301e-01 8.18392992e-01 -1.15180218e+00
-1.94997758e-01 -1.01313567e+00 -1.14615941e+00 5.72812021e-01
1.88588485e-01 -1.06441677e+00 -3.90614480e-01 -4.51049097e-02]
|
[10.418937683105469, -0.3842658996582031]
|
f9d2750e-350a-4609-90bd-961d9573ad21
|
multimodal-learning-with-channel-mixing-and
|
2209.12244
| null |
https://arxiv.org/abs/2209.12244v1
|
https://arxiv.org/pdf/2209.12244v1.pdf
|
Multimodal Learning with Channel-Mixing and Masked Autoencoder on Facial Action Unit Detection
|
Recent studies utilizing multi-modal data aimed at building a robust model for facial Action Unit (AU) detection. However, due to the heterogeneity of multi-modal data, multi-modal representation learning becomes one of the main challenges. On one hand, it is difficult to extract the relevant features from multi-modalities by only one feature extractor, on the other hand, previous studies have not fully explored the potential of multi-modal fusion strategies. For example, early fusion usually required all modalities to be present during inference, while late fusion and middle fusion increased the network size for feature learning. In contrast to a large amount of work on late fusion, there are few works on early fusion to explore the channel information. This paper presents a novel multi-modal network called Multi-modal Channel-Mixing (MCM), as a pre-trained model to learn a robust representation in order to facilitate the multi-modal fusion. We evaluate the learned representation on a downstream task of automatic facial action units detection. Specifically, it is a single stream encoder network that uses a channel-mixing module in early fusion, requiring only one modality in the downstream detection task. We also utilize the masked ViT encoder to learn features from the fusion image and reconstruct back two modalities with two ViT decoders. We have conducted extensive experiments on two public datasets, known as BP4D and DISFA, to evaluate the effectiveness and robustness of the proposed multimodal framework. The results show our approach is comparable or superior to the state-of-the-art baseline methods.
|
['Lijun Yin', 'Xiaotian Li', 'Taoyue Wang', 'Huiyuan Yang', 'Xiang Zhang']
|
2022-09-25
| null | null | null | null |
['action-unit-detection', 'facial-action-unit-detection']
|
['computer-vision', 'computer-vision']
|
[ 5.75982869e-01 -5.90641424e-02 -2.61884302e-01 -2.31848910e-01
-1.34511948e+00 -1.83650061e-01 8.49366605e-01 -4.70759720e-01
-3.61922085e-01 5.47044218e-01 4.52145338e-01 1.31157815e-01
3.06949914e-01 -4.72247481e-01 -7.32662797e-01 -8.95183206e-01
2.17320517e-01 -1.03210092e-01 3.06598246e-02 -1.73838511e-01
1.63457822e-02 2.53885984e-01 -2.16427493e+00 9.74539638e-01
4.59139436e-01 1.44424474e+00 1.85901076e-02 6.72571659e-01
9.32192132e-02 7.62592494e-01 -3.65784734e-01 -4.74455118e-01
2.90783018e-01 -7.69160092e-01 -6.04944170e-01 2.27746457e-01
3.76231611e-01 -6.69533610e-01 -3.49132895e-01 9.27613914e-01
9.04735863e-01 -4.65758294e-02 7.12705791e-01 -1.51865828e+00
-2.63717890e-01 5.30364513e-01 -8.99625301e-01 -1.23363234e-01
6.67901397e-01 1.60281718e-01 8.56226623e-01 -1.08469617e+00
4.54407543e-01 1.74994671e+00 3.05306703e-01 9.14123237e-01
-8.33492160e-01 -8.70394111e-01 -5.77163957e-02 3.74008685e-01
-1.33132923e+00 -1.02160501e+00 5.34603000e-01 -8.03189725e-02
7.80725300e-01 8.12173858e-02 3.50810140e-01 1.40427971e+00
1.74265116e-01 1.17377508e+00 1.31060708e+00 -4.68875915e-01
-8.31945240e-02 -4.31934297e-02 -3.44728291e-01 9.24957573e-01
-3.45259100e-01 2.01822445e-01 -9.37028885e-01 -1.62332579e-01
5.28722525e-01 8.57566297e-02 -7.22778589e-02 2.97957420e-01
-1.04971635e+00 8.36144924e-01 3.22436616e-02 3.20107311e-01
-4.46934909e-01 1.83061838e-01 4.86802578e-01 2.99038738e-01
3.57079089e-01 -3.82788390e-01 -2.12175861e-01 -2.73559958e-01
-1.02408278e+00 -7.74527714e-02 5.47080576e-01 6.21433556e-01
9.41658318e-01 -1.31280378e-01 -3.44898164e-01 7.86042273e-01
7.62472212e-01 6.19604409e-01 3.40302110e-01 -1.05731738e+00
6.11918509e-01 5.14510453e-01 -4.73935217e-01 -6.74634516e-01
-2.49722078e-01 3.45748454e-01 -1.00717080e+00 4.57917631e-01
3.21653426e-01 -4.42313582e-01 -1.12326777e+00 1.86705995e+00
2.28501335e-01 5.79136610e-01 3.17160755e-01 7.74454534e-01
1.08367717e+00 6.41201198e-01 9.09486189e-02 -3.47180516e-01
1.52016151e+00 -9.85785723e-01 -7.50732362e-01 -2.40874221e-03
5.00082135e-01 -9.95318651e-01 3.22770983e-01 3.93977076e-01
-1.20398343e+00 -7.02183127e-01 -9.36275423e-01 -1.01705417e-02
-3.59143496e-01 3.05968165e-01 7.40425289e-01 6.75113022e-01
-9.90847170e-01 2.45020851e-01 -8.55765700e-01 -4.64026302e-01
5.26939392e-01 3.67181540e-01 -8.61786783e-01 -3.68993104e-01
-1.23944688e+00 7.61842847e-01 3.61482292e-01 2.61428595e-01
-1.35170722e+00 -1.61002442e-01 -1.08012879e+00 -1.07806332e-01
3.28184366e-01 -7.09640384e-01 1.17095816e+00 -1.21823716e+00
-1.72077131e+00 5.40223241e-01 -5.12963951e-01 -1.36456281e-01
2.06025138e-01 -6.86352327e-02 -5.35759866e-01 5.32752752e-01
-1.04531936e-01 1.06282163e+00 1.31940281e+00 -1.16302490e+00
-1.04656923e+00 -3.22827578e-01 1.75415695e-01 3.35754246e-01
-3.84663194e-01 3.94488811e-01 -6.97967649e-01 -3.94769132e-01
-5.48572093e-02 -6.93408728e-01 6.88027069e-02 -1.05597883e-01
-2.29624867e-01 -2.37670571e-01 8.89001369e-01 -6.18042886e-01
1.22511542e+00 -2.34052515e+00 3.05621445e-01 3.65624540e-02
4.68550771e-02 1.74775064e-01 -5.32773614e-01 4.10941064e-01
-3.07030082e-02 -1.56426197e-03 -4.16894406e-02 -9.00649607e-01
-1.88553482e-02 2.86923051e-01 3.40195447e-02 5.87673426e-01
3.58732700e-01 7.23191321e-01 -5.56829929e-01 -8.20314288e-01
3.29853565e-01 7.76208460e-01 -6.12289667e-01 3.38911325e-01
1.09688573e-01 4.16649520e-01 -2.72813022e-01 1.33871555e+00
7.60737717e-01 1.24055929e-02 -1.58843026e-01 -4.94376779e-01
8.34569335e-02 -1.97790369e-01 -1.46525967e+00 2.05071139e+00
-4.21756595e-01 3.96078020e-01 2.31378496e-01 -9.23954070e-01
3.75172973e-01 8.70829046e-01 6.35148525e-01 -7.55136192e-01
4.92985547e-01 2.02661708e-01 -1.75778046e-01 -5.69457889e-01
4.05726522e-01 -2.75193632e-01 -8.87986124e-02 5.05872726e-01
6.44947529e-01 3.80380481e-01 2.31573358e-01 1.51412740e-01
9.96489108e-01 1.34644240e-01 1.38048917e-01 6.13612175e-01
7.44364738e-01 -5.90135515e-01 5.37983179e-01 5.00867128e-01
-2.66636580e-01 7.53057659e-01 4.27609384e-01 1.87960312e-01
-4.09772575e-01 -7.88141072e-01 7.61175454e-02 1.43419397e+00
4.59004492e-02 -6.36805952e-01 -6.70232356e-01 -8.22865129e-01
-2.12928161e-01 1.40428275e-01 -7.30223238e-01 -2.02135727e-01
-1.97994933e-01 -7.37123966e-01 9.50692654e-01 4.27351862e-01
8.07306468e-01 -9.58262801e-01 -3.35784823e-01 -9.17370766e-02
-5.22740602e-01 -1.18909705e+00 -3.32263380e-01 -1.63469408e-02
-5.17720461e-01 -1.14662480e+00 -7.24218488e-01 -4.91403341e-01
3.96213919e-01 2.73275912e-01 5.49134672e-01 -7.27685094e-02
-1.77058220e-01 5.54789007e-01 -5.55185199e-01 -3.28611851e-01
-4.47412819e-01 -1.15860231e-01 -6.27175272e-02 6.84048533e-01
4.26020145e-01 -2.64989227e-01 -4.93061364e-01 3.08945298e-01
-1.09841609e+00 2.31750845e-03 1.06341922e+00 8.44165564e-01
3.96901757e-01 -5.10757342e-02 5.16945720e-01 -3.30171436e-01
4.06378150e-01 -6.26834571e-01 3.67464684e-02 2.89200515e-01
-8.16681087e-02 -8.83995555e-03 1.87360525e-01 -3.80376965e-01
-1.35597658e+00 3.51469189e-01 -3.40000004e-01 -5.85158706e-01
-4.17749435e-01 5.32707572e-01 -3.86438519e-01 -1.05257288e-01
4.02880669e-01 1.99988991e-01 2.49127761e-01 -4.02606070e-01
6.34809077e-01 1.02513349e+00 3.01854700e-01 -3.86601865e-01
5.11751711e-01 6.15363777e-01 7.69581497e-02 -8.10131848e-01
-4.93004650e-01 -4.92148787e-01 -5.42176485e-01 -4.03537482e-01
9.99397278e-01 -1.23859191e+00 -8.45116675e-01 7.63687789e-01
-1.11088562e+00 1.10144973e-01 2.01997995e-01 6.01038277e-01
-5.71150184e-01 5.68395197e-01 -6.03126287e-01 -8.25414240e-01
-2.00611591e-01 -1.42986083e+00 1.51746964e+00 2.91022301e-01
1.21001691e-01 -6.84769571e-01 -9.39839557e-02 6.97013140e-01
3.09424609e-01 1.64795533e-01 4.07154649e-01 -2.56498069e-01
-4.45276648e-01 -9.80355218e-02 -3.21671814e-01 6.12037063e-01
2.06426769e-01 3.26731689e-02 -1.52989173e+00 -2.71450400e-01
-2.08528787e-01 -8.29059362e-01 1.34709299e+00 2.23318160e-01
8.64672303e-01 5.48134148e-02 -3.72363716e-01 5.10686457e-01
1.09694517e+00 -3.74940559e-02 8.43734384e-01 -7.42857680e-02
5.32887280e-01 3.46257597e-01 7.44264364e-01 6.48488164e-01
4.66978341e-01 4.54049230e-01 6.24682188e-01 -1.65993318e-01
-3.42612088e-01 -1.40389144e-01 9.15820956e-01 5.02784431e-01
-4.74340349e-01 -2.68441230e-01 -3.44840348e-01 2.67990202e-01
-1.94445682e+00 -1.20242929e+00 3.65599990e-01 1.82135272e+00
7.10983813e-01 -2.99889803e-01 2.30892003e-01 1.73935995e-01
6.05297327e-01 2.75247306e-01 -1.70037359e-01 -2.34461814e-01
-3.22882921e-01 2.46127322e-01 3.18082929e-01 3.03634346e-01
-1.30680442e+00 9.22672033e-01 6.08702326e+00 1.18358231e+00
-1.19718063e+00 3.83428782e-01 3.33333522e-01 -1.77808300e-01
5.67288920e-02 -1.01381667e-01 -1.09814620e+00 3.42686266e-01
1.01801741e+00 4.15823042e-01 3.64811361e-01 4.42365676e-01
6.97816163e-02 -4.62774187e-01 -1.14781785e+00 1.34553087e+00
5.36488950e-01 -9.51739192e-01 1.08424939e-01 9.85086933e-02
6.33152723e-01 1.27826128e-02 7.69732296e-02 4.32895213e-01
-1.05705902e-01 -1.13525975e+00 5.60022473e-01 7.63634443e-01
9.10107374e-01 -7.28349507e-01 7.87533343e-01 2.52072811e-01
-1.51726639e+00 -2.60320902e-01 -1.98844016e-01 5.24349287e-02
3.46060693e-01 1.70631424e-01 -5.27232528e-01 7.87534177e-01
5.30085504e-01 8.02037001e-01 -6.50751412e-01 7.00120509e-01
-1.92732066e-01 5.39744854e-01 -4.02795672e-01 4.40060079e-01
1.47358328e-01 3.29725683e-01 2.02273116e-01 1.15077579e+00
4.88797486e-01 -8.60748813e-02 1.15101784e-01 2.76380509e-01
-1.59283638e-01 2.37491485e-02 -6.33693635e-01 -9.18475017e-02
7.33796284e-02 1.47318685e+00 -3.02931547e-01 -2.12614089e-01
-1.00861382e+00 1.08137286e+00 2.42295284e-02 3.15059334e-01
-1.05032241e+00 -5.97531945e-02 5.78651845e-01 -3.86555016e-01
2.92131692e-01 3.57253067e-02 9.61138010e-02 -1.20401776e+00
-1.83784962e-01 -1.10568368e+00 7.76606977e-01 -6.39468670e-01
-1.19754577e+00 4.96438980e-01 5.35860471e-02 -1.28473258e+00
-5.27915955e-01 -6.24554098e-01 -3.01013142e-01 8.00007045e-01
-1.69160533e+00 -1.74644458e+00 -1.86631337e-01 1.25563502e+00
4.59292710e-01 -3.60159755e-01 9.01537240e-01 7.32896507e-01
-6.11728430e-01 9.47701156e-01 -2.82819271e-01 3.38038169e-02
1.02019691e+00 -6.93008840e-01 -4.88351583e-01 8.64694655e-01
2.19943095e-02 3.05132419e-01 1.76180542e-01 -4.16367322e-01
-1.77223098e+00 -8.91704500e-01 6.68629348e-01 -2.00198501e-01
3.36915106e-01 -2.78362244e-01 -4.99097317e-01 6.42595828e-01
6.01960540e-01 1.12540409e-01 8.66845012e-01 -1.35917068e-01
-2.69044816e-01 -1.54077947e-01 -1.05141056e+00 3.49110514e-01
8.57042432e-01 -6.84212565e-01 -3.49076301e-01 -1.22506522e-01
2.42098346e-01 -1.90463319e-01 -8.92782569e-01 5.40831745e-01
8.12388420e-01 -1.12032259e+00 7.09258258e-01 -2.58052886e-01
4.88959342e-01 -3.80785018e-01 -5.74685812e-01 -1.10361695e+00
2.49585370e-03 -5.59925735e-01 -3.68117452e-01 1.56268859e+00
4.55991268e-01 -4.86671776e-01 4.61913466e-01 2.59411454e-01
-6.53384253e-02 -7.32319534e-01 -1.20198727e+00 -2.89282739e-01
-3.27254593e-01 -6.15023613e-01 4.01488394e-01 5.37160397e-01
-6.37319535e-02 3.43376040e-01 -8.04186702e-01 1.91406399e-01
4.84206617e-01 -4.74748015e-03 8.53030086e-01 -8.00916731e-01
-3.02178532e-01 -3.28905582e-01 -5.57030141e-01 -9.79872227e-01
3.17026824e-01 -7.68732488e-01 5.82292601e-02 -1.49465632e+00
4.52108055e-01 1.27926499e-01 -4.00206387e-01 8.51131082e-01
-1.01146288e-01 6.39399767e-01 2.97639221e-01 -2.44554076e-02
-7.50062048e-01 7.19665647e-01 1.25870419e+00 -1.64453715e-01
3.51994671e-02 -1.24484211e-01 -7.99811721e-01 7.93694437e-01
4.41263348e-01 -2.45780081e-01 -2.90681213e-01 -3.24800581e-01
2.55168211e-02 2.18576580e-01 3.53440404e-01 -1.01337445e+00
4.27846074e-01 -7.42726773e-02 4.79867131e-01 -6.62014544e-01
8.55321407e-01 -8.67649317e-01 -1.07527010e-01 1.60860028e-02
-5.72081245e-02 -3.62018079e-01 2.69351304e-01 4.17223871e-01
-5.99865079e-01 6.18312918e-02 7.39953041e-01 -6.07991405e-02
-1.08221674e+00 3.20353448e-01 -4.85650808e-01 -3.56456906e-01
1.14250863e+00 -2.28192076e-01 -2.79295743e-01 -4.48142141e-01
-7.28586376e-01 8.76724944e-02 6.65777996e-02 6.04096472e-01
8.43355715e-01 -1.51980579e+00 -7.74915159e-01 4.68664408e-01
1.53141081e-01 -4.85383898e-01 5.59071541e-01 1.15412033e+00
1.11356631e-01 7.91903585e-02 -4.16022986e-01 -6.69639349e-01
-1.61989105e+00 3.71164799e-01 2.36214012e-01 5.74022066e-03
1.57904308e-02 8.00334275e-01 1.34597763e-01 -5.19736595e-02
3.34516466e-01 -2.86091920e-02 -3.81480664e-01 6.53068006e-01
8.77188981e-01 4.20692474e-01 -5.14216758e-02 -1.11135697e+00
-2.56861180e-01 6.69131756e-01 5.16421944e-02 -4.10861403e-01
1.11174762e+00 -3.87309313e-01 -2.29155555e-01 2.89475411e-01
1.42575788e+00 -3.04053426e-01 -1.13364458e+00 -2.48554558e-01
-6.14565194e-01 -3.80125701e-01 6.12681992e-02 -5.99885583e-01
-1.30802965e+00 1.07440710e+00 9.24593329e-01 -1.72679901e-01
1.65821731e+00 1.43151879e-01 7.12900579e-01 2.09125921e-01
2.43591413e-01 -1.13836873e+00 2.18408927e-01 4.31201041e-01
7.71401405e-01 -1.54822016e+00 -2.75046706e-01 -2.95871884e-01
-6.80632234e-01 1.13298607e+00 6.88186944e-01 3.28465223e-01
8.17132294e-01 2.29210153e-01 1.33024007e-01 -1.36322394e-01
-7.86323607e-01 -7.23517537e-01 3.31879735e-01 3.78511488e-01
4.15763795e-01 -2.02925444e-01 -8.04692954e-02 5.07602572e-01
3.65158468e-01 2.73703277e-01 9.76500139e-02 1.10353553e+00
-3.73458475e-01 -1.22730780e+00 -5.38770139e-01 3.09592336e-01
-6.62128448e-01 3.48879322e-02 -2.70074844e-01 5.77764809e-01
6.01030648e-01 1.45006776e+00 -3.39502543e-01 -7.05802798e-01
1.80574745e-01 2.99697042e-01 6.88880980e-01 -4.63050425e-01
-3.46528828e-01 3.28204870e-01 4.81452793e-02 -9.64733362e-01
-1.08868265e+00 -6.88533247e-01 -1.03883970e+00 -2.84509301e-01
-3.66770953e-01 -3.16370308e-01 6.00667894e-01 1.19514871e+00
5.09652317e-01 4.34503317e-01 6.69133186e-01 -1.17343462e+00
-2.26258188e-01 -1.25694466e+00 -6.06144071e-01 4.18791354e-01
3.36063236e-01 -9.62249339e-01 -3.23628932e-01 1.50146335e-01]
|
[13.537264823913574, 1.7722654342651367]
|
54eaffe2-2b3f-468f-b7b3-dfb4daf86804
|
scene-text-detection-for-augmented-reality
|
2101.01054
| null |
https://arxiv.org/abs/2101.01054v1
|
https://arxiv.org/pdf/2101.01054v1.pdf
|
Scene Text Detection for Augmented Reality -- Character Bigram Approach to reduce False Positive Rate
|
Natural scene text detection is an important aspect of scene understanding and could be a useful tool in building engaging augmented reality applications. In this work, we address the problem of false positives in text spotting. We propose improving the performace of sliding window text spotters by looking for character pairs (bigrams) rather than single characters. An efficient convolutional neural network is designed and trained to detect bigrams. The proposed detector reduces false positive rate by 28.16% on the ICDAR 2015 dataset. We demonstrate that detecting bigrams is a computationally inexpensive way to improve sliding window text spotters.
|
['Bharadwaj Amrutur', 'Sagar Gubbi']
|
2020-12-26
| null | null | null | null |
['text-spotting', 'scene-text-detection']
|
['computer-vision', 'computer-vision']
|
[ 9.53398645e-01 -2.02253878e-01 -6.48738369e-02 -2.63160944e-01
-7.03786731e-01 -2.40616053e-01 7.76682019e-01 1.39547884e-01
-6.56509042e-01 4.17927444e-01 3.84684652e-01 -4.83164787e-01
5.48353612e-01 -8.97905409e-01 -9.31728840e-01 -2.20082089e-01
2.86540329e-01 1.14970520e-01 6.25147283e-01 -1.47390664e-01
5.25560915e-01 4.77178484e-01 -1.56979787e+00 1.00435817e+00
6.19587064e-01 6.88353360e-01 5.44627368e-01 1.13312685e+00
-3.04371357e-01 9.36519861e-01 -8.23278010e-01 -3.28399092e-01
1.63947254e-01 -4.79175240e-01 -4.49290037e-01 3.22616845e-02
9.87343192e-01 -6.76391482e-01 -3.61073524e-01 1.08995569e+00
3.20741862e-01 -9.81234610e-02 2.27188095e-01 -6.16412044e-01
-3.41271549e-01 8.76319289e-01 -7.52413392e-01 4.15308625e-01
5.20556331e-01 -3.01352352e-01 9.80938613e-01 -1.30175638e+00
4.76526886e-01 1.04138625e+00 6.17077708e-01 4.02501374e-01
-5.94921291e-01 -7.50988722e-01 -1.18066490e-01 1.50078293e-02
-1.44127262e+00 -4.72029388e-01 3.44179660e-01 -3.98476154e-01
1.21903014e+00 6.50680602e-01 4.15496379e-01 9.42009330e-01
1.84035227e-01 1.17150235e+00 5.77651143e-01 -8.99477065e-01
8.32850188e-02 -1.49743417e-02 4.68672961e-01 7.80375004e-01
5.29076695e-01 -9.49860215e-02 -7.69231796e-01 -3.90542895e-02
8.39134514e-01 2.78480828e-01 -6.07073680e-02 2.31695533e-01
-1.40806961e+00 8.67761612e-01 2.02606559e-01 2.03158826e-01
-4.33677919e-02 3.49170536e-01 4.83053446e-01 1.44596845e-01
4.14929152e-01 4.97741461e-01 -1.30392546e-02 -4.35960710e-01
-9.39846873e-01 -4.20407988e-02 3.45696032e-01 9.43704724e-01
4.30126786e-01 1.27263755e-01 -1.65056184e-01 1.05913019e+00
3.93204577e-03 5.05162299e-01 5.16746402e-01 -2.29897380e-01
6.94520891e-01 7.64710844e-01 1.03207007e-01 -9.32088435e-01
-4.20392841e-01 1.29096150e-01 -7.03081727e-01 4.06246223e-02
2.43069038e-01 -1.30919442e-01 -1.22519791e+00 6.64544702e-01
9.95997339e-02 2.07793452e-02 -5.77701747e-01 7.29570329e-01
6.04595542e-01 8.87595832e-01 -3.15346390e-01 1.55621663e-01
1.44245398e+00 -7.28751659e-01 -8.43183756e-01 -6.17656946e-01
1.15144265e+00 -1.26360214e+00 1.34813356e+00 5.57458401e-01
-7.57041514e-01 -4.47074443e-01 -1.24090528e+00 -2.35981733e-01
-4.53963131e-01 7.63510466e-01 5.02686858e-01 1.05860078e+00
-6.59392416e-01 2.18797207e-01 -8.29377055e-01 -4.76468414e-01
4.38933104e-01 3.42838526e-01 -8.49877149e-02 1.39090801e-02
-8.18998933e-01 5.76369047e-01 5.77257216e-01 9.06158835e-02
-1.23344988e-01 -4.05848920e-01 -8.20966542e-01 5.17538823e-02
4.68903542e-01 -1.01417534e-01 1.35898459e+00 -8.25563312e-01
-1.27155936e+00 9.45040882e-01 -3.88061345e-01 -7.79445112e-01
6.88671649e-01 -7.79216886e-01 -6.84727013e-01 6.82403445e-02
-3.68039571e-02 3.43719304e-01 1.07481146e+00 -7.18500733e-01
-1.00139594e+00 -1.96990997e-01 -3.06974262e-01 7.44255409e-02
-4.25988495e-01 3.80941004e-01 -3.77327472e-01 -8.90898228e-01
1.47798434e-01 -7.49639511e-01 -1.36221632e-01 1.67404458e-01
-6.45795584e-01 -7.68862218e-02 9.57160592e-01 -6.52467430e-01
1.35206938e+00 -2.09017754e+00 -8.16425979e-01 1.69920668e-01
2.71237314e-01 5.37477314e-01 1.02535062e-01 3.44915360e-01
1.55380704e-02 1.44195184e-01 -8.00999776e-02 -2.15567008e-01
-3.36281866e-01 -2.86920648e-02 -8.05357277e-01 4.73264664e-01
3.13785225e-01 8.22242379e-01 -6.59003139e-01 -3.82051975e-01
8.39028239e-01 2.11952731e-01 -3.74052465e-01 -1.57716915e-01
-1.49880946e-01 -4.51630801e-01 -9.20074526e-03 3.82665604e-01
7.99101651e-01 -2.36175075e-01 -2.36730222e-02 9.83192176e-02
-3.11556190e-01 6.53230369e-01 -1.21716475e+00 1.25061333e+00
-2.50804961e-01 1.41486239e+00 -7.03995764e-01 -4.33784515e-01
9.15633917e-01 -1.85419753e-01 -1.17157148e-02 -1.00423038e+00
2.47084454e-01 -1.70453135e-02 -3.38264823e-01 -2.84659535e-01
1.32699656e+00 -3.46443728e-02 -4.94233556e-02 4.35434371e-01
-5.06234407e-01 6.98573515e-02 3.08444090e-02 2.72737533e-01
1.22469306e+00 -4.46118265e-01 5.35627663e-01 1.53475508e-01
1.78149432e-01 1.71486184e-01 -8.05103183e-02 1.12536132e+00
2.35988759e-02 9.26685274e-01 1.89758897e-01 -9.19738650e-01
-1.28495550e+00 -7.65640974e-01 -1.43960029e-01 1.26727891e+00
1.04388572e-01 -8.13617587e-01 -4.10098761e-01 -4.90298450e-01
-2.13421270e-01 8.54470611e-01 -6.52729332e-01 -7.86332265e-02
-8.58382583e-01 -6.44732296e-01 7.65124261e-01 7.83537269e-01
4.60780293e-01 -9.22779262e-01 -9.54544425e-01 -5.10608815e-02
-1.28649145e-01 -1.42891634e+00 -5.17828822e-01 3.32096964e-01
-7.16695845e-01 -8.16749156e-01 -7.47175872e-01 -8.27756107e-01
6.21712267e-01 8.13160539e-01 8.31717312e-01 1.45844012e-01
-6.42194629e-01 -4.37147804e-02 -7.16124475e-01 -6.11510336e-01
-6.12652421e-01 -1.09497178e-02 -3.09290618e-01 -1.80711731e-01
6.63306236e-01 3.58483225e-01 -4.29256648e-01 3.86940867e-01
-1.06508434e+00 5.24172664e-01 5.65803945e-01 8.20135832e-01
4.44677144e-01 -1.15335129e-01 -1.25936806e-01 -1.21493781e+00
5.18052816e-01 7.74834380e-02 -7.69814074e-01 1.73236594e-01
-1.92407846e-01 -1.28201649e-01 5.80785751e-01 -5.37519336e-01
-8.74231100e-01 5.37871182e-01 -3.89077246e-01 -4.36248556e-02
-3.79153699e-01 1.38432547e-01 3.67000669e-01 -1.44985050e-01
8.57058346e-01 3.99122477e-01 -5.65861821e-01 -4.22143459e-01
2.48564050e-01 9.14301038e-01 3.97717834e-01 8.43558013e-02
5.21959782e-01 6.94129765e-01 -4.77162927e-01 -1.64199686e+00
-5.70798695e-01 -1.00866127e+00 -6.27778590e-01 -1.26913115e-01
7.65979230e-01 -8.79324675e-01 -4.92878735e-01 3.66795480e-01
-1.23011994e+00 -2.82543331e-01 -1.62516102e-01 2.65966594e-01
-1.22451097e-01 6.20333850e-01 -4.05052900e-01 -9.84494448e-01
-3.26853156e-01 -5.31763852e-01 1.71362913e+00 2.92515289e-02
-4.19835031e-01 -5.21908283e-01 4.72780094e-02 2.04805106e-01
-5.39267547e-02 -1.23367757e-01 4.89643186e-01 -6.87795043e-01
-3.06872368e-01 -6.87535405e-01 -6.19036794e-01 -1.05059303e-01
1.13292530e-01 8.49344954e-02 -1.19164848e+00 6.19431697e-02
-2.41216496e-01 -7.08120642e-03 1.11580002e+00 3.58669311e-01
1.24421084e+00 -2.92793036e-01 -4.75382000e-01 2.84662366e-01
1.26907134e+00 2.47517288e-01 1.09792721e+00 4.09957021e-01
9.62993860e-01 1.94865707e-02 7.22829401e-01 8.48909974e-01
-8.83833691e-02 5.25261104e-01 8.47906694e-02 -1.31483868e-01
-1.88972980e-01 -3.64655346e-01 2.46745378e-01 3.84744495e-01
3.77852142e-01 -4.16580409e-01 -1.27743232e+00 6.09171748e-01
-1.80933118e+00 -1.03403199e+00 -7.57495403e-01 2.10442924e+00
5.12473881e-01 5.15875340e-01 -4.69253659e-02 4.07235444e-01
1.09037280e+00 1.74381971e-01 -1.38497695e-01 -4.97291803e-01
-3.09721529e-01 4.05181229e-01 7.97716439e-01 3.99636179e-01
-1.37482750e+00 1.22687387e+00 6.68980837e+00 6.21177375e-01
-1.05246711e+00 -2.52970219e-01 2.90362984e-01 -6.57314658e-02
1.81354746e-01 -4.88083005e-01 -1.13891923e+00 3.77002299e-01
7.61075675e-01 1.85940415e-01 -5.77067165e-03 1.08836222e+00
2.38586918e-01 -5.21621585e-01 -8.51532876e-01 1.33539605e+00
4.01950985e-01 -1.78640103e+00 2.65561610e-01 -1.92093834e-01
6.36680722e-01 2.23893434e-01 -2.02036247e-01 -5.81819899e-02
3.29214901e-01 -8.73176455e-01 7.18273640e-01 -3.98691297e-02
7.98265100e-01 -5.93062341e-01 6.58131659e-01 2.66599327e-01
-1.32969713e+00 1.01635613e-01 -6.56228662e-01 -1.96699724e-01
9.74712819e-02 5.40810823e-01 -1.61959529e+00 -2.00249195e-01
5.83163083e-01 5.12527168e-01 -7.48950005e-01 1.33679819e+00
5.78425489e-02 6.85780585e-01 -5.57657659e-01 -5.01002014e-01
1.08681731e-01 3.67406666e-01 3.64087790e-01 1.63324261e+00
2.64940083e-01 -1.11953989e-01 -6.68508261e-02 5.85226417e-01
-2.38583803e-01 1.37511954e-01 -8.92441094e-01 -2.33952969e-01
1.86899826e-01 9.49127316e-01 -1.31772912e+00 -5.49719453e-01
-3.36046308e-01 1.52790892e+00 -7.03971982e-02 -9.55805704e-02
-8.39348912e-01 -7.70001531e-01 2.74477154e-01 2.02037930e-01
5.50749302e-01 -3.93482625e-01 -7.74227083e-01 -1.18237591e+00
1.56301081e-01 -8.15662920e-01 3.89002919e-01 -8.50484014e-01
-7.41196275e-01 3.36127341e-01 -5.43688416e-01 -1.39875233e+00
4.93141524e-02 -8.12531769e-01 -7.09112406e-01 4.66291100e-01
-1.26931715e+00 -1.02834392e+00 -6.30844057e-01 6.35963261e-01
1.22063160e+00 2.20635265e-01 7.41302371e-01 3.22762936e-01
-4.93314028e-01 8.03116381e-01 2.29986727e-01 5.22751093e-01
6.79715037e-01 -9.90628481e-01 1.34935486e+00 1.32734406e+00
6.54567063e-01 5.66811919e-01 8.10249746e-01 -1.05129385e+00
-1.40849948e+00 -1.13853419e+00 8.93640757e-01 -3.84216607e-01
6.85238063e-01 -7.67143011e-01 -8.23311090e-01 6.17557883e-01
-1.15068853e-01 -2.87108034e-01 6.40842497e-01 -3.18833366e-02
-3.28093231e-01 4.74023640e-01 -7.32195139e-01 8.65451097e-01
8.59753609e-01 -7.44142711e-01 -6.19146943e-01 4.80937898e-01
7.29384661e-01 -7.05993056e-01 1.30811837e-02 3.75386477e-02
7.55415738e-01 -8.88013601e-01 8.70804787e-01 -3.73055935e-01
3.78916711e-01 -9.14721861e-02 -1.13652781e-01 -5.77515662e-01
8.92768279e-02 -6.57942832e-01 -6.82116151e-02 6.23749733e-01
4.41831976e-01 -2.61739701e-01 1.12881279e+00 4.61183518e-01
-2.08572879e-01 6.43003061e-02 -7.63968110e-01 -6.36407137e-01
-3.98176402e-01 -9.11215663e-01 1.87297389e-01 7.10332990e-01
6.72748387e-02 3.37163776e-01 -6.34213448e-01 2.98031688e-01
2.50122935e-01 -1.13227502e-01 9.02031004e-01 -1.07572675e+00
1.90163590e-02 -3.27186704e-01 -4.85218138e-01 -1.52034950e+00
-4.91769135e-01 -3.15036833e-01 4.14936930e-01 -1.31922638e+00
-3.18245292e-02 1.30453883e-02 1.47969529e-01 3.28891426e-01
-2.06600741e-01 6.12197757e-01 1.99153841e-01 1.22502102e-02
-6.72895133e-01 2.19686985e-01 7.43735075e-01 -8.03519189e-02
-2.36013874e-01 1.57718331e-01 -2.58561969e-01 1.00735891e+00
8.63798618e-01 -4.37273592e-01 3.82480733e-02 -5.11179328e-01
6.73405945e-01 -3.92102450e-01 2.81386346e-01 -1.11930192e+00
3.83835882e-01 -2.42092963e-02 5.97380280e-01 -1.40065622e+00
3.84180397e-01 -8.07678521e-01 -5.30560255e-01 4.98844028e-01
-5.34181297e-01 -6.64761410e-06 5.00666976e-01 5.90672195e-01
1.81315064e-01 -2.81280756e-01 5.60520828e-01 8.73269737e-02
-1.01241016e+00 -2.89565265e-01 -7.34275639e-01 -2.29997158e-01
9.17739511e-01 -6.23373151e-01 -4.98155415e-01 -1.21658146e-01
-8.89886320e-02 -2.91696608e-01 2.83484310e-01 6.43565476e-01
1.21998656e+00 -9.99603808e-01 -3.10107410e-01 4.54217434e-01
3.85918230e-01 -1.85846537e-01 -1.65397540e-01 3.32181811e-01
-1.06560862e+00 5.07153869e-01 -6.47697076e-02 -6.03011966e-01
-1.83176100e+00 2.38576457e-01 1.08116213e-02 1.70440555e-01
-1.00563669e+00 1.11712348e+00 6.49664700e-02 -4.20984626e-02
5.35648525e-01 -8.14040303e-01 7.93557614e-03 -9.39938650e-02
1.23822296e+00 5.32447278e-01 3.08247566e-01 -3.13311815e-01
-2.67740071e-01 4.96676177e-01 -4.88385558e-01 7.44399033e-04
1.18634868e+00 -2.00439975e-01 1.62541404e-01 6.00069642e-01
1.01179516e+00 7.72408098e-02 -6.74532712e-01 -1.55072719e-01
3.83751214e-01 -7.57217586e-01 1.98474258e-01 -6.52928531e-01
-3.80604625e-01 1.05145705e+00 7.82427907e-01 4.16547328e-01
7.81439185e-01 1.23843774e-02 1.13806677e+00 9.11899865e-01
1.27241641e-01 -1.32062888e+00 1.51803479e-01 6.71936631e-01
7.19042957e-01 -1.63867211e+00 4.68487330e-02 -5.09792864e-01
-6.45412326e-01 1.47800124e+00 5.76717496e-01 -1.70326397e-01
3.43590885e-01 5.58364272e-01 -1.30225018e-01 -3.89423966e-01
-3.06616306e-01 -2.97238201e-01 4.43605870e-01 2.87317872e-01
5.19041777e-01 1.21703267e-01 8.71886164e-02 -1.04952827e-02
-3.61742228e-01 2.02253181e-02 7.70411730e-01 1.06174672e+00
-8.88963640e-01 -6.34054601e-01 -4.66700405e-01 8.29790711e-01
-4.35833246e-01 -5.67416489e-01 -8.55784297e-01 4.95941222e-01
-2.73867577e-01 7.05610037e-01 2.64458060e-01 -6.43590033e-01
8.05451497e-02 -7.95942033e-04 1.06258780e-01 -8.07553411e-01
-3.94655913e-01 1.49360120e-01 2.66912282e-01 -4.41366851e-01
-3.65562513e-02 -2.55572349e-01 -1.26456118e+00 -4.36714530e-01
-7.88153112e-01 -6.46779001e-01 7.59116292e-01 8.38552594e-01
2.00562328e-01 3.68895799e-01 4.85731244e-01 -5.12981236e-01
-7.88871199e-02 -7.78508902e-01 -2.89216816e-01 1.41826913e-01
5.31378448e-01 -8.85384604e-02 -5.96195161e-02 2.83164054e-01]
|
[11.948110580444336, 2.2691969871520996]
|
9f4ee903-fa53-4651-983c-e3a52893ab55
|
audio-inpainting-with-generative-adversarial
|
2003.07704
| null |
https://arxiv.org/abs/2003.07704v1
|
https://arxiv.org/pdf/2003.07704v1.pdf
|
Audio inpainting with generative adversarial network
|
We study the ability of Wasserstein Generative Adversarial Network (WGAN) to generate missing audio content which is, in context, (statistically similar) to the sound and the neighboring borders. We deal with the challenge of audio inpainting long range gaps (500 ms) using WGAN models. We improved the quality of the inpainting part using a new proposed WGAN architecture that uses a short-range and a long-range neighboring borders compared to the classical WGAN model. The performance was compared with two different audio instruments (piano and guitar) and on virtuoso pianists together with a string orchestra. The objective difference grading (ODG) was used to evaluate the performance of both architectures. The proposed model outperforms the classical WGAN model and improves the reconstruction of high-frequency content. Further, we got better results for instruments where the frequency spectrum is mainly in the lower range where small noises are less annoying for human ear and the inpainting part is more perceptible. Finally, we could show that better test results for audio dataset were reached where a particular instrument is accompanist by other instruments if we train the network only on this particular instrument neglecting the other instruments.
|
['A. Eltelt', 'P. P. Ebner']
|
2020-03-13
| null | null | null | null |
['audio-inpainting']
|
['audio']
|
[ 1.85335517e-01 2.81051755e-01 5.78287363e-01 2.32845053e-01
-1.21075857e+00 -7.00642884e-01 2.60435581e-01 -3.19316655e-01
-2.10251167e-01 1.05397952e+00 3.78334314e-01 2.59508222e-01
-1.79960072e-01 -6.98884547e-01 -9.87362087e-01 -8.97299647e-01
-2.91799903e-01 2.70128936e-01 1.64522171e-01 -4.59695131e-01
-2.62588561e-01 2.99994111e-01 -1.57794356e+00 6.01197124e-01
6.05153143e-01 9.12569880e-01 1.69782996e-01 1.15846658e+00
3.50971371e-01 8.53625178e-01 -1.34602380e+00 -2.43953243e-01
5.32192826e-01 -8.62713814e-01 -6.58077180e-01 -3.71037573e-01
6.01938784e-01 -1.39566466e-01 -2.05440804e-01 9.64495480e-01
9.03988361e-01 2.86597162e-01 3.66850704e-01 -1.04256058e+00
-3.87350798e-01 1.08557343e+00 -3.61922830e-02 -3.96552682e-02
4.20225441e-01 3.38265330e-01 9.48961437e-01 -2.88617343e-01
7.32523978e-01 1.34487867e+00 1.10489070e+00 3.64524394e-01
-1.25356948e+00 -6.73524737e-01 -2.75068253e-01 3.57533932e-01
-1.28764999e+00 -9.59857460e-03 1.11148822e+00 -7.94615000e-02
4.74141687e-01 5.62620759e-01 6.96943164e-01 1.52864969e+00
8.73026624e-02 4.54067290e-01 1.10665476e+00 -4.68540996e-01
9.18299109e-02 -7.48255551e-02 -4.35589910e-01 1.84852228e-01
-4.55324620e-01 5.04539907e-01 -6.00640059e-01 -4.91082296e-02
5.50063133e-01 -5.45740902e-01 -4.81810540e-01 3.54128331e-01
-1.05679893e+00 6.67759418e-01 2.48177752e-01 7.06282198e-01
-5.15546083e-01 3.88798475e-01 4.10112441e-01 7.35380828e-01
1.75291285e-01 8.74807537e-01 -1.76714182e-01 -2.86265433e-01
-1.41815138e+00 5.68971932e-01 8.44834566e-01 4.84826565e-01
2.80777633e-01 7.56086707e-01 -4.14976060e-01 8.08909297e-01
-2.56010532e-01 4.11618978e-01 6.64150059e-01 -1.15534413e+00
2.97286540e-01 -3.38539869e-01 2.50289410e-01 -8.44794035e-01
-4.00660455e-01 -7.77869821e-01 -7.66699553e-01 6.50863588e-01
7.86173880e-01 -5.99592626e-01 -7.11764455e-01 1.83627820e+00
-4.00646292e-02 4.20858979e-01 1.13846853e-01 1.02257156e+00
8.08829665e-01 9.00159538e-01 -4.11162168e-01 -4.61368710e-02
1.14871585e+00 -9.70161259e-01 -9.89343762e-01 1.66048348e-01
9.82731432e-02 -1.36171246e+00 1.34725785e+00 1.07265496e+00
-1.36404073e+00 -1.35769713e+00 -1.29433823e+00 -4.59375530e-02
6.02678694e-02 4.77270335e-02 5.42550422e-02 6.69010282e-01
-1.03193915e+00 1.31897748e+00 -3.70869279e-01 2.10895136e-01
-1.82991832e-01 -6.99807331e-02 -2.94344753e-01 4.06038970e-01
-1.62117577e+00 7.55377412e-01 3.52452010e-01 1.19663840e-02
-1.18989503e+00 -9.69936192e-01 -4.72978354e-01 9.06444266e-02
9.65247676e-02 -2.77664065e-01 1.18135798e+00 -1.20247400e+00
-1.90171349e+00 3.88719022e-01 5.51610708e-01 -9.22576964e-01
9.57909942e-01 -5.58842242e-01 -7.06464052e-01 3.10970873e-01
-2.25669980e-01 5.28809667e-01 1.21584833e+00 -1.18778658e+00
-2.43753195e-01 1.80333614e-01 -1.08765543e-01 -1.09386928e-01
2.13508144e-01 -3.06749225e-01 1.83505595e-01 -1.28668213e+00
-2.79840678e-01 -9.24689531e-01 1.64340198e-01 -4.22117501e-01
-6.04621410e-01 2.82092601e-01 7.85358489e-01 -1.16689324e+00
9.32323575e-01 -2.30303216e+00 2.03083783e-01 1.16183445e-01
-3.65822315e-01 3.70564550e-01 -3.33258867e-01 8.10100794e-01
-4.40505028e-01 -4.64453362e-02 -1.24372113e-02 -4.02850538e-01
1.15015559e-01 4.83879866e-03 -7.05190778e-01 2.17540398e-01
2.85823978e-02 4.46999639e-01 -5.89325130e-01 -1.68373346e-01
-5.03863627e-03 6.78034365e-01 -5.56546926e-01 4.21491802e-01
-3.06804687e-01 6.62799656e-01 2.09608287e-01 3.14310282e-01
7.13308692e-01 8.08599591e-01 -3.52516294e-01 -3.76681715e-01
-1.91925708e-02 1.01206720e-01 -1.44637346e+00 1.88654125e+00
-7.30090439e-01 8.30992639e-01 2.64103085e-01 -6.94710195e-01
1.28052151e+00 5.89126408e-01 1.88289672e-01 -5.51945746e-01
1.03478372e-01 1.22674860e-01 3.02424729e-01 -5.83728254e-01
5.13722956e-01 -4.78365898e-01 9.76297422e-04 3.18776816e-01
3.49688202e-01 -7.91046143e-01 5.91308251e-03 -2.12639973e-01
1.03883457e+00 3.07135791e-01 -1.32693753e-01 -1.02976397e-01
2.90452987e-01 -3.77953023e-01 3.53417188e-01 8.78104746e-01
-9.47866067e-02 1.17375445e+00 6.85438633e-01 -1.28819525e-01
-1.12227774e+00 -1.15595031e+00 3.75261784e-01 8.40303242e-01
-2.90596962e-01 -3.35734546e-01 -1.04541719e+00 -2.74679780e-01
-2.14612126e-01 1.06851256e+00 -6.41562998e-01 -4.23002064e-01
-6.92196012e-01 -1.10062636e-01 1.22764564e+00 2.73946702e-01
4.45119411e-01 -1.42894840e+00 -3.01004320e-01 3.82859707e-01
-4.13962692e-01 -9.20052767e-01 -4.46952134e-01 2.13467941e-01
-5.91424942e-01 -6.39411867e-01 -9.05299366e-01 -6.48310840e-01
-3.13745648e-01 -5.81329048e-01 1.15296185e+00 -3.29617083e-01
-1.46650925e-01 4.39927876e-02 -4.36116725e-01 -8.26395512e-01
-8.34114730e-01 -9.42552760e-02 -1.29462793e-01 -4.45563383e-02
-4.47077155e-01 -1.05972731e+00 -5.28978765e-01 1.52599826e-01
-1.03053415e+00 -2.58507043e-01 3.29929501e-01 8.92132103e-01
3.88208777e-01 3.57272893e-01 7.26891398e-01 -5.71887672e-01
8.03604007e-01 -1.75441071e-01 -3.38828355e-01 -3.68236244e-01
7.73266256e-02 -1.32355774e-02 1.13883293e+00 -8.78484666e-01
-8.84972453e-01 -2.52267540e-01 -8.19639742e-01 -5.64808488e-01
-2.02840030e-01 -3.14631425e-02 -2.40503341e-01 1.53362766e-01
8.95155847e-01 -7.69914910e-02 -7.63236731e-02 -6.74637735e-01
2.63899833e-01 4.13350731e-01 9.32735205e-01 -5.22881627e-01
8.77673388e-01 3.15850735e-01 -9.62533802e-03 -7.55923867e-01
-5.47132552e-01 1.52410105e-01 -1.08646207e-01 -2.95995682e-01
7.49904990e-01 -6.65494978e-01 -5.57023168e-01 5.14978051e-01
-1.26498640e+00 -4.74424750e-01 -9.67467666e-01 7.61013150e-01
-8.84965003e-01 2.29966968e-01 -7.96217024e-01 -7.75633097e-01
-4.17003095e-01 -1.08347809e+00 8.46552134e-01 3.88641022e-02
-5.10322273e-01 -7.50429332e-01 3.03110421e-01 1.31193802e-01
3.89329553e-01 7.07369685e-01 9.00594056e-01 -4.59857911e-01
-7.44273961e-02 -1.31499738e-01 5.13792992e-01 9.41315830e-01
-7.85206929e-02 1.81954980e-01 -1.38882315e+00 -1.41542852e-01
3.98341060e-01 -2.93976009e-01 7.98461795e-01 4.36392725e-01
9.46741879e-01 -2.83314168e-01 5.75817943e-01 5.21638215e-01
1.27738690e+00 2.87020922e-01 1.03454459e+00 1.96729928e-01
3.35540026e-01 5.89137077e-01 6.48867667e-01 3.89953345e-01
-5.90091825e-01 8.10392082e-01 5.99885583e-01 -3.08997899e-01
-6.75271034e-01 -4.99658167e-01 6.50377750e-01 8.39362085e-01
-3.44411910e-01 -2.93586940e-01 -2.75262445e-01 4.96482104e-01
-1.29835367e+00 -1.10596669e+00 -1.84442997e-01 2.15973997e+00
1.09187305e+00 3.40597093e-01 4.12430376e-01 8.35449338e-01
6.52607620e-01 1.33740306e-01 -1.37147769e-01 -8.57160568e-01
-4.02232856e-01 9.17053342e-01 2.23247960e-01 6.62446678e-01
-7.93369055e-01 5.50769448e-01 6.30822515e+00 1.23865628e+00
-1.26172912e+00 2.07322970e-01 1.93076819e-01 -3.55610460e-01
-1.32840708e-01 -2.47023329e-01 -3.54363412e-01 6.58079386e-01
1.13617754e+00 1.26904503e-01 5.44891953e-01 6.80176377e-01
3.13668162e-01 4.87124808e-02 -9.47212875e-01 7.94300854e-01
-9.85427126e-02 -1.15163708e+00 -8.25716928e-02 -3.92416030e-01
7.69632936e-01 -4.60643917e-01 3.28625381e-01 4.31695104e-01
4.81512472e-02 -1.21209419e+00 1.18454230e+00 6.85485244e-01
7.48466969e-01 -1.07009649e+00 8.88032794e-01 2.70828187e-01
-8.08464408e-01 5.77953234e-02 -2.80582875e-01 -4.18783911e-02
5.44665419e-02 5.70533752e-01 -8.14176381e-01 7.33218431e-01
6.96850479e-01 3.72901373e-02 -1.94992095e-01 1.07999027e+00
-4.04624492e-01 9.45814669e-01 -1.83884203e-01 4.45792824e-01
3.34875077e-01 -4.93148789e-02 9.89704192e-01 1.17857635e+00
6.70353830e-01 -4.60949957e-01 -3.28062475e-01 8.21629405e-01
-8.08200091e-02 3.23585095e-03 -4.75958169e-01 2.74676353e-01
1.54668957e-01 1.02132380e+00 -3.42436552e-01 -3.79093178e-02
1.38864905e-01 9.31086719e-01 -2.85026044e-01 3.51361722e-01
-1.16371655e+00 -7.53180504e-01 3.83332640e-01 1.65823743e-01
4.50963110e-01 1.56187907e-01 2.06937641e-03 -5.63881397e-01
-1.20901726e-01 -1.15916705e+00 2.13274315e-01 -1.21152198e+00
-9.37058628e-01 9.47219491e-01 -2.66963184e-01 -1.53908408e+00
-5.61338246e-01 -1.85064763e-01 -8.11155438e-01 9.83035505e-01
-1.12511706e+00 -8.86201382e-01 -1.47211030e-01 5.54511607e-01
5.23360610e-01 -4.82414775e-02 9.70627964e-01 4.51208562e-01
6.05759658e-02 6.69321299e-01 1.32231608e-01 -2.11441219e-01
1.01579130e+00 -1.31864929e+00 -8.48879106e-03 8.15546036e-01
4.15790021e-01 4.87749930e-03 1.35740530e+00 -3.26798528e-01
-8.09540153e-01 -1.12870347e+00 5.31111419e-01 7.92493597e-02
4.99630660e-01 -3.28262150e-01 -9.93867099e-01 3.34845871e-01
6.64320290e-01 -1.91783607e-01 3.83953363e-01 -2.18159094e-01
-2.58725315e-01 -3.78845900e-01 -1.37040722e+00 3.06215107e-01
6.08914793e-01 -5.70379674e-01 -6.85988486e-01 1.79263070e-01
8.62646222e-01 -4.23826188e-01 -8.81908596e-01 1.66178197e-01
5.33656359e-01 -1.31507146e+00 8.70097578e-01 -4.07907873e-01
6.14204824e-01 -3.50436479e-01 -3.69495153e-02 -1.81253803e+00
2.72355862e-02 -1.33535695e+00 1.27443463e-01 1.46213830e+00
1.16909847e-01 -1.64679736e-01 4.63717699e-01 -3.31295133e-01
-3.64180714e-01 -3.27597588e-01 -1.02086961e+00 -1.10500431e+00
2.83961147e-01 -5.18112600e-01 5.61607897e-01 5.63079476e-01
-5.01415193e-01 6.76016435e-02 -9.56261337e-01 1.79049700e-01
2.55249321e-01 1.17857620e-01 6.18168175e-01 -1.02018952e+00
-8.90362322e-01 -2.35735759e-01 -4.51585352e-01 -3.35572541e-01
-1.08498596e-01 -5.49775600e-01 5.60685247e-02 -1.02900684e+00
-5.46719670e-01 -1.44847538e-02 -4.29259717e-01 5.23375832e-02
3.93401012e-02 6.74186647e-01 6.54422700e-01 -2.89343864e-01
-6.41501038e-06 3.18848968e-01 1.49167049e+00 -2.32983708e-01
-2.06340775e-01 3.16885859e-01 -2.88631469e-01 7.99781501e-01
7.86143720e-01 -6.54262006e-01 -3.80937308e-01 -4.45203148e-02
1.37702227e-01 4.18654799e-01 4.19684708e-01 -1.68097150e+00
-2.68195987e-01 5.06509125e-01 3.42532307e-01 -5.55066884e-01
6.28618360e-01 -8.01057160e-01 7.68882453e-01 6.28663301e-01
-4.78770196e-01 -2.88659781e-01 4.13616985e-01 1.09609433e-01
-6.54006064e-01 -3.83420110e-01 1.11134899e+00 -4.57821414e-02
-1.73601300e-01 -3.18940848e-01 -3.74133527e-01 1.43839017e-01
6.78224266e-01 -2.12983671e-03 2.70726159e-02 -9.52248275e-01
-1.15188301e+00 -5.47674656e-01 9.36738476e-02 2.16965348e-01
3.34463775e-01 -1.35404909e+00 -9.61562097e-01 1.46882400e-01
-5.07275403e-01 -3.07378948e-01 6.21380448e-01 6.06636107e-01
-5.60578942e-01 3.95003110e-02 -6.05956793e-01 -2.53153175e-01
-1.35218418e+00 4.30198401e-01 4.27392811e-01 -4.71273035e-01
-5.75680375e-01 1.00486112e+00 -1.19007891e-02 -1.33463055e-01
5.41122377e-01 -6.77221715e-01 -1.40963346e-01 2.27147356e-01
4.34638381e-01 6.69936478e-01 3.10944706e-01 -2.41537750e-01
7.14374632e-02 4.38379794e-01 4.34494674e-01 -3.77421826e-01
1.31285310e+00 3.52897465e-01 2.69687861e-01 6.92324877e-01
1.12430286e+00 7.51676857e-01 -1.21423590e+00 2.37918988e-01
-5.55899978e-01 -1.70365348e-01 -1.02267571e-01 -1.00418746e+00
-1.13176501e+00 9.79048848e-01 7.89451897e-01 5.83741784e-01
1.36574531e+00 -4.10806924e-01 9.12122726e-01 -3.09270732e-02
1.23158820e-01 -1.05468476e+00 2.49332160e-01 3.18880379e-01
1.46615720e+00 -3.69405955e-01 -2.84484893e-01 -1.35564312e-01
-6.88354671e-01 1.29113448e+00 1.03565328e-01 -5.66286802e-01
3.43251735e-01 5.25279582e-01 3.22104484e-01 4.11044002e-01
-4.98169750e-01 -4.39615957e-02 3.18679541e-01 7.98622131e-01
3.03337753e-01 -4.79307435e-02 -2.28921011e-01 9.18533266e-01
-9.84921575e-01 -2.02778697e-01 6.89000547e-01 3.08471113e-01
-6.21614978e-02 -1.10695922e+00 -8.60997915e-01 -7.20219463e-02
-9.28371608e-01 -1.07311748e-01 -2.46852532e-01 1.08858073e+00
6.92716122e-01 1.10568583e+00 -1.03130445e-01 -4.98757958e-01
5.46679854e-01 2.12202251e-01 5.25906920e-01 -1.45305753e-01
-1.15099621e+00 4.61072773e-01 2.32470483e-01 -7.87272692e-01
-1.65338516e-01 -4.08464640e-01 -1.04182220e+00 -1.77377716e-01
-1.75881028e-01 3.39933813e-01 6.27094805e-01 4.61982042e-01
6.60909042e-02 1.11367142e+00 5.74172556e-01 -9.10721719e-01
-5.58290064e-01 -1.34178352e+00 -9.85369146e-01 6.71294570e-01
5.53085864e-01 -3.05771440e-01 -5.51864803e-01 1.32928401e-01]
|
[15.566678047180176, 5.954543590545654]
|
cc20ca01-bebd-4359-9c33-4a6a4ad289ba
|
deep-learning-ensembles-for-skin-lesion
|
1808.08480
| null |
http://arxiv.org/abs/1808.08480v1
|
http://arxiv.org/pdf/1808.08480v1.pdf
|
Deep-Learning Ensembles for Skin-Lesion Segmentation, Analysis, Classification: RECOD Titans at ISIC Challenge 2018
|
This extended abstract describes the participation of RECOD Titans in parts 1
to 3 of the ISIC Challenge 2018 "Skin Lesion Analysis Towards Melanoma
Detection" (MICCAI 2018). Although our team has a long experience with melanoma
classification and moderate experience with lesion segmentation, the ISIC
Challenge 2018 was the very first time we worked on lesion attribute detection.
For each task we submitted 3 different ensemble approaches, varying
combinations of models and datasets. Our best results on the official testing
set, regarding the official metric of each task, were: 0.728 (segmentation),
0.344 (attribute detection) and 0.803 (classification). Those submissions
reached, respectively, the 56th, 14th and 9th places.
|
['Michel Fornaciali', 'Vinícius Ribeiro', 'Fábio Perez', 'Alceu Bissoto', 'Sandra Avila', 'Eduardo Valle']
|
2018-08-25
| null | null | null | null |
['skin-lesion-segmentation']
|
['medical']
|
[ 5.44639289e-01 1.91829711e-01 -1.29452303e-01 1.01495601e-01
-1.10800314e+00 -6.65309429e-01 8.68857026e-01 4.30120409e-01
-6.78539217e-01 7.49264956e-01 -3.57301980e-02 -4.04934019e-01
-8.27730671e-02 -4.18256760e-01 -1.68453142e-01 -8.69935274e-01
1.03963591e-01 3.31540197e-01 2.18762070e-01 -1.37977794e-01
3.86926919e-01 2.59188414e-02 -1.15923560e+00 7.48258770e-01
1.21181607e+00 8.06851864e-01 -3.25105518e-01 1.30532300e+00
1.65499598e-01 3.92886758e-01 -4.77867573e-01 -7.21067607e-01
6.41589111e-04 -3.96373153e-01 -1.04961014e+00 2.25834781e-03
6.67995751e-01 3.49436253e-01 4.00290415e-02 9.99724329e-01
4.91399974e-01 -5.39330184e-01 1.07645607e+00 -9.72406507e-01
-1.54097587e-01 4.52717781e-01 -8.51843774e-01 1.63965479e-01
3.92397255e-01 2.58520514e-01 4.50862736e-01 -5.82024634e-01
9.62619305e-01 5.34758747e-01 1.05794299e+00 1.04262602e+00
-8.77934933e-01 -9.96419787e-02 2.54077584e-01 2.56249040e-01
-1.42626190e+00 -3.60647291e-01 -5.64572178e-02 -7.54473329e-01
7.89745569e-01 9.41935301e-01 6.57994628e-01 1.45689130e+00
1.24858066e-01 8.30776930e-01 1.63752329e+00 -5.73428869e-01
9.72910151e-02 5.49669623e-01 4.11740601e-01 6.91970408e-01
2.09480003e-01 -1.08228810e-01 -3.14408690e-01 -1.56701371e-01
1.80506676e-01 -7.02148616e-01 -1.69141114e-01 4.05311227e-01
-1.15461671e+00 5.44669211e-01 1.06094912e-01 1.03386201e-01
-5.03025234e-01 8.84885807e-03 5.70142746e-01 3.76238763e-01
6.92331612e-01 2.63290793e-01 -1.73550174e-01 -1.55905008e-01
-8.37641716e-01 4.19323742e-02 8.55025828e-01 5.52882373e-01
-1.53374393e-02 -4.60825533e-01 -3.93904477e-01 9.32828009e-01
2.98199877e-02 3.31623793e-01 4.14350003e-01 -6.43344700e-01
2.24671215e-01 6.11755848e-01 -6.94052577e-02 -2.36414582e-01
-7.12299049e-01 -6.64139748e-01 -1.01009524e+00 2.62504965e-01
7.01049387e-01 -5.82516372e-01 -1.59161162e+00 1.09999311e+00
3.57484818e-02 2.61911377e-02 1.63178355e-01 6.38063550e-01
9.58684623e-01 3.29888463e-02 6.60680175e-01 8.05103555e-02
1.49726903e+00 -7.44260967e-01 -3.08867007e-01 7.33518898e-02
7.80833244e-01 -1.07341480e+00 4.15658385e-01 8.17796409e-01
-8.55494678e-01 -8.54396448e-02 -9.00373936e-01 4.94413584e-01
-7.12321520e-01 3.60178530e-01 6.78478777e-01 9.84045923e-01
-1.19456065e+00 1.62983909e-01 -7.01831996e-01 -1.02547073e+00
4.21663880e-01 1.53086379e-01 -3.46626163e-01 8.75385553e-02
-9.31238055e-01 1.17951739e+00 7.28376880e-02 -1.38267919e-01
-7.93983579e-01 -8.79948139e-01 -1.49474725e-01 -9.10569251e-01
1.83297396e-01 -7.79810488e-01 7.67450929e-01 -1.08094120e+00
-1.06902516e+00 1.62338316e+00 -8.14087167e-02 -8.83951485e-01
1.13194025e+00 -6.08461834e-02 -7.35021949e-01 4.67558019e-02
-3.14353444e-02 8.89180660e-01 5.37118316e-01 -1.09808695e+00
-1.23775280e+00 -4.35657799e-01 -3.69113870e-02 2.16093108e-01
-8.04900005e-02 1.51703879e-01 -4.14896220e-01 -5.80863178e-01
-2.73216456e-01 -9.90864933e-01 -4.08495605e-01 -1.83890790e-01
-7.78112888e-01 -3.57948661e-01 3.28602314e-01 -9.58035231e-01
1.22393596e+00 -1.92976665e+00 2.73351967e-01 3.89361769e-01
2.16635525e-01 4.31616485e-01 -7.25035667e-02 5.54627180e-01
-1.50677010e-01 5.71173847e-01 -4.03532475e-01 -5.87218344e-01
-1.49999037e-01 -1.91823050e-01 2.50056535e-01 4.57894683e-01
1.42666355e-01 6.31916463e-01 -5.21979570e-01 -6.85437024e-01
1.60642743e-01 4.24151927e-01 1.48287460e-01 -5.09490609e-01
2.04229102e-01 2.53050894e-01 -8.84263888e-02 1.09168077e+00
6.96354091e-01 5.75907975e-02 -6.47706091e-02 2.14168895e-02
-1.97638869e-01 -5.61434090e-01 -9.34131622e-01 1.55637896e+00
-3.23900074e-01 8.81809652e-01 1.88097969e-01 -4.43692595e-01
5.44850767e-01 4.83929068e-01 5.88497698e-01 -3.71812582e-01
-2.28064582e-01 4.52068537e-01 2.25268066e-01 -5.54808438e-01
2.94671923e-01 3.21308643e-01 4.57646959e-02 5.18412292e-02
-9.47290212e-02 5.24653196e-02 5.56562126e-01 3.24951231e-01
1.05444694e+00 -9.40085724e-02 4.48499590e-01 -5.16912043e-01
9.04419720e-01 2.97708452e-01 1.12311840e-01 7.19667554e-01
-6.96019411e-01 9.03547764e-01 7.80920625e-01 -5.44968784e-01
-6.93363309e-01 -1.12893152e+00 -6.41802847e-01 9.65526283e-01
-1.54102609e-01 -4.90251780e-01 -1.34416151e+00 -1.13361156e+00
-6.45165220e-02 4.66004610e-01 -1.34108067e+00 3.10336739e-01
-1.30909637e-01 -1.14352703e+00 1.07324636e+00 2.48920918e-01
4.79760170e-01 -8.46224546e-01 -1.63555250e-01 -1.75124645e-01
-1.48057729e-01 -8.76398563e-01 -1.67665437e-01 2.85769571e-02
-3.40733737e-01 -1.59765685e+00 -1.08232665e+00 -6.01051688e-01
7.25081801e-01 -2.72837490e-01 8.45264137e-01 2.54441828e-01
-9.40246880e-01 4.66210753e-01 -4.24585730e-01 -8.83821607e-01
-4.85024303e-01 4.80487227e-01 -2.97671318e-01 2.60185689e-01
4.52871829e-01 1.00983202e-01 -5.93856514e-01 1.86009392e-01
-7.52705753e-01 1.55771211e-01 8.21410000e-01 6.16127729e-01
5.32814562e-01 -5.42634189e-01 5.22806525e-01 -1.38568401e+00
6.07948482e-01 -2.49847412e-01 -2.49657989e-01 5.96762776e-01
-5.33528864e-01 -6.69691205e-01 1.77014545e-01 6.56334534e-02
-9.72562492e-01 3.13524157e-01 -2.61646807e-01 3.27683866e-01
-7.30802715e-01 2.48603061e-01 5.13294041e-01 -3.85574996e-01
1.05305231e+00 2.14215055e-01 1.48110613e-01 -2.01628834e-01
-1.01315394e-01 7.85204649e-01 4.70018744e-01 -1.81863025e-01
5.44548154e-01 4.65628177e-01 1.89530373e-01 -9.73310649e-01
-8.02562654e-01 -6.93434298e-01 -8.80134463e-01 -3.88362974e-01
1.20243645e+00 -8.49544346e-01 -2.13348135e-01 1.20924580e+00
-9.18882072e-01 -4.44597542e-01 -1.69003367e-01 1.62044898e-01
-1.96294844e-01 4.35976923e-01 -4.88779277e-01 -7.87025928e-01
-6.58930123e-01 -9.44603086e-01 1.13982201e+00 4.03794676e-01
-5.23127019e-01 -1.48914421e+00 2.73375332e-01 5.41972339e-01
5.57535887e-01 7.63403237e-01 3.92737478e-01 -4.17402297e-01
1.76478580e-01 -6.51011109e-01 -2.51522928e-01 2.76272774e-01
-2.72275265e-02 5.73536336e-01 -1.27054012e+00 -2.36738607e-01
-8.41181576e-01 -2.77116269e-01 1.57361555e+00 4.63438928e-01
1.37138867e+00 3.09252173e-01 -7.02561319e-01 6.80509925e-01
1.55770278e+00 -5.36270626e-02 9.65656579e-01 4.02954847e-01
3.21964502e-01 7.12900221e-01 5.57225347e-01 1.31637573e-01
1.91622883e-01 3.89711887e-01 4.43785906e-01 -3.96446884e-01
-6.95829868e-01 7.02064410e-02 7.60823786e-02 2.37795174e-01
-8.89298677e-01 -1.77051470e-01 -1.26063633e+00 6.13147259e-01
-1.53359556e+00 -4.78832453e-01 -8.99119318e-01 2.19903445e+00
5.77889025e-01 8.65806043e-02 2.72411913e-01 -1.38754705e-02
6.74820185e-01 -3.98456939e-02 -2.74209648e-01 -5.46604216e-01
-3.93835813e-01 2.68270493e-01 6.07095778e-01 5.66895723e-01
-1.81325209e+00 8.44273210e-01 7.08856440e+00 1.07967734e+00
-8.79265130e-01 -2.58019753e-02 9.23499644e-01 1.10357046e-01
-1.37707742e-03 -2.37779617e-01 -8.68199825e-01 5.45162916e-01
9.45085108e-01 1.28619090e-01 -1.57905594e-02 1.85243562e-01
-3.87851328e-01 -5.99543095e-01 -6.77449584e-01 7.38949597e-01
2.82512426e-01 -1.25286102e+00 -1.04729988e-01 4.71406937e-01
9.87859547e-01 1.97655633e-01 3.63281637e-01 8.10065493e-02
9.37115401e-02 -1.51727593e+00 2.08703697e-01 1.02864873e+00
1.20524013e+00 -5.96171141e-01 1.20804250e+00 5.34718595e-02
-8.32290113e-01 1.17697693e-01 -9.69646126e-02 3.16857576e-01
-3.37352037e-01 4.46662188e-01 -7.23615706e-01 8.39753449e-01
4.58266020e-01 7.11976588e-01 -1.38611591e+00 1.39137590e+00
-8.58947262e-02 8.29346120e-01 -2.04450712e-01 -3.14227432e-01
3.92082870e-01 1.74906760e-01 6.27529860e-01 1.69374549e+00
-1.38799436e-02 -2.58116633e-01 -1.30686238e-01 5.87705560e-02
2.51864523e-01 1.11019045e-01 -2.56069571e-01 3.07461947e-01
8.57746005e-02 1.80070794e+00 -7.21830547e-01 -1.83982879e-01
-1.04520418e-01 1.05002367e+00 -7.46571794e-02 3.47692251e-01
-8.21450233e-01 -5.56571126e-01 3.83825243e-01 -2.23596655e-02
-3.56637657e-01 4.18720216e-01 -7.22781956e-01 -8.11117351e-01
-9.65084806e-02 -8.53119016e-01 9.90915716e-01 -2.60803223e-01
-1.38857269e+00 6.85568988e-01 -3.68401289e-01 -9.55321491e-01
8.82065576e-03 -1.11095619e+00 -9.04767811e-01 1.07176888e+00
-1.33007586e+00 -1.53970814e+00 -6.17553830e-01 4.21398073e-01
4.31558847e-01 -6.69622838e-01 1.31202853e+00 -8.02475810e-02
-7.74904072e-01 1.15897548e+00 -3.41485925e-02 4.84921262e-02
1.03499758e+00 -1.69568861e+00 2.92771846e-01 6.03788793e-01
-9.06159207e-02 4.11810540e-03 4.11146879e-01 -3.75790358e-01
-9.62304354e-01 -1.22322381e+00 6.99688375e-01 -9.29318368e-01
7.44215190e-01 -1.57082044e-02 -4.47977066e-01 3.20397586e-01
5.82488358e-01 -2.91420102e-01 9.07337964e-01 3.36560637e-01
-9.60459113e-02 -4.62128926e-04 -1.25121284e+00 4.37917203e-01
7.50812054e-01 -3.63143712e-01 1.63815454e-01 8.40285480e-01
-3.99135537e-02 -4.79618013e-01 -1.11911619e+00 4.36724722e-01
7.12400258e-01 -6.73884034e-01 6.88656747e-01 -6.52915835e-01
4.75207835e-01 -3.06584872e-02 4.29636911e-02 -1.30864370e+00
-1.01683222e-01 -5.51606715e-01 1.56422317e-01 1.01908576e+00
1.19019663e+00 -5.88731945e-01 1.15360117e+00 1.06300190e-01
-2.64398903e-01 -1.06638312e+00 -9.72392857e-01 -4.15849864e-01
5.86617291e-01 -3.47650647e-01 -1.18275946e-02 7.97908545e-01
-2.50293493e-01 -1.33317977e-01 -8.44663903e-02 -9.28615779e-02
9.27424371e-01 -4.20700222e-01 6.75841868e-01 -1.24203050e+00
2.80527622e-02 -8.68399680e-01 -5.65545440e-01 -3.73120159e-02
-3.95385414e-01 -9.79951859e-01 -3.45495462e-01 -1.72180176e+00
5.32543659e-01 -2.42361918e-01 -4.52977180e-01 5.24768889e-01
-4.93172407e-01 7.00327873e-01 9.52503234e-02 1.28459021e-01
-6.30595028e-01 -6.78006291e-01 1.05810201e+00 -3.83189589e-01
1.41546771e-01 4.00728673e-01 -8.59976709e-01 7.72389948e-01
9.95471478e-01 1.52207270e-01 1.61278293e-01 -3.46342504e-01
-1.07224789e-02 -3.65732908e-01 6.47417784e-01 -1.16362679e+00
5.34147441e-01 -3.39300632e-02 7.01058447e-01 -5.48052430e-01
2.88945466e-01 -2.21160784e-01 3.32172997e-02 8.37206125e-01
-4.18039829e-01 -5.15888631e-01 3.01229864e-01 3.89964402e-01
-9.31177661e-02 -4.06040490e-01 8.99278224e-01 -1.47025391e-01
-1.03929806e+00 2.46422574e-01 -5.34705639e-01 -5.98535277e-02
1.54712486e+00 -3.73141319e-01 -5.73819399e-01 9.60517824e-02
-1.43045652e+00 4.60987985e-01 3.86832833e-01 1.45775720e-01
4.63097334e-01 -7.21430004e-01 -1.56430280e+00 -1.62886575e-01
5.27843833e-01 -4.13840383e-01 4.20232028e-01 1.43341148e+00
-6.56815171e-01 5.13265610e-01 -1.82449818e-01 -7.40350664e-01
-1.62455523e+00 -1.38741270e-01 6.91597760e-01 -6.39561057e-01
-2.35885367e-01 1.21032333e+00 -2.75541872e-01 -5.07042170e-01
3.36740077e-01 4.32000220e-01 -4.25647438e-01 3.24059218e-01
5.66683292e-01 6.99879885e-01 2.33902812e-01 -2.50961542e-01
-5.18170655e-01 7.98461437e-01 -5.33439755e-01 -7.00341240e-02
7.57106423e-01 2.94597268e-01 -2.16575444e-01 1.66526690e-01
7.77947068e-01 -2.12911651e-01 -9.30517197e-01 4.21981156e-01
2.23329991e-01 -7.21341819e-02 -1.58971891e-01 -1.68323934e+00
-7.51502514e-01 7.89917767e-01 1.09917736e+00 2.93909222e-01
1.31783390e+00 -2.68793821e-01 3.33989382e-01 3.71164195e-02
1.19133852e-01 -1.40021741e+00 -5.61137438e-01 4.25546736e-01
8.46812606e-01 -1.25453401e+00 2.40926698e-01 -7.57102728e-01
-9.62793648e-01 1.14016676e+00 6.74290419e-01 -9.53904241e-02
5.17538071e-01 2.79723763e-01 3.38039398e-01 -9.87453535e-02
-9.42084372e-01 -4.14531142e-01 5.56297481e-01 1.08363891e+00
6.10314906e-01 6.19079113e-01 -8.05104613e-01 2.02819273e-01
-4.93798256e-02 -2.96223700e-01 5.91460407e-01 3.23164672e-01
-2.39451334e-01 -1.18942165e+00 -2.26901859e-01 8.55199099e-01
-8.14763129e-01 -7.11337477e-02 -1.36580145e+00 1.29240298e+00
4.22517478e-01 8.34386647e-01 6.08660057e-02 -4.67474580e-01
2.89267629e-01 5.29631190e-02 4.65953439e-01 -2.82120764e-01
-9.54989195e-01 -1.59730747e-01 6.42666698e-01 -3.42214823e-01
-4.01338428e-01 -8.56386304e-01 -7.62164652e-01 -3.01777840e-01
2.62059551e-02 1.85357466e-01 8.52277279e-01 5.24243355e-01
4.09081466e-02 5.29371321e-01 3.87399763e-01 7.78591409e-02
-4.84402418e-01 -1.23489535e+00 -7.22554147e-01 2.38700435e-01
-5.57099767e-02 -1.35764882e-01 -1.64667070e-01 1.24885745e-01]
|
[15.721688270568848, -3.000516891479492]
|
fe0f41a1-515a-4cbc-b688-fc71945cbac9
|
bennettnlp-at-semeval-2021-task-5-toxic-spans
| null | null |
https://aclanthology.org/2021.semeval-1.128
|
https://aclanthology.org/2021.semeval-1.128.pdf
|
BennettNLP at SemEval-2021 Task 5: Toxic Spans Detection using Stacked Embedding Powered Toxic Entity Recognizer
|
With the rapid growth in technology, social media activity has seen a boom across all age groups. It is humanly impossible to check all the tweets, comments and status manually whether they follow proper community guidelines. A lot of toxicity is regularly posted on these social media platforms. This research aims to find toxic words in a sentence so that a healthy social community is built across the globe and the users receive censored content with specific warnings and facts. To solve this challenging problem, authors have combined concepts of Linked List for pre-processing and then used the idea of stacked embeddings like BERT Embeddings, Flair Embeddings and Word2Vec on the flairNLP framework to get the desired results. F1 metric was used to evaluate the model. The authors were able to produce a 0.74 F1 score on their test set.
|
['Vipul Mishra', 'Ambuje Gupta', 'Harsh Kataria']
|
2021-08-01
| null | null | null |
semeval-2021
|
['toxic-spans-detection']
|
['natural-language-processing']
|
[-4.32187766e-01 1.13356583e-01 1.37761841e-02 -3.04441065e-01
-3.42960387e-01 -2.96568573e-01 6.18410647e-01 8.53882492e-01
-4.66516763e-01 6.65167391e-01 5.79813421e-01 -1.02509923e-01
-2.43902504e-01 -8.43615294e-01 -6.69987947e-02 -2.17143759e-01
-1.12293065e-01 1.45126000e-01 1.52119786e-01 -3.28065395e-01
6.42023206e-01 2.08111286e-01 -1.47495306e+00 2.06617769e-02
1.03318799e+00 5.12562811e-01 -2.01411709e-01 6.17877603e-01
-3.90845299e-01 6.87438905e-01 -4.38712358e-01 -1.03622007e+00
-1.63115989e-02 -2.69488841e-02 -5.72159827e-01 -2.54110992e-01
3.62946689e-01 4.39239517e-02 -3.06926787e-01 1.42887974e+00
6.81851149e-01 8.22429657e-02 4.68683541e-01 -1.13733363e+00
-1.11502600e+00 8.09578061e-01 -7.92196512e-01 3.65827560e-01
5.55960417e-01 -1.57464981e-01 1.08354807e+00 -8.35582852e-01
4.03445899e-01 1.25436687e+00 6.54513001e-01 1.22370139e-01
-7.06026137e-01 -8.20651174e-01 -2.95756832e-02 3.77574116e-01
-1.34407055e+00 2.46612187e-02 3.75092894e-01 -6.93316102e-01
8.57581317e-01 3.37525755e-01 6.98066831e-01 1.28707230e+00
5.94786704e-01 2.32271120e-01 1.04998386e+00 -2.53082335e-01
-1.12673238e-01 4.70926821e-01 7.52603412e-01 7.51004994e-01
6.46539569e-01 -3.88619244e-01 -6.32678986e-01 -2.58691430e-01
-2.71480858e-01 3.90064180e-01 1.15525015e-01 3.95377874e-01
-8.67290616e-01 1.01696515e+00 3.91609013e-01 7.41118550e-01
-1.31936118e-01 -1.94226086e-01 5.20787060e-01 2.27993444e-01
1.05998719e+00 3.61000210e-01 -1.65352970e-01 -5.56056909e-02
-9.27460432e-01 2.45968953e-01 7.64610708e-01 4.05877650e-01
4.52535331e-01 -3.88839990e-01 -1.76598459e-01 8.46803904e-01
6.53261960e-01 5.77669561e-01 6.90478623e-01 -3.53844225e-01
4.54292566e-01 1.11347950e+00 -1.51682764e-01 -1.76291025e+00
-4.22789276e-01 -6.56262279e-01 -7.61999190e-01 -1.91243306e-01
2.15006366e-01 -5.21890461e-01 -6.80835128e-01 1.31453300e+00
4.13710952e-01 1.43216178e-01 -3.31846714e-01 6.82363689e-01
7.50558317e-01 9.65881705e-01 2.44345784e-01 -1.80576578e-01
1.43052840e+00 -6.12933755e-01 -1.07874966e+00 -5.55202775e-02
5.13263702e-01 -8.43034565e-01 6.36655092e-01 5.64373553e-01
-7.76638567e-01 -2.96806276e-01 -1.11293340e+00 -2.02600986e-01
-8.62023711e-01 -5.44365406e-01 6.74600542e-01 1.02460134e+00
-6.98990345e-01 8.01542580e-01 -4.67489839e-01 -7.97705829e-01
4.69589323e-01 1.28189355e-01 -6.09988034e-01 -2.07884789e-01
-1.48532903e+00 8.86937618e-01 2.95698315e-01 -7.87990466e-02
-3.85723323e-01 -7.42559969e-01 -6.11135662e-01 -7.74740661e-03
-1.74802199e-01 -2.59833127e-01 6.15372658e-01 -7.39512861e-01
-7.37880230e-01 9.50669587e-01 1.19365469e-01 -2.74380118e-01
3.87006998e-01 -4.98201638e-01 -8.78064930e-01 -2.57663071e-01
1.65478349e-01 5.77269681e-02 4.35939759e-01 -5.81449270e-01
-7.16465950e-01 -7.23906100e-01 -3.88370007e-01 -8.63037407e-02
-9.78736460e-01 5.55021584e-01 -1.26757562e-01 -5.42983949e-01
-1.02819681e-01 -7.17041790e-01 -4.08897921e-02 -3.29647720e-01
-5.77455163e-01 -5.43948829e-01 8.21154535e-01 -9.71219242e-01
1.64713645e+00 -2.00674844e+00 -2.11352631e-01 1.47292793e-01
4.32598323e-01 5.56856751e-01 2.30889678e-01 8.05805981e-01
-4.51448001e-03 7.75813460e-01 1.07354023e-01 -2.64510423e-01
3.09632607e-02 -4.05063704e-02 -8.61253217e-02 8.06452453e-01
-1.83624804e-01 3.97887379e-01 -1.08160603e+00 -4.55067188e-01
-1.73393965e-01 6.91393614e-01 -4.08420682e-01 7.41684958e-02
1.12228453e-01 -1.02971993e-01 -5.89140832e-01 3.89471114e-01
7.66071320e-01 -1.59789380e-02 -6.68274164e-02 1.28643304e-01
-4.03980851e-01 5.29692322e-02 -8.39552939e-01 1.18525577e+00
-6.11456186e-02 8.02554905e-01 -1.72312364e-01 -5.71545243e-01
1.02513814e+00 2.24592939e-01 3.13334346e-01 -4.46934640e-01
5.98519742e-01 -5.67730181e-02 4.69615757e-02 -9.54719424e-01
4.34243590e-01 -1.78586006e-01 -1.51922107e-01 3.80159557e-01
-1.89358592e-01 4.99385238e-01 2.09501684e-01 5.31110585e-01
1.27892637e+00 -3.84044051e-01 1.55383110e-01 -3.23744059e-01
5.44414103e-01 -1.44796491e-01 3.76853138e-01 3.61085802e-01
-2.84626007e-01 5.74409366e-01 4.77140605e-01 -2.81402558e-01
-8.54652643e-01 -1.01118386e+00 -1.91386387e-01 1.19151700e+00
-2.84759372e-01 -6.28961265e-01 -6.21319115e-01 -6.36376441e-01
1.68392986e-01 7.19754577e-01 -7.05234051e-01 -3.20183747e-02
2.16746470e-04 -7.66348958e-01 3.83121282e-01 -2.96143472e-01
3.81452978e-01 -8.90257180e-01 9.64102000e-02 1.36898622e-01
7.65556395e-02 -5.50745964e-01 -3.81526053e-01 -2.82714248e-01
-4.49205130e-01 -1.10197413e+00 -5.59574187e-01 -6.63900971e-01
5.77437758e-01 1.69844508e-01 6.23225808e-01 1.77385405e-01
-3.61722291e-01 -2.07964182e-01 -7.43244708e-01 -6.95244908e-01
7.53158797e-03 6.21012412e-02 2.35682920e-01 2.94235289e-01
7.33198166e-01 -6.39703691e-01 -4.12245989e-01 -2.09735394e-01
-7.36600161e-01 -4.61014330e-01 1.47989005e-01 2.22165287e-01
-1.47551805e-01 1.61146134e-01 6.87903881e-01 -1.08571470e+00
1.04564953e+00 -1.16119909e+00 -2.03485563e-01 -2.87556788e-03
-6.99354529e-01 -2.42250159e-01 4.03231829e-01 -2.04922289e-01
-6.69557452e-01 -4.38166052e-01 -3.30682665e-01 1.23508252e-01
1.41531929e-01 7.27191985e-01 4.28402163e-02 3.94352019e-01
5.75796247e-01 -3.01872045e-01 -1.67627856e-01 -5.81806958e-01
3.88609499e-01 1.32105029e+00 -2.32509300e-02 1.67097062e-01
9.38085496e-01 1.86472878e-01 -5.54513991e-01 -1.06561351e+00
-1.28567648e+00 -7.72536933e-01 -3.42160165e-01 -4.72020984e-01
1.16487026e+00 -6.87964976e-01 -7.91801155e-01 4.56026554e-01
-1.24797690e+00 6.78632438e-01 2.82854259e-01 5.12280643e-01
5.09659171e-01 4.87724543e-01 -4.89255965e-01 -1.02745569e+00
-5.09470403e-01 -4.24639463e-01 1.46613598e-01 3.91386896e-01
-4.63640749e-01 -1.04048431e+00 4.03008461e-01 6.42924249e-01
4.77775246e-01 4.41771120e-01 5.72215140e-01 -1.24490559e+00
-7.03347772e-02 -5.81419170e-01 -3.44403088e-01 4.70283359e-01
1.17488287e-01 3.01543027e-01 -9.24059331e-01 -1.29116043e-01
-1.16071351e-01 -1.16646290e-01 6.53341413e-01 -1.25330761e-01
8.96828651e-01 -6.32231474e-01 -2.78540581e-01 2.06944812e-02
1.42075467e+00 -1.43176436e-01 6.97958887e-01 2.03608632e-01
6.18544281e-01 7.05576360e-01 2.49146923e-01 6.33246183e-01
3.72979343e-01 8.03431645e-02 3.76064539e-01 4.00048435e-01
2.72967190e-01 -5.69997489e-01 5.87889731e-01 1.18503726e+00
3.20862025e-01 -5.27123272e-01 -1.03294349e+00 8.09014320e-01
-1.66826797e+00 -1.20728171e+00 -7.21108258e-01 1.99536562e+00
7.00220227e-01 3.19510221e-01 -1.53660789e-01 4.09170568e-01
8.84398341e-01 3.16929102e-01 3.94017585e-02 -8.82846713e-01
1.11303769e-01 6.39878288e-02 6.67010546e-01 5.47373056e-01
-9.86750662e-01 6.79826021e-01 5.85227680e+00 6.53231680e-01
-9.88538921e-01 4.26113158e-01 5.73452115e-01 -3.64958681e-02
-4.07425195e-01 -1.27348289e-01 -9.18578923e-01 7.58518279e-01
1.40621150e+00 -5.54669499e-01 -1.06883034e-01 4.82959270e-01
5.83204210e-01 -1.08455205e-02 -4.14331555e-01 8.47120106e-01
4.37360466e-01 -1.12596560e+00 -2.12155387e-01 2.03370050e-01
7.54002452e-01 3.20748866e-01 -4.82273959e-02 4.42187607e-01
4.12660688e-01 -1.13991606e+00 3.02189797e-01 7.38905787e-01
1.90917000e-01 -8.66371989e-01 7.56409645e-01 5.27936935e-01
-5.15049338e-01 -1.55999377e-01 -4.11634952e-01 -5.42626321e-01
1.88617840e-01 1.11870766e+00 -7.22693622e-01 4.07460243e-01
7.41257012e-01 8.32589865e-01 -7.64399707e-01 1.41234899e+00
-1.90761596e-01 9.88267541e-01 -2.31720835e-01 -6.53405428e-01
2.44824067e-01 -2.12968186e-01 4.29809660e-01 1.29894304e+00
4.63776052e-01 -1.75220013e-01 -8.05490389e-02 4.02726084e-01
-2.78055638e-01 7.20047951e-01 -7.94984341e-01 -4.76124227e-01
2.27054551e-01 1.51091599e+00 -4.85967517e-01 2.97214147e-02
-4.71575230e-01 8.32567990e-01 4.16121125e-01 -8.62136856e-02
-7.95973003e-01 -5.55261195e-01 6.49697840e-01 6.16902828e-01
-8.25127065e-02 -1.91635251e-01 -2.04019308e-01 -9.10079598e-01
-2.84114540e-01 -5.28791010e-01 3.50127071e-01 -6.30605519e-01
-1.82661080e+00 2.80516535e-01 -4.38573003e-01 -7.34113634e-01
5.25261581e-01 -3.91346723e-01 -7.77679265e-01 7.21064508e-01
-1.23821843e+00 -8.50121498e-01 -1.01617113e-01 2.79102713e-01
1.76088020e-01 -7.72671700e-02 5.83295763e-01 7.77696490e-01
-8.94252419e-01 4.72410709e-01 1.70487031e-01 3.04679424e-01
7.62401342e-01 -1.06220281e+00 5.72173260e-02 8.03050220e-01
2.37815008e-02 8.25932145e-01 8.79931927e-01 -9.84999776e-01
-1.04943573e+00 -1.22160661e+00 1.92446506e+00 -6.26226366e-01
1.15415585e+00 -4.44718838e-01 -8.00479233e-01 5.03444850e-01
7.38522232e-01 -2.54272789e-01 1.14234090e+00 4.08754796e-01
-3.45818460e-01 -2.87586242e-01 -1.33141589e+00 4.44635510e-01
7.17721283e-01 -1.64335594e-01 -1.00276577e+00 8.78033161e-01
6.69074774e-01 3.22132796e-01 -8.23805928e-01 -1.74515724e-01
3.24602157e-01 -9.34681892e-01 5.63244522e-01 -8.86691511e-01
4.94321823e-01 -1.18935645e-01 8.12252834e-02 -1.23051870e+00
-4.42935467e-01 -5.43118060e-01 2.16780648e-01 1.78886187e+00
7.41406977e-01 -6.41521871e-01 4.79056001e-01 7.06161201e-01
4.38219085e-02 -5.55204332e-01 -7.97402740e-01 -3.15043151e-01
1.52630508e-01 -4.04978365e-01 1.67767540e-01 1.20483637e+00
3.34053576e-01 5.68949044e-01 -4.24039394e-01 3.16094041e-01
6.52642667e-01 -5.65684617e-01 4.18226898e-01 -1.64292824e+00
4.73973513e-01 -2.14212492e-01 -6.50253654e-01 -1.88648924e-01
-5.42179011e-02 -1.16973281e+00 -4.64706570e-01 -1.90920067e+00
4.35065985e-01 -1.12639546e-01 -4.30255413e-01 2.21083596e-01
-1.09416857e-01 3.21912736e-01 -5.21439984e-02 -2.78216153e-01
-7.47447908e-01 3.76207620e-01 9.03613925e-01 -1.16152160e-01
1.42088801e-01 -2.20064506e-01 -1.19428968e+00 6.34106040e-01
8.60635221e-01 -6.50677085e-01 -2.43455857e-01 -1.89987004e-01
9.91654277e-01 -5.31301737e-01 1.79265030e-02 -1.09751379e+00
3.44289064e-01 5.47572412e-02 3.05965096e-01 -6.13274217e-01
-3.85369696e-02 -8.94216120e-01 1.13059499e-03 3.50884616e-01
-3.32450092e-01 1.94799915e-01 -1.85412765e-01 5.55345714e-01
-6.06314540e-02 -3.72946799e-01 6.24934316e-01 1.00446507e-01
-1.93360001e-01 3.14639837e-01 -5.94551206e-01 1.39647067e-01
1.25240993e+00 -3.70456092e-02 -3.95952046e-01 -2.03679830e-01
-7.07187831e-01 5.33129096e-01 9.71221328e-02 8.58133256e-01
6.26683950e-01 -1.22595084e+00 -1.01163852e+00 3.99570577e-02
5.39390836e-03 -4.63173151e-01 3.64944190e-01 8.13622117e-01
-6.57577813e-01 1.70148849e-01 6.43401444e-02 6.22080117e-02
-1.50967789e+00 6.00770414e-01 7.07102381e-03 -1.63163856e-01
-4.78805453e-01 8.44051361e-01 -6.56112492e-01 -4.29393888e-01
2.76500553e-01 -1.05444878e-01 -8.54787767e-01 7.57379353e-01
1.00942159e+00 5.26770055e-01 -1.31571561e-01 -7.52628505e-01
-4.26368803e-01 3.19608688e-01 -1.62381932e-01 1.52598113e-01
1.65902364e+00 -1.27066553e-01 -5.76933920e-01 5.82961559e-01
1.56781292e+00 5.65909386e-01 -2.46548980e-01 2.14118123e-01
2.20696539e-01 -6.17940128e-01 2.26091683e-01 -7.65357912e-01
-9.30191517e-01 7.34490454e-01 6.58538043e-01 7.27514029e-01
3.68770540e-01 -1.98577836e-01 9.65879261e-01 1.87773444e-02
-7.31332824e-02 -1.08024871e+00 -1.26290977e-01 7.50923991e-01
6.76040947e-01 -1.15381074e+00 2.42074803e-01 -1.97345659e-01
-5.11249900e-01 7.86190033e-01 2.54361093e-01 -2.57868081e-01
1.20628786e+00 -6.40887395e-02 -6.05660677e-02 -3.65942150e-01
-6.45164788e-01 -1.71310008e-02 9.97214913e-02 4.35055166e-01
7.82901645e-01 -2.11240873e-02 -1.08470213e+00 8.05384457e-01
-3.10579866e-01 -9.84731391e-02 6.94446206e-01 4.65830564e-01
-1.09017313e+00 -1.12863588e+00 -3.33468050e-01 6.35438740e-01
-1.21838295e+00 -6.59750327e-02 -4.30139661e-01 3.97270888e-01
4.65139359e-01 1.35871911e+00 -1.35689288e-01 -5.48768818e-01
1.39328435e-01 3.30835253e-01 -1.62483528e-01 -7.56945074e-01
-9.52822328e-01 -4.95256156e-01 2.86932349e-01 -2.41783857e-01
-2.27954388e-01 -7.22658396e-01 -1.16019869e+00 -7.44929552e-01
-2.14010745e-01 4.73737001e-01 8.94336939e-01 7.63738334e-01
2.39934072e-01 3.06183219e-01 7.44847775e-01 -1.50668040e-01
-1.96116358e-01 -1.14271712e+00 -5.40857911e-01 5.35748839e-01
9.17645767e-02 -3.68966550e-01 -6.26285970e-01 -3.75051588e-01]
|
[8.861722946166992, 10.513983726501465]
|
08402206-2acf-4f80-aedc-3067175dea43
|
blind-image-quality-assessment-using-a-deep
|
1907.02665
| null |
https://arxiv.org/abs/1907.02665v1
|
https://arxiv.org/pdf/1907.02665v1.pdf
|
Blind Image Quality Assessment Using A Deep Bilinear Convolutional Neural Network
|
We propose a deep bilinear model for blind image quality assessment (BIQA) that handles both synthetic and authentic distortions. Our model consists of two convolutional neural networks (CNN), each of which specializes in one distortion scenario. For synthetic distortions, we pre-train a CNN to classify image distortion type and level, where we enjoy large-scale training data. For authentic distortions, we adopt a pre-trained CNN for image classification. The features from the two CNNs are pooled bilinearly into a unified representation for final quality prediction. We then fine-tune the entire model on target subject-rated databases using a variant of stochastic gradient descent. Extensive experiments demonstrate that the proposed model achieves superior performance on both synthetic and authentic databases. Furthermore, we verify the generalizability of our method on the Waterloo Exploration Database using the group maximum differentiation competition.
|
['Zhou Wang', 'Kede Ma', 'Weixia Zhang', 'Jia Yan', 'Dexiang Deng']
|
2019-07-05
| null | null | null | null |
['blind-image-quality-assessment']
|
['computer-vision']
|
[-6.56451061e-02 -5.17356217e-01 -8.60711746e-03 -7.66171455e-01
-1.56402767e+00 -4.97867078e-01 2.63162971e-01 -3.53512287e-01
-3.53327960e-01 3.77144068e-01 5.10204136e-01 -2.78025389e-01
9.93929058e-02 -4.94511485e-01 -7.19191194e-01 -4.81201351e-01
-1.61033735e-01 -1.18671864e-01 -7.46146217e-02 -2.58536786e-01
1.72693059e-01 4.16169375e-01 -1.21080565e+00 7.25903809e-01
1.08237529e+00 1.44202697e+00 -1.57071546e-01 8.99273276e-01
6.12051964e-01 8.53977203e-01 -7.27329671e-01 -9.22666132e-01
8.64201427e-01 -2.49835864e-01 -6.91060722e-01 3.47298294e-01
8.49820316e-01 -8.48416328e-01 -8.95861328e-01 1.21194971e+00
8.77848148e-01 -2.65353248e-02 5.19167006e-01 -1.17496836e+00
-1.03512681e+00 2.15734228e-01 -3.42682630e-01 3.14108253e-01
3.26513022e-01 6.88459516e-01 1.20122731e+00 -1.22601128e+00
1.28132313e-01 1.31884921e+00 5.85070252e-01 4.07288492e-01
-1.03495908e+00 -7.37948298e-01 3.40672769e-02 5.98392248e-01
-1.27912796e+00 -7.53343463e-01 5.87573051e-01 -3.36910516e-01
6.91801190e-01 8.63417685e-02 4.50320005e-01 1.02767944e+00
1.11326337e-01 1.00002968e+00 1.16959512e+00 1.59254149e-01
2.09315524e-01 -4.55815606e-02 -1.64646387e-01 6.00541055e-01
-1.40394911e-01 2.43941545e-01 -6.35241568e-01 3.28862318e-03
6.98791862e-01 -2.97070742e-01 -6.30861998e-01 -3.53522420e-01
-1.25385380e+00 6.57860041e-01 8.73127341e-01 -2.08513990e-01
-3.25866938e-01 9.75350365e-02 2.78308421e-01 6.74303353e-01
3.56721342e-01 3.87414128e-01 -2.50322700e-01 3.57458703e-02
-1.11587381e+00 4.51334149e-01 4.45365518e-01 9.57178831e-01
6.13769233e-01 -1.48221599e-02 -6.15119636e-01 1.09345174e+00
3.97627890e-01 3.93023044e-01 4.82644945e-01 -1.20769751e+00
7.98489153e-01 2.54658043e-01 2.09141284e-01 -9.84343350e-01
-5.58108278e-02 -6.35238767e-01 -1.06781662e+00 3.47676188e-01
3.72376561e-01 1.32844463e-01 -1.16170967e+00 1.57175469e+00
-2.26134986e-01 -6.58565462e-02 -1.22397924e-02 1.34783244e+00
7.41103530e-01 5.74263513e-01 -1.70133397e-01 4.63759676e-02
1.13598287e+00 -1.17080402e+00 -5.45257270e-01 -7.67126903e-02
1.62139043e-01 -7.12397158e-01 1.17927516e+00 7.43986130e-01
-1.47068894e+00 -9.07154381e-01 -1.42065036e+00 -5.49775660e-01
-6.87292069e-02 4.35092062e-01 2.68679589e-01 9.74818468e-01
-1.47729421e+00 4.71069515e-01 -5.42982936e-01 3.09807181e-01
7.96975732e-01 4.18283552e-01 -1.99545205e-01 -4.87055629e-01
-1.32719517e+00 5.70501924e-01 -1.42016336e-01 3.60387206e-01
-1.62911117e+00 -7.15236485e-01 -8.73548210e-01 4.10411581e-02
-1.57607272e-01 -7.78119922e-01 1.31919730e+00 -9.46757793e-01
-1.60206163e+00 9.31595981e-01 -2.33422220e-01 -4.02251124e-01
7.20972955e-01 -1.72180459e-01 -6.30911946e-01 1.90284133e-01
8.46094415e-02 6.97637618e-01 1.06037998e+00 -1.51264000e+00
-6.40951037e-01 -2.80642182e-01 4.71658885e-01 2.77899086e-01
-5.63608110e-01 1.08942501e-01 -8.96563232e-01 -9.32911932e-01
1.85020357e-01 -6.21710777e-01 -2.61791468e-01 3.63508850e-01
-4.28839356e-01 1.87486351e-01 1.40401453e-01 -1.02389205e+00
1.20488739e+00 -2.21210003e+00 2.46448979e-01 3.50738019e-01
5.33593237e-01 8.21488053e-02 -3.33884537e-01 -1.47176543e-02
-7.61183575e-02 -2.82322280e-02 -3.94384861e-01 -7.16009319e-01
2.14494407e-01 -3.88684459e-02 -1.99660838e-01 6.01709366e-01
2.88118184e-01 9.04430032e-01 -6.74981654e-01 -2.71370143e-01
-1.72712788e-01 3.69174302e-01 -9.01993752e-01 6.31097913e-01
2.40779936e-01 2.26024911e-01 -6.85641915e-02 7.63934374e-01
1.11714959e+00 -2.92548567e-01 -9.73564312e-02 -5.61952472e-01
2.56283402e-01 4.26527947e-01 -1.19754279e+00 2.02547669e+00
-6.19166613e-01 5.17565370e-01 -3.01994551e-02 -7.72167087e-01
6.10207140e-01 4.10417736e-01 2.00392455e-01 -9.77119327e-01
-7.04816133e-02 2.62244165e-01 -6.05759211e-02 -3.19358379e-01
5.56798577e-01 2.00367749e-01 5.38149029e-02 4.68648404e-01
4.02951568e-01 -1.08681507e-01 5.64702414e-02 2.44635776e-01
8.79979432e-01 -3.23306531e-01 -1.79646730e-01 1.77884791e-02
5.97852767e-01 -4.90604192e-01 5.86469173e-01 7.58675933e-01
-5.54251313e-01 1.21177030e+00 4.43027556e-01 -4.14869666e-01
-1.20885718e+00 -1.43971014e+00 -1.50909886e-01 9.97089207e-01
3.59965503e-01 -3.12577546e-01 -6.99076891e-01 -9.35673237e-01
-4.84673791e-02 6.55058175e-02 -5.64570367e-01 -2.59525418e-01
-3.56767833e-01 -8.68147552e-01 4.63988006e-01 3.88362914e-01
9.22614932e-01 -7.47553587e-01 1.36661634e-01 -1.22477591e-01
-2.97062248e-01 -1.17341781e+00 -7.93992221e-01 -3.13299716e-01
-5.90524733e-01 -9.21805024e-01 -1.03869259e+00 -8.67365599e-01
5.56347966e-01 3.29970837e-01 1.24837065e+00 9.06831995e-02
1.68679014e-01 2.05992043e-01 -1.24181561e-01 1.14837959e-01
-3.36858392e-01 -1.37757376e-01 1.13022879e-01 4.80109334e-01
-1.40607178e-01 -5.56893408e-01 -1.10944712e+00 4.92384166e-01
-7.62068391e-01 -1.86636552e-01 5.95993042e-01 8.63955736e-01
4.61254179e-01 7.27991462e-02 4.24065888e-01 -3.23051900e-01
9.13527012e-01 -2.67656773e-01 -5.19194424e-01 2.51041949e-01
-6.29831135e-01 -1.21384904e-01 4.50941652e-01 -1.84131235e-01
-9.87031877e-01 -2.28477925e-01 -3.75830501e-01 -4.25733060e-01
-3.30515243e-02 2.91054249e-01 -6.77156866e-01 -4.16301757e-01
7.20756114e-01 1.71411440e-01 -3.62269193e-01 -4.82323974e-01
4.70231354e-01 8.71232927e-01 9.61102068e-01 -4.21303451e-01
1.03402853e+00 2.86508799e-01 -5.21017969e-01 -1.47089154e-01
-8.10351610e-01 -2.35048220e-01 -4.83287424e-01 -1.29080012e-01
7.88792133e-01 -1.46557796e+00 -6.98283672e-01 9.45842743e-01
-1.11848426e+00 -3.30539167e-01 -6.95658848e-02 3.70508045e-01
-5.25265336e-01 4.27974850e-01 -9.19744670e-01 -3.74464512e-01
-2.76667893e-01 -1.78542578e+00 1.07338822e+00 1.36863962e-01
4.61703122e-01 -7.25558579e-01 -9.67480615e-02 6.67403817e-01
4.17557657e-01 -2.35465720e-01 6.06378138e-01 -1.97783262e-01
-8.08806837e-01 -1.97223946e-01 -7.11831093e-01 8.60038519e-01
1.33283809e-01 -4.06734228e-01 -1.26840472e+00 -6.21736586e-01
5.98954558e-02 -7.34152257e-01 1.07643914e+00 2.08249792e-01
1.58716619e+00 -3.40212792e-01 4.71123844e-01 1.17598403e+00
1.25644135e+00 -4.69332561e-02 8.26148927e-01 4.15051788e-01
7.27719247e-01 1.16980992e-01 3.32764149e-01 2.89404005e-01
8.35138500e-01 7.23019958e-01 6.01901412e-01 -3.39279056e-01
-3.79904360e-01 -7.83246830e-02 4.73420382e-01 8.10146332e-01
-5.18498011e-02 -1.46607906e-01 -8.24815333e-01 6.74048543e-01
-1.36761153e+00 -7.15941906e-01 2.97524214e-01 2.05300784e+00
1.18524647e+00 2.33095542e-01 1.69687286e-01 2.71820277e-01
4.27683324e-01 2.50180393e-01 -5.17830670e-01 -2.25169197e-01
-2.59727925e-01 2.82043844e-01 4.66236144e-01 4.85597342e-01
-1.20831788e+00 6.72174037e-01 6.96466398e+00 7.75795221e-01
-1.00574136e+00 1.12640202e-01 1.07781935e+00 -1.92661002e-01
-3.10409844e-01 -4.33669388e-01 -2.37270862e-01 5.16312242e-01
8.77527595e-01 -1.46233872e-01 7.17308342e-01 5.45343697e-01
2.39925757e-01 2.00637817e-01 -1.10940802e+00 1.17210591e+00
8.78097117e-02 -1.14149022e+00 3.60621065e-01 -5.56027368e-02
1.05364943e+00 6.39353544e-02 7.41718471e-01 1.73534304e-01
4.41845447e-01 -1.25004685e+00 8.19644392e-01 5.11300266e-01
9.70015168e-01 -7.92028606e-01 7.10648537e-01 -1.20559245e-01
-9.75023925e-01 -3.81698996e-01 -5.63264966e-01 3.28986853e-01
3.85945812e-02 5.06753325e-01 -2.74466097e-01 7.03491807e-01
9.97445107e-01 8.14391196e-01 -1.02161038e+00 1.41109300e+00
-2.94970661e-01 5.43371677e-01 2.89095938e-01 8.27652395e-01
8.75927061e-02 5.02483696e-02 3.45005006e-01 1.05337191e+00
2.99086809e-01 6.96542114e-02 -1.68259349e-02 8.42869997e-01
-5.83687782e-01 -1.12466559e-01 2.47059930e-02 2.94846117e-01
1.70898721e-01 1.09124398e+00 1.54881984e-01 -4.27138120e-01
-6.05967820e-01 1.39052975e+00 2.88717568e-01 6.18707180e-01
-7.37344146e-01 -4.62492555e-01 1.05248046e+00 -4.12850767e-01
2.67248005e-01 -8.73012543e-02 -4.93629664e-01 -1.49363220e+00
2.33608529e-01 -1.52766442e+00 2.34777346e-01 -9.71435547e-01
-1.46241069e+00 8.99659336e-01 -4.22143131e-01 -1.60047221e+00
-4.94804531e-02 -6.71561778e-01 -5.13335347e-01 1.24497044e+00
-1.86556458e+00 -1.12013042e+00 -6.11038446e-01 1.05231714e+00
5.13264358e-01 -3.17265034e-01 5.56841493e-01 7.52803147e-01
-6.70971692e-01 1.27195644e+00 -1.03171892e-01 4.20861572e-01
1.08854842e+00 -1.43491960e+00 7.96202719e-01 1.23834705e+00
5.49419522e-02 4.21015233e-01 3.16382349e-01 -2.17240021e-01
-1.21033120e+00 -1.35013545e+00 5.67067027e-01 -2.84558445e-01
4.88592595e-01 -3.57337654e-01 -9.80069101e-01 4.15750593e-01
2.94687361e-01 3.90697181e-01 5.09374678e-01 -1.68627739e-01
-8.19188893e-01 -3.33270371e-01 -1.22001493e+00 4.61184174e-01
1.10089982e+00 -1.06885338e+00 -3.17419708e-01 4.19560313e-01
7.99554110e-01 -7.97291160e-01 -1.19305813e+00 3.51832151e-01
5.27964056e-01 -1.22535837e+00 1.23170006e+00 -5.86500943e-01
7.91251481e-01 -2.31100202e-01 -3.95331770e-01 -1.66657889e+00
-4.41968501e-01 -5.17740548e-01 -8.53766203e-02 8.40361774e-01
6.00788414e-01 -2.98902065e-01 6.20206833e-01 4.97635871e-01
-3.02510470e-01 -7.32480407e-01 -7.66167164e-01 -6.82436764e-01
1.58109069e-01 -6.47023797e-01 8.95891488e-01 7.26096034e-01
-3.78224522e-01 -1.89043388e-01 -6.83517218e-01 5.89297175e-01
8.59867215e-01 -3.62839065e-02 6.12837076e-01 -5.64855993e-01
-4.53580946e-01 -4.60940719e-01 -4.91431624e-01 -1.45994484e+00
-1.11398928e-01 -7.89707482e-01 1.19797587e-01 -1.42902720e+00
2.26439804e-01 -2.85776913e-01 -6.41832769e-01 2.29435161e-01
-5.52118838e-01 8.48930657e-01 1.34450763e-01 2.87150890e-01
-7.78608203e-01 7.51696408e-01 1.49518287e+00 -5.43654978e-01
-6.42136335e-02 9.67031866e-02 -8.54201674e-01 2.02185601e-01
6.01479888e-01 -8.68106559e-02 -3.40657651e-01 -9.98938859e-01
2.11064145e-01 6.48024231e-02 4.45405215e-01 -1.22776830e+00
3.05772692e-01 1.71115935e-01 6.75544620e-01 -1.84830755e-01
2.46274799e-01 -5.26479900e-01 -3.81934851e-01 9.44653451e-02
-6.64894462e-01 5.94522357e-02 2.62092408e-02 3.95450890e-01
-5.76783717e-01 3.43204916e-01 1.01438856e+00 8.94632041e-02
-5.23136258e-01 1.03549886e+00 2.54930649e-02 8.22337568e-02
4.32132214e-01 9.63596106e-02 -2.49894276e-01 -8.32309127e-01
-7.77808845e-01 4.99753326e-01 4.47796196e-01 5.35540819e-01
1.00125110e+00 -1.77768099e+00 -9.01093960e-01 2.95269847e-01
2.97907770e-01 -2.06611365e-01 4.60430831e-01 6.13087773e-01
-4.23709005e-01 7.21771643e-02 -4.24270630e-01 -4.86166477e-01
-9.01822388e-01 5.84803164e-01 7.87749350e-01 -2.66386539e-01
-2.23989293e-01 1.09718418e+00 2.75349081e-01 -3.74657780e-01
5.16786754e-01 -2.60404170e-01 -1.63785920e-01 -2.31294215e-01
7.59705305e-01 2.94090569e-01 4.15401399e-01 -7.30449855e-01
-1.64675638e-01 3.15607667e-01 -1.24115974e-01 -3.77031237e-01
1.21251714e+00 -3.97216082e-01 -5.40540442e-02 1.93191618e-02
1.52882147e+00 -9.78361741e-02 -1.48239291e+00 -4.13038522e-01
-3.31782043e-01 -8.14110637e-01 4.41745520e-01 -9.67698455e-01
-1.56734002e+00 9.89596963e-01 1.04181921e+00 -1.67529345e-01
1.72964358e+00 -3.74402225e-01 8.18777084e-01 2.88624555e-01
1.64711058e-01 -7.90124536e-01 2.60709435e-01 1.59896448e-01
1.32363212e+00 -1.47643900e+00 -1.23748019e-01 -8.22568461e-02
-6.60834491e-01 9.20376897e-01 6.41840518e-01 -1.42600596e-01
6.50286317e-01 -1.12201512e-01 3.84202927e-01 -1.91518627e-02
-6.31496906e-01 1.76457651e-02 8.43733490e-01 7.27950215e-01
1.62632987e-01 1.07966624e-02 9.20424685e-02 7.17355669e-01
-4.90089417e-01 -1.18135735e-02 5.07404447e-01 4.33223605e-01
-6.86081722e-02 -9.58167732e-01 -2.84669757e-01 3.48716527e-01
-4.29497033e-01 -2.94751167e-01 -1.56319439e-01 2.98165649e-01
2.45444223e-01 1.06485605e+00 1.16352094e-02 -6.79698467e-01
6.07804477e-01 -4.49387848e-01 4.00099546e-01 -2.50825346e-01
-7.48173594e-01 -2.21484199e-01 -9.76720229e-02 -1.09064269e+00
-1.38050139e-01 -4.34165061e-01 -5.19005835e-01 -3.83660644e-01
1.63243443e-01 -8.64429027e-02 5.42945385e-01 6.79748476e-01
1.62841082e-01 5.56213558e-01 1.22367835e+00 -9.37938154e-01
-7.08177626e-01 -8.57625663e-01 -5.80186725e-01 5.50578713e-01
7.88889408e-01 -3.29321891e-01 -3.97424161e-01 9.73736718e-02]
|
[11.88398551940918, -1.8052458763122559]
|
12620c92-a521-4fc7-a25d-b174f40d4913
|
bertective-language-models-and-contextual
| null | null |
https://aclanthology.org/2021.eacl-main.232
|
https://aclanthology.org/2021.eacl-main.232.pdf
|
BERTective: Language Models and Contextual Information for Deception Detection
|
Spotting a lie is challenging but has an enormous potential impact on security as well as private and public safety. Several NLP methods have been proposed to classify texts as truthful or deceptive. In most cases, however, the target texts{'} preceding context is not considered. This is a severe limitation, as any communication takes place in context, not in a vacuum, and context can help to detect deception. We study a corpus of Italian dialogues containing deceptive statements and implement deep neural models that incorporate various linguistic contexts. We establish a new state-of-the-art identifying deception and find that not all context is equally useful to the task. Only the texts closest to the target, if from the same speaker (rather than questions by an interlocutor), boost performance. We also find that the semantic information in language models such as BERT contributes to the performance. However, BERT alone does not capture the implicit knowledge of deception cues: its contribution is conditional on the concurrent use of attention to learn cues from BERT{'}s representations.
|
['Dirk Hovy', 'Massimo Poesio', 'Federico Bianchi', 'Tommaso Fornaciari']
|
2021-04-01
| null | null | null |
eacl-2021-2
|
['deception-detection']
|
['miscellaneous']
|
[ 1.54235987e-02 7.38192052e-02 -1.94121525e-02 -6.10429704e-01
-8.77230704e-01 -9.16666985e-01 8.38968694e-01 2.75398999e-01
-4.29674327e-01 7.46159911e-01 6.25665069e-01 -4.71301913e-01
1.66206256e-01 -4.40104008e-01 -4.45884019e-01 -4.66780573e-01
3.49806666e-01 1.43410072e-01 -1.84400864e-02 -5.20448446e-01
7.13968515e-01 5.40636003e-01 -7.30375350e-01 6.86629415e-01
7.68091142e-01 8.92740071e-01 -2.59331226e-01 5.13791263e-01
-7.47651905e-02 1.37839806e+00 -1.22466779e+00 -9.32354808e-01
-8.78561512e-02 -6.92663193e-01 -1.23989534e+00 -2.33229265e-01
5.21540284e-01 -6.95876002e-01 -4.49261457e-01 9.84378576e-01
3.98002028e-01 4.05119918e-02 6.15974665e-01 -1.00771117e+00
-8.61419797e-01 4.41220284e-01 -2.65980929e-01 6.15105510e-01
4.86007214e-01 2.11460084e-01 1.04411566e+00 -6.99200928e-01
4.06695187e-01 1.33315766e+00 5.85593104e-01 7.22367465e-01
-9.28955734e-01 -5.53571165e-01 4.83018756e-02 5.88519931e-01
-9.72717822e-01 -9.16742086e-01 1.04923594e+00 -3.62043560e-01
9.54954088e-01 3.39658380e-01 2.20922872e-01 1.76907599e+00
2.38925323e-01 9.24358487e-01 1.04154658e+00 -4.27582264e-01
1.22969687e-01 2.18025997e-01 3.66573364e-01 4.76518393e-01
-2.12315414e-02 4.30343580e-03 -8.11862767e-01 -6.10151649e-01
-2.98098521e-03 -4.18041646e-01 -5.73129237e-01 1.69204652e-01
-7.88206637e-01 1.26010621e+00 5.25971770e-01 5.78305542e-01
-1.58344135e-01 7.14278147e-02 6.59498394e-01 4.42581415e-01
5.38681090e-01 8.13040912e-01 -1.10605329e-01 -5.01464844e-01
-1.00524652e+00 1.19081780e-01 1.18298995e+00 3.02628368e-01
3.33101630e-01 -7.56204501e-02 -7.56531730e-02 7.12419152e-01
9.84034389e-02 3.21060002e-01 4.54593629e-01 -9.04150009e-01
7.63687134e-01 2.67863870e-01 4.22043025e-01 -1.67083597e+00
-3.94507438e-01 -4.09392685e-01 -3.11073363e-01 -1.35496885e-01
7.17157185e-01 -2.27501929e-01 -2.80042470e-01 1.64964449e+00
1.38921998e-02 -1.58419266e-01 4.00793478e-02 1.21841776e+00
6.88807309e-01 6.02988124e-01 6.78930730e-02 -4.36577983e-02
1.29172349e+00 -7.37081826e-01 -7.84614205e-01 -8.31046820e-01
8.11478198e-01 -8.39176118e-01 7.97439039e-01 4.15277749e-01
-9.81161714e-01 3.02396566e-02 -1.16617405e+00 -4.01577145e-01
-2.39859924e-01 -6.55029640e-02 4.19073164e-01 8.04542482e-01
-7.52978384e-01 6.23209655e-01 -3.88192564e-01 -2.95626730e-01
5.07804215e-01 -1.13475740e-01 -3.04484516e-01 -4.68691625e-02
-1.63453162e+00 1.63595605e+00 6.77790269e-02 3.91695201e-01
-8.03056657e-01 -1.19083986e-01 -9.73534822e-01 9.32631046e-02
4.67208713e-01 -3.86687756e-01 1.15492666e+00 -1.52728939e+00
-1.29061079e+00 1.13452923e+00 -5.61877251e-01 -7.23379731e-01
6.52864814e-01 -4.04723376e-01 -5.52745461e-01 5.68945348e-01
1.66614249e-01 2.11452901e-01 1.03496015e+00 -1.13137984e+00
-1.47654474e-01 -6.67875469e-01 3.92516881e-01 2.26309553e-01
-2.33249396e-01 5.39079309e-01 4.87808466e-01 -4.26837534e-01
-1.38711974e-01 -6.16612196e-01 5.14091194e-01 -2.94745564e-02
-4.43471998e-01 -3.32676530e-01 8.76324594e-01 -1.17533219e+00
1.23234487e+00 -2.21920109e+00 -1.54568627e-01 -6.02149442e-02
3.10698658e-01 5.59192002e-01 1.93217963e-01 7.48208344e-01
1.95846841e-01 4.58039314e-01 -3.01398307e-01 -3.31064671e-01
9.96868312e-02 1.55816287e-01 -4.97353584e-01 7.22659051e-01
2.67860144e-01 1.03917122e+00 -7.67692566e-01 -1.57826871e-01
-1.56393908e-02 2.43029118e-01 -2.55821615e-01 5.90043236e-03
2.43428517e-02 2.82342881e-01 -2.96128869e-01 2.94580221e-01
5.16989112e-01 -1.46937028e-01 2.31758300e-02 2.30032533e-01
2.72859573e-01 1.02552485e+00 -4.04899448e-01 1.38147140e+00
-3.50593984e-01 1.14314210e+00 5.03949046e-01 -1.16327929e+00
8.06582391e-01 3.34933877e-01 -4.32454348e-01 -6.78658664e-01
1.56947061e-01 2.72401154e-01 2.98532575e-01 -8.15395832e-01
5.80741227e-01 -5.59445858e-01 -1.41927898e-01 5.93854547e-01
-2.95249939e-01 -1.91195123e-02 -2.73765773e-01 5.96860826e-01
9.21111822e-01 -2.97104090e-01 3.21258128e-01 -2.27429450e-01
5.89294434e-01 7.01780170e-02 5.20062029e-01 8.37766111e-01
-6.94380999e-01 4.25294846e-01 1.07450879e+00 -1.95997015e-01
-6.95567191e-01 -6.30493283e-01 9.55322466e-04 1.16289568e+00
9.61862653e-02 -7.45446458e-02 -6.95683360e-01 -1.02846241e+00
-8.30324888e-02 1.55571043e+00 -7.04431593e-01 -6.13781989e-01
-6.50026143e-01 -4.33591157e-01 9.97531414e-01 2.55464286e-01
4.41730112e-01 -9.97566581e-01 -6.78213298e-01 1.16120018e-01
-6.00556374e-01 -1.11370277e+00 -6.80148482e-01 -5.87254763e-02
-4.61639017e-01 -1.09451914e+00 -2.18300074e-01 -3.87831956e-01
2.55597472e-01 1.91432565e-01 1.06713951e+00 4.22217578e-01
1.59017831e-01 2.98393875e-01 -4.34281230e-01 -3.59413326e-01
-8.00945222e-01 -1.86541170e-01 -3.38124394e-01 1.20450102e-01
6.30517125e-01 -2.81571895e-01 -4.83759582e-01 1.96596935e-01
-7.24867523e-01 -4.12521183e-01 1.24111831e-01 9.48942840e-01
-6.60099864e-01 -3.70703042e-01 8.56143773e-01 -8.34579110e-01
1.07002366e+00 -7.33969569e-01 -7.05982232e-03 -1.01950215e-02
-7.03813136e-02 -6.44126236e-02 6.29161716e-01 -1.35823920e-01
-1.13102233e+00 -7.11652994e-01 -3.50815684e-01 -1.42750489e-02
-3.87661755e-01 4.05699074e-01 -1.10856868e-01 6.51207790e-02
9.22615707e-01 1.45666793e-01 2.58038729e-01 -2.64749646e-01
2.89151650e-02 8.37178946e-01 2.07199723e-01 -5.46857178e-01
3.15515041e-01 3.80616188e-01 -2.98990160e-01 -9.69790936e-01
-1.31085980e+00 -4.30089325e-01 -4.73380059e-01 4.53953668e-02
5.99253297e-01 -4.82770592e-01 -8.46168816e-01 3.19487035e-01
-1.65107882e+00 3.47100385e-02 3.59713078e-01 2.48397619e-01
-2.25584403e-01 7.61561275e-01 -9.47512984e-01 -9.73890364e-01
-3.01270097e-01 -9.93023753e-01 7.88142323e-01 -3.36955965e-01
-5.65111697e-01 -1.07359076e+00 -4.21130806e-01 9.36170161e-01
4.57937688e-01 1.29955158e-01 1.02784824e+00 -1.43811870e+00
-8.92234668e-02 -3.06481391e-01 -2.50181079e-01 6.18901432e-01
-1.57272406e-02 -4.02491570e-01 -1.10221326e+00 -3.34669016e-02
6.69657826e-01 -7.46783614e-01 8.24234545e-01 -1.47646517e-01
7.66292870e-01 -9.54227269e-01 -5.73967583e-02 -1.78395528e-02
9.28429008e-01 1.47285357e-01 6.68089986e-01 7.87573215e-03
5.21201909e-01 9.23548818e-01 2.40606308e-01 1.96233869e-01
4.70845640e-01 5.13937116e-01 5.58605373e-01 4.90546912e-01
1.64805979e-01 -2.41844803e-01 7.34990180e-01 2.37059921e-01
5.31085849e-01 -4.07045692e-01 -1.06627369e+00 5.88411510e-01
-1.57305455e+00 -1.33174706e+00 -3.30816150e-01 2.03689790e+00
9.24714804e-01 7.17860907e-02 -4.66674529e-02 1.44634858e-01
7.28570163e-01 4.80963647e-01 -3.95201355e-01 -1.00376558e+00
-1.91258773e-01 -1.77080944e-01 -1.27323925e-01 9.42796826e-01
-8.70879352e-01 9.65228140e-01 6.09922171e+00 7.60571957e-01
-1.27513576e+00 2.61907279e-01 7.36419678e-01 -2.49889389e-01
-3.47230434e-01 -1.18043862e-01 -4.44917858e-01 6.68396115e-01
1.04282713e+00 7.15695927e-03 3.23116750e-01 6.13366067e-01
3.57777804e-01 -3.46456110e-01 -1.35899007e+00 7.83028007e-01
8.41309011e-01 -1.01154292e+00 -1.16616242e-01 -7.09027499e-02
1.99115947e-01 -3.80480707e-01 -4.37019579e-02 1.41519710e-01
-1.40084932e-02 -1.34010792e+00 9.56650078e-01 2.18004972e-01
1.51799127e-01 -8.64013791e-01 9.76018369e-01 9.04441595e-01
4.97236960e-02 -4.84137684e-02 -2.89137870e-01 -4.04732972e-01
1.55572176e-01 5.23611009e-01 -1.09652948e+00 2.52389282e-01
1.90740317e-01 4.59910750e-01 -5.26223481e-01 4.14656997e-01
-6.99925542e-01 9.47111487e-01 7.64816627e-03 -4.64236856e-01
6.23839617e-01 1.41485035e-01 9.06381309e-01 1.29406285e+00
-1.53437555e-01 2.42977574e-01 -1.12149129e-02 1.15676939e+00
-1.47264987e-01 5.31534525e-03 -7.51293004e-01 -1.65894583e-01
3.16201031e-01 8.82061064e-01 -1.19112082e-01 -2.33882323e-01
-3.11158597e-01 1.31742644e+00 5.32518923e-01 2.63217211e-01
-6.18445039e-01 -3.02700907e-01 5.40075421e-01 1.42335549e-01
-2.86262836e-02 -2.78966248e-01 -3.76002014e-01 -1.20514476e+00
2.79083014e-01 -9.45924103e-01 3.65967542e-01 -7.92656600e-01
-1.48668528e+00 2.50714034e-01 -3.79344881e-01 -5.01037180e-01
-4.22202647e-01 -6.61180139e-01 -6.15910113e-01 1.00470531e+00
-1.46867800e+00 -9.83196139e-01 6.18946180e-02 5.19571304e-01
4.56883878e-01 2.30455071e-01 6.93690181e-01 2.76153162e-02
-3.27397972e-01 5.77643752e-01 -2.17687905e-01 6.55721307e-01
8.46880853e-01 -9.58270848e-01 7.24082291e-02 9.71511424e-01
8.94897729e-02 8.45559359e-01 8.34737837e-01 -4.89182085e-01
-1.01916337e+00 -4.71039653e-01 1.47190142e+00 -9.93313730e-01
9.66911077e-01 -4.81947482e-01 -1.24564564e+00 7.97139287e-01
3.90424430e-01 -2.05225572e-01 7.51482189e-01 2.04573259e-01
-7.41440296e-01 2.48503163e-01 -1.46688175e+00 4.43675458e-01
9.51960921e-01 -1.09580350e+00 -1.28571057e+00 3.49906534e-01
5.09096920e-01 -2.79607952e-01 -1.26740888e-01 -1.98661804e-01
2.79852986e-01 -1.33166051e+00 6.21142209e-01 -9.39540148e-01
8.36461067e-01 1.98082238e-01 1.82074476e-02 -1.41630936e+00
1.50983250e-02 -4.59564149e-01 2.39642505e-02 1.18329108e+00
4.75378931e-01 -9.77716148e-01 5.27394593e-01 1.07194614e+00
-1.44855589e-01 -3.84586602e-01 -1.17823064e+00 -5.34275115e-01
5.19454420e-01 -4.18950200e-01 2.39375472e-01 1.24180746e+00
4.84405905e-01 6.32289767e-01 -4.89718527e-01 6.68486953e-02
3.71129364e-01 1.16984233e-01 2.45095611e-01 -9.11843002e-01
1.13167455e-02 -4.84608233e-01 -3.23763639e-01 -1.15394902e+00
7.56333113e-01 -9.82234597e-01 -4.09325697e-02 -1.23790026e+00
1.04418188e-01 -1.28184244e-01 2.11592138e-01 3.75955492e-01
-2.64204353e-01 -2.06109229e-02 2.79284507e-01 1.85580686e-01
-4.87858057e-01 5.61457694e-01 9.79840755e-01 -1.73077717e-01
2.21308365e-01 1.43782506e-02 -1.11506474e+00 8.97289336e-01
9.40138817e-01 -4.70670849e-01 -5.10591976e-02 -6.76047564e-01
2.62661994e-01 2.99389482e-01 8.08909774e-01 -3.22871506e-01
2.47459248e-01 -7.03331679e-02 3.71780127e-01 -6.31470084e-02
5.92123628e-01 -6.89315438e-01 -6.45643890e-01 3.28845859e-01
-6.69937670e-01 -1.29093200e-01 8.63652006e-02 5.62086344e-01
-2.32843295e-01 -8.64672482e-01 6.90082490e-01 -3.64394546e-01
-3.52157980e-01 -4.76093739e-01 -5.59498668e-01 4.42066312e-01
6.16117299e-01 -1.11892872e-01 -6.53536022e-01 -9.38012481e-01
-6.48464799e-01 5.45734763e-02 3.54016393e-01 3.87694925e-01
5.34828782e-01 -8.88199568e-01 -8.52342367e-01 -2.30734900e-01
-4.99930345e-02 -5.82597136e-01 3.60989757e-02 8.20668519e-01
-2.46703118e-01 6.20391846e-01 2.98650920e-01 -2.26691410e-01
-1.13531041e+00 5.23849964e-01 5.23097694e-01 2.54176650e-02
-1.97741553e-01 9.10068393e-01 1.36485338e-01 -1.75734460e-01
7.35577568e-02 1.86922207e-01 -1.15450643e-01 2.14534342e-01
7.41466165e-01 3.77068967e-01 1.76745489e-01 -1.00764656e+00
-5.03669679e-01 -1.33080766e-01 -2.31881663e-01 -2.00101346e-01
8.99760127e-01 -3.47922087e-01 -2.79359818e-01 5.90737104e-01
1.37573981e+00 2.81049848e-01 -6.40612185e-01 -2.10283190e-01
1.80524632e-01 -7.83061326e-01 -7.34461695e-02 -1.26465178e+00
-6.58899128e-01 1.21685064e+00 -3.13610524e-01 3.67714256e-01
5.26800990e-01 9.78003908e-03 7.71036625e-01 4.34141725e-01
4.31133062e-01 -9.68153954e-01 3.61076564e-01 6.20755374e-01
1.36835468e+00 -1.44479752e+00 -2.26539820e-01 -2.91968375e-01
-1.08083642e+00 1.11983240e+00 4.88706738e-01 -1.50160976e-02
2.37699836e-01 -2.29819030e-01 1.12033352e-01 -2.37177998e-01
-6.63444161e-01 2.84878612e-01 1.95427328e-01 3.85466278e-01
5.67803264e-01 -5.48207164e-02 -5.13613343e-01 7.10921943e-01
-4.20399725e-01 -5.78596294e-01 7.14026511e-01 7.64666796e-01
-4.53314960e-01 -6.72809780e-01 -5.00129163e-01 4.03801590e-01
-1.02139258e+00 -3.72084588e-01 -1.17312598e+00 5.77705503e-01
-2.21943274e-01 1.61101401e+00 -8.57823491e-02 7.87315667e-02
7.14218710e-03 5.13092577e-01 2.79432565e-01 -6.12344980e-01
-1.06020486e+00 -4.89557117e-01 6.80364370e-01 -5.09559929e-01
-1.16721869e-01 -6.73698008e-01 -1.06030130e+00 -7.07394242e-01
-3.80086571e-01 3.01930904e-01 5.15972018e-01 1.28910017e+00
4.27617788e-01 -1.45738766e-01 5.20402312e-01 -3.27713430e-01
-9.52401817e-01 -1.07771778e+00 -3.68295908e-01 6.37354374e-01
7.77077854e-01 -3.78502458e-01 -9.59643483e-01 -3.95206392e-01]
|
[8.190781593322754, 10.409466743469238]
|
c91a4d84-317b-4e0b-872c-44dda50e2c16
|
tleague-a-framework-for-competitive-self-play
|
2011.12895
| null |
https://arxiv.org/abs/2011.12895v2
|
https://arxiv.org/pdf/2011.12895v2.pdf
|
TLeague: A Framework for Competitive Self-Play based Distributed Multi-Agent Reinforcement Learning
|
Competitive Self-Play (CSP) based Multi-Agent Reinforcement Learning (MARL) has shown phenomenal breakthroughs recently. Strong AIs are achieved for several benchmarks, including Dota 2, Glory of Kings, Quake III, StarCraft II, to name a few. Despite the success, the MARL training is extremely data thirsty, requiring typically billions of (if not trillions of) frames be seen from the environment during training in order for learning a high performance agent. This poses non-trivial difficulties for researchers or engineers and prevents the application of MARL to a broader range of real-world problems. To address this issue, in this manuscript we describe a framework, referred to as TLeague, that aims at large-scale training and implements several main-stream CSP-MARL algorithms. The training can be deployed in either a single machine or a cluster of hybrid machines (CPUs and GPUs), where the standard Kubernetes is supported in a cloud native manner. TLeague achieves a high throughput and a reasonable scale-up when performing distributed training. Thanks to the modular design, it is also easy to extend for solving other multi-agent problems or implementing and verifying MARL algorithms. We present experiments over StarCraft II, ViZDoom and Pommerman to show the efficiency and effectiveness of TLeague. The code is open-sourced and available at https://github.com/tencent-ailab/tleague_projpage
|
['Zhengyou Zhang', 'Meng Fang', 'Jiawei Xu', 'Shuxing Li', 'Xinghai Sun', 'Lei Han', 'Jiechao Xiong', 'Peng Sun']
|
2020-11-25
| null | null | null | null |
['dota-2']
|
['playing-games']
|
[-5.47151685e-01 -4.82410401e-01 -2.12527990e-01 9.64496806e-02
-8.18278015e-01 -7.11922228e-01 6.37444615e-01 3.76671664e-02
-6.32147908e-01 1.06464231e+00 -4.69698459e-01 -3.64202559e-01
-1.55882731e-01 -7.12532938e-01 -8.04851770e-01 -1.09354019e+00
-4.13974613e-01 9.79420960e-01 4.71496284e-01 -4.29222137e-01
9.60788801e-02 2.91502148e-01 -1.64558315e+00 2.26492390e-01
5.24201334e-01 6.69613481e-01 4.02584940e-01 1.02209926e+00
4.10775334e-01 1.13056552e+00 -6.34473503e-01 -2.12583199e-01
4.77794588e-01 -3.42316121e-01 -8.15091908e-01 -2.14985564e-01
2.84911916e-02 -2.11597309e-01 -1.19564965e-01 7.29605794e-01
7.47055292e-01 2.22344518e-01 -6.63179606e-02 -1.75669110e+00
2.54647136e-01 4.46658701e-01 -6.87100112e-01 1.84518799e-01
3.19052935e-01 5.38508177e-01 1.00775492e+00 -4.37795460e-01
5.81472874e-01 9.87572014e-01 4.96730804e-01 4.70852047e-01
-8.83253574e-01 -6.89815700e-01 -1.61258310e-01 4.36543167e-01
-1.06194639e+00 -2.32782006e-01 4.44646508e-01 -3.09808217e-02
1.01940191e+00 3.61912012e-01 9.95535493e-01 1.19442761e+00
4.68829036e-01 1.09415877e+00 1.48499131e+00 -2.35414088e-01
6.78769410e-01 -2.80192763e-01 -3.08048159e-01 8.95249248e-01
1.85482547e-01 3.42824131e-01 -7.05615103e-01 -3.91782165e-01
7.91800261e-01 -2.68449545e-01 1.63326994e-01 -4.29006100e-01
-1.34784949e+00 1.11832952e+00 2.71402001e-01 1.89237967e-01
-3.57609540e-01 4.99631792e-01 6.97257519e-01 5.57714999e-01
9.66149345e-02 5.01691818e-01 -5.23473978e-01 -6.84878111e-01
-7.48255491e-01 6.99194789e-01 1.10280502e+00 6.60230100e-01
6.57451391e-01 1.97586611e-01 4.43596750e-01 3.38969678e-01
8.09420869e-02 3.64603341e-01 5.48958123e-01 -1.39714551e+00
1.73285946e-01 1.32705241e-01 1.12947665e-01 -5.26486874e-01
-6.11376166e-01 -2.20381171e-01 -7.78251350e-01 8.17079186e-01
5.34917772e-01 -6.06214106e-01 -3.35932285e-01 1.33420122e+00
7.59001791e-01 5.03105462e-01 3.37249190e-01 1.14654171e+00
5.63510835e-01 8.77049565e-01 -2.58761972e-01 -1.89202368e-01
1.31927741e+00 -1.42754364e+00 -3.45414169e-02 -3.64842296e-01
6.41719639e-01 -5.66539943e-01 8.65773261e-01 7.73236811e-01
-1.19260907e+00 -2.34620318e-01 -9.53787565e-01 4.91304636e-01
-1.72356784e-01 -3.78310472e-01 1.24482179e+00 4.03027564e-01
-1.00008953e+00 7.35272765e-01 -1.42846406e+00 -2.44056284e-01
2.93057173e-01 6.18611038e-01 -3.17359716e-01 -3.47397439e-02
-8.64707351e-01 8.64939034e-01 4.75108534e-01 -3.71917158e-01
-1.33181036e+00 -3.97277743e-01 -5.63174665e-01 -9.61353108e-02
7.41588116e-01 -7.25775242e-01 1.54678416e+00 -1.09649014e+00
-1.80101383e+00 5.97445786e-01 3.88857633e-01 -6.09326422e-01
7.07277775e-01 -4.94891172e-03 4.66053896e-02 1.04718171e-01
1.80144571e-02 4.30098355e-01 7.22251952e-01 -9.87582088e-01
-9.07202244e-01 -1.83346763e-01 4.60251480e-01 5.48118949e-01
1.58898681e-01 2.73985714e-01 -2.19352767e-01 -2.89715230e-01
-4.39490616e-01 -1.22303498e+00 -4.73041385e-01 -4.08264339e-01
1.83382452e-01 -3.42561275e-01 7.63517201e-01 -7.99278095e-02
4.57611769e-01 -1.90547323e+00 3.11607152e-01 -8.27912837e-02
1.78089604e-01 2.87770927e-01 -3.13419431e-01 8.24875236e-01
3.39192927e-01 -3.87083501e-01 1.87222883e-02 -2.09181219e-01
1.35712609e-01 6.99756980e-01 4.76637408e-02 6.79632246e-01
-1.36939913e-01 7.36482203e-01 -1.21525621e+00 -5.04221439e-01
2.24524125e-01 6.03136644e-02 -6.85858607e-01 2.26712540e-01
-4.50635105e-01 6.41099334e-01 -4.60571587e-01 7.23928094e-01
4.38802004e-01 -3.14355940e-01 4.25593525e-01 4.91328895e-01
-3.02934885e-01 1.08505093e-01 -1.41175377e+00 1.78300369e+00
-4.32991832e-01 3.61157447e-01 4.11740452e-01 -1.04166794e+00
4.31128979e-01 3.63065481e-01 7.82459617e-01 -7.20123351e-01
1.34046420e-01 3.40895563e-01 1.53304473e-01 -3.39562327e-01
4.68778223e-01 -8.00539041e-04 -2.28354990e-01 7.52126455e-01
1.82068229e-01 -3.32358122e-01 7.22716331e-01 1.09411895e-01
1.43347144e+00 2.22180948e-01 3.11760426e-01 -2.67804295e-01
3.14423084e-01 4.30679381e-01 7.66341209e-01 8.68199468e-01
-2.98863679e-01 2.34757382e-02 5.14923394e-01 -8.50777209e-01
-1.03686273e+00 -9.24937546e-01 2.91847706e-01 1.52326405e+00
2.21838459e-01 -6.24014974e-01 -6.07929051e-01 -6.20501041e-01
-1.94295403e-02 3.78819078e-01 -1.74522206e-01 2.05165163e-01
-7.97848701e-01 -8.81316364e-01 5.29141665e-01 2.32945979e-01
6.71626151e-01 -1.49063420e+00 -1.27701771e+00 2.87003756e-01
1.28103778e-01 -9.64209676e-01 -3.13105993e-02 4.28626060e-01
-6.02986038e-01 -1.18556106e+00 -3.99423391e-01 -7.04311192e-01
1.57048881e-01 1.79853782e-01 1.27234304e+00 3.24503243e-01
-2.74856329e-01 3.14787358e-01 -5.59125185e-01 -3.28242242e-01
-4.62463677e-01 1.92380697e-01 3.16669434e-01 -4.74343061e-01
-3.78630608e-02 -8.28540862e-01 -4.30996269e-01 2.89357364e-01
-7.23546684e-01 2.66299039e-01 5.03049493e-01 1.14065659e+00
3.33622277e-01 1.36325687e-01 3.57954174e-01 -5.94189525e-01
5.08487999e-01 -4.65241998e-01 -1.15239775e+00 -1.22623639e-02
-4.53781366e-01 -1.57680199e-01 8.50546479e-01 -4.57977235e-01
-5.44324100e-01 7.64992163e-02 -1.33917451e-01 -4.25470233e-01
-2.04551727e-01 4.52614248e-01 4.68312442e-01 -2.58641422e-01
4.66851294e-01 2.85844207e-01 1.72993630e-01 2.34511471e-03
1.41039178e-01 3.38901162e-01 3.79809916e-01 -1.04252505e+00
6.46242797e-01 3.27252030e-01 9.87124071e-03 -6.51198626e-01
-5.70702732e-01 -2.63684779e-01 -1.37753457e-01 -3.75825912e-01
4.83979404e-01 -8.57328236e-01 -1.50749409e+00 5.61757147e-01
-9.07243073e-01 -1.01662219e+00 -2.88363934e-01 5.23449838e-01
-8.83689165e-01 8.44513774e-02 -8.60441446e-01 -7.03185439e-01
-3.80294055e-01 -1.37441742e+00 8.18645597e-01 4.87327844e-01
2.66803950e-01 -8.75070155e-01 5.94853818e-01 5.49768507e-01
3.78557235e-01 2.78982908e-01 3.22489023e-01 -5.76775789e-01
-6.17017448e-01 2.06757918e-01 2.35395864e-01 2.29102895e-02
-3.13886434e-01 2.92174164e-02 -6.34235263e-01 -9.66942966e-01
-6.31676689e-02 -9.03088510e-01 3.11455935e-01 2.25573912e-01
7.68844485e-01 -4.29997981e-01 1.30435908e-02 5.78797936e-01
1.50134921e+00 2.69227803e-01 3.19432348e-01 8.79520535e-01
2.52874881e-01 2.70248186e-02 8.65629137e-01 9.50528264e-01
4.62016761e-01 6.64384425e-01 8.23641300e-01 -4.42413688e-02
3.70937347e-01 1.98484421e-01 6.85575843e-01 9.88846242e-01
-5.22705078e-01 -1.21650994e-01 -1.05093551e+00 2.79320925e-01
-2.18058801e+00 -1.16031599e+00 6.85115308e-02 1.95213628e+00
7.74162471e-01 2.67776684e-03 5.00887692e-01 -2.67557502e-02
3.23630571e-01 1.90245584e-01 -6.69890583e-01 -5.60755789e-01
-2.33302154e-02 4.26001608e-01 4.52057779e-01 4.16116506e-01
-9.47472095e-01 1.19222069e+00 5.43015480e+00 9.77701962e-01
-1.16133881e+00 4.45067286e-01 3.70153606e-01 -2.35016927e-01
4.10249621e-01 2.11313501e-01 -4.66881067e-01 4.34618384e-01
1.07106507e+00 -3.40602309e-01 1.05811310e+00 1.01067960e+00
9.12030861e-02 -5.16719997e-01 -7.92598844e-01 1.06337297e+00
-3.72993290e-01 -1.57459617e+00 -6.28148139e-01 8.29687417e-02
7.39804089e-01 7.73223996e-01 -2.17813492e-01 6.43756926e-01
9.82490659e-01 -8.93196344e-01 7.48537958e-01 -1.25942335e-01
3.10014635e-01 -1.06982923e+00 7.61819065e-01 7.82619834e-01
-1.02143109e+00 -1.88926458e-01 -3.96087229e-01 -4.46501791e-01
-1.32088080e-01 1.85019057e-02 -6.78074300e-01 7.29023159e-01
8.55817080e-01 4.40866023e-01 -2.86634624e-01 1.06163955e+00
-2.12256789e-01 6.10195577e-01 -4.35036182e-01 -3.12176526e-01
7.24138618e-01 -3.89064342e-01 4.53732401e-01 1.06434679e+00
1.50008246e-01 2.03753531e-01 6.09745622e-01 2.97143430e-01
7.92192742e-02 7.91747123e-03 -3.03547889e-01 1.40127642e-02
3.10695380e-01 1.62102973e+00 -8.13456476e-01 -2.96375930e-01
-4.30812538e-01 9.73284066e-01 4.62390125e-01 -5.68438768e-02
-1.17996812e+00 5.47102652e-03 7.31895208e-01 -3.42749834e-01
2.08239049e-01 -4.98127937e-01 1.90206453e-01 -1.21043229e+00
-2.49627724e-01 -1.51252556e+00 5.89537501e-01 -6.00184798e-01
-1.14956188e+00 5.92568755e-01 -2.46675983e-01 -1.06771684e+00
-4.13242847e-01 -5.97495317e-01 -7.81355500e-01 1.48586512e-01
-1.35679805e+00 -1.02860165e+00 -2.91912705e-01 8.06701839e-01
6.53371513e-01 -4.61781204e-01 9.81393218e-01 2.19213992e-01
-6.20880306e-01 1.66235000e-01 3.84971768e-01 -1.51580215e-01
5.01221538e-01 -1.37829411e+00 3.85638982e-01 4.79282737e-01
3.14277470e-01 -5.88586926e-02 8.77147794e-01 -3.21965426e-01
-2.01803756e+00 -6.24658704e-01 1.98470414e-01 -1.79171771e-01
1.00119889e+00 -3.50864083e-01 -4.33565587e-01 6.54599011e-01
5.18394589e-01 2.66830593e-01 3.25414062e-01 6.07070439e-02
-2.14981865e-02 -1.37075067e-01 -9.08792853e-01 4.70146537e-01
5.79568326e-01 -8.64785835e-02 -1.97727174e-01 7.87906826e-01
2.45693877e-01 -8.61513793e-01 -8.97212267e-01 4.04352285e-02
3.96013916e-01 -1.42412102e+00 6.74650252e-01 -5.40678918e-01
3.16159368e-01 -2.33401552e-01 -1.23691574e-01 -1.44840765e+00
-2.38226894e-02 -9.78160083e-01 -1.84018180e-01 7.06702113e-01
3.19847353e-02 -7.21443653e-01 1.06415915e+00 8.60926658e-02
-1.40158728e-01 -7.71263897e-01 -1.21793103e+00 -9.98108208e-01
9.05585438e-02 -2.82707155e-01 4.95318264e-01 1.00648558e+00
2.33218282e-01 3.57776374e-01 -6.35541201e-01 1.67884529e-01
6.74394667e-01 4.30888861e-01 1.08800936e+00 -8.98782313e-01
-9.13637221e-01 -2.24494517e-01 -2.48344690e-01 -6.63380325e-01
1.27458170e-01 -8.40382636e-01 7.12465681e-03 -1.20822418e+00
1.37964666e-01 -7.28128552e-01 -7.56070167e-02 7.71484733e-01
2.52511799e-01 1.66692927e-01 6.77601874e-01 3.19549620e-01
-1.20126724e+00 4.35388982e-01 1.26278102e+00 -2.35813372e-02
-1.62784442e-01 2.40015939e-01 2.94628516e-02 7.23117769e-01
1.15615547e+00 -7.07424521e-01 -1.11178949e-01 -3.77355427e-01
3.30422342e-01 5.63388467e-01 4.21638697e-01 -1.36225331e+00
4.06942248e-01 -5.39468646e-01 -7.80380219e-02 -1.50492758e-01
5.81182182e-01 -6.45993531e-01 3.45066369e-01 9.06391382e-01
1.08553402e-01 5.90708077e-01 1.43882334e-01 3.50469314e-02
-7.99143389e-02 -2.86422968e-01 9.06771541e-01 -4.67029244e-01
-7.22605705e-01 2.67731458e-01 -5.85637033e-01 2.20639274e-01
1.39162493e+00 2.56783009e-01 -5.60983956e-01 -3.72863948e-01
-4.05871630e-01 5.38712442e-01 6.45837605e-01 6.67694658e-02
2.77413487e-01 -9.99517977e-01 -8.30390930e-01 1.34080444e-02
-2.75655806e-01 -2.89221201e-02 2.16852590e-01 1.00723624e+00
-9.19643402e-01 6.84984103e-02 -6.79734468e-01 -5.51528513e-01
-1.38738048e+00 5.34981012e-01 3.94767702e-01 -6.32732332e-01
-7.68411934e-01 5.83937228e-01 -3.76319319e-01 -5.73816419e-01
6.74889758e-02 9.74024832e-02 3.67958248e-01 -2.73600459e-01
4.63828862e-01 4.99555081e-01 -1.23662930e-02 -2.81621933e-01
-3.66083562e-01 8.09528381e-02 1.15551807e-01 -2.18464673e-01
1.81750846e+00 4.26844895e-01 -1.84224486e-01 2.60339290e-01
5.36061347e-01 -1.99238271e-01 -1.56738532e+00 1.40894845e-01
-8.73001888e-02 -1.98629513e-01 -3.99051942e-02 -6.89449191e-01
-1.12454259e+00 5.03083825e-01 4.52319175e-01 3.10055941e-01
9.00159955e-01 -1.76353361e-02 8.18457842e-01 7.30557561e-01
1.07802117e+00 -1.25412858e+00 1.91423133e-01 7.55871356e-01
6.05658650e-01 -1.30736375e+00 8.25165585e-02 2.30704054e-01
-9.12104189e-01 1.05537415e+00 7.02919781e-01 -5.30914426e-01
1.01816289e-01 6.92405999e-01 -5.15162908e-02 -2.14999571e-01
-1.23013067e+00 -1.94708467e-01 -5.49541533e-01 4.02276099e-01
-6.46820590e-02 2.10043028e-01 -1.77436650e-01 2.23318636e-01
-3.13438356e-01 -5.50363101e-02 7.32917309e-01 1.30604672e+00
-4.24902022e-01 -1.31861591e+00 -4.41369057e-01 1.48039490e-01
-3.94788325e-01 1.41016558e-01 8.91267695e-03 1.01878345e+00
9.79579091e-02 7.90477216e-01 -3.64818759e-02 -2.17180774e-01
-3.80074866e-02 -2.73341179e-01 6.88462853e-01 -2.64562994e-01
-1.07437694e+00 2.17870940e-02 1.76167965e-01 -8.51612329e-01
-4.79309708e-01 -7.02329993e-01 -1.43683946e+00 -9.49420691e-01
-4.68509272e-02 5.05538702e-01 7.44624734e-01 8.83990228e-01
1.86505228e-01 2.96160728e-01 6.51265383e-01 -1.27050519e+00
-6.00845039e-01 -5.89709580e-01 -6.66430950e-01 -3.90171893e-02
2.47320067e-02 -5.49216747e-01 -1.46779940e-01 -4.38848078e-01]
|
[3.83294939994812, 1.6925885677337646]
|
888cff46-ced8-4f87-994b-9d1149f11687
|
bayesian-optimisation-for-mixed-variable
|
2202.04832
| null |
https://arxiv.org/abs/2202.04832v2
|
https://arxiv.org/pdf/2202.04832v2.pdf
|
Bayesian Optimisation for Mixed-Variable Inputs using Value Proposals
|
Many real-world optimisation problems are defined over both categorical and continuous variables, yet efficient optimisation methods such asBayesian Optimisation (BO) are not designed tohandle such mixed-variable search spaces. Recent approaches to this problem cast the selection of the categorical variables as a bandit problem, operating independently alongside a BO component which optimises the continuous variables. In this paper, we adopt a holistic view and aim to consolidate optimisation of the categorical and continuous sub-spaces under a single acquisition metric. We derive candidates from the ExpectedImprovement criterion, which we call value proposals, and use these proposals to make selections on both the categorical and continuous components of the input. We show that this unified approach significantly outperforms existing mixed-variable optimisation approaches across several mixed-variable black-box optimisation tasks.
|
['Benjamin Ward Muir', 'David Alexander', 'Iadine Chades', 'Amir Dezfouli', 'Yan Zuo']
|
2022-02-10
| null | null | null | null |
['bayesian-optimisation']
|
['methodology']
|
[ 3.36216152e-01 1.06755085e-01 -7.25643277e-01 -5.11391044e-01
-1.24427629e+00 -6.57430351e-01 9.49120164e-01 2.08728433e-01
-6.83853745e-01 1.09322727e+00 1.64451793e-01 -6.37436569e-01
-8.72506857e-01 -5.60552418e-01 -3.86655241e-01 -1.08856511e+00
-4.15983275e-02 9.07961428e-01 -3.90225112e-01 4.65995222e-02
3.94800663e-01 3.75818729e-01 -1.45165122e+00 -1.35866538e-01
9.40687776e-01 1.13427222e+00 -2.12927729e-01 1.00220394e+00
-2.33697176e-01 2.10172921e-01 -6.65001452e-01 -6.58639610e-01
2.66020119e-01 -5.44496357e-01 -8.20483923e-01 -2.52777189e-02
-1.19850442e-01 3.58287692e-01 4.96874273e-01 9.74986970e-01
5.05074561e-01 4.24425215e-01 7.01848447e-01 -1.36627066e+00
-3.39632571e-01 6.91489816e-01 -4.95868057e-01 3.78073715e-02
1.05375394e-01 3.20045322e-01 1.71899772e+00 -3.18031698e-01
2.50827968e-01 1.55874383e+00 4.58821923e-01 1.44468769e-01
-1.99099028e+00 -2.31085211e-01 5.62868714e-01 5.00444807e-02
-8.57480049e-01 -4.83234614e-01 6.27419829e-01 -3.10089529e-01
1.08202064e+00 8.98559928e-01 5.07642210e-01 7.44041324e-01
1.02508560e-01 7.95647085e-01 1.26201999e+00 -5.48156619e-01
5.23033321e-01 -5.04978970e-02 4.75899465e-02 -2.82147918e-02
1.65404081e-01 5.52330434e-01 -3.30974340e-01 -2.72149116e-01
4.72083479e-01 -1.19236775e-01 -9.67968032e-02 -6.57770753e-01
-1.27543807e+00 1.31822872e+00 1.80630848e-01 -1.47466660e-01
-6.33455992e-01 2.28680462e-01 2.96913177e-01 5.02494216e-01
4.54114020e-01 8.74066830e-01 -7.64304578e-01 -4.25806731e-01
-7.80870795e-01 4.34861600e-01 7.99642801e-01 6.20695114e-01
6.31122053e-01 -2.39324391e-01 -5.35741389e-01 1.07631135e+00
7.66797781e-01 3.29087913e-01 3.45333397e-01 -8.46920490e-01
6.72562242e-01 4.05082881e-01 3.02332580e-01 -5.86807132e-01
-6.72943234e-01 -4.73506480e-01 -5.42223811e-01 3.43043834e-01
4.25238281e-01 -3.72397274e-01 -1.27295136e+00 1.61697233e+00
7.06427932e-01 -3.83350641e-01 -8.50202665e-02 8.54128063e-01
4.37958807e-01 4.54889536e-01 2.79732227e-01 -4.33383822e-01
1.05274367e+00 -1.05000055e+00 -7.04706728e-01 -3.05290282e-01
6.53399587e-01 -4.37063426e-01 6.00333571e-01 7.02446699e-01
-1.35510135e+00 -2.29471922e-01 -9.49296653e-01 4.93193343e-02
-2.42520452e-01 -4.95672911e-01 8.34181726e-01 8.31127405e-01
-1.09858477e+00 5.86025000e-01 -6.90824866e-01 3.59231651e-01
1.81322485e-01 8.62855434e-01 1.13814516e-04 4.06806886e-01
-1.09377909e+00 9.71806288e-01 5.27627110e-01 3.10651720e-01
-5.94418287e-01 -3.58273685e-01 -8.90967727e-01 8.95999745e-03
6.46457851e-01 -8.02413285e-01 1.41004264e+00 -1.08574736e+00
-1.71968317e+00 6.19497716e-01 -1.20659210e-01 -5.43139458e-01
3.74468118e-01 -6.71361163e-02 -1.58482730e-01 -4.11407948e-01
-8.81364346e-02 4.76935148e-01 8.29005718e-01 -1.17136884e+00
-9.39750254e-01 -4.15100724e-01 1.60302565e-01 4.06003028e-01
3.52770418e-01 2.98392624e-01 -3.53180259e-01 -9.09276843e-01
7.11636385e-03 -8.19783866e-01 -7.60471463e-01 -4.20278847e-01
-3.62756729e-01 -2.08770081e-01 -1.16621859e-01 -4.10463750e-01
1.75755739e+00 -1.87563741e+00 8.87424827e-01 5.07468164e-01
1.33170247e-01 -7.64076412e-02 -6.91424459e-02 1.03925869e-01
-4.53416288e-01 3.93331319e-01 -6.57323301e-01 -6.71390176e-01
3.67375791e-01 5.69759786e-01 6.11798316e-02 7.04730153e-01
3.47886086e-01 1.01353765e+00 -9.33596492e-01 -5.14740705e-01
2.67087400e-01 -7.59240910e-02 -7.56830812e-01 2.38487929e-01
-5.12750924e-01 3.58291388e-01 -3.70842248e-01 5.91309190e-01
4.19506937e-01 7.19579756e-02 1.87731862e-01 7.29799807e-01
-1.95138484e-01 4.07543004e-01 -1.44584703e+00 1.28820026e+00
-4.23946589e-01 1.82241052e-01 4.45478112e-01 -1.27859139e+00
8.85659337e-01 3.00522566e-01 6.18499160e-01 -5.25036812e-01
2.55765915e-01 1.89066917e-01 2.46142328e-01 -3.87332171e-01
2.81187028e-01 -6.49010122e-01 -3.30417246e-01 2.20323786e-01
8.61253664e-02 -3.16912293e-01 1.64090499e-01 -4.59851176e-01
9.16656852e-01 1.39522508e-01 6.37460768e-01 -8.23095515e-02
6.13290668e-01 -1.09611556e-01 5.47007322e-01 1.20721209e+00
-9.51733664e-02 5.58577240e-01 8.55067492e-01 -2.48787820e-01
-6.77120984e-01 -8.31772625e-01 -4.99410063e-01 1.36722052e+00
-5.73472567e-02 -2.05069497e-01 -1.91790402e-01 -7.81450570e-01
3.82015914e-01 8.03482354e-01 -7.42838860e-01 1.82364702e-01
-3.27555090e-01 -1.17006409e+00 2.98690684e-02 2.57230520e-01
-2.21338987e-01 -9.70977783e-01 -6.27921939e-01 4.81619447e-01
8.68757740e-02 -2.36506820e-01 6.20894656e-02 9.65880513e-01
-8.46323729e-01 -5.05797505e-01 -5.76445580e-01 -3.92450899e-01
3.31472382e-02 -4.91271913e-01 1.24259520e+00 -5.59614077e-02
5.88825829e-02 2.19010422e-03 -3.54335785e-01 -6.07193112e-01
-1.19323045e-01 1.01598859e-01 -4.79920805e-01 1.89518735e-01
3.81777316e-01 -2.46349186e-01 -2.84394175e-01 2.48929605e-01
-7.10716009e-01 -2.66533405e-01 5.53235352e-01 1.38700020e+00
6.23420298e-01 3.33965495e-02 3.74374121e-01 -6.45783305e-01
6.57124937e-01 -8.26315165e-01 -8.61050248e-01 3.01721841e-01
-8.83400083e-01 6.37876928e-01 2.89913625e-01 -5.17285049e-01
-7.63929605e-01 -1.79168582e-01 -2.24465504e-01 -6.43745884e-02
-2.77482361e-01 8.86713088e-01 -4.05662864e-01 1.76862717e-01
2.97131181e-01 -3.76911342e-01 -1.35861844e-01 -6.25056088e-01
5.00696242e-01 7.71216094e-01 3.40290189e-01 -7.50746667e-01
5.51541448e-01 5.26502393e-02 1.45198390e-01 -2.77775466e-01
-5.05227685e-01 -8.13730001e-01 -5.22384524e-01 5.88926002e-02
9.37226117e-01 -5.18119693e-01 -6.25385225e-01 4.37694378e-02
-9.66001153e-01 -3.08519870e-01 -5.13001859e-01 5.99087834e-01
-9.22062933e-01 -3.49267870e-02 1.38577804e-01 -1.33060086e+00
2.94096898e-02 -1.49453151e+00 1.31375408e+00 -5.15378974e-02
-5.63699484e-01 -1.33993375e+00 3.56599182e-01 8.86200294e-02
1.74744964e-01 3.14867377e-01 9.15736139e-01 -8.79339635e-01
-3.45543802e-01 -2.42365152e-01 5.69734089e-02 -4.72996011e-02
-1.58716813e-01 -8.79170895e-02 -6.93033159e-01 -1.29837632e-01
3.93604875e-01 -6.34512007e-02 9.41966653e-01 8.05392087e-01
1.19837987e+00 -4.03320014e-01 -1.88882530e-01 1.03908575e+00
1.26899803e+00 4.65913624e-01 4.35841620e-01 8.50199640e-01
2.91421741e-01 7.13316917e-01 8.42307329e-01 5.40160060e-01
2.72136480e-01 9.03796852e-01 6.42673671e-01 -2.40922436e-01
6.05907500e-01 3.90147954e-01 1.38782382e-01 2.03324080e-01
-2.04505101e-02 -4.27989364e-01 -1.09313476e+00 8.37251484e-01
-2.09048176e+00 -8.29572856e-01 -1.08636320e-01 2.32690239e+00
1.14101756e+00 3.70352566e-01 4.56689864e-01 2.72002548e-01
6.58555567e-01 2.51114905e-01 -5.23027480e-01 -1.17184091e+00
4.62840050e-02 3.84772718e-01 8.11396182e-01 8.30796778e-01
-1.40180850e+00 4.31140482e-01 7.25045204e+00 6.40169084e-01
-8.56647551e-01 -1.02475017e-01 7.09710419e-01 -6.05855227e-01
-4.23100621e-01 1.37576625e-01 -7.83798695e-01 5.54642856e-01
1.27820182e+00 6.44845814e-02 6.81981266e-01 6.72833800e-01
8.60760063e-02 -8.56607184e-02 -1.14020038e+00 7.61734545e-01
-1.86600789e-01 -6.98922217e-01 -3.05889189e-01 4.68425483e-01
7.20389068e-01 -1.04283929e-01 1.54470459e-01 3.49634111e-01
9.39365745e-01 -1.52348661e+00 6.86339498e-01 4.44589168e-01
5.08550823e-01 -8.58328402e-01 7.98815429e-01 2.29263857e-01
-5.96769035e-01 -5.94870806e-01 1.59297585e-01 -9.75902379e-02
3.84692967e-01 3.80806267e-01 -4.62914318e-01 6.41489506e-01
4.68350798e-01 4.74547952e-01 -3.12992811e-01 1.44435692e+00
-2.67466962e-01 5.59760034e-01 -3.88541073e-01 -9.78416577e-03
7.08384097e-01 -6.01507366e-01 8.38430285e-01 1.02311623e+00
-6.99665621e-02 -1.46228731e-01 1.52626470e-01 7.17546463e-01
3.76445681e-01 -1.98193751e-02 -1.11581504e-01 -8.00167248e-02
2.17981517e-01 8.63804042e-01 -5.33486128e-01 -1.39778450e-01
-4.27895784e-01 1.91831559e-01 9.18438360e-02 4.94376808e-01
-8.70246291e-01 -4.92435426e-01 8.32322776e-01 -6.43881261e-01
5.96229792e-01 -6.00872859e-02 -1.10234845e+00 -8.75334918e-01
-1.49524314e-02 -1.19746709e+00 7.93239236e-01 -2.58518279e-01
-9.38025415e-01 1.86969817e-01 4.13267821e-01 -5.89250863e-01
-6.59206331e-01 -5.83114326e-01 -3.78383517e-01 1.26850677e+00
-1.64168954e+00 -7.58726418e-01 4.65364724e-01 2.17381716e-01
5.57123482e-01 1.09259740e-01 6.07973456e-01 -1.55034646e-01
-7.70622790e-01 4.99815762e-01 6.21988595e-01 -3.04414988e-01
4.48248059e-01 -1.87939298e+00 4.94216919e-01 6.00068629e-01
3.47997583e-02 6.74686670e-01 9.44615543e-01 -5.80774426e-01
-1.19928217e+00 -4.46333379e-01 1.03311133e+00 -7.59966016e-01
8.95377040e-01 -1.87238127e-01 -6.76712453e-01 5.55826128e-01
-4.17726338e-02 -4.12885785e-01 5.89936197e-01 8.97865832e-01
1.49397433e-01 1.24623835e-01 -9.27170753e-01 5.83178043e-01
6.46270394e-01 -1.00088865e-01 -9.02677238e-01 4.08012373e-03
5.78874588e-01 -1.94909096e-01 -1.19710112e+00 5.72136343e-01
4.82363313e-01 -5.36058962e-01 1.17384148e+00 -1.01457691e+00
3.19314688e-01 6.36773556e-02 -9.84643772e-02 -1.49343598e+00
-4.06701505e-01 -9.04146194e-01 -3.48580122e-01 1.01146913e+00
6.16633058e-01 -8.72385263e-01 4.96286243e-01 1.02482116e+00
-1.25429258e-02 -1.00583208e+00 -1.38111401e+00 -6.30207241e-01
3.23220968e-01 -5.75433314e-01 9.60563719e-01 5.87181211e-01
-5.52913845e-02 3.42679769e-01 -3.21002662e-01 -1.44159496e-01
4.70861882e-01 2.73472279e-01 7.47575343e-01 -1.21657658e+00
-7.40914583e-01 -1.15068626e+00 -4.06381160e-01 -9.60223675e-01
-8.00086185e-03 -6.55180871e-01 4.49363142e-01 -1.33550489e+00
-2.44938344e-01 -6.84055805e-01 -6.83407605e-01 4.52891856e-01
-6.16995037e-01 -2.73571432e-01 -3.32184359e-02 -2.89819557e-02
-6.73602521e-01 6.31539285e-01 7.78187752e-01 -6.03740215e-02
-4.83874500e-01 3.27430010e-01 -9.49492276e-01 3.66168290e-01
5.92717707e-01 -6.75734580e-01 -1.54962778e-01 -3.70265782e-01
3.88923883e-01 2.54479825e-01 2.15029135e-01 -3.47394407e-01
2.71774288e-02 -9.33390856e-01 2.10246488e-01 -3.89686018e-01
3.31736118e-01 -8.64419997e-01 2.88293272e-01 -1.45142712e-02
-6.98363781e-01 1.35136291e-01 1.11765461e-02 5.55015028e-01
-2.03233615e-01 -3.99413168e-01 6.79717362e-01 -1.50233318e-04
-9.72606912e-02 -9.06936526e-02 -1.19157873e-01 4.70598601e-02
1.05312753e+00 -3.48143846e-01 7.20274299e-02 -1.71522409e-01
-8.60753596e-01 8.48621845e-01 3.86921227e-01 2.90993899e-01
1.18963502e-01 -1.04672217e+00 -8.49006593e-01 1.95690036e-01
-1.45034239e-01 2.16522053e-01 -2.99852639e-01 1.19296753e+00
-9.28318053e-02 7.90982723e-01 2.51816213e-01 -7.35239267e-01
-1.26448917e+00 8.70207489e-01 3.21843833e-01 -6.14795089e-01
-1.11290455e-01 1.07344007e+00 8.29907581e-02 -6.42760038e-01
4.39517140e-01 -2.75928289e-01 -2.71402389e-01 5.45170367e-01
2.86533535e-01 4.08664972e-01 1.34758726e-01 -5.64781070e-01
-3.07084352e-01 2.45365217e-01 2.62874633e-01 -5.71125746e-01
1.52216661e+00 -3.75324130e-01 -1.00013979e-01 6.18443191e-01
1.10623074e+00 -2.73472130e-01 -1.17533767e+00 -2.29312181e-01
6.13506317e-01 -6.57976031e-01 3.67342293e-01 -9.30673897e-01
-5.95280647e-01 6.78338587e-01 9.81501415e-02 5.65404415e-01
1.23232615e+00 -1.77374676e-01 3.82420570e-02 1.17071554e-01
3.43436413e-02 -1.14481723e+00 -7.37356603e-01 4.97035444e-01
8.21817338e-01 -1.35380352e+00 1.25482872e-01 1.63634107e-01
-7.51412511e-01 7.46303499e-01 5.72442189e-02 1.40739009e-01
3.98815781e-01 -8.81515630e-03 -1.56043872e-01 -1.26882687e-01
-1.22115541e+00 -4.87229258e-01 6.76276326e-01 4.02760476e-01
3.21338922e-01 1.94017962e-01 -5.70305407e-01 4.10441279e-01
-3.36044222e-01 -5.24497986e-01 -8.21186826e-02 1.00333130e+00
-1.60774946e-01 -1.38827932e+00 -8.84279549e-01 7.88757503e-01
-7.38264501e-01 -1.25435948e-01 -4.51432317e-01 6.50788784e-01
-1.17639624e-01 1.18552136e+00 2.88153403e-02 -6.34940863e-02
1.89288288e-01 1.14248954e-01 8.93802494e-02 -5.87635636e-01
-1.05641365e+00 3.62622529e-01 3.90052795e-01 -4.41524059e-01
-3.93598974e-01 -1.05292094e+00 -4.31079179e-01 -4.59147105e-03
-8.25729966e-01 5.31407118e-01 9.41810846e-01 1.10141432e+00
-1.31813243e-01 4.82589245e-01 6.44567311e-01 -1.05997741e+00
-1.03300679e+00 -7.22144067e-01 -3.13359052e-01 2.99624890e-01
8.99555683e-01 -9.08121645e-01 -5.90559185e-01 -4.17060047e-01]
|
[6.320610523223877, 3.8592095375061035]
|
4da1b5b8-f900-425b-aa05-d8fe9b11372b
|
scale-adaptive-blind-deblurring
| null | null |
http://papers.nips.cc/paper/5566-scale-adaptive-blind-deblurring
|
http://papers.nips.cc/paper/5566-scale-adaptive-blind-deblurring.pdf
|
Scale Adaptive Blind Deblurring
|
The presence of noise and small scale structures usually leads to large kernel estimation errors in blind image deblurring empirically, if not a total failure. We present a scale space perspective on blind deblurring algorithms, and introduce a cascaded scale space formulation for blind deblurring. This new formulation suggests a natural approach robust to noise and small scale structures through tying the estimation across multiple scales and balancing the contributions of different scales automatically by learning from data. The proposed formulation also allows to handle non-uniform blur with a straightforward extension. Experiments are conducted on both benchmark dataset and real-world images to validate the effectiveness of the proposed method. One surprising finding based on our approach is that blur kernel estimation is not necessarily best at the finest scale.
|
['Jianchao Yang', 'Haichao Zhang']
|
2014-12-01
| null | null | null |
neurips-2014-12
|
['blind-image-deblurring']
|
['computer-vision']
|
[ 3.98111232e-02 -6.49349988e-01 1.31817177e-01 -1.18435258e-02
-4.33274388e-01 -6.53171241e-01 6.28007352e-01 -3.12453300e-01
-2.89700598e-01 6.89262629e-01 7.61199594e-01 -1.00588379e-02
-4.22576964e-01 -1.20189078e-01 -2.95727551e-01 -7.91100383e-01
6.05357401e-02 -1.86248735e-01 3.64250034e-01 6.14379160e-02
6.19167209e-01 4.67579901e-01 -1.06738949e+00 -1.17914386e-01
1.23373306e+00 5.78433096e-01 2.51334578e-01 1.02205241e+00
2.00702220e-01 7.70149887e-01 -5.62478721e-01 -2.28658512e-01
5.72175264e-01 -4.29262161e-01 -6.25320315e-01 3.27624649e-01
9.24843252e-01 -5.03535032e-01 -4.55799431e-01 1.50355363e+00
6.77035332e-01 1.10215835e-01 5.98537028e-01 -6.05137229e-01
-1.09200680e+00 1.76930949e-01 -9.20006752e-01 7.27973461e-01
1.81680322e-01 1.84158400e-01 6.55950129e-01 -8.72618616e-01
2.32221633e-01 1.17932916e+00 9.37657416e-01 1.01998612e-01
-1.19838214e+00 -2.72336870e-01 -1.28465980e-01 3.98785353e-01
-1.44597685e+00 -6.39438093e-01 6.86466098e-01 -6.17283940e-01
5.18894136e-01 2.69153357e-01 3.11132699e-01 7.47672021e-01
2.19001666e-01 3.74534220e-01 1.94576848e+00 -4.17690456e-01
3.19098651e-01 -6.66345432e-02 4.90775526e-01 3.75844359e-01
6.56431019e-01 2.93421626e-01 -4.20861214e-01 -3.52213711e-01
1.06791878e+00 -1.34448826e-01 -8.70631158e-01 -4.19433802e-01
-1.38227022e+00 4.53742057e-01 6.52635455e-01 3.07410955e-01
-2.96480238e-01 1.37149557e-01 1.68648556e-01 4.91533279e-01
6.50229216e-01 6.13757193e-01 -3.52444619e-01 -1.16057232e-01
-1.38675451e+00 2.83005591e-02 5.63408315e-01 6.82170570e-01
5.91375113e-01 -2.70815268e-02 -3.89979988e-01 9.63612139e-01
7.48015344e-02 5.48167109e-01 7.64664292e-01 -9.30732369e-01
8.95103738e-02 1.85321737e-02 6.74391806e-01 -9.35371757e-01
-2.46199921e-01 -4.38467264e-01 -1.02182996e+00 3.85912180e-01
7.56394446e-01 -1.94007039e-01 -9.35544133e-01 1.49072921e+00
1.46329284e-01 6.13953531e-01 -5.99705651e-02 1.36546314e+00
2.31068715e-01 3.07063490e-01 -2.57216573e-01 -3.59855562e-01
1.47700834e+00 -1.04684913e+00 -9.53050971e-01 -2.85438955e-01
-9.60270837e-02 -1.08456266e+00 8.17693293e-01 2.84266680e-01
-8.37720811e-01 -6.28722727e-01 -9.96731758e-01 -2.34828331e-02
-1.06309518e-01 2.92742133e-01 3.89943659e-01 6.86450422e-01
-1.38035905e+00 5.76506615e-01 -5.48356533e-01 -6.85138166e-01
1.60225108e-01 6.36472926e-02 -1.71421543e-01 -8.67537558e-02
-8.92937362e-01 1.23630118e+00 -1.00073339e-02 2.25660712e-01
-4.65006143e-01 -4.39840317e-01 -4.63189304e-01 -1.07990757e-01
9.57684033e-03 -8.54156494e-01 9.87226248e-01 -1.07051921e+00
-1.31897461e+00 4.79265779e-01 -4.92092252e-01 -4.85091656e-01
8.13421011e-01 -6.43355787e-01 -3.85731220e-01 1.67583674e-01
1.10719930e-02 2.10040137e-01 1.66087198e+00 -1.38645959e+00
-2.19525546e-01 -2.61454254e-01 -1.81365088e-01 2.59957284e-01
-3.74496520e-01 2.84967184e-01 -1.48916185e-01 -1.20833623e+00
1.71533108e-01 -8.50270927e-01 4.24942933e-02 -1.90295652e-02
-6.16567060e-02 2.06836432e-01 5.97610831e-01 -1.11402833e+00
1.35963976e+00 -2.08797884e+00 3.01657766e-01 -1.71238840e-01
3.83942634e-01 1.45742819e-01 -3.76613848e-02 1.33120507e-01
-2.78116882e-01 -2.58306354e-01 -3.26326519e-01 -1.16171151e-01
-1.78398803e-01 -2.30322540e-01 -4.45678622e-01 9.57055390e-01
-8.76915157e-02 7.57566214e-01 -8.07814002e-01 -1.85340181e-01
2.43531719e-01 4.18016791e-01 -2.69488871e-01 3.28409761e-01
3.87568295e-01 4.21299011e-01 9.05313641e-02 5.99599123e-01
1.08373594e+00 -3.75497818e-01 -3.05175871e-01 -5.06250024e-01
-1.94662884e-01 -2.98757136e-01 -1.25211990e+00 1.44810343e+00
-3.02198291e-01 9.61254001e-01 3.56482536e-01 -5.07899642e-01
4.34710592e-01 3.18287253e-01 6.86925352e-02 -1.15797326e-01
4.56527574e-03 2.66536117e-01 -7.38914264e-03 -5.99592268e-01
6.84426308e-01 -2.80900717e-01 4.14751709e-01 3.93713832e-01
-1.64894611e-02 -6.62106648e-02 -1.67953789e-01 5.19360639e-02
9.61497128e-01 -2.22921953e-01 5.24066389e-01 -6.56608343e-01
5.51858008e-01 -3.40026975e-01 9.29971263e-02 9.84300673e-01
-6.44704342e-01 8.58735919e-01 1.22479703e-02 -2.91934758e-01
-1.14681470e+00 -9.36063826e-01 -2.79001296e-01 8.56530368e-01
5.89569390e-01 -2.03635432e-02 -9.91920352e-01 -5.88989079e-01
-1.17917843e-01 2.98659921e-01 -7.04634011e-01 9.87771675e-02
-4.38387364e-01 -1.11764920e+00 3.35427880e-01 3.13326299e-01
6.64502621e-01 -5.57585120e-01 -2.78538823e-01 -9.96621624e-02
-1.21604674e-01 -1.19873917e+00 -1.06307101e+00 -2.48202220e-01
-9.52808976e-01 -8.47334146e-01 -1.36045003e+00 -7.14502573e-01
8.45959842e-01 8.93917382e-01 6.73645437e-01 -1.56821925e-02
-1.74315959e-01 3.19486022e-01 -2.45840743e-01 1.20181002e-01
-4.20026124e-01 -3.01162720e-01 3.44276488e-01 2.82120824e-01
1.27118856e-01 -4.57703322e-01 -9.32803750e-01 5.23018420e-01
-8.23390365e-01 -8.01625475e-02 7.68150210e-01 8.70417237e-01
-8.03754479e-02 3.04874748e-01 5.23245037e-01 -1.55083567e-01
1.03405130e+00 -2.73476005e-01 -6.51730895e-01 2.64702976e-01
-8.74369979e-01 9.65479314e-02 3.65075558e-01 -7.07447648e-01
-1.17087436e+00 -3.60862732e-01 5.50413013e-01 -4.76416111e-01
-2.31964499e-01 2.53421869e-02 2.20321313e-01 -5.31919539e-01
9.91526723e-01 2.67718107e-01 7.46058002e-02 -7.40182519e-01
5.90191662e-01 8.80786836e-01 6.17627740e-01 -2.85566360e-01
1.05640018e+00 5.24810374e-01 -3.52916181e-01 -9.16194975e-01
-5.97646534e-01 -6.68807387e-01 -6.86153710e-01 -3.76400873e-02
7.41404831e-01 -1.03654850e+00 -1.94448203e-01 1.01100051e+00
-1.19020724e+00 -1.88397139e-01 -5.23949899e-02 5.42605460e-01
-3.07665288e-01 8.68888021e-01 -8.45708549e-01 -6.43631756e-01
-3.27878207e-01 -1.18187141e+00 8.15160573e-01 4.25044805e-01
-1.57339200e-01 -1.14536178e+00 1.69858396e-01 4.57346678e-01
9.79260564e-01 -3.92056495e-01 4.61882651e-01 -1.98251635e-01
-4.62440640e-01 -1.95380032e-01 -9.44038928e-01 4.48784232e-01
6.71709001e-01 -4.07613248e-01 -1.17456627e+00 -7.19925761e-01
4.52561200e-01 2.41146013e-01 9.60788250e-01 6.17734253e-01
6.41658485e-01 -4.23891783e-01 -1.41124483e-02 6.15747809e-01
1.53650510e+00 -4.57344890e-01 5.25992334e-01 4.71710414e-01
8.57334077e-01 3.17625672e-01 2.00969025e-01 2.33063132e-01
1.33090511e-01 8.43337297e-01 8.65134671e-02 -1.40088767e-01
-8.36311579e-01 3.24418843e-01 5.53270519e-01 7.65960097e-01
-1.74563333e-01 3.32615785e-02 -6.62901640e-01 6.44796848e-01
-1.63955939e+00 -9.31489587e-01 -1.48491576e-01 2.32395148e+00
1.09503937e+00 -2.87393540e-01 2.21108254e-02 -1.86424285e-01
1.04174042e+00 1.49395630e-01 -3.54704469e-01 2.37651095e-02
-3.47606033e-01 -4.03326079e-02 8.82649660e-01 9.84508812e-01
-1.22629929e+00 9.74124432e-01 7.32835674e+00 6.84329450e-01
-1.18699956e+00 2.51157910e-01 1.31633043e-01 1.20181523e-01
1.14815980e-01 2.25934774e-01 -3.26619983e-01 6.23268247e-01
5.56441367e-01 -3.11204106e-01 9.62230563e-01 4.21061963e-01
4.70793605e-01 -2.56025106e-01 -6.37508392e-01 1.12122297e+00
9.42057818e-02 -1.01420331e+00 -1.66414201e-01 -8.57630596e-02
9.29837346e-01 2.03473419e-01 4.71169613e-02 -4.84301120e-01
3.59726489e-01 -8.87500226e-01 8.10737193e-01 6.36491776e-01
6.67133152e-01 -5.67813367e-02 7.45675385e-01 7.27394745e-02
-7.86292672e-01 -1.41887248e-01 -3.12758893e-01 -3.63857821e-02
-6.49019191e-03 7.18526900e-01 -7.28323758e-01 3.15746933e-01
6.26765311e-01 8.73649359e-01 -1.02761841e+00 1.62792385e+00
-1.61351502e-01 5.86741507e-01 2.41794474e-02 3.65953177e-01
-2.62917787e-01 -2.18409672e-01 1.01211655e+00 1.61231101e+00
4.92695779e-01 -1.22997440e-01 -3.84620905e-01 6.55486345e-01
2.04192311e-01 -1.46230444e-01 -2.34239191e-01 3.47878933e-01
4.70566392e-01 1.16258609e+00 -6.83450997e-01 -3.77080858e-01
-4.71899599e-01 1.61171424e+00 7.77124614e-02 8.95044923e-01
-6.62702799e-01 -2.60339439e-01 7.59380817e-01 2.07314976e-02
4.24612045e-01 -5.07214546e-01 -5.48183203e-01 -1.65721869e+00
-1.14399955e-01 -9.49200809e-01 1.39754876e-01 -1.05184901e+00
-1.68764961e+00 4.08486485e-01 -3.18290666e-02 -1.44741333e+00
2.89104640e-01 -4.89432216e-01 -5.00818372e-01 1.17984378e+00
-1.86778331e+00 -1.02649963e+00 -5.49016178e-01 5.48561811e-01
5.40017188e-01 1.70650787e-03 4.30834949e-01 1.40216902e-01
-3.17997634e-01 4.05620188e-01 3.50635558e-01 -1.46602735e-01
1.32567930e+00 -1.55212343e+00 4.01696265e-01 1.38507926e+00
-1.17392585e-01 9.71382856e-01 1.14896190e+00 -6.95762694e-01
-1.25212228e+00 -8.44651759e-01 4.21924651e-01 -7.26645231e-01
1.01393139e+00 1.17318973e-01 -1.07239771e+00 2.93471456e-01
5.63182294e-01 2.21450552e-01 2.69052923e-01 -2.40726247e-01
-5.58732688e-01 -3.44908759e-02 -1.11267483e+00 3.89855772e-01
6.49540663e-01 -6.89576387e-01 -8.15215528e-01 1.18345998e-01
4.88707989e-01 -2.41872758e-01 -7.28014290e-01 2.43155077e-01
3.86685401e-01 -1.03213084e+00 1.09961021e+00 -9.50753242e-02
8.79482478e-02 -7.16233194e-01 -5.77927977e-02 -1.64695203e+00
-7.95749009e-01 -8.89416695e-01 -3.46738786e-01 9.89689589e-01
-1.74633861e-01 -6.64957285e-01 2.02003062e-01 3.34456563e-01
1.97831869e-01 1.13115221e-01 -8.98386419e-01 -7.85540998e-01
-2.87455525e-02 -5.22456085e-03 3.16134930e-01 1.10252595e+00
-2.36492574e-01 1.50878474e-01 -8.34778428e-01 6.89436138e-01
1.11651731e+00 1.85590312e-01 5.39202154e-01 -1.01862395e+00
-3.51575583e-01 -5.47916114e-01 -3.95270228e-01 -1.26698625e+00
-1.90548614e-01 -4.11050200e-01 1.32166639e-01 -1.40415370e+00
3.84792596e-01 -1.43538788e-01 -3.41604173e-01 1.38982944e-02
-7.98458755e-01 3.84784251e-01 -6.16355278e-02 8.67338240e-01
-2.52810925e-01 1.50622413e-01 1.41891229e+00 1.40148010e-02
-6.52606264e-02 -9.69421491e-02 -7.87973404e-01 7.39035130e-01
5.47761023e-01 -1.03458479e-01 -2.46205121e-01 -6.45567358e-01
4.59123123e-03 -8.87672603e-02 6.94761217e-01 -9.51202571e-01
3.78110379e-01 -2.52512395e-01 4.42297459e-01 1.62718832e-01
2.01374777e-02 -8.05991292e-01 1.19620385e-02 1.21402234e-01
-1.53441504e-01 -1.95856825e-01 1.65700376e-01 8.45283151e-01
-8.78793076e-02 2.92841531e-03 1.19630861e+00 1.17211044e-01
-6.86045349e-01 -6.55198619e-02 -9.40441340e-02 -1.61710545e-01
5.63635707e-01 -2.90503025e-01 -6.41732991e-01 -4.13210660e-01
-5.64952791e-01 -4.22349930e-01 9.38767076e-01 4.97852296e-01
3.29990298e-01 -1.10892081e+00 -7.82785654e-01 2.23127693e-01
-6.04179390e-02 -6.05618954e-01 2.10881293e-01 1.15888441e+00
-4.98300135e-01 1.06845893e-01 -3.92819047e-02 -3.86624306e-01
-1.24831212e+00 5.26652038e-01 5.67717254e-01 2.07082003e-01
-5.76941490e-01 9.37628567e-01 1.82651699e-01 2.31015697e-01
-3.11831813e-02 -4.43298489e-01 -1.60673305e-01 -3.75047438e-02
7.31479406e-01 5.89858592e-01 6.67692497e-02 -8.97016943e-01
-1.80627733e-01 8.91441703e-01 -1.01012848e-01 -1.63201287e-01
9.80005383e-01 -8.10258508e-01 -5.24359524e-01 1.38559967e-01
9.11514282e-01 4.83329147e-01 -1.45626223e+00 -4.30585861e-01
3.43212821e-02 -9.50072229e-01 3.98058087e-01 -8.63686621e-01
-7.95359671e-01 5.59793890e-01 1.03881359e+00 3.73532444e-01
1.23002064e+00 -7.89357126e-02 6.62245750e-01 -1.02239497e-01
1.68001816e-01 -7.42069781e-01 1.56593934e-01 2.13148445e-01
1.01761627e+00 -1.25608873e+00 2.90434241e-01 -1.35433421e-01
-4.20269877e-01 1.03754413e+00 2.80119210e-01 -2.48943850e-01
5.77291489e-01 1.63524762e-01 2.30712622e-01 2.96058189e-02
-2.35920370e-01 -1.91925503e-02 5.53557932e-01 4.50342745e-01
3.48099947e-01 -7.74523523e-03 -2.93533176e-01 2.46049300e-01
7.14139268e-02 1.36208713e-01 6.92515790e-01 6.21808529e-01
-6.77424669e-01 -8.24466050e-01 -1.06808650e+00 1.52050138e-01
-5.21512389e-01 -4.01938677e-01 -2.55304009e-01 1.29297122e-01
-1.87722161e-01 1.02317619e+00 -2.51744807e-01 -1.06502794e-01
-1.53783783e-02 -1.60033286e-01 5.70816576e-01 -1.42524749e-01
-1.85262531e-01 3.94253343e-01 -4.11430627e-01 -3.40076715e-01
-5.35356164e-01 -6.35086179e-01 -6.12907827e-01 -1.87996969e-01
-7.02965319e-01 2.47998908e-02 4.71970677e-01 7.17267513e-01
4.63694423e-01 1.01960130e-01 5.82265139e-01 -8.82469594e-01
-8.64808202e-01 -1.34450173e+00 -5.90237796e-01 5.06865203e-01
1.13137484e+00 -5.87766767e-01 -9.94790673e-01 4.02192742e-01]
|
[11.630448341369629, -2.75104022026062]
|
c014c1f0-5879-47c6-a422-e2cb0c86a372
|
parameter-free-geometric-document-layout
| null | null |
https://ieeexplore.ieee.org/abstract/document/969115
|
https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=969115
|
Parameter-free Geometric Document Layout Analysis
|
Automatic transformation of paper documents into electronic documents requires geometric document layout analysis at the first stage. However, variations in character font sizes, text line spacing, and document layout structures have made it difficult to design a general-purpose document layout analysis algorithm for many years. The use of some parameters has therefore been unavoidable in previous methods. In this paper, we propose a parameter-free method for segmenting the document images into maximal homogeneous regions and identifying them as texts, images, tables, and ruling lines. A pyramidal quadtree structure is constructed for multiscale analysis and a periodicity measure is suggested to find a periodical attribute of text regions for page segmentation. To obtain robust page segmentation results, a confirmation procedure using texture analysis is applied to only ambiguous regions. Based on the proposed periodicity measure, multiscale analysis, and confirmation procedure, we could develop a robust method for geometric document layout analysis independent of character font sizes, text line spacing, and document layout structures. The proposed method was experimented with the document database from the University of Washington and the MediaTeam Document Database. The results of these tests have shown that the proposed method provides more accurate results than the previous ones.
|
['and Dae-Seok Ryu', 'IEEE', 'Senior Member', 'Seong-Whan Lee']
|
2001-11-01
| null | null | null |
ieee-transactions-on-pattern-analysis-and-19
|
['document-layout-analysis', 'texture-classification']
|
['computer-vision', 'computer-vision']
|
[ 4.07639205e-01 -5.41656733e-01 1.33053318e-01 -7.67211914e-02
-1.73119217e-01 -7.70288765e-01 4.09138411e-01 4.75075841e-01
-2.99890880e-02 5.14122963e-01 -1.21054016e-01 -5.86897373e-01
-4.61415827e-01 -7.38674283e-01 -1.88013941e-01 -4.70919758e-01
1.85379490e-01 2.06159934e-01 5.51269352e-01 1.31934896e-01
9.10643220e-01 7.71103859e-01 -1.23286533e+00 2.75696635e-01
1.04429162e+00 5.30996382e-01 3.86910111e-01 9.26369429e-01
-6.64448738e-01 2.91170210e-01 -8.19439113e-01 6.63306341e-02
1.57047376e-01 -6.59199655e-01 -6.66901827e-01 1.06903875e+00
1.69199511e-01 -3.12429398e-01 2.02076867e-01 1.03568935e+00
1.40078664e-01 1.53729722e-01 1.00815272e+00 -5.25307357e-01
-5.95950544e-01 5.15432239e-01 -1.13219774e+00 4.57330346e-02
1.87207311e-01 -4.76893485e-01 7.20597684e-01 -6.81442142e-01
8.02153647e-01 1.13454282e+00 2.96651930e-01 -2.30035767e-01
-1.09395635e+00 -1.35278492e-03 2.35124025e-02 1.28290623e-01
-1.38880229e+00 -6.66247755e-02 9.84786451e-01 -5.02654731e-01
5.72889209e-01 3.86156052e-01 3.22670639e-01 2.66804636e-01
5.17249405e-01 6.25833213e-01 1.14384639e+00 -1.15229654e+00
1.34194285e-01 1.96348786e-01 5.46267152e-01 9.05376613e-01
4.26812321e-01 -8.81031513e-01 2.09731072e-01 -3.37106250e-02
9.63166356e-01 -3.94620389e-01 -1.71010405e-01 -5.82325220e-01
-9.96621251e-01 4.30759072e-01 -2.34998912e-01 8.32471132e-01
-1.62322238e-01 -4.75243151e-01 3.94016176e-01 -6.48208037e-02
2.30507888e-02 2.20561147e-01 4.23208065e-02 -2.00913791e-02
-1.18213022e+00 -2.91225296e-02 6.46462977e-01 9.50700939e-01
3.09051633e-01 -9.79540274e-02 6.88235313e-02 1.01968455e+00
9.25941393e-02 4.40080017e-01 2.70290077e-01 -6.82454228e-01
6.84219480e-01 6.67388976e-01 2.45993406e-01 -1.49464703e+00
-2.50424802e-01 -3.54465544e-02 -7.44922161e-01 2.10356861e-01
4.77573365e-01 5.03426827e-02 -7.08904862e-01 7.10516632e-01
1.34964302e-01 -9.44859087e-01 -4.05113026e-02 4.28928852e-01
8.81672055e-02 8.55749249e-01 -4.92326468e-01 -3.84590089e-01
1.41807318e+00 -7.21650064e-01 -1.06967986e+00 4.86734569e-01
7.22924352e-01 -1.42163682e+00 1.21234989e+00 8.01920831e-01
-9.03768063e-01 -6.22683227e-01 -1.27882302e+00 1.86811268e-01
-2.71938920e-01 1.03810716e+00 1.22304425e-01 7.46485531e-01
-6.79237068e-01 2.83368349e-01 -5.78699052e-01 -6.28867984e-01
-2.68310398e-01 1.27612442e-01 -6.29297420e-02 2.16267154e-01
-3.80731910e-01 5.37533998e-01 2.56023288e-01 4.14246738e-01
4.10465330e-01 2.60129899e-01 -2.23917022e-01 7.96341226e-02
7.84178153e-02 -4.36175652e-02 8.77383649e-01 -7.96863914e-01
-1.35243177e+00 4.49102104e-01 -1.24555230e-01 5.09276427e-02
6.44090295e-01 8.99622515e-02 -6.43595755e-01 5.18076718e-01
-2.05732137e-02 1.75998107e-01 9.47525144e-01 -1.14910209e+00
-5.52749634e-01 -3.12721729e-01 -4.82884735e-01 1.10140570e-01
-2.49615893e-01 2.72854324e-02 -6.91746354e-01 -7.83703268e-01
5.51745057e-01 -7.51980364e-01 2.40507692e-01 -3.03005069e-01
-7.06258357e-01 -4.25662939e-03 1.27599263e+00 -9.35546398e-01
1.44129622e+00 -2.15624142e+00 -3.53074491e-01 9.88831818e-01
-2.67460704e-01 2.57188827e-02 8.03456828e-02 4.24440712e-01
1.10041246e-01 4.75558341e-02 -1.28997728e-01 2.51399726e-01
-1.69806749e-01 -9.70704705e-02 -1.71808064e-01 3.52130294e-01
-1.65280700e-01 6.51483759e-02 -1.43905252e-01 -9.63410974e-01
2.22779527e-01 2.01801464e-01 -1.81298509e-01 -2.18702272e-01
-1.47563770e-01 -4.47742827e-02 -7.12331295e-01 7.74035990e-01
9.05821204e-01 -8.24947667e-04 4.82452452e-01 -4.19649839e-01
-5.10594845e-01 -1.78177863e-01 -1.57789099e+00 9.52429891e-01
-1.32357910e-01 8.18655133e-01 3.95498099e-03 -7.18040764e-01
1.39318526e+00 6.62789419e-02 3.75894308e-01 -7.06590712e-01
2.35298350e-01 2.87532181e-01 -1.11044236e-02 -4.90714818e-01
9.65147376e-01 6.51184797e-01 4.91998419e-02 4.57071394e-01
-6.03147328e-01 -1.60651743e-01 8.85137081e-01 4.43007797e-02
6.48823440e-01 8.96894038e-02 3.14967670e-02 -5.28230965e-01
7.09381998e-01 2.06823662e-01 1.92163453e-01 5.03737271e-01
1.95188448e-01 5.55343747e-01 7.18805909e-01 -1.96301743e-01
-1.36709702e+00 -8.07924688e-01 -3.36681068e-01 4.36739594e-01
1.27869740e-01 -3.63522559e-01 -1.17757213e+00 -2.26720467e-01
-1.85433373e-01 4.62394118e-01 -1.45433635e-01 5.05611122e-01
-7.71148443e-01 -6.43807411e-01 2.54835606e-01 3.13220829e-01
8.42591882e-01 -8.22756529e-01 -4.80440617e-01 3.08196187e-01
-1.12766199e-01 -9.22146201e-01 -5.44022739e-01 4.82041528e-03
-1.02336061e+00 -1.06191027e+00 -9.85737324e-01 -1.04758692e+00
1.14667571e+00 5.22463620e-01 3.22804242e-01 5.95071949e-02
-3.53255063e-01 2.32440099e-01 -5.59860170e-01 4.62597162e-02
-5.95841765e-01 1.92568839e-01 -1.73900932e-01 4.25565206e-02
-8.04174542e-02 -2.77823023e-02 -4.04372573e-01 5.91049254e-01
-9.70704198e-01 2.43850350e-02 5.33767879e-01 3.50820869e-01
5.56929708e-01 7.37890184e-01 1.61087528e-01 -7.47108519e-01
1.04047191e+00 3.43941331e-01 -9.20684755e-01 6.28897011e-01
-5.69486201e-01 1.92738324e-01 8.19206774e-01 -1.59660757e-01
-1.28313577e+00 -1.16818406e-01 2.94199109e-01 2.77842343e-01
-2.28288203e-01 3.76247376e-01 -4.05179888e-01 -6.33614659e-02
4.30244386e-01 3.43641937e-01 -1.14146553e-01 -6.94325686e-01
4.64338288e-02 9.48876143e-01 5.58721125e-01 -5.10921299e-01
7.33937919e-01 1.13794759e-01 1.52238846e-01 -1.39351785e+00
-3.22803892e-02 -3.44898641e-01 -1.02468503e+00 -2.99943954e-01
8.81894290e-01 -1.59633085e-01 -4.08443362e-01 5.80121160e-01
-1.17505801e+00 5.18267713e-02 2.65452087e-01 4.46646869e-01
-2.99796999e-01 1.09921682e+00 -4.62971777e-01 -7.02477753e-01
-9.33095440e-02 -9.68350887e-01 5.90910316e-01 2.38126472e-01
-4.64128792e-01 -8.40053976e-01 -1.40905976e-01 2.23155394e-01
2.00976711e-02 1.47491992e-01 1.46607876e+00 -8.02997425e-02
-6.36556745e-01 -2.72064596e-01 -3.01864564e-01 2.04153061e-01
6.04734898e-01 8.63064587e-01 -3.98305327e-01 1.71909593e-02
-1.06686339e-01 3.79415751e-01 4.56311673e-01 3.99074584e-01
1.22239625e+00 -1.42881945e-01 -2.69349039e-01 2.39222571e-01
1.39800549e+00 9.04667139e-01 6.45698607e-01 7.75073349e-01
4.88872260e-01 3.87618691e-01 8.35295737e-01 5.30904412e-01
-1.44218341e-01 4.40825939e-01 -3.33296567e-01 -1.69626191e-01
7.60769472e-02 4.38213646e-02 3.09428927e-02 9.91347373e-01
-1.69838928e-02 -7.38071501e-01 -7.14354873e-01 3.56133044e-01
-1.47254419e+00 -7.54114568e-01 -4.65744227e-01 1.91038871e+00
6.70642316e-01 4.31157678e-01 6.29976243e-02 7.08617508e-01
1.03053176e+00 5.16592003e-02 -1.41884070e-02 -1.11451399e+00
-3.25182915e-01 -3.18133235e-01 4.82898682e-01 4.42323506e-01
-1.02978671e+00 5.97822905e-01 6.16336775e+00 8.94478023e-01
-1.01634669e+00 -7.35500395e-01 6.74359858e-01 4.30035591e-01
-1.92677230e-01 -8.28017145e-02 -8.03893745e-01 4.17102486e-01
2.94511139e-01 9.01864395e-02 1.40607342e-01 5.45246065e-01
5.66270232e-01 -7.66469300e-01 -5.07095814e-01 6.91393197e-01
-1.26088589e-01 -1.09981477e+00 3.47488731e-01 -5.37350960e-03
7.41153538e-01 -1.05668330e+00 1.17097958e-03 -5.84710658e-01
-4.44063157e-01 -4.86953497e-01 6.42571747e-01 5.60360909e-01
4.19590890e-01 -8.47875953e-01 5.16626477e-01 1.00287251e-01
-1.21172225e+00 1.95226714e-01 -3.21751922e-01 3.37801844e-01
-1.72907442e-01 6.16361678e-01 -8.41033399e-01 5.06096363e-01
2.33958438e-01 6.72022700e-02 -6.74427807e-01 1.09383869e+00
2.65083104e-01 4.29019481e-01 -3.70590448e-01 -3.14605862e-01
3.20574582e-01 -7.81921268e-01 2.66357005e-01 1.35337436e+00
5.84293187e-01 -1.38161704e-01 -6.14633709e-02 6.96224451e-01
2.82734215e-01 6.56082928e-01 -2.54115105e-01 -3.60021323e-01
3.21623266e-01 1.00195682e+00 -1.69334924e+00 -1.13854267e-01
-3.76502395e-01 9.58949268e-01 -4.17135715e-01 4.54781353e-01
-5.52945018e-01 -1.04155838e+00 -3.48673612e-01 3.55031371e-01
5.29471695e-01 -8.36522579e-01 -7.62104988e-01 -6.22935891e-01
3.49840343e-01 -9.15526927e-01 2.79573090e-02 -5.67212999e-01
-5.14989495e-01 4.63521421e-01 -4.29644324e-02 -1.30215418e+00
-1.04164898e-01 -6.91269577e-01 -6.92289054e-01 8.47136021e-01
-6.62069857e-01 -5.89528084e-01 -7.60258958e-02 2.92491615e-01
6.74557328e-01 -3.13445896e-01 6.14633024e-01 -8.86503085e-02
-6.86938405e-01 4.29821134e-01 8.37539673e-01 1.47169873e-01
5.65007269e-01 -1.08857334e+00 7.56674409e-02 1.09514821e+00
2.45278538e-03 6.10160708e-01 7.65755296e-01 -8.31934214e-01
-1.11017716e+00 -4.48454797e-01 8.91686738e-01 2.26613641e-01
4.82129127e-01 -3.71128857e-01 -9.31680262e-01 1.13637730e-01
3.24284673e-01 -8.21159840e-01 4.01721954e-01 -2.68770337e-01
1.78213686e-01 -2.37608790e-01 -8.87606680e-01 7.04769433e-01
1.95844233e-01 -1.95232585e-01 -5.49358547e-01 2.27737293e-01
-1.03457443e-01 -1.42979652e-01 -6.91622257e-01 -7.10270321e-03
6.47064447e-01 -8.77818704e-01 5.47728181e-01 2.96982706e-01
1.73572078e-01 -6.82418287e-01 1.59621999e-01 -6.50021791e-01
-2.97440350e-01 -5.61744392e-01 5.52210331e-01 1.39435387e+00
4.74892646e-01 -3.75909537e-01 6.11411870e-01 4.77140546e-01
1.29793838e-01 -3.84202212e-01 -2.45159790e-01 -7.95001745e-01
-4.82030898e-01 2.18152508e-01 2.93613374e-01 6.91262603e-01
-1.92720611e-02 -2.31947266e-02 -1.80117153e-02 6.92239404e-02
6.46911085e-01 6.08967900e-01 5.66498399e-01 -1.04625046e+00
-2.19538420e-01 -5.92056155e-01 -1.10830389e-01 -9.34047997e-01
-3.08614492e-01 -2.55018443e-01 -1.74411669e-01 -1.72753024e+00
-6.59603477e-02 -3.44315290e-01 1.10828824e-01 2.93407440e-02
1.86026782e-01 -1.64450139e-01 9.08464342e-02 3.54763716e-01
-1.34620011e-01 8.35324675e-02 1.34428287e+00 -2.07537696e-01
-5.65964580e-01 2.47718081e-01 -1.54357404e-01 6.89785659e-01
9.54281449e-01 -5.43848537e-02 -5.82195938e-01 -3.04159880e-01
-3.95597704e-02 1.90670567e-03 -1.92236230e-01 -9.74976361e-01
1.52135968e-01 -5.21970205e-02 8.19148242e-01 -1.30429780e+00
-1.40432924e-01 -7.57836342e-01 -4.19020280e-02 2.21408010e-01
-2.50758410e-01 4.94282752e-01 1.83461383e-01 2.96134472e-01
-1.08615614e-01 -6.95489466e-01 6.51622951e-01 9.38763320e-02
-5.85541666e-01 -4.85495389e-01 -8.84612918e-01 -5.51560283e-01
8.56238484e-01 -8.72676611e-01 -1.82389632e-01 1.00349505e-02
-5.72989583e-01 -2.05748200e-01 6.66390657e-01 1.62432030e-01
5.08758247e-01 -1.04079068e+00 -2.86750138e-01 3.04735482e-01
-2.74202555e-01 -4.17135715e-01 1.95195481e-01 5.81098199e-01
-1.30516469e+00 7.70548344e-01 -4.23061460e-01 -5.04511356e-01
-1.67170417e+00 2.85768360e-01 -1.99857205e-01 -1.14577964e-01
-7.00768292e-01 5.37989065e-02 -1.94749191e-01 3.17406148e-01
2.73580700e-01 -7.74597704e-01 -3.55775386e-01 1.15734868e-01
2.93807596e-01 4.75674093e-01 1.87106922e-01 -1.98418662e-01
-5.84325343e-02 1.10201311e+00 -2.25332141e-01 -1.85882851e-01
8.11084926e-01 -4.03844118e-01 -2.44082063e-01 4.29475188e-01
9.17358816e-01 5.70138454e-01 -7.96694875e-01 1.52924031e-01
3.41248214e-01 -5.35254359e-01 -1.90448195e-01 -4.39395487e-01
-5.04922867e-01 6.35255098e-01 4.68979269e-01 6.58028901e-01
1.24634707e+00 -6.30745411e-01 5.23091137e-01 5.83658457e-01
9.10106897e-02 -1.75207734e+00 -7.66286924e-02 4.57664549e-01
7.22115159e-01 -5.25411904e-01 1.71555355e-01 -5.18779457e-01
-4.02847737e-01 1.73881435e+00 4.57626045e-01 6.14516437e-02
2.76327461e-01 3.47796649e-01 -1.86922610e-01 1.54632956e-01
-1.46194428e-01 1.65023506e-01 2.59405404e-01 2.25053892e-01
6.77269697e-01 -2.03228425e-02 -7.68126786e-01 1.70130953e-01
-1.03668407e-01 -1.59414813e-01 8.97487164e-01 1.18302286e+00
-9.29572642e-01 -1.24357772e+00 -1.02228332e+00 2.86194652e-01
-3.84455293e-01 2.55013436e-01 -6.14282072e-01 1.02372777e+00
-1.48788258e-01 7.73417652e-01 1.90306216e-01 -6.13275170e-02
2.84342110e-01 6.87486529e-02 6.81381822e-01 -6.18403852e-02
-1.11273423e-01 6.56489849e-01 9.38646942e-02 1.25933692e-01
-2.32147112e-01 -6.38829291e-01 -1.44091690e+00 -2.60177165e-01
-3.95377129e-01 4.01203483e-01 8.60576570e-01 6.07513189e-01
2.26450469e-02 4.24164593e-01 6.96031511e-01 -2.39694998e-01
-2.02009052e-01 -6.80490911e-01 -9.00672317e-01 -2.03511789e-02
-1.92842960e-01 -2.74259001e-01 3.37781059e-03 4.56343293e-01]
|
[11.851146697998047, 2.58912992477417]
|
b527ea7a-22a1-45fb-9501-4c4eb0a85b1c
|
optimized-deep-encoder-decoder-methods-for
|
2008.06266
| null |
https://arxiv.org/abs/2008.06266v2
|
https://arxiv.org/pdf/2008.06266v2.pdf
|
Optimized Deep Encoder-Decoder Methods for Crack Segmentation
|
Surface crack segmentation poses a challenging computer vision task as background, shape, colour and size of cracks vary. In this work we propose optimized deep encoder-decoder methods consisting of a combination of techniques which yield an increase in crack segmentation performance. Specifically we propose a decoder-part for an encoder-decoder based deep learning architecture for semantic segmentation and study its components to achieve increased performance. We also examine the use of different encoder strategies and introduce a data augmentation policy to increase the amount of available training data. The performance evaluation of our method is carried out on four publicly available crack segmentation datasets. Additionally, we introduce two techniques into the field of surface crack segmentation, previously not used there: Generating results using test-time-augmentation and performing a statistical result analysis over multiple training runs. The former approach generally yields increased performance results, whereas the latter allows for more reproducible and better representability of a methods results. Using those aforementioned strategies with our proposed encoder-decoder architecture we are able to achieve new state of the art results in all datasets.
|
['Jacob König', 'Mark Jenkins', 'Peter Barrie', 'Mike Mannion', 'Gordon Morison']
|
2020-08-14
| null | null | null | null |
['crack-segmentation']
|
['computer-vision']
|
[ 5.19636393e-01 5.55348545e-02 4.14258718e-01 -2.94536531e-01
-1.08369899e+00 -2.62256265e-01 4.54822481e-01 4.57569808e-01
-6.97629273e-01 4.67559665e-01 -2.20177874e-01 -7.17365444e-02
2.44559512e-01 -8.70148242e-01 -8.54139268e-01 -7.06581652e-01
1.26321822e-01 6.51241899e-01 6.77568078e-01 -2.28655726e-01
6.66667819e-01 4.84621525e-01 -1.65946567e+00 3.27898264e-01
6.93140864e-01 8.65290761e-01 3.26969862e-01 9.30768609e-01
3.74076096e-03 4.47655797e-01 -6.18249536e-01 -2.07904249e-01
1.65536016e-01 -3.29424948e-01 -8.53992760e-01 1.77889943e-01
3.20914239e-01 -9.64478478e-02 1.71003088e-01 7.09109128e-01
6.90266311e-01 -1.10014625e-01 7.83550322e-01 -6.00335062e-01
-2.05497563e-01 5.53278208e-01 -5.14170825e-01 1.64023474e-01
2.59218246e-01 1.52979463e-01 8.42694104e-01 -7.03966200e-01
6.77901566e-01 8.21553111e-01 7.26968646e-01 5.18006861e-01
-1.36716127e+00 -6.39107451e-02 -3.22170705e-01 1.04920834e-01
-1.11480987e+00 -2.95700133e-01 9.98108327e-01 -7.03584135e-01
9.34359610e-01 -2.81419195e-02 4.59200799e-01 8.75754237e-01
-3.60157192e-02 7.32643306e-01 1.15265918e+00 -8.22750986e-01
4.25031811e-01 3.88696194e-02 2.86138564e-01 7.64679968e-01
2.17895031e-01 -1.05226874e-01 -1.17198713e-01 2.79537708e-01
7.44120896e-01 -3.74872923e-01 -1.64446920e-01 -1.78938583e-01
-9.20012414e-01 8.31971407e-01 3.33636880e-01 7.83438861e-01
-2.37188905e-01 4.29664642e-01 6.00309372e-01 1.87905818e-01
4.18517739e-01 4.52912331e-01 -3.65736067e-01 -3.14296447e-02
-1.27784526e+00 4.14885879e-01 7.25907743e-01 4.42283660e-01
8.92021120e-01 4.94874232e-02 -9.69107449e-03 9.13029492e-01
3.07861328e-01 1.52117878e-01 4.42875594e-01 -9.07530844e-01
5.27660072e-01 6.18777215e-01 -2.17787266e-01 -8.07072818e-01
-5.44976652e-01 -3.30987364e-01 -2.97680467e-01 5.54413795e-01
5.97184718e-01 -2.78553218e-01 -1.27436423e+00 1.23759997e+00
-1.98866241e-02 -4.86876667e-02 1.04228713e-01 6.91126525e-01
5.71368039e-01 4.73397553e-01 -1.85042359e-02 9.91615877e-02
1.23336470e+00 -7.99662054e-01 -5.85476577e-01 -7.39380866e-02
6.89725339e-01 -7.87268400e-01 1.14530456e+00 6.86720669e-01
-9.81884062e-01 -5.03805876e-01 -1.38337862e+00 -1.85426027e-02
-5.26301682e-01 5.04194498e-01 2.10131332e-01 8.27112436e-01
-1.05990887e+00 8.70315254e-01 -9.65990126e-01 -1.99887410e-01
4.97907639e-01 5.22920310e-01 -2.61606306e-01 9.25127491e-02
-7.72064328e-01 8.40158343e-01 5.34248173e-01 1.48548573e-01
-7.88839281e-01 -1.30889520e-01 -8.65827858e-01 -1.48673460e-01
2.27711633e-01 -3.89218748e-01 1.10185838e+00 -7.60258198e-01
-1.44747472e+00 1.12831128e+00 1.09669946e-01 -6.08943641e-01
5.62185585e-01 -3.27831686e-01 1.60860226e-01 2.58966148e-01
-7.57640153e-02 7.58837044e-01 8.46772730e-01 -1.67305589e+00
-3.28862041e-01 -3.41381878e-01 -1.71945035e-01 -3.07566643e-01
-1.31401598e-01 -2.72940934e-01 -5.13466954e-01 -6.74597383e-01
1.61184762e-02 -7.48388290e-01 -3.25472772e-01 -3.86588722e-01
-4.75708514e-01 -4.11279462e-02 8.42400312e-01 -7.88525879e-01
1.18342829e+00 -2.00609684e+00 2.99723983e-01 1.52824655e-01
4.45841923e-02 3.53651911e-01 7.79650584e-02 5.96615791e-01
-5.06181121e-02 2.03885168e-01 -1.00802863e+00 -6.80744052e-01
-2.44217724e-01 3.43069911e-01 3.37159425e-01 4.09642309e-01
4.28442091e-01 4.96142149e-01 -2.71198601e-01 -5.67910552e-01
2.04970166e-01 3.19612503e-01 -6.38417423e-01 2.01697007e-01
-5.37951350e-01 4.21750575e-01 -1.54214561e-01 6.67430997e-01
6.31541550e-01 9.61826220e-02 -3.05623233e-01 -8.89964867e-03
-2.71445543e-01 -1.48992285e-01 -1.07754457e+00 2.03450012e+00
-5.26501417e-01 7.77619004e-01 -4.58043143e-02 -1.15596247e+00
1.11440063e+00 3.38157654e-01 3.49193126e-01 -6.40161753e-01
6.60697877e-01 5.61743915e-01 -2.84709558e-02 -7.94444084e-01
5.00999987e-01 -1.97056770e-01 1.59205347e-01 2.73834080e-01
1.98976379e-02 -1.83462262e-01 5.32836676e-01 -2.10078180e-01
9.83618617e-01 1.76943481e-01 -3.25637251e-01 -3.66134644e-01
7.21432745e-01 2.72253573e-01 9.78639070e-03 5.17103016e-01
2.90714763e-02 1.13010037e+00 6.95345104e-01 -2.52092451e-01
-1.24059606e+00 -7.03907192e-01 -1.82462499e-01 6.32180095e-01
2.30746157e-02 -2.69216727e-02 -1.14248741e+00 -5.61083853e-01
-2.82981426e-01 5.88582993e-01 -8.10609519e-01 1.13564752e-01
-8.72741044e-01 -9.61750567e-01 6.85964823e-01 6.23315871e-01
4.24345553e-01 -1.20065880e+00 -1.10853159e+00 2.42172480e-01
8.60455260e-02 -1.06595743e+00 1.57176360e-01 4.52604055e-01
-9.48022902e-01 -1.19581211e+00 -9.45050657e-01 -8.52921188e-01
3.94313425e-01 -4.82885838e-01 9.32594359e-01 4.04565096e-01
-6.74583972e-01 1.83544055e-01 -7.06813335e-01 -2.34222487e-01
-5.76420426e-01 3.75376970e-01 -7.38241136e-01 1.32869743e-02
-1.53105751e-01 -3.22784066e-01 -6.81448281e-01 -1.46783516e-01
-1.14388490e+00 -1.16231889e-01 6.01441622e-01 5.74148297e-01
3.69120091e-01 -3.56328040e-02 5.08496225e-01 -1.09290874e+00
6.97937965e-01 -3.31041634e-01 -5.18797040e-01 -1.30251020e-01
-8.14929605e-01 3.72542441e-01 3.54729384e-01 -2.23607235e-02
-9.46817577e-01 2.37343684e-01 -8.85907590e-01 -1.42598853e-01
-3.95462096e-01 5.56309700e-01 3.87450606e-02 2.63576787e-02
7.84289718e-01 3.37773189e-02 9.98543054e-02 -7.90510297e-01
1.50711745e-01 6.50331497e-01 4.34929550e-01 -6.49026394e-01
4.71813858e-01 4.18732345e-01 6.04108274e-02 -8.02474320e-01
-3.33713979e-01 -4.58710492e-01 -9.01093781e-01 -3.91897857e-01
1.32949460e+00 -4.67907369e-01 -3.16178411e-01 8.39234889e-01
-1.30843472e+00 -4.80562598e-01 -2.18902379e-01 1.34044111e-01
-6.10585392e-01 4.98874754e-01 -6.81866109e-01 -7.76781142e-01
-3.51480097e-01 -1.68766618e+00 1.24187398e+00 2.47913301e-01
1.45104066e-01 -1.02565134e+00 3.43115509e-01 4.70709980e-01
3.24026734e-01 6.93740427e-01 9.82024550e-01 -6.49292409e-01
-4.04114574e-01 -3.30484569e-01 -2.43018806e-01 5.41022837e-01
-1.26211792e-01 1.33564293e-01 -1.01265025e+00 -1.06723599e-01
-2.78473467e-01 -1.97539121e-01 1.26553476e+00 3.41703981e-01
1.05100250e+00 3.58418494e-01 -1.76009297e-01 6.38581142e-02
1.91181493e+00 1.42121285e-01 8.17844272e-01 6.16183817e-01
6.29942954e-01 7.69058287e-01 4.63104308e-01 2.26115659e-01
1.51448309e-01 8.01786840e-01 7.26994634e-01 -2.05320060e-01
-2.75073826e-01 3.05901080e-01 1.52103513e-01 3.86922777e-01
-3.62113982e-01 -3.46355885e-01 -1.23450077e+00 8.46021473e-01
-1.62490714e+00 -5.82280993e-01 -8.12088847e-01 2.08410692e+00
6.72271669e-01 4.17769134e-01 2.75676906e-01 7.23937452e-01
5.30870199e-01 -2.59178579e-02 -1.45823866e-01 -6.89765215e-01
1.34794876e-01 5.68623722e-01 5.41086495e-01 5.31365573e-01
-1.06482828e+00 8.74673367e-01 6.28713179e+00 6.55240834e-01
-1.24184000e+00 1.69303522e-01 5.11043072e-01 3.34562331e-01
-2.51155108e-01 -6.54182360e-02 -4.15729642e-01 3.36309016e-01
8.98272753e-01 6.18381083e-01 -3.72801945e-02 5.95069468e-01
-9.96143650e-03 -4.84621346e-01 -9.47216213e-01 7.78847456e-01
1.22080170e-01 -1.47219837e+00 -3.07768255e-01 -2.54900772e-02
5.06446898e-01 9.66155306e-02 -1.11270428e-01 -4.44407500e-02
-2.65632659e-01 -9.52045977e-01 8.70346725e-01 4.41166788e-01
5.92485845e-01 -6.51627421e-01 9.64538157e-01 1.83798730e-01
-8.65065694e-01 -1.34189576e-01 3.78289334e-02 1.81541234e-01
2.36547127e-01 6.78559661e-01 -7.60749102e-01 7.15916574e-01
5.56581557e-01 5.00721872e-01 -8.08161259e-01 1.06427586e+00
2.03693639e-02 8.16841006e-01 -2.77598858e-01 2.41831943e-01
3.28145564e-01 -1.03653803e-01 3.63459945e-01 1.61395347e+00
1.87116027e-01 -4.56352293e-01 -3.38753723e-02 8.38601053e-01
1.24736093e-01 2.48388901e-01 -4.11087543e-01 1.61086962e-01
-6.24833331e-02 9.92232323e-01 -1.22640216e+00 -2.21279114e-01
-3.15165728e-01 8.94454598e-01 1.97649270e-01 -7.95504525e-02
-7.54431248e-01 -3.43064547e-01 1.37263909e-01 3.49513203e-01
3.74147862e-01 -5.41104794e-01 -7.38981783e-01 -7.81536758e-01
-1.11953944e-01 -6.48815215e-01 2.08495021e-01 -4.10791993e-01
-8.38836551e-01 6.82030618e-01 1.74501792e-01 -8.59167516e-01
-1.84472397e-01 -7.01538384e-01 -7.16639400e-01 6.76868677e-01
-1.34141707e+00 -1.18892717e+00 -2.25678012e-01 2.58604884e-01
1.01084518e+00 -3.40358540e-02 4.17247027e-01 5.57546258e-01
-7.71481037e-01 4.01334047e-01 -1.34569958e-01 1.11642979e-01
3.55261862e-01 -1.54490125e+00 2.52849013e-01 1.05522370e+00
1.43780664e-01 1.28172725e-01 9.13491070e-01 -5.55126011e-01
-9.13840532e-01 -8.02280307e-01 4.85372633e-01 -3.17831576e-01
3.97293895e-01 -2.52011567e-01 -1.12306631e+00 2.98944086e-01
4.10562038e-01 -4.34021503e-01 4.06594008e-01 -1.33618549e-01
-2.22618114e-02 1.70987263e-01 -1.07155097e+00 1.48410529e-01
3.06122720e-01 -1.10229805e-01 -6.33076310e-01 1.10614868e-02
3.66107136e-01 -4.38181818e-01 -8.37073743e-01 4.43975687e-01
3.87779355e-01 -1.39825976e+00 7.27573693e-01 1.06632933e-01
8.81118655e-01 -2.09407717e-01 -7.56993666e-02 -1.01503944e+00
2.64550865e-01 -1.59170642e-01 2.83708632e-01 1.20080841e+00
6.86596096e-01 -4.77456242e-01 1.00120068e+00 2.72801936e-01
-4.50347245e-01 -1.02331090e+00 -9.03045952e-01 -3.27848732e-01
3.59869003e-01 -5.03566146e-01 1.87614843e-01 5.22441745e-01
-4.79922384e-01 8.49521831e-02 -8.03172141e-02 9.99062061e-02
5.08554697e-01 -1.05171790e-03 4.95871693e-01 -1.33270931e+00
-2.15044811e-01 -5.69583714e-01 -3.76416057e-01 -5.64078271e-01
-1.01707019e-01 -6.74866736e-01 2.88752943e-01 -1.68047118e+00
-2.61562634e-02 -4.61344928e-01 -4.99587618e-02 4.80727673e-01
1.07437670e-02 5.71055293e-01 1.50451869e-01 1.07482515e-01
-7.90258646e-02 2.10086748e-01 1.11297715e+00 4.98361187e-03
-1.21320449e-01 5.33711948e-02 -1.19703010e-01 4.90078032e-01
9.08619046e-01 -6.61624014e-01 -1.91748425e-01 -5.45612454e-01
1.02600358e-01 1.49766142e-02 5.17234743e-01 -1.44607317e+00
-1.14819827e-02 4.40458924e-01 1.16301164e-01 -6.80832326e-01
3.02877337e-01 -8.41476977e-01 -9.08891410e-02 8.70975494e-01
-2.96219647e-01 -2.41657607e-02 2.96926528e-01 6.04701340e-01
-3.22296143e-01 -7.86785483e-01 1.11699307e+00 -2.42071763e-01
-5.62453091e-01 -2.03995779e-01 -6.05394840e-01 -1.92305326e-01
1.04651010e+00 -5.01665175e-01 1.04997372e-02 3.43509130e-02
-1.02812886e+00 -4.08756360e-02 5.78215361e-01 1.69485584e-01
4.90017384e-01 -6.52594030e-01 -7.35500038e-01 6.87162727e-02
3.48988362e-03 1.67010337e-01 1.57197848e-01 8.71895730e-01
-1.20770025e+00 9.55462828e-02 -3.09194744e-01 -7.79657602e-01
-1.29748154e+00 2.41891727e-01 4.00858641e-01 -4.37344939e-01
-6.56602740e-01 8.55088651e-01 -4.52859551e-01 -1.91550285e-01
-9.75327748e-06 -5.15432000e-01 -5.14138758e-01 3.15829158e-01
-1.31150022e-01 4.55171436e-01 4.08268064e-01 -5.19536257e-01
-1.93019733e-01 7.31864870e-01 -4.28349674e-02 -2.00858518e-01
1.53407574e+00 1.31417379e-01 -7.59182796e-02 4.31767792e-01
1.26640034e+00 -1.50378168e-01 -1.20724320e+00 3.01653773e-01
2.76363373e-01 -1.06580101e-01 1.63566425e-01 -6.91511691e-01
-1.30673909e+00 1.17953241e+00 1.05736113e+00 4.67541307e-01
1.21181929e+00 1.74187049e-02 9.31179404e-01 -4.86305915e-02
1.05680428e-01 -1.14273345e+00 2.06024393e-01 3.33378196e-01
6.80970728e-01 -1.41123986e+00 -1.63444728e-01 -3.90040815e-01
-5.31388044e-01 1.41185665e+00 3.70339721e-01 -4.09481704e-01
5.38109541e-01 4.45780843e-01 2.19727337e-01 -6.85030878e-01
-3.32481414e-01 -6.17930591e-01 4.42683175e-02 4.26647544e-01
6.04983568e-01 -2.78044313e-01 -7.35630810e-01 3.06361884e-01
1.37580723e-01 -1.25706598e-01 8.02925467e-01 1.08370280e+00
-6.22188449e-01 -1.61058772e+00 -4.25458580e-01 3.47630143e-01
-7.00435042e-01 1.97765455e-01 -4.36028093e-01 1.00320125e+00
3.31017762e-01 9.17819858e-01 -2.03246146e-01 -3.29120576e-01
3.64856392e-01 1.46849558e-01 6.02081656e-01 -8.24675679e-01
-8.49664867e-01 8.20217878e-02 8.05430189e-02 -2.25785315e-01
-5.78295887e-01 -7.36216128e-01 -1.47196543e+00 2.70695418e-01
-5.06361425e-01 4.23571914e-02 1.09955966e+00 1.19872630e+00
4.00663912e-02 8.27034652e-01 4.60682571e-01 -1.07732999e+00
-7.90512413e-02 -1.06165969e+00 -4.30352449e-01 3.69033962e-01
2.71574646e-01 -7.19203711e-01 -1.46783262e-01 5.01157582e-01]
|
[7.486310005187988, 1.5333507061004639]
|
bdce5d0e-eea2-4b93-bcf7-1e9cc44ab427
|
explainable-systematic-analysis-for-synthetic
|
2101.03134
| null |
https://arxiv.org/abs/2101.03134v3
|
https://arxiv.org/pdf/2101.03134v3.pdf
|
Explainable Systematic Analysis for Synthetic Aperture Sonar Imagery
|
In this work, we present an in-depth and systematic analysis using tools such as local interpretable model-agnostic explanations (LIME) (arXiv:1602.04938) and divergence measures to analyze what changes lead to improvement in performance in fine tuned models for synthetic aperture sonar (SAS) data. We examine the sensitivity to factors in the fine tuning process such as class imbalance. Our findings show not only an improvement in seafloor texture classification, but also provide greater insight into what features play critical roles in improving performance as well as a knowledge of the importance of balanced data for fine tuning deep learning models for seafloor classification in SAS imagery.
|
['Alina Zare', 'James Keller', 'Jeff Dale', 'Joshua Peeples', 'Sarah Walker']
|
2021-01-06
| null | null | null | null |
['texture-classification']
|
['computer-vision']
|
[ 9.01822466e-03 2.06843480e-01 2.72426307e-02 -6.11483634e-01
-6.16802394e-01 -6.37657464e-01 5.19192874e-01 1.69668332e-01
-1.40898243e-01 5.38780510e-01 6.88648641e-01 -6.65381014e-01
-9.48064983e-01 -8.22182775e-01 -7.41966724e-01 -7.85045862e-01
-2.25899577e-01 2.52836913e-01 -1.44282803e-01 -4.81620729e-01
5.21127164e-01 5.24923742e-01 -1.32293880e+00 3.47321004e-01
8.30538690e-01 8.46948326e-01 -2.54325330e-01 7.99441278e-01
3.86331826e-01 7.37248361e-01 -5.36933661e-01 -8.72350186e-02
3.14420342e-01 -3.05062294e-01 -7.75392890e-01 -1.43240795e-01
8.30192566e-01 -8.25568140e-02 -2.02615947e-01 7.42770314e-01
4.26712036e-01 1.56272218e-01 8.60709310e-01 -7.53996491e-01
-4.78901982e-01 8.58218014e-01 -3.34368408e-01 8.35381508e-01
-5.11207163e-01 4.15463328e-01 1.13333106e+00 -4.94645625e-01
2.44681150e-01 1.28161633e+00 9.66078043e-01 1.06342278e-01
-9.93450522e-01 -5.22813559e-01 -9.36846733e-02 3.10454816e-02
-1.06591201e+00 -5.94222009e-01 3.01473528e-01 -8.44248056e-01
9.27622855e-01 4.51215118e-01 5.42234540e-01 6.28660858e-01
8.88894796e-01 8.37193578e-02 1.23259640e+00 -4.43877965e-01
2.73445308e-01 3.45420428e-02 3.18271995e-01 7.32717276e-01
6.58453822e-01 4.33177710e-01 -6.29198134e-01 -2.24093765e-01
8.22671592e-01 -5.71040750e-01 -2.61207610e-01 -1.33519694e-01
-1.03361106e+00 1.20027912e+00 5.93139231e-01 -3.20605338e-01
-1.59743905e-01 3.00043106e-01 2.62386352e-01 3.12261194e-01
8.03878725e-01 1.39292097e+00 -8.55262816e-01 1.41041309e-01
-8.43490481e-01 3.77358437e-01 6.39999747e-01 -7.32175335e-02
5.67759037e-01 5.96458018e-01 2.50150949e-01 7.05487132e-01
3.19392085e-01 7.31078565e-01 3.49607974e-01 -1.32068944e+00
2.26511478e-01 8.75922963e-02 6.40018880e-02 -1.39011478e+00
-6.45631790e-01 -1.04429018e+00 -4.93254602e-01 3.82961869e-01
1.25176981e-01 -3.68268847e-01 -9.04234588e-01 1.40697157e+00
-1.57469928e-01 -6.82619736e-02 1.90660790e-01 9.51776862e-01
8.69260073e-01 5.58436632e-01 -3.67068201e-02 3.43952626e-01
1.13775706e+00 -9.28441286e-01 -1.48332492e-01 -6.84000492e-01
6.23456776e-01 -8.29643011e-01 1.00775635e+00 4.47969437e-01
-5.48182368e-01 -4.24078763e-01 -1.31663692e+00 3.30313325e-01
-1.46233916e-01 -8.12732503e-02 9.02123332e-01 6.39007509e-01
-7.52361417e-01 9.04445171e-01 -9.34165716e-01 -3.76088142e-01
3.63825113e-01 2.26419300e-01 -9.92796794e-02 2.36641437e-01
-1.09492707e+00 1.04991829e+00 1.59539625e-01 8.50835890e-02
-9.71023917e-01 -8.29672337e-01 -8.62348020e-01 1.85336173e-01
-1.61407632e-04 -7.18346536e-01 1.11962605e+00 -1.14982665e+00
-9.60361660e-01 6.04671121e-01 1.58542052e-01 -8.01791131e-01
1.47613034e-01 -1.11198135e-01 -2.67882943e-01 3.34316455e-02
-1.74066141e-01 5.44982791e-01 7.77098060e-01 -1.17880094e+00
-6.24625862e-01 -3.73133808e-01 2.59898841e-01 1.62220091e-01
-1.18115626e-01 -1.36473000e-01 4.73956913e-01 -7.25027144e-01
1.99143022e-01 -8.98559332e-01 -3.01923901e-01 -1.72850087e-01
1.92742632e-03 3.26281190e-01 3.10904831e-01 -8.06693077e-01
1.09276938e+00 -1.95018196e+00 -2.19229877e-01 3.29623580e-01
3.65728825e-01 8.21381584e-02 -2.18581453e-01 2.43405759e-01
-1.69317365e-01 4.32431698e-01 -1.19798116e-01 2.92504251e-01
-1.30987033e-01 2.62643367e-01 -5.15465379e-01 2.89153814e-01
4.22281861e-01 5.28054714e-01 -5.90677202e-01 8.74516815e-02
2.43118256e-01 1.95536822e-01 -7.38159537e-01 -2.75184900e-01
-7.24049583e-02 1.99113384e-01 -4.55048561e-01 5.61918616e-01
6.71103299e-01 -3.36519897e-01 -7.27867112e-02 -1.38581619e-01
-2.50311702e-01 5.12101471e-01 -8.17583501e-01 6.67703211e-01
-6.12335861e-01 1.15339267e+00 -7.86552299e-03 -7.23583341e-01
1.01648235e+00 -1.87557191e-01 -1.98614657e-01 -5.62939823e-01
-4.00761291e-02 2.66019404e-01 5.67403316e-01 -5.18376231e-01
4.67895180e-01 -3.75439823e-01 1.58532697e-04 3.65990013e-01
-2.17842132e-01 -3.76413703e-01 -1.84048012e-01 -1.06735714e-01
8.24468672e-01 3.16209644e-02 4.59640712e-01 -1.07915330e+00
-4.63799350e-02 6.13474071e-01 4.96271789e-01 1.05690300e+00
7.60414526e-02 6.60479248e-01 4.42870975e-01 -7.76551425e-01
-1.12822866e+00 -7.67725945e-01 -4.97675776e-01 8.34064007e-01
-1.31015465e-01 -4.34743434e-01 -4.99104738e-01 -2.56311834e-01
2.42177308e-01 8.82897496e-01 -1.13166690e+00 -5.46304941e-01
-2.13064268e-01 -1.40848029e+00 6.68173313e-01 7.16523767e-01
3.20969492e-01 -7.12883949e-01 -9.15535271e-01 -1.81519210e-01
-1.18220896e-01 -7.10851967e-01 2.61990577e-01 6.34155273e-01
-1.19326341e+00 -9.17251348e-01 -1.42772064e-01 -8.93632099e-02
5.33887208e-01 3.11578333e-01 1.24709451e+00 9.10566822e-02
-3.47324878e-01 1.57729149e-01 -3.81421894e-01 -6.58814609e-01
-5.16280890e-01 1.65310696e-01 1.11878872e-01 -5.73789477e-01
2.59338498e-01 -2.28960007e-01 -5.93702912e-01 5.53334832e-01
-7.85695314e-01 -6.58814237e-02 5.52434325e-01 9.15597618e-01
3.70064713e-02 3.12295526e-01 3.08614343e-01 -7.72367835e-01
5.38264155e-01 -5.90878785e-01 -5.80341041e-01 3.44460011e-02
-8.73110354e-01 4.18991089e-01 3.94693524e-01 7.65552968e-02
-1.14568329e+00 -6.42363012e-01 2.16249786e-02 -6.66351318e-02
-1.72637496e-03 1.08198488e+00 5.05295157e-01 -3.49311113e-01
1.28410685e+00 -3.98866475e-01 -6.58158734e-02 -3.99231613e-01
-3.03625017e-01 4.57153171e-01 1.87589854e-01 -5.79021215e-01
7.22593009e-01 4.70565468e-01 8.62407535e-02 -8.91247213e-01
-1.45037985e+00 -1.18648624e-02 -4.68773514e-01 1.69479996e-02
5.73355377e-01 -1.00642300e+00 -8.25471729e-02 3.24521363e-01
-3.90168488e-01 -8.78461719e-01 -1.36877775e-01 6.80806220e-01
-1.81919351e-01 -1.54517040e-01 -1.50259510e-01 -5.22699296e-01
-1.45994604e-01 -1.02702677e+00 7.79755175e-01 4.38650817e-01
-3.98028791e-01 -1.34835517e+00 1.95696890e-01 7.04490483e-01
5.82322299e-01 2.19125867e-01 1.18324804e+00 -6.18474364e-01
-3.95776421e-01 2.61935800e-01 -1.55826256e-01 4.90097046e-01
8.42273422e-03 2.85834908e-01 -1.15197003e+00 -1.50432289e-01
-1.96301132e-01 -2.77891994e-01 1.35542893e+00 8.44607055e-01
9.22378957e-01 -2.01721802e-01 -8.02314878e-02 9.59632397e-01
1.31283438e+00 -2.64349431e-02 4.84526694e-01 7.85588145e-01
3.28598350e-01 8.24205279e-01 8.56682241e-01 5.09006262e-01
7.66282454e-02 2.11034715e-01 4.14462686e-01 -1.90626919e-01
-8.32410380e-02 1.33822737e-02 1.94098111e-02 4.06830400e-01
2.50762124e-02 -3.42740536e-01 -1.36778450e+00 5.53795159e-01
-1.60129714e+00 -6.47228658e-01 -2.76506662e-01 1.89747870e+00
3.07298899e-01 2.45581970e-01 -4.98571187e-01 -2.26247132e-01
1.48553029e-01 4.89792883e-01 -3.66016537e-01 -7.41079509e-01
-4.25227970e-01 3.61169696e-01 8.52551758e-01 7.84334183e-01
-1.06654859e+00 9.11533773e-01 7.20714426e+00 5.91480672e-01
-1.35796833e+00 -3.19492877e-01 8.36944163e-01 -1.87111035e-01
-4.95356321e-01 2.40131542e-01 -5.61123133e-01 -1.38981538e-02
1.00812960e+00 2.29181886e-01 2.33328134e-01 6.64216697e-01
6.46726370e-01 -2.77640134e-01 -5.32333910e-01 3.81601959e-01
1.72750905e-01 -1.78197587e+00 8.64226893e-02 1.46254569e-01
9.08808708e-01 4.05352682e-01 2.32074305e-01 5.70294708e-02
3.16084266e-01 -1.14931846e+00 7.33090937e-01 5.21313548e-01
3.42830420e-01 -8.56113434e-01 1.20767069e+00 -1.22400619e-01
-3.79153013e-01 -3.12593907e-01 -5.95947027e-01 -5.13384342e-01
-4.72191870e-01 7.33614862e-01 -1.01819646e+00 3.21459353e-01
1.12057090e+00 4.30919439e-01 -9.47683752e-01 9.45309520e-01
-6.57420084e-02 1.35921061e+00 -2.54166722e-01 2.87029773e-01
4.02176976e-01 1.64825216e-01 6.01961553e-01 9.76332188e-01
3.11676145e-01 2.12759107e-01 -3.65233034e-01 5.99468827e-01
5.43510616e-01 -1.59620166e-01 -3.85038584e-01 -3.47589523e-01
3.33424211e-01 1.11054516e+00 -7.92965293e-01 -3.21658403e-02
6.20062347e-05 1.05971150e-01 -1.16347797e-01 2.25384921e-01
-5.70390701e-01 -1.63525060e-01 9.21157181e-01 3.89585108e-01
3.73039246e-01 -2.88442224e-01 -7.62231112e-01 -9.97512698e-01
-5.57958245e-01 -1.17736530e+00 5.00797153e-01 -1.22303176e+00
-9.85150635e-01 2.82058686e-01 -1.11496653e-02 -9.26287889e-01
3.32890800e-03 -8.96564782e-01 -8.30620468e-01 6.86493039e-01
-1.42903686e+00 -1.04732597e+00 -5.73645294e-01 -2.34736398e-01
3.43232155e-01 -2.62593687e-01 8.00314724e-01 -2.13707447e-01
-3.16226244e-01 2.71087646e-01 5.14000475e-01 -1.60634533e-01
9.55026209e-01 -1.13752687e+00 5.41912317e-01 7.61974633e-01
-7.87574239e-03 7.91713774e-01 1.15608633e+00 -6.51749909e-01
-9.51461673e-01 -7.70610631e-01 3.87317657e-01 -5.80437124e-01
8.22378337e-01 1.10183783e-01 -7.76029170e-01 7.72613883e-01
1.05235234e-01 -2.62852967e-01 9.27371144e-01 4.89853859e-01
-1.44378245e-01 -1.86679706e-01 -8.22794914e-01 3.84825826e-01
5.97323775e-01 1.43619373e-01 -7.38673270e-01 -3.95150967e-02
6.03349328e-01 -3.74846578e-01 -8.57873261e-01 7.66336620e-01
6.06746614e-01 -1.12577736e+00 7.33922243e-01 -8.68944347e-01
8.31592977e-01 -1.55722901e-01 -5.06651282e-01 -1.53609788e+00
-5.15503526e-01 -2.74314225e-01 6.38754964e-01 6.57561600e-01
7.41614819e-01 -8.46187115e-01 6.14287555e-01 1.21901341e-01
-4.50778335e-01 -7.52114773e-01 -3.83279502e-01 -2.86623836e-01
3.79185587e-01 -3.64618450e-01 4.63806450e-01 1.09496224e+00
-5.62121749e-01 1.08546637e-01 -1.40710399e-01 4.84685421e-01
5.67560852e-01 2.41799459e-01 6.96895182e-01 -1.49194467e+00
-3.85697871e-01 -2.74388492e-01 -3.36308599e-01 -3.00855607e-01
-1.65750921e-01 -7.18877196e-01 -4.66793962e-02 -1.34396362e+00
2.20991567e-01 -4.64827418e-01 -3.66982251e-01 4.34095621e-01
-1.09766200e-01 3.66592616e-01 -7.53537118e-02 1.94891512e-01
-1.83528230e-01 2.54192650e-01 9.57742631e-01 1.67768165e-01
6.80505112e-02 1.97515693e-02 -1.09471214e+00 1.04079843e+00
8.87850761e-01 -6.59136593e-01 -1.95084542e-01 -7.85978377e-01
8.06369841e-01 -1.54153809e-01 8.29964519e-01 -9.47144210e-01
-3.13495398e-01 -4.10602808e-01 7.32474923e-01 -2.80221671e-01
-3.45009714e-02 -1.66108802e-01 -1.76135391e-01 5.16069591e-01
-5.17192364e-01 -3.58085595e-02 7.86041796e-01 3.68719578e-01
-1.60589069e-01 -3.61358196e-01 9.26384628e-01 -1.44842893e-01
-8.04082155e-01 -3.45483601e-01 -5.50245941e-01 9.30749103e-02
4.30472225e-01 -1.58935443e-01 -7.20857620e-01 -4.32504654e-01
-8.31298649e-01 3.19968969e-01 4.64134037e-01 5.25076210e-01
1.78141937e-01 -7.75811195e-01 -9.45826173e-01 2.25654304e-01
9.54028070e-02 -2.72105873e-01 4.42774236e-01 9.00153279e-01
-9.83360469e-01 5.90277970e-01 -3.45449060e-01 -5.74930429e-01
-1.14103580e+00 -3.86970133e-01 8.68868649e-01 5.44850640e-02
-3.26272011e-01 1.14036214e+00 3.89227599e-01 -7.21874118e-01
-3.39808881e-01 -4.94382977e-01 -1.53117776e-02 1.75336733e-01
1.84428275e-01 4.64585066e-01 1.76010355e-01 -2.06025362e-01
-4.96836185e-01 5.56471109e-01 1.58734798e-01 -6.06296770e-02
1.58302128e+00 -1.16699979e-01 -1.38066327e-02 5.17951310e-01
7.50533342e-01 1.87952533e-01 -1.42034483e+00 5.73724657e-02
-3.17526162e-01 -5.36326528e-01 6.79479539e-01 -1.24684989e+00
-8.58836710e-01 9.53166902e-01 6.45405650e-01 4.67036059e-03
6.48907185e-01 -3.81216928e-02 2.53945857e-01 5.82084239e-01
-2.35288501e-01 -9.47840095e-01 5.78897595e-02 6.66584373e-01
9.20677483e-01 -1.33356273e+00 4.79481727e-01 2.92043611e-02
-8.00634384e-01 1.40068579e+00 6.16240323e-01 -2.23395407e-01
5.51218927e-01 3.69603664e-01 4.47035134e-01 -5.71151614e-01
-9.09205437e-01 2.22032234e-01 4.57861066e-01 3.41870457e-01
2.90618628e-01 1.07767075e-01 2.01653272e-01 5.55009425e-01
-7.50454187e-01 -4.16939646e-01 7.82396853e-01 5.53332686e-01
-8.04865241e-01 -3.28085184e-01 -6.60360873e-01 7.85054922e-01
-3.43484730e-01 -2.96921402e-01 -5.48130453e-01 1.09168065e+00
-2.18173824e-02 8.32630277e-01 5.41706860e-01 -2.32393786e-01
-5.31955771e-02 -1.10673875e-01 3.44293892e-01 -6.52808249e-01
-5.76717794e-01 -2.71824971e-02 5.93461514e-01 -1.46200404e-01
-1.81112617e-01 -6.79224491e-01 -1.08094597e+00 -2.05352679e-01
-5.62532842e-01 2.90150285e-01 6.81631565e-01 1.08422041e+00
4.00469452e-01 8.26856911e-01 3.43604654e-01 -7.07971692e-01
-6.92917585e-01 -1.04472411e+00 -5.75933099e-01 -2.26493254e-01
5.32112360e-01 -6.59336388e-01 -1.01798725e+00 -1.56579360e-01]
|
[6.804311752319336, 2.8775060176849365]
|
decc5c6e-75d4-4990-93ed-585cfceeeddd
|
natural-language-processing-for-policymaking
|
2302.03490
| null |
https://arxiv.org/abs/2302.03490v1
|
https://arxiv.org/pdf/2302.03490v1.pdf
|
Natural Language Processing for Policymaking
|
Language is the medium for many political activities, from campaigns to news reports. Natural language processing (NLP) uses computational tools to parse text into key information that is needed for policymaking. In this chapter, we introduce common methods of NLP, including text classification, topic modeling, event extraction, and text scaling. We then overview how these methods can be used for policymaking through four major applications including data collection for evidence-based policymaking, interpretation of political decisions, policy communication, and investigation of policy effects. Finally, we highlight some potential limitations and ethical concerns when using NLP for policymaking. This text is from Chapter 7 (pages 141-162) of the Handbook of Computational Social Science for Policy (2023). Open Access on Springer: https://doi.org/10.1007/978-3-031-16624-2
|
['Rada Mihalcea', 'Zhijing Jin']
|
2023-02-07
| null | null | null | null |
['event-extraction']
|
['natural-language-processing']
|
[ 1.80420995e-01 4.82500643e-01 -1.16625929e+00 -2.85746515e-01
-1.05190575e+00 -8.03886771e-01 1.07529747e+00 9.91862178e-01
-6.71416402e-01 9.92172480e-01 1.42070413e+00 -1.54495800e+00
-1.32321283e-01 -6.97668016e-01 -3.91723543e-01 -3.88196111e-01
3.43166083e-01 3.00558716e-01 -2.33780339e-01 1.44185960e-01
5.62513769e-01 3.75423610e-01 -9.41131175e-01 2.76528537e-01
9.94865537e-01 4.43554282e-01 -1.17568433e-01 4.93779778e-01
-6.82081938e-01 1.07433593e+00 -7.43542790e-01 -2.77671635e-01
-6.99292645e-02 -4.00951505e-01 -7.50510097e-01 -5.55080175e-01
1.63414389e-01 -3.83464508e-02 -3.42886671e-02 9.50709581e-01
4.91289794e-01 1.97492596e-02 6.00339651e-01 -6.53012037e-01
-4.69553977e-01 1.06038547e+00 -3.46510768e-01 5.13594031e-01
4.99850899e-01 1.58367790e-02 8.20974112e-01 -6.76242352e-01
9.10722077e-01 1.46273041e+00 3.90440971e-01 1.56836778e-01
-1.06789863e+00 -8.24435949e-01 3.65889996e-01 4.97032516e-02
-6.32879257e-01 -7.46663094e-01 3.53756309e-01 -9.72911656e-01
9.67875838e-01 4.33111101e-01 7.46341288e-01 1.13759458e+00
4.74955440e-01 2.01435298e-01 1.75214374e+00 -7.42240548e-01
4.17685360e-01 1.08114086e-01 3.09663534e-01 4.93530780e-02
5.54730356e-01 -1.28828585e-01 -1.70149073e-01 -8.18378985e-01
2.74029404e-01 -1.47379190e-01 3.20519447e-01 7.89683580e-01
-1.23990667e+00 1.23969412e+00 -2.37875059e-01 4.75393534e-01
-7.43293226e-01 3.80252711e-02 4.17199075e-01 1.88847035e-01
9.47712541e-01 2.57569432e-01 -3.34869504e-01 -6.51135743e-01
-6.35898590e-01 3.58260691e-01 1.10808170e+00 1.88458666e-01
-8.03082716e-03 -4.37611401e-01 -4.75683630e-01 6.72170639e-01
6.81913853e-01 8.77099633e-01 1.98750105e-02 -1.01634073e+00
5.04939675e-01 4.49448347e-01 3.70646298e-01 -1.22339559e+00
-4.49070752e-01 1.05495580e-01 -4.72501785e-01 -2.71880090e-01
4.79243696e-01 -5.60210824e-01 -6.71216846e-01 1.41408181e+00
5.58766723e-01 -4.46938127e-01 5.21709621e-02 2.18119234e-01
1.11586189e+00 9.62842226e-01 8.95012200e-01 -8.63581240e-01
1.92080450e+00 -3.57687682e-01 -1.17432249e+00 -5.03521487e-02
3.43182027e-01 -1.16201293e+00 8.16614091e-01 1.53318970e-02
-1.01735723e+00 3.01674366e-01 -6.88592494e-02 -2.22730801e-01
-6.23947382e-01 -9.34007540e-02 5.70558190e-01 5.91059685e-01
-5.55101931e-01 -3.16363201e-02 -6.89152658e-01 -6.89808190e-01
4.98855412e-01 -2.46269837e-01 2.86549062e-01 4.57966685e-01
-1.45647717e+00 1.11015916e+00 9.60304514e-02 -3.21481854e-01
-2.57303536e-01 -6.77506387e-01 -6.58717573e-01 -2.98025161e-01
5.46585143e-01 -4.58391756e-01 1.44187915e+00 -1.74025685e-01
-1.37382114e+00 9.86040831e-01 -4.36001480e-01 -2.74451643e-01
3.00246149e-01 -1.05191298e-01 -6.64810419e-01 5.79780079e-02
5.70061564e-01 3.22721377e-02 2.74893641e-01 -6.56441867e-01
-9.01166022e-01 -1.13712415e-01 1.06970519e-01 1.12575524e-01
-4.78894748e-02 1.18121445e+00 3.68456185e-01 -6.46180451e-01
-2.53304124e-01 -3.52794141e-01 -5.32013237e-01 -4.89188254e-01
-3.35576802e-01 -6.30663335e-01 2.73101717e-01 -1.15337133e+00
1.55111575e+00 -1.65346289e+00 -6.09259307e-01 1.30651832e-01
4.36150469e-02 -4.75013144e-02 4.04787123e-01 1.12597287e+00
-4.55760323e-02 8.36468399e-01 9.55216959e-02 2.97313869e-01
7.90820941e-02 9.23226327e-02 -6.71705186e-01 6.64651871e-01
-3.48310679e-01 9.57766891e-01 -1.03734767e+00 -3.74839067e-01
3.28969359e-01 2.71682739e-01 -4.52316478e-02 -7.52517939e-01
-1.75204739e-01 6.25885785e-01 -9.96001840e-01 4.95177329e-01
1.82911709e-01 -2.72697601e-02 4.77007002e-01 1.21030755e-01
-1.03271139e+00 1.04546440e+00 -7.10379720e-01 9.31734383e-01
-3.02196920e-01 8.87808323e-01 3.53654861e-01 -9.24601734e-01
4.45271015e-01 5.25128841e-01 6.98188066e-01 -6.14326000e-01
4.53288853e-01 1.44585231e-02 -8.28297585e-02 -6.31088197e-01
2.05537677e-01 -1.77004829e-01 -4.84885573e-01 9.61424232e-01
-6.63414955e-01 -1.00126818e-01 1.82016388e-01 1.34560630e-01
8.51914287e-01 -4.02172655e-01 1.17105925e+00 -6.52490020e-01
2.78649300e-01 5.62406838e-01 5.78901172e-01 7.88363039e-01
-1.79925382e-01 -5.90981662e-01 5.33586264e-01 -5.87543905e-01
-8.52931380e-01 -3.57455373e-01 -6.70421004e-01 1.11859906e+00
-6.04341567e-01 -6.21832967e-01 -3.93999279e-01 -1.19284995e-01
7.20054582e-02 1.24673808e+00 -5.60317993e-01 6.45666778e-01
-3.83156896e-01 -7.14442611e-01 1.68224722e-01 -2.07280248e-01
1.88873097e-01 -1.01380610e+00 -6.64331794e-01 3.35774750e-01
-4.09184724e-01 -9.55808520e-01 -6.58025965e-04 -2.73722857e-01
-5.99475205e-01 -1.28485274e+00 -3.50775778e-01 -1.74676448e-01
5.70647061e-01 3.60167585e-02 5.83340764e-01 -3.73407900e-01
-7.73231983e-02 7.10400403e-01 -3.27989638e-01 -9.88995790e-01
-1.03258395e+00 -1.42609477e-01 2.24299252e-01 -6.25996351e-01
7.87295282e-01 -3.67404521e-01 -4.10132885e-01 -3.49036604e-01
-6.48293614e-01 3.10955465e-01 3.20401907e-01 3.50118756e-01
3.18364859e-01 -2.50917345e-01 5.33153832e-01 -9.09898698e-01
1.28319871e+00 -6.97841465e-01 -7.74878502e-01 2.68388331e-01
-7.15229392e-01 -6.07599378e-01 2.26336733e-01 -2.45539546e-01
-1.20584095e+00 -8.63665700e-01 -6.90386444e-02 7.71661162e-01
-6.08769596e-01 8.97844195e-01 3.45689088e-01 4.33683127e-01
8.58059347e-01 -3.11825424e-01 8.76795575e-02 -5.89725196e-01
6.87359095e-01 9.44942772e-01 9.48733464e-02 -6.63759530e-01
4.04182464e-01 6.73681080e-01 -5.37389338e-01 -9.17742908e-01
-9.92471874e-01 -3.23377937e-01 -1.59753576e-01 -4.28073943e-01
9.33427513e-01 -8.81951869e-01 -8.25055420e-01 -1.41995028e-01
-9.66594160e-01 -4.74749386e-01 -5.96119523e-01 9.31338608e-01
-1.17878258e-01 -6.34723809e-03 -5.55210233e-01 -1.08947647e+00
-3.42801243e-01 -8.48149121e-01 4.39928442e-01 2.72413850e-01
-5.89125752e-01 -1.27165723e+00 2.94036210e-01 6.89443946e-01
4.70078975e-01 7.24128187e-01 1.02545989e+00 -7.13773549e-01
2.25830287e-01 4.14704159e-02 9.33035314e-02 -3.14781815e-01
1.83460191e-01 3.61282378e-01 -4.68132049e-01 1.56107709e-01
1.95161238e-01 4.62765507e-02 6.97369099e-01 1.06463969e+00
5.55966020e-01 -1.30648315e+00 -5.12559950e-01 -1.68737993e-01
1.05589032e+00 6.13882303e-01 3.04772794e-01 8.34100246e-01
1.38666317e-01 8.97001803e-01 5.49399853e-01 7.10153043e-01
6.88933730e-01 4.28407550e-01 -4.40080136e-01 5.83489798e-02
2.23242819e-01 -3.73432904e-01 4.25325632e-01 5.00666559e-01
-3.36699247e-01 1.28785521e-01 -1.58585739e+00 5.01718640e-01
-1.81170249e+00 -1.16636252e+00 -3.78615260e-01 1.72177827e+00
1.18077362e+00 2.87063904e-02 2.07140103e-01 -3.29359472e-01
6.50517523e-01 3.47158492e-01 -8.62348527e-02 -9.88180339e-01
1.39395863e-01 7.30428323e-02 7.14236915e-01 7.81605124e-01
-1.03214014e+00 1.13727438e+00 6.34374189e+00 6.02504432e-01
-8.82678986e-01 3.17444980e-01 8.43367994e-01 3.23873200e-02
-5.53258538e-01 3.97328645e-01 -7.15790153e-01 3.04225206e-01
1.23727632e+00 -1.01009560e+00 1.65500939e-01 5.18956125e-01
1.29288995e+00 -5.71898341e-01 -1.45724893e-01 3.48388493e-01
-3.29816371e-01 -1.67855394e+00 -2.20276192e-01 2.84797370e-01
8.53815079e-01 2.41889060e-01 -7.77393132e-02 2.18717344e-02
8.99654627e-01 -7.59941757e-01 7.81483769e-01 2.19004944e-01
9.01947558e-01 -1.44659251e-01 4.07152057e-01 3.22738320e-01
-6.22521520e-01 -2.51523495e-01 -4.28423911e-01 -6.11062825e-01
7.27744758e-01 1.19589245e+00 -6.32430077e-01 3.87680739e-01
5.16449511e-01 5.14069319e-01 -6.73805550e-02 8.17213178e-01
-6.42380536e-01 1.18728662e+00 -3.23301017e-01 -5.26272282e-02
4.04413968e-01 -9.06999335e-02 7.97007799e-01 1.03430223e+00
1.23408437e-01 7.46692955e-01 2.17783213e-01 3.08113903e-01
2.55428374e-01 6.02135599e-01 -6.19343579e-01 -5.47202587e-01
9.50448155e-01 9.42585528e-01 -9.89516020e-01 -5.81471086e-01
-4.66031730e-01 -3.41919571e-01 -1.58670917e-01 5.61131716e-01
-4.90661830e-01 2.49775082e-01 4.12769169e-01 3.75251800e-01
-4.63861138e-01 -1.02714144e-01 -2.60637850e-01 -8.94004762e-01
-4.61770177e-01 -1.11574686e+00 6.24024689e-01 -8.32043365e-02
-9.15121436e-01 -8.54032487e-02 6.89779401e-01 -4.88527268e-01
-9.54869688e-02 -2.73754179e-01 -5.95437050e-01 6.64961278e-01
-9.22315121e-01 -9.20700252e-01 3.33367974e-01 -1.20687433e-01
4.16347742e-01 -8.97465367e-03 8.10261309e-01 -4.33048308e-02
-4.52589840e-01 -4.09881085e-01 4.12107319e-01 -2.07446560e-01
6.66146457e-01 -9.26995873e-01 5.12744248e-01 5.76166928e-01
-4.29593891e-01 6.11067355e-01 6.86377466e-01 -1.01549888e+00
-8.33889842e-01 -8.08187127e-01 1.58087528e+00 -2.80917317e-01
1.09419298e+00 -3.28297526e-01 -1.80445537e-01 6.72585785e-01
5.12764215e-01 -9.68071103e-01 1.30882609e+00 3.57990503e-01
3.09156980e-02 3.69988978e-01 -1.13234675e+00 8.64073575e-01
6.17920876e-01 -4.07099068e-01 -1.06887543e+00 9.39369082e-01
4.46011424e-01 -2.41214469e-01 -1.05667126e+00 -6.98505268e-02
8.40535879e-01 -1.55586720e-01 8.25390041e-01 -9.76910114e-01
3.92782211e-01 1.62502646e-01 1.00962728e-01 -8.20518613e-01
-4.92438883e-01 -8.79960895e-01 3.88159543e-01 1.19329989e+00
5.34311235e-01 -1.24091280e+00 7.60827288e-02 9.08975780e-01
2.34445840e-01 -5.47756851e-01 -1.00834250e+00 -1.05785213e-01
3.83480251e-01 -5.97778141e-01 3.34203929e-01 1.47403109e+00
4.99999732e-01 1.41790673e-01 5.70740923e-02 -2.26584882e-01
3.83267790e-01 1.06287822e-01 4.73744810e-01 -1.13187051e+00
3.60575676e-01 -8.00630391e-01 1.74632251e-01 -5.13728440e-01
7.45101199e-02 -6.19036734e-01 -3.84811908e-01 -2.20632005e+00
2.76045382e-01 -2.14181364e-01 2.95424342e-01 5.19063830e-01
-9.34112594e-02 -5.51501572e-01 1.38019368e-01 4.69745964e-01
-6.86175674e-02 1.24415308e-01 1.03570199e+00 6.26596287e-02
-4.89717096e-01 8.92708525e-02 -1.27805293e+00 9.06689763e-01
1.23837793e+00 -8.12422335e-01 7.84629118e-03 -1.62294745e-01
7.64811099e-01 9.72789153e-02 1.23630494e-01 -1.19674265e-01
4.84116167e-01 -1.18105769e+00 -1.17867582e-01 -7.55302072e-01
-3.52709800e-01 -5.28049469e-01 2.59608448e-01 6.89922810e-01
-6.64638340e-01 3.35405648e-01 4.03528780e-01 4.23939914e-01
-3.50535475e-02 -1.71588555e-01 2.02566952e-01 -2.21140906e-01
-2.09855825e-01 -2.41470233e-01 -1.13718808e+00 3.38272363e-01
1.12143815e+00 2.65325218e-01 -9.92382944e-01 -5.43923557e-01
-7.74766803e-01 3.27972531e-01 1.10135704e-01 3.34035635e-01
-5.68865649e-02 -1.11089480e+00 -1.15779662e+00 -5.48195004e-01
-3.44663054e-01 -3.79174501e-01 1.08823575e-01 1.31036484e+00
-4.79168594e-01 1.15881813e+00 3.83324981e-01 1.25442162e-01
-1.19647408e+00 2.20383123e-01 -2.05651402e-01 -3.37264508e-01
-6.44047379e-01 3.21354270e-01 -6.42114803e-02 -1.33538127e-01
1.24904551e-01 -4.28346097e-01 -5.65482199e-01 5.37519693e-01
7.38397539e-01 5.21602452e-01 -7.00189233e-01 -8.02855670e-01
-7.24616885e-01 1.76920772e-01 -1.48232689e-03 -6.88735425e-01
1.32273972e+00 -3.01301569e-01 -6.55269206e-01 5.51888227e-01
5.85874140e-01 6.17600739e-01 -2.75795996e-01 -1.31388560e-01
3.35922420e-01 -2.49935210e-01 1.36473775e-01 -9.85888004e-01
-1.91477969e-01 3.42151910e-01 -3.76229808e-02 5.56522906e-01
6.46921933e-01 2.43319362e-01 9.66028273e-02 3.24315429e-02
-1.09564811e-01 -1.39801228e+00 -7.13141382e-01 3.07381988e-01
1.06919158e+00 -9.17682350e-01 5.39687276e-01 -2.39275366e-01
-3.39340985e-01 1.08089983e+00 -3.73101562e-01 5.97866952e-01
1.18099070e+00 1.54316679e-01 2.46807337e-01 -5.44212878e-01
-8.57856631e-01 1.36753425e-01 1.18777320e-01 2.97419220e-01
6.07509911e-01 5.63372076e-01 -1.53378415e+00 5.60701191e-01
-2.94241875e-01 5.16318865e-02 2.77049839e-01 9.07496870e-01
-5.06690025e-01 -1.17088377e+00 -8.40919554e-01 7.19808400e-01
-1.14031684e+00 -4.30184275e-01 -7.67834902e-01 7.51901150e-01
-6.72387779e-02 1.33096468e+00 8.34556296e-02 3.84766221e-01
3.70809399e-02 4.85984273e-02 -2.27822334e-01 -6.35337174e-01
-7.86168873e-01 2.50872105e-01 8.14931035e-01 -3.89332563e-01
-8.00615191e-01 -1.06891620e+00 -6.54782236e-01 -6.90649331e-01
-8.46107602e-02 5.49737275e-01 1.09219587e+00 7.59570897e-01
3.18845153e-01 3.45760405e-01 1.48574010e-01 -4.24177805e-03
-4.12721336e-01 -9.84816134e-01 -1.78186178e-01 -2.56383032e-01
7.52060190e-02 -4.37064707e-01 -4.30917889e-01 -2.73923855e-02]
|
[9.002686500549316, 9.863202095031738]
|
24860f01-b50f-457d-9656-3fa82856904b
|
more-complete-resultset-retrieval-from-large
| null | null |
https://dl.acm.org/doi/10.1145/3360901.3364436#d2419191e1
|
https://svn.aksw.org/papers/2019/KCAP2019_WIMUQ/public.pdf
|
More Complete Resultset Retrieval from Large Heterogeneous RDF Sources
|
Over the last years, the Web of Data has grown significantly. Various interfaces such as LOD Stats, LOD Laudromat, SPARQL endpoints provide access to the hundered of thousands of RDF datasets, representing billions of facts. These datasets are available in different formats such as raw data dumps and HDT files or directly accessible via SPARQL endpoints. Querying such large amount of distributed data is particularly challenging and many of these datasets cannot be directly queried using the SPARQL query language. In order to tackle these problems, we present WimuQ, an integrated query engine to execute SPARQL queries and retrieve results from large amount of heterogeneous RDF data sources. Presently, WimuQ is able to execute both federated and non-federated SPARQL queries over a total of 668,166 datasets from LOD Stats and LOD Laudromat as well as 559 active SPARQL endpoints. These data sources represent a total of 221.7 billion triples from more than 5 terabytes of information from datasets retrieved using the service "Where is My URI" (WIMU). Our evaluation on state-of-the-art real-data benchmarks shows that WimuQ retrieves more complete results for the benchmark queries.
|
['Andre Valdestilhas', 'Tommaso Soru', 'Muhammad Saleem']
|
2019-11-12
| null | null | null |
acm-10th-international-conference-on
|
['rdf-dataset-discovery']
|
['knowledge-base']
|
[-9.53253925e-01 6.51888624e-02 -2.92612016e-01 -6.95241690e-01
-7.90776432e-01 -7.72978127e-01 5.91730356e-01 8.60913038e-01
-2.58441150e-01 1.15852046e+00 5.33343017e-01 6.29009083e-02
-6.98972702e-01 -1.93261814e+00 -6.93011165e-01 -3.87832080e-03
-2.24722356e-01 1.09033263e+00 1.02990675e+00 -6.90508723e-01
-1.85546637e-01 3.11474532e-01 -2.38732076e+00 8.17882299e-01
9.77959991e-01 1.54374444e+00 -1.10332347e-01 9.60441083e-02
-8.40356469e-01 8.57043505e-01 -6.45486712e-01 -5.28317511e-01
3.45007367e-02 4.06234443e-01 -1.00307417e+00 -1.31037569e+00
5.56390762e-01 -1.97825059e-01 6.51660515e-03 1.00654030e+00
7.30720401e-01 -1.46465257e-01 -3.08423847e-01 -1.25815904e+00
-4.82615620e-01 1.02945852e+00 1.07960686e-01 1.62743777e-01
8.10112357e-01 -4.71480608e-01 9.04334903e-01 -8.37277472e-01
1.40297592e+00 1.29035127e+00 1.79370344e-01 -1.26364812e-01
-5.55040240e-01 -4.71072495e-01 -6.57867074e-01 3.02901983e-01
-1.46251225e+00 -5.25760174e-01 -2.33801678e-02 -2.33654790e-02
1.23356903e+00 1.02262938e+00 5.78302681e-01 2.00716570e-01
6.47240579e-02 2.54183441e-01 7.59220958e-01 1.36938870e-01
1.85712904e-01 1.56136900e-01 2.54463732e-01 4.69304234e-01
9.80912089e-01 -1.78331718e-01 -1.04993260e+00 -7.00166523e-01
9.49465781e-02 -4.92813736e-02 -9.10574943e-03 -2.28139162e-01
-1.23782241e+00 3.69457811e-01 6.38957798e-01 2.33493328e-01
-4.61342812e-01 -1.29802391e-01 8.56165528e-01 4.88782704e-01
1.09209262e-01 1.90294936e-01 -7.64953256e-01 -7.54321665e-02
-3.63936752e-01 7.19291925e-01 1.23779094e+00 1.50006187e+00
1.06943631e+00 -2.86950558e-01 -4.43994440e-02 7.67873526e-01
4.70289230e-01 9.01366353e-01 2.34677494e-01 -1.00914562e+00
8.67994547e-01 1.24733341e+00 4.14530993e-01 -5.79166234e-01
-3.29664856e-01 2.50081182e-01 -4.10886794e-01 2.42175937e-01
2.36024223e-02 4.02965397e-01 -4.57103103e-01 9.98424172e-01
9.20273125e-01 -7.68025339e-01 4.30167526e-01 8.79370213e-01
1.48415065e+00 5.47931790e-01 2.37695917e-01 -5.95034808e-02
1.63652611e+00 -1.89891666e-01 -9.03566778e-01 6.20736063e-01
4.74460065e-01 -8.33655715e-01 1.10359240e+00 1.84432998e-01
-1.53136611e+00 -2.13845268e-01 -9.74363029e-01 -5.48099935e-01
-1.25998056e+00 -7.09798872e-01 6.04210973e-01 1.13476098e-01
-7.69902945e-01 1.80193618e-01 -6.43890977e-01 -5.97044170e-01
1.48423776e-01 1.37701496e-01 -6.84518754e-01 -2.07383752e-01
-1.37793076e+00 6.16067529e-01 1.15002954e+00 -4.06863749e-01
-4.15010780e-01 -1.21493757e+00 -5.74204922e-01 2.12471083e-01
4.46861058e-01 -8.12557161e-01 8.18565607e-01 1.50773674e-01
-5.91669917e-01 8.41971040e-01 2.32290596e-01 -3.05311173e-01
3.46587032e-01 -1.05985060e-01 -1.31802106e+00 -7.88716376e-02
4.33803171e-01 -8.06248561e-02 -3.78213376e-01 -8.89565647e-01
-9.35447514e-01 -9.60469186e-01 4.57766876e-02 -1.84567004e-01
-1.17233925e-01 3.91989142e-01 -4.84710455e-01 -3.34050581e-02
-2.00149909e-01 -4.18797769e-02 5.25012970e-01 -1.48947075e-01
-2.18256071e-01 -5.45529723e-01 1.10617626e+00 -3.42487663e-01
1.43616045e+00 -1.71581078e+00 -3.09635401e-01 2.25703448e-01
7.95580000e-02 3.80059004e-01 1.76804662e-01 1.21531045e+00
4.09685194e-01 4.23792928e-01 1.49562180e-01 4.71228987e-01
3.03918719e-01 6.71226382e-01 -4.16648954e-01 -4.32867944e-01
-4.04478520e-01 9.56465542e-01 -7.92648733e-01 -8.22385550e-01
-8.60721096e-02 3.12522262e-01 -2.94684887e-01 3.74108642e-01
-8.80374730e-01 -9.48159397e-02 -7.51588523e-01 1.27405226e+00
1.02482939e+00 -4.25691813e-01 5.23041844e-01 -5.91191649e-01
-4.51931477e-01 4.36157495e-01 -1.19241130e+00 2.13062501e+00
-2.38419935e-01 -3.01795274e-01 4.35363725e-02 -1.40490830e-01
9.56511974e-01 4.67749268e-01 6.50980532e-01 -1.42305541e+00
-4.62768376e-01 8.18507791e-01 -8.94145072e-01 -1.05105782e+00
7.24531174e-01 5.16171753e-01 -4.56431538e-01 3.20860237e-01
1.06746648e-02 2.36953303e-01 1.02355516e+00 2.84274638e-01
1.40861833e+00 1.76390380e-01 3.24974954e-01 -3.99985790e-01
4.11528707e-01 4.51911658e-01 6.91178143e-01 4.05656397e-01
6.54975116e-01 -5.15194796e-02 3.69422466e-01 -8.89857292e-01
-9.11830068e-01 -1.74657702e+00 -5.90290070e-01 1.00470328e+00
2.48124450e-01 -1.05463934e+00 3.34422616e-03 -3.72859180e-01
6.72616005e-01 5.60929716e-01 -3.37130129e-01 4.16574478e-01
-4.81866181e-01 -5.62533081e-01 6.16879642e-01 2.51951545e-01
9.37985837e-01 -1.21909189e+00 -9.05879021e-01 2.32766792e-01
-1.90968245e-01 -9.33894157e-01 4.09814477e-01 -5.04323363e-01
-7.24576592e-01 -1.62537122e+00 3.20842236e-01 -1.63070392e-02
-5.50296381e-02 -1.29735842e-01 1.92780948e+00 3.35690141e-01
-4.29396719e-01 -1.73870064e-02 -4.38091189e-01 -3.80978227e-01
-1.87058613e-01 9.60742533e-02 -3.86807799e-01 -4.09449428e-01
5.16749382e-01 -7.24495173e-01 -5.21057010e-01 3.88906538e-01
-1.56737828e+00 -2.87336022e-01 1.95059646e-02 -3.02600786e-02
9.94069040e-01 9.85315666e-02 9.11414206e-01 -1.29207528e+00
5.46483517e-01 -1.20493281e+00 -1.21528304e+00 6.85099840e-01
-9.84180450e-01 2.33895108e-01 5.45389116e-01 5.44932723e-01
-1.14634013e+00 -5.16219795e-01 -1.68517977e-01 -2.82593537e-02
1.12683773e-01 1.06307745e+00 -7.31384218e-01 4.32844609e-01
6.12653077e-01 -3.04973871e-01 -1.79647744e-01 -1.23116851e+00
4.88548547e-01 6.72483563e-01 7.08982170e-01 -9.30929959e-01
5.14913142e-01 5.88940978e-01 -5.14743626e-02 -2.98675120e-01
-5.58665514e-01 -3.39166462e-01 -1.19514756e-01 -2.73112138e-03
7.12464154e-01 -8.05619061e-01 -1.06746602e+00 -7.19183907e-02
-9.46059585e-01 1.11108020e-01 -7.98291445e-01 5.53116761e-02
-2.70481676e-01 -2.75945157e-01 -3.19711149e-01 -2.22953334e-01
-6.69169307e-01 -6.65435970e-01 1.12249517e+00 2.31829956e-02
3.05085748e-01 -4.92422163e-01 4.87126440e-01 4.97148722e-01
8.34521830e-01 7.99405813e-01 1.27322519e+00 -6.57808006e-01
-9.69319046e-01 -3.08047891e-01 -4.37833428e-01 -3.79025072e-01
4.44457084e-02 5.78833260e-02 -5.36824405e-01 -1.03220932e-01
-7.74639785e-01 -4.84765291e-01 5.31694526e-03 -7.25711167e-01
8.69999051e-01 -2.45565236e-01 -3.68721753e-01 5.98203719e-01
2.21344876e+00 6.56330511e-02 1.03221762e+00 6.06946349e-01
3.16970378e-01 3.71211022e-01 8.96766067e-01 5.80369413e-01
7.81935930e-01 6.83557093e-01 9.21335220e-01 6.21836543e-01
-5.76326251e-02 -4.05452698e-01 -3.05146724e-01 8.01752329e-01
-1.29962295e-01 -3.63869786e-01 -1.10214996e+00 3.36515397e-01
-2.05914211e+00 -1.06956434e+00 -7.65465915e-01 2.50240779e+00
9.63620365e-01 -2.71839827e-01 -2.02989317e-02 -8.17930922e-02
3.63003373e-01 6.63500503e-02 -7.23482668e-01 -4.26901206e-02
-5.15164673e-01 2.87140012e-01 4.10336852e-01 5.77637628e-02
-4.98574108e-01 6.37782812e-01 5.83514166e+00 1.84014946e-01
-6.27575636e-01 4.87230927e-01 -3.94734412e-01 -4.38224077e-01
-8.30086291e-01 1.38290599e-01 -1.06623805e+00 8.16121459e-01
1.41346979e+00 -7.77410209e-01 2.71251649e-01 5.26780665e-01
-1.49125963e-01 -2.76750475e-01 -8.12363625e-01 6.10244393e-01
-4.92087573e-01 -2.03232837e+00 3.34209263e-01 2.62858033e-01
5.71802378e-01 6.41104162e-01 -7.88997710e-01 2.27862686e-01
6.94570839e-01 -6.28957629e-01 6.55172288e-01 1.27429473e+00
9.10793483e-01 -5.55926859e-01 6.45208180e-01 1.65273864e-02
-1.37571776e+00 -2.46905163e-02 -6.66066408e-01 5.55652916e-01
1.76170498e-01 1.21627748e+00 -3.28854561e-01 1.44120860e+00
1.55573642e+00 2.65456796e-01 -6.25126958e-01 1.13315463e+00
4.63608772e-01 -2.80890223e-02 -8.10327053e-01 2.75791258e-01
-5.76933384e-01 1.78665072e-02 6.76809102e-02 6.71917439e-01
6.71616495e-01 3.00679822e-02 -4.75496426e-02 6.73322558e-01
-5.79565406e-01 5.47638774e-01 -5.53084970e-01 7.32960030e-02
1.13509142e+00 1.07014525e+00 4.95928042e-02 -6.43997073e-01
-5.22790551e-01 -7.86057860e-03 5.47456980e-01 3.03854823e-01
-6.27674043e-01 -6.43083930e-01 8.72710228e-01 7.91130364e-01
2.73923180e-03 1.66065633e-01 4.43526506e-01 -1.31564188e+00
5.41103542e-01 -9.09475505e-01 1.17179739e+00 -1.24621534e+00
-1.35806286e+00 7.15580285e-01 2.53008097e-01 -9.20885086e-01
-4.66202021e-01 -8.36665332e-02 4.84762751e-02 8.58740568e-01
-1.60673702e+00 -1.16317523e+00 -9.72328722e-01 1.07813179e+00
-3.23873758e-01 -1.15424238e-01 1.13376582e+00 1.13684368e+00
-2.65143454e-01 -1.84086621e-01 2.95695394e-01 -1.05393291e-01
7.88021982e-01 -1.12372386e+00 2.39918038e-01 1.68848410e-01
-3.38391036e-01 6.52905107e-01 3.12702000e-01 -8.30460012e-01
-2.04983377e+00 -1.06967926e+00 8.92184734e-01 -3.87548476e-01
6.46772861e-01 -2.51248628e-01 -1.11754394e+00 7.76945293e-01
2.59655952e-01 9.20962572e-01 5.48675895e-01 -1.56449109e-01
-6.98417902e-01 -1.06165230e+00 -1.48862302e+00 -2.80854166e-01
1.12685978e+00 -2.74658322e-01 -5.34990788e-01 5.56581974e-01
7.64610410e-01 -3.25170875e-01 -2.16519046e+00 7.30872631e-01
5.29291034e-01 -1.40411115e+00 1.11541247e+00 -8.89484525e-01
-1.03565119e-02 -6.23250067e-01 -8.36539567e-01 -7.11033285e-01
3.11506212e-01 -2.53105760e-01 -5.99573970e-01 1.52886295e+00
2.86837459e-01 -9.92908359e-01 3.65276873e-01 7.69315839e-01
-3.35463248e-02 -4.58050549e-01 -1.03062117e+00 -7.81907499e-01
-4.65205640e-01 5.19104898e-02 1.81953561e+00 8.36568296e-01
-6.07767738e-02 -6.56690001e-02 2.70749927e-01 1.88291639e-01
6.87243283e-01 5.99380434e-01 9.84298646e-01 -1.94030690e+00
1.86932340e-01 2.29490712e-01 -5.05490422e-01 -1.01088487e-01
-4.19175953e-01 -1.08087957e+00 -8.77322137e-01 -2.12451744e+00
-3.52181643e-01 -1.08525252e+00 -2.41440147e-01 7.24103332e-01
5.58259249e-01 1.06892791e-02 -5.73085388e-03 6.37511075e-01
-7.59244204e-01 1.70985281e-01 8.83691072e-01 -6.94608837e-02
3.92442137e-01 -3.78030747e-01 -2.81657308e-01 1.71391919e-01
3.08765233e-01 -5.71944892e-01 -5.43083727e-01 -8.59994829e-01
1.39265525e+00 5.90779006e-01 4.39305902e-01 -1.20958912e+00
4.33843493e-01 -3.08085293e-01 6.76021799e-02 -9.84876096e-01
2.79226989e-01 -1.07378769e+00 1.28721809e+00 -7.83866867e-02
1.45080924e-01 3.57944727e-01 -1.00389540e-01 2.23570377e-01
-7.74346471e-01 3.88922058e-02 2.64378965e-01 -2.74916321e-01
-6.85361803e-01 4.92634326e-01 5.76233983e-01 6.73321724e-01
1.02375114e+00 5.33752561e-01 -1.28140414e+00 4.20496553e-01
-6.74602330e-01 5.42130232e-01 6.75766170e-01 5.10579288e-01
2.88157284e-01 -1.68175781e+00 -7.27541625e-01 3.17719966e-01
4.98594970e-01 1.38168305e-01 1.39133886e-01 3.88676733e-01
-8.62382770e-01 5.08486569e-01 -3.92257452e-01 -1.86636791e-01
-8.22102487e-01 7.41043329e-01 1.89309493e-01 -2.50532687e-01
-7.22225308e-01 -7.02595040e-02 -8.56097519e-01 -7.04594016e-01
-1.67064384e-01 -1.41865149e-01 2.25922093e-01 3.78790796e-01
8.16041827e-01 7.93802500e-01 7.30581760e-01 -2.28159025e-01
-6.51108861e-01 2.16381356e-01 4.04775858e-01 1.80158183e-01
1.58164191e+00 -4.61665727e-03 -7.71505535e-01 4.72471207e-01
1.12426043e+00 5.19311666e-01 -3.17013055e-01 -1.14560179e-01
2.50538677e-01 -7.40606546e-01 -4.70427305e-01 -1.34547937e+00
-1.29549634e+00 1.28391311e-01 3.86199027e-01 8.47076952e-01
9.85479236e-01 2.75173157e-01 9.76831675e-01 5.79526901e-01
1.28555274e+00 -8.76989901e-01 -8.62101912e-01 7.19960555e-02
1.05912340e+00 -9.51705098e-01 1.77530155e-01 -2.81927675e-01
-2.12585218e-02 9.86822605e-01 3.68298113e-01 2.22344950e-01
8.58482778e-01 4.70674545e-01 1.83267474e-01 -9.60045218e-01
-1.18403804e+00 -1.25261232e-01 -1.89736664e-01 2.02978551e-01
3.57142240e-01 -2.29193922e-02 -6.56192422e-01 3.42677206e-01
-2.05801338e-01 6.05407119e-01 7.62684420e-02 1.39465261e+00
-5.32597482e-01 -1.52628648e+00 -3.51762265e-01 7.72053480e-01
-4.72626567e-01 1.81268573e-01 3.08805138e-01 1.03063488e+00
1.73383176e-01 6.32618666e-01 3.42864454e-01 1.06644258e-01
9.26476836e-01 2.16451243e-01 -1.81663204e-02 -2.45552182e-01
-8.61649096e-01 -7.83682168e-01 5.24394512e-01 -1.07295823e+00
-2.45785892e-01 -3.48904252e-01 -1.89857435e+00 -6.57443106e-01
4.04403001e-01 7.12526441e-01 1.00731921e+00 2.17191964e-01
9.62983787e-01 1.38286054e-01 -6.60473993e-03 2.01067254e-01
-3.24108243e-01 -6.05631173e-01 -6.70899808e-01 5.94490767e-01
-3.97267878e-01 -5.51235735e-01 1.86352253e-01 -2.58929759e-01]
|
[9.186179161071777, 7.872093200683594]
|
124a5550-45b4-4f07-89e7-9ae5a914417b
|
deep-unfolding-as-iterative-regularization
|
2211.13452
| null |
https://arxiv.org/abs/2211.13452v1
|
https://arxiv.org/pdf/2211.13452v1.pdf
|
Deep unfolding as iterative regularization for imaging inverse problems
|
Recently, deep unfolding methods that guide the design of deep neural networks (DNNs) through iterative algorithms have received increasing attention in the field of inverse problems. Unlike general end-to-end DNNs, unfolding methods have better interpretability and performance. However, to our knowledge, their accuracy and stability in solving inverse problems cannot be fully guaranteed. To bridge this gap, we modified the training procedure and proved that the unfolding method is an iterative regularization method. More precisely, we jointly learn a convex penalty function adversarially by an input-convex neural network (ICNN) to characterize the distance to a real data manifold and train a DNN unfolded from the proximal gradient descent algorithm with this learned penalty. Suppose the real data manifold intersects the inverse problem solutions with only the unique real solution. We prove that the unfolded DNN will converge to it stably. Furthermore, we demonstrate with an example of MRI reconstruction that the proposed method outperforms conventional unfolding methods and traditional regularization methods in terms of reconstruction quality, stability and convergence speed.
|
['Dong Liang', 'Jing Cheng', 'Qingyong Zhu', 'Zhuo-Xu Cui']
|
2022-11-24
| null | null | null | null |
['mri-reconstruction']
|
['computer-vision']
|
[ 1.30521521e-01 5.77236593e-01 1.44741267e-01 -4.28657025e-01
-5.60815036e-01 -4.34110790e-01 6.73353225e-02 -4.78099287e-01
-4.29569304e-01 8.18783700e-01 1.07983097e-01 -2.51352489e-01
-3.28784078e-01 -4.27732140e-01 -1.27346170e+00 -8.43292236e-01
4.51853573e-02 5.23762941e-01 -3.70200127e-01 -1.73019171e-01
4.35305610e-02 4.97041941e-01 -6.47683799e-01 -2.87768453e-01
1.17654240e+00 7.88552225e-01 1.67407855e-01 1.82280719e-01
1.15555212e-01 6.52035356e-01 -7.25808367e-02 -7.64629319e-02
5.23090482e-01 -5.52850127e-01 -9.25168514e-01 -8.29950068e-03
3.33985627e-01 -4.11332518e-01 -6.03583694e-01 1.33681738e+00
5.60202241e-01 4.00357634e-01 5.70159554e-01 -9.40830052e-01
-1.07365394e+00 4.42234844e-01 -4.37032431e-01 -9.99318585e-02
-3.89801078e-02 -2.30993673e-01 6.70359194e-01 -1.30560505e+00
7.57514417e-01 9.13261354e-01 1.01350868e+00 9.99359429e-01
-1.35119081e+00 -3.30530077e-01 -6.66220160e-03 -9.33797807e-02
-1.13048017e+00 -2.96570152e-01 9.59051430e-01 -3.86842430e-01
2.25149289e-01 1.01693332e-01 4.73904163e-01 1.08074963e+00
-1.09400135e-02 7.83257425e-01 8.68045390e-01 -3.16785961e-01
2.77717501e-01 -7.68738464e-02 3.08372993e-02 8.78323615e-01
-3.92948389e-02 9.77265909e-02 -3.07563487e-02 5.12095690e-02
1.12802255e+00 2.01037839e-01 -5.99231839e-01 -4.86454070e-01
-1.04000747e+00 9.39176381e-01 8.48435402e-01 2.06235066e-01
-3.69321555e-01 2.08992600e-01 2.56901145e-01 3.26350898e-01
5.01700938e-01 2.68561572e-01 -2.50294447e-01 2.20342755e-01
-6.70577347e-01 1.22105420e-01 7.44901061e-01 6.63738430e-01
6.68726921e-01 2.58531868e-01 1.67420194e-01 8.46330047e-01
3.00648034e-01 4.39201921e-01 3.99682999e-01 -1.24514830e+00
4.17554110e-01 4.20917451e-01 3.60729508e-02 -9.83072042e-01
-5.82733214e-01 -6.57127321e-01 -1.21175110e+00 4.74948972e-01
6.63048446e-01 -3.56789440e-01 -6.85076118e-01 2.13500714e+00
3.08167338e-01 4.49386597e-01 5.12845889e-02 1.50804484e+00
5.48404038e-01 5.25615811e-01 -2.89198160e-01 -2.36942232e-01
7.76108444e-01 -7.76104629e-01 -7.69183815e-01 -1.53278127e-01
7.03083992e-01 -4.47557598e-01 1.13560534e+00 3.06534559e-01
-1.16550112e+00 -3.44124198e-01 -1.14454365e+00 -2.68858135e-01
2.56310225e-01 1.90406829e-01 1.86387584e-01 1.16211161e-01
-8.88028800e-01 1.07405102e+00 -1.13997912e+00 -2.88178492e-02
3.60883921e-01 3.48964006e-01 -3.35278451e-01 1.02859698e-02
-1.16007280e+00 8.29402685e-01 2.60809879e-03 6.08898640e-01
-9.98125672e-01 -9.38565016e-01 -5.80686748e-01 -2.19637036e-01
9.59530398e-02 -8.11338842e-01 9.47775066e-01 -1.29710746e+00
-1.69953024e+00 9.21039283e-01 5.76019809e-02 -5.13262928e-01
8.54540169e-01 -5.32700956e-01 -3.54577638e-02 9.28870589e-03
5.06123118e-02 3.37479532e-01 9.12155688e-01 -1.26798737e+00
6.30829558e-02 -5.04893422e-01 6.82333186e-02 1.27464965e-01
-2.18432918e-01 -5.30016005e-01 2.23973915e-02 -6.53724551e-01
6.03806853e-01 -9.47114408e-01 -4.89041537e-01 4.68499899e-01
-3.99365187e-01 4.71958779e-02 8.47483456e-01 -7.44008958e-01
8.37244391e-01 -2.27767396e+00 5.55928767e-01 2.22045973e-01
4.27884281e-01 2.32119173e-01 -1.28190398e-01 -7.65283629e-02
-4.43068802e-01 -1.42544448e-01 -6.06983721e-01 -2.59657919e-01
-6.27771989e-02 4.69332844e-01 -3.82106990e-01 1.08035207e+00
-7.95578063e-02 9.84362185e-01 -9.53446269e-01 -5.57440221e-02
3.61668393e-02 6.59302711e-01 -7.49565363e-01 3.39555502e-01
-8.49058703e-02 9.52962220e-01 -5.77456772e-01 3.50438431e-02
7.12721944e-01 -1.50572315e-01 -3.37437540e-02 -3.56466115e-01
1.50231838e-01 -1.92997620e-01 -1.00938594e+00 2.04446459e+00
-6.05616391e-01 6.19672179e-01 5.42345405e-01 -1.67804837e+00
1.08673120e+00 2.44627804e-01 6.13377273e-01 -4.98240978e-01
8.93787816e-02 5.24501324e-01 -2.12664396e-01 -8.11986089e-01
1.69748254e-02 -4.57936823e-01 4.19101745e-01 3.90526503e-01
-1.05946667e-01 1.54376641e-01 -4.11197305e-01 3.39619517e-02
8.82983983e-01 2.33852476e-01 -1.38926923e-01 -5.43874681e-01
6.87652469e-01 -2.76170045e-01 8.14956307e-01 4.58476484e-01
-1.86308518e-01 1.01539814e+00 4.95768070e-01 -7.03396142e-01
-1.27065992e+00 -1.35413611e+00 -4.02563900e-01 7.68844247e-01
1.66142061e-01 3.39388192e-01 -1.17694354e+00 -5.79567015e-01
-1.77401766e-01 5.02997220e-01 -7.43062019e-01 -2.37008721e-01
-9.31068480e-01 -4.84030604e-01 4.49676722e-01 3.94067615e-01
5.42701840e-01 -1.09289110e+00 -2.54324436e-01 2.16428623e-01
-4.30336833e-01 -1.05584097e+00 -8.13163877e-01 -9.09263119e-02
-1.28554666e+00 -9.72361743e-01 -1.02653766e+00 -1.11814332e+00
1.27701735e+00 -2.09641770e-01 9.44471896e-01 9.88977700e-02
-8.46324861e-02 2.42219046e-01 -2.45184861e-02 2.20240235e-01
-4.44539845e-01 -9.51558575e-02 3.34052771e-01 1.92114606e-01
-1.27298176e-01 -1.10768700e+00 -9.21412349e-01 3.54682416e-01
-9.60550487e-01 1.77448556e-01 4.33109522e-01 9.60263431e-01
7.28742421e-01 -4.82216358e-01 8.07005942e-01 -8.41789782e-01
6.78794384e-01 -4.13446397e-01 -5.92425168e-01 1.80331394e-01
-5.38009584e-01 4.84153122e-01 1.00267386e+00 -4.59658980e-01
-8.65910232e-01 3.22767347e-01 -4.47186500e-01 -7.41773725e-01
2.13937178e-01 5.64355373e-01 6.86727278e-03 -3.34775746e-01
1.06083763e+00 1.79447547e-01 3.70347053e-01 -5.12786508e-01
5.24479389e-01 1.19939506e-01 8.39929402e-01 -8.21949184e-01
7.31621504e-01 7.49375343e-01 1.72254264e-01 -6.14499867e-01
-1.18167698e+00 7.27798566e-02 -6.81702971e-01 -2.28952855e-01
8.37383389e-01 -6.11470222e-01 -7.65395164e-01 2.85674393e-01
-1.33296490e+00 -4.65803504e-01 -5.26217818e-01 5.73688805e-01
-8.25367093e-01 5.32359779e-01 -7.52064586e-01 -4.83784407e-01
-4.77898329e-01 -1.11752856e+00 5.93596220e-01 -1.20928504e-01
6.41431510e-02 -1.29132903e+00 7.67127350e-02 1.08419405e-02
2.23927423e-01 7.00802565e-01 7.14000881e-01 -2.65622318e-01
-2.88250566e-01 7.97028001e-03 -2.22766936e-01 6.66007817e-01
-7.05056190e-02 -4.17316318e-01 -6.68592751e-01 -3.06559086e-01
7.55865395e-01 -1.95155159e-01 6.57930851e-01 7.33939350e-01
1.30258512e+00 -4.74928230e-01 -4.08314355e-02 1.28630495e+00
1.50446796e+00 -1.57467097e-01 6.42566562e-01 2.96470284e-01
9.12026107e-01 5.57975352e-01 2.21022796e-02 1.85734168e-01
-4.95336279e-02 2.95116991e-01 6.28569841e-01 -8.47237110e-02
5.99960387e-02 -1.62015572e-01 3.18521023e-01 9.72794235e-01
-2.02533498e-01 4.38440114e-01 -7.26955116e-01 4.27759945e-01
-2.09317470e+00 -5.97158849e-01 -2.89659262e-01 2.05809069e+00
7.06762552e-01 1.66872069e-02 -2.79570401e-01 -5.96590489e-02
7.53896952e-01 -1.09062843e-01 -9.78509307e-01 -2.88171351e-01
-6.47675470e-02 -8.13010186e-02 4.78880733e-01 7.65062392e-01
-7.63894796e-01 5.87618887e-01 5.91026258e+00 5.97744823e-01
-1.06390357e+00 2.07447812e-01 6.41438782e-01 5.41842654e-02
-4.39987212e-01 -2.04521120e-01 -2.77427107e-01 1.67629883e-01
4.83793408e-01 -1.66898891e-01 8.13403130e-01 8.71248007e-01
4.92008746e-01 4.50427204e-01 -1.29152930e+00 1.02880299e+00
-1.04210898e-01 -1.41398764e+00 -1.46220461e-01 -3.05218436e-02
1.01406002e+00 6.18676879e-02 1.32440209e-01 1.86351053e-02
1.04844064e-01 -1.14892507e+00 5.59429646e-01 7.05347180e-01
6.53323531e-01 -8.42716813e-01 4.92704719e-01 5.71946084e-01
-7.45601535e-01 8.08579847e-02 -4.73486781e-01 -3.90347391e-02
3.31412584e-01 8.35392654e-01 -3.13291848e-01 2.87058443e-01
4.04392928e-01 8.28472733e-01 1.18963286e-01 6.47748291e-01
-3.06831032e-01 4.52928960e-01 -2.63379902e-01 3.84858876e-01
4.26819503e-01 -9.09547687e-01 7.42489219e-01 8.03189576e-01
3.88531089e-01 4.63682234e-01 -5.46634607e-02 1.32295799e+00
-3.73136073e-01 -8.47588778e-02 -6.58471227e-01 2.55266368e-01
-5.39415702e-02 1.16546309e+00 -5.31499147e-01 8.33754987e-02
-2.32271388e-01 1.06644619e+00 6.71470523e-01 7.43621290e-01
-8.85887384e-01 -2.75442570e-01 6.82249844e-01 1.19241297e-01
-4.26332504e-02 -2.95182109e-01 -5.90414703e-01 -1.23540640e+00
2.79410034e-01 -5.01414359e-01 1.65945236e-02 -7.18493164e-01
-1.41645324e+00 7.74573743e-01 -5.55335283e-01 -1.27298045e+00
-3.17800827e-02 -7.41469502e-01 -5.88321626e-01 8.40527356e-01
-1.09416175e+00 -5.34934759e-01 -2.97795415e-01 8.10133994e-01
2.98502147e-01 -6.97705448e-02 6.11073613e-01 5.23700356e-01
-4.99807209e-01 5.12131095e-01 5.92055321e-01 4.43843007e-01
4.03785050e-01 -1.19985378e+00 4.61722985e-02 8.66879940e-01
-2.09911048e-01 7.17173398e-01 8.84556711e-01 -3.69324774e-01
-1.48336220e+00 -1.13093388e+00 2.81148851e-01 -1.75346047e-01
7.96245694e-01 -2.23964125e-01 -1.18797755e+00 9.20461178e-01
-4.16662432e-02 5.28696120e-01 9.07540619e-02 -1.54774249e-01
-9.51309782e-03 -1.77371994e-01 -1.22462642e+00 6.87290311e-01
1.19866514e+00 -5.07698059e-01 -6.17880285e-01 5.95365226e-01
6.72766626e-01 -8.45649004e-01 -7.83458054e-01 4.52396512e-01
3.62044781e-01 -8.22834849e-01 1.15742278e+00 -7.15205848e-01
6.97425187e-01 -2.06395075e-01 -6.02480303e-03 -1.42825818e+00
-1.79931611e-01 -7.76886702e-01 -2.80700475e-01 7.16427207e-01
1.95433795e-01 -6.98306441e-01 1.03382957e+00 6.26027405e-01
-4.74993676e-01 -1.05010486e+00 -1.15255344e+00 -7.81008720e-01
4.94222581e-01 -4.31472808e-01 4.62067463e-02 8.92458856e-01
-1.98149309e-01 2.85717726e-01 -4.65764940e-01 2.43272871e-01
9.87420857e-01 -2.51922924e-02 2.66178310e-01 -1.04375589e+00
-1.63727254e-01 -3.29762578e-01 -1.75477579e-01 -1.50650597e+00
5.31929970e-01 -1.16823792e+00 2.27327213e-01 -1.50750589e+00
-1.27294332e-01 -5.00182390e-01 -1.35441765e-01 6.87284544e-02
5.98619357e-02 6.48891702e-02 -4.84158620e-02 3.86425436e-01
-1.90185577e-01 8.79307389e-01 1.77332330e+00 6.45617722e-03
-2.81150281e-01 1.18427634e-01 -4.62234497e-01 9.73424435e-01
6.55175388e-01 -5.78643084e-01 -5.44653893e-01 -7.89893389e-01
2.45747373e-01 3.29338044e-01 3.95140707e-01 -7.93374598e-01
2.88276583e-01 1.52083904e-01 2.53362030e-01 -1.50323972e-01
1.33092865e-01 -9.59235430e-01 5.04587144e-02 6.11200809e-01
-4.52305824e-01 -2.25177750e-01 -1.37614980e-01 6.55330479e-01
-2.42772792e-02 -4.10441399e-01 9.71382856e-01 7.71486759e-03
-2.23087698e-01 6.75990760e-01 -7.25591853e-02 4.15372938e-01
7.19055057e-01 -1.29288599e-01 2.29881763e-01 -4.26500916e-01
-1.31918764e+00 1.25604972e-01 5.41958958e-02 -2.47057504e-03
8.84782732e-01 -1.60842848e+00 -8.21138263e-01 2.57920235e-01
-5.29247403e-01 3.94421846e-01 1.24795198e-01 1.24694431e+00
-8.37131083e-01 -7.13937581e-02 -9.41963568e-02 -6.93891466e-01
-5.69012344e-01 5.78584552e-01 9.18757498e-01 -7.93344006e-02
-1.12123024e+00 6.80537403e-01 4.97915506e-01 -8.86394620e-01
4.09331828e-01 -2.24724144e-01 1.51355416e-01 -4.06877220e-01
3.37299079e-01 4.04714257e-01 -9.25743878e-02 -3.99621159e-01
-2.16178566e-01 6.53979123e-01 2.79643446e-01 -8.20896924e-02
1.67904341e+00 -2.23734424e-01 -3.12549531e-01 1.42743021e-01
1.75033820e+00 -3.25289398e-01 -1.61003995e+00 -3.87696773e-01
-3.62387337e-02 -1.64203331e-01 -4.27011028e-02 -2.37050116e-01
-1.69381261e+00 8.41902316e-01 7.07314432e-01 8.34492669e-02
1.07058454e+00 -2.34691381e-01 9.48242605e-01 5.87980747e-01
3.90496328e-02 -1.01739287e+00 1.17974676e-01 5.72181463e-01
1.25128877e+00 -1.19200444e+00 -1.72908694e-01 -1.07381895e-01
-2.88831115e-01 1.49731243e+00 5.03067970e-01 -8.47702563e-01
9.77424085e-01 1.45635083e-01 1.61106214e-01 -9.86235812e-02
-1.06497467e-01 4.00573045e-01 2.64467895e-01 3.63701135e-01
3.40380400e-01 -1.28870398e-01 -5.72708368e-01 4.20805693e-01
-4.98773642e-02 1.53346032e-01 4.11575437e-01 4.25791413e-01
-3.20746928e-01 -7.11424887e-01 -3.49183232e-01 1.52856067e-01
-4.06184763e-01 3.09181903e-02 1.62148461e-01 5.90642452e-01
-1.38340175e-01 4.06263888e-01 -1.90128461e-01 -2.00080767e-01
3.38846147e-01 -2.25509658e-01 5.60474873e-01 -1.57539070e-01
-7.94884190e-02 -6.56280294e-02 -4.03424680e-01 -5.74345887e-01
-2.48023108e-01 -4.59533185e-01 -1.71823847e+00 -3.99891973e-01
-4.72029075e-02 2.48765513e-01 5.30273438e-01 1.15255511e+00
2.51141489e-01 4.91188526e-01 8.63527358e-01 -9.18250918e-01
-8.68911326e-01 -6.74796164e-01 -6.29411876e-01 5.95241010e-01
6.25744760e-01 -4.62873608e-01 -6.41106486e-01 1.30949974e-01]
|
[11.90147590637207, -2.429399251937866]
|
ffdb8d58-abee-49f9-a973-1f52565679e4
|
multi-label-logo-recognition-and-retrieval
|
2205.05419
| null |
https://arxiv.org/abs/2205.05419v2
|
https://arxiv.org/pdf/2205.05419v2.pdf
|
Multi-Label Logo Recognition and Retrieval based on Weighted Fusion of Neural Features
|
Classifying logo images is a challenging task as they contain elements such as text or shapes that can represent anything from known objects to abstract shapes. While the current state of the art for logo classification addresses the problem as a multi-class task focusing on a single characteristic, logos can have several simultaneous labels, such as different colors. This work proposes a method that allows visually similar logos to be classified and searched from a set of data according to their shape, color, commercial sector, semantics, general characteristics, or a combination of features selected by the user. Unlike previous approaches, the proposal employs a series of multi-label deep neural networks specialized in specific attributes and combines the obtained features to perform the similarity search. To delve into the classification system, different existing logo topologies are compared and some of their problems are analyzed, such as the incomplete labeling that trademark registration databases usually contain. The proposal is evaluated considering 76,000 logos (7 times more than previous approaches) from the European Union Trademarks dataset, which is organized hierarchically using the Vienna ontology. Overall, experimentation attains reliable quantitative and qualitative results, reducing the normalized average rank error of the state-of-the-art from 0.040 to 0.018 for the Trademark Image Retrieval task. Finally, given that the semantics of logos can often be subjective, graphic design students and professionals were surveyed. Results show that the proposed methodology provides better labeling than a human expert operator, improving the label ranking average precision from 0.53 to 0.68.
|
['Antonio Pertusa', 'Antonio Javier Gallego', 'Marisa Bernabeu']
|
2022-05-11
| null | null | null | null |
['logo-recognition']
|
['computer-vision']
|
[-2.00490151e-02 -2.48089254e-01 -4.73144323e-01 -2.20889524e-01
-5.35172880e-01 -9.59984004e-01 6.09471798e-01 7.17404485e-01
-2.46384487e-01 4.29458767e-01 -3.35459918e-01 -1.14167534e-01
-6.49646699e-01 -8.35475206e-01 -4.49578941e-01 -6.07109427e-01
9.43282545e-02 1.09937954e+00 1.31908506e-01 -5.21320514e-02
4.86184895e-01 1.10778487e+00 -2.16632247e+00 2.91294694e-01
7.01397121e-01 1.61381888e+00 -2.39155963e-02 -8.86196420e-02
-4.19060588e-01 5.72752237e-01 -8.51952851e-01 -1.68067813e-01
1.92030489e-01 1.33291930e-01 -7.76105344e-01 2.35697255e-01
1.06610930e+00 3.59655842e-02 2.26975068e-01 1.15739179e+00
3.36073071e-01 4.25806604e-02 1.22211468e+00 -1.27946258e+00
-8.67080629e-01 2.72228897e-01 -2.86568969e-01 -2.27459714e-01
2.98756957e-01 -3.19714874e-01 1.22277355e+00 -7.89113283e-01
8.47413301e-01 1.15012264e+00 5.70505202e-01 -4.76877810e-03
-1.14124644e+00 -7.30652392e-01 -2.05367774e-01 5.25827169e-01
-1.59572160e+00 3.84083092e-02 8.40451002e-01 -7.96967149e-01
7.22621083e-01 3.04291219e-01 4.96438354e-01 6.31044328e-01
2.35343248e-01 1.65621221e-01 1.51838529e+00 -5.11242270e-01
2.19973132e-01 5.96687913e-01 3.21091771e-01 7.24130154e-01
5.58713436e-01 -1.58709586e-01 -3.38149339e-01 -1.00238379e-02
3.22461039e-01 -1.58258826e-01 2.44484410e-01 -7.12561786e-01
-6.78500652e-01 7.15330839e-01 4.92625892e-01 7.41233885e-01
-1.22788884e-01 -2.20788866e-01 3.76371741e-01 3.50626081e-01
2.33602494e-01 5.57479858e-01 -1.65047109e-01 4.78168309e-01
-8.36305082e-01 1.50445238e-01 9.44081724e-01 7.71927774e-01
9.26849484e-01 -1.66037828e-01 1.70681909e-01 9.33243036e-01
3.90368938e-01 4.36506540e-01 5.14688551e-01 -7.05876112e-01
4.39836755e-02 9.70254481e-01 -1.98651459e-02 -1.20887959e+00
-6.29277229e-01 -5.28430879e-01 -3.63823146e-01 7.78052151e-01
6.35289848e-01 6.49780571e-01 -1.05294156e+00 1.08967793e+00
5.53518273e-02 -7.42489874e-01 -1.49964929e-01 6.75680697e-01
1.14018464e+00 3.89630675e-01 2.19775975e-01 1.41814888e-01
1.71013141e+00 -9.10760224e-01 -7.08835006e-01 -1.48113489e-01
3.98045570e-01 -1.00061750e+00 1.01559842e+00 6.88771486e-01
-5.83607614e-01 -6.56630337e-01 -1.22306824e+00 2.03025207e-01
-1.08070827e+00 5.12523174e-01 5.62131405e-01 7.67805815e-01
-8.88088465e-01 6.43401504e-01 -1.91648871e-01 -5.87362349e-01
4.04891849e-01 5.22073150e-01 -4.04151082e-01 1.92734465e-01
-9.87072170e-01 9.39205706e-01 8.83608103e-01 -1.18745141e-01
-7.69981861e-01 -3.76765817e-01 -4.87248391e-01 3.15617830e-01
6.01404905e-01 -1.23680428e-01 7.06726253e-01 -1.05060005e+00
-1.19294095e+00 1.14530563e+00 4.60799158e-01 -7.56379738e-02
6.08732283e-01 3.52628413e-03 -8.22811663e-01 3.91003132e-01
2.88306206e-01 4.32282865e-01 7.00976968e-01 -1.67895746e+00
-7.41248846e-01 -2.81800270e-01 1.61567315e-01 5.89606876e-04
-4.75330859e-01 -3.33494581e-02 -3.13329637e-01 -6.22221708e-01
1.93937048e-01 -8.61385465e-01 2.78614253e-01 5.48399054e-02
-9.39198062e-02 -6.15541935e-01 6.00441217e-01 -3.87255430e-01
1.07909262e+00 -2.06236696e+00 -3.66316527e-01 6.24639392e-01
2.47144639e-01 5.28402179e-02 6.24065772e-02 5.33429027e-01
-2.47993663e-01 1.16951838e-01 -9.47334319e-02 1.03659496e-01
2.97825634e-01 -9.56120435e-03 2.03791931e-01 4.27683622e-01
-3.05511951e-01 5.82583010e-01 -5.09946883e-01 -7.84662187e-01
3.06782365e-01 3.48096848e-01 6.41085878e-02 -1.22455053e-01
-1.60036892e-01 -5.27472720e-02 -4.36265975e-01 1.18169081e+00
5.74262977e-01 -4.28613096e-01 2.61176854e-01 -4.59843129e-01
1.48141226e-02 -1.96101457e-01 -1.29802203e+00 1.31845915e+00
-7.07433701e-01 6.28023446e-01 -4.52086329e-01 -1.13860869e+00
1.29667425e+00 2.12827116e-01 7.51025081e-01 -7.21701086e-01
4.77127761e-01 7.33596623e-01 -3.66958201e-01 -5.58974981e-01
4.53796148e-01 -1.47601411e-01 -8.87957886e-02 3.51668596e-01
4.33905423e-02 1.13955870e-01 6.05890691e-01 -2.71182775e-01
6.84880614e-01 1.73093826e-01 3.50588053e-01 -3.89225781e-01
7.96341956e-01 2.61729538e-01 1.07421145e-01 6.36273205e-01
9.94358510e-02 4.92369533e-01 4.42505479e-01 -6.00625396e-01
-9.93309855e-01 -1.11841190e+00 -4.37760234e-01 1.15821040e+00
3.11104536e-01 8.81275162e-02 -6.04536355e-01 -7.61179626e-01
2.34566629e-01 3.59854847e-01 -6.45687222e-01 1.58191845e-01
-4.55027789e-01 -5.92629254e-01 3.69009346e-01 3.51827383e-01
3.72322112e-01 -1.32258320e+00 -6.66446686e-01 1.69578627e-01
1.94504559e-01 -9.34476912e-01 -1.37195230e-01 3.94496948e-01
-6.82119131e-01 -1.19009244e+00 -6.70673192e-01 -1.14969945e+00
7.05837727e-01 -1.82993665e-01 1.22312438e+00 -2.83815991e-02
-3.40909988e-01 2.84993768e-01 -4.64679778e-01 -1.57475591e-01
-5.84812760e-01 1.54690742e-01 -2.47273445e-01 2.59003401e-01
5.50454497e-01 -3.53684008e-01 -1.70554340e-01 4.46556598e-01
-1.24847651e+00 -5.71191013e-01 9.13350046e-01 6.68515325e-01
6.71048164e-01 3.15122038e-01 4.32226509e-01 -9.83870149e-01
5.62945664e-01 -2.32349217e-01 -9.01678681e-01 6.98625863e-01
-1.22608066e+00 1.79238886e-01 6.04291260e-01 -5.57720959e-01
-1.00691116e+00 1.79082021e-01 2.59901404e-01 -1.27269059e-01
-5.62959015e-01 6.03828788e-01 -3.94421279e-01 -4.76370335e-01
7.18106568e-01 7.67567661e-03 -1.11497864e-01 -8.90277028e-01
2.17256159e-01 7.74599493e-01 4.53865558e-01 -6.20661080e-01
7.45822251e-01 5.55037320e-01 1.99700728e-01 -4.51095819e-01
-6.95209622e-01 -5.79405963e-01 -6.92814887e-01 -5.21373332e-01
8.03648531e-01 -3.97602409e-01 -1.09903085e+00 4.87736017e-01
-9.50706244e-01 2.37242296e-01 -1.42203763e-01 3.34069580e-01
-3.92125100e-01 3.64588290e-01 -3.98892134e-01 -4.34185594e-01
-1.04421243e-01 -1.12350881e+00 8.60085607e-01 2.03264952e-01
-1.23259746e-01 -9.07811701e-01 -2.07379833e-01 5.43380022e-01
2.88396657e-01 4.26581889e-01 1.36701453e+00 -1.02351582e+00
-5.29237032e-01 -6.30554080e-01 -3.33858758e-01 1.58749804e-01
6.91276044e-02 -2.32924446e-01 -9.77058530e-01 -2.16254681e-01
-1.48597643e-01 -2.31721908e-01 7.37802565e-01 2.12831348e-02
9.50144351e-01 -1.55775085e-01 -2.22413659e-01 2.43082434e-01
1.96559501e+00 7.05527544e-01 2.63005048e-01 1.01484668e+00
6.34353518e-01 1.02227044e+00 6.85642421e-01 2.47362122e-01
-1.11883909e-01 7.99803138e-01 6.34177864e-01 2.31396360e-03
-2.61112124e-01 1.61712945e-01 1.05234712e-01 2.96376348e-01
-1.46911845e-01 -3.11869532e-01 -9.42113101e-01 4.44497287e-01
-1.46398807e+00 -5.65080702e-01 -7.37592205e-02 2.17740655e+00
2.33934984e-01 2.74158388e-01 1.69939578e-01 4.69945222e-01
9.52274382e-01 1.02571182e-01 -2.66624451e-01 -3.78985852e-01
-2.37904236e-01 1.83713108e-01 8.42268288e-01 3.19679379e-01
-1.20524848e+00 6.42164111e-01 5.47167730e+00 1.33842206e+00
-1.16224575e+00 4.58800234e-02 2.75879592e-01 1.29938498e-01
-1.45424709e-01 -2.71456242e-01 -6.51575506e-01 3.02401155e-01
5.22334874e-01 -1.28258884e-01 3.75052035e-01 9.37067688e-01
-3.27762425e-01 -2.44357973e-01 -1.09913290e+00 1.02751529e+00
4.94214058e-01 -1.14970815e+00 3.01551878e-01 3.83041918e-01
6.29629731e-01 -3.57422560e-01 2.56048590e-01 -1.42482668e-02
-9.41232666e-02 -1.23362052e+00 1.14170337e+00 4.52719837e-01
7.03584015e-01 -6.34585500e-01 9.54463005e-01 -3.26709509e-01
-1.47273850e+00 -2.66079217e-01 -5.60927019e-02 3.13019991e-01
-2.37205680e-02 3.21343571e-01 -5.48988104e-01 8.97480786e-01
8.35194051e-01 5.53460479e-01 -9.93298709e-01 1.18644476e+00
6.55520260e-02 6.09295331e-02 -4.13728029e-01 -3.72514367e-01
3.66226584e-01 -2.74165481e-01 2.86904961e-01 8.60644698e-01
5.36483824e-01 -6.49463952e-01 2.26191729e-01 6.67169154e-01
6.36797547e-02 6.58690989e-01 -4.02881354e-01 -5.77611476e-02
2.91999727e-01 1.26759028e+00 -1.35223532e+00 -4.70526427e-01
-4.13623393e-01 5.39232373e-01 -1.22533210e-01 3.74279588e-01
-6.74458802e-01 -6.43447697e-01 -1.62713807e-02 3.62370968e-01
2.16265872e-01 2.61949986e-01 -3.49038899e-01 -4.62771475e-01
3.20798725e-01 -6.58800781e-01 6.03276789e-01 -7.50089467e-01
-1.40849555e+00 8.32680464e-01 2.02787533e-01 -1.75992858e+00
1.87228695e-01 -1.17456603e+00 5.98172583e-02 4.82488275e-01
-1.17195797e+00 -1.47924495e+00 -4.18414176e-01 2.54161470e-03
2.86565393e-01 -5.99163651e-01 6.15835249e-01 8.28980923e-01
-1.34474877e-02 2.94392735e-01 4.31456685e-01 2.79697273e-02
7.86919653e-01 -1.27396238e+00 -3.67705196e-01 1.22103564e-01
3.24974060e-01 1.69227138e-01 6.05605841e-01 -4.62055802e-01
-5.55138946e-01 -6.08535886e-01 1.08991671e+00 -3.27046573e-01
9.02826905e-01 -3.29019949e-02 -8.42792988e-01 2.07492620e-01
7.20953941e-02 -4.87123877e-02 6.49562657e-01 -1.66205555e-01
-5.40024638e-01 -4.36740160e-01 -1.05717933e+00 1.19610883e-01
6.49227321e-01 -5.26701033e-01 -5.33900559e-01 2.94410348e-01
-1.14997670e-01 7.42492964e-04 -1.21499300e+00 2.57994920e-01
9.80216205e-01 -9.43057299e-01 9.43829417e-01 -3.30021024e-01
2.81565577e-01 -5.03623426e-01 -2.81355474e-02 -8.87980044e-01
-1.35396078e-01 1.93471238e-01 4.62905020e-01 1.33167684e+00
3.80061179e-01 -6.44832671e-01 1.00883925e+00 4.76094009e-03
-1.12221152e-01 -5.02719462e-01 -9.46084917e-01 -1.25814998e+00
-1.20771982e-01 -2.02873737e-01 6.97721243e-01 9.49394584e-01
-4.51817602e-01 3.36932242e-01 -1.23906903e-01 -4.24370170e-02
6.56457424e-01 4.77183640e-01 2.96047330e-01 -2.08099198e+00
-5.47088273e-02 -8.80477667e-01 -6.94660366e-01 -3.47812861e-01
2.78696030e-01 -1.23713028e+00 -1.91795334e-01 -1.61062646e+00
-2.68956330e-02 -5.67208111e-01 -5.92589676e-01 5.80191016e-01
5.49530327e-01 6.53515935e-01 2.44854540e-01 4.25965995e-01
-5.19845605e-01 6.09734803e-02 1.00599873e+00 -7.39015877e-01
-1.93146858e-02 -1.21790655e-01 -3.23080003e-01 8.14832509e-01
6.76352143e-01 -7.11527050e-01 1.36884581e-02 -9.49162170e-02
4.88460392e-01 -2.99413800e-01 2.34204292e-01 -1.29525375e+00
-4.89323102e-02 9.38568637e-02 2.57579207e-01 -6.29090071e-01
1.83364645e-01 -1.48622370e+00 6.21023297e-01 5.58728158e-01
-2.10304692e-01 -1.36807293e-01 6.12336919e-02 3.69623542e-01
-5.30087531e-01 -7.22735286e-01 7.19924390e-01 -1.12979710e-01
-1.06256187e+00 -4.10712697e-02 -3.65591854e-01 -2.03631014e-01
1.05411720e+00 -5.91101170e-01 -3.78090650e-01 -3.35137993e-02
-1.04876244e+00 -2.39218086e-01 5.18562913e-01 4.23760980e-01
6.13733009e-02 -1.56914914e+00 -3.04300636e-01 -1.81696713e-01
7.12100148e-01 -5.76155961e-01 7.97704794e-03 4.90796417e-01
-1.10857844e+00 5.39192021e-01 -7.26359844e-01 -4.24141824e-01
-1.44532895e+00 6.94614589e-01 3.04233283e-01 -3.99854034e-01
-3.06615472e-01 1.85955182e-01 3.48849781e-02 -2.56724656e-01
3.77452731e-01 -3.85565907e-02 -9.00651991e-01 9.37214792e-01
-9.40433443e-02 6.61692202e-01 4.55410391e-01 -1.03119230e+00
-4.23011035e-01 1.30841768e+00 2.42588326e-01 1.01197727e-01
1.31784594e+00 8.33140314e-02 -2.83380717e-01 5.05181432e-01
1.31281161e+00 1.00745708e-01 -6.18803084e-01 -1.45895526e-01
4.33564067e-01 -2.71597594e-01 -3.22274774e-01 -1.04866445e+00
-1.11454356e+00 7.42923617e-01 1.20751977e+00 6.32631063e-01
1.08834934e+00 2.52834439e-01 3.61233950e-01 5.88267088e-01
3.87344360e-01 -1.46060443e+00 1.19548909e-01 9.37235728e-02
8.36502492e-01 -1.14021075e+00 1.66555718e-01 -5.91978729e-01
-3.75790209e-01 1.38038969e+00 3.14234942e-01 -3.22444737e-02
5.20869374e-01 -5.61307147e-02 4.69128698e-01 -5.41729033e-01
1.18572198e-01 -2.71312237e-01 7.84002721e-01 5.57018280e-01
3.93064082e-01 5.45217246e-02 -8.50941300e-01 2.81766176e-01
-1.29571974e-01 -1.95644930e-01 3.28480989e-01 6.95160091e-01
-5.52402735e-01 -1.35638916e+00 -7.80600965e-01 5.50032258e-01
-5.16294360e-01 2.30193347e-01 -5.10783613e-01 1.11220074e+00
5.30301929e-01 7.72825480e-01 1.41194254e-01 -1.19459428e-01
4.36293721e-01 7.41941556e-02 2.57799238e-01 -3.54640692e-01
-8.25073063e-01 2.52761304e-01 1.04750112e-01 -1.83785260e-01
-6.38122141e-01 -4.52131093e-01 -9.74958897e-01 1.22433178e-01
-2.54646093e-01 2.66420841e-01 7.51806438e-01 8.33044648e-01
-3.41591418e-01 7.05565155e-01 4.40355659e-01 -6.75383091e-01
-2.49228969e-01 -8.05740654e-01 -9.85423446e-01 8.01380157e-01
-8.11481997e-02 -1.08480704e+00 -4.48893458e-01 1.02499641e-01]
|
[10.243022918701172, -0.10115228593349457]
|
cdc90c82-8435-46f4-bffa-0662c4cdd7ea
|
learning-word-embeddings-for-hyponymy-with
|
1710.02437
| null |
http://arxiv.org/abs/1710.02437v1
|
http://arxiv.org/pdf/1710.02437v1.pdf
|
Learning Word Embeddings for Hyponymy with Entailment-Based Distributional Semantics
|
Lexical entailment, such as hyponymy, is a fundamental issue in the semantics
of natural language. This paper proposes distributional semantic models which
efficiently learn word embeddings for entailment, using a recently-proposed
framework for modelling entailment in a vector-space. These models postulate a
latent vector for a pseudo-phrase containing two neighbouring word vectors. We
investigate both modelling words as the evidence they contribute about this
phrase vector, or as the posterior distribution of a one-word phrase vector,
and find that the posterior vectors perform better. The resulting word
embeddings outperform the best previous results on predicting hyponymy between
words, in unsupervised and semi-supervised experiments.
|
['James Henderson']
|
2017-10-06
| null | null | null | null |
['learning-word-embeddings']
|
['methodology']
|
[ 1.32869557e-01 2.75259078e-01 -7.06041276e-01 -6.40130758e-01
4.47675325e-02 -4.02378231e-01 9.46475625e-01 5.84561706e-01
-9.33027983e-01 3.13839912e-01 1.06613159e+00 -6.28106475e-01
-4.86333743e-02 -8.74786854e-01 -3.71531844e-01 -3.70472133e-01
9.56319943e-02 6.78014636e-01 -8.33369195e-02 -4.48821157e-01
5.16233981e-01 1.23016359e-02 -1.57454944e+00 2.85904318e-01
3.68386120e-01 4.94853705e-01 1.38108194e-01 4.29170609e-01
-6.41950548e-01 4.44546700e-01 -3.02217931e-01 -7.90130019e-01
-1.76282316e-01 -1.32497877e-01 -9.60175574e-01 -7.44714662e-02
1.23703331e-02 -1.54066458e-01 -2.12406710e-01 1.14179730e+00
-7.94934167e-04 1.76751301e-01 1.21837080e+00 -1.18667984e+00
-1.20110774e+00 7.95066178e-01 -1.52757064e-01 2.28722826e-01
6.42767727e-01 -4.07393605e-01 2.17356014e+00 -1.24497831e+00
5.51543057e-01 1.28651166e+00 6.49930894e-01 3.04306030e-01
-1.34752202e+00 -3.80096794e-03 -6.52483329e-02 5.78748405e-01
-1.36331296e+00 -7.82168806e-02 6.53547823e-01 -4.10369307e-01
1.80342138e+00 1.67555809e-01 7.01390088e-01 9.74103868e-01
5.20753801e-01 5.08827627e-01 5.34506738e-01 -7.73654461e-01
2.26429746e-01 1.36397198e-01 4.42971349e-01 4.41041201e-01
6.09938323e-01 -1.56007707e-01 -3.81360501e-01 -5.91129661e-01
3.41528237e-01 2.68159330e-01 4.78346013e-02 -4.42962915e-01
-9.32732046e-01 1.56560934e+00 7.18070380e-03 5.66368043e-01
-4.88562673e-01 3.53192270e-01 4.70475376e-01 2.81101704e-01
6.34852290e-01 3.67798537e-01 -6.45398080e-01 -7.37777501e-02
-7.36352026e-01 3.70633632e-01 8.77666295e-01 6.86800539e-01
9.33707654e-01 -2.23268002e-01 2.25470707e-01 7.33709157e-01
9.80584383e-01 1.10219710e-01 9.59986866e-01 -5.72151124e-01
1.28943831e-01 5.68547308e-01 3.29532027e-02 -9.51648474e-01
-2.19975576e-01 4.17256117e-01 -1.84156761e-01 -2.38080785e-01
-1.06591515e-01 2.92199582e-01 -5.39142847e-01 1.78774571e+00
2.13586837e-01 2.25501508e-01 3.60571265e-01 5.84301114e-01
6.45955145e-01 8.76734972e-01 2.68652439e-01 -2.99588680e-01
1.66930187e+00 -5.68498194e-01 -1.00973713e+00 -4.83860165e-01
1.12327123e+00 -9.55849349e-01 1.21615505e+00 -9.28948373e-02
-8.05670857e-01 -2.83021510e-01 -1.22527993e+00 -2.37513438e-01
-5.48930228e-01 -4.08230603e-01 6.98463798e-01 7.15093970e-01
-6.84515297e-01 6.39282346e-01 -3.56195271e-01 -4.87584919e-01
-1.79467443e-02 -2.62910202e-02 -5.38254857e-01 -2.74242133e-01
-1.69840956e+00 1.40136576e+00 6.60457313e-01 -3.84345204e-01
-2.96012938e-01 -4.83442038e-01 -1.55025852e+00 4.33954298e-02
-6.97057471e-02 -5.26749611e-01 8.50730360e-01 -4.74149913e-01
-8.01312447e-01 1.30358064e+00 -5.59346080e-01 -6.38997793e-01
-3.57521534e-01 -2.43167624e-01 -6.39770985e-01 -2.20950514e-01
3.15801531e-01 1.92683682e-01 9.74063396e-01 -8.64195406e-01
-2.17623249e-01 -3.44668984e-01 2.06520468e-01 3.76093686e-01
-6.22823596e-01 1.49415404e-01 3.26846063e-01 -6.28669083e-01
8.86982977e-02 -5.79523742e-01 -2.41045445e-01 -1.32438198e-01
-2.08394617e-01 -1.01978230e+00 4.30375397e-01 -2.69344151e-01
1.45596731e+00 -2.20618677e+00 1.16712198e-01 2.88077116e-01
2.91614890e-01 7.60646239e-02 -1.43488839e-01 8.48973334e-01
-4.88358349e-01 1.44008175e-01 -4.00341451e-01 -2.57860214e-01
6.24084532e-01 1.05814338e+00 -5.61274052e-01 7.46886790e-01
2.32411772e-01 9.79764998e-01 -8.38957012e-01 -3.83451402e-01
3.32515001e-01 2.66486883e-01 -7.01359451e-01 1.78130120e-01
-1.99011788e-01 -8.51744592e-01 7.96829164e-02 -1.65744685e-02
6.30182683e-01 -4.06570248e-02 7.12023437e-01 -3.41492951e-01
1.81507900e-01 7.30128229e-01 -1.33030391e+00 1.64308321e+00
-5.05981028e-01 4.83769864e-01 -5.72238564e-01 -1.26560831e+00
6.60197198e-01 6.27688229e-01 1.72154889e-01 -2.23368675e-01
1.93102449e-01 1.27728611e-01 4.20321710e-02 -7.71147311e-01
8.82599473e-01 -9.66784000e-01 -4.51858282e-01 9.45389509e-01
3.78646880e-01 -2.14271516e-01 1.95600614e-01 4.18772280e-01
8.28929126e-01 -1.14679504e-02 9.79624450e-01 -5.83379388e-01
3.48300040e-01 -3.43440950e-01 2.91856050e-01 3.04370731e-01
-1.81792751e-01 3.23666781e-01 4.95951951e-01 -4.22524184e-01
-1.06709349e+00 -1.43649733e+00 -5.35148263e-01 1.41762459e+00
1.34748653e-01 -1.18636394e+00 -2.45966837e-01 -5.95744550e-01
2.72289306e-01 1.38701165e+00 -7.29858220e-01 -5.16449988e-01
-2.80746043e-01 -5.99083602e-01 6.89770639e-01 7.56371200e-01
-7.63445199e-01 -1.09099126e+00 -5.31186581e-01 2.59580106e-01
-1.83808371e-01 -9.06847358e-01 -4.05355364e-01 5.13251543e-01
-5.45511663e-01 -9.58172679e-01 -2.30218902e-01 -9.21417773e-01
3.78573269e-01 -1.08860835e-01 1.32242632e+00 2.20959157e-01
-4.96356696e-01 3.50905329e-01 -3.58098567e-01 -1.80017650e-01
-2.78509706e-01 -6.12630904e-01 6.19502604e-01 -2.06891224e-01
1.31569564e+00 -3.95737797e-01 -1.17317684e-01 -2.85319656e-01
-1.38849592e+00 -5.97733080e-01 1.32069588e-01 1.07623506e+00
3.46452504e-01 -1.37503371e-01 1.79043040e-01 -9.59154546e-01
1.08153915e+00 -9.20320809e-01 2.53628761e-01 1.40313745e-01
-7.98084676e-01 5.71860075e-01 4.23753746e-02 -3.73739779e-01
-7.96392977e-01 -4.71345216e-01 -6.54738307e-01 1.53831001e-02
-3.14855278e-01 8.50271463e-01 -3.94273400e-02 7.39362597e-01
6.65336668e-01 1.30772486e-01 1.12201413e-02 -1.65934399e-01
1.20269597e+00 5.40221155e-01 9.68703181e-02 -6.32948875e-01
5.05265296e-01 3.55196506e-01 -1.77323341e-01 -1.31338060e+00
-7.93501139e-01 -1.06806123e+00 -6.96637809e-01 4.77495790e-01
9.90151525e-01 -7.92532027e-01 -9.43048969e-02 -3.75681698e-01
-1.62071574e+00 5.72827578e-01 -6.45567715e-01 7.44491160e-01
-5.89856863e-01 7.61381686e-01 -6.22235358e-01 -5.89578748e-01
-7.05797672e-02 -7.08064675e-01 7.89283872e-01 -4.38831836e-01
-1.43005466e+00 -1.73802686e+00 6.14938021e-01 -1.29020944e-01
2.73746490e-01 -4.25197303e-01 1.53517807e+00 -1.23116624e+00
4.43200827e-01 -2.95072675e-01 -2.96646915e-02 4.25902903e-01
2.71578074e-01 -1.03118397e-01 -9.48800325e-01 9.37565193e-02
3.18901211e-01 -4.61037725e-01 9.73635197e-01 2.62119234e-01
4.98471439e-01 -4.26982403e-01 -2.33436704e-01 9.10683647e-02
1.57791543e+00 -4.72869366e-01 6.37507975e-01 9.68880579e-03
6.35359585e-01 9.59385931e-01 4.15113658e-01 5.54218173e-01
2.79198915e-01 1.64872333e-01 4.65973228e-01 4.59592611e-01
3.71131063e-01 -3.34837347e-01 4.01078522e-01 1.07638121e+00
2.51408517e-01 -1.31705493e-01 -6.25761747e-01 9.37814951e-01
-1.81498873e+00 -1.00469208e+00 -4.51178640e-01 1.67967904e+00
9.45125699e-01 3.40591036e-02 -1.73512295e-01 4.91616726e-01
4.50976074e-01 6.60352528e-01 -2.91000530e-02 -1.00242949e+00
-4.38562818e-02 7.89029717e-01 1.18519254e-01 1.00238502e+00
-7.08219051e-01 1.07802641e+00 7.19666481e+00 7.52490044e-01
-2.27055296e-01 3.77884895e-01 -3.00688535e-01 1.80245772e-01
-1.37065113e+00 3.47143769e-01 -5.73029041e-01 1.84266955e-01
8.73581529e-01 -5.07025838e-01 -1.07976243e-01 8.12105656e-01
-2.03493521e-01 1.14602946e-01 -1.35033739e+00 7.32889175e-01
5.20405710e-01 -1.25910139e+00 4.50115412e-01 9.05240551e-02
3.68325084e-01 -7.97935054e-02 -8.58958289e-02 1.65998653e-01
7.49568790e-02 -1.13164914e+00 4.25111592e-01 9.05998349e-02
5.51418304e-01 -9.14462447e-01 1.00869799e+00 2.27731019e-01
-1.18622398e+00 3.69003594e-01 -9.64175105e-01 -6.16811156e-01
4.66328084e-01 6.95166647e-01 -5.45784295e-01 1.05966397e-01
1.94461286e-01 8.82915795e-01 -1.21583953e-01 2.88847685e-01
-7.70587325e-01 4.92836982e-01 -1.14343926e-01 -4.24149930e-01
3.47862512e-01 -2.08051234e-01 4.67936128e-01 1.45855284e+00
8.22447315e-02 -3.05374533e-01 -1.23029925e-01 9.81848001e-01
1.60712481e-01 3.65847707e-01 -1.16777396e+00 -3.46618831e-01
3.69030684e-01 1.02755582e+00 -4.61308211e-01 -4.68007654e-01
-6.30428791e-01 1.22164643e+00 3.06787193e-01 1.20692231e-01
-8.15724432e-01 -4.71029192e-01 1.42840755e+00 -8.04912671e-02
3.27492177e-01 -2.56596118e-01 -1.25289053e-01 -1.10978496e+00
-8.12851116e-02 -1.18965134e-01 4.15184736e-01 -4.92518485e-01
-1.89394176e+00 2.41760716e-01 1.27723143e-01 -7.55120635e-01
-5.64254940e-01 -1.07905042e+00 -8.52217138e-01 9.29974198e-01
-1.31342590e+00 -7.59699404e-01 7.25320101e-01 4.87422228e-01
4.77742553e-01 -1.01174176e-01 1.49665475e+00 -1.65880129e-01
1.48180053e-01 4.13198054e-01 -6.11822642e-02 -2.75852680e-01
5.89729249e-01 -1.41090751e+00 6.37761891e-01 5.77007174e-01
9.98754740e-01 1.17878270e+00 1.05601370e+00 -5.25099695e-01
-1.01824284e+00 -7.49182343e-01 2.06819320e+00 -7.48154581e-01
1.10827255e+00 -2.12180287e-01 -9.70495462e-01 8.27431142e-01
5.54261625e-01 -1.48736998e-01 1.64429855e+00 6.54360175e-01
-8.24711740e-01 7.45909929e-01 -1.03144336e+00 6.18556917e-01
8.67441535e-01 -1.13007319e+00 -1.88554025e+00 2.27023154e-01
1.18123841e+00 4.61389512e-01 -6.81670904e-01 1.15161143e-01
5.20409942e-01 -6.61780000e-01 1.02470994e+00 -1.28183663e+00
7.16007888e-01 -8.42284188e-02 -8.59240174e-01 -1.53028071e+00
-4.81653988e-01 5.38665913e-02 -2.70400882e-01 8.13277781e-01
4.03581411e-01 -5.78666270e-01 6.05330288e-01 3.20209652e-01
1.61431879e-01 -6.46134198e-01 -1.00206530e+00 -7.60789990e-01
3.66962910e-01 -1.02509892e+00 5.03411770e-01 1.08095360e+00
8.46738636e-01 8.19629788e-01 -2.08546221e-01 -1.53181285e-01
6.61043882e-01 -6.57020062e-02 -3.60969454e-02 -1.22789586e+00
-3.49052876e-01 -4.04903382e-01 -8.51339400e-01 -1.08079517e+00
8.37222695e-01 -1.36702001e+00 2.96159722e-02 -1.67093933e+00
6.94280043e-02 2.59109616e-01 -1.94812089e-01 1.43539399e-01
-2.69063979e-01 3.29537898e-01 -3.13667208e-01 -9.46419239e-02
-4.11414057e-01 7.75791168e-01 4.03447241e-01 8.87648668e-03
4.27273035e-01 -4.63051111e-01 -6.63523376e-01 1.18527842e+00
6.95282936e-01 -7.59322107e-01 -5.75504601e-01 -4.32176769e-01
8.54706705e-01 -5.40028036e-01 1.13291919e-01 -3.84410135e-02
-1.09781504e-01 -1.98591650e-01 -9.76666808e-03 -4.17807043e-01
4.36585695e-01 -8.37397933e-01 -4.75932509e-01 4.98387098e-01
-5.41630208e-01 4.20247227e-01 -1.61440909e-01 8.02896500e-01
-2.51649559e-01 -9.85035360e-01 3.26407641e-01 -1.55297652e-01
-1.04129040e+00 -5.00704870e-02 -8.91335964e-01 4.58198279e-01
9.28031743e-01 -1.28264040e-01 1.63770109e-01 -1.80243298e-01
-7.11589158e-01 -1.61084101e-01 2.43399948e-01 6.60027981e-01
1.26021886e+00 -1.73524332e+00 -6.43831193e-01 4.16767478e-01
7.39106596e-01 -6.73874378e-01 -1.40322492e-01 3.17371279e-01
-2.31173575e-01 2.69480705e-01 -1.60697307e-02 -3.01513582e-01
-1.25496340e+00 6.02792084e-01 -2.25796044e-01 -5.01834555e-03
-4.13248986e-01 1.05021799e+00 1.61291063e-02 -7.61174619e-01
-1.37476876e-01 -3.77866834e-01 -4.06850368e-01 3.08334976e-01
5.19669294e-01 1.06578752e-01 -2.34064817e-01 -1.03661108e+00
-5.75303972e-01 3.79668653e-01 -1.10799618e-01 -3.26590359e-01
1.12579727e+00 -1.11804388e-01 -6.69624269e-01 1.02155244e+00
1.68526447e+00 -1.27868131e-01 -1.34552225e-01 -5.04435480e-01
4.35536832e-01 -4.54421282e-01 -2.28120297e-01 2.44306669e-01
1.20900990e-03 7.99206018e-01 6.44524097e-02 3.14242989e-01
8.19545910e-02 6.12301111e-01 7.85930753e-01 3.38396400e-01
9.87257957e-02 -1.05158138e+00 1.35745123e-01 8.91589701e-01
5.49762905e-01 -1.03884792e+00 1.02012001e-01 -2.54788607e-01
-6.78039730e-01 1.08048856e+00 9.44865346e-02 -5.20264149e-01
1.28791189e+00 9.56291258e-02 -2.14865692e-02 -5.44540167e-01
-8.21680248e-01 -3.42621893e-01 3.33757430e-01 6.26472890e-01
9.65017915e-01 4.77765590e-01 -8.21043074e-01 7.83899665e-01
-3.35682571e-01 -6.74271405e-01 3.61162603e-01 9.27576780e-01
-5.89377463e-01 -1.21566880e+00 1.29520604e-02 6.38853908e-01
-5.42283833e-01 -7.43811429e-01 -2.16482043e-01 3.74662399e-01
2.62231022e-01 9.96827304e-01 4.03174430e-01 -2.53712505e-01
6.81973770e-02 5.08331418e-01 6.74089253e-01 -1.10062647e+00
-4.69650105e-02 -2.90024102e-01 2.27844507e-01 -4.49912637e-01
-3.89902145e-01 -3.82647008e-01 -1.48948526e+00 -3.53540927e-01
-4.28599626e-01 5.07843137e-01 5.27605534e-01 1.54804611e+00
-2.41029978e-01 1.44325152e-01 2.85108060e-01 -4.58847553e-01
-9.20439780e-01 -9.90118742e-01 -9.92684066e-01 9.68809664e-01
-2.77084280e-02 -4.44387436e-01 -6.81098163e-01 -3.78640220e-02]
|
[10.398630142211914, 8.80810546875]
|
15d422c1-53f7-4624-b2c9-647b851072cd
|
pose-mum-reinforcing-key-points-relationship
|
2203.07837
| null |
https://arxiv.org/abs/2203.07837v1
|
https://arxiv.org/pdf/2203.07837v1.pdf
|
Pose-MUM : Reinforcing Key Points Relationship for Semi-Supervised Human Pose Estimation
|
A well-designed strong-weak augmentation strategy and the stable teacher to generate reliable pseudo labels are essential in the teacher-student framework of semi-supervised learning (SSL). Considering these in mind, to suit the semi-supervised human pose estimation (SSHPE) task, we propose a novel approach referred to as Pose-MUM that modifies Mix/UnMix (MUM) augmentation. Like MUM in the dense prediction task, the proposed Pose-MUM makes strong-weak augmentation for pose estimation and leads the network to learn the relationship between each human key point much better than the conventional methods by adding the mixing process in intermediate layers in a stochastic manner. In addition, we employ the exponential-moving-average-normalization (EMAN) teacher, which is stable and well-suited to the SSL framework and furthermore boosts the performance. Extensive experiments on MS-COCO dataset show the superiority of our proposed method by consistently improving the performance over the previous methods following SSHPE benchmark.
|
['Jin Young Choi', 'Nojun Kwak', 'Jongkeun Na', 'Jaeseung Lim', 'Hwijun Lee', 'Jongmok Kim']
|
2022-03-15
| null | null | null | null |
['semi-supervised-human-pose-estimation']
|
['computer-vision']
|
[ 9.27794352e-02 3.46477598e-01 -5.97553253e-02 -4.74671066e-01
-8.96109521e-01 -1.65470824e-01 7.30658472e-01 -1.23208284e-01
-7.00187743e-01 6.11067951e-01 2.20605239e-01 1.90532357e-01
5.86942807e-02 -3.30084413e-01 -9.93846059e-01 -8.18421721e-01
2.51499772e-01 7.19981194e-01 4.03002799e-01 -2.98582613e-01
-4.18532938e-01 2.61537224e-01 -1.44466209e+00 -2.43242145e-01
9.89192963e-01 9.66606796e-01 3.05374175e-01 3.36940736e-01
2.67476708e-01 7.51493096e-01 -3.53637189e-01 -5.55032432e-01
2.43590683e-01 -2.86298394e-01 -7.45431483e-01 2.66759396e-01
5.86900234e-01 -2.54951447e-01 -2.03619614e-01 7.14757502e-01
7.49859571e-01 3.27522188e-01 6.03472352e-01 -1.15086269e+00
-4.23270650e-02 5.75684369e-01 -9.64874566e-01 -1.84937447e-01
2.47969627e-01 -1.71789199e-01 7.87301600e-01 -1.12377083e+00
4.51597720e-01 1.30589366e+00 9.00811315e-01 5.01454830e-01
-1.12587059e+00 -6.34617329e-01 2.15043411e-01 1.57917514e-02
-1.56020129e+00 -3.69858354e-01 8.49481702e-01 -1.98591903e-01
1.69575065e-01 1.12580277e-01 5.67679465e-01 1.28886700e+00
-9.34932232e-02 1.27253938e+00 1.17205715e+00 -4.13609028e-01
-1.07792169e-01 2.25043461e-01 -2.81531736e-02 7.17481315e-01
9.00773052e-03 -1.91879973e-01 -5.08693993e-01 7.08075091e-02
7.54745424e-01 2.19490603e-02 -1.14082471e-01 -9.00391459e-01
-1.51608074e+00 5.10846019e-01 6.60024285e-01 -8.37515071e-02
-4.34564203e-01 1.95878461e-01 4.18836176e-01 -1.97253272e-01
4.96623635e-01 1.63212851e-01 -4.77740526e-01 1.60601899e-01
-1.24311280e+00 3.33898097e-01 3.51130486e-01 9.64186013e-01
5.13761222e-01 4.31732237e-02 -3.06282490e-01 9.22818005e-01
3.59838635e-01 5.42535067e-01 4.21024084e-01 -6.56231105e-01
4.53817606e-01 3.35512072e-01 1.64721623e-01 -7.54443288e-01
-5.16080737e-01 -1.09440529e+00 -1.02392316e+00 8.76920894e-02
4.18240756e-01 -1.87794313e-01 -1.01557159e+00 2.10432339e+00
8.26860726e-01 4.49011505e-01 -8.55180323e-02 8.39687049e-01
9.23227310e-01 5.11667073e-01 2.70956635e-01 -2.10038483e-01
1.22754085e+00 -1.40341055e+00 -8.85653436e-01 -2.22012401e-01
5.87522924e-01 -7.18345404e-01 1.07875967e+00 3.01145613e-01
-9.89178717e-01 -1.05185914e+00 -1.04485476e+00 4.20795307e-02
1.45292077e-02 6.59591079e-01 5.49998879e-01 3.48705471e-01
-6.18638813e-01 4.88578290e-01 -1.06202424e+00 -1.40572473e-01
4.32001054e-01 3.62481266e-01 -6.13088012e-01 4.82761152e-02
-1.05364418e+00 8.43156636e-01 4.93922710e-01 5.05843461e-01
-8.85488153e-01 -4.88127112e-01 -1.01224637e+00 -2.54650086e-01
7.23526061e-01 -7.26850808e-01 1.06472909e+00 -7.74005949e-01
-1.74631631e+00 6.35150194e-01 8.16615745e-02 -4.10164595e-01
9.58886445e-01 -7.13454843e-01 2.23700069e-02 5.67222089e-02
4.30148691e-02 9.99515057e-01 9.99432445e-01 -1.31699991e+00
-4.50483859e-01 -5.05605102e-01 -2.00503692e-01 4.98920560e-01
-4.13316905e-01 -1.12708174e-01 -8.49117279e-01 -9.02612686e-01
2.94838905e-01 -1.31557691e+00 -4.27944213e-01 -1.42003775e-01
-5.25382400e-01 -3.79787326e-01 7.45381296e-01 -7.50318408e-01
1.24912882e+00 -1.86932099e+00 4.56123173e-01 1.94510654e-01
1.17795520e-01 4.79990125e-01 -2.31508762e-02 1.74659953e-01
-2.64202476e-01 -6.74665511e-01 -3.58968168e-01 -1.13233542e+00
-4.71708477e-02 1.44965917e-01 1.55270293e-01 7.31897473e-01
1.03527203e-01 1.04367280e+00 -7.75007427e-01 -8.21675897e-01
3.52340400e-01 6.52432024e-01 -3.91506106e-01 5.32838106e-01
-1.12113699e-01 7.68547416e-01 -2.83478647e-01 3.16131204e-01
7.46511281e-01 -3.31411451e-01 -5.81562556e-02 -3.45528811e-01
1.88490599e-01 -1.11598574e-01 -1.66239464e+00 1.96703279e+00
-3.89614016e-01 3.92695777e-02 9.70945805e-02 -8.18075180e-01
9.06013548e-01 3.55710983e-01 3.08025658e-01 -1.47925481e-01
2.03820601e-01 8.49030018e-02 -1.99326858e-01 -2.99329281e-01
2.47893035e-01 -2.60842871e-02 -7.93364272e-03 2.26010069e-01
4.22362447e-01 1.07603133e-01 8.51126611e-02 3.08404505e-01
5.09541988e-01 7.55338609e-01 2.69115508e-01 -2.06153154e-01
7.94087172e-01 -4.66446072e-01 6.78801835e-01 3.05964082e-01
-1.73250869e-01 8.42474818e-01 1.51319951e-01 -1.66782573e-01
-8.22462559e-01 -9.11489129e-01 -5.16935214e-02 1.31746924e+00
1.66129500e-01 -4.70187545e-01 -9.95988071e-01 -9.94194627e-01
-3.29256743e-01 2.91112781e-01 -6.16357327e-01 7.44815962e-03
-5.94781697e-01 -8.41356695e-01 2.94514149e-01 8.15585077e-01
5.72881281e-01 -9.67934072e-01 -1.49223030e-01 7.94840679e-02
-3.81044596e-01 -1.29517233e+00 -6.46473467e-01 2.86079943e-01
-5.79359353e-01 -8.64603221e-01 -1.31617367e+00 -9.69328761e-01
8.12149942e-01 -8.36106688e-02 6.98414981e-01 -3.38197201e-01
1.40167922e-01 7.40402117e-02 -3.89480472e-01 -2.83877283e-01
-6.06356375e-02 3.56507629e-01 3.12690049e-01 4.73597527e-01
-5.78607842e-02 -5.61218917e-01 -5.24508357e-01 3.63219678e-01
-6.43320203e-01 2.45154306e-01 9.00745213e-01 8.01118016e-01
7.68147647e-01 -1.13807410e-01 5.09803712e-01 -1.05987501e+00
1.98259920e-01 -8.70125443e-02 -4.25037056e-01 2.20493227e-01
-6.18288934e-01 1.68865219e-01 5.58500350e-01 -7.42409647e-01
-1.22493458e+00 7.31546104e-01 -4.09123272e-01 -5.39635122e-01
-1.93965867e-01 1.76131114e-01 -4.67656493e-01 -2.58221120e-01
5.37960768e-01 3.12370270e-01 4.11102995e-02 -7.60964692e-01
4.22455162e-01 4.11772758e-01 9.69313562e-01 -5.05801499e-01
1.28284228e+00 3.30910474e-01 4.47096750e-02 -5.74880362e-01
-1.10814226e+00 -5.05206764e-01 -1.07168555e+00 -1.46197319e-01
8.90094876e-01 -1.20308387e+00 -6.66099608e-01 7.38168001e-01
-9.15086508e-01 -1.93305716e-01 -1.87167972e-01 7.06924736e-01
-5.45394957e-01 5.01377523e-01 -5.36782324e-01 -8.90153408e-01
-3.63675922e-01 -1.28239155e+00 1.21990371e+00 1.91459358e-01
-1.05610691e-01 -7.67474711e-01 1.56056151e-01 5.36895990e-01
1.31430626e-01 4.79756892e-01 1.18318215e-01 -7.78588295e-01
-1.56261519e-01 -3.18996966e-01 -7.87078310e-03 5.81936896e-01
-1.29477084e-01 -2.72487968e-01 -9.40510631e-01 -4.56452817e-01
-1.49865374e-01 -7.26448715e-01 8.75619888e-01 3.48328501e-01
1.07237923e+00 -1.08870529e-01 -1.86037853e-01 7.74194837e-01
9.65678453e-01 -5.21830559e-01 3.67622137e-01 1.84599012e-01
1.12636447e+00 7.06879258e-01 9.05794978e-01 4.42743421e-01
6.29212618e-01 9.22419131e-01 2.44191483e-01 -6.43091321e-01
-1.24936573e-01 -5.93396544e-01 3.54069889e-01 9.95589733e-01
-2.03008071e-01 7.58441463e-02 -6.51225388e-01 2.57900238e-01
-1.90860426e+00 -5.44817686e-01 -1.08999945e-01 2.09994483e+00
9.30504262e-01 2.97441751e-01 5.56218326e-01 2.55031466e-01
6.81332290e-01 3.34642828e-01 -2.51279831e-01 3.22987258e-01
2.32599843e-02 2.08429933e-01 4.27111298e-01 4.02561247e-01
-1.49167216e+00 1.00249255e+00 5.23335171e+00 1.23143411e+00
-8.13217700e-01 2.80032098e-01 5.04653394e-01 2.40782097e-01
2.41257057e-01 -2.61383593e-01 -1.02046943e+00 3.44227970e-01
6.50727093e-01 4.92834836e-01 9.69579145e-02 1.01352489e+00
1.69431627e-01 -1.99846383e-02 -9.97980952e-01 1.01964283e+00
1.47324592e-01 -7.60159791e-01 -8.76844600e-02 -2.43035004e-01
7.15912759e-01 -2.96467423e-01 4.24908735e-02 3.41371477e-01
5.67721613e-02 -6.59544230e-01 9.38446701e-01 3.89868051e-01
7.34841585e-01 -1.00863862e+00 9.58968520e-01 7.52136827e-01
-1.39700878e+00 1.08579464e-01 -1.61939636e-01 3.04427177e-01
3.08899224e-01 3.89502913e-01 -6.48706317e-01 6.11692369e-01
6.40031040e-01 5.88495255e-01 -8.97319257e-01 9.00006831e-01
-6.53114557e-01 6.08706892e-01 -4.40517873e-01 2.26795658e-01
2.42139339e-01 -1.22130893e-01 4.98999685e-01 1.18250489e+00
-2.20381439e-01 -1.04679875e-01 4.41176683e-01 4.14681405e-01
1.63105354e-02 2.39392236e-01 -4.60894108e-02 2.35090375e-01
1.96971774e-01 1.55297315e+00 -8.50540400e-01 -2.96702802e-01
-6.61876276e-02 1.12246215e+00 3.12432230e-01 7.95347840e-02
-9.65353012e-01 -3.12779427e-01 -5.39511964e-02 2.93547839e-01
2.63663858e-01 -2.29865611e-01 8.62551332e-02 -9.21385527e-01
1.96558297e-01 -1.07114291e+00 3.41352254e-01 -5.64702928e-01
-1.00748241e+00 8.47607374e-01 3.08360547e-01 -1.43955445e+00
-2.61657655e-01 -2.50901669e-01 -4.37088400e-01 5.11225462e-01
-1.41150510e+00 -1.92125583e+00 -3.78291190e-01 6.70217633e-01
4.89956558e-01 4.36367095e-02 5.57326913e-01 4.67140466e-01
-7.24794626e-01 9.16920900e-01 -1.84395596e-01 1.23297632e-01
1.03619146e+00 -1.28248966e+00 2.96478599e-01 8.43984067e-01
2.08189324e-01 6.02312982e-01 7.93548405e-01 -6.01209223e-01
-9.88315880e-01 -1.19842589e+00 7.16927230e-01 -4.80321705e-01
3.71032834e-01 -6.86326802e-01 -7.57838666e-01 7.37789035e-01
7.54657164e-02 2.35697508e-01 4.53569323e-01 4.41521369e-02
-1.40621856e-01 -2.52961069e-01 -9.10154223e-01 3.30849648e-01
9.22091126e-01 -3.13022166e-01 -7.07568467e-01 4.88974303e-01
7.74123073e-01 -6.91485882e-01 -6.79477394e-01 7.13537574e-01
4.97475177e-01 -6.69070125e-01 1.09648311e+00 -4.32097048e-01
2.79235393e-01 -4.48209763e-01 1.14982985e-01 -1.17720449e+00
-2.47893795e-01 -7.60537207e-01 -2.45927036e-01 1.45604622e+00
2.71293759e-01 -2.25101933e-01 9.87965286e-01 1.77013859e-01
-1.38926674e-02 -1.09165549e+00 -5.72855473e-01 -6.91800714e-01
-2.05408201e-01 -2.15416372e-01 3.23972702e-01 7.70831823e-01
-4.08781886e-01 4.93687958e-01 -1.00103772e+00 3.02229494e-01
1.01029491e+00 -2.05678269e-01 1.15188634e+00 -1.21828878e+00
-5.89038849e-01 1.76599145e-01 -2.85208583e-01 -1.38538373e+00
2.62160838e-01 -5.90273440e-01 2.65205711e-01 -1.23945343e+00
3.22684854e-01 -4.22946185e-01 -2.58529842e-01 3.54522616e-01
-6.09181702e-01 5.52047491e-01 1.11105278e-01 2.13563308e-01
-1.08769441e+00 9.82093453e-01 1.34229243e+00 2.43097216e-01
-2.51435071e-01 5.38800776e-01 -3.82183254e-01 9.48331892e-01
1.90719858e-01 -5.08877337e-01 -4.17658865e-01 3.11840139e-02
-1.00907378e-01 -1.26818866e-01 1.80738494e-01 -1.29617906e+00
3.57335031e-01 1.86057553e-01 5.90586424e-01 -9.76457417e-01
4.89251196e-01 -8.30556393e-01 7.45832622e-02 4.34915572e-01
-5.13777912e-01 -8.73582736e-02 -1.75610781e-01 6.31907225e-01
-2.00478241e-01 -3.34369834e-03 1.10388851e+00 -7.22887879e-03
-5.52702427e-01 5.19541562e-01 2.52520353e-01 1.40984371e-01
9.75179493e-01 3.47489864e-02 3.54269594e-01 -4.39573139e-01
-9.19765353e-01 4.48495299e-01 1.98821962e-01 3.12823236e-01
3.25336814e-01 -1.49435532e+00 -6.81230545e-01 1.24942943e-01
1.63845390e-01 5.04154742e-01 3.18209648e-01 1.18895984e+00
-2.54645079e-01 1.43852085e-01 -4.66252975e-02 -6.23171091e-01
-1.25806093e+00 4.92255151e-01 1.57020926e-01 -4.54595536e-01
-5.71602881e-01 1.16503918e+00 2.73253053e-01 -6.89936876e-01
7.32064903e-01 -2.35950742e-02 -4.09615159e-01 1.98224857e-01
4.71414983e-01 4.60363537e-01 -4.10501361e-02 -1.05622685e+00
-3.83224189e-01 5.80188572e-01 -7.11747557e-02 -8.67132246e-02
1.52094352e+00 -1.35288700e-01 1.67380929e-01 3.71189862e-01
1.14958906e+00 2.92669218e-02 -1.39881647e+00 -5.48735440e-01
-1.07142292e-01 -3.27156782e-02 -1.50818741e-02 -5.01509845e-01
-1.00163722e+00 8.78451347e-01 6.26926124e-01 -4.34877008e-01
8.84906828e-01 -6.45330846e-02 8.90171647e-01 2.61668563e-01
2.17528835e-01 -1.20253336e+00 3.09345514e-01 2.17070892e-01
7.73587823e-01 -1.35148883e+00 9.27853361e-02 -5.47877550e-01
-8.91199231e-01 7.80828595e-01 8.52263033e-01 -1.54092431e-01
5.50153613e-01 2.91391760e-01 1.22004829e-01 7.07085282e-02
-2.32784897e-01 -3.38129073e-01 6.93799973e-01 2.75147557e-01
3.61986518e-01 -1.63782969e-01 -3.29265088e-01 7.13764548e-01
-1.34163037e-01 -7.58824721e-02 -1.83185786e-02 8.73559892e-01
-2.80209988e-01 -1.09195161e+00 -4.76262569e-01 9.17550549e-02
-4.31780905e-01 1.31041050e-01 -2.83076018e-01 8.82345557e-01
3.11902106e-01 5.73169053e-01 -2.86304951e-01 -4.31155622e-01
3.87398273e-01 9.95904505e-02 4.10686970e-01 -6.72024429e-01
-5.79168499e-01 5.50094306e-01 -1.83897406e-01 -6.08009338e-01
-6.44252777e-01 -5.23873806e-01 -1.25890434e+00 7.19589368e-02
-6.89190865e-01 2.08117127e-01 5.08500338e-01 1.24790847e+00
2.20396183e-02 5.92317760e-01 6.62882924e-01 -1.08533049e+00
-8.29833686e-01 -1.23924327e+00 -4.98817474e-01 5.63960731e-01
2.52724081e-01 -9.63737369e-01 -3.20979774e-01 -4.47412543e-02]
|
[7.195310592651367, -0.751413106918335]
|
2a8e1115-09a3-4ce6-a53c-0f9bd27fd968
|
end-to-end-abnormality-detection-in-medical
| null | null |
https://openreview.net/forum?id=rk1FQA0pW
|
https://openreview.net/pdf?id=rk1FQA0pW
|
End-to-End Abnormality Detection in Medical Imaging
|
Deep neural networks (DNN) have shown promising performance in computer vision. In medical imaging, encouraging results have been achieved with deep learning for applications such as segmentation, lesion detection and classification. Nearly all of the deep learning based image analysis methods work on reconstructed images, which are obtained from original acquisitions via solving inverse problems (reconstruction). The reconstruction algorithms are designed for human observers, but not necessarily optimized for DNNs which can often observe features that are incomprehensible for human eyes. Hence, it is desirable to train the DNNs directly from the original data which lie in a different domain with the images. In this paper, we proposed an end-to-end DNN for abnormality detection in medical imaging. To align the acquisition with the annotations made by radiologists in the image domain, a DNN was built as the unrolled version of iterative reconstruction algorithms to map the acquisitions to images, and followed by a 3D convolutional neural network (CNN) to detect the abnormality in the reconstructed images. The two networks were trained jointly in order to optimize the entire DNN for the detection task from the original acquisitions. The DNN was implemented for lung nodule detection in low-dose chest computed tomography (CT), where a numerical simulation was done to generate acquisitions from 1,018 chest CT images with radiologists' annotations. The proposed end-to-end DNN demonstrated better sensitivity and accuracy for the task compared to a two-step approach, in which the reconstruction and detection DNNs were trained separately. A significant reduction of false positive rate on suspicious lesions were observed, which is crucial for the known over-diagnosis in low-dose lung CT imaging. The images reconstructed by the proposed end-to-end network also presented enhanced details in the region of interest.
|
['Kyungsang Kim', 'Dufan Wu', 'Bin Dong', 'Quanzheng Li']
|
2018-01-01
| null | null | null |
iclr-2018-1
|
['lung-nodule-detection']
|
['medical']
|
[ 4.05114889e-01 4.86121833e-01 4.72418070e-02 -3.87797982e-01
-6.91347122e-01 -2.36822248e-01 2.25623146e-01 -1.25754327e-01
-7.41974831e-01 3.37094665e-01 -2.47517489e-02 -5.15663087e-01
-1.15695909e-01 -7.49056578e-01 -6.11526430e-01 -7.97629833e-01
4.44171391e-02 7.72178590e-01 3.81244510e-01 2.16211528e-01
-1.69954479e-01 6.08854413e-01 -1.04041505e+00 5.55077016e-01
5.01671135e-01 9.21173751e-01 6.67495668e-01 9.10979390e-01
-6.15001656e-03 8.78906906e-01 -1.68468624e-01 4.38271761e-02
3.95955294e-01 -7.02256382e-01 -7.89000273e-01 3.66910934e-01
2.15416588e-02 -6.83921874e-01 -4.60976452e-01 1.14901769e+00
6.34920180e-01 -1.50833175e-01 7.20233262e-01 -5.51964760e-01
-3.14899296e-01 4.61695850e-01 -5.37923455e-01 4.58706379e-01
-6.24132305e-02 2.17703104e-01 5.12426019e-01 -9.14523840e-01
3.56055498e-01 7.95321822e-01 7.25947857e-01 5.67128360e-01
-7.54122674e-01 -3.73620242e-01 -4.66367543e-01 4.06047627e-02
-1.06525028e+00 1.17699049e-01 4.67405528e-01 -5.01837194e-01
4.95435059e-01 3.10201496e-01 7.00841963e-01 8.14945340e-01
2.74931580e-01 6.32087231e-01 7.56890297e-01 -3.82812083e-01
-1.35067925e-01 2.25786477e-01 -8.55579227e-02 8.46646249e-01
3.62269402e-01 3.97255063e-01 3.95370662e-01 1.14113577e-01
1.25656354e+00 4.86986756e-01 -3.77657235e-01 -3.28773826e-01
-1.27512729e+00 8.28736544e-01 9.35603917e-01 5.84022760e-01
-8.62941921e-01 -7.68153742e-03 3.37096244e-01 -1.76213682e-01
8.92317370e-02 2.09206432e-01 -9.35142860e-02 3.83754462e-01
-8.04890633e-01 -2.76805341e-01 7.17398345e-01 3.54595065e-01
1.65997803e-01 5.88337444e-02 -4.69863296e-01 6.48496389e-01
4.53661531e-01 2.35612497e-01 9.58947718e-01 -5.90973258e-01
2.09419549e-01 6.22501969e-01 -1.08027712e-01 -6.94748282e-01
-7.35098898e-01 -9.04017508e-01 -1.34462202e+00 2.25541204e-01
5.37104666e-01 -1.58805564e-01 -1.40474045e+00 1.20528710e+00
4.80611056e-01 1.34454697e-01 1.98311552e-01 1.34798431e+00
9.82157111e-01 4.96349066e-01 1.31960101e-02 -1.05502568e-01
1.48700380e+00 -1.01564443e+00 -5.87601542e-01 -1.83473974e-01
6.91785455e-01 -8.65553200e-01 8.81108046e-01 3.04404348e-01
-1.17830229e+00 -7.49321818e-01 -9.74419475e-01 7.83978328e-02
1.91948876e-01 4.35242385e-01 1.05602980e-01 4.98176962e-01
-9.24598336e-01 5.35890341e-01 -1.05333114e+00 -2.49228254e-01
6.46392882e-01 5.74392676e-01 -1.75551802e-01 -1.78949878e-01
-1.04863191e+00 9.85170305e-01 6.54323518e-01 4.37871575e-01
-1.34218240e+00 -4.97642338e-01 -4.78015929e-01 5.03024273e-02
4.80161875e-01 -8.20943117e-01 1.47888422e+00 -1.20838428e+00
-1.30215204e+00 1.08042967e+00 2.92163730e-01 -6.54805303e-01
8.67487371e-01 1.15861990e-01 -2.05545738e-01 2.78393626e-01
1.71658620e-01 4.55908924e-01 7.57118762e-01 -1.19635558e+00
-6.92741156e-01 -2.50749916e-01 -1.77711874e-01 2.53650844e-01
1.35719767e-02 -3.78457338e-01 -5.62790453e-01 -5.43243229e-01
4.62159902e-01 -8.08634818e-01 -5.73993742e-01 2.50160038e-01
-5.10983229e-01 1.92730010e-01 9.30860698e-01 -9.00547802e-01
8.15815330e-01 -2.10081959e+00 -8.97567123e-02 2.58644521e-01
5.28444827e-01 4.70599711e-01 1.41676643e-03 -3.16775560e-01
-4.78357166e-01 -6.23062178e-02 -4.29189414e-01 -7.46140480e-02
-5.80722153e-01 4.68158841e-01 3.91625315e-01 7.08590567e-01
-3.19614820e-02 1.01026690e+00 -8.02344561e-01 -7.23288417e-01
4.51941818e-01 3.55458140e-01 -4.51354384e-01 6.32320583e-01
-8.14551488e-02 9.28838313e-01 -4.13757622e-01 3.64999682e-01
6.30470991e-01 -5.11757016e-01 1.48735955e-01 -4.21096206e-01
5.04906401e-02 -2.35742956e-01 -8.97227287e-01 1.42254746e+00
-5.29779196e-01 4.37836409e-01 3.45670521e-01 -1.20459354e+00
5.51700175e-01 7.36371517e-01 7.98559666e-01 -7.48076379e-01
4.91255850e-01 3.00406992e-01 6.00097239e-01 -1.07655168e+00
-1.48271099e-01 -5.08726001e-01 3.34003031e-01 4.87269133e-01
-1.87432334e-01 -3.23916644e-01 -8.07632580e-02 -2.53687203e-01
1.00614834e+00 -5.08489847e-01 4.43266928e-01 5.25531247e-02
8.65375698e-01 2.39524975e-01 2.56947398e-01 8.04105341e-01
-1.59910321e-01 8.74612331e-01 3.52523327e-01 -7.11415231e-01
-1.15779603e+00 -1.10336089e+00 -2.99831271e-01 4.47048217e-01
1.42004769e-02 5.06620288e-01 -7.85412014e-01 -8.49265516e-01
-4.90842074e-01 6.23569250e-01 -5.66766798e-01 -9.26547796e-02
-7.02667177e-01 -7.90108263e-01 4.57689762e-01 5.12767851e-01
8.15201640e-01 -1.18701279e+00 -8.79421413e-01 3.32698494e-01
-1.11111671e-01 -1.12667060e+00 -2.63550818e-01 4.68303710e-01
-1.18780386e+00 -1.33828568e+00 -9.82232273e-01 -1.02370536e+00
1.04448223e+00 6.47493079e-02 9.77475107e-01 2.50250518e-01
-5.96442401e-01 2.65873581e-01 -6.90782741e-02 -3.45231175e-01
-9.05401826e-01 -6.04276918e-02 -3.01834673e-01 -5.56407310e-03
-2.98173521e-02 -2.58786261e-01 -8.42496574e-01 4.19197202e-01
-1.31117153e+00 1.99517697e-01 1.22440410e+00 1.19581258e+00
8.43002319e-01 2.66818911e-01 1.57714978e-01 -1.12675667e+00
4.42290038e-01 -3.77726197e-01 -5.44340014e-01 7.29759708e-02
-2.83568382e-01 2.61782743e-02 6.81316674e-01 -2.59556293e-01
-1.10686195e+00 5.35010040e-01 -5.52890718e-01 -4.71268207e-01
-4.08123374e-01 4.22375560e-01 3.30842227e-01 -1.05667561e-01
7.83613145e-01 2.76160568e-01 3.47771466e-01 -1.40350103e-01
5.72015382e-02 6.30287707e-01 5.88752091e-01 -2.61037145e-02
6.46428525e-01 4.58698303e-01 1.72085419e-01 -6.42619491e-01
-1.07629085e+00 -4.80702013e-01 -8.89428258e-01 -1.88243330e-01
1.34220636e+00 -6.74397826e-01 -3.56018633e-01 2.89783657e-01
-1.15537596e+00 -1.33324146e-01 -6.25556886e-01 9.44515467e-01
-2.87906528e-01 4.05329883e-01 -6.27373874e-01 -3.31833780e-01
-5.02262712e-01 -1.46345866e+00 8.25530827e-01 3.66493106e-01
1.32264957e-01 -1.12691379e+00 -1.32495731e-01 3.40426147e-01
4.52189386e-01 2.03031570e-01 1.10440743e+00 -9.40604210e-01
-5.38011014e-01 -4.46144521e-01 -4.74570811e-01 7.10615635e-01
2.06180558e-01 -4.32544053e-01 -9.33461428e-01 -2.09759369e-01
5.98902464e-01 5.61438780e-03 6.22138977e-01 9.20232356e-01
1.50424671e+00 -1.33180827e-01 -3.08494389e-01 7.83799827e-01
1.49004650e+00 4.00188595e-01 5.12901664e-01 7.66976178e-02
7.95763552e-01 2.78698027e-01 3.07627976e-01 2.28536695e-01
-2.72721410e-01 2.72299528e-01 9.92962360e-01 -8.30742717e-01
-3.54753256e-01 8.14092085e-02 -1.35792464e-01 5.95069349e-01
-1.15823753e-01 -2.41231605e-01 -1.02493262e+00 5.13360023e-01
-1.51333725e+00 -5.26572049e-01 -5.52930295e-01 2.01256371e+00
4.26668912e-01 2.30462998e-01 -1.94128200e-01 -4.96062171e-03
8.01945329e-01 -1.76701695e-01 -6.36191607e-01 -6.36399686e-02
3.76784652e-01 3.03853422e-01 6.85349941e-01 3.26281041e-01
-1.14040220e+00 2.91188300e-01 5.80750132e+00 6.98288798e-01
-1.28018713e+00 4.50365603e-01 8.11871409e-01 9.93283093e-02
1.42812923e-01 -3.81932139e-01 -3.61071706e-01 1.46611422e-01
5.51631570e-01 1.51739061e-01 -1.86063405e-02 8.67131829e-01
5.13818979e-01 -2.47759625e-01 -1.18365228e+00 9.09477353e-01
-1.82503134e-01 -1.20878220e+00 -2.50137690e-02 4.03160527e-02
7.24740803e-01 1.68102309e-01 1.88310534e-01 1.54202953e-01
-6.69070557e-02 -1.17108154e+00 1.66641980e-01 3.98275405e-01
7.27082491e-01 -5.62411964e-01 1.29450440e+00 7.27886856e-01
-8.89155447e-01 2.93348217e-03 -3.43871355e-01 2.53086269e-01
2.72781670e-01 5.39346218e-01 -1.64340651e+00 5.04204571e-01
4.19156730e-01 5.28458357e-01 -4.61881161e-01 1.03138947e+00
-2.53806800e-01 5.78205466e-01 -2.16011390e-01 6.24286495e-02
5.36711693e-01 -5.59387468e-02 4.68518317e-01 1.09305060e+00
4.44313765e-01 3.20178121e-01 1.55182242e-01 1.08748424e+00
-1.10056892e-01 -1.61840454e-01 -6.10473275e-01 1.91470906e-01
-2.49855250e-01 1.48186743e+00 -1.04499125e+00 -5.09923398e-01
-3.09909880e-01 9.81406629e-01 -1.97531462e-01 1.91457137e-01
-1.05995047e+00 1.09515183e-01 -1.72321975e-01 5.48584104e-01
1.91279456e-01 1.74441114e-01 -2.57328957e-01 -6.64860070e-01
-1.99048728e-01 -6.37474418e-01 4.65528816e-01 -6.56769574e-01
-1.24911809e+00 8.92083049e-01 -6.63464740e-02 -1.40466571e+00
-3.43193889e-01 -8.55980158e-01 -6.70133114e-01 8.19208264e-01
-1.28087139e+00 -7.81756163e-01 -6.34584785e-01 6.89316571e-01
6.52763665e-01 6.65918738e-02 5.76869786e-01 5.48407853e-01
-4.20421839e-01 1.37174875e-01 4.81950752e-02 4.22953010e-01
2.22126067e-01 -1.24313939e+00 -1.90633997e-01 8.96907330e-01
-9.97095257e-02 1.31230159e-02 2.31576860e-01 -4.78293300e-01
-8.93172622e-01 -1.34254563e+00 2.64913380e-01 -9.31688249e-02
2.19536781e-01 1.99262887e-01 -9.22174037e-01 6.07149720e-01
1.38102800e-01 6.03899121e-01 2.36675277e-01 -8.29974473e-01
7.01739132e-01 1.28072470e-01 -1.33916140e+00 3.02087933e-01
5.40185630e-01 -2.06316054e-01 -5.63814402e-01 6.59381151e-01
3.75605971e-01 -8.84729505e-01 -6.28961384e-01 5.22700608e-01
2.40124285e-01 -9.91759717e-01 1.19953704e+00 -3.73512745e-01
5.70870757e-01 -1.73503593e-01 8.47640932e-02 -1.03834140e+00
-2.98663974e-01 1.97085425e-01 2.94894695e-01 4.07371044e-01
4.43442672e-01 -5.04963219e-01 1.01964772e+00 3.47501218e-01
-4.03901845e-01 -8.71774793e-01 -9.42930400e-01 -2.46130496e-01
-1.53510094e-01 -3.61051500e-01 -6.25493154e-02 5.84448099e-01
-9.54108119e-01 1.63896129e-01 -3.27335596e-02 4.89098698e-01
6.25987232e-01 -2.28006199e-01 3.10004175e-01 -1.05394399e+00
-4.86725539e-01 -3.48227531e-01 -2.39672229e-01 -1.14120615e+00
-2.98768640e-01 -1.14840138e+00 1.28165677e-01 -1.82952118e+00
3.54670435e-01 -2.64456511e-01 -2.98439056e-01 1.25588566e-01
-4.81164753e-02 1.57166317e-01 -1.12054706e-01 3.73872072e-01
-2.04450727e-01 1.09341413e-01 1.93692362e+00 -1.70083940e-01
-4.12041470e-02 6.49313629e-01 -1.82229161e-01 1.03834176e+00
5.48102677e-01 -8.04842114e-01 -3.19669873e-01 -4.32899863e-01
-3.39412421e-01 5.73832929e-01 6.69418156e-01 -1.14486945e+00
3.84650707e-01 3.20750743e-01 7.94741511e-01 -8.76132786e-01
1.88654467e-01 -1.34060681e+00 2.70375937e-01 1.15884292e+00
-2.79919326e-01 -9.94906425e-02 5.38497157e-02 5.56393266e-01
-4.96930778e-01 -6.41174853e-01 1.20350718e+00 -7.30530143e-01
-5.58434546e-01 3.47620279e-01 -4.02575612e-01 -2.24206984e-01
1.31334233e+00 -3.38037193e-01 3.46263707e-01 -2.97333717e-01
-1.19814932e+00 -7.33563676e-02 -1.47918224e-01 -1.92922428e-01
9.49604928e-01 -1.11176360e+00 -7.51742423e-01 3.76219332e-01
-3.90124947e-01 7.16280103e-01 4.47540909e-01 1.26438725e+00
-1.05282807e+00 4.66273785e-01 -2.33131960e-01 -1.11804795e+00
-1.14111280e+00 6.03663266e-01 1.10652566e+00 -8.24486852e-01
-6.01732969e-01 8.41260433e-01 5.33609152e-01 -4.96508211e-01
2.30072871e-01 -7.48954833e-01 -1.80043146e-01 -3.02871346e-01
8.58907104e-02 -5.32763302e-02 2.75863230e-01 -3.77777219e-01
-1.50629729e-01 6.02665961e-01 -2.44720921e-01 3.98485847e-02
1.28051150e+00 1.02272213e-01 5.99216595e-02 -4.02107947e-02
1.22621846e+00 -5.00980556e-01 -1.04902220e+00 -3.56275439e-01
-1.73658341e-01 -2.25061730e-01 3.86528909e-01 -7.89646924e-01
-1.48820448e+00 1.01042950e+00 1.16144979e+00 1.45649329e-01
1.25577879e+00 6.86639268e-03 7.67455041e-01 1.86237276e-01
-9.97563601e-02 -6.33816540e-01 4.22839671e-01 2.17288271e-01
6.61440253e-01 -1.56597531e+00 -3.46006383e-03 -2.94526845e-01
-6.46633208e-01 1.49798143e+00 7.08177984e-01 -1.55238256e-01
6.67050004e-01 2.28458256e-01 1.20771877e-01 -5.25905132e-01
-2.42594421e-01 -1.54471979e-01 3.88240963e-01 3.82773846e-01
4.40645516e-01 7.40945041e-02 -9.19366553e-02 5.79093218e-01
3.28399986e-01 1.51285753e-01 4.93235320e-01 8.28325808e-01
-4.24546897e-01 -7.88281858e-01 -6.37425244e-01 6.42425299e-01
-6.34351909e-01 1.17386855e-01 -2.93253250e-02 1.15263379e+00
3.76371145e-01 5.04737198e-01 -7.62328878e-03 -9.65218712e-03
4.99362290e-01 -3.65554541e-01 2.41619095e-01 -8.37637305e-01
-7.54142821e-01 1.59529313e-01 -2.75435477e-01 -3.16116303e-01
-2.66966015e-01 -3.12037766e-01 -1.52236354e+00 2.23682791e-01
-2.60405183e-01 -4.38074730e-02 6.13338947e-01 8.88175428e-01
-2.71418184e-01 1.09584451e+00 6.17693007e-01 -7.83624232e-01
-5.94371498e-01 -9.18191075e-01 -5.52231967e-01 3.61704469e-01
4.46014613e-01 -2.57545143e-01 -3.34414959e-01 3.81644443e-02]
|
[15.289369583129883, -2.142106056213379]
|
c53ed74e-4bf6-4e46-88b9-2350a9140fe0
|
generative-poisoning-using-random
|
2211.01086
| null |
https://arxiv.org/abs/2211.01086v1
|
https://arxiv.org/pdf/2211.01086v1.pdf
|
Generative Poisoning Using Random Discriminators
|
We introduce ShortcutGen, a new data poisoning attack that generates sample-dependent, error-minimizing perturbations by learning a generator. The key novelty of ShortcutGen is the use of a randomly-initialized discriminator, which provides spurious shortcuts needed for generating poisons. Different from recent, iterative methods, our ShortcutGen can generate perturbations with only one forward pass in a label-free manner, and compared to the only existing generative method, DeepConfuse, our ShortcutGen is faster and simpler to train while remaining competitive. We also demonstrate that integrating a simple augmentation strategy can further boost the robustness of ShortcutGen against early stopping, and combining augmentation and non-augmentation leads to new state-of-the-art results in terms of final validation accuracy, especially in the challenging, transfer scenario. Lastly, we speculate, through uncovering its working mechanism, that learning a more general representation space could allow ShortcutGen to work for unseen data.
|
['Martha Larson', 'Zhengyu Zhao', 'Zhuoran Liu', 'Alex Kolmus', 'Dirren van Vlijmen']
|
2022-11-02
| null | null | null | null |
['data-poisoning']
|
['adversarial']
|
[ 3.36193711e-01 3.00880075e-01 -3.07119638e-01 1.66833639e-01
-1.02094591e+00 -1.13327014e+00 9.63034213e-01 1.54824376e-01
-4.32713002e-01 1.05981255e+00 6.66464791e-02 -4.58741099e-01
3.45667720e-01 -8.06107581e-01 -8.45538914e-01 -8.06145668e-01
1.37892300e-02 6.97467625e-01 1.20672897e-01 -2.34367520e-01
1.74397618e-01 5.18424273e-01 -1.10416842e+00 2.27095902e-01
7.05452263e-01 2.69874901e-01 -5.89444757e-01 5.99541724e-01
4.43203270e-01 6.32330775e-01 -1.10681295e+00 -7.82874703e-01
5.10287404e-01 -6.63028181e-01 -9.28630531e-01 -1.16750047e-01
2.05911204e-01 -4.90171313e-01 -4.26405936e-01 7.82207906e-01
7.35744119e-01 -2.49967664e-01 6.71483397e-01 -1.53104150e+00
-6.32770777e-01 8.80170465e-01 -4.16987062e-01 1.19615354e-01
2.30821013e-01 9.06838834e-01 7.33362675e-01 -6.31980002e-01
7.28387356e-01 1.16970873e+00 8.61822128e-01 1.02949309e+00
-1.75490642e+00 -1.00837207e+00 -1.47055939e-01 -2.10345358e-01
-1.11182654e+00 -2.86358625e-01 4.74710763e-01 -2.75125653e-01
6.05752289e-01 4.85222489e-01 4.27427799e-01 2.14989066e+00
1.07976973e-01 8.02687049e-01 1.10034215e+00 -1.13483824e-01
4.48240995e-01 8.51729512e-02 -3.09761725e-02 4.40470785e-01
5.77690482e-01 6.47702873e-01 -2.97903776e-01 -8.56544077e-01
3.99150938e-01 -5.07367998e-02 -4.17831957e-01 -1.95513636e-01
-9.61469173e-01 1.33829963e+00 4.13872063e-01 7.03623593e-02
-5.29106557e-02 4.91886437e-01 5.69994986e-01 2.58959860e-01
4.68218505e-01 1.01202500e+00 -2.53809392e-01 -1.16170160e-01
-9.59063590e-01 4.65722859e-01 9.12706792e-01 6.98709548e-01
6.14881635e-01 1.82549343e-01 -6.47260666e-01 -2.55154893e-02
-8.74820724e-02 4.60985333e-01 3.22447002e-01 -7.26732969e-01
4.02425081e-01 3.03499430e-01 1.00314468e-01 -4.69362289e-01
-3.06352496e-01 -7.81514585e-01 -7.07635105e-01 2.99408257e-01
4.48801339e-01 -6.23028517e-01 -1.09190917e+00 2.04800224e+00
3.89093906e-01 3.71516198e-01 3.41381803e-02 6.03194356e-01
3.99455339e-01 3.60042989e-01 2.02953100e-01 4.01477516e-02
1.03576922e+00 -6.23211563e-01 -1.74249291e-01 -4.48081344e-02
8.42691720e-01 -2.16645762e-01 1.22776186e+00 5.39276719e-01
-8.67377520e-01 5.05138934e-02 -1.21790564e+00 1.82261631e-01
-4.14548695e-01 -5.78450859e-01 7.49526799e-01 1.07653439e+00
-7.42967427e-01 6.94452286e-01 -6.61322594e-01 -1.78349718e-01
9.59190249e-01 3.23842496e-01 -2.80736536e-01 -7.32124001e-02
-1.27252293e+00 7.20687509e-01 2.93163449e-01 -3.93890768e-01
-1.73917007e+00 -1.02609563e+00 -6.66383982e-01 -8.14361349e-02
3.57173771e-01 -1.01884902e+00 1.08405066e+00 -3.58839989e-01
-1.26471364e+00 4.67768162e-01 3.20219755e-01 -8.46539438e-01
9.65519190e-01 -8.73166993e-02 7.87948146e-02 1.73549369e-01
-3.89126912e-02 8.10086012e-01 8.95513058e-01 -1.49032736e+00
1.08835613e-02 -8.13975856e-02 -7.36300275e-02 -6.47191554e-02
-3.23084593e-01 -1.34794220e-01 1.45412341e-01 -7.47577786e-01
-6.03025377e-01 -8.74351382e-01 -6.18978798e-01 -3.61913770e-01
-1.05563748e+00 -1.33412957e-01 9.94645715e-01 -1.30539387e-01
9.02893305e-01 -1.93251514e+00 -9.43644252e-03 2.41261572e-01
4.91781205e-01 6.76291347e-01 -4.25331920e-01 8.25999022e-01
-8.24222639e-02 7.95890868e-01 -7.75150180e-01 -3.23529243e-01
-1.07871138e-01 2.74109822e-02 -8.51346493e-01 5.91225743e-01
3.76348972e-01 1.17660308e+00 -1.14786041e+00 6.83155581e-02
-1.03116803e-01 4.59150404e-01 -7.59506404e-01 6.86353967e-02
-5.65668523e-01 6.43052638e-01 -2.51389593e-01 5.52866042e-01
6.69283152e-01 -7.72781447e-02 -2.92433321e-01 2.14528605e-01
3.71383131e-01 2.78800577e-01 -7.41738081e-01 1.61085010e+00
-1.02671571e-01 4.29965109e-01 -3.29774201e-01 -6.56833768e-01
6.37425363e-01 3.36740553e-01 1.22036017e-01 -1.05594598e-01
3.23767126e-01 1.91280678e-01 -2.21262187e-01 -1.69890717e-01
2.02420816e-01 -3.27412844e-01 -2.91256696e-01 9.57025826e-01
1.56998280e-02 -2.98835486e-01 4.01422083e-02 5.67274153e-01
1.68096149e+00 -1.89505637e-01 1.95606630e-02 -9.39606503e-02
2.60957479e-02 1.01335958e-01 3.46998662e-01 1.32976329e+00
2.21152250e-02 9.05740976e-01 7.36180425e-01 -4.52570230e-01
-1.10242784e+00 -1.08843124e+00 9.14833043e-03 6.67492449e-01
-1.36006892e-01 -5.72855651e-01 -9.79236901e-01 -1.50768077e+00
3.27740729e-01 1.06197906e+00 -8.76907885e-01 -6.74318731e-01
-4.13228840e-01 -1.13751090e+00 1.47999012e+00 1.31942466e-01
2.80490011e-01 -1.16351771e+00 -4.61093575e-01 2.03464434e-01
1.93942152e-02 -7.17655063e-01 -4.94662762e-01 4.08169597e-01
-6.24238610e-01 -1.22249055e+00 -4.33580965e-01 -1.73987508e-01
5.91286004e-01 -6.19193316e-02 1.08968747e+00 3.03103328e-01
-4.75261152e-01 1.33399025e-01 -2.65143692e-01 -6.07237041e-01
-8.39948654e-01 5.04210651e-01 -5.17132208e-02 -2.68776596e-01
1.88839436e-01 -7.03069687e-01 -5.56949496e-01 1.98812068e-01
-1.20664549e+00 -4.60101902e-01 4.97869313e-01 9.36760128e-01
1.28513366e-01 -2.31251910e-01 8.96831691e-01 -1.40185547e+00
8.02080631e-01 -9.42402184e-01 -4.39303100e-01 -7.38331527e-02
-6.96115494e-01 2.26867363e-01 9.50647593e-01 -7.29464769e-01
-5.24338365e-01 -1.95146769e-01 -2.59144783e-01 -4.57606435e-01
-2.01560736e-01 -6.60107061e-02 -1.31210387e-01 -6.88641295e-02
1.35464931e+00 4.13112879e-01 1.46193177e-01 -3.95833522e-01
8.34939957e-01 3.83975953e-01 5.15145421e-01 -6.01445675e-01
1.31338429e+00 5.78233242e-01 8.64697918e-02 -3.61065894e-01
-7.11638927e-01 9.28811878e-02 -1.32708684e-01 3.24095339e-01
6.14530087e-01 -7.23954082e-01 -7.90810227e-01 6.90268338e-01
-1.24644268e+00 -5.49295545e-01 -7.94418275e-01 -6.30187988e-02
-4.07608002e-01 2.92460471e-01 -8.23896408e-01 -6.01169705e-01
-6.16268814e-01 -1.04402661e+00 8.93138289e-01 -2.40748987e-01
-3.39039773e-01 -7.78152704e-01 3.80200922e-01 1.69318423e-01
4.95050997e-01 7.91780710e-01 8.02361786e-01 -1.31811225e+00
-7.18687594e-01 -3.62955391e-01 9.25369635e-02 3.58200312e-01
-9.45053101e-02 -2.12344840e-01 -1.08142161e+00 -7.15534508e-01
1.00377806e-01 -8.58801663e-01 1.08509171e+00 -2.06105873e-01
1.21728766e+00 -9.56667364e-01 -3.93579364e-01 9.50894058e-01
1.22798574e+00 -1.74696013e-01 8.46430838e-01 2.29006886e-01
7.72666037e-01 9.30936560e-02 8.81480873e-02 4.28480029e-01
-1.78511515e-02 2.78685063e-01 7.50452936e-01 -2.83573568e-01
-1.74808323e-01 -5.29214382e-01 4.16708618e-01 -1.09496914e-01
5.68750978e-01 -4.30421561e-01 -6.11912310e-01 4.24061656e-01
-1.41205728e+00 -1.24762619e+00 -3.22826318e-02 2.21966887e+00
1.28503621e+00 1.70077488e-01 5.14040768e-01 3.09400648e-01
5.88027596e-01 4.33461145e-02 -8.61347198e-01 -3.00830781e-01
-1.75633892e-01 7.07259297e-01 7.05426514e-01 3.71865243e-01
-1.04774380e+00 9.92511094e-01 7.54660130e+00 9.70581114e-01
-1.09054756e+00 2.58826435e-01 8.21519554e-01 -3.96131784e-01
-7.28731096e-01 6.01447374e-02 -7.28962779e-01 6.28376007e-01
9.59943056e-01 -1.20201632e-01 4.42273289e-01 7.64650464e-01
-3.04392159e-01 3.26277763e-01 -1.26166511e+00 5.30232549e-01
1.11606441e-01 -1.63878775e+00 2.15478972e-01 2.95121819e-01
8.00892532e-01 2.84671694e-01 2.15610713e-01 2.28167206e-01
1.01629937e+00 -1.36711216e+00 6.24573529e-01 -1.08358704e-01
8.27876151e-01 -1.13851547e+00 5.26240885e-01 4.98492479e-01
-3.35461408e-01 -1.52257890e-01 -1.31338477e-01 2.11527601e-01
2.23944001e-02 8.35079074e-01 -1.13351345e+00 3.95154029e-01
1.98173434e-01 2.55576283e-01 -7.69623578e-01 1.07711387e+00
-5.97014427e-01 1.01731133e+00 -3.19860786e-01 1.23979770e-01
1.91708371e-01 2.86432981e-01 9.47734177e-01 1.31926179e+00
1.09830134e-01 -5.01386821e-02 2.14299306e-01 1.24806201e+00
-4.01957899e-01 -4.66637224e-01 -9.87134755e-01 -6.49625063e-02
7.78747380e-01 1.21035719e+00 -4.72444475e-01 -1.54165745e-01
3.59685868e-01 9.70428765e-01 3.10063392e-01 3.52315873e-01
-1.05176759e+00 -6.07420266e-01 5.93902051e-01 2.03765556e-01
2.29792923e-01 9.02770981e-02 -6.20688915e-01 -9.61737752e-01
-2.06649438e-01 -1.33911407e+00 5.74056387e-01 -2.11329937e-01
-1.53432643e+00 7.01749325e-01 -1.20540634e-02 -9.64469016e-01
-3.62879753e-01 -2.06924886e-01 -8.45921755e-01 7.21602261e-01
-1.11208844e+00 -1.11409092e+00 -4.83083762e-02 5.72185457e-01
1.77667156e-01 -1.46375448e-01 8.72155964e-01 -1.91916689e-01
-7.06081152e-01 1.18951583e+00 -2.53690898e-01 4.53088060e-02
6.19492829e-01 -1.20037055e+00 9.47544754e-01 1.07223892e+00
2.42862955e-01 6.27785027e-01 8.95608664e-01 -8.11203361e-01
-1.14631045e+00 -1.31830609e+00 3.10195684e-01 -9.89149928e-01
4.95709956e-01 -9.74258125e-01 -7.94712722e-01 5.89741349e-01
2.85574138e-01 -4.92171273e-02 6.95752740e-01 -1.67072058e-01
-9.79616106e-01 2.25054592e-01 -1.84921801e+00 6.70249522e-01
1.09176421e+00 -3.69043022e-01 -2.75179744e-01 4.55310792e-01
1.04166412e+00 -3.12020808e-01 -3.26617360e-01 1.81028679e-01
-2.02781372e-02 -9.57693100e-01 9.80154574e-01 -8.90717804e-01
3.23628068e-01 -2.37994045e-01 2.20471084e-01 -1.50584841e+00
-2.32694536e-01 -1.39143598e+00 -3.29742014e-01 1.34725654e+00
5.33993185e-01 -1.01504707e+00 9.53617752e-01 2.04372063e-01
3.68606783e-02 -7.74615288e-01 -8.83902669e-01 -1.05048072e+00
5.05164862e-01 -2.63694346e-01 8.81942809e-01 9.17717636e-01
4.81755985e-03 2.33647883e-01 -6.16158724e-01 4.82068248e-02
8.58746886e-01 -4.53395635e-01 9.07967985e-01 -1.09571052e+00
-6.12981081e-01 -2.52862543e-01 -3.20273876e-01 -5.55857062e-01
2.00469330e-01 -1.12165546e+00 1.49876513e-02 -1.05895555e+00
3.58173192e-01 -5.73389232e-01 -5.00617549e-02 8.67810726e-01
-3.85366321e-01 6.06326282e-01 2.23880321e-01 1.79884776e-01
-2.96426415e-01 5.72341859e-01 9.86073375e-01 -1.67191058e-01
-1.58127159e-01 -4.05329727e-02 -1.31671870e+00 2.15921372e-01
8.82641196e-01 -1.05836403e+00 -5.30710578e-01 -1.45270005e-01
-1.21434219e-02 -5.44834912e-01 5.51337123e-01 -9.86037135e-01
4.07714210e-03 -1.09411441e-01 3.33686739e-01 -3.27558219e-01
7.25762695e-02 -2.79908448e-01 1.23275504e-01 7.92513907e-01
-3.68596494e-01 3.93180139e-02 2.34928891e-01 5.97743928e-01
4.56060201e-01 -2.89613634e-01 8.50711405e-01 -1.27268851e-01
2.54121929e-01 5.08321524e-01 -4.56695706e-01 6.39519811e-01
1.14717782e+00 1.65619791e-01 -8.52666497e-01 -2.28534684e-01
-1.99170202e-01 -5.78533448e-02 7.08507180e-01 1.35053799e-01
4.59095925e-01 -1.23428261e+00 -1.05725241e+00 3.83879960e-01
5.61224259e-02 8.48737508e-02 -2.20704690e-01 4.12015021e-01
-2.57151157e-01 4.52628396e-02 2.23498657e-01 -4.92605597e-01
-7.17225254e-01 9.36255157e-01 2.51475990e-01 -3.51457089e-01
-6.41861498e-01 9.97888565e-01 1.98182359e-01 -3.92352313e-01
2.08233729e-01 1.36523828e-01 5.97584009e-01 -1.85312092e-01
6.81972742e-01 3.71277720e-01 6.01598658e-02 3.00975014e-02
-2.96265125e-01 -1.73967391e-01 -3.21690142e-01 -2.99065679e-01
1.15395951e+00 5.57228804e-01 1.09201767e-01 -7.98425302e-02
1.12513328e+00 2.31576025e-01 -1.31811821e+00 2.82995224e-01
-2.07336485e-01 -3.70763510e-01 -3.77826989e-01 -1.06833661e+00
-9.81685042e-01 6.20781243e-01 2.10221395e-01 1.67125106e-01
8.32980156e-01 4.36150655e-02 9.84947562e-01 3.32291335e-01
4.00198936e-01 -3.62564981e-01 3.80596638e-01 2.70561039e-01
8.27175379e-01 -8.66605401e-01 -1.17537878e-01 -2.25625396e-01
-5.46938419e-01 6.40715718e-01 5.38839102e-01 -3.08574706e-01
1.19662844e-01 5.38846493e-01 5.57483286e-02 -1.42295495e-01
-9.00865555e-01 9.17482898e-02 -1.43357053e-01 9.87917542e-01
-9.68981534e-02 -1.91177487e-01 -7.50718936e-02 4.20919657e-01
-3.90425771e-01 -1.15150549e-01 8.40389013e-01 7.84967959e-01
-1.38648391e-01 -1.57981849e+00 -3.20142418e-01 3.29093397e-01
-6.43382192e-01 -2.87952006e-01 -7.13631749e-01 8.83008540e-01
1.44739807e-01 1.04868519e+00 -2.70802468e-01 -5.55944622e-01
7.07758144e-02 1.36152387e-01 3.77231419e-01 -7.59178042e-01
-1.15050280e+00 -5.36144853e-01 -3.18583623e-02 -5.57839096e-01
4.04590845e-01 -4.79439676e-01 -1.03289294e+00 -4.52104449e-01
-3.85032028e-01 4.01715785e-01 3.87434483e-01 8.15229535e-01
5.95766664e-01 5.72968312e-02 1.01164901e+00 -6.96841478e-01
-1.12821078e+00 -7.51419425e-01 -1.24146961e-01 7.23058105e-01
5.17425179e-01 -3.95982683e-01 -8.79079878e-01 -4.23755467e-01]
|
[5.868206024169922, 7.660614967346191]
|
b621a0b2-e3ed-49e7-9ef1-ed9aa1c9988e
|
cgan-based-high-dimensional-imu-sensor-data
|
2302.07998
| null |
https://arxiv.org/abs/2302.07998v1
|
https://arxiv.org/pdf/2302.07998v1.pdf
|
cGAN-Based High Dimensional IMU Sensor Data Generation for Therapeutic Activities
|
Human activity recognition is a core technology for applications such as rehabilitation, ambient health monitoring, and human-computer interactions. Wearable devices, particularly IMU sensors, can help us collect rich features of human movements that can be leveraged in activity recognition. Developing a robust classifier for activity recognition has always been of interest to researchers. One major problem is that there is usually a deficit of training data for some activities, making it difficult and sometimes impossible to develop a classifier. In this work, a novel GAN network called TheraGAN was developed to generate realistic IMU signals associated with a particular activity. The generated signal is of a 6-channel IMU. i.e., angular velocities and linear accelerations. Also, by introducing simple activities, which are meaningful subparts of a complex full-length activity, the generation process was facilitated for any activity with arbitrary length. To evaluate the generated signals, besides perceptual similarity metrics, they were applied along with real data to improve the accuracy of classifiers. The results show that the maximum increase in the f1-score belongs to the LSTM classifier by a 13.27% rise when generated data were added. This shows the validity of the generated data as well as TheraGAN as a tool to build more robust classifiers in case of imbalanced data problem.
|
['Saeed Behzadipour', 'Alireza Taheri', 'Ali Ghadami', 'Mohammad Mohammadzadeh']
|
2023-02-16
| null | null | null | null |
['human-activity-recognition', 'human-activity-recognition']
|
['computer-vision', 'time-series']
|
[ 4.79361594e-01 -1.21560104e-01 -7.22069591e-02 -3.87687311e-02
-4.71767485e-01 -2.73824662e-01 4.75735098e-01 -4.81608175e-02
-2.27872849e-01 8.88865709e-01 4.59189206e-01 2.21062109e-01
-6.39917105e-02 -6.74388945e-01 -7.12906420e-01 -8.86856377e-01
-1.46677226e-01 -2.14985490e-01 -1.79124087e-01 -8.76152441e-02
-3.56714125e-03 3.35441202e-01 -1.83978832e+00 2.37845257e-01
7.53206193e-01 1.01815319e+00 2.36196503e-01 4.07099247e-01
-2.34284475e-02 5.23388445e-01 -1.16183925e+00 1.26502290e-01
1.00327663e-01 -8.25966120e-01 -5.28526962e-01 1.79122955e-01
-8.74886066e-02 5.22218421e-02 1.57039598e-01 6.32821620e-01
7.38298595e-01 4.19727087e-01 5.08512735e-01 -1.13838577e+00
-3.15252803e-02 4.59554404e-01 -3.02819163e-01 3.30055773e-01
8.69895697e-01 4.20096628e-02 4.53198075e-01 -3.86064380e-01
5.03260851e-01 6.70674860e-01 5.21712720e-01 4.39264596e-01
-8.82411301e-01 -6.55780911e-01 -1.59091413e-01 4.57624286e-01
-1.05032790e+00 -6.85693771e-02 1.07284796e+00 -4.75825399e-01
8.59978735e-01 5.41217268e-01 1.05442321e+00 1.57734692e+00
2.49987870e-01 8.37072790e-01 1.17791116e+00 -4.62957233e-01
2.79362947e-01 -7.95851499e-02 -7.75449425e-02 1.07045084e-01
1.96858212e-01 -1.69467762e-01 -4.37567532e-01 2.50077963e-01
5.19528985e-01 2.17622042e-01 -3.61848325e-01 4.41168919e-02
-1.32630122e+00 4.61977154e-01 4.71526623e-01 8.54165912e-01
-6.12379134e-01 -8.73240754e-02 3.72368723e-01 1.96735740e-01
3.38568538e-01 6.19212866e-01 -9.36855823e-02 -6.77546799e-01
-6.89545274e-01 5.21324947e-02 5.31717360e-01 6.38735116e-01
3.94170463e-01 3.67051870e-01 -2.83126056e-01 7.96589077e-01
-1.25375455e-02 5.94416738e-01 9.93813455e-01 -4.40724254e-01
4.14150774e-01 8.96979690e-01 -5.60868345e-02 -1.04783010e+00
-7.24168777e-01 -7.11157560e-01 -9.58554447e-01 7.96557143e-02
3.94572347e-01 -3.09816867e-01 -7.39768028e-01 1.64658272e+00
3.83970946e-01 3.23963344e-01 -2.14416999e-02 1.00038302e+00
5.95062315e-01 6.98220491e-01 -8.48469511e-02 -2.97631413e-01
1.23277771e+00 -5.49831927e-01 -8.47609580e-01 -2.01457232e-01
6.75151289e-01 -5.59990764e-01 1.10783935e+00 4.55438226e-01
-6.54663682e-01 -7.79950559e-01 -1.39573324e+00 6.30270302e-01
-2.58836985e-01 8.16701055e-02 5.41116178e-01 7.89257765e-01
-5.69440961e-01 7.56042957e-01 -7.92450130e-01 -5.00032008e-01
1.46704018e-01 2.48562172e-01 -5.12291789e-01 2.31254146e-01
-1.23277640e+00 8.69857132e-01 5.13819516e-01 1.80215478e-01
-5.11027277e-01 -1.24907963e-01 -6.61518455e-01 -1.95595145e-01
1.79891393e-01 -3.52550060e-01 8.17830324e-01 -1.08645415e+00
-1.47769833e+00 3.99139673e-01 1.71552584e-01 -3.59454840e-01
5.15645325e-01 -2.38807902e-01 -7.43620634e-01 -1.77639782e-01
-1.17731139e-01 2.46620364e-02 8.74969006e-01 -8.27898979e-01
-3.85927409e-01 -3.26904595e-01 -1.74354523e-01 3.58784705e-01
-4.02160168e-01 -1.48569599e-01 1.41754508e-01 -8.49615097e-01
1.08517677e-01 -9.69867170e-01 3.19862425e-01 -6.16008282e-01
-3.49923939e-01 3.13832872e-02 7.75741220e-01 -9.09366488e-01
1.19408906e+00 -2.07896256e+00 1.53800264e-01 3.81479651e-01
-1.86279699e-01 4.63816971e-01 9.35399756e-02 3.95345747e-01
-1.68299377e-01 -2.22645719e-02 -7.29093477e-02 7.67374691e-03
-2.23606825e-01 2.25462496e-01 1.92079559e-01 3.49867851e-01
1.06538683e-01 6.50584161e-01 -8.27649415e-01 -6.86696246e-02
4.27166283e-01 5.92349768e-01 -1.75080925e-01 3.61669779e-01
2.06509545e-01 9.33805585e-01 -3.59648138e-01 4.42649186e-01
1.69758320e-01 1.19971998e-01 -1.03605993e-01 -3.10037613e-01
1.08211964e-01 2.87802994e-01 -1.36870873e+00 1.68634164e+00
-6.59729660e-01 5.90162694e-01 -4.12767887e-01 -1.23702788e+00
1.08317125e+00 5.03630817e-01 8.31653237e-01 -8.90668631e-01
4.37062919e-01 7.82783926e-02 1.18117176e-01 -7.92043269e-01
7.14551881e-02 1.22223057e-01 -8.75249431e-02 4.81164217e-01
1.14051644e-02 2.95826763e-01 3.32353979e-01 -4.29745704e-01
1.07446384e+00 3.39055806e-01 3.61695826e-01 1.04911909e-01
6.33519411e-01 -3.12824100e-01 4.09321994e-01 4.36600596e-01
5.70729822e-02 5.77340424e-01 2.77365670e-02 -2.08764762e-01
-9.06599581e-01 -8.10504019e-01 4.26076911e-02 7.50088453e-01
-6.68908954e-02 -2.97298282e-01 -8.63014877e-01 -3.30896944e-01
-4.59991634e-01 4.58227634e-01 -5.47960341e-01 -4.24177289e-01
-5.40001631e-01 -9.44839656e-01 6.14430964e-01 6.71323597e-01
9.27691996e-01 -1.28603697e+00 -1.05058253e+00 3.27687353e-01
-4.92947400e-01 -8.94120753e-01 -2.03374490e-01 1.98952094e-01
-1.06959689e+00 -1.02702534e+00 -7.74611413e-01 -5.92757761e-01
3.84586334e-01 8.78182799e-02 6.32974684e-01 -1.60827368e-01
-2.40496710e-01 1.86593071e-01 -7.92313159e-01 -5.22664607e-01
-3.66741717e-01 1.06033064e-01 2.30685964e-01 2.31116548e-01
2.41407350e-01 -6.78683698e-01 -6.44287586e-01 5.09181678e-01
-7.10988343e-01 1.48894545e-02 6.53933525e-01 8.71774375e-01
4.43477958e-01 5.76066300e-02 6.30649388e-01 -2.93402791e-01
9.36341166e-01 -4.41485494e-01 1.06145531e-01 -1.05057154e-02
-2.37090677e-01 -1.31191239e-02 6.54856741e-01 -7.84946978e-01
-8.56623232e-01 -1.70654629e-03 -2.62274444e-01 -2.67013237e-02
-3.72327894e-01 5.42391837e-01 -4.04819131e-01 1.56785056e-01
9.82134461e-01 2.23753318e-01 8.11354965e-02 -4.62446749e-01
-8.51947442e-02 9.40811217e-01 3.01041931e-01 -2.95097560e-01
3.42359006e-01 4.55224700e-02 3.06014903e-03 -1.22110951e+00
-3.16478789e-01 -3.95766556e-01 -3.39933127e-01 -5.85490644e-01
8.59109879e-01 -6.66176915e-01 -5.08918822e-01 7.04530835e-01
-8.24792027e-01 -2.02179953e-01 -2.92547643e-01 8.40739787e-01
-4.41937685e-01 1.77115053e-01 -2.19662011e-01 -8.85887086e-01
-5.38009644e-01 -1.03345037e+00 7.09735274e-01 3.65518451e-01
-6.08752012e-01 -7.55345047e-01 -5.98688088e-02 5.07999063e-01
4.72310126e-01 8.28589320e-01 5.19584596e-01 -7.25402296e-01
-3.70704159e-02 -6.16948962e-01 4.46043700e-01 5.77709913e-01
5.05219936e-01 -4.11394328e-01 -1.01256549e+00 -1.84188560e-01
3.82742435e-01 -1.64474219e-01 1.18797995e-01 1.86166048e-01
9.50869739e-01 -4.18763429e-01 -1.89267650e-01 3.52591306e-01
9.20971096e-01 6.73492670e-01 9.44154441e-01 3.22697580e-01
6.57380462e-01 4.31917578e-01 6.13352478e-01 3.82898062e-01
-1.89763233e-01 9.48487878e-01 2.98777342e-01 -9.86052454e-02
-6.59730807e-02 -2.38362476e-01 4.51865464e-01 9.09958899e-01
-6.72898591e-01 -1.42120168e-01 -7.22828627e-01 1.24097653e-01
-1.60001028e+00 -1.16217148e+00 -1.60312235e-01 2.46930194e+00
4.75676775e-01 1.26141950e-01 3.94214958e-01 9.91071701e-01
6.64453030e-01 5.31982854e-02 -3.76505017e-01 -3.73230129e-01
-5.84856281e-03 2.48663157e-01 2.28795737e-01 -5.68492413e-02
-1.01983178e+00 1.89395279e-01 6.03473902e+00 7.03750849e-01
-1.51183867e+00 1.03672378e-01 2.95263141e-01 -1.39433786e-01
3.48933227e-02 -5.45685649e-01 -3.64506036e-01 9.62546051e-01
9.48101819e-01 4.43039574e-02 3.95598710e-01 7.56719470e-01
5.35619259e-01 -3.41200680e-01 -7.79194891e-01 1.25709665e+00
2.50112653e-01 -7.92358398e-01 -1.79903284e-01 7.95589313e-02
6.09433234e-01 -3.20350945e-01 -3.79019320e-01 7.76003301e-02
-4.21640337e-01 -1.07741642e+00 4.24949408e-01 5.86716175e-01
5.19523203e-01 -7.46382654e-01 1.07034600e+00 5.15925169e-01
-1.10305452e+00 3.18553634e-02 3.66185829e-02 -3.47522825e-01
1.70123547e-01 4.86980021e-01 -8.34828198e-01 6.47776723e-01
5.77408969e-01 6.53512537e-01 -4.72784430e-01 1.16805220e+00
-3.01965058e-01 7.46332049e-01 -4.59780127e-01 -4.20499027e-01
-1.42650336e-01 -2.61067182e-01 5.43915749e-01 1.08832395e+00
7.91022658e-01 -2.46856898e-01 -1.31172210e-01 5.10361671e-01
2.29715303e-01 3.23726088e-01 -8.96581769e-01 -7.12103248e-02
1.79808378e-01 1.07420802e+00 -6.47615016e-01 -4.45592031e-02
-7.50739574e-02 1.02110577e+00 -2.19924957e-01 1.78779781e-01
-9.71733332e-01 -4.56351012e-01 3.96470636e-01 2.45213941e-01
-2.79175401e-01 -1.57332316e-01 -1.99736848e-01 -9.76405263e-01
3.72099489e-01 -9.78735328e-01 2.96300501e-01 -8.45569253e-01
-8.06666195e-01 6.20002747e-01 1.70826837e-01 -1.81147265e+00
-8.03349137e-01 -3.77191216e-01 -7.21962333e-01 7.23631322e-01
-5.97898722e-01 -9.50955033e-01 -8.51038814e-01 6.17674828e-01
4.49819326e-01 -1.08688109e-01 8.61693144e-01 4.67938602e-01
-5.01526892e-01 4.67076093e-01 -1.48155823e-01 4.47629988e-02
3.14228058e-01 -9.64904487e-01 8.63532871e-02 1.01053417e+00
2.64111519e-01 5.07736742e-01 8.21183622e-01 -5.41039050e-01
-1.23246729e+00 -7.87574947e-01 5.52519679e-01 -2.11713731e-01
2.04421207e-01 -2.25172803e-01 -7.19349623e-01 4.25025433e-01
2.04123184e-02 -2.66062140e-01 8.74514282e-01 -2.03576326e-01
2.83210814e-01 -2.78694957e-01 -1.15982163e+00 5.25824428e-01
1.12746882e+00 -2.90883958e-01 -6.63247824e-01 1.38640590e-02
3.00899446e-02 -4.35519874e-01 -1.02257085e+00 4.20489192e-01
7.24508107e-01 -8.91928196e-01 7.92144537e-01 -2.71066129e-01
1.03454515e-01 -3.60396862e-01 1.04833625e-01 -1.61250341e+00
-7.84133300e-02 -3.83621693e-01 -2.77943105e-01 1.07592726e+00
2.71792591e-01 -6.15900636e-01 7.33771265e-01 1.79972276e-01
-1.36838168e-01 -5.96967280e-01 -8.91433775e-01 -9.23246205e-01
-5.23384929e-01 -6.61490381e-01 5.79278588e-01 8.58815372e-01
3.20964128e-01 3.73250902e-01 -7.05387533e-01 -4.16451424e-01
2.61701733e-01 -3.02694291e-01 9.67835546e-01 -1.24233556e+00
-2.81407267e-01 -2.30426058e-01 -1.02408218e+00 -7.77717590e-01
-3.41756463e-01 -7.44596720e-01 -8.21619034e-02 -1.66490555e+00
-3.72895926e-01 -1.33778632e-01 -3.52015287e-01 3.94674182e-01
1.17665692e-03 3.84889513e-01 9.35321450e-02 -2.15988494e-02
6.20179586e-02 4.13893998e-01 1.30596769e+00 -2.68744141e-01
-5.99449456e-01 4.78031129e-01 -3.11344415e-01 4.76091713e-01
1.04801965e+00 -2.49716803e-01 -6.92671239e-01 2.30392888e-02
7.31145069e-02 -1.82553530e-02 1.77918494e-01 -1.69031751e+00
-1.67951599e-01 7.45789930e-02 6.95313573e-01 -2.69039273e-01
5.17135322e-01 -8.78452837e-01 7.42270052e-01 5.71847975e-01
1.04111312e-02 2.03994419e-02 7.57429749e-02 2.85164177e-01
-3.21972460e-01 -1.64472729e-01 3.60473454e-01 5.65994484e-03
-6.72961593e-01 -1.63218319e-01 -4.07481372e-01 -1.02509104e-01
1.26914918e+00 -7.24677920e-01 2.83808913e-02 -5.24036646e-01
-7.47748315e-01 -1.86580241e-01 1.52217671e-01 6.88143671e-01
4.55733448e-01 -1.52286482e+00 -3.83509040e-01 3.11777234e-01
2.67257899e-01 -1.05593517e-01 3.15251917e-01 1.02778530e+00
-3.92870277e-01 1.86294734e-01 -7.65601695e-01 -6.19993746e-01
-1.34326494e+00 2.32690498e-01 3.44404966e-01 -3.95470187e-02
-7.20115185e-01 3.02371740e-01 -5.21538913e-01 5.01333922e-02
2.90472686e-01 -7.02915013e-01 -7.21809328e-01 4.49528456e-01
6.25310063e-01 6.06299758e-01 3.67079973e-01 -6.84004188e-01
-3.40265840e-01 5.71313024e-01 5.94167888e-01 -8.40895474e-02
1.12396920e+00 1.06352039e-01 3.72089773e-01 7.47229934e-01
9.62000906e-01 -1.33975700e-01 -1.04380846e+00 3.41350257e-01
5.40333763e-02 -4.07758921e-01 -2.63522416e-01 -7.21073091e-01
-9.77920949e-01 7.38354146e-01 1.00784135e+00 2.77152598e-01
1.28911400e+00 -3.42747450e-01 6.69174492e-01 2.13658363e-01
6.95764184e-01 -1.26060045e+00 1.77798390e-01 1.00288451e-01
9.04040694e-01 -9.58273947e-01 -2.41868168e-01 4.20062393e-02
-7.38909543e-01 8.51731300e-01 3.68867427e-01 7.32156411e-02
2.44126394e-01 8.73597562e-02 2.43820939e-02 2.36752540e-01
-5.48307374e-02 -2.42864564e-01 4.46214199e-01 8.36904347e-01
4.77584630e-01 8.40258226e-02 -7.73139060e-01 6.76294565e-01
-2.81238079e-01 2.08473012e-01 4.26440537e-01 1.04809153e+00
-3.63229215e-01 -9.28619146e-01 -5.69189191e-01 6.99198186e-01
-4.61792916e-01 5.30719280e-01 -1.12924263e-01 6.88303590e-01
3.74463648e-01 1.08191407e+00 -7.64061809e-02 -9.53958392e-01
5.53172410e-01 2.82532007e-01 4.85516191e-01 -3.59513223e-01
-6.19708776e-01 -2.24515274e-01 1.64269745e-01 -6.02836072e-01
-6.95715845e-01 -4.69383240e-01 -1.22298145e+00 -7.60138631e-02
-2.26534963e-01 1.67168245e-01 9.80616033e-01 1.07549596e+00
1.16116479e-01 7.90094852e-01 5.61292708e-01 -1.05306482e+00
-2.29388416e-01 -1.43863511e+00 -4.35632706e-01 9.52455163e-01
3.68682407e-02 -8.47529829e-01 -3.34358990e-01 2.38834731e-02]
|
[7.348008155822754, 0.5259703397750854]
|
ac2e46fa-68ed-4bac-8ae3-efd46798d996
|
explainable-artificial-intelligence-in
|
2203.16073
| null |
https://arxiv.org/abs/2203.16073v4
|
https://arxiv.org/pdf/2203.16073v4.pdf
|
Explainability in Process Outcome Prediction: Guidelines to Obtain Interpretable and Faithful Models
|
Although a recent shift has been made in the field of predictive process monitoring to use models from the explainable artificial intelligence field, the evaluation still occurs mainly through performance-based metrics, thus not accounting for the actionability and implications of the explanations. In this paper, we define explainability through the interpretability of the explanations and the faithfulness of the explainability model in the field of process outcome prediction. The introduced properties are analysed along the event, case, and control flow perspective which are typical for a process-based analysis. This allows comparing inherently created explanations with post-hoc explanations. We benchmark seven classifiers on thirteen real-life events logs, and these cover a range of transparent and non-transparent machine learning and deep learning models, further complemented with explainability techniques. Next, this paper contributes a set of guidelines named X-MOP which allows selecting the appropriate model based on the event log specifications, by providing insight into how the varying preprocessing, model complexity and explainability techniques typical in process outcome prediction influence the explainability of the model.
|
['Johannes De Smedt', 'Alexander Stevens']
|
2022-03-30
| null | null | null | null |
['predictive-process-monitoring']
|
['time-series']
|
[ 3.40620667e-01 7.87795603e-01 -2.31630221e-01 -5.94639838e-01
-3.97110470e-02 -3.94168288e-01 1.07447958e+00 9.22267854e-01
3.82339776e-01 4.75090414e-01 4.59438622e-01 -7.40656853e-01
-1.04329014e+00 -8.56350183e-01 -3.15522671e-01 -3.19376111e-01
-8.75147507e-02 6.86240673e-01 -3.33576918e-01 3.88094068e-01
3.54273260e-01 5.95649719e-01 -1.77115726e+00 6.99849129e-01
8.72931421e-01 1.02883136e+00 -2.14272946e-01 6.43720806e-01
-4.11885142e-01 1.27203321e+00 -5.21710038e-01 -2.86206156e-01
-1.28619835e-01 -6.22206628e-01 -8.83402705e-01 9.27243605e-02
-3.15749720e-02 9.32736546e-02 1.17758334e-01 3.28747779e-01
-8.93512368e-02 -1.21740162e-01 9.21346009e-01 -1.74849367e+00
-6.62252903e-01 1.06538570e+00 1.64379105e-01 3.99031378e-02
6.06521189e-01 3.70802343e-01 1.08825397e+00 -5.05164385e-01
5.21138251e-01 1.27745318e+00 6.50791466e-01 3.35231036e-01
-1.27801633e+00 -3.99110675e-01 2.82662600e-01 4.36734974e-01
-6.56826258e-01 -1.35400081e-02 2.91090757e-01 -8.28228235e-01
1.24913001e+00 7.39412248e-01 8.60946357e-01 1.04778099e+00
6.62523210e-01 4.74369884e-01 1.35721123e+00 -4.53844130e-01
6.88698471e-01 3.59643549e-01 6.01111293e-01 3.00552756e-01
6.52498007e-01 3.65956157e-01 -5.42275488e-01 3.87271307e-02
5.19956529e-01 4.86407012e-01 -2.01544061e-01 -3.98346633e-01
-1.22098529e+00 7.59078145e-01 2.89412856e-01 5.52942395e-01
-5.82344651e-01 1.18967786e-01 2.98910350e-01 5.43834209e-01
2.12872714e-01 9.19434607e-01 -7.87788153e-01 -2.21160859e-01
-8.15430164e-01 2.73647368e-01 1.26465166e+00 8.56282055e-01
7.61730134e-01 1.27127618e-01 -6.80990279e-01 -1.02601081e-01
5.08824706e-01 -1.23096205e-01 3.92824024e-01 -7.77628779e-01
1.56127065e-01 1.18539178e+00 3.11431438e-02 -6.68000579e-01
-6.22204185e-01 -4.14109647e-01 -5.80663443e-01 4.11383063e-01
5.13221323e-01 -3.07415035e-02 -8.76963198e-01 1.21641421e+00
-8.59060287e-02 -2.25238547e-01 2.00404897e-02 5.19783258e-01
7.13759720e-01 5.99924505e-01 5.30590713e-01 -3.76375318e-01
1.25830626e+00 -9.12986040e-01 -1.00167537e+00 -2.22173765e-01
6.09604597e-01 -2.17486501e-01 1.09845853e+00 6.00865066e-01
-1.05838168e+00 -6.76658273e-01 -1.01410711e+00 1.75985590e-01
-5.68743825e-01 -1.22114100e-01 9.68726337e-01 4.48993802e-01
-6.33254766e-01 1.05700946e+00 -9.37472999e-01 -5.44115603e-01
5.68225347e-02 4.25689280e-01 -2.70750046e-01 3.24776381e-01
-9.46890175e-01 1.27569902e+00 4.95180786e-01 -8.83281454e-02
-8.22012007e-01 -8.05727839e-01 -6.72506809e-01 9.07732248e-01
2.98051506e-01 -8.74756277e-01 1.30487573e+00 -9.15107846e-01
-1.20458066e+00 2.88370311e-01 -5.45118079e-02 -6.85589254e-01
8.21163535e-01 -2.65718609e-01 -5.04665673e-01 -2.41113722e-01
-1.35648280e-01 1.45883322e-01 4.39197570e-01 -1.22500813e+00
-7.98658550e-01 -2.29629233e-01 6.85978234e-02 -2.63793021e-01
1.10761046e-01 -1.83740258e-01 1.68510497e-01 -1.96507618e-01
1.84835345e-01 -5.95652759e-01 -3.00895244e-01 -3.09717417e-01
-6.01902783e-01 -1.74217120e-01 5.03587842e-01 -3.57104540e-01
1.31346083e+00 -1.94017494e+00 7.40135908e-02 3.18734020e-01
6.04570568e-01 -3.41605961e-01 3.37499142e-01 9.04108107e-01
-6.32018685e-01 4.71042693e-01 -1.09353244e-01 -3.13765645e-01
5.38127184e-01 1.84673652e-01 -3.78901929e-01 1.61520705e-01
4.05519843e-01 6.39707088e-01 -6.35036230e-01 -1.92731321e-01
6.36114955e-01 2.58665830e-01 -2.96280384e-01 3.43473673e-01
-1.96234688e-01 5.69258869e-01 -6.07023120e-01 6.12665594e-01
8.78586322e-02 -3.41522545e-01 1.00187056e-01 1.87463731e-01
-2.74413645e-01 4.21442538e-01 -1.17426646e+00 9.54199493e-01
-6.17122293e-01 6.65771961e-01 -5.65322459e-01 -5.82215548e-01
1.03956389e+00 5.82128644e-01 5.08694172e-01 -3.81387025e-01
1.33465916e-01 1.94795325e-01 2.53833920e-01 -5.59321404e-01
3.49183202e-01 -1.85064644e-01 1.86295345e-01 6.29429340e-01
2.54485067e-02 -9.99926478e-02 2.35346094e-01 -3.05130482e-01
1.36629355e+00 1.17305003e-01 1.00194252e+00 -2.94876695e-01
6.47082806e-01 1.19881816e-01 3.78667206e-01 8.79543424e-01
-8.53212178e-02 6.39960587e-01 9.32764471e-01 -8.36810291e-01
-8.42969596e-01 -9.41944778e-01 -2.70045578e-01 6.74180388e-01
-2.06299096e-01 -6.47256672e-01 -7.23983049e-01 -7.71059692e-01
1.79195479e-01 1.38941288e+00 -9.38902318e-01 -3.72948170e-01
-1.09705240e-01 -3.70314300e-01 7.42260143e-02 4.69830841e-01
1.47470534e-01 -1.29537106e+00 -9.20681238e-01 3.06280673e-01
3.45986277e-01 -6.60183012e-01 5.55746675e-01 7.06743360e-01
-1.23847461e+00 -1.28572237e+00 2.94487059e-01 1.37646392e-01
3.01630706e-01 -3.14229995e-01 1.37034059e+00 1.87831670e-01
5.37522882e-02 4.27883357e-01 -4.30000603e-01 -1.00931334e+00
-7.86508143e-01 1.65531665e-01 -2.99801111e-01 -1.17843576e-01
7.33708262e-01 -4.36477154e-01 -2.70547420e-01 1.85869485e-01
-8.85951757e-01 1.01097137e-01 8.06628823e-01 4.02275801e-01
5.28570414e-01 8.27338770e-02 4.15271401e-01 -1.08053076e+00
8.25821519e-01 -5.61302066e-01 -1.67115226e-01 3.22055608e-01
-1.49619770e+00 4.17226404e-01 5.22127986e-01 -1.70499921e-01
-1.07093334e+00 -9.84808207e-02 2.60743558e-01 -7.28019401e-02
-7.35677898e-01 6.06013894e-01 -3.72428149e-01 5.75856924e-01
7.38436878e-01 -1.65251240e-01 -2.65116721e-01 -2.95458883e-01
4.10667360e-01 2.78273791e-01 6.53643459e-02 -3.38730544e-01
7.79730916e-01 3.75496417e-01 1.61464766e-01 -1.42994672e-01
-4.70964104e-01 -7.06589520e-02 -5.44066727e-01 -3.69545043e-01
9.30778921e-01 -2.52756447e-01 -7.61498749e-01 -3.28875005e-01
-1.19767535e+00 -2.00089037e-01 -1.07630038e+00 7.12987781e-01
-9.02164340e-01 -1.73321053e-01 -5.45742214e-01 -9.26334739e-01
-2.52345294e-01 -1.34241319e+00 8.32373142e-01 1.17234714e-01
-1.18480277e+00 -1.14842248e+00 -1.21192016e-01 1.72632977e-01
3.42007339e-01 5.49150407e-01 1.37099707e+00 -1.49026012e+00
-8.02440405e-01 -5.79668164e-01 -6.50383085e-02 -4.21556644e-02
2.15619072e-01 2.04623714e-01 -1.09949028e+00 3.39385927e-01
2.82529026e-01 5.06147385e-01 5.33660650e-01 4.52911019e-01
1.05001736e+00 -3.85673732e-01 -4.21082199e-01 2.72603869e-01
1.29137576e+00 2.56748021e-01 5.62978625e-01 8.82837057e-01
5.14637470e-01 1.18331432e+00 6.12653792e-01 5.53409457e-01
2.92222518e-02 2.83710539e-01 7.35201716e-01 -5.35489880e-02
6.71545863e-02 -2.02675551e-01 2.29268387e-01 1.35959595e-01
-3.74720812e-01 -2.32709199e-01 -1.04069841e+00 2.28348747e-01
-2.05445743e+00 -1.10628402e+00 -8.68349314e-01 2.08693361e+00
1.39619425e-01 4.06504333e-01 -1.99227527e-01 6.49772882e-01
4.16902661e-01 -6.03909269e-02 -2.14881256e-01 -9.61580396e-01
1.83377296e-01 3.88274193e-02 3.03249747e-01 4.69868928e-01
-7.03867793e-01 2.06627637e-01 6.69502783e+00 2.42283821e-01
-8.03577483e-01 1.96515393e-04 4.82434243e-01 -9.01176631e-02
-5.90129197e-01 2.95023024e-01 -4.64649558e-01 1.83292240e-01
1.34193909e+00 -4.47512865e-01 1.68050781e-01 1.09424806e+00
8.20793152e-01 9.19539109e-02 -2.12046909e+00 4.74736154e-01
-3.48805845e-01 -1.28288293e+00 2.49514252e-01 3.51027966e-01
5.41493475e-01 -6.69745564e-01 -2.68283766e-03 2.51979440e-01
1.53655991e-01 -1.50614953e+00 1.05031562e+00 9.16401327e-01
-1.43106908e-01 -4.23130244e-01 9.78978336e-01 7.13499188e-02
-9.62075710e-01 -6.12338662e-01 5.82384132e-02 -5.54700851e-01
1.07467316e-01 6.72653437e-01 -1.04087639e+00 7.38514304e-01
6.59366906e-01 7.08204389e-01 -6.44121468e-01 9.07254517e-01
-5.08835495e-01 6.23152971e-01 2.19322994e-01 2.62703896e-02
9.67209637e-02 -4.51041125e-02 4.74347323e-01 1.38990033e+00
5.22839367e-01 -4.57041502e-01 -3.34125519e-01 1.42954969e+00
6.41944647e-01 -5.91469891e-02 -6.07448637e-01 7.89621398e-02
2.07655460e-01 1.14374697e+00 -7.75960088e-01 -4.14093226e-01
-2.82729715e-01 4.58565235e-01 -2.47790173e-01 2.47053832e-01
-7.63190091e-01 5.67480586e-02 6.56714082e-01 4.18603480e-01
-4.08611819e-02 1.60106450e-01 -9.69767809e-01 -5.85898101e-01
-7.98826963e-02 -9.17148292e-01 4.78594869e-01 -9.87615466e-01
-1.29415381e+00 6.14567935e-01 3.86963099e-01 -1.24891579e+00
-6.12665117e-01 -6.20580077e-01 -8.10762584e-01 9.50112998e-01
-1.08802390e+00 -1.16992390e+00 -6.68448985e-01 7.02734068e-02
5.58535218e-01 -7.07557797e-02 7.66866982e-01 -1.53963372e-01
-5.46403050e-01 -2.10061863e-01 -3.06188613e-01 -5.25426924e-01
4.18149978e-01 -1.78034461e+00 2.78016150e-01 5.51789105e-01
5.01728058e-02 8.12155843e-01 1.34355700e+00 -5.51496565e-01
-1.11060941e+00 -9.53414977e-01 1.25647902e+00 -9.30809855e-01
7.69900501e-01 1.31065890e-01 -1.12406075e+00 1.08145380e+00
2.90960133e-01 -5.69113731e-01 7.92993426e-01 3.00296456e-01
6.47294987e-03 -1.66518211e-01 -1.03187549e+00 5.71534812e-01
8.81987154e-01 -3.26848745e-01 -9.10928667e-01 1.96546406e-01
6.47676229e-01 -5.87619357e-02 -1.08746040e+00 3.28298450e-01
4.09444183e-01 -1.37377679e+00 6.37269378e-01 -8.84021997e-01
7.63215721e-01 -3.82561922e-01 1.42313287e-01 -1.11884880e+00
-5.32660127e-01 -3.53711396e-01 -4.44664925e-01 1.32817161e+00
7.37838686e-01 -5.17731667e-01 6.41381323e-01 1.09513760e+00
-2.81788349e-01 -5.98762572e-01 -6.33847356e-01 -5.58237731e-01
-2.03176305e-01 -7.24683523e-01 1.20342529e+00 9.21828985e-01
2.60049015e-01 2.57972419e-01 1.31392539e-01 1.05968781e-01
2.14643478e-01 1.52084172e-01 7.49027371e-01 -1.78702068e+00
-2.99529612e-01 -7.60072410e-01 -6.20781362e-01 -2.35455513e-01
-1.34153754e-01 -6.83632553e-01 -3.25436682e-01 -2.00115895e+00
1.49024844e-01 2.23234203e-02 -3.62980574e-01 5.21535993e-01
-7.81313330e-02 -6.62562191e-01 4.31251377e-01 4.70137805e-01
-2.52683997e-01 3.06631893e-01 8.97440493e-01 -1.79710969e-01
-4.08762515e-01 2.42237166e-01 -7.18345165e-01 8.50879729e-01
8.24713051e-01 -6.43207312e-01 -5.66153109e-01 -1.03817903e-01
3.05058360e-01 1.25522455e-02 5.24220347e-01 -1.28483248e+00
8.15828219e-02 -3.96107793e-01 3.83637547e-01 -1.74507901e-01
-5.93313463e-02 -1.36145687e+00 8.39293301e-01 7.31021941e-01
-8.05174649e-01 3.62290710e-01 1.26117080e-01 6.44774914e-01
-3.35568756e-01 -3.04363400e-01 1.87328488e-01 -1.26712769e-01
-6.13421082e-01 4.90371371e-03 -6.45560503e-01 -6.67045832e-01
1.24032450e+00 -7.22655654e-01 -2.10973278e-01 -4.62583601e-01
-1.17075181e+00 -1.33098900e-01 2.54077256e-01 4.29629564e-01
2.11530045e-01 -8.58992755e-01 -5.41535437e-01 -4.37926836e-02
3.80488813e-01 -2.84973323e-01 2.88044643e-02 7.99943030e-01
-5.31008840e-01 6.41335011e-01 -3.62330973e-01 -5.42254806e-01
-8.94727290e-01 7.89409280e-01 5.12013793e-01 -6.18739605e-01
-6.06034935e-01 -8.40321407e-02 2.43906796e-01 -2.95395285e-01
9.13244188e-02 -1.06711698e+00 -5.41330457e-01 3.92120928e-02
2.31527507e-01 6.46636188e-01 2.18306005e-01 -3.71408351e-02
-1.21505074e-01 1.28279090e-01 3.72743785e-01 -9.99444127e-02
1.26447916e+00 1.37340957e-02 -2.95941327e-02 9.62966859e-01
2.73333281e-01 -2.49665454e-01 -1.19757307e+00 4.96031314e-01
6.58997178e-01 -1.19240902e-01 -1.39598683e-01 -1.07107937e+00
-6.80436671e-01 1.06155014e+00 4.90541607e-01 1.05001366e+00
8.81098211e-01 3.20430845e-02 -1.92959413e-01 2.38457158e-01
6.97776973e-02 -7.54792988e-01 -3.32768291e-01 1.90417290e-01
1.14141309e+00 -8.34688008e-01 1.49295568e-01 -5.72786212e-01
-6.79117501e-01 1.40552223e+00 5.49088299e-01 2.95824856e-01
5.03423810e-01 6.42990917e-02 -1.56612545e-01 -4.85229999e-01
-8.60372126e-01 -2.19746809e-02 2.46515125e-01 6.84429169e-01
7.54738331e-01 2.56242365e-01 -4.59530294e-01 1.07594192e+00
-5.04989266e-01 2.36187533e-01 5.90080500e-01 7.70649433e-01
-3.44379812e-01 -9.81305182e-01 -5.72782099e-01 5.65826356e-01
-2.91903317e-01 1.91158935e-01 -8.13345432e-01 1.34701824e+00
2.43449450e-01 1.33820784e+00 1.44650921e-01 -4.44270670e-01
8.05987060e-01 2.73808420e-01 -1.44914007e-02 -7.50005841e-01
-1.15248120e+00 -4.93881226e-01 3.69464934e-01 -6.02507293e-01
-2.13686958e-01 -7.87502229e-01 -1.31719947e+00 -6.53248489e-01
-1.37623578e-01 2.42738739e-01 6.86496377e-01 1.35901487e+00
2.50542790e-01 1.17438698e+00 4.36479189e-02 -4.44026113e-01
-5.16351938e-01 -1.10864615e+00 -5.00885963e-01 5.47913730e-01
1.09299667e-01 -6.17462039e-01 -6.28273129e-01 1.00075848e-01]
|
[8.637079238891602, 5.902775287628174]
|
3f83c4e4-2306-4474-b105-b229b9751fb8
|
temporally-consistent-online-depth-estimation-1
|
2304.07435
| null |
https://arxiv.org/abs/2304.07435v2
|
https://arxiv.org/pdf/2304.07435v2.pdf
|
Temporally Consistent Online Depth Estimation Using Point-Based Fusion
|
Depth estimation is an important step in many computer vision problems such as 3D reconstruction, novel view synthesis, and computational photography. Most existing work focuses on depth estimation from single frames. When applied to videos, the result lacks temporal consistency, showing flickering and swimming artifacts. In this paper we aim to estimate temporally consistent depth maps of video streams in an online setting. This is a difficult problem as future frames are not available and the method must choose between enforcing consistency and correcting errors from previous estimations. The presence of dynamic objects further complicates the problem. We propose to address these challenges by using a global point cloud that is dynamically updated each frame, along with a learned fusion approach in image space. Our approach encourages consistency while simultaneously allowing updates to handle errors and dynamic objects. Qualitative and quantitative results show that our method achieves state-of-the-art quality for consistent video depth estimation.
|
['Lei Xiao', 'Douglas Lanman', 'Eric Penner', 'Numair Khan']
|
2023-04-15
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Khan_Temporally_Consistent_Online_Depth_Estimation_Using_Point-Based_Fusion_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Khan_Temporally_Consistent_Online_Depth_Estimation_Using_Point-Based_Fusion_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['3d-reconstruction']
|
['computer-vision']
|
[ 2.82655358e-01 -2.80418217e-01 1.90475732e-01 -1.27481133e-01
-5.47294021e-01 -4.81866777e-01 4.91040975e-01 -4.73436080e-02
-4.81675237e-01 7.45712042e-01 -9.65948217e-03 2.51559615e-01
1.16754584e-01 -5.71488142e-01 -7.06823230e-01 -7.28384912e-01
-1.89038496e-02 3.13698351e-01 8.41898263e-01 6.06541857e-02
5.22293270e-01 5.79377472e-01 -1.77734780e+00 1.55697107e-01
6.48941696e-01 9.99212861e-01 5.66347480e-01 1.01809883e+00
-1.05982795e-02 9.06748116e-01 -4.52890128e-01 -1.88337296e-01
4.95543689e-01 -3.67317230e-01 -5.06358683e-01 5.42830706e-01
7.96212375e-01 -8.72815311e-01 -2.17021510e-01 1.01330197e+00
3.60240906e-01 2.69780278e-01 1.96008489e-01 -1.21211410e+00
2.83884674e-01 -2.39497632e-01 -6.81009591e-01 3.86664540e-01
6.45300925e-01 1.78357005e-01 4.16951418e-01 -7.33071089e-01
9.78883505e-01 9.17329192e-01 5.71052074e-01 5.73253572e-01
-1.15636551e+00 -2.73743838e-01 3.08448970e-01 3.88971537e-01
-1.22647750e+00 -7.32342660e-01 7.80452371e-01 -4.68164086e-01
8.72161210e-01 9.41451862e-02 8.88079226e-01 7.57160306e-01
2.90032864e-01 4.11184728e-01 8.50355625e-01 -3.81173044e-01
5.50934851e-01 -1.70148820e-01 -4.67548609e-01 4.12051678e-01
1.31751448e-01 1.34431273e-01 -1.00937986e+00 7.30785131e-02
1.11662161e+00 4.01446298e-02 -5.17897785e-01 -5.59695542e-01
-1.32978332e+00 3.00974607e-01 -1.55627765e-02 -2.19732057e-02
-5.60433984e-01 3.87781709e-01 8.50842297e-02 2.95142382e-01
5.72550893e-01 1.07996456e-01 -3.98565292e-01 -5.57043195e-01
-1.16283691e+00 3.70517343e-01 5.82604766e-01 9.40732539e-01
7.65705943e-01 -5.47113419e-02 3.99640828e-01 4.10886049e-01
2.48732567e-01 4.65600610e-01 1.71014637e-01 -1.80095637e+00
3.28075469e-01 9.40482244e-02 3.56937975e-01 -1.10806525e+00
-1.24602899e-01 1.68887243e-01 -5.37984908e-01 9.02668297e-01
6.08667970e-01 8.47472548e-02 -6.54109597e-01 1.51764846e+00
7.30773926e-01 5.95667183e-01 -1.13795772e-01 1.07712257e+00
3.84564728e-01 5.83418608e-01 -4.17224288e-01 -6.05899513e-01
7.44848430e-01 -6.77848041e-01 -8.60982358e-01 -4.18130219e-01
6.25954494e-02 -7.14680135e-01 3.91647935e-01 8.81445050e-01
-1.48852694e+00 -3.69006306e-01 -9.55944598e-01 -2.23875284e-01
2.67492771e-01 -4.72836345e-01 2.45394006e-01 3.16428751e-01
-1.19173861e+00 7.06667185e-01 -1.10608494e+00 -3.64931613e-01
1.78444937e-01 3.18576455e-01 -5.83788872e-01 -4.26377922e-01
-6.38681531e-01 8.96844149e-01 1.64895788e-01 1.15203334e-03
-5.81213593e-01 -6.32769823e-01 -8.34467888e-01 -2.98467368e-01
5.57293057e-01 -8.53112042e-01 1.27562439e+00 -1.00635922e+00
-1.65369785e+00 6.99766815e-01 -4.96994406e-01 -3.07414174e-01
8.64487350e-01 -3.14428866e-01 1.32962212e-01 4.84956831e-01
-8.33048820e-02 6.64291859e-01 9.40006971e-01 -1.38601077e+00
-8.55951190e-01 -4.19101685e-01 1.44575119e-01 4.88264114e-01
1.02315977e-01 -3.22976857e-01 -7.99190998e-01 -3.39291722e-01
5.65282702e-01 -8.04722369e-01 -1.90217927e-01 7.97421336e-01
3.25526744e-01 4.42790627e-01 1.04754877e+00 -5.97514331e-01
9.54901516e-01 -1.91480124e+00 3.96385074e-01 -3.04735117e-02
2.69543290e-01 -5.99487871e-02 1.90181166e-01 2.15410903e-01
3.71199638e-01 -3.81664246e-01 -1.48630217e-01 -7.71535933e-01
-4.81053293e-01 5.47947049e-01 -9.00234878e-02 6.85661256e-01
-7.78416321e-02 4.04071242e-01 -1.09783757e+00 -5.59475660e-01
7.02344358e-01 6.48183167e-01 -7.58595288e-01 2.02401757e-01
-2.76997596e-01 7.97338188e-01 -8.29640701e-02 6.63386703e-01
9.23802376e-01 -1.62437916e-01 2.42837861e-01 -8.54218379e-02
-4.21237797e-01 2.63671614e-02 -1.57268047e+00 2.18909216e+00
-3.10397089e-01 9.54138577e-01 3.10498834e-01 -5.58922827e-01
5.04345238e-01 1.93623319e-01 7.85849750e-01 -6.73967063e-01
5.39430454e-02 1.99223801e-01 -2.66416878e-01 -6.22484148e-01
8.04813981e-01 -1.66407019e-01 5.82855046e-01 2.75442600e-01
-2.42428929e-01 -6.11373901e-01 1.25561684e-01 2.16776237e-01
1.24311161e+00 3.89327168e-01 2.57905960e-01 1.02891110e-01
2.84371465e-01 -7.15497881e-02 7.45090187e-01 4.02439624e-01
-2.79458880e-01 1.20335424e+00 2.12003902e-01 -3.76067072e-01
-1.16606903e+00 -9.28601861e-01 -2.51681488e-02 1.91049874e-01
6.16615236e-01 -4.33581173e-01 -4.03767496e-01 -2.43908539e-01
-2.01057389e-01 1.77162081e-01 -4.98381317e-01 2.57877588e-01
-6.88803315e-01 -1.88741609e-01 -2.50133961e-01 4.09297794e-01
4.84731823e-01 -5.87621152e-01 -1.33231413e+00 4.74453837e-01
-5.08423388e-01 -1.49209809e+00 -3.87515754e-01 -1.11630350e-01
-1.29224706e+00 -1.09434605e+00 -6.09407008e-01 -3.14937353e-01
6.13417387e-01 6.54531837e-01 1.24928892e+00 2.39327103e-01
-1.37392595e-01 6.23338223e-01 -2.98987389e-01 -1.42153502e-01
-3.01718414e-01 -5.93919933e-01 1.36235552e-02 -1.64299682e-02
-1.11102164e-01 -7.47774363e-01 -8.13167393e-01 3.97632509e-01
-1.14529014e+00 3.04574311e-01 4.35396098e-03 7.01821744e-01
7.29448855e-01 1.43363640e-01 -1.07867941e-01 -3.90567958e-01
-1.73920289e-01 -1.54774562e-01 -9.78274345e-01 -1.83243647e-01
-3.89080018e-01 -1.58408180e-01 2.90269375e-01 -5.40600955e-01
-9.81660604e-01 3.70148301e-01 8.95695537e-02 -8.48114669e-01
8.70016664e-02 2.45548002e-02 -4.58768010e-02 -3.00484955e-01
4.46788609e-01 1.12889878e-01 2.45553687e-01 -1.29726604e-01
-1.18079968e-02 2.32591972e-01 7.09055901e-01 -2.90196180e-01
4.55465853e-01 1.09422064e+00 2.34808147e-01 -9.81749833e-01
-4.66400772e-01 -5.68780005e-01 -8.20855498e-01 -7.44747221e-01
7.57386506e-01 -9.24264550e-01 -6.68751955e-01 6.23261690e-01
-1.41414416e+00 -5.38255572e-01 -2.22473711e-01 4.77979988e-01
-8.39199781e-01 6.80735052e-01 -4.49308425e-01 -9.31180537e-01
1.18590333e-01 -1.08365619e+00 1.33174253e+00 1.61039919e-01
-1.25973910e-01 -1.15524113e+00 2.56482452e-01 2.57017970e-01
2.45498389e-01 4.38312799e-01 -2.93889027e-02 5.59884429e-01
-1.18063891e+00 1.29320070e-01 1.02401167e-01 2.00336188e-01
7.46440962e-02 1.60021231e-01 -1.00297761e+00 -2.26565167e-01
2.14803055e-01 -1.33047691e-02 5.31617880e-01 6.54781580e-01
8.37187886e-01 -6.34829700e-02 -1.01297557e-01 8.23339820e-01
1.68771160e+00 2.46382743e-01 8.67621779e-01 4.60862130e-01
5.94794214e-01 7.75392473e-01 8.25706005e-01 6.43691123e-01
5.06465912e-01 9.94808316e-01 7.11957753e-01 2.83246428e-01
-1.91915587e-01 1.15053080e-01 2.15981275e-01 4.61066246e-01
-2.74525374e-01 -3.55136633e-01 -8.42042685e-01 4.81516063e-01
-2.05844069e+00 -1.18809569e+00 -1.95943490e-01 2.45155120e+00
6.00551307e-01 -4.84689511e-03 -1.91065267e-01 4.27253425e-01
4.24958944e-01 1.68628562e-02 -4.29999083e-01 1.02793150e-01
-6.35384545e-02 -1.59060314e-01 4.56172168e-01 8.26256156e-01
-7.49630094e-01 6.28071368e-01 6.31932974e+00 1.15361191e-01
-1.20152736e+00 9.30766538e-02 3.30892563e-01 -4.81778264e-01
-2.49483913e-01 1.38878122e-01 -4.29213345e-01 6.69892430e-01
5.03605843e-01 7.15667605e-02 3.23282629e-01 4.45820212e-01
4.23123389e-01 -9.94472504e-01 -1.12643230e+00 1.51320982e+00
3.58720630e-01 -1.35523117e+00 -4.23920304e-01 1.63754702e-01
1.00527406e+00 5.35129495e-02 -2.52226144e-01 -4.69911754e-01
-9.31985974e-02 -5.48980892e-01 8.70060086e-01 7.31921852e-01
6.26972854e-01 -5.36423981e-01 5.27562201e-01 4.31228131e-01
-1.07946336e+00 1.62868097e-01 -2.65985221e-01 -3.55812311e-01
6.92889810e-01 7.86951959e-01 -4.71687943e-01 4.13581550e-01
9.17324483e-01 1.01129174e+00 -2.48108745e-01 1.22165287e+00
1.01037070e-01 -1.51304334e-01 -6.11318052e-01 4.10227120e-01
-1.61076367e-01 -3.15838933e-01 6.42169893e-01 7.06114233e-01
5.90103328e-01 4.76684600e-01 2.40681581e-02 5.69562018e-01
3.68479252e-01 -4.44101125e-01 -7.45690286e-01 5.18883288e-01
3.80123794e-01 9.04272020e-01 -8.37065637e-01 -4.66919363e-01
-4.70294923e-01 1.28701818e+00 6.62267283e-02 2.64699757e-01
-5.83030164e-01 2.21101269e-01 8.24753582e-01 3.84053051e-01
3.01200062e-01 -7.43368089e-01 -2.30985135e-01 -1.57572532e+00
3.18795264e-01 -4.65835899e-01 1.98478654e-01 -1.06860256e+00
-8.06782722e-01 3.61437231e-01 5.16770929e-02 -1.64038408e+00
-6.32646382e-01 -3.13142091e-01 -3.31309617e-01 4.96557236e-01
-1.71805358e+00 -6.01259768e-01 -8.29082072e-01 6.49550021e-01
8.23479712e-01 4.52101827e-01 2.98923045e-01 3.96615952e-01
-1.46253509e-02 -7.60644674e-02 1.03671923e-01 -5.34146965e-01
7.79741943e-01 -9.57467675e-01 1.31669819e-01 1.23151433e+00
-7.84601644e-02 7.80068487e-02 1.00988448e+00 -6.32538736e-01
-1.57959414e+00 -5.78461945e-01 7.74158835e-01 -6.29451752e-01
2.88644373e-01 -1.02006808e-01 -9.53238010e-01 3.76542181e-01
-6.25517517e-02 4.62830603e-01 7.01047629e-02 -5.78125358e-01
1.23264333e-02 -1.13052316e-01 -1.19477904e+00 3.55146199e-01
1.13486707e+00 -4.11762834e-01 -9.28419456e-02 6.65594637e-02
2.97282815e-01 -9.13994551e-01 -5.15374243e-01 1.88758135e-01
7.78244078e-01 -1.56211209e+00 1.01149893e+00 3.21968704e-01
4.89309222e-01 -6.22719407e-01 -2.66660720e-01 -1.00793219e+00
1.88691273e-01 -5.90658665e-01 -4.89919811e-01 8.27966154e-01
-3.30271810e-01 -3.40611994e-01 1.00820303e+00 8.61957550e-01
3.80858034e-02 -3.85204136e-01 -1.15043402e+00 -7.24559367e-01
-4.46518779e-01 -6.09872341e-01 5.29417433e-02 8.45643342e-01
-2.71744817e-01 -1.78064674e-01 -5.37192941e-01 3.48354578e-01
1.10438025e+00 -1.24803051e-01 1.07551980e+00 -1.05294263e+00
-4.55422699e-01 -2.06948251e-01 -8.55523109e-01 -1.33461988e+00
-1.56422824e-01 -8.03384855e-02 3.62671137e-01 -1.55362678e+00
-1.38608217e-01 -3.11277002e-01 4.11803544e-01 -8.48037079e-02
-2.62111425e-02 4.48759556e-01 2.86209345e-01 1.94336861e-01
-6.00025594e-01 4.52951193e-01 1.09541678e+00 1.58826575e-01
-2.04351038e-01 -2.23933280e-01 6.69450536e-02 7.03717172e-01
4.93808061e-01 -3.91054988e-01 -5.80331922e-01 -7.84883678e-01
2.63156176e-01 4.75502968e-01 4.82906580e-01 -1.20507812e+00
5.40691555e-01 -2.95279831e-01 4.17778939e-01 -6.72484875e-01
9.12913322e-01 -1.16071784e+00 5.60330153e-01 3.73003185e-01
2.79417068e-01 3.15599591e-01 1.49632111e-01 7.17527449e-01
-3.68904829e-01 -1.62927821e-01 9.32269335e-01 -2.65836507e-01
-1.01243198e+00 4.89304543e-01 -2.20660001e-01 -1.18475586e-01
1.07043195e+00 -7.20389783e-01 -2.14755461e-02 -6.23813510e-01
-5.42875290e-01 1.87375098e-01 1.13524294e+00 2.47803167e-01
9.73385513e-01 -1.11234379e+00 -5.95510542e-01 3.18756282e-01
5.33967614e-02 4.90593344e-01 5.07144570e-01 8.51147413e-01
-9.90643919e-01 -1.95980802e-01 -1.62070557e-01 -1.17038536e+00
-1.51867878e+00 2.18639180e-01 4.07446235e-01 1.11035831e-01
-8.99908960e-01 8.30370784e-01 9.69617739e-02 2.57766187e-01
2.81050086e-01 -4.07419413e-01 1.32537544e-01 -7.31450990e-02
7.60052264e-01 5.05761623e-01 1.50141791e-01 -4.73548859e-01
-2.72691101e-01 1.01125252e+00 1.66300073e-01 -6.08742952e-01
1.18953812e+00 -7.74348378e-01 -5.47749288e-02 5.22349477e-01
1.02465856e+00 -5.45015605e-03 -1.93301964e+00 -1.69710308e-01
-2.78422326e-01 -1.22048521e+00 2.00036809e-01 -2.69716799e-01
-1.06948459e+00 7.14300036e-01 5.49983680e-01 -1.05187275e-01
1.26768529e+00 -1.57527104e-01 8.08257699e-01 1.23897783e-01
7.31820464e-01 -1.03869367e+00 2.60896951e-01 4.97441620e-01
6.05576217e-01 -1.54280567e+00 2.58246928e-01 -3.51209432e-01
-3.65378618e-01 1.26511681e+00 5.84545851e-01 -6.65648803e-02
5.42326272e-01 5.08003473e-01 1.41250625e-01 9.99015346e-02
-9.90675449e-01 -3.46882306e-02 -1.23861395e-01 6.36904418e-01
1.73541866e-02 -5.75458109e-01 -1.69334248e-01 -4.16507065e-01
1.52700886e-01 9.82516110e-02 9.41793144e-01 1.28528428e+00
-3.32776099e-01 -1.08719373e+00 -6.13083780e-01 9.09788255e-03
-2.15227008e-01 1.72870696e-01 7.96406642e-02 5.37691593e-01
7.45906085e-02 9.22039092e-01 3.54382128e-01 -1.44989848e-01
1.19051605e-01 -3.27537358e-01 1.00638521e+00 -4.23526824e-01
-1.64488867e-01 2.64238417e-01 -1.01886481e-01 -1.01364839e+00
-9.91681278e-01 -9.06494498e-01 -1.22342408e+00 -4.18509066e-01
-2.39275739e-01 -2.82024145e-01 9.32470441e-01 6.81526303e-01
2.43131101e-01 3.75728339e-01 5.85646987e-01 -1.47474813e+00
7.78161362e-02 -4.83323157e-01 -4.74745542e-01 3.83075148e-01
8.39590013e-01 -6.71507895e-01 -5.92666984e-01 4.26906139e-01]
|
[8.785134315490723, -2.2335731983184814]
|
6b7058ff-8250-4d39-867f-d54792770199
|
performance-analysis-of-a-foreground
|
2105.12311
| null |
https://arxiv.org/abs/2105.12311v1
|
https://arxiv.org/pdf/2105.12311v1.pdf
|
Performance Analysis of a Foreground Segmentation Neural Network Model
|
In recent years the interest in segmentation has been growing, being used in a wide range of applications such as fraud detection, anomaly detection in public health and intrusion detection. We present an ablation study of FgSegNet_v2, analysing its three stages: (i) Encoder, (ii) Feature Pooling Module and (iii) Decoder. The result of this study is a proposal of a variation of the aforementioned method that surpasses state of the art results. Three datasets are used for testing: CDNet2014, SBI2015 and CityScapes. In CDNet2014 we got an overall improvement compared to the state of the art, mainly in the LowFrameRate subset. The presented approach is promising as it produces comparable results with the state of the art (SBI2015 and Cityscapes datasets) in very different conditions, such as different lighting conditions.
|
['Bruno Faria', 'André Leite Ferreira', 'António Ramires Fernandes', 'Joel Tomás Morais']
|
2021-05-26
| null | null | null | null |
['foreground-segmentation']
|
['computer-vision']
|
[ 1.60159826e-01 -5.80584034e-02 1.75685346e-01 -2.90909022e-01
-3.48082632e-01 -2.61833370e-01 8.78163993e-01 4.58986014e-01
-1.00039959e+00 7.96545267e-01 8.54740441e-02 -3.44952315e-01
-7.11462200e-02 -8.61287951e-01 -3.60736340e-01 -4.08960640e-01
-1.75460652e-01 4.53154981e-01 9.53855634e-01 -1.99083626e-01
2.13155031e-01 5.96757352e-01 -1.90091109e+00 5.02117395e-01
6.44983709e-01 1.16866779e+00 -6.98068067e-02 6.08541250e-01
-1.61275074e-01 5.31825602e-01 -8.67182255e-01 -3.58246326e-01
1.35774434e-01 -4.00807261e-01 -9.53590751e-01 -8.67576525e-02
1.63556755e-01 2.43695956e-02 -7.72503763e-02 9.12947536e-01
6.82929814e-01 -4.70194705e-02 3.25310946e-01 -9.60253596e-01
1.08680695e-01 3.59904766e-01 -3.98770303e-01 5.63206911e-01
2.18708828e-01 2.73266509e-02 5.69801569e-01 -5.37587881e-01
7.54551530e-01 1.09935713e+00 6.76879883e-01 2.23478213e-01
-1.06124127e+00 -5.04218400e-01 -2.63978858e-02 5.85520267e-01
-1.47929859e+00 -2.20103204e-01 1.24342665e-01 -1.59090176e-01
1.21236527e+00 3.38629246e-01 6.04312599e-01 1.12323010e+00
-6.96482360e-02 9.06086564e-01 1.33392644e+00 -4.23038483e-01
4.18454140e-01 3.13640684e-01 1.24884576e-01 4.38904196e-01
2.50683546e-01 7.68548399e-02 -4.03672367e-01 6.84030205e-02
3.86210978e-01 -3.13750297e-01 -1.49794996e-01 3.20872098e-01
-1.04814041e+00 9.04884100e-01 1.49062768e-01 8.87065709e-01
-5.46074450e-01 -3.92811969e-02 6.66269958e-01 5.26598811e-01
6.29417002e-01 -1.16920937e-02 -4.20268089e-01 -4.17180121e-01
-1.29313624e+00 3.99639964e-01 1.04996288e+00 7.49132276e-01
4.06374246e-01 7.68520832e-02 -1.59650937e-01 7.51394808e-01
3.34484875e-01 1.99500695e-01 6.52999341e-01 -3.69876444e-01
3.71136248e-01 3.87416065e-01 -1.47222653e-01 -6.37401760e-01
-7.60122657e-01 -5.48435271e-01 -6.78323388e-01 2.09807336e-01
5.40189028e-01 -3.10988784e-01 -1.16913736e+00 1.46564329e+00
6.87794685e-02 3.46809506e-01 7.40586370e-02 6.23952329e-01
1.07774246e+00 6.56638443e-01 7.06095994e-02 5.92689142e-02
1.33273065e+00 -6.72983646e-01 -6.20846033e-01 6.67704791e-02
4.14494187e-01 -1.14879274e+00 4.38281506e-01 7.41896927e-01
-9.94882345e-01 -7.48241007e-01 -1.06770456e+00 2.86426276e-01
-7.48292208e-01 -3.57987918e-02 5.11131823e-01 1.10062456e+00
-1.26901257e+00 6.13820672e-01 -6.51409745e-01 -8.61323535e-01
4.27577436e-01 3.96972388e-01 -2.18628317e-01 1.69944420e-01
-1.28646541e+00 8.99960637e-01 6.35659516e-01 -1.21125922e-01
-5.69385231e-01 -3.74776542e-01 -7.83629954e-01 -1.75132323e-02
3.42628658e-01 -3.98276955e-01 1.10271466e+00 -8.70277941e-01
-1.34498167e+00 1.23148298e+00 1.34421200e-01 -1.10100484e+00
9.71495569e-01 -1.51160970e-01 -7.42639244e-01 1.91608384e-01
-2.05555901e-01 7.97602177e-01 5.69689691e-01 -7.33499289e-01
-1.14495885e+00 -3.26493144e-01 -2.07611889e-01 -2.05680341e-01
1.55172080e-01 1.43317237e-01 -5.78005493e-01 -7.25956380e-01
-2.35893339e-01 -7.37198114e-01 -6.19952567e-02 -4.07979846e-01
-5.62554061e-01 -2.00372994e-01 9.48426306e-01 -5.19154608e-01
1.21110773e+00 -2.03102422e+00 -2.36988157e-01 3.68114978e-01
-1.92778453e-01 8.76431465e-01 9.32640284e-02 4.39607650e-01
-1.29360139e-01 2.05307946e-01 -3.23985338e-01 -3.69433075e-01
-1.92171544e-01 2.64396131e-01 7.70049915e-02 6.54309511e-01
2.29278862e-01 6.15339279e-01 -6.26388907e-01 -4.21844512e-01
6.17763162e-01 5.41111231e-01 -2.86502570e-01 -1.47922859e-01
-1.66846998e-02 4.49184209e-01 -4.22512770e-01 5.48268437e-01
7.03833878e-01 2.21050307e-01 -2.72045821e-01 1.01419218e-01
-4.78329986e-01 1.22664727e-01 -1.41123796e+00 1.74337959e+00
-2.35520661e-01 7.28969336e-01 -9.91836339e-02 -1.23029554e+00
1.16138947e+00 6.44079566e-01 5.25187790e-01 -9.09944892e-01
4.55133140e-01 4.80625242e-01 1.96355119e-01 -5.59440076e-01
3.91837150e-01 -1.06940471e-01 1.30760387e-01 4.53370102e-02
4.55975503e-01 1.22006416e-01 5.76013565e-01 -2.35820085e-01
1.04767597e+00 1.85007881e-02 2.76214898e-01 -5.38294971e-01
9.69154894e-01 -2.80106723e-01 2.48608842e-01 6.98554993e-01
-5.02473474e-01 7.68143952e-01 5.90066791e-01 -3.53118479e-01
-8.59497726e-01 -9.47418630e-01 -3.14093888e-01 3.60280663e-01
-4.73124571e-02 -2.46748865e-01 -1.06152534e+00 -5.12039185e-01
-2.54828990e-01 7.03774452e-01 -6.25377715e-01 7.96340555e-02
-5.67061007e-01 -1.10092592e+00 8.53521585e-01 2.90182799e-01
1.20982611e+00 -1.52227616e+00 -1.05568922e+00 5.13112724e-01
-5.67079559e-02 -1.37219024e+00 2.34163269e-01 2.04536989e-01
-8.95461798e-01 -1.11240900e+00 -8.08551848e-01 -4.71202850e-01
-3.33584696e-02 -2.77619392e-01 1.23512328e+00 3.45296375e-02
-6.85506523e-01 9.00815576e-02 -6.49671674e-01 -9.05518413e-01
-3.85102242e-01 4.87606823e-01 -5.56861281e-01 8.51839781e-02
5.74495673e-01 -3.43545943e-01 -6.04737282e-01 3.32065463e-01
-1.13159227e+00 -3.49114060e-01 5.02500892e-01 4.48443532e-01
3.94800365e-01 5.61982580e-02 6.38767600e-01 -1.37030900e+00
4.86462563e-01 -3.86623383e-01 -7.14876235e-01 -1.53095350e-01
-6.17253602e-01 -2.60266662e-01 6.14369512e-01 2.18375593e-01
-9.36069429e-01 -1.92334011e-01 -8.55269194e-01 1.35632351e-01
-7.91375458e-01 8.91560838e-02 1.28029138e-02 6.36564642e-02
6.01479530e-01 -7.07040494e-03 -9.56102461e-02 -7.30396271e-01
-1.52364686e-01 7.08404839e-01 5.62358499e-01 -6.16878718e-02
4.00941640e-01 5.73583663e-01 4.48942073e-02 -1.24753022e+00
-1.34335056e-01 -7.66872823e-01 -4.89275694e-01 -1.24961898e-01
1.04695821e+00 -5.43651760e-01 -5.53531885e-01 8.35465491e-01
-1.05282211e+00 -1.10162735e-01 -5.50379038e-01 5.15197098e-01
-3.59261632e-01 1.41328231e-01 -5.03950655e-01 -6.79686666e-01
-3.07871848e-01 -1.25950480e+00 8.81095231e-01 4.72961754e-01
-4.57210392e-02 -9.78673875e-01 2.87474506e-02 -8.55577737e-03
6.60266519e-01 6.56508446e-01 5.44952154e-01 -9.77144301e-01
-2.80407131e-01 -4.08639282e-01 -2.52195686e-01 5.40611386e-01
-2.23337471e-01 1.31972000e-01 -1.35882127e+00 -3.61399055e-02
-4.81280535e-02 5.99857867e-02 1.18609488e+00 4.11105186e-01
1.07799566e+00 3.78469855e-01 -1.23043343e-01 4.99932140e-01
1.70674849e+00 6.44151151e-01 1.14264572e+00 6.80685699e-01
-4.34949249e-02 6.19060278e-01 4.90455359e-01 5.01947820e-01
7.72612765e-02 6.55678213e-01 5.75070322e-01 -4.17796671e-01
-2.02253252e-01 3.69113654e-01 1.15022704e-01 1.98991507e-01
-3.67641628e-01 -6.47268116e-01 -8.26658309e-01 7.43482828e-01
-1.79383492e+00 -1.00097942e+00 -4.09184575e-01 2.41789103e+00
6.08837865e-02 6.27318799e-01 5.35425544e-01 7.28564978e-01
6.52152598e-01 2.44168550e-01 -1.22647986e-01 -9.90456998e-01
-4.96448092e-02 9.54922140e-01 5.42205870e-01 1.48648098e-01
-1.30312574e+00 7.03809202e-01 5.79511929e+00 8.61814797e-01
-1.22596014e+00 2.69199580e-01 7.62623012e-01 1.21714838e-01
2.98073292e-01 -4.67718393e-01 -6.60266399e-01 6.16552591e-01
1.35984123e+00 2.97508508e-01 5.04414886e-02 3.09496462e-01
1.58027798e-01 -5.15463352e-01 -7.17881918e-01 8.36526632e-01
1.19578764e-01 -1.10759366e+00 -3.69330645e-01 1.22227736e-01
3.87227982e-01 3.99954677e-01 -4.85999733e-02 2.89526463e-01
-8.37304592e-02 -1.04960489e+00 7.86568046e-01 3.81160647e-01
6.59587085e-01 -1.10666645e+00 1.26416695e+00 4.27447520e-02
-1.00477064e+00 1.41306162e-01 -1.28404304e-01 2.50274271e-01
3.02917570e-01 7.08365679e-01 -6.47148490e-01 9.58687484e-01
9.10431087e-01 4.76090491e-01 -6.14484727e-01 1.50895798e+00
-2.52285093e-01 7.94769287e-01 -4.95817959e-01 2.17255689e-02
7.54861116e-01 2.77392492e-02 5.99607646e-01 1.70883095e+00
2.34324187e-01 -2.16483131e-01 -1.11890793e-01 5.40382743e-01
8.13714340e-02 2.60333836e-01 -3.46176028e-01 3.39805633e-01
-4.81133051e-02 1.18203783e+00 -1.27959335e+00 -4.90183771e-01
-7.45816112e-01 9.50579226e-01 -4.42989528e-01 1.66619644e-01
-9.88325477e-01 -5.10725796e-01 4.63081270e-01 6.19389750e-02
7.22541273e-01 2.72292554e-01 -1.54800102e-01 -7.33637333e-01
-1.27731591e-01 -6.11950576e-01 5.52763700e-01 -1.99822038e-01
-6.71202838e-01 1.01815104e+00 2.70142049e-01 -1.07971382e+00
-4.01386470e-01 -6.89709961e-01 -6.66160762e-01 7.98779547e-01
-1.65093017e+00 -7.43611574e-01 -2.44254351e-01 4.79604125e-01
4.95650709e-01 -2.76161075e-01 8.09981108e-01 7.36683905e-01
-5.37598252e-01 5.21939278e-01 4.08983268e-02 -4.49681170e-02
5.03975093e-01 -1.27743697e+00 6.80498242e-01 9.43182051e-01
1.85599059e-01 -5.30033223e-02 7.54342794e-01 -2.33486012e-01
-5.10179758e-01 -1.06192124e+00 9.93625641e-01 2.23148093e-01
2.64087826e-01 -2.70165116e-01 -7.20335901e-01 2.48030707e-01
4.74841058e-01 1.52183846e-01 4.10003453e-01 -1.27204478e-01
1.27400100e-01 -1.28829226e-01 -1.71095848e+00 9.04012546e-02
7.63789058e-01 -2.44344305e-02 -3.03176761e-01 -1.01387493e-01
9.04149264e-02 -3.67713660e-01 -6.97580576e-01 3.72138590e-01
5.58102071e-01 -1.76205480e+00 8.40694010e-01 -1.91356823e-01
1.60248399e-01 -3.24656665e-02 7.82002583e-02 -1.14802206e+00
2.28380069e-01 -5.42346895e-01 1.25276104e-01 1.17307031e+00
5.51605165e-01 -9.37450588e-01 7.76427388e-01 -3.53542492e-02
-1.38118446e-01 -8.70399535e-01 -1.19637418e+00 -7.62147188e-01
-7.45607466e-02 -4.95712578e-01 3.99359047e-01 5.45302093e-01
-5.65986693e-01 9.37005803e-02 7.21537247e-02 -3.01353037e-01
2.22923890e-01 -1.87900156e-01 5.37693739e-01 -1.39089310e+00
-2.13061526e-01 -6.08561218e-01 -1.00416672e+00 -3.95812899e-01
-3.43303412e-01 -6.98452890e-01 -2.21488446e-01 -1.52479994e+00
-1.65202573e-01 -2.13397279e-01 -3.25355798e-01 4.15574700e-01
1.37563139e-01 7.09454000e-01 3.11087728e-01 -2.49259800e-01
-3.52045387e-01 -8.80796462e-02 9.03480709e-01 1.11416236e-01
-1.32737905e-01 1.58651665e-01 -4.25367862e-01 7.71927536e-01
9.12351847e-01 -3.83478552e-01 -1.17669187e-01 -5.67860082e-02
-1.80939585e-01 -2.79701054e-01 3.26350391e-01 -1.58802688e+00
-2.34947093e-02 7.08142042e-01 3.74384224e-01 -7.21148014e-01
2.84582507e-02 -9.56380904e-01 3.36806253e-02 9.11967397e-01
6.34941906e-02 1.02607518e-01 5.37809491e-01 1.65666685e-01
-5.12481630e-01 -5.50585747e-01 1.06250834e+00 -2.16941670e-01
-1.09276509e+00 1.35088623e-01 -6.48084342e-01 8.31056386e-02
1.07444715e+00 -4.67130780e-01 -1.77194402e-01 -2.11564973e-01
-9.51519072e-01 8.70284736e-02 1.51481062e-01 3.90634656e-01
3.36516559e-01 -7.46577442e-01 -9.48049068e-01 3.37513238e-01
-4.49352674e-02 -1.95332289e-01 1.38818398e-01 1.12287819e+00
-8.65711987e-01 5.69024324e-01 -3.16408277e-01 -7.01960742e-01
-1.18519640e+00 2.68213958e-01 2.07353339e-01 -6.86100721e-01
-7.00831592e-01 6.95781767e-01 -1.91833913e-01 -1.86955184e-02
3.93344045e-01 -6.09848082e-01 -4.71047848e-01 3.87899935e-01
5.54010868e-01 6.48939371e-01 6.46685958e-01 -6.97140753e-01
-5.07925570e-01 2.61240721e-01 -1.04905032e-01 -2.75113463e-01
1.40872025e+00 1.53542176e-01 2.11073518e-01 2.73166448e-01
1.17183828e+00 -4.19158250e-01 -7.40950704e-01 7.92198628e-02
3.13914418e-01 -2.67425805e-01 4.07901667e-02 -9.18713212e-01
-1.26275849e+00 9.70580459e-01 1.16777635e+00 5.13058424e-01
1.37023270e+00 -3.28334272e-01 8.97493124e-01 -2.71560345e-03
4.91520703e-01 -1.25845647e+00 -5.48738897e-01 5.83014727e-01
4.92259800e-01 -9.69347119e-01 -1.07605025e-01 -3.06382120e-01
-4.58438098e-01 1.19256806e+00 7.69801214e-02 -3.80283952e-01
8.71304512e-01 4.72446740e-01 -9.43411980e-03 -3.05681378e-01
-4.00530994e-01 -7.47994542e-01 7.56956339e-02 6.40195370e-01
6.01832151e-01 -1.69460088e-01 -8.08817804e-01 1.58763692e-01
1.98642425e-02 1.62147671e-01 3.76792312e-01 9.32945430e-01
-3.32741708e-01 -1.23035920e+00 -3.88358980e-01 5.97069860e-01
-1.06371009e+00 2.46223286e-02 -4.23523158e-01 1.42925954e+00
5.38340271e-01 1.04438531e+00 2.79788613e-01 -1.21594362e-01
7.60571957e-01 1.10235624e-01 5.33017755e-01 -5.02670884e-01
-1.18956518e+00 2.04968601e-01 3.34134847e-01 -8.65013301e-01
-8.23824167e-01 -8.88973594e-01 -1.20661879e+00 -2.88692176e-01
-2.67038226e-01 1.04721375e-01 8.79524589e-01 9.77418423e-01
5.58428839e-03 9.05776143e-01 3.03477407e-01 -7.42997408e-01
1.27538636e-01 -1.03993452e+00 -6.42602146e-01 3.93555969e-01
8.28679055e-02 -5.03490865e-01 -3.87400202e-02 -5.66629246e-02]
|
[8.900558471679688, -0.712554395198822]
|
56dce76d-1d0a-46ce-8295-54db71620ffe
|
neural-adaptation-layers-for-cross-domain
|
1810.06368
| null |
http://arxiv.org/abs/1810.06368v1
|
http://arxiv.org/pdf/1810.06368v1.pdf
|
Neural Adaptation Layers for Cross-domain Named Entity Recognition
|
Recent research efforts have shown that neural architectures can be effective
in conventional information extraction tasks such as named entity recognition,
yielding state-of-the-art results on standard newswire datasets. However,
despite significant resources required for training such models, the
performance of a model trained on one domain typically degrades dramatically
when applied to a different domain, yet extracting entities from new emerging
domains such as social media can be of significant interest. In this paper, we
empirically investigate effective methods for conveniently adapting an
existing, well-trained neural NER model for a new domain. Unlike existing
approaches, we propose lightweight yet effective methods for performing domain
adaptation for neural models. Specifically, we introduce adaptation layers on
top of existing neural architectures, where no re-training using the source
domain data is required. We conduct extensive empirical studies and show that
our approach significantly outperforms state-of-the-art methods.
|
['Bill Yuchen Lin', 'Wei Lu']
|
2018-10-15
|
neural-adaptation-layers-for-cross-domain-1
|
https://aclanthology.org/D18-1226
|
https://aclanthology.org/D18-1226.pdf
|
emnlp-2018-10
|
['cross-domain-named-entity-recognition']
|
['natural-language-processing']
|
[-4.72755544e-02 3.36244971e-01 -4.15159464e-01 -6.23284101e-01
-8.54590952e-01 -6.42230809e-01 7.78048575e-01 2.63539493e-01
-1.09455609e+00 9.69778359e-01 3.19030643e-01 -2.55452722e-01
1.50189579e-01 -9.19302464e-01 -8.61935198e-01 1.69837791e-02
3.58965583e-02 6.11782193e-01 3.59471470e-01 -2.88064808e-01
-1.83977693e-01 4.93366003e-01 -9.03187811e-01 2.41906513e-02
8.35282505e-01 7.22973466e-01 -1.95104226e-01 2.16703624e-01
-4.61279213e-01 5.96393049e-01 -7.52701223e-01 -8.14108551e-01
1.58136502e-01 -3.81181464e-02 -1.15347254e+00 -3.93945217e-01
3.41016918e-01 -4.76994157e-01 -5.72511137e-01 9.80252564e-01
4.42523748e-01 3.03877354e-01 6.40462220e-01 -8.96552384e-01
-1.05944908e+00 9.67582405e-01 -1.66420281e-01 3.04551542e-01
-2.69097798e-02 -4.75151271e-01 7.81984746e-01 -8.40287030e-01
1.00106275e+00 6.42126560e-01 9.09061611e-01 7.70424902e-01
-1.04454637e+00 -9.62090015e-01 4.01720554e-01 4.17970680e-02
-1.17913973e+00 -7.20939577e-01 6.26495898e-01 -3.60547043e-02
1.19590402e+00 -4.50944752e-01 1.86536282e-01 1.28760505e+00
-2.60507882e-01 1.03543603e+00 5.45762300e-01 -3.84791046e-01
1.84154451e-01 2.07901537e-01 3.20588827e-01 2.60483265e-01
6.99819386e-01 -2.28903189e-01 -2.00219348e-01 -2.14042917e-01
4.84929502e-01 -1.72843203e-01 4.86911014e-02 -8.86162817e-02
-9.90289748e-01 6.14712119e-01 4.47939157e-01 7.24005401e-01
-6.23203516e-01 -1.56848431e-01 3.82062972e-01 2.07943335e-01
7.85627663e-01 8.72825503e-01 -1.01644278e+00 -2.06165016e-01
-1.20314980e+00 2.38815501e-01 1.15241528e+00 1.14524519e+00
7.00170934e-01 -9.08513367e-02 9.94402245e-02 9.70766842e-01
-1.02344446e-01 1.72193125e-01 4.36951280e-01 -7.23347127e-01
6.42231762e-01 7.56889403e-01 3.02512527e-01 -7.12338865e-01
-6.33816540e-01 -5.01690805e-01 -7.70337224e-01 -4.50647891e-01
4.73536789e-01 -8.35976303e-01 -1.06155121e+00 1.87830985e+00
3.40352923e-01 3.06490332e-01 3.95719230e-01 3.29615146e-01
1.00830007e+00 6.60150707e-01 4.46213245e-01 1.96676865e-01
1.23205459e+00 -9.74300563e-01 -6.20351791e-01 -7.87892759e-01
6.63697481e-01 -3.50395083e-01 4.94726807e-01 -2.25307699e-02
-9.86405015e-01 -2.62690097e-01 -9.27513182e-01 -2.29998901e-01
-9.27187204e-01 1.38622925e-01 6.59594655e-01 4.00213420e-01
-8.14498842e-01 5.30033588e-01 -8.53220940e-01 -7.62183070e-01
4.98498172e-01 3.34148914e-01 -5.75799048e-01 -1.26542598e-01
-1.58501160e+00 1.10629606e+00 8.23587894e-01 -1.57946572e-01
-6.47256196e-01 -7.74365902e-01 -8.50807130e-01 2.40737781e-01
3.39333236e-01 -5.25886536e-01 1.66876912e+00 -7.21896350e-01
-1.35196781e+00 7.75107682e-01 -2.35255465e-01 -6.77672923e-01
9.77880433e-02 -5.15632689e-01 -9.44014370e-01 1.29848197e-01
1.37364194e-01 5.93788207e-01 3.98122579e-01 -1.04951584e+00
-8.34720135e-01 -4.23748605e-02 2.89534092e-01 -5.37673980e-02
-1.07472920e+00 3.55026901e-01 -4.70324844e-01 -6.24957740e-01
-2.87466228e-01 -7.07728148e-01 -4.49384689e-01 -3.43113959e-01
-4.08392340e-01 -3.35885555e-01 5.31242788e-01 -8.10200751e-01
1.49135029e+00 -1.88118732e+00 -3.57125074e-01 3.60930078e-02
2.47614205e-01 6.35402381e-01 -2.28217691e-01 2.84536302e-01
7.57414997e-02 4.72661644e-01 -2.90356487e-01 -2.27817550e-01
5.07845245e-02 3.22705731e-02 -9.79750603e-02 -1.62637495e-02
4.59011823e-01 8.84924769e-01 -8.81796956e-01 -2.98098534e-01
-4.32510287e-01 5.70257843e-01 -4.56934184e-01 2.54348725e-01
-4.72291969e-02 5.28451242e-02 -6.44012630e-01 4.19511884e-01
6.12798870e-01 -6.21438205e-01 4.13932323e-01 -3.42262238e-02
-4.28416970e-04 7.76552200e-01 -9.69328821e-01 1.66264296e+00
-5.19087672e-01 5.74211240e-01 -1.80973727e-02 -1.08655477e+00
8.66042554e-01 5.04278958e-01 2.97841638e-01 -6.41309083e-01
9.93012860e-02 2.67568588e-01 -1.70656845e-01 -2.80854404e-01
7.31748104e-01 1.32771134e-02 -3.54097664e-01 4.89966840e-01
5.65145791e-01 6.06960416e-01 4.92860645e-01 1.48076221e-01
1.35133362e+00 -7.53102526e-02 5.77157378e-01 1.03720516e-01
6.57096207e-02 3.32354397e-01 7.18048513e-01 7.46957719e-01
-2.19117284e-01 1.94217891e-01 1.42360985e-01 -4.35332119e-01
-1.20700717e+00 -7.40890920e-01 -2.68153220e-01 1.46866238e+00
-7.51528963e-02 -2.37790182e-01 -9.26825404e-01 -1.18365586e+00
-7.83982873e-02 8.26420009e-01 -5.57730019e-01 3.92281711e-02
-7.55865395e-01 -8.96316290e-01 1.09300637e+00 8.38374853e-01
8.80444169e-01 -1.05277252e+00 -1.31456062e-01 6.72098696e-01
-2.50212938e-01 -1.48236942e+00 -1.08042464e-01 2.77579606e-01
-8.71122718e-01 -6.95981979e-01 -8.92519236e-01 -9.98349786e-01
7.07106769e-01 3.57991345e-02 1.45621860e+00 -2.03784227e-01
4.18921143e-01 1.79818928e-01 -3.32338333e-01 -4.83913779e-01
-4.72596079e-01 1.08210659e+00 6.84499219e-02 -3.38565886e-01
9.81189132e-01 -5.81268966e-01 -4.83695507e-01 2.09161267e-01
-1.04323947e+00 -3.21920455e-01 7.64463782e-01 7.56873846e-01
2.27551073e-01 -9.97984782e-03 1.02526557e+00 -1.40057015e+00
8.15933168e-01 -9.43526447e-01 -3.28770161e-01 4.43445146e-01
-6.45816863e-01 1.71790689e-01 7.89691925e-01 -5.50943851e-01
-1.48543501e+00 -2.42528357e-02 -3.56505990e-01 1.96860880e-01
-7.01606035e-01 9.06664491e-01 -1.58584490e-01 1.70495138e-01
9.60997581e-01 -1.00111775e-01 -7.30856180e-01 -8.13078642e-01
3.99963796e-01 8.78810406e-01 5.75309813e-01 -6.21424139e-01
8.87143314e-01 2.22079203e-01 -7.09581852e-01 -6.44691527e-01
-1.04284060e+00 -4.63810027e-01 -9.17364419e-01 3.53128403e-01
7.65613556e-01 -1.03267395e+00 -1.69349127e-02 4.03767347e-01
-1.25693452e+00 -2.07782462e-01 -9.77635756e-02 4.38144416e-01
7.38887116e-02 2.05542948e-02 -8.02572668e-01 -4.34891343e-01
-4.10277545e-01 -4.48709726e-01 5.84557295e-01 5.97037733e-01
-2.46597886e-01 -1.37279081e+00 3.72462034e-01 1.59571484e-01
7.24330306e-01 -6.47120625e-02 7.54783690e-01 -1.65505898e+00
-1.16742104e-01 -5.11828005e-01 -2.56996185e-01 2.43020505e-01
1.11191235e-02 -2.24261656e-01 -1.01400232e+00 -8.75893161e-02
-4.57318932e-01 -2.64254153e-01 7.91537285e-01 6.42876923e-02
7.03826904e-01 -4.66834217e-01 -7.54189432e-01 5.08226991e-01
1.22238398e+00 3.11686724e-01 4.08381045e-01 8.89498591e-01
4.52916026e-01 3.20350260e-01 2.57592946e-01 1.28209084e-01
7.72591114e-01 2.61793911e-01 -2.19082713e-01 -4.33021426e-01
1.38990760e-01 -4.73628253e-01 1.85450479e-01 6.21435821e-01
-7.82666430e-02 -4.79362100e-01 -1.21387792e+00 1.09807026e+00
-1.66796601e+00 -8.70008469e-01 5.39544940e-01 1.92611694e+00
1.21244836e+00 2.73550302e-01 1.00344969e-02 -5.75257182e-01
8.15870583e-01 8.84377956e-02 -9.04471159e-01 -1.64866835e-01
-7.95873106e-02 4.86088991e-01 6.63915515e-01 -1.13103986e-01
-1.48697913e+00 1.16126060e+00 7.13480997e+00 4.70358074e-01
-9.38975871e-01 2.58489996e-01 4.74863648e-01 3.67988758e-02
-1.97071090e-01 -5.15313167e-03 -1.26141906e+00 2.46706724e-01
1.44329250e+00 -4.64722246e-01 1.13220237e-01 1.24609327e+00
-3.73987406e-01 4.18486595e-01 -1.21089745e+00 4.75014299e-01
-4.23783250e-02 -1.35530972e+00 -3.33506092e-02 3.63466144e-02
7.20167458e-01 5.36149740e-01 -3.20088953e-01 8.41230690e-01
8.95678699e-01 -7.93003678e-01 2.07767501e-01 2.71948874e-01
5.74773669e-01 -7.03464329e-01 8.87032032e-01 4.92141902e-01
-9.91013408e-01 1.12555981e-01 -4.54578280e-01 1.68225557e-01
3.12708050e-01 6.76468432e-01 -9.20737922e-01 4.82013106e-01
7.13637471e-01 6.13464773e-01 -6.80016756e-01 1.06493592e+00
-4.49123502e-01 8.82904112e-01 -6.55361652e-01 -6.46418706e-02
2.08248287e-01 4.00790036e-01 2.82040477e-01 1.51038754e+00
3.37766886e-01 1.25817180e-01 -1.83164291e-02 6.67234957e-01
-9.19778526e-01 2.04069197e-01 -8.00727308e-01 -4.37194556e-01
7.95682490e-01 1.25243032e+00 -4.79945719e-01 -7.18179405e-01
-8.30442786e-01 8.67620289e-01 8.80946636e-01 6.72789156e-01
-7.27965713e-01 -8.23012769e-01 6.20880842e-01 5.89947961e-02
4.86425877e-01 -2.95222640e-01 -9.68255326e-02 -1.61742091e+00
1.72236541e-04 -6.43471897e-01 5.83816469e-01 -3.88431609e-01
-1.67795253e+00 8.15128803e-01 9.18948948e-02 -9.22768176e-01
-5.48548818e-01 -5.62780797e-01 -4.72407579e-01 5.87567925e-01
-1.88669467e+00 -1.12738204e+00 9.55918729e-02 4.96110469e-01
2.71464378e-01 -3.62699956e-01 1.05246925e+00 6.22677565e-01
-6.81727767e-01 9.08672631e-01 4.85558212e-01 1.01271534e+00
1.04602468e+00 -1.01280856e+00 1.00355673e+00 1.16719139e+00
2.85295576e-01 1.03454626e+00 3.20603400e-01 -7.07807481e-01
-8.54592681e-01 -1.15889430e+00 1.37599409e+00 -4.11977232e-01
8.08663726e-01 -2.73717552e-01 -1.10492957e+00 1.11339629e+00
4.80665684e-01 4.98815142e-02 1.07643878e+00 7.30470479e-01
-6.01224840e-01 8.20211396e-02 -1.16524768e+00 5.63100934e-01
1.11297619e+00 -4.65529114e-01 -1.04381502e+00 1.04487017e-01
7.05540895e-01 -2.46842444e-01 -1.26313210e+00 3.61430377e-01
4.23701793e-01 -4.03562695e-01 8.46413314e-01 -1.23046362e+00
4.57420647e-01 5.71404472e-02 1.48273245e-01 -1.50197721e+00
-4.79035974e-01 -3.13806623e-01 -4.51882303e-01 1.54139531e+00
1.04272223e+00 -7.30492175e-01 7.50105679e-01 1.22179627e+00
7.98443481e-02 -3.16814631e-01 -7.59634614e-01 -7.98919499e-01
5.61590314e-01 -2.71991938e-01 6.43253565e-01 1.44208848e+00
1.44889593e-01 6.73926115e-01 -1.39755636e-01 3.39554250e-01
3.90636861e-01 -1.16792083e-01 6.84766769e-01 -1.46402514e+00
-8.50379318e-02 -2.25483432e-01 -8.83286446e-02 -1.26760817e+00
4.19957489e-01 -8.27874541e-01 -1.93840563e-02 -1.89032829e+00
1.40940592e-01 -4.96037811e-01 -7.09925592e-01 9.89219606e-01
-2.10691333e-01 8.38652253e-02 3.76928523e-02 5.53904176e-02
-7.88989723e-01 3.52732301e-01 6.55136943e-01 -2.44771272e-01
-2.28083789e-01 -1.71711460e-01 -1.11307514e+00 8.59877586e-01
8.37403178e-01 -8.39680374e-01 -1.13444842e-01 -7.98978388e-01
4.01594937e-01 -3.52490664e-01 -1.62620530e-01 -1.01205373e+00
5.23491681e-01 3.13572213e-02 5.82732320e-01 -2.16237381e-01
-8.23267028e-02 -7.72431910e-01 -2.74335325e-01 -2.44277000e-01
-4.66936618e-01 -8.39686692e-02 4.38687354e-01 4.34339970e-01
-4.80791181e-01 -4.31038469e-01 5.31184912e-01 -9.87725109e-02
-9.79885459e-01 4.02103573e-01 -3.90321106e-01 5.50306737e-01
6.95363462e-01 7.31236115e-02 -5.05301893e-01 -2.57499725e-01
-6.80667341e-01 1.99913174e-01 3.24449003e-01 5.33685863e-01
2.45791763e-01 -1.27713025e+00 -6.70470476e-01 -1.32109493e-01
2.08501697e-01 1.37335911e-01 3.74715477e-02 2.48625547e-01
-1.27098531e-01 2.90146291e-01 -1.61329478e-01 8.76948163e-02
-7.48259664e-01 4.93116736e-01 1.82534039e-01 -7.95588851e-01
-4.38169360e-01 6.13600671e-01 1.95162613e-02 -8.56565893e-01
4.94266339e-02 -2.96206146e-01 -4.30644929e-01 1.29820019e-01
6.37009919e-01 1.63596869e-01 2.84940422e-01 -3.35606247e-01
-3.41815859e-01 8.51484537e-02 -6.26680732e-01 1.09409504e-02
1.87842691e+00 3.43872840e-03 3.07020545e-01 1.13286167e-01
1.09188509e+00 -5.80147132e-02 -1.01601219e+00 -6.63620889e-01
3.89280081e-01 2.09112093e-01 5.35892206e-04 -1.07365572e+00
-8.81961107e-01 7.24262893e-01 2.31998265e-01 2.32170507e-01
1.17371321e+00 -3.82685177e-02 1.34026706e+00 1.07690430e+00
4.25683975e-01 -1.46007454e+00 -5.55026293e-01 1.02574098e+00
2.35988215e-01 -1.32632244e+00 -4.67477851e-02 1.31979976e-02
-5.27632117e-01 1.04343832e+00 6.57262444e-01 -1.85151264e-01
1.00915790e+00 5.09421408e-01 8.55562165e-02 -4.51491289e-02
-6.12094879e-01 -2.77713388e-01 2.84852266e-01 6.66042805e-01
6.11753047e-01 -2.23457336e-01 -1.27802625e-01 1.10724044e+00
6.53243333e-04 3.13261330e-01 3.73958558e-01 1.11196792e+00
-3.30304146e-01 -1.30387211e+00 6.50050044e-02 4.09675479e-01
-9.42832470e-01 -4.74433392e-01 -4.08990115e-01 1.02031696e+00
-1.21731818e-01 7.02053070e-01 9.49148182e-03 -2.49221206e-01
6.39944315e-01 5.79364061e-01 -5.89897558e-02 -7.03625381e-01
-7.29172885e-01 -3.28030676e-01 6.10592186e-01 -2.09924042e-01
-7.09158063e-01 -4.87031847e-01 -1.31060076e+00 -1.35678560e-01
-1.42830864e-01 3.30138266e-01 6.32315218e-01 1.05230260e+00
8.14077079e-01 1.87209785e-01 9.01765674e-02 -5.90195537e-01
-2.35560402e-01 -1.11606419e+00 -2.67271847e-01 4.78178978e-01
1.23607688e-01 -5.87893367e-01 -3.38758007e-02 2.13001311e-01]
|
[9.676016807556152, 9.42609691619873]
|
0351e7a8-d3c6-4db1-b0fd-6cdd04a6cd11
|
interactive-matching-network-for-multi-turn
|
1901.01824
| null |
https://arxiv.org/abs/1901.01824v2
|
https://arxiv.org/pdf/1901.01824v2.pdf
|
Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
|
In this paper, we propose an interactive matching network (IMN) for the multi-turn response selection task. First, IMN constructs word representations from three aspects to address the challenge of out-of-vocabulary (OOV) words. Second, an attentive hierarchical recurrent encoder (AHRE), which is capable of encoding sentences hierarchically and generating more descriptive representations by aggregating with an attention mechanism, is designed. Finally, the bidirectional interactions between whole multi-turn contexts and response candidates are calculated to derive the matching information between them. Experiments on four public datasets show that IMN outperforms the baseline models on all metrics, achieving a new state-of-the-art performance and demonstrating compatibility across domains for multi-turn response selection.
|
['Zhen-Hua Ling', 'Jia-Chen Gu', 'Quan Liu']
|
2019-01-07
| null | null | null | null |
['conversational-response-selection']
|
['natural-language-processing']
|
[ 2.72737890e-01 -8.02197456e-02 -5.02923489e-01 -6.31457984e-01
-1.19048512e+00 -2.11456373e-01 5.04513264e-01 5.50180115e-02
-4.92125601e-01 4.90501791e-01 9.28171873e-01 -1.98496222e-01
1.18568957e-01 -6.85942411e-01 -3.62669408e-01 -1.31327301e-01
2.53769726e-01 5.64958811e-01 1.68627307e-01 -8.10228765e-01
4.10109401e-01 7.16200024e-02 -1.49811912e+00 9.49896574e-01
7.74040341e-01 9.53835964e-01 3.90029609e-01 6.79061234e-01
-3.06145549e-01 9.70693469e-01 -4.92714524e-01 -4.28337604e-01
-2.20200419e-01 -4.25151736e-01 -1.15007663e+00 -2.24039108e-01
2.66390890e-01 -2.53979832e-01 -6.74324095e-01 4.31491137e-01
8.15673232e-01 6.03373587e-01 6.89419806e-01 -6.43335938e-01
-1.17966866e+00 1.17192113e+00 -4.51049246e-02 1.91238523e-01
8.07687283e-01 1.30626053e-01 1.77069855e+00 -1.48501492e+00
7.35201776e-01 1.70911169e+00 1.68107510e-01 7.91115284e-01
-1.27256334e+00 -8.10448229e-01 4.92782444e-01 4.44662392e-01
-1.36286247e+00 -4.64098275e-01 5.55859387e-01 -2.79032350e-01
1.62938595e+00 4.48604286e-01 2.85599649e-01 1.63363087e+00
8.37939903e-02 1.25292003e+00 6.21998727e-01 -3.90191555e-01
-3.49279255e-01 7.78538212e-02 6.19243145e-01 1.58218488e-01
-5.54356098e-01 1.73195191e-02 -6.79249227e-01 -1.40710622e-01
2.74374813e-01 -1.20678000e-01 -1.72968313e-01 3.30866575e-01
-9.26024497e-01 1.26424575e+00 5.44872105e-01 2.05360010e-01
-3.80042881e-01 -1.18852537e-02 6.50641263e-01 4.97217774e-01
7.05584049e-01 5.38038373e-01 -1.59029841e-01 1.69885129e-01
-4.33927298e-01 3.16811234e-01 4.22153175e-01 1.14258456e+00
6.11524403e-01 1.50921524e-01 -1.03585589e+00 1.38919628e+00
2.90353566e-01 3.51775259e-01 5.72757185e-01 -2.94880450e-01
6.96358860e-01 8.85263324e-01 -2.10805714e-01 -9.13578451e-01
-3.31400633e-01 -3.90460670e-01 -1.06890190e+00 -7.33090937e-01
-4.33607399e-01 4.72613089e-02 -8.79843235e-01 1.79791665e+00
1.56613097e-01 -3.66115510e-01 8.42657238e-02 1.01390254e+00
1.68465650e+00 1.05263293e+00 4.00699764e-01 -1.23177581e-01
1.47033632e+00 -1.05376649e+00 -9.19224977e-01 -5.01194417e-01
5.67087948e-01 -6.18713856e-01 1.38488102e+00 -6.39623329e-02
-1.11537075e+00 -8.23704362e-01 -7.66737461e-01 -5.56585848e-01
-3.24486375e-01 6.98631033e-02 3.03331733e-01 -8.65092501e-02
-1.04997742e+00 8.66269842e-02 1.20933203e-03 -6.53811097e-02
-1.98587209e-01 3.45585436e-01 -2.78288484e-01 1.85598388e-01
-2.31158757e+00 1.01778150e+00 2.12272197e-01 -1.06962885e-04
-9.06624973e-01 -5.24809361e-01 -1.03147244e+00 2.22262427e-01
2.04017594e-01 -8.12068939e-01 1.24797189e+00 -9.26723957e-01
-1.62646329e+00 7.42427051e-01 -4.44795460e-01 -4.07755256e-01
-1.17331429e-03 -2.97528327e-01 -6.07088864e-01 -8.45511351e-03
1.55347824e-01 9.18089330e-01 7.00910926e-01 -8.47138643e-01
-3.26008856e-01 -1.55050442e-01 1.04786806e-01 4.84056562e-01
-4.10919368e-01 4.49934095e-01 -4.44709480e-01 -8.62992108e-01
-2.41310224e-01 -8.17881405e-01 -2.09133491e-01 -1.06243193e+00
-6.37542665e-01 -1.15183842e+00 3.68145436e-01 -6.95075870e-01
1.97589898e+00 -1.95180154e+00 5.05451798e-01 -5.40105440e-03
8.32515359e-02 1.14229240e-01 -5.68863392e-01 9.64459479e-01
-2.47710258e-01 2.54494399e-01 1.52235836e-01 -4.64456171e-01
6.23552054e-02 -2.61461943e-01 -8.58562887e-01 -9.91141573e-02
2.99386859e-01 1.24415874e+00 -7.11896837e-01 -4.20141220e-01
-1.67531595e-01 3.67293745e-01 -5.71975589e-01 6.31640255e-01
-1.24343611e-01 2.24846840e-01 -4.84875321e-01 3.53303164e-01
1.70007855e-01 -3.26617151e-01 3.12450409e-01 -1.07105430e-02
1.55656442e-01 1.05193210e+00 -7.49484062e-01 1.31808496e+00
-7.54057705e-01 5.43883741e-01 -5.01370251e-01 -6.44714713e-01
1.22169626e+00 2.95548379e-01 1.64193168e-01 -1.21654809e+00
1.59801334e-01 7.49290586e-02 -2.27636635e-01 -4.26553875e-01
1.05076885e+00 2.62138009e-01 -7.33297527e-01 5.88595986e-01
1.71477795e-01 9.90114212e-02 4.20183651e-02 5.79169035e-01
8.56425762e-01 -3.23389441e-01 4.66489494e-01 -2.29022130e-01
6.89553678e-01 -3.62853765e-01 5.26473343e-01 8.53329480e-01
3.03898066e-01 4.71984774e-01 3.93857777e-01 -5.86575985e-01
-7.96782136e-01 -7.15258241e-01 6.47909269e-02 1.78029919e+00
1.57318890e-01 -6.90456748e-01 -4.33568060e-01 -6.91544294e-01
-8.95914361e-02 9.82955337e-01 -7.60323584e-01 -4.19979423e-01
-8.60710561e-01 -2.71934748e-01 4.19430494e-01 5.81733525e-01
1.42000228e-01 -1.56406701e+00 -1.85012564e-01 4.62516218e-01
-8.28717530e-01 -9.01312888e-01 -9.45111632e-01 6.83979765e-02
-3.41170609e-01 -5.37972391e-01 -3.44954729e-01 -8.21902931e-01
1.98620290e-01 4.48097676e-01 1.58010578e+00 1.06478594e-01
-7.47720227e-02 3.92981023e-02 -8.52720082e-01 -1.67950258e-01
-2.21271411e-01 5.74145138e-01 -5.20073064e-02 1.08539909e-01
5.47463477e-01 -1.80407345e-01 -6.60142124e-01 2.72896886e-01
-7.45571136e-01 1.31975845e-01 4.83222157e-01 1.10860085e+00
6.55845523e-01 -9.27661300e-01 1.00373042e+00 -1.00224936e+00
1.16076124e+00 -6.22743428e-01 -2.95225754e-02 5.80529749e-01
-5.14008880e-01 2.40288265e-02 6.10326588e-01 -5.78284860e-01
-9.91859555e-01 -1.99204251e-01 -2.29514271e-01 -6.26841336e-02
2.04661548e-01 3.85557681e-01 -3.41298431e-02 5.28814852e-01
4.76656467e-01 4.48823333e-01 -4.47833061e-01 -2.82312185e-01
6.07459605e-01 1.10723400e+00 -4.08776887e-02 -3.05972040e-01
1.17750973e-01 -1.55715674e-01 -7.42738008e-01 -7.06736743e-01
-1.10342264e+00 -5.55990398e-01 -3.33679587e-01 -2.36268714e-01
9.49729919e-01 -9.23410416e-01 -6.13555789e-01 1.46753579e-01
-1.41358018e+00 -2.33555034e-01 5.70747666e-02 2.50932902e-01
-3.01140547e-01 1.10504530e-01 -8.73057902e-01 -9.18800950e-01
-9.56093252e-01 -1.33256876e+00 1.30826247e+00 6.74585998e-02
-6.42395914e-01 -6.41657650e-01 7.99785629e-02 4.66827601e-01
4.02366728e-01 -4.64045465e-01 1.02636981e+00 -1.03020275e+00
-3.87029290e-01 -7.64331073e-02 -1.06133841e-01 3.03437393e-02
-2.77839243e-01 -2.38795474e-01 -8.28387260e-01 -3.15549999e-01
-2.88989753e-01 -9.08694446e-01 1.30816555e+00 1.90660432e-01
1.45466292e+00 -4.85489398e-01 -2.31902167e-01 2.99059540e-01
9.72644985e-01 3.56372505e-01 7.79964745e-01 1.35776699e-01
5.50929964e-01 9.36453044e-01 7.11010158e-01 5.09117365e-01
7.95951664e-01 1.02305365e+00 2.62055188e-01 -4.79529798e-02
-7.59796128e-02 -4.46629167e-01 3.70546728e-01 1.26564109e+00
5.45643985e-01 -7.78903782e-01 -6.09230340e-01 6.23479545e-01
-1.93563759e+00 -9.65382040e-01 -2.69747794e-01 1.84972501e+00
9.97990429e-01 -1.23384297e-01 3.63974832e-02 -4.38728839e-01
7.43425786e-01 4.81878191e-01 -5.08943081e-01 -9.26376164e-01
-3.79440308e-01 3.38622749e-01 -7.34190419e-02 7.86240339e-01
-1.03194332e+00 1.40343893e+00 6.53624201e+00 1.00319481e+00
-9.35352743e-01 1.12225629e-01 9.35179889e-01 -1.44645069e-02
-8.29756975e-01 -1.05875351e-01 -1.36752474e+00 3.63827258e-01
8.98558140e-01 -1.12857908e-01 2.48908192e-01 6.01712406e-01
3.25067043e-02 4.16161060e-01 -9.52373981e-01 6.63554013e-01
2.96610177e-01 -1.26554215e+00 6.91962719e-01 -3.68807077e-01
5.07585049e-01 -1.37336448e-01 2.24411011e-01 1.00198138e+00
6.12489223e-01 -1.23958623e+00 4.54961985e-01 7.17096746e-01
9.67409253e-01 -9.69473660e-01 6.21343315e-01 2.87805408e-01
-1.33053637e+00 -1.76092803e-01 -3.76729876e-01 2.28844643e-01
2.50735998e-01 2.91148186e-01 -7.22691357e-01 5.54975450e-01
5.30014217e-01 5.45173645e-01 -5.12415707e-01 3.74454945e-01
-3.85057390e-01 4.99736190e-01 2.24189058e-01 -5.31726599e-01
5.00142872e-01 1.57193542e-01 4.91882175e-01 1.61239982e+00
-4.34934646e-02 3.90737414e-01 2.35674158e-01 7.14678526e-01
-3.23423594e-01 5.92987359e-01 -5.03501356e-01 1.39922112e-01
8.64132047e-01 1.12305677e+00 -2.00308021e-02 -4.97253925e-01
-1.59509525e-01 9.06927407e-01 8.98870766e-01 4.25949097e-01
-8.32830548e-01 -5.25567830e-01 7.68044770e-01 -1.18797377e-01
1.26572549e-01 3.26018631e-01 -7.18923360e-02 -1.09152198e+00
-7.15968311e-02 -1.11472356e+00 6.56050920e-01 -6.10512495e-01
-1.56620610e+00 1.12135255e+00 -6.58912659e-02 -1.27518284e+00
-5.66652477e-01 -7.64745772e-02 -4.27514464e-01 1.05147278e+00
-1.27325416e+00 -1.11016119e+00 1.19994201e-01 4.11282480e-01
1.32610273e+00 -1.75673306e-01 9.50348675e-01 2.65590906e-01
-5.89078367e-01 1.10825121e+00 -4.02549446e-01 9.12283063e-02
8.14661145e-01 -8.37150395e-01 6.73449337e-01 6.14877880e-01
9.90983024e-02 8.97965252e-01 2.12826282e-01 -6.82216525e-01
-1.23346245e+00 -1.11903024e+00 1.61474133e+00 -3.68984312e-01
4.59234327e-01 -7.01407790e-01 -1.02093160e+00 5.43945253e-01
7.50151575e-01 -4.55685705e-01 9.71064091e-01 2.10515320e-01
-6.42976522e-01 8.17634538e-02 -2.57758528e-01 7.39709854e-01
1.15625417e+00 -9.19206321e-01 -7.74769783e-01 4.17929620e-01
1.44072425e+00 -4.10074532e-01 -5.95491588e-01 5.70141256e-01
5.22644520e-01 -6.65579021e-01 1.12692451e+00 -1.19126546e+00
7.05577731e-01 3.31121385e-01 -2.84132361e-01 -1.49537754e+00
-8.95284653e-01 -6.71589673e-01 -6.10231608e-02 1.34109366e+00
6.89381421e-01 -3.22125822e-01 8.25185552e-02 2.17882887e-01
-4.19497252e-01 -1.15339887e+00 -7.77579010e-01 -2.42238775e-01
9.41618606e-02 -3.97842377e-01 6.05840743e-01 7.94082046e-01
1.00866809e-01 1.21131670e+00 -8.04585338e-01 -3.19957584e-01
1.08789645e-01 3.12309682e-01 5.37108779e-01 -8.61185789e-01
-4.67093475e-02 -6.23812556e-01 1.30046308e-01 -1.69621277e+00
8.54547381e-01 -1.04774773e+00 -7.84791599e-04 -1.57709825e+00
3.41399014e-01 -1.55593425e-01 -4.85177189e-01 2.58427203e-01
-6.13722026e-01 -9.14069545e-03 1.01685472e-01 1.39750659e-01
-9.55312848e-01 9.64353383e-01 1.17412281e+00 -3.85961592e-01
-2.09537581e-01 -1.86764449e-01 -8.44605386e-01 1.06434561e-01
5.50600350e-01 -5.38729191e-01 -4.63685125e-01 -6.85691953e-01
3.83614033e-01 4.32070076e-01 -1.71796352e-01 -1.58462092e-01
2.84402698e-01 -2.13489205e-01 8.71153399e-02 -9.49234426e-01
3.04591268e-01 -2.20028579e-01 -1.13060012e-01 8.91826749e-02
-1.36487019e+00 3.15509409e-01 -1.75552592e-01 5.64855695e-01
-4.85794663e-01 -7.24751502e-02 5.90706646e-01 -6.12006448e-02
-8.01133394e-01 3.76321197e-01 -6.43687069e-01 4.34809685e-01
5.82475245e-01 8.56857598e-02 -2.45366856e-01 -7.14703977e-01
-5.85750163e-01 7.00271904e-01 -2.95141459e-01 1.14658916e+00
1.18427598e+00 -1.69684446e+00 -1.15698421e+00 2.62462169e-01
6.37396634e-01 -2.80039877e-01 6.84071362e-01 5.13831794e-01
2.14161992e-01 6.13548219e-01 2.78697014e-01 -2.57221252e-01
-1.36938512e+00 4.56842124e-01 1.72127753e-01 -7.72416353e-01
-4.95205134e-01 1.14883626e+00 5.35311162e-01 -6.26918614e-01
2.42530569e-01 6.62013888e-02 -9.47058201e-01 5.62473774e-01
7.61320710e-01 1.03357531e-01 1.77163705e-02 -8.68477285e-01
-4.21142370e-01 2.18758941e-01 -4.46423680e-01 -9.91664478e-04
1.07043338e+00 -4.10274893e-01 -2.36772031e-01 3.61751825e-01
1.41901875e+00 -2.27814421e-01 -5.49541712e-01 -6.63309515e-01
1.07622884e-01 -2.26237401e-01 -2.10887462e-01 -6.36784673e-01
-7.10917413e-01 8.62430155e-01 8.93427581e-02 1.52476385e-01
9.84520376e-01 3.00146407e-03 1.14355803e+00 4.79450971e-01
2.97526903e-02 -1.23028421e+00 4.66722488e-01 1.17758715e+00
1.29447627e+00 -1.11955702e+00 -4.37870234e-01 -3.11525762e-01
-1.27645350e+00 1.09948123e+00 9.75901961e-01 -1.87736556e-01
3.68169904e-01 -1.12931944e-01 -3.00197974e-02 -1.90517023e-01
-1.61383271e+00 -3.46493006e-01 6.61868572e-01 7.56403580e-02
8.74278724e-01 1.97760537e-01 -5.83741128e-01 7.90380001e-01
-5.18452413e-02 -6.75481498e-01 1.72046095e-01 3.39437038e-01
-4.78798270e-01 -1.15971780e+00 2.53140450e-01 3.35682273e-01
-3.51561517e-01 -5.76597989e-01 -1.05087006e+00 3.95135134e-01
-3.48231286e-01 1.19110084e+00 3.91130447e-02 -9.72324848e-01
6.52599514e-01 1.63434550e-01 -2.12335065e-01 -8.69697571e-01
-1.17491961e+00 3.78928259e-02 5.43832779e-01 -7.48255551e-01
2.29157060e-02 -2.87492156e-01 -9.36815560e-01 5.97830825e-02
-2.66277373e-01 2.41332635e-01 1.30324483e-01 6.64949298e-01
6.59722805e-01 7.43488193e-01 1.17173386e+00 -4.27750140e-01
-8.39208424e-01 -1.35280228e+00 -1.21487141e-01 6.45289242e-01
1.69540435e-01 -5.78931808e-01 -1.81985766e-01 -5.97699344e-01]
|
[12.440510749816895, 7.8762078285217285]
|
7fc40c2e-a2b7-4328-b292-f1e0e19734eb
|
handwriting-styles-benchmarks-and-evaluation
|
1809.00862
| null |
http://arxiv.org/abs/1809.00862v1
|
http://arxiv.org/pdf/1809.00862v1.pdf
|
Handwriting styles: benchmarks and evaluation metrics
|
Evaluating the style of handwriting generation is a challenging problem,
since it is not well defined. It is a key component in order to develop in
developing systems with more personalized experiences with humans. In this
paper, we propose baseline benchmarks, in order to set anchors to estimate the
relative quality of different handwriting style methods. This will be done
using deep learning techniques, which have shown remarkable results in
different machine learning tasks, learning classification, regression, and most
relevant to our work, generating temporal sequences. We discuss the challenges
associated with evaluating our methods, which is related to evaluation of
generative models in general. We then propose evaluation metrics, which we find
relevant to this problem, and we discuss how we evaluate the evaluation
metrics. In this study, we use IRON-OFF dataset. To the best of our knowledge,
there is no work done before in generating handwriting (either in terms of
methodology or the performance metrics), our in exploring styles using this
dataset.
|
['Gerard Bailly', 'Damien Pellier', 'Omar Mohammed']
|
2018-09-04
| null | null | null | null |
['handwriting-generation']
|
['computer-vision']
|
[ 3.25124115e-01 -1.22133419e-01 -6.48486987e-02 -3.96320999e-01
-4.70890969e-01 -8.12139094e-01 9.46712077e-01 -2.37838164e-01
-1.98284090e-01 1.07980001e+00 4.11412686e-01 1.91174764e-02
-1.99192435e-01 -7.11478889e-01 -5.34320652e-01 -7.73732662e-01
1.11456640e-01 8.31351995e-01 1.37371466e-01 -4.22552615e-01
6.38182282e-01 7.67755032e-01 -1.62651527e+00 5.78536153e-01
5.43155968e-01 6.45320356e-01 2.45908052e-02 9.73360538e-01
-2.26190872e-02 9.46866810e-01 -1.22355640e+00 -7.00696468e-01
-5.32818809e-02 -8.62503231e-01 -1.01576114e+00 1.40668750e-01
5.25374532e-01 -2.24175051e-01 -5.64622581e-02 6.57089055e-01
7.65840709e-01 7.27592036e-02 8.75671864e-01 -1.11227965e+00
-7.66124964e-01 8.27394009e-01 -1.07423685e-01 1.06446810e-01
4.78566319e-01 2.53340542e-01 7.53998876e-01 -4.79455203e-01
1.04117596e+00 1.19663346e+00 5.51450014e-01 9.88683760e-01
-9.89144683e-01 -3.21159601e-01 -1.12882867e-01 1.76656153e-02
-8.13173950e-01 -2.28670046e-01 9.08882141e-01 -6.04857981e-01
6.44472063e-01 3.06409746e-01 5.41107416e-01 1.90031588e+00
2.68341959e-01 1.01842141e+00 1.58591223e+00 -6.58801615e-01
2.10194767e-01 2.59636730e-01 2.14863762e-01 5.20699263e-01
2.90769711e-03 4.47652079e-02 -6.81486487e-01 2.58111566e-01
8.93978536e-01 -5.08781195e-01 -1.83778256e-01 -7.85581842e-02
-1.30249238e+00 6.03618622e-01 -3.85840610e-02 8.00584078e-01
-1.73203453e-01 3.13068539e-01 4.14345711e-01 4.79855627e-01
3.12796414e-01 7.86217034e-01 -1.74842134e-01 -6.85626447e-01
-1.19242692e+00 7.33004749e-01 1.15780950e+00 1.22754788e+00
1.97745502e-01 6.90540746e-02 -7.06551373e-01 8.78019989e-01
-2.92456765e-02 2.67050028e-01 5.95335424e-01 -8.21695626e-01
4.49330568e-01 3.70955497e-01 1.27751291e-01 -7.31180668e-01
-4.22207028e-01 -1.75389722e-01 -7.89046645e-01 3.64436388e-01
6.45373881e-01 -3.59185427e-01 -8.90749872e-01 1.36722016e+00
-3.64629388e-01 -1.50075987e-01 -6.82888627e-02 8.41910243e-01
8.03540349e-01 5.27183414e-01 -2.18799695e-01 -1.09951794e-01
1.14552903e+00 -9.83487427e-01 -7.62039900e-01 1.31655470e-01
1.79958522e-01 -1.32992029e+00 1.31823516e+00 7.80801415e-01
-1.25007725e+00 -7.65827656e-01 -1.18657923e+00 -4.87837978e-02
-4.26591724e-01 4.18005019e-01 6.43077135e-01 8.42247486e-01
-1.25573635e+00 1.07675779e+00 -8.34063053e-01 -6.07635677e-01
8.96184519e-02 3.58993039e-02 -5.73309883e-02 3.07002306e-01
-1.06791031e+00 1.06745875e+00 3.09040934e-01 -1.87333003e-01
-9.15311098e-01 -3.99244785e-01 -3.61570239e-01 -3.97815704e-01
1.21764861e-01 -6.08108461e-01 1.71818781e+00 -8.67340267e-01
-1.76708686e+00 9.58683074e-01 -5.31899445e-02 -4.99289602e-01
1.10108042e+00 -2.47675881e-01 -4.51022476e-01 -3.96342516e-01
-1.44204766e-01 7.00738072e-01 7.52422273e-01 -1.39161980e+00
-5.32940507e-01 -1.87028378e-01 1.00089975e-01 -1.19143270e-01
-2.77793199e-01 9.44463983e-02 -4.85363722e-01 -1.11565948e+00
-4.55508411e-01 -1.10092652e+00 -1.14827938e-02 -6.00649357e-01
-3.50890726e-01 -4.34414774e-01 6.79649055e-01 -6.52616441e-01
1.46861887e+00 -1.58765674e+00 3.37774158e-01 6.55665994e-02
-9.41029340e-02 3.00747842e-01 -4.85313684e-03 7.56563842e-01
3.97651464e-01 2.28930458e-01 -2.32921407e-01 -5.56034744e-01
1.45921320e-01 2.82092661e-01 -6.07714772e-01 -1.67963669e-01
3.90369505e-01 7.91408002e-01 -9.89690900e-01 -3.95404339e-01
2.63263844e-02 4.05419767e-01 -1.31028414e-01 3.21457863e-01
-4.86326069e-01 4.38047558e-01 -2.52670288e-01 6.28073692e-01
2.68629819e-01 -2.02309750e-02 8.58533010e-02 -1.20115690e-01
-3.28201950e-01 1.25476241e-01 -1.01977754e+00 1.74086440e+00
-5.59051573e-01 9.79385555e-01 -8.15249383e-01 -4.61511642e-01
1.27162409e+00 3.22179556e-01 1.72257930e-01 -5.16457438e-01
1.19210184e-01 4.80506420e-01 -9.28159803e-03 -6.25395238e-01
8.88381124e-01 1.50264531e-01 2.66892612e-02 6.16244137e-01
3.74667495e-01 -2.80471146e-01 1.00104630e+00 -1.43373340e-01
9.00819421e-01 5.79056263e-01 5.38368225e-02 -3.14426422e-01
4.85847980e-01 4.18455116e-02 1.23446584e-01 8.35273683e-01
-8.49889070e-02 9.89062965e-01 8.31554472e-01 -6.93695545e-01
-1.43291593e+00 -8.33608329e-01 8.49179924e-02 7.76757896e-01
-1.84508920e-01 -3.45890254e-01 -1.05505455e+00 -6.36070848e-01
-2.41970256e-01 7.94094086e-01 -8.69158387e-01 2.07767755e-01
-9.10227001e-01 -7.11013317e-01 7.79926598e-01 7.19964802e-01
4.59973544e-01 -1.58475375e+00 -6.83155596e-01 3.69019568e-01
-1.23932892e-02 -9.03336346e-01 -3.64082575e-01 -3.00935954e-02
-9.21782911e-01 -8.91741097e-01 -1.08654642e+00 -6.62988305e-01
2.12356836e-01 -3.62497836e-01 1.47708476e+00 -1.58764925e-02
-1.99480757e-01 3.26272070e-01 -6.28112495e-01 -7.56500840e-01
-9.23276603e-01 5.86349189e-01 -3.26809645e-01 -2.37011731e-01
4.44538713e-01 -4.36675996e-01 -3.90113920e-01 4.45971638e-01
-1.02963316e+00 1.77423835e-01 8.52541387e-01 7.55331993e-01
3.89065653e-01 -3.62963468e-01 2.87674189e-01 -1.04176629e+00
1.23062599e+00 -9.66914147e-02 -4.52915490e-01 4.26516175e-01
-5.54606915e-01 3.37969363e-01 5.31652510e-01 -3.77211988e-01
-9.77902472e-01 -1.81629643e-01 -2.21875295e-01 -1.52069494e-01
-4.24454927e-01 1.95645601e-01 1.73005342e-01 6.25465810e-02
1.02652073e+00 1.85579941e-01 -1.44312292e-01 -6.89568162e-01
2.57182598e-01 6.84009016e-01 3.87092173e-01 -8.36094141e-01
4.29160714e-01 2.44874239e-01 1.47432894e-01 -8.39784920e-01
-6.11680150e-01 -1.59579739e-01 -9.10029173e-01 -4.25143033e-01
7.78831482e-01 -1.59913436e-01 -3.48580569e-01 6.81077003e-01
-1.40069330e+00 -5.47102690e-01 -1.68427065e-01 1.51052296e-01
-1.05736351e+00 1.68102130e-01 -7.96824336e-01 -7.75759935e-01
-3.30913156e-01 -1.34786212e+00 1.16015601e+00 2.59679437e-01
-6.65688813e-01 -1.15989423e+00 5.70206702e-01 2.56852597e-01
5.48607171e-01 6.15428984e-01 5.71439683e-01 -3.34213734e-01
-4.73070949e-01 1.82300955e-01 1.22222178e-01 4.86685395e-01
2.24836171e-01 4.84849453e-01 -1.11346090e+00 -1.30295202e-01
-3.65602612e-01 -3.43519926e-01 8.61432612e-01 1.95320487e-01
1.20688164e+00 -1.45215109e-01 -1.75982006e-02 3.71643990e-01
1.22680342e+00 4.48112398e-01 9.74476457e-01 5.18023014e-01
5.09376109e-01 6.76539540e-01 7.85101891e-01 5.36695957e-01
-3.27450968e-02 7.38173902e-01 -8.73665884e-02 8.17458183e-02
-6.27784312e-01 -2.18762904e-01 4.44002658e-01 6.63293660e-01
-9.36822057e-01 -5.36143303e-01 -1.04514730e+00 3.69940162e-01
-1.82198095e+00 -1.17903459e+00 -9.12832320e-02 1.91346645e+00
1.03241074e+00 1.90123603e-01 5.10648668e-01 3.55877846e-01
5.21813989e-01 8.40144828e-02 -1.48764029e-01 -8.28454614e-01
-1.45935237e-01 4.72502798e-01 2.06694797e-01 3.41120183e-01
-1.15301192e+00 1.07433915e+00 6.59179068e+00 6.19631410e-01
-1.35836530e+00 -2.32911304e-01 6.28878117e-01 3.75945158e-02
-1.66377649e-01 -2.52433509e-01 -1.04625809e+00 5.66779137e-01
7.82281280e-01 -5.34635484e-02 6.77531242e-01 8.70058477e-01
2.79895753e-01 3.55363458e-01 -1.52017176e+00 8.56853068e-01
4.69905436e-01 -1.35729921e+00 2.39096746e-01 -1.66916713e-01
1.04535520e+00 -3.32769901e-01 7.20717460e-02 1.88811854e-01
2.68847227e-01 -1.06922543e+00 9.96294916e-01 9.46987569e-01
4.53748405e-01 -7.61240482e-01 8.54055524e-01 -6.62409142e-03
-7.15492547e-01 3.12258959e-01 -2.54941374e-01 1.32349506e-01
1.40372127e-01 4.44444954e-01 -1.03052580e+00 5.79708040e-01
7.53695667e-01 6.29111707e-01 -7.62239933e-01 1.01443183e+00
-4.61215705e-01 6.11094117e-01 3.34637493e-01 -6.79306328e-01
1.81561202e-01 -4.25676405e-02 4.73661631e-01 1.64305520e+00
5.35347402e-01 -3.36272925e-01 -6.41064122e-02 1.10643363e+00
-2.63277665e-02 1.23727955e-01 -6.52905464e-01 -3.57689381e-01
1.21800140e-01 1.15794861e+00 -7.84778535e-01 -2.72544950e-01
-9.81262326e-02 1.09705341e+00 7.22389519e-02 2.06737086e-01
-7.64386415e-01 -5.94422877e-01 4.20480520e-01 1.46815673e-01
-2.10384414e-01 -3.85473609e-01 -4.75879192e-01 -9.70223010e-01
1.20327212e-01 -1.16146970e+00 1.82795897e-01 -7.41821110e-01
-1.49589372e+00 8.25236738e-01 1.86639547e-01 -1.22246826e+00
-7.39260375e-01 -9.42551196e-01 -6.64798200e-01 8.67998540e-01
-1.04946554e+00 -1.20978105e+00 -6.25671685e-01 2.76283026e-01
8.03608477e-01 -4.64669824e-01 6.83819532e-01 1.10984102e-01
-2.83417106e-01 6.41433537e-01 -1.25473425e-01 7.49013573e-02
9.72557366e-01 -1.56117105e+00 6.90538526e-01 7.08540618e-01
2.61323184e-01 6.87138319e-01 8.70246589e-01 -5.34703672e-01
-1.13236630e+00 -8.83819520e-01 9.25220072e-01 -7.81192243e-01
5.76625586e-01 -2.98308164e-01 -7.29838133e-01 5.10611475e-01
5.68951666e-01 -6.79539263e-01 5.93842983e-01 4.43208367e-02
3.89901735e-02 1.16884649e-01 -9.13642943e-01 7.12312341e-01
1.22685158e+00 -1.60588846e-01 -5.64161360e-01 3.01915407e-01
4.29514050e-01 -6.08229876e-01 -1.05824399e+00 4.42706555e-01
7.96787024e-01 -1.20009351e+00 6.08727872e-01 -7.08254814e-01
9.43709373e-01 -2.10667744e-01 1.54091299e-01 -1.57009876e+00
-2.13957042e-01 -7.51455605e-01 -1.56720787e-01 1.64397800e+00
5.72131515e-01 -2.22960025e-01 9.96388316e-01 2.38100812e-01
-2.28337348e-01 -8.94014299e-01 -2.64114529e-01 -1.02206838e+00
2.91374922e-01 -1.98784426e-01 8.22828114e-01 8.27652693e-01
-3.66190612e-01 1.57659605e-01 -6.33866966e-01 -4.08784270e-01
3.54691148e-01 2.54115820e-01 1.05830002e+00 -1.01661670e+00
-4.72995162e-01 -9.68450367e-01 -4.00169760e-01 -5.63452482e-01
-1.48075759e-01 -5.88597119e-01 -3.08220312e-02 -1.54332006e+00
5.34140319e-03 -2.65564114e-01 -4.91462573e-02 2.91141331e-01
-1.19487671e-02 2.10061282e-01 3.24298829e-01 1.08061865e-01
-3.32453519e-01 1.67754307e-01 1.53590548e+00 -3.24032128e-01
-2.54165441e-01 1.22272074e-01 -3.91251534e-01 4.56950277e-01
1.34677029e+00 -7.24008083e-02 -4.64994639e-01 -5.39448500e-01
1.97592616e-01 -3.07010233e-01 1.73716918e-01 -1.25328863e+00
-3.51175368e-02 -2.59496897e-01 5.65160275e-01 -6.10565782e-01
5.49882874e-02 -2.78685838e-01 1.55608639e-01 3.80921513e-01
-5.27849555e-01 3.84253681e-01 5.99340610e-02 -1.37985930e-01
-3.45853388e-01 -5.57337523e-01 7.76340425e-01 -2.90466309e-01
-8.85680377e-01 5.53822890e-02 -1.39858887e-01 -1.47347134e-02
9.42153215e-01 -4.67698239e-02 -2.26063132e-01 -4.04118448e-01
-7.01756060e-01 -9.56776738e-02 5.75010180e-01 8.50094378e-01
3.91250461e-01 -1.39745283e+00 -8.49061549e-01 -1.13361068e-01
1.82805240e-01 -5.12635767e-01 -8.41888413e-02 3.80963862e-01
-1.07961321e+00 5.36543667e-01 -7.34120965e-01 -5.52683771e-01
-1.30688131e+00 4.47947294e-01 1.16617359e-01 -4.86772448e-01
-7.08128437e-02 5.89915693e-01 -5.53997576e-01 -2.15164438e-01
3.95154089e-01 -3.82583410e-01 -4.98516858e-01 1.46307513e-01
6.11621559e-01 7.49880016e-01 4.77855921e-01 -1.97090104e-01
-7.45700323e-04 6.19158983e-01 -9.11595300e-02 -3.33652526e-01
1.16621447e+00 4.51636612e-01 -1.93505332e-01 6.42385960e-01
8.68764758e-01 -4.64421548e-02 -1.04861128e+00 2.49925420e-01
3.74462664e-01 -4.81671691e-01 -5.21242082e-01 -1.00170505e+00
-8.32258344e-01 9.50829625e-01 7.49164641e-01 3.81613433e-01
1.06847620e+00 -2.14638308e-01 6.30069435e-01 4.44881290e-01
5.93586028e-01 -1.44278979e+00 3.73281211e-01 6.74252391e-01
1.46889758e+00 -1.12009931e+00 -2.75430977e-01 -5.24103530e-02
-7.46568978e-01 1.57589471e+00 6.99344099e-01 -1.35161638e-01
2.04910040e-01 4.12333637e-01 2.57821441e-01 3.76063399e-02
-7.35333502e-01 -1.03250563e-01 4.94864136e-01 6.15539610e-01
1.12915516e+00 8.32854658e-02 -7.97459662e-01 2.68397987e-01
-7.40514338e-01 2.38080204e-01 5.98753929e-01 9.52149034e-01
-2.23932192e-01 -1.83254051e+00 -3.36873889e-01 4.47690815e-01
-4.25202399e-01 2.54680514e-01 -1.01735020e+00 8.41645718e-01
2.61339933e-01 7.92733550e-01 -1.47053733e-01 -5.63039362e-01
5.89599371e-01 4.58126962e-02 9.63162184e-01 -3.69684607e-01
-6.78926528e-01 -4.09414172e-01 2.86732614e-01 -2.97907501e-01
-4.59893644e-01 -7.76933074e-01 -6.08559370e-01 -5.16248882e-01
2.41058201e-01 -5.62835559e-02 8.18706095e-01 5.59045255e-01
-3.27090621e-02 6.37518227e-01 4.43643153e-01 -1.06197298e+00
-5.93189716e-01 -1.13566875e+00 -3.33383590e-01 7.65831470e-01
-1.13587797e-01 -5.52382112e-01 2.16614958e-02 3.08056116e-01]
|
[11.800114631652832, 2.2799978256225586]
|
eeefbac7-5a09-4856-a3a8-a3a36fbdafab
|
on-guiding-video-object-segmentation
|
1904.11256
| null |
http://arxiv.org/abs/1904.11256v1
|
http://arxiv.org/pdf/1904.11256v1.pdf
|
On guiding video object segmentation
|
This paper presents a novel approach for segmenting moving objects in
unconstrained environments using guided convolutional neural networks. This
guiding process relies on foreground masks from independent algorithms (i.e.
state-of-the-art algorithms) to implement an attention mechanism that
incorporates the spatial location of foreground and background to compute their
separated representations. Our approach initially extracts two kinds of
features for each frame using colour and optical flow information. Such
features are combined following a multiplicative scheme to benefit from their
complementarity. These unified colour and motion features are later processed
to obtain the separated foreground and background representations. Then, both
independent representations are concatenated and decoded to perform foreground
segmentation. Experiments conducted on the challenging DAVIS 2016 dataset
demonstrate that our guided representations not only outperform non-guided, but
also recent and top-performing video object segmentation algorithms.
|
["Noel E. O'Connor", 'José M. Martínez', 'Juan C. SanMiguel', 'Kevin McGuinness', 'Eric Arazo', 'Diego Ortego']
|
2019-04-25
| null | null | null | null |
['foreground-segmentation']
|
['computer-vision']
|
[ 9.03805375e-01 -2.98570007e-01 -1.57014355e-01 -4.53225702e-01
-4.58752245e-01 -7.46729076e-01 7.57834971e-01 -3.60492855e-01
-5.96262574e-01 6.09158099e-01 -9.24878847e-03 -2.23081842e-01
7.58844912e-02 -6.53537154e-01 -6.70480788e-01 -9.32973027e-01
-1.90876096e-01 6.58469945e-02 7.08136678e-01 1.96539849e-01
4.39737201e-01 6.65615439e-01 -1.66912067e+00 5.16727090e-01
9.43930745e-01 1.05182803e+00 3.21800858e-01 1.09759295e+00
-4.72873062e-01 1.13409114e+00 -5.27747095e-01 -7.74779618e-02
3.10071766e-01 -6.34313345e-01 -1.04407346e+00 4.99610394e-01
4.73740041e-01 -6.59982681e-01 -3.59428465e-01 7.91200876e-01
1.73208937e-01 4.37368214e-01 6.28386438e-01 -1.07105458e+00
-5.63507199e-01 4.57699060e-01 -7.09407389e-01 8.72194648e-01
4.85685095e-02 3.74459118e-01 6.56719446e-01 -7.31366456e-01
7.66638756e-01 1.17192304e+00 1.49638027e-01 4.89577413e-01
-1.08799899e+00 -4.49343175e-01 6.77341461e-01 1.02756888e-01
-8.87564600e-01 -4.67569649e-01 8.45431685e-01 -4.95624959e-01
7.93677986e-01 8.88906941e-02 7.80375838e-01 8.79268110e-01
-2.13146016e-01 1.13661742e+00 7.47095108e-01 -2.72256941e-01
1.67201012e-01 -2.79928625e-01 3.59548658e-01 8.17350626e-01
3.87685478e-01 7.58284982e-03 -4.87821966e-01 2.26393759e-01
8.69307339e-01 2.01773923e-02 -4.16948020e-01 -4.26966071e-01
-1.33719623e+00 5.86342335e-01 5.19994497e-01 3.26310486e-01
-5.34355819e-01 2.87436485e-01 1.78792432e-01 -3.41482371e-01
4.21292812e-01 -8.81095603e-02 -4.54921782e-01 -1.25555679e-01
-1.43968356e+00 1.13623321e-01 6.96161032e-01 8.07019532e-01
1.00393271e+00 1.13751426e-01 -5.23479342e-01 4.64331210e-01
4.99883384e-01 3.30351472e-01 1.64916098e-01 -1.16897380e+00
4.49716926e-01 7.17392325e-01 2.39616185e-01 -9.26877797e-01
-3.53379071e-01 -1.85222745e-01 -6.78044856e-01 2.94799238e-01
6.47037745e-01 -2.15886548e-01 -1.53639007e+00 1.35587132e+00
4.44798321e-01 6.84862494e-01 1.69988703e-02 1.21083343e+00
1.09801817e+00 8.66824329e-01 3.92008781e-01 3.92101966e-02
1.09110212e+00 -1.35961854e+00 -5.68953812e-01 -3.89788598e-01
3.70873719e-01 -6.75795257e-01 3.30831081e-01 2.36599103e-01
-1.26184368e+00 -8.59853864e-01 -1.09652996e+00 -1.49144411e-01
-3.97094339e-01 2.90005058e-01 7.28437901e-01 7.91134417e-01
-9.70570624e-01 7.97005355e-01 -1.08708346e+00 -1.18124783e-01
8.69665444e-01 4.35580730e-01 -5.85548915e-02 6.38588984e-03
-7.69843936e-01 3.99553269e-01 6.10197127e-01 5.06353498e-01
-1.12656760e+00 -3.56042892e-01 -8.73108745e-01 -1.63079739e-01
4.11598533e-01 -6.68755114e-01 1.08086741e+00 -1.43816268e+00
-1.69979072e+00 8.55571628e-01 -3.78534764e-01 -5.42419434e-01
4.42949742e-01 -5.36128461e-01 3.81478854e-02 7.59367883e-01
-1.09784342e-02 9.12954152e-01 1.14398837e+00 -1.37050879e+00
-1.07058883e+00 9.36707854e-02 7.37391934e-02 1.78002492e-01
5.35937175e-02 1.89119190e-01 -9.94980872e-01 -6.06107056e-01
1.74380839e-01 -6.11593664e-01 -3.99665773e-01 1.06424920e-01
-4.31749135e-01 -5.64531016e-04 1.21298242e+00 -8.49738300e-01
1.05796385e+00 -2.02797031e+00 3.69583964e-01 3.37437689e-02
3.11359674e-01 6.07171237e-01 -2.55895972e-01 -1.86553374e-01
1.42822474e-01 7.20611140e-02 -5.60444951e-01 -3.84233534e-01
-2.90272832e-01 2.72794217e-01 1.61576383e-02 6.01269305e-01
7.84864604e-01 9.75132585e-01 -1.03244936e+00 -6.95840418e-01
6.60079122e-01 6.62037373e-01 -4.04197603e-01 2.45829925e-01
-2.65603304e-01 6.85710013e-01 -4.69389826e-01 7.33003378e-01
8.78660142e-01 -7.21064955e-02 -9.10103135e-03 -5.25024813e-03
-1.54989153e-01 2.57652432e-01 -1.23874009e+00 1.54842889e+00
1.10659981e-02 7.45437145e-01 2.18580037e-01 -1.17768013e+00
7.58090913e-01 2.50196215e-02 4.77525681e-01 -1.86139867e-01
5.03391325e-01 -2.80265547e-02 7.98118263e-02 -4.70789492e-01
5.21115184e-01 4.06439036e-01 2.58682728e-01 3.52222711e-01
2.13817641e-01 1.62414923e-01 5.33480406e-01 2.35380203e-01
7.68742859e-01 8.72444928e-01 -2.27272622e-02 -1.74287751e-01
9.70877945e-01 5.11770733e-02 7.53370941e-01 7.70814896e-01
-6.80406809e-01 9.62124646e-01 3.15267682e-01 -5.14209688e-01
-6.05966628e-01 -8.72999191e-01 2.65880048e-01 1.27297640e+00
5.04293084e-01 1.23315416e-02 -1.00173783e+00 -9.00338709e-01
-1.33900985e-01 5.22557199e-01 -8.29411983e-01 9.93211046e-02
-9.31968212e-01 -7.02676356e-01 2.51753628e-01 9.48428094e-01
7.00989962e-01 -1.28750432e+00 -1.17551517e+00 2.59970218e-01
-2.15327173e-01 -1.32782340e+00 -2.58610606e-01 2.54065573e-01
-8.94141614e-01 -1.16290760e+00 -1.02939570e+00 -8.00208390e-01
7.54620492e-01 6.69063091e-01 9.12205279e-01 3.47489804e-01
-3.18365216e-01 3.31106484e-01 -4.05480474e-01 -1.52754381e-01
-2.23169640e-01 -1.65336914e-02 -4.64485347e-01 6.01972342e-01
2.71734357e-01 -1.53188616e-01 -8.74940634e-01 1.10986993e-01
-9.35522199e-01 1.59182966e-01 6.79563284e-01 4.14204627e-01
5.85366368e-01 -2.43606642e-01 2.29736678e-02 -5.90932786e-01
1.72013436e-02 -3.61271232e-01 -6.98070467e-01 1.86860189e-01
3.16778719e-01 -4.25846241e-02 3.22460234e-01 -5.03471494e-01
-1.39590478e+00 2.86486804e-01 2.63344079e-01 -3.95810544e-01
-6.50384784e-01 9.08165611e-03 -2.00655222e-01 -8.84572044e-02
1.95761278e-01 2.96147615e-01 -3.53819251e-01 -3.85820895e-01
7.40881979e-01 4.81861740e-01 7.78469503e-01 -5.06553531e-01
8.29663217e-01 8.21412802e-01 -1.87414303e-01 -8.58810067e-01
-7.43025482e-01 -7.32329130e-01 -1.30949926e+00 -3.77248704e-01
1.48909509e+00 -6.83442831e-01 -2.39271089e-01 7.57511258e-01
-1.34430611e+00 -4.81483281e-01 -1.54638529e-01 3.83180737e-01
-5.30794859e-01 4.49222088e-01 -6.27020240e-01 -1.17540264e+00
-2.41014481e-01 -1.20707881e+00 1.22925234e+00 8.25108945e-01
1.06944636e-01 -8.65419149e-01 -2.65641093e-01 3.49220663e-01
2.65330076e-01 3.96181077e-01 2.81997412e-01 -5.21877050e-01
-1.12573659e+00 8.77079517e-02 -6.21762872e-01 2.55751878e-01
1.45299062e-01 5.98148882e-01 -9.73974228e-01 7.77025521e-03
-4.23286527e-01 7.54664615e-02 1.45773530e+00 6.88475788e-01
1.02896249e+00 1.48686185e-01 -4.57557857e-01 1.00527835e+00
1.26150298e+00 4.49188501e-01 6.82412088e-01 4.89610732e-01
9.82843876e-01 5.27630627e-01 5.91866612e-01 2.41256297e-01
6.77650943e-02 2.07971126e-01 4.59702760e-01 -3.38924319e-01
-2.33067930e-01 3.68469000e-01 2.80875742e-01 2.69437402e-01
-5.34001410e-01 -1.56817332e-01 -8.51065278e-01 7.59692967e-01
-1.89393330e+00 -1.10715175e+00 -3.40147972e-01 1.86492467e+00
4.26330149e-01 1.72245368e-01 2.48159766e-01 8.23673531e-02
9.09379840e-01 4.49336201e-01 -5.10795951e-01 -3.24633986e-01
-2.81022310e-01 4.26861465e-01 4.91424650e-01 1.06834657e-01
-1.74117875e+00 1.13500905e+00 5.90869284e+00 3.98120552e-01
-9.07169461e-01 -1.13984063e-01 8.43494475e-01 1.04644038e-02
1.85763270e-01 3.97837199e-02 -7.84408689e-01 3.03475052e-01
7.83311605e-01 1.63270995e-01 3.83155704e-01 6.19852901e-01
1.03755835e-02 -3.87187719e-01 -8.55187714e-01 6.44115865e-01
-8.33730251e-02 -1.34558296e+00 4.66751168e-03 -1.81938007e-01
9.88888562e-01 -7.27751777e-02 -2.39342581e-02 1.11321211e-01
2.54218936e-01 -8.85370910e-01 9.59189534e-01 6.82019711e-01
1.88875750e-01 -7.81835675e-01 7.42233813e-01 1.40766367e-01
-1.53444517e+00 -2.02844873e-01 -2.62789130e-01 -9.21072289e-02
3.76072407e-01 3.32380354e-01 -1.99567035e-01 7.89900184e-01
7.87183523e-01 9.19891238e-01 -6.32479489e-01 1.25115275e+00
-2.25281805e-01 7.28132486e-01 -2.33755395e-01 1.67128652e-01
7.84350991e-01 -3.54137897e-01 3.32822531e-01 1.57676280e+00
4.86194296e-03 1.40200645e-01 1.56799540e-01 1.02790177e+00
1.63672715e-01 -1.25981182e-01 -2.67765552e-01 -2.52257973e-01
-3.49012315e-02 1.55746794e+00 -1.48045278e+00 -8.48064005e-01
-6.16773844e-01 1.27951467e+00 1.55117884e-01 7.76565015e-01
-1.08218122e+00 -3.39758068e-01 7.17764437e-01 -3.59922111e-01
1.02533293e+00 -4.75081831e-01 -1.95901841e-01 -9.87330139e-01
-3.25211257e-01 -5.14125049e-01 3.32579225e-01 -4.90973681e-01
-9.17239368e-01 5.29633701e-01 8.91476348e-02 -1.04219687e+00
4.43682037e-02 -7.39847362e-01 -7.70351470e-01 8.73321235e-01
-1.70399082e+00 -1.03532028e+00 -5.02205670e-01 5.08048236e-01
7.35598445e-01 1.39593005e-01 1.72018662e-01 6.95491508e-02
-9.97836828e-01 -1.21172322e-02 -8.51087943e-02 6.34504557e-01
3.18550527e-01 -1.19555533e+00 6.92068756e-01 1.41997111e+00
2.17391774e-01 5.37210464e-01 1.44382119e-01 -7.44938314e-01
-1.06327784e+00 -1.36341310e+00 3.74423563e-01 -3.73356581e-01
3.92095298e-01 -1.68891594e-01 -9.35739875e-01 5.38960516e-01
4.53438550e-01 3.89720738e-01 5.06673515e-01 -5.57358205e-01
1.52824717e-02 2.57196397e-01 -8.55157912e-01 4.94175851e-01
1.22241044e+00 -2.04729617e-01 -5.83087504e-01 -5.63484877e-02
5.89125812e-01 -4.05006409e-01 -1.71807393e-01 3.58045191e-01
4.02458847e-01 -1.19393039e+00 1.18034697e+00 -5.91593504e-01
3.99719477e-01 -6.11425817e-01 -1.14298977e-01 -6.96766198e-01
-3.60061526e-01 -7.60704219e-01 -4.31467831e-01 1.21923375e+00
4.12655734e-02 -3.95215631e-01 8.03415358e-01 4.33782130e-01
5.06870374e-02 -6.14476919e-01 -5.56869090e-01 -4.92865473e-01
-1.47561550e-01 -3.94341588e-01 3.55990142e-01 4.94853646e-01
-7.75599897e-01 -6.21400215e-02 -5.26304655e-02 2.35046029e-01
6.92424417e-01 3.00685525e-01 7.27939367e-01 -1.22552359e+00
5.99026605e-02 -7.47472703e-01 -5.09176552e-01 -1.25844467e+00
1.66983858e-01 -7.42915750e-01 2.89319813e-01 -1.78516042e+00
1.50865510e-01 -4.87929992e-02 -5.56085825e-01 2.88839966e-01
-7.29634106e-01 5.97106755e-01 4.84480500e-01 1.82369258e-02
-8.59165132e-01 2.92584181e-01 1.24185336e+00 -1.61944568e-01
-4.61320698e-01 -9.31653380e-02 -5.60752690e-01 9.52172458e-01
7.62760162e-01 -2.47014597e-01 -1.07382230e-01 -3.25387388e-01
-6.91624939e-01 -4.25119281e-01 6.98214233e-01 -1.24183524e+00
-2.76668463e-02 -1.70127302e-01 8.79380703e-01 -9.27888274e-01
1.54046878e-01 -5.71266055e-01 -3.19136858e-01 4.40954804e-01
-1.54899880e-01 -3.69612813e-01 3.06527138e-01 6.45995259e-01
-5.16803004e-02 -1.68427929e-01 6.95599794e-01 -2.04562336e-01
-1.19497526e+00 2.95567542e-01 -6.90142572e-01 -1.50318965e-02
1.22493136e+00 -6.43139482e-01 -1.17682256e-01 -2.40472674e-01
-7.80137956e-01 1.35834113e-01 1.77501291e-01 3.91490757e-01
6.27554059e-01 -1.00012887e+00 -5.55746078e-01 2.53161669e-01
-4.27078515e-01 1.74250335e-01 1.73850879e-01 9.59127545e-01
-8.43717992e-01 4.04047370e-01 -3.50863874e-01 -9.90260720e-01
-1.18613780e+00 6.04264319e-01 3.82449597e-01 -8.05022493e-02
-6.51014984e-01 1.15115857e+00 4.15705174e-01 2.14910135e-02
3.54956955e-01 -6.91599429e-01 -2.87682086e-01 2.41697907e-01
7.34727919e-01 4.25320655e-01 -3.33995908e-01 -1.04359627e+00
-5.92915833e-01 8.76386285e-01 4.98707630e-02 -1.36349171e-01
1.22659242e+00 -2.56020010e-01 6.60591722e-02 2.20511422e-01
1.09725010e+00 -3.17557007e-01 -1.84160888e+00 -3.59416381e-02
8.44961628e-02 -6.35249734e-01 5.94661385e-02 -5.07633328e-01
-1.50780272e+00 1.12845874e+00 4.46373492e-01 1.05763339e-01
1.38659871e+00 -1.87153503e-01 7.52869308e-01 1.74611174e-02
1.29814232e-02 -8.84187281e-01 1.23460544e-02 4.17582512e-01
4.10276681e-01 -1.07189131e+00 -2.25261878e-02 -5.37111759e-01
-5.88181973e-01 1.40743375e+00 5.48833311e-01 -2.13420019e-01
3.14663082e-01 2.88658589e-01 2.60572314e-01 1.47331193e-01
-3.54525149e-01 -9.42602813e-01 6.25560641e-01 8.98668826e-01
5.06916225e-01 -4.07636017e-01 -1.10354997e-01 4.73264247e-01
4.22592252e-01 8.58636852e-03 1.84369177e-01 1.21861315e+00
-5.65016031e-01 -7.64994442e-01 -5.85336030e-01 2.16192111e-01
-4.94730920e-01 1.54801682e-01 -5.81489325e-01 8.23533118e-01
4.58695859e-01 1.13270974e+00 2.20093414e-01 -1.55133277e-01
-7.07102567e-03 1.69062644e-01 4.91295427e-01 -4.03834581e-01
-6.33128881e-01 3.57557565e-01 -5.46154566e-02 -9.27383244e-01
-1.08478308e+00 -7.41000593e-01 -1.54884601e+00 1.96062088e-01
-3.26430678e-01 -3.23036224e-01 5.34807682e-01 9.65941250e-01
1.39679179e-01 1.02091587e+00 2.37811238e-01 -1.86965489e+00
2.16015965e-01 -5.28810382e-01 -2.16838270e-01 5.57032228e-01
5.90108037e-01 -7.27118134e-01 -3.44800264e-01 4.27891672e-01]
|
[9.163252830505371, -0.24157829582691193]
|
e768bab4-e5c1-4e14-bf3f-9bbe4453415d
|
mask-textspotter-an-end-to-end-trainable
|
1807.02242
| null |
http://arxiv.org/abs/1807.02242v2
|
http://arxiv.org/pdf/1807.02242v2.pdf
|
Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes
|
Recently, models based on deep neural networks have dominated the fields of
scene text detection and recognition. In this paper, we investigate the problem
of scene text spotting, which aims at simultaneous text detection and
recognition in natural images. An end-to-end trainable neural network model for
scene text spotting is proposed. The proposed model, named as Mask TextSpotter,
is inspired by the newly published work Mask R-CNN. Different from previous
methods that also accomplish text spotting with end-to-end trainable deep
neural networks, Mask TextSpotter takes advantage of simple and smooth
end-to-end learning procedure, in which precise text detection and recognition
are acquired via semantic segmentation. Moreover, it is superior to previous
methods in handling text instances of irregular shapes, for example, curved
text. Experiments on ICDAR2013, ICDAR2015 and Total-Text demonstrate that the
proposed method achieves state-of-the-art results in both scene text detection
and end-to-end text recognition tasks.
|
['Wenhao Wu', 'Pengyuan Lyu', 'Minghui Liao', 'Cong Yao', 'Xiang Bai']
|
2018-07-06
| null | null | null |
eccv-2018-9
|
['text-spotting']
|
['computer-vision']
|
[ 6.27645910e-01 -3.34323436e-01 3.01518023e-01 -4.45436001e-01
-6.12116754e-01 -3.44591647e-01 6.29471481e-01 -8.79039392e-02
-5.14478385e-01 8.66061524e-02 -5.17174825e-02 -1.67059451e-01
2.55233586e-01 -5.72769046e-01 -7.88626671e-01 -4.19718146e-01
8.68113220e-01 7.72997618e-01 2.52181590e-01 -3.98001671e-02
5.26684642e-01 2.31486887e-01 -1.36139548e+00 6.01542234e-01
9.79338288e-01 8.49152803e-01 6.30806923e-01 8.12594056e-01
-6.57334685e-01 9.08811033e-01 -6.17992997e-01 -4.56719995e-01
2.27997731e-02 -2.28672951e-01 -6.50747299e-01 6.46574438e-01
8.44793737e-01 -5.43184757e-01 -5.79581201e-01 9.66155589e-01
6.24769747e-01 -1.11066833e-01 7.00148940e-01 -7.76532173e-01
-6.99663043e-01 7.53969610e-01 -7.11542547e-01 -1.79595500e-01
2.80403793e-01 2.74079409e-03 6.51695371e-01 -1.47729170e+00
4.69997436e-01 1.08101606e+00 8.45856428e-01 3.95551980e-01
-7.77551651e-01 -3.46323460e-01 1.58018857e-01 -9.93177481e-03
-1.44341648e+00 -4.86645818e-01 8.31552267e-01 -5.54059982e-01
8.92839849e-01 3.92924607e-01 1.29379883e-01 9.75368619e-01
-5.99709116e-02 1.77058876e+00 5.70368171e-01 -6.95188582e-01
-9.09610093e-02 -1.46534946e-02 2.67142713e-01 8.20537984e-01
2.73238093e-01 -5.34457743e-01 -5.09590566e-01 3.11334491e-01
7.87568629e-01 2.33580410e-01 -2.71572322e-01 -1.37411401e-01
-1.53438866e+00 5.81073642e-01 9.70752165e-02 4.70812857e-01
-1.61731854e-01 4.68928739e-02 6.33247375e-01 -9.93513167e-02
6.50353193e-01 1.41721277e-03 -2.22093895e-01 1.27007551e-02
-1.40698874e+00 -1.26388013e-01 6.40003324e-01 9.87951040e-01
3.44040900e-01 4.68489796e-01 -5.63323140e-01 1.28881633e+00
2.24357277e-01 8.91626894e-01 7.32586741e-01 1.17780805e-01
8.91769826e-01 8.79390478e-01 -7.99007248e-03 -8.71767044e-01
-3.57058316e-01 -9.44140255e-02 -1.12740648e+00 -1.73419729e-01
3.17020923e-01 -7.32874721e-02 -1.66140628e+00 6.98420584e-01
1.17289506e-01 -6.74672574e-02 -1.57277837e-01 8.89070094e-01
1.36070895e+00 6.87364578e-01 -1.96442276e-01 3.74536812e-01
1.18017471e+00 -1.31760705e+00 -9.56430674e-01 -4.05338943e-01
8.25156152e-01 -1.14629853e+00 1.28264022e+00 4.54530895e-01
-8.57654035e-01 -6.50548220e-01 -8.61564219e-01 -4.60076541e-01
-6.21838450e-01 8.93355131e-01 1.80135220e-01 6.95486963e-01
-9.11037147e-01 1.33844823e-01 -7.68545151e-01 -7.28451312e-01
5.41901469e-01 2.52023160e-01 2.82922294e-03 1.93138141e-02
-6.10569537e-01 4.74202752e-01 6.44831896e-01 4.32766557e-01
-8.25711966e-01 -1.22191615e-01 -7.68183768e-01 5.35496883e-02
5.60784400e-01 -5.84232986e-01 1.20108807e+00 -1.07691085e+00
-1.71822286e+00 1.18037152e+00 -1.84254855e-01 -4.28175569e-01
1.07568121e+00 -6.45401359e-01 -2.26246029e-01 2.49775916e-01
1.28961027e-01 6.09198630e-01 1.29565477e+00 -1.07989836e+00
-5.45899630e-01 -2.39725441e-01 -6.47845507e-01 3.40426892e-01
-6.05914533e-01 2.25505933e-01 -8.65777552e-01 -9.71372008e-01
1.95285901e-01 -4.49399352e-01 3.26539986e-02 2.39317283e-01
-1.11211991e+00 -3.44592243e-01 1.50974381e+00 -8.00863385e-01
7.87152529e-01 -1.92702222e+00 -1.64068148e-01 -1.70864075e-01
2.43777812e-01 4.79702175e-01 -1.34466082e-01 3.69310349e-01
1.16822593e-01 5.38580008e-02 -3.27230930e-01 -8.98489892e-01
3.14410478e-01 -1.48225889e-01 -5.08693755e-01 6.10655785e-01
5.76822124e-02 1.12821484e+00 -2.23750368e-01 -9.00820792e-01
1.00449562e+00 4.07238275e-01 1.04002751e-01 1.43048748e-01
-5.24569273e-01 -9.11478102e-02 -5.26471734e-01 9.13924873e-01
8.59630287e-01 -3.72646272e-01 -3.16867143e-01 9.59630534e-02
-1.38860598e-01 -2.19667733e-01 -1.06523693e+00 1.57224536e+00
-7.75391161e-02 1.27073455e+00 1.75889879e-01 -1.11171544e+00
1.06549215e+00 1.73283353e-01 2.98179001e-01 -7.70745158e-01
5.27634799e-01 8.64707753e-02 -8.02743912e-01 -6.82474852e-01
1.10705185e+00 4.21889752e-01 3.67268138e-02 3.29259366e-01
-3.51696521e-01 -3.41064751e-01 2.07285136e-01 8.05999860e-02
6.69979692e-01 2.32241049e-01 -1.11686029e-01 -7.67178982e-02
6.53591990e-01 1.91544205e-01 -1.93750829e-01 1.06034768e+00
2.12666579e-02 1.03412199e+00 -2.29457933e-02 -7.15913653e-01
-1.17855895e+00 -7.17055976e-01 -2.30603620e-01 1.28587663e+00
3.99191350e-01 -2.63799459e-01 -1.03934693e+00 -6.16475165e-01
-2.16959327e-01 5.73612452e-01 -5.66548288e-01 4.07322735e-01
-6.90334916e-01 -6.06147170e-01 8.68656874e-01 4.45463359e-01
1.13504624e+00 -1.35057509e+00 -5.16152978e-01 1.01338811e-01
-4.69424754e-01 -1.64732003e+00 -7.30864465e-01 1.91786140e-01
-8.02064300e-01 -8.56127501e-01 -1.28645849e+00 -1.33263278e+00
8.13131452e-01 4.71280158e-01 8.47915113e-01 5.19635119e-02
-6.92532599e-01 4.62558001e-01 -5.05450904e-01 -3.91053528e-01
-3.21847409e-01 1.54565915e-01 -4.40561742e-01 4.28030074e-01
3.15943509e-01 4.08089280e-01 -4.86137092e-01 4.51267928e-01
-1.24828660e+00 5.62940657e-01 6.23511374e-01 7.63156176e-01
5.32882452e-01 1.51055828e-01 1.67516433e-02 -8.12574685e-01
5.40776670e-01 1.95196539e-01 -4.95455742e-01 3.95535469e-01
-2.31048584e-01 -2.53602058e-01 7.18213618e-01 -4.27218378e-01
-1.08230758e+00 5.65204382e-01 -1.25814945e-01 -5.28494596e-01
-5.63888431e-01 3.78399581e-01 2.87360288e-02 -5.83785549e-02
5.59665143e-01 1.00076175e+00 -4.36654091e-01 -5.43805063e-01
3.45084190e-01 1.18065619e+00 5.73951483e-01 -1.41514674e-01
7.52128303e-01 7.30031431e-01 -5.88731110e-01 -1.42596638e+00
-6.89554930e-01 -7.26090312e-01 -9.93801713e-01 -2.01787814e-01
1.27003682e+00 -8.83198082e-01 -4.26251620e-01 1.30411911e+00
-1.17612743e+00 -6.05464280e-01 -1.22436928e-02 1.72792599e-02
-5.58008790e-01 8.62297952e-01 -6.08723700e-01 -8.19134772e-01
-9.53054011e-01 -9.00016069e-01 1.90773082e+00 6.68871701e-02
2.36444339e-01 -1.00634837e+00 -3.90238553e-01 5.97317278e-01
1.92971691e-01 -3.19625204e-03 4.78308320e-01 -7.96084404e-01
-5.96009612e-01 -6.04819298e-01 -6.24721408e-01 2.19372615e-01
-4.54257205e-02 3.19150500e-02 -1.03409874e+00 -2.03742474e-01
-1.96762159e-01 -3.22663367e-01 1.15893173e+00 5.27044415e-01
1.31704032e+00 -1.40337810e-01 -4.91972774e-01 7.01531470e-01
1.41286111e+00 2.21708231e-02 8.21543336e-01 5.07613301e-01
1.37179184e+00 3.49971443e-01 4.52872723e-01 4.97623265e-01
2.76843160e-01 4.72464561e-01 3.32919776e-01 -7.68263757e-01
-2.43519783e-01 -1.89886674e-01 1.93979159e-01 5.61239660e-01
6.77568853e-01 -7.99451172e-01 -1.17339528e+00 5.72219014e-01
-1.94712245e+00 -6.54758453e-01 -5.47450900e-01 1.80552423e+00
3.92496049e-01 2.81118453e-01 4.39566001e-02 3.14103633e-01
1.21370590e+00 2.64154643e-01 -6.42180860e-01 -1.61340207e-01
-4.59934860e-01 -8.75308886e-02 6.20110571e-01 1.31629393e-01
-1.52596974e+00 1.55289948e+00 5.68728733e+00 1.23386884e+00
-1.29925370e+00 -1.32166639e-01 6.69780314e-01 3.59095514e-01
3.46080273e-01 -5.25664926e-01 -8.21508050e-01 3.45131218e-01
1.79902807e-01 1.64418429e-01 1.54812947e-01 9.26874876e-01
3.36460769e-01 -3.33010368e-02 -9.61012721e-01 1.31480289e+00
5.91345429e-01 -1.20716417e+00 3.92847508e-01 -4.48398322e-01
7.45321572e-01 1.99142933e-01 1.88774220e-03 1.56283766e-01
9.23437178e-02 -1.14326012e+00 9.36611414e-01 3.41912776e-01
1.01046145e+00 -4.42041248e-01 6.03060961e-01 4.40052032e-01
-1.38311017e+00 1.74985901e-01 -3.45125347e-01 2.90282309e-01
8.26900266e-03 5.91498256e-01 -1.19906032e+00 4.58879173e-01
5.51384091e-01 1.03581059e+00 -7.89786398e-01 1.19350326e+00
-2.27669906e-02 6.65740848e-01 -3.99237812e-01 -5.91682315e-01
4.02394444e-01 -5.26786782e-03 4.49474275e-01 1.69759512e+00
1.64546236e-01 -3.74298096e-01 4.92375433e-01 1.03087187e+00
-3.17617357e-01 4.15094316e-01 -4.87076461e-01 -2.46761978e-01
-4.55103349e-03 1.20027745e+00 -1.41950977e+00 -6.83413148e-01
-8.99698213e-02 1.52185440e+00 -3.82528715e-02 3.68882298e-01
-6.98353231e-01 -9.31721151e-01 -4.33616281e-01 -9.69421193e-02
7.13805556e-01 -2.70027995e-01 -8.48490953e-01 -1.26477253e+00
3.62233341e-01 -7.22855866e-01 1.63805008e-01 -1.14146101e+00
-1.12193799e+00 5.69695115e-01 -5.50206780e-01 -1.18132174e+00
2.53127307e-01 -8.71488571e-01 -7.13029325e-01 4.96709824e-01
-1.37088847e+00 -1.65878153e+00 -6.34053469e-01 8.51129293e-01
1.46954739e+00 -7.21295401e-02 3.68977726e-01 1.33439943e-01
-9.21440125e-01 7.48923242e-01 6.87452435e-01 6.99182212e-01
5.04158854e-01 -1.10414171e+00 8.31805289e-01 1.04647219e+00
2.71749765e-01 -4.10176143e-02 5.61803699e-01 -8.59777212e-01
-1.45467293e+00 -1.44218421e+00 5.60798347e-01 -4.90559697e-01
3.97841007e-01 -8.51014197e-01 -7.89332747e-01 5.72678506e-01
3.24720234e-01 -4.24507737e-01 -1.38662487e-01 -3.15464079e-01
-1.26075283e-01 2.76798040e-01 -9.69380021e-01 8.44520628e-01
8.66595626e-01 -4.60309178e-01 -4.67542380e-01 8.94704461e-01
5.41090667e-01 -7.17078149e-01 -1.66826978e-01 2.60524750e-01
3.24042678e-01 -6.53219461e-01 7.94708550e-01 4.82379757e-02
4.40552503e-01 -7.22393319e-02 6.61670119e-02 -5.96793056e-01
1.39861956e-01 -5.13070643e-01 2.71424234e-01 1.14998448e+00
2.40729436e-01 -3.33128572e-01 1.11198366e+00 2.57204056e-01
-4.26317513e-01 -1.79994047e-01 -8.54117453e-01 -6.22633040e-01
1.91451795e-02 -5.24044096e-01 2.23770037e-01 1.00594854e+00
-3.26559365e-01 3.48809749e-01 -6.50717676e-01 -4.46663722e-02
6.48980677e-01 1.96524769e-01 1.03958952e+00 -1.09752488e+00
1.23579629e-01 -7.41537392e-01 -2.62626767e-01 -1.68587196e+00
6.08679205e-02 -8.12418103e-01 4.96263474e-01 -1.97653270e+00
1.98642075e-01 2.27899682e-02 3.01637381e-01 5.43292761e-01
-3.20248902e-01 3.16381246e-01 3.32404196e-01 2.73724496e-01
-1.13623941e+00 7.12529719e-01 1.30609083e+00 -5.87166786e-01
-2.80963808e-01 -1.34356141e-01 -6.64759204e-02 7.60404289e-01
6.56651199e-01 -2.07717001e-01 -1.40044540e-02 -8.30104232e-01
-1.30044192e-01 5.09136580e-02 3.74227971e-01 -9.89026129e-01
4.80311900e-01 1.47498414e-01 6.16321981e-01 -1.59700119e+00
3.76436770e-01 -8.02109957e-01 -5.68754792e-01 3.58972728e-01
-4.74640191e-01 -3.16672027e-01 3.60017717e-01 4.98042375e-01
-4.74955514e-02 -3.98725659e-01 6.99439049e-01 2.31099855e-02
-7.54878402e-01 1.95446253e-01 -7.51696885e-01 -4.08683829e-02
7.84520507e-01 -7.33041823e-01 -5.08293092e-01 -3.09823990e-01
-4.39650327e-01 1.52227223e-01 3.10952127e-01 6.38508737e-01
1.05625415e+00 -7.97310293e-01 -9.92799640e-01 1.77246571e-01
2.91668057e-01 3.42452049e-01 2.35211194e-01 6.09838545e-01
-8.96788239e-01 6.78045034e-01 3.31008911e-01 -1.19589746e+00
-1.55626845e+00 4.86655027e-01 4.45657879e-01 -1.45920768e-01
-9.88172650e-01 7.14256585e-01 4.58294868e-01 -4.85011071e-01
5.90555370e-01 -4.85171586e-01 3.21769230e-02 -3.75407577e-01
3.71165216e-01 2.14827955e-01 1.21505365e-01 -5.75403273e-01
-1.16853468e-01 1.01294863e+00 -3.98547471e-01 1.60720766e-01
1.11980891e+00 -3.25038254e-01 -4.94655147e-02 3.51536304e-01
8.77118945e-01 -2.92367697e-01 -9.86156821e-01 -4.16980952e-01
5.91158401e-03 -3.93562466e-01 1.38796628e-01 -1.02468419e+00
-9.43039656e-01 1.09957933e+00 7.26838171e-01 3.11995447e-01
9.69676018e-01 -3.59664857e-01 9.87027705e-01 9.15084958e-01
-1.88425798e-02 -1.36020660e+00 4.03152108e-01 7.62007833e-01
7.69940317e-01 -1.57000160e+00 -1.39497072e-01 -3.67610723e-01
-5.07511556e-01 1.30108845e+00 7.14687049e-01 -6.10755058e-04
1.19249977e-01 3.11112911e-01 1.68473870e-01 -2.27432504e-01
-1.81996241e-01 -2.87117004e-01 3.44443679e-01 3.72000188e-01
2.39644304e-01 -9.19340272e-03 2.46826276e-01 5.15818149e-02
1.22124599e-02 2.65021506e-03 5.29264927e-01 9.63697016e-01
-7.45903611e-01 -4.82427269e-01 -7.75868118e-01 6.23138189e-01
-4.73190784e-01 -4.39271152e-01 -9.16093230e-01 7.71591783e-01
-3.58021796e-01 9.86967802e-01 8.24590847e-02 -2.34502137e-01
2.34671697e-01 3.11692152e-02 1.45401046e-01 -4.31759685e-01
-5.20292640e-01 3.51074815e-01 5.35709597e-03 2.30655633e-02
-1.45743057e-01 -2.95258880e-01 -1.32184744e+00 -1.59703791e-01
-8.42155337e-01 -3.90697479e-01 1.00541854e+00 1.01010287e+00
2.55901754e-01 5.88194907e-01 6.05852723e-01 -9.62269247e-01
-1.36243388e-01 -1.19384789e+00 -5.84245086e-01 5.03712654e-01
4.01522338e-01 -5.84466197e-02 -1.69049188e-01 6.83085680e-01]
|
[12.031879425048828, 2.3055598735809326]
|
d4b6044a-fbee-4ddd-9214-522b1ef59948
|
beyond-simple-meta-learning-multi-purpose
|
2201.05151
| null |
https://arxiv.org/abs/2201.05151v2
|
https://arxiv.org/pdf/2201.05151v2.pdf
|
Beyond Simple Meta-Learning: Multi-Purpose Models for Multi-Domain, Active and Continual Few-Shot Learning
|
Modern deep learning requires large-scale extensively labelled datasets for training. Few-shot learning aims to alleviate this issue by learning effectively from few labelled examples. In previously proposed few-shot visual classifiers, it is assumed that the feature manifold, where classifier decisions are made, has uncorrelated feature dimensions and uniform feature variance. In this work, we focus on addressing the limitations arising from this assumption by proposing a variance-sensitive class of models that operates in a low-label regime. The first method, Simple CNAPS, employs a hierarchically regularized Mahalanobis-distance based classifier combined with a state of the art neural adaptive feature extractor to achieve strong performance on Meta-Dataset, mini-ImageNet and tiered-ImageNet benchmarks. We further extend this approach to a transductive learning setting, proposing Transductive CNAPS. This transductive method combines a soft k-means parameter refinement procedure with a two-step task encoder to achieve improved test-time classification accuracy using unlabelled data. Transductive CNAPS achieves state of the art performance on Meta-Dataset. Finally, we explore the use of our methods (Simple and Transductive) for "out of the box" continual and active learning. Extensive experiments on large scale benchmarks illustrate robustness and versatility of this, relatively speaking, simple class of models. All trained model checkpoints and corresponding source codes have been made publicly available.
|
['Frank Wood', 'Leonid Sigal', 'Jan-Willem van de Meent', 'Vaden Masrani', 'Raghav Goyal', 'Jarred Barber', 'Peyman Bateni']
|
2022-01-13
| null | null | null | null |
['cross-domain-few-shot']
|
['computer-vision']
|
[ 4.95251119e-01 2.13332623e-01 -4.28443849e-01 -4.35082376e-01
-1.04563463e+00 -2.77849823e-01 9.55557108e-01 8.38961601e-02
-5.89663625e-01 6.79358482e-01 -5.12574278e-02 1.18658431e-01
-3.29174489e-01 -6.01458073e-01 -6.61193073e-01 -1.06304836e+00
3.33240144e-02 4.33496088e-01 3.21426660e-01 -2.21961930e-01
2.61686653e-01 2.04778329e-01 -1.94154429e+00 3.18693370e-01
5.34106076e-01 1.20637727e+00 -9.32433382e-02 7.39692390e-01
-1.43341571e-01 1.16872954e+00 -4.29507464e-01 -3.60380232e-01
3.70886087e-01 -4.48716134e-01 -8.68757427e-01 3.03007573e-01
3.42778295e-01 -1.91752687e-02 1.84381112e-01 8.60375822e-01
6.38375998e-01 4.69965518e-01 9.77682114e-01 -1.46439195e+00
-6.62094891e-01 2.92677909e-01 -4.00641501e-01 2.99828231e-01
-1.70299292e-01 4.90289897e-01 9.49065387e-01 -1.15068698e+00
7.18670011e-01 9.62403953e-01 7.84474790e-01 7.97416627e-01
-1.38285792e+00 -3.59398633e-01 -4.25147861e-02 4.22398508e-01
-1.32144856e+00 -6.72810197e-01 8.15497339e-01 -5.68622053e-01
1.23323679e+00 2.02675000e-01 5.86616158e-01 1.28433275e+00
1.07127406e-01 8.23105752e-01 1.31927586e+00 -8.10812831e-01
8.61817300e-01 3.00408483e-01 3.61971915e-01 7.53233016e-01
-8.48898441e-02 2.16247454e-01 -5.18122375e-01 -1.98446617e-01
2.67660886e-01 1.22339211e-01 -6.58548325e-02 -9.25178051e-01
-9.16595697e-01 1.14974606e+00 4.52797055e-01 3.35255712e-01
-1.86361566e-01 1.50646135e-01 6.67895675e-01 4.53718424e-01
9.35003161e-01 4.12456751e-01 -3.50102246e-01 -1.61944836e-01
-8.98715675e-01 -2.26489723e-01 7.54006565e-01 9.27593231e-01
1.00325572e+00 7.50667229e-02 -5.12205720e-01 1.07719350e+00
1.32488608e-01 1.26304254e-01 6.74888432e-01 -8.39675725e-01
-2.17816141e-02 5.52639484e-01 -3.08744013e-01 -4.18053359e-01
-2.87640363e-01 -4.64942127e-01 -7.59952545e-01 5.32324910e-01
3.59140038e-02 2.10606828e-02 -1.49954271e+00 1.40297270e+00
4.06550616e-01 5.00646412e-01 3.37302312e-02 7.39887238e-01
8.50623608e-01 5.29625535e-01 3.04066181e-01 -4.38509345e-01
9.18231964e-01 -1.12085819e+00 -3.74125689e-01 -8.98340493e-02
9.60077286e-01 -2.83276707e-01 1.16695094e+00 1.50984764e-01
-6.59241021e-01 -5.60526609e-01 -1.19327283e+00 3.44796292e-02
-7.94688344e-01 -2.54290760e-01 5.52097738e-01 6.93814874e-01
-1.07591307e+00 7.04825938e-01 -7.06537962e-01 -5.56408167e-01
9.09966409e-01 2.54346162e-01 -4.09004360e-01 -1.56315729e-01
-9.89670694e-01 7.05030084e-01 5.26949048e-01 -2.36644089e-01
-1.18760908e+00 -7.38529801e-01 -9.01944876e-01 -1.16336094e-02
5.82468390e-01 -6.32305264e-01 1.38391697e+00 -1.25063121e+00
-1.65135467e+00 1.08611393e+00 2.20493883e-01 -4.86937612e-01
5.12159646e-01 -6.41531572e-02 -4.77675907e-02 2.12855101e-01
4.02422212e-02 8.08818936e-01 1.15322268e+00 -1.29618466e+00
-4.68589395e-01 -2.57702321e-01 -2.21775278e-01 1.55652925e-01
-5.37231982e-01 -6.30811378e-02 -2.49335155e-01 -3.51284176e-01
-3.13424975e-01 -9.21629429e-01 -2.69792050e-01 5.53117841e-02
-1.26046449e-01 -3.79557073e-01 8.47504616e-01 1.71413422e-01
8.34853590e-01 -2.09492970e+00 3.11317779e-02 -2.03949753e-02
2.63218731e-01 6.12778604e-01 -1.66469097e-01 3.96248549e-01
-1.26951784e-01 -8.49251598e-02 -4.10684407e-01 -5.98953128e-01
2.80792322e-02 3.00133288e-01 -2.42582440e-01 7.39041984e-01
5.21209836e-01 1.29504573e+00 -1.00854278e+00 -5.85819423e-01
4.22647983e-01 4.67510581e-01 -1.53130412e-01 1.53236657e-01
-4.38203737e-02 4.16678526e-02 2.69197412e-02 7.75122046e-01
3.33496034e-01 -4.29493368e-01 -3.36675227e-01 1.25981584e-01
1.72958896e-03 -3.19126964e-01 -8.23909163e-01 1.81603312e+00
-3.12674493e-01 5.95038652e-01 -3.76992285e-01 -1.41412854e+00
8.59055638e-01 2.15324596e-01 4.49867368e-01 -7.94601321e-01
3.65880609e-01 3.35336104e-02 -2.83935785e-01 -4.41381514e-01
2.11949438e-01 -3.60902071e-01 -3.46873552e-02 2.25963071e-01
6.28744960e-01 1.10231955e-02 2.52708465e-01 1.95130631e-01
1.27862680e+00 1.52882844e-01 6.21488214e-01 -3.01286966e-01
1.55669630e-01 1.56698272e-01 4.43148524e-01 1.02883673e+00
-5.86919427e-01 7.06146896e-01 3.07597101e-01 -3.45917463e-01
-1.09962833e+00 -9.95487750e-01 -1.99134246e-01 1.53569138e+00
-1.05807923e-01 -2.74834603e-01 -6.64326072e-01 -1.04264700e+00
-9.83868390e-02 8.19768250e-01 -1.25835407e+00 -6.28196120e-01
-1.11125678e-01 -8.24650228e-01 4.11874920e-01 5.93819261e-01
3.27324092e-01 -1.08228815e+00 -8.03018987e-01 8.72961655e-02
4.10931051e-01 -6.83956087e-01 6.75413832e-02 7.98177004e-01
-6.82291150e-01 -9.94725764e-01 -9.46678340e-01 -7.56750464e-01
4.77799445e-01 3.09372157e-01 1.04829466e+00 -1.81703493e-01
-7.06794679e-01 6.45857334e-01 -7.01041639e-01 -4.45363671e-01
-2.93994099e-01 3.08889765e-02 -1.49349377e-01 1.97906300e-01
5.48995376e-01 -3.15145493e-01 -4.83880550e-01 1.62790194e-01
-8.96607995e-01 7.46405730e-03 6.93822443e-01 1.20996094e+00
5.26836991e-01 -2.69273162e-01 6.23083174e-01 -1.08414280e+00
3.44826162e-01 -7.33212531e-01 -1.57011002e-01 3.44738513e-01
-8.76856923e-01 3.60960397e-03 6.14759564e-01 -6.71567261e-01
-9.96360064e-01 3.03193331e-01 9.70929563e-02 -8.23785901e-01
-4.11187500e-01 1.94159538e-01 1.63809866e-01 -1.96217120e-01
1.12488425e+00 2.85750806e-01 -7.50610605e-02 -2.71164954e-01
5.67520916e-01 8.07177544e-01 2.62469977e-01 -1.84266716e-01
6.73370302e-01 6.54292226e-01 1.52690290e-02 -8.95772874e-01
-9.85622406e-01 -8.44349325e-01 -8.10153365e-01 -3.91095430e-01
7.05601275e-01 -7.84432352e-01 -2.93348283e-01 5.89920819e-01
-7.05326498e-01 -6.50462329e-01 -8.18061292e-01 3.63444686e-01
-9.22146857e-01 7.94716030e-02 -4.20271754e-01 -8.96400154e-01
-4.82838631e-01 -8.75498354e-01 1.05961907e+00 1.61254510e-01
-1.00182377e-01 -1.11734700e+00 4.79858935e-01 2.23636851e-01
4.71488297e-01 2.02074558e-01 5.63999176e-01 -9.83703613e-01
-7.52957538e-02 -2.14296684e-01 -1.53961346e-01 4.13054347e-01
-1.28717860e-02 -9.62136835e-02 -1.48592877e+00 -3.94824475e-01
-1.46444678e-01 -1.00241172e+00 1.36883497e+00 2.90420681e-01
8.40448260e-01 -2.04383433e-02 -2.46555388e-01 5.30762792e-01
1.53716588e+00 -4.79178727e-02 5.63852608e-01 3.07936013e-01
7.03502536e-01 5.64947009e-01 8.17647755e-01 3.89863372e-01
1.13219902e-01 5.54368913e-01 4.28922504e-01 -1.72778949e-01
-1.61487684e-01 1.30810261e-01 2.19070211e-01 7.11294711e-01
-4.08722199e-02 -1.47942528e-01 -1.05607665e+00 5.73814332e-01
-2.06377912e+00 -1.10498643e+00 2.50776052e-01 2.17931247e+00
8.06567490e-01 3.59462857e-01 1.54159024e-01 4.08852398e-01
6.64982021e-01 8.67485180e-02 -6.75274074e-01 -4.61312801e-01
2.45059356e-02 2.64579028e-01 4.88707185e-01 2.00405434e-01
-1.42562056e+00 8.54751885e-01 5.56245422e+00 1.14150155e+00
-1.21445417e+00 4.51588571e-01 7.61201978e-01 -3.34825575e-01
1.47231832e-01 7.31710196e-02 -7.66007781e-01 3.51851493e-01
1.06690836e+00 -1.99809708e-02 1.04156710e-01 1.20341134e+00
-2.27171063e-01 -1.20153330e-01 -1.13924742e+00 9.87262785e-01
3.76873583e-01 -1.45530283e+00 -2.57803410e-01 -3.38253519e-03
8.84585023e-01 2.78171897e-01 1.14219882e-01 8.82441759e-01
2.60246396e-01 -9.13937211e-01 6.44261718e-01 6.54009104e-01
9.40908551e-01 -7.51697123e-01 5.61672211e-01 3.24353218e-01
-1.06334078e+00 -4.14457411e-01 -4.22956795e-01 -5.32238856e-02
-2.42740333e-01 5.02970934e-01 -9.11443412e-01 3.28633904e-01
6.90420806e-01 8.11698139e-01 -9.04469669e-01 1.27850080e+00
2.21778870e-01 7.67970145e-01 -2.83080246e-02 -2.24698737e-01
4.86105621e-01 3.58408570e-01 4.17962730e-01 1.34747779e+00
-1.09517545e-01 -6.88008666e-02 2.10831940e-01 6.07333004e-01
3.97507548e-02 2.22515479e-01 -8.25414538e-01 3.44214737e-02
2.60141075e-01 1.49854219e+00 -1.18598294e+00 -6.38445854e-01
-3.86745423e-01 1.06016445e+00 7.04631209e-01 1.87914833e-01
-5.97439706e-01 -4.60023224e-01 1.67849675e-01 -4.09588590e-02
4.21937376e-01 1.53753057e-01 -1.01633035e-01 -1.04645658e+00
-3.23444009e-01 -5.13221979e-01 4.39882994e-01 -4.61284339e-01
-1.50209379e+00 4.24613029e-01 1.82308525e-01 -1.49455416e+00
-4.13776875e-01 -5.65952957e-01 -6.48862362e-01 3.59480083e-01
-1.39156342e+00 -1.23990691e+00 -4.11136210e-01 6.76599145e-01
8.75008345e-01 -2.83584386e-01 8.63357008e-01 2.73626763e-02
-5.83127022e-01 6.76884830e-01 3.71125489e-01 1.41987391e-03
7.27694988e-01 -1.39659345e+00 3.74690384e-01 5.42132735e-01
2.81060576e-01 2.66784817e-01 5.56210399e-01 -4.16589200e-01
-1.19285500e+00 -1.35738766e+00 3.79400641e-01 -4.09152180e-01
6.24223232e-01 -5.67856252e-01 -9.50045764e-01 4.60315883e-01
1.89483643e-01 6.32830381e-01 8.98127437e-01 -3.26781832e-02
-4.26770747e-01 -4.33823057e-02 -1.11189818e+00 2.24787697e-01
9.95727777e-01 -4.58441973e-01 -6.27155721e-01 3.46011758e-01
6.85750782e-01 8.31691548e-02 -6.40444219e-01 4.52284932e-01
3.43502998e-01 -1.02237535e+00 7.81580806e-01 -6.71653926e-01
3.88203472e-01 1.28698245e-01 -6.21782988e-02 -1.51974702e+00
-4.83471602e-01 -4.88483340e-01 -3.65605146e-01 1.18732667e+00
2.20807478e-01 -5.26484072e-01 8.25076163e-01 3.90687376e-01
-3.04772347e-01 -1.03106427e+00 -1.07978082e+00 -1.01142454e+00
7.56795853e-02 -3.59355897e-01 -9.42002982e-02 1.16961062e+00
9.62030794e-03 6.49011075e-01 -4.00228143e-01 -5.72874486e-01
9.30650413e-01 -2.15206385e-01 6.47485435e-01 -1.39691055e+00
-3.06651026e-01 -3.82831216e-01 -7.61414766e-01 -3.00598145e-01
2.47499347e-01 -9.37363744e-01 2.38755360e-01 -1.28205943e+00
3.84846091e-01 -3.29582363e-01 -5.32775283e-01 7.14225590e-01
-1.34948701e-01 7.16153383e-01 9.86856893e-02 2.64190674e-01
-1.13574433e+00 7.19082892e-01 7.85525501e-01 -2.52780765e-01
-2.35399663e-01 -2.08997294e-01 -3.43342602e-01 5.24319589e-01
7.14691222e-01 -5.30676842e-01 -6.53434694e-01 9.44339484e-02
-6.25669863e-03 -4.43927616e-01 5.38568079e-01 -1.18095970e+00
3.03212374e-01 -9.08442661e-02 4.73604083e-01 -1.85723245e-01
6.03208721e-01 -6.75554514e-01 -2.40141496e-01 4.61246997e-01
-6.39914393e-01 -5.94162166e-01 -6.59328997e-02 7.92700231e-01
3.59979048e-02 -3.09964657e-01 1.19956720e+00 -1.72629356e-01
-1.11866915e+00 3.05758506e-01 -2.61749923e-01 3.04542005e-01
1.55064249e+00 -4.78876144e-01 -5.11934936e-01 -3.13464627e-02
-7.10571289e-01 -7.08546788e-02 4.36970055e-01 4.25975502e-01
6.27764225e-01 -1.34268165e+00 -6.20135605e-01 1.50461569e-01
7.28607833e-01 -3.08799386e-01 2.37362579e-01 1.05779266e+00
-1.15755774e-01 1.46587178e-01 -2.09061250e-01 -8.32737505e-01
-1.27413368e+00 9.05899286e-01 3.20610791e-01 1.89068932e-02
-6.92101002e-01 9.79825497e-01 -3.19203250e-02 -3.16111028e-01
2.50951409e-01 1.64132863e-01 -2.62696624e-01 4.22145694e-01
6.91168904e-01 5.60369968e-01 3.60380977e-01 -7.14771330e-01
-3.31231892e-01 3.94572020e-01 -3.38693082e-01 9.18407273e-03
1.35654426e+00 -4.59561795e-02 3.89086217e-01 1.03275764e+00
1.67084372e+00 -6.59053385e-01 -1.54194951e+00 -5.37004769e-01
3.75250578e-02 -2.66168475e-01 3.28972667e-01 -6.21986747e-01
-8.51056635e-01 1.01301336e+00 1.03046799e+00 3.58217239e-01
8.94776165e-01 1.16111360e-01 3.32095623e-01 4.77761656e-01
1.72904000e-01 -1.45135772e+00 3.07094216e-01 3.84206355e-01
4.34956223e-01 -1.67374003e+00 -4.93282340e-02 -4.90400791e-02
-8.70409846e-01 9.46352780e-01 6.02144897e-01 -1.66225672e-01
6.95348203e-01 7.86594301e-02 1.41520873e-01 -3.86784226e-01
-1.13785708e+00 -6.01608276e-01 3.45820904e-01 6.48644984e-01
1.47457540e-01 -1.45545542e-01 3.82120721e-02 2.04651862e-01
2.74485052e-01 1.27958581e-01 2.37871751e-01 1.34526241e+00
-7.04086721e-01 -4.46236074e-01 -7.87947327e-02 5.94021618e-01
-1.21509880e-01 -1.63517505e-01 -4.81319129e-01 6.86113298e-01
6.27834648e-02 9.75939572e-01 1.50837451e-01 -4.80041444e-01
9.16046128e-02 4.10402179e-01 4.52397227e-01 -8.72465193e-01
-5.26447594e-01 -1.65237144e-01 -1.24853663e-01 -5.56265414e-01
-5.17275572e-01 -3.98238331e-01 -9.68263865e-01 -2.70990189e-02
-4.71858233e-01 -5.93927130e-02 5.68019927e-01 1.16177416e+00
3.33026737e-01 3.91812652e-01 8.60786915e-01 -1.16355252e+00
-8.40143025e-01 -1.01844704e+00 -5.67626119e-01 4.60649043e-01
3.09751809e-01 -9.28280354e-01 -6.50009990e-01 -6.45776391e-02]
|
[9.917192459106445, 2.812528610229492]
|
f779404d-63d6-4d57-a6fa-54fa3919a74e
|
danish-fungi-2020-not-just-another-image
|
2103.10107
| null |
https://arxiv.org/abs/2103.10107v4
|
https://arxiv.org/pdf/2103.10107v4.pdf
|
Danish Fungi 2020 -- Not Just Another Image Recognition Dataset
|
We introduce a novel fine-grained dataset and benchmark, the Danish Fungi 2020 (DF20). The dataset, constructed from observations submitted to the Atlas of Danish Fungi, is unique in its taxonomy-accurate class labels, small number of errors, highly unbalanced long-tailed class distribution, rich observation metadata, and well-defined class hierarchy. DF20 has zero overlap with ImageNet, allowing unbiased comparison of models fine-tuned from publicly available ImageNet checkpoints. The proposed evaluation protocol enables testing the ability to improve classification using metadata -- e.g. precise geographic location, habitat, and substrate, facilitates classifier calibration testing, and finally allows to study the impact of the device settings on the classification performance. Experiments using Convolutional Neural Networks (CNN) and the recent Vision Transformers (ViT) show that DF20 presents a challenging task. Interestingly, ViT achieves results superior to CNN baselines with 80.45% accuracy and 0.743 macro F1 score, reducing the CNN error by 9% and 12% respectively. A simple procedure for including metadata into the decision process improves the classification accuracy by more than 2.95 percentage points, reducing the error rate by 15%. The source code for all methods and experiments is available at https://sites.google.com/view/danish-fungi-dataset.
|
['Tobias Frøslev', 'Thomas Læssøe', 'Thomas S. Jeppesen', 'Jacob Heilmann-Clausen', 'Jiří Matas', 'Milan Šulc', 'Lukáš Picek']
|
2021-03-18
| null | null | null | null |
['fine-grained-image-recognition', 'classifier-calibration', 'classifier-calibration']
|
['computer-vision', 'computer-vision', 'miscellaneous']
|
[-9.10779834e-03 -3.44690681e-01 -1.00513034e-01 -5.94575033e-02
-4.33591574e-01 -9.45605040e-01 9.60771143e-01 2.77814955e-01
-4.80096787e-01 7.48327315e-01 -8.22955072e-02 -2.22974628e-01
-2.21332878e-01 -9.31282401e-01 -7.61749744e-01 -7.72130132e-01
-1.91307604e-01 3.27041268e-01 2.80472428e-01 3.45769078e-01
1.22425944e-01 5.01426101e-01 -1.58448601e+00 2.67026335e-01
5.61835945e-01 1.30946660e+00 5.36389410e-01 6.40670717e-01
1.63500816e-01 4.94484335e-01 -9.19643939e-01 -4.28534776e-01
3.51965487e-01 3.97375643e-01 -5.62055886e-01 -3.19812417e-01
6.67795599e-01 -3.25248241e-01 -3.42270185e-04 8.51886868e-01
4.27553862e-01 -2.58232385e-01 9.56116855e-01 -1.27723753e+00
-6.27263069e-01 3.95717859e-01 -1.54073849e-01 2.34393507e-01
-8.98064971e-02 4.70716804e-01 7.20030904e-01 -5.55542886e-01
6.80959404e-01 8.76274943e-01 1.25194621e+00 2.20052898e-01
-1.20894682e+00 -7.95925677e-01 6.78088591e-02 1.29477024e-01
-1.72037983e+00 -1.79289579e-01 5.23200370e-02 -9.58389759e-01
1.23441410e+00 4.20402229e-01 3.94716054e-01 1.46644723e+00
1.54034778e-01 2.94093005e-02 1.12511671e+00 4.87408750e-02
2.00346917e-01 2.87603028e-02 -3.88438553e-02 5.56401134e-01
4.20102209e-01 5.68945646e-01 -5.04586875e-01 1.47828266e-01
7.89771497e-01 4.94795665e-02 -2.09046468e-01 3.98177467e-02
-1.45611048e+00 3.66723657e-01 7.92761266e-01 2.97616534e-02
-4.02246267e-01 8.04237202e-02 5.36547840e-01 1.48692966e-01
4.53973055e-01 5.57392597e-01 -9.06325281e-01 -1.78654447e-01
-6.45572603e-01 1.52091891e-01 7.59918451e-01 7.44202852e-01
6.02641046e-01 1.40946796e-02 -5.10775372e-02 9.30218816e-01
-6.10913597e-02 7.72267997e-01 2.44422689e-01 -8.96011233e-01
2.63417006e-01 6.63354576e-01 9.64234844e-02 -8.46146584e-01
-4.01749402e-01 -8.50239277e-01 -1.04811597e+00 1.91324621e-01
5.03256619e-01 4.40787487e-02 -1.08139777e+00 1.72714460e+00
1.35062233e-01 4.91422974e-02 -1.63656473e-01 4.75279868e-01
8.02647889e-01 4.19297665e-01 1.94606289e-01 1.74681142e-01
1.12851942e+00 -7.59510040e-01 -1.96603820e-01 1.11686550e-01
2.46903345e-01 -5.79352617e-01 1.08674300e+00 4.73712355e-01
-3.94037157e-01 -6.96187615e-01 -9.25819457e-01 3.61375183e-01
-1.12762427e+00 3.33552092e-01 5.64654529e-01 4.67665672e-01
-1.00675273e+00 6.74105823e-01 -8.18831503e-01 -8.03356647e-01
5.99350929e-01 4.63337749e-01 -6.21038735e-01 1.40713438e-01
-7.53822982e-01 6.27762496e-01 5.58299065e-01 -1.69657990e-01
-1.41420913e+00 -1.14169168e+00 -4.07265514e-01 5.39427809e-02
7.54345953e-02 -4.88458693e-01 1.29676342e+00 -6.12800121e-01
-1.25638640e+00 8.75150323e-01 2.61716276e-01 -5.25381684e-01
6.18786156e-01 -2.02775002e-01 -3.98452312e-01 -1.51133537e-01
3.47483307e-01 1.00298977e+00 2.07846045e-01 -9.91290629e-01
-9.89121735e-01 -3.83133978e-01 1.25943661e-01 -3.68163854e-01
-4.63412136e-01 -3.02998602e-01 -3.83576989e-01 -8.67805183e-01
-4.04356271e-01 -9.61758077e-01 1.02466922e-02 6.13970496e-03
-3.43971491e-01 -4.60400246e-03 6.30051076e-01 -7.12147236e-01
8.28789949e-01 -2.00426579e+00 -2.46912494e-01 -2.64716268e-01
8.68693441e-02 3.57487768e-01 -3.39182824e-01 3.29519778e-01
-9.20849107e-03 3.52547437e-01 -1.26664534e-01 -3.26794945e-02
-6.84206188e-02 1.16785944e-01 -2.00639114e-01 3.38587850e-01
3.68900895e-01 7.49657214e-01 -7.61909485e-01 1.30666807e-01
7.29126811e-01 7.17079461e-01 -4.45933372e-01 3.27103078e-01
-3.84183884e-01 4.09006208e-01 3.50887328e-02 1.07951570e+00
5.60310900e-01 -3.08402658e-01 -4.66600917e-02 -4.65143353e-01
-3.24917674e-01 1.64866999e-01 -9.41753447e-01 1.53517640e+00
-6.05786562e-01 7.95576751e-01 -1.15335681e-01 -5.51155627e-01
8.97420466e-01 1.47962809e-01 3.85045648e-01 -6.75545454e-01
9.04623643e-02 1.68970600e-01 -2.15527788e-01 -2.71509349e-01
1.17590263e-01 4.08232808e-01 1.13490388e-01 3.15437093e-02
1.90424576e-01 -6.73263147e-02 2.00776577e-01 -2.21595570e-01
1.21072996e+00 1.88172549e-01 3.55874866e-01 -4.29778904e-01
1.75175309e-01 1.54150411e-01 5.10651767e-01 8.09462845e-01
-2.76130378e-01 5.14677525e-01 1.32736459e-01 -7.97839344e-01
-9.62242961e-01 -1.30089092e+00 -4.71264303e-01 1.05197001e+00
-7.43396804e-02 -3.81280184e-01 -8.55583489e-01 -6.24228120e-01
2.46407241e-01 3.74456644e-01 -9.07926857e-01 3.22069563e-02
-1.58510415e-03 -8.88287663e-01 8.76304567e-01 7.89871454e-01
8.86543572e-01 -1.04532790e+00 -4.42643017e-01 9.81779769e-02
-6.77999184e-02 -1.08107960e+00 1.59685820e-01 7.24930346e-01
-5.52518308e-01 -1.36342180e+00 -4.81039375e-01 -6.84173584e-01
2.01712534e-01 1.46618620e-01 1.20409226e+00 -2.03750208e-01
-3.91645104e-01 -1.09257035e-01 -3.34660709e-01 -4.29955989e-01
-1.76476747e-01 6.02180541e-01 1.54997647e-01 -5.17987072e-01
5.55642784e-01 -6.93878055e-01 -5.92477560e-01 4.63606894e-01
-7.72965372e-01 1.54732596e-02 5.64982772e-01 7.86682308e-01
7.87775755e-01 1.69395193e-01 6.12741292e-01 -6.30926073e-01
1.05835751e-01 -6.69003069e-01 -6.90495372e-01 2.88315266e-01
-6.64662004e-01 -3.06421131e-01 8.71803820e-01 -3.74020338e-01
-8.56854439e-01 1.14706203e-01 -4.92297933e-02 -3.18134606e-01
-6.97099328e-01 2.35684797e-01 -3.50139499e-01 -1.28663212e-01
8.57923090e-01 2.80784070e-02 -3.12520444e-01 -7.51223505e-01
2.81884253e-01 1.08457661e+00 1.06232011e+00 -6.11589909e-01
5.63989758e-01 4.97290939e-01 -1.34803370e-01 -6.51832342e-01
-7.35571861e-01 -5.24842560e-01 -7.99932420e-01 -1.06468372e-01
6.18296504e-01 -1.20709431e+00 -7.71690786e-01 9.10497487e-01
-9.28540289e-01 -7.76028037e-01 -8.99545178e-02 2.30778694e-01
-2.72786558e-01 -8.72717202e-02 -5.02825141e-01 -3.42385143e-01
-5.18432021e-01 -1.05773544e+00 1.22670496e+00 9.03386176e-02
-1.59390673e-01 -9.48904753e-01 -7.09572528e-03 2.12346822e-01
5.80339611e-01 6.20886326e-01 5.95998824e-01 -5.74249268e-01
-4.45513457e-01 -2.57866800e-01 -4.79374200e-01 4.04591262e-01
5.24537981e-01 4.50825781e-01 -1.35228264e+00 -3.46632242e-01
-5.62786281e-01 -1.69302285e-01 1.03695250e+00 3.93887371e-01
1.50831592e+00 -1.78191632e-01 -4.05113846e-01 9.99022186e-01
1.61154175e+00 2.18515635e-01 2.84323752e-01 6.17483735e-01
6.82926595e-01 3.48384589e-01 4.47624922e-01 5.07067442e-01
4.04522300e-01 6.81015372e-01 1.03631485e+00 7.84279332e-02
-5.77133060e-01 -1.87420156e-02 1.64000899e-01 5.29921293e-01
-1.17528141e-01 -4.53859806e-01 -1.05238378e+00 7.82817900e-01
-1.53349388e+00 -7.58289456e-01 -1.41751841e-01 2.26050758e+00
6.32514536e-01 9.09664668e-03 5.74885197e-02 1.42692238e-01
7.70505667e-01 -2.09743410e-01 -5.96791267e-01 -3.94054390e-02
-3.81166458e-01 3.48214924e-01 9.11123097e-01 4.31640595e-01
-1.40197253e+00 1.12887514e+00 6.45626259e+00 9.42297459e-01
-1.32919645e+00 -9.84796286e-02 6.21355414e-01 -1.76130742e-01
3.74344438e-01 -4.54485238e-01 -7.80237913e-01 6.93140209e-01
1.01383126e+00 2.19182387e-01 6.81243896e-01 7.86741734e-01
1.60387442e-01 4.15002555e-02 -8.83481085e-01 7.72442460e-01
-2.35219553e-01 -1.55772376e+00 2.11747751e-01 2.92478919e-01
7.28701651e-01 6.92922533e-01 1.15294494e-01 8.91876072e-02
7.59060979e-01 -1.09073329e+00 7.84218311e-01 3.49082947e-01
1.30657887e+00 -6.24668181e-01 8.30649257e-01 -2.10356694e-02
-1.35873592e+00 -2.10495830e-01 -3.67794514e-01 -2.71639258e-01
-3.77499431e-01 7.91509807e-01 -9.32500660e-01 6.47354126e-01
1.36779153e+00 8.85324001e-01 -9.89082932e-01 1.06076396e+00
-5.79993837e-02 7.89565921e-01 -6.81967735e-01 6.65614828e-02
1.06407635e-01 1.07158981e-01 2.60455906e-01 1.45471621e+00
3.79785448e-01 -4.67715472e-01 1.91124365e-01 5.56640506e-01
-1.41918242e-01 -1.86051890e-01 -5.86814821e-01 -3.16375084e-02
7.53051698e-01 1.18638515e+00 -5.88571250e-01 -1.24172807e-01
-5.54462150e-02 7.12218761e-01 3.61181974e-01 1.46716937e-01
-8.02740216e-01 -3.38497132e-01 1.09129846e+00 2.78337579e-02
3.84469658e-01 -7.70795569e-02 -4.17018622e-01 -8.97593975e-01
-1.62148811e-02 -7.46619344e-01 5.12592673e-01 -8.93575013e-01
-1.44647801e+00 7.53645241e-01 -3.69526684e-01 -9.88987446e-01
-1.04595907e-01 -1.13822591e+00 -3.33327979e-01 7.43400156e-01
-1.50244725e+00 -1.62525046e+00 -9.86326456e-01 3.46373141e-01
4.29947585e-01 -3.80193025e-01 1.16768479e+00 3.74054551e-01
-7.47832477e-01 4.77068633e-01 4.02122736e-01 -2.21923441e-02
7.03724563e-01 -1.45711875e+00 5.97509325e-01 6.27203703e-01
4.11434732e-02 3.57020974e-01 3.73990923e-01 -5.94267190e-01
-8.37717652e-01 -1.74200892e+00 4.95174706e-01 -6.93874538e-01
7.95046151e-01 -2.64652729e-01 -6.55796051e-01 4.85344172e-01
4.12094332e-02 -8.17456245e-02 6.74362183e-01 7.38043189e-02
-8.19376945e-01 -5.30348420e-01 -1.32645035e+00 3.53221357e-01
1.11723232e+00 -5.53608716e-01 6.44284114e-02 6.03595734e-01
7.93167770e-01 -1.01834968e-01 -1.21222425e+00 6.99450791e-01
8.74837041e-01 -8.41986120e-01 9.17854369e-01 -5.12848139e-01
4.35901552e-01 -4.80109781e-01 -7.39216983e-01 -1.35264754e+00
-5.82016945e-01 -1.02973059e-01 3.21159922e-02 1.45156455e+00
2.62779891e-01 -5.46489775e-01 6.69853151e-01 -3.56195629e-01
-1.86420023e-01 -5.51554263e-01 -7.24465668e-01 -1.13643253e+00
2.66082615e-01 -5.23639262e-01 9.50144112e-01 9.46248055e-01
-5.74422657e-01 -2.43641287e-02 -3.10295429e-02 6.09983444e-01
4.80173409e-01 -1.15889609e-01 9.28552091e-01 -1.53129745e+00
-6.56348392e-02 -6.57416165e-01 -5.25781214e-01 -3.58120799e-01
-4.34803627e-02 -8.47631156e-01 -7.83021823e-02 -1.40175962e+00
1.65037289e-01 -6.14769399e-01 -4.76837039e-01 8.25443268e-01
-9.44256932e-02 6.33575559e-01 -7.63976052e-02 2.93154091e-01
-3.72213870e-01 1.19605958e-01 6.02916777e-01 -3.46870691e-01
2.49432370e-01 -1.93599418e-01 -4.52580750e-01 5.25032222e-01
1.14163542e+00 -2.45697960e-01 9.94773675e-03 -7.53179848e-01
-2.40987733e-01 -7.19119608e-01 6.72899961e-01 -1.47321105e+00
-1.24191679e-01 -2.27955475e-01 5.68394244e-01 -4.43474710e-01
2.72179842e-01 -8.55743170e-01 7.03838825e-01 6.64398372e-01
-3.18755545e-02 -3.46288756e-02 4.48345989e-01 3.24227184e-01
3.77803259e-02 1.89204291e-01 8.02536309e-01 -1.19677283e-01
-7.62882352e-01 4.55044955e-01 -1.92014843e-01 -3.03982556e-01
7.81816185e-01 -1.99826702e-01 -7.88716733e-01 2.24497497e-01
-5.04446983e-01 -1.47768050e-01 8.16558838e-01 5.60236752e-01
-2.47230694e-01 -1.08431554e+00 -6.64688468e-01 1.67385101e-01
5.16824901e-01 1.37906954e-01 -9.85148698e-02 3.00818473e-01
-8.72825146e-01 6.59127712e-01 -5.52381158e-01 -8.35364163e-01
-1.10923862e+00 3.07372212e-01 6.44257486e-01 -2.15886518e-01
-1.36580765e-01 8.74076009e-01 4.00966108e-01 -8.85548234e-01
4.25298125e-01 -5.21838605e-01 2.63677035e-02 3.88929695e-02
5.13837159e-01 5.14329255e-01 4.76442397e-01 -3.96029413e-01
-4.41263884e-01 3.77201617e-01 1.77135602e-01 2.88688123e-01
1.62900412e+00 1.96006954e-01 5.08271642e-02 2.01271012e-01
9.16490734e-01 -3.79885793e-01 -1.48592114e+00 6.22266755e-02
4.00895774e-02 -4.07236278e-01 8.43895823e-02 -1.71799302e+00
-1.17914104e+00 8.35979819e-01 9.76032615e-01 3.62969369e-01
9.73293841e-01 -1.76432371e-01 3.34240049e-01 4.36515093e-01
4.16242421e-01 -5.89841604e-01 -1.49293855e-01 5.46288967e-01
9.16371644e-01 -1.17499328e+00 -2.72539139e-01 -4.15049374e-01
-1.15740933e-02 8.27304363e-01 8.38771045e-01 1.09905258e-01
5.40482163e-01 6.02804542e-01 2.62412220e-01 -7.40979761e-02
-7.81641960e-01 -2.22032592e-01 -5.88453449e-02 1.26360476e+00
2.22883075e-01 5.87867022e-01 3.95347029e-01 6.53940856e-01
-5.58993697e-01 5.15925996e-02 1.13053106e-01 5.06060421e-01
-3.24175566e-01 -6.88057303e-01 -2.68834800e-01 5.45449615e-01
-3.81625354e-01 -9.61953998e-02 -7.41807818e-01 7.78015673e-01
6.50689363e-01 9.82823789e-01 3.37806195e-01 -6.75673783e-01
2.77597457e-01 -2.00696513e-01 3.55286717e-01 -3.84722322e-01
-8.83485019e-01 -4.42394227e-01 1.12621278e-01 -5.58744013e-01
-2.66994029e-01 -5.63968599e-01 -5.87467551e-01 -6.70931101e-01
-4.48627442e-01 -1.64727330e-01 9.72643137e-01 6.26010358e-01
5.50351739e-01 5.81295073e-01 3.99998814e-01 -8.11971366e-01
-2.32076928e-01 -1.32676423e+00 -4.66707081e-01 3.29213440e-01
1.15217239e-01 -8.72569323e-01 -4.87859249e-01 3.72661740e-01]
|
[9.527166366577148, 2.144723415374756]
|
6087374d-547d-4197-90bd-dfb18c1216e5
|
deepremaster-temporal-source-reference
|
2009.08692
| null |
https://arxiv.org/abs/2009.08692v1
|
https://arxiv.org/pdf/2009.08692v1.pdf
|
DeepRemaster: Temporal Source-Reference Attention Networks for Comprehensive Video Enhancement
|
The remastering of vintage film comprises of a diversity of sub-tasks including super-resolution, noise removal, and contrast enhancement which aim to restore the deteriorated film medium to its original state. Additionally, due to the technical limitations of the time, most vintage film is either recorded in black and white, or has low quality colors, for which colorization becomes necessary. In this work, we propose a single framework to tackle the entire remastering task semi-interactively. Our work is based on temporal convolutional neural networks with attention mechanisms trained on videos with data-driven deterioration simulation. Our proposed source-reference attention allows the model to handle an arbitrary number of reference color images to colorize long videos without the need for segmentation while maintaining temporal consistency. Quantitative analysis shows that our framework outperforms existing approaches, and that, in contrast to existing approaches, the performance of our framework increases with longer videos and more reference color images.
|
['Edgar Simo-Serra', 'Satoshi Iizuka']
|
2020-09-18
| null | null | null | null |
['video-enhancement']
|
['computer-vision']
|
[ 3.59875262e-01 -4.57840204e-01 1.28633425e-01 8.61370713e-02
-6.40873194e-01 -6.92942560e-01 1.24343552e-01 -3.20650548e-01
-2.44438693e-01 4.68280911e-01 -1.10015951e-01 -2.00668141e-01
6.87734084e-03 -4.50984180e-01 -6.80256546e-01 -5.71938634e-01
6.01833798e-02 -2.13032097e-01 6.00880682e-01 -3.35448414e-01
2.39562348e-01 6.79385543e-01 -1.43817198e+00 3.64552945e-01
8.31431508e-01 1.00441134e+00 3.38133067e-01 1.13082004e+00
1.66123465e-01 1.01178801e+00 -5.57093143e-01 -3.67049634e-01
3.57830584e-01 -4.49805140e-01 -6.87798977e-01 4.74880010e-01
6.11796796e-01 -8.60420167e-01 -5.95125198e-01 1.10015512e+00
2.18379617e-01 3.70844066e-01 1.79232702e-01 -1.08012092e+00
-1.12588906e+00 4.02579933e-01 -8.21948230e-01 4.31672305e-01
4.97042358e-01 1.53673574e-01 6.98111296e-01 -8.65972042e-01
7.99401700e-01 1.04607356e+00 4.72434074e-01 5.81328034e-01
-1.28795099e+00 -4.73628223e-01 5.58422148e-01 3.83809596e-01
-1.23283064e+00 -4.36070025e-01 8.12202692e-01 -3.05012017e-01
5.99413276e-01 3.47386599e-01 7.61557162e-01 9.24328625e-01
8.39441195e-02 6.34482443e-01 9.53334749e-01 -1.25621349e-01
1.01081014e-01 -1.25101969e-01 -2.39257857e-01 3.48677218e-01
-1.51665568e-01 -1.35028720e-01 -5.07655084e-01 3.79414260e-01
1.12346709e+00 1.21406920e-01 -6.79693103e-01 -3.30339164e-01
-1.02950954e+00 8.07508156e-02 2.69100040e-01 3.83533925e-01
-2.92172521e-01 1.24351971e-01 3.03782463e-01 3.11992496e-01
6.55583322e-01 2.98152924e-01 -2.82295018e-01 -3.02074343e-01
-1.45863581e+00 3.31615359e-02 3.34850669e-01 1.10459960e+00
3.34604472e-01 3.05538744e-01 -7.38341659e-02 7.23132908e-01
-1.66680187e-01 2.04886287e-01 1.67071912e-02 -1.45156300e+00
4.96140093e-01 2.29656965e-01 5.85072160e-01 -9.90173995e-01
-2.18378335e-01 -1.58761591e-01 -7.87578166e-01 6.08596742e-01
4.20361608e-01 -7.21204877e-02 -9.81217861e-01 1.36526752e+00
9.64971557e-02 1.26499832e-01 -1.91927344e-01 1.35801268e+00
5.73923945e-01 8.63115311e-01 -1.40754119e-01 -4.37489629e-01
9.87310052e-01 -1.09740043e+00 -1.12739325e+00 6.78661317e-02
-2.83610165e-01 -8.06473732e-01 1.23840809e+00 8.97896230e-01
-1.63428438e+00 -7.23782837e-01 -1.25026441e+00 -3.82451415e-01
-1.60339922e-01 4.69992645e-02 2.88696706e-01 4.22449738e-01
-1.36261368e+00 7.90708065e-01 -8.74127746e-01 -7.54547119e-02
3.31861347e-01 2.39790186e-01 -3.03048640e-01 -2.74326384e-01
-1.01259077e+00 5.21801353e-01 2.33621255e-01 4.59468514e-01
-1.00440490e+00 -7.00361073e-01 -5.04863799e-01 1.20319329e-01
5.04970253e-01 -3.39525431e-01 1.09107268e+00 -1.55353332e+00
-1.72319877e+00 5.60144842e-01 8.19566473e-02 -7.58405402e-02
9.64331269e-01 -5.19370496e-01 -6.31066144e-01 6.12337053e-01
-3.83726656e-01 3.71784061e-01 1.04367971e+00 -1.56004024e+00
-6.52505994e-01 -5.45257032e-02 5.43399513e-01 2.64864117e-01
-3.81978720e-01 3.11923653e-01 -1.42569482e+00 -6.76560819e-01
2.73639057e-02 -7.06028342e-01 -1.31025225e-01 2.59480596e-01
-3.23626816e-01 4.38960999e-01 1.09871304e+00 -1.07888317e+00
1.35675931e+00 -2.26754594e+00 5.20855010e-01 -1.53053477e-01
4.52351511e-01 2.16020808e-01 -2.62806565e-01 2.24578261e-01
-1.06230445e-01 1.20532751e-01 -1.53392583e-01 -2.57031113e-01
-3.75754774e-01 -2.17379585e-01 -5.90638556e-02 3.44870239e-01
1.96749270e-01 5.92889011e-01 -8.54678452e-01 -4.36491400e-01
3.12972814e-01 6.92330837e-01 -4.30936873e-01 2.89994806e-01
-2.00752735e-01 5.12957156e-01 4.87644412e-02 7.79599249e-01
1.06515563e+00 -1.51958942e-01 2.93125689e-01 -4.74966466e-01
-2.07876474e-01 -2.79208928e-01 -1.17555177e+00 1.80954492e+00
-3.34212333e-01 9.65434194e-01 4.68067616e-01 -4.04215723e-01
4.77704793e-01 2.23919541e-01 5.98594546e-01 -8.01263750e-01
1.28339892e-02 -1.04962319e-01 -2.55647093e-01 -6.91620648e-01
1.03420150e+00 3.41436565e-01 2.61127859e-01 3.48497182e-01
-3.24289709e-01 5.44155249e-04 5.05470097e-01 3.76281291e-01
9.03085947e-01 5.05769491e-01 -3.50232363e-01 1.66170195e-01
3.27020139e-01 -1.36400804e-01 5.58106542e-01 4.44627613e-01
-1.50777936e-01 1.00561047e+00 5.35376132e-01 -2.06865251e-01
-1.32719147e+00 -1.08071148e+00 1.66365981e-01 1.10562205e+00
6.27773583e-01 -9.61906910e-02 -8.97154152e-01 -2.74099678e-01
-4.35518682e-01 4.53414112e-01 -6.09122038e-01 1.04101583e-01
-7.94159174e-01 -4.67968196e-01 3.63005400e-02 6.70562148e-01
5.63907444e-01 -9.43103492e-01 -6.84329808e-01 1.98560938e-01
-2.88398594e-01 -1.25591683e+00 -8.70872259e-01 -1.54209539e-01
-8.32985699e-01 -1.10653555e+00 -1.07485139e+00 -6.44960999e-01
6.39683127e-01 5.36984801e-01 1.00130129e+00 1.77252725e-01
-2.45160714e-01 4.27491188e-01 -4.96489018e-01 4.18385029e-01
-3.39816272e-01 -1.92147911e-01 -3.99094343e-01 2.27685809e-01
-2.26396635e-01 -6.48770809e-01 -9.28425670e-01 2.36021727e-01
-1.42205536e+00 4.17044103e-01 4.11421031e-01 5.58985770e-01
5.83796859e-01 2.19541550e-01 2.06243649e-01 -8.05342674e-01
5.44997394e-01 -2.35704005e-01 -5.96335471e-01 3.52978915e-01
-2.92146236e-01 -4.03599620e-01 7.84818172e-01 -7.22736657e-01
-1.37384033e+00 -1.46594644e-01 1.61941260e-01 -7.94717729e-01
6.55768216e-02 9.44580510e-02 -2.42947012e-01 -2.18596477e-02
2.48808935e-01 2.15260670e-01 -2.34980568e-01 -5.39630651e-01
4.60036576e-01 4.62685913e-01 8.38669360e-01 -1.75912961e-01
7.77209699e-01 5.02448142e-01 -4.15789634e-01 -5.95053017e-01
-5.00636220e-01 2.77665090e-02 -8.83759737e-01 -6.08500659e-01
9.60961282e-01 -7.70249963e-01 -7.92234540e-01 6.59450352e-01
-9.18055236e-01 -7.92086005e-01 -7.44647905e-02 8.84230733e-02
-3.91881794e-01 6.08521640e-01 -1.14052701e+00 -7.79898703e-01
-1.86589956e-01 -1.16848278e+00 8.03271830e-01 3.39853495e-01
2.68114001e-01 -6.48002386e-01 -3.84109020e-01 4.10446167e-01
6.26194596e-01 3.40950400e-01 7.79400468e-01 2.80931324e-01
-9.53929543e-01 -5.78730218e-02 -5.40517449e-01 5.39182365e-01
1.54298395e-01 4.94919479e-01 -7.52635777e-01 -4.02623802e-01
-1.71928748e-01 -7.77305216e-02 8.34366024e-01 3.58470917e-01
1.47712433e+00 -1.26279682e-01 1.80893973e-01 7.55648315e-01
1.45998168e+00 5.41405022e-01 8.44707072e-01 4.77934718e-01
7.60250270e-01 3.58891606e-01 7.74284601e-01 4.29934919e-01
2.69942552e-01 6.57314003e-01 6.47313237e-01 -5.45877814e-01
-3.95944625e-01 7.23649412e-02 4.58718061e-01 6.59251928e-01
-3.71230602e-01 -5.04089534e-01 -4.33789611e-01 6.42151535e-01
-1.69681978e+00 -9.40923274e-01 -1.51694864e-01 2.05882454e+00
6.49185598e-01 -7.16466270e-03 1.87990591e-01 1.01693101e-01
8.20295453e-01 2.08092615e-01 -7.84646749e-01 -3.66268992e-01
-1.97792500e-01 -2.49810353e-01 4.31982756e-01 3.35388571e-01
-9.47809637e-01 7.22149670e-01 6.86083364e+00 5.56328058e-01
-1.21948624e+00 6.03035055e-02 7.44785011e-01 -5.02800584e-01
-3.01618963e-01 -2.02152848e-01 -1.26602590e-01 5.80370843e-01
5.77635586e-01 1.47272289e-01 8.42634082e-01 3.61184359e-01
5.56648433e-01 -2.58294195e-01 -9.26862359e-01 9.38750565e-01
2.08278954e-01 -1.14891958e+00 -1.38189033e-01 -2.93066055e-01
8.79066229e-01 -4.27710712e-01 4.35978383e-01 -1.27251104e-01
1.38168871e-01 -8.24248493e-01 1.20585823e+00 6.99558914e-01
1.19368064e+00 -7.55710661e-01 3.14641237e-01 -2.65721112e-01
-1.12130308e+00 -2.92263806e-01 -1.28613368e-01 3.66422236e-01
4.03952271e-01 2.51557708e-01 -5.15848808e-02 6.93781972e-01
1.05411375e+00 6.83952868e-01 -6.79495096e-01 9.43844616e-01
-5.49165308e-02 3.09664607e-01 8.35308880e-02 5.30683994e-01
-5.26958331e-02 -2.73595721e-01 3.95043433e-01 1.17067730e+00
3.85849357e-01 1.25804573e-01 3.42603624e-02 8.68259549e-01
-1.27429232e-01 -1.52140543e-01 -7.39248190e-03 -1.36229452e-02
1.50751993e-01 1.26557422e+00 -7.64397562e-01 -3.63153309e-01
-5.64496756e-01 1.65033114e+00 6.23528473e-02 8.93952549e-01
-1.01963162e+00 -4.37299818e-01 3.78215790e-01 3.26067954e-01
4.43843603e-01 -3.70205849e-01 -9.16751698e-02 -1.23377204e+00
1.97223365e-01 -8.55350673e-01 2.12578237e-01 -1.35101926e+00
-9.84036565e-01 9.65244889e-01 -8.45864937e-02 -1.39311349e+00
6.85901344e-02 -3.76962245e-01 -4.60976005e-01 5.45566201e-01
-1.58361661e+00 -1.09914184e+00 -6.40447855e-01 7.62968183e-01
8.55360806e-01 2.51525551e-01 1.60790831e-01 6.61957264e-01
-7.19860494e-01 4.50630665e-01 2.60444909e-01 -1.91436568e-03
8.04904997e-01 -1.27414536e+00 2.47196436e-01 1.35066402e+00
-2.63954490e-01 2.22413912e-01 8.95150423e-01 -5.52183270e-01
-1.39560568e+00 -9.92103159e-01 7.92529136e-02 -1.79524925e-02
5.52051127e-01 -3.57925653e-01 -1.10627961e+00 5.03878891e-01
4.72473115e-01 7.55939260e-02 2.03359738e-01 -3.05573314e-01
-2.65399903e-01 -3.24916542e-01 -9.64250684e-01 7.71762192e-01
8.91326249e-01 -3.96646321e-01 1.31932676e-01 4.26496901e-02
8.12568605e-01 -7.00712919e-01 -9.45997834e-01 3.71752423e-03
6.12236679e-01 -1.15380383e+00 8.86844516e-01 -1.70036376e-01
7.24738121e-01 -6.94461048e-01 8.64021108e-02 -8.95376086e-01
-2.99299687e-01 -8.27511251e-01 -1.68135107e-01 1.29026973e+00
2.38509655e-01 -1.05490915e-01 5.51090062e-01 9.58704412e-01
-9.37253833e-02 -3.42960358e-01 -7.31359065e-01 -4.93983299e-01
-1.83483481e-01 -3.80404234e-01 4.38688904e-01 6.91792727e-01
-3.93762648e-01 -1.61549985e-01 -8.63134921e-01 3.05533320e-01
5.41489005e-01 2.36437500e-01 5.00981808e-01 -7.04543352e-01
-3.83281589e-01 -3.28734457e-01 -8.06834474e-02 -1.02520585e+00
-3.51899475e-01 -2.35927820e-01 3.75149325e-02 -1.50032449e+00
2.97987998e-01 -1.81079879e-01 -4.90548700e-01 1.76742435e-01
-1.96674168e-01 6.75420582e-01 5.49643159e-01 1.04986779e-01
-9.68296945e-01 3.33272099e-01 1.56390989e+00 -1.64675325e-01
-4.81122941e-01 -2.85482377e-01 -5.93313575e-01 4.89230722e-01
5.69420576e-01 -6.58165812e-02 -3.43476415e-01 -7.57953227e-01
2.41826549e-01 4.83660012e-01 3.79639059e-01 -9.28698659e-01
2.25357413e-01 -1.99490041e-01 7.19016969e-01 -5.82050025e-01
4.51090127e-01 -9.18711782e-01 4.34444547e-01 1.85060687e-03
-3.49400818e-01 3.31878573e-01 3.00810009e-01 5.92443466e-01
-2.08282471e-01 -3.19033414e-02 9.95987117e-01 -2.08191741e-02
-7.64154792e-01 2.79491365e-01 -5.90740144e-01 -2.11217880e-01
1.10761786e+00 -4.72160339e-01 -1.96342066e-01 -5.42627752e-01
-9.31907237e-01 1.09994687e-01 9.36773777e-01 4.62607980e-01
7.65641570e-01 -1.24153972e+00 -5.68163037e-01 -3.77682410e-02
-3.15566629e-01 -7.91058317e-02 1.03537822e+00 8.07758987e-01
-8.83181214e-01 -1.20193690e-01 -3.81829143e-01 -3.63715112e-01
-1.27153468e+00 9.62292612e-01 3.73814315e-01 -9.88327861e-02
-9.79033113e-01 4.87215698e-01 2.25308672e-01 4.13789034e-01
2.93716967e-01 -3.72280389e-01 -2.39433676e-01 -5.24095297e-02
7.79264927e-01 5.67974508e-01 -1.90578890e-03 -5.08953393e-01
-2.10459717e-02 6.28130496e-01 -2.72401363e-01 -1.77110404e-01
1.41940117e+00 -7.56498218e-01 -2.26175301e-02 3.41606528e-01
8.42688143e-01 -8.20051064e-04 -2.03298998e+00 -1.56436302e-02
-3.36993486e-01 -8.23255658e-01 1.77321464e-01 -9.83576238e-01
-1.60459661e+00 7.89130092e-01 7.65049636e-01 2.02592388e-01
1.78105342e+00 -4.13008273e-01 8.55826557e-01 -1.72571197e-01
4.49755117e-02 -1.35587466e+00 2.63685316e-01 7.33221620e-02
9.93834019e-01 -1.02307928e+00 5.95010892e-02 -2.73442626e-01
-8.06162298e-01 1.31309867e+00 7.85148203e-01 -1.28664479e-01
1.44613683e-01 4.16520089e-01 1.54799223e-01 5.32371625e-02
-7.64693081e-01 3.32421390e-04 3.32606018e-01 5.69901288e-01
2.96305358e-01 -3.92226934e-01 3.39841470e-02 4.61002320e-01
3.67666304e-01 5.36862016e-02 1.02378345e+00 6.38907254e-01
-2.00472668e-01 -8.35162461e-01 -4.39162970e-01 1.02979392e-01
-7.50958562e-01 -7.04356804e-02 -4.15066816e-02 7.39496946e-01
-2.38124244e-02 1.14910781e+00 5.74416965e-02 -2.39949360e-01
3.62035841e-01 -3.58984530e-01 4.86541361e-01 -6.40231371e-02
-5.43126047e-01 4.86126155e-01 -2.17396304e-01 -8.93558443e-01
-5.66682518e-01 -3.60124320e-01 -9.63422596e-01 -3.83111656e-01
-5.63476756e-02 -2.88728207e-01 4.41534817e-01 4.73269403e-01
2.40646437e-01 1.00715971e+00 5.55082798e-01 -1.15823877e+00
1.35866210e-01 -6.73549712e-01 -8.53441894e-01 5.31711698e-01
6.34056032e-01 -2.71629572e-01 -1.39704540e-01 4.71932501e-01]
|
[10.991805076599121, -1.2680683135986328]
|
b365e2c3-d750-4c47-b6a5-148d19e614e3
|
stylestegan-leak-free-style-transfer-based-on
|
2307.00225
| null |
https://arxiv.org/abs/2307.00225v1
|
https://arxiv.org/pdf/2307.00225v1.pdf
|
StyleStegan: Leak-free Style Transfer Based on Feature Steganography
|
In modern social networks, existing style transfer methods suffer from a serious content leakage issue, which hampers the ability to achieve serial and reversible stylization, thereby hindering the further propagation of stylized images in social networks. To address this problem, we propose a leak-free style transfer method based on feature steganography. Our method consists of two main components: a style transfer method that accomplishes artistic stylization on the original image and an image steganography method that embeds content feature secrets on the stylized image. The main contributions of our work are as follows: 1) We identify and explain the phenomenon of content leakage and its underlying causes, which arise from content inconsistencies between the original image and its subsequent stylized image. 2) We design a neural flow model for achieving loss-free and biased-free style transfer. 3) We introduce steganography to hide content feature information on the stylized image and control the subsequent usage rights. 4) We conduct comprehensive experimental validation using publicly available datasets MS-COCO and Wikiart. The results demonstrate that StyleStegan successfully mitigates the content leakage issue in serial and reversible style transfer tasks. The SSIM performance metrics for these tasks are 14.98% and 7.28% higher, respectively, compared to a suboptimal baseline model.
|
['Xinpeng Zhang', 'Zhenxing Qian', 'Qichao Ying', 'Bingshan Liu', 'Xiujian Liang']
|
2023-07-01
| null | null | null | null |
['style-transfer', 'image-steganography']
|
['computer-vision', 'computer-vision']
|
[ 7.45456398e-01 -7.61284605e-02 -9.75777954e-02 8.01943764e-02
-1.28835320e-01 -6.92019522e-01 8.10410976e-01 -6.57544672e-01
-2.35556185e-01 8.86250794e-01 2.71394819e-01 -2.98989713e-01
6.49702787e-01 -9.44752872e-01 -8.82529438e-01 -5.84895492e-01
2.13826194e-01 -2.71490991e-01 1.65610269e-01 -3.92283916e-01
4.00424451e-01 2.73698866e-01 -9.60781395e-01 2.35661566e-01
9.85976338e-01 8.86743486e-01 -1.24922232e-03 6.44153893e-01
-2.74508059e-01 9.57497835e-01 -9.12762046e-01 -7.15383053e-01
3.63614351e-01 -9.95481789e-01 -7.94103146e-01 3.01022381e-01
2.66547501e-01 -6.55125976e-01 -7.39204466e-01 1.31465936e+00
4.93309975e-01 -3.13820213e-01 6.68368578e-01 -1.52378130e+00
-1.39653957e+00 5.41918457e-01 -7.62714982e-01 -9.49087143e-02
3.63161005e-02 4.65109140e-01 6.23923063e-01 -4.64616954e-01
8.86928737e-01 1.45104599e+00 3.64015818e-01 8.89291883e-01
-1.23702347e+00 -1.21369815e+00 -2.22659439e-01 -1.30096152e-01
-1.00737679e+00 -4.27983552e-01 9.52157676e-01 -1.44915715e-01
2.40355015e-01 2.57670552e-01 8.45267713e-01 1.38148618e+00
4.94355112e-01 9.95971203e-01 1.57813156e+00 -2.10774109e-01
-1.65709585e-01 4.12456036e-01 -6.25188112e-01 8.71986866e-01
2.50696361e-01 1.66214049e-01 -6.94114447e-01 1.51271317e-02
1.02062023e+00 -1.84648544e-01 -4.07283127e-01 -1.61740646e-01
-1.07459188e+00 9.24881637e-01 5.75262964e-01 1.29953012e-01
6.36761412e-02 4.21209663e-01 5.19426882e-01 8.09654117e-01
6.96108937e-01 3.05231214e-01 1.64332449e-01 7.99020454e-02
-9.52968061e-01 -7.58242384e-02 8.32717001e-01 1.17155910e+00
7.16419637e-01 2.55400091e-01 -2.51233131e-01 3.92260820e-01
2.74105579e-01 1.03839207e+00 3.88414234e-01 -7.63848364e-01
5.45461893e-01 3.25964719e-01 -2.09673718e-01 -1.40718496e+00
4.46111172e-01 -1.84856623e-01 -1.09366417e+00 1.69960007e-01
1.79487303e-01 -3.96777689e-01 -7.38830388e-01 1.68657160e+00
-1.51686057e-01 3.05747002e-01 -8.26734230e-02 5.64655840e-01
7.88943946e-01 7.86657631e-01 -1.65889468e-02 5.27777448e-02
1.28580451e+00 -1.03975260e+00 -1.05221546e+00 -1.28383279e-01
4.68363911e-01 -9.93164062e-01 1.17978525e+00 -2.55666804e-02
-1.13707709e+00 -3.04024160e-01 -1.36681044e+00 7.41126016e-03
-3.06381822e-01 -2.71690279e-01 3.10020536e-01 1.05183363e+00
-1.11879683e+00 6.44475400e-01 -2.56274492e-01 -6.75660074e-02
1.03045535e+00 2.24516809e-01 -3.26101661e-01 1.87284470e-01
-1.53010035e+00 4.79010344e-01 1.76241308e-01 -2.15305507e-01
-7.60154903e-01 -8.88707399e-01 -8.14464271e-01 -7.35225901e-02
1.76151231e-01 -6.93689108e-01 7.81247616e-01 -1.60198319e+00
-1.79046547e+00 1.15516508e+00 -5.60089909e-02 -2.94134080e-01
9.70723808e-01 -1.58850417e-01 -4.73938286e-01 1.76301628e-01
7.15972409e-02 7.48675764e-01 1.44473994e+00 -1.46812141e+00
-5.25209010e-01 1.59352720e-01 -2.57262558e-01 1.31439179e-01
-7.55677879e-01 -1.12204134e-01 -5.15989721e-01 -1.09216893e+00
-4.65262383e-01 -1.04228032e+00 2.01545894e-01 2.83618838e-01
-7.53943920e-01 5.51647961e-01 1.32594407e+00 -8.14478397e-01
1.16227436e+00 -2.19353962e+00 1.91793427e-01 2.45479748e-01
5.52313626e-01 6.55342937e-01 -2.66533166e-01 4.06757325e-01
1.59789830e-01 5.48653960e-01 -2.73785412e-01 -3.83079618e-01
-2.04237610e-01 4.77784388e-02 -7.40298867e-01 4.25411910e-01
2.04486609e-01 1.37442851e+00 -8.62398922e-01 -4.91434276e-01
-5.37581034e-02 5.11888802e-01 -4.10305291e-01 2.61240602e-01
3.75080556e-02 6.94818199e-01 -3.47584784e-01 3.77244800e-01
1.01501656e+00 -1.66567117e-01 6.37284592e-02 -7.24970251e-02
1.80901736e-01 -3.94469965e-03 -7.32649505e-01 1.38043487e+00
-4.33938980e-01 9.11123991e-01 -1.09946571e-01 -2.58939087e-01
8.95271122e-01 1.81787595e-01 1.15739055e-01 -7.16891050e-01
3.73943835e-01 2.51745701e-01 -5.71345806e-01 -2.60182112e-01
7.01400876e-01 -3.30506444e-01 -2.02504650e-01 8.07435632e-01
-9.49643180e-02 -3.31085682e-01 -2.56101876e-01 4.27129239e-01
9.04336751e-01 -1.16342353e-02 -3.39208730e-02 -2.03616306e-01
8.50254655e-01 -6.18319333e-01 3.40534955e-01 7.05780149e-01
-2.96869248e-01 3.45722109e-01 9.46554661e-01 -1.47958949e-01
-1.30044246e+00 -9.30061936e-01 4.33350652e-01 6.53170466e-01
4.54151750e-01 -3.05143267e-01 -1.07349205e+00 -1.01255262e+00
-2.68243868e-02 5.35550117e-01 -5.99237561e-01 -4.66219842e-01
-6.81374907e-01 -2.88335174e-01 1.11016345e+00 1.81105863e-02
1.39036584e+00 -1.41427886e+00 -1.01895258e-01 -9.06707719e-02
-5.50409794e-01 -9.99778628e-01 -1.19136417e+00 -7.08555400e-01
-7.58406699e-01 -8.10167193e-01 -9.77942765e-01 -7.06679404e-01
7.91306913e-01 5.26463985e-01 8.33837032e-01 4.03435588e-01
1.42701566e-02 2.48156086e-01 -1.12051517e-01 -4.55492496e-01
-8.44807148e-01 1.74507841e-01 -2.60464907e-01 3.31371367e-01
-4.77807708e-02 -4.65063006e-01 -6.98116660e-01 4.20009226e-01
-1.37331140e+00 1.82266057e-01 5.02394319e-01 9.11917388e-01
1.55762091e-01 3.24713029e-02 4.82570350e-01 -1.22751915e+00
7.90077686e-01 -3.37274998e-01 -4.88186717e-01 1.81242689e-01
-7.64065385e-01 1.62833750e-01 6.04740918e-01 -4.59874064e-01
-1.25124598e+00 -3.60197365e-01 9.85453427e-02 -3.49618286e-01
3.19091529e-01 -1.05187096e-01 -3.68203372e-01 -5.23842812e-01
3.08425695e-01 6.86495066e-01 5.02183676e-01 -9.10617709e-02
5.02696097e-01 7.24238336e-01 3.11180264e-01 -2.72215992e-01
1.40238309e+00 9.05406058e-01 -2.91249827e-02 -8.44179809e-01
-4.59565878e-01 2.01514110e-01 -2.92467147e-01 -2.35708013e-01
6.27146840e-01 -8.62203062e-01 -6.90563560e-01 1.20524704e+00
-1.15234578e+00 -5.10484636e-01 -3.73701125e-01 -2.13711828e-01
-6.24502420e-01 8.08036029e-01 -8.56369972e-01 -3.38876307e-01
-6.75964475e-01 -1.05992806e+00 6.70783699e-01 1.72688231e-01
-1.16800461e-02 -1.18684006e+00 -1.97204411e-01 3.26517314e-01
7.06345320e-01 4.31913674e-01 7.01403201e-01 8.96114111e-02
-8.57525408e-01 1.06896393e-01 -6.18237078e-01 5.57402372e-01
4.42604423e-01 -1.94388792e-01 -8.23047996e-01 -6.43597305e-01
5.55723384e-02 -2.52847254e-01 1.04643106e+00 -8.07942078e-02
1.02618122e+00 -6.64537370e-01 -1.76582962e-01 1.01088548e+00
1.29960811e+00 -3.21205729e-03 1.16497874e+00 3.99436831e-01
9.56652105e-01 5.93097270e-01 1.29233718e-01 1.43830940e-01
9.98871922e-02 2.06730455e-01 3.46026748e-01 -2.68536985e-01
-5.21610796e-01 -8.38941634e-01 6.32972956e-01 7.91875124e-01
1.40097022e-01 -7.03212678e-01 -1.07535765e-01 4.45787638e-01
-1.49416423e+00 -8.37997913e-01 -4.03681137e-02 1.83492923e+00
1.00623012e+00 2.11582437e-01 -5.73717840e-02 4.77389172e-02
9.48012888e-01 6.98789001e-01 -5.07225335e-01 -5.01525760e-01
-3.51730973e-01 8.15260932e-02 9.45425570e-01 4.54131037e-01
-9.48405087e-01 1.35144734e+00 5.82594681e+00 1.04178548e+00
-1.28639889e+00 1.13248520e-01 7.33917236e-01 4.24267612e-02
-6.30207658e-01 -2.65685588e-01 -5.17205238e-01 9.35684502e-01
6.88874424e-01 -2.40049064e-01 7.33725131e-01 1.69517949e-01
4.55616228e-02 2.23759845e-01 -5.55891395e-01 8.86675954e-01
2.30721772e-01 -1.55883670e+00 5.16041398e-01 2.54880250e-01
9.94540691e-01 -5.32105625e-01 7.55418181e-01 -1.03288069e-01
2.40765750e-01 -9.50913131e-01 7.84662902e-01 3.29105079e-01
1.53380549e+00 -9.36278403e-01 4.02435213e-01 -1.24866828e-01
-9.08646524e-01 2.01420784e-01 -1.33358046e-01 2.06721142e-01
1.45431295e-01 4.20080155e-01 -3.83938044e-01 5.07297873e-01
3.31896424e-01 9.72765267e-01 -4.73195881e-01 4.06845719e-01
-7.47456312e-01 8.73552680e-01 1.96888894e-01 3.95817757e-02
4.15985078e-01 -1.02153115e-01 8.31194222e-01 1.27329433e+00
3.59842092e-01 -3.71675670e-01 -5.61837196e-01 1.22282493e+00
-8.11970413e-01 -3.65668684e-02 -8.69746864e-01 -1.48327291e-01
3.96336079e-01 9.56798673e-01 -7.11524069e-01 -2.77749628e-01
-2.14400813e-01 1.73749208e+00 -8.52484852e-02 4.67238069e-01
-8.04113448e-01 -8.13554525e-01 6.83632493e-01 1.32194608e-01
3.90206724e-01 -5.88724250e-03 -4.77004737e-01 -1.42040515e+00
-7.35812262e-02 -1.00595832e+00 -6.79334700e-02 -5.34024954e-01
-1.20003784e+00 3.61397684e-01 -3.97477835e-01 -1.11141002e+00
3.11820954e-01 -2.06795350e-01 -8.42864513e-01 1.06643963e+00
-1.94457030e+00 -1.34990633e+00 -1.70337483e-01 6.68096423e-01
2.98586994e-01 -4.48265731e-01 5.25699317e-01 2.69049019e-01
-2.57643223e-01 9.80186641e-01 2.18132451e-01 3.09005290e-01
1.08669436e+00 -1.01225567e+00 9.16113138e-01 9.60923195e-01
-3.40392560e-01 3.68649513e-01 5.50985336e-01 -1.07850420e+00
-1.23295903e+00 -1.27550006e+00 9.69677210e-01 -2.79041678e-01
6.94998503e-01 -5.33645332e-01 -8.08031619e-01 7.04285979e-01
6.39173627e-01 -3.24787617e-01 5.93076646e-01 -8.29094291e-01
-4.67398137e-01 6.64465874e-02 -1.44976711e+00 9.34781015e-01
1.26814604e+00 -7.15589702e-01 -1.66014984e-01 -7.48784095e-02
9.37105358e-01 -1.09940805e-01 -4.58547920e-01 -2.68693835e-01
6.07524753e-01 -9.68073130e-01 8.59602451e-01 -3.42819154e-01
1.01573408e+00 -1.10097021e-01 2.50319123e-01 -1.45632088e+00
-1.99295580e-01 -1.36362648e+00 -9.34589356e-02 1.47087145e+00
3.20502557e-02 -8.40687394e-01 8.19536626e-01 1.37650698e-01
5.30790269e-01 -1.67338029e-01 -5.71003020e-01 -7.41956294e-01
2.28996605e-01 7.14670643e-02 8.79425883e-01 1.04397810e+00
-4.06963140e-01 1.61013588e-01 -1.13211977e+00 -2.72318184e-01
1.01117730e+00 -2.41879031e-01 7.84962177e-01 -7.03485250e-01
8.84214044e-02 -3.39699060e-01 -1.68566644e-01 -1.11822772e+00
3.53150964e-01 -9.56560791e-01 -3.88521433e-01 -8.44132066e-01
1.55123591e-01 -3.05617362e-01 -1.13390259e-01 6.57046959e-02
-1.81527525e-01 7.77779102e-01 5.63194931e-01 4.18617964e-01
-3.16669822e-01 6.79029047e-01 2.11266232e+00 -2.91991949e-01
-4.64805849e-02 -1.50429249e-01 -1.07226968e+00 4.11841303e-01
8.66887093e-01 -5.30522287e-01 -4.90490645e-01 -3.77114505e-01
3.27580482e-01 -1.09493844e-01 4.37246561e-01 -6.54354572e-01
-1.75639600e-01 -7.99164362e-03 3.06100190e-01 -1.12975180e-01
-2.58056596e-02 -7.31864214e-01 -1.23530328e-01 9.31907356e-01
-4.39599812e-01 -2.93024808e-01 5.57405725e-02 8.81973565e-01
8.44253823e-02 3.03466488e-02 1.06803024e+00 -2.89310720e-02
-5.40028751e-01 3.84755552e-01 -5.94982564e-01 4.62779738e-02
9.59941506e-01 -1.78729624e-01 -3.46566349e-01 -8.03810239e-01
-3.03684384e-01 5.61833940e-02 6.86452329e-01 6.60645008e-01
7.88307965e-01 -1.70673847e+00 -7.66754091e-01 6.48139954e-01
-1.47638187e-01 -5.59550643e-01 1.62467137e-01 3.86577725e-01
-8.36287916e-01 1.94246456e-01 -6.52403057e-01 -2.58664880e-02
-1.34703243e+00 4.74921316e-01 2.02863857e-01 -3.19626540e-01
-5.34704745e-01 7.81150103e-01 3.84397566e-01 -2.18394503e-01
-5.03430963e-02 1.80083692e-01 -3.89511511e-02 -6.15360476e-02
5.88150442e-01 6.22705698e-01 -5.78482747e-01 -7.56365955e-01
1.42103955e-01 4.77763444e-01 -2.08726943e-01 -5.01854494e-02
9.54328716e-01 -7.20422208e-01 -4.50688481e-01 1.71802528e-02
1.51156390e+00 4.25930992e-02 -1.35210562e+00 -6.92174971e-01
-4.82229978e-01 -8.75367582e-01 -4.64479625e-02 -5.86929679e-01
-1.39533615e+00 8.59008610e-01 4.79928702e-01 1.40349120e-01
1.03072476e+00 -4.56793845e-01 1.75740814e+00 -2.83777853e-03
1.26598850e-01 -1.04782987e+00 4.32381690e-01 3.88905436e-01
8.39850962e-01 -1.14446878e+00 -1.52804151e-01 -4.55773294e-01
-7.68679798e-01 8.69413555e-01 2.54285097e-01 -2.85869867e-01
6.12512708e-01 5.67082502e-02 6.11372255e-02 -4.75970916e-02
-1.89945444e-01 3.78726572e-01 2.21507773e-01 6.30459368e-01
-3.56393047e-02 -4.59062010e-02 -1.71229526e-01 1.61230743e-01
-2.23457426e-01 8.50205719e-02 7.09340930e-01 8.68790627e-01
-1.17854103e-01 -1.25585842e+00 -3.86999071e-01 2.83977985e-01
-8.59824538e-01 -1.55275270e-01 -7.30316877e-01 4.85002428e-01
-6.04741536e-02 9.08868551e-01 -9.77917090e-02 -5.80373704e-01
4.68878821e-02 -1.50221944e-01 3.09711039e-01 -1.61538273e-01
-6.85264349e-01 -4.47336696e-02 -1.84814572e-01 -4.44553226e-01
-3.36670995e-01 -2.70750135e-01 -7.56353021e-01 -1.23514044e+00
-2.21806899e-01 -1.49986207e-01 4.16714340e-01 6.96278334e-01
3.39277744e-01 5.86092174e-01 1.02127266e+00 -6.20212436e-01
-3.01325649e-01 -5.30928433e-01 -8.53084743e-01 6.63135946e-01
5.29433668e-01 -1.81541100e-01 -5.04283905e-01 3.16471100e-01]
|
[11.647695541381836, -0.5964611172676086]
|
df654152-f1fd-47b2-ab16-a96eee956cf3
|
quantifying-the-lidar-sim-to-real-domain
|
2303.01899
| null |
https://arxiv.org/abs/2303.01899v1
|
https://arxiv.org/pdf/2303.01899v1.pdf
|
Quantifying the LiDAR Sim-to-Real Domain Shift: A Detailed Investigation Using Object Detectors and Analyzing Point Clouds at Target-Level
|
LiDAR object detection algorithms based on neural networks for autonomous driving require large amounts of data for training, validation, and testing. As real-world data collection and labeling are time-consuming and expensive, simulation-based synthetic data generation is a viable alternative. However, using simulated data for the training of neural networks leads to a domain shift of training and testing data due to differences in scenes, scenarios, and distributions. In this work, we quantify the sim-to-real domain shift by means of LiDAR object detectors trained with a new scenario-identical real-world and simulated dataset. In addition, we answer the questions of how well the simulated data resembles the real-world data and how well object detectors trained on simulated data perform on real-world data. Further, we analyze point clouds at the target-level by comparing real-world and simulated point clouds within the 3D bounding boxes of the targets. Our experiments show that a significant sim-to-real domain shift exists even for our scenario-identical datasets. This domain shift amounts to an average precision reduction of around 14 % for object detectors trained with simulated data. Additional experiments reveal that this domain shift can be lowered by introducing a simple noise model in simulation. We further show that a simple downsampling method to model real-world physics does not influence the performance of the object detectors.
|
['Markus Lienkamp', 'Esteban Rivera', 'Luca Scalerandi', 'Sebastian Huch']
|
2023-03-03
| null | null | null | null |
['synthetic-data-generation', 'synthetic-data-generation']
|
['medical', 'miscellaneous']
|
[ 1.81299806e-01 -6.08306453e-02 3.58272135e-01 -5.20653725e-01
-4.82874781e-01 -4.68746483e-01 5.59822798e-01 3.93567055e-01
-8.26349318e-01 6.48570716e-01 -6.31537259e-01 -3.48102838e-01
1.50395244e-01 -1.06641769e+00 -1.14362204e+00 -3.91747653e-01
-1.83505014e-01 9.72469330e-01 7.86595762e-01 -1.23786308e-01
-1.16111271e-01 1.03532302e+00 -2.09472156e+00 -1.17785729e-01
7.78875053e-01 7.20611870e-01 2.83176541e-01 5.23332834e-01
-1.73275247e-01 -1.29024342e-01 -6.93737209e-01 1.29270274e-02
8.08150113e-01 1.03496037e-01 6.51103556e-02 -1.20908186e-01
8.45306098e-01 -3.52815866e-01 -1.62498668e-01 1.03083479e+00
4.89775360e-01 3.03275008e-02 6.58175826e-01 -1.39846003e+00
1.97473124e-01 2.43991148e-02 -3.25285703e-01 -4.77972515e-02
-1.12261370e-01 5.92523932e-01 1.74677923e-01 -6.69577479e-01
4.66119438e-01 1.17280769e+00 7.65347421e-01 2.21323535e-01
-1.34636891e+00 -9.58564699e-01 -2.03656465e-01 -7.06773102e-02
-1.58700776e+00 -5.62695265e-02 6.53183460e-01 -5.83430946e-01
7.03668237e-01 -1.29924774e-01 6.66131794e-01 7.66201079e-01
1.52099743e-01 4.34066989e-02 1.01642406e+00 -3.90955657e-01
5.64499438e-01 4.56113845e-01 5.43615967e-02 4.23134476e-01
9.42081213e-01 8.16107273e-01 -1.12750359e-01 1.06921718e-02
3.79062742e-01 -2.16890335e-01 -4.82681133e-02 -8.15726340e-01
-9.89917338e-01 8.94498885e-01 6.19615436e-01 -1.95969611e-01
-1.82157725e-01 6.45038038e-02 2.97113150e-01 4.18496057e-02
1.74864963e-01 4.51080054e-01 -3.62029016e-01 3.69648904e-01
-8.04989517e-01 5.47695518e-01 5.91796100e-01 9.62822378e-01
1.01137221e+00 2.63576806e-01 1.92491665e-01 3.18290532e-01
2.14335069e-01 1.00671911e+00 2.29240939e-01 -9.47942138e-01
3.67624253e-01 6.58376515e-01 4.30785239e-01 -6.60647213e-01
-4.41830069e-01 -3.54013264e-01 -3.52413386e-01 9.42127466e-01
7.27292776e-01 -8.26146603e-02 -1.14106739e+00 1.49917459e+00
3.23639035e-01 1.15795746e-01 1.39405802e-01 7.16094613e-01
5.04742742e-01 4.65811312e-01 7.93093964e-02 1.03051275e-01
1.01580083e+00 -3.41456592e-01 -1.31175220e-01 -4.77943927e-01
5.31996310e-01 -4.73019063e-01 1.08761322e+00 3.85866174e-03
-6.14115298e-01 -8.56162131e-01 -1.33383763e+00 4.31180954e-01
-6.38673425e-01 2.12989971e-02 2.50281394e-01 7.52425492e-01
-6.09646559e-01 6.16318524e-01 -7.41755366e-01 -5.34614682e-01
3.49622339e-01 3.37440163e-01 -1.60208136e-01 -1.61183625e-02
-9.17846501e-01 1.17918992e+00 7.61969030e-01 -2.51811892e-01
-7.78975248e-01 -8.79731119e-01 -7.89396584e-01 -3.77279148e-02
2.71972895e-01 -3.52134824e-01 1.26414180e+00 -4.25905198e-01
-9.53044534e-01 7.90569842e-01 -1.51422182e-02 -8.01945508e-01
6.93752348e-01 8.31664652e-02 -5.14084518e-01 -3.16554546e-01
1.47446468e-01 9.76210296e-01 5.77079654e-01 -1.70538104e+00
-7.45172322e-01 -4.22328740e-01 -1.49553254e-01 -5.96178770e-02
3.03214967e-01 -3.91522437e-01 -5.25680296e-02 -3.71207856e-02
1.19508252e-01 -1.04660666e+00 -3.06584746e-01 3.16251576e-01
9.08847749e-02 3.16503644e-01 9.71004128e-01 1.38292043e-03
3.92086834e-01 -2.13825393e+00 -9.27682400e-01 1.96129039e-01
-1.43075183e-01 4.49971437e-01 -2.25159541e-01 2.11502522e-01
-7.98221007e-02 -1.23477057e-01 -3.25628966e-01 1.14092123e-04
-1.06243089e-01 4.16551083e-01 -3.96681190e-01 3.79783809e-01
3.26036870e-01 4.56824929e-01 -7.26239264e-01 -3.46112669e-01
6.61102474e-01 4.59046155e-01 -2.96919972e-01 4.05134447e-02
-4.09249276e-01 3.93409431e-01 -6.66994005e-02 2.96951056e-01
1.09789741e+00 2.93174416e-01 -1.17658049e-01 -5.66566288e-02
-3.16079021e-01 1.62039697e-01 -1.27561533e+00 9.89353120e-01
-5.21220148e-01 9.02312994e-01 -1.29193231e-01 -5.79898059e-01
1.24937546e+00 -2.87716866e-01 1.65286884e-01 -9.66102779e-01
1.54811546e-01 2.98931211e-01 3.75021517e-01 -1.82892472e-01
6.36299193e-01 -4.99507248e-01 1.09269321e-01 1.63227767e-01
-2.13654354e-01 -7.10389435e-01 2.34883428e-01 -1.70055896e-01
8.18704307e-01 -3.16560641e-02 -6.32061809e-02 -1.91500038e-01
9.73787606e-02 5.77355444e-01 4.56239462e-01 8.00131202e-01
-1.87572077e-01 5.72725296e-01 1.72918782e-01 -5.90539932e-01
-1.39154744e+00 -1.21585357e+00 -4.85402256e-01 3.59109014e-01
4.59489644e-01 6.06932975e-02 -6.69200420e-01 -3.70577276e-01
4.91216928e-01 1.25115633e+00 -3.73405963e-01 -2.80369520e-01
-4.20490652e-01 -8.22132349e-01 4.70590353e-01 5.45220852e-01
4.29373026e-01 -8.18520129e-01 -1.36280203e+00 5.83078377e-02
4.18542475e-01 -1.52083313e+00 4.79152024e-01 3.47907662e-01
-8.91498029e-01 -1.01113892e+00 -1.51695818e-01 -1.66776672e-01
5.94276011e-01 3.43480021e-01 1.06763864e+00 -1.08983658e-01
-4.04008210e-01 1.16409119e-02 -7.17935339e-02 -1.12596118e+00
-8.08410168e-01 -3.16950589e-01 1.84171617e-01 -5.16508102e-01
4.46435571e-01 -4.55769479e-01 -2.71881491e-01 4.72835183e-01
-7.49807656e-01 -1.20174646e-01 5.14557362e-01 3.48030716e-01
5.48901021e-01 1.87990546e-01 9.19777304e-02 -4.77979124e-01
4.09830302e-01 3.69568951e-02 -1.37425387e+00 -1.50388733e-01
-6.85351610e-01 2.74556190e-01 5.04880786e-01 -4.56134588e-01
-7.71259427e-01 4.37672585e-01 2.47328192e-01 -6.92204177e-01
-5.42880654e-01 2.10510585e-02 -2.36903578e-02 -1.63675547e-01
1.20784009e+00 -1.27437770e-01 -2.48532407e-02 -2.43987948e-01
9.92860347e-02 4.69459087e-01 5.12000084e-01 -4.27532405e-01
1.24641120e+00 8.39507043e-01 2.89613217e-01 -8.53810966e-01
-4.78526801e-01 -2.54019111e-01 -6.42946362e-01 -2.73413092e-01
5.76750815e-01 -8.48116875e-01 -5.07106543e-01 2.55250931e-01
-1.13627410e+00 -5.64899564e-01 -7.29848921e-01 8.29784513e-01
-3.88938189e-01 7.67012164e-02 2.21672967e-01 -8.93100083e-01
1.20651893e-01 -1.19154251e+00 1.00137496e+00 2.20577165e-01
1.04346454e-01 -4.67116207e-01 5.01237512e-02 -1.94625258e-01
2.95605898e-01 4.16768104e-01 7.95245707e-01 -4.64494884e-01
-7.20087409e-01 -4.51213062e-01 -5.59466600e-01 3.03245187e-01
-1.53842896e-01 2.72171289e-01 -1.14843297e+00 -1.53515145e-01
-5.62802367e-02 -7.82420412e-02 8.77361834e-01 2.87212342e-01
8.36103976e-01 3.60492676e-01 -5.46752870e-01 2.97129095e-01
1.52411807e+00 3.49891961e-01 4.21385407e-01 3.40761840e-01
3.50085676e-01 7.55723953e-01 8.51359487e-01 1.07451826e-01
8.96075219e-02 7.08794296e-01 6.72061205e-01 -9.91692990e-02
-2.47464359e-01 -3.97676200e-01 2.92281136e-02 -8.94734189e-02
3.23208719e-01 -1.55292213e-01 -1.39904892e+00 5.42955339e-01
-1.51810169e+00 -7.79064953e-01 -4.03915584e-01 2.80504012e+00
1.36311531e-01 7.68502474e-01 8.16031694e-02 6.50142655e-02
7.86484301e-01 -1.67504713e-01 -7.13733137e-01 -2.63338983e-01
3.20078246e-02 2.53129750e-01 1.13475227e+00 4.48211104e-01
-8.41526568e-01 6.87959373e-01 6.15094948e+00 4.77660954e-01
-1.53007579e+00 -1.33092189e-02 3.07898987e-02 -5.18523939e-02
9.68057737e-02 9.86421257e-02 -9.39944386e-01 5.26796043e-01
1.20149422e+00 -3.05706501e-01 -7.28937797e-03 1.06502593e+00
4.94493395e-01 -5.28848648e-01 -1.04238760e+00 8.08781385e-01
-2.68543929e-01 -1.26252830e+00 -8.36277008e-02 2.23539591e-01
3.87805730e-01 5.65288186e-01 -2.64671773e-01 4.04328585e-01
4.29860890e-01 -7.62561381e-01 8.43598545e-01 2.19634041e-01
8.02850783e-01 -6.30318940e-01 7.73726046e-01 7.29726851e-01
-1.13673103e+00 6.98272809e-02 -6.57432199e-01 -1.38382643e-01
5.20316623e-02 6.28081083e-01 -1.43792427e+00 2.95671910e-01
6.75848067e-01 -2.49233237e-03 -8.06707621e-01 1.24256504e+00
-1.15244351e-01 5.47987461e-01 -8.48260820e-01 -1.11287586e-01
-3.04076225e-02 -1.38020605e-01 4.37677622e-01 1.05314469e+00
3.37548465e-01 -3.63379627e-01 9.68421772e-02 1.09151351e+00
2.54091412e-01 -4.49958116e-01 -9.70728457e-01 3.10043633e-01
5.81104457e-01 9.88736093e-01 -7.48372316e-01 -2.23844662e-01
-7.54784569e-02 1.59100741e-01 8.31620768e-04 3.37514073e-01
-9.49252784e-01 -2.91286737e-01 6.54199481e-01 7.13197410e-01
2.81177789e-01 -4.59371597e-01 -4.41138029e-01 -5.71137607e-01
4.51927632e-02 -3.16663265e-01 -1.25460505e-01 -9.30259407e-01
-9.54130709e-01 4.28583354e-01 4.83592123e-01 -1.53939950e+00
-3.14340740e-01 -8.50781083e-01 -6.25999808e-01 9.02265728e-01
-1.48720634e+00 -7.59918451e-01 -8.22314918e-01 5.14234975e-02
2.02318326e-01 1.93222836e-02 5.76344132e-01 2.23062038e-01
-1.02737933e-01 1.08410932e-01 1.01877987e-01 4.88080829e-02
4.87636030e-01 -9.12258446e-01 7.87809432e-01 6.70366049e-01
7.13282973e-02 1.35382473e-01 1.16139567e+00 -6.23439431e-01
-8.41730535e-01 -1.41550767e+00 2.94802517e-01 -3.78569037e-01
4.36901063e-01 -4.54379141e-01 -1.03432643e+00 3.30077499e-01
-2.83283710e-01 1.91177428e-01 8.63035023e-02 -3.10943872e-01
-2.98420727e-01 -3.27196866e-01 -1.44724321e+00 4.24224794e-01
8.62219453e-01 -2.18529403e-01 -4.63930249e-01 3.37452918e-01
6.79618299e-01 -3.80136490e-01 -2.42980167e-01 9.99253929e-01
4.70553279e-01 -1.16886234e+00 7.98242033e-01 -3.49430948e-01
4.66001630e-02 -6.63255632e-01 -2.02640906e-01 -1.27711749e+00
2.40530059e-01 1.99485078e-01 3.80215108e-01 7.12922812e-01
4.30148065e-01 -7.44669437e-01 1.03220356e+00 4.75959599e-01
-2.51649953e-02 -2.34720618e-01 -9.48298693e-01 -1.31749761e+00
1.78250626e-01 -8.99380624e-01 5.97229540e-01 5.90822279e-01
-8.27377319e-01 2.49433443e-01 5.35012722e-01 4.46830511e-01
7.15141892e-01 1.79154411e-01 1.19658244e+00 -1.67564988e+00
1.63797289e-01 -1.74173430e-01 -8.16335678e-01 -4.12738770e-01
1.25864714e-01 -6.12576425e-01 2.15042427e-01 -1.46360099e+00
-2.40194812e-01 -7.62296855e-01 2.17437476e-01 3.40377763e-02
2.56175429e-01 1.01002574e-01 2.92246044e-01 2.82315705e-02
7.97899961e-02 5.24582267e-01 7.81516492e-01 -6.73195049e-02
-1.15403220e-01 1.70312658e-01 1.11784995e-01 9.53568518e-01
9.34091985e-01 -7.68953919e-01 -3.89326304e-01 -4.12290573e-01
2.03763810e-03 -2.55146444e-01 7.34691381e-01 -1.81988478e+00
-1.84497274e-02 -1.42172694e-01 4.59632725e-01 -9.73683238e-01
6.64374530e-01 -1.07264364e+00 8.78346786e-02 7.14970648e-01
1.81202069e-01 -5.88587932e-02 8.00879180e-01 4.51036245e-01
-6.28465042e-02 -3.81736279e-01 1.08459771e+00 -1.24866709e-01
-7.32375205e-01 -1.28392819e-02 -2.98773289e-01 -8.12496766e-02
1.20657218e+00 -6.85793996e-01 -2.82599419e-01 -2.32965201e-01
-3.56608540e-01 7.12177530e-02 7.40477920e-01 3.00796330e-01
3.49254578e-01 -9.03929412e-01 -5.81788242e-01 4.38683957e-01
3.94713104e-01 4.23173606e-01 -1.32819444e-01 2.28345647e-01
-6.50180042e-01 3.86540592e-01 -2.67083526e-01 -9.99832511e-01
-1.17857337e+00 5.26343048e-01 6.20918870e-01 7.19069839e-02
-2.82895118e-01 4.20838058e-01 1.80194631e-01 -8.97416651e-01
5.76926842e-02 -6.90244794e-01 3.74500126e-01 -1.75048053e-01
6.49324432e-02 1.89714313e-01 3.91361892e-01 -5.37966907e-01
-2.84877330e-01 5.58508873e-01 3.46973181e-01 -1.89360052e-01
9.71280038e-01 4.01919782e-01 5.31471431e-01 5.15280902e-01
6.04333997e-01 4.79673520e-02 -1.29214704e+00 -2.75112018e-02
-1.15043178e-01 -4.79566783e-01 -3.15199047e-02 -5.94648898e-01
-7.02986598e-01 1.15617514e+00 1.22857606e+00 -4.13004830e-02
6.90946519e-01 -7.51657039e-02 2.66945213e-01 7.32750714e-01
6.79411709e-01 -9.23272252e-01 -2.93666929e-01 5.61120808e-01
7.43485272e-01 -1.34253371e+00 9.41644236e-02 -4.51612353e-01
-3.55207890e-01 7.63102770e-01 9.16992545e-01 -2.49641821e-01
5.47362745e-01 4.85398591e-01 2.20405862e-01 -1.32763878e-01
-2.80010104e-01 -3.62525284e-01 -1.00135304e-01 8.38568449e-01
-2.50001162e-01 2.31256336e-01 1.33024141e-01 -5.86501099e-02
-5.47186852e-01 8.45795274e-02 5.82002223e-01 8.75124335e-01
-7.72894561e-01 -9.02957857e-01 -7.93092310e-01 4.03992951e-01
4.27120805e-01 2.24262103e-01 -3.15527290e-01 1.37057531e+00
4.66384053e-01 7.05627859e-01 6.22403026e-01 -3.22541654e-01
7.27280796e-01 5.96030205e-02 4.49525595e-01 -7.25098193e-01
-1.26972169e-01 -4.91839826e-01 -4.29801308e-02 -2.30202734e-01
-8.43049213e-02 -4.73873943e-01 -1.54815769e+00 -1.73128784e-01
-3.68210167e-01 -9.17432830e-02 1.25942016e+00 6.48873925e-01
4.06416893e-01 4.56120253e-01 2.62003034e-01 -1.12284827e+00
-6.25514984e-01 -8.96223664e-01 -3.14783603e-01 2.37028524e-01
2.20358133e-01 -9.92935538e-01 -4.40732270e-01 -2.78897047e-01]
|
[7.948729515075684, -2.476039171218872]
|
a28ce1c5-9ade-4fdc-8434-ca468e3ea29e
|
evolving-losses-for-unsupervised-video
|
2002.12177
| null |
https://arxiv.org/abs/2002.12177v1
|
https://arxiv.org/pdf/2002.12177v1.pdf
|
Evolving Losses for Unsupervised Video Representation Learning
|
We present a new method to learn video representations from large-scale unlabeled video data. Ideally, this representation will be generic and transferable, directly usable for new tasks such as action recognition and zero or few-shot learning. We formulate unsupervised representation learning as a multi-modal, multi-task learning problem, where the representations are shared across different modalities via distillation. Further, we introduce the concept of loss function evolution by using an evolutionary search algorithm to automatically find optimal combination of loss functions capturing many (self-supervised) tasks and modalities. Thirdly, we propose an unsupervised representation evaluation metric using distribution matching to a large unlabeled dataset as a prior constraint, based on Zipf's law. This unsupervised constraint, which is not guided by any labeling, produces similar results to weakly-supervised, task-specific ones. The proposed unsupervised representation learning results in a single RGB network and outperforms previous methods. Notably, it is also more effective than several label-based methods (e.g., ImageNet), with the exception of large, fully labeled video datasets.
|
['Michael S. Ryoo', 'Anelia Angelova', 'AJ Piergiovanni']
|
2020-02-26
|
evolving-losses-for-unsupervised-video-1
|
http://openaccess.thecvf.com/content_CVPR_2020/html/Piergiovanni_Evolving_Losses_for_Unsupervised_Video_Representation_Learning_CVPR_2020_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2020/papers/Piergiovanni_Evolving_Losses_for_Unsupervised_Video_Representation_Learning_CVPR_2020_paper.pdf
|
cvpr-2020-6
|
['self-supervised-action-recognition']
|
['computer-vision']
|
[ 6.52108669e-01 -8.74365345e-02 -5.90074301e-01 -4.63794470e-01
-9.56575155e-01 -4.46981758e-01 6.98455930e-01 6.81030527e-02
-5.43109059e-01 8.02262127e-01 2.72292793e-01 3.07288110e-01
-2.65167058e-01 -6.46233916e-01 -7.30644107e-01 -8.68187785e-01
2.30818465e-01 4.94522750e-01 2.61525095e-01 6.01844303e-02
1.81891844e-01 1.36857271e-01 -1.95834875e+00 2.85965890e-01
6.81195080e-01 1.26759839e+00 2.03049973e-01 2.38261223e-01
-9.98998806e-02 1.17203307e+00 -2.00415879e-01 -3.86295557e-01
1.85785964e-01 -6.85646176e-01 -1.00059879e+00 5.09921312e-01
4.41564769e-01 -1.43399641e-01 -2.29406267e-01 1.09725666e+00
3.50648880e-01 6.84442759e-01 1.11232138e+00 -1.31168699e+00
-5.93981743e-01 3.26251477e-01 -5.84358513e-01 -1.32621795e-01
3.43647748e-01 -1.09948784e-01 1.12027919e+00 -6.60952389e-01
8.98638487e-01 1.19767320e+00 3.84255052e-01 6.51554227e-01
-1.14751112e+00 -2.13719666e-01 2.01955140e-01 4.38276947e-01
-1.34699786e+00 -2.80921429e-01 8.95203114e-01 -6.45761669e-01
5.74713230e-01 2.31843144e-02 3.02550942e-01 1.32836998e+00
-2.08891869e-01 9.81982827e-01 9.82354522e-01 -5.21320403e-01
3.94006044e-01 2.09087774e-01 -7.54447952e-02 7.26999879e-01
5.09307608e-02 -1.03393935e-01 -6.72029018e-01 -8.05983096e-02
7.01718450e-01 3.61251563e-01 -2.08910927e-01 -9.38285172e-01
-1.20658624e+00 9.80894864e-01 1.50362417e-01 4.50848639e-01
-3.20104241e-01 3.78908664e-01 6.45092368e-01 4.11518633e-01
4.69883025e-01 1.87225699e-01 -4.26007599e-01 -2.18594283e-01
-8.06480229e-01 -6.10089451e-02 4.12230790e-01 9.19939101e-01
1.18578053e+00 3.37319039e-02 -3.58137518e-01 1.03106344e+00
3.93296123e-01 2.94768095e-01 8.53248298e-01 -1.12510073e+00
3.02856624e-01 4.17505682e-01 -7.16369599e-02 -7.00318933e-01
-2.71746874e-01 8.06425661e-02 -7.52942443e-01 1.46818981e-01
3.55624527e-01 -1.17081515e-01 -8.72369707e-01 1.97679198e+00
1.52465150e-01 4.30322617e-01 1.49766296e-01 8.65405381e-01
8.03124070e-01 3.51379126e-01 1.60411224e-01 -4.85846847e-01
1.07310581e+00 -1.22430658e+00 -6.84302986e-01 5.28213829e-02
7.28655696e-01 -3.28305334e-01 8.00909817e-01 3.67496431e-01
-8.27529609e-01 -5.83596587e-01 -7.89637148e-01 1.66988209e-01
-4.30047244e-01 1.21293746e-01 8.26326549e-01 6.34806573e-01
-8.58128071e-01 6.96344316e-01 -5.81193030e-01 -8.08047652e-01
4.91264790e-01 2.03253046e-01 -5.17832518e-01 -3.13222021e-01
-1.01717699e+00 9.23019171e-01 7.64993131e-01 -2.73870617e-01
-1.19712496e+00 -2.00536206e-01 -1.13845217e+00 -2.62569547e-01
5.55854917e-01 -6.54382944e-01 8.96812081e-01 -1.52558374e+00
-1.73744893e+00 1.05064809e+00 -5.27104810e-02 -3.48304182e-01
3.08073759e-01 -1.04628265e-01 -3.95894706e-01 4.32731777e-01
-6.94499537e-02 7.27130532e-01 1.19593215e+00 -1.36427736e+00
-4.39047992e-01 -2.35895753e-01 2.27126330e-01 2.85047650e-01
-6.89261734e-01 -8.29664320e-02 -4.18508619e-01 -7.36831427e-01
-1.34794235e-01 -7.63236344e-01 -2.75209218e-01 1.82609543e-01
-1.57753110e-01 -2.94769138e-01 7.12035596e-01 -2.81735480e-01
9.70879912e-01 -2.13591933e+00 7.77314007e-01 1.09130405e-01
-3.48387100e-02 5.26824445e-02 -3.46218139e-01 3.57366890e-01
6.96528926e-02 -1.09621406e-01 -4.93481904e-01 -3.21571499e-01
-2.86490028e-03 4.33635741e-01 1.55994192e-01 6.32459104e-01
2.48669147e-01 7.87817001e-01 -1.13322902e+00 -5.83279967e-01
3.48359108e-01 2.93298095e-01 -3.85987312e-01 3.14019769e-01
-3.40270221e-01 5.84246755e-01 -4.45095301e-01 8.31355095e-01
2.60229915e-01 -4.63420123e-01 1.48095071e-01 -2.06611082e-01
1.71783656e-01 -4.93185997e-01 -1.21014154e+00 2.40925312e+00
-5.07150352e-01 3.86440009e-01 -3.01466912e-01 -1.58562970e+00
1.01968908e+00 3.75408113e-01 8.45352411e-01 -4.02862936e-01
2.50326127e-01 1.63429335e-01 -4.73559797e-01 -7.73948550e-01
3.33550245e-01 -3.39814931e-01 -1.03077725e-01 5.92929244e-01
8.21346700e-01 2.72186864e-02 4.42616194e-01 -8.71836953e-03
9.98040557e-01 4.54775214e-01 4.67024118e-01 -1.86003689e-02
5.67721784e-01 -3.56283456e-01 6.44960999e-01 7.21867979e-01
-1.87580585e-01 8.92645359e-01 2.21930504e-01 -2.84653574e-01
-8.38946342e-01 -8.80650878e-01 -4.03324068e-02 1.33693516e+00
2.97738492e-01 -3.63381475e-01 -4.72778738e-01 -8.36377561e-01
4.53065038e-02 3.22645396e-01 -7.20220029e-01 -3.39392990e-01
-1.26142800e-01 -4.82434601e-01 3.42962384e-01 5.79685152e-01
4.33246881e-01 -1.28135502e+00 -5.36846578e-01 8.37902650e-02
-1.14326827e-01 -1.01870251e+00 -2.22509503e-01 3.40468168e-01
-9.01036680e-01 -1.14588761e+00 -1.12468600e+00 -7.38664448e-01
6.00925386e-01 1.19683325e-01 9.15900290e-01 -1.53319975e-02
-1.96449310e-01 1.10595989e+00 -7.69969881e-01 8.93591866e-02
3.17560807e-02 -1.62014112e-01 1.47344187e-01 5.35638392e-01
4.68445331e-01 -5.44652224e-01 -3.62316787e-01 1.79726943e-01
-1.17940938e+00 -2.59271264e-01 5.80704153e-01 1.08046591e+00
8.52882504e-01 -1.92492500e-01 6.63978696e-01 -8.66813958e-01
3.53868127e-01 -6.63240433e-01 -2.00310275e-01 6.35785758e-01
-3.49309266e-01 2.46956393e-01 5.46171010e-01 -6.76158130e-01
-1.12521660e+00 4.06041116e-01 2.34722197e-01 -8.70029032e-01
-4.01672542e-01 5.01475215e-01 -1.20923266e-01 -2.77071983e-01
5.67712009e-01 2.42576286e-01 3.46446857e-02 -4.77061689e-01
7.20883548e-01 4.94152516e-01 2.75246084e-01 -7.69779205e-01
5.68199933e-01 5.41310370e-01 -3.44949961e-02 -7.73434818e-01
-8.00566196e-01 -6.74958527e-01 -9.49817717e-01 -4.73634928e-01
1.12803745e+00 -7.97142208e-01 -4.13957447e-01 4.68357354e-01
-9.96255815e-01 -2.76376575e-01 -6.01492167e-01 8.02295148e-01
-1.11602032e+00 5.74727535e-01 -4.17786360e-01 -8.26591730e-01
1.60376564e-01 -9.44563568e-01 1.08494687e+00 2.67309308e-01
5.95264956e-02 -1.22157228e+00 3.20211828e-01 4.84618157e-01
2.88575441e-01 3.89572918e-01 6.90002441e-01 -7.06059873e-01
-3.06895256e-01 2.00051501e-01 -1.84333369e-01 6.55335188e-01
2.58354604e-01 -1.01604216e-01 -1.02977395e+00 -4.49427336e-01
-2.57358700e-01 -1.17804694e+00 1.26860845e+00 3.49541157e-01
1.40842676e+00 -3.13302167e-02 -1.02137230e-01 6.24899328e-01
1.56811452e+00 5.19178100e-02 6.61106825e-01 3.59281689e-01
7.57629693e-01 7.15996504e-01 5.66900313e-01 6.62161410e-01
1.94745287e-01 6.35100961e-01 4.48952317e-01 1.42755643e-01
3.93656194e-02 -1.59665212e-01 4.36888099e-01 7.63844132e-01
-5.29689789e-01 -1.86078891e-01 -4.87962306e-01 4.74017024e-01
-2.22555280e+00 -1.19429147e+00 3.63169879e-01 2.30772114e+00
5.35089195e-01 -1.53842151e-01 2.32502103e-01 -1.57582298e-01
8.52800667e-01 3.31238598e-01 -7.21183240e-01 -1.71716437e-01
-2.57928461e-01 2.31791556e-01 4.03557688e-01 3.53287272e-02
-1.44418311e+00 9.30643797e-01 6.22018623e+00 8.74624372e-01
-9.01542604e-01 3.48525435e-01 3.80260050e-01 -2.45554149e-02
-4.44325596e-01 -2.38587439e-01 -4.32638347e-01 3.72853994e-01
7.41622031e-01 -2.07543552e-01 4.21822339e-01 8.17611575e-01
-9.85016003e-02 -7.62072811e-03 -1.11587322e+00 1.32064474e+00
5.85030138e-01 -1.23247886e+00 2.22543776e-01 -1.57370329e-01
1.02203465e+00 -1.01333328e-01 -1.75580196e-02 4.17187840e-01
2.03268334e-01 -8.47567677e-01 5.77816069e-01 7.04550028e-01
8.56528401e-01 -5.97173989e-01 6.17791235e-01 9.52345133e-02
-1.34228694e+00 -3.10477644e-01 -4.44591701e-01 1.25026837e-01
4.23179805e-01 2.05518320e-01 1.27619639e-01 7.51703918e-01
4.86685723e-01 1.29503858e+00 -5.27252555e-01 1.14349484e+00
-2.00887442e-01 3.24875891e-01 -2.63240524e-02 1.16973497e-01
3.30065936e-01 -3.87841105e-01 2.80592293e-01 1.04322267e+00
4.05985743e-01 -1.30955040e-01 4.18329656e-01 4.03537720e-01
-3.32670718e-01 3.19970340e-01 -9.20133293e-01 -2.29627863e-01
1.66952103e-01 1.18283546e+00 -7.06166446e-01 -4.38152254e-01
-7.56461084e-01 1.19210327e+00 4.73616898e-01 5.73207676e-01
-6.38792694e-01 -1.63730294e-01 3.90549898e-01 -2.98432440e-01
5.60067773e-01 6.66249394e-02 2.77832657e-01 -1.33253360e+00
-6.80349991e-02 -7.16362894e-01 7.05890834e-01 -5.86522937e-01
-1.57438242e+00 4.12556440e-01 1.40495107e-01 -1.68752527e+00
-5.76141477e-01 -7.41022110e-01 -5.19466162e-01 2.45620817e-01
-1.70549941e+00 -1.23118234e+00 -2.07053810e-01 9.89609241e-01
5.61291814e-01 -5.91838002e-01 9.83789027e-01 4.80909944e-01
-4.28310692e-01 5.28157413e-01 3.59229296e-01 1.57661538e-03
9.20018375e-01 -1.20667076e+00 -6.43064141e-01 5.75447559e-01
3.59924972e-01 1.94089442e-01 3.06373775e-01 -2.62049526e-01
-1.34747314e+00 -1.16037834e+00 2.92463392e-01 -1.51114926e-01
7.11308897e-01 -9.53127146e-02 -8.41184437e-01 6.32609189e-01
1.10809624e-01 4.09311414e-01 1.04526854e+00 1.14295244e-01
-4.82627511e-01 -8.84140134e-02 -1.23172128e+00 -6.40475899e-02
1.18197417e+00 -5.70886612e-01 -4.75911200e-01 6.07921839e-01
5.83726287e-01 4.12236825e-02 -1.00979817e+00 4.14969027e-01
5.01225710e-01 -7.77600884e-01 8.05853784e-01 -1.00172269e+00
5.72013140e-01 -2.07369849e-01 -4.78480667e-01 -1.28033590e+00
-3.31099093e-01 -4.33534086e-01 -4.47471648e-01 1.15422904e+00
3.30210984e-01 -3.02062213e-01 6.69419646e-01 2.83418238e-01
-1.45991743e-02 -5.87055981e-01 -9.47592974e-01 -1.05637157e+00
-7.67023489e-02 -2.31751055e-01 1.00945830e-01 1.11985457e+00
-1.71444640e-02 1.09148592e-01 -8.34994256e-01 -2.80691028e-01
8.95786405e-01 1.73963338e-01 4.25233692e-01 -1.27790534e+00
-4.68844235e-01 -3.49747241e-01 -8.24188948e-01 -9.77357924e-01
5.77618241e-01 -1.06346905e+00 5.99588491e-02 -1.45632672e+00
4.83745933e-01 -1.84739813e-01 -7.97682822e-01 4.88957375e-01
1.22895837e-01 3.27404946e-01 2.89898098e-01 1.73868835e-01
-1.31989312e+00 9.52976763e-01 1.09258294e+00 -3.70066553e-01
-1.26315989e-02 -1.37720704e-01 -4.45111036e-01 8.02969754e-01
6.25881910e-01 -3.95940751e-01 -5.01397610e-01 -3.94181788e-01
-5.81842251e-02 -4.99135368e-02 1.67657524e-01 -1.00776398e+00
2.41560638e-01 -3.63048851e-01 1.26188308e-01 -2.38263443e-01
4.42617685e-01 -7.96433032e-01 -4.79912414e-04 9.05772299e-02
-5.20961761e-01 -3.52701992e-01 -2.60703713e-01 8.48690033e-01
-5.69398046e-01 -6.59261525e-01 9.25379276e-01 -3.41428906e-01
-1.40033650e+00 5.23839653e-01 -2.66678035e-01 2.14273050e-01
1.32051444e+00 -2.18305111e-01 -7.95992464e-02 -3.02899331e-01
-8.58297408e-01 1.86291993e-01 4.05660272e-01 5.37172139e-01
7.42829680e-01 -1.71746457e+00 -6.33627474e-01 -9.51578319e-02
5.96473515e-01 -3.17534626e-01 3.83462995e-01 6.53668404e-01
-1.11376919e-01 2.27397516e-01 -5.30421317e-01 -6.55113876e-01
-1.01965904e+00 6.76760197e-01 1.10702008e-01 -1.73994422e-01
-4.72000390e-01 7.77535975e-01 -1.06831685e-01 -5.46485424e-01
3.78120542e-01 1.07995413e-01 -4.64430779e-01 4.01774108e-01
6.70172811e-01 4.29512203e-01 -2.60503978e-01 -8.31592381e-01
-2.39422917e-01 8.07309270e-01 1.31782547e-01 1.21668726e-02
1.42887354e+00 -1.23998508e-01 3.43096368e-02 8.35149109e-01
1.46279621e+00 -7.38833368e-01 -1.45403743e+00 -4.54130322e-01
-6.37319032e-03 -5.14303267e-01 -1.44413024e-01 -3.31732541e-01
-1.30364907e+00 7.77927518e-01 7.07087934e-01 1.75360460e-02
1.08261490e+00 2.28765920e-01 4.12147850e-01 6.66732013e-01
7.07045376e-01 -1.42130709e+00 5.82886398e-01 4.22569007e-01
6.29132926e-01 -1.60976875e+00 -6.19021244e-02 6.52744155e-03
-8.93843710e-01 1.22913837e+00 7.55796313e-01 -4.78833355e-02
6.07411206e-01 -2.08093435e-01 -1.73473835e-01 -1.13382094e-01
-6.13214731e-01 -7.15807259e-01 3.21230650e-01 8.94960999e-01
3.87970388e-01 -1.22616701e-01 -3.38335514e-01 2.40457267e-01
6.10054851e-01 2.06412211e-01 2.54374415e-01 1.01822293e+00
-5.39947748e-01 -1.22927618e+00 -2.07538847e-02 5.96669912e-01
-1.58176020e-01 2.29066834e-01 -1.04175143e-01 4.83686715e-01
1.72483817e-01 7.99369574e-01 6.73070922e-03 -3.81190836e-01
6.93723783e-02 1.18418247e-01 7.24043071e-01 -8.53152394e-01
-1.71941027e-01 -1.03240073e-01 -1.30430683e-01 -5.55397570e-01
-1.24623466e+00 -7.11039901e-01 -9.03202474e-01 1.88740566e-01
-3.14611822e-01 2.27657612e-02 2.75237650e-01 1.09945071e+00
3.16436216e-02 3.51860434e-01 7.61648238e-01 -1.00585723e+00
-4.23963279e-01 -8.64928186e-01 -9.80760872e-01 9.23832357e-01
-3.64835672e-02 -1.22438300e+00 -3.73445153e-01 3.61570179e-01]
|
[8.574716567993164, 0.7938348650932312]
|
405aaee0-67db-451d-9dd2-12c891383856
|
the-power-of-tiling-for-small-object
| null | null |
https://openaccess.thecvf.com/content_CVPRW_2019/papers/UAVision/Unel_The_Power_of_Tiling_for_Small_Object_Detection_CVPRW_2019_paper.pdf
|
https://openaccess.thecvf.com/content_CVPRW_2019/papers/UAVision/Unel_The_Power_of_Tiling_for_Small_Object_Detection_CVPRW_2019_paper.pdf
|
The Power of Tiling for Small Object Detection
|
Deep neural network based techniques are state-of-the-art for object detection and classification with the help ofthe development in computational power and memory ef-ficiency. Although these networks are adapted for mobileplatforms with sacrifice in accuracy; the resolution increasein visual sources makes the problem even harder by raisingthe expectations to leverage all the details in images. Real-time small object detection in low power mobile devices hasbeen one of the fundamental problems of surveillance ap-plications. In this study, we address the detection of pedes-trians and vehicles onboard a micro aerial vehicle (MAV)with high-resolution imagery. For this purpose, we exploitPeleeNet, to our best knowledge the most efficient networkmodel on mobile GPUs, as the backbone of an SSD networkas well as 38x38 feature map in the earlier layer. After illus-trating the low accuracy of state-of-the-art object detectorsunder the MAV scenario, we introduce a tiling based ap-proach that is applied in both training and inference phases.The proposed technique limits the detail loss in object de-tection while feeding the network with a fixed size input.The improvements provided by the proposed approach areshown by in-depth experiments performed along Nvidia Jet-son TX1 and TX2 using the VisDrone2018 dataset.
|
['F. Ozge UnelBurak OzkalayciCevahir Cigla']
|
2019-06-11
| null | null | null |
computer-vision-and-pattern-recognition-2019-1
|
['small-object-detection']
|
['computer-vision']
|
[ 1.61212951e-01 -1.89558923e-01 -2.11048368e-02 1.23364732e-01
-1.31901443e-01 -5.30856073e-01 6.20723069e-01 -2.10739851e-01
-7.31303155e-01 3.68476540e-01 -5.91995537e-01 -5.06962717e-01
-9.90201458e-02 -9.64984477e-01 -8.08673203e-01 -6.69547081e-01
-2.68292069e-01 1.38347656e-01 8.69146883e-01 -2.42240012e-01
-1.69798270e-01 9.42672729e-01 -1.98793209e+00 5.06102443e-01
6.97621346e-01 1.42036581e+00 3.46764654e-01 1.03092825e+00
4.79225814e-01 8.15317154e-01 -6.29172206e-01 -3.58614892e-01
5.62013805e-01 1.41723573e-01 -2.70311713e-01 7.73699358e-02
1.05055094e+00 -8.27470124e-01 -3.90760750e-01 1.02888250e+00
4.10180718e-01 -1.45526126e-01 4.07901645e-01 -1.04700184e+00
3.91996294e-01 1.28928602e-01 -5.80529630e-01 8.81033540e-01
-2.98669636e-01 1.73232675e-01 3.15685391e-01 -8.55062187e-01
5.04588366e-01 9.95959938e-01 1.06788445e+00 1.71401054e-01
-8.77332270e-01 -6.73756778e-01 -6.85857013e-02 3.22694719e-01
-1.14488280e+00 -3.75642359e-01 4.05991614e-01 -4.31535274e-01
1.33014119e+00 3.40769112e-01 7.55226135e-01 8.43401611e-01
4.64444399e-01 6.20571733e-01 8.02007318e-01 -1.99520439e-01
9.49770510e-02 2.34226003e-01 3.57033700e-01 1.08362448e+00
9.25160527e-01 4.75852311e-01 -2.72873789e-01 -1.96096033e-01
9.05832350e-01 1.48459241e-01 -2.27150723e-01 -2.15838552e-01
-6.54831886e-01 7.05332398e-01 6.47985160e-01 8.87888074e-02
-6.79572523e-01 2.39368007e-01 5.26665270e-01 1.53980345e-01
2.74793029e-01 1.66000426e-01 -3.72842550e-01 6.80182204e-02
-1.16254592e+00 1.07561603e-01 7.04006910e-01 5.92399836e-01
3.71831059e-01 5.36155939e-01 1.08067274e-01 2.53397137e-01
3.22924703e-01 9.03843462e-01 2.10591078e-01 -7.94867456e-01
4.33759481e-01 5.76679885e-01 3.19382139e-02 -1.22136092e+00
-5.05279243e-01 -6.78370714e-01 -8.97065938e-01 9.59577620e-01
3.90168697e-01 -5.53123772e-01 -1.04855478e+00 1.07757831e+00
4.76744413e-01 1.07174627e-01 -3.42698544e-02 9.38884199e-01
7.41585016e-01 8.53471279e-01 -1.14923805e-01 2.57076621e-01
1.58797252e+00 -7.18398392e-01 -1.73451990e-01 -3.52854252e-01
5.76120436e-01 -5.02816319e-01 4.20182139e-01 6.66303158e-01
-7.27240801e-01 -6.60180688e-01 -1.39395487e+00 2.19216019e-01
-3.33791941e-01 4.52387333e-01 6.36915445e-01 8.49716485e-01
-9.41769183e-01 6.22507989e-01 -1.21253955e+00 -7.05687702e-01
7.34378934e-01 6.59815669e-01 -1.54440388e-01 2.49391332e-01
-5.68588257e-01 7.76790500e-01 4.97775048e-01 3.17537016e-03
-1.08597898e+00 -5.81125617e-01 -3.61429065e-01 1.63623616e-01
6.02264822e-01 -5.05149424e-01 1.11367118e+00 -1.07959068e+00
-1.23263645e+00 5.96575916e-01 5.08143306e-01 -9.92690384e-01
5.49683154e-01 -3.23840648e-01 -2.32423782e-01 4.12562460e-01
-3.65048558e-01 7.44819641e-01 1.06499135e+00 -8.01808596e-01
-1.08738649e+00 -5.13803005e-01 3.90210241e-01 8.81832987e-02
-4.09576148e-01 1.19874299e-01 -8.71162638e-02 -4.71553475e-01
-1.79368183e-01 -8.57220352e-01 2.81033423e-02 6.21365070e-01
-2.39883319e-01 1.43282101e-01 1.43148077e+00 -5.39753735e-01
7.84753978e-01 -2.03346634e+00 -2.50309259e-01 -1.37583702e-03
1.60852522e-01 1.09267485e+00 -2.49002315e-02 7.51137137e-02
1.41278118e-01 -5.47566891e-01 -2.37610331e-03 -8.66227970e-02
-3.85067731e-01 1.80966988e-01 -4.30983782e-01 5.87796748e-01
7.70043880e-02 3.80586565e-01 -5.23242652e-01 -4.98897284e-01
3.89409095e-01 5.86097896e-01 -3.91655475e-01 1.81141749e-01
-1.03834942e-01 -3.63690376e-01 -3.33785146e-01 8.36891472e-01
7.79747009e-01 -7.72255659e-02 -2.43539177e-02 -3.49569321e-01
-4.88241464e-01 -7.86719248e-02 -1.20039320e+00 1.18231380e+00
-2.75649905e-01 1.03540003e+00 6.91850603e-01 -6.83609068e-01
5.75684845e-01 1.58169791e-01 -1.55062387e-02 -3.05593282e-01
6.14907324e-01 1.87513322e-01 1.66988850e-01 -4.47241873e-01
6.98769093e-01 3.28646749e-01 4.95659739e-01 -5.14186779e-03
1.07125126e-01 1.86329708e-01 1.88147634e-01 -1.56652853e-01
9.85509396e-01 3.33519317e-02 1.97614655e-01 -5.86998343e-01
3.40519160e-01 4.00938630e-01 8.40425864e-02 1.01720536e+00
-4.43602465e-02 2.27995396e-01 2.48884633e-01 -7.55350292e-01
-9.68068063e-01 -6.31381035e-01 -3.52462173e-01 1.06879616e+00
-9.74026248e-02 -2.58440852e-01 -9.83983934e-01 -7.32800305e-01
-1.95468977e-01 2.57355720e-01 -5.05828857e-01 3.09405953e-01
-5.80749929e-01 -1.13965583e+00 8.69685173e-01 6.61691606e-01
1.11652732e+00 -7.28294969e-01 -1.87067223e+00 1.39312074e-01
4.87602085e-01 -1.18267715e+00 2.10840642e-01 5.28130949e-01
-1.13158131e+00 -1.33948100e+00 -4.17685270e-01 -5.64573646e-01
5.41777968e-01 5.94590962e-01 8.98718536e-01 7.55070969e-02
-5.70889533e-01 1.08246796e-01 -4.14587855e-02 -7.02451825e-01
-8.48544389e-02 2.17281088e-01 9.95316207e-02 -2.41164103e-01
4.36373562e-01 -2.47078791e-01 -5.89632273e-01 1.61681697e-01
-7.85636306e-01 1.83183309e-02 7.13572502e-01 6.91262484e-01
2.58056760e-01 1.43007547e-01 -3.24006885e-01 -4.95907485e-01
3.64420936e-02 -6.00809865e-02 -1.44336283e+00 8.51961551e-04
-2.07592800e-01 -2.29224890e-01 7.26149499e-01 -4.84883547e-01
-7.75752068e-01 2.59273052e-01 -1.37786895e-01 -4.50429827e-01
-2.41617098e-01 -4.34658118e-02 -3.51241231e-02 -6.41154051e-01
9.11633015e-01 1.28163781e-03 -1.09997869e-01 -1.80168584e-01
-1.00407518e-01 6.59666657e-01 5.33532441e-01 7.67153203e-02
6.58932090e-01 7.73349941e-01 4.54923600e-01 -1.35389531e+00
-6.28900349e-01 -1.96649596e-01 -5.15311301e-01 -2.03305736e-01
8.02341580e-01 -9.97205675e-01 -7.93877244e-01 8.24027300e-01
-1.46785891e+00 -2.73372173e-01 -1.30539894e-01 3.57168764e-01
3.90479229e-02 1.93688914e-01 -5.37085235e-01 -9.18094337e-01
-6.41642511e-01 -1.03485262e+00 9.26719904e-01 3.55200708e-01
3.16133827e-01 -6.52993321e-01 -2.54869342e-01 -1.92908928e-01
4.44368809e-01 2.64426649e-01 3.72294188e-01 -3.96671444e-01
-9.30533290e-01 -4.69722807e-01 -6.57996714e-01 3.39805394e-01
-3.61280292e-01 -1.38435960e-01 -1.13800573e+00 -3.94584119e-01
2.35761508e-01 -2.58522457e-03 1.04379821e+00 4.42468762e-01
8.65147710e-01 -2.67845601e-01 -5.01622200e-01 9.49242949e-01
1.73855090e+00 2.53293514e-01 5.62838197e-01 5.24613619e-01
7.17429578e-01 3.54381114e-01 4.62842733e-01 4.98100609e-01
-5.92360757e-02 7.84148097e-01 8.78520787e-01 -1.97629690e-01
-2.73196459e-01 2.52455413e-01 5.15538335e-01 8.86894390e-02
-3.27543736e-01 -5.60604990e-01 -9.75169837e-01 2.04360634e-01
-1.65576661e+00 -9.27133381e-01 -4.65296507e-01 2.07121539e+00
3.74711938e-02 3.40041220e-01 7.06781521e-02 6.11332878e-02
4.66171801e-01 3.61999750e-01 -3.99915159e-01 -8.46459344e-02
-1.25770181e-01 -9.83148441e-02 1.01156569e+00 3.08656603e-01
-1.49693668e+00 6.95194662e-01 5.20384884e+00 7.92713046e-01
-1.43532240e+00 2.56698668e-01 2.95559257e-01 -1.80181444e-01
8.50426555e-01 -4.73807961e-01 -1.31910765e+00 4.10873413e-01
9.72482860e-01 5.17722964e-01 2.07310081e-01 1.17379844e+00
-2.33759973e-02 -4.45631504e-01 -7.09247768e-01 1.06481910e+00
4.43668097e-01 -1.24848115e+00 -7.12273419e-02 4.33124751e-02
2.83110112e-01 5.75062275e-01 1.56988233e-01 5.09530120e-02
-1.62998643e-02 -6.53873026e-01 4.73669052e-01 4.22726683e-02
7.36306012e-01 -8.31540108e-01 1.18674183e+00 3.78053099e-01
-1.07372153e+00 -3.84511679e-01 -6.52981937e-01 -2.17107549e-01
-8.04293528e-02 2.94838965e-01 -9.47142661e-01 1.19314224e-01
1.04033983e+00 3.53888988e-01 -5.85523844e-01 1.14189196e+00
-3.88563983e-02 6.29020929e-01 -7.78115511e-01 -1.32014662e-01
4.79278445e-01 8.58376399e-02 8.01185906e-01 1.32437897e+00
5.77992737e-01 3.13128203e-01 1.40497610e-01 2.88461953e-01
7.21175447e-02 -4.10724342e-01 -8.47873449e-01 1.53472096e-01
1.15938649e-01 1.47512674e+00 -1.03550506e+00 -5.04657388e-01
-3.36676121e-01 8.29347312e-01 -1.07150905e-01 -1.39510214e-01
-8.35907638e-01 -5.14329851e-01 5.89710355e-01 6.18151188e-01
8.23147357e-01 -3.05298865e-01 -2.12627605e-01 -1.00275481e+00
-1.59448653e-01 -7.74527192e-01 2.04452530e-01 -6.12138808e-01
-5.38810492e-01 1.10176194e+00 5.74953705e-02 -1.16163766e+00
-1.91402495e-01 -1.20771837e+00 -4.55871969e-01 4.34148699e-01
-1.47716177e+00 -1.23344469e+00 -7.86614776e-01 2.05989838e-01
6.65631115e-01 -5.28244317e-01 7.84656465e-01 2.51324892e-01
-5.50062239e-01 2.81284302e-01 2.24873140e-01 7.04169348e-02
6.30421489e-02 -7.96691179e-01 3.86126369e-01 1.39103734e+00
-8.09233934e-02 2.01302141e-01 7.74755418e-01 -8.01264107e-01
-1.57001388e+00 -1.21908462e+00 2.41394699e-01 -3.52460086e-01
4.98030514e-01 -5.52890420e-01 -8.55963945e-01 4.32214409e-01
3.13314766e-01 1.55050516e-01 9.74800959e-02 -3.46097618e-01
8.05452019e-02 -2.10296288e-01 -1.16424477e+00 2.93539852e-01
8.21906984e-01 -1.57585070e-01 -3.90942067e-01 4.40968871e-01
2.33129427e-01 -6.38643742e-01 -2.45216638e-01 2.74670780e-01
5.43647647e-01 -1.25943518e+00 1.01734960e+00 -1.86969757e-01
7.80364573e-02 -4.04634267e-01 -2.16953710e-01 -8.20152819e-01
-1.80752322e-01 -4.52535778e-01 -4.36699033e-01 6.39464915e-01
-9.66366567e-03 -4.72195119e-01 1.30063081e+00 -2.89625470e-02
-2.13745192e-01 -5.17606080e-01 -9.82705891e-01 -9.84660864e-01
-5.38878143e-01 -3.08682203e-01 3.73077095e-01 2.96631128e-01
-7.47955382e-01 2.60730624e-01 -3.04035425e-01 6.38337910e-01
7.69162595e-01 -2.34007761e-01 7.88928449e-01 -1.43263698e+00
-3.70735645e-01 1.95215680e-02 -8.10557246e-01 -9.43172157e-01
-3.51439178e-01 -2.98873335e-01 -1.54906427e-02 -1.03900862e+00
-8.84952918e-02 -7.49769583e-02 1.80641040e-01 3.74088228e-01
2.37767115e-01 4.95650202e-01 4.18087006e-01 4.02314253e-02
-5.80770850e-01 1.28180891e-01 5.55409193e-01 -2.97504738e-02
2.49233693e-02 -7.01078996e-02 8.66283849e-02 1.00514030e+00
5.76391041e-01 -6.70721769e-01 -3.77159655e-01 -7.92075813e-01
8.34233761e-02 -7.25723431e-02 8.95657659e-01 -1.72800899e+00
5.23047030e-01 4.71581042e-01 7.58149922e-01 -9.67666745e-01
7.35302210e-01 -1.29615724e+00 8.06854144e-02 1.00963759e+00
4.68958825e-01 3.71375859e-01 6.70759797e-01 3.16391230e-01
8.15370455e-02 -3.36732090e-01 8.90559256e-01 -1.06516801e-01
-8.91250312e-01 1.26473710e-01 -6.74291492e-01 -4.10536081e-01
1.06463337e+00 -5.25646389e-01 -5.37629545e-01 1.66268274e-01
-1.28432035e-01 -2.13224679e-01 4.99790996e-01 1.05036013e-01
5.15332460e-01 -7.25391984e-01 -5.77964544e-01 4.31627989e-01
-2.85678655e-01 -3.43658812e-02 3.16957533e-01 7.69053102e-01
-1.11717618e+00 7.22051978e-01 -5.08564711e-01 -8.20656121e-01
-1.72383976e+00 4.70697999e-01 6.17050469e-01 -6.20580278e-02
-5.87545574e-01 8.90997112e-01 2.23455027e-01 6.35317639e-02
3.55256766e-01 -5.76692879e-01 -3.04372400e-01 1.39503673e-01
9.43780959e-01 7.84985900e-01 3.02303582e-01 -4.18295115e-01
-4.30522591e-01 5.28927088e-01 -1.66410327e-01 2.66198188e-01
1.40194356e+00 3.11673641e-01 7.29745999e-02 -1.30504236e-01
7.24537551e-01 -5.11105776e-01 -1.45846617e+00 2.18196645e-01
-3.37048858e-01 -4.92410928e-01 6.59209609e-01 -6.71690047e-01
-1.07136607e+00 9.63199496e-01 1.47907126e+00 2.35501692e-01
1.03085637e+00 -4.76881802e-01 6.50732696e-01 7.68708885e-01
5.14400601e-01 -9.33051646e-01 -1.09364755e-01 5.74678898e-01
5.04470170e-01 -1.38308990e+00 1.90483421e-01 -5.59623837e-01
-1.69668123e-01 1.36156261e+00 8.78533125e-01 -4.23384905e-01
3.98498416e-01 6.69542015e-01 -1.51883900e-01 -4.01939958e-01
-4.47965324e-01 -1.58551242e-02 7.99267665e-02 8.15851271e-01
-1.06018163e-01 -1.10782482e-01 1.77895337e-01 1.72789514e-01
1.07402362e-01 9.04661417e-02 5.60204506e-01 1.20150912e+00
-7.08287120e-01 -4.96198386e-01 -7.62608647e-01 6.19451344e-01
-6.03641927e-01 -1.99464366e-01 -5.27812466e-02 1.02643132e+00
4.34781313e-01 7.44273484e-01 2.66714215e-01 -3.52418840e-01
2.90817678e-01 -6.70523047e-01 3.88885438e-01 -4.11273211e-01
-8.84694874e-01 -2.29829267e-01 3.00547063e-01 -4.54649925e-01
-2.52202719e-01 -5.24702013e-01 -8.73717248e-01 -2.67155796e-01
-5.97282112e-01 -2.54009426e-01 9.25726116e-01 7.06159115e-01
5.39913177e-01 5.90396523e-01 1.59364436e-02 -1.24642003e+00
-4.75894123e-01 -7.74612248e-01 -3.67726952e-01 -4.19049799e-01
4.18280721e-01 -5.63414097e-01 -6.37424663e-02 -8.31633806e-02]
|
[8.46604061126709, -0.9556877613067627]
|
24bc625c-54d4-40cb-bac2-613b3ba33515
|
ualberta-at-semeval-2023-task-1-context
|
2306.14067
| null |
https://arxiv.org/abs/2306.14067v1
|
https://arxiv.org/pdf/2306.14067v1.pdf
|
UAlberta at SemEval-2023 Task 1: Context Augmentation and Translation for Multilingual Visual Word Sense Disambiguation
|
We describe the systems of the University of Alberta team for the SemEval-2023 Visual Word Sense Disambiguation (V-WSD) Task. We present a novel algorithm that leverages glosses retrieved from BabelNet, in combination with text and image encoders. Furthermore, we compare language-specific encoders against the application of English encoders to translated texts. As the contexts given in the task datasets are extremely short, we also experiment with augmenting these contexts with descriptions generated by a language model. This yields substantial improvements in accuracy. We describe and evaluate additional V-WSD methods which use image generation and text-conditioned image segmentation. Overall, the results of our official submission rank us 18 out of 56 teams. Some of our unofficial results are even better than the official ones. Our code is publicly available at https://github.com/UAlberta-NLP/v-wsd.
|
['Grzegorz Kondrak', 'Ning Shi', 'Talgat Omarov', 'Bradley Hauer', 'Michael Ogezi']
|
2023-06-24
| null | null | null | null |
['image-generation', 'word-sense-disambiguation']
|
['computer-vision', 'natural-language-processing']
|
[ 1.82628468e-01 8.84616300e-02 -2.24092737e-01 -3.42565477e-01
-1.12725055e+00 -9.72332358e-01 1.01962245e+00 2.43055355e-02
-8.10732663e-01 8.80274117e-01 5.10658443e-01 -3.50070208e-01
3.44768554e-01 -4.48008031e-01 -8.45573604e-01 -2.55797148e-01
3.91880304e-01 7.51603723e-01 1.08129814e-01 -3.52051109e-01
3.21020633e-02 1.19798603e-02 -1.22617626e+00 6.33640349e-01
6.71766460e-01 7.83209741e-01 5.16328096e-01 8.14416170e-01
-1.52544990e-01 5.94741046e-01 -4.88803893e-01 -7.90684342e-01
2.73688555e-01 -1.47559509e-01 -1.20634520e+00 1.19579751e-02
9.81176019e-01 -1.90313518e-01 -4.28443521e-01 1.09641027e+00
4.63680536e-01 1.00601651e-01 6.47210360e-01 -1.09965336e+00
-1.01530254e+00 5.96161962e-01 -3.59371692e-01 2.25926042e-01
4.73710537e-01 2.38124654e-01 1.37640119e+00 -9.89330173e-01
1.27767205e+00 1.33076930e+00 4.30737197e-01 6.29056454e-01
-1.27102363e+00 -5.89516878e-01 2.69406199e-01 3.96412432e-01
-1.55973387e+00 -6.54354036e-01 1.81834742e-01 -5.27996898e-01
1.39432454e+00 1.61443099e-01 5.65583408e-01 1.47077012e+00
-2.02644885e-01 1.16991484e+00 1.15962625e+00 -5.81460595e-01
-1.12949610e-01 1.07890181e-01 -2.66984329e-02 8.79410088e-01
3.55058461e-01 1.32900596e-01 -5.19989073e-01 1.40784338e-01
4.96825695e-01 -5.48992395e-01 -2.26694405e-01 -9.84071568e-02
-1.40251553e+00 5.83573282e-01 4.18204039e-01 3.87742698e-01
-2.02154234e-01 2.47658998e-01 5.38505495e-01 1.65254220e-01
6.73812628e-01 6.17227435e-01 -4.71558869e-01 9.49816927e-02
-1.17190957e+00 5.76670706e-01 7.25170016e-01 1.24922287e+00
5.55800796e-01 -2.27285817e-01 -3.93660843e-01 8.91205072e-01
2.57808179e-01 6.50981963e-01 3.01718682e-01 -1.22904384e+00
3.69538069e-01 1.12742923e-01 1.81836784e-01 -7.97219336e-01
-1.63537309e-01 -2.23909482e-01 -3.30993652e-01 1.09461814e-01
4.86567497e-01 -1.21092059e-01 -1.31152511e+00 1.75377393e+00
-3.67922941e-03 5.94260264e-03 9.48748142e-02 9.70124543e-01
9.22105908e-01 5.67610621e-01 5.12033582e-01 1.06509365e-01
1.58532870e+00 -1.11364615e+00 -7.47733355e-01 -4.14440781e-01
4.09799188e-01 -1.05536044e+00 1.14017093e+00 3.30137551e-01
-9.81604040e-01 -5.64097762e-01 -9.84385848e-01 -4.39738780e-01
-6.43269897e-01 3.58005971e-01 3.83952111e-01 4.22529429e-02
-1.38742471e+00 3.57603461e-01 -7.27623999e-01 -8.26503932e-01
4.35579449e-01 -2.98890751e-02 -4.04532194e-01 -5.22784851e-02
-1.30348003e+00 1.21198380e+00 5.97062588e-01 -2.53771871e-01
-1.06590819e+00 -3.71677876e-01 -1.04226983e+00 -4.91272688e-01
5.05082607e-01 -8.71302366e-01 1.70073855e+00 -9.85387743e-01
-8.90683532e-01 1.41968369e+00 -2.58363277e-01 -6.95222139e-01
6.45024657e-01 -4.06444877e-01 -4.23851758e-01 3.45018566e-01
6.17684066e-01 1.31090283e+00 4.68082756e-01 -1.29395974e+00
-6.99400127e-01 -1.27416492e-01 2.13170782e-01 4.60607916e-01
1.08208187e-01 1.29449084e-01 -9.82449234e-01 -8.38629663e-01
-5.35779059e-01 -9.07164574e-01 -2.13782072e-01 5.71432300e-02
-6.97371006e-01 -2.76577532e-01 4.55327332e-01 -1.08066773e+00
9.29316580e-01 -2.09208465e+00 -2.43813302e-02 -1.34982824e-01
2.40333676e-01 2.76108891e-01 -4.03543621e-01 3.16005975e-01
-7.34572671e-03 4.02366877e-01 -3.96453202e-01 -6.22084379e-01
2.44502649e-01 3.78940463e-01 -3.59239101e-01 1.23002850e-01
4.01450485e-01 9.84986007e-01 -1.19678414e+00 -6.88418925e-01
3.15811187e-01 5.31761467e-01 -3.01175624e-01 -6.52684048e-02
-5.92873335e-01 1.82443112e-01 -3.63215417e-01 4.78488684e-01
2.80363351e-01 -2.94633597e-01 1.95918798e-01 -3.86503667e-01
-1.93670943e-01 5.04213631e-01 -8.20086300e-01 2.02714324e+00
-4.27946359e-01 9.16145086e-01 7.26120686e-03 -5.84680855e-01
3.52654636e-01 4.01558399e-01 1.72436073e-01 -7.81732142e-01
-5.00978380e-02 2.59037733e-01 -4.54524666e-01 -4.68157589e-01
7.32390106e-01 -1.04335777e-01 4.82031405e-02 9.05478746e-02
5.77444494e-01 -3.76226455e-01 7.50997007e-01 5.24109244e-01
9.52894688e-01 5.57927132e-01 4.26768899e-01 -1.33149564e-01
1.44829273e-01 4.27850515e-01 3.11805010e-01 7.88963079e-01
-2.19606414e-01 7.14080036e-01 2.40933061e-01 -2.40387693e-01
-1.26805639e+00 -1.14949882e+00 -1.75544798e-01 9.15178418e-01
5.15934750e-02 -8.50072026e-01 -8.16807687e-01 -7.52973557e-01
-1.34282276e-01 1.15919280e+00 -5.21535158e-01 2.58966535e-01
-2.17423543e-01 -3.54293078e-01 8.84587526e-01 6.41255200e-01
4.36343312e-01 -1.13171363e+00 -4.07987058e-01 4.22532186e-02
-3.10544997e-01 -1.65398860e+00 -6.52940571e-01 3.98631506e-02
-2.28115872e-01 -9.77813542e-01 -8.61485124e-01 -8.16868603e-01
3.53721976e-01 -4.46549058e-02 1.50746858e+00 4.81998874e-03
-4.47831064e-01 5.09525001e-01 -4.15661573e-01 -5.58664322e-01
-4.55491573e-01 1.80195123e-01 -7.95436725e-02 -4.71863806e-01
6.23325884e-01 1.17189279e-02 -4.22413319e-01 -1.34869531e-01
-9.33528960e-01 3.02480429e-01 5.33827841e-01 6.69226110e-01
8.34431469e-01 -5.15421748e-01 1.71847820e-01 -1.11574638e+00
5.37422955e-01 -1.76152453e-01 -7.27611959e-01 3.13883215e-01
-7.45329320e-01 1.03407994e-01 6.33707345e-02 3.84710953e-02
-1.07094097e+00 2.70890504e-01 -2.50108540e-01 -5.07760346e-01
-4.23007190e-01 2.87514776e-01 1.63156509e-01 3.95523906e-01
7.09787905e-01 -7.83220530e-02 -3.41415256e-01 -4.76049870e-01
8.32423091e-01 6.20347977e-01 7.87254095e-01 -7.41491497e-01
5.56103408e-01 3.72324675e-01 -5.73009849e-01 -7.28452623e-01
-1.05333531e+00 -5.40080667e-01 -5.82358122e-01 -1.20661438e-01
1.38664424e+00 -1.13893569e+00 -7.31719136e-02 1.82892516e-01
-1.50548875e+00 -5.40163398e-01 -2.99128354e-01 2.82027185e-01
-4.82095510e-01 1.93459496e-01 -6.44591749e-01 -6.28004193e-01
-4.36655372e-01 -1.25093269e+00 1.37206733e+00 1.05263919e-01
-7.01051474e-01 -9.72491860e-01 8.52318779e-02 4.78508115e-01
1.60665363e-01 1.32235467e-01 5.81147194e-01 -8.52668583e-01
-3.39536071e-01 8.11598003e-02 -4.09174263e-01 4.25418824e-01
-1.12257868e-01 8.48876312e-02 -1.11819601e+00 -1.32253751e-01
-7.56866932e-01 -5.88081598e-01 1.13294828e+00 3.82465899e-01
8.69366527e-01 -2.86281049e-01 -4.36777204e-01 3.67340952e-01
1.47799242e+00 1.67625118e-02 3.64296049e-01 4.80833232e-01
6.49296224e-01 4.53092247e-01 5.75740874e-01 1.30307987e-01
5.80226541e-01 6.71936750e-01 1.89054936e-01 -1.20003074e-01
-6.83062971e-01 -3.82064581e-01 3.25792968e-01 6.20986938e-01
-2.04442322e-01 -4.99161720e-01 -1.16878402e+00 1.04950559e+00
-1.67324591e+00 -8.97619486e-01 -1.36623317e-02 1.81759906e+00
1.09073937e+00 1.15161560e-01 -1.97915789e-02 -4.87940550e-01
7.17003167e-01 3.08372229e-01 -2.39800662e-01 -2.99118936e-01
-3.29008728e-01 5.07956505e-01 7.05423594e-01 9.69354153e-01
-1.28944838e+00 1.57305813e+00 6.52780771e+00 9.04686153e-01
-7.05878675e-01 3.24724108e-01 5.76152265e-01 -2.47906670e-01
-4.59109634e-01 4.32471111e-02 -7.67959297e-01 1.40618056e-01
8.39305699e-01 -7.53740296e-02 5.60063303e-01 5.71442485e-01
-3.37281115e-02 -2.42650136e-02 -9.56633270e-01 8.64082754e-01
1.82995826e-01 -1.37898517e+00 1.67909116e-01 -1.12713382e-01
8.22695732e-01 6.67390168e-01 1.77097246e-02 2.92895585e-01
7.19790995e-01 -1.10177422e+00 9.83737826e-01 4.59090471e-01
1.10052371e+00 -4.19864118e-01 5.71070671e-01 -2.46232629e-01
-8.98577034e-01 5.82546175e-01 -3.41252506e-01 3.09057951e-01
2.50197917e-01 5.22518098e-01 -1.09455419e+00 5.79003572e-01
7.85845578e-01 8.58193755e-01 -7.33577609e-01 6.75671041e-01
-6.59904242e-01 6.10114336e-01 -3.02312076e-01 8.52298737e-02
3.59509945e-01 9.31691453e-02 8.34026217e-01 1.74525201e+00
9.79354009e-02 -9.68753770e-02 2.62695193e-01 9.14509356e-01
-4.39634144e-01 1.48965806e-01 -9.10070360e-01 -3.59117895e-01
5.15126467e-01 1.25457382e+00 -5.31879365e-01 -7.13173807e-01
-4.18646634e-01 1.27453947e+00 2.93262988e-01 5.82579792e-01
-6.84319079e-01 -3.57324332e-01 6.77642167e-01 -1.16415463e-01
3.12557817e-01 -2.46818662e-01 -2.70800620e-01 -1.26753807e+00
-1.21427074e-01 -9.46604133e-01 4.80194569e-01 -1.27854824e+00
-1.38248956e+00 7.41311133e-01 8.04499462e-02 -8.27007651e-01
-3.35902095e-01 -9.39215124e-01 2.49476898e-02 7.34888613e-01
-1.50641716e+00 -1.10937715e+00 -3.15496773e-02 3.85945112e-01
8.41793895e-01 -1.32103324e-01 8.27109516e-01 3.46812814e-01
-2.52651453e-01 4.56538677e-01 -1.58266917e-01 3.94398272e-01
1.13357651e+00 -1.59077549e+00 7.46703804e-01 9.46171403e-01
6.58473134e-01 5.20786643e-01 8.45656574e-01 -8.36519957e-01
-7.31391668e-01 -1.23255777e+00 1.27912724e+00 -7.40708172e-01
8.36839020e-01 -3.00280869e-01 -5.55733979e-01 1.14307082e+00
9.85091507e-01 -1.27148479e-01 4.55842078e-01 -4.56862934e-02
-4.83018845e-01 2.64522076e-01 -8.50031912e-01 8.18569243e-01
1.16231763e+00 -7.87482500e-01 -8.10237467e-01 5.43619990e-01
8.67972434e-01 -6.66923761e-01 -6.22833192e-01 2.07728609e-01
3.68778110e-01 -4.62836266e-01 9.56398070e-01 -7.44088233e-01
7.55289555e-01 -4.80732948e-01 -4.55852270e-01 -1.34839129e+00
-1.21236883e-01 -4.16349828e-01 2.85648853e-01 1.19850636e+00
8.22981894e-01 -2.49471858e-01 1.97062820e-01 2.48310164e-01
-2.21613064e-01 -4.55965698e-01 -7.28845954e-01 -5.99150360e-01
1.15373068e-01 -7.69227326e-01 1.73787579e-01 1.09771979e+00
-2.39124462e-01 5.70210576e-01 -1.48323670e-01 -6.38994714e-03
5.78162611e-01 -1.61338151e-01 3.70965689e-01 -6.03702247e-01
-3.78307074e-01 -4.38019663e-01 -2.90522903e-01 -8.43797743e-01
3.26576769e-01 -1.17647815e+00 2.41974249e-01 -1.90442681e+00
1.94071785e-01 9.37592685e-02 -4.43424612e-01 1.02882528e+00
-2.49209017e-01 6.28839135e-01 3.98158699e-01 7.35690966e-02
-9.00076509e-01 1.18167698e-01 1.19956458e+00 -3.94409895e-01
3.57006192e-01 -6.01405382e-01 -7.57534981e-01 6.53320611e-01
7.99017549e-01 -3.31308752e-01 -2.51842588e-01 -9.98813212e-01
1.10045180e-01 -3.53491008e-01 5.34240007e-01 -7.01237679e-01
5.60462251e-02 2.79592406e-02 4.65587497e-01 -3.87112170e-01
2.75440574e-01 -3.60656470e-01 -6.61232769e-02 2.38544852e-01
-5.58986127e-01 1.63836777e-01 4.12474513e-01 2.29237974e-01
-2.79296547e-01 -1.52102858e-01 5.97602665e-01 -4.43781704e-01
-1.09922779e+00 1.43502742e-01 -4.48667139e-01 4.48366463e-01
6.94470525e-01 2.48544514e-01 -5.56011081e-01 -4.34384704e-01
-9.24734831e-01 4.00411010e-01 6.34629786e-01 5.41796982e-01
4.24335927e-01 -1.30455256e+00 -8.86614382e-01 -3.08603317e-01
3.73662382e-01 -3.74002568e-02 -2.26677492e-01 7.32270181e-01
-6.46482348e-01 5.19281447e-01 -5.62990718e-02 -3.34835589e-01
-1.14890885e+00 4.28205580e-01 2.27748603e-01 -1.85659185e-01
-5.25305510e-01 9.61356223e-01 1.37968972e-01 -3.47242147e-01
5.16627505e-02 -1.58117056e-01 -6.87451884e-02 6.53542355e-02
5.18406451e-01 -2.06616297e-02 7.47317597e-02 -6.64399207e-01
-4.75110203e-01 3.69958669e-01 -1.91142529e-01 -6.42888665e-01
1.18920588e+00 -1.69538721e-01 6.50351048e-02 3.84698033e-01
1.06818509e+00 1.24910392e-01 -1.04148269e+00 -3.02727550e-01
2.22309902e-01 -1.75227582e-01 7.23681822e-02 -1.22533774e+00
-9.11522329e-01 8.03630769e-01 4.52316880e-01 -1.69126987e-01
1.06092644e+00 3.08351517e-01 7.57485628e-01 4.58304375e-01
1.27512112e-01 -1.19597840e+00 -3.02134156e-01 7.49794424e-01
1.01454496e+00 -1.20644379e+00 -8.12038258e-02 -2.58784533e-01
-9.01560366e-01 8.98434758e-01 5.42320192e-01 -1.45346612e-01
1.35856792e-01 3.38256687e-01 3.97602409e-01 -2.59603232e-01
-7.44884193e-01 -8.29029202e-01 5.10927320e-01 7.40064323e-01
7.43503690e-01 1.57517448e-01 -4.50827122e-01 4.64091748e-01
-3.99739891e-01 7.65111744e-02 1.87371880e-01 7.47480750e-01
-1.59613878e-01 -1.15162075e+00 -6.21632710e-02 3.21718156e-01
-4.67881441e-01 -6.29555821e-01 -7.79304266e-01 9.64580357e-01
2.24909067e-01 7.53022015e-01 1.01198107e-01 -3.91669482e-01
2.67562330e-01 3.23397756e-01 6.08797193e-01 -7.24061370e-01
-2.81859487e-01 1.70618162e-01 7.49401450e-01 -6.94942653e-01
-5.97706139e-01 -7.55901039e-01 -1.34177208e+00 -9.39074717e-03
2.38839999e-01 -4.69740704e-02 5.98920941e-01 9.57085669e-01
2.53188342e-01 4.68539536e-01 -2.20245421e-01 -5.64430058e-01
-1.09908469e-01 -9.26998019e-01 -2.33681664e-01 6.25259459e-01
2.21701235e-01 -4.62415785e-01 6.49914816e-02 5.68482161e-01]
|
[10.790388107299805, 1.5670826435089111]
|
a8ba1f28-0815-4cf9-9474-c7c8396a748d
|
self-supervision-versus-synthetic-datasets
|
2204.11493
| null |
https://arxiv.org/abs/2204.11493v1
|
https://arxiv.org/pdf/2204.11493v1.pdf
|
Self-supervision versus synthetic datasets: which is the lesser evil in the context of video denoising?
|
Supervised training has led to state-of-the-art results in image and video denoising. However, its application to real data is limited since it requires large datasets of noisy-clean pairs that are difficult to obtain. For this reason, networks are often trained on realistic synthetic data. More recently, some self-supervised frameworks have been proposed for training such denoising networks directly on the noisy data without requiring ground truth. On synthetic denoising problems supervised training outperforms self-supervised approaches, however in recent years the gap has become narrower, especially for video. In this paper, we propose a study aiming to determine which is the best approach to train denoising networks for real raw videos: supervision on synthetic realistic data or self-supervision on real data. A complete study with quantitative results in case of natural videos with real motion is impossible since no dataset with clean-noisy pairs exists. We address this issue by considering three independent experiments in which we compare the two frameworks. We found that self-supervision on the real data outperforms supervision on synthetic data, and that in normal illumination conditions the drop in performance is due to the synthetic ground truth generation, not the noise model.
|
['Pablo Arias', 'Gabriele Facciolo', 'Aranud Barral', 'Valéry Dewil']
|
2022-04-25
| null | null | null | null |
['video-denoising']
|
['computer-vision']
|
[ 3.66104454e-01 -9.05573368e-02 3.79178554e-01 -5.21574378e-01
-8.44482422e-01 -2.67045557e-01 6.13945425e-01 -4.62111682e-02
-6.75626099e-01 9.97935593e-01 1.26657009e-01 2.08414346e-01
-7.87728466e-03 -6.91171885e-01 -8.30793798e-01 -1.00176740e+00
9.67544410e-03 2.80086249e-01 1.07060425e-01 -2.79174924e-01
-6.44804835e-02 6.65934160e-02 -1.90518606e+00 2.86552995e-01
8.15989196e-01 7.53861189e-01 4.83842522e-01 6.12967551e-01
1.02787383e-01 5.27731538e-01 -7.73863018e-01 -2.17162117e-01
5.46341538e-01 -6.67371154e-01 -3.40781301e-01 3.18940490e-01
6.48382246e-01 -2.17352882e-01 -1.23338684e-01 1.30439460e+00
6.76489472e-01 3.25207002e-02 3.38164508e-01 -1.10046601e+00
-3.52659106e-01 3.47837985e-01 -3.03890467e-01 2.65586134e-02
3.23101461e-01 3.01865667e-01 5.68244755e-01 -5.83448410e-01
8.68632197e-01 1.12577891e+00 8.42298269e-01 6.57264590e-01
-1.49353337e+00 -4.00675714e-01 -2.66215503e-01 2.38903150e-01
-1.07480395e+00 -6.48145318e-01 8.89679670e-01 -2.66285807e-01
3.05939704e-01 6.41633943e-02 4.16768193e-01 1.68576944e+00
-1.88479394e-01 5.21826267e-01 1.55569959e+00 -5.71633577e-01
3.12759012e-01 2.76329488e-01 -8.44863504e-02 1.32027954e-01
4.46178168e-01 3.59614193e-01 -5.20229459e-01 1.68763250e-01
6.59597635e-01 -3.60567957e-01 -6.70671582e-01 -3.82004857e-01
-1.06791055e+00 6.90129101e-01 1.69697031e-01 8.30141544e-01
-5.38483322e-01 2.21851673e-02 6.18407726e-01 7.48955011e-01
6.35284960e-01 3.33212823e-01 -3.01908761e-01 -1.91780061e-01
-1.28263056e+00 2.01619670e-01 8.92879486e-01 5.14331877e-01
6.15004957e-01 3.76402348e-01 2.53531694e-01 8.24799001e-01
6.67294934e-02 4.86680955e-01 5.37171781e-01 -1.05337059e+00
4.64287937e-01 -6.57635406e-02 2.30119705e-01 -1.06261504e+00
-8.38339999e-02 -5.17301083e-01 -1.12207377e+00 4.99973565e-01
8.47635448e-01 -3.20696563e-01 -8.55647922e-01 1.77333570e+00
2.50364118e-03 1.65909663e-01 3.23080271e-01 1.10921001e+00
7.45677292e-01 6.62142456e-01 -2.46909961e-01 -5.31121850e-01
6.91108882e-01 -7.44560480e-01 -1.08569896e+00 -7.93296769e-02
4.97171938e-01 -9.55682576e-01 8.98335516e-01 1.05873561e+00
-9.53781009e-01 -9.17950988e-01 -1.00967121e+00 3.50106865e-01
-2.93482155e-01 2.51106411e-01 5.33930166e-03 7.28365362e-01
-1.15038323e+00 9.97670114e-01 -7.59419203e-01 -6.61809206e-01
2.30775073e-01 1.62136838e-01 -7.33994603e-01 -3.09145361e-01
-1.22355831e+00 9.27235425e-01 3.33596408e-01 4.08347964e-01
-1.18553293e+00 -2.17331871e-01 -8.79608572e-01 -2.42917284e-01
4.34579134e-01 -4.24957544e-01 9.25201297e-01 -1.58859539e+00
-1.40460467e+00 7.67751873e-01 9.64826643e-02 -6.92387760e-01
8.70139837e-01 -3.58980954e-01 -5.00299811e-01 1.52260497e-01
7.11246878e-02 4.22600538e-01 1.18472636e+00 -1.86529315e+00
-1.41597688e-01 -2.39475369e-01 -8.63768607e-02 -2.00417742e-01
-1.90612569e-01 -2.34106317e-01 -1.05362549e-01 -8.29876125e-01
3.93170379e-02 -7.32363760e-01 -2.76970565e-01 -9.13881958e-02
-2.19569311e-01 2.17044130e-01 9.88575757e-01 -6.25045121e-01
7.57325888e-01 -2.25422263e+00 1.24183096e-01 8.43754336e-02
-1.37276530e-01 6.13149464e-01 -3.45034689e-01 4.92820680e-01
-3.90821904e-01 7.31128454e-03 -4.23601031e-01 -6.12993062e-01
-3.40569437e-01 5.89998722e-01 -5.67291006e-02 4.80619699e-01
1.35400712e-01 2.16310784e-01 -1.01797628e+00 -5.25956631e-01
2.94440687e-01 6.84569001e-01 -2.50022620e-01 3.31390291e-01
-9.19207111e-02 7.91382313e-01 -4.44603860e-02 3.20274115e-01
7.83348918e-01 3.37203592e-01 8.97295401e-02 -5.01005352e-01
-4.23060767e-02 -2.02335984e-01 -1.51706362e+00 1.77368903e+00
-4.80311900e-01 8.94369304e-01 4.92767125e-01 -1.40208411e+00
1.01301622e+00 6.56881928e-01 3.80154699e-01 -6.56489432e-01
2.20223770e-01 4.11592096e-01 -5.03404438e-02 -8.68895888e-01
2.03849211e-01 -4.40174162e-01 4.67308164e-01 3.89421470e-02
2.89433569e-01 -2.04061329e-01 6.29909694e-01 -6.92430437e-02
1.01879752e+00 4.36551780e-01 -7.09244981e-02 -3.07514071e-01
6.49290085e-01 -2.24980786e-02 5.99588335e-01 8.55889738e-01
-6.77870363e-02 1.09256113e+00 4.57668334e-01 -3.46838325e-01
-1.07817578e+00 -7.12667942e-01 -6.32150620e-02 3.73953402e-01
6.09354563e-02 -2.64427751e-01 -1.04032230e+00 -5.32711923e-01
-4.87734824e-01 5.50500572e-01 -5.95094502e-01 1.63818896e-01
-5.39255083e-01 -8.31738412e-01 4.15015161e-01 1.54921472e-01
6.54696524e-01 -1.11044765e+00 -4.03122663e-01 2.80374438e-01
-2.94892192e-01 -1.35852683e+00 1.97391450e-01 2.96479970e-01
-1.12489438e+00 -1.18427956e+00 -7.94270217e-01 -5.98468006e-01
6.61093295e-01 2.69301027e-01 1.21537602e+00 2.23235026e-01
-3.74002419e-02 3.13627481e-01 -6.12107635e-01 -1.66883200e-01
-7.29141891e-01 -3.43885154e-01 1.00441009e-01 2.54830241e-01
1.98358507e-03 -7.70242810e-01 -4.58480775e-01 4.82571632e-01
-1.44930482e+00 -2.73182243e-01 4.88994300e-01 1.07516706e+00
4.08385456e-01 3.29934150e-01 5.15703559e-01 -9.39161777e-01
5.53255796e-01 -2.80846685e-01 -5.86255014e-01 -9.37300250e-02
-3.18231255e-01 7.97698349e-02 9.21149671e-01 -2.55611688e-01
-1.21270633e+00 1.11523613e-01 -4.65782821e-01 -4.35917020e-01
-6.70410275e-01 4.95999515e-01 -1.64656192e-01 -4.98122685e-02
1.00180662e+00 3.23214754e-02 2.16622129e-01 -7.20456839e-01
-5.62566891e-02 5.29453933e-01 5.89990437e-01 -4.61558759e-01
1.01497579e+00 5.83583593e-01 6.21915981e-03 -1.20861316e+00
-7.54855871e-01 -3.87940496e-01 -6.00736797e-01 -2.28713006e-01
9.22442853e-01 -8.59164953e-01 -2.69324705e-02 6.65077984e-01
-1.13901472e+00 -4.59686875e-01 -4.37446386e-01 6.67466640e-01
-5.66721320e-01 6.90429151e-01 -5.71246982e-01 -7.43559122e-01
-4.23791818e-02 -1.30622816e+00 8.22937489e-01 -1.76401194e-02
1.18090317e-01 -1.03719401e+00 1.65256634e-01 4.19724822e-01
4.58716035e-01 5.10408521e-01 3.16015959e-01 -4.17081803e-01
-3.80876005e-01 -3.40403736e-01 -1.17299231e-02 1.21639013e+00
2.11483732e-01 -5.51135987e-02 -9.75608945e-01 -3.33338380e-01
5.18628836e-01 -4.99716789e-01 8.70487034e-01 4.03940290e-01
9.74774420e-01 1.55420169e-01 1.84791729e-01 3.62742990e-01
1.70902503e+00 -1.71634406e-01 9.35785353e-01 3.35003763e-01
4.70142573e-01 7.22940862e-01 7.93281019e-01 2.89908070e-02
-2.68028051e-01 5.90186477e-01 7.06818402e-01 -3.34107101e-01
-2.83131927e-01 5.84676713e-02 5.51227748e-01 6.81885242e-01
-3.12895864e-01 -4.95067686e-01 -6.91044509e-01 6.66570842e-01
-1.61058569e+00 -9.51593220e-01 -5.64000845e-01 2.43667555e+00
6.26896501e-01 3.56133163e-01 -4.90841866e-02 6.23385191e-01
6.49722159e-01 8.16840455e-02 2.59655248e-02 -2.74376553e-02
-4.48065728e-01 1.04024738e-01 2.75613666e-01 5.62934995e-01
-1.10506046e+00 4.04703081e-01 5.59255266e+00 7.66651154e-01
-1.20006037e+00 3.60492438e-01 6.05402350e-01 1.15039416e-01
5.61399870e-02 -8.01045895e-02 -3.09251398e-01 5.31513870e-01
8.20735037e-01 4.91442949e-01 1.02289759e-01 5.16427517e-01
7.16695607e-01 -4.32254136e-01 -1.01927471e+00 1.11141384e+00
1.06297426e-01 -9.65938509e-01 -8.66996348e-02 -6.13067448e-02
9.41325724e-01 -1.79098472e-02 -3.51518691e-01 -7.87830949e-02
-4.26927917e-02 -9.22531784e-01 3.40358615e-01 6.91510916e-01
3.80718291e-01 -4.75560874e-01 1.30148447e+00 5.68332791e-01
-5.73701739e-01 1.58887044e-01 -3.09284836e-01 5.48100360e-02
3.97500545e-01 1.03671551e+00 -3.31524879e-01 7.66207099e-01
7.74746776e-01 7.95316339e-01 -5.42293072e-01 1.13942599e+00
-3.19192499e-01 8.87316287e-01 -3.22290480e-01 3.87099624e-01
2.56806403e-01 -4.81573045e-01 6.93546832e-01 1.21610498e+00
4.65725601e-01 -2.11598366e-01 -9.49716866e-02 5.77472627e-01
2.80229062e-01 -1.00411857e-02 -7.84021497e-01 1.42872050e-01
-2.65501648e-01 1.20409739e+00 -7.41490781e-01 -3.25920314e-01
-3.57599914e-01 9.20027614e-01 -2.45045304e-01 4.88816142e-01
-5.77695251e-01 -2.04220504e-01 2.21204862e-01 3.54746521e-01
2.08706722e-01 -3.43421429e-01 -6.11451417e-02 -1.19362116e+00
7.44506866e-02 -1.14304769e+00 8.42532218e-02 -9.82678652e-01
-1.38547671e+00 9.35291648e-01 1.14749730e-01 -1.39337957e+00
-3.65144819e-01 -6.76605761e-01 -5.92116654e-01 6.66071773e-01
-1.31150913e+00 -7.07083225e-01 -6.65025532e-01 5.23576200e-01
5.83899796e-01 -5.58121279e-02 5.55103660e-01 7.24603057e-01
-3.65000010e-01 1.13257676e-01 4.34556425e-01 1.40543416e-01
1.01795506e+00 -1.12710655e+00 -6.89965189e-02 1.09208477e+00
2.66895860e-01 2.28222266e-01 1.28625441e+00 -4.50130284e-01
-1.08505893e+00 -6.88728869e-01 5.79611540e-01 -4.90374714e-02
4.77716744e-01 -2.68682599e-01 -1.22250926e+00 1.90813258e-01
5.03419042e-01 2.79405475e-01 2.61761010e-01 -2.58484811e-01
4.44412343e-02 -4.68656987e-01 -1.22446442e+00 2.46455297e-01
8.68099749e-01 -1.38041854e-01 -4.49381858e-01 3.43957782e-01
2.50732750e-01 -7.61679336e-02 -8.40063214e-01 4.62845713e-01
7.34780282e-02 -1.56129825e+00 7.63914645e-01 -1.38371497e-01
5.88630199e-01 -5.66724300e-01 -1.28250033e-01 -1.61317301e+00
3.76860976e-01 -4.15206134e-01 3.07512015e-01 1.30066574e+00
1.87007681e-01 -4.08394873e-01 8.17519724e-01 -5.59204035e-02
3.06731220e-02 -2.95502961e-01 -7.98321009e-01 -9.02175486e-01
-2.16940388e-01 -5.16392350e-01 2.65091825e-02 9.00896251e-01
-6.46065652e-01 1.17535979e-01 -6.75593257e-01 -9.43368077e-02
9.03617322e-01 -1.72458306e-01 9.58399653e-01 -1.13403428e+00
-3.15977752e-01 5.83364144e-02 -6.57929778e-01 -7.30365396e-01
7.07200617e-02 -2.73288012e-01 4.23773795e-01 -1.46415603e+00
-1.85995921e-01 -2.05440253e-01 1.26860097e-01 1.36301994e-01
4.32146490e-02 6.59653902e-01 8.76391679e-02 8.43872726e-02
-3.90772790e-01 5.69205821e-01 1.25467420e+00 1.03032058e-02
5.15813269e-02 -1.76284518e-02 -1.94026306e-01 8.11549067e-01
9.44855094e-01 -5.97686529e-01 -4.96845186e-01 -5.86330295e-01
1.01518810e-01 6.93543106e-02 5.31580448e-01 -1.46459627e+00
5.22709191e-02 3.19265485e-01 2.05674499e-01 -2.73199618e-01
3.70594025e-01 -9.83709753e-01 2.58334517e-01 3.91234070e-01
-6.33047670e-02 -7.47448057e-02 -1.93776041e-02 6.80169046e-01
-8.20566297e-01 -7.37814784e-01 8.92178118e-01 -3.26825052e-01
-5.90851367e-01 -2.21939087e-01 -3.80471170e-01 3.09334248e-02
7.03014195e-01 -4.74480718e-01 1.88262120e-01 -8.49434912e-01
-1.06585085e+00 -2.18335152e-01 5.06256759e-01 1.35078385e-01
4.38587725e-01 -1.10226989e+00 -9.31041181e-01 1.19259059e-01
-2.22835734e-01 1.14797510e-01 2.06832528e-01 8.72399092e-01
-6.69767559e-01 -1.09516099e-01 -3.01299155e-01 -7.80895710e-01
-1.40586901e+00 5.38355410e-01 3.31416547e-01 -1.96980774e-01
-2.75178134e-01 3.91439438e-01 -1.47245377e-01 -2.87669390e-01
3.63442004e-01 -2.34275430e-01 -1.48252830e-01 1.25588924e-01
2.71748394e-01 3.88981789e-01 2.91671544e-01 -7.75119603e-01
1.57936603e-01 4.96329278e-01 3.21445853e-01 -1.86000198e-01
1.57427835e+00 -3.47143300e-02 -1.78478152e-01 4.48619783e-01
1.22430408e+00 -1.12382621e-01 -1.31079638e+00 -5.87420445e-03
3.69816720e-02 -5.80393553e-01 1.26785219e-01 -4.73014623e-01
-1.42147791e+00 8.33718240e-01 8.87700379e-01 5.29019415e-01
1.42993295e+00 -4.80459809e-01 5.07883847e-01 4.85015422e-01
3.98220181e-01 -1.14940441e+00 1.74029335e-01 2.42070556e-01
1.00747669e+00 -1.56317556e+00 -1.22814044e-01 -5.54881573e-01
-3.52219790e-01 1.18244255e+00 4.28175926e-01 -3.73482615e-01
6.43817008e-01 2.28090674e-01 4.08651620e-01 -1.13087714e-01
-3.08614522e-01 -2.96594024e-01 -1.06966086e-01 7.22552180e-01
4.81081009e-01 -4.45768774e-01 -6.07498944e-01 4.09434587e-02
2.46244743e-02 3.43421817e-01 7.32153058e-01 9.24963713e-01
-1.32668227e-01 -1.56472647e+00 -8.61203134e-01 3.63394707e-01
-6.40713871e-01 1.05787493e-01 -1.58725604e-01 9.56127524e-01
4.74366814e-01 1.30551827e+00 -2.96634763e-01 -1.36776745e-01
4.39757079e-01 -1.90521106e-01 5.67053914e-01 -4.19111997e-01
-5.25685191e-01 1.44742206e-01 3.38061452e-01 -5.32347739e-01
-1.08490968e+00 -6.96443558e-01 -7.14791119e-01 9.41644423e-03
-4.37603652e-01 5.11471868e-01 8.21919382e-01 8.34170878e-01
-2.00694874e-02 3.68585140e-01 4.68569785e-01 -1.31406057e+00
-5.25767863e-01 -1.13546574e+00 -5.92208803e-01 9.24381793e-01
4.87281591e-01 -5.52495778e-01 -6.78212404e-01 3.75629991e-01]
|
[11.478289604187012, -2.402952194213867]
|
0ec83662-3c0a-4506-a117-f1a4aeddcceb
|
1st-place-solution-for-youtubevos-challenge-1
|
2212.14679
| null |
https://arxiv.org/abs/2212.14679v1
|
https://arxiv.org/pdf/2212.14679v1.pdf
|
1st Place Solution for YouTubeVOS Challenge 2022: Referring Video Object Segmentation
|
The task of referring video object segmentation aims to segment the object in the frames of a given video to which the referring expressions refer. Previous methods adopt multi-stage approach and design complex pipelines to obtain promising results. Recently, the end-to-end method based on Transformer has proved its superiority. In this work, we draw on the advantages of the above methods to provide a simple and effective pipeline for RVOS. Firstly, We improve the state-of-the-art one-stage method ReferFormer to obtain mask sequences that are strongly correlated with language descriptions. Secondly, based on a reliable and high-quality keyframe, we leverage the superior performance of video object segmentation model to further enhance the quality and temporal consistency of the mask results. Our single model reaches 70.3 J &F on the Referring Youtube-VOS validation set and 63.0 on the test set. After ensemble, we achieve 64.1 on the final leaderboard, ranking 1st place on CVPR2022 Referring Youtube-VOS challenge. Code will be available at https://github.com/Zhiweihhh/cvpr2022-rvos-challenge.git.
|
['Jinfeng Bai', 'Zhilong Ji', 'Yuan Gao', 'Bo Chen', 'Zhiwei Hu']
|
2022-12-27
| null | null | null | null |
['referring-video-object-segmentation', 'video-object-segmentation', 'video-semantic-segmentation']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[-1.70554593e-01 -2.03353390e-01 -4.87370938e-01 -2.43029296e-01
-1.14423287e+00 -5.51233292e-01 3.59529912e-01 -5.02123475e-01
-2.71453679e-01 2.69337654e-01 1.26821488e-01 -7.83714354e-02
2.60448098e-01 -1.33683488e-01 -6.78529024e-01 -1.20332502e-01
3.26103568e-01 1.07108660e-01 5.12637258e-01 -1.49247617e-01
1.18977226e-01 1.29103318e-01 -1.38923824e+00 4.82667059e-01
6.78037345e-01 1.31425989e+00 3.89365852e-01 6.42425537e-01
-1.46263540e-01 1.00388658e+00 -4.31808770e-01 -5.22747755e-01
2.49637663e-01 -6.30370557e-01 -9.60893512e-01 1.47151276e-01
7.23828316e-01 -2.72258997e-01 -6.02912009e-01 9.90393400e-01
5.13782144e-01 5.98018318e-02 1.71185747e-01 -1.39290214e+00
-6.06023252e-01 7.33748913e-01 -5.12997448e-01 6.57226205e-01
4.70571160e-01 4.82660860e-01 1.31605637e+00 -1.07174885e+00
9.46015835e-01 1.06902063e+00 3.60626221e-01 5.97202182e-01
-7.85315037e-01 -8.53182852e-01 3.89036804e-01 6.06969655e-01
-1.59331834e+00 -8.05668831e-01 6.57007754e-01 -4.85863149e-01
8.69181991e-01 3.66757661e-01 8.07616174e-01 1.09700537e+00
-2.45103016e-01 1.41246760e+00 8.52144718e-01 5.61160482e-02
-1.95726022e-01 -4.56983224e-02 -5.54634929e-02 6.57238841e-01
-4.04761076e-01 -2.01234058e-01 -9.24224496e-01 5.28143942e-01
4.44140851e-01 -1.60756424e-01 -6.23456836e-01 1.72603592e-01
-1.32481909e+00 4.26830769e-01 4.66450751e-01 3.51617336e-01
-4.11850661e-01 3.35377157e-01 4.55842674e-01 8.35697874e-02
6.11111403e-01 2.72244513e-01 -3.46221685e-01 -6.17272615e-01
-1.47729528e+00 2.18067944e-01 4.99782771e-01 1.48238695e+00
2.55998582e-01 5.65010644e-02 -6.11335635e-01 7.98286140e-01
4.21043634e-01 4.44791704e-01 1.10445753e-01 -1.28007102e+00
5.50852656e-01 3.20623040e-01 -1.35401428e-01 -5.95550835e-01
-1.57062209e-03 -5.09939253e-01 -3.97096783e-01 -2.64623612e-01
2.15415731e-01 1.94769695e-01 -9.53431487e-01 1.43685341e+00
1.72919795e-01 3.93562049e-01 -3.21202546e-01 1.24406612e+00
1.55996203e+00 6.91655636e-01 1.22245461e-01 -1.02306101e-02
1.49128497e+00 -1.42148721e+00 -8.84153783e-01 -1.51514173e-01
3.79202694e-01 -8.94422948e-01 1.10712600e+00 3.91957045e-01
-1.41550016e+00 -8.62217605e-01 -8.35438311e-01 -2.80161262e-01
1.21879034e-01 2.29075417e-01 3.84864360e-01 2.77127326e-01
-1.13475585e+00 5.24857700e-01 -7.09350765e-01 -2.69019961e-01
8.69947731e-01 1.57346621e-01 -1.05784670e-01 -5.21940589e-02
-1.11482334e+00 4.31645930e-01 1.84680715e-01 1.02483578e-01
-1.02736735e+00 -8.41338098e-01 -5.90821803e-01 -1.11479536e-01
7.00732470e-01 -6.32679403e-01 1.54328191e+00 -9.15861726e-01
-1.39658904e+00 1.15249574e+00 -5.33908188e-01 -4.02114064e-01
8.58528793e-01 -4.28550392e-01 -3.46119553e-01 5.48217058e-01
4.31505084e-01 1.20623600e+00 7.63904214e-01 -9.61504757e-01
-1.03627014e+00 1.83431491e-01 1.06163479e-01 4.15602699e-02
4.25571240e-02 3.56581897e-01 -1.51229191e+00 -5.24742186e-01
9.93171930e-02 -8.60134482e-01 2.05906063e-01 -1.32335484e-01
-6.06140673e-01 -5.04997134e-01 7.58178592e-01 -8.59324336e-01
1.49407482e+00 -2.48596215e+00 2.93167233e-01 -3.10707808e-01
5.02964556e-01 2.09746718e-01 -6.69668540e-02 1.17241368e-01
2.81227436e-02 4.70420301e-01 9.50820222e-02 -4.96730685e-01
-2.21161190e-02 -1.35727853e-01 -2.70612150e-01 5.17410517e-01
4.24110293e-01 1.26126635e+00 -8.97779644e-01 -7.40540564e-01
9.25724134e-02 3.03759336e-01 -4.02085036e-01 3.57369870e-01
-2.32665271e-01 7.76391864e-01 -4.78569776e-01 9.17987883e-01
3.79992157e-01 -4.78539199e-01 -2.81243861e-01 -4.33927983e-01
-1.59910604e-01 4.03153658e-01 -8.36603165e-01 1.92469847e+00
-1.84000060e-01 1.06255960e+00 -5.81056811e-02 -5.71897686e-01
7.28466272e-01 4.12750095e-01 6.52744710e-01 -8.06658864e-01
2.35227361e-01 3.18815708e-01 -1.25020266e-01 -7.08125293e-01
5.43964446e-01 2.29311273e-01 8.96306932e-02 -8.74135867e-02
2.25571051e-01 -2.59863526e-01 6.95914745e-01 5.60885847e-01
8.59458268e-01 7.17514157e-01 -8.80551189e-02 -1.34965435e-01
6.20049953e-01 -1.52826682e-02 7.14182556e-01 5.99067688e-01
-5.63462317e-01 9.83741105e-01 6.41907454e-01 -3.43258306e-02
-9.40775096e-01 -8.35427046e-01 -8.56114104e-02 9.45638895e-01
3.02558243e-01 -8.36393595e-01 -8.04958582e-01 -6.74602985e-01
-3.55691612e-01 7.32944012e-01 -4.49796915e-01 2.09549665e-01
-6.48940623e-01 -7.84955826e-03 6.46486044e-01 5.65332949e-01
7.31581032e-01 -1.02887654e+00 -4.76029456e-01 1.15842737e-01
-6.94573402e-01 -1.73648214e+00 -7.40483999e-01 -3.06899518e-01
-5.77791035e-01 -1.01402462e+00 -9.03425992e-01 -6.69485092e-01
1.39214814e-01 3.33496928e-01 1.36959267e+00 1.53428018e-01
-2.18832083e-02 4.51852888e-01 -6.36590123e-01 -6.97653219e-02
-2.47518152e-01 1.76893562e-01 -4.12357509e-01 -9.59440134e-03
3.32887173e-01 -1.16664723e-01 -6.98772788e-01 4.79058385e-01
-6.15083516e-01 3.99983674e-01 2.41005704e-01 4.97523725e-01
8.49964917e-01 -4.33849901e-01 3.94429088e-01 -5.63434124e-01
1.00043416e-01 -4.90883052e-01 -5.71565509e-01 1.17548004e-01
-4.41137105e-01 -4.40486938e-01 3.65381002e-01 -2.99641252e-01
-6.80193663e-01 -4.38962057e-02 -3.23299319e-01 -9.18075144e-01
2.74105966e-02 2.45711386e-01 -1.50779381e-01 2.76355445e-01
2.52924651e-01 2.33448148e-01 -3.52143198e-01 -4.42334503e-01
4.67006564e-01 8.02103579e-01 6.00346386e-01 -4.13044691e-01
5.64560592e-01 4.43597227e-01 -2.02610120e-01 -6.65978312e-01
-9.83630061e-01 -8.20255399e-01 -4.60449755e-01 -6.39780462e-01
1.14719248e+00 -1.37230086e+00 -7.47576952e-01 3.88064802e-01
-1.20304334e+00 -5.83839417e-01 -2.40967512e-01 3.32618415e-01
-5.70340276e-01 3.30530196e-01 -6.57539010e-01 -5.41614711e-01
-4.07760173e-01 -1.51347876e+00 1.22535825e+00 3.24271142e-01
-3.36498797e-01 -4.93881345e-01 -4.59426045e-01 7.01563895e-01
3.32320809e-01 -1.10502824e-01 4.84240837e-02 -5.55282891e-01
-1.01034582e+00 7.65914097e-02 -4.26277578e-01 2.89801925e-01
-2.39277765e-01 3.49846810e-01 -1.00294018e+00 -1.23244442e-01
-2.28756323e-01 -2.29748428e-01 1.01309466e+00 4.44373280e-01
1.20054364e+00 1.77078635e-01 -1.91654682e-01 8.80679011e-01
1.09035957e+00 1.83619827e-01 5.81750751e-01 4.14038867e-01
9.83876050e-01 4.39629972e-01 1.01311636e+00 3.40755880e-01
6.65620029e-01 1.00741529e+00 3.94563258e-01 -7.37198256e-03
-6.50925338e-01 -1.52033135e-01 4.53191876e-01 9.34238493e-01
-9.11564380e-02 -4.93424445e-01 -9.15020585e-01 7.34144747e-01
-1.84765422e+00 -9.49956000e-01 -4.80753630e-01 1.68226707e+00
6.38681710e-01 3.15027177e-01 1.96741700e-01 -2.37010464e-01
6.65485799e-01 4.46739554e-01 -4.80771363e-01 9.75086689e-02
-2.82402039e-01 -4.09515202e-02 2.97224671e-01 3.17338198e-01
-1.10293865e+00 1.42079866e+00 4.89177608e+00 1.10306513e+00
-1.17152834e+00 4.29759085e-01 7.56029069e-01 -4.76623684e-01
-1.73850119e-01 -3.45699824e-02 -1.10605800e+00 5.67856669e-01
1.00941503e+00 -2.10389658e-03 4.05293912e-01 5.28846860e-01
5.43234646e-01 -9.08170044e-02 -1.12362492e+00 1.37229562e+00
1.01637214e-01 -1.49770606e+00 -2.90848643e-01 -2.47003764e-01
7.14011073e-01 4.22132760e-01 1.77376181e-01 3.53278607e-01
-2.55930603e-01 -8.57612193e-01 1.40630627e+00 4.16269809e-01
1.03219557e+00 -4.76009160e-01 5.61247408e-01 -1.13838337e-01
-1.47296917e+00 2.28519425e-01 1.14821449e-01 3.32117110e-01
6.31835520e-01 1.98899850e-01 -5.90208828e-01 5.83888769e-01
1.03435302e+00 1.20831025e+00 -6.17473423e-01 1.22922683e+00
-6.02099001e-01 7.63937712e-01 -3.22497606e-01 1.26940221e-01
4.71094340e-01 3.47847939e-02 6.14814222e-01 1.42115235e+00
1.35098204e-01 1.54265463e-01 1.49906665e-01 9.94212091e-01
-4.26410019e-01 1.37010783e-01 -2.02814505e-01 -1.87516466e-01
3.93394113e-01 1.39444005e+00 -8.74859214e-01 -5.49336433e-01
-6.00125074e-01 1.01694703e+00 -2.81808432e-02 4.19467360e-01
-1.42037344e+00 -1.03750769e-02 5.00307024e-01 1.28747702e-01
7.12544382e-01 -1.69296652e-01 -1.15377583e-01 -1.21847034e+00
1.96516648e-01 -1.01663446e+00 2.42339000e-01 -9.83765364e-01
-9.62205291e-01 9.72507000e-01 2.10883282e-02 -1.52860713e+00
-1.42406635e-02 -3.55757058e-01 -3.84694606e-01 8.63600075e-01
-1.59323263e+00 -1.26153016e+00 -5.23244321e-01 4.84747380e-01
1.16066647e+00 -2.20191196e-01 1.83195502e-01 6.70134366e-01
-7.36925960e-01 6.01358831e-01 -3.71641070e-01 2.34608665e-01
7.67628849e-01 -9.46703672e-01 6.19790435e-01 1.20135224e+00
3.93848687e-01 3.52932513e-01 5.38664401e-01 -6.83605015e-01
-1.33329535e+00 -1.11091399e+00 8.48360538e-01 -5.03678858e-01
7.92899132e-01 -1.17108651e-01 -6.86069191e-01 6.44143224e-01
4.17739004e-01 2.85626322e-01 3.93251687e-01 -2.59192884e-01
-2.74778605e-01 1.36431092e-02 -6.74053550e-01 6.77965224e-01
1.32816720e+00 -6.40613437e-01 -3.05491000e-01 3.76914322e-01
8.95311594e-01 -8.14881861e-01 -9.82590735e-01 3.07802647e-01
3.63615841e-01 -9.41163123e-01 7.90809095e-01 -5.62369108e-01
7.73512006e-01 -5.17634451e-01 -3.16571862e-01 -7.06263423e-01
-2.06458583e-01 -9.08140004e-01 -9.84220058e-02 1.53180015e+00
4.09840554e-01 -5.91575019e-02 6.30702555e-01 3.99638802e-01
-4.36360359e-01 -9.57990885e-01 -9.39891875e-01 -7.60242701e-01
-3.73043597e-01 -8.10466468e-01 1.76085159e-01 6.31806374e-01
-2.44339347e-01 5.00666380e-01 -3.95204127e-01 -9.03144181e-02
4.21465933e-01 7.89142326e-02 8.60389531e-01 -6.23122573e-01
-4.32539463e-01 -6.38530970e-01 -3.92386764e-01 -1.64906216e+00
2.70796239e-01 -1.00410461e+00 3.75902206e-02 -1.59717000e+00
2.13075027e-01 -2.07439721e-01 -2.21846044e-01 3.60054195e-01
-3.86184007e-01 6.18123829e-01 8.15178812e-01 5.02557755e-01
-1.27956140e+00 4.62533891e-01 1.36067855e+00 -1.00530095e-01
-1.45542428e-01 -8.29305500e-02 -5.99460542e-01 5.59447408e-01
8.12196851e-01 -4.97562677e-01 -1.90198153e-01 -5.71812093e-01
6.62600098e-04 1.00914633e-03 3.23725075e-01 -7.89915740e-01
1.82867423e-01 1.09318040e-01 8.25795978e-02 -8.86267424e-01
4.81427312e-01 -4.76214707e-01 2.22423464e-01 2.95858264e-01
-1.69740349e-01 6.18387870e-02 1.97065189e-01 2.61783808e-01
-4.31512535e-01 -1.91294879e-01 7.22099721e-01 -7.09996372e-02
-1.19739914e+00 5.29350877e-01 -6.47313446e-02 5.44200361e-01
9.38933074e-01 -1.90132305e-01 -2.77067989e-01 -5.17969728e-01
-6.35541141e-01 4.58605766e-01 4.20285761e-01 6.40710890e-01
6.97770953e-01 -1.08801711e+00 -9.29790199e-01 -2.43713409e-01
1.80695668e-01 -4.85562496e-02 3.67282569e-01 1.42392635e+00
-5.36442101e-01 4.49293792e-01 -6.37031347e-02 -1.00826907e+00
-1.34603572e+00 3.57214630e-01 2.44834989e-01 3.80264930e-02
-7.99646139e-01 1.35340345e+00 1.10676803e-01 3.30518097e-01
1.58485100e-01 -3.22712272e-01 -2.78716117e-01 2.14620799e-01
5.08294761e-01 3.18843752e-01 -1.19149648e-01 -1.09316850e+00
-4.98048156e-01 6.22595847e-01 -7.83953629e-03 -2.20312700e-01
1.15945733e+00 -3.69950265e-01 7.65717924e-02 4.06179696e-01
1.24984789e+00 9.96704102e-02 -1.45893085e+00 -1.50304094e-01
-1.29067764e-01 -6.99740052e-01 1.70674458e-01 -4.49287474e-01
-1.41774070e+00 8.40482652e-01 4.30782765e-01 1.11189976e-01
1.11199141e+00 3.66246164e-01 1.00340176e+00 -2.56423771e-01
1.36117756e-01 -7.99567521e-01 1.15283327e-02 4.25461560e-01
8.69744897e-01 -1.39365220e+00 5.05443066e-02 -7.21692741e-01
-9.28973198e-01 1.11851394e+00 5.60181916e-01 6.92020282e-02
2.48797774e-01 -2.73319660e-03 2.05909938e-01 -1.64415225e-01
-8.11383367e-01 -5.36377847e-01 6.35111392e-01 2.02425972e-01
5.61322689e-01 -1.68828800e-01 -2.07672343e-01 5.60062587e-01
-1.71201989e-01 2.14681894e-01 3.47612143e-01 4.75174397e-01
-5.29119559e-02 -8.04923594e-01 -3.99883240e-02 1.51095450e-01
-8.59249949e-01 -3.13565314e-01 -1.38692349e-01 7.06149936e-01
-3.96949705e-03 1.11545205e+00 -1.10647388e-01 -3.50047112e-01
5.20147920e-01 -1.40417099e-01 4.14366990e-01 -4.08506066e-01
-7.57794261e-01 6.42145097e-01 3.12579334e-01 -1.00454640e+00
-6.01559997e-01 -7.06044912e-01 -1.32728255e+00 -2.88757831e-01
-1.63025990e-01 2.24251577e-04 4.16498542e-01 8.97099912e-01
4.32959080e-01 8.64119947e-01 3.49040180e-01 -8.87870073e-01
-9.96304154e-02 -8.66560638e-01 -1.89141855e-01 4.79810119e-01
7.83292726e-02 -5.35334587e-01 -2.05518186e-01 3.64297748e-01]
|
[9.513982772827148, 0.35508227348327637]
|
afd85748-dba2-4c8f-895e-1f25394cf2cb
|
incremental-boosting-convolutional-neural
|
1707.05395
| null |
http://arxiv.org/abs/1707.05395v1
|
http://arxiv.org/pdf/1707.05395v1.pdf
|
Incremental Boosting Convolutional Neural Network for Facial Action Unit Recognition
|
Recognizing facial action units (AUs) from spontaneous facial expressions is
still a challenging problem. Most recently, CNNs have shown promise on facial
AU recognition. However, the learned CNNs are often overfitted and do not
generalize well to unseen subjects due to limited AU-coded training images. We
proposed a novel Incremental Boosting CNN (IB-CNN) to integrate boosting into
the CNN via an incremental boosting layer that selects discriminative neurons
from the lower layer and is incrementally updated on successive mini-batches.
In addition, a novel loss function that accounts for errors from both the
incremental boosted classifier and individual weak classifiers was proposed to
fine-tune the IB-CNN. Experimental results on four benchmark AU databases have
demonstrated that the IB-CNN yields significant improvement over the
traditional CNN and the boosting CNN without incremental learning, as well as
outperforming the state-of-the-art CNN-based methods in AU recognition. The
improvement is more impressive for the AUs that have the lowest frequencies in
the databases.
|
['Zibo Meng', 'Yan Tong', 'Shizhong Han', 'Ahmed Shehab Khan']
|
2017-07-17
|
incremental-boosting-convolutional-neural-1
|
http://papers.nips.cc/paper/6258-incremental-boosting-convolutional-neural-network-for-facial-action-unit-recognition
|
http://papers.nips.cc/paper/6258-incremental-boosting-convolutional-neural-network-for-facial-action-unit-recognition.pdf
|
neurips-2016-12
|
['facial-action-unit-detection']
|
['computer-vision']
|
[ 1.65138036e-01 6.51491284e-02 -2.95921922e-01 -8.82995009e-01
-6.54150128e-01 6.27602488e-02 2.82016009e-01 -6.06504977e-01
-3.41832042e-01 7.42384672e-01 3.82192321e-02 3.54683787e-01
5.33465683e-01 -5.68856359e-01 -8.69329691e-01 -9.22735572e-01
-1.66479781e-01 -4.38929461e-02 -4.03369553e-02 -4.19852197e-01
-4.78036582e-01 6.73260331e-01 -1.90106928e+00 6.36881888e-01
4.51280117e-01 1.85963988e+00 -5.74573517e-01 3.41629565e-01
1.82153031e-01 1.23482370e+00 -4.98051226e-01 -6.25363588e-01
2.76893526e-01 -5.42272508e-01 -5.23064554e-01 -4.47556674e-02
7.95596123e-01 -8.05839479e-01 -4.52582359e-01 7.83975840e-01
5.92358947e-01 6.28397465e-02 3.13951075e-01 -1.50482774e+00
-6.84713900e-01 6.34096488e-02 -5.67608893e-01 -9.10462365e-02
1.81028247e-01 -1.39358804e-01 8.65030348e-01 -1.17550194e+00
2.97582328e-01 1.38090885e+00 9.28378701e-01 1.27624905e+00
-9.61485624e-01 -1.30313516e+00 2.45806143e-01 4.31824356e-01
-1.35102057e+00 -6.28116786e-01 8.42220724e-01 -1.93185285e-01
1.03373623e+00 -9.73378941e-02 5.92407167e-01 1.37512171e+00
-1.83992699e-01 1.18362129e+00 9.93878305e-01 -2.52577633e-01
5.41304350e-02 -1.58061102e-01 -2.32296064e-02 1.03418624e+00
-3.03992778e-01 1.62757918e-01 -6.64009571e-01 -1.69556484e-01
6.62778378e-01 1.53303565e-02 9.38788354e-02 -3.56107987e-02
-3.20327282e-01 8.13717186e-01 7.97047138e-01 2.07511991e-01
-5.11244833e-01 3.39784890e-01 7.03445137e-01 2.40487233e-01
7.96583533e-01 -6.03935160e-02 -5.65786183e-01 -2.37033784e-01
-7.28231728e-01 1.93310559e-01 3.83776963e-01 5.83028197e-01
8.82686079e-01 3.50758493e-01 -2.96195209e-01 1.24636245e+00
-2.10404620e-01 3.08240622e-01 4.89304036e-01 -6.41100228e-01
2.44023666e-01 7.79560685e-01 -1.32864729e-01 -8.54357958e-01
-5.23706019e-01 -2.47019902e-01 -9.94972765e-01 4.44677651e-01
3.20973963e-01 -3.35566700e-01 -1.12442255e+00 2.07002306e+00
2.72924930e-01 5.48096418e-01 -4.17934023e-02 7.59336710e-01
1.28703332e+00 5.01317084e-01 2.46109158e-01 -1.33181781e-01
9.02652025e-01 -1.22005737e+00 -6.77482426e-01 -5.24868704e-02
5.43108225e-01 -2.61501551e-01 6.80287004e-01 4.34611887e-01
-8.92918706e-01 -9.96571004e-01 -1.00262952e+00 1.50740802e-01
-8.61385390e-02 4.14472282e-01 8.40940118e-01 5.78648329e-01
-1.04759812e+00 5.60439348e-01 -7.79254019e-01 -7.24268854e-02
1.18929064e+00 6.16188943e-01 -6.80232286e-01 -2.01505069e-02
-1.22228718e+00 7.26943612e-01 9.17841643e-02 5.11858284e-01
-1.08537602e+00 -5.77946782e-01 -9.37758029e-01 -4.70642783e-02
1.18557796e-01 -1.33779898e-01 1.37434423e+00 -2.02382612e+00
-2.04882908e+00 9.17060912e-01 -3.19722205e-01 -3.73198211e-01
4.45819676e-01 -3.08720231e-01 -3.89079958e-01 1.02450892e-01
-2.07950085e-01 9.39813614e-01 1.14074421e+00 -8.46812010e-01
-5.44378340e-01 -5.63867450e-01 9.51751992e-02 5.04873507e-02
-5.60452640e-01 1.50458410e-01 -1.11578129e-01 -4.99789298e-01
-2.46488616e-01 -9.19598043e-01 2.33005881e-02 2.37403646e-01
2.52281725e-01 -7.48576224e-01 1.00293362e+00 -4.99931514e-01
9.97724235e-01 -2.11615157e+00 3.52593884e-02 -9.51974168e-02
8.58139619e-02 6.98493719e-01 -4.06546712e-01 -1.29146218e-01
-3.81371439e-01 -5.03673255e-01 1.52743742e-01 -4.43414360e-01
-3.73894036e-01 3.02341491e-01 2.47567091e-02 3.44168663e-01
5.80031753e-01 1.03625703e+00 -8.98177743e-01 -9.82122645e-02
-1.34701520e-01 5.13104796e-01 -5.29216707e-01 6.76266670e-01
-9.06436592e-02 4.48043287e-01 -1.09489724e-01 1.12878823e+00
7.76092231e-01 -1.11785784e-01 -1.22095890e-01 -2.22044677e-01
3.58162344e-01 -2.46191829e-01 -6.12971544e-01 1.50552309e+00
-3.79341215e-01 4.57439542e-01 1.16943337e-01 -1.38152564e+00
1.25870478e+00 5.31018853e-01 4.38020796e-01 -8.58900249e-01
3.30519378e-01 2.31126115e-01 -1.49432331e-01 -5.28331518e-01
1.72701385e-02 -4.20927644e-01 1.67230844e-01 5.01699597e-02
6.13790035e-01 3.79712075e-01 -1.28591835e-01 -4.51272815e-01
9.58626747e-01 3.28194976e-01 3.59801501e-01 1.68426067e-01
8.11540723e-01 -3.94044012e-01 9.21768129e-01 4.70228136e-01
-5.95465839e-01 5.37362993e-01 5.84608555e-01 -1.08086932e+00
-5.11251092e-01 -7.25128651e-01 -8.93496796e-02 1.62520564e+00
-3.21954280e-01 -4.66521531e-02 -8.88914227e-01 -1.21757972e+00
1.75248638e-01 -4.01737355e-02 -1.27761769e+00 -3.06750566e-01
-4.55963284e-01 -8.74976099e-01 7.52539039e-01 9.99892056e-01
7.54250109e-01 -1.38389099e+00 -2.94195622e-01 1.33511007e-01
3.36421952e-02 -1.17267358e+00 -9.97060686e-02 2.70425946e-01
-8.17865968e-01 -1.02143908e+00 -9.06544566e-01 -8.12829554e-01
7.71466911e-01 -1.50999635e-01 8.92200768e-01 9.88127589e-02
-3.80615771e-01 3.35917026e-02 -5.38614154e-01 -7.06392050e-01
-1.83188900e-01 -8.68297666e-02 1.55141920e-01 6.99927092e-01
7.87840009e-01 -3.52946758e-01 -5.22660792e-01 3.46141785e-01
-4.91388619e-01 -2.32006460e-01 5.80718637e-01 1.44332361e+00
3.55158687e-01 -5.86593688e-01 9.11093175e-01 -7.09562719e-01
2.23323509e-01 -3.66933048e-01 -3.70433450e-01 7.42973760e-02
-2.18573600e-01 -3.29163134e-01 5.92107058e-01 -6.86610043e-01
-1.12103081e+00 3.84756833e-01 -5.54790795e-01 -5.75607598e-01
-2.06085473e-01 2.18668744e-01 -9.80799720e-02 -5.54648995e-01
6.37835860e-01 4.31030709e-03 4.59784329e-01 -2.44545192e-01
4.81155850e-02 6.76164806e-01 3.58767807e-01 -4.19795215e-01
2.43960917e-01 5.66646338e-01 5.47974790e-03 -8.36801112e-01
-1.24510264e+00 -2.65896499e-01 -7.64240265e-01 -5.65197647e-01
7.58744836e-01 -1.20300317e+00 -8.03449333e-01 1.05534053e+00
-1.12426507e+00 -5.08663476e-01 -1.44558281e-01 1.27224207e-01
-5.39281487e-01 -4.17709202e-02 -8.38217616e-01 -7.88579404e-01
-5.04446566e-01 -9.75858808e-01 1.15534461e+00 4.70131099e-01
-2.15114951e-01 -5.86741984e-01 5.01237959e-02 3.62834573e-01
4.03072029e-01 5.16432822e-01 4.10049736e-01 -3.73454571e-01
4.19049561e-02 -5.63388824e-01 -4.15299058e-01 1.05549049e+00
3.50274235e-01 -1.81747712e-02 -1.56424236e+00 -3.32735628e-01
-3.43666852e-01 -1.32785106e+00 8.57830226e-01 3.27738315e-01
1.41721570e+00 -2.93313056e-01 -5.31740785e-02 7.84716129e-01
9.88555431e-01 3.08659405e-01 8.15269828e-01 1.34472415e-01
6.54218078e-01 5.36638319e-01 5.16700566e-01 4.19700533e-01
-3.78956124e-02 7.95290470e-01 6.51962817e-01 -4.57536608e-01
-4.25405800e-02 -5.45350909e-02 5.50386608e-01 3.87460947e-01
-5.29852927e-01 4.98714060e-01 -4.48224247e-01 4.10793275e-01
-1.94606543e+00 -1.04083991e+00 5.34158170e-01 1.78046763e+00
8.67766142e-01 -1.57198846e-01 4.91461396e-01 -1.31564856e-01
3.45502645e-01 2.00437278e-01 -6.23571098e-01 -6.14826500e-01
-2.37005264e-01 6.47677958e-01 4.64536399e-02 1.32523134e-01
-1.31734288e+00 1.23730588e+00 6.88868999e+00 7.89545000e-01
-1.56855392e+00 2.83700734e-01 1.18486702e+00 -1.65535897e-01
6.35139227e-01 -6.82816982e-01 -1.04986691e+00 1.92520663e-01
7.84779787e-01 3.77964348e-01 8.94659832e-02 1.47352242e+00
-4.65094708e-02 2.48685151e-01 -1.00802851e+00 1.26047897e+00
5.80861807e-01 -1.15164769e+00 3.69472951e-02 -2.48788089e-01
1.08752739e+00 1.19265176e-01 -4.43640836e-02 6.04046285e-01
1.91979691e-01 -1.36348832e+00 6.06301367e-01 4.56820935e-01
8.89921129e-01 -8.63148332e-01 1.15315962e+00 -4.08934802e-02
-1.05301774e+00 -3.59558880e-01 -5.58032751e-01 -4.02947813e-01
-4.08471555e-01 9.31618959e-02 -6.00998878e-01 1.64497733e-01
1.08489537e+00 1.03166246e+00 -4.40430880e-01 4.83080745e-01
-3.34622175e-01 6.14168882e-01 -1.06443996e-02 -3.20027024e-01
4.77306753e-01 1.10574896e-02 -5.33001162e-02 1.23858798e+00
1.36355475e-01 2.04033166e-01 4.30876538e-02 4.68907237e-01
-4.20548439e-01 2.47182786e-01 -4.45365041e-01 1.49308905e-01
-1.93876639e-01 1.42059422e+00 -8.36976152e-03 -2.83855170e-01
-8.10238421e-01 1.03886318e+00 9.47075009e-01 3.20936829e-01
-7.62791812e-01 -5.80894127e-02 1.15194178e+00 -5.65835088e-02
5.15419900e-01 1.78367183e-01 7.54614845e-02 -9.52492476e-01
8.74740332e-02 -1.05054629e+00 4.48478222e-01 -7.00941861e-01
-1.43392181e+00 1.07543135e+00 -2.77738541e-01 -1.19291306e+00
-2.81235665e-01 -1.07980478e+00 -4.07059878e-01 5.90367973e-01
-1.51488113e+00 -1.51497948e+00 -5.59837997e-01 8.12775552e-01
3.95569980e-01 -4.97520179e-01 1.17259765e+00 3.78783047e-01
-7.31713831e-01 1.12603426e+00 -2.24925727e-01 6.46958888e-01
6.87503457e-01 -7.86968291e-01 -1.54816195e-01 4.82035130e-01
-1.33621156e-01 1.92419171e-01 5.44349328e-02 -1.41218245e-01
-8.93362701e-01 -1.20725572e+00 6.43634796e-01 -1.82727978e-01
5.51624894e-01 -5.79055250e-01 -8.30406010e-01 6.80380762e-01
1.27635626e-02 7.28411615e-01 8.46899033e-01 2.87784755e-01
-8.41384351e-01 -7.97905445e-01 -1.05831027e+00 2.60872483e-01
9.93761361e-01 -4.76208061e-01 -1.08601384e-01 5.46917282e-02
1.12721205e-01 -4.54329401e-01 -6.43064439e-01 8.12874913e-01
1.05238128e+00 -1.01983881e+00 6.66710436e-01 -1.07793331e+00
5.64398885e-01 2.20953837e-01 -4.91204299e-02 -1.19477737e+00
-3.58408928e-01 -4.85934585e-01 -3.33834499e-01 9.75138545e-01
3.85451823e-01 -4.32925999e-01 1.26776993e+00 4.20484483e-01
8.46178755e-02 -1.09881985e+00 -1.21116078e+00 -7.46986628e-01
-1.08008087e-01 -3.97088706e-01 5.43603063e-01 7.89137423e-01
-1.40282689e-02 3.54932129e-01 -7.80429721e-01 -2.29568139e-01
4.37648445e-01 -1.18932262e-01 9.40519571e-01 -1.08154547e+00
-1.60106439e-02 -4.42672044e-01 -9.27861810e-01 -9.73923981e-01
5.92693150e-01 -6.45999372e-01 2.78748095e-01 -9.51538146e-01
2.90678293e-01 -2.97107637e-01 -6.02775097e-01 1.07106757e+00
-4.03536528e-01 7.03086495e-01 -1.70986623e-01 -1.43815309e-01
-6.26756430e-01 9.51134920e-01 1.17854023e+00 -2.83926457e-01
-3.87171395e-02 1.67692855e-01 -4.92015213e-01 7.13230073e-01
5.44272840e-01 -3.18293840e-01 -1.09613240e-01 -9.69912335e-02
-1.14055045e-01 -2.77018666e-01 2.34328866e-01 -1.01428092e+00
-8.24820623e-02 9.27702412e-02 6.32794738e-01 -3.15611631e-01
5.63500464e-01 -6.85943365e-01 -4.04514760e-01 3.38214368e-01
-2.31251493e-01 -2.76309699e-01 4.49721575e-01 3.20084214e-01
-5.80049157e-01 2.06617132e-01 1.23482227e+00 -3.69647704e-02
-8.15329313e-01 7.70124137e-01 -1.73395887e-01 -2.51707852e-01
1.08605027e+00 -2.69586056e-01 1.50827825e-01 -6.74923360e-01
-8.03241313e-01 -3.25311087e-02 -1.47968784e-01 6.42818749e-01
7.11242378e-01 -1.69919014e+00 -8.99237394e-01 4.46057171e-01
3.64402086e-01 -1.97690949e-01 2.73728579e-01 7.00592816e-01
-3.91029656e-01 2.22136244e-01 -7.15735972e-01 -6.21465445e-01
-1.72996962e+00 2.41805226e-01 6.30484760e-01 -1.80428684e-01
-1.34137988e-01 1.42314076e+00 2.61846900e-01 -4.24653590e-01
6.30387843e-01 2.27456577e-02 -3.26249599e-01 1.06899582e-01
1.01847529e+00 1.76041275e-01 3.30756962e-01 -1.01311457e+00
-3.36997360e-01 4.74238127e-01 -2.31830075e-01 4.52055305e-01
1.57846510e+00 3.63620669e-01 -2.70697653e-01 2.65956610e-01
1.55641329e+00 -7.38156319e-01 -1.61786139e+00 -5.02262056e-01
-3.01406592e-01 -3.58723909e-01 -1.68456331e-01 -5.85505426e-01
-1.38171554e+00 8.81408691e-01 7.22139716e-01 -3.91405791e-01
1.44582331e+00 -9.19470191e-02 7.17425406e-01 4.15913999e-01
3.32297117e-01 -1.18795812e+00 5.32972574e-01 6.11168802e-01
1.20261049e+00 -1.50447726e+00 -1.88051388e-01 -2.32911617e-01
-6.27425969e-01 1.42154098e+00 1.11629891e+00 -3.97362769e-01
8.87825489e-01 1.73133224e-01 4.95711654e-01 -5.32511808e-02
-6.93944573e-01 -1.94981650e-01 4.35952514e-01 5.15358150e-01
5.90295553e-01 -2.62598023e-02 -2.05918085e-02 7.03864992e-01
8.39879438e-02 4.94716316e-01 -1.84599146e-01 8.54148805e-01
-3.21562365e-02 -8.83354068e-01 -1.03940867e-01 3.53374839e-01
-6.21835291e-01 1.45754084e-01 -6.08474493e-01 7.18517065e-01
4.71626014e-01 6.16485655e-01 2.20970422e-01 -6.61340714e-01
4.19798374e-01 2.37435400e-01 7.53828406e-01 -4.62636501e-01
-6.37814999e-01 -2.56124109e-01 8.49277154e-02 -1.03589606e+00
-7.17136323e-01 -5.74113488e-01 -8.14771593e-01 4.65928623e-03
-2.81092793e-01 -7.33873695e-02 3.63542676e-01 8.70275855e-01
2.48821646e-01 2.53267705e-01 9.38872099e-01 -1.12924147e+00
-5.01103461e-01 -1.36068678e+00 -4.52422857e-01 5.80939829e-01
4.90693957e-01 -9.37800765e-01 -2.59085119e-01 -5.23604490e-02]
|
[13.601037979125977, 1.674950122833252]
|
967ff50d-c5da-4029-a6a9-99ea80398a4e
|
recent-advancements-in-end-to-end-autonomous
|
2307.04370
| null |
https://arxiv.org/abs/2307.04370v1
|
https://arxiv.org/pdf/2307.04370v1.pdf
|
Recent Advancements in End-to-End Autonomous Driving using Deep Learning: A Survey
|
End-to-End driving is a promising paradigm as it circumvents the drawbacks associated with modular systems, such as their overwhelming complexity and propensity for error propagation. Autonomous driving transcends conventional traffic patterns by proactively recognizing critical events in advance, ensuring passengers' safety and providing them with comfortable transportation, particularly in highly stochastic and variable traffic settings. This paper presents a comprehensive review of the End-to-End autonomous driving stack. It provides a taxonomy of automated driving tasks wherein neural networks have been employed in an End-to-End manner, encompassing the entire driving process from perception to control, while addressing key challenges encountered in real-world applications. Recent developments in End-to-End autonomous driving are analyzed, and research is categorized based on underlying principles, methodologies, and core functionality. These categories encompass sensorial input, main and auxiliary output, learning approaches ranging from imitation to reinforcement learning, and model evaluation techniques. The survey incorporates a detailed discussion of the explainability and safety aspects. Furthermore, it assesses the state-of-the-art, identifies challenges, and explores future possibilities. We maintained the latest advancements and their corresponding open-source implementations at https://github.com/Pranav-chib/Recent-Advancements-in-End-to-End-Autonomous-Driving-using-Deep-Learning.
|
['Pravendra Singh', 'Pranav Singh Chib']
|
2023-07-10
| null | null | null | null |
['autonomous-driving']
|
['computer-vision']
|
[-3.35537374e-01 6.52397722e-02 -4.38561380e-01 -5.66797554e-01
-7.00904369e-01 -6.83403313e-01 5.31058669e-01 -2.07933575e-01
-5.06187260e-01 5.57243288e-01 -1.04387350e-01 -5.55132747e-01
-2.02717304e-01 -6.60803139e-01 -7.85980999e-01 -6.36445343e-01
-2.56495386e-01 2.78082639e-01 2.53466189e-01 -7.99934864e-01
1.27908304e-01 4.65631336e-01 -2.20664120e+00 -2.16976583e-01
7.32557476e-01 9.72055912e-01 1.60017028e-01 9.95434284e-01
3.31484944e-01 5.92906475e-01 -2.68185586e-01 -2.99123436e-01
3.45867604e-01 1.92149699e-01 -1.66550666e-01 -2.56219655e-01
2.36053109e-01 -3.71634781e-01 -1.03004074e+00 5.67979813e-01
6.23944163e-01 3.26114684e-01 4.36277866e-01 -2.04407144e+00
-2.13026837e-01 -8.49111751e-02 1.30386606e-01 3.30192119e-01
1.78113326e-01 7.70953894e-01 5.57823539e-01 -7.53467143e-01
2.14329630e-01 9.27038789e-01 6.78657055e-01 8.09114218e-01
-9.05018926e-01 -7.27521956e-01 2.68755764e-01 5.19197106e-01
-1.20481253e+00 -8.28428805e-01 6.87359452e-01 -6.34082198e-01
1.14411879e+00 5.11514619e-02 6.74933910e-01 1.17626941e+00
7.38812029e-01 1.07110786e+00 4.96285200e-01 2.29031727e-01
2.71170110e-01 8.39986429e-02 1.09913638e-02 4.98024195e-01
8.20515230e-02 1.08278656e+00 -5.14529169e-01 3.14418346e-01
-7.30845183e-02 -1.02493770e-01 3.80595028e-01 -5.34257233e-01
-1.13659310e+00 7.25854933e-01 2.70326465e-01 -2.96720147e-01
-5.17511547e-01 5.09185374e-01 6.56107783e-01 2.93424726e-01
9.31060500e-03 5.96192777e-02 -4.16572332e-01 -6.73507214e-01
-7.18659639e-01 8.38837445e-01 6.43366456e-01 1.25236619e+00
7.95105219e-01 5.14670551e-01 -8.03534687e-02 5.55854499e-01
2.57557482e-01 6.46434665e-01 1.94311023e-01 -1.46781266e+00
4.34774369e-01 -4.75699641e-02 3.48443002e-01 -6.85141444e-01
-7.38789320e-01 -4.80841100e-01 -4.22763318e-01 8.33128929e-01
-6.06432930e-02 -6.13577247e-01 -7.63635576e-01 1.63254952e+00
2.03775033e-01 7.88583308e-02 2.67230839e-01 8.47907603e-01
7.85314322e-01 4.84968871e-01 2.54419237e-01 3.38636816e-01
1.17322576e+00 -1.20144069e+00 -7.89862335e-01 -6.63019896e-01
3.22788835e-01 -4.04017359e-01 7.26553798e-01 4.12866890e-01
-1.11163473e+00 -9.64452803e-01 -1.28172398e+00 -1.65557444e-01
-7.90116787e-01 -1.25850320e-01 4.07312304e-01 5.46132803e-01
-1.08085120e+00 3.26280177e-01 -7.57075071e-01 -2.21019462e-01
3.31891805e-01 4.58969295e-01 -1.35295287e-01 1.73113063e-01
-1.57848990e+00 1.27640665e+00 6.06097691e-02 6.84815226e-03
-1.47843456e+00 -7.35265791e-01 -1.04024780e+00 -3.53984237e-01
3.49510968e-01 -7.27676630e-01 1.95202506e+00 -3.37602869e-02
-1.63932395e+00 4.15505201e-01 -4.37809914e-01 -7.88786948e-01
7.44136930e-01 -3.09866071e-01 -8.72857809e-01 -1.78134158e-01
4.12188262e-01 1.20362234e+00 5.90365112e-01 -1.19634652e+00
-1.29866481e+00 7.61003979e-03 1.25376344e-01 2.38407001e-01
3.60471457e-01 -4.39717650e-01 -2.40069285e-01 -1.00306198e-02
-3.99606705e-01 -1.08775091e+00 -4.12611425e-01 1.58657268e-01
-8.53326470e-02 -3.52880150e-01 1.09564590e+00 -1.36128947e-01
1.02525055e+00 -2.28530765e+00 -3.42877597e-01 -2.07979798e-01
2.14901507e-01 2.04953894e-01 -8.40242356e-02 7.74957895e-01
2.15083018e-01 -3.97142082e-01 -6.11819737e-02 -4.90297973e-01
3.57861757e-01 4.34342444e-01 -4.12448287e-01 4.18507755e-01
2.88888514e-01 1.17017817e+00 -1.15204751e+00 -2.29671761e-01
8.92514408e-01 4.65304583e-01 -2.00996563e-01 1.79327384e-01
-2.73682480e-03 5.77180386e-01 -3.04193437e-01 6.70445919e-01
4.95943159e-01 5.72948396e-01 -7.06050336e-01 2.86293089e-01
-5.53660333e-01 4.85424578e-01 -9.45097744e-01 1.47944725e+00
-6.20453894e-01 1.10538995e+00 2.83895016e-01 -9.94526863e-01
9.76503134e-01 3.49120587e-01 4.66148794e-01 -1.01274836e+00
2.25152388e-01 4.09727514e-01 -9.69575420e-02 -8.32993329e-01
7.91559815e-01 1.08233675e-01 -4.27821338e-01 -1.26775831e-01
-1.61267102e-01 -2.44387299e-01 4.82110232e-01 -3.14838439e-02
8.91052365e-01 3.87586415e-01 -6.70420825e-02 -3.82347740e-02
5.81810236e-01 3.00343096e-01 4.07744706e-01 7.85136104e-01
-1.11763597e+00 9.05217230e-02 1.78056374e-01 -5.38985968e-01
-8.20522189e-01 -1.26354492e+00 -1.39246479e-01 1.29378819e+00
4.97821897e-01 -9.56789851e-02 -6.38339639e-01 -3.53679627e-01
4.23702329e-01 1.21219873e+00 -6.05877459e-01 -3.67260188e-01
-4.30628747e-01 -9.04364511e-02 7.66753316e-01 7.63584971e-01
4.99268860e-01 -8.63132894e-01 -9.31202888e-01 4.15209472e-01
-2.14322656e-01 -1.37620306e+00 -7.00478405e-02 2.17719227e-01
-5.23491383e-01 -7.35489249e-01 -1.61944702e-01 -5.86472273e-01
6.25834465e-02 6.30897105e-01 9.30427909e-01 -3.71269405e-01
-1.80616770e-02 4.05664444e-01 1.64163798e-01 -9.19198692e-01
-4.49972600e-01 1.74652878e-02 3.99489462e-01 -2.99306244e-01
7.58299351e-01 -3.94598186e-01 -8.03608596e-01 5.82198858e-01
-2.98527807e-01 -1.91881359e-01 5.39690614e-01 4.82149065e-01
4.39867973e-01 -1.01623058e-01 8.18304539e-01 -1.50426179e-02
7.21254647e-01 -6.14983320e-01 -6.70481920e-01 -4.91973549e-01
-7.36720204e-01 -2.86610335e-01 6.51482880e-01 -7.16334507e-02
-7.29489505e-01 1.56336203e-01 -5.95704973e-01 -2.63837039e-01
-7.42287099e-01 6.98000267e-02 9.49999988e-02 -5.55886179e-02
7.94374228e-01 4.23745662e-01 5.52760363e-01 1.30224571e-01
5.66214263e-01 9.01321828e-01 8.02010655e-01 -3.27143759e-01
8.51609170e-01 5.85342884e-01 4.88571152e-02 -6.21618688e-01
-4.34263676e-01 -5.37268639e-01 -5.57852447e-01 -7.48712599e-01
6.55564070e-01 -1.23387742e+00 -1.17090476e+00 3.65209430e-01
-8.74282420e-01 -6.07564867e-01 -4.61790442e-01 6.08852744e-01
-1.08160996e+00 -9.47432518e-02 -4.26217884e-01 -9.52464998e-01
-5.25546819e-02 -1.51019883e+00 1.08406270e+00 3.34594339e-01
-2.54333258e-01 -8.12359691e-01 1.08581096e-01 4.79506433e-01
7.68684924e-01 1.98339134e-01 2.28035703e-01 -2.86380172e-01
-4.56050277e-01 -6.01647735e-01 2.14302942e-01 2.48970032e-01
-3.04030061e-01 -4.20267582e-02 -1.19719362e+00 -1.69773638e-01
-5.26710451e-01 -9.48859528e-02 6.94654465e-01 5.47440708e-01
8.54592979e-01 1.83450595e-01 -5.56567550e-01 3.16453934e-01
1.09893596e+00 4.68204021e-01 4.91112381e-01 6.52280509e-01
2.24170268e-01 9.21127200e-01 1.01536930e+00 1.38710991e-01
9.95138347e-01 4.64267403e-01 8.66396248e-01 -2.99005606e-03
-7.09015578e-02 -3.47522765e-01 6.43146455e-01 4.33941901e-01
2.29508951e-01 -2.82395899e-01 -6.35631025e-01 8.09335947e-01
-2.02588654e+00 -9.84427154e-01 -2.60214537e-01 1.94273806e+00
1.52830824e-01 5.96374989e-01 5.86027205e-01 2.02086776e-01
6.44730151e-01 1.87162057e-01 -1.05569315e+00 -9.12032127e-01
3.02857637e-01 -3.97092491e-01 8.61079812e-01 6.72495127e-01
-1.11360860e+00 9.39417601e-01 7.46058750e+00 5.92052817e-01
-1.08053553e+00 1.86776906e-01 2.45472848e-01 -2.63572723e-01
1.09703332e-01 -2.44260639e-01 -1.06525671e+00 5.46305776e-01
1.46660495e+00 -8.95656943e-02 5.10262132e-01 1.21877384e+00
9.02618945e-01 -1.90669253e-01 -1.05044019e+00 7.81469166e-01
-3.08145374e-01 -1.21979415e+00 -5.58410764e-01 -3.58871408e-02
3.59474838e-01 8.76866221e-01 2.50247508e-01 8.26283336e-01
1.57442287e-01 -8.45952272e-01 1.17914557e+00 4.12634194e-01
6.89805090e-01 -9.07608032e-01 5.69811106e-01 4.36625361e-01
-1.33111572e+00 -5.60096681e-01 -1.06682688e-01 -4.31412309e-01
5.47232628e-01 2.62354255e-01 -4.54854488e-01 3.16212237e-01
7.25609660e-01 7.55240381e-01 -1.98603854e-01 9.64349031e-01
-2.62980461e-01 4.03579652e-01 -1.45413727e-01 -3.80242705e-01
6.47416592e-01 -1.07347788e-02 8.07850182e-01 1.47911370e+00
1.52621672e-01 -3.09352338e-01 1.46935031e-01 8.19907665e-01
4.77655679e-01 -6.34180009e-01 -1.02922618e+00 4.33955103e-01
6.23422325e-01 1.17625153e+00 3.42929363e-02 -3.16354126e-01
-4.79673356e-01 3.47048461e-01 -7.55781084e-02 4.88321960e-01
-1.15718508e+00 -8.36238205e-01 1.40942001e+00 1.92062914e-01
1.33444592e-01 -6.74984694e-01 -4.11736310e-01 -2.98987359e-01
1.21270895e-01 -5.10229170e-01 -2.77953018e-02 -8.78190637e-01
-7.25407302e-01 3.94488305e-01 2.06635788e-01 -1.55978835e+00
-5.12576699e-01 -7.19345093e-01 -6.39102936e-01 6.42382562e-01
-2.00668168e+00 -8.36230397e-01 -4.64074731e-01 3.97480935e-01
9.40489590e-01 -2.69149989e-01 4.84092325e-01 5.56764841e-01
-5.45530558e-01 6.70266390e-01 1.92349851e-01 -1.71535745e-01
4.03598964e-01 -9.73281205e-01 8.15424621e-01 7.53942430e-01
-7.15437591e-01 9.42562521e-02 1.07075155e+00 -2.02888906e-01
-1.62184298e+00 -1.28457928e+00 7.26513445e-01 -6.46851122e-01
6.15584314e-01 -5.18745124e-01 -3.90874982e-01 6.66763544e-01
4.19498146e-01 -1.15845129e-01 2.51195818e-01 -2.32969776e-01
3.94811891e-02 -5.42583406e-01 -1.14015281e+00 9.01371300e-01
1.00277150e+00 -3.58699530e-01 -3.50010455e-01 1.28346294e-01
6.44660830e-01 -6.34469032e-01 -4.78085041e-01 3.26729774e-01
8.05689454e-01 -9.25133705e-01 7.54577875e-01 -4.42617297e-01
2.42841184e-01 -5.89786530e-01 -4.69305702e-02 -1.38368738e+00
-5.41143775e-01 -1.05679238e+00 -3.98506880e-01 5.58815181e-01
5.37292063e-01 -9.50510085e-01 7.36320853e-01 6.63572490e-01
-8.07455778e-01 -8.64798784e-01 -1.13997304e+00 -9.93768930e-01
2.25111246e-01 -1.08040893e+00 5.74864984e-01 1.98940828e-01
5.06991111e-02 2.30543032e-01 -2.05247894e-01 2.19085366e-01
6.03149533e-01 -4.98040646e-01 1.06731176e+00 -9.43902910e-01
3.15616667e-01 -6.41028106e-01 -7.32253790e-01 -1.40307510e+00
5.40244691e-02 -5.55459559e-01 5.25427401e-01 -1.68743849e+00
-7.14104950e-01 -2.19273612e-01 -2.40723625e-01 3.26827109e-01
2.88188636e-01 8.33278596e-02 -3.42857745e-03 -1.34366438e-01
-6.39494717e-01 5.75836062e-01 1.09211457e+00 -2.23460093e-01
-1.74462989e-01 2.63006628e-01 -8.15372527e-01 4.17784631e-01
1.24568856e+00 -2.88778543e-01 -5.45214772e-01 -3.92790973e-01
3.78464274e-02 -8.44129547e-02 5.34081995e-01 -1.33485949e+00
5.83401501e-01 -2.42479339e-01 -1.24781102e-01 -1.02341056e+00
7.15234399e-01 -8.06708872e-01 -2.11806104e-01 6.78894997e-01
-2.86650985e-01 3.57650816e-01 6.60604239e-01 7.77429044e-01
-2.23045900e-01 8.50984752e-02 7.37699747e-01 1.73982531e-01
-1.36651194e+00 2.73769051e-01 -1.26833415e+00 -2.98613496e-02
1.41651642e+00 -5.64651906e-01 -2.69448996e-01 -6.12948239e-01
-7.27075160e-01 9.46149409e-01 -1.44307479e-01 1.07504964e+00
5.70254862e-01 -1.23873186e+00 -7.87215292e-01 4.24454004e-01
5.29350102e-01 -1.92373648e-01 3.71349812e-01 1.00416374e+00
-2.40527332e-01 9.23150361e-01 -3.38878721e-01 -8.47560883e-01
-6.80194974e-01 5.46119869e-01 6.86159670e-01 4.29418683e-01
-5.69048464e-01 3.76873761e-01 -2.86070835e-02 -7.73642957e-01
3.97250026e-01 -9.82586369e-02 -1.07343666e-01 -4.25429732e-01
2.90660262e-01 6.89364195e-01 2.84997195e-01 -7.83962965e-01
-4.21132177e-01 4.78057981e-01 1.62228525e-01 -2.46590659e-01
8.41617942e-01 -6.35960996e-01 6.70210421e-01 4.78164226e-01
1.00803530e+00 -6.00448966e-01 -1.71464217e+00 2.44410992e-01
-3.96645755e-01 -2.04575717e-01 2.48286843e-01 -7.95079291e-01
-8.34909856e-01 1.11881006e+00 9.07444298e-01 1.29616842e-01
6.64552450e-01 -3.21651816e-01 1.17698073e+00 5.16816795e-01
4.74847615e-01 -1.38799357e+00 -3.63661528e-01 7.98675716e-01
7.81766653e-01 -1.37512290e+00 -5.60254216e-01 -5.27962521e-02
-8.23934436e-01 9.44346428e-01 7.58660436e-01 -1.74803659e-01
9.91634190e-01 3.63813758e-01 4.01024878e-01 -8.22842568e-02
-1.04590440e+00 -4.85660195e-01 -5.95865473e-02 1.00882959e+00
6.89283945e-03 3.31611812e-01 -3.89202125e-02 1.48311913e-01
-4.33074743e-01 5.98101877e-02 4.28991944e-01 1.10690773e+00
-5.24652660e-01 -7.13193595e-01 -9.03285146e-02 2.94198364e-01
-3.13068666e-02 3.95582795e-01 1.13283195e-01 8.65736306e-01
3.53256673e-01 1.61966443e+00 1.43636972e-01 -8.29040766e-01
8.97413969e-01 1.88357895e-03 -8.10946897e-02 -5.61060011e-03
-3.70957553e-01 -3.70198250e-01 3.71390581e-01 -7.90557027e-01
-4.24951725e-02 -7.38505363e-01 -1.36337864e+00 -6.73608601e-01
7.66143650e-02 -6.65573925e-02 1.25145125e+00 8.61653209e-01
9.54254150e-01 7.69541502e-01 7.87166357e-01 -1.27762938e+00
-5.02852619e-01 -5.39447665e-01 -8.23405087e-02 -1.60668582e-01
9.28798914e-01 -9.79741871e-01 -4.42459613e-01 -2.59949625e-01]
|
[5.694730281829834, 0.9634681940078735]
|
33e92fad-a947-41e5-b138-b1e1b7f0b3e8
|
feds-filtered-edit-distance-surrogate
|
2103.04635
| null |
https://arxiv.org/abs/2103.04635v2
|
https://arxiv.org/pdf/2103.04635v2.pdf
|
FEDS -- Filtered Edit Distance Surrogate
|
This paper proposes a procedure to train a scene text recognition model using a robust learned surrogate of edit distance. The proposed method borrows from self-paced learning and filters out the training examples that are hard for the surrogate. The filtering is performed by judging the quality of the approximation, using a ramp function, enabling end-to-end training. Following the literature, the experiments are conducted in a post-tuning setup, where a trained scene text recognition model is tuned using the learned surrogate of edit distance. The efficacy is demonstrated by improvements on various challenging scene text datasets such as IIIT-5K, SVT, ICDAR, SVTP, and CUTE. The proposed method provides an average improvement of $11.2 \%$ on total edit distance and an error reduction of $9.5\%$ on accuracy.
|
['Jiri Matas', 'Yash Patel']
|
2021-03-08
| null | null | null | null |
['scene-text-recognition']
|
['computer-vision']
|
[ 6.04140460e-01 -2.04604924e-01 1.04458451e-01 -8.02458286e-01
-8.01475763e-01 -4.76276964e-01 6.32248223e-01 3.89932752e-01
-7.23691523e-01 3.46835881e-01 -1.34170547e-01 -6.00473545e-02
-9.38162282e-02 -6.92815959e-01 -7.93889046e-01 -4.54389304e-01
5.09199873e-02 3.45233679e-01 1.05306515e-02 -1.46654099e-02
7.38643885e-01 1.86139777e-01 -1.48926592e+00 3.87966514e-01
1.21189916e+00 1.19263422e+00 1.89564645e-01 8.29634130e-01
-4.05517310e-01 6.85507834e-01 -8.12079608e-01 -5.19324183e-01
4.11726445e-01 -1.77292183e-01 -5.74307621e-01 7.48792291e-02
7.64323592e-01 -1.38579652e-01 -3.91074061e-01 9.06792760e-01
4.27200943e-01 2.37210333e-01 6.90312386e-01 -7.85552621e-01
-5.28492391e-01 6.09609008e-01 -3.46061677e-01 3.53727452e-02
3.08594435e-01 1.84186697e-01 7.08513021e-01 -1.13889503e+00
4.56923991e-01 9.74347413e-01 7.84383655e-01 3.50276291e-01
-1.05132627e+00 -5.77859104e-01 2.02355027e-01 1.52355060e-01
-1.60357785e+00 -5.59640408e-01 6.22917116e-01 -4.84282792e-01
1.17883193e+00 2.23991454e-01 2.48101175e-01 9.39636946e-01
1.15538709e-01 7.61559248e-01 1.03075850e+00 -8.07416618e-01
4.79998291e-01 2.81447381e-01 2.44240522e-01 6.80940211e-01
1.05369866e-01 -1.24312997e-01 -8.11695158e-01 1.59579337e-01
3.09819192e-01 -3.49692494e-01 -1.20412573e-01 -3.92349064e-01
-9.86450136e-01 7.22189069e-01 3.63216877e-01 5.51683120e-02
4.80281375e-02 8.43842328e-03 7.55289316e-01 3.46395642e-01
5.53544044e-01 4.59150434e-01 -2.75295049e-01 -4.75404024e-01
-1.28484142e+00 -2.26186514e-01 5.82411587e-01 1.15106571e+00
3.09889138e-01 2.92992026e-01 -3.91001195e-01 1.07064950e+00
2.98197120e-01 5.56546748e-01 5.65727830e-01 -3.86213779e-01
7.96780586e-01 8.06779385e-01 6.59480393e-02 -8.50737274e-01
-2.92708986e-02 -3.93569380e-01 -7.10059345e-01 1.97489917e-01
2.33258620e-01 -3.86211723e-02 -1.18784010e+00 1.32987726e+00
2.29960158e-01 -3.96232633e-03 1.81754664e-01 6.22181535e-01
5.93978703e-01 7.73456335e-01 7.30732232e-02 -1.48366868e-01
8.57062876e-01 -1.11802864e+00 -6.16295040e-01 -1.85066894e-01
7.33165324e-01 -9.39000964e-01 1.57246578e+00 6.91336155e-01
-8.56252968e-01 -8.65695536e-01 -1.26046932e+00 1.63420945e-01
-4.14030373e-01 7.16221571e-01 3.14244330e-01 8.03150773e-01
-7.60246098e-01 6.02578104e-01 -6.69539154e-01 -6.09047592e-01
4.29864794e-01 2.05409914e-01 7.92341009e-02 -8.94365162e-02
-7.28317976e-01 6.76927209e-01 6.20494425e-01 9.46353078e-02
-9.11100090e-01 -7.80990362e-01 -7.11685896e-01 -9.24805552e-02
1.04310013e-01 -2.75261730e-01 1.05281878e+00 -9.92836535e-01
-1.75986409e+00 8.74912441e-01 1.14782840e-01 -6.78849697e-01
9.80557263e-01 -5.54500520e-01 -6.38586342e-01 -5.16828448e-02
-1.99946299e-01 4.31650043e-01 1.22354114e+00 -9.56170201e-01
-6.41588271e-01 -2.20127732e-01 -1.93253145e-01 3.92772198e-01
-6.55359030e-01 -1.66141853e-01 -5.31217098e-01 -8.51427495e-01
-1.31293893e-01 -7.71762073e-01 3.67862210e-02 1.35055259e-01
-3.89087886e-01 3.32822911e-02 1.07187545e+00 -5.18737733e-01
1.19068682e+00 -2.29002953e+00 -2.28093818e-01 2.65351146e-01
-1.79999590e-01 4.14866418e-01 -1.50773540e-01 3.74717355e-01
1.23503499e-01 -1.26771390e-01 -4.23162848e-01 -2.99807519e-01
-4.45821993e-02 -1.72901392e-01 -3.34381908e-01 4.55943257e-01
3.28261815e-02 5.12165487e-01 -7.43852079e-01 -2.93972582e-01
5.11306703e-01 4.09661233e-01 -3.66744846e-01 3.01941514e-01
-2.78674453e-01 -1.45264473e-02 -1.34570181e-01 4.86967295e-01
7.52842665e-01 -1.33550406e-01 -2.64021400e-02 -2.08984867e-01
-1.78036109e-01 8.95640068e-03 -1.35338235e+00 1.86971927e+00
-5.40393472e-01 8.38614702e-01 -3.37565899e-01 -9.24492061e-01
1.49545002e+00 -2.18420386e-01 2.42674887e-01 -7.19387412e-01
1.45107895e-01 -7.64176771e-02 -2.77291685e-01 -3.00602049e-01
7.95470297e-01 3.90168071e-01 -1.50885522e-01 2.43762612e-01
-4.32322472e-02 -2.69440681e-01 2.50534028e-01 1.46201089e-01
1.05142283e+00 2.39544362e-01 7.26479962e-02 -4.12639439e-01
7.07823396e-01 2.45326713e-01 1.62969649e-01 8.40716004e-01
-7.67946169e-02 5.12515247e-01 1.94233302e-02 -3.45983326e-01
-1.19557202e+00 -9.50508833e-01 -3.40105772e-01 1.07877862e+00
1.44080713e-01 -5.00516653e-01 -9.02737975e-01 -7.94108212e-01
3.00630461e-02 1.06839383e+00 -5.27250409e-01 -4.71947968e-01
-4.29760784e-01 -5.92097819e-01 7.21665502e-01 6.14057481e-01
9.53975081e-01 -8.32384348e-01 -7.88337350e-01 5.89739718e-02
1.35185465e-01 -1.07693863e+00 -6.62633240e-01 3.05015415e-01
-9.77071822e-01 -8.59688163e-01 -4.28447902e-01 -8.40629220e-01
8.71491194e-01 5.10410368e-02 8.09172273e-01 -1.05161592e-01
-5.30094445e-01 3.80554289e-01 -4.85063374e-01 -4.33251053e-01
-4.87490952e-01 1.78970233e-01 1.24071918e-01 1.51517153e-01
3.65384340e-01 -3.24123502e-02 -3.94569397e-01 3.66694093e-01
-7.26116776e-01 8.22638646e-02 3.49912524e-01 9.58483875e-01
6.86812937e-01 1.24706410e-01 1.65827647e-01 -8.72360766e-01
6.61786437e-01 -4.17628847e-02 -1.00582075e+00 4.72387791e-01
-1.04280388e+00 1.33180261e-01 1.00632560e+00 -5.28858900e-01
-1.34606802e+00 3.39873105e-01 1.84135541e-01 -4.33723450e-01
-2.41682872e-01 3.52508456e-01 4.83548909e-04 -3.67042243e-01
9.50454712e-01 4.49222833e-01 -3.80482703e-01 -3.49378198e-01
3.39345604e-01 8.10184538e-01 6.30994678e-01 -5.35131395e-01
8.84032130e-01 2.34006122e-01 -4.19983447e-01 -8.41986895e-01
-6.91729724e-01 -2.99076527e-01 -7.17536032e-01 -3.18278283e-01
5.00184417e-01 -9.02754605e-01 -4.87674803e-01 7.44280994e-01
-7.16496289e-01 -6.21275067e-01 -3.66498232e-01 4.76860315e-01
-5.46068251e-01 3.53042036e-01 -4.13372427e-01 -6.95552766e-01
-8.10014844e-01 -9.33993399e-01 1.01673412e+00 2.17681244e-01
-1.78813457e-01 -8.21514904e-01 -1.08166821e-02 2.74754465e-01
4.80220795e-01 1.16973836e-02 9.14422154e-01 -6.30694628e-01
-3.87493670e-01 -3.77423137e-01 -2.51359373e-01 4.41263497e-01
1.52341336e-01 3.49289209e-01 -9.96810853e-01 -4.20337319e-01
-1.73338071e-01 -5.18557012e-01 6.57183230e-01 2.01632723e-01
1.33348703e+00 -1.49604216e-01 -1.18296936e-01 9.34593976e-01
1.64897883e+00 3.94591808e-01 6.56615973e-01 4.66527492e-01
7.04748750e-01 1.41643465e-01 9.12348032e-01 6.53521538e-01
5.27224541e-02 6.06817484e-01 1.81697428e-01 1.75877064e-01
-1.68335453e-01 -4.03812975e-01 4.42681372e-01 8.62327397e-01
4.38689083e-01 -2.66938418e-01 -1.07444930e+00 4.39852059e-01
-1.65557826e+00 -6.66926205e-01 -6.49919435e-02 2.44731593e+00
8.47112060e-01 5.56763947e-01 -1.50420696e-01 3.03017765e-01
6.64010108e-01 7.11941421e-02 -8.58713210e-01 -7.15376318e-01
-1.37103230e-01 2.03228310e-01 5.77970088e-01 4.65330005e-01
-1.18924451e+00 1.24229050e+00 6.49597454e+00 8.65629137e-01
-1.40181303e+00 -3.21711302e-01 6.12495184e-01 -1.20508306e-01
2.03398690e-01 -1.57460406e-01 -7.23060191e-01 4.80829805e-01
9.46198940e-01 -3.23455274e-01 4.07945186e-01 1.11284542e+00
2.23455757e-01 -2.15991110e-01 -1.17081022e+00 1.21383631e+00
3.31938863e-01 -1.22229648e+00 5.18447347e-02 -3.78376096e-01
7.33405232e-01 -3.08464076e-02 1.09989591e-01 4.29584682e-01
2.10219115e-01 -9.02995288e-01 9.39032376e-01 4.81794357e-01
1.08061337e+00 -7.97069132e-01 4.36493576e-01 2.77183950e-01
-1.19406855e+00 -8.05166811e-02 -4.66294527e-01 3.13143730e-01
-2.25756660e-01 6.22567117e-01 -1.13862002e+00 4.39739585e-01
7.33684301e-01 6.86242104e-01 -8.21094930e-01 1.14722562e+00
-7.65749738e-02 8.56321812e-01 -2.44268686e-01 -2.89228469e-01
1.10105783e-01 -2.76026875e-01 4.15262192e-01 1.69486082e+00
3.06476414e-01 -1.78085878e-01 -6.81498721e-02 6.48620665e-01
-1.52785748e-01 3.83485407e-01 -5.23207486e-01 4.75310313e-04
5.41099131e-01 1.19156289e+00 -7.47388422e-01 -4.62150186e-01
3.62173505e-02 1.37241030e+00 1.63208589e-01 1.10886045e-01
-1.10081840e+00 -7.51211941e-01 1.90969288e-01 -3.03384811e-01
5.78748465e-01 -1.60813659e-01 -6.10577881e-01 -1.05179214e+00
3.00749660e-01 -1.04024577e+00 4.21024591e-01 -6.24330223e-01
-1.05304706e+00 6.91005111e-01 -1.28624335e-01 -1.33505285e+00
-1.68630388e-02 -5.69065869e-01 -5.42573333e-01 5.09698212e-01
-1.06006074e+00 -7.40648866e-01 -7.35465765e-01 5.10174394e-01
9.56444144e-01 -3.99342924e-01 7.07590103e-01 2.34488726e-01
-7.13903487e-01 1.24096692e+00 4.62014526e-01 2.64888778e-02
8.07843566e-01 -1.13869929e+00 5.63569009e-01 8.73489380e-01
1.73632994e-01 3.58670622e-01 7.67989576e-01 -5.38279593e-01
-1.55072343e+00 -1.46525121e+00 4.93359506e-01 -3.72517258e-01
3.89305472e-01 -6.23184204e-01 -8.37244630e-01 3.40262979e-01
9.24089029e-02 1.11693200e-02 2.58499771e-01 -1.70324191e-01
-6.25023246e-01 -4.54743534e-01 -1.52745938e+00 5.31498909e-01
1.17502987e+00 -4.49262649e-01 -4.08059865e-01 3.76868039e-01
5.08747280e-01 -5.74556053e-01 -9.56700683e-01 3.21071923e-01
3.63629282e-01 -6.11423492e-01 6.12349093e-01 -3.29288960e-01
3.10185462e-01 -2.22938493e-01 -3.71357054e-01 -1.31511772e+00
-8.12480822e-02 -4.65198070e-01 -4.67769541e-02 1.32203972e+00
4.48815912e-01 -2.56831527e-01 9.02501583e-01 4.82970744e-01
-1.64615199e-01 -5.81295848e-01 -8.47573578e-01 -9.27542567e-01
-1.85604587e-01 -5.02210379e-01 2.57072300e-01 9.49142098e-01
-1.21036775e-01 2.17617080e-01 -3.04543585e-01 3.68588865e-02
8.25416923e-01 -1.02614351e-01 8.45346510e-01 -9.60146010e-01
-1.86060593e-01 -3.23848367e-01 -4.26759750e-01 -1.05993533e+00
-1.04070999e-01 -7.55621076e-01 2.34244213e-01 -1.19977856e+00
-3.41240801e-02 -4.03488070e-01 -1.69154242e-01 3.24426532e-01
3.98937650e-02 -7.10215867e-02 1.87475011e-01 -4.48499434e-03
-7.47818053e-01 7.00101495e-01 6.83379412e-01 -4.17946041e-01
-3.31495106e-01 -2.27470025e-01 -3.88184220e-01 5.81656635e-01
8.95733297e-01 -4.53748137e-01 -5.90847611e-01 -6.34401977e-01
-1.07640989e-01 -3.79026145e-01 -2.48296428e-02 -1.26926613e+00
3.63731176e-01 -1.35039046e-01 5.51923275e-01 -7.14953244e-01
3.12912196e-01 -7.60151803e-01 -9.66979265e-02 4.30264443e-01
-6.72064304e-01 -3.84199023e-02 4.40071732e-01 5.59663832e-01
-1.13112917e-02 -1.90017700e-01 1.05169857e+00 3.13878387e-01
-8.40286970e-01 2.98825223e-02 -9.22956765e-02 8.02067742e-02
1.13212180e+00 -5.10915339e-01 -2.66062975e-01 -2.19438836e-01
-2.50510067e-01 9.81908366e-02 4.28140670e-01 4.65395182e-01
8.43320608e-01 -1.16034269e+00 -6.58698618e-01 2.60849386e-01
5.42550445e-01 -1.56405285e-01 2.58247573e-02 5.10138333e-01
-5.55273831e-01 3.68071437e-01 -8.03884044e-02 -8.26912999e-01
-1.36216950e+00 4.06821042e-01 3.80852282e-01 -1.44909263e-01
-6.55323148e-01 9.04042423e-01 -4.15792495e-01 -4.42572206e-01
7.50394523e-01 -4.33717847e-01 1.05738528e-01 -2.17502773e-01
4.17087942e-01 4.05279458e-01 3.79769623e-01 -3.44346672e-01
-3.95544052e-01 6.96677625e-01 -3.57456386e-01 2.88400985e-02
1.17441916e+00 -3.27063501e-02 3.44989121e-01 4.45237666e-01
1.21369576e+00 -1.13795005e-01 -1.39842498e+00 -2.89705455e-01
1.51355803e-01 -6.36310160e-01 4.80982699e-02 -9.92385447e-01
-8.80889535e-01 6.44180715e-01 1.10134673e+00 -1.25234842e-01
1.20779121e+00 -4.81701434e-01 5.07794738e-01 7.85757184e-01
2.99808532e-01 -1.49848020e+00 2.23790407e-01 7.50539005e-01
8.72867227e-01 -1.19872677e+00 -3.35105637e-04 -1.27038017e-01
-7.24658191e-01 1.22256994e+00 8.65470886e-01 -1.82948172e-01
3.94786954e-01 4.17777479e-01 1.80472240e-01 -9.70471278e-02
-7.11214960e-01 2.68337786e-01 2.35830620e-01 3.53924364e-01
4.02491540e-01 -1.14285581e-01 -2.39498049e-01 8.33800584e-02
-2.33803391e-01 1.29884360e-02 2.95226783e-01 9.33646619e-01
-4.61402267e-01 -7.63470292e-01 -3.43328774e-01 5.95816493e-01
-7.58593231e-02 -1.80990592e-01 -4.82650578e-01 6.60099447e-01
-1.38743415e-01 9.38231170e-01 9.92557034e-02 -5.49672782e-01
6.49660110e-01 1.11870147e-01 3.66989613e-01 -3.59917939e-01
-6.26231551e-01 -2.04523712e-01 7.88999572e-02 -4.00018394e-01
-4.90418039e-02 -5.13865650e-01 -1.31213593e+00 -1.64356813e-01
-3.90112042e-01 2.18748804e-02 9.41801190e-01 6.83408856e-01
3.68040174e-01 4.52353984e-01 9.74526942e-01 -3.53004634e-01
-7.84796596e-01 -8.89646471e-01 -1.45074889e-01 5.17509460e-01
4.42352518e-02 -3.93004358e-01 -3.46973032e-01 3.41061205e-01]
|
[11.868977546691895, 2.25020694732666]
|
c02af192-ca17-4f40-99d3-a97ad14bcba6
|
read-and-reap-the-rewards-learning-to-play
|
2302.04449
| null |
https://arxiv.org/abs/2302.04449v2
|
https://arxiv.org/pdf/2302.04449v2.pdf
|
Read and Reap the Rewards: Learning to Play Atari with the Help of Instruction Manuals
|
High sample complexity has long been a challenge for RL. On the other hand, humans learn to perform tasks not only from interaction or demonstrations, but also by reading unstructured text documents, e.g., instruction manuals. Instruction manuals and wiki pages are among the most abundant data that could inform agents of valuable features and policies or task-specific environmental dynamics and reward structures. Therefore, we hypothesize that the ability to utilize human-written instruction manuals to assist learning policies for specific tasks should lead to a more efficient and better-performing agent. We propose the Read and Reward framework. Read and Reward speeds up RL algorithms on Atari games by reading manuals released by the Atari game developers. Our framework consists of a QA Extraction module that extracts and summarizes relevant information from the manual and a Reasoning module that evaluates object-agent interactions based on information from the manual. Auxiliary reward is then provided to a standard A2C RL agent, when interaction is detected. When assisted by our design, A2C improves on 4 games in the Atari environment with sparse rewards, and requires 1000x less training frames compared to the previous SOTA Agent 57 on Skiing, the hardest game in Atari.
|
['Tom M. Mitchell', 'Yuanzhi Li', 'Amos Azaria', 'Paul Pu Liang', 'Yewen Fan', 'Yue Wu']
|
2023-02-09
| null | null | null | null |
['atari-games']
|
['playing-games']
|
[-3.49324420e-02 2.71838546e-01 -1.67777076e-01 -1.14981927e-01
-7.60277510e-01 -7.92394340e-01 5.72698295e-01 -1.69491872e-01
-8.33656609e-01 9.22221005e-01 2.05348328e-01 -2.03049362e-01
-1.90181166e-01 -5.51865697e-01 -6.00729406e-01 -4.50670481e-01
-3.52731571e-02 9.49806213e-01 5.20807683e-01 -5.65165639e-01
2.01594532e-01 2.45467767e-01 -2.12027025e+00 1.40970707e-01
7.29181588e-01 7.72451222e-01 9.59265947e-01 1.08207381e+00
3.94231528e-02 1.56340945e+00 -9.35731709e-01 2.13545099e-01
2.81239599e-01 -3.22631478e-01 -9.23016131e-01 -3.84311676e-01
-1.29695311e-01 -9.67495978e-01 -3.56010973e-01 8.45498204e-01
3.89575243e-01 5.33323109e-01 4.02862161e-01 -1.34336221e+00
-1.23119988e-01 9.65878129e-01 -1.47446752e-01 1.14919439e-01
5.19197881e-01 7.86200643e-01 9.08349872e-01 -1.58339128e-01
8.08456421e-01 1.17683530e+00 -2.13341713e-01 7.72752464e-01
-6.74204290e-01 -6.57851398e-01 2.56689042e-01 5.52122712e-01
-7.63916612e-01 -4.43282276e-01 5.51968873e-01 -4.25212413e-01
1.36137390e+00 6.28286898e-02 7.93473601e-01 1.29564941e+00
-5.09051280e-03 1.32896984e+00 8.01237404e-01 -3.02161306e-01
4.93838847e-01 -1.18548058e-01 4.63804565e-02 9.98451114e-01
3.46330628e-02 4.77871060e-01 -7.68202364e-01 1.18029222e-01
5.67457318e-01 -2.89505571e-01 -2.03827411e-01 -3.38125646e-01
-1.20565140e+00 8.26122582e-01 -5.70417941e-02 8.24894607e-02
-5.75746596e-01 4.75682765e-01 5.27892947e-01 4.68456149e-01
-2.25469887e-01 8.88206482e-01 -7.17440426e-01 -1.07996726e+00
-2.32899353e-01 6.73363209e-01 7.44603336e-01 1.03195214e+00
7.30939448e-01 2.02105120e-01 -1.56512097e-01 4.19762224e-01
3.06613415e-01 9.08628345e-01 6.84718728e-01 -1.49531817e+00
6.45612299e-01 6.95296228e-01 4.76728112e-01 -4.48315233e-01
-6.73481107e-01 -2.07854912e-01 1.52763054e-01 7.12440848e-01
5.34492373e-01 -4.43925649e-01 -4.76772249e-01 1.53949142e+00
3.22441667e-01 -2.07928121e-01 3.26032519e-01 1.14344013e+00
9.68098760e-01 4.73875225e-01 3.19178030e-02 2.27371063e-02
1.40136158e+00 -1.11238706e+00 -9.01500583e-01 -4.62482065e-01
1.09028459e+00 -3.46192181e-01 1.44185424e+00 7.77617931e-01
-1.08284903e+00 -2.89899141e-01 -9.59206820e-01 -1.56419381e-01
-3.25731695e-01 2.75329381e-01 6.93188906e-01 1.94260672e-01
-8.96294534e-01 4.51802790e-01 -9.32783723e-01 -1.90121740e-01
1.33949831e-01 5.39037406e-01 -8.92473292e-03 1.45499066e-01
-9.51673865e-01 1.23859024e+00 4.48931038e-01 -3.01326007e-01
-1.23790622e+00 -2.55457669e-01 -1.10862911e+00 8.48771408e-02
1.19452620e+00 -2.46331871e-01 1.93007684e+00 -6.80379570e-01
-2.31697321e+00 2.46315524e-01 3.01600456e-01 -6.04855955e-01
1.94577724e-01 -4.98208016e-01 8.63630995e-02 1.82841763e-01
-1.29336212e-02 8.32296193e-01 6.73222005e-01 -9.71210182e-01
-1.17146766e+00 -2.90511936e-01 7.41777062e-01 6.31549478e-01
-7.72422366e-03 -3.52136306e-02 -3.67315352e-01 3.51835117e-02
-7.54174292e-01 -1.09217429e+00 -4.49356884e-02 -6.15293503e-01
-1.18733933e-02 -5.57299435e-01 8.92095983e-01 -6.66790903e-01
9.74251509e-01 -1.87510800e+00 4.02705342e-01 2.11168714e-02
3.21789026e-01 2.11981639e-01 -3.39239925e-01 9.53967497e-02
4.58114415e-01 -4.52486843e-01 5.33568144e-01 1.70132786e-01
3.01995367e-01 3.61115843e-01 -1.05839297e-01 4.52913158e-02
-1.17490917e-01 9.37005937e-01 -1.37512207e+00 -4.47978258e-01
3.82089972e-01 9.74993035e-02 -5.59074223e-01 5.13517618e-01
-7.64219463e-01 3.96306396e-01 -7.59213209e-01 3.63478869e-01
-2.90853918e-01 -2.92045414e-01 2.72853732e-01 3.64540517e-01
-1.66778371e-01 8.13175440e-01 -9.49971378e-01 1.94507325e+00
-5.89290619e-01 7.14993417e-01 1.27583057e-01 -5.98728001e-01
6.53691173e-01 2.70231366e-01 4.49086636e-01 -1.05381525e+00
2.97451526e-01 -4.15472686e-02 1.88069910e-01 -6.46456659e-01
7.09982336e-01 8.73714209e-01 -2.17860609e-01 7.98418105e-01
-1.67526335e-01 -3.65129322e-01 5.74517548e-01 4.34970230e-01
1.68887186e+00 7.94506967e-01 5.19197524e-01 1.06211290e-01
2.29391038e-01 6.56712890e-01 2.13260561e-01 9.65431809e-01
-3.05044800e-01 -2.21145689e-01 3.34756792e-01 -5.58458865e-01
-6.16822660e-01 -5.90456665e-01 7.48106539e-01 1.69716394e+00
6.08339384e-02 -6.79115057e-01 -7.68169284e-01 -1.03940618e+00
-3.90133649e-01 1.03809106e+00 -3.55671972e-01 -1.81662947e-01
-6.58299863e-01 7.69720078e-02 2.59549946e-01 4.93192166e-01
4.67917889e-01 -1.76468515e+00 -1.70646238e+00 2.89014518e-01
-2.82987833e-01 -9.00054157e-01 -5.32468915e-01 6.03426158e-01
-3.41274679e-01 -1.24217522e+00 -1.44786626e-01 -4.52461213e-01
3.89757991e-01 3.41004103e-01 1.41100967e+00 1.12543657e-01
-1.90057322e-01 8.38064253e-01 -6.78105235e-01 -6.37961090e-01
-5.34594476e-01 9.54567343e-02 1.87704653e-01 -9.53695059e-01
3.69289219e-01 -7.40334690e-02 -2.36204147e-01 2.57709384e-01
-4.00277257e-01 3.99410963e-01 6.97579086e-01 7.74600446e-01
2.08832249e-01 4.31493148e-02 2.54004329e-01 -7.24455595e-01
7.92601526e-01 -7.78832436e-02 -1.29042792e+00 1.46004096e-01
-4.89961654e-01 6.50304794e-01 6.73203528e-01 -4.97847110e-01
-1.07172489e+00 4.46593136e-01 2.51041859e-01 -1.47569120e-01
-3.33115965e-01 4.41679507e-01 -2.55952105e-02 2.48049662e-01
8.38549912e-01 2.10112825e-01 3.99517491e-02 -4.58143465e-02
4.68525469e-01 6.31951213e-01 5.64550102e-01 -8.07476640e-01
8.15378070e-01 -7.66730309e-02 -3.12978685e-01 -6.08238339e-01
-6.54731929e-01 -4.77710098e-01 -1.67334199e-01 -5.68449736e-01
9.04007256e-01 -7.75216758e-01 -1.64980435e+00 4.12296616e-02
-1.01039457e+00 -1.22822368e+00 -5.19921780e-01 5.70340872e-01
-1.09344661e+00 -1.20561659e-01 -4.40131545e-01 -9.21798110e-01
-2.27102488e-01 -1.55815947e+00 9.39414918e-01 5.65306723e-01
-5.10392666e-01 -5.22136092e-01 1.98358998e-01 5.19460022e-01
2.35323757e-01 -4.58569437e-01 6.05881453e-01 -7.22139537e-01
-7.44477689e-01 2.21476674e-01 1.59231737e-01 -1.03524148e-01
1.48524836e-01 -3.62573713e-01 -7.89010167e-01 -1.36721775e-01
-3.17711562e-01 -9.14618671e-01 2.61427969e-01 3.08372200e-01
8.90019000e-01 -4.10530895e-01 -1.26289651e-01 4.97093610e-02
7.90500402e-01 7.27716863e-01 4.85186040e-01 7.10573971e-01
4.00944769e-01 4.86624986e-01 1.17631698e+00 5.54124951e-01
6.37393594e-01 1.10120881e+00 6.89024448e-01 3.46841455e-01
-5.56571409e-02 -4.39726152e-02 9.70914304e-01 3.97901982e-01
-2.82346159e-01 -5.83135597e-02 -6.78601980e-01 1.72108650e-01
-2.13417959e+00 -1.03185296e+00 2.83014476e-01 1.88472807e+00
9.72371876e-01 1.06788866e-01 3.83105069e-01 -2.00378373e-01
-5.43031283e-02 -1.60336599e-01 -9.09242451e-01 -1.03788286e-01
3.73198479e-01 1.29474118e-01 4.62921888e-01 5.86281896e-01
-8.43307436e-01 1.31226373e+00 5.70941639e+00 6.71573579e-01
-6.25977278e-01 -1.65027559e-01 -5.39591722e-02 -4.64936346e-01
3.42850864e-01 -3.53322446e-01 -9.13205266e-01 1.94316611e-01
1.04645109e+00 -8.38170052e-02 1.12096548e+00 1.34834146e+00
4.97855186e-01 -5.39368212e-01 -1.19468760e+00 9.39572453e-01
-6.70349896e-02 -1.25786221e+00 -3.78029704e-01 1.45112857e-01
2.08164364e-01 1.56442970e-01 6.28495067e-02 8.76245260e-01
1.53446341e+00 -8.05622458e-01 8.37957203e-01 2.15775982e-01
4.50126290e-01 -8.34315896e-01 7.90345967e-01 5.28093994e-01
-1.05694032e+00 -1.70762017e-01 -2.26135999e-01 -2.55095929e-01
-2.72295535e-01 -4.23221678e-01 -1.37767470e+00 3.19632620e-01
7.27040172e-01 6.06855392e-01 -4.96006578e-01 5.56605041e-01
-6.44653320e-01 4.16076243e-01 -1.41070679e-01 -5.86150348e-01
2.58466303e-01 -1.19618133e-01 4.79563296e-01 6.34210527e-01
1.05619922e-01 4.69954818e-01 4.50631529e-01 5.25446713e-01
2.57579118e-01 -1.49158329e-01 -7.70620942e-01 -2.02956200e-01
4.84626323e-01 1.36527014e+00 -5.14944851e-01 -5.09019017e-01
-3.59341174e-01 7.95760036e-01 5.21649659e-01 3.22198689e-01
-8.28780413e-01 -3.24238867e-01 6.67954981e-01 -1.60065427e-01
2.43766248e-01 -3.08670342e-01 3.20624232e-01 -8.41097772e-01
-3.31751257e-01 -1.50161970e+00 4.01724696e-01 -1.19058681e+00
-5.07790864e-01 5.14089644e-01 8.06454048e-02 -1.09170806e+00
-1.09355092e+00 -8.29555035e-01 -1.64919123e-01 3.64591211e-01
-1.39018679e+00 -5.46632588e-01 -4.97466147e-01 7.60753095e-01
9.10553455e-01 -6.15301073e-01 8.56917679e-01 -3.04778069e-01
-5.23648202e-01 1.24280773e-01 -2.76837260e-01 7.14611560e-02
5.75621247e-01 -1.48503470e+00 1.10988848e-01 4.76911962e-01
1.27984732e-01 3.51944149e-01 8.66763711e-01 -7.41952002e-01
-1.80736125e+00 -7.19419777e-01 7.22603351e-02 -5.94819963e-01
5.82775891e-01 -9.42394808e-02 -4.21781987e-01 7.50940263e-01
3.33500326e-01 -5.51250339e-01 3.82496148e-01 9.90596116e-02
-7.94083551e-02 9.48148295e-02 -8.43947113e-01 9.42630589e-01
1.06291521e+00 -2.63561189e-01 -7.82991886e-01 3.99075240e-01
8.34216058e-01 -7.47416139e-01 -3.76158267e-01 -3.83396268e-01
3.81610990e-01 -6.34554446e-01 8.22682142e-01 -6.78957283e-01
2.49426141e-01 -6.13063335e-01 -1.67622103e-03 -1.63307369e+00
-7.20788762e-02 -9.15059626e-01 -3.83573234e-01 5.74635148e-01
3.65966529e-01 -3.17618638e-01 7.81267762e-01 6.92466021e-01
-2.43821189e-01 -1.47352114e-01 -4.58185762e-01 -7.94874966e-01
-5.81841230e-01 -6.07738256e-01 5.84480405e-01 3.15735608e-01
5.26943266e-01 6.21655881e-01 -2.60106653e-01 5.12459502e-02
4.64991540e-01 -8.63353908e-02 1.26994669e+00 -1.20023298e+00
-5.76454282e-01 -4.01553422e-01 1.06528386e-01 -1.20865953e+00
4.19588417e-01 -5.88455617e-01 6.53340876e-01 -1.57622898e+00
6.38872236e-02 -3.50321770e-01 2.02649474e-01 8.78714919e-01
2.74768993e-02 -3.76623750e-01 2.98971742e-01 1.77566737e-01
-1.40249097e+00 4.16170299e-01 1.36157620e+00 -2.45246544e-01
-8.02123904e-01 -1.30156368e-01 -4.59953815e-01 8.15022767e-01
8.11082900e-01 -2.58914649e-01 -6.87271953e-01 -3.05530190e-01
5.11676252e-01 3.35981488e-01 1.28393695e-01 -1.21553397e+00
4.07340378e-01 -5.27712405e-01 2.63617653e-02 -6.46293402e-01
4.58520293e-01 -9.38008845e-01 -1.64068803e-01 5.06117105e-01
-6.47178411e-01 1.03490360e-01 1.90693647e-01 2.67014474e-01
3.11586916e-01 -4.63542879e-01 2.67546862e-01 -4.39674675e-01
-1.00234234e+00 -7.38893300e-02 -9.88169909e-01 1.16241500e-02
1.05371833e+00 7.71291405e-02 -6.93115711e-01 -8.07810605e-01
-4.41968113e-01 5.61356187e-01 3.43834341e-01 4.08823937e-01
5.87808788e-01 -9.38281894e-01 -2.78854012e-01 -4.15905491e-02
9.36793387e-02 1.00173421e-01 -1.03956744e-01 4.16787118e-01
-2.89672613e-01 5.66406727e-01 -4.26495641e-01 -2.61905283e-01
-1.48077643e+00 4.30477381e-01 1.00107655e-01 -5.87164581e-01
-8.76870811e-01 4.36588675e-01 1.90280393e-01 -3.73837441e-01
6.19855106e-01 -5.48049092e-01 -7.69426763e-01 -2.04717591e-01
9.46226418e-01 3.26852143e-01 -1.31592438e-01 -1.44996062e-01
-2.38794759e-01 9.74144638e-02 -3.26736569e-01 -3.78631353e-01
1.46791244e+00 4.96330187e-02 4.98841912e-01 2.70526320e-01
2.07756937e-01 -5.03022932e-02 -1.78722358e+00 -1.55315280e-01
1.62966505e-01 -1.96357369e-01 1.81733072e-01 -1.28394437e+00
-8.28834236e-01 5.08419693e-01 3.59515339e-01 8.94168466e-02
7.71882117e-01 1.12175785e-01 5.51520944e-01 1.16908109e+00
9.54036593e-01 -1.48399425e+00 6.89243674e-01 7.55535245e-01
7.98158705e-01 -1.22237456e+00 -6.49115816e-02 2.16734827e-01
-1.05502164e+00 1.00675857e+00 1.14549363e+00 3.32830429e-01
-2.31061921e-01 5.24556339e-01 1.26549900e-01 -4.65241283e-01
-1.23342586e+00 -5.39513767e-01 -6.43786862e-02 1.12743330e+00
6.52781352e-02 1.65546000e-01 2.85551608e-01 7.76593089e-01
-4.49526876e-01 7.51264542e-02 8.46943259e-01 1.12780833e+00
-7.01427996e-01 -8.87259185e-01 -2.86350936e-01 4.74102259e-01
8.60486552e-02 1.61271423e-01 -4.17640150e-01 9.00069416e-01
-3.11868876e-01 1.15341473e+00 -8.94402787e-02 -4.80889976e-01
3.01693588e-01 -6.81083649e-02 4.82526243e-01 -8.05875003e-01
-8.43607426e-01 -2.39059016e-01 3.97332191e-01 -1.06072474e+00
-1.96647376e-01 -7.07185566e-01 -1.81720781e+00 -2.03462169e-01
-2.36438081e-01 1.22713484e-01 6.48747861e-01 1.10868454e+00
2.67724365e-01 9.02707696e-01 -3.27447639e-03 -9.39011216e-01
-5.86625278e-01 -8.74595046e-01 -1.44264907e-01 6.99961931e-02
9.56200287e-02 -1.02183986e+00 -1.55323952e-01 5.61330793e-03]
|
[4.036898612976074, 1.436249017715454]
|
36506e60-b4c4-408d-9284-9f6945771452
|
3d-intracranial-aneurysm-classification-and
|
2201.02198
| null |
https://arxiv.org/abs/2201.02198v2
|
https://arxiv.org/pdf/2201.02198v2.pdf
|
3D Intracranial Aneurysm Classification and Segmentation via Unsupervised Dual-branch Learning
|
Intracranial aneurysms are common nowadays and how to detect them intelligently is of great significance in digital health. While most existing deep learning research focused on medical images in a supervised way, we introduce an unsupervised method for the detection of intracranial aneurysms based on 3D point cloud data. In particular, our method consists of two stages: unsupervised pre-training and downstream tasks. As for the former, the main idea is to pair each point cloud with its jittered counterpart and maximise their correspondence. Then we design a dual-branch contrastive network with an encoder for each branch and a subsequent common projection head. As for the latter, we design simple networks for supervised classification and segmentation training. Experiments on the public dataset (IntrA) show that our unsupervised method achieves comparable or even better performance than some state-of-the-art supervised techniques, and it is most prominent in the detection of aneurysmal vessels. Experiments on the ModelNet40 also show that our method achieves the accuracy of 90.79\% which outperforms existing state-of-the-art unsupervised models.
|
['Xiao Liu', 'Xuequan Lu', 'Di Shao']
|
2022-01-06
| null | null | null | null |
['unsupervised-pre-training']
|
['methodology']
|
[-2.18443781e-01 4.77207959e-01 -5.31107970e-02 -5.39716959e-01
-5.05984068e-01 -1.99598051e-03 5.04206836e-01 1.30064413e-01
-5.34671843e-01 2.12711498e-01 1.09619103e-01 -5.51986814e-01
-1.51572386e-02 -6.48864150e-01 -6.26964986e-01 -7.90883899e-01
-4.23453480e-01 1.02905846e+00 5.77919483e-01 1.03456445e-01
2.06920564e-01 7.60243833e-01 -1.00463796e+00 1.88212633e-01
7.74279475e-01 1.22917366e+00 7.34335557e-02 2.58447140e-01
-3.95544231e-01 8.31736803e-01 -5.49584627e-01 -4.94153678e-01
3.64429861e-01 1.25381663e-01 -9.33506310e-01 3.22796702e-01
1.99782431e-01 -4.76985067e-01 -3.21187973e-01 1.07990253e+00
6.85704827e-01 -3.78194422e-01 8.94597650e-01 -8.60773861e-01
-3.26559514e-01 5.64414263e-01 -7.51608133e-01 5.36400855e-01
-3.50187004e-01 -1.23588473e-01 6.22306526e-01 -7.18630612e-01
5.44709802e-01 8.73210847e-01 6.58880234e-01 3.85421127e-01
-8.97632480e-01 -7.01993465e-01 6.40817434e-02 1.43919334e-01
-1.10042417e+00 -3.19290102e-01 7.87078977e-01 -6.95180774e-01
6.11967444e-01 -2.14053482e-01 7.50496805e-01 8.77895176e-01
2.48416007e-01 9.21688080e-01 8.21602881e-01 -2.97926664e-01
4.34297532e-01 -2.04253402e-02 1.67718694e-01 7.31368721e-01
1.06010295e-01 2.64275491e-01 -3.46665233e-02 7.13898093e-02
1.14410746e+00 7.58904070e-02 1.58001035e-01 -7.67171443e-01
-1.10921657e+00 1.01017702e+00 9.18718338e-01 4.00916934e-01
-5.46402931e-01 -1.45903587e-01 4.85804975e-01 -7.84000382e-02
6.01783693e-01 4.93167460e-01 -4.63233203e-01 2.40248486e-01
-1.08867168e+00 4.78954725e-02 7.54524767e-01 9.91945982e-01
7.22411647e-02 -2.89255768e-01 1.74487993e-01 7.29617894e-01
3.47728878e-01 4.20467481e-02 7.00668693e-01 -7.14412868e-01
2.74975985e-01 5.30308783e-01 -2.71713763e-01 -8.08517814e-01
-8.72002482e-01 -8.92746627e-01 -1.06144154e+00 4.18153137e-01
4.96974885e-01 -1.99262500e-01 -1.29736125e+00 1.09592152e+00
2.86628932e-01 3.06590945e-01 -1.15729831e-01 9.54069078e-01
1.34592867e+00 2.40495250e-01 1.39216986e-02 1.18231168e-02
1.44228590e+00 -1.20099103e+00 -3.22571874e-01 -1.46942616e-01
8.04539621e-01 -5.42663872e-01 3.33997726e-01 3.75792891e-01
-1.09224319e+00 -3.53134453e-01 -7.32830226e-01 1.09435216e-01
-3.88647050e-01 6.24143556e-02 7.56825864e-01 4.66006190e-01
-1.00927114e+00 2.64256984e-01 -1.21367311e+00 -1.25836179e-01
9.29183900e-01 3.35500687e-01 -2.55818665e-01 3.97036374e-01
-8.38462234e-01 9.73957837e-01 2.57576704e-01 -4.24300954e-02
-6.47620142e-01 -7.94342816e-01 -5.78300357e-01 -1.21611625e-01
-5.67233823e-02 -7.91038334e-01 1.43763566e+00 -7.92777002e-01
-1.54770958e+00 1.28400373e+00 8.45810175e-02 -1.11298001e+00
8.11308086e-01 -2.28258297e-01 -2.50046849e-01 5.98801851e-01
3.88966978e-01 1.02516925e+00 7.91584671e-01 -1.26499760e+00
-9.32918668e-01 -4.13765877e-01 -2.41670683e-01 -8.41919780e-02
-2.56321847e-01 1.72875315e-01 -5.95586777e-01 -6.06503785e-01
5.70819497e-01 -8.15161705e-01 -7.49951184e-01 1.96430504e-01
-6.61688507e-01 -2.68935919e-01 6.71222627e-01 -7.26974010e-01
7.39529610e-01 -2.00669861e+00 -1.08376905e-01 3.01795930e-01
8.56324375e-01 3.04930031e-01 4.28907514e-01 -3.35556179e-01
-4.33912575e-01 -1.47050977e-01 -4.58765864e-01 -6.41395152e-01
-3.95386815e-01 4.31172848e-02 -1.52920365e-01 6.79363430e-01
9.38309804e-02 8.69184196e-01 -8.17756772e-01 -7.38178313e-01
4.03794289e-01 4.45346326e-01 -6.43919647e-01 1.90969817e-02
2.74092257e-01 1.05656898e+00 -5.77843547e-01 5.55626094e-01
7.30501890e-01 -6.90850735e-01 -2.97785074e-01 7.40197226e-02
-2.09466830e-01 2.03121021e-01 -7.28762388e-01 1.71691358e+00
-2.18766123e-01 5.17767489e-01 -1.47310272e-01 -1.59338629e+00
1.02689564e+00 4.19122607e-01 9.36854899e-01 -6.92013502e-01
3.87373030e-01 5.30610144e-01 1.47852629e-01 -6.64338708e-01
-1.41099408e-01 8.50910097e-02 2.64804691e-01 -2.88581997e-02
-1.72104109e-02 -8.20208527e-03 -7.26172626e-02 1.48739606e-01
1.09091318e+00 -3.18036318e-01 1.86584115e-01 -1.92730337e-01
5.34952581e-01 6.38334155e-02 3.67225885e-01 7.09848225e-01
-4.60487992e-01 9.14061248e-01 5.85069597e-01 -8.29646885e-01
-1.03385639e+00 -1.08234310e+00 -7.67072320e-01 3.88000429e-01
1.97437167e-01 -1.22434855e-01 -7.54376352e-01 -8.45947504e-01
-1.34737138e-02 4.51670945e-01 -5.75665832e-01 1.73444629e-01
-8.24526250e-01 -8.56679082e-01 2.06593856e-01 7.57374704e-01
7.29847610e-01 -1.09843791e+00 -9.52294469e-01 2.47007757e-01
6.05607070e-02 -1.34096336e+00 1.76759884e-01 1.88854337e-01
-1.18374670e+00 -9.14632618e-01 -1.14706469e+00 -1.18877423e+00
8.86854172e-01 -1.61474928e-01 1.21248615e+00 -5.03537990e-02
-2.77341008e-01 -1.76003143e-01 -3.84616971e-01 -7.81848550e-01
-1.52595043e-01 3.00442934e-01 -3.74060988e-01 -9.50795114e-02
5.99979758e-01 -7.10043073e-01 -9.62807894e-01 1.82532132e-01
-3.59564185e-01 -7.15451613e-02 8.02679896e-01 7.57473469e-01
7.09998846e-01 -1.62377119e-01 6.26912832e-01 -9.25686419e-01
2.35106111e-01 -5.32952905e-01 -6.23375893e-01 -3.47828448e-01
-4.88760173e-01 -3.17685544e-01 5.55821240e-01 -1.57745499e-02
-7.16028392e-01 4.12666529e-01 -4.67598766e-01 -4.97989953e-01
-5.94540656e-01 3.20416629e-01 3.26827586e-01 -5.44347726e-02
6.02121174e-01 2.83430032e-02 2.83421606e-01 -6.14044249e-01
3.33147198e-01 9.06328201e-01 7.41109848e-01 -1.96084172e-01
3.73419821e-01 8.96287084e-01 1.04670133e-02 -7.62944281e-01
-7.54439890e-01 -8.61585617e-01 -9.78634298e-01 -4.66707461e-02
1.11941576e+00 -6.97846055e-01 -3.89859766e-01 5.46892643e-01
-1.32244337e+00 5.83225265e-02 -3.58930588e-01 6.59737051e-01
-6.10224962e-01 3.35661381e-01 -7.38319695e-01 -3.70835334e-01
-5.67105651e-01 -1.60670221e+00 1.12917590e+00 2.29773857e-03
1.06511757e-01 -9.45752800e-01 -1.05370618e-01 2.25168556e-01
3.93086851e-01 1.35441676e-01 7.38047957e-01 -1.06841397e+00
-3.97070885e-01 -3.94739836e-01 -4.65190023e-01 2.78644770e-01
-1.18367463e-01 -2.14182377e-01 -8.07471931e-01 -8.41403082e-02
2.04937905e-01 2.27232188e-01 9.91327524e-01 1.02733827e+00
1.60986626e+00 1.02211617e-01 -7.54446983e-01 9.60324585e-01
9.91299391e-01 2.86474109e-01 6.45086408e-01 8.43246222e-01
5.84780872e-01 5.01954019e-01 -5.14416117e-03 2.56522000e-01
2.47404397e-01 5.71056664e-01 7.50330985e-01 -5.68212390e-01
-7.66870454e-02 3.23569715e-01 -3.46030593e-01 9.18984413e-01
-1.72986373e-01 3.46558392e-01 -1.19151998e+00 5.82517922e-01
-1.83445334e+00 -7.21044898e-01 -2.86649734e-01 2.03299594e+00
5.32278657e-01 4.61752921e-01 2.96744704e-01 1.49346711e-02
7.19728112e-01 -7.36214668e-02 -4.34471369e-01 -1.17376529e-01
4.35430259e-02 4.43937480e-01 7.92895019e-01 -1.21309094e-01
-1.81344008e+00 7.28259861e-01 6.45910263e+00 3.69991899e-01
-1.33239162e+00 1.71673089e-01 1.01228023e+00 2.33818516e-01
5.08921862e-01 -4.06146556e-01 -6.51409626e-01 6.65339649e-01
7.18916118e-01 2.53771365e-01 -3.80712003e-01 1.20721388e+00
1.39771923e-01 2.46778011e-01 -8.77444983e-01 1.03456652e+00
-2.80717045e-01 -1.55926836e+00 -1.64168745e-01 1.61690071e-01
6.75600708e-01 4.76430625e-01 1.19755805e-01 1.11556508e-01
1.44066438e-01 -9.53647792e-01 6.70818090e-01 3.77279311e-01
4.68716681e-01 -6.87024653e-01 1.03312206e+00 3.44199449e-01
-9.31981564e-01 -4.09546085e-02 -3.99891049e-01 2.41175190e-01
4.94235963e-01 8.99614513e-01 -8.73058081e-01 3.64664555e-01
9.90886271e-01 9.08714354e-01 -4.29765046e-01 1.94506359e+00
-5.18166542e-01 6.18614614e-01 -3.42922479e-01 3.83177906e-01
4.89422232e-01 -1.27468854e-01 7.27226436e-01 1.26055741e+00
8.02047253e-02 -1.61571503e-01 2.79250771e-01 7.80321419e-01
6.19795471e-02 1.90553814e-01 -3.44960839e-01 6.68601871e-01
2.93186814e-01 1.13137746e+00 -9.96395350e-01 -6.17481530e-01
-6.75353587e-01 5.67162752e-01 3.32714021e-01 -1.69055723e-02
-7.72603214e-01 -2.82840669e-01 2.15367660e-01 1.58349544e-01
4.92931306e-01 -7.18755350e-02 -7.32123911e-01 -9.42078054e-01
-2.16558743e-02 -4.71403569e-01 3.33867699e-01 -4.83401477e-01
-1.57127070e+00 7.64884889e-01 -1.29974589e-01 -1.50883269e+00
-1.96709111e-01 -8.15020204e-01 -8.39351654e-01 6.32471442e-01
-1.71514881e+00 -1.07025445e+00 -2.61918187e-01 4.72021312e-01
6.00403905e-01 -3.68946791e-01 7.17949092e-01 4.12187397e-01
-6.00415409e-01 3.71190190e-01 1.51828006e-01 6.93344891e-01
5.27318895e-01 -1.37819064e+00 7.31576741e-01 8.33040416e-01
-1.33159727e-01 4.15774167e-01 4.49079007e-01 -5.19279182e-01
-4.54293102e-01 -1.23941946e+00 7.02710807e-01 -2.23544523e-01
6.16922736e-01 1.00874193e-01 -8.34375918e-01 7.54549623e-01
4.48394418e-02 3.76181960e-01 3.21846128e-01 -1.67371422e-01
-5.50715998e-02 1.77658468e-01 -1.12815797e+00 2.90332824e-01
1.04518604e+00 2.85701044e-02 -8.87249887e-01 7.65323818e-01
6.98844492e-01 -7.21953213e-01 -7.50014186e-01 5.26588619e-01
3.15272123e-01 -1.05990851e+00 1.16909480e+00 -6.85833752e-01
8.83888841e-01 -2.53315829e-02 4.74223435e-01 -1.28072882e+00
-3.16614479e-01 -3.25646758e-01 -2.13495150e-01 5.10977149e-01
4.88257974e-01 -8.80798697e-01 1.24542665e+00 2.32554853e-01
-5.98545730e-01 -1.07423949e+00 -1.03331745e+00 -7.62405157e-01
5.65280497e-01 -2.89647400e-01 5.54308593e-01 9.30492997e-01
-9.30296630e-02 2.16505509e-02 2.07221806e-01 4.68642503e-01
8.72481942e-01 8.96095634e-02 5.75534284e-01 -1.44583654e+00
7.08255777e-03 -8.63151968e-01 -9.00055826e-01 -1.43971431e+00
-5.81157729e-02 -1.00876129e+00 -8.16160813e-02 -1.62283969e+00
-6.12265579e-02 -6.59153223e-01 -3.16696495e-01 4.20174934e-02
1.77798793e-01 -1.48783535e-01 -2.54949391e-01 5.75355411e-01
-3.73461902e-01 1.91577747e-01 1.36052799e+00 -1.29732206e-01
-2.99322546e-01 3.90296966e-01 -3.85127157e-01 1.08212698e+00
8.25260758e-01 -3.61618578e-01 -4.34626117e-02 -5.27090013e-01
-4.60633427e-01 -1.18327066e-01 5.28151572e-01 -1.27430999e+00
6.05060518e-01 5.94048142e-01 3.63138288e-01 -8.95273626e-01
7.33342171e-02 -7.93987453e-01 -4.80397433e-01 6.46217823e-01
-2.12600172e-01 -7.83174932e-02 -1.25133947e-01 3.38500410e-01
-4.02231246e-01 -2.58767754e-01 1.07978928e+00 -3.61918598e-01
-6.49458468e-01 5.83745003e-01 -3.08716267e-01 -9.83047560e-02
1.06112993e+00 -1.31625503e-01 -2.03328267e-01 -1.08782336e-01
-1.08476627e+00 2.42756799e-01 -5.82778677e-02 2.88722575e-01
5.84704816e-01 -8.74872625e-01 -7.35160053e-01 3.84833187e-01
1.12231500e-01 6.49463296e-01 -1.10261329e-01 1.23829782e+00
-8.74271154e-01 6.05260015e-01 -2.85750479e-01 -1.07453084e+00
-7.91064739e-01 5.37563682e-01 7.42854357e-01 -2.34477267e-01
-1.33935916e+00 1.10313034e+00 4.41737115e-01 -4.07739550e-01
6.03170693e-01 -3.12359303e-01 -5.37730277e-01 -2.65262365e-01
4.59440589e-01 1.31860241e-01 4.50969905e-01 -7.08162963e-01
-3.01640064e-01 5.41703820e-01 -5.16555905e-01 1.87300131e-01
1.65226829e+00 2.59165764e-01 -8.03225338e-02 -8.19239318e-02
1.38376582e+00 -4.97236520e-01 -9.06519830e-01 -3.64632756e-01
-2.73485035e-02 -1.58011287e-01 3.26637298e-01 -4.93220896e-01
-1.65529895e+00 9.87933636e-01 8.38225305e-01 3.10847580e-01
8.66423965e-01 2.75590628e-01 9.56869721e-01 2.27411538e-01
3.44396889e-01 -7.81545639e-01 -2.14020506e-01 4.34349000e-01
7.23663628e-01 -1.38827825e+00 -1.48784161e-01 -7.25535989e-01
-2.49445260e-01 1.36788201e+00 4.54820454e-01 -4.64623183e-01
1.23750114e+00 2.27087408e-01 1.60779387e-01 -7.26261735e-01
-1.46438494e-01 -1.35564283e-01 1.68937787e-01 4.58071887e-01
2.98386604e-01 1.69519156e-01 -2.41712257e-01 5.12569368e-01
-5.68665028e-01 1.02201872e-01 1.57810897e-01 1.05732346e+00
-4.97548312e-01 -7.56802917e-01 -2.19497338e-01 8.30206752e-01
-5.85191250e-01 -4.01999727e-02 -2.44672056e-02 7.14297473e-01
1.45233676e-01 6.43428802e-01 5.31812668e-01 -9.49240401e-02
4.11075681e-01 -2.59301364e-01 1.02507643e-01 -5.20067573e-01
-5.72986782e-01 2.17860445e-01 -1.68202847e-01 -5.39063394e-01
-3.74794900e-01 -7.46503532e-01 -1.39308178e+00 1.10989235e-01
-3.38131219e-01 -8.82446095e-02 8.49243402e-01 9.93067682e-01
4.05094534e-01 7.14276969e-01 5.92167139e-01 -9.26640987e-01
-5.36513805e-01 -8.14095616e-01 -4.95959073e-01 5.07639766e-01
2.47704327e-01 -9.01223421e-01 -4.23546970e-01 1.68509990e-01]
|
[14.468547821044922, -2.260490655899048]
|
e69cea76-435c-4f0f-9c23-50b982b016e4
|
deep-compositional-captioning-describing
|
1511.05284
| null |
http://arxiv.org/abs/1511.05284v2
|
http://arxiv.org/pdf/1511.05284v2.pdf
|
Deep Compositional Captioning: Describing Novel Object Categories without Paired Training Data
|
While recent deep neural network models have achieved promising results on
the image captioning task, they rely largely on the availability of corpora
with paired image and sentence captions to describe objects in context. In this
work, we propose the Deep Compositional Captioner (DCC) to address the task of
generating descriptions of novel objects which are not present in paired
image-sentence datasets. Our method achieves this by leveraging large object
recognition datasets and external text corpora and by transferring knowledge
between semantically similar concepts. Current deep caption models can only
describe objects contained in paired image-sentence corpora, despite the fact
that they are pre-trained with large object recognition datasets, namely
ImageNet. In contrast, our model can compose sentences that describe novel
objects and their interactions with other objects. We demonstrate our model's
ability to describe novel concepts by empirically evaluating its performance on
MSCOCO and show qualitative results on ImageNet images of objects for which no
paired image-caption data exist. Further, we extend our approach to generate
descriptions of objects in video clips. Our results show that DCC has distinct
advantages over existing image and video captioning approaches for generating
descriptions of new objects in context.
|
['Lisa Anne Hendricks', 'Kate Saenko', 'Raymond Mooney', 'Trevor Darrell', 'Subhashini Venugopalan', 'Marcus Rohrbach']
|
2015-11-17
|
deep-compositional-captioning-describing-1
|
http://openaccess.thecvf.com/content_cvpr_2016/html/Hendricks_Deep_Compositional_Captioning_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Hendricks_Deep_Compositional_Captioning_CVPR_2016_paper.pdf
|
cvpr-2016-6
|
['novel-concepts']
|
['reasoning']
|
[ 5.18913686e-01 2.95515150e-01 2.64410049e-01 -5.95777214e-01
-9.65618908e-01 -5.06926477e-01 1.12250102e+00 9.71858948e-02
-3.22716177e-01 8.26088071e-01 5.78921795e-01 5.94720952e-02
3.71254951e-01 -5.56960642e-01 -1.42065322e+00 -1.90212294e-01
1.35009065e-01 6.15012288e-01 2.49988854e-01 -3.08928043e-01
4.52934280e-02 1.90897569e-01 -1.68292618e+00 1.00240922e+00
1.49314120e-01 7.81121254e-01 5.77259123e-01 6.97935879e-01
-2.10587874e-01 1.14871478e+00 -7.20346689e-01 -5.03583252e-01
1.33215249e-01 -6.83056056e-01 -1.01338434e+00 5.60644567e-01
1.00733888e+00 -6.63505018e-01 -5.30430853e-01 7.34418571e-01
2.79814750e-01 3.29947546e-02 6.81735992e-01 -1.58662558e+00
-1.37405217e+00 8.39170516e-01 2.39879545e-02 1.83649704e-01
4.91399467e-01 4.16528106e-01 8.37285101e-01 -8.75399530e-01
1.08877993e+00 1.22168386e+00 4.24907655e-01 1.01203287e+00
-1.43938208e+00 -4.52588111e-01 9.13995877e-02 3.89563322e-01
-1.17426932e+00 -5.12001753e-01 7.67016053e-01 -3.74457151e-01
1.12565410e+00 9.17138830e-02 5.69362223e-01 1.52172244e+00
-3.77692372e-01 9.09597099e-01 6.25889242e-01 -3.67351264e-01
2.43337035e-01 3.52306873e-01 -4.48198199e-01 2.03198463e-01
-3.81816551e-02 -6.14740374e-03 -4.12971705e-01 2.50372916e-01
8.42945218e-01 -1.00094654e-01 -2.01298162e-01 -2.87287384e-01
-1.56008196e+00 6.45547032e-01 6.10046446e-01 4.30398464e-01
-4.91915792e-01 8.11624348e-01 5.40243626e-01 -5.30061238e-02
2.18692735e-01 6.49501145e-01 -1.58651784e-01 3.82545590e-02
-8.43952119e-01 6.23703182e-01 4.90448833e-01 1.57470536e+00
6.14660323e-01 4.26534712e-02 -5.01496434e-01 5.80584586e-01
-2.04601973e-01 5.53112507e-01 6.12933874e-01 -1.16072714e+00
5.22276163e-01 3.43591720e-01 1.37544975e-01 -9.69424307e-01
3.00208814e-02 -1.39834419e-01 -5.19867897e-01 -3.13912481e-01
-1.40193075e-01 2.22258717e-01 -8.61210823e-01 1.86986482e+00
-2.18639284e-01 2.52971560e-01 5.99527121e-01 9.58094358e-01
1.27200079e+00 9.43041563e-01 5.62456310e-01 8.93060863e-02
1.25023985e+00 -1.10072112e+00 -4.87769246e-01 -4.25404608e-01
5.11362195e-01 -5.01719832e-01 9.98839974e-01 -3.61079931e-01
-1.47703075e+00 -9.10910428e-01 -6.22645676e-01 -2.72198021e-01
-4.37201351e-01 -1.34484962e-01 3.45842063e-01 9.42202359e-02
-1.44171357e+00 1.34782419e-01 -3.64483088e-01 -4.45362061e-01
6.51860714e-01 1.24561422e-01 -5.88453650e-01 -2.82351941e-01
-1.03894532e+00 9.64643776e-01 8.09875131e-01 -2.72720814e-01
-1.45773995e+00 -8.35343122e-01 -1.21906459e+00 3.17880124e-01
1.75326392e-01 -9.38794792e-01 1.59203243e+00 -1.63553548e+00
-6.89604938e-01 1.02361190e+00 3.62334400e-02 -8.79596829e-01
2.05764040e-01 2.09462062e-01 -2.72387832e-01 8.00347388e-01
4.04674679e-01 1.95223105e+00 4.98327643e-01 -1.85340440e+00
-5.69512129e-01 1.53535515e-01 5.91810942e-01 2.35498786e-01
-3.58919859e-01 2.46923976e-02 -4.87143308e-01 -6.14964843e-01
-4.74251986e-01 -7.40902841e-01 -2.28609964e-01 6.47863448e-02
-2.49270365e-01 2.86230780e-02 9.25973713e-01 -4.23802465e-01
5.16373098e-01 -2.11518526e+00 -1.12154730e-01 -4.71601456e-01
-5.84705323e-02 3.18463445e-01 -6.95142567e-01 6.44614398e-01
-2.49004364e-01 2.82077998e-01 -3.50882173e-01 -5.85114419e-01
-2.95431651e-02 4.66842771e-01 -5.78097999e-01 -1.30768314e-01
8.31602335e-01 1.46784663e+00 -1.12832153e+00 -7.73349941e-01
1.71998322e-01 5.48318326e-01 -6.94550812e-01 2.14127883e-01
-7.93631315e-01 3.39742035e-01 -1.62074417e-01 3.82508039e-01
4.36395407e-01 -4.83719885e-01 -3.42245284e-03 -4.09674346e-01
2.43188933e-01 -8.21814165e-02 -4.29046005e-01 1.77824318e+00
-4.50766146e-01 9.69699264e-01 -3.76479149e-01 -1.23498833e+00
7.09990561e-01 6.96478546e-01 4.18561339e-01 -8.00477386e-01
4.08913795e-04 -3.18849436e-03 -2.38613889e-01 -9.42524552e-01
6.68435156e-01 -5.21718323e-01 -1.77593917e-01 5.46160340e-01
3.09343934e-01 -5.10731339e-01 5.02890825e-01 6.66903198e-01
1.10813677e+00 2.66086966e-01 -1.41495228e-01 6.92760050e-02
4.96165305e-01 4.23174232e-01 -3.58413234e-02 9.30612504e-01
-2.01008305e-01 1.21455646e+00 2.27944925e-01 -4.37644154e-01
-1.60583246e+00 -1.27348268e+00 9.07588452e-02 8.08026195e-01
2.48387232e-01 -3.32297564e-01 -6.34898305e-01 -6.02376759e-01
-2.67455041e-01 8.56386900e-01 -8.22186351e-01 -1.52950242e-01
-5.66676319e-01 8.39473214e-03 3.47235173e-01 7.81192899e-01
5.72789788e-01 -1.75098026e+00 -5.74158728e-01 3.69573802e-01
-5.74765444e-01 -1.79881871e+00 -4.61990625e-01 -3.77241373e-01
-6.13607407e-01 -9.47720408e-01 -8.95217597e-01 -1.44863415e+00
1.06485260e+00 3.83875281e-01 1.58466601e+00 4.36535552e-02
-3.58844995e-01 1.12269127e+00 -6.10563934e-01 -3.87759924e-01
-8.42889369e-01 -2.21751705e-01 -2.31061369e-01 -1.10807963e-01
2.26244062e-01 -5.30094922e-01 -4.68660772e-01 8.52163434e-02
-1.59367132e+00 8.72654080e-01 6.56858206e-01 7.31043398e-01
4.47263807e-01 -3.84146154e-01 9.24566269e-01 -6.35603726e-01
4.53920215e-01 -5.97566664e-01 -1.73177436e-01 3.71646225e-01
7.00494424e-02 2.25728527e-02 5.58041930e-01 -6.23685062e-01
-1.09220040e+00 3.70108932e-01 1.02283172e-01 -6.55531883e-01
-4.61999744e-01 4.50030863e-01 1.26307085e-01 3.43488634e-01
6.52079523e-01 7.14873254e-01 4.82163280e-02 -3.59285064e-02
5.86424112e-01 4.61064488e-01 1.03531528e+00 -7.56529331e-01
7.04802811e-01 6.43348396e-01 -1.76615372e-01 -5.22063076e-01
-9.16624129e-01 -2.60159016e-01 -5.50154030e-01 -4.19442266e-01
1.19910538e+00 -1.19126642e+00 -4.40856367e-01 -9.47915241e-02
-1.63505507e+00 -1.48742229e-01 -5.32777905e-01 2.22149223e-01
-1.16241229e+00 1.74094528e-01 -4.37406778e-01 -3.96512091e-01
-2.71616787e-01 -1.14161873e+00 1.37522483e+00 4.72734533e-02
-1.64190918e-01 -9.76110399e-01 -2.61805952e-01 5.17949104e-01
4.96186197e-01 4.36725080e-01 7.06955671e-01 -6.70884311e-01
-8.53224874e-01 -7.64103904e-02 -4.40673679e-01 5.64561844e-01
4.84253392e-02 -3.52158040e-01 -6.88985288e-01 -9.51752067e-02
-3.30517441e-01 -7.08571792e-01 5.59242666e-01 6.77841008e-02
1.30656326e+00 -5.24755836e-01 -2.49132067e-01 8.38862956e-02
1.61416817e+00 1.65753692e-01 7.76180625e-01 2.82494128e-01
5.42809606e-01 8.54742646e-01 2.91963279e-01 9.99911651e-02
6.11051500e-01 6.56825483e-01 5.56963682e-01 -2.21741155e-01
-5.03480077e-01 -4.51506853e-01 3.50533336e-01 4.58465844e-01
2.28125751e-01 -3.36673200e-01 -9.63933051e-01 1.13181758e+00
-1.90453720e+00 -1.28438580e+00 3.85293290e-02 1.40264893e+00
9.31363046e-01 -6.01342581e-02 -1.45113140e-01 -4.97139782e-01
8.00537109e-01 -3.07951719e-01 -4.40944880e-01 -2.78759778e-01
-3.42680156e-01 8.32108110e-02 1.37374848e-02 1.61679506e-01
-9.83219266e-01 9.27095890e-01 6.58877230e+00 3.55534524e-01
-7.63323307e-01 1.22761779e-01 6.54170036e-01 -3.16598982e-01
-3.81207019e-01 -5.52044697e-02 -5.23779929e-01 3.42385620e-01
9.56960261e-01 -2.61777908e-01 8.19139630e-02 7.23231494e-01
1.12425514e-01 3.98375057e-02 -1.58853912e+00 1.13462675e+00
7.97210932e-01 -1.90795112e+00 8.08986187e-01 -3.11778665e-01
1.03115320e+00 -1.72615066e-01 4.82956730e-02 4.87417370e-01
1.03592113e-01 -9.45959687e-01 1.07964242e+00 4.82330501e-01
6.53697014e-01 -2.81085163e-01 6.93561316e-01 1.22735173e-01
-7.24404395e-01 -6.73955828e-02 -2.72342056e-01 -1.20677322e-01
4.97703850e-01 3.80024454e-03 -1.17957008e+00 1.91365957e-01
6.41495168e-01 8.42179835e-01 -8.61379445e-01 8.93716097e-01
-1.08287960e-01 2.62828529e-01 7.20251203e-02 8.56889226e-03
5.95861077e-01 2.38051653e-01 3.07336837e-01 1.14441359e+00
2.37128288e-01 2.21296400e-01 -3.61964814e-02 1.34258783e+00
-3.02920461e-01 -1.09106936e-01 -8.22252214e-01 -2.59457082e-01
2.16969803e-01 9.48667765e-01 -5.10267317e-01 -9.14432585e-01
-5.88674247e-01 8.72940779e-01 1.64700538e-01 4.74338800e-01
-9.79327142e-01 5.65044731e-02 2.40410462e-01 3.29975605e-01
4.55676258e-01 -1.72613457e-01 4.29383963e-02 -1.15743041e+00
1.90486759e-01 -7.02825129e-01 -3.06416163e-03 -1.81268084e+00
-1.39162147e+00 7.97024250e-01 5.60153365e-01 -1.25817871e+00
-4.48725730e-01 -5.27116299e-01 -4.25386667e-01 4.29353714e-01
-1.24957955e+00 -1.76879883e+00 -4.44068760e-01 6.10751629e-01
9.22643960e-01 -1.25291362e-01 8.15166056e-01 3.32820743e-01
1.75602920e-02 1.83917135e-01 -1.78147301e-01 1.52120382e-01
5.40776074e-01 -9.74792600e-01 5.71889520e-01 6.29828811e-01
4.15743142e-01 3.41017783e-01 8.86763990e-01 -5.62779486e-01
-9.53883350e-01 -1.31254315e+00 8.42523396e-01 -6.40234351e-01
4.52444434e-01 -6.18170083e-01 -8.49897504e-01 9.83337104e-01
5.72072029e-01 1.11996196e-01 5.77188730e-01 -5.75294256e-01
-4.38861877e-01 1.04708374e-01 -9.25709248e-01 7.87274182e-01
1.16299117e+00 -5.64881504e-01 -9.13153112e-01 7.56312430e-01
1.03284848e+00 -7.21268281e-02 -6.66060984e-01 3.33540142e-01
2.97553539e-01 -6.65221095e-01 1.24736989e+00 -1.00998700e+00
1.29536021e+00 -2.60093123e-01 -3.51528049e-01 -1.03812587e+00
-4.53599319e-02 -7.94919580e-02 2.35771507e-01 1.38457334e+00
4.62342143e-01 3.31183965e-03 7.41219521e-01 9.13962007e-01
-5.21530092e-01 -3.09365302e-01 -5.62005341e-01 -9.16193902e-01
2.94222659e-03 -3.33722711e-01 5.29250324e-01 8.37765276e-01
-3.75304893e-02 4.80402619e-01 -2.46597067e-01 -1.71867684e-01
3.82861853e-01 1.21974811e-01 9.10527170e-01 -4.69698757e-01
-2.63440132e-01 -2.27922648e-01 -7.71798849e-01 -6.96296930e-01
5.76151431e-01 -1.11967480e+00 3.81837457e-01 -1.87035418e+00
7.61831105e-01 -9.36352685e-02 1.59594640e-01 6.42124474e-01
1.55832563e-02 7.50287056e-01 5.54717481e-01 3.08456719e-01
-1.08119106e+00 7.02907622e-01 1.44609880e+00 -5.72885275e-01
2.04812840e-01 -8.00660253e-01 -6.91997170e-01 3.12471479e-01
6.12731695e-01 -4.40041304e-01 -4.79600728e-01 -8.35332334e-01
3.11765615e-02 -4.51445989e-02 8.74374449e-01 -1.18854296e+00
5.92649281e-02 -1.24643862e-01 4.80159491e-01 -4.72953469e-01
5.87700129e-01 -8.80069196e-01 3.93200934e-01 3.35231543e-01
-9.48139787e-01 1.11145273e-01 5.85285544e-01 5.72439492e-01
-5.63729405e-01 -3.42154264e-01 4.86760408e-01 -4.98863429e-01
-1.04534233e+00 1.98719651e-01 -4.79118437e-01 -4.36915345e-02
1.31006074e+00 -1.85748205e-01 -6.17586672e-01 -7.97278345e-01
-8.86239648e-01 2.80201942e-01 5.49613535e-01 8.18384171e-01
1.00964296e+00 -1.67774034e+00 -1.03458524e+00 -1.36290163e-01
7.25664437e-01 -8.42435379e-03 3.71779054e-01 2.38034204e-01
-7.00981736e-01 5.59393704e-01 -5.46306491e-01 -6.78369880e-01
-1.09077537e+00 9.78842378e-01 2.12439790e-01 3.38478625e-01
-6.54804707e-01 5.35478711e-01 6.85579181e-01 -6.68627173e-02
-9.12081171e-03 -1.35911524e-01 -1.47563368e-02 -4.01745379e-01
6.23922527e-01 -5.44311523e-01 -3.93434227e-01 -1.00661039e+00
-1.57693885e-02 2.02657536e-01 -2.20841542e-01 -2.82620341e-01
1.41610503e+00 -2.52528995e-01 8.02787486e-03 1.78613260e-01
1.43611038e+00 -7.92945087e-01 -1.22165060e+00 -2.40033448e-01
-1.81135684e-01 -2.70875394e-01 -3.59839529e-01 -8.94550920e-01
-7.97180414e-01 7.46150196e-01 3.42925191e-01 -6.46846294e-02
1.07834303e+00 5.73364377e-01 9.10144091e-01 5.03177106e-01
1.41859815e-01 -8.02469373e-01 7.80346334e-01 3.38880658e-01
1.38860297e+00 -1.42284930e+00 -3.10885847e-01 -2.00141042e-01
-9.11588788e-01 9.65986550e-01 9.59513366e-01 -8.63361880e-02
1.72770932e-01 -1.64125502e-01 2.55130231e-03 -2.82076746e-01
-9.88857388e-01 -2.77703553e-01 1.17241912e-01 8.36768031e-01
1.59460098e-01 -1.19428754e-01 1.89106483e-02 5.54422498e-01
-1.18372947e-01 8.06891322e-02 9.58286703e-01 9.87814009e-01
-1.44067496e-01 -8.39434683e-01 -2.05219075e-01 2.34203801e-01
-2.00479701e-01 -2.59293348e-01 -4.23055887e-01 8.21750283e-01
1.83664739e-01 8.02219391e-01 5.96300960e-01 -7.04659447e-02
2.01760858e-01 -1.20503418e-01 6.53086364e-01 -9.88623500e-01
-4.97111738e-01 -4.36048299e-01 1.33449763e-01 -3.33777845e-01
-9.44461048e-01 -5.79425991e-01 -1.19734347e+00 2.15713397e-01
2.07981259e-01 1.22749679e-01 1.00150716e+00 8.98420453e-01
4.48652774e-01 5.33128858e-01 2.06122443e-01 -1.02868986e+00
-1.07518665e-01 -8.83201838e-01 -1.99912161e-01 1.21343851e+00
2.45419651e-01 -3.27395707e-01 -8.54518786e-02 1.04722989e+00]
|
[10.968169212341309, 1.0231789350509644]
|
0c352f1a-2ee0-4948-8d42-8099c38c76e0
|
modeling-diverse-chemical-reactions-for
|
2208.05482
| null |
https://arxiv.org/abs/2208.05482v1
|
https://arxiv.org/pdf/2208.05482v1.pdf
|
Modeling Diverse Chemical Reactions for Single-step Retrosynthesis via Discrete Latent Variables
|
Single-step retrosynthesis is the cornerstone of retrosynthesis planning, which is a crucial task for computer-aided drug discovery. The goal of single-step retrosynthesis is to identify the possible reactants that lead to the synthesis of the target product in one reaction. By representing organic molecules as canonical strings, existing sequence-based retrosynthetic methods treat the product-to-reactant retrosynthesis as a sequence-to-sequence translation problem. However, most of them struggle to identify diverse chemical reactions for a desired product due to the deterministic inference, which contradicts the fact that many compounds can be synthesized through various reaction types with different sets of reactants. In this work, we aim to increase reaction diversity and generate various reactants using discrete latent variables. We propose a novel sequence-based approach, namely RetroDVCAE, which incorporates conditional variational autoencoders into single-step retrosynthesis and associates discrete latent variables with the generation process. Specifically, RetroDVCAE uses the Gumbel-Softmax distribution to approximate the categorical distribution over potential reactions and generates multiple sets of reactants with the variational decoder. Experiments demonstrate that RetroDVCAE outperforms state-of-the-art baselines on both benchmark dataset and homemade dataset. Both quantitative and qualitative results show that RetroDVCAE can model the multi-modal distribution over reaction types and produce diverse reactant candidates.
|
['Feng Wu', 'Yunfei Liu', 'Jie Wang', 'Huarui He']
|
2022-08-10
| null | null | null | null |
['retrosynthesis']
|
['medical']
|
[ 4.28224027e-01 -1.43226549e-01 -4.49113101e-01 -1.73660778e-02
-6.85593247e-01 -1.13325763e+00 9.50309753e-01 1.00969598e-01
-2.68968850e-01 1.17876029e+00 3.67451698e-01 -5.27740240e-01
4.75072324e-01 -9.15749669e-01 -1.01422071e+00 -1.20624542e+00
6.04858279e-01 6.00844681e-01 -2.05412686e-01 -1.90759659e-01
2.68374413e-01 5.55121660e-01 -1.03036177e+00 4.92924184e-01
6.46761894e-01 5.56448877e-01 4.06910866e-01 6.49557948e-01
-2.69531518e-01 5.25231361e-01 -6.83374763e-01 -5.81462860e-01
9.04608518e-02 -6.61520183e-01 -5.67448318e-01 -3.88406724e-01
-2.37431452e-01 -3.91850509e-02 -1.86294004e-01 9.91497576e-01
6.32861435e-01 3.74135882e-01 1.35581660e+00 -9.09149766e-01
-7.50744820e-01 1.03567576e+00 -6.99001476e-02 -2.95130134e-01
3.15033942e-01 4.04551774e-01 1.17731535e+00 -1.11166692e+00
7.75714517e-01 1.34651339e+00 1.67487338e-01 7.68663049e-01
-1.41326427e+00 -7.97106743e-01 1.08415306e-01 -9.78482813e-02
-1.40568137e+00 -3.71198595e-01 6.22524381e-01 -5.31388581e-01
1.25410521e+00 1.40858829e-01 6.43422723e-01 1.70599341e+00
6.19276881e-01 6.82787359e-01 4.67015833e-01 9.56321955e-02
6.01159751e-01 -1.74271837e-01 -7.09855735e-01 6.22342229e-01
-2.12513715e-01 4.11541790e-01 -4.15736139e-01 -2.07676478e-02
7.23653197e-01 2.13958189e-01 -7.76993632e-02 1.24280572e-01
-1.48854220e+00 1.22012508e+00 4.45145756e-01 1.72018215e-01
-6.83170736e-01 1.55049831e-01 5.53752482e-01 -3.28023195e-01
1.71164915e-01 1.00689983e+00 -5.20816028e-01 2.43980631e-01
-1.05622542e+00 4.17922527e-01 7.49021471e-01 1.15763259e+00
4.38524246e-01 3.37940663e-01 -5.60790062e-01 3.86961222e-01
4.03995812e-01 1.81532934e-01 4.28313941e-01 -5.59402466e-01
3.01893204e-01 3.02636176e-01 -9.49187111e-03 -2.93067724e-01
-3.22498381e-01 -2.57233411e-01 -9.06559765e-01 -2.38164410e-01
2.05016539e-01 -3.20416301e-01 -1.34238756e+00 1.73911810e+00
4.66577053e-01 -2.29254048e-02 4.18789238e-01 6.91575825e-01
1.03614736e+00 1.64385760e+00 5.33743858e-01 -7.09204853e-01
1.04548979e+00 -1.12216425e+00 -7.51420259e-01 2.62946904e-01
4.65672851e-01 -8.70262444e-01 4.29037899e-01 6.05736196e-01
-1.19353926e+00 -3.25034618e-01 -1.11109471e+00 -2.80696988e-01
-7.10609317e-01 5.87905608e-02 6.11076534e-01 4.71901923e-01
-4.32445228e-01 6.97367311e-01 -6.26173675e-01 3.59296352e-01
2.02144548e-01 3.36016685e-01 -2.04570502e-01 1.78813368e-01
-1.36742294e+00 8.14216614e-01 1.11680591e+00 2.36489758e-01
-1.80879021e+00 -1.03853953e+00 -9.20139074e-01 1.88469350e-01
7.22864330e-01 -9.94502246e-01 1.36315334e+00 -5.90323806e-01
-1.95415866e+00 8.74481127e-02 -2.57462978e-01 -1.61262944e-01
4.38846320e-01 3.46510798e-01 -1.71200961e-01 -2.18008667e-01
5.05439006e-02 1.02771115e+00 7.33698726e-01 -1.07303357e+00
-3.04186732e-01 1.60447359e-01 -3.63386929e-01 1.50895849e-01
5.37896931e-01 -3.14366430e-01 -2.94906288e-01 -7.67626226e-01
-2.63520628e-01 -8.28208864e-01 -5.42759836e-01 -4.34094191e-01
-8.72226536e-01 -4.35529739e-01 1.45283952e-01 -4.16381627e-01
1.08844566e+00 -1.52367210e+00 8.97889256e-01 6.74464405e-02
1.93474498e-02 2.56035328e-01 -6.12676069e-02 7.63403416e-01
-4.08970028e-01 4.60557342e-01 -3.50939423e-01 -1.09733857e-01
1.25857785e-01 -5.62043265e-02 -3.43842208e-01 2.59884715e-01
2.67737925e-01 1.06151712e+00 -1.15977001e+00 -4.81722593e-01
3.65998775e-01 6.21216595e-01 -6.26164615e-01 2.24806845e-01
-1.20792961e+00 5.66154361e-01 -5.28477669e-01 8.46412957e-01
3.08928579e-01 -1.08653747e-01 5.03364682e-01 -4.14916128e-01
-3.70492607e-01 3.11792135e-01 -7.36483932e-01 1.92330885e+00
-3.43514413e-01 -5.20834811e-02 -7.23334074e-01 -7.20412910e-01
7.43257165e-01 7.68863916e-01 4.32923794e-01 -3.84571612e-01
3.41272503e-01 1.73758477e-01 -6.99668080e-02 -7.52122775e-02
7.09373415e-01 -8.18840563e-01 -2.54626900e-01 1.54535053e-02
2.92800248e-01 -3.65397245e-01 4.21891361e-01 1.02012902e-01
4.18738514e-01 5.31497896e-01 6.91414833e-01 7.99646080e-02
6.61493003e-01 -5.95886707e-02 6.66938424e-01 4.52608973e-01
1.56595558e-01 4.63327885e-01 6.03271246e-01 -3.59412342e-01
-1.48383880e+00 -1.26720214e+00 2.12109491e-01 7.16104150e-01
-2.54713714e-01 -4.04294550e-01 -6.85194314e-01 -6.77135885e-01
-4.37538296e-01 1.04559147e+00 -5.08388281e-01 -2.03418642e-01
-2.74009049e-01 -7.86761403e-01 5.63315570e-01 4.33323890e-01
-5.15885949e-02 -1.01803768e+00 2.08699822e-01 6.47519588e-01
-8.65411907e-02 -6.57849133e-01 -9.40308928e-01 4.14113611e-01
-3.80540848e-01 -8.41511130e-01 -1.08465075e+00 -8.25688243e-01
4.66558188e-01 -1.90191194e-01 8.10607493e-01 -5.44891894e-01
-3.73437017e-01 -3.93391728e-01 -1.58337932e-02 -5.08686185e-01
-1.06753552e+00 7.47808591e-02 -5.44566754e-03 -2.39733174e-01
-8.56646374e-02 -2.89955884e-01 -7.86361992e-01 -2.06119679e-02
-1.02377355e+00 1.90990582e-01 7.07078636e-01 9.16652381e-01
1.15246737e+00 2.02715188e-01 5.64450741e-01 -6.22085154e-01
5.04800200e-01 -7.82906830e-01 -8.29617620e-01 5.05129039e-01
-4.84922856e-01 7.91718304e-01 1.09194529e+00 -7.12405860e-01
-1.13323402e+00 4.69508678e-01 -3.62103552e-01 -4.50837046e-01
1.35815471e-01 6.75903440e-01 -5.81476152e-01 5.19763589e-01
5.96217215e-01 7.45835483e-01 -3.39984357e-01 2.74249837e-02
8.12126458e-01 2.27215350e-01 2.50759006e-01 -8.03761959e-01
2.57746279e-01 1.04049295e-02 4.31903243e-01 -7.26810217e-01
-4.72184688e-01 -3.07013094e-01 -2.04370499e-01 6.50022104e-02
1.24696732e+00 -9.38564599e-01 -1.29986155e+00 1.96769118e-01
-1.47990417e+00 -3.19181830e-01 1.15569659e-01 2.86727667e-01
-6.95131361e-01 2.44997919e-01 -5.43647468e-01 -5.65461874e-01
-6.39143705e-01 -1.95630193e+00 1.16538715e+00 6.64932579e-02
-2.60422736e-01 -7.89752305e-01 1.48164168e-01 -4.69641294e-03
-3.42600733e-01 4.58791673e-01 1.09005606e+00 -6.09800100e-01
-8.12079191e-01 2.94440091e-02 2.52647251e-01 5.04778512e-02
3.14168006e-01 2.72559673e-01 -5.45244694e-01 1.18693979e-02
-5.70918739e-01 -2.40106478e-01 9.80555296e-01 6.16810858e-01
1.09531343e+00 -4.94699538e-01 -4.09358591e-01 4.76321906e-01
1.32868350e+00 8.56205642e-01 5.74610651e-01 -1.45688370e-01
9.11198735e-01 4.09770906e-01 4.32530254e-01 4.28949416e-01
-3.15747298e-02 4.00872976e-01 4.68493760e-01 2.91122973e-01
2.44173810e-01 -7.72176087e-01 5.74167907e-01 2.98776060e-01
3.91820930e-02 -9.31334376e-01 -5.64271629e-01 2.13331789e-01
-1.56723154e+00 -1.10774398e+00 1.06377289e-01 2.04441619e+00
1.26326036e+00 -1.41108826e-01 3.00816089e-01 -1.25160620e-01
7.09394395e-01 2.06792071e-01 -7.02504158e-01 -5.28025925e-01
1.17464177e-01 5.45406818e-01 6.54360175e-01 5.52866042e-01
-8.76251996e-01 1.34303832e+00 6.11570930e+00 9.78017509e-01
-1.41737199e+00 -3.13242495e-01 7.15651691e-01 -2.24171206e-01
-4.33499783e-01 -1.83699772e-01 -1.06491816e+00 6.78076625e-01
1.14219725e+00 -1.69147089e-01 5.88513613e-01 6.18068099e-01
4.16644424e-01 4.07455295e-01 -1.46352124e+00 7.48674512e-01
-3.40772361e-01 -2.12337518e+00 5.53668201e-01 -1.16003066e-01
8.61123681e-01 -4.67199653e-01 2.02877939e-01 2.87357718e-01
5.28203785e-01 -1.14198327e+00 1.08613741e+00 2.85362363e-01
8.00725639e-01 -1.03475702e+00 1.50352478e-01 1.68532193e-01
-1.17968595e+00 1.13040417e-01 7.15261698e-03 6.37412846e-01
3.35541368e-01 2.02838793e-01 -1.48251069e+00 6.86393261e-01
-1.26768634e-01 6.55253828e-01 4.10549670e-01 5.16556203e-01
-5.51350415e-01 2.84668446e-01 -1.50865912e-01 -6.61148369e-01
5.46631277e-01 -3.96483928e-01 3.09788883e-01 1.14717603e+00
5.47968864e-01 1.20706938e-01 2.11950988e-01 1.49054468e+00
-4.31306154e-01 -7.78134121e-03 -4.24504071e-01 -8.32338810e-01
5.42197108e-01 7.91388690e-01 -6.84302390e-01 -4.19887632e-01
4.58754711e-02 1.03113103e+00 -1.81151912e-01 4.94216114e-01
-1.12350917e+00 -2.32886285e-01 5.27409554e-01 -4.94800717e-01
4.24096465e-01 -2.53574848e-01 -2.02450249e-03 -9.02844727e-01
-5.70287406e-01 -1.19561374e+00 3.07473570e-01 -7.30478525e-01
-8.40783358e-01 2.56701022e-01 -1.50767475e-01 -8.42809260e-01
-7.70381913e-02 -4.70963955e-01 -4.48322415e-01 1.02080965e+00
-9.98078585e-01 -1.07516301e+00 5.98751724e-01 2.57828832e-01
1.14791739e+00 -7.78272524e-02 8.23622465e-01 8.21431056e-02
-8.03900778e-01 4.59465563e-01 1.52961105e-01 -3.00159931e-01
6.99831903e-01 -1.30140603e+00 5.97268462e-01 7.29555666e-01
1.18183447e-02 9.05120134e-01 9.77518797e-01 -9.08806264e-01
-1.59219038e+00 -1.40191126e+00 7.11190462e-01 -1.88814789e-01
4.86051619e-01 -5.66636860e-01 -4.61646050e-01 4.99619663e-01
9.48287398e-02 -4.53224868e-01 7.17280149e-01 -4.21575189e-01
-3.12317431e-01 3.87153506e-01 -9.12603796e-01 1.04838526e+00
6.10034585e-01 -3.61135960e-01 -2.61281192e-01 4.72269177e-01
1.26139641e+00 -6.37922227e-01 -1.06334817e+00 -8.70782975e-03
5.24893641e-01 -3.00294727e-01 1.27088535e+00 -7.94680357e-01
1.01747727e+00 -4.32631880e-01 -2.13529646e-01 -1.49492657e+00
-2.67623127e-01 -8.99063289e-01 -4.80381213e-02 8.62399817e-01
9.48633254e-01 -3.15268219e-01 7.82565117e-01 2.90579408e-01
-4.30872828e-01 -6.88737869e-01 -6.18299127e-01 -6.35359406e-01
4.07883167e-01 -1.72258437e-01 8.83506954e-01 6.38621867e-01
-1.64609879e-01 6.17452323e-01 -4.91282761e-01 1.57624349e-01
2.95165777e-01 -1.17813058e-01 3.56298804e-01 -6.63565278e-01
-3.71827543e-01 -6.05296135e-01 2.59365976e-01 -1.02712238e+00
2.66694546e-01 -1.16302383e+00 4.76105273e-01 -1.37243533e+00
6.85785860e-02 -1.16553167e-02 -2.04269245e-01 2.20887780e-01
-2.78638303e-01 -4.29006487e-01 -4.85693850e-02 -2.27672905e-01
-2.25535944e-01 9.38055515e-01 1.42491853e+00 -6.80826485e-01
-4.99085277e-01 -1.17083304e-01 -8.08954298e-01 5.38390353e-02
5.09727001e-01 -6.17627144e-01 -5.81277132e-01 2.33459070e-01
6.29562616e-01 6.75721288e-01 -1.22937009e-01 -3.75980407e-01
6.28152192e-02 -6.49284005e-01 3.90844494e-01 -8.94885957e-01
4.49999779e-01 -4.69931453e-01 5.80734074e-01 6.08120859e-01
-6.76217794e-01 -2.97723621e-01 1.14232369e-01 8.75705838e-01
1.27071619e-01 -8.87472555e-02 5.41004717e-01 -5.26905715e-01
-3.18888158e-01 8.13342869e-01 -8.61373723e-01 -4.43990558e-01
1.03414655e+00 1.08493112e-01 -4.98664640e-02 2.11005230e-02
-7.07107186e-01 1.35820657e-01 1.40475467e-01 3.36497009e-01
6.10910058e-01 -1.14619946e+00 -5.60287297e-01 -2.45766774e-01
6.66137934e-02 4.92475033e-01 1.71962276e-01 2.50594497e-01
-6.43825948e-01 7.78732598e-01 1.92351252e-01 -3.95731837e-01
-1.14760733e+00 1.20697367e+00 3.81184578e-01 -1.76331967e-01
9.56437439e-02 1.07148421e+00 4.79500681e-01 -3.55413288e-01
1.05399348e-01 -5.31308830e-01 6.44321218e-02 2.32154682e-01
3.59059185e-01 8.60872492e-02 -6.81945262e-03 -3.44520777e-01
-1.39674783e-01 1.00139603e-01 -3.37826461e-01 -3.52950655e-02
1.10762846e+00 5.20053685e-01 1.66747853e-01 1.24365941e-01
1.35073233e+00 -4.86753196e-01 -1.21110010e+00 4.52060886e-02
-4.29372907e-01 2.55816877e-01 2.35295355e-01 -9.16555107e-01
-7.47223973e-01 8.50871623e-01 7.29307383e-02 -4.32145923e-01
6.80417061e-01 -1.97321996e-01 9.18633044e-01 5.03318131e-01
-9.49797034e-02 -7.72370219e-01 1.25761494e-01 2.42728695e-01
9.62452888e-01 -1.04565406e+00 1.17143400e-01 -2.65156120e-01
-7.41785944e-01 1.25944757e+00 1.41539574e-01 4.41470385e-01
1.56691328e-01 -3.42171714e-02 -4.34446812e-01 -9.66865011e-03
-7.65960574e-01 1.97855592e-01 2.00954795e-01 2.26509914e-01
6.11152947e-01 3.17163229e-01 -3.37786227e-01 5.27036905e-01
-3.97188179e-02 -3.10173109e-02 2.83941090e-01 8.30266356e-01
-1.07692115e-01 -1.48482275e+00 -2.11255550e-01 -2.42876112e-01
-5.24950325e-01 -5.03289461e-01 -6.96057022e-01 3.99685562e-01
1.79645479e-01 8.25889051e-01 -3.87044340e-01 -2.22751305e-01
-1.59987547e-02 2.59447485e-01 7.04948723e-01 -7.00157344e-01
-6.92745507e-01 4.49399263e-01 7.64596909e-02 -2.97029227e-01
-1.48403838e-01 -6.37791455e-01 -1.49553549e+00 -2.56543785e-01
-3.25106561e-01 3.83633733e-01 8.75338912e-01 9.01514471e-01
2.54999548e-01 8.93640995e-01 7.06343293e-01 -8.54976356e-01
-4.58878160e-01 -5.73090136e-01 -1.10380538e-01 -1.13546878e-01
3.54056805e-01 -4.35954869e-01 1.03587247e-01 2.07885608e-01]
|
[4.5003557205200195, 6.105567932128906]
|
8cbc3ba7-5e65-4115-b0eb-83325406ffe1
|
emrel-joint-representation-of-entities-and
| null | null |
https://openreview.net/forum?id=2csU2MGpRbN
|
https://openreview.net/pdf?id=2csU2MGpRbN
|
EmRel: Joint Representation of Entities and Embedded Relations for Multi-triple Extraction
|
Multi-triple extraction is a challenging task due to the existence of informative inter-triple correlations and consequently rich interactions across the constituent entities and relations.
While existing works only explore cross-entity interactions, we propose to explicitly introduce relation representation, jointly represent it with entities, and novelly align them to identify valid triples.
We perform comprehensive experiments on document-level relation extraction and joint entity and relation extraction along with detailed ablations to demonstrate the advantage of the proposed method.
|
['Anonymous']
|
2021-11-16
| null | null | null |
acl-arr-november-2021-11
|
['document-level-relation-extraction', 'joint-entity-and-relation-extraction']
|
['natural-language-processing', 'natural-language-processing']
|
[ 9.37710330e-02 5.87038338e-01 -6.85333073e-01 -2.82637239e-01
-8.34806442e-01 -7.02464759e-01 6.71643376e-01 6.46376789e-01
-1.08679961e-02 1.07481468e+00 4.39854383e-01 -2.35170007e-01
-3.76167297e-01 -8.47387195e-01 -5.93494833e-01 -3.41420411e-04
-5.36191285e-01 5.80613375e-01 2.67496198e-01 -1.67277634e-01
-2.56793082e-01 2.20384985e-01 -1.25755894e+00 3.76431227e-01
1.13271487e+00 6.27638459e-01 -7.03576684e-01 1.31674781e-01
-1.51315123e-01 8.58866990e-01 -5.23500383e-01 -9.35808182e-01
6.06692173e-02 -1.18143380e-01 -1.12619472e+00 -5.89275248e-02
3.68558243e-02 1.67764008e-01 -2.23782167e-01 9.05490577e-01
1.73365757e-01 -1.64619505e-01 6.89856350e-01 -1.29543567e+00
-4.88054723e-01 1.22137022e+00 -8.31428289e-01 1.77070662e-01
7.00714588e-01 -4.37936395e-01 1.72965312e+00 -1.22456837e+00
9.46297109e-01 1.00623631e+00 7.83298314e-01 -9.41260010e-02
-1.07979786e+00 -6.08477414e-01 1.81821004e-01 3.41423869e-01
-1.67648876e+00 -5.68161309e-01 4.82940704e-01 -2.31437966e-01
1.54380608e+00 3.62765461e-01 5.36922038e-01 8.37013960e-01
-2.78291970e-01 7.27955520e-01 6.39764011e-01 -5.59856236e-01
-3.89437944e-01 -1.84070542e-02 6.08723640e-01 7.11642146e-01
9.48438287e-01 -2.10175484e-01 -7.45989919e-01 -2.90521324e-01
5.37127107e-02 -5.25078118e-01 -1.50428504e-01 -2.66732126e-01
-1.19451916e+00 2.88678139e-01 4.91245478e-01 4.64697331e-01
-4.81513888e-01 1.87038518e-02 4.82121915e-01 -5.65427244e-02
4.06743348e-01 5.95544815e-01 -7.27277100e-01 8.39390419e-03
-6.02327824e-01 3.93881708e-01 9.19903219e-01 1.45421267e+00
7.87264526e-01 -4.28717524e-01 -3.63433480e-01 7.94201732e-01
1.73135132e-01 1.13077603e-01 -6.34529591e-02 -3.54059398e-01
8.57997000e-01 1.20794761e+00 3.63342799e-02 -1.24094498e+00
-5.67319751e-01 -3.23909491e-01 -6.91648126e-01 -7.11807251e-01
-6.25170320e-02 -2.79223382e-01 -7.39449084e-01 1.46848094e+00
7.46365190e-01 2.35224105e-02 2.53056169e-01 2.90916920e-01
1.38375664e+00 6.66524842e-02 3.41812789e-01 -4.99560624e-01
1.59614873e+00 -8.75635386e-01 -1.25562346e+00 -2.98775375e-01
1.04550552e+00 -4.52620536e-01 2.80930996e-01 -3.60476524e-01
-1.00013816e+00 8.33838806e-02 -1.07371652e+00 -2.65171498e-01
-6.39996767e-01 2.13077217e-01 1.20037115e+00 3.35176319e-01
-3.81009847e-01 3.92462671e-01 -7.20053852e-01 -1.95626631e-01
4.64973539e-01 5.74288726e-01 -7.95185208e-01 2.25078359e-01
-1.63423848e+00 1.15911603e+00 9.61670220e-01 4.80192363e-01
1.63285524e-01 -8.70339394e-01 -1.26116240e+00 1.37229726e-01
9.02972400e-01 -8.91785622e-01 8.39978993e-01 1.03531359e-02
-7.19106019e-01 7.23324478e-01 -6.02748752e-01 -3.95875871e-01
-1.12350598e-01 -3.85948211e-01 -8.59537005e-01 -1.85747281e-01
1.89017564e-01 1.72217578e-01 -1.26244411e-01 -1.19723916e+00
-7.23652482e-01 -3.02825481e-01 1.34631515e-01 2.09915549e-01
-4.63659823e-01 2.87080675e-01 -7.36741781e-01 -5.50335169e-01
3.84027034e-01 -7.48946726e-01 -1.77887045e-02 -9.72810268e-01
-1.24787498e+00 -5.76374710e-01 6.37287438e-01 -5.88934183e-01
1.88646781e+00 -1.66920424e+00 1.33438393e-01 5.15334785e-01
6.07019484e-01 2.01847881e-01 1.36399835e-01 7.27680385e-01
-4.48089033e-01 5.45833051e-01 -2.06997082e-01 -1.83558822e-01
8.92107934e-02 2.40443662e-01 -2.13312373e-01 9.03555900e-02
5.93329847e-01 1.43452978e+00 -8.28542829e-01 -8.68798673e-01
-2.84526706e-01 3.52736294e-01 -1.76633239e-01 -1.83232084e-01
-1.15940414e-01 -5.45977950e-02 -6.56190515e-01 9.39266503e-01
6.44909024e-01 -4.63536948e-01 8.28400970e-01 -9.72453892e-01
1.52563855e-01 7.13558733e-01 -1.18544900e+00 1.04684913e+00
-1.92568481e-01 2.39105180e-01 -3.68416637e-01 -6.31724775e-01
5.87314963e-01 3.60354006e-01 7.65076458e-01 -4.23293740e-01
-9.32534859e-02 2.82671243e-01 3.38521749e-02 -6.60034180e-01
7.96036720e-01 2.07936659e-01 -1.92084059e-01 1.79911301e-01
1.91412106e-01 4.00250167e-01 7.83809602e-01 6.37340486e-01
1.34660292e+00 1.80694342e-01 1.00696552e+00 5.22389561e-02
4.94710028e-01 -2.26466209e-02 7.51815081e-01 2.29287475e-01
3.73725176e-01 2.20434114e-01 7.10927546e-01 -2.22872928e-01
-5.98505259e-01 -7.72583961e-01 -1.44750103e-01 5.35773277e-01
1.94143519e-01 -1.37632239e+00 -6.24824055e-02 -1.24305022e+00
1.45137057e-01 4.61023480e-01 -7.95041382e-01 1.78838186e-02
-7.52233982e-01 -1.03275275e+00 5.89962065e-01 6.47420406e-01
3.43515098e-01 -6.41138196e-01 1.28105178e-01 3.51887345e-02
-5.64495027e-01 -1.88789546e+00 -6.78957254e-02 1.97894916e-01
-4.64742392e-01 -1.47325659e+00 2.45112643e-01 -5.09037614e-01
7.05492675e-01 -1.86624721e-01 1.57184243e+00 1.88299462e-01
-9.23027471e-02 -6.59278482e-02 -4.02551979e-01 -7.95499980e-02
-4.82116938e-02 5.00348568e-01 -9.62423310e-02 -3.83959949e-01
6.60608053e-01 -6.43304229e-01 -8.21385309e-02 3.55470404e-02
-5.90671599e-01 8.30347165e-02 6.57579064e-01 6.46787703e-01
6.36245966e-01 2.82543570e-01 4.48514283e-01 -1.75654888e+00
6.38545334e-01 -4.69817072e-01 -2.22651109e-01 9.36943471e-01
-7.73970425e-01 1.63006708e-01 -2.08193976e-02 1.89165115e-01
-9.86763179e-01 4.41963188e-02 1.52706191e-01 2.80152559e-01
1.67636693e-01 1.14837456e+00 -4.82878327e-01 1.60278708e-01
3.74249071e-01 -4.39236760e-01 -8.79609048e-01 -2.01818004e-01
7.15835392e-01 3.71415645e-01 4.58458841e-01 -7.85259068e-01
1.06378698e+00 1.41315833e-01 2.55661577e-01 -4.63527143e-01
-1.14490628e+00 -4.53403801e-01 -9.48555410e-01 2.73097694e-01
5.11692882e-01 -1.09275389e+00 -8.72050226e-01 -1.38246432e-01
-1.33346605e+00 4.42473561e-01 -2.21799716e-01 3.11296910e-01
2.49529272e-01 3.90357733e-01 -5.70873916e-01 -6.58248007e-01
-4.77122188e-01 -7.65845537e-01 9.70497847e-01 8.54726359e-02
-5.31544566e-01 -9.48015392e-01 2.54156440e-01 2.92284191e-01
-1.52213141e-01 5.21908820e-01 9.28108990e-01 -1.01660419e+00
-5.73682785e-01 -4.99327481e-01 -5.82285166e-01 -4.51181203e-01
6.26247466e-01 4.15865034e-01 -5.60082734e-01 2.40214095e-01
-9.06516373e-01 -2.29682013e-01 7.51614869e-01 -1.80296734e-01
6.03113592e-01 -4.13022906e-01 -9.31253374e-01 2.21249744e-01
1.11966074e+00 -3.50452542e-01 4.64210778e-01 2.70339161e-01
1.13017178e+00 7.55031466e-01 6.62413299e-01 1.86580092e-01
1.04131055e+00 9.14247394e-01 9.55560952e-02 -1.24090917e-01
-4.79761809e-02 -2.30148152e-01 5.00164218e-02 9.90115464e-01
-3.15504968e-01 -2.69152254e-01 -9.23355877e-01 6.35678291e-01
-2.00809646e+00 -8.13150883e-01 -7.50361741e-01 1.63323367e+00
1.33448005e+00 1.35945186e-01 1.56241387e-01 1.38849750e-01
5.79736471e-01 4.49183099e-02 1.34565933e-02 -6.71092793e-02
-4.95018840e-01 3.63468856e-01 4.50412065e-01 4.91646469e-01
-1.35879159e+00 1.14422846e+00 7.06649256e+00 7.49242663e-01
-3.27912211e-01 -5.81586026e-02 6.02393821e-02 1.67017058e-01
-6.73660755e-01 4.79216486e-01 -1.09134102e+00 -8.89353827e-02
5.38459837e-01 -3.75655890e-01 -9.33586508e-02 3.65821123e-01
-4.27022070e-01 -1.50801748e-01 -1.24093628e+00 6.03479445e-01
-2.25847483e-01 -1.33493006e+00 1.14618361e-01 6.23242073e-02
8.62387836e-01 -3.05654228e-01 -4.19823945e-01 3.78225863e-01
6.53311968e-01 -9.70361829e-01 2.21158445e-01 3.00790370e-01
5.71592927e-01 -6.60458922e-01 7.77128816e-01 -2.12686196e-01
-1.77858341e+00 2.40447268e-01 1.84931830e-01 1.88057989e-01
2.44909182e-01 1.20355380e+00 -8.67436647e-01 1.50492597e+00
4.25273538e-01 1.01631296e+00 -8.51666927e-01 7.13376820e-01
-5.69479704e-01 3.93533409e-01 -5.80436707e-01 2.34548792e-01
-2.63534367e-01 -7.17447624e-02 4.79005635e-01 1.48817909e+00
5.31585775e-02 2.68257797e-01 9.93279517e-02 7.30373561e-01
-4.20750588e-01 2.60270566e-01 -5.81347525e-01 -4.65443313e-01
7.24914372e-01 1.41114664e+00 -5.25731444e-01 -3.81689459e-01
-6.78351462e-01 5.34933627e-01 8.69580090e-01 4.25180286e-01
-7.29729593e-01 -4.69224751e-01 4.95952964e-01 -2.18454823e-01
3.99106324e-01 -2.89992243e-01 -4.49433267e-01 -1.40516698e+00
5.80246210e-01 -6.13535285e-01 8.42724264e-01 -3.37412864e-01
-1.19592738e+00 6.46913886e-01 2.16687813e-01 -9.92312610e-01
-3.62601757e-01 -1.16623268e-01 -2.18622863e-01 6.52710736e-01
-1.48679841e+00 -1.72691500e+00 1.01505872e-02 2.73079723e-01
-3.06980729e-01 2.72784173e-01 9.74892378e-01 6.84188843e-01
-9.97646153e-01 8.99926305e-01 -5.28573990e-01 4.61250037e-01
5.54362535e-01 -1.36696494e+00 5.53353846e-01 9.47840154e-01
5.12374222e-01 1.21988428e+00 6.04972064e-01 -1.02368271e+00
-1.21140254e+00 -9.93448019e-01 1.59316659e+00 -6.97712839e-01
9.59860861e-01 -4.83768255e-01 -8.86374116e-01 1.16811085e+00
4.11611259e-01 4.26296443e-01 9.75229204e-01 9.84704971e-01
-6.84136808e-01 2.96667498e-02 -8.04590464e-01 7.22513616e-01
1.38644004e+00 -5.37035465e-01 -4.97901082e-01 3.33744496e-01
7.69524038e-01 -5.29051125e-01 -1.24999368e+00 1.21511292e+00
4.64291066e-01 -6.66320384e-01 1.05863988e+00 -8.52052867e-01
3.93609285e-01 -4.90440696e-01 1.05186060e-01 -9.37826395e-01
-1.33784324e-01 -5.44385254e-01 -1.00549006e+00 1.91938150e+00
1.16752291e+00 -3.50063413e-01 5.95215619e-01 8.56543064e-01
8.46012011e-02 -9.15147483e-01 -7.18136489e-01 -6.05348706e-01
-4.12169814e-01 -1.84283882e-01 6.80288315e-01 1.23277104e+00
6.95109487e-01 8.36969495e-01 -4.03233290e-01 5.60950160e-01
4.86182362e-01 6.45133317e-01 7.64965355e-01 -1.06200492e+00
-2.53355145e-01 -1.61953241e-01 -2.29779735e-01 -6.72752202e-01
2.78604031e-01 -9.57131565e-01 -1.08263761e-01 -1.68008435e+00
3.80250633e-01 -5.82681537e-01 -9.11473632e-02 7.26739526e-01
-7.13252187e-01 1.62949175e-01 -2.78288931e-01 9.89985019e-02
-1.01376271e+00 6.23307467e-01 8.81682038e-01 -7.52395466e-02
-1.09679133e-01 -2.17161804e-01 -9.03469980e-01 4.09625232e-01
1.68064579e-01 -4.47401226e-01 -2.54730761e-01 -2.25981295e-01
8.49023938e-01 -1.34882078e-01 -2.31246697e-03 -4.92984086e-01
3.58662307e-01 -9.72090885e-02 1.11397110e-01 -6.71265185e-01
1.91385165e-01 -8.28533947e-01 4.26571429e-01 -1.81355923e-01
-2.38305986e-01 -1.12179890e-01 1.86276108e-01 4.53785896e-01
-6.01653993e-01 9.66130942e-02 -9.75089222e-02 2.82392293e-01
-5.22233367e-01 2.71091610e-01 4.14673418e-01 2.16788501e-01
1.08131766e+00 7.78405070e-02 -5.89440048e-01 6.56633899e-02
-9.83665884e-01 4.66454268e-01 8.39037821e-02 3.48198503e-01
3.56785476e-01 -1.66233492e+00 -5.92437863e-01 -1.92863703e-01
4.68513340e-01 6.38472363e-02 -2.95778718e-02 1.08511138e+00
1.05896227e-01 3.72059017e-01 3.76727164e-01 -1.02146752e-01
-1.54714096e+00 5.03633380e-01 8.86731520e-02 -9.47061777e-01
-4.32243794e-01 7.88958848e-01 -1.44702062e-01 -4.70824063e-01
-1.80027694e-01 -2.19265059e-01 -5.42445481e-01 1.90913156e-01
2.19711810e-01 9.97760594e-02 3.26814979e-01 -7.31661499e-01
-8.86839867e-01 4.15197998e-01 -2.72384226e-01 2.71810681e-01
1.22035778e+00 -1.59558937e-01 -6.12119734e-01 2.75436819e-01
8.74787033e-01 5.36673963e-01 -2.33084947e-01 -4.69823182e-01
7.84933448e-01 -3.53451371e-01 -4.43407238e-01 -7.71981001e-01
-8.80261362e-01 2.26555228e-01 -4.80670959e-01 3.32441032e-01
7.88779497e-01 2.67654628e-01 7.53531635e-01 4.31383967e-01
1.65066123e-01 -6.49980664e-01 -5.08204579e-01 4.94623572e-01
7.51414597e-01 -1.28288853e+00 7.58606255e-01 -1.52697623e+00
-5.03383219e-01 8.06157887e-01 7.80944586e-01 3.36386800e-01
7.56083429e-01 6.15137100e-01 -3.85337442e-01 -5.81732094e-01
-8.87893379e-01 -7.24327326e-01 7.15295076e-01 4.13610280e-01
7.88376033e-01 -4.43195850e-02 -6.22834921e-01 8.19659173e-01
-3.41968574e-02 -9.77491662e-02 1.36661917e-01 9.32470620e-01
2.74197221e-01 -1.66941273e+00 2.18458638e-01 4.83077288e-01
-4.56902921e-01 -4.25135404e-01 -8.45162153e-01 8.60680521e-01
1.23664528e-01 8.50555241e-01 -4.21660572e-01 -7.09441185e-01
6.78035796e-01 -2.73507070e-02 5.63316405e-01 -7.71378875e-01
-7.42345333e-01 -1.28623664e-01 1.03312159e+00 -4.60808843e-01
-7.98136830e-01 -6.89887583e-01 -1.51993012e+00 -1.99799851e-01
-7.55982578e-01 3.47342551e-01 1.81712002e-01 1.36698544e+00
7.65234351e-01 7.27773726e-01 2.72472978e-01 -5.49909994e-02
9.78497788e-02 -1.07969224e+00 -2.66713321e-01 4.38891172e-01
3.86407860e-02 -9.61433351e-01 -5.95717691e-04 -1.34516761e-01]
|
[9.219468116760254, 8.580756187438965]
|
90cfc217-0a98-4cc8-a08a-fce7ec25bfb1
|
yolo-z-improving-small-object-detection-in
|
2112.11798
| null |
https://arxiv.org/abs/2112.11798v4
|
https://arxiv.org/pdf/2112.11798v4.pdf
|
YOLO-Z: Improving small object detection in YOLOv5 for autonomous vehicles
|
As autonomous vehicles and autonomous racing rise in popularity, so does the need for faster and more accurate detectors. While our naked eyes are able to extract contextual information almost instantly, even from far away, image resolution and computational resources limitations make detecting smaller objects (that is, objects that occupy a small pixel area in the input image) a genuinely challenging task for machines and a wide-open research field. This study explores how the popular YOLOv5 object detector can be modified to improve its performance in detecting smaller objects, with a particular application in autonomous racing. To achieve this, we investigate how replacing certain structural elements of the model (as well as their connections and other parameters) can affect performance and inference time. In doing so, we propose a series of models at different scales, which we name `YOLO-Z', and which display an improvement of up to 6.9% in mAP when detecting smaller objects at 50% IOU, at the cost of just a 3ms increase in inference time compared to the original YOLOv5. Our objective is to inform future research on the potential of adjusting a popular detector such as YOLOv5 to address specific tasks and provide insights on how specific changes can impact small object detection. Such findings, applied to the broader context of autonomous vehicles, could increase the amount of contextual information available to such systems.
|
['Andrew Bradley', 'Fabio Cuzzolin', 'Izzeddin Teeti', 'Aduen Benjumea']
|
2021-12-22
| null | null | null | null |
['small-object-detection']
|
['computer-vision']
|
[ 1.95554346e-01 3.06194555e-02 1.85333472e-02 -1.66888788e-01
-2.01024994e-01 -6.79589868e-01 3.91346931e-01 1.65119380e-01
-6.85192108e-01 3.37769091e-01 -3.65683645e-01 -3.66368204e-01
2.16381341e-01 -9.57936227e-01 -8.63268197e-01 -3.84657502e-01
9.17749628e-02 1.56143457e-01 1.01870334e+00 -2.44402498e-01
3.12662065e-01 7.54671812e-01 -2.04412699e+00 1.31469473e-01
7.48582840e-01 8.46361995e-01 4.14816260e-01 7.54722238e-01
3.32696050e-01 5.22981584e-01 -7.33564734e-01 -7.53851235e-02
2.99388349e-01 1.65084228e-01 -4.68278795e-01 3.25655937e-02
8.46175194e-01 -6.42573714e-01 -3.68263394e-01 8.01870644e-01
2.88020045e-01 1.03269629e-01 3.18380862e-01 -1.32650852e+00
-3.17918509e-01 2.38599926e-01 -5.21763563e-01 6.88842177e-01
-5.09809889e-02 7.10288465e-01 7.22355604e-01 -7.02127755e-01
5.11863947e-01 1.26828980e+00 6.97806239e-01 2.40845338e-01
-1.25333834e+00 -6.29385412e-01 2.20991403e-01 4.85707849e-01
-1.58045304e+00 -6.81035578e-01 2.40955070e-01 -2.94089496e-01
1.01879287e+00 2.28720024e-01 2.86213458e-01 6.46641910e-01
2.90181994e-01 5.94932079e-01 1.04454935e+00 -1.19262747e-01
2.21271232e-01 3.49161208e-01 -4.49673049e-02 6.81858242e-01
6.96842134e-01 -1.11634076e-01 -5.11474311e-01 2.61401653e-01
6.71112061e-01 -2.28657946e-01 2.67030627e-01 -4.86217618e-01
-1.25890303e+00 8.14731658e-01 5.29274762e-01 1.62025560e-02
-1.26910761e-01 4.69903290e-01 2.29850575e-01 1.45423979e-01
1.92293718e-01 7.28596210e-01 -2.42006436e-01 -6.16359226e-02
-6.38030887e-01 4.56872106e-01 3.82372707e-01 9.55955982e-01
9.76620972e-01 6.93695918e-02 -1.00692146e-01 4.71014977e-01
5.32069728e-02 8.88206959e-01 -1.06789201e-01 -1.33398461e+00
4.17472750e-01 6.86536968e-01 2.36652583e-01 -1.10847354e+00
-6.63048565e-01 -3.51862788e-01 -2.31843777e-02 5.26952088e-01
6.33046925e-01 -2.61485338e-01 -6.69151187e-01 1.50309753e+00
3.17376316e-01 4.43151779e-02 -1.90459237e-01 1.01933253e+00
5.63876092e-01 6.38278782e-01 1.32391527e-01 3.68066996e-01
1.60219836e+00 -8.48805249e-01 -1.93112358e-01 -8.03212225e-01
8.46776426e-01 -7.16958702e-01 1.10301983e+00 2.05131799e-01
-6.99360073e-01 -6.22636199e-01 -1.25636840e+00 -2.13723049e-01
-5.31443596e-01 2.65924513e-01 5.98075449e-01 7.40151823e-01
-9.74265873e-01 1.96878299e-01 -8.24056625e-01 -6.65489972e-01
4.51840341e-01 3.86188149e-01 -1.30538002e-01 -3.69959101e-02
-8.96490276e-01 1.25188661e+00 2.89836168e-01 -6.60541207e-02
-6.64352059e-01 -4.48475778e-01 -6.65914476e-01 1.41525730e-01
9.66028869e-01 -3.96494567e-01 1.10586035e+00 -7.04515457e-01
-1.07968235e+00 6.74409509e-01 -1.62495852e-01 -7.99117565e-01
3.80688548e-01 -2.46829867e-01 -2.49583274e-01 2.12386027e-01
2.97552019e-01 1.18815386e+00 6.92515850e-01 -1.05773401e+00
-1.03773808e+00 -5.21815956e-01 3.99216622e-01 2.67312437e-01
-1.21163405e-01 -1.56337664e-01 -6.30695939e-01 -1.54530525e-01
-1.32088885e-01 -1.35895348e+00 -2.62454093e-01 3.33966047e-01
1.11404821e-01 -2.21061721e-01 1.16350663e+00 -2.55250961e-01
9.27159607e-01 -2.16701984e+00 -3.30501348e-01 1.56315733e-02
3.48141730e-01 5.22466242e-01 -1.95198253e-01 3.06062073e-01
6.05843365e-01 -3.28377858e-02 -9.24468338e-02 1.00474261e-01
-2.48510942e-01 2.55082369e-01 -1.44417003e-01 5.25502861e-01
4.34262812e-01 9.16905046e-01 -9.36052203e-01 -4.79895413e-01
6.00213528e-01 2.84439981e-01 -6.95338368e-01 -3.28730434e-01
-2.75407374e-01 2.94187248e-01 -4.12472516e-01 6.13949358e-01
5.19567132e-01 -2.44466454e-01 -3.70105729e-02 -1.16942916e-02
-4.01187986e-01 3.46210264e-02 -1.13027740e+00 1.01715291e+00
-4.18289185e-01 1.46795690e+00 3.56005579e-02 -8.23010266e-01
8.65305603e-01 -2.38231614e-01 1.93475202e-01 -9.14036393e-01
9.79463458e-02 -4.83096484e-03 3.89348894e-01 -5.33605337e-01
1.03101957e+00 3.82914126e-01 -3.80033553e-02 1.19973741e-01
-3.86344522e-01 -1.12606093e-01 3.63305420e-01 2.04137251e-01
1.20720172e+00 -9.09744427e-02 -3.74795906e-02 -2.72344768e-01
1.68341592e-01 4.27626461e-01 2.72537827e-01 1.10679221e+00
-5.15382528e-01 2.31583908e-01 4.04680252e-01 -3.02582502e-01
-1.10016310e+00 -1.00528586e+00 -2.98424959e-01 1.27265513e+00
6.67618096e-01 -2.59644657e-01 -8.09850812e-01 -3.78036022e-01
3.52447957e-01 8.19632351e-01 -4.22994137e-01 -2.58787572e-01
-6.25719428e-01 -7.08661020e-01 6.19535387e-01 6.99560285e-01
6.92902148e-01 -9.06063795e-01 -1.48411489e+00 4.73409146e-02
6.86359033e-02 -1.49597394e+00 3.13951783e-02 2.51770820e-02
-8.47459793e-01 -9.54673588e-01 -2.54476100e-01 -4.95028824e-01
5.47388911e-01 8.19734871e-01 8.95916224e-01 6.22237995e-02
-4.07689780e-01 3.86899590e-01 -3.46318215e-01 -6.53069973e-01
-2.30532438e-01 2.84175217e-01 7.98462033e-02 -1.61486521e-01
6.28608108e-01 -1.61851272e-01 -6.62168741e-01 6.03977680e-01
-8.03425193e-01 2.41457466e-02 8.32030654e-01 2.62168080e-01
2.02133209e-01 -8.67630467e-02 5.04939198e-01 -8.39457035e-01
2.85144329e-01 -3.42546701e-01 -9.34844971e-01 -1.83112130e-01
-7.77894258e-01 1.27343908e-01 5.12292087e-01 -3.30156773e-01
-7.65817165e-01 4.18967307e-02 2.02616394e-01 -1.83707580e-01
-6.86782673e-02 4.14677076e-02 2.39041373e-01 -2.65467942e-01
9.32968438e-01 -7.87176564e-02 4.99595590e-02 -1.70352653e-01
5.44129908e-01 6.49294555e-01 4.73236531e-01 -2.30567336e-01
7.99264252e-01 7.81253517e-01 8.44626129e-02 -1.07060862e+00
-4.17165101e-01 -4.57040846e-01 -5.70110023e-01 -4.12357211e-01
8.75338674e-01 -9.17530060e-01 -9.26244974e-01 2.54224986e-01
-9.42401350e-01 -4.83124822e-01 -9.82058868e-02 3.98660600e-01
-2.79512376e-01 1.47370622e-01 -3.27700943e-01 -6.02464855e-01
2.01272279e-01 -1.32460010e+00 1.01578057e+00 3.89136851e-01
-1.04482219e-01 -4.52631116e-01 -4.67599034e-01 5.40727496e-01
5.65249681e-01 -9.81458798e-02 6.01567209e-01 -1.95462376e-01
-1.01641047e+00 -1.90398812e-01 -6.28170371e-01 2.51019597e-01
-1.66227177e-01 2.05379929e-02 -1.00067019e+00 -2.18842074e-01
-2.98255593e-01 -3.22828367e-02 9.65724349e-01 2.00734988e-01
9.56202328e-01 9.45245624e-02 -5.74267507e-01 1.81292206e-01
1.27200663e+00 2.00508684e-01 5.36165118e-01 4.78628099e-01
6.41568661e-01 6.40338123e-01 8.82891595e-01 1.83966726e-01
4.37317371e-01 9.73702013e-01 5.82746029e-01 -1.82470661e-02
-2.76903242e-01 2.47890875e-02 2.44708672e-01 3.07731181e-01
-8.68052468e-02 -2.68647879e-01 -9.89042401e-01 5.82825482e-01
-1.72193480e+00 -9.42454338e-01 -1.36755213e-01 2.22906566e+00
3.49718601e-01 4.82753724e-01 1.10481337e-01 -1.33054495e-01
6.73014820e-01 3.02386671e-01 -8.35733414e-01 -4.98974413e-01
-4.90444154e-02 -2.71758407e-01 1.21887469e+00 3.62540305e-01
-9.68606651e-01 1.10081995e+00 6.79237318e+00 5.99277079e-01
-1.31717503e+00 -8.47855359e-02 6.32356405e-01 -2.76009887e-01
3.26078117e-01 1.59773022e-01 -1.02144694e+00 5.12273490e-01
1.02157080e+00 1.37969792e-01 3.93847197e-01 1.09231794e+00
4.00119007e-01 -7.24376202e-01 -1.05076981e+00 8.52805436e-01
7.96652436e-02 -1.37989545e+00 -1.69236615e-01 1.05930716e-01
5.63925982e-01 2.71377832e-01 1.70161113e-01 3.00793022e-01
2.09856242e-01 -9.31259632e-01 7.97119141e-01 2.62570590e-01
5.40821075e-01 -6.10279977e-01 7.74058998e-01 2.93427944e-01
-1.07546389e+00 -2.33108819e-01 -4.50723410e-01 -3.42244118e-01
-8.47471803e-02 6.74655214e-02 -1.14370453e+00 -2.80455112e-01
8.94744039e-01 2.70475537e-01 -1.08260059e+00 9.73609030e-01
5.86823151e-02 6.53423488e-01 -4.36266810e-01 -2.60938525e-01
3.64244670e-01 1.44342422e-01 5.73523402e-01 1.21781766e+00
9.19358581e-02 5.81888817e-02 -1.62003383e-01 6.91910386e-01
4.61472943e-02 -2.83732057e-01 -6.89836204e-01 2.29773492e-01
9.30589914e-01 1.16066897e+00 -1.08263588e+00 -4.38644528e-01
-5.82852781e-01 6.98341608e-01 1.85362771e-01 2.92327493e-01
-1.03150690e+00 -3.18768352e-01 7.14208841e-01 5.43056726e-01
5.45940280e-01 -4.71456647e-01 -5.82246184e-01 -6.67906046e-01
4.55728471e-02 -8.03907931e-01 -9.77904797e-02 -8.62054825e-01
-5.64517438e-01 1.56288952e-01 1.21765219e-01 -1.08068097e+00
-5.27696386e-02 -7.85543382e-01 -2.62256444e-01 3.16465467e-01
-1.32792377e+00 -7.44521916e-01 -5.96063077e-01 1.14773609e-01
7.19970763e-01 1.79828405e-01 2.34783903e-01 2.24993169e-01
-4.41873819e-01 4.65250880e-01 2.27670968e-01 6.86333552e-02
7.27775276e-01 -7.30910957e-01 6.64797127e-01 9.54334259e-01
9.54968035e-02 3.77160966e-01 8.53872597e-01 -3.61373633e-01
-1.61496902e+00 -1.11010218e+00 4.04124349e-01 -1.04221237e+00
4.20706511e-01 -4.93392944e-01 -7.15806186e-01 7.42883861e-01
-1.36016816e-01 4.95986715e-02 -1.15849096e-02 1.23002902e-02
-2.25640669e-01 -2.98822671e-01 -9.97108221e-01 8.72460008e-01
9.91525829e-01 -3.92660648e-01 -2.66626269e-01 -2.25245822e-02
7.17961252e-01 -2.65155852e-01 -5.51653504e-01 4.44904089e-01
6.81699157e-01 -1.09562516e+00 9.95977163e-01 -3.21701095e-02
6.36074599e-03 -6.54006779e-01 -1.19725525e-01 -8.62064719e-01
-2.99271077e-01 -1.14772409e-01 1.72363088e-01 7.39805818e-01
4.38136339e-01 -7.91458547e-01 8.35269451e-01 5.48603952e-01
-1.76319480e-01 -6.34926975e-01 -8.92928898e-01 -8.24293017e-01
-2.29448780e-01 -5.99435627e-01 2.57782906e-01 5.67236423e-01
-3.70423406e-01 4.36825961e-01 -1.08164549e-01 4.16865051e-01
2.33482629e-01 -4.94504385e-02 1.21573567e+00 -1.14470327e+00
1.25364393e-01 -3.99033338e-01 -8.57182443e-01 -1.23663151e+00
-4.66007829e-01 -4.83314037e-01 1.82420716e-01 -1.28753531e+00
-1.02466077e-01 -7.63651788e-01 8.02890286e-02 2.55774885e-01
-1.61489412e-01 6.17264450e-01 4.71040457e-01 1.83354869e-01
-8.46001923e-01 1.07584544e-01 9.23590243e-01 -3.48486714e-02
-1.35156676e-01 -8.74677151e-02 -5.43182194e-01 7.48467326e-01
7.02827990e-01 -4.10884261e-01 -2.71913081e-01 -6.74409449e-01
2.26228520e-01 -3.66163880e-01 5.79942584e-01 -1.45054328e+00
3.14337939e-01 -1.79955035e-01 3.94651860e-01 -5.39954424e-01
5.60485899e-01 -6.85430765e-01 -1.58713698e-01 5.67188978e-01
-5.58683500e-02 1.23444319e-01 5.95279336e-01 5.78645468e-01
1.34254679e-01 -8.99991766e-02 7.12719738e-01 1.85452297e-03
-1.35072935e+00 -2.41974205e-01 -8.86815369e-01 -1.24677159e-01
1.28420699e+00 -5.99990666e-01 -6.76946163e-01 -2.10672066e-01
-2.19755769e-01 3.47924471e-01 7.85442352e-01 7.24104881e-01
3.27996105e-01 -8.18389893e-01 -5.97775578e-01 1.23376787e-01
3.55018318e-01 -5.46317408e-03 1.56346306e-01 7.93867886e-01
-8.12713444e-01 7.16938972e-01 -3.19310635e-01 -9.08941984e-01
-1.31821549e+00 4.69524622e-01 1.13912016e-01 3.47136378e-01
-6.66243017e-01 7.00175107e-01 3.43321323e-01 -5.60222305e-02
2.08629034e-02 -5.98932266e-01 -5.82840554e-02 7.67867407e-03
3.29228222e-01 6.63618445e-01 -8.88559073e-02 -6.98358834e-01
-5.26165187e-01 4.49497044e-01 -5.73132150e-02 1.16112262e-01
7.99560785e-01 -3.66603643e-01 2.34658137e-01 4.50159222e-01
7.37664521e-01 1.69479504e-01 -1.65960431e+00 -2.05205485e-01
-1.96615756e-02 -5.68146110e-01 1.51168019e-01 -4.93553519e-01
-8.31306279e-01 7.33616710e-01 1.02765703e+00 1.80977866e-01
7.27370441e-01 3.72176617e-01 6.79042101e-01 7.32896507e-01
4.39297438e-01 -1.29126024e+00 -2.43036933e-02 4.43349749e-01
3.92817467e-01 -1.47771287e+00 2.00432092e-01 -3.92145216e-01
-5.98118663e-01 8.65305305e-01 8.92981648e-01 -2.93715656e-01
2.17765495e-02 2.28584021e-01 2.29492262e-02 -2.97853202e-01
-7.63019085e-01 -4.43806320e-01 -5.34969717e-02 6.27389014e-01
-7.80725479e-02 1.55068353e-01 -9.61747542e-02 -1.43881232e-01
-1.28934622e-01 -2.07221627e-01 5.72228134e-01 8.41412485e-01
-9.68940914e-01 -4.48752403e-01 -5.75259209e-01 7.44244933e-01
-1.29454330e-01 6.37403131e-02 -2.54120201e-01 1.02713823e+00
4.24683720e-01 1.21147585e+00 4.20318425e-01 -3.85756910e-01
3.56678069e-01 -3.44432235e-01 2.68329442e-01 -6.73683167e-01
-1.26130059e-01 -4.72418219e-01 1.89007491e-01 -7.08653867e-01
-1.94640011e-01 -7.11595774e-01 -1.20891142e+00 -5.16217291e-01
-4.93283868e-01 -3.70272100e-01 7.95371473e-01 9.56639469e-01
6.19481802e-01 4.48709100e-01 1.92821413e-01 -9.24421608e-01
-2.60160804e-01 -7.39683628e-01 -1.04777977e-01 7.07237273e-02
2.32800722e-01 -8.77047420e-01 -1.29175961e-01 -3.26580703e-02]
|
[8.218766212463379, -1.1341255903244019]
|
9b1242c7-3b3a-4c5a-85de-6db4f63f8141
|
news-driven-stock-prediction-with-attention
|
2004.01878
| null |
https://arxiv.org/abs/2004.01878v1
|
https://arxiv.org/pdf/2004.01878v1.pdf
|
News-Driven Stock Prediction With Attention-Based Noisy Recurrent State Transition
|
We consider direct modeling of underlying stock value movement sequences over time in the news-driven stock movement prediction. A recurrent state transition model is constructed, which better captures a gradual process of stock movement continuously by modeling the correlation between past and future price movements. By separating the effects of news and noise, a noisy random factor is also explicitly fitted based on the recurrent states. Results show that the proposed model outperforms strong baselines. Thanks to the use of attention over news events, our model is also more explainable. To our knowledge, we are the first to explicitly model both events and noise over a fundamental stock value state for news-driven stock movement prediction.
|
['He-Yan Huang', 'Yue Zhang', 'Xiao Liu', 'Changsen Yuan']
|
2020-04-04
| null | null | null | null |
['stock-prediction']
|
['time-series']
|
[-5.36578298e-01 -1.20498903e-01 -7.86851823e-01 -2.85191298e-01
-5.08048475e-01 -4.02964056e-01 1.06692684e+00 -9.50341821e-02
-2.02537686e-01 7.93000758e-01 1.05664706e+00 -1.74063519e-01
2.63797760e-01 -1.10710943e+00 -8.45741868e-01 -1.88808575e-01
-2.53565341e-01 3.02367985e-01 4.02102798e-01 -6.61596537e-01
2.57052213e-01 -6.95333257e-02 -1.21091568e+00 9.92167462e-03
2.53655344e-01 1.01024091e+00 5.47244921e-02 4.83903915e-01
-5.39380193e-01 1.73822844e+00 -4.50904071e-01 -4.36171860e-01
5.27413785e-01 -5.92497230e-01 -4.44626838e-01 8.69761109e-02
-6.43550158e-02 -6.41565144e-01 -6.66324675e-01 1.06117022e+00
-2.29068846e-01 1.18979309e-02 4.43876803e-01 -1.05143547e+00
-1.02024841e+00 1.50464773e+00 -4.43855286e-01 8.54541779e-01
-6.66640252e-02 2.39203215e-01 1.55138171e+00 -5.75028837e-01
4.95047659e-01 1.09023511e+00 6.76733375e-01 2.63193071e-01
-1.16555738e+00 -6.23493016e-01 9.36935782e-01 -1.65767521e-02
-4.70762730e-01 -1.81263581e-01 1.09316790e+00 -5.83136380e-01
1.09337950e+00 2.53396165e-02 1.21016169e+00 1.53140557e+00
6.99112236e-01 1.21858323e+00 9.78201509e-01 8.89935344e-02
2.22775996e-01 -3.00392449e-01 5.38328707e-01 -4.51625185e-03
2.42035836e-01 5.17036736e-01 -6.91314816e-01 -2.36071452e-01
8.17428470e-01 7.06564903e-01 1.39407352e-01 1.92397267e-01
-1.24568331e+00 7.58878112e-01 6.46918491e-02 2.38471016e-01
-1.01082420e+00 5.82208216e-01 2.27148175e-01 6.03348792e-01
1.05993307e+00 2.29880616e-01 -9.07570302e-01 -5.55792570e-01
-1.26708591e+00 5.87334454e-01 8.69095743e-01 8.40391457e-01
3.49941880e-01 8.19202363e-01 -2.51420557e-01 7.18793347e-02
4.49116260e-01 8.18016469e-01 8.55033100e-01 -5.07023931e-01
3.75196159e-01 2.15395883e-01 5.68786860e-01 -7.35012233e-01
-4.48337168e-01 -8.28655779e-01 -4.97024298e-01 -6.63808957e-02
3.16690594e-01 -3.58725637e-01 -1.03685808e+00 1.72658265e+00
-2.60283142e-01 9.80180562e-01 2.12060571e-01 3.37881893e-01
5.31605184e-01 1.05048454e+00 3.66123579e-02 -4.18242484e-01
1.26767421e+00 -1.00572813e+00 -1.24102032e+00 -3.36865753e-01
2.33261451e-01 -3.00554842e-01 4.00308430e-01 1.79451704e-02
-1.31947303e+00 -3.29868287e-01 -8.04041266e-01 3.52222323e-01
-2.52227724e-01 -4.66273338e-01 5.20621181e-01 -9.20941774e-03
-7.32553661e-01 6.48423851e-01 -1.45237494e+00 3.26191068e-01
-2.11520381e-02 -1.80509269e-01 4.01956081e-01 6.39233112e-01
-1.62404823e+00 9.46343303e-01 2.75913656e-01 3.15919407e-02
-8.77671659e-01 -9.59903896e-01 -7.43648767e-01 1.42466098e-01
3.26393098e-01 -4.92244065e-01 1.65531766e+00 -8.76117885e-01
-1.66212881e+00 1.40360221e-01 -1.95445865e-01 -1.31024241e+00
5.86733580e-01 -3.73840362e-01 -7.59341478e-01 -4.08471227e-01
-5.15703708e-02 -4.22469266e-02 6.70278609e-01 -6.92395687e-01
-6.34773374e-01 8.64767879e-02 -2.28095636e-01 -5.03637232e-02
1.20233402e-01 4.55112371e-04 -3.02752089e-02 -1.37762809e+00
-7.81381652e-02 -9.18466449e-01 -2.52055496e-01 -7.18861282e-01
-2.76860118e-01 -4.66461107e-02 4.23633039e-01 -8.50057006e-01
1.47597539e+00 -1.97946906e+00 -3.94592226e-01 -2.38553640e-02
-7.12881386e-02 -2.26274848e-01 -1.62193924e-01 7.12319195e-01
-1.49206504e-01 1.20838076e-01 7.76510611e-02 -4.45469379e-01
2.61555433e-01 2.76170552e-01 -1.27861559e+00 3.23268175e-01
1.13290638e-01 1.59032202e+00 -6.49124026e-01 4.71813351e-01
-1.27684996e-01 1.96686640e-01 -2.94381678e-01 -3.77626941e-02
-7.38155961e-01 2.16571808e-01 -4.65365618e-01 2.18690604e-01
4.72760767e-01 -5.15930653e-01 -1.45562217e-01 4.27797079e-01
-3.72130632e-01 9.91881430e-01 -1.13629556e+00 9.19448495e-01
-1.38221934e-01 3.00276071e-01 -5.99825740e-01 -6.08004272e-01
1.02888131e+00 4.77825642e-01 6.24840975e-01 -7.58975089e-01
1.40946805e-01 -1.10820770e-01 -2.84697805e-02 7.90637732e-02
6.88066065e-01 -6.36531949e-01 -2.84824401e-01 8.69314492e-01
-1.75012097e-01 1.51673630e-01 -3.34179662e-02 5.39827347e-02
8.77192795e-01 1.20400980e-01 2.58273542e-01 -2.11714298e-01
-1.17456049e-01 -3.01838033e-02 1.00818193e+00 7.69159198e-01
-1.43008843e-01 2.45236740e-01 5.60091078e-01 -9.12539482e-01
-1.06335497e+00 -9.50599730e-01 1.93579420e-01 1.04530144e+00
-1.13666058e-01 -3.53973150e-01 1.37268482e-02 -3.03484946e-01
4.09480274e-01 1.08276212e+00 -9.85763431e-01 -3.52169946e-02
-6.40599549e-01 -1.06073129e+00 1.01788104e-01 7.87768066e-01
3.10596853e-01 -1.09490800e+00 -1.86974272e-01 8.44860554e-01
5.28396666e-02 -1.02525842e+00 -6.52325153e-01 1.21560149e-01
-9.52632010e-01 -5.01536727e-01 -7.86161125e-01 -2.02158481e-01
-1.24742985e-01 -5.21951243e-02 1.26642287e+00 -1.87253311e-01
5.49634993e-01 3.24585080e-01 -1.97376743e-01 -9.01311874e-01
-5.22242725e-01 -2.05703937e-02 -1.23869441e-01 2.31515244e-01
3.08643878e-01 -5.42045653e-01 -6.16104841e-01 -3.52574512e-02
-8.30560565e-01 6.17148764e-02 2.73843974e-01 6.63547814e-01
5.06241620e-01 9.74703804e-02 5.74710846e-01 -7.88137496e-01
4.01178539e-01 -9.15818989e-01 -9.10669863e-01 1.79589987e-02
-7.34525800e-01 3.20729584e-01 2.62521088e-01 -6.52367949e-01
-1.25551271e+00 -3.11826259e-01 -1.16422758e-01 -1.25736937e-01
6.10964466e-03 8.71856332e-01 5.99754035e-01 8.64635468e-01
-1.69825509e-01 6.98434114e-01 -5.05895913e-03 -6.16678655e-01
2.28595048e-01 -1.41707480e-01 3.88056427e-01 -9.54207312e-03
8.54578793e-01 6.05577111e-01 -3.36649895e-01 -4.45140272e-01
-1.16352808e+00 -2.61269540e-01 -1.42624334e-01 6.83741570e-02
7.00863838e-01 -1.47998321e+00 -2.50823468e-01 9.86497462e-01
-8.87554169e-01 -5.09174526e-01 -7.71375597e-01 8.41331542e-01
-5.03192544e-01 -4.92930152e-02 -1.46252894e+00 -1.07526302e+00
-1.55112401e-01 -6.17049873e-01 7.40782797e-01 2.85137771e-03
-1.55756071e-01 -1.40367806e+00 5.52783012e-01 -2.43264735e-01
7.93144047e-01 2.21431330e-01 3.97143364e-01 -1.09599423e+00
-8.32297862e-01 -4.18713152e-01 4.08810496e-01 5.40910885e-02
3.06326210e-01 4.39171121e-02 -4.85613495e-01 1.31165143e-02
3.14707249e-01 1.58876687e-01 1.27290559e+00 8.12687337e-01
1.49398878e-01 -6.37628078e-01 5.60856275e-02 4.28390056e-01
1.10372019e+00 3.69224101e-01 5.36730170e-01 7.20214307e-01
2.59493113e-01 2.33663797e-01 3.94354701e-01 7.64431119e-01
8.37610245e-01 3.01649243e-01 4.65034992e-01 8.03757180e-03
3.03632557e-01 -7.25517333e-01 6.99360371e-01 1.02742958e+00
-1.39974691e-02 -2.84537107e-01 -6.39079809e-01 6.47983015e-01
-1.88863730e+00 -1.68060148e+00 -2.61058688e-01 1.58517289e+00
7.09830701e-01 6.99684858e-01 5.21635711e-01 -2.61066973e-01
4.55056727e-01 8.48080158e-01 -7.09156811e-01 2.74427116e-01
-3.98636609e-01 -2.27306172e-01 8.79597425e-01 5.91171026e-01
-1.12025523e+00 9.54387903e-01 7.87201548e+00 2.61578619e-01
-1.05669260e+00 -1.49999738e-01 5.08986950e-01 -3.25754225e-01
-6.97379589e-01 5.65092601e-02 -1.23875833e+00 9.05564487e-01
1.17396581e+00 -5.39548814e-01 7.31062815e-02 7.57249832e-01
5.22098124e-01 4.29838151e-01 -6.39536917e-01 3.11224788e-01
-4.52524513e-01 -1.52372050e+00 2.69785315e-01 2.83987850e-01
9.65350628e-01 4.40021932e-01 3.62214953e-01 4.01775628e-01
1.12667942e+00 -3.66069525e-01 1.41613495e+00 1.13123775e+00
-2.82346547e-01 -6.98435783e-01 7.23404467e-01 6.15718782e-01
-1.33116341e+00 -2.62137502e-01 -5.51483668e-02 -5.42574227e-01
7.45583534e-01 5.66059172e-01 -8.39327797e-02 1.58527926e-01
2.80584991e-01 1.38716221e+00 -1.75330758e-01 6.16811752e-01
-1.95402503e-01 1.21719587e+00 -1.55565202e-01 1.66119486e-01
5.34848809e-01 -1.47115171e-01 6.11962497e-01 9.04948771e-01
5.34182191e-01 1.14363529e-01 2.08121970e-01 9.14901793e-01
4.64839227e-02 -2.28621438e-01 -4.42233503e-01 -1.72001064e-01
1.14696026e-01 3.72599661e-01 -6.74527168e-01 -6.13853514e-01
-8.93781185e-01 6.17448211e-01 -1.35062054e-01 4.61124361e-01
-9.17972207e-01 2.93154538e-01 1.09450483e+00 2.41194844e-01
8.51414621e-01 -3.71151865e-01 -1.52717188e-01 -1.68729067e+00
-1.84995890e-01 -4.55411881e-01 3.52092206e-01 -5.28232634e-01
-1.64727557e+00 5.13697028e-01 -4.59736511e-02 -1.25906742e+00
-6.08855724e-01 -1.44090727e-01 -9.25046861e-01 8.07173371e-01
-1.89000320e+00 -7.74358571e-01 6.73537731e-01 3.82660806e-01
7.95586288e-01 -2.46330485e-01 2.59661615e-01 -3.78719479e-01
-4.90299881e-01 -8.79481882e-02 3.58766139e-01 5.06496429e-01
1.70005500e-01 -1.31566644e+00 1.54786670e+00 1.08034468e+00
4.55040097e-01 5.43695092e-01 8.43507230e-01 -1.29902935e+00
-1.00292218e+00 -1.12940466e+00 9.74909246e-01 -5.61917424e-01
1.44467962e+00 -1.15150586e-01 -1.22714365e+00 1.33454776e+00
2.98685551e-01 -5.89298308e-02 5.16096532e-01 -2.25499094e-01
-4.05233741e-01 1.83180347e-01 -4.29451168e-01 4.10207748e-01
6.90397620e-01 -4.82301056e-01 -1.10303271e+00 -1.10787697e-01
1.18605673e+00 -2.56652445e-01 -7.55881608e-01 1.09607428e-01
3.77598405e-01 -7.41826952e-01 8.44666660e-01 -8.06825519e-01
3.74102443e-01 -6.99237511e-02 1.36309683e-01 -1.52490807e+00
-8.08678687e-01 -7.23306656e-01 -6.84790254e-01 1.18681407e+00
6.68369591e-01 -9.11387503e-01 7.83077955e-01 7.34783292e-01
2.14969516e-01 -2.81951100e-01 -7.52121329e-01 -1.10947776e+00
1.74480990e-01 -6.53841674e-01 1.17459285e+00 8.64109039e-01
-1.22107513e-01 -2.29718094e-03 -9.82489705e-01 1.65599585e-01
3.40385318e-01 3.70328754e-01 4.19935346e-01 -1.31702816e+00
-5.13724506e-01 -6.49738550e-01 -1.12809390e-01 -1.46394241e+00
3.15409929e-01 -3.57859582e-01 -1.60029139e-02 -1.21118486e+00
1.18338533e-01 3.28024954e-01 -7.33450770e-01 6.36245981e-02
-3.16553921e-01 -9.62807238e-02 5.06704330e-01 5.27947128e-01
-3.76508951e-01 8.61851990e-01 1.14369118e+00 -3.43699791e-02
-2.87491113e-01 4.81380135e-01 -7.48912752e-01 8.29654872e-01
9.05865490e-01 -6.96368814e-01 -1.65303007e-01 -9.61698964e-02
6.21100426e-01 4.14869219e-01 9.48136300e-02 -4.59889323e-01
1.99142799e-01 -5.22108972e-01 4.50754613e-02 -1.02348578e+00
2.19174117e-01 -5.02178550e-01 5.89794099e-01 6.71462297e-01
-6.42689466e-01 5.78345954e-01 1.11185551e-01 1.09860742e+00
-5.13127446e-01 2.26091221e-01 4.04662699e-01 -3.03636879e-01
-6.67486548e-01 5.50689459e-01 -9.39200103e-01 6.32128716e-02
8.17625046e-01 1.36093795e-01 -1.79006279e-01 -1.04515898e+00
-9.98516083e-01 2.43781760e-01 8.97039548e-02 6.57586038e-01
2.81878352e-01 -1.45372915e+00 -8.35988462e-01 1.23447850e-01
-2.81312644e-01 -5.38083255e-01 1.76595584e-01 6.86716437e-01
-5.07874694e-03 3.56450021e-01 1.23469844e-01 -2.36289781e-02
-5.68076968e-01 8.14176440e-01 2.42515847e-01 -7.49501288e-01
-1.03382230e+00 3.86816561e-01 -5.37732914e-02 1.33292362e-01
9.14783627e-02 -9.60275888e-01 -4.00253832e-01 5.14414787e-01
1.01104689e+00 3.70907158e-01 -4.32640791e-01 -7.13227928e-01
8.12959298e-02 4.49649394e-01 -2.22079352e-01 -3.91177863e-01
1.78201950e+00 -4.16707516e-01 1.90187827e-01 1.24827838e+00
7.82006919e-01 -1.32622540e-01 -1.79654825e+00 -6.33234859e-01
5.97264588e-01 3.52979302e-02 1.99010789e-01 -5.38877070e-01
-1.23913467e+00 2.03725159e-01 4.03655274e-03 6.76151276e-01
5.56207836e-01 -3.87565196e-02 1.27351785e+00 8.77192765e-02
3.65032047e-01 -8.97541225e-01 -1.49931222e-01 1.03442681e+00
7.03686357e-01 -1.17043591e+00 -2.79371262e-01 9.43873972e-02
-9.38564956e-01 8.53504360e-01 -1.13934772e-02 -7.50170052e-01
1.36002088e+00 7.34942913e-01 3.21691602e-01 -3.62848490e-02
-1.38632977e+00 -3.87778491e-01 1.80860087e-01 -6.38815090e-02
2.13704124e-01 9.51107405e-03 9.78801697e-02 9.49331284e-01
-5.18279791e-01 1.30959794e-01 7.71107018e-01 1.08463049e+00
-4.82009768e-01 -6.98985815e-01 -2.62403488e-01 5.22486269e-01
-1.05053401e+00 -3.15584093e-01 -4.63725701e-02 8.17131460e-01
-5.32781065e-01 6.50225043e-01 6.05553925e-01 -9.07548293e-02
4.77416217e-01 3.98805708e-01 -2.93740451e-01 -4.82909173e-01
-6.14565611e-01 4.71200168e-01 -2.27213889e-01 -4.49335337e-01
-3.69927138e-01 -9.16113257e-01 -1.05440271e+00 -4.47075546e-01
-7.55131394e-02 8.22919533e-02 2.43325248e-01 9.10463810e-01
2.56152153e-01 6.21887505e-01 7.12675631e-01 -7.93590546e-01
-9.02026355e-01 -9.57927346e-01 -1.20449531e+00 3.69827211e-01
9.11785483e-01 -5.15675485e-01 -6.70038879e-01 6.01498485e-02]
|
[4.450096130371094, 4.25547456741333]
|
d68c9fa1-2f0e-4ca4-b573-8ffa3761f92a
|
exploiting-unlabeled-data-for-neural
|
1611.08987
| null |
http://arxiv.org/abs/1611.08987v2
|
http://arxiv.org/pdf/1611.08987v2.pdf
|
Exploiting Unlabeled Data for Neural Grammatical Error Detection
|
Identifying and correcting grammatical errors in the text written by
non-native writers has received increasing attention in recent years. Although
a number of annotated corpora have been established to facilitate data-driven
grammatical error detection and correction approaches, they are still limited
in terms of quantity and coverage because human annotation is labor-intensive,
time-consuming, and expensive. In this work, we propose to utilize unlabeled
data to train neural network based grammatical error detection models. The
basic idea is to cast error detection as a binary classification problem and
derive positive and negative training examples from unlabeled data. We
introduce an attention-based neural network to capture long-distance
dependencies that influence the word being detected. Experiments show that the
proposed approach significantly outperforms SVMs and convolutional networks
with fixed-size context window.
|
['Yang Liu', 'Zhuoran Liu']
|
2016-11-28
| null | null | null | null |
['grammatical-error-detection']
|
['natural-language-processing']
|
[ 2.45581746e-01 -1.13609806e-01 -6.35989383e-02 -8.73130560e-01
-5.47002971e-01 -1.54362321e-01 6.91803843e-02 6.51127160e-01
-8.24881852e-01 9.37439442e-01 -9.14054140e-02 -4.42464411e-01
3.25736851e-01 -6.78655803e-01 -5.48045516e-01 -7.76936710e-02
3.64696920e-01 2.34805942e-01 6.80418536e-02 -1.78791419e-01
6.94551408e-01 1.06831253e-01 -1.36303830e+00 2.58895811e-02
1.63404465e+00 6.62869930e-01 4.07648712e-01 6.07516408e-01
-4.35181409e-01 8.74146760e-01 -8.49799871e-01 -5.76249540e-01
-3.42784762e-01 -7.30345905e-01 -1.04784060e+00 -2.50512332e-01
4.21378940e-01 -1.02892868e-01 -7.52816126e-02 1.36388195e+00
3.57863218e-01 2.25019559e-01 3.92973930e-01 -7.18287587e-01
-1.24099541e+00 5.31008899e-01 -1.77967951e-01 6.67276502e-01
2.09811345e-01 -2.12573215e-01 9.48880374e-01 -1.22677815e+00
4.26903993e-01 9.68867958e-01 7.49381959e-01 9.17096734e-01
-9.60864842e-01 -5.65084040e-01 2.73863107e-01 4.62117046e-01
-1.17860401e+00 -3.34062845e-01 6.78654552e-01 -3.57790023e-01
1.48395860e+00 9.87544507e-02 4.52000678e-01 9.65346396e-01
2.28962898e-01 7.42153466e-01 7.47152388e-01 -1.09026563e+00
1.15420386e-01 8.43654051e-02 6.37785852e-01 8.60261261e-01
3.85445118e-01 -1.67800799e-01 -4.77722466e-01 1.38276204e-01
3.85568470e-01 -4.63925824e-02 -3.00396085e-01 4.16812629e-01
-5.74037850e-01 9.56781507e-01 3.28550339e-01 5.28389871e-01
-1.60926029e-01 1.42516524e-01 6.21763289e-01 4.88712430e-01
9.51842487e-01 4.67185438e-01 -5.32721639e-01 -4.54540074e-01
-8.14059079e-01 5.97247481e-03 5.33601284e-01 7.81296730e-01
5.74504435e-01 1.85015813e-01 2.73884777e-02 1.23054612e+00
4.24245536e-01 2.08647311e-01 7.41286814e-01 -1.66077331e-01
7.12876081e-01 8.23959947e-01 -2.62039085e-03 -1.09820700e+00
-3.24278653e-01 -3.24043691e-01 -6.39390230e-01 -1.22043602e-02
5.06073058e-01 -2.78510213e-01 -8.40085208e-01 1.41019034e+00
-3.65864225e-02 -1.72639802e-01 -1.00205697e-01 9.66225326e-01
7.72009134e-01 5.53243339e-01 4.85173762e-01 9.19321738e-03
1.02632344e+00 -1.10370719e+00 -9.59885359e-01 -6.03121340e-01
1.02534020e+00 -7.04608381e-01 1.09552813e+00 1.89764619e-01
-9.35422242e-01 -5.57733357e-01 -1.03759909e+00 -4.47638839e-01
-2.86774516e-01 3.20022941e-01 3.10034782e-01 6.29422069e-01
-6.32497907e-01 6.85303748e-01 -8.90894353e-01 -5.36396027e-01
4.91969943e-01 2.26542071e-01 -3.51374179e-01 6.16721483e-03
-1.22863758e+00 1.23294878e+00 5.05842865e-01 2.81017512e-01
-3.47533166e-01 -3.38327706e-01 -9.11901534e-01 9.17995051e-02
-1.57198068e-02 -1.38456702e-01 1.22775757e+00 -1.27234554e+00
-1.10940063e+00 9.65240657e-01 -4.73214626e-01 -4.41732407e-01
2.35334381e-01 -4.02781934e-01 -5.63142836e-01 -2.93911099e-01
8.28914195e-02 2.25372359e-01 6.38980031e-01 -7.43390560e-01
-9.00406361e-01 -6.28729939e-01 -2.18494385e-01 6.24310151e-02
-5.20108163e-01 6.22383356e-01 -2.41730943e-01 -8.85812879e-01
8.67875591e-02 -6.05527818e-01 -6.49083406e-02 -2.73136318e-01
-3.76011103e-01 -6.50840580e-01 6.34032190e-01 -1.14387417e+00
1.69810593e+00 -1.79808569e+00 6.97556660e-02 -9.25220996e-02
1.27375618e-01 7.82678246e-01 -4.04730439e-03 3.41976523e-01
-7.83586502e-03 2.75429159e-01 -4.45834845e-01 -4.43956137e-01
-3.37488234e-01 2.02319339e-01 -6.39668033e-02 3.12243223e-01
5.54277360e-01 7.19396055e-01 -1.10591924e+00 -4.20195043e-01
1.04183480e-01 3.25260341e-01 -5.26945591e-01 4.04354066e-01
-2.56104141e-01 5.58244646e-01 -2.90484905e-01 7.91857243e-01
4.36694682e-01 -1.22725159e-01 1.41500652e-01 3.96896809e-01
-1.22509999e-02 7.90448666e-01 -7.84545004e-01 1.55300689e+00
-6.38108492e-01 9.04438019e-01 -2.90223151e-01 -1.25028467e+00
1.33297145e+00 4.88048829e-02 -4.65644479e-01 -7.20921636e-01
3.33680838e-01 5.89218318e-01 1.40022919e-01 -8.75330091e-01
8.12399149e-01 -3.12410861e-01 -5.40613309e-02 4.17589009e-01
1.55061275e-01 3.46722305e-01 2.29320988e-01 -1.94080055e-01
1.12960279e+00 7.11344108e-02 3.68426770e-01 -2.96950489e-02
7.69360244e-01 3.26521173e-02 8.41380060e-01 4.69438821e-01
-3.30086470e-01 3.98806095e-01 2.14719862e-01 -6.15587175e-01
-9.38133597e-01 -2.56757587e-01 -2.11493284e-01 1.29707623e+00
1.31682707e-02 -2.18036681e-01 -8.35691571e-01 -1.08867478e+00
3.05791479e-02 7.75981784e-01 -5.18978119e-01 -1.77570835e-01
-8.74901235e-01 -6.12826645e-01 5.25343955e-01 7.95060575e-01
4.42568928e-01 -1.38051128e+00 -3.99506956e-01 4.07344520e-01
-2.09446520e-01 -6.40574157e-01 -3.91483784e-01 4.21058208e-01
-9.53345716e-01 -1.15312719e+00 -3.84145468e-01 -1.14442587e+00
1.17689621e+00 -1.83285460e-01 1.05920088e+00 7.95731843e-01
-3.78389806e-01 -2.16351017e-01 -6.73656046e-01 -4.80105340e-01
-5.05654097e-01 3.63036245e-01 -3.36077102e-02 -3.07948440e-01
1.04134464e+00 -8.43464360e-02 -1.77855372e-01 -9.02446732e-03
-5.19836128e-01 -2.33565643e-01 3.08694601e-01 1.33553374e+00
2.36424163e-01 -4.43794996e-01 9.23353612e-01 -1.15054536e+00
9.36041355e-01 -4.48405117e-01 -5.00962377e-01 4.04030591e-01
-9.43900704e-01 1.20257959e-01 7.09607124e-01 -2.99322288e-02
-1.19939518e+00 -2.25281015e-01 -5.44861853e-01 8.84041786e-02
-2.71695942e-01 7.85060942e-01 -7.24786939e-03 -1.24672651e-01
6.84872866e-01 5.56504913e-02 -2.33301640e-01 -7.51688182e-01
-2.92898156e-02 1.21780360e+00 4.08886969e-01 -4.69978690e-01
2.73497254e-01 -3.60769600e-01 -5.79003692e-01 -5.82772553e-01
-1.03217602e+00 -1.97899669e-01 -8.86327803e-01 -1.04989342e-01
7.23126888e-01 -5.89826345e-01 -3.38283479e-01 5.97764432e-01
-1.56371260e+00 -3.79070975e-02 5.50280362e-02 4.40363765e-01
4.99919653e-02 3.73014241e-01 -8.04843068e-01 -9.29057479e-01
-6.04488432e-01 -8.87531221e-01 6.46114886e-01 4.11838740e-01
-3.82868737e-01 -1.18726349e+00 3.00967377e-02 3.99125159e-01
4.53931838e-01 -1.81950927e-01 1.06994987e+00 -9.92739856e-01
-2.56989717e-01 -4.73496377e-01 -3.29751819e-01 6.97398365e-01
7.17779398e-02 -6.06844015e-02 -8.19365084e-01 -1.75376497e-02
-3.15340161e-01 -4.10990417e-01 8.24363172e-01 6.96256608e-02
1.22191584e+00 -8.57833177e-02 -2.72327632e-01 3.56736988e-01
1.29822600e+00 4.55192417e-01 4.22699183e-01 4.04370666e-01
7.36973345e-01 5.87585390e-01 7.82831609e-01 2.60665745e-01
2.96415716e-01 4.33042377e-01 3.34964335e-01 1.90035686e-01
9.82957110e-02 -2.87723571e-01 1.37701377e-01 1.00317144e+00
-1.82126872e-02 -5.06948054e-01 -1.11071098e+00 8.93078208e-01
-1.89183176e+00 -7.80854166e-01 -4.71742839e-01 2.05958796e+00
1.04546964e+00 -1.25008048e-02 -3.95851791e-01 2.19958991e-01
9.87808347e-01 -1.95103630e-01 -4.00567204e-01 -8.55674326e-01
4.23962511e-02 5.42919397e-01 1.83163196e-01 6.52465641e-01
-1.03062475e+00 1.17138970e+00 6.21268511e+00 4.48431641e-01
-1.17735493e+00 3.33740771e-01 6.11668110e-01 1.65917724e-01
-1.89370945e-01 -1.23515762e-01 -1.04278862e+00 6.56569839e-01
1.01338780e+00 -3.01155914e-02 -2.65134200e-02 9.18649495e-01
1.20894819e-01 -1.05609402e-01 -1.02718031e+00 8.71735334e-01
3.25648755e-01 -1.06647706e+00 -3.64103466e-01 -3.42716545e-01
6.43234193e-01 -6.54655546e-02 -3.63679200e-01 4.88517225e-01
2.11164609e-01 -1.18115664e+00 5.89439034e-01 4.60119128e-01
6.65451527e-01 -7.49080598e-01 1.12330198e+00 4.11564499e-01
-5.48083603e-01 -1.39595404e-01 -6.88180387e-01 -6.54514015e-01
-3.43264043e-02 6.91632450e-01 -6.16469264e-01 1.93401221e-02
7.64359832e-01 7.45633423e-01 -5.30965507e-01 9.25669909e-01
-7.41042972e-01 8.65399778e-01 1.35066763e-01 -5.57114661e-01
-2.29624659e-02 -1.99230149e-01 1.08089618e-01 1.35141134e+00
3.78010958e-01 1.67787820e-02 -4.09055725e-02 9.19716895e-01
-4.85792100e-01 4.70988035e-01 -4.29069877e-01 -1.55189514e-01
5.55059314e-01 1.00994122e+00 -3.84055883e-01 -1.65164068e-01
-6.25866532e-01 1.29436040e+00 1.04840660e+00 -6.25555441e-02
-5.83488882e-01 -9.66196299e-01 6.50551617e-01 -1.19119942e-01
2.91383862e-01 -9.38696265e-02 -4.77979511e-01 -1.17942405e+00
1.77684963e-01 -5.91281652e-01 2.93096244e-01 -4.20558870e-01
-1.41047382e+00 6.86832666e-01 -6.18536472e-01 -8.11731160e-01
-5.18424697e-02 -7.53061891e-01 -7.33248830e-01 1.25730586e+00
-1.56775546e+00 -7.93883264e-01 -4.66102839e-01 1.97173525e-02
8.49999487e-01 -1.80908725e-01 1.03324783e+00 5.62008560e-01
-1.03670752e+00 9.13759172e-01 2.05887765e-01 5.79766870e-01
6.76583469e-01 -1.22540677e+00 5.34599662e-01 1.15909660e+00
-1.74451582e-02 8.89651775e-01 5.15471041e-01 -8.69468033e-01
-9.38348591e-01 -1.19753718e+00 1.83770442e+00 -2.91206747e-01
4.60285008e-01 -1.76816434e-01 -1.32809997e+00 7.39765584e-01
8.48020762e-02 4.04706210e-01 7.79623330e-01 3.94748896e-01
-1.84684649e-01 1.63462356e-01 -1.15999532e+00 9.85891521e-02
9.35334742e-01 -6.28987491e-01 -8.52524281e-01 2.45007277e-01
2.83296734e-01 -4.91532147e-01 -6.45731211e-01 1.57006398e-01
1.86626896e-01 -7.55979180e-01 1.96033880e-01 -9.58882987e-01
7.85319269e-01 -9.80051458e-02 2.24964455e-01 -1.51791871e+00
-1.65276662e-01 -7.91645050e-02 5.46103083e-02 1.41514242e+00
6.07130766e-01 -5.27379096e-01 7.01268911e-01 7.89297462e-01
-4.33006674e-01 -7.26139605e-01 -9.16577458e-01 -5.47392309e-01
1.96440965e-01 -3.62751782e-01 3.56746912e-01 1.14681399e+00
4.01525170e-01 3.92518789e-01 -4.29799616e-01 2.81330664e-02
-2.23665382e-03 -2.39200965e-01 2.75230199e-01 -1.42442977e+00
6.60549477e-02 -2.47223243e-01 -4.55018908e-01 -9.93172944e-01
4.74495620e-01 -9.76623654e-01 4.11251843e-01 -1.41070306e+00
1.02617014e-02 -7.21253633e-01 -2.46366039e-01 7.10307539e-01
-7.52936602e-01 2.19324023e-01 -2.65544385e-01 1.11435562e-01
-4.71248567e-01 6.65650010e-01 5.90306163e-01 -1.27157271e-01
-5.23884445e-02 -1.31603345e-01 -5.11626422e-01 7.76525855e-01
8.87816191e-01 -6.53634369e-01 -2.85511017e-02 -8.90679061e-01
4.49386984e-01 -2.29779720e-01 -4.25939597e-02 -7.40098715e-01
6.28791451e-02 8.47111363e-03 2.48935819e-01 -2.92848498e-01
-3.58958453e-01 -5.81917882e-01 -5.32006323e-01 4.24257636e-01
-5.50105453e-01 2.74314880e-01 1.06992513e-01 4.94201362e-01
-4.51655626e-01 -9.86730278e-01 6.91908777e-01 -1.31506398e-01
-7.56903768e-01 -1.28856702e-02 -3.54712158e-01 2.13678911e-01
9.13788140e-01 -1.43466085e-01 -4.01025683e-01 6.04760461e-02
-5.72890818e-01 1.29431307e-01 3.73986691e-01 6.23795092e-01
6.88719332e-01 -1.10735333e+00 -6.61356986e-01 2.70230621e-01
4.61874694e-01 -6.88625202e-02 9.02779028e-02 5.40277719e-01
-7.86853731e-01 4.26338941e-01 -1.53803334e-01 -1.88376784e-01
-1.55595291e+00 1.76358640e-01 3.37673664e-01 -1.78839430e-01
-4.98816997e-01 1.23398554e+00 -6.12127006e-01 -6.18346870e-01
4.65625972e-01 -3.00237000e-01 -2.96915621e-01 -1.89371869e-01
6.85294986e-01 3.96848798e-01 4.48080242e-01 -5.66064775e-01
-3.23705733e-01 2.03282490e-01 -3.04296315e-01 5.14590263e-01
1.30812502e+00 -2.16629893e-01 -2.63891369e-01 4.35317904e-01
8.56941760e-01 -1.90346256e-01 -6.02438569e-01 -3.06738675e-01
5.17957628e-01 -6.79631591e-01 1.95305392e-01 -8.92146528e-01
-9.53165770e-01 1.28614211e+00 5.28837621e-01 8.13743919e-02
7.72860706e-01 -2.77675569e-01 9.36764717e-01 3.87784690e-01
2.52497703e-01 -1.50628603e+00 -2.23096564e-01 9.11990047e-01
7.14955091e-01 -1.67980683e+00 -4.18084115e-01 -3.64507258e-01
-3.39883775e-01 1.33206844e+00 1.19707835e+00 -7.81196877e-02
3.13269436e-01 -1.11454129e-02 1.36390314e-01 -6.88403621e-02
-5.10974705e-01 -1.02231242e-02 1.95052847e-01 5.39410353e-01
1.19100952e+00 -2.16071934e-01 -9.88668144e-01 7.13481009e-01
1.05995536e-01 6.02322258e-02 4.18347657e-01 1.13843858e+00
-7.28832006e-01 -1.46196306e+00 -2.21450120e-01 6.76079512e-01
-7.28609562e-01 -3.37449998e-01 -3.02575767e-01 3.77999425e-01
3.07530075e-01 1.04476261e+00 1.96628198e-01 -2.57521421e-01
2.26874024e-01 5.67679346e-01 3.98945689e-01 -1.07565987e+00
-9.40849960e-01 -6.32927656e-01 3.71536613e-01 -8.02932829e-02
-2.00317845e-01 -4.64570671e-01 -1.49100721e+00 -1.33631244e-01
-7.99245059e-01 1.65620282e-01 7.66078889e-01 1.20348275e+00
3.89066637e-01 5.17005384e-01 2.65085340e-01 -3.37538362e-01
-6.05775774e-01 -1.51454985e+00 -3.60787958e-01 4.28078681e-01
2.09274337e-01 -5.30951142e-01 -1.81609854e-01 4.42717373e-02]
|
[10.995911598205566, 10.753350257873535]
|
f975a332-2999-40c0-8373-2df47ea1e5b2
|
context-aware-group-activity-recognition
| null | null |
https://lear.inrialpes.fr/people/alahari/papers/dasgupta20.pdf
|
https://lear.inrialpes.fr/people/alahari/papers/dasgupta20.pdf
|
Context Aware Group Activity Recognition
|
This paper addresses the task of group activity recognition in multi-person videos. Existing approaches decompose this task into feature learning and relational reasoning. Despite showing progress, these methods only rely on appearance features for people and overlook the available contextual information, which can play an important role in group activity understanding. In this work, we focus on the feature learning aspect and propose a two-stream architecture that not only considers person-level appearance features, but also makes use of contextual information present in videos for group activity recognition. In particular, we propose to use two types of contextual information beneficial for two different scenarios: pose context and scene context that provide crucial cues for group activity understanding. We combine appearance and contextual features to encode each person with an enriched representation. Finally, these combined features are used in relational reasoning for predicting group activities. We evaluate our method on two benchmarks, Volleyball and Collective Activity and show that joint modeling of contextual information with appearance features benefits in group activity understanding.
|
['Karteek Alahari', 'C. V. Jawahar', 'Avijit Dasgupta']
|
2021-01-01
| null | null | null |
icpr-2021-1
|
['group-activity-recognition', 'relational-reasoning']
|
['computer-vision', 'natural-language-processing']
|
[ 3.18678141e-01 -2.25091517e-01 -2.88849354e-01 -5.28672814e-01
-3.78574610e-01 -3.66378278e-01 8.56099069e-01 4.87921268e-01
-2.59316087e-01 4.28585231e-01 7.30958521e-01 3.07511777e-01
-2.72574514e-01 -7.13146210e-01 -5.28133333e-01 -5.78960598e-01
6.33628219e-02 1.32812783e-01 1.30912319e-01 -2.34722078e-01
2.31683955e-01 3.71540517e-01 -1.96406054e+00 6.13703549e-01
6.27297819e-01 9.49817479e-01 -1.19008176e-01 6.11130059e-01
1.50004551e-02 1.36068499e+00 -4.52385366e-01 -2.45667219e-01
1.97501630e-01 -6.69889510e-01 -6.92209005e-01 6.67330801e-01
7.95030832e-01 -3.38448524e-01 -3.75749081e-01 5.59303999e-01
2.36925438e-01 4.66904163e-01 6.07313871e-01 -1.33443677e+00
-1.54024735e-01 3.34946543e-01 -4.02725756e-01 3.08349192e-01
9.36931610e-01 2.56908804e-01 1.10359716e+00 -7.60544538e-01
6.46862268e-01 1.15741622e+00 5.55684447e-01 2.98998505e-01
-9.82637107e-01 -2.99489111e-01 7.60716498e-01 7.55044401e-01
-1.09442639e+00 -5.13973117e-01 9.45602477e-01 -7.13611662e-01
7.72850871e-01 2.69950241e-01 1.21195149e+00 1.07268047e+00
-1.99384153e-01 1.21001399e+00 1.07782125e+00 -3.95158589e-01
-1.57760028e-02 -2.07726735e-05 3.13058913e-01 8.53875637e-01
6.51991367e-02 -1.16508357e-01 -9.76450801e-01 9.43510458e-02
6.79898560e-01 6.90295756e-01 -2.70913154e-01 -4.15167481e-01
-1.39345634e+00 6.47964180e-01 3.57957631e-01 4.21146035e-01
-4.59029675e-01 3.56969118e-01 2.72555172e-01 2.85397559e-01
3.80895704e-01 1.90332592e-01 5.83594739e-02 -3.45455199e-01
-6.40685737e-01 1.34448081e-01 6.47736311e-01 5.66328287e-01
9.29031432e-01 -4.41517860e-01 -5.30855417e-01 8.87292862e-01
3.18774492e-01 2.41406694e-01 1.05970681e-01 -1.03044808e+00
4.98263210e-01 1.13946748e+00 -1.11772297e-02 -1.27346647e+00
-4.09328908e-01 -3.09049904e-01 -4.63877559e-01 2.50847302e-02
6.20970666e-01 1.33304760e-01 -5.99792957e-01 1.58759797e+00
6.51722550e-01 3.82140368e-01 -2.02266127e-01 9.58295941e-01
9.36301231e-01 2.67092466e-01 1.27697945e-01 -1.77159123e-02
1.49328208e+00 -1.53185058e+00 -7.19579577e-01 -2.66818225e-01
6.26275778e-01 -5.61243355e-01 6.96587503e-01 4.32489932e-01
-1.03885043e+00 -7.94353724e-01 -8.71870339e-01 -3.72992083e-02
-4.38701600e-01 2.78379500e-01 8.39415193e-01 4.75376040e-01
-7.24152863e-01 6.58744633e-01 -7.09311724e-01 -6.32941425e-01
5.21452725e-01 2.76484907e-01 -7.01870978e-01 -3.31087291e-01
-8.23112965e-01 6.56596124e-01 -1.81975719e-02 1.56171322e-01
-7.41794825e-01 -4.41890121e-01 -8.74083221e-01 -7.31295645e-02
6.92464292e-01 -9.19686019e-01 8.47764194e-01 -1.17851782e+00
-1.20549822e+00 6.93491459e-01 -2.02184558e-01 -3.31980705e-01
6.83096349e-01 -4.66986269e-01 -2.60597140e-01 5.84086835e-01
-1.42816365e-01 3.41164440e-01 8.38759542e-01 -1.17175746e+00
-8.21867645e-01 -5.02320826e-01 5.83201826e-01 6.01867080e-01
-5.51235974e-01 2.66028419e-02 -7.30770528e-01 -5.24938583e-01
1.15386602e-02 -8.03707242e-01 -1.00446463e-01 1.68045253e-01
-7.00755939e-02 -3.41860473e-01 5.92605948e-01 -7.81177342e-01
1.13350034e+00 -1.91663182e+00 4.07301128e-01 2.50373811e-01
3.09352428e-01 -5.70655949e-02 -4.86214794e-02 6.38022065e-01
1.92716643e-01 -1.93303406e-01 2.13760361e-01 -6.34901404e-01
-6.50795549e-02 2.03568235e-01 1.86543927e-01 3.88624072e-01
1.54088348e-01 9.64049995e-01 -1.00168610e+00 -6.95149660e-01
4.16116774e-01 5.92313349e-01 -6.53311610e-01 1.93867683e-01
5.48863634e-02 7.17017174e-01 -6.15877628e-01 7.31053770e-01
2.93029368e-01 -1.38796970e-01 2.75990486e-01 -3.25665325e-01
1.17764607e-01 -1.80797443e-01 -1.29484785e+00 1.91976452e+00
-4.51006770e-01 3.53680581e-01 -7.93649107e-02 -9.60605860e-01
7.68015623e-01 3.05168722e-02 8.65992188e-01 -5.68743348e-01
4.05752026e-02 -2.42681026e-01 -1.71359435e-01 -6.34716272e-01
3.60509962e-01 2.08817944e-01 -4.43326943e-02 5.95086336e-01
1.55943766e-01 3.84630710e-01 2.98742622e-01 3.25137466e-01
1.23956203e+00 4.85445291e-01 4.29600120e-01 2.03101575e-01
7.97575653e-01 -2.60284990e-01 5.86517334e-01 7.60146856e-01
-3.07099938e-01 5.37530124e-01 4.42572773e-01 -7.29058146e-01
-6.27741575e-01 -7.56264091e-01 6.12883925e-01 1.45518363e+00
4.59961295e-01 -8.40900481e-01 -4.86869007e-01 -1.01168060e+00
-4.32913564e-02 1.23971559e-01 -1.01156032e+00 -1.91436380e-01
-6.73928022e-01 -4.12672162e-01 8.49647522e-02 8.40270400e-01
4.87978101e-01 -1.02574646e+00 -4.90261048e-01 -2.76603941e-02
-3.57957006e-01 -1.13842714e+00 -4.60937202e-01 -2.04212174e-01
-7.24272370e-01 -1.43436360e+00 -2.40642115e-01 -3.61855984e-01
8.02964568e-01 5.18005431e-01 1.09765041e+00 5.12675345e-01
-5.29452860e-01 1.26143682e+00 -6.28915787e-01 -1.58539906e-01
-1.15138479e-02 -1.03598021e-01 -1.24392293e-01 6.90551221e-01
3.63633752e-01 -6.80795729e-01 -7.95713842e-01 5.13668358e-01
-5.91391921e-01 3.61025810e-01 6.46297991e-01 6.96156442e-01
4.84163463e-01 -2.87037790e-01 2.29375422e-01 -7.73505867e-01
1.64039567e-01 -3.15564901e-01 1.06688619e-01 5.72911322e-01
-9.34259668e-02 -1.72352344e-01 2.40657359e-01 -3.53981137e-01
-1.25685084e+00 3.01679939e-01 2.13669866e-01 -4.74216372e-01
-3.10623020e-01 1.76709771e-01 -3.83703709e-01 -1.05043791e-01
3.20461571e-01 1.74983934e-01 -1.26521289e-01 -4.79477435e-01
2.16983661e-01 1.86908543e-01 2.72358358e-01 -8.24264467e-01
7.23460495e-01 7.51342297e-01 1.76581487e-01 -7.57067978e-01
-9.66717482e-01 -9.75590289e-01 -1.04201424e+00 -7.20417440e-01
9.92731392e-01 -1.17159319e+00 -8.55877817e-01 3.74644518e-01
-7.84924209e-01 -1.66181341e-01 -4.37111676e-01 3.89411271e-01
-6.67465210e-01 6.45607650e-01 -3.25055778e-01 -8.82527471e-01
-1.13005808e-03 -8.40241551e-01 1.22220695e+00 2.29988053e-01
-1.38652269e-02 -1.12701511e+00 4.91044745e-02 1.03973043e+00
1.37439206e-01 5.72062552e-01 2.28164539e-01 -7.19347775e-01
-7.79221833e-01 -3.61238241e-01 -1.33875191e-01 1.35881782e-01
3.70771065e-02 -1.32009879e-01 -9.81116772e-01 -6.25128001e-02
-3.48953247e-01 -3.15963268e-01 1.17521834e+00 1.10968091e-01
1.09714675e+00 -1.03567638e-01 -4.05602634e-01 4.50807512e-01
1.15105665e+00 -1.39475435e-01 5.70464075e-01 1.23169564e-01
1.09417272e+00 8.60504329e-01 1.00712061e+00 6.37241364e-01
6.38231635e-01 9.44825828e-01 5.13032615e-01 -1.83963496e-02
-2.93511689e-01 -3.69557261e-01 4.79118645e-01 6.25518203e-01
-9.43861306e-01 1.69413224e-01 -7.18031883e-01 3.80039632e-01
-2.49958396e+00 -1.36772907e+00 -7.13581890e-02 1.92141533e+00
3.85520816e-01 -1.90904006e-01 4.49500084e-01 1.67239934e-01
6.43039465e-01 3.09019208e-01 -2.14537412e-01 -1.09336404e-02
-1.96729526e-02 -1.94366604e-01 -7.03661470e-03 1.66060507e-01
-1.27425039e+00 6.29603684e-01 5.27838087e+00 6.75348639e-01
-4.03316051e-01 1.15235247e-01 3.66337389e-01 -2.28941202e-01
-1.19352929e-01 2.77265124e-02 -7.34890997e-01 1.35161772e-01
4.63602394e-01 2.65407801e-01 2.56875008e-01 8.67647946e-01
1.88908558e-02 -3.42988640e-01 -1.38765013e+00 1.09960246e+00
4.63057160e-01 -1.11496341e+00 -2.31443103e-02 2.46328264e-01
6.29735649e-01 -5.52377522e-01 -2.75818139e-01 4.47051376e-01
5.70922419e-02 -9.01504338e-01 7.11827159e-01 1.03433478e+00
7.59698972e-02 -7.18735516e-01 6.78696215e-01 2.26453945e-01
-1.54101908e+00 -3.70939702e-01 -3.32718119e-02 -2.82927066e-01
1.93702579e-01 4.11851138e-01 -5.61868429e-01 8.66620898e-01
6.08799100e-01 1.39981616e+00 -9.93993759e-01 1.05111217e+00
-2.55595148e-01 2.29547381e-01 -1.52799785e-01 2.31465176e-01
-6.11730404e-02 -2.53422916e-01 3.33807468e-01 1.16310716e+00
1.29642680e-01 4.17739414e-02 7.80878067e-01 3.61726254e-01
1.19308755e-01 2.05063403e-01 -5.86928844e-01 -9.48108640e-03
6.91622496e-02 1.43097174e+00 -7.36244500e-01 -3.92573774e-01
-7.60273397e-01 1.02255189e+00 4.37067091e-01 1.09587580e-01
-7.93728352e-01 1.49481058e-01 7.74725258e-01 2.28692800e-01
2.84019053e-01 -2.68161595e-01 1.95161775e-01 -1.50002885e+00
8.39127414e-03 -8.47262323e-01 7.26804912e-01 -5.58742404e-01
-1.16231203e+00 1.11346446e-01 5.29570058e-02 -1.29644561e+00
-2.17006862e-01 -4.96481687e-01 -7.51534104e-01 1.70606256e-01
-1.42674685e+00 -1.83031511e+00 -9.54921484e-01 8.61009598e-01
7.85313964e-01 -4.98765074e-02 5.01737177e-01 2.90137917e-01
-6.20688140e-01 3.22139055e-01 -4.13805902e-01 2.22819105e-01
7.70288765e-01 -1.24785745e+00 -5.26528619e-02 7.18730688e-01
4.95670468e-01 6.90910518e-01 4.22292173e-01 -5.34413159e-01
-1.53391111e+00 -9.21958029e-01 6.97897375e-01 -8.18182766e-01
4.64109629e-01 -4.14767593e-01 -6.41806185e-01 6.72587693e-01
-2.42958143e-02 2.52521843e-01 1.06694710e+00 5.39430559e-01
-5.15926242e-01 -1.15634069e-01 -8.56016338e-01 5.50713122e-01
1.67855847e+00 -6.11328363e-01 -5.67310572e-01 1.38550177e-01
3.20802808e-01 -1.00193350e-02 -9.79978919e-01 2.12155104e-01
8.76435280e-01 -1.14862037e+00 1.21240914e+00 -6.11580670e-01
4.28928852e-01 -4.18223768e-01 -1.52648613e-01 -1.03987896e+00
-2.23038867e-01 -3.20177138e-01 -3.26589823e-01 1.32935965e+00
-1.18317686e-01 -1.82862684e-01 7.10528970e-01 5.26489317e-01
2.62017958e-02 -6.03802860e-01 -4.51261580e-01 -4.41715926e-01
-6.65920675e-01 -4.12366956e-01 3.95196915e-01 9.21633959e-01
2.28392601e-01 2.88762718e-01 -6.95427060e-01 -2.07134113e-01
5.47544360e-01 3.98397774e-01 1.23644769e+00 -1.34909284e+00
-4.71008390e-01 -2.38462046e-01 -7.30805457e-01 -7.75481701e-01
5.84599078e-02 -6.86659813e-01 -2.19419718e-01 -1.78558779e+00
6.56736135e-01 -3.27039659e-02 -4.85834420e-01 6.24505579e-01
-3.52427810e-01 4.76339519e-01 5.20486951e-01 1.62295595e-01
-1.35022283e+00 4.99139428e-01 1.17551494e+00 -2.91319668e-01
-1.08313039e-01 -1.56659111e-02 -5.51304936e-01 8.83376718e-01
6.27938688e-01 2.92679351e-02 -4.21033680e-01 -1.93571746e-01
2.34993502e-01 -1.24849100e-02 7.38963485e-01 -1.38137460e+00
2.69923359e-01 -3.76853526e-01 6.28982663e-01 -5.15371978e-01
6.29702985e-01 -9.23499763e-01 1.03470720e-01 3.07267249e-01
-3.28873158e-01 -3.00085098e-01 -2.82769591e-01 1.06500053e+00
-3.87373924e-01 2.57365406e-02 2.89342642e-01 -2.63744056e-01
-1.11550164e+00 3.85771841e-01 -1.77306727e-01 -3.27652127e-01
1.20091999e+00 -5.80397725e-01 -2.74362952e-01 -6.50525689e-01
-1.27397668e+00 3.18991572e-01 5.19413710e-01 6.61056459e-01
5.27761161e-01 -1.51639247e+00 -6.19642079e-01 9.24755409e-02
5.81862628e-01 -3.83354068e-01 7.40149021e-01 1.27413678e+00
-1.60570040e-01 2.86935568e-02 -4.38930601e-01 -7.86082208e-01
-1.70148933e+00 5.09457290e-01 1.31816939e-01 -4.50487405e-01
-5.40068388e-01 6.86169147e-01 4.64304030e-01 -9.13980976e-02
2.26889715e-01 -2.48056114e-01 -7.35809088e-01 4.81681019e-01
6.74294889e-01 4.78261352e-01 -2.04824880e-01 -9.11877155e-01
-3.91222954e-01 8.61859500e-01 1.03710271e-01 1.51704162e-01
1.47348869e+00 -2.70232618e-01 9.34988409e-02 4.42347348e-01
9.40807700e-01 1.18552744e-01 -1.46626341e+00 -5.36616147e-01
2.27458440e-02 -7.73137450e-01 -1.98936701e-01 -7.05581307e-01
-1.10630262e+00 1.10347617e+00 4.62695092e-01 -3.33109461e-02
1.13680744e+00 1.13558687e-01 4.69418377e-01 3.71884882e-01
5.89640021e-01 -1.31370533e+00 7.39359260e-01 1.48250118e-01
8.39940786e-01 -1.35987103e+00 4.40938026e-01 -6.55649722e-01
-8.22829843e-01 1.13538373e+00 6.45925164e-01 -4.50524166e-02
3.12942982e-01 -2.68517226e-01 -2.92830020e-01 -2.88198024e-01
-8.75962913e-01 -7.08139658e-01 6.57596469e-01 7.24129856e-01
2.60608733e-01 -4.56486903e-02 -1.78726733e-01 7.96344101e-01
2.62345135e-01 -4.77749109e-02 2.66455203e-01 1.09537876e+00
-5.55178344e-01 -1.17854345e+00 -4.29820478e-01 3.21332932e-01
-2.57636458e-01 3.88791084e-01 -6.15868568e-01 7.71278739e-01
5.52156985e-01 8.39694023e-01 -1.66474446e-03 -4.22232956e-01
3.32323074e-01 1.50925592e-01 8.55253100e-01 -6.78480148e-01
-9.45778251e-01 -1.21295629e-02 4.36780334e-01 -9.20929253e-01
-9.37386155e-01 -9.34849620e-01 -8.40818286e-01 -1.14133097e-01
-4.42108028e-02 -1.25312567e-01 4.64644641e-01 1.09348679e+00
1.83503762e-01 5.73745012e-01 4.49250460e-01 -9.37493086e-01
6.96927086e-02 -7.70907342e-01 -4.76073176e-01 9.18865681e-01
1.99993148e-01 -9.84711289e-01 -2.29615033e-01 3.53584945e-01]
|
[8.193720817565918, 0.5783317685127258]
|
74b80057-f347-4cc2-86fd-3eeb6ed1a9d4
|
decomposing-the-generalization-gap-in
|
2307.03659
| null |
https://arxiv.org/abs/2307.03659v1
|
https://arxiv.org/pdf/2307.03659v1.pdf
|
Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulation
|
What makes generalization hard for imitation learning in visual robotic manipulation? This question is difficult to approach at face value, but the environment from the perspective of a robot can often be decomposed into enumerable factors of variation, such as the lighting conditions or the placement of the camera. Empirically, generalization to some of these factors have presented a greater obstacle than others, but existing work sheds little light on precisely how much each factor contributes to the generalization gap. Towards an answer to this question, we study imitation learning policies in simulation and on a real robot language-conditioned manipulation task to quantify the difficulty of generalization to different (sets of) factors. We also design a new simulated benchmark of 19 tasks with 11 factors of variation to facilitate more controlled evaluations of generalization. From our study, we determine an ordering of factors based on generalization difficulty, that is consistent across simulation and our real robot setup.
|
['Chelsea Finn', 'Ted Xiao', 'Lisa Lee', 'Annie Xie']
|
2023-07-07
| null | null | null | null |
['imitation-learning']
|
['methodology']
|
[ 1.0246364e-01 -2.5384340e-01 -1.6728611e-01 -8.2401820e-02
-4.8566356e-01 -9.5145363e-01 7.2100419e-01 -2.2023262e-01
-5.6264412e-01 6.7850679e-01 1.5601376e-01 -5.6732118e-01
-4.3317404e-01 -2.0380294e-01 -8.4483773e-01 -5.4481304e-01
-2.1528545e-01 3.3835900e-01 2.5423676e-01 -1.5801759e-01
3.3006826e-01 6.8290329e-01 -1.5616077e+00 -9.8377556e-02
5.9879613e-01 5.1477587e-01 5.4989141e-01 5.2846342e-01
3.0358461e-01 4.6894810e-01 -6.8913239e-01 3.5264847e-01
7.6689589e-01 -4.9111524e-01 -6.2430269e-01 4.1384381e-01
2.2181594e-01 -4.4145510e-01 -2.8697950e-01 1.0141352e+00
2.4731292e-01 3.3930820e-01 8.7655705e-01 -1.4889737e+00
-4.3340468e-01 4.5084023e-01 -2.1403751e-01 2.3432358e-01
4.3219250e-01 7.8005588e-01 6.6766888e-01 -3.5866407e-01
5.9498632e-01 1.5246338e+00 3.6800158e-01 3.6034960e-01
-1.2994790e+00 -5.4862928e-01 2.5260338e-01 -3.1353690e-02
-1.2389059e+00 -2.7895695e-01 3.0371982e-01 -6.2737775e-01
7.9940248e-01 -2.1334545e-01 4.2426214e-01 1.2917534e+00
4.3589792e-01 3.1187248e-01 1.4180194e+00 -3.9129013e-01
5.1571810e-01 1.1115476e-01 -2.0176062e-01 2.4613000e-01
3.3274770e-01 4.4552219e-01 -7.7659994e-02 -1.5398145e-01
1.1110901e+00 -2.7927095e-01 -6.0818863e-01 -9.6287137e-01
-1.2374029e+00 5.9851760e-01 3.3204302e-01 8.2417548e-02
-2.3617835e-01 3.7872928e-01 2.8472582e-01 8.3469236e-01
-3.8986024e-01 1.0519924e+00 -5.6670272e-01 -3.9521611e-01
-3.6670204e-02 5.7555145e-01 1.0586753e+00 1.1888248e+00
7.4775016e-01 5.1459324e-02 1.9604494e-01 6.9809794e-01
4.8079178e-02 3.0622318e-01 2.3124298e-01 -1.4792010e+00
4.8624727e-01 3.4442478e-01 6.3309610e-01 -7.6769918e-01
-6.3246918e-01 -1.1253681e-01 1.3890019e-01 9.9866617e-01
9.9031502e-01 -4.2471132e-01 -8.7001830e-01 2.0532110e+00
3.3475775e-02 -3.5494512e-01 -2.1374931e-03 1.2339711e+00
1.9725522e-02 2.8901118e-01 1.3198167e-01 -8.8177174e-02
1.0201167e+00 -8.6297244e-01 -3.2941931e-01 -5.4731417e-01
7.4162704e-01 -6.6511494e-01 1.4129092e+00 4.0984946e-01
-7.5425321e-01 -3.8503492e-01 -1.2454972e+00 2.6482067e-01
-9.2764847e-02 -5.9408192e-02 4.4873822e-01 4.1167548e-01
-1.0346850e+00 7.5807285e-01 -9.1668022e-01 -8.9680415e-01
-3.0073056e-01 4.4157001e-01 -3.7865371e-01 -8.9014702e-02
-8.4014910e-01 1.5732399e+00 4.0178192e-01 -1.6804484e-01
-1.2522303e+00 -3.1828785e-01 -6.6261268e-01 -4.8718184e-02
7.3858529e-01 -6.8387103e-01 1.5424441e+00 -9.2941558e-01
-1.5373951e+00 4.2548341e-01 1.2316080e-01 -2.1871986e-02
5.3475177e-01 1.2332434e-04 1.7722939e-01 -1.2463440e-01
7.5219661e-02 6.4140058e-01 9.5598668e-01 -1.5977803e+00
-3.5020581e-01 -4.4696522e-01 6.3006765e-01 5.2026242e-01
4.8302030e-03 3.3500943e-02 -9.8890029e-02 -4.2702276e-01
-1.8334650e-01 -1.6183770e+00 -2.9700604e-01 2.1266738e-01
3.5533798e-01 -1.0649788e-01 5.7610863e-01 -1.7884801e-01
4.6587956e-01 -2.2001023e+00 5.0618726e-01 -1.9295691e-02
-2.4064628e-02 -1.0783543e-01 -2.9156393e-01 7.0818120e-01
-1.3158049e-01 -5.9341021e-02 -7.0639469e-02 3.7169661e-02
1.0891759e-01 5.1628977e-01 -3.0711025e-01 4.6316838e-01
7.7039912e-02 5.6412888e-01 -1.0839058e+00 -3.0468617e-02
8.3892159e-02 2.4407916e-01 -5.2992350e-01 -1.0625004e-01
-2.1704245e-01 3.7121436e-01 -5.2494770e-01 1.8038143e-01
2.0893042e-01 9.0524681e-02 1.9714752e-01 2.5470281e-01
-4.3492496e-02 1.0422944e-01 -1.2070129e+00 1.3883835e+00
-4.3412900e-01 5.2517569e-01 2.4698144e-01 -6.1217606e-01
6.6829401e-01 7.5352341e-02 1.8732986e-01 -3.1601942e-01
2.4313247e-01 2.7237403e-01 6.7372119e-01 -7.3067784e-01
5.8467561e-01 -1.6601022e-01 5.8820039e-02 5.1022494e-01
1.0822249e-02 -6.0181910e-01 5.7347488e-02 -2.1779040e-01
1.3082211e+00 2.7495369e-01 2.0786192e-01 -4.0342087e-01
-9.8166734e-02 3.2574710e-01 3.2229692e-01 9.7338784e-01
-6.2530261e-01 3.6172241e-01 7.1562594e-01 -5.4138534e-02
-1.2798586e+00 -1.1968143e+00 -4.3498926e-02 8.7151980e-01
3.6789468e-01 -2.3897885e-01 -7.2551703e-01 -3.5471940e-01
2.3623581e-01 9.4531387e-01 -5.0515246e-01 -2.1451584e-01
-3.6938277e-01 -4.0560073e-01 2.7740163e-01 6.3315964e-01
-2.6321512e-02 -1.0644156e+00 -1.0786555e+00 -1.1614624e-01
1.9693632e-01 -1.2744050e+00 -3.6549932e-01 4.7639614e-01
-7.5143337e-01 -1.1143237e+00 -4.0893453e-01 -7.4630505e-01
7.3497266e-01 5.3107983e-01 7.4322337e-01 -1.3187651e-01
-1.1846310e-01 8.3089393e-01 -3.9868474e-01 -4.7407678e-01
-8.5650331e-01 -1.1183043e-01 3.2723880e-01 -8.8560051e-01
1.1942583e-01 -6.1439192e-01 -4.2839170e-01 5.9588677e-01
-7.9693699e-01 -4.3411434e-01 6.6953462e-01 8.2603955e-01
-3.1235497e-03 3.6310461e-01 4.0398043e-01 -2.1146043e-01
1.2242373e+00 -5.3905082e-01 -7.2471803e-01 -2.2041446e-02
-4.9618846e-01 1.4556980e-01 6.1315596e-01 -7.8515333e-01
-7.4698216e-01 8.6549371e-02 5.3902745e-01 -6.1712205e-01
-2.7480602e-01 4.4655067e-01 7.3445208e-02 -1.6256857e-01
8.1896001e-01 -1.1081941e-01 3.0024174e-01 -2.1459049e-01
3.3598930e-01 5.1909465e-01 3.0603382e-01 -1.0617043e+00
6.9745129e-01 1.3260092e-01 -5.9123483e-02 -8.7924755e-01
-2.1388019e-02 2.3809357e-02 -4.4362262e-01 -9.6741050e-02
6.1566836e-01 -7.7828705e-01 -6.8620324e-01 4.4393286e-01
-9.3327832e-01 -1.0238333e+00 -8.1950746e-02 6.7918581e-01
-1.1211976e+00 3.5326344e-01 -4.6113867e-01 -6.9722378e-01
5.4228163e-01 -1.8454089e+00 8.3418941e-01 2.6911821e-02
-5.6673771e-01 -6.2440360e-01 -2.3138867e-01 -6.3218653e-02
4.6471623e-01 4.4464193e-02 1.1597592e+00 -3.8661718e-01
-5.6866550e-01 4.6728615e-02 -8.2808726e-02 3.5787401e-01
2.9284507e-01 -2.1329391e-01 -5.5092299e-01 -5.8797705e-01
3.2702875e-01 -5.5796975e-01 3.1658396e-01 1.6103604e-01
8.3739877e-01 9.6314013e-02 -3.1435496e-01 2.3484090e-01
1.3500113e+00 5.1931906e-01 4.3158644e-01 6.7688161e-01
4.0967542e-01 5.9777492e-01 5.8525556e-01 2.5376871e-01
2.4391012e-01 7.7161002e-01 3.9283922e-01 4.4615558e-01
1.1457667e-01 -1.0565622e-01 4.8164287e-01 6.0198262e-02
3.3809040e-02 -2.5662407e-01 -9.0592974e-01 3.9035517e-01
-1.5493017e+00 -6.7975789e-01 4.7950557e-01 2.2652507e+00
3.0846858e-01 3.2358462e-01 2.1768534e-01 -2.8321264e-02
5.2807033e-01 -7.2959572e-02 -7.0682538e-01 -5.1025295e-01
2.8649902e-01 -4.6208873e-01 5.7936537e-01 7.1820134e-01
-5.6841767e-01 8.2899588e-01 7.0920582e+00 4.8376924e-01
-1.2370485e+00 -3.9953706e-01 5.4396655e-02 -2.5959091e-02
-3.7533537e-02 2.3325369e-01 -4.1012907e-01 3.8188073e-01
5.4411924e-01 -1.6547032e-01 1.0406827e+00 9.2712677e-01
3.8855934e-01 -4.6385312e-01 -1.5468884e+00 6.0848993e-01
-5.0112858e-02 -3.0173346e-01 -2.0111535e-01 1.9700286e-01
5.1201844e-01 3.0889589e-01 1.3752805e-01 5.2421170e-01
5.7355934e-01 -9.5785540e-01 6.6755432e-01 2.3797411e-02
3.2943863e-01 -4.5117170e-01 2.8697681e-01 6.4627254e-01
-5.7630467e-01 -4.5838648e-01 -2.3670895e-01 -5.9247023e-01
-1.9887766e-02 -3.9347455e-01 -1.0656739e+00 4.8670791e-02
5.7566136e-01 6.1651710e-02 -3.3515635e-01 8.8132888e-01
-1.8681084e-01 2.9777077e-01 -4.2783734e-01 -2.6172724e-01
4.1401127e-01 -1.1803151e-01 7.0038658e-01 7.6156086e-01
2.5078169e-01 1.6297039e-01 2.9890957e-01 8.5788643e-01
3.0678010e-01 -3.6280832e-01 -9.5249587e-01 -9.8136552e-02
7.1466833e-01 8.5468018e-01 -8.5602742e-01 2.0764418e-01
-3.9660969e-01 6.6361064e-01 5.1704431e-01 7.6913410e-01
-7.8817803e-01 -1.4019209e-01 9.1565555e-01 5.6922957e-03
4.0873155e-01 -9.1167337e-01 -1.3784055e-01 -9.2640984e-01
2.6647580e-01 -1.2390367e+00 -1.0393888e-01 -9.2627031e-01
-1.1143689e+00 2.9801673e-01 4.6121058e-01 -1.2478726e+00
-4.7757670e-01 -9.5637572e-01 -3.4609994e-01 8.3443403e-01
-1.0651042e+00 -4.4670334e-01 -3.2873854e-01 3.2795286e-01
5.6715274e-01 -9.4358698e-03 6.8811804e-01 -1.4553903e-01
-3.6269248e-01 4.1836488e-01 4.3572973e-02 -3.5506853e-01
8.9655751e-01 -1.0403976e+00 3.4462836e-01 4.4118086e-01
-3.2533142e-01 9.3435258e-01 1.1921537e+00 -4.6016416e-01
-1.6836565e+00 -5.1490170e-01 8.8320605e-02 -8.8265628e-01
8.9995307e-01 -2.0530325e-01 -8.3886415e-01 9.5613080e-01
8.8948183e-02 -2.7085853e-01 -4.7060832e-02 9.7883083e-02
-3.2249647e-01 2.8807139e-01 -9.9422240e-01 1.0505053e+00
1.2499689e+00 -4.4950321e-01 -6.8062645e-01 6.8682820e-02
8.1582695e-01 -5.1554418e-01 -8.4176195e-01 4.3766943e-01
4.6687347e-01 -7.8940946e-01 7.0159101e-01 -4.1495880e-01
5.0822961e-01 -4.5281845e-01 -2.5082257e-01 -1.8901933e+00
-3.6940131e-01 -4.5974731e-01 5.4627967e-01 8.0124593e-01
3.7665093e-01 -8.1603545e-01 4.7164735e-01 7.2474438e-01
-2.3956765e-01 -8.8294959e-01 -7.0440906e-01 -1.2182316e+00
5.3168052e-01 -4.4814393e-01 3.8611731e-01 7.0016611e-01
3.0869305e-01 3.1075501e-01 1.4548282e-02 2.0562284e-01
1.5122478e-01 -1.8512870e-01 1.0470034e+00 -6.6323292e-01
-5.1690608e-01 -7.1428442e-01 -6.4458430e-01 -1.1374394e+00
2.8399438e-01 -5.8003443e-01 3.8617104e-01 -1.2221825e+00
1.3669224e-01 -5.4671395e-01 8.2674205e-02 3.4474269e-01
-7.0458114e-02 -4.2101657e-01 5.0634176e-01 1.4204738e-01
-2.8584048e-01 5.4746044e-01 1.3626621e+00 1.0146671e-01
-2.0351589e-01 -2.8727543e-01 -5.6733668e-01 7.1689403e-01
8.0633509e-01 -1.8809901e-01 -8.3010799e-01 -8.4993225e-01
-1.2799822e-01 1.8042435e-01 2.8107333e-01 -9.1842252e-01
4.4447385e-02 -4.5484498e-01 1.4256743e-01 1.6874848e-01
3.4649077e-01 -9.6177340e-01 -3.6891989e-02 5.0380701e-01
-4.0458202e-01 3.5641721e-01 5.3186333e-01 5.9189630e-01
2.5435320e-01 -2.7800119e-01 6.3337892e-01 -2.0087588e-01
-7.6640677e-01 -3.0980486e-01 -6.6909695e-01 2.4589176e-01
1.0536706e+00 -2.8077263e-01 -3.1998006e-01 -6.1426342e-01
-5.0473881e-01 5.1535308e-01 1.0687433e+00 8.4466726e-01
1.5008394e-01 -8.9797950e-01 -3.6084262e-01 1.0920822e-03
1.3905616e-01 -3.7000543e-01 -2.4886765e-01 6.2318856e-01
-4.3340561e-01 2.8334466e-01 -3.9262298e-01 -5.9133643e-01
-8.6691743e-01 8.2660049e-01 4.4366822e-01 2.6618880e-01
-4.9509311e-01 4.5211452e-01 3.8003141e-01 -5.2524453e-01
4.5526037e-01 -6.5108997e-01 1.9851971e-01 -4.4471285e-01
8.5783228e-02 3.7556249e-01 -2.1493134e-01 -3.9523485e-01
-1.9894187e-01 5.6436336e-01 4.1401118e-02 -2.2947058e-01
7.8886610e-01 -1.8266582e-01 2.3563871e-01 5.8989686e-01
9.4257879e-01 -3.7349391e-01 -1.6395223e+00 1.1305182e-01
-7.7816509e-02 -5.1514488e-01 -4.1786894e-01 -7.2245789e-01
-4.9013984e-01 4.5838946e-01 4.0621215e-01 8.6292028e-02
8.2395673e-01 -3.8606867e-02 2.8233200e-01 9.1801870e-01
9.0244699e-01 -9.7114420e-01 2.0831256e-01 6.8231928e-01
1.2381690e+00 -1.3286287e+00 -4.5135461e-02 -2.5423935e-01
-7.0614570e-01 9.1818649e-01 8.2631201e-01 -3.4762660e-01
4.3704659e-01 3.6709800e-01 3.0813351e-01 -2.6686979e-03
-6.1612856e-01 -4.9530915e-03 -2.6692447e-01 6.8879932e-01
5.3791303e-02 -5.2842284e-03 -3.3307898e-01 -3.3890024e-02
-4.6058905e-01 1.3597053e-01 7.7445233e-01 1.0806501e+00
-4.6304202e-01 -9.0851271e-01 -4.2579532e-01 9.0112709e-02
-3.8120847e-02 4.4222382e-01 -3.8896656e-01 1.2947030e+00
-1.1226349e-01 9.1677219e-01 -1.4955267e-01 -4.3253541e-01
5.7613522e-01 -8.7956898e-02 9.8308718e-01 -7.2345865e-01
-1.1275141e-01 -2.2593866e-01 1.6270129e-01 -5.3132176e-01
1.3192608e-02 -9.0969712e-01 -1.2716571e+00 -7.7505126e-03
-1.6664854e-01 -1.2579392e-01 7.5187111e-01 1.0531704e+00
-9.5324581e-03 3.7182635e-01 3.2201058e-01 -1.2384837e+00
-1.2585537e+00 -8.4517211e-01 -5.3311795e-01 5.8140516e-01
3.7706688e-01 -1.0345782e+00 -8.8896281e-01 -3.1032762e-01]
|
[4.503927230834961, 1.0684868097305298]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.