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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
46eb8c4e-7ae2-4ced-a979-261c02ce1ac6 | on-permutation-invariant-training-for-speech | 2102.04945 | null | https://arxiv.org/abs/2102.04945v2 | https://arxiv.org/pdf/2102.04945v2.pdf | On permutation invariant training for speech source separation | We study permutation invariant training (PIT), which targets at the permutation ambiguity problem for speaker independent source separation models. We extend two state-of-the-art PIT strategies. First, we look at the two-stage speaker separation and tracking algorithm based on frame level PIT (tPIT) and clustering, which was originally proposed for the STFT domain, and we adapt it to work with waveforms and over a learned latent space. Further, we propose an efficient clustering loss scalable to waveform models. Second, we extend a recently proposed auxiliary speaker-ID loss with a deep feature loss based on "problem agnostic speech features", to reduce the local permutation errors made by the utterance level PIT (uPIT). Our results show that the proposed extensions help reducing permutation ambiguity. However, we also note that the studied STFT-based models are more effective at reducing permutation errors than waveform-based models, a perspective overlooked in recent studies. | ['Jordi Pons', 'Xiaoyu Liu'] | 2021-02-09 | null | null | null | null | ['speaker-separation'] | ['speech'] | [ 5.00045240e-01 3.06223333e-02 7.70888925e-02 -4.11262989e-01
-1.35619032e+00 -6.03457570e-01 5.32787919e-01 -3.35339963e-01
-1.72234863e-01 4.26060289e-01 3.72399569e-01 -9.06505883e-02
-5.65747619e-01 2.94437017e-02 -5.49696505e-01 -9.90135968e-01
-3.68364990e-01 4.36554283e-01 1.98429704e-01 -1.21394731e-01
1.86599135e-01 2.74189323e-01 -1.52947354e+00 2.01286301e-01
9.02221620e-01 7.52440274e-01 7.03083128e-02 6.44692183e-01
5.86725911e-03 3.88335049e-01 -7.47944474e-01 -2.05749676e-01
2.89300412e-01 -5.47859907e-01 -5.80564797e-01 -6.83491230e-02
3.70168328e-01 3.57959062e-01 -1.92955360e-01 9.67037022e-01
8.41884017e-01 3.14876348e-01 5.65670490e-01 -1.45517778e+00
-3.32688510e-01 1.03741014e+00 -5.43630719e-01 3.71633232e-01
-4.93447520e-02 -3.54260683e-01 9.81666863e-01 -7.90198147e-01
8.29911157e-02 1.49109304e+00 1.01844347e+00 4.69311059e-01
-1.24783123e+00 -8.10813785e-01 3.58271331e-01 4.62798923e-01
-1.45342374e+00 -9.88929331e-01 1.07485080e+00 -3.07829112e-01
6.77293539e-01 4.49011803e-01 -5.84564246e-02 1.16535938e+00
-2.16042757e-01 7.01534510e-01 8.48757386e-01 -6.76924229e-01
2.24171907e-01 -1.01911753e-01 4.21570688e-01 4.00443465e-01
2.87756901e-02 -2.19949279e-02 -8.03340733e-01 -1.74445510e-01
2.64536679e-01 -4.08478260e-01 -4.64573681e-01 -4.89083588e-01
-1.02651429e+00 7.93999851e-01 -1.56823456e-01 5.23204207e-01
-5.25904410e-02 9.30037815e-03 4.02398080e-01 2.70239145e-01
6.31091952e-01 1.09577462e-01 -6.12417102e-01 -2.60896236e-01
-1.38696015e+00 -3.19783166e-02 8.12395692e-01 8.93182695e-01
5.45143068e-01 3.75572771e-01 -6.48378432e-01 1.05893219e+00
5.23662210e-01 5.16605020e-01 7.14129150e-01 -9.71051931e-01
4.36352164e-01 -8.26292410e-02 -1.53379649e-01 -6.73386335e-01
-3.16156209e-01 -8.15349460e-01 -5.76665163e-01 -1.11061558e-01
1.14333421e-01 -2.63558090e-01 -9.09974396e-01 2.23849320e+00
1.59080729e-01 6.55047476e-01 1.83369331e-02 5.77243924e-01
3.97039056e-01 4.27935213e-01 -2.31410846e-01 -5.73126197e-01
1.42531049e+00 -1.26525724e+00 -9.92847383e-01 -2.34490156e-01
2.50448436e-01 -8.51450205e-01 8.47161233e-01 6.76573277e-01
-1.07453322e+00 -6.68948412e-01 -1.02083004e+00 2.53741711e-01
-8.43227878e-02 4.13713992e-01 3.18725854e-01 1.22468579e+00
-1.29849839e+00 4.70276088e-01 -9.55442727e-01 -3.37481707e-01
1.57489732e-01 3.55560064e-01 -3.33607718e-02 2.17246711e-01
-1.06822824e+00 4.36468989e-01 1.33521631e-01 1.05939075e-01
-8.32653403e-01 -7.07062960e-01 -8.68038476e-01 3.86273295e-01
4.48735744e-01 -4.29374605e-01 1.25019228e+00 -8.53115559e-01
-2.04488063e+00 2.90866017e-01 -8.51149023e-01 -6.20091796e-01
2.05626503e-01 -2.91383654e-01 -6.30652785e-01 3.16129178e-01
5.73092774e-02 3.64558399e-01 1.39907932e+00 -1.31153250e+00
-3.63144547e-01 -2.56108999e-01 -5.48370659e-01 6.67656511e-02
-5.26049972e-01 3.58761847e-01 -3.98896307e-01 -8.81484032e-01
2.02236429e-01 -8.53912294e-01 2.84640025e-02 -5.09824932e-01
-4.53371763e-01 -1.91925734e-01 9.47044730e-01 -9.10482168e-01
1.26844585e+00 -2.55506468e+00 2.63371289e-01 1.82383984e-01
-1.67696163e-01 4.67733443e-01 -3.39960635e-01 4.70220804e-01
-2.67303556e-01 -2.99039092e-02 -6.88353837e-01 -1.02107549e+00
4.97300863e-01 1.51810154e-01 -4.65280831e-01 3.85979980e-01
1.75555840e-01 3.94398868e-01 -4.92547631e-01 -3.10772240e-01
4.50936407e-02 5.64350486e-01 -7.07820892e-01 -1.12322539e-01
9.65514779e-02 5.25584102e-01 1.25559747e-01 2.82805473e-01
1.05431855e+00 2.52911180e-01 1.14537880e-01 -1.82793096e-01
-2.62724966e-01 6.29882276e-01 -1.37506139e+00 1.81061840e+00
-2.05194518e-01 5.75828731e-01 4.18187380e-01 -1.25201511e+00
7.54462838e-01 6.29432380e-01 3.75943482e-01 -9.79648232e-02
-2.00824905e-02 1.38108864e-01 8.42845142e-02 -1.25711918e-01
2.97996372e-01 -1.31156266e-01 1.69653390e-02 2.18379602e-01
3.06428164e-01 3.09620708e-01 9.32483375e-02 2.90393848e-02
1.04090416e+00 2.16975734e-02 -2.36631613e-02 -4.20450330e-01
8.04057240e-01 -5.14874458e-01 8.70406091e-01 9.40999448e-01
-3.61046344e-01 7.82926083e-01 4.01049703e-01 3.90737593e-01
-5.82261980e-01 -1.24243462e+00 -2.42762223e-01 1.26786089e+00
-1.33160800e-01 -3.70482653e-01 -8.16372931e-01 -5.97872674e-01
-2.22211927e-01 8.02195370e-01 -2.61707038e-01 -1.21395648e-01
-8.24059606e-01 -8.36897612e-01 8.35355759e-01 4.48395491e-01
4.57798421e-01 -5.34325421e-01 -3.98528576e-02 3.14289659e-01
-3.11518759e-01 -1.16162133e+00 -7.51812994e-01 5.25614798e-01
-5.28307438e-01 -5.76589644e-01 -6.67714119e-01 -8.63145769e-01
1.97690368e-01 3.27462047e-01 5.65472603e-01 -6.04327500e-01
1.18808620e-01 5.75037479e-01 -5.76339364e-01 -3.92185539e-01
-3.60879391e-01 1.96094915e-01 3.91479045e-01 4.62501019e-01
2.96288520e-01 -7.64015138e-01 -3.34464401e-01 3.41107547e-01
-6.59473002e-01 -3.73622984e-01 4.41865146e-01 7.46417940e-01
2.47777686e-01 4.68296111e-01 9.53822732e-01 -4.81334180e-01
3.65707964e-01 -2.36362353e-01 -3.02698016e-01 1.82935983e-01
-4.33753848e-01 2.13228509e-01 5.59787989e-01 -5.14026105e-01
-1.32767212e+00 -1.29086465e-01 -3.74515027e-01 -6.24420345e-01
-2.65698463e-01 1.56699851e-01 -5.25483787e-01 -1.70182645e-01
2.41417170e-01 3.57652128e-01 2.16067280e-03 -8.45928788e-01
2.50329554e-01 7.33637929e-01 5.48680663e-01 -4.43941236e-01
9.65267777e-01 4.65654135e-01 -1.82223096e-01 -8.70312691e-01
-7.25893021e-01 -7.01944590e-01 -6.19187951e-01 4.39332575e-01
7.08585322e-01 -8.24808836e-01 -6.10486627e-01 3.65961134e-01
-1.15848804e+00 -9.48791131e-02 -2.96037614e-01 6.71755731e-01
-5.67127824e-01 8.81192982e-01 -4.27171171e-01 -1.14910543e+00
-2.85861999e-01 -1.02962232e+00 1.31234705e+00 -7.97188580e-02
8.79141539e-02 -9.06442523e-01 3.77379984e-01 3.27702671e-01
5.28035879e-01 -4.05019879e-01 6.79996669e-01 -1.19517410e+00
-2.84957260e-01 2.75082290e-01 4.98967059e-02 5.52391350e-01
1.82739571e-01 -2.64455527e-01 -1.39386845e+00 -5.77733517e-01
6.85610950e-01 3.73600900e-01 1.05234408e+00 6.44252658e-01
9.72383618e-01 -3.68566453e-01 -5.02727509e-01 6.55801594e-01
1.05966008e+00 2.45020166e-01 4.17326480e-01 -1.14468057e-02
5.82768083e-01 4.99961287e-01 2.71433711e-01 3.82887065e-01
3.53470802e-01 1.04733682e+00 -1.10332392e-01 6.38058335e-02
-3.08431685e-01 6.12881174e-03 7.66102731e-01 1.13872659e+00
3.03098321e-01 -3.57454062e-01 -6.23978615e-01 6.71709657e-01
-1.91530716e+00 -1.05902267e+00 1.45537592e-02 2.24789262e+00
7.01774955e-01 6.37902990e-02 3.06284904e-01 5.03362596e-01
1.07983947e+00 1.33864954e-01 -1.25079826e-01 -4.11543846e-01
-2.16642603e-01 4.95298207e-01 3.73074830e-01 5.89421630e-01
-1.34056640e+00 7.67704189e-01 6.11244297e+00 1.24497497e+00
-1.02796960e+00 5.11363208e-01 5.80980256e-02 -9.39885303e-02
-1.98356465e-01 -2.27842685e-02 -9.30776536e-01 5.21182477e-01
1.27738428e+00 -2.66730618e-02 3.33080769e-01 5.05886316e-01
3.67462635e-01 3.30022812e-01 -1.22028673e+00 9.33879018e-01
4.78878438e-01 -7.53707170e-01 -6.78091645e-02 -9.99982432e-02
2.94276625e-01 -1.80808946e-01 1.85742736e-01 4.74920869e-01
4.70840484e-02 -7.08926141e-01 7.59759307e-01 2.20524192e-01
3.27555954e-01 -8.14797461e-01 5.18008351e-01 1.65138409e-01
-1.37669063e+00 -2.91446745e-01 -2.50376552e-01 2.36008450e-01
3.88463318e-01 7.18159795e-01 -6.94599867e-01 8.98033202e-01
4.48146164e-01 4.02836412e-01 -4.40057367e-01 1.18443894e+00
-1.50143757e-01 1.07244921e+00 -4.06692296e-01 4.82222110e-01
2.48166434e-02 1.02270216e-01 1.14779842e+00 1.44605434e+00
4.91587162e-01 -3.78831089e-01 1.09691143e-01 6.29172981e-01
2.43919075e-01 -1.52126119e-01 -3.35820317e-01 1.67170539e-01
7.28930235e-01 1.20418036e+00 -5.67688882e-01 -1.08154088e-01
-2.28734359e-01 1.04762661e+00 -2.63797268e-02 3.13989639e-01
-1.01149356e+00 -5.86190581e-01 6.53845727e-01 -2.54477888e-01
1.00276411e+00 -2.99635231e-01 -8.17276090e-02 -1.15788233e+00
5.77290170e-02 -7.65079618e-01 2.05676213e-01 -2.63306916e-01
-1.41881466e+00 5.48084855e-01 6.86867982e-02 -1.21252489e+00
-2.09455147e-01 -3.27815115e-01 -8.16171944e-01 5.87925076e-01
-1.73069727e+00 -1.15265930e+00 3.21586013e-01 6.90296113e-01
7.96849132e-01 -2.07227424e-01 6.71023369e-01 6.14229441e-01
-7.31540859e-01 1.20777202e+00 3.64380240e-01 -5.02448119e-02
1.00384712e+00 -1.16821659e+00 4.14860994e-01 1.24401057e+00
3.97488534e-01 7.32063949e-01 7.57938802e-01 -3.25017989e-01
-1.15301383e+00 -1.12469900e+00 9.86745119e-01 -4.28483903e-01
5.13696492e-01 -7.14081228e-01 -9.47711468e-01 7.73639441e-01
3.00878048e-01 -4.46448594e-01 8.57265294e-01 1.89857319e-01
-3.00462216e-01 -1.91532016e-01 -1.15152419e+00 3.56003284e-01
1.18511832e+00 -2.81705409e-01 -9.66241717e-01 1.09310217e-01
1.00051630e+00 -8.29817075e-03 -3.88122708e-01 4.88829851e-01
3.04548502e-01 -8.80134463e-01 1.12174618e+00 -1.27643272e-01
-4.49214727e-01 -3.52389425e-01 -1.61671102e-01 -1.43439603e+00
-6.65310860e-01 -1.11781943e+00 8.40246007e-02 1.89401472e+00
3.17297339e-01 -8.05981517e-01 5.33529162e-01 1.06436685e-01
-7.66748965e-01 1.62127867e-01 -1.57610059e+00 -1.19749129e+00
-7.20570087e-02 -5.31348288e-01 4.74744469e-01 7.80911565e-01
1.27488915e-02 3.27665150e-01 -3.65927100e-01 6.64032042e-01
7.72668779e-01 1.15260379e-02 4.40569371e-01 -1.18327308e+00
-6.81029320e-01 -4.62539762e-01 -2.35057607e-01 -9.53428745e-01
3.65784883e-01 -8.02981853e-01 3.08501810e-01 -1.10480237e+00
-9.10156891e-02 -2.69973040e-01 -5.67850828e-01 3.65353405e-01
-1.99435428e-01 1.22442268e-01 3.47634077e-01 1.32971078e-01
-6.51453376e-01 7.11548388e-01 4.92079884e-01 -4.35649306e-02
-2.62414277e-01 2.73363590e-01 -6.96133614e-01 7.24735379e-01
7.57417023e-01 -6.20488644e-01 -4.17943716e-01 -2.76022404e-01
-4.27772164e-01 -9.65635255e-02 2.76986361e-01 -1.36277854e+00
3.52080822e-01 3.27188432e-01 -6.97789341e-02 -6.25306189e-01
5.24952888e-01 -7.09607184e-01 -4.02678661e-02 2.94449002e-01
-3.33301693e-01 -4.56623107e-01 3.64796102e-01 7.00151384e-01
-2.48259798e-01 -3.22020739e-01 8.36742043e-01 3.45375419e-01
-3.48236114e-01 -6.84292316e-02 -5.52316844e-01 4.99869809e-02
5.58488548e-01 -4.10238802e-02 -1.50298417e-01 -2.69790381e-01
-8.05796504e-01 6.24910090e-03 -1.97036847e-01 2.59648651e-01
2.50949293e-01 -1.15991592e+00 -7.84696996e-01 5.31334221e-01
-1.09238453e-01 -3.56963784e-01 3.41326237e-01 1.18725789e+00
2.07584918e-01 6.94518387e-01 1.13354765e-01 -7.20327973e-01
-1.34290874e+00 5.72649598e-01 9.42851454e-02 -2.97159314e-01
-4.64432955e-01 1.01545763e+00 4.69101429e-01 -5.75571954e-01
4.55813110e-01 -2.97497272e-01 -1.26803100e-01 3.93463932e-02
4.21026528e-01 5.65239549e-01 3.36924970e-01 -5.54585814e-01
-7.15672672e-01 6.38082027e-01 -2.89665386e-02 -4.82240915e-01
1.30714238e+00 -4.38077986e-01 1.29183978e-01 3.87848943e-01
1.04442739e+00 6.38208747e-01 -1.18695700e+00 -2.69437611e-01
1.55852020e-01 -1.81219906e-01 -5.41674867e-02 -6.85220957e-01
-6.87177837e-01 9.53689396e-01 7.68819988e-01 3.34728122e-01
1.22768390e+00 -1.04933746e-01 8.19972515e-01 1.19596034e-01
2.59778053e-01 -7.66267717e-01 -1.14276558e-01 6.52932942e-01
6.07565761e-01 -8.45045805e-01 -5.88407993e-01 -4.48696405e-01
-5.16245008e-01 8.22334468e-01 2.48073682e-01 1.36945099e-01
6.31666958e-01 4.16636527e-01 -1.71388030e-01 2.62909502e-01
-5.66505253e-01 -3.44536215e-01 3.51679206e-01 7.88380563e-01
2.71225840e-01 -1.21615551e-01 -3.04498643e-01 9.77995694e-01
-2.74593920e-01 -2.33791858e-01 2.51581252e-01 6.19984269e-01
-6.27980590e-01 -1.47051072e+00 -6.98550940e-01 5.79706114e-03
-4.43930507e-01 -3.99479926e-01 -1.86793268e-01 4.23651099e-01
1.17405295e-01 1.24767613e+00 -7.68775493e-02 -2.32706875e-01
1.05725892e-01 6.33719146e-01 4.50710952e-01 -7.20100105e-01
-4.95781302e-01 5.42090297e-01 7.69981444e-02 -1.21415772e-01
-5.22231400e-01 -9.52885985e-01 -9.96561348e-01 1.70644894e-01
-3.73447090e-01 4.25055206e-01 6.94694936e-01 1.00777674e+00
6.71539128e-01 7.38245487e-01 8.30051899e-01 -9.80392218e-01
-6.65544629e-01 -1.07393992e+00 -6.17100954e-01 5.10593737e-03
5.99391460e-01 -7.74081886e-01 -7.70525694e-01 8.62634741e-03] | [14.946768760681152, 5.842853546142578] |
f79386a9-093e-422d-bf74-601f3ce86b28 | a-baseline-for-3d-multi-object-tracking | 1907.03961 | null | https://arxiv.org/abs/1907.03961v5 | https://arxiv.org/pdf/1907.03961v5.pdf | 3D Multi-Object Tracking: A Baseline and New Evaluation Metrics | 3D multi-object tracking (MOT) is an essential component for many applications such as autonomous driving and assistive robotics. Recent work on 3D MOT focuses on developing accurate systems giving less attention to practical considerations such as computational cost and system complexity. In contrast, this work proposes a simple real-time 3D MOT system. Our system first obtains 3D detections from a LiDAR point cloud. Then, a straightforward combination of a 3D Kalman filter and the Hungarian algorithm is used for state estimation and data association. Additionally, 3D MOT datasets such as KITTI evaluate MOT methods in the 2D space and standardized 3D MOT evaluation tools are missing for a fair comparison of 3D MOT methods. Therefore, we propose a new 3D MOT evaluation tool along with three new metrics to comprehensively evaluate 3D MOT methods. We show that, although our system employs a combination of classical MOT modules, we achieve state-of-the-art 3D MOT performance on two 3D MOT benchmarks (KITTI and nuScenes). Surprisingly, although our system does not use any 2D data as inputs, we achieve competitive performance on the KITTI 2D MOT leaderboard. Our proposed system runs at a rate of $207.4$ FPS on the KITTI dataset, achieving the fastest speed among all modern MOT systems. To encourage standardized 3D MOT evaluation, our system and evaluation code are made publicly available at https://github.com/xinshuoweng/AB3DMOT. | ['David Held', 'Kris Kitani', 'Xinshuo Weng', 'Jianren Wang'] | 2019-07-09 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-2.06449002e-01 -4.53120947e-01 7.45083094e-02 -9.73043442e-02
-4.35575426e-01 -4.38394099e-01 5.97962022e-01 -6.29823059e-02
-7.03365266e-01 6.47580862e-01 -8.19060862e-01 -5.73642790e-01
-9.14565171e-04 -8.94669235e-01 -6.45410419e-01 -6.68526590e-01
-1.32625014e-01 9.88287687e-01 8.73063862e-01 -4.49592263e-01
9.55316797e-02 5.99973202e-01 -1.90109932e+00 -5.15619278e-01
6.87238991e-01 9.14977908e-01 3.35663587e-01 1.00985372e+00
8.10044445e-03 1.15767606e-01 -6.15011275e-01 1.63967371e-01
5.88677883e-01 -1.00207254e-01 -1.75675035e-01 -3.85570973e-01
5.82273543e-01 -3.81793767e-01 -2.98955292e-01 7.82560647e-01
8.20749402e-01 -1.02414250e-01 5.03576040e-01 -1.56066263e+00
5.52057251e-02 3.04065142e-02 -4.54779476e-01 2.78907269e-01
3.59906763e-01 4.24381196e-01 6.69097185e-01 -7.35566914e-01
3.85201007e-01 1.19290245e+00 7.08375931e-01 3.32923234e-01
-1.07831085e+00 -9.88522053e-01 -2.40843639e-01 1.21299841e-01
-1.52039695e+00 -4.28438455e-01 4.19943601e-01 -3.10878217e-01
1.06941676e+00 2.75482982e-01 8.22855353e-01 7.96107292e-01
7.24604547e-01 5.44671476e-01 1.23587894e+00 -2.08274368e-03
2.20553026e-01 8.77951737e-03 1.79232210e-01 8.25414479e-01
6.44853592e-01 4.67297792e-01 -5.84192932e-01 3.76558416e-02
5.16973615e-01 7.13542551e-02 3.85331780e-01 -6.89500391e-01
-1.27685118e+00 4.74361539e-01 3.92059922e-01 5.88061512e-02
-1.72229111e-01 4.83783275e-01 1.37847677e-01 4.42650259e-01
2.63892949e-01 -6.69338107e-02 -1.09140575e-01 -3.62843573e-01
-6.60375237e-01 3.84209007e-01 6.92437053e-01 1.39491045e+00
1.06107461e+00 -1.44507721e-01 1.90790027e-01 2.42900908e-01
4.32664156e-01 1.38478410e+00 -2.43093930e-02 -9.52211082e-01
2.27238581e-01 7.05253422e-01 2.55687296e-01 -7.32385457e-01
-6.56185865e-01 -2.08851591e-01 -5.39779484e-01 7.24370301e-01
2.51554579e-01 5.54825366e-02 -6.21830106e-01 1.20298517e+00
7.46913552e-01 8.95360783e-02 8.42796341e-02 8.25333714e-01
1.09633231e+00 3.82282823e-01 -4.79729205e-01 8.86875764e-02
1.22493923e+00 -5.67207217e-01 -4.69396561e-01 -3.92509639e-01
7.86327302e-01 -6.20253563e-01 7.17087924e-01 5.35219684e-02
-8.74350369e-01 -6.21688962e-01 -1.20077384e+00 -1.74033076e-01
-4.63035315e-01 -4.22884598e-02 4.57501054e-01 5.90672135e-01
-1.10364830e+00 2.74801850e-01 -1.48264229e+00 -9.39771712e-01
2.65083104e-01 7.77664602e-01 -1.59068093e-01 -1.04199655e-01
-7.22331107e-01 1.23316491e+00 2.92204380e-01 -2.17984915e-01
-9.46904182e-01 -2.34516621e-01 -6.98857844e-01 -4.91109997e-01
3.76330167e-01 -9.02661443e-01 1.18507326e+00 3.94858688e-01
-1.50716770e+00 9.05543566e-01 -3.27305824e-01 -5.64348340e-01
4.15719539e-01 -3.74552011e-01 -2.25883812e-01 -2.27515414e-01
3.03679913e-01 6.76269352e-01 5.20112097e-01 -9.04176414e-01
-1.07625604e+00 -3.50833088e-01 9.84619036e-02 1.68433994e-01
8.50635320e-02 -3.48151833e-01 -4.88688648e-01 3.76315594e-01
8.34891796e-02 -1.31137359e+00 -6.65472597e-02 1.68015659e-01
-1.61626577e-01 -4.47003752e-01 1.21581280e+00 3.53695512e-01
7.57289529e-01 -2.06737065e+00 -1.16719946e-01 -2.68890679e-01
3.07288200e-01 1.79776460e-01 1.44316837e-01 4.07397777e-01
9.43225265e-01 -3.59408289e-01 -5.69738969e-02 -6.30839527e-01
1.86332956e-01 3.96700323e-01 1.41585529e-01 7.89406657e-01
1.09566711e-01 8.58223617e-01 -1.04894745e+00 -6.13705158e-01
7.39163816e-01 5.09906828e-01 -5.18571198e-01 -1.15611449e-01
-1.82146467e-02 4.23238128e-01 -5.41049719e-01 1.00563455e+00
8.25345099e-01 -2.89372355e-01 -8.77264664e-02 5.55588678e-02
-6.81139171e-01 5.57483673e-01 -1.12684631e+00 1.83414316e+00
-2.80656844e-01 7.01283693e-01 6.49225637e-02 -6.15455687e-01
1.16242206e+00 -2.17413098e-01 7.05080807e-01 -5.98755002e-01
3.67690027e-01 5.39003849e-01 -3.36844176e-02 -2.44119819e-02
9.59264934e-01 1.01972781e-01 -4.21893060e-01 2.68262058e-01
-2.13949680e-01 -7.05638409e-01 2.15819523e-01 1.96532160e-01
1.58852398e+00 3.34077209e-01 2.85562485e-01 -4.84845608e-01
2.40135416e-01 6.11728728e-01 5.61229765e-01 8.13261569e-01
-6.06024086e-01 3.45011204e-02 -1.85037091e-01 -2.61033446e-01
-7.45929420e-01 -1.26104677e+00 -3.46947819e-01 6.44406021e-01
1.08697629e+00 -6.63917005e-01 2.96943239e-03 -5.26458502e-01
6.28342211e-01 3.48053485e-01 -3.21037650e-01 1.27259538e-01
-6.61536098e-01 -6.08106494e-01 6.07939124e-01 2.21974224e-01
6.19255722e-01 -5.69941461e-01 -1.35661376e+00 2.73565024e-01
1.80947140e-01 -1.22194910e+00 -8.21312517e-02 4.11440521e-01
-8.76838267e-01 -1.10332227e+00 -1.10836931e-01 -4.57928360e-01
3.99662495e-01 9.35602009e-01 1.00888467e+00 2.62684643e-01
-1.68998986e-01 2.32640341e-01 -2.82060504e-01 -6.40804350e-01
-2.04228327e-01 2.22164586e-01 6.81308091e-01 -7.94270396e-01
6.85815036e-01 -5.23719788e-01 -4.52106059e-01 7.75990665e-01
-3.02003741e-01 1.24846004e-01 6.59201920e-01 4.85755563e-01
8.04856181e-01 -1.13709264e-01 -3.85894850e-02 -3.77273649e-01
-1.10050164e-01 -4.00821179e-01 -1.14463830e+00 -5.10493517e-01
-5.73020577e-01 -7.39384517e-02 3.31003338e-01 -2.09775954e-01
-3.66461098e-01 3.39854628e-01 -1.82906866e-01 -4.61166143e-01
-1.29368946e-01 6.07490633e-03 2.48578072e-01 -3.46663773e-01
6.22951150e-01 8.67233723e-02 8.60157162e-02 -4.64831889e-01
1.08215354e-01 9.10902083e-01 4.70127404e-01 -4.02524889e-01
1.26445448e+00 9.33999836e-01 5.72274268e-01 -1.05713952e+00
-4.37459916e-01 -6.05277181e-01 -7.48729467e-01 -2.99817324e-01
5.97167671e-01 -1.08597422e+00 -1.09213960e+00 3.11195999e-01
-8.34876478e-01 -4.77666557e-01 -1.91855565e-01 6.37890160e-01
-5.82705796e-01 -1.44678997e-02 -1.74299791e-01 -8.84582698e-01
-2.52082765e-01 -1.33421159e+00 1.45746982e+00 2.32457474e-01
-5.49858361e-02 -5.61958909e-01 3.36257786e-01 1.60299018e-01
4.64603662e-01 2.81616569e-01 1.25573784e-01 -3.00453514e-01
-1.07879281e+00 -1.58257738e-01 -1.61263704e-01 -3.61974865e-01
2.87581921e-01 -4.86841202e-02 -7.89181530e-01 -6.44094348e-01
-3.98653209e-01 -1.04859062e-01 5.97644091e-01 2.59970218e-01
3.12963009e-01 5.20283043e-01 -9.01649952e-01 6.64675534e-01
1.29533768e+00 2.34976068e-01 1.88904867e-01 5.03175139e-01
5.10812998e-01 1.24847360e-01 1.18106282e+00 2.38601193e-01
8.23356450e-01 1.02361381e+00 8.28814983e-01 2.40224481e-01
-2.45755792e-01 -1.61461055e-01 4.98021007e-01 8.92955184e-01
-1.21455863e-01 -3.02726656e-01 -9.95242834e-01 5.36065042e-01
-1.88263547e+00 -6.38921022e-01 -5.48873007e-01 2.27437329e+00
1.54436901e-01 6.19390666e-01 2.26807043e-01 4.05393429e-02
4.06540453e-01 -7.29707163e-03 -7.33209133e-01 6.48313239e-02
8.79098475e-02 9.15764943e-02 1.01620805e+00 5.15495837e-01
-1.09030497e+00 1.09600222e+00 5.47693396e+00 5.84307790e-01
-1.04076159e+00 1.09064192e-01 -5.07368326e-01 -5.05537927e-01
2.40013197e-01 2.09854811e-01 -1.58728027e+00 3.56657565e-01
1.11353505e+00 -1.77732259e-01 2.30064467e-01 7.81242669e-01
1.88889146e-01 -3.30588520e-01 -1.05111158e+00 1.02494168e+00
-3.34801704e-01 -1.27981615e+00 -4.83724058e-01 6.58059299e-01
3.45848531e-01 7.20219314e-01 -2.53635943e-01 4.96447802e-01
4.97254342e-01 -5.53399682e-01 7.72211730e-01 1.46571577e-01
9.32323456e-01 -5.30729413e-01 7.21995234e-01 6.68968022e-01
-1.62908721e+00 3.03364098e-01 -5.43407083e-01 -4.60438073e-01
2.48170078e-01 4.90497589e-01 -9.21067536e-01 7.71656513e-01
1.13267791e+00 9.58232522e-01 -4.09582376e-01 1.03558469e+00
3.85348611e-02 1.76908866e-01 -1.00577962e+00 -5.70015311e-01
8.78547803e-02 -5.32946400e-02 9.91406679e-01 8.44688535e-01
5.87034285e-01 8.67260844e-02 2.79987037e-01 6.63308024e-01
5.19780666e-02 -3.99053961e-01 -1.08508444e+00 1.78264648e-01
7.73419023e-01 1.28448188e+00 -6.81569934e-01 -3.29706579e-01
-3.25561076e-01 4.76358175e-01 -2.69496478e-02 -4.30314094e-01
-8.65097284e-01 -4.30425853e-01 1.17800832e+00 1.80704117e-01
5.00806510e-01 -8.74367833e-01 -2.40276024e-01 -9.22023118e-01
-1.99253306e-01 -3.23729753e-01 -1.63673554e-02 -5.34085095e-01
-8.05825710e-01 4.22490984e-01 1.95704877e-01 -1.67470276e+00
-2.52433270e-01 -5.90407491e-01 -2.76118755e-01 5.56124389e-01
-1.77560592e+00 -8.87401581e-01 -7.06965148e-01 4.29803252e-01
3.04602325e-01 1.15219697e-01 4.59829479e-01 4.99269992e-01
-4.31776732e-01 2.90598303e-01 2.08292499e-01 -3.82944137e-01
6.84742093e-01 -1.11006105e+00 8.65202487e-01 7.60477722e-01
3.16399001e-02 4.21129405e-01 8.38119864e-01 -8.87617707e-01
-2.22737503e+00 -9.97772694e-01 7.18543768e-01 -9.29028153e-01
6.27651393e-01 -6.29329145e-01 -4.54563558e-01 6.33933544e-01
-3.72545309e-02 1.57009140e-01 1.60691932e-01 -2.37211078e-01
3.77033949e-01 -2.78696984e-01 -1.12911701e+00 3.35811287e-01
1.61217654e+00 -2.10918188e-02 -3.00236613e-01 2.76971787e-01
6.57401681e-01 -8.95405173e-01 -8.58552217e-01 6.28760993e-01
6.70325816e-01 -1.05778134e+00 8.86782348e-01 2.22035378e-01
-3.91678959e-01 -1.02807295e+00 -4.01662350e-01 -8.59239578e-01
-1.10048808e-01 -6.20239437e-01 -3.13495517e-01 7.40145326e-01
3.38817528e-03 -8.64774764e-01 9.46988881e-01 -2.25027159e-01
-5.19571066e-01 -4.31663334e-01 -1.37879765e+00 -1.26561213e+00
-2.81945288e-01 -4.93293822e-01 7.13552415e-01 2.99520314e-01
-3.21550101e-01 4.66976464e-01 -2.08396837e-01 5.31689882e-01
1.03946698e+00 4.46016282e-01 1.66420269e+00 -1.50376666e+00
9.07168388e-02 -7.46963248e-02 -7.39495635e-01 -1.52040386e+00
-1.48717880e-01 -8.82507801e-01 2.73690283e-01 -1.52724218e+00
-1.73723072e-01 -8.80553424e-01 -1.71427786e-01 3.78173620e-01
2.89805740e-01 5.69097102e-01 2.21554950e-01 4.86526996e-01
-8.95067334e-01 6.39145315e-01 1.09001303e+00 -1.22972138e-01
-2.93155849e-01 1.27923027e-01 -2.96630055e-01 4.59563673e-01
6.32193625e-01 -7.34196126e-01 -1.07899075e-02 -5.60960472e-01
-9.42390338e-02 -8.54123458e-02 5.89444220e-01 -1.60954511e+00
5.66894650e-01 -8.66229087e-02 2.09348891e-02 -1.39687288e+00
9.57061291e-01 -9.77283061e-01 4.00249243e-01 9.11602497e-01
6.69738233e-01 1.89187497e-01 4.50342655e-01 3.38859111e-01
2.88644612e-01 3.08296919e-01 7.23130882e-01 1.05507612e-01
-8.79909277e-01 3.60945761e-01 -3.46389115e-01 -2.56340414e-01
1.21576643e+00 -5.20619571e-01 -7.19553411e-01 1.59556679e-02
1.07564658e-01 6.33214712e-01 1.05777025e+00 3.23172927e-01
5.19770265e-01 -1.31444681e+00 -6.14717960e-01 2.52353817e-01
2.91635245e-01 2.28708208e-01 -1.92032650e-01 1.16462719e+00
-6.10490322e-01 7.17997849e-01 -2.66393632e-01 -1.33388340e+00
-1.44304514e+00 9.40199569e-02 1.73790291e-01 5.10841310e-02
-8.15134883e-01 5.28927922e-01 -1.47920132e-01 -6.39089823e-01
4.13448736e-02 -6.09289587e-01 3.96981120e-01 -3.93909454e-01
1.72787830e-01 4.51717824e-01 1.14063866e-01 -7.07691252e-01
-9.36290443e-01 8.14175785e-01 2.58567840e-01 1.59997538e-01
1.18207800e+00 -2.74375051e-01 3.64134431e-01 5.55837095e-01
6.52216434e-01 -4.94151823e-02 -1.21007729e+00 -1.60143554e-01
-1.38396263e-01 -6.08282626e-01 1.05408370e-01 -2.77686507e-01
-3.78479362e-01 7.99442172e-01 1.03575301e+00 2.63540477e-01
7.67248750e-01 2.26588964e-01 9.10163760e-01 7.42068410e-01
1.09579992e+00 -6.34252191e-01 -3.12876374e-01 8.40333402e-01
1.93466455e-01 -1.50017703e+00 2.90486813e-01 -5.05445711e-02
2.13787816e-02 6.31469369e-01 6.81015551e-01 -1.68332130e-01
3.95043939e-01 5.51563025e-01 1.47103161e-01 -3.29513013e-01
-7.65913665e-01 -6.73198104e-01 -1.06128529e-01 5.94600737e-01
-1.70037206e-02 9.49569233e-03 -1.77821666e-01 -1.62099406e-01
-4.35295463e-01 -1.12470146e-02 3.80606294e-01 1.49505699e+00
-7.57302165e-01 -1.17975390e+00 -4.80553031e-01 3.92561287e-01
8.59307498e-02 2.06115469e-01 -2.32039869e-01 1.23546600e+00
7.54122809e-02 1.00844264e+00 1.61946416e-01 -9.34559643e-01
5.96511960e-01 -3.61762643e-01 5.59101880e-01 -6.01491332e-01
-3.85895163e-01 -1.35848507e-01 1.33030610e-02 -6.72919452e-01
-5.83506763e-01 -8.38168740e-01 -1.43437064e+00 -7.65446067e-01
-5.41787744e-01 3.85315716e-02 1.07620049e+00 8.75749409e-01
5.71764827e-01 3.71276766e-01 3.62842083e-01 -1.19631457e+00
-1.71145648e-01 -8.07739139e-01 -4.86120969e-01 -2.58963466e-01
5.30074835e-01 -1.21470785e+00 -3.53902638e-01 -5.41141450e-01] | [6.7227678298950195, -2.2536988258361816] |
c5e4a1ff-07d5-427f-8834-f98b13abde29 | modeling-multi-hop-question-answering-as | null | null | https://openreview.net/forum?id=C1XEENowywW | https://openreview.net/pdf?id=C1XEENowywW | Modeling Multi-hop Question Answering as Single Sequence Prediction | Fusion-in-decoder (Fid) (Izacard and Grave, 2020) is a generative question answering (QA) model that leverages passage retrieval with a pre-trained transformer and pushed the state of the art on single-hop QA. However, the complexity of multi-hop QA hinders the effectiveness of the generative QA approach. In this work, we propose a simple generative approach (PathFid) that extends the task beyond just answer generation by explicitly modeling the reasoning process to resolve the answer for multi-hop questions. By linearizing the hierarchical reasoning path of supporting passages, their key sentences, and finally the factoid answer, we cast the problem as a single sequence prediction task. To facilitate complex reasoning with multiple clues, we further extend the unified flat representation of multiple input documents by encoding cross-passage interactions. Our extensive experiments demonstrate that PathFid leads to strong performance gains on two multi-hop QA datasets: HotpotQA and IIRC. Besides the performance gains, PathFid is more interpretable, which in turn yields answers that are more faithfully grounded to the supporting passages and facts compared to the baseline Fid model. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['multi-hop-question-answering', 'generative-question-answering'] | ['knowledge-base', 'natural-language-processing'] | [-1.64659947e-01 5.72726130e-01 1.65149987e-01 -2.08409280e-01
-1.69665611e+00 -7.76507556e-01 8.46621513e-01 6.48983046e-02
8.79535675e-02 9.70142126e-01 9.47264612e-01 -6.37941539e-01
-1.31609932e-01 -9.90723372e-01 -8.25522006e-01 -1.47434762e-02
5.05673826e-01 8.50523591e-01 5.63930631e-01 -8.59613180e-01
6.78480789e-02 -3.81477743e-01 -1.24680889e+00 8.90247822e-01
1.28801930e+00 6.07299268e-01 7.53810480e-02 9.16089594e-01
-5.69218993e-01 1.48092413e+00 -7.46330142e-01 -1.07173932e+00
-1.12925276e-01 -8.13698649e-01 -1.43445003e+00 -4.17154104e-01
4.94426072e-01 -6.21393383e-01 -4.76833463e-01 6.58059120e-01
3.06425691e-01 1.30380809e-01 5.80311775e-01 -9.59804595e-01
-1.21997666e+00 9.22330022e-01 -1.23385675e-02 2.94364631e-01
9.70835268e-01 1.52005449e-01 1.60987830e+00 -9.30938601e-01
7.22815633e-01 1.48817623e+00 5.17605186e-01 4.62590426e-01
-9.18702066e-01 -9.62851495e-02 1.71030685e-01 5.95670938e-01
-9.50305760e-01 -2.35585779e-01 4.77094769e-01 -1.07462943e-01
1.25221848e+00 4.55554247e-01 4.36234802e-01 1.11122859e+00
1.02186054e-01 1.07772744e+00 7.02560008e-01 -3.49747568e-01
-1.53869828e-02 -2.49386400e-01 4.42629725e-01 8.08428943e-01
-2.07125004e-02 -4.81218070e-01 -4.49937761e-01 -1.54055521e-01
2.22661391e-01 -3.00804496e-01 -4.37145174e-01 1.46964520e-01
-9.68358517e-01 1.00201821e+00 4.30583239e-01 1.01703241e-01
-4.65775162e-01 -3.64019014e-02 2.58057147e-01 4.51539397e-01
1.42646819e-01 7.70709336e-01 -4.26434577e-01 -2.15847462e-01
-8.25827241e-01 6.86454177e-01 1.18645608e+00 1.01309109e+00
5.39776504e-01 -3.88738424e-01 -9.86552119e-01 5.54224491e-01
3.55094850e-01 5.29213130e-01 1.74441263e-01 -1.21922147e+00
8.72713745e-01 8.34438622e-01 1.98643103e-01 -8.48017693e-01
-6.83586821e-02 -4.74666715e-01 -3.76796395e-01 -7.56658673e-01
3.45909745e-01 -5.65805957e-02 -7.61245906e-01 1.63336134e+00
2.84845799e-01 -3.60246301e-01 4.08072621e-01 8.45914543e-01
1.22460508e+00 1.02357495e+00 -5.91519941e-03 1.79840222e-01
1.59890485e+00 -1.52674735e+00 -7.99339592e-01 -3.87949795e-01
5.18822014e-01 -6.93479061e-01 1.43838739e+00 7.60874972e-02
-1.28456271e+00 -3.94553751e-01 -7.87868440e-01 -8.42313707e-01
-6.82113469e-02 4.66361530e-02 5.20231485e-01 8.78071710e-02
-1.24279237e+00 -1.75477471e-02 -3.69018942e-01 -1.78624809e-01
1.17218144e-01 -2.72085518e-01 3.43114175e-02 -6.88895643e-01
-1.74146414e+00 1.02701986e+00 2.08819985e-01 -1.12066835e-01
-8.36087465e-01 -9.71328795e-01 -9.17596757e-01 3.74384552e-01
6.43858135e-01 -1.51564753e+00 1.67140317e+00 -8.36157054e-02
-1.50875223e+00 3.19690645e-01 -5.81655622e-01 -5.59748888e-01
4.04959887e-01 -6.27874255e-01 -4.18759435e-01 5.24134278e-01
4.85435486e-01 7.01877058e-01 5.10645330e-01 -1.11206138e+00
-4.10150826e-01 -4.27452363e-02 8.25012743e-01 2.66023308e-01
1.15985431e-01 -1.17158696e-01 -6.79073274e-01 -2.95275241e-01
-1.14468180e-01 -4.86696482e-01 1.09197507e-02 -5.01780391e-01
-5.94194829e-01 -5.33139408e-01 4.45685297e-01 -1.17256141e+00
1.52842712e+00 -1.71240425e+00 3.57198030e-01 -2.58773029e-01
2.20698044e-01 9.68551263e-02 -3.80648792e-01 9.75507855e-01
3.60253930e-01 -1.52288759e-02 -2.36194789e-01 -2.79798418e-01
4.62809414e-01 4.43785846e-01 -1.09282386e+00 -5.60238183e-01
4.83204961e-01 1.61921930e+00 -9.89071190e-01 -5.77273428e-01
-4.19726789e-01 2.70833760e-01 -9.34730113e-01 3.58984619e-01
-8.30104887e-01 3.44426304e-01 -6.02358043e-01 5.93951702e-01
2.28514135e-01 -7.75616288e-01 1.35233486e-02 -1.75763488e-01
4.09683883e-01 1.15955079e+00 -5.15566587e-01 1.84924746e+00
-4.58449215e-01 3.19183946e-01 -2.72419304e-01 -3.59713405e-01
7.52930582e-01 4.99316156e-01 -4.57983352e-02 -1.02447808e+00
-3.13249648e-01 1.90807164e-01 -1.38791353e-01 -7.18560755e-01
9.47612584e-01 -1.91711895e-02 -2.46084750e-01 7.10215807e-01
2.10020676e-01 -3.61400604e-01 6.08557343e-01 9.49166775e-01
1.37588441e+00 3.18578482e-01 -6.02627231e-04 1.03183433e-01
6.14658952e-01 3.17009300e-01 1.79799229e-01 9.23116803e-01
2.03012094e-01 5.79681277e-01 5.47272146e-01 -9.66411829e-02
-8.16514313e-01 -1.27118933e+00 3.41692865e-01 9.38030839e-01
9.84692499e-02 -9.08815145e-01 -7.62635946e-01 -9.36487496e-01
-1.10187337e-01 1.20933247e+00 -5.20427227e-01 -2.44982928e-01
-8.40908885e-01 -3.98636252e-01 8.25335622e-01 5.95293343e-01
6.01718307e-01 -7.97445416e-01 -3.04403335e-01 3.13615382e-01
-1.17878103e+00 -1.12188017e+00 -4.28166658e-01 -3.67031872e-01
-5.71750045e-01 -9.78911698e-01 -6.02460802e-01 -3.89822304e-01
3.37772071e-01 1.72488227e-01 1.76980436e+00 2.05145672e-01
1.56101406e-01 6.15290284e-01 -6.95890009e-01 -3.71392593e-02
-5.98671854e-01 1.87453374e-01 -6.02920532e-01 -3.76476794e-01
1.23703890e-01 -4.05936539e-01 -6.94067240e-01 -3.07665486e-02
-8.93038094e-01 2.36909762e-01 5.48247218e-01 8.88031840e-01
4.22133803e-01 -2.93196589e-01 7.53782034e-01 -8.00511897e-01
1.10314703e+00 -7.36390352e-01 -1.00656167e-01 8.56255472e-01
-4.34769511e-01 4.16462839e-01 6.48067355e-01 6.37302548e-03
-1.52533722e+00 -8.33414912e-01 -5.46281636e-01 8.27971622e-02
3.25606495e-01 7.46263683e-01 -6.84183016e-02 6.06291413e-01
5.83331287e-01 3.68211359e-01 -3.25978756e-01 -3.94130737e-01
8.56046021e-01 2.38825396e-01 7.25546539e-01 -9.59071696e-01
8.14102292e-01 1.03956252e-01 -3.14786911e-01 -8.74305516e-02
-1.44429994e+00 -2.82941788e-01 -1.66738942e-01 -4.94360104e-02
8.75857115e-01 -9.00346339e-01 -5.41321874e-01 -2.20646247e-01
-1.58453774e+00 -2.83264935e-01 -4.27432179e-01 -5.27422130e-02
-4.57897276e-01 4.53155756e-01 -9.91225779e-01 -5.84124923e-01
-7.73349524e-01 -8.59031141e-01 1.14326775e+00 2.13235304e-01
-5.00342727e-01 -9.93635654e-01 2.53983378e-01 1.02081728e+00
5.06250560e-01 -2.32365653e-01 1.49225008e+00 -5.23793578e-01
-9.68411386e-01 -5.09138741e-02 -2.56216675e-01 5.15343137e-02
-1.05838120e-01 -2.16415465e-01 -8.54276240e-01 2.18224451e-01
7.54970536e-02 -7.04119563e-01 9.37796116e-01 -1.71003237e-01
6.36533201e-01 -6.83520555e-01 3.80239449e-02 -2.75655035e-02
1.02989495e+00 -2.38244943e-02 8.76810730e-01 2.73758590e-01
5.29445767e-01 5.97616851e-01 5.27639747e-01 -3.41245458e-02
1.28535783e+00 4.68047202e-01 2.67620534e-01 1.06763951e-01
-5.81996322e-01 -8.57490480e-01 4.37828183e-01 1.20585525e+00
1.76597014e-01 -5.11025250e-01 -8.39800298e-01 5.95062375e-01
-1.89841163e+00 -1.14375508e+00 -3.12283456e-01 1.56671119e+00
1.13667870e+00 -1.34898767e-01 -1.54825568e-01 -2.22164299e-02
1.39077753e-01 4.93799075e-02 -2.70541608e-01 -3.76666307e-01
-3.23352218e-01 4.30249244e-01 -4.35266703e-01 9.16296482e-01
-5.34172714e-01 9.17809069e-01 6.53288031e+00 7.34783649e-01
-4.40463185e-01 1.96319908e-01 2.81395346e-01 1.04766250e-01
-1.11004829e+00 2.55531341e-01 -9.53079104e-01 2.07902551e-01
9.44729626e-01 -2.40198582e-01 4.03905839e-01 3.33670080e-01
-2.82557845e-01 -1.26623407e-01 -1.10361278e+00 3.15802097e-01
2.61339843e-01 -1.50222003e+00 7.06872106e-01 -3.09199184e-01
6.56391323e-01 -2.63975203e-01 3.61850858e-03 9.24138546e-01
6.39775097e-01 -8.54481280e-01 7.90724277e-01 7.15284169e-01
3.32773298e-01 -4.09322232e-01 6.96108937e-01 4.29260522e-01
-8.50614309e-01 -2.26319283e-01 -1.92776799e-01 -3.54476757e-02
6.92888319e-01 5.12479842e-01 -8.45120728e-01 1.23241949e+00
4.94721562e-01 3.36508989e-01 -6.91473484e-01 6.77943349e-01
-8.62571597e-01 7.67612517e-01 -5.79735301e-02 -3.68790515e-02
4.51633751e-01 -5.05027957e-02 4.17351186e-01 1.02822399e+00
4.19031382e-01 3.24973226e-01 -1.22272149e-01 1.20433593e+00
-1.77319258e-01 -1.20358013e-01 -2.15071276e-01 -1.64171949e-01
4.69017297e-01 9.21403527e-01 -1.01144724e-01 -5.93268931e-01
-4.47399408e-01 1.09283006e+00 5.33665359e-01 5.07511258e-01
-8.22738409e-01 -2.21151128e-01 2.76259005e-01 -1.55276656e-01
4.25730646e-01 -1.52349123e-03 -1.10415541e-01 -1.44750202e+00
2.72371650e-01 -1.24338901e+00 8.46954882e-01 -1.13902855e+00
-1.33519113e+00 7.74757922e-01 -3.73605080e-02 -8.39668810e-01
-7.05417931e-01 -2.43482918e-01 -6.14854336e-01 1.01603878e+00
-1.83124447e+00 -1.38294899e+00 -5.74958464e-03 6.46469057e-01
7.06251383e-01 3.96725595e-01 1.02724957e+00 2.98190951e-01
-3.17037225e-01 5.42381108e-01 -4.35731024e-01 5.41862957e-02
5.51822186e-01 -1.32695782e+00 5.74495912e-01 1.04113603e+00
3.35758328e-01 9.89520550e-01 6.03338361e-01 -6.62672937e-01
-1.53835046e+00 -9.12522793e-01 1.40912020e+00 -9.81376410e-01
7.79586196e-01 -1.56993255e-01 -1.11626470e+00 7.81606317e-01
5.77771723e-01 -6.39281571e-01 6.49075925e-01 1.84713691e-01
-7.22657502e-01 9.12592635e-02 -6.93466306e-01 8.29309642e-01
8.02659631e-01 -8.99786592e-01 -1.29224825e+00 3.34080577e-01
1.46650505e+00 -5.63535571e-01 -7.71886766e-01 3.06483775e-01
2.23875001e-01 -7.71425426e-01 1.03354585e+00 -8.16132069e-01
1.05389929e+00 -4.54463482e-01 -3.01630944e-01 -1.10061908e+00
-3.44529480e-01 -6.27336919e-01 -7.66316950e-01 1.32374120e+00
8.17925215e-01 -3.62727135e-01 2.80687243e-01 6.49173558e-01
-4.75317627e-01 -8.83039594e-01 -7.69184589e-01 -4.42850530e-01
2.16926634e-01 -3.33095551e-01 8.85796368e-01 4.64240313e-01
1.21479549e-01 9.24879789e-01 -1.34462789e-01 1.72248706e-01
2.18454301e-01 5.21134138e-01 6.79053068e-01 -7.36368418e-01
-8.08851361e-01 -1.51335135e-01 3.69204998e-01 -1.61138344e+00
-2.90138517e-02 -9.51743603e-01 1.03877909e-01 -2.29713440e+00
1.99204549e-01 -6.05401583e-02 1.39868930e-01 3.68391275e-01
-7.43452728e-01 -1.00603491e-01 1.91132024e-01 2.01185569e-01
-1.00237191e+00 8.50086749e-01 1.64510775e+00 -2.58791000e-01
1.94631115e-01 -3.02577078e-01 -1.10807335e+00 2.69961566e-01
4.79104578e-01 -1.51288465e-01 -8.20906162e-01 -1.07417703e+00
7.19484687e-01 4.57305431e-01 4.75209028e-01 -6.20632231e-01
5.17680407e-01 2.15350464e-01 -1.83730144e-02 -6.76765561e-01
5.01861811e-01 -1.87364951e-01 -1.29025444e-01 1.94059074e-01
-5.68248689e-01 2.60374278e-01 2.39860922e-01 5.61129749e-01
-4.97728854e-01 -1.03669524e-01 -7.31387511e-02 -1.77119300e-01
-4.76875067e-01 -3.18665951e-02 -2.38029540e-01 6.20626569e-01
3.88054878e-01 1.94218650e-01 -1.03509390e+00 -8.37050676e-01
-4.02872860e-01 5.62252641e-01 -8.91573075e-03 3.81071180e-01
6.55504048e-01 -1.18657875e+00 -8.90077353e-01 -2.23295376e-01
2.36543447e-01 1.45108551e-01 4.94300783e-01 8.96939099e-01
-2.63591647e-01 9.32341337e-01 2.33337075e-01 -2.02287629e-01
-8.01762938e-01 4.79040951e-01 2.62367994e-01 -8.35558891e-01
-6.33900166e-01 1.02627933e+00 1.27740055e-01 -5.74456453e-01
-1.36761039e-01 -4.18038249e-01 -3.09731156e-01 -1.86018320e-03
7.19477773e-01 2.36872345e-01 1.71604261e-01 -1.40642554e-01
-4.70436290e-02 1.74823642e-01 -3.83225918e-01 -3.40630889e-01
1.01654172e+00 -3.38090986e-01 -4.54452753e-01 1.49433643e-01
7.08432972e-01 7.48081133e-02 -9.95046556e-01 -4.40329194e-01
7.06502348e-02 -8.40505064e-02 -3.42344820e-01 -1.35219824e+00
-4.99045849e-01 9.11889672e-01 -4.83157188e-01 2.12508515e-01
1.01695549e+00 3.75964999e-01 1.38783157e+00 6.23898506e-01
4.79061574e-01 -5.45685053e-01 4.18146163e-01 1.02568638e+00
1.22896671e+00 -8.03959846e-01 -4.47455466e-01 -4.54786330e-01
-7.47684896e-01 9.37106848e-01 5.92891753e-01 3.69358003e-01
1.40505526e-02 8.62635020e-03 1.30203620e-01 -4.12526548e-01
-1.23794913e+00 -2.27524877e-01 4.60928053e-01 3.69068861e-01
3.11469972e-01 -1.64108053e-01 -2.84804463e-01 1.05119240e+00
-5.84954739e-01 6.18191771e-02 3.22171360e-01 7.95022786e-01
-3.90644670e-01 -1.01891601e+00 -1.89298183e-01 2.63811231e-01
-2.98747301e-01 -6.57045543e-01 -5.38504541e-01 6.02332056e-01
-2.88911521e-01 1.37704456e+00 -2.05656156e-01 -3.08829099e-01
3.92664492e-01 4.17621940e-01 5.02853870e-01 -6.02836609e-01
-6.04073763e-01 -4.51195210e-01 5.80307305e-01 -5.22856832e-01
-4.00488935e-02 -3.00515056e-01 -1.32794893e+00 -3.06589603e-01
-4.34004925e-02 6.55459642e-01 9.12775695e-02 1.25567520e+00
7.61011779e-01 7.99786031e-01 1.21775679e-01 1.82135403e-01
-8.25108349e-01 -9.73505557e-01 1.72502264e-01 2.21110106e-01
2.78568596e-01 -2.81273097e-01 -2.34049186e-02 1.80010274e-02] | [11.126029968261719, 7.951905727386475] |
03abe157-d310-4482-bc54-7117233b0c32 | improving-deep-policy-gradients-with-value | 2302.10145 | null | https://arxiv.org/abs/2302.10145v1 | https://arxiv.org/pdf/2302.10145v1.pdf | Improving Deep Policy Gradients with Value Function Search | Deep Policy Gradient (PG) algorithms employ value networks to drive the learning of parameterized policies and reduce the variance of the gradient estimates. However, value function approximation gets stuck in local optima and struggles to fit the actual return, limiting the variance reduction efficacy and leading policies to sub-optimal performance. This paper focuses on improving value approximation and analyzing the effects on Deep PG primitives such as value prediction, variance reduction, and correlation of gradient estimates with the true gradient. To this end, we introduce a Value Function Search that employs a population of perturbed value networks to search for a better approximation. Our framework does not require additional environment interactions, gradient computations, or ensembles, providing a computationally inexpensive approach to enhance the supervised learning task on which value networks train. Crucially, we show that improving Deep PG primitives results in improved sample efficiency and policies with higher returns using common continuous control benchmark domains. | ['Christopher Amato', 'Enrico Marchesini'] | 2023-02-20 | null | null | null | null | ['value-prediction', 'continuous-control'] | ['computer-code', 'playing-games'] | [-1.21398583e-01 8.36726874e-02 -5.80691397e-01 -1.10546030e-01
-7.78080940e-01 -6.34180069e-01 6.40938342e-01 2.72043705e-01
-6.95477307e-01 1.02727592e+00 1.55549422e-01 -4.71455842e-01
-2.49965951e-01 -7.66337156e-01 -9.92716551e-01 -7.24456787e-01
-2.67131120e-01 3.35355192e-01 1.42562285e-01 -2.76299030e-01
4.59806889e-01 4.33908135e-01 -1.49223232e+00 7.02604428e-02
7.67445505e-01 1.22300375e+00 2.35930644e-03 6.98988736e-01
6.15575798e-02 5.79799891e-01 -8.13946009e-01 -4.15050536e-02
5.57610214e-01 -3.24905694e-01 -2.87558705e-01 -4.66913313e-01
4.10344094e-01 -6.35701299e-01 3.99338156e-02 9.91110861e-01
6.21399641e-01 4.10093129e-01 3.86094838e-01 -1.19095719e+00
-3.46456766e-01 8.58273387e-01 -3.02923590e-01 2.32653975e-01
-2.46143341e-01 7.88902581e-01 7.91990042e-01 -1.19797796e-01
5.76116145e-01 1.38176036e+00 8.25512826e-01 6.92785919e-01
-1.46410978e+00 -3.55054170e-01 2.89280117e-01 -3.16280693e-01
-7.57384241e-01 -4.20257956e-01 2.97422171e-01 -1.40257195e-01
1.10149431e+00 -1.50130421e-01 9.34012413e-01 1.03614545e+00
3.14666837e-01 5.94230890e-01 1.10045898e+00 -8.01296681e-02
8.42167377e-01 1.13633484e-01 -1.61120787e-01 6.32717729e-01
4.17398721e-01 8.49721968e-01 -2.90593475e-01 -2.82085389e-01
7.51313210e-01 -1.20485462e-01 -2.47059911e-01 -6.67968094e-01
-6.77049339e-01 7.07129300e-01 4.83943522e-01 -6.11926652e-02
-6.55637920e-01 7.93979764e-01 4.14341211e-01 4.46412832e-01
2.16836408e-01 1.21797204e+00 -7.57344663e-01 -6.52743518e-01
-6.51318192e-01 7.68350244e-01 8.14831614e-01 4.03908759e-01
6.20460272e-01 3.07920694e-01 -7.13443100e-01 4.96982753e-01
-1.55622318e-01 6.48807824e-01 5.08310735e-01 -1.55198729e+00
4.56563950e-01 4.47069734e-01 5.69445193e-01 -6.82966888e-01
-2.70998746e-01 -8.25206816e-01 -1.24177663e-02 6.85201406e-01
8.35090518e-01 -7.00426757e-01 -8.41162384e-01 1.84711576e+00
2.46003270e-01 -6.82870373e-02 -6.75229728e-02 6.76232040e-01
-2.04226658e-01 4.57499802e-01 6.54785559e-02 -1.02094010e-01
5.51280499e-01 -8.30854595e-01 -2.66504407e-01 -1.94297299e-01
7.07390070e-01 -7.84289315e-02 1.46107459e+00 2.39982814e-01
-1.23193371e+00 -2.26655126e-01 -9.70807135e-01 4.19911474e-01
-2.43654042e-01 -2.94196129e-01 3.87591749e-01 5.77327490e-01
-1.01816499e+00 1.37248409e+00 -1.01795352e+00 1.20152362e-01
7.54750609e-01 4.90265816e-01 3.65525872e-01 2.18147919e-01
-7.07599580e-01 1.02371454e+00 5.23901224e-01 -3.04126501e-01
-1.19461215e+00 -1.12483215e+00 -3.78664255e-01 4.11116451e-01
5.82882702e-01 -4.44557101e-01 1.51307631e+00 -1.01332259e+00
-1.91780853e+00 -6.25819936e-02 2.64344782e-01 -1.02731943e+00
8.54461014e-01 -2.89533854e-01 2.91217268e-01 -3.88900667e-01
-3.83412451e-01 7.72017300e-01 9.84280467e-01 -1.12000906e+00
-7.30063140e-01 -2.18025118e-01 2.48357523e-02 2.75731325e-01
-4.34301972e-01 -5.99388361e-01 2.22108603e-01 -7.61740059e-02
-5.58310747e-01 -8.05013657e-01 -4.21592295e-01 -1.94440216e-01
1.59955993e-02 -1.73490912e-01 5.17952085e-01 -3.75576258e-01
1.06985819e+00 -1.98461437e+00 -9.22188982e-02 5.15693605e-01
5.71480766e-02 3.35032791e-01 -3.72289807e-01 1.74417570e-01
4.43969607e-01 1.73628807e-01 4.14273143e-02 -3.89385186e-02
2.53059059e-01 3.60592127e-01 -4.55024153e-01 4.11051571e-01
1.03150479e-01 9.54294562e-01 -1.16917586e+00 7.02154683e-03
1.79826133e-02 3.03918093e-01 -1.01228440e+00 1.25522465e-01
-7.38844872e-01 3.14148754e-01 -5.00123084e-01 2.68248320e-01
8.77195895e-02 -3.66349262e-03 2.14738935e-01 3.11212689e-01
-3.45365137e-01 3.87404412e-01 -9.01686668e-01 1.04928303e+00
-6.14382505e-01 4.64318782e-01 6.62804022e-02 -9.53238904e-01
1.03210306e+00 -3.42394769e-01 3.80455643e-01 -8.60963762e-01
3.15548122e-01 3.54614139e-01 2.19351992e-01 -3.94618139e-02
5.48805118e-01 9.13510472e-02 2.97678709e-01 3.98672312e-01
-2.18248293e-01 -2.90401399e-01 3.27936143e-01 -4.63670820e-01
1.07011616e+00 4.23367679e-01 -4.17584628e-02 -4.20094609e-01
-2.88209459e-03 1.81990832e-01 4.90825117e-01 1.14514792e+00
-2.44437069e-01 -9.04295295e-02 1.11754441e+00 -4.75099027e-01
-1.31999016e+00 -1.05812716e+00 7.54005983e-02 1.29205394e+00
-5.59483804e-02 -7.52290785e-02 -7.42791772e-01 -6.91514492e-01
5.56318462e-01 1.07264364e+00 -6.14456594e-01 -5.80070853e-01
-7.04856813e-01 -4.04843092e-01 5.08067191e-01 5.71739316e-01
4.61607188e-01 -9.60946202e-01 -1.12009072e+00 3.81234795e-01
4.61340934e-01 -4.92553800e-01 -3.86659890e-01 3.56025100e-01
-1.19254684e+00 -7.68871903e-01 -6.18365943e-01 -2.44510189e-01
5.86709201e-01 -4.32286263e-01 1.08105266e+00 -2.24534180e-02
4.68453132e-02 3.35757375e-01 2.09877118e-01 -5.99628568e-01
-5.19776225e-01 3.12986821e-01 1.15046129e-02 -5.62744617e-01
-1.92183077e-01 -4.43060517e-01 -6.09723389e-01 5.72943613e-02
-4.30841446e-01 -5.30782461e-01 3.81292850e-01 1.03072488e+00
5.68654835e-01 1.86100267e-02 4.90960658e-01 -3.84843886e-01
1.33302486e+00 -3.08656543e-01 -1.18542290e+00 2.29393259e-01
-1.12300992e+00 8.44604492e-01 1.02038014e+00 -7.78929114e-01
-9.62392628e-01 -2.69863755e-01 3.91202331e-01 -4.52709138e-01
4.15197015e-01 2.10609883e-01 5.57535589e-01 1.81511380e-02
1.11385071e+00 -5.86597286e-02 5.16015053e-01 -3.07555765e-01
3.95919710e-01 9.93806198e-02 3.23211581e-01 -1.07598710e+00
3.04272830e-01 1.64915800e-01 2.34762609e-01 -2.12797150e-01
-5.77506781e-01 2.20280141e-01 9.25364345e-02 -1.58687919e-01
2.69690990e-01 -4.60100710e-01 -1.01672375e+00 1.77860647e-01
-6.30255580e-01 -1.18596292e+00 -8.39299798e-01 3.72774810e-01
-7.46260166e-01 -2.91998982e-01 -3.91529232e-01 -8.78461123e-01
-3.03107888e-01 -1.12744057e+00 6.92848504e-01 4.55594867e-01
-4.94024083e-02 -9.12619472e-01 2.12840199e-01 -3.95861924e-01
8.64439309e-01 2.86914259e-01 1.01019788e+00 -5.18345058e-01
-4.03884619e-01 6.68473616e-02 -1.51763838e-02 4.49273646e-01
-2.28287935e-01 -4.86231335e-02 -6.13371670e-01 -3.56837302e-01
-2.15279281e-01 -3.12591255e-01 7.70909309e-01 6.33137047e-01
1.07616067e+00 -6.81489527e-01 -2.35132575e-01 7.30446100e-01
1.42009926e+00 3.77831489e-01 2.90110588e-01 7.08283067e-01
2.86226213e-01 3.87159377e-01 5.39028287e-01 7.94604063e-01
-1.98144257e-01 2.89881885e-01 5.05217969e-01 3.38611811e-01
9.72140655e-02 -5.61694205e-01 5.58787167e-01 -3.29721600e-01
-9.78684574e-02 7.48900846e-02 -9.63871717e-01 4.77716208e-01
-1.83812284e+00 -9.61969018e-01 6.56550705e-01 2.50368094e+00
1.09221530e+00 3.78616154e-01 4.94400442e-01 -4.45982963e-01
5.08658409e-01 -6.63329512e-02 -1.24978042e+00 -6.79835200e-01
3.90922934e-01 2.43603975e-01 1.06405127e+00 6.32102966e-01
-5.59330583e-01 9.34011102e-01 6.78240919e+00 4.95547473e-01
-1.43823373e+00 -2.51376927e-01 9.00885224e-01 -6.99329495e-01
-2.90014684e-01 -2.22410917e-01 -8.19056392e-01 5.94640374e-01
1.17743039e+00 -3.51012021e-01 1.20267999e+00 1.28563046e+00
3.83536816e-01 -5.94555140e-02 -1.07711065e+00 5.91030538e-01
-7.01295316e-01 -1.53368831e+00 -3.39859813e-01 3.31444174e-01
9.90525424e-01 2.80187160e-01 3.70341539e-01 7.57827103e-01
8.54081452e-01 -1.13796306e+00 7.56668508e-01 4.97124285e-01
4.05299515e-01 -1.13309467e+00 6.41271770e-01 3.44786167e-01
-4.67597246e-01 -7.88930714e-01 -3.67915154e-01 -2.80865252e-01
-1.12736218e-01 1.80640519e-01 -1.08775699e+00 -5.01646698e-01
6.45847023e-01 8.35729837e-02 -3.65577102e-01 8.61997902e-01
-1.07133299e-01 4.92353737e-01 -6.43303454e-01 -7.44305849e-01
6.71873152e-01 -2.24471450e-01 4.89528120e-01 7.08542526e-01
2.41568625e-01 -5.60481191e-01 3.03399861e-01 1.04718649e+00
-1.20243244e-02 -7.96351507e-02 -5.42506933e-01 -3.31677735e-01
6.88303709e-01 6.72681332e-01 -5.80948234e-01 -3.98128957e-01
2.69124269e-01 3.05459470e-01 6.10994756e-01 4.65770096e-01
-7.55991817e-01 -2.77974844e-01 1.02163708e+00 9.89212170e-02
4.55715656e-01 -2.19675392e-01 -5.23297727e-01 -6.08929455e-01
-6.16249517e-02 -1.07267928e+00 -5.83444983e-02 -1.14595316e-01
-9.67073977e-01 4.63247783e-02 2.35637426e-01 -5.81223190e-01
-8.13917100e-01 -5.81792712e-01 -6.06379390e-01 6.88704967e-01
-1.25790727e+00 -2.16594100e-01 1.02145553e-01 9.61107463e-02
8.44723508e-02 -1.21970408e-01 3.05217117e-01 -2.56683201e-01
-4.67574298e-01 7.50108004e-01 6.35474324e-01 -3.35138083e-01
4.00111139e-01 -1.36552680e+00 3.00518960e-01 4.72736478e-01
-5.54909348e-01 6.52097404e-01 9.96980131e-01 -5.90324938e-01
-1.34621644e+00 -8.08072329e-01 7.92202279e-02 -3.47518235e-01
8.03344488e-01 6.27664756e-03 -4.99934793e-01 2.30703637e-01
-1.19404539e-01 1.44831389e-02 -1.92482337e-01 1.49366796e-01
-1.31535053e-01 -3.84553701e-01 -1.41688669e+00 1.04496658e+00
1.08192575e+00 -2.01730698e-01 2.80178618e-02 3.75605300e-02
7.27265060e-01 -5.42434871e-01 -9.02364016e-01 1.98470518e-01
6.67590797e-01 -9.95610774e-01 8.73261869e-01 -9.49390233e-01
3.68981093e-01 1.42136231e-01 -2.84311175e-02 -1.81492400e+00
2.33862288e-02 -8.06099653e-01 -4.61473763e-01 8.52541387e-01
4.18392032e-01 -9.54505682e-01 9.13495302e-01 9.01013792e-01
-5.72331138e-02 -1.24515510e+00 -7.63853788e-01 -1.06551778e+00
4.84632820e-01 -1.71600729e-01 9.00860071e-01 4.42451358e-01
-1.39865115e-01 3.09290970e-03 -2.16534231e-02 -2.88447082e-01
7.24416733e-01 -1.25810048e-02 6.31059468e-01 -6.89762712e-01
-7.19067633e-01 -1.07298422e+00 1.05688199e-01 -9.47399199e-01
9.54823121e-02 -4.56004381e-01 1.51301980e-01 -1.26289797e+00
-3.07163119e-01 -5.69890738e-01 -2.48420566e-01 5.43654442e-01
-1.67044923e-01 -5.01312494e-01 4.55273151e-01 -1.28303781e-01
-4.17530805e-01 6.74532413e-01 1.32954061e+00 -3.25411111e-02
-9.10822868e-01 4.71672043e-02 -6.94642305e-01 4.27476138e-01
1.30121577e+00 -3.59067082e-01 -6.58826292e-01 -3.32802802e-01
2.62090445e-01 -2.71556098e-02 1.45497322e-01 -1.08603251e+00
-1.79805472e-01 -4.04552877e-01 5.46951830e-01 -8.60650986e-02
-4.93146367e-02 -4.49194372e-01 -4.13718253e-01 7.89250910e-01
-7.81279504e-01 2.48844802e-01 3.74217629e-01 4.94664460e-01
4.11662787e-01 -3.70103240e-01 8.75969470e-01 -2.20025554e-01
-4.40465838e-01 1.52011961e-01 -3.05043489e-01 3.77409607e-01
7.52557158e-01 -4.94915731e-02 -3.26305211e-01 -2.69254357e-01
-2.91744977e-01 5.28351665e-01 5.94841659e-01 2.48545066e-01
1.23918243e-01 -9.48473990e-01 -3.32051992e-01 8.45570583e-03
-3.89492273e-01 -7.23434091e-02 -3.24883670e-01 4.20561165e-01
-3.59324396e-01 1.54880613e-01 -2.04001054e-01 -4.59486455e-01
-6.84624434e-01 3.50604117e-01 9.49235439e-01 -5.03427744e-01
-3.79196674e-01 6.09700501e-01 -4.94924456e-01 -5.60019314e-01
6.84237182e-01 -6.95329487e-01 2.53205597e-01 -1.66635215e-01
4.15892690e-01 7.88796365e-01 -2.46822596e-01 5.29399812e-01
2.86128540e-02 1.16809346e-02 5.36483750e-02 -3.86371881e-01
1.36197424e+00 2.37139866e-01 2.23604038e-01 1.11348070e-01
1.07165205e+00 -3.51694822e-01 -2.26076126e+00 5.53072058e-02
2.02966571e-01 -3.60543281e-01 3.53979051e-01 -9.50478494e-01
-9.20396686e-01 4.42075491e-01 8.89452040e-01 1.60720542e-01
7.20197618e-01 -4.13266569e-01 6.74259543e-01 8.84164572e-01
3.56380820e-01 -1.53299272e+00 1.13498561e-01 7.55568027e-01
7.16333807e-01 -1.03155124e+00 2.36948691e-02 6.25581920e-01
-5.41233778e-01 9.74886179e-01 8.67344260e-01 -3.21828187e-01
3.29074293e-01 3.84865791e-01 -4.62486073e-02 1.84244812e-01
-1.00094008e+00 -1.34273082e-01 -2.46735707e-01 7.20386326e-01
-6.69000205e-03 4.71601635e-02 -2.00164586e-01 2.54346598e-02
-3.07784468e-01 4.76474427e-02 1.33445203e-01 9.41013873e-01
-6.87916279e-01 -9.91777718e-01 -3.56690407e-01 5.91662347e-01
-4.85081851e-01 -3.38835754e-02 -9.16612968e-02 9.12666380e-01
-1.35123938e-01 5.75314999e-01 3.56928766e-01 -3.25210318e-02
2.85568565e-01 -4.34131362e-02 5.94164610e-01 -1.82931319e-01
-7.95326889e-01 -2.94945449e-01 2.52598733e-01 -8.47993791e-01
4.18070108e-01 -5.44919670e-01 -1.33632302e+00 -4.53668445e-01
1.30580142e-01 9.24457014e-02 7.51192033e-01 6.86478257e-01
5.63536227e-01 4.57260162e-01 4.99227256e-01 -8.64562571e-01
-1.69693899e+00 -6.12102985e-01 -1.12681746e-01 1.80882290e-01
5.39585054e-01 -6.10469043e-01 -3.46351951e-01 -6.56783283e-01] | [4.088036060333252, 2.255612850189209] |
b4bb5ee5-1cdf-4d1b-9eee-991b16399ceb | optimal-proposal-learning-for-deployable-end | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Song_Optimal_Proposal_Learning_for_Deployable_End-to-End_Pedestrian_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Song_Optimal_Proposal_Learning_for_Deployable_End-to-End_Pedestrian_Detection_CVPR_2023_paper.pdf | Optimal Proposal Learning for Deployable End-to-End Pedestrian Detection | End-to-end pedestrian detection focuses on training a pedestrian detection model via discarding the Non-Maximum Suppression (NMS) post-processing. Though a few methods have been explored, most of them still suffer from longer training time and more complex deployment, which cannot be deployed in the actual industrial applications. In this paper, we intend to bridge this gap and propose an Optimal Proposal Learning (OPL) framework for deployable end-to-end pedestrian detection. Specifically, we achieve this goal by using CNN-based light detector and introducing two novel modules, including a Coarse-to-Fine (C2F) learning strategy for proposing precise positive proposals for the Ground-Truth (GT) instances by reducing the ambiguity of sample assignment/output in training/testing respectively, and a Completed Proposal Network (CPN) for producing extra information compensation to further recall the hard pedestrian samples. Extensive experiments are conducted on CrowdHuman, TJU-Ped and Caltech, and the results show that our proposed OPL method significantly outperforms the competing methods. | ['Honggang Zhang', 'Xuansong Xie', 'Yifeng Geng', 'Biao Wang', 'Jun-Yan He', 'Pengyu Li', 'Binghui Chen', 'Xiaolin Song'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['pedestrian-detection'] | ['computer-vision'] | [ 3.64874192e-02 -1.61854684e-01 2.58663237e-01 -4.21563625e-01
-5.72485864e-01 -1.16591163e-01 2.90184587e-01 4.92519476e-02
-6.45321131e-01 6.69696033e-01 -2.85130262e-01 -1.72161609e-01
6.21835291e-01 -7.45541930e-01 -6.60746038e-01 -7.29376912e-01
2.82258093e-01 2.13263884e-01 1.04318714e+00 -1.79196727e-02
8.52355510e-02 2.21284658e-01 -1.50048018e+00 2.52918065e-01
7.86114097e-01 9.16494310e-01 5.96165538e-01 5.25100768e-01
1.54771775e-01 3.59063357e-01 -6.29348874e-01 -7.69608140e-01
5.03381431e-01 -3.30615938e-02 -1.15928994e-02 4.17436719e-01
5.97143412e-01 -6.44363880e-01 -1.94975406e-01 1.25950634e+00
9.74945545e-01 8.17155540e-02 4.89082694e-01 -1.48139524e+00
-2.27195993e-01 1.48469239e-01 -9.92960215e-01 2.79414058e-01
1.19220644e-01 9.26135063e-01 4.39155430e-01 -1.20567083e+00
1.19055346e-01 1.63735151e+00 8.01159143e-01 6.05183303e-01
-8.05173039e-01 -6.96217239e-01 2.70571411e-01 3.76315922e-01
-1.33344781e+00 -5.29804528e-01 7.79801488e-01 -4.33984578e-01
4.94490325e-01 1.13524701e-02 8.95768166e-01 1.09189236e+00
-8.15937389e-03 1.09174263e+00 1.13257623e+00 -4.00199234e-01
2.44529620e-01 3.54436904e-01 1.76353633e-01 9.11491990e-01
6.89258873e-01 4.24038291e-01 -1.21624477e-01 1.51698634e-01
6.58721209e-01 2.11228598e-02 -6.33155555e-02 -2.16613427e-01
-7.52123415e-01 5.83421528e-01 6.33174539e-01 -3.63687038e-01
-1.63767755e-01 -1.25220507e-01 4.76284593e-01 -2.92574108e-01
3.12182665e-01 -4.64426041e-01 -3.52337360e-01 4.91825461e-01
-8.89122605e-01 3.42003167e-01 2.88682908e-01 1.22967267e+00
6.37418628e-01 9.27644074e-02 -6.39719784e-01 7.09319651e-01
6.96669221e-01 8.13993335e-01 3.01278625e-02 -5.25021017e-01
7.19945252e-01 6.13258362e-01 4.33661580e-01 -8.69578063e-01
-4.26767826e-01 -8.36692810e-01 -7.94754326e-01 3.57838005e-01
4.39330637e-01 -4.05853957e-01 -8.29386175e-01 1.21076560e+00
5.85355520e-01 2.85077631e-01 -2.81421617e-02 1.48484194e+00
1.00095999e+00 6.33496404e-01 5.92840075e-01 -1.32007584e-01
1.65154207e+00 -1.33399844e+00 -2.39994511e-01 -6.09226465e-01
2.42201984e-01 -1.00430965e+00 9.10373688e-01 4.41555321e-01
-9.55687404e-01 -1.14696705e+00 -1.10452187e+00 -1.47900224e-01
-1.62275299e-01 9.36202705e-01 5.36842108e-01 9.60214853e-01
-6.48549438e-01 -2.32651662e-02 -4.81662124e-01 -4.64522451e-01
7.00757444e-01 1.68940857e-01 -2.90606543e-03 -2.07718104e-01
-8.14431667e-01 6.04587913e-01 5.80620229e-01 6.60695374e-01
-1.13972330e+00 -3.84645700e-01 -5.26200652e-01 -7.06352741e-02
4.33067262e-01 -9.25242007e-01 1.02947366e+00 -6.63860917e-01
-1.35993266e+00 8.11498940e-01 -7.53214508e-02 -5.63495994e-01
8.30053508e-01 -2.20629379e-01 -3.55662614e-01 -1.55621290e-01
3.11616659e-01 1.05777407e+00 8.14139903e-01 -1.44079566e+00
-1.33331251e+00 -2.63518721e-01 -2.22736131e-02 1.12287119e-01
-9.20169428e-02 1.46742076e-01 -6.24023438e-01 -5.38785100e-01
1.76869668e-02 -6.76555693e-01 -4.11514223e-01 3.35649550e-01
-6.30627990e-01 -4.01119053e-01 9.01742876e-01 -6.38270259e-01
7.49641418e-01 -1.95321643e+00 -6.02414668e-01 -1.52247950e-01
4.64989729e-02 8.30310762e-01 -1.48911551e-01 -1.18113935e-01
3.75009418e-01 -4.91106570e-01 2.89419480e-02 -8.71726513e-01
8.41324218e-03 4.14815266e-03 -1.74879149e-01 4.74258363e-01
3.84454578e-01 5.94797909e-01 -9.73626196e-01 -8.44700456e-01
6.79092646e-01 4.24624741e-01 -4.66036588e-01 2.02436730e-01
-1.27989456e-01 3.03015143e-01 -4.89667684e-01 8.40523779e-01
1.32886302e+00 2.47023135e-01 -2.98944294e-01 -4.28144783e-01
-4.26402986e-01 -2.00171992e-01 -1.67685449e+00 1.17207122e+00
-1.61756799e-02 3.68450165e-01 1.17737733e-01 -7.37170696e-01
9.93769288e-01 -2.11676225e-01 -1.54434845e-01 -6.47376418e-01
2.77332842e-01 1.63226247e-01 -8.81366208e-02 -6.48289740e-01
5.32362044e-01 2.54172802e-01 2.28486478e-01 -3.12934041e-01
-1.04173049e-01 4.57383931e-01 3.64192158e-01 -4.51565869e-02
3.64524782e-01 8.80865082e-02 2.56070625e-02 -2.53982395e-01
9.81321275e-01 -2.96647418e-02 1.02563679e+00 8.02561760e-01
-7.72149801e-01 5.36825061e-01 3.10829915e-02 -6.05704546e-01
-1.09873915e+00 -1.21317065e+00 5.77111728e-02 1.04084373e+00
4.87399191e-01 -1.16067931e-01 -9.55627859e-01 -8.06443512e-01
-1.99293077e-01 5.92735350e-01 -3.22009884e-02 4.12137099e-02
-7.15118706e-01 -9.80812192e-01 4.78256255e-01 6.36455297e-01
1.07815230e+00 -8.39451909e-01 -8.34210217e-01 1.95418477e-01
-2.70625651e-01 -1.21081054e+00 -5.71081102e-01 -2.00123683e-01
-5.29680431e-01 -1.12132478e+00 -7.14452207e-01 -9.32604790e-01
6.55040264e-01 8.62352312e-01 7.81316221e-01 8.70673805e-02
-4.85919595e-01 -9.53193754e-02 -8.50998014e-02 -7.60724068e-01
-8.67668190e-04 -2.23031417e-01 4.93685938e-02 2.75843918e-01
5.46730697e-01 4.62961346e-02 -1.17081249e+00 5.11142790e-01
-3.10822099e-01 2.16006726e-01 9.40744221e-01 7.48979986e-01
4.95006621e-01 1.15706258e-01 5.07361114e-01 -4.43215907e-01
2.10397512e-01 -1.87929291e-02 -8.57935488e-01 1.27757847e-01
-3.82460684e-01 -4.04437810e-01 5.25297105e-01 -3.04646105e-01
-1.36704063e+00 2.74173409e-01 -2.90673792e-01 -2.45048553e-01
-5.39115787e-01 -4.69092399e-01 -4.88950372e-01 -1.01228639e-01
5.84827304e-01 2.78458923e-01 -3.87760699e-01 -3.86072099e-01
1.72531188e-01 7.13734746e-01 6.59612834e-01 -4.81979936e-01
7.46005774e-01 5.16891539e-01 8.93103257e-02 -8.42838049e-01
-8.62546504e-01 -6.12824917e-01 -5.25713384e-01 -5.59892893e-01
9.24208999e-01 -1.20569813e+00 -8.60723972e-01 4.49017555e-01
-1.38540566e+00 6.19222671e-02 -1.32918119e-01 5.69382191e-01
-8.25099051e-02 6.38807893e-01 -4.76497084e-01 -1.21941733e+00
-5.22463500e-01 -1.27456117e+00 1.33387327e+00 6.67605579e-01
5.34397364e-01 -3.44942808e-01 -4.44056958e-01 4.06848580e-01
2.81623363e-01 -4.42799181e-03 2.14718372e-01 -1.61121160e-01
-9.35126543e-01 -1.81740955e-01 -7.25742400e-01 3.49873453e-01
-5.09548664e-01 -3.15422192e-02 -1.18030119e+00 -5.01839519e-01
-2.17241049e-01 -7.61153772e-02 1.16544795e+00 6.11822307e-01
7.56563127e-01 -9.12448168e-02 -5.50050855e-01 4.47670847e-01
1.33350420e+00 -7.71679282e-02 5.70924282e-01 2.75467068e-01
5.25381982e-01 5.75361669e-01 8.78177583e-01 5.62581360e-01
5.52002907e-01 7.64236748e-01 5.35109997e-01 -3.40327412e-01
-6.49516702e-01 -3.78722936e-01 4.35641706e-01 6.11705147e-02
-1.01446711e-01 -2.82986790e-01 -4.50996399e-01 3.08844298e-01
-1.88986897e+00 -9.99070942e-01 -6.97086990e-01 2.18514919e+00
3.44278723e-01 6.30742192e-01 5.40714264e-01 8.72844383e-02
1.13420391e+00 -6.78074285e-02 -4.08939600e-01 2.86392003e-01
-2.29885161e-01 -1.43861949e-01 6.07564270e-01 4.46984738e-01
-1.37750506e+00 1.08131254e+00 5.00583506e+00 1.01820743e+00
-8.34022641e-01 3.03528905e-01 7.68625021e-01 2.20676854e-01
4.76538748e-01 -1.64371394e-02 -1.49158466e+00 6.03875041e-01
2.67953813e-01 5.97256780e-01 -1.24974020e-01 1.14862120e+00
5.88401437e-01 -2.44099349e-01 -7.06421733e-01 1.26424456e+00
-9.73418951e-02 -8.46039116e-01 -1.19090118e-02 -3.74849856e-01
3.36112678e-01 -2.21262917e-01 3.64836790e-02 3.70326161e-01
-3.78395654e-02 -3.02393496e-01 1.02926040e+00 5.69577515e-01
3.07404220e-01 -7.04900801e-01 8.84795129e-01 5.71989536e-01
-1.46074927e+00 -3.76108795e-01 -9.32704449e-01 -1.23954592e-02
2.69911796e-01 7.44910777e-01 -8.26271355e-01 4.99347121e-01
7.05630243e-01 3.21685255e-01 -9.49477553e-01 1.58943510e+00
-4.09252971e-01 5.64016461e-01 -2.64882118e-01 -2.73834825e-01
7.82643110e-02 -2.47194734e-03 6.96594834e-01 1.62490273e+00
1.39055058e-01 6.64603710e-03 7.44694591e-01 8.36741269e-01
3.87849420e-01 -6.09764233e-02 5.99140674e-02 8.33664238e-01
4.49036926e-01 1.52236235e+00 -9.05687392e-01 -5.32711804e-01
-2.94901967e-01 8.88076901e-01 1.77706443e-02 2.53965169e-01
-1.00233567e+00 -2.34182864e-01 3.02975923e-01 3.53691518e-01
2.49296591e-01 -2.03341246e-01 -1.50516495e-01 -9.53778386e-01
4.03155118e-01 -4.09845829e-01 3.25711668e-01 -7.28084445e-01
-1.54479408e+00 4.34615612e-01 -1.15034752e-01 -1.34755063e+00
2.64562488e-01 -7.58601487e-01 -6.55956686e-01 9.37872410e-01
-1.74708045e+00 -1.57750785e+00 -7.00979710e-01 5.85923791e-01
8.88751626e-01 -1.74655795e-01 8.27792436e-02 7.47679293e-01
-8.40955794e-01 8.46475422e-01 -5.38181663e-01 2.97275633e-01
7.45011806e-01 -9.10483301e-01 2.72766501e-01 1.37023950e+00
-4.75537270e-01 2.84153789e-01 8.20949197e-01 -6.57757461e-01
-1.04337907e+00 -1.57085288e+00 6.17252052e-01 -9.23832804e-02
8.33801087e-03 -3.40875387e-01 -3.19040656e-01 4.04307514e-01
1.40532300e-01 3.50228369e-01 1.30876631e-01 -4.06438857e-01
-6.68310300e-02 -5.27853847e-01 -1.33011925e+00 6.48561776e-01
1.18267334e+00 2.62697875e-01 -3.86861622e-01 4.91287589e-01
5.14099061e-01 -3.38271230e-01 3.92679088e-02 3.91042709e-01
4.65262860e-01 -1.19769311e+00 1.25033128e+00 -2.05958202e-01
-1.25324488e-01 -1.02195334e+00 -1.88448474e-01 -8.56657624e-01
-5.16164780e-01 -3.66590887e-01 2.26226784e-02 1.46644258e+00
-1.60909727e-01 -5.62412858e-01 9.46648777e-01 2.89075494e-01
-4.74515289e-01 -4.28591877e-01 -8.93967092e-01 -7.21442878e-01
-4.38289702e-01 -3.35762501e-01 4.15848106e-01 2.69063652e-01
-6.95742905e-01 3.82743835e-01 -4.91585791e-01 6.45182312e-01
1.35516822e+00 -1.79339662e-01 1.19089806e+00 -1.16091454e+00
-3.15173805e-01 -3.85854244e-01 -6.20298505e-01 -1.45398009e+00
-3.98016840e-01 -3.53161216e-01 3.25168043e-01 -1.45565224e+00
3.31722885e-01 -5.07996261e-01 8.53592381e-02 1.45289868e-01
-3.87175798e-01 3.11787397e-01 2.51037776e-01 6.12795800e-02
-9.07877088e-01 4.75981504e-01 1.08350456e+00 -2.84228116e-01
-2.44970191e-02 3.56919914e-01 -4.16710377e-01 9.13146496e-01
5.67513347e-01 -2.92073250e-01 -2.04881817e-01 -3.77832651e-01
-4.17773813e-01 -1.24370344e-01 1.08567214e+00 -1.64142513e+00
5.19841135e-01 7.54062012e-02 8.62509608e-01 -1.21648777e+00
3.56385887e-01 -7.38125622e-01 -3.50984335e-01 7.38643169e-01
8.59420374e-02 -1.20406963e-01 6.57603219e-02 7.85009265e-01
1.75540924e-01 -2.46277392e-01 1.12010407e+00 -1.61096036e-01
-9.50217485e-01 3.39378655e-01 -1.53320193e-01 -1.81933716e-01
1.23830163e+00 -4.00314540e-01 -5.06388545e-01 3.96542490e-01
-5.03324270e-01 4.10221428e-01 8.54092371e-03 2.51733065e-01
8.15552890e-01 -1.21688139e+00 -9.36834633e-01 4.86221880e-01
1.29185021e-01 2.54070833e-02 6.48588002e-01 6.90447152e-01
-3.92147988e-01 4.48325783e-01 -1.88149825e-01 -8.57468069e-01
-1.25155282e+00 6.56271040e-01 3.81257266e-01 -2.52575800e-03
-6.51979506e-01 8.86079669e-01 3.59544903e-01 -2.27848426e-01
4.33388919e-01 -2.51560032e-01 -1.72442541e-01 -3.13747197e-01
6.70962334e-01 6.70594871e-01 -1.75787032e-01 -6.17628634e-01
-3.52773279e-01 3.92322749e-01 -1.17696121e-01 1.46358475e-01
8.40842783e-01 -3.48940074e-01 2.59088576e-01 -1.75310418e-01
6.09830081e-01 -1.62200674e-01 -1.69805694e+00 -3.63851637e-02
-4.42696273e-01 -6.57402277e-01 -1.77192077e-01 -6.21091187e-01
-1.00267661e+00 1.04767823e+00 1.22011912e+00 -1.02811411e-01
9.84336019e-01 -3.51615697e-01 9.61265802e-01 3.76412958e-01
4.68164384e-01 -1.16489363e+00 9.51573178e-02 2.82246888e-01
4.25007880e-01 -1.35672045e+00 -6.71551656e-03 -7.69271851e-01
-3.04318696e-01 1.19088113e+00 1.09021783e+00 -2.60773361e-01
5.17886162e-01 8.77775550e-02 -1.94209620e-01 1.76989034e-01
-3.00785482e-01 -4.86373544e-01 -3.62543687e-02 9.77891386e-01
-3.51663865e-02 1.08576283e-01 -3.43729913e-01 7.13081598e-01
6.00663014e-02 1.37864780e-02 2.52796203e-01 6.24141991e-01
-1.04383385e+00 -7.96170413e-01 -6.77773654e-01 2.24645466e-01
5.04507385e-02 -1.53532124e-03 6.52778596e-02 6.70236170e-01
9.39743638e-01 1.30856240e+00 -2.00747401e-01 -3.30823064e-01
5.51409543e-01 -2.94467896e-01 4.33447450e-01 -3.65499496e-01
-3.78350586e-01 2.03659847e-01 -1.44432876e-02 -2.30038360e-01
-4.73253548e-01 -6.74824834e-01 -8.62487674e-01 -1.78502619e-01
-5.43046892e-01 -1.65012687e-01 2.95175344e-01 8.31500053e-01
2.54128695e-01 5.94429672e-01 4.03009534e-01 -1.15963864e+00
-5.09300709e-01 -1.04635084e+00 -3.11633378e-01 2.77938396e-01
8.81947502e-02 -9.32919502e-01 4.74281497e-02 1.34517267e-01] | [8.068449020385742, -0.5954602360725403] |
ec181da1-ef20-4dc5-82e0-8dca8b065749 | multilingual-controllable-transformer-based | 2307.02120 | null | https://arxiv.org/abs/2307.02120v1 | https://arxiv.org/pdf/2307.02120v1.pdf | Multilingual Controllable Transformer-Based Lexical Simplification | Text is by far the most ubiquitous source of knowledge and information and should be made easily accessible to as many people as possible; however, texts often contain complex words that hinder reading comprehension and accessibility. Therefore, suggesting simpler alternatives for complex words without compromising meaning would help convey the information to a broader audience. This paper proposes mTLS, a multilingual controllable Transformer-based Lexical Simplification (LS) system fined-tuned with the T5 model. The novelty of this work lies in the use of language-specific prefixes, control tokens, and candidates extracted from pre-trained masked language models to learn simpler alternatives for complex words. The evaluation results on three well-known LS datasets -- LexMTurk, BenchLS, and NNSEval -- show that our model outperforms the previous state-of-the-art models like LSBert and ConLS. Moreover, further evaluation of our approach on the part of the recent TSAR-2022 multilingual LS shared-task dataset shows that our model performs competitively when compared with the participating systems for English LS and even outperforms the GPT-3 model on several metrics. Moreover, our model obtains performance gains also for Spanish and Portuguese. | ['Horacio Saggion', 'Kim Cheng SHEANG'] | 2023-07-05 | null | null | null | null | ['lexical-simplification', 'reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.48154892e-04 3.22729230e-01 -2.03705013e-01 -7.31729120e-02
-1.19224513e+00 -5.18261671e-01 7.68132687e-01 6.24613166e-01
-9.91315544e-01 1.00418377e+00 7.15243876e-01 -5.77489793e-01
4.65893447e-02 -4.94884968e-01 -7.13864386e-01 -1.24001160e-01
5.52212775e-01 8.13092113e-01 3.09255123e-01 -9.94067967e-01
1.28576398e-01 1.66081309e-01 -1.48545706e+00 5.46781301e-01
1.38625157e+00 4.51142102e-01 6.37456179e-01 1.69102579e-01
-4.24381077e-01 7.60315359e-01 -6.50104046e-01 -7.38315642e-01
1.08528279e-01 -1.48378193e-01 -8.92002821e-01 -5.64003289e-01
7.38395512e-01 6.15852885e-02 8.29926580e-02 8.19140792e-01
7.89590716e-01 2.05063865e-01 5.59011340e-01 -2.51348525e-01
-7.98426867e-01 1.29567218e+00 -1.14884146e-01 2.23714590e-01
7.24862993e-01 2.00309362e-02 1.27982008e+00 -1.09760892e+00
8.89897406e-01 1.58155775e+00 6.97497308e-01 4.09437567e-01
-1.20532382e+00 -3.15577120e-01 2.65740782e-01 5.71688414e-01
-1.63731122e+00 -5.70804954e-01 3.40164572e-01 2.32956819e-02
1.66011810e+00 5.72242975e-01 5.08554220e-01 1.34758759e+00
9.73796621e-02 1.06350017e+00 1.53486824e+00 -1.24705708e+00
-2.96605796e-01 3.23950648e-01 6.08868524e-02 5.53866148e-01
1.75656006e-01 -2.52319694e-01 -7.79750466e-01 4.28141028e-01
-9.77348238e-02 -5.44128656e-01 -5.40683210e-01 1.80054501e-01
-1.42916572e+00 6.68290377e-01 -5.88721188e-04 6.15810812e-01
-2.47333065e-01 -3.58770370e-01 4.66324449e-01 5.60266852e-01
6.71471477e-01 7.26570785e-01 -8.75774503e-01 -1.48142174e-01
-1.06566632e+00 4.03047174e-01 8.54824483e-01 1.06086957e+00
5.34340203e-01 -7.40481317e-02 -6.43568277e-01 1.10881245e+00
-9.82792825e-02 9.62196767e-01 5.74443579e-01 -4.36406821e-01
7.88374424e-01 4.53150004e-01 -1.12746060e-01 -6.48842871e-01
-3.80300462e-01 -6.98855579e-01 -5.94919443e-01 -2.98329115e-01
4.96948808e-01 1.50638789e-01 -5.86339235e-01 1.69351757e+00
-6.39736131e-02 -7.01267600e-01 2.19294569e-03 4.26122159e-01
9.28385079e-01 7.33397186e-01 1.70935050e-01 -1.83716655e-01
1.34722698e+00 -1.08509350e+00 -1.07684255e+00 -3.19081098e-01
9.18561161e-01 -1.19582927e+00 1.81298256e+00 5.64531267e-01
-1.29190230e+00 -5.68154991e-01 -9.29496884e-01 -7.24188983e-01
-9.14330363e-01 4.30455953e-01 7.95527101e-02 6.83876872e-01
-9.55098569e-01 4.38817143e-01 -3.40374500e-01 -7.10015714e-01
8.29363149e-03 3.62318126e-03 -3.09991509e-01 -3.66529226e-01
-1.61110055e+00 1.63901317e+00 5.32932520e-01 -1.42884612e-01
-4.13460821e-01 -8.44570875e-01 -8.50329280e-01 8.98512779e-04
4.16687191e-01 -6.57942653e-01 1.36218834e+00 -5.99819720e-01
-1.69241023e+00 1.06318748e+00 -4.04605657e-01 -7.03553021e-01
6.93935215e-01 -8.17890227e-01 -4.78091121e-01 -3.29209656e-01
2.52832890e-01 7.30014622e-01 6.91039383e-01 -7.16875970e-01
-7.21978962e-01 1.52712420e-01 -3.10431533e-02 5.05470157e-01
-3.57574970e-01 9.62196663e-02 -2.26346269e-01 -9.98119116e-01
-4.72046942e-01 -7.34612346e-01 8.47556666e-02 -6.05077028e-01
-6.18653893e-01 -4.57964361e-01 3.42671245e-01 -1.26643920e+00
1.74435520e+00 -1.79081643e+00 5.00514328e-01 -6.55036941e-02
-1.21395953e-01 6.39624774e-01 -1.67606071e-01 8.68362248e-01
3.90633233e-02 3.08051229e-01 -4.58679497e-02 -5.30770838e-01
2.41568819e-01 1.28066495e-01 -3.63733381e-01 7.19387606e-02
-9.49883685e-02 1.20452917e+00 -8.52742791e-01 -4.06499028e-01
5.20971835e-01 2.69166201e-01 -3.93193185e-01 -1.03619121e-01
-4.93956685e-01 3.62071723e-01 -2.72933934e-02 4.50107515e-01
2.69039333e-01 2.48189881e-01 -7.56892934e-02 -1.76355273e-01
-4.74116892e-01 1.00843048e+00 -7.62233913e-01 1.86573994e+00
-9.77203310e-01 6.49072945e-01 -6.12990916e-01 -4.07145053e-01
5.76618373e-01 2.32565463e-01 -2.39676878e-01 -1.23571229e+00
1.40312448e-01 5.97559929e-01 -1.78211294e-02 -3.93844754e-01
7.86743283e-01 1.32637039e-01 -3.11180383e-01 2.26239800e-01
8.88023376e-02 -3.89007002e-01 5.73596895e-01 1.31191939e-01
9.18550014e-01 2.29962230e-01 7.74908006e-01 -6.75120831e-01
9.64362621e-01 -8.50788355e-02 7.20980912e-02 9.56513941e-01
2.81102628e-01 3.14014763e-01 2.63683889e-02 -2.11240560e-01
-9.76612866e-01 -1.00733364e+00 -1.46757886e-01 1.40375853e+00
-3.24711144e-01 -8.05504382e-01 -1.00461411e+00 -6.40131056e-01
-1.53550997e-01 1.62278211e+00 -2.94479817e-01 5.85723929e-02
-8.87390137e-01 -3.57947469e-01 7.31829762e-01 -6.56894296e-02
4.47233588e-01 -1.21742630e+00 -2.34800577e-01 3.94715875e-01
-7.14007318e-01 -1.18627381e+00 -3.39430958e-01 5.83339557e-02
-3.14792126e-01 -7.43701100e-01 -6.90582156e-01 -8.56863856e-01
1.20495312e-01 1.63241699e-02 1.48340344e+00 -1.14281978e-02
-8.60381126e-02 1.02469856e-02 -6.14886999e-01 -8.14518809e-01
-6.17601216e-01 7.61909664e-01 -2.12349027e-01 -3.41238141e-01
5.60949981e-01 -1.15104653e-01 1.41541800e-02 -2.80540735e-01
-7.17897952e-01 2.37421781e-01 6.25441551e-01 6.80280447e-01
6.72456801e-01 -4.05573398e-01 5.88217735e-01 -1.20869017e+00
9.04692054e-01 -1.17528088e-01 -2.73143262e-01 6.45071387e-01
-5.21730840e-01 3.03359658e-01 9.06044364e-01 -3.01223427e-01
-1.24165320e+00 -5.04207790e-01 -5.46749890e-01 2.87546754e-01
-7.80222863e-02 6.39239550e-01 -4.90747958e-01 1.11963138e-01
7.96957672e-01 2.66802490e-01 -4.22595859e-01 -9.28233325e-01
8.97171199e-01 5.24902046e-01 2.75664955e-01 -5.59214890e-01
7.11985290e-01 -4.27685976e-02 -2.58386612e-01 -1.16693187e+00
-1.26747775e+00 -3.08096319e-01 -5.67927301e-01 5.07604033e-02
5.13168156e-01 -9.20462251e-01 -5.54346666e-02 6.17695451e-01
-1.39183092e+00 -2.71434963e-01 -6.19144857e-01 1.59965992e-01
-1.63484842e-01 3.36963356e-01 -5.47055900e-01 -2.35960484e-01
-4.67940778e-01 -1.13486564e+00 1.09761035e+00 -1.87854752e-01
-6.84814572e-01 -1.20799446e+00 4.99231145e-02 3.77310723e-01
7.96932995e-01 -1.73682645e-01 1.43517756e+00 -9.00470614e-01
-3.61709148e-01 -1.72972283e-03 4.47396562e-02 5.85150957e-01
-5.53435739e-03 -2.69946396e-01 -8.47906530e-01 -1.19645134e-01
-3.03344727e-01 -4.86975193e-01 8.86547804e-01 2.20143005e-01
9.53716516e-01 -4.37974721e-01 -3.18488330e-02 3.73578995e-01
1.13066244e+00 -4.37874883e-01 6.73209488e-01 5.56316257e-01
8.06140542e-01 6.27767384e-01 6.18719697e-01 7.53574148e-02
7.88953841e-01 7.59104788e-01 3.86082493e-02 -1.60278812e-01
-7.17265129e-01 -4.54257280e-01 6.25709653e-01 1.65102375e+00
9.37818289e-02 -5.91197312e-01 -8.35807323e-01 5.93312323e-01
-1.56609964e+00 -5.86736560e-01 -4.15437996e-01 2.09732461e+00
1.22364044e+00 2.05546632e-01 -1.71060666e-01 2.05432057e-01
3.35002005e-01 2.18927950e-01 1.41791612e-01 -6.33235753e-01
-8.87779415e-01 8.12439740e-01 5.88265777e-01 9.10603225e-01
-8.08779120e-01 1.73106802e+00 5.69250822e+00 1.64893901e+00
-9.37275052e-01 4.26674932e-01 2.64000148e-01 -5.46901524e-02
-5.37634969e-01 6.14405749e-03 -1.14188838e+00 1.81158558e-01
9.90514457e-01 -3.41119200e-01 5.82569897e-01 3.38424236e-01
2.15648726e-01 -2.21450791e-01 -1.08634686e+00 7.82474935e-01
3.16701055e-01 -1.15827417e+00 6.53961658e-01 -5.19149303e-01
7.59778917e-01 2.13895604e-01 1.83706969e-01 6.46456361e-01
1.44671217e-01 -1.11524999e+00 1.24578249e+00 7.01428711e-01
8.69840860e-01 -8.01223218e-01 7.13856637e-01 5.56069434e-01
-9.02491808e-01 3.10859412e-01 -3.46810728e-01 -2.35147610e-01
1.46797687e-01 6.37909055e-01 -6.51439726e-01 6.61825538e-01
5.03016949e-01 6.87149107e-01 -1.12924099e+00 7.59391427e-01
-9.60415184e-01 8.44013214e-01 -2.58954376e-01 -3.46293658e-01
2.54210770e-01 1.48730174e-01 8.74292076e-01 1.88792670e+00
2.99478531e-01 -4.26589519e-01 -1.79927181e-02 4.74346548e-01
-1.04816467e-01 9.86806512e-01 -5.46721458e-01 1.06334947e-01
5.71540475e-01 1.00733769e+00 -1.77417830e-01 -3.68918598e-01
-4.37157482e-01 9.64929521e-01 7.42400229e-01 1.76788360e-01
-4.69912678e-01 -5.66550910e-01 1.66140586e-01 1.58489168e-01
2.30053768e-01 -1.09458111e-01 -2.84762800e-01 -1.17608821e+00
1.68406039e-01 -1.58068097e+00 2.65209883e-01 -6.00938261e-01
-1.21877992e+00 8.65282714e-01 5.53400256e-02 -8.88703406e-01
-2.61333376e-01 -6.80721283e-01 -3.77384014e-02 1.20683062e+00
-1.80302417e+00 -1.49170578e+00 2.02032074e-01 4.96186495e-01
8.77824605e-01 -3.28652799e-01 1.10701418e+00 3.47262532e-01
-3.54423434e-01 8.62123132e-01 2.90950060e-01 -3.30396891e-01
9.13654983e-01 -1.33122742e+00 5.13626814e-01 9.91028130e-01
2.89328903e-01 5.50835907e-01 8.27323794e-01 -5.78007698e-01
-9.51812029e-01 -1.08529282e+00 1.88892984e+00 -6.85995281e-01
7.80302823e-01 -5.76900303e-01 -9.87339735e-01 7.03832448e-01
8.57202828e-01 -7.90503383e-01 3.68799716e-01 1.92827612e-01
-2.87639409e-01 -1.19995981e-01 -8.43947530e-01 1.11086476e+00
1.05137539e+00 -5.91067314e-01 -1.07331109e+00 5.82358718e-01
9.61524248e-01 -4.51487422e-01 -6.15114689e-01 2.69890159e-01
1.91370711e-01 -8.24402869e-01 7.89816558e-01 -4.17806417e-01
1.56912193e-01 -1.74129196e-02 -2.32750729e-01 -1.87735677e+00
-1.78370655e-01 -9.39968407e-01 1.47981659e-01 1.26743746e+00
8.44080031e-01 -5.40689647e-01 -5.46990894e-02 -8.30791593e-02
-2.75478393e-01 -5.72368205e-01 -1.19698524e+00 -8.12870979e-01
5.28696537e-01 -6.17220879e-01 6.77819371e-01 6.90583527e-01
-8.89260024e-02 8.05079222e-01 -1.41683355e-01 -1.85348079e-01
3.55046719e-01 -4.79243279e-01 6.08120024e-01 -9.29475546e-01
-2.17929065e-01 -5.79395950e-01 1.28286332e-01 -1.10165167e+00
4.97693032e-01 -1.22004068e+00 -1.87266186e-01 -1.48779798e+00
-7.63177946e-02 -3.47559541e-01 -7.73339812e-03 5.57314277e-01
-3.35628569e-01 1.54045774e-02 2.55633175e-01 -7.48128220e-02
-6.60632908e-01 6.70842648e-01 1.24227607e+00 -9.92405266e-02
1.53296674e-02 -3.10190290e-01 -4.78298604e-01 7.05243647e-01
6.76553726e-01 -4.70831096e-01 -1.56403854e-01 -8.22577655e-01
5.23473442e-01 -3.95859331e-01 -7.58636147e-02 -8.65307570e-01
6.70041069e-02 2.07530320e-01 -2.76502948e-02 -7.47763336e-01
1.50826067e-01 -5.34806430e-01 -1.54357508e-01 4.59238827e-01
-3.69459242e-01 3.15030068e-01 4.49564606e-01 -1.66710883e-01
-1.35385156e-01 -4.16580051e-01 5.12591898e-01 -1.58191010e-01
-6.61388576e-01 -3.24680120e-01 -6.82058454e-01 6.28897905e-01
5.01183867e-01 1.37088433e-01 -4.13755327e-01 -3.99762988e-01
-3.75225604e-01 -4.21525761e-02 2.40207478e-01 6.84948444e-01
2.25528702e-01 -9.52259600e-01 -1.15186965e+00 1.58294812e-01
3.54529411e-01 -2.76052773e-01 -4.03136499e-02 6.66174591e-01
-5.90709984e-01 9.56670463e-01 -1.55204313e-03 -2.26400584e-01
-1.07210386e+00 3.70485753e-01 2.43637249e-01 -7.83274710e-01
-4.82323080e-01 9.06856537e-01 -1.29627347e-01 -9.07140017e-01
2.45660007e-01 -6.28368020e-01 -2.88612574e-01 1.07142560e-01
5.75680435e-01 6.41940892e-01 6.40807033e-01 -7.45396078e-01
-1.22867532e-01 3.76239091e-01 -5.01747012e-01 -3.09317321e-01
1.07364190e+00 -5.03273129e-01 -4.26293820e-01 5.87548912e-01
8.25704396e-01 6.41388059e-01 -3.35609674e-01 -6.36927724e-01
4.66271549e-01 3.67635898e-02 2.29623560e-02 -1.38317037e+00
-4.01212782e-01 1.01283765e+00 1.37213022e-01 -1.82307720e-01
1.00214231e+00 -1.64545700e-01 9.00330663e-01 7.58836925e-01
4.65944469e-01 -1.32967341e+00 -2.70806074e-01 1.16577923e+00
1.12382650e+00 -1.00963342e+00 -1.60254031e-01 -4.00356531e-01
-5.95348597e-01 8.78849626e-01 3.09083551e-01 2.23196447e-01
2.69954234e-01 -1.23777837e-01 2.24050045e-01 1.06567800e-01
-6.81932092e-01 -3.49633276e-01 7.17889905e-01 5.22159994e-01
7.34339714e-01 2.63161123e-01 -7.80724883e-01 5.60168386e-01
-7.83958495e-01 -3.85854751e-01 3.36995304e-01 7.46026933e-01
-4.73575175e-01 -1.53644371e+00 -1.97981149e-01 2.63489872e-01
-6.59410894e-01 -1.00506067e+00 -4.77384210e-01 1.12244225e+00
2.34818220e-01 8.66089702e-01 -5.54730177e-01 7.74496868e-02
7.33055770e-01 2.66274601e-01 6.98244095e-01 -9.54216480e-01
-9.32450175e-01 -5.55819273e-02 4.83719736e-01 -4.47237313e-01
-2.49735713e-02 -7.63644457e-01 -8.79049599e-01 -3.99873495e-01
-2.17962578e-01 -1.34646650e-02 6.14500761e-01 1.24375927e+00
6.92835972e-02 5.83742917e-01 8.88620093e-02 -6.23078167e-01
-5.94267845e-01 -1.30669200e+00 -2.11659133e-01 2.59273827e-01
1.27592802e-01 -3.37510079e-01 -1.75684139e-01 -7.32667223e-02] | [10.951156616210938, 10.388480186462402] |
15bfd45a-0202-4848-aba7-3e0481d646b3 | radon-features-and-barcodes-for-medical-image | 1604.04675 | null | http://arxiv.org/abs/1604.04675v1 | http://arxiv.org/pdf/1604.04675v1.pdf | Radon Features and Barcodes for Medical Image Retrieval via SVM | For more than two decades, research has been performed on content-based image
retrieval (CBIR). By combining Radon projections and the support vector
machines (SVM), a content-based medical image retrieval method is presented in
this work. The proposed approach employs the normalized Radon projections with
corresponding image category labels to build an SVM classifier, and the Radon
barcode database which encodes every image in a binary format is also generated
simultaneously to tag all images. To retrieve similar images when a query image
is given, Radon projections and the barcode of the query image are generated.
Subsequently, the k-nearest neighbor search method is applied to find the
images with minimum Hamming distance of the Radon barcode within the same class
predicted by the trained SVM classifier that uses Radon features. The
performance of the proposed method is validated by using the IRMA 2009 dataset
with 14,410 x-ray images in 57 categories. The results demonstrate that our
method has the capacity to retrieve similar responses for the correctly
identified query image and even for those mistakenly classified by SVM. The
approach further is very fast and has low memory requirement. | ['H. R. Tizhoosh', 'Shujin Zhu'] | 2016-04-16 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 6.19009316e-01 -1.22656502e-01 -5.18508434e-01 -5.01526892e-01
-1.27479231e+00 -3.19248676e-01 5.73780715e-01 6.57354653e-01
-4.96083707e-01 2.93690473e-01 1.33813322e-01 -3.66640896e-01
-4.79079157e-01 -1.02992892e+00 -1.65951490e-01 -9.01499212e-01
1.50403157e-01 3.63504261e-01 5.04504502e-01 2.17182979e-01
7.26468384e-01 5.66133916e-01 -1.76308167e+00 6.75604820e-01
5.23923516e-01 1.38607800e+00 5.07176518e-01 6.14686906e-01
-1.96108654e-01 9.71195102e-01 -4.73772138e-01 -1.31040839e-02
1.92581818e-01 -4.28431243e-01 -8.61629069e-01 -1.14907682e-01
4.25158292e-01 -2.97719806e-01 -6.11411035e-01 9.76989865e-01
4.75206167e-01 2.80975610e-01 9.96391296e-01 -7.14933395e-01
-9.43011642e-01 2.22382098e-01 -2.51335531e-01 5.52657962e-01
5.07499576e-01 -6.31012976e-01 7.18344271e-01 -9.47517097e-01
4.61028188e-01 7.79012740e-01 5.60232759e-01 3.25838715e-01
-5.88146627e-01 -4.69792843e-01 -6.68145061e-01 5.19917965e-01
-1.38858140e+00 1.66956753e-01 8.01788867e-01 -5.79289377e-01
5.62880576e-01 9.12404180e-01 5.36262333e-01 2.80980766e-01
5.49346864e-01 5.24450064e-01 1.50526023e+00 -6.73787177e-01
4.44832653e-01 5.57267010e-01 6.05868936e-01 1.00245881e+00
-2.58465502e-02 3.23745869e-02 -4.09944318e-02 -6.15394294e-01
3.83212298e-01 5.72000802e-01 -1.02516122e-01 -2.99956799e-01
-9.23036098e-01 1.14524090e+00 7.90294766e-01 5.97690880e-01
-3.67412835e-01 -2.28551045e-01 3.92229497e-01 -1.27982795e-01
2.51694232e-01 1.95060223e-01 5.37638403e-02 5.05844891e-01
-1.03466296e+00 -2.44194582e-01 4.52056259e-01 4.22772378e-01
3.42467099e-01 -5.20204961e-01 -3.64119589e-01 1.12701368e+00
4.53918308e-01 7.90635645e-01 1.05954826e+00 -4.49672490e-01
1.24242730e-01 5.92859209e-01 -3.82542133e-01 -1.43230712e+00
-2.07142934e-01 -9.69145298e-02 -5.87062299e-01 2.06528269e-02
-2.86082197e-02 8.29163253e-01 -1.16262686e+00 6.44536555e-01
2.69907534e-01 9.53763723e-04 4.70183700e-01 9.07127798e-01
1.46062362e+00 7.04226375e-01 5.09346984e-02 2.17564614e-03
1.66730201e+00 -1.09383416e+00 -4.83830452e-01 1.24905780e-01
5.23621082e-01 -8.38670433e-01 7.98795402e-01 1.58394933e-01
-5.70630610e-01 -6.16448104e-01 -1.10293806e+00 2.85658658e-01
-4.89692122e-01 6.15886092e-01 4.27719861e-01 7.59586751e-01
-7.23509729e-01 3.30628246e-01 -5.68166077e-01 -4.77420568e-01
2.91125864e-01 3.11158717e-01 -4.42916662e-01 -4.52942371e-01
-9.86649930e-01 1.07097363e+00 4.82579231e-01 -1.81881934e-01
-2.42187142e-01 -3.20789397e-01 -9.96542931e-01 -8.86252299e-02
-2.79764086e-01 -1.29251392e-03 7.86793292e-01 -7.06968307e-01
-9.17829096e-01 1.08816636e+00 5.95320128e-02 -2.99111575e-01
3.24052542e-01 3.87653202e-01 -6.77966714e-01 8.60286653e-01
3.29692900e-01 4.28131133e-01 5.16581476e-01 -1.22245228e+00
-6.43013537e-01 -5.20770133e-01 -4.17161882e-01 2.37681121e-01
-2.87936836e-01 1.89695269e-01 -3.13361138e-01 -4.61878687e-01
8.37389529e-01 -8.26160610e-01 -1.83971617e-02 -7.57236555e-02
-2.95975387e-01 -2.94683188e-01 7.66076744e-01 -8.57533634e-01
9.86621439e-01 -2.18408227e+00 -5.85115135e-01 8.03159177e-01
-2.26880074e-01 1.09391995e-01 1.35174140e-01 2.50097305e-01
-2.14064151e-01 -3.64315301e-01 -3.75340283e-01 3.28737229e-01
-6.20050132e-01 2.07663268e-01 -4.48871940e-01 5.86412311e-01
-1.72100678e-01 6.67640150e-01 -6.70086563e-01 -9.72242177e-01
4.85270917e-01 5.68436205e-01 -5.87438084e-02 6.03145435e-02
5.38004577e-01 9.16287899e-02 -5.59131444e-01 7.53718436e-01
7.58281887e-01 -5.06563246e-01 -7.19980821e-02 -4.96043205e-01
3.18724424e-01 -3.89469564e-01 -9.17235434e-01 1.12575924e+00
-3.60941082e-01 4.73751873e-01 -9.31330800e-01 -1.33004940e+00
1.19062448e+00 2.90126115e-01 7.82607675e-01 -1.06228626e+00
1.13018714e-01 5.74202240e-01 -4.56536293e-01 -8.61962676e-01
3.53403777e-01 5.23369526e-03 -2.82772239e-02 3.65482956e-01
-2.78710991e-01 1.51810959e-01 -1.60481215e-01 -8.61291885e-02
1.00239193e+00 -3.64418268e-01 4.54503000e-01 -5.47128357e-02
9.81607318e-01 5.19724488e-01 -2.17968941e-01 7.47564971e-01
-9.23766494e-02 7.34959424e-01 -3.94693851e-01 -5.37007630e-01
-9.92658973e-01 -1.16388965e+00 -8.26680243e-01 7.07992256e-01
3.78077030e-01 2.84326613e-01 -6.39327466e-01 -5.93563974e-01
6.22476935e-02 5.81814468e-01 -6.32855594e-01 -1.62445888e-01
-5.05643725e-01 -6.36197388e-01 2.32004434e-01 3.50416332e-01
6.30558670e-01 -1.13025749e+00 -8.02300453e-01 -8.81988406e-02
-9.59375277e-02 -7.84452558e-01 -2.28713185e-01 3.03003918e-02
-8.97240162e-01 -1.37421894e+00 -1.10192704e+00 -1.26570356e+00
1.26003432e+00 4.18219566e-01 2.68599331e-01 1.13741875e-01
-8.06115687e-01 5.71357250e-01 -4.87185001e-01 4.06005308e-02
-6.24336421e-01 -4.58346426e-01 -2.39759013e-01 -1.45543337e-01
5.46718419e-01 3.36196482e-01 -7.88447917e-01 2.80173689e-01
-9.75635886e-01 -2.50679791e-01 7.08852768e-01 9.86380458e-01
8.92728806e-01 2.28936628e-01 2.47888699e-01 -7.46204257e-01
4.16982442e-01 -5.16105771e-01 -4.29246604e-01 5.69094777e-01
-8.15616190e-01 -1.40240863e-01 4.43375051e-01 -3.99646342e-01
-6.40574396e-01 3.64195891e-02 -3.90123315e-02 -1.11099653e-01
-4.86241914e-02 6.54700220e-01 5.90509892e-01 -2.64768571e-01
8.89049530e-01 8.52821350e-01 2.66799688e-01 -3.21028978e-01
4.12311628e-02 1.36160553e+00 7.65448987e-01 2.26632990e-02
5.85465014e-01 5.96643031e-01 1.24669515e-01 -7.78880656e-01
-7.15894997e-01 -9.81488228e-01 -2.39618987e-01 -2.18115881e-01
9.97803271e-01 -8.61645579e-01 -3.61996800e-01 1.51492596e-01
-7.02982247e-01 6.21663094e-01 -4.82000373e-02 9.26645279e-01
-5.43451667e-01 4.98808175e-01 -5.75751007e-01 -8.84833872e-01
-6.29961014e-01 -1.38152528e+00 9.51897502e-01 3.17187518e-01
-6.69255061e-03 -5.45761704e-01 1.59372121e-01 7.59258628e-01
3.82920951e-01 -2.59922761e-02 1.16942275e+00 -1.18821084e+00
-3.51691663e-01 -8.18370700e-01 -5.00577271e-01 2.63474137e-01
2.87177294e-01 -3.88601869e-01 -7.80206621e-01 -2.33189240e-01
2.51119792e-01 3.31945601e-03 8.71831298e-01 2.63501585e-01
1.34504724e+00 -3.48597676e-01 -6.16292000e-01 1.60182551e-01
1.79123878e+00 7.76764452e-01 6.68344080e-01 6.02882385e-01
2.71411300e-01 4.12345022e-01 9.99418318e-01 2.81244032e-02
7.34314919e-02 5.94802737e-01 1.64141163e-01 -2.10778028e-01
6.37736991e-02 -1.57986879e-01 -2.51864016e-01 7.10213542e-01
2.31116965e-01 2.30144054e-01 -9.65696454e-01 3.40941370e-01
-1.37695158e+00 -1.18815601e+00 -9.44301952e-03 2.15173197e+00
6.72172427e-01 -3.71737331e-01 -2.70454049e-01 6.56178892e-01
8.30824554e-01 -9.04635862e-02 -9.44480896e-02 -2.36009285e-01
1.90700680e-01 5.61228573e-01 6.55338168e-01 3.99385929e-01
-1.39658523e+00 3.74891192e-01 6.46175385e+00 1.01229036e+00
-1.43452013e+00 1.09636262e-01 6.94226086e-01 4.75638688e-01
-2.67781556e-01 -2.02441365e-01 -4.68266070e-01 5.59016049e-01
6.05065942e-01 7.80302957e-02 -8.91427509e-03 1.19401109e+00
9.78213083e-03 -5.23121595e-01 -6.37772679e-01 1.20337844e+00
7.52360284e-01 -1.44706893e+00 1.73801154e-01 -1.50612101e-01
5.06666541e-01 -2.28575781e-01 3.38016123e-01 -1.15234628e-01
-3.90251487e-01 -1.01798093e+00 3.51343900e-01 7.84326255e-01
8.43647301e-01 -5.19982696e-01 9.96604562e-01 1.75625086e-01
-9.19507384e-01 -3.93622249e-01 -5.87734580e-01 5.98362148e-01
-5.17350614e-01 3.16766232e-01 -1.13553703e+00 4.44262207e-01
7.16542780e-01 3.75480950e-01 -7.21972585e-01 1.44655287e+00
3.11382383e-01 3.93264979e-01 -1.37587219e-01 -2.02170953e-01
1.48844212e-01 -1.98207468e-01 1.48993969e-01 1.10564303e+00
6.33915305e-01 3.38504881e-01 2.40796119e-01 1.90485477e-01
4.59009230e-01 8.53889406e-01 -6.23059809e-01 2.74324358e-01
4.01221722e-01 9.86807942e-01 -1.24339676e+00 -5.92965782e-01
-3.70564401e-01 1.06121314e+00 -3.74466896e-01 4.56083752e-02
-6.99830532e-01 -6.19952023e-01 -2.61549145e-01 1.19497783e-01
2.54469275e-01 3.93230498e-01 -6.72878101e-02 -5.32474637e-01
8.36353824e-02 -5.50754905e-01 7.74382889e-01 -1.13594520e+00
-1.07597792e+00 8.22714150e-01 8.22290778e-04 -1.58214569e+00
8.49962905e-02 -5.56992888e-01 -2.67553508e-01 7.88098276e-01
-1.37185538e+00 -1.05491316e+00 -3.60785633e-01 5.36840141e-01
4.01030272e-01 -4.93414879e-01 1.09199333e+00 4.45207685e-01
3.37281153e-02 3.92439991e-01 6.97187304e-01 4.02411252e-01
4.83468860e-01 -9.45976138e-01 -6.79840624e-01 1.99374035e-01
2.27721855e-01 5.13381004e-01 4.77012604e-01 -6.75983965e-01
-1.12345684e+00 -1.04039800e+00 1.13017774e+00 -1.50939345e-01
1.23411402e-01 3.43075275e-01 -7.36881375e-01 1.41332954e-01
-3.46323013e-01 3.49799573e-01 1.08689523e+00 -9.12220120e-01
-3.39473933e-01 -2.11178571e-01 -1.81870842e+00 2.37350259e-02
-4.08206321e-03 -8.41145515e-01 -8.88476193e-01 7.39975810e-01
3.04269254e-01 -4.98695463e-01 -1.02974665e+00 3.07947963e-01
5.13430834e-01 -6.56908095e-01 1.44192350e+00 -3.44199926e-01
6.12042010e-01 -4.28192317e-01 -6.99386060e-01 -6.41942382e-01
-3.33177112e-02 8.03090215e-01 4.19413805e-01 5.80504000e-01
2.20148832e-01 -5.43953061e-01 8.06897998e-01 3.65587384e-01
1.96523935e-01 -9.08274472e-01 -1.01283967e+00 -7.71979690e-01
-3.00567299e-01 -3.79231982e-02 3.04760516e-01 8.85658801e-01
1.81424141e-01 -2.91365951e-01 -2.23930348e-02 2.31435508e-01
6.40953839e-01 3.93564939e-01 2.48818677e-02 -1.07058203e+00
-2.37309083e-01 -1.69146210e-01 -1.01410961e+00 -4.62870806e-01
-6.76162019e-02 -1.57063591e+00 1.03355218e-02 -1.73849607e+00
6.99922264e-01 -7.36717641e-01 -6.81347728e-01 4.63172644e-01
-1.13121504e-02 7.80597508e-01 -4.91670445e-02 6.21701539e-01
-1.58369035e-01 2.26739477e-02 8.79648328e-01 -5.41318059e-01
-2.54670046e-02 2.69104660e-01 -4.84441489e-01 4.17362422e-01
4.96813565e-01 -7.04146981e-01 -4.20119494e-01 1.96195707e-01
-4.18960899e-01 3.77786785e-01 4.78341252e-01 -1.24111652e+00
2.86639869e-01 -1.89543199e-02 6.95718586e-01 -1.00264478e+00
3.50666195e-01 -1.15147591e+00 3.27952236e-01 1.09195459e+00
-6.71609461e-01 2.75392272e-02 -2.12990776e-01 6.55068457e-01
-3.78668725e-01 -7.27583468e-01 8.97450507e-01 -3.19090746e-02
-7.22382009e-01 1.01267748e-01 -3.67077887e-01 -5.65658867e-01
1.47263050e+00 -6.96682632e-01 -1.32621273e-01 4.46270891e-02
-7.65955627e-01 -4.83902693e-01 1.15101404e-01 2.85226524e-01
1.14541376e+00 -1.55330098e+00 -3.17585438e-01 1.47803813e-01
6.82560742e-01 -6.70589864e-01 2.61002123e-01 5.35544813e-01
-9.46191013e-01 6.58160210e-01 -1.31346121e-01 -8.47951293e-01
-1.90566003e+00 8.15331876e-01 1.79549202e-01 -7.44534358e-02
-8.56001854e-01 5.27573466e-01 -4.30547655e-01 -2.83550948e-01
1.60465557e-02 -7.77246952e-02 -8.64127815e-01 3.67213413e-02
6.94372535e-01 1.74083129e-01 2.67198235e-01 -1.01639938e+00
-4.95616496e-01 1.09449983e+00 2.14808490e-02 -7.40061775e-02
1.17938054e+00 3.41708094e-01 -3.57882082e-01 1.73537046e-01
1.85621238e+00 -1.23113312e-01 -4.53760549e-02 -5.56271195e-01
-4.60674092e-02 -7.22720146e-01 1.57175034e-01 -7.48159587e-01
-1.04203224e+00 5.07971227e-01 1.50597656e+00 -3.85919735e-02
9.09582436e-01 2.64473259e-01 6.28940046e-01 4.96056944e-01
4.54816073e-01 -9.98744607e-01 1.17728822e-01 -1.81672856e-01
5.97621918e-01 -1.31662583e+00 1.88678980e-01 -5.05453885e-01
-5.72794378e-01 1.25651836e+00 4.45016660e-02 -2.62725145e-01
9.66638744e-01 -2.99283773e-01 3.53641182e-01 -3.37904394e-01
1.60833880e-01 2.01652139e-01 3.74419212e-01 6.01202190e-01
1.01710714e-01 6.08089156e-02 -8.19261134e-01 1.59546211e-01
-1.55086651e-01 1.39180878e-02 1.12064719e-01 1.10258651e+00
-8.78721535e-01 -7.92977035e-01 -8.29446793e-01 9.02968109e-01
-5.60764670e-01 -2.34805513e-02 3.22787911e-01 4.12672281e-01
-2.83945292e-01 9.11331713e-01 2.02953264e-01 -2.63740897e-01
4.28046770e-02 9.56154327e-05 3.60277623e-01 -3.74199539e-01
-2.77186543e-01 -3.23708981e-01 -3.05516511e-01 -2.58994251e-01
-3.97587031e-01 -4.12656486e-01 -1.60175538e+00 4.06757146e-01
-3.21553975e-01 4.77628201e-01 1.03639710e+00 8.00992668e-01
-2.15487272e-01 1.88695505e-01 9.26102340e-01 -1.05654955e-01
-8.58610749e-01 -5.28723657e-01 -5.13271570e-01 5.81661820e-01
1.15838990e-01 -5.36405325e-01 -3.01137775e-01 -3.12846899e-02] | [14.275508880615234, -1.4590497016906738] |
9d034f30-a19d-4277-b60a-d41c58626d6f | learning-metric-graphs-for-neuron | 1902.00100 | null | http://arxiv.org/abs/1902.00100v1 | http://arxiv.org/pdf/1902.00100v1.pdf | Learning Metric Graphs for Neuron Segmentation In Electron Microscopy Images | In the deep metric learning approach to image segmentation, a convolutional
net densely generates feature vectors at the pixels of an image. Pairs of
feature vectors are trained to be similar or different, depending on whether
the corresponding pixels belong to same or different ground truth segments. To
segment a new image, the feature vectors are computed and clustered. Both
empirically and theoretically, it is unclear whether or when deep metric
learning is superior to the more conventional approach of directly predicting
an affinity graph with a convolutional net. We compare the two approaches using
brain images from serial section electron microscopy images, which constitute
an especially challenging example of instance segmentation. We first show that
seed-based postprocessing of the feature vectors, as originally proposed,
produces inferior accuracy because it is difficult for the convolutional net to
predict feature vectors that remain uniform across large objects. Then we
consider postprocessing by thresholding a nearest neighbor graph followed by
connected components. In this case, segmentations from a "metric graph" turn
out to be competitive or even superior to segmentations from a directly
predicted affinity graph. To explain these findings theoretically, we invoke
the property that the metric function satisfies the triangle inequality. Then
we show with an example where this constraint suppresses noise, causing
connected components to more robustly segment a metric graph than an
unconstrained affinity graph. | ['Kyle Luther', 'H. Sebastian Seung'] | 2019-01-31 | null | null | null | null | ['electron-microscopy-image-segmentation'] | ['computer-vision'] | [ 5.45667470e-01 4.59533095e-01 2.75626928e-01 -5.53657234e-01
-6.29116535e-01 -7.71818459e-01 3.80972177e-01 4.12730068e-01
-6.23965740e-01 5.93642116e-01 -2.75824457e-01 -1.81370199e-01
-4.30605710e-01 -7.11166859e-01 -8.11826944e-01 -9.68491673e-01
-1.09736212e-01 8.08348298e-01 4.96485680e-01 1.30287081e-01
6.09444559e-01 8.57581377e-01 -1.26245666e+00 1.67762414e-01
4.97112781e-01 8.64836514e-01 2.41425321e-01 7.56797314e-01
-1.42405108e-01 2.01922402e-01 -3.87326956e-01 -2.32009292e-01
5.34927368e-01 -7.46207893e-01 -1.26172841e+00 1.30568922e-01
5.43347716e-01 -9.18665528e-03 2.83956416e-02 1.28525186e+00
1.58822238e-01 -2.38325391e-02 1.06965601e+00 -1.13652480e+00
-5.34179568e-01 6.87859952e-01 -4.84113634e-01 4.17339087e-01
6.20795116e-02 -8.32350925e-02 1.21836889e+00 -5.45260787e-01
9.11601603e-01 7.29144275e-01 8.09783757e-01 3.70143443e-01
-1.75327635e+00 -1.64680168e-01 -6.20440468e-02 2.10566595e-01
-1.34875250e+00 -8.48691091e-02 7.48149216e-01 -6.46688223e-01
6.61001027e-01 2.18680277e-01 8.33742261e-01 5.28668165e-01
4.17087436e-01 4.91035610e-01 1.09840775e+00 -2.75493890e-01
3.52578223e-01 -1.50383860e-01 2.23997280e-01 7.93619692e-01
2.12790444e-01 -4.88087460e-02 2.05426905e-02 -3.84264439e-02
6.87812865e-01 -6.78807944e-02 -4.23363030e-01 -7.73085833e-01
-1.58665836e+00 9.06849861e-01 8.89965534e-01 7.20541775e-01
-3.69697392e-01 2.01892361e-01 8.74509737e-02 2.57457852e-01
3.97436805e-02 8.86948705e-01 -3.85718465e-01 3.18195462e-01
-1.21605027e+00 1.71945155e-01 7.30262041e-01 6.44982517e-01
1.13101399e+00 -4.10440534e-01 3.84149365e-02 2.63102442e-01
1.34522229e-01 -3.47944535e-02 4.07090932e-01 -1.26943648e+00
-1.17574498e-01 7.52550125e-01 -2.01660097e-01 -1.08757722e+00
-8.18461239e-01 -1.19646423e-01 -7.77936757e-01 3.71356636e-01
1.09199238e+00 7.67761841e-02 -1.01990950e+00 1.65847814e+00
1.05168745e-01 -5.76038770e-02 -1.96172848e-01 1.06360102e+00
3.70613933e-01 3.19584817e-01 -3.49465489e-01 -1.13346867e-01
9.28856015e-01 -4.39724833e-01 -2.64108658e-01 -2.36017536e-02
7.95145750e-01 -7.10591614e-01 9.44911122e-01 3.44906002e-01
-8.05800617e-01 -2.40934491e-01 -1.21954799e+00 6.34795353e-02
-5.29726803e-01 -3.12058538e-01 5.66258729e-01 3.45738143e-01
-1.29236233e+00 1.11932862e+00 -8.87694299e-01 -5.90072155e-01
6.66604280e-01 6.79582894e-01 -4.35453892e-01 2.17638925e-01
-7.24255085e-01 8.09809327e-01 3.43303412e-01 -1.14376985e-01
-6.19220257e-01 -5.06432056e-01 -6.13708317e-01 1.47804946e-01
-2.44747162e-01 -4.60530847e-01 8.32649291e-01 -1.29364502e+00
-9.20163512e-01 1.17813826e+00 4.71614115e-02 -5.79879642e-01
5.30194163e-01 6.32252097e-01 2.48339832e-01 3.73715669e-01
2.37676859e-01 9.82093394e-01 5.81450939e-01 -1.32486987e+00
-3.95870298e-01 -5.31322837e-01 -1.21415127e-03 -1.40472353e-01
1.18274100e-01 -1.32983238e-01 4.09585238e-02 -3.92272741e-01
5.96202075e-01 -9.35726881e-01 -4.90614176e-01 6.57313839e-02
-7.15329289e-01 -7.30445459e-02 5.97751498e-01 -1.81196913e-01
6.32028878e-01 -2.02159858e+00 4.62079421e-02 7.36524463e-01
5.28351784e-01 -2.73233861e-01 -1.00947335e-01 -3.58970985e-02
-4.32828695e-01 2.59878069e-01 -7.29457438e-01 1.63612843e-01
3.93316709e-03 2.94357300e-01 1.69570282e-01 9.80919182e-01
1.93706438e-01 7.92598307e-01 -7.60965347e-01 -5.97362101e-01
-7.84217659e-03 3.36413831e-01 -4.61146027e-01 -1.83915287e-01
-7.05719590e-02 4.82807994e-01 -1.89412281e-01 3.43513548e-01
6.16376460e-01 -4.26219225e-01 7.61094540e-02 -3.20316672e-01
-2.22885627e-02 -9.37822759e-02 -1.10819268e+00 1.61550033e+00
4.92563508e-02 1.04204905e+00 -1.63283333e-01 -1.55218673e+00
8.22327912e-01 -1.78699680e-02 7.72461236e-01 -5.25646925e-01
3.75145137e-01 3.69176656e-01 5.05047262e-01 -1.57349318e-01
1.68090060e-01 -3.85925740e-01 1.63477108e-01 5.49145222e-01
9.20223296e-02 -2.06935272e-01 3.08186263e-01 2.51594543e-01
1.38696599e+00 -1.54189095e-01 4.54813540e-02 -7.68321216e-01
3.69963408e-01 2.59574115e-01 4.46219504e-01 6.58537865e-01
-3.94372970e-01 1.11814117e+00 9.52466607e-01 -5.12224078e-01
-1.17373705e+00 -1.28663814e+00 -5.37116826e-01 6.60212219e-01
3.98277700e-01 3.01047657e-02 -1.17565882e+00 -7.54577756e-01
-1.61580816e-01 1.96207643e-01 -9.59049106e-01 -1.15038849e-01
-5.49846172e-01 -7.54654884e-01 3.17011148e-01 4.32552308e-01
2.20529065e-01 -1.06855989e+00 -7.75254250e-01 2.17260405e-01
1.12711877e-01 -7.76947379e-01 -3.64734143e-01 7.79265642e-01
-8.18010926e-01 -1.33111215e+00 -7.24608183e-01 -1.12917113e+00
1.16448903e+00 -9.54383798e-03 1.02375400e+00 3.81840318e-01
-4.95951921e-01 1.18922770e-01 -2.32125700e-01 2.25208905e-02
-3.37088943e-01 1.90806195e-01 -1.31110251e-01 -7.23595768e-02
4.03097391e-01 -5.71008801e-01 -7.86337197e-01 2.82268614e-01
-9.08313930e-01 -2.07685634e-01 5.33297837e-01 8.77749383e-01
9.84325409e-01 -5.17725274e-02 3.95046741e-01 -8.83863151e-01
2.73210168e-01 -3.22834074e-01 -5.19083619e-01 1.40794948e-01
-6.65638983e-01 2.50648975e-01 6.14801705e-01 -3.10269237e-01
-2.11428538e-01 5.38480103e-01 -2.64036562e-02 -1.00683182e-01
-2.46281326e-01 3.42560560e-01 2.44853701e-02 -3.77587408e-01
8.87314320e-01 8.70709717e-02 2.21495733e-01 -1.14604548e-01
2.58334547e-01 3.03583235e-01 5.51382720e-01 -3.24906766e-01
5.73557079e-01 7.17665732e-01 1.44304752e-01 -5.59505701e-01
-4.98456746e-01 -3.56959850e-01 -1.10790706e+00 -1.39843047e-01
1.29814529e+00 3.86655470e-03 -5.14174223e-01 2.74345912e-02
-1.25669825e+00 -5.03000081e-01 -5.91649592e-01 4.46174800e-01
-1.02521944e+00 2.22651318e-01 -5.65803826e-01 -8.60768110e-02
7.08395466e-02 -1.43420768e+00 8.57722461e-01 2.22037643e-01
-2.86216974e-01 -9.98837948e-01 7.15189353e-02 -3.48969176e-02
2.09904104e-01 3.38520139e-01 1.24416721e+00 -1.10325634e+00
-4.30042267e-01 -3.05942267e-01 -4.15138900e-01 1.55188084e-01
2.71369934e-01 2.83826321e-01 -7.15770185e-01 -1.46752089e-01
-5.61277904e-02 6.16002902e-02 8.03866565e-01 5.36771417e-01
1.19492924e+00 -1.49990782e-01 -5.19264877e-01 7.22507358e-01
1.55162859e+00 1.96170479e-01 6.36263132e-01 4.16580856e-01
7.66771913e-01 8.56942475e-01 1.59590319e-01 -2.30687466e-02
4.17889915e-02 4.35081005e-01 4.69283283e-01 -3.29790384e-01
-2.42978334e-02 1.94953442e-01 -1.02640264e-01 5.48274279e-01
1.90136954e-01 8.39217454e-02 -1.09008098e+00 7.30018854e-01
-1.55280793e+00 -8.58623385e-01 -4.38248813e-01 2.15861678e+00
6.73405886e-01 2.68996716e-01 1.94450989e-01 1.96472555e-01
9.00434375e-01 -2.84462810e-01 -7.10011005e-01 -5.35116911e-01
-1.64048418e-01 2.78028399e-01 7.04540908e-01 5.22281289e-01
-1.00915909e+00 7.56068766e-01 6.48987389e+00 4.81398106e-01
-1.05107713e+00 -5.86772747e-02 9.49910998e-01 5.90179861e-03
-3.19130063e-01 1.12865917e-01 -2.85893291e-01 4.15307194e-01
7.10088074e-01 -1.25351086e-01 4.07270819e-01 5.11740983e-01
1.38330245e-02 -2.77317256e-01 -1.44452345e+00 8.55820894e-01
-1.85537115e-01 -1.51694524e+00 5.11661954e-02 2.75005072e-01
8.10804844e-01 2.38775983e-01 -1.45291552e-01 -3.54036361e-01
4.07268554e-01 -1.25754678e+00 5.87539494e-01 5.65126061e-01
4.24027681e-01 -7.12082684e-01 8.13253582e-01 1.66401356e-01
-9.57614899e-01 2.29286402e-01 -5.68983436e-01 2.05033720e-01
2.35483572e-02 5.52701652e-01 -8.11441839e-01 1.35832429e-01
6.12991929e-01 3.93473297e-01 -6.44331574e-01 1.23837483e+00
1.83409706e-01 3.74671459e-01 -3.72320473e-01 1.10486366e-01
4.59623814e-01 -4.38118935e-01 3.63209605e-01 1.13343132e+00
1.99503943e-01 -7.85436258e-02 -3.68932746e-02 1.14204264e+00
3.87994535e-02 1.64159551e-01 -8.72193575e-01 8.78402069e-02
1.10810332e-01 1.30355477e+00 -1.65799832e+00 -2.14962110e-01
-3.22282195e-01 1.03877091e+00 3.81094754e-01 2.98401386e-01
-6.38212740e-01 -5.72837651e-01 6.05739355e-01 3.58497471e-01
4.26736504e-01 -1.71379253e-01 -6.32253587e-01 -6.70710802e-01
-1.42500848e-01 -4.09478247e-01 1.38828784e-01 -5.82884073e-01
-1.25931501e+00 6.61164165e-01 -2.31213540e-01 -9.60478544e-01
-9.63771343e-02 -7.05427885e-01 -6.94884181e-01 5.89808881e-01
-1.00790823e+00 -6.03205144e-01 -5.01488261e-02 5.04336238e-01
9.69107747e-02 3.02634835e-01 5.77885807e-01 -2.27251346e-03
-2.70840377e-01 2.65113533e-01 2.59296983e-01 2.58647740e-01
3.35493803e-01 -1.68200016e+00 1.32099107e-01 7.91043580e-01
5.52262306e-01 4.49234068e-01 6.76920652e-01 -4.40285802e-01
-8.87405992e-01 -9.81199682e-01 7.57985175e-01 -4.30080891e-01
7.48891830e-01 -3.46081167e-01 -1.09040797e+00 6.43805683e-01
1.59732252e-01 3.10945809e-01 5.66704452e-01 -3.90913159e-01
-7.69560933e-02 6.99807256e-02 -1.34306765e+00 4.67803955e-01
9.40794110e-01 -3.21505040e-01 -4.96519446e-01 5.46776354e-01
3.67770046e-01 1.09530672e-01 -9.55488741e-01 2.96482891e-01
3.90433103e-01 -1.26233268e+00 6.94856167e-01 -6.19980872e-01
3.56447846e-01 -4.71393198e-01 -3.28022987e-01 -1.29814863e+00
-3.73044103e-01 -3.25887263e-01 6.17131770e-01 8.90257478e-01
7.46441782e-01 -4.50169533e-01 1.08784389e+00 5.70133507e-01
-1.82689041e-01 -9.55768585e-01 -1.10053360e+00 -9.19132769e-01
5.29978871e-01 -2.25536048e-01 4.80181247e-01 1.05997825e+00
9.02826190e-02 1.96793631e-01 5.15190423e-01 2.38684961e-03
5.14976919e-01 2.75241524e-01 2.93426931e-01 -1.50889564e+00
-3.27330567e-02 -9.26245868e-01 -1.02801633e+00 -8.33948672e-01
2.50320464e-01 -1.24300385e+00 4.49797451e-01 -1.64126992e+00
1.89528689e-01 -5.01532614e-01 -3.29857528e-01 3.39152306e-01
1.58960059e-01 6.13806725e-01 -9.79157835e-02 2.48314679e-01
-4.77661848e-01 2.37032678e-02 1.25582016e+00 -3.24442267e-01
-1.24820650e-01 -1.45276681e-01 -5.96301556e-01 8.65735650e-01
9.14840937e-01 -6.25612080e-01 -2.43566096e-01 -2.84200996e-01
2.80081779e-01 -9.61278155e-02 3.94889325e-01 -1.05434740e+00
3.27278435e-01 1.30994067e-01 5.63541114e-01 -2.00256854e-01
-5.73292375e-02 -8.93088877e-01 9.99684706e-02 5.91238797e-01
-4.95171547e-01 1.25552446e-01 -2.00482175e-01 3.44984412e-01
1.36875678e-02 -5.16639531e-01 1.17161191e+00 -1.64992288e-01
-3.74729186e-01 3.64113927e-01 -6.47113562e-01 -2.77800765e-02
1.11211514e+00 -7.59439766e-01 -1.34278536e-01 -1.43114865e-01
-1.06260455e+00 -2.01009259e-01 8.14872801e-01 -2.26321489e-01
5.93393028e-01 -1.32492924e+00 -4.76808965e-01 2.06593230e-01
6.97312877e-02 1.50541276e-01 -2.21910685e-01 1.14982200e+00
-8.47612441e-01 5.05429059e-02 -3.63136023e-01 -8.90870094e-01
-9.86603498e-01 6.33545637e-01 5.98801017e-01 4.79689874e-02
-6.33132279e-01 9.19740975e-01 3.72343630e-01 -3.28035861e-01
-4.90354858e-02 -6.48979485e-01 -5.74594438e-02 6.94485903e-02
8.69080126e-02 6.22227378e-02 1.78420886e-01 -8.41498435e-01
-4.27123487e-01 8.03537428e-01 -1.64764404e-01 -6.14317432e-02
1.25664914e+00 -6.78947866e-02 -2.30186865e-01 4.39148188e-01
1.55665326e+00 -3.48917872e-01 -1.42995858e+00 -9.26308520e-03
3.36468250e-01 -3.03742677e-01 -8.84460509e-02 -5.16763628e-01
-1.39417696e+00 7.65512049e-01 7.59727895e-01 6.69943333e-01
8.78798783e-01 4.93400604e-01 6.61507964e-01 3.22665215e-01
1.95516139e-01 -9.07863200e-01 4.86403098e-03 3.00714463e-01
7.14110732e-01 -1.23383105e+00 -6.75906986e-02 -2.44408660e-02
-2.45711252e-01 1.23875248e+00 4.55434859e-01 -5.14285922e-01
7.08242714e-01 2.86080897e-01 1.38667002e-01 -6.16961718e-01
-2.27300122e-01 -3.56771767e-01 1.39803275e-01 7.60689735e-01
4.12664890e-01 1.31980302e-02 -3.32683891e-01 2.53738552e-01
-4.82989281e-01 -4.39656764e-01 6.27639830e-01 6.54772401e-01
-5.92242599e-01 -8.27644587e-01 -1.83509558e-01 8.36468220e-01
-3.42135221e-01 -1.04844928e-01 -7.35978842e-01 8.47018719e-01
1.70722440e-01 6.18563890e-01 5.99899173e-01 -3.76975596e-01
1.45606384e-01 -1.69865251e-01 6.97889745e-01 -5.85784853e-01
-4.50576961e-01 -1.15017116e-01 -5.49087346e-01 -5.40169418e-01
-4.53978151e-01 -8.00210059e-01 -1.86702991e+00 -1.48499623e-01
-3.73157948e-01 1.00385047e-01 6.51073694e-01 1.13712990e+00
9.82988626e-02 2.41562188e-01 5.84293664e-01 -1.01629877e+00
-8.96488130e-02 -6.46546781e-01 -7.41595149e-01 4.62526232e-01
2.58133858e-01 -5.94027996e-01 -6.02101862e-01 2.23088399e-01] | [14.213846206665039, -2.9566657543182373] |
c4aa67dc-8c39-4d40-92b9-d6ec10a3c536 | span-based-discontinuous-constituency-parsing | 2003.13785 | null | https://arxiv.org/abs/2003.13785v1 | https://arxiv.org/pdf/2003.13785v1.pdf | Span-based discontinuous constituency parsing: a family of exact chart-based algorithms with time complexities from O(n^6) down to O(n^3) | We introduce a novel chart-based algorithm for span-based parsing of discontinuous constituency trees of block degree two, including ill-nested structures. In particular, we show that we can build variants of our parser with smaller search spaces and time complexities ranging from $\mathcal O(n^6)$ down to $\mathcal O(n^3)$. The cubic time variant covers 98\% of constituents observed in linguistic treebanks while having the same complexity as continuous constituency parsers. We evaluate our approach on German and English treebanks (Negra, Tiger and Discontinuous PTB) and report state-of-the-art results in the fully supervised setting. We also experiment with pre-trained word embeddings and \bert{}-based neural networks. | ['Caio Corro'] | 2020-03-30 | null | null | null | null | ['constituency-parsing'] | ['natural-language-processing'] | [ 9.57491025e-02 5.65935552e-01 -6.78616017e-02 -4.24628288e-01
-1.44722474e+00 -1.02842629e+00 -8.22282806e-02 6.17404759e-01
-7.01374590e-01 7.33729839e-01 3.10447097e-01 -1.20938694e+00
2.15935871e-01 -9.89370346e-01 -5.58674276e-01 -2.63041735e-01
-5.87189615e-01 4.89965320e-01 4.85251576e-01 -6.17621422e-01
6.51299069e-03 2.51888782e-01 -9.19604063e-01 2.60268450e-01
6.81013465e-01 7.89815903e-01 1.03738643e-01 1.17499864e+00
-6.11998022e-01 5.55680513e-01 -3.93108428e-01 -9.19862688e-01
2.64434546e-01 -8.20968896e-02 -1.09578753e+00 -4.34094489e-01
7.47615159e-01 1.36684567e-01 -2.16145471e-01 9.97350574e-01
2.68742889e-01 -1.08767413e-01 6.72986954e-02 -3.89107555e-01
-4.59311277e-01 1.26573241e+00 -3.55810702e-01 7.48815894e-01
2.51681268e-01 -2.47880340e-01 1.88262260e+00 -5.52913666e-01
6.55274093e-01 1.32695842e+00 7.89704978e-01 4.97828513e-01
-1.40365362e+00 -5.24113595e-01 5.26315093e-01 -2.27524251e-01
-7.84276783e-01 -4.70922261e-01 5.36889493e-01 -4.01065014e-02
1.62659144e+00 1.13077208e-01 3.65059465e-01 5.94621956e-01
3.26800585e-01 5.62771499e-01 1.09125030e+00 -9.44986641e-01
1.24114072e-02 -4.72838968e-01 7.77063191e-01 1.23376822e+00
4.52554286e-01 -7.82676972e-04 -3.93156797e-01 -1.75190330e-01
6.11140311e-01 -7.51642287e-01 2.26620287e-01 2.72581697e-01
-8.18142354e-01 1.23614836e+00 3.85896750e-02 5.93379378e-01
3.55603099e-02 3.91781449e-01 6.66048765e-01 4.87068415e-01
5.38535953e-01 2.10083857e-01 -9.94353116e-01 -1.97349787e-01
-5.98861039e-01 1.22252181e-01 1.06028867e+00 1.33819950e+00
6.92710161e-01 6.34034052e-02 2.87103564e-01 7.94885457e-01
-1.13295525e-01 2.58590370e-01 1.69854432e-01 -8.12267005e-01
1.19397295e+00 3.78825486e-01 -1.13737144e-01 -1.04715168e-01
-7.04131484e-01 -8.44293982e-02 -4.49817955e-01 -6.12249151e-02
8.43262851e-01 -5.11874855e-01 -8.29047263e-01 1.86701691e+00
3.46969992e-01 -4.24247593e-01 9.01476443e-02 1.46631986e-01
4.96008754e-01 7.78366983e-01 3.27568769e-01 -3.85742903e-01
2.00473261e+00 -7.63222039e-01 -6.53331816e-01 -6.99694991e-01
1.07574129e+00 -7.75527537e-01 9.72848594e-01 2.82663345e-01
-1.61491084e+00 -3.73976558e-01 -9.79947686e-01 -4.67000097e-01
-2.67475039e-01 -1.30808517e-01 1.05062413e+00 1.08484650e+00
-1.23058760e+00 6.93989456e-01 -1.15270150e+00 4.05958816e-02
1.00237072e-01 3.44197989e-01 -4.41517472e-01 3.68106738e-02
-1.05488455e+00 7.81561434e-01 3.93447965e-01 6.63217306e-02
-2.60499835e-01 -6.10634267e-01 -1.23143506e+00 7.39962310e-02
1.71899334e-01 -1.01583442e-02 1.48576033e+00 -2.10953176e-01
-1.33319855e+00 1.17168689e+00 -3.94046396e-01 -6.16847336e-01
-7.77197257e-02 -2.96060443e-01 -4.38679546e-01 8.61782581e-02
2.63580710e-01 3.71385604e-01 -1.59328338e-02 -5.70214570e-01
-1.17138112e+00 -3.62839371e-01 3.69308889e-01 -1.81089357e-01
-1.04084715e-01 5.14500916e-01 -1.53574198e-01 -4.81942147e-01
5.27054131e-01 -6.73430920e-01 -5.46090245e-01 -4.89077330e-01
-2.81877607e-01 -7.10600853e-01 1.41099095e-02 -9.77516949e-01
1.52170706e+00 -1.97387588e+00 -5.91757037e-02 -1.12436317e-01
2.58616731e-02 1.76612124e-01 -3.42394829e-01 6.24645293e-01
-6.43922761e-02 3.27471018e-01 -4.11888957e-01 -3.14491093e-01
2.57726699e-01 5.16006589e-01 -1.44217551e-01 2.24204972e-01
4.42339182e-01 7.55982220e-01 -8.54451120e-01 -6.56923711e-01
-1.36194393e-01 -1.92244306e-01 -6.39491796e-01 -2.07116790e-02
-3.97998124e-01 -1.09057583e-01 -2.89991915e-01 7.72713125e-01
6.82917118e-01 2.99807161e-01 9.51850653e-01 2.15859443e-01
-4.59349543e-01 1.22110856e+00 -1.09358025e+00 1.72598028e+00
-8.67139637e-01 3.48887146e-01 5.78383446e-01 -1.01632357e+00
7.78021812e-01 3.36588323e-01 6.97639883e-02 -5.44199407e-01
1.19283147e-01 4.19866443e-01 1.67165771e-01 -1.66623980e-01
5.77933431e-01 -3.93219829e-01 -8.96952808e-01 2.71349072e-01
2.64188141e-01 -1.70032263e-01 6.41819060e-01 -3.25014838e-03
1.66999698e+00 2.03166716e-02 5.31001747e-01 -5.69034815e-01
2.80367225e-01 1.70701474e-01 9.12219465e-01 6.59850538e-01
-3.67579490e-01 -2.71027386e-02 1.03757453e+00 -8.25372636e-01
-1.05446267e+00 -1.30054295e+00 -3.65406066e-01 1.80936480e+00
-2.85085678e-01 -5.34166694e-01 -7.33486176e-01 -8.21816742e-01
-4.77576792e-01 7.66919971e-01 -4.78980213e-01 6.40087366e-01
-1.70847023e+00 -6.70435309e-01 7.94450998e-01 9.69889998e-01
2.19130605e-01 -1.32951617e+00 -6.98271215e-01 9.57251728e-01
4.36393805e-02 -1.49637234e+00 -4.56160367e-01 1.07394075e+00
-1.25881636e+00 -1.14088690e+00 1.75790772e-01 -1.60630977e+00
2.78120458e-01 -3.59279215e-01 1.64214659e+00 -3.95412669e-02
-1.92524090e-01 -3.28867197e-01 -5.87659895e-01 -8.37414786e-02
-6.08440936e-01 4.15841222e-01 -2.74260104e-01 -1.02819586e+00
6.09082103e-01 -6.84043705e-01 -1.52937278e-01 -2.87530690e-01
-5.61747432e-01 -3.18434566e-01 4.08130199e-01 1.08459103e+00
5.82988381e-01 -1.69027805e-01 1.88339800e-01 -1.33580697e+00
4.53225583e-01 -9.90092307e-02 -1.09853184e+00 2.48690993e-01
-1.62955284e-01 3.75272274e-01 1.00667703e+00 -8.51517320e-02
-9.70213890e-01 3.00580293e-01 -8.72364461e-01 5.25413513e-01
-1.56182617e-01 4.01209682e-01 -1.13827534e-01 2.89748669e-01
5.97587168e-01 -1.02301314e-01 -6.66902721e-01 -6.89589143e-01
7.45267928e-01 2.52631903e-01 6.23903334e-01 -1.08094203e+00
6.01901531e-01 7.93691948e-02 -9.99186262e-02 -7.60569692e-01
-8.56939614e-01 -2.58722723e-01 -6.63260698e-01 6.98564887e-01
1.05657327e+00 -8.41059566e-01 -5.20899177e-01 -1.53579220e-01
-1.52530193e+00 -4.72010702e-01 -2.56360203e-01 1.80047929e-01
-4.97367859e-01 5.05070806e-01 -1.28289533e+00 -8.14845383e-01
-5.54432571e-01 -7.92757273e-01 1.01370561e+00 -1.29927978e-01
-1.05248421e-01 -9.74389970e-01 2.59338289e-01 6.69792369e-02
-1.37258142e-01 2.31673732e-01 1.55501592e+00 -7.36645579e-01
-4.51353759e-01 -6.78260773e-02 -1.39257222e-01 2.18315169e-01
4.47924286e-02 -4.40299273e-01 -6.15008771e-01 -5.23129329e-02
-1.95333213e-01 -1.82057828e-01 9.09474552e-01 3.58350366e-01
5.06462991e-01 -4.03200120e-01 -1.97030097e-01 4.74294603e-01
1.54550993e+00 1.68772444e-01 4.41393375e-01 2.45373383e-01
3.18430334e-01 6.10444605e-01 5.03594577e-01 9.43989158e-02
4.39350069e-01 1.27120167e-01 2.28880569e-01 2.78474241e-01
5.08873761e-02 -2.66452879e-01 5.54591835e-01 1.14558125e+00
-6.44643605e-03 -6.43031597e-02 -8.75297785e-01 1.09556103e+00
-1.36772668e+00 -6.12877846e-01 -2.51139581e-01 1.65451574e+00
1.23943949e+00 6.88092172e-01 8.94031525e-02 3.78803641e-01
6.82183564e-01 3.63098830e-01 1.50246382e-01 -1.52340662e+00
6.78609535e-02 1.56370473e+00 7.57999182e-01 8.41306150e-01
-1.34591115e+00 1.75168085e+00 6.68318796e+00 6.77479565e-01
-4.88890946e-01 3.96325976e-01 4.76752102e-01 1.40663281e-01
-3.24794829e-01 2.79500216e-01 -1.18240416e+00 -4.94396351e-02
1.35719848e+00 2.11732134e-01 2.06614509e-01 9.34738696e-01
-3.44425976e-01 3.42422398e-03 -1.16425049e+00 4.16749537e-01
-6.11618400e-01 -1.46962690e+00 -7.07413495e-01 -4.79553416e-02
3.51722240e-01 3.19806725e-01 -3.88628811e-01 5.40500104e-01
1.07775915e+00 -7.51115024e-01 7.38713384e-01 -6.63854957e-01
1.02566361e+00 -7.27247715e-01 6.33874416e-01 2.18790948e-01
-1.84078920e+00 -6.52952269e-02 -5.80465674e-01 -3.10483038e-01
6.54899061e-01 4.82210398e-01 -4.92248446e-01 4.75332141e-01
7.00662136e-01 -1.04867242e-01 -1.71660736e-01 3.39499652e-01
-6.50289774e-01 1.00903201e+00 -7.43185699e-01 -4.70607400e-01
6.01494431e-01 -1.38673559e-01 2.42812008e-01 1.64425528e+00
6.09139390e-02 6.17780328e-01 1.56945452e-01 2.09682018e-01
-1.77565590e-01 2.54492402e-01 -3.19084167e-01 2.97803991e-02
6.52810037e-01 1.11844456e+00 -8.40481818e-01 -2.53147840e-01
-5.61127901e-01 6.71276867e-01 9.86745000e-01 -2.67357618e-01
-5.92393279e-01 -6.68628156e-01 6.38387978e-01 -1.01098321e-01
1.03584766e+00 -7.42636204e-01 -4.68547523e-01 -9.29781199e-01
2.80103415e-01 -7.49878049e-01 8.06279659e-01 1.40442789e-01
-1.15687680e+00 1.03831291e+00 -2.33804565e-02 -4.42773461e-01
-1.25460997e-01 -1.15746045e+00 -6.97977364e-01 8.41949105e-01
-1.57742774e+00 -9.68857110e-01 5.63838959e-01 4.82955053e-02
5.64807653e-01 6.05579391e-02 1.35388970e+00 8.34582150e-02
-4.36822146e-01 8.54382455e-01 -1.96433976e-01 5.52109957e-01
4.86309752e-02 -1.61189055e+00 1.34050179e+00 1.03627443e+00
5.09539545e-01 5.17166078e-01 5.84702313e-01 -3.13696682e-01
-1.39784038e+00 -8.90620589e-01 1.65232992e+00 -3.78487527e-01
1.11870098e+00 -8.14346313e-01 -5.68381488e-01 9.70652461e-01
3.73576641e-01 1.90174446e-01 6.49483025e-01 7.59386897e-01
-6.79566026e-01 -1.81650603e-03 -1.07777047e+00 4.87120748e-01
1.52854562e+00 -4.05269355e-01 -9.89592969e-01 3.41348976e-01
1.05063438e+00 -6.63189769e-01 -1.04792643e+00 6.49992516e-03
4.04402941e-01 -8.66093159e-01 6.28136635e-01 -7.11166561e-01
2.85713613e-01 3.49289149e-01 -4.15176868e-01 -8.79871607e-01
-4.35372442e-01 -9.26785052e-01 1.17820181e-01 1.06599140e+00
9.16919708e-01 -6.67667866e-01 9.61181641e-01 2.42364153e-01
-5.20502746e-01 -7.03364015e-01 -1.59395862e+00 -7.20875680e-01
9.76985633e-01 -7.04897642e-01 4.24787521e-01 3.76774698e-01
2.45076776e-01 4.08277392e-01 1.14168786e-01 3.22572917e-01
3.60049903e-01 2.51354486e-01 1.65591061e-01 -9.21341062e-01
-4.41277444e-01 -3.54842871e-01 -2.15881348e-01 -1.24131739e+00
2.65746683e-01 -9.66527998e-01 2.18468085e-01 -1.32806170e+00
-4.35697556e-01 -9.34633851e-01 -2.07726911e-01 6.55136883e-01
-7.78896809e-02 -5.15662059e-02 7.18715265e-02 -6.09491944e-01
-2.73149699e-01 7.77390273e-03 6.61991119e-01 -3.24848928e-02
-1.13856941e-01 -3.12590569e-01 -7.49034286e-01 7.99535096e-01
8.12911153e-01 -7.98940957e-01 1.93046287e-01 -9.60227728e-01
3.88375133e-01 5.19299567e-01 -2.24594533e-01 -7.45786488e-01
-1.10096075e-01 1.15150502e-02 -2.43335634e-01 -6.75903678e-01
9.04974937e-02 -3.29664171e-01 -7.16819525e-01 5.46375871e-01
-2.11841464e-01 8.04619372e-01 5.39015770e-01 3.36034477e-01
-1.08890228e-01 -6.26600325e-01 5.53196609e-01 -4.67050612e-01
-4.03371513e-01 2.67540455e-01 -4.29483473e-01 6.28668964e-01
4.78071481e-01 2.84678694e-02 -1.94917500e-01 2.86680013e-01
-8.14620495e-01 1.17713042e-01 -1.58652842e-01 -3.18360515e-02
1.89650372e-01 -8.33190143e-01 -7.02319682e-01 -3.02613266e-02
-1.59787014e-01 2.07569152e-01 1.21756028e-02 2.65806407e-01
-1.11394203e+00 5.61347961e-01 1.70815244e-01 -4.82130013e-02
-1.10950482e+00 3.08773845e-01 -5.92471994e-02 -9.98103201e-01
-7.09276617e-01 1.34918928e+00 -2.15244040e-01 -6.15849018e-01
4.36643958e-02 -9.84123290e-01 8.96420032e-02 -2.18107942e-02
3.22981060e-01 3.41947079e-02 2.09005058e-01 -3.65184426e-01
-6.49072349e-01 4.55533981e-01 -1.04130305e-01 -2.61226505e-01
1.52673364e+00 3.28492731e-01 -2.53332287e-01 6.99501038e-02
1.08854926e+00 4.45909172e-01 -9.76411700e-01 -1.83118448e-01
6.21284068e-01 2.60670990e-01 -3.41749072e-01 -4.30562735e-01
-6.62866950e-01 9.18828130e-01 2.32583404e-01 5.79507172e-01
1.00207436e+00 2.36218229e-01 1.29190004e+00 5.91019452e-01
6.59451604e-01 -1.21902013e+00 -4.74946290e-01 8.95691872e-01
2.99121976e-01 -9.23236668e-01 -2.32098401e-01 -7.21173584e-01
-2.15043426e-01 1.31088781e+00 1.79251879e-01 -6.78236604e-01
8.14045370e-01 7.40730584e-01 6.44517690e-02 1.19304089e-02
-9.58919406e-01 -2.65321791e-01 -3.70614022e-01 3.68105799e-01
8.62659872e-01 4.11343724e-01 -9.09390569e-01 9.33776498e-01
-8.07202399e-01 -6.20336950e-01 3.64678144e-01 1.36777735e+00
-6.89599395e-01 -1.83977711e+00 -1.25101209e-02 1.09901875e-01
-9.87235606e-01 -7.58362353e-01 8.38045627e-02 1.06426382e+00
1.44830391e-01 1.12491918e+00 1.55903816e-01 -6.76273415e-03
4.93296236e-01 4.84009355e-01 9.11510646e-01 -9.26195204e-01
-1.02828395e+00 9.16283056e-02 9.99325812e-01 -4.78506446e-01
-1.31691441e-01 -7.19156086e-01 -1.50955498e+00 -4.09202099e-01
-3.81910533e-01 2.40030438e-01 4.87531602e-01 8.13635349e-01
-1.47298023e-01 3.30285966e-01 4.28575486e-01 -3.79152238e-01
-7.14950681e-01 -1.08437264e+00 -4.64210838e-01 -8.47510900e-03
2.25954354e-02 -1.40839860e-01 -1.29561365e-01 -4.03827615e-02] | [10.332148551940918, 9.74126148223877] |
7ee459b7-d20b-422c-ba13-4815caef1dba | towards-tractable-mathematical-reasoning | 2111.05364 | null | https://arxiv.org/abs/2111.05364v1 | https://arxiv.org/pdf/2111.05364v1.pdf | Towards Tractable Mathematical Reasoning: Challenges, Strategies, and Opportunities for Solving Math Word Problems | Mathematical reasoning would be one of the next frontiers for artificial intelligence to make significant progress. The ongoing surge to solve math word problems (MWPs) and hence achieve better mathematical reasoning ability would continue to be a key line of research in the coming time. We inspect non-neural and neural methods to solve math word problems narrated in a natural language. We also highlight the ability of these methods to be generalizable, mathematically reasonable, interpretable, and explainable. Neural approaches dominate the current state of the art, and we survey them highlighting three strategies to MWP solving: (1) direct answer generation, (2) expression tree generation for inferring answers, and (3) template retrieval for answer computation. Moreover, we discuss technological approaches, review the evolution of intuitive design choices to solve MWPs, and examine them for mathematical reasoning ability. We finally identify several gaps that warrant the need for external knowledge and knowledge-infused learning, among several other opportunities in solving MWPs. | ['Aditi Avasthi', 'Manas Gaur', 'Prashant Kikani', 'Amit Sheth', 'Keyur Faldu'] | 2021-10-29 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 3.28995585e-01 5.08028746e-01 6.09544925e-02 -3.58415544e-01
-3.68180156e-01 -8.91616106e-01 2.91868418e-01 8.26208964e-02
2.22061053e-02 7.78950155e-01 2.00500816e-01 -6.82661831e-01
-7.05874205e-01 -1.48353803e+00 -6.23200893e-01 -1.64368257e-01
3.28979790e-01 5.07137537e-01 -2.74699569e-01 -5.54547429e-01
9.15741563e-01 7.19633579e-01 -1.49082851e+00 4.41271335e-01
1.25840771e+00 8.22156310e-01 2.98787594e-01 7.69314945e-01
-7.95254171e-01 1.39089680e+00 -6.66632056e-01 -7.62814701e-01
-1.80899456e-01 -4.93625045e-01 -1.40079367e+00 -6.25830173e-01
5.29517412e-01 -2.33106300e-01 -3.72608662e-01 8.48827243e-01
2.72838444e-01 3.22300196e-01 5.78026235e-01 -1.02197862e+00
-1.41630840e+00 5.78878582e-01 2.38678560e-01 2.92316675e-01
7.22499073e-01 3.62681329e-01 8.52828205e-01 -6.69920564e-01
5.63613534e-01 1.29041004e+00 5.72428405e-01 8.61514747e-01
-9.28864598e-01 -4.20729607e-01 1.03533335e-01 6.16717637e-01
-1.02378511e+00 -1.71978667e-01 7.87848473e-01 -3.10455412e-01
1.35883224e+00 5.16033769e-01 1.05089092e+00 6.56935930e-01
1.34180993e-01 6.81468427e-01 1.10240149e+00 -6.16394520e-01
1.91769361e-01 1.25958651e-01 5.28532088e-01 1.08203351e+00
2.68102050e-01 -2.95970663e-02 -9.17877495e-01 1.15335837e-01
1.01750743e+00 -4.80044693e-01 -3.42111439e-01 7.30831102e-02
-9.30401444e-01 1.05772281e+00 3.22963774e-01 3.30351412e-01
-2.19968632e-01 4.53018963e-01 3.36735770e-02 3.23901922e-01
-7.23731071e-02 1.53235888e+00 -5.37020802e-01 -2.26158366e-01
-6.92911446e-01 7.57651031e-01 1.03310013e+00 7.75354207e-01
4.95308965e-01 4.64392811e-01 -1.18208803e-01 4.05022711e-01
6.08956330e-02 3.98649246e-01 2.05344141e-01 -1.47029507e+00
4.71634299e-01 7.93767869e-01 -7.78505057e-02 -1.38464296e+00
-3.76091510e-01 -3.91919076e-01 -4.40394610e-01 1.18754804e-01
6.21729851e-01 -6.94974810e-02 -4.43487436e-01 1.56100082e+00
1.36665320e-02 -3.13977271e-01 1.69271097e-01 6.89946175e-01
1.38736379e+00 1.05746555e+00 -1.73176732e-02 8.38104784e-02
1.24155200e+00 -8.78885388e-01 -8.60875249e-01 -2.07245991e-01
5.36672473e-01 -4.21840191e-01 1.10456598e+00 3.05802256e-01
-1.58787453e+00 -5.07158875e-01 -9.39635456e-01 -4.42692012e-01
-7.51225293e-01 1.29843457e-02 1.32204080e+00 7.74086058e-01
-1.09327185e+00 7.58017600e-01 -3.56699198e-01 -1.77595481e-01
3.93411368e-01 5.34811795e-01 1.19895309e-01 -6.89770132e-02
-1.44351590e+00 1.43719637e+00 4.78964180e-01 2.68110752e-01
-1.75482705e-01 -1.09953523e+00 -9.12662804e-01 2.12088645e-01
3.59098971e-01 -1.23180711e+00 1.16896725e+00 -6.95482135e-01
-1.73086107e+00 1.05121982e+00 -4.62826669e-01 -4.06757683e-01
5.66012738e-03 -1.52099654e-01 -1.81567252e-01 4.47206050e-01
-3.38614851e-01 1.04070091e+00 3.95050406e-01 -1.17885375e+00
-1.82672665e-01 -9.25473347e-02 5.70826173e-01 1.28144667e-01
-2.06757605e-01 -3.12385652e-02 2.63830572e-01 -5.34089029e-01
3.05810571e-01 -3.40678602e-01 -6.47783047e-03 3.78254890e-01
-9.07972977e-02 -7.63982713e-01 1.12718709e-01 -7.47086465e-01
1.22644150e+00 -1.46586955e+00 4.95724052e-01 1.62832096e-01
5.97113609e-01 2.65328348e-01 -1.97335213e-01 5.45867383e-01
-1.83985069e-01 2.37383634e-01 7.44196847e-02 3.88789803e-01
4.04819071e-01 1.25832081e-01 -6.59590542e-01 -2.84170210e-01
4.90479141e-01 1.70294082e+00 -9.13702369e-01 -1.44848078e-01
2.98443526e-01 2.48801708e-01 -5.92510283e-01 2.69684404e-01
-5.96060336e-01 2.37107471e-01 -5.09652853e-01 5.90288639e-01
4.30057019e-01 -2.80790895e-01 5.18705994e-02 -1.70399860e-01
-2.31376588e-01 5.27152121e-01 -9.34919894e-01 1.44043255e+00
-5.19386947e-01 9.41479862e-01 -1.82795882e-01 -1.10118759e+00
1.21926880e+00 1.16990052e-01 -2.16111317e-01 -7.62869835e-01
-1.01435361e-02 2.79094934e-01 1.77792042e-01 -9.79702771e-01
3.90967816e-01 -3.75121146e-01 2.32834324e-01 4.76998836e-01
8.14847648e-03 -1.01338768e+00 2.46540859e-01 5.96660487e-02
8.07017565e-01 2.41387174e-01 2.76007652e-01 -3.35602432e-01
7.07703531e-01 1.10793829e-01 -1.25610279e-02 9.33498263e-01
5.89106604e-02 1.46434709e-01 4.34812725e-01 -1.09170568e+00
-7.40263581e-01 -1.15779495e+00 1.76423341e-01 8.31816316e-01
-1.03572883e-01 -3.02125186e-01 -9.26819742e-01 8.02387074e-02
-2.89086074e-01 1.24344528e+00 -3.59273016e-01 -2.61331230e-01
-8.01852643e-01 -2.47081980e-01 6.38962924e-01 6.06450617e-01
6.32753015e-01 -1.50754583e+00 -9.04729784e-01 7.38453716e-02
-5.01528084e-01 -8.13517869e-01 5.35504639e-01 1.52647525e-01
-1.15914786e+00 -7.58038700e-01 -5.57362795e-01 -9.32884514e-01
5.49680948e-01 -4.62023355e-02 1.41532493e+00 5.98367572e-01
-2.91832060e-01 6.08780444e-01 6.67559281e-02 -4.60901380e-01
-2.83001997e-02 5.53262718e-02 -1.54063776e-01 -8.58663499e-01
4.71506923e-01 -6.80072129e-01 -3.47318381e-01 -3.14485103e-01
-7.45636582e-01 2.34023198e-01 3.19695413e-01 4.86688286e-01
5.79432920e-02 1.99602157e-01 7.26414382e-01 -4.90732312e-01
1.07396114e+00 -3.29197764e-01 -4.02790248e-01 5.36302805e-01
-3.14620465e-01 2.70904571e-01 6.08065605e-01 -2.42102280e-01
-1.24525702e+00 -3.89709234e-01 -3.45306426e-01 3.24094832e-01
-2.83913493e-01 6.29116893e-01 4.38361801e-02 -4.37416792e-01
7.43551672e-01 3.42185050e-01 -1.91465989e-01 1.19186573e-01
5.70532560e-01 -2.85033751e-02 5.17841756e-01 -1.28586555e+00
9.31102633e-01 5.42818708e-03 1.91721201e-01 -8.30958605e-01
-9.93001580e-01 4.27667141e-01 -3.14519405e-01 -3.33439291e-01
8.69692326e-01 -3.11098784e-01 -1.12711549e+00 1.67369932e-01
-1.76033056e+00 -3.14353466e-01 -3.06398571e-01 2.08130494e-01
-6.78289771e-01 9.65791792e-02 -8.13931048e-01 -1.00423872e+00
-4.37168211e-01 -1.00242376e+00 6.59400284e-01 7.70918965e-01
-1.06597221e+00 -1.24128234e+00 -2.18862981e-01 8.40222836e-01
6.90659285e-01 1.86107501e-01 1.90377176e+00 -3.32789332e-01
-1.02788937e+00 -1.24659454e-02 -5.15327156e-01 -1.55395284e-01
-3.11591476e-01 -5.40779047e-02 -9.12333131e-01 6.52843595e-01
3.06895077e-01 -7.48581469e-01 5.40154874e-01 3.01778436e-01
1.56719077e+00 -4.22248662e-01 1.44964885e-02 3.69202584e-01
1.24465621e+00 3.04277658e-01 9.54958558e-01 2.35404551e-01
2.33625486e-01 1.15101135e+00 1.99213717e-02 -8.56473520e-02
4.81014341e-01 1.31072775e-01 2.03106239e-01 1.87098831e-01
-1.34617373e-01 -3.55026811e-01 -9.48689729e-02 1.08320582e+00
-3.89411420e-01 -7.44824260e-02 -1.04622328e+00 2.50230521e-01
-1.53392744e+00 -1.14534640e+00 -3.67353052e-01 1.56118453e+00
8.29750419e-01 -4.65300400e-03 -2.86718905e-01 3.34883660e-01
3.03847700e-01 -1.61268726e-01 -5.20323813e-01 -9.23660517e-01
-1.98850930e-01 1.16604674e+00 -4.44023520e-01 6.62245154e-01
-4.64803815e-01 1.21299946e+00 7.35934973e+00 7.47519016e-01
-7.49548852e-01 -5.49677849e-01 6.31491780e-01 2.68225878e-01
-7.48587012e-01 -2.57773727e-01 -3.74768436e-01 -9.63608623e-02
7.98617959e-01 -9.18352604e-02 9.60465372e-01 6.08805478e-01
-1.49033368e-01 -2.29853898e-01 -1.21349573e+00 1.04184377e+00
6.26792461e-02 -1.93850887e+00 3.83665144e-01 -4.93780077e-01
5.63166559e-01 -8.85837853e-01 1.58366606e-01 4.15981591e-01
5.18195219e-02 -1.69953680e+00 5.63237846e-01 9.54372704e-01
3.52216721e-01 -7.33298779e-01 4.17966902e-01 2.55978942e-01
-1.03352690e+00 -1.03816882e-01 -4.30752963e-01 -9.52896893e-01
7.17656314e-03 3.86318356e-01 -3.53551954e-02 5.23388684e-01
5.63904643e-01 2.52574414e-01 -4.21099871e-01 9.10369158e-01
-6.55421555e-01 2.73620039e-01 -1.37166724e-01 -9.82292295e-01
1.13551728e-01 -1.39191553e-01 2.56431073e-01 7.87982583e-01
2.35318214e-01 6.58305287e-01 -5.30359805e-01 1.69327283e+00
2.45583981e-01 -1.92829981e-01 -5.84550619e-01 -5.64514399e-01
4.57852840e-01 9.37499523e-01 -8.53862822e-01 -1.99293852e-01
-1.41783521e-01 8.87251198e-01 5.80127478e-01 3.54832172e-01
-7.37251401e-01 -6.92710280e-01 4.04548824e-01 -2.69880354e-01
-1.73645958e-01 -3.77182931e-01 -9.36286092e-01 -1.02036440e+00
3.47052254e-02 -1.03187728e+00 9.64673460e-02 -1.36969757e+00
-1.31429946e+00 2.63130218e-01 5.25440313e-02 -2.39398479e-01
-8.65886435e-02 -1.30804491e+00 -7.75912285e-01 9.66693461e-01
-1.32350230e+00 -7.87205994e-01 -2.64866590e-01 1.97171569e-02
5.93967021e-01 3.86978649e-02 1.38154912e+00 -1.40505597e-01
-2.65648693e-01 2.80530155e-01 -5.68976700e-01 -3.49733122e-02
-2.20872268e-01 -9.96137023e-01 2.87557751e-01 2.09121928e-01
-4.92983647e-02 1.17852008e+00 6.29149318e-01 -3.69553536e-01
-1.89330697e+00 -4.16975707e-01 1.25160193e+00 -6.65795207e-01
6.18809938e-01 -2.05627754e-01 -8.49186063e-01 3.65798116e-01
3.84960234e-01 -8.18109035e-01 8.28131258e-01 3.17195244e-02
-2.50906855e-01 2.30781361e-01 -1.13527226e+00 1.00428891e+00
1.20044160e+00 -6.56471312e-01 -1.18279958e+00 4.41788346e-01
8.75376344e-01 -5.49358964e-01 -5.77555001e-01 2.84982353e-01
8.22076857e-01 -8.69437337e-01 1.27770019e+00 -9.11922812e-01
1.09527624e+00 -2.42487639e-01 6.30383864e-02 -1.02369559e+00
-4.23370510e-01 -7.14493096e-01 -3.60828370e-01 8.95348489e-01
5.04983425e-01 -5.74558198e-01 8.79444957e-01 1.22052324e+00
5.75964563e-02 -1.25890481e+00 -5.96096098e-01 -4.84932661e-01
7.11857319e-01 -5.09961307e-01 6.35064304e-01 9.04846966e-01
4.99045134e-01 5.06604552e-01 1.14784554e-01 -2.39199206e-01
2.54939705e-01 4.02476788e-01 4.43057001e-01 -1.16989470e+00
-4.07020420e-01 -9.38550770e-01 -1.64178744e-01 -1.14976752e+00
4.96015221e-01 -1.15925932e+00 -4.08215046e-01 -2.06259775e+00
-7.93712065e-02 1.81622431e-01 4.83272374e-01 1.41671434e-01
-7.12215006e-02 -1.30274832e-01 2.22101301e-01 -4.10814077e-01
-1.97683692e-01 3.61525029e-01 1.56674123e+00 -8.45871568e-02
1.58145651e-02 -3.14506263e-01 -1.00845659e+00 8.43040049e-01
1.09929097e+00 -6.49571940e-02 -5.67852080e-01 -6.77833617e-01
9.71031785e-01 8.67682621e-02 3.21506530e-01 -8.83515298e-01
3.42331618e-01 -5.46328187e-01 7.25015521e-01 -3.45488876e-01
4.48524207e-01 -5.62886596e-01 -1.79815441e-01 6.30948961e-01
-5.96283972e-01 2.70120472e-01 5.60974598e-01 -2.70597879e-02
3.82641889e-02 -7.78834045e-01 5.36210477e-01 -6.00762963e-01
-6.85114086e-01 -3.19600642e-01 -8.76485229e-01 1.88965574e-01
7.09339380e-01 -4.40196127e-01 -5.39801419e-01 -5.50426662e-01
-5.20364583e-01 3.36495399e-01 -3.73680770e-01 1.47081345e-01
1.00038743e+00 -1.13120377e+00 -3.65375936e-01 -3.16018611e-01
-2.43363515e-01 -3.45368423e-02 3.34418148e-01 2.86108583e-01
-1.08792531e+00 8.98832917e-01 -1.54263839e-01 6.43093362e-02
-8.14257145e-01 5.18270195e-01 6.60772145e-01 -2.69212663e-01
-2.99502492e-01 1.16377211e+00 -1.63195692e-02 -9.22052145e-01
4.12207037e-01 -4.93611336e-01 -3.80012125e-01 -2.91083664e-01
6.93558872e-01 6.11338437e-01 -2.29395360e-01 2.98305780e-01
-8.36093426e-02 8.96104276e-01 4.42718089e-01 1.95076119e-03
1.34304273e+00 3.07028174e-01 -6.73506379e-01 3.38654935e-01
5.18307030e-01 -3.16893101e-01 -2.86797076e-01 1.52732238e-01
7.36927986e-02 -2.28987560e-02 -3.48420531e-01 -1.16914272e+00
-7.14143813e-01 1.20376980e+00 -9.90089998e-02 1.94250226e-01
9.66605306e-01 -7.48667419e-02 8.19323063e-01 1.11896682e+00
1.27363130e-01 -9.95066285e-01 3.08358133e-01 7.56718874e-01
1.13937747e+00 -6.16891861e-01 2.31311526e-02 -7.65508592e-01
-1.58119366e-01 1.98293066e+00 8.72685730e-01 -9.41410661e-02
3.00138593e-01 3.35348397e-01 -1.89069271e-01 -5.12702644e-01
-5.79991102e-01 -1.16087191e-01 3.02672654e-01 8.10108364e-01
8.08863282e-01 -2.19436958e-01 -2.85099834e-01 5.78038752e-01
-8.90007913e-01 1.96353823e-01 2.91505396e-01 8.91297638e-01
-5.99493027e-01 -9.58761156e-01 -5.59813619e-01 4.12111670e-01
-6.05135672e-02 -3.70597571e-01 -8.57523501e-01 5.80254495e-01
5.01596443e-02 1.08105743e+00 -2.02543616e-01 -1.72073413e-02
1.61437705e-01 4.43181455e-01 1.27238595e+00 -4.36620653e-01
-7.13950098e-01 -9.13810551e-01 2.34267458e-01 -3.44778597e-01
-2.07457483e-01 -2.65382409e-01 -1.48253417e+00 -7.76032209e-01
-6.41847253e-02 5.42966053e-02 3.87897819e-01 1.20793068e+00
1.56133071e-01 6.41651332e-01 -3.30184817e-01 -5.78324795e-01
-3.20251286e-01 -4.43070710e-01 1.64498031e-01 -2.72908121e-01
-1.46513641e-01 -3.09282392e-01 -8.73066708e-02 1.92951486e-02] | [9.444352149963379, 7.250920295715332] |
00c6616e-3a26-46e2-a0f3-c8a53613dc21 | dytanvo-joint-refinement-of-visual-odometry | 2209.08430 | null | https://arxiv.org/abs/2209.08430v4 | https://arxiv.org/pdf/2209.08430v4.pdf | DytanVO: Joint Refinement of Visual Odometry and Motion Segmentation in Dynamic Environments | Learning-based visual odometry (VO) algorithms achieve remarkable performance on common static scenes, benefiting from high-capacity models and massive annotated data, but tend to fail in dynamic, populated environments. Semantic segmentation is largely used to discard dynamic associations before estimating camera motions but at the cost of discarding static features and is hard to scale up to unseen categories. In this paper, we leverage the mutual dependence between camera ego-motion and motion segmentation and show that both can be jointly refined in a single learning-based framework. In particular, we present DytanVO, the first supervised learning-based VO method that deals with dynamic environments. It takes two consecutive monocular frames in real-time and predicts camera ego-motion in an iterative fashion. Our method achieves an average improvement of 27.7% in ATE over state-of-the-art VO solutions in real-world dynamic environments, and even performs competitively among dynamic visual SLAM systems which optimize the trajectory on the backend. Experiments on plentiful unseen environments also demonstrate our method's generalizability. | ['Sebastian Scherer', 'Wenshan Wang', 'Yilin Cai', 'Shihao Shen'] | 2022-09-17 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [-1.70300752e-01 -1.26174614e-01 -3.60114813e-01 -2.35697299e-01
-5.74829698e-01 -7.54547119e-01 3.94802868e-01 -2.74787158e-01
-5.00582278e-01 4.90239888e-01 1.33775756e-01 -1.19099922e-01
2.78905630e-01 -2.54202753e-01 -7.62160122e-01 -5.52395642e-01
-1.58445776e-01 9.55570221e-01 8.66671443e-01 -8.97831321e-02
3.26586962e-02 4.13973808e-01 -1.44412255e+00 -2.36484200e-01
6.99623346e-01 4.95648682e-01 6.32410407e-01 8.93275499e-01
1.39839463e-02 9.00416672e-01 -6.78068250e-02 3.98544269e-03
4.24088478e-01 -1.61289439e-01 -8.22232187e-01 4.97104287e-01
8.94376099e-01 -3.61077487e-01 -7.96340406e-01 1.13423014e+00
3.74757767e-01 3.04171741e-01 2.29145661e-01 -1.28751063e+00
2.78124660e-02 -9.16843340e-02 -5.12286842e-01 2.55177796e-01
5.11559904e-01 4.27677184e-01 9.91777182e-01 -8.41751039e-01
1.07761037e+00 1.16290438e+00 7.83036768e-01 4.39236194e-01
-1.10783315e+00 -2.21437827e-01 5.55604458e-01 2.77196229e-01
-1.41757572e+00 -5.71814656e-01 4.74894196e-01 -6.19363487e-01
1.27183700e+00 -1.36141926e-01 9.25755501e-01 9.53269005e-01
2.48679608e-01 1.20235491e+00 6.11865580e-01 -2.59910780e-03
1.92585438e-01 -2.87151396e-01 -1.57065004e-01 1.02571881e+00
3.67384166e-01 7.66720697e-02 -6.87416434e-01 1.70882970e-01
5.88364959e-01 -5.80096617e-02 -2.91668564e-01 -1.27721870e+00
-1.50509048e+00 6.97868526e-01 5.14034092e-01 -1.88604966e-01
-5.29082492e-02 5.89245677e-01 2.35570952e-01 1.69365242e-01
3.62625599e-01 2.73497730e-01 -5.53776264e-01 -3.50588351e-01
-1.01562905e+00 2.95849472e-01 7.73600280e-01 1.35807598e+00
1.02637935e+00 2.52673566e-01 4.10684615e-01 4.69146788e-01
2.84570873e-01 7.82578826e-01 3.69645685e-01 -1.35881579e+00
4.30294663e-01 2.90896744e-01 3.89760345e-01 -8.60377908e-01
-7.40504324e-01 -2.61588037e-01 -4.18495387e-01 4.14844928e-03
3.68447363e-01 -1.58018604e-01 -1.32267439e+00 1.50952637e+00
4.63253647e-01 5.09236038e-01 -3.44468467e-02 1.28550673e+00
5.78704476e-01 4.04792279e-01 -1.00023530e-01 -1.16482332e-01
7.50320613e-01 -1.46739733e+00 -4.89882529e-01 -9.54830825e-01
5.74905038e-01 -5.37659883e-01 6.55772507e-01 2.95775712e-01
-7.17862010e-01 -4.65901434e-01 -9.77005839e-01 -2.06467956e-01
-1.75061911e-01 -2.03507900e-01 8.53621006e-01 2.66682625e-01
-1.30807078e+00 2.65511215e-01 -1.26681781e+00 -8.74501228e-01
1.82499006e-01 4.02415991e-01 -4.73136723e-01 -3.14874142e-01
-6.36050820e-01 8.73087585e-01 3.96073282e-01 -1.75994206e-02
-1.32232642e+00 -2.79379994e-01 -1.37550759e+00 -4.10244018e-01
6.75225973e-01 -9.86430287e-01 1.26583338e+00 -7.21665859e-01
-1.27954364e+00 8.40814054e-01 -6.13611698e-01 -5.63570201e-01
9.40918267e-01 -5.14909506e-01 -1.97264940e-01 9.41436216e-02
3.65115076e-01 9.28750634e-01 6.38618588e-01 -1.37906408e+00
-1.00116062e+00 -2.55963445e-01 -7.55564659e-04 6.77611887e-01
2.89635539e-01 -6.32822871e-01 -1.23532557e+00 -1.11517049e-01
8.46008301e-01 -1.44650507e+00 -6.66107178e-01 -2.58938931e-02
-3.12986523e-01 3.32912683e-01 8.47303689e-01 -3.40981364e-01
7.85215199e-01 -1.82272577e+00 5.28825045e-01 -1.02833003e-01
3.27309877e-01 5.65039506e-03 1.03988647e-01 9.83321071e-02
6.45236135e-01 -4.34910774e-01 -1.60127953e-01 -7.23712921e-01
-1.24561356e-03 5.93199968e-01 -1.43270209e-01 9.10830677e-01
-2.43039966e-01 1.05810177e+00 -1.32670796e+00 -5.55813193e-01
6.90001607e-01 -3.34787183e-02 -6.81387663e-01 2.85301115e-02
-3.26950073e-01 6.29903197e-01 -2.11591244e-01 8.57207239e-01
5.09315431e-01 -2.78468698e-01 3.57669443e-01 9.95787382e-02
-7.03103170e-02 9.47764069e-02 -1.28215504e+00 2.26772475e+00
-1.84336677e-01 1.09008300e+00 9.47285220e-02 -5.77869058e-01
5.69581330e-01 -2.85786420e-01 5.45621932e-01 -5.64298630e-01
4.95333690e-03 2.51559764e-01 -3.50857079e-01 -4.87493187e-01
1.05710256e+00 8.84385854e-02 -3.57929707e-01 -1.29305184e-01
2.27305204e-01 -4.78818625e-01 9.31802839e-02 2.79160768e-01
1.00763214e+00 6.69580936e-01 4.47698116e-01 -1.91926092e-01
3.27046782e-01 5.28747439e-01 9.37208831e-01 8.75228941e-01
-7.30074644e-01 7.33413875e-01 1.67180635e-02 -7.46602893e-01
-1.14278352e+00 -1.20084465e+00 2.71150768e-01 8.25546741e-01
1.17131853e+00 -4.11857963e-01 -3.57316971e-01 -7.48004138e-01
2.63609529e-01 4.04538274e-01 -2.76884079e-01 1.46838516e-01
-6.63953543e-01 -5.79375505e-01 2.90718824e-01 5.01689970e-01
4.80539441e-01 -6.81170404e-01 -8.41075242e-01 3.64721209e-01
-4.86666381e-01 -1.66638970e+00 -4.00699586e-01 2.55143106e-01
-8.53261352e-01 -1.06523788e+00 -3.81884277e-01 -6.01779878e-01
3.97708714e-01 8.13682735e-01 1.16823447e+00 -2.54171699e-01
-1.60637334e-01 4.95404750e-01 -1.68841019e-01 -7.73691982e-02
-4.66302931e-02 2.60619968e-01 4.52603787e-01 -2.85279334e-01
5.25058389e-01 -3.42701972e-01 -5.68368018e-01 3.85784924e-01
-3.58998686e-01 1.57788217e-01 2.75301903e-01 7.53326833e-01
8.35014462e-01 -4.02233839e-01 -2.30596095e-01 -7.17256308e-01
-2.96719193e-01 -5.47910094e-01 -8.14587653e-01 -3.77598442e-02
-6.26217127e-01 1.33528620e-01 1.60225675e-01 -2.62288868e-01
-9.50672626e-01 5.74352860e-01 2.00590149e-01 -7.96460569e-01
-1.71671286e-01 7.98936188e-02 -4.43159156e-02 -3.87476683e-01
5.38856149e-01 2.75053173e-01 -3.37432176e-01 -2.95372188e-01
4.87834811e-01 2.42471442e-01 9.44480360e-01 -2.08159983e-01
1.04681385e+00 1.03408992e+00 -7.36640245e-02 -8.20103705e-01
-8.71823370e-01 -1.18347692e+00 -9.45066273e-01 -3.34307462e-01
9.52219129e-01 -1.48263657e+00 -3.49677593e-01 5.53012013e-01
-9.12957489e-01 -7.08606482e-01 -3.16658393e-02 6.33474469e-01
-9.00432229e-01 5.92952251e-01 -5.26958466e-01 -6.44022822e-01
1.17053106e-01 -1.06427610e+00 1.25635278e+00 1.51989788e-01
-4.84132431e-02 -9.85595882e-01 3.04475754e-01 2.91677386e-01
-8.28758776e-02 3.15181047e-01 4.18898985e-02 -2.22153708e-01
-1.18139350e+00 -5.83089655e-03 -6.64702207e-02 -2.10126847e-01
-1.79991394e-01 -3.90462011e-01 -9.02559578e-01 -5.79273224e-01
-4.77589190e-01 -1.94555953e-01 1.21123278e+00 4.34342891e-01
4.71956015e-01 8.08229372e-02 -6.56874180e-01 1.30293655e+00
1.63994932e+00 -9.03721079e-02 4.74866509e-01 5.95115304e-01
1.26014733e+00 3.10946018e-01 9.27337110e-01 4.40379143e-01
8.12054396e-01 7.65789688e-01 9.15471017e-01 -5.36826737e-02
-1.29239872e-01 -5.36515892e-01 3.74699712e-01 8.44714224e-01
2.80859135e-02 -3.15400630e-01 -1.06505334e+00 1.09366405e+00
-2.28044605e+00 -9.27533984e-01 -2.95849383e-01 2.00013018e+00
1.11724161e-01 3.31080317e-01 3.17305722e-03 -5.56587040e-01
4.05153185e-01 6.04011834e-01 -8.33124936e-01 3.51707935e-02
-3.22000563e-01 -5.34682393e-01 1.25385106e+00 8.84312987e-01
-1.37961936e+00 1.78586435e+00 6.23181152e+00 2.56003767e-01
-9.44240987e-01 3.31550390e-01 -4.27207984e-02 -3.13763171e-01
-1.10726826e-01 3.68389249e-01 -9.38750744e-01 2.11301997e-01
6.24491155e-01 4.33385512e-03 4.75881100e-01 1.10816312e+00
1.29431501e-01 -4.51963633e-01 -9.11155999e-01 1.32278001e+00
1.21013157e-01 -1.26132846e+00 -2.76902795e-01 1.36758447e-01
1.07950807e+00 1.03813696e+00 -3.45142901e-01 1.78481713e-01
8.03906977e-01 -6.67602181e-01 1.15706325e+00 2.71793634e-01
5.21663487e-01 -6.12479389e-01 7.31044173e-01 5.43965042e-01
-1.54055440e+00 -1.79514229e-01 -4.83650863e-01 -1.87972039e-01
5.91454268e-01 2.40846172e-01 -9.16612625e-01 6.43299639e-01
9.21718180e-01 1.14536154e+00 -5.37704647e-01 1.28192782e+00
-3.30690622e-01 8.02252293e-02 -4.64185625e-01 2.70834684e-01
6.77784085e-01 -3.76167633e-02 8.73689950e-01 1.34983587e+00
3.24007392e-01 3.83478380e-03 7.44850814e-01 3.53983551e-01
1.95272431e-01 -4.73749429e-01 -9.16781604e-01 3.22353333e-01
3.74368370e-01 9.53483343e-01 -6.22981668e-01 -4.23850298e-01
-4.43526506e-01 1.50584280e+00 3.80201936e-01 4.90484446e-01
-7.71176994e-01 8.61068815e-02 1.07499897e+00 -1.49467355e-03
5.68088710e-01 -8.98535252e-01 -2.01593656e-02 -1.69489884e+00
-1.41151905e-01 -3.20513338e-01 2.66504496e-01 -8.10390472e-01
-7.55641878e-01 4.76941943e-01 -1.37196854e-01 -1.45986974e+00
-5.21836817e-01 -4.29019302e-01 -2.80833486e-02 3.94947439e-01
-1.62872112e+00 -1.13781774e+00 -8.30757260e-01 6.64518833e-01
9.26376760e-01 2.31185347e-01 3.53662699e-01 1.60065979e-01
-2.36506864e-01 3.57122496e-02 2.10150018e-01 4.80818935e-02
7.06952393e-01 -1.38772154e+00 8.47287297e-01 1.31635201e+00
4.89994377e-01 6.73652589e-02 9.42781866e-01 -7.01812923e-01
-1.60260808e+00 -1.19901741e+00 8.74855280e-01 -8.33865762e-01
5.89045584e-01 -5.11631727e-01 -5.70960283e-01 1.11691999e+00
-1.21190481e-01 3.68834227e-01 1.35809695e-03 -1.71350867e-01
-2.00326532e-01 1.90607548e-01 -7.73977220e-01 7.51583695e-01
1.72706532e+00 -4.98176068e-01 -4.01026517e-01 2.59494931e-01
8.08309317e-01 -1.01986456e+00 -3.45918089e-01 4.30736363e-01
4.61781919e-01 -1.13878751e+00 1.09981894e+00 -3.96821856e-01
-2.98555166e-01 -7.92193115e-01 -3.80954117e-01 -1.05563176e+00
-3.44105929e-01 -8.52553725e-01 -3.56597960e-01 5.10538936e-01
1.42650917e-01 -3.28240365e-01 1.10677457e+00 3.59228134e-01
-3.07760864e-01 -9.31878760e-02 -9.03452098e-01 -1.03465676e+00
-5.41990995e-01 -7.13061869e-01 1.59687445e-01 9.18781936e-01
-2.63222218e-01 1.97790727e-01 -7.41404653e-01 6.07297897e-01
9.73192871e-01 8.25712010e-02 1.39918208e+00 -1.07139468e+00
-3.31832319e-01 -4.92410921e-02 -9.15422797e-01 -1.78132653e+00
3.08321178e-01 -6.02833271e-01 5.56745946e-01 -1.67626059e+00
1.01132385e-01 -2.85805911e-01 4.34816927e-02 2.02215537e-01
-1.44973904e-01 3.23612601e-01 3.21034074e-01 6.03738785e-01
-1.20413530e+00 4.82632935e-01 1.00812542e+00 -2.40743026e-01
-5.12039840e-01 -2.36958385e-01 1.07442802e-02 1.09676385e+00
4.08021569e-01 -3.80225331e-01 -6.21013820e-01 -7.98210323e-01
-4.21471298e-02 1.80096496e-02 4.48001921e-01 -1.21137464e+00
6.41808808e-01 -4.23542619e-01 2.77207822e-01 -8.28135908e-01
5.46996474e-01 -6.98461056e-01 2.83564419e-01 4.41337764e-01
2.99854249e-01 1.26771584e-01 1.99849084e-01 1.02205205e+00
-2.43246883e-01 2.34975219e-01 6.04368329e-01 -3.96306783e-01
-1.98555398e+00 5.99861801e-01 -2.79736727e-01 2.47413233e-01
9.67407048e-01 -5.33742905e-01 -1.93211406e-01 -4.88189906e-01
-8.18278909e-01 5.07327616e-01 1.08499956e+00 6.74398839e-01
5.77672958e-01 -1.02007222e+00 -2.91105479e-01 4.97834384e-02
3.98565441e-01 4.10333008e-01 2.91420728e-01 9.89683807e-01
-9.81009364e-01 4.97419298e-01 -1.65796926e-04 -1.26124370e+00
-1.01147699e+00 4.85655218e-01 4.47784424e-01 -9.03515890e-02
-9.54800367e-01 8.21687102e-01 3.08957368e-01 -3.85611087e-01
1.72688484e-01 -2.50524074e-01 2.25704968e-01 -1.94018230e-01
1.39057979e-01 2.93907762e-01 -1.18721262e-01 -1.01002181e+00
-6.59548521e-01 8.08839381e-01 1.58205718e-01 -2.47821301e-01
1.05283177e+00 -8.12663436e-01 2.20449969e-01 5.78130126e-01
1.09870183e+00 -9.67049226e-02 -2.07289481e+00 -3.56274664e-01
2.33413547e-01 -7.42469251e-01 -1.34324804e-02 -1.82436243e-01
-8.94067407e-01 6.47316933e-01 4.62129474e-01 -3.30960721e-01
6.74865603e-01 2.98919082e-01 8.27673614e-01 5.86889029e-01
1.06689095e+00 -1.17482662e+00 -1.70679111e-02 8.93502653e-01
1.73854962e-01 -1.53049839e+00 9.59979817e-02 -4.84798193e-01
-8.61975372e-01 8.26350570e-01 7.41312563e-01 -1.77402720e-01
1.64961398e-01 1.82354599e-01 1.93290040e-01 -1.36169940e-01
-6.65385067e-01 -6.72774971e-01 6.43972307e-02 7.60004818e-01
-3.07498366e-01 5.18169068e-02 2.80114174e-01 -1.93849266e-01
-6.83942288e-02 -1.95321202e-01 5.80736101e-01 9.90214348e-01
-7.06964612e-01 -6.30157053e-01 -1.39294207e-01 -8.92692134e-02
1.63052361e-02 4.89239246e-02 -2.19764933e-01 1.14600551e+00
7.45976418e-02 7.91306436e-01 1.16505034e-01 -5.15378296e-01
2.06595585e-01 -1.21511534e-01 4.36710447e-01 -4.53766346e-01
-6.05659299e-02 2.13883936e-01 2.95470357e-01 -1.22306418e+00
-4.67706561e-01 -9.22504485e-01 -1.27850091e+00 -4.32112277e-01
-5.73816523e-02 -2.45934322e-01 5.51850855e-01 1.00084829e+00
3.79368424e-01 2.74447381e-01 2.21497312e-01 -1.33620977e+00
-1.38525024e-01 -4.16444868e-01 -3.83261830e-01 4.21821147e-01
8.35450888e-01 -8.32157850e-01 -3.91441643e-01 1.49819031e-01] | [8.10881519317627, -2.1210713386535645] |
4c3489b8-8dab-46af-980d-dece0ebda42a | recursive-generalization-transformer-for | 2303.06373 | null | https://arxiv.org/abs/2303.06373v2 | https://arxiv.org/pdf/2303.06373v2.pdf | Recursive Generalization Transformer for Image Super-Resolution | Transformer architectures have exhibited remarkable performance in image super-resolution (SR). Since the quadratic computational complexity of the self-attention (SA) in Transformer, existing methods tend to adopt SA in a local region to reduce overheads. However, the local design restricts the global context exploitation, which is critical for accurate image reconstruction. In this work, we propose the Recursive Generalization Transformer (RGT) for image SR, which can capture global spatial information and is suitable for high-resolution images. Specifically, we propose the recursive-generalization self-attention (RG-SA). It recursively aggregates input features into representative feature maps, and then utilizes cross-attention to extract global information. Meanwhile, the channel dimensions of attention matrices (query, key, and value) are further scaled for a better trade-off between computational overheads and performance. Furthermore, we combine the RG-SA with local self-attention to enhance the exploitation of the global context, and propose the hybrid adaptive integration (HAI) for module integration. The HAI allows the direct and effective fusion between features at different levels (local or global). Extensive experiments demonstrate that our RGT outperforms recent state-of-the-art methods. | ['Xiaokang Yang', 'Linghe Kong', 'Jinjin Gu', 'Yulun Zhang', 'Zheng Chen'] | 2023-03-11 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 1.64740697e-01 -3.24497521e-01 -6.76443428e-02 -2.61791974e-01
-8.65807891e-01 4.74715009e-02 3.52424055e-01 -1.99564472e-01
-1.17582813e-01 3.81109267e-01 4.95515436e-01 4.94640172e-02
-3.43760103e-01 -1.00676477e+00 -5.25783837e-01 -8.24835718e-01
1.25695020e-01 -2.79512703e-01 5.05897105e-01 -3.15931082e-01
1.95403114e-01 5.55672109e-01 -1.63408279e+00 3.77560794e-01
1.21060562e+00 1.29200089e+00 9.57400620e-01 2.43780121e-01
-2.20868498e-01 8.18745613e-01 -3.55612248e-01 -1.59144893e-01
2.03233406e-01 -3.53804082e-01 -6.14453018e-01 -8.42254758e-02
4.76175696e-01 -5.54576159e-01 -4.15821761e-01 1.19551802e+00
6.86940730e-01 1.12021372e-01 1.49583161e-01 -7.28269994e-01
-1.00868928e+00 5.60985327e-01 -8.74162614e-01 6.69242918e-01
1.48384884e-01 4.42589633e-02 9.53261137e-01 -1.14850605e+00
2.83531070e-01 1.52548158e+00 3.88285458e-01 1.64571613e-01
-1.13511038e+00 -9.39333677e-01 3.26231867e-01 5.13229966e-01
-1.71529698e+00 -4.62896228e-01 6.93402290e-01 -1.40376240e-01
8.69867623e-01 2.79517889e-01 5.62447548e-01 6.32541478e-01
1.54324859e-01 7.51838624e-01 1.21069252e+00 -1.72284991e-01
-9.96087957e-03 9.47301239e-02 -1.38793975e-01 5.40638447e-01
-5.90900034e-02 2.31030658e-02 -6.73285961e-01 1.51709482e-01
1.53508198e+00 1.94769576e-01 -5.54761291e-01 -1.84433714e-01
-1.18706155e+00 6.35480583e-01 9.39363599e-01 3.75713170e-01
-6.02398276e-01 -1.10349702e-02 2.32120812e-01 1.49676427e-01
3.07373822e-01 2.69923478e-01 -2.08412111e-01 2.94167638e-01
-7.34407425e-01 -8.51501301e-02 -2.45230779e-01 9.46180880e-01
8.78407121e-01 6.40663505e-02 -5.58601439e-01 1.03504121e+00
-9.42648351e-02 2.65745878e-01 7.06376910e-01 -7.98441350e-01
4.54876333e-01 7.90274620e-01 3.24261189e-03 -1.15785086e+00
-1.82510093e-01 -8.73806119e-01 -1.07985270e+00 -6.78647086e-02
-3.31928790e-01 3.50070596e-01 -7.69910216e-01 1.69521248e+00
2.64285952e-01 1.96074203e-01 3.76782604e-02 1.17124271e+00
9.76039112e-01 7.21660435e-01 1.27209514e-01 -2.65862435e-01
1.44125021e+00 -1.03713858e+00 -8.10516715e-01 -2.15444490e-01
2.08613984e-02 -6.40604436e-01 1.18154407e+00 1.08500272e-01
-1.12491655e+00 -1.12735975e+00 -1.11044109e+00 -4.57152247e-01
-1.98699534e-01 3.23593676e-01 4.05438602e-01 2.37988323e-01
-1.06295872e+00 3.86186004e-01 -4.89581466e-01 -1.03466757e-01
5.44839919e-01 4.97316122e-01 -2.29376420e-01 -1.45386755e-01
-1.22900391e+00 4.83969241e-01 2.68876076e-01 8.56469572e-02
-6.00357652e-01 -6.41543686e-01 -8.75881314e-01 3.50671440e-01
4.00850326e-01 -5.43073416e-01 8.93587112e-01 -7.98367977e-01
-1.44122291e+00 3.48136812e-01 -3.78650784e-01 -2.62887776e-01
6.61626831e-02 -3.03541034e-01 -4.70888108e-01 3.05304021e-01
2.30046675e-01 6.42585039e-01 1.10585392e+00 -1.17529893e+00
-7.87936985e-01 -5.22566020e-01 2.11083367e-01 5.25627851e-01
-6.87115252e-01 1.35340825e-01 -6.92074358e-01 -8.87617826e-01
3.96874934e-01 -3.38174373e-01 -1.72007814e-01 -3.19340438e-01
-1.23700798e-01 4.44963723e-02 9.07060623e-01 -7.41215110e-01
1.78247452e+00 -2.44758487e+00 3.16536784e-01 7.29037728e-03
3.92546415e-01 2.41334736e-01 -9.92145687e-02 -4.49855253e-02
-9.01407748e-02 5.71006499e-02 -5.55123165e-02 -2.24205807e-01
-4.75640476e-01 5.61296344e-02 -2.47481212e-01 5.59295230e-02
3.16646576e-01 9.73421693e-01 -7.40181148e-01 -6.74582481e-01
1.82230815e-01 6.57245278e-01 -5.74059486e-01 2.60918081e-01
2.59724021e-01 5.19433737e-01 -7.74873614e-01 6.52318418e-01
8.92689884e-01 -5.79847395e-01 1.13910567e-02 -8.07193696e-01
-3.63745332e-01 1.41772419e-01 -1.13894248e+00 1.53505754e+00
-8.21870685e-01 2.41728038e-01 7.57500753e-02 -5.47608614e-01
1.03048158e+00 -8.99396688e-02 2.42719501e-01 -1.21320724e+00
1.04662990e-02 8.84403959e-02 -1.25809625e-01 -2.54869670e-01
7.41120398e-01 1.58494338e-01 1.37382358e-01 2.66338941e-02
-1.05241247e-01 5.15292704e-01 -1.98177695e-01 1.01577766e-01
6.67665064e-01 1.05854541e-01 5.27912080e-01 -3.57545972e-01
9.27268863e-01 -4.63624120e-01 7.95775235e-01 4.49549586e-01
1.40724005e-02 7.22394884e-01 1.99800953e-01 -4.78094906e-01
-7.73574591e-01 -9.64741409e-01 -1.36832207e-01 1.16202307e+00
6.73849702e-01 -4.90613431e-01 -6.96880043e-01 -5.30417144e-01
-2.00208649e-01 2.99368531e-01 -6.52025700e-01 -1.20404281e-01
-8.06448936e-01 -6.57195687e-01 -1.41564328e-02 8.59708905e-01
1.14745617e+00 -9.75662947e-01 -8.37779582e-01 2.52768904e-01
-3.12712312e-01 -9.83798862e-01 -7.46041119e-01 -2.19938546e-01
-8.07479382e-01 -6.01025939e-01 -7.11100698e-01 -5.84583521e-01
5.51088214e-01 7.23289788e-01 8.17160785e-01 6.46081641e-02
-1.05841279e-01 -6.08333386e-03 -4.22392607e-01 1.34316117e-01
2.25529030e-01 3.87529507e-02 -1.38349116e-01 4.11645293e-01
9.68208164e-02 -7.28693664e-01 -1.07715952e+00 4.93509322e-01
-9.35591757e-01 2.99010366e-01 1.12166238e+00 1.06091821e+00
9.84224200e-01 1.78255320e-01 6.94677472e-01 -6.20128810e-01
4.20523286e-01 -3.66780013e-01 -4.37394261e-01 2.82559693e-01
-4.66017842e-01 3.36177126e-02 7.11800694e-01 -4.00138438e-01
-1.35050595e+00 -1.76712021e-01 -9.87261161e-03 -7.08227098e-01
3.16766918e-01 3.80858481e-01 -5.64398527e-01 -3.45582575e-01
2.23963842e-01 5.99341750e-01 -1.97965071e-01 -7.11159945e-01
2.85927296e-01 6.80421054e-01 5.00892818e-01 -3.35468382e-01
4.76528674e-01 3.84697109e-01 -1.83022931e-01 -6.55425668e-01
-7.66301453e-01 -2.36652568e-01 -3.54699939e-01 7.66490921e-02
8.34336340e-01 -1.15253794e+00 -4.95803326e-01 3.80817592e-01
-8.12274039e-01 -8.18392932e-02 -2.21319869e-01 2.27767348e-01
-3.90720576e-01 2.80864090e-01 -5.89183569e-01 -7.20216990e-01
-5.62119186e-01 -1.46242857e+00 1.13765693e+00 5.96608698e-01
3.25816900e-01 -4.68320787e-01 -5.93866944e-01 2.29559615e-01
9.18937922e-01 -9.35985595e-02 7.92663574e-01 -1.31002963e-01
-1.08679557e+00 2.95618564e-01 -9.80369985e-01 2.14203864e-01
2.80355513e-01 -5.54609656e-01 -8.70436728e-01 -3.84811908e-01
-4.21556830e-02 4.06606421e-02 8.87986302e-01 3.30792636e-01
1.51098931e+00 -4.48554695e-01 -3.58713329e-01 8.42607796e-01
1.61121058e+00 1.97557420e-01 8.52086723e-01 5.15213668e-01
8.86657953e-01 2.72424370e-01 9.83515978e-01 5.06609976e-01
5.26844084e-01 1.01541770e+00 3.73454273e-01 -2.97901332e-01
-3.95451754e-01 -2.83190906e-01 2.94026971e-01 7.77399600e-01
-1.62142441e-01 2.27715477e-01 -4.87251610e-01 6.16425693e-01
-1.75401676e+00 -8.01562369e-01 4.28232163e-01 2.30369186e+00
6.93911314e-01 3.26297595e-03 -1.11028165e-01 -1.21038876e-01
6.87322974e-01 3.48308980e-01 -5.93298197e-01 -4.70623598e-02
-2.92743683e-01 2.46109247e-01 5.00128984e-01 3.66196305e-01
-1.07567835e+00 8.46876979e-01 5.45797110e+00 1.36339521e+00
-1.21134126e+00 3.45629156e-01 7.25110292e-01 1.62292900e-03
-3.64887208e-01 -1.55367732e-01 -9.45037961e-01 5.11423290e-01
5.15249848e-01 -1.26497477e-01 3.79693180e-01 6.71975970e-01
9.85630322e-04 9.21401531e-02 -5.10813713e-01 1.01644909e+00
-2.79868580e-02 -1.18020272e+00 4.00374115e-01 6.51772246e-02
6.92370474e-01 -1.96234643e-01 3.44105422e-01 2.98073947e-01
-8.17637295e-02 -8.21329534e-01 6.07131541e-01 5.50341666e-01
1.12012887e+00 -9.46046412e-01 6.97671533e-01 9.73698944e-02
-1.79235685e+00 -7.12286055e-01 -4.04746175e-01 3.87749046e-01
-1.02252746e-02 4.40399110e-01 -5.53323254e-02 9.27993774e-01
1.11941433e+00 7.51837850e-01 -8.53632271e-01 7.09232688e-01
1.87201664e-01 -7.42406445e-03 -1.66271195e-01 3.85641307e-01
1.69923469e-01 -1.86323896e-01 5.58063030e-01 9.80416059e-01
4.96605903e-01 3.35926920e-01 1.40074551e-01 8.25010240e-01
1.50116637e-01 2.43033379e-01 -1.67482898e-01 3.46813053e-01
7.28192389e-01 1.22196305e+00 -3.68065059e-01 -4.17964458e-01
-4.53320473e-01 1.03402495e+00 4.96349424e-01 3.24298382e-01
-6.98542714e-01 -4.83847737e-01 7.57294893e-01 3.03244501e-01
7.35844910e-01 -3.85075659e-02 -2.93980300e-01 -1.05594647e+00
4.52447347e-02 -9.07582045e-01 3.99752557e-01 -8.28839123e-01
-9.57295239e-01 1.00628674e+00 5.20579033e-02 -1.47672999e+00
2.38263920e-01 -7.67145380e-02 -3.94508988e-01 1.04329538e+00
-1.68640387e+00 -1.31545866e+00 -6.07434154e-01 7.66039550e-01
7.07471967e-01 -4.75568660e-02 3.98343414e-01 4.66915786e-01
-8.07324052e-01 8.55180025e-01 -1.64900750e-01 -1.62997067e-01
5.36772728e-01 -9.19901609e-01 1.99391201e-01 9.48991001e-01
-2.05813244e-01 9.11957920e-01 3.30950469e-01 -5.53369582e-01
-1.32006204e+00 -1.15789998e+00 3.57571721e-01 1.45054772e-01
2.79137850e-01 -2.70671770e-02 -1.14270496e+00 4.92337048e-01
-3.88492905e-02 2.36247852e-01 2.51346529e-01 -2.32977554e-01
-5.72917104e-01 -6.05476320e-01 -1.15088224e+00 6.65839851e-01
1.12654173e+00 -5.73712349e-01 -2.24089444e-01 -1.62031546e-01
1.07526994e+00 -3.79827261e-01 -1.14047205e+00 6.55576110e-01
5.56917012e-01 -1.30792141e+00 1.29488480e+00 1.32474616e-01
4.78907317e-01 -7.25862622e-01 -4.33041304e-01 -9.49740171e-01
-1.00718307e+00 -3.28157246e-01 -2.19844043e-01 1.29251456e+00
3.22169177e-02 -6.98322415e-01 3.07464749e-01 9.93487388e-02
-1.15671858e-01 -1.08475184e+00 -8.98357689e-01 -6.35553956e-01
-3.92917097e-01 1.51660368e-01 1.09849715e+00 7.27788866e-01
-3.28898966e-01 3.90831053e-01 -4.58071142e-01 3.61706972e-01
6.65049016e-01 4.28965837e-01 4.71353680e-01 -8.61003160e-01
-3.11538160e-01 -5.59282839e-01 -4.72713947e-01 -1.40061545e+00
-4.45163816e-01 -4.01365489e-01 -2.66064107e-01 -1.44784403e+00
4.30456489e-01 -4.93951619e-01 -6.89914823e-01 4.19211775e-01
-5.51804066e-01 3.23979557e-01 1.92310765e-01 4.35217798e-01
-6.71816766e-01 7.59210110e-01 1.44369245e+00 1.17556877e-01
-3.18889797e-01 -4.01616722e-01 -9.91438270e-01 4.29425001e-01
7.12757230e-01 4.88612764e-02 -4.44117367e-01 -5.04327118e-01
-1.66457608e-01 8.87658074e-02 2.83580720e-01 -1.08680010e+00
3.12457055e-01 -1.70441605e-02 5.53754866e-01 -7.76523113e-01
4.02223170e-01 -7.61663556e-01 1.50313035e-01 1.02300964e-01
-1.36444077e-01 1.29704416e-01 1.93198934e-01 6.02231503e-01
-4.60193425e-01 3.25013906e-01 9.67953503e-01 -1.36399895e-01
-9.43811715e-01 5.65258980e-01 2.17486069e-01 -4.48543489e-01
7.97353566e-01 -2.35430434e-01 -3.60459387e-01 -2.90009260e-01
-4.41866994e-01 2.34843701e-01 5.66229522e-01 3.74032319e-01
9.93668079e-01 -1.55934882e+00 -6.28507972e-01 5.21202445e-01
1.23556755e-01 1.29518643e-01 9.58101392e-01 9.97662485e-01
-2.26599231e-01 2.21821770e-01 -4.51275527e-01 -5.51837862e-01
-1.20555639e+00 8.60644996e-01 1.82904989e-01 -4.45049435e-01
-9.16953087e-01 7.21603274e-01 7.99336851e-01 1.45017445e-01
-7.78447762e-02 3.68612334e-02 -6.12697363e-01 -1.28064558e-01
1.03599060e+00 3.36316079e-01 -1.01618871e-01 -8.46221745e-01
-1.39771268e-01 9.52251315e-01 -4.65053201e-01 2.61008233e-01
1.07705235e+00 -7.56538570e-01 -2.44267419e-01 9.10821110e-02
1.03049934e+00 1.79490343e-01 -1.25264847e+00 -6.91397667e-01
-3.18062961e-01 -9.28130627e-01 5.39787710e-01 -5.69799244e-01
-1.41294730e+00 7.01839089e-01 7.34766483e-01 8.31732005e-02
1.87894452e+00 -9.78634730e-02 9.53583419e-01 -1.77679405e-01
4.95995730e-01 -7.00021744e-01 2.72408873e-02 2.06996158e-01
1.12373996e+00 -9.46104169e-01 2.25038886e-01 -6.63555622e-01
-6.23726606e-01 7.93788254e-01 1.00443816e+00 4.48761135e-02
4.17837590e-01 1.60144478e-01 -1.94120705e-01 -1.62254274e-02
-6.07167780e-01 -4.33715373e-01 3.78327549e-01 4.49987561e-01
2.10828677e-01 -1.25597036e-02 -1.51719421e-01 8.86797011e-01
2.37155054e-02 -6.13474622e-02 1.52410775e-01 6.68378770e-01
-4.82522011e-01 -8.09603035e-01 -3.29818517e-01 3.98126334e-01
-4.12253976e-01 -3.36586624e-01 5.29355444e-02 5.26239276e-01
2.61113763e-01 6.70057416e-01 8.05716366e-02 -6.61296129e-01
3.04106116e-01 -5.04588127e-01 4.63274181e-01 -2.62751877e-01
-4.68173057e-01 4.18361664e-01 -2.94588774e-01 -8.34279120e-01
-3.55002254e-01 -3.08291286e-01 -9.92236376e-01 -1.17606401e-01
-3.62932444e-01 -9.20297205e-02 3.05321872e-01 6.17898583e-01
7.99313009e-01 8.40909421e-01 8.48624587e-01 -9.00162458e-01
-2.72595376e-01 -9.28625286e-01 -4.92844671e-01 1.19672984e-01
3.90019655e-01 -9.15023804e-01 -1.08709507e-01 -2.87531167e-01] | [10.889854431152344, -1.8447080850601196] |
dc5e007e-0c9d-4f56-9899-f78694b03b71 | ts-sep-joint-diarization-and-separation | 2303.03849 | null | https://arxiv.org/abs/2303.03849v2 | https://arxiv.org/pdf/2303.03849v2.pdf | TS-SEP: Joint Diarization and Separation Conditioned on Estimated Speaker Embeddings | Since diarization and source separation of meeting data are closely related tasks, we here propose an approach to perform the two objectives jointly. It builds upon the target-speaker voice activity detection (TS-VAD) diarization approach, which assumes that initial speaker embeddings are available. We replace the final combined speaker activity estimation network of TS-VAD with a network that produces speaker activity estimates at a time-frequency resolution. Those act as masks for source extraction, either via masking or via beamforming. The technique can be applied both for single-channel and multi-channel input and, in both cases, achieves a new state-of-the-art word error rate (WER) on the LibriCSS meeting data recognition task. We further compute speaker-aware and speaker-agnostic WERs to isolate the contribution of diarization errors to the overall WER performance. | ['Jonathan Le Roux', 'Reinhold Haeb-Umbach', 'Gordon Wichern', 'Aswin Shanmugam Subramanian', 'Christoph Boeddeker'] | 2023-03-07 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 4.83866304e-01 1.93097144e-01 1.13049164e-01 -3.93860281e-01
-1.66435087e+00 -6.42445922e-01 7.67630577e-01 3.79319638e-02
-4.07535255e-01 3.13951850e-01 6.83959544e-01 -2.79258072e-01
2.53823735e-02 -1.54151976e-01 -2.51695693e-01 -8.74795437e-01
2.78155133e-02 3.16838883e-02 5.08791767e-02 1.10514835e-01
-9.34386924e-02 5.10203660e-01 -1.62682354e+00 3.11477691e-01
6.17435336e-01 8.49122643e-01 -6.72187377e-03 1.17432570e+00
-1.62875786e-01 4.12443787e-01 -1.20367420e+00 1.77828074e-01
-1.33097116e-02 -6.92950964e-01 -5.08097470e-01 2.33177543e-01
2.74807662e-01 8.88685882e-02 -4.32691127e-01 8.31373215e-01
9.42452252e-01 1.62879284e-02 6.94593072e-01 -1.02830005e+00
1.77994817e-01 9.05957937e-01 -2.48706609e-01 5.30351281e-01
4.52603370e-01 -3.83113138e-02 8.67157400e-01 -1.19364452e+00
3.97969261e-02 1.28119683e+00 3.47201526e-01 5.55177152e-01
-1.46415091e+00 -7.77256489e-01 7.10705370e-02 1.39922025e-02
-1.66461694e+00 -1.38562512e+00 7.91043758e-01 -3.83917570e-01
1.22733414e+00 5.40032804e-01 1.63904279e-01 1.26375210e+00
-4.47300911e-01 8.65943670e-01 6.34856820e-01 -7.25582600e-01
3.34086031e-01 1.62208706e-01 2.48530060e-01 1.91478878e-01
-1.27697468e-01 -8.68798641e-04 -1.18825924e+00 -1.67344406e-01
3.66575599e-01 -7.14594126e-01 -6.23196661e-01 1.11594744e-01
-1.24773073e+00 5.48052073e-01 -2.32244045e-01 5.76257586e-01
-3.28731507e-01 1.17006063e-01 2.40489706e-01 4.22221929e-01
6.82716906e-01 2.60075688e-01 -3.19433063e-01 -2.29513377e-01
-1.33250535e+00 3.83061245e-02 9.07832801e-01 6.45239234e-01
3.36653262e-01 6.17783964e-01 -4.66662884e-01 8.95690382e-01
5.47228038e-01 5.85040748e-01 4.85333949e-01 -5.68973720e-01
4.40615416e-01 5.96237602e-03 6.29642755e-02 -3.31168950e-01
-3.24361295e-01 -5.84223390e-01 -3.90074492e-01 -2.42259204e-02
4.67153311e-01 -4.22239751e-01 -1.08403289e+00 1.87107563e+00
3.48451167e-01 4.74034250e-01 3.10474366e-01 6.88556015e-01
8.24685276e-01 8.45200777e-01 -2.03077838e-01 -5.66723883e-01
1.33859599e+00 -7.64925182e-01 -1.19083750e+00 -3.23772609e-01
2.13039383e-01 -7.78772235e-01 6.46065533e-01 5.91663122e-01
-1.26171923e+00 -4.81637388e-01 -1.29440784e+00 2.97031373e-01
-6.54651523e-02 1.40128210e-01 -2.53885239e-03 1.30797637e+00
-1.11287344e+00 2.21609131e-01 -8.44090581e-01 -6.00171648e-03
-9.18550044e-02 4.68899935e-01 1.80632472e-02 4.08675462e-01
-1.16972458e+00 6.88399851e-01 -1.91399515e-01 1.91880353e-02
-1.22162855e+00 -6.11298919e-01 -1.01271915e+00 2.31589094e-01
2.96659201e-01 6.04674742e-02 1.62684715e+00 -8.94972026e-01
-1.89962053e+00 6.21590495e-01 -7.40363896e-01 -3.83387178e-01
2.27591977e-01 -6.30288795e-02 -9.02115703e-01 1.31815067e-02
-2.66399801e-01 2.47800350e-01 1.23006916e+00 -1.13557327e+00
-4.80675936e-01 -2.12012008e-01 -6.38123572e-01 2.86926359e-01
-4.89572883e-01 3.68417203e-01 -4.88949776e-01 -6.88512564e-01
2.08854496e-01 -4.97093081e-01 3.74111980e-01 -4.53386337e-01
-6.17592633e-01 -3.06849837e-01 5.79157472e-01 -9.75887954e-01
1.43640459e+00 -2.39799428e+00 2.84743845e-01 5.03077924e-01
-4.33965353e-04 2.84027219e-01 -3.59620243e-01 1.87309623e-01
-3.04145873e-01 -8.51765946e-02 -2.08956972e-01 -9.00271833e-01
1.40712455e-01 -3.08574200e-01 -9.65756848e-02 5.20941257e-01
4.23732579e-01 4.14940685e-01 -5.55545151e-01 -2.33276457e-01
2.21024200e-01 7.66952693e-01 -3.42419982e-01 4.16758448e-01
2.38227591e-01 2.87382245e-01 1.35439262e-01 5.19585490e-01
4.55815017e-01 5.18436790e-01 2.57783502e-01 -6.94859028e-02
-3.02125573e-01 9.09304142e-01 -1.62867022e+00 1.57086468e+00
-6.02306247e-01 9.92592394e-01 6.64526165e-01 -7.04751432e-01
7.27747023e-01 9.13619220e-01 3.49529356e-01 -3.62223685e-01
2.49812186e-01 3.10920268e-01 2.16382444e-01 -1.52203172e-01
3.15339446e-01 -1.99416894e-02 -1.04328200e-01 3.55600476e-01
6.03920579e-01 6.58317879e-02 -4.58226241e-02 -5.57443462e-02
1.11447835e+00 -3.75464797e-01 1.50353864e-01 -3.51860583e-01
6.98901892e-01 -5.83047330e-01 2.89903969e-01 5.62154651e-01
-2.66485870e-01 6.17526472e-01 3.21765333e-01 5.10604024e-01
-4.97066975e-01 -1.27684712e+00 -1.55388534e-01 1.23095679e+00
-3.05108994e-01 -3.64829332e-01 -1.07209158e+00 -3.99799138e-01
-2.20037699e-01 9.22726393e-01 -3.35613787e-01 -1.75899118e-01
-6.35153890e-01 -6.20234609e-01 9.21971262e-01 4.21589971e-01
-6.29708990e-02 -6.81981564e-01 -8.44156817e-02 3.35022748e-01
-4.39766437e-01 -1.13278425e+00 -8.60392869e-01 8.14549804e-01
-3.00278366e-01 -5.55094540e-01 -6.69874549e-01 -6.35986865e-01
1.67434707e-01 1.94193125e-01 7.72041380e-01 -4.83545363e-01
-6.83522373e-02 4.34088558e-01 -1.23921029e-01 -6.94531083e-01
-7.68268943e-01 4.80109006e-02 3.43200326e-01 5.91708243e-01
5.87691128e-01 -5.89329243e-01 -1.73077956e-01 2.91995972e-01
-5.64009905e-01 -4.61445719e-01 4.52433139e-01 3.39154959e-01
1.86401159e-01 1.53846061e-02 7.91533113e-01 -2.39883527e-01
7.50811636e-01 -3.08626026e-01 -4.54322129e-01 -9.77729186e-02
-3.46407056e-01 8.44863057e-02 1.54810160e-01 -6.09735727e-01
-1.11682332e+00 1.17118075e-01 -5.33415735e-01 -2.72698015e-01
-3.91560942e-01 2.51065344e-01 -7.24744976e-01 2.27039680e-01
7.88435459e-01 3.23511153e-01 -1.99494302e-01 -5.81551373e-01
1.00300230e-01 1.37158048e+00 6.28239870e-01 1.14574954e-01
7.42813647e-01 2.26420555e-02 -6.72496259e-01 -1.45050919e+00
-4.03580695e-01 -8.84548426e-01 -3.26271594e-01 -1.69874549e-01
8.25663507e-01 -1.20097601e+00 -4.02045250e-01 5.94839990e-01
-1.22554481e+00 -2.24255383e-01 -2.65988976e-01 6.06135607e-01
-2.36166835e-01 9.72749218e-02 -3.08510303e-01 -1.38559484e+00
-2.49554962e-01 -9.49945152e-01 1.18331397e+00 -1.17735639e-01
-5.64959049e-01 -6.43277287e-01 9.23570991e-02 2.46068209e-01
3.97438645e-01 -1.93733171e-01 4.84284401e-01 -1.17163599e+00
-9.89298970e-02 -1.59465075e-01 2.48271048e-01 5.72092831e-01
2.47198507e-01 -3.23036492e-01 -1.89069569e+00 -1.56483889e-01
3.11920434e-01 2.28887707e-01 8.39822292e-01 6.22336507e-01
7.89729238e-01 -2.29595616e-01 -1.71044737e-01 3.10647994e-01
8.96420538e-01 1.59120902e-01 4.98620182e-01 -4.10824627e-01
4.74065632e-01 5.15931726e-01 5.04973121e-02 4.66531575e-01
-4.92845755e-03 8.88449848e-01 1.47124574e-01 -6.22257963e-02
-5.44668794e-01 5.45633659e-02 9.07554984e-01 1.09139252e+00
2.71394014e-01 -7.27583051e-01 -8.13189983e-01 8.92917097e-01
-1.37955403e+00 -9.32956457e-01 -2.32259408e-01 2.38358927e+00
9.70130384e-01 1.58017620e-01 4.41328228e-01 8.48076105e-01
8.33694994e-01 3.16423923e-01 -3.24171036e-01 -4.89019811e-01
5.60790710e-02 4.51232493e-01 4.75438744e-01 8.42207313e-01
-1.04861522e+00 4.71986920e-01 6.86053133e+00 8.37933719e-01
-1.03079653e+00 4.16362554e-01 1.10109322e-01 -5.61875582e-01
-2.57005841e-01 -6.94651842e-01 -9.03407335e-01 2.74806678e-01
1.74344218e+00 8.44255090e-02 5.06789625e-01 5.11487067e-01
3.89574498e-01 -6.43832833e-02 -1.43283916e+00 1.23634732e+00
4.03784424e-01 -8.16250920e-01 -4.21267658e-01 1.14261605e-01
2.31609523e-01 -4.07267213e-02 -1.59497336e-01 2.16155872e-01
9.73519534e-02 -9.45313513e-01 9.61110890e-01 8.16129297e-02
9.06536460e-01 -7.75286019e-01 5.09827793e-01 3.61542702e-01
-1.36302137e+00 -4.13737148e-02 3.21331978e-01 2.94903606e-01
2.56709844e-01 6.70297086e-01 -1.06511199e+00 3.11153144e-01
3.40414137e-01 3.15879397e-02 -5.19611388e-02 7.81490803e-01
-2.78613687e-01 1.00568271e+00 -4.22581106e-01 1.00837447e-01
-1.20375626e-01 5.13011873e-01 8.97919536e-01 1.60209882e+00
3.10123354e-01 -2.12298900e-01 -5.79961054e-02 6.95109248e-01
-1.77405715e-01 -1.08471535e-01 -1.68660596e-01 -6.41363636e-02
8.67653668e-01 1.10051346e+00 -4.34701979e-01 -2.02030271e-01
-2.89002925e-01 1.08942187e+00 -2.10560918e-01 5.22061288e-01
-8.23606312e-01 -6.48979843e-01 1.06227732e+00 -2.53659599e-02
4.05705780e-01 -2.64933556e-01 9.99759790e-03 -8.91099989e-01
-1.13783956e-01 -9.11772549e-01 1.11409053e-02 -2.75313079e-01
-7.59619713e-01 5.46971202e-01 -1.11491643e-01 -8.96099806e-01
-4.10512805e-01 -4.49257672e-01 -7.37926185e-01 1.22059178e+00
-1.70867956e+00 -6.19130313e-01 -1.78233311e-01 6.73203886e-01
9.72335637e-01 -8.00194666e-02 1.01067567e+00 5.70348740e-01
-6.92995548e-01 9.64985371e-01 -7.56986588e-02 1.87967628e-01
6.73629582e-01 -1.31421626e+00 5.73856890e-01 1.14188635e+00
3.14866662e-01 2.05136612e-01 8.56669664e-01 -1.46232486e-01
-1.29408491e+00 -1.15222025e+00 1.28596151e+00 -5.68071365e-01
2.91866153e-01 -9.31480348e-01 -7.30825365e-01 3.84834886e-01
3.51717800e-01 -3.64401788e-01 8.14960778e-01 1.21564545e-01
-8.14105794e-02 -6.24586008e-02 -8.79786313e-01 3.31194341e-01
8.29740226e-01 -7.96409190e-01 -7.98013091e-01 1.49610266e-02
6.37229145e-01 -2.38017365e-01 -5.04861653e-01 -3.85228395e-02
2.98178166e-01 -6.45961225e-01 9.52932954e-01 -1.05717979e-01
-3.54272723e-01 -2.77905911e-01 -3.11031461e-01 -1.45127523e+00
-1.16455004e-01 -1.19492328e+00 -2.37671256e-01 1.61547041e+00
6.74406409e-01 -4.49570507e-01 2.95745969e-01 2.90047497e-01
-3.86219233e-01 7.64421150e-02 -1.32431614e+00 -9.35088933e-01
-3.53566706e-01 -8.66104186e-01 4.69288230e-01 6.01819217e-01
2.40596443e-01 6.25192642e-01 -3.77935082e-01 5.35408139e-01
4.71289545e-01 -5.91187060e-01 4.70685005e-01 -1.21303296e+00
-3.73177618e-01 -4.60899055e-01 -7.07273260e-02 -1.00891840e+00
9.81043875e-02 -6.88643873e-01 7.41104066e-01 -1.22807157e+00
-4.94303077e-01 8.07323605e-02 -4.63952392e-01 2.69371539e-01
3.73967108e-03 1.64988935e-01 -1.56852320e-01 -5.86201958e-02
-1.65040717e-01 3.80948782e-01 4.13735896e-01 -9.33457911e-02
-6.86877847e-01 8.65103826e-02 -5.45527935e-01 6.33389413e-01
5.78045130e-01 -6.07713580e-01 -2.70951599e-01 -2.27111861e-01
-3.97176504e-01 2.04818577e-01 1.52492478e-01 -1.20047617e+00
1.85491666e-01 2.43257418e-01 1.46288708e-01 -4.62149531e-01
7.12474763e-01 -5.06274104e-01 -8.50585923e-02 2.63616323e-01
-5.46253800e-01 -4.88139898e-01 2.42097184e-01 5.29990494e-01
-3.06960016e-01 -1.22107364e-01 7.33838558e-01 1.94953337e-01
-1.68349981e-01 -1.78785190e-01 -1.00954378e+00 -1.33279935e-01
6.06535375e-01 -2.67703891e-01 1.21488087e-01 -5.02249002e-01
-8.04133534e-01 -1.09342508e-01 -3.08564872e-01 4.16090280e-01
4.02752906e-01 -1.29720688e+00 -9.58679616e-01 5.39792001e-01
3.44171785e-02 -2.41304934e-01 7.22809285e-02 9.76791441e-01
2.81105697e-01 4.77213264e-01 5.63299775e-01 -6.30794764e-01
-1.54536176e+00 1.58424273e-01 6.17615342e-01 1.02336006e-02
-1.43445238e-01 1.16783404e+00 1.57923982e-01 -1.00725636e-01
8.34005952e-01 -4.17594135e-01 -2.41378888e-01 4.45947528e-01
9.07717049e-01 4.09248084e-01 5.74786842e-01 -7.97083139e-01
-6.76129520e-01 1.22630559e-01 4.00476724e-01 -7.94549346e-01
1.13550580e+00 -2.84682721e-01 3.96679312e-01 6.99684501e-01
1.09827280e+00 5.10122538e-01 -9.81785655e-01 -3.08753043e-01
5.19667156e-02 -1.13388754e-01 6.05618358e-01 -8.54131043e-01
-6.01833224e-01 1.03579164e+00 7.57137001e-01 4.08831865e-01
1.07453978e+00 -1.03390310e-02 5.03677905e-01 9.54059437e-02
-2.78091639e-01 -1.16933036e+00 1.24848358e-01 4.27144825e-01
8.37682128e-01 -8.69423270e-01 -5.14131486e-01 -4.39208388e-01
-2.96423525e-01 9.56796706e-01 1.85104117e-01 2.05580682e-01
6.56210244e-01 6.06433928e-01 2.46897966e-01 9.31390822e-02
-6.25622511e-01 -3.68631542e-01 5.93636453e-01 5.00472784e-01
6.08640552e-01 -7.07854182e-02 9.28998664e-02 7.95963824e-01
-3.22125778e-02 -5.63833773e-01 4.04238373e-01 6.59308195e-01
-6.02485597e-01 -1.01122665e+00 -7.62984872e-01 1.19239077e-01
-4.59283888e-01 -2.85009891e-01 -6.47241592e-01 3.56397599e-01
-2.52669334e-01 1.56917465e+00 2.08541766e-01 -4.88981098e-01
6.05521441e-01 8.27945232e-01 2.23452404e-01 -7.73507297e-01
-8.25864494e-01 8.63888860e-01 4.86092508e-01 -3.44143480e-01
-4.29900557e-01 -8.88604164e-01 -1.03174376e+00 1.99446850e-03
-7.12597132e-01 1.82763651e-01 1.08083391e+00 9.60307240e-01
3.15885007e-01 1.20541239e+00 8.23205411e-01 -1.04281116e+00
-5.58209002e-01 -1.31803548e+00 -6.25887990e-01 -1.31258190e-01
7.98902333e-01 -2.86746770e-01 -7.90845871e-01 2.15731356e-02] | [14.703490257263184, 6.096660614013672] |
fb4aada6-ca87-44e4-a35f-b83726bf7d95 | fedvmr-a-new-federated-learning-method-for | 2210.15977 | null | https://arxiv.org/abs/2210.15977v1 | https://arxiv.org/pdf/2210.15977v1.pdf | FedVMR: A New Federated Learning method for Video Moment Retrieval | Despite the great success achieved, existing video moment retrieval (VMR) methods are developed under the assumption that data are centralizedly stored. However, in real-world applications, due to the inherent nature of data generation and privacy concerns, data are often distributed on different silos, bringing huge challenges to effective large-scale training. In this work, we try to overcome above limitation by leveraging the recent success of federated learning. As the first that is explored in VMR field, the new task is defined as video moment retrieval with distributed data. Then, a novel federated learning method named FedVMR is proposed to facilitate large-scale and secure training of VMR models in decentralized environment. Experiments on benchmark datasets demonstrate its effectiveness. This work is the very first attempt to enable safe and efficient VMR training in decentralized scene, which is hoped to pave the way for further study in the related research field. | ['Xin-Shun Xu', 'Meng Liu', 'Peng-Fei Zhang', 'Zhen-Duo Chen', 'Xin Luo', 'Yan Wang'] | 2022-10-28 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [-1.41081676e-01 -5.06082714e-01 -4.77281749e-01 -2.38316640e-01
-6.32140636e-01 -5.32515466e-01 7.44631827e-01 5.25481738e-02
-3.20162654e-01 7.40517855e-01 -1.28611242e-02 -3.77684414e-01
-3.03058326e-01 -6.69203401e-01 -4.69709426e-01 -8.32315803e-01
-2.56943077e-01 1.31646529e-01 3.29904035e-02 -8.67686868e-02
3.00649017e-01 5.20088255e-01 -1.52364242e+00 4.76907752e-02
4.94083256e-01 9.00535405e-01 -7.19529092e-02 3.27446997e-01
-2.19361231e-01 1.00440526e+00 -4.54348594e-01 -7.80420065e-01
6.36071861e-01 -1.21929944e-01 -7.52629220e-01 -1.46994814e-01
2.65271366e-01 -5.37147462e-01 -6.76320791e-01 9.26889181e-01
6.06238842e-01 2.87287682e-01 2.69134223e-01 -1.48643970e+00
-4.94229108e-01 4.06523108e-01 -5.44295669e-01 1.82346463e-01
3.01362246e-01 -8.33470821e-02 9.34703171e-01 -4.75958437e-01
8.11277032e-01 7.70664811e-01 1.22987889e-01 6.21029675e-01
-4.94299859e-01 -7.92374611e-01 1.67328253e-01 3.30113530e-01
-1.52565694e+00 -4.82087106e-01 9.09121752e-01 -4.30046022e-02
6.78855360e-01 4.89470780e-01 4.18420732e-01 7.18782663e-01
1.18191428e-01 1.05147195e+00 1.12551200e+00 -2.45746300e-01
3.47219050e-01 3.35139573e-01 -2.39077300e-01 5.36706626e-01
2.24992961e-01 7.79908756e-03 -6.30312324e-01 -4.41191822e-01
3.23030621e-01 4.60511237e-01 -2.90095896e-01 -6.02271438e-01
-1.07787323e+00 7.01222658e-01 2.06568405e-01 4.78775680e-01
-2.48188481e-01 -2.25631386e-01 7.69470513e-01 6.43303752e-01
4.62715924e-01 -5.70882484e-02 -3.63814831e-01 -1.63949445e-01
-1.20843971e+00 1.87565610e-01 7.96497583e-01 6.82266712e-01
7.11052418e-01 7.40604699e-02 2.45866373e-01 5.93299866e-01
4.16036427e-01 4.80426162e-01 8.37278783e-01 -6.68702066e-01
4.98610914e-01 4.74424303e-01 -4.66214865e-02 -1.27563286e+00
2.61662155e-01 -1.11872792e-01 -6.98957384e-01 -9.64976847e-02
-1.48804151e-02 -5.70618324e-02 -5.55405200e-01 1.25246191e+00
8.33905160e-01 4.72016692e-01 3.91449958e-01 9.64859605e-01
6.15266263e-01 6.02035999e-01 -1.53011661e-02 -3.70205313e-01
9.56210196e-01 -7.86913216e-01 -6.40053630e-01 4.59484637e-01
7.46516109e-01 -9.04510796e-01 4.16892141e-01 3.92488331e-01
-8.22920024e-01 -6.30420372e-02 -9.60137188e-01 2.35338256e-01
-4.73126262e-01 -2.45541930e-01 1.09161842e+00 6.50020719e-01
-1.03475988e+00 3.13857585e-01 -7.27002501e-01 -5.65290630e-01
6.51880920e-01 4.99035835e-01 -8.06313634e-01 -2.93705106e-01
-1.06758702e+00 3.95774215e-01 3.39010298e-01 9.48214531e-02
-1.17907691e+00 -2.89970487e-01 -4.38525617e-01 -1.74278602e-01
5.97814202e-01 -4.40657914e-01 1.09939790e+00 -8.43988836e-01
-1.34073877e+00 9.05981541e-01 2.28865594e-01 -7.18468249e-01
6.53814614e-01 5.82740270e-03 -4.58310902e-01 2.18260601e-01
-1.85201079e-01 1.25964329e-01 1.12438047e+00 -1.13471782e+00
-7.61569560e-01 -5.56898654e-01 -9.61974077e-03 2.44332075e-01
-8.64282072e-01 2.20746577e-01 -3.64409924e-01 -3.74481410e-01
-1.16001405e-01 -7.83997536e-01 -2.48066008e-01 -3.51134181e-01
2.99480483e-02 -3.57907861e-01 1.32383847e+00 -3.30719024e-01
1.22200668e+00 -2.24943185e+00 -1.23783559e-01 2.28666782e-01
4.22255218e-01 7.42785990e-01 -4.91010994e-02 8.39572608e-01
7.39222839e-02 1.79980576e-01 2.81399995e-01 -2.66756564e-01
4.25912300e-03 1.06850386e-01 -7.38066316e-01 7.65811026e-01
-3.22935164e-01 6.86567247e-01 -6.96996331e-01 -1.00326741e+00
3.65883321e-01 4.31473553e-01 -2.76496291e-01 2.49140605e-01
3.20669450e-02 2.58385628e-01 -1.03417230e+00 1.00125182e+00
7.83638358e-01 -8.61011371e-02 2.85682410e-01 -7.56979808e-02
-2.60191768e-01 -3.59470725e-01 -1.17900360e+00 1.70702016e+00
-1.85963109e-01 3.27985972e-01 3.63181829e-01 -1.26844990e+00
8.49627972e-01 7.43236184e-01 1.06056952e+00 -5.56435347e-01
3.30868572e-01 3.33603323e-01 -4.67638463e-01 -5.68918407e-01
7.67850995e-01 1.19922064e-01 1.59881160e-01 6.72311902e-01
-8.87269601e-02 2.92095125e-01 2.17077024e-02 6.28396034e-01
1.10797524e+00 9.66033116e-02 1.66026577e-01 2.04569176e-01
8.03258717e-01 -6.77713677e-02 5.16971707e-01 5.37678540e-01
-5.89110494e-01 1.09652191e-01 -3.61511426e-04 -7.62049377e-01
-8.84751618e-01 -8.03205848e-01 -3.00364476e-02 9.88978446e-01
1.97252899e-01 -7.33042419e-01 -4.42648947e-01 -8.29070091e-01
6.99281618e-02 -3.11179031e-02 -1.44284025e-01 7.57511482e-02
-5.73143899e-01 -4.60353822e-01 5.56607008e-01 -4.01771115e-03
6.51871860e-01 -9.52829957e-01 -6.84246719e-01 6.69530034e-02
-1.56930819e-01 -9.94782150e-01 -1.97656319e-01 -3.13665777e-01
-9.83589888e-01 -1.26255894e+00 -7.82424033e-01 -7.44058728e-01
6.65841758e-01 8.92421961e-01 8.89191568e-01 5.54307759e-01
-5.47030389e-01 7.57449448e-01 -7.29361475e-01 -4.07382488e-01
-4.53092009e-02 5.64429946e-02 2.21552864e-01 3.37843418e-01
5.17105579e-01 -4.67163831e-01 -7.69598305e-01 3.45223576e-01
-1.42447734e+00 -3.32122445e-01 4.64757025e-01 7.75234342e-01
4.80887353e-01 4.10845190e-01 5.44038177e-01 -7.82322526e-01
5.09048045e-01 -7.02736259e-01 -7.76449859e-01 6.72913671e-01
-7.40810633e-01 -2.68184811e-01 6.95469022e-01 -3.08169574e-01
-1.09428227e+00 -8.37748696e-04 1.60374939e-01 -6.61560237e-01
-8.86700451e-02 4.73434091e-01 4.08297554e-02 -4.01968569e-01
1.72783017e-01 5.96455276e-01 9.58710760e-02 -3.69071037e-01
2.23843575e-01 1.07386994e+00 1.28518134e-01 -7.31752098e-01
1.10067928e+00 7.76971996e-01 6.68008700e-02 -7.27179766e-01
-2.51375258e-01 -7.32985854e-01 -1.19729549e-01 -2.74171501e-01
5.31662643e-01 -1.07855570e+00 -8.62943113e-01 2.86823988e-01
-6.88175976e-01 1.19600028e-01 -9.99199226e-03 6.08062983e-01
-3.83392185e-01 6.18227184e-01 -3.27268481e-01 -1.04408836e+00
-8.68620217e-01 -8.42513740e-01 7.85974860e-01 1.20574862e-01
4.21480238e-01 -8.58920813e-01 3.48225474e-01 7.12705016e-01
6.54862523e-01 -7.40108406e-03 4.35653836e-01 -8.65212202e-01
-1.11227691e+00 -6.60678983e-01 3.10926000e-03 2.70842761e-01
1.46864071e-01 2.18618020e-01 -7.50795722e-01 -7.64916837e-01
1.74821779e-01 -6.75218463e-01 4.48450506e-01 -2.10892007e-01
9.50982869e-01 -3.84905934e-01 -3.23831886e-01 6.02854013e-01
1.81198597e+00 -3.64025868e-02 5.03595948e-01 4.01976317e-01
4.26779270e-01 3.86301041e-01 9.06931639e-01 7.73489892e-01
3.43790025e-01 2.99581170e-01 5.85460424e-01 2.31527120e-01
3.26433897e-01 -3.71288836e-01 2.22267553e-01 9.73785162e-01
-1.58413071e-02 -1.71255916e-01 -7.04128325e-01 5.49971819e-01
-1.97274590e+00 -1.11197877e+00 4.37707335e-01 2.28979349e+00
3.85060042e-01 -4.55237716e-01 -1.01563849e-01 4.49269749e-02
5.45723498e-01 6.29439950e-01 -3.61453325e-01 -3.06302726e-01
-8.00018087e-02 -4.37018126e-02 4.68788296e-01 -1.77687094e-01
-1.09583449e+00 1.03780425e+00 5.24337816e+00 9.62788284e-01
-1.41946781e+00 2.19105542e-01 4.27040637e-01 -1.62092090e-01
-6.45892024e-02 2.37022385e-01 -6.50557637e-01 4.12981570e-01
7.91442633e-01 -5.30898213e-01 7.29571819e-01 9.56332684e-01
-5.92985824e-02 1.08381055e-01 -7.35216081e-01 1.27038348e+00
1.20788023e-01 -1.24511755e+00 4.02529649e-02 2.57995188e-01
8.75560522e-01 2.95729458e-01 2.36430392e-01 3.06956649e-01
1.08897671e-01 -7.00633645e-01 1.33437723e-01 2.93192625e-01
5.37399590e-01 -8.11790824e-01 6.64245009e-01 4.73063558e-01
-1.22880340e+00 -3.34211349e-01 -6.33296132e-01 1.31652817e-01
-1.07491538e-01 2.53734976e-01 -7.28306234e-01 8.62421691e-01
7.29751825e-01 5.71368098e-01 -4.28965569e-01 1.17445219e+00
2.77872175e-01 6.02790534e-01 -2.73028076e-01 3.75266597e-02
2.67934173e-01 -1.59243807e-01 4.30569082e-01 7.64224470e-01
1.86543599e-01 6.06842041e-02 2.50718862e-01 3.97534184e-02
-3.30021888e-01 7.96244681e-01 -1.19538867e+00 -3.16584736e-01
4.74305660e-01 1.61068726e+00 -5.54099143e-01 -2.36560062e-01
-5.04418373e-01 7.75865197e-01 3.26401770e-01 3.07997912e-01
-7.22325861e-01 -2.14734823e-01 4.85739112e-01 -5.25735393e-02
3.94816935e-01 -2.23114058e-01 4.79477257e-01 -1.56615031e+00
7.34815970e-02 -1.29424167e+00 6.93007708e-01 -2.20316678e-01
-1.26497650e+00 5.69636762e-01 -1.82010055e-01 -1.42723048e+00
-2.42388666e-01 -2.96320856e-01 -6.02883220e-01 3.19146842e-01
-1.78595066e+00 -1.39203548e+00 -1.27913365e-02 1.22471213e+00
3.02851528e-01 -6.07255995e-01 8.84741485e-01 5.88387251e-01
-4.71690118e-01 6.41757905e-01 5.92977941e-01 2.36615330e-01
9.06846464e-01 -5.79055309e-01 -2.39653170e-01 7.31409073e-01
4.99532700e-01 8.82062614e-01 4.16415781e-01 -6.27712667e-01
-2.17114687e+00 -1.02604830e+00 7.18269050e-01 -6.30319491e-02
7.32934892e-01 -1.81536734e-01 -5.80242276e-01 6.45275772e-01
2.78976649e-01 4.00360346e-01 9.41238999e-01 -8.01174194e-02
-3.71633857e-01 -6.60129488e-01 -1.31055832e+00 3.68320644e-01
5.42966902e-01 -6.68431520e-01 -2.91319132e-01 6.65136099e-01
5.73805928e-01 -1.38637528e-01 -9.99638021e-01 3.35892260e-01
4.07581240e-01 -8.98022652e-01 7.85889983e-01 -7.55966127e-01
-1.01296075e-01 -2.63009220e-01 -3.62975746e-01 -6.81636631e-01
2.89380640e-01 -9.82146502e-01 -3.55456144e-01 1.45485294e+00
-3.25465471e-01 -7.17071712e-01 1.15498078e+00 7.33960509e-01
4.53308970e-01 -6.32593572e-01 -1.17389834e+00 -7.18764961e-01
-1.85479432e-01 -1.77665040e-01 8.50804925e-01 1.08758605e+00
-1.66294038e-01 -1.96709245e-01 -7.21451700e-01 -3.24429087e-02
8.13674748e-01 3.12531978e-01 1.06796324e+00 -9.60989714e-01
-1.58530012e-01 7.86874220e-02 -7.19670773e-01 -6.05347335e-01
2.40289420e-01 -8.12897205e-01 -3.85010898e-01 -1.00968432e+00
2.39427745e-01 -6.13506138e-01 -6.73868835e-01 3.92968893e-01
2.48429775e-01 2.63595372e-01 5.20674169e-01 5.93260288e-01
-1.18339348e+00 5.20589471e-01 9.73118722e-01 -7.37811401e-02
8.17306191e-02 1.13786586e-01 -4.32250977e-01 3.52084637e-01
1.10298514e+00 -3.59344125e-01 -7.10245073e-01 -4.82153505e-01
8.01420659e-02 2.25565210e-01 3.26309443e-01 -9.53436553e-01
5.50463319e-01 -3.81860167e-01 -1.49731152e-02 -7.70082533e-01
1.79136187e-01 -1.28940248e+00 2.04618201e-01 1.05644308e-01
-6.44537583e-02 1.90295175e-01 -3.78875077e-01 6.91939592e-01
-5.05570233e-01 9.68595129e-03 3.24449897e-01 -4.15803403e-01
-6.99247003e-01 7.62358665e-01 1.54098824e-01 -1.02538392e-01
1.51002884e+00 -6.35118335e-02 -2.90629268e-01 -3.86561662e-01
-7.19945207e-02 3.91808569e-01 5.75021446e-01 3.68057609e-01
8.98205221e-01 -1.12351716e+00 -5.26004374e-01 1.81704551e-01
9.25046392e-03 -1.16920002e-01 4.99457091e-01 5.70749640e-01
-4.37268764e-01 6.92490876e-01 -1.10868126e-01 -3.53097588e-01
-1.61337411e+00 1.06448686e+00 -8.59826133e-02 -2.42411032e-01
-6.61230326e-01 3.74245793e-01 -3.66089880e-01 -2.35764697e-01
5.91590941e-01 7.15550721e-01 2.66552847e-02 -1.04217455e-01
6.41714633e-01 4.05273467e-01 9.35562029e-02 -7.51953125e-01
-3.86873841e-01 2.65720844e-01 -4.24664021e-01 7.16177449e-02
1.44102812e+00 -2.43610099e-01 -5.37275970e-01 1.20839126e-01
1.42301965e+00 1.55272052e-01 -9.62354302e-01 -2.50037909e-01
-1.50606588e-01 -9.28804755e-01 6.93118125e-02 -2.43276417e-01
-1.46087527e+00 5.79228342e-01 7.11697280e-01 7.37590268e-02
1.13015246e+00 -3.80761415e-01 1.14915442e+00 6.21048748e-01
9.20664907e-01 -1.18812466e+00 -9.49437320e-02 6.01677969e-02
1.39354050e-01 -1.39663386e+00 2.81838268e-01 2.19305605e-01
-5.17965198e-01 8.90274942e-01 6.75835907e-01 2.07017036e-03
6.96390867e-01 2.63473438e-03 2.87112415e-01 -1.92809999e-01
-9.05000746e-01 2.62064487e-01 -2.17586011e-01 4.46515411e-01
1.79584891e-01 -1.59264088e-01 -6.40526533e-01 1.01947831e-02
3.70280623e-01 3.36354136e-01 1.04748599e-01 1.50178456e+00
-3.32676679e-01 -1.54605985e+00 -4.21846598e-01 2.89822876e-01
-9.36955750e-01 1.46525368e-01 -1.55461907e-01 6.59120917e-01
-1.87308863e-01 9.47023332e-01 -3.82950306e-01 -3.92588407e-01
-3.11192989e-01 -4.99241613e-03 1.55699477e-01 -3.06771882e-02
-6.53663516e-01 -7.93786719e-02 -6.02712035e-01 -5.93993425e-01
-5.48148870e-01 -3.51766974e-01 -1.01375496e+00 -5.67713499e-01
-4.01632965e-01 7.44300067e-01 1.08314168e+00 5.96350431e-01
4.04817343e-01 -1.13429189e-01 1.25000083e+00 -3.97336245e-01
-1.13656354e+00 -2.55519956e-01 -6.87879920e-01 5.33372045e-01
1.91666260e-01 -1.98014483e-01 -3.04782242e-01 -2.93262005e-01] | [5.862910270690918, 6.3329596519470215] |
e632393c-9f50-4a06-a365-8923bde37731 | clustering-ensemble-meets-low-rank-tensor | 2012.08916 | null | https://arxiv.org/abs/2012.08916v1 | https://arxiv.org/pdf/2012.08916v1.pdf | Clustering Ensemble Meets Low-rank Tensor Approximation | This paper explores the problem of clustering ensemble, which aims to combine multiple base clusterings to produce better performance than that of the individual one. The existing clustering ensemble methods generally construct a co-association matrix, which indicates the pairwise similarity between samples, as the weighted linear combination of the connective matrices from different base clusterings, and the resulting co-association matrix is then adopted as the input of an off-the-shelf clustering algorithm, e.g., spectral clustering. However, the co-association matrix may be dominated by poor base clusterings, resulting in inferior performance. In this paper, we propose a novel low-rank tensor approximation-based method to solve the problem from a global perspective. Specifically, by inspecting whether two samples are clustered to an identical cluster under different base clusterings, we derive a coherent-link matrix, which contains limited but highly reliable relationships between samples. We then stack the coherent-link matrix and the co-association matrix to form a three-dimensional tensor, the low-rankness property of which is further explored to propagate the information of the coherent-link matrix to the co-association matrix, producing a refined co-association matrix. We formulate the proposed method as a convex constrained optimization problem and solve it efficiently. Experimental results over 7 benchmark data sets show that the proposed model achieves a breakthrough in clustering performance, compared with 12 state-of-the-art methods. To the best of our knowledge, this is the first work to explore the potential of low-rank tensor on clustering ensemble, which is fundamentally different from previous approaches. | ['Qingfu Zhang', 'Junhui Hou', 'Hui Liu', 'Yuheng Jia'] | 2020-12-16 | null | null | null | null | ['clustering-ensemble'] | ['graphs'] | [-1.42611772e-01 -5.24680376e-01 3.45012136e-02 -1.83195323e-01
-5.62434494e-01 -5.49993753e-01 1.73951462e-01 -5.20504937e-02
-8.91461894e-02 2.79367417e-01 1.62771925e-01 8.22633803e-02
-6.50270760e-01 -4.36006844e-01 -1.97106466e-01 -1.25893652e+00
-2.56481707e-01 5.14478803e-01 -1.75827593e-02 -4.81847078e-02
2.97752649e-01 7.92406797e-02 -1.38077486e+00 3.96457821e-01
1.27190065e+00 8.91678452e-01 6.90790042e-02 1.93547457e-01
1.39253987e-02 3.52390528e-01 -3.22035283e-01 -1.27263784e-01
1.45572290e-01 -3.74465525e-01 -7.77768135e-01 4.84329700e-01
1.02689847e-01 2.72771150e-01 -3.50260437e-02 1.21365714e+00
2.44420886e-01 2.27784261e-01 4.91068780e-01 -1.43457448e+00
-5.54845512e-01 7.61808097e-01 -8.21486115e-01 -5.19976579e-03
3.56694967e-01 -9.25046355e-02 1.10171115e+00 -1.10017645e+00
3.41217309e-01 1.22572374e+00 4.87985402e-01 -1.30798677e-02
-1.49230003e+00 -6.64076090e-01 4.64208990e-01 3.15173060e-01
-1.87989128e+00 -1.63971856e-01 9.34452534e-01 -6.09872162e-01
3.80988330e-01 4.86159623e-01 6.72967970e-01 6.17830813e-01
-1.66083500e-01 6.70502782e-01 1.43091798e+00 -2.17981241e-03
2.85593957e-01 -1.82904974e-01 3.76475185e-01 5.75651765e-01
3.55571508e-01 -3.06015372e-01 -4.26359028e-01 -3.68674815e-01
1.77301377e-01 2.72517890e-01 -6.11442626e-01 -4.32734847e-01
-1.57588196e+00 5.12530744e-01 6.87879384e-01 6.43064082e-01
-4.70000893e-01 -2.97335654e-01 6.70098364e-02 -1.95230208e-02
4.47830677e-01 1.76629737e-01 -3.32287289e-02 2.60178804e-01
-1.04645813e+00 8.67842585e-02 7.46536493e-01 6.83516800e-01
9.84426677e-01 -3.03521663e-01 6.44253939e-02 8.44934762e-01
6.60618246e-01 3.57080102e-01 3.70926917e-01 -8.81319642e-01
5.50697088e-01 9.24699664e-01 -8.02597627e-02 -1.59838307e+00
-2.86651373e-01 -6.73107684e-01 -1.34440887e+00 -3.63664925e-01
2.72530764e-01 2.80266106e-02 -3.54257077e-01 1.52592385e+00
6.00317001e-01 7.60134697e-01 -3.38882133e-02 1.16132867e+00
5.48389792e-01 5.76208651e-01 -4.20206815e-01 -7.64799953e-01
1.07329905e+00 -8.36791992e-01 -7.64866829e-01 3.35750580e-01
4.82185066e-01 -9.49911833e-01 4.83082265e-01 6.11856282e-01
-8.38965416e-01 -5.00426471e-01 -1.12062073e+00 4.28381830e-01
-1.04934692e-01 1.26364142e-01 4.34629917e-01 3.94824743e-01
-8.40598166e-01 5.22410214e-01 -7.81106532e-01 -2.60538131e-01
-4.90838382e-03 3.63381326e-01 -3.45345050e-01 -3.65898252e-01
-8.05769861e-01 4.38118428e-01 5.84687471e-01 4.91003573e-01
-4.90249544e-01 -5.08074939e-01 -3.18516642e-01 -3.44363451e-01
7.15643287e-01 -5.97397387e-01 3.74007761e-01 -5.50897896e-01
-1.11556792e+00 3.11057717e-01 -4.50004548e-01 2.66867340e-01
1.06407188e-01 -1.25742316e-01 -7.05678642e-01 -2.80534867e-02
2.74120420e-01 1.14276923e-01 6.82004750e-01 -1.92897379e+00
-5.97042263e-01 -4.93493825e-01 -2.56147414e-01 2.83379376e-01
-5.08054316e-01 -1.89277917e-01 -8.98523808e-01 -6.59870207e-01
8.56581092e-01 -1.14781415e+00 -4.54535276e-01 -5.77792287e-01
-7.44810700e-01 -3.45487714e-01 1.05352879e+00 -5.21005392e-01
1.80880916e+00 -2.16412902e+00 9.30712402e-01 7.22461462e-01
5.51317275e-01 2.22198069e-02 -4.63780388e-02 4.82406944e-01
-2.41114721e-01 2.39799663e-01 -5.86407065e-01 -2.73616910e-01
-1.64966643e-01 3.42637151e-01 -6.96892887e-02 5.64987302e-01
-5.69772087e-02 4.00837809e-01 -1.35249329e+00 -7.35044003e-01
1.17760435e-01 3.94090503e-01 -2.93514758e-01 2.91260481e-01
3.04276556e-01 5.91781497e-01 -5.08102059e-01 6.13467038e-01
9.47763741e-01 -2.83002168e-01 6.16327167e-01 -9.36431587e-01
-1.43266201e-01 -2.91169018e-01 -1.78086579e+00 1.60251403e+00
1.26745000e-01 -5.70562668e-02 3.06188941e-01 -1.30221140e+00
9.02556896e-01 2.95944631e-01 1.03306293e+00 1.07768243e-02
3.70552987e-02 3.81198138e-01 2.87320316e-01 -4.86436993e-01
4.01851833e-01 -2.24503521e-02 3.08653079e-02 4.67617035e-01
-2.24302962e-01 2.70180166e-01 6.35338783e-01 4.44841355e-01
7.85948336e-01 -4.85840216e-02 -1.81866400e-02 -4.04313862e-01
9.02590454e-01 -1.09348521e-01 7.98160553e-01 2.88996845e-01
3.90981250e-02 4.46714729e-01 2.77465492e-01 -1.57331407e-01
-7.92222381e-01 -1.12817681e+00 -9.91279110e-02 6.10470057e-01
3.11652303e-01 -8.14269662e-01 -7.17177033e-01 -7.38235056e-01
-8.63727331e-02 1.29543751e-01 -4.43977773e-01 -2.51475070e-02
-4.96611029e-01 -1.26238418e+00 2.02983558e-01 2.70360768e-01
6.17524683e-01 -2.83433765e-01 3.64981949e-01 2.46210799e-01
-6.67271137e-01 -9.16181147e-01 -6.27035499e-01 -2.07072094e-01
-1.03499079e+00 -1.27098536e+00 -3.81147772e-01 -6.56300962e-01
8.36441159e-01 7.12684929e-01 8.54951501e-01 5.22408545e-01
2.44984001e-01 2.79507667e-01 -6.42132342e-01 3.43599379e-01
3.02911904e-02 -7.80558214e-02 5.16526222e-01 7.95799255e-01
2.82329649e-01 -8.15203130e-01 -5.45106292e-01 7.02027917e-01
-9.85113442e-01 -1.03634298e-01 6.50037229e-01 9.61888969e-01
8.46418738e-01 7.47356713e-01 4.05077040e-01 -6.92590117e-01
7.21212864e-01 -8.02857220e-01 -1.68038175e-01 3.55150551e-01
-6.45966172e-01 -2.48561781e-02 6.89580977e-01 -4.46237922e-01
-8.41088951e-01 1.25060588e-01 4.68547672e-01 -6.81195796e-01
1.55389115e-01 1.01675236e+00 -3.36540103e-01 2.48168921e-03
2.13880196e-01 2.88274556e-01 -1.88658923e-01 -5.73776186e-01
4.42185938e-01 6.41634583e-01 5.85407495e-01 -9.37601328e-01
1.14002943e+00 6.56001925e-01 1.47559032e-01 -7.20563769e-01
-8.97403181e-01 -9.35375273e-01 -1.09247649e+00 -3.54271293e-01
7.65163302e-01 -7.48397470e-01 -7.72176147e-01 2.82975376e-01
-8.28035474e-01 3.74978513e-01 2.28030547e-01 6.75737143e-01
-6.80885613e-02 8.13322902e-01 -4.02017295e-01 -7.77974367e-01
1.13036849e-01 -1.16776025e+00 8.36694777e-01 -6.16782978e-02
-5.30856326e-02 -9.11029756e-01 1.94789439e-01 7.15446651e-01
-2.00946540e-01 4.76952136e-01 6.24062479e-01 -3.70100826e-01
-6.59879565e-01 -6.14721961e-02 -7.07055926e-02 3.29652607e-01
2.30762497e-01 3.11481148e-01 -5.45002759e-01 -6.26045406e-01
-5.67474738e-02 2.09063426e-01 8.63846421e-01 3.09356619e-02
1.14659071e+00 -3.38059992e-01 -5.60973287e-01 6.80376589e-01
1.41615546e+00 1.65907711e-01 4.01869833e-01 5.25511727e-02
1.21667361e+00 6.61886692e-01 7.41316974e-01 4.87714142e-01
6.79046810e-01 7.47643948e-01 2.19529018e-01 1.06380209e-01
3.85576814e-01 8.81595314e-02 2.83275425e-01 1.93505728e+00
-6.58455789e-01 1.06735580e-01 -9.32617009e-01 5.13418436e-01
-2.35108924e+00 -9.87630725e-01 -7.30157554e-01 2.07131553e+00
6.19018197e-01 -2.59038895e-01 2.92138755e-01 4.43604171e-01
9.35734749e-01 2.02102903e-02 -2.39905477e-01 1.58035263e-01
-7.67059028e-02 -1.71498269e-01 1.26389384e-01 4.10096854e-01
-1.11869097e+00 4.70391154e-01 5.60239267e+00 8.63637149e-01
-7.90805638e-01 1.66946521e-03 2.24731773e-01 4.11537383e-03
-2.12852120e-01 2.98893809e-01 -3.80114764e-01 7.88375258e-01
5.43380380e-01 -1.63323656e-01 7.52224267e-01 4.33176249e-01
2.21387625e-01 8.94175544e-02 -1.09011507e+00 9.55692589e-01
1.30240455e-01 -9.04639363e-01 -6.08879551e-02 2.97026068e-01
1.01143134e+00 -3.35243315e-01 5.79623654e-02 7.73093700e-02
4.99175847e-01 -7.80994713e-01 3.48203659e-01 7.21887946e-01
3.38346481e-01 -8.75595212e-01 9.36069906e-01 3.13711196e-01
-1.72365963e+00 1.82178691e-02 -2.34441295e-01 1.52232856e-01
8.81518796e-02 1.14940774e+00 -4.99965906e-01 1.36891508e+00
1.05446780e+00 1.25527751e+00 -5.80600381e-01 1.03461015e+00
1.29122332e-01 6.33151829e-01 -3.53971153e-01 2.34942570e-01
2.37859026e-01 -9.13218737e-01 6.90248668e-01 1.10895538e+00
4.60683495e-01 4.27524805e-01 7.70112574e-01 7.00445652e-01
8.04205835e-02 1.70833349e-01 -3.31804127e-01 5.95958270e-02
6.60765052e-01 1.62631977e+00 -7.48385906e-01 -4.28254813e-01
-2.40887940e-01 8.70139599e-01 3.98652613e-01 5.28579116e-01
-6.09561503e-01 -7.47194216e-02 5.38682103e-01 -3.04057598e-01
2.48168185e-01 -4.06531692e-01 -3.75754327e-01 -1.38707805e+00
2.46738493e-01 -1.01292133e+00 5.87666810e-01 -5.51635921e-01
-1.65122986e+00 4.91554379e-01 1.10207297e-01 -1.56454742e+00
1.69713125e-01 -4.61857527e-01 -5.54379046e-01 6.79163814e-01
-9.29811180e-01 -1.07477152e+00 -4.75727439e-01 7.73451388e-01
-4.62883934e-02 -7.82808438e-02 6.41883194e-01 3.96587014e-01
-1.13166797e+00 3.10187757e-01 4.89604235e-01 2.16049738e-02
6.83211505e-01 -1.29811966e+00 -6.32901430e-01 9.78669643e-01
-2.13467386e-02 1.11474669e+00 4.55616534e-01 -7.53459036e-01
-1.52544343e+00 -1.26691031e+00 5.10271847e-01 -3.88522923e-01
7.91218638e-01 -8.31285715e-02 -1.07134569e+00 4.74929184e-01
9.94785875e-02 -6.87405542e-02 1.07937038e+00 4.77397293e-01
-5.05590796e-01 -4.10563409e-01 -9.05842483e-01 6.14769697e-01
1.07750976e+00 -4.33609724e-01 -5.07114708e-01 2.54327029e-01
4.48063880e-01 1.89502612e-02 -1.53114271e+00 4.87156600e-01
3.41576934e-01 -1.03776634e+00 1.05620646e+00 -3.20440441e-01
2.49246150e-01 -1.28849018e+00 -4.56551999e-01 -1.57657993e+00
-7.72220969e-01 -4.48441982e-01 -2.36018747e-01 1.52642500e+00
1.04437776e-01 -4.95250255e-01 4.49646086e-01 1.60554603e-01
-2.63951808e-01 -9.10233378e-01 -7.91673899e-01 -7.65167773e-01
-2.12759078e-01 -2.65432388e-01 7.20222950e-01 1.51480496e+00
2.07071021e-01 4.47344989e-01 -3.71457130e-01 5.33267736e-01
1.15923274e+00 3.51101100e-01 6.99248493e-01 -1.24552214e+00
-2.88989425e-01 -4.88423616e-01 -3.02177131e-01 -8.38329732e-01
2.09884606e-02 -1.19908917e+00 4.68562059e-02 -1.28670275e+00
5.12248814e-01 -6.97131515e-01 -4.98855442e-01 2.15342462e-01
-5.60619056e-01 3.80809933e-01 3.23984534e-01 7.09776759e-01
-9.38674390e-01 5.59886277e-01 1.27984142e+00 -2.70830393e-01
-2.29282916e-01 -4.23869848e-01 -8.27894807e-01 6.95221722e-01
5.03278673e-01 -3.44575822e-01 -3.63676518e-01 -8.61965790e-02
2.05820948e-02 -1.15777925e-01 3.34666707e-02 -1.07710743e+00
5.99862218e-01 -3.50680083e-01 2.72483170e-01 -8.46495628e-01
2.52382725e-01 -1.01691139e+00 6.27459407e-01 2.51526654e-01
9.39342380e-02 4.33957763e-03 -3.57343704e-01 7.78677464e-01
-4.64515984e-01 1.65600955e-01 4.59970444e-01 -1.47201279e-02
-4.28983063e-01 4.36672956e-01 -1.34418458e-01 -1.18301272e-01
1.06694221e+00 -3.13937157e-01 -2.68405080e-01 1.86625551e-02
-9.04713392e-01 5.89507341e-01 4.88129616e-01 1.96041986e-01
6.44480705e-01 -1.86233091e+00 -9.38683033e-01 -7.78200179e-02
5.99899776e-02 2.17640415e-01 3.20753127e-01 1.39477825e+00
-9.24227461e-02 1.49568558e-01 4.74746749e-02 -1.04604340e+00
-1.28802705e+00 7.61858940e-01 7.79695064e-02 -4.54451948e-01
-1.99505627e-01 5.94919920e-01 4.61991616e-02 -5.30247211e-01
-1.49421737e-01 -4.86714998e-03 -3.30979824e-01 2.70373464e-01
3.27741891e-01 7.58919954e-01 -7.29152709e-02 -1.12411988e+00
-6.09696150e-01 1.05701101e+00 4.59721461e-02 -1.38770044e-02
1.24847627e+00 -2.78443277e-01 -9.34786141e-01 6.14497423e-01
1.26038015e+00 -7.89549574e-02 -7.01423883e-01 -3.73682767e-01
1.28574893e-01 -6.59764230e-01 3.46898064e-02 -3.53221476e-01
-1.32963216e+00 4.34789091e-01 3.17737520e-01 3.88808668e-01
1.35539639e+00 -1.23814613e-01 5.91971457e-01 3.31066847e-01
3.22892100e-01 -1.11993504e+00 3.02374482e-01 3.02383482e-01
8.50338042e-01 -1.07866776e+00 4.03467536e-01 -8.25958908e-01
-5.62143147e-01 9.17288184e-01 6.68490231e-01 -1.92438468e-01
8.30466270e-01 -1.28490716e-01 -2.35904753e-01 -3.42279285e-01
-5.94446480e-01 -2.45912269e-01 6.13606215e-01 4.86772835e-01
5.14894485e-01 4.97389019e-01 -4.37430054e-01 5.34961700e-01
-1.42626598e-01 -2.82373518e-01 2.48467237e-01 6.11570358e-01
-2.36252114e-01 -1.22535503e+00 -9.17093039e-01 4.90676790e-01
-8.43367726e-02 1.54129356e-01 -5.57710648e-01 3.01779091e-01
6.39206886e-01 1.46728277e+00 -9.18443352e-02 -1.03882861e+00
1.39447778e-01 -8.20590928e-02 2.08117679e-01 -4.75329757e-01
-4.25013900e-01 3.65335196e-01 2.20265351e-02 -5.68449616e-01
-9.77144659e-01 -8.17447126e-01 -1.04157829e+00 -5.35585523e-01
-5.11060357e-01 4.62267786e-01 3.33874106e-01 8.36698651e-01
2.54479289e-01 5.19016981e-01 1.02685058e+00 -9.28445101e-01
-2.01778784e-01 -8.77943933e-01 -6.51466846e-01 7.64131248e-01
1.22668728e-01 -8.17492664e-01 -4.96188372e-01 1.30144671e-01] | [7.960501194000244, 4.654808521270752] |
56853ae2-bf5f-4381-b172-af0923de2c2f | sass-data-and-methods-for-subject-aware | 2303.14589 | null | https://arxiv.org/abs/2303.14589v1 | https://arxiv.org/pdf/2303.14589v1.pdf | SASS: Data and Methods for Subject Aware Sentence Simplification | Sentence simplification tends to focus on the generic simplification of sentences by making them more readable and easier to understand. This paper provides a dataset aimed at training models that perform subject aware sentence simplifications rather than simplifying sentences as a whole. We also test models on that dataset which are inspired by model architecture used in abstractive summarization. We hand generated portions of the data and augment the dataset by further manipulating those hand written simplifications. Our results show that data-augmentation, data-masking, and model architecture choices used in summarization provide a solid baseline for comparison on subject aware simplification. | ['Anand Tyagi', 'Luke Martin', 'Brad Windsor'] | 2023-03-26 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 3.10082406e-01 7.15831578e-01 2.30174419e-02 -7.99994707e-01
-9.24262822e-01 -4.36237693e-01 6.81864262e-01 5.62887967e-01
-6.04815185e-01 8.70765507e-01 1.11382830e+00 -1.95755899e-01
1.38338938e-01 -5.44313431e-01 -4.72302616e-01 4.14799750e-02
2.75959104e-01 6.98130667e-01 -3.81586224e-01 -7.13779747e-01
5.07335544e-01 4.68419313e-01 -1.19257128e+00 5.44533491e-01
1.34629774e+00 -6.63183481e-02 4.73656058e-02 1.09582627e+00
-3.41425417e-03 9.80340481e-01 -1.50837946e+00 -7.09879279e-01
1.44339710e-01 -5.69501758e-01 -1.19262040e+00 -1.39837787e-01
1.04597592e+00 -1.90972239e-01 -4.87961322e-01 7.94473469e-01
6.29920661e-01 4.87295717e-01 6.96613729e-01 -6.61382675e-01
-9.43437219e-01 1.12093270e+00 -4.57911119e-02 4.99038249e-01
5.99969864e-01 1.01684511e-01 8.73489678e-01 -4.69511271e-01
9.85861897e-01 1.54543042e+00 7.65377104e-01 8.68285596e-01
-1.33941674e+00 -1.57421589e-01 3.22804719e-01 8.57085884e-02
-9.50112700e-01 -9.64483440e-01 7.02685595e-01 -9.19969752e-02
1.68244433e+00 1.00777996e+00 6.70295537e-01 9.19811845e-01
5.32000184e-01 7.18244255e-01 3.77399087e-01 -5.92769206e-01
-3.81617658e-02 2.72070896e-02 7.47902393e-01 7.74639025e-02
5.76416671e-01 -5.95945597e-01 -2.94075489e-01 3.04387752e-02
-6.86921328e-02 -2.05812886e-01 -4.00572896e-01 2.72126079e-01
-8.94577146e-01 7.89987087e-01 2.58321434e-01 1.10934600e-01
-5.23357034e-01 -1.25262037e-01 8.46727788e-01 6.20105803e-01
7.72281706e-01 1.46449447e+00 -5.55461228e-01 -1.38287008e-01
-1.33486998e+00 8.04969847e-01 1.11649513e+00 1.23766041e+00
5.91359615e-01 3.30845624e-01 -9.06567931e-01 7.64697075e-01
-5.16784370e-01 3.86833757e-01 5.30398965e-01 -1.19044280e+00
9.27425027e-01 6.03158534e-01 -6.21596463e-02 -6.52322471e-01
-6.60783887e-01 -4.23472792e-01 -9.77919757e-01 -2.12397337e-01
-2.49547526e-01 -3.60718220e-01 -1.00488985e+00 1.73507953e+00
-3.11761528e-01 -9.11956549e-01 2.52638817e-01 3.98889363e-01
1.25268066e+00 6.59652352e-01 3.59008402e-01 -4.56905663e-01
9.80454683e-01 -1.14779639e+00 -1.12734473e+00 -2.48572573e-01
1.32294250e+00 -8.85007679e-01 1.29018271e+00 3.01873982e-01
-1.55261290e+00 -6.23604357e-01 -1.00068319e+00 -1.00070167e+00
-3.31835151e-01 -5.38244471e-02 5.36174357e-01 4.23005670e-01
-1.25934887e+00 9.13069963e-01 -5.93621373e-01 -6.03022635e-01
5.93425930e-01 2.98073828e-01 -3.71773362e-01 1.35400608e-01
-1.08996105e+00 1.67795157e+00 4.85986382e-01 -3.71423483e-01
-4.04065758e-01 -1.02604175e+00 -1.14873302e+00 3.64409745e-01
-3.56927738e-02 -1.33630693e+00 1.69241047e+00 -5.80346644e-01
-1.28922510e+00 7.92210519e-01 -5.77140987e-01 -1.04574251e+00
2.64279544e-01 -8.04126918e-01 -3.49695146e-01 -4.30804007e-02
1.23682059e-01 7.66938329e-01 5.11050105e-01 -9.15027738e-01
-1.28923923e-01 4.26146910e-02 2.82583833e-01 5.02203763e-01
-3.33039343e-01 3.68527919e-02 2.10911594e-02 -9.98611212e-01
-5.65199018e-01 -4.80730921e-01 -3.24934185e-01 -8.32733393e-01
-1.18246865e+00 -1.15893558e-01 7.18398213e-01 -1.11972117e+00
1.96655297e+00 -1.54284227e+00 4.44802582e-01 -4.22193706e-01
4.38328773e-01 6.59343421e-01 -5.33335030e-01 9.94821727e-01
-4.28097934e-01 8.59598160e-01 -4.05023366e-01 -1.15539670e+00
-1.34683535e-01 2.23128885e-01 -5.26807189e-01 9.46875196e-03
2.54326433e-01 1.06940091e+00 -6.54976189e-01 -6.29683435e-01
3.26989383e-01 2.49352921e-02 -1.06171739e+00 5.14459237e-02
-2.01435521e-01 9.19991806e-02 -7.64663815e-02 2.03426436e-01
6.67656958e-01 5.54790318e-01 -2.93477148e-01 -2.35961333e-01
-1.52718142e-01 9.87432837e-01 -4.78398681e-01 1.87243652e+00
-4.76048499e-01 1.02908373e+00 -2.80255646e-01 -6.23592794e-01
6.13072157e-01 8.08321238e-02 -1.60096973e-01 -5.67352951e-01
1.28027216e-01 -3.79882336e-01 2.89150745e-01 -7.13312924e-01
1.16746414e+00 -1.21508956e-01 -2.58999825e-01 4.75983322e-01
8.99455622e-02 -1.17270267e+00 1.03362834e+00 8.30125809e-01
1.16613567e+00 -3.93305361e-01 5.97388268e-01 -5.03039062e-01
4.47837830e-01 2.21488446e-01 3.90622705e-01 1.03991830e+00
2.51033962e-01 9.15539682e-01 5.40654123e-01 -4.06789750e-01
-1.36969674e+00 -6.99581563e-01 1.32612512e-01 8.68662715e-01
-4.77374762e-01 -1.23111248e+00 -1.19305348e+00 -6.19774342e-01
-1.75202221e-01 1.88052559e+00 -6.75956845e-01 -5.23694813e-01
-9.35042977e-01 -4.78770167e-01 6.27811849e-01 2.50123948e-01
2.36959487e-01 -1.12027359e+00 -4.28908795e-01 -1.40355946e-02
-3.96947384e-01 -5.93934298e-01 -5.81659496e-01 1.46017626e-01
-1.17310452e+00 -8.37922990e-01 -2.26600930e-01 -5.54838419e-01
6.03565753e-01 2.52148002e-01 1.69198382e+00 2.58557022e-01
5.48142903e-02 7.14714453e-02 -3.92762750e-01 -1.17010903e+00
-8.18063498e-01 8.09904933e-01 3.12430225e-02 -9.25732911e-01
1.75188348e-01 -5.95288455e-01 -1.19006932e-01 -6.68373168e-01
-1.17624640e+00 1.79600194e-01 5.33847570e-01 5.04376948e-01
2.47444779e-01 -2.88618028e-01 5.02177715e-01 -1.55960858e+00
1.45829809e+00 -1.63501740e-01 1.41746044e-01 3.12934577e-01
-2.93855786e-01 1.65874839e-01 1.02294695e+00 -5.40489182e-02
-1.16514611e+00 -4.47658628e-01 -4.93468523e-01 5.87644503e-02
-1.86223045e-01 6.54491127e-01 -2.35766321e-01 3.20782512e-01
1.02096438e+00 -5.92707545e-02 -1.40433833e-01 -8.70335281e-01
6.48144543e-01 4.89996493e-01 5.85411608e-01 -1.54571503e-01
8.50335121e-01 8.85718539e-02 -1.57078460e-01 -1.25929821e+00
-1.14125109e+00 -4.48861457e-02 -8.64476979e-01 3.94182354e-01
5.24144828e-01 -6.77892268e-01 4.02506143e-02 1.16984896e-01
-1.66692126e+00 -2.24377885e-01 -1.08163905e+00 8.75182971e-02
-4.15925175e-01 6.23687387e-01 -5.37786365e-01 -3.72560233e-01
-1.04183221e+00 -5.93566775e-01 9.84675705e-01 1.42082140e-01
-1.35885358e+00 -1.05876553e+00 3.28517169e-01 1.86625317e-01
5.59949100e-01 1.39467716e-01 1.19739270e+00 -1.06886160e+00
2.54169311e-02 -4.87431556e-01 2.20056623e-01 6.62018359e-01
2.27629483e-01 4.15013939e-01 -8.04754853e-01 -4.52057011e-02
3.10728550e-01 -1.89035878e-01 1.09500873e+00 6.34812236e-01
1.25927055e+00 -8.80087614e-01 -8.64113122e-02 7.54446447e-01
8.65231633e-01 -2.17693195e-01 9.77185667e-01 3.12379926e-01
8.22947800e-01 5.16901374e-01 2.73609072e-01 2.94384539e-01
5.15460193e-01 3.53673369e-01 1.21148005e-02 -3.49010170e-01
-6.23488545e-01 -1.75669849e-01 7.66820647e-03 1.18057084e+00
1.41803518e-01 -7.78938115e-01 -5.97690344e-01 6.10684156e-01
-1.49677682e+00 -1.03069067e+00 -3.87232929e-01 1.62489820e+00
1.16801870e+00 3.04535449e-01 -1.51471794e-01 1.79307640e-01
3.24644238e-01 2.56372303e-01 -2.72597492e-01 -1.34391367e+00
-5.72270751e-01 4.56163734e-01 2.42613524e-01 1.07464826e+00
-1.09647429e+00 1.15208960e+00 7.54264164e+00 5.75047314e-01
-7.24454165e-01 -1.51875794e-01 4.43922013e-01 -7.18171060e-01
-6.23775303e-01 -7.32352957e-03 -6.83856249e-01 2.16679856e-01
1.08291090e+00 -7.45857954e-01 2.04376236e-01 5.09256303e-01
6.77839875e-01 -2.51459897e-01 -1.53641152e+00 5.80778718e-01
3.49208951e-01 -1.53553569e+00 1.06385660e+00 -4.76750761e-01
9.09737766e-01 -1.46146074e-01 -2.24969462e-01 6.17316484e-01
2.07264617e-01 -1.24844050e+00 7.74506450e-01 1.04278636e+00
3.80175143e-01 -7.14971662e-01 9.66395557e-01 4.31334138e-01
-2.56452858e-01 1.95125356e-01 -3.68092179e-01 -8.62498343e-01
4.16769326e-01 6.14834964e-01 -6.51460052e-01 6.02002621e-01
2.62998104e-01 7.97100365e-01 -1.29361951e+00 9.21881676e-01
-4.21500832e-01 5.29076934e-01 -1.79934114e-01 1.11297175e-01
-2.20140442e-01 -3.03520821e-02 9.73663092e-01 1.81364143e+00
-1.88676864e-01 4.72312570e-02 -2.63553083e-01 4.79928017e-01
-4.07488585e-01 1.68785065e-01 -8.30999672e-01 -3.04229017e-02
5.91882408e-01 9.43885326e-01 8.05344339e-03 -8.05892050e-01
-8.39270130e-02 8.52370739e-01 6.32959425e-01 4.07284588e-01
-5.07462561e-01 -9.27300096e-01 7.52019286e-01 -1.43945396e-01
3.05881239e-02 -1.10880345e-01 -1.14179754e+00 -1.48681974e+00
1.95495635e-02 -1.07197428e+00 1.71191365e-01 -7.14457393e-01
-9.44497347e-01 6.58024430e-01 3.21964234e-01 -7.60984540e-01
-3.39199334e-01 1.48696408e-01 -1.15860879e+00 1.15284443e+00
-1.28452694e+00 -9.10101473e-01 -1.01512268e-01 -3.26111056e-02
9.80171800e-01 1.14757866e-01 1.10886729e+00 3.11357267e-02
-8.83696973e-01 8.00380528e-01 -5.69121130e-02 -6.17607906e-02
8.25987339e-01 -1.31412184e+00 1.16676271e+00 1.13438094e+00
-1.50902987e-01 1.25580955e+00 1.37422037e+00 -8.33698034e-01
-8.87678921e-01 -1.22016621e+00 1.64898527e+00 -1.01166332e+00
1.18890002e-01 -4.10612583e-01 -1.06837153e+00 9.98294473e-01
8.96602511e-01 -1.12951124e+00 5.60846686e-01 2.96602190e-01
9.49613452e-02 -7.49743804e-02 -1.13623965e+00 9.78928387e-01
1.15489388e+00 -2.67936796e-01 -1.43469369e+00 7.00718462e-01
1.05649853e+00 -5.32885969e-01 -6.59451544e-01 3.01329464e-01
-1.14803113e-01 -6.64743125e-01 7.27505445e-01 -1.25198412e+00
8.55590045e-01 1.45927042e-01 1.22436799e-01 -1.87064958e+00
-5.55135012e-01 -9.18364048e-01 -2.37545431e-01 1.54678929e+00
5.42750657e-01 -2.60789722e-01 4.95698035e-01 8.92485142e-01
-9.01655316e-01 -4.51336354e-01 -6.50387526e-01 -5.74687600e-01
5.83496094e-01 -5.26819080e-02 6.36172891e-01 7.11974859e-01
4.60942760e-02 8.97211730e-01 -8.98580775e-02 -4.66791749e-01
2.62976497e-01 -1.93403274e-01 1.13220167e+00 -1.10216963e+00
3.01663786e-01 -7.26135015e-01 5.34037575e-02 -9.26221311e-01
1.99137419e-01 -9.77868855e-01 -2.59628952e-01 -2.10818243e+00
2.08624557e-01 2.47581407e-01 4.91370052e-01 4.63952988e-01
-5.58733284e-01 -2.28364039e-02 5.69559157e-01 1.52693585e-01
-7.12362349e-01 8.49901855e-01 1.36923218e+00 -2.89750099e-01
-3.42540920e-01 -1.72042459e-01 -1.40426373e+00 8.12492311e-01
9.43434238e-01 -5.03422678e-01 -3.94175977e-01 -1.11469555e+00
-1.72838885e-02 -4.55691367e-01 -7.80371651e-02 -1.09616971e+00
1.51673406e-01 1.80430617e-02 2.24901602e-01 -7.53256977e-01
2.97278404e-01 -2.58634597e-01 -1.92367524e-01 5.43515742e-01
-8.56351197e-01 2.59693265e-01 5.62455595e-01 -2.32434962e-02
-2.32294351e-01 -5.38365304e-01 6.74098670e-01 -3.20059478e-01
-1.39130726e-01 -2.62528032e-01 -5.79096317e-01 5.22099018e-01
5.29930830e-01 -8.15796014e-03 -6.41440868e-01 -8.04191768e-01
-6.53874576e-01 4.10792947e-01 6.77375495e-01 2.59861499e-01
3.67223710e-01 -7.11599886e-01 -1.09585643e+00 1.37333088e-02
-2.39084929e-01 2.09714696e-01 2.43590891e-01 4.04446334e-01
-8.37470829e-01 8.67446899e-01 -1.19178526e-01 -5.87827452e-02
-1.32744741e+00 6.17438138e-01 2.09104314e-01 -4.89727169e-01
-5.45662165e-01 8.25610638e-01 8.07820410e-02 -6.76649451e-01
1.38302952e-01 -8.22948933e-01 -3.88789713e-01 1.49527803e-01
6.36535048e-01 6.46087110e-01 4.59683329e-01 -2.47742549e-01
-1.58623576e-01 2.07976475e-01 -6.18568599e-01 2.12282553e-01
1.40318274e+00 -3.33326727e-01 -2.73819357e-01 2.66344190e-01
1.08746076e+00 2.03927010e-01 -5.03365636e-01 -2.31154151e-02
-8.67161453e-02 2.96465345e-02 -2.43687183e-01 -1.04784536e+00
-3.79847288e-01 7.95767009e-01 -4.58054692e-01 4.77749109e-01
1.16576016e+00 -3.32652181e-01 7.40644097e-01 1.07592797e+00
-3.87178481e-01 -1.13429070e+00 -4.99236047e-01 1.03732240e+00
1.53181219e+00 -9.03327584e-01 8.54521096e-01 -5.19729912e-01
-8.20247829e-01 1.05274820e+00 6.17562771e-01 -3.81069988e-01
1.51944935e-01 3.93349379e-02 1.34506244e-02 -1.62095115e-01
-9.88568723e-01 7.61618689e-02 2.80290037e-01 8.48572552e-01
6.30351245e-01 -7.17795938e-02 -7.41571128e-01 7.70593524e-01
-1.14103198e+00 -1.94629073e-01 1.01212811e+00 8.53351653e-01
-3.24702621e-01 -1.07243860e+00 -1.26697391e-01 1.01466680e+00
-3.42073143e-01 -6.39191628e-01 -1.13532412e+00 1.03343034e+00
-1.63645476e-01 1.12764263e+00 6.28689751e-02 -2.43018597e-01
9.86593604e-01 1.94593474e-01 4.78145361e-01 -1.31242406e+00
-1.15694046e+00 -5.74634492e-01 8.24059665e-01 -1.75475791e-01
-1.46801043e-02 -7.10401773e-01 -9.08894122e-01 -6.98033690e-01
-6.26195893e-02 3.76972288e-01 3.65943134e-01 9.88710403e-01
7.38658190e-01 1.02106941e+00 2.13025242e-01 -1.09574854e+00
-6.15758538e-01 -1.72061133e+00 -1.50660262e-01 7.80677855e-01
5.91665268e-01 9.98980999e-02 -3.26282650e-01 2.89204806e-01] | [11.313506126403809, 10.124588012695312] |
b1d8c5c3-c554-48d3-8fc8-9aa01db0ad0c | adapting-marbert-for-improved-arabic-dialect | 2103.01065 | null | https://arxiv.org/abs/2103.01065v1 | https://arxiv.org/pdf/2103.01065v1.pdf | Adapting MARBERT for Improved Arabic Dialect Identification: Submission to the NADI 2021 Shared Task | In this paper, we tackle the Nuanced Arabic Dialect Identification (NADI) shared task (Abdul-Mageed et al., 2021) and demonstrate state-of-the-art results on all of its four subtasks. Tasks are to identify the geographic origin of short Dialectal (DA) and Modern Standard Arabic (MSA) utterances at the levels of both country and province. Our final model is an ensemble of variants built on top of MARBERT that achieves an F1-score of 34.03% for DA at the country-level development set -- an improvement of 7.63% from previous work. | ['Khaled Essam', 'Muhammad ElNokrashy', 'Mohamed Gabr', 'Badr AlKhamissi'] | 2021-03-01 | null | https://aclanthology.org/2021.wanlp-1.29 | https://aclanthology.org/2021.wanlp-1.29.pdf | eacl-wanlp-2021-4 | ['dialect-identification'] | ['natural-language-processing'] | [-4.47990477e-01 -9.54215899e-02 1.21107750e-01 -6.15185022e-01
-1.28753495e+00 -8.58378232e-01 9.64122713e-01 -2.74466842e-01
-3.70598942e-01 6.27757370e-01 2.66724885e-01 -5.16732931e-01
9.47997421e-02 -3.80365491e-01 -2.27198154e-01 -5.65619648e-01
-3.28767896e-01 8.32207620e-01 -2.86240242e-02 -1.02825928e+00
9.25450772e-02 3.97997022e-01 -8.48646939e-01 5.78812897e-01
1.32020617e+00 4.53966975e-01 -2.43424594e-01 6.38699412e-01
2.44270816e-01 1.04392159e+00 -6.17052019e-01 -8.71616423e-01
-6.60474673e-02 -4.49044943e-01 -1.36661744e+00 -2.30549172e-01
7.95765758e-01 -6.36149883e-01 -1.95586562e-01 8.25000167e-01
5.17913103e-01 -3.06160361e-01 1.02981317e+00 -9.29983914e-01
-1.12633586e+00 1.29360390e+00 -8.55376542e-01 5.86466789e-02
3.04125518e-01 -1.22684419e-01 1.15878224e+00 -1.56939101e+00
5.60985744e-01 1.44250619e+00 9.94545996e-01 4.07953024e-01
-9.88624811e-01 -6.06780529e-01 1.92791700e-01 2.89244086e-01
-1.62395382e+00 -8.59008968e-01 3.84236246e-01 -3.25873166e-01
9.71566498e-01 4.06595767e-02 -3.61246504e-02 8.03887665e-01
-3.20960820e-01 1.24253452e+00 1.26531959e+00 -6.93842053e-01
-3.41440409e-01 -1.60836503e-01 6.06866181e-01 9.77134824e-01
-2.70105839e-01 -6.42295897e-01 -5.11006534e-01 -1.32938623e-01
3.84305656e-01 -8.89188528e-01 -8.28196481e-02 3.47252399e-01
-1.27128077e+00 9.49398577e-01 2.74579704e-01 3.23181123e-01
-4.26269025e-01 -4.45148706e-01 1.78492993e-01 6.35717928e-01
7.88774073e-01 1.04194276e-01 -6.83886290e-01 -4.15386915e-01
-6.85577333e-01 4.87806380e-01 8.87595892e-01 7.35578418e-01
4.15808141e-01 4.04758334e-01 3.25608850e-01 1.42934465e+00
5.02899766e-01 8.43534052e-01 3.98240954e-01 -6.13741577e-01
5.89596629e-01 3.17121863e-01 8.46000910e-02 -3.73283565e-01
-4.60448295e-01 -3.55556272e-02 -6.24266863e-01 8.17862451e-02
1.29535627e+00 -6.89899564e-01 -1.10445976e+00 1.62850451e+00
-8.50571599e-03 -2.27313682e-01 2.31181026e-01 8.28537047e-01
7.37343967e-01 9.40598667e-01 -1.15474023e-01 2.28139684e-01
1.40507162e+00 -1.03099990e+00 -6.07533336e-01 -3.52005243e-01
8.15414071e-01 -9.70585346e-01 1.05922663e+00 4.82307702e-01
-1.07799065e+00 -1.27951622e-01 -9.70269203e-01 9.21629295e-02
-6.03456557e-01 5.05577087e-01 2.72610456e-01 1.15863347e+00
-1.33948076e+00 -8.63892362e-02 -4.86142635e-01 -4.96736765e-01
1.38795614e-01 2.07742423e-01 -5.00670135e-01 -2.94988006e-01
-1.39076829e+00 1.16827834e+00 9.80027765e-02 1.26596451e-01
-7.88636386e-01 -7.48766065e-01 -8.04678917e-01 -4.60434616e-01
-1.61273107e-01 2.56254405e-01 1.23348486e+00 -1.12006748e+00
-1.64909208e+00 1.32718062e+00 -5.45728542e-02 -4.22574133e-01
2.19596513e-02 -4.63168353e-01 -8.46303344e-01 -1.25405520e-01
1.24213018e-01 4.84943867e-01 7.32592583e-01 -1.20070994e+00
-9.96867299e-01 -7.25688636e-01 4.28728275e-02 3.37469727e-01
-1.27802968e-01 8.36367071e-01 -1.67558566e-01 -9.38435078e-01
4.98756692e-02 -1.27107179e+00 1.12061270e-01 -9.86657917e-01
-1.85793713e-01 -2.46998534e-01 7.21838772e-01 -1.83892870e+00
1.32295418e+00 -1.92745388e+00 2.06359312e-01 3.32327634e-01
2.08854470e-02 4.49896306e-01 -5.01256883e-01 3.63298953e-01
2.87764389e-02 -2.71363020e-01 -5.35434127e-01 -2.63699681e-01
1.69975460e-01 -1.57222971e-01 -1.95510060e-01 5.49891055e-01
5.57102203e-01 7.78083801e-01 -7.93624699e-01 1.45391569e-01
-3.91012073e-01 2.49372914e-01 -1.49861082e-01 -8.12614635e-02
-3.36866006e-02 3.35598618e-01 3.47445399e-01 1.16276670e+00
9.82356012e-01 4.29295421e-01 2.50165820e-01 2.45830908e-01
-3.76564950e-01 5.67414999e-01 -9.23979998e-01 1.49502563e+00
-3.83662283e-01 6.54157639e-01 6.09576583e-01 -4.69496250e-01
8.72790813e-01 2.59533614e-01 1.19585350e-01 -5.08816183e-01
-1.04681075e-01 8.48685384e-01 5.25362432e-01 1.18891455e-01
6.23153567e-01 3.75859469e-01 -4.31333661e-01 7.75362313e-01
2.28042081e-01 1.37057543e-01 1.31283700e-01 8.58519897e-02
4.69705909e-01 1.34704262e-01 2.45800197e-01 -8.88052464e-01
9.05107558e-01 -9.24092159e-03 1.00102074e-01 6.92729592e-01
-3.82388413e-01 7.30816424e-01 5.02709389e-01 -4.30080742e-01
-8.47160161e-01 -1.36718845e+00 -2.48638928e-01 1.77029765e+00
-6.33376718e-01 -2.51261711e-01 -1.06417286e+00 -1.04885983e+00
-1.50661469e-01 6.87515378e-01 -6.00342751e-01 4.34465110e-01
-1.16033125e+00 -1.36986208e+00 1.43795252e+00 3.38999212e-01
5.69961607e-01 -7.43292630e-01 2.33153135e-01 2.39308104e-01
-5.19445360e-01 -8.09808850e-01 -5.38793504e-01 -3.12459588e-01
-1.12858221e-01 -8.43141258e-01 -1.31672156e+00 -1.34434855e+00
2.22127095e-01 -4.60160039e-02 1.41997683e+00 -3.03714365e-01
3.62470925e-01 -1.96329565e-04 -5.86683512e-01 -6.45319223e-01
-8.09060872e-01 5.76595843e-01 3.56918760e-02 3.06914002e-01
4.89908844e-01 1.65436685e-01 4.01265472e-02 3.46956253e-01
-2.28015035e-01 -3.52715224e-01 1.44918799e-01 9.03659940e-01
-3.09686437e-02 -5.38315654e-01 1.07654190e+00 -9.57440794e-01
4.58281577e-01 -6.35179579e-01 -5.52400053e-01 4.96384233e-01
-4.25109357e-01 -3.09824556e-01 4.08527285e-01 2.54475381e-02
-1.34213853e+00 7.15275630e-02 -6.85059071e-01 6.89375579e-01
-2.66488165e-01 7.73592293e-01 -2.61677295e-01 -5.81426099e-02
6.45916879e-01 3.16872507e-01 1.40193015e-01 -5.15954196e-01
6.34966314e-01 1.53090572e+00 6.61462545e-01 -7.12162137e-01
3.43346000e-01 5.39252795e-02 -6.06911242e-01 -1.19718468e+00
-6.54135823e-01 -2.75290161e-01 -1.24957752e+00 -2.91860640e-01
6.33013666e-01 -1.37815022e+00 -1.47637531e-01 1.57516873e+00
-1.03755426e+00 -7.29691684e-01 7.45903611e-01 1.49137065e-01
-2.68935174e-01 2.35043868e-01 -1.10537636e+00 -7.23749340e-01
-4.58440185e-01 -1.09248424e+00 8.24542999e-01 -3.18727016e-01
-2.15961918e-01 -1.09002364e+00 1.85286701e-01 1.45015478e-01
4.96785879e-01 2.20876448e-02 1.11986828e+00 -8.62648547e-01
1.35297626e-01 1.72101706e-01 -3.43704134e-01 1.48813039e-01
3.26065779e-01 2.25909144e-01 -1.11622298e+00 -4.62175041e-01
-4.95659798e-01 -4.34213728e-01 6.44780040e-01 3.78625751e-01
1.86290950e-01 -3.09787780e-01 4.41974759e-01 2.61168450e-01
1.04334176e+00 2.23771304e-01 4.41142231e-01 6.08910620e-01
7.79832661e-01 8.51643562e-01 7.01775014e-01 3.81613314e-01
1.14229631e+00 6.23203456e-01 1.45271108e-01 2.14510486e-02
-1.41983554e-01 3.38431060e-01 7.80578494e-01 1.01232374e+00
-1.96697384e-01 8.40204358e-02 -1.66930115e+00 1.23809969e+00
-1.74573588e+00 -7.72518039e-01 -7.07626700e-01 2.06631756e+00
1.15730703e+00 -3.62748116e-01 9.45976138e-01 -4.82446849e-02
6.32982552e-01 6.29791096e-02 -2.98314691e-02 -6.99531674e-01
-5.93671024e-01 9.56347436e-02 3.82678807e-01 9.37492490e-01
-1.41387475e+00 1.58921611e+00 6.62729931e+00 7.97481894e-01
-9.13784742e-01 1.19389057e-01 8.04325938e-01 3.55198890e-01
9.26958025e-02 -4.44928735e-01 -9.55939293e-01 3.44484329e-01
1.57546604e+00 1.74158197e-02 8.00113201e-01 4.02629167e-01
-1.98638424e-01 1.50534868e-01 -7.18430579e-01 6.04308963e-01
3.09728295e-01 -8.66135299e-01 2.58334000e-02 -9.80207697e-03
8.94355774e-01 5.12573242e-01 4.41439956e-01 4.01245296e-01
8.23685825e-01 -1.15910089e+00 1.15930796e+00 2.56611407e-02
9.52906966e-01 -1.24918616e+00 9.35344636e-01 2.03024268e-01
-9.13437724e-01 -8.34165215e-02 -4.59084399e-02 1.62299901e-01
1.29560698e-02 3.61636840e-02 -8.63644481e-01 4.90129352e-01
7.54796326e-01 7.66144693e-01 -8.16615283e-01 3.48114938e-01
-1.69855326e-01 1.42432308e+00 -3.12503397e-01 1.73112258e-01
9.40004647e-01 -3.91347349e-01 3.60533446e-01 1.74635935e+00
2.30870619e-01 -1.18267454e-01 -2.23976269e-01 1.26912788e-01
-1.02288527e-02 3.08552235e-01 -5.27580023e-01 1.51119813e-01
4.80293512e-01 8.48973691e-01 -1.39760166e-01 -1.08557492e-01
-6.33920610e-01 1.13603485e+00 5.13820112e-01 2.04934582e-01
-6.33997202e-01 -7.78606474e-01 1.04161489e+00 -4.72002268e-01
3.63913268e-01 -4.98222917e-01 -3.79263401e-01 -9.20460582e-01
-3.47054809e-01 -1.38370955e+00 8.29386950e-01 -2.68224150e-01
-1.40573740e+00 8.23247731e-01 -3.35751265e-01 -7.89558649e-01
-8.48283768e-01 -9.72307146e-01 -3.26863676e-01 1.40912104e+00
-1.40438151e+00 -2.09403729e+00 2.00028583e-01 8.20373833e-01
5.80850899e-01 -9.58023489e-01 1.12510920e+00 4.45623040e-01
-5.14345527e-01 1.18998516e+00 7.03240573e-01 9.36876774e-01
1.20446634e+00 -1.63325667e+00 1.10391009e+00 1.09169018e+00
1.76044315e-01 4.27095592e-01 1.14414349e-01 -3.14585775e-01
-1.14100993e+00 -1.04253781e+00 1.48022366e+00 -9.32776034e-01
1.03434980e+00 -4.29596514e-01 -8.24010074e-01 9.38196957e-01
5.37444830e-01 -7.80668259e-01 5.97006798e-01 4.79191601e-01
-7.82972991e-01 1.27140209e-01 -1.20710659e+00 8.68062794e-01
6.15469277e-01 -7.05159605e-01 -3.45064372e-01 2.23175943e-01
4.35932904e-01 -3.32699835e-01 -1.15400314e+00 -4.99633886e-02
7.80320764e-01 -8.45490634e-01 9.57390249e-01 -6.63461685e-01
2.24278390e-01 3.71592864e-02 -3.26418161e-01 -1.80720222e+00
-4.13467169e-01 -6.44693017e-01 2.13393465e-01 1.53359032e+00
8.14761758e-01 -7.36535609e-01 2.74932474e-01 9.89408046e-02
-3.41261536e-01 -3.04247260e-01 -8.19377303e-01 -7.51337707e-01
9.04366493e-01 -3.07441533e-01 8.64306808e-01 1.39453161e+00
1.15738876e-01 5.71480453e-01 -3.64062339e-01 5.19547284e-01
5.04181743e-01 -2.76602805e-01 4.89886075e-01 -9.09955204e-01
2.59309590e-01 -7.48648763e-01 -1.58251654e-02 -6.79403543e-01
5.52952051e-01 -9.74648774e-01 1.05154840e-02 -1.19288290e+00
-5.57093084e-01 -4.32832062e-01 -9.60790440e-02 6.88835382e-01
-2.41566494e-01 6.10663712e-01 3.03136200e-01 2.37624973e-01
-2.63477445e-01 1.14365995e-01 4.99201298e-01 -2.18724489e-01
-2.58636206e-01 1.04623839e-01 -6.60838187e-01 6.08365476e-01
1.16578841e+00 6.71504587e-02 1.26894325e-01 -9.34760094e-01
-1.66964039e-01 -2.05874354e-01 -2.79355943e-01 -6.11680329e-01
-3.10671359e-01 1.54487295e-02 -1.21505044e-01 -6.65444374e-01
8.12856704e-02 -2.31101617e-01 -5.97897053e-01 4.38320041e-01
-1.54865965e-01 7.09623873e-01 3.44792575e-01 -1.97753966e-01
-2.48159781e-01 -1.97267786e-01 9.74988520e-01 6.86847493e-02
-1.00488555e+00 1.43611193e-01 -1.02562690e+00 2.14642584e-01
4.83228028e-01 2.21173450e-01 -5.25411546e-01 -5.00750780e-01
-2.42472395e-01 2.17269868e-01 1.78791985e-01 4.75269735e-01
3.04696470e-01 -1.30675459e+00 -2.03709030e+00 3.35822314e-01
3.47564429e-01 -5.54036319e-01 -1.98224500e-01 7.89874434e-01
-9.77569103e-01 5.58415830e-01 -4.70375419e-01 -2.08040625e-01
-1.26772928e+00 -3.42297047e-01 2.47257903e-01 -5.47541790e-02
-9.33812857e-02 9.62604284e-01 -2.92455047e-01 -1.33638799e+00
-3.37893777e-02 4.02021319e-01 -5.29628992e-01 3.78850579e-01
7.75842249e-01 7.74910927e-01 3.07999074e-01 -1.39509058e+00
-6.34001553e-01 3.00553232e-01 -5.09718478e-01 -7.15414762e-01
1.26565945e+00 -2.56481707e-01 -6.96141541e-01 5.16621292e-01
1.01429331e+00 4.53053921e-01 -7.91649997e-01 -4.46814060e-01
4.76231664e-01 -1.10136203e-01 -2.68376201e-01 -1.53276646e+00
-5.78618586e-01 6.77003562e-01 5.92306316e-01 1.09301366e-01
8.19401443e-01 -7.10284337e-02 8.16782236e-01 3.91401082e-01
2.59038091e-01 -1.13570428e+00 -5.65126836e-01 1.67801189e+00
8.27368855e-01 -1.37207305e+00 -6.46704078e-01 -2.53138095e-01
-9.99903142e-01 1.09336257e+00 4.63151425e-01 2.54210867e-02
9.60615337e-01 6.68153763e-02 8.94737720e-01 2.64686216e-02
-2.55981266e-01 -4.55343246e-01 4.08219516e-01 1.05131233e+00
9.78562415e-01 4.07103628e-01 -2.56978363e-01 8.14471960e-01
-4.20733720e-01 -8.14887404e-01 6.86739743e-01 6.69996202e-01
-3.17585677e-01 -9.40157592e-01 -5.14589369e-01 3.12170804e-01
-6.37152374e-01 -5.44233441e-01 -7.63993800e-01 8.73558998e-01
-3.30369651e-01 1.21782434e+00 2.55306065e-01 -4.04045284e-01
1.40271470e-01 3.82814944e-01 4.63623405e-01 -2.65551269e-01
-9.53939557e-01 -9.95877013e-02 6.81260526e-01 -4.55715507e-02
-2.26218790e-01 -1.08986688e+00 -1.21146786e+00 -6.07525885e-01
1.62802674e-02 -6.21129014e-02 5.82726538e-01 8.39349151e-01
9.58921909e-02 -2.79155821e-01 7.94345379e-01 -6.79925740e-01
-5.93274176e-01 -1.29135168e+00 -8.30385625e-01 1.54405147e-01
4.84693766e-01 -8.08016732e-02 3.10771856e-02 1.35921910e-01] | [10.17410659790039, 10.77118968963623] |
71000dac-8334-4b05-a03a-b9287aded1ef | an-efficient-probabilistically-sound | cs/9905007 | null | https://arxiv.org/abs/cs/9905007v1 | https://arxiv.org/pdf/cs/9905007v1.pdf | An Efficient, Probabilistically Sound Algorithm for Segmentation and Word Discovery | This paper presents a model-based, unsupervised algorithm for recovering word boundaries in a natural-language text from which they have been deleted. The algorithm is derived from a probability model of the source that generated the text. The fundamental structure of the model is specified abstractly so that the detailed component models of phonology, word-order, and word frequency can be replaced in a modular fashion. The model yields a language-independent, prior probability distribution on all possible sequences of all possible words over a given alphabet, based on the assumption that the input was generated by concatenating words from a fixed but unknown lexicon. The model is unusual in that it treats the generation of a complete corpus, regardless of length, as a single event in the probability space. Accordingly, the algorithm does not estimate a probability distribution on words; instead, it attempts to calculate the prior probabilities of various word sequences that could underlie the observed text. Experiments on phonemic transcripts of spontaneous speech by parents to young children suggest that this algorithm is more effective than other proposed algorithms, at least when utterance boundaries are given and the text includes a substantial number of short utterances. Keywords: Bayesian grammar induction, probability models, minimum description length (MDL), unsupervised learning, cognitive modeling, language acquisition, segmentation | ['Michael R. Brent'] | 1999-05-12 | null | null | null | null | ['language-acquisition'] | ['natural-language-processing'] | [ 6.20467603e-01 3.29657465e-01 -2.57842124e-01 -5.07807016e-01
-4.96908873e-01 -5.25240898e-01 2.90937603e-01 2.64244646e-01
-4.27680939e-01 7.19312012e-01 2.88790107e-01 -6.15276754e-01
-2.25923851e-01 -7.38954425e-01 -6.03510737e-01 -5.67759514e-01
2.20243663e-01 9.82416868e-01 2.59455711e-01 -4.29840311e-02
5.01775920e-01 2.32266992e-01 -1.69248784e+00 -9.72651616e-02
9.70256209e-01 -1.08065084e-01 8.42221677e-01 8.05322170e-01
-5.96428394e-01 3.71395439e-01 -7.63889849e-01 -4.11568671e-01
-3.39658827e-01 -8.06836963e-01 -1.06396389e+00 5.46355724e-01
3.35887750e-03 -2.69384325e-01 4.98109050e-02 1.14757609e+00
-1.13429185e-02 7.34702796e-02 1.11031497e+00 -6.30782664e-01
-4.53078210e-01 1.23263085e+00 -4.63274419e-02 1.43646210e-01
6.02435887e-01 -4.06143725e-01 9.23817098e-01 -6.68881178e-01
5.19912243e-01 1.37886488e+00 1.73424318e-01 5.28520882e-01
-1.31052601e+00 -3.57023299e-01 1.38357818e-01 -3.09829295e-01
-1.55871940e+00 -3.20070326e-01 4.93060082e-01 -4.98093247e-01
1.21568429e+00 4.90740053e-02 7.25364268e-01 8.84974837e-01
3.88630122e-01 6.59889281e-01 7.97038853e-01 -1.32715189e+00
3.42326075e-01 5.68108000e-02 4.47244257e-01 6.54806852e-01
5.59241951e-01 6.66224882e-02 -6.10425055e-01 -1.80283666e-01
6.19075894e-01 -6.55708194e-01 4.14670669e-02 8.91805887e-02
-7.96686411e-01 8.51589799e-01 -9.68441963e-01 5.69605350e-01
-2.36142680e-01 -1.97195396e-01 1.85801182e-02 -1.15679577e-03
3.29460979e-01 9.80787799e-02 -7.24997997e-01 -2.69684255e-01
-9.35458064e-01 2.10502669e-01 9.91145372e-01 1.20576382e+00
6.68448746e-01 7.27847964e-02 3.92225504e-01 9.00284588e-01
7.39265919e-01 7.38451242e-01 6.60066605e-01 -7.17476070e-01
2.13990614e-01 2.74818599e-01 -1.45038307e-01 -5.37246943e-01
-4.59875688e-02 1.14428245e-01 -6.53241109e-03 -2.23918155e-01
6.59476161e-01 -1.95147038e-01 -1.22594142e+00 1.99834418e+00
2.36289889e-01 -2.09780067e-01 2.70011991e-01 7.91979060e-02
6.13275886e-01 9.18659687e-01 2.63081282e-01 -7.29529202e-01
1.32932055e+00 -2.69917816e-01 -9.78661895e-01 -3.19630563e-01
5.02930224e-01 -8.83820236e-01 9.07670200e-01 7.69714713e-01
-1.20613229e+00 -4.09475327e-01 -8.46355319e-01 1.47225380e-01
-2.15798691e-01 -8.80800188e-02 6.52357340e-01 1.11808050e+00
-8.61196637e-01 3.95355791e-01 -7.72107661e-01 -3.75992298e-01
-4.20639426e-01 5.00927806e-01 -1.31754532e-01 1.66366845e-01
-1.11418188e+00 6.44197702e-01 9.73929822e-01 -1.28735468e-01
-6.22254372e-01 1.42229781e-01 -9.66573358e-01 -1.19011253e-02
1.43721074e-01 -2.30356976e-01 1.44683433e+00 -9.44462419e-01
-1.58035529e+00 8.74861240e-01 -7.10301638e-01 -2.61927433e-02
-2.38046050e-01 -2.76131667e-02 -3.93282324e-01 2.28690654e-01
9.08697024e-02 4.32936728e-01 9.73719895e-01 -1.02057970e+00
-7.17042029e-01 -2.12674826e-01 -6.13462806e-01 1.65648907e-01
1.36539429e-01 3.05840105e-01 -3.82457912e-01 -7.17677534e-01
5.48275292e-01 -7.59910941e-01 -6.96677342e-02 -1.10466290e+00
-2.34064609e-01 -6.93036377e-01 -2.76470929e-02 -8.30623686e-01
1.27059472e+00 -2.13866282e+00 2.00759619e-02 3.17899406e-01
-2.78079927e-01 -7.52291530e-02 1.18970238e-01 6.35627151e-01
-1.20554201e-01 2.93859810e-01 -6.12788796e-01 -7.29339942e-02
-5.11614941e-02 7.37196505e-01 -3.92347425e-01 2.24496216e-01
1.22570939e-01 4.40025717e-01 -7.91570425e-01 -7.21291542e-01
8.56548026e-02 2.20866486e-01 -4.15225238e-01 1.51054159e-01
-4.89910096e-01 1.09102048e-01 -3.41808289e-01 3.39743823e-01
4.01086122e-01 3.52700144e-01 4.49441224e-01 7.49831498e-01
-2.62371302e-01 7.69854009e-01 -1.27669179e+00 1.51207221e+00
-2.62307853e-01 5.26629269e-01 -3.55911523e-01 -8.83562386e-01
9.62565362e-01 6.76108837e-01 -5.44583872e-02 2.40200832e-02
3.98401409e-01 3.43680829e-01 5.17126203e-01 -6.92843735e-01
5.30494034e-01 -2.61769325e-01 -3.10047060e-01 6.15037918e-01
5.17852783e-01 -4.41602170e-01 6.69852555e-01 8.34140033e-02
6.73784435e-01 1.74204126e-01 5.64707041e-01 -2.48300970e-01
2.50204355e-01 -1.07467420e-01 6.02849603e-01 9.20466959e-01
2.62124568e-01 4.83496070e-01 4.41090673e-01 1.75100043e-01
-8.68934810e-01 -1.27415454e+00 -4.46663588e-01 1.28672087e+00
-1.26249909e-01 -2.41395324e-01 -1.18066120e+00 -3.08527738e-01
-5.71803093e-01 1.48040843e+00 -1.33424923e-01 6.82817996e-02
-6.48973763e-01 -7.99347997e-01 4.36080992e-01 1.97288454e-01
-2.22962230e-01 -1.59741473e+00 -3.42636347e-01 6.18023396e-01
-2.71530330e-01 -9.64566350e-01 -3.25994402e-01 3.26648891e-01
-1.07478189e+00 -7.42113411e-01 -2.13674769e-01 -1.29071438e+00
9.21156764e-01 -2.63986528e-01 7.87170470e-01 8.21734145e-02
-9.03212130e-02 5.82105577e-01 -4.61032450e-01 -6.96900010e-01
-9.38900709e-01 -1.34140387e-01 1.48744777e-01 -2.63219386e-01
8.78810048e-01 -5.74655831e-01 3.82857770e-01 -2.04820737e-01
-8.90419900e-01 -2.54048079e-01 5.09811401e-01 6.32201195e-01
4.95247632e-01 3.86755615e-01 4.43769425e-01 -9.64162767e-01
5.48938215e-01 -3.93722355e-01 -6.07939243e-01 2.27152064e-01
-4.56889540e-01 1.46647826e-01 3.64042997e-01 -5.50441265e-01
-1.43751836e+00 1.60384193e-01 -4.78588909e-01 5.54465830e-01
-1.02434289e+00 3.84838283e-01 -5.47152996e-01 6.16336226e-01
3.92941117e-01 8.25566888e-01 -1.30005151e-01 -6.16404235e-01
3.28687489e-01 8.13123286e-01 5.77464402e-01 -7.24143803e-01
4.06708837e-01 -2.03049973e-01 -3.87113154e-01 -1.38964367e+00
-4.34348792e-01 -4.54915434e-01 -1.01673222e+00 -3.39239947e-02
8.54507148e-01 -5.75732529e-01 -3.07849050e-01 5.51900089e-01
-1.46567285e+00 -9.96471792e-02 -2.06428319e-01 1.10189545e+00
-7.39422143e-01 5.52564383e-01 -4.23199117e-01 -1.17864799e+00
4.58679348e-02 -9.26744223e-01 7.07763374e-01 3.13550651e-01
-7.26034105e-01 -1.10956419e+00 1.33936152e-01 2.13653862e-01
-3.74603420e-01 -4.01226580e-01 1.45261705e+00 -1.21794391e+00
-2.60596246e-01 -8.38396326e-02 5.48636794e-01 5.46031654e-01
4.66727942e-01 5.68843961e-01 -6.33521676e-01 5.79598583e-02
3.10526639e-01 -9.29051638e-02 6.09347343e-01 4.95152831e-01
5.79287648e-01 -3.22395444e-01 -2.56232589e-01 9.92697477e-02
1.23268592e+00 8.39537680e-01 4.05688763e-01 -2.39711747e-01
3.90223473e-01 9.84133244e-01 3.44894439e-01 9.91205648e-02
2.67409891e-01 -1.21355720e-01 -7.02677527e-04 4.01122868e-01
9.40242261e-02 -3.20668489e-01 4.83115852e-01 1.52767336e+00
1.77917302e-01 -6.42171383e-01 -1.05990732e+00 1.12749696e+00
-1.24755383e+00 -8.12699258e-01 -3.22543561e-01 2.26409912e+00
1.15384710e+00 5.26269317e-01 1.66992899e-02 4.73015934e-01
1.03865778e+00 -1.76828340e-01 -4.62136902e-02 -9.72972810e-01
1.46558965e-02 5.35685480e-01 2.95830786e-01 1.06483257e+00
-5.74606419e-01 1.20350146e+00 8.01158905e+00 8.60264003e-01
-5.29075027e-01 -1.41870067e-01 2.12315962e-01 4.54157263e-01
-6.13821685e-01 2.52856493e-01 -1.22416675e+00 5.53446770e-01
1.30998445e+00 -3.53945851e-01 2.51578093e-01 5.14725327e-01
2.00759828e-01 -6.45654976e-01 -1.03311992e+00 3.90244186e-01
2.27166563e-01 -6.75811291e-01 3.20048213e-01 5.19444123e-02
4.20622170e-01 -2.89510459e-01 -3.00600171e-01 1.79868657e-02
4.18324590e-01 -8.27050924e-01 8.25709522e-01 2.58615911e-01
5.99551320e-01 -8.74144137e-01 5.25225222e-01 8.87309670e-01
-6.71224535e-01 3.87317091e-01 -4.28078026e-01 -4.00788456e-01
1.31033108e-01 4.03852493e-01 -1.07356560e+00 1.05664462e-01
8.50884393e-02 -1.84969351e-01 -2.05886334e-01 8.41032684e-01
-7.36521184e-01 1.41360319e+00 -4.67329830e-01 -4.06493753e-01
4.68586050e-02 -3.67569089e-01 7.02984869e-01 1.48479879e+00
2.69194782e-01 4.65650499e-01 -4.14794311e-02 5.43413043e-01
3.51042926e-01 6.08888268e-01 -6.17985606e-01 -2.76083261e-01
7.32252538e-01 6.41234338e-01 -1.28448868e+00 -3.20897549e-01
-4.81551170e-01 5.82907915e-01 -2.86065228e-02 3.17860037e-01
-3.23882908e-01 -5.54798901e-01 1.62515059e-01 -4.41062450e-02
5.36976755e-01 -3.46915454e-01 -2.38668948e-01 -6.13270223e-01
-1.48733258e-01 -6.65818512e-01 1.12252831e-01 -4.01088119e-01
-1.04176223e+00 4.54695940e-01 5.63313603e-01 -4.39315706e-01
-8.47444892e-01 -5.57092547e-01 -5.90952277e-01 1.12053859e+00
-7.28824377e-01 -4.88746047e-01 6.93907619e-01 2.27955714e-01
1.00272012e+00 -3.85410070e-01 1.05116999e+00 -4.00101006e-01
-3.83999050e-01 3.68797421e-01 1.21794753e-01 1.88975900e-01
1.10340178e-01 -1.46692991e+00 4.28638548e-01 1.05889094e+00
4.99518126e-01 8.33456635e-01 8.83920372e-01 -1.11053157e+00
-7.87280440e-01 -5.65464497e-01 1.67089832e+00 -3.44671309e-01
5.24163723e-01 -5.91471136e-01 -1.05206287e+00 7.38066494e-01
3.03538591e-01 -7.74448872e-01 9.93357360e-01 1.44511391e-03
1.94881439e-01 3.32131416e-01 -9.03464496e-01 4.96036649e-01
7.20013320e-01 -2.73554265e-01 -1.54697144e+00 3.50999534e-01
9.27569807e-01 -1.77523389e-01 -5.28715134e-01 -1.28231104e-02
4.74709243e-01 -6.92521751e-01 3.30498964e-01 -4.82839376e-01
1.45602480e-01 -7.88526982e-02 -3.79364602e-02 -1.04864573e+00
-2.71301389e-01 -8.46475720e-01 2.87096322e-01 1.50296581e+00
6.93288267e-01 -6.39190435e-01 5.96212447e-01 4.15054202e-01
-1.74484342e-01 -2.45642021e-01 -9.02501583e-01 -6.41665816e-01
1.54655039e-01 -8.50919247e-01 5.80618441e-01 5.79155803e-01
1.31101951e-01 5.06738305e-01 -1.13081381e-01 2.57359773e-01
7.22705662e-01 -1.83747336e-01 9.62067395e-02 -1.28233254e+00
-3.88643891e-01 -1.47140473e-01 -1.31236643e-01 -1.16784406e+00
4.05310929e-01 -5.70430458e-01 7.05471992e-01 -1.36145961e+00
-1.06119581e-01 -1.92604557e-01 1.51802689e-01 2.48905063e-01
-1.38094500e-01 -4.50833678e-01 -2.32277244e-01 -4.71527390e-02
5.10351881e-02 2.39064738e-01 7.97953665e-01 2.34181210e-01
-4.42302942e-01 2.86207139e-01 -5.34100235e-01 1.29941380e+00
8.19991708e-01 -1.06092191e+00 -7.44641483e-01 -3.96872431e-01
1.06694676e-01 3.10142130e-01 -3.54212195e-01 -4.49946672e-01
1.77263230e-01 -4.14360732e-01 2.03190595e-01 -8.70320737e-01
-9.93262231e-02 -5.30829787e-01 4.81125340e-02 2.77406156e-01
-2.91794270e-01 -6.90946504e-02 7.04555660e-02 3.23019743e-01
-1.03535384e-01 -1.26789153e+00 5.32855213e-01 -2.89230973e-01
-6.68803811e-01 -2.05908418e-01 -1.40305042e+00 1.03414841e-01
7.45832443e-01 -5.32863021e-01 3.67941439e-01 -1.74880624e-01
-9.68535662e-01 -3.09979439e-01 2.23641381e-01 2.01904833e-01
6.43829167e-01 -8.11183453e-01 -7.55133450e-01 4.65027988e-01
-2.84605205e-01 1.09940320e-01 -1.83414891e-01 1.40477464e-01
-7.41712034e-01 5.43544412e-01 1.05814077e-01 -4.00961816e-01
-1.10442102e+00 4.17493463e-01 -2.20963895e-01 1.84066698e-01
-3.59080404e-01 9.77263212e-01 9.49898213e-02 -2.23940879e-01
4.23922300e-01 -4.23637778e-01 -3.99847388e-01 -8.67290571e-02
4.10972327e-01 1.64455757e-01 -2.41739288e-01 -8.85395944e-01
-9.92303267e-02 4.87877280e-01 -2.26931289e-01 -5.66973388e-01
9.62969363e-01 -4.39726800e-01 -4.15356964e-01 1.11189067e+00
6.86487079e-01 4.35183972e-01 -6.04884565e-01 -8.66432935e-02
2.61093020e-01 -1.29635915e-01 -2.75958657e-01 -4.75545675e-01
-2.23430350e-01 5.98808765e-01 5.98845854e-02 4.13132638e-01
9.83475029e-01 2.71158010e-01 5.44626892e-01 2.69387335e-01
1.83806092e-01 -1.24655902e+00 -3.38696092e-01 8.42074394e-01
4.71969962e-01 -8.00834775e-01 -1.59270614e-01 -7.97877550e-01
-3.22124958e-01 1.28284323e+00 3.00476015e-01 -1.21361297e-03
6.91966951e-01 2.72714227e-01 -2.15652227e-01 9.46043134e-02
-7.44795263e-01 -3.44902009e-01 1.43793911e-01 7.56127000e-01
6.63128674e-01 1.42775387e-01 -9.73443627e-01 5.66514194e-01
-8.28299820e-01 -4.46153879e-01 9.03809905e-01 9.32817459e-01
-9.54183340e-01 -1.42467308e+00 -6.23856544e-01 2.91689306e-01
-7.66812086e-01 -3.74150693e-01 -2.86545992e-01 7.82253206e-01
5.56769252e-01 1.13101411e+00 3.10156524e-01 5.97919747e-02
-6.58927187e-02 5.69907427e-01 7.56276786e-01 -1.22928989e+00
-4.13135439e-02 5.73697209e-01 2.27197617e-01 2.77478784e-01
-2.70841449e-01 -1.14654398e+00 -1.57361662e+00 2.40398690e-01
-5.55037737e-01 6.37666106e-01 8.28926265e-01 1.46927845e+00
-5.97873926e-01 1.25789747e-01 4.60625261e-01 -2.93865442e-01
-3.02517295e-01 -1.08244348e+00 -8.98874998e-01 -1.26705185e-01
-1.59950182e-02 -3.45239729e-01 -5.25269508e-01 4.99722242e-01] | [10.50473403930664, 9.693554878234863] |
d098a2ac-b140-4d13-bea0-00ca21069f2b | when-geometric-deep-learning-meets-pretrained | 2212.03447 | null | https://arxiv.org/abs/2212.03447v1 | https://arxiv.org/pdf/2212.03447v1.pdf | When Geometric Deep Learning Meets Pretrained Protein Language Models | Geometric deep learning has recently achieved great success in non-Euclidean domains, and learning on 3D structures of large biomolecules is emerging as a distinct research area. However, its efficacy is largely constrained due to the limited quantity of structural data. Meanwhile, protein language models trained on substantial 1D sequences have shown burgeoning capabilities with scale in a broad range of applications. Nevertheless, no preceding studies consider combining these different protein modalities to promote the representation power of geometric neural networks. To address this gap, we make the foremost step to integrate the knowledge learned by well-trained protein language models into several state-of-the-art geometric networks. Experiments are evaluated on a variety of protein representation learning benchmarks, including protein-protein interface prediction, model quality assessment, protein-protein rigid-body docking, and binding affinity prediction, leading to an overall improvement of 20% over baselines and the new state-of-the-art performance. Strong evidence indicates that the incorporation of protein language models' knowledge enhances geometric networks' capacity by a significant margin and can be generalized to complex tasks. | ['Jinbo Xu', 'Dragomir Radev', 'Yu Tao', 'Fang Wu'] | 2022-12-07 | null | null | null | null | ['protein-interface-prediction'] | ['miscellaneous'] | [ 1.65736869e-01 1.95540085e-01 -2.93095678e-01 -4.57215965e-01
-9.00107145e-01 -5.84409535e-01 4.53243971e-01 5.67765594e-01
-4.39064533e-01 6.97561145e-01 1.71512589e-01 -5.62422395e-01
4.93479483e-02 -5.28502285e-01 -1.08149230e+00 -8.24955761e-01
-1.88597456e-01 7.96084881e-01 1.25869825e-01 -3.28154892e-01
1.54106066e-01 9.37734425e-01 -7.39564955e-01 2.62423575e-01
6.36569202e-01 7.63850152e-01 1.60046667e-01 1.40095457e-01
-2.52041668e-01 3.98627400e-01 -1.34229586e-01 -2.54892617e-01
4.18084785e-02 -4.83230203e-02 -8.59985888e-01 -4.76504028e-01
4.67244804e-01 -2.99399439e-02 -3.23423445e-01 7.14909911e-01
6.98774517e-01 7.87344053e-02 6.70953929e-01 -3.48047227e-01
-1.01886010e+00 5.64751439e-02 -6.97890937e-01 5.22502996e-02
3.40738684e-01 3.06108445e-01 1.06653440e+00 -1.09686756e+00
9.97883976e-01 1.26529598e+00 8.10903370e-01 4.10342872e-01
-1.65190971e+00 -4.69139725e-01 2.77817756e-01 5.88883609e-02
-1.35393643e+00 -1.90242469e-01 5.21575093e-01 -4.55582559e-01
1.77089524e+00 -4.86395270e-01 3.63817930e-01 1.04355288e+00
5.15569329e-01 5.01639724e-01 7.11330235e-01 -1.69968754e-01
-2.11527254e-02 -5.81929564e-01 5.44282161e-02 9.55885887e-01
3.28296602e-01 -5.40828593e-02 -4.92236793e-01 -3.22244257e-01
7.65864551e-01 1.25356793e-01 -1.27125651e-01 -9.36231911e-01
-1.00715518e+00 8.87915432e-01 8.45997572e-01 1.83056071e-01
-3.51041079e-01 1.42245498e-02 5.42965889e-01 9.30651426e-02
6.80595756e-01 7.70500541e-01 -7.24777341e-01 -7.95618594e-02
-6.37851298e-01 6.31331980e-01 5.57009459e-01 6.57238305e-01
5.28097928e-01 -1.81684390e-01 3.69469792e-01 6.98870838e-01
3.71951818e-01 3.92643958e-01 2.27603614e-02 -4.18095887e-01
4.64837104e-01 7.31688499e-01 -1.57455204e-03 -8.61183584e-01
-8.85323942e-01 -3.07937354e-01 -8.31629992e-01 3.68159324e-01
7.43728518e-01 1.61766753e-01 -9.07244027e-01 1.91226935e+00
2.12011620e-01 -2.13872030e-01 -1.70612969e-02 9.01432097e-01
8.04857731e-01 6.88210070e-01 3.65860611e-01 3.79771218e-02
1.09737515e+00 -7.41109252e-01 -2.07890704e-01 -8.61731991e-02
1.03903222e+00 -7.27286160e-01 7.26236463e-01 3.33043575e-01
-1.14075577e+00 -5.29437423e-01 -1.16266704e+00 -4.61902678e-01
-6.61493838e-01 -2.60868222e-02 8.67752850e-01 1.72554404e-01
-9.31109786e-01 7.28366494e-01 -1.04837728e+00 -5.88838518e-01
7.25786388e-01 8.16679716e-01 -6.82302713e-01 -2.03568980e-01
-1.09850729e+00 1.33756173e+00 3.85775954e-01 -1.31990746e-01
-6.46584511e-01 -1.06077945e+00 -7.27572203e-01 -1.44212440e-01
1.51043385e-01 -6.37630403e-01 9.54761565e-01 -4.84036148e-01
-1.17726636e+00 1.14463651e+00 -3.31535906e-01 -7.15689719e-01
1.60681576e-01 -2.97125041e-01 -1.24987461e-01 1.43327728e-01
-2.76094496e-01 9.29578900e-01 8.73188376e-02 -9.27279532e-01
-2.02537887e-02 -7.32013226e-01 1.38463065e-01 3.19141239e-01
1.63611367e-01 2.69354042e-03 -1.19707987e-01 -5.08670807e-01
1.74089447e-01 -9.37363803e-01 -4.82248396e-01 2.85023838e-01
-7.93261081e-02 -3.85305703e-01 5.78845620e-01 -4.81983334e-01
6.65424705e-01 -1.85906374e+00 4.98144180e-01 1.31409153e-01
4.65531707e-01 5.38478911e-01 -2.70328283e-01 5.87916911e-01
-3.14474493e-01 2.59798951e-02 -1.53166518e-01 -1.01597473e-01
-2.06401229e-01 -7.77764395e-02 -2.38270432e-01 7.08755672e-01
5.11212766e-01 1.26807904e+00 -7.01513052e-01 -1.05582111e-01
3.70214194e-01 8.86367917e-01 -6.23629510e-01 1.11366712e-01
-5.54121494e-01 4.97484654e-01 -6.25382721e-01 4.17179048e-01
6.70901358e-01 -8.44288230e-01 3.72793704e-01 -4.61010158e-01
6.81499690e-02 5.32207370e-01 -5.30958951e-01 2.13152719e+00
3.34332958e-02 3.22531193e-01 -1.70443982e-01 -1.09751499e+00
9.99778032e-01 1.01071186e-01 7.87661910e-01 -6.50705814e-01
-1.15519732e-01 1.44697189e-01 3.83494705e-01 -2.57357240e-01
2.18879297e-01 -2.81536400e-01 2.63731211e-01 3.64717007e-01
1.45545930e-01 2.00221464e-01 -1.81537479e-01 8.44639540e-02
9.79401529e-01 7.24806070e-01 3.74797195e-01 -3.19710016e-01
4.37273741e-01 3.06164891e-01 1.93949744e-01 2.78664559e-01
-1.52043030e-01 5.22888303e-01 4.15447652e-01 -9.69072938e-01
-1.36239851e+00 -1.04515815e+00 -2.82587916e-01 1.44425464e+00
-2.80546993e-02 -4.38487023e-01 -6.63623333e-01 -4.79009926e-01
2.56569475e-01 4.80378233e-02 -5.91777444e-01 -3.12504292e-01
-9.02719557e-01 -1.20216954e+00 6.26264036e-01 5.83847404e-01
1.59584597e-01 -1.00146604e+00 -2.09395111e-01 4.67515081e-01
2.79084384e-01 -1.03228235e+00 -2.27909446e-01 4.59919572e-01
-1.04784536e+00 -1.10114574e+00 -7.64922380e-01 -9.69616115e-01
5.54807067e-01 3.81506413e-01 1.17932606e+00 -2.29451925e-01
-3.16906631e-01 -1.42170787e-01 4.11424926e-03 -3.96170795e-01
-2.24448308e-01 4.86239076e-01 1.93130419e-01 -5.83390415e-01
8.35909545e-01 -5.38556933e-01 -7.51712680e-01 1.10116109e-01
-7.54890859e-01 4.30583395e-02 6.66222930e-01 7.65781760e-01
8.44996095e-01 -6.87055469e-01 6.74262166e-01 -7.70928442e-01
6.19577885e-01 -2.53855616e-01 -4.63543355e-01 2.64727145e-01
-6.34620845e-01 5.13282716e-01 7.36941814e-01 -2.78743327e-01
-8.30459177e-01 1.86169669e-02 -3.25218678e-01 -1.86031833e-01
-2.38935426e-01 7.28359401e-01 -6.46396130e-02 -4.68556851e-01
8.73830855e-01 4.45050187e-02 4.23327148e-01 -6.03692472e-01
5.71936429e-01 -9.28915630e-04 2.75557786e-01 -9.08776045e-01
5.85255444e-01 4.26485032e-01 3.64720434e-01 -6.85215414e-01
-8.01132739e-01 -2.75098592e-01 -8.85738969e-01 4.42791164e-01
1.02810109e+00 -9.38642383e-01 -9.90207613e-01 2.74002969e-01
-1.24362016e+00 -2.86861837e-01 1.62885368e-01 4.06610280e-01
-4.24518526e-01 5.79466820e-01 -7.16622591e-01 -1.46673962e-01
-5.19635677e-01 -1.51182997e+00 1.21065092e+00 -2.05544621e-01
-3.07763189e-01 -1.09734845e+00 2.12426275e-01 2.71280646e-01
4.17100519e-01 3.51166487e-01 1.36212504e+00 -9.17720616e-01
-7.54915833e-01 -1.34944007e-01 -4.93785292e-01 -3.50507274e-02
4.01347280e-02 -3.49824041e-01 -6.58311784e-01 -4.88392770e-01
-4.85577464e-01 -7.52825499e-01 9.05860126e-01 5.28603375e-01
9.83673453e-01 2.13610694e-01 -5.19157648e-01 6.60939455e-01
1.31075907e+00 7.07387328e-02 3.87754083e-01 2.20931515e-01
9.79361534e-01 4.41642195e-01 2.37772927e-01 3.17757986e-02
1.69627413e-01 8.42400968e-01 5.57127774e-01 -3.96827906e-01
6.81768134e-02 -2.68148273e-01 1.32527724e-01 5.69472790e-01
-4.48727340e-01 -3.94881591e-02 -1.19138181e+00 -4.56262156e-02
-1.75529957e+00 -9.25952137e-01 2.02386603e-01 2.11603618e+00
8.13671529e-01 4.31892425e-01 8.05377886e-02 -5.75520873e-01
2.74869055e-01 4.27354366e-01 -1.01851714e+00 -2.28601068e-01
-3.15145314e-01 4.40577447e-01 5.23417294e-01 3.54725540e-01
-1.07447076e+00 1.18943262e+00 6.47080135e+00 6.95663869e-01
-1.26698494e+00 -3.27487260e-01 8.70853543e-01 3.54601070e-02
-2.42154729e-02 -2.74549991e-01 -7.89765239e-01 -3.54662798e-02
8.99354696e-01 7.29283914e-02 2.74937570e-01 7.16612816e-01
4.34263527e-01 3.25656950e-01 -1.26360130e+00 1.00163126e+00
-3.30487192e-02 -1.84516573e+00 2.42970169e-01 3.23419988e-01
6.05094075e-01 6.79145694e-01 2.83220053e-01 1.91362068e-01
2.39333972e-01 -1.60480618e+00 1.55955270e-01 4.57404643e-01
8.65385234e-01 -7.68420458e-01 5.96616626e-01 4.84278724e-02
-1.15878439e+00 6.19484246e-01 -7.20635593e-01 -1.89995468e-02
1.52760595e-01 2.45114028e-01 -7.46476889e-01 4.40273136e-01
4.60007787e-01 9.62899506e-01 -3.15966249e-01 6.47691905e-01
2.77000159e-01 3.79338086e-01 -2.17282280e-01 -3.01897656e-02
6.29419804e-01 -4.93284404e-01 5.51441461e-02 1.11069310e+00
-1.74362674e-01 3.28674585e-01 3.93808335e-01 1.16278088e+00
-4.16471303e-01 3.97484183e-01 -7.25436985e-01 -2.25904286e-01
1.85494706e-01 8.57264340e-01 -4.50371861e-01 -2.87794806e-02
-7.79552400e-01 5.76388955e-01 7.37838328e-01 4.79151428e-01
-8.28481019e-01 -9.10086483e-02 1.14165652e+00 2.03943670e-01
1.25287056e-01 -6.64919078e-01 -1.92274794e-01 -8.38865399e-01
-1.58328906e-01 -9.67326343e-01 1.98907312e-03 -6.87209547e-01
-1.35278642e+00 2.22585320e-01 -2.87384927e-01 -5.90511799e-01
2.01221123e-01 -1.28951228e+00 -1.56196624e-01 1.18387842e+00
-1.42904747e+00 -1.24982536e+00 9.89379287e-02 1.48156077e-01
2.55809069e-01 -3.73573244e-01 1.13274646e+00 2.40938470e-01
-4.40906793e-01 6.07893348e-01 5.25995016e-01 -7.62204593e-03
9.36117113e-01 -1.02303839e+00 7.63406694e-01 2.37887129e-01
-2.94764731e-02 1.09237123e+00 6.14322245e-01 -5.54829359e-01
-1.68951595e+00 -1.13598573e+00 8.91524971e-01 -6.82506919e-01
5.71638167e-01 -3.70166987e-01 -1.36753201e+00 6.53249443e-01
-7.77585804e-02 1.48494929e-01 7.83930123e-01 3.25603724e-01
-7.36831963e-01 1.75284356e-01 -9.01917636e-01 5.18863559e-01
1.24754775e+00 -7.74465322e-01 -4.78528529e-01 5.90781987e-01
8.10900271e-01 -7.02376366e-01 -1.21429217e+00 7.54365265e-01
6.00690186e-01 -6.96654558e-01 1.47729933e+00 -1.29078090e+00
4.28451359e-01 -1.45076990e-01 -8.99781957e-02 -8.89945567e-01
-6.01866543e-01 -4.76533651e-01 -2.31303647e-02 4.97692078e-01
6.80488944e-01 -4.91057843e-01 1.06683719e+00 5.02529383e-01
-3.51354301e-01 -1.27603984e+00 -7.36009538e-01 -4.54730958e-01
8.22978556e-01 -8.90084207e-02 4.31982338e-01 7.98417747e-01
1.89225897e-01 7.21893251e-01 -1.39462247e-01 -1.72587514e-01
3.84422362e-01 1.34455770e-01 8.03201437e-01 -1.30511868e+00
-1.64891869e-01 -6.19611442e-01 -4.49456036e-01 -1.70673645e+00
3.63961041e-01 -1.28346169e+00 -3.00514936e-01 -1.64493883e+00
4.54444468e-01 -1.84197217e-01 -4.68429983e-01 3.90494645e-01
-8.79732668e-02 1.92924365e-02 -1.35603949e-01 1.95256323e-01
-8.85397911e-01 6.12826109e-01 1.37278569e+00 -2.48679578e-01
-6.42238781e-02 -2.72819549e-01 -7.55889475e-01 5.92988610e-01
7.74407744e-01 6.08046874e-02 -3.02936763e-01 -3.81858200e-01
3.61344218e-01 -2.60570705e-01 1.92059711e-01 -7.94840932e-01
-1.02390997e-01 -9.54961404e-02 5.26295722e-01 -5.00870049e-01
3.50131005e-01 -6.36457264e-01 -7.38965720e-02 3.90744716e-01
-4.78643835e-01 2.71497548e-01 4.47634041e-01 7.21551955e-01
9.29003283e-02 5.34291267e-01 1.04916000e+00 -1.03871152e-01
-3.74416262e-01 6.81620777e-01 -1.79537028e-01 1.88450962e-01
6.80128336e-01 -2.37019330e-01 -4.12339181e-01 2.41304338e-02
-7.67796934e-01 -2.14062363e-01 8.79474819e-01 3.18577796e-01
5.33898652e-01 -9.90890324e-01 -3.78451228e-01 1.09225392e-01
3.01861495e-01 1.52409479e-01 1.78340096e-02 6.43147290e-01
-8.72354388e-01 1.07953238e+00 -1.90664425e-01 -6.46861136e-01
-1.12766123e+00 7.17603326e-01 5.07064998e-01 -5.07358253e-01
-6.58280432e-01 6.66110396e-01 5.40043831e-01 -5.09358108e-01
2.34850988e-01 -5.66510141e-01 4.36170623e-02 -4.55709368e-01
2.14765549e-01 7.55457729e-02 2.08480835e-01 -9.32921946e-01
-4.25677806e-01 7.90665388e-01 -4.78298485e-01 5.95931411e-01
1.48267841e+00 3.40101689e-01 4.90422435e-02 1.17686339e-01
1.43834329e+00 -6.10102236e-01 -1.45805383e+00 -6.68933928e-01
1.62672326e-01 2.21652269e-01 -4.54804778e-01 -7.86258101e-01
-6.01413131e-01 1.14483726e+00 6.08531654e-01 -5.22705615e-01
4.35355544e-01 1.99957490e-01 7.66034365e-01 9.13022935e-01
5.75429678e-01 -6.63785338e-01 1.93262562e-01 9.09422278e-01
7.62865782e-01 -1.58908403e+00 2.55293310e-01 -1.48968250e-01
-3.28807682e-01 1.12318707e+00 5.14290750e-01 -2.67346889e-01
6.65612817e-01 4.45105396e-02 -4.26575057e-02 -5.98043978e-01
-5.61701357e-01 2.07760911e-02 4.57549989e-01 6.16839767e-01
1.16714501e+00 -8.70849937e-02 -7.47632459e-02 3.35401088e-01
5.99139482e-02 -2.14415804e-01 -2.58761525e-01 8.00056458e-01
-8.23444664e-01 -1.26674163e+00 2.05939170e-02 2.78529137e-01
-7.65066743e-01 -3.03063661e-01 -6.50928319e-01 8.55726898e-01
-1.33534297e-01 4.44745988e-01 -2.53299505e-01 4.24203947e-02
4.09956127e-01 3.15759510e-01 7.61402428e-01 -5.14168322e-01
-3.89052033e-01 -6.29423857e-02 -1.39393538e-01 -6.27847493e-01
-5.77997506e-01 -3.87527257e-01 -1.70865071e+00 -4.65745896e-01
-7.87703022e-02 3.17375027e-02 3.64745498e-01 1.01979041e+00
8.23767662e-01 3.12397510e-01 -1.42291427e-01 -1.02750885e+00
-6.58867776e-01 -8.64092410e-01 -2.33101770e-01 5.87705970e-01
1.20389216e-01 -7.58438528e-01 2.87623852e-01 -6.08222708e-02] | [4.9024858474731445, 5.697126865386963] |
24f29e20-5e06-4d30-a7db-8bd6becefcac | hsolo-homography-from-a-single-affine-aware | 2009.05004 | null | https://arxiv.org/abs/2009.05004v1 | https://arxiv.org/pdf/2009.05004v1.pdf | HSolo: Homography from a single affine aware correspondence | The performance of existing robust homography estimation algorithms is highly dependent on the inlier rate of feature point correspondences. In this paper, we present a novel procedure for homography estimation that is particularly well suited for inlier-poor domains. By utilizing the scale and rotation byproducts created by affine aware feature detectors such as SIFT and SURF, we obtain an initial homography estimate from a single correspondence pair. This estimate allows us to filter the correspondences to an inlier-rich subset for use with a robust estimator. Especially at low inlier rates, our novel algorithm provides dramatic performance improvements. | ['Tony Perkins', 'Cara Monical', 'Antonio Gonzales'] | 2020-09-10 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-1.06869027e-01 -3.79813403e-01 -2.11188048e-01 -3.45606320e-02
-1.05632675e+00 -8.18460226e-01 5.62623799e-01 8.05007070e-02
-2.25899890e-01 7.47923017e-01 1.36510059e-01 3.93890530e-01
-3.13589685e-02 -6.80076838e-01 -6.26181066e-01 -3.42107445e-01
2.45146349e-01 6.56007528e-01 2.40540192e-01 -2.89631605e-01
5.77998161e-01 1.08299530e+00 -1.21197581e+00 -3.32345605e-01
5.59332073e-01 6.91457331e-01 -1.95085987e-01 8.03634942e-01
4.06353801e-01 -5.55731682e-03 -2.00964779e-01 -2.24014521e-01
8.12353015e-01 -3.71453881e-01 -2.25200608e-01 3.96409243e-01
9.35150087e-01 -6.62953079e-01 -6.10432863e-01 1.07660389e+00
3.29088688e-01 1.64305776e-01 5.96273601e-01 -8.99229646e-01
3.04801285e-01 -1.14054680e-01 -7.37978935e-01 -1.22215018e-01
9.13846195e-01 -1.56937867e-01 1.00806475e+00 -1.49184239e+00
1.01292169e+00 1.11048174e+00 1.00653362e+00 -2.10608214e-01
-1.16847503e+00 -3.75222713e-01 -4.94356245e-01 -2.10864291e-01
-1.70646870e+00 -8.63123775e-01 9.70199227e-01 -2.16501027e-01
8.14395845e-01 1.83755279e-01 6.30770326e-01 2.02308044e-01
5.59594631e-01 2.63350874e-01 5.20811617e-01 -6.21888399e-01
-9.41303149e-02 3.86670008e-02 -2.71411002e-01 8.61405313e-01
5.87749839e-01 3.06951255e-01 -5.43367505e-01 -4.13182616e-01
1.62118924e+00 8.72091278e-02 -3.22190553e-01 -1.13801789e+00
-1.51401317e+00 8.13732266e-01 5.70608854e-01 5.18687181e-02
-5.24343431e-01 3.86025250e-01 1.35608315e-01 3.14125776e-01
1.61306128e-01 5.22746444e-01 1.04464553e-01 -6.46124780e-02
-1.06133306e+00 1.50834173e-01 8.89575183e-01 1.46672571e+00
1.05599725e+00 -1.35528088e-01 4.36955243e-01 5.87657869e-01
3.50687176e-01 6.76083148e-01 2.40666531e-02 -1.10835767e+00
5.53516507e-01 3.39129925e-01 3.22063804e-01 -1.47702742e+00
-2.57750481e-01 -3.92225057e-01 -5.71727693e-01 1.15174584e-01
2.64868736e-01 2.80240364e-03 -6.53974175e-01 9.65098560e-01
5.64637363e-01 -1.35836467e-01 -8.66138712e-02 7.50009298e-01
1.82668537e-01 3.40699255e-01 -8.07187438e-01 -2.12741509e-01
9.08337355e-01 -6.85134649e-01 -6.36463225e-01 -1.61021814e-01
4.48398918e-01 -1.50069547e+00 3.28706533e-01 2.84959763e-01
-1.26474380e+00 -3.58957440e-01 -1.39621162e+00 -2.49855235e-01
8.65850449e-02 1.89931273e-01 2.12649107e-01 2.87774682e-01
-9.52255964e-01 9.39122021e-01 -8.52334619e-01 -6.02370560e-01
-2.00656578e-02 7.14852333e-01 -6.88477576e-01 6.42623156e-02
-5.47508121e-01 1.13409364e+00 1.97608873e-01 -6.66451827e-02
-2.88258642e-01 -3.95364374e-01 -9.65929568e-01 -1.78015530e-02
1.63286880e-01 -8.56139243e-01 1.15942156e+00 -3.71697605e-01
-1.25395501e+00 6.73806787e-01 -3.65818769e-01 -2.62365133e-01
8.81906211e-01 -3.86942923e-01 5.55479564e-02 5.65060019e-01
1.50267497e-01 3.06587547e-01 1.24154127e+00 -1.26136649e+00
-5.03284514e-01 -4.32508737e-01 -4.69751060e-01 5.09269118e-01
1.14337474e-01 -2.01548412e-01 -6.03008091e-01 -5.48546791e-01
8.38488400e-01 -9.55878735e-01 -2.80994892e-01 3.63930404e-01
-2.44473979e-01 4.53965873e-01 7.20612109e-01 -5.53063929e-01
9.04011011e-01 -2.13762164e+00 -9.40773934e-02 8.57189000e-01
3.04545879e-01 -7.64952451e-02 2.21095234e-01 9.29676950e-01
1.11232169e-01 -4.97568518e-01 2.20501035e-01 -1.60543934e-01
-3.09806585e-01 -8.11278895e-02 -2.59946942e-01 1.21832550e+00
-5.13335131e-02 3.94594342e-01 -8.46497118e-01 -5.86851478e-01
7.23020434e-01 5.76466918e-01 -5.85384727e-01 9.38341841e-02
5.04531205e-01 2.25440815e-01 -5.05533278e-01 7.73371577e-01
7.11512506e-01 5.69114499e-02 -1.56158358e-01 -6.51759744e-01
-4.46355581e-01 1.42789511e-02 -1.63312030e+00 1.57052732e+00
-4.29677397e-01 7.50486434e-01 1.67454988e-01 -2.01167256e-01
1.27042878e+00 2.46057495e-01 8.41246784e-01 -1.19120903e-01
4.11243945e-01 7.40995467e-01 -2.68068433e-01 2.89475292e-01
7.42022753e-01 -8.67890939e-02 1.63677663e-01 9.70458388e-02
-4.99106273e-02 -7.18670964e-01 6.10199049e-02 7.48904273e-02
9.30881262e-01 1.67289358e-02 1.09016895e+00 -2.70505667e-01
4.83791888e-01 1.83294758e-01 4.00283754e-01 5.62616289e-01
-1.62396535e-01 1.06053734e+00 -9.79880393e-02 -3.93634766e-01
-1.58193660e+00 -1.02548873e+00 -4.91817385e-01 -6.75627291e-02
4.69353825e-01 -5.58053553e-01 -3.66957456e-01 -2.25191817e-01
4.44596082e-01 4.01607007e-02 -2.18149185e-01 -8.47630873e-02
-8.83270383e-01 3.08286436e-02 2.03216504e-02 5.49757123e-01
6.13444090e-01 -1.92836210e-01 -4.49590176e-01 4.82339501e-01
2.66892940e-01 -1.01367366e+00 -8.10940146e-01 4.36647721e-02
-1.26635516e+00 -1.14334846e+00 -8.59830081e-01 -7.67382205e-01
1.08596730e+00 6.98541760e-01 7.84495294e-01 -5.73493652e-02
-1.51458785e-01 5.05488217e-01 -1.91252947e-01 3.66077498e-02
-2.14682505e-01 -2.22597837e-01 2.20355168e-01 -1.07224360e-01
1.82457194e-01 -3.81524861e-01 -6.62687242e-01 7.84122050e-01
-2.55235165e-01 -3.74245703e-01 4.33919251e-01 1.01274455e+00
8.99721444e-01 -1.40385032e-01 -1.04644053e-01 -5.53942859e-01
2.72214681e-01 8.57216269e-02 -1.14336264e+00 -2.39165630e-02
-5.09471297e-01 6.92345276e-02 5.30702531e-01 -2.08387554e-01
-8.26276362e-01 8.77598524e-01 1.26820197e-02 -7.20132947e-01
2.59198129e-01 1.27791673e-01 -4.78378199e-02 -1.08519042e+00
8.35440516e-01 -4.54955809e-02 7.62657821e-02 -4.37324971e-01
4.43390161e-01 4.15504128e-01 9.38554645e-01 -2.09597707e-01
1.34290004e+00 6.87282681e-01 4.80263770e-01 -1.14084911e+00
-1.68162480e-01 -1.10411263e+00 -1.25203562e+00 -4.42969091e-02
3.21828187e-01 -9.93034542e-01 -4.00268525e-01 4.41108756e-02
-1.15259063e+00 6.06441319e-01 -3.55392396e-01 9.95671391e-01
-1.02156532e+00 6.81241393e-01 -2.71687776e-01 -5.57332516e-01
-2.08731711e-01 -1.10430241e+00 1.13416684e+00 1.68211628e-02
-5.18184304e-01 -1.01328063e+00 3.64819765e-01 -2.06358448e-01
-7.50838071e-02 2.85793781e-01 1.77589685e-01 -2.73205787e-01
-8.28227401e-01 -9.35413420e-01 -1.42145783e-01 -8.67341980e-02
1.70824841e-01 1.33590147e-01 -5.83873272e-01 -6.01236701e-01
1.46638215e-01 9.09908712e-02 5.31532288e-01 4.32026714e-01
1.62878230e-01 -8.68857950e-02 -6.57869637e-01 1.04740691e+00
1.57198262e+00 1.82055488e-01 5.18489182e-01 5.63055575e-01
7.36796141e-01 9.13857520e-02 1.01533365e+00 5.92064559e-01
-4.25870856e-03 7.62931645e-01 8.12434703e-02 -1.25990242e-01
-1.48205385e-01 -6.10718846e-01 -7.77225494e-02 7.50616133e-01
-8.25587511e-02 1.86866775e-01 -6.80808961e-01 5.37630260e-01
-1.92086101e+00 -5.93087971e-01 -2.43362814e-01 2.59817386e+00
4.44370031e-01 -1.15775526e-01 9.54943523e-02 1.83141112e-01
8.05076361e-01 -2.97253653e-02 -4.03419167e-01 -2.19041005e-01
9.28777829e-02 -1.06228963e-01 1.18459594e+00 7.94169068e-01
-1.04519844e+00 8.57029021e-01 7.41610670e+00 2.14797974e-01
-5.17561555e-01 -4.74433959e-01 -4.40823525e-01 2.34102875e-01
-6.54144064e-02 5.45916200e-01 -1.08681417e+00 -9.58439335e-02
3.39254260e-01 -5.96540689e-01 2.37483218e-01 9.73898888e-01
-7.65124522e-03 -1.98550701e-01 -1.13689828e+00 1.49322641e+00
3.97235721e-01 -1.27758121e+00 -1.04001738e-01 2.80558884e-01
8.21305335e-01 -1.80980772e-01 -1.30690292e-01 -5.94095945e-01
1.22069716e-01 -3.04984391e-01 1.94562256e-01 2.60972321e-01
6.82686985e-01 -9.34017420e-01 6.46373928e-01 4.12932970e-02
-1.41924584e+00 2.83333749e-01 -6.95596635e-01 2.01279208e-01
4.33880717e-01 4.95906949e-01 -1.21291673e+00 6.25503004e-01
3.78444418e-02 9.82238114e-01 -2.55434930e-01 1.69073474e+00
-3.13371234e-02 -1.99486837e-01 -7.94400871e-01 3.64626139e-01
7.11975098e-02 -4.53633457e-01 8.29700887e-01 8.64157677e-01
8.15006554e-01 2.94640541e-01 2.74488121e-01 3.50935310e-01
-7.66120702e-02 2.19310731e-01 -1.20060551e+00 1.87164381e-01
6.47280753e-01 1.16635966e+00 -7.00499356e-01 -2.36791432e-01
-4.47304219e-01 1.08566666e+00 3.01832519e-02 4.63424958e-02
-2.52800912e-01 -8.38605583e-01 3.79785329e-01 2.94707865e-01
2.48737693e-01 -7.72819579e-01 -3.45325381e-01 -1.53978169e+00
-2.94354185e-02 -7.35118210e-01 2.16183677e-01 -5.18673182e-01
-8.34647834e-01 3.77397209e-01 7.87727758e-02 -1.75221586e+00
-7.16873169e-01 -3.90544385e-01 -6.42666101e-01 8.20840538e-01
-1.11680019e+00 -9.12802577e-01 -2.60023445e-01 7.27796257e-01
4.70346302e-01 -1.12105541e-01 5.03823459e-01 1.64958805e-01
4.45867609e-03 5.17090976e-01 4.50719357e-01 -1.72370195e-01
9.77966547e-01 -1.04357040e+00 6.29028857e-01 1.00179446e+00
2.17569649e-01 9.48793948e-01 1.11955738e+00 -9.88231897e-01
-1.73486531e+00 -4.18258399e-01 9.95080829e-01 -4.80968177e-01
6.46958709e-01 -3.71639356e-02 -6.30080462e-01 9.52840090e-01
-3.55920017e-01 -5.81125133e-02 1.73234902e-02 -1.75114319e-01
-6.89611882e-02 -1.63633734e-01 -1.28828371e+00 4.16086972e-01
7.89731503e-01 -5.75263500e-01 -6.43717945e-01 4.38420534e-01
-6.65976107e-03 -8.35731447e-01 -1.01121151e+00 2.72989482e-01
8.69764209e-01 -8.92046452e-01 1.23814487e+00 1.26378685e-01
-1.92983255e-01 -3.87566119e-01 -1.66921318e-01 -1.20528150e+00
-4.73166704e-01 -1.05785465e+00 1.21904805e-01 9.06879187e-01
-1.38133526e-01 -6.91803515e-01 9.34650064e-01 4.94604468e-01
5.96882664e-02 -1.38193825e-02 -9.02399838e-01 -1.03605306e+00
-4.81344849e-01 3.20625961e-01 4.52466100e-01 6.77779436e-01
9.51090977e-02 8.79497230e-02 -4.61784422e-01 2.76186585e-01
1.22540426e+00 2.94817001e-01 1.16384506e+00 -1.14351881e+00
-2.84330435e-02 7.12000066e-03 -8.95359814e-01 -1.36165512e+00
-1.71141133e-01 -4.99459833e-01 2.02919573e-01 -1.02268589e+00
-1.29753590e-01 -1.34764895e-01 2.80060649e-01 -1.52088897e-02
2.06680104e-01 4.14905131e-01 1.31689876e-01 6.79805577e-01
-2.64244899e-02 4.53417987e-01 1.14770496e+00 4.12653893e-01
-4.62316006e-01 8.34777281e-02 -4.30524908e-02 9.66789305e-01
5.47317207e-01 -3.73616934e-01 -1.17779285e-01 -2.40288645e-01
-2.04401910e-01 3.33217680e-01 1.31887391e-01 -1.01283157e+00
5.93179703e-01 -3.43789421e-02 7.96476781e-01 -1.07810521e+00
6.91611171e-01 -9.68598425e-01 3.96752059e-01 4.09170330e-01
-5.45778312e-03 5.06826520e-01 -8.02012682e-02 4.54088330e-01
-3.01962972e-01 -3.60487878e-01 9.30448890e-01 -2.54258335e-01
-5.26453793e-01 3.54611129e-01 8.10863897e-02 -4.52348679e-01
1.01114035e+00 -7.35476732e-01 -1.00126234e-03 -6.81638062e-01
-4.80824947e-01 -1.24027468e-01 1.28825617e+00 4.74874116e-02
9.09165084e-01 -1.48050880e+00 -4.42932248e-01 5.88574231e-01
1.67530477e-01 -2.26836458e-01 -2.69180596e-01 7.19724178e-01
-1.03245044e+00 5.06259024e-01 -4.40614194e-01 -6.45075798e-01
-1.43004429e+00 4.03378278e-01 1.37872621e-01 2.45671526e-01
-8.67998481e-01 4.17708188e-01 -3.51870880e-02 7.22964928e-02
3.41017507e-02 5.70610464e-02 2.40211040e-01 -2.07652166e-01
3.53867352e-01 8.31236899e-01 1.21822461e-01 -1.12437320e+00
-3.58839720e-01 1.18894649e+00 -1.10319436e-01 -3.63606900e-01
1.04162443e+00 -3.17795783e-01 1.10621214e-01 1.16744384e-01
1.63120317e+00 4.07221168e-01 -1.40822494e+00 -1.27765508e-02
-1.91540748e-01 -1.37920225e+00 2.28813231e-01 1.10066056e-01
-7.21977830e-01 6.31700337e-01 2.34591350e-01 -1.41143516e-01
6.98409796e-01 -2.10977733e-01 8.05073380e-01 4.62126970e-01
5.57113826e-01 -9.64018106e-01 -2.05324039e-01 3.06385458e-01
1.06417215e+00 -1.22024226e+00 7.56919026e-01 -7.84275115e-01
-9.68644768e-02 1.55888999e+00 1.15050234e-01 -8.44872534e-01
5.21175623e-01 1.39146224e-01 1.99241308e-03 -2.33391985e-01
-1.89527959e-01 -1.17613979e-01 4.84449446e-01 5.93538940e-01
2.48525321e-01 -4.23320353e-01 -4.00540382e-01 -6.46157920e-01
-3.16516399e-01 -2.36847252e-01 7.41712272e-01 1.13022602e+00
-8.54046524e-01 -1.31195760e+00 -9.13197696e-01 2.52401888e-01
7.81074911e-02 2.25496277e-01 -4.80314583e-01 1.02704227e+00
-7.31982112e-01 6.85629964e-01 2.38729380e-02 -1.79588020e-01
6.48409188e-01 -3.95338833e-01 7.47660100e-01 -4.08835620e-01
-4.35055256e-01 7.29010582e-01 2.19582736e-01 -9.12357867e-01
-9.62565094e-02 -9.44291592e-01 -1.12817514e+00 -4.11578834e-01
-6.07964277e-01 -3.13600339e-02 7.40121663e-01 4.79792297e-01
2.14972511e-01 -3.63407373e-01 9.71170127e-01 -1.13706493e+00
-8.21316481e-01 -2.48176098e-01 -7.67828763e-01 3.36530745e-01
6.17475986e-01 -8.36520016e-01 -6.11780047e-01 -1.07092746e-01] | [7.8930134773254395, -2.3348515033721924] |
925c6212-c1b9-480e-9639-20b969021eaa | polysemous-visual-semantic-embedding-for-1 | 1906.04402 | null | https://arxiv.org/abs/1906.04402v2 | https://arxiv.org/pdf/1906.04402v2.pdf | Polysemous Visual-Semantic Embedding for Cross-Modal Retrieval | Visual-semantic embedding aims to find a shared latent space where related visual and textual instances are close to each other. Most current methods learn injective embedding functions that map an instance to a single point in the shared space. Unfortunately, injective embedding cannot effectively handle polysemous instances with multiple possible meanings; at best, it would find an average representation of different meanings. This hinders its use in real-world scenarios where individual instances and their cross-modal associations are often ambiguous. In this work, we introduce Polysemous Instance Embedding Networks (PIE-Nets) that compute multiple and diverse representations of an instance by combining global context with locally-guided features via multi-head self-attention and residual learning. To learn visual-semantic embedding, we tie-up two PIE-Nets and optimize them jointly in the multiple instance learning framework. Most existing work on cross-modal retrieval focuses on image-text data. Here, we also tackle a more challenging case of video-text retrieval. To facilitate further research in video-text retrieval, we release a new dataset of 50K video-sentence pairs collected from social media, dubbed MRW (my reaction when). We demonstrate our approach on both image-text and video-text retrieval scenarios using MS-COCO, TGIF, and our new MRW dataset. | ['Yale Song', 'Mohammad Soleymani'] | 2019-06-11 | polysemous-visual-semantic-embedding-for | http://openaccess.thecvf.com/content_CVPR_2019/html/Song_Polysemous_Visual-Semantic_Embedding_for_Cross-Modal_Retrieval_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Song_Polysemous_Visual-Semantic_Embedding_for_Cross-Modal_Retrieval_CVPR_2019_paper.pdf | cvpr-2019-6 | ['video-text-retrieval'] | ['computer-vision'] | [ 1.93350241e-01 -3.61131430e-01 -4.67729837e-01 -3.65638554e-01
-1.12967479e+00 -5.68141639e-01 8.57762814e-01 1.01193063e-01
-4.02699560e-01 3.87752950e-01 4.93884474e-01 6.76600412e-02
-3.78572762e-01 -4.10209805e-01 -6.56957746e-01 -6.07568800e-01
3.79656963e-02 4.18548822e-01 -1.80013478e-01 -1.09711565e-01
1.66726738e-01 -1.61478482e-02 -1.54422355e+00 9.44585204e-01
3.40735734e-01 7.97956288e-01 3.77779067e-01 6.33334756e-01
-2.89834529e-01 7.52628624e-01 -6.01743817e-01 -5.70412695e-01
8.86025280e-02 -1.96986735e-01 -8.54993582e-01 4.00535725e-02
1.04118955e+00 -2.31643915e-01 -6.51531339e-01 8.78546417e-01
5.44543564e-01 2.33720362e-01 9.15186048e-01 -1.59789062e+00
-1.31482959e+00 4.05817837e-01 -6.07965469e-01 2.39935488e-01
4.69211608e-01 -5.65283466e-03 1.51059639e+00 -1.13284862e+00
8.41712236e-01 1.55417287e+00 3.58373046e-01 4.38328892e-01
-1.21802652e+00 -4.96933281e-01 3.47922951e-01 6.26302421e-01
-1.49079919e+00 -2.31777892e-01 7.68283486e-01 -3.59250009e-01
1.07049537e+00 3.22016746e-01 6.80090487e-01 1.55156112e+00
8.46568197e-02 1.23532271e+00 8.21274996e-01 -1.37648076e-01
-2.21822113e-01 5.66911846e-02 1.14384070e-01 4.64653730e-01
-9.79171693e-03 -2.65549779e-01 -8.28386962e-01 -5.44769056e-02
5.37137151e-01 5.03104806e-01 -3.85803044e-01 -5.67283034e-01
-1.55102766e+00 8.07412148e-01 6.08559132e-01 5.85187912e-01
-1.62406564e-01 4.90066916e-01 4.50198501e-01 6.17085636e-01
4.52576041e-01 3.59888494e-01 -1.68245032e-01 8.31054989e-03
-8.68888378e-01 2.60692686e-01 5.10577261e-01 1.05200875e+00
8.18092942e-01 -3.21959317e-01 -3.81785125e-01 1.07542717e+00
4.05816346e-01 5.59054732e-01 7.23956108e-01 -6.67278111e-01
6.05278313e-01 5.26539683e-01 -1.33604214e-01 -1.38542652e+00
2.25786231e-02 1.52570317e-02 -5.65418303e-01 -4.32779938e-01
-3.53980623e-02 3.98615122e-01 -9.24596429e-01 1.74769163e+00
-6.80287853e-02 3.44398826e-01 2.38600463e-01 1.11198711e+00
1.09523177e+00 9.65467811e-01 1.65730834e-01 7.24249557e-02
1.40427446e+00 -1.26653695e+00 -8.94178748e-01 -4.68566358e-01
4.69632417e-01 -7.67246485e-01 1.31732023e+00 -7.89077282e-02
-1.04315615e+00 -3.60530496e-01 -9.83835876e-01 -5.49150467e-01
-8.19466949e-01 -6.98188366e-03 3.68529558e-01 -7.89097790e-03
-1.07698226e+00 2.64091820e-01 -4.16305363e-01 -5.13971925e-01
2.69863367e-01 1.49297789e-01 -6.87521040e-01 -4.36109334e-01
-1.32299304e+00 6.76974058e-01 2.90986985e-01 -4.37830463e-02
-9.41074193e-01 -5.16450226e-01 -9.37561929e-01 1.01925455e-01
6.41129851e-01 -8.42028260e-01 7.58034348e-01 -1.27414024e+00
-9.26601410e-01 1.20674074e+00 -2.59613454e-01 -7.29580596e-02
1.71251297e-01 -4.16227609e-01 -5.02681315e-01 4.49843615e-01
3.89897376e-01 9.04883325e-01 1.34665835e+00 -1.44770634e+00
-8.46745670e-02 -4.66541141e-01 4.07427728e-01 4.34203684e-01
-5.78905523e-01 -4.40647751e-02 -9.28256512e-01 -8.61970365e-01
-8.54892097e-03 -8.08709621e-01 2.69348502e-01 1.39170036e-01
-2.46755600e-01 -3.00362468e-01 9.53641176e-01 -4.96556431e-01
1.26767325e+00 -2.33999300e+00 6.42347574e-01 -7.41979778e-02
3.56434286e-01 9.29435715e-02 -5.97988069e-01 7.94002831e-01
-3.37010771e-01 2.47357368e-01 -5.89993596e-02 -6.30788505e-01
1.72090665e-01 5.60375571e-01 -3.44326466e-01 3.32982183e-01
2.65916675e-01 1.33337557e+00 -1.07119048e+00 -7.29946196e-01
1.72033221e-01 7.24685729e-01 -5.22300959e-01 1.28525943e-01
-1.03595674e-01 -5.02536260e-03 -4.29085314e-01 7.17577994e-01
2.21775740e-01 -5.38440168e-01 9.88822430e-02 -4.72652912e-01
3.37475449e-01 -1.90633744e-01 -9.70141232e-01 2.10226989e+00
-6.66761220e-01 9.99677122e-01 -2.07746893e-01 -1.17119443e+00
5.11183500e-01 3.82744610e-01 5.03758609e-01 -8.61822128e-01
-1.37080476e-01 -6.21271925e-03 -5.17161191e-01 -8.88700068e-01
8.56538236e-01 -2.27276519e-01 -2.10198596e-01 5.41502535e-01
3.95818025e-01 2.28110209e-01 1.43154860e-02 6.32282138e-01
8.83074105e-01 -1.02858983e-01 9.26876590e-02 5.58962822e-02
3.78197551e-01 -4.24775779e-01 5.47995754e-02 8.47902536e-01
-1.12642542e-01 9.11504269e-01 5.69402158e-01 -4.48933393e-01
-9.85459089e-01 -1.23501813e+00 -4.01918544e-03 1.43425310e+00
4.49908018e-01 -8.66903305e-01 -1.77197754e-01 -6.04853153e-01
6.84572160e-02 2.74329782e-01 -7.02699542e-01 -2.16533825e-01
-3.02379638e-01 -3.39551747e-01 2.51570702e-01 4.00774270e-01
1.38545945e-01 -1.20596504e+00 -1.67838529e-01 -3.41875590e-02
-5.40253282e-01 -1.17460859e+00 -7.78551638e-01 -2.20613629e-01
-4.59352583e-01 -1.10092294e+00 -9.45752919e-01 -9.50838149e-01
5.74348152e-01 6.19678199e-01 1.35339141e+00 3.49399418e-01
-5.58481157e-01 1.10872424e+00 -5.77297568e-01 9.29935351e-02
-3.54935490e-02 -7.85614848e-02 -1.85329810e-01 3.38178933e-01
3.95877630e-01 -1.65343434e-01 -6.06328070e-01 2.09217936e-01
-1.41838026e+00 -6.09789304e-02 3.81310672e-01 1.23556709e+00
6.80210650e-01 -3.30729634e-01 4.19977993e-01 -4.72151935e-01
7.54176915e-01 -7.73272991e-01 -2.80050505e-02 7.59601712e-01
-1.50795311e-01 9.48497728e-02 1.51055828e-01 -6.34978533e-01
-6.17666662e-01 -2.71880537e-01 3.62505466e-01 -1.12063730e+00
1.72729120e-01 6.46364629e-01 -4.48257290e-02 3.64822745e-01
2.99872220e-01 1.72007158e-01 -8.27761516e-02 -3.18534493e-01
6.40672684e-01 6.15064204e-01 3.30931634e-01 -6.47806644e-01
5.49566925e-01 4.59913731e-01 -3.38599026e-01 -1.00429583e+00
-1.00461233e+00 -7.95475006e-01 -5.75909555e-01 -2.10058734e-01
1.16975880e+00 -1.05471635e+00 -4.85662460e-01 7.97457770e-02
-1.13977003e+00 -1.44712955e-01 -2.34266803e-01 3.04390579e-01
-6.42523646e-01 4.72440213e-01 -4.27467257e-01 -3.85457754e-01
-1.47983342e-01 -1.30430365e+00 1.70428383e+00 -2.92797089e-01
-9.27510485e-02 -1.23865533e+00 3.02393809e-02 4.72586244e-01
3.59109819e-01 1.07330624e-02 8.44701707e-01 -4.33558941e-01
-6.75913453e-01 -2.52402723e-01 -4.02143151e-01 1.31654173e-01
-6.24106750e-02 1.26812432e-03 -8.06203604e-01 -5.96550405e-01
-3.59313905e-01 -7.50927091e-01 1.20680034e+00 2.57492572e-01
1.38742101e+00 -4.29067820e-01 -3.72970462e-01 5.82818031e-01
1.54129434e+00 -3.03177416e-01 6.49384856e-01 4.12502766e-01
8.20389330e-01 5.97419441e-01 4.73438561e-01 2.70468324e-01
5.18867731e-01 8.14434230e-01 4.12946135e-01 5.73533028e-02
3.63875506e-03 -3.05850089e-01 4.50300694e-01 9.09291327e-01
1.64625585e-01 -6.20932221e-01 -8.03122938e-01 8.50821197e-01
-2.07219505e+00 -1.35553801e+00 3.19013387e-01 1.92030048e+00
5.56816757e-01 -3.17115307e-01 -2.40142360e-01 -3.25154871e-01
8.36029887e-01 6.02745473e-01 -3.32660407e-01 -2.10554197e-01
-1.93037212e-01 -3.01874820e-02 1.32536422e-02 5.17735720e-01
-1.21974456e+00 9.45521235e-01 5.65863276e+00 9.29042876e-01
-1.09077537e+00 3.12732697e-01 3.50883335e-01 -4.53801543e-01
-6.30485296e-01 -8.99948180e-02 -3.89707893e-01 3.33120465e-01
6.29925907e-01 -1.36622459e-01 3.97888809e-01 5.42246699e-01
-3.51132005e-01 1.90906972e-01 -1.33261895e+00 1.49673724e+00
6.31289124e-01 -1.24250960e+00 5.78756094e-01 -1.51954472e-01
7.14262128e-01 -4.77066124e-03 4.29998815e-01 4.44929898e-01
-1.62892833e-01 -1.15406275e+00 5.61260760e-01 6.39597237e-01
9.45418894e-01 -5.03510416e-01 4.77696002e-01 5.32178860e-03
-1.25351942e+00 -1.31760359e-01 -5.06946921e-01 4.05072540e-01
1.40727550e-01 2.68354695e-02 -3.82914066e-01 6.15475416e-01
8.45201612e-01 1.33561444e+00 -7.48278916e-01 7.18598664e-01
-2.54884847e-02 2.06730864e-03 -4.13764641e-02 6.46701306e-02
5.25194883e-01 -2.07482185e-02 7.07826316e-01 1.28897476e+00
1.98191553e-01 -9.26752687e-02 2.09056973e-01 8.68931830e-01
-2.60031432e-01 1.13721758e-01 -9.60525453e-01 -2.78747380e-01
2.47064963e-01 1.18278408e+00 -4.45630938e-01 -5.41452706e-01
-5.99054217e-01 1.26920795e+00 4.83392835e-01 6.64680064e-01
-8.60275567e-01 -2.01600224e-01 8.48492682e-01 -2.34034970e-01
2.13294193e-01 -1.63987905e-01 4.01202053e-01 -1.66555572e+00
1.45997241e-01 -7.47627020e-01 5.85170209e-01 -1.11199558e+00
-1.84853971e+00 4.62120920e-01 2.53331125e-01 -1.25426304e+00
-1.70987993e-01 -6.75281882e-01 -3.81346703e-01 5.06211162e-01
-1.55023396e+00 -1.30180526e+00 -2.72238910e-01 9.26840425e-01
9.05192912e-01 -1.69432253e-01 7.04662681e-01 4.56249982e-01
-2.75666922e-01 7.01548159e-01 2.07893342e-01 4.55324091e-02
1.08315527e+00 -1.17034161e+00 -3.32365148e-02 2.94181049e-01
6.74503148e-01 8.82109344e-01 3.72567743e-01 -4.35752839e-01
-1.64544284e+00 -9.80458319e-01 8.80661845e-01 -6.60544157e-01
9.80570614e-01 -4.53186035e-01 -9.49390173e-01 9.71523225e-01
5.42579412e-01 1.76219225e-01 7.74347842e-01 1.63244367e-01
-5.05013406e-01 1.46813527e-01 -6.21948838e-01 8.02862287e-01
9.81540322e-01 -1.16227329e+00 -8.88310909e-01 5.66587269e-01
9.12419498e-01 -1.54020160e-01 -9.15611148e-01 1.71059281e-01
6.87010109e-01 -6.73509598e-01 1.37794530e+00 -1.02650869e+00
7.87949443e-01 -2.91589666e-02 -5.26284993e-01 -1.25524807e+00
-1.22506745e-01 -2.93670833e-01 1.84190664e-02 1.11146724e+00
1.69138089e-01 -3.07031512e-01 3.23358029e-01 4.28694636e-01
-3.96070583e-03 -7.16383398e-01 -8.84062111e-01 -6.20009959e-01
6.43630624e-02 -5.33960402e-01 4.10940975e-01 1.38001382e+00
-2.88385954e-02 3.85492593e-01 -6.44824624e-01 5.52354977e-02
4.86105710e-01 2.54477233e-01 5.89227498e-01 -7.90554643e-01
-3.07159841e-01 -4.11107779e-01 -7.84517348e-01 -9.76381123e-01
5.74506879e-01 -1.24779856e+00 -2.15266511e-01 -1.46517789e+00
6.24775708e-01 1.47297280e-02 -5.39949179e-01 5.13511956e-01
-2.28139207e-01 4.96109366e-01 5.22949338e-01 2.80346364e-01
-1.19118893e+00 7.37149715e-01 1.26874101e+00 -6.12051964e-01
1.77549332e-01 -7.92452812e-01 -4.40565467e-01 4.28178310e-01
4.44971830e-01 -4.26270068e-01 -4.94030714e-01 -8.50995004e-01
5.77019870e-01 2.24413157e-01 7.28016198e-01 -5.88666201e-01
9.04466286e-02 1.56803243e-02 3.41704696e-01 -6.85926795e-01
7.26298451e-01 -9.95322347e-01 3.52942199e-02 -9.83788893e-02
-7.54752934e-01 2.59591877e-01 -5.38570322e-02 9.28598166e-01
-6.93845153e-01 -3.01899701e-01 1.73945740e-01 -2.13299662e-01
-9.72555220e-01 5.32475412e-01 -1.91913530e-01 2.44575247e-01
9.54373837e-01 -1.46470442e-01 -4.62709397e-01 -5.54454029e-01
-1.14860141e+00 5.01301944e-01 5.65102398e-01 1.09905601e+00
9.64998960e-01 -1.67420232e+00 -5.76481104e-01 1.44227728e-01
8.00511837e-01 -3.39769214e-01 5.52440822e-01 6.55414820e-01
-1.07082896e-01 4.68496710e-01 -1.22582197e-01 -8.69362712e-01
-1.37461603e+00 7.74595797e-01 1.69084668e-01 -1.03409179e-01
-6.66970789e-01 7.24266589e-01 5.77557504e-01 -3.48972380e-01
1.80543676e-01 -1.69536322e-02 -2.73249030e-01 3.50876838e-01
5.27926862e-01 6.35336414e-02 -3.30518275e-01 -1.01208591e+00
-2.33697623e-01 7.93495953e-01 -1.14492960e-01 -1.57855637e-02
1.24691927e+00 -3.40254724e-01 -2.93999553e-01 8.36889505e-01
2.07397914e+00 -4.35689270e-01 -9.40321088e-01 -4.18131590e-01
-1.51644275e-01 -8.10066581e-01 -4.39538024e-02 -3.97649020e-01
-1.14956903e+00 1.05966198e+00 6.18453801e-01 1.09521471e-01
9.04547393e-01 5.15074909e-01 5.93522787e-01 6.54831231e-01
2.71888338e-02 -1.09441268e+00 7.98177779e-01 4.12133306e-01
1.19470859e+00 -1.47825503e+00 -1.14550233e-01 1.37041911e-01
-8.14067364e-01 1.10566020e+00 5.31344235e-01 -1.78783070e-02
4.96989906e-01 -3.64209622e-01 -1.87957883e-01 -6.31286144e-01
-1.07206786e+00 -1.88115090e-01 6.73373938e-01 3.00350308e-01
4.63095129e-01 6.32961243e-02 -8.61024410e-02 2.88087845e-01
2.04288960e-01 -3.30814928e-01 2.70586520e-01 9.16813970e-01
-1.11121632e-01 -1.03115022e+00 -2.80294359e-01 4.09882039e-01
-4.83177722e-01 -1.97976172e-01 -4.51472580e-01 8.19819033e-01
-2.07917005e-01 5.88175237e-01 3.01852554e-01 -1.36773616e-01
9.17752609e-02 3.34101349e-01 4.54710931e-01 -5.48488557e-01
-2.84530252e-01 1.22221159e-02 -2.43752748e-01 -7.47506678e-01
-7.53415823e-01 -5.31577229e-01 -8.45144689e-01 2.19961572e-02
-1.45844415e-01 -1.67139053e-01 5.95560491e-01 7.86221385e-01
3.84908855e-01 4.65511888e-01 3.82574260e-01 -9.58883822e-01
-2.53790259e-01 -6.14994645e-01 -5.52024662e-01 9.34046924e-01
5.01158297e-01 -8.14013124e-01 -4.55560982e-01 1.36694163e-01] | [10.486713409423828, 1.1839659214019775] |
0f7ea0b4-82c5-4d02-9ea8-8b9431a9a82f | vlg-net-video-language-graph-matching-network | 2011.10132 | null | https://arxiv.org/abs/2011.10132v2 | https://arxiv.org/pdf/2011.10132v2.pdf | VLG-Net: Video-Language Graph Matching Network for Video Grounding | Grounding language queries in videos aims at identifying the time interval (or moment) semantically relevant to a language query. The solution to this challenging task demands understanding videos' and queries' semantic content and the fine-grained reasoning about their multi-modal interactions. Our key idea is to recast this challenge into an algorithmic graph matching problem. Fueled by recent advances in Graph Neural Networks, we propose to leverage Graph Convolutional Networks to model video and textual information as well as their semantic alignment. To enable the mutual exchange of information across the modalities, we design a novel Video-Language Graph Matching Network (VLG-Net) to match video and query graphs. Core ingredients include representation graphs built atop video snippets and query tokens separately and used to model intra-modality relationships. A Graph Matching layer is adopted for cross-modal context modeling and multi-modal fusion. Finally, moment candidates are created using masked moment attention pooling by fusing the moment's enriched snippet features. We demonstrate superior performance over state-of-the-art grounding methods on three widely used datasets for temporal localization of moments in videos with language queries: ActivityNet-Captions, TACoS, and DiDeMo. | ['Sisi Qu', 'Mengmeng Xu', 'Bernard Ghanem', 'Jesper Tegner', 'Mattia Soldan'] | 2020-11-19 | null | null | null | null | ['video-grounding', 'moment-retrieval', 'natural-language-moment-retrieval'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.28614132e-02 -1.98261559e-01 -5.64745009e-01 -3.23614955e-01
-8.35728228e-01 -5.92883170e-01 9.04709518e-01 5.63077390e-01
-2.46772185e-01 3.53111066e-02 8.28556836e-01 1.46418393e-01
-3.14907357e-02 -5.83926558e-01 -7.84585059e-01 -1.75863892e-01
-4.24639463e-01 7.84928817e-03 3.28890920e-01 -1.18999235e-01
-6.55645970e-03 1.44173756e-01 -1.63477337e+00 9.21364129e-01
4.01490211e-01 1.39737761e+00 1.04056276e-01 7.03521609e-01
-3.92743409e-01 1.54344654e+00 -1.61325291e-01 -5.25777578e-01
-1.13097675e-01 -5.09981751e-01 -9.24135268e-01 1.46512687e-01
6.93710685e-01 -1.96433827e-01 -1.09668338e+00 9.48051095e-01
3.17094117e-01 4.73870516e-01 2.60375500e-01 -1.77085769e+00
-7.45918810e-01 7.26878524e-01 -3.19582105e-01 5.87159932e-01
1.13652122e+00 6.74166232e-02 1.30201113e+00 -5.82829058e-01
1.01779711e+00 1.22710431e+00 4.52751160e-01 2.73607671e-01
-8.63777995e-01 -3.76489043e-01 3.76651734e-01 7.05425143e-01
-1.58429205e+00 -5.33027351e-01 9.46965694e-01 -5.59635341e-01
1.23541486e+00 1.87176019e-01 9.06528294e-01 1.36446083e+00
-5.65588363e-02 9.10807908e-01 3.84044915e-01 3.87544148e-02
-1.75386861e-01 -4.95665282e-01 -1.85700431e-01 1.08631074e+00
-3.90909582e-01 -2.63191104e-01 -1.15710962e+00 -2.17784613e-01
5.55365384e-01 4.18053746e-01 -1.65887296e-01 -3.35882157e-01
-1.70761490e+00 7.04066575e-01 6.23076022e-01 5.20765960e-01
-5.66123307e-01 6.19830787e-01 7.04553604e-01 2.47739688e-01
4.74298149e-01 2.05007538e-01 3.52430679e-02 -9.13723111e-02
-9.04641747e-01 3.56927305e-01 7.24350393e-01 1.18953907e+00
7.19994366e-01 -3.25031757e-01 -7.94794917e-01 5.62274218e-01
4.42174792e-01 2.01693535e-01 3.63886356e-01 -8.09856176e-01
8.47095072e-01 8.80421162e-01 -1.93908885e-01 -1.53418899e+00
-1.68163255e-01 -2.60142498e-02 -4.37395245e-01 -8.25233161e-01
1.47956520e-01 2.11376771e-01 -8.43250811e-01 1.93797529e+00
2.52223402e-01 8.35297763e-01 -2.32680112e-01 9.62383151e-01
1.21883714e+00 5.80010414e-01 5.39976418e-01 9.21453908e-02
1.66105783e+00 -8.48833799e-01 -7.50612140e-01 -1.37928396e-01
6.03296995e-01 -5.15628040e-01 8.59137893e-01 -3.20533603e-01
-1.04716599e+00 -4.79341000e-01 -6.73777223e-01 -3.39597046e-01
-7.26444185e-01 -3.23323220e-01 6.28866136e-01 -2.87424833e-01
-1.14545262e+00 4.19638187e-01 -6.78537071e-01 -7.41079509e-01
4.94179249e-01 1.43084154e-01 -6.34328783e-01 -1.57257110e-01
-1.39422977e+00 3.94167960e-01 5.23379087e-01 1.13528259e-02
-9.94963050e-01 -4.92512614e-01 -1.28616560e+00 7.29046464e-02
7.39413738e-01 -1.00631046e+00 1.05998087e+00 -9.72362697e-01
-9.24883187e-01 1.14727437e+00 -3.97191525e-01 -6.70429885e-01
2.12900788e-01 -3.56073789e-02 -6.71303093e-01 6.44634485e-01
3.89193624e-01 8.10128868e-01 8.02195549e-01 -6.58176482e-01
-6.58152163e-01 -3.50131154e-01 4.89779472e-01 3.04097801e-01
-1.58536151e-01 3.04046333e-01 -1.10354483e+00 -7.63090372e-01
9.72443596e-02 -6.77205086e-01 1.06900677e-01 -2.76041567e-01
-3.48106474e-01 -4.98716086e-01 9.30521190e-01 -8.70008945e-01
1.54381251e+00 -2.12530255e+00 4.01118189e-01 8.97326618e-02
6.46607935e-01 -3.81925315e-01 -4.02759016e-01 7.99124002e-01
1.71047412e-02 -1.70807131e-02 3.50406021e-01 -4.03460324e-01
2.53980488e-01 2.08554134e-01 -3.42597306e-01 4.34607446e-01
1.64431646e-01 1.49815023e+00 -1.31241202e+00 -8.62170637e-01
7.34810755e-02 5.95665872e-01 -6.38281524e-01 5.03798425e-01
-6.00744724e-01 3.44925165e-01 -5.64970553e-01 9.56754744e-01
-1.28839538e-01 -6.90655172e-01 8.25203881e-02 -9.29312587e-01
2.87464380e-01 1.14709832e-01 -8.00362289e-01 2.43861556e+00
-2.39364222e-01 6.89762890e-01 -1.43446967e-01 -1.10973883e+00
4.56974000e-01 5.71790099e-01 1.08899951e+00 -8.78587008e-01
7.15099201e-02 -2.32885092e-01 -6.42763853e-01 -9.95436549e-01
6.12538517e-01 3.32152069e-01 -3.97096217e-01 2.63217866e-01
6.89090550e-01 1.76078662e-01 2.38664612e-01 8.40588748e-01
1.18700302e+00 2.29996949e-01 1.23659432e-01 1.95677921e-01
4.89631504e-01 -1.87579408e-01 2.21346051e-01 5.81795931e-01
-2.30501443e-01 3.71633291e-01 6.64743125e-01 -4.47029650e-01
-5.56388199e-01 -8.83075953e-01 8.13383758e-01 1.70292401e+00
3.81293178e-01 -8.89062047e-01 -5.05811810e-01 -7.83943295e-01
8.49357471e-02 2.05209196e-01 -8.47868979e-01 -1.45487696e-01
-4.29098666e-01 1.14950072e-02 4.77230638e-01 4.53123689e-01
4.90430027e-01 -8.79910827e-01 -2.46047705e-01 1.41135558e-01
-7.62528479e-01 -1.82078779e+00 -1.02949584e+00 -3.98138553e-01
-3.21884096e-01 -1.46336949e+00 -4.62394625e-01 -6.61099970e-01
3.97209197e-01 4.52792227e-01 1.51456177e+00 1.72297820e-01
-1.93736553e-01 1.34390545e+00 -5.21080136e-01 7.88918808e-02
-5.47345765e-02 -8.09195563e-02 -1.68536425e-01 5.31029820e-01
5.06013393e-01 -5.20854712e-01 -7.98976839e-01 1.77722827e-01
-1.12790155e+00 1.80417299e-01 1.08903408e-01 4.75867122e-01
7.78128803e-01 -3.70771140e-01 2.60857046e-01 -4.31308031e-01
5.93119144e-01 -1.01663685e+00 -4.08040613e-01 5.38185835e-01
-2.01058146e-02 5.54315224e-02 2.04989627e-01 -5.52553117e-01
-6.83398485e-01 -1.52517026e-02 3.51180434e-01 -1.04467511e+00
3.38593312e-02 7.65990078e-01 -1.06795028e-01 5.15461247e-03
3.03546607e-01 1.46275893e-01 -2.63599008e-01 -1.51411548e-01
8.28864574e-01 2.31407896e-01 9.45210278e-01 -5.86440086e-01
4.10106510e-01 6.99570537e-01 7.36644566e-02 -5.71696401e-01
-9.24707413e-01 -9.56742585e-01 -5.11195540e-01 -5.87121189e-01
1.50039268e+00 -1.17746222e+00 -1.02791822e+00 1.06390893e-01
-1.17066371e+00 -2.21122667e-01 -1.13483891e-01 4.15147364e-01
-8.48693609e-01 4.88591582e-01 -4.79206651e-01 -2.74692774e-01
-1.95312217e-01 -1.01153886e+00 1.42666864e+00 5.13943881e-02
-1.07664041e-01 -1.24480021e+00 -1.84914451e-02 5.38091183e-01
2.35212713e-01 6.39260650e-01 5.33621013e-01 -8.87758613e-01
-8.81155729e-01 -2.23224834e-01 -4.69180107e-01 -2.28083670e-01
1.96863357e-02 -1.80179089e-01 -7.39531934e-01 -1.54541820e-01
-4.05314982e-01 -3.22764099e-01 9.14055169e-01 2.44727015e-01
1.28029656e+00 -5.55008948e-01 -5.03916442e-01 8.18239033e-01
1.24611247e+00 -1.84679180e-01 2.83307254e-01 -4.73408587e-03
1.15019882e+00 4.52467620e-01 4.04161036e-01 5.73362291e-01
8.41196001e-01 8.39433312e-01 7.39350259e-01 2.58601397e-01
-2.64209807e-01 -7.20431209e-01 3.77988100e-01 6.67915642e-01
-2.23172791e-02 -4.32000577e-01 -8.63906205e-01 7.58523941e-01
-2.24663973e+00 -1.54557025e+00 2.98080564e-01 1.77778828e+00
3.90843302e-01 -3.21900904e-01 2.41213679e-01 -6.18015647e-01
7.55795121e-01 7.24993646e-01 -3.03686023e-01 2.60525018e-01
-1.51417002e-01 -3.21077377e-01 3.24139744e-01 2.37239689e-01
-1.32631707e+00 9.15975273e-01 5.09129286e+00 6.54585779e-01
-9.62816477e-01 2.89665639e-01 3.31181854e-01 -2.27072477e-01
-4.01714772e-01 7.00336173e-02 -3.99014384e-01 4.58535582e-01
9.50019181e-01 -1.94645718e-01 8.00665140e-01 4.10515815e-01
1.68845356e-01 8.50936919e-02 -1.47833371e+00 1.48823369e+00
4.74559277e-01 -1.87279034e+00 2.62000561e-01 -3.94416422e-01
5.61112881e-01 3.80652040e-01 -2.21601993e-01 4.14891869e-01
8.12294707e-02 -8.38474333e-01 9.64467645e-01 8.64516854e-01
7.27357924e-01 -2.84757167e-01 4.43821490e-01 -1.70116097e-01
-2.00488520e+00 -7.50742108e-02 1.79210499e-01 2.85525829e-01
3.20275456e-01 -1.40106771e-02 -7.33862936e-01 8.84034574e-01
6.93523169e-01 1.29050088e+00 -7.10809827e-01 8.60209882e-01
8.43965039e-02 1.77341327e-01 -9.83359814e-02 2.95031548e-01
4.96732473e-01 2.02480610e-02 5.30655563e-01 1.45048654e+00
3.27878892e-01 1.59055945e-02 4.99899477e-01 9.19890404e-01
-4.79770035e-01 3.16206925e-02 -8.86462986e-01 -6.71056449e-01
4.50911343e-01 1.12411463e+00 -6.62076354e-01 -6.10980332e-01
-8.71846557e-01 1.17584276e+00 3.59830260e-01 7.02613652e-01
-1.14040852e+00 -1.55450881e-01 7.68066049e-01 1.30022526e-01
1.89751655e-01 -1.92805693e-01 5.87111771e-01 -1.54293811e+00
-5.89749310e-03 -6.54993534e-01 1.02153313e+00 -1.05194664e+00
-1.44003177e+00 5.28880775e-01 2.77327091e-01 -1.20062578e+00
-5.29184043e-01 -2.46964812e-01 -5.04186809e-01 5.22671044e-01
-1.29624677e+00 -1.80344260e+00 -7.32074678e-01 1.33106863e+00
4.23867702e-01 -9.39610228e-02 4.81421590e-01 6.03076935e-01
-1.60563856e-01 3.27555865e-01 -7.63942003e-01 4.87405211e-01
4.93184119e-01 -8.67088079e-01 2.32578576e-01 7.57047951e-01
6.22862279e-01 4.09295589e-01 3.91629457e-01 -7.70738542e-01
-1.91034305e+00 -1.31038320e+00 9.80748892e-01 -5.60693502e-01
1.20524907e+00 -3.78449887e-01 -8.34820926e-01 9.49207246e-01
2.85545498e-01 4.94216979e-01 6.03123248e-01 -1.53875515e-01
-7.65025556e-01 4.08700332e-02 -6.97029173e-01 6.75492287e-01
1.57380128e+00 -1.53726506e+00 -5.56107879e-01 6.43490314e-01
1.15252900e+00 -5.56106389e-01 -1.09049022e+00 2.42393702e-01
4.12810147e-01 -7.62227356e-01 1.16026855e+00 -1.02918518e+00
1.90216690e-01 -1.65844858e-01 -5.24805665e-01 -7.94979155e-01
5.72220720e-02 -1.14406967e+00 -5.43029964e-01 1.19566822e+00
-2.64768098e-02 -3.12052928e-02 6.06176615e-01 5.01888573e-01
-2.43139878e-01 -4.45121139e-01 -8.65965843e-01 -5.05747139e-01
-9.25353587e-01 -8.16473484e-01 7.14750051e-01 1.22386396e+00
3.52384239e-01 3.44901502e-01 -3.05284917e-01 2.06438094e-01
3.74022812e-01 1.56469166e-01 6.83262050e-01 -8.87120426e-01
-2.12129116e-01 -5.94646215e-01 -7.72006273e-01 -9.78052080e-01
6.37907982e-01 -1.22638392e+00 -2.10911691e-01 -1.73614705e+00
2.82363504e-01 2.43623629e-01 -5.04539371e-01 4.93913054e-01
-3.16188596e-02 1.71412095e-01 3.54136825e-01 1.23672970e-01
-1.48935783e+00 4.18867737e-01 1.06084847e+00 -3.65195811e-01
-1.13729201e-01 -4.38773721e-01 -3.27312261e-01 4.60093349e-01
2.32877940e-01 -1.49991095e-01 -6.22509480e-01 -6.34776473e-01
6.32524610e-01 5.78721046e-01 8.98246169e-01 -9.26842332e-01
5.57111084e-01 -2.13511497e-01 -4.45639975e-02 -5.78799903e-01
4.69539762e-01 -8.95625889e-01 3.46266747e-01 -1.30914971e-01
-5.88093638e-01 4.20499504e-01 1.26462862e-01 1.04832304e+00
-6.44674003e-01 4.43018943e-01 1.38265684e-01 -1.96879417e-01
-1.24405468e+00 9.33822393e-01 1.80342682e-02 2.39132896e-01
1.02575254e+00 -7.51783550e-02 -4.20968592e-01 -7.78383553e-01
-1.12674677e+00 3.90585005e-01 3.41341674e-01 8.42118263e-01
6.99438870e-01 -1.70055532e+00 -3.50074351e-01 -1.57170817e-02
6.87381029e-01 -4.75602567e-01 5.36650717e-01 1.08073843e+00
-9.09984112e-02 5.15880525e-01 6.16657883e-02 -6.64717972e-01
-1.19164383e+00 7.86129415e-01 3.34114224e-01 -1.40750617e-01
-3.73587042e-01 9.65129852e-01 1.40991464e-01 1.50556892e-01
3.27321738e-01 -6.64617002e-01 -1.85528904e-01 3.47967774e-01
3.78887475e-01 7.82969072e-02 -1.95953161e-01 -1.11962843e+00
-4.69291836e-01 4.62035298e-01 4.69217718e-01 -4.28351574e-03
1.10875916e+00 -3.42808366e-01 -2.98427165e-01 3.17770839e-01
1.52915764e+00 -2.03827441e-01 -1.11504734e+00 -6.47110879e-01
1.85853601e-01 -3.93332332e-01 -8.39821622e-02 -3.40583384e-01
-1.09529519e+00 6.21053159e-01 1.91249862e-01 3.23939472e-01
1.10204399e+00 6.22684419e-01 8.29437613e-01 2.95392543e-01
1.77167431e-01 -8.20258856e-01 4.81946677e-01 3.50837052e-01
9.15774167e-01 -1.36193573e+00 -2.09745035e-01 -1.36134252e-01
-6.05120778e-01 9.64138627e-01 4.31846112e-01 1.28424004e-01
6.55690730e-01 -3.55581909e-01 -3.83899242e-01 -7.77442276e-01
-9.06541109e-01 -6.02323592e-01 1.05317950e+00 4.58075881e-01
3.47140998e-01 -9.01867300e-02 2.11605623e-01 5.89371026e-01
3.33880723e-01 1.23122513e-01 -1.27807692e-01 7.82261729e-01
-5.34832627e-02 -6.56275272e-01 1.58501826e-02 3.74061435e-01
-5.57805181e-01 -9.30577666e-02 -2.92863488e-01 5.58530331e-01
4.90901954e-02 9.05792832e-01 3.43938172e-01 -6.33858621e-01
1.66384667e-01 1.22255169e-01 4.17418212e-01 -5.02070248e-01
-5.77545166e-01 1.01174728e-03 1.34890378e-01 -1.42639101e+00
-9.35448647e-01 -3.81906360e-01 -1.19704163e+00 -1.44352272e-01
2.41159275e-01 7.03771785e-02 3.38491708e-01 9.44484115e-01
8.01031888e-01 5.86806476e-01 1.75506309e-01 -8.78959715e-01
1.55398905e-01 -5.83767891e-01 -3.27145427e-01 1.09291279e+00
3.10531914e-01 -7.73850024e-01 -9.69016925e-02 4.10354167e-01] | [10.058099746704102, 0.8284958004951477] |
67658ba9-d0bd-49b7-86f9-48d3fa797e6f | deepportraitdrawing-generating-human-body | 2205.02070 | null | https://arxiv.org/abs/2205.02070v2 | https://arxiv.org/pdf/2205.02070v2.pdf | DeepPortraitDrawing: Generating Human Body Images from Freehand Sketches | Researchers have explored various ways to generate realistic images from freehand sketches, e.g., for objects and human faces. However, how to generate realistic human body images from sketches is still a challenging problem. It is, first because of the sensitivity to human shapes, second because of the complexity of human images caused by body shape and pose changes, and third because of the domain gap between realistic images and freehand sketches. In this work, we present DeepPortraitDrawing, a deep generative framework for converting roughly drawn sketches to realistic human body images. To encode complicated body shapes under various poses, we take a local-to-global approach. Locally, we employ semantic part auto-encoders to construct part-level shape spaces, which are useful for refining the geometry of an input pre-segmented hand-drawn sketch. Globally, we employ a cascaded spatial transformer network to refine the structure of body parts by adjusting their spatial locations and relative proportions. Finally, we use a global synthesis network for the sketch-to-image translation task, and a face refinement network to enhance facial details. Extensive experiments have shown that given roughly sketched human portraits, our method produces more realistic images than the state-of-the-art sketch-to-image synthesis techniques. | ['Shi-Min Hu', 'Song-Hai Zhang', 'Ariel Shamir', 'Hongbo Fu', 'Chen Wang', 'Xian Wu'] | 2022-05-04 | null | null | null | null | ['sketch-to-image-translation'] | ['computer-vision'] | [ 2.89758652e-01 2.17760861e-01 1.73720002e-01 -4.63457853e-01
-2.61785775e-01 -5.45198321e-01 6.49187744e-01 -8.35880280e-01
1.39866814e-01 5.50265670e-01 2.35545039e-01 1.83303297e-01
1.66939840e-01 -1.03919673e+00 -7.11854279e-01 -3.39009970e-01
5.14055431e-01 6.22505069e-01 1.41572997e-01 -4.24259841e-01
1.02517111e-02 8.65071833e-01 -1.46217227e+00 2.53616363e-01
4.87104088e-01 8.55679035e-01 -2.52116829e-01 6.86465442e-01
-2.43015394e-01 2.98669606e-01 -5.70333481e-01 -1.06495225e+00
4.53815341e-01 -7.69035697e-01 -4.45947528e-01 4.86981899e-01
8.42023075e-01 -7.82797873e-01 -4.67615724e-01 8.64019215e-01
4.94228214e-01 -5.54376729e-02 7.54113913e-01 -1.39998829e+00
-8.87100995e-01 4.14762467e-01 -7.93840766e-01 -6.74724162e-01
3.66436601e-01 3.27778995e-01 5.80268145e-01 -9.74593163e-01
9.18198466e-01 1.72829235e+00 6.46757364e-01 9.58197117e-01
-1.29021645e+00 -1.10649157e+00 3.88695635e-02 -3.85389626e-01
-1.40325665e+00 -5.84559023e-01 1.09756458e+00 -4.43900466e-01
1.51144758e-01 1.55326128e-01 9.93521869e-01 1.05150104e+00
4.37560864e-02 6.28090620e-01 8.19086671e-01 -2.50789940e-01
-1.20070748e-01 -2.14153543e-01 -7.78561890e-01 1.05389798e+00
1.02341147e-02 1.32186189e-01 -4.19570059e-01 -1.48087546e-01
1.78109181e+00 1.06535338e-01 -1.70453683e-01 -5.91124356e-01
-1.18433142e+00 7.00861692e-01 4.41438138e-01 2.52818894e-02
-3.37291420e-01 5.33791959e-01 6.26671240e-02 -2.23401804e-02
2.47546956e-01 3.54740709e-01 -2.58127023e-02 2.08471179e-01
-1.20612013e+00 7.46634066e-01 6.26935840e-01 1.09673011e+00
5.77928603e-01 1.60182729e-01 -2.84103811e-01 8.90979171e-01
2.78064609e-01 6.61074340e-01 6.94492180e-03 -1.09716856e+00
3.21935922e-01 5.68453252e-01 1.62039235e-01 -1.11744511e+00
-3.86256091e-02 2.62082145e-02 -1.10215795e+00 4.38323528e-01
4.17380005e-01 2.89413687e-02 -1.11573446e+00 1.71444881e+00
4.72061336e-01 -1.82458088e-01 -3.98218960e-01 1.24711072e+00
9.53472793e-01 4.86790776e-01 1.13831215e-01 3.20290774e-01
1.42663968e+00 -1.04104924e+00 -7.60078192e-01 -8.15605000e-02
-3.34855199e-01 -7.68654048e-01 1.06647587e+00 7.32072815e-02
-1.56000626e+00 -7.60162234e-01 -8.40376079e-01 -4.52230781e-01
-1.40803561e-01 3.45436633e-01 2.92789072e-01 4.99031544e-01
-9.38047469e-01 6.32852614e-01 -5.35442412e-01 -7.52192810e-02
6.21001542e-01 2.09084690e-01 -5.18637538e-01 -5.36735170e-03
-1.03586328e+00 5.96630394e-01 2.84617972e-02 1.16695657e-01
-7.06094742e-01 -7.83137619e-01 -1.07949495e+00 1.54544532e-01
1.95911273e-01 -1.11300707e+00 1.13158262e+00 -1.11128044e+00
-1.79978418e+00 1.02837157e+00 -3.66114788e-02 1.76328659e-01
8.22638452e-01 1.26490772e-01 -1.76230296e-01 1.61034405e-01
-9.60786268e-02 1.13001573e+00 1.28457236e+00 -1.35024798e+00
-1.05336130e-01 -3.11629832e-01 -2.98623126e-02 2.07168818e-01
2.11420462e-01 -1.23184742e-02 -8.43479693e-01 -1.17901862e+00
1.10219933e-01 -9.12792265e-01 -2.54533142e-01 9.06748891e-01
-4.32530135e-01 8.56606737e-02 7.72231996e-01 -8.52769375e-01
9.15559947e-01 -2.00539041e+00 4.21472639e-01 4.06599104e-01
1.93619221e-01 1.92961559e-01 -4.66542065e-01 4.35298711e-01
-7.27504492e-02 1.74061805e-01 -3.30807567e-01 -5.91337621e-01
1.51979789e-01 2.25210458e-01 -4.13545728e-01 1.68415487e-01
3.80632609e-01 1.28694880e+00 -7.73717999e-01 -6.18036509e-01
3.01032811e-01 7.67939508e-01 -6.45556688e-01 3.58446091e-01
-2.40917787e-01 5.40086746e-01 -3.53203624e-01 7.18590736e-01
8.22181463e-01 1.02783494e-01 2.00297404e-02 -4.58338559e-01
1.91976577e-01 -2.26562127e-01 -1.10943842e+00 1.79205370e+00
-4.53600109e-01 2.79698372e-01 3.09788674e-01 -3.31042200e-01
1.10629332e+00 3.02149117e-01 2.00920075e-01 -6.20496452e-01
1.33296117e-01 1.93009555e-01 -2.87445188e-01 -9.90744084e-02
3.74246866e-01 -6.51582897e-01 -8.01897235e-03 5.58538377e-01
-7.96423554e-02 -8.38838220e-01 -1.76581182e-02 -7.77980611e-02
4.02471095e-01 4.93673772e-01 1.20168030e-01 1.50944656e-02
4.32852119e-01 -5.16281068e-01 2.28386104e-01 1.13821268e-01
2.56051600e-01 1.21738112e+00 5.02984583e-01 -7.41783559e-01
-1.32430351e+00 -1.25867164e+00 2.15079799e-01 8.24697256e-01
8.89093801e-02 -2.62943089e-01 -9.69489813e-01 -3.74935538e-01
9.22106504e-02 3.39764744e-01 -6.74380124e-01 -1.14729859e-01
-7.89162338e-01 -5.96688269e-03 5.94508886e-01 6.89342558e-01
6.00492060e-01 -1.35876167e+00 -6.39027774e-01 1.78471804e-01
-1.06997430e-01 -1.01182556e+00 -1.15814042e+00 -6.93758547e-01
-6.89037025e-01 -9.07774627e-01 -1.44415653e+00 -7.25753486e-01
9.49538529e-01 -1.01580888e-01 1.15044129e+00 3.22794557e-01
-5.03948390e-01 1.74284935e-01 6.77261055e-02 -2.25014284e-01
-5.23820400e-01 -2.17122585e-01 -1.02265246e-01 1.36656806e-01
-4.78676260e-01 -6.97759688e-01 -9.13782418e-01 5.57775140e-01
-1.09577799e+00 6.09654367e-01 5.29368937e-01 8.85871172e-01
6.77535832e-01 -2.72594124e-01 1.74335986e-01 -6.90231144e-01
6.89598918e-01 1.37232035e-01 -4.70370144e-01 3.21981639e-01
4.44636717e-02 1.49298966e-01 5.40419340e-01 -6.38579488e-01
-1.06453300e+00 2.05593526e-01 -3.33239794e-01 -8.53630602e-01
-1.54788699e-02 -1.52940392e-01 -3.53989929e-01 -1.18532889e-01
4.30641770e-01 1.71345010e-01 2.49677837e-01 -3.04855019e-01
7.64485359e-01 3.59294474e-01 8.54450822e-01 -9.40330267e-01
1.02472174e+00 4.94369000e-01 4.34557535e-02 -7.35435605e-01
-3.97314399e-01 1.94405004e-01 -8.63536716e-01 -2.39997253e-01
8.43922675e-01 -7.16313481e-01 -5.55178940e-01 6.10266328e-01
-1.28725255e+00 -5.60374498e-01 -4.16592389e-01 -3.03202629e-01
-7.21665323e-01 2.09447101e-01 -6.52208984e-01 -5.98827779e-01
-5.90482771e-01 -1.10830212e+00 1.75473189e+00 2.81666428e-01
-3.63386571e-01 -5.63896596e-01 -1.47591904e-01 1.86985657e-01
4.09216553e-01 6.70223594e-01 8.07389557e-01 4.78237629e-01
-6.05917394e-01 -1.01646550e-01 -3.67442340e-01 6.69357479e-02
3.71883988e-01 3.22068036e-01 -7.01728821e-01 -2.17095658e-01
-7.07992852e-01 -2.92972505e-01 3.04576993e-01 2.40764409e-01
1.37318528e+00 -6.38975978e-01 -2.72225380e-01 8.96513641e-01
1.03029335e+00 5.91459125e-02 9.07834888e-01 -4.94628549e-01
9.19620574e-01 6.89224184e-01 2.81989157e-01 4.73979682e-01
3.00852388e-01 9.33311641e-01 1.10885009e-01 -4.80959743e-01
-6.13767266e-01 -9.65059400e-01 -5.41013032e-02 3.64708424e-01
-4.08747941e-01 4.44961041e-02 -4.52375501e-01 3.45403641e-01
-1.35462511e+00 -9.20587301e-01 2.46889189e-01 2.02791333e+00
9.36594129e-01 -3.14525396e-01 3.91898006e-01 -1.01257913e-01
6.74578071e-01 4.94246595e-02 -5.64350426e-01 -3.82614970e-01
1.39735848e-01 6.01279795e-01 5.61082736e-02 3.53342265e-01
-7.62568295e-01 1.11907911e+00 6.01931667e+00 6.54092252e-01
-1.29161954e+00 -9.88816917e-02 6.88465536e-01 -3.86588881e-03
-4.41579968e-01 -4.08621877e-01 -2.93103427e-01 4.14806545e-01
1.58626035e-01 4.07651551e-02 6.09128654e-01 6.14707053e-01
-1.10280961e-01 3.45659822e-01 -1.00015020e+00 1.37045562e+00
1.34241879e-01 -1.32726157e+00 5.89973688e-01 -2.34783050e-02
8.85487080e-01 -7.20571756e-01 8.91585574e-02 9.00629908e-03
2.35677928e-01 -1.27674782e+00 1.24696028e+00 6.83949232e-01
1.43247652e+00 -8.63641143e-01 1.58975184e-01 2.02320859e-01
-1.40780532e+00 2.99911469e-01 -3.48095953e-01 1.55922920e-01
3.44618410e-01 2.21671909e-01 -4.49515879e-01 1.54922768e-01
4.54368323e-01 2.62098730e-01 -3.13079357e-01 6.20021164e-01
-4.91649687e-01 -1.18995421e-01 -2.62637168e-01 1.60229072e-01
-6.75238855e-03 -3.44185472e-01 1.57277927e-01 1.05904675e+00
5.24729550e-01 4.90126401e-01 -1.92102537e-01 1.57010543e+00
-3.18426520e-01 -4.60728668e-02 -7.01385975e-01 -6.94474280e-02
4.45511073e-01 1.29971325e+00 -6.80423975e-01 -4.35122222e-01
8.41219500e-02 1.46614397e+00 8.53102803e-02 4.13108349e-01
-8.83184969e-01 -3.72604370e-01 6.92443371e-01 5.48310935e-01
2.44323850e-01 -2.53371388e-01 -1.91148862e-01 -1.03805912e+00
-1.51248090e-02 -8.97547424e-01 -1.63198695e-01 -1.22148156e+00
-1.10896027e+00 6.48695469e-01 -3.22496071e-02 -9.65594888e-01
-2.01867074e-01 -3.93308401e-01 -7.07655907e-01 1.02505064e+00
-9.71760094e-01 -1.69675827e+00 -4.76608872e-01 6.27922118e-01
6.21894836e-01 2.12953418e-01 6.45113647e-01 2.85592526e-01
-6.13050126e-02 8.12163830e-01 -7.15537965e-01 4.24810737e-01
6.76721990e-01 -8.76010001e-01 1.04008424e+00 4.36745942e-01
1.12939335e-01 6.84456348e-01 4.11337525e-01 -6.97075546e-01
-1.36592782e+00 -1.07361281e+00 6.04258776e-01 -4.41409409e-01
-2.08551157e-02 -5.52593291e-01 -6.76254451e-01 5.29694915e-01
7.04318807e-02 2.24749461e-01 1.94666147e-01 -4.80836362e-01
-4.37763810e-01 5.05915806e-02 -1.33763659e+00 9.75506485e-01
1.43527734e+00 -5.32264590e-01 -4.07239497e-01 -1.49694204e-01
3.54588419e-01 -7.86325097e-01 -8.35849285e-01 3.38474274e-01
1.25608051e+00 -8.71258020e-01 1.34490252e+00 -4.14839417e-01
7.80400932e-01 -3.83946657e-01 2.36377358e-01 -1.35098577e+00
-4.04535145e-01 -7.18246937e-01 9.38756540e-02 1.13522470e+00
4.82234433e-02 -2.36763775e-01 7.37659872e-01 8.13358724e-01
2.16378599e-01 -7.44718015e-01 -6.18706644e-01 -4.35170621e-01
1.30271047e-01 8.51238295e-02 1.04743230e+00 7.63200641e-01
-4.22574997e-01 4.46586385e-02 -6.77127719e-01 -2.49756813e-01
5.79391301e-01 4.87181455e-01 1.17076397e+00 -1.06006515e+00
-1.62034079e-01 -6.30103350e-01 -2.33025923e-01 -1.09284556e+00
6.46483451e-02 -5.47822356e-01 2.26351023e-02 -1.41988480e+00
5.04089519e-02 -4.71447051e-01 4.91331220e-01 5.50309896e-01
-1.43094093e-01 7.96187699e-01 5.58456898e-01 -7.38682002e-02
5.68657443e-02 6.91081464e-01 2.12077022e+00 -1.47651121e-01
-1.13517255e-01 -1.42861471e-01 -5.65913022e-01 6.72259092e-01
4.73361313e-01 -1.13204397e-01 -3.55260670e-01 -3.61179799e-01
1.06911816e-01 3.74617159e-01 6.55943930e-01 -7.77436793e-01
-7.40107819e-02 -3.08933765e-01 7.52530992e-01 -5.85052788e-01
6.62297368e-01 -6.55468404e-01 7.17007399e-01 5.06540596e-01
-2.33571544e-01 1.06528893e-01 1.50651082e-01 1.33099169e-01
-4.19659242e-02 1.83496788e-01 1.10899925e+00 -3.98872137e-01
-2.54665345e-01 6.97312474e-01 1.28804311e-01 -6.78535849e-02
6.62989676e-01 -3.46753418e-01 5.81837930e-02 -5.90487242e-01
-6.66566849e-01 -5.84704541e-02 7.63661742e-01 5.98638117e-01
9.49454367e-01 -1.74528134e+00 -7.77636051e-01 6.93824768e-01
-4.46415059e-02 2.09003031e-01 3.00259531e-01 3.85777652e-01
-8.09120536e-01 9.23813973e-03 -5.95910311e-01 -3.06590527e-01
-1.36332417e+00 4.75025654e-01 4.84782338e-01 1.41403794e-01
-7.80594826e-01 7.89400876e-01 7.51787126e-01 -6.20615363e-01
1.07755296e-01 -6.01209283e-01 3.07540148e-01 -3.07552308e-01
5.22725284e-01 1.60251603e-01 -4.17562932e-01 -7.63352752e-01
-8.48129541e-02 1.19557810e+00 5.65428972e-01 -2.25360155e-01
1.05148041e+00 2.22813189e-01 1.60092488e-02 -1.80471674e-01
9.22215879e-01 1.03696413e-01 -1.55271518e+00 1.08008564e-01
-7.35069692e-01 -6.95674419e-01 -3.95596743e-01 -8.48239243e-01
-1.45238554e+00 1.08241332e+00 2.68774867e-01 -4.29929912e-01
9.42212820e-01 -6.69451728e-02 9.75011051e-01 -6.16049916e-02
4.45574671e-01 -7.53910780e-01 2.34613180e-01 2.13712186e-01
1.68586981e+00 -8.10697496e-01 -8.68746638e-02 -5.90381861e-01
-5.84906697e-01 1.26611078e+00 6.57448411e-01 -1.57368869e-01
4.88330066e-01 3.41878533e-01 -6.95066601e-02 -2.51966208e-01
-1.85789421e-01 1.14881486e-01 7.03210056e-01 5.66345811e-01
3.39639872e-01 2.05287054e-01 -1.93270913e-03 6.03973269e-01
-5.57309031e-01 3.18072408e-01 6.86911419e-02 4.12512451e-01
7.52475411e-02 -1.17578602e+00 -6.12818420e-01 1.08609602e-01
-7.09846318e-02 1.84484780e-01 -7.41823554e-01 8.74934018e-01
4.56669986e-01 2.86326915e-01 1.07491150e-01 -1.32958680e-01
6.20813370e-01 -4.02135961e-02 1.04654121e+00 -4.96217996e-01
-3.77686650e-01 4.62116562e-02 -1.44262493e-01 -5.70754170e-01
-1.45372733e-01 -3.03367674e-01 -1.15340567e+00 -3.66703033e-01
6.32467940e-02 -3.95886868e-01 5.40339649e-01 4.68394041e-01
2.82845736e-01 4.85316783e-01 2.67865390e-01 -1.49122071e+00
-3.04374486e-01 -6.68195069e-01 -5.81035316e-01 6.36472583e-01
1.80394873e-01 -7.50028074e-01 1.50294572e-01 3.02636623e-01] | [12.041792869567871, -0.4376649260520935] |
e7b0e463-7b74-4634-ba25-13ea0daaf5f7 | smug-towards-robust-mri-reconstruction-by | 2303.12735 | null | https://arxiv.org/abs/2303.12735v1 | https://arxiv.org/pdf/2303.12735v1.pdf | SMUG: Towards robust MRI reconstruction by smoothed unrolling | Although deep learning (DL) has gained much popularity for accelerated magnetic resonance imaging (MRI), recent studies have shown that DL-based MRI reconstruction models could be oversensitive to tiny input perturbations (that are called 'adversarial perturbations'), which cause unstable, low-quality reconstructed images. This raises the question of how to design robust DL methods for MRI reconstruction. To address this problem, we propose a novel image reconstruction framework, termed SMOOTHED UNROLLING (SMUG), which advances a deep unrolling-based MRI reconstruction model using a randomized smoothing (RS)-based robust learning operation. RS, which improves the tolerance of a model against input noises, has been widely used in the design of adversarial defense for image classification. Yet, we find that the conventional design that applies RS to the entire DL process is ineffective for MRI reconstruction. We show that SMUG addresses the above issue by customizing the RS operation based on the unrolling architecture of the DL-based MRI reconstruction model. Compared to the vanilla RS approach and several variants of SMUG, we show that SMUG improves the robustness of MRI reconstruction with respect to a diverse set of perturbation sources, including perturbations to the input measurements, different measurement sampling rates, and different unrolling steps. Code for SMUG will be available at https://github.com/LGM70/SMUG. | ['Sijia Liu', 'Saiprasad Ravishankar', 'Yuguang Yao', 'Shijun Liang', 'Jinghan Jia', 'Hui Li'] | 2023-03-14 | null | null | null | null | ['adversarial-defense', 'unrolling', 'mri-reconstruction'] | ['adversarial', 'computer-vision', 'computer-vision'] | [ 3.05906951e-01 9.04006511e-02 2.20350191e-01 -1.51821837e-01
-1.17983556e+00 -5.16344905e-01 3.42454076e-01 -1.64872959e-01
-4.18578446e-01 4.97307748e-01 2.47019351e-01 -5.32933414e-01
1.36378957e-02 -5.37553668e-01 -9.03157830e-01 -1.05852365e+00
-2.62004852e-01 -1.55758828e-01 3.35955888e-01 -2.53580183e-01
7.91398883e-02 7.50524521e-01 -6.72484398e-01 1.59020379e-01
5.73846757e-01 7.92444646e-01 3.83647941e-02 7.11900711e-01
3.56293797e-01 7.46530235e-01 -6.77403212e-01 -8.23545083e-02
4.07906532e-01 -3.78881812e-01 -6.15818858e-01 -3.64548266e-01
2.21738592e-01 -4.21896607e-01 -6.84947431e-01 1.33801508e+00
1.10016084e+00 7.40793645e-02 3.65642399e-01 -9.98346150e-01
-5.16624987e-01 7.24841833e-01 -5.21395504e-01 4.66533393e-01
4.09838930e-02 1.87219754e-01 2.63337910e-01 -6.17657065e-01
4.79081154e-01 1.00705624e+00 9.28506792e-01 9.88766968e-01
-1.49792969e+00 -7.31421351e-01 -2.38812596e-01 3.79570238e-02
-1.24745631e+00 -4.50016886e-01 9.48068976e-01 -3.90786290e-01
3.12262923e-01 5.32108963e-01 6.56056851e-02 1.47320199e+00
7.46966362e-01 8.10835481e-01 1.44649649e+00 -1.93516403e-01
4.81458634e-01 -2.52302468e-01 -2.05007289e-02 4.47746187e-01
2.97953617e-02 3.40602338e-01 7.66152814e-02 -5.64961910e-01
1.02625740e+00 -2.35578835e-01 -7.00703502e-01 -2.59529531e-01
-1.13958395e+00 8.79575193e-01 6.03431702e-01 3.09991986e-01
-2.57694602e-01 2.62416989e-01 7.69412816e-01 3.96452010e-01
3.28018606e-01 5.90752363e-01 -3.60470146e-01 2.35558152e-01
-6.68712318e-01 1.83027908e-01 4.75027502e-01 4.01675135e-01
1.42577961e-01 5.10491073e-01 -2.08986387e-01 7.13233173e-01
-8.74445736e-02 4.88941252e-01 9.16919231e-01 -9.05639112e-01
2.88608909e-01 -2.33887583e-01 -2.38934204e-01 -1.06935489e+00
-7.18473136e-01 -6.06841207e-01 -1.17642105e+00 2.35372066e-01
3.01394731e-01 -4.09271419e-01 -8.15883815e-01 1.98133230e+00
3.28595817e-01 4.44144219e-01 -6.22966513e-02 1.04647744e+00
8.44889879e-01 3.39835912e-01 -8.92963409e-02 -1.53270423e-01
1.00994074e+00 -5.36328554e-01 -7.39700377e-01 -1.74691081e-01
3.19761574e-01 -5.84817231e-01 1.26370192e+00 4.37172979e-01
-9.53333020e-01 -3.46498907e-01 -1.20637703e+00 2.57447511e-01
1.47419879e-02 -3.92902762e-01 3.94060493e-01 8.99357319e-01
-1.05474496e+00 6.75273776e-01 -1.07634115e+00 1.98560774e-01
3.34635407e-01 3.51693690e-01 -4.69001710e-01 4.71632667e-02
-1.41737199e+00 8.69431973e-01 -7.35647678e-02 1.98122978e-01
-1.16846025e+00 -8.36443365e-01 -8.71522486e-01 -3.17363352e-01
2.99802095e-01 -4.49632078e-01 1.17509186e+00 -7.68445969e-01
-1.69012368e+00 6.70447886e-01 2.26593390e-01 -6.36077106e-01
7.83775210e-01 -6.63492084e-02 -4.76582319e-01 3.05916518e-01
-1.78469807e-01 1.65250883e-01 1.26044428e+00 -1.34542203e+00
5.14669120e-01 -3.46299648e-01 -9.67405736e-02 -3.06022704e-01
1.48094311e-01 2.06256211e-01 -7.93723539e-02 -1.22468519e+00
2.35135794e-01 -1.11189151e+00 -7.49870598e-01 -1.62253469e-01
-6.16588950e-01 5.40267110e-01 5.24223626e-01 -7.82771170e-01
1.02513671e+00 -2.26511359e+00 6.75312281e-02 2.94232756e-01
2.85452068e-01 4.42982256e-01 -2.74805456e-01 1.04018047e-01
-5.38235247e-01 8.48842114e-02 -5.64414740e-01 -1.48864672e-01
-2.16227978e-01 2.22158313e-01 -5.35777450e-01 9.42366481e-01
-2.22150877e-01 9.69092369e-01 -9.19089079e-01 -1.18272915e-01
1.74909830e-01 5.48863292e-01 -6.49046481e-01 2.49639571e-01
2.62474686e-01 1.02973783e+00 -5.19660413e-01 4.62978840e-01
9.06302571e-01 8.56855363e-02 1.16031215e-01 -4.01967496e-01
3.38826984e-01 -5.08913137e-02 -9.86684680e-01 1.38513446e+00
-4.01149333e-01 4.58219022e-01 3.62368047e-01 -1.16197097e+00
6.85077429e-01 3.35357487e-01 6.10601842e-01 -6.63400590e-01
3.39171499e-01 2.26688728e-01 -1.20044295e-02 -5.41045487e-01
9.88469049e-02 -4.95415688e-01 -3.26874793e-01 4.06672060e-01
-1.27821326e-01 -2.99980521e-01 -5.92930198e-01 3.35647374e-01
1.38001013e+00 -2.50973552e-01 2.55923718e-01 -3.66590351e-01
4.91164207e-01 -5.06559730e-01 7.53465772e-01 1.08123970e+00
-5.17969489e-01 1.02035069e+00 4.20411915e-01 -3.37073743e-01
-9.20509219e-01 -1.35028028e+00 -4.43068326e-01 5.75415850e-01
6.59024715e-02 -1.38282046e-01 -9.76069927e-01 -6.81622267e-01
-8.21818262e-02 5.16714156e-01 -7.07135856e-01 -5.55250585e-01
-8.87454927e-01 -1.16614151e+00 9.95747507e-01 3.61997634e-01
4.33102310e-01 -9.00174737e-01 -3.58979374e-01 2.45400652e-01
-2.87505567e-01 -1.07796395e+00 -6.39735699e-01 1.37134075e-01
-8.57873857e-01 -9.60593224e-01 -6.94228232e-01 -2.15376645e-01
7.60644138e-01 -1.96810476e-02 7.38798022e-01 2.13975143e-02
-2.75205791e-01 3.82981330e-01 -3.31302583e-01 -2.40387991e-01
-9.44103062e-01 -2.88388580e-01 3.87023121e-01 -2.69587375e-02
-3.07761550e-01 -8.17063749e-01 -7.61927366e-01 3.09324563e-01
-1.47815323e+00 -2.46142805e-01 3.13713908e-01 9.70202804e-01
5.91500223e-01 -9.27916244e-02 7.61479795e-01 -9.47048783e-01
8.13408852e-01 -6.03420556e-01 -3.93621624e-01 -1.17073111e-01
-2.74888426e-01 2.34553054e-01 1.01954532e+00 -8.24832618e-01
-5.28120160e-01 -2.24287719e-01 -8.55985045e-01 -5.70051730e-01
2.88786516e-02 3.28405708e-01 -9.73992273e-02 -6.92100346e-01
9.55284238e-01 2.86101758e-01 3.20731014e-01 -5.80393374e-01
2.25450113e-01 4.02104825e-01 7.38938570e-01 -5.81959069e-01
8.62290263e-01 6.05072141e-01 1.13751506e-02 -6.52719796e-01
-6.21266127e-01 5.06696291e-02 -1.44323945e-01 -2.24547148e-01
5.77768862e-01 -5.58893979e-01 -4.63164598e-01 6.62970841e-01
-7.67696798e-01 -5.88211536e-01 -2.14837700e-01 4.59906459e-01
-7.40363955e-01 7.26355135e-01 -9.68188643e-01 -3.19111049e-01
-5.12095571e-01 -1.69462669e+00 7.88878322e-01 -1.32301599e-01
-1.21483088e-01 -8.41446459e-01 -3.01598422e-02 1.92127660e-01
7.43499279e-01 7.02177644e-01 8.95128846e-01 -6.95447743e-01
-2.24634171e-01 -3.31419557e-01 1.67402238e-01 7.80399799e-01
-4.03990224e-02 -5.37135780e-01 -8.98191631e-01 -6.99770451e-01
7.96477437e-01 -2.91306764e-01 5.17215312e-01 6.06642902e-01
1.60998642e+00 -5.24348319e-01 7.47372629e-03 1.09652960e+00
1.37494385e+00 3.39170843e-02 9.99110579e-01 5.68889141e-01
6.46001756e-01 3.90825495e-02 1.00527011e-01 2.45823815e-01
-1.44625381e-01 7.46921718e-01 4.56882477e-01 -2.40424111e-01
-1.28730610e-01 -5.24322465e-02 5.44378519e-01 8.94114077e-01
2.19850257e-01 2.29727700e-01 -8.20909739e-01 1.79996088e-01
-1.45295262e+00 -7.54193187e-01 1.24181129e-01 2.26065063e+00
9.43598986e-01 1.32844061e-01 -1.91679135e-01 1.47137925e-01
6.22692168e-01 3.59502524e-01 -7.49077916e-01 -3.79833758e-01
-4.02352512e-02 3.13946187e-01 8.59008610e-01 5.66610157e-01
-1.05845153e+00 6.29216611e-01 6.81644678e+00 1.06994689e+00
-1.56318605e+00 5.47415376e-01 6.15615964e-01 5.87792136e-02
-3.13069046e-01 -3.10073435e-01 -1.95741385e-01 6.90254331e-01
8.20309520e-01 -1.34330750e-01 5.69615602e-01 7.48503149e-01
4.41319853e-01 2.10406423e-01 -6.48966730e-01 8.67330432e-01
1.27660215e-01 -1.30268896e+00 -9.11829397e-02 -2.03486845e-01
6.18835449e-01 2.38225505e-01 2.06736341e-01 2.49656945e-01
2.32360676e-01 -9.46628630e-01 5.40449142e-01 3.56900871e-01
8.16516280e-01 -6.94112003e-01 7.06055462e-01 3.76586050e-01
-5.66270113e-01 -3.40514593e-02 -3.20122302e-01 4.48827237e-01
3.11078668e-01 7.23869562e-01 -5.13802648e-01 4.96526480e-01
5.93100131e-01 -3.07102073e-02 -5.01939237e-01 8.40904772e-01
-3.09181601e-01 8.14039528e-01 -1.53727412e-01 6.11308992e-01
1.07215405e-01 9.82516855e-02 1.06207597e+00 1.01766729e+00
1.85918212e-02 1.08470529e-01 9.98165831e-02 6.53450966e-01
3.81463841e-02 -1.76847160e-01 -6.14102006e-01 4.77877557e-01
3.15610588e-01 1.04156089e+00 -5.45442462e-01 -8.51636231e-02
-4.55092601e-02 9.00785446e-01 1.11506611e-01 4.17491943e-01
-9.98553216e-01 -2.28969470e-01 6.46835327e-01 2.23718911e-01
-4.14600782e-02 -3.28809589e-01 -2.42767707e-01 -1.32239544e+00
-6.09290637e-02 -1.38770950e+00 2.83961177e-01 -6.53126061e-01
-1.36184800e+00 8.87169421e-01 -2.59837341e-02 -1.20738316e+00
-1.18806690e-01 -3.49383652e-01 -6.23620987e-01 5.64009786e-01
-1.19553435e+00 -8.16815257e-01 7.90180080e-03 8.20305347e-01
2.79879987e-01 -3.31908017e-02 6.85516119e-01 3.76640975e-01
-5.66882730e-01 8.53867948e-01 2.82832086e-01 2.17644036e-01
6.33663177e-01 -1.05064011e+00 3.87182385e-01 1.14969337e+00
-2.27305725e-01 7.64097571e-01 1.03112960e+00 -5.89045882e-01
-1.46214271e+00 -1.15785396e+00 1.23338535e-01 -2.39218757e-01
8.60913157e-01 -3.75505000e-01 -1.14125943e+00 7.59334266e-01
-1.93351880e-01 4.61580783e-01 5.45460641e-01 -5.58371961e-01
-2.44371220e-01 -1.19877167e-01 -1.61445951e+00 9.07397389e-01
8.22685003e-01 -6.94688201e-01 -6.23699963e-01 3.68952513e-01
8.40345740e-01 -7.49653041e-01 -9.67650294e-01 7.12467253e-01
3.13417405e-01 -7.36957371e-01 1.38502192e+00 -4.55847412e-01
1.82490140e-01 -1.80803508e-01 -1.60425290e-01 -1.56945074e+00
-3.19871217e-01 -9.10110354e-01 -5.74479513e-02 5.97267926e-01
2.98513770e-02 -9.60081398e-01 4.89281774e-01 6.34300232e-01
-2.55417556e-01 -8.47814322e-01 -1.27082610e+00 -9.28306162e-01
4.38497096e-01 -5.92848957e-01 5.28355241e-01 1.00226367e+00
-1.96598366e-01 -2.37462997e-01 -5.44469476e-01 5.35923004e-01
7.90121615e-01 -2.90157259e-01 5.37768543e-01 -4.23299223e-01
-5.52977145e-01 -2.24804893e-01 -5.08690119e-01 -7.27058291e-01
1.32455736e-01 -9.48323190e-01 1.50105223e-01 -9.60573435e-01
-1.87648535e-01 -5.24301112e-01 -4.09438491e-01 2.83835888e-01
-4.27533954e-01 4.14079249e-01 3.26259017e-01 1.61453858e-01
-7.93927684e-02 4.15658444e-01 1.31008053e+00 -2.27572005e-02
2.74353884e-02 1.03272781e-01 -6.92300856e-01 7.52279341e-01
9.83298302e-01 -6.07507646e-01 -1.85695276e-01 -3.76361459e-01
-1.22384153e-01 2.51912802e-01 5.26480258e-01 -1.04213881e+00
1.00137353e-01 9.20708477e-02 1.19616203e-01 -7.31964633e-02
5.51220551e-02 -6.69950366e-01 3.62788647e-01 7.35345423e-01
-3.93185467e-01 8.64249095e-02 2.93709815e-01 3.34934503e-01
3.32173966e-02 -2.69265175e-01 1.25090361e+00 -2.00016975e-01
-3.52672368e-01 3.49424720e-01 -5.01289785e-01 5.11018857e-02
7.59045362e-01 1.71649456e-01 -2.38763243e-01 -3.72266144e-01
-9.28812385e-01 -2.57106721e-01 3.10003877e-01 2.31479973e-01
7.57814467e-01 -1.24924350e+00 -6.49840891e-01 4.26928401e-01
-3.64204019e-01 -1.72045261e-01 5.32118678e-01 1.08680117e+00
-4.48374450e-01 -1.36558712e-01 -6.80330172e-02 -4.51120526e-01
-9.81744230e-01 8.23619008e-01 6.48332179e-01 -3.97826672e-01
-1.07933319e+00 7.67766178e-01 2.50315934e-01 -7.47407377e-01
1.43631682e-01 -1.80693075e-01 1.53972358e-01 -4.93619353e-01
7.65723944e-01 1.83056220e-01 3.24891150e-01 -4.33669269e-01
-3.44975203e-01 2.91450232e-01 1.81302968e-02 -1.40291285e-02
1.27651393e+00 -2.88017225e-02 4.08321731e-02 2.33268738e-01
1.39160323e+00 3.18285227e-01 -1.19133234e+00 -1.66927367e-01
-2.88163453e-01 -4.60042715e-01 2.96479076e-01 -5.94336629e-01
-1.42106628e+00 5.73257208e-01 6.03406250e-01 6.08903989e-02
1.21575499e+00 -2.50323474e-01 1.06728959e+00 -5.90489767e-02
7.02664256e-01 -5.97617984e-01 6.94893003e-02 3.07039022e-01
1.29713070e+00 -9.67994452e-01 -6.08290359e-02 -1.08737983e-01
-6.48986578e-01 7.57414579e-01 3.27593893e-01 -3.23287040e-01
6.91721201e-01 7.48008430e-01 3.22783381e-01 -8.90192464e-02
-2.14127153e-01 2.93108970e-01 -1.09520089e-03 6.16293490e-01
-2.30926499e-02 1.69382766e-01 -3.65461975e-01 6.92631185e-01
-7.51452297e-02 -5.00658900e-02 6.32708907e-01 9.30185854e-01
9.05961916e-03 -1.11126566e+00 -8.00587714e-01 3.66792500e-01
-9.25481856e-01 -1.06911696e-01 1.84599355e-01 5.42112589e-01
-1.48831040e-01 5.84897995e-01 -4.02358890e-01 -4.97678071e-01
2.74754971e-01 -3.11577231e-01 4.64499503e-01 -2.62576103e-01
-6.44262791e-01 -4.12461013e-02 -1.06716186e-01 -8.94591510e-01
-1.15863748e-01 -4.50538129e-01 -1.15946627e+00 -3.63019526e-01
-1.72166198e-01 2.77631935e-02 5.62240064e-01 8.40057075e-01
2.53054529e-01 6.28122091e-01 9.35373664e-01 -1.00018001e+00
-1.12055349e+00 -6.39507592e-01 -6.77897394e-01 5.85089505e-01
6.31034970e-01 -5.55200219e-01 -6.18977308e-01 -3.32027018e-01] | [13.53173828125, -2.2964131832122803] |
0d4e1569-17f9-4b7c-aec1-aaa9145c0686 | candid-correspondence-alignment-for-deep | 2306.09887 | null | https://arxiv.org/abs/2306.09887v1 | https://arxiv.org/pdf/2306.09887v1.pdf | CANDID: Correspondence AligNment for Deep-burst Image Denoising | With the advent of mobile phone photography and point-and-shoot cameras, deep-burst imaging is widely used for a number of photographic effects such as depth of field, super-resolution, motion deblurring, and image denoising. In this work, we propose to solve the problem of deep-burst image denoising by including an optical flow-based correspondence estimation module which aligns all the input burst images with respect to a reference frame. In order to deal with varying noise levels the individual burst images are pre-filtered with different settings. Exploiting the established correspondences one network block predicts a pixel-wise spatially-varying filter kernel to smooth each image in the original and prefiltered bursts before fusing all images to generate the final denoised output. The resulting pipeline achieves state-of-the-art results by combining all available information provided by the burst. | ['Hendrik PA Lensch', 'Raphael Braun', 'Arijit Mallick'] | 2023-06-16 | null | null | null | null | ['deblurring', 'optical-flow-estimation', 'image-denoising', 'super-resolution'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.33774596e-01 -3.34762841e-01 3.13249052e-01 -2.36105040e-01
-5.50036609e-01 -3.56005698e-01 4.48920429e-01 -2.56223440e-01
-7.51337409e-01 5.83707631e-01 3.32904607e-01 2.19360948e-01
1.22655623e-01 -5.45820355e-01 -6.04208231e-01 -7.66778529e-01
3.78980726e-01 -1.07810885e-01 5.46001017e-01 -1.06487028e-01
2.13249773e-01 6.15898967e-01 -1.43613183e+00 3.20352972e-01
7.74791300e-01 8.74500096e-01 5.06575763e-01 8.63455236e-01
1.95140511e-01 9.59221601e-01 -2.76477247e-01 -2.85426825e-01
5.04157484e-01 -5.48717976e-01 -4.03327554e-01 3.22731614e-01
8.94402087e-01 -9.84795451e-01 -8.81121814e-01 1.20150149e+00
3.20034444e-01 4.65494752e-01 4.66856025e-02 -7.34576583e-01
-2.37226382e-01 -2.64951419e-02 -4.66204703e-01 7.03803897e-01
2.53330886e-01 5.41839361e-01 3.70048314e-01 -7.40768671e-01
1.05046868e+00 1.10787225e+00 6.34664357e-01 2.80537575e-01
-1.40439487e+00 -3.03241581e-01 -1.36578247e-01 4.23801780e-01
-9.70412970e-01 -6.40236735e-01 6.28942132e-01 -3.46410602e-01
7.73926973e-01 3.00269132e-03 6.93541348e-01 7.55052567e-01
3.35448056e-01 1.29428655e-01 9.58358347e-01 -2.68999994e-01
7.65881091e-02 -3.48433197e-01 -2.07189187e-01 4.39131230e-01
2.89020035e-02 1.72322959e-01 -7.56982327e-01 2.31066123e-01
1.15291965e+00 1.18482761e-01 -6.49146795e-01 8.24236274e-02
-1.25930488e+00 3.12254220e-01 4.37513620e-01 3.50338727e-01
-7.64081538e-01 3.73244554e-01 -3.22312564e-02 1.12458140e-01
6.69927359e-01 2.23612934e-01 -1.08925447e-01 -1.32046863e-01
-1.21037531e+00 2.17092335e-01 3.49358499e-01 4.14533138e-01
9.71904516e-01 1.61471069e-01 -2.09602267e-02 6.23319745e-01
6.99999626e-04 3.69361877e-01 3.57858509e-01 -1.68078375e+00
3.33733469e-01 6.38027787e-02 3.91467810e-01 -1.05900812e+00
-3.39475244e-01 -3.55692446e-01 -9.81046855e-01 4.10344839e-01
4.93400931e-01 9.35183838e-02 -8.54825020e-01 1.34152722e+00
5.80671847e-01 8.25607300e-01 -1.44762874e-01 1.23712802e+00
5.27967215e-01 8.13489676e-01 -1.83676422e-01 -4.09535438e-01
1.09701514e+00 -8.30896080e-01 -9.04253781e-01 -4.58872765e-01
1.53948933e-01 -9.00588214e-01 5.46270609e-01 6.41970813e-01
-1.43927193e+00 -6.48816228e-01 -8.00298035e-01 -5.61121345e-01
2.24652395e-01 -1.31925240e-01 -7.57041201e-02 8.75211507e-02
-1.08094299e+00 8.65633368e-01 -1.02739012e+00 -9.96348485e-02
3.34347874e-01 9.49856415e-02 -6.48579717e-01 -5.41132450e-01
-8.50468874e-01 1.05780292e+00 6.95519075e-02 2.91885108e-01
-8.94835472e-01 -9.80701685e-01 -7.68713236e-01 4.86142002e-02
2.61329144e-01 -9.07719493e-01 9.32200968e-01 -8.81364048e-01
-1.61286283e+00 7.15920389e-01 -3.27360004e-01 -5.50139844e-01
6.55113459e-01 -3.21226895e-01 -1.41050741e-01 5.82535625e-01
6.40745014e-02 4.66466874e-01 1.08081543e+00 -9.40920353e-01
-6.58028841e-01 -1.23927727e-01 -2.04468802e-01 3.19656342e-01
1.22078180e-01 5.38115948e-02 -6.36130273e-01 -7.39416420e-01
2.25779191e-01 -6.73755825e-01 -3.32155854e-01 2.26155877e-01
-1.32456779e-01 6.69851780e-01 9.44367826e-01 -1.11842275e+00
9.47653055e-01 -2.29213524e+00 3.99683475e-01 -3.46339971e-01
4.68951911e-01 4.00276273e-01 -3.65098864e-01 2.46551171e-01
-5.87102063e-02 -5.17860532e-01 -2.55248696e-01 -8.31814766e-01
-8.67965758e-01 2.04914808e-01 -2.05009028e-01 8.57221723e-01
6.18930012e-02 7.58782446e-01 -1.02517498e+00 -1.92822307e-01
9.52459455e-01 7.21612215e-01 -5.12690067e-01 3.48645926e-01
-4.94017042e-02 1.04537821e+00 3.22167389e-02 1.80968016e-01
1.01464236e+00 2.69774012e-02 -1.48220941e-01 -5.71040511e-01
-4.69911814e-01 4.34270538e-02 -1.31647241e+00 1.91328287e+00
-6.22084200e-01 1.17817867e+00 4.88789529e-01 -6.54549003e-01
5.60830414e-01 5.29319420e-02 7.25331128e-01 -6.95213616e-01
1.32594764e-01 1.39316320e-01 -2.45715082e-01 -6.32893205e-01
7.25415230e-01 -8.73156860e-02 5.97453296e-01 2.31893420e-01
6.30801022e-02 -3.52224588e-01 1.59428827e-02 1.67703301e-01
1.20755017e+00 3.07714306e-02 5.03359549e-02 1.98845789e-01
6.40360177e-01 -1.79606870e-01 5.09398401e-01 5.13869703e-01
-4.74550165e-02 1.36201680e+00 2.03538373e-01 -6.22965932e-01
-1.41813707e+00 -9.11140025e-01 1.61748081e-01 4.52027321e-01
3.98709655e-01 -3.14098001e-01 -8.05470288e-01 -8.55999887e-02
-2.62413621e-01 5.90403080e-01 -4.02863920e-01 -7.34134689e-02
-8.54290366e-01 -2.64629155e-01 2.04221040e-01 2.68822983e-02
8.09600770e-01 -6.93738759e-01 -8.81151438e-01 5.38203418e-01
-6.00615323e-01 -1.62104940e+00 -6.00790143e-01 -1.08217239e-01
-7.68957615e-01 -9.78314042e-01 -7.33476222e-01 -3.49143565e-01
4.96942818e-01 6.90074146e-01 8.90667677e-01 3.70860063e-02
-3.75783414e-01 -2.47514457e-03 -1.13782093e-01 3.98146838e-01
-5.84050298e-01 -4.54661340e-01 -2.05723390e-01 5.93140364e-01
-4.77636248e-01 -8.30825865e-01 -9.70443904e-01 2.71117300e-01
-1.42374885e+00 2.91361481e-01 3.00964803e-01 8.51346314e-01
5.06229460e-01 6.19665198e-02 -1.41705558e-01 -4.59868878e-01
2.80506730e-01 -1.72683194e-01 -8.23686421e-01 -1.31862342e-01
-1.65733591e-01 -1.16444945e-01 5.99169135e-01 -5.93863368e-01
-1.24542737e+00 2.02814743e-01 2.10120180e-03 -8.20949912e-01
-7.84483477e-02 2.07305133e-01 5.23142032e-02 -1.83103874e-01
6.63923562e-01 2.70199478e-01 3.74954492e-01 -4.64108616e-01
5.50998390e-01 3.42267603e-01 1.23454487e+00 2.26894140e-01
6.62763596e-01 1.12258112e+00 2.65057802e-01 -1.14860427e+00
-8.74588966e-01 -6.92566752e-01 -6.92470193e-01 -4.45462406e-01
1.15407014e+00 -1.06930864e+00 -4.96250898e-01 9.94551301e-01
-1.49092019e+00 -5.47240078e-01 -4.51227695e-01 4.85761911e-01
-4.73641247e-01 6.19767547e-01 -7.89371312e-01 -2.95363098e-01
-1.47578821e-01 -1.33989108e+00 1.12484241e+00 4.34740305e-01
-4.26330604e-02 -7.66415358e-01 1.24115564e-01 2.33395293e-01
4.32941854e-01 -1.43854227e-02 3.46800238e-01 3.91161293e-01
-9.98932898e-01 -1.34658337e-01 -1.83786079e-01 8.45759332e-01
1.82498638e-02 -2.23976716e-01 -9.04848576e-01 -2.25900095e-02
3.70651752e-01 2.66112626e-01 8.84703040e-01 8.08413744e-01
7.66484499e-01 -8.95161480e-02 1.46404639e-01 1.29654205e+00
1.45757747e+00 -1.25697866e-01 1.23521173e+00 4.34730440e-01
9.55335379e-01 5.12105107e-01 3.27092052e-01 2.23876089e-01
2.03469589e-01 9.79663908e-01 5.74498117e-01 -1.62356570e-01
-6.66444182e-01 9.95668173e-02 2.79786587e-01 1.74131289e-01
-1.95018630e-02 -3.02617878e-01 -5.03364265e-01 4.97815579e-01
-1.79837584e+00 -1.11681592e+00 -4.68685001e-01 2.29362369e+00
6.26534939e-01 -3.62189263e-01 -2.59002090e-01 -1.35547549e-01
5.81142068e-01 4.14456338e-01 -4.32072818e-01 4.81145680e-02
-4.81897980e-01 9.32362452e-02 7.08625436e-01 1.07539880e+00
-8.32032800e-01 9.52441692e-01 5.45861149e+00 6.33018792e-01
-1.40438795e+00 3.06202978e-01 6.30918443e-01 -2.38717780e-01
-6.60677850e-02 2.11103797e-01 -3.78968149e-01 5.98703444e-01
8.41258407e-01 -4.86300774e-02 6.96031034e-01 1.90883875e-01
7.14486718e-01 -8.20375800e-01 -7.02325106e-01 1.35121298e+00
7.57799223e-02 -1.69464386e+00 -2.65888005e-01 2.65813973e-02
1.01450562e+00 2.12756231e-01 -1.31741792e-01 -6.36389852e-01
-1.22853054e-03 -8.50501180e-01 7.91307151e-01 8.00034881e-01
7.19211757e-01 -4.76370990e-01 5.19705296e-01 3.79902869e-01
-7.37337589e-01 -7.18741119e-02 -2.08942607e-01 -3.98498885e-02
7.69759357e-01 1.08051431e+00 -3.93403053e-01 5.90116918e-01
7.53030539e-01 1.07841575e+00 -2.48363048e-01 1.09080458e+00
-2.73791909e-01 2.58208513e-01 -3.93883407e-01 8.84906471e-01
1.64864808e-02 -6.28195643e-01 7.48000860e-01 9.10717487e-01
4.79826719e-01 3.38340759e-01 -2.71352947e-01 8.11071098e-01
1.88406453e-01 -6.71122968e-01 -3.39506954e-01 5.43712258e-01
1.38845429e-01 1.31325245e+00 -6.51769280e-01 -5.03402650e-01
-3.70770544e-01 1.58812284e+00 -1.25013039e-01 6.45705044e-01
-4.92876053e-01 9.71654709e-03 9.08515036e-01 5.50221086e-01
4.47884500e-01 -5.17646253e-01 -2.39152551e-01 -1.40198708e+00
1.80770066e-02 -6.12872720e-01 -1.57968383e-02 -1.26262009e+00
-8.03646982e-01 6.39220774e-01 9.08998423e-04 -1.20040190e+00
-2.99664438e-01 -2.69134820e-01 -6.67896032e-01 1.18372595e+00
-1.57897639e+00 -7.99634755e-01 -8.07862937e-01 5.63979149e-01
5.18873990e-01 2.93580860e-01 1.42547414e-01 4.90143389e-01
-3.75246823e-01 -2.42358625e-01 3.09308976e-01 -2.00264573e-01
9.16200459e-01 -7.19142437e-01 5.19003510e-01 1.53711247e+00
-1.03410773e-01 2.50128448e-01 8.85174930e-01 -7.23944545e-01
-1.21166253e+00 -1.11513281e+00 6.40496016e-01 -1.82883069e-01
5.00614226e-01 5.02851121e-02 -1.09671187e+00 4.98975158e-01
1.26337498e-01 7.22807884e-01 -3.46490234e-01 -1.00505626e+00
-6.43882602e-02 -3.64100814e-01 -9.62707698e-01 3.78579825e-01
8.05512249e-01 -5.41430950e-01 -6.89366758e-02 2.86465120e-02
4.37772751e-01 -7.74911642e-01 -4.42814022e-01 -8.90097618e-02
3.40764105e-01 -1.49040818e+00 1.29782426e+00 1.10684209e-01
6.78743422e-01 -5.69328427e-01 1.65400207e-01 -1.48994088e+00
-2.06561640e-01 -1.14998460e+00 1.18749859e-02 9.02049780e-01
-4.31196123e-01 -6.68938935e-01 6.59836352e-01 4.26512986e-01
-2.74277866e-01 -1.85717195e-01 -1.12247896e+00 -3.16890597e-01
-5.24645686e-01 -3.25526834e-01 6.36388082e-03 6.80226386e-01
-6.78353012e-01 9.04149711e-02 -6.82368755e-01 2.12979764e-01
9.87335026e-01 -2.81094402e-01 8.35394025e-01 -8.13375533e-01
-2.96517551e-01 -1.59741417e-01 -6.12812400e-01 -1.37200105e+00
3.60348411e-02 -4.49333310e-01 1.90248564e-01 -1.45777261e+00
-2.77244002e-01 2.43591368e-02 3.51217806e-01 -1.27852812e-01
-2.74832696e-01 5.46997905e-01 3.29587311e-01 3.23217571e-01
-2.26019263e-01 3.52428287e-01 1.19106555e+00 2.51381278e-01
-3.67037743e-01 -1.94480896e-01 -1.72972754e-01 6.56383038e-01
3.52077067e-01 -3.65888357e-01 -1.82202309e-01 -8.24406087e-01
-6.30834932e-03 4.34108287e-01 8.12009037e-01 -9.84309971e-01
5.39881647e-01 -6.56318590e-02 3.92888635e-01 -4.20688778e-01
6.72154069e-01 -7.86876500e-01 6.38915777e-01 3.29892397e-01
-1.74141303e-01 7.22936615e-02 -4.44721207e-02 6.26092732e-01
-3.99080276e-01 -5.87080903e-02 1.21070385e+00 -1.08019717e-01
-7.31081665e-01 2.65602648e-01 -4.02872026e-01 2.88217328e-02
8.18684876e-01 -3.87134343e-01 -4.12807912e-01 -6.74340606e-01
-5.05989492e-01 -1.62894964e-01 7.21191585e-01 4.15846258e-02
6.86377168e-01 -7.81638622e-01 -7.64923632e-01 3.73540193e-01
-4.71251935e-01 3.75666827e-01 8.91207159e-01 1.08716059e+00
-9.77025926e-01 -1.73728645e-01 -3.74459058e-01 -7.10526347e-01
-1.15045857e+00 2.77209818e-01 7.26115406e-01 -1.68516174e-01
-1.14926863e+00 8.30245972e-01 2.95716316e-01 3.82912189e-01
-1.50430845e-02 -4.55357254e-01 1.52162433e-01 -8.78462419e-02
8.77421916e-01 5.93733668e-01 3.01580787e-01 -8.06125700e-01
-8.78442898e-02 7.15167463e-01 2.14887097e-01 -4.57870752e-01
1.63286495e+00 -6.13945842e-01 -2.62977034e-01 1.34283140e-01
1.40455616e+00 5.66908307e-02 -1.99081838e+00 -2.45856896e-01
-3.97723347e-01 -9.60814476e-01 5.76671958e-01 -2.93405354e-01
-1.38098717e+00 7.31613398e-01 5.37275195e-01 -2.06131086e-01
1.39676487e+00 -3.06835204e-01 9.89867449e-01 -2.09608376e-01
7.29996758e-03 -7.32058764e-01 1.00452423e-01 3.12084317e-01
6.63286984e-01 -9.66016173e-01 3.39978278e-01 -5.44645250e-01
-3.50198030e-01 1.29640329e+00 1.76852584e-01 -4.03208762e-01
2.77001709e-01 3.69034588e-01 1.47116467e-01 1.27024159e-01
-5.76168180e-01 -9.32517797e-02 2.50964135e-01 7.40728498e-01
-1.00003779e-01 -6.17989719e-01 -8.79671238e-03 -1.70683503e-01
1.15148082e-01 1.99337810e-01 1.05871511e+00 5.21291852e-01
-2.42143750e-01 -9.50388908e-01 -6.20002091e-01 2.04973873e-02
-2.42423281e-01 -2.77094185e-01 2.40136310e-01 3.71321082e-01
1.39391541e-01 9.30900097e-01 2.78761119e-01 -7.15259910e-02
4.42445576e-01 -4.37851340e-01 5.92994571e-01 -3.51899475e-01
-6.59091353e-01 2.67849237e-01 -6.65164366e-02 -1.11368811e+00
-5.30612350e-01 -7.36950397e-01 -9.52192724e-01 -2.68339276e-01
6.26266748e-02 -4.56419736e-01 6.60053730e-01 7.01541424e-01
4.10167247e-01 3.69423568e-01 3.89930636e-01 -1.51765990e+00
-6.76270053e-02 -7.51171887e-01 -4.79011804e-01 7.01923847e-01
9.15792704e-01 -2.86414564e-01 -7.01538026e-01 5.46693146e-01] | [10.892706871032715, -1.7559303045272827] |
8bff86b6-8687-45ae-86f6-b988597b2bac | clustering-noisy-signals-with-structured | 1510.05214 | null | http://arxiv.org/abs/1510.05214v1 | http://arxiv.org/pdf/1510.05214v1.pdf | Clustering Noisy Signals with Structured Sparsity Using Time-Frequency Representation | We propose a simple and efficient time-series clustering framework
particularly suited for low Signal-to-Noise Ratio (SNR), by simultaneous
smoothing and dimensionality reduction aimed at preserving clustering
information. We extend the sparse K-means algorithm by incorporating structured
sparsity, and use it to exploit the multi-scale property of wavelets and group
structure in multivariate signals. Finally, we extract features invariant to
translation and scaling with the scattering transform, which corresponds to a
convolutional network with filters given by a wavelet operator, and use the
network's structure in sparse clustering. By promoting sparsity, this transform
can yield a low-dimensional representation of signals that gives improved
clustering results on several real datasets. | ['Or Zuk', 'Tom Hope', 'Avishai Wagner'] | 2015-10-18 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 1.83441445e-01 -3.50297421e-01 1.05117977e-01 -2.83859074e-01
-5.50479591e-01 -4.66865331e-01 5.30432045e-01 -1.13431722e-01
-2.10780427e-01 1.96671620e-01 5.85494757e-01 3.09803963e-01
-5.95060229e-01 -6.35610044e-01 -2.05876067e-01 -1.20777404e+00
-8.32170725e-01 -1.73226833e-01 2.54602879e-02 1.17359748e-02
4.96633686e-02 7.23473072e-01 -1.63047945e+00 5.16333222e-01
5.16291678e-01 1.15014756e+00 1.92481637e-01 5.29144764e-01
1.41531289e-01 6.52690649e-01 -5.02089262e-01 3.86166394e-01
1.72740117e-01 -4.58875388e-01 -3.66162747e-01 5.36844611e-01
2.18507051e-02 4.73127306e-01 -6.00337505e-01 1.05612159e+00
4.69124377e-01 5.51564395e-01 4.80669528e-01 -8.94417048e-01
-4.39246655e-01 6.12525463e-01 -5.57242990e-01 1.92517623e-01
1.32247612e-01 -2.54433781e-01 6.34525836e-01 -8.06358099e-01
5.59594750e-01 1.26083422e+00 9.94306922e-01 4.89689261e-02
-1.55759704e+00 -2.55571753e-01 -3.20755690e-01 3.00549179e-01
-1.58461845e+00 -6.26036704e-01 1.22736537e+00 -2.41845518e-01
8.33601236e-01 4.25067961e-01 5.28307736e-01 8.45975816e-01
-5.93187809e-02 5.45420587e-01 9.41229880e-01 -4.75633681e-01
5.71176469e-01 -5.19800127e-01 -2.74451822e-02 3.81251216e-01
-1.68932274e-01 -2.21490767e-02 -5.68510771e-01 -2.10148260e-01
7.09846973e-01 2.92987823e-01 -3.41642588e-01 -3.99479359e-01
-1.42475820e+00 9.10933554e-01 3.85750145e-01 8.68516624e-01
-6.48598909e-01 9.37541127e-02 5.86265087e-01 5.79504669e-01
3.96202534e-01 3.51503700e-01 -1.96059391e-01 3.14552665e-01
-1.42464805e+00 -1.70271069e-01 6.81511462e-01 5.27550638e-01
8.27043951e-01 4.64714825e-01 -1.52018011e-01 7.38765419e-01
1.28083423e-01 5.05865574e-01 7.64160573e-01 -1.54458928e+00
-1.29082188e-01 1.39820009e-01 -3.36129308e-01 -1.45676899e+00
-7.55850732e-01 -7.33997881e-01 -1.71168697e+00 -2.98936255e-02
1.66010678e-01 7.65129924e-02 -5.99485397e-01 1.61906278e+00
1.54368043e-01 7.26318479e-01 1.34623393e-01 9.01501417e-01
5.91543674e-01 5.77651799e-01 -3.63246202e-01 -7.96707928e-01
1.27427793e+00 -5.40370166e-01 -9.66030359e-01 2.85367727e-01
3.89340580e-01 -7.87410557e-01 6.97250009e-01 6.50450230e-01
-1.00883377e+00 -7.34250784e-01 -7.58925259e-01 3.04094970e-01
-2.52265990e-01 7.17138425e-02 5.55591881e-01 4.70302105e-01
-1.16999817e+00 9.69256103e-01 -1.10437346e+00 -2.97083020e-01
4.37500626e-01 2.56114483e-01 -4.10548478e-01 -1.16777703e-01
-9.46875751e-01 3.55843842e-01 1.84949741e-01 8.60036165e-03
-4.73763198e-01 -7.16013074e-01 -7.54020870e-01 2.07029343e-01
2.05843057e-02 -3.82502496e-01 4.04999942e-01 -7.59262323e-01
-1.35358739e+00 3.54999810e-01 -4.49376434e-01 -6.49230540e-01
-2.69606382e-01 1.61362320e-01 -9.11974847e-01 8.73598874e-01
1.36153817e-01 1.23760588e-01 1.52517569e+00 -1.09865642e+00
-1.93984509e-01 -3.40360194e-01 -8.49306226e-01 -2.08810464e-01
-7.33127236e-01 1.22898579e-01 -1.92931831e-01 -1.11872101e+00
7.21903205e-01 -4.57701027e-01 -6.81279719e-01 -3.07544768e-01
-8.36333558e-02 5.88220693e-02 1.30580151e+00 -8.05999517e-01
1.30896270e+00 -2.85100436e+00 3.12481701e-01 8.37417006e-01
4.46021646e-01 1.77171212e-02 -2.90367991e-01 6.52833283e-01
-2.81177849e-01 -4.03512150e-01 -3.26703906e-01 -4.35714126e-01
-2.15766907e-01 3.21977228e-01 -1.81689858e-01 8.56852353e-01
4.40637581e-02 5.85997641e-01 -7.73606479e-01 -3.29445213e-01
4.95798886e-01 8.04460347e-01 -4.16351944e-01 -3.04475110e-02
3.54102999e-01 6.58322811e-01 -2.26328909e-01 4.58343953e-01
8.15956771e-01 -3.05532753e-01 1.49929196e-01 -6.58983707e-01
-2.55101204e-01 -2.46563569e-01 -1.51818061e+00 1.99298930e+00
-2.39120260e-01 6.89504981e-01 7.67940283e-01 -1.67639852e+00
1.02038527e+00 4.07576174e-01 1.05660379e+00 -4.93793935e-01
2.06749573e-01 1.29271388e-01 -2.86564559e-01 -4.28195775e-01
1.76903792e-02 -1.02854311e-01 1.11836515e-01 2.85756916e-01
2.75569588e-01 -2.67854799e-02 1.34642482e-01 2.67678380e-01
1.28743529e+00 -6.67336524e-01 5.62476702e-02 -6.13284826e-01
6.50366962e-01 -4.18232203e-01 2.51218349e-01 5.70563972e-01
5.46392500e-02 6.77782774e-01 -2.82205231e-02 -4.02563423e-01
-1.08314741e+00 -9.59416986e-01 -2.18737289e-01 8.11214745e-01
-3.51535559e-01 -3.71089697e-01 -4.03709143e-01 -4.81678918e-02
-8.67118035e-03 3.72327954e-01 -5.26018441e-01 -3.12810183e-01
-6.25864506e-01 -7.08491147e-01 4.08933073e-01 3.00583005e-01
3.71463895e-01 -7.81383276e-01 -3.91659409e-01 4.13064569e-01
-2.39688367e-01 -1.14415300e+00 -3.99147600e-01 6.77348137e-01
-1.04507160e+00 -9.24120188e-01 -5.66504776e-01 -9.79351699e-01
4.72745657e-01 6.03673160e-01 8.27446878e-01 5.19048274e-02
-4.27551508e-01 4.72567618e-01 -6.76250935e-01 1.68180272e-01
-1.29176334e-01 -2.38064289e-01 2.42567629e-01 5.44823825e-01
3.19955498e-01 -1.28417361e+00 -5.48227906e-01 2.06356883e-01
-1.04718983e+00 -5.60562849e-01 3.55866253e-01 8.56548905e-01
4.95613277e-01 1.04963279e+00 5.20010650e-01 -2.72900999e-01
5.58789551e-01 -3.47833663e-01 -1.69184759e-01 -1.97916240e-01
-2.02031374e-01 -1.06256820e-01 1.10124123e+00 -5.03369510e-01
-6.94194674e-01 3.37184012e-01 1.18669057e-02 -8.10492575e-01
-2.58033752e-01 4.48124856e-01 8.25059041e-02 -4.02244508e-01
7.36629367e-01 5.67255020e-01 3.32653910e-01 -7.23074079e-01
6.06109262e-01 5.45180142e-01 8.07343781e-01 -3.72614622e-01
1.12105942e+00 1.15463257e+00 3.31834465e-01 -1.36968207e+00
-6.72332764e-01 -8.89852524e-01 -9.26648498e-01 -3.97649296e-02
7.59058058e-01 -1.01208043e+00 -5.50211489e-01 2.24339128e-01
-7.58077264e-01 1.74245521e-01 -7.36735284e-01 8.93727183e-01
-5.97018421e-01 8.07463646e-01 -6.30069792e-01 -6.02299213e-01
-1.40513763e-01 -5.38791895e-01 9.36246693e-01 -2.61756480e-01
-2.27759391e-01 -1.20110428e+00 -1.73638731e-01 -2.70980746e-01
7.00684130e-01 4.75658745e-01 7.57644296e-01 -4.97181624e-01
-3.31803411e-02 -1.87296733e-01 8.70065168e-02 5.62450111e-01
2.11855352e-01 -1.86715722e-01 -7.53819168e-01 -6.94161594e-01
7.54521668e-01 5.90885580e-02 1.13855040e+00 9.65821087e-01
1.32509637e+00 -2.43157953e-01 -2.65246749e-01 9.30132926e-01
1.36225498e+00 -1.76948413e-01 5.28150856e-01 5.19145206e-02
6.50839627e-01 6.02412581e-01 6.13207817e-02 8.11253607e-01
-1.14939407e-01 3.39348018e-01 1.46035179e-01 -3.32202673e-01
-2.92060792e-01 2.66781986e-01 1.94016576e-01 1.47000229e+00
-9.69334468e-02 4.39713478e-01 -5.32929182e-01 6.52878761e-01
-1.80337811e+00 -1.32031524e+00 -4.17111725e-01 1.95835090e+00
5.68402052e-01 -2.52763152e-01 2.23737091e-01 8.00116003e-01
7.90501595e-01 3.90727222e-01 -6.69664741e-02 1.43790364e-01
-6.87862456e-01 5.43270111e-01 3.91571611e-01 2.88729489e-01
-1.25353909e+00 5.97191036e-01 7.16223574e+00 1.10432386e+00
-1.04946876e+00 9.96088684e-02 5.66942170e-02 -1.22700613e-02
-9.75542292e-02 -2.11406663e-01 3.05985734e-02 4.26060796e-01
9.79943693e-01 -1.25502512e-01 7.49490499e-01 5.04783273e-01
6.83544159e-01 3.96506667e-01 -7.56435871e-01 1.03182352e+00
-6.35368973e-02 -1.42169094e+00 -1.80495396e-01 -1.97302386e-01
6.71334982e-01 -1.49492353e-01 1.63522139e-01 -2.51087129e-01
1.31205857e-01 -9.76089478e-01 3.98227304e-01 5.50114274e-01
5.09747326e-01 -8.92793298e-01 5.46101272e-01 1.48869634e-01
-1.61183846e+00 -3.59654754e-01 -5.39709985e-01 -8.22090879e-02
2.51276553e-01 1.28652930e+00 -2.33599126e-01 6.46864891e-01
9.16672289e-01 1.29269731e+00 -1.85213566e-01 1.06690121e+00
2.29393587e-01 7.63385415e-01 -4.94452298e-01 4.04791325e-01
2.90987909e-01 -3.51053923e-01 6.54688895e-01 1.50207114e+00
6.48982584e-01 1.53771624e-01 4.68384027e-01 3.87361646e-01
2.09504813e-01 -2.52242833e-02 -4.82811421e-01 1.01685807e-01
6.06244266e-01 1.43364859e+00 -1.15721834e+00 -3.91484350e-01
-3.81179750e-01 9.32758689e-01 -3.39149475e-01 5.79904795e-01
-2.09416777e-01 -5.16931057e-01 5.58099091e-01 -2.42398940e-02
8.56545806e-01 -5.85072577e-01 -3.14563304e-01 -1.02439678e+00
-2.15386167e-01 -9.45461392e-01 3.71721834e-01 -4.76804852e-01
-1.61367214e+00 3.94862562e-01 -3.47935110e-01 -1.50482607e+00
1.02762319e-02 -2.76675761e-01 -5.65805912e-01 5.85281074e-01
-1.38482571e+00 -1.06425118e+00 -1.41179293e-01 1.41063690e+00
2.47772187e-01 -3.58118683e-01 8.67946804e-01 4.51040894e-01
-9.00302827e-02 1.38443291e-01 5.93255103e-01 1.26805622e-02
5.24523258e-01 -1.17008829e+00 1.31964509e-03 8.56521130e-01
1.38744235e-01 6.95487618e-01 7.17484832e-01 -4.02285963e-01
-1.67371678e+00 -1.22540462e+00 5.81261158e-01 9.17166844e-02
1.02231216e+00 -3.20909172e-01 -1.06311667e+00 2.27630466e-01
1.64293841e-01 2.99618661e-01 1.01718879e+00 4.20186669e-02
-4.55930591e-01 -1.02216765e-01 -1.24311090e+00 1.30189911e-01
8.91083241e-01 -7.07628310e-01 -6.13294482e-01 4.67912793e-01
7.21239567e-01 2.42880091e-01 -1.27382576e+00 1.42913401e-01
1.87349185e-01 -7.83119261e-01 1.41936135e+00 -3.48415047e-01
-1.89801045e-02 -5.95325112e-01 -5.04580736e-01 -1.54130781e+00
-1.19195521e+00 -1.03624713e+00 -2.93840319e-01 9.93123233e-01
-1.79320067e-01 -3.19676697e-01 6.99775279e-01 -4.17395264e-01
-3.24987590e-01 -2.79094428e-02 -1.21221876e+00 -1.00127673e+00
-2.20898822e-01 -4.38444644e-01 4.18179125e-01 1.44371808e+00
1.60408810e-01 1.35773383e-02 -5.33312321e-01 5.08719027e-01
1.14207494e+00 1.91016734e-01 3.06457877e-01 -1.43046331e+00
-1.64161772e-01 -4.21643317e-01 -4.95647162e-01 -6.74317837e-01
2.50531316e-01 -1.00375259e+00 -1.28717437e-01 -1.05909145e+00
-2.90259778e-01 4.84417267e-02 -3.69081289e-01 1.86792806e-01
3.07460099e-01 5.28871655e-01 7.39629120e-02 5.73720813e-01
-6.89917505e-01 5.31800389e-01 1.01415527e+00 2.49631703e-02
-2.85904527e-01 -1.43053725e-01 -4.48415428e-01 6.87991738e-01
5.56782067e-01 -4.03520852e-01 -8.10009837e-02 -1.96727827e-01
-2.59852678e-01 1.13451272e-01 2.91434526e-01 -1.43453825e+00
6.00576878e-01 2.13706478e-01 5.99071264e-01 -5.27402639e-01
3.41191620e-01 -1.29931676e+00 3.06767702e-01 5.51730633e-01
-2.85702646e-01 -2.77021170e-01 -1.37974918e-01 7.00541675e-01
-7.22016454e-01 1.24001950e-01 9.61673260e-01 -6.74434155e-02
-5.72358310e-01 4.17209379e-02 -6.55619085e-01 -4.35390443e-01
7.65289426e-01 -2.38817245e-01 2.10505918e-01 -3.90788883e-01
-1.18152368e+00 4.30362448e-02 1.21168166e-01 -7.41989166e-02
6.67489469e-01 -1.65360796e+00 -7.68399537e-01 6.02708101e-01
-2.29358450e-01 -3.06742400e-01 4.36571956e-01 1.23165512e+00
-1.26345992e-01 -8.78977403e-03 -1.53847679e-01 -8.40604365e-01
-1.18757296e+00 7.72031903e-01 2.70155072e-03 -2.47437935e-02
-9.87183273e-01 6.11274242e-01 -4.26022500e-01 -1.57017618e-01
2.23874211e-01 -1.61862224e-01 -3.92916739e-01 3.96379799e-01
7.92640686e-01 5.13348997e-01 1.60309225e-01 -8.24277222e-01
-3.86098623e-01 8.29360485e-01 4.88621086e-01 1.09127603e-01
1.87244320e+00 -2.84443587e-01 -6.09161019e-01 3.40953082e-01
1.61675692e+00 9.60245952e-02 -1.02727199e+00 -5.41792691e-01
2.16568381e-01 -5.19107699e-01 3.24130684e-01 -1.10466920e-01
-1.28225529e+00 7.12782204e-01 5.57781756e-01 8.64153862e-01
1.52146530e+00 1.46858795e-02 6.98996961e-01 5.06939173e-01
1.55990347e-01 -1.21784902e+00 2.40027308e-01 4.78495777e-01
5.92671514e-01 -8.15403104e-01 3.16441245e-02 -4.71134633e-01
-1.84594244e-01 1.26477921e+00 -4.65786934e-01 -5.34243226e-01
1.13534236e+00 5.82045496e-01 4.97023389e-03 -4.07044321e-01
-4.36778575e-01 -2.46241882e-01 3.43164057e-01 1.02951276e+00
1.97100222e-01 1.30517572e-01 -9.55333486e-02 5.21687090e-01
-2.88108498e-01 -2.03501359e-01 1.48494080e-01 6.49127305e-01
-8.84295940e-01 -8.38265717e-01 -8.44210267e-01 4.82035965e-01
-3.22817534e-01 1.05222715e-02 1.81452818e-02 2.87382007e-01
-4.72026467e-02 1.22836530e+00 1.00431778e-01 -5.43962121e-01
2.11357117e-01 1.72205537e-03 3.11839897e-02 -3.59516889e-01
-5.70226014e-01 9.05699253e-01 -2.44166210e-01 -7.52731919e-01
-9.28301275e-01 -8.10122967e-01 -1.23070490e+00 -4.17623043e-01
1.24359615e-02 3.79304796e-01 5.49645543e-01 6.98303223e-01
4.16594893e-01 6.60123467e-01 1.22228372e+00 -1.13877940e+00
-3.65750045e-01 -8.13645661e-01 -1.28637159e+00 6.83236718e-01
5.22888005e-01 -3.44398737e-01 -7.93529987e-01 3.49703401e-01] | [11.712676048278809, -2.2940471172332764] |
6948850e-a293-43b2-976e-720bff32730c | food-recommendation-framework-existing | 1905.06269 | null | https://arxiv.org/abs/1905.06269v2 | https://arxiv.org/pdf/1905.06269v2.pdf | Food Recommendation: Framework, Existing Solutions and Challenges | A growing proportion of the global population is becoming overweight or obese, leading to various diseases (e.g., diabetes, ischemic heart disease and even cancer) due to unhealthy eating patterns, such as increased intake of food with high energy and high fat. Food recommendation is of paramount importance to alleviate this problem. Unfortunately, modern multimedia research has enhanced the performance and experience of multimedia recommendation in many fields such as movies and POI, yet largely lags in the food domain. This article proposes a unified framework for food recommendation, and identifies main issues affecting food recommendation including building the personal model, analyzing unique food characteristics, incorporating various context and domain knowledge. We then review existing solutions for these issues, and finally elaborate research challenges and future directions in this field. To our knowledge, this is the first survey that targets the study of food recommendation in the multimedia field and offers a collection of research studies and technologies to benefit researchers in this field. | ['Weiqing Min', 'Ramesh Jain', 'Shuqiang Jiang'] | 2019-05-15 | null | null | null | null | ['food-recommendation'] | ['miscellaneous'] | [ 1.50858283e-01 -1.57510489e-01 -1.03297913e+00 -2.39711151e-01
2.06857264e-01 -3.42466354e-01 -1.86719507e-01 9.17141557e-01
-3.69575806e-02 2.55562305e-01 6.71373546e-01 -1.56816438e-01
-1.05605111e-01 -1.18903410e+00 -2.54516155e-01 -5.53100586e-01
-8.51973891e-02 -2.47113422e-01 1.39320329e-01 -3.24510306e-01
2.31191114e-01 -3.16309184e-01 -1.92984045e+00 1.18529588e-01
9.35739160e-01 7.62068629e-01 4.98944551e-01 1.33911252e-01
-3.37673455e-01 2.12514207e-01 -5.66873252e-02 -3.82747382e-01
-6.75266683e-02 -6.08683169e-01 -3.02966237e-01 1.34382680e-01
2.24836528e-01 -6.61928892e-01 -4.20458019e-02 1.35220206e+00
4.65413660e-01 3.50139081e-01 3.13708395e-01 -9.71755683e-01
-1.22874415e+00 1.04496896e+00 -4.94221359e-01 2.01704681e-01
6.21730208e-01 -4.41968054e-01 5.98169386e-01 -4.80527043e-01
9.68943834e-02 1.08473313e+00 4.53844100e-01 2.85949171e-01
-6.57490790e-01 -7.65039027e-01 5.79439163e-01 2.18708053e-01
-1.05753446e+00 -1.82809234e-02 5.71777463e-01 -2.77585536e-01
5.53439200e-01 4.60853934e-01 1.20486486e+00 8.23524535e-01
1.16137102e-01 5.00843763e-01 8.90339255e-01 -5.18747389e-01
2.23085493e-01 2.23589718e-01 5.42355239e-01 3.93593937e-01
8.13808680e-01 2.46148836e-02 -2.80365765e-01 1.61291957e-01
6.27088428e-01 5.57213604e-01 -7.13728443e-02 -1.30285054e-01
-1.04130244e+00 1.04915750e+00 3.11703563e-01 4.94306892e-01
-4.54657853e-01 -4.70254660e-01 1.59846842e-01 2.42370605e-01
6.35050714e-01 -9.14666131e-02 -4.00015980e-01 3.86279464e-01
-5.38886011e-01 2.95322001e-01 8.00066590e-01 9.71769631e-01
4.73624259e-01 -4.22803871e-02 1.80084750e-01 8.27213824e-01
9.08694804e-01 9.39431667e-01 4.63834643e-01 -7.35606611e-01
-2.61540487e-02 4.32918161e-01 2.37445027e-01 -1.71100986e+00
-6.94934189e-01 -4.81465250e-01 -8.22490692e-01 -3.24953109e-01
3.91809046e-01 -1.25790268e-01 -3.29023540e-01 1.51773059e+00
6.76542640e-01 -6.95322752e-02 -2.51183927e-01 1.06489348e+00
1.37721455e+00 7.74647951e-01 8.61661255e-01 -5.25251389e-01
1.86272705e+00 -9.66591239e-01 -1.03689015e+00 -7.38769174e-02
3.75407636e-01 -1.06117606e+00 8.56826305e-01 4.42656189e-01
-1.03187108e+00 -7.29222715e-01 -9.68365133e-01 6.92059770e-02
-5.25685847e-01 -5.14174551e-02 8.32467079e-01 1.08507419e+00
-7.25894928e-01 4.53021348e-01 -5.31379163e-01 -1.28398395e+00
-4.44691367e-02 1.23195518e-02 2.43742287e-01 -5.99858344e-01
-1.36916387e+00 5.88161588e-01 5.99360824e-01 -3.70069951e-01
-7.52983764e-02 -8.85077596e-01 -9.62622762e-01 -2.12121010e-01
3.90742600e-01 -1.00249493e+00 1.11347103e+00 -9.77414548e-01
-1.43913424e+00 5.03049195e-01 1.41726062e-01 -1.35292932e-01
-2.78264761e-01 -5.67842245e-01 -1.24604106e+00 -3.41279954e-02
1.47660002e-01 5.61203063e-01 4.61782157e-01 -8.85971129e-01
-1.03163362e+00 -4.55979288e-01 3.18744123e-01 3.05176795e-01
-6.85085535e-01 3.13836902e-01 -7.84976408e-02 -8.46554399e-01
3.78124863e-02 -6.01348937e-01 -4.36050892e-01 7.08624795e-02
-8.55061039e-02 -2.25816175e-01 5.51795661e-01 -6.43343389e-01
1.81495786e+00 -2.28572965e+00 -2.41592616e-01 5.88178597e-02
-2.24008001e-02 1.75150916e-01 -3.56527790e-02 5.98157823e-01
1.96522310e-01 3.23354483e-01 5.68123519e-01 5.63045979e-01
1.03199549e-01 1.78873718e-01 1.91416666e-01 4.04420048e-01
-6.53283238e-01 4.53594506e-01 -1.39048886e+00 -4.14555013e-01
5.37897348e-01 6.22057557e-01 -6.56829834e-01 -2.27941841e-01
-4.66813922e-01 2.74631917e-01 -8.84875774e-01 6.72150850e-01
7.74232924e-01 -4.07040477e-01 3.94317180e-01 -6.63265169e-01
-6.25813782e-01 1.50460899e-01 -1.18416083e+00 1.68357551e+00
-7.39689320e-02 -3.24669003e-01 -5.76104522e-02 -7.99627900e-01
7.09300518e-01 3.22475553e-01 7.87363350e-01 -8.00724566e-01
3.31160516e-01 8.75460431e-02 -4.01913282e-03 -1.10616255e+00
5.97783625e-01 6.97407722e-02 1.90248296e-01 3.68089944e-01
-4.38462019e-01 8.19629788e-01 5.46471894e-01 1.24482969e-02
3.06316525e-01 1.99382201e-01 7.61152148e-01 -5.94476759e-01
4.94925052e-01 1.42299205e-01 2.43490472e-01 2.42230698e-01
-2.87768334e-01 -1.53580204e-01 -6.65248394e-01 -3.55780423e-01
-8.30417991e-01 -8.70801151e-01 -1.26174435e-01 1.40957367e+00
5.03964543e-01 -6.30550086e-01 -3.99199456e-01 -3.17771584e-01
5.83803132e-02 7.46646702e-01 -2.36550897e-01 -1.57401428e-01
-4.26009983e-01 -8.41576338e-01 -2.98237115e-01 3.29328597e-01
8.15923393e-01 -6.60719275e-01 -4.85789031e-01 4.33756113e-01
-6.15596235e-01 -4.94101465e-01 -6.95938349e-01 -4.30157244e-01
-1.12193859e+00 -1.05994105e+00 -7.74374843e-01 -1.00607812e+00
5.33018708e-01 1.22076476e+00 1.27229393e+00 4.32095438e-01
4.15396355e-02 4.55720901e-01 -1.04096174e+00 -5.59959531e-01
-1.53455243e-01 -1.59767359e-01 8.55925903e-02 -3.19604903e-01
9.85292315e-01 -3.39600742e-01 -1.33474469e+00 4.73562360e-01
-1.03879142e+00 -7.54803568e-02 2.68093199e-01 5.28211035e-02
7.68897593e-01 6.31563604e-01 7.70245671e-01 -7.75998533e-01
2.88480818e-01 -1.24725306e+00 -6.70502260e-02 -1.05523625e-02
-6.69423521e-01 -7.07370162e-01 3.11313033e-01 -5.86577892e-01
-8.66797864e-01 -2.71994561e-01 -3.96871984e-01 6.54458106e-01
-5.09121060e-01 7.59774804e-01 -2.39632457e-01 1.39451504e-01
5.07628441e-01 -3.12084466e-01 -1.55985922e-01 -7.47072279e-01
4.50386047e-01 5.69452703e-01 9.73386616e-02 -2.31549621e-01
3.07370782e-01 1.70681223e-01 -2.05732152e-01 -9.95095074e-01
-1.22179449e+00 -6.28344297e-01 -7.09632114e-02 -5.42878330e-01
9.24308658e-01 -9.22405183e-01 -4.42494661e-01 1.42395824e-01
-3.44987482e-01 8.12206641e-02 8.91194418e-02 1.00594401e+00
7.62639046e-02 5.72769642e-01 -6.98946774e-01 -4.75524634e-01
-3.92759442e-01 -5.63824356e-01 3.42505753e-01 6.28908336e-01
-4.01417524e-01 -1.15461171e+00 -4.25252207e-02 4.35156196e-01
6.60887480e-01 2.29656726e-01 9.48853493e-01 -1.76302224e-01
-2.10305437e-01 1.86370224e-01 -9.05239061e-02 -2.63762653e-01
4.46197301e-01 -1.00704357e-01 -3.91666561e-01 -3.22820663e-01
-6.76232204e-02 2.00872630e-01 5.97979903e-01 9.99957204e-01
6.86971009e-01 -2.61040956e-01 -6.30076468e-01 2.81869043e-02
1.55846167e+00 5.33310831e-01 5.47619641e-01 5.22499681e-01
3.56015474e-01 6.93690896e-01 9.21119273e-01 5.82201838e-01
7.95389175e-01 4.86449838e-01 4.79243517e-01 -7.33296648e-02
-3.92130703e-01 -3.48308086e-01 4.34883356e-01 1.17148268e+00
-1.22789383e-01 -3.84972751e-01 -1.44273236e-01 3.96557122e-01
-1.65162086e+00 -1.02736878e+00 -5.73360622e-01 2.22311187e+00
5.30517042e-01 -5.07435739e-01 5.97926497e-01 2.03388110e-01
8.09762776e-01 -2.81940818e-01 -3.76637399e-01 -3.89321931e-02
1.28059953e-01 -9.51402336e-02 4.08091217e-01 1.04648851e-01
-1.38181806e+00 5.10930836e-01 6.81368589e+00 3.23838383e-01
-9.62581396e-01 1.19531699e-01 3.74369711e-01 1.62184745e-01
-4.56721663e-01 -5.70164323e-01 -8.57776463e-01 6.72636330e-01
9.53919828e-01 -1.97195530e-01 3.45368713e-01 7.39858747e-01
4.91015792e-01 -2.70521849e-01 -5.94205856e-01 7.19437718e-01
2.30607092e-01 -8.02096665e-01 -1.42908394e-01 1.92934591e-02
6.45658910e-01 -4.20339584e-01 2.02482894e-01 1.51681483e-01
1.67341530e-01 -3.86024266e-01 5.67372382e-01 1.55819401e-01
3.50385427e-01 -7.04867721e-01 4.79447216e-01 1.60433918e-01
-1.28700507e+00 -9.48487141e-04 -7.27991819e-01 -3.86478722e-01
2.90643334e-01 1.07606554e+00 2.59671360e-01 6.42534018e-01
1.11649609e+00 1.16487920e+00 -1.42381728e-01 1.38390625e+00
4.89655286e-02 5.51327109e-01 -3.13646466e-01 -7.51013160e-02
-2.05372363e-01 -4.81351286e-01 2.08893344e-01 9.51696813e-01
9.78311360e-01 6.79826200e-01 8.27260315e-01 1.41223431e-01
2.87112832e-01 9.91290152e-01 -4.18912143e-01 3.31851877e-02
3.25360060e-01 1.20332599e+00 -9.60992634e-01 -2.11065620e-01
-9.92135525e-01 4.92684335e-01 -3.65318507e-01 6.77859262e-02
-7.48689055e-01 3.14006388e-01 5.47779679e-01 3.03486973e-01
1.91767454e-01 8.96643568e-03 1.42643273e-01 -1.11099970e+00
-6.34556711e-01 -8.17985594e-01 8.10016930e-01 -3.88099253e-01
-1.39802134e+00 -8.73347819e-02 2.09589675e-01 -1.29166007e+00
2.92600513e-01 -3.44334602e-01 -5.94960079e-02 3.89114469e-01
-1.74053359e+00 -1.07575047e+00 -2.47986168e-01 5.08041978e-01
5.11812568e-01 2.96773404e-01 1.02993333e+00 1.05600941e+00
-5.45439601e-01 4.23401088e-01 1.57892093e-01 -5.36217272e-01
8.46553266e-01 -6.66129231e-01 -3.86472821e-01 4.03456807e-01
-3.06187809e-01 6.88053250e-01 8.92138600e-01 -9.28458869e-01
-1.58801103e+00 -1.25597823e+00 1.01440084e+00 1.63059905e-01
4.41439599e-01 3.93224686e-01 -4.65971619e-01 4.25248116e-01
4.37845856e-01 -6.45263433e-01 1.47291458e+00 1.59101531e-01
-1.21383160e-01 -3.76652218e-02 -1.56520498e+00 6.68894470e-01
1.05606616e+00 4.01023507e-01 -1.26562506e-01 4.23735678e-01
6.71629727e-01 -1.24616452e-01 -1.38970721e+00 7.74536431e-02
9.19978738e-01 -8.71644855e-01 1.27108026e+00 -3.54243606e-01
4.45922941e-01 -3.83552670e-01 -4.30004984e-01 -1.25862527e+00
-1.08088648e+00 6.08865954e-02 -2.66514987e-01 1.30789244e+00
2.08718125e-02 -3.07678580e-01 6.00317478e-01 4.42244768e-01
-2.19990358e-01 -2.44048521e-01 1.48676395e-01 -4.73542660e-01
-5.53942844e-02 -6.17338605e-02 8.24585915e-01 1.01049805e+00
3.36231083e-01 1.59571752e-01 -5.89626372e-01 2.13229626e-01
8.55169952e-01 2.46811524e-01 3.30694377e-01 -1.31122434e+00
-1.50653392e-01 -3.60203266e-01 1.90122034e-02 -1.18253922e+00
-4.98957723e-01 -7.90704012e-01 -4.03502584e-01 -1.76434457e+00
3.80425990e-01 -1.62750736e-01 -5.86770535e-01 1.51770696e-01
-1.80183515e-01 5.12273788e-01 4.37651426e-02 -7.49084651e-02
-6.32937014e-01 -2.34599933e-01 1.65106654e+00 -1.95427150e-01
-2.18989924e-01 -5.70647232e-02 -1.38412011e+00 8.94246578e-01
1.31871510e+00 -3.03773791e-01 -7.16207325e-01 -4.23975855e-01
7.13116527e-01 -1.34405121e-01 -1.62172645e-01 -7.22976446e-01
-1.53562486e-01 -6.46266282e-01 4.19588774e-01 -6.52534366e-01
-3.36366266e-01 -1.20564508e+00 6.51470721e-01 7.34660983e-01
-2.02551648e-01 -8.54545459e-02 5.35561480e-02 5.87111712e-01
2.53877819e-01 -3.54281873e-01 5.50897360e-01 -2.86688298e-01
-1.09723115e+00 2.29510173e-01 -6.74337745e-01 -3.94669175e-01
1.13432467e+00 -7.46192336e-02 -4.02968317e-01 -2.10589737e-01
-9.49090064e-01 2.75341690e-01 3.70343298e-01 6.14941776e-01
3.84958506e-01 -1.59501958e+00 -6.65966570e-01 1.05057113e-01
2.82649368e-01 -7.54547596e-01 7.19756246e-01 8.19046199e-01
-2.37440407e-01 2.75696844e-01 -3.15305740e-01 -1.30514145e-01
-1.29861677e+00 1.20828474e+00 -2.50563949e-01 -7.74481818e-02
-7.48335242e-01 2.08092496e-01 6.88072518e-02 -3.86815183e-02
3.49073708e-01 -4.96468693e-01 -9.36226726e-01 6.32852968e-03
1.19341719e+00 8.11831653e-01 -3.27486783e-01 -5.60607255e-01
-7.80197233e-02 7.34054983e-01 5.73490299e-02 7.86981821e-01
9.00808811e-01 -9.50759172e-01 -9.86804888e-02 3.22005242e-01
6.05909884e-01 -6.63303062e-02 -4.82405841e-01 -7.93898776e-02
-3.56378049e-01 -5.95015049e-01 2.49680653e-01 -1.01954782e+00
-1.27664542e+00 3.46826226e-01 9.61804926e-01 6.31429493e-01
1.22541821e+00 -5.08335494e-02 1.30865073e+00 -1.43196225e-01
6.72584951e-01 -9.29893851e-01 -8.90577137e-02 -3.88408676e-02
3.52551132e-01 -1.18538845e+00 1.79643303e-01 -1.02427959e+00
-4.09167521e-02 1.05927014e+00 1.85047045e-01 1.16463173e-02
1.57570696e+00 2.24721618e-02 2.30925083e-01 2.83138733e-02
-2.92338058e-02 -3.74440908e-01 2.10516587e-01 8.59811425e-01
1.12501693e+00 3.37585539e-01 -1.23807728e+00 8.10581744e-01
-6.86389357e-02 5.30124903e-01 -2.51468848e-02 8.96438181e-01
-1.12885654e+00 -1.54218161e+00 -5.75731337e-01 5.85510969e-01
-8.23505044e-01 -1.45190405e-02 4.25960451e-01 4.74142879e-01
6.22068286e-01 1.68461645e+00 -5.56927547e-02 -2.68529654e-01
3.29054564e-01 -2.11251959e-01 5.23688078e-01 -4.05706286e-01
-5.76830626e-01 6.47140801e-01 2.76989013e-01 -2.35548839e-01
-1.26276577e+00 -5.96585214e-01 -1.06586003e+00 -8.88424695e-01
-1.86523840e-01 7.21816644e-02 5.77043593e-01 6.40189111e-01
1.14020653e-01 5.99812925e-01 2.92898715e-01 -5.87666452e-01
-2.38886520e-01 -7.30245590e-01 -9.82498169e-01 6.41990721e-01
-1.37505159e-01 -7.71026552e-01 2.17681214e-01 2.96374142e-01] | [11.545733451843262, 4.458858966827393] |
c5ab727f-1cf3-4468-9081-b0724039ebbc | on-the-model-based-stochastic-value-gradient | 2008.12775 | null | https://arxiv.org/abs/2008.12775v3 | https://arxiv.org/pdf/2008.12775v3.pdf | On the model-based stochastic value gradient for continuous reinforcement learning | For over a decade, model-based reinforcement learning has been seen as a way to leverage control-based domain knowledge to improve the sample-efficiency of reinforcement learning agents. While model-based agents are conceptually appealing, their policies tend to lag behind those of model-free agents in terms of final reward, especially in non-trivial environments. In response, researchers have proposed model-based agents with increasingly complex components, from ensembles of probabilistic dynamics models, to heuristics for mitigating model error. In a reversal of this trend, we show that simple model-based agents can be derived from existing ideas that not only match, but outperform state-of-the-art model-free agents in terms of both sample-efficiency and final reward. We find that a model-free soft value estimate for policy evaluation and a model-based stochastic value gradient for policy improvement is an effective combination, achieving state-of-the-art results on a high-dimensional humanoid control task, which most model-based agents are unable to solve. Our findings suggest that model-based policy evaluation deserves closer attention. | ['Samuel Stanton', 'Brandon Amos', 'Andrew Gordon Wilson', 'Denis Yarats'] | 2020-08-28 | null | null | null | null | ['humanoid-control'] | ['robots'] | [-2.52048165e-01 8.41019154e-02 -5.75237513e-01 4.46353927e-02
-7.63931632e-01 -5.77991068e-01 8.69332790e-01 1.81065693e-01
-1.00622118e+00 1.10383105e+00 2.29792580e-01 -3.41025770e-01
-3.73281926e-01 -5.00165701e-01 -6.05776131e-01 -6.31077051e-01
-1.84871882e-01 8.72350574e-01 1.69002295e-01 -5.35035372e-01
4.08747733e-01 2.29851186e-01 -1.38293672e+00 -3.97133052e-01
1.01571214e+00 5.86026311e-01 2.47017726e-01 4.93205070e-01
4.49518651e-01 8.34622502e-01 -3.17576647e-01 3.33981141e-02
3.04015398e-01 -4.22068179e-01 -4.37402219e-01 -2.76961774e-01
2.24446915e-02 -6.57676637e-01 -4.03173745e-01 9.30783391e-01
8.09948742e-01 5.09849489e-01 5.74422717e-01 -1.24640393e+00
-3.65079731e-01 6.18856251e-01 -3.84665102e-01 6.60610870e-02
3.74181010e-02 9.85872567e-01 8.70281398e-01 -1.28923550e-01
5.10774434e-01 1.77981114e+00 5.07696986e-01 8.73521149e-01
-1.37751782e+00 -4.81235057e-01 4.99815375e-01 1.36829078e-01
-7.21760154e-01 -2.60406971e-01 4.19805527e-01 -5.09256065e-01
1.24123406e+00 -2.36396343e-01 1.09843516e+00 1.11412859e+00
4.35166359e-01 7.51621246e-01 1.47792876e+00 -3.38716775e-01
7.55410373e-01 -2.54302830e-01 -3.73301685e-01 6.41525745e-01
3.80008280e-01 1.13435996e+00 -3.62313718e-01 -3.10823411e-01
8.81201982e-01 -2.04756215e-01 1.00899376e-01 -7.80188084e-01
-1.13980579e+00 1.08480930e+00 3.70312840e-01 -2.30726719e-01
-7.88482249e-01 6.97207034e-01 3.44300002e-01 1.50875717e-01
1.24753430e-01 1.30947590e+00 -5.25436461e-01 -7.24147856e-01
-6.99415565e-01 1.11259127e+00 8.14646482e-01 4.66654748e-01
3.14166158e-01 3.63801032e-01 -2.78556287e-01 4.95958120e-01
4.76026595e-01 5.64776421e-01 4.70060229e-01 -1.67889762e+00
2.39348009e-01 3.57291013e-01 7.30836213e-01 -4.53042209e-01
-6.54143691e-01 -4.60210115e-01 -1.90002576e-01 1.00322104e+00
4.68444765e-01 -6.71426952e-01 -8.85973394e-01 2.16198969e+00
4.42622095e-01 6.98211417e-02 9.26229283e-02 1.00667906e+00
-2.13945165e-01 4.65608716e-01 2.57517129e-01 -2.36371577e-01
1.00531387e+00 -1.13330054e+00 -4.49570805e-01 -3.98662746e-01
4.89106208e-01 -2.27297217e-01 1.10923100e+00 4.83740687e-01
-1.36524844e+00 -2.24433929e-01 -1.04041457e+00 5.88466287e-01
-3.02060768e-02 -5.51519394e-01 6.42568111e-01 5.11464298e-01
-1.00670004e+00 9.87086415e-01 -1.29463434e+00 -3.75538737e-01
3.46090555e-01 3.58305454e-01 2.41518572e-01 2.18527451e-01
-1.02061856e+00 1.62789571e+00 4.41228956e-01 -2.89850950e-01
-1.47383082e+00 -8.08227479e-01 -6.08205497e-01 1.97099969e-02
9.51338172e-01 -9.78452921e-01 1.89669299e+00 -6.75766349e-01
-1.99217188e+00 -1.55807920e-02 1.49985030e-01 -8.70325387e-01
6.45666242e-01 -1.45035520e-01 7.98354745e-02 4.19309735e-02
-5.85337766e-02 7.91280568e-01 9.41218078e-01 -1.42677283e+00
-9.12324488e-01 -1.16318211e-01 3.22725445e-01 6.14343047e-01
-3.20007354e-02 -1.48936763e-01 3.45221698e-01 -4.81482208e-01
-6.24799490e-01 -1.17301166e+00 -7.78331578e-01 -4.16931957e-02
1.61538467e-01 -4.25431043e-01 5.55151105e-01 -2.74901897e-01
1.15549195e+00 -1.40472794e+00 2.25841448e-01 -4.70923297e-02
2.11106688e-01 4.95778501e-01 -2.61178881e-01 5.80650032e-01
4.42897975e-01 1.49273798e-01 -2.09541798e-01 -6.20379411e-02
4.76356000e-01 3.34414303e-01 -2.53049821e-01 3.04199547e-01
2.77927667e-01 9.25530255e-01 -1.34824312e+00 -2.91550398e-01
3.05362552e-01 1.53605685e-01 -8.94540668e-01 1.56817257e-01
-7.94466496e-01 4.72997725e-01 -7.57603765e-01 1.41821325e-01
1.37991428e-01 -4.59179096e-02 1.91146955e-01 4.81314540e-01
-2.28329927e-01 2.19164312e-01 -1.04499304e+00 1.31523633e+00
-2.97955513e-01 2.11908594e-01 1.55542701e-01 -9.11418974e-01
5.20890176e-01 1.48199931e-01 6.03352308e-01 -7.91190803e-01
4.34272066e-02 5.20336516e-02 4.72536862e-01 -4.02052641e-01
6.78795516e-01 -2.84577787e-01 4.05787565e-02 6.35744452e-01
-1.77019879e-01 -5.97891927e-01 5.54329157e-01 -6.03225641e-02
1.16264391e+00 7.47302055e-01 2.96906590e-01 -3.92964453e-01
-2.27991436e-02 2.25238711e-01 6.20933354e-01 1.15338767e+00
-5.06088853e-01 -2.69014519e-02 4.59220290e-01 -1.42816603e-01
-1.21366572e+00 -8.86178851e-01 2.34324679e-01 1.12727308e+00
1.19507827e-01 -3.00567865e-01 -7.91346967e-01 -4.26065683e-01
4.05328393e-01 8.90677392e-01 -6.48962140e-01 -3.70412976e-01
-7.38324761e-01 -5.08336902e-01 1.77595735e-01 5.10921240e-01
2.48427391e-01 -1.21201265e+00 -1.31098592e+00 7.40866303e-01
3.19263309e-01 -6.19652987e-01 -3.60846072e-01 2.14092910e-01
-8.44671130e-01 -9.25218165e-01 -8.11975241e-01 -1.23756126e-01
3.09464872e-01 -5.94993681e-02 1.16940677e+00 -2.79372726e-02
1.47388950e-01 7.33620703e-01 -1.06527492e-01 -7.31239080e-01
-6.45478129e-01 -6.18833229e-02 4.57234323e-01 -7.00488389e-01
-8.04185197e-02 -4.08938736e-01 -7.62228370e-01 2.25774586e-01
-6.64416134e-01 -1.07544713e-01 5.36627471e-01 1.12433755e+00
4.28656399e-01 -1.11174434e-01 7.80927837e-01 -3.13464314e-01
1.27067351e+00 -3.53204370e-01 -9.91701961e-01 9.48188677e-02
-1.12765145e+00 5.23626626e-01 6.88240886e-01 -8.82069588e-01
-9.70910430e-01 -3.15067619e-01 1.00343876e-01 -3.18000376e-01
1.29560065e-02 5.02539515e-01 5.19007564e-01 2.60935235e-03
7.77016103e-01 -2.01282371e-02 5.81111193e-01 -2.21535340e-01
4.19352919e-01 9.05141458e-02 1.37362123e-01 -1.08805680e+00
5.40151417e-01 1.60037637e-01 1.52166173e-01 -3.85549456e-01
-5.28711915e-01 8.75737965e-02 1.34826615e-01 -2.58057863e-01
8.40408504e-01 -8.54788303e-01 -1.17276371e+00 3.16552788e-01
-7.19307959e-01 -1.08987153e+00 -6.96912825e-01 6.60878897e-01
-1.22849238e+00 -6.45242482e-02 -4.86412048e-01 -1.13156044e+00
-1.40343867e-02 -1.33638120e+00 6.93683624e-01 6.79518461e-01
-1.83696955e-01 -1.06933010e+00 5.45780361e-01 -1.31147280e-01
7.54337192e-01 1.55577034e-01 9.73512769e-01 -5.00853002e-01
-4.21234131e-01 1.46062404e-01 2.18888178e-01 2.15125963e-01
-1.60923198e-01 -2.63575613e-01 -5.08406222e-01 -6.96200430e-01
-2.09564835e-01 -6.11899316e-01 6.59766793e-01 7.99513936e-01
6.21166527e-01 -5.10847986e-01 -3.58336270e-01 -3.70925805e-03
1.28163099e+00 4.66981322e-01 1.67816296e-01 6.07276738e-01
2.37057954e-01 3.34135205e-01 7.39360750e-01 7.84513175e-01
6.86179042e-01 7.95595586e-01 6.48816168e-01 1.93257600e-01
9.42969695e-02 -4.36753213e-01 5.11645377e-01 2.17753112e-01
-1.56085432e-01 -7.28014261e-02 -8.81881773e-01 5.21274388e-01
-2.26178360e+00 -1.29734087e+00 5.91956675e-01 2.13848257e+00
8.97493124e-01 3.66430044e-01 8.40285480e-01 -3.75620484e-01
4.33336347e-01 8.72337446e-02 -1.23557651e+00 -3.23372960e-01
2.69143671e-01 -6.12064749e-02 4.87717181e-01 5.97268760e-01
-8.53165746e-01 1.18601656e+00 6.98752165e+00 8.05171788e-01
-9.92084742e-01 -1.00680448e-01 3.69615585e-01 -3.74979705e-01
-1.71173051e-01 1.72074169e-01 -6.60657644e-01 5.85810900e-01
1.03206074e+00 -4.43705738e-01 9.99514222e-01 1.05346036e+00
8.26711416e-01 -4.45556432e-01 -1.07142901e+00 6.02625072e-01
-4.42508370e-01 -1.33835244e+00 -2.87405401e-01 4.08516139e-01
9.05504584e-01 2.75549829e-01 1.70930833e-01 8.18946123e-01
1.17987299e+00 -9.88129735e-01 1.06242192e+00 6.70901835e-01
3.21090110e-02 -8.52705061e-01 4.96995449e-01 5.51298022e-01
-7.74016857e-01 -4.65647757e-01 -2.13270247e-01 -4.65156376e-01
3.03874671e-01 -1.36651918e-01 -7.78692484e-01 6.56037182e-02
4.08593565e-01 4.27885264e-01 -2.30600461e-01 1.23754311e+00
-5.58655597e-02 6.56379521e-01 -3.75342876e-01 -5.29121518e-01
7.03428149e-01 -2.03511253e-01 6.80889845e-01 7.54909098e-01
1.91372007e-01 1.37022853e-01 5.94263554e-01 8.42448533e-01
4.03631747e-01 -4.43575382e-01 -6.53955758e-01 -2.80916065e-01
6.35421395e-01 8.94057214e-01 -6.19904935e-01 -3.69861275e-01
8.58886465e-02 2.34422863e-01 4.31232661e-01 5.30065417e-01
-8.73181820e-01 -5.30731119e-02 9.95628893e-01 -1.16446607e-01
4.07432407e-01 -6.53702974e-01 -3.66666690e-02 -9.04264152e-01
-2.73515165e-01 -1.35942125e+00 2.09374446e-02 -5.35057962e-01
-1.25720286e+00 1.04311347e-01 2.69579351e-01 -1.09977376e+00
-1.01615953e+00 -5.76546609e-01 -3.68775278e-01 5.31203389e-01
-1.38723600e+00 -6.66553557e-01 3.36222142e-01 2.56362081e-01
4.78987694e-01 -2.49397665e-01 6.59482479e-01 -3.05758208e-01
-3.94701958e-01 2.70804435e-01 3.95175725e-01 -3.42721373e-01
5.89405477e-01 -1.27306175e+00 2.94952124e-01 6.02003694e-01
-2.55183548e-01 4.43489015e-01 1.19392049e+00 -5.57829380e-01
-1.74784470e+00 -5.32516122e-01 -8.40494633e-02 -6.02511406e-01
7.34574616e-01 6.97641447e-02 -5.48759818e-01 2.79501677e-01
2.38453627e-01 -3.33765596e-01 -8.85376185e-02 6.46095127e-02
5.99128902e-02 9.04438123e-02 -1.19235718e+00 1.01703429e+00
9.87058878e-01 -1.09683633e-01 -6.24224126e-01 2.58858074e-02
7.20715880e-01 -4.88831043e-01 -8.79347563e-01 3.05667341e-01
7.40437746e-01 -7.88909018e-01 9.07018423e-01 -1.06955612e+00
3.07884514e-01 -2.57585794e-01 -3.33284922e-02 -2.00428200e+00
-4.86318797e-01 -1.04170620e+00 -4.22383219e-01 6.56420708e-01
2.99188823e-01 -8.03148627e-01 7.32069492e-01 8.39006782e-01
-1.84568912e-01 -9.01679575e-01 -8.58822942e-01 -1.14873087e+00
5.87004960e-01 -3.02372485e-01 5.31470239e-01 5.58541596e-01
2.11060360e-01 2.95620799e-01 -4.18865800e-01 -3.31019759e-01
7.00328469e-01 -1.39447644e-01 7.47584581e-01 -1.04080498e+00
-6.37394905e-01 -1.08082175e+00 7.43888021e-02 -1.06864917e+00
3.37230176e-01 -2.78074712e-01 2.48084068e-01 -1.79006422e+00
1.68101072e-01 -6.26716733e-01 -2.85140783e-01 4.68592972e-01
-2.68468559e-01 -5.85319102e-01 6.61352873e-01 -5.55705093e-02
-7.63255835e-01 9.32959437e-01 1.55217302e+00 -6.78487867e-02
-4.03606534e-01 -1.41850160e-02 -8.20385456e-01 7.59584427e-01
9.55447912e-01 -6.27810895e-01 -6.28423691e-01 -3.35906565e-01
2.04085305e-01 3.44635218e-01 2.98732311e-01 -1.04958487e+00
1.32082775e-01 -9.22769129e-01 1.23770900e-01 -7.44696483e-02
4.17225450e-01 -4.86077487e-01 -3.58513892e-01 1.01106608e+00
-6.57084882e-01 3.95716250e-01 3.22518468e-01 8.10781956e-01
2.80135214e-01 -2.54490167e-01 8.34052920e-01 -3.10871333e-01
-7.02606499e-01 3.05812448e-01 -7.15171039e-01 4.48384643e-01
9.04185891e-01 -4.03538942e-02 -5.07306814e-01 -6.71591043e-01
-3.56816769e-01 5.90386927e-01 4.67642635e-01 4.16053742e-01
3.84556532e-01 -1.12515378e+00 -6.95607662e-01 -2.65108556e-01
-2.76295722e-01 -3.72307062e-01 2.39369413e-03 5.85241199e-01
-6.74206838e-02 5.14281154e-01 -4.33400959e-01 -3.03414315e-01
-5.63996971e-01 6.22740030e-01 5.41429222e-01 -6.91697180e-01
-4.20232117e-01 4.65589643e-01 -5.49290627e-02 -5.38542032e-01
3.64798546e-01 -4.21625525e-01 1.93574578e-01 -1.40249133e-01
4.03758973e-01 4.69913691e-01 -4.76078063e-01 2.12247968e-02
-1.63516030e-01 2.88096756e-01 6.65944889e-02 -7.80406058e-01
1.37158692e+00 8.06277096e-02 5.74439287e-01 1.77718043e-01
2.83688247e-01 -6.38969362e-01 -2.27481842e+00 -6.38967454e-02
9.48744193e-02 -3.17987353e-01 1.60035715e-01 -1.18119776e+00
-4.73109335e-01 6.81286037e-01 7.14448094e-01 2.58813977e-01
7.32728601e-01 -3.96614045e-01 3.77142459e-01 7.48205185e-01
7.88389206e-01 -1.42323780e+00 3.58229250e-01 8.15683603e-01
7.80582845e-01 -1.35366142e+00 1.46988034e-01 7.42375553e-01
-8.92592072e-01 7.62241781e-01 8.19759071e-01 -4.10026103e-01
4.82126564e-01 2.67669827e-01 -2.07567036e-01 1.62279889e-01
-1.27821827e+00 -4.89497364e-01 -1.45440459e-01 9.90048468e-01
1.31031582e-02 9.60991010e-02 -2.93961018e-01 1.76849559e-01
-1.26713425e-01 1.78254768e-03 4.27300036e-01 1.24682510e+00
-7.92109072e-01 -1.21323502e+00 -2.47318059e-01 4.23075199e-01
-7.51124993e-02 9.78736877e-02 1.27566576e-01 1.02053344e+00
-4.77436125e-01 1.09895551e+00 -1.90417934e-02 -1.19485915e-01
3.63548458e-01 -1.19444191e-01 8.73515844e-01 -3.70084792e-01
-6.05840087e-01 -7.79074570e-03 2.27654397e-01 -7.99850464e-01
-2.01505825e-01 -6.87585950e-01 -1.29583275e+00 -4.27606285e-01
-2.32146960e-02 8.88055041e-02 6.89903378e-01 9.89315689e-01
5.56568384e-01 4.96838182e-01 4.63344425e-01 -1.33764064e+00
-1.51623237e+00 -8.10321510e-01 -3.00355583e-01 5.88525645e-02
5.48748791e-01 -1.13425362e+00 -1.67382900e-02 -4.60692763e-01] | [4.1407952308654785, 1.9398424625396729] |
6cc7fc78-4b7d-4c8b-928d-e24fb51f6273 | destruction-and-construction-learning-for | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Chen_Destruction_and_Construction_Learning_for_Fine-Grained_Image_Recognition_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Chen_Destruction_and_Construction_Learning_for_Fine-Grained_Image_Recognition_CVPR_2019_paper.pdf | Destruction and Construction Learning for Fine-Grained Image Recognition | Delicate feature representation about object parts plays a critical role in fine-grained recognition. For example, experts can even distinguish fine-grained objects relying only on object parts according to professional knowledge. In this paper, we propose a novel "Destruction and Construction Learning" (DCL) method to enhance the difficulty of fine-grained recognition and exercise the classification model to acquire expert knowledge. Besides the standard classification backbone network, another "destruction and construction" stream is introduced to carefully "destruct" and then "reconstruct" the input image, for learning discriminative regions and features. More specifically, for "destruction", we first partition the input image into local regions and then shuffle them by a Region Confusion Mechanism (RCM). To correctly recognize these destructed images, the classification network has to pay more attention to discriminative regions for spotting the differences. To compensate the noises introduced by RCM, an adversarial loss, which distinguishes original images from destructed ones, is applied to reject noisy patterns introduced by RCM. For "construction", a region alignment network, which tries to restore the original spatial layout of local regions, is followed to model the semantic correlation among local regions. By jointly training with parameter sharing, our proposed DCL injects more discriminative local details to the classification network. Experimental results show that our proposed framework achieves state-of-the-art performance on three standard benchmarks. Moreover, our proposed method does not need any external knowledge during training, and there is no computation overhead at inference time except the standard classification network feed-forwarding. Source code: https://github.com/JDAI-CV/DCL.
| [' Tao Mei', ' Wei Zhang', ' Yalong Bai', 'Yue Chen'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 1.80695012e-01 -1.53372750e-01 -2.05712304e-01 -3.95446002e-01
-5.82763672e-01 -7.63059378e-01 3.61615747e-01 -4.88462932e-02
-2.38560945e-01 4.79916811e-01 -1.28471985e-01 -1.87436327e-01
-1.76160961e-01 -9.41546500e-01 -9.27461505e-01 -9.29192305e-01
4.03418243e-01 1.88436538e-01 2.16800809e-01 9.31586623e-02
3.25665385e-01 9.57413137e-01 -1.36819994e+00 5.08303940e-01
7.25222409e-01 1.47408748e+00 3.90119433e-01 1.99635684e-01
-2.93718576e-01 7.49355376e-01 -6.36003673e-01 -1.36240035e-01
3.79995733e-01 -9.09184590e-02 -8.77669871e-01 3.41106355e-01
5.08695424e-01 -2.54173785e-01 -4.09096718e-01 1.33070779e+00
3.13049257e-01 1.81102410e-01 5.59807062e-01 -1.02870107e+00
-7.43638992e-01 5.07365346e-01 -6.50667727e-01 1.99619979e-01
3.14409211e-02 3.01798373e-01 7.83679783e-01 -9.52553630e-01
2.83375621e-01 1.13937795e+00 4.57606167e-01 6.66853189e-01
-1.13051438e+00 -9.20245945e-01 6.07007027e-01 2.35141695e-01
-1.67991006e+00 -3.39621007e-01 1.11643052e+00 -2.57105649e-01
3.78758699e-01 6.06399894e-01 4.38728452e-01 1.11687851e+00
5.89858070e-02 8.30824971e-01 1.24303925e+00 -1.36508048e-01
1.33427352e-01 1.83066353e-01 1.20171063e-01 6.90990329e-01
4.98356447e-02 7.71870017e-02 -2.78429091e-02 2.36944422e-01
9.33380663e-01 4.86925662e-01 -4.75586742e-01 -4.53529656e-01
-1.04060435e+00 4.47602808e-01 8.51261377e-01 5.18957078e-01
-2.15558052e-01 4.30049486e-02 3.17463964e-01 2.19982475e-01
2.95070767e-01 2.28331447e-01 -3.51161033e-01 2.88929075e-01
-7.56750226e-01 -1.94428280e-01 3.88383895e-01 8.03742290e-01
1.14629745e+00 -2.71275550e-01 -3.33101302e-01 9.66053545e-01
1.40117213e-01 2.52129078e-01 5.89270115e-01 -6.40972674e-01
6.18580580e-01 9.37495708e-01 1.68914516e-02 -1.26526964e+00
-1.01660296e-01 -8.00467789e-01 -1.42737639e+00 1.40364975e-01
3.67292523e-01 4.34017032e-01 -1.02333486e+00 1.38774037e+00
3.11269641e-01 2.57729441e-01 -2.87332386e-01 1.16333961e+00
6.24961853e-01 5.19254744e-01 -1.80482283e-01 1.90239429e-01
1.37426960e+00 -1.23877466e+00 -1.17957078e-01 -3.71282429e-01
3.98812294e-01 -6.05895519e-01 1.28432083e+00 3.97865802e-01
-7.44176984e-01 -7.24465549e-01 -9.57363605e-01 9.99322161e-02
-6.42684340e-01 3.30428690e-01 4.80941713e-01 3.77514035e-01
-4.93207812e-01 5.21096766e-01 -4.92960870e-01 1.22467965e-01
8.62602830e-01 2.72863954e-01 -5.63142419e-01 -4.04516339e-01
-9.40217614e-01 4.28154677e-01 4.72054988e-01 5.64784884e-01
-1.11509526e+00 -8.36027861e-01 -7.84362972e-01 1.44718841e-01
4.27605271e-01 -3.88759464e-01 7.92314827e-01 -1.23424852e+00
-1.17263961e+00 8.25411379e-01 5.04049808e-02 -9.30454284e-02
5.52585185e-01 3.63676995e-02 -4.95499432e-01 8.32622498e-02
1.59571871e-01 4.10201073e-01 1.08367455e+00 -1.51949024e+00
-6.51915133e-01 -4.07118887e-01 1.69427678e-01 -4.80790697e-02
-2.07754046e-01 -1.90186292e-01 -7.32607305e-01 -1.11376822e+00
1.20888948e-01 -6.37633264e-01 -2.45047197e-01 1.84158668e-01
-6.47433400e-01 -1.40655518e-01 8.69830549e-01 -5.61380148e-01
1.11186588e+00 -2.43597555e+00 -1.87040493e-02 7.03712225e-01
3.86184186e-01 2.57680297e-01 -3.17768008e-01 -1.01381108e-01
-1.26207083e-01 8.21658820e-02 -3.06136519e-01 -1.78885266e-01
3.21034975e-02 2.20283896e-01 -4.18667585e-01 5.42176485e-01
1.31620854e-01 9.69200313e-01 -5.71360946e-01 -3.87508512e-01
3.97334099e-01 3.22165400e-01 -3.35784048e-01 1.98448747e-01
-9.19445232e-02 4.17300522e-01 -6.73509836e-01 8.36976647e-01
1.06368542e+00 -4.94026840e-01 -3.03789862e-02 -6.45983338e-01
9.42190662e-02 -1.39107630e-01 -1.26116264e+00 1.74336541e+00
-5.87171674e-01 -3.58418236e-03 1.48229092e-01 -1.34185934e+00
1.01390052e+00 -1.79407805e-01 -3.09884027e-02 -9.59784746e-01
2.60298550e-01 1.18864164e-01 -2.53530890e-01 -1.71521306e-01
6.17015511e-02 1.26585960e-01 -4.49137270e-01 2.69998044e-01
-2.17647582e-01 3.37804884e-01 -2.49094144e-01 -7.13664517e-02
9.45931494e-01 -9.95046794e-02 1.64549425e-01 -1.69607520e-01
9.18993950e-01 -1.28510743e-01 7.44857132e-01 6.80479705e-01
-1.18880734e-01 5.05066097e-01 1.51262626e-01 -5.69455504e-01
-6.85636699e-01 -1.08836293e+00 9.57320929e-02 9.93434668e-01
7.04232454e-01 7.02172443e-02 -7.43250191e-01 -1.09915042e+00
1.19403757e-01 2.40464836e-01 -9.09742355e-01 -3.99063051e-01
-5.71529329e-01 -2.73839504e-01 6.62416101e-01 6.34826839e-01
9.23850179e-01 -1.20567226e+00 -2.50164926e-01 4.99200895e-02
-1.33476734e-01 -7.56537795e-01 -7.22305357e-01 2.77054876e-01
-6.22560084e-01 -1.03319633e+00 -7.28042781e-01 -1.02378261e+00
1.14325738e+00 4.99681771e-01 8.52554977e-01 2.64962137e-01
-4.95583683e-01 -9.87462327e-02 -3.46840531e-01 3.16436052e-01
-7.67853186e-02 9.82311890e-02 -2.52537131e-01 3.77368748e-01
1.14433065e-01 -6.38719916e-01 -9.14033711e-01 6.92592144e-01
-1.04000723e+00 -3.61044258e-02 1.10616171e+00 1.13456321e+00
8.67756248e-01 4.83526409e-01 3.37581724e-01 -9.41308022e-01
2.60437906e-01 -3.31433773e-01 -5.61582565e-01 4.14963305e-01
-4.36624348e-01 1.91170797e-01 1.14269292e+00 -5.82817197e-01
-1.07808876e+00 -1.14560477e-04 3.03023271e-02 -9.34070349e-01
-6.27453446e-01 7.76525661e-02 -6.83703125e-01 -4.01906878e-01
3.17107052e-01 5.52665174e-01 -2.97219157e-01 -9.23399806e-01
4.33777392e-01 5.69402337e-01 7.24901617e-01 -7.32983649e-01
1.06495297e+00 5.76134682e-01 -3.89658332e-01 -1.52073905e-01
-8.60226333e-01 -4.29648489e-01 -4.66326773e-01 7.08451197e-02
6.49015427e-01 -7.78098524e-01 -7.18594968e-01 5.90677202e-01
-8.79220247e-01 -2.43405461e-01 -4.05861020e-01 1.14874654e-01
-1.77394986e-01 3.81694347e-01 -5.58751643e-01 -3.56467843e-01
-2.73687512e-01 -1.01854932e+00 9.87391651e-01 4.55482423e-01
3.26654434e-01 -7.82317817e-01 -4.24430013e-01 4.82685447e-01
3.22334498e-01 4.99110930e-02 9.21108007e-01 -3.37254196e-01
-8.49444628e-01 -1.95511714e-01 -7.22600639e-01 5.26583374e-01
4.58233625e-01 -5.78800678e-01 -9.39329982e-01 -4.01589096e-01
-1.58735797e-01 -2.69039929e-01 9.68844175e-01 -1.43772423e-01
1.94958854e+00 -7.27599382e-01 -5.07528007e-01 8.52285564e-01
1.49786901e+00 1.38212666e-01 6.95100069e-01 4.00256991e-01
1.06728625e+00 6.05512440e-01 6.34284437e-01 3.83579493e-01
1.13429829e-01 4.23818260e-01 5.36801755e-01 -2.54857957e-01
-2.51590908e-01 -4.05213475e-01 7.85020813e-02 4.78093833e-01
1.59927145e-01 -1.40427545e-01 -3.88650715e-01 5.24899542e-01
-1.54542184e+00 -8.39783430e-01 4.06372339e-01 2.01435065e+00
8.08926404e-01 9.11321938e-02 -2.70317435e-01 1.14025377e-01
9.03200388e-01 4.06299114e-01 -7.93592870e-01 -5.83880879e-02
-2.40473211e-01 3.13929558e-01 6.21196687e-01 2.32550725e-01
-1.22855258e+00 1.00261855e+00 4.01913214e+00 1.57342255e+00
-1.42322135e+00 2.04920564e-02 8.10663879e-01 2.04676285e-01
-3.68276983e-01 -1.13845192e-01 -6.20323598e-01 6.69817746e-01
1.22910962e-01 4.67507064e-01 7.15259314e-01 9.14392173e-01
6.17229007e-02 2.00219929e-01 -8.48752201e-01 9.91933465e-01
-6.82630856e-03 -1.42080438e+00 3.25965822e-01 -9.61836427e-02
5.84266603e-01 -3.76682431e-01 8.86435509e-02 1.64307356e-01
1.10381609e-02 -1.13340902e+00 7.96914458e-01 6.56418681e-01
1.05854225e+00 -7.77786016e-01 6.93641841e-01 4.26017702e-01
-1.55704856e+00 -1.51017725e-01 -5.35229445e-01 2.93075114e-01
-3.77550721e-01 7.84300745e-01 -3.10345292e-01 7.32628584e-01
9.00864840e-01 5.39212346e-01 -7.28731573e-01 7.94556022e-01
-2.41594449e-01 3.03528786e-01 -2.93847360e-02 3.26127112e-01
1.80033818e-01 -6.67652562e-02 2.30826944e-01 1.12503672e+00
1.44240633e-01 -1.35318428e-01 3.26559365e-01 8.87106001e-01
-3.41284454e-01 -1.03128225e-01 -3.44260126e-01 2.24985421e-01
5.41131854e-01 1.40960979e+00 -8.51116359e-01 -3.61937314e-01
-2.47753337e-01 1.43055403e+00 5.46574056e-01 2.65149206e-01
-8.80739808e-01 -6.09826922e-01 5.55838406e-01 6.49221614e-02
7.20475018e-01 2.29267597e-01 -1.70021206e-01 -1.21060157e+00
3.60709280e-01 -9.36584830e-01 3.85101259e-01 -5.96700847e-01
-1.61754620e+00 5.44113517e-01 -3.43986958e-01 -1.34123254e+00
3.64585459e-01 -5.00923812e-01 -6.06323421e-01 8.16903710e-01
-1.66127503e+00 -1.36403334e+00 -6.64341331e-01 1.02027285e+00
5.40976644e-01 9.97382123e-03 5.39088130e-01 6.28409028e-01
-7.06267059e-01 1.07211196e+00 -1.36033073e-01 3.91262889e-01
5.90913773e-01 -1.07229364e+00 1.24119394e-01 7.43854046e-01
-9.54117030e-02 6.20991409e-01 1.83646739e-01 -4.90744263e-01
-1.24883831e+00 -1.73743391e+00 4.27537680e-01 4.18916456e-02
4.15341973e-01 -5.95734179e-01 -1.06496310e+00 5.62159061e-01
-3.18917245e-01 6.19156659e-01 2.88610548e-01 -3.60125482e-01
-9.41345036e-01 -6.22514427e-01 -1.35904074e+00 3.56515855e-01
1.11868834e+00 -7.28982508e-01 -5.27031660e-01 1.01835370e-01
6.80632412e-01 -1.29240245e-01 -8.82183433e-01 4.95022893e-01
6.02008581e-01 -8.03592443e-01 1.06609941e+00 -3.39568138e-01
1.47572383e-01 -7.71986067e-01 -1.84974551e-01 -1.22481549e+00
-6.21709049e-01 -1.80932343e-01 2.30091158e-02 1.37610471e+00
2.11908847e-01 -8.68909240e-01 9.97531235e-01 9.31630358e-02
-2.02408135e-01 -8.55529428e-01 -8.67784142e-01 -8.85755956e-01
-3.39132249e-02 -2.54294306e-01 8.87715995e-01 9.89921808e-01
-7.24941969e-01 -3.60435136e-02 -2.82766610e-01 5.34489870e-01
6.27659976e-01 5.96210778e-01 5.41080892e-01 -9.92925882e-01
-3.88855308e-01 -5.59620261e-01 -3.76032352e-01 -1.29371607e+00
1.73927337e-01 -8.72820854e-01 1.22858696e-01 -1.15191114e+00
2.65820801e-01 -8.90251219e-01 -7.31038749e-01 7.16653049e-01
-3.93707938e-02 6.41415060e-01 9.04402062e-02 3.72919500e-01
-7.91367590e-01 5.18884420e-01 1.53535986e+00 -5.62925816e-01
9.83891860e-02 -8.08809176e-02 -1.00048184e+00 5.96071422e-01
8.59124005e-01 -3.24033976e-01 -2.34357104e-01 -3.21403503e-01
-1.38986617e-01 -1.58602998e-01 7.56424725e-01 -9.94169950e-01
2.17483088e-01 -2.24244922e-01 7.60659933e-01 -4.73784328e-01
1.55876130e-01 -1.04574740e+00 2.12485045e-01 3.81717563e-01
-1.10647634e-01 -2.69663811e-01 1.63392216e-01 5.04478037e-01
-3.64991665e-01 1.47705749e-02 9.91043329e-01 -1.96139172e-01
-9.36818004e-01 5.35542130e-01 -3.33510414e-02 -5.35510853e-02
1.12027097e+00 -2.54200131e-01 -6.28016293e-01 2.03827173e-01
-6.21672869e-01 8.81550238e-02 5.68055570e-01 3.91715169e-01
7.63516784e-01 -1.46058869e+00 -3.00285101e-01 4.85045880e-01
3.70220929e-01 3.48131359e-01 7.62741387e-01 6.38787627e-01
-3.95758629e-01 2.68785506e-01 -1.92446440e-01 -3.75593841e-01
-9.93806720e-01 9.23064828e-01 4.30428803e-01 -4.24316406e-01
-6.37827933e-01 8.50818574e-01 8.02606523e-01 -5.49926460e-01
2.64191091e-01 -4.16416496e-01 9.28703230e-03 -1.68127015e-01
4.48524415e-01 1.42751321e-01 1.32991672e-01 -4.51577663e-01
-5.64298749e-01 7.87034452e-01 -4.32375342e-01 4.99897927e-01
1.01994407e+00 -2.06599429e-01 -1.49424285e-01 8.84763747e-02
1.37732208e+00 1.63988337e-01 -1.31264091e+00 -3.03473562e-01
-3.47651988e-01 -5.89456081e-01 3.71239856e-02 -1.01400363e+00
-1.45929325e+00 8.44844759e-01 7.88879037e-01 1.11923721e-02
1.52976680e+00 1.77076951e-01 8.56620491e-01 2.25432321e-01
3.34047854e-01 -7.84497023e-01 1.47882095e-02 3.70274276e-01
8.67963314e-01 -1.02067673e+00 -2.15028167e-01 -4.48691994e-01
-2.60324925e-01 7.96011150e-01 7.82407880e-01 -4.41444188e-01
6.32040560e-01 1.24490105e-01 -5.48351705e-02 -1.32889062e-01
-3.68384928e-01 2.39779837e-02 4.11208779e-01 4.79873896e-01
-2.74877846e-01 2.37710506e-01 1.24215297e-01 8.58437240e-01
9.17803571e-02 -2.05016598e-01 -2.08894685e-01 7.15947807e-01
-4.06704068e-01 -1.01790380e+00 -4.76399601e-01 4.31121111e-01
-2.52816647e-01 -1.27280444e-01 -3.59662205e-01 6.10375166e-01
6.44396842e-01 6.77846134e-01 2.80430853e-01 -5.64184427e-01
3.10411751e-01 -3.85225683e-01 3.69720906e-01 -3.77001584e-01
-6.17760301e-01 -4.04869951e-02 -2.37614140e-01 -8.40524316e-01
-1.40396923e-01 -1.49430811e-01 -1.16649377e+00 -3.04520518e-01
-2.11499199e-01 3.22752565e-01 3.35419446e-01 9.87733662e-01
4.85427916e-01 7.79794276e-01 1.07259345e+00 -7.42728293e-01
-3.91089529e-01 -7.41608322e-01 -6.81506038e-01 4.71129924e-01
3.53738040e-01 -6.49034321e-01 -5.02625704e-01 -1.80843562e-01] | [9.641033172607422, 2.0254297256469727] |
dc09f51f-53d0-4a86-a4b6-0dfa2a7afa85 | silk-simple-learned-keypoints | 2304.06194 | null | https://arxiv.org/abs/2304.06194v1 | https://arxiv.org/pdf/2304.06194v1.pdf | SiLK -- Simple Learned Keypoints | Keypoint detection & descriptors are foundational tech-nologies for computer vision tasks like image matching, 3D reconstruction and visual odometry. Hand-engineered methods like Harris corners, SIFT, and HOG descriptors have been used for decades; more recently, there has been a trend to introduce learning in an attempt to improve keypoint detectors. On inspection however, the results are difficult to interpret; recent learning-based methods employ a vast diversity of experimental setups and design choices: empirical results are often reported using different backbones, protocols, datasets, types of supervisions or tasks. Since these differences are often coupled together, it raises a natural question on what makes a good learned keypoint detector. In this work, we revisit the design of existing keypoint detectors by deconstructing their methodologies and identifying the key components. We re-design each component from first-principle and propose Simple Learned Keypoints (SiLK) that is fully-differentiable, lightweight, and flexible. Despite its simplicity, SiLK advances new state-of-the-art on Detection Repeatability and Homography Estimation tasks on HPatches and 3D Point-Cloud Registration task on ScanNet, and achieves competitive performance to state-of-the-art on camera pose estimation in 2022 Image Matching Challenge and ScanNet. | ['Matt Feiszli', 'Weiyao Wang', 'Pierre Gleize'] | 2023-04-12 | null | null | null | null | ['point-cloud-registration', 'keypoint-detection', '3d-reconstruction', 'homography-estimation', 'visual-odometry'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'robots'] | [-2.81349599e-01 -3.79646361e-01 -5.43622851e-01 -1.30859286e-01
-5.74685395e-01 -6.55162096e-01 7.81183720e-01 7.56767858e-03
-4.24838632e-01 -2.02876702e-02 -1.13030791e-01 -1.58621550e-01
-2.06498861e-01 -3.83806825e-01 -1.01703393e+00 -4.58242506e-01
-2.08262682e-01 4.15896773e-01 4.90218639e-01 -3.12995374e-01
5.63357830e-01 8.81625831e-01 -1.57143438e+00 -4.60293293e-01
4.77742136e-01 9.55361605e-01 9.20580775e-02 5.29207587e-01
3.80987264e-02 4.54202771e-01 -1.96055487e-01 -4.54932868e-01
7.47453034e-01 1.94444004e-02 -4.76305217e-01 -1.07185595e-01
1.15650833e+00 -3.44558179e-01 -7.75995016e-01 1.03137720e+00
5.27359247e-01 -2.61756271e-01 4.20461953e-01 -1.65179908e+00
-5.09799361e-01 2.39501894e-01 -6.67940259e-01 -3.53866033e-02
4.32913572e-01 2.41614312e-01 1.02548099e+00 -1.11534989e+00
7.58247793e-01 1.09852314e+00 1.26765180e+00 3.71463150e-01
-1.18778741e+00 -8.34863365e-01 3.49671394e-02 2.74396151e-01
-1.49583602e+00 -3.47854823e-01 7.01355219e-01 -3.89881194e-01
1.05547345e+00 -9.67644155e-03 7.39577889e-01 9.76149440e-01
3.13394755e-01 8.59340668e-01 9.25473273e-01 -3.96733820e-01
-9.58085060e-02 5.72676845e-02 -5.51677532e-02 8.74933898e-01
5.31260848e-01 3.67314816e-01 -6.64740503e-01 -1.66594207e-01
9.63850856e-01 2.13684455e-01 -1.91565290e-01 -1.31664205e+00
-1.57176077e+00 8.00788581e-01 5.88331103e-01 -1.13160774e-01
-2.19121143e-01 4.85136181e-01 3.34164858e-01 6.45082235e-01
-4.92338091e-02 4.87244278e-01 -3.83218259e-01 -1.47280276e-01
-8.13830733e-01 5.05257666e-01 8.52888584e-01 1.38293982e+00
1.04622424e+00 -2.34834090e-01 3.04373473e-01 4.53432173e-01
4.36948061e-01 8.29316735e-01 2.76255548e-01 -8.68435800e-01
3.65223080e-01 6.51078463e-01 -3.04892510e-02 -1.35687125e+00
-4.62630540e-01 -2.02790469e-01 -6.27134621e-01 2.74655223e-01
2.47975349e-01 2.21024871e-01 -9.43864644e-01 1.38989675e+00
3.65152091e-01 1.52230889e-01 -3.57271701e-01 9.63992655e-01
6.46084249e-01 9.99099240e-02 -2.96697199e-01 3.76561850e-01
1.17144382e+00 -9.34083760e-01 -1.47493005e-01 -3.43445063e-01
6.02210701e-01 -9.61402774e-01 7.90340960e-01 2.63231963e-01
-7.17583895e-01 -6.51865900e-01 -1.23065066e+00 -1.34840325e-01
-3.03639650e-01 -7.14583769e-02 8.41220915e-01 5.73795497e-01
-1.04659212e+00 6.59401834e-01 -7.76091993e-01 -7.92351663e-01
2.35950723e-02 5.76344609e-01 -6.63325727e-01 -6.51923046e-02
-8.72465372e-01 1.08764243e+00 1.21384159e-01 -1.56219274e-01
-6.79153562e-01 -9.77115333e-01 -8.22154343e-01 -3.11552852e-01
4.59549069e-01 -1.04852092e+00 1.24204361e+00 -3.70017022e-01
-1.39606106e+00 1.15472448e+00 1.53710634e-01 -5.56499243e-01
7.65998423e-01 -4.21384305e-01 -2.28858337e-01 7.09149688e-02
1.58481255e-01 9.76019561e-01 1.14694488e+00 -1.10980177e+00
-9.78293121e-01 -3.10896605e-01 -7.74368197e-02 8.98066387e-02
9.82015505e-02 -7.12885708e-02 -7.48425305e-01 -3.84279221e-01
5.28993189e-01 -1.37614655e+00 -1.71564415e-01 4.14277017e-01
-1.96893215e-01 -1.54033303e-01 8.30116570e-01 -1.94325112e-02
8.02718580e-01 -2.18781161e+00 -3.76325808e-02 2.36560643e-01
3.57660413e-01 1.12687044e-01 8.48998427e-02 5.65843165e-01
3.73490676e-02 -1.08347103e-01 2.35906079e-01 -3.77475798e-01
4.10915971e-01 1.43453151e-01 -5.36397219e-01 9.11458731e-01
8.14646184e-02 9.50689912e-01 -8.84923816e-01 -4.70840394e-01
7.30176449e-01 5.04157126e-01 -5.42489111e-01 2.18065064e-02
1.44391298e-01 1.50480747e-01 -3.85670513e-01 1.02619314e+00
7.23665297e-01 -1.66769817e-01 -3.32852423e-01 -5.75371802e-01
-3.93213123e-01 1.50590584e-01 -1.63017082e+00 1.97379267e+00
-2.41713449e-01 6.97207510e-01 -1.03126876e-01 -7.51231790e-01
9.10722256e-01 -2.39566006e-02 7.39994049e-01 -6.08442366e-01
7.81233832e-02 5.01972795e-01 -2.75562853e-01 -2.31128767e-01
7.47898877e-01 3.94623905e-01 -2.05761507e-01 1.17135391e-01
1.01395227e-01 -4.18675065e-01 -8.00766349e-02 8.09169933e-02
1.21310341e+00 2.36447647e-01 4.79320019e-01 -1.62040770e-01
3.29529881e-01 3.41271549e-01 3.40428472e-01 8.89343202e-01
-3.33401620e-01 8.14483404e-01 2.06702083e-01 -7.31869519e-01
-1.17485058e+00 -1.03372097e+00 -2.82734662e-01 7.67589569e-01
5.48170209e-01 -5.94134569e-01 -3.24179530e-01 -5.94736159e-01
6.74005628e-01 -2.26266488e-01 -5.00285923e-01 -1.37207150e-01
-7.93929636e-01 -9.64558125e-02 7.66493142e-01 5.29474497e-01
5.74926078e-01 -5.17693222e-01 -8.73326898e-01 7.46560469e-02
2.80438900e-01 -1.30815995e+00 -5.92910588e-01 2.74623483e-01
-7.27157950e-01 -1.22024500e+00 -5.65450907e-01 -7.69388080e-01
5.91427684e-01 9.49572146e-01 1.28437030e+00 7.11071491e-02
-4.61392313e-01 8.41938019e-01 -2.51316845e-01 -4.18501735e-01
8.35572090e-03 2.75103539e-01 4.73383278e-01 -3.25790733e-01
5.10235310e-01 -5.27223647e-01 -8.97060573e-01 6.41710699e-01
-5.23413479e-01 -3.28797907e-01 1.02750456e+00 7.76078463e-01
7.04164684e-01 -6.25343144e-01 -3.24373245e-01 -4.40238148e-01
2.18183652e-01 -5.43677844e-02 -1.01237285e+00 1.92014202e-01
-8.83148015e-01 3.56104612e-01 3.36270958e-01 -4.35586601e-01
-2.27437109e-01 3.15398097e-01 4.20649536e-02 -8.02871168e-01
-8.97827595e-02 7.35956132e-02 1.81970999e-01 -1.01054788e+00
7.28514791e-01 1.60147101e-01 2.65276786e-02 -3.85672957e-01
5.15259504e-01 5.13068318e-01 7.90337622e-01 -5.69584489e-01
1.31597316e+00 7.49223351e-01 2.28024170e-01 -9.21890140e-01
-4.17313457e-01 -9.70532477e-01 -8.29400361e-01 2.31793337e-02
5.65187514e-01 -1.16359234e+00 -1.03804028e+00 5.28711557e-01
-1.09467697e+00 3.95740904e-02 -1.65542886e-01 5.40136218e-01
-6.79075539e-01 4.45151180e-01 -1.30781889e-01 -2.54741549e-01
-4.45224613e-01 -1.25260472e+00 1.50374532e+00 2.17141390e-01
-7.09158182e-02 -7.36779511e-01 3.40579480e-01 -1.36554345e-01
4.79728699e-01 2.04657361e-01 3.77748132e-01 -2.57354796e-01
-1.08622944e+00 -4.81417060e-01 -3.85862440e-01 5.62633537e-02
-1.19096391e-01 7.43166059e-02 -8.01276386e-01 -6.22035205e-01
-3.42198372e-01 -2.67856866e-01 7.58884430e-01 2.08234489e-01
6.05773449e-01 1.46369323e-01 -6.26448214e-01 1.27652836e+00
1.47904849e+00 -3.33742976e-01 4.83490556e-01 8.73609006e-01
8.65760744e-01 3.66695017e-01 6.13343775e-01 2.47592434e-01
6.81725681e-01 1.01260614e+00 5.80321312e-01 -8.43640566e-02
-5.72108328e-02 -5.72698951e-01 4.04917806e-01 5.61488986e-01
-1.64446533e-02 4.87720311e-01 -9.86367702e-01 4.18648362e-01
-1.84709799e+00 -7.78048217e-01 1.66394729e-02 2.20353293e+00
3.87944579e-01 2.03624412e-01 1.64480776e-01 -1.25378653e-01
4.63925481e-01 1.66468158e-01 -5.97763598e-01 1.57370910e-01
-9.28733200e-02 -3.74969095e-02 1.28967655e+00 5.10093234e-02
-1.23835325e+00 1.01370490e+00 6.36790514e+00 4.81893241e-01
-1.44070828e+00 -2.72032231e-01 -2.66994029e-01 4.07696247e-01
3.50477621e-02 4.25072670e-01 -1.25319171e+00 1.03666522e-01
5.20175159e-01 -1.08292818e-01 1.82269275e-01 1.21120512e+00
-3.18719059e-01 8.33833888e-02 -1.35373831e+00 1.53768241e+00
1.95736721e-01 -1.44305277e+00 -1.44714430e-01 1.22577362e-01
5.54078877e-01 7.10896552e-01 1.17372602e-01 2.15300426e-01
2.88660824e-01 -6.79130614e-01 8.59278798e-01 1.36216849e-01
6.49625123e-01 -3.83005947e-01 5.21411836e-01 1.14686243e-01
-1.39975739e+00 -4.14930023e-02 -5.71507335e-01 1.18205868e-01
-6.00690767e-02 1.78198338e-01 -7.08689690e-01 5.58022738e-01
9.74636495e-01 9.03952539e-01 -7.44263053e-01 1.54416358e+00
-3.27796727e-01 2.90776212e-02 -5.67695677e-01 1.10855684e-01
3.06837022e-01 3.26885581e-02 6.46424174e-01 1.18634832e+00
2.79203057e-01 -3.14523429e-01 4.42614377e-01 6.00771546e-01
-5.15511818e-02 -1.44121736e-01 -8.11436892e-01 3.16828966e-01
7.43121207e-01 1.35613823e+00 -6.64038062e-01 8.05743784e-03
-7.38442004e-01 7.91410446e-01 7.16736391e-02 -3.08448672e-02
-8.42884839e-01 -4.82581198e-01 1.16143358e+00 2.97746003e-01
4.55672741e-01 -6.71940923e-01 6.14719093e-02 -1.19735742e+00
1.69662058e-01 -8.87562573e-01 1.59235775e-01 -3.93875986e-01
-1.12669933e+00 2.13578492e-01 1.37864202e-01 -1.72411859e+00
-2.91644096e-01 -8.63202512e-01 -3.66558522e-01 4.57365930e-01
-1.88132215e+00 -1.31061399e+00 -7.13591158e-01 7.34957755e-01
3.51092309e-01 -1.37868613e-01 4.95352477e-01 3.82840842e-01
-7.52404630e-02 6.65397942e-01 5.71604334e-02 3.17677826e-01
1.16358495e+00 -1.18203783e+00 1.05517435e+00 8.14043820e-01
4.94821668e-01 8.99880469e-01 7.31750190e-01 -3.17963779e-01
-2.38928080e+00 -6.36121809e-01 5.70207059e-01 -7.64313579e-01
8.95156980e-01 -5.19580662e-01 -5.66319227e-01 6.81535661e-01
-7.77419582e-02 4.25903618e-01 -1.99768078e-02 5.83241060e-02
-6.92793965e-01 -3.96923274e-01 -8.14723253e-01 5.89495361e-01
1.23430324e+00 -6.54200792e-01 -6.10176980e-01 1.50554851e-01
7.72177815e-01 -8.08566868e-01 -7.75460064e-01 4.28644091e-01
7.52079844e-01 -9.66937244e-01 1.38154817e+00 -1.14767954e-01
-2.80866563e-01 -5.56034267e-01 -1.36222869e-01 -1.06154811e+00
-4.14847583e-01 -9.45546627e-01 -6.08467124e-03 8.22503746e-01
-6.96929917e-02 -6.01858139e-01 1.08629739e+00 3.49383473e-01
-2.31997594e-01 -3.62651139e-01 -9.15226042e-01 -1.06743121e+00
-2.11053252e-01 -4.06063408e-01 5.48818171e-01 8.57136726e-01
-2.33692110e-01 1.77820683e-01 -5.10933816e-01 2.48952359e-01
9.46422756e-01 1.09845102e-01 1.58671427e+00 -1.37470138e+00
-5.99569567e-02 -5.17580926e-01 -1.09404540e+00 -1.46360791e+00
-1.18109286e-01 -6.09305978e-01 -1.21654779e-01 -1.05069268e+00
-1.68739140e-01 -5.08571088e-01 -8.68633091e-02 4.03560668e-01
1.89386785e-01 2.27614090e-01 4.19535637e-01 6.56917751e-01
-7.14101374e-01 1.43015414e-01 8.25363755e-01 -2.47695550e-01
-2.69423902e-01 -1.63481885e-03 -4.71850693e-01 5.68896592e-01
3.52307618e-01 -4.38086867e-01 -1.38168335e-01 -6.37066185e-01
4.43101108e-01 -3.39428902e-01 7.25133061e-01 -1.20215952e+00
8.49265516e-01 3.89132686e-02 1.94201991e-01 -8.00091684e-01
1.22389831e-01 -1.01667619e+00 2.41042599e-01 6.10276878e-01
1.45200223e-01 6.48704469e-01 9.33213905e-03 5.53174198e-01
-2.77317435e-01 1.90406606e-01 6.26986384e-01 -2.03177691e-01
-1.38030231e+00 6.15592301e-01 3.83445293e-01 5.05505390e-02
9.60179150e-01 -6.64539695e-01 -3.94100189e-01 -5.07433563e-02
8.05468038e-02 1.57838136e-01 9.34877694e-01 7.59372115e-01
6.15569711e-01 -1.38634062e+00 -5.76867223e-01 3.96365434e-01
6.38945699e-01 -1.08526079e-02 -9.43585038e-02 1.07530105e+00
-8.26065540e-01 6.04628265e-01 -3.16269219e-01 -1.14156473e+00
-1.10838401e+00 5.93278646e-01 2.15398252e-01 3.86599638e-02
-8.17828238e-01 6.42306030e-01 -1.45108745e-01 -5.18144786e-01
4.03419375e-01 -5.24326265e-01 2.80948132e-01 2.98619121e-02
2.40647212e-01 3.03216547e-01 2.90383846e-01 -7.04045534e-01
-6.10392213e-01 1.36014712e+00 -2.24908471e-01 3.78115445e-01
1.26867092e+00 -1.30728066e-01 1.37450263e-01 8.59136954e-02
1.33592916e+00 -5.31847589e-02 -1.38505709e+00 -4.68609720e-01
2.49475375e-01 -7.50918269e-01 -1.11399643e-01 -1.48482695e-01
-8.13990116e-01 6.58827484e-01 7.65093863e-01 3.25740129e-02
6.92360699e-01 2.31270805e-01 7.43074656e-01 7.37923861e-01
8.92098963e-01 -1.03842747e+00 6.23110645e-02 6.45646632e-01
5.60827017e-01 -1.55473292e+00 2.80328125e-01 -2.77719140e-01
-3.62853289e-01 1.17246783e+00 4.01933402e-01 -4.52057362e-01
5.91146708e-01 2.82457262e-01 2.02430964e-01 -3.82339597e-01
-4.35175061e-01 -3.10858309e-01 3.78823549e-01 7.72600591e-01
7.78004602e-02 -2.35479534e-01 7.52352923e-02 -5.73151782e-02
-2.92532563e-01 -1.81701824e-01 1.47506893e-01 1.15226710e+00
-5.06232917e-01 -1.22423363e+00 -5.70058227e-01 2.29012147e-02
-2.85338778e-02 1.19890263e-02 -3.40820312e-01 1.38640714e+00
-9.61311087e-02 2.85365939e-01 -1.16734236e-01 -6.89870000e-01
8.50688994e-01 -3.44035000e-01 6.04122043e-01 -2.03917205e-01
-4.86404479e-01 -2.89112121e-01 -4.42294031e-01 -9.84880865e-01
-4.47641343e-01 -7.67972052e-01 -8.29826713e-01 -6.06089354e-01
-4.26605076e-01 -1.37549028e-01 9.05460000e-01 6.79088056e-01
5.82394660e-01 -1.33758128e-01 7.11277008e-01 -1.21937931e+00
-9.33467925e-01 -4.18773979e-01 -3.00841182e-01 4.64303702e-01
5.34024715e-01 -8.92217517e-01 -4.10632312e-01 -1.82081550e-01] | [7.863284587860107, -2.171964168548584] |
cfbee086-7bf5-4231-bb1b-153c7fd07fc4 | sim-to-real-6d-object-pose-estimation-via | 2204.07049 | null | https://arxiv.org/abs/2204.07049v2 | https://arxiv.org/pdf/2204.07049v2.pdf | Sim-to-Real 6D Object Pose Estimation via Iterative Self-training for Robotic Bin Picking | In this paper, we propose an iterative self-training framework for sim-to-real 6D object pose estimation to facilitate cost-effective robotic grasping. Given a bin-picking scenario, we establish a photo-realistic simulator to synthesize abundant virtual data, and use this to train an initial pose estimation network. This network then takes the role of a teacher model, which generates pose predictions for unlabeled real data. With these predictions, we further design a comprehensive adaptive selection scheme to distinguish reliable results, and leverage them as pseudo labels to update a student model for pose estimation on real data. To continuously improve the quality of pseudo labels, we iterate the above steps by taking the trained student model as a new teacher and re-label real data using the refined teacher model. We evaluate our method on a public benchmark and our newly-released dataset, achieving an ADD(-S) improvement of 11.49% and 22.62% respectively. Our method is also able to improve robotic bin-picking success by 19.54%, demonstrating the potential of iterative sim-to-real solutions for robotic applications. | ['Qi Dou', 'Pieter Abbeel', 'Yun-hui Liu', 'Yichuan Li', 'Stephen James', 'Rui Cao', 'Kai Chen'] | 2022-04-14 | null | null | null | null | ['6d-pose-estimation', 'robotic-grasping'] | ['computer-vision', 'robots'] | [ 2.97433615e-01 2.18708158e-01 -8.54764059e-02 -4.43180591e-01
-1.03700376e+00 -7.55012512e-01 2.49648303e-01 -8.28837529e-02
-5.10083616e-01 5.79980016e-01 -2.85569340e-01 -1.44217968e-01
5.64765707e-02 -6.53933525e-01 -1.22173059e+00 -6.48253262e-01
-2.02194974e-01 1.13573527e+00 4.02369261e-01 -2.62503803e-01
3.92260104e-01 7.72960961e-01 -1.60400140e+00 1.23919494e-01
9.59152102e-01 1.05291653e+00 8.24362457e-01 5.27063966e-01
1.59752354e-01 4.75435674e-01 -5.24769127e-01 -1.29012838e-01
6.00757718e-01 1.81083232e-01 -8.02839279e-01 2.24467009e-01
1.82603985e-01 -7.02446401e-01 -1.60041481e-01 7.33123243e-01
6.24447167e-01 9.75494385e-02 5.78492701e-01 -1.22974527e+00
-1.91214800e-01 7.57314384e-01 -5.56003392e-01 -6.26438141e-01
5.07450938e-01 3.22931021e-01 6.89126670e-01 -9.68301356e-01
8.62495005e-01 1.33838165e+00 5.13279498e-01 6.54794753e-01
-1.13712478e+00 -7.43235290e-01 1.90156192e-01 1.10033236e-01
-9.68384743e-01 -2.12413624e-01 6.81910455e-01 -2.32421771e-01
6.17600322e-01 -4.01335061e-02 7.84394562e-01 1.21765828e+00
-4.84396741e-02 1.12376380e+00 7.48276055e-01 -3.07755411e-01
2.41316274e-01 -1.50102496e-01 -2.44424358e-01 7.14348018e-01
9.02156755e-02 -1.06134815e-02 -2.21848860e-01 -7.48615116e-02
8.39347959e-01 -5.80571480e-02 -1.57392666e-01 -1.08237886e+00
-1.41302478e+00 4.90018040e-01 8.83844793e-01 -3.48500699e-01
-5.87002635e-01 3.06011736e-01 2.39121810e-01 2.18246561e-02
7.90212005e-02 7.56973684e-01 -5.76054573e-01 -5.26789464e-02
-2.57540852e-01 5.42236686e-01 8.03533137e-01 1.36954474e+00
7.25900531e-01 -3.89340073e-01 -1.33161530e-01 9.50603366e-01
2.47239232e-01 6.83692157e-01 2.63997465e-01 -1.30953157e+00
4.95640725e-01 5.38543522e-01 5.00536859e-01 -5.63046753e-01
-5.04783750e-01 -3.80991876e-01 -2.30385989e-01 3.43840748e-01
4.28831637e-01 1.33837223e-01 -1.34274662e+00 1.58739686e+00
7.60681152e-01 -9.81602669e-02 2.45830715e-02 1.06507230e+00
5.65076709e-01 3.28458101e-01 5.68603491e-03 1.89708665e-01
8.95322382e-01 -1.19049656e+00 -9.63888541e-02 -1.87428638e-01
4.81756955e-01 -7.33827293e-01 1.01279843e+00 7.39083409e-01
-9.66773152e-01 -5.73816657e-01 -9.66036677e-01 8.05380195e-02
-6.64881095e-02 5.17647088e-01 5.94079912e-01 3.85329011e-03
-7.14510918e-01 1.03578103e+00 -1.18320155e+00 -1.95848837e-01
5.82523167e-01 6.01353943e-01 -4.29291159e-01 -3.66250604e-01
-4.74187165e-01 1.00572336e+00 5.87314308e-01 5.27646318e-02
-1.46606421e+00 -6.74409986e-01 -7.08723187e-01 -3.84675980e-01
7.06452072e-01 -4.94212061e-01 1.52516103e+00 -3.12516361e-01
-1.76379085e+00 6.56782568e-01 1.47268057e-01 -1.20355211e-01
9.01329935e-01 -3.70032758e-01 4.06291097e-01 9.95374918e-02
2.19922230e-01 1.19461739e+00 7.29387701e-01 -1.94828320e+00
-7.41840363e-01 -4.99863625e-01 3.78204346e-01 2.76649535e-01
-8.53968635e-02 -4.95360583e-01 -7.40039051e-01 -2.70434409e-01
7.26867020e-01 -1.39287925e+00 -2.77482301e-01 3.11212331e-01
-4.75084931e-01 -3.50165218e-01 6.01293862e-01 -4.85513210e-01
6.04801849e-02 -1.81223667e+00 3.87104183e-01 1.70444772e-01
1.92788228e-01 1.16808005e-01 -4.71455723e-01 2.46533632e-01
1.93240806e-01 -5.05518019e-01 -4.24494930e-02 -6.50179446e-01
1.33004934e-02 3.05898488e-01 -2.41010025e-01 4.73482758e-01
3.43684673e-01 1.01336837e+00 -1.16677690e+00 -3.22067231e-01
3.82495195e-01 2.30780080e-01 -6.76935196e-01 5.76383054e-01
-5.93288124e-01 6.26234174e-01 -4.49019849e-01 7.62665272e-01
8.68675292e-01 -2.67221302e-01 2.37957478e-01 -4.76961136e-01
9.88847837e-02 1.30687684e-01 -1.15560853e+00 2.09339857e+00
-4.63750154e-01 9.69452262e-02 -5.80338482e-03 -8.83624017e-01
1.14546478e+00 -1.60575166e-01 6.60186350e-01 -3.06249768e-01
3.65930140e-01 3.58733445e-01 -6.49041533e-02 -5.18297255e-01
4.55939502e-01 2.98022240e-01 6.60219118e-02 2.08334044e-01
3.44962865e-01 -6.57906711e-01 1.14821196e-01 -9.92657058e-03
9.59619105e-01 9.27147210e-01 -1.98595926e-01 -8.96006674e-02
2.33609051e-01 3.76255751e-01 2.68610716e-01 5.88124275e-01
-1.03551425e-01 7.16255188e-01 2.59486556e-01 -3.40513378e-01
-1.30622101e+00 -1.08766079e+00 1.94115303e-02 1.06805086e+00
6.67569458e-01 -5.91568425e-02 -5.37935257e-01 -9.24032211e-01
5.13786852e-01 6.50341451e-01 -6.25364006e-01 -2.38060027e-01
-6.53898418e-01 -4.06210482e-01 6.07295372e-02 7.35894203e-01
1.45877019e-01 -1.08927274e+00 -7.41616905e-01 3.34050596e-01
-7.24741742e-02 -1.11095655e+00 6.64030537e-02 4.39161301e-01
-7.18892097e-01 -1.12163055e+00 -7.31543183e-01 -8.67383361e-01
9.49101925e-01 3.90109956e-01 7.17680752e-01 7.49528855e-02
-3.15154970e-01 9.91396233e-02 -6.34747922e-01 -5.19973695e-01
-6.56869411e-01 1.88966841e-01 1.60616010e-01 -6.10233009e-01
-2.59778410e-01 -3.75399619e-01 -6.95298731e-01 6.08875871e-01
-5.83064735e-01 1.80000469e-01 7.02633739e-01 8.76772821e-01
6.03305161e-01 -6.41754031e-01 6.39231145e-01 -3.85237753e-01
2.95800626e-01 -2.03335315e-01 -7.92510152e-01 2.13295177e-01
-4.57398266e-01 2.34768316e-01 5.69974959e-01 -8.31393182e-01
-8.66323054e-01 5.80597281e-01 -2.04973698e-01 -7.95937657e-01
4.20843028e-02 2.34181195e-01 -8.61667097e-04 -3.86122227e-01
7.42050111e-01 -1.30596772e-01 3.01488101e-01 -4.43061084e-01
4.73910600e-01 6.61290765e-01 7.69341052e-01 -9.43001807e-01
8.21292102e-01 3.05323929e-01 -1.27553299e-01 -5.18390834e-02
-8.56656253e-01 -4.10974711e-01 -7.50690162e-01 -2.55998224e-01
4.34184253e-01 -9.10148442e-01 -1.16943252e+00 4.94334370e-01
-1.12442708e+00 -7.94298589e-01 -2.11373538e-01 5.40617347e-01
-8.56312096e-01 3.30538094e-01 -5.03938854e-01 -7.29423106e-01
-3.34004611e-01 -1.52467060e+00 1.52758873e+00 1.53314516e-01
1.40269652e-01 -2.14391828e-01 -2.08493486e-01 3.51804465e-01
7.78022334e-02 2.90591955e-01 3.98148805e-01 -5.49593925e-01
-5.74751794e-01 -3.03731382e-01 -3.89675319e-01 1.58630341e-01
5.21362610e-02 -2.58557886e-01 -6.95669711e-01 -6.23124778e-01
-4.33211178e-01 -9.63602006e-01 5.98138392e-01 -7.63126463e-02
1.39404738e+00 2.40465149e-01 -5.36418974e-01 4.95943397e-01
1.10662806e+00 6.48806840e-02 3.56904835e-01 3.62431377e-01
7.06810892e-01 4.26765680e-01 1.20854926e+00 5.79659343e-01
4.75787431e-01 6.44464195e-01 1.09742033e+00 2.95610428e-01
-5.87584451e-02 -4.89993513e-01 4.26558778e-02 3.59693229e-01
-1.91798344e-01 -1.02690630e-01 -8.88899982e-01 4.22320217e-01
-1.83700776e+00 -3.64597857e-01 2.34705329e-01 2.12840247e+00
9.00933146e-01 2.65483201e-01 5.77748492e-02 -1.02981329e-01
5.42589545e-01 -3.81233275e-01 -9.68963146e-01 4.05591309e-01
4.78222460e-01 1.19735755e-01 6.40746057e-01 2.60971099e-01
-1.01146472e+00 1.37838912e+00 5.31236601e+00 6.73518598e-01
-1.27805614e+00 -2.19275817e-01 2.12381139e-01 8.22042301e-02
-8.32925215e-02 -1.49794951e-01 -6.58077776e-01 1.15547404e-01
3.77220303e-01 1.86734349e-01 7.81671107e-01 1.29981232e+00
-3.82677652e-02 -1.70574978e-01 -1.31830490e+00 7.58925974e-01
1.08413972e-01 -9.74301279e-01 -2.12744940e-02 -1.64183766e-01
6.94227695e-01 2.13803977e-01 -3.06730896e-01 4.82581407e-01
6.55176282e-01 -8.16027939e-01 1.01740479e+00 2.59529680e-01
7.64963984e-01 -5.36527455e-01 6.87291980e-01 6.95894897e-01
-8.74736249e-01 -1.59319326e-01 -5.91496348e-01 1.79792836e-01
1.23226829e-01 1.37753233e-01 -1.50410628e+00 7.38771439e-01
7.01258659e-01 4.80933756e-01 -3.49045813e-01 1.24053395e+00
-3.34282935e-01 -4.59723547e-02 -5.35195649e-01 -2.75514901e-01
-1.15235280e-02 1.61862701e-01 4.64350551e-01 5.89532435e-01
2.54603446e-01 4.60402109e-02 5.34146726e-01 8.52446735e-01
-6.83391169e-02 -1.21769801e-01 -9.94977802e-02 1.73457414e-01
6.66651070e-01 1.41259980e+00 -8.19624245e-01 -3.23398799e-01
3.90048921e-01 1.09001648e+00 6.41843677e-01 -6.39477791e-03
-8.12391937e-01 -1.38964042e-01 3.91128451e-01 -3.31136078e-01
3.55835497e-01 -3.23441744e-01 -1.06280588e-01 -9.22163010e-01
1.63510174e-01 -8.07304084e-01 -1.54807881e-01 -1.01638031e+00
-1.06993949e+00 6.72108114e-01 8.31386372e-02 -1.36350465e+00
-5.06430566e-01 -8.81769836e-01 -1.22509785e-01 6.84086800e-01
-1.42791307e+00 -1.30911613e+00 -9.06992614e-01 4.31556031e-02
6.21263206e-01 1.49470223e-02 7.93966949e-01 -1.11707143e-01
-1.54066354e-01 5.70975363e-01 -6.95212781e-02 -1.46480843e-01
8.10175598e-01 -1.26364815e+00 3.79090935e-01 1.61526337e-01
-1.64225921e-01 3.55043501e-01 8.13754618e-01 -7.13096023e-01
-1.79914546e+00 -1.07779634e+00 -4.92768288e-02 -5.90959072e-01
4.95118618e-01 -4.78892118e-01 -5.70417821e-01 6.15008354e-01
-4.82157081e-01 1.69899017e-01 -1.10538758e-01 -3.13229829e-01
-1.28461152e-01 -7.14678094e-02 -1.43178201e+00 5.24203658e-01
1.38612854e+00 1.42195478e-01 -5.72440982e-01 5.73283851e-01
1.01314545e+00 -1.21390891e+00 -9.60240781e-01 9.73603129e-01
9.06657100e-01 -4.70170766e-01 1.12115788e+00 -4.88554746e-01
6.12875938e-01 -2.64305085e-01 -1.13115251e-01 -1.40527022e+00
-6.46926984e-02 -3.87593955e-01 3.11857555e-02 7.76766837e-01
2.89934784e-01 -4.73002046e-01 1.19834411e+00 3.10388148e-01
-3.47105294e-01 -1.14389002e+00 -5.99963844e-01 -6.73908174e-01
-4.55530323e-02 -4.90483731e-01 7.14148521e-01 3.85446280e-01
-1.26414835e-01 -3.39161381e-02 -5.10849394e-02 3.01669121e-01
7.29030192e-01 4.02767807e-01 1.36317623e+00 -1.15924323e+00
-1.69830099e-01 -1.25041291e-01 -1.81149527e-01 -1.56252170e+00
2.35329747e-01 -7.93607533e-01 7.34138608e-01 -1.63124251e+00
2.17016548e-01 -9.19862211e-01 3.46476175e-02 7.00588763e-01
-1.88634023e-01 4.12083715e-01 3.37755293e-01 2.78777868e-01
-5.96527994e-01 8.02504659e-01 1.71435106e+00 -8.89900476e-02
-3.17573398e-01 1.74978703e-01 -2.68662274e-01 5.52588284e-01
7.33563542e-01 -2.35801294e-01 -2.45123729e-01 -4.14628923e-01
-3.03329706e-01 4.49299328e-02 5.19074380e-01 -1.01141512e+00
5.36920689e-02 -2.35662311e-01 5.78104138e-01 -8.41944337e-01
6.19530261e-01 -9.81788635e-01 -2.37892777e-01 6.16684735e-01
-2.34697074e-01 -2.81112403e-01 1.64394215e-01 5.84508061e-01
4.60907876e-01 -3.98308426e-01 6.04497254e-01 -1.10732198e-01
-6.00458980e-01 3.79905194e-01 2.68155962e-01 -3.07078987e-01
1.14129233e+00 -4.40672301e-02 -2.04541713e-01 -2.73680538e-01
-6.93723679e-01 6.27757132e-01 6.13431990e-01 6.15733206e-01
5.74840844e-01 -1.08361864e+00 -5.71219385e-01 1.22335218e-01
3.48669261e-01 6.85345709e-01 -3.92261818e-02 5.00536382e-01
-8.06801677e-01 2.24173758e-02 -2.29060620e-01 -1.06709492e+00
-9.26224828e-01 5.15599847e-01 4.82814051e-02 6.15213774e-02
-6.43772244e-01 9.36074257e-01 -2.51567364e-01 -1.07979751e+00
6.54983997e-01 -4.93300855e-01 5.50074056e-02 -4.43260819e-01
2.19208859e-02 2.74249285e-01 1.62424579e-01 -2.69666046e-01
-2.87859529e-01 3.59592468e-01 -9.88309830e-03 -7.53545761e-02
1.69106650e+00 2.30628729e-01 8.28677490e-02 1.92847520e-01
1.11115932e+00 -1.68128163e-01 -1.75639749e+00 -5.45857288e-02
-1.51808634e-01 -5.19840002e-01 -4.89689738e-01 -1.05221725e+00
-9.23026621e-01 4.59097087e-01 4.51772749e-01 -2.22578153e-01
6.66146815e-01 3.80006373e-01 7.12324679e-01 9.05276597e-01
8.98037076e-01 -9.22486603e-01 4.61281210e-01 5.98951876e-01
1.31630814e+00 -1.38097298e+00 -5.14711924e-02 -6.61884189e-01
-4.53541517e-01 1.17518592e+00 9.90387380e-01 -4.30718869e-01
1.75113663e-01 2.92715073e-01 6.45883232e-02 2.35506031e-03
-4.06045645e-01 2.15612669e-02 1.97267532e-01 6.33407652e-01
-1.17269054e-01 1.05648838e-01 -4.56205308e-02 3.72241706e-01
-5.68078756e-01 1.38818592e-01 2.26884872e-01 1.05619979e+00
-5.33613145e-01 -1.12247252e+00 -3.14456016e-01 4.02314514e-01
3.25533718e-01 5.00326633e-01 -1.20358624e-01 6.75442815e-01
8.92870966e-03 6.51279986e-01 -8.53964686e-02 -7.01336682e-01
5.61435997e-01 -1.12655520e-01 8.08398843e-01 -6.11279607e-01
-2.39592656e-01 -6.90283775e-02 3.87271829e-02 -7.81620681e-01
-1.16852485e-01 -5.97104251e-01 -1.55172455e+00 4.63171601e-02
-5.86819351e-01 -1.05901547e-01 1.13467133e+00 8.48946750e-01
3.48226398e-01 5.91332316e-01 6.74814641e-01 -1.81667221e+00
-1.30253720e+00 -1.17508161e+00 -4.03105728e-02 5.17238915e-01
4.24269959e-02 -1.21114314e+00 -2.04292849e-01 -2.39275724e-01] | [5.826952934265137, -0.8953045606613159] |
c56aa890-74fd-4740-9c5c-9a669b45b4b6 | carfi-rider-localization-using-wi-fi-csi | 2301.01592 | null | https://arxiv.org/abs/2301.01592v1 | https://arxiv.org/pdf/2301.01592v1.pdf | CarFi: Rider Localization Using Wi-Fi CSI | With the rise of hailing services, people are increasingly relying on shared mobility (e.g., Uber, Lyft) drivers to pick up for transportation. However, such drivers and riders have difficulties finding each other in urban areas as GPS signals get blocked by skyscrapers, in crowded environments (e.g., in stadiums, airports, and bars), at night, and in bad weather. It wastes their time, creates a bad user experience, and causes more CO2 emissions due to idle driving. In this work, we explore the potential of Wi-Fi to help drivers to determine the street side of the riders. Our proposed system is called CarFi that uses Wi-Fi CSI from two antennas placed inside a moving vehicle and a data-driven technique to determine the street side of the rider. By collecting real-world data in realistic and challenging settings by blocking the signal with other people and other parked cars, we see that CarFi is 95.44% accurate in rider-side determination in both line of sight (LoS) and non-line of sight (nLoS) conditions, and can be run on an embedded GPU in real-time. | ['Shahriar Nirjon', 'Shan Lin', 'Mahathir Monjur', 'Shiwei Fang', 'Hongkai Chen', 'Sirajum Munir'] | 2022-12-21 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [-5.10172904e-01 -3.50745201e-01 -2.68551767e-01 5.82925044e-02
-4.35426205e-01 -6.24203622e-01 2.48003095e-01 -5.82604110e-01
-3.98234636e-01 8.35439503e-01 -2.02136219e-01 -7.69248426e-01
-1.99251100e-02 -1.13068831e+00 -4.33920920e-01 -7.43586004e-01
-6.70718178e-02 2.33345315e-01 6.22434795e-01 -3.70674044e-01
-5.49112968e-02 5.42930186e-01 -1.70198214e+00 -2.56889015e-01
8.77063572e-01 7.49867976e-01 1.79715693e-01 5.84813118e-01
-1.36694655e-01 5.60770273e-01 -4.61999983e-01 -2.30910212e-01
1.52615651e-01 -9.31640714e-03 1.23419262e-01 -4.97947037e-01
2.67934531e-01 -3.43609184e-01 -7.35839248e-01 4.49566483e-01
6.14236116e-01 3.18259507e-01 1.64325461e-01 -1.87455022e+00
2.18748584e-01 -1.85075536e-01 -7.36494064e-01 2.74791986e-01
4.90305990e-01 2.87146688e-01 1.18643135e-01 -4.51392829e-01
1.84892893e-01 1.04130566e+00 6.90540016e-01 1.25981331e-01
-5.80465198e-01 -1.12377310e+00 -9.24619362e-02 4.40553755e-01
-1.79064393e+00 -7.39974737e-01 3.27177823e-01 -1.33651450e-01
5.50522149e-01 6.41625583e-01 4.32279676e-01 8.15976024e-01
2.11430103e-01 6.08062685e-01 6.91631377e-01 -3.52587067e-02
8.82059485e-02 2.68489480e-01 7.29236826e-02 2.69037396e-01
5.78519285e-01 2.83244491e-01 -5.65341711e-01 -2.78715432e-01
1.94894850e-01 1.54690146e-01 -3.43607277e-01 8.07459354e-02
-1.07630002e+00 4.22412485e-01 4.76471752e-01 2.06422403e-01
-4.28072691e-01 2.51089066e-01 -1.03114896e-01 2.46562418e-02
4.31885608e-02 -2.10089296e-01 -5.89262545e-02 -6.60809696e-01
-8.11959088e-01 2.97194093e-01 7.56534219e-01 1.24912488e+00
6.05293512e-01 -2.17395186e-01 6.49579912e-02 5.31137943e-01
3.64852220e-01 1.46886444e+00 -1.03091091e-01 -6.46905065e-01
8.30321252e-01 -1.08931400e-01 6.27971530e-01 -1.23426020e+00
-7.36944437e-01 -7.27722645e-01 -6.28765047e-01 -8.07069242e-02
5.41728556e-01 -9.04383779e-01 -5.40782034e-01 1.31031251e+00
4.21446085e-01 7.84596264e-01 -1.73281386e-01 1.08424902e+00
6.00598097e-01 5.83431363e-01 -2.44614720e-01 3.42927165e-02
1.16227186e+00 -8.22051287e-01 -9.29655373e-01 -3.97389501e-01
7.79907823e-01 -7.22427666e-01 4.53699082e-01 5.01097441e-01
-7.59608030e-01 -4.83523577e-01 -8.40925097e-01 3.61823797e-01
-3.80669028e-01 -3.61331590e-02 2.39136860e-01 1.15604055e+00
-8.60993266e-01 -1.55146092e-01 -7.36838579e-01 -2.16520801e-01
2.75858760e-01 3.12963486e-01 -1.07080199e-01 -5.15778065e-01
-1.24039900e+00 6.56221390e-01 -5.47144771e-01 3.15916121e-01
-2.07598209e-01 -6.46892667e-01 -5.07616401e-01 -3.79602872e-02
3.49878997e-01 -5.06822109e-01 8.20169210e-01 -4.74007964e-01
-1.07871795e+00 2.44556274e-03 -7.83835530e-01 -3.76467764e-01
4.98746574e-01 -9.52014606e-03 -1.37628317e+00 -4.27750021e-01
3.38694990e-01 2.51028299e-01 3.18282664e-01 -1.22977352e+00
-1.17879057e+00 -4.26722139e-01 -2.84975301e-03 -8.23404938e-02
4.47387278e-01 -2.29847431e-01 -4.80769038e-01 2.62049198e-01
1.21330125e-02 -1.47206581e+00 -8.79613087e-02 -4.55980688e-01
-5.12836695e-01 2.27889605e-02 1.02775228e+00 -3.24323297e-01
1.35367537e+00 -2.34106135e+00 -1.12905061e+00 7.23503709e-01
-4.49374691e-03 4.54728752e-01 2.14295357e-01 4.81058776e-01
5.00783205e-01 -2.14714080e-01 3.13469201e-01 -1.28495023e-01
4.34979238e-02 3.33354980e-01 -1.69689983e-01 6.32048309e-01
-5.07026434e-01 7.10481167e-01 -1.19877517e+00 -9.57702175e-02
4.07268912e-01 6.53766215e-01 -3.04849029e-01 -2.18877494e-01
9.28280890e-01 6.75277531e-01 -5.17786205e-01 4.11482841e-01
1.49466455e+00 4.52727199e-01 -2.36283392e-01 4.31176305e-01
-8.64660323e-01 4.18613374e-01 -1.41678774e+00 7.50346005e-01
-9.30424452e-01 1.15948045e+00 2.48796269e-01 -3.42194259e-01
6.22381866e-01 2.45196879e-01 6.62404746e-02 -1.23764336e+00
-1.02243330e-02 4.36767220e-01 -6.39432902e-03 -5.64793110e-01
5.43051481e-01 2.73891389e-01 -5.71027771e-02 2.79070586e-01
-8.79653871e-01 3.80064249e-01 3.85961868e-02 -1.26555011e-01
1.25658989e+00 -5.42755067e-01 -1.77909210e-01 -4.36795019e-02
5.63306451e-01 -1.40044272e-01 3.58919263e-01 8.03908765e-01
-3.86446327e-01 1.50942951e-01 -2.87266932e-02 -1.20619252e-01
-3.72953385e-01 -1.18906999e+00 -2.84161061e-01 7.93180048e-01
7.23346889e-01 -9.67211574e-02 -3.19793642e-01 -8.86262208e-02
4.04880702e-01 1.12244248e+00 9.17743370e-02 -5.07460013e-02
-4.83036757e-01 -3.50408912e-01 6.19052887e-01 1.65295452e-01
8.36450160e-01 -1.63494825e-01 -2.66678184e-01 4.07096982e-01
-5.77533901e-01 -1.19151759e+00 -3.54077905e-01 -3.61743540e-01
-2.74326000e-02 -6.47783041e-01 -5.53711534e-01 -2.65946537e-01
5.58000565e-01 1.60423493e+00 6.70829237e-01 5.68624102e-02
2.94151694e-01 1.79728985e-01 1.73589975e-01 -6.38246536e-01
2.76289165e-01 -6.55151829e-02 8.04020613e-02 3.67925316e-01
7.22611368e-01 -3.72598737e-01 -9.08565342e-01 1.20017517e+00
5.20255277e-03 -1.57804176e-01 5.11328541e-02 1.03783673e-02
1.20236471e-01 6.15829527e-01 6.98106945e-01 -3.30669910e-01
2.25627273e-01 -1.20213997e+00 -7.57445455e-01 -3.59250665e-01
-9.21066478e-02 -5.31003475e-01 4.87188429e-01 1.88540146e-01
-1.03945363e+00 -1.80109233e-01 -4.50666368e-01 1.52481556e-01
-4.21897501e-01 -5.64302653e-02 -2.84993500e-01 -2.24986285e-01
4.31999117e-01 -2.01131210e-01 -3.57274771e-01 -8.34200308e-02
2.57433295e-01 1.30204189e+00 3.95073086e-01 3.96197319e-01
9.47193384e-01 9.03054237e-01 2.13152722e-01 -1.24687397e+00
-1.28338918e-01 -1.12034822e+00 -1.66931495e-01 -5.01842380e-01
4.00756210e-01 -1.12940145e+00 -1.14942443e+00 4.12640095e-01
-9.28003013e-01 -1.03412539e-01 5.21143079e-01 7.79072940e-01
6.39426187e-02 -1.31105520e-02 2.37762481e-01 -1.25590909e+00
3.42235804e-01 -1.08654428e+00 8.02223802e-01 4.68869239e-01
-1.01105168e-01 -8.32365990e-01 -7.61905834e-02 7.24438965e-01
8.31267476e-01 5.55799045e-02 -1.08897857e-01 5.81464954e-02
-8.77741218e-01 -6.70988441e-01 -2.45596856e-01 -3.15716863e-01
2.35707939e-01 -4.08391207e-01 -1.28081369e+00 2.42920294e-02
-4.53499615e-01 7.14429975e-01 4.66095030e-01 5.68511128e-01
7.83422589e-01 -2.03911662e-01 -1.15445495e+00 4.55673814e-01
1.08870077e+00 3.58480424e-01 1.05185366e+00 1.18002959e-01
5.80673695e-01 2.44307101e-01 8.31792831e-01 3.36124361e-01
7.87825882e-01 9.44469154e-01 5.65263391e-01 -2.89972216e-01
1.04460102e-02 -1.08324178e-01 2.48714611e-01 1.50624469e-01
-1.42405510e-01 -9.46287215e-01 -7.71959543e-01 6.50215745e-01
-1.72972894e+00 -9.86632943e-01 -1.14920390e+00 2.74711299e+00
-1.40149295e-01 1.32979169e-01 3.00642252e-01 5.31030223e-02
8.62712860e-01 -3.25315416e-01 -1.30437553e-01 -3.52056414e-01
9.08978283e-02 -2.30346888e-01 1.62174177e+00 8.29440057e-01
-7.07164705e-01 4.45142746e-01 5.09680510e+00 9.05614436e-01
-1.09310853e+00 2.57646441e-01 2.73533046e-01 -3.13010842e-01
-4.11977261e-01 -2.32358590e-01 -9.36051846e-01 9.71614599e-01
1.21971703e+00 1.11362234e-01 5.74160814e-01 6.05263531e-01
9.16906416e-01 -8.72929573e-01 -4.50342327e-01 1.23284125e+00
-1.89921156e-01 -1.08206975e+00 -1.09118986e+00 5.36612988e-01
5.55937767e-01 6.02872789e-01 -1.67470379e-03 1.84391096e-01
-1.61267184e-02 -7.64281273e-01 4.92276490e-01 4.02944356e-01
3.59379828e-01 -1.10313308e+00 8.94513667e-01 5.80184996e-01
-1.36594808e+00 -1.04679093e-01 4.67492491e-02 -1.95376799e-01
7.34103024e-01 8.99713218e-01 -7.86268473e-01 3.05278420e-01
3.47641319e-01 1.07463123e-02 -6.11995012e-02 1.56176686e+00
-2.35257164e-01 7.83523500e-01 -7.37363756e-01 -1.34849668e-01
4.52383935e-01 -2.49473229e-01 6.32314920e-01 1.12016356e+00
7.69927263e-01 2.05852371e-02 -1.86914146e-01 5.44740319e-01
2.84613669e-01 -4.17067498e-01 -1.04266798e+00 7.63298392e-01
6.24918461e-01 1.21124065e+00 -4.82717305e-01 1.88555848e-02
-5.37250578e-01 4.75159287e-01 -6.52430296e-01 8.33661675e-01
-1.19657218e+00 -7.41519809e-01 1.10258389e+00 6.60549104e-01
2.12792218e-01 -7.75221825e-01 -1.88672826e-01 -5.87275624e-01
9.10173208e-02 -5.31328917e-02 -3.13958287e-01 -6.38811707e-01
-5.74713528e-01 8.53982046e-02 -5.16193569e-01 -1.61353743e+00
5.98441437e-02 -2.09440604e-01 -6.86234117e-01 1.03537512e+00
-1.91228020e+00 -4.19656128e-01 -5.28675437e-01 6.86166465e-01
-3.72215547e-02 1.89369634e-01 2.03552827e-01 1.26923323e+00
-4.66494650e-01 6.70875013e-01 6.72565401e-01 -1.12667084e-01
6.08614624e-01 -5.96885622e-01 6.83824360e-01 6.64764166e-01
-1.66925907e-01 4.60736483e-01 6.20056152e-01 -4.79955047e-01
-1.72249031e+00 -1.10524702e+00 1.27911675e+00 -3.85488272e-01
3.25644940e-01 -6.54430389e-01 -2.28340209e-01 2.87112355e-01
-1.30707219e-01 1.31017908e-01 7.02482045e-01 1.12109058e-01
4.17869657e-01 -2.75939822e-01 -1.22926462e+00 8.80805016e-01
1.17677224e+00 -2.02498406e-01 1.96518555e-01 4.23165709e-01
6.57306111e-04 -4.20422196e-01 8.04050565e-02 -2.92394102e-01
7.79074669e-01 -1.03090703e+00 1.01108468e+00 1.80881947e-01
-8.00950646e-01 -4.90970939e-01 1.77869871e-01 -1.38242745e+00
-2.66297609e-01 -8.58769894e-01 2.86101252e-01 9.43872631e-01
4.73681659e-01 -1.37776029e+00 4.26578462e-01 8.02286565e-01
-1.13506593e-01 -2.21373379e-01 -1.29414928e+00 -1.02291453e+00
-6.24368429e-01 -1.11968529e+00 1.07022405e+00 4.21811938e-01
9.63528361e-03 2.24061891e-01 -4.57654953e-01 6.93078339e-01
4.15496707e-01 -3.77527416e-01 1.26403570e+00 -8.51922214e-01
3.31831157e-01 -1.96729764e-01 -5.59720039e-01 -1.56154191e+00
-1.63046002e-01 -5.83962798e-01 1.31997079e-01 -1.59795654e+00
-7.10366547e-01 -1.21023273e+00 4.33498174e-02 2.71804724e-02
2.31069237e-01 2.79187530e-01 -8.98285210e-02 -7.44211748e-02
-5.25592089e-01 1.03105322e-01 1.06840205e+00 -5.26351333e-02
-3.73093694e-01 9.54353809e-01 -4.13454652e-01 6.50666416e-01
7.99114525e-01 -3.42034578e-01 -4.51020658e-01 -3.72950822e-01
6.92879081e-01 1.03046156e-01 5.57147145e-01 -1.49751496e+00
4.50127304e-01 -2.29009613e-01 -4.11430985e-04 -1.04563940e+00
5.30114412e-01 -1.26546597e+00 3.91073585e-01 3.92652541e-01
6.18025422e-01 -4.21989411e-01 3.33208650e-01 5.39893031e-01
2.00917035e-01 1.55223042e-01 3.50815713e-01 4.18085396e-01
-5.73143780e-01 2.84054607e-01 -8.81821871e-01 -2.10470393e-01
1.08173156e+00 -4.73707646e-01 -6.51889145e-01 -8.54126334e-01
-1.20268330e-01 8.11694086e-01 1.08622879e-01 4.39628512e-01
2.10301250e-01 -1.52031434e+00 -3.30060452e-01 4.77395892e-01
1.14367396e-01 -5.21926284e-01 4.77825195e-01 1.30541897e+00
-3.51502210e-01 1.13116574e+00 1.12030014e-01 -7.62425423e-01
-1.02308583e+00 2.01159745e-01 3.35134119e-01 4.42708015e-01
-3.69863659e-01 6.39975429e-01 2.53249019e-01 -1.31253481e-01
2.87822559e-02 -4.74329382e-01 3.42317149e-02 -8.86933953e-02
8.14368188e-01 1.05440998e+00 2.49763027e-01 -1.12189233e+00
-6.84001207e-01 5.10371745e-01 6.59496665e-01 -6.75925538e-02
5.46510220e-01 -9.86741126e-01 4.60717499e-01 2.15902820e-01
1.02116179e+00 7.13851154e-01 -7.71146834e-01 -6.05465621e-02
-4.32454675e-01 -9.74286675e-01 4.28592980e-01 -5.25167942e-01
-1.03411865e+00 5.92238784e-01 8.49180162e-01 4.47843701e-01
6.16249681e-01 -1.91367209e-01 1.38356674e+00 3.22611362e-01
1.01217759e+00 -1.02121460e+00 -1.01341259e+00 4.28630918e-01
2.90891994e-02 -1.22541273e+00 -4.98534501e-01 -5.42952955e-01
-4.67951834e-01 7.82364130e-01 1.40237644e-01 2.79743254e-01
8.94689679e-01 2.67063797e-01 1.57911032e-01 -1.13398030e-01
-2.53834665e-01 -7.30828106e-01 -2.11010888e-01 1.04313290e+00
1.78352632e-02 5.93675911e-01 6.26278147e-02 4.19794291e-01
-4.04268682e-01 8.37784074e-03 5.95979035e-01 8.48957181e-01
-6.20446563e-01 -7.17085838e-01 -7.79595912e-01 4.84409928e-01
-5.53340428e-02 2.27770269e-01 2.79151022e-01 6.66590214e-01
7.34818518e-01 1.78154469e+00 3.54830533e-01 -2.44543225e-01
6.79744899e-01 -2.75152922e-01 -3.68606448e-02 1.03033543e-01
-4.77187857e-02 -2.20464602e-01 6.26622677e-01 -6.34759903e-01
-8.36129710e-02 -6.48224354e-01 -1.36613369e+00 -8.98624718e-01
-5.02388000e-01 1.84321791e-01 9.78684783e-01 1.18780315e+00
7.33010530e-01 5.00622988e-01 1.11545503e+00 -6.10028625e-01
5.44646680e-01 -3.32923055e-01 -7.91829824e-01 -3.15598220e-01
7.02144027e-01 -9.73439515e-01 -4.24785823e-01 -7.27088511e-01] | [6.243051052093506, 1.031459093093872] |
5b60858e-b74b-47d1-8354-feabb721f7df | a-tale-of-two-latent-flows-learning-latent | 2301.09300 | null | https://arxiv.org/abs/2301.09300v1 | https://arxiv.org/pdf/2301.09300v1.pdf | A Tale of Two Latent Flows: Learning Latent Space Normalizing Flow with Short-run Langevin Flow for Approximate Inference | We study a normalizing flow in the latent space of a top-down generator model, in which the normalizing flow model plays the role of the informative prior model of the generator. We propose to jointly learn the latent space normalizing flow prior model and the top-down generator model by a Markov chain Monte Carlo (MCMC)-based maximum likelihood algorithm, where a short-run Langevin sampling from the intractable posterior distribution is performed to infer the latent variables for each observed example, so that the parameters of the normalizing flow prior and the generator can be updated with the inferred latent variables. We show that, under the scenario of non-convergent short-run MCMC, the finite step Langevin dynamics is a flow-like approximate inference model and the learning objective actually follows the perturbation of the maximum likelihood estimation (MLE). We further point out that the learning framework seeks to (i) match the latent space normalizing flow and the aggregated posterior produced by the short-run Langevin flow, and (ii) bias the model from MLE such that the short-run Langevin flow inference is close to the true posterior. Empirical results of extensive experiments validate the effectiveness of the proposed latent space normalizing flow model in the tasks of image generation, image reconstruction, anomaly detection, supervised image inpainting and unsupervised image recovery. | ['Ping Li', 'Dingcheng Li', 'Yifei Xu', 'Yaxuan Zhu', 'Jianwen Xie'] | 2023-01-23 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 5.47129452e-01 3.09571803e-01 -1.29593164e-01 -4.50733453e-02
-7.87185431e-01 -2.69713163e-01 9.16185260e-01 -4.86527920e-01
-2.61029869e-01 6.95919514e-01 3.33890885e-01 -1.58154830e-01
-2.64100969e-01 -7.37003207e-01 -8.40425968e-01 -1.04357457e+00
1.36203170e-01 7.95768261e-01 -3.20499510e-01 2.86048353e-01
3.50670516e-01 4.73557651e-01 -1.07998192e+00 -2.69805826e-02
8.96844923e-01 4.31611478e-01 1.70161113e-01 1.07377017e+00
-7.78542226e-03 8.60027671e-01 -3.20107222e-01 -1.79964796e-01
3.34574938e-01 -1.01487124e+00 -6.64880037e-01 5.60910344e-01
1.08970255e-01 -5.34409165e-01 -2.70534605e-01 1.32979012e+00
1.72695875e-01 1.87747180e-01 1.14594579e+00 -1.05884206e+00
-4.04937059e-01 2.87807226e-01 -8.87867928e-01 -1.10973073e-02
1.85453728e-01 4.19099778e-01 7.50143170e-01 -7.79108167e-01
1.02643466e+00 1.47843015e+00 4.58434194e-01 5.14269710e-01
-1.83963907e+00 -4.05444026e-01 -1.52730599e-01 -3.62109005e-01
-1.36109769e+00 -4.18355167e-01 7.81549156e-01 -8.14213872e-01
4.42628950e-01 -1.87460780e-01 6.22543216e-01 1.14852989e+00
5.80830753e-01 6.46827996e-01 9.45970476e-01 -6.34107888e-01
5.13377011e-01 2.22261827e-02 -3.86818379e-01 8.51180255e-01
2.34018974e-02 2.91389704e-01 -6.03621542e-01 -5.60790241e-01
1.10052311e+00 -1.53523654e-01 -4.69626263e-02 -4.23881292e-01
-1.07416320e+00 1.08158910e+00 -1.19014814e-01 -1.75718889e-01
-6.10282242e-01 5.44037580e-01 4.72959429e-02 8.92020985e-02
6.55252695e-01 3.26663941e-01 6.45660609e-02 4.18016911e-02
-1.46887457e+00 4.89065081e-01 8.32791686e-01 9.31350589e-01
9.50729430e-01 1.97523102e-01 -3.55540127e-01 3.02599669e-01
6.74005985e-01 8.17200601e-01 2.68840373e-01 -1.48405421e+00
2.90429741e-01 -9.93036702e-02 3.01185995e-01 -8.26863110e-01
4.08863842e-01 -5.05629718e-01 -9.41680312e-01 3.16613227e-01
5.04822552e-01 -1.30821884e-01 -8.85137081e-01 2.04480910e+00
2.16920659e-01 5.13293266e-01 -7.24113137e-02 4.48977411e-01
-2.11948171e-01 1.12805891e+00 -1.42837912e-01 -8.45578492e-01
1.06888318e+00 -6.47413015e-01 -8.75098348e-01 -1.50818467e-01
4.11999345e-01 -8.74045849e-01 8.71274114e-01 2.08862901e-01
-1.43947077e+00 -5.09453237e-01 -9.89863336e-01 2.02862933e-01
4.99634355e-01 6.97692558e-02 9.00245011e-02 3.17035019e-01
-7.52036512e-01 7.86907434e-01 -1.25227225e+00 -1.47301719e-01
4.36270356e-01 -1.61944464e-01 -3.12560387e-02 1.07084550e-01
-6.75617993e-01 5.85191905e-01 3.22572559e-01 9.63196903e-02
-1.49499261e+00 -5.73647857e-01 -5.22009850e-01 -1.13216497e-01
2.24856734e-01 -1.11295378e+00 1.04386866e+00 -1.05476582e+00
-1.64871204e+00 7.90169299e-01 -6.20804846e-01 -4.15092468e-01
1.10036349e+00 -1.40774295e-01 2.52947748e-01 3.25095326e-01
2.20818117e-01 8.39511991e-01 1.62296844e+00 -1.29617131e+00
-2.67688990e-01 -1.06236450e-02 -5.64419150e-01 2.46294037e-01
2.74419188e-01 -4.16160524e-01 -2.59809703e-01 -6.62742853e-01
2.17438102e-01 -9.85700250e-01 -3.54844719e-01 5.83912656e-02
-3.78800482e-01 1.64353192e-01 6.06921434e-01 -7.56954730e-01
1.17737639e+00 -2.09007549e+00 4.54808414e-01 2.94497222e-01
-4.12343675e-03 -3.08015436e-01 2.84254309e-02 4.83252555e-01
-7.16622397e-02 -5.75395674e-02 -7.89408684e-01 -6.52851760e-01
-2.87118942e-01 4.58231866e-01 -1.02042890e+00 8.62803638e-01
1.70294881e-01 8.60013604e-01 -1.01731074e+00 -5.84556460e-01
1.02572560e-01 3.69941503e-01 -8.98620486e-01 4.24833477e-01
-4.57582742e-01 8.94118369e-01 -3.22639138e-01 3.04406206e-03
6.95996583e-01 -3.92459691e-01 1.30347550e-01 -1.47831485e-01
-3.12597714e-02 -2.46966690e-01 -1.38979006e+00 1.84979701e+00
-2.06117392e-01 5.00079274e-01 -1.70702711e-01 -6.60693467e-01
6.97375476e-01 2.83225834e-01 3.37774098e-01 -9.31459740e-02
-1.96400806e-01 1.36780009e-01 -3.10361296e-01 -5.28201282e-01
3.40788901e-01 -6.50677919e-01 1.59078211e-01 8.79337609e-01
2.12432802e-01 -3.19014460e-01 1.80063665e-01 5.80680072e-01
6.86573148e-01 5.42330384e-01 4.51695621e-01 -1.56237215e-01
4.95417207e-01 -2.77714998e-01 5.00375628e-01 1.18065786e+00
5.49018644e-02 7.38306940e-01 7.61077881e-01 -1.39894277e-01
-1.68551326e+00 -1.35287738e+00 -9.71982479e-02 4.66817617e-01
-2.57985055e-01 -2.61785805e-01 -1.02005839e+00 -3.95710558e-01
-3.86668891e-01 8.21969151e-01 -4.96315926e-01 -1.33187026e-01
-6.65497601e-01 -8.90162289e-01 3.63802344e-01 1.24767171e-02
5.07206678e-01 -1.04753268e+00 -4.58146721e-01 3.87506247e-01
-4.35767502e-01 -8.71233284e-01 -6.38358891e-01 -2.74435043e-01
-1.34483373e+00 -6.16834044e-01 -7.20910966e-01 -3.78523022e-01
1.05928552e+00 -3.62514824e-01 9.44237173e-01 -2.09220290e-01
-5.50964847e-02 2.51674056e-01 2.56015033e-01 2.29888320e-01
-9.64520097e-01 -1.56572476e-01 -7.61907473e-02 4.27280039e-01
-3.05428416e-01 -8.56075168e-01 -4.67351019e-01 1.37390923e-02
-1.10888743e+00 4.62059408e-01 4.95540470e-01 1.03645384e+00
8.71280491e-01 -2.41752695e-02 2.29360938e-01 -9.32624698e-01
4.47650850e-01 -3.40729207e-01 -1.03506744e+00 1.15305118e-01
-7.98771977e-01 6.84797823e-01 3.95791799e-01 -5.40107965e-01
-1.34093392e+00 1.84671536e-01 -4.56402786e-02 -6.88486457e-01
-3.20668519e-02 3.09315085e-01 -3.65755931e-02 4.29187566e-01
5.73930442e-01 4.40697402e-01 1.44443214e-01 -4.11809474e-01
6.92120254e-01 1.65907905e-01 9.17786002e-01 -8.84258270e-01
1.01267707e+00 1.02265143e+00 4.41873938e-01 -7.48762727e-01
-9.45464849e-01 -1.01463087e-01 -6.38488531e-01 -1.68723509e-01
1.11422598e+00 -7.75165081e-01 -3.81725341e-01 7.00478256e-01
-1.37724292e+00 -2.13768139e-01 -8.60513091e-01 5.19557893e-01
-1.16763484e+00 4.63499665e-01 -6.58648908e-01 -1.20323443e+00
-4.51967586e-03 -1.22776461e+00 1.13751209e+00 1.04254730e-01
-2.71886051e-01 -1.20895088e+00 4.65240180e-01 -1.98485702e-02
-1.07914828e-01 3.77793461e-01 1.04773927e+00 -4.25334573e-02
-1.04153013e+00 -9.06716138e-02 4.79020774e-02 3.28958303e-01
-2.03416824e-01 2.43453801e-01 -1.04828858e+00 -3.96339118e-01
4.44410294e-01 6.85535669e-02 8.95102680e-01 9.25673544e-01
6.34640932e-01 -6.33809388e-01 -1.70493737e-01 1.00833440e+00
1.65829504e+00 6.10343665e-02 7.81855881e-01 -2.01931149e-01
5.12586236e-01 5.10390520e-01 1.91601232e-01 5.03061831e-01
-2.79317856e-01 3.17354649e-01 1.97188675e-01 3.13296974e-01
-1.13702312e-01 -8.33757102e-01 5.18893719e-01 8.22468102e-01
1.06964737e-01 -1.76337779e-01 -6.57396674e-01 4.87102717e-01
-1.93657994e+00 -1.25959396e+00 4.77887690e-02 2.37268710e+00
8.96289587e-01 2.19525844e-01 -8.71465504e-02 -2.15491310e-01
7.37303674e-01 2.90044665e-01 -5.89606166e-01 -6.89403266e-02
2.97683477e-02 5.18244430e-02 3.73297095e-01 1.19141746e+00
-7.23643363e-01 6.68294728e-01 6.89017534e+00 1.04964006e+00
-5.94192147e-01 2.83903748e-01 6.44817948e-01 3.81424204e-02
-5.91365099e-01 5.83028436e-01 -8.14157903e-01 6.65976942e-01
7.89687335e-01 -2.84290344e-01 6.20110333e-01 4.01420087e-01
5.56760490e-01 -3.46294016e-01 -1.07939303e+00 8.59902799e-01
-8.04103017e-02 -1.53141403e+00 4.42141235e-01 6.05031610e-01
1.10815287e+00 -3.33378732e-01 1.95680130e-02 -2.98030376e-01
4.09340471e-01 -7.80956268e-01 1.02281022e+00 9.81509686e-01
8.17131102e-01 -7.36902595e-01 2.54997015e-01 7.82236755e-01
-8.48380506e-01 9.94474962e-02 -1.97620526e-01 1.11918852e-01
6.82537436e-01 8.85582268e-01 -6.75569832e-01 1.07323587e-01
-5.41743217e-03 7.04123616e-01 1.33341281e-02 5.35579681e-01
-4.49297965e-01 7.68482924e-01 -2.56406575e-01 5.02314210e-01
1.50475457e-01 -9.74581599e-01 1.06542671e+00 8.66854966e-01
2.89723963e-01 -2.77089745e-01 1.09077357e-01 1.56571960e+00
-4.63987067e-02 -3.75228852e-01 -4.60873485e-01 -1.61350027e-01
2.96138912e-01 9.29707706e-01 -1.02500689e+00 -4.13745254e-01
1.93747178e-01 1.09436536e+00 5.07466868e-02 8.87773097e-01
-7.52012432e-01 -2.52016336e-02 4.02089119e-01 -2.22816076e-02
1.91197827e-01 -3.02615732e-01 -3.60944599e-01 -1.28357160e+00
-2.21283048e-01 -5.00560760e-01 2.95553803e-01 -9.38550770e-01
-1.13808322e+00 1.70251325e-01 2.96556234e-01 -1.12890625e+00
-8.59271109e-01 -1.07761286e-01 -7.54950643e-01 1.24302018e+00
-1.19598544e+00 -7.87463069e-01 2.31765777e-01 3.42695773e-01
4.83315378e-01 -1.34672925e-01 5.70758879e-01 -2.66040474e-01
-4.64935362e-01 7.22169802e-02 4.12445247e-01 -1.41499132e-01
3.89615953e-01 -1.03884816e+00 1.77588120e-01 1.23840690e+00
9.18858722e-02 5.78174114e-01 1.19766712e+00 -1.03799605e+00
-9.66945767e-01 -9.30720687e-01 8.07621002e-01 -3.86889964e-01
6.13551915e-01 -2.89708942e-01 -8.07711899e-01 9.15751636e-01
9.06740651e-02 -1.62639737e-01 1.73151106e-01 -5.65324247e-01
-4.40714620e-02 1.56955197e-02 -1.09857237e+00 5.31481385e-01
6.13632321e-01 -6.95463419e-01 -4.18022156e-01 3.88502538e-01
4.80616599e-01 -2.31071055e-01 -3.53282094e-01 -1.37398392e-01
5.75147033e-01 -7.39803016e-01 9.51420605e-01 -4.53469068e-01
7.72182465e-01 -3.94999117e-01 -4.72672842e-02 -1.10095704e+00
-1.93329632e-01 -1.23914838e+00 -5.26321471e-01 1.11061335e+00
3.97655256e-02 -4.09631193e-01 7.94468820e-01 9.34212878e-02
2.50063956e-01 -4.79193985e-01 -8.61377120e-01 -6.40574932e-01
8.02243650e-02 -4.64661092e-01 2.05190107e-01 4.02814031e-01
-7.46353924e-01 2.61361957e-01 -7.02124238e-01 2.67219931e-01
1.24774647e+00 2.32236069e-02 7.50702560e-01 -8.97705138e-01
-9.31539237e-01 -6.81146234e-02 6.14835247e-02 -1.28876948e+00
1.84604779e-01 -8.54299426e-01 2.39541486e-01 -1.30438936e+00
5.28440297e-01 3.62706371e-02 3.57003868e-01 -2.85591662e-01
-1.16617471e-01 -8.45168531e-02 2.22603396e-01 7.49495685e-01
-7.76519626e-02 6.49733484e-01 1.41182530e+00 1.94227263e-01
-2.00677797e-01 8.49220008e-02 -3.32485229e-01 9.46653903e-01
1.70032501e-01 -1.02543557e+00 -6.55403972e-01 -1.54975280e-01
6.99809670e-01 4.54625994e-01 4.79346156e-01 -6.23459697e-01
2.55374312e-01 -3.18662554e-01 2.97357023e-01 -6.37211025e-01
2.40644127e-01 -4.01005298e-01 4.35582131e-01 6.12337947e-01
-6.16759598e-01 -9.91785750e-02 -4.25988406e-01 8.48541498e-01
1.44299105e-01 -7.35945284e-01 1.26067877e+00 -2.34273061e-01
4.53888327e-02 4.73089784e-01 -6.36916876e-01 2.54048347e-01
6.48726046e-01 -9.50569510e-02 2.53558040e-01 -7.35687375e-01
-8.63115013e-01 -1.82255819e-01 4.46981847e-01 -3.88073504e-01
5.46345294e-01 -1.42732179e+00 -7.37086713e-01 6.10300064e-01
-3.05621833e-01 1.85626790e-01 1.40527114e-01 8.18386316e-01
-5.03548265e-01 -1.47535563e-01 1.22541383e-01 -9.76370931e-01
-6.19826913e-01 2.91311383e-01 4.95339423e-01 -7.47781277e-01
-7.81511247e-01 6.94997013e-01 3.82148564e-01 -1.90841570e-01
4.78396080e-02 -1.25518171e-02 3.16396385e-01 -5.99121442e-03
4.10708368e-01 6.21912897e-01 -5.19230664e-01 -4.52158600e-01
2.18336612e-01 6.29395306e-01 6.17572516e-02 -1.00681365e+00
1.14481688e+00 -4.69376981e-01 -1.70708299e-01 6.03153110e-01
1.21477103e+00 -2.62030419e-02 -1.85600269e+00 -8.29374269e-02
-2.74263024e-01 -5.80510974e-01 2.78135724e-02 -1.85222745e-01
-7.70268142e-01 9.83486116e-01 4.43313837e-01 -1.00494862e-01
9.21003163e-01 -3.36442739e-02 5.35281181e-01 8.24375004e-02
-1.27407163e-01 -1.05340433e+00 3.16760987e-01 3.68269980e-01
6.66506052e-01 -6.26705170e-01 2.16886848e-01 2.74545513e-02
-3.61044973e-01 1.01904058e+00 7.66162798e-02 -5.25642157e-01
7.34702647e-01 2.47748464e-01 -4.69215691e-01 -8.34236890e-02
-7.71839380e-01 3.60651165e-01 1.28389612e-01 4.54794705e-01
-5.91318458e-02 -1.84692636e-01 2.43975054e-02 -9.54186916e-02
-8.95057172e-02 2.02946693e-01 6.56770170e-01 7.52737045e-01
-2.38009870e-01 -9.61133063e-01 -4.46095556e-01 1.23920552e-01
-1.29231438e-01 -7.37614110e-02 2.98372775e-01 3.60593826e-01
3.92082334e-02 4.56672728e-01 2.92367250e-01 4.73468482e-01
-1.36696696e-01 3.60501558e-01 6.80825055e-01 -3.01294148e-01
2.53429800e-01 7.37418175e-01 -4.86427605e-01 -4.31427568e-01
-3.72644365e-01 -9.06816840e-01 -9.87886488e-01 -3.29416454e-01
-2.39058390e-01 1.12294681e-01 5.40489256e-01 1.18232965e+00
8.00391883e-02 2.92097807e-01 8.31781983e-01 -8.39817703e-01
-8.68343055e-01 -8.40738058e-01 -6.86417103e-01 5.11859834e-01
4.24583435e-01 -2.47392207e-01 -7.85980463e-01 6.14847839e-01] | [7.039623260498047, 3.7757680416107178] |
c3e9bf11-cc7b-4d38-bb60-fc9b35de6af9 | contrastive-masked-autoencoders-for-self | 2211.11210 | null | https://arxiv.org/abs/2211.11210v2 | https://arxiv.org/pdf/2211.11210v2.pdf | Contrastive Masked Autoencoders for Self-Supervised Video Hashing | Self-Supervised Video Hashing (SSVH) models learn to generate short binary representations for videos without ground-truth supervision, facilitating large-scale video retrieval efficiency and attracting increasing research attention. The success of SSVH lies in the understanding of video content and the ability to capture the semantic relation among unlabeled videos. Typically, state-of-the-art SSVH methods consider these two points in a two-stage training pipeline, where they firstly train an auxiliary network by instance-wise mask-and-predict tasks and secondly train a hashing model to preserve the pseudo-neighborhood structure transferred from the auxiliary network. This consecutive training strategy is inflexible and also unnecessary. In this paper, we propose a simple yet effective one-stage SSVH method called ConMH, which incorporates video semantic information and video similarity relationship understanding in a single stage. To capture video semantic information for better hashing learning, we adopt an encoder-decoder structure to reconstruct the video from its temporal-masked frames. Particularly, we find that a higher masking ratio helps video understanding. Besides, we fully exploit the similarity relationship between videos by maximizing agreement between two augmented views of a video, which contributes to more discriminative and robust hash codes. Extensive experiments on three large-scale video datasets (i.e., FCVID, ActivityNet and YFCC) indicate that ConMH achieves state-of-the-art results. Code is available at https://github.com/huangmozhi9527/ConMH. | ['Shutao Xia', 'Ziyun Zeng', 'Bin Chen', 'Jinpeng Wang', 'Yuting Wang'] | 2022-11-21 | null | null | null | null | ['video-similarity'] | ['computer-vision'] | [-3.93178985e-02 -1.67614207e-01 -5.83103359e-01 -3.96621615e-01
-8.15991879e-01 -3.70923042e-01 4.04213399e-01 -2.17505526e-02
-2.81392336e-01 4.70708400e-01 3.62868518e-01 1.09073393e-01
2.69321859e-01 -6.38306379e-01 -9.22209024e-01 -7.66830444e-01
-1.28406852e-01 3.24256755e-02 3.27638507e-01 5.73581643e-03
1.20115392e-01 -3.54716107e-02 -1.61641085e+00 4.75372314e-01
5.60163319e-01 1.19669819e+00 4.39626753e-01 3.60599488e-01
2.46949986e-01 1.02073169e+00 -8.36162344e-02 -3.63410890e-01
2.29623526e-01 -5.53086698e-01 -7.12103128e-01 7.94948041e-02
2.83326507e-01 -7.47153282e-01 -9.61276770e-01 9.47498620e-01
3.87131363e-01 7.51221702e-02 3.40638250e-01 -1.51593804e+00
-5.58086395e-01 4.78550345e-01 -4.79334056e-01 1.60796314e-01
5.90937674e-01 2.92534083e-01 1.17484236e+00 -9.20250595e-01
6.16224229e-01 9.58848536e-01 5.26084363e-01 4.29290771e-01
-9.18583095e-01 -9.53187525e-01 -1.51685476e-01 7.23346412e-01
-1.83949232e+00 -4.72934842e-01 7.33290613e-01 -2.28936508e-01
6.42092645e-01 6.10054359e-02 8.11360300e-01 9.54117537e-01
-1.27451003e-01 1.15502679e+00 7.01290250e-01 -7.19589442e-02
1.71530142e-01 7.85436407e-02 -1.50925636e-01 9.33305144e-01
3.15544195e-02 3.37779708e-02 -7.20859528e-01 4.90039438e-02
7.71373868e-01 5.14707506e-01 -6.05207741e-01 -4.50867087e-01
-1.21741176e+00 9.52829003e-01 6.61960006e-01 2.72749454e-01
-3.03054214e-01 1.46092162e-01 4.54848260e-01 2.64060169e-01
7.37390071e-02 3.78794894e-02 -1.04143016e-01 -9.82384235e-02
-1.23640180e+00 3.75684239e-02 5.59009314e-01 1.02863514e+00
1.18174553e+00 -2.07462639e-01 -9.23443586e-02 5.91266930e-01
4.21488851e-01 5.14603436e-01 6.04770720e-01 -9.42861915e-01
4.55248564e-01 3.70402843e-01 -3.34818423e-01 -1.25638461e+00
1.01586312e-01 -4.77197766e-03 -7.27006078e-01 -4.81492579e-01
-3.48236188e-02 4.30631042e-01 -8.39884222e-01 1.73252654e+00
3.36734235e-01 7.25541353e-01 1.35696679e-02 1.15139747e+00
8.79563868e-01 9.35830116e-01 5.16407229e-02 -1.78363845e-01
1.30155957e+00 -1.11475289e+00 -5.84230483e-01 -7.86430836e-02
7.24097550e-01 -5.82823455e-01 8.89185190e-01 1.08386822e-01
-9.11294162e-01 -6.86787248e-01 -1.15678465e+00 -3.54686916e-01
-1.25269726e-01 1.02683812e-01 3.59422535e-01 1.81745902e-01
-9.58365023e-01 4.71237689e-01 -8.64464879e-01 -2.91576385e-01
5.23939073e-01 2.82390058e-01 -5.78995347e-01 -5.39595068e-01
-1.48607492e+00 3.94402534e-01 5.28990924e-01 -8.60040039e-02
-1.29797077e+00 -4.28317457e-01 -1.17817664e+00 1.25951976e-01
4.52847958e-01 -2.97309041e-01 9.61174369e-01 -8.36421669e-01
-1.05685151e+00 7.54679561e-01 -2.26523176e-01 -4.71476942e-01
2.54816502e-01 -1.55888572e-01 -2.63660669e-01 8.86590123e-01
1.75353095e-01 1.03587043e+00 1.15334141e+00 -1.12053990e+00
-5.88086545e-01 -2.35403016e-01 8.85871872e-02 2.27664068e-01
-6.00883007e-01 -2.08581477e-01 -1.09676027e+00 -7.71391094e-01
4.75649871e-02 -1.03820515e+00 1.34298891e-01 2.53310621e-01
-2.27637798e-01 -1.07528463e-01 1.10938978e+00 -9.21045423e-01
1.40544319e+00 -2.34312415e+00 2.88075209e-01 9.19376686e-02
2.24372864e-01 2.71938145e-01 -2.38131151e-01 4.50960457e-01
3.29560116e-02 -1.24498978e-01 -2.09698528e-01 -5.08432746e-01
-6.29348531e-02 2.44733348e-01 -3.09712231e-01 6.39136016e-01
2.26448715e-01 9.88148510e-01 -9.27607000e-01 -8.10146093e-01
2.61701554e-01 6.91262066e-01 -7.86845744e-01 5.41158617e-01
-9.21093449e-02 2.47936130e-01 -3.81544024e-01 8.17271113e-01
5.37209928e-01 -5.40710747e-01 1.99404716e-01 -4.98925090e-01
1.43225208e-01 8.30299184e-02 -1.00643837e+00 2.08703256e+00
-9.47451591e-02 6.62266731e-01 -1.25106066e-01 -1.16893578e+00
6.35585129e-01 3.61164510e-01 5.72369218e-01 -7.66203701e-01
1.44030675e-01 1.57820851e-01 -5.32687962e-01 -5.42005599e-01
3.48549277e-01 1.45605281e-01 1.43793020e-02 3.78071487e-01
2.96261668e-01 2.02757001e-01 1.04042850e-01 5.26552081e-01
8.88033807e-01 1.17186755e-01 3.87024850e-01 4.99437675e-02
6.92451000e-01 -2.45282099e-01 6.89471602e-01 4.22095716e-01
-3.96615207e-01 9.03853357e-01 2.33792573e-01 -2.72414148e-01
-9.97516990e-01 -8.83424461e-01 7.02581108e-02 9.50590253e-01
6.22156143e-01 -8.07878017e-01 -7.72512138e-01 -6.90796018e-01
-2.22359002e-02 1.83438346e-01 -6.38099670e-01 -4.90541786e-01
-5.86183250e-01 -2.23678187e-01 4.47412521e-01 4.56434548e-01
7.82899261e-01 -1.01412487e+00 -4.56899315e-01 -2.61262842e-02
-6.01742089e-01 -1.27382147e+00 -7.66213894e-01 -1.43562362e-01
-6.88377440e-01 -1.16969931e+00 -7.20177174e-01 -1.02153051e+00
5.37069798e-01 8.70442688e-01 8.73136818e-01 5.47713041e-01
-3.11696380e-01 2.72007048e-01 -6.38670385e-01 3.92821372e-01
-6.63773641e-02 9.68824849e-02 1.01557188e-02 1.00579947e-01
4.85998720e-01 -4.65868622e-01 -9.37424123e-01 5.73708355e-01
-1.19064045e+00 8.02968964e-02 6.29560590e-01 8.87430847e-01
7.65985370e-01 -1.90681256e-02 1.49001420e-01 -3.74051213e-01
-1.62787482e-01 -7.29935467e-01 -2.93639302e-01 2.76580483e-01
-3.91173571e-01 -1.43783577e-02 6.50811315e-01 -3.82080913e-01
-6.46209061e-01 1.72042802e-01 -6.60103709e-02 -1.10062265e+00
4.50303815e-02 3.42889786e-01 -3.31477135e-01 -1.38591573e-01
8.32525864e-02 7.50235081e-01 2.37764493e-01 -1.72599241e-01
2.86442876e-01 6.55967832e-01 6.21233582e-01 -3.44347954e-01
8.89929891e-01 6.04373753e-01 -4.14137036e-01 -7.01989174e-01
-7.30018318e-01 -7.04125404e-01 -4.92258906e-01 -1.67380214e-01
9.47793782e-01 -1.41697145e+00 -5.88979125e-01 3.99525434e-01
-8.55210841e-01 -2.18372911e-01 7.56395087e-02 5.46568394e-01
-6.02450967e-01 6.00108504e-01 -8.13445568e-01 -3.48802507e-01
-2.78746247e-01 -1.37268484e+00 1.24261594e+00 2.45193511e-01
7.79438093e-02 -6.74130082e-01 -3.46056521e-02 6.13161445e-01
1.85811967e-01 -1.85852274e-01 5.65855861e-01 -5.73351204e-01
-1.03267610e+00 3.03077381e-02 -4.04893935e-01 4.10897285e-01
1.31510850e-02 -1.81349844e-01 -8.85625184e-01 -6.74440384e-01
-9.25074983e-03 -7.44316339e-01 9.96502638e-01 1.18159324e-01
1.29690337e+00 -3.87436450e-01 -2.30517820e-01 8.86945367e-01
1.47557855e+00 -6.97216950e-03 8.88277769e-01 2.33111292e-01
8.77747774e-01 3.51046443e-01 8.30092013e-01 5.53475559e-01
7.73269296e-01 7.51846611e-01 3.53868127e-01 1.40525838e-02
-1.47871584e-01 -6.80674553e-01 5.85677981e-01 1.10306811e+00
1.82191655e-01 -1.48169011e-01 -5.63273847e-01 4.59597260e-01
-1.89121497e+00 -1.12045777e+00 5.05676448e-01 2.11751390e+00
8.51130664e-01 -5.27016371e-02 1.58624262e-01 1.75134018e-01
8.50396037e-01 5.02912700e-01 -5.15281141e-01 4.46559280e-01
-1.57060586e-02 -1.96213260e-01 2.55144447e-01 2.42867291e-01
-1.19345903e+00 9.76619542e-01 4.79723501e+00 1.05305874e+00
-1.14688563e+00 2.27449954e-01 6.39660597e-01 -1.75410286e-01
-1.42377421e-01 8.54819864e-02 -7.54265904e-01 7.80048370e-01
7.89163709e-01 5.77295981e-02 5.12626290e-01 8.78230870e-01
-6.39947355e-02 6.38378710e-02 -1.09701097e+00 1.35289168e+00
4.31576490e-01 -1.45608819e+00 3.09086621e-01 -4.86418605e-02
5.29350042e-01 5.72208408e-03 -6.40540421e-02 4.17669058e-01
-3.24027836e-01 -6.82885468e-01 7.96743572e-01 1.67626321e-01
9.00458932e-01 -6.39677882e-01 7.72902071e-01 3.41049284e-02
-1.68955505e+00 -1.23321533e-01 -3.79083663e-01 3.44285518e-01
2.39814714e-01 1.78423762e-01 -4.50697243e-01 4.60335135e-01
1.00606596e+00 1.25433874e+00 -5.53674042e-01 9.20263350e-01
-2.11834371e-01 4.53482240e-01 -2.02591136e-01 2.38356620e-01
3.18303347e-01 -5.36687188e-02 1.87408403e-01 1.13116992e+00
3.10810149e-01 3.45991731e-01 3.38220060e-01 4.59799975e-01
-2.55683243e-01 -1.41652515e-02 -5.49577594e-01 -2.08885342e-01
7.18770027e-01 1.17209613e+00 -6.90674961e-01 -4.24177766e-01
-6.44833624e-01 1.14698625e+00 2.24660605e-01 2.62603700e-01
-1.19348907e+00 -2.40910247e-01 7.05161154e-01 1.11288384e-01
7.21153080e-01 -1.06946580e-01 3.69040281e-01 -1.53800702e+00
6.42997324e-02 -1.14471471e+00 5.94085813e-01 -6.70790613e-01
-8.49581242e-01 5.56382835e-01 -8.87328908e-02 -1.54728365e+00
-1.36587709e-01 -1.90519422e-01 -3.60998660e-01 1.19069889e-01
-1.73676372e+00 -1.10223269e+00 -6.61836684e-01 1.03293192e+00
6.40398800e-01 -4.30366658e-02 4.49841559e-01 5.95296502e-01
-6.45733058e-01 7.57610083e-01 -3.16585377e-02 3.59240651e-01
8.39065969e-01 -6.94679320e-01 1.44396171e-01 7.54518509e-01
2.13478044e-01 6.34632409e-01 3.85652035e-01 -5.80214679e-01
-1.78169119e+00 -1.18242359e+00 5.62496185e-01 2.45585144e-02
5.94581664e-01 -3.50019366e-01 -1.15150142e+00 5.19639671e-01
1.41562715e-01 3.05372775e-01 7.42271245e-01 -4.90613252e-01
-6.48133814e-01 -2.65553057e-01 -8.82464051e-01 3.55863631e-01
1.19443083e+00 -1.17179227e+00 -5.42256176e-01 2.01591372e-01
8.74864936e-01 -1.50907472e-01 -8.21775734e-01 3.79404664e-01
7.02136099e-01 -1.05638611e+00 1.05971336e+00 -2.57088751e-01
6.88245416e-01 -5.14640331e-01 -4.56756502e-01 -7.84931064e-01
-1.92542106e-01 -5.93817592e-01 -5.04119396e-01 1.06723261e+00
-6.09893538e-02 -2.47947663e-01 7.92938709e-01 3.05740893e-01
8.86131451e-02 -8.60861480e-01 -6.85269952e-01 -6.69085264e-01
-5.54095924e-01 -1.71519861e-01 4.91843522e-01 1.02161145e+00
4.87212874e-02 8.54399651e-02 -7.30580211e-01 2.41549715e-01
7.56835043e-01 2.50691652e-01 7.25081503e-01 -6.72976851e-01
-2.78683096e-01 -2.53540394e-03 -7.49435604e-01 -1.39021313e+00
2.23615512e-01 -7.29641974e-01 1.69566035e-01 -1.05438769e+00
6.49950027e-01 -2.57428110e-01 -5.43424070e-01 5.95849872e-01
-3.15370560e-01 6.23127878e-01 3.06868017e-01 6.90342367e-01
-1.21981299e+00 9.62576091e-01 9.35056984e-01 -1.01810418e-01
1.52852207e-01 -5.54299653e-01 -5.03061116e-01 2.95032740e-01
5.49477637e-01 -4.88178134e-01 -6.36184275e-01 -4.24026102e-01
-2.17988536e-01 2.02469900e-01 4.96169686e-01 -1.11784780e+00
4.51346695e-01 1.06237225e-01 3.92702073e-01 -5.72335899e-01
4.57525611e-01 -7.64285922e-01 1.71900600e-01 5.84732652e-01
-2.52532959e-01 1.60650671e-01 -2.01750711e-01 7.08671510e-01
-6.77745521e-01 -3.17075141e-02 7.45522439e-01 -4.02616113e-02
-1.08441365e+00 6.82351232e-01 1.22047022e-01 -2.69395299e-02
1.11926603e+00 -2.25466758e-01 -3.67330372e-01 -6.65753007e-01
-2.50387400e-01 4.35632944e-01 7.98007369e-01 5.38062274e-01
1.00301242e+00 -1.46493793e+00 -4.38968539e-01 3.95264417e-01
2.55050570e-01 -1.36117220e-01 5.15088320e-01 7.85761416e-01
-4.64908481e-01 4.46180344e-01 -2.54432917e-01 -7.36157060e-01
-1.27566433e+00 7.18079627e-01 -1.27468944e-01 4.59502898e-02
-5.30139506e-01 8.17828238e-01 3.06883901e-01 1.05229013e-01
3.90065253e-01 -3.84672061e-02 -1.90065466e-02 7.28706196e-02
8.65265965e-01 9.49986652e-02 -2.55658656e-01 -8.50297213e-01
-4.05100405e-01 5.33074617e-01 -3.14711452e-01 1.75980166e-01
1.23121524e+00 -2.83889264e-01 -1.16422419e-02 4.26579826e-02
1.86769867e+00 -4.42787528e-01 -1.51930249e+00 -4.54615831e-01
-2.78093159e-01 -7.60355234e-01 1.16306975e-01 -4.24891338e-03
-1.42037070e+00 8.09948862e-01 5.52800119e-01 -1.33858725e-01
1.37259555e+00 1.70390666e-01 1.34842932e+00 3.03095311e-01
4.57518727e-01 -8.98337543e-01 4.67336059e-01 1.58898816e-01
5.63560843e-01 -1.48355091e+00 9.83379781e-02 -3.37419719e-01
-7.69244730e-01 8.46887648e-01 5.94592392e-01 -2.72550583e-02
5.69775224e-01 -1.32961318e-01 -1.51383862e-01 -1.06762454e-01
-7.57126749e-01 -9.02748629e-02 2.06129044e-01 2.42750257e-01
1.13722295e-01 -2.99527198e-01 -1.11495733e-01 3.90920252e-01
1.38517961e-01 7.01303110e-02 1.06243558e-01 1.14353895e+00
-4.02820528e-01 -8.05309415e-01 -1.08829148e-01 1.54550895e-01
-2.56942511e-01 -2.85934836e-01 -2.65413038e-02 7.06331849e-01
-9.66617763e-02 7.38459766e-01 -6.70751259e-02 -8.10548663e-01
-9.33928937e-02 -2.54331201e-01 2.02889159e-01 -3.92868936e-01
-1.47299320e-01 2.13131130e-01 -3.18716168e-01 -1.04178131e+00
-5.80928147e-01 -5.89778900e-01 -1.33058584e+00 -3.07891011e-01
-2.87105680e-01 3.41342241e-01 3.87353152e-01 8.18673849e-01
4.98673975e-01 -2.23921947e-02 9.91771340e-01 -9.34570432e-01
-4.10789311e-01 -5.76737523e-01 -4.43546414e-01 6.48296177e-01
4.55864936e-01 -7.50826061e-01 -4.68194842e-01 2.78163999e-01] | [10.063794136047363, 0.7104232907295227] |
7cd18a96-2733-470e-859c-85587eb5ea4a | motion-aware-transformer-for-occluded-person | 2202.04243 | null | https://arxiv.org/abs/2202.04243v2 | https://arxiv.org/pdf/2202.04243v2.pdf | Motion-Aware Transformer For Occluded Person Re-identification | Recently, occluded person re-identification(Re-ID) remains a challenging task that people are frequently obscured by other people or obstacles, especially in a crowd massing situation. In this paper, we propose a self-supervised deep learning method to improve the location performance for human parts through occluded person Re-ID. Unlike previous works, we find that motion information derived from the photos of various human postures can help identify major human body components. Firstly, a motion-aware transformer encoder-decoder architecture is designed to obtain keypoints heatmaps and part-segmentation maps. Secondly, an affine transformation module is utilized to acquire motion information from the keypoint detection branch. Then the motion information will support the segmentation branch to achieve refined human part segmentation maps, and effectively divide the human body into reasonable groups. Finally, several cases demonstrate the efficiency of the proposed model in distinguishing different representative parts of the human body, which can avoid the background and occlusion disturbs. Our method consistently achieves state-of-the-art results on several popular datasets, including occluded, partial, and holistic. | ['Xiai Chen', 'Wei Hong', 'Zhekun Lv', 'Hongye Liu', 'Mi Zhou'] | 2022-02-09 | null | null | null | null | ['human-part-segmentation'] | ['computer-vision'] | [-1.68565989e-01 -1.84960216e-01 -1.91225275e-01 -1.97322711e-01
-5.29328704e-01 -2.61669546e-01 2.79537976e-01 -4.30639237e-01
-4.16632503e-01 6.01296306e-01 5.49200177e-01 5.80773771e-01
3.82041126e-01 -5.31779110e-01 -4.73093629e-01 -6.60755575e-01
2.94937819e-01 5.43955743e-01 3.94534290e-01 -9.71745774e-02
-1.26680091e-01 2.41609469e-01 -1.57087564e+00 -2.14817822e-02
8.03910017e-01 7.05600142e-01 4.26945426e-02 4.19003844e-01
1.42753229e-01 4.28712398e-01 -7.31593788e-01 -4.87973869e-01
4.36462671e-01 -4.88870114e-01 -8.15715313e-01 6.00477695e-01
6.65835798e-01 -7.26189077e-01 -6.54629350e-01 9.79392886e-01
8.58917236e-01 1.81140542e-01 5.59671164e-01 -1.25900280e+00
-3.02545130e-01 3.00370157e-01 -1.01603270e+00 1.30439430e-01
5.25005281e-01 2.84446597e-01 2.76829898e-01 -7.10678279e-01
5.41344047e-01 1.50568056e+00 9.05417144e-01 7.90423930e-01
-8.47021222e-01 -7.17776418e-01 3.22904885e-01 4.54187095e-01
-1.74877441e+00 -4.28166091e-01 8.57312441e-01 -4.44683075e-01
3.09036851e-01 1.57790333e-01 1.07604039e+00 1.00293553e+00
-2.11914778e-02 1.19485998e+00 6.18094027e-01 -2.03755975e-01
-2.06320748e-01 -3.03658750e-02 4.04894650e-02 7.97276020e-01
5.90577960e-01 -2.77175516e-01 -3.02871048e-01 1.66836858e-01
8.60269308e-01 2.95046240e-01 -3.65695596e-01 -4.17952925e-01
-1.40351200e+00 2.24934921e-01 4.53980535e-01 1.19427435e-01
-3.94207716e-01 8.64514261e-02 4.86816436e-01 -3.74985874e-01
2.44810745e-01 -3.30974102e-01 -2.08367750e-01 9.74828601e-02
-1.14329445e+00 3.97479445e-01 4.60015655e-01 1.08576405e+00
5.89736581e-01 -5.08614555e-02 -5.42579949e-01 7.68247485e-01
4.84475732e-01 7.54097342e-01 5.27818799e-01 -8.50003064e-01
5.30175805e-01 7.45816588e-01 3.97782922e-01 -9.76650715e-01
-5.66994607e-01 -4.85988766e-01 -1.16661859e+00 -1.15381069e-01
6.83252573e-01 -2.20048681e-01 -1.02855480e+00 1.47521639e+00
8.59842777e-01 -2.58881226e-02 -1.68500185e-01 1.38715279e+00
9.21820581e-01 3.46474737e-01 2.70996660e-01 1.86331815e-03
1.76900840e+00 -1.33241129e+00 -7.11557686e-01 -4.54811931e-01
3.19708347e-01 -5.03999949e-01 6.07947052e-01 8.55995119e-02
-1.03824544e+00 -1.05084229e+00 -8.10788631e-01 -2.36135140e-01
9.87158250e-03 6.59091830e-01 2.26677924e-01 6.70648873e-01
-6.75937295e-01 3.42333704e-01 -7.77421594e-01 -4.66838419e-01
6.04646444e-01 2.78818369e-01 -4.00154889e-01 -1.71463788e-01
-1.08745027e+00 5.86928070e-01 2.37462088e-01 5.26542604e-01
-9.04245019e-01 -2.66297668e-01 -9.75562215e-01 -1.60651237e-01
1.98972166e-01 -1.13272536e+00 1.00586641e+00 -8.56779575e-01
-1.20487118e+00 1.00269330e+00 -4.92164820e-01 -2.77544945e-01
1.08530486e+00 -4.89611119e-01 -3.58126134e-01 1.95281461e-01
3.72961640e-01 8.54397535e-01 9.03361499e-01 -1.14580297e+00
-9.76207376e-01 -6.18184686e-01 -3.86366606e-01 5.45143723e-01
-6.05743006e-02 -9.26920250e-02 -1.15701807e+00 -7.74971306e-01
2.17174858e-01 -1.03404033e+00 -2.51745820e-01 7.29276687e-02
-8.30426812e-01 -3.18486094e-01 7.19297767e-01 -1.28096771e+00
1.08873284e+00 -1.99947453e+00 2.67292798e-01 4.29351889e-02
3.40031475e-01 2.75364637e-01 2.22330853e-01 -1.16478145e-01
1.45107448e-01 -1.14907689e-01 -9.22560096e-02 -5.85622489e-01
-4.24023382e-02 -5.43622449e-02 2.72743285e-01 8.82168591e-01
-3.13572168e-01 1.08418655e+00 -5.42447746e-01 -1.04119647e+00
4.72179145e-01 5.10619521e-01 -6.42097667e-02 1.87563196e-01
3.62349480e-01 7.07094371e-01 -2.67029196e-01 9.15428996e-01
8.00336480e-01 -1.17144853e-01 -2.21747726e-01 -6.37384892e-01
1.00898758e-01 -2.26926804e-01 -1.37587726e+00 1.82547092e+00
1.66408300e-01 4.95983541e-01 4.33580652e-02 -7.01403379e-01
6.54717624e-01 2.09878087e-01 6.97414994e-01 -5.92541218e-01
2.18033254e-01 -1.50438800e-01 -2.76009619e-01 -5.26481152e-01
2.92364568e-01 4.04130220e-01 1.06100766e-02 1.69305205e-01
-2.70285040e-01 6.17046833e-01 1.67465836e-01 -9.79957581e-02
4.09774989e-01 3.39422882e-01 3.04940104e-01 -1.97482869e-01
8.44929993e-01 3.83157730e-02 8.49455118e-01 4.58718836e-01
-8.49190056e-01 8.50028992e-01 -1.95858717e-01 -7.76594281e-01
-1.06071806e+00 -9.73686635e-01 7.45298713e-02 8.68092597e-01
8.98136675e-01 -1.74167767e-01 -1.10353029e+00 -6.26300216e-01
-1.35284409e-01 6.30205795e-02 -5.42757988e-01 -1.03141822e-01
-8.68237019e-01 -7.96256006e-01 6.08136058e-01 7.95686662e-01
1.39794886e+00 -7.96290338e-01 -6.69745803e-01 1.45764470e-01
-9.24542487e-01 -1.08814740e+00 -9.67394412e-01 -5.52723169e-01
-5.70042014e-01 -1.17788815e+00 -1.63927138e+00 -1.13254285e+00
7.63498008e-01 6.73915267e-01 7.22318590e-01 7.38747343e-02
-5.25254250e-01 3.84524882e-01 -3.79263982e-02 -1.11482449e-01
1.77603737e-01 6.15752302e-02 2.25598484e-01 1.75321445e-01
5.12411237e-01 -9.52377543e-02 -1.26960170e+00 8.05491507e-01
-2.37902284e-01 2.68314958e-01 4.58793849e-01 5.35482883e-01
7.29334593e-01 2.19431788e-01 1.45879820e-01 -2.37361625e-01
5.23228757e-02 -9.69377309e-02 -1.40430123e-01 4.63050544e-01
-1.53375432e-01 -2.92108983e-01 1.73745021e-01 -5.65778911e-01
-1.16568136e+00 4.66428697e-01 -8.64360482e-02 -2.49859199e-01
-4.67827559e-01 -4.03432459e-01 -7.04924703e-01 -2.54983120e-02
3.54536235e-01 4.05324221e-01 -1.25019312e-01 -6.17666900e-01
3.67077082e-01 7.66540945e-01 9.74207997e-01 -2.38808841e-01
8.58120024e-01 8.03989053e-01 -2.42315531e-01 -7.74388134e-01
-6.02560818e-01 -7.38156259e-01 -1.03778207e+00 -5.34783006e-01
1.28321493e+00 -1.35294020e+00 -6.80063844e-01 8.54699433e-01
-1.17469382e+00 -1.14730872e-01 -2.10712090e-01 4.24544036e-01
-1.88819259e-01 7.08544910e-01 -6.98501468e-01 -8.24682355e-01
-7.07959950e-01 -1.01020503e+00 1.51412058e+00 7.11859763e-01
-1.57919243e-01 -5.58951437e-01 -1.89045563e-01 5.80442965e-01
-7.83048794e-02 2.10711569e-01 1.64859727e-01 -3.57188553e-01
-5.11728108e-01 -3.35514486e-01 -2.19414413e-01 1.42766284e-02
2.13365763e-01 -4.06237513e-01 -8.62911701e-01 -3.89892817e-01
-2.63563216e-01 1.68835223e-01 7.58053005e-01 6.03579283e-01
8.61430228e-01 -4.11900580e-01 -9.04445648e-01 8.33944023e-01
7.70434558e-01 1.02202736e-01 7.09851503e-01 4.67022836e-01
1.30570495e+00 8.37463498e-01 6.79427624e-01 4.46042091e-01
8.53459001e-01 8.32397997e-01 5.96830854e-03 -4.37805951e-01
-4.61105347e-01 -4.50280458e-01 2.55275339e-01 3.91359389e-01
-3.41449946e-01 -5.31660691e-02 -7.36456156e-01 6.10221148e-01
-1.86432755e+00 -1.04186654e+00 -3.06035697e-01 2.20888424e+00
4.24551070e-01 -2.67090708e-01 6.33037746e-01 1.33088127e-01
1.29891372e+00 8.18524044e-03 -6.88974619e-01 7.56465137e-01
-1.57709256e-01 -6.10132039e-01 6.88668966e-01 2.34144002e-01
-1.56895983e+00 9.47172403e-01 5.95604134e+00 7.90482283e-01
-6.37724578e-01 1.23309739e-01 6.25192821e-01 4.85490710e-02
2.16521904e-01 -4.45419937e-01 -1.13476288e+00 7.09545195e-01
9.39414203e-02 -3.25479396e-02 8.11159983e-02 1.00708449e+00
1.03821822e-01 -7.45844394e-02 -7.96183169e-01 1.61571741e+00
3.83384228e-01 -6.93899095e-01 -1.83530316e-01 2.57192440e-02
7.24347293e-01 -4.63918775e-01 -1.35701731e-01 3.49898450e-02
2.69354172e-02 -8.07214797e-01 8.52073312e-01 6.42519474e-01
6.63486421e-01 -8.23352098e-01 7.60965288e-01 2.56692618e-01
-1.75843608e+00 -2.24146079e-02 -4.87467408e-01 2.23815903e-01
5.30104578e-01 3.12068790e-01 -4.62351501e-01 5.42452455e-01
1.06730735e+00 7.16380298e-01 -9.06464517e-01 1.33182526e+00
-2.30254620e-01 1.88636184e-01 -2.86035538e-01 3.05475265e-01
-3.56129348e-01 -5.92604559e-03 4.97760564e-01 1.19835234e+00
1.58723205e-01 7.34930784e-02 4.11508650e-01 5.17641366e-01
2.32291728e-01 1.19688120e-02 -1.93741441e-01 5.36093473e-01
2.24300072e-01 1.21475661e+00 -8.98302555e-01 -6.09973431e-01
-1.69595718e-01 1.68174517e+00 -9.64815766e-02 4.91311461e-01
-1.11746895e+00 -2.88750708e-01 6.03136837e-01 2.87672013e-01
2.63079315e-01 -2.34447777e-01 3.88571247e-02 -1.38797104e+00
9.59873050e-02 -7.97228813e-01 6.10421896e-01 -7.19018161e-01
-1.12010801e+00 4.48678583e-01 4.21694666e-02 -1.44784546e+00
-2.09136546e-01 -2.14700386e-01 -4.04786408e-01 7.80350149e-01
-8.73479247e-01 -1.45519519e+00 -9.31397915e-01 8.20919991e-01
7.52567410e-01 -1.37975156e-01 2.30916798e-01 6.08796716e-01
-9.74684656e-01 8.07811677e-01 -2.37274945e-01 5.94999194e-01
8.02500308e-01 -9.35007274e-01 6.31583571e-01 1.14747405e+00
-5.06864376e-02 4.50859725e-01 4.35033202e-01 -1.06313348e+00
-1.08255470e+00 -1.35236037e+00 5.73279440e-01 -4.45876658e-01
-3.24063599e-01 -3.07456970e-01 -7.18770087e-01 6.04634225e-01
-1.75973371e-01 3.20074446e-02 4.24223423e-01 -3.59934956e-01
8.78461823e-02 -1.59247428e-01 -1.03962231e+00 7.33414888e-01
1.53165758e+00 -2.15964019e-01 -6.37192726e-01 2.27788493e-01
4.18343872e-01 -5.93156457e-01 -5.69714487e-01 4.56170887e-01
8.48635733e-01 -7.23312080e-01 1.48639166e+00 -3.54877770e-01
2.71805073e-03 -7.40946412e-01 1.56997472e-01 -9.22231436e-01
-5.70861638e-01 -4.26303178e-01 -2.83999056e-01 1.32023001e+00
-3.54481876e-01 -3.70418161e-01 1.03964806e+00 7.79992163e-01
2.06667677e-01 -2.92958409e-01 -8.00297558e-01 -4.89263982e-01
-3.92606914e-01 7.57856146e-02 6.10367537e-01 5.59177220e-01
-3.12739760e-01 8.98985192e-02 -8.62572968e-01 2.73829550e-01
9.76383567e-01 1.09160021e-01 1.23191082e+00 -1.07265246e+00
8.02986175e-02 -1.91574380e-01 -6.77935779e-01 -1.65823710e+00
-3.60111184e-02 -4.75109190e-01 1.26842678e-01 -1.88239992e+00
4.32984531e-01 -9.01605114e-02 3.54890451e-02 3.51592451e-01
-5.20784795e-01 4.63585258e-01 1.62794918e-01 6.95002675e-01
-8.37373495e-01 6.24800384e-01 1.43734825e+00 -4.54031229e-01
-2.97490567e-01 1.90618172e-01 -4.92266685e-01 9.32776213e-01
6.06269419e-01 -1.12802856e-01 -1.63938701e-01 -4.01823968e-01
-4.78595883e-01 -1.39426380e-01 8.44700873e-01 -1.43548524e+00
5.39235890e-01 1.31172270e-01 1.15423930e+00 -1.09595811e+00
3.49673420e-01 -7.48263955e-01 3.99157226e-01 8.12176883e-01
7.21499175e-02 -2.03126278e-02 3.58711183e-02 6.53558075e-01
3.85477245e-02 8.22436064e-02 7.04147816e-01 -2.58069992e-01
-9.78289783e-01 5.70174575e-01 -1.69336110e-01 6.07060529e-02
1.08972490e+00 -4.63657856e-01 -1.46241203e-01 -3.42103601e-01
-6.26605511e-01 4.48982269e-01 5.35310566e-01 4.86561954e-01
5.47295928e-01 -1.48676240e+00 -7.43484199e-01 2.48243153e-01
-2.05054712e-02 1.73873350e-01 7.39558816e-01 9.08124447e-01
-6.39088690e-01 3.15106273e-01 -3.05454105e-01 -8.16653192e-01
-1.46939015e+00 6.80337965e-01 6.91508055e-01 1.52807206e-01
-1.01618719e+00 7.01397061e-01 4.90036428e-01 5.14268130e-02
4.22065526e-01 -9.93812010e-02 -2.93026119e-01 1.06007727e-02
8.62523854e-01 9.05573249e-01 -5.14445901e-01 -1.30801368e+00
-5.83383203e-01 1.10504436e+00 5.08642383e-02 -3.29221599e-02
5.34537435e-01 -7.40516841e-01 2.45827496e-01 3.02249007e-02
1.03931332e+00 -1.08969569e-01 -1.56736422e+00 -2.34385848e-01
-4.98683929e-01 -5.57808816e-01 -5.16093194e-01 -4.37134057e-01
-1.20450628e+00 9.61199105e-01 1.08947301e+00 -3.16345781e-01
9.76677358e-01 -7.12756440e-02 1.18166471e+00 7.89605454e-02
5.79246998e-01 -1.20068777e+00 -6.73879460e-02 -3.09771243e-02
5.32342315e-01 -1.29298127e+00 1.87112942e-01 -5.27408004e-01
-9.44846690e-01 7.82476544e-01 9.28319752e-01 1.74224555e-01
2.52439857e-01 -1.57655686e-01 1.35652497e-01 1.10583201e-01
2.89178789e-01 -6.56465411e-01 5.77736020e-01 1.01560986e+00
3.57310474e-02 -2.21508984e-02 -1.83643892e-01 8.83129299e-01
-2.38515764e-01 -1.97876900e-01 -1.10983089e-01 5.91001213e-01
-4.58158463e-01 -6.87174439e-01 -7.98118532e-01 1.39331639e-01
-3.52790385e-01 3.80346805e-01 -3.42774481e-01 8.73226166e-01
5.40367663e-01 8.57109189e-01 3.80676016e-02 -4.42704499e-01
3.36505860e-01 -1.77938938e-02 3.89092207e-01 -1.39563128e-01
-2.85279274e-01 2.68476546e-01 -1.04278579e-01 -6.36008918e-01
-5.16559482e-01 -6.17832959e-01 -1.09510374e+00 -2.71854162e-01
-1.67366162e-01 -1.01464972e-01 1.64024428e-01 9.62920189e-01
2.81610042e-01 5.82463622e-01 2.37774074e-01 -1.22371519e+00
-7.85478577e-02 -8.86614978e-01 -4.45685476e-01 7.10199893e-01
2.99072027e-01 -7.91433454e-01 -6.96266666e-02 4.34517056e-01] | [14.595158576965332, 0.8008497953414917] |
03ee7b78-9030-4c90-86ed-4388ffe19da6 | a-comprehensive-survey-on-deep-graph | 2304.05055 | null | https://arxiv.org/abs/2304.05055v2 | https://arxiv.org/pdf/2304.05055v2.pdf | A Comprehensive Survey on Deep Graph Representation Learning | Graph representation learning aims to effectively encode high-dimensional sparse graph-structured data into low-dimensional dense vectors, which is a fundamental task that has been widely studied in a range of fields, including machine learning and data mining. Classic graph embedding methods follow the basic idea that the embedding vectors of interconnected nodes in the graph can still maintain a relatively close distance, thereby preserving the structural information between the nodes in the graph. However, this is sub-optimal due to: (i) traditional methods have limited model capacity which limits the learning performance; (ii) existing techniques typically rely on unsupervised learning strategies and fail to couple with the latest learning paradigms; (iii) representation learning and downstream tasks are dependent on each other which should be jointly enhanced. With the remarkable success of deep learning, deep graph representation learning has shown great potential and advantages over shallow (traditional) methods, there exist a large number of deep graph representation learning techniques have been proposed in the past decade, especially graph neural networks. In this survey, we conduct a comprehensive survey on current deep graph representation learning algorithms by proposing a new taxonomy of existing state-of-the-art literature. Specifically, we systematically summarize the essential components of graph representation learning and categorize existing approaches by the ways of graph neural network architectures and the most recent advanced learning paradigms. Moreover, this survey also provides the practical and promising applications of deep graph representation learning. Last but not least, we state new perspectives and suggest challenging directions which deserve further investigations in the future. | ['Ming Zhang', 'Xiao Luo', 'Yusheng Zhao', 'Jingyang Yuan', 'Junwei Yang', 'Zhiping Xiao', 'Fang Sun', 'Jianhao Shen', 'Yifang Qin', 'Ziyue Qiao', 'Qingqing Long', 'Zequn Liu', 'Yiyang Gu', 'Zheng Fang', 'Wei Ju'] | 2023-04-11 | null | null | null | null | ['graph-embedding'] | ['graphs'] | [ 3.33481096e-02 3.46254498e-01 -5.74732840e-01 -9.44072530e-02
2.15134807e-02 -2.09341139e-01 4.74188685e-01 4.74703759e-01
2.49434467e-02 3.48325014e-01 2.79079229e-01 -2.77454555e-01
-3.36396515e-01 -1.07585835e+00 -3.93022269e-01 -8.06295991e-01
-4.65157241e-01 4.41874951e-01 -1.04657233e-01 -3.13373774e-01
5.84271103e-02 5.25355041e-01 -1.28819215e+00 4.78880294e-02
5.51508009e-01 8.42361569e-01 4.23997752e-02 3.79106164e-01
-3.57312560e-01 9.41905022e-01 -4.40177679e-01 -3.65072668e-01
-1.40614167e-03 -4.81399447e-01 -5.69138348e-01 2.18313448e-02
3.24545473e-01 1.38747026e-04 -1.30331147e+00 1.19722533e+00
4.27822053e-01 2.36304954e-01 6.04904950e-01 -1.28229725e+00
-1.21100223e+00 5.94119608e-01 -6.02040887e-01 2.02544212e-01
2.08685696e-01 -1.59140706e-01 1.31495929e+00 -4.77669686e-01
4.76881385e-01 1.08466458e+00 6.32784128e-01 3.51039171e-01
-1.14008975e+00 -5.68309426e-01 5.91666877e-01 3.97054851e-01
-1.37804735e+00 -4.53389017e-03 1.21808648e+00 -3.45666200e-01
1.15321279e+00 3.03519354e-03 9.13369536e-01 8.51082563e-01
2.44289532e-01 6.36213362e-01 5.95250845e-01 -3.61451536e-01
9.63992625e-03 -2.57669896e-01 2.86586463e-01 1.08115184e+00
5.61832130e-01 2.42628660e-02 -1.57718673e-01 -3.49447876e-02
7.77866185e-01 4.15239781e-01 -1.21889912e-01 -8.76495481e-01
-1.08792078e+00 1.24224973e+00 1.11089706e+00 7.15149403e-01
-2.17262194e-01 3.23973805e-01 6.59068704e-01 4.87362951e-01
3.33443612e-01 3.41133863e-01 2.01537479e-02 3.84472281e-01
-5.06161034e-01 6.82302192e-02 5.67546308e-01 8.65300238e-01
8.84274125e-01 5.91933250e-01 1.71868186e-02 8.37208390e-01
3.27718556e-01 4.29892614e-02 4.87467140e-01 -2.33808666e-01
5.94188750e-01 9.93873119e-01 -7.09017217e-01 -1.72153652e+00
-5.73267579e-01 -5.74264884e-01 -1.52953291e+00 -4.37002853e-02
-8.77207145e-02 8.75062346e-02 -8.10251772e-01 1.49856341e+00
-1.67801663e-01 2.59039849e-01 -1.53816096e-03 6.67613268e-01
1.38003361e+00 6.85134053e-01 2.15807408e-02 -8.52654949e-02
1.00546753e+00 -9.48642850e-01 -7.91644871e-01 -3.99446905e-01
7.90949583e-01 -2.57504255e-01 7.82314658e-01 1.72818005e-02
-8.79069388e-01 -4.71736938e-01 -1.19460094e+00 -1.60595983e-01
-6.86943114e-01 -1.20269008e-01 1.31155741e+00 3.97112012e-01
-1.20050430e+00 5.81712604e-01 -8.18752289e-01 -5.01462102e-01
6.39170349e-01 5.49681485e-01 -5.98588705e-01 -3.78705621e-01
-1.28834426e+00 5.31727254e-01 4.38624114e-01 1.85340434e-01
-6.77495182e-01 -2.61614263e-01 -1.40779638e+00 3.61360133e-01
2.04936609e-01 -6.39818072e-01 6.17483139e-01 -5.90393543e-01
-1.17638576e+00 8.40248644e-01 1.79943904e-01 -4.88945246e-01
-1.18250996e-01 1.67088360e-01 -4.90206927e-01 1.39329821e-01
-1.59893751e-01 2.64141113e-01 5.61196983e-01 -9.76386547e-01
-1.88137427e-01 -4.94171560e-01 1.81766167e-01 2.51724154e-01
-8.01115870e-01 -2.65402853e-01 -3.93504530e-01 -8.82918060e-01
2.03704536e-01 -6.89530671e-01 -5.43667555e-01 -8.99637267e-02
-3.22810352e-01 -3.98476928e-01 8.90976489e-01 -2.31083959e-01
1.66591787e+00 -2.11989594e+00 5.55281460e-01 8.73660445e-02
8.77654314e-01 4.85988468e-01 -3.56899559e-01 9.22306657e-01
-3.88850242e-01 7.61248693e-02 -2.54976898e-01 -7.96598569e-02
-1.09108284e-01 2.66186237e-01 -3.02314371e-01 7.34800339e-01
5.55314170e-03 1.33730161e+00 -1.05068159e+00 -2.28740811e-01
5.13862491e-01 6.52046502e-01 -3.97404164e-01 3.01527798e-01
-2.85159200e-02 1.17627576e-01 -4.95218962e-01 3.97684544e-01
4.00061369e-01 -5.91230214e-01 4.85993952e-01 -2.68370599e-01
2.64042646e-01 3.11629444e-01 -1.00297999e+00 1.82072866e+00
-1.88144520e-01 6.63565636e-01 -5.26042767e-02 -1.91831827e+00
1.10478270e+00 2.15963081e-01 6.48400664e-01 -6.95125520e-01
7.35308081e-02 -3.17014232e-02 2.10435271e-01 -1.52212143e-01
2.37956628e-01 -1.92274407e-01 -6.98904246e-02 4.69962835e-01
1.41048104e-01 2.35732291e-02 2.06239358e-01 4.76059824e-01
1.18440866e+00 -4.20077354e-01 6.32900119e-01 -3.93809751e-02
4.63935673e-01 -3.71720761e-01 2.84683198e-01 3.79783958e-01
-1.58424556e-01 4.42288131e-01 8.20425332e-01 -7.06455886e-01
-6.27641618e-01 -8.45816910e-01 2.24065095e-01 8.60090196e-01
1.30844221e-01 -7.82926142e-01 -2.08789214e-01 -8.00956309e-01
6.29733130e-02 2.48337477e-01 -9.18566108e-01 -6.42803729e-01
-5.65479755e-01 -7.25231409e-01 3.39337945e-01 7.00020134e-01
2.78665185e-01 -1.08249545e+00 1.99701086e-01 1.67473614e-01
3.60018462e-01 -9.08446848e-01 -1.93517432e-01 1.71529263e-01
-1.12694490e+00 -1.15624607e+00 -7.56282806e-01 -1.37767446e+00
8.87668133e-01 7.66336620e-01 1.15474463e+00 6.28641069e-01
-1.36627093e-01 2.28429765e-01 -4.58617955e-01 5.77651244e-03
-6.85483813e-02 3.48742485e-01 -8.52166116e-02 -1.95812434e-01
3.65841806e-01 -8.71068358e-01 -3.97917122e-01 -1.04149848e-01
-8.53080750e-01 -1.13643669e-01 6.25826418e-01 1.05193365e+00
6.73042119e-01 4.19345766e-01 7.22225904e-01 -1.08377111e+00
8.09556544e-01 -6.79137468e-01 -5.08264899e-01 1.32856131e-01
-6.79314494e-01 2.34446637e-02 9.04867053e-01 -1.14279494e-01
-3.17954302e-01 -2.04865187e-01 -1.50633320e-01 -6.11667156e-01
1.77422374e-01 1.01884460e+00 -2.54520565e-01 -2.40471214e-01
3.86597812e-01 3.24127406e-01 1.35531217e-01 -3.98832917e-01
5.63884854e-01 2.65488535e-01 1.12028092e-01 -2.38523409e-01
7.79370368e-01 2.45539844e-01 2.56886512e-01 -8.71030271e-01
-7.39598870e-01 -3.78558367e-01 -7.07974732e-01 3.41251157e-02
6.38474345e-01 -7.47652709e-01 -4.59913701e-01 2.63132483e-01
-9.07294571e-01 -2.63590962e-01 -1.28720343e-01 4.56493616e-01
-3.69613796e-01 7.79843867e-01 -7.52263844e-01 -3.95822346e-01
-3.41245323e-01 -1.01279712e+00 7.64490187e-01 1.43428817e-01
6.62019700e-02 -1.56975520e+00 2.15350464e-01 1.78870396e-03
3.62031490e-01 3.72195214e-01 1.34934044e+00 -6.41930521e-01
-3.64544302e-01 -5.60871720e-01 -4.26323414e-01 2.47849941e-01
4.34132189e-01 -2.43091166e-01 -5.17469943e-01 -6.52949274e-01
-5.28375030e-01 -2.76536465e-01 1.03893018e+00 4.16621804e-01
1.39874446e+00 -2.11670578e-01 -5.59226692e-01 8.31716716e-01
1.40201652e+00 -8.47041830e-02 5.97004890e-01 -8.92281439e-03
1.29046488e+00 5.48531175e-01 6.65268824e-02 5.86702526e-02
4.25171614e-01 4.59293038e-01 6.94090366e-01 -3.23591024e-01
-3.19498181e-01 -4.15054888e-01 1.64891273e-01 1.30911660e+00
-1.27424985e-01 -4.21911865e-01 -7.53762960e-01 4.91411507e-01
-1.90769875e+00 -1.02600682e+00 -2.99017178e-03 2.06984186e+00
1.75642729e-01 1.45475000e-01 1.81690812e-01 2.80091435e-01
7.44911969e-01 8.18733931e-01 -4.71674055e-01 -4.76596236e-01
-1.39416754e-01 2.79056251e-01 2.23894030e-01 2.70534456e-01
-1.08033788e+00 1.09983552e+00 6.62124920e+00 4.60511059e-01
-1.15761447e+00 -2.10859060e-01 3.69539559e-01 2.18194664e-01
-4.73907262e-01 -1.28008693e-01 -2.26170376e-01 2.12591097e-01
7.05567479e-01 -3.66570503e-01 4.81853724e-01 8.52322102e-01
-3.25859666e-01 6.11501932e-01 -1.13962948e+00 1.14407504e+00
2.66624272e-01 -1.62523615e+00 4.58309352e-01 2.00088888e-01
6.22971833e-01 2.31343716e-01 8.38542953e-02 6.45603776e-01
2.09320128e-01 -1.42208123e+00 -1.02715656e-01 2.02007607e-01
9.53578413e-01 -9.37120914e-01 7.46794522e-01 8.53599608e-02
-1.69615209e+00 -1.34832457e-01 -7.64405668e-01 -4.45171297e-01
9.12111327e-02 6.22045338e-01 -3.50432664e-01 8.86756718e-01
3.18070918e-01 1.62333190e+00 -5.22550225e-01 9.91097629e-01
-4.85728353e-01 5.17126799e-01 1.35912269e-01 -1.07777059e-01
5.16328931e-01 -6.16876841e-01 3.77372235e-01 1.08471358e+00
1.31153062e-01 -6.12645783e-02 3.30440879e-01 7.58494139e-01
-4.53839988e-01 9.67448875e-02 -1.10818231e+00 -7.57706523e-01
2.88107544e-01 1.14253163e+00 -7.79148042e-01 -1.67759389e-01
-8.72159958e-01 7.23888040e-01 8.19870234e-01 4.49761003e-01
-5.18834412e-01 -6.30041718e-01 7.74604380e-01 9.81603786e-02
2.83015072e-01 -5.55509686e-01 -6.55088574e-02 -1.21983433e+00
-1.24482274e-01 -8.07348251e-01 8.38118136e-01 -3.04845989e-01
-1.49298000e+00 7.35153556e-01 -2.47137815e-01 -1.14343929e+00
-1.24320291e-01 -7.96285808e-01 -7.51673400e-01 6.66161239e-01
-1.57349288e+00 -1.21652257e+00 -3.26586604e-01 5.96651614e-01
2.52313405e-01 -5.46894372e-01 1.11014855e+00 4.72937405e-01
-7.93547809e-01 6.15172148e-01 1.31115094e-01 5.94906569e-01
2.37503752e-01 -1.16230977e+00 6.83599114e-01 6.16478860e-01
5.19775510e-01 7.44088173e-01 1.33633956e-01 -5.65953493e-01
-1.79939854e+00 -1.22762001e+00 7.43576229e-01 7.29779825e-02
8.67486417e-01 -5.92280269e-01 -1.18725991e+00 8.94346118e-01
8.10468048e-02 4.54205155e-01 7.77503669e-01 4.60043818e-01
-5.50434709e-01 -1.61251083e-01 -6.94275379e-01 5.57485819e-01
1.12399197e+00 -6.86013401e-01 -4.42154437e-01 4.87199247e-01
6.76506996e-01 -1.21487103e-01 -8.76862705e-01 3.07982355e-01
3.84126365e-01 -7.95562685e-01 1.01094115e+00 -1.03357053e+00
1.52411506e-01 -1.81421340e-01 -3.87496687e-02 -1.41169298e+00
-8.47863436e-01 -4.58296865e-01 -6.52834117e-01 8.60994756e-01
-7.32385591e-02 -7.45582938e-01 9.02408183e-01 8.12682435e-02
-2.22759694e-01 -1.16518974e+00 -6.77444577e-01 -6.77145183e-01
2.99072057e-01 -1.73111737e-01 5.21770716e-01 1.15415287e+00
2.41025314e-01 7.27761745e-01 -3.54629427e-01 -2.74321306e-02
5.02857566e-01 3.06257308e-01 6.39792442e-01 -1.39378166e+00
-2.49332543e-02 -8.65677416e-01 -1.06863236e+00 -1.16615224e+00
5.59643030e-01 -1.45623183e+00 -5.36935925e-01 -2.25835443e+00
1.37746304e-01 -3.75588834e-01 -6.37676656e-01 5.05803049e-01
-1.26163468e-01 1.25212386e-01 -2.87516322e-02 1.07811555e-01
-6.55104816e-01 8.67669940e-01 1.28326356e+00 -4.66390520e-01
-5.20507544e-02 -8.04151371e-02 -1.02469158e+00 3.96214664e-01
7.09985495e-01 -2.71405846e-01 -7.85320461e-01 -6.08775735e-01
3.84412646e-01 -1.20664969e-01 1.59934059e-01 -8.29603076e-01
1.31385833e-01 5.67520373e-02 1.57421842e-01 -3.81137222e-01
6.89162612e-02 -8.00597429e-01 -1.85383052e-01 6.45543456e-01
-2.17574060e-01 2.30045587e-01 4.00300100e-02 9.52079535e-01
-4.71726835e-01 1.42173648e-01 7.97130764e-01 -9.12411287e-02
-8.66995513e-01 9.77938533e-01 -2.94879079e-01 -1.43777952e-01
1.01615775e+00 -1.68909520e-01 -1.05444700e-01 -5.40476203e-01
-4.67611194e-01 1.74059585e-01 2.99797475e-01 5.97539842e-01
7.48392463e-01 -1.56985593e+00 -6.47134840e-01 2.88598835e-01
3.67225081e-01 4.20937985e-02 2.83007085e-01 5.73376238e-01
-4.09797579e-01 5.29275835e-01 -2.16247812e-01 -3.43531132e-01
-8.75677407e-01 1.08058834e+00 2.68329948e-01 -3.72623950e-01
-1.08307600e+00 7.47514188e-01 4.76975739e-01 -3.64203602e-01
3.31700832e-01 -2.21974537e-01 -5.38645983e-01 -9.75985974e-02
4.70242292e-01 2.01123506e-01 1.80020750e-01 -7.11453438e-01
-4.71533895e-01 5.40937483e-01 -2.29122296e-01 8.18305314e-01
1.49558985e+00 1.33409396e-01 -2.73197889e-01 3.46934795e-01
1.52462947e+00 -2.95952588e-01 -6.58806145e-01 -3.65270644e-01
-2.74550058e-02 -2.63700336e-01 2.86698282e-01 -1.54733688e-01
-1.60634863e+00 1.18850720e+00 1.91906720e-01 4.25625801e-01
1.03723502e+00 1.65887251e-01 8.13423097e-01 4.49709088e-01
2.87306190e-01 -7.11450100e-01 2.33953595e-01 5.36731362e-01
8.89061451e-01 -1.20280266e+00 3.67834151e-01 -4.42242622e-01
-3.11274827e-01 1.20832825e+00 4.86003160e-01 -6.20893955e-01
8.76612484e-01 -1.95119549e-02 -2.59910673e-01 -6.77105248e-01
-4.78937328e-01 -2.60439634e-01 4.92760986e-01 8.84218812e-01
8.40794623e-01 4.53124158e-02 -1.98475257e-01 6.26646459e-01
-3.39329652e-02 -4.63000000e-01 2.02501774e-01 6.38051987e-01
-2.01826170e-01 -1.21587288e+00 2.50195682e-01 8.20771277e-01
3.23990919e-02 -5.32608926e-02 -5.03537118e-01 8.25776100e-01
-2.75518954e-01 7.70413697e-01 4.08097580e-02 -4.82311487e-01
3.42823833e-01 -4.92736608e-01 4.30179745e-01 -1.01543546e+00
-3.03616881e-01 -2.72781312e-01 -1.58709466e-01 -4.95807707e-01
-3.45659226e-01 -3.04882586e-01 -1.25406420e+00 -4.16171312e-01
-1.92785159e-01 2.54719406e-01 1.91131040e-01 6.69205964e-01
5.64489186e-01 7.58544743e-01 5.22184670e-01 -8.00145388e-01
-1.21580996e-01 -7.92507052e-01 -9.10153210e-01 2.61797756e-01
2.58436590e-01 -9.20497954e-01 -2.91347235e-01 -4.66699392e-01] | [7.103440761566162, 6.280629634857178] |
db8ed22f-6b37-46e3-80b8-5309f5f84859 | gensyn-a-multi-stage-framework-for-generating | 2212.05975 | null | https://arxiv.org/abs/2212.05975v1 | https://arxiv.org/pdf/2212.05975v1.pdf | GenSyn: A Multi-stage Framework for Generating Synthetic Microdata using Macro Data Sources | Individual-level data (microdata) that characterizes a population, is essential for studying many real-world problems. However, acquiring such data is not straightforward due to cost and privacy constraints, and access is often limited to aggregated data (macro data) sources. In this study, we examine synthetic data generation as a tool to extrapolate difficult-to-obtain high-resolution data by combining information from multiple easier-to-obtain lower-resolution data sources. In particular, we introduce a framework that uses a combination of univariate and multivariate frequency tables from a given target geographical location in combination with frequency tables from other auxiliary locations to generate synthetic microdata for individuals in the target location. Our method combines the estimation of a dependency graph and conditional probabilities from the target location with the use of a Gaussian copula to leverage the available information from the auxiliary locations. We perform extensive testing on two real-world datasets and demonstrate that our approach outperforms prior approaches in preserving the overall dependency structure of the data while also satisfying the constraints defined on the different variables. | ['Huzefa Rangwala', 'Sanmay Das', 'Siddhartha Sikdar', 'Angeela Acharya'] | 2022-12-08 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [-6.31648228e-02 8.06341022e-02 -2.36303672e-01 -4.08593684e-01
-1.07366025e+00 -4.98011053e-01 6.54480875e-01 6.52722895e-01
-2.10210890e-01 1.52043808e+00 4.72245783e-01 -1.10475667e-01
-3.57085735e-01 -1.40632010e+00 -7.68760920e-01 -5.56101382e-01
-9.22873691e-02 6.07787013e-01 -2.02370688e-01 -1.10887345e-02
-1.63111046e-01 4.95536268e-01 -1.35393274e+00 -9.30053294e-02
1.34675527e+00 5.21637678e-01 9.22085568e-02 2.82171845e-01
6.05136342e-02 1.79373175e-01 -7.53616571e-01 -7.55569041e-01
3.11553448e-01 -6.93961456e-02 -1.40842676e-01 -6.04211688e-02
1.10621423e-01 -2.20061824e-01 5.69131300e-02 9.42761540e-01
2.69934416e-01 1.26037419e-01 8.29489946e-01 -1.45402837e+00
-5.89079857e-01 5.50813556e-01 -8.50284815e-01 1.05743771e-02
6.83425486e-01 -1.15984239e-01 7.73577094e-01 -5.17681897e-01
5.07895052e-01 1.13981009e+00 7.46639729e-01 -2.95665383e-01
-1.86379158e+00 -4.62755173e-01 -8.65668058e-02 -4.68340963e-01
-1.82887840e+00 -3.56179386e-01 5.05657911e-01 -3.28277141e-01
3.72952700e-01 2.77397960e-01 4.90611196e-01 1.19163275e+00
1.68834940e-01 3.14669371e-01 9.96396601e-01 -1.91192195e-01
4.02979374e-01 3.61668527e-01 4.24641743e-02 7.71083981e-02
9.48697746e-01 -5.31838611e-02 -4.18056667e-01 -9.03765380e-01
4.91814166e-01 2.23544136e-01 -9.36342776e-02 -3.70007843e-01
-9.32181895e-01 7.47022867e-01 3.48784566e-01 3.89863923e-02
-7.11989939e-01 -1.48495033e-01 -2.13740319e-01 -9.95878130e-02
6.12786889e-01 1.14532925e-01 -3.94291282e-01 -1.37578860e-01
-1.00920391e+00 7.38768935e-01 9.95545447e-01 1.10599220e+00
1.04393411e+00 -1.39824584e-01 -1.81967810e-01 5.61238408e-01
1.77726224e-01 1.08123899e+00 -7.15418309e-02 -5.13600528e-01
9.96061325e-01 6.71400547e-01 6.68613315e-01 -1.29529953e+00
-2.97047377e-01 -1.66949496e-01 -1.07152319e+00 -3.21965396e-01
8.40183437e-01 -6.45308375e-01 -6.56150579e-01 1.97749913e+00
6.28677785e-01 1.16490856e-01 6.66793734e-02 3.66037637e-01
2.62977451e-01 6.45249307e-01 -1.40954144e-02 -2.03965291e-01
1.19270706e+00 8.35024267e-02 -6.13036454e-01 7.91391358e-02
3.46134692e-01 -1.93884552e-01 7.62166262e-01 5.36487028e-02
-8.01870167e-01 -3.49348992e-01 -6.06139600e-01 1.30520523e-01
-8.47829938e-01 -5.57485549e-03 5.82074583e-01 7.93612957e-01
-7.89394975e-01 1.83143869e-01 -7.34351575e-01 -3.75160217e-01
3.79599869e-01 2.77664363e-01 -5.96263230e-01 -1.21493123e-01
-1.30874550e+00 3.36223811e-01 2.02476010e-01 -3.42881419e-02
-1.29829258e-01 -1.28494644e+00 -1.16143227e+00 3.60214680e-01
4.66089189e-01 -6.45431638e-01 4.18977648e-01 -1.48736611e-01
-6.13894224e-01 2.08286449e-01 -2.17448801e-01 -3.84588778e-01
4.35250312e-01 5.40488251e-02 -4.32439268e-01 -1.89497381e-01
5.22776842e-01 1.94517851e-01 5.17397583e-01 -1.29133391e+00
-7.54241347e-01 -5.87298095e-01 -7.74423033e-02 -9.85510498e-02
-7.69908726e-02 -2.65282810e-01 -3.63703191e-01 -7.80357957e-01
-3.33691120e-01 -6.56439662e-01 -3.19408476e-01 -5.76978326e-01
-7.44372427e-01 2.71072865e-01 2.78524190e-01 -9.45954621e-01
1.51353705e+00 -1.99941504e+00 2.31891014e-02 7.80153632e-01
6.28014207e-02 -2.45283127e-01 1.83888748e-01 8.35099459e-01
3.07067841e-01 1.87599406e-01 -4.42479819e-01 -4.50865239e-01
7.20237643e-02 1.85583636e-01 -3.28387171e-01 2.88617164e-01
2.05993757e-01 7.93071926e-01 -7.58683443e-01 -4.20510501e-01
1.06673852e-01 5.82776964e-01 -6.19331002e-01 2.37287674e-02
-3.75786908e-02 2.46738210e-01 -4.79338229e-01 7.81252861e-01
9.26261842e-01 -5.01263477e-02 2.95210510e-01 1.54876202e-01
-2.54900344e-02 -2.87600309e-02 -1.49726617e+00 1.29077876e+00
-3.27386737e-01 6.71429485e-02 9.63284522e-02 -6.52780831e-01
9.04169202e-01 -8.30029100e-02 5.39937079e-01 -3.30749959e-01
-2.44063407e-01 4.56574410e-02 -2.95581102e-01 -1.41506791e-01
6.27621293e-01 3.06726247e-01 -7.21869648e-01 4.98043597e-01
-2.71952271e-01 5.40996827e-02 4.85008836e-01 1.06203079e-01
9.21698213e-01 -1.18788354e-01 5.62649906e-01 -2.99551904e-01
2.47884989e-01 4.91811112e-02 6.95425391e-01 8.81287336e-01
2.92427480e-01 6.11482739e-01 8.67095351e-01 -2.93138295e-01
-1.06922793e+00 -1.38883233e+00 -5.12567401e-01 4.81386453e-01
-2.92971194e-01 -3.63193780e-01 -7.21236944e-01 -5.69590688e-01
6.05315149e-01 8.11852276e-01 -9.39629197e-01 1.31447360e-01
-1.50121868e-01 -1.34776688e+00 4.71396536e-01 2.96958715e-01
3.52847159e-01 -4.99961406e-01 -5.26314795e-01 2.53256589e-01
-3.61478597e-01 -9.29550767e-01 -2.50792742e-01 -1.22701526e-01
-4.87590909e-01 -9.57359970e-01 -5.95108628e-01 -1.32960044e-02
8.43899608e-01 -9.95974243e-03 1.06420898e+00 -3.65759283e-01
-3.16516384e-02 9.90665108e-02 -2.41768986e-01 -4.37352121e-01
-4.22735400e-02 3.12837511e-01 1.59257546e-01 4.19390291e-01
6.96477711e-01 -8.83788049e-01 -1.28401771e-01 4.67108898e-02
-1.02242422e+00 -2.32553795e-01 4.36523914e-01 7.18378425e-01
6.14043891e-01 2.81345010e-01 9.01831865e-01 -8.83763194e-01
6.79111779e-01 -1.11431098e+00 -8.65143359e-01 3.98438394e-01
-2.92712033e-01 6.72171637e-02 5.70499003e-01 -3.65337908e-01
-1.12992036e+00 -4.41957302e-02 3.32799703e-01 1.42402917e-01
-3.58833522e-01 9.52141464e-01 -6.62561476e-01 4.78103518e-01
3.59861642e-01 8.78136978e-02 -7.21748695e-02 -5.76034188e-01
2.97951162e-01 8.77753615e-01 6.07987881e-01 -8.18911672e-01
9.93281603e-01 6.14977896e-01 2.52552152e-01 -8.71367395e-01
-6.35148406e-01 -6.48194328e-02 -8.98276031e-01 2.87505358e-01
4.79873061e-01 -9.86323535e-01 -5.58628559e-01 3.41140598e-01
-6.48222268e-01 -5.12516312e-02 -3.80950719e-01 5.24488807e-01
-3.01719487e-01 5.41676655e-02 5.26188798e-02 -9.70537901e-01
9.22259763e-02 -8.71161699e-01 1.19259155e+00 3.22403014e-01
-2.30988577e-01 -1.21749663e+00 1.93984479e-01 1.74050063e-01
3.99132967e-01 8.51645708e-01 8.38887215e-01 -6.81430340e-01
-6.63420439e-01 -5.19666374e-01 -2.92312086e-01 -1.87463015e-01
6.02343202e-01 5.00985235e-02 -6.12308919e-01 -2.44384840e-01
-3.64923537e-01 5.20360582e-02 3.77449006e-01 5.08553445e-01
9.35791790e-01 -4.63566303e-01 -4.44590688e-01 6.72118843e-01
1.48025489e+00 -1.59580976e-01 5.50649941e-01 1.66182928e-02
4.75214124e-01 8.90021324e-01 6.13460779e-01 8.58943999e-01
9.43729937e-01 7.00200498e-01 -3.91575322e-02 -2.40057204e-02
4.57270771e-01 -6.60176754e-01 -3.74377258e-02 -2.05111075e-02
6.93724230e-02 -2.70497173e-01 -9.27294075e-01 8.45556915e-01
-1.95312989e+00 -1.01217461e+00 -6.84604868e-02 2.73275328e+00
8.82510066e-01 -1.45397246e-01 4.45011318e-01 -1.86755463e-01
6.72504365e-01 -1.11251175e-01 -4.99018490e-01 1.98286042e-01
-2.21082255e-01 9.37041175e-03 7.47101486e-01 5.12066305e-01
-9.61424053e-01 3.78621697e-01 6.37792015e+00 6.74857557e-01
-5.16002059e-01 -2.42988914e-01 7.86561370e-01 -9.96429101e-02
-4.58389670e-01 -4.70654182e-02 -9.77525115e-01 8.19336891e-01
1.27244544e+00 -6.32191241e-01 4.10808831e-01 4.88299519e-01
3.20309728e-01 -4.42649692e-01 -1.00135672e+00 6.76357746e-01
-6.70061782e-02 -1.07812738e+00 3.34409624e-02 7.79401481e-01
7.11348832e-01 -3.63116533e-01 1.43907145e-01 6.91102669e-02
5.99938452e-01 -9.74602163e-01 3.64651680e-01 8.83122623e-01
7.71957815e-01 -1.22069097e+00 9.02037859e-01 4.95729476e-01
-1.22066176e+00 -7.88604766e-02 -4.06617165e-01 -9.83677655e-02
2.77844369e-01 9.01443362e-01 -7.55909562e-01 1.00199652e+00
7.28941441e-01 3.99404764e-01 -7.28677928e-01 9.09193397e-01
6.79189190e-02 4.15487736e-01 -8.53309631e-01 1.36268601e-01
-2.22013116e-01 -6.65811896e-01 2.57530063e-01 8.99734259e-01
7.41283298e-01 -2.03933284e-01 2.48891339e-02 1.15333092e+00
-1.51352748e-01 4.36075777e-03 -9.48518157e-01 2.15751469e-01
9.59232092e-01 1.09620476e+00 -2.68547654e-01 -1.38986096e-01
-5.85653663e-01 2.31494144e-01 4.53119546e-01 6.48920238e-01
-6.70393825e-01 -4.70800608e-01 7.62810051e-01 3.25506449e-01
2.55876064e-01 -2.35579774e-01 -3.15267295e-01 -1.24067438e+00
2.55482763e-01 -6.71382964e-01 4.64956522e-01 -6.28778338e-01
-1.42476833e+00 9.83608142e-02 6.45586491e-01 -9.85611558e-01
-6.54629946e-01 -1.33395791e-01 -3.02012593e-01 1.44705439e+00
-1.24876142e+00 -1.31595194e+00 -2.13670641e-01 6.21223211e-01
-2.41989315e-01 -6.79443479e-02 7.71937490e-01 1.86252862e-01
-6.87516809e-01 7.16723919e-01 4.72288430e-01 2.65940607e-01
4.74860013e-01 -1.37454545e+00 5.49767792e-01 9.33305264e-01
-5.09271398e-02 8.94772053e-01 6.30271196e-01 -1.13704324e+00
-1.22420418e+00 -1.18336833e+00 1.06212294e+00 -7.91297197e-01
6.66758776e-01 -7.51042545e-01 -8.38974714e-01 8.20104778e-01
-1.58810467e-01 -1.97862223e-01 9.41281199e-01 3.19711149e-01
-4.65396419e-02 -4.21721369e-01 -1.73542452e+00 5.36869287e-01
6.05604351e-01 -3.09633523e-01 -5.29057980e-01 -3.69195119e-02
5.66016555e-01 -9.40058902e-02 -1.23621798e+00 1.02841675e-01
4.62048709e-01 -8.61958802e-01 1.05770087e+00 -4.57002729e-01
1.57622889e-01 -3.84408265e-01 -4.32021618e-01 -1.55450511e+00
-7.25564957e-02 -5.08276105e-01 -3.70272659e-02 1.73851395e+00
7.03261316e-01 -1.03529215e+00 6.63686216e-01 1.27993393e+00
5.49275637e-01 -3.30901563e-01 -1.05129218e+00 -5.99773467e-01
-1.92525052e-02 -2.28063315e-01 1.43495631e+00 7.32768297e-01
5.52656874e-03 2.40268297e-02 -6.02803469e-01 6.35572910e-01
1.02687085e+00 1.82325587e-01 1.30687761e+00 -1.39558887e+00
-1.21362068e-01 3.09641898e-01 -2.70227820e-01 -4.60428298e-01
-2.34253369e-02 -2.82426029e-01 -3.90932381e-01 -1.36233902e+00
2.19972447e-01 -6.82860911e-01 1.60882875e-01 3.06856513e-01
-2.58594245e-01 9.87595841e-02 -1.07261255e-01 -6.20300844e-02
5.36591448e-02 4.03984278e-01 6.59229338e-01 1.02797128e-01
-5.31772494e-01 1.96502462e-01 -1.09139371e+00 4.98527676e-01
6.75799727e-01 -4.78844136e-01 -4.75368649e-01 -1.59204960e-01
4.18630123e-01 3.76465946e-01 2.69184470e-01 -9.60454404e-01
1.51972964e-01 -5.22964597e-01 6.01201653e-01 -9.90385711e-01
4.44509089e-01 -1.06982434e+00 7.32314825e-01 -1.27790630e-01
-4.17262167e-02 1.72836155e-01 1.00945212e-01 7.23117352e-01
-1.87238932e-01 4.60503139e-02 2.09407136e-01 1.24330193e-01
1.65793121e-01 2.69942850e-01 -1.75896138e-01 3.17277759e-02
1.04201603e+00 -2.46781055e-02 -3.94392222e-01 -4.77391601e-01
-5.24359763e-01 5.09934485e-01 8.03625047e-01 1.49950206e-01
3.20920020e-01 -1.37471783e+00 -9.48770046e-01 6.39949381e-01
1.46335125e-01 2.44852915e-01 1.00474663e-01 6.68403685e-01
-1.82888910e-01 4.49421793e-01 -1.85072079e-01 -3.04812223e-01
-8.19888592e-01 7.50419199e-01 -3.89857851e-02 -4.22438860e-01
-3.46432686e-01 9.58585814e-02 -5.61848469e-02 -6.23005092e-01
-1.85809076e-01 -4.80034560e-01 -2.84352992e-03 3.85842532e-01
6.40182555e-01 5.52108049e-01 -1.99047089e-01 -6.89721584e-01
-1.37545362e-01 1.59799144e-01 1.02834769e-01 -3.24008226e-01
1.49770260e+00 -5.06843328e-01 3.25987302e-02 4.38279033e-01
7.04798877e-01 6.59379065e-01 -1.15948617e+00 -2.59979516e-01
1.10425189e-01 -9.32000458e-01 -5.87242842e-01 -6.06769919e-01
-7.23276973e-01 3.76141936e-01 -9.62154418e-02 6.57292366e-01
1.04047585e+00 -1.48644760e-01 3.78209800e-01 2.50781298e-01
7.08820403e-01 -8.90037954e-01 -7.65356898e-01 -5.64476438e-02
7.40222812e-01 -1.13107717e+00 8.34866539e-02 -4.56430703e-01
-3.82503450e-01 5.43848515e-01 3.03540617e-01 8.60017985e-02
9.28512931e-01 2.71293670e-01 -2.68003196e-01 1.31543145e-01
-6.70478582e-01 -2.83965230e-01 1.12230882e-01 1.05702698e+00
-5.72844930e-02 2.39023149e-01 -1.08248487e-01 8.82357359e-01
-4.95642394e-01 -1.57604724e-01 9.36041176e-01 7.01681256e-01
6.83037043e-02 -1.24153805e+00 -8.17684948e-01 9.58441257e-01
-5.60460865e-01 -1.80686921e-01 -2.88136989e-01 1.05638671e+00
4.64575812e-02 9.77153599e-01 4.23533857e-01 2.05195192e-02
3.54525685e-01 -2.57904008e-02 -9.65263024e-02 -4.28832322e-01
-1.72196269e-01 -6.79823384e-02 1.01095147e-01 -2.17620417e-01
-3.58623028e-01 -1.19284105e+00 -7.04684913e-01 -7.64698505e-01
1.53446585e-01 3.71316314e-01 4.90386575e-01 6.52238369e-01
6.99310720e-01 1.53494909e-01 6.65197611e-01 -7.98174739e-01
-4.24439847e-01 -9.88889396e-01 -9.28677976e-01 4.86881077e-01
4.95841473e-01 -7.40322948e-01 -3.32941175e-01 -1.15507275e-01] | [6.653600692749023, 4.241783618927002] |
b7c62b56-77a5-4419-81cc-5cb9219a23af | histred-a-historical-document-level-relation | 2307.04285 | null | https://arxiv.org/abs/2307.04285v1 | https://arxiv.org/pdf/2307.04285v1.pdf | HistRED: A Historical Document-Level Relation Extraction Dataset | Despite the extensive applications of relation extraction (RE) tasks in various domains, little has been explored in the historical context, which contains promising data across hundreds and thousands of years. To promote the historical RE research, we present HistRED constructed from Yeonhaengnok. Yeonhaengnok is a collection of records originally written in Hanja, the classical Chinese writing, which has later been translated into Korean. HistRED provides bilingual annotations such that RE can be performed on Korean and Hanja texts. In addition, HistRED supports various self-contained subtexts with different lengths, from a sentence level to a document level, supporting diverse context settings for researchers to evaluate the robustness of their RE models. To demonstrate the usefulness of our dataset, we propose a bilingual RE model that leverages both Korean and Hanja contexts to predict relations between entities. Our model outperforms monolingual baselines on HistRED, showing that employing multiple language contexts supplements the RE predictions. The dataset is publicly available at: https://huggingface.co/datasets/Soyoung/HistRED under CC BY-NC-ND 4.0 license. | ['Jaegul Choo', 'Youngwoo Cho', 'Minseok Choi', 'Soyoung Yang'] | 2023-07-10 | null | null | null | null | ['document-level-relation-extraction', 'relation-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [-3.91856909e-01 -2.50187796e-02 -7.09446549e-01 -2.77691036e-01
-8.68900537e-01 -8.49966049e-01 8.86013329e-01 1.38701499e-01
-3.19335580e-01 1.00642335e+00 7.29076028e-01 -4.74021256e-01
1.11624740e-01 -7.50686407e-01 -5.43386459e-01 -1.09138474e-01
1.41333088e-01 2.84502894e-01 -2.49113724e-01 -4.01129514e-01
-3.89598384e-02 2.94456780e-01 -6.09025836e-01 2.58535057e-01
1.00301015e+00 3.30350578e-01 1.52425066e-01 3.88507426e-01
2.86459982e-01 5.50401688e-01 -4.12371457e-01 -9.13466215e-01
2.59630293e-01 -4.29325432e-01 -1.02503955e+00 -4.46579576e-01
4.76716459e-01 -7.77235031e-02 -6.26354814e-01 8.08315814e-01
3.94258022e-01 -7.74407014e-02 3.36364627e-01 -1.04607356e+00
-1.38218880e+00 1.42789829e+00 -2.77910739e-01 3.52829397e-01
4.26488698e-01 -2.50202954e-01 1.51062381e+00 -1.09396946e+00
1.27867603e+00 8.02937031e-01 7.28263915e-01 3.22477758e-01
-9.62340891e-01 -9.87120152e-01 1.00805245e-01 3.71567816e-01
-1.45760465e+00 -5.80741942e-01 5.80819011e-01 -3.35152060e-01
1.21329820e+00 1.73693597e-01 6.43105149e-01 1.39043772e+00
1.17749192e-01 4.77136016e-01 1.09223008e+00 -5.41193247e-01
-3.45917493e-01 -1.24636134e-02 1.73516855e-01 6.69979513e-01
3.31426084e-01 -2.82305777e-01 -8.96610618e-01 -6.06072843e-02
4.35425460e-01 -9.33055058e-02 -3.44883323e-01 4.10683572e-01
-1.46045077e+00 6.18671596e-01 3.16145599e-01 5.77494681e-01
3.99242304e-02 -2.14646429e-01 3.96644711e-01 3.83578926e-01
6.13333523e-01 6.16510510e-01 -7.19995856e-01 -1.46898612e-01
-7.81510293e-01 2.50376821e-01 8.49098444e-01 1.43193173e+00
5.15828371e-01 -4.94416535e-01 7.66957179e-02 1.03640437e+00
-1.47047594e-01 5.08776486e-01 5.21335483e-01 -6.19782388e-01
9.96454239e-01 6.55818164e-01 -3.16891611e-01 -1.03326702e+00
-2.98288256e-01 -4.14735377e-01 -8.95015955e-01 -1.00895667e+00
1.15174219e-01 -1.90862581e-01 -4.67214793e-01 1.70516312e+00
2.34109610e-01 9.47229713e-02 3.47100079e-01 6.40946507e-01
1.15230763e+00 6.34791732e-01 -6.75311685e-02 -1.28540769e-01
1.40256810e+00 -1.11191213e+00 -6.95533156e-01 -3.09082806e-01
8.98220718e-01 -8.58549535e-01 9.62278903e-01 1.01837926e-01
-8.68636191e-01 -2.43983865e-01 -1.01165783e+00 -5.73231518e-01
-5.78415155e-01 3.84458691e-01 8.15573275e-01 2.45443612e-01
-7.44755566e-01 4.48544174e-01 -8.34112823e-01 -7.02516615e-01
2.79068142e-01 -4.51192595e-02 -6.99138284e-01 -1.23474143e-01
-1.64235198e+00 1.02748501e+00 4.91434366e-01 3.57180417e-01
-5.41314900e-01 -8.20697010e-01 -9.31297123e-01 -2.60466039e-01
5.72535396e-01 -4.80917513e-01 1.11815989e+00 -2.79846787e-01
-1.01192510e+00 1.03556168e+00 -2.74239182e-01 -3.12467039e-01
4.01604146e-01 -5.90025306e-01 -8.45487714e-01 -9.03059617e-02
3.88157517e-01 6.74222857e-02 -1.11356294e-02 -8.39864850e-01
-3.76509011e-01 -3.58080655e-01 -1.38530284e-01 -1.43979788e-02
-4.08041388e-01 3.69973660e-01 -6.59094155e-01 -1.10057938e+00
6.85228081e-03 -1.16848063e+00 -5.51750697e-02 -8.17852497e-01
-8.95917058e-01 -2.52832979e-01 4.70114589e-01 -1.08161056e+00
1.76368988e+00 -1.78362858e+00 7.77700022e-02 2.26257965e-02
1.63562462e-01 1.62114706e-02 -1.41915172e-01 1.01827359e+00
-1.20278597e-01 5.52356601e-01 -1.68426931e-01 -1.68964326e-01
-2.26749644e-01 7.19747171e-02 -4.57885593e-01 2.98176497e-01
1.55332938e-01 1.15874910e+00 -1.04385185e+00 -4.35776144e-01
-5.53331137e-01 1.68323845e-01 -1.97200254e-01 -5.97686656e-02
4.50789556e-02 3.44737411e-01 -5.08514225e-01 9.61857617e-01
3.30315560e-01 -2.47464791e-01 8.34017396e-01 -2.34714612e-01
-3.92062157e-01 8.18911612e-01 -4.95168954e-01 1.71869159e+00
-3.42992663e-01 7.96133876e-01 -3.38472664e-01 -4.92765665e-01
8.90039146e-01 4.24024493e-01 3.27074528e-01 -6.21710479e-01
-1.24923810e-01 4.96256262e-01 -8.35673660e-02 -3.96715343e-01
1.01310289e+00 9.04633924e-02 -3.92649680e-01 4.51747417e-01
3.96906398e-02 -3.36418711e-02 5.00131190e-01 6.14147246e-01
1.15277672e+00 3.42837870e-01 7.36440063e-01 -1.27115816e-01
2.57088006e-01 2.92217165e-01 1.11232579e+00 3.59616935e-01
7.58492872e-02 3.98683518e-01 4.54860538e-01 -1.06998511e-01
-1.09708595e+00 -7.23321319e-01 -4.60599989e-01 8.05521965e-01
9.87633094e-02 -1.09742367e+00 -4.29171443e-01 -9.70523536e-01
3.96059155e-02 6.27242088e-01 -6.68141365e-01 2.26811990e-01
-8.82818758e-01 -7.70865023e-01 9.08019364e-01 4.57405329e-01
5.51926613e-01 -8.63434255e-01 1.03951566e-01 1.20288014e-01
-4.28432435e-01 -1.56997693e+00 -6.90943837e-01 -9.47051570e-02
-6.07671380e-01 -1.12671971e+00 -4.43756551e-01 -7.82826066e-01
2.92930216e-01 2.24838816e-02 1.37965572e+00 -1.43940881e-01
-1.75832272e-01 2.60161340e-01 -6.00311756e-01 -1.59831017e-01
-3.40013325e-01 6.68561995e-01 2.04969928e-01 -6.02867067e-01
5.72653592e-01 -6.94143057e-01 -2.84352750e-01 1.91457290e-02
-5.00812709e-01 2.23061100e-01 4.50257629e-01 8.20785701e-01
3.95090550e-01 -1.62732333e-01 8.04660439e-01 -1.30928278e+00
5.03129601e-01 -9.04523790e-01 -2.57097155e-01 7.21417785e-01
-7.88400590e-01 -1.31875321e-01 6.21666372e-01 -3.26127917e-01
-1.05961859e+00 -5.11154115e-01 1.41497940e-01 1.50781900e-01
1.92541003e-01 8.31599414e-01 -2.77594000e-01 4.75778550e-01
3.62405807e-01 -6.96795881e-02 -7.04137802e-01 -6.04592562e-01
5.99392772e-01 8.42045069e-01 7.15662718e-01 -7.86648870e-01
7.64943480e-01 -4.17595394e-02 -1.81582540e-01 -6.73358321e-01
-1.05488193e+00 -1.48057863e-01 -9.32019889e-01 2.45921873e-02
8.04859579e-01 -1.11162758e+00 -3.07366401e-01 3.58363748e-01
-1.09673345e+00 -3.53332311e-01 1.38033956e-01 5.05596042e-01
2.13054359e-01 1.91281512e-01 -1.16772091e+00 -3.12995791e-01
-3.88686419e-01 -6.82969213e-01 5.56715131e-01 2.27277189e-01
-4.42335606e-01 -1.01754189e+00 2.84164548e-01 5.38854897e-01
-1.84897959e-01 1.73478544e-01 1.10975659e+00 -8.87983799e-01
-4.43678766e-01 9.34444219e-02 -8.38181376e-02 -1.14883415e-01
3.09748024e-01 2.27880076e-01 -6.67482078e-01 -5.96182011e-02
-4.51860398e-01 -3.03615928e-01 7.16292858e-01 -2.90190995e-01
6.66040421e-01 -5.18207490e-01 -5.13810992e-01 5.93817770e-01
1.24882126e+00 1.31317019e-01 4.63178009e-01 7.92628586e-01
9.92855251e-01 5.87053955e-01 6.13006115e-01 2.24147812e-01
9.77084458e-01 6.21800900e-01 -2.46907294e-01 6.63729385e-02
-1.67535737e-01 -6.19596183e-01 4.09736723e-01 1.42763531e+00
-1.81317419e-01 -2.28483841e-01 -1.27224493e+00 9.75555003e-01
-1.69907534e+00 -9.57172096e-01 -1.66975215e-01 1.81525350e+00
1.53522038e+00 -1.76623136e-01 -1.39793888e-01 -4.58882898e-01
6.72599196e-01 1.60471722e-01 -3.79341573e-01 -1.71489656e-01
-5.42097509e-01 2.50463694e-01 4.80629086e-01 3.32960606e-01
-9.30990279e-01 1.22679913e+00 5.48246479e+00 6.38718724e-01
-1.02062774e+00 1.29220739e-01 4.60883826e-01 -8.74904394e-02
-5.52656591e-01 5.30500472e-01 -1.12474251e+00 4.07372624e-01
8.80277336e-01 -5.41736424e-01 2.99460471e-01 5.79031646e-01
-5.51725850e-02 1.07017308e-01 -1.23564315e+00 7.01709211e-01
2.43796453e-01 -1.37559974e+00 -2.07833797e-01 5.32017015e-02
8.32937539e-01 2.05441147e-01 -1.03093833e-01 4.72495794e-01
5.33764839e-01 -1.03246260e+00 6.18013263e-01 4.54960167e-01
9.16750669e-01 -5.89850128e-01 5.25465727e-01 2.08219469e-01
-1.19247878e+00 2.14867160e-01 -1.20397553e-01 1.22989908e-01
2.27986649e-01 7.39218056e-01 -1.00871539e+00 1.09030735e+00
7.20685899e-01 1.22539496e+00 -8.56147647e-01 3.27903479e-01
-5.66505194e-01 8.96440387e-01 -5.80684468e-02 2.04708949e-01
-2.24040952e-02 -4.37004477e-01 5.58312595e-01 1.58647454e+00
5.26630878e-01 4.20931607e-01 8.64691958e-02 7.39877939e-01
-6.26936436e-01 3.32955182e-01 -6.61104918e-01 -4.66207355e-01
9.43974733e-01 1.39617264e+00 -4.36672449e-01 -1.30210727e-01
-5.12563884e-01 9.74399924e-01 7.14641213e-01 3.72844994e-01
-6.13090813e-01 -4.04410511e-01 3.45839292e-01 3.55404951e-02
1.17091620e-02 -4.69691247e-01 -1.68424204e-01 -1.67651105e+00
1.54904753e-01 -9.00265276e-01 5.80234230e-01 -6.38074279e-01
-1.41552484e+00 8.24411333e-01 7.69629031e-02 -1.00133109e+00
-1.47558495e-01 -4.03524071e-01 -3.02136362e-01 8.42488229e-01
-1.17220891e+00 -1.70695317e+00 6.89503774e-02 1.73941225e-01
3.80056083e-01 -3.30317885e-01 7.97431886e-01 4.64930594e-01
-1.08529329e+00 7.73220599e-01 6.30802736e-02 6.51892483e-01
1.22250080e+00 -1.08704293e+00 5.42481303e-01 1.12842262e+00
3.97657007e-01 1.28657532e+00 2.54059076e-01 -1.12856734e+00
-1.18190038e+00 -9.58475947e-01 1.77859807e+00 -6.54031336e-01
1.08910048e+00 -5.28742135e-01 -9.48151946e-01 1.25300908e+00
4.85595763e-01 -1.93573982e-01 1.00373518e+00 6.68683827e-01
-6.51088893e-01 3.27121206e-02 -6.39716506e-01 9.20390546e-01
1.36351728e+00 -8.68381739e-01 -7.49717474e-01 2.32638732e-01
6.47996485e-01 -4.37675387e-01 -1.40103447e+00 4.01210755e-01
5.24258792e-01 -4.95973647e-01 5.92968881e-01 -7.88505971e-01
8.27927291e-01 -1.05069689e-01 -2.57167757e-01 -1.22307384e+00
-2.65952021e-01 -6.07032299e-01 -2.17634797e-01 1.91081429e+00
7.81384051e-01 -5.49444079e-01 2.68538266e-01 5.69170177e-01
-2.26659924e-01 -9.44494367e-01 -6.30336583e-01 -6.91477060e-01
3.43335330e-01 -4.42830205e-01 7.73427486e-01 1.63430738e+00
2.30082899e-01 7.38720357e-01 -3.69066983e-01 8.84600282e-02
1.18267648e-01 4.36063796e-01 5.87595940e-01 -8.64989698e-01
-2.03864858e-01 -3.41006994e-01 3.64771765e-03 -8.34021926e-01
5.18869758e-01 -1.35473394e+00 -3.37967545e-01 -1.35025394e+00
4.98115331e-01 -7.90279090e-01 -3.66621390e-02 9.21571910e-01
-4.51615840e-01 2.53638268e-01 1.28326967e-01 5.95456481e-01
-3.90053123e-01 4.89406049e-01 9.75410402e-01 9.29854438e-02
-1.42442822e-01 -4.38217133e-01 -9.54184353e-01 5.10477066e-01
7.06082702e-01 -5.76586723e-01 -4.42464575e-02 -4.68692422e-01
5.45832038e-01 -1.38284236e-01 6.47427514e-02 -3.60842705e-01
2.51947612e-01 -3.54881734e-01 2.80641943e-01 -4.93336678e-01
-5.88251837e-02 -2.59140372e-01 2.96697140e-01 -1.06714562e-01
-5.12476325e-01 4.89515007e-01 3.44977193e-02 2.11272180e-01
-2.31580734e-01 -3.75697240e-02 1.39639840e-01 -1.01580881e-02
-6.91217005e-01 1.52245924e-01 9.29448530e-02 4.31432933e-01
7.18173027e-01 1.64419129e-01 -7.44086623e-01 -1.05867438e-01
-4.48825538e-01 3.90388876e-01 4.48469460e-01 5.78226030e-01
2.49620840e-01 -1.37980545e+00 -1.05582118e+00 -1.01126924e-01
3.42576861e-01 -9.97006521e-02 1.26760617e-01 7.76130855e-01
-3.30093443e-01 4.65643406e-01 2.47630447e-01 1.27017468e-01
-1.24779189e+00 3.59360576e-01 -6.79343492e-02 -5.33042371e-01
-7.66517580e-01 6.25365555e-01 -2.40072146e-01 -7.36408293e-01
-3.53241235e-01 -1.33401200e-01 -2.60955870e-01 8.36983472e-02
2.20304772e-01 5.16655922e-01 2.51153260e-02 -6.74137056e-01
-4.59997416e-01 2.69240320e-01 -4.15517420e-01 -3.27571034e-01
1.54058146e+00 -2.30078653e-01 -5.91049969e-01 5.64607263e-01
9.43009734e-01 1.06898189e+00 -4.78089035e-01 -3.55199665e-01
5.37384450e-01 -2.39789411e-01 -2.60114521e-01 -1.12520516e+00
-8.72453988e-01 2.80798554e-01 -3.15890133e-01 -1.97359622e-01
1.01063049e+00 2.63247192e-01 1.06881011e+00 4.09886926e-01
2.93323547e-01 -9.53576744e-01 -4.19902444e-01 8.37567985e-01
9.35631394e-01 -1.11983371e+00 3.69387776e-01 -5.31844616e-01
-9.52625871e-01 8.57812107e-01 5.59711158e-01 2.16471359e-01
5.88387787e-01 3.55106592e-01 1.49701655e-01 -6.73377961e-02
-8.06676924e-01 -1.65733993e-02 4.31428254e-01 2.34068722e-01
8.90760183e-01 1.98291153e-01 -6.59162223e-01 1.20042586e+00
-8.15397024e-01 -1.66810185e-01 5.00588655e-01 8.82455945e-01
4.26780730e-01 -1.64660251e+00 -1.06401201e-02 3.10683668e-01
-7.92685747e-01 -7.16482699e-01 -7.22086906e-01 9.34269130e-01
2.75111556e-01 8.90353799e-01 -2.59922415e-01 -4.49548483e-01
1.13848202e-01 2.03644927e-03 3.52167517e-01 -8.04220736e-01
-8.96198392e-01 -1.46973521e-01 6.19536161e-01 -1.53976083e-01
-3.74856740e-01 -8.04895759e-01 -1.19022441e+00 -4.55813676e-01
-2.03911573e-01 4.52338845e-01 3.09193045e-01 8.65376353e-01
5.04268527e-01 1.94874346e-01 4.20566171e-01 -9.31973830e-02
9.88992229e-02 -1.16517627e+00 -6.15062773e-01 1.24205776e-01
3.52704115e-02 -3.14135164e-01 -1.62866116e-01 1.43435776e-01] | [9.689759254455566, 8.982097625732422] |
befcfa6c-ce9c-435f-9078-dc42b7a37b2c | verifying-results-of-the-ibm-qiskit-quantum | 2009.02376 | null | https://arxiv.org/abs/2009.02376v1 | https://arxiv.org/pdf/2009.02376v1.pdf | Verifying Results of the IBM Qiskit Quantum Circuit Compilation Flow | Realizing a conceptual quantum algorithm on an actual physical device necessitates the algorithm's quantum circuit description to undergo certain transformations in order to adhere to all constraints imposed by the hardware. In this regard, the individual high-level circuit components are first synthesized to the supported low-level gate-set of the quantum computer, before being mapped to the target's architecture---utilizing several optimizations in order to improve the compilation result. Specialized tools for this complex task exist, e.g., IBM's Qiskit, Google's Cirq, Microsoft's QDK, or Rigetti's Forest. However, to date, the circuits resulting from these tools are hardly verified, which is mainly due to the immense complexity of checking if two quantum circuits indeed realize the same functionality. In this paper, we propose an efficient scheme for quantum circuit equivalence checking---specialized for verifying results of the IBM Qiskit quantum circuit compilation flow. To this end, we combine characteristics unique to quantum computing, e.g., its inherent reversibility, and certain knowledge about the compilation flow into a dedicated equivalence checking strategy. Experimental evaluations confirm that the proposed scheme allows to verify even large circuit instances with tens of thousands of operations within seconds or even less, whereas state-of-the-art techniques frequently time-out or require substantially more runtime. A corresponding open source implementation of the proposed method is publicly available at https://github.com/iic-jku/qcec. | ['Lukas Burgholzer', 'Robert Wille', 'Rudy Raymond'] | 2020-09-04 | null | null | null | null | ['quantum-circuit-equivalence-checking'] | ['methodology'] | [ 3.66421252e-01 -6.86850958e-03 1.18607014e-01 -7.72029832e-02
-7.64716387e-01 -9.58685637e-01 3.28668654e-01 4.17072147e-01
-5.02301678e-02 7.89619327e-01 -6.91168427e-01 -9.17846084e-01
3.22076119e-02 -1.23989081e+00 -9.39067483e-01 -6.69691443e-01
7.33799636e-02 3.79445672e-01 1.48091182e-01 -4.77990627e-01
7.05118477e-01 3.91899914e-01 -1.45417047e+00 1.27205566e-01
8.44908655e-01 8.43801141e-01 -7.48802274e-02 9.14996684e-01
3.94652858e-02 4.69018877e-01 -2.61566162e-01 -6.09655559e-01
4.68153715e-01 -8.35137308e-01 -8.02877843e-01 -5.11326492e-01
8.91692787e-02 -2.41536275e-01 -5.88274896e-01 1.75930393e+00
5.36638312e-02 -2.62890637e-01 2.13081807e-01 -1.27297556e+00
-2.04615012e-01 7.66029596e-01 2.28254646e-01 -1.97202578e-01
3.87374073e-01 4.28721279e-01 1.20003378e+00 -1.13163047e-01
6.83093309e-01 6.47621453e-01 2.88598955e-01 4.47230637e-01
-1.43257582e+00 -7.47661829e-01 -7.83956468e-01 3.31216812e-01
-1.75780714e+00 -4.94462162e-01 6.12043202e-01 -6.10672608e-02
1.06519568e+00 3.18461001e-01 5.94585061e-01 5.02153039e-01
7.68593073e-01 -1.51739433e-01 1.30833817e+00 -6.12269878e-01
5.04371166e-01 2.51567304e-01 3.10916334e-01 8.95982146e-01
6.91936672e-01 2.88603812e-01 -1.15713157e-01 -2.60282934e-01
4.22941029e-01 -3.98216158e-01 -7.41040111e-02 -3.36314768e-01
-1.13197243e+00 4.98744726e-01 3.06430578e-01 6.38633132e-01
2.27894653e-02 5.33428252e-01 5.67956030e-01 6.74100995e-01
-3.32493126e-01 5.53100646e-01 -2.26394773e-01 -2.93397248e-01
-6.32145047e-01 2.17402324e-01 8.92151892e-01 1.10657609e+00
1.18179452e+00 -3.31545502e-01 3.50053281e-01 -4.26917732e-01
-1.07523091e-01 3.59340727e-01 -4.13199998e-02 -6.35714352e-01
2.58938015e-01 5.22652149e-01 -3.31699289e-02 -6.43956602e-01
-3.53473157e-01 -2.64946073e-01 -9.17860925e-01 1.05852708e-01
3.05971414e-01 2.56182641e-01 -2.46870086e-01 1.59810877e+00
2.03193024e-01 2.02827230e-01 4.75342304e-01 7.48714924e-01
2.71818191e-01 7.04074979e-01 -4.24562812e-01 -2.30115518e-01
1.48531294e+00 -4.80554253e-01 -5.86076140e-01 2.52220958e-01
1.17953265e+00 -5.82138121e-01 7.79430330e-01 4.35214758e-01
-1.04372454e+00 -5.24823844e-01 -1.57341552e+00 -1.72735527e-02
-3.67371082e-01 7.34316334e-02 8.63790333e-01 1.06303203e+00
-1.03485048e+00 9.51669097e-01 -7.07525671e-01 -5.37136570e-02
9.16216075e-02 7.71044374e-01 -4.31684434e-01 -4.44932692e-02
-1.18338668e+00 7.75110245e-01 7.90295601e-01 1.38799995e-01
-7.10877061e-01 -4.25011665e-01 -4.85169351e-01 2.72987753e-01
2.95597881e-01 -7.39354014e-01 1.24918795e+00 -5.52630723e-01
-1.87615848e+00 8.20346296e-01 -1.00492202e-02 -6.55502617e-01
2.21865296e-01 6.59573495e-01 -6.65794134e-01 1.51174888e-01
-9.19581577e-02 -1.74723491e-02 3.86373997e-01 -6.61563814e-01
-5.17634571e-01 -4.16555911e-01 6.59154296e-01 -4.50684547e-01
1.32181585e-01 -2.32103899e-01 -1.81640074e-01 2.32940748e-01
1.84367627e-01 -1.29449785e+00 -3.12233627e-01 -4.88513112e-01
-6.20921493e-01 2.66359281e-02 2.73797333e-01 3.53854187e-02
1.09747374e+00 -2.17948270e+00 1.40212834e-01 5.40616751e-01
8.61202627e-02 2.20961437e-01 1.49601609e-01 6.92361712e-01
-4.44748029e-02 -8.16434920e-02 -1.89131305e-01 1.17955990e-01
2.68979460e-01 2.81327236e-02 -2.52962619e-01 8.08036268e-01
-2.15797238e-02 9.84389246e-01 -8.37230146e-01 -3.12672406e-01
4.33487445e-01 -4.62107435e-02 -8.05864990e-01 -2.34280452e-01
-2.96293437e-01 6.40153706e-01 -4.33063149e-01 4.27228361e-01
8.65670741e-01 -3.10583502e-01 6.72761679e-01 -2.43261755e-01
-3.91191155e-01 6.67627811e-01 -1.28239834e+00 1.89280200e+00
-4.69021797e-01 4.32532370e-01 -2.00561166e-01 -8.77208352e-01
6.50482059e-01 2.11549431e-01 -6.25856668e-02 -8.58950257e-01
5.95436156e-01 8.34147155e-01 3.94043714e-01 -1.42715732e-02
7.16845036e-01 -3.35114658e-01 -4.10342276e-01 6.00317717e-01
-3.85559872e-02 -6.86198771e-01 4.40379113e-01 2.42003307e-01
1.29661202e+00 -6.79986402e-02 5.37760854e-01 -4.69189227e-01
9.70950127e-01 2.24533275e-01 3.85839909e-01 8.32852125e-01
-1.51361257e-01 7.86824245e-03 6.70885682e-01 -1.83085904e-01
-1.20045948e+00 -8.85233819e-01 -2.63744205e-01 1.91500813e-01
5.39750397e-01 -8.44052374e-01 -1.01108873e+00 -2.20148116e-01
-3.66805136e-01 8.42845857e-01 -1.46688297e-01 -5.33983886e-01
-4.10227299e-01 -4.30226207e-01 8.86970520e-01 5.52565977e-02
5.50979137e-01 -6.99531138e-01 -6.46510482e-01 3.05692822e-01
2.03399017e-01 -1.19914448e+00 8.50027874e-02 2.89480716e-01
-8.09815228e-01 -1.09689713e+00 2.45280087e-01 -4.52693999e-01
8.52175832e-01 -3.65241915e-02 7.69149840e-01 2.32628837e-01
-4.39232618e-01 -2.72259742e-01 -2.58215338e-01 1.78594485e-01
-1.02874076e+00 -6.41715666e-03 2.21400321e-01 -2.14922845e-01
1.25784382e-01 -5.55474937e-01 -4.31057215e-01 -1.91068705e-02
-9.93993461e-01 6.09974675e-02 5.18349826e-01 6.98866904e-01
4.55975622e-01 7.44281530e-01 -4.33715135e-02 -9.73361433e-01
1.43066704e-01 1.23869449e-01 -1.45965958e+00 2.21426859e-01
-4.93645370e-01 3.93993884e-01 1.20121324e+00 -5.32712303e-02
-4.26925451e-01 1.04059711e-01 -9.53117833e-02 -1.49332406e-02
3.16028930e-02 2.50898719e-01 -3.25246036e-01 -6.04824126e-01
6.03564084e-01 3.48991454e-01 -4.13803995e-01 2.56784260e-01
3.24202597e-01 5.71955919e-01 6.79123700e-01 -7.58811593e-01
1.14674354e+00 4.67868149e-01 7.27645457e-01 -3.62309903e-01
-3.76201779e-01 7.05499947e-03 -4.80261624e-01 1.50420487e-01
4.77193326e-01 -5.62698960e-01 -1.36381674e+00 4.03651088e-01
-1.28018367e+00 -1.92157269e-01 -9.55264419e-02 4.98238593e-01
-5.25904059e-01 4.14757907e-01 -7.32795894e-01 -8.59755993e-01
-3.22613120e-01 -1.57436144e+00 8.66452992e-01 2.51983553e-01
9.92612988e-02 -6.52866423e-01 6.10971712e-02 7.21365809e-02
2.45262519e-01 -8.52806643e-02 1.08310187e+00 -2.99221992e-01
-1.06089640e+00 -4.18709844e-01 -2.59628296e-01 2.96762645e-01
-5.12839630e-02 -3.29459347e-02 -8.86797011e-01 -4.19530928e-01
-2.73998939e-02 -1.35850266e-01 2.44966090e-01 -2.06683397e-01
1.06054044e+00 6.33253753e-02 -1.19463488e-01 5.36179364e-01
1.84629965e+00 1.29811019e-01 9.13710237e-01 1.58442527e-01
1.93473250e-01 1.11167401e-01 6.12433374e-01 3.24674547e-01
8.62445235e-02 7.51903892e-01 4.17329311e-01 6.07650936e-01
3.62420261e-01 -2.24025771e-01 3.76761854e-01 1.02890801e+00
-1.12127863e-01 1.89093843e-01 -5.92510104e-01 -1.38324931e-01
-1.46663022e+00 -8.96628320e-01 -5.08186877e-01 2.58557487e+00
6.23888075e-01 5.03916442e-01 -4.75183666e-01 5.55484056e-01
5.87643445e-01 -3.20210814e-01 -1.11071177e-01 -7.12551951e-01
1.25225812e-01 7.27077723e-01 8.53128135e-01 4.89879817e-01
-6.39586031e-01 1.06460309e+00 5.01883459e+00 9.65860248e-01
-1.19915426e+00 2.38503665e-01 8.26808717e-03 4.23609644e-01
-4.31199402e-01 1.02913904e+00 -5.66948116e-01 3.95830363e-01
1.34718299e+00 -4.45226192e-01 8.51089656e-01 5.39584756e-01
-1.10392049e-01 -2.57185310e-01 -1.28159213e+00 1.05840456e+00
-3.81822109e-01 -1.48823726e+00 -3.54931951e-01 2.26099834e-01
5.21465778e-01 -2.68918216e-01 -1.62808567e-01 4.58020836e-01
-1.16724364e-01 -6.94303751e-01 9.40286100e-01 2.30429366e-01
1.06527972e+00 -8.88950467e-01 8.81800056e-01 2.67514557e-01
-1.18864048e+00 -3.44746909e-03 -4.88987714e-01 -3.41766924e-01
-3.39968652e-02 6.86456800e-01 -6.65631711e-01 1.01799488e+00
1.79946512e-01 9.40693319e-02 -3.53207141e-01 6.21598423e-01
-2.43626624e-01 3.00398827e-01 -1.49903387e-01 -3.28099638e-01
2.26605251e-01 -7.42667079e-01 3.78248096e-01 8.27756643e-01
6.14162803e-01 3.39521915e-01 -3.48656952e-01 1.09021080e+00
-1.91727400e-01 -1.00153014e-01 -6.21373475e-01 -4.10451353e-01
3.99863034e-01 1.31682575e+00 -9.16211426e-01 -5.31525254e-01
-1.90408245e-01 1.13498187e+00 6.68106228e-02 -1.82197824e-01
-1.24725997e+00 -6.14909470e-01 3.65239888e-01 -3.04073066e-01
3.39043029e-02 -3.96836281e-01 -3.11424762e-01 -1.34826267e+00
6.93720877e-02 -8.40890646e-01 -1.34819880e-01 -5.11953294e-01
-6.19498789e-01 5.77854693e-01 -4.43344653e-01 -1.22761166e+00
-1.18422858e-01 -7.14238226e-01 -4.46665972e-01 7.58427620e-01
-1.29308176e+00 -7.50559032e-01 2.83365436e-02 5.50469518e-01
-3.12840700e-01 1.69724435e-01 1.09736252e+00 4.77399558e-01
-5.05180776e-01 6.63146555e-01 2.96587765e-01 -2.21717343e-01
3.78735244e-01 -9.49283242e-01 5.19898057e-01 9.55534101e-01
1.76618382e-01 8.17725658e-01 9.21148717e-01 -4.30925399e-01
-2.38991237e+00 -6.59592271e-01 1.00731874e+00 -1.73178777e-01
1.06424284e+00 -5.90839565e-01 -4.81854141e-01 5.71767807e-01
-1.79954052e-01 7.95265660e-02 3.52122217e-01 -7.10139945e-02
-5.14260352e-01 -2.99055666e-01 -1.15819407e+00 6.03479207e-01
9.56019342e-01 -1.25147378e+00 -1.66786835e-01 4.33922797e-01
4.36625779e-01 -5.79981387e-01 -8.94465387e-01 2.07529038e-01
5.84553421e-01 -1.15598214e+00 5.72242320e-01 -3.28008920e-01
2.98294306e-01 -7.49333382e-01 -4.33719784e-01 -7.67377377e-01
8.07095319e-02 -9.18162882e-01 1.51229933e-01 9.01608348e-01
2.77124226e-01 -1.06538546e+00 4.98197138e-01 3.85414004e-01
-2.78045714e-01 -1.41915798e-01 -1.27256775e+00 -8.16375196e-01
-8.95871222e-02 -6.35336041e-01 8.32410038e-01 8.66728544e-01
4.53331679e-01 2.43236735e-01 -7.15475306e-02 6.84575438e-01
6.13113582e-01 6.12235904e-01 9.75883186e-01 -9.11220193e-01
-6.84132040e-01 -4.80837494e-01 -9.95996654e-01 -6.19952857e-01
2.99104810e-01 -1.23428476e+00 -2.40222290e-02 -8.18465173e-01
2.40689740e-01 -5.56889534e-01 -2.25195006e-01 1.12521730e-01
4.14827853e-01 3.47157657e-01 1.59371585e-01 9.82170925e-02
-5.90245426e-01 2.87553430e-01 1.02989972e+00 -1.30909100e-01
-7.96061084e-02 -1.52292559e-02 -4.02422607e-01 1.09369069e-01
8.33487451e-01 -6.34505510e-01 -1.43019497e-01 1.35451913e-01
6.13298774e-01 4.46095705e-01 5.39893627e-01 -1.40009916e+00
4.87121195e-01 1.55468985e-01 -6.18763685e-01 -2.79294461e-01
2.60387510e-01 -8.10343325e-01 7.16585934e-01 1.00319171e+00
1.06443554e-01 -5.88157028e-02 8.86849090e-02 2.05178663e-01
-1.38883188e-01 -4.78726327e-01 7.92505503e-01 1.43797323e-01
-5.56710482e-01 9.94665772e-02 -2.60207385e-01 -4.19627696e-01
1.03515899e+00 7.30816200e-02 -6.24107778e-01 9.47236866e-02
-3.43126088e-01 -1.74718007e-01 1.05709743e+00 -1.82807878e-01
2.20963463e-01 -1.03490782e+00 -2.40466967e-01 3.58994067e-01
2.95729071e-01 -4.12186891e-01 5.00546694e-01 1.22188461e+00
-1.07698226e+00 8.63407373e-01 -3.77466142e-01 -4.74015564e-01
-9.16106403e-01 6.60431206e-01 2.89622813e-01 -2.72367835e-01
-4.69990373e-01 2.16212511e-01 -3.80630493e-02 -5.84586024e-01
-6.22168005e-01 -6.92087293e-01 7.41504788e-01 -6.54167175e-01
3.75908345e-01 8.27803463e-02 4.94672149e-01 -5.23185611e-01
-5.00264764e-01 5.91741800e-01 2.45011240e-01 -1.45277470e-01
7.09503651e-01 8.53235573e-02 -7.45618701e-01 2.01617718e-01
1.34339166e+00 1.56021118e-01 -2.48543501e-01 7.28452355e-02
-1.62649125e-01 -1.53968841e-01 3.35116126e-02 -1.88594714e-01
-6.20161355e-01 8.64401877e-01 5.20530522e-01 3.18246901e-01
1.12068558e+00 -1.95038602e-01 8.20824265e-01 9.77769017e-01
1.39278448e+00 -9.16151166e-01 -8.06485176e-01 4.19535458e-01
-1.29228547e-01 -7.82108188e-01 -1.42548289e-02 -4.89135593e-01
4.19461317e-02 1.26915002e+00 -2.96273176e-02 -2.01996148e-01
2.85925180e-01 2.64993995e-01 -4.83706295e-01 -2.09895968e-02
-5.57476878e-01 -1.77346140e-01 -1.92116588e-01 -1.48027673e-01
7.78233632e-02 4.69233185e-01 -5.15908480e-01 3.68044108e-01
-4.26709861e-01 1.81740522e-01 9.32720840e-01 8.78226221e-01
-1.50602892e-01 -1.55738389e+00 -3.72976959e-01 1.32676929e-01
-1.43250436e-01 -2.92127877e-01 -1.00879870e-01 7.11113214e-01
1.95460603e-01 9.44917858e-01 -2.19521210e-01 -6.80872977e-01
2.19397709e-01 -4.89677936e-02 9.64152575e-01 -4.16222155e-01
-6.82130516e-01 -3.03820670e-01 9.15452689e-02 -7.88050115e-01
-2.63335764e-01 -4.73885477e-01 -1.73614657e+00 -8.97559404e-01
-5.92443585e-01 5.17257810e-01 8.56220961e-01 8.61100316e-01
2.65216380e-01 3.76819700e-01 7.16248155e-01 -6.00386381e-01
-5.85931480e-01 -4.21044379e-01 -8.55393231e-01 -2.87963077e-02
6.59444705e-02 -2.35050619e-01 -3.65556926e-01 -3.50872189e-01] | [5.590461730957031, 4.942792892456055] |
f0763fa8-1cc0-465b-95ea-d6359940126f | unsupervised-multi-stream-highlight-detection | 1910.06189 | null | https://arxiv.org/abs/1910.06189v2 | https://arxiv.org/pdf/1910.06189v2.pdf | Unsupervised Multi-stream Highlight detection for the Game "Honor of Kings" | With the increasing popularity of E-sport live, Highlight Flashback has been a critical functionality of live platforms, which aggregates the overall exciting fighting scenes in a few seconds. In this paper, we introduce a novel training strategy without any additional annotation to automatically generate highlights for game video live. Considering that the existing manual edited clips contain more highlights than long game live videos, we perform pair-wise ranking constraints across clips from edited and long live videos. A multi-stream framework is also proposed to fuse spatial, temporal as well as audio features extracted from videos. To evaluate our method, we test on long game live videos with an average length of about 15 minutes. Extensive experimental results on videos demonstrate its satisfying performance on highlights generation and effectiveness by the fusion of three streams. | ['Hui Zhan', 'Wentao Yao', 'Li Wang', 'Chengwei Zhu', 'Zixun Sun'] | 2019-10-14 | null | null | null | null | ['highlight-detection'] | ['computer-vision'] | [ 1.09003946e-01 -2.88639128e-01 5.16597666e-02 -9.78842005e-02
-1.02003908e+00 -6.72945678e-01 5.11256516e-01 8.00438747e-02
-4.80448604e-01 6.05654120e-01 3.12881291e-01 2.71483302e-01
-2.06610173e-01 -4.07741815e-01 -6.33546650e-01 -4.04304445e-01
-5.42406738e-01 -4.72021163e-01 7.34603286e-01 -1.29208177e-01
3.73028815e-01 1.20497361e-01 -1.96934938e+00 6.61376476e-01
6.46372259e-01 1.12233400e+00 1.73288509e-01 9.09984469e-01
3.58469971e-02 1.25527978e+00 -1.00268459e+00 -4.97489929e-01
4.17181253e-01 -4.36873227e-01 -1.79739103e-01 2.95302738e-02
5.95713794e-01 -5.57273567e-01 -3.80379796e-01 8.50627601e-01
7.67174304e-01 4.82624143e-01 1.17816851e-02 -1.40589595e+00
2.35328838e-01 8.75676095e-01 -5.71808338e-01 6.32368326e-01
7.30717123e-01 1.67358324e-01 1.04437530e+00 -1.07590592e+00
7.85750508e-01 6.51159167e-01 5.68784356e-01 4.61594090e-02
-6.19459748e-01 -1.01452887e+00 1.22161180e-01 6.03816628e-01
-1.43660772e+00 -5.82690299e-01 8.91594052e-01 -4.14186567e-01
6.33040190e-01 3.85913104e-01 9.24637616e-01 7.98914135e-01
2.17463255e-01 6.63697779e-01 6.30063593e-01 -1.28961220e-01
2.52862778e-02 -3.58587116e-01 -3.67884099e-01 2.40496859e-01
-1.82366669e-01 1.11872479e-01 -1.26475668e+00 -1.37566268e-01
6.34066820e-01 3.19656245e-02 -2.88570702e-01 4.59942445e-02
-1.29411781e+00 6.02358222e-01 -3.92075330e-02 1.61396973e-02
-5.47340691e-01 1.24635682e-01 8.62921119e-01 2.82145500e-01
4.74545896e-01 3.83834511e-01 4.73133959e-02 -8.14971507e-01
-1.36449921e+00 3.51175070e-01 5.48417091e-01 9.59539175e-01
4.37904269e-01 7.22590759e-02 -6.13025725e-01 7.58088589e-01
-4.06225652e-01 2.55136967e-01 1.67577043e-01 -1.06227171e+00
4.46202397e-01 2.30545834e-01 1.44458860e-01 -1.40436125e+00
-1.75381288e-01 -4.66722190e-01 -4.72726583e-01 4.08873521e-02
6.96163252e-02 -2.95827031e-01 -2.30758563e-01 1.43242669e+00
1.96109131e-01 8.56617391e-01 -4.39002246e-01 1.00447762e+00
9.10605967e-01 8.69693458e-01 -4.87953871e-02 -6.05982959e-01
1.12375665e+00 -9.86977100e-01 -9.08193946e-01 2.57045180e-01
1.23750627e-01 -1.01649225e+00 1.15494943e+00 8.29904199e-01
-1.26572597e+00 -6.13795042e-01 -1.18275380e+00 1.39431685e-01
1.89077705e-01 -7.10842311e-02 4.01698977e-01 3.09540391e-01
-6.41169429e-01 4.73150998e-01 -4.07493919e-01 6.13036007e-03
2.24186793e-01 6.62639067e-02 -3.52176785e-01 1.19504429e-01
-1.10356951e+00 1.17333889e-01 5.16966522e-01 -5.76108657e-02
-1.03501546e+00 -9.99348879e-01 -6.50954604e-01 5.48398085e-02
8.52936745e-01 -3.41442004e-02 1.23874855e+00 -1.00798714e+00
-1.56268120e+00 5.91342151e-01 1.53413251e-01 -3.51118863e-01
5.37555516e-01 -7.58837998e-01 -5.47525764e-01 6.99923694e-01
1.68979373e-02 4.91874784e-01 7.39279747e-01 -9.53233361e-01
-1.23055291e+00 1.88013926e-01 6.16399705e-01 4.63605493e-01
-6.58878088e-01 4.67318594e-01 -8.43951285e-01 -8.14048052e-01
-3.60879302e-01 -7.23553777e-01 1.14262998e-01 -1.20781176e-01
-1.65699273e-01 2.73696899e-01 7.37305701e-01 -6.26579106e-01
1.80638576e+00 -2.41932440e+00 1.14936560e-01 2.42609128e-01
2.10360110e-01 -1.78985484e-02 -1.29710779e-01 4.88622457e-01
1.10115647e-01 -3.52314234e-01 2.32477412e-01 -5.04396409e-02
-9.25806090e-02 -1.23611256e-01 -4.78678882e-01 9.57442671e-02
-2.90381480e-02 3.47218841e-01 -1.13875020e+00 -7.52992094e-01
1.94706872e-01 4.38262045e-01 -7.29560375e-01 3.93912047e-01
5.93013279e-02 4.41140443e-01 -6.59196824e-02 6.06295109e-01
5.01492143e-01 2.31013268e-01 -1.50238439e-01 -2.36053020e-01
-5.11321425e-01 -3.57634132e-03 -1.27618587e+00 2.01402783e+00
-1.88243985e-01 6.46502376e-01 -6.15959093e-02 -5.36908090e-01
7.22336590e-01 4.39595580e-01 8.72051656e-01 -9.08269048e-01
1.53682709e-01 5.09857275e-02 -2.61198044e-01 -7.37295866e-01
9.87582862e-01 9.04561430e-02 -2.60146022e-01 2.46188328e-01
1.05842300e-01 1.37883022e-01 8.25659513e-01 4.53176558e-01
1.24283206e+00 3.08209240e-01 1.03804775e-01 2.26423115e-01
3.18119675e-01 -6.88740313e-02 7.12300837e-01 5.98443329e-01
-2.13391304e-01 9.75523055e-01 6.49214625e-01 -1.96282849e-01
-7.33314097e-01 -9.73878503e-01 4.16752070e-01 1.54067588e+00
3.88152778e-01 -1.17393494e+00 -7.11743653e-01 -3.63303751e-01
-5.39905846e-01 2.65976071e-01 -3.63697350e-01 -1.96721658e-01
-6.62803590e-01 -4.28525180e-01 5.63770592e-01 3.20848614e-01
3.52423459e-01 -1.15909278e+00 -1.04460120e+00 4.27808464e-01
-4.25922453e-01 -1.24528635e+00 -8.01268041e-01 -3.18940639e-01
-2.87971497e-01 -1.25328791e+00 -7.79141665e-01 -5.49427152e-01
1.01309419e-01 5.80918729e-01 1.00684297e+00 -1.49166705e-02
-3.89510900e-01 3.48192751e-01 -9.08219039e-01 -2.84819245e-01
1.42322510e-01 -6.56989887e-02 -1.65819854e-01 1.70791060e-01
-1.08101685e-02 -6.42135978e-01 -6.56502187e-01 2.95205921e-01
-1.10005832e+00 2.27323905e-01 4.50594276e-01 6.77935302e-01
7.55093038e-01 2.52145976e-01 4.05152112e-01 -6.13712192e-01
6.92247510e-01 -5.05460978e-01 -2.97086567e-01 -2.28536800e-02
1.93318516e-01 -7.60399997e-01 6.05080903e-01 -5.70908785e-01
-9.83810306e-01 -6.59886971e-02 7.95617104e-02 -7.99223781e-01
2.37917244e-01 7.46779621e-01 -3.57120335e-02 1.82421252e-01
5.45188248e-01 5.38663985e-03 -3.12856257e-01 -2.09180593e-01
2.00713798e-01 4.49589670e-01 9.80665147e-01 -4.82815564e-01
7.40178823e-01 3.82987589e-01 -3.19804877e-01 -8.28052998e-01
-7.61601985e-01 -6.40435040e-01 -3.19314331e-01 -9.49113309e-01
6.37242913e-01 -1.14579284e+00 -7.83410609e-01 3.85630637e-01
-9.26214874e-01 -3.03104252e-01 -4.06168759e-01 6.42628193e-01
-4.48922336e-01 2.68492609e-01 -5.90907574e-01 -6.11584663e-01
-3.62044662e-01 -8.53533208e-01 9.25540626e-01 3.66611332e-01
-1.61410451e-01 -3.75962377e-01 2.16732815e-01 1.39537081e-01
1.98272660e-01 5.40471017e-01 -4.31875251e-02 -1.65066764e-01
-4.91064698e-01 -3.30797017e-01 -1.77399531e-01 5.42344004e-02
7.35956803e-02 3.72243106e-01 -8.02474022e-01 -2.76022196e-01
-3.19956988e-01 -1.92389548e-01 7.32255816e-01 3.53907853e-01
1.24824190e+00 -1.86669663e-01 1.65112033e-01 6.58906758e-01
1.19060731e+00 3.55659097e-01 7.25948572e-01 3.28548729e-01
5.54533362e-01 6.09880865e-01 1.45448816e+00 1.12997878e+00
1.59739051e-02 7.39065588e-01 5.24252772e-01 -9.43413153e-02
6.78970106e-03 -3.00088793e-01 7.81010032e-01 9.90372896e-01
-4.22984004e-01 -2.19658896e-01 -5.09968460e-01 4.48490620e-01
-1.80185759e+00 -1.64218712e+00 -1.63185298e-02 2.45368576e+00
6.08679473e-01 3.72742146e-01 5.28462291e-01 3.73418033e-01
8.72356594e-01 4.09507036e-01 -2.57891476e-01 -1.73718482e-01
-1.64359286e-01 3.42959166e-01 5.85900880e-02 3.07912201e-01
-1.12878633e+00 6.75653756e-01 6.36154318e+00 1.23651826e+00
-1.33580542e+00 1.76679924e-01 3.04240793e-01 -1.01285303e+00
-1.95078894e-01 -1.38329893e-01 -3.47828805e-01 7.22552478e-01
7.44068027e-01 -4.95406628e-01 1.79144695e-01 6.46524370e-01
6.64216757e-01 -1.41047001e-01 -6.29105330e-01 1.24577081e+00
3.83386374e-01 -1.45439875e+00 -1.05153337e-01 -1.27869681e-01
6.43473983e-01 -3.91640753e-01 2.13577636e-02 3.31383616e-01
-2.21287563e-01 -6.94096804e-01 1.17895317e+00 4.64679718e-01
1.09466338e+00 -1.13955855e+00 5.37738144e-01 3.29345390e-02
-1.64820194e+00 -1.78950876e-01 -2.15065882e-01 -2.47201040e-01
4.58700329e-01 6.20339870e-01 -2.74982512e-01 7.12422371e-01
8.57156038e-01 9.23268259e-01 -5.52566767e-01 1.54488945e+00
-1.33593976e-01 6.25901818e-01 -1.66641340e-01 2.91314542e-01
3.04678250e-02 -4.26768512e-03 7.24464715e-01 1.41976440e+00
6.59772277e-01 3.43119293e-01 3.39021742e-01 2.49177083e-01
2.09494829e-01 3.93922567e-01 -4.40087259e-01 -6.58532837e-03
4.34099436e-01 1.64151824e+00 -7.74265409e-01 -3.68060350e-01
-4.46830064e-01 9.09298062e-01 -1.30368739e-01 8.91792923e-02
-1.45970154e+00 -6.55997276e-01 4.12486047e-01 2.69739687e-01
2.40607232e-01 -2.04622343e-01 2.85712004e-01 -1.16852176e+00
2.30369180e-01 -8.12731683e-01 4.78896081e-01 -9.08527076e-01
-8.59875381e-01 7.94653952e-01 5.20581715e-02 -1.85505438e+00
-1.46698445e-01 1.84389241e-02 -9.56783175e-01 7.37072229e-02
-1.20930040e+00 -8.70354652e-01 -6.87492013e-01 7.27928042e-01
7.84272611e-01 -4.82878536e-02 4.42965984e-01 8.59041452e-01
-6.53111756e-01 4.92171079e-01 -1.96550518e-01 -2.06308410e-01
1.06878686e+00 -9.20077682e-01 -1.50402561e-01 1.01013780e+00
1.34611920e-01 2.18981713e-01 9.70222473e-01 -5.80774546e-01
-1.22116482e+00 -9.25037861e-01 3.58535111e-01 1.68230519e-01
6.39816701e-01 -2.12746084e-01 -6.24601483e-01 2.25494191e-01
4.05862778e-01 -4.13539782e-02 1.03188694e+00 -4.43304330e-02
-9.50884074e-02 -3.65130991e-01 -6.00429952e-01 4.77658153e-01
1.25800300e+00 -4.38637853e-01 -3.33808750e-01 6.98893797e-03
5.46658576e-01 -6.64276838e-01 -7.29906201e-01 3.47218692e-01
8.45978200e-01 -1.44253933e+00 7.49532104e-01 1.11566801e-02
6.41560256e-01 -4.79492664e-01 2.18985360e-02 -1.10206473e+00
-6.72341138e-02 -1.04957485e+00 -1.20179625e-02 1.51678836e+00
-5.97644150e-02 1.66788429e-01 6.53403461e-01 2.10536961e-02
-3.45595479e-01 -5.51687658e-01 -7.13697135e-01 -5.76854706e-01
-8.62872481e-01 -6.74089789e-01 3.23437333e-01 7.39827991e-01
3.13466877e-01 9.32687223e-02 -8.36757004e-01 -7.64821097e-02
4.27046955e-01 1.47170067e-01 1.05041838e+00 -1.19192171e+00
-4.69542831e-01 -2.27937847e-01 -5.45851111e-01 -9.02558327e-01
-1.19554326e-01 -3.05454254e-01 1.61107153e-01 -1.07483160e+00
5.23771167e-01 -2.17954703e-02 -6.76046252e-01 1.35117993e-01
-2.72304147e-01 7.20483780e-01 3.56802016e-01 1.04294680e-01
-1.29363644e+00 4.20692652e-01 1.02698553e+00 9.53697935e-02
-3.67411077e-01 -1.99841052e-01 -4.42818701e-01 6.51497662e-01
4.82849717e-01 -3.69435161e-01 -6.28737569e-01 -1.98030844e-01
5.41452467e-01 2.71505862e-01 1.23970948e-01 -1.48675692e+00
2.22657278e-01 -2.39609346e-01 -7.54457805e-03 -7.46345937e-01
5.74925125e-01 -5.13644934e-01 4.34357375e-01 9.05055404e-02
-3.80653590e-01 7.27831870e-02 3.88644546e-01 4.01689857e-01
-7.92655826e-01 1.59345359e-01 4.74025965e-01 2.16004252e-01
-9.24017966e-01 4.12057042e-01 -3.02162796e-01 6.11881949e-02
1.26678181e+00 -5.04566610e-01 -7.93875828e-02 -5.96685767e-01
-6.99074984e-01 2.18164936e-01 3.79566133e-01 5.30826092e-01
7.50216782e-01 -1.40880585e+00 -6.59251511e-01 -4.32393551e-02
2.64444113e-01 -2.13332146e-01 8.02742064e-01 9.27701890e-01
-7.37448573e-01 -1.03713326e-01 -5.65062821e-01 -4.58262414e-01
-1.61997032e+00 4.22537625e-01 -2.05309793e-01 -1.85195193e-01
-7.49294579e-01 8.54001164e-01 3.40990692e-01 4.00373787e-01
3.13072681e-01 -2.79477835e-02 -4.72071528e-01 5.33429563e-01
1.01524794e+00 4.90230620e-01 -1.27907529e-01 -7.84599066e-01
-1.15036450e-01 3.48649383e-01 4.17829677e-02 -3.50091070e-01
1.38976157e+00 -1.97298467e-01 3.21151406e-01 6.13926888e-01
8.28154624e-01 5.59042394e-01 -1.54985130e+00 9.08657834e-02
-1.92574486e-01 -9.27907109e-01 -1.77397415e-01 -2.44961470e-01
-1.24550843e+00 7.01879978e-01 3.62998694e-01 2.09483519e-01
1.47002172e+00 -4.06368583e-01 1.10495842e+00 -1.22623540e-01
5.47562897e-01 -1.24999106e+00 5.43980598e-01 3.02684069e-01
1.06216395e+00 -7.89993465e-01 2.19855774e-02 -6.16667032e-01
-7.65945435e-01 1.05379951e+00 6.01039588e-01 -2.24937588e-01
3.70820075e-01 4.10444796e-01 1.90427005e-02 -1.71424091e-01
-8.88657033e-01 -1.24646775e-01 7.62538090e-02 3.46834362e-01
4.12631750e-01 -2.24056974e-01 -5.24321854e-01 7.36780703e-01
-2.78273702e-01 1.83539882e-01 8.21748614e-01 1.00326836e+00
-5.50318718e-01 -8.26777995e-01 -4.63662475e-01 2.42791340e-01
-8.04751694e-01 2.48375088e-02 -2.22835392e-01 5.23031235e-01
2.76394248e-01 1.06613767e+00 2.02735052e-01 -6.83575928e-01
4.43976074e-01 -3.04171443e-01 2.90504545e-01 -4.61780936e-01
-9.75048959e-01 5.47788978e-01 1.57213032e-01 -9.86022174e-01
-5.55611491e-01 -5.50000131e-01 -1.21081448e+00 -3.77926469e-01
-2.77765870e-01 3.70868653e-01 2.67523885e-01 4.78803724e-01
4.69555348e-01 8.27334702e-01 8.48204195e-01 -1.23065937e+00
1.97558179e-01 -7.51986742e-01 -7.98054338e-01 5.89387238e-01
6.87213093e-02 -7.37465322e-01 -2.80960798e-01 3.72519135e-01] | [10.095919609069824, 0.417460173368454] |
24567745-c8a9-4e20-b7ad-02b630a5a8e8 | a-combined-pca-mlp-network-for-early-breast | 2206.09128 | null | https://arxiv.org/abs/2206.09128v1 | https://arxiv.org/pdf/2206.09128v1.pdf | A Combined PCA-MLP Network for Early Breast Cancer Detection | Breast cancer is the second most responsible for all cancer types and has been the cause of numerous deaths over the years, especially among women. Any improvisation of the existing diagnosis system for the detection of cancer can contribute to minimizing the death ratio. Moreover, cancer detection at an early stage has recently been a prime research area in the scientific community to enhance the survival rate. Proper choice of machine learning tools can ensure early-stage prognosis with high accuracy. In this paper, we have studied different machine learning algorithms to detect whether a patient is likely to face breast cancer or not. Due to the implicit behavior of early-stage features, we have implemented a multilayer perception model with the integration of PCA and suggested it to be more viable than other detection algorithms. Our 4 layers MLP-PCA network has obtained the best accuracy of 100% with a mean of 90.48% accuracy on the BCCD dataset. | ['Md. Saif Hassan Onim', 'Arunima Dey Pooja', 'Md. Wahiduzzaman Khan Arnob'] | 2022-06-18 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 2.04168633e-01 3.34376365e-01 -5.52311361e-01 -2.51225561e-01
-3.62913370e-01 3.39071393e-01 5.54635584e-01 5.87134659e-01
-5.28499365e-01 5.77577055e-01 -4.44111042e-02 -4.07448202e-01
-1.89982399e-01 -1.02888501e+00 -1.62904829e-01 -9.33996499e-01
-5.87823875e-02 4.28554416e-01 -1.21617518e-01 6.70826882e-02
3.09981406e-01 8.74596238e-01 -1.43683398e+00 4.70978409e-01
5.94649255e-01 1.01825166e+00 2.77136147e-01 5.82090199e-01
-4.61653098e-02 8.24790359e-01 -3.12087178e-01 -3.53823274e-01
-1.95273846e-01 -1.50647178e-01 -6.52048647e-01 -3.36763769e-01
-2.26689517e-01 -1.43017322e-01 5.95718157e-03 9.31356430e-01
4.00007099e-01 -5.23277521e-01 9.57862198e-01 -9.16522622e-01
-8.59448984e-02 4.24818188e-01 -7.39652753e-01 -1.29647776e-02
1.85287133e-01 -3.35929841e-01 3.84180456e-01 -7.88641632e-01
1.32059515e-01 1.01436031e+00 7.25805104e-01 8.84762287e-01
-1.18890572e+00 -5.38674891e-01 -5.56779683e-01 5.76448739e-01
-1.40501106e+00 -3.71010840e-01 6.94368780e-01 -2.89182842e-01
6.70875072e-01 4.96516764e-01 8.06726992e-01 9.71154690e-01
6.70635879e-01 4.94425684e-01 1.16628480e+00 -9.14030969e-01
1.12444676e-01 4.54527140e-01 2.44455442e-01 7.73655653e-01
7.09099412e-01 1.91182107e-01 -1.32459536e-01 -1.79905415e-01
5.45913637e-01 2.36069813e-01 1.33459330e-01 -1.08529352e-01
-7.30553567e-01 7.39484847e-01 4.09275681e-01 6.68528855e-01
-4.01658118e-01 8.01380128e-02 1.65021166e-01 -9.83543321e-02
9.86821949e-02 1.97065905e-01 -3.70995820e-01 9.67329443e-02
-6.46440387e-01 -2.43532330e-01 7.81223834e-01 -4.50774245e-02
3.49332064e-01 -4.50243443e-01 -8.71268809e-02 7.48740375e-01
3.34989011e-01 4.09778178e-01 5.71942687e-01 -7.18105078e-01
-2.87656665e-01 1.00784266e+00 -1.63665220e-01 -1.39960575e+00
-6.85715020e-01 -7.67735422e-01 -1.30224979e+00 4.06306118e-01
4.28698033e-01 4.22343165e-02 -5.95635355e-01 1.27171183e+00
1.61255643e-01 -7.60405958e-02 2.70939201e-01 4.39417005e-01
5.94804049e-01 6.22434676e-01 4.16690826e-01 -2.04535857e-01
1.50450516e+00 -4.47495252e-01 -5.91738224e-01 -2.27435485e-01
5.26589811e-01 -4.28016216e-01 3.50430280e-01 5.67359507e-01
-4.63226497e-01 -4.12457973e-01 -8.63744378e-01 2.38007829e-01
-2.76510984e-01 7.39135921e-01 9.99038041e-01 9.72113967e-01
-9.66292679e-01 4.59277213e-01 -1.07703030e+00 -8.19241643e-01
4.68867511e-01 4.29670483e-01 -5.12491822e-01 -1.62774831e-01
-6.99579358e-01 1.19998729e+00 5.14668107e-01 2.76125193e-01
-8.06177139e-01 -3.85888815e-01 -2.21606836e-01 7.04584643e-02
1.15341749e-02 -4.80097175e-01 8.94673169e-01 -1.18096566e+00
-1.02977908e+00 1.10890210e+00 -4.50500399e-01 -4.82538342e-01
3.77559692e-01 -8.08746293e-02 -3.92142028e-01 1.72384173e-01
-2.00220764e-01 6.32831097e-01 5.75216770e-01 -1.10835326e+00
-1.09363282e+00 -2.95767039e-01 -4.14092183e-01 -2.37399507e-02
-6.69186711e-01 -9.76280794e-02 -2.43314236e-01 -3.98524031e-02
2.30279505e-01 -7.19718039e-01 -4.08815742e-01 1.81647003e-01
-1.20676547e-01 -3.59018177e-01 6.65449858e-01 -8.95900428e-01
1.08658040e+00 -2.22040606e+00 3.01329116e-03 2.08320245e-01
8.32401961e-02 1.13391623e-01 4.19623435e-01 1.93390042e-01
-6.17926680e-02 9.18365717e-02 3.73341963e-02 -1.13656327e-01
-5.86402535e-01 3.55782092e-01 2.34707102e-01 5.11853635e-01
4.23128009e-01 3.64454836e-01 -7.15124786e-01 -7.16074288e-01
2.53668487e-01 7.09156096e-01 -1.73227400e-01 2.30871558e-01
2.20860541e-01 2.00735956e-01 -5.84807456e-01 9.11271513e-01
6.34405971e-01 -2.10378125e-01 2.55969822e-01 -4.70661633e-02
4.49355580e-02 -3.07177454e-01 -6.67941511e-01 1.13695014e+00
-1.62642077e-01 6.43495262e-01 9.99403521e-02 -1.08899617e+00
8.49461675e-01 5.33881545e-01 6.07757092e-01 -5.91043353e-01
2.96102077e-01 3.58608425e-01 2.33321562e-01 -7.42410541e-01
5.25437593e-02 -1.83668897e-01 6.19164892e-02 -2.49268666e-01
-3.72685194e-01 3.21037650e-01 -7.52596781e-02 -6.31864965e-02
1.13663685e+00 -3.72164249e-01 8.09668601e-01 -3.16824287e-01
8.16382289e-01 3.24009001e-01 6.29025042e-01 6.64446592e-01
-5.16767561e-01 1.44235909e-01 6.57446802e-01 -4.11104143e-01
-5.25227189e-01 -7.87296236e-01 -5.47232747e-01 4.63542134e-01
-1.74968272e-01 2.42251173e-01 -3.80704910e-01 -4.43063855e-01
-3.73697025e-03 7.08695531e-01 -7.45056331e-01 -3.21430057e-01
-2.54706115e-01 -1.25779307e+00 4.79294747e-01 2.41501316e-01
4.95881468e-01 -1.01128578e+00 -6.40954316e-01 1.59058705e-01
-4.15987223e-02 -4.90363121e-01 6.79269254e-01 5.68060577e-01
-1.06340885e+00 -1.31291950e+00 -7.59180427e-01 -9.94953036e-01
1.01445460e+00 5.64412549e-02 6.95705056e-01 3.29873532e-01
-5.21030664e-01 9.56667811e-02 -3.39151353e-01 -5.70800602e-01
-8.96476746e-01 1.12187974e-01 -1.33709460e-01 -4.84330356e-02
6.74098849e-01 -9.07831714e-02 -5.39630592e-01 -1.59843326e-01
-4.73672986e-01 1.73627242e-01 1.09254694e+00 7.35172987e-01
4.86280292e-01 5.48435986e-01 4.86421198e-01 -8.37386727e-01
4.72048581e-01 -5.76797605e-01 -3.36724043e-01 1.26580358e-01
-6.53114438e-01 -1.24537461e-01 2.66895682e-01 -2.29493678e-01
-1.02896857e+00 4.12491798e-01 -3.53063136e-01 2.46933743e-01
-6.45934165e-01 5.08005142e-01 -9.47339088e-02 -1.72456339e-01
4.27978098e-01 1.21481605e-01 2.35639617e-01 -3.85861814e-01
-4.33873534e-01 7.13726938e-01 3.76122445e-01 -5.64205870e-02
3.28159422e-01 4.44517374e-01 6.23567998e-01 -1.11736023e+00
-6.64182305e-01 -6.20888054e-01 -3.76683742e-01 -3.44794512e-01
9.68114495e-01 -7.53223360e-01 -8.68529677e-01 8.27891588e-01
-1.01866353e+00 7.41868466e-02 3.53435934e-01 4.99069393e-01
1.14736319e-01 1.95979923e-01 -5.98169506e-01 -1.09316123e+00
-4.56443965e-01 -8.58842790e-01 4.89049226e-01 5.05213201e-01
-2.37441123e-01 -9.70333040e-01 2.72816084e-02 2.17522502e-01
5.63447297e-01 4.29193377e-01 1.15286720e+00 -5.13231516e-01
-1.29625604e-01 -6.16864920e-01 -3.41389298e-01 1.94541678e-01
3.06393355e-01 3.06664884e-01 -1.31165373e+00 -1.95161045e-01
1.45943791e-01 8.44924822e-02 9.90996242e-01 5.47312558e-01
1.26237082e+00 -1.40240133e-01 -1.09826434e+00 2.73177266e-01
1.60916007e+00 5.47573984e-01 6.49190843e-01 4.27469224e-01
2.80308813e-01 6.29886270e-01 5.40156841e-01 1.92081869e-01
1.60254359e-01 1.48829728e-01 8.35237086e-01 -3.47161442e-01
3.73391286e-02 -1.92051642e-02 1.23509541e-01 5.91339827e-01
-1.78829238e-01 -8.41652825e-02 -1.12690187e+00 3.72409582e-01
-1.38115466e+00 -1.02705920e+00 -3.76016885e-01 2.06811047e+00
6.13640189e-01 3.14535737e-01 -3.01883489e-01 7.38215506e-01
5.79060793e-01 -4.89833444e-01 7.81277344e-02 -3.95302147e-01
6.65435717e-02 -4.68300134e-02 4.88025576e-01 1.80598497e-01
-1.35391068e+00 2.72699594e-01 6.45933962e+00 6.54955924e-01
-1.47670674e+00 -2.05481723e-01 1.19456041e+00 5.10281324e-01
2.37925440e-01 -2.05502063e-01 -8.61975849e-01 3.75919312e-01
8.87158394e-01 2.62030512e-01 -1.42129987e-01 9.96260583e-01
2.53827780e-01 -6.57443523e-01 -1.05540991e+00 5.60381651e-01
-8.34597554e-03 -9.75557446e-01 -1.94548681e-01 3.21742482e-02
2.75420368e-01 -4.68413383e-01 -9.63777006e-02 1.41401604e-01
-2.56515414e-01 -1.21714532e+00 1.95296407e-01 8.40015292e-01
6.44224107e-01 -8.14806938e-01 1.18410611e+00 7.83834159e-01
-6.88744545e-01 -2.82735974e-01 -1.87432379e-01 -2.63670981e-01
-3.63045335e-01 7.35146821e-01 -1.39459229e+00 2.10221842e-01
7.38381803e-01 5.17221332e-01 -8.19641769e-01 1.30932641e+00
-1.64781120e-02 8.30101848e-01 -4.53940123e-01 -5.41593254e-01
-3.07496697e-01 1.31176785e-01 3.02337408e-01 1.06158435e+00
4.84304488e-01 1.10609188e-04 -1.40282571e-01 2.35696375e-01
4.50662434e-01 6.81271851e-02 -3.59196246e-01 5.01255617e-02
2.99071640e-01 1.33163631e+00 -8.01858664e-01 4.94716242e-02
-3.64940256e-01 6.84010029e-01 2.28903681e-01 -1.30519703e-01
-3.95441085e-01 -9.18193758e-02 2.35711187e-01 2.41100356e-01
-1.30491570e-01 2.92106658e-01 -4.67765480e-01 -5.46201289e-01
-4.54730093e-01 -5.72844684e-01 4.89228219e-01 -3.49099785e-01
-1.10579360e+00 5.27771354e-01 -2.95128971e-01 -7.57726312e-01
3.93963559e-03 -9.80363011e-01 -7.25376904e-01 6.73390508e-01
-1.50571847e+00 -1.08689201e+00 -3.62035692e-01 1.17289767e-01
4.19013426e-02 -7.35931322e-02 1.37321150e+00 2.87279516e-01
-9.62667465e-01 3.91567558e-01 3.67638767e-02 1.46175683e-01
5.85817575e-01 -1.17009556e+00 -7.18933463e-01 6.19799614e-01
-5.15299618e-01 1.34094492e-01 8.24209332e-01 -5.87884188e-01
-1.39502048e+00 -9.09767389e-01 1.18682230e+00 -1.16965935e-01
2.81946987e-01 1.27001852e-01 -7.10356236e-01 1.24380499e-01
-1.50466226e-02 -3.17392349e-01 6.94846213e-01 -6.56719804e-02
3.66356224e-01 -6.63504541e-01 -1.35056305e+00 4.36042398e-01
1.39126971e-01 6.59818389e-03 -3.00810128e-01 2.87102442e-02
1.32116050e-01 5.03764451e-02 -7.63682425e-01 5.72487593e-01
6.73987865e-01 -1.05864894e+00 1.00551367e+00 -1.38657065e-02
5.69052279e-01 -8.15107003e-02 5.26110269e-02 -8.60612631e-01
-6.32299542e-01 2.97047049e-01 -4.54724021e-03 1.25459087e+00
5.23056805e-01 -6.57516658e-01 1.17243552e+00 5.91056585e-01
2.66950548e-01 -9.69726145e-01 -1.03663254e+00 -7.44896531e-02
-8.46151486e-02 -2.81888157e-01 1.12263605e-01 6.92755282e-01
-1.13491833e-01 -1.08665049e-01 -1.77827731e-01 3.60635281e-01
8.11872959e-01 -2.43759975e-01 3.99309039e-01 -1.50669611e+00
-2.37704337e-01 -5.98214269e-01 -6.01472855e-01 -1.36150867e-01
-3.40713382e-01 -7.57846355e-01 -1.71902865e-01 -1.59337592e+00
6.58893228e-01 -6.31844223e-01 -5.87507129e-01 6.73743844e-01
-1.88871995e-01 7.31209964e-02 -3.47782105e-01 8.16597119e-02
1.37918308e-01 -1.09114051e-01 1.06301355e+00 -1.61694869e-01
-3.15926550e-03 3.64773899e-01 -7.43024707e-01 8.77659917e-01
9.68353093e-01 -6.02411807e-01 1.79471031e-01 -2.67368928e-02
1.35195956e-01 3.36777776e-01 3.45879465e-01 -1.22519982e+00
2.57320762e-01 -3.28952491e-01 9.80915964e-01 -6.48126960e-01
5.28068066e-01 -1.37564993e+00 3.71590585e-01 1.38039052e+00
-7.02822208e-02 -5.43486357e-01 1.88483730e-01 5.00099182e-01
-1.17027946e-01 -4.64989007e-01 7.55052030e-01 -4.85911220e-02
-8.01775932e-01 -2.28913814e-01 -8.52825940e-01 -8.53855431e-01
1.42086887e+00 -1.66503966e-01 -1.51612222e-01 4.94553633e-02
-3.22203964e-01 6.80161417e-02 2.50106573e-01 1.48410380e-01
5.90037227e-01 -9.06250060e-01 -7.28395343e-01 2.42076248e-01
1.08135670e-01 -2.84555793e-01 1.04410224e-01 8.93173695e-01
-8.68147135e-01 5.26021183e-01 -3.45128149e-01 -5.39230168e-01
-1.65317607e+00 4.52798963e-01 4.45975631e-01 -4.85557795e-01
-3.11514854e-01 8.98428261e-01 -3.09327692e-01 8.98979232e-02
4.77324188e-01 -3.62208188e-02 -8.09674084e-01 -1.34682227e-02
4.99180377e-01 5.70144296e-01 -5.33316098e-02 -4.41381693e-01
-5.59616923e-01 3.41979176e-01 -1.47478655e-01 3.62246782e-01
1.20005238e+00 2.83937305e-01 -4.66736972e-01 3.02367479e-01
9.00382757e-01 -3.01590208e-02 -5.68300128e-01 4.47349250e-01
2.45260388e-01 -2.13476956e-01 3.45645219e-01 -1.05540216e+00
-8.98484647e-01 7.58243263e-01 1.03220129e+00 2.92466223e-01
1.38001263e+00 -1.19851895e-01 1.74703270e-01 4.25763279e-01
3.61199588e-01 -6.83318257e-01 -1.85135871e-01 -3.70443985e-02
5.71832538e-01 -1.67849886e+00 1.59188613e-01 -4.87001449e-01
-2.07139701e-01 1.40487432e+00 4.38894659e-01 -2.04262406e-01
7.82336295e-01 3.69114816e-01 2.14085519e-01 1.19281746e-01
-7.73866773e-01 2.59548724e-01 1.64161846e-01 6.56985462e-01
5.81745028e-01 4.26472157e-01 -6.85742736e-01 6.66834056e-01
2.91783273e-01 2.34936431e-01 1.97418571e-01 7.77339578e-01
-7.67632604e-01 -1.23444092e+00 -7.11200953e-01 8.54938388e-01
-8.12438309e-01 3.62408072e-01 -4.51698393e-01 7.44597197e-01
3.34288538e-01 9.86067355e-01 6.28335625e-02 -2.20695376e-01
-2.06564501e-01 -1.83392297e-02 3.11896473e-01 -2.36988544e-01
-4.60239857e-01 -1.46750689e-01 1.34060577e-01 -1.85773790e-01
-5.40082932e-01 -6.44625545e-01 -1.28578889e+00 -2.12772176e-01
-2.17100546e-01 1.16583049e-01 9.13391292e-01 8.41380537e-01
2.13186219e-02 5.44263244e-01 4.30886239e-01 -4.32572156e-01
-5.10587811e-01 -9.38782573e-01 -4.85139042e-01 -1.47396386e-01
1.52003512e-01 -3.92795175e-01 -3.86490673e-01 -4.90099788e-02] | [15.287921905517578, -2.771505355834961] |
49edcabf-2108-4eab-95d3-fee606786e5d | livable-exploring-long-tailed-classification | 2306.06935 | null | https://arxiv.org/abs/2306.06935v1 | https://arxiv.org/pdf/2306.06935v1.pdf | LIVABLE: Exploring Long-Tailed Classification of Software Vulnerability Types | Prior studies generally focus on software vulnerability detection and have demonstrated the effectiveness of Graph Neural Network (GNN)-based approaches for the task. Considering the various types of software vulnerabilities and the associated different degrees of severity, it is also beneficial to determine the type of each vulnerable code for developers. In this paper, we observe that the distribution of vulnerability type is long-tailed in practice, where a small portion of classes have massive samples (i.e., head classes) but the others contain only a few samples (i.e., tail classes). Directly adopting previous vulnerability detection approaches tends to result in poor detection performance, mainly due to two reasons. First, it is difficult to effectively learn the vulnerability representation due to the over-smoothing issue of GNNs. Second, vulnerability types in tails are hard to be predicted due to the extremely few associated samples.To alleviate these issues, we propose a Long-taIled software VulnerABiLity typE classification approach, called LIVABLE. LIVABLE mainly consists of two modules, including (1) vulnerability representation learning module, which improves the propagation steps in GNN to distinguish node representations by a differentiated propagation method. A sequence-to-sequence model is also involved to enhance the vulnerability representations. (2) adaptive re-weighting module, which adjusts the learning weights for different types according to the training epochs and numbers of associated samples by a novel training loss. | ['Qing Liao', 'Ge Li', 'Haoyu Wang', 'Feng Luo', 'Cuiyun Gao', 'Xin-Cheng Wen'] | 2023-06-12 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-8.45731869e-02 -1.25581309e-01 -1.22201130e-01 -1.30015954e-01
-1.65705413e-01 -5.70526481e-01 7.55349249e-02 4.21454728e-01
9.46014712e-04 3.18203807e-01 -3.32009681e-02 -7.07381546e-01
-9.74289700e-02 -1.20537925e+00 -2.78293550e-01 -5.59281766e-01
-2.62955904e-01 -1.20463803e-01 6.38095140e-01 -4.36289608e-01
4.11793053e-01 1.01629376e-01 -1.37584543e+00 1.22277945e-01
1.20632744e+00 9.19826686e-01 1.54270515e-01 5.14281392e-01
-7.09228039e-01 7.06707418e-01 -8.88914049e-01 -4.51301783e-01
4.24128920e-02 -2.69429356e-01 -5.23455083e-01 -5.38716197e-01
-5.72403260e-02 -1.48002565e-01 -2.62396485e-01 1.72280896e+00
2.41102561e-01 -1.00139871e-01 7.69238472e-01 -1.49388015e+00
-6.57763362e-01 8.92915845e-01 -9.53496575e-01 3.29800487e-01
2.57958472e-01 -1.73382424e-02 1.02852774e+00 -4.27945912e-01
1.91016749e-01 1.21470273e+00 8.91341329e-01 4.53012884e-01
-7.74562120e-01 -7.30935276e-01 2.87876636e-01 2.14943156e-01
-1.29866183e+00 2.55185068e-01 1.01974463e+00 -8.39319289e-01
7.70767152e-01 1.95808530e-01 3.58005524e-01 1.00018048e+00
3.84671032e-01 3.17848265e-01 6.76651120e-01 -1.73190296e-01
3.29089522e-01 1.07732229e-01 5.21432817e-01 4.17734534e-01
4.23296928e-01 -6.79086596e-02 2.74545699e-01 -5.67328095e-01
4.14021760e-01 3.13563377e-01 -2.70465732e-01 -1.26909157e-02
-4.99338120e-01 1.06381524e+00 6.80057943e-01 2.69333571e-01
-2.81819940e-01 -1.08480200e-01 8.84114683e-01 4.50791329e-01
4.78535175e-01 2.29002863e-01 -5.31187236e-01 -7.12809339e-02
-8.67548227e-01 8.07559192e-02 6.80760682e-01 8.35454583e-01
9.08217967e-01 3.00184935e-01 -1.23784319e-01 8.47594142e-01
4.77239698e-01 2.20229919e-03 6.34775221e-01 -3.14535238e-02
8.11967015e-01 1.22903347e+00 -3.70744407e-01 -1.32454538e+00
-4.61472511e-01 -5.18725276e-01 -1.05982304e+00 4.56718445e-01
3.92490804e-01 -3.76556724e-01 -8.56045127e-01 1.76234698e+00
1.41852021e-01 9.52190235e-02 -3.65707912e-02 5.02377033e-01
8.92333806e-01 5.97678721e-01 2.59154618e-01 4.15873602e-02
1.27307129e+00 -7.48198986e-01 -3.67230266e-01 -1.98791057e-01
7.85695493e-01 -4.82656121e-01 1.16914618e+00 1.69463262e-01
-5.97462893e-01 -3.30908388e-01 -9.84020710e-01 4.92006123e-01
-6.64763212e-01 -2.17245847e-01 6.24975920e-01 9.52852726e-01
-9.59213257e-01 5.25382698e-01 -5.00446379e-01 -6.73965663e-02
4.07275140e-01 6.92999959e-02 -3.17866430e-02 6.44835234e-02
-1.53721023e+00 4.21670020e-01 6.58547342e-01 -3.45371813e-02
-7.62676835e-01 -6.97150767e-01 -9.36050832e-01 4.25582081e-01
3.89024407e-01 -1.67784616e-01 7.66220212e-01 -1.13021290e+00
-8.53804588e-01 1.51392594e-01 3.73102605e-01 3.80877027e-04
3.51391912e-01 2.22900763e-01 -6.70153797e-01 -1.25371367e-01
-1.60963133e-01 -3.73445414e-02 8.90709460e-01 -9.00723875e-01
-5.72716832e-01 -7.02962205e-02 4.43352282e-01 -2.63384312e-01
-1.07339394e+00 2.20322728e-01 -2.38921613e-01 -8.83368671e-01
-1.67661473e-01 -6.17180407e-01 -3.12774658e-01 -3.86660278e-01
-5.78102767e-01 -3.53327245e-01 8.00711274e-01 -8.21270585e-01
2.04163098e+00 -2.03225422e+00 -4.13493589e-02 5.18226445e-01
5.81591725e-01 5.19642711e-01 -3.65002036e-01 5.29349983e-01
-3.88245821e-01 3.36869627e-01 -4.42753911e-01 1.59231082e-01
-1.23885944e-01 -1.30487308e-01 -2.25746796e-01 -3.55792902e-02
2.23089784e-01 4.73374814e-01 -9.97886419e-01 -2.78930992e-01
-2.53895223e-01 3.78379852e-01 -5.17793894e-01 2.85306036e-01
-1.82785243e-01 -8.33459124e-02 -7.32935965e-01 7.56446958e-01
9.23366487e-01 -1.08497970e-01 5.72880544e-03 -9.79415420e-03
-1.38842519e-02 1.32288590e-01 -1.04269350e+00 9.38785374e-01
-4.84714836e-01 2.80161142e-01 -1.19622029e-01 -9.66486096e-01
1.17618310e+00 2.78384775e-01 -1.38424737e-02 -2.05655932e-01
1.51335791e-01 1.95537791e-01 3.18111807e-01 -6.32942080e-01
3.71615857e-01 2.73285866e-01 -1.53785512e-01 4.29612339e-01
-3.38602215e-01 3.22230250e-01 1.84327915e-01 2.36372158e-01
1.61348224e+00 -1.13240846e-01 2.08633631e-01 -1.79884151e-01
7.04674065e-01 -3.62372249e-01 9.38977540e-01 4.22217906e-01
-1.78582311e-01 4.56014127e-01 1.27034080e+00 -3.37216377e-01
-7.84011006e-01 -6.78287446e-01 5.19103333e-02 9.56338286e-01
1.54183917e-02 -4.95278805e-01 -9.65275943e-01 -1.15977693e+00
-1.14571504e-01 5.98916471e-01 -7.52036631e-01 -5.95471442e-01
-4.34659213e-01 -1.04861045e+00 6.62189126e-01 8.06502342e-01
1.72519162e-01 -1.33437216e+00 -3.16592097e-01 1.55672148e-01
6.52592331e-02 -5.45549631e-01 -5.27637005e-01 1.05213083e-01
-8.41971815e-01 -1.18274033e+00 -7.37040222e-01 -7.01571822e-01
9.08385694e-01 1.51704982e-01 1.05717623e+00 7.32829034e-01
-1.20599426e-01 -2.39969119e-01 -7.29602635e-01 -1.97026983e-01
-5.90866446e-01 1.68844864e-01 -2.41670370e-01 -1.19855896e-01
3.67990673e-01 -6.65960848e-01 -4.94883478e-01 2.08127886e-01
-9.92555201e-01 -7.15518534e-01 6.44512534e-01 7.42303431e-01
1.32382408e-01 6.47761703e-01 7.99461424e-01 -9.68986630e-01
1.04923427e+00 -1.08978140e+00 -6.09353602e-01 4.25564051e-01
-8.05960536e-01 -1.21678866e-01 1.04842973e+00 -6.78735971e-01
-8.65957797e-01 -3.90564173e-01 -3.68689239e-01 -2.74563253e-01
1.23610057e-01 9.96406436e-01 -3.89845222e-01 -3.29710752e-01
5.71373999e-01 1.20851181e-01 -3.05062961e-02 -3.68569195e-01
-1.22956961e-01 7.37527132e-01 5.55503108e-02 -5.03483593e-01
9.99575853e-01 -3.56480837e-01 -1.82478458e-01 -5.34883320e-01
-3.58954161e-01 -2.11926118e-01 -2.37407953e-01 -3.43898162e-02
6.88682973e-01 -5.12524664e-01 -3.79315972e-01 6.44589782e-01
-8.81767809e-01 -3.49579811e-01 2.27628008e-01 1.37544841e-01
8.26532766e-02 8.62038851e-01 -8.04914951e-01 -7.82399535e-01
-6.02196276e-01 -1.34890163e+00 4.71649379e-01 4.68932450e-01
-1.43916696e-01 -1.23079872e+00 3.60419080e-02 -2.36169621e-01
5.72389603e-01 2.52801210e-01 1.51142347e+00 -6.61069930e-01
-2.16600761e-01 -4.22518671e-01 -4.17972028e-01 5.60368359e-01
1.76319659e-01 4.95034754e-01 -6.62180960e-01 -3.96560282e-01
-2.12691188e-01 -1.41544789e-01 7.77550399e-01 1.58802032e-01
1.38106501e+00 -2.48165458e-01 -2.88304180e-01 6.12309873e-01
1.46836984e+00 2.60398328e-01 8.90792429e-01 4.95031774e-01
9.14264977e-01 6.77886188e-01 5.36172688e-01 3.99439543e-01
3.92418236e-01 3.96482229e-01 8.32045138e-01 7.30126351e-02
1.65622249e-01 -8.48671945e-04 4.93084490e-01 9.01943684e-01
-1.67485862e-03 -3.68021786e-01 -1.14659274e+00 4.61629570e-01
-1.68033552e+00 -7.71906972e-01 -3.22586000e-01 2.27591634e+00
8.52088690e-01 4.82476890e-01 3.27295303e-01 3.53224725e-01
1.00196767e+00 1.46511108e-01 -4.95839775e-01 -5.84233403e-01
2.07088336e-01 -1.60307661e-01 1.57421261e-01 8.17033201e-02
-9.55780685e-01 6.02774501e-01 5.35640574e+00 1.12714612e+00
-1.05065858e+00 -1.24996856e-01 4.04263139e-01 5.33763289e-01
-6.46031260e-01 1.72330916e-01 -6.69815361e-01 9.50814486e-01
7.41814733e-01 -1.38685301e-01 1.47874102e-01 8.91315579e-01
-3.14074993e-01 2.33541653e-01 -6.35182440e-01 5.13365388e-01
-5.73602915e-02 -7.87654817e-01 -8.30700397e-02 -1.12688981e-01
4.75571066e-01 -2.25980833e-01 4.75417916e-03 6.21091187e-01
2.45585799e-01 -7.50457466e-01 6.34943783e-01 3.61853361e-01
6.42185032e-01 -9.70784783e-01 9.95067120e-01 4.11858410e-01
-1.57859504e+00 -6.14657223e-01 -7.45604157e-01 2.00675968e-02
-1.79958284e-01 8.66842985e-01 -4.58252370e-01 6.98485553e-01
8.20413053e-01 4.54927266e-01 -8.56316924e-01 1.33188260e+00
-4.15711939e-01 8.04433942e-01 2.64455862e-02 -1.62366629e-01
1.67568251e-01 -2.06569046e-01 4.63935643e-01 1.10771120e+00
5.39833844e-01 -3.18962693e-01 1.77934930e-01 9.39600527e-01
8.29692483e-02 7.69894759e-05 -2.49057308e-01 -8.47677141e-02
6.95956349e-01 1.60551107e+00 -8.48347068e-01 -3.16718146e-02
-7.70796120e-01 3.32748353e-01 4.18707579e-01 4.38684314e-01
-7.03685224e-01 -1.13532448e+00 5.38199127e-01 4.83758971e-02
3.49654615e-01 4.09856178e-02 -1.47609547e-01 -9.70224738e-01
6.55886531e-02 -8.63456190e-01 6.42149031e-01 -2.52254009e-01
-1.58891857e+00 1.05477881e+00 -1.50841743e-01 -1.52349579e+00
5.39890630e-03 -4.50353324e-01 -1.37154305e+00 1.06106162e+00
-1.54655981e+00 -8.99432003e-01 -4.43270296e-01 3.71543884e-01
1.22048631e-01 -3.69049072e-01 5.28176069e-01 5.33393919e-01
-1.01861131e+00 1.23274302e+00 1.60496170e-03 3.52431387e-01
4.39427286e-01 -1.34478927e+00 5.84293127e-01 9.13877010e-01
-4.70649332e-01 7.53864706e-01 4.51729476e-01 -1.03666723e+00
-9.07366574e-01 -1.38452983e+00 5.59269786e-01 -2.20350608e-01
9.79373217e-01 -3.00160468e-01 -1.46371830e+00 4.15744454e-01
-1.83726326e-01 1.05483361e-01 5.79123676e-01 1.65740296e-01
-6.29356325e-01 8.28389227e-02 -1.26336265e+00 5.55281758e-01
6.20247662e-01 -4.61919636e-01 -4.27801698e-01 -1.88670997e-02
7.20995724e-01 -7.68070342e-03 -8.41196954e-01 3.33036780e-01
2.05977395e-01 -9.11734343e-01 8.31458092e-01 -3.37852359e-01
5.68483293e-01 -2.71884739e-01 1.08178683e-01 -1.43852913e+00
-5.49252391e-01 -2.02048406e-01 -3.11846495e-01 1.62897265e+00
4.23366070e-01 -1.02726293e+00 7.50752091e-01 3.54338259e-01
-1.06913082e-01 -9.75317061e-01 -5.64179242e-01 -8.37332666e-01
1.92468971e-01 -1.16631575e-01 1.06497681e+00 1.01847923e+00
8.79628658e-02 -8.15521702e-02 -8.04069787e-02 2.68860608e-01
6.04146123e-01 1.07993394e-01 2.42684707e-01 -1.41873813e+00
-3.82461131e-01 -7.15294003e-01 -6.43635690e-01 -4.38593626e-01
2.81542361e-01 -8.99737716e-01 -4.24271561e-02 -1.24251199e+00
2.68394262e-01 -8.13189030e-01 -6.58442318e-01 5.73338926e-01
-7.78302968e-01 2.98269093e-02 -2.18554705e-01 2.02900335e-01
-2.42635161e-01 2.89252371e-01 8.65195394e-01 -2.10389391e-01
-1.10166796e-01 4.15749282e-01 -8.21292400e-01 8.84932101e-01
8.80513012e-01 -6.05581403e-01 -7.02494264e-01 -2.76334912e-01
6.93512917e-01 7.48379752e-02 1.71369627e-01 -9.12696600e-01
-1.02466671e-02 -1.01423852e-01 8.56243744e-02 -4.51567382e-01
-3.24815124e-01 -5.48043668e-01 -1.45157278e-01 5.59867442e-01
-1.97221845e-01 2.52433121e-01 1.38888642e-01 5.75294197e-01
-2.17478678e-01 -8.37208927e-01 5.14527917e-01 3.49565484e-02
-7.59010315e-01 5.92145741e-01 -3.63783628e-01 2.32837379e-01
9.43773508e-01 -2.27334097e-01 -5.66518009e-01 -1.57019541e-01
-1.91368878e-01 3.18684965e-01 5.48441172e-01 5.37724495e-01
6.30725861e-01 -1.25157189e+00 -5.87464809e-01 3.07959169e-01
2.76653975e-01 -1.09527953e-01 5.79573691e-01 6.59644425e-01
-3.66957009e-01 -3.51306766e-01 -2.82064319e-01 -8.80226865e-02
-1.18073487e+00 8.21962357e-01 1.19775385e-01 -3.67688298e-01
-6.58028364e-01 9.20503616e-01 4.21561241e-01 -5.34346163e-01
3.10737193e-01 -8.84784609e-02 -9.17526245e-01 2.00761765e-01
7.88565934e-01 5.43301225e-01 -5.87482704e-03 -4.58903879e-01
-4.51252937e-01 6.69712365e-01 -3.97212356e-01 7.35848308e-01
1.16220820e+00 2.64737129e-01 -4.38815862e-01 7.02913180e-02
1.14154422e+00 -3.88219068e-03 -1.14517951e+00 -2.11247858e-02
1.99890643e-01 -3.20521355e-01 -2.30131112e-02 -6.91574693e-01
-1.47674739e+00 1.06591654e+00 4.06106234e-01 7.33230889e-01
1.30354142e+00 -3.25098008e-01 1.00554156e+00 -2.14297641e-02
2.94833809e-01 -7.51382530e-01 -5.20235933e-02 5.75907290e-01
6.01747572e-01 -9.38666999e-01 -1.34960070e-01 -5.08883059e-01
-3.78338486e-01 1.31548452e+00 7.98142135e-01 -5.58911636e-02
6.93961620e-01 5.41572690e-01 9.63085145e-03 -7.09413439e-02
-4.84799117e-01 2.41744161e-01 2.69489348e-01 8.23452890e-01
3.49031568e-01 1.01247653e-01 -3.00604314e-01 9.41047132e-01
4.59606759e-02 -4.96285826e-01 7.62604415e-01 7.49148488e-01
-4.55620497e-01 -1.29958355e+00 -3.92394841e-01 7.87324846e-01
-6.05645537e-01 -3.59883696e-01 -1.24975406e-01 3.82386506e-01
8.33249614e-02 9.77300286e-01 -2.21715972e-01 -8.47760797e-01
3.45369697e-01 -2.02977404e-01 -3.35024297e-02 -6.98403895e-01
-9.45482075e-01 -3.07597518e-01 -2.61841983e-01 -3.18904877e-01
3.73361170e-01 -3.92106563e-01 -1.17852795e+00 -2.43561268e-01
-4.90981311e-01 2.23433271e-01 3.25305253e-01 5.72136283e-01
1.98979795e-01 1.00436807e+00 8.69260669e-01 -5.24779439e-01
-9.86805260e-01 -9.26615953e-01 -7.40011036e-01 3.34148288e-01
1.98540509e-01 -7.31000423e-01 -7.07367063e-01 -5.88545144e-01] | [7.075276851654053, 7.759525775909424] |
d64c43c8-970e-4ae3-ab9c-5e0751e01334 | a-preliminary-exploration-of-gans-for | null | null | https://aclanthology.org/2020.emnlp-main.645 | https://aclanthology.org/2020.emnlp-main.645.pdf | A Preliminary Exploration of GANs for Keyphrase Generation | We introduce a new keyphrase generation approach using Generative Adversarial Networks (GANs). For a given document, the generator produces a sequence of keyphrases, and the discriminator distinguishes between human-curated and machine-generated keyphrases. We evaluated this approach on standard benchmark datasets. We observed that our model achieves state-of-the-art performance in the generation of abstractive keyphrases and is comparable to the best performing extractive techniques. Although we achieve promising results using GANs, they are not significantly better than the state-of-the-art generative models. To our knowledge, this is one of the first works that use GANs for keyphrase generation. We present a detailed analysis of our observations and expect that these findings would help other researchers to further study the use of GANs for the task of keyphrase generation. | ['Amanda Stent', 'Rajiv Ratn Shah', 'Rakesh Gosangi', 'Debanjan Mahata', 'Haimin Zhang', 'Avinash Swaminathan'] | null | null | null | null | emnlp-2020-11 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 3.64698410e-01 2.88919985e-01 -6.94450513e-02 2.70620346e-01
-1.13181365e+00 -1.10857105e+00 1.29882264e+00 2.14202821e-01
-2.90190816e-01 9.34913635e-01 4.42635089e-01 -5.35554409e-01
2.87899256e-01 -1.22082102e+00 -9.70035195e-01 -5.91389954e-01
1.28702149e-01 5.37714720e-01 1.08560458e-01 -5.02211511e-01
4.60670024e-01 4.34155047e-01 -1.28031087e+00 3.62549573e-01
5.70848346e-01 6.29459143e-01 -2.14607909e-01 1.11005449e+00
-9.18070748e-02 9.22632635e-01 -1.18472719e+00 -7.21221268e-01
4.37282115e-01 -7.78439641e-01 -9.66482103e-01 -1.62699133e-01
6.56583309e-01 -7.38951802e-01 -5.35398006e-01 8.54943395e-01
3.50582659e-01 1.67104714e-02 7.90113509e-01 -1.48573232e+00
-1.03994215e+00 1.00216365e+00 -1.72366351e-01 1.78504840e-01
4.99933869e-01 5.42240553e-02 1.22136819e+00 -5.94162643e-01
8.42906594e-01 1.03129327e+00 2.61809140e-01 6.78866208e-01
-1.28911710e+00 -8.59030128e-01 -1.93875983e-01 -2.10386276e-01
-9.95326221e-01 -1.49000630e-01 6.98149979e-01 -1.00759536e-01
5.43674111e-01 4.30231750e-01 7.79732883e-01 1.36286008e+00
3.66877526e-01 1.06465530e+00 1.39178586e+00 -4.47878718e-01
1.28981158e-01 2.07734723e-02 -2.51998603e-01 7.06811428e-01
3.64671290e-01 1.93821162e-01 -5.87331295e-01 -5.11435270e-01
8.75306427e-01 -1.28483340e-01 -2.09912047e-01 5.59062585e-02
-1.66942036e+00 9.94455934e-01 1.36211485e-01 4.12784785e-01
-5.80952942e-01 4.25853610e-01 2.08167180e-01 6.24863982e-01
5.57626128e-01 1.02758479e+00 -2.82478154e-01 -3.90475571e-01
-1.30936587e+00 9.33703542e-01 1.02206647e+00 8.30274224e-01
4.72598851e-01 1.47606328e-01 -5.96128464e-01 3.27026695e-01
-2.84668118e-01 5.61038613e-01 2.94059604e-01 -8.09761226e-01
3.59702826e-01 3.30491871e-01 1.67968631e-01 -8.30131590e-01
2.67416716e-01 -1.47336870e-01 -6.82818532e-01 6.17937632e-02
3.04330409e-01 -4.30678755e-01 -1.22373033e+00 1.37561190e+00
-1.58786789e-01 1.00464344e-01 1.67730182e-01 3.12622935e-01
9.84917521e-01 9.54898477e-01 -2.02299133e-01 2.08084099e-02
1.38972747e+00 -9.50408041e-01 -7.41192222e-01 1.21101983e-01
6.66971281e-02 -1.12195790e+00 1.00311267e+00 3.54645222e-01
-1.24804306e+00 -4.14205641e-01 -1.07008040e+00 1.21929675e-01
-5.71925282e-01 -1.86599549e-02 8.75925839e-01 6.24661505e-01
-1.17665088e+00 5.29761136e-01 -5.48397422e-01 -4.45538908e-01
2.66337186e-01 -1.73323005e-02 -3.21891993e-01 1.90965369e-01
-1.25604475e+00 6.35735512e-01 5.15429676e-01 -4.83076274e-01
-1.30260372e+00 -9.00252223e-01 -4.94526833e-01 -1.73453182e-01
6.27817035e-01 -9.74779189e-01 1.57273352e+00 -5.03223360e-01
-1.39654958e+00 7.65178025e-01 1.28566418e-02 -6.81581378e-01
3.85482132e-01 -3.45251739e-01 -4.02869999e-01 3.24835211e-01
8.15082863e-02 7.12507486e-01 1.18177247e+00 -1.35103452e+00
-5.61741173e-01 1.85447574e-01 3.89076233e-01 -7.22361356e-02
-3.59816581e-01 1.32826179e-01 -3.42779130e-01 -1.43275702e+00
-4.06914413e-01 -1.11252058e+00 -6.42270595e-02 -4.01925743e-01
-7.82106340e-01 -3.70445959e-02 8.64705682e-01 -7.27358103e-01
1.25738084e+00 -1.76573610e+00 5.19342832e-02 6.52633160e-02
2.41525859e-01 2.29606241e-01 -2.74896950e-01 1.09731138e+00
3.60866934e-02 5.98190606e-01 -1.27747834e-01 -2.24492908e-01
1.00596078e-01 9.48419794e-03 -1.14953160e+00 -4.09662686e-02
1.28739610e-01 1.21882176e+00 -9.87202048e-01 -3.11446607e-01
-1.81941032e-01 3.08097869e-01 -3.55213165e-01 5.53294361e-01
-6.12592578e-01 2.27613330e-01 -5.10571182e-01 3.59150589e-01
3.51974159e-01 -8.25911686e-02 -1.39515758e-01 -2.73538500e-01
1.66648433e-01 4.81625438e-01 -8.89487088e-01 1.45711899e+00
-4.21919286e-01 8.04782867e-01 -3.37058932e-01 -3.67847681e-01
7.95238137e-01 5.74428320e-01 2.82643646e-01 -1.60220981e-01
-3.02189775e-02 1.25518546e-01 -2.41621643e-01 3.17305982e-01
1.22026801e+00 3.56275663e-02 -3.17571193e-01 1.17694235e+00
2.01648280e-01 -9.91441667e-01 4.90868688e-01 8.14995289e-01
1.21957278e+00 -2.92566624e-02 2.70004362e-01 1.17092133e-02
1.68006584e-01 -1.12891480e-01 -3.48660171e-01 1.22326624e+00
5.18437982e-01 7.95317709e-01 6.81601882e-01 -5.28596818e-01
-1.21581388e+00 -1.08652484e+00 4.39134032e-01 1.00327218e+00
-3.12922567e-01 -1.08122718e+00 -9.84810054e-01 -1.03072786e+00
-5.76459952e-02 9.70069110e-01 -8.52391422e-01 -4.57111448e-02
-5.15038848e-01 -5.75172842e-01 9.78043735e-01 5.47451317e-01
4.10752833e-01 -1.38208437e+00 -3.66548121e-01 1.41329303e-01
-1.35174945e-01 -1.13877857e+00 -7.41825104e-01 -7.70701319e-02
-6.41777635e-01 -1.00818193e+00 -1.13834929e+00 -4.45363671e-01
8.24642360e-01 5.24648689e-02 1.49021578e+00 2.78478652e-01
7.75696486e-02 6.42195761e-01 -6.32026911e-01 -7.88622677e-01
-1.24790442e+00 5.01084924e-01 -3.03885520e-01 -2.34068111e-01
-5.71226254e-02 -5.23106515e-01 -5.35570920e-01 -2.07777485e-01
-1.44740081e+00 2.09773660e-01 7.25089848e-01 7.60823429e-01
3.47687751e-01 4.07139331e-01 1.07644103e-01 -1.15446615e+00
1.41067898e+00 -8.71677846e-02 -4.96646553e-01 1.60148293e-01
-8.52215290e-01 5.03353298e-01 9.23374534e-01 -4.75718468e-01
-7.89580286e-01 -5.89557230e-01 6.23240427e-04 3.62312756e-02
-1.61051765e-01 1.13424428e-01 3.99449259e-01 -2.18838509e-02
6.65001392e-01 7.31969595e-01 -3.60778153e-01 -3.24558288e-01
8.08966994e-01 4.04921234e-01 7.28083074e-01 -8.43795419e-01
1.44864392e+00 4.57868397e-01 5.82303805e-03 -5.24505794e-01
-6.78290546e-01 7.63693731e-03 -1.02519110e-01 1.60899177e-01
6.37871861e-01 -7.60664880e-01 -3.57544035e-01 7.49466121e-01
-1.21888649e+00 -3.96018118e-01 -6.04058385e-01 -1.28468394e-01
-5.86735010e-01 3.25398117e-01 -9.02466834e-01 -3.99767488e-01
-9.36955988e-01 -8.98702204e-01 1.26972079e+00 2.39124417e-01
-4.94216502e-01 -8.81563127e-01 2.04072833e-01 1.77374169e-01
4.85263437e-01 4.23985153e-01 9.03725922e-01 -8.72379124e-01
-7.89930403e-01 -4.21195000e-01 1.99296214e-02 3.20104808e-01
3.81782323e-01 1.94120318e-01 -7.03630865e-01 -3.15804005e-01
-3.28631014e-01 -4.99520600e-01 1.02753174e+00 -6.08498901e-02
1.09233224e+00 -8.29057634e-01 -2.09833100e-01 3.76510054e-01
1.08088362e+00 1.60882354e-01 9.07394767e-01 4.16656822e-01
6.13800228e-01 -3.61361429e-02 2.80375779e-01 2.59285510e-01
4.83638465e-01 4.18703824e-01 1.89981401e-01 -2.44187400e-01
-2.27431878e-01 -8.36856484e-01 3.18834960e-01 5.08384109e-01
-2.79633671e-01 -6.29563272e-01 -5.36889672e-01 6.77504241e-01
-1.50222385e+00 -1.10272431e+00 2.72917956e-01 1.78194058e+00
1.17202854e+00 3.10566157e-01 2.73340136e-01 1.67580828e-01
2.95592278e-01 6.99112415e-01 -7.82451704e-02 -5.27757823e-01
1.72813386e-02 1.01717174e+00 6.81337714e-01 2.32054219e-01
-1.03728306e+00 1.17504764e+00 7.31586933e+00 7.13724554e-01
-1.00198078e+00 -1.34990737e-01 4.71222967e-01 2.00782344e-01
-8.01863492e-01 1.19711682e-01 -7.05217123e-01 4.98603582e-01
8.37575138e-01 -2.92305410e-01 6.84865594e-01 6.62306905e-01
-3.49588215e-01 -8.73087067e-03 -1.08407378e+00 6.66706026e-01
1.02388568e-01 -1.63447940e+00 4.95970935e-01 1.53677508e-01
1.16603351e+00 -3.17397624e-01 1.51247248e-01 1.59226894e-01
1.00480986e+00 -1.05357051e+00 6.88118458e-01 5.44827282e-01
6.33330941e-01 -8.78435373e-01 4.96749640e-01 3.26277204e-02
-8.55186939e-01 3.19948107e-01 -1.07450057e-02 2.40075678e-01
5.94589971e-02 5.78950524e-01 -1.10032654e+00 6.45104229e-01
3.60526502e-01 2.84516275e-01 -7.77882099e-01 5.16441822e-01
-7.21092880e-01 8.73812735e-01 -2.61259209e-02 -2.77545244e-01
2.99874097e-01 1.35349259e-01 6.57256901e-01 1.23453367e+00
3.22649568e-01 -1.42502457e-01 1.36914670e-01 9.85111654e-01
-4.81845111e-01 -4.80873995e-02 -6.84145808e-01 -9.70720232e-01
3.67219001e-01 1.40380228e+00 -9.23354030e-01 -7.42325962e-01
-1.18931606e-01 1.28518593e+00 -1.53855696e-01 3.51474673e-01
-5.56236267e-01 -5.79782128e-01 4.84318405e-01 1.00514419e-01
4.37589467e-01 -4.25283343e-01 -6.65681958e-02 -1.21637797e+00
-4.27234434e-02 -1.54634857e+00 2.66372383e-01 -8.40662062e-01
-1.24878061e+00 6.75956786e-01 2.81612396e-01 -8.34020793e-01
-7.29668558e-01 -2.21471041e-01 -7.36639917e-01 7.77981579e-01
-1.19772625e+00 -1.36199260e+00 -1.83253556e-01 3.82044226e-01
4.10596430e-01 -3.53843689e-01 1.14984024e+00 -4.03620005e-01
2.59111375e-01 5.96469223e-01 -9.31579769e-02 4.92861331e-01
7.39866316e-01 -1.49185550e+00 1.42258549e+00 1.03447258e+00
6.72220349e-01 7.63361156e-01 9.39621389e-01 -8.09382796e-01
-1.41245961e+00 -8.05228114e-01 6.07138097e-01 -6.54703498e-01
7.92740464e-01 -5.10235250e-01 -5.24370372e-01 7.57757246e-01
8.28797877e-01 -2.93387502e-01 4.61457163e-01 -2.97097832e-01
-4.05186623e-01 1.79341406e-01 -8.62727642e-01 9.98062193e-01
7.74694026e-01 -7.13439524e-01 -7.16697514e-01 2.28715882e-01
1.00273812e+00 -6.86488867e-01 -7.67184556e-01 2.55191416e-01
6.06995642e-01 -9.09151673e-01 9.57125187e-01 -5.06740391e-01
8.39148879e-01 -1.47925094e-01 1.45066053e-01 -1.74282110e+00
-1.29461437e-01 -1.01472509e+00 -5.14702082e-01 1.25539613e+00
3.29851210e-01 -5.72890580e-01 7.44524419e-01 7.49958977e-02
3.30882102e-01 -5.71538985e-01 -1.41343579e-01 -7.63733089e-01
2.17150688e-01 -1.75513178e-01 1.02815640e+00 7.90900052e-01
-5.25569260e-01 2.71936864e-01 -6.32042706e-01 -1.89046979e-01
4.73892123e-01 1.34697720e-01 1.36122108e+00 -8.34470987e-01
-5.70240140e-01 -3.62089396e-01 -2.26637325e-03 -8.02639484e-01
1.56764403e-01 -7.73371339e-01 -2.62216717e-01 -1.72722065e+00
5.65641448e-02 4.34016511e-02 -9.38900164e-04 5.69836795e-01
-4.34103429e-01 4.81328666e-01 5.62504649e-01 3.40223871e-02
-5.48770651e-02 3.06953698e-01 1.58218014e+00 -3.44295263e-01
-5.07601649e-02 7.10601285e-02 -1.23511338e+00 2.75055170e-01
8.55603397e-01 -6.73742771e-01 -4.98411924e-01 3.25897671e-02
3.86509508e-01 -1.55387893e-01 4.93910760e-01 -7.88171113e-01
2.24806983e-02 -3.01041156e-01 2.93565989e-01 -5.42255402e-01
1.02072045e-01 -3.67249370e-01 1.17097810e-01 3.40853453e-01
-4.87835020e-01 3.30074400e-01 3.38999510e-01 2.20974967e-01
-1.28071293e-01 -8.50790516e-02 2.62024492e-01 -5.92492878e-01
-3.41081113e-01 2.90162802e-01 -6.20587230e-01 2.20393643e-01
5.92600524e-01 1.53515041e-01 -4.85852569e-01 -1.06671917e+00
1.05320187e-02 -4.84693527e-01 6.70492172e-01 5.35520196e-01
5.69663584e-01 -1.29805374e+00 -9.44050431e-01 1.70588374e-01
3.43071483e-02 -1.84997886e-01 -4.11132365e-01 2.00921353e-02
-5.71518004e-01 4.01311696e-01 -7.79432729e-02 1.48840711e-01
-9.92531002e-01 5.21151721e-01 -7.42168203e-02 -8.57991815e-01
-4.69557762e-01 6.09451830e-01 2.39626318e-02 -2.14921668e-01
-1.76699072e-01 -3.44449341e-01 2.51512527e-01 -5.63599765e-02
7.83951819e-01 1.93839490e-01 7.85334408e-02 -1.95984423e-01
7.49757886e-02 1.94729343e-01 -4.62583840e-01 -4.92732793e-01
1.22955775e+00 6.04758024e-01 -2.37342551e-01 1.73495069e-01
9.67357635e-01 5.94459534e-01 -8.48340750e-01 -1.23859957e-01
-4.86705691e-01 -2.29536548e-01 -8.75366479e-02 -9.87715423e-01
-1.01892722e+00 3.54240417e-01 1.41671404e-01 5.53880870e-01
1.17286885e+00 2.27483753e-02 1.25204587e+00 3.97536486e-01
3.54688317e-01 -8.15420389e-01 5.48196197e-01 3.96735907e-01
1.07643926e+00 -8.34335148e-01 3.10887247e-01 -6.74477071e-02
-4.86245155e-01 1.11144781e+00 2.51111418e-01 -3.49331111e-01
2.79849231e-01 5.50760746e-01 2.40683034e-02 -1.61554858e-01
-8.85504901e-01 -4.96664718e-02 6.91873789e-01 5.95816433e-01
4.10452336e-01 1.16972923e-01 -3.30323517e-01 2.20136210e-01
-8.76732230e-01 1.08181164e-01 7.80566156e-01 1.13456845e+00
3.20920907e-02 -1.63742959e+00 -5.10629535e-01 5.12347877e-01
-9.31286335e-01 -5.87808430e-01 -9.21606481e-01 8.37568283e-01
-2.66006768e-01 8.70133460e-01 -1.27955541e-01 -5.49087524e-01
1.44187808e-01 1.73594832e-01 7.10501730e-01 -6.61570191e-01
-9.13472652e-01 -2.78926134e-01 1.40182242e-01 -2.33586878e-01
-2.84065694e-01 -2.16480345e-01 -8.42188060e-01 -6.42096400e-01
-1.34845506e-02 3.30793470e-01 6.10423028e-01 6.67968333e-01
3.60312074e-01 5.34518123e-01 5.90895355e-01 -5.70744455e-01
-3.62840027e-01 -9.67522264e-01 -3.48252594e-01 6.27337575e-01
2.93147594e-01 3.44131030e-02 -1.46578103e-01 4.17682886e-01] | [12.288890838623047, 8.899931907653809] |
00e7716c-47c1-422e-8d23-d44b282c700d | crepe-open-domain-question-answering-with | 2211.17257 | null | https://arxiv.org/abs/2211.17257v1 | https://arxiv.org/pdf/2211.17257v1.pdf | CREPE: Open-Domain Question Answering with False Presuppositions | Information seeking users often pose questions with false presuppositions, especially when asking about unfamiliar topics. Most existing question answering (QA) datasets, in contrast, assume all questions have well defined answers. We introduce CREPE, a QA dataset containing a natural distribution of presupposition failures from online information-seeking forums. We find that 25% of questions contain false presuppositions, and provide annotations for these presuppositions and their corrections. Through extensive baseline experiments, we show that adaptations of existing open-domain QA models can find presuppositions moderately well, but struggle when predicting whether a presupposition is factually correct. This is in large part due to difficulty in retrieving relevant evidence passages from a large text corpus. CREPE provides a benchmark to study question answering in the wild, and our analyses provide avenues for future work in better modeling and further studying the task. | ['Hannaneh Hajishirzi', 'Luke Zettlemoyer', 'Sewon Min', 'Xinyan Velocity Yu'] | 2022-11-30 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [ 2.38678068e-01 7.48195410e-01 9.86488238e-02 -3.40660930e-01
-1.69116330e+00 -1.01369011e+00 6.88206255e-01 4.22993839e-01
-2.72635877e-01 8.67002726e-01 6.38895750e-01 -8.47880661e-01
-3.98727626e-01 -7.11979330e-01 -8.97162437e-01 1.09041654e-01
4.52532113e-01 9.19386566e-01 8.88984263e-01 -8.96165609e-01
4.71477896e-01 -2.88011581e-01 -1.05990016e+00 7.39096463e-01
1.26429832e+00 6.24713063e-01 7.66662285e-02 7.19750881e-01
-4.36999202e-01 1.32525992e+00 -6.45541489e-01 -1.32039452e+00
-1.73339382e-01 -3.65953565e-01 -1.96234620e+00 -5.87979972e-01
1.06427848e+00 -4.08513725e-01 -4.67513874e-02 9.31836247e-01
1.52985618e-01 -1.98441874e-02 5.86197078e-01 -1.06403255e+00
-1.02243209e+00 6.19899392e-01 3.35306257e-01 7.36670852e-01
1.06588149e+00 7.33748302e-02 1.68743825e+00 -7.88827896e-01
7.12968767e-01 1.14521110e+00 7.97613978e-01 6.07249439e-01
-1.04608035e+00 1.15648769e-01 -2.72222519e-01 8.98759365e-01
-9.24682975e-01 -5.61063766e-01 6.15807414e-01 -1.82982862e-01
1.10489321e+00 8.00689876e-01 -1.42942786e-01 1.15312505e+00
3.60992476e-02 8.60264421e-01 1.04430163e+00 -7.24425554e-01
-1.01862557e-01 4.67478298e-02 6.00990057e-01 4.34970170e-01
1.22989170e-01 -5.55436671e-01 -4.16977555e-01 -5.79946458e-01
2.12044716e-01 -7.17413604e-01 -6.55749261e-01 6.65152222e-02
-9.29783106e-01 6.23914957e-01 3.52199674e-01 4.57804292e-01
-4.84038502e-01 -2.11138085e-01 -1.59857616e-01 8.92131448e-01
-6.33925274e-02 1.11766303e+00 -8.90349448e-01 -3.72170776e-01
-6.60868466e-01 8.78394961e-01 1.51854467e+00 7.79267073e-01
6.24078393e-01 -1.06544113e+00 -3.43373507e-01 9.19614434e-01
3.33205611e-01 3.87073308e-01 4.04434383e-01 -1.70271325e+00
8.15954864e-01 6.43360019e-01 8.13972712e-01 -1.12561464e+00
-1.45318761e-01 -9.36484113e-02 2.15849996e-01 -7.84395874e-01
1.14456093e+00 -1.33016676e-01 -2.31602900e-02 1.65033686e+00
2.87270874e-01 -1.92436263e-01 1.32341206e-01 6.04688048e-01
8.21531057e-01 4.96980876e-01 1.50159597e-01 -6.69968948e-02
1.85951650e+00 -7.49998033e-01 -8.14555466e-01 -6.35241449e-01
6.60949767e-01 -1.01932967e+00 1.57149982e+00 2.22934946e-01
-1.27165985e+00 -4.22365889e-02 -5.46740472e-01 -7.30233014e-01
-1.42577469e-01 -4.18595672e-01 2.19491065e-01 6.17427170e-01
-9.34457958e-01 2.73988515e-01 -2.59709984e-01 -3.48167509e-01
2.44182959e-01 -3.63420129e-01 1.12743817e-01 -4.40482080e-01
-1.76146877e+00 1.21509695e+00 -1.64982360e-02 -1.14443123e-01
-6.10451519e-01 -8.87748480e-01 -8.39036167e-01 1.98624581e-01
8.28971267e-01 -7.62519717e-01 2.17872620e+00 -6.02657139e-01
-9.53029096e-01 9.82737958e-01 -5.65268815e-01 -5.93540788e-01
3.42290401e-01 -5.83777905e-01 -5.13391376e-01 3.06599885e-01
5.80447793e-01 3.01349968e-01 6.52005076e-01 -9.04417515e-01
-5.49290001e-01 -2.80207694e-01 8.25967789e-01 3.34611535e-01
6.39331788e-02 3.66071939e-01 -8.72715190e-02 -1.21052541e-01
3.65987033e-01 -3.06925148e-01 1.56290129e-01 -2.36834705e-01
-5.13219833e-01 -1.06309295e+00 1.74139470e-01 -9.16632056e-01
1.33157241e+00 -1.51347625e+00 -3.37258190e-01 -7.74846226e-02
2.29701295e-01 1.21820129e-01 -1.51069537e-01 8.39602053e-01
1.60109237e-01 4.37457651e-01 -1.20977961e-01 1.41494557e-01
3.77588868e-01 2.52570480e-01 -1.00281942e+00 -1.35992110e-01
4.79906201e-01 1.22631514e+00 -9.83952582e-01 -7.50992239e-01
-3.53553146e-01 -2.86410391e-01 -5.69785893e-01 2.73808450e-01
-1.10966229e+00 -1.01800211e-01 -4.62546557e-01 5.79858840e-01
5.63288093e-01 -6.81434095e-01 8.11481625e-02 1.43035486e-01
3.46916676e-01 1.51872933e+00 -7.98535943e-01 1.45555496e+00
-2.00046360e-01 6.33110285e-01 9.44044292e-02 -4.50005889e-01
1.71276450e-01 4.88985270e-01 -1.27441093e-01 -9.64499593e-01
3.06842662e-02 3.19303572e-01 -7.76773021e-02 -1.00715065e+00
6.77337408e-01 -2.20926166e-01 1.15965605e-01 6.41540408e-01
3.47998249e-03 -3.18107873e-01 2.35107183e-01 6.90943301e-01
1.36934388e+00 -1.88643366e-01 7.51319751e-02 -3.66413325e-01
8.38456690e-01 5.03767371e-01 9.17245075e-02 1.17883015e+00
-4.58408117e-01 2.61977941e-01 7.17448235e-01 -1.08775981e-01
-6.03372157e-01 -1.36420369e+00 -5.55474579e-01 1.29622543e+00
4.27706912e-02 -6.16280913e-01 -7.41556942e-01 -8.96692514e-01
-2.12817505e-01 1.39683962e+00 -2.99824268e-01 2.09076464e-01
-8.48894656e-01 -1.30792364e-01 7.58089304e-01 1.54752627e-01
3.75684977e-01 -1.09393787e+00 -3.19837093e-01 2.87896574e-01
-1.17427826e+00 -1.11048532e+00 -7.76771009e-02 -2.61390805e-01
-8.04297149e-01 -1.71650088e+00 -4.20914829e-01 -9.63494003e-01
1.51181087e-01 1.66329637e-01 1.96204531e+00 6.65594041e-01
3.67857903e-01 7.81860888e-01 -4.11032170e-01 -3.73389453e-01
-6.13669097e-01 -1.71854813e-02 -5.45757949e-01 -7.56965339e-01
9.12380397e-01 -3.00911516e-01 -5.05442381e-01 4.23811644e-01
-8.63287807e-01 -3.95681351e-01 6.33467063e-02 7.58037031e-01
1.26076415e-01 -2.21230924e-01 8.11139405e-01 -1.19758952e+00
1.23626924e+00 -1.03335285e+00 -3.85759026e-01 6.66651070e-01
-3.98265421e-01 9.44144651e-02 4.37119663e-01 1.44619942e-01
-1.51501644e+00 -9.89781260e-01 -8.21012259e-01 7.00536430e-01
-2.58558840e-01 5.89524090e-01 7.34710619e-02 2.65652180e-01
1.51240659e+00 7.02618733e-02 -1.82184294e-01 -5.35235643e-01
6.85637176e-01 5.48199952e-01 6.52680993e-01 -1.01526964e+00
8.52518439e-01 9.74521339e-02 -5.97626984e-01 -7.75664389e-01
-1.71000469e+00 -7.40809858e-01 -1.02900997e-01 1.90164670e-01
3.11565876e-01 -5.43211639e-01 -7.53636956e-01 -1.25098927e-02
-1.28488553e+00 -3.31540018e-01 -1.60148472e-01 -1.85454831e-01
-5.82611501e-01 1.00193870e+00 -9.02359247e-01 -7.30440438e-01
-2.89240986e-01 -3.49652976e-01 6.46774530e-01 2.87956148e-01
-7.22822428e-01 -1.08305883e+00 3.32202792e-01 1.21669614e+00
5.28285980e-01 -5.20348966e-01 1.40477812e+00 -1.07556522e+00
-6.24763846e-01 1.39321923e-01 -6.96849152e-02 2.34353617e-01
-1.00949965e-01 -4.52374548e-01 -7.83263803e-01 3.41302693e-01
4.85010237e-01 -9.53412354e-01 5.85823059e-01 -1.73835352e-01
7.28851378e-01 -8.84320676e-01 4.47328761e-02 -3.82583499e-01
1.06307793e+00 -5.42086005e-01 9.33661044e-01 5.61026275e-01
-1.48776799e-01 9.43972647e-01 6.99275434e-01 -1.09266251e-01
1.01697135e+00 1.99214071e-01 2.68487394e-01 7.79000163e-01
2.88695786e-02 -4.51169670e-01 1.34403780e-01 3.34239244e-01
5.16506016e-01 -3.15611422e-01 -1.17944050e+00 1.00417483e+00
-1.61374331e+00 -1.13380849e+00 -5.98908246e-01 1.79374230e+00
1.64254367e+00 1.43748805e-01 -1.51317105e-01 -3.76001038e-02
2.96236783e-01 -5.49674518e-02 -3.04172277e-01 -2.89377254e-02
-1.15362585e-01 5.95053554e-01 -3.79779041e-01 9.38448429e-01
-8.08263898e-01 8.21672738e-01 6.63802290e+00 5.58401167e-01
-1.35379553e-01 2.05110446e-01 3.48502189e-01 4.49432194e-01
-1.03077984e+00 3.07269514e-01 -6.02165163e-01 2.30769917e-01
1.09140241e+00 -1.25126079e-01 3.06796819e-01 7.66361892e-01
-5.52745104e-01 -2.79302090e-01 -1.25634146e+00 4.56254631e-01
2.20637128e-01 -1.28892148e+00 -1.70247257e-01 -6.54952407e-01
4.37868595e-01 1.89702362e-01 -1.01901114e-01 7.25543797e-01
3.72672558e-01 -9.10961807e-01 4.36984986e-01 5.18018842e-01
1.50682579e-03 -2.16008008e-01 8.08971047e-01 9.95604753e-01
-1.83839694e-01 1.39451884e-02 -4.67964709e-01 -3.65392923e-01
3.21450114e-01 1.84660777e-01 -9.65693295e-01 2.85777241e-01
7.65703559e-01 -6.01127148e-02 -1.03329897e+00 1.03381646e+00
-8.41334522e-01 1.03102279e+00 -5.21830499e-01 -4.26678598e-01
6.88793436e-02 1.60939306e-01 5.65747976e-01 6.74780548e-01
-1.28756270e-01 2.42063835e-01 -4.99222189e-01 1.16594315e+00
-3.19525063e-01 1.52439937e-01 -1.71895444e-01 -1.25919431e-02
6.64668679e-01 1.00020933e+00 1.38958484e-01 -3.21624875e-01
-5.85275888e-01 9.34906363e-01 6.76811635e-01 3.18716913e-01
-5.57860255e-01 -3.95289660e-01 2.97984600e-01 1.51655748e-01
7.54519850e-02 -8.49463325e-03 -7.59874955e-02 -1.41161239e+00
6.10441506e-01 -1.24299824e+00 8.56706500e-01 -1.09710014e+00
-1.85024846e+00 1.44902751e-01 -1.76793352e-01 -3.95297498e-01
-5.29106379e-01 -7.49356508e-01 -5.57851017e-01 9.46836412e-01
-2.01707983e+00 -5.79387844e-01 -2.14317009e-01 6.72320962e-01
4.36962843e-01 6.24958336e-01 9.15380597e-01 6.19938038e-02
8.35194662e-02 4.53200728e-01 -3.19426268e-01 1.31253242e-01
1.07608545e+00 -1.44739485e+00 5.02277911e-01 8.20965886e-01
5.42647064e-01 1.26187193e+00 1.05692637e+00 -6.05961859e-01
-1.26415396e+00 -4.39769506e-01 1.81120801e+00 -1.71231711e+00
9.89206135e-01 6.16606791e-03 -1.53132570e+00 1.02189970e+00
9.37248468e-01 -5.14813006e-01 9.51490164e-01 7.28548825e-01
-5.41713238e-01 3.58270198e-01 -1.16731989e+00 7.30045855e-01
8.09597313e-01 -8.57596278e-01 -2.05258584e+00 9.57985044e-01
1.01768768e+00 -5.53845346e-01 -6.92885220e-01 2.02929989e-01
-2.02052426e-02 -1.03674269e+00 9.38905418e-01 -1.09737897e+00
5.54355323e-01 -7.41284192e-02 -4.84913550e-02 -7.97454834e-01
-9.91666615e-02 -6.37406170e-01 -3.59703690e-01 1.22485268e+00
8.90968978e-01 -4.18858439e-01 7.91162550e-01 1.32496393e+00
-8.93678740e-02 -4.07407910e-01 -1.14146447e+00 -4.53713298e-01
5.26847243e-01 -7.71816373e-01 4.37542289e-01 9.67621744e-01
5.19191027e-01 8.27416539e-01 3.57254416e-01 1.97468579e-01
4.00186270e-01 5.45520447e-02 5.02609193e-01 -1.07683730e+00
-3.47582608e-01 -2.79329747e-01 3.87685686e-01 -1.69988835e+00
9.31028649e-02 -7.26541221e-01 1.62599906e-01 -1.76491237e+00
-3.57941054e-02 -4.54663485e-01 1.74800783e-01 3.65278155e-01
-6.34932280e-01 -1.78777739e-01 -2.34306142e-01 2.14827925e-01
-1.09471679e+00 4.50773239e-01 1.27445900e+00 5.41972276e-03
2.58070409e-01 1.93666652e-01 -1.25862002e+00 8.17595720e-01
7.64059961e-01 -3.11876774e-01 -5.21900833e-01 -7.12474942e-01
1.13134718e+00 1.76222593e-01 5.87565243e-01 -6.68723762e-01
5.26052117e-01 -3.34387988e-01 -9.62192342e-02 -5.14673948e-01
2.99864560e-01 -7.57858336e-01 -7.63406575e-01 1.00843057e-01
-6.60919964e-01 5.32796718e-02 9.00724381e-02 5.12176156e-01
-2.13796884e-01 -9.16041136e-01 7.41923302e-02 -3.47951114e-01
-5.16718447e-01 -3.81154001e-01 -5.93952835e-01 1.18459010e+00
2.62801796e-01 4.07255858e-01 -1.08597088e+00 -8.35486889e-01
-4.49027479e-01 6.34565592e-01 1.86843604e-01 3.21667612e-01
5.06195486e-01 -9.66635823e-01 -6.98828578e-01 -2.61158615e-01
3.16617012e-01 5.03012985e-02 2.99331635e-01 7.27892816e-01
-4.78219807e-01 8.74004185e-01 3.11009496e-01 -2.20390633e-01
-8.34034503e-01 2.85403579e-01 5.03252387e-01 -3.09619009e-01
2.58835498e-02 1.22410774e+00 -4.00108904e-01 -8.07395101e-01
4.94418433e-03 -3.67943019e-01 -3.21431011e-01 -1.43267527e-01
8.50009561e-01 1.53197378e-01 2.85531402e-01 -1.77877307e-01
-2.86367327e-01 -2.18905240e-01 -2.13932544e-01 -1.88810706e-01
8.55507910e-01 -3.93480152e-01 -5.73238134e-01 1.18663713e-01
6.50336385e-01 4.02258188e-01 -4.98331010e-01 -6.95524395e-01
5.36579132e-01 -5.11169672e-01 -5.60517251e-01 -1.32515621e+00
2.58907020e-01 6.24814451e-01 -2.75835872e-01 7.24719822e-01
7.90363491e-01 5.72875142e-01 1.07368779e+00 1.12734950e+00
3.65227312e-01 -9.64041293e-01 3.34709793e-01 9.95596111e-01
1.18347454e+00 -1.33441353e+00 -5.19448698e-01 -5.12461662e-01
-3.89158070e-01 9.14914429e-01 8.78068805e-01 1.25314444e-01
3.28947961e-01 -4.42709476e-01 2.82435745e-01 -6.71782374e-01
-1.05563962e+00 -1.25671998e-01 4.21952575e-01 5.09558141e-01
6.37700796e-01 -3.92155558e-01 -4.63442177e-01 1.03388667e+00
-8.93239915e-01 -2.80895770e-01 4.55522388e-01 9.07658577e-01
-6.98488355e-01 -1.10702670e+00 -3.62032980e-01 5.54363728e-01
-7.71082640e-01 -6.52865469e-01 -7.14058340e-01 4.26048189e-01
-5.05575776e-01 1.61888623e+00 -1.28867954e-01 1.60946205e-01
3.63548040e-01 6.06042266e-01 6.89359367e-01 -6.61385000e-01
-7.16845751e-01 -7.70358562e-01 7.17723608e-01 -5.26220262e-01
-1.18693367e-01 -7.31384814e-01 -1.06708646e+00 -3.04073960e-01
-4.19834942e-01 5.96328139e-01 3.61755900e-02 1.26634419e+00
4.84295607e-01 -1.84977710e-01 -6.61549270e-02 5.85052729e-01
-1.23562372e+00 -1.24651110e+00 1.03399776e-01 8.10406089e-01
4.18532014e-01 -1.63563043e-01 -5.04791498e-01 -9.78718922e-02] | [11.175873756408691, 8.001626014709473] |
a94809cb-aa0d-456b-8cf0-910a8fb0b8ec | fast-and-accurate-capitalization-and | 1908.02404 | null | https://arxiv.org/abs/1908.02404v1 | https://arxiv.org/pdf/1908.02404v1.pdf | Fast and Accurate Capitalization and Punctuation for Automatic Speech Recognition Using Transformer and Chunk Merging | In recent years, studies on automatic speech recognition (ASR) have shown outstanding results that reach human parity on short speech segments. However, there are still difficulties in standardizing the output of ASR such as capitalization and punctuation restoration for long-speech transcription. The problems obstruct readers to understand the ASR output semantically and also cause difficulties for natural language processing models such as NER, POS and semantic parsing. In this paper, we propose a method to restore the punctuation and capitalization for long-speech ASR transcription. The method is based on Transformer models and chunk merging that allows us to (1), build a single model that performs punctuation and capitalization in one go, and (2), perform decoding in parallel while improving the prediction accuracy. Experiments on British National Corpus showed that the proposed approach outperforms existing methods in both accuracy and decoding speed. | ['The-Loc Nguyen', 'Hien Nguyen', 'Binh Nguyen', 'Vu Bao Hung Nguyen', 'Quoc Truong Do', 'Luong Chi Mai', 'Pham Ngoc Phuong'] | 2019-08-07 | null | null | null | null | ['punctuation-restoration'] | ['natural-language-processing'] | [ 4.61777538e-01 2.93844491e-01 1.68091938e-01 -4.21313852e-01
-8.35745394e-01 -4.50447112e-01 3.06929708e-01 3.96387070e-01
-4.97948378e-01 6.31495833e-01 5.49686968e-01 -7.99953401e-01
3.20339918e-01 -3.39881063e-01 -4.57777888e-01 -3.31860185e-01
4.55176443e-01 4.33956116e-01 4.62282240e-01 -3.53082448e-01
3.15350741e-01 3.28434139e-01 -1.20746362e+00 4.60519731e-01
1.05668497e+00 5.39396286e-01 8.01429749e-01 8.00363541e-01
-5.95050871e-01 8.30689669e-01 -9.80177283e-01 -5.68352938e-01
-1.28122661e-02 -4.11892503e-01 -1.09301722e+00 7.20153898e-02
-2.58869529e-01 1.63250059e-01 -1.87667802e-01 1.01098597e+00
4.14109677e-01 2.34029368e-02 3.73921514e-01 -5.67410529e-01
-4.92071599e-01 1.33309674e+00 -1.50881648e-01 3.21862221e-01
4.84985709e-01 -3.01081568e-01 9.59782958e-01 -7.53268898e-01
3.98996502e-01 1.22601974e+00 5.00423551e-01 8.13923180e-01
-8.25419068e-01 -3.31316799e-01 1.12738699e-01 1.11115262e-01
-1.49160182e+00 -9.07374322e-01 5.22310317e-01 -1.37891635e-01
1.51121521e+00 4.48910117e-01 4.03622150e-01 6.20785654e-01
-3.70624773e-02 8.59611928e-01 9.74980831e-01 -8.79086792e-01
4.13549840e-02 1.73012857e-04 3.35694075e-01 3.90412331e-01
-1.28573433e-01 -2.69781440e-01 -4.13678020e-01 4.17348295e-01
4.90399659e-01 -2.82164961e-01 -2.32340828e-01 7.13085473e-01
-1.05755210e+00 5.21370769e-01 -1.71290830e-01 7.69618869e-01
-4.80788916e-01 -3.10930789e-01 5.58600724e-01 2.67802387e-01
4.72673237e-01 2.37680286e-01 -5.14041066e-01 -4.99584317e-01
-1.21187913e+00 -1.77210778e-01 7.21819222e-01 1.08645678e+00
5.48747122e-01 2.67954290e-01 -5.97208105e-02 1.37175095e+00
3.29579890e-01 5.81858516e-01 8.82132530e-01 -4.84586746e-01
8.37996840e-01 4.38017011e-01 -4.56901602e-02 -4.11289930e-01
-1.13091148e-01 -1.84621319e-01 -6.76221788e-01 -3.16774338e-01
3.48551124e-01 -1.13351844e-01 -1.41782093e+00 1.35488689e+00
1.67519460e-03 -3.81157920e-02 5.91483355e-01 5.60670674e-01
7.48937547e-01 1.21906602e+00 2.55552113e-01 -3.97723287e-01
1.43907249e+00 -1.10735357e+00 -1.08160174e+00 -4.00602192e-01
9.49089825e-01 -1.20603645e+00 1.10823953e+00 3.76371503e-01
-1.41908181e+00 -6.52402639e-01 -9.52148259e-01 -2.72403508e-01
-3.47948581e-01 4.18948412e-01 -3.03236879e-02 8.83826852e-01
-1.11756051e+00 4.76478368e-01 -8.87320876e-01 -3.64767790e-01
-1.73883244e-01 2.72878170e-01 -1.99905723e-01 1.13021828e-01
-1.20524740e+00 9.57075119e-01 7.01984942e-01 3.46535325e-01
-3.46670449e-01 -2.08934262e-01 -8.67217600e-01 2.58281529e-01
4.14607488e-02 -2.10866258e-01 1.66276193e+00 -1.11792195e+00
-1.87787247e+00 6.84937358e-01 -6.60008013e-01 -5.92627823e-01
3.27927507e-02 -2.96398431e-01 -6.38373613e-01 -6.16767108e-02
-3.06666434e-01 4.15485591e-01 5.46223760e-01 -9.12265420e-01
-1.03065288e+00 -5.19046962e-01 -4.43742633e-01 5.24444461e-01
-3.91510613e-02 6.89792693e-01 -3.45421582e-01 -8.21524441e-01
4.02104139e-01 -6.91131949e-01 -2.26109818e-01 -1.12426925e+00
-3.18142295e-01 -5.02771139e-01 5.37331820e-01 -1.50268507e+00
1.78497636e+00 -2.04816055e+00 7.22906888e-02 9.12407562e-02
-4.99313354e-01 8.02474618e-01 8.98604989e-02 5.52699029e-01
-1.70533270e-01 2.76040018e-01 -2.82412529e-01 -6.18312895e-01
-2.03564465e-01 5.80010772e-01 -5.78750849e-01 -7.68705234e-02
1.68560117e-01 8.09711099e-01 -5.71131527e-01 -4.14549440e-01
2.28342459e-01 3.16857785e-01 -2.00610369e-01 3.56833369e-01
-6.14899099e-02 1.12619959e-01 -2.04254929e-02 5.32145143e-01
3.72616410e-01 5.71858644e-01 3.05527091e-01 2.01548442e-01
-4.23401356e-01 1.36990321e+00 -1.16938901e+00 1.34757161e+00
-6.22639775e-01 3.71517450e-01 6.80910572e-02 -1.18893552e+00
1.03007066e+00 7.78273225e-01 -1.15965106e-01 -7.19038725e-01
1.72935307e-01 6.25003576e-01 1.17636360e-01 -3.31736326e-01
1.06329906e+00 -3.52861553e-01 -1.73501283e-01 1.38392821e-01
-4.95658778e-02 -2.55977631e-01 1.08923085e-01 3.86202782e-02
8.31132531e-01 -1.03007585e-01 4.80816603e-01 4.34643701e-02
8.52308095e-01 6.57510459e-02 6.15397394e-01 3.02685887e-01
1.53385401e-01 6.54640019e-01 2.97329903e-01 8.09832811e-02
-1.32676625e+00 -8.10067058e-01 2.30643973e-01 9.74784255e-01
-2.58223027e-01 -4.57680315e-01 -1.13691866e+00 -4.99454856e-01
-5.66508889e-01 1.12129748e+00 2.06262976e-01 2.07471907e-01
-1.25112414e+00 -3.34426671e-01 9.66503859e-01 5.95700204e-01
1.88702747e-01 -1.38996577e+00 8.77878293e-02 6.54795587e-01
-4.93219137e-01 -1.29793108e+00 -4.49931741e-01 3.83827418e-01
-9.95389044e-01 -3.75244558e-01 -7.83577144e-01 -1.18689704e+00
4.81545299e-01 2.27662340e-01 5.79362154e-01 3.25878173e-01
1.81564689e-01 -1.97505116e-01 -8.52651119e-01 -2.90125191e-01
-1.20371175e+00 4.33345228e-01 -6.00039884e-02 -2.43464902e-01
2.67231286e-01 -3.30489606e-01 1.21784084e-01 2.95073211e-01
-8.22014093e-01 -6.88008815e-02 6.59416735e-01 5.63037753e-01
5.29211700e-01 2.18722578e-02 7.95111835e-01 -9.17484045e-01
6.46724343e-01 -1.00908346e-01 -4.58190560e-01 4.19317484e-01
-6.63624644e-01 1.17031828e-01 9.17806745e-01 -9.09675434e-02
-1.42378020e+00 5.74641936e-02 -1.03825152e+00 6.78970069e-02
-3.24764550e-01 5.12309968e-01 -5.14332235e-01 3.27451348e-01
3.95529151e-01 6.32897556e-01 -1.26598999e-01 -9.38563466e-01
2.15036198e-01 1.55080795e+00 5.71248233e-01 -3.37403297e-01
5.05711436e-01 -3.56327325e-01 -7.42980778e-01 -1.31092250e+00
-5.98195851e-01 -7.78433144e-01 -6.22180820e-01 8.30208510e-02
8.14601243e-01 -8.11625957e-01 -2.34956264e-01 7.56085157e-01
-1.44093335e+00 -1.51085332e-01 -1.55624703e-01 5.09170115e-01
-3.49138826e-01 8.11254203e-01 -8.54690313e-01 -8.90574753e-01
-5.25339425e-01 -1.05143929e+00 8.03764403e-01 2.30647743e-01
-1.94572061e-01 -7.90955305e-01 -1.10153288e-01 6.90509558e-01
2.97625005e-01 -9.14302707e-01 8.64222169e-01 -8.39176416e-01
-4.18068528e-01 8.03905204e-02 1.56155694e-03 8.97099853e-01
6.29161149e-02 -3.93698849e-02 -8.34475935e-01 4.43003513e-02
-1.05113424e-01 1.20337285e-01 5.64606488e-01 1.69208661e-01
8.39490235e-01 -4.46610063e-01 -1.22614764e-01 3.79314721e-01
9.90415394e-01 7.18465567e-01 1.19701576e+00 1.22513235e-01
4.93372828e-01 6.59457028e-01 7.23064065e-01 5.74391223e-02
4.01873469e-01 4.86294150e-01 -7.55093843e-02 9.12606716e-02
-4.23493385e-01 -5.39981842e-01 7.53684759e-01 1.82073140e+00
-3.87256294e-02 -4.31891322e-01 -9.77134764e-01 7.50342727e-01
-1.71362960e+00 -8.52673888e-01 -3.09819192e-01 2.18725657e+00
1.02992260e+00 2.29651332e-01 -6.34046271e-02 4.09821421e-01
9.96597230e-01 1.56895712e-01 3.05025220e-01 -1.10818827e+00
-2.10315183e-01 5.23202181e-01 5.70300162e-01 9.07893479e-01
-6.43891275e-01 1.66004443e+00 6.25711727e+00 1.06709647e+00
-1.18570828e+00 2.34040588e-01 3.70761901e-01 5.35595179e-01
-3.85061294e-01 1.98761478e-01 -1.25755978e+00 4.19001460e-01
1.49091756e+00 -5.69141470e-02 5.53175688e-01 6.61360979e-01
2.97968328e-01 2.91072670e-02 -7.00553715e-01 9.15649414e-01
6.05076365e-02 -1.02858806e+00 8.59516412e-02 -3.81874442e-01
2.96148241e-01 3.21928039e-02 -5.46475112e-01 4.24419910e-01
9.83443707e-02 -9.65588868e-01 9.98537838e-01 8.27951357e-02
6.37601852e-01 -7.82083988e-01 8.51439297e-01 5.65774262e-01
-1.26244009e+00 7.43222535e-02 -3.77039313e-01 -2.01482147e-01
5.40954709e-01 2.92879671e-01 -1.22828925e+00 6.26937807e-01
2.52270877e-01 2.78445631e-01 -3.67916763e-01 1.08186233e+00
-5.99372625e-01 1.10248494e+00 -2.69142300e-01 -2.74760664e-01
1.99910447e-01 -3.26213509e-01 5.42736411e-01 1.69347131e+00
5.82875788e-01 3.42769355e-01 -1.34787247e-01 2.29590818e-01
6.95947558e-02 6.34208500e-01 -1.20388061e-01 -3.81849945e-01
6.58531964e-01 8.06913197e-01 -8.05037916e-01 -4.89640266e-01
-2.61193365e-01 1.12042499e+00 3.39855790e-01 2.25488380e-01
-5.50618768e-01 -6.24904156e-01 5.26664197e-01 1.96028158e-01
4.48553801e-01 -6.12893045e-01 -4.26264822e-01 -1.03282845e+00
2.57996112e-01 -8.64865422e-01 1.41684726e-01 -6.77735507e-01
-6.72687471e-01 9.15719032e-01 -3.21997046e-01 -6.68699920e-01
-2.25982368e-01 -5.64252079e-01 -4.28573042e-01 1.09400654e+00
-1.69886541e+00 -9.35693443e-01 3.26889843e-01 3.15049887e-01
1.01799273e+00 -9.35075954e-02 7.61061251e-01 5.19754529e-01
-6.81990802e-01 5.56255639e-01 9.69198998e-03 2.16832235e-01
3.23515743e-01 -1.15768766e+00 6.34563506e-01 1.22993064e+00
2.95610547e-01 5.46565294e-01 5.96705556e-01 -7.09994435e-01
-9.40173864e-01 -8.97222817e-01 1.85065389e+00 -1.93220805e-02
5.68164587e-01 -3.36589426e-01 -1.06834424e+00 7.47714102e-01
2.83212423e-01 -5.90068877e-01 4.76562858e-01 5.64483479e-02
-4.34495090e-03 -1.65263459e-01 -8.82597506e-01 5.11385739e-01
8.60889137e-01 -6.55650556e-01 -1.27098989e+00 2.58993246e-02
8.67484331e-01 -3.55842084e-01 -6.43419445e-01 7.83313885e-02
2.95423090e-01 -6.35512829e-01 4.05205131e-01 -3.99029881e-01
3.57558168e-02 -4.21170652e-01 -2.70491689e-01 -1.41850758e+00
-9.90135446e-02 -9.70200002e-01 3.70004714e-01 1.58984566e+00
7.65664458e-01 -5.84988713e-01 4.62608069e-01 4.34832543e-01
-8.90833139e-01 -3.71406913e-01 -1.05199313e+00 -8.79933298e-01
-1.73456252e-01 -6.09573543e-01 5.76310813e-01 4.30465341e-01
2.63654321e-01 5.24475932e-01 -1.91634059e-01 3.17947745e-01
2.22107396e-02 -2.83530295e-01 3.30874741e-01 -9.80874240e-01
-1.39412254e-01 -2.91333795e-01 -2.98140138e-01 -1.47864830e+00
2.18077138e-01 -8.58531713e-01 4.07742262e-01 -1.66571951e+00
-2.64766097e-01 -6.71547592e-01 -9.47032273e-02 7.56454468e-01
-3.53064276e-02 -1.15670674e-01 2.62947857e-01 7.59562477e-02
-2.62814134e-01 5.32415867e-01 7.12012887e-01 2.53689289e-01
-3.88987690e-01 2.10966557e-01 -5.73791862e-01 6.24274373e-01
9.29790735e-01 -6.71776235e-01 -5.17617166e-02 -5.70837915e-01
-6.44882470e-02 4.48560208e-01 -3.28634292e-01 -9.16099727e-01
1.41916409e-01 4.29628929e-03 -1.14101164e-01 -7.72305369e-01
1.42150864e-01 -6.90258563e-01 6.09378815e-02 3.24394047e-01
-1.43997520e-01 2.46876538e-01 2.99062673e-02 -1.07869469e-02
-5.11846781e-01 -8.45316172e-01 8.12195182e-01 -4.15438041e-02
-7.17207670e-01 -2.27791354e-01 -7.66057909e-01 1.13480709e-01
6.90389395e-01 -4.31420416e-01 -1.62112653e-01 -1.26790747e-01
-8.02874446e-01 -1.44416720e-01 3.52703005e-01 5.18620670e-01
6.13773704e-01 -7.98889339e-01 -5.94674885e-01 4.53694880e-01
-2.32161835e-01 4.47528064e-02 8.41601118e-02 5.89630663e-01
-7.66802669e-01 7.17902005e-01 5.05222529e-02 -1.57797903e-01
-1.63278818e+00 3.38310570e-01 1.05262846e-01 -3.18617493e-01
-4.82632667e-01 7.41274536e-01 -2.41652608e-01 -3.10294390e-01
4.00925040e-01 -6.74883008e-01 -3.30930650e-01 -6.92874268e-02
6.00169361e-01 3.30929935e-01 5.11881471e-01 -1.02758503e+00
-3.54398549e-01 3.41196448e-01 -2.33263567e-01 -3.31718922e-01
1.10080802e+00 -5.86744368e-01 -1.55399755e-01 3.97008240e-01
7.81488240e-01 4.00301993e-01 -5.49842298e-01 -6.16928637e-02
3.85868728e-01 -1.32464275e-01 3.46049257e-02 -7.85650253e-01
-5.98549247e-01 1.08143604e+00 2.50021908e-02 4.64675426e-01
1.01203215e+00 -3.12318914e-02 1.38786232e+00 4.91569161e-01
6.13138378e-02 -1.52836597e+00 -5.49421012e-01 1.17268634e+00
5.99259615e-01 -7.24991024e-01 -6.58686876e-01 -5.26664495e-01
-7.77900636e-01 1.18868327e+00 2.73649663e-01 2.43882596e-01
4.21787083e-01 3.59054297e-01 2.47186840e-01 3.57385367e-01
-6.14507914e-01 -3.16027731e-01 7.24135488e-02 3.85199398e-01
6.40543401e-01 3.16304445e-01 -8.34887028e-01 7.02853262e-01
-6.43154740e-01 -2.10665926e-01 6.69855118e-01 8.14560652e-01
-1.01143563e+00 -1.56859255e+00 -4.06081647e-01 3.11625414e-02
-8.64987135e-01 -4.32025820e-01 -2.28652358e-01 3.27377528e-01
-4.81516384e-02 1.13710570e+00 3.60996276e-03 -3.03793073e-01
4.85169411e-01 5.15843093e-01 3.19269001e-01 -9.80544329e-01
-7.32444525e-01 3.37011844e-01 5.28664231e-01 -1.68556288e-01
-4.87183146e-02 -7.02163696e-01 -1.67727470e+00 7.23747984e-02
-4.53096211e-01 6.42503560e-01 1.07458055e+00 1.05477154e+00
2.38598302e-01 4.24788743e-01 6.61096990e-01 -4.76446599e-01
-5.84402919e-01 -1.32681692e+00 -6.39863431e-01 9.95589718e-02
1.02505639e-01 -3.55837531e-02 -3.64491105e-01 2.43496642e-01] | [14.24354076385498, 7.102929592132568] |
938a2c09-9123-4d3d-9350-e83d4859f630 | towards-automated-survey-variable-search-and | 2209.06804 | null | https://arxiv.org/abs/2209.06804v1 | https://arxiv.org/pdf/2209.06804v1.pdf | Towards Automated Survey Variable Search and Summarization in Social Science Publications | Nowadays there is a growing trend in many scientific disciplines to support researchers by providing enhanced information access through linking of publications and underlying datasets, so as to support research with infrastructure to enhance reproducibility and reusability of research results. In this research note, we present an overview of an ongoing research project, named VADIS (VAriable Detection, Interlinking and Summarization), that aims at developing technology and infrastructure for enhanced information access in the Social Sciences via search and summarization of publications on the basis of automatic identification and indexing of survey variables in text. We provide an overview of the overarching vision underlying our project, its main components, and related challenges, as well as a thorough discussion of how these are meant to address the limitations of current information access systems for publications in the Social Sciences. We show how this goal can be concretely implemented in an end-user system by presenting a search prototype, which is based on user requirements collected from qualitative interviews with empirical Social Science researchers. | ['Andrea Zielinski', 'Benjamin Zapilko', 'Simone Paolo Ponzetto', 'Philipp Mayr', 'Henning Kroll', 'Kai Eckert', 'Tornike Tsereteli', 'Sotaro Takeshita', 'Yavuz Selim Kartal'] | 2022-09-14 | null | null | null | null | ['variable-detection'] | ['natural-language-processing'] | [ 8.70786384e-02 2.29178861e-01 -5.01346052e-01 -2.11387873e-01
-5.79649389e-01 -9.40368950e-01 7.66930342e-01 8.47562611e-01
-3.73931229e-01 7.41537154e-01 6.97162211e-01 -7.07609832e-01
-6.68914497e-01 -4.76413637e-01 4.36477624e-02 -1.67096406e-02
4.66682613e-01 5.69609344e-01 -1.32838622e-01 -3.54706608e-02
9.76222157e-01 6.66474104e-01 -1.78869474e+00 -1.10097669e-01
8.55203152e-01 4.29928929e-01 4.60923910e-01 4.11533266e-01
-8.96160841e-01 1.83260530e-01 -8.62734079e-01 -3.22364688e-01
-3.35124284e-01 -3.06723475e-01 -1.08296847e+00 -2.80712605e-01
4.76552963e-01 1.78515792e-01 2.51785427e-01 1.16032505e+00
5.45806646e-01 1.06404396e-03 6.18147969e-01 -1.30774224e+00
-7.26179540e-01 8.34561348e-01 -1.97811142e-01 1.74051493e-01
9.77608025e-01 -4.53371435e-01 1.08446407e+00 -7.90185392e-01
1.27757812e+00 1.36007571e+00 3.46846670e-01 5.59497736e-02
-1.27947950e+00 -7.46567130e-01 2.08534971e-02 2.59133548e-01
-1.10971808e+00 -4.73020732e-01 7.43102074e-01 -7.39048064e-01
9.12546158e-01 7.64484227e-01 5.31718314e-01 1.01753223e+00
-3.89321484e-02 2.10868120e-01 1.02930927e+00 -6.71373725e-01
2.52141982e-01 6.98096275e-01 7.32216775e-01 2.72100359e-01
8.12689126e-01 -5.64308643e-01 -5.93719423e-01 -5.18761754e-01
5.73823810e-01 -1.14543840e-01 3.70078534e-02 -3.28792155e-01
-1.39749002e+00 6.60330355e-01 -1.90600529e-02 8.19886804e-01
-7.13341951e-01 -5.90570390e-01 3.52458239e-01 5.50348580e-01
6.74120903e-01 9.02352571e-01 -4.81159121e-01 -4.51283067e-01
-1.37502027e+00 5.66621125e-01 1.24833930e+00 1.06699705e+00
5.99170446e-01 -4.98469621e-01 -3.96141857e-01 8.70366752e-01
5.87033629e-01 6.77839875e-01 3.41993004e-01 -1.29152131e+00
2.67656177e-01 1.14817822e+00 1.09276779e-01 -1.28785408e+00
-6.22341931e-01 -3.79455805e-01 -5.07407248e-01 -3.97092029e-02
2.13828713e-01 -1.23961553e-01 -3.84871274e-01 1.25738370e+00
3.90490681e-01 -7.05320418e-01 -2.42912665e-01 5.22008359e-01
1.58198667e+00 2.10976496e-01 1.48078546e-01 -6.28683627e-01
1.67181206e+00 -2.36703664e-01 -1.37033045e+00 1.59998029e-01
5.63929498e-01 -1.05309057e+00 8.16001296e-01 1.41277701e-01
-1.28543878e+00 -3.44697833e-01 -6.31684601e-01 -3.21059883e-01
-1.05900025e+00 -2.16245860e-01 4.01258945e-01 6.60477400e-01
-1.19919348e+00 4.65396523e-01 -5.11185825e-01 -9.66512203e-01
3.13263327e-01 -1.42786175e-01 -3.33746135e-01 3.75145108e-01
-8.96919191e-01 9.18228984e-01 3.79524231e-02 -5.38805127e-01
1.63036868e-01 -9.77589190e-01 -6.15531623e-01 6.79833442e-02
3.24533612e-01 -6.72265470e-01 1.01944447e+00 -5.63452423e-01
-1.01609099e+00 1.01934814e+00 -4.36109513e-01 2.33919173e-01
2.74206042e-01 -3.75062339e-02 -4.12113607e-01 1.45894840e-01
5.76004207e-01 5.02083153e-02 -5.26544303e-02 -1.22814322e+00
-4.61325854e-01 -8.22111189e-01 -5.21530151e-01 -8.12002122e-02
-7.93810368e-01 8.80058229e-01 -4.28710401e-01 -6.67337358e-01
-5.88837340e-02 -4.29617882e-01 -2.09443703e-01 -4.33971398e-02
-1.72363818e-01 -4.77046371e-01 6.36661530e-01 -9.96367812e-01
1.77411973e+00 -1.67710173e+00 3.61610681e-01 6.43343866e-01
4.51547921e-01 1.45929679e-01 6.30806983e-02 1.06457138e+00
2.04290077e-01 6.02920234e-01 -4.68602404e-03 -2.88210630e-01
1.99440837e-01 -8.81276950e-02 -1.96534991e-01 2.49018446e-01
-3.70931178e-01 6.53841436e-01 -1.21429384e+00 -8.04703295e-01
2.47264609e-01 4.43088442e-01 -9.64981541e-02 3.34205218e-02
-2.09873896e-02 1.84835404e-01 -5.98497570e-01 6.45392656e-01
6.31389081e-01 -1.25681847e-01 4.02536571e-01 3.04979712e-01
-1.02677333e+00 3.79826397e-01 -1.40187407e+00 1.42023313e+00
-1.40190154e-01 8.66378546e-01 3.85405242e-01 -1.04123843e+00
1.03145945e+00 4.53065783e-01 6.69973314e-01 -6.24259472e-01
5.33939302e-02 3.48267287e-01 -6.56063974e-01 -7.68276751e-01
7.97845542e-01 6.75432205e-01 2.27191411e-02 6.13875210e-01
4.33872044e-02 -2.08545163e-01 5.95302701e-01 4.42929447e-01
7.73544014e-01 4.74616438e-02 5.54420412e-01 -5.82404077e-01
7.99877286e-01 5.88152632e-02 9.54345688e-02 9.27643597e-01
9.07775685e-02 1.17233858e-01 6.34782195e-01 -8.24542195e-02
-1.07922244e+00 -4.46403533e-01 -5.61067402e-01 9.68943775e-01
-3.92789453e-01 -9.06302691e-01 -7.81533659e-01 -1.65198311e-01
4.70464192e-02 8.32827330e-01 -5.34368217e-01 5.35744011e-01
8.78636390e-02 -2.31125414e-01 -4.06680144e-02 -1.67835906e-01
2.96950817e-01 -1.26673484e+00 -7.57625163e-01 5.37274852e-02
-1.95711598e-01 -8.25816274e-01 3.13973159e-01 -4.03177798e-01
-8.24951947e-01 -1.03186345e+00 -9.73887801e-01 -7.12656140e-01
4.03308153e-01 3.62943381e-01 1.22745025e+00 1.04069240e-01
-3.40172768e-01 6.11702442e-01 -4.54755217e-01 -7.83174992e-01
-3.01358491e-01 5.05650401e-01 -2.75273055e-01 -8.54306459e-01
7.06415653e-01 -4.46821958e-01 -3.89593363e-01 -8.92372578e-02
-9.88203228e-01 -2.78535873e-01 2.80368835e-01 3.29311550e-01
2.28983819e-01 -2.98862815e-01 8.07989597e-01 -9.84950602e-01
1.07369304e+00 -9.06838715e-01 -7.73878694e-01 3.62906069e-01
-1.29124510e+00 -6.52900413e-02 7.66042843e-02 1.03931174e-01
-9.84370828e-01 -6.14084601e-01 6.89115841e-03 3.73586118e-01
-2.86808103e-01 1.30147839e+00 9.40583497e-02 -1.01421058e-01
5.27264059e-01 -3.53546947e-01 2.10598171e-01 -1.10175276e+00
4.62325513e-01 1.36964130e+00 1.30491167e-01 -2.97497898e-01
3.57010394e-01 1.00359611e-01 1.06114089e-01 -1.01659417e+00
-4.39720958e-01 -8.50458384e-01 -5.19035697e-01 -5.05793869e-01
3.84149283e-01 -5.00464261e-01 -8.21528018e-01 -8.19301382e-02
-9.84939933e-01 1.58400759e-01 -3.21195394e-01 5.25159597e-01
-7.79528618e-02 1.99801341e-01 6.32670075e-02 -8.80203605e-01
-5.10326743e-01 -8.91479850e-01 8.31822634e-01 3.37252319e-01
-8.50734055e-01 -1.25140107e+00 4.18874443e-01 6.36361241e-01
7.72351265e-01 2.97323495e-01 8.95142376e-01 -7.16865361e-01
3.91574614e-02 -2.19695032e-01 -3.04889143e-01 -2.15147421e-01
-4.94493730e-02 6.59874141e-01 -6.32457972e-01 -1.40264124e-01
-5.01060903e-01 9.59990919e-02 4.04213071e-01 4.94700253e-01
1.08967602e+00 -3.82301748e-01 -6.65637195e-01 9.20474827e-02
1.36006248e+00 -1.46223664e-01 3.30502182e-01 9.21786070e-01
4.09540862e-01 1.05516243e+00 4.79626507e-01 6.15235150e-01
5.39689064e-01 5.69691360e-01 1.18379556e-01 -1.36312917e-01
-9.12373811e-02 1.54777840e-01 -2.08831057e-01 8.31602633e-01
-2.91323304e-01 2.41784696e-02 -1.19957757e+00 8.59917521e-01
-1.73076308e+00 -9.56549346e-01 -6.28875077e-01 2.20541930e+00
9.87734556e-01 -3.47325116e-01 2.88908452e-01 7.54297599e-02
5.92342079e-01 1.49510838e-02 5.44168316e-02 -8.15906286e-01
1.15426108e-01 3.25344503e-01 5.09271085e-01 5.90709448e-01
-5.06161094e-01 7.45501578e-01 7.06102467e+00 3.22627664e-01
-6.43244445e-01 -1.01456746e-01 3.20944339e-02 9.44659337e-02
-8.48963261e-01 1.80723295e-01 -7.27288663e-01 2.31007069e-01
1.16868520e+00 -8.06201935e-01 3.37716103e-01 5.53637922e-01
6.16970360e-01 -1.78742986e-02 -5.16601622e-01 4.74129766e-01
-3.23652625e-02 -1.67334235e+00 4.90511442e-03 1.29273131e-01
7.55101025e-01 -1.51358955e-02 -2.28949532e-01 3.59647200e-02
3.20057094e-01 -6.57212913e-01 5.23335636e-01 7.02579856e-01
7.58920729e-01 -5.06886601e-01 7.18746185e-01 -1.40628535e-02
-7.90815771e-01 1.35137156e-01 -2.55857985e-02 -2.46237919e-01
-1.61165744e-02 7.38857031e-01 -5.16659796e-01 9.30354178e-01
9.57475066e-01 9.32414353e-01 -6.24656379e-01 1.31083965e+00
-8.72366596e-03 5.05281150e-01 1.14993073e-01 -4.37510341e-01
-1.99741870e-01 -2.80353457e-01 8.50748420e-01 1.54753768e+00
3.14259261e-01 -1.03311762e-01 -1.76167205e-01 9.36280906e-01
-3.52785736e-02 5.30825078e-01 -5.83622634e-01 -3.35034817e-01
1.07382405e+00 1.39989495e+00 -7.12031722e-01 -3.11649859e-01
-6.01529300e-01 1.53174803e-01 1.23179190e-01 5.18231153e-01
3.55244651e-02 -9.76711214e-01 4.55116808e-01 1.10056341e-01
3.13826352e-02 -2.72881210e-01 -6.82974041e-01 -7.97049284e-01
-1.19954698e-01 -1.08037949e+00 4.65315193e-01 -6.59852803e-01
-8.92124712e-01 1.86101615e-01 4.81449485e-01 -4.82141674e-01
-3.21517318e-01 -2.46037200e-01 2.23401748e-02 1.14113522e+00
-1.14101744e+00 -8.60958517e-01 -4.28471059e-01 1.87865362e-01
2.00994655e-01 -5.93666136e-02 9.63080704e-01 4.27908182e-01
-6.30272329e-01 1.42458230e-01 4.85756159e-01 -3.38022351e-01
8.70355606e-01 -1.30653703e+00 3.25819999e-01 4.43377584e-01
-2.52178609e-01 1.03110504e+00 9.50388014e-01 -9.85048115e-01
-1.31444108e+00 -4.71853256e-01 1.71633899e+00 -4.57019240e-01
1.01512456e+00 -1.70814052e-01 -8.86767805e-01 5.10035455e-01
6.13423705e-01 -9.41996515e-01 9.21244502e-01 5.61233163e-01
9.58524197e-02 2.03015551e-01 -1.19811690e+00 7.40242302e-01
8.67045164e-01 -3.23886842e-01 -7.84291625e-01 3.98475468e-01
3.95456433e-01 1.38313277e-03 -1.35519052e+00 -9.05945431e-03
7.83141315e-01 -3.11853945e-01 9.53466415e-01 -5.52324653e-01
3.92089754e-01 5.45808189e-02 2.50812352e-01 -1.01335764e+00
-5.16002774e-01 -7.62206435e-01 9.24163386e-02 1.85703278e+00
2.82779932e-01 -7.78233111e-01 2.71260798e-01 9.13088799e-01
1.05294816e-01 -7.81192482e-02 -8.72131526e-01 -2.46022508e-01
1.62660137e-01 -2.60004222e-01 6.83676004e-01 1.33470631e+00
5.31552792e-01 2.67352343e-01 3.44846934e-01 -2.23103806e-01
6.40858054e-01 4.13219333e-02 8.38906288e-01 -1.95934248e+00
5.83899558e-01 -9.00767207e-01 -3.52309525e-01 -2.62683719e-01
3.41388620e-02 -8.82447839e-01 -5.30283034e-01 -2.22614026e+00
5.20138562e-01 -7.23047704e-02 -1.92973003e-01 2.47272670e-01
-9.49766263e-02 -3.37283537e-02 -1.45343542e-01 4.27369684e-01
-5.06436050e-01 -4.49217716e-03 9.66207922e-01 4.06823903e-01
-5.01551032e-01 -2.19387919e-01 -1.19301987e+00 3.57740223e-01
5.54545164e-01 -5.03992856e-01 -2.01698747e-02 -3.88498418e-02
8.55865479e-01 -1.58565953e-01 3.25831503e-01 -5.76581240e-01
3.54729623e-01 -5.40483654e-01 8.08732361e-02 -8.89144957e-01
-3.26713711e-01 -8.46220136e-01 3.98605585e-01 1.63656905e-01
-7.31173933e-01 2.97876745e-01 2.64947623e-01 -2.89839245e-02
-7.96202421e-02 -4.98525947e-01 2.00785156e-02 -6.24220073e-02
-2.17973739e-01 -2.42482319e-01 -7.05184340e-01 -5.64181060e-02
7.42173612e-01 2.05270853e-02 -6.58833921e-01 -3.14726472e-01
-3.02896231e-01 4.67480391e-01 5.84778190e-01 5.39197922e-01
2.47351453e-01 -8.74715090e-01 -7.68653393e-01 1.76215157e-01
1.25437990e-01 -4.80202287e-01 -1.77438810e-01 6.59777701e-01
-4.23146158e-01 6.70348644e-01 -4.10244465e-01 2.72175912e-02
-1.61820459e+00 4.83983666e-01 -2.95675009e-01 -1.01838134e-01
-4.90606517e-01 3.69031101e-01 -6.78439558e-01 -3.33018333e-01
5.91423810e-01 -1.06547594e-01 -9.95139003e-01 6.24717355e-01
6.69528425e-01 8.50372076e-01 1.11823268e-01 -5.59819043e-01
-5.33226311e-01 4.00114775e-01 1.20867267e-01 -5.29552817e-01
1.66356707e+00 -5.36708474e-01 -6.63707197e-01 7.67376602e-01
9.60436761e-01 2.62457490e-01 -2.72786826e-01 -3.81122112e-01
3.72538209e-01 -3.28412443e-01 6.23401284e-01 -9.27529275e-01
-7.78201520e-01 3.67986083e-01 3.51876318e-01 8.76244426e-01
7.02915072e-01 3.00636202e-01 -3.61090302e-02 3.42927426e-01
-1.35255694e-01 -1.45707178e+00 -5.43105245e-01 1.53517038e-01
1.22231889e+00 -1.08562446e+00 6.83495462e-01 -3.21573377e-01
-2.71297634e-01 1.41845858e+00 -3.07064801e-01 5.44443667e-01
9.62502003e-01 2.22531669e-02 -1.24959918e-02 -8.60721469e-01
-4.45349783e-01 -2.13838533e-01 3.99542958e-01 5.34886599e-01
9.74924445e-01 -1.13498889e-01 -1.26389754e+00 3.94175023e-01
-3.79813462e-01 3.25962663e-01 3.10024083e-01 9.63121414e-01
-4.39079791e-01 -1.30272055e+00 -6.94079041e-01 6.81080163e-01
-9.10708487e-01 1.86446868e-02 -9.88733292e-01 7.53510773e-01
-3.24732542e-01 1.26898885e+00 9.51611698e-02 1.38302088e-01
5.00229657e-01 -5.81130497e-02 5.95977604e-02 -5.83677351e-01
-8.85002732e-01 -1.25339046e-01 3.95956904e-01 -2.39613101e-01
-7.17466414e-01 -1.22888184e+00 -7.69524276e-01 -5.61054230e-01
-1.84459046e-01 5.22183776e-01 1.36323977e+00 6.65030837e-01
7.40256190e-01 6.64949238e-01 3.35936487e-01 -6.80760562e-01
-1.18403204e-01 -1.17690969e+00 -5.40947795e-01 1.34834245e-01
2.40235627e-01 -6.61348939e-01 -4.84342545e-01 7.19832480e-02] | [9.649813652038574, 8.287088394165039] |
4e3ffd9a-1d5b-4fe9-8817-1bd38a22a810 | lissnas-locality-based-iterative-search-space | 2307.03110 | null | https://arxiv.org/abs/2307.03110v1 | https://arxiv.org/pdf/2307.03110v1.pdf | LISSNAS: Locality-based Iterative Search Space Shrinkage for Neural Architecture Search | Search spaces hallmark the advancement of Neural Architecture Search (NAS). Large and complex search spaces with versatile building operators and structures provide more opportunities to brew promising architectures, yet pose severe challenges on efficient exploration and exploitation. Subsequently, several search space shrinkage methods optimize by selecting a single sub-region that contains some well-performing networks. Small performance and efficiency gains are observed with these methods but such techniques leave room for significantly improved search performance and are ineffective at retaining architectural diversity. We propose LISSNAS, an automated algorithm that shrinks a large space into a diverse, small search space with SOTA search performance. Our approach leverages locality, the relationship between structural and performance similarity, to efficiently extract many pockets of well-performing networks. We showcase our method on an array of search spaces spanning various sizes and datasets. We accentuate the effectiveness of our shrunk spaces when used in one-shot search by achieving the best Top-1 accuracy in two different search spaces. Our method achieves a SOTA Top-1 accuracy of 77.6\% in ImageNet under mobile constraints, best-in-class Kendal-Tau, architectural diversity, and search space size. | ['Yiran Chen', 'Tunhou Zhang', 'Arjun Sridhar', 'Bhavna Gopal'] | 2023-07-06 | null | null | null | null | ['efficient-exploration', 'architecture-search'] | ['methodology', 'methodology'] | [ 1.72284693e-02 -7.89479762e-02 -5.42606711e-01 -3.38191211e-01
-8.50115359e-01 -5.94088614e-01 3.22447181e-01 -3.32588553e-01
-6.77236259e-01 6.78171217e-01 4.39759083e-02 -4.75239456e-01
-8.31568897e-01 -3.77859354e-01 -5.01219928e-01 -7.16899514e-01
-2.54563570e-01 5.56718588e-01 3.48699689e-01 -2.02319205e-01
4.07064974e-01 6.21894002e-01 -1.38990557e+00 3.51052098e-02
7.53224909e-01 1.35079789e+00 2.35330299e-01 3.58895212e-01
9.15480480e-02 3.42696607e-01 -5.44755101e-01 -2.30710268e-01
6.39478743e-01 -2.12922245e-01 -6.89999938e-01 -5.56364059e-01
8.41828525e-01 -3.20973285e-02 -3.42912972e-01 9.21985567e-01
5.88867128e-01 1.18585071e-02 4.02516454e-01 -1.20137870e+00
-3.36016983e-01 9.45145369e-01 -5.74799299e-01 8.88332844e-01
-4.76079792e-01 1.65798366e-01 1.56841648e+00 -9.85531330e-01
6.94595635e-01 9.28017855e-01 7.64501989e-01 4.62423027e-01
-1.56160772e+00 -1.08492386e+00 3.25633168e-01 1.10307030e-01
-1.57693410e+00 -7.76370645e-01 6.92876875e-01 -1.20461993e-01
1.36927176e+00 5.81927419e-01 9.01446402e-01 9.53649223e-01
1.95819870e-01 7.98258543e-01 5.80901563e-01 -2.87091106e-01
6.16419673e-01 -5.08532263e-02 1.36553973e-01 7.07601786e-01
3.48421782e-01 1.02376945e-01 -9.53606129e-01 -3.94984871e-01
6.50438249e-01 2.41840351e-02 -9.34154168e-02 -6.53235912e-01
-9.58520472e-01 9.18853223e-01 7.40861475e-01 4.19151753e-01
-2.63099670e-01 3.04760039e-01 1.81529507e-01 4.30186898e-01
3.19033712e-01 1.16984940e+00 -6.49856925e-01 -1.41054273e-01
-1.35486889e+00 4.07852441e-01 6.77963018e-01 8.14856112e-01
4.62968767e-01 2.99585521e-01 -1.23634376e-01 9.94958997e-01
1.95882544e-01 1.85357586e-01 5.86588860e-01 -1.03423643e+00
5.63368499e-01 6.61919653e-01 -3.88385117e-01 -1.05673981e+00
-6.31170928e-01 -1.29321098e+00 -6.17104888e-01 1.57498583e-01
1.37594774e-01 -5.48417941e-02 -9.82317150e-01 1.97302580e+00
7.78202433e-03 -3.32967579e-01 -3.46558690e-01 6.94647014e-01
4.67431605e-01 1.82150736e-01 -5.50114848e-02 -1.93479061e-01
1.26570618e+00 -1.14963770e+00 -9.07153860e-02 -7.22291172e-01
8.17441702e-01 -1.76041707e-01 1.11380911e+00 2.97969341e-01
-1.38478279e+00 -3.81024964e-02 -1.23449779e+00 1.96693704e-01
-1.96247593e-01 -8.33419058e-03 6.89420223e-01 7.63069570e-01
-1.47947598e+00 6.12939537e-01 -7.35420108e-01 -9.84108075e-02
1.05775058e+00 7.52184033e-01 -1.14235997e-01 2.30494097e-01
-8.06558728e-01 7.36659586e-01 4.31498766e-01 -1.21852010e-01
-8.57447267e-01 -8.60293388e-01 -2.69506425e-01 4.14976358e-01
6.17487609e-01 -6.96414769e-01 9.11132872e-01 -6.83378339e-01
-1.07033587e+00 6.32268429e-01 3.96641046e-02 -9.20559108e-01
2.03362375e-01 -2.80548204e-02 -2.66258001e-01 -2.17026904e-01
-5.15387356e-02 8.88820648e-01 7.49029696e-01 -7.76069701e-01
-5.99549651e-01 -6.66626692e-01 -3.12567383e-01 3.20329845e-01
-9.92211223e-01 5.38347028e-02 -4.93883044e-01 -7.59987533e-01
5.58303118e-01 -1.11812794e+00 -4.51433510e-01 -1.09534770e-01
-4.55354393e-01 -1.01881862e-01 3.62800479e-01 -2.38610238e-01
1.91574645e+00 -1.92738092e+00 1.81513414e-01 7.38209784e-01
6.52682841e-01 1.23565309e-01 -3.08428466e-01 -2.87955880e-01
-6.61512390e-02 4.82447088e-01 2.34520864e-02 -1.20044775e-01
3.48883905e-02 -1.83480028e-02 -1.39081225e-01 2.75929093e-01
-8.84210616e-02 1.07851839e+00 -4.09117728e-01 -3.95322144e-01
-5.02249837e-01 1.24538727e-01 -8.94488394e-01 -4.10372943e-01
1.65019035e-02 -1.62637845e-01 -6.53238714e-01 1.07649016e+00
2.20261365e-01 -6.01865590e-01 1.50143415e-01 6.92079356e-03
-1.03156291e-01 2.55954027e-01 -9.50263262e-01 1.64126730e+00
-9.06994864e-02 6.73558891e-01 2.64638811e-01 -7.86399543e-01
1.05613446e+00 -3.77015442e-01 4.42438334e-01 -9.87006664e-01
1.43590793e-01 6.56428695e-01 3.88552368e-01 7.05312714e-02
3.44174534e-01 3.19159806e-01 2.08541110e-01 4.28628832e-01
9.61168762e-03 2.14287445e-01 2.00998515e-01 -5.94224147e-02
1.51501763e+00 -5.54908335e-01 1.15674280e-01 -7.69153297e-01
-1.23624876e-01 4.46898900e-02 6.27184391e-01 9.47715700e-01
-3.81753564e-01 2.87420452e-01 3.38316500e-01 -8.48369241e-01
-1.23554361e+00 -9.46396708e-01 -1.70541137e-01 1.65186155e+00
-2.69217521e-01 -4.22768265e-01 -6.17721319e-01 -6.32196367e-01
1.16083364e-03 4.64004099e-01 -6.27135217e-01 -2.83857018e-01
-8.53671849e-01 -9.22080576e-01 8.19991887e-01 5.90531588e-01
3.64306331e-01 -9.35619652e-01 -1.09092522e+00 -5.68537898e-02
3.47084701e-01 -4.83593166e-01 -5.41784525e-01 7.57817268e-01
-1.32802022e+00 -6.55623496e-01 -7.62789667e-01 -9.18408453e-01
4.72459048e-01 2.94471756e-02 1.36132419e+00 1.98236957e-01
-4.85123217e-01 -2.91429847e-01 1.45191595e-01 -3.81974220e-01
1.32335275e-01 8.99708271e-01 2.42193535e-01 -5.38500309e-01
1.86177209e-01 -6.64775789e-01 -8.26158047e-01 5.19854069e-01
-3.44167769e-01 -1.50426745e-01 8.19844246e-01 1.11090863e+00
7.15518594e-01 1.23188198e-01 6.06117010e-01 -4.09227967e-01
7.95356989e-01 -5.21671116e-01 -6.04243875e-01 4.08903778e-01
-1.51636279e+00 4.31028187e-01 2.91532248e-01 -4.84836906e-01
-5.67647517e-01 -4.08739559e-02 2.82718837e-01 -6.49694324e-01
3.30405593e-01 3.37407976e-01 2.15869278e-01 -3.43217552e-01
1.11914015e+00 1.86160207e-01 8.36210921e-02 -5.92956185e-01
1.15044244e-01 1.73164159e-01 2.70579606e-01 -3.93859237e-01
5.79053581e-01 2.39401117e-01 -3.53577100e-02 -3.81480634e-01
-5.76259613e-01 -3.04597110e-01 -3.35951269e-01 3.14049870e-01
3.08802783e-01 -7.23080695e-01 -4.37309116e-01 -1.28580984e-02
-5.02158761e-01 -2.45754555e-01 -4.79448646e-01 1.63222522e-01
-3.27660024e-01 -2.95519501e-01 -2.54723221e-01 -7.37939835e-01
-8.59681249e-01 -1.41204584e+00 7.33537436e-01 3.35855216e-01
-5.23779273e-01 -7.40352213e-01 -1.46855228e-02 1.52670890e-01
7.55551755e-01 -2.99579471e-01 1.10178363e+00 -1.18795514e+00
-7.12571681e-01 6.98155388e-02 -2.15837941e-01 -9.28707048e-02
-4.01924610e-01 -3.12197059e-01 -6.77528918e-01 -5.45821011e-01
-2.59494871e-01 -3.10594797e-01 1.08294725e+00 7.18010008e-01
1.39657760e+00 -4.71679866e-01 -7.12992787e-01 1.11668789e+00
1.11559033e+00 6.32902384e-01 3.02419960e-01 8.67288589e-01
3.76471102e-01 3.26135606e-01 1.96491301e-01 4.29229647e-01
-1.79415107e-01 7.35269845e-01 5.36965549e-01 1.48034915e-02
1.55761912e-01 1.96081232e-02 -1.76868230e-01 4.66406584e-01
1.71289127e-02 -1.72978431e-01 -1.26650977e+00 7.10317194e-01
-1.72373438e+00 -8.10206175e-01 6.53874815e-01 1.88206816e+00
9.80896413e-01 5.87403476e-01 2.71343470e-01 -1.55654967e-01
3.31219077e-01 2.80970454e-01 -1.15623987e+00 -2.82421857e-01
-3.14962119e-01 2.00761512e-01 8.39141786e-01 4.89320830e-02
-9.29053247e-01 7.37648666e-01 7.53987694e+00 1.31992733e+00
-8.73749435e-01 -6.09367490e-02 1.15545440e+00 -8.99438560e-01
-4.53810364e-01 -1.49017334e-01 -1.17737877e+00 2.39029884e-01
1.11410356e+00 -3.41269448e-02 6.45823240e-01 1.24955690e+00
-2.62562960e-01 2.10530758e-01 -1.00888908e+00 1.32812166e+00
1.41022587e-02 -1.91952300e+00 -1.96595669e-01 3.08298469e-01
8.46735597e-01 4.82631654e-01 5.44262648e-01 2.51095206e-01
1.77390829e-01 -1.41185474e+00 8.23014796e-01 6.61969483e-02
8.40009987e-01 -6.57638371e-01 3.46367836e-01 1.37818366e-01
-1.11010921e+00 -7.92342544e-01 -1.89831659e-01 3.47241670e-01
-2.47584134e-01 2.35066637e-01 -9.44093466e-01 -3.26841712e-01
1.10102820e+00 2.90221900e-01 -7.95334280e-01 1.28222013e+00
4.72327203e-01 4.42964464e-01 -6.57697916e-01 -4.86284465e-01
4.25954431e-01 4.90332395e-02 7.32560277e-01 9.39355373e-01
4.36547130e-01 -2.70690560e-01 -1.39038309e-01 1.04485476e+00
-2.79640615e-01 7.32707903e-02 -5.08624017e-01 -1.80515662e-01
9.89883184e-01 9.53594685e-01 -8.40495646e-01 -5.55464923e-02
1.70318201e-01 3.14760149e-01 3.95922363e-01 2.57548273e-01
-6.36564910e-01 -3.38504493e-01 6.98213696e-01 2.01407284e-01
2.57001787e-01 -4.00103964e-02 -9.69650209e-01 -6.34967089e-01
-2.25389730e-02 -1.11240637e+00 5.66679001e-01 -2.61533350e-01
-9.04467404e-01 9.81357098e-01 -5.88998124e-02 -8.38504434e-01
-2.07779005e-01 -4.85751748e-01 -3.61080647e-01 6.96804225e-01
-1.00488508e+00 -8.20196867e-01 -1.06049709e-01 2.46624187e-01
9.13218737e-01 -1.05925262e+00 5.48457086e-01 2.29382083e-01
-7.67443478e-01 1.20994091e+00 3.83016050e-01 -3.20773989e-01
3.64136636e-01 -8.12829971e-01 5.76116860e-01 4.78309423e-01
5.21160185e-01 9.41146672e-01 5.35693765e-01 -5.75635374e-01
-1.16430199e+00 -4.54435199e-01 5.17165959e-01 -4.44203615e-01
7.85711288e-01 -2.45807931e-01 -6.41198635e-01 4.83653694e-01
-1.74419641e-01 -1.38189301e-01 4.78415191e-01 6.82191849e-01
-5.96480906e-01 -3.00704569e-01 -9.55908775e-01 8.55709851e-01
1.49724936e+00 -2.85568863e-01 -1.11089669e-01 5.21022677e-02
7.56787360e-01 -3.26855361e-01 -6.70420647e-01 4.91577595e-01
8.50818932e-01 -9.26398218e-01 1.27967942e+00 -7.47398555e-01
3.30077447e-02 3.13660383e-01 -3.08359057e-01 -1.24154460e+00
-5.21506071e-01 -8.09525728e-01 -2.92710960e-01 3.90873343e-01
1.08263040e+00 -6.53607190e-01 1.39096546e+00 8.71687770e-01
-2.12439239e-01 -1.62550390e+00 -1.19794595e+00 -9.06039834e-01
1.03089355e-01 -2.10189655e-01 8.02766204e-01 7.05439031e-01
-1.13721222e-01 3.55964094e-01 -3.09710968e-02 -5.09242177e-01
5.12056351e-01 -9.21890605e-03 2.44129106e-01 -1.44854188e+00
-3.27864170e-01 -1.28512383e+00 -1.82955146e-01 -9.23722148e-01
-1.51355341e-01 -7.86393046e-01 -1.44794464e-01 -8.99139106e-01
4.07602757e-01 -9.89362538e-01 -5.05470574e-01 7.52414405e-01
3.57478112e-01 1.55214638e-01 1.15184180e-01 6.99833453e-01
-7.91942418e-01 4.18727219e-01 8.76147389e-01 -1.55279055e-01
-4.45116401e-01 -1.32130265e-01 -1.00038052e+00 7.04736292e-01
8.69822979e-01 -3.35952699e-01 -7.98742294e-01 -5.94341159e-01
5.40312052e-01 -4.11901921e-01 1.14892982e-01 -1.12377739e+00
4.09618497e-01 -1.90652832e-01 3.87220979e-01 -5.55177212e-01
4.02363420e-01 -5.67794979e-01 1.31243125e-01 5.37705481e-01
-7.45283306e-01 5.44834554e-01 2.83887863e-01 4.84486043e-01
1.67261392e-01 -1.72481105e-01 8.14134002e-01 -3.94555442e-02
-7.86663830e-01 3.40393573e-01 3.25132072e-01 3.11028153e-01
8.68909538e-01 -6.28833175e-01 -4.67820942e-01 1.48338631e-01
-7.68025041e-01 2.99078465e-01 2.37494946e-01 4.30661261e-01
6.41715705e-01 -1.23815382e+00 -3.53892803e-01 4.65820760e-01
3.98768112e-02 -8.73794109e-02 5.70748225e-02 8.31774890e-01
-4.34247911e-01 6.52731299e-01 -2.80847192e-01 -6.21885955e-01
-1.45127892e+00 1.28159374e-01 4.81574267e-01 -3.89748245e-01
-5.32884836e-01 1.62820601e+00 2.02742383e-01 -1.36142075e-01
5.75635314e-01 -4.82119881e-02 -1.98750272e-02 1.88544914e-01
5.21029592e-01 5.90838373e-01 1.93033203e-01 -1.94556773e-01
-5.74819624e-01 2.91628331e-01 -4.24637347e-01 -1.32618025e-01
1.54656649e+00 -2.16628704e-02 8.31604972e-02 1.22935914e-01
1.23985052e+00 -3.73331755e-01 -1.29848385e+00 -2.15837285e-01
2.87933648e-01 -4.34608072e-01 4.84402627e-01 -9.42867160e-01
-1.30636156e+00 4.14245039e-01 8.88148487e-01 2.36061603e-01
1.14943731e+00 2.08717257e-01 6.60683632e-01 6.79168344e-01
4.62089717e-01 -1.43246210e+00 3.00685763e-01 6.00811899e-01
8.43854308e-01 -1.07670474e+00 9.02622566e-02 1.86617330e-01
-4.23657894e-01 8.40324342e-01 9.74528909e-01 1.43223584e-01
6.82721972e-01 3.92206281e-01 -3.95492941e-01 -6.16552770e-01
-1.13682008e+00 2.06445485e-01 3.88652474e-01 2.46504188e-01
8.86496529e-02 -1.41472677e-02 3.64958234e-02 5.70024192e-01
-4.98455614e-01 -4.64757293e-01 -2.41591319e-01 8.61059487e-01
-7.99046218e-01 -6.34346187e-01 -1.60098225e-01 8.92956614e-01
-3.32327545e-01 -4.50896949e-01 -3.71247947e-01 6.89164639e-01
-2.07950085e-01 4.01920140e-01 2.52938569e-01 -6.87789679e-01
1.14231080e-01 2.23539859e-01 1.93674967e-01 -2.43064120e-01
-6.16514862e-01 1.52524590e-01 2.31046870e-01 -8.41511488e-01
3.45095456e-01 -5.44746578e-01 -1.04187322e+00 -2.90130705e-01
-5.79039931e-01 -7.62076862e-03 7.08674490e-01 6.87096477e-01
7.26178050e-01 5.10173023e-01 3.52155089e-01 -4.89532948e-01
-9.97902572e-01 -6.32213771e-01 -4.69768763e-01 -1.46609783e-01
2.68205911e-01 -7.71682382e-01 -2.98657477e-01 -6.91509724e-01] | [8.603718757629395, 3.2441904544830322] |
59a759df-0845-4167-9606-1cda36efe08d | meta-learning-a-real-time-tabular-automl | 2207.01848 | null | https://arxiv.org/abs/2207.01848v5 | https://arxiv.org/pdf/2207.01848v5.pdf | TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second | We present TabPFN, a trained Transformer that can do supervised classification for small tabular datasets in less than a second, needs no hyperparameter tuning and is competitive with state-of-the-art classification methods. TabPFN is fully entailed in the weights of our network, which accepts training and test samples as a set-valued input and yields predictions for the entire test set in a single forward pass. TabPFN is a Prior-Data Fitted Network (PFN) and is trained offline once, to approximate Bayesian inference on synthetic datasets drawn from our prior. This prior incorporates ideas from causal reasoning: It entails a large space of structural causal models with a preference for simple structures. On the 18 datasets in the OpenML-CC18 suite that contain up to 1 000 training data points, up to 100 purely numerical features without missing values, and up to 10 classes, we show that our method clearly outperforms boosted trees and performs on par with complex state-of-the-art AutoML systems with up to 70$\times$ speedup. This increases to a 3200$\times$ speedup when a GPU is available. We also validate these results on an additional 67 small numerical datasets from OpenML. We provide all our code, the trained TabPFN, an interactive browser demo and a Colab notebook at https://github.com/automl/TabPFN. | ['Frank Hutter', 'Katharina Eggensperger', 'Samuel Müller', 'Noah Hollmann'] | 2022-07-05 | null | null | null | null | ['classification'] | ['methodology'] | [ 1.10182464e-01 6.46446705e-01 -5.44403493e-01 -7.47459888e-01
-9.93990123e-01 -4.11031246e-01 8.08482826e-01 6.94233179e-02
5.11054769e-02 1.33691895e+00 1.28233120e-01 -8.15103233e-01
-3.46465737e-01 -9.30943549e-01 -1.15856183e+00 -5.33232570e-01
-3.01869363e-01 1.18508363e+00 1.67517066e-01 2.36775890e-01
-3.18327770e-02 1.53560743e-01 -1.52704597e+00 6.53800786e-01
6.43914580e-01 1.15458107e+00 -3.32157582e-01 6.67603195e-01
8.91425461e-02 1.01466596e+00 -4.73811805e-01 -6.15944922e-01
-9.66711789e-02 -1.48699418e-01 -7.29069650e-01 -3.92114967e-01
7.25431442e-01 -2.78791904e-01 -2.05604658e-01 6.70554638e-01
2.44910091e-01 -1.95434436e-01 7.59326339e-01 -1.60497081e+00
-2.48626187e-01 1.20232570e+00 -3.63861710e-01 -1.55520558e-01
3.18008244e-01 2.88680524e-01 1.16699910e+00 -8.35999072e-01
6.72640622e-01 1.70325530e+00 1.02705777e+00 2.33107284e-01
-1.79455900e+00 -7.74990499e-01 -7.95488209e-02 2.08660573e-01
-1.14842618e+00 -2.50427872e-01 2.39557996e-01 -4.97679830e-01
1.16329551e+00 3.75426739e-01 7.21821010e-01 1.47814608e+00
4.07262862e-01 6.83249891e-01 9.55893695e-01 -2.50651330e-01
4.63318288e-01 -1.37760192e-01 2.89228380e-01 8.90813589e-01
2.97319919e-01 3.43251199e-01 -6.78556144e-01 -6.77564263e-01
5.69715202e-01 -4.08472747e-01 -1.87154245e-02 -3.41218710e-01
-1.05043876e+00 9.02913570e-01 2.93083966e-01 -5.58105052e-01
-1.26847147e-03 7.70441651e-01 2.01608658e-01 -4.92147505e-02
4.22441691e-01 8.69703740e-02 -1.00542319e+00 -2.52840042e-01
-8.07031274e-01 5.63809931e-01 1.19727612e+00 1.04685140e+00
6.64133966e-01 -8.56507495e-02 5.62049001e-02 8.96255255e-01
3.63643885e-01 5.51845551e-01 1.65913603e-03 -1.40371346e+00
3.03095549e-01 5.25458694e-01 -2.72071306e-02 -5.08641422e-01
-4.83547330e-01 -4.33940768e-01 -8.32569122e-01 1.12473458e-01
7.75444508e-01 -2.39122823e-01 -9.16635633e-01 1.77394331e+00
4.84909266e-01 -8.30330849e-02 -2.13099539e-01 3.69618446e-01
6.30261362e-01 6.55931413e-01 9.64018106e-02 -9.16723162e-02
1.27442610e+00 -6.46573842e-01 -4.05730218e-01 -3.88014823e-01
7.27643430e-01 -5.83805561e-01 1.05345309e+00 8.72574210e-01
-8.98444772e-01 -1.23511925e-01 -1.08992743e+00 -9.65871587e-02
-4.59592938e-01 -1.25223082e-02 1.19806266e+00 6.07066810e-01
-7.38333344e-01 1.03404224e+00 -9.98464644e-01 -3.07009220e-01
6.08305395e-01 2.34229758e-01 -1.29302293e-01 -2.62741059e-01
-1.18090022e+00 6.68426633e-01 6.46688402e-01 -4.50014211e-02
-1.16641009e+00 -1.23797059e+00 -7.78550982e-01 9.33625549e-02
6.21838570e-01 -7.58450389e-01 1.53421509e+00 -2.79137850e-01
-1.22908497e+00 3.22916269e-01 5.79452068e-02 -5.98160684e-01
6.69279873e-01 -1.57117531e-01 -2.13768542e-01 -3.45343292e-01
8.88428316e-02 8.96854818e-01 3.32192421e-01 -1.06453645e+00
-6.40057027e-01 -1.14814326e-01 -7.97164142e-02 -3.34194660e-01
1.69450641e-01 -2.55614579e-01 -2.88500428e-01 -5.54460466e-01
-6.98413327e-02 -7.28227615e-01 -1.59905583e-01 5.91417886e-02
-7.17187405e-01 -3.62883747e-01 5.22423744e-01 -4.30593669e-01
1.11466765e+00 -1.57685113e+00 -6.61059543e-02 1.69060320e-01
5.60807213e-02 -3.60405862e-01 1.25009596e-01 3.70104283e-01
-5.00043750e-01 3.37553680e-01 -7.37700999e-01 -1.37228966e-01
3.65719616e-01 5.35454273e-01 -4.12599802e-01 3.48105431e-01
1.99955940e-01 6.84701860e-01 -8.53641748e-01 -6.45093739e-01
1.80632159e-01 1.40522450e-01 -7.95569718e-01 8.47100466e-02
-9.85971570e-01 -1.90334871e-01 -9.80632529e-02 7.09494650e-01
5.67928731e-01 -5.07003129e-01 5.51292121e-01 -3.38281915e-02
8.12013969e-02 6.83138907e-01 -1.23370767e+00 1.66839683e+00
-3.26226056e-01 4.80704039e-01 -8.99886340e-02 -7.68183231e-01
6.58123493e-01 2.23963723e-01 2.63226628e-01 -1.51728794e-01
-2.46256590e-02 2.43476257e-01 -4.45283987e-02 -4.45673913e-02
-6.42052516e-02 1.13534361e-01 -1.38030767e-01 6.68290734e-01
3.96442115e-01 -3.83288562e-01 6.39907300e-01 3.36117893e-01
1.54952824e+00 6.80373490e-01 3.23690712e-01 -4.76772010e-01
1.54934637e-02 3.58740509e-01 7.47648239e-01 1.00186646e+00
5.43515563e-01 2.45778948e-01 1.09759295e+00 -8.30702960e-01
-1.08735359e+00 -1.11325407e+00 -6.50223076e-01 1.00064719e+00
-6.28507853e-01 -8.15680981e-01 -6.73995554e-01 -7.14985669e-01
3.31332177e-01 1.43838167e+00 -7.63395548e-01 2.09352329e-01
-3.47926319e-01 -1.27588582e+00 7.46561050e-01 5.33596218e-01
3.69845361e-01 -9.67184544e-01 -3.28177422e-01 2.84931749e-01
-3.03181093e-02 -8.60063970e-01 3.36881071e-01 7.53665030e-01
-9.99290287e-01 -1.29578495e+00 1.57557681e-01 3.14705037e-02
3.81265879e-01 -7.79077828e-01 1.58064675e+00 -1.54195040e-01
-3.74760121e-01 -2.29489431e-01 1.50998637e-01 -4.36186254e-01
-7.09570110e-01 2.17977986e-01 -1.88928932e-01 -7.58433819e-01
2.28097320e-01 -9.63757336e-01 -1.52397856e-01 4.05763566e-01
-4.99192327e-01 6.16143584e-01 3.28182995e-01 1.14873028e+00
4.57563400e-01 9.43899155e-02 1.48112938e-01 -1.13130677e+00
2.52517283e-01 -7.41175234e-01 -1.04259479e+00 2.22865403e-01
-8.74487221e-01 3.26503664e-01 7.09991336e-01 -5.73456138e-02
-1.11800647e+00 9.67699811e-02 -2.33606741e-01 -1.50527209e-01
-2.86435246e-01 8.26752663e-01 -1.13851979e-01 6.38570666e-01
9.79963124e-01 -4.28952456e-01 -4.99203987e-02 -6.23575270e-01
5.22706628e-01 4.11475897e-01 6.06302619e-01 -1.07957220e+00
5.04953504e-01 1.82712972e-01 3.27619761e-01 -2.50734776e-01
-1.14656270e+00 3.50439191e-01 -5.89729846e-01 -9.01984870e-02
2.96568871e-01 -6.39597297e-01 -1.13543618e+00 2.05889046e-01
-1.09082973e+00 -1.04254329e+00 -2.20289215e-01 2.10439146e-01
-7.87497997e-01 -2.93817252e-01 -6.65155232e-01 -6.55983210e-01
-8.23474005e-02 -7.58788347e-01 8.45919073e-01 -3.45407307e-01
-3.39540720e-01 -9.16058064e-01 -1.00532718e-01 2.11395994e-01
1.24547805e-03 3.33740771e-01 1.41185832e+00 -6.17464602e-01
-6.51265681e-01 -2.05792293e-01 -2.94875205e-01 -7.87133053e-02
-2.12671325e-01 4.84816343e-01 -1.02037144e+00 1.12421103e-01
-3.68650526e-01 -7.85686493e-01 1.04755652e+00 3.88535887e-01
1.62753975e+00 -5.07531345e-01 -6.78346217e-01 6.36628270e-01
1.16622508e+00 2.56940685e-02 5.32464206e-01 1.63007602e-01
5.86598635e-01 5.44540942e-01 6.59651816e-01 5.12320518e-01
2.77386516e-01 3.96099180e-01 7.13449597e-01 7.62055963e-02
-2.60554180e-02 -4.95977581e-01 1.36850074e-01 2.64007002e-01
1.67694643e-01 -6.16062880e-01 -1.44311726e+00 2.17388660e-01
-2.18649220e+00 -7.38072217e-01 -4.49196815e-01 2.09267139e+00
1.28922260e+00 4.78242278e-01 -2.48849437e-01 2.24913597e-01
3.50900412e-01 -1.52169347e-01 -6.59907699e-01 -2.50887513e-01
9.34969261e-02 4.06379402e-01 6.53624117e-01 7.85358906e-01
-9.70863938e-01 8.59122515e-01 6.77287006e+00 1.30738091e+00
-5.58806658e-01 9.21716318e-02 1.09383750e+00 -2.48263389e-01
-3.33086252e-01 2.62353897e-01 -8.67535114e-01 2.89745569e-01
1.57420015e+00 -7.71862566e-02 5.99112630e-01 9.97065902e-01
3.96688990e-02 -4.28827196e-01 -1.59099483e+00 4.64681953e-01
-4.36918497e-01 -1.69286335e+00 -1.38992101e-01 7.78075010e-02
4.13742989e-01 2.51624703e-01 -3.02303046e-01 3.60186726e-01
1.45674300e+00 -1.46857464e+00 8.30470145e-01 3.20408881e-01
8.69893372e-01 -6.61612451e-01 6.25437796e-01 3.63499343e-01
-7.25992382e-01 -3.50756496e-02 -3.60221118e-01 -2.88058460e-01
2.36967057e-02 9.51888919e-01 -1.09263325e+00 5.88899970e-01
9.42165494e-01 8.86112094e-01 -5.85868180e-01 6.25148594e-01
-7.95330405e-01 1.18094671e+00 -8.14059079e-01 -7.47126192e-02
-7.80573189e-02 2.50329971e-01 1.75878271e-01 1.14745235e+00
2.02787176e-01 5.35030812e-02 -1.42625824e-01 1.25123847e+00
5.99974282e-02 -3.69684488e-01 -2.94219673e-01 1.68111563e-01
6.08816624e-01 1.14089835e+00 -5.17183602e-01 -5.68300366e-01
8.73957872e-02 1.58872917e-01 4.32378471e-01 1.34175956e-01
-1.07949471e+00 -2.32350454e-02 3.94849807e-01 4.22382876e-02
1.33109868e-01 5.07269613e-02 -4.65773374e-01 -8.56205702e-01
-4.39939499e-01 -9.75999475e-01 7.44028568e-01 -1.17784595e+00
-1.43612826e+00 1.58329844e-01 5.07035911e-01 -6.18685305e-01
-6.79859459e-01 -1.04881847e+00 -1.50150806e-01 8.20581496e-01
-8.35632205e-01 -9.26661432e-01 -3.24038535e-01 2.54792273e-01
2.31104732e-01 1.52154326e-01 1.11831880e+00 1.20630696e-01
-8.12243938e-01 4.13540035e-01 3.52158360e-02 -3.13804522e-02
7.68676877e-01 -1.44876492e+00 6.72380447e-01 4.72114831e-01
-1.31496713e-01 5.80135345e-01 8.79882038e-01 -9.10822988e-01
-1.26868963e+00 -1.14726388e+00 6.61839604e-01 -7.01692045e-01
1.04605961e+00 -7.09128439e-01 -7.01374948e-01 1.12489331e+00
1.93130195e-01 6.19425662e-02 5.55232763e-01 6.00066602e-01
-4.34034407e-01 -1.86224982e-01 -1.04576373e+00 4.79055375e-01
1.20003355e+00 1.26369283e-01 -4.03563410e-01 5.32571614e-01
6.48142874e-01 -7.00890243e-01 -1.10030687e+00 5.57519794e-01
8.52507591e-01 -1.07522535e+00 7.34065950e-01 -5.49117386e-01
1.00406015e+00 -1.35680586e-01 -3.85362506e-01 -1.21882677e+00
-8.89675841e-02 -5.44624925e-01 -1.31272569e-01 1.22374415e+00
8.72444928e-01 -7.40290225e-01 8.65083516e-01 6.85235083e-01
-1.58180311e-01 -7.97659516e-01 -9.28063869e-01 -5.29557943e-01
2.46672511e-01 -1.18326211e+00 7.01181948e-01 8.37174952e-01
-8.99903551e-02 3.47300380e-01 -1.50638342e-01 2.48196200e-02
1.11756063e+00 9.46768373e-03 8.37717116e-01 -1.50146782e+00
-4.89047140e-01 -2.99072444e-01 1.23518027e-01 -7.76069820e-01
2.40473941e-01 -1.00150907e+00 -3.10419537e-02 -1.34520555e+00
3.10169667e-01 -7.65229106e-01 1.35303363e-01 1.25273335e+00
2.36839801e-01 1.08653128e-01 -2.75589108e-01 -4.30733599e-02
-1.36063576e-01 4.45485085e-01 6.98809206e-01 -1.63386121e-01
3.32973778e-01 -2.50627667e-01 -3.56062323e-01 1.02980578e+00
6.68282032e-01 -8.72758746e-01 -2.53753781e-01 -1.49597079e-01
5.76146841e-01 4.97789383e-01 7.13982642e-01 -8.72736335e-01
8.59616771e-02 -3.85661155e-01 7.10860372e-01 -1.01837349e+00
3.72460455e-01 -5.80184221e-01 6.32096171e-01 3.92857373e-01
-4.32980239e-01 -2.17386797e-01 5.87768972e-01 3.44219923e-01
1.92516297e-01 -2.35675737e-01 4.64610040e-01 -1.78220287e-01
-6.26368523e-02 -5.51907569e-02 -3.67256731e-01 -3.76269855e-02
5.42694032e-01 2.70389438e-01 -8.12407792e-01 -1.90485954e-01
-6.94488287e-01 4.33871865e-01 2.49370098e-01 7.22074732e-02
9.55724344e-02 -1.17092633e+00 -6.80308104e-01 7.22444355e-02
-1.57955006e-01 2.60118604e-01 -1.35583267e-01 5.55431306e-01
-5.06040037e-01 6.61337137e-01 2.09478438e-02 -7.96005011e-01
-8.81045043e-01 3.00955415e-01 4.43867654e-01 -3.42334837e-01
-2.44065747e-01 8.07360113e-01 -5.37846498e-02 -8.77227366e-01
2.04954132e-01 -4.80101079e-01 4.53210860e-01 -5.53083345e-02
1.23945370e-01 4.92067099e-01 1.07131541e-01 4.06715512e-01
-2.44147465e-01 -3.37367281e-02 1.38112143e-01 -3.40387702e-01
1.53655505e+00 4.63829190e-01 -4.78436857e-01 7.58040786e-01
5.80250800e-01 -2.50922799e-01 -1.42243552e+00 1.38356835e-01
-3.19420956e-02 -2.27247804e-01 -3.42619829e-02 -1.53320813e+00
-7.45088041e-01 7.99377203e-01 1.59481660e-01 1.72131076e-01
6.84898973e-01 4.40084599e-02 -8.09798986e-02 5.29384136e-01
5.17932773e-01 -6.03202224e-01 -2.65797108e-01 5.19059777e-01
1.07561934e+00 -9.84464467e-01 4.00685132e-01 -5.97240686e-01
2.78773271e-02 1.18940723e+00 6.33783460e-01 1.29186148e-02
5.88647306e-01 8.41745615e-01 -3.46200079e-01 -2.32741043e-01
-1.73159420e+00 4.55186397e-01 -4.54539321e-02 3.86310071e-01
5.85842907e-01 1.46431565e-01 -1.20417960e-02 6.08910799e-01
-7.58666694e-01 2.08218917e-01 4.59711313e-01 6.64665997e-01
-1.58212855e-01 -1.21518707e+00 -4.55571443e-01 8.59529853e-01
-2.22913072e-01 -4.84525323e-01 -1.36415169e-01 1.16554928e+00
8.74620304e-02 7.29675293e-01 1.66695535e-01 -2.60025591e-01
-2.39076409e-02 3.81356448e-01 5.05780578e-01 -5.13442993e-01
-3.46132845e-01 -2.47528274e-02 6.94738150e-01 -9.17758405e-01
1.85024049e-02 -9.10745561e-01 -1.13742542e+00 -7.04735577e-01
4.31820974e-02 7.13950321e-02 6.50916159e-01 7.04677284e-01
1.23768948e-01 7.31023192e-01 -1.74580455e-01 -8.37989330e-01
-5.21951795e-01 -1.37175250e+00 -2.39620581e-01 -2.20683560e-01
-2.12568671e-01 -8.58719051e-01 -6.47567034e-01 5.72242588e-02] | [8.725162506103516, 6.400200366973877] |
2516e737-55fb-43f8-b50f-7a6f7fb94bcb | cap-vstnet-content-affinity-preserved | 2303.17867 | null | https://arxiv.org/abs/2303.17867v1 | https://arxiv.org/pdf/2303.17867v1.pdf | CAP-VSTNet: Content Affinity Preserved Versatile Style Transfer | Content affinity loss including feature and pixel affinity is a main problem which leads to artifacts in photorealistic and video style transfer. This paper proposes a new framework named CAP-VSTNet, which consists of a new reversible residual network and an unbiased linear transform module, for versatile style transfer. This reversible residual network can not only preserve content affinity but not introduce redundant information as traditional reversible networks, and hence facilitate better stylization. Empowered by Matting Laplacian training loss which can address the pixel affinity loss problem led by the linear transform, the proposed framework is applicable and effective on versatile style transfer. Extensive experiments show that CAP-VSTNet can produce better qualitative and quantitative results in comparison with the state-of-the-art methods. | ['Changqing Zou', 'Chengying Gao', 'Linfeng Wen'] | 2023-03-31 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wen_CAP-VSTNet_Content_Affinity_Preserved_Versatile_Style_Transfer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wen_CAP-VSTNet_Content_Affinity_Preserved_Versatile_Style_Transfer_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-style-transfer', 'image-matting'] | ['computer-vision', 'computer-vision'] | [ 3.47079664e-01 -6.87142164e-02 -2.05548942e-01 1.21339438e-02
-3.31823677e-01 -3.79288942e-01 3.45497787e-01 -8.81250739e-01
-1.80964857e-01 1.18102717e+00 2.90452510e-01 1.34169877e-01
2.46909663e-01 -6.95972860e-01 -7.78948188e-01 -7.81593502e-01
7.04504073e-01 8.23588148e-02 1.83827147e-01 -3.40654701e-01
2.31663615e-01 5.77101350e-01 -1.09372640e+00 2.99316376e-01
1.16913760e+00 9.73147452e-01 5.24876267e-02 1.42280420e-03
-4.12359685e-01 8.37614715e-01 -4.93064612e-01 -6.93236172e-01
3.49285692e-01 -8.00673246e-01 -6.57588720e-01 -8.63213763e-02
5.18355370e-01 -4.57397968e-01 -6.18250608e-01 1.14065635e+00
7.23564088e-01 1.70983613e-01 7.64136195e-01 -1.17546797e+00
-1.36790466e+00 3.93813342e-01 -7.72501945e-01 2.24762335e-02
1.02307580e-01 6.69182241e-02 7.70879507e-01 -9.81034875e-01
8.58428121e-01 1.62658238e+00 6.92933261e-01 8.54452848e-01
-1.34475684e+00 -8.23101223e-01 7.10239634e-02 1.46989360e-01
-1.33220172e+00 -2.68185019e-01 1.11036742e+00 -1.21395595e-01
1.60668597e-01 3.34379852e-01 6.87664986e-01 1.31715214e+00
4.07098889e-01 1.12148356e+00 1.18265986e+00 -1.01030782e-01
-1.73997104e-01 1.90987855e-01 -4.08569038e-01 6.44932985e-01
-3.26707661e-02 3.00524738e-02 -4.64686632e-01 4.16940272e-01
1.40154219e+00 2.04350874e-01 -5.20018399e-01 -5.15789211e-01
-1.03980672e+00 5.43758452e-01 8.19485903e-01 2.52804518e-01
-6.99649975e-02 2.22570702e-01 4.54103917e-01 4.51246083e-01
7.42898643e-01 2.81362027e-01 -3.03748064e-02 -4.74689156e-02
-8.81782353e-01 -6.38393406e-03 4.15520042e-01 1.17813432e+00
7.66977370e-01 4.53019917e-01 -7.12350309e-01 8.45139980e-01
1.39611870e-01 6.82434261e-01 4.08018887e-01 -9.24092412e-01
5.62173486e-01 7.03000009e-01 -5.81475459e-02 -1.05428684e+00
6.02806471e-02 -4.01751041e-01 -1.23857641e+00 2.40728423e-01
-2.01730840e-02 -1.01033337e-01 -8.97594392e-01 1.86847448e+00
1.22344628e-01 3.05979908e-01 -2.65474260e-01 7.65636206e-01
1.07577038e+00 7.23415196e-01 -4.55108508e-02 -7.47052357e-02
8.31416368e-01 -1.06664312e+00 -1.06671858e+00 7.64856264e-02
2.18918055e-01 -9.93432820e-01 1.41165340e+00 7.85150900e-02
-1.33289480e+00 -6.91843212e-01 -9.26392198e-01 -6.09279454e-01
-1.12928644e-01 2.10240737e-01 3.92736316e-01 5.06026924e-01
-1.04193497e+00 8.82239878e-01 -3.28835130e-01 -2.27148280e-01
7.39524782e-01 3.34046185e-01 -3.54223281e-01 7.68614113e-02
-1.19901252e+00 8.45487356e-01 1.36679605e-01 3.63229126e-01
-5.02575099e-01 -9.63433623e-01 -7.70302653e-01 8.91215354e-03
1.47987679e-01 -9.56367195e-01 6.47982657e-01 -1.22492874e+00
-2.07378626e+00 9.18367088e-01 -1.63069770e-01 8.80218372e-02
1.11569464e+00 -3.82771522e-01 -2.07048982e-01 -9.89418402e-02
-8.25284421e-03 9.55707192e-01 1.02923381e+00 -1.27795470e+00
-3.71938109e-01 -3.70959342e-02 -1.37515932e-01 3.70410740e-01
-4.89291877e-01 -1.39381215e-01 -6.76088274e-01 -1.13350070e+00
-8.34358782e-02 -9.23021138e-01 2.82023937e-01 4.74715471e-01
-4.31228638e-01 4.16146405e-02 1.20903277e+00 -8.24134529e-01
1.11376035e+00 -2.16089606e+00 6.69548035e-01 1.70492288e-02
8.58296677e-02 4.44695443e-01 -3.42211187e-01 8.71859938e-02
-1.14845581e-01 2.86248565e-01 -2.77624905e-01 -4.32660669e-01
-1.26995176e-01 2.25967765e-01 -4.92639571e-01 3.31611902e-01
2.61258632e-01 1.30882299e+00 -7.50674844e-01 -5.50753176e-01
3.76992941e-01 9.11128104e-01 -3.68933856e-01 1.45718277e-01
5.44626042e-02 7.32033730e-01 -4.21784163e-01 5.81119895e-01
1.12810826e+00 1.28750533e-01 -4.12134856e-01 -4.30942446e-01
-4.02567210e-03 -1.88285425e-01 -8.36743355e-01 1.89688563e+00
-4.36018080e-01 7.35772729e-01 -4.15196717e-02 -4.39117432e-01
1.17839289e+00 5.43607660e-02 2.13057771e-01 -7.96223104e-01
4.03404504e-01 2.46850178e-01 -5.62881827e-01 -2.51976103e-01
4.89524752e-01 -4.01052415e-01 3.12369585e-01 3.56086046e-02
-2.49124300e-02 -2.05211550e-01 -2.04486623e-01 1.27679065e-01
6.89327121e-01 7.24107623e-01 -3.96793038e-01 -4.54624742e-01
9.44406509e-01 -5.80476463e-01 9.18283463e-01 5.26938327e-02
-4.23778929e-02 9.94110107e-01 6.88361824e-01 -3.61640930e-01
-1.33789814e+00 -1.09551287e+00 3.67600061e-02 7.31733918e-01
6.40024126e-01 -6.93082670e-03 -7.84478784e-01 -7.88451910e-01
-7.98013136e-02 5.30739009e-01 -9.00133073e-01 -4.64227766e-01
-6.15065396e-01 -3.48983735e-01 3.75961363e-01 6.32464767e-01
1.11597717e+00 -1.21068680e+00 3.59729558e-01 -7.52494186e-02
-3.56172204e-01 -7.74810076e-01 -1.12366664e+00 -4.22075719e-01
-9.68385398e-01 -5.65206409e-01 -1.19564855e+00 -8.72501731e-01
8.85110140e-01 2.72618741e-01 1.00931561e+00 1.12719648e-01
3.92228067e-02 4.84711751e-02 -2.15014189e-01 -1.89004391e-01
-2.88012922e-01 3.37563694e-01 -2.57815540e-01 3.64858508e-01
-1.82883397e-01 -8.09523940e-01 -8.56506169e-01 5.84076822e-01
-1.22767961e+00 1.75011680e-01 7.46104121e-01 1.05718970e+00
6.81667268e-01 -3.96013618e-01 4.22482461e-01 -1.06123126e+00
5.91660738e-01 -5.77838533e-02 -3.61740887e-01 4.43101794e-01
-5.51583529e-01 1.94843143e-01 8.19730341e-01 -4.64054406e-01
-1.46994388e+00 -2.28624776e-01 -9.57026705e-02 -8.94788802e-01
4.79498297e-01 -1.51308522e-01 -5.82774758e-01 -6.23332441e-01
4.35776412e-01 2.61381239e-01 1.65091902e-01 -6.44692779e-01
3.45353782e-01 4.11846817e-01 5.31001270e-01 -5.17830193e-01
1.21402943e+00 5.58281064e-01 2.59305984e-01 -6.09192491e-01
-6.52135491e-01 1.53512387e-02 -8.02078068e-01 -1.46860346e-01
7.72113383e-01 -8.82060647e-01 -5.59153557e-01 6.96193218e-01
-1.20380723e+00 -3.00793350e-01 -5.98926902e-01 8.41784924e-02
-5.76415598e-01 4.70571280e-01 -8.56786072e-01 -4.02110636e-01
-5.93949914e-01 -9.21845675e-01 9.48984861e-01 4.60048407e-01
2.23488465e-01 -1.13101411e+00 -3.35590690e-02 2.91750729e-01
6.15075469e-01 3.88755083e-01 7.20486939e-01 1.56465113e-01
-6.91491604e-01 1.21092580e-01 -4.97420400e-01 6.96169317e-01
3.42293113e-01 4.61502224e-02 -8.32098067e-01 -4.61428821e-01
-1.46496296e-01 -1.44763127e-01 1.03815734e+00 7.88128823e-02
1.11004674e+00 -2.02046171e-01 -1.15866356e-01 1.14089501e+00
1.32312334e+00 2.62743514e-02 1.37383199e+00 3.40788007e-01
1.30396247e+00 3.78976911e-01 4.54989612e-01 1.38087541e-01
1.30417058e-02 5.85269213e-01 1.57288209e-01 -5.85350573e-01
-5.94618380e-01 -6.11962914e-01 4.16805565e-01 7.96110332e-01
-3.35188597e-01 -2.89194018e-01 -7.46729672e-02 2.99876004e-01
-1.87416828e+00 -8.88444304e-01 2.17606053e-02 1.93060327e+00
8.86675239e-01 -9.36471522e-02 -8.48243684e-02 -2.34832056e-02
1.04189360e+00 1.86237872e-01 -8.07433903e-01 -4.85190123e-01
-4.78407174e-01 2.26794258e-01 5.41542232e-01 2.53435642e-01
-8.24940503e-01 1.26312220e+00 6.32349062e+00 1.33523870e+00
-1.29026139e+00 3.96031700e-02 8.71275663e-01 8.60001072e-02
-8.10838938e-01 -1.57256678e-01 -4.93990064e-01 6.73949778e-01
1.70306191e-01 -1.83022171e-01 5.54286659e-01 5.99790215e-01
1.92358986e-01 4.28634942e-01 -9.29380894e-01 1.15412390e+00
1.89721137e-01 -1.24770689e+00 6.23774290e-01 -2.16330793e-02
9.10573602e-01 -6.12072468e-01 4.95283872e-01 1.49426505e-01
7.05732629e-02 -9.63991106e-01 7.00755537e-01 6.93161905e-01
1.24575889e+00 -9.19025183e-01 7.67484725e-01 -2.69318789e-01
-1.30098677e+00 1.39152139e-01 -6.25635266e-01 2.95568377e-01
1.62175447e-01 4.68122005e-01 1.06441155e-01 8.12232375e-01
6.00703299e-01 1.26576090e+00 -5.33441663e-01 9.78705168e-01
-5.24026990e-01 3.30311179e-01 5.03732376e-02 1.60603747e-01
1.45854652e-01 -7.09308445e-01 6.02929294e-01 7.87134588e-01
4.85543311e-01 -2.04408869e-01 -3.02436918e-01 1.13205791e+00
-5.68128169e-01 4.31181133e-01 -6.42066360e-01 2.97007650e-01
1.50359586e-01 1.31260920e+00 -5.84247291e-01 -2.64037430e-01
-2.44189098e-01 1.55014408e+00 2.97678530e-01 5.54920852e-01
-9.40245688e-01 -6.43251359e-01 4.49634850e-01 1.12713747e-01
3.42607975e-01 2.38786876e-01 -4.30281788e-01 -1.40716767e+00
7.96090662e-02 -6.60964012e-01 -3.29923965e-02 -9.95087683e-01
-1.39464808e+00 5.22358298e-01 -3.75086933e-01 -1.47893310e+00
4.06961888e-01 -5.54421067e-01 -8.94032001e-01 9.56645489e-01
-1.55827880e+00 -1.61730838e+00 -5.07136941e-01 7.20647335e-01
4.02181238e-01 -2.12975696e-01 2.72868544e-01 5.09921432e-01
-7.89114177e-01 9.55705047e-01 3.86855632e-01 -3.16850431e-02
1.12009645e+00 -1.02255702e+00 2.62522221e-01 7.45102406e-01
-1.99867070e-01 5.04710972e-01 5.11437833e-01 -6.61424458e-01
-1.14879251e+00 -1.33964801e+00 3.33580911e-01 -2.00991258e-01
3.88596237e-01 -2.35019505e-01 -9.47912216e-01 6.99095786e-01
4.10295665e-01 -2.72484627e-02 2.90834099e-01 -4.04402405e-01
-2.61606246e-01 -4.99415487e-01 -1.45080090e+00 9.16993618e-01
1.53860331e+00 -6.14472091e-01 -3.97103995e-01 -5.74163571e-02
8.22001517e-01 -3.92203480e-01 -7.34846711e-01 5.06987214e-01
5.33437431e-01 -8.54430556e-01 9.59836006e-01 -2.65852064e-01
7.03835428e-01 -2.65563101e-01 3.79638284e-01 -1.37879074e+00
-6.30772293e-01 -1.05818880e+00 2.54604191e-01 1.72912514e+00
2.40110248e-01 -6.17250443e-01 7.92717159e-01 5.99666834e-01
-1.64138407e-01 -4.66521561e-01 -1.03294468e+00 -8.91256094e-01
4.51687932e-01 4.98445392e-01 5.92596054e-01 8.61531377e-01
-5.86528540e-01 4.25875187e-01 -9.28135872e-01 -6.40463471e-01
6.58783317e-01 -1.74099907e-01 7.12924719e-01 -1.17391682e+00
2.72794813e-01 -5.08545637e-01 -5.07789493e-01 -1.20253861e+00
3.49646866e-01 -7.08360612e-01 -8.78028050e-02 -1.39747858e+00
2.78048813e-01 -2.57576257e-01 -4.94211465e-01 2.67224550e-01
-2.92429715e-01 7.63360918e-01 2.05141470e-01 3.75691026e-01
-3.59117955e-01 1.26686156e+00 2.02675152e+00 -3.87699038e-01
-2.90053993e-01 -2.46188611e-01 -7.56392777e-01 5.04784465e-01
6.81932688e-01 -1.35656744e-01 -5.57495296e-01 -4.49766845e-01
1.70373693e-01 -3.64810377e-01 1.49033517e-01 -9.65055346e-01
-1.18643954e-01 -7.84502849e-02 6.18293941e-01 -4.03815627e-01
3.18123639e-01 -7.73892105e-01 2.80153543e-01 3.82116497e-01
-3.65716338e-01 -1.02908360e-02 3.28069806e-01 5.54577231e-01
-2.54843533e-01 -2.57743318e-02 1.02944481e+00 1.28050642e-05
-5.66399157e-01 5.43493986e-01 1.48304746e-01 1.14851534e-01
1.04030168e+00 -3.67133588e-01 -2.73690283e-01 -3.30642879e-01
-4.87975866e-01 9.23705753e-03 7.65815914e-01 5.37929356e-01
6.87313616e-01 -1.93359518e+00 -6.63736582e-01 8.79988223e-02
-2.81161487e-01 -8.62312391e-02 4.56946254e-01 6.52924776e-01
-8.03333342e-01 1.22268405e-02 -7.69927800e-01 -2.59320617e-01
-1.17457068e+00 5.56776583e-01 2.99220681e-01 -1.78827539e-01
-7.09848642e-01 9.28700566e-01 6.43869817e-01 -4.65548396e-01
8.63426179e-02 -7.24510923e-02 -2.62304634e-01 -1.93528444e-01
2.62788504e-01 6.01563096e-01 -3.77612233e-01 -7.42426395e-01
-7.33682662e-02 1.15185094e+00 -4.48549949e-02 1.87091250e-02
1.21554375e+00 -4.00522202e-01 -4.32526082e-01 2.13069960e-01
1.21972477e+00 8.81343260e-02 -1.53707623e+00 -1.79515451e-01
-6.55820310e-01 -8.36582243e-01 1.30986338e-02 -5.49222112e-01
-1.58145523e+00 9.71936464e-01 7.16472089e-01 -4.04386401e-01
1.08124697e+00 -4.85114455e-01 1.18422914e+00 1.01817764e-01
2.48459369e-01 -1.06936991e+00 3.30401182e-01 2.70987898e-01
1.27107871e+00 -1.11540616e+00 1.71682350e-02 -6.19878113e-01
-6.71658754e-01 1.00816047e+00 8.50175023e-01 -2.94703126e-01
3.13592076e-01 -1.02042504e-01 3.07434015e-02 2.97120064e-01
-2.37405643e-01 6.65783584e-02 4.65253145e-01 4.65907782e-01
3.95742685e-01 -4.33627427e-01 -4.80590731e-01 2.04684526e-01
-1.54088819e-02 2.44039208e-01 2.25427762e-01 6.59632564e-01
3.43812704e-02 -1.38712835e+00 -2.65214324e-01 1.79910272e-01
-3.39799404e-01 -1.33759871e-01 -6.12842381e-01 6.63591027e-01
1.38799369e-01 5.10526776e-01 -8.19300301e-03 -6.37044132e-01
5.59523523e-01 -1.17413454e-01 5.59465468e-01 -6.42377734e-02
-6.28599584e-01 3.39555107e-02 -2.99967021e-01 -5.67489445e-01
-3.90365690e-01 -2.19113052e-01 -9.78886724e-01 -7.97350347e-01
-2.83444375e-01 -8.61587748e-02 2.34551996e-01 6.47677898e-01
4.54638720e-01 7.47597218e-01 8.58464777e-01 -9.22790706e-01
-1.63940117e-01 -9.11428213e-01 -7.75596440e-01 7.60462224e-01
2.72367913e-02 -7.45813370e-01 -4.60503638e-01 1.45795673e-01] | [11.537464141845703, -0.7198633551597595] |
27f403f5-1478-4d81-b210-be92b6a2cffa | embedding-based-entity-alignment-using | null | null | https://ieeexplore.ieee.org/document/9194492/authors | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9194492 | Embedding-Based Entity Alignment Using Relation Structural Similarity | Entity alignment aims to find entities in different knowledge graphs that semantically represent the same real-world entity. Recently, embedding-based entity alignment methods, which represent knowledge graphs as low-dimensional embeddings and perform entity alignments by measuring the similarity between entity embeddings, have achieved promising results. Most of these methods mainly focus on improving the knowledge graph embedding model or leveraging attributes to obtain more semantic information. However, the structural similarity between the two relations (considering all entities attached on the two relations) in different KGs has not been utilized in the existing methods. In this paper, we propose a novel embedding-based entity alignment method that takes the advantages of relation structural similarity. Specifically, our method first jointly learns the embeddings of two knowledge graphs in a uniform vector space, using the entity pairs regarding to the seed alignments (the alignments already known) that each shares the same embedding. Then, it iteratively computes the structural similarity between the relations in different knowledge graphs according to the seed alignments and the alignments with high reliability generated during training, which makes the embeddings of relations with high similarity closer to each other. Experimental results on five widely used real-world datasets show that the proposed approach significantly outperforms the state-of-the-art embedding-based ones for entity alignment. | ['Yanhui Peng; Jing Zhang; Cangqi Zhou; Jian Xu'] | 2022-09-11 | null | null | null | 2020-ieee-international-conference-on-4 | ['knowledge-graph-embedding', 'entity-alignment', 'entity-embeddings', 'entity-alignment'] | ['graphs', 'knowledge-base', 'methodology', 'natural-language-processing'] | [-4.69553828e-01 3.48412037e-01 -3.46772194e-01 -1.67528898e-01
-1.47961095e-01 -3.65587294e-01 5.31052947e-01 9.04622912e-01
-4.02263165e-01 4.24670726e-01 3.25990975e-01 1.45352468e-01
-6.40055954e-01 -1.28328216e+00 -3.27437997e-01 -6.05630040e-01
-1.47904783e-01 6.18928850e-01 3.45776618e-01 -2.75223494e-01
3.88560519e-02 1.71989530e-01 -1.24572873e+00 -1.87730327e-01
1.08670461e+00 7.45792568e-01 -5.78934029e-02 -5.28574884e-02
-4.67456609e-01 4.45902109e-01 -4.19659644e-01 -8.46467614e-01
4.46566641e-02 -2.74043620e-01 -8.56843352e-01 -2.17026934e-01
9.56154987e-02 2.57349163e-01 -6.33109450e-01 1.18805933e+00
3.77193868e-01 1.96883038e-01 6.62769258e-01 -1.42038167e+00
-1.36014259e+00 5.61249554e-01 -3.86857599e-01 1.89675376e-01
3.81319433e-01 -5.07840574e-01 1.41282558e+00 -8.20663393e-01
7.45719016e-01 1.09280372e+00 5.57516038e-01 -8.33666697e-03
-1.04114974e+00 -4.09319043e-01 1.22032300e-01 9.21447456e-01
-1.83237064e+00 1.21533900e-01 8.26969802e-01 -5.00521123e-01
1.06818354e+00 -8.07113796e-02 8.26975226e-01 5.10934114e-01
-5.66957071e-02 2.91138083e-01 5.51634908e-01 -4.66587991e-01
6.84416667e-02 4.68743801e-01 2.48602524e-01 5.97092628e-01
6.64133132e-01 -4.23576444e-01 -3.42166513e-01 -2.90090263e-01
3.67284596e-01 -7.13765472e-02 -4.98879880e-01 -9.22139406e-01
-1.35703897e+00 8.39366972e-01 9.57070947e-01 9.23876524e-01
-5.86575031e-01 -1.77830875e-01 3.36722910e-01 5.14188111e-02
3.38480204e-01 6.42316580e-01 -4.19138014e-01 3.09503853e-01
-2.15165645e-01 -8.70877579e-02 7.18598127e-01 1.02613842e+00
1.12520945e+00 -2.75845468e-01 1.32948056e-01 5.47350764e-01
6.31990671e-01 6.21178150e-02 6.95436895e-01 -3.89306284e-02
6.31504595e-01 1.41432381e+00 1.02798544e-01 -1.92859602e+00
-2.20197648e-01 -3.73291850e-01 -5.70691466e-01 -4.11700875e-01
4.65532318e-02 8.94527808e-02 -3.37212831e-01 1.71595514e+00
7.56349206e-01 8.69586825e-01 6.61730111e-01 7.14692831e-01
1.02251577e+00 6.64910197e-01 1.15542173e-01 7.44926790e-03
1.47578824e+00 -9.16017950e-01 -8.40218067e-01 6.23482242e-02
1.08257091e+00 -6.29702210e-01 5.92280746e-01 -4.44080412e-01
-4.36858147e-01 -4.98407423e-01 -1.31294394e+00 1.91800982e-01
-8.66114914e-01 1.35293722e-01 5.54904699e-01 4.77359205e-01
-8.65076959e-01 5.62526047e-01 -4.86292094e-01 -6.49254858e-01
5.26030688e-03 2.96403527e-01 -8.68475795e-01 4.15657237e-02
-1.64630854e+00 1.15501380e+00 1.14612162e+00 1.23176925e-01
-2.45092418e-02 -5.19833505e-01 -1.19483173e+00 2.60856658e-01
5.66014469e-01 -6.58391297e-01 2.72444218e-01 -4.37485427e-01
-9.10578072e-01 7.42053032e-01 3.77692021e-02 -2.09588960e-01
-1.72630653e-01 -1.35576338e-01 -9.62169468e-01 -4.65051979e-02
1.06458704e-03 1.42332226e-01 3.35588962e-01 -1.21729171e+00
-5.37909985e-01 -5.28268695e-01 2.32731327e-01 4.55896288e-01
-9.47038829e-01 -3.52719277e-01 -4.22458619e-01 -2.98407376e-01
2.77862251e-01 -6.78405344e-01 -5.60995133e-04 -3.31095695e-01
-2.93409884e-01 -6.27659440e-01 7.06793606e-01 -5.75803936e-01
1.57248759e+00 -2.17839694e+00 6.00807011e-01 4.16646361e-01
3.24999958e-01 5.20793855e-01 -1.72550336e-01 7.84280360e-01
-1.66546822e-01 1.48421019e-01 6.86063096e-02 1.91519082e-01
4.28529009e-02 2.84810364e-01 1.16797693e-01 3.18766475e-01
1.35829806e-01 9.25661564e-01 -1.33035076e+00 -7.47096062e-01
1.81277186e-01 4.05543625e-01 -1.73922136e-01 4.04503465e-01
1.69111118e-01 5.10293469e-02 -6.17945731e-01 2.10010305e-01
4.87552404e-01 -2.25827470e-01 7.85411716e-01 -7.91343570e-01
2.80047506e-01 -1.00748487e-01 -1.48155785e+00 1.60153842e+00
-3.87712747e-01 2.04811826e-01 -6.04195535e-01 -1.22548676e+00
1.21851218e+00 3.77418667e-01 4.87278223e-01 -4.89601880e-01
2.86705792e-02 3.78571957e-01 4.67798747e-02 -6.66918576e-01
6.79555237e-01 2.01234996e-01 5.67303337e-02 1.14300273e-01
4.66888070e-01 1.84965640e-01 1.64710373e-01 3.59808385e-01
8.54121506e-01 -1.42197251e-01 6.81514502e-01 -1.53395412e-02
7.89633036e-01 -3.32946420e-01 5.71462929e-01 -1.05453387e-03
1.51927043e-02 -1.65182743e-02 4.19245809e-01 -2.71576554e-01
-1.02419913e+00 -1.16605759e+00 3.11680287e-02 4.79869455e-01
7.44963050e-01 -9.49584007e-01 -3.37701380e-01 -9.71343160e-01
2.68734932e-01 4.25474316e-01 -8.19381177e-01 -4.60901916e-01
-4.01071429e-01 -5.11863232e-01 3.38208526e-01 6.15552187e-01
3.12312424e-01 -7.46968985e-01 2.31156480e-02 3.48660380e-01
-5.32850288e-02 -1.14214945e+00 -1.54365748e-01 -2.60750711e-01
-6.53956175e-01 -1.28989446e+00 -4.58592594e-01 -1.18782997e+00
9.04967248e-01 1.61387652e-01 7.73383617e-01 3.69645685e-01
-2.00856663e-02 2.90252715e-01 -7.36897469e-01 1.10514723e-01
-1.66281208e-01 2.69081563e-01 2.32444257e-01 3.45672101e-01
5.90745926e-01 -7.04179227e-01 -3.37824583e-01 3.72557074e-01
-8.91191185e-01 -2.45127156e-01 5.89972913e-01 8.43001783e-01
6.34611011e-01 2.72240251e-01 6.58385158e-01 -7.83711135e-01
7.71334469e-01 -8.71721029e-01 -2.54297376e-01 8.08173597e-01
-9.01842415e-01 4.79027092e-01 6.03924751e-01 -3.83204699e-01
-6.47293746e-01 -5.05142391e-01 2.31336713e-01 -4.37371641e-01
1.44384250e-01 1.02884960e+00 -2.99639821e-01 -1.48844961e-02
4.27546024e-01 2.38686055e-01 -3.93004417e-01 -3.45483929e-01
8.75362337e-01 5.51071465e-01 2.89712667e-01 -5.31960130e-01
1.06421638e+00 1.35967270e-01 4.08484936e-02 -4.28284347e-01
-5.47036767e-01 -6.94250345e-01 -8.57217550e-01 2.17305720e-01
9.28050876e-01 -7.22375154e-01 -4.45534259e-01 7.88840726e-02
-1.21563137e+00 5.34987271e-01 -3.05098653e-01 9.02818620e-01
-1.38448685e-01 6.70311630e-01 -2.31289655e-01 -4.71742898e-01
-2.31234610e-01 -8.03105056e-01 6.20379388e-01 5.82747698e-01
-2.08958060e-01 -1.19779539e+00 5.13182998e-01 -9.00269747e-02
2.36314997e-01 3.38617206e-01 1.41744387e+00 -1.20045280e+00
-4.94929671e-01 -3.23996127e-01 -3.84945422e-01 2.87600439e-02
6.03405654e-01 2.40062065e-02 -4.19665992e-01 -2.22009107e-01
-5.85319936e-01 8.63701180e-02 3.87572587e-01 -2.58393914e-01
3.38003337e-01 -3.22914392e-01 -7.53608823e-01 3.43827933e-01
1.63617253e+00 2.16898948e-01 6.45399034e-01 5.54380357e-01
1.02982950e+00 6.00726724e-01 8.68065953e-01 1.39964953e-01
4.57808793e-01 1.00551462e+00 3.31959486e-01 1.33863240e-01
2.57486969e-01 -5.13325810e-01 -3.38952877e-02 1.49710929e+00
-1.76287234e-01 -2.79942572e-01 -6.70189798e-01 1.04815388e+00
-2.12546015e+00 -8.20305407e-01 -2.07184464e-01 2.26292157e+00
6.48535430e-01 -2.02918462e-02 -2.50294000e-01 5.17496094e-02
8.91641378e-01 3.34348977e-01 -1.81047484e-01 -3.07441205e-01
-3.62730287e-02 2.28814021e-01 2.13154659e-01 3.80284518e-01
-8.98471177e-01 9.69181538e-01 4.42924166e+00 8.16287339e-01
-5.33173800e-01 2.32895017e-01 -2.38897443e-01 5.14318645e-01
-5.72740376e-01 3.77840817e-01 -6.95448160e-01 4.47553784e-01
5.63764632e-01 -8.25058043e-01 1.17965743e-01 8.43599617e-01
-4.45867330e-01 2.31777415e-01 -1.01785767e+00 8.67894948e-01
2.07925752e-01 -1.19976878e+00 2.88904995e-01 -1.07363418e-01
8.51193786e-01 -5.12199163e-01 -1.51704386e-01 3.19089532e-01
2.71939784e-02 -7.90814757e-01 1.62792504e-01 6.62286937e-01
4.00504619e-01 -7.97731936e-01 1.35377371e+00 -1.48937881e-01
-1.88377357e+00 2.49534175e-01 -6.46318257e-01 3.17826331e-01
2.50906587e-01 6.18937969e-01 -9.87729788e-01 1.55315304e+00
3.95251244e-01 9.34989452e-01 -7.40619659e-01 1.07352781e+00
-6.53442323e-01 2.30306417e-01 -4.66054827e-02 -1.37298256e-01
4.42320034e-02 -5.56961894e-01 4.21958864e-01 8.33219528e-01
4.71224993e-01 -3.64889950e-02 5.20411208e-02 6.53495669e-01
-2.61303157e-01 7.81936228e-01 -7.83152640e-01 -3.10371697e-01
9.54034507e-01 1.29568005e+00 -3.49538386e-01 -4.46209460e-01
-6.80798769e-01 9.23135877e-01 7.56453335e-01 1.20356821e-01
-8.65038097e-01 -7.61834919e-01 8.64009082e-01 -2.34266698e-01
2.80369401e-01 -2.65636295e-01 3.62470746e-01 -1.13767862e+00
2.68508434e-01 -3.16595554e-01 5.44345319e-01 -6.62259936e-01
-1.43436503e+00 7.39002585e-01 3.03350072e-02 -1.25583017e+00
-1.99810714e-02 -3.08999896e-01 -5.23505926e-01 8.32762003e-01
-1.42979670e+00 -1.12514400e+00 -4.61044103e-01 4.67614442e-01
-2.31279537e-01 -1.57489851e-01 1.23259127e+00 4.82555211e-01
-6.13607407e-01 6.69002175e-01 2.46209189e-01 3.75946909e-01
6.35944009e-01 -1.21202719e+00 3.19144696e-01 6.27063513e-01
5.15818119e-01 8.83010387e-01 3.47612590e-01 -7.75731742e-01
-1.11461627e+00 -9.81857300e-01 1.17784321e+00 -1.98396385e-01
9.73678529e-01 1.00902796e-01 -1.24222660e+00 7.59047747e-01
3.63536298e-01 1.42975032e-01 9.60511923e-01 4.35082048e-01
-5.13578773e-01 -1.79170713e-01 -9.63763416e-01 5.35440326e-01
1.08653784e+00 -5.47374070e-01 -1.01963174e+00 1.08489573e-01
8.63330245e-01 -1.31463334e-01 -1.45764220e+00 4.77906108e-01
4.21292812e-01 -4.69182312e-01 9.81897652e-01 -1.04275024e+00
2.39774987e-01 -7.27232456e-01 -1.50629282e-01 -1.78333676e+00
-4.30518508e-01 1.70990124e-01 -3.60152781e-01 1.56323481e+00
3.65839303e-01 -9.10792828e-01 5.65458953e-01 1.58061579e-01
2.45664164e-01 -9.70956683e-01 -9.36283588e-01 -1.07102489e+00
-3.36381495e-01 4.18670774e-01 1.09514534e+00 1.54888821e+00
3.69591624e-01 4.49451208e-01 -1.06824242e-01 5.66169858e-01
2.93945611e-01 3.06408435e-01 8.26519489e-01 -1.50792253e+00
-7.51488879e-02 -2.12736413e-01 -1.55375898e+00 -4.20288622e-01
3.37546915e-01 -1.27766955e+00 -4.86242831e-01 -1.87851322e+00
1.88304678e-01 -5.95559120e-01 -6.28081024e-01 4.16734517e-01
-6.54916644e-01 -3.00090879e-01 1.94951575e-02 5.73765747e-02
-5.70838511e-01 9.48269725e-01 7.98628151e-01 -3.07480037e-01
-1.05219528e-01 -5.76889217e-01 -6.29674494e-01 5.11198878e-01
6.05674684e-01 -5.83700180e-01 -5.57888329e-01 -4.67562497e-01
3.73124301e-01 -4.04920459e-01 5.62269203e-02 -1.07616663e+00
3.27539444e-01 -9.38007534e-02 5.48332259e-02 -6.02886453e-02
1.98962107e-01 -1.20014775e+00 4.93884385e-01 2.63234645e-01
-1.98118147e-02 1.99728504e-01 -8.96363892e-03 9.64992940e-01
-5.68237901e-01 -3.93670440e-01 3.61160606e-01 2.97823876e-01
-1.05616581e+00 3.27032775e-01 3.62713963e-01 1.67375520e-01
1.51972067e+00 -1.28224581e-01 -3.63743842e-01 -9.21971425e-02
-7.44879425e-01 2.08533764e-01 5.00208676e-01 6.08801663e-01
6.33513272e-01 -1.85324407e+00 -6.56030238e-01 -2.31001541e-01
7.08923101e-01 -1.54885128e-01 1.84139326e-01 7.34490931e-01
-2.40483418e-01 2.17747718e-01 -2.08470285e-01 -2.54961699e-01
-1.32649672e+00 8.44936192e-01 1.12768516e-01 -6.55505896e-01
-3.59555393e-01 7.44203448e-01 -9.35819447e-02 -4.41901147e-01
-1.79025486e-01 -4.36481014e-02 -6.69004321e-01 3.68711829e-01
1.41265571e-01 2.33249754e-01 7.86054805e-02 -9.94994998e-01
-6.76222444e-01 9.50784743e-01 -1.49852082e-01 2.06828326e-01
1.23501432e+00 1.58583194e-01 -1.95162177e-01 3.23421508e-01
1.45432544e+00 3.08218509e-01 -2.25559160e-01 -8.14198136e-01
2.54230976e-01 -8.07054102e-01 -4.20767218e-01 -1.03937760e-01
-1.18188572e+00 5.67012787e-01 4.51552004e-01 3.28796297e-01
7.88981795e-01 2.39959463e-01 7.33517110e-01 4.36318755e-01
6.00412548e-01 -8.09326649e-01 5.66660501e-02 1.54748395e-01
4.12451535e-01 -8.90237093e-01 1.96256042e-01 -7.00499415e-01
-5.50408542e-01 1.06806445e+00 8.53676498e-01 -7.54602849e-02
6.97981119e-01 -4.97635871e-01 -2.03003228e-01 -5.07412434e-01
-1.95859566e-01 -4.72123712e-01 5.04454076e-01 7.50874698e-01
1.69532180e-01 2.12280363e-01 -8.88672590e-01 5.37333548e-01
3.32275815e-02 -3.26922685e-01 1.70560598e-01 7.67259061e-01
-3.77814502e-01 -1.33298707e+00 9.83485430e-02 3.55134457e-01
1.25599205e-01 -2.73730252e-02 -4.30253923e-01 7.58006752e-01
2.82369196e-01 8.04832935e-01 -1.28781170e-01 -7.52845645e-01
5.54371595e-01 1.08924210e-01 3.99289370e-01 -7.19398797e-01
-1.80546001e-01 -7.72129834e-01 9.84289423e-02 -3.08766246e-01
-5.36880195e-01 -2.19391257e-01 -1.38184869e+00 -1.64187163e-01
-9.15943146e-01 6.30484045e-01 3.83069426e-01 9.68880713e-01
6.60467029e-01 4.63788539e-01 6.87013268e-01 -2.04546869e-01
-2.14839131e-01 -8.84437799e-01 -8.53216708e-01 9.03240860e-01
-4.45693970e-01 -1.14165974e+00 -2.14992836e-01 -2.55289495e-01] | [8.840375900268555, 7.816086769104004] |
609ac1a7-5443-4cab-97d2-122130a51b0b | lexical-simplification-benchmarks-for-english | 2209.05301 | null | https://arxiv.org/abs/2209.05301v1 | https://arxiv.org/pdf/2209.05301v1.pdf | Lexical Simplification Benchmarks for English, Portuguese, and Spanish | Even in highly-developed countries, as many as 15-30\% of the population can only understand texts written using a basic vocabulary. Their understanding of everyday texts is limited, which prevents them from taking an active role in society and making informed decisions regarding healthcare, legal representation, or democratic choice. Lexical simplification is a natural language processing task that aims to make text understandable to everyone by replacing complex vocabulary and expressions with simpler ones, while preserving the original meaning. It has attracted considerable attention in the last 20 years, and fully automatic lexical simplification systems have been proposed for various languages. The main obstacle for the progress of the field is the absence of high-quality datasets for building and evaluating lexical simplification systems. We present a new benchmark dataset for lexical simplification in English, Spanish, and (Brazilian) Portuguese, and provide details about data selection and annotation procedures. This is the first dataset that offers a direct comparison of lexical simplification systems for three languages. To showcase the usability of the dataset, we adapt two state-of-the-art lexical simplification systems with differing architectures (neural vs.\ non-neural) to all three languages (English, Spanish, and Brazilian Portuguese) and evaluate their performances on our new dataset. For a fairer comparison, we use several evaluation measures which capture varied aspects of the systems' efficacy, and discuss their strengths and weaknesses. We find a state-of-the-art neural lexical simplification system outperforms a state-of-the-art non-neural lexical simplification system in all three languages. More importantly, we find that the state-of-the-art neural lexical simplification systems perform significantly better for English than for Spanish and Portuguese. | ['Horacio Saggion', 'Marcos Zampieri', 'Kai North', 'Matthew Shardlow', 'Daniel Ferres', 'Sanja Stajner'] | 2022-09-12 | null | null | null | null | ['lexical-simplification'] | ['natural-language-processing'] | [ 8.29776004e-02 2.53227085e-01 -3.14288467e-01 -3.10521901e-01
-4.70187783e-01 -2.87018538e-01 6.29879117e-01 6.36287808e-01
-1.04612732e+00 8.24268699e-01 4.55914617e-01 -3.55522096e-01
6.29626438e-02 -8.02464664e-01 -1.50160015e-01 -2.12400749e-01
6.03434861e-01 7.36326039e-01 -4.44255918e-02 -7.77355909e-01
-5.99549338e-02 5.92745245e-01 -1.50246489e+00 2.44102091e-01
9.90895331e-01 4.19371337e-01 -2.38752328e-02 3.19706053e-01
-4.14780080e-01 6.30099475e-01 -7.24649906e-01 -8.06554377e-01
1.11349955e-01 -2.06861243e-01 -9.19385493e-01 -6.63435698e-01
5.66085696e-01 -3.49765643e-02 -3.77857804e-01 1.11850119e+00
7.97970653e-01 1.57774523e-01 5.86912930e-01 -6.19525731e-01
-7.46529520e-01 8.36248279e-01 -1.06404524e-03 1.86042294e-01
6.20833874e-01 7.40447789e-02 9.54725683e-01 -6.52910531e-01
1.10048425e+00 1.40925694e+00 1.00864410e+00 6.76083028e-01
-1.03519869e+00 -6.10700011e-01 1.67631418e-01 8.69024917e-02
-1.68361807e+00 -6.42586946e-01 4.06310230e-01 -2.88267523e-01
1.61105728e+00 3.98473054e-01 6.10225558e-01 1.02230489e+00
4.07203346e-01 4.93457347e-01 7.22497404e-01 -7.73548961e-01
-1.21511184e-01 1.47744671e-01 3.53890866e-01 6.21757567e-01
5.37279248e-01 -2.96658486e-01 -3.68747175e-01 1.69550779e-03
7.11734593e-02 -2.35359222e-01 -2.10680738e-01 1.33659557e-01
-1.18249357e+00 9.65153635e-01 9.72210169e-02 7.22969770e-01
-2.29343578e-01 -2.50528544e-01 8.22427392e-01 4.66343373e-01
4.87775236e-01 7.53168225e-01 -5.21837831e-01 -1.58853129e-01
-1.01753795e+00 6.07351720e-01 1.14298344e+00 8.33106816e-01
5.11392176e-01 1.47159681e-01 -2.52912462e-01 9.81231391e-01
-1.14957228e-01 6.91018343e-01 5.79901516e-01 -6.75580978e-01
5.21289825e-01 6.38837397e-01 -2.48850092e-01 -9.30478692e-01
-8.92444730e-01 -9.80649069e-02 -1.17852795e+00 2.07056804e-03
3.73175621e-01 -7.12119266e-02 -7.93574572e-01 1.76142812e+00
-4.48816195e-02 -6.83220744e-01 3.04495823e-02 3.85138422e-01
1.37757897e+00 5.08461773e-01 3.40740055e-01 -4.39138144e-01
1.40688276e+00 -6.66421413e-01 -9.85134423e-01 -7.30737895e-02
7.87302196e-01 -9.63678956e-01 1.32699931e+00 4.92879182e-01
-1.20692790e+00 -5.10885954e-01 -9.39838171e-01 -7.25858629e-01
-9.57802951e-01 -8.53060409e-02 5.51087797e-01 7.30206966e-01
-7.85426378e-01 7.00027168e-01 -5.51484346e-01 -7.51899660e-01
2.42654532e-01 4.74751383e-01 -6.18732810e-01 -2.53209081e-02
-1.53777647e+00 1.49610460e+00 4.79857147e-01 -3.49319547e-01
-1.76045328e-01 -7.47004509e-01 -1.00270665e+00 -2.84407903e-02
2.76298225e-01 -9.93050635e-01 1.22208107e+00 -7.38387644e-01
-1.39383721e+00 1.21414196e+00 -1.93360865e-01 -5.24362385e-01
4.54868287e-01 -3.52693111e-01 -5.30630887e-01 -2.93651611e-01
2.00975180e-01 7.06018507e-01 4.95884985e-01 -7.25482166e-01
-6.39325082e-01 -1.24457628e-01 1.44968227e-01 3.13627958e-01
-1.83288544e-01 3.19233984e-01 -5.16465902e-01 -1.03333151e+00
-5.19297838e-01 -8.48914921e-01 -1.56329677e-01 -1.91603333e-01
-4.24723476e-01 -3.69973660e-01 4.53850448e-01 -8.96140754e-01
1.88185036e+00 -2.01779819e+00 3.58693331e-01 -7.88215399e-02
3.17364365e-01 8.18858087e-01 -4.99648713e-02 4.95972931e-01
-1.42686740e-01 5.93605399e-01 -3.52868140e-01 -6.72128737e-01
2.48568639e-01 4.43981528e-01 -4.15013283e-02 3.30631346e-01
-4.71065268e-02 9.36596811e-01 -6.23032749e-01 -5.21239758e-01
3.80326092e-01 6.74785376e-01 -6.42590940e-01 -2.25378215e-01
-8.61777887e-02 1.77284211e-01 -7.08616003e-02 4.39743131e-01
2.61466771e-01 3.28301817e-01 8.69203582e-02 -1.71412513e-01
-1.91496447e-01 4.88943458e-01 -1.08282924e+00 1.62596905e+00
-6.21375501e-01 8.61280322e-01 -1.10673279e-01 -7.27260709e-01
4.39560175e-01 2.41191536e-01 1.72165826e-01 -1.09057832e+00
3.88852447e-01 3.39128166e-01 1.54780179e-01 -4.04617131e-01
8.19869518e-01 -3.70315135e-01 -4.14465845e-01 4.08789426e-01
-3.78378183e-02 -4.64965284e-01 8.60489309e-01 -4.40611616e-02
8.08899939e-01 -2.26047590e-01 1.05053759e+00 -4.21939552e-01
7.96633005e-01 -2.05957085e-01 4.34332371e-01 7.92747736e-01
-8.49346593e-02 6.17308378e-01 3.00891757e-01 -8.05633247e-01
-1.14190435e+00 -7.35647082e-01 -1.97493702e-01 1.28220022e+00
-3.90350342e-01 -7.61242270e-01 -9.42755520e-01 -3.82736057e-01
3.75344455e-02 1.03826189e+00 -7.07991898e-01 -1.12176865e-01
-8.99729490e-01 -6.47162437e-01 1.02554798e+00 1.34195656e-01
3.13084811e-01 -1.57020950e+00 -5.08175313e-01 1.57622799e-01
-3.83374602e-01 -1.08319485e+00 -2.93658793e-01 1.37357339e-01
-4.84140307e-01 -1.00094879e+00 -4.52531397e-01 -6.78452730e-01
2.36289799e-01 -1.97426245e-01 1.52590477e+00 1.65876344e-01
-1.86414421e-01 7.75685683e-02 -2.46455520e-01 -8.51374686e-01
-8.14995527e-01 7.82501101e-01 2.41161719e-01 -6.86357975e-01
5.55206954e-01 -3.56687397e-01 -4.56833420e-03 -2.31092691e-01
-1.17114925e+00 -1.20081924e-01 4.66154695e-01 6.80075049e-01
6.85356379e-01 6.70045102e-03 3.61520499e-01 -1.23150206e+00
1.19112527e+00 -1.03694543e-01 -2.50484526e-01 1.42183453e-01
-5.56818545e-01 2.43767306e-01 1.00707018e+00 -2.89619863e-01
-8.67612422e-01 -2.89643288e-01 -7.23569512e-01 1.58812582e-01
-2.25324675e-01 7.32187569e-01 -2.69457757e-01 2.91002691e-02
9.24323618e-01 -1.71380728e-01 -1.79322526e-01 -6.37511671e-01
6.20179236e-01 6.23816550e-01 5.10751307e-01 -4.18712527e-01
5.65958202e-01 3.60236406e-01 -9.70321372e-02 -1.46891725e+00
-9.24223661e-01 -3.31181318e-01 -8.83445978e-01 4.33048040e-01
8.46728504e-01 -6.16613984e-01 -5.15244484e-01 4.51813430e-01
-1.35471487e+00 -1.99525595e-01 -7.93150067e-01 1.88287690e-01
-2.56989717e-01 4.82021987e-01 -6.07627749e-01 -1.54644862e-01
-7.31182218e-01 -1.18225300e+00 1.01744342e+00 2.22061854e-02
-9.64865625e-01 -1.17270768e+00 2.11363763e-01 6.48913532e-02
6.12579644e-01 1.68051049e-01 1.28720403e+00 -7.92774677e-01
2.41648763e-01 -3.12129855e-01 -1.69985920e-01 4.14586246e-01
1.16618343e-01 1.30600296e-02 -6.95684314e-01 -1.36090294e-01
-1.01067178e-01 -1.03239141e-01 8.46006751e-01 4.06149000e-01
7.67049491e-01 -2.37921774e-01 -1.69567198e-01 6.72135055e-01
1.02829361e+00 -9.07661766e-02 6.84406817e-01 4.73602384e-01
8.85047853e-01 4.05548424e-01 2.52249897e-01 1.41953498e-01
6.46052480e-01 7.78429687e-01 -4.62732166e-02 -2.13314489e-01
-4.56625193e-01 2.03998759e-01 2.13783309e-01 1.18718100e+00
-1.78562492e-01 -3.02230060e-01 -9.95122790e-01 4.80105549e-01
-1.60808301e+00 -7.78530061e-01 6.94816606e-03 2.14353013e+00
1.20441389e+00 2.77594447e-01 1.21987211e-02 2.73456633e-01
3.41947109e-01 1.99037537e-01 -2.86018997e-01 -9.53284502e-01
-5.12974262e-01 6.45301104e-01 2.63036430e-01 8.12548816e-01
-1.28439510e+00 1.24877107e+00 6.61471748e+00 1.08321512e+00
-1.18923819e+00 2.38487329e-02 3.61244410e-01 -1.06363349e-01
-2.33635023e-01 -4.28861320e-01 -9.12649214e-01 8.16706344e-02
1.00369918e+00 -3.58318597e-01 5.82305789e-01 4.94026303e-01
2.50577688e-01 6.88110888e-02 -1.16157246e+00 1.07310152e+00
2.53170580e-01 -1.14388108e+00 5.03465414e-01 -2.16174930e-01
5.79399824e-01 2.32105851e-01 -1.88821793e-01 5.51048875e-01
1.28683209e-01 -1.39726079e+00 7.43341565e-01 7.34736681e-01
9.81251478e-01 -8.62581611e-01 9.87186015e-01 3.76095355e-01
-9.99820113e-01 2.10001454e-01 -3.20427179e-01 -3.59449357e-01
2.08695129e-01 5.23096979e-01 -2.82471687e-01 4.92360979e-01
6.97329462e-01 6.29453480e-01 -7.04230368e-01 5.23626387e-01
-3.53269041e-01 2.45886654e-01 -4.77446824e-01 -6.23223633e-02
1.11181043e-01 -2.39370540e-02 6.61286056e-01 1.76491559e+00
-1.15444297e-02 6.86875582e-02 9.02357027e-02 3.68605524e-01
-2.54277736e-01 8.00697446e-01 -7.04562604e-01 -1.06842853e-01
3.59967828e-01 9.43139017e-01 -5.51727414e-01 -4.65197146e-01
-5.21305799e-01 7.31342971e-01 4.89860773e-01 4.59660329e-02
-7.15051889e-01 -7.37379491e-01 7.42477894e-01 5.32673858e-02
-6.97196752e-04 -1.10600427e-01 -6.18754864e-01 -1.26648200e+00
-3.20978425e-02 -1.41354871e+00 4.06211197e-01 -4.10374910e-01
-1.25696290e+00 8.11695457e-01 5.93084544e-02 -5.99313736e-01
-3.07371080e-01 -6.52687132e-01 -1.17435366e-01 1.00260711e+00
-1.41532230e+00 -1.16817975e+00 -1.50707886e-01 4.67550933e-01
5.49106896e-01 -3.41929585e-01 1.22240603e+00 6.17093265e-01
-6.23615563e-01 7.41731286e-01 3.86710605e-03 4.02667187e-02
8.74854863e-01 -1.08389819e+00 7.22363472e-01 7.31379509e-01
1.59873031e-02 8.60748768e-01 7.47571647e-01 -5.67625046e-01
-7.86701739e-01 -9.81929004e-01 1.67300749e+00 -6.09850585e-01
5.32072008e-01 -3.35916609e-01 -9.28990722e-01 6.56244934e-01
4.90821540e-01 -4.99244362e-01 6.79934680e-01 2.50413567e-01
-2.42607594e-01 3.96959968e-02 -9.89638448e-01 1.10767591e+00
1.12996054e+00 -5.62266886e-01 -9.49440897e-01 2.28156328e-01
8.16538930e-01 -4.27907139e-01 -8.69361877e-01 4.81761009e-01
6.79721057e-01 -8.79502535e-01 9.36416149e-01 -7.10711479e-01
1.67516664e-01 -5.37633970e-02 -1.32323831e-01 -1.16900003e+00
-3.56834918e-01 -8.70614648e-01 1.40049487e-01 1.04502833e+00
4.86995101e-01 -6.85933650e-01 4.43053305e-01 3.44966054e-01
-2.05317229e-01 -6.54288232e-01 -9.88300681e-01 -3.95368695e-01
7.36565053e-01 -6.31106734e-01 6.29720867e-01 1.03241968e+00
-1.59497991e-01 7.62399077e-01 -1.70324326e-01 -4.02096897e-01
-2.59123277e-02 -2.88721889e-01 9.17811394e-01 -1.48023927e+00
2.15267852e-01 -1.07431126e+00 -2.86886007e-01 -5.75917125e-01
5.19287765e-01 -9.92328405e-01 -4.19513077e-01 -1.77614701e+00
2.22199876e-02 -1.17453203e-01 1.25525191e-01 7.27932990e-01
-2.28901342e-01 3.28734845e-01 4.13789451e-01 7.38654658e-02
-5.76162934e-01 4.58713502e-01 9.82956529e-01 -2.81857371e-01
-2.55967885e-01 -3.98759246e-01 -9.26492631e-01 1.02702475e+00
7.38403678e-01 -5.31739473e-01 -1.94696367e-01 -5.22116125e-01
5.76199353e-01 -5.60498595e-01 -3.05587679e-01 -8.76534879e-01
9.67545435e-02 8.93620551e-02 -1.14001641e-02 -3.12010765e-01
1.17718093e-01 -5.83924890e-01 8.05817619e-02 5.04002452e-01
-1.01774931e-01 2.70596743e-01 5.07029295e-01 -2.44905233e-01
-2.15580538e-01 -2.34314442e-01 9.51399624e-01 -1.92401305e-01
-5.19301713e-01 -4.97389250e-02 -6.23671114e-01 3.98323625e-01
6.75820768e-01 -1.60885677e-01 -3.02468598e-01 -3.85544777e-01
-8.41784835e-01 -6.59591109e-02 5.77672601e-01 5.65587878e-01
2.23207936e-01 -9.82590556e-01 -9.29387212e-01 2.23841444e-01
4.98932414e-02 6.17307127e-02 -5.24632670e-02 6.96298957e-01
-1.02615607e+00 8.56621027e-01 -1.98213398e-01 -9.06053483e-02
-1.33224905e+00 6.17455602e-01 4.30662364e-01 -8.63386333e-01
-5.18378079e-01 4.23226684e-01 7.79795721e-02 -9.89798784e-01
2.62850195e-01 -8.39816570e-01 -4.79501456e-01 2.55521119e-01
4.90040421e-01 6.12361848e-01 4.72567320e-01 -1.03451002e+00
-3.58159691e-01 7.00269401e-01 -2.27000549e-01 1.90102011e-01
1.21984649e+00 -8.84070620e-02 -3.79658014e-01 3.76628608e-01
9.07729506e-01 3.64349574e-01 -2.75744237e-02 -1.91887602e-01
-7.50899613e-02 2.18840480e-01 -1.56472877e-01 -8.44306707e-01
-7.44585812e-01 8.40739012e-01 2.17955619e-01 2.21797347e-01
1.17378044e+00 -3.98578823e-01 8.68249059e-01 8.23964536e-01
1.65814251e-01 -1.14369249e+00 -5.92844307e-01 1.11010277e+00
9.09079373e-01 -1.08944893e+00 4.45381939e-01 -4.89150941e-01
-5.43150306e-01 1.03743720e+00 1.09158620e-01 5.59591204e-02
7.00203300e-01 1.45674720e-01 2.76045442e-01 -2.47123986e-01
-3.76213610e-01 -3.15079063e-01 4.60583985e-01 5.95300853e-01
7.07701683e-01 1.48914233e-01 -7.95554459e-01 7.45433331e-01
-7.82020032e-01 -1.07997037e-01 4.07002956e-01 7.52356350e-01
-1.99125394e-01 -1.45120406e+00 -1.88502252e-01 4.95552182e-01
-8.31414938e-01 -5.24937928e-01 -7.21689701e-01 1.41112196e+00
4.31046844e-01 7.59099722e-01 -2.56765544e-01 1.94567107e-02
9.16451395e-01 1.75933763e-01 4.54421848e-01 -8.44166279e-01
-1.00698113e+00 -1.94175452e-01 4.46778595e-01 -4.56293643e-01
-3.05850327e-01 -7.24293828e-01 -1.14736915e+00 -7.41347909e-01
4.79711220e-02 -1.73240080e-02 4.15626645e-01 1.09379876e+00
2.68234432e-01 5.88684738e-01 -2.76946694e-01 -9.12856758e-01
-2.43762285e-01 -1.04876578e+00 -2.96809047e-01 6.31360829e-01
1.26022965e-01 -4.99105275e-01 -1.27355859e-01 3.74714807e-02] | [10.91717529296875, 10.415300369262695] |
df85f62b-32d1-42de-9f7c-f84db28072cc | unsupervised-3d-object-learning-through | 2302.11622 | null | https://arxiv.org/abs/2302.11622v1 | https://arxiv.org/pdf/2302.11622v1.pdf | Unsupervised 3D Object Learning through Neuron Activity aware Plasticity | We present an unsupervised deep learning model for 3D object classification. Conventional Hebbian learning, a well-known unsupervised model, suffers from loss of local features leading to reduced performance for tasks with complex geometric objects. We present a deep network with a novel Neuron Activity Aware (NeAW) Hebbian learning rule that dynamically switches the neurons to be governed by Hebbian learning or anti-Hebbian learning, depending on its activity. We analytically show that NeAW Hebbian learning relieves the bias in neuron activity, allowing more neurons to attend to the representation of the 3D objects. Empirical results show that the NeAW Hebbian learning outperforms other variants of Hebbian learning and shows higher accuracy over fully supervised models when training data is limited. | ['Saibal Mukhopadhyay', 'Biswadeep Chakraborty', 'Beomseok Kang'] | 2023-02-22 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [-1.11078009e-01 9.37733874e-02 -2.66566038e-01 -3.75846177e-01
5.59849560e-01 -4.98393416e-01 8.34835947e-01 1.21556394e-01
-5.73274374e-01 6.27960980e-01 1.40909076e-01 -2.02008843e-01
-5.42090416e-01 -8.38483274e-01 -8.64931464e-01 -1.17630649e+00
-2.21740618e-01 8.13848376e-01 6.92508280e-01 -3.65010314e-02
3.75816047e-01 1.16868401e+00 -1.25062895e+00 2.95431435e-01
2.94984490e-01 9.53364074e-01 1.80963576e-01 7.14765012e-01
-8.58576000e-02 1.16454804e+00 -7.47920692e-01 3.02504659e-01
3.76896679e-01 -1.71232089e-01 -8.86743784e-01 -3.13639641e-01
4.76039588e-01 -1.59316406e-01 -2.95318127e-01 8.30702782e-01
2.90623009e-01 3.55720162e-01 1.10352385e+00 -1.19133615e+00
-5.83923459e-01 5.67151010e-01 -1.95541263e-01 8.26524794e-01
-3.93184453e-01 1.62476614e-01 8.97152364e-01 -9.02472079e-01
3.73726904e-01 1.28403556e+00 6.23800457e-01 7.20511377e-01
-1.52853203e+00 -7.22495556e-01 3.89004350e-01 4.03223872e-01
-1.31701660e+00 -8.87910500e-02 8.30373347e-01 -6.08576477e-01
1.34099054e+00 -8.70066732e-02 1.30096483e+00 8.86839449e-01
3.35213333e-01 9.53643024e-01 1.08346713e+00 -5.54389834e-01
5.10659277e-01 -2.61291116e-01 5.01218617e-01 7.14633167e-01
6.23614788e-01 3.51629794e-01 -5.70257723e-01 2.53838450e-01
1.22311676e+00 1.65809050e-01 1.04018405e-01 -1.00718105e+00
-9.56679881e-01 6.44958615e-01 9.09777641e-01 5.59969664e-01
-5.70413411e-01 5.22418559e-01 -3.98700796e-02 3.75200152e-01
2.92083118e-02 6.52098596e-01 -5.18404782e-01 2.09242985e-01
-5.87500274e-01 3.71418774e-01 4.48837519e-01 5.27933717e-01
8.97206604e-01 3.71201247e-01 -2.85928100e-01 7.77863145e-01
3.95117968e-01 2.81886429e-01 5.10623276e-01 -9.04963970e-01
-2.83589303e-01 9.67795730e-01 -1.28900751e-01 -7.92553961e-01
-6.22639596e-01 -6.18577778e-01 -8.58493924e-01 8.48040342e-01
2.97927022e-01 3.24597687e-01 -1.18448770e+00 1.49636614e+00
-1.56161144e-01 -6.44299686e-02 1.27870724e-01 8.87481093e-01
9.15525317e-01 6.86262012e-01 8.91252384e-02 1.29327804e-01
6.58101380e-01 -9.36379433e-01 -5.29337943e-01 -2.05217406e-01
4.78911579e-01 8.65022242e-02 9.37624812e-01 4.93161708e-01
-9.72276270e-01 -6.14922285e-01 -1.30067039e+00 1.27171859e-01
-6.20302439e-01 -4.31722552e-01 7.17254817e-01 4.09625888e-01
-1.24779427e+00 4.93267953e-01 -9.90767837e-01 -5.52903593e-01
1.04013968e+00 9.05897081e-01 -1.52813464e-01 3.74150068e-01
-7.65180826e-01 1.20711613e+00 6.04696393e-01 -1.50119662e-01
-1.34973776e+00 -5.49084187e-01 -3.97598773e-01 2.19108552e-01
-6.45633265e-02 -6.30682886e-01 1.19146740e+00 -1.17369330e+00
-1.67308450e+00 1.09781837e+00 2.44195983e-01 -8.91876638e-01
2.12059729e-02 -1.73159137e-01 2.20669135e-02 1.36166930e-01
-5.03752172e-01 1.32018518e+00 7.36936331e-01 -1.60007024e+00
-4.39248443e-01 -4.39609855e-01 9.64784324e-02 2.76069194e-01
-3.04061145e-01 -6.09828293e-01 6.57285824e-02 -7.62542903e-01
2.51934171e-01 -8.35474372e-01 -1.25927284e-01 4.06249017e-02
-1.48827761e-01 -6.48868024e-01 9.52795386e-01 4.68644768e-01
7.57900298e-01 -1.94874215e+00 1.82592403e-02 4.97066438e-01
4.09183592e-01 3.82626504e-01 -1.58172831e-01 -1.05218053e-01
-1.97104022e-01 -1.62581578e-01 -7.61997253e-02 4.34914321e-01
-3.10073942e-01 8.49298537e-01 -3.03693891e-01 3.82146209e-01
2.17810404e-02 1.03127837e+00 -1.05805564e+00 -1.97903603e-01
3.41562748e-01 4.46476676e-02 -6.86753154e-01 8.11964795e-02
7.98433274e-03 4.05079424e-01 -1.29034996e-01 3.73965591e-01
3.53456587e-01 -1.83904454e-01 -1.23032175e-01 -1.74906254e-01
-2.31120829e-03 4.99259174e-01 -7.33141363e-01 1.30975699e+00
-2.15822369e-01 9.24744070e-01 -2.34099016e-01 -1.56589758e+00
1.20992303e+00 5.24680205e-02 4.35838610e-01 -5.31289935e-01
3.04209352e-01 4.45421524e-02 8.15571249e-01 -1.05620459e-01
-2.94414133e-01 -3.50541659e-02 3.51634443e-01 4.12212849e-01
6.74023271e-01 -2.01147914e-01 -4.99812700e-02 -6.14832491e-02
1.15582037e+00 -1.20469853e-01 2.40979508e-01 -8.74984920e-01
2.86960274e-01 -3.25169563e-02 3.52458596e-01 1.17508090e+00
-4.32625651e-01 1.16496161e-01 3.89283508e-01 -9.76201177e-01
-6.05831504e-01 -1.56339324e+00 -1.16875442e-02 1.23436534e+00
4.45085526e-01 2.06898078e-01 -6.39327586e-01 -7.02321172e-01
9.81647521e-02 9.14910734e-01 -1.03913999e+00 -7.10707307e-01
-6.74922109e-01 -8.06891799e-01 4.17623281e-01 6.19972706e-01
7.18249321e-01 -1.49254942e+00 -1.28521001e+00 4.76138145e-02
6.03143275e-01 -6.50659978e-01 2.09547412e-02 1.07918215e+00
-1.47065675e+00 -9.23556983e-01 -4.91446555e-01 -1.02864480e+00
1.00696445e+00 2.39968672e-02 9.74134386e-01 7.58733484e-04
-1.99374318e-01 4.84173417e-01 -1.04994848e-01 -8.13059509e-01
-2.97405481e-01 2.49724075e-01 1.10720888e-01 -5.71171194e-02
8.24249446e-01 -8.51376951e-01 -7.52622962e-01 2.49325871e-01
-6.69610322e-01 1.89607944e-02 6.95978880e-01 8.67716968e-01
5.97506821e-01 -8.12327042e-02 5.92149019e-01 -8.84027123e-01
3.19061041e-01 -6.48764744e-02 -4.89489377e-01 -2.18379095e-01
-9.39106941e-01 3.74121457e-01 2.35256687e-01 -6.32724524e-01
-8.89888465e-01 3.88401560e-02 3.41383368e-02 -5.55792809e-01
-3.64454359e-01 2.98993886e-02 3.90464999e-02 -2.65765905e-01
9.54423785e-01 3.27176154e-01 -2.05507815e-01 -4.06700045e-01
8.24250951e-02 7.47398138e-02 6.06660604e-01 -4.13517982e-01
5.22399604e-01 6.44481242e-01 1.00800633e-01 -5.65864027e-01
-9.62346673e-01 -2.38536045e-01 -1.06184363e+00 -4.37470555e-01
8.31178248e-01 -4.26629394e-01 -7.95347154e-01 6.29262805e-01
-1.11887598e+00 -1.08536029e+00 -6.81811631e-01 5.02230644e-01
-7.65938461e-01 -4.55964118e-01 -6.03676319e-01 -6.99906170e-01
-1.71198130e-01 -6.25984907e-01 2.96214193e-01 1.85451433e-01
-4.38422829e-01 -1.10627496e+00 2.94982523e-01 -1.78733781e-01
2.32583851e-01 -1.94352150e-01 1.39372277e+00 -9.21706736e-01
-3.60236526e-01 -7.73092061e-02 -3.20441090e-02 2.31752634e-01
1.31972760e-01 -2.27243587e-01 -8.90042543e-01 4.67001982e-02
-9.09289047e-02 -2.04224750e-01 1.22276390e+00 6.11585736e-01
1.40786552e+00 -6.01808190e-01 -3.22444439e-01 4.77906048e-01
1.08536613e+00 5.22837698e-01 5.80911517e-01 5.09274423e-01
4.70678478e-01 3.72924328e-01 9.78443678e-03 1.51498720e-01
1.83922425e-01 2.69979715e-01 9.87422645e-01 -1.56456113e-01
-2.22098053e-01 -3.10935169e-01 9.76051837e-02 6.99743032e-01
-4.80165780e-01 -1.11925483e-01 -1.05697381e+00 6.34137154e-01
-1.84880853e+00 -1.14361382e+00 1.38207227e-01 1.90082526e+00
8.69798124e-01 5.28229713e-01 8.64765272e-02 3.14351648e-01
4.29646134e-01 -5.52629195e-02 -7.68225372e-01 -6.35985851e-01
-4.05540079e-01 4.54094321e-01 7.20199704e-01 3.22223186e-01
-1.03696072e+00 1.13160133e+00 7.36055994e+00 5.47029436e-01
-1.32677972e+00 2.49924630e-01 3.60191494e-01 -2.50820249e-01
1.21858902e-01 -3.09361488e-01 -8.57215405e-01 1.15517750e-01
7.35913813e-01 3.92016590e-01 3.07252049e-01 6.46551669e-01
1.64262995e-01 -3.00311029e-01 -1.32362127e+00 1.26360905e+00
-9.75177251e-03 -1.65288818e+00 5.27891874e-01 3.01888473e-02
1.01928520e+00 3.29135001e-01 2.68433243e-02 3.10931832e-01
1.08447015e+00 -1.13412988e+00 8.47004831e-01 4.41440880e-01
-1.86599791e-02 -5.15605509e-01 5.09830654e-01 5.52076459e-01
-4.95253056e-01 -4.41244811e-01 -4.93993223e-01 -2.52128363e-01
-3.41339767e-01 3.22926879e-01 -5.69745600e-01 -6.12790227e-01
1.01288581e+00 7.31281519e-01 -5.49485743e-01 1.09865761e+00
-4.29654062e-01 8.38542044e-01 -2.19510347e-01 -4.27069247e-01
6.32853329e-01 8.37539360e-02 5.41621268e-01 1.24511480e+00
-1.12301880e-03 3.11757654e-01 3.12074414e-03 9.19632733e-01
-1.75480932e-01 -1.88003838e-01 -6.12354934e-01 3.69914770e-01
3.71735454e-01 8.94176066e-01 -9.71581280e-01 -4.22972560e-01
5.42586893e-02 7.43414462e-01 4.86919194e-01 5.89574158e-01
-2.65711129e-01 -1.80454910e-01 3.48062158e-01 3.59696358e-01
4.52920079e-01 -3.55245411e-01 -6.89285398e-01 -4.69933718e-01
-8.33023846e-01 -2.32215837e-01 4.82814699e-01 -9.09603119e-01
-1.05699301e+00 4.94541079e-01 3.29413675e-02 -8.71788323e-01
1.89789068e-02 -1.13059556e+00 -6.82539642e-01 2.38092422e-01
-1.22508657e+00 -8.74871254e-01 -3.42516035e-01 8.43111873e-01
2.88839132e-01 -4.69213545e-01 7.49896228e-01 -7.59229213e-02
-2.77458042e-01 3.50984573e-01 -6.07430525e-02 2.99722761e-01
4.12607193e-01 -1.39298201e+00 -1.89864188e-01 2.51444459e-01
5.50142646e-01 4.76505071e-01 4.98446286e-01 -1.81168586e-01
-8.48132670e-01 -1.13582659e+00 4.87822205e-01 -5.31187236e-01
2.50839978e-01 -3.55445534e-01 -9.82453287e-01 7.70370960e-01
1.58561006e-01 3.06549281e-01 7.59504199e-01 4.41443622e-02
-4.34321463e-01 -6.88418150e-01 -9.81449783e-01 6.27105236e-01
1.13724482e+00 -1.66454211e-01 -1.02560675e+00 1.70691293e-02
9.60620344e-02 1.48815125e-01 -4.63020772e-01 6.81585670e-01
5.70078909e-01 -1.03054142e+00 9.15869713e-01 -7.49344826e-01
-6.20981492e-02 -3.81107092e-01 -1.11681819e-01 -1.53582299e+00
-7.85984814e-01 -1.13879509e-01 -6.20683789e-01 4.40850884e-01
3.35278571e-01 -5.31768382e-01 9.86122131e-01 -1.27537414e-01
-2.93545693e-01 -7.88228571e-01 -1.02249527e+00 -1.02265239e+00
3.32513630e-01 -1.67333364e-01 1.88147381e-01 7.93468475e-01
-9.19283256e-02 5.22201955e-01 2.72955418e-01 -1.52488634e-01
4.42285001e-01 2.31279526e-02 4.99421239e-01 -1.58554411e+00
3.10512390e-02 -1.19594646e+00 -7.21527517e-01 -1.30212033e+00
6.66974112e-02 -1.13543689e+00 1.94964021e-01 -1.61365855e+00
2.53286660e-01 -3.52404803e-01 -8.95346820e-01 8.80855560e-01
4.95927721e-01 6.01675391e-01 -8.86681825e-02 4.06672686e-01
-4.98738021e-01 2.76829988e-01 1.27241504e+00 -5.01279354e-01
-5.11729717e-01 7.81668629e-03 -2.96480417e-01 1.02878881e+00
7.59998441e-01 -4.55679536e-01 -4.25227672e-01 -5.94903409e-01
8.22512284e-02 -8.65920246e-01 7.16144383e-01 -1.19674218e+00
4.02732253e-01 -7.36709163e-02 9.78256643e-01 -7.64269114e-01
1.05457783e-01 -7.92214453e-01 -3.90349597e-01 1.02608514e+00
-7.86089897e-01 -2.46610254e-01 3.05123031e-01 3.23607653e-01
2.92487646e-04 -3.26798379e-01 1.34808719e+00 -2.70553589e-01
-8.71136606e-01 3.46795559e-01 -9.90066707e-01 -2.38010719e-01
9.73104417e-01 -7.33946562e-01 -1.21568874e-01 -2.12130159e-01
-9.76514101e-01 -3.55056934e-02 8.72482881e-02 4.08112347e-01
6.48265183e-01 -1.44948936e+00 -3.97925109e-01 3.41604352e-01
2.51722820e-02 1.50693595e-01 -1.03810191e-01 7.40118861e-01
-7.21350312e-01 6.17863655e-01 -7.58869290e-01 -9.55624044e-01
-8.18773091e-01 6.31954908e-01 6.83974981e-01 -1.96977511e-01
-4.59075928e-01 9.88446534e-01 5.64513743e-01 -4.15938437e-01
3.77742916e-01 -3.23061973e-01 -5.01674056e-01 -2.61267573e-02
8.62686709e-02 4.35741216e-01 1.73146859e-01 -4.83905882e-01
-3.98215353e-01 3.90802383e-01 -7.77662843e-02 1.43317729e-01
1.37385356e+00 1.16040893e-01 2.79385187e-02 8.22442174e-01
8.10104907e-01 -6.02716744e-01 -1.42543888e+00 -2.57688373e-01
2.09629163e-01 1.46559253e-01 2.73848712e-01 -9.81154203e-01
-1.02092874e+00 1.32733440e+00 9.28906977e-01 -1.04964614e-01
9.47966754e-01 2.96916068e-01 1.16765633e-01 1.03529322e+00
2.53611982e-01 -1.16610420e+00 5.42771518e-01 9.52310026e-01
8.91722739e-01 -8.95459175e-01 -1.81178480e-01 6.24863617e-02
-1.62217826e-01 1.13807034e+00 1.20096743e+00 -8.37032020e-01
1.13963866e+00 2.35647857e-01 1.60243198e-01 -3.27608317e-01
-8.25681448e-01 -1.36349469e-01 3.18961591e-01 8.46240103e-01
2.46693287e-02 -1.97313428e-01 2.86887139e-01 6.53894842e-01
-1.56373337e-01 -1.16094537e-01 2.62186825e-01 8.12378585e-01
-7.74117708e-01 -4.44143146e-01 -1.80800408e-01 6.08553350e-01
-2.74237953e-02 1.10323370e-01 -5.43446243e-01 7.34107316e-01
4.70789015e-01 4.12975013e-01 7.79262900e-01 1.64692700e-02
1.93055600e-01 1.38775986e-02 1.14297116e+00 -7.86994636e-01
-5.97381711e-01 6.92116395e-02 -7.00083494e-01 -1.57104447e-01
-7.14616537e-01 -2.33460650e-01 -1.70553279e+00 -7.79382363e-02
-8.54004398e-02 -4.00718441e-03 1.43780470e-01 9.98148978e-01
2.63793766e-01 6.13535285e-01 3.73424292e-01 -1.00326478e+00
-1.85576096e-01 -9.67551410e-01 -6.16502225e-01 4.24172610e-01
3.76763374e-01 -1.08789301e+00 -8.79162177e-02 2.59046078e-01] | [9.703516960144043, 2.35276198387146] |
7dd4f7b8-7995-48d9-89f6-a68233904528 | cross-modality-multi-atlas-segmentation-using-1 | 2202.02000 | null | https://arxiv.org/abs/2202.02000v3 | https://arxiv.org/pdf/2202.02000v3.pdf | Cross-Modality Multi-Atlas Segmentation via Deep Registration and Label Fusion | Multi-atlas segmentation (MAS) is a promising framework for medical image segmentation. Generally, MAS methods register multiple atlases, i.e., medical images with corresponding labels, to a target image; and the transformed atlas labels can be combined to generate target segmentation via label fusion schemes. Many conventional MAS methods employed the atlases from the same modality as the target image. However, the number of atlases with the same modality may be limited or even missing in many clinical applications. Besides, conventional MAS methods suffer from the computational burden of registration or label fusion procedures. In this work, we design a novel cross-modality MAS framework, which uses available atlases from a certain modality to segment a target image from another modality. To boost the computational efficiency of the framework, both the image registration and label fusion are achieved by well-designed deep neural networks. For the atlas-to-target image registration, we propose a bi-directional registration network (BiRegNet), which can efficiently align images from different modalities. For the label fusion, we design a similarity estimation network (SimNet), which estimates the fusion weight of each atlas by measuring its similarity to the target image. SimNet can learn multi-scale information for similarity estimation to improve the performance of label fusion. The proposed framework was evaluated by the left ventricle and liver segmentation tasks on the MM-WHS and CHAOS datasets, respectively. Results have shown that the framework is effective for cross-modality MAS in both registration and label fusion. | ['Liqin Huang', 'Xiahai Zhuang', 'Lei LI', 'Wangbin Ding'] | 2022-02-04 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 1.03262044e-01 -1.20662510e-01 -8.69152546e-02 -5.69616616e-01
-9.91837025e-01 -4.37855512e-01 3.44896257e-01 1.43832088e-01
-5.51523924e-01 3.27222854e-01 2.42938638e-01 2.11329833e-01
8.67726952e-02 -7.67748535e-01 -3.14600438e-01 -1.06216729e+00
2.88967639e-01 4.68381882e-01 4.11519140e-01 -1.57531053e-01
-2.09352091e-01 3.99826348e-01 -9.51278687e-01 1.39190704e-01
9.70741749e-01 1.03417277e+00 3.44089210e-01 4.40671444e-02
-2.40815237e-01 3.72443914e-01 -4.15455699e-01 -7.11585879e-02
2.63270110e-01 -6.07988477e-01 -7.66824961e-01 -4.08753604e-02
2.45913699e-01 -1.32139057e-01 -3.70390803e-01 1.39676654e+00
7.27991998e-01 -1.38090281e-02 4.93948787e-01 -1.14599085e+00
-2.59582251e-01 7.08735287e-01 -6.43002808e-01 2.90827453e-01
-2.16427427e-02 7.46253459e-03 4.32074100e-01 -5.73584557e-01
3.90874267e-01 9.57714558e-01 6.79721057e-01 3.44984859e-01
-8.22343946e-01 -8.85376692e-01 -2.11739391e-01 -1.03188172e-01
-1.55844641e+00 -2.42866158e-01 8.25616956e-01 -4.78184253e-01
1.53610826e-01 2.82805741e-01 7.24203467e-01 3.33376944e-01
4.88267541e-01 6.23888791e-01 1.43451369e+00 -2.59775259e-02
-1.03480488e-01 -2.39258185e-01 1.91148426e-02 8.74490261e-01
-6.29461780e-02 1.44879669e-01 3.46488692e-03 -1.24977969e-01
7.62052059e-01 2.57941574e-01 -4.55351770e-01 -2.11998343e-01
-1.86613369e+00 7.29562461e-01 9.82673824e-01 9.15262997e-01
-5.21838665e-01 -1.85405776e-01 5.71193278e-01 -3.69262286e-02
2.80109346e-01 1.80805340e-01 -8.49509686e-02 5.13109982e-01
-1.14271891e+00 -9.15849581e-02 3.32258224e-01 6.66679084e-01
6.14000976e-01 -6.93995357e-02 -4.62376684e-01 8.00773382e-01
6.15444601e-01 5.65136671e-01 1.08118415e+00 -6.78026497e-01
1.66231230e-01 7.45778322e-01 -3.52134585e-01 -1.02373624e+00
-7.88866520e-01 -4.66154814e-01 -1.38033271e+00 2.62668543e-02
3.49116147e-01 5.99694159e-03 -1.15228391e+00 1.98289776e+00
6.98272407e-01 5.07405519e-01 1.35030538e-01 1.18970609e+00
1.37394559e+00 5.25876045e-01 2.60235935e-01 -5.46159625e-01
1.58232534e+00 -1.02529895e+00 -8.83809745e-01 1.27697647e-01
6.88389957e-01 -9.69252527e-01 7.15000331e-01 -1.46263763e-01
-1.14226913e+00 -5.78166187e-01 -9.97720420e-01 2.15505719e-01
-1.54505000e-01 6.27114400e-02 3.24137300e-01 3.17315489e-01
-8.68063450e-01 3.14509422e-01 -1.01580250e+00 -1.42289340e-01
4.22923177e-01 6.66296899e-01 -4.91017699e-01 -7.83423632e-02
-1.28206468e+00 1.03818476e+00 6.97996557e-01 2.19794601e-01
-7.05608904e-01 -6.20952606e-01 -9.68263030e-01 -7.30973929e-02
5.85404187e-02 -6.46151245e-01 1.07417452e+00 -9.32549536e-01
-1.18567288e+00 1.09929752e+00 2.48716716e-02 -9.62199196e-02
2.42431208e-01 4.09285814e-01 -7.30943859e-01 1.73627600e-01
3.96265417e-01 8.58900368e-01 4.66156572e-01 -1.11634791e+00
-5.31011105e-01 -3.88340145e-01 -3.13416064e-01 4.39929157e-01
5.34394495e-02 1.65798903e-01 -4.33775961e-01 -7.90253401e-01
6.43810153e-01 -9.29782271e-01 -3.06965619e-01 -3.78096372e-01
-3.39880228e-01 -4.51051705e-02 7.35848606e-01 -8.59121025e-01
1.07083893e+00 -2.09410286e+00 1.89562246e-01 3.71067584e-01
4.77628201e-01 3.47319156e-01 -1.08552635e-01 -3.77601862e-01
-1.83599547e-01 -2.26504833e-01 -5.25883734e-01 -1.03613295e-01
-3.38571906e-01 2.99114883e-01 2.42351890e-01 6.52125180e-01
-4.03927833e-01 1.05318534e+00 -8.24390292e-01 -1.04841006e+00
4.46029961e-01 4.61966306e-01 -1.28053963e-01 3.98502111e-01
3.24973196e-01 1.18593931e+00 -4.46776628e-01 6.50732160e-01
7.83562481e-01 -2.43243262e-01 -7.64836296e-02 -8.77256095e-01
5.87779358e-02 -3.29579979e-01 -1.08228672e+00 2.04471946e+00
-4.36666578e-01 -2.61603184e-02 1.95752174e-01 -9.47515965e-01
9.23330605e-01 6.84367061e-01 1.15110648e+00 -7.83782661e-01
6.02342010e-01 4.88774449e-01 1.34954706e-01 -4.10543829e-01
-2.59586703e-02 -4.15120721e-01 -1.14706464e-01 4.25863862e-01
1.04483418e-01 -1.55744880e-01 7.06772655e-02 -1.31917551e-01
7.17075169e-01 -2.88278937e-01 3.34062517e-01 -3.32866907e-01
8.98189545e-01 -1.13264538e-01 8.94037366e-01 3.41407061e-01
-3.82434249e-01 6.03630245e-01 -1.14207469e-01 -5.74058235e-01
-8.85811329e-01 -1.16775727e+00 -3.92231941e-01 5.96850216e-01
5.40683210e-01 -1.38373360e-01 -9.84450102e-01 -8.09070766e-01
-3.81953478e-01 2.68159229e-02 -3.27104717e-01 -2.43744597e-01
-6.83951318e-01 -1.03155077e+00 6.27898932e-01 2.93463111e-01
8.68413746e-01 -1.10843801e+00 -3.28153491e-01 4.27024633e-01
-4.43484455e-01 -1.04144847e+00 -1.10322583e+00 -2.90995508e-01
-9.40162241e-01 -1.05254924e+00 -1.06741619e+00 -1.07737684e+00
8.69911969e-01 2.43494138e-02 8.74532819e-01 2.49929354e-01
3.35625634e-02 1.49067293e-03 4.50949912e-04 4.42480966e-02
-6.82693541e-01 5.01302779e-02 1.12062529e-01 1.54464841e-01
7.60581493e-02 -5.30904949e-01 -6.52434707e-01 5.98088503e-01
-1.23882580e+00 3.99816155e-01 5.97731113e-01 9.65533435e-01
8.43323886e-01 -1.63138583e-01 6.13601863e-01 -6.40062749e-01
5.22807896e-01 -3.99575114e-01 -6.89763546e-01 5.33705771e-01
-4.89351124e-01 -2.24543974e-01 3.16814125e-01 -4.02584016e-01
-8.53397965e-01 2.99079597e-01 -4.42768604e-01 -3.73656034e-01
-1.20123841e-01 6.77115262e-01 -1.82126716e-01 -5.13689816e-01
3.18419218e-01 2.33064696e-01 2.47006923e-01 -1.99634477e-01
3.09678793e-01 5.77562273e-01 8.59842479e-01 -3.54927778e-01
5.04432499e-01 2.64899492e-01 1.62527233e-01 -6.56823590e-02
-7.20283985e-01 -3.41221869e-01 -7.41590023e-01 -2.16186851e-01
1.29815626e+00 -9.50079560e-01 -6.86593354e-01 7.12268114e-01
-1.11306190e+00 6.82681277e-02 -1.03469558e-01 9.25485492e-01
-4.18697357e-01 3.43516499e-01 -8.09063971e-01 1.99259803e-01
-7.16159821e-01 -1.84722638e+00 9.89970267e-01 7.21148133e-01
1.23039670e-01 -1.19694746e+00 1.37780502e-01 2.89226592e-01
5.03109694e-01 4.42130238e-01 6.34957850e-01 -8.31975639e-01
-3.94095778e-01 -1.99723914e-01 -4.24590319e-01 2.32931316e-01
5.36449373e-01 -5.51004946e-01 -6.71986282e-01 -4.11636353e-01
2.71558911e-01 -9.89213400e-03 5.05966067e-01 7.43778944e-01
9.32056248e-01 -8.39618072e-02 -4.32126790e-01 8.77732337e-01
1.24798059e+00 4.09032255e-01 3.14087272e-01 4.55192029e-02
1.07505238e+00 2.21958712e-01 4.95453179e-01 1.46676870e-02
5.54915786e-01 8.12456071e-01 1.17739111e-01 -7.49440372e-01
-2.13315994e-01 1.58054397e-01 -1.20503388e-01 1.53763390e+00
1.55444846e-01 1.55280620e-01 -1.05966187e+00 3.24742645e-01
-1.72487378e+00 -5.63695610e-01 -5.10381348e-02 2.06491804e+00
9.85555649e-01 -2.78401285e-01 -5.82234114e-02 -4.02834147e-01
1.10600305e+00 -1.32971955e-02 -4.49198782e-01 2.14765579e-01
-1.05087630e-01 1.42996192e-01 4.42419291e-01 5.69762766e-01
-1.30025709e+00 6.90187037e-01 6.00051975e+00 9.48753774e-01
-1.42199850e+00 6.94004059e-01 7.81411350e-01 2.77235955e-01
-1.10800065e-01 -1.10689692e-01 -4.13692862e-01 7.16370821e-01
6.00516558e-01 -2.80172408e-01 2.50392407e-01 4.14062291e-01
2.82385349e-02 -5.88927530e-02 -8.02457452e-01 1.12415862e+00
4.78817336e-02 -1.40231407e+00 -1.75987676e-01 -1.70073390e-01
8.25421870e-01 2.15958714e-01 -1.60884023e-01 1.21723175e-01
1.65590897e-01 -9.33849216e-01 2.07951471e-01 6.32556140e-01
8.78388047e-01 -5.59280694e-01 1.16814542e+00 1.86404824e-01
-1.49305308e+00 4.40958112e-01 -7.99483135e-02 6.14639699e-01
3.81980032e-01 5.02907336e-01 -4.98786837e-01 9.48208630e-01
4.34307456e-01 6.76563442e-01 -6.24314845e-01 1.16469336e+00
1.24587938e-01 1.94398120e-01 -4.58425462e-01 6.30682886e-01
2.94997126e-01 -4.84020442e-01 3.01654518e-01 8.75893056e-01
5.90567887e-01 2.01780096e-01 8.23220313e-01 7.80950189e-01
-1.73691526e-01 3.69695812e-01 -3.98791403e-01 3.49867433e-01
3.88518900e-01 1.64872921e+00 -1.02668989e+00 -5.68804145e-01
-4.08539712e-01 7.49140799e-01 -1.91254064e-01 1.60397008e-01
-8.74439836e-01 -2.99318116e-02 -1.85652226e-02 -2.87531197e-01
-4.13446367e-01 1.42692044e-01 -2.82414794e-01 -1.25380290e+00
-2.10465536e-01 -7.62965798e-01 6.38445914e-01 -6.82719946e-01
-1.25111103e+00 9.51115072e-01 3.49866860e-02 -1.42949212e+00
8.46611485e-02 -1.23908045e-02 -7.34029591e-01 1.15643919e+00
-1.25507534e+00 -1.45038950e+00 -5.92989266e-01 8.13552976e-01
5.75272832e-03 -2.60838240e-01 7.82867134e-01 8.00853848e-01
-3.07074636e-01 5.25629222e-01 6.19536825e-02 3.76186579e-01
8.24640751e-01 -1.00241292e+00 -5.89724965e-02 6.42211020e-01
-6.05776571e-02 4.07512486e-01 2.11099625e-01 -7.45663702e-01
-8.00665021e-01 -1.03850913e+00 4.63930637e-01 5.77269867e-02
3.25815618e-01 1.99179500e-01 -1.03734457e+00 6.44599319e-01
2.70108372e-01 6.08557403e-01 6.83449209e-01 -4.56288099e-01
2.34219447e-01 -2.34118536e-01 -1.50401402e+00 4.00035113e-01
4.77457255e-01 -3.57326865e-01 -5.90721548e-01 4.90667254e-01
7.79208302e-01 -9.98192132e-01 -1.59699166e+00 8.07651997e-01
4.29018676e-01 -6.31634057e-01 9.30059314e-01 -7.09706247e-02
1.52929705e-02 -7.47007906e-01 -4.88714725e-02 -1.49162138e+00
-2.28258282e-01 -2.47546267e-02 3.67524862e-01 1.30900013e+00
-5.25013693e-02 -6.20101035e-01 4.80866373e-01 4.43146914e-01
-4.37910795e-01 -7.61007845e-01 -1.25851393e+00 -4.47621644e-01
1.60252735e-01 2.92460918e-02 1.06156242e+00 1.25363564e+00
-3.38388145e-01 1.46625519e-01 -1.81776494e-01 2.84280211e-01
5.89310050e-01 7.01543763e-02 2.80506790e-01 -1.15177071e+00
6.88003823e-02 -6.29557371e-01 -4.07134444e-01 -5.20171344e-01
2.39623472e-01 -1.45183659e+00 5.80090731e-02 -1.53274572e+00
1.98109925e-01 -6.72142148e-01 -7.40789592e-01 5.01147807e-01
-2.56690115e-01 5.75095356e-01 1.67423353e-01 4.36645418e-01
-4.64614570e-01 5.41937590e-01 1.75251245e+00 -3.28670055e-01
-1.29476190e-01 -2.72954889e-02 -2.68638313e-01 7.61334896e-01
7.31938839e-01 -5.71722329e-01 -1.68271646e-01 -2.70887464e-01
-4.03769195e-01 5.46181321e-01 1.77832529e-01 -1.17631507e+00
3.94187540e-01 8.25826228e-02 3.85782450e-01 -4.71953332e-01
3.70689780e-02 -1.02379704e+00 4.76777792e-01 6.51773274e-01
-2.40507722e-01 2.01715827e-01 5.32580912e-02 5.25261052e-02
-6.03895903e-01 -1.37983978e-01 1.14855433e+00 -1.84241623e-01
-4.61358637e-01 8.34299803e-01 3.70413177e-02 -1.13844723e-01
1.12113225e+00 1.34271577e-01 -1.31846666e-01 6.10321872e-02
-1.08077657e+00 3.37765038e-01 3.09237182e-01 2.48971894e-01
5.67897201e-01 -1.96458209e+00 -6.25561893e-01 2.15322196e-01
9.64818373e-02 2.76781440e-01 4.96082693e-01 1.52744401e+00
-4.74769592e-01 1.50261149e-02 -5.35668612e-01 -9.13608551e-01
-1.30705249e+00 5.77023983e-01 8.34430277e-01 -5.46496570e-01
-5.37101746e-01 4.47773874e-01 4.76101071e-01 -7.94508636e-01
-1.96061686e-01 -3.08629751e-01 -4.17188883e-01 -1.05445936e-01
4.11670238e-01 -9.54869092e-02 1.41448542e-01 -1.28431118e+00
-3.46199334e-01 9.76257622e-01 -5.63683249e-02 2.37707309e-02
8.78252149e-01 -1.31063133e-01 -5.49554646e-01 2.52470106e-01
1.31455171e+00 -2.44404122e-01 -8.23863804e-01 -7.63338149e-01
-1.78881362e-01 -2.40170732e-01 3.18841070e-01 -7.67514169e-01
-1.71876383e+00 6.62794352e-01 1.20819247e+00 -1.79697752e-01
1.38275468e+00 4.57204878e-02 1.15451849e+00 -1.73358813e-01
2.97031641e-01 -6.97887123e-01 -3.26121420e-01 1.37193322e-01
6.35990143e-01 -1.47650564e+00 -5.68883047e-02 -2.40022615e-01
-5.89642882e-01 1.04075575e+00 5.72676122e-01 7.13889077e-02
8.06211293e-01 2.02120975e-01 4.13431764e-01 -4.58604366e-01
1.69344798e-01 -1.12389149e-02 5.90875506e-01 2.56797701e-01
4.77147043e-01 2.60418385e-01 -5.27679265e-01 6.19644821e-01
8.57520103e-02 7.65668824e-02 5.24384715e-02 6.75396979e-01
-2.85755098e-01 -1.23740172e+00 -6.49658203e-01 4.90116119e-01
-5.04578054e-01 -5.66271096e-02 3.24587286e-01 5.92392683e-01
5.10935843e-01 6.79830849e-01 3.24747190e-02 -4.03344542e-01
1.61602676e-01 -1.25052005e-01 4.28306580e-01 -3.46221179e-01
-8.64382684e-01 3.73056293e-01 -5.87923229e-01 -4.63612109e-01
-7.98795879e-01 -4.88757998e-01 -1.54735255e+00 -2.30226547e-01
-2.21156269e-01 2.16280833e-01 4.99432117e-01 1.22386086e+00
2.88346870e-04 7.15932250e-01 7.44264722e-01 -7.37386048e-01
-5.07896207e-02 -9.79127765e-01 -5.96185505e-01 6.89163387e-01
1.58841893e-01 -7.36725211e-01 1.60307169e-01 -2.29812115e-02] | [14.306962966918945, -2.51255202293396] |
3b15d10c-774b-4b95-b79b-618d24a6c8b5 | cross-lingual-lexical-sememe-prediction | null | null | https://aclanthology.org/D18-1033 | https://aclanthology.org/D18-1033.pdf | Cross-lingual Lexical Sememe Prediction | Sememes are defined as the minimum semantic units of human languages. As important knowledge sources, sememe-based linguistic knowledge bases have been widely used in many NLP tasks. However, most languages still do not have sememe-based linguistic knowledge bases. Thus we present a task of cross-lingual lexical sememe prediction, aiming to automatically predict sememes for words in other languages. We propose a novel framework to model correlations between sememes and multi-lingual words in low-dimensional semantic space for sememe prediction. Experimental results on real-world datasets show that our proposed model achieves consistent and significant improvements as compared to baseline methods in cross-lingual sememe prediction. The codes and data of this paper are available at \url{https://github.com/thunlp/CL-SP}. | ['Ruobing Xie', 'Yankai Lin', 'Fanchao Qi', 'Maosong Sun', 'Hao Zhu', 'Zhiyuan Liu'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['multilingual-word-embeddings', 'learning-word-embeddings'] | ['methodology', 'methodology'] | [-3.78050655e-01 -2.73048580e-01 -6.52619481e-01 -3.86173636e-01
-6.57738149e-01 -5.56538105e-01 4.84367073e-01 2.45631188e-01
-5.85137188e-01 8.52396488e-01 3.98038447e-01 -4.44612414e-01
2.90966369e-02 -8.87589455e-01 -5.60603917e-01 -1.76168337e-01
3.21003109e-01 6.04338825e-01 4.62745540e-02 -3.58407617e-01
3.44955981e-01 -5.05690202e-02 -1.53875518e+00 4.22706485e-01
1.26391828e+00 6.85672045e-01 5.97220778e-01 -1.04106247e-01
-7.72060752e-01 3.65565509e-01 -6.97458535e-02 -5.73353171e-01
1.27710149e-01 -6.14568949e-01 -7.90605307e-01 -2.17578575e-01
2.04416350e-01 6.04069173e-01 1.49885625e-01 1.50617206e+00
1.08665355e-01 2.80892670e-01 8.37342978e-01 -6.58860147e-01
-1.09265220e+00 1.05474079e+00 -3.48692417e-01 8.11633840e-02
3.29109520e-01 -6.32589340e-01 1.41588497e+00 -1.22755170e+00
6.63924575e-01 1.46068656e+00 3.38867426e-01 5.23928702e-01
-9.20857012e-01 -7.56330729e-01 3.76579538e-02 7.65563250e-01
-1.62067652e+00 -4.72135931e-01 8.40315819e-01 -1.49153769e-01
1.26968503e+00 1.49375051e-01 4.28725570e-01 1.01601946e+00
2.14920312e-01 6.69318199e-01 1.20239174e+00 -7.85971582e-01
3.21593769e-02 4.99075204e-01 3.06581289e-01 7.46506333e-01
5.48926950e-01 -2.15739474e-01 -7.20227242e-01 2.89849669e-01
4.32025135e-01 -3.73779148e-01 1.37777865e-01 -2.24170480e-02
-1.11773121e+00 9.58271384e-01 2.61269696e-02 7.94605374e-01
-2.97421664e-01 -5.29322922e-01 5.33231139e-01 2.31212065e-01
5.28390288e-01 2.27843523e-01 -7.06208229e-01 -1.21897735e-01
-4.13701832e-01 -2.26087511e-01 7.87284732e-01 1.07930827e+00
1.00362694e+00 -7.22969323e-02 6.01854086e-01 1.37222874e+00
3.13108832e-01 6.49872959e-01 1.04509914e+00 -3.13709497e-01
4.69113350e-01 8.43549967e-01 -2.04850897e-01 -1.01500595e+00
-3.79167110e-01 9.16819349e-02 -5.23846805e-01 -5.18569946e-01
6.48333803e-02 -7.21752793e-02 -5.27401507e-01 1.52054918e+00
3.10747236e-01 1.60144255e-01 3.47724915e-01 6.84918582e-01
8.82417381e-01 9.58041728e-01 4.01218086e-01 -3.71580869e-01
1.45876753e+00 -8.64064813e-01 -8.79667938e-01 -4.14754242e-01
1.01355422e+00 -9.15234685e-01 1.47522843e+00 2.67422825e-01
-8.91907990e-01 -5.53861499e-01 -7.42774725e-01 -4.15261596e-01
-1.04457045e+00 2.45915949e-01 6.52273417e-01 7.07550466e-01
-7.56897390e-01 1.78943366e-01 -8.18273067e-01 -5.61515033e-01
-1.19452015e-01 4.51989360e-02 -3.42965186e-01 1.77021489e-01
-1.54191935e+00 1.18590975e+00 1.26367474e+00 -1.75675347e-01
-1.15525030e-01 -3.14361274e-01 -1.09470296e+00 -1.94704309e-01
4.51156616e-01 -1.90966487e-01 8.06040108e-01 -8.87791336e-01
-1.21966958e+00 1.18879402e+00 -6.43708229e-01 -3.04986387e-01
-1.99194685e-01 -1.98349267e-01 -1.04656589e+00 -4.42257434e-01
2.66116768e-01 3.89147848e-01 2.26410136e-01 -1.11618114e+00
-9.77246940e-01 -4.09599960e-01 -4.12979692e-01 4.81038988e-01
-8.27667058e-01 2.26360172e-01 -6.00854695e-01 -5.67241549e-01
-1.67788155e-02 -7.30023205e-01 1.13645390e-01 -9.08385515e-01
-4.06184852e-01 -7.55373061e-01 3.74859810e-01 -7.71903336e-01
1.60995853e+00 -1.82321608e+00 1.11279033e-01 2.03837410e-01
-3.53885531e-01 2.52763748e-01 -1.58640034e-02 3.56856108e-01
9.73939970e-02 2.06824660e-01 -6.31116033e-02 -1.55386001e-01
2.91017890e-01 3.78387243e-01 -3.34705636e-02 9.05001238e-02
-2.83865005e-01 1.00186145e+00 -9.59450066e-01 -7.04768240e-01
5.76926410e-01 4.96117383e-01 -2.49018088e-01 -2.69838750e-01
-3.18959951e-01 1.07647575e-01 -5.02889991e-01 5.98464549e-01
3.66977453e-01 -1.33912638e-01 7.48886168e-01 -2.22196788e-01
-3.40065926e-01 8.62334311e-01 -1.12888300e+00 1.84790039e+00
-7.69295037e-01 1.88196555e-01 -5.35150945e-01 -9.78697121e-01
9.56062257e-01 1.25144869e-01 2.12833464e-01 -9.33256805e-01
3.70684564e-01 6.55193567e-01 8.11843351e-02 -2.13465959e-01
6.18903518e-01 -4.49981779e-01 -3.26084554e-01 7.74199143e-02
1.00277886e-01 4.30691764e-02 7.23256528e-01 -2.50962555e-01
3.45328569e-01 8.48481953e-02 9.67890680e-01 -8.09203386e-01
9.72342789e-01 1.86878115e-01 7.20991910e-01 3.30877453e-02
1.29347324e-01 -2.05645606e-01 -1.63184807e-01 -1.74401492e-01
-7.25542307e-01 -1.25008333e+00 -5.60220420e-01 1.29761541e+00
2.93713182e-01 -9.15626407e-01 -7.29788423e-01 -5.24156868e-01
-1.17900819e-01 1.30439460e+00 -4.08955455e-01 1.23552330e-01
-4.86320525e-01 -6.52027607e-01 2.92038977e-01 3.20877731e-01
2.21489578e-01 -1.41427350e+00 6.98024482e-02 1.83096677e-01
-3.54141533e-01 -1.20583618e+00 -5.37675142e-01 4.64803539e-02
-6.38937891e-01 -9.62006390e-01 6.48552552e-02 -1.25386131e+00
4.75178689e-01 5.86955994e-02 1.13260877e+00 -1.29639909e-01
-2.52897352e-01 -4.23850343e-02 -5.44806063e-01 -6.58059120e-01
-4.93023247e-01 1.52818412e-01 4.66275305e-01 -3.08431834e-01
1.48401046e+00 -3.84686470e-01 -1.68776095e-01 1.02422081e-01
-6.14585996e-01 1.28135711e-01 3.18635225e-01 4.71524417e-01
1.10173476e+00 1.79966062e-01 5.82954824e-01 -1.16689265e+00
7.05681622e-01 -6.68340862e-01 -7.92589068e-01 4.29428816e-01
-7.64593899e-01 1.81436971e-01 9.02021646e-01 -8.44251662e-02
-1.19653869e+00 -1.12004042e-01 -2.93421388e-01 1.62321866e-01
-4.22354192e-01 8.17319572e-01 -4.90318000e-01 3.92493993e-01
3.76460522e-01 5.62207162e-01 -5.02824187e-01 -7.93083429e-01
4.91789132e-01 7.23996639e-01 2.56049335e-01 -7.20685303e-01
4.07593668e-01 2.59494364e-01 -3.57405603e-01 -1.22086501e+00
-1.11731040e+00 -6.75512016e-01 -9.40100849e-01 1.55151516e-01
1.04820704e+00 -9.67593968e-01 -4.67778355e-01 -2.01162118e-02
-1.02575731e+00 1.13172255e-01 3.53209637e-02 7.11783051e-01
-4.45404083e-01 3.66267145e-01 -5.09113073e-01 -3.87527347e-01
-3.69963259e-01 -6.75403714e-01 7.59080410e-01 1.12456597e-01
-4.21145529e-01 -1.78220224e+00 3.32061201e-01 3.55849415e-01
-1.83768235e-02 -3.03052992e-01 1.20999336e+00 -1.00218213e+00
-2.22119242e-01 3.30643862e-01 5.09855486e-02 4.88996953e-01
7.09470809e-01 -3.66918534e-01 -5.01761973e-01 -7.28622079e-02
-1.31957769e-01 -3.25993776e-01 5.85175395e-01 2.57504702e-01
1.03790009e+00 -1.87472939e-01 -3.17038864e-01 5.64579725e-01
1.69876051e+00 3.29340339e-01 2.47218817e-01 4.76477474e-01
8.61827672e-01 5.33690095e-01 7.35825121e-01 3.35811824e-01
7.63031185e-01 5.20630479e-01 -3.00639063e-01 3.96829963e-01
-1.09090731e-01 -4.93893445e-01 6.35243535e-01 1.94591391e+00
-2.99389325e-02 -3.64543140e-01 -1.17331648e+00 8.97733271e-01
-1.76989758e+00 -6.77010596e-01 -3.98168564e-01 1.98787665e+00
1.09885371e+00 -1.92354303e-02 -8.26309770e-02 -2.84406930e-01
6.92108929e-01 7.09003285e-02 -3.53171468e-01 -6.25702500e-01
-4.49371248e-01 3.94528389e-01 4.55532461e-01 7.41221368e-01
-9.52981591e-01 1.81822395e+00 5.22572565e+00 1.20679796e+00
-8.24904263e-01 3.18415284e-01 -5.11511043e-02 1.42705396e-01
-5.02708018e-01 -1.45418108e-01 -1.32076895e+00 6.31821752e-01
1.19000638e+00 -8.84421706e-01 3.39131027e-01 6.71895325e-01
2.62620926e-01 1.84584528e-01 -8.95663142e-01 1.14302540e+00
2.87074655e-01 -1.20467746e+00 5.60685098e-01 -4.31503877e-02
7.66823113e-01 2.64441997e-01 -1.01594158e-01 3.01373303e-01
5.20674467e-01 -9.14820910e-01 4.10584748e-01 1.33849174e-01
6.61975503e-01 -1.10852206e+00 3.18794817e-01 4.42891896e-01
-1.56678963e+00 4.87404048e-01 -9.09238040e-01 1.40146360e-01
2.80488670e-01 5.03137767e-01 -5.66616058e-01 7.33861446e-01
6.18381977e-01 1.15508628e+00 -4.37364280e-01 4.65427607e-01
-5.24450839e-01 6.74122870e-01 -1.55474305e-01 -4.14695531e-01
2.27525100e-01 -7.02866375e-01 3.46703947e-01 1.55679584e+00
5.55613399e-01 6.63069040e-02 2.22114295e-01 6.79957092e-01
-1.48697719e-01 1.18382919e+00 -6.40016198e-01 -3.90548795e-01
7.00392485e-01 1.03002131e+00 -7.28762507e-01 -2.81631678e-01
-8.30882847e-01 1.05535722e+00 6.44840658e-01 -2.37412341e-02
-6.52437508e-01 -3.01574230e-01 9.36534166e-01 -1.83757469e-01
1.66725844e-01 -2.38170028e-01 -3.21180731e-01 -1.41563046e+00
5.51018864e-02 -6.90840542e-01 6.69366062e-01 -1.82954133e-01
-1.70604634e+00 6.36804163e-01 -2.37973034e-01 -9.82604742e-01
-3.15186590e-01 -1.14433634e+00 -1.25312865e-01 8.18264544e-01
-1.59314179e+00 -1.10055315e+00 3.14424425e-01 8.08652580e-01
8.20054531e-01 -4.95781928e-01 1.28985548e+00 3.48206013e-01
-3.66438985e-01 5.11112869e-01 5.26952267e-01 2.84985811e-01
7.05008864e-01 -1.23393857e+00 3.43408883e-01 8.28602552e-01
7.29844749e-01 8.19411695e-01 4.91400659e-01 -9.11269367e-01
-1.35648596e+00 -1.12356615e+00 1.58980596e+00 -6.91637814e-01
1.18494701e+00 -4.52801585e-01 -8.87665153e-01 9.57118273e-01
4.96828496e-01 -5.00048578e-01 1.20528674e+00 5.15788317e-01
-2.40142867e-01 8.34038705e-02 -5.95944226e-01 7.49255240e-01
1.24708080e+00 -6.64120436e-01 -1.07564270e+00 4.08943653e-01
5.01615584e-01 -1.00927852e-01 -8.20624650e-01 1.06128104e-01
2.38037512e-01 -6.87950969e-01 8.97296727e-01 -5.67570984e-01
2.51488000e-01 -2.01163083e-01 -4.33906436e-01 -1.60870838e+00
-3.16685081e-01 -1.29003415e-03 -4.55755778e-02 1.06700647e+00
7.05821276e-01 -5.85814774e-01 5.63871443e-01 -4.96633761e-02
-1.93142340e-01 -2.82880187e-01 -8.13327074e-01 -1.14695740e+00
5.20045102e-01 -6.12792671e-01 4.35743362e-01 1.30075514e+00
4.84715998e-01 7.43275225e-01 -2.81103969e-01 6.95431158e-02
8.19181204e-01 2.39937246e-01 1.54508561e-01 -1.15057003e+00
4.81749587e-02 -5.72827935e-01 -3.24607581e-01 -8.78292501e-01
1.00690734e+00 -1.63071895e+00 -1.45556942e-01 -1.62655866e+00
2.85841733e-01 -2.40456432e-01 -6.29015565e-01 5.37863374e-01
-1.82869315e-01 -1.35628798e-03 1.50717199e-02 -9.43820551e-02
-6.41506553e-01 5.63692153e-01 9.25803185e-01 2.00702995e-01
-1.81330830e-01 -5.30413687e-01 -7.22638249e-01 9.33906257e-01
1.14508784e+00 -4.15510356e-01 -5.71435392e-01 -6.37615383e-01
2.60355681e-01 -5.31293929e-01 -3.11548948e-01 -7.44213879e-01
1.77098930e-01 -5.62346458e-01 5.54239191e-02 -5.29560208e-01
2.93841124e-01 -8.63739192e-01 1.02797724e-01 2.12263077e-01
-6.37259260e-02 2.49895424e-01 3.86790276e-01 1.39963672e-01
-4.97383684e-01 -1.98071241e-01 7.15853035e-01 -2.22130269e-01
-1.41000175e+00 2.06481799e-01 -4.11061347e-01 4.25845772e-01
1.08187795e+00 -4.91539128e-02 -2.19072461e-01 1.80902705e-01
-5.61762333e-01 1.96117297e-01 4.67779160e-01 1.04415357e+00
7.05818057e-01 -1.48637724e+00 -7.51908779e-01 1.07772760e-01
4.07271892e-01 -6.21548295e-01 7.74497017e-02 3.76785398e-01
-3.15348029e-01 8.78019989e-01 -2.41871431e-01 -1.26750574e-01
-1.32624054e+00 6.17492199e-01 1.18859388e-01 4.02060375e-02
-4.20187473e-01 7.45981872e-01 2.32020736e-01 -9.77150023e-01
-7.41114374e-04 -6.41102195e-02 -3.45007658e-01 9.39211249e-02
4.40113008e-01 2.74999410e-01 1.04516514e-01 -1.11572731e+00
-5.91394842e-01 4.91785556e-01 -1.79621980e-01 8.84893239e-02
1.31982243e+00 -5.32719076e-01 -4.44930732e-01 1.06520307e+00
1.24912441e+00 4.04354423e-01 -1.91407919e-01 -4.92183298e-01
6.77763641e-01 -2.63197482e-01 -1.83381978e-02 -7.51748621e-01
-6.20563447e-01 7.98555136e-01 3.10534954e-01 -2.83481985e-01
7.36046851e-01 1.81559473e-01 9.54048038e-01 6.12392068e-01
6.86460018e-01 -1.60593164e+00 -4.77212995e-01 1.00666475e+00
4.99729276e-01 -1.31197298e+00 -3.80767375e-01 -6.00850999e-01
-7.93063521e-01 8.16907942e-01 7.18733370e-01 7.26552457e-02
9.70620930e-01 1.20111043e-02 4.16158795e-01 -1.92000851e-01
-7.80096650e-01 -4.54570532e-01 5.64622045e-01 2.97670692e-01
1.09213340e+00 5.77049494e-01 -1.02932167e+00 7.95055628e-01
-7.43011177e-01 -3.53138417e-01 9.00460705e-02 5.51221550e-01
-6.96093738e-01 -1.76643634e+00 -5.41297160e-02 4.81699258e-01
-3.85505319e-01 -7.23076999e-01 -4.49010491e-01 7.82222986e-01
2.55438685e-01 8.91957939e-01 5.07994853e-02 -2.39878118e-01
2.28917211e-01 4.41153049e-01 4.38350201e-01 -9.22647595e-01
-1.19918823e-01 4.02030461e-02 2.22914219e-01 -4.08923060e-01
-5.46999097e-01 -7.93202996e-01 -1.81965041e+00 -3.93212438e-01
-9.40601749e-04 5.83090067e-01 5.41621327e-01 9.41532254e-01
1.62263885e-01 1.68843716e-01 2.90169418e-01 -1.49759233e-01
-6.27715886e-02 -8.29581976e-01 -1.00591028e+00 4.99653906e-01
-4.32583183e-01 -5.94514966e-01 -2.43047744e-01 2.00206906e-01] | [10.701483726501465, 9.63391399383545] |
39a3276a-85e7-40ff-9f99-6721ed69b8ac | atca-an-arc-trajectory-based-model-with | 2208.00856 | null | https://arxiv.org/abs/2208.00856v1 | https://arxiv.org/pdf/2208.00856v1.pdf | ATCA: an Arc Trajectory Based Model with Curvature Attention for Video Frame Interpolation | Video frame interpolation is a classic and challenging low-level computer vision task. Recently, deep learning based methods have achieved impressive results, and it has been proven that optical flow based methods can synthesize frames with higher quality. However, most flow-based methods assume a line trajectory with a constant velocity between two input frames. Only a little work enforces predictions with curvilinear trajectory, but this requires more than two frames as input to estimate the acceleration, which takes more time and memory to execute. To address this problem, we propose an arc trajectory based model (ATCA), which learns motion prior from only two consecutive frames and also is lightweight. Experiments show that our approach performs better than many SOTA methods with fewer parameters and faster inference speed. | ['Jie Yang', 'Lingtong Kong', 'Jinfeng Liu'] | 2022-08-01 | null | null | null | null | ['video-frame-interpolation'] | ['computer-vision'] | [-2.68398315e-01 -3.73780549e-01 -4.02353346e-01 -1.31740168e-01
-3.53536189e-01 -1.94390729e-01 6.05850756e-01 -2.45567724e-01
-3.76014292e-01 8.07100058e-01 7.82020018e-02 -3.04060668e-01
2.81612813e-01 -7.91255414e-01 -8.89462113e-01 -4.73103404e-01
-2.63219960e-02 5.66332228e-02 6.43465161e-01 2.75008380e-02
3.13553661e-01 4.28696394e-01 -1.28108513e+00 1.40166923e-01
7.79204190e-01 8.56813014e-01 1.31180510e-01 8.90128016e-01
-1.51669711e-01 1.20250309e+00 -3.51357251e-01 -2.93403983e-01
4.94871527e-01 -5.36422133e-01 -8.95474851e-01 -5.63506000e-02
7.44077206e-01 -9.78517950e-01 -8.42023551e-01 7.82448590e-01
1.91234842e-01 3.92262071e-01 2.99120456e-01 -1.42472589e+00
-5.81118345e-01 1.74532637e-01 -6.64682031e-01 4.12228227e-01
4.12254900e-01 5.62376797e-01 6.61396265e-01 -8.01984370e-01
6.65814340e-01 1.37222469e+00 8.15307021e-01 5.93091667e-01
-1.13642621e+00 -6.03907228e-01 1.66561589e-01 5.95555365e-01
-1.16023099e+00 -2.47530445e-01 6.49772823e-01 -4.03746992e-01
7.29474604e-01 7.76275322e-02 1.13788271e+00 8.35637093e-01
1.44399017e-01 1.01778400e+00 8.24662685e-01 -7.56444260e-02
1.38548106e-01 -4.73042309e-01 -2.67192751e-01 8.71855855e-01
3.96596268e-02 1.41310081e-01 -3.61810058e-01 1.01278216e-01
1.34893453e+00 1.19079299e-01 -4.63016838e-01 -7.03000724e-02
-1.43872273e+00 8.52425456e-01 3.71414185e-01 1.11125566e-01
-2.37361684e-01 6.30645156e-01 3.28148007e-01 1.03574231e-01
3.39122951e-01 -3.72683555e-02 -2.55831599e-01 -5.37494600e-01
-1.14910376e+00 6.35306656e-01 6.80518270e-01 1.16770506e+00
1.02486289e+00 3.57681334e-01 -1.89318806e-01 4.30071086e-01
2.78384626e-01 4.15801644e-01 2.68040538e-01 -1.55379021e+00
3.41996491e-01 1.47635296e-01 3.80354345e-01 -1.37528765e+00
-1.75492033e-01 8.87699798e-02 -9.67346311e-01 4.75502878e-01
7.46224582e-01 -2.50371128e-01 -6.60831928e-01 1.43501437e+00
3.90349597e-01 1.01226544e+00 -3.02192330e-01 1.25503087e+00
4.30597514e-01 1.09171546e+00 -1.16527140e-01 -3.48824471e-01
8.57061923e-01 -1.45657611e+00 -8.19333434e-01 4.39832769e-02
4.07101482e-01 -8.33101630e-01 1.08471620e+00 3.76589239e-01
-1.43578541e+00 -9.10989344e-01 -9.88352180e-01 -4.53231454e-01
1.97685018e-01 -1.06764235e-01 8.86071801e-01 4.28317279e-01
-1.02422059e+00 9.34309423e-01 -1.23596835e+00 2.47074421e-02
6.01863682e-01 2.30627507e-01 -8.40889066e-02 -1.06033489e-01
-9.73480105e-01 7.11808681e-01 1.41212761e-01 2.35239401e-01
-7.58133054e-01 -9.99056578e-01 -9.44876671e-01 -2.99452478e-03
1.89281002e-01 -9.46902633e-01 1.29583621e+00 -9.90275621e-01
-2.01711011e+00 1.98420063e-01 -6.09006345e-01 -6.91224635e-01
9.84951198e-01 -4.99380440e-01 -9.11160111e-02 2.68031031e-01
-2.28690892e-01 9.15831804e-01 9.65422451e-01 -9.39714253e-01
-8.68853331e-01 2.18135610e-01 4.32547569e-01 -7.98637122e-02
-4.73262779e-02 -6.13756068e-02 -5.93473434e-01 -6.78109765e-01
-1.96619734e-01 -8.76725554e-01 -3.33952397e-01 6.87592447e-01
-7.12261349e-02 -4.23307329e-01 1.00950778e+00 -5.67820847e-01
1.33706808e+00 -1.79434180e+00 -3.56781930e-02 -3.17109227e-01
1.50425017e-01 5.64267516e-01 7.19532147e-02 6.63887784e-02
2.57678360e-01 -3.57960686e-02 -2.35375717e-01 -3.53964478e-01
-1.78006411e-01 2.59723753e-01 -4.89981949e-01 4.97191966e-01
1.26192302e-01 8.14612448e-01 -1.19237173e+00 -7.18424916e-01
6.06607914e-01 7.59984910e-01 -8.87225330e-01 2.42780164e-01
-3.24206531e-01 6.52569473e-01 -2.29872957e-01 1.25821277e-01
7.81516135e-01 -2.34967336e-01 -8.66897330e-02 -3.56121480e-01
-3.49352568e-01 8.75864625e-02 -1.25841057e+00 1.93312991e+00
-3.98098439e-01 1.09372640e+00 -3.23753923e-01 -8.55534017e-01
7.14565635e-01 3.30044419e-01 6.96110964e-01 -3.86137009e-01
1.28873646e-01 -6.52444735e-02 -1.30243316e-01 -7.96518147e-01
4.26949620e-01 1.08502798e-01 6.41146898e-01 1.82038084e-01
-3.03985178e-01 -1.96275175e-01 2.94654042e-01 9.32708010e-02
8.23771417e-01 7.96202183e-01 2.46307049e-02 5.37135266e-02
7.84018993e-01 1.71725020e-01 1.08955419e+00 4.64899629e-01
-4.16183650e-01 8.33390713e-01 3.58268529e-01 -1.01443863e+00
-1.33527935e+00 -8.32132220e-01 9.63902026e-02 4.57574844e-01
5.37091374e-01 -5.36660194e-01 -7.87244081e-01 -3.00280392e-01
-1.36717290e-01 3.83834183e-01 -3.66561025e-01 1.80958837e-01
-1.33884716e+00 -2.96856374e-01 3.14280927e-01 7.44936943e-01
7.19042718e-01 -6.67283118e-01 -7.01137006e-01 3.91860545e-01
-2.40100935e-01 -1.46075213e+00 -8.57351840e-01 -8.63858163e-01
-1.02057147e+00 -1.03201926e+00 -8.49712014e-01 -8.31990302e-01
6.53957963e-01 4.02988672e-01 9.41004038e-01 4.67426330e-01
-2.57324278e-01 4.04711962e-02 -1.25306427e-01 -1.01015672e-01
-2.35405907e-01 -3.16793233e-01 -1.07224457e-01 7.21184015e-02
2.10674748e-01 -4.82180268e-01 -9.66033936e-01 3.08152497e-01
-8.04311395e-01 5.35166979e-01 9.31315348e-02 8.20061386e-01
6.62007928e-01 -2.67266631e-01 2.28308007e-01 -4.47479963e-01
2.56391048e-01 -2.38021493e-01 -8.31801891e-01 -1.12213537e-01
-2.88662642e-01 -3.31405364e-02 1.09706748e+00 -4.80416805e-01
-1.05202138e+00 2.21234202e-01 -1.51808634e-01 -9.20938432e-01
-9.94247478e-03 1.54365495e-01 2.51513422e-01 -6.16136901e-02
4.18605983e-01 1.98946163e-01 1.57767668e-01 -1.85381129e-01
2.81433046e-01 1.93781167e-01 7.44898617e-01 -5.05496264e-01
8.95576715e-01 8.71953309e-01 2.06274658e-01 -8.31592202e-01
-7.00955570e-01 -2.45583534e-01 -7.67308295e-01 -4.50636595e-01
9.82656419e-01 -8.02064478e-01 -1.15369856e+00 4.71133202e-01
-1.53938258e+00 -6.47329330e-01 6.06189929e-02 7.85890520e-01
-8.06557357e-01 6.86839998e-01 -8.79745960e-01 -5.62230706e-01
-2.12858841e-01 -1.35886288e+00 6.70805156e-01 3.63216609e-01
-7.13921711e-02 -1.15028274e+00 1.51320413e-01 3.01105343e-03
3.69268358e-01 4.51023519e-01 3.66754264e-01 4.26301152e-01
-1.14301646e+00 1.17321379e-01 -3.16651642e-01 1.35622591e-01
1.45456105e-01 4.84683007e-01 -6.38888240e-01 -3.44134539e-01
-8.34240168e-02 1.21292733e-02 6.67948544e-01 5.41170418e-01
1.50489700e+00 -2.93595642e-01 -2.05765009e-01 1.01522219e+00
1.35276341e+00 1.88909993e-01 7.95235515e-01 2.45401725e-01
9.35217679e-01 2.64766812e-01 6.94984376e-01 3.10828030e-01
6.23830557e-01 6.99650586e-01 3.89585286e-01 -2.49703508e-02
-3.39950413e-01 -2.68875510e-01 3.37599456e-01 8.98472786e-01
-6.15938842e-01 1.39246322e-02 -6.57842755e-01 4.87141490e-01
-2.13789749e+00 -1.38103223e+00 -4.75534499e-01 2.01367259e+00
8.95322561e-01 1.29934266e-01 2.13611782e-01 1.53484628e-01
5.86632371e-01 2.75391072e-01 -6.31776035e-01 -2.16750368e-01
1.63305387e-01 9.84384120e-02 4.56599921e-01 7.99594283e-01
-1.05755937e+00 1.05697668e+00 6.95472479e+00 6.59063280e-01
-1.28349400e+00 -6.34194836e-02 5.48050463e-01 -1.35325417e-01
-2.30130613e-01 3.23097795e-01 -6.50965035e-01 7.09198713e-01
7.02741683e-01 -3.03011924e-01 4.42093998e-01 7.55122602e-01
6.54669642e-01 -9.19452533e-02 -1.14740360e+00 1.27754045e+00
-7.50366598e-02 -1.92670238e+00 8.87284502e-02 -2.07511604e-01
8.22790086e-01 -1.58292070e-01 -2.49816597e-01 1.16591886e-01
1.93071574e-01 -8.96330953e-01 6.28207684e-01 6.52236760e-01
5.75247526e-01 -6.93178415e-01 3.16302627e-01 3.10275912e-01
-1.41442204e+00 2.43514910e-01 -7.10117579e-01 -4.23377693e-01
5.59565544e-01 4.50167865e-01 -3.47252160e-01 4.00967389e-01
6.85952187e-01 1.06928396e+00 -1.31923303e-01 1.36058545e+00
-2.40492180e-01 6.10159993e-01 -2.98301786e-01 3.83325964e-02
4.71355915e-01 -4.51889843e-01 4.37535673e-01 1.12056649e+00
3.68630856e-01 1.42624006e-01 4.98085320e-01 7.37144828e-01
6.87232018e-02 -2.15892315e-01 -4.03125584e-01 4.77126479e-01
3.29291195e-01 1.12354779e+00 -5.85152209e-01 -5.13342321e-01
-8.15236568e-01 9.23429787e-01 1.31866470e-01 4.04290646e-01
-1.26538658e+00 -3.28856617e-01 9.47241366e-01 9.53533575e-02
2.30201870e-01 -8.01342249e-01 -5.08669205e-02 -1.39671361e+00
3.53858317e-03 -4.30261880e-01 3.53847072e-02 -8.07758152e-01
-8.41037273e-01 3.56303841e-01 -2.43779719e-01 -1.60891759e+00
-3.54828805e-01 -4.58622903e-01 -6.96443856e-01 6.98108912e-01
-1.84573591e+00 -9.07053769e-01 -5.83716512e-01 7.93524981e-01
8.81471157e-01 1.50497228e-01 4.01028425e-01 5.22540689e-01
-5.08939505e-01 4.10362482e-01 -8.34202692e-02 3.64113212e-01
6.59819186e-01 -1.08435166e+00 6.55554056e-01 9.37740624e-01
9.39499810e-02 3.96560103e-01 7.59865642e-01 -4.40944552e-01
-1.75680447e+00 -1.06774068e+00 6.38137817e-01 -2.73580045e-01
5.57737350e-01 1.94752350e-01 -1.03553092e+00 7.86147833e-01
4.67527479e-01 5.68822920e-01 9.74546745e-02 -6.19175613e-01
-6.30666316e-02 -1.83514401e-01 -8.42988074e-01 8.06006551e-01
1.13032353e+00 -3.47606689e-01 -2.42453218e-01 1.65400177e-01
7.58508205e-01 -7.70820260e-01 -8.77803981e-01 1.48762703e-01
5.22110820e-01 -1.12611198e+00 1.08889997e+00 -5.15738130e-01
6.63678408e-01 -7.47756660e-01 2.80939430e-01 -9.86340225e-01
-3.61954510e-01 -1.15115881e+00 -6.53045714e-01 7.83429205e-01
-1.06662750e-01 -3.06397140e-01 9.17888522e-01 6.06519401e-01
-1.21696994e-01 -9.33880448e-01 -6.03783131e-01 -9.15736496e-01
1.54700935e-01 -4.89329875e-01 6.12574577e-01 9.21373606e-01
-2.47430339e-01 7.78536592e-03 -6.38951898e-01 9.82382298e-02
5.77238500e-01 2.79435277e-01 1.06203258e+00 -9.42706704e-01
-1.24752283e-01 -4.16531861e-01 -4.48072106e-01 -1.63533616e+00
2.28579432e-01 -5.44441938e-01 6.57912763e-03 -1.55375993e+00
-1.66898027e-01 -3.95796359e-01 1.37312680e-01 1.42043903e-01
-3.53685439e-01 2.44829595e-01 2.83804536e-01 2.79696316e-01
-3.90774190e-01 5.01784205e-01 1.61373425e+00 -1.72635660e-01
-1.10378444e-01 -7.59199634e-02 5.17324470e-02 1.11384714e+00
6.85320914e-01 -1.65933415e-01 -4.73240584e-01 -9.48634863e-01
4.98130172e-02 3.54207337e-01 4.12987053e-01 -1.16908562e+00
4.49812025e-01 -4.75021660e-01 4.74277318e-01 -7.21589148e-01
3.53054225e-01 -6.93670988e-01 8.75408128e-02 6.49021685e-01
-1.76922664e-01 2.59581298e-01 5.73604107e-02 5.51013589e-01
-2.39273474e-01 -1.66518047e-01 9.23143983e-01 -2.67136618e-02
-8.11650515e-01 9.97749150e-01 -3.15964371e-01 3.56052332e-02
1.07747304e+00 -3.20011526e-01 -1.22098578e-02 -4.37195837e-01
-5.23128271e-01 1.05070204e-01 4.42649335e-01 4.72084403e-01
7.34186769e-01 -1.58285892e+00 -7.17361927e-01 4.13534790e-02
-5.98295093e-01 3.78760487e-01 1.87982306e-01 7.47999370e-01
-1.21992910e+00 2.93914855e-01 -2.01059461e-01 -8.55374753e-01
-9.12889302e-01 6.73816502e-01 2.26195663e-01 7.97341093e-02
-1.16609955e+00 5.95266998e-01 -2.12451108e-02 2.10844293e-01
2.08888352e-01 -4.49694663e-01 -1.23676620e-02 -3.21106523e-01
9.00499463e-01 7.81449735e-01 -4.26073283e-01 -5.29103696e-01
6.26449510e-02 8.15280914e-01 1.09408952e-01 2.99632661e-02
1.06054676e+00 -1.67092681e-01 8.56983587e-02 2.28930295e-01
1.29646957e+00 -2.70841748e-01 -2.14613199e+00 -1.40034765e-01
-1.50389209e-01 -1.01794255e+00 2.54479772e-03 7.96894431e-02
-1.31599784e+00 1.10325646e+00 1.95875421e-01 1.24607518e-01
8.53476644e-01 -7.15290189e-01 1.43584514e+00 2.42743641e-01
5.19306421e-01 -1.05219901e+00 9.99017209e-02 5.95165730e-01
6.14773035e-01 -1.24464631e+00 5.82721420e-02 -4.82950866e-01
-2.95576006e-01 1.48892927e+00 9.50034618e-01 -4.73706990e-01
5.79204679e-01 2.69753933e-01 -8.49897340e-02 5.13614058e-01
-7.16536343e-01 1.36853516e-01 2.91671306e-02 5.72856009e-01
4.21179682e-01 -3.92607212e-01 -4.60367322e-01 -1.54196605e-01
-2.04911474e-02 3.37474823e-01 6.83034241e-01 7.61538982e-01
-3.33381146e-01 -1.06957293e+00 -2.50559419e-01 9.19230059e-02
-4.75927085e-01 9.47319791e-02 4.89440471e-01 7.20685840e-01
-1.09466322e-01 8.64531577e-01 3.81943345e-01 -1.16313528e-03
1.27992168e-01 -3.54777902e-01 6.11433208e-01 -8.96597132e-02
-3.26720327e-01 4.48226817e-02 -3.04242104e-01 -9.60677683e-01
-7.84648359e-01 -4.92280215e-01 -1.56295800e+00 -8.12182188e-01
-1.26395598e-02 -6.44743219e-02 3.02895874e-01 9.33259368e-01
3.64677846e-01 4.35292125e-01 7.00019419e-01 -1.19791877e+00
-4.77372147e-02 -5.31549335e-01 5.37347607e-02 5.15165925e-01
5.70087075e-01 -5.50353885e-01 -1.63449138e-01 5.89110255e-01] | [10.610309600830078, -1.3519113063812256] |
53a80f8e-688a-4f97-b05d-823b51aa6257 | learning-multi-subset-of-classes-for-fine | null | null | https://dl.acm.org/doi/abs/10.1145/3552484.3555754 | https://dl.acm.org/doi/abs/10.1145/3552484.3555754 | Learning Multi-Subset of Classes for Fine-Grained Food Recognition | Food image recognition is a complex computer vision task, because of the large number of fine-grained food classes. Fine-grained recognition tasks focus on learning subtle discriminative details to distinguish similar classes. In this paper, we introduce a new method to improve the classification of classes that are more difficult to discriminate based on Multi-Subsets learning. Using a pre-trained network, we organize classes in multiple subsets using a clustering technique. Later, we embed these subsets in a multi-head model structure. This structure has three distinguishable parts. First, we use several shared blocks to learn the generalized representation of the data. Second, we use multiple specialized blocks focusing on specific subsets that are difficult to distinguish. Lastly, we use a fully connected layer to weight the different subsets in an end-to-end manner by combining the neuron outputs. We validated our proposed method using two recent state-of-the-art vision transformers on three public food recognition datasets. Our method was successful in learning the confused classes better and we outperformed the state-of-the-art on the three datasets. | ['Petia Radeva', 'Marc Bolaños', 'Bhalaji Nagarajan', 'Javier Ródenas'] | 2022-10-10 | null | null | null | 30th-acm-international-conference-on | ['food-recognition', 'fine-grained-image-classification'] | ['computer-vision', 'computer-vision'] | [ 4.02397782e-01 -3.00770015e-01 -1.07689977e-01 -6.52880847e-01
-4.77966458e-01 -6.37262404e-01 4.36540216e-01 6.05169773e-01
-5.24557650e-01 3.26704264e-01 1.05185047e-01 3.00427765e-01
-3.84460129e-02 -6.70228720e-01 -9.01885033e-01 -9.74715233e-01
-5.27502336e-02 2.63240606e-01 3.81678492e-01 -5.97227216e-02
2.85042703e-01 1.25348866e-01 -1.85072565e+00 9.15881395e-01
8.88339639e-01 1.19118404e+00 3.88585836e-01 3.04092050e-01
-7.81183094e-02 6.21209860e-01 -2.79362708e-01 3.19772735e-02
1.95549712e-01 -2.73471564e-01 -4.17176604e-01 2.38957450e-01
6.75043344e-01 -4.23492156e-02 2.18246907e-01 1.25541687e+00
2.57666498e-01 2.76983809e-02 8.74361932e-01 -9.54559505e-01
-9.25703645e-01 8.93386841e-01 -7.14543521e-01 2.94284850e-01
2.55674750e-01 1.42230943e-01 6.95969582e-01 -8.33361566e-01
1.21788410e-02 1.34884632e+00 7.82147467e-01 3.85483295e-01
-1.42500496e+00 -8.98037553e-01 5.74865401e-01 4.14363205e-01
-1.24062800e+00 -4.00510818e-01 8.12458396e-01 -6.32453978e-01
8.67367148e-01 1.68569759e-01 6.36374414e-01 9.43532050e-01
8.00308958e-02 8.92039478e-01 1.43263972e+00 -4.01698500e-01
1.22558065e-01 -5.21182381e-02 9.76285040e-01 8.01489532e-01
4.92230564e-01 2.65599728e-01 -2.42900625e-01 1.03612974e-01
2.79687583e-01 4.20075715e-01 -1.95303485e-01 -7.39516675e-01
-1.20116079e+00 1.14087439e+00 9.00623739e-01 3.88966173e-01
-4.19316739e-01 -2.81989157e-01 2.45815992e-01 5.32724261e-02
4.44278747e-01 3.19457829e-01 -4.64601696e-01 6.83039904e-01
-8.34364533e-01 -1.30999818e-01 7.68856645e-01 3.83205563e-01
7.75394261e-01 -2.83595264e-01 -3.29623133e-01 9.83472228e-01
5.15147686e-01 5.16318858e-01 7.49346495e-01 -1.33208990e-01
3.53393614e-01 7.08487511e-01 -6.48894086e-02 -1.00948608e+00
-8.59166384e-01 -5.00047028e-01 -1.15154922e+00 2.70537853e-01
5.75568616e-01 2.97544181e-01 -1.30401087e+00 1.50025856e+00
1.59386829e-01 1.89692199e-01 1.58171266e-01 1.00173092e+00
1.19228721e+00 5.01060009e-01 3.87309134e-01 2.62399375e-01
1.44581914e+00 -1.45441699e+00 -1.78382903e-01 -3.47553730e-01
4.28007364e-01 -7.13927805e-01 8.78528655e-01 4.65645820e-01
-5.83813131e-01 -9.43259180e-01 -1.38526607e+00 2.64291584e-01
-8.24022114e-01 3.89435023e-01 6.16666734e-01 5.00751019e-01
-6.41146719e-01 7.25005150e-01 -9.43805158e-01 -3.57392699e-01
6.03232920e-01 2.30628312e-01 -4.28413451e-01 -2.72367895e-01
-8.61566603e-01 7.48442471e-01 7.72668600e-01 2.31225342e-01
-9.42665637e-01 -5.85293293e-01 -1.25227547e+00 1.67131081e-01
7.18287230e-02 -6.81124747e-01 8.50645185e-01 -1.05651486e+00
-1.37382770e+00 8.85801673e-01 3.17724407e-01 -2.77098596e-01
9.68966931e-02 -1.07668459e-01 -2.93341368e-01 4.70125563e-02
2.51376182e-01 7.83461750e-01 8.81372333e-01 -1.23400640e+00
-8.68564785e-01 -6.46862149e-01 -7.33268559e-02 -1.01447828e-01
-9.44941267e-02 -2.22857371e-01 -7.51849487e-02 -6.97383821e-01
2.00501427e-01 -9.94538069e-01 -2.30709806e-01 -1.66533798e-01
-3.53223622e-01 -4.39707220e-01 4.92875606e-01 -4.35733795e-01
8.88142943e-01 -2.33341742e+00 2.14765444e-01 -3.16831544e-02
1.98092178e-01 4.26840305e-01 -3.49812746e-01 -7.74514303e-02
-2.72744391e-02 -2.12244377e-01 -1.49514019e-01 -7.16474354e-02
2.04548702e-01 2.71661263e-02 1.72054499e-01 5.21767378e-01
4.24670100e-01 7.45673120e-01 -9.68944252e-01 -1.83935791e-01
4.44152474e-01 4.17292148e-01 -3.43107462e-01 2.33937830e-01
-2.12204829e-01 2.25282460e-01 -4.59776789e-01 7.34414697e-01
8.21159482e-01 -3.61478418e-01 -5.47400564e-02 -7.48052716e-01
-8.36641714e-02 -2.09975958e-01 -1.28318822e+00 1.63718379e+00
-1.09276570e-01 -1.74830690e-01 -1.07810177e-01 -1.29820228e+00
8.20893526e-01 -3.75874490e-01 5.40954806e-02 -5.78035712e-01
2.64900565e-01 1.97935417e-01 1.95845813e-01 -4.24076349e-01
4.08807546e-02 1.74909849e-02 -2.37128183e-01 3.94373238e-01
2.45077774e-01 3.06759953e-01 4.79758203e-01 -4.70718563e-01
6.07130587e-01 6.08850382e-02 4.37505215e-01 -5.46991825e-01
4.75027949e-01 1.12912714e-01 3.83318454e-01 8.95046294e-01
-5.04190087e-01 6.09707296e-01 -2.19910577e-01 -8.12605441e-01
-5.82477212e-01 -9.25753951e-01 -5.79164065e-02 1.46976733e+00
2.75631189e-01 -7.68877566e-02 -6.06527090e-01 -9.60002422e-01
3.09142053e-01 3.78591895e-01 -9.95946944e-01 -3.74458700e-01
-3.71113807e-01 -1.13795829e+00 3.73903304e-01 7.38881290e-01
8.00916970e-01 -1.25177848e+00 -8.66255581e-01 2.62755692e-01
-1.02142235e-02 -7.15313971e-01 -6.42229259e-01 9.73150313e-01
-6.14546180e-01 -1.43157530e+00 -8.18004072e-01 -1.27376831e+00
5.49062371e-01 7.51716435e-01 1.20705128e+00 1.37021914e-02
-4.47287738e-01 -2.83581298e-02 -5.51978052e-01 -4.52100784e-01
-1.51568428e-01 1.24545477e-01 -1.40329763e-01 1.45699725e-01
8.99488568e-01 -1.14785768e-01 -7.25249767e-01 3.05666655e-01
-8.37622583e-01 6.73673349e-05 7.92152882e-01 1.00465202e+00
9.05498326e-01 5.74385002e-02 3.80478203e-01 -6.94405973e-01
3.53234798e-01 -4.45696771e-01 -2.99557507e-01 4.00102705e-01
-2.76654333e-01 4.25241917e-01 7.85333037e-01 -7.82836735e-01
-6.07241809e-01 5.19637227e-01 -2.67157797e-02 -1.14282511e-01
-6.07976735e-01 5.12582719e-01 -1.58091530e-01 -1.06677301e-01
5.52046001e-01 1.00222588e-01 -9.22157019e-02 -7.92447686e-01
4.41053480e-01 5.43693364e-01 4.39770311e-01 -4.37086999e-01
3.32203120e-01 1.24652527e-01 -3.96001011e-01 -6.21361613e-01
-1.13603473e+00 -7.55698740e-01 -9.55588281e-01 2.22945258e-01
1.13267374e+00 -9.77269530e-01 -4.39350396e-01 8.79800022e-01
-7.19704211e-01 -4.32930470e-01 4.70824838e-02 5.36588132e-01
-2.43723318e-01 3.73037934e-01 -5.49867690e-01 -2.32039467e-01
-4.80916083e-01 -1.13155675e+00 1.12071967e+00 5.58246255e-01
1.29813952e-02 -6.51430726e-01 1.28298342e-01 2.56931454e-01
3.68061334e-01 4.28720444e-01 9.14474487e-01 -7.91164696e-01
7.99487717e-03 1.72798291e-01 -4.72693413e-01 1.38514370e-01
4.70616072e-01 -2.47056857e-01 -9.07812595e-01 -5.49780548e-01
-2.06127971e-01 -5.21603644e-01 1.63933086e+00 3.93380493e-01
1.16392481e+00 1.23157473e-02 -6.03557646e-01 7.59119749e-01
1.42022383e+00 2.70904064e-01 1.98418111e-01 5.51128507e-01
8.13752770e-01 4.87259388e-01 4.68968958e-01 1.70938075e-01
5.40002167e-01 5.83314955e-01 4.40051556e-01 -3.63085926e-01
-3.34964931e-01 -7.88627006e-03 3.70791763e-01 6.39172614e-01
2.28338793e-01 7.75201470e-02 -3.54507923e-01 3.62750620e-01
-1.98696339e+00 -8.61609280e-01 1.04790941e-01 1.88565421e+00
7.91305542e-01 8.30718204e-02 3.26756716e-01 1.97374448e-01
8.12357306e-01 2.06630342e-04 -7.62065709e-01 -1.98813453e-01
-1.69076577e-01 3.71386379e-01 4.94451046e-01 1.28522590e-01
-1.85386336e+00 8.33309710e-01 5.59365988e+00 5.92762291e-01
-1.23989761e+00 -1.13156267e-01 6.55251443e-01 2.46939585e-01
2.84851134e-01 -7.08405256e-01 -8.62715304e-01 5.27447104e-01
5.99335432e-01 5.44147193e-01 2.30521977e-01 7.70505965e-01
-3.55179459e-01 2.12607551e-02 -1.34199798e+00 9.49963927e-01
4.42504585e-01 -1.04177332e+00 1.20241657e-01 -5.15361607e-01
6.00861132e-01 8.36890042e-02 -9.60302725e-02 5.32658100e-01
5.29703200e-01 -1.00456798e+00 8.99520457e-01 3.07280302e-01
4.15401220e-01 -5.10979891e-01 7.22167194e-01 2.76006222e-01
-1.45709479e+00 -2.26599231e-01 -7.52876103e-01 -1.23046696e-01
-5.19622326e-01 4.67641324e-01 -2.66356528e-01 4.40538794e-01
1.06816375e+00 7.73802936e-01 -9.21135902e-01 1.27422988e+00
-1.10661596e-01 2.78201401e-01 -3.73695523e-01 -1.52344897e-01
4.74350959e-01 -7.02085495e-02 -2.17891708e-01 1.50839007e+00
1.21609159e-01 1.87262550e-01 8.71302843e-01 5.97385466e-01
2.53520787e-01 -1.73438005e-02 -3.37817699e-01 -4.02658479e-03
-1.44392803e-01 1.49113345e+00 -1.07158864e+00 -3.21660787e-01
-4.52845335e-01 1.01345706e+00 5.53287506e-01 -3.73040885e-02
-7.11144865e-01 -4.42660719e-01 3.84313494e-01 -3.44010085e-01
9.45407510e-01 1.90395303e-02 1.22395694e-01 -1.27505553e+00
-2.46706977e-01 -1.02984107e+00 8.78102899e-01 -3.86894643e-01
-1.83695590e+00 8.18922639e-01 -1.39675185e-01 -9.10983562e-01
5.85968830e-02 -9.16067481e-01 -3.13974082e-01 5.63345969e-01
-1.82932127e+00 -1.51227641e+00 -5.40768325e-01 7.55123734e-01
5.06122172e-01 -4.89046834e-02 1.22707021e+00 4.30292308e-01
-6.33770049e-01 6.85288489e-01 2.06370130e-01 2.58115113e-01
8.18971395e-01 -1.24618328e+00 -2.31600758e-02 5.94505131e-01
3.18816364e-01 3.89278054e-01 3.93944144e-01 -4.70522016e-01
-1.16334701e+00 -1.37539482e+00 5.85376322e-01 -8.81057680e-02
4.79349136e-01 -5.09833097e-01 -9.16958451e-01 4.03637767e-01
1.62754193e-01 2.98253983e-01 9.99888659e-01 1.52297586e-01
-8.53187561e-01 -3.40903848e-01 -1.30948317e+00 -6.78660721e-02
7.34898090e-01 -1.94146976e-01 -9.58967984e-01 1.44784048e-01
5.24909019e-01 -6.42700791e-02 -5.55413246e-01 5.37369251e-01
7.11391628e-01 -7.36604154e-01 9.61196482e-01 -6.76014304e-01
2.44850785e-01 -6.57169759e-01 -4.40039814e-01 -1.89089239e+00
-1.03175092e+00 2.17968032e-01 9.68830734e-02 9.96949375e-01
3.02132636e-01 -4.92563516e-01 6.26630485e-01 3.59666459e-02
-2.70586908e-01 -4.81964290e-01 -4.26018894e-01 -6.47245526e-01
1.17243677e-01 3.17502618e-01 6.35852993e-01 9.13753569e-01
-1.19736932e-01 5.49262881e-01 -1.22578263e-01 2.64441341e-01
7.97641635e-01 8.74428511e-01 3.01808715e-01 -1.50110745e+00
-3.58904958e-01 -4.91979867e-01 -4.14467692e-01 -8.59966218e-01
5.34719378e-02 -1.09228528e+00 4.51845199e-01 -1.45730865e+00
7.38481998e-01 -4.17872965e-01 -9.56224501e-01 8.65706027e-01
-4.72564042e-01 7.31449068e-01 4.84236211e-01 3.74413580e-02
-8.60525310e-01 3.03146839e-01 9.62535381e-01 -1.01000154e+00
-1.84276372e-01 -1.50751740e-01 -1.02630019e+00 7.12734818e-01
8.81387353e-01 -3.71470362e-01 -9.07603428e-02 -5.09241045e-01
-4.88522798e-01 -6.71193600e-01 3.22679907e-01 -1.19164038e+00
1.12868929e-02 -7.60175148e-03 9.58885670e-01 -6.90406442e-01
-2.46575959e-02 -8.01873207e-01 9.85946804e-02 9.10608768e-01
-2.37559721e-01 -3.14663291e-01 3.39473367e-01 3.66712302e-01
-1.33100420e-01 -1.47636294e-01 9.92653608e-01 -3.18985969e-01
-1.03940260e+00 2.23408476e-01 -1.36843696e-01 -2.67101735e-01
9.98753190e-01 9.11283027e-03 -4.11931902e-01 4.31848705e-01
-8.75905395e-01 2.07336575e-01 1.74118057e-01 6.46643043e-01
5.27344882e-01 -1.45564854e+00 -8.32474589e-01 5.28624177e-01
5.72090149e-01 -2.90948957e-01 1.87131777e-01 4.74819034e-01
-2.69537032e-01 5.12035906e-01 -6.55210972e-01 -9.14054334e-01
-1.39760673e+00 1.07388699e+00 3.06811243e-01 -4.34854954e-01
-5.73447347e-01 7.55614221e-01 7.18178749e-01 -3.76713157e-01
3.73290718e-01 -9.22723591e-01 -8.29074383e-01 2.65327156e-01
9.30661976e-01 5.12100349e-04 1.26476943e-01 -7.04484403e-01
-5.38963020e-01 9.62325931e-01 -3.09474170e-01 9.03681278e-01
1.61588764e+00 -8.38110894e-02 1.89192772e-01 2.58211166e-01
1.25299764e+00 -5.52985191e-01 -1.37343109e+00 -2.15523973e-01
-4.61065210e-02 -2.02636331e-01 -3.65924127e-02 -1.18553805e+00
-1.27698338e+00 8.11798990e-01 1.16555798e+00 3.59053284e-01
1.10802782e+00 3.56460758e-03 6.78551555e-01 2.60941625e-01
3.88204128e-01 -7.76377201e-01 -1.99016053e-02 5.76078653e-01
6.92378879e-01 -1.68047428e+00 -1.91810265e-01 -4.68579978e-01
-2.96751410e-01 1.27541947e+00 5.15099347e-01 -3.90985668e-01
9.10640895e-01 1.16975382e-01 2.58229315e-01 -2.72530109e-01
-2.68413603e-01 -4.38535452e-01 5.56978583e-01 8.37402046e-01
2.71806091e-01 4.48203564e-01 -1.56701952e-01 1.15703464e+00
1.57579973e-01 -3.34137143e-03 8.27627033e-02 7.05141544e-01
-7.87538171e-01 -8.97406399e-01 -4.18725342e-01 3.98751765e-01
-3.18601698e-01 -1.77019686e-01 -2.92206109e-01 4.97804224e-01
5.76649547e-01 1.26360226e+00 4.02101241e-02 -5.45138180e-01
4.77833897e-01 2.62791142e-02 8.74780178e-01 -6.38277233e-01
-9.50602889e-01 1.61112368e-01 -1.12139449e-01 -5.61138928e-01
-7.22053945e-01 -5.44415295e-01 -9.34329450e-01 8.91928226e-02
-2.52366602e-01 2.46107262e-02 3.60796630e-01 8.77266347e-01
1.72729164e-01 6.09447837e-01 5.54605663e-01 -1.28593194e+00
-7.51231790e-01 -1.03830624e+00 -6.60094798e-01 7.83411741e-01
4.33137208e-01 -9.50253725e-01 -3.15831304e-01 1.41911909e-01] | [11.528765678405762, 4.35015869140625] |
cf8f530e-3553-4ddf-a9a9-a08d9b2b9420 | on-a-two-truths-phenomenon-in-spectral-graph | 1808.07801 | null | http://arxiv.org/abs/1808.07801v3 | http://arxiv.org/pdf/1808.07801v3.pdf | On a 'Two Truths' Phenomenon in Spectral Graph Clustering | Clustering is concerned with coherently grouping observations without any
explicit concept of true groupings. Spectral graph clustering - clustering the
vertices of a graph based on their spectral embedding - is commonly approached
via K-means (or, more generally, Gaussian mixture model) clustering composed
with either Laplacian or Adjacency spectral embedding (LSE or ASE). Recent
theoretical results provide new understanding of the problem and solutions, and
lead us to a 'Two Truths' LSE vs. ASE spectral graph clustering phenomenon
convincingly illustrated here via a diffusion MRI connectome data set: the
different embedding methods yield different clustering results, with LSE
capturing left hemisphere/right hemisphere affinity structure and ASE capturing
gray matter/white matter core-periphery structure. | ['Youngser Park', 'Eric Bridgeford', 'Minh Tang', 'Joshua Cape', 'John M. Conroy', 'Vince Lyzinski', 'Joshua T. Vogelstein', 'Carey E. Priebe', 'Avanti Athreya'] | 2018-08-23 | null | null | null | null | ['spectral-graph-clustering'] | ['graphs'] | [-1.34926111e-01 2.80992121e-01 -6.67999759e-02 -9.40446183e-02
4.67075892e-02 -6.78365648e-01 5.87790132e-01 1.94807962e-01
-1.31817207e-01 1.39428318e-01 5.72652698e-01 -2.38767743e-01
-7.00503647e-01 -4.36371118e-01 6.44332124e-03 -1.07499981e+00
-6.55635178e-01 6.61461473e-01 5.88804297e-02 8.08519050e-02
5.39094843e-02 3.62340093e-01 -7.58965135e-01 -4.41566631e-02
8.16213489e-01 3.12787712e-01 4.37222607e-02 5.06575167e-01
-2.45739877e-01 2.92189866e-01 -9.23942924e-02 -3.15130323e-01
3.50989282e-01 -5.48529804e-01 -7.18299627e-01 3.01049531e-01
6.57666326e-02 5.83268285e-01 -6.10048711e-01 1.41043115e+00
3.57917190e-01 9.54440311e-02 1.25583196e+00 -1.71233463e+00
-1.08386588e+00 8.25334013e-01 -1.20248902e+00 3.39480579e-01
1.42021254e-01 -8.81213173e-02 1.14150214e+00 -5.57209492e-01
7.68704832e-01 1.34066224e+00 6.50783062e-01 2.41085917e-01
-1.97456121e+00 -1.45125315e-01 8.21385682e-02 3.68085861e-01
-1.72942960e+00 -1.16670989e-02 1.06260645e+00 -9.85512793e-01
6.95346832e-01 1.93688616e-01 8.56998086e-01 9.01951432e-01
2.63820559e-01 4.24558818e-01 1.39698327e+00 -1.33117169e-01
3.14620137e-01 9.34408978e-03 4.99911398e-01 5.88814497e-01
4.82357115e-01 -1.46365970e-01 1.28382435e-02 -5.96754611e-01
7.62111485e-01 1.28266186e-01 -4.78901505e-01 -1.05134463e+00
-1.44467080e+00 9.90633011e-01 7.17107236e-01 6.66200757e-01
-4.67030436e-01 1.19328544e-01 4.60757077e-01 5.34507096e-01
1.66908175e-01 9.14966017e-02 9.24869031e-02 5.09044349e-01
-1.08455002e+00 -3.60732228e-01 5.55368125e-01 7.97118485e-01
8.48681808e-01 3.13513801e-02 4.95748997e-01 6.90755606e-01
8.03595722e-01 2.47161582e-01 4.69530523e-01 -9.60572541e-01
9.39094052e-02 6.96173310e-01 -3.91814351e-01 -1.43109024e+00
-8.33222508e-01 -1.65903285e-01 -1.50392473e+00 1.11529209e-01
3.58032078e-01 -1.40911713e-01 -5.25310874e-01 1.82255483e+00
1.48217693e-01 3.23562741e-01 -1.73308298e-01 1.06586218e+00
4.94903296e-01 4.03590351e-01 1.60163313e-01 -5.02404988e-01
1.31203890e+00 -5.24636984e-01 -8.10034275e-01 5.52737527e-02
3.50759745e-01 -4.77839619e-01 7.23213196e-01 9.86858755e-02
-7.85982311e-01 -2.69595891e-01 -7.37544835e-01 2.55567670e-01
-4.46425289e-01 -3.45818162e-01 6.39568448e-01 7.15537548e-01
-1.78086567e+00 4.98771399e-01 -7.84888327e-01 -6.90480173e-01
7.92051256e-02 3.03176224e-01 -8.50017369e-01 6.50530532e-02
-7.95623243e-01 5.56696117e-01 4.83300149e-01 -9.23015326e-02
-3.46786380e-01 -5.62064528e-01 -6.85345888e-01 -7.48212114e-02
2.55460571e-02 -8.38007510e-01 5.28088994e-02 -1.00364363e+00
-1.06973886e+00 1.06518257e+00 -6.73520863e-02 -1.61427990e-01
1.51691258e-01 5.27426243e-01 -6.08944535e-01 5.50775468e-01
1.46599919e-01 6.22728169e-01 9.11606371e-01 -1.60354400e+00
3.01765859e-01 -7.81260967e-01 -7.79855669e-01 1.03537142e-01
-1.48019120e-01 -5.91322258e-02 9.57760066e-02 -6.16542935e-01
1.02550578e+00 -1.14214957e+00 -4.88770843e-01 -2.87264943e-01
-8.12302291e-01 -3.94864798e-01 8.30039680e-01 -7.37121463e-01
1.36121869e+00 -2.23359108e+00 6.00305676e-01 8.81429374e-01
1.12787962e+00 -2.81997412e-01 -1.45556241e-01 9.28897619e-01
-9.69935775e-01 3.36796135e-01 -4.37594116e-01 1.34769276e-01
2.66632646e-01 4.58351262e-02 7.70306662e-02 1.27802169e+00
-1.71341807e-01 9.21202660e-01 -1.17951429e+00 -6.65377975e-01
2.36145914e-01 4.79192287e-01 -3.33564699e-01 -1.91176608e-01
5.92493415e-01 3.31857562e-01 -2.22016007e-01 2.65953630e-01
5.74004114e-01 -5.81418455e-01 6.96282566e-01 -5.21608710e-01
2.51264691e-01 -4.04511452e-01 -1.37115633e+00 1.45809853e+00
4.16176796e-01 7.22912610e-01 4.41746563e-01 -1.42163670e+00
6.11340046e-01 3.76615286e-01 1.01609182e+00 -9.64284316e-03
1.08231463e-01 -2.07758144e-01 6.62175655e-01 -3.81892592e-01
-1.62906885e-01 -1.29290819e-01 7.08020329e-02 8.00591409e-01
2.02476010e-01 2.66716480e-01 -5.53819574e-02 8.60743344e-01
1.36448979e+00 -6.38952136e-01 2.64155269e-01 -9.59633887e-01
2.37929299e-01 -1.80435255e-01 2.55000412e-01 3.31769168e-01
-6.53799653e-01 5.23750126e-01 6.74170494e-01 3.96268182e-02
-1.20741141e+00 -1.86311436e+00 -1.46359071e-01 7.12492168e-01
1.51199982e-01 -6.14212453e-01 -8.61443579e-01 -2.97929883e-01
1.30996943e-01 2.30504185e-01 -8.16849768e-01 -2.17686534e-01
-1.94020838e-01 -1.18367314e+00 2.61498719e-01 2.23622605e-01
7.96931088e-02 -7.63037682e-01 -7.85937824e-04 5.45553267e-02
8.48081987e-03 -7.69861758e-01 -6.39232218e-01 7.51368701e-02
-1.07704031e+00 -1.21716177e+00 -7.82329142e-01 -9.11197841e-01
1.05529499e+00 4.83349502e-01 9.81756806e-01 9.24096182e-02
-5.66387117e-01 8.27775002e-01 -3.24515194e-01 4.63650048e-01
-1.82851627e-01 -3.83829325e-01 5.82028866e-01 2.65631825e-01
3.76658469e-01 -1.13609231e+00 -8.22768271e-01 1.84541211e-01
-9.13248420e-01 -2.37025186e-01 5.65032840e-01 4.88857180e-01
1.86511382e-01 1.74336508e-01 4.99096692e-01 -6.61689460e-01
1.10056913e+00 -8.31869781e-01 -3.60844396e-02 1.43130645e-01
-7.64593482e-01 -1.03948109e-01 4.05364275e-01 -2.97955841e-01
-3.62235308e-01 -4.42074351e-02 4.54102993e-01 -6.83134258e-01
-2.75745600e-01 4.72126096e-01 -1.26175925e-01 5.31838723e-02
6.82236850e-01 3.67310822e-01 2.91828871e-01 -3.10549647e-01
9.80148852e-01 5.17481923e-01 5.33783734e-01 -2.81580329e-01
8.93129110e-01 7.91337192e-01 1.19491918e-02 -9.85225916e-01
1.93410158e-01 -8.56743872e-01 -1.37637079e+00 -3.32551956e-01
1.37710238e+00 -5.39694369e-01 -8.53317380e-01 -3.96961020e-03
-6.27604902e-01 -1.16306372e-01 -6.91710263e-02 6.91717505e-01
-4.90034342e-01 1.02779055e+00 -9.69031096e-01 -6.16434634e-01
-4.70597148e-02 -9.19447064e-01 5.44286251e-01 -1.88555941e-01
-4.30061191e-01 -1.58715880e+00 2.73909718e-01 2.23705620e-02
2.04537600e-01 4.58817482e-01 1.33896232e+00 -5.86973667e-01
-1.18338510e-01 1.05903879e-01 -5.07395744e-01 -8.33221525e-02
2.37164542e-01 1.57251343e-01 -4.24747795e-01 -5.67773223e-01
5.08741327e-02 4.07124788e-01 6.78664267e-01 7.93494642e-01
6.85261309e-01 -2.06987187e-01 -5.77144861e-01 2.68053770e-01
1.68934762e+00 -2.56522186e-02 5.44340670e-01 -1.94979906e-01
9.68710840e-01 7.96796381e-01 -3.61972302e-01 1.85849369e-01
2.96590805e-01 2.43092403e-01 7.36817122e-02 -1.51071087e-01
-1.08831517e-01 6.47459179e-02 3.48590523e-01 1.40093851e+00
-2.05682680e-01 7.35114291e-02 -1.22773838e+00 6.93848252e-01
-1.90735662e+00 -1.30086374e+00 -8.51254165e-01 1.93635046e+00
4.47328031e-01 -2.08559573e-01 5.29462516e-01 -1.08034909e-02
1.18981647e+00 1.51666805e-01 -3.27620685e-01 -7.70130754e-03
-2.52840489e-01 -3.37842792e-01 3.63822609e-01 5.30821621e-01
-8.30587804e-01 8.43587935e-01 7.52182388e+00 7.83069432e-01
-7.59796798e-01 2.40514249e-01 3.95778745e-01 1.81098402e-01
-3.71954292e-01 1.34876639e-01 2.93090880e-01 4.51075375e-01
9.13281441e-01 -3.82530928e-01 7.25561619e-01 4.14019287e-01
2.57991672e-01 1.15874344e-02 -9.04738009e-01 1.41866720e+00
3.41891916e-03 -1.14570045e+00 4.67599556e-02 5.24990857e-01
7.41376281e-01 1.22639574e-01 -5.42240851e-02 -3.16069007e-01
8.29391360e-01 -1.01516056e+00 2.59254783e-01 5.34886658e-01
6.78212047e-01 -4.72007394e-01 3.46766412e-01 2.13103324e-01
-1.36430323e+00 1.23181202e-01 -5.59767485e-01 2.58663476e-01
2.84590840e-01 7.68626809e-01 -6.23663962e-01 5.35563350e-01
6.61098480e-01 8.92144144e-01 -6.60103083e-01 9.98608589e-01
9.97672901e-02 6.18722558e-01 -2.06928402e-01 3.19065899e-01
3.40260893e-01 -1.33159113e+00 7.79390454e-01 1.19836724e+00
-1.10377736e-01 3.11669260e-01 2.42539510e-01 1.23426116e+00
2.80577093e-01 2.40470678e-01 -7.48201489e-01 -2.42417961e-01
4.38622981e-01 1.65060914e+00 -1.70380759e+00 -3.11918855e-01
-3.33304435e-01 1.16266084e+00 2.83754915e-01 7.09905922e-01
-5.46561182e-01 -8.07412118e-02 6.03604734e-01 2.75953829e-01
-1.84023276e-01 -6.44743204e-01 -2.11276740e-01 -1.04868937e+00
-5.80777109e-01 -4.10932302e-01 3.59215826e-01 -6.82411313e-01
-2.01559138e+00 4.20880586e-01 1.07296243e-01 -8.62753689e-01
2.65277565e-01 -5.41946173e-01 -9.33685958e-01 7.37824321e-01
-6.93175375e-01 -7.13666797e-01 -4.04806733e-02 9.59564745e-01
-1.92842975e-01 -1.94555342e-01 5.56630731e-01 3.07068497e-01
-5.42996228e-01 3.18613052e-02 5.19125581e-01 4.05909359e-01
5.93392372e-01 -1.83370006e+00 1.05540588e-01 4.46539402e-01
3.06561917e-01 1.11901987e+00 6.86116397e-01 -8.15031826e-01
-1.47168648e+00 -7.86962390e-01 4.69995171e-01 -5.30803561e-01
1.38795877e+00 -3.93882930e-01 -8.71435940e-01 7.33581603e-01
6.40217960e-01 -1.66988730e-01 1.13693810e+00 1.22109868e-01
-3.22638959e-01 4.10675168e-01 -1.13092887e+00 7.02820241e-01
1.21509194e+00 -8.03117335e-01 -6.66397810e-01 5.96869409e-01
4.33565408e-01 7.21115828e-01 -1.40605199e+00 1.02627695e-01
-2.52856547e-03 -1.10613751e+00 1.08061838e+00 -6.59848154e-01
-1.65203705e-01 -5.01547158e-01 -3.19821648e-02 -1.61943877e+00
-1.11087275e+00 -5.58676243e-01 2.73932248e-01 9.28180158e-01
9.99656543e-02 -6.36422336e-01 6.67994142e-01 2.11871475e-01
-1.87186241e-01 -3.11374784e-01 -1.06057250e+00 -7.38582909e-01
1.86457127e-01 -9.74881500e-02 7.34232366e-02 1.83286428e+00
6.09920084e-01 5.20533741e-01 2.21987963e-01 1.94071367e-01
1.03051543e+00 -9.76637006e-02 3.14257354e-01 -1.52771795e+00
9.45700705e-02 -1.03736866e+00 -8.60026419e-01 -3.45548004e-01
3.42958212e-01 -1.39890516e+00 -5.22274792e-01 -1.60329747e+00
4.63072985e-01 -1.78936079e-01 -2.33489484e-01 1.00638866e-01
-3.07488106e-02 2.34870017e-01 -5.89813944e-03 4.92833942e-01
-7.95713305e-01 3.27592492e-01 8.64691615e-01 -5.87194897e-02
-1.23195060e-01 -5.82693100e-01 -5.04904807e-01 7.05516398e-01
5.90021789e-01 -2.21034348e-01 -6.33441150e-01 1.09593593e-01
2.05000177e-01 9.34905037e-02 4.91868436e-01 -7.94184268e-01
3.03507417e-01 1.27916053e-01 3.23957086e-01 -2.35507727e-01
-1.73768084e-02 -9.12425935e-01 6.34705305e-01 4.95970666e-01
-3.10503077e-02 4.17401135e-01 -1.85860574e-01 1.00665522e+00
-6.87013119e-02 3.48141752e-02 8.04839671e-01 -5.28778918e-02
-6.35586679e-01 3.61762285e-01 -7.04713762e-01 -1.74435526e-02
1.28161240e+00 -4.23173338e-01 -2.79187918e-01 -4.00646716e-01
-1.60047972e+00 7.10519105e-02 5.94042242e-01 2.06896022e-01
6.83862507e-01 -1.67994952e+00 -6.90436006e-01 -5.60590364e-02
-1.16181366e-01 -8.23608339e-01 3.03681850e-01 1.56094921e+00
-4.06858295e-01 1.59698114e-01 -2.84839690e-01 -8.27801228e-01
-1.12251210e+00 1.14475691e+00 2.95802772e-01 7.47625306e-02
-8.90403807e-01 7.19094694e-01 4.07676816e-01 -3.99857134e-01
-3.05795550e-01 1.44321263e-01 -2.36821622e-01 4.50668454e-01
1.14678428e-01 6.24971032e-01 -4.79650259e-01 -1.06214595e+00
-4.35579211e-01 6.08435571e-01 2.81339705e-01 -1.16320752e-01
1.17904806e+00 -4.21022415e-01 -8.19331110e-01 6.81131005e-01
1.39067912e+00 -9.87298228e-03 -6.55209780e-01 -1.90559387e-01
4.37026411e-01 -1.72122926e-01 1.62627250e-01 -2.90561438e-01
-1.17588854e+00 8.54174256e-01 7.63510466e-01 9.06429529e-01
7.76174188e-01 4.94841635e-01 2.76589930e-01 2.29937118e-02
2.74498194e-01 -9.45474863e-01 7.09350705e-02 -1.41275525e-01
7.96322346e-01 -1.07258296e+00 -5.73738031e-02 -5.11243284e-01
-8.05781305e-01 9.32308018e-01 1.04591824e-01 -6.31613195e-01
1.22173941e+00 -7.87515044e-02 -4.81531732e-02 -9.35733318e-01
-3.27583253e-01 -2.67974794e-01 5.94218254e-01 8.26848924e-01
4.50590372e-01 4.15304512e-01 -2.81521231e-01 3.41622114e-01
-2.30863377e-01 -9.67690647e-01 2.92654276e-01 2.94415116e-01
-5.23888946e-01 -8.64807546e-01 -5.48453689e-01 7.67176688e-01
9.54934284e-02 -2.19339982e-01 -9.48041081e-01 7.88190722e-01
-1.60947159e-01 8.72318983e-01 1.69711798e-01 -4.59893495e-01
-2.43424594e-01 2.17420936e-01 4.47731763e-01 -5.37562668e-01
-2.19136342e-01 6.61763668e-01 -6.13747597e-01 -5.93060434e-01
-6.13670707e-01 -9.43695784e-01 -1.47331822e+00 -6.88775063e-01
-2.83080906e-01 5.01708329e-01 2.40732133e-01 5.21334767e-01
4.31873560e-01 2.90199727e-01 4.80347216e-01 -8.49588752e-01
1.87216997e-02 -9.95147407e-01 -1.31756318e+00 8.88283372e-01
5.17866574e-02 -6.07571185e-01 -6.80241168e-01 2.06860110e-01] | [7.115475654602051, 5.202324867248535] |
1ad3b85c-d6a5-4e49-8431-385e1fdf633c | gravl-bert-graphical-visual-linguistic | null | null | https://aclanthology.org/2022.coling-1.22 | https://aclanthology.org/2022.coling-1.22.pdf | GRAVL-BERT: Graphical Visual-Linguistic Representations for Multimodal Coreference Resolution | Learning from multimodal data has become a popular research topic in recent years. Multimodal coreference resolution (MCR) is an important task in this area. MCR involves resolving the references across different modalities, e.g., text and images, which is a crucial capability for building next-generation conversational agents. MCR is challenging as it requires encoding information from different modalities and modeling associations between them. Although significant progress has been made for visual-linguistic tasks such as visual grounding, most of the current works involve single turn utterances and focus on simple coreference resolutions. In this work, we propose an MCR model that resolves coreferences made in multi-turn dialogues with scene images. We present GRAVL-BERT, a unified MCR framework which combines visual relationships between objects, background scenes, dialogue, and metadata by integrating Graph Neural Networks with VL-BERT. We present results on the SIMMC 2.0 multimodal conversational dataset, achieving the rank-1 on the DSTC-10 SIMMC 2.0 MCR challenge with F1 score 0.783. Our code is available at https://github.com/alexa/gravl-bert. | ['Mohit Bansal', 'Tagyoung Chung', 'Chien-Wei Lin', 'Arijit Biswas', 'Shuyang Gao', 'Jiun-Yu Kao', 'Sanchit Agarwal', 'Arpit Gupta', 'Danfeng Guo'] | null | null | null | null | coling-2022-10 | ['coreference-resolution'] | ['natural-language-processing'] | [ 1.99104935e-01 1.57126456e-01 -2.79317647e-01 -1.74597278e-01
-1.05201137e+00 -5.54523885e-01 9.94995475e-01 8.79194662e-02
-2.82667071e-01 6.37161136e-01 6.35459065e-01 -1.81707978e-01
2.97473609e-01 -3.66488934e-01 -4.23756093e-01 -5.11172116e-01
2.49466330e-01 7.64277518e-01 4.67726052e-01 -6.26119912e-01
8.55400786e-02 1.89935729e-01 -1.50084245e+00 1.06980193e+00
6.66294396e-01 6.67192817e-01 3.58983010e-01 1.00447297e+00
-3.65589172e-01 1.12984252e+00 -4.92160022e-01 -8.51143420e-01
-4.80974495e-01 -4.64992881e-01 -1.19542599e+00 -6.88542575e-02
6.64822519e-01 -5.44325076e-02 -4.55535173e-01 1.05823433e+00
5.35235941e-01 2.99358457e-01 5.22833705e-01 -1.69608676e+00
-4.88838315e-01 6.29560828e-01 -5.91193676e-01 1.65111408e-01
8.23065519e-01 7.97361061e-02 1.01116896e+00 -9.15680170e-01
9.06908214e-01 1.88493824e+00 2.50654191e-01 8.30752134e-01
-1.02743542e+00 -5.28523147e-01 8.48200992e-02 8.16779256e-01
-1.22002363e+00 -7.60877848e-01 6.66724145e-01 -2.39013642e-01
9.44133341e-01 6.35985315e-01 3.52212846e-01 1.19759214e+00
-1.50479630e-01 1.23909295e+00 9.91430700e-01 -5.60892582e-01
-2.21520334e-01 6.74177185e-02 -1.10038675e-01 6.77632689e-01
-3.51900816e-01 -3.55236918e-01 -8.35660517e-01 1.29271239e-01
5.36127627e-01 -1.40961379e-01 -5.69190204e-01 -4.83003318e-01
-1.10916984e+00 7.09589839e-01 6.25173509e-01 3.73072356e-01
1.95687860e-01 1.21224495e-02 4.37166601e-01 8.36711079e-02
1.93332762e-01 1.36675164e-01 2.87615985e-01 -1.77209720e-01
-4.24475700e-01 4.44631800e-02 7.54545093e-01 1.08080590e+00
4.18034822e-01 -4.56780255e-01 -3.46385300e-01 1.23505020e+00
3.83385390e-01 5.55024326e-01 -6.11511245e-02 -1.23271012e+00
8.85382652e-01 8.06921899e-01 -2.53973082e-02 -1.22872007e+00
-5.86702466e-01 3.26926447e-02 -1.01575112e+00 -8.77020434e-02
3.22239399e-01 7.21120164e-02 -5.18668830e-01 1.65330470e+00
4.55970734e-01 2.54711062e-02 4.56332803e-01 1.05545628e+00
1.81698394e+00 7.05707848e-01 1.88703746e-01 -2.52341181e-01
1.59195328e+00 -1.24314427e+00 -9.87091899e-01 -2.48876557e-01
3.47638696e-01 -1.00702691e+00 1.02799916e+00 4.93891798e-02
-1.09936702e+00 -3.62925440e-01 -7.42849827e-01 -5.85864902e-01
-4.47299153e-01 -1.24578841e-01 3.86566907e-01 1.52707875e-01
-1.19722724e+00 -1.98279276e-01 -3.65696251e-01 -7.45743811e-01
1.19680271e-01 7.78002068e-02 -5.38986742e-01 -3.46856862e-01
-1.36148012e+00 1.07947302e+00 3.38211477e-01 1.98921397e-01
-7.62152016e-01 -1.36393175e-01 -9.92201984e-01 -6.05134629e-02
7.85076380e-01 -5.68127573e-01 1.62008882e+00 -8.10697675e-01
-1.35781467e+00 1.07691479e+00 -3.36159199e-01 -2.33889893e-01
5.97700417e-01 -3.71679515e-02 -7.21273601e-01 4.62147057e-01
6.47187829e-02 9.70218182e-01 4.96640176e-01 -1.66006243e+00
-9.50097203e-01 -3.10695678e-01 6.00921631e-01 6.96676195e-01
2.33120453e-02 1.49249673e-01 -1.13136387e+00 -6.26662225e-02
-9.35793296e-02 -1.03192651e+00 2.08850086e-01 -3.28894287e-01
-4.47995275e-01 -5.07243156e-01 1.01113403e+00 -6.11044943e-01
9.77976799e-01 -1.92438471e+00 5.92109561e-01 -3.17194134e-01
3.00989360e-01 2.07888067e-01 -3.71198028e-01 5.56164265e-01
1.82679921e-01 -1.74192578e-01 -1.53700024e-01 -6.75543904e-01
9.22672674e-02 8.28179196e-02 -1.40412241e-01 1.96919546e-01
-1.10774348e-02 9.91597116e-01 -9.94807541e-01 -8.96381915e-01
2.84189701e-01 6.65555060e-01 -1.47889048e-01 3.67706120e-01
-4.39800441e-01 4.90155965e-01 -1.57074425e-02 6.55965507e-01
6.02890015e-01 -4.44369793e-01 5.30712545e-01 -6.36763990e-01
-1.80226296e-01 -6.11033924e-02 -8.92087042e-01 1.95794296e+00
-3.24865133e-01 1.10748005e+00 4.16887105e-01 -6.26551867e-01
5.53911984e-01 4.08738971e-01 5.82433641e-02 -1.04848421e+00
4.21641022e-02 -2.81654358e-01 -2.53349870e-01 -7.02181518e-01
8.95729721e-01 1.66127175e-01 -3.74336928e-01 2.60264218e-01
-5.79486899e-02 -1.38045579e-01 4.69676912e-01 7.94496179e-01
6.28479660e-01 1.17816497e-03 2.87952125e-01 4.17659014e-01
8.09225917e-01 3.02954406e-01 3.11732113e-01 5.19330144e-01
-2.84796834e-01 5.40980518e-01 6.75481856e-01 -7.85353854e-02
-5.51366389e-01 -1.00616801e+00 2.40749240e-01 1.44273901e+00
7.06204116e-01 -6.06997073e-01 -6.26557171e-01 -5.73748648e-01
-3.89500767e-01 6.39720500e-01 -4.73009676e-01 1.80124909e-01
-6.07820272e-01 -3.06380898e-01 6.98296249e-01 3.57122600e-01
7.04124331e-01 -1.15651274e+00 -3.18655312e-01 -1.88639969e-01
-1.06780744e+00 -1.46260476e+00 -4.84937608e-01 -4.28058594e-01
-2.36281767e-01 -1.49110758e+00 -5.44438481e-01 -7.64131546e-01
3.46912712e-01 7.15054452e-01 1.36081243e+00 2.01808780e-01
-4.13586378e-01 8.84809196e-01 -4.27454859e-01 -1.10800393e-01
-4.33963686e-01 8.45967210e-04 -3.36016893e-01 1.55306086e-02
2.84341842e-01 1.20203219e-01 -5.12279272e-01 4.29961622e-01
-7.70341635e-01 7.33245790e-01 3.88226151e-01 6.26932323e-01
4.01860595e-01 -3.48791063e-01 2.84808606e-01 -7.43768990e-01
6.88581049e-01 -3.22192460e-01 -3.93190563e-01 6.46673918e-01
1.54677495e-01 -2.60969222e-01 1.26098925e-02 -1.92797109e-01
-1.49889863e+00 -1.99309871e-01 9.87210572e-02 -2.91567922e-01
-2.82057464e-01 3.08431566e-01 -3.96167755e-01 2.80027568e-01
2.56470650e-01 2.09736098e-02 -1.01888761e-01 -2.64188170e-01
8.68091524e-01 9.18982983e-01 1.04251909e+00 -6.49003386e-01
3.40988934e-01 5.10836363e-01 -1.68526486e-01 -1.05402124e+00
-8.48485649e-01 -7.34649301e-01 -6.73693717e-01 -7.89347410e-01
1.09206593e+00 -9.63684916e-01 -1.10107350e+00 2.23781496e-01
-1.45664036e+00 -4.76547956e-01 5.48865199e-01 7.71140903e-02
-4.86487359e-01 5.57161272e-01 -5.10571957e-01 -7.64554083e-01
-3.17174256e-01 -1.25461078e+00 9.48489368e-01 6.07076883e-01
-2.83987582e-01 -9.07286584e-01 1.81278512e-01 1.20879412e+00
1.53019518e-01 1.55787364e-01 7.15015054e-01 -3.25730085e-01
-7.19453573e-01 3.21457922e-01 -6.85612679e-01 -4.14391577e-01
-3.81734669e-02 4.12996560e-02 -1.00436175e+00 -4.18473482e-01
-6.37777925e-01 -6.42294466e-01 9.49904203e-01 1.12031683e-01
7.03091443e-01 -4.35025617e-02 -5.47837377e-01 2.40145907e-01
9.15525317e-01 2.95396447e-01 5.79263449e-01 2.87545800e-01
9.61192906e-01 8.75679255e-01 7.26852357e-01 9.38121602e-02
9.79429662e-01 1.05103648e+00 7.69584835e-01 -3.01127851e-01
-4.71983075e-01 -1.62728041e-01 2.30593294e-01 7.32855618e-01
-1.89348847e-01 -4.63680834e-01 -1.10056376e+00 4.86747831e-01
-2.41416359e+00 -1.28359020e+00 -3.18467200e-01 1.68429339e+00
8.09050500e-01 -3.23673278e-01 5.36757112e-02 -4.63005602e-01
1.09289145e+00 3.55512410e-01 -3.90016705e-01 -2.28390306e-01
-3.86706829e-01 -5.32983065e-01 -2.51157522e-01 8.27939212e-01
-1.16403174e+00 1.10713077e+00 4.75238514e+00 5.40689886e-01
-8.03832650e-01 1.75768241e-01 3.81707937e-01 -1.41282991e-01
-2.97141075e-01 -1.35496810e-01 -8.61436844e-01 5.71098253e-02
5.45067549e-01 1.27756998e-01 4.52969342e-01 4.04841691e-01
2.47982368e-02 -3.11832279e-01 -1.07652259e+00 1.34510827e+00
6.03368759e-01 -1.46401286e+00 2.24075064e-01 -2.83720315e-01
5.54991543e-01 -1.19784169e-01 4.94813845e-02 4.74842310e-01
4.21027571e-01 -1.20060253e+00 4.34416920e-01 5.79530418e-01
9.40074027e-01 -7.88663208e-01 7.21814930e-01 1.07241176e-01
-1.46125066e+00 1.84122443e-01 -1.34306282e-01 3.21658820e-01
2.35222459e-01 -6.32503182e-02 -1.03299248e+00 6.58741653e-01
8.62580180e-01 7.49139845e-01 -5.78503549e-01 8.72985184e-01
-2.59910166e-01 -5.02804741e-02 9.25998390e-02 2.23552845e-02
7.73016810e-02 1.59183010e-01 5.80597997e-01 1.59276676e+00
-2.36434303e-02 2.78570443e-01 2.34338239e-01 5.16100943e-01
-2.87640095e-01 -4.59939614e-02 -4.80671912e-01 1.11533076e-01
6.55138552e-01 1.45116210e+00 -5.10960937e-01 -2.35372201e-01
-6.47231042e-01 9.69111621e-01 5.93871832e-01 4.41698551e-01
-7.35914350e-01 -1.73591450e-01 5.39032996e-01 -3.56138080e-01
-3.58758122e-02 -1.14709705e-01 4.08404380e-01 -1.20190835e+00
-4.52190638e-01 -1.14255214e+00 1.00182712e+00 -1.19764149e+00
-1.07138574e+00 6.60754383e-01 1.33472577e-01 -1.01050794e+00
-1.93589464e-01 -5.84581494e-01 -3.95328581e-01 5.37114799e-01
-1.41487849e+00 -1.39239323e+00 -6.10178947e-01 9.18549597e-01
9.52536464e-01 -1.05934322e-01 8.47041547e-01 2.31180452e-02
-8.34867299e-01 4.08676207e-01 -2.53540576e-01 3.26922387e-01
1.10646164e+00 -1.18757010e+00 -4.28911224e-02 6.57305300e-01
2.23265544e-01 3.38536561e-01 6.85507596e-01 -5.39827228e-01
-1.46038437e+00 -8.01131964e-01 8.08198512e-01 -4.36323851e-01
5.52818477e-01 -3.25226128e-01 -8.55038583e-01 5.93317866e-01
1.04394007e+00 -4.83518720e-01 7.04259872e-01 1.82525560e-01
-4.77568358e-01 2.27922946e-02 -7.72817433e-01 1.06295121e+00
9.98831868e-01 -7.54837692e-01 -7.39822626e-01 2.34913900e-01
6.98948920e-01 -8.44552040e-01 -7.85313308e-01 3.76681745e-01
3.16986352e-01 -9.56275642e-01 1.07383776e+00 -4.84769106e-01
4.49165314e-01 -4.33184534e-01 -3.60235929e-01 -1.24151051e+00
-3.87097299e-02 -6.18803501e-01 -1.21896408e-01 1.50658190e+00
2.45093852e-01 2.75278650e-02 4.28058624e-01 4.93183821e-01
-6.14956431e-02 -6.89568594e-02 -1.04397249e+00 -1.16456328e-02
-4.12061751e-01 -3.23342770e-01 3.93537194e-01 1.02856898e+00
4.16362226e-01 9.01861012e-01 -4.62745756e-01 2.85126150e-01
7.12828159e-01 4.37990963e-01 1.17200994e+00 -1.14673817e+00
1.03981689e-01 -5.41724861e-01 3.57032232e-02 -9.67097282e-01
3.72342676e-01 -8.82393599e-01 7.14115873e-02 -2.05323434e+00
6.52666867e-01 7.51966164e-02 -6.29505292e-02 4.97811377e-01
-4.10202518e-03 2.44137749e-01 5.68973899e-01 2.46507078e-01
-1.41837835e+00 5.00489712e-01 1.35489202e+00 -5.79030156e-01
-4.30940166e-02 -4.21304852e-01 -4.78149503e-01 7.82803535e-01
8.93597543e-01 8.69756117e-02 -4.34580922e-01 -4.10441846e-01
1.50005326e-01 6.35668457e-01 5.07830083e-01 -5.75011313e-01
7.12013304e-01 -4.14506674e-01 2.15252653e-01 -1.00557065e+00
7.72961259e-01 -6.81229353e-01 1.20202601e-01 8.30159485e-02
-5.05424678e-01 1.39277756e-01 4.12351638e-01 6.15383744e-01
-3.84305447e-01 1.16937608e-01 5.50471604e-01 -1.90875113e-01
-1.37997842e+00 -1.24593519e-01 -1.48818240e-01 2.91718364e-01
7.87119448e-01 1.66586623e-01 -1.21297669e+00 -6.96301877e-01
-8.75845015e-01 7.61169672e-01 4.37214315e-01 8.91485214e-01
8.89536381e-01 -1.37530017e+00 -7.94976234e-01 -4.59788322e-01
5.96075475e-01 4.12648208e-02 6.32340252e-01 8.80607128e-01
-2.76156217e-01 7.08224833e-01 -2.63047487e-01 -9.95734155e-01
-2.13288331e+00 2.87294030e-01 3.37742150e-01 -5.24025075e-02
-4.33395356e-01 6.74675822e-01 3.16303670e-01 -4.38957781e-01
4.93949145e-01 3.08023214e-01 -8.06924522e-01 3.62003028e-01
7.77276576e-01 3.56811494e-01 -3.04893374e-01 -1.19440830e+00
-5.98966479e-01 5.66191137e-01 -1.81990787e-01 -2.55057216e-01
1.03986299e+00 -6.16571784e-01 -3.43244404e-01 4.39504147e-01
9.25103962e-01 -2.01384693e-01 -1.13780320e+00 -5.22701681e-01
-1.38024762e-01 -2.30728269e-01 -1.62968770e-01 -8.69104981e-01
-9.55273747e-01 9.15283561e-01 5.78280807e-01 1.30730063e-01
9.11798835e-01 5.00153184e-01 2.78484613e-01 4.95517284e-01
2.22295463e-01 -1.02470541e+00 4.94281411e-01 6.71900332e-01
1.39257467e+00 -1.55869842e+00 -5.34470826e-02 -3.90352190e-01
-1.10114789e+00 1.02209151e+00 8.72624934e-01 8.29298556e-01
-1.52354389e-01 -1.43886684e-02 4.02086735e-01 -2.99080431e-01
-9.32825923e-01 -5.48282325e-01 4.53855902e-01 6.81687534e-01
5.26464641e-01 1.87100500e-01 1.45316914e-01 2.32325703e-01
1.05402134e-01 -3.70956421e-01 4.72827673e-01 6.03850663e-01
-4.97459561e-01 -1.00516856e+00 -6.03838086e-01 -2.61076361e-01
-7.48173594e-02 -1.33205652e-01 -8.14636052e-01 1.02406263e+00
-1.77490681e-01 1.55254257e+00 2.25575358e-01 -1.64290383e-01
3.42348963e-01 -1.14346839e-01 4.21272546e-01 -4.16514844e-01
-2.09134623e-01 3.87669504e-02 6.51050091e-01 -6.24664068e-01
-1.04088902e+00 -5.83449781e-01 -1.32059944e+00 -4.79935497e-01
1.29511759e-01 1.46625265e-01 3.42691898e-01 7.62458086e-01
1.69834554e-01 5.15553832e-01 1.15693904e-01 -7.92183638e-01
4.22605366e-01 -8.59327018e-01 -5.70579320e-02 5.41065097e-01
3.28069091e-01 -7.18236744e-01 -1.08944803e-01 1.56368092e-01] | [10.88027286529541, 1.4546972513198853] |
acda3f09-5f28-4be5-ab5a-ae22c41e5f0a | comparison-of-multiple-features-and-modeling | 1707.04373 | null | http://arxiv.org/abs/1707.04373v2 | http://arxiv.org/pdf/1707.04373v2.pdf | Comparison of Multiple Features and Modeling Methods for Text-dependent Speaker Verification | Text-dependent speaker verification is becoming popular in the speaker
recognition society. However, the conventional i-vector framework which has
been successful for speaker identification and other similar tasks works
relatively poorly in this task. Researchers have proposed several new methods
to improve performance, but it is still unclear that which model is the best
choice, especially when the pass-phrases are prompted during enrollment and
test. In this paper, we introduce four modeling methods and compare their
performance on the newly published RedDots dataset. To further explore the
influence of different frame alignments, Viterbi and forward-backward
algorithms are both used in the HMM-based models. Several bottleneck features
are also investigated. Our experiments show that, by explicitly modeling the
lexical content, the HMM-based modeling achieves good results in the
fixed-phrase condition. In the prompted-phrase condition, GMM-HMM and
i-vector/HMM are not as successful. In both conditions, the forward-backward
algorithm brings more benefits to the i-vector/HMM system. Additionally, we
also find that even though bottleneck features perform well for
text-independent speaker verification, they do not outperform MFCCs on the most
challenging Imposter-Correct trials on RedDots. | ['Yi Liu', 'Zhuzi Chen', 'Michael T. Johnson', 'Liang He', 'Yao Tian', 'Jia Liu'] | 2017-07-14 | null | null | null | null | ['text-independent-speaker-verification', 'text-dependent-speaker-verification'] | ['speech', 'speech'] | [-1.24170281e-01 -5.82793355e-01 -1.68638974e-01 -6.13694906e-01
-1.17015672e+00 -4.87753063e-01 4.57442433e-01 -2.17675045e-01
-5.50628185e-01 5.60340703e-01 2.60438263e-01 -8.20852280e-01
2.37483665e-01 -3.89585481e-03 -1.99377090e-01 -8.90664577e-01
3.04558963e-01 2.11217239e-01 1.49797320e-01 -3.03455651e-01
4.32684034e-01 3.54209632e-01 -1.52730381e+00 4.38964702e-02
6.72459781e-01 4.70420629e-01 2.08429992e-01 8.95666420e-01
-2.67947108e-01 4.31447804e-01 -9.33204532e-01 -4.81027663e-01
-9.27327722e-02 -5.57799399e-01 -7.10830867e-01 -1.01975322e-01
4.06388164e-01 -3.10040057e-01 -1.88279420e-01 8.01826298e-01
1.02534008e+00 1.28504345e-02 4.39695567e-01 -1.11031115e+00
-2.59418011e-01 7.37744510e-01 -3.28205913e-01 5.14902234e-01
7.67057717e-01 -4.94147092e-02 7.12972760e-01 -1.07643282e+00
2.09289983e-01 1.46446168e+00 6.91451252e-01 6.74524844e-01
-1.19024646e+00 -9.10384953e-01 1.92976058e-01 5.47549009e-01
-1.70925045e+00 -1.06438220e+00 6.97797537e-01 -2.27439255e-01
1.11063516e+00 5.03103673e-01 2.03280151e-01 1.20219946e+00
-3.17986719e-02 7.88617432e-01 1.08032322e+00 -8.71632695e-01
3.69082429e-02 4.28131849e-01 5.79322338e-01 1.80818185e-01
-2.52405375e-01 2.03082681e-01 -8.75848711e-01 -2.18756661e-01
1.76799014e-01 -4.20612156e-01 -4.07044113e-01 2.86778957e-01
-1.04932201e+00 9.21328545e-01 -4.36070412e-01 6.26861215e-01
-6.78745881e-02 -3.77253473e-01 3.32810432e-01 3.91545713e-01
3.09051275e-01 -3.46780829e-02 -4.46770906e-01 -5.49130857e-01
-1.33711529e+00 1.93793446e-01 9.57516849e-01 7.48792231e-01
3.74068230e-01 3.49352539e-01 -1.64158806e-01 1.10746896e+00
6.51931286e-01 6.35786772e-01 6.76906884e-01 -4.42264050e-01
5.33085763e-01 -6.03733538e-03 6.03854023e-02 -3.94279361e-01
-3.44733417e-01 -3.55258197e-01 -4.16461527e-01 -1.50228426e-01
7.28146195e-01 -1.61758795e-01 -1.06701410e+00 1.61465120e+00
2.85033554e-01 2.16748178e-01 -2.35196948e-02 6.54636860e-01
8.04871500e-01 7.67773807e-01 2.34964192e-01 -5.49658477e-01
1.42928374e+00 -7.45436132e-01 -1.29244041e+00 -1.99153200e-01
7.58230329e-01 -1.35774767e+00 9.47490692e-01 2.15948135e-01
-9.54417348e-01 -7.53488839e-01 -1.02891302e+00 3.36740762e-01
-2.71757096e-01 5.39700091e-02 -2.29060021e-03 1.51812088e+00
-1.23595643e+00 7.30939209e-02 -7.46941209e-01 -4.57527161e-01
-1.84525937e-01 4.67491060e-01 -2.76401192e-01 -1.10828809e-01
-1.29720175e+00 1.26534581e+00 -2.96257362e-02 3.94896828e-02
-5.68062246e-01 -2.95944154e-01 -8.31584156e-01 -4.61545251e-02
-4.24736366e-02 -4.31714877e-02 1.51698852e+00 -4.29554254e-01
-1.79015255e+00 5.24732947e-01 -1.07107782e+00 -3.36016625e-01
1.79039299e-01 4.26714234e-02 -8.13745439e-01 -1.44326925e-01
-2.14147344e-01 4.69198644e-01 1.04246914e+00 -1.02918172e+00
-6.58954561e-01 -3.67561549e-01 -4.99173015e-01 2.77061820e-01
-3.16786855e-01 7.34670520e-01 -3.80269378e-01 -4.66189355e-01
2.31554985e-01 -1.00748539e+00 1.63258553e-01 -7.13388562e-01
-4.61006850e-01 -3.97052526e-01 1.02717745e+00 -9.88625109e-01
1.57896411e+00 -2.52157378e+00 -2.36012146e-01 1.67699054e-01
-4.89634126e-01 4.98151541e-01 1.16390146e-01 5.04202306e-01
-1.36076003e-01 3.48781526e-01 3.10405008e-02 -7.47578204e-01
5.55047244e-02 2.08783761e-01 -7.79706836e-02 3.51301521e-01
-7.27389157e-02 5.36517262e-01 -1.63456082e-01 -7.11447716e-01
3.43904167e-01 7.01218545e-01 -3.05832863e-01 8.29542279e-02
5.24635673e-01 5.41551352e-01 5.74305542e-02 6.64246500e-01
8.32302749e-01 2.89906442e-01 1.30249873e-01 1.52671546e-01
-3.88483614e-01 7.96939135e-01 -1.28843856e+00 1.28621602e+00
-3.42015058e-01 8.53029430e-01 2.58581400e-01 -7.90647805e-01
7.68638194e-01 8.87781084e-01 -1.28449395e-01 -5.52630365e-01
-2.46963166e-02 2.93301374e-01 4.57090765e-01 -5.08440018e-01
5.93782544e-01 -5.37600815e-01 2.39109844e-01 2.12528855e-01
-3.11572861e-04 -1.06474541e-01 1.82208836e-01 -5.82341198e-03
5.85422039e-01 -2.92725176e-01 1.38871998e-01 -1.40544221e-01
7.91042805e-01 -2.60428280e-01 6.54709041e-01 8.12331319e-01
-5.74793577e-01 7.36293733e-01 -2.87327543e-03 2.14444354e-01
-5.63425183e-01 -7.45234370e-01 -4.24960941e-01 1.26088107e+00
-1.54100552e-01 -4.37161595e-01 -9.33501303e-01 -4.83992994e-01
-2.53613770e-01 1.10161006e+00 -2.27908865e-02 1.54212356e-01
-6.23215675e-01 -7.74157941e-01 8.63707781e-01 3.67050856e-01
2.61640549e-01 -7.85922647e-01 1.88797023e-02 4.51797426e-01
-5.70401967e-01 -1.12398684e+00 -7.62920141e-01 2.53919572e-01
-7.38825679e-01 -4.62908804e-01 -9.26095009e-01 -1.05311346e+00
2.33608112e-01 4.11054075e-01 5.23365140e-01 2.74732001e-02
1.96546718e-01 2.35788986e-01 -4.47925985e-01 -4.65358406e-01
-7.86860526e-01 1.45013824e-01 3.21152389e-01 5.20803183e-02
7.24576831e-01 -6.98185265e-02 -2.49050841e-01 7.27996111e-01
-4.90999043e-01 -5.10220408e-01 2.48552144e-01 9.90621448e-01
-7.56303268e-03 5.60420193e-02 7.02657700e-01 -3.38602006e-01
5.32711506e-01 -6.99916407e-02 -1.91256344e-01 2.25043282e-01
-5.01939952e-01 -6.44316152e-02 2.09581733e-01 -6.38111353e-01
-1.16433918e+00 4.77286242e-02 -1.04400539e+00 1.71284226e-03
-5.21807075e-01 4.95083481e-01 -3.75378311e-01 -1.76529109e-01
5.39326489e-01 5.91906428e-01 1.58391312e-01 -7.04790413e-01
-2.44937316e-01 1.22933090e+00 7.93416724e-02 -1.49658591e-01
7.57880032e-01 -6.54700398e-02 -7.66277552e-01 -1.46699631e+00
-2.06395820e-01 -8.51492047e-01 -5.19197822e-01 -5.12318797e-02
8.14810395e-01 -1.00053310e+00 -6.17914438e-01 8.17734718e-01
-1.20212805e+00 -1.97845101e-02 1.97705552e-01 9.70470965e-01
-1.64645061e-01 7.29824007e-01 -6.62135720e-01 -1.30681479e+00
-1.07704192e-01 -1.74208653e+00 9.21658039e-01 2.31220677e-01
-4.61244941e-01 -8.57594073e-01 -1.21845476e-01 6.18017375e-01
7.14171290e-01 -8.51616740e-01 8.01924706e-01 -1.01330531e+00
-5.75811900e-02 -2.66250551e-01 2.41343141e-01 5.30475259e-01
2.45879546e-01 1.40839502e-01 -1.37567413e+00 -4.17571694e-01
2.62508452e-01 1.70683056e-01 5.07853866e-01 5.78096509e-01
6.18799031e-01 -2.02017114e-01 -2.70848244e-01 1.55488014e-01
8.85988593e-01 7.10284412e-01 7.03636527e-01 4.35431451e-02
3.24561983e-01 5.35998285e-01 4.51144546e-01 9.01707858e-02
6.19020104e-01 1.03631985e+00 -2.21327052e-01 -6.07939623e-03
-9.79896188e-02 -1.22014649e-01 8.81312847e-01 1.25375175e+00
2.17020571e-01 -3.69995296e-01 -9.94083881e-01 4.97528523e-01
-1.25602901e+00 -1.14613223e+00 -2.34774858e-01 2.43999624e+00
8.56944799e-01 1.75449461e-01 3.78557533e-01 7.94244707e-01
1.01912129e+00 3.81423570e-02 -5.13006151e-02 -5.47221780e-01
-3.11360210e-01 1.74400046e-01 1.65660337e-01 8.60814333e-01
-9.49136257e-01 7.74841130e-01 6.98796844e+00 9.02168572e-01
-1.35357940e+00 3.06291014e-01 4.12386954e-01 1.04863361e-01
1.10360831e-02 4.87786941e-02 -1.52953148e+00 6.06429517e-01
1.50016725e+00 4.39399853e-02 4.31687571e-02 6.21608615e-01
4.20633376e-01 -2.66681761e-01 -9.66086209e-01 1.20238006e+00
3.07166100e-01 -8.54974747e-01 -2.25636795e-01 1.37599483e-01
2.10288167e-01 -2.29115769e-01 1.00781515e-01 5.15716612e-01
-3.15640301e-01 -7.51469791e-01 9.11642253e-01 -1.20422743e-01
5.59391081e-01 -5.69782495e-01 8.44565451e-01 3.52180362e-01
-1.09384787e+00 2.17528045e-02 -3.75162996e-02 5.67175634e-02
4.71188694e-01 2.01687619e-01 -1.08733714e+00 3.69998246e-01
4.76967931e-01 1.14941865e-01 -4.98027742e-01 1.23824680e+00
-5.87011650e-02 1.30575490e+00 -3.04541439e-01 -1.37792140e-01
-7.66329095e-02 1.86520427e-01 6.82345212e-01 1.58181369e+00
3.78066838e-01 -1.79246634e-01 -5.26591428e-02 1.07826546e-01
4.46781933e-01 3.12846303e-01 -3.38785738e-01 6.48666099e-02
6.50598466e-01 7.78611779e-01 -4.23412174e-01 -3.11152726e-01
-5.34939229e-01 8.03416073e-01 -2.26419464e-01 5.70712626e-01
-7.99347162e-01 -3.67459655e-01 7.47107506e-01 5.72281145e-02
3.39114159e-01 -5.15519738e-01 -1.66275293e-01 -1.04518199e+00
-9.31141227e-02 -1.17911458e+00 1.89953417e-01 -3.98675174e-01
-9.80987191e-01 6.08792424e-01 4.05064188e-02 -1.03568041e+00
-5.18042982e-01 -4.25112933e-01 -6.79110527e-01 1.18713975e+00
-1.58759117e+00 -8.06473970e-01 2.98420787e-01 7.25303829e-01
9.09160495e-01 -1.12552673e-01 1.03425610e+00 6.81717336e-01
-8.60469460e-01 1.12562048e+00 1.76460430e-01 1.16188452e-01
1.01716459e+00 -1.00459158e+00 5.07709324e-01 9.36034381e-01
3.45110655e-01 8.81885171e-01 8.97315443e-01 -3.38272423e-01
-1.30420017e+00 -4.24645633e-01 1.40659440e+00 -2.33240023e-01
1.95386469e-01 -4.42513555e-01 -1.11873853e+00 5.15852273e-01
3.58036548e-01 -6.56223178e-01 9.96130168e-01 3.83098155e-01
-2.36077547e-01 5.84695749e-02 -1.06355035e+00 4.89275217e-01
5.37890792e-01 -8.18323195e-01 -7.38324165e-01 2.30353341e-01
5.44325829e-01 -3.09329450e-01 -6.16880894e-01 1.53473184e-01
5.54632723e-01 -8.90999377e-01 8.54688883e-01 -2.71692365e-01
-4.08370286e-01 -3.11221391e-01 -3.53510171e-01 -1.26038945e+00
-1.95460364e-01 -6.96895778e-01 2.38259554e-01 1.65274966e+00
9.22786355e-01 -9.42024291e-01 7.14051783e-01 6.59167528e-01
-3.65099832e-02 -1.59811482e-01 -1.38795328e+00 -1.08833516e+00
-3.81591171e-02 -7.31358051e-01 5.08171737e-01 7.51658618e-01
1.34800121e-01 4.32236671e-01 -3.67591262e-01 2.31953025e-01
3.95776331e-01 -3.30291122e-01 6.58405304e-01 -9.04554963e-01
-7.37911686e-02 -3.80130678e-01 -4.05391604e-01 -1.16032004e+00
1.95230603e-01 -5.45143545e-01 2.09979355e-01 -1.14616442e+00
-6.29170015e-02 -3.93601090e-01 -1.59259200e-01 3.29875857e-01
-5.47971606e-01 1.41958818e-02 3.60253662e-01 1.16613261e-01
3.96461599e-02 3.35945278e-01 6.68255091e-01 -2.04925105e-01
-4.77186859e-01 4.75232273e-01 -4.22688603e-01 3.58084649e-01
9.24376726e-01 -4.79619771e-01 -5.17795458e-02 -1.93041965e-01
-6.56500340e-01 2.59581864e-01 -7.50479847e-02 -9.66785491e-01
4.76293504e-01 1.16465099e-01 2.57125109e-01 -7.63725758e-01
7.24339545e-01 -5.15376329e-01 7.56311119e-02 4.30353492e-01
-1.52242720e-01 1.91686869e-01 4.82193440e-01 2.15205371e-01
-4.49126989e-01 -4.87377584e-01 8.75875354e-01 3.15238237e-01
-4.14555371e-01 -1.66718364e-01 -9.91467237e-01 -2.18817353e-01
5.82448423e-01 -4.70941603e-01 6.34184554e-02 -7.23829508e-01
-7.11986005e-01 -1.82573467e-01 1.42909423e-01 5.03594816e-01
5.28242171e-01 -9.86487448e-01 -9.50682044e-01 4.93250728e-01
-4.81319129e-02 -7.25286186e-01 3.74637634e-01 1.08207595e+00
-1.67630628e-01 7.59167135e-01 2.21797094e-01 -9.24584270e-01
-1.98522973e+00 2.46149987e-01 2.06249729e-01 3.38115506e-02
-4.86511812e-02 9.85326946e-01 -5.90977147e-02 -3.51508439e-01
6.29736900e-01 -1.29617944e-01 -1.85481265e-01 2.48386905e-01
7.11902261e-01 2.69076645e-01 3.94721299e-01 -1.14382398e+00
-6.38754308e-01 4.11841631e-01 -4.68816876e-01 -6.59877956e-01
7.45130479e-01 -3.86512309e-01 3.11612368e-01 6.02852941e-01
1.15001404e+00 2.55781174e-01 -6.04095578e-01 -9.96672064e-02
1.67424425e-01 -3.29949766e-01 1.17172271e-01 -5.94543338e-01
-6.42441750e-01 1.09348607e+00 7.70318031e-01 3.27155948e-01
8.71114612e-01 -4.57791120e-01 6.06269598e-01 8.24660957e-02
2.72240043e-01 -9.85371053e-01 -5.08735240e-01 6.79205239e-01
6.49859309e-01 -1.28773057e+00 -3.45036805e-01 -4.11343992e-01
-7.57897496e-01 8.84370863e-01 3.38988692e-01 8.46511900e-01
8.51065993e-01 1.93052948e-01 4.89050657e-01 4.74966645e-01
-4.78220552e-01 -1.76480487e-01 9.87434387e-02 5.01497865e-01
7.57893860e-01 4.58987653e-02 -3.65587741e-01 2.84600228e-01
-4.12557960e-01 -4.33213919e-01 4.15119946e-01 9.70218837e-01
-3.75079364e-01 -1.53666759e+00 -9.99005616e-01 2.34868616e-01
-6.64863348e-01 -2.01831803e-01 -3.68174493e-01 7.28858232e-01
-3.38490278e-01 1.76156783e+00 -2.75657862e-01 -5.76837897e-01
2.44920090e-01 8.83698285e-01 1.84688896e-01 -5.68597615e-01
-8.30003977e-01 3.87733102e-01 2.36079305e-01 -8.13618377e-02
-3.24611306e-01 -1.04034746e+00 -9.60944235e-01 -4.36330676e-01
-1.08358562e+00 5.42151332e-01 1.27909517e+00 1.03481793e+00
1.77149281e-01 1.84804693e-01 7.61079192e-01 -7.14619815e-01
-7.70297468e-01 -1.35377228e+00 -5.03997445e-01 2.33937539e-02
4.93572474e-01 -4.92470711e-01 -7.38762975e-01 5.29958345e-02] | [14.343689918518066, 6.196404457092285] |
a04db6a2-fe8f-4508-9325-4352a0a282c0 | active-learning-for-transition-state | 2108.04698 | null | https://arxiv.org/abs/2108.04698v2 | https://arxiv.org/pdf/2108.04698v2.pdf | Active Learning for Saddle Point Calculation | The saddle point (SP) calculation is a grand challenge for computationally intensive energy function in computational chemistry area, where the saddle point may represent the transition state (TS). The traditional methods need to evaluate the gradients of the energy function at a very large number of locations. To reduce the number of expensive computations of the true gradients, we propose an active learning framework consisting of a statistical surrogate model, Gaussian process regression (GPR) for the energy function, and a single-walker dynamics method, gentle accent dynamics (GAD), for the saddle-type transition states. SP is detected by the GAD applied to the GPR surrogate for the gradient vector and the Hessian matrix. Our key ingredient for efficiency improvements is an active learning method which sequentially designs the most informative locations and takes evaluations of the original model at these locations to train GPR. We formulate this active learning task as the optimal experimental design problem and propose a very efficient sample-based sub-optimal criterion to construct the optimal locations. We show that the new method significantly decreases the required number of energy or force evaluations of the original model. | ['Xiang Zhou', 'Hongqiao Wang', 'Shuting Gu'] | 2021-08-10 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 1.32061794e-01 -2.52305437e-02 -7.14472532e-02 1.09402779e-02
-1.45700216e+00 -5.07360518e-01 4.20757473e-01 4.95400339e-01
-6.98335588e-01 8.76923025e-01 -2.18484655e-01 -4.05555338e-01
-7.37355277e-02 -6.34108543e-01 -8.96927357e-01 -1.47639310e+00
-1.33441299e-01 5.03555715e-01 3.20024192e-01 1.22215319e-02
7.62070417e-01 6.67063415e-01 -8.90991926e-01 -4.14460629e-01
1.22665298e+00 7.72490203e-01 3.04298371e-01 7.24021435e-01
9.19392407e-02 2.67823130e-01 -5.83824441e-02 1.36954814e-01
1.28027305e-01 -5.76715469e-01 -6.00848138e-01 -5.80705583e-01
-9.61802825e-02 1.76387787e-01 1.91097781e-01 1.05706728e+00
9.81362581e-01 7.42834270e-01 1.07126641e+00 -8.22485805e-01
3.21502751e-03 7.06788525e-02 -6.27439022e-01 8.23674351e-02
-7.88414106e-02 3.37223172e-01 8.82747948e-01 -1.37730110e+00
7.61643589e-01 9.33842659e-01 7.81517506e-01 1.56337395e-01
-1.60009515e+00 -4.69557315e-01 1.46424947e-02 1.79696664e-01
-1.46610916e+00 -5.77601075e-01 9.99115109e-01 -5.29956877e-01
1.04912531e+00 6.50112182e-02 6.86683536e-01 7.73986757e-01
5.31416297e-01 4.44909692e-01 8.44029129e-01 -5.72011948e-01
1.05580604e+00 -1.52639478e-01 1.45711288e-01 8.34991336e-01
2.09278345e-01 2.71695942e-01 -5.89702487e-01 -8.04423273e-01
4.96443450e-01 -4.49388362e-02 -9.96555686e-02 -8.83624673e-01
-6.20686710e-01 1.10294843e+00 1.46772370e-01 -3.63783419e-01
-5.82837880e-01 1.05172023e-01 5.90583757e-02 -2.89395481e-01
5.79082191e-01 9.09119189e-01 -6.47937715e-01 -3.57982218e-01
-8.66734564e-01 4.09409493e-01 8.70721757e-01 2.80082375e-01
8.92647028e-01 -2.89750159e-01 -2.93971598e-02 4.06297654e-01
7.65043139e-01 2.13735044e-01 -1.18956596e-01 -7.85683215e-01
3.83555442e-01 7.40516186e-01 6.02187574e-01 -6.45898938e-01
-2.48978704e-01 -2.43565127e-01 -5.46891630e-01 4.27794576e-01
4.94304985e-01 -5.91632903e-01 -6.01975620e-01 1.51709819e+00
9.42284107e-01 6.83525205e-02 -1.65988013e-01 5.65796971e-01
2.90400684e-01 1.15229547e+00 3.39054316e-01 -8.31625700e-01
5.31658232e-01 -9.51179624e-01 -3.46999377e-01 -4.71564643e-02
8.41141582e-01 -5.43109179e-01 9.42409158e-01 2.04721749e-01
-1.41055489e+00 -2.42294341e-01 -1.15165114e+00 5.97343855e-02
-3.67208004e-01 1.58559471e-01 4.76502329e-01 2.03945473e-01
-8.18070114e-01 1.14813673e+00 -1.32400846e+00 2.93235421e-01
2.87860483e-01 7.28348672e-01 1.01760782e-01 6.38638139e-01
-8.84281695e-01 9.08555627e-01 3.03062022e-01 4.24411595e-01
-1.04634166e+00 -1.07248700e+00 -6.90445900e-01 -7.27481470e-02
4.16689754e-01 -3.53106499e-01 9.01865125e-01 -7.14453816e-01
-1.61990750e+00 3.97495300e-01 -5.07849932e-01 -2.04890460e-01
6.39556348e-01 -1.05626673e-01 2.18491971e-01 5.52844033e-02
-2.29515702e-01 4.12732184e-01 8.90640974e-01 -1.13905585e+00
-9.40735266e-02 -3.87165397e-01 -5.97526073e-01 2.55950779e-01
8.03842470e-02 -9.83230099e-02 3.51564330e-03 -2.11580783e-01
3.28666240e-01 -1.02118635e+00 -7.25417376e-01 -8.48459750e-02
-5.36593139e-01 -4.98400152e-01 6.88622832e-01 -6.67709351e-01
1.23225915e+00 -1.82839692e+00 4.48107630e-01 4.85626966e-01
3.90625060e-01 1.38074800e-01 2.74700493e-01 5.65020859e-01
-1.50502861e-01 1.75362285e-02 -4.33579862e-01 -3.35580498e-01
-2.46826768e-01 -2.23085523e-01 -1.14803396e-01 6.93465650e-01
3.59624535e-01 9.66470718e-01 -1.13226032e+00 -4.86963958e-01
3.82382162e-02 6.22946620e-01 -5.12809634e-01 3.46851468e-01
-1.79930240e-01 6.75518215e-01 -5.94540179e-01 1.12193994e-01
7.00827062e-01 -3.57200265e-01 2.63803005e-01 -2.18921944e-01
-8.45830381e-01 3.28844339e-01 -1.33706558e+00 1.51223779e+00
-4.48887572e-02 2.74075009e-02 3.44402827e-02 -1.22977173e+00
1.01937222e+00 1.58452168e-01 8.50753248e-01 -3.16609293e-01
-1.54117867e-01 3.64225686e-01 -3.76314409e-02 -3.14621717e-01
1.84559107e-01 -1.13155849e-01 1.72034606e-01 2.50634402e-01
-2.04963982e-01 -1.01388827e-01 5.46503402e-02 4.37880531e-02
9.40517068e-01 4.88469630e-01 4.12809312e-01 -5.91658175e-01
5.32015383e-01 1.05300605e-01 5.68844914e-01 6.61069691e-01
-1.67020857e-01 2.06378952e-01 6.84177220e-01 -3.52781236e-01
-1.16152906e+00 -9.24666524e-01 6.18710369e-02 8.42981517e-01
6.61543384e-02 -4.81726408e-01 -7.34918773e-01 -4.49248195e-01
-1.92518219e-01 4.95073289e-01 -5.27239680e-01 -4.98888761e-01
-8.54730189e-01 -1.24182796e+00 -1.94670469e-01 3.29604834e-01
2.76720703e-01 -7.84562528e-01 -5.22325873e-01 4.59757626e-01
2.78431833e-01 -1.84118390e-01 -5.24558723e-01 7.19928682e-01
-1.00988877e+00 -1.05912685e+00 -5.05539000e-01 -6.87229812e-01
8.45920801e-01 -2.22514689e-01 6.74470663e-01 -2.97353089e-01
-1.69083759e-01 -1.60526663e-01 1.39760092e-01 -2.35037223e-01
-3.59852016e-01 2.86113890e-03 -1.28084242e-01 -1.01464987e-01
-7.31698965e-05 -4.82935369e-01 -9.33400035e-01 1.05831049e-01
-1.60490915e-01 5.80959320e-02 4.10622984e-01 8.06748390e-01
1.09447789e+00 -6.36143237e-02 3.65029365e-01 -5.72879195e-01
6.04043126e-01 -4.54347372e-01 -1.02212369e+00 2.55890965e-01
-7.92189300e-01 5.82394719e-01 5.57323813e-01 -4.82402563e-01
-9.40827250e-01 5.86864531e-01 -7.11590201e-02 -1.22142397e-01
5.20401657e-01 5.56780636e-01 -2.19039813e-01 -4.16456133e-01
6.70710623e-01 3.27103019e-01 4.71614581e-03 -4.71479148e-01
4.43644226e-02 2.26007044e-01 1.29799470e-01 -9.46850061e-01
5.77154458e-01 4.78135645e-02 5.51599503e-01 -9.15677249e-01
-6.81490898e-01 -6.63853049e-01 -7.52505660e-01 -6.20077811e-02
6.81146443e-01 -6.03267014e-01 -1.09528542e+00 2.04509631e-01
-1.15385056e+00 -5.32331586e-01 -3.07051420e-01 4.14623678e-01
-5.89760125e-01 4.41708207e-01 -3.51939976e-01 -1.22508955e+00
-6.54156625e-01 -1.46065784e+00 1.12666357e+00 1.61778763e-01
-2.90067494e-01 -1.17987478e+00 7.59105921e-01 2.72704698e-02
1.90624624e-01 4.71475512e-01 1.17858613e+00 -4.38976675e-01
-4.83143717e-01 -3.11683297e-01 3.89950424e-01 4.38830927e-02
-1.78780392e-01 3.92716378e-01 -6.19230568e-01 -4.13393915e-01
1.13218538e-01 -2.41141450e-02 7.39899993e-01 8.47412229e-01
1.07518232e+00 -2.19566926e-01 -5.67857146e-01 4.93339062e-01
1.50735688e+00 6.80192947e-01 5.41733503e-01 -5.36454134e-02
7.03887582e-01 2.42856741e-01 3.42740834e-01 3.88992578e-01
-6.16632439e-02 6.26385629e-01 -2.03811064e-01 -2.82522023e-01
4.51401889e-01 -4.77789670e-01 6.65030479e-01 7.46603310e-01
-4.15390700e-01 9.34420079e-02 -1.18125582e+00 -3.30258943e-02
-1.97017646e+00 -8.84302497e-01 -2.19964758e-01 2.70443916e+00
1.24786305e+00 2.11121723e-01 1.44934520e-01 1.60074860e-01
5.89773595e-01 -9.29068699e-02 -1.09656835e+00 -2.78666824e-01
2.36468539e-01 4.55662221e-01 4.82279241e-01 9.09499586e-01
-1.11856854e+00 7.18478560e-01 7.25583029e+00 1.05841529e+00
-1.14811432e+00 1.46169022e-01 7.27306128e-01 1.63760513e-01
-3.37219462e-02 5.04729748e-01 -1.15928113e+00 6.55305803e-01
1.07760608e+00 -1.16068879e-02 8.95285681e-02 9.40239489e-01
1.06210768e+00 -3.14823210e-01 -8.89821172e-01 8.46538782e-01
-5.44176996e-01 -1.64103889e+00 -3.46453130e-01 2.21791402e-01
7.71650553e-01 -3.42750102e-01 -1.73853502e-01 -1.53730961e-03
2.58421063e-01 -7.87004173e-01 6.12794936e-01 7.79901206e-01
4.57981646e-01 -7.89919794e-01 1.03317432e-01 4.60468233e-01
-1.22560525e+00 2.19766036e-01 -3.51210862e-01 1.08294658e-01
9.67108458e-02 4.66009051e-01 -5.87226450e-01 -1.07656166e-01
4.30884175e-02 5.78632772e-01 -3.28489780e-01 1.15305531e+00
5.57516776e-02 8.11311960e-01 -5.50892532e-01 -4.60149109e-01
1.42637596e-01 -9.60357368e-01 7.59204686e-01 7.61160195e-01
1.33822009e-01 -1.28136396e-01 3.21143925e-01 1.21671343e+00
2.53716290e-01 1.75343573e-01 -1.60549507e-01 -3.44646186e-01
6.48423731e-01 1.12786591e+00 -9.85930502e-01 -9.12350491e-02
1.92415230e-02 6.65364146e-01 5.48402131e-01 5.91340303e-01
-6.54091120e-01 -3.89650881e-01 3.56467515e-01 2.72935331e-01
8.36731419e-02 -6.19624138e-01 -1.90812632e-01 -8.11089337e-01
-1.73286095e-01 -3.36106271e-01 1.47766173e-01 -5.95474541e-01
-8.71755898e-01 -2.39903867e-01 -2.01330590e-03 -7.31890678e-01
-1.20059527e-01 -4.55768198e-01 -9.92583394e-01 1.08463240e+00
-1.14340341e+00 -4.85447884e-01 1.57761857e-01 3.32106322e-01
3.97662729e-01 2.00220227e-01 7.09989429e-01 -1.61637517e-03
-1.00052869e+00 3.91116530e-01 7.66760468e-01 -2.37750009e-01
3.83597702e-01 -1.37680840e+00 2.66330332e-01 8.29821825e-01
-4.60242897e-01 8.37720692e-01 9.75520432e-01 -1.03702831e+00
-1.81952441e+00 -6.12914920e-01 6.49976254e-01 -3.72114807e-01
7.11121857e-01 -5.13489723e-01 -7.60294795e-01 1.74878910e-01
-4.09593552e-01 -1.38824955e-01 4.87006307e-01 -1.63102239e-01
3.91705990e-01 1.07461408e-01 -1.09758091e+00 6.61704659e-01
7.67372370e-01 -2.31869251e-01 -9.46999788e-02 5.71246982e-01
5.56951582e-01 -2.62684822e-01 -7.46285677e-01 1.67274371e-01
2.40596071e-01 -3.14336151e-01 1.17787254e+00 -8.87864292e-01
2.44679496e-01 -3.00934762e-01 2.00189531e-01 -1.10208440e+00
-2.42750749e-01 -1.29258680e+00 -7.00795531e-01 7.99639642e-01
8.34517717e-01 -6.44557714e-01 1.01937032e+00 9.28926647e-01
-2.57917494e-01 -1.36333203e+00 -1.00070465e+00 -4.79009569e-01
2.97550261e-01 7.89641887e-02 -2.39465870e-02 6.09271169e-01
-9.52384919e-02 5.80971718e-01 -1.97451655e-02 8.55251215e-04
9.83146966e-01 2.03070068e-03 3.86642396e-01 -1.34684134e+00
-2.99731880e-01 -1.84844449e-01 6.73523620e-02 -1.06711376e+00
-1.23573549e-01 -5.17737627e-01 1.95296928e-01 -1.39431226e+00
2.01321483e-01 -4.97887284e-01 3.09848022e-02 1.63636848e-01
-3.90253216e-01 -6.45579755e-01 -4.08202440e-01 3.63389194e-01
-4.95852411e-01 8.84906709e-01 1.27053380e+00 1.18462801e-01
-7.49716103e-01 1.15269981e-01 -5.56659758e-01 6.67028666e-01
5.10402739e-01 -6.50459766e-01 -3.19040328e-01 4.75566059e-01
6.88045144e-01 8.31293240e-02 3.20651472e-01 -6.50872946e-01
5.42596579e-01 -3.89906496e-01 5.05209565e-01 -8.68844807e-01
5.03209293e-01 -3.89485866e-01 2.00875416e-01 5.82587004e-01
-5.15993953e-01 -2.08203844e-03 -2.31548816e-01 7.08168685e-01
2.72785991e-01 -4.62452769e-01 1.05135810e+00 -9.19517353e-02
-2.23474696e-01 5.08193254e-01 -3.73399645e-01 9.80450362e-02
9.28554177e-01 -3.64402920e-01 1.23671696e-01 1.51079372e-01
-9.06023562e-01 1.09568551e-01 3.26950639e-01 -3.43522906e-01
4.64971960e-01 -1.04760575e+00 -2.59899676e-01 4.58440371e-02
-4.28013712e-01 3.23574036e-01 2.24952921e-02 1.15757453e+00
-6.84143484e-01 1.14794597e-01 4.27534640e-01 -3.83561730e-01
-9.42603409e-01 2.29851589e-01 7.92316914e-01 -4.84718382e-01
-5.16411185e-01 7.96456873e-01 -2.93283015e-01 -2.26281494e-01
-2.29239110e-02 -1.52562350e-01 2.26428434e-02 2.20803276e-01
2.56360143e-01 7.93637693e-01 1.61107078e-01 -4.09151822e-01
-2.78552860e-01 6.78394556e-01 -1.77815668e-02 -3.86813805e-02
1.62400341e+00 1.49646267e-01 -5.18154688e-02 4.41204846e-01
1.41467345e+00 -6.97956681e-02 -1.77529395e+00 6.92740679e-02
1.54971942e-01 9.87832844e-02 5.01863062e-01 -3.75682861e-01
-4.33854997e-01 7.11908877e-01 8.30004632e-01 -1.42011672e-01
5.07284701e-01 -2.87316859e-01 5.52086830e-01 5.82487941e-01
3.69548388e-02 -1.48713839e+00 9.13480520e-02 4.62062389e-01
5.18503010e-01 -1.10426092e+00 1.79976016e-01 -3.00509334e-01
-3.31457555e-01 1.13018918e+00 2.51401603e-01 -2.72604972e-01
8.97638321e-01 2.82482207e-01 -5.29944122e-01 -4.49566633e-01
-5.84330022e-01 2.21035331e-01 2.70519465e-01 1.09545469e-01
3.36064309e-01 -1.08410127e-01 -4.98091400e-01 3.49044859e-01
2.22061649e-01 -2.71855831e-01 -6.13364466e-02 1.22196209e+00
-6.72662020e-01 -1.16664350e+00 -1.18390024e-01 1.83405235e-01
-5.70229776e-02 -1.67826071e-01 -4.07148033e-01 4.13004458e-01
-2.36123770e-01 6.64281070e-01 -1.07572936e-02 1.40117154e-01
8.69670510e-02 3.61751825e-01 4.34490323e-01 -4.45806593e-01
-4.21273202e-01 1.61895722e-01 -7.48941898e-02 -4.46187645e-01
-2.24709719e-01 -6.93082213e-01 -1.43432450e+00 -2.67926723e-01
-5.80121160e-01 6.47497058e-01 6.07732117e-01 1.15874028e+00
4.83868122e-01 -1.63666710e-01 8.37719440e-01 -1.01850140e+00
-8.27191532e-01 -6.69776559e-01 -2.72261560e-01 1.66432053e-01
2.16328472e-01 -9.71551299e-01 -6.13797784e-01 -1.81970581e-01] | [5.339611530303955, 5.050850868225098] |
36c77a97-8248-4725-9019-e6dcad798fb7 | predicting-tasks-in-goal-oriented-spoken | null | null | https://aclanthology.org/W13-4038 | https://aclanthology.org/W13-4038.pdf | Predicting Tasks in Goal-Oriented Spoken Dialog Systems using Semantic Knowledge Bases | null | ['er', 'Aasish Pappu', 'Alex Rudnicky'] | 2013-08-01 | null | null | null | ws-2013-8 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.172306537628174, 3.7500662803649902] |
1227d887-dcc2-4960-9918-d33253b7bfab | a-framework-of-meta-functional-learning-for | 2203.14840 | null | https://arxiv.org/abs/2203.14840v1 | https://arxiv.org/pdf/2203.14840v1.pdf | A Framework of Meta Functional Learning for Regularising Knowledge Transfer | Machine learning classifiers' capability is largely dependent on the scale of available training data and limited by the model overfitting in data-scarce learning tasks. To address this problem, this work proposes a novel framework of Meta Functional Learning (MFL) by meta-learning a generalisable functional model from data-rich tasks whilst simultaneously regularising knowledge transfer to data-scarce tasks. The MFL computes meta-knowledge on functional regularisation generalisable to different learning tasks by which functional training on limited labelled data promotes more discriminative functions to be learned. Based on this framework, we formulate three variants of MFL: MFL with Prototypes (MFL-P) which learns a functional by auxiliary prototypes, Composite MFL (ComMFL) that transfers knowledge from both functional space and representational space, and MFL with Iterative Updates (MFL-IU) which improves knowledge transfer regularisation from MFL by progressively learning the functional regularisation in knowledge transfer. Moreover, we generalise these variants for knowledge transfer regularisation from binary classifiers to multi-class classifiers. Extensive experiments on two few-shot learning scenarios, Few-Shot Learning (FSL) and Cross-Domain Few-Shot Learning (CD-FSL), show that meta functional learning for knowledge transfer regularisation can improve FSL classifiers. | ['Shaogang Gong', 'Yanwei Fu', 'Pan Li'] | 2022-03-28 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 5.58744848e-01 1.57340795e-01 -4.04209644e-01 -4.14234042e-01
-6.78740382e-01 -2.12750807e-01 8.72311175e-01 2.16354772e-01
-6.05284512e-01 1.05158961e+00 1.21946938e-01 2.04180539e-01
-7.08037376e-01 -1.00573540e+00 -8.30596685e-01 -6.61089599e-01
6.88328058e-04 2.21263900e-01 5.97702682e-01 -3.44263583e-01
5.76050133e-02 3.32135528e-01 -1.98350799e+00 7.80372500e-01
7.97007143e-01 1.07924271e+00 4.30375606e-01 5.04211009e-01
-3.70684832e-01 1.00834537e+00 -4.11836356e-01 -2.27791131e-01
1.86933577e-01 -4.16825354e-01 -1.14171898e+00 1.24987923e-02
6.71687201e-02 2.67686844e-01 3.87341902e-02 7.45330036e-01
4.97973412e-01 8.35607052e-01 9.90093529e-01 -1.34999287e+00
-7.97472537e-01 3.00771475e-01 -8.64935219e-02 3.54052871e-01
2.31786057e-01 1.10289581e-01 7.80940652e-01 -1.01094532e+00
7.02869236e-01 1.38354528e+00 9.61575925e-01 8.06907415e-01
-1.25770175e+00 -1.54065415e-01 9.47620794e-02 6.05419934e-01
-1.02484024e+00 -3.73089224e-01 4.32257503e-01 -6.40538752e-01
1.25459051e+00 7.62213301e-03 4.85040396e-01 1.19909060e+00
-1.36173544e-02 8.61088455e-01 1.30194914e+00 -7.89055645e-01
6.80531442e-01 1.98267356e-01 4.08822030e-01 4.34884936e-01
-4.48498540e-02 3.82493794e-01 -7.83958375e-01 -1.89570323e-01
4.17641729e-01 3.02934706e-01 -2.84965634e-01 -4.24960136e-01
-9.34174061e-01 9.85615432e-01 2.43097842e-01 5.77542245e-01
-4.19348717e-01 -7.66273960e-02 7.64214218e-01 8.66172016e-01
7.35655844e-01 3.57051134e-01 -7.67316282e-01 -1.32100239e-01
-6.59056425e-01 -1.52013645e-01 8.35399091e-01 1.01030707e+00
1.39726281e+00 1.68881685e-01 -5.62913299e-01 1.20311797e+00
-1.54551029e-01 1.31904691e-01 1.27270973e+00 -5.86257756e-01
5.60957007e-02 6.90940917e-01 -1.08197384e-01 -3.76601741e-02
-2.48374656e-01 -1.88193023e-01 -6.59632206e-01 1.41347587e-01
-1.36221334e-01 -1.47349373e-01 -1.12436652e+00 1.56526983e+00
3.42588335e-01 4.56195503e-01 2.25329518e-01 3.87680411e-01
9.40328062e-01 7.73808300e-01 3.97581905e-01 -5.17944932e-01
1.07997704e+00 -1.01863456e+00 -5.23990095e-01 -1.24376737e-01
9.83595729e-01 -2.07989901e-01 1.21421874e+00 2.31169537e-01
-7.13939130e-01 -1.11147761e+00 -1.10527372e+00 2.42635086e-01
-9.91368175e-01 -4.30201888e-01 4.77578610e-01 5.87473214e-01
-1.05795348e+00 9.19854820e-01 -5.48811376e-01 -6.20956600e-01
6.34709418e-01 2.34656587e-01 -6.18917406e-01 -2.99359292e-01
-1.59911370e+00 1.31477654e+00 1.20621324e+00 -3.89729500e-01
-9.91503119e-01 -1.15460014e+00 -1.07978547e+00 -2.12744940e-02
5.06179631e-01 -6.73054039e-01 1.23902154e+00 -1.27501023e+00
-1.61240864e+00 8.63740861e-01 2.42471442e-01 -5.49068987e-01
2.62748569e-01 -1.56473696e-01 -5.54648399e-01 -1.17199637e-01
5.29238284e-02 5.57018876e-01 1.20428979e+00 -1.02281809e+00
-8.17282975e-01 -7.58202672e-02 -5.20627573e-02 2.02733904e-01
-5.11485875e-01 -2.60111779e-01 7.62659237e-02 -5.40662646e-01
-7.99349129e-01 -3.77050072e-01 1.11191561e-02 -1.16272964e-01
3.61096680e-01 -5.40949166e-01 9.58635092e-01 -2.07423225e-01
1.21214569e+00 -2.08499908e+00 3.92846972e-01 -1.80285592e-02
-2.70750880e-01 8.59632015e-01 -4.64325249e-01 4.97657061e-01
-1.92990556e-01 -4.11141694e-01 -5.27326882e-01 2.79411599e-02
-8.73588845e-02 8.82110357e-01 -1.98369008e-02 3.59749466e-01
3.71607393e-01 1.40973365e+00 -1.17150438e+00 -3.47829431e-01
4.62200135e-01 2.82104552e-01 -1.03636757e-01 4.50620174e-01
-2.57535636e-01 6.89467788e-02 -3.17119330e-01 5.56925356e-01
3.47118020e-01 -4.04587947e-02 -1.67836502e-01 -1.26351073e-01
-6.32671714e-02 -3.84478062e-01 -1.13962281e+00 1.95297623e+00
-8.91836643e-01 1.38713688e-01 -1.28382176e-01 -1.71366060e+00
9.14735496e-01 4.31944877e-01 4.70741808e-01 -6.49941385e-01
3.24085681e-03 6.98786005e-02 -3.29231590e-01 -6.46224201e-01
-1.81808765e-03 -8.60117316e-01 2.54610516e-02 3.67302477e-01
1.22715211e+00 -8.48616567e-03 2.49361306e-01 -2.29665905e-01
1.04901922e+00 3.92873049e-01 1.00980747e+00 -2.25448370e-01
7.22115159e-01 -1.71810016e-01 5.21641135e-01 7.97839642e-01
-4.27643359e-01 1.85541257e-01 -8.28426480e-02 -4.66411382e-01
-7.51797736e-01 -9.39431489e-01 -2.94398725e-01 1.97633326e+00
-3.53960514e-01 -2.96788216e-01 -4.63804364e-01 -1.00951314e+00
2.92833418e-01 8.92259002e-01 -1.13899028e+00 -7.44538188e-01
-2.81472564e-01 -8.91525030e-01 3.31093669e-01 6.26804233e-01
4.86856103e-01 -1.45711160e+00 -8.35913241e-01 3.35464656e-01
4.58223313e-01 -5.17142892e-01 -1.64779574e-01 8.68305147e-01
-8.09673846e-01 -1.17869556e+00 -9.83322203e-01 -7.57854044e-01
2.82267898e-01 2.01831490e-01 1.02765989e+00 -3.91230464e-01
-6.75849915e-01 8.67579222e-01 -8.20984364e-01 -5.03339410e-01
-5.20805180e-01 -4.37214822e-02 1.74127251e-01 2.51827061e-01
4.88374770e-01 -7.09328890e-01 -6.98909312e-02 1.64673626e-01
-1.12899101e+00 -2.62693733e-01 8.22379053e-01 1.28636122e+00
5.19145191e-01 -1.26846105e-01 9.97584343e-01 -1.31166553e+00
6.03582323e-01 -9.02676761e-01 2.71946713e-02 6.65867448e-01
-6.82081938e-01 2.64726818e-01 6.09753013e-01 -6.57972932e-01
-1.48921633e+00 6.70409650e-02 -3.84608060e-02 -6.92394137e-01
-4.04884070e-01 5.71759045e-01 -1.76207006e-01 -1.55503526e-01
1.41427195e+00 3.97775352e-01 -3.00777727e-03 -5.72651923e-01
8.94852042e-01 6.69044495e-01 4.27883208e-01 -7.43985116e-01
3.23033422e-01 2.62885213e-01 -1.11527070e-01 -9.46514666e-01
-1.27093089e+00 -9.45490599e-01 -1.12831676e+00 -2.58279949e-01
7.48244226e-01 -6.42037272e-01 -1.52527794e-01 2.78549016e-01
-7.45438635e-01 -5.18125534e-01 -1.18862104e+00 3.77166748e-01
-9.96581078e-01 4.30144742e-02 -2.91805744e-01 -7.37633407e-01
-3.48577917e-01 -4.87814218e-01 7.40210772e-01 -5.27654104e-02
1.31243756e-02 -1.51800764e+00 6.34076893e-01 -1.90283254e-01
5.08820832e-01 3.10057521e-01 8.95545661e-01 -9.30594206e-01
3.99106622e-01 -8.52474496e-02 2.70416518e-03 8.43235850e-01
3.57343674e-01 -6.61085129e-01 -1.36701739e+00 -4.84711498e-01
8.86144191e-02 -9.27933455e-01 1.24771082e+00 3.33302617e-01
8.02273333e-01 -2.20028371e-01 -1.95718974e-01 5.49421847e-01
1.53196251e+00 1.47634447e-01 4.80497152e-01 2.70261168e-01
5.03732443e-01 5.68714678e-01 7.52874374e-01 4.04986024e-01
-1.22628622e-01 4.96569216e-01 -8.05577710e-02 2.91813731e-01
-5.04718840e-01 -7.38091171e-02 4.85353649e-01 7.79962957e-01
-3.72931927e-01 3.46317768e-01 -7.72639573e-01 5.18548191e-01
-2.29799366e+00 -1.18035603e+00 4.39761549e-01 2.04451752e+00
1.12641156e+00 1.80071648e-02 1.69729605e-01 2.44813398e-01
8.33130181e-01 6.17473572e-02 -6.85594916e-01 -5.46364784e-01
-1.42230261e-02 7.12682962e-01 2.02241153e-01 2.99285084e-01
-1.23677528e+00 8.20065558e-01 5.70216846e+00 1.42901003e+00
-8.63561869e-01 8.12760532e-01 1.07460096e-01 1.73052996e-01
3.88026834e-02 -1.48794606e-01 -7.45086670e-01 3.60226780e-01
1.34130347e+00 -4.53900516e-01 2.61892647e-01 1.11013424e+00
-4.35520768e-01 5.21514937e-02 -1.30092371e+00 1.02181876e+00
1.99266329e-01 -1.36417794e+00 3.09918493e-01 -3.29226941e-01
9.05456305e-01 2.91392170e-02 -2.40076214e-01 1.24866378e+00
3.49951059e-01 -8.37824523e-01 4.26357895e-01 9.27915633e-01
1.06832922e+00 -7.61968374e-01 6.41633332e-01 6.87650740e-01
-1.32065034e+00 -5.15690327e-01 -7.74799824e-01 -9.46077779e-02
-5.43489940e-02 4.05742735e-01 -5.70567548e-01 8.78912449e-01
4.82829750e-01 1.14885473e+00 -5.27708828e-01 9.15984392e-01
-1.01018466e-01 4.84680116e-01 2.08783820e-01 1.34233981e-01
3.04366648e-01 2.43006274e-01 1.83817834e-01 1.53696799e+00
2.06809491e-01 9.15631056e-02 3.62017393e-01 5.88255584e-01
1.46680072e-01 1.09898537e-01 -7.30882823e-01 2.06376359e-01
2.44983926e-01 1.23614478e+00 -2.67764062e-01 -4.56411749e-01
-6.33471370e-01 1.19267714e+00 6.99117541e-01 2.50769258e-01
-2.88932711e-01 -4.30419624e-01 3.47206444e-01 -8.86971951e-02
4.18830395e-01 4.01944250e-01 2.99870402e-01 -1.15449274e+00
-3.14134181e-01 -3.70620489e-01 9.46952343e-01 -5.06393731e-01
-1.86497092e+00 3.20247173e-01 4.04903054e-01 -1.23781109e+00
-3.50547284e-01 -8.44276607e-01 -6.48517072e-01 8.49046230e-01
-1.76069915e+00 -1.45484161e+00 -1.21382512e-01 9.67348754e-01
8.00572753e-01 -5.78242064e-01 1.19658184e+00 6.83162287e-02
-1.08859189e-01 4.48256552e-01 1.88174263e-01 -4.26831663e-01
7.72921145e-01 -1.31121051e+00 1.74085665e-02 2.99409896e-01
2.37108860e-02 2.87195176e-01 1.13504872e-01 -5.58784842e-01
-1.17619967e+00 -1.55612338e+00 5.39278746e-01 -3.73257786e-01
6.48191154e-01 -4.77784984e-02 -1.27020299e+00 4.77929533e-01
-9.61391777e-02 5.73822081e-01 9.64646697e-01 1.24292731e-01
-4.94887263e-01 -9.92029533e-02 -1.24403846e+00 -2.00915158e-01
9.35123742e-01 -7.12255895e-01 -1.25833786e+00 2.83041358e-01
6.07799113e-01 2.11440250e-01 -1.03520441e+00 4.35075492e-01
3.69771004e-01 -8.32774878e-01 1.05042112e+00 -1.01770854e+00
2.46268809e-02 4.77070361e-02 -2.44190723e-01 -1.74011445e+00
-6.64102852e-01 -4.08162117e-01 -6.49809420e-01 9.64156926e-01
1.78459380e-02 -4.14603353e-01 2.61580557e-01 1.59360766e-01
-4.47896928e-01 -6.87113166e-01 -1.26999831e+00 -1.32883477e+00
2.48185888e-01 -5.47220886e-01 1.42733067e-01 1.23935699e+00
1.76622465e-01 4.85908657e-01 -4.61037993e-01 -6.05975330e-01
5.12076139e-01 -1.03005111e-01 4.03316438e-01 -1.62708127e+00
-5.39585710e-01 -3.11684847e-01 -6.18589103e-01 -2.75744647e-01
5.63401282e-01 -1.43819666e+00 -2.03063060e-02 -1.47344971e+00
4.26284492e-01 -4.14442867e-02 -7.39944637e-01 8.97828639e-01
-3.06220084e-01 3.10331676e-02 9.37636942e-02 1.50973856e-01
-9.00408745e-01 7.68566310e-01 1.11235642e+00 -1.24398701e-01
-2.85664678e-01 1.56369835e-01 -3.36836457e-01 6.42587721e-01
4.61920142e-01 -2.50947148e-01 -7.47986674e-01 2.93300837e-01
-1.16727896e-01 -3.34598541e-01 2.20298931e-01 -9.97716069e-01
1.30380690e-01 -2.87975460e-01 4.03052330e-01 -9.20432433e-03
3.16546649e-01 -6.15975976e-01 -1.58067018e-01 6.06744885e-01
-3.74103397e-01 -7.19420135e-01 1.74592078e-01 8.98150623e-01
-2.76209235e-01 -6.62139416e-01 1.07945716e+00 -5.81396461e-01
-1.49110854e+00 3.02579582e-01 -2.97441959e-01 4.19990331e-01
1.28333366e+00 -4.24191952e-01 -1.35440946e-01 1.57675803e-01
-1.14762616e+00 -4.21696045e-02 -1.33533180e-02 4.16736633e-01
6.79353356e-01 -1.60615778e+00 -6.20303154e-01 2.83483863e-01
5.94255209e-01 -2.24264368e-01 4.67704386e-01 5.79565287e-01
1.93351209e-01 3.73452663e-01 -5.75499892e-01 -3.39043975e-01
-8.38955343e-01 1.00040126e+00 1.81860223e-01 -1.73906043e-01
-6.56918705e-01 1.03888285e+00 7.35807642e-02 -7.36010849e-01
1.76976025e-01 -1.32892430e-01 -3.38820815e-01 6.25818193e-01
6.82914019e-01 8.24580967e-01 2.82230973e-01 -5.52215338e-01
-1.39346167e-01 3.88914227e-01 -2.36714613e-02 1.22983545e-01
1.49219191e+00 1.72151834e-01 1.42957374e-01 1.03929043e+00
1.40231478e+00 -8.89939666e-01 -1.47103190e+00 -8.12338054e-01
3.81044567e-01 -1.04457118e-01 2.23695245e-02 -8.73830795e-01
-3.98890674e-01 9.73672986e-01 5.91329157e-01 -3.49315815e-02
1.04580593e+00 5.69116278e-03 3.12072128e-01 7.35107422e-01
4.34016228e-01 -1.52171302e+00 3.25645089e-01 7.59670198e-01
8.58704209e-01 -1.31780303e+00 -3.32060345e-02 1.10589683e-01
-7.39465475e-01 1.25059617e+00 5.88415980e-01 -2.08283914e-03
1.10863447e+00 -5.14539070e-02 -4.78796542e-01 -1.60936013e-01
-1.00345135e+00 -6.56780124e-01 6.78618729e-01 9.59926844e-01
1.82111293e-01 -1.71199236e-02 -1.56892151e-01 1.03749788e+00
4.34389263e-01 3.79140407e-01 -1.13953881e-01 1.41779304e+00
-9.68132854e-01 -1.10419405e+00 -4.38659526e-02 6.64031804e-01
3.61971967e-02 4.05211300e-02 -1.65956825e-01 7.33291149e-01
6.96780801e-01 5.76192141e-01 -5.22785969e-02 -2.02933937e-01
5.77244520e-01 8.01719666e-01 6.93442464e-01 -1.32028210e+00
-6.98303521e-01 -3.21752340e-01 -8.50603655e-02 -5.18257499e-01
-8.99219751e-01 -3.30070734e-01 -8.70772958e-01 3.15776229e-01
-3.27783525e-01 1.77040085e-01 1.58590049e-01 1.27788401e+00
9.54001099e-02 6.11429095e-01 4.67496842e-01 -9.44250584e-01
-8.29760551e-01 -1.27518058e+00 -8.83113086e-01 5.74482381e-01
1.63270056e-01 -1.08692729e+00 -2.54231572e-01 3.59887987e-01] | [10.014633178710938, 3.1050620079040527] |
7aabe2ff-bfbc-4da9-8ab4-b6f4d5189221 | interpretable-machine-learning-of-amino-acid | 2303.15228 | null | https://arxiv.org/abs/2303.15228v1 | https://arxiv.org/pdf/2303.15228v1.pdf | Interpretable machine learning of amino acid patterns in proteins: a statistical ensemble approach | Explainable and interpretable unsupervised machine learning helps understand the underlying structure of data. We introduce an ensemble analysis of machine learning models to consolidate their interpretation. Its application shows that restricted Boltzmann machines compress consistently into a few bits the information stored in a sequence of five amino acids at the start or end of $\alpha$-helices or $\beta$-sheets. The weights learned by the machines reveal unexpected properties of the amino acids and the secondary structure of proteins: (i) His and Thr have a negligible contribution to the amphiphilic pattern of $\alpha$-helices; (ii) there is a class of $\alpha$-helices particularly rich in Ala at their end; (iii) Pro occupies most often slots otherwise occupied by polar or charged amino acids, and its presence at the start of helices is relevant; (iv) Glu and especially Asp on one side, and Val, Leu, Iso, and Phe on the other, display the strongest tendency to mark amphiphilic patterns, i.e., extreme values of an "effective hydrophobicity", though they are not the most powerful (non) hydrophobic amino acids. | ['Marco Baiesi', 'Enzo Orlandini', 'Anna Braghetto'] | 2023-03-27 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 1.95131168e-01 4.03559685e-01 -3.22281033e-01 -3.80966574e-01
1.00434318e-01 -5.09573042e-01 4.87122476e-01 5.29858232e-01
-5.80742717e-01 1.15057790e+00 4.27003354e-01 -6.33347988e-01
-4.93349470e-02 -5.76160371e-01 -5.99417806e-01 -1.31843042e+00
-5.81494272e-01 5.43662369e-01 7.93733373e-02 -4.59910393e-01
4.20017213e-01 4.74574655e-01 -1.49371469e+00 3.33344698e-01
9.71241832e-01 8.07039201e-01 4.03369009e-01 3.65481079e-01
-5.49041212e-01 7.63118327e-01 -3.22295800e-02 -1.90020576e-01
3.85142788e-02 -2.97440857e-01 -7.72212684e-01 -3.41809243e-01
6.97711930e-02 8.57470259e-02 1.97188452e-01 1.20304453e+00
2.20754206e-01 7.37460107e-02 1.10946083e+00 -5.44078410e-01
-4.23815221e-01 5.55767775e-01 -4.62761402e-01 8.66943784e-03
1.82639614e-01 2.59756267e-01 1.27673101e+00 -8.96308064e-01
7.44432271e-01 9.34516013e-01 4.08387989e-01 5.31330049e-01
-1.20588493e+00 -9.83735994e-02 3.39984089e-01 2.46810541e-02
-1.06754255e+00 -2.01274887e-01 6.74270809e-01 -5.56911170e-01
1.29280901e+00 6.20490909e-01 6.16555035e-01 8.36198926e-01
7.38030195e-01 3.18373054e-01 1.27015913e+00 -4.20632303e-01
5.92180371e-01 1.64915323e-01 5.29139221e-01 6.61023080e-01
2.37667695e-01 3.80045138e-02 -3.32745552e-01 -5.85113108e-01
8.16045254e-02 1.52601659e-01 -2.49029785e-01 -7.17487097e-01
-1.11561489e+00 8.01739454e-01 1.22264855e-01 3.91860366e-01
-3.94792676e-01 -1.59286141e-01 2.39471376e-01 1.11541348e-02
2.17433795e-01 5.17428935e-01 -8.68103266e-01 -3.04650158e-01
-6.91035569e-01 2.82972932e-01 8.70539844e-01 4.21541870e-01
9.69298482e-01 -1.59675732e-01 7.13054061e-01 7.66414523e-01
1.87763900e-01 2.81195790e-01 5.37873924e-01 -5.50594330e-01
3.83234769e-01 6.01188958e-01 4.82402563e-01 -6.29000902e-01
-5.48632205e-01 -5.79261780e-02 -9.90594804e-01 4.16136920e-01
4.15751278e-01 3.03831510e-02 -6.91614628e-01 1.73392761e+00
-5.92552535e-02 -6.63967788e-01 5.93155473e-02 6.60705805e-01
5.74891984e-01 7.58920431e-01 4.82151747e-01 -7.56162465e-01
1.35466993e+00 -4.53689784e-01 -4.70731407e-01 -8.13800916e-02
4.53021586e-01 -3.76914173e-01 8.59411359e-01 2.54482776e-01
-1.17616546e+00 -4.82925475e-01 -1.05693030e+00 1.11495301e-01
-6.44187629e-01 -5.48943639e-01 6.32913768e-01 4.41086143e-01
-6.80701256e-01 1.01093221e+00 -6.51215136e-01 -9.88747478e-02
-9.52242687e-02 4.41707313e-01 -5.37554204e-01 4.59696680e-01
-1.01305318e+00 1.21006501e+00 5.40865839e-01 -2.13420510e-01
-4.72921938e-01 -4.35050339e-01 -7.41297662e-01 -2.03647558e-02
-1.25968307e-01 -4.00847942e-01 5.32963693e-01 -1.33751166e+00
-1.26975453e+00 8.59258533e-01 -3.89950871e-01 -3.80313188e-01
1.51400179e-01 -2.04625532e-01 -2.75168806e-01 -3.34265269e-02
-5.69073617e-01 7.03333020e-01 5.63810050e-01 -1.27310300e+00
-3.92125458e-01 -5.09799838e-01 -1.73348919e-01 1.43049583e-01
4.37686741e-02 4.39417586e-02 5.98255455e-01 -4.97312605e-01
4.38862890e-01 -7.36324906e-01 -5.37483990e-01 -3.15152228e-01
-3.15798342e-01 -2.12830350e-01 5.91899455e-01 -5.27156770e-01
1.21475255e+00 -1.97335267e+00 3.26835752e-01 5.77306330e-01
3.63212138e-01 2.52488822e-01 4.42989767e-01 7.98150182e-01
-8.47578108e-01 1.59140781e-01 -5.68009138e-01 7.84952641e-01
3.26412693e-02 3.58971447e-01 -4.70101535e-01 4.78298903e-01
-9.73043293e-02 4.86958951e-01 -7.96911240e-01 -1.84488714e-01
2.20827371e-01 2.00342715e-01 -3.70627224e-01 -5.31402938e-02
-3.77401501e-01 2.17846751e-01 -3.79870832e-01 4.95082229e-01
5.98281384e-01 -3.31421703e-01 5.95983803e-01 -3.00258212e-02
-5.14647245e-01 6.32713914e-01 -7.37017870e-01 9.15225208e-01
4.40957129e-01 4.54624742e-01 -8.09039697e-02 -1.03161526e+00
1.08550537e+00 2.67895311e-01 5.89764774e-01 -4.40666348e-01
-1.60376027e-01 1.78116873e-01 5.60669243e-01 -4.70469862e-01
4.31002736e-01 -5.16783714e-01 1.71647891e-01 5.07080853e-01
-2.42640123e-01 3.98181565e-02 -3.00526787e-02 4.67529111e-02
9.59770799e-01 9.71446857e-02 7.36557603e-01 -8.56749117e-01
6.29084289e-01 -4.96002883e-01 8.75657618e-01 3.03341389e-01
-2.09599152e-01 3.84403020e-01 5.85757077e-01 -9.21915412e-01
-1.50112820e+00 -1.07039464e+00 -2.64491409e-01 1.34725893e+00
1.23547077e-01 -3.52641106e-01 -7.84156024e-01 -3.87776583e-01
5.04824854e-02 7.42261529e-01 -6.01048946e-01 3.94097529e-02
-5.79911113e-01 -1.03213191e+00 -2.47501537e-01 3.90875101e-01
1.86179265e-01 -1.42013204e+00 -7.89797783e-01 2.42338642e-01
-1.12436041e-01 -2.27672279e-01 2.56658215e-02 1.05941808e+00
-1.17686760e+00 -9.24347043e-01 -5.98086417e-01 -4.95428532e-01
7.54132271e-01 -2.52719551e-01 1.20130956e+00 -3.06849889e-02
-1.13846146e-01 -3.02438110e-01 -2.31504545e-01 -5.90838909e-01
-2.90916592e-01 -2.36377344e-01 3.89086366e-01 -3.70090067e-01
5.77561259e-01 -8.60560596e-01 -7.13811994e-01 3.13174009e-01
-5.85793555e-01 -7.55201429e-02 3.77025038e-01 6.53519452e-01
6.16248310e-01 -2.68457383e-01 1.97228253e-01 -9.07927930e-01
4.09747988e-01 -4.35624540e-01 -1.27923056e-01 4.04219888e-02
-5.46718776e-01 5.16743660e-01 6.08132541e-01 -1.86927423e-01
-7.65213728e-01 -9.45200995e-02 -2.91061938e-01 3.46874386e-01
-3.98501426e-01 3.62931252e-01 -5.26817203e-01 2.18431205e-01
6.44877315e-01 5.05687416e-01 1.99315455e-02 -5.90823591e-01
5.46183698e-02 4.89378035e-01 1.52091652e-01 -6.78892255e-01
4.30924237e-01 5.47812700e-01 -2.34889805e-01 -1.28800118e+00
-1.23541981e-01 -1.52006119e-01 -6.44029319e-01 1.22107103e-01
8.32135797e-01 -3.98336530e-01 -5.45562446e-01 1.92634448e-01
-7.97322154e-01 -1.47021085e-01 -4.31749016e-01 6.18826747e-01
-8.64582479e-01 8.97648811e-01 -5.13413608e-01 -9.17369246e-01
-3.27821910e-01 -1.00864363e+00 2.78768837e-01 1.01081423e-01
-9.14221823e-01 -8.69026661e-01 4.36739355e-01 8.13701749e-02
1.48429975e-01 1.24022439e-01 1.80133080e+00 -9.86570954e-01
-4.41266857e-02 2.32445523e-01 9.30944737e-03 3.41786444e-01
3.22188959e-02 3.24453384e-01 -6.52228713e-01 -2.89697587e-01
-2.85492331e-01 -3.41543742e-02 9.80201006e-01 3.78333718e-01
1.12289202e+00 -3.75383437e-01 -4.03333396e-01 3.54231387e-01
1.12759674e+00 8.10481012e-01 7.53003895e-01 2.93910712e-01
8.61277953e-02 7.93650210e-01 3.24315608e-01 3.09830636e-01
-1.96137160e-01 3.90715748e-01 6.27598464e-01 6.12880029e-02
6.77147448e-01 -4.40100916e-02 4.48487401e-01 9.02810156e-01
-6.46087587e-01 -5.20321652e-02 -8.41985226e-01 5.18969357e-01
-1.54178882e+00 -9.97580707e-01 -5.15841655e-02 2.24603629e+00
1.11817658e+00 4.32511955e-01 6.24937471e-03 1.21045765e-02
5.66679180e-01 4.14373666e-01 -5.14519274e-01 -9.27104890e-01
-2.42813632e-01 5.59540033e-01 6.28791749e-01 5.39109349e-01
-9.32745397e-01 6.31438375e-01 7.39041185e+00 6.31995976e-01
-9.28046107e-01 -5.10751903e-01 3.47984105e-01 -1.32608619e-02
-7.96598136e-01 3.36224705e-01 -6.18739069e-01 1.02863991e+00
8.71578574e-01 3.38048846e-01 7.72685260e-02 8.95114005e-01
2.64434457e-01 -1.99112713e-01 -1.00039732e+00 6.09807789e-01
-5.35103023e-01 -1.26239049e+00 2.18710124e-01 4.44504112e-01
4.77151483e-01 6.89247530e-03 1.36254644e-02 -2.35188663e-01
5.38850248e-01 -1.22443557e+00 4.42736030e-01 6.81461692e-01
5.71736336e-01 -1.03996503e+00 5.32289207e-01 5.45259118e-01
-5.38998663e-01 6.44735545e-02 -8.51526380e-01 -3.75944763e-01
-4.06536460e-02 7.95207918e-01 -2.87424028e-01 2.22547725e-02
6.93263352e-01 1.93138123e-01 -9.72898230e-02 4.28883433e-01
-1.56857535e-01 6.57605886e-01 -2.60916382e-01 -2.85914958e-01
5.27321458e-01 -7.43995786e-01 7.03779936e-01 1.20766127e+00
-8.56986344e-02 5.79326689e-01 -1.85375556e-01 4.27848399e-01
2.49411687e-01 2.48858541e-01 -6.15874171e-01 -4.67945412e-02
7.73021504e-02 8.57003510e-01 -6.75168157e-01 -3.77471536e-01
-2.59900093e-01 4.43570912e-01 2.79954076e-01 4.53177661e-01
-4.00268584e-01 -4.57857728e-01 1.08224702e+00 1.60429478e-01
3.62442404e-01 -1.38999045e-01 -6.51269555e-02 -9.02045190e-01
-1.12674803e-01 -9.09489393e-01 2.24470854e-01 -6.22328162e-01
-9.81617153e-01 4.27995205e-01 -2.95013964e-01 -6.73726082e-01
-3.76341015e-01 -7.89173961e-01 -5.72139561e-01 8.99906814e-01
-1.01295888e+00 -4.42170620e-01 3.58739555e-01 4.25334960e-01
1.37442693e-01 -4.41001743e-01 1.14262748e+00 -3.84823889e-01
-1.44493893e-01 1.82813883e-01 6.97444201e-01 -3.46338041e-02
3.86899471e-01 -1.45433605e+00 2.24420831e-01 2.46576667e-01
1.14730269e-01 1.03717124e+00 1.30284572e+00 -5.84566951e-01
-8.64340961e-01 -5.90861976e-01 1.16271102e+00 -1.79087266e-01
1.60622180e-01 -1.29172921e-01 -1.09391940e+00 5.66806078e-01
1.84368864e-01 -4.09551561e-01 1.37356973e+00 3.64824802e-01
-4.94991601e-01 2.34349445e-01 -1.08024979e+00 6.83405101e-01
5.83582044e-01 -3.47661585e-01 -1.03818619e+00 2.99081326e-01
2.92325795e-01 2.14186743e-01 -5.82434654e-01 3.81801546e-01
9.89033461e-01 -1.52888060e+00 8.27825963e-01 -1.09721410e+00
4.77896601e-01 -1.26282379e-01 -2.88773894e-01 -1.03411388e+00
-5.59082210e-01 -4.96253878e-01 2.50498857e-02 3.93205911e-01
6.56308889e-01 -7.64153719e-01 9.37717438e-01 6.53406799e-01
1.28529780e-02 -8.87142539e-01 -1.10335696e+00 -3.22149158e-01
3.23500395e-01 -5.09292725e-03 3.98531646e-01 8.99761736e-01
6.74479604e-01 -9.20462757e-02 -2.40056843e-01 -4.81473118e-01
3.77157837e-01 2.66148388e-01 2.19075561e-01 -1.22770643e+00
-2.45854050e-01 -3.78585786e-01 -3.30764621e-01 -9.98507679e-01
-1.84487417e-01 -6.52371645e-01 -2.34245956e-01 -9.27522838e-01
4.26682413e-01 -3.06216747e-01 -6.39139175e-01 2.62232423e-01
1.08405933e-01 -3.21542501e-01 -1.02758901e-02 2.23728150e-01
-3.50264728e-01 2.40923181e-01 5.96746624e-01 3.81880463e-03
-3.26468796e-02 -1.05083600e-01 -7.31305122e-01 1.54131377e+00
7.91280091e-01 -1.67708173e-01 -1.12295084e-01 1.05330311e-01
5.91194510e-01 1.59699395e-02 -1.03052035e-01 -5.48967659e-01
-3.71669680e-01 -4.80297059e-01 5.59792101e-01 -6.64317369e-01
2.29324624e-01 -1.07952940e+00 2.28790358e-01 7.95310080e-01
-3.35378230e-01 1.11675456e-01 -2.47084916e-01 4.21914369e-01
1.57885611e-01 -3.43183547e-01 1.03140640e+00 -4.72100705e-01
-4.76372153e-01 4.65878323e-02 -9.48913455e-01 -1.57055050e-01
1.02979159e+00 -6.24748290e-01 7.44105428e-02 -3.13533098e-01
-1.01294255e+00 -3.31437200e-01 8.32559288e-01 1.33915558e-01
5.22520959e-01 -7.79699385e-01 -4.35679615e-01 3.09233576e-01
6.26427010e-02 -4.29125905e-01 9.64149013e-02 3.90160859e-01
-8.47438812e-01 4.64282930e-01 -5.74356616e-01 -2.79067606e-01
-1.41903865e+00 7.16705441e-01 2.55183160e-01 -4.33884710e-01
-5.41751146e-01 8.27656090e-01 3.79183680e-01 -3.11108708e-01
3.79035920e-02 -3.47993612e-01 -4.07717407e-01 1.15516700e-01
4.99483556e-01 4.55265492e-01 -1.18568495e-01 -7.43204355e-01
-5.77203572e-01 4.08932656e-01 -5.35631657e-01 3.52290869e-01
1.38887119e+00 1.72045991e-01 -5.67387819e-01 4.78009641e-01
1.06309175e+00 2.28743792e-01 -1.00060713e+00 1.08676113e-01
1.06606789e-01 -2.58539189e-02 -6.23686194e-01 -7.24109948e-01
-4.33077246e-01 9.33603168e-01 4.66770530e-01 1.68812618e-01
5.35007000e-01 -1.75407361e-02 5.57336390e-01 4.26017731e-01
6.08603895e-01 -1.23001504e+00 -4.40192133e-01 5.65550864e-01
4.82980371e-01 -7.31401443e-01 2.64423639e-01 -8.16951245e-02
-5.42402864e-01 1.13166118e+00 8.05434510e-02 -1.28665185e-02
5.66677332e-01 1.35204524e-01 -2.22171828e-01 -2.19785735e-01
-1.00852466e+00 1.78361326e-01 1.24630537e-02 3.74164402e-01
7.42480099e-01 2.70604253e-01 -9.77392256e-01 3.69980454e-01
-3.59868616e-01 -5.32520831e-01 1.80307165e-01 8.68532240e-01
-1.10924375e+00 -9.87359703e-01 -2.44613826e-01 4.01133448e-01
-6.33737087e-01 -4.04442161e-01 -5.88067174e-01 6.91231310e-01
4.33584660e-01 5.89028418e-01 2.61840165e-01 -1.52411446e-01
-2.24431470e-01 6.35698855e-01 3.13383728e-01 -1.80970624e-01
-6.10579073e-01 -1.51503198e-02 -1.61155656e-01 -5.41500628e-01
-2.63252169e-01 -3.87999177e-01 -1.44377065e+00 -3.56103927e-01
-1.31381720e-01 8.50943387e-01 5.46268582e-01 9.19865429e-01
2.03190461e-01 -6.62465841e-02 2.61252403e-01 -4.67107356e-01
-5.96609592e-01 -8.96691859e-01 -1.08516574e+00 4.40772653e-01
3.33704144e-01 -5.64688385e-01 -5.52119493e-01 3.48957740e-02] | [4.781103134155273, 5.347103595733643] |
1ebbaae3-fc42-490b-b7dd-bc43ee54435a | deep-learning-of-physical-laws-from-scarce | 2005.03448 | null | https://arxiv.org/abs/2005.03448v3 | https://arxiv.org/pdf/2005.03448v3.pdf | Physics-informed learning of governing equations from scarce data | Harnessing data to discover the underlying governing laws or equations that describe the behavior of complex physical systems can significantly advance our modeling, simulation and understanding of such systems in various science and engineering disciplines. This work introduces a novel physics-informed deep learning framework to discover governing partial differential equations (PDEs) from scarce and noisy data for nonlinear spatiotemporal systems. In particular, this approach seamlessly integrates the strengths of deep neural networks for rich representation learning, physics embedding, automatic differentiation and sparse regression to (1) approximate the solution of system variables, (2) compute essential derivatives, as well as (3) identify the key derivative terms and parameters that form the structure and explicit expression of the PDEs. The efficacy and robustness of this method are demonstrated, both numerically and experimentally, on discovering a variety of PDE systems with different levels of data scarcity and noise accounting for different initial/boundary conditions. The resulting computational framework shows the potential for closed-form model discovery in practical applications where large and accurate datasets are intractable to capture. | ['Hao Sun', 'Zhao Chen', 'Yang Liu'] | 2020-05-05 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [-2.40390494e-01 -5.10275364e-01 3.65271658e-01 1.94885954e-01
-4.79567111e-01 -7.79424906e-01 6.11292779e-01 1.23313896e-01
9.11795422e-02 9.81759727e-01 7.96493888e-02 -4.20594364e-01
-5.47362924e-01 -4.44554657e-01 -5.43879449e-01 -1.06333649e+00
-3.82711411e-01 2.63752222e-01 -2.99038559e-01 -3.86869788e-01
7.06804693e-02 1.18584692e+00 -1.11132264e+00 -5.32102287e-01
8.67388189e-01 1.23167825e+00 -1.93465963e-01 6.84532225e-01
1.35341406e-01 8.51478755e-01 -6.83286339e-02 3.49519163e-01
3.24879795e-01 -2.82999575e-01 -4.00880843e-01 -1.86114535e-01
1.77105233e-01 -4.45687294e-01 -9.12841499e-01 7.50028312e-01
5.24221003e-01 5.05495250e-01 8.59917462e-01 -8.73547077e-01
-9.36143398e-01 -2.50352353e-01 -3.40869367e-01 6.25307858e-01
1.99138504e-02 8.73757482e-01 7.87343442e-01 -8.15400004e-01
3.77413452e-01 8.91534626e-01 1.10256708e+00 2.66932249e-01
-1.38964987e+00 -3.07391018e-01 -2.34806672e-01 -2.23631248e-01
-1.44137883e+00 -3.56018811e-01 9.21628475e-01 -8.36658597e-01
6.96460426e-01 7.42800012e-02 7.11534500e-01 1.09235716e+00
5.17934620e-01 2.52220422e-01 1.00276864e+00 2.40261599e-01
4.77345586e-01 -7.88386911e-02 2.16243684e-01 8.07583690e-01
3.02849770e-01 3.81690651e-01 -2.36962721e-01 -5.65643847e-01
1.14733744e+00 1.20578475e-01 -3.85012150e-01 -1.96443468e-01
-9.21637774e-01 7.44888365e-01 4.41492587e-01 2.05337524e-01
-8.34945321e-01 4.24025029e-01 1.23206645e-01 8.78731906e-02
3.85039240e-01 1.00527012e+00 -6.47483349e-01 -4.63151038e-02
-7.99405158e-01 7.30517149e-01 1.01177371e+00 6.90456569e-01
7.98803985e-01 5.28707445e-01 -6.77509680e-02 3.89221787e-01
3.07634659e-02 7.89298177e-01 5.25178500e-02 -1.32837927e+00
-1.93511888e-01 4.78141755e-01 6.72708571e-01 -1.15102351e+00
-6.08316123e-01 -5.63794553e-01 -1.15401900e+00 4.39536124e-02
2.81880140e-01 -5.61660647e-01 -7.68899024e-01 1.51873398e+00
6.03576601e-01 4.61850643e-01 -3.37753780e-02 1.12594020e+00
7.79552460e-01 9.92879212e-01 -1.99967727e-01 5.10032885e-02
1.12627661e+00 -3.13192338e-01 -4.10732239e-01 2.55737513e-01
4.59167451e-01 -3.83752882e-01 9.94963109e-01 -6.22383989e-02
-1.21409822e+00 -4.60044444e-01 -7.29336500e-01 -2.61713535e-01
-3.31588387e-01 -9.10570193e-03 7.63016880e-01 -1.00412071e-01
-1.08537877e+00 1.02607930e+00 -1.12534416e+00 3.82282995e-02
6.51749611e-01 2.89218754e-01 -5.70849925e-02 1.08685076e-01
-1.21127498e+00 6.89422309e-01 -3.82363319e-01 4.15377617e-01
-1.18551612e+00 -1.93198693e+00 -7.21356511e-01 3.51089031e-01
8.32541510e-02 -1.08369040e+00 8.46524954e-01 -3.07007462e-01
-1.50473344e+00 2.96628714e-01 -1.51089489e-01 -3.26965302e-01
5.09560049e-01 -1.58524603e-01 -1.78511277e-01 1.51111811e-01
2.65714619e-02 8.92435983e-02 6.70283496e-01 -1.13319969e+00
-2.00786465e-03 -5.40533960e-02 -8.15255344e-02 -2.22275853e-01
-2.72009317e-02 -5.46753168e-01 9.52984393e-02 -6.01404607e-01
-1.13050766e-01 -8.50605071e-01 -5.26238322e-01 5.09200513e-01
-2.07339063e-01 -1.03509672e-01 8.22237611e-01 -8.10397625e-01
6.65264904e-01 -1.96138084e+00 3.96237135e-01 2.27518693e-01
5.60903013e-01 3.15224648e-01 2.13546097e-01 8.77723277e-01
1.61977515e-01 9.93043259e-02 -5.12067735e-01 -6.58722147e-02
-4.04439121e-02 2.81085193e-01 -6.51404560e-01 7.78244555e-01
7.18389452e-01 1.18923700e+00 -7.89878786e-01 5.65457307e-02
4.40123707e-01 1.01940143e+00 -5.79073071e-01 3.69283736e-01
-2.22862825e-01 1.09973967e+00 -8.19413483e-01 3.52998614e-01
5.37291765e-01 -6.06423080e-01 -4.29647446e-01 -2.73713619e-01
-3.19505900e-01 1.74752235e-01 -1.19668555e+00 1.26211262e+00
-6.06995165e-01 7.91601121e-01 6.41762197e-01 -1.28963697e+00
6.22190714e-01 2.78050721e-01 1.02220535e+00 -6.19413376e-01
3.01718950e-01 2.18629852e-01 -2.79606562e-02 -8.56448889e-01
1.73690125e-01 -2.78889388e-01 1.03759756e-02 3.70620579e-01
3.26225115e-03 -4.10448253e-01 7.01231137e-02 1.94094360e-01
1.39616549e+00 -1.41249627e-01 -6.04177117e-02 -8.32448065e-01
7.48674214e-01 1.10229641e-01 3.30221236e-01 6.05019152e-01
-1.16342656e-01 1.88565925e-01 5.26726544e-01 -6.96395814e-01
-1.26361382e+00 -9.38690901e-01 -3.29545736e-01 3.47328007e-01
1.30236804e-01 2.41310626e-01 -3.56484473e-01 4.03987199e-01
5.52649736e-01 2.86139786e-01 -7.17354953e-01 -5.28566837e-01
-9.26410794e-01 -5.90813935e-01 3.56605053e-01 4.67906773e-01
2.70754218e-01 -8.37823868e-01 -5.61546087e-01 3.37593555e-01
2.86173791e-01 -1.38952994e+00 -1.37253851e-01 -1.29740573e-02
-6.27708495e-01 -1.00263393e+00 -6.70954168e-01 -4.47891742e-01
5.08348584e-01 -1.20780349e-01 8.22390199e-01 1.13848731e-01
-7.64250219e-01 6.52294040e-01 2.39368275e-01 1.12138070e-01
-4.71024364e-01 -1.09025344e-01 2.55232811e-01 5.62080108e-02
-5.08573830e-01 -1.13269770e+00 -7.76150167e-01 -2.75458358e-02
-8.66431415e-01 -4.09281731e-01 3.19498867e-01 6.48324370e-01
5.83640099e-01 9.72008705e-02 5.33181906e-01 -3.85613114e-01
9.90805030e-01 -8.16239834e-01 -1.01443183e+00 -1.25785932e-01
-5.09014487e-01 2.69888163e-01 1.16261041e+00 -4.87209827e-01
-7.94145107e-01 -1.77076250e-01 5.46814762e-02 -7.75695920e-01
-9.08230320e-02 5.73349833e-01 3.04249138e-01 -4.19149935e-01
6.00365341e-01 4.29724753e-01 -8.41141418e-02 -6.46544516e-01
2.57379919e-01 2.06401646e-02 8.56040120e-01 -9.78352904e-01
8.42997909e-01 6.84967160e-01 7.26021647e-01 -1.18950021e+00
-5.59837222e-01 -5.68688512e-01 -5.56981623e-01 5.37406728e-02
3.81944448e-01 -8.38200390e-01 -1.19345999e+00 5.27469516e-01
-1.07093143e+00 -4.78013635e-01 -5.32868981e-01 2.99908489e-01
-2.95176536e-01 1.33834630e-01 -8.57569098e-01 -8.93318236e-01
-3.01840365e-01 -8.41278374e-01 1.18033850e+00 3.20948452e-01
-2.38256812e-01 -1.40722978e+00 3.01181376e-01 -1.50719374e-01
7.97097921e-01 7.70371616e-01 1.11389899e+00 -4.84345928e-02
-5.41798115e-01 -3.05186301e-01 -3.00708592e-01 1.11259721e-01
1.63930833e-01 2.82604426e-01 -7.50946581e-01 -2.43094280e-01
1.98182940e-01 -9.73008722e-02 6.30996644e-01 7.38571227e-01
7.92713761e-01 -3.89248401e-01 -1.01184204e-01 1.00575960e+00
1.48119998e+00 -1.67769685e-01 1.22381583e-01 -6.09741151e-01
1.02422607e+00 4.06904429e-01 -4.08730865e-01 7.79804409e-01
2.49351040e-01 3.10973644e-01 1.82222620e-01 -1.69236362e-01
2.01254077e-02 -9.31084063e-03 1.29069671e-01 7.56723940e-01
1.72693759e-01 2.54633650e-02 -1.07719946e+00 7.22942531e-01
-1.39057076e+00 -7.45562196e-01 -1.76498443e-01 1.64822078e+00
7.81769216e-01 -3.33930820e-01 9.12151039e-02 -1.83509648e-01
3.61559987e-01 2.51000002e-02 -1.29658651e+00 -2.89678067e-01
-1.43120915e-01 4.20760751e-01 4.05628234e-01 4.70377296e-01
-9.48258102e-01 7.33723879e-01 6.49279976e+00 1.91293463e-01
-1.59503758e+00 -9.01743472e-02 4.00492519e-01 -1.61724575e-02
-3.15571785e-01 -1.13264345e-01 -4.58360761e-01 3.09344292e-01
1.02762032e+00 -3.05715621e-01 7.75272429e-01 3.47382963e-01
6.97164118e-01 2.16493279e-01 -9.77100551e-01 8.22462261e-01
-7.01303184e-01 -1.89560044e+00 -1.08472981e-01 2.89182186e-01
1.12961972e+00 2.68986851e-01 1.54777557e-01 -1.82842076e-01
4.59205151e-01 -1.22677791e+00 5.23491621e-01 1.02810895e+00
5.58083236e-01 -2.60018051e-01 3.54448259e-01 2.90933788e-01
-1.07659781e+00 -1.07962430e-01 -2.15808436e-01 -4.39440101e-01
2.58053809e-01 8.58543456e-01 -1.16993368e-01 2.71509469e-01
7.22222626e-01 1.14347851e+00 -1.85063528e-03 9.12523448e-01
1.58809751e-01 7.83796012e-01 -7.09820688e-01 -6.29348084e-02
5.03283083e-01 -4.68114078e-01 8.03843439e-01 8.22467566e-01
2.77300090e-01 5.26465714e-01 1.53356316e-02 1.67882538e+00
-7.38677233e-02 -3.22919607e-01 -5.49496174e-01 -4.44796503e-01
4.18368757e-01 1.26886451e+00 -5.58508337e-01 -5.70448451e-02
-1.38706878e-01 4.43706155e-01 1.62473217e-01 9.51515734e-01
-8.46717060e-01 -1.31804692e-02 1.21884167e+00 1.90521866e-01
5.56473613e-01 -8.22217464e-01 -3.07839543e-01 -1.06713033e+00
2.34791011e-01 -4.25350219e-01 1.19348457e-02 -6.24315858e-01
-1.58150232e+00 2.38812074e-01 -3.29555757e-02 -7.46753752e-01
-6.63431808e-02 -7.85012126e-01 -9.59799886e-01 1.07747984e+00
-1.56005359e+00 -7.55267620e-01 -2.57742167e-01 7.04799473e-01
-1.12295493e-01 3.54583673e-02 5.95600426e-01 3.28728020e-01
-9.12006438e-01 1.56076706e-03 7.94203222e-01 1.63141385e-01
-1.99165344e-01 -1.02067769e+00 3.70323002e-01 5.88318706e-01
-2.71168679e-01 6.50744200e-01 8.93354475e-01 -4.10859764e-01
-2.07017875e+00 -9.48954821e-01 1.30075678e-01 -3.63984048e-01
1.10353816e+00 -3.70248646e-01 -1.20776629e+00 1.54905826e-01
-3.16945136e-01 6.52520776e-01 2.62558639e-01 -3.51103723e-01
-7.88265988e-02 -1.61365971e-01 -1.11824906e+00 3.67274076e-01
6.80986643e-01 -6.38280928e-01 -8.90665948e-02 3.83784622e-01
5.34363985e-01 -3.68578494e-01 -1.14964104e+00 2.71010786e-01
4.97169495e-01 -4.60409701e-01 1.11087871e+00 -1.02961791e+00
7.34726310e-01 -9.50492471e-02 4.74332049e-02 -1.20140922e+00
-3.26663345e-01 -1.23388982e+00 -6.55336499e-01 8.69291365e-01
2.27752075e-01 -8.86333048e-01 4.02467787e-01 9.68453109e-01
-1.27574041e-01 -1.09710896e+00 -9.81318951e-01 -7.02374697e-01
5.96660793e-01 -2.81562626e-01 6.19324088e-01 1.07126868e+00
-5.06292641e-01 -8.28153342e-02 -1.89006820e-01 5.68627715e-01
4.22585160e-01 3.27863514e-01 5.27885795e-01 -1.29641473e+00
-1.99880108e-01 -6.20939434e-01 -2.58130074e-01 -8.89579535e-01
2.50646979e-01 -6.55545950e-01 -1.07885063e-01 -1.39244986e+00
-5.18351912e-01 -6.25193059e-01 -2.51064092e-01 -3.58421654e-02
2.01894902e-02 -2.15188712e-01 -1.39654234e-01 3.91740441e-01
1.46065969e-02 9.21710491e-01 1.34847081e+00 -8.31273422e-02
-1.52929932e-01 -1.61066681e-01 -7.77270138e-01 4.83707845e-01
5.72497189e-01 -3.72359484e-01 -2.55167067e-01 -3.66966397e-01
7.98283368e-02 1.34752616e-01 1.13358128e+00 -1.05358183e+00
3.45382988e-01 -3.48099351e-01 3.57342482e-01 1.85692273e-02
3.33913654e-01 -7.41335690e-01 1.12115622e-01 4.29189593e-01
-3.41841012e-01 5.27640469e-02 7.31319666e-01 6.01014972e-01
9.88409370e-02 3.67036849e-01 7.77126253e-01 7.28708832e-03
-4.88320291e-01 6.71936750e-01 -5.18783987e-01 4.75107849e-01
7.34028041e-01 1.78613618e-01 -7.53492787e-02 -5.49065948e-01
-4.89833057e-01 3.14238995e-01 1.88040391e-01 2.12562699e-02
3.37705135e-01 -1.19163299e+00 -7.39357710e-01 3.57786924e-01
-4.39332753e-01 1.11361638e-01 3.76440138e-01 9.43849742e-01
-8.16898406e-01 3.23766947e-01 -6.90899864e-02 -6.73660576e-01
-1.92132220e-01 2.81638384e-01 8.29145908e-01 -2.21329957e-01
-8.80720735e-01 7.76701748e-01 8.52341652e-02 -4.95268464e-01
-1.90004751e-01 -7.04624176e-01 3.28977823e-01 -4.27853256e-01
2.24689350e-01 5.91934264e-01 -2.13551238e-01 -7.38564253e-01
-3.91495556e-01 6.91887677e-01 4.66139883e-01 2.22957224e-01
1.64158428e+00 3.78211588e-02 -1.76832393e-01 4.01926339e-01
1.56496227e+00 -2.66660988e-01 -1.84516001e+00 -3.38961422e-01
-3.07610691e-01 5.38482666e-02 4.64785457e-01 -4.64020967e-01
-1.33315456e+00 1.09995651e+00 2.68822908e-01 4.05376613e-01
7.64453888e-01 -1.70737132e-02 1.29097009e+00 6.19621038e-01
-6.18751347e-02 -7.03884840e-01 -3.07515591e-01 5.95197439e-01
8.76122892e-01 -9.55984294e-01 2.01629177e-01 -5.94853647e-02
-3.76253203e-02 1.09560084e+00 2.20615104e-01 -7.40461528e-01
1.26426411e+00 4.89585221e-01 -1.77003890e-01 -5.16702771e-01
-6.09125376e-01 1.54126048e-01 3.71518910e-01 2.58139700e-01
1.30995125e-01 -2.26480618e-01 3.39304924e-01 5.30498028e-01
-6.53214157e-02 -3.54654640e-02 3.71066123e-01 7.25254416e-01
1.11563317e-01 -3.18694621e-01 8.22985079e-03 4.39493537e-01
-2.05941141e-01 -8.99069607e-02 -3.20284128e-01 8.21262717e-01
-1.74118936e-01 5.37847340e-01 -8.20303410e-02 1.77996695e-01
3.20450842e-01 -1.86378956e-01 2.06546351e-01 -2.09855646e-01
-4.34963167e-01 -1.83696970e-01 -4.70049858e-01 -5.53145111e-01
-6.70275241e-02 -6.74601793e-01 -1.36986780e+00 -7.02643573e-01
7.14987889e-02 1.17630653e-01 3.83703381e-01 1.28407454e+00
9.70091760e-01 6.63372040e-01 5.68392634e-01 -1.21269965e+00
-5.89121222e-01 -4.97671813e-01 -6.73019767e-01 3.74716818e-01
1.07788825e+00 -7.95058966e-01 -6.26361430e-01 2.35440377e-02] | [6.548013210296631, 3.4209301471710205] |
7a5a8ab2-f385-4a70-8739-51b4fed6d30e | time-space-transformers-for-video-panoptic | 2210.03546 | null | https://arxiv.org/abs/2210.03546v1 | https://arxiv.org/pdf/2210.03546v1.pdf | Time-Space Transformers for Video Panoptic Segmentation | We propose a novel solution for the task of video panoptic segmentation, that simultaneously predicts pixel-level semantic and instance segmentation and generates clip-level instance tracks. Our network, named VPS-Transformer, with a hybrid architecture based on the state-of-the-art panoptic segmentation network Panoptic-DeepLab, combines a convolutional architecture for single-frame panoptic segmentation and a novel video module based on an instantiation of the pure Transformer block. The Transformer, equipped with attention mechanisms, models spatio-temporal relations between backbone output features of current and past frames for more accurate and consistent panoptic estimates. As the pure Transformer block introduces large computation overhead when processing high resolution images, we propose a few design changes for a more efficient compute. We study how to aggregate information more effectively over the space-time volume and we compare several variants of the Transformer block with different attention schemes. Extensive experiments on the Cityscapes-VPS dataset demonstrate that our best model improves the temporal consistency and video panoptic quality by a margin of 2.2%, with little extra computation. | ['Sergiu Nedevschi', 'Andra Petrovai'] | 2022-10-07 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 5.76777533e-02 -2.33556554e-01 -2.03105539e-01 -3.03996801e-01
-7.46763170e-01 -4.62036461e-01 5.06250024e-01 -1.69899777e-01
-1.48739204e-01 3.16748500e-01 6.48721084e-02 -2.04692572e-01
2.89425626e-02 -1.09900129e+00 -8.46260905e-01 -7.15138972e-01
-3.24691921e-01 3.70419979e-01 9.71487761e-01 2.46209744e-02
-1.38444011e-03 3.52098316e-01 -1.55986691e+00 6.53806746e-01
8.23607922e-01 1.51013613e+00 2.10012227e-01 9.73996639e-01
-2.59839177e-01 1.01043773e+00 -4.40042317e-01 -1.54035971e-01
7.65456259e-01 -2.10158437e-01 -7.65871227e-01 4.20403540e-01
1.14505601e+00 -6.29033089e-01 -5.14800787e-01 7.92187035e-01
3.53704672e-03 1.53154671e-01 2.56084919e-01 -1.04488802e+00
-2.01288313e-01 6.41539693e-01 -7.78456807e-01 8.45316112e-01
-2.46153787e-01 5.49728215e-01 1.20141935e+00 -2.45499983e-01
6.19354367e-01 1.23638046e+00 8.26765597e-01 1.35855645e-01
-1.15666854e+00 -4.32938278e-01 5.25649309e-01 3.19141559e-02
-1.19570446e+00 -2.33057037e-01 1.77967697e-01 -4.63894129e-01
1.09478974e+00 5.17819107e-01 9.24626470e-01 4.49031383e-01
1.23921908e-01 8.60163569e-01 9.89325702e-01 3.29514384e-01
1.24354251e-01 -3.36180359e-01 1.13775611e-01 6.97261870e-01
-1.30485222e-01 -9.99467894e-02 -6.08930662e-02 2.01804504e-01
9.48099554e-01 1.12361796e-02 -4.58343714e-01 1.38492207e-04
-1.15170765e+00 5.81214011e-01 8.40264976e-01 3.24333668e-01
-4.95400280e-01 6.00342035e-01 2.75980800e-01 2.00728744e-01
1.10043192e+00 4.11778480e-01 -7.45092571e-01 1.21943071e-01
-1.74705601e+00 3.74960929e-01 6.54275060e-01 8.51099432e-01
9.07979906e-01 1.43877357e-01 -4.18432951e-01 5.67352295e-01
2.43737042e-01 4.73707229e-01 9.11377147e-02 -1.29347420e+00
6.33420289e-01 4.44096059e-01 2.40891315e-02 -8.45935524e-01
-4.98919994e-01 -2.78841466e-01 -6.69616401e-01 1.65163595e-02
4.10292447e-01 -2.07081228e-01 -1.33801889e+00 1.46620345e+00
4.79837567e-01 1.08272564e+00 -2.70476013e-01 1.12779582e+00
6.72235489e-01 1.39222014e+00 1.55980587e-01 -1.87331215e-01
1.42671609e+00 -1.34369373e+00 -3.19205165e-01 -1.81419879e-01
1.98778361e-01 -6.31848931e-01 8.28912437e-01 -3.72226187e-03
-1.19363415e+00 -6.20159447e-01 -9.86329615e-01 -6.05621822e-02
-4.81235743e-01 -2.23165913e-03 6.38169587e-01 2.89315552e-01
-1.41458261e+00 9.06416714e-01 -9.83415246e-01 -3.80379856e-01
4.19150978e-01 2.32139304e-01 4.21832912e-02 2.96302378e-01
-1.18971705e+00 3.38906020e-01 4.39267874e-01 4.17283438e-02
-7.47943580e-01 -1.24677074e+00 -8.50758672e-01 4.62001264e-01
3.65810364e-01 -9.51835155e-01 1.34098303e+00 -1.13624501e+00
-1.58779657e+00 7.72969782e-01 1.36982232e-01 -8.99895549e-01
5.17172098e-01 -2.87012249e-01 -3.80973935e-01 7.21461296e-01
1.29205763e-01 1.19596565e+00 1.04775953e+00 -8.63296270e-01
-1.19827521e+00 -8.89704153e-02 1.82092354e-01 1.28558427e-01
1.58185184e-01 -8.30721930e-02 -1.09087396e+00 -7.02662528e-01
5.27052432e-02 -9.12384033e-01 -5.63739836e-01 3.39407995e-02
-1.06338590e-01 7.01592639e-02 8.46987784e-01 -8.00231636e-01
1.34380758e+00 -1.95608306e+00 4.07387037e-03 7.39586651e-02
3.45208406e-01 4.06152368e-01 -1.43933281e-01 -2.18318358e-01
1.19047128e-01 1.87151253e-01 -6.40069127e-01 -4.59511012e-01
-2.19815135e-01 9.86810997e-02 -6.42165661e-01 3.37581873e-01
2.28134722e-01 7.03825772e-01 -7.31002927e-01 -4.71879929e-01
4.67743307e-01 2.40937158e-01 -6.40899122e-01 3.23596984e-01
-1.00946212e+00 3.57748657e-01 -4.48453635e-01 6.02631032e-01
1.08653498e+00 -5.44194221e-01 -7.72133917e-02 -1.49902612e-01
-5.84362626e-01 5.42366385e-01 -1.00147176e+00 1.66807044e+00
-4.82488513e-01 9.41623926e-01 2.23699749e-01 -7.23709166e-01
4.64070618e-01 2.64723390e-01 6.27045155e-01 -7.46963084e-01
-1.14737965e-01 6.04290552e-02 -5.85547268e-01 -4.94883657e-01
9.15725589e-01 4.14489776e-01 5.88751435e-02 1.71100736e-01
-6.81528598e-02 -3.28183800e-01 4.83846247e-01 2.28705987e-01
8.54389369e-01 3.51413250e-01 -5.45848869e-02 -4.84291285e-01
3.54576945e-01 3.06977183e-01 6.55701995e-01 6.61220193e-01
-1.19481437e-01 9.47757125e-01 6.30231440e-01 -8.57526600e-01
-1.14171875e+00 -9.51313615e-01 -1.96386680e-01 9.86568749e-01
2.34494835e-01 -6.65270746e-01 -8.97079945e-01 -4.72492099e-01
-1.56592578e-01 4.31785375e-01 -6.25264585e-01 6.84010208e-01
-9.52060103e-01 -1.28077877e+00 3.84180039e-01 5.55565357e-01
1.02723062e+00 -7.49333739e-01 -9.54723656e-01 3.39958549e-01
-1.99557781e-01 -1.49655414e+00 -5.49619317e-01 -2.42372721e-01
-9.36402619e-01 -9.70597506e-01 -7.91271269e-01 -3.15161467e-01
5.98057397e-02 4.39993382e-01 1.43943000e+00 -4.47681919e-02
-8.52587521e-02 1.88919529e-01 -8.99925232e-02 8.77334252e-02
-2.24900901e-01 1.99050605e-01 -6.71936512e-01 2.33431339e-01
2.43990961e-02 -7.09893227e-01 -9.33036745e-01 2.17056334e-01
-1.00357985e+00 5.43191552e-01 3.18216413e-01 2.88548797e-01
7.68793106e-01 -7.80442432e-02 -2.01087117e-01 -9.55087304e-01
-1.75658941e-01 -7.71061182e-01 -1.39357209e+00 1.78937793e-01
-3.92814904e-01 -5.82010597e-02 5.36074877e-01 -3.24568227e-02
-1.12311578e+00 7.73904026e-02 -1.79911181e-01 -7.72807121e-01
-7.36181810e-02 2.46934414e-01 2.16340348e-01 3.36979240e-01
2.25373536e-01 1.10670060e-01 -4.55753773e-01 -5.59306920e-01
4.74954963e-01 4.17232126e-01 6.73469067e-01 -2.01882392e-01
4.51616973e-01 8.29434514e-01 -1.88828707e-01 -8.81036162e-01
-9.47579384e-01 -6.53183103e-01 -5.79977572e-01 -1.93343312e-01
1.26771724e+00 -1.37928700e+00 -4.13864762e-01 7.04681814e-01
-1.10537601e+00 -8.72837901e-01 -1.54460624e-01 3.03834200e-01
-5.55791974e-01 2.40499005e-01 -9.06420231e-01 -3.09126675e-01
-5.39103508e-01 -1.24295282e+00 1.50264108e+00 3.11736226e-01
1.14846975e-01 -7.30686426e-01 2.69835472e-01 3.02852482e-01
5.37078321e-01 2.26581365e-01 6.34768903e-01 -1.43186480e-01
-1.47438514e+00 5.36990643e-01 -6.78761065e-01 1.02313332e-01
-2.14027703e-01 5.52930236e-01 -1.08733642e+00 2.90927291e-02
-1.30506203e-01 2.23709032e-01 1.61276758e+00 9.22356367e-01
1.26742530e+00 -4.46529567e-01 -2.98567742e-01 1.27626550e+00
1.61753631e+00 9.34474990e-02 8.66045475e-01 3.71616393e-01
8.89144778e-01 3.57691705e-01 4.75790769e-01 3.02427381e-01
4.89562571e-01 7.00927079e-01 4.73309726e-01 -1.67306319e-01
-2.52090216e-01 6.13848716e-02 3.15346330e-01 6.84350014e-01
-6.87093958e-02 -6.26653671e-01 -7.93932974e-01 8.49200189e-01
-2.13771725e+00 -1.03175795e+00 -2.35030144e-01 1.96931386e+00
5.90560138e-01 2.52923220e-02 1.41568899e-01 -1.97836891e-01
6.10013306e-01 1.01537836e+00 -2.46846184e-01 -1.94145203e-01
-1.03018351e-01 2.46642932e-01 1.00197101e+00 6.44370139e-01
-1.73231041e+00 1.30194139e+00 6.43231249e+00 8.93678844e-01
-1.36188221e+00 1.98940217e-01 1.13386059e+00 -2.13676140e-01
-7.19957203e-02 7.40326121e-02 -9.79996920e-01 7.23002553e-01
1.32104921e+00 3.66661012e-01 2.77852446e-01 7.06055939e-01
3.57118905e-01 -1.44476101e-01 -7.53671646e-01 7.19870627e-01
-2.80360550e-01 -2.05739355e+00 2.88734794e-01 -2.41243929e-01
1.03354955e+00 7.68125355e-01 2.75480648e-04 7.53679723e-02
3.11226815e-01 -6.05766773e-01 7.28763819e-01 3.72190446e-01
5.97806334e-01 -5.31853259e-01 3.29580754e-01 -2.19764292e-01
-1.73306108e+00 -7.63270035e-02 -2.56554455e-01 2.56465942e-01
4.12593722e-01 6.75872803e-01 -3.51003945e-01 4.74660933e-01
1.14927924e+00 1.12080169e+00 -5.40025055e-01 1.19825280e+00
-6.25473633e-02 8.11191499e-01 -6.45403683e-01 6.43464684e-01
1.07265604e+00 -3.85194719e-01 6.74473286e-01 1.55839586e+00
3.73802364e-01 1.85263574e-01 2.42427602e-01 9.45627153e-01
-1.59819245e-01 -1.26498833e-01 -3.14844042e-01 2.44549066e-01
1.74935997e-01 1.22375941e+00 -9.82865930e-01 -9.57332313e-01
-5.28000772e-01 7.80708075e-01 1.16159983e-01 3.10620993e-01
-1.10774457e+00 1.84676081e-01 9.68096733e-01 2.29060695e-01
8.25510681e-01 -6.31659478e-02 -2.94950008e-01 -1.31760859e+00
-2.65253097e-01 -6.32185161e-01 6.93070292e-01 -1.04879308e+00
-7.50927687e-01 8.13188910e-01 8.75902995e-02 -1.13234830e+00
-1.01329513e-01 -2.59369493e-01 -7.75195777e-01 5.29629409e-01
-1.78095996e+00 -1.07877588e+00 -4.63489294e-01 4.79315549e-01
1.00778949e+00 3.00242513e-01 2.40950212e-01 3.71152222e-01
-6.65697217e-01 -2.04234459e-02 1.02672517e-01 3.78120481e-03
2.31258810e-01 -1.35240722e+00 9.67935026e-01 1.15336680e+00
1.25016868e-01 -1.28759995e-01 5.66647768e-01 -4.73222643e-01
-7.38618851e-01 -1.76411486e+00 8.36046815e-01 -1.99201360e-01
8.08160067e-01 -5.68374284e-02 -9.92360055e-01 9.06326771e-01
4.22271580e-01 8.12299475e-02 -1.67020541e-02 -4.03669834e-01
-3.15798879e-01 -4.43487585e-01 -8.57396841e-01 7.25867033e-01
7.82930017e-01 -2.93199897e-01 -4.77120131e-01 4.29423660e-01
1.28810143e+00 -5.35424471e-01 -7.88241982e-01 4.06465203e-01
1.78168386e-01 -1.15832460e+00 9.74025548e-01 -1.15556918e-01
5.29211223e-01 -5.24739027e-01 -7.20803961e-02 -1.06985927e+00
-3.22134197e-01 -8.68121624e-01 -8.83804187e-02 8.73628795e-01
2.92177647e-01 -2.69609272e-01 5.86144090e-01 3.60561401e-01
-4.04606670e-01 -4.95781869e-01 -9.81286943e-01 -5.15213549e-01
-2.10498989e-01 -5.05631149e-01 8.11160147e-01 7.25991666e-01
-5.69755614e-01 1.90785125e-01 -4.75202978e-01 6.23628259e-01
3.69074136e-01 7.34013259e-01 5.01586139e-01 -8.86691391e-01
-3.54705244e-01 -6.40261889e-01 -2.63967484e-01 -1.69095480e+00
-1.29294838e-03 -4.23416167e-01 6.09981120e-02 -1.38856375e+00
-5.43594956e-02 -2.90540934e-01 -2.87019223e-01 2.05336958e-01
-1.10945955e-01 3.93411160e-01 3.30843806e-01 3.35316390e-01
-7.49708772e-01 3.06460053e-01 1.19653821e+00 -2.81566024e-01
-4.66652095e-01 1.40560091e-01 -4.19758186e-02 8.27144384e-01
7.37044871e-01 -4.70933020e-01 -4.21162248e-01 -8.56687129e-01
-8.74670893e-02 1.89246431e-01 3.64974767e-01 -1.24756503e+00
1.62345618e-01 -2.14839652e-01 4.99424897e-02 -1.00214195e+00
4.43398893e-01 -7.64560461e-01 2.84692526e-01 3.20499212e-01
-1.67334288e-01 2.81973362e-01 5.28146088e-01 5.73082685e-01
-3.42700988e-01 2.86699057e-01 1.08965456e+00 -4.59941208e-01
-1.19322419e+00 7.92126000e-01 -5.66741288e-01 9.12289321e-03
9.10261810e-01 6.04192764e-02 -5.05887210e-01 -1.26933843e-01
-6.02055967e-01 5.31653881e-01 3.01561475e-01 3.74395847e-01
7.44281560e-02 -8.15357924e-01 -6.55086577e-01 1.11021981e-01
-2.43855417e-01 1.09472133e-01 4.88019109e-01 9.92913187e-01
-1.31371319e+00 6.49950981e-01 -1.06325902e-01 -9.37377274e-01
-9.54915643e-01 4.00493860e-01 6.38769567e-01 -6.00214899e-01
-1.01931226e+00 7.37819850e-01 6.07886374e-01 5.15858009e-02
-6.04340322e-02 -1.07335269e+00 -2.26534694e-01 2.58608520e-01
6.97854817e-01 3.08247894e-01 -1.12427048e-01 -5.61052442e-01
-5.50497919e-02 6.08516455e-01 9.04961154e-02 -1.11645512e-01
1.54374993e+00 -3.61916512e-01 -9.50487107e-02 3.11694533e-01
1.01091194e+00 -4.63560939e-01 -2.01329327e+00 -2.44404539e-01
-1.60817683e-01 -4.98093635e-01 3.96866202e-01 -5.65896988e-01
-1.81050646e+00 9.16561246e-01 5.09040952e-01 2.44500652e-01
1.34601772e+00 -2.70763457e-01 1.28238869e+00 8.24449509e-02
-7.58572519e-02 -1.00314784e+00 -4.30554390e-01 4.46648628e-01
4.81392324e-01 -1.06611085e+00 6.54821619e-02 -5.57099104e-01
-4.59013402e-01 9.69886303e-01 7.21360862e-01 -3.16248834e-01
7.72726059e-01 2.09275961e-01 7.37939179e-02 -3.01404983e-01
-1.05690658e+00 -3.65596443e-01 3.40055317e-01 6.11946732e-02
4.51060273e-02 2.73473542e-02 8.26496854e-02 9.97224748e-02
1.33233234e-01 2.87148710e-02 2.58335650e-01 3.72554749e-01
-5.82838774e-01 -4.20686841e-01 -2.29668528e-01 3.25818598e-01
-5.22419155e-01 -4.78688359e-01 2.23996311e-01 9.74883318e-01
2.30515733e-01 5.91896176e-01 9.52875555e-01 -1.19327739e-01
-8.31061006e-02 -2.63229042e-01 8.32250565e-02 -1.82726622e-01
-8.44485343e-01 4.29534227e-01 1.47745222e-01 -1.03411782e+00
-9.24244165e-01 -4.79130536e-01 -9.67075169e-01 -7.15278745e-01
2.69012988e-01 1.18751144e-02 2.32505724e-01 8.33417118e-01
4.71310258e-01 5.87857127e-01 5.61988652e-01 -1.10068202e+00
1.09676674e-01 -8.13846409e-01 -5.40572941e-01 2.53037155e-01
3.99197012e-01 -1.56954348e-01 -2.91526526e-01 3.08609277e-01] | [9.355246543884277, -0.009507892653346062] |
e6ba8826-ab8e-4026-baba-f56f0bc01af4 | enhancing-low-light-images-using-infrared | 2307.04122 | null | https://arxiv.org/abs/2307.04122v1 | https://arxiv.org/pdf/2307.04122v1.pdf | Enhancing Low-Light Images Using Infrared-Encoded Images | Low-light image enhancement task is essential yet challenging as it is ill-posed intrinsically. Previous arts mainly focus on the low-light images captured in the visible spectrum using pixel-wise loss, which limits the capacity of recovering the brightness, contrast, and texture details due to the small number of income photons. In this work, we propose a novel approach to increase the visibility of images captured under low-light environments by removing the in-camera infrared (IR) cut-off filter, which allows for the capture of more photons and results in improved signal-to-noise ratio due to the inclusion of information from the IR spectrum. To verify the proposed strategy, we collect a paired dataset of low-light images captured without the IR cut-off filter, with corresponding long-exposure reference images with an external filter. The experimental results on the proposed dataset demonstrate the effectiveness of the proposed method, showing better performance quantitatively and qualitatively. The dataset and code are publicly available at https://wyf0912.github.io/ELIEI/ | ['Bihan Wen', 'Alex C. Kot', 'Wenhan Yang', 'Renjie Wan', 'YuFei Wang', 'Shulin Tian'] | 2023-07-09 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 7.48106360e-01 -3.88841838e-01 8.49744231e-02 -2.01205462e-02
-4.82276499e-01 -4.20211375e-01 2.32477590e-01 -3.33799988e-01
-4.88389403e-01 6.79781258e-01 1.23088390e-01 -2.08302382e-02
9.58261415e-02 -8.04536164e-01 -5.51656246e-01 -1.16028380e+00
3.62416923e-01 -6.96372747e-01 2.68607914e-01 2.68900413e-02
1.20307036e-01 1.25820190e-01 -1.77656674e+00 1.04588427e-01
1.07887256e+00 1.02778494e+00 6.52469277e-01 2.65634865e-01
5.29713370e-02 5.95697761e-01 -9.29915309e-02 -1.53594822e-01
7.09354997e-01 -4.71374661e-01 -1.05592176e-01 1.70385651e-02
4.70836401e-01 -7.65411496e-01 -5.37691474e-01 1.42969263e+00
4.36019361e-01 2.41540551e-01 2.11193487e-01 -8.65710735e-01
-5.95489502e-01 -2.59235620e-01 -6.78532362e-01 9.61881131e-02
3.81309837e-01 5.18349409e-01 6.09298646e-01 -8.20709646e-01
4.26568240e-01 8.17824244e-01 3.86596262e-01 3.35262507e-01
-1.16077840e+00 -8.94276977e-01 -3.55101824e-01 2.80619413e-01
-1.20744932e+00 -5.95872581e-01 1.10270214e+00 -1.05757371e-01
2.85239577e-01 2.45206222e-01 5.28046548e-01 7.13664472e-01
2.87876666e-01 2.44812086e-01 1.77645361e+00 -5.31427622e-01
-9.41192284e-02 2.73913592e-01 -5.08427806e-02 6.02664351e-01
5.77900529e-01 6.73643947e-01 -4.20399934e-01 1.58807248e-01
6.98228002e-01 4.83983696e-01 -8.47738028e-01 6.55111298e-02
-8.41470718e-01 9.54160765e-02 5.47229707e-01 2.48702243e-01
-1.49944991e-01 -1.66012436e-01 -1.06534719e-01 1.17833957e-01
4.91483122e-01 5.31580336e-02 -3.88642680e-03 2.84429044e-01
-4.83837634e-01 -3.49515349e-01 1.57855958e-01 5.83562136e-01
8.90867591e-01 -2.28233024e-01 -2.07162336e-01 8.95440221e-01
3.66792679e-01 8.43076885e-01 5.99034615e-02 -9.30006087e-01
3.79398644e-01 3.06860745e-01 4.51015681e-01 -7.94220686e-01
-9.10995677e-02 -4.05875623e-01 -9.37829614e-01 6.01431191e-01
4.47887838e-01 -8.69892836e-02 -7.10118115e-01 1.49871683e+00
4.26934540e-01 2.26572886e-01 1.73966780e-01 1.25383162e+00
7.96837032e-01 7.32163131e-01 -1.79056197e-01 -7.19341636e-01
1.34341729e+00 -7.12053716e-01 -8.53327811e-01 -1.98108971e-01
-4.87061925e-02 -1.08526826e+00 1.19368899e+00 4.71890181e-01
-1.20282328e+00 -7.05665350e-01 -1.00289941e+00 -1.80432126e-01
1.37797043e-01 4.90390807e-01 3.04739803e-01 4.99402821e-01
-7.47939706e-01 4.10687745e-01 -4.90955383e-01 -1.69130266e-01
2.57092178e-01 -8.59178603e-02 -2.68078148e-01 -6.46530509e-01
-9.78207946e-01 4.40595895e-01 2.02742577e-01 2.91320354e-01
-5.13567686e-01 -5.82840741e-01 -5.82292020e-01 -2.66388822e-02
4.41863567e-01 -3.99252117e-01 6.89629376e-01 -8.79796505e-01
-1.66826427e+00 7.12424636e-01 -2.40672067e-01 8.36258531e-02
3.48631829e-01 -1.43106967e-01 -3.68143260e-01 5.06339490e-01
-1.42567828e-01 1.88985560e-02 8.15910161e-01 -1.57484722e+00
-4.40798014e-01 -4.37200755e-01 1.99671745e-01 3.35856199e-01
-4.04651582e-01 3.28790136e-02 -5.15304089e-01 -4.20974731e-01
1.67154729e-01 -7.55960941e-01 1.24086000e-01 1.79458797e-01
-2.34697729e-01 4.29362178e-01 8.78037632e-01 -6.21814966e-01
9.92029846e-01 -2.43481398e+00 -6.64336085e-01 -3.06245387e-01
-3.63751650e-02 4.47610319e-01 -1.89309180e-01 5.12167335e-01
8.86550639e-03 -3.09671938e-01 -3.86595398e-01 -9.38157365e-02
-6.67436838e-01 -8.11171457e-02 -2.30995696e-02 7.51906276e-01
-1.10911891e-01 5.07946372e-01 -8.52521539e-01 -4.96112883e-01
5.78427315e-01 6.37174249e-01 -3.73359099e-02 2.86524087e-01
8.39880556e-02 6.96170449e-01 -2.59197772e-01 6.68842554e-01
1.10406148e+00 9.07799304e-02 -4.66867574e-02 -5.63519478e-01
-5.25030315e-01 -1.59517705e-01 -1.14902985e+00 1.43043351e+00
-5.66660106e-01 4.54737544e-01 2.97184646e-01 -3.98180127e-01
9.29037690e-01 6.91386908e-02 3.31257582e-01 -1.10260320e+00
1.85256690e-01 1.66154504e-01 -3.97380799e-01 -7.28000522e-01
1.35159373e-01 -3.39876056e-01 5.38198233e-01 2.13329837e-01
-3.97483885e-01 2.97215711e-02 1.80928171e-01 -1.12067625e-01
6.29009187e-01 1.56867146e-01 1.02907792e-01 -1.43960556e-02
7.47835219e-01 -4.26087976e-01 5.43311059e-01 5.44773757e-01
-1.54784411e-01 5.96373498e-01 -2.95727968e-01 -6.89186156e-02
-8.72209132e-01 -1.04515970e+00 -3.38436335e-01 5.27308524e-01
8.15549493e-01 -1.65497698e-02 -3.69532257e-01 -4.39484492e-02
-3.38858038e-01 6.30611002e-01 -1.08064272e-01 -1.01163365e-01
-2.45517462e-01 -7.91369796e-01 -1.53009206e-01 -3.48485745e-02
1.09666395e+00 -8.48051429e-01 -5.72576880e-01 -1.75068095e-01
-4.33878779e-01 -1.23561203e+00 -2.61890143e-01 -2.81679511e-01
-7.27016449e-01 -1.25097644e+00 -6.06048346e-01 -5.70734084e-01
6.52107894e-01 8.81601095e-01 7.38588393e-01 1.63270459e-01
-7.26766706e-01 1.81886971e-01 -3.48144948e-01 -2.53567457e-01
-7.61394426e-02 -6.61149561e-01 -3.44734073e-01 4.58781034e-01
5.71057312e-02 -4.31933671e-01 -1.17982495e+00 3.40024084e-01
-1.04563713e+00 4.12984669e-01 8.10259044e-01 9.06786978e-01
5.63353598e-01 4.44394499e-01 2.20865369e-01 -5.65421641e-01
2.26608709e-01 1.16503283e-01 -9.19174314e-01 -6.28315029e-04
-6.40399396e-01 -3.98920059e-01 6.98193669e-01 -1.97233424e-01
-1.73395395e+00 -6.02082291e-04 6.57817870e-02 -1.79314300e-01
-3.14068258e-01 -6.58668950e-02 -3.77817065e-01 -3.40496182e-01
4.95924532e-01 5.14052570e-01 -6.59063272e-03 -5.11795580e-01
1.11680552e-01 7.94392943e-01 4.60802168e-01 -1.68032229e-01
8.98284912e-01 1.05364764e+00 2.87764043e-01 -1.08965445e+00
-9.05204594e-01 -6.64956093e-01 -1.22255616e-01 -5.78088522e-01
8.04918110e-01 -1.15105522e+00 -8.96566570e-01 6.79510236e-01
-8.91972601e-01 -1.75256461e-01 -1.36400625e-01 8.12533736e-01
-1.29677534e-01 7.14595616e-01 -7.78360128e-01 -9.97301400e-01
-5.96494615e-01 -9.02761102e-01 9.26714957e-01 5.69763958e-01
6.78520441e-01 -6.24049962e-01 -9.94901061e-02 6.21941149e-01
4.12912488e-01 1.58748224e-01 5.88453114e-01 6.48620605e-01
-8.79954219e-01 -1.59658670e-01 -7.19452024e-01 7.15189040e-01
2.35391453e-01 -1.74590349e-01 -1.27428389e+00 -3.24478865e-01
3.02336723e-01 -1.03002809e-01 9.93085146e-01 4.08206105e-01
1.17400301e+00 -5.26738353e-02 -1.59978539e-01 8.66354823e-01
1.99763060e+00 6.25755340e-02 9.60587740e-01 3.06457818e-01
5.22755802e-01 5.94053447e-01 9.73162472e-01 3.91693056e-01
-3.34703759e-03 6.79373384e-01 7.78159976e-01 -6.95089638e-01
-5.07147372e-01 1.47505045e-01 4.06538069e-01 4.03730869e-01
-2.89054811e-01 -3.26622903e-01 -3.56970787e-01 3.13827366e-01
-1.40890729e+00 -1.11825728e+00 -7.94172466e-01 2.55028462e+00
8.81080925e-01 -2.68454909e-01 -3.44299525e-01 8.28743502e-02
8.47296774e-01 3.16622891e-02 -3.60644400e-01 2.93599427e-01
-4.34457958e-01 2.03361943e-01 7.08442748e-01 6.79677427e-01
-7.64814913e-01 4.33811069e-01 5.21375895e+00 8.09277773e-01
-1.20893407e+00 1.52963698e-01 4.95060265e-01 -2.62916684e-01
-2.57355183e-01 3.82491127e-02 -5.87177575e-01 8.02422941e-01
4.28873330e-01 8.62020329e-02 4.77545887e-01 1.50047734e-01
6.71502054e-01 -6.34878278e-01 -4.73484278e-01 1.13542986e+00
-3.62030067e-03 -7.37922192e-01 -2.50322878e-01 8.98964927e-02
6.10168159e-01 -1.13896586e-01 1.68097705e-01 -4.87970859e-01
-2.77218282e-01 -5.57768226e-01 3.42685461e-01 7.03007042e-01
1.17710233e+00 -6.94245636e-01 6.17465377e-01 2.31304005e-01
-9.83645678e-01 -1.76711023e-01 -5.31267643e-01 -1.39860973e-01
1.54879242e-01 1.01808667e+00 -2.79936969e-01 7.06265688e-01
8.84735048e-01 8.00731897e-01 -4.38603044e-01 1.08520424e+00
-5.24949372e-01 5.07928550e-01 -2.66722083e-01 3.94598931e-01
-2.90079504e-01 -9.41875100e-01 6.02233589e-01 9.10878778e-01
3.11722219e-01 5.18968999e-01 1.74069360e-01 1.03388953e+00
7.18932599e-02 -5.99868894e-02 -6.04030788e-01 2.67633319e-01
1.12467259e-01 1.57770395e+00 -5.56205809e-01 7.95088485e-02
-7.44833350e-01 9.71339107e-01 -4.17653531e-01 6.87995672e-01
-6.35521889e-01 -6.86822355e-01 4.37819749e-01 4.25186366e-01
-1.23222485e-01 -1.79351211e-01 -2.84260124e-01 -1.11460531e+00
4.44893062e-01 -5.07973850e-01 2.43571863e-01 -1.27035022e+00
-1.12901771e+00 2.96340436e-01 -1.85559466e-01 -1.45806372e+00
7.07467198e-01 -4.53041106e-01 -5.11755347e-01 1.12120998e+00
-2.11771560e+00 -1.15305030e+00 -9.24421847e-01 8.47473025e-01
2.78686881e-01 5.01369894e-01 4.51439142e-01 5.46073735e-01
-4.56626564e-01 1.09353885e-01 5.81348062e-01 -1.58716351e-01
9.27360594e-01 -5.70553899e-01 -4.46560621e-01 1.25717616e+00
-4.32094634e-01 4.95213330e-01 5.80494165e-01 -5.06482780e-01
-1.51365685e+00 -9.39638495e-01 2.51174718e-01 -2.24354770e-03
2.47386530e-01 -3.45989913e-01 -7.49152064e-01 1.85457215e-01
3.71097863e-01 1.68130711e-01 5.24059534e-01 -6.23205125e-01
-3.36740091e-02 -5.04035413e-01 -1.24764669e+00 5.31527400e-01
9.26965058e-01 -4.71343219e-01 -1.64260760e-01 2.33248755e-01
4.67800051e-01 -6.53937310e-02 -5.07504642e-01 6.12154841e-01
6.77153289e-01 -1.30624628e+00 1.13644755e+00 5.01193941e-01
4.68419015e-01 -7.17646539e-01 -6.91874027e-02 -9.09995139e-01
-8.29316750e-02 -4.05098706e-01 3.41560304e-01 1.44707298e+00
-8.19992572e-02 -1.08076429e+00 5.20302713e-01 1.61974594e-01
6.13553897e-02 -4.99440819e-01 -5.28025508e-01 -7.46017218e-01
-5.61499536e-01 -1.04581863e-01 4.90497649e-02 6.48647547e-01
-4.53227490e-01 7.64531121e-02 -4.21925217e-01 4.61512566e-01
1.14051545e+00 4.94234622e-01 7.45339215e-01 -1.01415348e+00
-2.72078484e-01 2.09336534e-01 -1.09397851e-01 -8.90282750e-01
-2.12776735e-01 -6.35672152e-01 1.76342562e-01 -1.57541645e+00
5.23113251e-01 -2.46659711e-01 -1.68899432e-01 2.40340158e-01
-3.28174084e-01 7.34452724e-01 2.24814177e-01 3.96532774e-01
-3.84233117e-01 6.47204995e-01 1.47005630e+00 9.96123906e-03
-2.79338598e-01 -1.42593741e-01 -5.83072722e-01 5.97087622e-01
7.28977203e-01 -3.81095558e-01 -3.70691836e-01 -3.66550833e-01
3.09433043e-02 -1.43584073e-01 6.14358544e-01 -1.13034046e+00
6.48199320e-02 -3.41845095e-01 5.88558257e-01 -2.76020765e-01
5.54812133e-01 -1.04678404e+00 3.19706976e-01 4.84892845e-01
7.64139518e-02 -7.21258342e-01 4.55208197e-02 7.11710513e-01
-1.80937663e-01 -2.56516248e-01 1.17180848e+00 -2.38732472e-01
-6.08688653e-01 1.29191160e-01 5.76801822e-02 -2.09533304e-01
1.11893296e+00 -4.06797409e-01 -8.74307752e-01 -2.63753891e-01
-1.02608025e-01 -1.85981825e-01 9.20600235e-01 -2.88102566e-03
6.66032493e-01 -9.69378531e-01 -4.77884501e-01 3.71027946e-01
3.15112591e-01 -2.90527135e-01 7.07915127e-01 1.04677463e+00
-6.23100877e-01 3.41859646e-02 -2.92166203e-01 -4.82181340e-01
-1.62654889e+00 4.41772103e-01 3.05000335e-01 8.63964781e-02
-8.91799271e-01 5.26251316e-01 5.74136853e-01 1.95098504e-01
-1.24263637e-01 -1.46940008e-01 -1.05484858e-01 -4.35964406e-01
6.85190618e-01 4.70031023e-01 -3.18573564e-02 -5.17719984e-01
-2.29291469e-02 9.18220460e-01 3.33479911e-01 1.47469833e-01
1.41995180e+00 -7.65949488e-01 -1.80212930e-01 2.97868490e-01
1.03726411e+00 3.37129235e-01 -1.29242063e+00 -2.60182232e-01
-8.41108143e-01 -1.22782016e+00 5.06651044e-01 -7.34904706e-01
-1.03674889e+00 8.68060827e-01 8.37524533e-01 -4.85479496e-02
1.68327320e+00 -2.70880908e-01 9.02241826e-01 -1.11011095e-01
3.64524812e-01 -9.60017860e-01 1.34045482e-01 -1.06059171e-01
6.13896787e-01 -1.40043080e+00 1.31442010e-01 -7.91308045e-01
-2.26122379e-01 1.11762869e+00 4.86727327e-01 1.89602152e-01
3.97898167e-01 1.23887584e-01 1.54278696e-01 -1.91123299e-02
-3.15746546e-01 -5.59791446e-01 -5.49318716e-02 6.12781763e-01
3.36034745e-01 -1.95014000e-01 -4.57236886e-01 2.32460022e-01
4.18134183e-01 2.41350591e-01 7.09454060e-01 7.40043938e-01
-4.80121464e-01 -7.52324283e-01 -5.35727620e-01 3.10191244e-01
-6.61136627e-01 -1.89365163e-01 -7.17398187e-04 3.69565248e-01
2.46803597e-01 1.39686966e+00 -2.78629988e-01 5.82136512e-02
2.77513564e-01 -3.48777741e-01 5.16979635e-01 -4.31122750e-01
-2.84007750e-03 3.53032351e-01 -2.88855970e-01 -6.24111772e-01
-6.75617218e-01 -4.13657784e-01 -9.84025538e-01 -1.24225751e-01
-5.56072474e-01 -1.83198988e-01 6.76083326e-01 3.77333969e-01
3.10407639e-01 3.28706443e-01 8.09199393e-01 -7.27795005e-01
-3.67930233e-02 -7.30508327e-01 -9.36986327e-01 7.17018962e-01
5.63571036e-01 -6.01876199e-01 -8.44453752e-01 1.27172217e-01] | [10.746853828430176, -2.5354561805725098] |
464a905c-c586-4780-9fc8-3af8217d658d | forensic-dental-age-estimation-using-modified | 2208.09799 | null | https://arxiv.org/abs/2208.09799v1 | https://arxiv.org/pdf/2208.09799v1.pdf | Forensic Dental Age Estimation Using Modified Deep Learning Neural Network | Dental age is one of the most reliable methods to identify an individual's age. By using dental panoramic radiography (DPR) images, physicians and pathologists in forensic sciences try to establish the chronological age of individuals with no valid legal records or registered patients. The current methods in practice demand intensive labor, time, and qualified experts. The development of deep learning algorithms in the field of medical image processing has improved the sensitivity of predicting truth values while reducing the processing speed of imaging time. This study proposed an automated approach to estimate the forensic ages of individuals ranging in age from 8 to 68 using 1,332 DPR images. Initially, experimental analyses were performed with the transfer learning-based models, including InceptionV3, DenseNet201, EfficientNetB4, MobileNetV2, VGG16, and ResNet50V2; and accordingly, the best-performing model, InceptionV3, was modified, and a new neural network model was developed. Reducing the number of the parameters already available in the developed model architecture resulted in a faster and more accurate dental age estimation. The performance metrics of the results attained were as follows: mean absolute error (MAE) was 3.13, root mean square error (RMSE) was 4.77, and correlation coefficient R$^2$ was 87%. It is conceivable to propose the new model as potentially dependable and practical ancillary equipment in forensic sciences and dental medicine. | ['Yahya Dogan', 'Musa Atas', 'Cuneyt Ozdemir', 'Isa Atas'] | 2022-08-21 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-9.82192010e-02 2.14421690e-01 8.21372196e-02 -2.28893578e-01
-5.95921814e-01 1.64346084e-01 -3.60263675e-03 2.71563560e-01
-1.05360210e+00 7.50945032e-01 -1.58112749e-01 -3.81367952e-01
-3.31702858e-01 -1.02201366e+00 -3.21822524e-01 -7.71721303e-01
-2.37063035e-01 6.54871225e-01 4.94562425e-02 2.61542320e-01
4.91617829e-01 6.99559152e-01 -1.55836725e+00 -2.16860548e-01
1.01044083e+00 1.07507181e+00 3.23550165e-01 5.55688322e-01
9.71083343e-03 7.21953690e-01 -5.60011446e-01 -7.17514336e-01
-3.45552005e-02 2.18363598e-01 -7.25963354e-01 -1.77761674e-01
3.86084706e-01 -8.23054492e-01 -3.91915530e-01 9.42664266e-01
8.53591204e-01 5.81986911e-04 7.12757647e-01 -9.03464556e-01
-8.41194749e-01 5.16431272e-01 -5.60364962e-01 4.80576217e-01
8.14788118e-02 1.19079873e-01 3.98304194e-01 -5.60596347e-01
6.16780162e-01 1.14137268e+00 9.48198140e-01 8.05734277e-01
-5.86123884e-01 -8.38275850e-01 -3.85552704e-01 8.10397923e-01
-1.34290004e+00 -1.90877780e-01 3.83976042e-01 -5.40828168e-01
5.36352158e-01 -3.55153978e-02 8.66209209e-01 8.15614104e-01
4.79885370e-01 3.83523971e-01 1.32450664e+00 -3.32109541e-01
3.50532383e-01 -2.98917502e-01 2.36358926e-01 7.00300574e-01
6.57297134e-01 1.02803715e-01 -2.35660449e-01 -4.00764644e-02
7.87117422e-01 -5.29888868e-02 1.78776875e-01 3.82670909e-01
-5.31932652e-01 1.00330746e+00 3.24743986e-01 4.44898576e-01
-6.37231112e-01 2.83758361e-02 3.61296237e-01 8.76295790e-02
4.73679662e-01 2.11668059e-01 -1.85971305e-01 5.14019327e-03
-1.05443525e+00 7.92663768e-02 2.63558745e-01 3.00706744e-01
2.91103631e-01 1.12086922e-01 3.68486792e-01 9.98187065e-01
4.14753884e-01 6.53598964e-01 7.31656134e-01 -1.15098488e+00
-3.76238562e-02 5.89836836e-01 -5.04879475e-01 -9.05322134e-01
-5.23320436e-01 -2.00157445e-02 -8.01532567e-01 2.21676707e-01
7.23167479e-01 1.72524173e-02 -1.51235878e+00 1.31648684e+00
3.26021522e-01 -9.48922113e-02 -1.64800748e-01 5.01457214e-01
9.79328692e-01 4.27798212e-01 4.13684517e-01 -5.13385125e-02
1.48043644e+00 -2.38592476e-01 -4.62518066e-01 -1.09946467e-01
4.11845505e-01 -6.07070804e-01 5.15531361e-01 5.56612313e-01
-1.19527221e+00 -5.68304777e-01 -8.88752639e-01 -8.65447745e-02
-3.18191588e-01 1.26269117e-01 7.87580669e-01 8.29167783e-01
-1.14490616e+00 8.69291365e-01 -9.93259192e-01 -2.59351641e-01
5.63599825e-01 7.27207541e-01 -5.07975280e-01 -2.57371634e-01
-1.01386333e+00 1.05368960e+00 3.33254993e-01 5.47366619e-01
-6.58145845e-01 -6.69718087e-01 -6.18588686e-01 -2.19520211e-01
-2.28668168e-01 -7.01609433e-01 9.91506159e-01 -5.25806189e-01
-1.22698557e+00 1.25056076e+00 3.06211621e-01 -4.72474009e-01
7.32979953e-01 -2.03483433e-01 -6.21255517e-01 6.27067506e-01
1.61359563e-01 7.74057090e-01 4.07677859e-01 -8.77240181e-01
-5.60908079e-01 -1.06580997e+00 -3.14142048e-01 -3.32461387e-01
-2.93524742e-01 1.12178892e-01 -3.33879441e-01 -3.08588147e-01
5.40555596e-01 -8.12320173e-01 -2.49920711e-01 3.16714734e-01
-2.06302702e-02 -4.14841324e-01 2.94181406e-01 -1.39619517e+00
1.07313991e+00 -1.97133601e+00 -2.63860077e-01 4.59023237e-01
3.27181786e-01 3.84067833e-01 1.76216856e-01 -8.86829346e-02
-7.84397125e-02 -1.37133077e-01 -2.53836513e-01 -5.84469847e-02
-6.02004170e-01 1.96084842e-01 5.14226794e-01 7.02431977e-01
-4.66378838e-01 4.22946841e-01 -6.59387410e-01 -1.06113446e+00
3.59688252e-01 7.40417659e-01 -4.81533498e-01 1.95606649e-02
3.89028460e-01 3.21648359e-01 -5.08655012e-01 1.00770450e+00
7.77799964e-01 -7.65935332e-02 7.25701498e-03 -1.89558432e-01
1.31580532e-01 -2.00949326e-01 -1.08915818e+00 1.44378138e+00
-6.18258536e-01 4.84789431e-01 1.48139298e-01 -8.24444532e-01
1.08125305e+00 3.69712293e-01 7.96608150e-01 -9.36356604e-01
6.05811119e-01 5.22861302e-01 1.95113897e-01 -9.10523653e-01
3.94777775e-01 -1.50491714e-01 2.85850435e-01 2.29110509e-01
7.79307708e-02 1.98078021e-01 2.74338841e-01 -2.80064326e-02
7.39877701e-01 -2.80370265e-01 9.89661664e-02 -1.15691781e-01
5.96426308e-01 -5.06663978e-01 4.64081019e-01 4.09387201e-01
-4.28665072e-01 5.31471193e-01 -6.46444187e-02 -7.17365801e-01
-1.05427659e+00 -1.26696789e+00 -4.28389341e-01 5.18023252e-01
-2.93391615e-01 2.35501960e-01 -9.20479357e-01 -1.80891305e-01
2.76161004e-02 5.35582185e-01 -7.12737620e-01 -5.34798317e-02
-8.42732012e-01 -8.80266905e-01 5.13562083e-01 6.49134815e-01
5.83957255e-01 -1.33118343e+00 -5.54914296e-01 2.66640306e-01
-8.41639936e-02 -8.73331606e-01 7.05316961e-02 -1.91246957e-01
-1.43526280e+00 -1.35647452e+00 -1.08688223e+00 -7.16672659e-01
8.05979371e-01 -4.23596859e-01 7.26360857e-01 4.17546660e-01
-5.38413107e-01 2.24015355e-01 -5.09031154e-02 -3.54515672e-01
-4.82133776e-01 1.56572357e-01 2.02024840e-02 -3.19967717e-01
6.00367188e-01 -6.60260022e-01 -9.29689765e-01 -1.83222070e-01
-7.95431137e-01 -5.05485117e-01 6.83717012e-01 5.51422298e-01
3.66502941e-01 -3.74994099e-01 7.90179431e-01 -8.00330222e-01
5.22930026e-01 -2.88237333e-01 -5.01679599e-01 8.96322355e-02
-1.01707220e+00 -3.63633782e-02 2.65779793e-01 -3.98144960e-01
-9.45730567e-01 -1.66429043e-01 -6.74403250e-01 -8.69348943e-02
-1.37474492e-01 3.06351304e-01 9.90104303e-02 6.45873025e-02
6.15638614e-01 1.71195880e-01 2.73794502e-01 -6.09683335e-01
-4.96075582e-03 9.33863521e-01 9.29922283e-01 -5.38607597e-01
3.56522799e-01 5.27275920e-01 1.42844198e-02 -1.10400307e+00
-3.89027297e-01 -2.49827772e-01 -5.04064739e-01 -5.43706536e-01
1.17132914e+00 -8.98081362e-01 -9.73160326e-01 1.16660714e+00
-9.60204065e-01 6.19421266e-02 9.65543762e-02 8.54296982e-01
-2.37685651e-01 7.79670000e-01 -1.01468825e+00 -6.26702487e-01
-1.09827960e+00 -1.13478613e+00 6.59221232e-01 4.19104934e-01
-2.35508069e-01 -9.00340557e-01 -1.17528327e-01 8.03379238e-01
3.26290041e-01 2.75909126e-01 1.07966959e+00 -4.69072789e-01
-8.26857537e-02 -5.25665581e-01 -1.22159526e-01 6.38523340e-01
-3.83015238e-02 2.42164895e-01 -8.81195366e-01 -7.41336271e-02
-1.35181742e-02 -9.59400367e-03 7.22563744e-01 8.42502117e-01
1.48456681e+00 8.29505771e-02 -1.39160588e-01 5.54405451e-01
1.34985065e+00 8.26491296e-01 1.05640686e+00 7.10292459e-01
6.41426086e-01 5.80357492e-01 1.68341368e-01 1.84135616e-01
5.82707226e-01 1.69727251e-01 4.20987844e-01 1.09054416e-01
-1.89011529e-01 -2.25037653e-02 7.86090866e-02 1.10466516e+00
-6.84975684e-01 1.68481261e-01 -1.01012921e+00 7.18806088e-01
-1.09623480e+00 -8.35219085e-01 -1.33526772e-01 2.12415934e+00
6.09268486e-01 1.98105082e-01 1.16949640e-01 5.52410483e-01
1.09020674e+00 -2.30162710e-01 -4.86117750e-01 -4.69413757e-01
2.90903360e-01 6.39795423e-01 6.56035900e-01 2.22181499e-01
-7.31839418e-01 5.12341917e-01 6.82045555e+00 6.61478162e-01
-1.28705204e+00 -3.06717702e-03 7.47267663e-01 -7.56808221e-02
1.38082251e-01 -4.36372578e-01 -6.57915056e-01 7.33757555e-01
1.18325126e+00 7.26222470e-02 4.60604280e-02 1.10568726e+00
1.36665870e-02 -3.27528805e-01 -7.33864069e-01 9.87443864e-01
7.04635456e-02 -1.11168909e+00 -1.87094554e-01 2.75540203e-01
2.41897881e-01 -1.89343929e-01 1.38702974e-01 1.04845211e-01
1.28447920e-01 -9.23027933e-01 4.35077161e-01 5.18329263e-01
7.34776020e-01 -8.82525861e-01 8.68861258e-01 -8.78062695e-02
-7.79749691e-01 -2.67569125e-01 -3.33203048e-01 1.81004386e-02
3.07925969e-01 5.69315493e-01 -7.49297857e-01 2.27550402e-01
9.35642898e-01 2.13634789e-01 -5.85828066e-01 1.10545838e+00
-2.68735290e-01 4.18048263e-01 -3.34333062e-01 4.46121097e-02
4.28671986e-02 -1.37047455e-01 -4.59134718e-03 5.96892476e-01
6.70906961e-01 1.72934145e-01 -5.81763506e-01 2.92287707e-01
1.39683057e-02 2.34719500e-01 -1.50941208e-01 3.56994599e-01
6.20279551e-01 1.10922444e+00 -1.00840211e+00 -3.99008244e-01
-1.57823533e-01 3.03339511e-01 -5.71411140e-02 -2.70407408e-01
-6.32580042e-01 -1.75516084e-01 2.43957013e-01 4.73821312e-01
3.69218141e-02 4.10671271e-02 -4.21379983e-01 -7.67374694e-01
-2.76060700e-01 -7.55654454e-01 6.73217356e-01 -8.46904218e-01
-1.29121017e+00 6.09986246e-01 1.64043024e-01 -5.85204720e-01
-2.39143401e-01 -7.93907881e-01 -5.67786276e-01 6.25803649e-01
-1.02301860e+00 -1.03348327e+00 -2.71621346e-01 6.18229270e-01
3.13459605e-01 -3.55285376e-01 5.89428842e-01 8.46115112e-01
-6.04790926e-01 6.30725622e-01 1.50281817e-01 2.55087137e-01
3.34891617e-01 -1.08523560e+00 9.00888070e-02 7.29802966e-01
-4.45004791e-01 5.69407821e-01 5.70616841e-01 -7.73318172e-01
-8.96090508e-01 -6.17682099e-01 9.77638543e-01 2.89070513e-02
4.42551374e-01 2.55104542e-01 -7.54252791e-01 3.75089318e-01
-1.91576242e-01 -3.77030879e-01 6.99678242e-01 -4.57484573e-02
-3.00376058e-01 -2.08888963e-01 -1.65576887e+00 3.61969858e-01
5.94918489e-01 -3.73499542e-01 -6.60390615e-01 1.25319824e-01
2.78725147e-01 -2.38711014e-01 -1.34177399e+00 2.76438087e-01
9.68775511e-01 -8.39592993e-01 1.29482925e+00 -7.23049417e-02
5.66848338e-01 2.36435503e-01 4.57467400e-02 -7.03965604e-01
-1.76258804e-03 3.00821275e-01 -4.23633903e-02 9.54679072e-01
1.10726342e-01 -6.10202193e-01 1.14648330e+00 6.71544671e-01
-4.12389815e-01 -8.81469309e-01 -1.25806439e+00 -6.53075278e-01
4.06945348e-01 -3.99349630e-01 5.13185501e-01 5.72134793e-01
-6.10074997e-01 -1.73695222e-01 -1.80731878e-01 6.42464310e-02
8.64986956e-01 -3.95188034e-01 2.19408795e-01 -1.66075742e+00
-1.91835444e-02 -2.46224999e-01 -7.28555679e-01 -1.68940738e-01
-1.02720849e-01 -6.10485971e-01 -3.36703837e-01 -1.84912145e+00
-5.90336975e-03 -3.69287223e-01 -3.12617004e-01 3.02127093e-01
-6.45828471e-02 2.60662317e-01 8.03305060e-02 1.49138376e-01
2.80400902e-01 1.11818492e-01 1.34238565e+00 -1.14936702e-01
-4.99673933e-02 2.47735441e-01 -5.11466563e-01 1.02214742e+00
8.21987510e-01 -5.68538010e-01 -1.64074630e-01 -4.68910724e-01
6.54771402e-02 2.53259540e-02 1.71508580e-01 -1.30766797e+00
2.39151761e-01 2.38045707e-01 5.66286206e-01 -7.73771346e-01
4.71664011e-01 -6.91081226e-01 4.19093519e-01 1.06983078e+00
1.78852037e-01 1.75051525e-01 -6.09568208e-02 1.70213338e-02
-5.60989007e-02 -6.35592997e-01 1.01029861e+00 -3.23087782e-01
-8.62233877e-01 4.41692352e-01 -4.15905088e-01 -3.60304922e-01
9.48934615e-01 -6.47747278e-01 -3.32440019e-01 -1.36670649e-01
-1.08596539e+00 -1.73995614e-01 1.50445178e-01 1.11597382e-01
7.38098204e-01 -9.81063306e-01 -5.97581267e-01 -1.90716997e-01
-1.95699647e-01 -7.96690211e-02 9.14685667e-01 5.82001150e-01
-1.32326889e+00 3.94161679e-02 -7.21286297e-01 -4.27427381e-01
-1.40608144e+00 3.37948978e-01 1.40821591e-01 -2.39359438e-01
-4.96204942e-01 8.33213985e-01 -4.84146208e-01 -2.60195613e-01
1.32741421e-01 -1.62001118e-01 -5.10395944e-01 4.23601508e-01
4.75758046e-01 1.03306174e+00 1.73430517e-01 -7.37079263e-01
-3.61640215e-01 9.10867333e-01 -3.87982130e-01 8.51482749e-02
1.75916469e+00 8.37530047e-02 -1.63868248e-01 -8.09151605e-02
1.14483881e+00 -4.39135075e-01 -6.79565966e-01 1.99095368e-01
-6.63346723e-02 -2.52679914e-01 2.44597778e-01 -4.98008639e-01
-1.63797534e+00 1.03101456e+00 1.45133734e+00 -1.14042513e-01
1.26649845e+00 -5.06470203e-02 1.00276470e+00 1.76854044e-01
3.33328038e-01 -1.44688821e+00 -9.55996662e-02 4.85146604e-02
3.12601000e-01 -1.16927731e+00 3.17342848e-01 -1.26943767e-01
-1.27176031e-01 1.29067278e+00 5.82343042e-01 -1.19303666e-01
8.10335517e-01 -1.07847750e-01 3.10986847e-01 -4.86787915e-01
2.95074787e-02 2.33267739e-01 1.84943862e-02 8.92549634e-01
4.63712603e-01 -2.51209252e-02 -9.59588945e-01 4.29956168e-01
-5.37412286e-01 3.30348879e-01 5.36672294e-01 7.69195020e-01
-4.17951912e-01 -9.71258879e-01 -5.94475806e-01 6.35889888e-01
-9.84986544e-01 1.19816005e-01 3.12503368e-01 1.04154766e+00
3.80351722e-01 8.70090842e-01 2.65588731e-01 -1.98205784e-01
6.39452413e-02 -2.15331212e-01 5.30476511e-01 -1.39596358e-01
-5.39361358e-01 -2.20073670e-01 -8.30384344e-02 -2.28305921e-01
-5.19300342e-01 -6.65731013e-01 -1.13748837e+00 -7.27371812e-01
-2.78834641e-01 1.68450102e-02 1.11651468e+00 9.07787561e-01
9.57604218e-03 4.52647299e-01 3.19652230e-01 -5.40082395e-01
-2.93939620e-01 -1.25938523e+00 -7.38269925e-01 1.69441357e-01
-2.60246843e-01 -6.75266385e-01 -3.21467429e-01 1.18492320e-01] | [14.154430389404297, -1.773532509803772] |
d13717c8-cd0d-4920-8dc5-79bafcf8e681 | a-hybrid-event-detection-approach-for-non | 1903.09180 | null | http://arxiv.org/abs/1903.09180v1 | http://arxiv.org/pdf/1903.09180v1.pdf | A Hybrid Event Detection Approach for Non-Intrusive Load Monitoring | Non-Intrusive Load Monitoring (NILM) is a practical method to provide
appliance-level electricity consumption information. Event detection, as an
important part of event-based NILM methods, has a direct impact on the accuracy
of the ultimate load disaggregation results in the entire NILM framework. This
paper presents a hybrid event detection approach for relatively complex
household load datasets that include appliances with long transients, high
fluctuations, and/or near-simultaneous actions. The proposed approach includes
a base algorithm based on moving average change with time limit, and two
auxiliary algorithms based on derivative analysis and filtering analysis. The
structure, steps, and working principle of this approach are described in
detail. The proposed approach does not require additional information about
household appliances, nor does it require any training sets. Case studies on
different datasets are conducted to evaluate the performance of the proposed
approach in comparison with several existing approaches including log
likelihood ratio detector with maxima (LLD-Max) approach, active window-based
(AWB) approach, and generalized likelihood ratio (GLR) approach. Results show
that the proposed approach works well in detecting events in complex household
load datasets and performs better than the existing approaches. | [] | 2019-03-21 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 1.25649467e-01 -3.25152040e-01 8.99389908e-02 -3.15897524e-01
-6.41681910e-01 -4.51423258e-01 6.76680565e-01 6.86658204e-01
-2.90139187e-02 9.64455128e-01 6.27724528e-02 -2.13617146e-01
-3.56422037e-01 -1.03091240e+00 -2.89279427e-02 -8.70502532e-01
-2.44925484e-01 3.03544521e-01 2.24726632e-01 1.69627413e-01
3.12430095e-02 6.15276158e-01 -1.41974390e+00 -1.26369819e-01
9.56751347e-01 9.68607008e-01 2.30042621e-01 5.32064378e-01
1.92230359e-01 7.54975796e-01 -8.43638837e-01 2.87238181e-01
1.38302073e-02 -5.07167220e-01 -4.12247926e-01 1.75659627e-01
-5.88576913e-01 -5.67070186e-01 3.56263518e-01 8.64902556e-01
8.06219518e-01 2.40811631e-01 8.45825970e-01 -1.58398509e+00
3.65078226e-02 6.86126351e-01 -7.56141424e-01 5.50431192e-01
8.52123082e-01 1.18217781e-01 3.44925433e-01 -2.21991941e-01
5.54508083e-02 1.00778890e+00 7.55020380e-01 -5.01891136e-01
-1.35078335e+00 -6.62681520e-01 -5.25635853e-02 4.92685586e-01
-1.56120956e+00 -1.76991031e-01 7.35982776e-01 -2.58241504e-01
1.50004518e+00 6.01853728e-01 4.39698875e-01 4.48418409e-01
2.12320060e-01 6.25285685e-01 1.29749596e+00 -5.71054935e-01
5.56594968e-01 2.84521341e-01 2.95189172e-01 -1.32158220e-01
5.51344275e-01 -1.05249725e-01 3.20217200e-03 -5.40034175e-01
7.75728524e-02 1.41737834e-01 -1.85335919e-01 3.08143735e-01
-5.80756128e-01 7.12426901e-01 -2.60678083e-01 6.64524853e-01
-8.99543464e-01 -3.53200346e-01 4.50573981e-01 3.24846730e-02
5.20492375e-01 -4.08860803e-01 -3.54218483e-01 -1.90919787e-01
-1.24891889e+00 1.23337783e-01 9.28548515e-01 7.32520759e-01
4.03231382e-01 2.22733185e-01 -3.55296284e-01 2.98358947e-01
5.37262201e-01 4.37250823e-01 4.91023093e-01 -1.77408129e-01
3.04335684e-01 7.19936311e-01 3.21821004e-01 -6.32518888e-01
-7.89614022e-01 -3.63555476e-02 -7.44562149e-01 9.01262164e-02
4.52814437e-02 -4.61137801e-01 -3.61432046e-01 1.44434988e+00
5.92077136e-01 1.75669283e-01 -4.89183292e-02 9.30253603e-03
4.20971453e-01 1.07048583e+00 3.26963156e-01 -1.06411445e+00
1.22320580e+00 -6.30319398e-03 -1.30021083e+00 6.34926379e-01
3.08258772e-01 -1.08995664e+00 4.64146912e-01 4.31706846e-01
-9.80243862e-01 -2.24474832e-01 -1.06199825e+00 6.46450818e-01
-6.53493226e-01 6.04813062e-02 2.80580610e-01 8.25653791e-01
-6.39204264e-01 5.40881336e-01 -8.39890361e-01 -9.29290116e-01
1.68800861e-01 4.20477271e-01 1.95292205e-01 3.96349937e-01
-9.50481117e-01 8.56875479e-01 8.63755226e-01 -6.49968460e-02
-3.98070574e-01 -5.41596889e-01 -7.13229120e-01 4.02827971e-02
-3.99239212e-02 1.00314990e-01 1.13812292e+00 -2.07631484e-01
-1.34180272e+00 1.40847236e-01 -1.30888805e-01 -6.51111662e-01
5.92401564e-01 6.62381873e-02 -8.37108552e-01 5.37106395e-02
4.80162352e-02 -1.60802826e-01 3.39655071e-01 -7.01106250e-01
-9.94043529e-01 -3.79909664e-01 -4.45795864e-01 3.54596712e-02
9.93832797e-02 2.60900199e-01 5.82333326e-01 -5.41701496e-01
-1.66045293e-01 -3.44054013e-01 2.42372885e-01 -7.86717355e-01
-3.99552673e-01 -8.85083020e-01 1.28350675e+00 -9.52101290e-01
1.58148932e+00 -1.97511435e+00 -1.06166148e+00 5.01773834e-01
-7.35306025e-01 2.91498005e-01 5.40995240e-01 1.15490520e+00
-3.05199564e-01 -3.55986536e-01 -1.91550076e-01 4.99278195e-02
3.85672718e-01 5.62473983e-02 -9.24752504e-02 7.41805673e-01
7.87405297e-03 5.87448835e-01 -8.04558277e-01 -4.58265424e-01
1.07886291e+00 5.01552105e-01 2.15895787e-01 2.11357906e-01
-4.48215567e-02 2.12644100e-01 -2.96007812e-01 6.45874858e-01
8.33928049e-01 1.63976371e-01 1.89718753e-01 -4.79902387e-01
-4.95906800e-01 2.70903707e-01 -1.98130167e+00 7.06978142e-01
-3.32118481e-01 3.68877113e-01 -2.13536099e-01 -1.21181440e+00
8.07637274e-01 1.06525278e+00 7.92243063e-01 -5.22129834e-01
3.11717510e-01 1.38351873e-01 -2.72752136e-01 -5.89694262e-01
-2.00370867e-02 1.55721784e-01 2.41049021e-01 6.01218700e-01
-4.38232832e-02 2.97754079e-01 6.94836557e-01 -2.45578468e-01
9.78787601e-01 1.27643168e-01 1.20741212e+00 -4.41720992e-01
7.17628002e-01 -3.59690934e-01 6.73048317e-01 4.25853699e-01
-2.92874753e-01 -2.13511616e-01 2.71615803e-01 -5.64658232e-02
-6.06269121e-01 -9.24110353e-01 -5.10810256e-01 5.06282985e-01
-1.30833939e-01 -1.53471276e-01 -6.05772793e-01 -1.14681870e-01
-7.45846331e-02 1.54472804e+00 -1.93175912e-01 8.79274085e-02
-4.26574528e-01 -1.72636533e+00 7.58500919e-02 3.12958211e-01
8.75564754e-01 -1.17909789e+00 -1.05501938e+00 8.40414464e-01
-8.47334638e-02 -7.69404590e-01 -9.77461487e-02 5.54629683e-01
-7.83187866e-01 -1.19896245e+00 -1.36127502e-01 -5.28980494e-01
5.41181564e-01 -1.17488638e-01 1.03978860e+00 -4.30999726e-01
-6.17128432e-01 3.30484331e-01 -4.09614623e-01 -7.87914634e-01
-3.97542417e-01 -1.69458479e-01 -9.88751426e-02 1.03344910e-01
9.62833703e-01 -1.02725458e+00 -5.78464150e-01 2.31326535e-01
-9.09138322e-01 -6.07168674e-01 4.21037316e-01 1.98220327e-01
4.35089380e-01 9.21458900e-01 1.25433087e+00 -6.89077199e-01
7.65736103e-01 -7.18096673e-01 -1.07061696e+00 2.58331776e-01
-1.02913916e+00 -4.05181885e-01 5.05714595e-01 -4.23267126e-01
-1.23078048e+00 1.82889238e-01 -1.95538536e-01 4.50368166e-01
-6.87500238e-01 -4.94124694e-03 -5.06680965e-01 4.48147953e-01
1.81628630e-01 2.85974830e-01 -5.41178763e-01 -7.90783167e-01
-9.26185846e-02 7.84332812e-01 4.24552768e-01 -1.27029970e-01
5.12931883e-01 2.25396395e-01 -2.55382229e-02 -8.80909562e-01
-2.02615317e-02 -6.31183803e-01 -5.31745076e-01 -1.55037016e-01
9.44418728e-01 -7.63291955e-01 -1.14607430e+00 7.46557951e-01
-8.85441601e-01 -2.04597637e-02 -4.47632253e-01 6.18345261e-01
-3.28296810e-01 2.20032141e-01 -5.43075025e-01 -1.63017201e+00
-5.84060669e-01 -8.15093219e-01 6.23640656e-01 3.80612016e-01
-5.59252203e-01 -1.10605967e+00 -1.85336601e-02 -1.81402922e-01
4.51391071e-01 9.42615926e-01 9.50962186e-01 -9.89050388e-01
-1.95883840e-01 -6.25973523e-01 1.31831378e-01 1.65097609e-01
8.25643778e-01 -7.60072991e-02 -9.29653287e-01 -4.34509844e-01
3.58420998e-01 2.75295228e-01 1.36104569e-01 5.85059047e-01
5.77802598e-01 -4.79799658e-01 -4.76648867e-01 6.40067756e-02
1.98223853e+00 8.53451371e-01 5.66459179e-01 1.16503641e-01
-5.07744141e-02 1.42039925e-01 5.15511453e-01 1.07700455e+00
3.77507776e-01 4.25314695e-01 3.33970666e-01 -7.06813708e-02
2.60545611e-01 -9.69868600e-02 5.23392856e-01 5.21730959e-01
2.29641140e-01 -5.63157737e-01 -1.99722871e-01 5.93194366e-01
-1.85668170e+00 -1.31123161e+00 -2.46856093e-01 2.24056292e+00
7.18156636e-01 1.01688415e-01 5.19516170e-01 1.03406239e+00
9.29220021e-01 -7.55995512e-02 -5.81005991e-01 -5.91023028e-01
-2.44322587e-02 3.04487109e-01 6.08731687e-01 3.53147417e-01
-1.18709505e+00 -1.66074678e-01 6.00263214e+00 8.19597900e-01
-5.94123065e-01 3.65080088e-01 2.77686507e-01 1.99225381e-01
2.73185819e-01 -2.87931234e-01 -8.23567390e-01 8.55540991e-01
1.12481713e+00 -3.85658383e-01 2.53826469e-01 5.93721271e-01
9.88506734e-01 -9.81306076e-01 -1.11661327e+00 9.37065959e-01
-2.51880705e-01 -5.65405786e-01 -3.69561464e-01 1.66085124e-01
8.13792765e-01 -3.27017397e-01 -6.66338980e-01 1.10426713e-02
3.58096391e-01 -4.51088965e-01 2.85705239e-01 5.21522641e-01
5.62984087e-02 -1.00793529e+00 1.13751805e+00 4.48814332e-01
-1.70385003e+00 -7.30191544e-03 1.12827443e-01 -1.42386749e-01
5.55404186e-01 9.46476579e-01 -8.67641747e-01 7.91496396e-01
7.04819560e-01 2.76438326e-01 -2.81539440e-01 1.06534255e+00
-9.21797194e-03 1.06371117e+00 -1.03728473e+00 -3.30345407e-02
-2.40598083e-01 -2.06675842e-01 4.04820919e-01 1.22533739e+00
3.95620167e-01 2.77068824e-01 1.85717538e-01 7.26526558e-01
4.61235166e-01 3.18429828e-01 -4.13497984e-01 3.93592358e-01
6.02874875e-01 1.25435698e+00 -1.07009196e+00 -5.72795272e-01
-6.10515177e-01 7.33327508e-01 -6.12527490e-01 2.06251755e-01
-7.42750168e-01 -5.33368230e-01 1.36111245e-01 7.09710717e-02
3.80808353e-01 2.20535785e-01 5.00666797e-02 -4.32144910e-01
6.33371994e-03 -5.12147307e-01 8.84569287e-01 -3.35761845e-01
-1.30354679e+00 -4.73855399e-02 6.26515090e-01 -1.04724932e+00
-7.24677980e-01 3.85382138e-02 -1.34348428e+00 9.86380219e-01
-1.18212914e+00 -8.16288233e-01 -8.55657384e-02 7.87675440e-01
5.78191578e-01 2.72703916e-01 9.17889357e-01 6.27348840e-01
-7.02529371e-01 3.56397986e-01 4.98219579e-01 -1.77749053e-01
5.03149740e-02 -1.35199690e+00 -2.01350302e-01 1.00974369e+00
-3.96279931e-01 -2.30954699e-02 9.63927925e-01 -8.64721417e-01
-9.08814907e-01 -1.10310709e+00 9.27450180e-01 1.08520471e-01
2.90495038e-01 -1.68377504e-01 -7.02152550e-01 7.29090095e-01
6.07233524e-01 -5.04180908e-01 7.65973389e-01 -4.37447697e-01
6.74943805e-01 -4.55664247e-01 -1.96621525e+00 -1.38056532e-01
1.20428979e-01 -8.17777514e-02 -7.59584725e-01 4.87282306e-01
-5.18041924e-02 1.62814200e-01 -1.15656078e+00 5.14385462e-01
2.05170870e-01 -1.02321517e+00 5.54616451e-01 3.26996773e-01
-8.73880088e-01 -7.26301670e-01 -1.54121488e-01 -9.88814533e-01
-1.44353062e-01 -9.43047762e-01 -3.05997372e-01 2.10918593e+00
6.18527830e-02 -1.05084860e+00 -2.07238998e-02 4.94396299e-01
3.06840271e-01 -3.14152449e-01 -8.89676332e-01 -7.15241194e-01
-6.18340194e-01 -4.24607038e-01 1.09837472e+00 6.58180296e-01
1.94061115e-01 1.04901921e-02 -5.03141657e-02 5.32174766e-01
7.54080534e-01 -1.19807683e-02 2.75874257e-01 -1.20043063e+00
-2.46660486e-02 -1.28666401e-01 -5.71003675e-01 -6.16962947e-02
-4.38982725e-01 -4.41954255e-01 -1.99078754e-01 -1.70243216e+00
7.48847052e-02 1.06552728e-01 -5.55758357e-01 4.94360864e-01
-5.90935089e-02 2.03391999e-01 -6.70461431e-02 -6.57406226e-02
-1.51514709e-01 3.44026953e-01 2.57119667e-02 1.14924088e-01
-5.42784274e-01 4.67168689e-01 5.74017987e-02 8.10754299e-01
1.19856560e+00 -3.43423337e-01 -6.22309566e-01 4.96623576e-01
-1.46099672e-01 -1.17338732e-01 1.81335360e-01 -1.14273739e+00
-1.60444938e-02 -3.54024284e-02 7.73382485e-01 -1.63767660e+00
-1.74238980e-01 -1.11466801e+00 8.90969157e-01 7.14460611e-01
2.31202409e-01 5.95265210e-01 1.90167367e-01 2.68854439e-01
1.22077346e-01 -4.05492127e-01 7.28265464e-01 -7.23935291e-02
-5.10895014e-01 -2.50709534e-01 -6.36051536e-01 -4.99068499e-01
1.52382886e+00 -2.72892267e-01 4.50964533e-02 -2.11871073e-01
-7.45175719e-01 3.25256616e-01 1.46145439e-02 1.44535109e-01
-1.72147483e-01 -1.44593716e+00 -5.79052448e-01 9.36441794e-02
-2.37152427e-01 -3.47624511e-01 -5.09947501e-02 8.68428528e-01
-3.21255833e-01 5.17442584e-01 1.29750267e-01 -5.32213628e-01
-1.16186523e+00 6.35936558e-01 1.37757316e-01 -4.84960645e-01
-5.74921429e-01 -1.66684628e-01 -3.19680601e-01 2.42262796e-01
3.61179650e-01 -6.34055138e-01 -3.82347763e-01 4.94119763e-01
6.64424121e-01 1.14872932e+00 3.59755367e-01 -5.62367022e-01
-4.90373433e-01 4.30138111e-01 3.48681062e-01 8.06858018e-02
1.24880505e+00 -5.34058213e-01 -2.10013077e-01 8.44571352e-01
9.72240210e-01 -2.25283876e-01 -7.45715499e-01 1.40099302e-01
2.86070555e-01 7.18633160e-02 -2.16004048e-02 -9.18960035e-01
-5.69470644e-01 3.20819616e-01 1.25094342e+00 8.56672525e-01
1.67231095e+00 -2.16822445e-01 6.71178341e-01 -3.30192372e-02
3.82548481e-01 -1.40145981e+00 -8.78277838e-01 -3.56996208e-01
5.14932156e-01 -1.12247288e+00 3.00619632e-01 -1.52001023e-01
1.82754189e-01 8.72402489e-01 1.71282023e-01 8.08082744e-02
1.20372963e+00 4.16969180e-01 -3.58386129e-01 1.46098465e-01
-5.97520947e-01 -4.25494015e-01 -2.84025311e-01 7.18898356e-01
4.07338917e-01 3.11199218e-01 -9.49261725e-01 4.26588625e-01
2.56038439e-02 3.31295162e-01 1.00191712e-01 1.17822087e+00
-3.48244041e-01 -8.75497878e-01 -7.21492112e-01 7.30689824e-01
-7.47533441e-01 2.85988361e-01 3.05769205e-01 1.00516737e+00
5.59181929e-01 1.55409896e+00 2.23218277e-01 1.34458274e-01
5.23847997e-01 4.24055964e-01 2.70824492e-01 -1.34831548e-01
-4.88673031e-01 2.52054691e-01 -6.98725954e-02 -3.23223233e-01
-5.78423738e-01 -1.08071613e+00 -1.21248269e+00 -3.79631609e-01
-5.73187888e-01 2.71462888e-01 5.26452303e-01 1.13335884e+00
-5.87025620e-02 4.53374207e-01 1.03736484e+00 -7.39688635e-01
-4.05352145e-01 -1.34471667e+00 -9.08390999e-01 4.12329674e-01
2.61738181e-01 -4.97295439e-01 -4.88845289e-01 1.57501072e-01] | [6.0070013999938965, 2.583630084991455] |
ad2dee00-524f-4499-8e0b-52828a673250 | preserving-privacy-in-domain-transfer-of | 2306.06503 | null | https://arxiv.org/abs/2306.06503v1 | https://arxiv.org/pdf/2306.06503v1.pdf | Preserving privacy in domain transfer of medical AI models comes at no performance costs: The integral role of differential privacy | Developing robust and effective artificial intelligence (AI) models in medicine requires access to large amounts of patient data. The use of AI models solely trained on large multi-institutional datasets can help with this, yet the imperative to ensure data privacy remains, particularly as membership inference risks breaching patient confidentiality. As a proposed remedy, we advocate for the integration of differential privacy (DP). We specifically investigate the performance of models trained with DP as compared to models trained without DP on data from institutions that the model had not seen during its training (i.e., external validation) - the situation that is reflective of the clinical use of AI models. By leveraging more than 590,000 chest radiographs from five institutions, we evaluated the efficacy of DP-enhanced domain transfer (DP-DT) in diagnosing cardiomegaly, pleural effusion, pneumonia, atelectasis, and in identifying healthy subjects. We juxtaposed DP-DT with non-DP-DT and examined diagnostic accuracy and demographic fairness using the area under the receiver operating characteristic curve (AUC) as the main metric, as well as accuracy, sensitivity, and specificity. Our results show that DP-DT, even with exceptionally high privacy levels (epsilon around 1), performs comparably to non-DP-DT (P>0.119 across all domains). Furthermore, DP-DT led to marginal AUC differences - less than 1% - for nearly all subgroups, relative to non-DP-DT. Despite consistent evidence suggesting that DP models induce significant performance degradation for on-domain applications, we show that off-domain performance is almost not affected. Therefore, we ardently advocate for the adoption of DP in training diagnostic medical AI models, given its minimal impact on performance. | ['Daniel Truhn', 'Georgios Kaissis', 'Sven Nebelung', 'Christiane Kuhl', 'Peter Isfort', 'Marwin Saehn', 'Teresa Nolte', 'Mahshad Lotfinia', 'Soroosh Tayebi Arasteh'] | 2023-06-10 | null | null | null | null | ['image-classification-with-dp', 'medical-diagnosis', 'domain-generalization', 'specificity'] | ['computer-vision', 'medical', 'methodology', 'natural-language-processing'] | [ 2.51400799e-01 6.01667166e-01 -1.37468085e-01 -4.67466474e-01
-8.94248903e-01 -6.36622310e-01 3.01675677e-01 5.00337005e-01
-7.36984789e-01 6.72338903e-01 2.53318012e-01 -8.38386476e-01
-3.37319434e-01 -6.71455026e-01 -6.87660933e-01 -4.91057843e-01
8.77509266e-02 6.83754206e-01 -1.77353099e-01 5.57769716e-01
-2.01974198e-01 4.27402794e-01 -6.99781239e-01 3.43869239e-01
9.74263370e-01 7.35463202e-01 -6.83181763e-01 4.29879040e-01
2.92544216e-01 7.73842633e-01 -5.76371551e-01 -7.73719728e-01
5.62881052e-01 -3.09175670e-01 -6.28375173e-01 -2.18039826e-01
4.02724355e-01 -8.81826162e-01 -1.85912877e-01 4.89917040e-01
5.64328194e-01 -4.13733393e-01 7.71944463e-01 -1.21759999e+00
-5.97124875e-01 4.23858374e-01 -2.13982210e-01 5.05483486e-02
2.76591629e-01 4.86271590e-01 7.02696502e-01 -2.65137225e-01
8.54464233e-01 6.03847742e-01 1.15275371e+00 4.74301457e-01
-1.42553115e+00 -7.43394136e-01 -3.31662595e-01 -5.30942798e-01
-1.09396625e+00 -3.48099947e-01 2.85820186e-01 -6.09186471e-01
6.39599800e-01 3.34590524e-01 7.36700773e-01 9.94378924e-01
2.56838948e-01 3.27238917e-01 1.11302531e+00 -2.05954611e-01
2.42920846e-01 5.53497255e-01 2.20239207e-01 4.68829930e-01
7.28523433e-01 1.23238660e-01 2.59836763e-02 -9.76387441e-01
5.44636130e-01 3.65584679e-02 -9.78809819e-02 -4.65062946e-01
-1.28119326e+00 6.34648383e-01 1.26495749e-01 1.97854549e-01
-4.95827675e-01 -4.01328683e-01 4.02039826e-01 3.35816860e-01
2.29301244e-01 6.14779592e-01 -5.25680244e-01 -4.49198298e-02
-8.93203974e-01 2.54062414e-01 1.01117814e+00 6.86555564e-01
9.59138796e-02 -3.41524273e-01 -1.97073177e-01 5.70832431e-01
-1.08419269e-01 3.30155283e-01 3.81276757e-01 -1.01893413e+00
5.17133296e-01 6.23482466e-01 2.04669267e-01 -9.19202983e-01
-5.13295114e-01 -2.69183517e-01 -8.36229622e-01 -1.41367659e-01
7.42360771e-01 -6.42200947e-01 -7.63382018e-01 1.71766222e+00
3.32224548e-01 -1.99485004e-01 1.42259732e-01 5.60125768e-01
4.57318634e-01 1.70431972e-01 5.89516222e-01 -1.01087444e-01
1.39434290e+00 -1.75080791e-01 -4.26328540e-01 -1.52679726e-01
1.01466966e+00 -2.37830475e-01 8.02352071e-01 1.51304498e-01
-9.37161565e-01 4.87708412e-02 -5.94811499e-01 6.80469200e-02
-1.96520329e-01 -2.17194840e-01 3.12296540e-01 9.64401066e-01
-9.44394231e-01 4.60097402e-01 -8.80626559e-01 -5.94438553e-01
9.11299467e-01 5.34670651e-01 -7.98801363e-01 -2.08997712e-01
-1.05583000e+00 8.11810553e-01 1.11400478e-01 -2.75224566e-01
-3.58129680e-01 -1.22657561e+00 -4.92712051e-01 9.59002450e-02
2.10188888e-03 -1.10018146e+00 9.66724157e-01 -8.14161897e-01
-7.09239841e-01 1.16146040e+00 1.84061974e-01 -7.41320074e-01
1.04034638e+00 -8.32216740e-02 -4.54747289e-01 2.16569215e-01
3.94960074e-03 6.02393925e-01 3.41427952e-01 -1.02388513e+00
-3.78565699e-01 -5.27961671e-01 -3.30992103e-01 2.61593349e-02
-5.15162170e-01 1.21587925e-01 1.25716805e-01 -5.15489817e-01
-2.43770733e-01 -1.13922620e+00 -3.61519665e-01 3.41276705e-01
-1.67618603e-01 2.81665504e-01 7.26815701e-01 -1.02143264e+00
1.15968394e+00 -2.37694144e+00 -6.32588565e-01 4.35262710e-01
5.08987248e-01 6.08749032e-01 2.15096757e-01 1.40841395e-01
-1.16396487e-01 3.69269431e-01 -6.39571011e-01 8.21107402e-02
-1.63547650e-01 8.49494264e-02 5.67034185e-02 4.89200294e-01
2.84280151e-01 9.02451992e-01 -5.86983025e-01 -7.08428085e-01
1.63057402e-01 2.94356704e-01 -9.30192947e-01 1.29907891e-01
2.15498462e-01 5.69547892e-01 -4.80277121e-01 6.52945161e-01
6.45787120e-01 -5.59788406e-01 4.86607015e-01 2.82393157e-01
1.45492613e-01 7.81285763e-02 -5.98872542e-01 1.02883720e+00
4.42469604e-02 3.40249658e-01 2.94682741e-01 -5.91539383e-01
7.90239573e-01 6.32478595e-01 8.37714016e-01 -5.11113405e-01
-3.05468533e-02 2.70668298e-01 4.13598061e-01 -4.26296830e-01
-6.81394860e-02 -4.15766478e-01 5.80774248e-02 5.83348811e-01
-3.61479372e-01 1.24169528e-01 -5.64570367e-01 1.32137779e-02
1.46293569e+00 -4.72744644e-01 4.42975134e-01 -2.80007601e-01
1.19385131e-01 2.77704507e-01 6.40068471e-01 7.70666480e-01
-6.26348615e-01 7.95525789e-01 7.46402681e-01 -3.17394525e-01
-1.10978711e+00 -1.06836140e+00 -6.84978127e-01 4.38892841e-01
-6.32048845e-01 -6.29973784e-02 -7.46849597e-01 -1.00195253e+00
6.31739378e-01 8.65113080e-01 -5.47224939e-01 -3.56644660e-01
-4.36300337e-01 -8.66863728e-01 1.02535558e+00 3.21287483e-01
4.18135226e-01 -7.35268056e-01 -1.02455294e+00 2.02386603e-01
3.17322575e-02 -9.39634085e-01 -3.24784815e-01 -3.66456360e-02
-9.38588500e-01 -1.08272147e+00 -1.02461207e+00 -3.85335088e-01
7.48360634e-01 -3.03319484e-01 1.05882192e+00 -7.66651407e-02
-3.57461065e-01 8.51601779e-01 7.18902722e-02 -5.93905389e-01
-5.78766823e-01 7.52325803e-02 -7.78269023e-02 -2.70931631e-01
8.11856389e-01 -4.57665145e-01 -8.49345684e-01 1.84912607e-01
-9.70720947e-01 -3.99796218e-01 7.06971824e-01 7.93428481e-01
2.46836796e-01 -6.04607224e-01 7.29710102e-01 -1.59339905e+00
6.21232092e-01 -7.56761014e-01 -3.02607775e-01 3.93336177e-01
-1.17510664e+00 -3.36849093e-01 5.43240726e-01 -4.52976912e-01
-9.90974665e-01 3.88615578e-02 1.72209054e-01 -4.37359452e-01
-3.51519972e-01 3.65098327e-01 1.45679951e-01 6.28708825e-02
7.59673536e-01 -1.89728916e-01 5.32382548e-01 -3.07113767e-01
-1.68675452e-01 9.97901678e-01 3.33872437e-01 -5.51115513e-01
4.97988194e-01 4.16827351e-01 -1.08164281e-01 -6.40793562e-01
-4.32065457e-01 -2.83105284e-01 -4.08614218e-01 2.43053988e-01
7.91522741e-01 -9.04565215e-01 -6.58785224e-01 2.29038149e-01
-5.26227593e-01 -1.70271873e-01 -4.85603750e-01 7.50237107e-01
-2.60771692e-01 2.83179849e-01 -4.40528482e-01 -7.30744123e-01
-5.16530871e-01 -7.97921181e-01 5.47087789e-01 -1.42890915e-01
-8.04925382e-01 -1.05427587e+00 1.58884913e-01 5.98950446e-01
4.37320232e-01 7.35177219e-01 1.26027107e+00 -1.32013988e+00
-2.54316896e-01 -6.80160701e-01 -2.31872216e-01 3.63482565e-01
3.28596681e-01 5.28509310e-03 -9.57877338e-01 -3.18354547e-01
1.32206395e-01 -1.98087677e-01 2.67671227e-01 4.87382621e-01
1.14233744e+00 -5.87023020e-01 -4.01624024e-01 5.58637500e-01
1.14837050e+00 4.52769637e-01 5.03395140e-01 3.36707950e-01
3.42902273e-01 6.35281801e-01 3.75920236e-01 6.12196505e-01
3.97154093e-01 2.46963412e-01 -1.27279058e-01 -3.30502629e-01
3.44419330e-01 -3.48590940e-01 -1.00552756e-02 1.76489934e-01
2.63733089e-01 1.57210976e-02 -1.46690750e+00 6.52949035e-01
-1.41639173e+00 -6.77041471e-01 9.18429941e-02 2.49166894e+00
8.25856090e-01 1.72758862e-01 3.73892486e-01 -1.72574028e-01
5.19291401e-01 -4.60943207e-02 -8.39340150e-01 -6.90016508e-01
-4.68335599e-02 1.66777477e-01 7.80412555e-01 -3.77109274e-02
-8.77406061e-01 1.65105730e-01 7.10861111e+00 -5.44851013e-02
-9.41055119e-01 -6.71693310e-02 1.09816539e+00 -1.61139086e-01
-3.52969170e-01 -4.25507799e-02 -1.68672740e-01 5.25610089e-01
1.40823412e+00 -5.55412769e-01 3.85590382e-02 9.67926741e-01
3.53403613e-02 -7.75691122e-02 -1.23377240e+00 4.37983900e-01
-2.16922089e-01 -1.18443704e+00 -1.45036891e-01 4.69470859e-01
6.16484463e-01 -1.05107583e-01 3.56816173e-01 1.20708324e-01
4.06294584e-01 -9.80498254e-01 1.77961066e-01 2.86950260e-01
1.14514208e+00 -5.59064567e-01 8.94120812e-01 2.92321026e-01
-4.05409813e-01 -2.37780452e-01 -4.13408726e-02 2.34005898e-01
-2.36379683e-01 5.94236493e-01 -1.44220054e+00 4.79629099e-01
6.66451454e-01 3.82409483e-01 -4.03014541e-01 8.24498713e-01
4.90613192e-01 8.87902737e-01 -5.25145948e-01 4.75020707e-01
1.11736275e-01 -1.49023235e-01 2.69170731e-01 1.19641781e+00
3.06396574e-01 5.02904773e-01 -3.50883901e-01 6.26229286e-01
-2.44178593e-01 1.03174120e-01 -1.02744901e+00 -3.92620891e-01
6.29765391e-01 7.53901422e-01 -2.14588165e-01 -2.47499064e-01
-6.95705295e-01 5.36956191e-01 6.14352711e-02 6.93902969e-02
-6.10752046e-01 -1.53644174e-01 7.08097041e-01 6.11831486e-01
-1.35867402e-01 3.16319466e-01 -6.67048573e-01 -7.89336205e-01
3.66826430e-02 -1.19054031e+00 1.03517318e+00 -3.56569946e-01
-1.27702510e+00 2.24152818e-01 -2.33669709e-02 -1.21863031e+00
-4.49767560e-01 -2.85626650e-01 -1.73316747e-01 9.05320227e-01
-1.17311013e+00 -9.05364990e-01 -5.34177423e-02 4.76015508e-01
-4.26717877e-01 3.77817154e-02 1.04833448e+00 3.41389894e-01
-4.85533297e-01 1.15982890e+00 3.41404140e-01 4.26201046e-01
1.01339889e+00 -8.76191378e-01 1.71664447e-01 5.78469574e-01
-4.19333130e-01 8.70368004e-01 3.17845255e-01 -8.09049010e-01
-1.07586682e+00 -1.08220422e+00 9.10512686e-01 -8.81228924e-01
3.04761082e-01 -4.37390013e-03 -1.06808901e+00 1.09173477e+00
-3.10532153e-01 -1.01762772e-01 1.10233212e+00 4.60924394e-02
-3.09442729e-01 -1.04298003e-01 -1.95220077e+00 3.53519171e-01
7.88807571e-01 -4.64687347e-01 -7.08651245e-01 1.72531679e-01
6.35639966e-01 -1.40130013e-01 -1.45703089e+00 5.74255586e-01
7.42070377e-01 -9.80684578e-01 9.06747818e-01 -8.60705614e-01
4.53055918e-01 1.70496196e-01 8.10585022e-02 -7.79207230e-01
-2.48793140e-01 -4.70651269e-01 2.57860184e-01 1.20579922e+00
5.41980982e-01 -1.13705552e+00 1.02873886e+00 1.78418517e+00
2.87464947e-01 -5.37340820e-01 -1.05427980e+00 -5.72100878e-01
4.87108052e-01 -3.00927274e-02 7.41636634e-01 1.49455798e+00
-7.75275305e-02 -4.60417926e-01 -1.44622445e-01 2.47430816e-01
6.47919536e-01 -2.09276259e-01 7.17569292e-01 -1.19143081e+00
-1.92350671e-01 -3.07008196e-02 -3.19950432e-01 -9.37829688e-02
-8.62815157e-02 -8.18307757e-01 -5.83269358e-01 -1.40677547e+00
3.19449008e-01 -7.23067045e-01 -4.74780351e-01 7.71223962e-01
-2.39791811e-01 -3.05484027e-01 1.80165544e-01 4.28697169e-01
1.87268376e-01 -2.56469566e-02 1.01099956e+00 2.26943523e-01
-1.88316688e-01 1.97772577e-01 -1.02310336e+00 3.68126512e-01
7.20825374e-01 -7.80475676e-01 -3.63713652e-01 -2.85760313e-01
-4.16425884e-01 3.01281929e-01 4.86558557e-01 -1.07224107e+00
1.09073468e-01 -9.16532055e-02 5.18241525e-01 -2.03314692e-01
6.03733324e-02 -1.24201107e+00 7.09184527e-01 9.25152719e-01
-5.53587317e-01 9.32293236e-02 3.82857621e-01 4.48780626e-01
-5.30818179e-02 1.91581622e-01 7.09242642e-01 -1.64345071e-01
1.19055994e-01 1.83778957e-01 -3.13849837e-01 3.19560468e-01
1.24762142e+00 -5.24368346e-01 -4.79545176e-01 -2.41525665e-01
-6.02335870e-01 1.96241707e-01 8.69824171e-01 -7.41791278e-02
2.68476188e-01 -7.47290194e-01 -7.46269464e-01 4.16456878e-01
3.12983423e-01 -2.83504929e-02 3.41977119e-01 8.90841544e-01
-5.84208190e-01 6.76176667e-01 -1.63854703e-01 -4.40133482e-01
-1.29657972e+00 6.27546012e-01 3.43968838e-01 -3.12583238e-01
-8.09422672e-01 4.47021455e-01 2.58474171e-01 -6.59866154e-01
1.95868611e-01 -3.05523753e-01 4.98338521e-01 -1.31813824e-01
1.88032642e-01 2.71887004e-01 1.06589682e-01 -1.76320106e-01
-5.36740065e-01 6.00296780e-02 -3.43953431e-01 -5.20509072e-02
1.37684464e+00 2.57052511e-01 9.34475064e-02 2.38754824e-01
1.12780809e+00 -3.70719507e-02 -1.14588308e+00 -2.53340602e-02
4.62267064e-02 -5.83671629e-01 -3.69730920e-01 -1.08626664e+00
-7.55655348e-01 6.03800714e-01 7.32287705e-01 1.93575118e-02
1.07767403e+00 -3.00050139e-01 6.08976543e-01 1.88317388e-01
3.06188881e-01 -7.69160211e-01 -5.97675443e-01 -2.22785518e-01
3.69150728e-01 -1.14151800e+00 2.88115680e-01 4.44682725e-02
-1.04532647e+00 4.97114986e-01 5.06374180e-01 1.97824910e-01
7.75366724e-01 1.97702363e-01 2.09465072e-01 8.52827542e-03
-8.11956525e-01 5.81971407e-01 -9.69854742e-02 7.50964880e-01
2.70789295e-01 3.16166013e-01 -3.04292768e-01 6.52010024e-01
-7.60079995e-02 5.46769977e-01 4.00608659e-01 1.24614418e+00
1.95793778e-01 -1.02674448e+00 -3.87956262e-01 1.14993703e+00
-7.18834758e-01 2.27346038e-03 -6.86481059e-01 1.04197097e+00
-1.81524809e-02 5.68227768e-01 2.48110026e-01 -2.30425015e-01
4.71506953e-01 4.56776172e-01 -1.43465355e-01 -4.27546769e-01
-1.15773785e+00 -4.25897330e-01 2.54518390e-01 -2.91683704e-01
-8.75454620e-02 -1.16374052e+00 -8.45646203e-01 -5.50624549e-01
1.45472452e-01 7.23318607e-02 3.38401824e-01 4.46757227e-01
9.05983150e-01 9.54697132e-02 4.92691219e-01 2.86356717e-01
-8.86192083e-01 -5.77886879e-01 -4.20642674e-01 6.02123797e-01
4.67171997e-01 -7.39094913e-02 -5.28854430e-01 -2.13888615e-01] | [6.343343734741211, 6.671206951141357] |
a27b229e-dd94-4355-abe6-b91f7e2f6835 | dinf-dynamic-instance-noise-filter-for | 2301.05565 | null | https://arxiv.org/abs/2301.05565v1 | https://arxiv.org/pdf/2301.05565v1.pdf | DINF: Dynamic Instance Noise Filter for Occluded Pedestrian Detection | Occlusion issue is the biggest challenge in pedestrian detection. RCNN-based detectors extract instance features by cropping rectangle regions of interest in the feature maps. However, the visible pixels of the occluded objects are limited, making the rectangle instance feature mixed with a lot of instance-irrelevant noise information. Besides, by counting the number of instances with different degrees of overlap of CrowdHuman dataset, we find that the number of severely overlapping objects and the number of slightly overlapping objects are unbalanced, which may exacerbate the challenges posed by occlusion issues. Regarding to the noise issue, from the perspective of denoising, an iterable dynamic instance noise filter (DINF) is proposed for the RCNN-based pedestrian detectors to improve the signal-noise ratio of the instance feature. Simulating the wavelet denoising process, we use the instance feature vector to generate dynamic convolutional kernels to transform the RoIs features to a domain in which the near-zero values represent the noise information. Then, soft thresholding with channel-wise adaptive thresholds is applied to convert the near-zero values to zero to filter out noise information. For the imbalance issue, we propose an IoU-Focal factor (IFF) to modulate the contributions of the well-regressed boxes and the bad-regressed boxes to the loss in the training process, paying more attention to the minority severely overlapping objects. Extensive experiments conducted on CrowdHuman and CityPersons demonstrate that our methods can help RCNN-based pedestrian detectors achieve state-of-the-art performance. | ['Xiao Jiajie', 'Luo Haibo', 'He Miao', 'Li Xiang'] | 2023-01-13 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [-1.02863483e-01 -2.54579097e-01 2.72314548e-01 -4.07357365e-01
-3.90674263e-01 -1.58399791e-01 2.38388166e-01 -1.18903987e-01
-7.69685686e-01 5.50726414e-01 4.77002263e-02 2.19936427e-02
3.38183165e-01 -9.98299837e-01 -6.00330830e-01 -9.93261993e-01
4.35615666e-02 -3.36953551e-01 6.81051791e-01 -1.68608099e-01
-1.80591986e-01 5.24067104e-01 -1.62550569e+00 6.71979129e-01
9.07819808e-01 1.21038914e+00 1.49231255e-01 1.70504898e-01
-1.24864601e-01 4.87720877e-01 -9.41949368e-01 -4.83781278e-01
6.50188148e-01 -8.40995535e-02 2.13668361e-01 1.33775115e-01
5.22952139e-01 -5.81990063e-01 -6.31674111e-01 1.35834086e+00
6.62376404e-01 1.58545971e-01 5.00859559e-01 -1.28068721e+00
-3.84809107e-01 2.79421896e-01 -1.05160391e+00 6.67471707e-01
-5.16233332e-02 6.16634965e-01 4.02334273e-01 -1.17953634e+00
2.85787076e-01 1.60180104e+00 6.67937040e-01 2.47693866e-01
-1.11705780e+00 -8.25184584e-01 4.72359180e-01 3.37992996e-01
-1.57612133e+00 -2.44070187e-01 7.39484012e-01 -3.47131759e-01
4.49756950e-01 2.87663877e-01 6.92084670e-01 8.96806657e-01
2.63289977e-02 7.08198309e-01 8.73957336e-01 -9.27074999e-02
3.05108856e-02 2.49016955e-01 1.49544179e-01 4.56224501e-01
6.69201016e-01 1.33411065e-01 -1.30802825e-01 9.47200730e-02
7.10458994e-01 3.02712262e-01 -1.68282419e-01 5.22845704e-03
-9.01095450e-01 6.74525201e-01 8.91156495e-01 1.75039425e-01
-2.72605002e-01 -5.00005670e-02 3.46383840e-01 -4.81983945e-02
3.44342202e-01 -1.08925141e-01 -3.56362492e-01 5.37302136e-01
-9.26503122e-01 1.94714770e-01 1.96866900e-01 8.33532453e-01
8.09550107e-01 1.50796294e-01 -7.29717672e-01 7.08478928e-01
1.78519040e-01 6.33681715e-01 1.26306027e-01 -5.39589047e-01
8.51312995e-01 9.55192089e-01 6.47806600e-02 -1.13493884e+00
-3.53387982e-01 -8.02068233e-01 -9.62207556e-01 3.48522902e-01
8.75403643e-01 -2.59877414e-01 -9.39372003e-01 1.42915261e+00
4.01466548e-01 1.64480284e-02 -2.79960543e-01 1.39027762e+00
1.17355692e+00 4.24400359e-01 4.64827210e-01 -1.02677047e-01
1.76874757e+00 -6.28319561e-01 -4.32634830e-01 -4.73458260e-01
2.57686466e-01 -7.40631640e-01 7.97264099e-01 -2.70155650e-02
-7.02074051e-01 -9.76244926e-01 -1.01460266e+00 1.77626945e-02
-3.19980562e-01 6.06565833e-01 3.56779873e-01 8.04431379e-01
-3.78228366e-01 3.00212085e-01 -6.27386749e-01 -1.25163258e-03
8.07456672e-01 8.62544775e-02 -2.55237162e-01 -2.07527444e-01
-1.05564666e+00 7.22929060e-01 3.06001723e-01 5.33819079e-01
-7.13771582e-01 -7.05219209e-01 -6.42009258e-01 2.19072923e-01
4.60079253e-01 -3.89327168e-01 4.36254412e-01 -1.00363934e+00
-7.71623611e-01 5.64548910e-01 -1.33068517e-01 -4.07690167e-01
7.93600857e-01 -1.29806668e-01 -3.48250628e-01 -4.35090847e-02
1.88635260e-01 6.79635167e-01 1.07892621e+00 -1.11291552e+00
-8.35241199e-01 -5.02649367e-01 -1.11744039e-01 9.30022523e-02
-3.15431446e-01 1.77431434e-01 -5.05919337e-01 -7.97783911e-01
4.21404034e-01 -5.06328106e-01 -3.31757307e-01 2.97835708e-01
-4.11875814e-01 -4.52271588e-02 9.36058998e-01 -7.21892834e-01
1.19741201e+00 -2.40550709e+00 -4.95700359e-01 2.05384031e-01
3.23912412e-01 3.39105606e-01 -2.65211552e-01 -2.86598116e-01
-1.09821208e-01 -4.85048182e-02 -9.75850672e-02 2.22375393e-02
-3.41926843e-01 -7.19910711e-02 -1.40538707e-01 5.19475281e-01
6.00336015e-01 6.65002882e-01 -7.24930644e-01 -5.79941452e-01
3.53159904e-01 6.49713993e-01 -3.10654730e-01 7.87344500e-02
7.46141523e-02 2.66199648e-01 -6.72802985e-01 7.73730397e-01
1.30273259e+00 3.23579043e-01 -4.59860265e-01 -4.68095750e-01
-2.26168260e-01 -2.25343853e-01 -1.62746835e+00 8.24414015e-01
1.49212837e-01 6.30768716e-01 2.80355126e-01 -8.13959956e-01
1.06641769e+00 -4.91409861e-02 2.60250568e-01 -7.77799785e-01
3.52358073e-01 7.16222636e-03 2.47327000e-01 -5.98735750e-01
3.63877594e-01 1.75408319e-01 1.84688941e-01 -2.11660013e-01
-2.62088597e-01 3.97574902e-01 8.10110793e-02 -5.00306860e-02
9.31235373e-01 -1.80661574e-01 2.64958125e-02 -1.79437503e-01
6.90037131e-01 -1.98401973e-01 8.34810197e-01 8.25416803e-01
-6.92307591e-01 7.72555232e-01 5.39541066e-01 -6.44628048e-01
-8.99249792e-01 -1.08990037e+00 -2.64099121e-01 9.82396603e-01
1.59881547e-01 5.21387644e-02 -9.48206186e-01 -5.47146797e-01
6.99064434e-02 3.63505274e-01 -5.89624882e-01 -1.77052036e-01
-8.24422836e-01 -1.12784326e+00 4.65114057e-01 6.18697464e-01
8.72518539e-01 -9.13544893e-01 -7.82916665e-01 2.03239709e-01
-3.63560051e-01 -1.16613054e+00 -4.14207876e-01 2.03541040e-01
-5.67221105e-01 -1.01835763e+00 -8.28452229e-01 -6.21965766e-01
1.00072801e+00 7.65460849e-01 7.82562256e-01 2.70264477e-01
-6.12014890e-01 -1.89281330e-01 -1.06266677e-01 -3.84895861e-01
1.96116447e-01 -3.28444451e-01 -2.06396118e-01 1.55484691e-01
6.03819728e-01 -1.22901537e-01 -9.76306975e-01 5.09192407e-01
-6.87500119e-01 -2.83214837e-01 3.99001539e-01 9.08761263e-01
4.58692878e-01 4.10736680e-01 2.97739029e-01 -2.26348922e-01
3.06216449e-01 -8.64430293e-02 -7.31873453e-01 -1.72992364e-01
1.98521748e-01 -2.30163753e-01 5.74010909e-01 -8.36413085e-01
-1.11896598e+00 1.33485958e-01 1.38452068e-01 -3.76201987e-01
-1.94338441e-01 -2.84372896e-01 -6.57057524e-01 1.28308982e-02
7.84256816e-01 3.53726856e-02 -2.77021855e-01 -2.58535147e-01
2.39660174e-01 5.19228518e-01 4.55316991e-01 -2.72762239e-01
1.02367568e+00 7.80625820e-01 -5.65001033e-02 -8.19239676e-01
-7.33368695e-01 -5.38610637e-01 -4.99915749e-01 -3.47062737e-01
9.36373770e-01 -1.18667996e+00 -5.69025815e-01 5.02457201e-01
-1.21988833e+00 1.36889592e-01 -3.56924713e-01 3.91687602e-01
1.79053605e-01 2.04956204e-01 -5.19665778e-01 -1.12078333e+00
-1.84327945e-01 -1.33101618e+00 8.64836514e-01 7.28085816e-01
2.34572858e-01 -9.71892551e-02 -7.02848792e-01 2.26553902e-01
3.36135089e-01 1.20386807e-02 7.67052412e-01 -4.76135403e-01
-9.02685642e-01 -3.91484529e-01 -7.75985181e-01 4.23272997e-01
-2.54301190e-01 1.71818733e-02 -1.17097712e+00 -9.65297222e-02
-9.61786788e-03 2.87419468e-01 1.29382527e+00 8.30791175e-01
1.12315655e+00 2.05831155e-02 -4.85979706e-01 4.94754523e-01
1.21543670e+00 2.03091241e-02 8.00515890e-01 3.51083189e-01
6.73498392e-01 7.64743984e-01 4.79339510e-01 4.77695405e-01
6.87314644e-02 6.08861864e-01 3.50926280e-01 -3.28391671e-01
-3.99120718e-01 -1.26714513e-01 2.75913417e-01 -1.08670369e-01
-1.72745645e-01 2.37326194e-02 -4.19670194e-01 3.75502974e-01
-1.63477695e+00 -1.08894575e+00 -3.69176716e-01 2.28191090e+00
4.43515599e-01 3.07841659e-01 1.57690436e-01 9.07614082e-02
1.08361042e+00 1.18740477e-01 -4.97745663e-01 1.80907741e-01
-4.97973561e-01 -7.45195374e-02 6.09895229e-01 1.34292548e-03
-1.47967982e+00 8.04867506e-01 4.91002893e+00 1.01234567e+00
-8.34626913e-01 1.70129957e-03 9.95769083e-01 -3.21989805e-01
2.29395211e-01 -1.49285451e-01 -1.16731393e+00 7.46459544e-01
2.42710859e-01 4.45506513e-01 1.99827030e-01 1.00615585e+00
5.73576331e-01 -3.20535958e-01 -5.87098479e-01 9.24910545e-01
-2.05348387e-01 -7.21357405e-01 -2.22105518e-01 -3.14142145e-02
5.66572130e-01 -2.38203734e-01 2.05347821e-01 3.69729578e-01
-3.15514207e-02 -8.09902370e-01 8.60818565e-01 4.86570150e-01
2.27420390e-01 -7.68236041e-01 9.87443566e-01 3.28688383e-01
-1.45635819e+00 -4.03928190e-01 -1.04319501e+00 -3.05912867e-02
-5.44672236e-02 1.12745881e+00 -3.28606576e-01 9.17589217e-02
1.02918363e+00 1.07052810e-01 -7.70509899e-01 1.33586395e+00
-3.00377667e-01 3.30176294e-01 -4.95893866e-01 3.90317477e-02
-9.21921805e-02 -1.69305444e-01 6.08316123e-01 1.34708464e+00
2.70346999e-01 2.81779557e-01 2.47739851e-01 1.12503624e+00
-1.14359949e-02 -1.84204038e-02 -2.39208132e-01 7.29574800e-01
4.30762500e-01 1.48131549e+00 -8.56147408e-01 -4.47875768e-01
-5.10458529e-01 6.52056575e-01 1.11561440e-01 7.21073091e-01
-9.18631852e-01 -3.51132721e-01 7.11624563e-01 4.25344616e-01
7.05834568e-01 -3.21423784e-02 -6.60048246e-01 -1.00433910e+00
5.82751930e-01 -7.15999484e-01 2.75863290e-01 -3.47422957e-01
-1.29933095e+00 4.85631436e-01 -1.67213038e-01 -1.40641785e+00
4.85510230e-01 -4.83830899e-01 -8.75777006e-01 9.64571655e-01
-1.44689512e+00 -9.04570222e-01 -7.31523871e-01 3.78609478e-01
4.17146504e-01 -9.75031555e-02 7.33817816e-02 5.82071185e-01
-9.72158730e-01 7.34922230e-01 -2.38859341e-01 5.94330668e-01
5.98364234e-01 -5.47459066e-01 3.05560321e-01 1.22485471e+00
-3.36040676e-01 5.46689630e-01 6.52930558e-01 -8.14508200e-01
-8.49155664e-01 -1.25025046e+00 5.03017008e-01 -1.53799817e-01
2.64513433e-01 -4.82818812e-01 -1.01438618e+00 1.14586435e-01
-4.25208688e-01 6.31597042e-01 9.31934863e-02 -3.35580349e-01
-5.33628225e-01 -2.90121496e-01 -1.26230538e+00 7.45829761e-01
1.01787031e+00 -1.27702281e-01 -4.28005934e-01 1.21185645e-01
5.37277043e-01 -1.17991030e-01 -2.88133472e-01 4.69818056e-01
4.01509941e-01 -1.14965987e+00 1.27935529e+00 -3.37185532e-01
2.91975707e-01 -7.76162207e-01 -9.75444838e-02 -9.46308911e-01
-4.84216422e-01 6.33875281e-02 1.36874184e-01 1.31332409e+00
1.77130342e-01 -4.55421716e-01 8.39769602e-01 6.91716015e-01
7.59544224e-02 -4.79859471e-01 -1.30744064e+00 -5.59277594e-01
-1.94268256e-01 -2.45808199e-01 5.85445344e-01 3.71451139e-01
-5.50768614e-01 -1.27898410e-01 -8.22868198e-02 3.56679827e-01
7.16859639e-01 -2.24217176e-01 7.70443380e-01 -1.21540904e+00
8.21020976e-02 -3.43358725e-01 -5.50468743e-01 -9.45616245e-01
-2.82816023e-01 -4.03034836e-01 7.57206529e-02 -1.24285448e+00
3.35978627e-01 -3.93351883e-01 -2.90621787e-01 2.22254559e-01
-6.69076502e-01 3.09059203e-01 4.57989663e-01 9.62327570e-02
-4.37904626e-01 4.99640882e-01 1.25965106e+00 -3.81975561e-01
-2.95782208e-01 1.13616243e-01 -5.58729231e-01 1.00797546e+00
4.98818487e-01 -6.07616425e-01 1.58698604e-01 -2.83136159e-01
-3.67224753e-01 -4.69153017e-01 8.32421720e-01 -1.34132850e+00
2.90896982e-01 2.52350010e-02 1.32115257e+00 -8.61820877e-01
3.17114711e-01 -8.90824318e-01 -1.88859269e-01 5.68278730e-01
7.78752342e-02 -2.20779464e-01 2.19689891e-01 4.86947089e-01
-2.73999497e-02 -1.82421863e-01 1.03199005e+00 -2.47892410e-01
-6.47938371e-01 2.24460512e-01 -5.15224874e-01 4.70865108e-02
9.98572886e-01 -5.49373507e-01 -5.10677218e-01 8.37905519e-03
-4.44885254e-01 8.38388130e-02 1.33318052e-01 2.89192170e-01
7.07142055e-01 -1.18913615e+00 -8.75133038e-01 6.91008270e-01
-1.81767829e-02 -1.47719653e-02 6.33025944e-01 6.31594658e-01
-1.63022727e-02 1.58757478e-01 -1.72734246e-01 -6.10706031e-01
-1.21219206e+00 5.44174135e-01 4.59744781e-01 -6.08709715e-02
-6.96005762e-01 9.06466365e-01 5.80477715e-01 7.62976939e-03
4.69454885e-01 -4.13158536e-01 -3.35856348e-01 2.26333573e-01
8.67495894e-01 5.97323418e-01 -1.24023311e-01 -6.71739101e-01
-4.98381764e-01 3.81367832e-01 -8.53452235e-02 2.46197566e-01
1.04794812e+00 -3.13306823e-02 6.30159229e-02 -2.43587986e-01
9.16145623e-01 -9.03396215e-03 -1.72565770e+00 -2.41363481e-01
-3.13764095e-01 -5.41070700e-01 2.96236053e-02 -4.25313443e-01
-1.32122707e+00 8.61194611e-01 1.01254833e+00 -1.62138551e-01
1.10442078e+00 -1.46149740e-01 6.32886648e-01 9.78375450e-02
7.84395412e-02 -1.25339282e+00 7.27330968e-02 3.38962406e-01
6.36268258e-01 -1.31953180e+00 2.27571860e-01 -6.81649327e-01
-5.01389742e-01 1.15715230e+00 9.23555791e-01 -1.81554168e-01
4.80522186e-01 4.00132328e-01 9.05352756e-02 1.64388061e-01
-2.52164572e-01 -5.60555041e-01 2.66740710e-01 8.77514184e-01
4.86517027e-02 8.02128688e-02 -2.00134218e-01 1.05574334e+00
1.56470668e-02 -2.49998078e-01 2.55905092e-01 4.44712400e-01
-8.33453536e-01 -5.09802103e-01 -9.53212798e-01 6.36412561e-01
-3.24066758e-01 -2.14747176e-01 -2.23945100e-02 7.07089961e-01
7.03350186e-01 1.10875416e+00 3.01746488e-01 -1.93545148e-01
7.40042269e-01 -2.48535201e-01 9.79388431e-02 -2.14513689e-01
-7.68357158e-01 3.28838855e-01 -1.89980283e-01 -4.83862907e-01
4.48618867e-02 -5.16754329e-01 -1.16110122e+00 -1.26765564e-01
-5.37288129e-01 -3.63048673e-01 3.16470236e-01 7.12042451e-01
4.19182554e-02 7.51840472e-01 4.30388838e-01 -8.16054225e-01
-5.54722071e-01 -1.01725328e+00 -3.88793379e-01 5.06952047e-01
2.11942226e-01 -7.83590436e-01 -5.27840674e-01 -1.88588589e-01] | [8.089515686035156, -0.5832346081733704] |
57d4c742-fc60-4105-b390-d24808849a48 | learning-canonical-3d-object-representation | 2108.04628 | null | https://arxiv.org/abs/2108.04628v1 | https://arxiv.org/pdf/2108.04628v1.pdf | Learning Canonical 3D Object Representation for Fine-Grained Recognition | We propose a novel framework for fine-grained object recognition that learns to recover object variation in 3D space from a single image, trained on an image collection without using any ground-truth 3D annotation. We accomplish this by representing an object as a composition of 3D shape and its appearance, while eliminating the effect of camera viewpoint, in a canonical configuration. Unlike conventional methods modeling spatial variation in 2D images only, our method is capable of reconfiguring the appearance feature in a canonical 3D space, thus enabling the subsequent object classifier to be invariant under 3D geometric variation. Our representation also allows us to go beyond existing methods, by incorporating 3D shape variation as an additional cue for object recognition. To learn the model without ground-truth 3D annotation, we deploy a differentiable renderer in an analysis-by-synthesis framework. By incorporating 3D shape and appearance jointly in a deep representation, our method learns the discriminative representation of the object and achieves competitive performance on fine-grained image recognition and vehicle re-identification. We also demonstrate that the performance of 3D shape reconstruction is improved by learning fine-grained shape deformation in a boosting manner. | ['Kwanghoon Sohn', 'Ig-Jae Kim', 'Minsu Kim', 'Seungryong Kim', 'Sunghun Joung'] | 2021-08-10 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Joung_Learning_Canonical_3D_Object_Representation_for_Fine-Grained_Recognition_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Joung_Learning_Canonical_3D_Object_Representation_for_Fine-Grained_Recognition_ICCV_2021_paper.pdf | iccv-2021-1 | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 2.00510353e-01 -1.75922349e-01 1.24075646e-02 -6.67520285e-01
-8.08867931e-01 -1.06722319e+00 8.27157855e-01 -4.29054260e-01
6.91830122e-04 -6.43302873e-02 -9.17767659e-02 -3.42422873e-02
2.32840836e-01 -8.20300996e-01 -1.30599427e+00 -6.20713711e-01
5.04013896e-01 7.57351518e-01 8.85683969e-02 -2.26001777e-02
1.80441305e-01 1.28883338e+00 -1.78869486e+00 2.52642721e-01
3.40242088e-01 1.13031340e+00 -4.19410802e-02 5.41403532e-01
-1.00977056e-01 5.40369898e-02 -3.65331739e-01 -3.31307441e-01
8.03995192e-01 3.56294736e-02 -5.94615400e-01 8.95282328e-01
1.27631593e+00 -5.99884152e-01 -3.84892672e-01 9.57792222e-01
1.26160488e-01 -4.24266467e-03 9.31925893e-01 -1.00687802e+00
-1.06727862e+00 -3.06602269e-01 -4.11004603e-01 -1.85383916e-01
1.71012074e-01 3.23905230e-01 7.05104291e-01 -1.09895360e+00
6.61531210e-01 1.72197759e+00 8.01245987e-01 6.37617469e-01
-1.66417503e+00 -4.98722076e-01 4.70803857e-01 -1.86993256e-01
-1.34734237e+00 -4.68913704e-01 1.00972402e+00 -6.88055933e-01
8.57088327e-01 1.66674927e-01 5.23686707e-01 9.52192724e-01
-1.25159591e-01 5.70122719e-01 9.22286928e-01 -2.71797389e-01
1.84840783e-02 -1.39780343e-01 4.29821201e-02 8.17087412e-01
1.70861140e-01 3.84465814e-01 -2.11288035e-01 -7.56850988e-02
1.04977751e+00 3.95624459e-01 4.73275036e-02 -1.02468073e+00
-1.03112113e+00 5.57238281e-01 5.27954400e-01 -1.77474663e-01
-1.91839203e-01 3.95049602e-01 -1.67590789e-02 3.46555024e-01
5.58105528e-01 1.76119164e-01 -4.51401651e-01 1.93742335e-01
-5.44891596e-01 3.72609794e-01 4.76969093e-01 9.42634344e-01
1.23637044e+00 3.37126017e-01 -1.26364708e-01 5.91180980e-01
4.99669224e-01 1.03703034e+00 3.07334602e-01 -1.14230263e+00
3.23608369e-02 8.30101252e-01 3.79821807e-02 -7.94209361e-01
-5.29699326e-02 -4.62002218e-01 -5.88817298e-01 7.07872748e-01
2.12761864e-01 3.99049848e-01 -1.24170709e+00 1.67550969e+00
5.69544256e-01 2.67698765e-01 -1.40128568e-01 8.50396574e-01
7.26734281e-01 3.44245315e-01 -3.80561680e-01 4.43287373e-01
1.16558194e+00 -7.22684801e-01 1.08207852e-01 -1.34854779e-01
3.24984521e-01 -5.23376942e-01 9.45666432e-01 4.29791920e-02
-7.82229006e-01 -7.24318922e-01 -1.00176013e+00 -1.54894844e-01
-3.49530011e-01 -4.77790348e-02 2.36588046e-01 7.10326731e-01
-1.11344469e+00 4.55072045e-01 -8.62878680e-01 -2.67357945e-01
5.73915362e-01 3.45918387e-01 -6.95391893e-01 -1.09142788e-01
-4.16065425e-01 8.23870420e-01 -6.13879599e-02 -1.24288067e-01
-1.06576025e+00 -8.58328938e-01 -1.08051515e+00 -3.51013571e-01
-5.73150963e-02 -1.00809371e+00 1.10117960e+00 -8.79700005e-01
-1.41224635e+00 1.37312424e+00 -3.85273695e-01 -1.69269621e-01
4.04431671e-01 1.09866291e-01 -2.24823635e-02 -8.66216123e-02
2.00408995e-02 4.68982816e-01 1.55637574e+00 -1.65995526e+00
-2.52613783e-01 -8.89509261e-01 1.19549774e-01 1.46273658e-01
2.31371701e-01 -3.84907275e-01 -5.19566476e-01 -7.15704501e-01
4.49083805e-01 -1.07070506e+00 -1.61011070e-01 4.23466235e-01
-8.65002424e-02 4.43471447e-02 1.18290830e+00 -4.24691916e-01
2.60499209e-01 -2.17583227e+00 1.13731839e-01 2.46279925e-01
3.00289899e-01 1.42586201e-01 -5.61644077e-01 -1.68159232e-01
-7.50484541e-02 2.20635936e-01 -9.78470370e-02 -5.34081519e-01
2.06975579e-01 3.88045460e-01 -5.92092156e-01 6.11050129e-01
7.77286828e-01 1.27176881e+00 -7.02346325e-01 2.17784062e-01
4.03910100e-01 6.59652829e-01 -7.84980655e-01 3.05625856e-01
-1.81169480e-01 7.12231338e-01 -6.16719723e-01 7.43536055e-01
9.06012595e-01 -2.33876571e-01 -2.04687417e-01 -5.86707473e-01
1.26682986e-02 -7.69214407e-02 -1.08207333e+00 1.51931345e+00
-4.73629445e-01 3.69715929e-01 9.36258361e-02 -1.12456882e+00
1.33565474e+00 -8.35433900e-02 4.92066920e-01 -7.02360332e-01
2.54039317e-02 4.21499386e-02 -4.16290343e-01 -1.75331041e-01
3.49433631e-01 -9.92235765e-02 -2.38237247e-01 5.68498373e-01
6.10655360e-02 -6.85793638e-01 -6.26001775e-01 -1.66496888e-01
9.17633116e-01 3.38802010e-01 -7.13130608e-02 -7.33841881e-02
2.40935713e-01 -2.42041945e-01 3.37433189e-01 7.11622119e-01
7.48106390e-02 7.58281887e-01 -1.55535862e-01 -8.94218743e-01
-1.41015613e+00 -1.32627964e+00 -3.36210549e-01 8.18175256e-01
3.21442157e-01 1.13892831e-01 -6.24836206e-01 -6.82964385e-01
7.06353605e-01 2.92695194e-01 -8.33537877e-01 -3.27978671e-01
-6.85589790e-01 -2.90143877e-01 4.87330705e-01 4.09653842e-01
5.40258527e-01 -4.92374688e-01 -3.80307913e-01 -1.06813267e-01
2.26380602e-01 -1.14431584e+00 -6.78570390e-01 3.02066356e-02
-9.11142409e-01 -1.05783105e+00 -3.10836464e-01 -7.46968031e-01
9.65732634e-01 5.83728075e-01 9.55169141e-01 8.96894485e-02
-4.52913821e-01 1.04886711e+00 -1.87087610e-01 -1.17869407e-01
-5.50686181e-01 -3.69212508e-01 1.92468941e-01 5.02729297e-01
2.68811882e-01 -7.18480468e-01 -3.11868459e-01 4.96789783e-01
-7.23784864e-01 -7.98447505e-02 3.99448842e-01 1.06945992e+00
9.96845126e-01 -2.64140546e-01 2.96927273e-01 -8.13369751e-01
-1.03612013e-01 -9.42870378e-02 -8.66039395e-01 6.51361197e-02
-3.54263097e-01 3.93305600e-01 3.08886766e-01 -4.81182396e-01
-9.76267695e-01 4.76306260e-01 -8.13183263e-02 -1.02685606e+00
-4.69521612e-01 -1.37638107e-01 -4.76100773e-01 -5.99210799e-01
7.14282811e-01 2.65575558e-01 2.99662769e-01 -8.59250128e-01
6.17817521e-01 3.99456918e-01 6.92620218e-01 -7.64307737e-01
1.37348235e+00 7.17347741e-01 1.43241718e-01 -6.51301503e-01
-7.83635557e-01 -3.90655249e-01 -1.20715034e+00 2.39501633e-02
7.76829600e-01 -1.14694917e+00 -6.94870770e-01 6.26496792e-01
-1.19390130e+00 -3.64191324e-01 -5.40497243e-01 7.87179619e-02
-6.64271712e-01 3.35368186e-01 -1.49347961e-01 -4.69071835e-01
2.52266973e-02 -1.28182638e+00 1.82142460e+00 -2.21018359e-01
8.76778662e-02 -8.19821835e-01 -1.16632178e-01 3.72170508e-01
2.98093319e-01 4.78571862e-01 8.92660737e-01 -2.40113080e-01
-1.02183163e+00 -2.82545447e-01 -3.66049260e-01 3.43731791e-01
4.30889010e-01 -1.62927806e-01 -1.09515440e+00 -3.80242646e-01
5.06999046e-02 -2.61403441e-01 9.37226951e-01 6.37910292e-02
1.22780812e+00 -3.91241044e-01 -2.66506851e-01 1.18496466e+00
1.20774412e+00 -1.44440621e-01 2.95137435e-01 1.85836226e-01
1.15250611e+00 4.41002518e-01 1.34189308e-01 3.90708834e-01
5.36642373e-01 9.68868554e-01 7.00853229e-01 -7.49990996e-03
-6.30830109e-01 -4.43169415e-01 1.54587984e-01 2.93029249e-01
-8.58096927e-02 2.60457009e-01 -7.80264020e-01 3.22372377e-01
-1.56052673e+00 -1.00544608e+00 2.22834125e-01 2.25976872e+00
6.52620375e-01 -2.77323157e-01 1.62538700e-02 -2.37951234e-01
5.73917866e-01 6.70659840e-02 -1.12267673e+00 -1.37257770e-01
-2.99726427e-01 6.84513375e-02 5.85532486e-01 6.19875312e-01
-1.10215604e+00 1.04020274e+00 6.22766113e+00 3.72697860e-01
-1.39420784e+00 -1.26427218e-01 4.26810712e-01 1.08485073e-01
-8.37826490e-01 -7.20811859e-02 -9.02779520e-01 7.61241168e-02
3.73744577e-01 4.45757955e-02 6.38351679e-01 7.49380350e-01
-1.21736087e-01 6.57131016e-01 -1.40264344e+00 1.33830726e+00
3.80349129e-01 -1.52321601e+00 5.80611110e-01 1.79835826e-01
8.94903541e-01 1.15925014e-01 2.15848073e-01 1.46499813e-01
4.26485091e-01 -9.85506713e-01 1.13503706e+00 6.60093009e-01
1.18787909e+00 -1.36686772e-01 1.18625157e-01 3.17035168e-01
-1.21462440e+00 6.69229925e-02 -2.81576037e-01 1.41877577e-01
-1.77785292e-01 2.28362069e-01 -9.23260331e-01 4.36610281e-01
6.28999174e-01 9.09846365e-01 -6.22621417e-01 7.51579046e-01
1.77045628e-01 2.43800983e-01 -2.83620328e-01 4.78534967e-01
-1.15941659e-01 -7.71064609e-02 8.02242219e-01 9.66025054e-01
2.55309194e-01 9.46425721e-02 4.95469689e-01 1.08692777e+00
-2.91679591e-01 -5.04081249e-01 -7.90151715e-01 1.98582456e-01
5.29575646e-01 9.39577162e-01 -2.87999421e-01 -1.35389432e-01
-2.95492828e-01 9.89725590e-01 3.51429880e-01 4.95600522e-01
-4.09548432e-01 2.00158849e-01 1.01300085e+00 2.49648497e-01
7.92504251e-01 -4.32383984e-01 -3.08961004e-01 -1.37519443e+00
1.66387528e-01 -7.92265475e-01 -6.68376163e-02 -7.61998773e-01
-1.46942449e+00 6.11932397e-01 -8.29315260e-02 -1.30115008e+00
-3.88563991e-01 -9.74894106e-01 -2.07134664e-01 1.02061462e+00
-1.52962637e+00 -1.68784010e+00 -4.65384901e-01 7.17490435e-01
4.89597589e-01 -2.03806445e-01 8.72099102e-01 7.82980621e-02
-1.92071721e-02 8.15650284e-01 8.71828478e-03 1.47101864e-01
6.37134850e-01 -1.05279231e+00 7.68631220e-01 6.09499514e-01
1.70980036e-01 3.99219513e-01 3.18328261e-01 -5.99259138e-01
-2.04682493e+00 -1.54687345e+00 2.34130859e-01 -1.00393248e+00
3.73688728e-01 -5.35872221e-01 -9.31741595e-01 8.58211458e-01
-6.58029377e-01 5.57669938e-01 2.52401054e-01 -1.76545642e-02
-1.21650028e+00 -3.17443401e-01 -1.30996311e+00 3.18688661e-01
1.41506565e+00 -9.67530549e-01 -5.62098980e-01 5.60597479e-02
8.28528821e-01 -6.76737428e-01 -7.94284880e-01 5.42906344e-01
6.58585131e-01 -4.90060389e-01 1.33359706e+00 -8.18479717e-01
-1.58448175e-01 -6.02360487e-01 -7.12361753e-01 -1.24306643e+00
-6.30391181e-01 -3.10564756e-01 -8.43691751e-02 9.15586233e-01
5.57387955e-02 -7.35165656e-01 9.06462312e-01 7.48222888e-01
-4.42038208e-01 -4.21132803e-01 -9.08937275e-01 -7.96799421e-01
2.42003500e-01 -5.60924292e-01 1.04105723e+00 6.78768337e-01
-9.12289619e-01 8.37116912e-02 -1.46322593e-01 5.89955986e-01
8.62051725e-01 6.75016880e-01 1.23445702e+00 -1.50609040e+00
-2.46106103e-01 -5.14422417e-01 -9.41722572e-01 -1.45068908e+00
6.25724673e-01 -1.10870647e+00 1.01130553e-01 -1.08040988e+00
2.42621243e-01 -6.18248641e-01 8.67859498e-02 6.56145275e-01
1.57723531e-01 7.94640899e-01 2.87716776e-01 4.15851593e-01
-3.88162762e-01 6.87614501e-01 1.45706511e+00 -5.55956960e-01
8.36154670e-02 -9.49144885e-02 -7.38240063e-01 6.15320385e-01
3.53619456e-01 -9.67654213e-02 -1.36929631e-01 -9.26910639e-01
-3.08222234e-01 -3.31483036e-01 1.00857770e+00 -4.29883987e-01
-3.91634628e-02 -1.71170056e-01 6.77716315e-01 -6.45059228e-01
5.86848736e-01 -1.00805080e+00 4.24711078e-01 6.46065862e-04
-1.34350330e-01 -2.40610659e-01 3.79955173e-01 6.40786529e-01
2.85879467e-02 9.84800756e-02 8.18530023e-01 -2.63274945e-02
-8.32516551e-01 8.42330515e-01 5.95959723e-02 -1.85379058e-01
7.33228385e-01 -6.48101985e-01 -1.49572238e-01 3.25657874e-02
-6.46194220e-01 -1.77797452e-01 1.13692033e+00 6.24560058e-01
6.61212862e-01 -1.81619012e+00 -7.18432128e-01 9.51964915e-01
3.13942820e-01 2.04991847e-01 3.58297497e-01 1.44471422e-01
-1.66895270e-01 2.19527006e-01 -2.48452365e-01 -1.08681273e+00
-1.00268281e+00 4.57524359e-01 7.54133344e-01 3.03173065e-01
-6.64191186e-01 6.32882357e-01 7.80834496e-01 -9.52604413e-01
1.04897946e-01 -3.98777366e-01 3.33371043e-01 -3.49847168e-01
4.34796035e-01 -1.68927982e-01 3.60522538e-01 -1.08699632e+00
-3.34968626e-01 1.26499188e+00 9.93718877e-02 1.60836205e-01
1.33010185e+00 -3.45701694e-01 -1.27915852e-02 4.24942553e-01
1.57601619e+00 3.46807092e-02 -1.86375225e+00 -5.89945912e-01
-3.43685895e-01 -8.44909549e-01 3.46347541e-02 -6.35859609e-01
-9.89223003e-01 6.99487448e-01 5.60754657e-01 -1.94359958e-01
8.36318433e-01 2.96688735e-01 3.60722363e-01 6.54781640e-01
6.11753404e-01 -3.83152276e-01 2.38742620e-01 8.59224737e-01
1.12737536e+00 -1.41771853e+00 -2.14131251e-01 -1.88359126e-01
-1.68595940e-01 1.28195143e+00 3.37317318e-01 -4.72001731e-01
8.52397859e-01 1.80166960e-01 2.26355586e-02 -2.13375032e-01
-4.54927355e-01 -5.44091649e-02 8.72623622e-01 1.03077912e+00
-5.70704527e-02 1.98597744e-01 8.86967659e-01 4.37638104e-01
-1.80827066e-01 -4.69631881e-01 6.10139444e-02 4.98252988e-01
-3.84967744e-01 -1.19957817e+00 -5.25943518e-01 3.32645178e-01
9.94721204e-02 2.45504022e-01 -4.10598367e-01 4.56461936e-01
1.02594167e-01 3.53064597e-01 4.61909205e-01 -5.15069306e-01
5.88491917e-01 2.45756544e-02 8.79089594e-01 -7.93820024e-01
-2.15993114e-02 -9.02950242e-02 -2.21940055e-01 -7.19477296e-01
-2.58738548e-01 -9.04509068e-01 -1.00084066e+00 -1.90835759e-01
4.34450470e-02 -2.73903191e-01 7.16392279e-01 9.22392368e-01
6.17539823e-01 1.94484562e-01 1.00022066e+00 -1.52228951e+00
-6.99827015e-01 -3.37883800e-01 -4.91894215e-01 7.11414158e-01
9.46369469e-01 -9.85175133e-01 -3.15010518e-01 4.50974882e-01] | [8.38293170928955, -3.231828451156616] |
29cd10d2-f2f5-4289-8a01-dc3a690f4520 | jazznet-a-dataset-of-fundamental-piano | 2302.08632 | null | https://arxiv.org/abs/2302.08632v1 | https://arxiv.org/pdf/2302.08632v1.pdf | jazznet: A Dataset of Fundamental Piano Patterns for Music Audio Machine Learning Research | This paper introduces the jazznet Dataset, a dataset of fundamental jazz piano music patterns for developing machine learning (ML) algorithms in music information retrieval (MIR). The dataset contains 162520 labeled piano patterns, including chords, arpeggios, scales, and chord progressions with their inversions, resulting in more than 26k hours of audio and a total size of 95GB. The paper explains the dataset's composition, creation, and generation, and presents an open-source Pattern Generator using a method called Distance-Based Pattern Structures (DBPS), which allows researchers to easily generate new piano patterns simply by defining the distances between pitches within the musical patterns. We demonstrate that the dataset can help researchers benchmark new models for challenging MIR tasks, using a convolutional recurrent neural network (CRNN) and a deep convolutional neural network. The dataset and code are available via: https://github.com/tosiron/jazznet. | ['Tosiron Adegbija'] | 2023-02-17 | null | null | null | null | ['music-information-retrieval'] | ['music'] | [ 1.24431349e-01 -2.87968993e-01 -1.42698869e-01 2.16412187e-01
-6.36550963e-01 -9.51837003e-01 3.60288203e-01 -4.54452395e-01
1.76793635e-01 3.40894729e-01 4.97761905e-01 -1.29276574e-01
-3.79852384e-01 -8.38160455e-01 -3.82483482e-01 -4.63116288e-01
-8.93337652e-02 4.51060623e-01 -7.40330592e-02 -5.40873230e-01
4.31337774e-01 2.29532659e-01 -1.69961739e+00 6.83846414e-01
5.05967857e-03 8.57922912e-01 5.49912006e-02 9.67991114e-01
1.92041765e-03 1.04346061e+00 -1.04775083e+00 -3.49092990e-01
3.70623946e-01 -8.29098642e-01 -8.53594899e-01 -6.19342804e-01
5.10079920e-01 1.13932885e-01 -5.60051620e-01 6.54969394e-01
9.00677919e-01 3.25003475e-01 3.80858928e-01 -1.24759793e+00
-7.04253852e-01 1.54570961e+00 -1.90239817e-01 8.56313258e-02
2.80360550e-01 4.79071997e-02 1.56343687e+00 -7.69757867e-01
6.29328251e-01 8.28362226e-01 9.72360075e-01 6.65207744e-01
-9.68647182e-01 -1.08770287e+00 -7.97020614e-01 4.43934500e-01
-1.49248910e+00 -4.32889402e-01 1.06255400e+00 -4.21954215e-01
6.48094296e-01 5.46898007e-01 1.09072983e+00 1.21194494e+00
-1.94395363e-01 9.74185109e-01 6.56879544e-01 -4.18218672e-01
-2.67962255e-02 -7.80511737e-01 -3.03454418e-02 3.41671586e-01
-5.42957723e-01 1.42760530e-01 -8.56257677e-01 -1.49015442e-01
9.31760430e-01 -3.87500793e-01 -1.97377335e-03 1.40041992e-01
-1.70168626e+00 4.05709445e-01 4.68376845e-01 5.90377450e-01
-8.28288049e-02 3.56051892e-01 5.44678569e-01 5.36226749e-01
-2.32682630e-01 1.05729985e+00 -2.87159920e-01 -6.97275162e-01
-1.11144686e+00 9.24735665e-01 6.89252436e-01 7.47203112e-01
1.82384759e-01 4.39529717e-01 -5.71162283e-01 1.34225512e+00
-1.72304317e-01 5.05793154e-01 8.25813174e-01 -1.11623073e+00
2.15708688e-01 2.40166843e-01 -2.80540615e-01 -1.03402436e+00
-3.40792507e-01 -6.44201398e-01 -9.44174647e-01 -4.12090831e-02
4.45111573e-01 1.28263496e-02 -3.92593771e-01 1.65953994e+00
-1.42766073e-01 3.48427534e-01 -3.09682041e-01 8.10023785e-01
1.29925263e+00 7.58254766e-01 -5.72273910e-01 2.56638855e-01
1.23228836e+00 -1.24527311e+00 -4.41954285e-01 3.80448401e-01
3.23449641e-01 -1.39665377e+00 1.60486078e+00 9.18035924e-01
-1.37497115e+00 -9.14430499e-01 -1.17742932e+00 -2.39422470e-01
-1.02033362e-01 3.63568157e-01 4.60296243e-01 1.94583848e-01
-8.50552142e-01 1.06848359e+00 -3.38450015e-01 1.72550827e-01
1.29644424e-01 1.66251078e-01 2.43362218e-01 7.37050772e-01
-1.08724344e+00 9.29602757e-02 2.93920726e-01 -4.72397506e-02
-7.20268667e-01 -1.04648685e+00 -3.25322986e-01 2.16856465e-01
-2.80380368e-01 -5.87430775e-01 1.81477046e+00 -8.35073531e-01
-1.98293877e+00 9.24925506e-01 3.04559886e-01 -5.00956297e-01
8.76378417e-02 -1.72402814e-01 -5.12683153e-01 -1.54610410e-01
-5.86813651e-02 7.98375070e-01 6.65782332e-01 -5.95080912e-01
-4.58490908e-01 1.74648330e-01 -2.52053797e-01 4.40519862e-02
-6.56707883e-02 1.99236810e-01 -4.25874084e-01 -1.46672213e+00
6.64534196e-02 -1.25451720e+00 8.93615559e-02 -6.44794643e-01
-9.00453985e-01 -4.33071047e-01 3.37016732e-01 -5.46744406e-01
1.73434329e+00 -2.28508711e+00 1.65881559e-01 2.68333972e-01
8.92812759e-03 1.96576759e-01 -5.44686079e-01 5.13148904e-01
-3.30716133e-01 -1.02720007e-01 5.00033610e-02 -2.02021077e-02
3.85274082e-01 -3.10008615e-01 -8.78815770e-01 -3.31150740e-01
-3.83411139e-01 1.22780287e+00 -8.18121374e-01 7.61111081e-02
-1.12638529e-02 3.12678397e-01 -5.74922740e-01 1.02765523e-01
-4.84180897e-01 6.55398071e-01 5.24532236e-02 5.69261730e-01
1.31965861e-01 -1.01054072e-01 1.28774330e-01 -1.97557583e-01
-1.32223219e-01 1.09658623e+00 -1.12248194e+00 1.95380032e+00
-3.40598434e-01 7.51117289e-01 -5.75774789e-01 -3.86143059e-01
1.30021989e+00 3.41388762e-01 6.76873624e-01 -7.34102011e-01
5.43321408e-02 2.82712579e-01 2.93246001e-01 -2.41447594e-02
7.97235191e-01 8.29995349e-02 -4.28072035e-01 9.30661738e-01
-1.71802174e-02 -3.35407585e-01 5.86591363e-01 -1.38732746e-01
1.27073359e+00 1.41797453e-01 -5.58896130e-03 1.22452803e-01
3.32159191e-01 2.06412543e-02 6.40212476e-01 7.57221162e-01
3.07227850e-01 8.49843144e-01 3.02094340e-01 -9.41151559e-01
-1.34240150e+00 -1.38455081e+00 -4.60055843e-02 1.29240263e+00
-4.52115953e-01 -1.13500595e+00 -4.09678519e-01 7.92837441e-02
-2.70752579e-01 2.33260542e-01 -2.58864522e-01 -4.33525033e-02
-1.01490498e+00 -4.19357747e-01 1.43740392e+00 1.40880898e-01
6.51124537e-01 -2.05887938e+00 -1.28235087e-01 3.29151988e-01
-5.02086222e-01 -3.26078415e-01 -7.35169351e-01 -2.48054430e-01
-4.89931464e-01 -1.12265837e+00 -7.29162633e-01 -1.12536061e+00
-2.82122731e-01 -1.14289761e-01 1.52600157e+00 -9.86682698e-02
-4.61218327e-01 -1.47001490e-01 -5.42111456e-01 -4.37266469e-01
-3.81043315e-01 5.24447143e-01 9.30651426e-02 -3.98635834e-01
2.78043091e-01 -1.04554391e+00 -5.46793520e-01 2.42321163e-01
-7.87472725e-01 2.27202162e-01 2.98397809e-01 6.64036870e-01
9.10975039e-01 -2.08972380e-01 6.71282232e-01 -6.77034497e-01
8.79651904e-01 -1.58425972e-01 -3.62576574e-01 -8.73222426e-02
-2.08220065e-01 -3.12679946e-01 5.70385993e-01 -6.35357320e-01
-3.96181375e-01 -1.76421180e-01 -4.01331931e-01 -4.64856118e-01
-6.83564544e-02 4.49156523e-01 1.36986628e-01 3.58901232e-01
9.40699935e-01 3.39347064e-01 -3.05618614e-01 -8.59924316e-01
6.43619359e-01 7.10740924e-01 1.04294789e+00 -6.87393546e-01
7.03120112e-01 -1.38065992e-02 -1.27934337e-01 -5.26866376e-01
-1.08692491e+00 -1.89809233e-01 -4.86529201e-01 -3.05406272e-01
4.13577229e-01 -7.33143151e-01 -7.98222363e-01 5.25526047e-01
-8.35688412e-01 -6.63976192e-01 -7.16007948e-01 2.30126485e-01
-9.55072105e-01 -4.45512496e-03 -1.22236156e+00 -2.16682851e-01
-9.55503881e-01 -6.10551596e-01 7.95065761e-01 1.91252172e-01
-1.05289173e+00 -3.37668121e-01 7.05237150e-01 3.29384595e-01
4.82998550e-01 8.54228884e-02 1.01894808e+00 -3.41417879e-01
-5.33713639e-01 1.16518535e-01 2.88911939e-01 5.11965193e-02
1.78227335e-01 2.02756420e-01 -9.78920937e-01 1.48356417e-02
-5.38153112e-01 -5.01814961e-01 7.27553487e-01 2.99051434e-01
1.55471206e+00 -6.21320844e-01 2.90092140e-01 1.09641600e+00
6.52986944e-01 3.39442223e-01 8.09032798e-01 5.53613603e-01
7.16709852e-01 2.01839536e-01 2.47486129e-01 5.82686484e-01
2.16108769e-01 1.04451132e+00 -3.39452252e-02 1.41975835e-01
-7.54156232e-01 -6.10669196e-01 3.43240231e-01 1.70095730e+00
-1.69833079e-01 9.03052539e-02 -8.05219173e-01 5.49150229e-01
-1.76403081e+00 -1.44524038e+00 1.75284296e-02 2.04033422e+00
1.35842907e+00 -2.53153265e-01 5.63799918e-01 7.81659067e-01
5.02546787e-01 2.89888978e-01 -3.82388145e-01 -5.38763940e-01
-3.06339979e-01 9.50844288e-01 -1.10749729e-01 1.65637195e-01
-1.17049813e+00 1.12166214e+00 6.61623812e+00 1.14652574e+00
-1.17792201e+00 -2.56300718e-01 3.09308410e-01 -5.46162367e-01
-2.42097631e-01 -2.12905526e-01 -5.44173360e-01 5.01055777e-01
8.11397851e-01 -3.05252075e-01 9.83655810e-01 7.55052567e-01
1.79055288e-01 9.45432007e-01 -7.77484715e-01 1.40362930e+00
2.57016197e-02 -1.84191048e+00 4.33569580e-01 -2.65569597e-01
8.22743356e-01 3.21616083e-01 2.60795593e-01 4.61351842e-01
4.42998618e-01 -1.16996038e+00 8.73071074e-01 7.16951787e-01
9.33098078e-01 -8.23408127e-01 1.66957766e-01 -1.63493544e-01
-1.38229680e+00 1.86804833e-03 -2.31764391e-01 -2.86871195e-01
-1.25217438e-01 4.33760345e-01 -7.90605605e-01 3.36157352e-01
8.62086654e-01 1.24192059e+00 -5.96230507e-01 1.32031131e+00
-3.07196170e-01 9.56872225e-01 -3.15192133e-01 -9.93584655e-03
-1.29623368e-01 -2.54831523e-01 7.10168540e-01 1.01549888e+00
6.51502609e-01 -4.21911031e-01 -8.99659544e-02 1.00967789e+00
-3.15355420e-01 3.40297073e-01 -4.21118528e-01 -2.81600744e-01
7.54372180e-01 1.32565618e+00 -5.32158256e-01 -2.15921819e-01
4.02862072e-01 6.35156989e-01 -8.71128344e-04 1.36647984e-01
-5.31688213e-01 -7.55395949e-01 7.85553277e-01 1.30022243e-02
2.00763196e-01 -1.29511476e-01 -3.93982232e-01 -1.05400276e+00
-5.17954230e-02 -1.42315519e+00 5.61899126e-01 -1.09670627e+00
-1.45353687e+00 7.49131382e-01 -5.62166870e-01 -1.76435876e+00
-5.06155133e-01 -3.30161721e-01 -9.62482929e-01 4.61555392e-01
-9.10656333e-01 -1.09917676e+00 -5.53738214e-02 6.00238264e-01
4.64135230e-01 -6.26500726e-01 1.19708157e+00 6.47490084e-01
-2.08121091e-01 7.28654504e-01 -1.20179188e-02 6.65074825e-01
7.93484151e-01 -1.16885090e+00 7.33591616e-01 2.21529663e-01
1.15759408e+00 4.78829056e-01 3.00223261e-01 -1.43333867e-01
-1.04572511e+00 -1.01860809e+00 1.00786293e+00 -4.82971638e-01
8.67274106e-01 -2.47888431e-01 -3.83658499e-01 5.88337004e-01
1.20221451e-01 -6.03108525e-01 1.06211948e+00 3.46035689e-01
-5.78773499e-01 -2.56506532e-01 -2.20964819e-01 9.74478304e-01
1.28981352e+00 -5.21961272e-01 -6.66323483e-01 2.98332781e-01
5.70376992e-01 -6.17862642e-01 -8.55803013e-01 7.00188220e-01
9.11277592e-01 -8.47557306e-01 1.08850479e+00 -5.09743750e-01
7.86623418e-01 -6.36842847e-01 1.33033559e-01 -1.19092250e+00
-6.29619122e-01 -1.02353013e+00 -2.03211442e-01 1.09396148e+00
5.67265749e-01 2.34377608e-01 8.62677217e-01 -5.46867251e-01
-2.39951000e-01 -4.69246954e-01 -7.45460391e-01 -5.84668934e-01
1.26127511e-01 -7.74240732e-01 1.15541899e+00 1.24193454e+00
2.12847367e-01 5.49324453e-01 -5.77753246e-01 -5.14092207e-01
1.73120558e-01 8.41421962e-01 1.19956613e+00 -1.08094239e+00
-9.07804906e-01 -1.19718671e+00 -3.70552629e-01 -1.02373362e+00
-2.13140324e-01 -1.57070160e+00 -3.59312207e-01 -1.08343172e+00
-2.75704563e-01 -7.16517031e-01 -6.73305273e-01 9.74787354e-01
4.27109808e-01 1.00320601e+00 6.17605865e-01 6.47933304e-01
-5.21344125e-01 5.59938490e-01 1.24431801e+00 -4.66838956e-01
-5.69604158e-01 1.98128328e-01 -5.04082680e-01 7.32929230e-01
1.23610914e+00 -4.04310554e-01 -2.34172612e-01 -3.94909978e-01
9.20169115e-01 -1.27364010e-01 1.78765878e-01 -1.39807403e+00
7.79865980e-02 2.98431486e-01 2.77372658e-01 -7.42347538e-01
4.25281703e-01 1.26255691e-01 5.23325026e-01 2.97340870e-01
-8.20743680e-01 2.97336817e-01 2.01288640e-01 -2.44934738e-01
-4.63855147e-01 -9.07717571e-02 2.70348698e-01 -1.55388251e-01
-3.44170958e-01 2.43389040e-01 -2.63373137e-01 1.14473782e-01
2.13796318e-01 2.84204274e-01 -2.66548246e-01 -5.69760859e-01
-1.05424464e+00 -3.74407977e-01 8.00815895e-02 9.45540011e-01
4.41739112e-01 -1.93806541e+00 -7.87484825e-01 2.98190475e-01
1.72884226e-01 -2.48178020e-01 1.01062804e-01 3.05395693e-01
-8.31378520e-01 2.72384226e-01 -4.38831389e-01 -2.66811609e-01
-1.30580330e+00 7.76845142e-02 2.94945449e-01 -2.27979049e-01
-9.32517648e-01 8.84897530e-01 -3.29681426e-01 -9.39111352e-01
4.01814282e-01 -3.86595935e-01 -3.77470285e-01 -1.75366700e-01
7.85816729e-01 2.77190208e-01 -7.31159523e-02 -4.14084792e-01
2.34911010e-01 6.34600103e-01 2.20294073e-01 -1.99930280e-01
1.26748347e+00 4.90382552e-01 -4.03725207e-01 8.22965860e-01
5.50878346e-01 5.57976127e-01 -5.43853998e-01 -2.14428186e-01
3.18678707e-01 -1.93114594e-01 -4.88687724e-01 -9.58485544e-01
-9.74725068e-01 6.17048383e-01 2.94938177e-01 9.34839770e-02
9.75067317e-01 -1.13595715e-02 1.30712140e+00 7.07233846e-01
1.72872871e-01 -8.90550137e-01 3.67564261e-01 9.03107047e-01
1.18654263e+00 -2.58511931e-01 -4.57932204e-01 1.29487589e-01
-5.53116083e-01 1.10325754e+00 3.04739892e-01 -5.55772305e-01
7.80677855e-01 2.53901422e-01 4.42688078e-01 -1.11225337e-01
-9.12301958e-01 -2.06962079e-02 5.79712749e-01 1.60499856e-01
8.03236485e-01 3.82164836e-01 -1.36092171e-01 1.10058475e+00
-1.77440989e+00 3.63001041e-02 4.71903116e-01 3.74758780e-01
-1.13553569e-01 -1.60526168e+00 -2.07496494e-01 3.65312248e-01
-5.75830042e-01 -5.71193278e-01 -7.45904028e-01 3.17968577e-01
3.85801375e-01 6.56427383e-01 3.36413980e-01 -7.68037558e-01
2.16499612e-01 2.36139849e-01 4.76024240e-01 -5.42177677e-01
-8.32301080e-01 1.69008940e-01 1.05443098e-01 -4.46264476e-01
-2.56399423e-01 -3.53974700e-01 -1.10706306e+00 -3.30153286e-01
3.49023014e-01 3.00664842e-01 3.68644118e-01 3.89541715e-01
6.35321081e-01 7.11908340e-01 8.09927881e-01 -8.56424689e-01
-4.86382917e-02 -1.27639127e+00 -5.72712600e-01 5.56119740e-01
-8.51191431e-02 -1.51918218e-01 -1.02376394e-01 2.58132100e-01] | [16.013181686401367, 5.4868950843811035] |
35aeea29-4f73-42f6-aead-475db1507eeb | sage-ndvi-a-stereotype-breaking-evaluation | 2306.06288 | null | https://arxiv.org/abs/2306.06288v1 | https://arxiv.org/pdf/2306.06288v1.pdf | SAGE-NDVI: A Stereotype-Breaking Evaluation Metric for Remote Sensing Image Dehazing Using Satellite-to-Ground NDVI Knowledge | Image dehazing is a meaningful low-level computer vision task and can be applied to a variety of contexts. In our industrial deployment scenario based on remote sensing (RS) images, the quality of image dehazing directly affects the grade of our crop identification and growth monitoring products. However, the widely used peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) provide ambiguous visual interpretation. In this paper, we design a new objective metric for RS image dehazing evaluation. Our proposed metric leverages a ground-based phenology observation resource to calculate the vegetation index error between RS and ground images at a hazy date. Extensive experiments validate that our metric appropriately evaluates different dehazing models and is in line with human visual perception. | ['Jui-Hsin Lai', 'Jun Yu', 'Mei Han', 'Yibing Wei', 'Andy Wong', 'Mingye Zhu', 'Zhicheng Yang', 'Zepeng Liu'] | 2023-06-09 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 8.20701480e-01 -3.31546336e-01 1.99952256e-02 6.16409332e-02
-3.05332810e-01 -7.08876967e-01 5.41940689e-01 5.63473642e-01
-1.11134529e-01 4.34274226e-01 -2.58713871e-01 -4.65988398e-01
-2.24695042e-01 -1.16154206e+00 -5.75611234e-01 -8.31185520e-01
-1.84287325e-01 -7.61632442e-01 3.26354682e-01 -5.38815081e-01
3.39536548e-01 6.29347622e-01 -1.78508937e+00 -2.29406133e-02
1.12185848e+00 1.29179537e+00 4.92194653e-01 1.08972025e+00
5.04019260e-01 5.03133416e-01 -7.75105476e-01 -2.15901539e-01
6.26919687e-01 -1.88140199e-01 -4.01836932e-01 3.41930926e-01
6.45764232e-01 -6.49023056e-01 3.33869900e-03 1.58342314e+00
3.14115554e-01 -2.50155538e-01 6.39974475e-01 -1.31386888e+00
-8.75828803e-01 1.32453531e-01 -8.42476368e-01 5.27200162e-01
-2.94815586e-03 4.31099117e-01 9.26975131e-01 -7.68745303e-01
2.02516869e-01 7.72948742e-01 8.52871180e-01 -4.52055931e-01
-1.28027070e+00 -4.08828348e-01 4.33249585e-02 2.43815482e-01
-1.67156756e+00 -7.12298900e-02 5.52407742e-01 -4.53212112e-01
2.76803762e-01 3.62447768e-01 5.95211744e-01 3.06263030e-01
5.18200219e-01 3.62884820e-01 1.20742393e+00 -3.97224247e-01
2.43437126e-01 -2.82175392e-01 -1.68182567e-01 4.92205352e-01
6.98250473e-01 4.86830980e-01 -2.07314402e-01 1.91018298e-01
8.86040866e-01 2.13909701e-01 -5.64233303e-01 -1.61021978e-01
-1.02072799e+00 6.28638864e-01 7.67654836e-01 1.16165645e-01
-6.35630012e-01 3.48728478e-01 -1.86304703e-01 3.78132969e-01
3.63025963e-01 6.07157528e-01 -1.58613905e-01 2.89583415e-01
-1.19653475e+00 9.64402258e-02 3.91345382e-01 7.29941547e-01
8.45958292e-01 3.60017836e-01 -2.03640148e-01 7.57286191e-01
2.38818154e-01 1.06018841e+00 -1.53073475e-01 -1.14184999e+00
-2.07667246e-01 2.71372586e-01 2.87921458e-01 -1.56947327e+00
1.27373457e-01 -5.52396178e-01 -9.46403146e-01 7.36989081e-01
1.41052440e-01 2.50996262e-01 -8.53451908e-01 1.18270099e+00
-1.55283093e-01 2.29407087e-01 1.22501468e-02 9.93572772e-01
5.37395954e-01 7.71384597e-01 3.27618346e-02 -1.06245860e-01
1.20698631e+00 -5.66459775e-01 -6.85287833e-01 -4.31167781e-01
4.10028584e-02 -8.00272763e-01 1.02456629e+00 6.51998222e-01
-9.63089108e-01 -5.16804636e-01 -1.60124207e+00 3.85894775e-01
-4.91882712e-01 2.64686167e-01 3.78576845e-01 7.86546350e-01
-1.26373172e+00 7.12145448e-01 -5.80797434e-01 -4.93077040e-01
2.42112771e-01 -2.04726040e-01 -1.03728689e-01 -1.64479777e-01
-8.73926461e-01 1.04390895e+00 3.54303092e-01 3.00887257e-01
-1.08433604e+00 -7.84526467e-01 -6.69075489e-01 3.06763053e-01
2.38808528e-01 -7.49560297e-02 8.70675206e-01 -9.02910471e-01
-1.13522160e+00 9.10253763e-01 4.27841961e-01 -6.40003145e-01
1.95285499e-01 -1.45541370e-01 -6.15657508e-01 4.98493671e-01
-9.93304998e-02 5.11956155e-01 1.37931836e+00 -1.81908309e+00
-8.14460456e-01 -3.63600850e-01 6.20019175e-02 7.41697624e-02
-2.26903260e-01 -2.02623874e-01 2.49144882e-02 -1.04379988e+00
3.86496156e-01 -6.73110485e-01 6.34902045e-02 5.70209563e-01
-7.59270564e-02 9.63724077e-01 7.47084200e-01 -8.70457709e-01
1.17482591e+00 -2.28322387e+00 -4.70040947e-01 1.66497990e-01
2.36845046e-01 5.24321616e-01 -2.78446883e-01 3.48620236e-01
-3.28113958e-02 2.61294454e-01 -6.73891962e-01 5.08122861e-01
-3.17292392e-01 6.86273128e-02 -4.62378621e-01 7.01522648e-01
5.57229459e-01 6.72777951e-01 -9.72512424e-01 -1.90904453e-01
5.53914726e-01 3.73028010e-01 -6.47193938e-02 1.96715578e-01
9.93437767e-02 -8.62389430e-02 9.12856683e-02 1.03493905e+00
1.14663315e+00 -5.77291427e-03 -1.89495713e-01 -3.63244504e-01
-2.81026095e-01 -4.32053924e-01 -8.74749899e-01 1.10186672e+00
-4.46934044e-01 1.00429428e+00 2.37752125e-01 -7.73597300e-01
1.00424635e+00 -1.70942713e-02 1.45691931e-01 -5.21062016e-01
-2.68138871e-02 8.23714435e-02 -2.32300192e-01 -1.49544358e-01
7.96823204e-01 1.96235627e-01 3.44391584e-01 5.83678931e-02
-1.55250087e-01 -6.40821457e-01 -1.48792654e-01 -5.00441454e-02
1.01925993e+00 -4.94928360e-02 6.01887405e-01 -5.62026680e-01
3.05089176e-01 2.27426365e-01 1.92459971e-01 7.83546805e-01
-5.56907892e-01 8.97581160e-01 9.12587270e-02 -2.16861367e-01
-1.05403352e+00 -1.39772940e+00 -3.02788883e-01 7.93037415e-01
6.54597402e-01 3.04224640e-02 -7.15892136e-01 -1.04732044e-01
1.13376893e-01 7.73594439e-01 -4.87348527e-01 -3.88416290e-01
1.12889916e-01 -7.54961550e-01 7.08758593e-01 1.57672957e-01
8.77884865e-01 -6.87616885e-01 -9.88487124e-01 1.78222179e-01
-1.31839635e-02 -1.16610539e+00 -2.68095702e-01 -4.91196811e-02
-7.49940991e-01 -9.72232699e-01 -8.54012966e-01 -3.67321610e-01
5.61615765e-01 1.19926786e+00 1.22932363e+00 2.33429775e-01
-6.70998633e-01 4.36334431e-01 -7.64789462e-01 -6.74676836e-01
-5.36359131e-01 -4.18949962e-01 -3.36253166e-01 3.02586015e-02
2.27754012e-01 -6.13711655e-01 -1.06475270e+00 3.81548822e-01
-1.36982822e+00 -2.65552372e-01 8.12458873e-01 6.76628768e-01
4.48195189e-01 5.75811207e-01 -1.33085266e-01 -1.76942110e-01
4.38190699e-01 -2.72599161e-01 -9.35273349e-01 3.50064486e-01
-9.72313225e-01 -2.92234838e-01 1.37413144e-01 -2.11200804e-01
-6.63012087e-01 -1.79848880e-01 4.49462801e-01 -3.47483277e-01
-1.98390797e-01 7.46641159e-01 -3.82140353e-02 -4.27095056e-01
8.38167191e-01 2.49434173e-01 -3.05564348e-02 -1.24294356e-01
3.47936273e-01 6.32292211e-01 1.18329644e+00 4.79648523e-02
1.31254351e+00 6.19254947e-01 1.86115295e-01 -1.38180971e+00
-4.29179698e-01 -4.86680239e-01 -2.99045086e-01 -4.42953378e-01
7.45668054e-01 -1.15873468e+00 -6.85385585e-01 8.07883322e-01
-1.01489842e+00 -2.42322251e-01 -4.86243814e-02 4.02230531e-01
-1.57102242e-01 5.84380150e-01 -2.31054530e-01 -9.52908695e-01
-3.35108876e-01 -8.36027026e-01 1.17054617e+00 1.40305925e-02
3.47739220e-01 -8.61289084e-01 -3.39726299e-01 -7.67099410e-02
5.95803142e-01 7.13952720e-01 6.42515719e-01 3.75264376e-01
-7.30279326e-01 -1.42320648e-01 -8.31335545e-01 7.23577857e-01
3.12998861e-01 4.48415935e-01 -1.12695837e+00 -2.14079052e-01
-2.96588615e-02 2.88666189e-01 8.92959595e-01 7.68119991e-01
9.78130221e-01 -1.04989573e-01 1.62746608e-01 7.78780460e-01
1.95431983e+00 -5.50721809e-02 1.03217769e+00 4.14827138e-01
5.72829068e-01 7.75393546e-01 1.05615115e+00 6.07135892e-01
-1.72639757e-01 4.80275691e-01 1.06775105e+00 -5.06104589e-01
-1.18021771e-01 9.44275968e-03 4.73286897e-01 2.16736421e-01
-1.48418963e-01 -4.30029273e-01 -1.10691297e+00 6.14486217e-01
-1.46153820e+00 -9.95858908e-01 -1.48664787e-02 2.36409283e+00
4.76483226e-01 -8.26799572e-02 -2.99977392e-01 3.34373474e-01
9.07795787e-01 4.15239364e-01 -3.57920915e-01 -1.87496424e-01
-5.00187576e-01 3.96969244e-02 1.63632238e+00 3.42588991e-01
-1.19310176e+00 9.69447255e-01 6.73331404e+00 6.49934649e-01
-1.24039125e+00 -8.88735354e-02 4.61650014e-01 4.16826755e-01
-1.81323767e-01 -8.26674849e-02 -1.51485130e-01 2.00978369e-01
7.51667082e-01 -4.24334168e-01 4.02837276e-01 6.85054362e-01
4.15813446e-01 -4.22518760e-01 -6.29762530e-01 1.13620746e+00
3.51606216e-03 -1.25826561e+00 4.01213616e-02 1.52181029e-01
6.26744628e-01 -1.64267406e-01 5.71754158e-01 -4.24682915e-01
5.41724443e-01 -1.12725377e+00 7.66734362e-01 2.87027955e-01
9.43646193e-01 -5.86836636e-01 6.35077417e-01 -1.00627631e-01
-1.40810406e+00 -2.42434945e-02 -3.71565878e-01 -2.77811964e-03
-1.34150490e-01 7.68680692e-01 -7.54035950e-01 6.23615742e-01
9.45471466e-01 6.76097870e-01 -7.01784432e-01 1.06308138e+00
-2.80045182e-01 6.68185055e-01 -1.52508244e-01 4.88787591e-01
2.53187925e-01 -3.16663235e-01 6.40739739e-01 1.01843560e+00
6.58368826e-01 1.95582062e-01 -1.03952579e-01 9.56757188e-01
3.07646751e-01 -5.99226840e-02 -7.74968743e-01 -2.03641623e-01
5.81641257e-01 1.07098198e+00 -7.69981802e-01 1.20577991e-01
-1.82526127e-01 1.11658251e+00 -6.32967532e-01 4.18404251e-01
-6.23178482e-01 -5.37673831e-01 9.32173729e-01 1.78078618e-02
4.00674999e-01 -3.95320445e-01 -3.19772542e-01 -8.02649498e-01
-1.61174238e-02 -8.25236082e-01 -1.18571043e-01 -1.17759788e+00
-1.12123001e+00 2.59700447e-01 -4.30545062e-02 -1.68964052e+00
2.08837628e-01 -8.25590670e-01 -5.58560729e-01 9.13196146e-01
-1.97572303e+00 -1.14460039e+00 -1.17236960e+00 1.50499195e-01
3.79249662e-01 -2.85307970e-03 6.97984457e-01 2.23899987e-02
-3.07794809e-01 3.26314658e-01 2.72284031e-01 -8.77716467e-02
4.94745076e-01 -1.18888354e+00 5.66110849e-01 1.65766227e+00
-2.83431292e-01 1.14902958e-01 1.15271187e+00 -4.50079411e-01
-1.26443279e+00 -1.39591932e+00 1.07677482e-01 -2.16889009e-02
6.73485577e-01 3.93039465e-01 -8.83085608e-01 2.09834352e-01
2.90496442e-02 1.87913358e-01 3.79608721e-01 -5.94868839e-01
-3.89305264e-01 -2.89412141e-01 -1.26444256e+00 6.50153041e-01
7.18983531e-01 -6.10103846e-01 -1.30177587e-01 3.61465216e-02
6.57685697e-01 4.55308035e-02 -1.00291431e+00 4.91135955e-01
6.15644693e-01 -1.01785588e+00 1.29400134e+00 8.05013999e-02
5.90514660e-01 -7.17553616e-01 -6.69720232e-01 -1.38625109e+00
-5.40820599e-01 -6.26374543e-01 2.57614702e-01 9.34942901e-01
1.77000418e-01 -3.35769594e-01 3.13511193e-01 6.27422780e-02
1.60701960e-01 8.16970617e-02 -2.37110481e-01 -1.19994032e+00
-1.41549587e-01 -3.39334548e-01 7.27544665e-01 8.81850421e-01
-5.73041916e-01 -4.18716133e-01 -4.15720195e-01 1.11899102e+00
9.56022143e-01 -1.57395735e-01 5.90691864e-01 -1.45925689e+00
-7.01820254e-02 -5.00322700e-01 -8.93563211e-01 -5.02969682e-01
-3.79604936e-01 -3.72660369e-01 3.76737624e-01 -1.29327667e+00
-3.15197438e-01 -4.31624129e-02 -3.54874730e-01 2.50288844e-01
-4.62448269e-01 6.44869566e-01 2.83164024e-01 3.37087572e-01
1.68042481e-01 4.10182565e-01 7.83163965e-01 -5.04993379e-01
-9.99657288e-02 -1.39103606e-01 -6.21267557e-01 6.78490341e-01
8.89242530e-01 -4.07029271e-01 -2.92658538e-01 -5.22009790e-01
1.15577020e-01 -2.77204603e-01 8.02266121e-01 -1.17020845e+00
-1.47998735e-01 -4.39834893e-01 1.48214579e-01 -5.00697792e-01
8.63074139e-02 -1.05063677e+00 9.31951702e-02 6.56224132e-01
-1.09159179e-01 6.32652938e-02 7.61886016e-02 6.77615285e-01
-3.53505582e-01 -2.37567857e-01 9.18849885e-01 4.53325510e-02
-1.18398106e+00 2.75276631e-01 -4.33345944e-01 -4.62622911e-01
9.43957925e-01 -6.04219258e-01 -5.55517852e-01 -4.93282557e-01
-7.89343342e-02 -2.02782691e-01 7.89280832e-01 2.60516047e-01
8.11349213e-01 -1.04821837e+00 -1.04127276e+00 1.78536326e-01
7.74505496e-01 -4.07808244e-01 -1.38474882e-01 6.38528705e-01
-1.35632730e+00 -7.94419944e-02 -5.35800278e-01 -7.41599739e-01
-1.30887747e+00 4.45536345e-01 7.90672451e-02 1.48045123e-01
-4.35239464e-01 6.40262187e-01 2.13111103e-01 2.24100903e-01
-9.95806828e-02 -4.98152465e-01 -5.84583506e-02 -1.36549950e-01
8.21288943e-01 4.68869954e-01 2.42894426e-01 -6.72603548e-01
-1.92413136e-01 3.42149287e-01 3.27087760e-01 -2.70040184e-01
1.23483801e+00 -4.45473015e-01 -8.03581849e-02 2.61763215e-01
9.06454086e-01 -1.96322620e-01 -1.48178220e+00 -2.33540207e-01
8.38560164e-02 -1.08946049e+00 8.29769671e-01 -5.31394601e-01
-1.27145481e+00 8.60453367e-01 1.23639560e+00 6.49071515e-01
1.59420037e+00 -7.21243620e-01 4.41345394e-01 4.78622615e-01
3.44367534e-01 -9.68272328e-01 2.73655225e-02 7.57776573e-02
9.09433663e-01 -1.40970027e+00 3.83900136e-01 -7.05664039e-01
-6.32224441e-01 8.93024862e-01 1.07028596e-01 -6.39445558e-02
6.06891215e-01 2.85527527e-01 4.61632609e-01 -4.52489443e-02
-2.48937517e-01 -6.36088192e-01 9.08501148e-02 1.17691815e+00
3.99904251e-02 9.36998874e-02 -1.83606334e-02 -9.27114636e-02
-4.12418731e-02 -8.26185644e-02 8.22220325e-01 1.03843462e+00
-9.84985352e-01 -4.48845685e-01 -7.26615131e-01 1.29947633e-01
-3.37446779e-01 -3.44587982e-01 -1.74278304e-01 4.11119878e-01
-1.27945364e-01 1.34530568e+00 -5.12990914e-03 -3.75447810e-01
1.98935598e-01 -6.67207778e-01 3.11524659e-01 -4.20132130e-01
-2.68704325e-01 -1.07186399e-01 -3.60700637e-01 -5.00816405e-01
-5.54680765e-01 -4.60646003e-01 -6.07737601e-01 -6.25073195e-01
-4.67235804e-01 -4.40641075e-01 9.44361746e-01 5.39646506e-01
2.20702544e-01 5.30113459e-01 8.39953244e-01 -9.11544979e-01
-3.49235982e-01 -6.56948447e-01 -1.27454984e+00 3.18758458e-01
6.86869860e-01 -6.53191149e-01 -4.76744056e-01 5.84581196e-01] | [10.922576904296875, -2.965663194656372] |
ca5f551b-fc79-490c-b015-a8eef066fb11 | multilingual-training-for-software | 2112.02043 | null | https://arxiv.org/abs/2112.02043v4 | https://arxiv.org/pdf/2112.02043v4.pdf | Multilingual training for Software Engineering | Well-trained machine-learning models, which leverage large amounts of open-source software data, have now become an interesting approach to automating many software engineering tasks. Several SE tasks have all been subject to this approach, with performance gradually improving over the past several years with better models and training methods. More, and more diverse, clean, labeled data is better for training; but constructing good-quality datasets is time-consuming and challenging. Ways of augmenting the volume and diversity of clean, labeled data generally have wide applicability. For some languages (e.g., Ruby) labeled data is less abundant; in others (e.g., JavaScript) the available data maybe more focused on some application domains, and thus less diverse. As a way around such data bottlenecks, we present evidence suggesting that human-written code in different languages (which performs the same function), is rather similar, and particularly preserving of identifier naming patterns; we further present evidence suggesting that identifiers are a very important element of training data for software engineering tasks. We leverage this rather fortuitous phenomenon to find evidence that available multilingual training data (across different languages) can be used to amplify performance. We study this for 3 different tasks: code summarization, code retrieval, and function naming. We note that this data-augmenting approach is broadly compatible with different tasks, languages, and machine-learning models. | ['Premkumar Devanbu', 'Toufique Ahmed'] | 2021-12-03 | null | null | null | null | ['type-prediction'] | ['computer-code'] | [-1.70521975e-01 -1.17792428e-01 -4.47557360e-01 -3.18251133e-01
-8.29949081e-01 -9.46841359e-01 5.28631330e-01 4.60633695e-01
-3.84818584e-01 6.54827714e-01 4.99509275e-01 -6.36478066e-01
3.47856246e-02 -4.65642095e-01 -7.33020067e-01 -3.06885183e-01
1.48443893e-01 7.53702670e-02 -7.07035512e-02 -5.91542006e-01
5.02507865e-01 7.54729360e-02 -1.64669323e+00 4.56054062e-01
1.07144701e+00 7.80473500e-02 2.97271371e-01 3.67544293e-01
-6.52161837e-01 9.67729926e-01 -8.66248488e-01 -4.19284999e-01
1.77077323e-01 -1.19376153e-01 -1.11519444e+00 -2.48810530e-01
5.87466359e-01 3.30976576e-01 1.45066425e-01 1.05413556e+00
3.18414569e-01 -2.73392618e-01 2.18760282e-01 -1.28651583e+00
-9.33338761e-01 1.01908588e+00 -6.81332469e-01 2.17569452e-02
3.93603534e-01 2.26683825e-01 1.11321795e+00 -6.53097987e-01
9.53285992e-01 7.43167758e-01 8.49556565e-01 5.63200116e-01
-1.26364720e+00 -6.05345964e-01 -1.79664612e-01 -1.31860688e-01
-1.21999681e+00 -4.92536485e-01 5.05877316e-01 -8.55061829e-01
1.10213220e+00 4.04947340e-01 3.34356964e-01 1.11335957e+00
2.24834159e-01 4.12286252e-01 9.55833197e-01 -8.45551372e-01
-1.10757858e-01 5.20770788e-01 3.31356108e-01 6.20325387e-01
8.76977086e-01 -3.90349805e-01 -4.48968977e-01 -5.66354692e-01
5.15069999e-02 2.60899812e-01 -3.67805749e-01 -4.70514268e-01
-1.28654778e+00 6.28176749e-01 -9.79117230e-02 8.59470129e-01
-9.54535976e-02 -5.30495457e-02 8.29701304e-01 7.23444045e-01
4.69422340e-01 1.29661119e+00 -9.24115658e-01 -7.19755828e-01
-1.11710930e+00 1.76786885e-01 1.33229351e+00 1.32943976e+00
1.18554640e+00 1.65291363e-03 3.50778699e-01 1.06629455e+00
7.38650709e-02 2.91244656e-01 7.98868954e-01 -1.00617790e+00
8.01117241e-01 9.79011357e-01 -1.69239327e-01 -5.82547843e-01
-2.56992996e-01 -2.29833990e-01 -3.74407560e-01 1.93769231e-01
5.23934305e-01 -1.41285121e-01 -5.92261136e-01 1.70275807e+00
4.91235312e-03 -7.41568923e-01 -3.18068564e-02 3.06539923e-01
6.03361249e-01 2.71531105e-01 1.34372771e-01 9.99930222e-03
1.31641257e+00 -8.34352553e-01 -4.90849942e-01 -4.38378274e-01
1.28115690e+00 -1.07499516e+00 1.27443838e+00 3.71754408e-01
-7.07187116e-01 -3.74628574e-01 -9.31571305e-01 -3.20529729e-01
-6.24238729e-01 4.18884531e-02 9.02650416e-01 8.82216692e-01
-1.09695399e+00 6.52413905e-01 -5.47542870e-01 -6.62748218e-01
3.84545714e-01 -9.49184895e-02 -7.08958864e-01 -2.64623612e-01
-6.75006390e-01 9.79006648e-01 3.20630729e-01 -5.07173955e-01
-5.79406857e-01 -8.70416820e-01 -9.23704147e-01 9.44791138e-02
5.53562880e-01 -2.94760197e-01 1.23662686e+00 -1.07398009e+00
-8.54224324e-01 1.15104198e+00 -1.00333825e-01 6.45363554e-02
-1.11092720e-02 -2.98187256e-01 -4.48232472e-01 -5.93158424e-01
4.55050170e-01 9.83247682e-02 7.14014053e-01 -1.33619618e+00
-5.67353129e-01 -4.46864098e-01 1.97993681e-01 -3.99941444e-01
-5.62043965e-01 6.42231822e-01 2.75317039e-02 -5.28489649e-01
-3.96301866e-01 -9.01371241e-01 -6.98748976e-02 -3.53376329e-01
-2.43675128e-01 -3.62172186e-01 5.01851678e-01 -8.72262120e-01
1.73567796e+00 -2.10563922e+00 2.09140629e-01 -4.47354950e-02
6.74580634e-01 3.49202901e-01 -2.63075680e-01 8.21121395e-01
-3.59948397e-01 7.65866518e-01 -4.91920382e-01 -9.43183601e-02
-1.14447035e-01 1.00938261e-01 -1.68078750e-01 2.64691472e-01
2.43682548e-01 6.34102583e-01 -8.45946074e-01 -3.44079524e-01
-3.47042292e-01 4.95016724e-02 -6.54177487e-01 5.87989800e-02
-1.83485538e-01 1.75096601e-01 -2.33868986e-01 8.07757854e-01
2.54516244e-01 -4.15269554e-01 2.39930987e-01 2.16783926e-01
-5.87107599e-01 4.04116809e-01 -9.43126082e-01 1.92912591e+00
-8.42485070e-01 9.91522253e-01 -1.45077361e-02 -7.92830110e-01
7.88772583e-01 2.30590984e-01 4.14111316e-01 -4.41652298e-01
-1.54308870e-01 4.70668286e-01 3.60915750e-01 -9.81415749e-01
7.22087026e-01 2.50307202e-01 -3.03024113e-01 8.35419118e-01
7.98171610e-02 -2.98293918e-01 5.40454566e-01 3.97507876e-01
1.46829748e+00 2.87179083e-01 5.93341708e-01 -4.82796967e-01
4.44525063e-01 4.27673787e-01 6.18150711e-01 6.27583265e-01
6.78135827e-02 3.87433827e-01 6.02417648e-01 -3.17779183e-01
-1.50634813e+00 -4.49696302e-01 -2.34313428e-01 1.46818936e+00
-5.03090143e-01 -1.00569987e+00 -6.80377364e-01 -6.83063507e-01
6.97467402e-02 7.15189934e-01 -4.34925467e-01 -1.70113653e-01
-7.53706336e-01 -5.09849727e-01 9.47022200e-01 4.01770890e-01
-6.83842599e-02 -1.00701809e+00 -5.16744077e-01 1.71162546e-01
-1.61686778e-01 -7.46814907e-01 -3.57112765e-01 3.59645188e-01
-7.61160195e-01 -1.24128425e+00 -5.68868399e-01 -6.62266970e-01
5.61000645e-01 3.80255550e-01 1.83092225e+00 5.69146693e-01
-4.14767861e-01 5.37333846e-01 -6.59911573e-01 -5.85049808e-01
-1.11435318e+00 3.68765920e-01 -1.41867056e-01 -6.34356439e-01
7.09607005e-01 -4.42355603e-01 7.23225176e-02 3.49996202e-02
-9.83057618e-01 -2.70606637e-01 7.67179668e-01 6.22252345e-01
-5.44733815e-02 -4.24367636e-01 5.58062077e-01 -1.46995771e+00
7.29512095e-01 -1.08608067e+00 -3.50141793e-01 4.53367174e-01
-8.11631382e-01 3.88105601e-01 8.47230077e-01 -4.03663874e-01
-9.05796826e-01 -2.84499943e-01 -3.48787270e-02 8.57573841e-03
-2.57623076e-01 7.95950830e-01 5.81867136e-02 -7.41204321e-02
1.26251233e+00 1.41287432e-03 2.46830106e-01 -8.38334799e-01
4.65843648e-01 1.10515952e+00 5.88536225e-02 -9.34015572e-01
8.61665428e-01 2.41146401e-01 -5.82440972e-01 -9.63847816e-01
-3.24577063e-01 -5.96715510e-01 -4.58290398e-01 2.37713590e-01
5.13642132e-01 -7.78518796e-01 -1.27841756e-01 3.32898587e-01
-1.21046305e+00 -2.88879275e-01 -4.13962483e-01 3.07991803e-01
-1.95867866e-01 4.61389780e-01 -3.87165248e-01 -4.11403745e-01
-2.14206681e-01 -1.17108643e+00 9.89832878e-01 4.71179970e-02
-6.31371915e-01 -9.18898225e-01 4.66293067e-01 3.54260266e-01
7.13571668e-01 1.14043288e-01 1.39561367e+00 -1.07815051e+00
-4.96009201e-01 -1.74694106e-01 -1.33564383e-01 5.97653031e-01
5.20212173e-01 4.41581547e-01 -8.59455943e-01 -3.01611543e-01
3.41260172e-02 -4.46478665e-01 4.85006839e-01 -3.25850010e-01
1.01568806e+00 -1.34968832e-01 -2.82017469e-01 3.04240823e-01
1.48143494e+00 -2.61564255e-02 4.53012824e-01 4.40492690e-01
1.02950037e+00 8.93697619e-01 2.84012467e-01 2.19443992e-01
4.51365769e-01 4.39358056e-01 -1.58911526e-01 5.95170669e-02
-2.34687269e-01 1.50166348e-01 4.45241243e-01 1.41583824e+00
-9.92833003e-02 8.77943039e-02 -1.35952866e+00 7.93552816e-01
-1.54160678e+00 -9.08773899e-01 -5.21137118e-01 2.09601569e+00
1.31797266e+00 -1.51093990e-01 1.24546681e-02 -3.35596316e-02
4.18447524e-01 5.20018674e-02 -2.18385413e-01 -4.49535787e-01
-3.01007498e-02 3.39855969e-01 3.04579347e-01 1.31035494e-02
-5.14595091e-01 5.22815228e-01 6.60816193e+00 6.37596846e-01
-1.07519090e+00 3.21258664e-01 7.12403934e-03 1.51824921e-01
-5.99177778e-01 3.79151493e-01 -5.73043942e-01 7.23302245e-01
1.13006485e+00 -4.78158653e-01 5.32448828e-01 1.11134112e+00
-1.19034767e-01 -1.58765867e-01 -1.51370776e+00 7.51432896e-01
2.36570299e-01 -1.32969642e+00 -2.51167804e-01 1.23847947e-01
8.11837018e-01 4.69533741e-01 -4.17714834e-01 6.71941578e-01
6.50693119e-01 -7.44654238e-01 7.33044684e-01 3.21299702e-01
7.68262923e-01 -1.74115390e-01 6.91497922e-01 4.50305164e-01
-8.99977744e-01 -1.23507775e-01 -3.39840591e-01 -8.76698121e-02
-4.37867314e-01 5.99554181e-01 -4.43826020e-01 5.13863683e-01
8.39384198e-01 7.70527780e-01 -1.26269639e+00 9.44658577e-01
-5.18220626e-02 5.68338752e-01 7.07187206e-02 -8.48302543e-02
-5.60133941e-02 -5.00547811e-02 4.31914002e-01 1.43001282e+00
1.85455784e-01 -5.93751729e-01 5.42949699e-02 1.00458443e+00
-2.15042278e-01 2.80401260e-01 -1.20510077e+00 -4.18856025e-01
6.10690296e-01 1.25775588e+00 -3.68550867e-01 -3.35214227e-01
-1.07407391e+00 4.77059722e-01 4.83975768e-01 2.05064312e-01
-5.72348356e-01 -6.69555008e-01 8.18666160e-01 3.01230568e-02
-1.21256307e-01 -2.59600282e-01 -2.20467418e-01 -1.39484680e+00
3.54781747e-01 -1.61810255e+00 3.13585460e-01 -6.35233164e-01
-1.40356064e+00 3.82681549e-01 -8.94620046e-02 -1.20816791e+00
-2.58907914e-01 -4.94847953e-01 -3.92360508e-01 9.10920918e-01
-1.34082472e+00 -8.63139033e-01 -2.92470783e-01 3.46438825e-01
6.08234584e-01 -4.75697279e-01 7.34744668e-01 6.80789351e-01
-3.97655785e-01 4.61732447e-01 5.16550839e-01 3.32616359e-01
1.22815430e+00 -1.39586270e+00 5.62571883e-01 9.03941929e-01
3.46014619e-01 1.49510205e+00 4.30429131e-01 -6.70644403e-01
-1.66595864e+00 -8.34096849e-01 9.46750998e-01 -1.12124944e+00
8.84031832e-01 -4.53456759e-01 -1.30596280e+00 7.59799838e-01
3.82306486e-01 -2.98370630e-01 9.89118755e-01 6.63011432e-01
-9.10892069e-01 -6.15006350e-02 -8.81170630e-01 4.08296019e-01
9.96527791e-01 -8.12377155e-01 -7.75155127e-01 3.83187085e-01
7.25778282e-01 -1.39486089e-01 -1.10998976e+00 -1.59741998e-01
3.42960536e-01 -8.02565813e-01 4.78579730e-01 -1.05436599e+00
7.02615619e-01 -1.53450772e-01 -3.94413322e-02 -1.50018525e+00
-1.93350434e-01 -7.38571227e-01 2.66213566e-01 1.56674445e+00
7.10834861e-01 -7.96045601e-01 4.30385381e-01 7.61599660e-01
-4.24717873e-01 -3.71855080e-01 -5.04777014e-01 -9.95324671e-01
4.38706905e-01 -3.56906623e-01 6.44639075e-01 1.43886125e+00
3.66352260e-01 4.03497487e-01 -2.17071958e-02 -4.31827247e-01
2.54127443e-01 9.39644575e-02 1.17413843e+00 -1.41024923e+00
-2.08533496e-01 -4.56574768e-01 -3.08086634e-01 -4.84625936e-01
3.46945673e-01 -1.29560852e+00 3.50122969e-03 -1.31383348e+00
4.40269619e-01 -7.62054741e-01 1.21954672e-01 7.48227417e-01
-3.37308586e-01 -2.38881528e-01 1.99273750e-01 5.69763899e-01
-4.80643660e-01 -2.36573428e-01 6.53471410e-01 -1.42423183e-01
-1.08859353e-01 -2.86920965e-01 -1.16507995e+00 7.31925845e-01
6.39430881e-01 -7.92230904e-01 -7.79392123e-02 -7.00677514e-01
6.89326704e-01 -3.42143476e-01 -1.32200778e-01 -8.88440251e-01
1.86660647e-01 -1.23783685e-01 -2.47996658e-01 2.48516917e-01
-4.40962642e-01 -9.11933184e-01 1.65407926e-01 4.45416987e-01
-2.56247997e-01 3.07132214e-01 2.87550598e-01 2.20721126e-01
-1.50238305e-01 -7.21782625e-01 5.48550546e-01 -4.82819468e-01
-8.11248004e-01 -1.43250331e-01 -4.17686492e-01 7.36597955e-01
7.99623907e-01 -1.84458271e-01 -7.91079164e-01 9.78303328e-02
-1.29696816e-01 -3.06835007e-02 9.92858171e-01 8.88089716e-01
-8.36832300e-02 -9.53937590e-01 -8.57273757e-01 1.93160012e-01
6.29204750e-01 -3.40544760e-01 -2.44394064e-01 6.51321411e-01
-4.73213971e-01 3.19523722e-01 -2.76015610e-01 -4.80812103e-01
-1.10652518e+00 6.23317122e-01 3.72656919e-02 -2.26207912e-01
-4.04590517e-01 4.91645426e-01 -1.86810061e-01 -7.26661146e-01
-1.45276383e-01 -3.94522101e-01 -5.77890426e-02 1.90175265e-01
4.38809186e-01 4.94473726e-01 3.89178813e-01 -3.41010928e-01
-3.94603014e-01 3.55198562e-01 -2.66684562e-01 3.91726822e-01
1.51959205e+00 1.77257180e-01 -8.56491327e-01 7.62859941e-01
1.26900625e+00 7.35683262e-01 -5.34993649e-01 -2.53116786e-01
5.76321661e-01 -5.84575295e-01 -5.05943179e-01 -7.26474404e-01
-8.64240348e-01 6.70285285e-01 8.04895014e-02 6.74056292e-01
6.79301858e-01 2.04936534e-01 2.76908934e-01 4.96196568e-01
8.49804044e-01 -1.02214420e+00 -1.67478472e-01 4.90821749e-01
6.73659205e-01 -1.37877834e+00 1.85652822e-01 -1.91152006e-01
-4.13412511e-01 1.14942455e+00 6.83824599e-01 3.32949996e-01
3.76203597e-01 4.64240521e-01 8.55918080e-02 -3.30951542e-01
-6.87088966e-01 -2.50970963e-02 -1.32779911e-01 6.10110104e-01
1.03100276e+00 -1.43735960e-01 -5.06894290e-01 4.84853566e-01
-9.42267030e-02 -1.36896491e-01 1.03749514e+00 1.26013565e+00
-2.91123867e-01 -1.50634682e+00 -4.24065471e-01 8.21442306e-01
-6.73602045e-01 -4.68043625e-01 -5.12256563e-01 9.66531277e-01
2.26804137e-01 9.55253243e-01 -2.90958017e-01 -1.06041566e-01
3.71960402e-01 4.68019903e-01 2.92832822e-01 -1.33790910e+00
-1.06967866e+00 -6.97373986e-01 1.81723535e-01 -4.66926962e-01
-3.49057436e-01 -6.35863364e-01 -8.68607998e-01 -4.84486848e-01
-5.16006827e-01 3.26907128e-01 7.96293855e-01 7.22283840e-01
6.70057774e-01 4.95506376e-01 3.77878487e-01 -3.99598241e-01
-6.19648337e-01 -9.99567568e-01 -3.02243859e-01 6.97752476e-01
2.03741610e-01 -4.20717418e-01 -4.69281763e-01 3.64011079e-01] | [7.694273471832275, 7.906487464904785] |
30034962-c245-4a85-a9dc-d489f874e9fe | language-free-compositional-action-generation | 2307.03538 | null | https://arxiv.org/abs/2307.03538v1 | https://arxiv.org/pdf/2307.03538v1.pdf | Language-free Compositional Action Generation via Decoupling Refinement | Composing simple elements into complex concepts is crucial yet challenging, especially for 3D action generation. Existing methods largely rely on extensive neural language annotations to discern composable latent semantics, a process that is often costly and labor-intensive. In this study, we introduce a novel framework to generate compositional actions without reliance on language auxiliaries. Our approach consists of three main components: Action Coupling, Conditional Action Generation, and Decoupling Refinement. Action Coupling utilizes an energy model to extract the attention masks of each sub-action, subsequently integrating two actions using these attentions to generate pseudo-training examples. Then, we employ a conditional generative model, CVAE, to learn a latent space, facilitating the diverse generation. Finally, we propose Decoupling Refinement, which leverages a self-supervised pre-trained model MAE to ensure semantic consistency between the sub-actions and compositional actions. This refinement process involves rendering generated 3D actions into 2D space, decoupling these images into two sub-segments, using the MAE model to restore the complete image from sub-segments, and constraining the recovered images to match images rendered from raw sub-actions. Due to the lack of existing datasets containing both sub-actions and compositional actions, we created two new datasets, named HumanAct-C and UESTC-C, and present a corresponding evaluation metric. Both qualitative and quantitative assessments are conducted to show our efficacy. | ['Ser-Nam Lim', 'Guangrun Wang', 'Yansong Tang', 'Guangyi Chen', 'Xiao Liu'] | 2023-07-07 | null | null | null | null | ['action-generation'] | ['computer-vision'] | [ 6.74763918e-01 2.78035581e-01 1.22629806e-01 -2.18358666e-01
-9.24517095e-01 -6.04802072e-01 1.01472771e+00 -4.26077783e-01
-1.90121643e-02 4.69596356e-01 5.86380839e-01 2.43656840e-02
3.52264762e-01 -8.45959604e-01 -8.94501567e-01 -5.99559307e-01
3.70022655e-01 2.06367195e-01 7.79409632e-02 -1.32097930e-01
8.65821466e-02 2.97585785e-01 -1.64496648e+00 6.03404343e-01
9.90040898e-01 8.24632704e-01 2.90037215e-01 5.34352422e-01
-1.61072373e-01 9.08027112e-01 -5.98818123e-01 -4.21229541e-01
3.17876875e-01 -8.92957211e-01 -8.02465260e-01 6.17985547e-01
2.05952644e-01 -6.17875278e-01 2.56638173e-02 8.85765553e-01
2.60766357e-01 2.78892875e-01 8.09579372e-01 -1.28996480e+00
-1.02826798e+00 4.73394126e-01 -3.06889236e-01 -5.45035660e-01
7.28346050e-01 4.58380699e-01 1.01854372e+00 -8.62916827e-01
8.20449889e-01 1.55377090e+00 4.42121893e-01 8.45666051e-01
-1.47306883e+00 -7.03725517e-01 3.06947857e-01 -1.34811476e-01
-1.15542185e+00 -3.82803172e-01 8.83806467e-01 -5.67967176e-01
9.18743253e-01 9.38768983e-02 7.29910672e-01 1.45446920e+00
-9.70430896e-02 9.27881241e-01 1.27141321e+00 -3.97719920e-01
2.28591427e-01 -1.06598750e-01 -5.08901238e-01 5.42686522e-01
-2.49041498e-01 9.24770012e-02 -4.81460184e-01 1.98576108e-01
1.10982907e+00 -7.51444884e-03 -2.19182074e-01 -2.78618693e-01
-1.37913764e+00 6.25476778e-01 3.54605913e-01 1.19341426e-01
-5.92829466e-01 2.38363370e-01 -7.06490502e-02 -4.81012389e-02
5.26906848e-01 4.43169564e-01 -8.74138251e-02 6.96944594e-02
-8.84039700e-01 4.33716983e-01 4.96269315e-01 1.16967201e+00
7.72957027e-01 -4.16273326e-02 -5.83641052e-01 6.81934476e-01
2.46111169e-01 3.24119210e-01 3.81218195e-01 -1.12338769e+00
4.65269864e-01 8.55340183e-01 1.23331264e-01 -7.51532733e-01
6.40816316e-02 -2.31345594e-01 -7.86524355e-01 1.58646896e-01
-4.01655473e-02 8.90691951e-02 -1.21820509e+00 1.95716321e+00
3.88973564e-01 3.03903431e-01 9.10929665e-02 9.01674092e-01
6.66072130e-01 6.64385498e-01 3.36565793e-01 5.66755496e-02
1.20883191e+00 -1.21462226e+00 -5.35019636e-01 -1.00756764e-01
4.91513044e-01 -7.40254819e-01 1.41542399e+00 1.29945111e-02
-1.37065017e+00 -8.08052063e-01 -1.02178609e+00 -3.73847902e-01
-1.73479304e-01 2.90890902e-01 4.44666147e-01 9.38575342e-02
-9.43785846e-01 4.80310887e-01 -8.60769928e-01 -4.25946154e-02
5.49965262e-01 -7.71736056e-02 -3.61515611e-01 -8.52859691e-02
-1.10107017e+00 6.95516706e-01 4.02009666e-01 -4.19035408e-04
-1.30758369e+00 -7.26347506e-01 -1.13009953e+00 -4.11193706e-02
4.47234780e-01 -1.04549778e+00 1.11181700e+00 -1.09429085e+00
-1.76959789e+00 8.95682573e-01 -7.24675208e-02 -2.33051822e-01
6.97788417e-01 -2.76626110e-01 -1.58745199e-01 6.90063536e-02
4.40645248e-01 1.03507733e+00 1.10330319e+00 -1.51599002e+00
-4.53918129e-01 4.74063568e-02 3.96967441e-01 4.55069721e-01
-7.30027929e-02 -2.28678942e-01 -8.10226738e-01 -1.04333413e+00
1.55172318e-01 -9.41073060e-01 -2.45311826e-01 -1.56334803e-01
-6.29765332e-01 -5.15766665e-02 4.94819671e-01 -7.24389732e-01
1.14488447e+00 -2.17178130e+00 6.47454917e-01 -2.67253648e-02
2.36076847e-01 -6.44084066e-02 -4.69328582e-01 2.82116771e-01
-1.12955041e-01 2.40202367e-01 -5.81037045e-01 -7.66034245e-01
2.07889885e-01 2.27527127e-01 -5.32612860e-01 -2.77348273e-02
6.46253288e-01 1.18576634e+00 -9.92790937e-01 -3.16280544e-01
3.55838627e-01 6.35471225e-01 -7.68197715e-01 5.11460721e-01
-6.76451445e-01 7.48096287e-01 -4.83428895e-01 4.47625101e-01
5.13636410e-01 -3.75602454e-01 1.34763241e-01 -4.51584339e-01
2.15523317e-02 3.40542227e-01 -1.03809845e+00 2.13327932e+00
-5.48905671e-01 1.22209527e-01 -2.07009390e-01 -9.38367188e-01
7.73016751e-01 2.70947337e-01 4.21841502e-01 -6.91388428e-01
1.19017102e-01 5.79056032e-02 -4.91488606e-01 -5.23896873e-01
4.52955484e-01 -3.09281826e-01 -2.47363299e-01 7.40591466e-01
1.01830050e-01 -5.96591055e-01 2.16897219e-01 3.74596149e-01
8.90161037e-01 9.59850073e-01 1.02915689e-01 2.21286118e-01
6.14371896e-01 -2.10241199e-01 4.72003311e-01 3.86658549e-01
1.16515361e-01 1.04678094e+00 5.17462790e-01 -1.94752678e-01
-9.48671997e-01 -1.24608064e+00 5.04344404e-01 7.91577160e-01
3.34460109e-01 -5.73632360e-01 -1.04028153e+00 -8.37126374e-01
-2.34475851e-01 9.79432642e-01 -8.00997734e-01 -3.48253965e-01
-4.61499155e-01 -4.28647965e-01 3.90519172e-01 5.91971576e-01
6.40015483e-01 -1.32865739e+00 -6.40441895e-01 1.96168631e-01
-4.54664260e-01 -1.32732856e+00 -6.90440178e-01 -2.72828132e-01
-5.96300602e-01 -9.66348052e-01 -7.92417169e-01 -6.54038131e-01
9.70197916e-01 1.70113742e-01 1.23857474e+00 2.30980158e-01
-1.28679276e-01 3.38064194e-01 -3.78209412e-01 5.84590286e-02
-7.39922285e-01 -1.37293980e-01 -2.85295457e-01 2.24786118e-01
-3.89710218e-02 -7.29700983e-01 -6.98100388e-01 2.57275671e-01
-1.17581153e+00 1.03217781e+00 8.62159550e-01 6.89293444e-01
8.90204132e-01 -2.70678788e-01 3.46759439e-01 -8.25678289e-01
4.72308546e-01 -3.54449600e-01 -3.21088284e-01 2.41606206e-01
-1.64063230e-01 1.86021611e-01 4.77543861e-01 -5.18856943e-01
-1.40014410e+00 1.64408714e-01 2.04324885e-03 -6.90408647e-01
-2.61956155e-01 2.96555310e-01 -5.73124826e-01 5.88999152e-01
5.62911034e-01 4.02020246e-01 -1.50802255e-01 -4.27657396e-01
8.20978761e-01 4.06459451e-01 7.83883989e-01 -8.07789087e-01
9.44656849e-01 5.37148595e-01 -1.97880387e-01 -4.12200451e-01
-1.07565856e+00 4.35928851e-02 -6.78646982e-01 -2.04628691e-01
1.44333482e+00 -9.90506232e-01 -1.59388348e-01 6.10675871e-01
-1.27011645e+00 -6.83712363e-01 -4.05453056e-01 1.77210718e-01
-8.30712199e-01 2.18214422e-01 -4.86405075e-01 -4.89264399e-01
-1.65664464e-01 -1.25993598e+00 1.52499485e+00 4.22883928e-02
-2.38584965e-01 -6.99242651e-01 -6.33823872e-02 5.85830986e-01
2.60370880e-01 6.37045085e-01 8.46534908e-01 -1.94581017e-01
-9.06864643e-01 -1.63504598e-03 -2.46724591e-01 5.32910883e-01
3.74836385e-01 -7.94665068e-02 -8.44982922e-01 6.63951086e-03
-2.10192963e-01 -4.81185198e-01 6.96779370e-01 1.12986542e-01
1.21833897e+00 -3.03518593e-01 -1.49763510e-01 6.04636133e-01
1.10763323e+00 2.11893860e-02 8.12460601e-01 1.07075565e-01
9.52314496e-01 6.65206432e-01 4.72748876e-01 1.71689421e-01
6.43726230e-01 6.85943007e-01 2.70360410e-01 -9.78614986e-02
-5.84803820e-01 -6.95933223e-01 4.11987364e-01 7.70678639e-01
-3.41432780e-01 -1.76032215e-01 -5.97040832e-01 4.82573718e-01
-1.75896287e+00 -9.19426024e-01 3.09077978e-01 1.94460034e+00
1.02988756e+00 1.45883918e-01 -1.45352721e-01 -1.18362628e-01
6.15457475e-01 2.12970570e-01 -4.74432081e-01 9.04892385e-02
-5.94199225e-02 3.43404442e-01 -1.68298587e-01 4.93224233e-01
-9.50443327e-01 1.18317950e+00 5.50807810e+00 7.40331888e-01
-1.08408940e+00 1.50148883e-01 5.98075390e-01 -1.37327939e-01
-7.55926013e-01 2.47912675e-01 -4.21522170e-01 4.39674169e-01
4.27141696e-01 1.17299028e-01 5.20443201e-01 5.39010823e-01
2.36068293e-01 6.82542473e-02 -1.18651760e+00 7.88190007e-01
2.32656687e-01 -1.33368516e+00 5.34238756e-01 -1.31077826e-01
1.04785693e+00 -5.73518336e-01 -1.63812973e-02 4.50933307e-01
5.09716749e-01 -9.96359050e-01 1.09965622e+00 6.52082860e-01
1.01498771e+00 -4.91755575e-01 2.18647689e-01 2.26226047e-01
-1.28579140e+00 1.81458488e-01 8.50392431e-02 6.80315122e-02
4.56807166e-01 2.91377157e-01 -2.82853305e-01 6.79496765e-01
4.65227872e-01 8.60791385e-01 -4.38859671e-01 2.83927083e-01
-7.83801079e-01 2.91751295e-01 4.46318090e-02 4.33171839e-01
1.52467638e-01 -3.96104425e-01 5.13909638e-01 9.28404033e-01
3.23975086e-01 1.45951420e-01 1.39222935e-01 1.48803973e+00
-1.61424190e-01 -1.71210214e-01 -3.63368899e-01 -1.71247885e-01
3.58152598e-01 1.24127519e+00 -7.10968971e-01 -6.17233217e-01
-4.93419200e-01 1.57354021e+00 2.44362637e-01 5.80681384e-01
-1.07906055e+00 -2.22205117e-01 5.83596230e-01 9.22875032e-02
2.46618584e-01 -1.30095005e-01 -1.94795474e-01 -1.32501948e+00
1.17206857e-01 -9.26527560e-01 7.68451318e-02 -1.10372782e+00
-1.10960186e+00 6.58474088e-01 3.15396428e-01 -1.38962507e+00
-4.71525252e-01 -2.07546994e-01 -5.42928338e-01 1.01881802e+00
-1.26359892e+00 -1.69549429e+00 -6.61550641e-01 4.94585663e-01
7.21783340e-01 7.71130323e-02 6.38854027e-01 1.35633856e-01
-5.81244409e-01 4.47877288e-01 -7.33155727e-01 -7.00645298e-02
5.17615616e-01 -1.18256021e+00 6.38127565e-01 9.79389012e-01
9.99723896e-02 5.82560301e-01 4.16366428e-01 -7.33070731e-01
-1.02367663e+00 -1.28774166e+00 6.74585283e-01 -6.29792809e-01
3.95270228e-01 -6.38629913e-01 -8.53367031e-01 7.23464131e-01
9.79929119e-02 -1.39723808e-01 3.98272336e-01 -4.75551873e-01
-3.99718374e-01 3.73707384e-01 -8.14970315e-01 1.01440978e+00
1.54691625e+00 -6.74600661e-01 -6.82970703e-01 1.92437228e-02
1.02636886e+00 -5.02825797e-01 -6.83161020e-01 4.80020195e-01
4.59027737e-01 -1.09171093e+00 1.06603968e+00 -4.73375529e-01
1.05526614e+00 -4.75936294e-01 -1.92313224e-01 -1.25801396e+00
-1.78136840e-01 -5.95969379e-01 -1.29983366e-01 1.32569480e+00
3.29138815e-01 -3.30632299e-01 7.06412375e-01 5.97307384e-01
-3.87387782e-01 -7.91797578e-01 -4.08440441e-01 -5.85278511e-01
-5.73855303e-02 -5.26020348e-01 8.98101389e-01 9.56782699e-01
-4.93480116e-01 4.54759806e-01 -3.84438276e-01 -7.47847855e-02
4.95396107e-01 2.53221124e-01 1.07931840e+00 -6.80705070e-01
-4.30397511e-01 -4.90876168e-01 -3.20550278e-02 -1.23030996e+00
3.72544080e-01 -9.01329517e-01 2.00466588e-01 -1.77242601e+00
1.64639264e-01 -2.94618338e-01 2.62611434e-02 6.34650230e-01
-4.58410740e-01 4.10257161e-01 2.70266294e-01 3.07252973e-01
-4.44078207e-01 1.00856149e+00 1.68533874e+00 -9.36844200e-02
-3.59270692e-01 -3.46609920e-01 -7.49595404e-01 7.42076039e-01
4.57121789e-01 -1.53295115e-01 -8.00212085e-01 -4.21955347e-01
4.23321389e-02 -8.15515369e-02 6.62100852e-01 -9.36524093e-01
-2.12400138e-01 -3.46464545e-01 3.58316422e-01 -4.70031619e-01
3.86653334e-01 -5.62812984e-01 3.95155072e-01 2.11345032e-01
-4.56420779e-01 -2.12263048e-01 7.32512772e-02 4.98535275e-01
-1.87447295e-01 9.84098166e-02 5.94915807e-01 -1.84134975e-01
-6.10620856e-01 3.60749960e-01 -5.24944887e-02 7.69599974e-02
1.14862430e+00 -2.44514495e-01 1.59853715e-02 -3.94898713e-01
-7.97625184e-01 1.43233746e-01 6.80928171e-01 6.74122274e-01
5.86922288e-01 -1.74069285e+00 -7.20230103e-01 4.24462795e-01
1.09945543e-01 4.19029534e-01 5.13602197e-01 6.94840193e-01
-4.15631473e-01 6.83666617e-02 -2.16704383e-01 -5.10603726e-01
-9.62920189e-01 5.88292122e-01 1.59143135e-01 -2.85199523e-01
-7.66854405e-01 7.39274263e-01 7.28052378e-01 -4.97554183e-01
2.49972157e-02 -4.74839389e-01 -1.23305926e-02 -9.15732607e-02
3.46252173e-01 5.25885336e-02 -3.50953668e-01 -8.10157359e-01
-4.68398184e-02 6.50952220e-01 1.62683904e-01 -4.52104717e-01
1.19461930e+00 -1.58219039e-01 2.17829607e-02 1.12611458e-01
1.11960649e+00 -1.30340815e-01 -1.79105604e+00 -1.68745592e-01
-2.53103405e-01 -3.89667243e-01 -3.06108981e-01 -7.57200420e-01
-9.70518529e-01 7.25162148e-01 1.43278241e-01 -1.26041085e-01
1.40731287e+00 1.27406955e-01 8.58073115e-01 -1.86232954e-01
2.19596162e-01 -9.22594607e-01 7.15864003e-01 2.45708942e-01
1.35088706e+00 -9.49375987e-01 -8.71169344e-02 -5.96489131e-01
-7.38830984e-01 6.11357868e-01 7.63745725e-01 -1.12781249e-01
2.85289466e-01 1.67268552e-02 2.09043529e-02 -9.26089808e-02
-7.06170797e-01 -2.10309789e-01 6.11286342e-01 4.96722817e-01
3.14707398e-01 4.92310673e-02 -1.83103338e-01 7.53475785e-01
-2.58303106e-01 1.09769881e-01 2.01365694e-01 9.26990211e-01
1.16965011e-01 -1.19805217e+00 -1.29257828e-01 1.53824151e-01
-1.45044938e-01 -1.28367230e-01 -5.22600651e-01 6.82069302e-01
4.20379221e-01 6.54143751e-01 8.85153860e-02 -5.04799068e-01
4.75446403e-01 7.58789629e-02 6.02703929e-01 -8.47315609e-01
-2.65129656e-01 1.57241121e-01 -8.86435285e-02 -8.52476478e-01
-6.92389905e-01 -4.76320207e-01 -1.35917687e+00 3.26339453e-02
-1.17074419e-02 -3.01791579e-01 4.41340894e-01 1.01872408e+00
5.10403931e-01 8.30813468e-01 5.51406503e-01 -1.24800611e+00
-3.68002862e-01 -8.96230757e-01 -9.47212428e-02 1.00514710e+00
4.93543334e-02 -8.18576038e-01 -3.51654619e-01 5.50341666e-01] | [11.461967468261719, -0.4728357791900635] |
b3be3669-3479-45ed-9912-d02554543959 | the-construction-of-a-chinese-collocational | null | null | https://aclanthology.org/C16-1307 | https://aclanthology.org/C16-1307.pdf | The Construction of a Chinese Collocational Knowledge Resource and Its Application for Second Language Acquisition | The appropriate use of collocations is a challenge for second language acquisition. However, high quality and easily accessible Chinese collocation resources are not available for both teachers and students. This paper presents the design and construction of a large scale resource of Chinese collocational knowledge, and a web-based application (OCCA, Online Chinese Collocation Assistant) which offers free and convenient collocation search service to end users. We define and classify collocations based on practical language acquisition needs and utilize a syntax based method to extract nine types of collocations. Totally 37 extraction rules are compiled with word, POS and dependency relation features, 1,750,000 collocations are extracted from a corpus for L2 learning and complementary Wikipedia data, and OCCA is implemented based on these extracted collocations. By comparing OCCA with two traditional collocation dictionaries, we find OCCA has higher entry coverage and collocation quantity, and our method achieves quite low error rate at less than 5{\%}. We also discuss how to apply collocational knowledge to grammatical error detection and demonstrate comparable performance to the best results in 2015 NLP-TEA CGED shared task. The preliminary experiment shows that the collocation knowledge is helpful in detecting all the four types of grammatical errors. | ['Kuang-hua Chen', 'Jiayong Chen', 'Renfen Hu'] | 2016-12-01 | the-construction-of-a-chinese-collocational-1 | https://aclanthology.org/C16-1307 | https://aclanthology.org/C16-1307.pdf | coling-2016-12 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-7.11593151e-01 -3.40970278e-01 5.50702959e-02 -2.13302523e-02
-9.38365638e-01 -7.03432202e-01 -3.38556796e-01 6.12958491e-01
-6.74015224e-01 1.22377849e+00 2.20869556e-01 -6.35764301e-01
9.38358754e-02 -6.62627280e-01 -4.51742530e-01 -2.36417070e-01
1.28499061e-01 3.35419834e-01 2.37808794e-01 -6.06189787e-01
4.89517748e-01 -1.46780148e-01 -1.40504575e+00 1.60775900e-01
2.31455398e+00 3.29194963e-01 9.15917814e-01 3.10798109e-01
-3.65618736e-01 2.93299437e-01 -9.23611641e-01 -7.84174979e-01
-5.29932916e-01 -3.87235880e-01 -1.09269655e+00 -5.51020145e-01
2.82886326e-01 1.21938244e-01 3.44670922e-01 1.24804497e+00
7.33663619e-01 -9.59046334e-02 6.79202855e-01 -7.14002430e-01
-1.30035877e+00 1.33232248e+00 6.49604481e-03 1.81311548e-01
7.36742914e-01 -5.28835952e-01 1.18191743e+00 -1.42694139e+00
6.13820255e-01 7.05608487e-01 9.63564634e-01 5.36823869e-01
-4.61550266e-01 -9.76438522e-01 1.37312755e-01 2.76718259e-01
-1.78658342e+00 3.94651387e-03 1.48850128e-01 -4.15878296e-01
1.35262513e+00 -4.68806438e-02 7.69491613e-01 6.54479861e-01
-4.68909964e-02 8.91244113e-01 1.12502277e+00 -1.34126985e+00
-2.35535026e-01 2.43409663e-01 5.71669102e-01 6.58503056e-01
5.26437819e-01 -3.35520416e-01 -6.01464927e-01 1.12003841e-01
1.51930749e-01 -5.81217110e-01 -5.67771971e-01 6.04036570e-01
-1.09855354e+00 8.19327474e-01 -1.73185568e-03 8.98026407e-01
1.79242596e-01 -3.34142536e-01 2.36793861e-01 5.54740787e-01
2.47679368e-01 6.58799052e-01 -9.28603411e-01 -4.17347014e-01
-3.49356771e-01 2.24879578e-01 7.24778771e-01 1.74452651e+00
4.75775242e-01 1.04489394e-01 9.25226733e-02 1.29024446e+00
3.24050605e-01 8.57176781e-01 9.62547600e-01 -1.28290653e-01
6.31509483e-01 5.14898360e-01 -5.63797317e-02 -9.11076009e-01
-2.02008650e-01 -2.42017135e-02 -2.51802534e-01 -4.51784521e-01
5.13774693e-01 -7.03591406e-01 -4.64639753e-01 1.53980529e+00
2.85141826e-01 -2.33379334e-01 2.25985005e-01 3.46645147e-01
1.32351398e+00 6.30160570e-01 4.49984998e-01 -1.56166062e-01
1.39707673e+00 -6.71231508e-01 -1.11239648e+00 2.49032214e-01
1.40653527e+00 -1.37345803e+00 1.34936833e+00 6.08610749e-01
-1.04234409e+00 -4.79131520e-01 -1.07919157e+00 -9.39305499e-02
-9.16616440e-01 6.15093946e-01 6.18094504e-01 1.04712760e+00
-6.93199515e-01 4.33981240e-01 -4.18733418e-01 -2.32675537e-01
-5.83099714e-03 6.05196580e-02 -2.80985385e-01 -1.15229212e-01
-1.64088905e+00 1.18908608e+00 9.30703819e-01 -2.15366602e-01
3.46201137e-02 -6.63866520e-01 -1.14040661e+00 -1.90252855e-01
-7.33799785e-02 2.71785498e-01 1.12634194e+00 -5.87758005e-01
-1.43985379e+00 8.48639071e-01 1.00423574e-01 2.04618379e-01
-3.97178829e-01 -5.44731021e-01 -8.53694975e-01 -3.09205621e-01
4.97195035e-01 2.81229824e-01 -6.95787519e-02 -5.86065173e-01
-1.33468890e+00 -3.86778302e-02 -3.86397451e-01 5.49008250e-01
-5.94408512e-01 7.15655267e-01 -5.14139473e-01 -9.66961026e-01
2.20614672e-01 -7.75728226e-01 3.94436002e-01 -8.01495433e-01
-2.28582874e-01 -1.20957649e+00 1.38987839e-01 -1.02441645e+00
2.12965798e+00 -1.72553575e+00 -1.16528369e-01 2.12140307e-01
-1.76571161e-01 5.61029613e-01 1.09470151e-01 6.14147663e-01
1.47715598e-01 5.20065188e-01 5.54153137e-02 8.29254240e-02
7.52338991e-02 9.42681953e-02 2.07220897e-01 -4.20147777e-02
8.85747969e-02 7.64472783e-01 -1.36031878e+00 -7.49551415e-01
-8.92401338e-02 -9.73019972e-02 -6.93753898e-01 3.11516553e-01
3.98127027e-02 1.97006971e-01 -4.36487705e-01 9.03525352e-01
4.65208620e-01 2.26374343e-01 6.39001906e-01 4.87313271e-01
-5.02189875e-01 1.08431065e+00 -1.31930268e+00 1.68191409e+00
-7.45510340e-01 4.55125839e-01 -4.32909071e-01 -4.42963421e-01
7.54956722e-01 7.76270092e-01 -2.12021589e-01 -4.20751423e-01
2.25305393e-01 9.33924615e-01 3.27447116e-01 -6.68197930e-01
7.94045210e-01 3.79798710e-01 -4.98294920e-01 2.58047491e-01
5.26199460e-01 -3.98212343e-01 4.39061761e-01 2.07540050e-01
6.04187012e-01 2.63559371e-01 7.39447415e-01 -8.52612674e-01
4.33425128e-01 5.16498744e-01 6.36073768e-01 4.61397380e-01
-1.04894219e-02 2.89876580e-01 -1.72017768e-01 1.88134819e-01
-4.44275498e-01 -8.72152746e-01 -5.71449101e-01 1.28580356e+00
-1.16043165e-01 -8.09844255e-01 -8.74602199e-01 -8.12667370e-01
-1.70649186e-01 6.49566174e-01 -8.13613981e-02 3.18766743e-01
-7.84754157e-01 -5.55467248e-01 7.49473512e-01 3.69009852e-01
3.63412052e-01 -1.53307533e+00 3.94313902e-01 6.06388927e-01
-6.39545977e-01 -8.80133271e-01 -9.06577408e-01 -4.62941118e-02
-3.10727566e-01 -1.18525100e+00 -4.22916681e-01 -1.81812215e+00
4.73095387e-01 1.06913768e-01 1.39976442e+00 8.12565863e-01
5.55458404e-02 2.69627534e-02 -1.15177882e+00 -7.92525053e-01
-3.62212211e-01 2.47187153e-01 3.92110109e-01 -9.12493646e-01
1.13455677e+00 -4.07773644e-01 7.29389116e-02 -2.36621752e-01
-1.30258456e-01 -2.33252913e-01 1.97634503e-01 9.50381100e-01
3.08186859e-01 -8.60826150e-02 1.16203225e+00 -1.10554492e+00
7.15215683e-01 -6.83322549e-01 -6.19315565e-01 3.56198698e-01
-7.10156560e-01 -3.31355125e-01 5.86458683e-01 -3.44530374e-01
-1.12873375e+00 -2.96889916e-02 -6.63091421e-01 4.18175638e-01
-1.98096573e-01 1.05056286e+00 -3.56760144e-01 -1.17196992e-01
6.98731720e-01 8.78766179e-02 -7.16465473e-01 -7.08085716e-01
3.98317665e-01 1.09546828e+00 3.90768945e-01 -1.23312891e+00
1.75261036e-01 -8.56073797e-01 -1.00200129e+00 -8.99135828e-01
-1.09307289e+00 -6.31643355e-01 -9.74216104e-01 -5.37468456e-02
8.71822000e-01 -1.22574556e+00 -6.64822638e-01 7.57581949e-01
-1.34047520e+00 -1.18732944e-01 3.09690386e-01 9.72335696e-01
2.17491239e-02 3.97988707e-01 -8.77927840e-01 -2.95016915e-01
-4.15991306e-01 -9.10301387e-01 3.01755637e-01 5.22335172e-01
-1.74520567e-01 -1.12005532e+00 1.43713504e-01 2.01225474e-01
8.23042393e-02 -2.27733433e-01 1.08622324e+00 -7.98727691e-01
-1.24312587e-01 3.56315613e-01 1.01547122e-01 4.77058291e-01
2.63543993e-01 1.18723474e-01 -4.31817919e-01 -6.77836090e-02
-6.92467332e-01 -7.93507218e-01 9.17937085e-02 4.68477421e-02
9.75078046e-01 -2.59086877e-01 -1.27591431e-01 1.86505526e-01
1.57905304e+00 3.20315540e-01 4.37808216e-01 7.10409731e-02
8.70724559e-01 3.45080853e-01 8.39180052e-01 5.19970357e-02
9.55793023e-01 4.34482813e-01 -3.38164747e-01 6.37447476e-01
-3.27071637e-01 -5.04240990e-01 5.04448593e-01 2.41794229e+00
-4.19559442e-02 -1.57382369e-01 -1.23508513e+00 1.33580101e+00
-1.42010021e+00 -3.78955662e-01 -7.43115604e-01 2.07063818e+00
2.11602592e+00 -2.05599651e-01 -2.53009588e-01 -8.34021792e-02
8.32153678e-01 -9.30283546e-01 4.13433969e-01 -4.33085918e-01
-2.11808786e-01 1.02494442e+00 1.05671154e-03 6.20657146e-01
-8.12318861e-01 1.78350830e+00 5.96238041e+00 1.46365321e+00
-6.67447448e-01 3.38909060e-01 -2.27345318e-01 7.96315908e-01
-3.45497191e-01 1.95506215e-02 -1.50919616e+00 7.26854444e-01
7.70493150e-01 -1.72886461e-01 5.46769388e-02 7.95876205e-01
-5.25688887e-01 -1.68022722e-01 -7.65021622e-01 8.77778709e-01
1.38522517e-02 -1.29564464e+00 -1.05055429e-01 -5.07893741e-01
1.18247831e+00 1.49966300e-01 -5.60858965e-01 7.41828561e-01
1.03661120e+00 -9.47804987e-01 6.56824231e-01 5.48947640e-02
1.04061139e+00 -9.00792003e-01 6.54357195e-01 2.54326820e-01
-1.36895680e+00 4.51453030e-01 -4.22169596e-01 -2.99693316e-01
-2.85954982e-01 2.53000766e-01 -5.65490365e-01 5.62742352e-01
9.39099669e-01 5.53810060e-01 -6.63766801e-01 1.15724790e+00
-9.20769751e-01 1.26642096e+00 -3.94285053e-01 -8.49292040e-01
1.29469290e-01 -3.01680803e-01 2.22595006e-01 1.66364646e+00
6.22487247e-01 5.83426297e-01 2.65952229e-01 4.86425608e-01
-1.05344625e-02 1.02191782e+00 -3.48878622e-01 9.17762965e-02
1.55646205e+00 9.86926138e-01 -2.30622038e-01 -2.60170460e-01
-7.50619948e-01 8.11858356e-01 1.02602983e+00 1.96031034e-02
-3.49865705e-01 -1.20756412e+00 5.60444236e-01 -4.69924837e-01
2.53777593e-01 -4.08366472e-01 -2.84407169e-01 -1.39944530e+00
-8.23036060e-02 -1.17828333e+00 4.70317274e-01 -5.72997332e-01
-1.61799014e+00 4.23146516e-01 -2.36634865e-01 -1.22865939e+00
-1.70157343e-01 -8.64228964e-01 -7.90561616e-01 1.30203700e+00
-1.41324937e+00 -1.03641319e+00 1.10870138e-01 7.97269344e-01
5.07600784e-01 -3.22564572e-01 1.18478513e+00 7.02317417e-01
-6.08523190e-01 1.24924719e+00 1.61198050e-01 6.48895442e-01
8.46594751e-01 -1.76108658e+00 -4.71387710e-03 9.58602905e-01
2.30991006e-01 7.71697938e-01 1.01162255e-01 -9.54045355e-01
-9.40191567e-01 -9.72189784e-01 2.00348115e+00 -6.77500665e-01
8.65441203e-01 -1.95083320e-01 -1.07690048e+00 7.06339777e-01
5.57781756e-01 -2.96045423e-01 1.01827133e+00 5.31449318e-01
1.81044936e-01 4.60107207e-01 -7.26867676e-01 6.00208044e-01
9.83170152e-01 -3.93651783e-01 -1.05601811e+00 4.51566488e-01
9.54062879e-01 -7.41678059e-01 -1.26930642e+00 6.63718358e-02
5.84844984e-02 -2.97939569e-01 2.87572503e-01 -6.25497282e-01
3.61408591e-01 -3.88756484e-01 -1.34190740e-02 -1.86876285e+00
-2.32125655e-01 -5.44490218e-01 3.49766463e-01 1.68970203e+00
9.71674502e-01 -4.34050977e-01 1.27450898e-01 8.28417242e-02
-8.77644897e-01 -6.84226155e-01 -8.30299437e-01 -6.69231117e-01
8.82974625e-01 -4.57329720e-01 5.03283024e-01 1.69587231e+00
9.36525762e-01 3.78932208e-01 1.82658821e-01 4.38424975e-01
6.73874393e-02 3.24174925e-03 2.32834667e-01 -9.78479028e-01
2.20688686e-01 -1.29698887e-01 1.09799363e-01 -9.18721974e-01
1.64640471e-01 -1.23359251e+00 2.32190207e-01 -1.13752675e+00
-3.27796757e-01 -1.41841149e+00 7.63921589e-02 7.51938164e-01
-8.62060487e-01 1.06176555e-01 1.41642466e-02 -1.93805516e-01
-3.30630392e-01 4.89338577e-01 1.13008082e+00 4.14628714e-01
-3.66825551e-01 -2.74033964e-01 -6.63996458e-01 8.00660491e-01
7.74580598e-01 -6.83198988e-01 1.05626069e-01 -7.11828887e-01
5.51081836e-01 -3.03942710e-01 -6.62855148e-01 -7.86039352e-01
3.26546371e-01 -3.49918872e-01 1.31356820e-01 -6.50323808e-01
-5.00463068e-01 -4.60538983e-01 -4.98977780e-01 1.84054717e-01
2.11554438e-01 2.89264679e-01 4.17563170e-01 -2.73997873e-01
-3.30355227e-01 -9.33729053e-01 5.27179956e-01 -1.25212044e-01
-9.38541770e-01 3.86904813e-02 -7.19025135e-01 7.39162207e-01
7.09607244e-01 7.83110783e-02 -1.99563503e-01 1.59014881e-01
-5.75164497e-01 5.42162418e-01 -1.34352207e-01 5.02057731e-01
4.87283796e-01 -1.71115065e+00 -1.10382688e+00 1.93130359e-01
4.28938210e-01 -4.44867536e-02 -1.07854061e-01 2.34117582e-01
-7.95943379e-01 4.41962481e-01 -1.71701744e-01 -1.70582101e-01
-1.11767840e+00 2.94276769e-03 1.50568753e-01 -1.67989179e-01
-2.66935706e-01 1.18276787e+00 -4.64862585e-01 -1.09128475e+00
2.03088030e-01 -4.83500689e-01 -8.35923672e-01 -3.55349444e-02
5.48014581e-01 2.13229164e-01 4.66952384e-01 -6.57377601e-01
-1.51009619e-01 5.41236401e-01 -2.15417773e-01 1.65997297e-01
8.56900632e-01 -3.97630364e-01 -3.07618737e-01 2.95007467e-01
6.79480970e-01 1.00137115e+00 -2.39831373e-01 -4.01212752e-01
3.77668709e-01 -2.17990711e-01 -3.33386183e-01 -1.08593571e+00
-6.56580865e-01 5.61386704e-01 1.72555327e-01 -8.79751518e-03
8.35117757e-01 -5.69452494e-02 9.84015346e-01 5.93224227e-01
6.07711375e-01 -1.63445663e+00 -2.62696028e-01 1.37273324e+00
7.49586940e-01 -1.34288490e+00 -3.15104485e-01 -9.30393755e-01
-6.35389090e-01 9.43616390e-01 1.23374736e+00 -1.02479406e-01
1.17808890e+00 1.65054295e-02 1.59547165e-01 -1.16190203e-01
-3.60582948e-01 -4.88801509e-01 4.61880356e-01 7.69709289e-01
1.13170993e+00 5.26733041e-01 -1.32570946e+00 1.29640877e+00
-8.41243684e-01 -5.12243629e-01 6.61286831e-01 8.38881016e-01
-6.28338337e-01 -1.49644816e+00 -7.75675401e-02 2.94182569e-01
-8.49638522e-01 -1.05285501e+00 1.54208289e-02 9.75695848e-01
6.46054626e-01 1.17960477e+00 -1.43457025e-01 -1.53769627e-01
2.44523212e-01 2.49117538e-01 5.12074292e-01 -1.38329363e+00
-9.74908471e-01 -4.16805476e-01 3.74025524e-01 -5.38146123e-02
-2.22033396e-01 -6.02631986e-01 -1.66968274e+00 -2.76081651e-01
-1.18124986e+00 7.13284373e-01 4.64206725e-01 1.11346388e+00
-3.12101183e-04 6.43399283e-02 4.45347160e-01 -7.81765133e-02
-4.74620536e-02 -1.43452156e+00 -7.03371286e-01 1.83843210e-01
-2.81981051e-01 -5.01397610e-01 -7.44125172e-02 1.38655767e-01] | [11.017912864685059, 10.765830993652344] |
62507329-a2b9-4901-993e-0a1fd717aefa | enhanced-multimodal-representation-learning-1 | 2306.07646 | null | https://arxiv.org/abs/2306.07646v1 | https://arxiv.org/pdf/2306.07646v1.pdf | Enhanced Multimodal Representation Learning with Cross-modal KD | This paper explores the tasks of leveraging auxiliary modalities which are only available at training to enhance multimodal representation learning through cross-modal Knowledge Distillation (KD). The widely adopted mutual information maximization-based objective leads to a short-cut solution of the weak teacher, i.e., achieving the maximum mutual information by simply making the teacher model as weak as the student model. To prevent such a weak solution, we introduce an additional objective term, i.e., the mutual information between the teacher and the auxiliary modality model. Besides, to narrow down the information gap between the student and teacher, we further propose to minimize the conditional entropy of the teacher given the student. Novel training schemes based on contrastive learning and adversarial learning are designed to optimize the mutual information and the conditional entropy, respectively. Experimental results on three popular multimodal benchmark datasets have shown that the proposed method outperforms a range of state-of-the-art approaches for video recognition, video retrieval and emotion classification. | ['Ya zhang', 'Yu Wang', 'Linyu Xing', 'Mengxi Chen'] | 2023-06-13 | enhanced-multimodal-representation-learning | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_Enhanced_Multimodal_Representation_Learning_With_Cross-Modal_KD_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_Enhanced_Multimodal_Representation_Learning_With_Cross-Modal_KD_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-recognition', 'video-retrieval', 'emotion-classification', 'emotion-classification'] | ['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 4.50376898e-01 3.57288122e-01 -2.20174178e-01 -2.68517643e-01
-1.00956714e+00 -3.40607613e-01 5.47210872e-01 1.01047106e-01
-6.11755729e-01 7.05182910e-01 -1.69581342e-02 -1.61661580e-01
-2.13252261e-01 -4.35481966e-01 -8.05344820e-01 -1.21707547e+00
2.27031842e-01 -3.35373916e-02 -2.67863065e-01 -5.39667793e-02
1.85981505e-02 1.11837395e-01 -1.53789675e+00 1.05711333e-01
1.35689902e+00 1.15353358e+00 6.05113767e-02 2.86479890e-01
-2.41755117e-02 9.39229012e-01 -4.88537133e-01 -6.66754127e-01
-1.46052018e-01 -5.60330808e-01 -7.86522925e-01 1.07041806e-01
2.00033337e-01 -1.05092973e-01 -6.12775624e-01 1.14634490e+00
6.59542024e-01 5.56894422e-01 8.11494350e-01 -1.42869973e+00
-5.99039257e-01 6.39530897e-01 -7.36771524e-01 -2.76924968e-02
4.67654407e-01 -1.96887434e-01 8.26494992e-01 -7.38047302e-01
2.38638610e-01 1.08235204e+00 4.17775393e-01 7.41738558e-01
-1.16018546e+00 -8.17126393e-01 1.84406564e-01 3.92775208e-01
-1.46601164e+00 -3.87311786e-01 1.05690324e+00 -2.06776485e-01
3.11151862e-01 2.82236576e-01 2.63144016e-01 1.10596251e+00
-2.86196351e-01 1.22733176e+00 1.18159080e+00 -5.79809546e-01
2.82914620e-02 4.83861327e-01 -1.18807420e-01 8.44546616e-01
-4.35772687e-01 -5.36306165e-02 -5.77441633e-01 -1.79102868e-01
2.71679997e-01 -1.43876523e-01 -5.81264794e-01 -5.79964399e-01
-9.98168111e-01 7.91418076e-01 2.22448871e-01 1.63205504e-01
-1.94239303e-01 -3.17849249e-01 3.39294046e-01 4.03680742e-01
2.63044924e-01 1.34752870e-01 -3.70451748e-01 -1.40609473e-01
-7.58614361e-01 -1.36907384e-01 6.83388472e-01 6.15934014e-01
9.28662479e-01 1.88962482e-02 -2.25318223e-01 9.19065237e-01
5.44642150e-01 4.67601120e-01 5.61254561e-01 -6.75721884e-01
6.15950346e-01 6.43777788e-01 -1.31502092e-01 -9.25206482e-01
1.50255561e-01 -2.40100235e-01 -7.90820658e-01 -1.87050357e-01
4.00929153e-01 -3.76409888e-01 -9.84293997e-01 2.16447854e+00
5.57307541e-01 6.17222905e-01 4.31607664e-01 9.38378930e-01
1.08348155e+00 5.64882934e-01 2.06796631e-01 -3.72288108e-01
8.25691164e-01 -1.02925229e+00 -8.18115830e-01 4.79207980e-03
6.09569132e-01 -7.09221363e-01 7.64175415e-01 1.94535673e-01
-1.17896235e+00 -4.08313215e-01 -9.79991615e-01 2.29996756e-01
-3.41684967e-01 -4.00776938e-02 4.60704714e-01 5.47829628e-01
-6.91599548e-01 3.59809071e-01 -8.18041146e-01 6.27379715e-02
3.18304330e-01 5.01336336e-01 -6.22178137e-01 -1.67898491e-01
-1.36462104e+00 9.55643296e-01 6.52495801e-01 6.20349869e-02
-1.04344976e+00 -6.94017828e-01 -1.02557468e+00 9.33512375e-02
6.14090025e-01 -4.77186173e-01 8.31496060e-01 -1.31652701e+00
-1.72977901e+00 7.23866343e-01 1.01007111e-01 -1.20973140e-01
3.72042596e-01 -1.92811117e-01 -3.23587209e-01 4.29383546e-01
-3.07448715e-01 7.76882410e-01 1.00945175e+00 -1.48842704e+00
-4.15734470e-01 -3.00576240e-01 1.72115475e-01 6.30154073e-01
-8.71638656e-01 -3.61168861e-01 -6.11074507e-01 -5.82095087e-01
-6.27349969e-03 -8.88686895e-01 -8.11963249e-03 -4.44028109e-01
-3.51686478e-01 -3.86876434e-01 1.03953981e+00 -7.62686074e-01
1.18205988e+00 -2.21573138e+00 6.69098735e-01 3.39048862e-01
-6.88804081e-03 2.97703236e-01 -5.16541302e-01 1.49856538e-01
-2.52114892e-01 -1.00751191e-01 -2.83694685e-01 -4.02243912e-01
5.57615124e-02 3.67003888e-01 -1.94324777e-01 4.24438715e-01
2.84785479e-01 7.36281514e-01 -8.79742086e-01 -5.91347098e-01
2.84053296e-01 9.11377728e-01 -6.06090188e-01 6.64641738e-01
1.17486402e-01 6.14505231e-01 -3.82996678e-01 5.13174474e-01
7.66059995e-01 -3.98998223e-02 2.30437398e-01 -3.40022236e-01
2.56791234e-01 -2.08443310e-02 -1.15428221e+00 1.92975926e+00
-4.19744909e-01 2.45020390e-01 2.17710435e-01 -1.62714803e+00
7.20340848e-01 4.99119997e-01 6.53417468e-01 -4.75900531e-01
2.02300265e-01 -9.34276655e-02 -1.51016667e-01 -6.24145508e-01
2.58860230e-01 -3.09282929e-01 -7.88617134e-02 2.18222350e-01
5.24121225e-01 1.59004573e-02 -7.36626387e-02 3.20513129e-01
6.15588188e-01 1.50418878e-01 -1.05252929e-01 2.78277602e-02
8.60823750e-01 -7.32518733e-01 5.25396407e-01 5.28511643e-01
-2.53329307e-01 2.86024064e-01 5.15004575e-01 3.73911917e-01
-5.12749791e-01 -1.03516638e+00 9.06753913e-02 1.23974597e+00
3.42932254e-01 -2.70266473e-01 -7.74483144e-01 -1.10729098e+00
-1.20374955e-01 5.44387043e-01 -8.00227404e-01 -6.40924096e-01
-9.95768085e-02 -5.71872234e-01 6.76138341e-01 2.78023154e-01
6.12647653e-01 -6.19258761e-01 -2.21551120e-01 -2.88573772e-01
-6.10053301e-01 -1.02154362e+00 -4.42236066e-01 2.27768645e-01
-7.09666669e-01 -9.22968030e-01 -8.65963161e-01 -7.85250008e-01
7.38143086e-01 -1.77218113e-02 7.80437231e-01 -2.76064995e-04
5.87832257e-02 9.43907619e-01 -3.54005724e-01 4.23926022e-03
-1.18699774e-01 -1.27431422e-01 7.45647820e-03 2.83494383e-01
2.63115019e-01 -4.55824047e-01 -4.51716781e-01 2.64788210e-01
-1.24944937e+00 -7.85386786e-02 6.11680329e-01 1.26687884e+00
5.01497865e-01 1.99835703e-01 6.65014923e-01 -3.77412051e-01
4.85015273e-01 -5.64288855e-01 -1.26300797e-01 6.04434669e-01
-5.78146338e-01 3.07498127e-01 3.34402740e-01 -9.53472376e-01
-1.50998938e+00 -6.15572482e-02 4.96663190e-02 -7.24973381e-01
-8.89241397e-02 8.35974097e-01 -5.04284620e-01 -2.31043503e-01
1.48602100e-02 4.42353666e-01 4.50759195e-03 -2.99618989e-01
4.72482622e-01 5.43398917e-01 5.73507905e-01 -7.64281929e-01
6.63272560e-01 2.50970095e-01 -7.82578290e-02 -8.00245345e-01
-9.20270503e-01 -4.16047275e-01 -4.78365928e-01 -3.01497161e-01
7.49802470e-01 -9.39616263e-01 -8.40747178e-01 5.16314805e-01
-8.19265008e-01 -2.54631303e-02 -7.62523413e-02 7.70662248e-01
-5.67094922e-01 5.77192366e-01 -4.15910780e-01 -9.51143682e-01
-1.19758300e-01 -1.09730566e+00 7.72415519e-01 6.13428712e-01
3.04444939e-01 -1.30824816e+00 3.38788331e-02 6.80755138e-01
1.60440207e-01 2.18241602e-01 8.85243297e-01 -8.47896278e-01
-2.23067001e-01 -4.84087132e-02 -1.00350402e-01 6.99011207e-01
-3.68356630e-02 -1.61346361e-01 -1.03062904e+00 -3.85271758e-01
-2.33157147e-02 -9.78640437e-01 1.18055391e+00 -1.70909911e-02
1.09127569e+00 -4.03696686e-01 -2.28083223e-01 5.20051718e-01
1.25425684e+00 1.25248358e-01 7.32979238e-01 1.39083773e-01
8.50847065e-01 7.30978429e-01 6.28623366e-01 4.37883854e-01
4.53308374e-01 4.22763348e-01 4.96230572e-01 6.93580285e-02
2.66693443e-01 -3.81155908e-01 4.98726308e-01 1.03094435e+00
-1.13158738e-02 -2.08396211e-01 -5.37143171e-01 3.28703701e-01
-1.82407570e+00 -8.95794034e-01 5.23536861e-01 2.38673949e+00
1.12398589e+00 -2.98957676e-01 -2.11142048e-01 7.04628676e-02
5.88207364e-01 1.68100312e-01 -4.17143285e-01 -9.02184471e-02
-1.64283589e-01 4.73878980e-02 2.34464724e-02 5.06752670e-01
-1.35903704e+00 7.19289243e-01 5.31401587e+00 1.08761239e+00
-1.17733884e+00 5.24474569e-02 4.26104844e-01 1.17365845e-01
-4.04397070e-01 -1.89233318e-01 -4.15769637e-01 4.70396280e-01
7.66967177e-01 -2.36803889e-02 5.11317372e-01 5.55137098e-01
-2.10507140e-01 -2.19763488e-01 -9.60175335e-01 1.07015288e+00
4.22163546e-01 -5.64330459e-01 6.96447045e-02 -3.69568355e-02
9.50370431e-01 -3.18646491e-01 4.01048958e-01 7.02031672e-01
5.14856959e-03 -8.89149845e-01 3.22827101e-01 7.81106532e-01
4.75025386e-01 -1.14443314e+00 7.98275948e-01 4.50351477e-01
-9.38477039e-01 -4.82124975e-03 -9.59753171e-02 4.39093560e-01
-2.45506287e-01 3.39826912e-01 -2.39642799e-01 7.35518575e-01
4.49613512e-01 5.23304284e-01 -4.84959394e-01 7.52801716e-01
-2.82702148e-01 4.86189574e-01 -2.77550906e-01 3.13859940e-01
1.76715136e-01 -3.39262843e-01 6.93950534e-01 1.15080762e+00
2.14222595e-01 1.92020759e-01 2.88354635e-01 4.83949870e-01
-3.10525388e-01 1.58778012e-01 -5.94251215e-01 -3.21509480e-01
2.69988507e-01 1.21939242e+00 -1.90758377e-01 -3.10005784e-01
-4.98073459e-01 1.22235465e+00 5.11546135e-01 6.29997432e-01
-9.12448764e-01 -4.06363487e-01 4.64385539e-01 -5.90138018e-01
4.11001265e-01 9.55438018e-02 2.51572907e-01 -1.37341535e+00
1.99007033e-03 -9.29923236e-01 7.15157747e-01 -4.61089373e-01
-1.24850821e+00 2.65910089e-01 -1.26016838e-02 -1.10676110e+00
-2.77111679e-01 -3.40278625e-01 -5.67708194e-01 8.10910583e-01
-1.79241848e+00 -1.07632959e+00 -5.93254305e-02 1.00112903e+00
-3.86742800e-02 -2.03965902e-01 7.92581677e-01 5.93506932e-01
-6.99470758e-01 1.11111414e+00 2.69296438e-01 1.24901168e-01
8.52059007e-01 -1.07797825e+00 -1.02969921e+00 4.24416751e-01
-9.94542241e-02 5.35994411e-01 6.79844975e-01 -2.86308706e-01
-1.62938344e+00 -7.22472787e-01 4.88274276e-01 -1.33990929e-01
6.45101547e-01 9.42708403e-02 -1.08871174e+00 4.22467351e-01
4.66529667e-01 -1.54856041e-01 1.03312981e+00 -2.16131844e-02
-4.56796080e-01 -1.12548478e-01 -1.16764319e+00 4.50120866e-01
2.16125831e-01 -8.50084126e-01 -6.51053369e-01 9.76892002e-03
4.57666725e-01 -2.39349023e-01 -1.13417709e+00 8.34227443e-01
6.52407944e-01 -7.11173952e-01 9.36402857e-01 -8.52111816e-01
5.73350191e-01 -4.82658744e-02 -2.33778954e-01 -1.56145251e+00
2.64371246e-01 -4.24468160e-01 -5.80797374e-01 1.55635691e+00
3.20860654e-01 -2.55859077e-01 6.60436153e-01 7.92981803e-01
8.83988887e-02 -9.05465662e-01 -1.01953852e+00 -5.51121831e-01
1.41073480e-01 -2.06091955e-01 1.15746625e-01 1.40609562e+00
3.88450205e-01 3.71252865e-01 -5.59475482e-01 2.09839568e-01
6.85185850e-01 -1.65143814e-02 5.80363154e-01 -8.45839143e-01
-2.53288269e-01 -3.28196466e-01 -2.27534503e-01 -1.07151258e+00
6.98061943e-01 -8.09308410e-01 1.09937668e-01 -1.05378103e+00
5.31470358e-01 -1.57804519e-01 -8.15824568e-01 3.72068137e-01
-5.73950231e-01 1.53006971e-01 2.63184816e-01 -7.72397071e-02
-8.18707287e-01 1.10043824e+00 1.46712208e+00 -3.93463731e-01
-2.34088048e-01 -1.87557675e-02 -6.22274399e-01 7.80951202e-01
5.50638914e-01 -3.91007036e-01 -4.93822217e-01 -2.10888073e-01
3.64739746e-02 2.62953818e-01 2.31238231e-01 -5.33202291e-01
2.92379260e-01 -1.74072102e-01 2.73873866e-01 -4.14819360e-01
6.42986238e-01 -1.02051544e+00 -3.61260623e-01 2.04134390e-01
-6.40964925e-01 -4.43781585e-01 2.30490208e-01 7.44130075e-01
-4.93357718e-01 -3.75969380e-01 8.16822886e-01 1.63587257e-01
-5.00689387e-01 2.02321753e-01 -2.17370749e-01 3.58813070e-02
1.02302730e+00 3.03185254e-01 -2.93232709e-01 -6.33264601e-01
-8.69345903e-01 5.49948215e-01 7.78574422e-02 4.15900022e-01
8.09551358e-01 -1.58748186e+00 -5.90047181e-01 7.43203685e-02
1.27730697e-01 -3.42925847e-01 5.29671609e-01 1.16918194e+00
3.79040569e-01 3.36125851e-01 -1.53019100e-01 -4.37943995e-01
-1.55489731e+00 6.50679231e-01 2.78314531e-01 -5.01604319e-01
6.47770539e-02 8.20931792e-01 2.67785132e-01 -7.65383184e-01
6.54390693e-01 3.23335648e-01 -4.18827534e-01 3.36689621e-01
4.49760288e-01 3.99407953e-01 -2.48847008e-01 -8.91288936e-01
-3.80127519e-01 2.10502967e-01 -8.57499093e-02 -2.17070922e-01
1.01207995e+00 -4.43527073e-01 -8.56060386e-02 3.44783187e-01
1.53716362e+00 -1.95442811e-01 -1.19282341e+00 -4.36326444e-01
-2.67072946e-01 -3.98292363e-01 2.49095216e-01 -8.03031325e-01
-1.21216977e+00 1.06488359e+00 8.24489176e-01 7.95755535e-02
1.35082996e+00 4.36089747e-02 6.08167708e-01 4.63315666e-01
-1.48693994e-01 -1.23283088e+00 3.05662841e-01 5.45501649e-01
5.28466165e-01 -1.53199399e+00 -1.15562133e-01 -2.07621604e-01
-6.83214128e-01 9.28321600e-01 8.41335475e-01 2.56411254e-01
6.29825890e-01 -2.48207208e-02 4.43304814e-02 1.38228416e-01
-6.74181283e-01 -2.03895822e-01 7.93182850e-01 3.59671026e-01
4.92580950e-01 -1.69619936e-02 -2.37046272e-01 6.22842789e-01
4.06906158e-01 -1.59895554e-01 4.33922186e-02 9.83853996e-01
-1.84091821e-01 -9.87074792e-01 -4.12767768e-01 2.17185393e-01
-7.19589293e-01 -9.46361944e-02 -3.81980449e-01 5.56185901e-01
5.58978319e-02 1.06420982e+00 -2.11889282e-01 -5.92997134e-01
-4.43447083e-02 3.19176227e-01 6.80950403e-01 -1.09336913e-01
-2.30867207e-01 2.70591676e-02 -2.46662930e-01 -2.50948370e-01
-9.26358283e-01 -3.35019946e-01 -8.45059812e-01 -4.36785258e-02
-8.28393579e-01 5.34715414e-01 5.21695197e-01 1.25221717e+00
-7.58590642e-03 4.55800444e-01 8.62234950e-01 -7.88015008e-01
-7.02531338e-01 -9.61333692e-01 -5.01756310e-01 4.94344831e-01
3.62585545e-01 -8.05697978e-01 -5.91602087e-01 1.04776829e-01] | [13.088095664978027, 5.067083835601807] |
330b0ed4-7b2b-48d3-8dc3-077005fe231d | knowlege-graph-embedding-by-flexible | 1505.05253 | null | http://arxiv.org/abs/1505.05253v2 | http://arxiv.org/pdf/1505.05253v2.pdf | Knowlege Graph Embedding by Flexible Translation | Knowledge graph embedding refers to projecting entities and relations in
knowledge graph into continuous vector spaces. State-of-the-art methods, such
as TransE, TransH, and TransR build embeddings by treating relation as
translation from head entity to tail entity. However, previous models can not
deal with reflexive/one-to-many/many-to-one/many-to-many relations properly, or
lack of scalability and efficiency. Thus, we propose a novel method, flexible
translation, named TransF, to address the above issues. TransF regards relation
as translation between head entity vector and tail entity vector with flexible
magnitude. To evaluate the proposed model, we conduct link prediction and
triple classification on benchmark datasets. Experimental results show that our
method remarkably improve the performance compared with several
state-of-the-art baselines. | ['Minlie Huang', 'Mantong Zhou', 'Jun Feng', 'Xiaoyan Zhu', 'Yu Hao'] | 2015-05-20 | null | null | null | null | ['triple-classification'] | ['graphs'] | [-4.04905677e-01 1.47151709e-01 -6.02949679e-01 -1.88798830e-01
-1.65877983e-01 -5.33178747e-01 7.96258628e-01 1.15590796e-01
-2.32805565e-01 7.12154925e-01 4.30629462e-01 -4.75614548e-01
-3.57209295e-01 -1.07940674e+00 -7.05203712e-01 -2.28291377e-01
-3.01475793e-01 7.11185932e-01 3.09442759e-01 -4.94283170e-01
-3.62175673e-01 5.59204631e-02 -8.72259438e-01 5.32201864e-02
7.07896829e-01 7.08095431e-01 -5.73824048e-01 1.79514706e-01
-3.77789080e-01 6.96717381e-01 -8.60292241e-02 -1.15184760e+00
1.81870759e-01 -8.43051076e-02 -9.24977183e-01 -4.26294148e-01
1.13411039e-01 -9.42933857e-02 -8.71980011e-01 9.32260215e-01
3.36527556e-01 -4.58563566e-02 7.36666441e-01 -1.55481160e+00
-1.68658030e+00 8.81919682e-01 -5.90603054e-01 6.40765578e-02
5.85321307e-01 -4.58897740e-01 1.64674485e+00 -1.46424818e+00
8.13754380e-01 1.24090445e+00 7.37575650e-01 5.71142212e-02
-1.22446573e+00 -5.36910474e-01 1.37414530e-01 6.00761950e-01
-1.75215626e+00 -1.15060456e-01 6.44991517e-01 -4.69948918e-01
1.25082958e+00 3.39245349e-01 6.18666828e-01 8.50072563e-01
-7.77377421e-03 5.58097363e-01 6.86870694e-01 -3.21982443e-01
-3.80593777e-01 2.73694307e-01 5.31964779e-01 7.78518498e-01
6.74443245e-01 -9.53743830e-02 -4.56252038e-01 -2.54393160e-01
5.27334094e-01 -6.11656206e-03 -4.80929524e-01 -6.42569005e-01
-1.52625525e+00 8.36093724e-01 6.49119854e-01 2.51141876e-01
-3.71585071e-01 1.54082656e-01 4.99981195e-01 4.92582411e-01
3.98553759e-01 3.56376439e-01 -6.56135857e-01 5.61227500e-02
-1.66216731e-01 8.66978988e-02 9.91140068e-01 1.29966164e+00
7.80872345e-01 -2.77905762e-01 -1.38089642e-01 6.90183997e-01
5.38899183e-01 -4.20066854e-03 3.65654945e-01 -2.85798937e-01
7.79357016e-01 1.13407385e+00 -1.13949217e-01 -1.19980180e+00
-3.69238377e-01 -3.01477849e-01 -7.87600756e-01 -5.68828881e-01
-2.16464996e-01 -3.93470004e-02 -5.97016275e-01 1.50286913e+00
6.19004250e-01 4.88842010e-01 3.50301981e-01 7.95510709e-01
1.26359069e+00 6.97256505e-01 9.82417352e-03 -2.84862161e-01
1.51569951e+00 -1.20929992e+00 -9.82486248e-01 7.96164125e-02
1.05265641e+00 -7.64913082e-01 1.00286722e+00 -3.51601988e-01
-7.87811995e-01 -3.10099512e-01 -8.63515973e-01 -2.19704270e-01
-9.41962302e-01 9.55840722e-02 1.03766966e+00 4.54469532e-01
-9.35541451e-01 4.00059283e-01 -6.44975126e-01 -5.70845187e-01
1.14082560e-01 3.84749770e-01 -1.05830264e+00 -7.68401921e-02
-1.77779627e+00 1.10164165e+00 7.86011755e-01 1.03284746e-01
-1.17913984e-01 -8.60659659e-01 -1.20514798e+00 1.81998089e-01
5.21531582e-01 -9.88133192e-01 5.68817437e-01 -2.82196142e-02
-1.15993428e+00 6.40351772e-01 -1.13595828e-01 -2.95363843e-01
7.27498019e-03 -3.58410805e-01 -9.86411870e-01 -4.43821669e-01
4.71421257e-02 1.29920930e-01 2.23915800e-01 -1.32860529e+00
-3.63055319e-01 -2.67511934e-01 2.21385777e-01 2.97858924e-01
-7.46675968e-01 1.35295163e-03 -7.95029044e-01 -4.14279133e-01
2.07872570e-01 -7.72277653e-01 2.05826119e-01 -2.38589019e-01
-8.21882904e-01 -6.67647600e-01 1.03436685e+00 -4.40597355e-01
1.79769862e+00 -1.87502158e+00 4.75425720e-01 2.66783893e-01
4.59094465e-01 4.78949130e-01 -8.89720768e-02 1.02010667e+00
-5.11094570e-01 3.29549640e-01 1.91983730e-01 -1.01229966e-01
2.33730584e-01 4.78286147e-01 -1.98586941e-01 4.21131581e-01
1.55886397e-01 1.41034722e+00 -9.54414904e-01 -7.03326106e-01
-2.56373715e-02 5.90488970e-01 -3.30248386e-01 1.87239334e-01
2.30131328e-01 -2.40136713e-01 -6.54452920e-01 8.06401908e-01
4.99815404e-01 -4.76857513e-01 5.10933757e-01 -6.86857343e-01
2.23337084e-01 3.92838657e-01 -1.11965263e+00 1.37022793e+00
-2.02431545e-01 1.89211398e-01 -6.02601528e-01 -9.67634916e-01
8.91735256e-01 4.50901717e-01 4.42468375e-01 -3.58005255e-01
1.97936855e-02 1.60367444e-01 7.74406046e-02 -6.30109668e-01
7.42958009e-01 6.60858452e-02 4.97398060e-03 1.10762276e-01
2.33989462e-01 5.08224487e-01 1.79416001e-01 5.49187541e-01
1.14804292e+00 2.38796011e-01 3.86170447e-01 1.13486879e-01
4.74359870e-01 -3.48392725e-01 7.96124935e-01 1.02079801e-01
-1.02341928e-01 9.36079100e-02 8.81508172e-01 -2.73771226e-01
-8.00336361e-01 -1.20216346e+00 6.03882596e-02 1.03946054e+00
2.72416383e-01 -1.07195294e+00 -1.98376358e-01 -9.83622074e-01
3.97555530e-01 6.69359863e-01 -7.88786829e-01 -3.83717179e-01
-4.92958307e-01 -6.90663576e-01 4.98718441e-01 6.89048707e-01
2.43733138e-01 -6.78690255e-01 5.00035286e-01 2.91108996e-01
-9.51873362e-02 -1.38908958e+00 -6.88517272e-01 -1.42953828e-01
-4.44968551e-01 -1.16387677e+00 -2.91576028e-01 -1.04899669e+00
5.50544560e-01 2.34030217e-01 1.15703726e+00 -8.27452242e-02
5.76275028e-02 2.91415393e-01 -5.82909882e-01 1.50336236e-01
1.46303311e-01 5.52252606e-02 3.40870589e-01 1.53065085e-01
4.60155487e-01 -6.10583901e-01 -3.97011697e-01 3.17693859e-01
-7.89764285e-01 -2.52224594e-01 4.49782968e-01 1.02506781e+00
6.93079948e-01 -2.78690197e-02 6.02578282e-01 -1.19856393e+00
9.60546970e-01 -6.48724854e-01 -1.62731752e-01 8.42557609e-01
-9.66333747e-01 1.13867342e-01 5.68013608e-01 -3.95969748e-01
-5.56793571e-01 -3.40938836e-01 2.27047607e-01 -7.09662735e-01
4.97696817e-01 1.08896661e+00 -4.08633143e-01 -1.87125430e-01
4.59376425e-01 2.47216672e-01 -4.94867653e-01 -2.44687945e-01
1.13219082e+00 5.31459391e-01 3.96420479e-01 -4.62682724e-01
1.25292897e+00 1.29206330e-01 3.16071808e-02 -4.74534750e-01
-5.88680923e-01 -6.30366206e-01 -6.42903388e-01 3.23031038e-01
8.21148932e-01 -8.80104601e-01 -6.76773190e-01 -5.58161624e-02
-1.18396401e+00 2.53549159e-01 -1.69525057e-01 6.97494030e-01
-1.22925043e-01 4.09637034e-01 -6.07119501e-01 -3.81321907e-01
-5.58121741e-01 -7.75545239e-01 6.61502600e-01 3.74126211e-02
-1.95801139e-01 -1.37982881e+00 1.51553929e-01 3.08916211e-01
3.99959385e-01 1.45840734e-01 1.01089120e+00 -9.25373256e-01
-4.34901774e-01 -4.00862068e-01 -4.84703869e-01 2.02267513e-01
2.98378766e-01 8.21690634e-02 -2.52172768e-01 -2.39978358e-01
-8.15617442e-01 -2.65580714e-01 8.21026385e-01 -3.58838618e-01
7.54814029e-01 -4.42539454e-01 -8.14099193e-01 7.80993223e-01
1.38111973e+00 -2.28823259e-01 6.77384079e-01 3.73760074e-01
1.30542254e+00 2.43828461e-01 5.90861559e-01 4.49484922e-02
1.23303723e+00 7.83647835e-01 2.16983572e-01 6.55733496e-02
-1.29486784e-01 -5.97615123e-01 1.64385140e-01 1.37884462e+00
-3.25608879e-01 -2.73335874e-01 -7.53635287e-01 8.53049517e-01
-2.07563710e+00 -8.18980515e-01 -3.96074206e-01 1.91182327e+00
9.68847990e-01 1.32182866e-01 -1.38259277e-01 -3.18703353e-02
5.67665458e-01 1.81176081e-01 -5.76972514e-02 -2.70502508e-01
-8.71888623e-02 6.18786179e-02 5.30930459e-01 4.33986038e-01
-1.04430604e+00 1.43694210e+00 5.78211117e+00 6.53041184e-01
-8.58227491e-01 2.05115005e-01 -2.00258285e-01 3.50731403e-01
-7.86021173e-01 3.62771571e-01 -8.73334587e-01 1.67917728e-01
6.06651604e-01 -5.45531750e-01 4.63054657e-01 5.95790148e-01
-4.21204716e-01 7.09467471e-01 -1.33146083e+00 1.00531387e+00
1.50771216e-01 -1.34832168e+00 2.94772923e-01 -1.72526147e-02
7.46773422e-01 -1.39871061e-01 -1.07910618e-01 8.60973358e-01
3.77035469e-01 -1.01622140e+00 1.12501733e-01 4.24656540e-01
7.45502353e-01 -5.95719814e-01 8.49162161e-01 -1.92181185e-01
-1.79959691e+00 3.19009274e-01 -4.51962113e-01 2.54642248e-01
2.41204143e-01 4.82399046e-01 -8.31976950e-01 1.34137368e+00
3.99435759e-01 9.18665767e-01 -5.10399222e-01 6.98977411e-01
-4.36461627e-01 4.91040915e-01 -1.04199000e-01 -4.76102792e-02
1.12986341e-01 -3.78881156e-01 3.89024019e-01 1.27114987e+00
2.60961473e-01 6.06179535e-02 2.41804779e-01 6.18770063e-01
-5.85281909e-01 4.12766963e-01 -8.32613468e-01 -3.77997130e-01
9.25755382e-01 1.31234407e+00 -3.60709950e-02 -2.60220915e-01
-9.22318697e-01 9.02823508e-01 8.65357161e-01 6.97622120e-01
-9.80986893e-01 -6.83067977e-01 7.90059149e-01 -8.52519125e-02
4.64761376e-01 -2.45860755e-01 6.62027970e-02 -1.29851091e+00
4.27679330e-01 -4.70642537e-01 6.19742870e-01 -5.48432767e-01
-1.50060213e+00 4.66217130e-01 -5.31727523e-02 -1.03189707e+00
1.37717649e-01 -4.77583379e-01 -6.04197025e-01 6.96590722e-01
-1.64803696e+00 -1.80769217e+00 6.02956377e-02 6.37936473e-01
-2.12369800e-01 -3.07686090e-01 1.02358651e+00 6.26107156e-01
-7.20391989e-01 1.05840707e+00 1.08652316e-01 5.51840663e-01
6.47305787e-01 -1.30805111e+00 4.96121556e-01 5.21167874e-01
3.44579756e-01 1.12684071e+00 3.68616939e-01 -5.87123930e-01
-1.83070505e+00 -1.38907623e+00 1.47218418e+00 -5.76058090e-01
1.28455412e+00 -2.77396142e-01 -1.07423687e+00 1.37104118e+00
3.50984633e-01 5.85688710e-01 9.30990338e-01 7.48767734e-01
-8.67255270e-01 -1.28434211e-01 -6.88919067e-01 6.49106741e-01
1.30584168e+00 -7.53675580e-01 -9.13410664e-01 4.06523019e-01
1.17657101e+00 -3.32071900e-01 -1.63510382e+00 5.84481061e-01
5.15749693e-01 -4.05035734e-01 1.22236764e+00 -1.10407376e+00
4.38826382e-01 -4.18511540e-01 -3.93556684e-01 -1.56237900e+00
-6.30170286e-01 -4.97416645e-01 -8.80616784e-01 1.51078820e+00
7.39172041e-01 -1.01400352e+00 4.84499782e-01 2.90074676e-01
-1.25777990e-01 -1.22407591e+00 -1.07934403e+00 -1.05952442e+00
7.94609636e-02 3.20804268e-02 9.35608625e-01 1.69750702e+00
4.41137046e-01 7.55043030e-01 -4.73129809e-01 3.95999998e-01
4.67387885e-01 3.19991261e-01 9.22724783e-01 -1.17419600e+00
-2.68930137e-01 -3.91916186e-01 -9.78701115e-01 -8.09361041e-01
3.14849347e-01 -1.16562724e+00 -6.56543911e-01 -2.09551096e+00
2.15863526e-01 -3.39849114e-01 -5.44586658e-01 6.62682593e-01
-5.54629028e-01 3.79085578e-02 -9.91782472e-02 1.03581011e-01
-7.67152786e-01 9.82723951e-01 1.27192974e+00 -1.95885107e-01
-4.36022878e-03 -5.45055687e-01 -9.00921047e-01 3.74079436e-01
4.62629765e-01 -3.71342421e-01 -5.60732603e-01 -4.64913219e-01
5.40438712e-01 -1.05189696e-01 9.77343842e-02 -4.23604697e-01
3.54652703e-01 -1.42338127e-01 -5.37450761e-02 -3.53139818e-01
4.19721961e-01 -6.59312069e-01 2.66806722e-01 3.45877674e-03
-1.45984814e-01 1.59955457e-01 -3.06846380e-01 7.55493164e-01
-4.99067754e-01 2.72450358e-01 -8.25846451e-04 3.82841736e-01
-8.43559802e-01 6.81252003e-01 6.30200863e-01 2.20628396e-01
1.25417209e+00 1.72169253e-01 -8.25866520e-01 -1.60833389e-01
-6.95877194e-01 6.74135745e-01 4.88338917e-02 9.05092776e-01
6.95432246e-01 -2.02949047e+00 -8.08991015e-01 -3.80256847e-02
5.91382980e-01 -6.19236417e-02 -1.44204289e-01 1.14019322e+00
-2.75961965e-01 5.06494105e-01 2.80879557e-01 -5.30952960e-02
-1.25847101e+00 7.88332522e-01 1.78607330e-01 -4.75571126e-01
-5.63589096e-01 1.00896597e+00 7.17060268e-02 -9.54986036e-01
8.01588148e-02 -2.14872628e-01 -3.04463923e-01 8.03552046e-02
1.21730760e-01 2.84050226e-01 -7.77309760e-02 -8.47114742e-01
-6.73313856e-01 7.22488284e-01 -2.13775247e-01 3.04582268e-01
1.15557182e+00 6.54318184e-02 -5.06460369e-01 3.34309489e-01
1.54424536e+00 4.70313318e-02 -3.03232521e-01 -5.96288323e-01
-4.40581376e-03 -5.22073448e-01 -6.03225343e-02 -4.20346946e-01
-1.01505089e+00 4.49822098e-01 1.19220257e-01 3.84340465e-01
5.37767410e-01 2.10843429e-01 9.18391109e-01 4.88963187e-01
4.09136504e-01 -8.60888004e-01 -1.91038832e-01 5.58483839e-01
8.43762279e-01 -1.22902131e+00 1.84323579e-01 -7.75047839e-01
-7.30887711e-01 8.08251441e-01 8.14873636e-01 7.66183212e-02
1.01746488e+00 -1.79849073e-01 -8.93602073e-02 -4.76578891e-01
-9.99332130e-01 -3.91861588e-01 6.40329480e-01 5.81405938e-01
6.49281979e-01 4.45178896e-01 -6.77421033e-01 4.88026291e-01
-2.09185198e-01 -6.68608919e-02 2.22900957e-01 8.64592969e-01
-4.05435860e-02 -1.48147404e+00 9.17572230e-02 5.91465294e-01
-1.51668683e-01 -3.13415796e-01 -5.24953127e-01 1.02996206e+00
4.21637483e-02 8.69120836e-01 -3.16594988e-01 -9.87497449e-01
6.88282311e-01 2.25599602e-01 2.43848711e-01 -5.66913962e-01
-3.23401272e-01 -3.31895441e-01 4.08598095e-01 -4.57676739e-01
-1.08528025e-01 -4.09163803e-01 -1.37810397e+00 -4.15449142e-01
-7.98275113e-01 2.65956670e-01 2.79593945e-01 7.27500021e-01
5.43935597e-01 7.72842109e-01 5.19801557e-01 -6.94286376e-02
-6.29584670e-01 -1.01788700e+00 -7.39597797e-01 6.86066508e-01
-1.57668032e-02 -9.24655020e-01 -1.17217429e-01 -2.82261133e-01] | [8.752547264099121, 7.87985897064209] |
b2b709e1-c581-4a8e-8052-af42404edc19 | self-supervised-contrastive-learning-for-eeg | 2109.07839 | null | https://arxiv.org/abs/2109.07839v1 | https://arxiv.org/pdf/2109.07839v1.pdf | Self-supervised Contrastive Learning for EEG-based Sleep Staging | EEG signals are usually simple to obtain but expensive to label. Although supervised learning has been widely used in the field of EEG signal analysis, its generalization performance is limited by the amount of annotated data. Self-supervised learning (SSL), as a popular learning paradigm in computer vision (CV) and natural language processing (NLP), can employ unlabeled data to make up for the data shortage of supervised learning. In this paper, we propose a self-supervised contrastive learning method of EEG signals for sleep stage classification. During the training process, we set up a pretext task for the network in order to match the right transformation pairs generated from EEG signals. In this way, the network improves the representation ability by learning the general features of EEG signals. The robustness of the network also gets improved in dealing with diverse data, that is, extracting constant features from changing data. In detail, the network's performance depends on the choice of transformations and the amount of unlabeled data used in the training process of self-supervised learning. Empirical evaluations on the Sleep-edf dataset demonstrate the competitive performance of our method on sleep staging (88.16% accuracy and 81.96% F1 score) and verify the effectiveness of SSL strategy for EEG signal analysis in limited labeled data regimes. All codes are provided publicly online. | ['Zhiyong Yuan', 'Bo Du', 'Jianhui Zhao', 'Xue Jiang'] | 2021-09-16 | null | null | null | null | ['sleep-staging', 'eeg-based-sleep-staging'] | ['medical', 'time-series'] | [ 2.92074531e-01 -1.48720637e-01 -2.36874819e-01 -7.00700521e-01
-3.66578341e-01 -2.75408745e-01 1.59281820e-01 -2.64838953e-02
-6.81387365e-01 1.01100588e+00 -1.44716026e-02 8.26244205e-02
-1.62298828e-01 -4.27738070e-01 -3.41840982e-01 -9.61188018e-01
1.72803216e-02 1.17359154e-01 -5.52663393e-02 -1.36875913e-01
9.71717015e-02 2.89585292e-01 -1.40976238e+00 7.15551525e-02
1.13557065e+00 1.10617745e+00 3.85439962e-01 3.27720270e-02
-2.31150255e-01 6.70568466e-01 -5.59705675e-01 1.82539269e-01
3.56997401e-02 -6.44709706e-01 -6.26308024e-01 1.94110394e-01
-2.41969183e-01 2.73605019e-01 -3.19573402e-01 1.19208336e+00
5.46631634e-01 -2.78285779e-02 6.60952687e-01 -1.37558031e+00
-3.40445161e-01 6.75188541e-01 -3.89440060e-01 7.60749698e-01
1.56364411e-01 -5.80083812e-03 6.62132740e-01 -9.23852503e-01
1.86351776e-01 5.75814486e-01 4.60368782e-01 7.64977992e-01
-1.15641654e+00 -1.18395877e+00 4.22345065e-02 4.84959960e-01
-1.51407540e+00 -5.34358740e-01 9.63565290e-01 -3.35266143e-01
5.02434194e-01 1.01855323e-01 7.12608516e-01 1.06924307e+00
4.68949169e-01 6.22492433e-01 1.41877627e+00 -4.61139113e-01
4.81381714e-01 3.74174595e-01 4.94675368e-01 4.67840672e-01
-6.48934841e-02 3.42959613e-02 -7.30041921e-01 2.23707065e-01
5.42751253e-01 7.63542503e-02 -4.47874159e-01 -1.93145588e-01
-1.20593107e+00 5.49352527e-01 4.82083797e-01 6.90389693e-01
-1.58658832e-01 -4.06204611e-01 3.71614814e-01 5.66002131e-01
6.96546733e-01 5.84166765e-01 -5.37908077e-01 -1.67575583e-01
-1.03489506e+00 -4.02208686e-01 4.37013417e-01 8.33625436e-01
7.61503458e-01 2.54604906e-01 -1.33283615e-01 1.00853741e+00
-1.52539918e-02 3.74944746e-01 1.12607157e+00 -3.45847756e-01
3.84542763e-01 8.35366070e-01 -2.54973829e-01 -7.22808480e-01
-8.11126232e-01 -6.15307868e-01 -1.13893533e+00 -2.28788659e-01
2.32206970e-01 -2.66322017e-01 -6.82868898e-01 1.69143164e+00
-1.06506772e-01 3.01856428e-01 1.18367314e-01 6.48257971e-01
9.02713716e-01 6.97259128e-01 -9.34150293e-02 -7.81702697e-01
1.05234659e+00 -7.80237377e-01 -9.66454148e-01 -4.80792046e-01
4.49898988e-01 -2.46724993e-01 1.07509589e+00 5.08753240e-01
-6.37663186e-01 -7.49801636e-01 -1.09046841e+00 3.49491924e-01
-2.63833851e-01 3.12791288e-01 5.03544450e-01 4.87884700e-01
-8.45457137e-01 4.17150557e-01 -7.86868751e-01 -3.67918342e-01
6.44084275e-01 8.79087031e-01 -4.40124869e-01 1.24038428e-01
-1.18778026e+00 7.85579264e-01 6.04070008e-01 1.59445181e-01
-5.71633816e-01 -5.31074643e-01 -5.84752083e-01 2.05166847e-01
2.10648641e-01 -5.98415174e-02 8.77381206e-01 -1.27955759e+00
-1.34203780e+00 6.73490405e-01 -1.38921320e-01 -4.94893879e-01
2.11569473e-01 2.16209918e-01 -7.54158795e-01 9.35914293e-02
7.71124139e-02 6.00478888e-01 1.04231536e+00 -8.98518145e-01
-5.65538764e-01 -6.21201158e-01 -4.70070392e-01 1.68025449e-01
-8.13776016e-01 -1.67138755e-01 -6.79847673e-02 -4.83792484e-01
2.21557930e-01 -9.02814865e-01 5.01653999e-02 -2.64421403e-01
-7.98546374e-02 -4.46922690e-01 6.82661831e-01 -3.77192646e-01
1.19047320e+00 -2.53710389e+00 -2.66971290e-02 3.58341455e-01
1.41314551e-01 2.72059202e-01 2.61467993e-02 1.23875635e-02
-3.67155880e-01 -2.88646311e-01 -4.09515977e-01 -7.30518475e-02
-3.01370203e-01 1.93960458e-01 -9.84942019e-02 4.81886446e-01
9.88770947e-02 8.94632816e-01 -9.45485830e-01 -5.33724368e-01
3.56099159e-02 6.94506243e-02 -1.42376661e-01 2.85137415e-01
3.16862881e-01 9.76144791e-01 -2.20063686e-01 3.31651747e-01
2.75648504e-01 -1.88828975e-01 -7.36111123e-03 -2.57519841e-01
-8.67013782e-02 1.49302974e-01 -9.81475353e-01 1.84845471e+00
-5.55206358e-01 7.60361493e-01 -4.49965835e-01 -1.25248122e+00
1.24715459e+00 3.13813955e-01 5.76174736e-01 -8.57719779e-01
3.48682553e-01 8.00884366e-02 3.64631772e-01 -6.86353326e-01
-1.72125041e-01 -5.89482710e-02 2.17485353e-02 5.04805923e-01
3.92837405e-01 6.71173213e-04 1.48541182e-01 -1.00019023e-01
8.62194359e-01 -2.17842609e-01 4.78619993e-01 -5.08477569e-01
6.24539077e-01 -1.75132781e-01 7.36498058e-01 4.97189820e-01
-1.43422320e-01 1.62059844e-01 1.73374221e-01 -3.46910030e-01
-6.41127765e-01 -6.80922508e-01 -4.39251810e-01 8.84191871e-01
2.13159591e-01 -2.33873397e-01 -6.56871378e-01 -6.36493385e-01
-4.83093023e-01 5.45306206e-01 -5.32617569e-01 -7.87085831e-01
-2.47490332e-01 -1.11464560e+00 3.57258648e-01 5.34436047e-01
7.71456480e-01 -1.35922289e+00 -3.75691593e-01 1.38728723e-01
-5.83143085e-02 -1.09888947e+00 -2.75926620e-01 6.05249107e-01
-1.06403363e+00 -9.61003661e-01 -4.39585507e-01 -1.21785223e+00
1.07531822e+00 2.33029798e-01 5.94404876e-01 -3.81883932e-03
-3.74521837e-02 1.92858465e-02 -3.95830721e-01 -4.91848618e-01
-3.36165220e-01 2.29585111e-01 4.83423024e-01 3.74909103e-01
6.12353325e-01 -8.17591548e-01 -4.55439836e-01 4.14342374e-01
-7.79597580e-01 6.27771467e-02 6.28253579e-01 1.08970189e+00
5.43500721e-01 5.71970940e-01 9.94048834e-01 -8.54516983e-01
6.96777821e-01 -4.14006591e-01 -3.47688884e-01 2.09674045e-01
-8.49193037e-01 8.07170346e-02 9.08853233e-01 -6.20403171e-01
-1.00963414e+00 1.63016900e-01 -8.83847177e-02 -1.25549883e-01
-2.26551622e-01 4.97666776e-01 -2.62877673e-01 -2.25550398e-01
7.95284867e-01 6.18575215e-01 -1.10929392e-01 -2.53081590e-01
-1.56846091e-01 1.02433097e+00 5.54636240e-01 -1.76275998e-01
6.50463045e-01 1.79691792e-01 -3.15498710e-01 -7.58630395e-01
-9.64283407e-01 -4.29871917e-01 -8.64949226e-01 -2.44771212e-01
7.03805149e-01 -8.32311809e-01 -4.48858738e-01 4.72186297e-01
-6.64217174e-01 -3.07098925e-01 -2.45372891e-01 7.27553666e-01
-2.59821147e-01 1.04870848e-01 -2.63367146e-01 -6.50934041e-01
-5.32763183e-01 -9.65824425e-01 4.68433261e-01 3.09911281e-01
-1.45092770e-01 -8.54965091e-01 -8.99909735e-02 1.64019004e-01
2.66826749e-01 -2.46620297e-01 1.07370424e+00 -9.10614073e-01
9.97576416e-02 -8.28016773e-02 -6.34612069e-02 5.90479672e-01
6.41778767e-01 -5.83576679e-01 -1.12188923e+00 -3.77739906e-01
5.20197093e-01 -3.10104191e-01 6.76862121e-01 1.67056188e-01
1.36122084e+00 -1.94164097e-01 -1.70615330e-01 5.42546511e-01
1.15300214e+00 5.93780458e-01 4.46840316e-01 2.99649864e-01
5.82818270e-01 4.90010113e-01 4.19183642e-01 2.39106447e-01
1.20037474e-01 2.78672397e-01 7.13359714e-02 -1.40759632e-01
8.52002650e-02 -7.49920681e-02 2.91570038e-01 1.26924455e+00
4.90217134e-02 3.27101976e-01 -8.87256086e-01 1.68605745e-01
-1.63127458e+00 -6.50979161e-01 8.26716125e-02 2.21922970e+00
1.12783575e+00 3.54841053e-01 -2.64637209e-02 6.58389568e-01
6.23452485e-01 -2.43242860e-01 -7.89105415e-01 8.77977535e-02
-1.13627315e-01 4.12480891e-01 2.24324331e-01 1.60246789e-02
-9.38242257e-01 7.34136701e-01 5.30675268e+00 8.40916395e-01
-1.28880763e+00 3.03229779e-01 5.09136856e-01 -3.76106277e-02
2.17066869e-01 -3.37802142e-01 -5.87807357e-01 8.45132649e-01
8.61866534e-01 -1.69172227e-01 6.70062184e-01 5.84230602e-01
4.77657080e-01 -9.71550494e-02 -1.24815691e+00 1.51787853e+00
1.99414268e-01 -1.02185261e+00 -2.20294669e-01 -3.45955998e-01
7.69020379e-01 -3.99170890e-02 -1.67165950e-01 4.19803232e-01
-3.45045924e-01 -8.80394995e-01 3.94623160e-01 3.68477434e-01
7.98579097e-01 -8.18715751e-01 8.38296592e-01 6.78395748e-01
-8.66529584e-01 -4.22070593e-01 -4.21569616e-01 -7.78466538e-02
-3.52307618e-01 5.06268978e-01 -6.92313790e-01 3.10654700e-01
7.38729358e-01 1.08110797e+00 -9.59052920e-01 1.06969011e+00
-4.57754344e-01 9.13980126e-01 5.69529273e-02 -1.54077038e-01
-3.52323316e-02 -2.59027570e-01 1.34791121e-01 9.82809961e-01
1.07959792e-01 2.29821995e-01 1.39627615e-02 7.14599490e-01
-1.45033255e-01 2.20642507e-01 -4.43195581e-01 8.02143365e-02
4.47100639e-01 1.20859933e+00 -8.45763564e-01 -7.83826783e-02
-1.76389068e-01 7.55867124e-01 4.46064383e-01 3.02842170e-01
-6.04538858e-01 -2.33892009e-01 -9.41446945e-02 -1.91765487e-01
-1.34241983e-01 -8.79381225e-02 -3.68537366e-01 -1.32399249e+00
-1.26983377e-03 -7.59605765e-01 1.38750553e-01 -7.19512582e-01
-1.42241371e+00 9.37199652e-01 -7.08861323e-03 -1.47372139e+00
6.72959387e-02 -2.28032947e-01 -6.52934968e-01 3.84287417e-01
-1.50246370e+00 -6.67616785e-01 -6.05002105e-01 1.01727366e+00
6.45024598e-01 -5.63025534e-01 8.23006332e-01 3.61358464e-01
-8.33480954e-01 6.04237556e-01 2.55867928e-01 1.79819271e-01
8.21150184e-01 -1.01370287e+00 -6.43851608e-02 6.78805113e-01
4.60094452e-01 6.10577583e-01 3.34256798e-01 -2.43344098e-01
-1.18243098e+00 -1.06492639e+00 7.23756790e-01 2.19221115e-01
7.12688804e-01 -6.20677888e-01 -1.01432157e+00 3.20781827e-01
9.25974846e-02 1.93813965e-02 1.00509715e+00 -3.22197676e-02
8.74025896e-02 -7.44964778e-01 -9.68480110e-01 3.61660272e-01
8.43953848e-01 -5.57275236e-01 -9.09495890e-01 4.28328216e-01
2.48832077e-01 -1.43939584e-01 -6.88292742e-01 3.18703562e-01
2.21343234e-01 -6.17731154e-01 4.54369068e-01 -4.25308466e-01
5.11516333e-02 -2.37150192e-01 2.24870756e-01 -1.59038901e+00
-3.83161843e-01 -3.40335727e-01 1.10382140e-01 1.20496571e+00
2.99197763e-01 -7.48180151e-01 5.58304548e-01 4.73866791e-01
-2.79376693e-02 -8.30365658e-01 -8.37848186e-01 -7.34679520e-01
-3.59646976e-01 -3.55952293e-01 4.78744388e-01 9.63335395e-01
3.72381240e-01 7.01356530e-01 -2.82826960e-01 2.95004193e-02
4.07099277e-01 3.89385223e-02 4.17831719e-01 -1.50227988e+00
2.64523514e-02 -9.00417715e-02 -4.31810349e-01 -6.36600077e-01
5.76390982e-01 -1.14428270e+00 7.01711401e-02 -1.27658200e+00
2.06397742e-01 -6.45153463e-01 -7.25311100e-01 7.33597577e-01
-1.36875451e-01 3.82455111e-01 7.05135986e-02 5.28477252e-01
-6.58550918e-01 6.43657506e-01 1.09025407e+00 -3.19182634e-01
-6.22760713e-01 2.47673288e-01 -5.54066360e-01 6.95778787e-01
1.20268667e+00 -7.67968297e-01 -7.98035979e-01 -2.41164997e-01
-3.83334495e-02 -1.68388337e-01 6.86743185e-02 -1.21232235e+00
4.49266225e-01 1.42744839e-01 6.69001818e-01 -1.61142781e-01
1.05216697e-01 -9.83676076e-01 -1.25447571e-01 4.49095100e-01
-5.39967656e-01 -1.14859380e-01 1.12005293e-01 4.36368972e-01
-2.52620816e-01 -4.06360149e-01 7.35023499e-01 4.65510897e-02
-7.87638903e-01 3.62633556e-01 -4.56101418e-01 5.87225333e-02
8.63191962e-01 -2.73184478e-01 -7.86941946e-02 -2.28987142e-01
-7.91745484e-01 2.43045643e-01 9.76066366e-02 4.73005295e-01
6.85305119e-01 -1.24084330e+00 -4.41654503e-01 6.74242735e-01
3.18232030e-01 -1.18113175e-01 4.71374094e-02 1.05587471e+00
-2.53157020e-02 1.46763816e-01 -5.48096359e-01 -7.04072952e-01
-1.00889981e+00 4.96199012e-01 1.97775975e-01 -2.73709428e-02
-5.09037435e-01 5.32011747e-01 -8.66861269e-03 -6.06559850e-02
4.37925935e-01 -3.02890301e-01 -6.08894706e-01 1.51726305e-01
6.10086560e-01 6.00952096e-02 4.33075070e-01 -4.38886434e-01
-3.52080911e-01 2.54265696e-01 -1.16679281e-01 2.75556922e-01
1.53099179e+00 -9.42427590e-02 -2.88344949e-01 8.77156377e-01
1.22751379e+00 -3.43550086e-01 -1.05987358e+00 -3.58552784e-01
1.86525717e-01 -3.84557843e-02 1.40807822e-01 -6.89622998e-01
-1.20864296e+00 7.18560576e-01 9.54596996e-01 4.97710258e-02
1.39531291e+00 -1.68621868e-01 5.27743518e-01 6.54824793e-01
6.80846810e-01 -1.02533829e+00 1.65852960e-02 2.15940967e-01
5.15777707e-01 -1.34917212e+00 1.88602079e-02 -7.79893771e-02
-7.23277092e-01 1.14280844e+00 6.51465535e-01 -1.85938448e-01
6.90012574e-01 1.96662441e-01 3.96042056e-02 -1.39403746e-01
-3.84813011e-01 8.46334323e-02 3.59076113e-01 4.32160348e-01
8.76613483e-02 -1.08351357e-01 -3.99797797e-01 9.50250745e-01
-4.42424297e-01 1.47391334e-01 3.36609066e-01 8.76614571e-01
-3.67415935e-01 -9.98990893e-01 -9.53958258e-02 8.31283808e-01
-2.35667601e-01 -9.70174819e-02 -1.78560570e-01 4.98795122e-01
3.38015199e-01 1.22608078e+00 1.90312974e-02 -5.41288018e-01
1.67026073e-01 1.36789829e-01 4.53746825e-01 -8.86416078e-01
-5.99415898e-01 -4.44701985e-02 -4.18533653e-01 -1.71566799e-01
-8.12501609e-01 -4.10837531e-01 -1.31857681e+00 3.69964451e-01
-4.58350807e-01 5.30291855e-01 4.17112023e-01 1.35367620e+00
2.13546440e-01 5.66563845e-01 1.06011117e+00 -5.88265479e-01
-3.53973210e-01 -1.05548751e+00 -8.23594630e-01 3.07156682e-01
3.13046008e-01 -7.82759964e-01 -4.03688699e-01 2.48356521e-01] | [13.18875789642334, 3.4714975357055664] |
17dfbe77-6af9-4289-965e-846779d885d7 | segmentation-of-drilled-holes-in-texture | null | null | https://www.mdpi.com/1424-8220/21/11/3633 | https://www.mdpi.com/1424-8220/21/11/3633 | Segmentation of Drilled Holes in Texture Wooden Furniture Panels Using Deep Neural Network | Drilling operations are an essential part of furniture from MDF laminated boards required for product assembly. Faults in the process might introduce adverse effects to the furniture. Inspection of the drilling quality can be challenging due to a big variety of board surface textures, dust, or woodchips in the manufacturing process, milling cutouts, and other kinds of defects. Intelligent computer vision methods can be engaged for global contextual analysis with local information attention for automated object detection and segmentation. In this paper, we propose blind and through drilled holes segmentation on textured wooden furniture panel images using the UNet encoder-decoder modifications enhanced with residual connections, atrous spatial pyramid pooling, squeeze and excitation module, and CoordConv layers for better segmentation performance. We show that even a lightweight architecture is capable to perform on a range of complex textures and is able to distinguish the holes drilling operations’ semantical information from the rest of the furniture board and conveyor context. The proposed model configurations yield better results in more complex cases with a not significant or small bump in processing time. Experimental results demonstrate that our best-proposed solution achieves a Dice score of up to 97.89% compared to the baseline U-Net model’s Dice score of 94.50%. Statistical, visual, and computational properties of each convolutional neural network architecture are addressed. | ['Tadas Surgailis', 'Arūnas Lipnickas', 'Rytis Augustauskas'] | 2021-05-23 | null | null | null | mdpi-sensors-2021-5 | ['unet-segmentation', '2d-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 5.35731137e-01 8.96398947e-02 2.07524911e-01 -1.92236498e-01
-3.07346076e-01 -4.46447760e-01 9.84025151e-02 -1.29579101e-02
5.72016314e-02 1.85696498e-01 -2.44310781e-01 -3.98120843e-02
-4.54002708e-01 -7.78426051e-01 -7.44182467e-01 -5.92188478e-01
2.09363043e-01 3.57317209e-01 3.62276256e-01 -1.76900163e-01
3.28637689e-01 8.15976739e-01 -1.52679253e+00 8.56849551e-01
5.30582011e-01 1.51540184e+00 8.55862975e-01 6.71925366e-01
1.02647364e-01 5.91638446e-01 -6.45973384e-01 -5.02125263e-01
3.83389592e-01 -3.44837829e-02 -5.69331527e-01 7.79202521e-01
5.86199820e-01 -4.56201673e-01 -6.24284409e-02 1.09645557e+00
2.25082800e-01 -2.99544055e-02 7.38839090e-01 -7.52294183e-01
-7.73584723e-01 4.43403184e-01 -6.49813831e-01 -1.14841871e-01
-3.24576609e-02 5.03132269e-02 8.50175202e-01 -8.91932428e-01
4.02870119e-01 1.35288906e+00 5.91784239e-01 8.45179632e-02
-8.06952715e-01 -1.98024690e-01 1.44758448e-01 1.32004529e-01
-7.43320584e-01 -4.82838266e-02 6.33162022e-01 -4.88010734e-01
7.87472010e-01 4.29703027e-01 4.65261787e-01 7.86566198e-01
6.18571699e-01 8.64586234e-01 9.75344062e-01 -2.61794657e-01
4.04732376e-02 -1.63746640e-01 2.60226846e-01 1.00513017e+00
2.42942095e-01 -7.85333589e-02 -1.98108315e-01 6.05058968e-01
1.22503400e+00 1.90969035e-01 -2.66453743e-01 -2.77261525e-01
-1.16121340e+00 4.05729681e-01 5.94014764e-01 7.22316578e-02
-4.36274558e-01 2.23317116e-01 3.23817074e-01 3.45902681e-01
2.86841482e-01 6.20396316e-01 -6.69013679e-01 1.41604483e-01
-9.14647102e-01 -2.83516854e-01 7.28413820e-01 1.01325428e+00
4.79074925e-01 3.28852050e-02 -1.51616201e-01 1.08138573e+00
1.43643215e-01 3.45419914e-01 9.23139378e-02 -8.66526365e-01
4.60237771e-01 4.82667267e-01 4.67717974e-03 -1.25847435e+00
-4.74582553e-01 -4.41963702e-01 -1.01840186e+00 3.46087307e-01
1.93669483e-01 -5.94530925e-02 -1.42846918e+00 7.71631837e-01
-9.67765227e-02 -4.64317381e-01 -3.86232764e-01 1.15891016e+00
8.44406009e-01 4.83518332e-01 -3.68845642e-01 4.85453784e-01
1.60061145e+00 -1.16462290e+00 -9.37130928e-01 -4.92134213e-01
2.65918789e-03 -1.12475646e+00 8.99709880e-01 7.22161114e-01
-1.24461710e+00 -7.40710557e-01 -1.48968422e+00 -1.59369126e-01
-4.16348368e-01 9.84678984e-01 8.22943628e-01 5.90022922e-01
-8.03105056e-01 8.21846962e-01 -9.54160392e-01 -3.44284177e-01
5.46862602e-01 4.62085396e-01 -2.94136524e-01 -3.45301777e-01
-5.80614686e-01 8.73951018e-01 2.41794482e-01 9.02853668e-01
-9.51790512e-01 -2.95698643e-01 -8.44052732e-01 1.88705310e-01
5.94469309e-01 -5.10364115e-01 9.86203015e-01 -5.30618608e-01
-1.57564127e+00 5.38846910e-01 2.65007734e-01 -4.78466190e-02
5.44065177e-01 -5.72477043e-01 -1.00691833e-01 1.38515204e-01
-1.23123661e-01 6.30171835e-01 9.64506507e-01 -1.16516531e+00
-3.67664397e-01 -4.47542906e-01 -5.39178215e-02 6.16678363e-03
1.45117402e-01 -1.93794414e-01 -6.75665855e-01 -8.32931817e-01
9.33165312e-01 -6.27239108e-01 -2.49252841e-01 1.01123378e-01
-5.40152311e-01 4.13874775e-01 8.81478548e-01 -1.09059703e+00
5.92384219e-01 -2.12253118e+00 -1.36421502e-01 2.58127272e-01
-2.13378027e-01 -5.41788712e-02 -1.01781823e-01 1.93526059e-01
-4.50382233e-02 -1.04395196e-01 -2.44090900e-01 -5.56026176e-02
1.10147886e-01 1.97353959e-01 -2.18991339e-02 3.90144080e-01
4.50393021e-01 8.09239984e-01 -5.09812713e-01 1.02371730e-01
4.95838165e-01 2.17838436e-01 -5.12935638e-01 -3.70320193e-02
-5.52892685e-02 9.54509825e-02 -2.33355984e-01 1.17828429e+00
7.03710914e-01 -7.96839595e-02 -3.72599252e-02 -5.42750359e-01
1.58040762e-01 1.16841076e-02 -1.30136621e+00 1.89135933e+00
-6.78321600e-01 7.99391747e-01 6.62033081e-01 -7.42780328e-01
1.19074929e+00 5.43091558e-02 -7.27900788e-02 -5.02502561e-01
2.93777913e-01 2.16975451e-01 2.01608557e-02 -6.53618157e-01
6.95298910e-01 4.19813506e-02 -8.39683414e-02 -1.65554509e-01
4.79446314e-02 -5.57859540e-01 -4.00618091e-02 -3.81457806e-01
1.09617698e+00 3.77429947e-02 -2.73428887e-01 -3.01577717e-01
1.75995201e-01 -3.23410332e-01 3.50253403e-01 6.95375264e-01
1.05668850e-01 9.85047042e-01 4.47723716e-01 -5.60555100e-01
-1.00678313e+00 -1.06759906e+00 -2.19650865e-02 7.82346725e-01
3.55451465e-01 2.89406419e-01 -7.17682004e-01 -3.42840344e-01
-8.67747813e-02 3.76552582e-01 -5.40164709e-01 -9.10092890e-02
-4.26155865e-01 -7.28110254e-01 7.28284791e-02 6.35016501e-01
1.09661794e+00 -1.03837430e+00 -8.36024284e-01 2.70776421e-01
-2.78740277e-04 -1.26119781e+00 -3.21915179e-01 6.71851158e-01
-1.00028074e+00 -1.03802741e+00 -7.14236915e-01 -1.14958298e+00
6.86642587e-01 1.61816195e-01 9.10810411e-01 -4.30232853e-01
-5.69879234e-01 1.04932666e-01 -2.59654403e-01 -1.51564240e-01
-2.13714182e-01 -9.83009711e-02 -5.22608757e-01 -2.12929174e-02
1.80809349e-01 -1.04483962e-01 -5.93699872e-01 6.87453687e-01
-7.80979276e-01 2.13914812e-01 9.38359380e-01 1.01252162e+00
4.16837037e-01 5.44767678e-01 1.99772581e-01 -6.97679937e-01
5.20576239e-01 -1.06228329e-01 -4.91972297e-01 2.32195958e-01
-1.35102436e-01 -1.96857288e-01 2.00804487e-01 3.36689390e-02
-1.15681672e+00 -1.86935484e-01 4.69109192e-02 -2.70381719e-01
-5.48546970e-01 3.27016950e-01 -5.59991360e-01 2.22841561e-01
8.54889154e-02 -3.92350197e-01 -1.56524897e-01 -6.21313512e-01
1.54645965e-01 6.54283643e-01 7.45474756e-01 -3.68723899e-01
3.53668272e-01 4.16234493e-01 -1.01365149e-01 -8.41579914e-01
-6.50626242e-01 -3.17059100e-01 -6.72906160e-01 -3.25160176e-01
1.19026768e+00 -6.23801410e-01 -5.90083361e-01 7.31122971e-01
-1.26761949e+00 -1.79982245e-01 -5.11717796e-02 5.68566144e-01
-4.87296790e-01 3.65172863e-01 -1.08366811e+00 -6.29092455e-01
-2.10441604e-01 -1.41861320e+00 1.07413507e+00 -1.61833763e-02
-1.89330980e-01 -4.14911985e-01 -9.03522313e-01 7.52839148e-01
1.62877813e-01 3.80236924e-01 1.13706863e+00 -2.42284283e-01
-7.10613549e-01 -3.77433330e-01 -5.08919954e-01 7.17612565e-01
3.08890045e-01 -6.03151508e-02 -1.08900201e+00 -2.97035277e-02
2.83599734e-01 2.42352225e-02 1.17684889e+00 6.60226226e-01
1.30097699e+00 2.32366361e-02 -1.25956014e-01 4.31586176e-01
1.29856479e+00 5.56790173e-01 7.52169609e-01 1.31405875e-01
7.09659815e-01 9.17131901e-01 9.00096655e-01 3.19821268e-01
-3.57117265e-01 3.18890750e-01 9.02094483e-01 -1.88998073e-01
-3.05367947e-01 2.44960692e-02 4.15390551e-01 8.11237037e-01
-9.55344141e-02 -4.67748970e-01 -8.06623161e-01 7.61309743e-01
-1.56377637e+00 -5.79870105e-01 -2.99925178e-01 1.98306382e+00
2.45566338e-01 3.51016819e-01 -7.20474482e-01 3.40292126e-01
9.49684441e-01 -1.11248866e-01 -3.63147318e-01 -8.53606999e-01
-1.76160336e-01 2.61980504e-01 8.27415884e-01 3.05946648e-01
-1.35974097e+00 7.56686926e-01 5.50836039e+00 8.98894846e-01
-7.57016957e-01 -1.12215474e-01 7.18360960e-01 -7.39844590e-02
1.58923909e-01 -5.59353411e-01 -3.75721425e-01 2.54204720e-01
4.23239917e-01 1.25784159e+00 3.55693519e-01 5.46199381e-01
9.80741978e-02 -3.69462639e-01 -1.05622017e+00 1.05894876e+00
1.15158193e-01 -1.17512226e+00 -1.81090280e-01 -1.38814300e-01
8.52065384e-01 -1.05308987e-01 3.70274365e-01 -2.65136987e-01
2.32261598e-01 -9.96988118e-01 8.50781918e-01 3.44938904e-01
5.25755405e-01 -7.92116463e-01 1.14324248e+00 -1.80044025e-01
-8.49679649e-01 -3.21351111e-01 -4.48361635e-01 -2.13999841e-02
1.46096637e-02 9.34952557e-01 -8.13507318e-01 5.27065039e-01
9.49359775e-01 5.89316249e-01 -2.44083241e-01 1.18068075e+00
-2.89080918e-01 3.84996891e-01 -3.10632169e-01 1.63328871e-02
6.13618076e-01 -3.59121114e-01 4.22406524e-01 1.01624084e+00
5.89045644e-01 -2.41214231e-01 -4.81292568e-02 8.56570423e-01
9.70536321e-02 -2.81643301e-01 -4.47713763e-01 3.61709855e-02
-3.80908772e-02 1.21984756e+00 -1.34734237e+00 -1.66547876e-02
-2.11302236e-01 1.31658304e+00 -2.82019854e-01 4.30478007e-01
-8.62949669e-01 -5.01997650e-01 3.92168730e-01 -8.66972096e-03
8.15494657e-01 -3.50920260e-01 -7.84720361e-01 -6.19855642e-01
3.05494368e-01 -6.30492747e-01 -6.12062588e-02 -1.11427379e+00
-1.17987955e+00 3.58863831e-01 -4.15722758e-01 -9.99659061e-01
5.82665443e-01 -1.18880200e+00 -1.98988527e-01 7.02977836e-01
-1.19387269e+00 -1.24470103e+00 -4.08547163e-01 6.62847459e-02
9.72641528e-01 -1.78112745e-01 4.97650683e-01 1.69921130e-01
-5.86853325e-01 1.67859882e-01 1.45954683e-01 1.68511629e-01
4.03155714e-01 -1.22927129e+00 2.45816991e-01 8.35964561e-01
-1.61978945e-01 2.51812965e-01 5.17214298e-01 -6.80314064e-01
-1.34875703e+00 -1.19864988e+00 3.03654879e-01 -1.90068573e-01
4.03418064e-01 -4.54186469e-01 -5.70894420e-01 4.68922913e-01
3.44767272e-01 -4.70634520e-01 3.91241282e-01 -5.40687367e-02
6.27370998e-02 2.07113847e-03 -1.18043137e+00 5.74767947e-01
7.21152604e-01 -3.84511888e-01 -6.76001847e-01 2.43446320e-01
2.33717039e-01 -5.53000033e-01 -8.56158495e-01 4.81707096e-01
5.87042570e-01 -9.68741417e-01 9.03139591e-01 -1.04249895e-01
8.15897703e-01 -3.02366048e-01 -4.13345486e-01 -9.81463492e-01
-2.80347407e-01 -4.92618442e-01 4.15775955e-01 1.04372382e+00
4.33183402e-01 -3.60653609e-01 8.76812875e-01 5.45948803e-01
-7.82520413e-01 -6.31450534e-01 -7.50078440e-01 -6.08760536e-01
-3.88699383e-01 -5.87311804e-01 3.78976256e-01 6.60986900e-01
-3.75903964e-01 1.73550963e-01 -1.73783958e-01 5.01391292e-01
2.71222293e-01 2.02812791e-01 2.46888682e-01 -1.22376800e+00
-1.63387701e-01 -4.69095230e-01 -7.03135848e-01 -1.23818707e+00
-3.47764164e-01 -6.59890652e-01 3.58694375e-01 -1.79731262e+00
-1.29152685e-01 -3.14339697e-01 -2.88823396e-01 4.62718695e-01
2.85859853e-01 3.33922476e-01 1.54552430e-01 -2.55409122e-01
-1.50568888e-01 3.60280663e-01 1.77168477e+00 -5.85436642e-01
-3.20073664e-02 6.70001507e-02 -1.77750185e-01 8.62985313e-01
7.68121898e-01 -1.15339085e-01 -3.12617540e-01 -8.69814277e-01
1.60320878e-01 2.59260111e-03 6.71512842e-01 -1.15269518e+00
1.14163034e-01 4.21570778e-01 6.56199455e-01 -9.75696445e-01
4.71417010e-01 -1.34031940e+00 1.05695605e-01 5.04728258e-01
-1.07199632e-01 -1.18711635e-01 3.80618125e-01 5.86990058e-01
-4.40636486e-01 -3.64580214e-01 6.50967062e-01 -3.11276734e-01
-8.25585961e-01 -1.41691133e-01 -5.92059672e-01 -6.66351438e-01
7.96728611e-01 -7.21518695e-01 -1.67916954e-01 -4.87459376e-02
-1.02017128e+00 -2.30215136e-02 3.41660321e-01 4.52217132e-01
8.01645994e-01 -9.99506593e-01 -3.89564723e-01 5.82181871e-01
-3.25329691e-01 2.41729349e-01 5.01420021e-01 7.62231410e-01
-1.10359251e+00 5.59638560e-01 -3.90381783e-01 -6.58265710e-01
-8.60355556e-01 2.16461420e-01 2.62118876e-01 -1.14144340e-01
-3.75456095e-01 1.20489514e+00 3.34179729e-01 -2.80859083e-01
1.33652285e-01 -1.24424589e+00 2.26261374e-02 2.73539454e-01
-3.76638398e-02 6.27574921e-01 6.07820630e-01 -1.30757257e-01
-7.89548978e-02 6.84457600e-01 -1.41319245e-01 2.84543127e-01
1.28477168e+00 -1.10989161e-01 -1.55171216e-01 2.97474891e-01
9.59348083e-01 -3.23569953e-01 -1.46149838e+00 2.17072010e-01
-1.10324770e-01 -3.97444189e-01 3.49746555e-01 -1.24082780e+00
-1.52430248e+00 1.14079905e+00 7.37289786e-01 1.47872925e-01
1.23335528e+00 -1.31940514e-01 7.01580703e-01 5.15292883e-01
3.38110656e-01 -1.20479143e+00 9.01904032e-02 5.18806159e-01
1.11944604e+00 -1.10703969e+00 -1.59551874e-01 -7.33130157e-01
-5.73133051e-01 1.34343195e+00 5.51482916e-01 -2.00936198e-01
5.28526187e-01 3.63960922e-01 -1.43121667e-02 -4.77404028e-01
-3.15572888e-01 1.39901685e-02 4.17099684e-01 4.95208412e-01
1.96764097e-01 6.28889650e-02 3.65997374e-01 8.14795494e-01
-1.33453295e-01 -3.39268118e-01 4.36323464e-01 9.70814943e-01
-5.10744989e-01 -3.53617221e-01 -5.40686011e-01 5.83947182e-01
-5.81009865e-01 -1.61253870e-01 -2.93181062e-01 7.95779884e-01
3.13090265e-01 1.09782469e+00 3.45154554e-01 -2.56929666e-01
3.55081677e-01 -1.72672793e-01 4.81815666e-01 -5.62442660e-01
-7.78463364e-01 4.53706235e-01 1.68133736e-01 -6.03932858e-01
-3.21701378e-01 -4.35585946e-01 -9.29520726e-01 1.19233906e-01
-5.83504021e-01 -3.45291138e-01 7.09628403e-01 6.89732015e-01
8.99229571e-02 1.06947696e+00 2.48678535e-01 -8.84778380e-01
7.14028552e-02 -9.91335273e-01 -1.01526761e+00 1.21904999e-01
2.17019126e-01 -6.41465783e-01 -2.84415577e-03 2.05038041e-01] | [7.490850448608398, 1.7877473831176758] |
cf3523ff-0aa1-4fea-881f-ef3d4ba69611 | hsmd-an-object-motion-detection-algorithm | 2109.04119 | null | https://arxiv.org/abs/2109.04119v1 | https://arxiv.org/pdf/2109.04119v1.pdf | HSMD: An object motion detection algorithm using a Hybrid Spiking Neural Network Architecture | The detection of moving objects is a trivial task performed by vertebrate retinas, yet a complex computer vision task. Object-motion-sensitive ganglion cells (OMS-GC) are specialised cells in the retina that sense moving objects. OMS-GC take as input continuous signals and produce spike patterns as output, that are transmitted to the Visual Cortex via the optic nerve. The Hybrid Sensitive Motion Detector (HSMD) algorithm proposed in this work enhances the GSOC dynamic background subtraction (DBS) algorithm with a customised 3-layer spiking neural network (SNN) that outputs spiking responses akin to the OMS-GC. The algorithm was compared against existing background subtraction (BS) approaches, available on the OpenCV library, specifically on the 2012 change detection (CDnet2012) and the 2014 change detection (CDnet2014) benchmark datasets. The results show that the HSMD was ranked overall first among the competing approaches and has performed better than all the other algorithms on four of the categories across all the eight test metrics. Furthermore, the HSMD proposed in this paper is the first to use an SNN to enhance an existing state of the art DBS (GSOC) algorithm and the results demonstrate that the SNN provides near real-time performance in realistic applications. | ['T. M. McGinnity', 'Joao Filipe Ferreira', 'Andreas Oikonomou', 'Pedro Machado'] | 2021-09-09 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 6.09377563e-01 -7.15169013e-01 7.36899257e-01 1.61658615e-01
-1.15957167e-02 -5.33497691e-01 8.06115448e-01 -3.05262983e-01
-1.09746277e+00 7.39106417e-01 -2.74173230e-01 -7.76387081e-02
2.72983342e-01 -3.04939300e-01 -7.12572157e-01 -1.14041877e+00
4.85201068e-02 -2.88606972e-01 1.43502295e+00 -2.06713095e-01
5.09681463e-01 6.03357971e-01 -2.10227990e+00 4.30464417e-01
3.86225522e-01 1.29789364e+00 5.53688705e-01 9.90103602e-01
4.37735170e-02 7.06935704e-01 -5.62007129e-01 -1.16416581e-01
5.99958301e-01 -5.47720909e-01 -2.41447285e-01 -5.86464882e-01
6.61333144e-01 2.39487633e-01 -1.04273722e-01 1.14881933e+00
8.10176671e-01 -4.24233526e-02 3.20317358e-01 -1.12865305e+00
-2.39113405e-01 6.80702701e-02 -4.45838451e-01 1.27628446e+00
-3.37208100e-02 6.11785173e-01 9.08402577e-02 -9.17404532e-01
9.68566895e-01 1.29254663e+00 8.32307994e-01 7.17229009e-01
-1.59471571e+00 -6.52789950e-01 2.83365720e-03 2.52136707e-01
-1.12747419e+00 -5.40252805e-01 1.62887663e-01 -7.10872769e-01
1.35495484e+00 3.74172717e-01 9.14791942e-01 1.01161802e+00
5.51634014e-01 6.26303315e-01 1.32874596e+00 -1.07045092e-01
8.74154091e-01 -4.80987281e-01 3.53335649e-01 8.86175260e-02
6.53066754e-01 2.20823511e-01 -7.08968163e-01 5.64322695e-02
8.32490802e-01 -3.12912971e-01 -4.32177603e-01 -1.46413162e-01
-1.52446246e+00 2.50531048e-01 5.39654493e-01 1.75987095e-01
-4.38154727e-01 4.70064104e-01 4.20789212e-01 1.67135075e-01
5.44205084e-02 5.02204858e-02 -3.73306334e-01 -9.34813023e-02
-1.01509786e+00 3.67166519e-01 6.77921712e-01 4.49654132e-01
4.68130410e-01 2.62744427e-01 -5.23899853e-01 3.91927779e-01
2.48876214e-01 6.93588197e-01 6.67088568e-01 -1.03369653e+00
-1.62145924e-02 7.44603097e-01 4.57319543e-02 -6.92913175e-01
-5.19190907e-01 -3.12019497e-01 -8.59240472e-01 1.03105760e+00
4.90845561e-01 6.08844422e-02 -1.48470342e+00 1.52302766e+00
-3.67416106e-02 5.83347261e-01 1.45844921e-01 9.69914854e-01
1.13844192e+00 5.49661398e-01 2.28935778e-01 -1.94913387e-01
1.06282842e+00 -4.98644084e-01 -4.76420939e-01 -5.81877351e-01
9.29168612e-02 -6.03829801e-01 5.32224417e-01 3.89194041e-01
-9.74359214e-01 -4.99159515e-01 -1.21063423e+00 2.53006041e-01
-6.09165967e-01 -3.37530114e-02 5.14718056e-01 5.50887048e-01
-1.61564720e+00 4.27982062e-01 -1.15805113e+00 -6.90586329e-01
6.86727822e-01 6.13376737e-01 -2.30401501e-01 2.40501106e-01
-5.75160801e-01 1.05089819e+00 3.74194533e-01 5.54595068e-02
-9.05426025e-01 -4.71235245e-01 -3.28859419e-01 -4.48342174e-01
-1.32524133e-01 -7.72613764e-01 1.20725381e+00 -1.33999372e+00
-1.38463461e+00 1.35650611e+00 -1.71241224e-01 -9.36644435e-01
4.94711667e-01 2.22389415e-01 -4.16739672e-01 1.15573898e-01
-1.56012541e-02 1.21573985e+00 5.97283483e-01 -1.09487402e+00
-8.99219155e-01 -3.16013455e-01 -4.03835446e-01 -2.71428287e-01
5.37823081e-01 3.55884612e-01 -3.57408702e-01 -7.72975028e-01
1.91788957e-01 -8.71525466e-01 -2.04322487e-01 2.72577345e-01
1.35686565e-02 1.66544497e-01 9.20973122e-01 -5.24525702e-01
9.46385145e-01 -2.25568414e+00 1.78947479e-01 -2.61219293e-01
-7.61204660e-02 1.01236916e+00 -1.70375630e-01 4.46026847e-02
-4.04068604e-02 -5.41681945e-01 -7.57763684e-01 -9.10779759e-02
-4.04198259e-01 1.13722339e-01 -1.95228279e-01 6.10368431e-01
3.18034887e-01 8.62456441e-01 -7.14375317e-01 9.00579430e-03
2.55820841e-01 4.39206809e-01 -1.87246278e-01 -4.73913789e-01
3.59895751e-02 4.16109562e-01 4.15448993e-01 8.10054064e-01
9.29833829e-01 4.21743840e-02 -6.16550028e-01 1.99516211e-02
-7.14081466e-01 -9.74339172e-02 -1.33561039e+00 1.53919327e+00
3.64619315e-01 1.25192368e+00 3.00305299e-02 -7.65992582e-01
9.93071437e-01 -1.56681105e-01 3.08766048e-02 -8.47497523e-01
2.10118607e-01 4.55344111e-01 4.41494733e-01 -3.07257324e-01
2.54416522e-02 1.79924533e-01 3.34094197e-01 -3.26452523e-01
1.80863127e-01 1.30561262e-01 4.59698290e-01 8.20776746e-02
1.62523139e+00 3.11680615e-01 3.81497473e-01 -5.02029061e-01
6.04244411e-01 1.14990883e-01 8.31945956e-01 7.29391158e-01
-7.74187207e-01 7.74044156e-01 1.33486450e-01 -3.36892098e-01
-5.60690999e-01 -1.48738134e+00 -1.30959645e-01 6.48119450e-01
3.25569719e-01 2.92021483e-01 -8.11905622e-01 2.06467316e-01
2.46390849e-02 5.01342297e-01 -5.65044403e-01 -6.98744133e-02
-3.60453248e-01 -1.11887491e+00 6.87888205e-01 5.32635689e-01
7.64286578e-01 -1.51751328e+00 -1.60167694e+00 5.25797248e-01
1.50262341e-01 -1.08756006e+00 2.13519379e-01 5.55205584e-01
-7.67095029e-01 -9.73730981e-01 -8.18757415e-01 -9.56491888e-01
3.79204065e-01 3.83328348e-01 7.00817823e-01 -4.63110954e-01
-9.56803977e-01 2.10666969e-01 -1.57133386e-01 -1.01074147e+00
-9.75145474e-02 -7.91617692e-01 3.42004411e-02 7.55527914e-02
9.07576919e-01 -5.65693915e-01 -7.79569089e-01 1.79201573e-01
-1.03887677e+00 9.64731351e-02 4.96360570e-01 2.28737399e-01
6.42769277e-01 -4.88379031e-01 3.31321388e-01 -2.33914912e-01
5.02964295e-02 -1.42928824e-01 -1.16046500e+00 -1.80696771e-01
-9.54933912e-02 -1.98213130e-01 2.18576699e-01 -5.36951005e-01
-7.64420807e-01 5.42667806e-01 1.89617231e-01 -8.56384635e-02
-2.76720971e-01 -2.49737591e-01 1.14921503e-01 -6.45715892e-01
1.01746595e+00 4.48870271e-01 -2.88221866e-01 -1.77926496e-01
-3.20684314e-01 3.87686878e-01 1.32704878e+00 5.32916248e-01
3.81291568e-01 1.20798874e+00 1.47387445e-01 -8.59738171e-01
-4.34997082e-02 -9.26483035e-01 -6.10899210e-01 -4.84620541e-01
9.84505713e-01 -1.00946701e+00 -5.70304930e-01 1.47284281e+00
-1.31891811e+00 -6.56079710e-01 -5.77970669e-02 4.86330718e-01
-6.81276917e-01 7.93096051e-02 -4.08603281e-01 -1.03497744e+00
-2.63831705e-01 -8.18515658e-01 9.07973588e-01 6.07041895e-01
-4.27008457e-02 -4.72728193e-01 3.35633397e-01 -2.45074540e-01
7.76582122e-01 6.84196472e-01 1.97767720e-01 -2.82878399e-01
-7.51682639e-01 8.68084468e-03 -4.07364070e-01 2.08531946e-01
-7.81508982e-02 4.75732893e-01 -1.29062438e+00 -1.02692917e-01
1.67191848e-02 2.89424300e-01 1.63540959e+00 1.08775699e+00
5.12131155e-01 3.93435329e-01 -3.95125002e-01 7.90718079e-01
1.95438838e+00 5.59653580e-01 1.22403467e+00 9.21074390e-01
1.02109060e-01 3.99505377e-01 1.84761047e-01 4.01673228e-01
-8.91393125e-02 4.71251041e-01 9.10473347e-01 -2.43605580e-02
-4.59233493e-01 6.68642700e-01 6.60580397e-01 2.34202698e-01
-3.47747922e-01 -4.73482460e-02 -1.03243446e+00 7.32261479e-01
-2.01282287e+00 -1.11855197e+00 -8.73344064e-01 2.37872291e+00
5.83303809e-01 3.25440586e-01 1.90845743e-01 2.45475829e-01
9.43412721e-01 -3.14441204e-01 -7.56529212e-01 -2.04962239e-01
-1.01858175e+00 5.48888206e-01 9.23144102e-01 -7.85413980e-02
-1.23039889e+00 8.56011450e-01 5.81127930e+00 1.84769541e-01
-1.36414933e+00 1.13774650e-01 2.86513716e-02 -2.94658691e-01
6.38229072e-01 -9.56172571e-02 -9.79474008e-01 8.13314795e-01
1.11226213e+00 -9.93120074e-02 6.48974121e-01 3.22703689e-01
4.09075975e-01 -8.43394816e-01 -7.00099528e-01 1.31418037e+00
-9.08310321e-05 -1.51243341e+00 -2.46025175e-01 -2.18851298e-01
7.95972288e-01 7.56588221e-01 -4.45515849e-02 1.14656687e-01
2.18492165e-01 -7.63951361e-01 9.29694295e-01 9.50537145e-01
2.45645180e-01 -3.10240299e-01 8.63917410e-01 2.20972449e-01
-9.65719938e-01 -2.62092799e-01 -7.49801874e-01 -2.46818766e-01
5.60113182e-03 4.86160517e-01 -1.80682093e-01 7.66850868e-03
1.34578204e+00 6.87974155e-01 -1.03376794e+00 2.08433247e+00
3.04541618e-01 5.22437274e-01 -3.55348557e-01 -1.06492236e-01
3.46657902e-01 4.09956649e-02 1.03548205e+00 1.70052350e+00
3.31215262e-01 -5.71512990e-02 -6.18034244e-01 8.96681964e-01
2.92633712e-01 -1.69287905e-01 -4.10389870e-01 1.77418023e-01
1.03871234e-01 1.37210453e+00 -1.30427623e+00 -4.44058508e-01
-3.63447070e-01 1.04108918e+00 -2.56402344e-01 4.14380461e-01
-6.16250396e-01 -4.45758492e-01 9.21652555e-01 1.65158827e-02
7.95270979e-01 -2.00370476e-02 -3.42377603e-01 -7.24246204e-01
9.27715935e-03 -4.19419229e-01 2.65794247e-01 -1.15774715e+00
-1.06175673e+00 5.22579670e-01 -2.63699561e-01 -1.30741191e+00
2.47129813e-01 -1.22043419e+00 -8.09824586e-01 6.90339983e-01
-1.45779669e+00 -6.35484219e-01 -7.26291180e-01 7.30745137e-01
3.96967530e-01 -1.34630308e-01 5.03132522e-01 2.17691839e-01
-5.65036356e-01 -1.08510874e-01 2.66025275e-01 -1.47491861e-02
6.55549288e-01 -1.07771397e+00 8.90509427e-01 1.50865257e+00
-1.57480150e-01 2.85852283e-01 7.94795930e-01 -6.11679137e-01
-1.08616209e+00 -1.34506750e+00 3.03528547e-01 -3.57459515e-01
5.17272830e-01 -3.56510520e-01 -1.11312687e+00 3.30848396e-01
1.02822825e-01 4.18546170e-01 3.42518628e-01 -1.32597816e+00
-9.61798728e-02 -7.08101913e-02 -1.16650605e+00 4.93832022e-01
1.19184721e+00 1.06586898e-02 -7.24461973e-01 -2.84272075e-01
3.77383322e-01 -3.13241005e-01 -5.30122519e-02 4.33642328e-01
3.49923909e-01 -1.37657332e+00 8.64488184e-01 -2.12777898e-01
-5.06053045e-02 -1.07328856e+00 -3.94144088e-01 -1.04269493e+00
-4.07768995e-01 -6.02024913e-01 -3.55581976e-02 1.00360262e+00
9.79737937e-02 -9.48608994e-01 4.02034461e-01 8.91220495e-02
-3.34746599e-01 -3.54069509e-02 -1.40835464e+00 -1.18967426e+00
-3.44627827e-01 -4.69445050e-01 -1.10324128e-02 5.38977683e-01
-6.40543938e-01 -1.50183961e-01 3.62603873e-01 2.15469956e-01
7.19385922e-01 -3.00146729e-01 5.79382122e-01 -1.53935421e+00
-9.44830384e-03 -8.39214325e-01 -1.57997239e+00 -3.22946787e-01
-4.65949178e-01 -8.16560268e-01 3.66885692e-01 -1.78493583e+00
5.54932617e-02 4.21690941e-01 -6.82483613e-01 3.77199024e-01
-1.55518264e-01 7.12993681e-01 1.42376482e-01 1.60267949e-01
-5.92725158e-01 -1.31698415e-01 6.78428173e-01 7.36910552e-02
-3.23834985e-01 8.76874924e-02 -2.52254874e-01 7.95877218e-01
6.03265822e-01 -5.24282157e-01 2.47918084e-01 -1.16782904e-01
-3.67016979e-02 -7.44785130e-01 1.18566954e+00 -1.80086517e+00
6.53620005e-01 1.19937025e-01 5.83795846e-01 -7.08613396e-01
1.41239092e-01 -4.52620327e-01 4.98092592e-01 1.10406709e+00
2.29866460e-01 -1.50469095e-01 8.31830144e-01 7.03722298e-01
2.56156437e-02 1.71725765e-01 1.36758840e+00 -1.33122444e-01
-1.54149830e+00 -2.02073723e-01 -1.03829670e+00 9.37893614e-02
1.16063929e+00 -9.36089218e-01 -6.39647722e-01 2.96597362e-01
-4.25943613e-01 -2.31392637e-01 5.71426392e-01 3.53905886e-01
5.23913860e-01 -1.04739034e+00 -6.71474338e-01 1.42032176e-01
2.47278720e-01 -9.34651494e-02 -7.44727999e-02 1.29223251e+00
-9.73151624e-01 5.22821367e-01 -9.62047696e-01 -1.12607086e+00
-1.35756457e+00 4.38305527e-01 4.47612524e-01 5.38193703e-01
-5.91048002e-01 1.22386432e+00 3.01444888e-01 1.68469787e-01
2.47817293e-01 -8.15798819e-01 -1.95586011e-01 2.18286943e-02
8.89012575e-01 7.42147684e-01 3.26144308e-01 -6.05087698e-01
-7.48488128e-01 5.19238114e-01 5.21152377e-01 -1.96305409e-01
1.44159222e+00 -5.95066585e-02 -3.36394966e-01 7.10472107e-01
7.42823780e-01 -5.61854064e-01 -1.68186271e+00 -7.47803450e-02
3.41563672e-01 -3.22110206e-01 4.38146340e-03 -9.31921542e-01
-8.20388496e-01 8.37654829e-01 1.31332159e+00 -3.77644189e-02
1.56681085e+00 -2.96783268e-01 3.75230879e-01 2.85166025e-01
4.26005512e-01 -1.08306527e+00 -1.18842535e-01 3.60680312e-01
7.86401808e-01 -9.74341333e-01 -3.95562708e-01 -2.00028956e-01
-7.11527407e-01 1.09304762e+00 7.20869780e-01 -4.31380808e-01
2.91895092e-01 5.77751219e-01 2.04325497e-01 5.93210794e-02
-8.25600505e-01 -6.68971479e-01 4.20961231e-02 8.49354863e-01
-1.97898541e-02 -2.77506351e-01 -1.61277860e-01 2.39055857e-01
2.43528053e-01 3.85139793e-01 6.03217304e-01 1.19013035e+00
-8.82100046e-01 -3.66901428e-01 -4.27979141e-01 4.61409986e-01
-4.01616156e-01 -1.73572794e-01 -8.16734314e-01 5.79259813e-01
3.72895181e-01 7.97070265e-01 2.88228840e-01 -1.38279386e-02
4.02964294e-01 4.61445302e-02 5.25912285e-01 -4.58831280e-01
-7.93042898e-01 -9.26586017e-02 -4.45792913e-01 -7.84319639e-01
-9.25815046e-01 -7.25524426e-01 -1.54160643e+00 -8.61299690e-03
-7.00132772e-02 -7.63869941e-01 9.80746269e-01 6.33311450e-01
3.71631682e-01 6.54221356e-01 -3.02252974e-02 -1.18394792e+00
1.47626698e-01 -7.51185238e-01 -6.12806141e-01 2.03735396e-01
4.66729254e-01 -7.18702734e-01 -6.92322075e-01 4.72359538e-01] | [8.671516418457031, -1.110706090927124] |
a20c3442-d763-4a73-b7d3-87f56633ea83 | 2012-11655 | 2012.11655 | null | https://arxiv.org/abs/2012.11655v3 | https://arxiv.org/pdf/2012.11655v3.pdf | Learning Dynamic Network Using a Reuse Gate Function in Semi-supervised Video Object Segmentation | Current state-of-the-art approaches for Semi-supervised Video Object Segmentation (Semi-VOS) propagates information from previous frames to generate segmentation mask for the current frame. This results in high-quality segmentation across challenging scenarios such as changes in appearance and occlusion. But it also leads to unnecessary computations for stationary or slow-moving objects where the change across frames is minimal. In this work, we exploit this observation by using temporal information to quickly identify frames with minimal change and skip the heavyweight mask generation step. To realize this efficiency, we propose a novel dynamic network that estimates change across frames and decides which path -- computing a full network or reusing previous frame's feature -- to choose depending on the expected similarity. Experimental results show that our approach significantly improves inference speed without much accuracy degradation on challenging Semi-VOS datasets -- DAVIS 16, DAVIS 17, and YouTube-VOS. Furthermore, our approach can be applied to multiple Semi-VOS methods demonstrating its generality. The code is available in https://github.com/HYOJINPARK/Reuse_VOS. | ['Nojun Kwak', 'Ganesh Venkatesh', 'Seohyeong Jeong', 'Jayeon Yoo', 'Hyojin Park'] | 2020-12-21 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Park_Learning_Dynamic_Network_Using_a_Reuse_Gate_Function_in_Semi-Supervised_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Park_Learning_Dynamic_Network_Using_a_Reuse_Gate_Function_in_Semi-Supervised_CVPR_2021_paper.pdf | cvpr-2021-1 | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 1.67325616e-01 -1.06482007e-01 -3.71542573e-01 -4.43183988e-01
-4.76354748e-01 -4.54257488e-01 1.35232329e-01 -1.90790296e-01
-4.34559852e-01 6.52372956e-01 -1.95212752e-01 -8.03496987e-02
2.24812999e-01 -4.51848060e-01 -7.04063594e-01 -6.17357135e-01
1.58940945e-02 3.05076361e-01 9.94283795e-01 1.85538054e-01
-3.53680477e-02 4.98474479e-01 -1.51700878e+00 2.07488254e-01
7.11448312e-01 9.09724534e-01 3.74399334e-01 8.21551442e-01
-1.97788417e-01 7.26477504e-01 -2.80048430e-01 -1.94222089e-02
5.34808874e-01 -4.08535391e-01 -8.83385956e-01 4.20815736e-01
5.95476925e-01 -4.99332190e-01 -3.04302901e-01 9.65885699e-01
3.47656727e-01 2.38648564e-01 2.23314345e-01 -1.33825946e+00
1.47716384e-02 4.74463850e-01 -8.74980390e-01 5.10489047e-01
2.16713682e-01 4.29648817e-01 7.86693752e-01 -6.67687654e-01
1.08738792e+00 1.24512434e+00 7.23869383e-01 6.37068331e-01
-1.14147258e+00 -5.24067879e-01 4.93523896e-01 1.88606769e-01
-1.27633893e+00 -6.59902751e-01 7.54118800e-01 -4.10606414e-01
5.68502545e-01 2.37034217e-01 8.65316570e-01 7.55332768e-01
-1.01635732e-01 1.28091085e+00 9.36842561e-01 -1.78046808e-01
9.77491736e-02 -1.11549430e-01 1.34736747e-01 9.36788857e-01
1.09151058e-01 -2.14615941e-01 -5.20210385e-01 8.72125775e-02
7.74755418e-01 -5.65343536e-02 -4.07552183e-01 -3.00529927e-01
-1.22409618e+00 4.13279563e-01 4.50615957e-02 2.43376210e-01
-2.75787830e-01 3.04491341e-01 3.92734915e-01 3.24340224e-01
7.01770484e-01 -1.31388664e-01 -5.99588871e-01 -3.63303781e-01
-1.47036839e+00 2.95491368e-01 7.04075575e-01 9.46795344e-01
8.62674773e-01 1.00081451e-01 -1.94006637e-01 6.94311142e-01
3.98767710e-01 2.33199239e-01 2.12616593e-01 -1.41233611e+00
3.57010424e-01 3.95516664e-01 1.25781953e-01 -8.73828351e-01
-3.95792931e-01 -5.05847931e-02 -5.13335586e-01 2.76570112e-01
7.20753491e-01 -1.77006021e-01 -1.16024530e+00 1.56739902e+00
8.36747289e-01 5.28738916e-01 -3.43624860e-01 1.02769625e+00
8.95315468e-01 5.49829602e-01 9.06908233e-03 -4.22423214e-01
1.10601318e+00 -1.22853017e+00 -7.36175418e-01 -1.73992634e-01
3.88013542e-01 -6.43518746e-01 9.52740848e-01 2.69907624e-01
-1.31266010e+00 -5.77313304e-01 -9.42451358e-01 1.35921106e-01
-8.22695568e-02 -2.18058135e-02 4.84463394e-01 4.56435323e-01
-1.17233241e+00 9.72936094e-01 -1.23500085e+00 -3.92755270e-01
7.48635590e-01 5.14619768e-01 -5.66710345e-02 2.63887104e-02
-8.87606442e-01 3.90886605e-01 4.99375135e-01 -6.29957579e-03
-8.82887244e-01 -8.32130194e-01 -7.42869794e-01 -1.91673175e-01
7.29369164e-01 -6.97803974e-01 1.27862501e+00 -1.43212569e+00
-1.56882179e+00 6.75157487e-01 -5.41685998e-01 -5.15343785e-01
9.87504900e-01 -1.70252174e-01 -1.07529469e-01 4.35788363e-01
1.03246704e-01 1.06098247e+00 1.09492207e+00 -1.17617023e+00
-7.53900170e-01 -1.59708276e-01 1.35684907e-01 2.37502158e-01
-1.93038970e-01 2.09358051e-01 -1.04080319e+00 -6.02246523e-01
2.04146326e-01 -1.09023809e+00 -4.58081186e-01 4.16062921e-01
-4.75157738e-01 -2.14736402e-01 1.05011773e+00 -5.63855708e-01
1.29756963e+00 -2.11788321e+00 -2.82543171e-02 1.07218407e-01
2.52174765e-01 5.64849913e-01 -3.40465307e-02 -1.56024154e-02
2.48011097e-01 5.33374511e-02 -3.89302760e-01 -6.01342440e-01
-3.39093745e-01 3.20282429e-01 1.78158075e-01 6.21387959e-01
1.26351148e-01 8.92221868e-01 -9.17155385e-01 -9.93537724e-01
4.43403780e-01 4.19528633e-01 -3.93422633e-01 1.11174002e-01
-4.10043120e-01 5.35413563e-01 -3.41695487e-01 8.42812777e-01
6.88737094e-01 -4.05335218e-01 -1.57640018e-02 -1.88184917e-01
-9.98648256e-02 -4.61641364e-02 -1.53877831e+00 1.75817204e+00
-7.82829300e-02 7.85904706e-01 1.69253543e-01 -7.98318982e-01
5.00660002e-01 2.03079492e-01 7.16603696e-01 -3.02000135e-01
2.57191747e-01 -9.39061213e-03 -1.31879017e-01 -6.32999063e-01
3.45846057e-01 3.77463251e-01 5.52972794e-01 2.87855178e-01
3.10961287e-02 2.43276954e-02 8.32307398e-01 3.27821016e-01
7.66720176e-01 6.57927752e-01 2.11644486e-01 -2.72207201e-01
5.60199201e-01 2.35137921e-02 9.94882524e-01 7.01688290e-01
-4.80132192e-01 7.69861758e-01 3.42989981e-01 -4.99258727e-01
-7.11195886e-01 -8.39514315e-01 3.06274761e-02 9.05281007e-01
5.28666556e-01 -5.19987404e-01 -9.46820378e-01 -7.86931992e-01
-3.89706232e-02 3.14632595e-01 -4.98895168e-01 3.47204417e-01
-6.53784752e-01 -5.60136616e-01 2.94742495e-01 5.51822484e-01
5.10989428e-01 -1.12371480e+00 -7.78175533e-01 2.86178380e-01
-2.55876899e-01 -1.31951475e+00 -7.03329265e-01 -2.20666274e-01
-1.05728269e+00 -1.15398872e+00 -7.41247892e-01 -6.32367551e-01
7.93098688e-01 3.83614808e-01 9.32104766e-01 2.30306670e-01
-4.48899835e-01 3.74484986e-01 -1.76641569e-01 -6.49970099e-02
-4.29638803e-01 -5.17142601e-02 -8.95331129e-02 2.14412510e-01
8.15349072e-02 -4.86412406e-01 -7.94962764e-01 4.59532171e-01
-7.09806263e-01 2.93603927e-01 7.93637931e-02 3.96428049e-01
9.46687818e-01 -8.80609378e-02 2.75904506e-01 -9.74259675e-01
-5.22572473e-02 -3.12967837e-01 -7.57081628e-01 2.04171792e-01
-5.26793420e-01 -2.58375555e-01 4.26234126e-01 -6.45886004e-01
-1.07477045e+00 3.26203197e-01 3.53163257e-02 -6.26038790e-01
-3.28111619e-01 3.63791585e-02 1.19367406e-01 -1.85249195e-01
1.85977891e-01 5.89249842e-02 8.45504110e-04 -4.31817591e-01
2.80810624e-01 3.94815564e-01 5.08312881e-01 -4.08763915e-01
7.05459118e-01 9.11226571e-01 -9.45218354e-02 -8.69719565e-01
-7.81736910e-01 -7.68773794e-01 -7.91229904e-01 -6.53941512e-01
9.70554531e-01 -9.22230065e-01 -6.38267696e-01 6.00444555e-01
-1.16018856e+00 -7.21568346e-01 -3.12345177e-01 3.27721864e-01
-4.45038974e-01 6.08028889e-01 -8.16216886e-01 -7.03563452e-01
-2.99400061e-01 -1.24008214e+00 9.07706738e-01 5.38957119e-01
-2.50431210e-01 -1.07327068e+00 -1.61682636e-01 3.08569342e-01
1.42473608e-01 3.86087477e-01 5.78320362e-02 -1.84613004e-01
-7.83808947e-01 2.51566410e-01 -3.14637005e-01 2.63789743e-01
2.23317847e-01 5.26339650e-01 -7.28039801e-01 -2.80387312e-01
-3.82805824e-01 -5.68452068e-02 9.63207424e-01 7.42475808e-01
1.24284756e+00 -3.24651599e-02 -3.48308861e-01 8.19747865e-01
1.35899043e+00 2.34788179e-01 3.80255073e-01 3.58632386e-01
9.92819428e-01 3.76810163e-01 8.15217853e-01 5.03627837e-01
3.85600239e-01 7.12266386e-01 3.36036980e-01 -2.41053000e-01
-5.65161407e-01 7.66542368e-03 4.25709873e-01 7.01136351e-01
-2.18483776e-01 -3.81869674e-01 -7.41565168e-01 8.12429309e-01
-2.07577205e+00 -9.63797927e-01 -3.42786938e-01 1.98872364e+00
9.04189348e-01 2.30588257e-01 3.53377759e-01 -6.68710023e-02
8.84890735e-01 4.21469241e-01 -6.61339223e-01 -1.11441210e-01
-1.27285607e-02 -3.26349363e-02 5.61710536e-01 5.78296125e-01
-1.20776725e+00 1.17192090e+00 5.66283655e+00 8.59042764e-01
-1.05927908e+00 2.95322865e-01 7.49654651e-01 -3.59682769e-01
-1.03544546e-02 4.70863134e-02 -8.62657309e-01 5.21302342e-01
5.91656089e-01 8.78857225e-02 3.72534096e-01 6.65605426e-01
6.88741624e-01 -5.58996737e-01 -8.44667614e-01 9.20651615e-01
6.25414923e-02 -1.38194680e+00 -7.01312721e-02 -3.69837850e-01
8.98406029e-01 2.18334213e-01 -3.74949157e-01 -3.20144325e-01
2.43685931e-01 -3.84997666e-01 9.03550863e-01 2.32052863e-01
6.87827289e-01 -5.01747608e-01 3.38993669e-01 1.67891830e-01
-1.52725554e+00 2.59914041e-01 -1.49017751e-01 1.76541924e-01
5.65175831e-01 5.85780263e-01 -7.19805300e-01 4.11383808e-01
8.53781104e-01 9.48349893e-01 -5.26945114e-01 1.09417045e+00
-1.82239577e-01 8.02266657e-01 -5.10314524e-01 2.77878612e-01
3.77660483e-01 -2.75809437e-01 7.05108404e-01 1.51819301e+00
6.34286851e-02 9.62531641e-02 5.32238603e-01 5.66523135e-01
1.00793429e-01 3.01883984e-02 -2.56254196e-01 1.64075762e-01
2.75160819e-01 1.25181448e+00 -1.42868125e+00 -6.36816323e-01
-4.13737804e-01 1.08711207e+00 7.89456591e-02 4.89141405e-01
-1.02937841e+00 -1.62724584e-01 5.91131508e-01 2.87462294e-01
5.50228894e-01 -2.93997675e-01 -5.80375195e-02 -1.12110853e+00
1.31733185e-02 -5.45956612e-01 4.14102465e-01 -5.38514137e-01
-7.20372736e-01 4.88526255e-01 1.34891942e-01 -1.14675117e+00
-2.80789416e-02 -1.95350319e-01 -6.28646135e-01 2.33942613e-01
-1.67041719e+00 -1.00413108e+00 -4.35498387e-01 7.52032399e-01
1.14028168e+00 3.07569623e-01 3.59962136e-03 4.70443934e-01
-7.54018128e-01 3.59327346e-01 -2.02306673e-01 1.98239982e-01
6.12562597e-01 -1.13988137e+00 4.41296041e-01 1.11086011e+00
2.09806785e-01 1.64826885e-01 6.21006012e-01 -7.05208480e-01
-1.08588850e+00 -1.13989210e+00 6.60902023e-01 -1.13636423e-02
5.55780292e-01 -1.32532209e-01 -9.34846222e-01 5.48958004e-01
1.58913031e-01 4.81866300e-01 2.67477065e-01 -4.06052589e-01
-6.66759070e-03 -3.59378867e-02 -1.12406385e+00 7.14947581e-01
1.30821824e+00 -1.22223966e-01 -6.49689734e-02 3.33728671e-01
7.39229023e-01 -7.12700367e-01 -7.14248121e-01 3.40732962e-01
5.82937956e-01 -1.05633569e+00 8.43963385e-01 -3.86509657e-01
2.23731652e-01 -7.07511246e-01 1.87315032e-01 -8.26090872e-01
-3.12040653e-02 -1.09159338e+00 -4.06053543e-01 1.27295804e+00
2.02909663e-01 -4.07884657e-01 1.01013827e+00 6.17807090e-01
-7.66365277e-03 -8.71345639e-01 -8.73324871e-01 -8.24496925e-01
-6.76451087e-01 -5.09071231e-01 3.26280445e-01 8.79041910e-01
-3.82584870e-01 -3.24436367e-01 -3.56164902e-01 2.22167224e-01
7.96983898e-01 1.84239239e-01 8.51604402e-01 -1.03443766e+00
-3.56595367e-01 -2.33884200e-01 -3.37071270e-01 -1.30030215e+00
2.00603172e-01 -6.43346667e-01 1.79108143e-01 -1.43967795e+00
1.35471806e-01 -4.79546547e-01 -1.66867167e-01 6.28597736e-01
-3.08268338e-01 5.63694358e-01 3.83255780e-01 2.07066834e-01
-1.03923512e+00 1.98379457e-01 1.37554073e+00 1.69970689e-03
-5.52854896e-01 1.15564287e-01 -1.17499575e-01 1.03364432e+00
9.70481396e-01 -6.81659520e-01 -4.47029740e-01 -3.27619195e-01
-2.09478676e-01 1.19676463e-01 3.03766489e-01 -1.01163733e+00
2.23709866e-01 -3.00246835e-01 1.96678534e-01 -5.75879753e-01
1.78632841e-01 -6.55424356e-01 2.22141922e-01 4.75353986e-01
-7.55314603e-02 -9.31261294e-03 2.20480323e-01 6.13527596e-01
-1.12676650e-01 -3.65203261e-01 9.31757927e-01 -1.99204221e-01
-1.04294300e+00 6.58607662e-01 -3.37307900e-01 2.97279298e-01
1.18798935e+00 -5.30832767e-01 9.80282500e-02 -2.96643913e-01
-8.87587190e-01 4.47594255e-01 5.31769574e-01 4.45765555e-01
5.61495602e-01 -1.01583219e+00 -6.19461715e-01 2.86478680e-02
-3.14844280e-01 2.14900404e-01 2.75910914e-01 1.12238109e+00
-7.30358183e-01 -5.05548306e-02 2.03268770e-02 -9.51311827e-01
-1.68910718e+00 2.62767345e-01 2.03679308e-01 -9.82339606e-02
-7.82682300e-01 9.00009036e-01 5.81852198e-02 1.56062692e-01
2.58683413e-01 -4.14412260e-01 -5.54611273e-02 1.30761385e-01
4.57227081e-01 5.35087228e-01 -2.07991704e-01 -7.14574933e-01
-3.49016339e-01 6.87714934e-01 -1.09792188e-01 -5.06973863e-02
1.21573687e+00 -4.44038063e-01 3.92804807e-03 5.20989478e-01
1.12224066e+00 -1.76090777e-01 -1.92169976e+00 -2.98497975e-01
-2.86985673e-02 -7.41267920e-01 3.13292556e-02 -2.75282025e-01
-1.49248779e+00 5.03540099e-01 5.21334708e-01 -9.31912661e-03
1.05038500e+00 9.03245732e-02 9.67575669e-01 1.07414976e-01
2.65633076e-01 -1.21564305e+00 -9.34173688e-02 1.85047895e-01
3.71358573e-01 -1.31253636e+00 2.02184066e-01 -7.04717755e-01
-6.68516457e-01 1.11885273e+00 6.81084752e-01 -5.53380623e-02
7.10114419e-01 3.86252970e-01 1.06969483e-01 -1.00088008e-01
-5.39057076e-01 -3.04026037e-01 2.06602812e-01 3.89469773e-01
1.81958660e-01 -2.49407724e-01 -2.84878582e-01 -1.44400656e-01
2.57981777e-01 1.87611744e-01 5.40543556e-01 9.93512034e-01
-3.12731266e-01 -9.91686344e-01 -1.77342042e-01 4.86460656e-01
-6.28256261e-01 7.28219971e-02 -2.46892646e-02 7.33735740e-01
2.19548017e-01 8.78779650e-01 6.79515451e-02 -1.01440283e-03
-2.11368012e-03 -5.13281040e-02 4.46489662e-01 -4.07750905e-01
-4.79915947e-01 3.76048446e-01 9.76372883e-02 -1.03345191e+00
-1.10417449e+00 -1.06809211e+00 -1.49927688e+00 -2.54266649e-01
-4.57014948e-01 -1.04757033e-01 5.10241985e-01 8.19329083e-01
3.88886631e-01 5.42784572e-01 4.18439180e-01 -1.13893282e+00
5.35617443e-03 -5.29683590e-01 -3.46102744e-01 4.41687584e-01
3.86026293e-01 -5.31763315e-01 -3.12104404e-01 5.24119973e-01] | [9.18741226196289, -0.1111103817820549] |
36bac4ff-13e3-4acc-bc37-f1e738835059 | parallel-chinese-english-entities-relations | null | null | https://aclanthology.org/L16-1589 | https://aclanthology.org/L16-1589.pdf | Parallel Chinese-English Entities, Relations and Events Corpora | This paper introduces the parallel Chinese-English Entities, Relations and Events (ERE) corpora developed by Linguistic Data Consortium under the DARPA Deep Exploration and Filtering of Text (DEFT) Program. Original Chinese newswire and discussion forum documents are annotated for two versions of the ERE task. The texts are manually translated into English and then annotated for the same ERE tasks on the English translation, resulting in a rich parallel resource that has utility for performers within the DEFT program, for participants in NIST{'}s Knowledge Base Population evaluations, and for cross-language projection research more generally. | ['Justin Mott', 'Zhiyi Song', 'Ann Bies', 'Stephanie Strassel'] | 2016-05-01 | parallel-chinese-english-entities-relations-1 | https://aclanthology.org/L16-1589 | https://aclanthology.org/L16-1589.pdf | lrec-2016-5 | ['knowledge-base-population'] | ['natural-language-processing'] | [ 1.55420229e-01 5.67603409e-01 -2.99366057e-01 -8.11920285e-01
-1.31566226e+00 -8.03005099e-01 8.37202907e-01 2.27251932e-01
-1.18579412e+00 1.03207755e+00 1.16680741e+00 -4.99729902e-01
6.68780953e-02 -5.55260718e-01 -3.61090958e-01 5.66584058e-02
2.32408762e-01 1.16388905e+00 1.54287577e-01 -4.62969929e-01
-2.81546861e-02 2.77945817e-01 -7.51981139e-01 7.58625805e-01
8.06914628e-01 5.11214495e-01 -5.22188507e-02 2.41314292e-01
-6.20155521e-02 5.78002989e-01 -5.27089238e-01 -1.23199093e+00
3.20583224e-01 -1.08019441e-01 -1.32984090e+00 -5.35057127e-01
2.36768380e-01 2.47630656e-01 -3.17794412e-01 1.17879593e+00
5.99682033e-01 2.79797852e-01 4.29298341e-01 -6.15354061e-01
-8.01255465e-01 1.43793416e+00 -3.48302811e-01 5.05354702e-01
6.35599613e-01 -5.85030794e-01 1.41849756e+00 -1.45698690e+00
1.45123327e+00 1.32069743e+00 6.94876015e-01 4.41460371e-01
-8.97920907e-01 -9.14688706e-01 9.98197868e-02 2.13414878e-01
-1.41093802e+00 -6.48278177e-01 1.31960839e-01 -1.41674504e-01
1.87343073e+00 3.29116166e-01 4.13840383e-01 1.43842554e+00
5.13875149e-02 9.93629217e-01 7.82391965e-01 -8.07453394e-01
-1.79846793e-01 3.95252138e-01 5.03596008e-01 2.78882653e-01
2.57543623e-01 9.42706615e-02 -7.88005114e-01 -2.54300684e-01
1.18528850e-01 -8.77817094e-01 -2.25369200e-01 2.74683654e-01
-1.46912289e+00 6.09504044e-01 4.83849943e-02 3.35617810e-01
-5.61319292e-01 -3.10142368e-01 5.25354683e-01 4.28981245e-01
7.32804716e-01 5.68882406e-01 -7.83641934e-01 -5.18187582e-01
-4.95382279e-01 5.84664226e-01 1.16675687e+00 1.33046591e+00
3.50050658e-01 -5.13795018e-01 -2.71828651e-01 9.82131600e-01
4.17476356e-01 3.38603050e-01 5.33120275e-01 -5.25836170e-01
1.44481730e+00 5.73336422e-01 3.27037722e-01 -6.23480439e-01
-6.42189085e-01 -2.35499889e-01 -1.30781695e-01 -4.98649865e-01
3.59113157e-01 -6.99600339e-01 -5.74650586e-01 1.67608643e+00
3.31261009e-01 -5.76974034e-01 6.78723395e-01 5.39068758e-01
8.71660233e-01 6.75662637e-01 5.15435338e-01 -3.76937613e-02
1.49540937e+00 -6.99170053e-01 -1.09556270e+00 -4.29088950e-01
8.05937588e-01 -1.12999976e+00 8.65983963e-01 1.33612141e-01
-1.08012879e+00 -4.62968767e-01 -8.74954104e-01 -3.24451715e-01
-8.42453241e-01 4.25935626e-01 7.79584110e-01 8.75647068e-01
-7.55359232e-01 1.28306374e-01 -9.26239192e-01 -7.06355631e-01
2.63483912e-01 3.72091472e-01 -4.45100486e-01 -1.27682626e-01
-1.95568681e+00 1.51759636e+00 8.64021420e-01 2.80333072e-01
-1.82485819e-01 -5.27969241e-01 -8.73098016e-01 -2.07253650e-01
8.30274820e-02 -2.71941423e-01 1.41491878e+00 -4.17004734e-01
-9.68681514e-01 1.01587808e+00 -2.10125983e-01 -6.34313107e-01
3.86412799e-01 -7.56330729e-01 -1.19566846e+00 -3.76172096e-01
5.41651189e-01 3.94374520e-01 -2.99517840e-01 -4.66622412e-01
-1.08773971e+00 -4.54902828e-01 -1.85076930e-02 7.63023913e-01
-3.44194472e-01 9.79802251e-01 -7.58431077e-01 -8.08560967e-01
-2.00597495e-01 -1.04990327e+00 -2.00365186e-01 -9.43942249e-01
-8.59360516e-01 -6.15732908e-01 5.98129928e-01 -1.24570596e+00
1.35329521e+00 -1.93599355e+00 -1.02836544e-04 3.15363020e-01
-1.86704934e-01 1.68509260e-01 3.07654850e-02 6.94627404e-01
-2.24586368e-01 3.00948650e-01 2.49308228e-01 -2.02478394e-01
1.20495185e-01 7.67450109e-02 -4.64459628e-01 2.78128564e-01
-2.77614668e-02 9.15526211e-01 -9.47685778e-01 -2.45358571e-01
-2.08127961e-01 2.33829215e-01 -2.68280298e-01 -3.85792673e-01
6.48041219e-02 2.21930236e-01 -8.13197672e-01 3.71310204e-01
7.10209459e-02 3.36950868e-01 4.17302608e-01 -2.89381564e-01
-3.83986980e-01 1.15305436e+00 -1.03064489e+00 1.56306827e+00
-2.82485843e-01 8.04405689e-01 -1.95183411e-01 -4.23241228e-01
6.78910136e-01 7.75394320e-01 1.73024192e-01 -7.17813551e-01
1.48753569e-01 4.57572103e-01 6.94423616e-02 -4.77624804e-01
1.16275406e+00 -9.66513678e-02 -4.78633910e-01 5.04414737e-01
2.78459609e-01 3.09116006e-01 5.75850844e-01 2.61460960e-01
1.02360880e+00 2.85571963e-01 3.93476456e-01 -3.14425975e-01
3.09618741e-01 6.19587600e-01 7.36125648e-01 5.01886368e-01
1.26311406e-01 1.85685888e-01 -1.10490814e-01 -3.23999077e-01
-1.02109182e+00 -9.64277506e-01 -4.88122255e-01 1.03642309e+00
-1.29343197e-01 -7.30799615e-01 -5.74227989e-01 -9.12703633e-01
-2.77892172e-01 1.36038029e+00 -4.55648452e-01 1.64549991e-01
-9.23179626e-01 -8.83234620e-01 1.10241473e+00 7.62871981e-01
5.75444221e-01 -1.10882294e+00 -1.50719061e-01 5.62925339e-01
-8.06650281e-01 -1.48648608e+00 -6.22878790e-01 2.47526243e-01
-1.73528910e-01 -8.68588150e-01 -4.13824379e-01 -1.02296388e+00
3.52976918e-02 -4.47779745e-01 1.17594922e+00 -8.43032300e-01
5.33319972e-02 2.67309099e-01 -4.91934508e-01 -8.36114168e-01
-6.06803358e-01 2.28170112e-01 1.69899598e-01 -3.99303973e-01
1.32432270e+00 -9.40803140e-02 1.52426213e-01 2.89342016e-01
-2.48611033e-01 6.38200939e-02 4.73407388e-01 8.66333485e-01
5.36120832e-01 6.87171370e-02 6.51961744e-01 -1.52961314e+00
1.03320122e+00 -3.82089734e-01 -5.45853317e-01 5.91418087e-01
-5.53782523e-01 -1.35390028e-01 6.70105591e-02 -9.56621096e-02
-2.15323782e+00 -1.13989197e-01 -3.55815351e-01 5.55072069e-01
-1.90576568e-01 9.78111506e-01 -5.48202634e-01 4.29149896e-01
1.06384039e+00 -2.52049595e-01 -1.01737130e+00 -4.44903910e-01
5.73641241e-01 1.09230316e+00 7.84739912e-01 -6.62676334e-01
6.44146085e-01 3.84190828e-02 -8.06690753e-01 -7.44287133e-01
-9.85062599e-01 -6.07350826e-01 -7.03214049e-01 1.47968844e-01
1.10674071e+00 -1.26228607e+00 7.73012114e-04 1.57174230e-01
-1.37822664e+00 2.67445836e-02 -4.95289743e-01 9.02833462e-01
-3.50778662e-02 -1.53049836e-02 -7.57062137e-01 -4.73311692e-01
-3.49616051e-01 -7.78081059e-01 9.23667789e-01 1.56870797e-01
-7.32209563e-01 -1.06160343e+00 4.90490288e-01 5.56626201e-01
6.25972971e-02 -2.49165520e-01 6.73583925e-01 -1.31963634e+00
-2.61103678e-02 -5.45575798e-01 -8.33563879e-02 8.99707600e-02
-1.55872568e-01 -3.80505294e-01 -7.40294933e-01 -5.37652336e-02
-2.88317889e-01 -4.43052351e-01 5.00712991e-01 2.44687889e-02
2.85608232e-01 -2.39185393e-01 -6.25524223e-01 1.30393043e-01
1.04977202e+00 4.59926844e-01 5.98707676e-01 3.02754372e-01
4.59650218e-01 8.86375725e-01 7.02651441e-01 -6.47831708e-02
6.27532601e-01 8.61681640e-01 -7.36311257e-01 1.29658118e-01
-1.16884403e-01 -3.65129948e-01 5.20103872e-01 9.31189597e-01
-2.89802611e-01 -4.19313133e-01 -1.17386091e+00 8.10760379e-01
-1.73152304e+00 -1.03335428e+00 7.75049906e-03 1.63020504e+00
1.25714505e+00 3.00381690e-01 -4.34763759e-01 -4.94179130e-01
7.00227141e-01 -2.49021068e-01 -9.27633271e-02 -3.08954924e-01
-5.09004653e-01 6.19373918e-01 6.00686729e-01 3.94277364e-01
-1.11901450e+00 1.28820217e+00 6.21399784e+00 8.71353090e-01
-4.05944943e-01 4.25727904e-01 2.72582173e-01 2.72782534e-01
-3.01983535e-01 2.42446706e-01 -1.58159256e+00 -8.52340534e-02
1.45101070e+00 -9.99178410e-01 1.49637744e-01 6.14205003e-01
-7.46072754e-02 3.08158174e-02 -1.21128225e+00 6.43659174e-01
-1.42933160e-01 -1.63650918e+00 -2.16427267e-01 1.29118443e-01
7.39380658e-01 9.77442741e-01 -2.18588874e-01 7.81936049e-01
8.66152346e-01 -6.25766039e-01 8.54948461e-01 1.40278593e-01
8.74469042e-01 -6.04788661e-01 9.76411402e-01 6.62440956e-02
-1.17524600e+00 1.04366422e-01 -3.26353163e-01 3.73634040e-01
6.97514772e-01 2.69950926e-01 -9.44221020e-01 8.23357642e-01
7.81057060e-01 5.05010486e-01 -5.27256429e-01 6.70327365e-01
-3.02847207e-01 8.51907670e-01 -4.53577638e-01 -1.66392431e-01
1.44256309e-01 -2.19365414e-02 9.22183275e-01 1.71918929e+00
9.91702452e-02 3.16440493e-01 1.82555154e-01 6.44642234e-01
-5.92940509e-01 3.94883096e-01 -4.21223998e-01 -4.72335935e-01
7.86580086e-01 1.09824383e+00 -6.31363392e-01 -3.51408392e-01
-8.30530107e-01 8.26549411e-01 3.11209351e-01 5.47496140e-01
-4.26529944e-01 -7.47327805e-01 3.87136102e-01 -3.69946808e-01
2.60371938e-02 5.91720156e-02 -1.04442134e-01 -1.36287045e+00
3.03791054e-02 -7.58273661e-01 6.09416902e-01 -6.50819421e-01
-1.56216812e+00 1.07627809e+00 1.85104117e-01 -8.69105816e-01
-5.20232797e-01 -7.14615107e-01 -2.68941224e-01 1.33531797e+00
-8.09300005e-01 -1.30672991e+00 3.15498531e-01 7.31323481e-01
6.99750006e-01 -5.67153811e-01 1.16510320e+00 7.46978819e-01
-4.45401162e-01 6.89555645e-01 -8.45743865e-02 6.91977382e-01
7.87958622e-01 -1.34084976e+00 9.43225920e-01 1.09678686e+00
5.99860430e-01 1.02288854e+00 6.78940654e-01 -1.15669119e+00
-9.34061468e-01 -1.25286341e+00 1.67609632e+00 -1.06613386e+00
9.69608963e-01 -4.98525798e-01 -4.63779181e-01 1.37376642e+00
5.61021328e-01 -5.57820261e-01 8.83719027e-01 5.88980198e-01
-1.81052342e-01 3.22987586e-01 -8.68476093e-01 6.85972750e-01
8.89923632e-01 -9.15655315e-01 -1.24871147e+00 6.80272758e-01
7.70832598e-01 -8.04533958e-01 -1.07912719e+00 2.59993464e-01
4.85522389e-01 -4.56425808e-02 7.50451684e-01 -8.38115990e-01
-2.25117411e-02 -1.84765249e-01 -5.57635128e-01 -1.38150048e+00
-6.05414249e-02 -6.24750316e-01 3.24994922e-01 1.44571996e+00
1.39843082e+00 -3.97334337e-01 5.31416059e-01 7.86756158e-01
-4.32178974e-01 -2.63819069e-01 -1.22822988e+00 -5.30461669e-01
1.08949900e-01 -8.82871509e-01 3.72181982e-01 1.20333350e+00
3.98709804e-01 1.04526901e+00 -1.05527334e-01 2.47565746e-01
4.05166805e-01 -2.81697422e-01 3.62273723e-01 -1.05991244e+00
-2.13565767e-01 -2.18882576e-01 -7.32565671e-02 -7.64538109e-01
8.49070325e-02 -1.24212754e+00 -5.37019260e-02 -1.50956249e+00
1.31722301e-01 -4.20768499e-01 1.06032073e-01 5.74243367e-01
-3.93604487e-02 -1.94245037e-02 -9.68103036e-02 2.62576491e-01
-3.80459279e-01 5.12483597e-01 6.27201974e-01 4.42002621e-03
-3.16234440e-01 9.90699157e-02 -9.02483761e-01 6.19241416e-01
3.73562664e-01 -4.45421994e-01 -1.22026958e-01 -6.04900062e-01
5.08282423e-01 1.63783552e-04 -2.61017829e-01 -7.27152169e-01
3.45414251e-01 1.12878129e-01 5.74811637e-01 -9.57282066e-01
1.48311228e-01 -6.28185511e-01 3.27324331e-01 3.62955295e-02
-6.69664562e-01 3.86856586e-01 5.03414035e-01 3.76582950e-01
-4.04918522e-01 -1.33353740e-01 2.88332671e-01 -1.17673106e-01
-7.41466820e-01 -5.83023727e-02 -5.03484309e-01 3.58287156e-01
6.38002932e-01 -1.31916869e-02 -4.87078875e-01 -2.62259901e-01
-1.12112248e+00 5.05748570e-01 -2.58125216e-01 7.38833010e-01
2.31830433e-01 -1.37660766e+00 -1.10719073e+00 -1.14508579e-02
3.15055132e-01 -4.15654033e-01 -3.33042368e-02 7.42059529e-01
-2.14552000e-01 7.13921666e-01 -3.76240276e-02 1.08950123e-01
-8.77124727e-01 8.31322595e-02 1.43076763e-01 -7.32400179e-01
-5.96005976e-01 9.20222223e-01 -2.09710717e-01 -8.60503018e-01
1.30589515e-01 -6.12338670e-02 -4.51785505e-01 1.50362134e-01
8.72646034e-01 3.21186185e-01 2.80550599e-01 -9.20014977e-01
-2.76600003e-01 -2.47705936e-01 -5.70734024e-01 -8.22109044e-01
1.39405608e+00 -2.94954091e-01 -5.82966842e-02 3.35655719e-01
8.95737827e-01 7.03552306e-01 -2.58628547e-01 -5.39554477e-01
8.53772521e-01 2.02854440e-01 -4.81748916e-02 -1.30125117e+00
-5.10234356e-01 3.03360045e-01 3.66863400e-01 -1.77107230e-01
6.65872097e-01 1.77502930e-01 7.11776376e-01 8.72683823e-01
5.57572424e-01 -1.52424014e+00 -8.86051834e-01 8.05657029e-01
9.47276115e-01 -7.54027009e-01 -1.88566059e-01 -5.21088719e-01
-8.82884860e-01 1.04172599e+00 5.21030307e-01 3.49302649e-01
8.84564936e-01 4.30467129e-01 1.40819103e-01 -3.96536827e-01
-6.45272076e-01 3.77002060e-02 4.32488829e-01 7.04469442e-01
8.52762699e-01 -2.45756935e-02 -6.73224330e-01 1.00001681e+00
-5.41208208e-01 -9.10291895e-02 2.25877300e-01 7.33938694e-01
-1.19230248e-01 -1.36835277e+00 1.15332842e-01 3.39837343e-01
-6.75876558e-01 -6.12430155e-01 -6.27699494e-01 1.05801392e+00
1.62040874e-01 8.47022414e-01 -1.14335887e-01 -3.53747517e-01
8.88442457e-01 2.75509894e-01 1.83721304e-01 -1.10867691e+00
-9.43557501e-01 1.13448955e-01 1.35274136e+00 -5.61063230e-01
-6.10825419e-01 -1.17863357e+00 -1.29695451e+00 7.96451867e-02
-3.57110739e-01 6.25649452e-01 5.45314491e-01 1.14579093e+00
7.35807717e-02 2.40596727e-01 -6.48318231e-02 -3.08718204e-01
-2.20259160e-01 -1.45438111e+00 -4.33559269e-01 3.52847129e-01
-5.65251946e-01 -4.11343932e-01 1.97179750e-01 9.88304615e-02] | [9.618907928466797, 9.37723159790039] |
e4f7a41f-b15a-4e2d-b4d9-0a5bfd24b87a | on-the-anatomy-of-latent-variable-generative | null | null | https://openreview.net/forum?id=29mx8VNszc | https://openreview.net/pdf?id=29mx8VNszc | On the Anatomy of Latent-variable Generative Models for Conditional Text Generation | Conditional text generation is a non-trivial task, which is until now predominantly performed with latent-variable generative models. In this work, we intend to explore several choices that are shown to affect the two essential aspects of model performance: expressivity and controllability. We propose to experiment with a series of latent-variable models built around simple design changes under a general unified framework, with a particular focus on prior distributions based on Energy-Based Models instead of the usual standard Gaussian. Our experiments validate the claim that this richer prior allows for a better representational power, but it exhibits difficult training. We provide a comprehensive analysis of these difficulties and a close comparison with recent work on EBM-based priors for conditional text generation. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['conditional-text-generation'] | ['natural-language-processing'] | [ 4.90492046e-01 3.17928702e-01 -9.02973861e-02 6.64265677e-02
-6.09457374e-01 -5.77613592e-01 1.19943368e+00 -3.51717085e-01
-3.71085145e-02 1.01272237e+00 2.95151651e-01 -4.03465003e-01
-1.95922047e-01 -7.47901499e-01 -4.19601172e-01 -8.15696716e-01
2.98953921e-01 5.53913593e-01 -1.31957397e-01 -2.03665793e-01
1.48009330e-01 3.76914144e-01 -1.50271678e+00 -1.28971696e-01
8.95096481e-01 4.64905679e-01 2.63633579e-01 7.72559702e-01
1.23160012e-01 4.40737695e-01 -6.62548423e-01 -5.83884895e-01
-2.86554605e-01 -6.08914435e-01 -7.40922809e-01 -4.44999449e-02
-2.24658757e-01 -1.21920086e-01 2.30132088e-01 7.46461034e-01
6.07997537e-01 1.28928542e-01 1.17149949e+00 -1.19327426e+00
-7.12211192e-01 6.52235806e-01 -1.75470293e-01 -5.96294552e-02
3.90779167e-01 2.66371310e-01 1.12225151e+00 -5.66617072e-01
7.07748950e-01 1.34986234e+00 7.04967678e-01 7.39831388e-01
-1.76487625e+00 -3.20733070e-01 3.89731884e-01 -1.63635537e-01
-1.27822566e+00 -6.44337356e-01 9.32344615e-01 -4.81527418e-01
9.19242978e-01 3.92940789e-01 4.52746242e-01 1.74599195e+00
4.74741697e-01 5.03252029e-01 1.16475892e+00 -7.98702478e-01
4.63108391e-01 4.66518521e-01 -9.88749787e-02 3.28712463e-01
3.63494337e-01 1.89636916e-01 -4.87096608e-01 -2.84054101e-01
7.98581660e-01 -4.38677549e-01 -2.95726091e-01 -7.01952696e-01
-1.03492963e+00 1.18840909e+00 -1.79237396e-01 5.43795824e-01
-3.22782993e-01 5.04315317e-01 -1.95962012e-01 7.94976652e-02
6.56026244e-01 5.98566949e-01 -3.35408121e-01 -3.56471747e-01
-1.06832778e+00 4.08676028e-01 9.68323171e-01 1.00688910e+00
5.89297891e-01 2.65789688e-01 -6.23525143e-01 4.16680515e-01
7.70257831e-01 4.45163369e-01 2.58269101e-01 -7.64647305e-01
1.46024257e-01 -4.35631201e-02 4.00469691e-01 -6.59589887e-01
-2.06534341e-01 -4.61496651e-01 -6.23561263e-01 2.70635754e-01
4.56589282e-01 -5.66947401e-01 -9.35618281e-01 2.00009322e+00
5.82965352e-02 -2.94750214e-01 -2.03828841e-01 4.56388712e-01
3.98939997e-01 5.36883175e-01 3.46484542e-01 -5.02232432e-01
1.11016369e+00 -6.05774879e-01 -1.07060766e+00 7.65642058e-03
1.89960748e-01 -9.38176394e-01 1.09880280e+00 3.33265007e-01
-1.47163939e+00 -3.88856292e-01 -8.77869487e-01 9.02743340e-02
-3.10053408e-01 8.75085369e-02 9.88236845e-01 1.19754243e+00
-1.19031203e+00 6.83146417e-01 -9.64423835e-01 -3.69627446e-01
-3.17556299e-02 3.67097259e-02 1.54947698e-01 4.27328497e-01
-1.23566532e+00 1.03273594e+00 -2.62424815e-02 -1.57805026e-01
-1.03360438e+00 -5.33443809e-01 -8.07129145e-01 1.77910611e-01
2.10946515e-01 -1.18694270e+00 1.46467924e+00 -6.70038402e-01
-2.23303390e+00 5.42780340e-01 -3.18259150e-01 -3.23463827e-01
6.94998622e-01 -1.87246189e-01 -1.80677861e-01 -1.27593100e-01
-1.51019424e-01 5.95714092e-01 1.01026011e+00 -1.52988636e+00
1.18187733e-01 9.05018002e-02 6.36614934e-02 -7.80886412e-03
-1.90607250e-01 7.15764686e-02 -2.03844860e-01 -8.59319031e-01
-2.55145580e-01 -8.93498361e-01 -3.22322726e-01 -3.64148259e-01
-6.14085674e-01 -1.69337854e-01 3.28019142e-01 -2.74358988e-01
1.49958086e+00 -1.57560861e+00 5.78126669e-01 8.66936594e-02
-2.22599600e-02 -1.28670007e-01 1.17604762e-01 7.47082055e-01
-3.29363614e-01 5.45732260e-01 -1.80939078e-01 -8.82973433e-01
5.16031921e-01 4.00486477e-02 -5.67605138e-01 3.02641243e-01
2.48650804e-01 1.01680112e+00 -4.90927815e-01 -5.33898473e-01
2.37725243e-01 8.82839382e-01 -6.46768451e-01 1.30797431e-01
-6.31072581e-01 5.23413837e-01 -5.13797522e-01 3.41600537e-01
4.00413871e-01 -3.16140383e-01 2.51175046e-01 1.43141478e-01
-2.62516916e-01 4.30605263e-01 -1.18708909e+00 1.64575994e+00
-5.24754345e-01 5.77054083e-01 -1.20767308e-02 -6.20296299e-01
5.19183099e-01 5.31897008e-01 3.29682499e-01 -2.73395807e-01
2.48484418e-01 -1.85968891e-01 -2.45139524e-01 -2.49655500e-01
6.29690170e-01 -6.95485234e-01 -1.06534615e-01 5.97655714e-01
2.59337753e-01 -5.58947563e-01 2.17647925e-01 1.88587353e-01
5.99215865e-01 6.13561332e-01 3.25779647e-01 -5.92011452e-01
7.02309534e-02 -3.49883914e-01 1.86746806e-01 8.45916152e-01
2.36070350e-01 5.98653376e-01 7.04813480e-01 2.01348618e-01
-1.03661537e+00 -1.22071707e+00 -4.38620359e-01 7.57790864e-01
-1.39735892e-01 -7.73725927e-01 -8.22200000e-01 -3.29846174e-01
-3.74957472e-01 1.44991434e+00 -7.68194616e-01 -6.37812912e-03
-1.90664053e-01 -1.23801064e+00 2.59675562e-01 4.25115258e-01
-1.08790793e-01 -8.43024075e-01 -8.10979128e-01 2.17793196e-01
-1.25744879e-01 -6.31716311e-01 5.24339788e-02 4.85683382e-01
-6.92820430e-01 -3.59164983e-01 -9.45778608e-01 -2.50112474e-01
4.37369615e-01 -3.47225070e-01 1.50644386e+00 -2.39412338e-01
-1.06315441e-01 8.33314955e-01 -3.49951804e-01 -5.71259797e-01
-7.68971145e-01 1.41313747e-01 -2.29918316e-01 -2.20228791e-01
-2.11028196e-02 -7.47428358e-01 -4.59134966e-01 2.18713924e-01
-9.20755029e-01 2.61436611e-01 5.26955366e-01 8.27074885e-01
4.32967395e-01 1.15672097e-01 3.12046498e-01 -7.92817712e-01
1.02703273e+00 -2.58092493e-01 -5.55429041e-01 2.20916644e-01
-8.84278059e-01 4.41791385e-01 3.85927737e-01 -4.87316459e-01
-1.50027275e+00 -1.39916971e-01 -2.45433390e-01 -4.27806824e-02
-2.12008670e-01 5.51829219e-01 -3.82979482e-01 2.37413645e-01
5.78660131e-01 1.83874682e-01 -1.82089537e-01 -5.32784522e-01
6.53304160e-01 2.39327684e-01 -1.02671245e-02 -8.34043503e-01
9.45089757e-01 2.83553898e-01 1.25313848e-01 -6.43169343e-01
-3.30286980e-01 1.69057384e-01 -4.86881405e-01 -1.15415074e-01
1.05297685e+00 -6.54038846e-01 -3.03773999e-01 2.29712084e-01
-1.09266281e+00 -7.40969896e-01 -5.89252174e-01 3.58775109e-01
-8.56591105e-01 3.34186047e-01 -5.73775589e-01 -1.07851088e+00
-5.86378984e-02 -1.17061973e+00 1.22330630e+00 4.49688658e-02
-5.60604036e-01 -1.52100897e+00 4.37876523e-01 -5.21449633e-02
8.70634973e-01 2.54759490e-01 1.08137584e+00 -2.73415953e-01
-5.92608333e-01 6.16229661e-02 3.42351228e-01 -1.94003936e-02
3.79916318e-02 4.38165873e-01 -1.18179178e+00 -2.78987259e-01
2.64258832e-01 3.01133078e-02 8.75878334e-01 6.97048008e-01
8.67401004e-01 -2.94879600e-02 -5.97611606e-01 4.36971337e-01
1.37256503e+00 -1.05015583e-01 7.35838830e-01 -2.97475383e-02
3.40043873e-01 3.99914891e-01 1.77349582e-01 6.00896120e-01
6.97913766e-02 8.67968440e-01 3.07032377e-01 -2.75499783e-02
-9.50521752e-02 -3.46877426e-01 3.54871631e-01 5.66469193e-01
-2.15773478e-01 -8.31952095e-01 -7.11704671e-01 4.01783228e-01
-1.70650864e+00 -1.02638519e+00 -2.21829221e-01 2.02917695e+00
8.99539292e-01 1.52322859e-01 -1.11148261e-01 6.93746880e-02
6.08018339e-01 2.51767576e-01 -1.41379043e-01 -3.47436786e-01
4.59364653e-02 3.76208842e-01 3.48259449e-01 7.52229273e-01
-1.01409435e+00 6.13972902e-01 8.42762661e+00 1.02934039e+00
-8.75967562e-01 1.83729038e-01 6.15345359e-01 -1.32784054e-01
-1.06968331e+00 2.25936025e-01 -9.83079672e-01 5.70514321e-01
1.06239009e+00 -3.05662751e-01 3.26552361e-01 6.38517618e-01
2.94919193e-01 -5.68219721e-02 -1.23984754e+00 5.73988736e-01
9.73454341e-02 -1.18034303e+00 2.01267302e-01 1.06731758e-01
8.24425042e-01 -5.60203433e-01 3.95787835e-01 3.25066149e-01
4.96618390e-01 -1.32822192e+00 9.09457982e-01 7.28067160e-01
7.28323221e-01 -4.66173142e-01 2.66666621e-01 2.44048595e-01
-8.90482366e-01 3.13982636e-01 -3.21766846e-02 -1.87203214e-01
6.21707737e-01 7.36120641e-01 -5.38078606e-01 4.72665876e-01
2.97494143e-01 3.19849670e-01 -2.90065974e-01 7.38525689e-01
-6.06221378e-01 8.01839352e-01 -1.96712255e-01 -1.86655730e-01
-6.07781410e-02 -8.88330862e-02 8.84574234e-01 1.28220105e+00
4.69693631e-01 -3.23106796e-01 -3.49351764e-01 1.50386083e+00
3.96905452e-01 -1.09927788e-01 -5.79781532e-01 1.50998328e-02
1.48757756e-01 1.17710805e+00 -9.28445697e-01 -1.93724394e-01
-1.34332374e-01 8.28678370e-01 1.66440867e-02 5.19027174e-01
-1.21465397e+00 8.82309303e-03 3.79522622e-01 1.61483079e-01
4.62807655e-01 -4.34480965e-01 -4.56890583e-01 -1.27097142e+00
-3.76494020e-01 -7.14991450e-01 -6.10304140e-02 -9.83631492e-01
-1.17603290e+00 3.41163009e-01 5.92740715e-01 -8.08117747e-01
-8.67288470e-01 -5.36992192e-01 -8.12873602e-01 1.36893511e+00
-1.46328413e+00 -1.20635176e+00 8.80052075e-02 4.48670596e-01
3.09274435e-01 1.18379407e-01 1.13480818e+00 1.65131153e-03
-6.03066981e-01 6.06990755e-01 1.33745983e-01 -6.12415135e-01
5.32038748e-01 -1.50153756e+00 5.75690269e-01 9.19322193e-01
1.49067193e-01 1.01670504e+00 1.44485176e+00 -6.05777085e-01
-1.28817368e+00 -6.58841848e-01 8.90616238e-01 -9.40911710e-01
5.39642096e-01 -5.79521477e-01 -3.44170570e-01 8.48062336e-01
6.66851699e-01 -9.05435562e-01 7.76445866e-01 4.24730182e-01
1.51350945e-01 5.97801507e-01 -8.43717277e-01 8.17599893e-01
8.77470374e-01 -4.01449800e-01 -4.23440814e-01 3.07528347e-01
7.16254056e-01 -1.04579546e-01 -7.50270009e-01 3.07631195e-01
3.13359827e-01 -9.08006251e-01 9.17676508e-01 -3.46902519e-01
2.94618726e-01 -1.62379205e-01 -6.42249919e-03 -1.39102674e+00
-5.58692932e-01 -1.18270636e+00 -1.28094658e-01 1.51778030e+00
5.81132233e-01 -5.57104945e-01 6.04048610e-01 8.55110466e-01
5.78739159e-02 -5.74879229e-01 -8.67869139e-01 -6.95863247e-01
5.45963347e-01 -5.20169616e-01 4.23951775e-01 6.95091665e-01
-1.03522435e-01 7.37871766e-01 -4.28064913e-01 -2.11219087e-01
3.60972255e-01 1.85428172e-01 7.58572936e-01 -1.33987570e+00
-6.15399599e-01 -6.55952573e-01 6.57650903e-02 -1.28405333e+00
4.57169786e-02 -5.69478810e-01 1.28450722e-01 -1.55907989e+00
3.47795188e-01 -2.44060487e-01 5.19134365e-02 1.04468241e-01
-1.04010388e-01 -1.52147794e-02 -3.70633453e-02 -1.03676297e-01
-4.42662448e-01 9.26777661e-01 1.17431974e+00 1.26742944e-01
-2.16608837e-01 2.66920805e-01 -9.07644391e-01 6.09300077e-01
7.23155439e-01 -3.89824569e-01 -6.92257345e-01 -3.32253426e-01
4.70499486e-01 1.04163013e-01 5.60868323e-01 -7.27611780e-01
-6.49694772e-03 -2.30405703e-01 3.39674026e-01 -4.47353840e-01
6.02342963e-01 -6.21284306e-01 6.51868284e-01 1.79288864e-01
-3.64443153e-01 -4.34720740e-02 3.02663773e-01 5.73464513e-01
6.53663278e-02 -5.18738449e-01 5.91562271e-01 -1.37891397e-01
-1.51125118e-01 -8.55504945e-02 -7.51999378e-01 -2.23974362e-01
8.69582593e-01 -1.05528422e-01 -1.13900930e-01 -7.32099116e-01
-1.08434367e+00 -2.90525049e-01 5.54172873e-01 2.05954760e-01
1.06845476e-01 -1.19427657e+00 -5.40475190e-01 1.06420733e-01
-2.39360124e-01 -3.44101489e-01 2.00538695e-01 6.83535874e-01
-3.69046442e-02 6.43967748e-01 -3.97551544e-02 -5.75749874e-01
-9.99029517e-01 7.47758090e-01 2.23097429e-01 -7.10663021e-01
-3.88661563e-01 7.23801196e-01 2.25031763e-01 -8.97966996e-02
-6.85047265e-03 -3.90261680e-01 -1.39267698e-01 1.35194778e-01
-4.54462618e-02 1.83015719e-01 -2.29229346e-01 -3.81850839e-01
-1.27189651e-01 4.40471292e-01 2.83589780e-01 -5.93448162e-01
1.12011492e+00 -3.55776161e-01 1.42718688e-01 5.63882768e-01
6.85854554e-01 2.52375484e-01 -1.38184536e+00 2.47830123e-01
-3.98847371e-01 -2.03875378e-01 1.33197322e-01 -8.82470250e-01
-5.51481366e-01 1.02545476e+00 4.23961997e-01 5.65988302e-01
9.05738533e-01 -2.61736568e-02 1.76824570e-01 -2.63792556e-02
3.51159066e-01 -1.07294893e+00 7.95018822e-02 1.40029043e-01
9.24549162e-01 -7.18511999e-01 1.97327480e-01 -3.92438024e-01
-5.33843517e-01 8.76301169e-01 2.15951473e-01 2.79214263e-01
7.54745364e-01 3.55666637e-01 -4.29844081e-01 -5.99185750e-02
-9.68412399e-01 -1.67145714e-01 2.53239721e-01 7.00196862e-01
7.70683408e-01 7.86021203e-02 -3.96514595e-01 7.11009443e-01
-3.47418189e-01 -5.25536686e-02 4.06419426e-01 7.92958081e-01
-1.66819721e-01 -1.54636598e+00 -3.84640068e-01 3.30812261e-02
-5.34274518e-01 -2.30593428e-01 -2.68817335e-01 8.61720741e-01
-4.56314944e-02 1.18741953e+00 -1.03473417e-01 6.06347881e-02
4.87758691e-04 3.25566620e-01 7.08223045e-01 -4.90629107e-01
-3.60038042e-01 3.14587325e-01 1.16754360e-01 -1.79157570e-01
-5.87521195e-01 -9.59886551e-01 -5.61788917e-01 -2.20839724e-01
-6.44059181e-01 1.80158138e-01 6.94255173e-01 7.52711356e-01
2.48926476e-01 8.77170324e-01 1.35468364e-01 -1.09284842e+00
-6.57286286e-01 -1.07560241e+00 -4.58010435e-01 2.16160312e-01
6.99838176e-02 -9.26047027e-01 -5.31074226e-01 2.98023105e-01] | [6.993789196014404, 3.8926377296447754] |
4a0caed8-b961-4796-aba6-54c4b13f2709 | a-storytelling-robot-managing-persuasive-and | 2107.12845 | null | https://arxiv.org/abs/2107.12845v2 | https://arxiv.org/pdf/2107.12845v2.pdf | A Storytelling Robot managing Persuasive and Ethical Stances via ACT-R: an Exploratory Study | We present a storytelling robot, controlled via the ACT-R cognitive architecture, able to adopt different persuasive techniques and ethical stances while conversing about some topics concerning COVID-19. The main contribution of the paper consists in the proposal of a needs-driven model that guides and evaluates, during the dialogue, the use (if any) of persuasive techniques available in the agent procedural memory. The portfolio of persuasive techniques tested in such a model ranges from the use of storytelling, to framing techniques and rhetorical-based arguments. To the best of our knowledge, this represents the first attempt of building a persuasive agent able to integrate a mix of explicitly grounded cognitive assumptions about dialogue management, storytelling and persuasive techniques as well as ethical attitudes. The paper presents the results of an exploratory evaluation of the system on 63 participants | ['Antonio Lieto', 'Manuel Gentile', 'Giuseppe Città', 'Agnese Augello'] | 2021-07-27 | null | null | null | null | ['dialogue-management'] | ['natural-language-processing'] | [-2.39372849e-02 1.18320537e+00 1.50784656e-01 -6.02869749e-01
9.45187733e-02 -6.87109530e-01 1.33414447e+00 4.35979784e-01
-4.06272352e-01 1.04285097e+00 9.93893564e-01 -6.06565475e-01
-3.49436492e-01 -6.91630960e-01 -1.55376896e-01 -1.75625935e-01
1.66855142e-01 7.92826355e-01 1.68722019e-01 -1.03993261e+00
7.74436533e-01 2.46887073e-01 -1.38592768e+00 2.46281549e-01
6.59359217e-01 -5.36031555e-03 2.85044223e-01 7.43685484e-01
-1.83962300e-01 1.82041252e+00 -8.33988786e-01 -2.57917017e-01
-5.45750141e-01 -5.09192824e-01 -1.37256086e+00 -1.54760674e-01
-4.49078351e-01 -4.68063623e-01 4.25892591e-01 7.68324256e-01
4.35292035e-01 2.24708736e-01 6.60638988e-01 -9.49047685e-01
-4.57924634e-01 1.32492673e+00 4.26584631e-01 1.72228348e-02
1.13909388e+00 2.63997734e-01 4.70350146e-01 -2.91647375e-01
8.15264463e-01 1.70825052e+00 5.28066278e-01 9.14819479e-01
-9.42985594e-01 -3.49500030e-01 -2.91046854e-02 4.30619001e-01
-6.44900680e-01 -4.70085770e-01 1.06585217e+00 -7.06512809e-01
9.81415749e-01 4.30301100e-01 1.06307065e+00 1.45761323e+00
2.34124750e-01 3.36722404e-01 1.71668518e+00 -1.00713384e+00
6.92265213e-01 9.74069774e-01 7.08483994e-01 3.90020251e-01
7.46515393e-02 9.36066657e-02 -5.92843056e-01 -2.35067219e-01
3.52250189e-01 -1.00812888e+00 -1.39513329e-01 -2.39230692e-01
-8.87404323e-01 1.09601641e+00 -1.50168583e-01 9.20570612e-01
-5.64796627e-01 -1.86167493e-01 7.36705363e-01 3.29641163e-01
2.08099931e-01 8.10156822e-01 -1.83072537e-01 -7.49981344e-01
3.83744724e-02 4.39015090e-01 1.55186546e+00 4.09746021e-01
2.04320729e-01 2.30353829e-02 -2.19843745e-01 5.89446604e-01
9.91385102e-01 2.79221714e-01 2.39587650e-01 -1.16318703e+00
-1.90751433e-01 6.63509250e-01 4.21991050e-01 -1.11151719e+00
-5.96262336e-01 2.02519163e-01 1.29485130e-01 6.91384077e-01
1.62813514e-01 -5.90346038e-01 8.98204073e-02 1.54787958e+00
6.20424986e-01 -5.41196227e-01 4.93363827e-01 8.76861751e-01
1.20056653e+00 4.36869174e-01 6.43797874e-01 -4.54146296e-01
1.59206903e+00 -4.73720849e-01 -1.22910893e+00 -7.30360076e-02
6.36125743e-01 -6.74039721e-01 9.86212790e-01 3.44238311e-01
-1.16871119e+00 -2.28221253e-01 -1.26571870e+00 -9.38244835e-02
-2.79233903e-01 -3.58065188e-01 6.34426773e-01 1.07909739e+00
-1.09120071e+00 4.63503413e-02 -1.66530624e-01 -8.85497868e-01
-3.97481978e-01 -2.27344602e-01 -7.47926235e-02 5.22157669e-01
-1.62164724e+00 1.68461978e+00 8.04869056e-01 -1.05306968e-01
-4.94208157e-01 -2.44440928e-01 -7.16494739e-01 -1.76728472e-01
5.10157466e-01 -6.97106302e-01 1.61954713e+00 -1.10805559e+00
-2.36270046e+00 9.30663526e-01 6.09586418e-01 -6.70632839e-01
7.37731159e-01 -4.28740233e-01 -2.99688309e-01 2.22800151e-01
5.38897999e-02 4.81919795e-01 2.50355661e-01 -1.47254658e+00
-4.39439982e-01 -7.57885128e-02 6.59134984e-01 4.76506978e-01
9.57501382e-02 4.62312907e-01 5.37588954e-01 -2.83727407e-01
-8.30061316e-01 -5.89527667e-01 -6.59602582e-02 -4.73009616e-01
-1.55688345e-01 -4.18783575e-01 5.00031114e-01 -4.62180465e-01
1.14072919e+00 -1.72261393e+00 7.76631832e-02 -3.85061018e-02
2.19822764e-01 4.29546565e-01 4.54495877e-01 1.08976078e+00
1.59385279e-02 5.87456711e-02 3.50190133e-01 2.60383159e-01
5.09128153e-01 2.39188626e-01 -2.30111480e-01 1.09923994e-02
-3.76126856e-01 4.02953178e-01 -9.21092689e-01 -7.26801395e-01
7.65591145e-01 5.35547972e-01 -1.98174492e-01 1.95119426e-01
-4.68945205e-01 2.24241257e-01 -7.91361213e-01 -2.64778256e-01
2.95101434e-01 3.09424311e-01 6.77869260e-01 2.87395149e-01
-6.39983535e-01 9.00729895e-01 -7.29305863e-01 1.35901785e+00
-3.31890583e-01 5.09395003e-01 2.81659007e-01 -4.80198115e-01
9.62241530e-01 7.27693617e-01 -1.21843554e-01 -5.97571433e-01
6.99098229e-01 -1.57291181e-02 8.48123953e-02 -8.79643381e-01
4.00202602e-01 -3.35154563e-01 -9.40730646e-02 9.45283234e-01
-2.33092606e-01 -2.83317685e-01 1.99490096e-02 3.09139609e-01
7.81132162e-01 4.82081145e-01 1.01975834e+00 -6.52340710e-01
8.76908898e-01 3.59294534e-01 5.27703911e-02 6.13292158e-01
-1.70732141e-01 -6.24627709e-01 5.07525921e-01 -7.06333101e-01
-6.69986308e-01 -4.64812785e-01 3.13868105e-01 1.10701501e+00
-6.59333691e-02 -4.23708171e-01 -1.42008126e+00 -6.32816911e-01
-8.65139484e-01 1.99072897e+00 -5.40956318e-01 -6.64340705e-02
-4.90916848e-01 -2.48853222e-01 4.30805773e-01 -2.87443012e-01
8.59331071e-01 -1.61358869e+00 -1.78330743e+00 5.46289384e-01
-2.38493800e-01 -5.72928607e-01 4.71622199e-01 -1.84036251e-02
-5.76909721e-01 -9.76965785e-01 3.16612959e-01 -4.96033818e-01
1.18154198e-01 -6.39029816e-02 8.86547506e-01 2.93816984e-01
3.32155675e-01 6.57006443e-01 -7.83425689e-01 -9.28929210e-01
-1.18238044e+00 -2.64526248e-01 -4.78126526e-01 -5.69886267e-01
3.30378622e-01 -6.52189851e-01 -2.68967032e-01 1.71539322e-01
-5.25430381e-01 6.11831069e-01 1.50823385e-01 3.74158144e-01
-4.60027665e-01 -1.46491736e-01 3.83439064e-01 -1.06953955e+00
1.44610012e+00 -2.41057739e-01 -5.22430800e-02 1.11560918e-01
-6.37021840e-01 -1.31173208e-01 2.68584073e-01 -3.13014477e-01
-1.63046193e+00 -4.62998360e-01 -4.95700449e-01 6.60941124e-01
-6.72351956e-01 3.21013629e-01 -2.23088294e-01 1.48145603e-02
1.10210741e+00 -1.94701180e-01 4.68806565e-01 -1.43888563e-01
4.92239505e-01 5.34922421e-01 1.73860818e-01 -8.84326637e-01
3.80472809e-01 6.71733469e-02 -5.57624757e-01 -8.43261182e-01
-2.39279911e-01 3.71256769e-02 -1.36725858e-01 -1.00418293e+00
8.11806858e-01 -3.27928960e-01 -1.12407506e+00 2.40755379e-02
-1.23269141e+00 -6.47523880e-01 -4.91506875e-01 4.42264557e-01
-1.01951969e+00 3.43570530e-01 -4.00753558e-01 -1.42415595e+00
-6.16428971e-01 -6.57001317e-01 2.90469408e-01 1.17937833e-01
-1.25210869e+00 -1.13728142e+00 3.01901817e-01 7.12848246e-01
5.39365530e-01 3.71476412e-01 1.08272529e+00 -9.35204327e-01
1.31833673e-01 1.16416745e-01 4.56552118e-01 -2.38224044e-01
-1.15766451e-01 -2.98376139e-02 -7.90789902e-01 4.21204329e-01
7.35823810e-01 -5.89901090e-01 -2.97058076e-01 7.79432133e-02
-9.09084529e-02 -8.35439324e-01 -1.08285673e-01 -6.38797820e-01
8.50841999e-01 8.19838226e-01 9.45762455e-01 9.59421992e-01
-1.25375837e-01 1.13311768e+00 1.19665313e+00 4.81420130e-01
6.26222134e-01 7.26890743e-01 2.58969277e-01 2.95852274e-01
3.62562039e-03 -1.85665831e-01 4.51398253e-01 1.98510528e-01
-3.11994016e-01 -1.08111568e-01 -9.95747864e-01 2.18204539e-02
-1.99810612e+00 -1.24532664e+00 -2.41220936e-01 1.62664485e+00
6.98891878e-01 3.49071413e-01 1.55424729e-01 4.22969043e-01
7.14419067e-01 3.16174507e-01 1.77267626e-01 -1.13582706e+00
3.45487237e-01 -2.02530757e-01 -2.65503854e-01 1.20204246e+00
-6.61603689e-01 9.07364428e-01 6.42744589e+00 2.47347519e-01
-6.67943954e-01 2.86134154e-01 3.47745776e-01 3.47872019e-01
-3.53674650e-01 -1.15100794e-01 -1.34355485e-01 1.87136129e-01
1.02156460e+00 -3.52581561e-01 3.07827920e-01 8.61601293e-01
7.27300584e-01 -6.05220377e-01 -7.56672502e-01 2.25967795e-01
2.10467726e-01 -1.32853770e+00 -8.88095349e-02 -1.95206791e-01
-3.27253759e-01 -6.41201079e-01 -3.54143828e-01 1.74778044e-01
7.72536337e-01 -6.13770783e-01 1.32623041e+00 4.25388336e-01
-8.56057331e-02 -6.13992691e-01 9.52688098e-01 6.40758991e-01
-1.76986426e-01 5.05787097e-02 3.03841472e-01 -7.44705021e-01
4.49329764e-01 -2.26928711e-01 -1.27536130e+00 1.08488545e-01
5.45927405e-01 1.70304701e-02 -4.09884490e-02 3.17138523e-01
-6.12963498e-01 5.83900630e-01 -3.11975963e-02 -8.89207661e-01
8.44581425e-02 -4.17565107e-01 9.55275178e-01 1.25372100e+00
-1.59430578e-01 6.48536623e-01 -4.42438684e-02 6.39964640e-01
9.41562176e-01 5.24497211e-01 -7.58163035e-01 1.58884794e-01
6.31541610e-01 9.78963912e-01 -7.10735023e-01 -3.12665641e-01
6.01062104e-02 2.80701309e-01 2.64887065e-02 6.95683137e-02
-6.43807530e-01 2.77603507e-01 2.16586217e-01 2.70151913e-01
-2.89814711e-01 1.00368284e-01 -5.80616236e-01 -1.74305633e-01
-3.34565580e-01 -1.19597113e+00 2.13102579e-01 -1.19383824e+00
-6.92530334e-01 5.85519552e-01 6.69482410e-01 -4.02967423e-01
-7.63026595e-01 -3.20643276e-01 -6.89184427e-01 5.89358032e-01
-8.41218531e-01 -1.16800559e+00 -1.96234122e-01 1.40423805e-01
5.83173931e-01 -1.30853325e-01 1.26628745e+00 -3.83505017e-01
3.68959717e-02 -3.85294855e-01 -7.97746658e-01 -4.40098464e-01
4.02580172e-01 -1.05327582e+00 -1.28922731e-01 3.17253411e-01
-5.63566446e-01 6.86034620e-01 1.71647608e+00 -6.68950140e-01
-1.13293707e+00 -1.62236974e-01 1.03310776e+00 -2.03159809e-01
7.12678671e-01 -1.70629341e-02 -7.82292247e-01 4.43311721e-01
1.08387578e+00 -1.30306149e+00 9.14932609e-01 3.93059105e-02
3.95761468e-02 3.69623303e-01 -1.60956705e+00 9.09409106e-01
8.75714183e-01 -3.05599988e-01 -1.61990213e+00 1.94008753e-01
5.45803905e-01 -2.18360662e-01 -5.43208301e-01 -3.91098969e-02
4.65001643e-01 -1.35278201e+00 7.76486695e-01 -2.12539285e-01
2.54657865e-01 -1.89518526e-01 7.96808079e-02 -1.33283210e+00
-1.39622167e-01 -1.07581198e+00 3.37010801e-01 1.50091171e+00
3.24049860e-01 -9.02490258e-01 2.68588305e-01 9.77806687e-01
-1.58795044e-01 -1.09265232e-02 -8.85953903e-01 2.34597355e-01
-9.69010293e-02 -5.12943447e-01 3.97335351e-01 9.06218410e-01
9.86652255e-01 8.91722083e-01 -2.02225342e-01 -2.07438365e-01
1.01807773e-01 -4.64658618e-01 8.29001069e-01 -1.52703786e+00
-1.52647913e-01 -4.50002551e-01 1.19218687e-02 -2.79688478e-01
1.07963309e-01 -2.82649040e-01 1.38093472e-01 -1.89562738e+00
-2.52959073e-01 2.64977254e-02 7.07340240e-01 2.63972282e-01
1.69713676e-01 -6.75095379e-01 2.84554183e-01 -1.09635055e-01
-3.94062519e-01 3.15657258e-01 7.93516159e-01 2.62014866e-01
-5.01114964e-01 -4.26927537e-01 -1.13617051e+00 1.19438028e+00
1.29885650e+00 -2.02664912e-01 -7.94915736e-01 2.09713459e-01
5.45680702e-01 3.02324682e-01 2.16393083e-01 -6.89779460e-01
2.31655642e-01 -3.73647451e-01 -3.17777693e-01 -2.56982028e-01
2.87381083e-01 -1.12902546e+00 7.35594034e-01 7.59373605e-01
-6.14006162e-01 -4.01697345e-02 2.10402608e-01 2.14161485e-01
1.84678748e-01 -5.08959651e-01 4.77930188e-01 -1.61964044e-01
-6.37871206e-01 -1.16798413e+00 -1.47363913e+00 -3.24852288e-01
1.47535110e+00 -6.48048401e-01 -6.00312531e-01 -6.29293919e-01
-6.57508790e-01 2.60176003e-01 4.44175839e-01 3.54703605e-01
4.48706001e-01 -7.13746369e-01 -5.52032530e-01 -2.85500228e-01
-1.05837546e-01 -5.91513336e-01 2.44862080e-01 4.34396595e-01
-9.76654589e-01 4.51812565e-01 -8.58480155e-01 8.47442597e-02
-1.63108385e+00 6.44899845e-01 3.35970551e-01 -7.67591521e-02
-8.50254953e-01 2.54304618e-01 6.16117343e-02 -2.76583225e-01
1.27782241e-01 2.09657118e-01 -1.25911319e+00 2.49039024e-01
7.68677652e-01 6.03719294e-01 -4.14763868e-01 -5.83577812e-01
-1.60516649e-01 3.05677921e-01 -9.90125071e-03 -8.06181312e-01
9.85674918e-01 -4.79300737e-01 -2.55839884e-01 7.68901050e-01
-1.07857078e-01 5.99763021e-02 -7.39690304e-01 3.94470990e-01
3.45279247e-01 -1.97950348e-01 -1.90837532e-01 -1.48388457e+00
1.12344503e-01 3.80153656e-01 4.49527532e-01 7.54355431e-01
7.32677817e-01 -8.60037208e-02 -1.05063535e-01 5.45781434e-01
6.90915942e-01 -1.54299045e+00 -1.32724941e-01 3.63949031e-01
1.48484766e+00 -5.50342977e-01 9.24143493e-02 -8.96642387e-01
-1.05770516e+00 1.18500280e+00 5.21386862e-01 1.50883093e-01
4.14421797e-01 4.74636197e-01 5.10114551e-01 -7.38528907e-01
-1.03917682e+00 2.05537051e-01 -5.35140455e-01 1.08585155e+00
6.18951023e-01 3.63881677e-01 -1.48518491e+00 2.41180241e-01
-5.63288510e-01 2.46515870e-01 7.58728564e-01 1.11923623e+00
-7.20900357e-01 -9.60691929e-01 -7.95037448e-01 -2.69055188e-01
-2.64088720e-01 4.39171940e-01 -1.32923663e+00 1.10755813e+00
-6.07057102e-02 1.56757963e+00 -4.35434103e-01 -1.17400542e-01
5.40527105e-01 2.59613752e-01 2.32207686e-01 -3.70386451e-01
-1.40017891e+00 -1.10595658e-01 1.32998312e+00 -2.22968444e-01
-8.34635437e-01 -8.23713303e-01 -1.28892469e+00 -6.89320028e-01
1.53354704e-01 6.05011582e-01 6.74482048e-01 1.04604220e+00
-1.64723292e-01 5.55754900e-01 1.33344889e-01 -6.28088534e-01
-2.29707628e-01 -1.34168887e+00 -2.30123717e-02 9.33647826e-02
-4.45226766e-02 -6.17909849e-01 -1.09863065e-01 -3.35185796e-01] | [12.896756172180176, 7.926074028015137] |
a63f81a5-23e5-4e38-98cc-49ffe0c40da3 | p-3-lm-probabilistically-permuted-prophet | 2210.12339 | null | https://arxiv.org/abs/2210.12339v1 | https://arxiv.org/pdf/2210.12339v1.pdf | P$^3$LM: Probabilistically Permuted Prophet Language Modeling for Generative Pre-Training | Conventional autoregressive left-to-right (L2R) sequence generation faces two issues during decoding: limited to unidirectional target sequence modeling, and constrained on strong local dependencies. To address the aforementioned problem, we propose P$^3$LM, a probabilistically permuted prophet language model, which strengthens the modeling of bidirectional information and long token dependencies for sequence generation. Specifically, P$^3$LM learns to generate tokens in permuted order upon an order-aware transformer decoder, as well as to generate the corresponding future $N$ tokens with a multi-stream attention mechanism. Extensive experiments are conducted on the GLGE benchmark, which includes four datasets for summarization, two for question generation, one for conversational question answering, and one for dialog response generation, where P$^3$LM achieves state-of-the-art results compared with strong publicly available generative pre-training methods. | ['Xiaodong He', 'Youzheng Wu', 'Jing Zhao', 'Yeyun Gong', 'Jiangyong Ying', 'Yifan Wang', 'Junwei Bao'] | 2022-10-22 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 6.95378482e-01 6.91827357e-01 1.45598436e-02 -3.19618046e-01
-1.37790537e+00 -6.15814924e-01 8.33330393e-01 -3.01694125e-01
-6.09208681e-02 1.17178357e+00 8.69383276e-01 -7.10096478e-01
4.24142092e-01 -8.26610386e-01 -9.83113289e-01 -5.70903182e-01
1.67287767e-01 8.57332289e-01 -2.37660870e-01 -6.51842117e-01
7.80516490e-02 -2.17712671e-01 -9.49884713e-01 7.90434539e-01
9.00541246e-01 5.93964100e-01 3.45989674e-01 1.14454806e+00
-3.52042049e-01 1.23140419e+00 -8.12280416e-01 -7.51332045e-01
-2.40733072e-01 -1.13049746e+00 -1.14712596e+00 -7.79885352e-02
4.31359522e-02 -4.02653575e-01 -3.44966352e-01 5.52421391e-01
7.53294230e-01 8.62496495e-02 7.49726474e-01 -9.54460621e-01
-1.07538462e+00 1.51723588e+00 -2.03507274e-01 1.03307143e-01
7.41673231e-01 5.10361373e-01 1.51839221e+00 -8.45608950e-01
6.15052760e-01 1.43566358e+00 2.07051307e-01 9.09806788e-01
-1.20560527e+00 -5.07635415e-01 3.05142611e-01 1.67460471e-01
-1.11528230e+00 -5.74848652e-01 6.66542649e-01 -1.86740652e-01
1.44517601e+00 3.49590272e-01 3.78148228e-01 1.64851522e+00
9.65295807e-02 1.18580139e+00 7.07734287e-01 -2.94881970e-01
-1.54475961e-02 -8.92894939e-02 -7.76410326e-02 5.30147195e-01
-3.62684071e-01 -1.60311729e-01 -7.71323025e-01 -2.04006687e-01
4.40972477e-01 -6.95216835e-01 -3.44024479e-01 5.04316747e-01
-1.30761325e+00 9.08087611e-01 1.10776655e-01 1.12264141e-01
-4.44416493e-01 2.28684828e-01 3.35760087e-01 2.29152575e-01
5.24481773e-01 5.36317706e-01 -4.43335265e-01 -4.13909584e-01
-6.76083982e-01 5.12744904e-01 9.83025014e-01 1.24513781e+00
5.09261072e-01 4.66810584e-01 -9.15720642e-01 8.17716122e-01
2.54552364e-01 5.52213013e-01 5.86937010e-01 -5.76803565e-01
9.23173964e-01 1.78057522e-01 1.23443007e-01 -3.17553133e-01
6.70744851e-02 -3.20474714e-01 -1.02076769e+00 -6.87881649e-01
1.82232440e-01 -5.81597686e-01 -9.26719606e-01 1.90613198e+00
-3.38780843e-02 -1.00838980e-02 4.82336313e-01 4.56832647e-01
1.26988578e+00 1.20181346e+00 7.27702305e-02 -1.25564143e-01
1.31194365e+00 -1.26619053e+00 -6.37776971e-01 -4.45130020e-01
4.88284707e-01 -6.98480487e-01 1.17604887e+00 4.79201749e-02
-1.45619321e+00 -4.67914611e-01 -5.55023611e-01 -3.60780001e-01
9.65935066e-02 5.05079105e-02 3.71552259e-01 4.33287084e-01
-1.28120804e+00 1.98489144e-01 -4.11932677e-01 1.86345384e-01
3.66935283e-02 -1.01758037e-02 2.51781810e-02 2.66390969e-03
-1.63853467e+00 8.01877320e-01 2.99096465e-01 3.56501698e-01
-1.07161045e+00 -7.81607032e-01 -9.80674088e-01 2.97233760e-01
-7.99475908e-02 -1.16276348e+00 1.57033384e+00 -5.70796132e-01
-2.19498467e+00 6.68511152e-01 -5.98604500e-01 -7.91158140e-01
6.14974320e-01 -1.56325474e-01 -1.06489368e-01 -2.57933661e-02
-3.12312134e-02 8.13043416e-01 6.57980382e-01 -1.13479614e+00
-3.10848117e-01 1.29941940e-01 -7.90542513e-02 4.02038872e-01
1.89315096e-01 -1.13039732e-01 -1.11634508e-01 -7.30890930e-01
-3.90383661e-01 -7.68284321e-01 -3.55274737e-01 -1.25855160e+00
-1.03861094e+00 -4.87881064e-01 3.26829344e-01 -1.17236447e+00
1.22371876e+00 -1.54946911e+00 4.19403285e-01 -1.47617608e-01
-3.51831675e-01 2.00120121e-01 -5.74448049e-01 8.65143180e-01
-2.31831133e-01 2.45866179e-01 -3.67943376e-01 -5.59819102e-01
2.94449806e-01 -3.58717665e-02 -9.28218246e-01 -3.83506268e-01
5.79668403e-01 1.38970673e+00 -9.79896545e-01 -2.62932032e-01
-2.32278079e-01 2.84659177e-01 -7.47346520e-01 7.05809295e-01
-8.22301745e-01 5.16295552e-01 -3.31121355e-01 3.73795629e-01
1.96960494e-01 -4.41474050e-01 4.41484034e-01 1.96466714e-01
2.41468370e-01 1.07084405e+00 -3.88889313e-01 1.56471300e+00
-7.12679625e-01 3.67497563e-01 -2.86088258e-01 -7.24476874e-01
1.01487350e+00 7.67856538e-01 4.16875444e-02 -5.32838166e-01
-7.34736444e-03 1.22791506e-01 -4.89518158e-02 -3.85540694e-01
9.99172270e-01 -3.30875754e-01 -3.39560956e-01 6.41386986e-01
1.47269696e-01 -3.61095726e-01 3.15333575e-01 6.37377977e-01
1.18759334e+00 2.10286841e-01 1.12883754e-01 1.16409853e-01
4.53061908e-01 -2.37482250e-01 2.86800295e-01 9.14067745e-01
4.47040111e-01 7.18230128e-01 7.23031998e-01 9.58109945e-02
-1.05773759e+00 -1.12062693e+00 4.93614435e-01 1.16577339e+00
-2.79481918e-01 -4.11299169e-01 -8.38339031e-01 -7.35182345e-01
-4.14381951e-01 1.48798144e+00 -3.57912481e-01 -2.96071231e-01
-1.09203160e+00 -8.72707248e-01 7.83710539e-01 5.73504031e-01
3.13710123e-01 -1.66189456e+00 -6.09972775e-02 6.56584561e-01
-1.02425838e+00 -1.07200265e+00 -8.19533825e-01 -1.18724450e-01
-5.76719582e-01 -4.97515231e-01 -8.17269742e-01 -8.87657106e-01
6.45723760e-01 -1.69250533e-01 1.63504195e+00 -9.55091119e-02
1.18449382e-01 1.65629685e-01 -5.73843539e-01 -2.53179193e-01
-1.02128983e+00 5.17156720e-01 -6.04544640e-01 -8.47107843e-02
-9.95223820e-02 -5.83197057e-01 -5.45248210e-01 -2.03891248e-02
-8.61673772e-01 3.79409671e-01 7.61642456e-01 1.11219990e+00
3.74749094e-01 -6.95449650e-01 1.23409557e+00 -1.03448689e+00
1.09066892e+00 -5.11603475e-01 -2.82552183e-01 5.09777129e-01
-2.21588239e-01 3.17837268e-01 8.70424151e-01 -2.75697768e-01
-1.47864377e+00 -3.89648348e-01 -7.95760632e-01 1.32538676e-01
1.13960259e-01 5.16900897e-01 -3.67094636e-01 7.32468367e-01
4.41756129e-01 9.73417222e-01 -3.18965882e-01 -3.43299448e-01
8.86752546e-01 6.22875452e-01 5.45605421e-01 -7.00097501e-01
6.65106714e-01 -1.01002127e-01 -5.07948339e-01 -6.91252530e-01
-7.36136794e-01 -7.83242285e-02 -1.66691855e-01 1.71153247e-01
8.50410819e-01 -8.50433767e-01 -6.73480392e-01 6.03960514e-01
-1.64080811e+00 -7.37134635e-01 -4.75879699e-01 -1.33118387e-02
-7.71702111e-01 3.47544193e-01 -9.96262610e-01 -8.48186195e-01
-9.91188645e-01 -1.04622507e+00 1.22664237e+00 6.67233914e-02
-4.26528841e-01 -8.37017119e-01 1.61279142e-01 7.15278387e-01
4.92897391e-01 -1.54454350e-01 1.24168003e+00 -6.98156476e-01
-9.31434453e-01 1.03948109e-01 2.96334904e-02 4.31706101e-01
-9.71935615e-02 -2.64529884e-01 -6.51191056e-01 -1.19198687e-01
-1.53428882e-01 -5.08157074e-01 1.03360283e+00 2.39274800e-01
1.08213341e+00 -9.86543477e-01 -2.73965374e-02 3.07244778e-01
6.99945688e-01 3.06073993e-01 9.24996495e-01 -2.26960868e-01
7.39342093e-01 4.37062174e-01 1.45418704e-01 4.92109984e-01
8.97639692e-01 4.64338481e-01 2.34737560e-01 5.33643439e-02
-1.87200665e-01 -8.87795389e-01 6.46717191e-01 1.28947592e+00
1.28502637e-01 -9.39099073e-01 -4.76614892e-01 6.30155861e-01
-1.55794847e+00 -1.13785541e+00 -2.79115047e-02 1.84522307e+00
1.35048187e+00 -1.68016225e-01 -6.21473566e-02 -3.29031944e-01
6.84181273e-01 6.73454940e-01 -3.38046640e-01 -5.56211054e-01
-1.74213797e-01 5.63791692e-01 1.35261267e-01 7.05002487e-01
-5.65022290e-01 1.25300837e+00 6.53339100e+00 9.65050161e-01
-8.23480010e-01 -4.88583138e-03 8.95792663e-01 -3.37008499e-02
-1.02974474e+00 8.54278654e-02 -1.04473925e+00 6.48085654e-01
1.18676794e+00 -3.70578885e-01 4.43533003e-01 5.48673391e-01
1.37721255e-01 3.53318244e-01 -1.12382603e+00 5.44749081e-01
2.27359116e-01 -1.60447776e+00 6.01076365e-01 -2.87051082e-01
7.69183636e-01 -1.49406210e-01 1.27563179e-01 6.29792213e-01
9.53420818e-01 -1.23837543e+00 6.78850293e-01 3.51981372e-01
7.48003900e-01 -6.44679785e-01 4.88739848e-01 5.27238667e-01
-1.00100935e+00 3.11323792e-01 -1.54717892e-01 1.82417125e-01
1.06327355e+00 4.18581158e-01 -1.45324504e+00 7.25415766e-01
1.61290154e-01 3.85397315e-01 2.62947491e-04 2.91939199e-01
-7.42044449e-01 1.11086893e+00 1.12099066e-01 -4.25945371e-01
2.55635262e-01 -1.49930134e-01 6.81858361e-01 1.40464652e+00
4.75284934e-01 1.57250330e-01 -2.02065215e-01 9.32375252e-01
-6.25640690e-01 -2.38593016e-02 -3.35789949e-01 -2.95517087e-01
3.95928591e-01 1.03110492e+00 -1.29974812e-01 -5.87326407e-01
7.42221475e-02 1.18521118e+00 2.52447486e-01 4.97176796e-01
-9.01967049e-01 -3.85263413e-01 3.92257690e-01 -1.04452513e-01
4.66219902e-01 -2.31403694e-01 6.19597733e-03 -1.30658615e+00
-6.68432564e-02 -1.30412817e+00 4.53376085e-01 -8.43140483e-01
-1.35705781e+00 9.18340206e-01 -1.19053178e-01 -6.80969417e-01
-1.01198387e+00 -2.90501453e-02 -8.36529613e-01 1.19232714e+00
-1.62042236e+00 -1.33235216e+00 1.04227334e-01 3.36213380e-01
9.59394157e-01 -9.21997577e-02 8.77863705e-01 9.04548690e-02
-4.14389491e-01 7.64138937e-01 -1.22123636e-01 2.06168637e-01
4.79110926e-01 -1.23385406e+00 1.14847422e+00 9.70400929e-01
1.76167473e-01 5.71894348e-01 5.86151958e-01 -5.43979824e-01
-1.30977130e+00 -1.25232601e+00 1.41787755e+00 -5.15180588e-01
5.19417822e-01 -5.55277765e-01 -8.83889496e-01 1.01790333e+00
7.95366824e-01 -7.27045774e-01 7.71693051e-01 -1.95044935e-01
-2.99074441e-01 2.99742252e-01 -7.43107021e-01 8.09964657e-01
1.15288687e+00 -4.48087841e-01 -6.18534565e-01 5.80457687e-01
1.32448983e+00 -6.51957691e-01 -5.55000663e-01 1.87719494e-01
2.57330149e-01 -6.90960467e-01 9.86712873e-01 -8.37698162e-01
9.56679642e-01 1.12755828e-01 3.93552743e-02 -1.65036833e+00
-1.41543135e-01 -1.29372597e+00 -1.71813015e-02 1.44420040e+00
1.01282108e+00 -6.84846044e-01 6.78776920e-01 3.13774258e-01
-5.15733123e-01 -6.98147416e-01 -8.68785679e-01 -4.59635943e-01
3.23886126e-01 -1.22336462e-01 8.31995308e-01 4.89271760e-01
5.31859621e-02 9.78343010e-01 -8.86490345e-01 -2.26561040e-01
2.65017331e-01 2.87139058e-01 8.57076108e-01 -2.13217035e-01
-7.50813901e-01 -5.05045652e-01 4.66889411e-01 -1.83564830e+00
4.50402528e-01 -1.05123925e+00 6.15658224e-01 -1.86419773e+00
1.01559669e-01 -2.45034516e-01 2.66727537e-01 1.88566923e-01
-6.65405512e-01 -6.29912019e-02 1.14529192e-01 -1.52331501e-01
-3.77463967e-01 1.15283239e+00 1.36244929e+00 -1.51269868e-01
-4.25894298e-02 1.64907172e-01 -9.89497304e-01 2.71561444e-01
6.72953069e-01 -2.54996568e-01 -4.50042397e-01 -6.46071732e-01
3.73923033e-01 6.11615419e-01 5.29682115e-02 -2.53562391e-01
-3.21423188e-02 -8.85282457e-02 2.30860580e-02 -7.41807580e-01
5.09081960e-01 2.13183030e-01 -8.77850428e-02 2.95000076e-01
-1.01920676e+00 2.18564838e-01 -1.77317828e-01 4.74529684e-01
-2.27561861e-01 -2.10135207e-01 5.03886223e-01 -4.24394011e-01
-2.49653667e-01 2.65741706e-01 -5.05952775e-01 6.26186728e-01
4.83679175e-01 3.41361731e-01 -6.60514355e-01 -9.61351812e-01
-4.58382398e-01 4.43924516e-01 -2.46145919e-01 5.85415542e-01
6.53528214e-01 -1.13296998e+00 -1.32352376e+00 -6.00600764e-02
-1.83212608e-01 1.88642517e-01 5.81709683e-01 3.94411713e-01
-2.23215774e-01 6.01599753e-01 3.22706461e-01 -1.97718382e-01
-1.01991951e+00 1.33988664e-01 1.61637917e-01 -1.03831303e+00
-3.87946427e-01 1.30895984e+00 2.61860728e-01 -5.73021770e-01
-1.48136005e-01 -2.77584106e-01 -1.46053001e-01 -1.70358777e-01
4.21706796e-01 1.74867690e-01 -7.72350878e-02 -4.68989670e-01
-7.74487033e-02 -7.42897093e-02 -2.93831557e-01 -5.30992568e-01
1.04246819e+00 -1.46817848e-01 -2.45243609e-01 8.46660957e-02
9.93842185e-01 6.54187379e-03 -1.07040942e+00 -3.27576369e-01
-2.19652265e-01 -9.03077871e-02 -6.90543294e-01 -9.71775830e-01
-6.75289512e-01 9.33287084e-01 -4.87926900e-01 2.25706413e-01
7.18027830e-01 1.40546188e-01 1.38508677e+00 3.05594832e-01
9.41223949e-02 -6.57406569e-01 5.23302734e-01 1.05403566e+00
1.26750612e+00 -7.96682358e-01 -5.86984396e-01 -1.76543608e-01
-1.13753247e+00 6.93211615e-01 6.82175815e-01 1.12292938e-01
6.08443990e-02 1.00262813e-01 4.43967897e-03 2.44961858e-01
-1.34959888e+00 2.54517943e-02 1.07665353e-01 4.79171008e-01
7.38772154e-01 1.39201656e-01 -2.91666210e-01 6.93500638e-01
-8.51942301e-01 -1.86230406e-01 5.70196033e-01 5.12071848e-01
-5.35029322e-02 -1.46612418e+00 1.11648217e-02 3.26245248e-01
-5.50198853e-01 -7.77351141e-01 -4.02654171e-01 3.34411800e-01
-4.99331415e-01 1.17027771e+00 9.59530696e-02 -2.20040768e-01
2.11209700e-01 3.58668059e-01 3.75321746e-01 -7.09124029e-01
-8.80919278e-01 -9.99670550e-02 6.61415398e-01 -1.06563859e-01
-1.11407423e-02 -5.07330179e-01 -1.24899089e+00 -1.53086945e-01
-1.11264914e-01 4.76361722e-01 4.23702687e-01 9.28976476e-01
6.35926545e-01 8.67415011e-01 6.97194517e-01 -6.37119949e-01
-9.86616254e-01 -1.35057664e+00 -1.27992347e-01 3.21675688e-01
8.86019394e-02 2.17426553e-01 -2.54772633e-01 2.77645856e-01] | [11.9098482131958, 8.958085060119629] |
515d3961-2d7b-45db-9796-d0bfd73d65c0 | federated-self-learning-with-weak-supervision | 2306.12015 | null | https://arxiv.org/abs/2306.12015v1 | https://arxiv.org/pdf/2306.12015v1.pdf | Federated Self-Learning with Weak Supervision for Speech Recognition | Automatic speech recognition (ASR) models with low-footprint are increasingly being deployed on edge devices for conversational agents, which enhances privacy. We study the problem of federated continual incremental learning for recurrent neural network-transducer (RNN-T) ASR models in the privacy-enhancing scheme of learning on-device, without access to ground truth human transcripts or machine transcriptions from a stronger ASR model. In particular, we study the performance of a self-learning based scheme, with a paired teacher model updated through an exponential moving average of ASR models. Further, we propose using possibly noisy weak-supervision signals such as feedback scores and natural language understanding semantics determined from user behavior across multiple turns in a session of interactions with the conversational agent. These signals are leveraged in a multi-task policy-gradient training approach to improve the performance of self-learning for ASR. Finally, we show how catastrophic forgetting can be mitigated by combining on-device learning with a memory-replay approach using selected historical datasets. These innovations allow for 10% relative improvement in WER on new use cases with minimal degradation on other test sets in the absence of strong-supervision signals such as ground-truth transcriptions. | ['Jasha Droppo', 'Ariya Rastrow', 'Anirudh Raju', 'Anit Kumar Sahu', 'Gautam Tiwari', 'Gopinath Chennupati', 'Milind Rao'] | 2023-06-21 | null | null | null | null | ['incremental-learning', 'self-learning', 'automatic-speech-recognition'] | ['methodology', 'natural-language-processing', 'speech'] | [ 5.04389107e-01 6.21408463e-01 -2.24910393e-01 -4.00520653e-01
-1.42111492e+00 -4.49147969e-01 7.11089075e-01 -1.26827732e-01
-4.95032430e-01 7.43745744e-01 6.31967783e-01 -5.80775142e-01
2.84590095e-01 2.18515247e-02 -9.00956213e-01 -4.45961028e-01
-6.71592308e-03 4.77618366e-01 -1.58902749e-01 -9.70297828e-02
-4.83812153e-01 1.18901432e-01 -1.09604108e+00 5.90779483e-01
5.18990040e-01 1.02110314e+00 -2.31894888e-02 9.63083923e-01
3.41866732e-01 1.10845411e+00 -7.99238324e-01 -6.29639149e-01
5.04869044e-01 -2.77786732e-01 -4.19488460e-01 3.21980685e-01
1.82703167e-01 -7.59433627e-01 -8.57182562e-01 5.59924662e-01
6.98557496e-01 2.05539376e-01 1.14919759e-01 -1.01287341e+00
-3.74680877e-01 6.55535758e-01 -2.47154474e-01 -1.45278340e-02
2.52488941e-01 5.11002243e-01 6.89394474e-01 -5.98070383e-01
4.60590601e-01 1.09456563e+00 6.83023334e-01 1.07349706e+00
-1.19268405e+00 -6.21121526e-01 1.54276773e-01 -6.41482100e-02
-9.73575592e-01 -1.43443453e+00 5.07251918e-01 1.66767195e-01
1.34753299e+00 4.64169800e-01 2.62650102e-01 1.80755198e+00
-2.11670920e-01 1.14119077e+00 8.10485959e-01 -2.92501479e-01
4.50638205e-01 7.46234119e-01 -1.22054301e-01 5.73049307e-01
-1.08006410e-01 2.48764023e-01 -1.28272867e+00 -5.34921706e-01
2.02754900e-01 2.78312743e-01 -2.38621265e-01 -1.38236508e-01
-7.11556792e-01 4.73959565e-01 -1.74516395e-01 1.20809726e-01
-6.21149540e-01 1.31042391e-01 5.80452561e-01 8.20708990e-01
9.85308409e-01 2.00770557e-01 -1.05234206e+00 -4.93173599e-01
-9.22499716e-01 -8.38283226e-02 9.28543031e-01 1.11491001e+00
5.33777058e-01 1.52951747e-01 -2.72399902e-01 7.72444427e-01
4.51525487e-02 5.56337476e-01 8.52568269e-01 -1.02954471e+00
6.68708861e-01 -1.38756027e-02 1.86693370e-01 -1.18073322e-01
-2.97820400e-02 -6.26689970e-01 -5.51325798e-01 -1.05093375e-01
2.92110890e-01 -6.78830564e-01 -8.88160229e-01 1.69346714e+00
3.12171102e-01 6.93846524e-01 4.55926061e-01 3.81604195e-01
1.89032018e-01 5.26261747e-01 -3.17576379e-01 -6.46062911e-01
1.13098085e+00 -1.08461201e+00 -7.29473770e-01 -1.94508567e-01
1.01249504e+00 -1.57908618e-01 1.11149526e+00 4.50382829e-01
-1.01567197e+00 2.03497391e-02 -8.62643480e-01 1.31083593e-01
-9.03410837e-02 1.27999643e-02 3.55113506e-01 1.11108494e+00
-1.33951771e+00 4.31054294e-01 -1.07225430e+00 -3.27698678e-01
4.39501852e-01 4.58734035e-01 -2.64462650e-01 5.59922680e-03
-1.01985312e+00 7.36470640e-01 -2.14067146e-01 -1.55585095e-01
-1.28685164e+00 -8.54945898e-01 -6.46574914e-01 -1.08265325e-01
6.71083033e-01 -5.87255538e-01 1.85847366e+00 -1.14034891e+00
-2.31637096e+00 6.40930355e-01 -4.85920310e-01 -1.24667072e+00
8.61458182e-01 -3.86169434e-01 -5.36752641e-01 -1.60802007e-01
-4.55503523e-01 1.79548800e-01 1.23452592e+00 -9.75725532e-01
-7.45390892e-01 -6.60538375e-01 -1.19739167e-01 3.16283882e-01
-6.71841979e-01 -1.05701052e-01 -1.64569899e-01 -4.24071401e-01
-4.73153293e-01 -1.13965869e+00 -2.26123303e-01 -3.21341544e-01
-2.13172823e-01 -2.20619723e-01 1.08109450e+00 -1.17233562e+00
1.05968916e+00 -2.24610567e+00 -4.24296051e-01 1.97056115e-01
3.62759084e-03 4.76796210e-01 -1.88819304e-01 1.54692978e-01
2.67051756e-01 -1.08560666e-01 8.75814483e-02 -1.02295184e+00
-2.27493718e-01 -2.08201390e-02 -5.08472264e-01 2.57152528e-01
-1.02119476e-01 8.25785279e-01 -7.87952483e-01 1.67955637e-01
8.82440582e-02 5.40573120e-01 -4.41923767e-01 4.88532364e-01
-4.33268517e-01 5.12196839e-01 -1.52101845e-01 4.02318567e-01
1.37876093e-01 -1.40378416e-01 4.29506898e-01 2.01648429e-01
3.70859206e-01 7.34345794e-01 -6.71944320e-01 1.99066591e+00
-1.04729295e+00 4.24369007e-01 5.42017162e-01 -6.24601543e-01
7.07888484e-01 8.98492157e-01 2.64010340e-01 -8.12326372e-01
-6.31025434e-03 1.38812765e-01 -4.00860816e-01 -3.97933841e-01
3.68863195e-01 -3.88668850e-02 -4.91755717e-02 7.52178013e-01
2.35596448e-01 2.41215169e-01 -6.93378508e-01 3.55478048e-01
1.74296331e+00 -3.58552068e-01 5.73230647e-02 4.39998448e-01
-2.56676264e-02 -4.38389301e-01 3.34431112e-01 1.03916240e+00
-2.67108411e-01 3.54431212e-01 -9.50141475e-02 -3.95707637e-01
-1.03126466e+00 -6.50330901e-01 2.24009410e-01 1.43628490e+00
-5.41558206e-01 -4.43945915e-01 -9.11100984e-01 -9.81991947e-01
-9.03801098e-02 1.19073105e+00 -4.33887422e-01 -4.52517271e-01
-4.44608033e-01 -6.60163999e-01 5.91955841e-01 2.50499189e-01
2.83856452e-01 -7.49477684e-01 -4.62602586e-01 3.39489490e-01
-6.77704066e-02 -1.12053633e+00 -8.44270527e-01 4.81600165e-01
-9.01923239e-01 -2.33117968e-01 -6.00664437e-01 -1.47914842e-01
3.18949074e-01 3.60330969e-01 7.95135856e-01 -3.62986118e-01
1.52410343e-01 9.41249013e-01 -1.65448010e-01 -4.05859590e-01
-8.23852897e-01 3.27722728e-01 5.35638690e-01 5.05317450e-01
2.85195291e-01 -8.62659931e-01 -3.16564113e-01 1.51154101e-01
-4.61198807e-01 -1.99929148e-01 6.08777523e-01 9.01600301e-01
5.09256758e-02 -3.36867869e-01 9.03203487e-01 -1.01389349e+00
8.82151902e-01 -5.62551677e-01 -2.99001485e-01 4.87525910e-01
-7.51377821e-01 2.93547660e-01 5.50867975e-01 -7.80589938e-01
-1.39356411e+00 1.82981014e-01 5.10079265e-02 -8.00349474e-01
-6.43317625e-02 -4.53836098e-02 -3.36260706e-01 2.59098351e-01
9.04613912e-01 4.23419923e-01 1.14757992e-01 -4.26062107e-01
6.50990069e-01 1.37260044e+00 6.58926129e-01 -4.14283842e-01
3.42703283e-01 2.77106255e-01 -6.88587129e-01 -8.92516673e-01
-4.79173541e-01 -4.81028974e-01 -5.77519685e-02 8.45078938e-03
1.91509262e-01 -1.38920569e+00 -8.55323911e-01 4.81808037e-01
-1.04389977e+00 -7.71144509e-01 -5.43544292e-01 3.25718820e-01
-7.13938117e-01 -1.61039364e-02 -6.99380219e-01 -1.56001663e+00
-8.09552610e-01 -6.36153519e-01 1.08921063e+00 7.07998797e-02
-3.48515332e-01 -7.52154469e-01 4.42696922e-03 7.25945592e-01
5.34415960e-01 -5.97545981e-01 2.99628496e-01 -1.31717205e+00
-6.17572069e-01 -4.22155976e-01 4.18999046e-01 4.67382789e-01
2.24542052e-01 -7.31775939e-01 -1.52496147e+00 -8.73626471e-01
3.51545274e-01 -6.40833914e-01 6.75037265e-01 1.98645860e-01
9.47326064e-01 -9.88242328e-01 -3.73174876e-01 3.01324934e-01
6.54387772e-01 4.56434697e-01 2.52982616e-01 -2.10158899e-01
4.27034885e-01 3.62843454e-01 2.08736539e-01 6.44322991e-01
3.24329972e-01 7.14520454e-01 -1.24044277e-01 2.67530680e-01
-1.00450613e-01 -7.25147426e-01 9.42671478e-01 7.47039914e-01
3.30582023e-01 -3.49278212e-01 -5.64383864e-01 6.79143727e-01
-1.91587353e+00 -8.13847303e-01 6.98372066e-01 2.65602589e+00
8.63038957e-01 1.93247959e-01 4.51269537e-01 -4.01075721e-01
6.29506886e-01 9.32102352e-02 -1.25158525e+00 -2.05962360e-01
-9.12817603e-04 8.48581865e-02 9.08320785e-01 6.51390374e-01
-6.86495721e-01 8.00188780e-01 5.78221750e+00 8.57768059e-01
-1.17477024e+00 8.63107145e-01 1.11064208e+00 -6.76712513e-01
-2.91889131e-01 -2.03313380e-01 -6.66712165e-01 3.99169296e-01
1.87072587e+00 -1.16697386e-01 9.19170976e-01 9.24101591e-01
3.58187705e-01 2.04945490e-01 -1.24429393e+00 1.22855484e+00
1.64619938e-01 -1.25056028e+00 -4.47522640e-01 1.36288241e-01
6.12919569e-01 5.77338934e-01 4.33861703e-01 3.68047625e-01
8.98607075e-01 -6.21878743e-01 5.25498986e-01 3.58968526e-01
1.06589162e+00 -4.73657012e-01 4.67826396e-01 7.26884604e-01
-5.64795911e-01 -3.95502239e-01 1.31334394e-01 1.19157441e-01
6.26141429e-02 4.10129964e-01 -1.64046443e+00 2.24736676e-01
4.60930586e-01 3.65992934e-01 -1.11226998e-01 4.13937837e-01
7.98466057e-02 1.08668232e+00 -7.10734546e-01 -1.24326520e-01
-8.66742879e-02 2.34810874e-01 6.83928311e-01 1.06302798e+00
3.18319529e-01 1.04234174e-01 -1.19770803e-01 1.80473208e-01
-6.48626268e-01 -3.22312047e-03 -9.10559237e-01 2.88705099e-02
8.32028031e-01 9.02058423e-01 5.57881929e-02 -5.09038091e-01
-2.91304141e-01 1.26591957e+00 3.46171618e-01 3.98641735e-01
-3.64649177e-01 2.51415282e-01 7.58823335e-01 3.99405420e-01
9.05423239e-02 -7.72246048e-02 -1.13576695e-01 -1.30771971e+00
2.51910746e-01 -1.26991057e+00 4.36465055e-01 -5.22379100e-01
-1.18244386e+00 6.08063638e-01 -4.42959160e-01 -7.76886165e-01
-7.69173563e-01 4.77651507e-02 -5.35685182e-01 5.28246880e-01
-9.97559190e-01 -1.06363392e+00 4.51787502e-01 6.51791871e-01
9.50050652e-01 -6.80455744e-01 9.32892025e-01 2.65343964e-01
-4.53356624e-01 1.26609325e+00 5.05844712e-01 -1.11789636e-01
7.46197701e-01 -8.73268604e-01 7.44803548e-01 7.10072160e-01
4.68373537e-01 4.89370465e-01 6.80658698e-01 -7.20149279e-01
-1.45945501e+00 -1.35582340e+00 6.59058571e-01 -7.91211009e-01
5.38463950e-01 -9.03471768e-01 -8.16823900e-01 1.07931924e+00
2.87017941e-01 -1.78798549e-02 5.71806490e-01 3.58347237e-01
-4.27658170e-01 -3.36414933e-01 -1.37645149e+00 4.92804736e-01
1.22545981e+00 -1.12138867e+00 -1.21505290e-01 4.35982317e-01
1.22938001e+00 -3.72975230e-01 -3.89885962e-01 4.41601761e-02
4.32226658e-01 -4.11879301e-01 4.15243059e-01 -8.89698684e-01
-4.71912354e-01 1.12240054e-01 -1.24084167e-01 -1.42126369e+00
3.59925359e-01 -1.81230247e+00 -5.69629073e-01 1.26984704e+00
7.12976873e-01 -6.57453477e-01 1.26100910e+00 1.03339207e+00
4.02133912e-02 -3.68466020e-01 -1.26845515e+00 -6.71081662e-01
-5.52751184e-01 -6.81272864e-01 4.48006630e-01 7.03638554e-01
2.68523753e-01 7.96262860e-01 -1.09328771e+00 2.41216898e-01
5.40683866e-01 -7.19666541e-01 1.03041863e+00 -5.71873248e-01
-7.23190844e-01 3.08044344e-01 6.25490956e-03 -1.37089336e+00
3.26463103e-01 -7.82381892e-01 8.30966383e-02 -6.64998591e-01
-9.90030020e-02 -2.26103038e-01 -2.41744339e-01 4.96392548e-01
1.24796942e-01 -5.26366413e-01 1.20821543e-01 4.61505689e-02
-1.08854687e+00 7.48844862e-01 3.52709651e-01 -2.07762152e-01
-6.53998196e-01 6.19767487e-01 -6.01919532e-01 3.68453234e-01
6.54524446e-01 -6.86248124e-01 -6.62555695e-01 -2.20306560e-01
-2.77724862e-01 5.35815895e-01 7.93075114e-02 -7.78236330e-01
6.18027568e-01 3.39968741e-01 -8.18168148e-02 -9.10002887e-02
5.22210717e-01 -1.02149487e+00 1.66388392e-01 2.60848254e-01
-8.95465851e-01 -1.99854404e-01 3.02352142e-02 1.03803837e+00
4.58664417e-01 1.17262319e-01 3.57199192e-01 -1.23167187e-02
-1.29341809e-02 2.75348008e-01 -4.56043541e-01 -2.39470974e-01
6.69504166e-01 1.63500175e-01 -1.18189633e-01 -1.09940255e+00
-8.78620267e-01 1.75561219e-01 1.24356754e-01 5.26002824e-01
4.06486779e-01 -9.13809836e-01 -7.24093795e-01 1.27203301e-01
-7.24913105e-02 -9.76823121e-02 2.50797182e-01 5.49010873e-01
6.42615080e-01 3.50773871e-01 4.22887415e-01 -3.20710689e-01
-1.45281112e+00 5.64375162e-01 6.01206183e-01 -4.36856627e-01
-4.36147094e-01 9.27238643e-01 -3.68350893e-02 -5.80842376e-01
7.81204641e-01 -1.81002781e-01 6.59621000e-01 -2.75961637e-01
8.24107111e-01 3.65397424e-01 5.90965807e-01 -8.54299888e-02
-7.16471225e-02 -6.40925705e-01 -5.11341631e-01 -6.38750196e-01
1.38742793e+00 -4.95691866e-01 5.46119034e-01 6.19291842e-01
1.23900950e+00 1.52288556e-01 -1.74533045e+00 -9.44823921e-01
-1.32373003e-02 -1.96367621e-01 7.16706887e-02 -1.12986624e+00
-8.11023951e-01 5.83343267e-01 9.98711944e-01 7.24308193e-02
8.54224801e-01 1.26208901e-01 1.14547706e+00 7.05528140e-01
6.81881845e-01 -1.21315587e+00 1.37893334e-01 1.89241767e-01
5.22759199e-01 -1.21432543e+00 -5.32157958e-01 2.96019822e-01
-8.49479973e-01 5.70119143e-01 2.23931447e-01 5.04794419e-01
6.68617725e-01 5.36517322e-01 3.24772179e-01 2.22562104e-01
-1.42154658e+00 9.92061496e-02 -2.26298511e-01 7.79088378e-01
-9.25101787e-02 2.88858175e-01 4.55034256e-01 7.76406765e-01
-1.70956075e-01 3.99025492e-02 2.57024944e-01 7.42343664e-01
-5.42110987e-02 -9.51411724e-01 -4.48044352e-02 6.12789214e-01
-4.30494785e-01 -2.48963133e-01 -3.28118443e-01 -1.10317431e-01
-3.51521164e-01 1.24186099e+00 -6.62042275e-02 -7.68799961e-01
3.64335179e-01 5.10315776e-01 1.23404369e-01 -4.73670691e-01
-7.79001653e-01 -7.61082992e-02 5.82667232e-01 -4.71089244e-01
7.41291791e-02 -1.04346466e+00 -7.26972878e-01 -2.97552496e-01
-4.34162498e-01 3.11540663e-01 8.20086479e-01 1.01179576e+00
9.56483066e-01 2.63638377e-01 1.02744138e+00 -5.28087080e-01
-1.09376395e+00 -1.10936201e+00 -2.57277220e-01 1.62641838e-01
6.46222591e-01 1.35289475e-01 -5.04424393e-01 -1.94476694e-02] | [14.264229774475098, 6.493922233581543] |
8b5a6209-7b94-4484-8c34-4d664f14f450 | convnets-with-smooth-adaptive-activation | null | null | http://proceedings.mlr.press/v54/hou17a.html | http://proceedings.mlr.press/v54/hou17a/hou17a.pdf | ConvNets with Smooth Adaptive Activation Functions for Regression | Within Neural Networks (NN), the parameters of Adaptive Activation Functions (AAF) control the shapes of activation functions.
These parameters are trained along with other parameters in the NN. AAFs have improved performance of Convolutional Neural Networks (CNN) in multiple classification tasks. In this paper, we propose and apply AAFs on CNNs for regression tasks. We argue that applying AAFs in the regression (second-to-last) layer of a NN can significantly decrease the bias of the regression NN. However, using existing AAFs may lead to overfitting. To address this problem, we propose a Smooth Adaptive Activation Function (SAAF) with a piecewise polynomial form which can approximate any continuous function to arbitrary degree of error, while having a bounded Lipschitz constant for given bounded model parameters. As a result, NNs with SAAF can avoid overfitting by simply regularizing model parameters. We empirically evaluated CNNs with SAAFs and achieved state-of-the-art results on age and pose estimation datasets. | ['Le Hou ; Dimitris Samaras ; Tahsin M. Kurc ; Yi Gao ; Joel H. Saltz'] | 2017-01-01 | null | null | null | null | ['age-and-gender-classification'] | ['computer-vision'] | [ 5.32506965e-03 3.39553982e-01 -1.12784050e-01 -7.71228433e-01
-2.09300548e-01 -3.44569236e-01 9.03652236e-02 -3.29783469e-01
-7.45730519e-01 5.33949375e-01 -2.12086588e-01 -1.17934957e-01
8.08634683e-02 -6.71686590e-01 -1.15525770e+00 -6.32423222e-01
1.53665856e-01 5.31065799e-02 3.44602108e-01 -2.01142877e-01
-1.23220630e-01 7.99208403e-01 -1.09480941e+00 1.79638103e-01
8.84371161e-01 1.43621242e+00 -1.55848920e-01 4.48368073e-01
3.36795449e-02 6.12737238e-01 -7.13581085e-01 -5.22801280e-01
4.60701883e-01 1.00203507e-01 -5.01728117e-01 -3.72571409e-01
6.94480121e-01 -3.30325067e-01 -4.53831017e-01 8.97219658e-01
3.53161454e-01 3.03427100e-01 7.50153244e-01 -1.22081637e+00
-7.18668163e-01 4.95917827e-01 -4.43674266e-01 -3.56170051e-02
-5.00550091e-01 -1.36910841e-01 5.97590625e-01 -9.71770942e-01
1.83303773e-01 1.21964371e+00 1.25589049e+00 1.00332296e+00
-1.19621801e+00 -8.95386338e-01 5.78835964e-01 -2.77508438e-01
-1.57289660e+00 -4.28546071e-01 5.26133418e-01 -2.04576939e-01
7.64360785e-01 1.46414682e-01 6.18337572e-01 1.00476789e+00
1.76744953e-01 6.83101058e-01 5.60526133e-01 -1.41768172e-01
2.86156703e-02 -6.18937798e-02 2.79323440e-02 9.16426420e-01
2.38093585e-01 -2.09558487e-01 -1.69018060e-01 -2.06642039e-02
1.31041825e+00 -6.27120212e-02 -1.16519161e-01 -2.03391016e-01
-6.65131569e-01 8.65698814e-01 9.33061957e-01 -5.28696179e-02
-2.27510989e-01 6.84704006e-01 3.35194528e-01 1.51242971e-01
5.72614431e-01 5.40695667e-01 -7.76337862e-01 2.49056935e-01
-5.05111873e-01 3.42719346e-01 5.44969559e-01 9.40141559e-01
6.42490387e-01 1.39507219e-01 -3.33201081e-01 1.35012829e+00
3.30673844e-01 3.03020835e-01 2.49304116e-01 -9.36892629e-01
4.71127272e-01 5.91063321e-01 -2.42151976e-01 -1.02759361e+00
-6.14565074e-01 -7.01388955e-01 -1.05551231e+00 1.53756186e-01
7.90844560e-01 -3.16604078e-01 -1.11485708e+00 1.87616670e+00
2.65002131e-01 1.60639405e-01 -2.12274700e-01 9.72306073e-01
8.47138226e-01 4.79122162e-01 8.76061842e-02 1.63578570e-01
1.03777528e+00 -1.07720900e+00 -5.49247921e-01 -4.54606980e-01
6.32085323e-01 -4.01365310e-01 1.29904807e+00 3.68368119e-01
-9.80702281e-01 -6.18039846e-01 -1.07217979e+00 -1.80710435e-01
-2.08605185e-01 5.92182636e-01 5.56836426e-01 4.30330753e-01
-1.05129886e+00 8.18933249e-01 -1.04917324e+00 1.13944612e-01
7.48425484e-01 7.79280961e-01 -3.19377273e-01 1.20774679e-01
-1.01067388e+00 6.95936441e-01 2.11835146e-01 6.88403845e-01
-5.68215072e-01 -8.83428991e-01 -6.69109404e-01 -2.70421747e-02
2.93590248e-01 -5.63181698e-01 1.35228503e+00 -1.23241591e+00
-1.65581024e+00 4.75561321e-01 5.27511872e-02 -4.95548666e-01
5.62755585e-01 -7.21532702e-01 -1.79674715e-01 -1.93880305e-01
-3.89929831e-01 1.01596427e+00 1.11523139e+00 -1.13127840e+00
-2.53480285e-01 -3.47831428e-01 9.52640027e-02 -1.14354901e-01
-7.03861296e-01 -8.93140286e-02 -5.66717923e-01 -7.78651655e-01
3.48560512e-01 -9.98789787e-01 -4.24073398e-01 4.84253079e-01
-1.70160711e-01 -3.82023692e-01 7.71332681e-01 -5.12935758e-01
1.42348170e+00 -2.12128639e+00 4.78752814e-02 2.46628746e-01
3.33376110e-01 4.65314180e-01 -2.16574505e-01 -4.37114030e-01
1.06926877e-02 2.26864234e-01 -9.42943171e-02 -4.01399195e-01
-3.21262240e-01 4.79209989e-01 1.44502579e-03 5.10354638e-01
5.49369156e-01 9.41198766e-01 -5.84217012e-01 -1.91338554e-01
-1.92034006e-01 7.86678672e-01 -7.76367784e-01 1.23152941e-01
-1.71673909e-01 1.58816352e-01 -5.65472126e-01 4.82033163e-01
6.97035551e-01 -1.00832999e-01 -2.67106533e-01 -4.70055193e-01
2.62106732e-02 1.05067387e-01 -9.58227098e-01 1.26550925e+00
-4.85788077e-01 7.23260641e-01 1.70020089e-01 -9.33071136e-01
1.31194735e+00 -3.66264582e-03 3.21466625e-01 -3.64202023e-01
4.67860937e-01 2.57617444e-01 2.53864199e-01 -2.78445363e-01
2.17971504e-01 1.97809502e-01 2.58528024e-01 -2.12449238e-01
-4.66838991e-03 3.60869974e-01 -2.63610423e-01 -4.54907835e-01
7.63367534e-01 9.12288502e-02 -1.80296853e-01 -4.36660349e-01
5.70814252e-01 -5.85325897e-01 7.44277656e-01 6.55540109e-01
-2.15430796e-01 8.82788360e-01 6.21321201e-01 -7.97749698e-01
-1.11068857e+00 -7.83303440e-01 -3.54633719e-01 1.17678666e+00
-2.86369473e-01 -1.95944831e-01 -1.03900492e+00 -9.48149145e-01
6.28106594e-02 1.15625612e-01 -9.00811970e-01 -2.45005220e-01
-1.07109034e+00 -6.15370870e-01 9.32016730e-01 1.17105329e+00
7.33643174e-01 -7.69294560e-01 -3.91435802e-01 1.62353471e-01
2.20071569e-01 -1.24194884e+00 -7.40798593e-01 3.61976385e-01
-1.02551937e+00 -7.78645098e-01 -8.67083430e-01 -8.05348516e-01
1.13957286e+00 -2.58039474e-01 8.16333830e-01 9.21423510e-02
1.58729434e-01 -1.02274798e-01 -5.49414642e-02 -6.36777222e-01
-7.69128799e-02 5.22633731e-01 2.07417160e-02 8.52194130e-02
4.85347100e-02 -4.95179057e-01 -6.87292516e-01 6.01841748e-01
-8.88464212e-01 -9.35340151e-02 5.72264850e-01 8.19675684e-01
6.63196504e-01 -4.63336051e-01 5.92228711e-01 -8.26256573e-01
4.84146088e-01 -1.18640080e-01 -7.13054597e-01 1.12035431e-01
-4.30985272e-01 3.29153031e-01 9.45582807e-01 -9.92125988e-01
-6.55984223e-01 2.11865157e-01 -2.78305501e-01 -9.16383684e-01
1.48924083e-01 3.47915381e-01 -1.43548744e-02 -4.78329390e-01
9.67911243e-01 -3.83142442e-01 2.06080750e-01 -4.68358070e-01
6.95735738e-02 3.51659060e-01 5.87538600e-01 -7.19980717e-01
6.08689129e-01 2.08716705e-01 3.06375235e-01 -5.86701930e-01
-1.20524728e+00 -1.32489264e-01 -5.94212532e-01 -3.42812270e-01
5.36040485e-01 -7.62133300e-01 -7.04823971e-01 6.46331966e-01
-1.14418280e+00 -5.86640358e-01 -7.43391961e-02 4.44885939e-01
-1.82946011e-01 -2.73630887e-01 -5.07760525e-01 -7.39736199e-01
-3.81257981e-01 -1.04387176e+00 7.50836849e-01 3.45014840e-01
-1.30325809e-01 -9.56388295e-01 -4.82012779e-01 -4.41223793e-02
5.95575392e-01 3.70491564e-01 5.69548190e-01 -6.21012151e-01
-8.84867087e-02 -2.83235669e-01 -2.26567641e-01 9.03497636e-01
3.67036532e-03 2.13134199e-01 -1.13931429e+00 -2.50577927e-01
-9.19103548e-02 -2.69120842e-01 1.00859487e+00 8.30049038e-01
1.78627586e+00 -5.80733061e-01 -1.89572975e-01 1.15006804e+00
1.08037734e+00 2.71129627e-02 7.57833898e-01 5.41147530e-01
9.43303049e-01 4.02431607e-01 3.07507247e-01 1.38045028e-01
1.44658014e-01 5.15483856e-01 5.04447341e-01 -2.00759888e-01
-1.00580547e-02 -1.62808731e-01 3.00604254e-01 3.90525281e-01
-4.55713838e-01 1.03407755e-01 -8.85143936e-01 2.53166825e-01
-1.83565974e+00 -2.52187878e-01 -2.51451105e-01 2.03816581e+00
9.45941389e-01 3.65636736e-01 8.55624229e-02 -1.10058725e-01
5.36750495e-01 -3.90453450e-02 -7.24795043e-01 -4.79358912e-01
2.70907953e-02 2.48665541e-01 1.01814198e+00 2.87655175e-01
-1.24632430e+00 8.70809197e-01 6.25964355e+00 8.64182293e-01
-1.35108852e+00 -1.89464942e-01 8.53673160e-01 -9.31993201e-02
1.62171852e-03 -3.50007862e-01 -1.13241100e+00 2.49438226e-01
7.73530245e-01 5.78339577e-01 5.01256466e-01 1.14456928e+00
8.51017088e-02 4.05862331e-01 -1.05173504e+00 7.21402168e-01
-3.11243087e-01 -1.17972267e+00 1.85584687e-02 -1.23631872e-01
7.66277492e-01 7.31980205e-02 1.27826914e-01 4.13749695e-01
-7.91976377e-02 -1.43634856e+00 8.16732824e-01 4.98067379e-01
8.14935148e-01 -9.65165079e-01 8.76298964e-01 2.25526884e-01
-1.15104699e+00 -2.16054171e-01 -5.63039422e-01 4.97724712e-02
-4.08004314e-01 4.63956237e-01 -7.91212738e-01 5.40812351e-02
7.68065810e-01 4.78433520e-01 -6.70549154e-01 9.45594728e-01
-1.76661447e-01 5.50794482e-01 -5.42857349e-01 -2.47524202e-01
3.08115274e-01 -1.96648300e-01 1.06122933e-01 1.17532158e+00
1.68351233e-01 -3.46827693e-03 -1.29592359e-01 8.56303334e-01
-3.56419891e-01 7.79525423e-03 -1.79239497e-01 1.24925561e-02
4.82820958e-01 1.18618345e+00 -5.30165851e-01 1.10344410e-01
-2.46722057e-01 4.51394111e-01 6.65357828e-01 4.87020969e-01
-1.05337274e+00 -4.66093630e-01 7.86285579e-01 5.07530034e-01
2.15602651e-01 -2.30858862e-01 -7.00298131e-01 -8.46969843e-01
3.00433099e-01 -6.85402513e-01 2.13660076e-01 -4.09552515e-01
-1.20749915e+00 7.28966832e-01 -2.84886882e-02 -9.62323487e-01
1.68724358e-01 -1.05289650e+00 -4.59741116e-01 7.18876421e-01
-1.32933581e+00 -1.16650641e+00 -4.32761252e-01 5.28050184e-01
2.92573303e-01 -1.35008156e-01 5.98970771e-01 3.96581084e-01
-8.02001595e-01 1.12296200e+00 -2.20831275e-01 5.56288838e-01
6.30923688e-01 -9.82010961e-01 5.03135741e-01 5.01968443e-01
-3.74577701e-01 8.17238688e-01 5.78860164e-01 -3.49018186e-01
-9.59853888e-01 -1.30292475e+00 4.89421964e-01 -1.04678825e-01
5.12863219e-01 -4.00398165e-01 -1.20364594e+00 7.93386102e-01
-4.08036023e-01 6.93653703e-01 2.84102827e-01 2.12944463e-01
-5.12663662e-01 -4.45328146e-01 -1.08543301e+00 5.90217590e-01
1.09083390e+00 -1.14269070e-01 1.07735489e-02 -5.78195453e-02
7.46094227e-01 -8.92006516e-01 -1.00072181e+00 8.46214592e-01
9.34023082e-01 -5.80634236e-01 1.11442411e+00 -7.62196779e-01
5.48696995e-01 -1.67741582e-01 1.93730630e-02 -1.29195058e+00
-2.17754140e-01 -5.12319505e-01 -5.36015332e-01 9.35686290e-01
5.57439744e-01 -7.23914444e-01 8.70715737e-01 7.88911760e-01
-4.08422887e-01 -1.56445718e+00 -7.89373338e-01 -9.30873811e-01
4.23690766e-01 -3.56166452e-01 5.02607763e-01 6.66366279e-01
-6.88205183e-01 1.34044603e-01 -1.33877799e-01 1.96588740e-01
3.19961935e-01 -8.82575035e-01 6.61335468e-01 -1.35966253e+00
-1.25322565e-01 -6.33639157e-01 -2.13318452e-01 -1.25622809e+00
3.01157176e-01 -5.88944256e-01 1.60201013e-01 -1.19764960e+00
-2.97011763e-01 -8.06752086e-01 -2.89327949e-01 7.77922213e-01
-1.81904942e-01 1.64342836e-01 3.27716544e-02 3.17826271e-02
-1.01010583e-01 5.50384939e-01 1.52767015e+00 7.26707131e-02
-3.42417836e-01 2.21555904e-01 -5.24525583e-01 1.03150463e+00
9.96280491e-01 -5.50025761e-01 -4.09201711e-01 -6.88620150e-01
1.80789232e-01 -5.20672619e-01 4.33220744e-01 -9.21735287e-01
2.38157272e-01 -1.38933390e-01 8.01840901e-01 -3.34961206e-01
3.41596574e-01 -7.15290010e-01 -1.45426661e-01 3.43688965e-01
-5.19052565e-01 3.71061452e-02 4.18873131e-01 2.27917388e-01
-1.24522280e-02 -2.91600138e-01 1.12432122e+00 1.79262578e-01
-1.93803936e-01 5.87358356e-01 -2.36134473e-02 2.65196268e-03
6.43514752e-01 -1.88903496e-01 -2.29000956e-01 -1.10146709e-01
-5.57614326e-01 3.34833026e-01 9.01401490e-02 3.94547552e-01
6.94623411e-01 -1.51260793e+00 -5.07620811e-01 4.10002679e-01
-1.09291218e-01 7.07775474e-01 -1.06347419e-01 7.98425853e-01
-5.66189408e-01 1.95053935e-01 -1.16413020e-01 -5.54861069e-01
-1.06488025e+00 1.10928290e-01 8.47772777e-01 -1.24771550e-01
-4.26663429e-01 1.28037572e+00 2.12173983e-01 -6.11449659e-01
7.56444514e-01 -8.46924424e-01 -2.39674404e-01 -1.99105471e-01
3.29157412e-01 2.39716381e-01 1.87548727e-01 -3.21925074e-01
-3.53394300e-01 6.19294107e-01 -2.61579990e-01 2.02762663e-01
1.45177114e+00 2.83006668e-01 1.26837805e-01 3.04794133e-01
1.39751720e+00 -2.96326965e-01 -1.76688552e+00 -1.51481435e-01
-1.19371474e-01 -2.77026176e-01 2.89234728e-01 -6.36898279e-01
-1.44666064e+00 6.21982157e-01 4.57708508e-01 3.41525748e-02
1.11170363e+00 -9.72426608e-02 7.76556075e-01 5.59861958e-01
-6.67716265e-02 -1.25430465e+00 2.14109495e-01 8.15623581e-01
1.05019677e+00 -9.36716855e-01 -9.70300809e-02 -5.17669439e-01
-4.60130155e-01 1.54665148e+00 1.17871797e+00 -4.91265684e-01
6.47861421e-01 5.03966808e-01 1.00633867e-01 6.58853948e-02
-5.93095601e-01 2.54300207e-01 7.78397262e-01 4.45093304e-01
6.04599774e-01 -1.78670421e-01 -2.05986992e-01 9.34501290e-01
-2.50276685e-01 -7.75091723e-02 7.35735074e-02 7.35631883e-01
-3.39572757e-01 -8.16430688e-01 -3.19057137e-01 4.92413402e-01
-5.02232671e-01 -2.47288775e-02 -3.38438213e-01 8.79838109e-01
1.72921464e-01 4.16216373e-01 3.01131457e-01 -4.61379558e-01
6.37641847e-01 -7.71744549e-02 7.05175281e-01 -3.15455705e-01
-7.16477513e-01 1.30354194e-02 7.16373976e-03 -5.45226514e-01
-1.66992784e-01 -4.95229155e-01 -1.46754360e+00 -2.00965747e-01
-4.86448258e-01 -3.09046030e-01 7.28840828e-01 8.56069326e-01
1.16093315e-01 7.75957048e-01 5.56642532e-01 -6.10288382e-01
-8.61619234e-01 -1.11564469e+00 -2.57361770e-01 2.81178318e-02
6.12117290e-01 -7.80033648e-01 -2.57118911e-01 -1.78058892e-01] | [8.927109718322754, 2.491614580154419] |
ae219131-ebe1-48f8-adf8-98a662c07225 | physics-informed-machine-learning-techniques | 2205.07838 | null | https://arxiv.org/abs/2205.07838v1 | https://arxiv.org/pdf/2205.07838v1.pdf | Physics-informed machine learning techniques for edge plasma turbulence modelling in computational theory and experiment | Edge plasma turbulence is critical to the performance of magnetic confinement fusion devices. Towards better understanding edge turbulence in both theory and experiment, a custom-built physics-informed deep learning framework constrained by partial differential equations is developed to accurately learn turbulent fields consistent with the two-fluid theory from partial observations of electron pressure. This calculation is not otherwise possible using conventional equilibrium models. With this technique, the first direct quantitative comparisons of turbulent fields between electrostatic two-fluid theory and electromagnetic gyrokinetic modelling are demonstrated with good overall agreement found in magnetized helical plasmas at low normalized pressure. To translate these computational techniques to experimental fusion plasmas, a novel method to translate brightness measurements of HeI line radiation into local plasma fluctuations is demonstrated via a newly created deep learning framework that integrates neutral transport physics and collisional radiative theory for the $3^3 D - 2^3 P$ transition in atomic helium. Using fast camera data on the Alcator C-Mod tokamak, this thesis presents the first 2-dimensional time-dependent experimental measurements of the turbulent electron density, electron temperature, and neutral density in a fusion plasma using a single spectral line. With this experimentally inferred data, initial estimates of the 2-dimensional turbulent electric field consistent with drift-reduced Braginskii theory under the framework of an axisymmetric fusion plasma with purely toroidal field are calculated. The inclusion of atomic helium effects on particle and energy sources are found to strengthen correlations between the electric field and electron pressure while broadening turbulent field amplitudes which impact ${\bf E \times B}$ flows and shearing rates. | ['Abhilash Mathews'] | 2022-05-16 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-2.14269966e-01 -3.90377581e-01 3.85192961e-01 -1.43754676e-01
7.72990212e-02 -4.42915186e-02 9.42158043e-01 -3.73937696e-01
-5.87073326e-01 1.01328623e+00 -7.08062649e-02 -7.57476330e-01
-3.03018540e-01 -6.90882683e-01 -1.84339076e-01 -1.13328111e+00
-2.27473587e-01 1.14539540e+00 -3.91805053e-01 -5.99531770e-01
1.99802905e-01 9.78407204e-01 -1.30125403e+00 1.80569395e-01
3.60293657e-01 1.16240442e+00 1.46724274e-02 6.96087837e-01
-2.51536518e-01 9.31988895e-01 -1.42777368e-01 2.22508505e-01
3.85189235e-01 -5.29999018e-01 -7.08955288e-01 -1.42330050e-01
-3.53807807e-01 5.83567396e-02 -4.49802697e-01 9.25820231e-01
5.08290052e-01 6.12997591e-01 1.03005266e+00 -5.24756551e-01
-2.28888482e-01 6.52782202e-01 -6.72658145e-01 8.76307070e-01
-3.62966895e-01 5.48985779e-01 3.92929018e-01 -7.40283489e-01
6.47845805e-01 7.82362103e-01 5.63324630e-01 5.37020683e-01
-8.85779440e-01 -4.43936795e-01 -1.00250125e+00 -5.08292578e-02
-1.08779871e+00 -3.63120615e-01 7.93932140e-01 -9.44796205e-01
1.54760182e+00 2.83725709e-01 6.82568312e-01 5.12772441e-01
7.00528264e-01 -7.13318586e-02 1.29632294e+00 -7.36738801e-01
3.93459201e-01 4.97451007e-01 -1.36303660e-02 5.04510462e-01
1.04964584e-01 1.07886636e+00 -4.17291462e-01 -7.07223490e-02
5.83903134e-01 -7.79301703e-01 -4.45716321e-01 -3.64048809e-01
-5.91396093e-01 7.35782385e-01 2.58339345e-01 1.02406263e+00
-4.18645978e-01 -1.41041994e-01 7.04166174e-01 1.18260439e-02
5.77894151e-01 8.32984328e-01 -6.59362674e-01 -6.97085202e-01
-1.06786752e+00 5.55986643e-01 8.74433219e-01 2.57589132e-01
7.21403599e-01 5.67235351e-01 1.48697257e-01 6.78887144e-02
1.18616499e-01 7.29740560e-01 6.85364366e-01 -5.59964240e-01
-2.65194595e-01 9.35222134e-02 3.19235325e-01 -4.33034897e-01
-6.27385676e-01 -1.00313418e-01 -9.57817197e-01 6.38708413e-01
1.05498560e-01 -7.98476517e-01 -8.49559963e-01 1.15885651e+00
4.29206997e-01 8.36645160e-03 5.82723439e-01 1.13247776e+00
7.09088862e-01 6.02683961e-01 -2.00517714e-01 -6.87473297e-01
9.54822183e-01 -6.45093679e-01 -5.89650214e-01 -1.61114946e-01
9.18556750e-01 -5.27679920e-01 2.44220331e-01 -1.94940194e-01
-1.23667157e+00 -5.47800422e-01 -9.78816748e-01 1.19685851e-01
-3.03424239e-01 -5.82044661e-01 1.13065493e+00 7.77628124e-01
-9.73387599e-01 1.09723055e+00 -9.94304836e-01 2.33852193e-01
1.46127073e-02 2.48371363e-01 -3.09614539e-01 9.06476676e-01
-1.26668835e+00 7.82139719e-01 5.27767599e-01 4.88383509e-02
-3.49139571e-01 -1.06792653e+00 -6.26178741e-01 -1.27106044e-03
-3.81295025e-01 -7.03972280e-01 1.64052677e+00 -5.17415941e-01
-1.75220108e+00 7.47624695e-01 -3.21867108e-01 -9.10528421e-01
4.25284535e-01 -5.37139662e-02 -5.51927865e-01 2.85223752e-01
-3.04672588e-02 6.81140497e-02 3.56941074e-01 -1.11558855e+00
-3.29371601e-01 -2.31194943e-02 -5.74090123e-01 2.49794483e-01
3.48669112e-01 1.92437053e-01 4.59230334e-01 3.15859199e-01
-7.00647831e-02 -7.27530479e-01 -3.90785992e-01 -9.70763385e-01
-1.46862015e-01 -5.63483946e-02 1.29881668e+00 -3.09262246e-01
3.31849694e-01 -1.71050239e+00 1.93460897e-01 1.45304576e-01
2.13281021e-01 3.61869007e-01 9.51710522e-01 2.29514018e-01
-5.61236978e-01 -6.74341759e-03 -6.09840035e-01 -1.15800485e-01
-1.40663505e-01 -3.27894747e-01 -5.45782626e-01 6.02138281e-01
-3.90458643e-01 9.41367209e-01 -7.27122188e-01 1.58184886e-01
8.79233956e-01 6.17377758e-01 -4.10793245e-01 4.53538783e-02
-9.18036401e-02 9.35167611e-01 -1.89436257e-01 -5.95380366e-02
1.15310574e+00 2.26564616e-01 -1.48508415e-01 -3.52565348e-02
-7.37497807e-01 -7.92890564e-02 -5.67605734e-01 1.20252550e+00
-4.86618340e-01 7.03361750e-01 9.37407494e-01 -1.29949582e+00
9.28795218e-01 3.32238287e-01 8.77773821e-01 -9.00950611e-01
7.22762346e-01 2.73272365e-01 3.88374388e-01 -4.43099827e-01
7.65785575e-01 -1.48999882e+00 2.33551025e-01 1.94846183e-01
1.35675833e-01 -1.05371749e+00 3.50500876e-03 2.34644949e-01
8.74032676e-01 -5.58689460e-02 -1.42916784e-01 -1.17486703e+00
6.42727911e-01 -2.11757079e-01 1.70376778e-01 7.12804139e-01
-1.55310690e-01 3.51085253e-02 4.09582466e-01 -7.40500808e-01
-1.31053984e+00 -7.96892464e-01 -9.47264254e-01 9.24770683e-02
1.91718787e-01 -6.02055371e-01 -5.88228226e-01 6.67266995e-02
5.22720180e-02 8.06043148e-01 -3.88202131e-01 -2.81113923e-01
-3.42209011e-01 -1.38529575e+00 2.49124378e-01 3.29521328e-01
6.39407396e-01 -1.46037996e+00 -9.09716547e-01 -1.69001028e-01
4.66470450e-01 -1.32248342e+00 5.58463395e-01 6.55281961e-01
-3.86187226e-01 -9.67298508e-01 3.40525173e-02 -1.99635863e-01
4.06134278e-02 -6.44103229e-01 1.06059659e+00 2.64433343e-02
-7.79887259e-01 3.13665420e-01 -1.65520355e-01 -6.17488384e-01
-7.48661280e-01 -4.85303581e-01 9.18022513e-01 -6.36179268e-01
9.20808315e-01 -3.89838934e-01 -2.96775401e-01 -2.31452733e-01
-6.26215041e-01 1.34694010e-01 1.17717274e-01 7.25049555e-01
2.06255332e-01 8.66431773e-01 5.90763800e-02 -4.27938163e-01
5.45380771e-01 -2.51791835e-01 -7.02873707e-01 -6.38045073e-01
-5.13380826e-01 8.54222551e-02 7.98639059e-01 2.13034138e-01
-1.61853218e+00 -7.27134824e-01 -4.17369008e-01 -3.24693590e-01
-4.32174385e-01 3.24710488e-01 3.98602635e-01 -2.41061330e-01
9.79819655e-01 3.27935033e-02 -1.23533420e-01 -9.93690044e-02
-1.08456619e-01 6.42761588e-01 7.52357066e-01 -8.82882178e-01
7.54900098e-01 7.65292287e-01 5.20358980e-01 -1.54948974e+00
-5.36689758e-01 -4.62490141e-01 -7.64652371e-01 -1.02186911e-01
1.04648459e+00 -6.70507014e-01 -1.01016700e+00 4.94103014e-01
-7.25554824e-01 -5.34415305e-01 -8.68129253e-01 9.16472018e-01
-7.48860478e-01 5.81569076e-01 -9.93439853e-01 -1.22947001e+00
-6.63830519e-01 -1.26187229e+00 7.17289507e-01 4.50691849e-01
2.83925422e-02 -1.34718525e+00 2.16132641e-01 1.16207510e-01
7.97283828e-01 2.88564086e-01 8.26289535e-01 -1.26430228e-01
-3.74053083e-02 2.63217628e-01 -5.55091947e-02 3.08593333e-01
7.03752190e-02 -1.55937508e-01 -9.58740354e-01 -4.68259513e-01
1.17210543e+00 -2.51982838e-01 1.05811560e+00 9.36900795e-01
8.95342350e-01 5.82550168e-01 -6.84692740e-01 1.12871456e+00
1.45539999e+00 3.79141062e-01 1.04431599e-01 3.81435364e-01
6.09034777e-01 7.22642019e-02 -4.53367345e-02 5.74563861e-01
-6.13329828e-01 3.75687569e-01 -1.40197739e-01 -2.31299683e-01
2.80092001e-01 5.40297687e-01 1.22701749e-01 4.53797728e-01
-4.22467202e-01 1.48868680e-01 -9.68845010e-01 2.19806880e-01
-1.15574837e+00 -1.19617629e+00 -4.75260019e-01 1.96591997e+00
4.79456544e-01 4.09137994e-01 -6.27066731e-01 -1.80235561e-02
2.90393561e-01 -6.20378815e-02 -4.38639790e-01 -1.07423234e+00
-2.44854480e-01 9.40886021e-01 7.89508224e-01 1.07534862e+00
-7.25413382e-01 8.76330614e-01 6.65583229e+00 3.18912923e-01
-1.66759837e+00 1.15843154e-01 5.38605571e-01 -3.45496595e-01
-3.52738976e-01 1.35693982e-01 -7.17443049e-01 5.19649982e-01
1.34510732e+00 -6.45848811e-01 4.12718445e-01 4.73269284e-01
2.85544574e-01 -5.25481701e-01 -5.97536385e-01 1.08540487e+00
-3.73284340e-01 -1.83060765e+00 -4.64790106e-01 2.05263540e-01
8.22286189e-01 5.62884033e-01 -1.75479144e-01 5.47744155e-01
-1.12676702e-01 -8.65262210e-01 4.43083942e-01 5.48472643e-01
1.06478000e+00 -8.86057317e-01 8.03799808e-01 4.46353972e-01
-8.40866566e-01 1.08021311e-01 -6.30219579e-01 -5.57613969e-01
8.97168696e-01 1.05427945e+00 -8.88277650e-01 8.35329056e-01
9.96075630e-01 3.79060209e-01 1.14437692e-01 2.51168728e-01
2.93432474e-01 3.86403531e-01 -6.88493192e-01 1.72913864e-01
3.73994380e-01 -9.02261019e-01 6.79846585e-01 1.21241355e+00
3.65537107e-01 4.84927297e-01 -5.98262697e-02 1.03703296e+00
1.64414823e-01 -9.50106531e-02 -8.46008301e-01 -2.35305086e-01
-4.66954261e-01 1.75492096e+00 -1.02325249e+00 -2.69266218e-01
-4.76756804e-02 4.21234846e-01 3.12144738e-02 1.21252336e-01
-6.87096775e-01 -1.61947861e-01 9.04774010e-01 3.97488534e-01
1.82664141e-01 -4.26215827e-01 -2.46676803e-01 -1.10923028e+00
-5.40513515e-01 1.84839159e-01 -2.75806189e-01 -7.98667967e-01
-1.24990869e+00 8.15239072e-01 4.98482853e-01 -6.74471483e-02
-4.59664434e-01 -1.23583984e+00 -1.12863481e+00 1.55544567e+00
-1.34382629e+00 -3.21849048e-01 4.26373705e-02 2.94374853e-01
1.09387767e-02 -4.51291919e-01 7.45114207e-01 -1.53044671e-01
-1.28172129e-01 -3.06084216e-01 5.91113389e-01 -2.67217517e-01
2.43578643e-01 -1.53654659e+00 4.55592036e-01 7.65817404e-01
-3.87812048e-01 4.29429114e-01 1.39015222e+00 -8.41830075e-01
-1.24610007e+00 -5.13313413e-01 2.90509194e-01 -4.06725138e-01
8.21553528e-01 -4.25066918e-01 -7.51307487e-01 4.22996372e-01
7.38910735e-01 3.65077674e-01 5.61793983e-01 -3.54669057e-02
7.43542194e-01 4.62932289e-01 -1.34765875e+00 2.45863959e-01
7.25505829e-01 -7.82964945e-01 -6.43159032e-01 6.39604986e-01
3.46918643e-01 -6.70406222e-01 -5.49968779e-01 1.10795796e+00
-1.72161758e-01 -1.42040324e+00 4.14621502e-01 -7.07421362e-01
4.73410934e-02 -4.41677384e-02 3.51955555e-02 -1.35972893e+00
-1.40007526e-01 -8.01476002e-01 9.87920016e-02 6.66514933e-01
2.12391421e-01 -2.22022533e-01 1.00910556e+00 5.20643711e-01
-4.25870121e-01 -7.56936669e-01 -1.24344754e+00 -2.97515064e-01
7.62141585e-01 -6.26218617e-01 1.75873432e-02 8.74624014e-01
4.44416255e-01 2.67916590e-01 4.44855355e-02 2.03556299e-01
2.92232722e-01 1.23753674e-01 2.63555318e-01 -1.28673553e+00
-3.32329124e-01 -2.19143853e-01 -4.27442551e-01 -3.98033649e-01
7.99727142e-01 -1.06766903e+00 3.76327224e-02 -1.06607223e+00
-2.23104462e-01 -5.27583599e-01 6.85983747e-02 -3.93776983e-01
4.75026727e-01 -1.58306584e-01 -2.67763019e-01 -4.63850722e-02
2.67110288e-01 6.97116375e-01 7.24285364e-01 1.23853482e-01
-3.41789693e-01 -3.18087548e-01 -1.21888407e-01 7.98922062e-01
9.28628385e-01 2.71464199e-01 -2.97132969e-01 7.70524070e-02
7.55618140e-02 1.56934723e-01 3.25879641e-02 -1.69027114e+00
1.12833224e-01 1.55690670e-01 6.47658110e-01 -1.42978579e-01
3.28369200e-01 -4.36034471e-01 -2.99523249e-02 1.69692218e-01
1.74908891e-01 -1.66432466e-02 7.70915270e-01 7.83475712e-02
-1.85012192e-01 -3.85697186e-01 1.24428093e+00 -7.72341847e-01
-6.12623632e-01 3.73267204e-01 -6.55559540e-01 -1.99742138e-01
1.20141280e+00 2.29358405e-01 -5.19577973e-02 -7.16389567e-02
-6.78139746e-01 -2.37425685e-01 6.31602108e-01 1.10005051e-01
1.86010450e-01 -7.78087795e-01 -4.70795631e-01 6.62524462e-01
-6.71748817e-01 2.62755007e-01 7.98906624e-01 7.38690674e-01
-1.08347499e+00 5.88635445e-01 -2.21891865e-01 -7.26231635e-01
-7.86706448e-01 7.26464391e-01 1.32497728e+00 -3.12978387e-01
-8.52883577e-01 1.11417627e+00 2.67908663e-01 -4.89833623e-01
-7.74573505e-01 -9.07170624e-02 -1.01530328e-04 -5.22721589e-01
6.12657249e-01 -1.44367993e-01 4.85098749e-01 -1.15085220e+00
2.55516991e-02 1.97885364e-01 2.71130111e-02 -1.73160791e-01
1.11660540e+00 7.68405646e-02 -3.73959482e-01 3.30403447e-01
1.00826693e+00 1.77055880e-01 -1.17283475e+00 2.98565179e-01
-2.81758606e-01 -1.18347913e-01 7.37663388e-01 -5.86964130e-01
-9.20856416e-01 7.88664103e-01 6.58077955e-01 4.45489764e-01
6.42495096e-01 9.25630182e-02 4.70877498e-01 2.64848202e-01
1.86471045e-01 -1.07191980e+00 -6.42894506e-01 9.40769017e-01
3.87632757e-01 -1.01604259e+00 -6.07504584e-02 -6.58401847e-02
-5.01196146e-01 1.30003369e+00 1.04201841e+00 1.57211453e-01
1.25532579e+00 9.34877813e-01 1.00338355e-01 -5.26847482e-01
-6.11020386e-01 2.45062634e-01 -2.63635367e-01 2.24844232e-01
6.73132837e-01 9.30021778e-02 -2.97195137e-01 9.97691751e-02
-7.14088321e-01 -1.80622533e-01 7.52969205e-01 8.32441628e-01
-3.89633596e-01 -8.75990272e-01 -2.39870951e-01 4.14626062e-01
-6.83796823e-01 2.63555963e-02 3.76423955e-01 6.42324805e-01
3.47901642e-01 5.95130801e-01 6.44796550e-01 4.38978747e-02
-3.18128727e-02 5.96809506e-01 3.00601780e-01 -2.91483998e-01
-1.69843778e-01 -2.11979419e-01 -2.62371093e-01 1.77478287e-02
-6.25217199e-01 -4.97219503e-01 -1.76994443e+00 -4.46508497e-01
-2.83932269e-01 1.31289387e+00 9.14191186e-01 1.39448750e+00
-7.54280835e-02 5.44226110e-01 2.85578161e-01 -1.56909728e+00
2.32738443e-03 -1.13333166e+00 -1.38637078e+00 3.82043600e-01
3.58519226e-01 -8.53884339e-01 -1.02953732e+00 -5.60017467e-01] | [6.45668888092041, 3.506392478942871] |
94c7508c-5f9f-42c3-b2ac-bff39ec5df1a | a-novel-metric-for-evaluating-semantics | 2110.01176 | null | https://arxiv.org/abs/2110.01176v3 | https://arxiv.org/pdf/2110.01176v3.pdf | Contextualized Semantic Distance between Highly Overlapped Texts | Overlapping frequently occurs in paired texts in natural language processing tasks like text editing and semantic similarity evaluation. Better evaluation of the semantic distance between the overlapped sentences benefits the language system's understanding and guides the generation. Since conventional semantic metrics are based on word representations, they are vulnerable to the disturbance of overlapped components with similar representations. This paper aims to address the issue with a mask-and-predict strategy. We take the words in the longest common sequence (LCS) as neighboring words and use masked language modeling (MLM) from pre-trained language models (PLMs) to predict the distributions on their positions. Our metric, Neighboring Distribution Divergence (NDD), represent the semantic distance by calculating the divergence between distributions in the overlapped parts. Experiments on Semantic Textual Similarity show NDD to be more sensitive to various semantic differences, especially on highly overlapped paired texts. Based on the discovery, we further implement an unsupervised and training-free method for text compression, leading to a significant improvement on the previous perplexity-based method. The high scalability of our method even enables NDD to outperform the supervised state-of-the-art in domain adaption by a huge margin. Further experiments on syntax and semantics analyses verify the awareness of internal sentence structures, indicating the high potential of NDD for further studies. | ['Hai Zhao', 'Zuchao Li', 'Letian Peng'] | 2021-10-04 | null | null | null | null | ['predicate-detection', 'sentence-compression', 'text-compression'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 4.96994525e-01 4.43156287e-02 -1.16065763e-01 -5.48340976e-01
-5.05825162e-01 -3.32449526e-01 7.17219710e-01 6.25309885e-01
-4.75510329e-01 5.69461823e-01 6.25670254e-01 -3.14685106e-01
-1.71358109e-01 -7.89132833e-01 -3.54888052e-01 -4.26759183e-01
1.22010849e-01 5.67854047e-01 3.22825491e-01 -6.17049634e-01
7.99774170e-01 1.09947138e-01 -1.73695385e+00 7.03591585e-01
1.18868876e+00 5.82705498e-01 9.11162853e-01 3.10934395e-01
-1.06334472e+00 4.15744752e-01 -8.59530449e-01 -3.39765966e-01
-2.58872174e-02 -6.47725821e-01 -9.75729287e-01 -8.05879012e-02
1.16578251e-01 1.17700048e-01 -4.16977219e-02 1.20972538e+00
5.02874732e-01 2.18067914e-01 7.62848020e-01 -7.73358226e-01
-6.34910941e-01 9.69500184e-01 -5.58858812e-01 4.05236870e-01
6.94504917e-01 -4.75489259e-01 1.05563462e+00 -9.42321539e-01
7.16464460e-01 1.40394723e+00 6.37959719e-01 3.10089827e-01
-9.79756355e-01 -2.36592725e-01 8.44567493e-02 3.50642025e-01
-1.51425481e+00 -2.55313098e-01 5.00884354e-01 -2.61351615e-01
1.31167471e+00 4.03303653e-01 1.23439714e-01 7.79605508e-01
1.28125668e-01 6.15503252e-01 6.96186483e-01 -8.90357256e-01
1.42037854e-01 3.33738059e-01 1.03069916e-01 3.45823824e-01
1.74904630e-01 -4.54979628e-01 -3.99190903e-01 -1.67174056e-01
1.75062314e-01 -3.60297188e-02 -2.35236377e-01 4.16588411e-02
-1.20259714e+00 9.12281752e-01 -1.59222350e-01 9.00026619e-01
-2.58792549e-01 -3.93859953e-01 6.93475008e-01 3.68456751e-01
6.32713556e-01 5.53947389e-01 -4.71360743e-01 -2.11131528e-01
-1.27315748e+00 1.52960032e-01 7.91529417e-01 9.65071738e-01
6.32444441e-01 -4.72189039e-01 -3.35744828e-01 1.14968073e+00
9.21122879e-02 2.72942901e-01 1.24621189e+00 -4.68967527e-01
6.51062131e-01 7.40706801e-01 -3.31975967e-01 -1.18169677e+00
-1.64011076e-01 -2.42992535e-01 -6.98113084e-01 -3.05445701e-01
2.66546607e-01 1.76524833e-01 -5.29282808e-01 1.58104336e+00
1.21261939e-01 -9.82081965e-02 1.79632425e-01 6.39234900e-01
5.31009555e-01 8.31744611e-01 -5.08673191e-02 -4.06807840e-01
1.31247926e+00 -6.01003647e-01 -8.88770521e-01 -1.61764547e-01
1.15669858e+00 -1.28254902e+00 1.28625238e+00 1.23854868e-01
-8.80584002e-01 -5.08077204e-01 -9.92238045e-01 6.21763691e-02
-4.68990982e-01 1.37147263e-01 2.74019659e-01 6.01540446e-01
-8.26820910e-01 9.99083102e-01 -3.21875632e-01 -7.29493618e-01
-5.52146174e-02 1.63547155e-02 -1.43733591e-01 1.74282283e-01
-1.48144734e+00 1.00877225e+00 7.44797468e-01 -4.46764380e-01
1.01469666e-01 -6.21607184e-01 -7.14858770e-01 9.99762267e-02
2.24617824e-01 -3.95965189e-01 9.65869129e-01 -1.09937954e+00
-1.29687178e+00 1.04145074e+00 -5.19923508e-01 -5.70608377e-01
2.14477286e-01 -5.30632026e-03 -6.27603352e-01 1.21572703e-01
2.69585729e-01 4.44596648e-01 6.64940000e-01 -7.70939052e-01
-4.82472539e-01 -2.35250503e-01 -5.50510526e-01 3.97573918e-01
-6.46316588e-01 2.54910976e-01 -2.30133221e-01 -1.00564051e+00
2.98452646e-01 -6.12206578e-01 -1.34860948e-02 -2.68317163e-01
-2.05864713e-01 -4.64157850e-01 6.62845612e-01 -8.71633410e-01
1.82507432e+00 -2.22594881e+00 1.48815680e-02 3.00234139e-01
-8.20110962e-02 3.75666946e-01 -2.45849386e-01 8.89890492e-01
1.49368374e-02 3.85072678e-02 -4.01735187e-01 -2.14511275e-01
5.56572713e-03 2.82378346e-01 -5.38209140e-01 1.43489540e-01
1.15631394e-01 5.54979801e-01 -8.02500010e-01 -7.10345685e-01
2.38370728e-02 1.19675435e-02 -3.27208310e-01 1.17530271e-01
-2.72165269e-01 -1.81734841e-02 -2.14339152e-01 2.18070120e-01
6.03543580e-01 -5.70772178e-02 4.67699558e-01 -3.09104566e-02
-1.16755383e-03 7.83871293e-01 -1.17128837e+00 1.89753258e+00
-5.35607457e-01 5.39700389e-01 -5.47469318e-01 -1.15299118e+00
1.36739397e+00 1.11375287e-01 1.46501243e-01 -8.99287164e-01
-1.03291102e-01 3.97724837e-01 4.18902226e-02 -6.13913119e-01
1.05679405e+00 -1.26775727e-01 -1.37775004e-01 5.22916377e-01
-1.10174961e-01 -1.23997875e-01 4.04999137e-01 4.89352942e-01
8.84461582e-01 -1.83260769e-01 5.95446348e-01 -4.21128750e-01
7.65344799e-01 -2.03827560e-01 1.98411450e-01 6.09367669e-01
7.96796307e-02 6.09818995e-01 5.76795220e-01 2.89691091e-02
-1.19731188e+00 -9.42299068e-01 -1.86287418e-01 1.14952159e+00
2.55915761e-01 -7.51781762e-01 -8.54136467e-01 -6.31713271e-01
-2.51943246e-03 1.33608186e+00 -2.83086121e-01 -2.44473070e-01
-5.60384512e-01 -6.82071090e-01 6.54623866e-01 4.55798805e-01
5.71346842e-02 -9.72463906e-01 -1.65090814e-01 2.95804769e-01
-4.17895615e-01 -9.71554399e-01 -5.46530187e-01 3.05823013e-02
-9.34215963e-01 -7.19956875e-01 -6.54307961e-01 -9.35729027e-01
5.11510372e-01 3.21207076e-01 9.46348011e-01 -2.52452306e-02
-1.64732441e-01 4.14730720e-02 -7.34242916e-01 -3.00030619e-01
-8.40142071e-01 4.45629619e-02 1.03421785e-01 -2.20820785e-01
7.38778293e-01 -6.02134466e-01 -2.77649850e-01 2.50171840e-01
-1.01414609e+00 -1.13686807e-01 5.53563774e-01 6.92717135e-01
4.33549196e-01 7.72769377e-02 6.30858660e-01 -9.06170845e-01
1.08012760e+00 -6.39228463e-01 -1.60635859e-01 4.57001299e-01
-9.15426612e-01 5.80164254e-01 7.79546618e-01 -4.97657001e-01
-1.15599108e+00 -3.61827105e-01 -5.10203242e-02 -1.35699296e-02
-1.01246767e-01 4.98458862e-01 -5.89537285e-02 5.07347286e-01
6.52845025e-01 5.74724495e-01 1.92998439e-01 -6.58354938e-01
4.40761536e-01 1.04155707e+00 3.49418789e-01 -3.56584907e-01
3.50865632e-01 2.15123624e-01 -3.61247122e-01 -9.78091836e-01
-6.20355785e-01 -7.45589316e-01 -5.87873340e-01 1.07504576e-01
5.50232410e-01 -6.91197217e-01 -1.66763097e-01 6.69474676e-02
-1.41144264e+00 3.16434234e-01 -2.61210054e-01 5.15808821e-01
-1.83701679e-01 1.06060386e+00 -3.70993942e-01 -5.51789105e-01
-3.69640619e-01 -7.09623754e-01 1.02042353e+00 -3.05279456e-02
-8.09958577e-01 -1.09827483e+00 2.18435347e-01 2.20477417e-01
4.57143426e-01 -4.27338153e-01 1.25962448e+00 -1.32022345e+00
1.28752962e-01 -1.57617167e-01 -6.87551498e-02 5.15779197e-01
1.51047915e-01 -1.46664619e-01 -6.05654418e-01 5.89750055e-03
1.00736998e-01 6.53980896e-02 8.36516619e-01 1.13776557e-01
1.25527871e+00 -2.39090711e-01 -3.04849714e-01 1.51065364e-01
1.12455964e+00 7.67536238e-02 7.67029643e-01 3.70217770e-01
3.25270861e-01 9.18034494e-01 8.74283075e-01 7.56009400e-01
2.94591393e-02 6.84012711e-01 4.63523529e-02 2.12582827e-01
-5.95015921e-02 -4.08168703e-01 4.73203182e-01 1.36005652e+00
4.49908227e-01 -4.75521892e-01 -7.26585984e-01 5.03248513e-01
-1.65959764e+00 -1.00536942e+00 -3.02496433e-01 2.33473468e+00
1.04154444e+00 1.86616004e-01 -1.98690534e-01 2.75262982e-01
1.01870430e+00 3.15687865e-01 -4.54263613e-02 -8.73677552e-01
-3.62091511e-01 2.42877513e-01 3.57951850e-01 4.63719308e-01
-6.24057114e-01 9.46432114e-01 5.72366858e+00 1.52353883e+00
-8.33421111e-01 -5.36016561e-02 2.95722067e-01 2.50632279e-02
-6.99059904e-01 8.69662985e-02 -9.19965684e-01 8.72100472e-01
1.00594366e+00 -4.61444616e-01 1.70854524e-01 6.95741594e-01
3.09769243e-01 -3.21234345e-01 -8.66817355e-01 8.86002004e-01
4.74372208e-01 -1.19158173e+00 4.42212015e-01 -2.23861441e-01
6.15942061e-01 -2.81962723e-01 -2.58917242e-01 1.07022658e-01
-2.58137375e-01 -8.01054120e-01 5.79353273e-01 5.08408129e-01
6.71912193e-01 -7.12566376e-01 9.60180521e-01 5.00596881e-01
-9.83598351e-01 2.77484477e-01 -8.49573791e-01 -1.41250297e-01
1.68666333e-01 9.57565665e-01 -1.15015817e+00 4.76703972e-01
2.72067428e-01 6.19131625e-01 -5.09898722e-01 6.27274454e-01
-1.35335466e-02 5.04873455e-01 -1.86453432e-01 -5.11629462e-01
1.87662110e-01 -3.86784911e-01 7.05260515e-01 1.60375059e+00
5.92381656e-01 -2.25834921e-01 -1.62352592e-01 8.66005480e-01
9.55945104e-02 7.25726604e-01 -5.41500986e-01 -2.79884726e-01
8.01358759e-01 9.22577500e-01 -7.57647693e-01 -5.42192161e-01
-3.79036427e-01 1.42130411e+00 2.31973156e-01 -1.21252343e-01
-6.24331713e-01 -7.09830225e-01 5.64499736e-01 1.20914787e-01
4.41223949e-01 -2.33567767e-02 -4.42959279e-01 -9.17766213e-01
2.75905788e-01 -7.77334571e-01 3.63145500e-01 -5.07717609e-01
-1.50703013e+00 6.26504660e-01 1.97296441e-02 -1.40211487e+00
-3.63905758e-01 -3.63370776e-01 -6.02419794e-01 9.23473716e-01
-1.12397194e+00 -4.08035606e-01 1.64033741e-01 4.76138234e-01
8.04192960e-01 -4.33706999e-01 9.60799515e-01 2.52941102e-01
-1.86374828e-01 7.29882240e-01 5.36034226e-01 -1.46019846e-01
8.48920643e-01 -1.11170197e+00 3.66266340e-01 7.76360571e-01
3.33418161e-01 6.37184441e-01 7.63223767e-01 -7.95330465e-01
-8.43536377e-01 -8.82517755e-01 1.81224811e+00 -1.19101599e-01
7.00345576e-01 -2.99752057e-01 -1.18250453e+00 1.23810217e-01
2.01969966e-01 -7.17381895e-01 8.36429954e-01 4.83294353e-02
-3.82515103e-01 6.58116341e-02 -9.13887680e-01 6.39444709e-01
1.04212987e+00 -6.35707498e-01 -1.33599079e+00 4.49924111e-01
9.90343988e-01 9.83978361e-02 -5.49857199e-01 1.15554564e-01
1.39482170e-01 -1.11248195e+00 6.89893126e-01 -3.93180609e-01
5.77260852e-01 -2.02796519e-01 -3.77717227e-01 -1.22520781e+00
-5.21880165e-02 -4.71763015e-01 2.16122150e-01 1.56091750e+00
4.73262280e-01 -5.88235974e-01 4.30407524e-01 1.45145431e-01
-1.39987409e-01 -5.65029621e-01 -8.35121453e-01 -9.97238815e-01
-1.38249537e-02 -4.54012185e-01 6.93152428e-01 1.02424216e+00
5.16655266e-01 3.39834332e-01 -9.18789357e-02 -1.49391919e-01
8.20070803e-02 4.84985597e-02 1.95740804e-01 -9.84654129e-01
-3.58848244e-01 -5.72995663e-01 -5.23708344e-01 -1.15586436e+00
3.52094591e-01 -1.28591239e+00 -1.64915936e-03 -1.15877903e+00
2.85573751e-01 -1.74550533e-01 -4.77426089e-02 1.45813981e-02
-2.67118871e-01 -2.00340673e-01 3.85106951e-02 2.71889985e-01
-6.50352776e-01 8.07721019e-01 8.45377088e-01 5.82020096e-02
-2.78830200e-01 -8.78157765e-02 -5.86230159e-01 6.84555292e-01
8.70769918e-01 -5.91192305e-01 -4.07860845e-01 -2.19557419e-01
1.79371938e-01 -2.56575942e-01 -1.44713461e-01 -8.21686685e-01
1.95969909e-01 6.68734894e-04 3.98462117e-02 -6.95734501e-01
-6.72929883e-02 -6.49136543e-01 -2.14938909e-01 5.19686580e-01
-7.23047435e-01 2.97101527e-01 3.95376096e-03 3.77676457e-01
-5.07575989e-01 -7.71404803e-01 4.17907953e-01 -6.88439161e-02
-8.34969282e-01 -3.34891796e-01 -6.15531206e-01 3.00606638e-01
8.81944299e-01 -4.83847290e-01 -1.07084237e-01 -3.17121625e-01
-3.17473561e-01 -1.73657969e-01 3.49073470e-01 5.34214795e-01
6.75935686e-01 -1.11336565e+00 -7.63362586e-01 3.75191033e-01
2.42656201e-01 -4.23900157e-01 1.33580998e-01 7.01861918e-01
-4.71428812e-01 4.38567609e-01 8.98786411e-02 -3.91526192e-01
-1.54438782e+00 4.74530309e-01 -1.56410728e-02 -3.12007427e-01
-4.58762169e-01 7.38233447e-01 -5.02402671e-02 -5.32797635e-01
2.69587904e-01 -3.40526760e-01 -2.73969740e-01 2.05444396e-01
6.21934116e-01 5.13256133e-01 2.12426648e-01 -4.83531952e-01
-3.68340313e-01 3.14554513e-01 -3.03481072e-01 -1.50533780e-01
1.04827094e+00 -5.84315836e-01 -4.92812008e-01 4.36708838e-01
1.40395868e+00 2.62398660e-01 -4.37573016e-01 -3.94076526e-01
5.37147641e-01 -5.61063826e-01 -1.70912832e-01 -6.03962421e-01
-3.90525669e-01 7.90986121e-01 3.33128333e-01 2.33079031e-01
9.76809382e-01 2.36238644e-01 1.01082003e+00 3.26595336e-01
-2.02275589e-02 -1.53080320e+00 5.56465276e-02 8.17249358e-01
7.61138976e-01 -9.26655114e-01 -5.19762449e-02 -4.23802465e-01
-9.71420109e-01 1.27768672e+00 2.54204035e-01 2.57221043e-01
4.01410669e-01 1.86327517e-01 5.16203046e-03 4.06884067e-02
-6.55968964e-01 -1.20965175e-01 3.86159003e-01 3.67738485e-01
7.84220576e-01 4.78529297e-02 -1.03922212e+00 4.94032025e-01
-4.68379140e-01 -4.60639119e-01 3.26886386e-01 7.34037459e-01
-7.74737418e-01 -1.50557590e+00 -3.47357959e-01 5.34555137e-01
-3.41772020e-01 -5.92422068e-01 -3.84039342e-01 3.91457081e-01
1.60314113e-01 9.46580887e-01 3.47239316e-01 -3.72873485e-01
1.17841728e-01 3.46429586e-01 3.31688881e-01 -6.15211964e-01
-4.24616814e-01 -8.08768421e-02 1.23003803e-01 -4.70083386e-01
-2.40380883e-01 -8.56247365e-01 -1.53682053e+00 -3.54614526e-01
-4.09231693e-01 2.03615189e-01 7.67899096e-01 1.11969662e+00
4.49600428e-01 2.88525760e-01 6.56117916e-01 -3.55402082e-01
-9.82872367e-01 -1.14202785e+00 -8.23095262e-01 9.13443685e-01
-2.61799455e-01 -2.42120191e-01 -4.47308928e-01 -1.07671902e-01] | [10.975997924804688, 8.962203025817871] |
6e0704e9-5b7f-4d61-8a88-211bf12f4f98 | accurate-and-efficient-event-based-semantic | 2304.11857 | null | https://arxiv.org/abs/2304.11857v2 | https://arxiv.org/pdf/2304.11857v2.pdf | Accurate and Efficient Event-based Semantic Segmentation Using Adaptive Spiking Encoder-Decoder Network | Leveraging the low-power, event-driven computation and the inherent temporal dynamics, spiking neural networks (SNNs) are potentially ideal solutions for processing dynamic and asynchronous signals from event-based sensors. However, due to the challenges in training and the restrictions in architectural design, there are limited examples of competitive SNNs in the realm of event-based dense prediction when compared to artificial neural networks (ANNs). In this paper, we present an efficient spiking encoder-decoder network designed for large-scale event-based semantic segmentation tasks. This is achieved by optimizing the encoder using a hierarchical search method. To enhance learning from dynamic event streams, we harness the inherent adaptive threshold of spiking neurons to modulate network activation. Moreover, we introduce a dual-path Spiking Spatially-Adaptive Modulation (SSAM) block, specifically designed to enhance the representation of sparse events, thereby considerably improving network performance. Our proposed network achieves a 72.57% mean intersection over union (MIoU) on the DDD17 dataset and a 57.22% MIoU on the recently introduced, larger DSEC-Semantic dataset. This performance surpasses the current state-of-the-art ANNs by 4%, whilst consuming significantly less computational resources. To the best of our knowledge, this is the first study demonstrating SNNs outperforming ANNs in demanding event-based semantic segmentation tasks, thereby establishing the vast potential of SNNs in the field of event-based vision. Our source code will be made publicly accessible. | ['Ran Cheng', 'Jiangxing Liao', 'Qinghai Guo', 'Jie Cheng', 'Hu Zhang', 'Kaiwei Che', 'Luziwei Leng', 'Rui Zhang'] | 2023-04-24 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 8.36948037e-01 -1.82326198e-01 2.80268788e-01 -1.66353613e-01
-5.86891294e-01 -2.94696569e-01 4.48579699e-01 3.50666702e-01
-8.79116178e-01 9.08277929e-01 -1.59635589e-01 1.02799706e-01
-1.10431470e-01 -6.96671724e-01 -9.90717947e-01 -9.32062626e-01
-3.50387484e-01 1.95791826e-01 7.52633750e-01 3.99173656e-03
2.77565062e-01 6.56029880e-01 -1.92663491e+00 6.16717219e-01
5.89668274e-01 1.45819020e+00 5.68607807e-01 4.88699347e-01
-5.07301204e-02 7.50418425e-01 -5.02847850e-01 1.22707464e-01
-4.42429744e-02 -4.78851229e-01 -2.06163630e-01 -4.08107609e-01
8.20994750e-02 2.00616401e-02 -4.39179927e-01 5.72504222e-01
8.96658421e-01 -1.26233026e-01 3.62102598e-01 -9.90373433e-01
-7.20268264e-02 6.55754626e-01 -9.84426811e-02 6.36152685e-01
-4.07579094e-02 2.14627743e-01 6.99952900e-01 -8.27945530e-01
7.49950111e-01 6.76590145e-01 7.68170118e-01 6.47676885e-01
-1.46638072e+00 -9.41727817e-01 2.33482774e-02 3.77021655e-02
-1.19321346e+00 -5.77997625e-01 5.26872635e-01 -2.02645108e-01
1.59854352e+00 -1.96998894e-01 1.15961170e+00 1.26477218e+00
3.06975693e-01 7.89377272e-01 1.19300687e+00 7.02763125e-02
7.73053467e-01 -5.89727700e-01 -1.48318246e-01 3.15455139e-01
1.96098283e-01 1.89042047e-01 -1.20655835e+00 8.94642472e-02
8.77191663e-01 6.09895438e-02 2.38928553e-02 1.38843089e-01
-1.30962741e+00 4.03429449e-01 5.18975794e-01 2.64094591e-01
-6.60208642e-01 5.87920547e-01 4.31209326e-01 -6.86496198e-02
3.32470872e-02 4.68283415e-01 -4.89479929e-01 -3.42838436e-01
-1.20242190e+00 2.68682420e-01 8.23062122e-01 5.55163324e-01
4.26971674e-01 4.22960788e-01 1.99634880e-02 5.63203990e-01
1.51799530e-01 3.72644514e-01 4.71853465e-01 -9.96770561e-01
-1.57363843e-02 7.19977438e-01 -2.77966261e-01 -4.55313087e-01
-6.86450183e-01 -5.13327360e-01 -8.51522446e-01 3.94281834e-01
4.58072871e-01 1.35451257e-01 -1.04828370e+00 1.72196996e+00
-8.99133533e-02 5.06995022e-01 9.03720632e-02 7.02110529e-01
6.61373556e-01 6.60423100e-01 3.96237671e-01 -3.27490002e-01
1.23687601e+00 -1.86491817e-01 -4.97623950e-01 -5.07968128e-01
1.46818295e-01 -3.69961828e-01 4.96570200e-01 4.05902565e-01
-1.16452134e+00 -2.02156022e-01 -1.20515859e+00 1.26049280e-01
-4.02511090e-01 -6.10921271e-02 8.87182951e-01 2.41767436e-01
-1.03475034e+00 5.83690166e-01 -1.37552226e+00 -5.16701400e-01
1.08656812e+00 8.50360453e-01 -1.21867657e-02 4.10056710e-01
-9.99860287e-01 5.32296598e-01 5.36666453e-01 -1.19417191e-01
-1.09551799e+00 -8.29259932e-01 -4.38013881e-01 -3.19629721e-02
-1.94001719e-01 -4.63260442e-01 1.14118540e+00 -7.68335819e-01
-1.44184434e+00 9.65414166e-01 -3.91601980e-01 -1.19353449e+00
-2.79718097e-02 2.08561480e-01 -1.08346440e-01 4.75993603e-01
-9.43325832e-02 1.20060349e+00 4.11461920e-01 -8.30490232e-01
-6.15240991e-01 -5.16918778e-01 -3.76950175e-01 -1.96994301e-02
-3.27582985e-01 -6.03735633e-02 -1.74426556e-01 -7.94480443e-01
2.00935543e-01 -8.38851750e-01 -3.62687051e-01 3.06428969e-01
1.76275536e-01 -8.93017463e-03 7.93495953e-01 -2.03148022e-01
9.91581380e-01 -2.23367763e+00 2.45073903e-02 -8.69241729e-02
1.08338371e-01 3.38262349e-01 2.35707425e-02 3.42004895e-01
1.76752046e-01 -2.91362584e-01 -7.41539180e-01 -2.00511858e-01
-4.38526750e-01 3.54358375e-01 -3.56479257e-01 2.74971455e-01
7.08800793e-01 1.12022722e+00 -7.34659910e-01 -3.38844061e-01
1.85327888e-01 5.18328428e-01 -4.11728531e-01 -9.33647603e-02
-4.30074692e-01 6.01052105e-01 -2.39361808e-01 9.12995815e-01
2.79364139e-01 -3.90335470e-01 9.06091332e-02 -1.58747524e-01
-5.32002091e-01 4.03443575e-01 -9.61171508e-01 1.98115015e+00
-1.49862662e-01 7.84962237e-01 -1.07910119e-01 -1.36464357e+00
1.19280469e+00 2.44878262e-01 8.79480839e-01 -1.19467556e+00
3.28613400e-01 6.55648470e-01 1.39884144e-01 3.35554294e-02
-3.46293021e-03 2.07029674e-02 -3.34295183e-02 2.79418319e-01
4.43672091e-01 1.14218511e-01 3.77270550e-01 8.32195580e-02
1.46407759e+00 1.48344159e-01 1.02250688e-01 -3.67495388e-01
-3.16372663e-02 2.68433057e-02 5.09821832e-01 7.94971406e-01
-2.59212881e-01 5.88933468e-01 3.26374054e-01 -3.01558644e-01
-1.02230108e+00 -1.30575979e+00 -5.82734644e-01 9.11604941e-01
2.37122357e-01 -1.38707608e-01 -5.84874272e-01 1.78272322e-01
-6.39631301e-02 2.95147181e-01 -3.46546680e-01 -1.00466192e-01
-6.78264141e-01 -1.18472230e+00 1.07005692e+00 7.49351263e-01
5.34596503e-01 -1.31740367e+00 -1.22395408e+00 9.56866801e-01
1.73350930e-01 -1.34981251e+00 2.61417449e-01 1.06195259e+00
-1.11328924e+00 -6.33513629e-01 -4.98402208e-01 -1.02767611e+00
3.10410559e-01 -2.44444177e-01 8.43835533e-01 -4.65886056e-01
-6.83449149e-01 -2.02373066e-03 -6.44473135e-02 -6.30640626e-01
1.64839804e-01 1.34487003e-01 -2.55361736e-01 -2.50974596e-01
4.84831363e-01 -1.24461663e+00 -8.90595496e-01 1.93915293e-01
-1.01774895e+00 2.60336578e-01 7.17686176e-01 9.18189585e-01
1.23846066e+00 -3.32831293e-01 1.02330375e+00 -5.11597991e-01
-1.04416072e-01 -5.59394538e-01 -6.90082133e-01 -2.60857821e-01
-5.86570740e-01 1.14693016e-01 6.58702672e-01 -5.43927670e-01
-8.79245281e-01 5.03577411e-01 -2.64983803e-01 -1.75668765e-03
-2.76339173e-01 3.12381506e-01 2.89794773e-01 -3.47491831e-01
7.40596890e-01 4.53620702e-01 4.54857713e-03 3.24904057e-03
-2.69126505e-01 3.87370199e-01 6.30469501e-01 -4.23942566e-01
-8.21555406e-02 9.38605487e-01 3.27796191e-01 -6.63230002e-01
-5.50667703e-01 -3.83906394e-01 -2.11871505e-01 -2.54291087e-01
7.76576817e-01 -1.09774327e+00 -8.53420258e-01 1.04437089e+00
-1.17064059e+00 -5.52050769e-01 -3.75342816e-01 3.53463560e-01
-8.64340961e-01 -3.40995401e-01 -8.40212882e-01 -8.38474154e-01
-4.88298714e-01 -9.08982813e-01 1.18026173e+00 4.69785780e-01
-1.73461795e-01 -5.87387919e-01 -1.06887154e-01 -3.11223604e-03
4.94165719e-01 5.11805356e-01 4.61955220e-01 -7.90746093e-01
-8.88353169e-01 2.35374309e-02 -2.65859306e-01 -1.28064647e-01
-3.33805740e-01 -2.32874274e-01 -1.14200330e+00 5.86178489e-02
-1.47857949e-01 -5.09879947e-01 1.15393698e+00 6.90053761e-01
1.23762155e+00 2.29800075e-01 -5.03432572e-01 7.40196526e-01
1.63681293e+00 4.22569066e-01 7.25918591e-01 3.81560594e-01
2.12403148e-01 2.49021351e-01 1.32656693e-01 8.30008686e-01
2.14035079e-01 4.14401323e-01 6.98881984e-01 1.56657979e-01
-3.20019990e-01 -1.24077082e-01 2.58368582e-01 6.67443812e-01
1.28216401e-01 -3.80145311e-01 -1.02959776e+00 7.80801296e-01
-1.93564451e+00 -9.56736088e-01 -1.34349227e-01 1.96030581e+00
1.00425398e+00 3.49911630e-01 -1.16774946e-01 2.32307762e-01
5.44411242e-01 1.26559734e-01 -1.03062677e+00 -2.19075963e-01
-7.57532597e-01 8.38769436e-01 5.71650147e-01 -2.23802269e-01
-9.67590630e-01 8.45713556e-01 5.64910412e+00 7.54520953e-01
-1.26077294e+00 2.58229468e-02 4.13480908e-01 -5.05630970e-01
2.28072405e-02 -1.27117664e-01 -1.13816404e+00 7.07320392e-01
1.39582646e+00 1.06949732e-01 6.03837430e-01 2.36912027e-01
2.32778251e-01 -3.59825820e-01 -8.39528501e-01 9.74012673e-01
-2.74124533e-01 -1.75471091e+00 -2.18847424e-01 -1.33516833e-01
7.25322902e-01 6.56578422e-01 -5.20258322e-02 -9.22348574e-02
6.91333413e-02 -1.12024581e+00 7.35387743e-01 5.25258005e-01
6.97085381e-01 -5.95803320e-01 4.93669003e-01 2.35099092e-01
-1.28463888e+00 -1.17817692e-01 -1.18440904e-01 -1.64179474e-01
3.60197186e-01 8.13846111e-01 -3.93428892e-01 -1.21029802e-01
9.92012203e-01 1.04711378e+00 -2.74636656e-01 1.23873973e+00
3.10516745e-01 7.47663915e-01 -9.39567745e-01 -2.37739518e-01
2.44978428e-01 2.61702776e-01 4.88305241e-01 1.36011910e+00
4.73136514e-01 1.77862063e-01 -2.26512760e-01 1.00693369e+00
-1.71508744e-01 -3.74254614e-01 -4.71711516e-01 -3.14204901e-01
6.39720917e-01 1.09228778e+00 -1.08378339e+00 -4.21125814e-02
-2.34804198e-01 8.57738793e-01 2.83541113e-01 1.97324842e-01
-7.99026906e-01 -4.29102540e-01 4.84445363e-01 2.05162372e-02
6.66491151e-01 -3.11548501e-01 -8.42222691e-01 -7.09066153e-01
-3.92028391e-02 -5.09036481e-01 2.50257164e-01 -6.41939282e-01
-9.46265697e-01 5.77555478e-01 -5.00559568e-01 -9.05752778e-01
9.07844584e-03 -6.75776005e-01 -3.74849796e-01 4.58359778e-01
-1.44982791e+00 -9.74428892e-01 -1.69962302e-01 5.30420244e-01
4.86439854e-01 3.14121135e-02 8.96736562e-01 3.83143306e-01
-3.20680261e-01 2.14851975e-01 1.69912696e-01 -1.84759691e-01
2.81184018e-01 -9.63511527e-01 2.88791060e-01 7.04870999e-01
1.83549523e-01 1.73175111e-01 4.37339395e-01 -5.53657532e-01
-1.45657396e+00 -1.16840959e+00 7.11122990e-01 5.92526011e-02
7.50113070e-01 -5.61662078e-01 -9.44502771e-01 3.06057781e-01
-1.03014685e-01 2.34196767e-01 7.08033383e-01 -4.98609632e-01
-2.00559244e-01 -3.20123255e-01 -1.00661755e+00 4.52269912e-01
1.38843119e+00 -4.28589612e-01 -2.35931680e-01 -7.45223556e-03
5.18334329e-01 -2.59674132e-01 -9.76033747e-01 6.80415630e-01
7.16076910e-01 -8.83354366e-01 9.51970279e-01 -9.04378891e-02
3.46012354e-01 -4.02952492e-01 -2.41165757e-01 -7.45288372e-01
-8.70295987e-02 -5.05226016e-01 -3.80555093e-01 1.13770139e+00
3.65091652e-01 -6.38878465e-01 9.07596409e-01 7.21765533e-02
-3.19336444e-01 -1.01213324e+00 -1.24586976e+00 -6.58353209e-01
-2.75401205e-01 -4.32148188e-01 4.89263237e-02 1.43907219e-01
-1.54526293e-01 -3.13315205e-02 9.86070558e-02 5.87535910e-02
7.19848156e-01 4.02690880e-02 -1.51974723e-01 -1.12697744e+00
-1.90560877e-01 -4.96901900e-01 -6.45483077e-01 -8.90226185e-01
2.13353913e-02 -9.99275506e-01 3.13607603e-01 -1.34935510e+00
-4.95717686e-04 -3.80674601e-01 -6.63207233e-01 4.63661700e-01
2.09516808e-01 5.87115288e-01 1.48073277e-02 2.14419112e-01
-7.80943751e-01 4.47727263e-01 8.34616125e-01 2.03481525e-01
-5.91422953e-02 -2.60597229e-01 -3.84053856e-01 4.96719927e-01
1.14513075e+00 -6.47203624e-01 -2.11901963e-01 -3.71245295e-01
1.33918658e-01 -3.67069803e-02 6.35599017e-01 -1.50864184e+00
7.83943892e-01 1.61474898e-01 5.16440511e-01 -5.33584893e-01
5.04228413e-01 -5.83955586e-01 2.76270956e-01 7.46115029e-01
-3.40569109e-01 -2.45112225e-01 7.19087422e-01 7.23650694e-01
-2.92121500e-01 1.35974586e-01 9.46051419e-01 -6.15258925e-02
-9.62445498e-01 4.22770649e-01 -9.74438131e-01 1.63477898e-01
1.05060697e+00 -5.14684856e-01 -4.59191889e-01 2.22938687e-01
-5.35424590e-01 -1.33094937e-01 2.93409586e-01 -8.45571160e-02
6.91603839e-01 -1.04487908e+00 -3.53865176e-01 3.26466352e-01
3.43325287e-02 1.84833273e-01 1.78413272e-01 9.04236019e-01
-4.31471258e-01 4.07728076e-01 -6.50829017e-01 -8.48731220e-01
-6.85619056e-01 -3.36070987e-03 2.55527407e-01 -1.62462160e-01
-4.56516564e-01 9.68354464e-01 -1.48289111e-02 4.04163115e-02
2.53910094e-01 -1.19362280e-01 -3.46366167e-02 7.03351572e-02
3.02990198e-01 2.56436557e-01 2.07526565e-01 -1.50957420e-01
-5.30006289e-01 3.12027425e-01 1.08174928e-01 -1.48616984e-01
1.67191219e+00 1.01982743e-01 -1.41306117e-01 6.23961270e-01
9.61205125e-01 -7.97052503e-01 -1.81389415e+00 -1.51048005e-01
2.07586095e-01 2.42737263e-01 7.24956915e-02 -8.95602167e-01
-9.80237186e-01 9.97526586e-01 7.38863707e-01 -9.60885286e-02
1.51561558e+00 1.52001336e-01 1.19257522e+00 3.08437586e-01
5.15459239e-01 -1.02022433e+00 2.04053327e-01 6.33612156e-01
5.22413373e-01 -9.18927372e-01 -4.94407088e-01 -1.19019233e-01
-3.53618503e-01 1.04772139e+00 6.04484081e-01 -4.62529033e-01
4.81036037e-01 1.12026107e+00 -3.71831954e-01 -1.85356319e-01
-1.14877009e+00 -2.85318106e-01 -2.68071890e-01 6.28693044e-01
2.02681422e-01 -8.23059008e-02 -2.20024437e-01 7.67850101e-01
1.77412495e-01 5.82289875e-01 4.32942621e-02 1.09752035e+00
-7.16479123e-01 -7.95598805e-01 1.75387248e-01 8.14553440e-01
-5.50892889e-01 -3.76402915e-01 -2.68946253e-02 2.31892884e-01
1.93723321e-01 5.91802776e-01 4.53197241e-01 -8.93639624e-02
1.83782130e-01 9.85897183e-02 5.79822719e-01 -3.23728472e-01
-6.82098866e-01 2.94699222e-02 -1.86878830e-01 -6.75611377e-01
-6.01199210e-01 -8.55586827e-01 -1.88312685e+00 -9.45048109e-02
2.20464189e-02 -5.67814648e-01 8.22078347e-01 8.24070156e-01
8.37305725e-01 7.67171860e-01 9.08012912e-02 -8.73838723e-01
-5.74809201e-02 -4.17533576e-01 -4.67362344e-01 7.91996568e-02
9.56413373e-02 -5.59611857e-01 -2.53010184e-01 3.11192989e-01] | [8.217541694641113, 2.4153876304626465] |
f7b230c9-9f72-4bf2-8d23-debc234039b6 | recent-progress-in-the-cuhk-dysarthric-speech | 2201.05845 | null | https://arxiv.org/abs/2201.05845v2 | https://arxiv.org/pdf/2201.05845v2.pdf | Recent Progress in the CUHK Dysarthric Speech Recognition System | Despite the rapid progress of automatic speech recognition (ASR) technologies in the past few decades, recognition of disordered speech remains a highly challenging task to date. Disordered speech presents a wide spectrum of challenges to current data intensive deep neural networks (DNNs) based ASR technologies that predominantly target normal speech. This paper presents recent research efforts at the Chinese University of Hong Kong (CUHK) to improve the performance of disordered speech recognition systems on the largest publicly available UASpeech dysarthric speech corpus. A set of novel modelling techniques including neural architectural search, data augmentation using spectra-temporal perturbation, model based speaker adaptation and cross-domain generation of visual features within an audio-visual speech recognition (AVSR) system framework were employed to address the above challenges. The combination of these techniques produced the lowest published word error rate (WER) of 25.21% on the UASpeech test set 16 dysarthric speakers, and an overall WER reduction of 5.4% absolute (17.6% relative) over the CUHK 2018 dysarthric speech recognition system featuring a 6-way DNN system combination and cross adaptation of out-of-domain normal speech data trained systems. Bayesian model adaptation further allows rapid adaptation to individual dysarthric speakers to be performed using as little as 3.06 seconds of speech. The efficacy of these techniques were further demonstrated on a CUDYS Cantonese dysarthric speech recognition task. | ['Helen Meng', 'Xunying Liu', 'Jianwei Yu', 'Mingyu Cui', 'Xurong Xie', 'Shoukang Hu', 'Mengzhe Geng', 'Shansong Liu'] | 2022-01-15 | null | null | null | null | ['audio-visual-speech-recognition'] | ['speech'] | [ 1.20214663e-01 7.40375966e-02 5.62137067e-01 -2.30896518e-01
-1.18131936e+00 -2.81878918e-01 6.03296280e-01 -6.11496210e-01
-5.36919653e-01 3.87547851e-01 4.69845384e-01 -4.06677365e-01
7.00757205e-02 3.35267670e-02 -2.09870979e-01 -7.40853131e-01
1.27561748e-01 6.56097651e-01 4.05863151e-02 -5.13340771e-01
-2.60631982e-02 8.29373777e-01 -1.77158082e+00 2.87258208e-01
4.59260911e-01 8.07429492e-01 4.74499106e-01 1.20755541e+00
1.56472430e-01 4.05034214e-01 -1.14979780e+00 1.85172930e-01
2.96321392e-01 -1.98564708e-01 -6.07109487e-01 1.95552021e-01
5.67884803e-01 -6.50874376e-01 -1.02227139e+00 8.20914626e-01
1.39556086e+00 3.76351207e-01 3.43620300e-01 -4.54776198e-01
-1.04536557e+00 1.37248948e-01 1.95022166e-01 8.61348391e-01
1.04850814e-01 4.34465021e-01 5.67318082e-01 -1.18486595e+00
3.00113320e-01 1.40929925e+00 1.91626906e-01 1.35027242e+00
-1.02232373e+00 -8.45830977e-01 -1.83317199e-01 4.80272651e-01
-1.24584389e+00 -1.38270366e+00 5.68066597e-01 -2.30499476e-01
1.91848803e+00 1.87726229e-01 8.10932815e-01 1.47456384e+00
-3.86703372e-01 6.72288299e-01 9.12261128e-01 -5.53716600e-01
3.78539681e-01 -2.29118899e-01 -6.83250353e-02 2.75072247e-01
-3.71333361e-01 4.74217325e-01 -1.06511760e+00 -1.25337005e-01
5.81671417e-01 -6.45981491e-01 -2.68020809e-01 4.10050005e-01
-8.19693863e-01 6.63089216e-01 -2.36673623e-01 3.19027603e-01
-3.45962197e-01 -2.50609547e-01 6.86320186e-01 7.64073670e-01
5.39931774e-01 3.12827945e-01 -6.50006235e-01 -6.91527665e-01
-8.95751655e-01 3.28108333e-02 4.58251327e-01 7.18049824e-01
-2.49948665e-01 1.19423795e+00 6.91372007e-02 1.84760404e+00
5.55898249e-01 7.13432968e-01 1.21035564e+00 -6.03907526e-01
4.51826900e-01 2.09894344e-01 -5.12421727e-01 -6.45553768e-02
-8.75111967e-02 -2.60641873e-01 -4.24008250e-01 4.64512140e-01
1.28825560e-01 2.13618148e-02 -1.78171039e+00 1.46526933e+00
1.02410421e-01 -2.22901702e-01 4.82026517e-01 8.35447013e-01
7.34066963e-01 9.71869826e-01 -1.54421419e-01 -2.33289838e-01
9.89659250e-01 -7.10284591e-01 -7.75926411e-01 -3.20578188e-01
4.96557087e-01 -8.83535683e-01 1.23500991e+00 7.17394412e-01
-1.18636191e+00 -3.70827854e-01 -1.11784256e+00 -5.92275001e-02
-2.25464240e-01 2.09992662e-01 -1.76672086e-01 1.04018664e+00
-1.48478019e+00 -1.01565987e-01 -1.07572365e+00 -4.98869687e-01
3.99850249e-01 8.05156589e-01 -5.22431970e-01 -7.20174517e-03
-1.11926937e+00 1.00343585e+00 2.20524713e-01 -1.60276636e-01
-1.25808370e+00 -8.67859960e-01 -7.52054453e-01 -2.04699010e-01
-4.53431830e-02 -8.63782712e-04 1.48958921e+00 -9.05375779e-01
-2.02267504e+00 1.00600266e+00 -2.08520770e-01 -6.69878960e-01
2.41332084e-01 -3.31076264e-01 -1.02865994e+00 1.58641100e-01
-4.76240635e-01 3.22663695e-01 8.60120356e-01 -6.68323278e-01
-4.64639157e-01 -6.86100245e-01 -8.46206725e-01 5.04702985e-01
-3.85342807e-01 6.14551187e-01 -1.84530854e-01 -9.82365012e-01
1.71335012e-01 -8.28242779e-01 4.15545702e-01 -1.49518266e-01
-1.34854391e-01 -3.44130576e-01 1.20037520e+00 -1.28681898e+00
1.13671875e+00 -2.29257154e+00 8.50945860e-02 -9.00436379e-03
-4.34636027e-02 1.11499059e+00 -2.22958431e-01 1.93853378e-01
-3.16279978e-01 -2.13290825e-01 -3.19263905e-01 -3.84759724e-01
-1.73407868e-01 2.97052771e-01 -1.99731275e-01 5.84172606e-01
1.35080263e-01 3.42069894e-01 -2.78777033e-01 2.03488439e-01
3.87371123e-01 8.49496305e-01 -3.51058066e-01 5.06828249e-01
7.91053548e-02 1.43285319e-01 2.86154211e-01 9.59814012e-01
3.37003499e-01 6.16357565e-01 -2.78632730e-01 3.83837342e-01
-1.23373484e-02 6.69692516e-01 -8.17162275e-01 1.49076688e+00
-3.57482910e-01 9.27822471e-01 5.65438569e-01 -8.87435138e-01
1.24838543e+00 8.07551682e-01 7.66215101e-02 -9.19876277e-01
-5.95721863e-02 6.14235044e-01 6.01519942e-01 -5.77221930e-01
9.45499837e-02 -2.43062094e-01 5.63977003e-01 5.46293159e-04
3.79431903e-01 -2.01925058e-02 -2.82867759e-01 -1.82401493e-01
1.41410458e+00 -5.41422009e-01 6.27620071e-02 -1.17982112e-01
6.73062146e-01 -1.91216603e-01 3.28137964e-01 4.95631784e-01
-6.36178195e-01 1.00216079e+00 -8.12470913e-02 -1.55242756e-01
-1.51827455e+00 -1.34635460e+00 -1.55026436e-01 9.72431421e-01
-8.99647593e-01 -1.69357181e-01 -6.02884591e-01 -2.12116897e-01
-2.19057202e-01 7.62719095e-01 -5.13941571e-02 -2.24578261e-01
-8.55064929e-01 -5.41818440e-01 1.15238380e+00 3.90011668e-01
3.25392216e-01 -1.34787738e+00 1.50348783e-01 2.78032213e-01
4.72420931e-01 -1.12004662e+00 -3.94466102e-01 9.81930867e-02
-4.30993736e-01 -6.82642937e-01 -1.15587890e+00 -1.03493392e+00
1.53784081e-01 -1.02262035e-01 5.04466295e-01 -3.34564060e-01
-5.77478051e-01 5.54492176e-01 -2.59930015e-01 -2.92862087e-01
-1.19228470e+00 -2.37657890e-01 9.62435246e-01 -3.18547159e-01
7.08205104e-01 -6.24672830e-01 -3.68751168e-01 4.60254960e-02
-6.86347961e-01 -4.39521343e-01 6.45011008e-01 1.14432156e+00
5.31167984e-01 -4.34066445e-01 1.04429245e+00 -1.20225258e-01
9.87847507e-01 -3.55181813e-01 -6.23090625e-01 1.97711010e-02
-6.88885093e-01 -2.82365561e-01 7.00782239e-01 -6.90300405e-01
-1.05518842e+00 6.56782314e-02 -8.27342153e-01 -7.73276806e-01
-4.58660275e-01 1.02797545e-01 -2.36684009e-01 3.03425547e-02
8.48668456e-01 8.40216279e-01 6.12918377e-01 -7.64958203e-01
-3.87638249e-02 1.69920707e+00 6.52710617e-01 -1.63297772e-01
4.19750005e-01 -2.24214420e-02 -6.41274691e-01 -1.79661298e+00
-2.07692776e-02 -5.65351009e-01 -3.15883428e-01 -6.85113892e-02
7.63348639e-01 -9.94504750e-01 -2.73231387e-01 1.07260394e+00
-9.63100791e-01 -3.47030759e-01 -1.25296548e-01 6.17052436e-01
-6.17987692e-01 3.15595567e-01 -4.32757735e-01 -1.07899559e+00
-8.26913416e-01 -1.50110626e+00 7.76707113e-01 -1.96331546e-01
-3.21430326e-01 -4.30321544e-01 2.54032522e-01 7.77987778e-01
6.94865108e-01 -6.86875999e-01 9.70015466e-01 -1.45656204e+00
-4.50385660e-02 -9.23331007e-02 1.01852313e-01 1.34608400e+00
3.41943920e-01 -2.74573773e-01 -1.20005596e+00 -4.17857528e-01
1.45027265e-01 -3.76690388e-01 4.70861584e-01 2.66268969e-01
7.98067331e-01 -3.23374420e-01 3.91780525e-01 3.96286041e-01
7.03103006e-01 8.45392168e-01 6.35586858e-01 3.13009232e-01
5.60825765e-01 2.73544997e-01 6.94735646e-02 1.83197185e-01
-1.62953541e-01 8.59090984e-01 -4.32448648e-02 4.48151469e-01
-9.37038541e-01 1.97345346e-01 9.47343767e-01 1.32354844e+00
2.51320630e-01 -1.43824503e-01 -1.27026761e+00 9.37427819e-01
-1.05552709e+00 -8.28377426e-01 3.18923473e-01 2.19657731e+00
7.86846697e-01 -1.56072617e-01 2.56752670e-01 4.29673195e-01
7.23424077e-01 1.04771972e-01 -8.21112752e-01 -7.23918378e-01
-3.25396329e-01 5.77299058e-01 2.64035136e-01 4.44045603e-01
-5.86491048e-01 1.05541992e+00 6.12865639e+00 9.59574819e-01
-1.17899883e+00 1.61856368e-01 2.92602867e-01 -8.27233911e-01
2.89449573e-01 -8.15583408e-01 -8.41206014e-01 4.80932176e-01
1.72752452e+00 1.72994304e-02 8.75982881e-01 8.86513710e-01
5.47707558e-01 4.23932582e-01 -7.19904244e-01 1.38431013e+00
4.18910503e-01 -1.13562143e+00 1.33810505e-01 1.66756764e-01
3.43373507e-01 8.17886055e-01 2.46445537e-01 2.82892644e-01
1.63254421e-02 -1.34637070e+00 5.42918503e-01 1.47294745e-01
1.14779186e+00 -9.70635176e-01 2.43467242e-01 6.46903515e-02
-5.79957724e-01 -2.77359784e-01 -2.13095322e-01 3.12148690e-01
-8.60802680e-02 -3.91454948e-03 -1.38907862e+00 -3.05256963e-01
8.31557989e-01 1.58718720e-01 -1.40310839e-01 7.88636804e-01
7.32008666e-02 1.22558534e+00 -5.38726330e-01 3.78993750e-02
9.37485397e-02 4.24561709e-01 1.17409754e+00 1.05598009e+00
5.79861164e-01 9.68717039e-02 -5.72135985e-01 2.97521174e-01
-1.01482905e-01 4.75955643e-02 -5.19601464e-01 -2.85500348e-01
6.28208399e-01 4.37751591e-01 2.10859433e-01 -1.43763542e-01
-4.65750039e-01 9.07615423e-01 2.75110990e-01 3.47295135e-01
-1.10056981e-01 -3.46744090e-01 1.24503386e+00 1.41421497e-01
1.93332836e-01 -5.19293129e-01 -7.06325769e-02 -7.71451235e-01
2.67016292e-01 -1.56855488e+00 1.19022124e-01 -6.64613366e-01
-1.16733909e+00 8.83411407e-01 -2.50429720e-01 -8.90007913e-01
-5.21308839e-01 -9.33359444e-01 -3.15716445e-01 1.29138660e+00
-1.13972163e+00 -7.86789954e-01 2.84494519e-01 6.79823995e-01
1.42319822e+00 -1.34831679e+00 1.00743079e+00 4.15308148e-01
-6.92473829e-01 7.38969743e-01 5.11267602e-01 -2.60333531e-02
5.39193571e-01 -1.07619035e+00 7.90024161e-01 1.03632498e+00
-3.83480638e-02 4.47609603e-01 5.89278638e-01 -7.19195426e-01
-1.27450895e+00 -8.98887098e-01 7.24513888e-01 -1.26743123e-01
7.55896449e-01 -5.21246791e-01 -1.22609627e+00 3.83700579e-01
2.98226811e-02 -2.40458248e-04 6.29455268e-01 -3.87939334e-01
-3.22605759e-01 -7.01030567e-02 -1.16942036e+00 5.98893642e-01
9.82080400e-01 -8.39685440e-01 -8.09527218e-01 2.66464353e-01
6.70189381e-01 -4.28030580e-01 -6.31662607e-01 2.74703145e-01
3.93512428e-01 -5.91341496e-01 9.76225674e-01 -6.48276985e-01
-2.04623088e-01 -1.38119176e-01 -5.65030217e-01 -1.48882306e+00
2.97086149e-01 -1.03619087e+00 -5.24429679e-01 1.15529919e+00
4.12312299e-01 -6.61198080e-01 7.33424842e-01 5.23905456e-01
-7.23620415e-01 -6.49259150e-01 -1.73721552e+00 -1.05859196e+00
9.37206224e-02 -7.54784167e-01 9.99414474e-02 3.98809165e-01
-2.75567859e-01 9.31927636e-02 -3.29994917e-01 3.27334553e-01
2.46663988e-01 -1.23410606e+00 2.55346358e-01 -8.73981774e-01
-1.76117495e-01 -3.57538372e-01 -8.65826249e-01 -7.19260633e-01
2.19974011e-01 -8.28337193e-01 2.67832540e-02 -1.30303073e+00
-5.75403512e-01 -1.65460240e-02 -1.53424829e-01 3.74922812e-01
3.24075073e-01 -1.79446578e-01 -6.24331832e-02 2.38748342e-01
2.71850705e-01 8.16796422e-01 8.00550818e-01 -1.83515444e-01
-4.49179322e-01 1.94381759e-01 -2.46584177e-01 3.62650871e-01
9.95556593e-01 -4.22557712e-01 -5.71624875e-01 -4.33426768e-01
-4.94779706e-01 -9.64830294e-02 8.73062834e-02 -1.09747279e+00
2.85905421e-01 3.72642219e-01 2.25870982e-01 -4.38926250e-01
8.97730887e-01 -2.95197964e-01 -2.49590456e-01 3.88991058e-01
-2.01463342e-01 8.57442990e-03 6.02958083e-01 3.51777554e-01
-2.20960900e-01 4.33672294e-02 1.27997875e+00 -6.58313707e-02
-7.26096451e-01 -1.29221082e-02 -1.01761711e+00 -4.31002118e-02
7.39206612e-01 -3.65236312e-01 -2.72053510e-01 -3.21298867e-01
-1.09467399e+00 -2.75405109e-01 -6.11646138e-02 7.09965825e-01
1.01071644e+00 -1.01440525e+00 -9.85846877e-01 7.04002559e-01
9.32169929e-02 -2.16618165e-01 4.73034620e-01 3.85244161e-01
-5.74642241e-01 6.32092237e-01 -1.97008088e-01 -4.04548049e-01
-1.82991028e+00 -9.40819010e-02 6.29220188e-01 3.37271661e-01
-9.68455970e-01 1.11483800e+00 -2.07562298e-01 -4.29277211e-01
6.10609412e-01 1.94902167e-01 -6.61280081e-02 -2.29178146e-01
6.80683553e-01 6.21829569e-01 7.25692511e-01 -1.00771987e+00
-3.96996260e-01 -5.21792248e-02 -5.87545455e-01 -6.23084188e-01
1.40461922e+00 -1.28981292e-01 4.21811253e-01 3.20057124e-01
1.24409866e+00 -1.08480342e-01 -9.91212606e-01 -1.32761329e-01
-2.43129611e-01 -1.55102164e-01 5.29911041e-01 -1.33609331e+00
-8.40177298e-01 9.14208472e-01 1.21836150e+00 1.50732696e-01
1.10349309e+00 2.02761680e-01 9.41120148e-01 6.04841173e-01
-3.15402895e-01 -1.38467085e+00 -1.14787973e-01 8.14584374e-01
1.38305664e+00 -9.65779603e-01 -4.89398450e-01 3.62860978e-01
-6.60618603e-01 1.25399864e+00 6.42462671e-01 -5.13467975e-02
5.02000511e-01 3.12133253e-01 6.02111518e-01 -1.18837349e-01
-8.26767564e-01 -1.11270659e-01 3.88873011e-01 1.01984942e+00
2.74801672e-01 -1.90609366e-01 1.34786829e-01 4.33198482e-01
-5.88331521e-01 -4.75312233e-01 4.76246566e-01 7.41061091e-01
-5.41392505e-01 -8.96481097e-01 -5.64738572e-01 5.92361510e-01
-7.99413204e-01 -3.41110557e-01 -3.99586260e-01 6.37536407e-01
-5.07197201e-01 8.87752652e-01 1.51288599e-01 -2.75530934e-01
4.93898898e-01 4.63162452e-01 1.31572887e-01 -8.55486691e-01
-5.38707316e-01 1.93915173e-01 3.77499819e-01 -2.83550054e-01
1.74908504e-01 -8.60726476e-01 -1.23248148e+00 7.32264295e-02
-2.71940529e-02 -3.64503890e-01 1.15118337e+00 1.01672268e+00
4.99650210e-01 5.63022256e-01 2.99636841e-01 -4.97117281e-01
-8.82232368e-01 -1.42825520e+00 -5.60041785e-01 -9.39683840e-02
8.37232471e-01 -4.02116299e-01 -7.77027130e-01 -1.01281002e-01] | [14.571558952331543, 6.438146114349365] |
e68406f9-2000-4496-a444-c83eae974ddd | how-to-train-your-event-camera-neural-network | 2003.09078 | null | https://arxiv.org/abs/2003.09078v5 | https://arxiv.org/pdf/2003.09078v5.pdf | Reducing the Sim-to-Real Gap for Event Cameras | Event cameras are paradigm-shifting novel sensors that report asynchronous, per-pixel brightness changes called 'events' with unparalleled low latency. This makes them ideal for high speed, high dynamic range scenes where conventional cameras would fail. Recent work has demonstrated impressive results using Convolutional Neural Networks (CNNs) for video reconstruction and optic flow with events. We present strategies for improving training data for event based CNNs that result in 20-40% boost in performance of existing state-of-the-art (SOTA) video reconstruction networks retrained with our method, and up to 15% for optic flow networks. A challenge in evaluating event based video reconstruction is lack of quality ground truth images in existing datasets. To address this, we present a new High Quality Frames (HQF) dataset, containing events and ground truth frames from a DAVIS240C that are well-exposed and minimally motion-blurred. We evaluate our method on HQF + several existing major event camera datasets. | ['Tom Drummond', 'Nick Barnes', 'Timo Stoffregen', 'Davide Scaramuzza', 'Cedric Scheerlinck', 'Robert Mahony', 'Lindsay Kleeman'] | 2020-03-20 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/5855_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123720528.pdf | eccv-2020-8 | ['video-reconstruction'] | ['computer-vision'] | [ 2.36186236e-01 -5.90812743e-01 1.39064565e-01 -4.90271837e-01
-4.75210518e-01 -2.92065531e-01 6.38917565e-01 -4.20953274e-01
-7.23172545e-01 6.98656857e-01 4.81441647e-01 -5.09668849e-02
2.62601614e-01 -6.31422698e-01 -1.23186982e+00 -3.67528379e-01
-2.97014922e-01 -2.49264821e-01 6.94656134e-01 2.42060162e-02
7.33042657e-02 6.39240265e-01 -1.70751894e+00 7.55544841e-01
-6.32492155e-02 1.12134922e+00 4.00315970e-02 1.36907792e+00
3.30273032e-01 1.53538394e+00 -5.69435596e-01 -2.89106011e-01
3.87466580e-01 -1.94640532e-01 -6.88141167e-01 2.68190959e-03
1.22400773e+00 -1.14377046e+00 -1.11792839e+00 7.20504224e-01
4.51089650e-01 2.77207792e-01 -7.63436332e-02 -1.42410934e+00
-4.53803122e-01 2.56259292e-01 -3.73196483e-01 9.60953772e-01
5.34156203e-01 6.50424421e-01 4.21247125e-01 -8.90740871e-01
9.80719805e-01 1.23008597e+00 9.31586564e-01 6.74635828e-01
-1.01377702e+00 -7.74561882e-01 -2.74244040e-01 4.87424582e-01
-1.03109610e+00 -1.07713187e+00 1.73238590e-01 -1.76055059e-01
1.86322665e+00 1.52095000e-03 9.02484834e-01 1.45220697e+00
5.88185906e-01 5.05497217e-01 5.22690535e-01 1.26346368e-02
7.20082894e-02 -6.39334977e-01 -3.22439641e-01 6.86505675e-01
2.18854114e-01 5.94847560e-01 -1.23764265e+00 4.45655167e-01
1.19074333e+00 1.96024030e-01 -6.88951969e-01 6.13757223e-02
-1.79961181e+00 3.13623875e-01 3.41835052e-01 -1.77189007e-01
-4.46851403e-01 1.01442480e+00 6.20796382e-01 2.92886674e-01
1.87149629e-01 7.48386458e-02 -4.84302074e-01 -5.91420233e-01
-1.17512548e+00 3.30825984e-01 6.76264942e-01 1.03529668e+00
6.94280326e-01 5.33056498e-01 -1.54894903e-01 1.03972159e-01
1.57056391e-01 5.62061608e-01 3.12314600e-01 -1.84681106e+00
3.84479761e-01 -1.30546570e-01 3.65356565e-01 -8.32731485e-01
-3.11432123e-01 2.33418465e-01 -6.56097472e-01 5.84662974e-01
4.57534015e-01 -1.67826176e-01 -8.83936465e-01 1.54493988e+00
-5.89893609e-02 8.38663757e-01 1.18516564e-01 1.28370118e+00
1.01200056e+00 1.06282091e+00 -7.97002017e-03 -2.79533535e-01
1.22442234e+00 -7.75417864e-01 -7.27459908e-01 -1.14574127e-01
-4.28269356e-02 -9.11839187e-01 7.39986539e-01 6.40162051e-01
-1.46841967e+00 -6.61941588e-01 -1.19171238e+00 -3.28938097e-01
-7.45948777e-02 -2.48139828e-01 8.30271363e-01 4.66431737e-01
-1.50035131e+00 7.49794662e-01 -1.10101008e+00 -4.94216084e-01
6.46528065e-01 4.08135802e-01 -6.84617877e-01 -1.60624310e-01
-9.17062104e-01 5.81172705e-01 5.04052222e-01 -1.17800571e-01
-1.53685093e+00 -9.50993955e-01 -7.52765417e-01 2.93159597e-02
1.12304322e-01 -6.84543729e-01 1.59262586e+00 -1.10471582e+00
-1.49243104e+00 7.55807519e-01 -5.83870225e-02 -9.79574859e-01
3.04035038e-01 -2.18758658e-01 -6.18000448e-01 6.81690753e-01
3.30003612e-02 1.31058288e+00 8.08444381e-01 -6.99189782e-01
-9.56410706e-01 2.66610444e-01 1.65031478e-01 2.31679808e-02
-7.28673413e-02 4.57996845e-01 -5.70718586e-01 -5.49432635e-01
-4.55588222e-01 -6.03740931e-01 1.36651486e-01 4.87057030e-01
1.73401356e-01 2.38350257e-01 1.17605472e+00 -4.14956480e-01
7.20728815e-01 -2.00905871e+00 -3.87633353e-01 -6.28156722e-01
2.78083563e-01 3.42203498e-01 -1.27870396e-01 -5.50203174e-02
-1.78875834e-01 -2.07333907e-01 2.32474491e-01 -4.38008845e-01
-4.70336080e-01 4.20700759e-01 -4.74793017e-01 4.89831239e-01
4.71635759e-01 8.62564385e-01 -9.50110197e-01 -4.13665086e-01
8.81537557e-01 7.22836912e-01 -6.94071710e-01 2.41604179e-01
-1.69431865e-01 3.31833810e-01 3.24561685e-01 7.78108120e-01
5.37932456e-01 -4.36333060e-01 -2.01919273e-01 -6.89369500e-01
-3.51951331e-01 1.34931266e-01 -1.08928168e+00 2.14435649e+00
-3.72073837e-02 1.60142505e+00 -1.73041150e-01 -3.68350536e-01
3.06040525e-01 5.62031984e-01 7.89317667e-01 -8.03023934e-01
1.96780369e-01 -8.67895111e-02 -3.46160054e-01 -5.27153969e-01
1.04885745e+00 1.38225675e-01 3.37756366e-01 7.08970875e-02
5.03483534e-01 1.19267702e-01 3.66160303e-01 3.13750982e-01
1.45680368e+00 3.17032009e-01 -7.98077211e-02 5.04532419e-02
4.47441218e-03 1.01959318e-01 6.79179966e-01 9.19406533e-01
-7.07351387e-01 1.22868097e+00 3.28112841e-02 -1.00361609e+00
-1.43526506e+00 -1.38592422e+00 -2.09035445e-02 6.73097491e-01
2.53968269e-01 -5.60203612e-01 -5.79986632e-01 -1.11277595e-01
-3.99491340e-01 2.63502598e-01 -3.45119447e-01 8.84328969e-03
-8.06811452e-01 -7.61513710e-01 9.13683712e-01 8.43109071e-01
1.01696622e+00 -9.91474092e-01 -1.29689801e+00 5.30178368e-01
-3.47864807e-01 -1.66724050e+00 -4.39919412e-01 1.08649559e-01
-8.89588952e-01 -1.29214275e+00 -3.94847631e-01 -4.73222733e-01
2.70083081e-02 6.24500930e-01 1.55985785e+00 -3.38337570e-01
-4.21096355e-01 4.93131101e-01 -1.72519431e-01 -3.26487005e-01
-1.06597386e-01 -6.48052752e-01 4.42979895e-02 -1.31116793e-01
3.98589134e-01 -4.25767064e-01 -1.04645276e+00 1.27082959e-01
-1.21978676e+00 3.00233006e-01 1.72071487e-01 4.93570387e-01
4.41226453e-01 -1.47254363e-01 9.42773297e-02 -2.50610828e-01
-9.85962688e-05 -4.08912361e-01 -7.62827814e-01 -1.02341473e-01
-3.33513469e-01 -2.82149166e-01 6.15897298e-01 -4.64616030e-01
-1.37362659e+00 2.10480466e-01 1.28846601e-01 -1.02846980e+00
-1.61140785e-01 -1.11291304e-01 4.57609206e-01 -2.81533986e-01
1.00476503e+00 -1.44926280e-01 -1.10424250e-01 2.79937863e-01
1.84255287e-01 3.81414890e-01 1.37661791e+00 -3.09041172e-01
1.98221982e-01 1.14323997e+00 2.99224537e-02 -5.25135696e-01
-3.95409197e-01 -4.01337594e-01 -7.32059106e-02 -7.32262075e-01
1.04072368e+00 -1.51370668e+00 -9.68773603e-01 7.95198143e-01
-1.41995192e+00 -5.43808877e-01 -4.65944231e-01 7.82432079e-01
-6.83103144e-01 1.77468687e-01 -1.12935078e+00 -2.41966829e-01
-3.22664946e-01 -1.21969175e+00 1.21323419e+00 6.46037877e-01
1.18297018e-01 -7.50688195e-01 1.56758036e-02 1.03290863e-01
7.98074067e-01 5.32251239e-01 -1.91745713e-01 2.97969580e-01
-1.19888461e+00 1.15639947e-01 -6.19040012e-01 2.50733376e-01
-1.59030750e-01 2.99889982e-01 -1.11080647e+00 -3.38476896e-01
-8.52962583e-02 -3.57665896e-01 9.44945812e-01 7.27823853e-01
1.07347488e+00 8.11255649e-02 5.30109107e-02 1.17656815e+00
1.59142268e+00 2.28558749e-01 1.21151805e+00 4.00257885e-01
6.62053883e-01 -2.10405737e-01 2.66915590e-01 7.02780664e-01
4.07911986e-01 5.23554683e-01 5.62176168e-01 1.11648021e-02
-7.43797123e-01 9.32973623e-03 8.78584981e-01 3.05493087e-01
-3.05419654e-01 -5.76155484e-01 -7.55341470e-01 6.39218688e-01
-1.83698952e+00 -1.37560761e+00 -4.22130734e-01 1.83128583e+00
8.77670050e-01 7.32232258e-02 -1.61278486e-01 -8.93774256e-02
6.79090977e-01 2.62804478e-01 -4.22416955e-01 -2.18540251e-01
-4.56062019e-01 3.96646976e-01 8.88962686e-01 1.61219031e-01
-1.19027400e+00 8.99868548e-01 7.21346045e+00 3.65646899e-01
-1.33520079e+00 1.38754100e-01 5.11084139e-01 -7.63053238e-01
2.91460633e-01 6.96214810e-02 -7.40426660e-01 4.90645260e-01
1.70572960e+00 1.14864558e-01 6.26717091e-01 4.95015830e-01
5.92395961e-01 -3.92946780e-01 -1.10458338e+00 1.46150804e+00
2.26570442e-01 -2.01026154e+00 -2.11646017e-02 -3.15184504e-01
9.49399114e-01 5.72197139e-01 -2.34874070e-01 -6.60220385e-02
3.98193210e-01 -8.59760404e-01 9.93113756e-01 7.83738494e-01
1.29906404e+00 -5.76501250e-01 5.69260478e-01 -4.27509010e-01
-1.24392891e+00 1.40538752e-01 -4.07438695e-01 -1.71782523e-01
5.68661153e-01 4.20031220e-01 -5.51243067e-01 1.74039602e-01
1.46326101e+00 1.24692488e+00 -5.53863406e-01 1.11211610e+00
2.50969648e-01 5.77267468e-01 -4.57416385e-01 5.00626981e-01
1.84374139e-01 4.29301202e-01 4.40603286e-01 1.39878404e+00
3.21467221e-01 -5.09088188e-02 -1.89353332e-01 6.16676629e-01
-3.37408811e-01 -8.36640418e-01 -7.04279900e-01 2.95910358e-01
2.20459789e-01 9.88052726e-01 -6.72661304e-01 -6.21213615e-01
-8.34237933e-01 1.30141401e+00 -6.22467436e-02 3.96388978e-01
-1.18193448e+00 -2.90662199e-01 9.48497832e-01 -4.18970697e-02
3.18096071e-01 -3.63359302e-01 1.09954670e-01 -1.65445089e+00
-1.67597786e-01 -7.23017395e-01 2.94396818e-01 -1.48963165e+00
-1.03082204e+00 2.66058207e-01 1.05650745e-01 -1.29063714e+00
-2.69478887e-01 -8.49432230e-01 -3.36752087e-01 3.48437369e-01
-1.69503987e+00 -8.79492223e-01 -9.22268033e-01 1.15893626e+00
7.92014420e-01 -1.52882099e-01 4.65944231e-01 7.62217164e-01
-4.53835458e-01 1.22608475e-01 -3.35183367e-02 2.26949468e-01
1.08042514e+00 -1.07072282e+00 5.97825706e-01 1.47864485e+00
6.47018403e-02 2.48823509e-01 5.94886005e-01 -4.88994151e-01
-1.92829382e+00 -1.46672928e+00 4.37697977e-01 -5.34818351e-01
4.91995662e-01 -1.09995119e-02 -5.31625330e-01 1.07205176e+00
5.61970234e-01 6.67367220e-01 6.90528378e-02 -7.04923570e-01
-3.05620193e-01 -4.73345488e-01 -9.62511122e-01 4.54981655e-01
1.07284296e+00 -6.15639925e-01 -2.28301004e-01 1.11906588e-01
6.58621311e-01 -8.58438373e-01 -8.19045007e-01 3.66857976e-01
5.85861981e-01 -1.43132615e+00 1.17124403e+00 -3.35961103e-01
6.98632956e-01 -6.93386495e-01 -2.08519518e-01 -6.37464345e-01
-1.08715154e-01 -8.48675072e-01 -5.26787698e-01 8.15356255e-01
-3.67468476e-01 -1.55403182e-01 7.89378822e-01 4.75182891e-01
-3.42775851e-01 8.22271779e-02 -1.05444682e+00 -5.85120559e-01
-5.76259375e-01 -8.38198721e-01 2.51225144e-01 7.33754396e-01
-5.24987102e-01 -2.52491944e-02 -7.15167999e-01 8.79744515e-02
6.85598016e-01 -3.32582295e-01 6.83644593e-01 -5.79431713e-01
-2.18492255e-01 -7.26227760e-02 -8.74458015e-01 -1.03103352e+00
-1.68904081e-01 -3.01040649e-01 1.92413270e-01 -1.08291364e+00
1.29600748e-01 3.45110714e-01 -1.56544343e-01 2.94622719e-01
1.41698122e-01 7.58341432e-01 1.29096791e-01 2.76213437e-01
-1.16266203e+00 2.00243816e-01 8.13426137e-01 1.65692732e-01
2.36811787e-01 -9.02001023e-01 -1.49371892e-01 4.83866245e-01
5.06838024e-01 -2.71280289e-01 -2.47479707e-01 -1.04821408e+00
2.12830722e-01 4.78089988e-01 1.02721894e+00 -1.61817861e+00
7.01159835e-01 -1.11186653e-01 9.64450479e-01 -6.01842046e-01
3.89805168e-01 -7.12592781e-01 6.96321785e-01 2.15751991e-01
-8.04770738e-02 4.74970073e-01 4.68677700e-01 6.75354481e-01
-2.92111665e-01 3.06799233e-01 1.01181507e+00 -3.56711626e-01
-1.36482894e+00 3.81003499e-01 -6.70004249e-01 7.81936571e-02
8.15665305e-01 -3.46541703e-01 -7.93031096e-01 -4.27794755e-01
-2.40361825e-01 -1.81933820e-01 6.69280767e-01 5.11766911e-01
1.00725460e+00 -1.35157657e+00 -6.25932574e-01 2.26692587e-01
8.12269456e-04 -1.99184325e-02 3.15330386e-01 5.54436207e-01
-1.38104427e+00 3.13312441e-01 -7.81199396e-01 -9.53866303e-01
-9.94844913e-01 3.23596895e-01 5.54900348e-01 4.29457545e-01
-8.20913970e-01 7.67281830e-01 -1.76118106e-01 3.96282792e-01
2.05392882e-01 -6.19905889e-01 3.76305729e-01 -3.99566799e-01
8.81358266e-01 4.29371208e-01 4.06116173e-02 -4.02317077e-01
-2.58447766e-01 1.54856712e-01 -2.44989153e-02 -1.64688870e-01
1.27692795e+00 -3.42305899e-01 2.06518456e-01 2.59998113e-01
1.15486085e+00 -5.95003426e-01 -2.03930712e+00 2.35580467e-02
-4.98585910e-01 -6.91198349e-01 4.49466079e-01 -4.78809148e-01
-1.31419015e+00 7.53383160e-01 1.04876506e+00 -2.19825879e-01
1.47048748e+00 -3.77958685e-01 1.13660693e+00 3.88699949e-01
2.93610930e-01 -8.93921435e-01 2.77605027e-01 5.79565883e-01
3.78447533e-01 -1.48955226e+00 -7.72916377e-02 5.40867820e-02
-1.92898348e-01 1.65123820e+00 7.07707286e-01 -2.34468475e-01
2.56697178e-01 6.16483390e-01 -1.62550688e-01 -1.52148113e-01
-1.15502274e+00 -1.06741928e-01 -2.22841308e-01 5.81632793e-01
3.21021616e-01 -3.47414017e-01 3.96396667e-01 -1.19150035e-01
1.69559434e-01 6.60012722e-01 9.78274763e-01 1.03637445e+00
-2.06979915e-01 -4.76884693e-01 -4.33352321e-01 2.36748412e-01
-7.19779909e-01 -2.91190296e-01 2.99679011e-01 6.91359103e-01
2.08823666e-01 1.08442092e+00 6.29080474e-01 -3.87732208e-01
3.02674890e-01 -3.48182023e-01 5.92577219e-01 -5.89527860e-02
-6.78488672e-01 -2.23579079e-01 2.17316478e-01 -1.38022316e+00
-9.26615834e-01 -8.23379040e-01 -1.21140051e+00 -1.02335668e+00
1.46973252e-01 -6.29603922e-01 5.60057342e-01 5.87113023e-01
5.70624292e-01 6.47410929e-01 2.34848753e-01 -1.20360947e+00
2.14028537e-01 -5.61556280e-01 -1.27174065e-01 6.61852658e-01
5.55626094e-01 -3.23304564e-01 -4.40662831e-01 9.15783763e-01] | [8.639253616333008, -1.190571904182434] |
bf63f9ea-ff1c-4c5c-968c-a153ce25387f | bootstrapped-representations-in-reinforcement | 2306.10171 | null | https://arxiv.org/abs/2306.10171v1 | https://arxiv.org/pdf/2306.10171v1.pdf | Bootstrapped Representations in Reinforcement Learning | In reinforcement learning (RL), state representations are key to dealing with large or continuous state spaces. While one of the promises of deep learning algorithms is to automatically construct features well-tuned for the task they try to solve, such a representation might not emerge from end-to-end training of deep RL agents. To mitigate this issue, auxiliary objectives are often incorporated into the learning process and help shape the learnt state representation. Bootstrapping methods are today's method of choice to make these additional predictions. Yet, it is unclear which features these algorithms capture and how they relate to those from other auxiliary-task-based approaches. In this paper, we address this gap and provide a theoretical characterization of the state representation learnt by temporal difference learning (Sutton, 1988). Surprisingly, we find that this representation differs from the features learned by Monte Carlo and residual gradient algorithms for most transition structures of the environment in the policy evaluation setting. We describe the efficacy of these representations for policy evaluation, and use our theoretical analysis to design new auxiliary learning rules. We complement our theoretical results with an empirical comparison of these learning rules for different cumulant functions on classic domains such as the four-room domain (Sutton et al, 1999) and Mountain Car (Moore, 1990). | ['Will Dabney', 'Marc G. Bellemare', 'Rishabh Agarwal', 'Anna Harutyunyan', 'Mark Rowland', 'Stephen Tu', 'Charline Le Lan'] | 2023-06-16 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 3.25901471e-02 8.58919322e-02 -4.62387145e-01 -8.80143791e-02
-8.54348004e-01 -8.20665121e-01 8.75327110e-01 1.26795173e-01
-7.64944673e-01 1.30361974e+00 2.96281189e-01 -6.39530003e-01
-3.13502014e-01 -5.04462242e-01 -7.01455355e-01 -8.17843258e-01
-4.24275428e-01 5.37501156e-01 5.53281941e-02 -5.49155295e-01
3.77775699e-01 5.38899720e-01 -1.46056354e+00 -1.52408391e-01
7.97439575e-01 8.62065613e-01 1.80252165e-01 7.46062398e-01
2.38399357e-01 9.80022967e-01 -4.79803950e-01 2.22763509e-01
3.89514178e-01 -5.83473980e-01 -8.64797831e-01 -1.69870764e-01
-9.88398865e-02 -4.15151745e-01 -3.86916578e-01 8.60599399e-01
4.50395137e-01 6.79814339e-01 8.95541370e-01 -1.10657799e+00
-9.34038982e-02 5.12288392e-01 -1.59061760e-01 3.77384722e-01
3.45890932e-02 4.86682713e-01 1.00470912e+00 -2.65645951e-01
4.80359465e-01 1.29597938e+00 4.96919006e-01 7.53145218e-01
-1.40074527e+00 -4.87403601e-01 4.65266019e-01 1.38950408e-01
-7.08432376e-01 -5.41909158e-01 5.98382950e-01 -2.63817370e-01
9.53829765e-01 -1.64185651e-02 7.84333229e-01 1.39234543e+00
2.96199828e-01 1.05839300e+00 1.43118799e+00 -4.47092265e-01
6.72044456e-01 1.67652331e-02 -6.61842450e-02 6.07928216e-01
1.36500373e-01 9.21690226e-01 -9.64801237e-02 -4.19230521e-01
8.57523620e-01 -1.93630204e-01 -2.12534696e-01 -8.13276410e-01
-8.87981713e-01 1.17101371e+00 2.13873610e-01 1.24592796e-01
-5.45282483e-01 4.46146578e-01 5.26037037e-01 4.95509565e-01
2.27290139e-01 6.59497201e-01 -6.58252776e-01 -4.87905294e-01
-4.40492541e-01 6.23104155e-01 8.80596697e-01 4.72908705e-01
7.10449815e-01 3.49515587e-01 -2.05576286e-01 3.92277092e-01
2.90389925e-01 4.41463590e-01 6.38752401e-01 -1.36808300e+00
2.70343065e-01 -5.04489765e-02 5.83880723e-01 -2.42697239e-01
-4.93596017e-01 -5.72252035e-01 -4.48668212e-01 6.61432862e-01
6.94231749e-01 -5.18704116e-01 -8.58017802e-01 2.13604903e+00
2.58209646e-01 1.19992517e-01 3.18041056e-01 6.61842048e-01
-2.13042945e-01 5.83139300e-01 3.58003378e-03 -4.60650533e-01
8.05086374e-01 -7.49887526e-01 -5.14046907e-01 -4.44535047e-01
7.20300078e-01 -3.23838681e-01 9.62137520e-01 2.88595706e-01
-9.20465231e-01 -4.27848130e-01 -8.96826923e-01 3.91047537e-01
-9.55989137e-02 -2.09263265e-01 8.45529974e-01 5.52957714e-01
-9.36049223e-01 1.16817701e+00 -1.13212073e+00 -2.85962194e-01
2.87629217e-01 3.36870074e-01 3.10648587e-02 4.35225487e-01
-1.32050884e+00 1.38988340e+00 4.52129006e-01 -1.42811343e-01
-1.34298611e+00 -1.67888030e-01 -6.51374876e-01 -4.44903821e-02
5.76234162e-01 -4.76323634e-01 1.87694561e+00 -1.06893277e+00
-1.99974716e+00 1.82500228e-01 -5.11037558e-03 -7.53233135e-01
3.22481006e-01 -9.10314992e-02 -3.53811644e-02 1.32178199e-02
-1.77156329e-02 3.57207388e-01 1.10929358e+00 -1.24290335e+00
-7.54242659e-01 -2.30551749e-01 4.28221643e-01 4.04610753e-01
2.60392100e-01 -3.47487003e-01 4.37481821e-01 -4.66763556e-01
-3.94982219e-01 -1.13583636e+00 -8.10741603e-01 -4.21208858e-01
3.90335359e-02 -1.97019428e-01 5.47186732e-01 -3.92994761e-01
8.90119135e-01 -1.88923144e+00 1.76737174e-01 3.00494492e-01
-1.14270434e-01 3.44906032e-01 -1.70993194e-01 5.75091541e-01
-1.44029215e-01 -2.84103543e-01 -2.85247445e-01 -2.66433358e-02
2.58006930e-01 5.58024406e-01 -7.48397112e-01 6.10477209e-01
2.91141868e-02 9.26817536e-01 -1.09099185e+00 -1.51681945e-01
2.88083345e-01 1.14909820e-01 -6.94324732e-01 1.97612643e-01
-4.83710170e-01 6.90979719e-01 -8.81545663e-01 7.94410333e-03
-1.96180698e-02 -7.21381381e-02 3.78757715e-01 4.02218282e-01
-1.38641089e-01 6.75429702e-01 -1.21468806e+00 1.36968744e+00
-4.76530343e-01 3.13854307e-01 1.75443619e-01 -1.51272166e+00
6.24520123e-01 3.06447834e-01 5.01460731e-01 -6.08890533e-01
1.81207582e-01 2.76855715e-02 3.82657498e-01 -2.84921974e-01
4.28048283e-01 -4.06736970e-01 -2.15794533e-01 8.50082397e-01
1.53663158e-02 -3.84425908e-01 1.91411525e-01 -5.11580445e-02
1.05395150e+00 5.86896420e-01 5.91514230e-01 -3.59344721e-01
1.48260459e-01 2.00195312e-01 5.58987677e-01 1.05410409e+00
-5.13603210e-01 -1.15175560e-01 7.47850716e-01 -5.13313293e-01
-9.82398152e-01 -1.02584183e+00 -2.72444542e-02 1.16501319e+00
-2.45158792e-01 -2.30109334e-01 -6.23468339e-01 -7.47848034e-01
4.12386376e-03 9.68274891e-01 -7.10940659e-01 -4.66343105e-01
-5.74298859e-01 -6.71419084e-01 3.14992815e-01 5.10876536e-01
2.05149487e-01 -1.14590669e+00 -1.22878122e+00 5.70726156e-01
7.45597035e-02 -5.90236247e-01 -2.19313622e-01 8.43356490e-01
-1.08975220e+00 -1.12329316e+00 -4.47279423e-01 -3.09801519e-01
2.52730608e-01 -1.59799278e-01 1.03397715e+00 -4.06739682e-01
-1.54907191e-02 7.30312884e-01 -2.25315779e-01 -3.26321930e-01
-6.69634163e-01 -3.96631286e-02 3.98506701e-01 -3.84987652e-01
1.22491419e-02 -6.04471624e-01 -5.61320961e-01 5.60083613e-02
-7.90746987e-01 -2.06155747e-01 6.06118679e-01 1.13992095e+00
3.49932373e-01 -2.24530101e-02 4.80067849e-01 -5.74526727e-01
1.08011377e+00 -3.95960987e-01 -9.13098752e-01 1.21804319e-01
-6.94605052e-01 9.47798789e-01 7.87791312e-01 -6.89774513e-01
-9.61199284e-01 4.59115766e-02 -1.03114478e-01 -2.95154095e-01
-1.28159255e-01 4.08692986e-01 3.58146608e-01 1.57051563e-01
7.22977400e-01 2.71566480e-01 3.06851089e-01 -3.44926506e-01
4.65546519e-01 2.55854219e-01 1.39824390e-01 -1.22294545e+00
7.45605230e-01 1.85193464e-01 9.85947475e-02 -5.12485445e-01
-9.18471694e-01 2.45391261e-02 -2.34620824e-01 -5.43820746e-02
6.37312472e-01 -5.86744428e-01 -9.69944298e-01 1.63716562e-02
-7.28699803e-01 -9.99078155e-01 -8.68519247e-01 6.90097272e-01
-1.30353773e+00 2.34193534e-01 -4.90883768e-01 -1.21536088e+00
9.73757431e-02 -1.28602886e+00 8.47232163e-01 1.05147518e-01
-1.38635918e-01 -1.06766403e+00 6.14846349e-01 -2.96292126e-01
4.45989460e-01 6.06032461e-02 1.04831851e+00 -6.75475478e-01
-2.41780192e-01 1.93927094e-01 3.69398564e-01 3.72163713e-01
6.64787292e-02 -2.35208675e-01 -9.04600978e-01 -6.05050087e-01
2.73312092e-01 -7.05023289e-01 9.72112536e-01 6.44003332e-01
9.83472526e-01 -3.12291533e-01 -1.03361040e-01 2.08320200e-01
1.16220605e+00 5.16943693e-01 4.23891813e-01 4.76537734e-01
2.64797546e-02 5.68971276e-01 6.17604911e-01 5.72177768e-01
2.41737306e-01 5.68870068e-01 4.35073495e-01 3.52547646e-01
5.19605637e-01 -4.99637365e-01 7.71624386e-01 2.98340082e-01
-2.09151193e-01 1.27859697e-01 -6.79426610e-01 1.07806824e-01
-2.09554029e+00 -1.34755671e+00 5.34459531e-01 2.38014579e+00
8.60113025e-01 3.49342406e-01 3.90822828e-01 -2.89742127e-02
4.94776577e-01 2.73169160e-01 -9.97549415e-01 -5.38559198e-01
1.68273434e-01 4.87700284e-01 5.59627354e-01 6.64997280e-01
-1.09004736e+00 9.86597598e-01 6.98668861e+00 7.57506669e-01
-9.18906748e-01 -6.19947799e-02 6.49151325e-01 6.54983521e-02
-1.29818255e-02 3.12733918e-01 -7.67390370e-01 2.15240389e-01
1.30542862e+00 -7.34362528e-02 8.83926988e-01 8.92648458e-01
3.72703493e-01 -2.74914563e-01 -1.24740338e+00 7.14856148e-01
-4.72538918e-01 -1.09065461e+00 -3.67834777e-01 2.69740760e-01
8.10075819e-01 1.29499048e-01 2.66184419e-01 9.50765014e-01
1.03283250e+00 -1.09715009e+00 6.73321903e-01 3.02837163e-01
3.61673385e-01 -7.74412572e-01 2.87520528e-01 6.05415881e-01
-8.67086470e-01 -5.36729872e-01 -2.98238456e-01 -4.41931516e-01
-5.04992828e-02 -6.20114841e-02 -7.23080218e-01 5.45487627e-02
2.03845456e-01 4.43509847e-01 -1.78780988e-01 7.19480455e-01
-2.34342828e-01 6.81765258e-01 -4.17728364e-01 -2.79179960e-01
7.24174261e-01 -2.79111773e-01 5.10989070e-01 5.35962403e-01
1.53250575e-01 -2.27495879e-02 3.48830879e-01 7.26601064e-01
3.46239865e-01 -3.11757028e-01 -8.21461856e-01 -3.12301934e-01
1.71317473e-01 8.34410250e-01 -6.21850550e-01 -2.80352443e-01
-2.94924885e-01 5.93372822e-01 5.34980774e-01 6.53004467e-01
-7.43040800e-01 -7.02961683e-02 1.01598597e+00 -1.26870736e-01
4.41177964e-01 -4.15366590e-01 2.70346105e-01 -1.19704759e+00
-2.42483765e-01 -1.15380228e+00 2.84776241e-01 -3.08133990e-01
-1.03489077e+00 2.07064405e-01 3.42259437e-01 -1.12922978e+00
-1.20145166e+00 -7.56920636e-01 -5.91217518e-01 6.57413304e-01
-1.39442885e+00 -4.42100167e-01 6.50875747e-01 6.90814137e-01
5.14998257e-01 -2.27336168e-01 9.58732545e-01 -4.22556400e-01
-4.29736346e-01 1.60815462e-01 7.56816983e-01 2.32649148e-02
2.99175262e-01 -1.41878378e+00 4.28915590e-01 5.38492382e-01
1.98460549e-01 5.73799908e-01 1.03055465e+00 -3.66963774e-01
-1.55268347e+00 -7.43175089e-01 1.20540820e-02 -4.91693288e-01
8.54323089e-01 -6.02869838e-02 -5.74072838e-01 8.65594804e-01
5.65367267e-02 1.29603606e-03 1.85452253e-01 2.83985943e-01
-1.10080123e-01 1.46983191e-01 -8.47618222e-01 8.01392138e-01
8.44997108e-01 -5.95579267e-01 -8.28216434e-01 1.10114306e-01
3.95168334e-01 -3.71681750e-01 -4.55839694e-01 3.07551861e-01
4.93502468e-01 -8.91523182e-01 8.56080413e-01 -1.24964976e+00
5.60033135e-02 -4.15795781e-02 -1.10851906e-01 -1.95979857e+00
-2.91797221e-01 -9.47625518e-01 -4.54366952e-01 4.73870426e-01
3.32374513e-01 -8.94001901e-01 7.81539559e-01 4.35875833e-01
-5.62212877e-02 -7.52937913e-01 -1.03235328e+00 -1.01803303e+00
6.16022348e-01 -4.85052437e-01 2.82799184e-01 5.57114363e-01
1.68056339e-01 4.42317903e-01 -2.61544675e-01 -1.38223499e-01
6.29816055e-01 3.03739160e-01 7.16364324e-01 -1.05043387e+00
-6.95139945e-01 -6.57304406e-01 -3.22784521e-02 -1.24690080e+00
6.00805640e-01 -6.41496420e-01 2.14723796e-01 -1.28011286e+00
-7.11842030e-02 -5.07578135e-01 -4.92388636e-01 3.24103028e-01
-7.36142322e-02 -6.35328412e-01 1.74341500e-01 -1.24322228e-01
-5.22375286e-01 9.01073515e-01 1.25924444e+00 6.68770354e-03
-3.17627788e-01 3.99979472e-01 -6.41434312e-01 7.41617501e-01
1.14360011e+00 -4.64964718e-01 -7.07784653e-01 -8.20116401e-02
1.62796050e-01 3.82887810e-01 3.11700404e-01 -8.01529169e-01
-1.23643950e-01 -6.32226944e-01 3.24904561e-01 -1.52676314e-01
4.46672231e-01 -6.75307691e-01 -3.16503316e-01 9.08735514e-01
-6.90813541e-01 8.35876465e-02 2.19097957e-01 7.78510928e-01
8.79387558e-02 -4.60331738e-01 8.95398200e-01 -3.61784846e-01
-6.04573727e-01 3.11181158e-01 -7.85299420e-01 6.23058498e-01
8.33448946e-01 1.13136880e-01 -1.44727036e-01 -7.60580480e-01
-7.51951575e-01 1.99140549e-01 2.46077865e-01 1.80355832e-01
4.63168502e-01 -1.18387115e+00 -3.95235509e-01 9.90239754e-02
-2.12335110e-01 -2.29229733e-01 -2.32536569e-01 7.11916029e-01
-9.53210518e-03 5.67836583e-01 -3.35105509e-01 -1.89746007e-01
-5.99431992e-01 6.52490795e-01 4.92578000e-01 -7.46354580e-01
-5.16051650e-01 3.39455724e-01 1.96100876e-01 -3.85080010e-01
2.79964715e-01 -5.48979402e-01 -1.19627528e-02 -6.29772320e-02
4.01208311e-01 1.27801359e-01 -2.99485028e-01 -2.84383446e-01
7.86686242e-02 1.36513516e-01 -5.41376099e-02 -6.46159112e-01
1.41931450e+00 6.22643679e-02 5.09844601e-01 6.37664437e-01
7.99836278e-01 -4.00291681e-01 -1.80179811e+00 -2.34462261e-01
9.79867801e-02 -1.42530933e-01 6.37501404e-02 -4.93702352e-01
-6.28039300e-01 8.85856628e-01 7.22798049e-01 2.64813453e-01
7.29820311e-01 -2.25655630e-01 4.08379555e-01 9.65049148e-01
6.22644305e-01 -1.44145358e+00 1.52810648e-01 8.85272086e-01
6.64212823e-01 -1.22899401e+00 -1.08274274e-01 7.22176731e-01
-8.59733701e-01 9.70239282e-01 4.12574232e-01 -3.95646930e-01
5.49123764e-01 2.28178784e-01 -3.03915948e-01 1.28427938e-01
-1.14704323e+00 -6.41999185e-01 -1.32178143e-01 7.25777090e-01
9.04238075e-02 6.36209026e-02 -3.37313488e-02 1.62039503e-01
-4.55295108e-02 -2.03983247e-01 3.92667949e-01 1.33677316e+00
-7.98246861e-01 -1.25933516e+00 -3.01035434e-01 5.27416289e-01
-3.52203250e-01 3.41178551e-02 -3.13217565e-02 8.93243253e-01
-4.88023996e-01 7.48252392e-01 -1.47830889e-01 -1.98935326e-02
1.30434021e-01 3.52252126e-01 8.14054072e-01 -6.19330645e-01
-4.28170532e-01 2.61326376e-02 2.05450818e-01 -7.21073449e-01
-2.69366652e-01 -7.72541225e-01 -1.16577280e+00 -1.49373710e-01
-1.03250593e-01 5.40601552e-01 5.15731633e-01 1.13015950e+00
3.22615989e-02 4.53075826e-01 7.77361214e-01 -1.12143731e+00
-1.54485464e+00 -8.52251530e-01 -5.88909566e-01 2.77869225e-01
7.48088181e-01 -1.00285316e+00 -4.08104390e-01 -3.66086692e-01] | [4.07340145111084, 1.8553258180618286] |
766055f3-aaec-4608-8112-c7679ba7a564 | diffusion-models-and-semi-supervised-learners | 2302.10586 | null | https://arxiv.org/abs/2302.10586v2 | https://arxiv.org/pdf/2302.10586v2.pdf | Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labels | In an effort to further advance semi-supervised generative and classification tasks, we propose a simple yet effective training strategy called dual pseudo training (DPT), built upon strong semi-supervised learners and diffusion models. DPT operates in three stages: training a classifier on partially labeled data to predict pseudo-labels; training a conditional generative model using these pseudo-labels to generate pseudo images; and retraining the classifier with a mix of real and pseudo images. Empirically, DPT consistently achieves SOTA performance of semi-supervised generation and classification across various settings. In particular, with one or two labels per class, DPT achieves a Fr\'echet Inception Distance (FID) score of 3.08 or 2.52 on ImageNet 256x256, surpassing strong diffusion models with full labels, such as IDDPM, CDM, ADM, and LDM. Besides, DPT outperforms competitive semi-supervised baselines substantially on ImageNet classification tasks, achieving top-1 accuracies of 59.0 (+2.8), 69.5 (+3.0), and 74.4 (+2.0) with one, two, or five labels per class, respectively. Notably, our results demonstrate that diffusion can generate realistic images with only a few labels (e.g., <0.1%) and generative augmentation remains viable for semi-supervised classification. | ['Jun Zhu', 'Chongxuan Li', 'Jiacheng Sun', 'Fan Bao', 'Yong Zhong', 'Zebin You'] | 2023-02-21 | null | null | null | null | ['semi-supervised-image-classification', 'conditional-image-generation'] | ['computer-vision', 'computer-vision'] | [ 4.72117871e-01 4.67758983e-01 -4.17479455e-01 -3.86637568e-01
-1.06036079e+00 -5.28553724e-01 1.07918572e+00 -3.14412832e-01
-3.78743440e-01 9.25513804e-01 -9.93963610e-03 -4.26097363e-01
3.19718510e-01 -7.64059246e-01 -7.79485106e-01 -7.43404567e-01
1.56957090e-01 7.48578966e-01 -1.39646605e-01 1.30511567e-01
-1.54055893e-01 6.31474108e-02 -1.20138264e+00 3.67405742e-01
9.73911703e-01 9.51341152e-01 8.98701325e-02 7.50260413e-01
-3.77094969e-02 1.06479716e+00 -6.47653699e-01 -5.56812167e-01
2.74806172e-01 -5.92791080e-01 -8.26919079e-01 5.41423142e-01
3.92972857e-01 -5.70856512e-01 -3.62366706e-01 8.94520640e-01
4.64428037e-01 -2.01048046e-01 1.14986789e+00 -1.28289628e+00
-1.27159667e+00 8.99160981e-01 -8.19365501e-01 2.37444621e-02
5.63172065e-02 5.28435469e-01 6.95318162e-01 -1.10268259e+00
6.12828672e-01 1.11937964e+00 6.59634590e-01 8.34165037e-01
-1.49813497e+00 -8.62684608e-01 -1.39143601e-01 -5.63987434e-01
-1.51581800e+00 -4.21507120e-01 5.23740470e-01 -5.49504578e-01
6.02720916e-01 -1.83489114e-01 2.67727584e-01 1.52098322e+00
2.76062842e-02 8.15654457e-01 1.67777216e+00 -3.93576473e-01
2.14328840e-01 3.24223250e-01 -5.35528734e-02 7.29129910e-01
4.73335199e-02 1.76071018e-01 -2.58120894e-01 -1.82581291e-01
7.39534795e-01 -7.89645910e-02 -1.46226380e-02 1.51118547e-01
-1.31124198e+00 1.00339484e+00 6.08127415e-01 -1.60146244e-02
-4.26542401e-01 2.66696841e-01 5.55792302e-02 1.00596689e-01
9.13434386e-01 1.74070552e-01 -3.42395566e-02 6.77980185e-02
-9.41926241e-01 -8.41052458e-02 4.99371201e-01 1.18289304e+00
7.98982263e-01 4.55899656e-01 -3.36738139e-01 9.54697132e-01
1.21718034e-01 7.44583249e-01 5.57307363e-01 -9.75657582e-01
3.39133799e-01 2.02296078e-01 1.44591015e-02 -3.02522868e-01
-6.75009787e-02 -7.36033559e-01 -1.09539258e+00 1.18462831e-01
2.63195366e-01 -4.73721802e-01 -1.53151453e+00 1.76664734e+00
2.26701479e-02 3.66983116e-01 2.79584557e-01 4.59524661e-01
8.89094472e-01 7.73883045e-01 4.50253367e-01 -5.90624847e-02
1.02330327e+00 -1.23529351e+00 -3.83172393e-01 -5.18725514e-01
7.62958527e-01 -7.96328902e-01 1.09064674e+00 2.40237489e-01
-1.03060055e+00 -8.79307151e-01 -8.78922224e-01 3.78125280e-01
-8.65232125e-02 3.57868701e-01 5.89982390e-01 6.90939128e-01
-1.36821377e+00 4.72089082e-01 -7.65056431e-01 -1.76810101e-02
8.81518066e-01 2.38482937e-01 -2.05344439e-01 -5.75192332e-01
-9.40546215e-01 5.90062559e-01 3.73416603e-01 -3.00554603e-01
-1.49655688e+00 -7.21002162e-01 -8.11873496e-01 -4.02102858e-01
-1.58326387e-01 -6.06476963e-01 1.17371035e+00 -9.72953558e-01
-1.22274482e+00 1.15869915e+00 1.39724374e-01 -7.64664114e-01
8.15622628e-01 -2.79878061e-02 -4.54547107e-01 1.30144387e-01
5.37290931e-01 1.45804548e+00 9.20439482e-01 -1.59223866e+00
-4.56064999e-01 -5.69464266e-02 -6.11295253e-02 1.13721259e-01
-2.84070551e-01 -3.35554391e-01 -2.69931614e-01 -8.62171352e-01
2.41680741e-02 -1.28121698e+00 -2.45432571e-01 -2.42358267e-01
-8.15258920e-01 -1.61752120e-01 8.77126217e-01 -4.51009035e-01
6.73768401e-01 -2.24055815e+00 -2.26936877e-01 -6.35838360e-02
4.89381224e-01 4.83037651e-01 -3.94936562e-01 4.81684618e-02
-7.18754828e-02 2.18207508e-01 -4.38638240e-01 -7.81640291e-01
-1.72820330e-01 2.72882074e-01 -3.72249842e-01 2.10821882e-01
3.85897100e-01 1.20482266e+00 -9.83671308e-01 -4.11771357e-01
8.00574198e-02 6.85087979e-01 -3.05654794e-01 3.05211078e-02
-2.18251809e-01 5.64580023e-01 -1.35784790e-01 6.22122467e-01
7.08393157e-01 -8.13926756e-01 -1.42360246e-02 1.52921557e-01
3.25986952e-01 3.51201370e-02 -7.45351493e-01 1.51705933e+00
-5.44801295e-01 6.56391323e-01 -3.44320029e-01 -5.73245406e-01
1.02616405e+00 2.10114211e-01 2.79817909e-01 -5.49013495e-01
7.46288290e-03 2.37115279e-01 -2.38977939e-01 -5.37412204e-02
2.93525994e-01 -2.61714250e-01 3.04802228e-02 9.19089735e-01
3.73149902e-01 -6.17676042e-03 1.09060124e-01 4.89725202e-01
1.01299167e+00 1.09133296e-01 -1.87012807e-01 -7.21913204e-02
-1.45964967e-02 6.90715909e-02 2.61785805e-01 8.06649148e-01
-6.19894899e-02 7.69151509e-01 2.69471765e-01 -2.48769194e-01
-1.10051596e+00 -1.22619379e+00 -2.93349206e-01 8.92682672e-01
-1.63589872e-03 -2.49544144e-01 -9.18057561e-01 -1.09655416e+00
-4.01586220e-02 8.24986875e-01 -7.76081026e-01 -2.90177733e-01
-2.23736137e-01 -1.10056615e+00 8.03879023e-01 6.34980559e-01
8.74415040e-01 -9.83555079e-01 1.62324905e-01 3.07572931e-02
-1.22991636e-01 -1.16032350e+00 -4.72698361e-01 2.17424631e-01
-7.62639284e-01 -6.49153352e-01 -9.82929409e-01 -9.10868227e-01
1.00964022e+00 4.21540856e-01 1.17704213e+00 -1.16445452e-01
3.70559953e-02 1.60314396e-01 -3.65846097e-01 -1.45336995e-02
-9.44013953e-01 1.02953203e-01 -2.74363253e-02 6.00945875e-02
6.92823827e-02 -5.86206973e-01 -7.44734287e-01 4.23442096e-01
-7.99751639e-01 3.43649060e-01 8.97485614e-01 9.79157209e-01
7.14504838e-01 -5.73210381e-02 8.24696183e-01 -1.37303650e+00
5.32319307e-01 -7.73119748e-01 -2.17554957e-01 1.00289196e-01
-1.07792091e+00 3.68452910e-03 6.01058722e-01 -8.36777210e-01
-1.18996298e+00 2.76774764e-02 -1.92384437e-01 -6.79018021e-01
-2.73616612e-01 3.24079722e-01 2.34166160e-01 5.74513860e-02
1.18501520e+00 4.49146330e-01 -7.83391949e-03 -1.91315338e-01
5.58979630e-01 8.36156428e-01 5.94714642e-01 -3.83649886e-01
9.32107985e-01 5.49431860e-01 -3.41197968e-01 -4.10359472e-01
-1.29803669e+00 -9.38413106e-03 -4.84393179e-01 1.19228975e-03
9.04104948e-01 -1.32762647e+00 1.62896052e-01 6.47090614e-01
-6.79608524e-01 -8.53728652e-01 -4.67893690e-01 4.41576928e-01
-3.43677789e-01 1.05481707e-01 -9.73275959e-01 -5.07936597e-01
-3.89432490e-01 -1.08820438e+00 1.04575169e+00 7.14152977e-02
-1.43079191e-01 -1.20340896e+00 -1.62370980e-01 6.27075613e-01
4.91535783e-01 3.37709934e-01 5.44620752e-01 -6.37656450e-01
-5.09464681e-01 -2.50847220e-01 -3.82942557e-01 7.92580903e-01
2.04206586e-01 -2.18454868e-01 -1.12577426e+00 -3.62931013e-01
-2.86281526e-01 -9.51841056e-01 9.81530786e-01 2.65215576e-01
1.03130591e+00 -2.38013983e-01 -4.15483564e-01 6.51910543e-01
1.17471802e+00 5.63178770e-02 5.67241490e-01 -1.44297302e-01
6.75657690e-01 1.60303652e-01 4.23668772e-01 3.51897866e-01
4.67415571e-01 2.34879345e-01 2.40614250e-01 -3.39372069e-01
-8.38065624e-01 -6.78850234e-01 3.62171113e-01 5.80630004e-01
1.86825365e-01 -2.97786295e-01 -9.59020078e-01 3.95767003e-01
-1.34625614e+00 -7.86438048e-01 -1.89274669e-01 2.07748961e+00
1.10892856e+00 5.14207602e-01 9.24461056e-03 9.62124672e-03
8.78782034e-01 9.95882973e-02 -7.14006662e-01 2.12211579e-01
-3.00203174e-01 4.25372094e-01 5.11998892e-01 3.25017095e-01
-1.27791226e+00 1.11019421e+00 6.65344429e+00 9.05512452e-01
-1.18588650e+00 2.91871458e-01 1.34599352e+00 2.24105150e-01
-2.90991753e-01 2.74161430e-04 -9.38092887e-01 7.48075187e-01
1.13990057e+00 6.14131838e-02 2.82760471e-01 1.00458384e+00
-3.16269338e-01 -2.65188273e-02 -8.58361840e-01 1.01406038e+00
2.82182395e-01 -1.40779412e+00 2.80430764e-01 3.15439284e-01
1.50100839e+00 4.27266359e-01 5.94308913e-01 5.30996084e-01
9.64427471e-01 -1.25191319e+00 7.12173522e-01 1.58111632e-01
1.37787795e+00 -5.04169524e-01 6.41956687e-01 6.49551332e-01
-7.61182070e-01 1.05803862e-01 -1.72418222e-01 1.70997068e-01
1.04404703e-01 6.93030715e-01 -9.92912054e-01 1.89529330e-01
3.59943748e-01 7.46063948e-01 -7.44593740e-01 5.94177485e-01
-5.86189508e-01 1.02571177e+00 -2.13830397e-01 3.19877654e-01
3.89277667e-01 -2.68094125e-03 -1.20007575e-01 1.24565208e+00
3.52158636e-01 7.20355436e-02 2.78405279e-01 9.26114619e-01
-6.19779110e-01 -4.03969318e-01 -4.72377807e-01 -2.56321710e-02
5.58095694e-01 1.21314478e+00 -8.59250009e-01 -6.39196336e-01
-1.51396856e-01 1.09465802e+00 2.60824025e-01 4.99406993e-01
-1.14159369e+00 -1.46427304e-01 9.31123197e-02 9.95284170e-02
9.19515267e-02 -1.17678074e-02 -2.58885473e-01 -1.09374595e+00
-3.40027303e-01 -6.82728887e-01 3.00831586e-01 -1.00596952e+00
-1.38188779e+00 9.99946594e-01 -1.68886870e-01 -1.16989875e+00
-3.51865560e-01 -3.49255443e-01 -4.93426859e-01 8.88046026e-01
-1.47325373e+00 -1.42837811e+00 -5.62145233e-01 5.01685262e-01
5.84249496e-01 -2.76208401e-01 7.89630234e-01 3.19680303e-01
-5.36140621e-01 6.45027816e-01 1.01609277e-02 3.68664026e-01
7.26079047e-01 -1.29736185e+00 6.37384415e-01 6.83361471e-01
4.35679615e-01 2.79199839e-01 3.69333357e-01 -6.41181767e-01
-8.08203399e-01 -1.64373541e+00 7.25055993e-01 -5.85795999e-01
4.10620064e-01 -3.29934359e-01 -6.65530205e-01 9.24130440e-01
3.07703465e-01 3.30076844e-01 7.52175272e-01 -2.51220971e-01
-5.54250360e-01 9.03616473e-02 -1.21669209e+00 4.36990708e-01
1.12705243e+00 -5.47168612e-01 -9.91164446e-02 8.46416771e-01
6.68972611e-01 -4.13903236e-01 -8.51385057e-01 4.34870422e-01
1.02279000e-01 -8.50892365e-01 9.79836166e-01 -2.79767811e-01
7.23063052e-01 -2.82805711e-02 -2.84704547e-02 -1.39175355e+00
-3.04612070e-01 -6.69441462e-01 -9.01588276e-02 1.32265007e+00
7.06259668e-01 -6.49704158e-01 9.90323246e-01 4.28441435e-01
-1.87340245e-01 -8.14928353e-01 -4.74063396e-01 -8.18041623e-01
2.65479535e-01 -4.02652234e-01 2.97995985e-01 1.16533208e+00
-6.05910063e-01 6.23727202e-01 -4.90047187e-01 -1.09704621e-01
8.67184103e-01 5.36784343e-02 7.35502005e-01 -9.46371496e-01
-3.83409709e-01 -2.33612299e-01 -2.17927783e-03 -1.31015956e+00
4.25137162e-01 -1.26401198e+00 -1.82286605e-01 -1.58137870e+00
3.24635088e-01 -1.06165802e+00 -2.35463262e-01 7.48311639e-01
-2.22947970e-01 9.68455672e-01 -2.06216842e-01 7.00158834e-01
-4.24249321e-01 5.94765186e-01 1.37167203e+00 -2.81913579e-01
1.47182848e-02 8.72637555e-02 -8.75384152e-01 6.08581364e-01
9.69763577e-01 -5.69849193e-01 -6.86674297e-01 -5.29244184e-01
-2.04569727e-01 -4.11417633e-02 3.89976859e-01 -1.01754522e+00
-1.61344241e-02 7.60824000e-03 5.83170950e-01 -3.63646358e-01
3.97513509e-01 -1.42838076e-01 1.95852146e-01 4.67780948e-01
-4.78209972e-01 -1.30567998e-01 -8.68382305e-02 6.15585268e-01
-1.91319183e-01 -8.46891627e-02 1.04197299e+00 -1.13153502e-01
-6.32183671e-01 4.54039782e-01 -8.41531232e-02 3.20053518e-01
1.13612771e+00 -9.36408341e-02 -6.07634842e-01 -4.23561841e-01
-8.32290411e-01 1.26620233e-01 3.48201782e-01 2.63277799e-01
7.18883276e-01 -1.43226922e+00 -9.30239499e-01 2.46247694e-01
7.50118718e-02 1.58367008e-01 2.30520189e-01 5.78706801e-01
-4.52153057e-01 2.19801933e-01 -1.28075415e-02 -7.20345497e-01
-7.49923348e-01 4.29979503e-01 1.01871617e-01 -3.36348534e-01
-4.59359467e-01 1.07119334e+00 3.15854371e-01 -4.57527816e-01
-5.04713915e-02 1.21903598e-01 1.86666816e-01 -2.23332822e-01
3.65929365e-01 1.83883369e-01 -2.15505630e-01 -6.86879277e-01
1.07894920e-01 2.22326800e-01 -2.81906754e-01 -2.65605628e-01
1.11300099e+00 1.83738649e-01 2.80577868e-01 1.73912510e-01
1.16593695e+00 -1.26366720e-01 -1.80614197e+00 -3.58308971e-01
-3.44817370e-01 -1.82224721e-01 -4.47625555e-02 -1.06307340e+00
-1.12984884e+00 6.87026739e-01 4.14228022e-01 5.18821217e-02
9.25975502e-01 3.33325654e-01 7.61350095e-01 4.82405256e-03
4.06089157e-01 -6.46575928e-01 5.51055074e-01 2.27277428e-01
6.28500640e-01 -1.39936543e+00 -1.34806782e-01 -3.78179312e-01
-9.33302581e-01 4.78628337e-01 7.37977862e-01 -1.43918619e-01
6.20955229e-01 2.91432649e-01 2.24082202e-01 1.23741068e-02
-8.42027962e-01 -1.13956943e-01 1.43198609e-01 6.32527769e-01
3.72577906e-01 2.51369298e-01 2.20330179e-01 9.49904844e-02
-1.95488736e-01 -3.86312194e-02 3.34096223e-01 8.52063954e-01
-2.27577761e-01 -1.11969137e+00 -1.22116514e-01 6.71237946e-01
-3.50821316e-01 -2.70060748e-01 -2.39679322e-01 6.86730444e-01
1.55196756e-01 9.40905869e-01 1.97750945e-02 -6.24574304e-01
-1.49598554e-01 1.11810930e-01 2.15280861e-01 -7.59575605e-01
-2.83904284e-01 1.64687797e-01 -2.81723235e-02 -3.07576321e-02
-4.94793087e-01 -4.96867418e-01 -1.09200227e+00 -2.98837572e-01
-3.74481946e-01 1.11869521e-01 4.88727331e-01 7.27793097e-01
5.16847610e-01 3.38999331e-01 6.93392575e-01 -5.85292041e-01
-7.43339717e-01 -1.27329421e+00 -5.50423026e-01 4.58085537e-01
8.32985342e-03 -4.80667651e-01 -3.98546159e-01 3.57674897e-01] | [11.466998100280762, -0.1411079466342926] |
ffd4e97e-cecd-4757-a9ab-9e2f74ed7240 | correction-of-errors-in-preference-ratings | 2306.03866 | null | https://arxiv.org/abs/2306.03866v1 | https://arxiv.org/pdf/2306.03866v1.pdf | Correction of Errors in Preference Ratings from Automated Metrics for Text Generation | A major challenge in the field of Text Generation is evaluation: Human evaluations are cost-intensive, and automated metrics often display considerable disagreement with human judgments. In this paper, we propose a statistical model of Text Generation evaluation that accounts for the error-proneness of automated metrics when used to generate preference rankings between system outputs. We show that existing automated metrics are generally over-confident in assigning significant differences between systems in this setting. However, our model enables an efficient combination of human and automated ratings to remedy the error-proneness of the automated metrics. We show that using this combination, we only require about 50% of the human annotations typically used in evaluations to arrive at robust and statistically significant results while yielding the same evaluation outcome as the pure human evaluation in 95% of cases. We showcase the benefits of approach for three text generation tasks: dialogue systems, machine translation, and text summarization. | ['Mark Cieliebak', 'Don Tuggener', 'Pius von Däniken', 'Jan Deriu'] | 2023-06-06 | null | null | null | null | ['text-summarization'] | ['natural-language-processing'] | [ 4.00586545e-01 6.11111641e-01 1.31920844e-01 -4.56690311e-01
-1.19099057e+00 -9.35090780e-01 1.10136759e+00 4.75691408e-01
-4.38315630e-01 1.14110243e+00 4.53559756e-01 -3.15896869e-01
1.27480999e-01 -4.57986027e-01 -2.04094440e-01 -2.89012522e-01
4.25789773e-01 8.24190319e-01 9.31144208e-02 -3.76226693e-01
6.63523316e-01 -1.30654633e-01 -1.52675366e+00 4.20203060e-01
1.22979069e+00 7.51303732e-01 -9.35858116e-03 1.12174118e+00
-1.07115857e-01 8.51499856e-01 -1.11322796e+00 -7.35603988e-01
5.73082380e-02 -6.53640926e-01 -1.22426522e+00 -1.94758940e-02
3.43544066e-01 -2.05165610e-01 3.44824255e-01 1.00551915e+00
6.63319230e-01 5.58519997e-02 1.21014392e+00 -1.11477411e+00
-4.86508787e-01 9.10803974e-01 -6.73936382e-02 -1.41172469e-01
9.09710050e-01 -7.30835870e-02 1.24003756e+00 -8.69258523e-01
5.45420110e-01 1.25138104e+00 5.20396173e-01 3.96512628e-01
-1.15962207e+00 -1.81264669e-01 -2.29874372e-01 -1.96948886e-01
-1.00421834e+00 -6.48771584e-01 1.70686170e-01 -6.46547973e-01
9.23077404e-01 4.55069751e-01 1.94475800e-01 8.95312965e-01
4.10149768e-02 5.98806500e-01 1.14061177e+00 -6.56549096e-01
2.17890933e-01 6.02466345e-01 3.07709247e-01 3.78201246e-01
4.01566148e-01 -3.66573215e-01 -3.49380493e-01 -4.48947370e-01
3.09587002e-01 -6.29879057e-01 -3.44946504e-01 1.09174680e-02
-1.41705394e+00 8.70948195e-01 -1.16321184e-01 4.45214361e-01
-3.90321255e-01 -1.75143644e-01 4.39718306e-01 3.82476836e-01
6.15788221e-01 1.16057265e+00 -3.30408961e-01 -5.45466900e-01
-1.01374209e+00 6.33838177e-01 1.18698752e+00 1.00736535e+00
5.66280842e-01 -1.95783496e-01 -6.48845196e-01 1.03316140e+00
-4.59588654e-02 3.74730289e-01 8.59322250e-01 -1.08081925e+00
5.44347167e-01 5.58027387e-01 7.07818449e-01 -1.01039600e+00
-2.26085097e-01 -6.70358762e-02 -6.04495347e-01 1.05585657e-01
5.25312483e-01 -4.29210544e-01 -6.34849846e-01 1.63838685e+00
-5.82091212e-02 -8.15980852e-01 1.99833259e-01 7.15602815e-01
5.40236592e-01 5.93388200e-01 -6.93357587e-02 -5.23536623e-01
1.22738147e+00 -9.43566978e-01 -7.49181151e-01 1.02181979e-01
8.26172531e-01 -1.06441045e+00 1.29388630e+00 3.51823747e-01
-1.43340886e+00 -4.30432916e-01 -9.98070836e-01 2.03809477e-02
-1.97845623e-01 1.88259602e-01 3.29314381e-01 6.72224581e-01
-1.24783552e+00 8.58045638e-01 -3.32405627e-01 -3.96778405e-01
-2.99088150e-01 2.99149454e-01 -1.19913325e-01 3.44073147e-01
-1.15728855e+00 1.32233274e+00 2.30815277e-01 -1.29084364e-01
-2.13004544e-01 -3.11753333e-01 -4.88552719e-01 4.16243225e-02
1.89329490e-01 -6.63067520e-01 1.96925259e+00 -1.09199870e+00
-1.61669338e+00 6.73790991e-01 -1.54120326e-01 -2.81374931e-01
8.70362222e-01 -4.05768037e-01 -7.15340078e-02 -6.56031966e-02
2.29108542e-01 4.61103827e-01 2.87570566e-01 -1.17350948e+00
-6.85235858e-01 -4.01008263e-04 -3.94248143e-02 5.12818158e-01
-1.20341092e-01 9.34176594e-02 2.90255360e-02 -5.21949172e-01
-3.01606566e-01 -9.03766155e-01 -2.41576776e-01 -6.40592694e-01
-4.31544721e-01 -4.19660807e-01 1.71984941e-01 -5.62031925e-01
1.43701613e+00 -1.60927057e+00 1.74899548e-01 1.59790460e-02
1.12651668e-01 3.23274225e-01 -6.11888953e-02 6.97942197e-01
2.08469748e-01 5.16218901e-01 -2.46204212e-01 -4.04623836e-01
2.82233328e-01 -1.71072811e-01 -3.89827043e-01 -1.50594816e-01
2.77393371e-01 7.09992647e-01 -1.09379864e+00 -6.14568770e-01
-8.22408795e-02 9.85739157e-02 -3.94960761e-01 4.47669655e-01
-1.92893565e-01 1.93322189e-02 -3.42583597e-01 2.86992490e-01
8.95199627e-02 -2.00102746e-01 2.46464610e-01 -1.25489876e-01
-2.00609997e-01 6.48329318e-01 -1.05224192e+00 1.30634797e+00
-4.08400148e-01 6.97158039e-01 -2.31725797e-01 -5.14893949e-01
8.06205332e-01 5.58762789e-01 4.70984839e-02 -3.80714417e-01
2.26443231e-01 4.48383570e-01 1.37154713e-01 -2.57822305e-01
1.16004288e+00 1.67561527e-02 -1.68863773e-01 1.04014909e+00
7.71978199e-02 -6.12717330e-01 4.87653464e-01 3.64880532e-01
1.02611518e+00 -1.12758361e-01 5.55761755e-01 -3.42357308e-01
3.99829060e-01 1.60648927e-01 2.55113780e-01 1.01425791e+00
-1.08829573e-01 8.24491203e-01 6.07528448e-01 -1.77107304e-01
-1.22750998e+00 -6.84239626e-01 1.85673594e-01 1.11901891e+00
-1.55211046e-01 -6.65393770e-01 -1.17652678e+00 -8.14000905e-01
-4.03388172e-01 1.08845842e+00 -5.36301255e-01 -1.01565886e-02
-2.70055145e-01 -9.09646988e-01 7.21579134e-01 2.81282037e-01
1.07235759e-01 -1.11515892e+00 -6.48130953e-01 2.81384945e-01
-5.13040841e-01 -9.66348767e-01 -4.44485247e-01 -1.18493088e-01
-6.19154692e-01 -9.83857453e-01 -7.57513762e-01 -4.69093502e-01
6.33337736e-01 1.20165624e-01 1.44285798e+00 1.59385204e-01
2.53096223e-01 2.07972303e-01 -5.36596954e-01 -5.24649203e-01
-1.05059981e+00 3.24473470e-01 7.71626309e-02 -3.34105700e-01
1.00524947e-01 -1.84477195e-01 -4.92653072e-01 3.19915712e-01
-7.68325567e-01 6.22443296e-02 5.81314921e-01 9.42816079e-01
-5.83892874e-02 -3.56232405e-01 8.45544159e-01 -1.14606833e+00
1.40961218e+00 -1.94283456e-01 -2.31907651e-01 5.38457751e-01
-1.10955513e+00 3.98177058e-01 6.46592736e-01 -3.15202415e-01
-9.79583979e-01 -3.25139552e-01 2.67708212e-01 2.61658967e-01
-7.22855702e-02 5.80829144e-01 2.49317676e-01 2.71131426e-01
1.00197697e+00 -2.42894962e-01 5.70497802e-03 -1.59814745e-01
3.93809676e-01 9.37048674e-01 3.57669950e-01 -6.14907920e-01
5.33802688e-01 -2.71662623e-01 -4.16104704e-01 -4.68330890e-01
-7.93861151e-01 -2.17376769e-01 -5.03744125e-01 -1.82595491e-01
6.18056178e-01 -6.35185719e-01 -2.29999125e-01 2.04881772e-01
-1.35040367e+00 -2.50109881e-01 -3.40919673e-01 3.34733486e-01
-7.22776115e-01 6.40849531e-01 -4.92858976e-01 -1.07837498e+00
-7.30994403e-01 -1.17213726e+00 1.14118659e+00 1.84069559e-01
-1.08118141e+00 -1.03150892e+00 3.73518169e-01 3.58051330e-01
4.90262657e-01 1.50866657e-01 9.53113496e-01 -1.03552079e+00
1.58528700e-01 -4.18569922e-01 -2.50767499e-01 3.42707753e-01
2.11568713e-01 3.13928425e-01 -9.75392759e-01 1.98718347e-02
-3.23184639e-01 -4.96623486e-01 5.82958400e-01 1.23458371e-01
6.71817183e-01 -5.57861269e-01 2.14598142e-02 -3.34532797e-01
9.42721188e-01 -1.18887296e-03 5.22849977e-01 1.65802002e-01
4.22114521e-01 9.77220416e-01 8.74453723e-01 4.15313303e-01
4.38879400e-01 7.63345063e-01 -1.62134081e-01 1.21679986e-02
1.63294986e-01 -2.41871849e-01 4.70603198e-01 9.71247017e-01
-3.15389395e-01 -5.72443545e-01 -8.81923437e-01 5.95045626e-01
-2.07538176e+00 -8.52531374e-01 -2.25441396e-01 2.47346139e+00
1.11237097e+00 7.90141150e-02 2.49362051e-01 2.58072972e-01
9.06397402e-01 -1.73416719e-01 8.51241648e-02 -9.18852866e-01
-4.91711870e-02 1.11259334e-01 2.66076028e-01 7.77141631e-01
-7.78784275e-01 6.98131144e-01 7.86106110e+00 5.10618806e-01
-5.88084817e-01 -1.08064234e-01 5.94592154e-01 1.26162274e-02
-4.38001066e-01 4.31430563e-02 -3.84744823e-01 4.64168757e-01
1.22801900e+00 -5.98038197e-01 1.90858364e-01 5.71006358e-01
2.23807037e-01 -1.28069609e-01 -1.35722232e+00 7.23556161e-01
1.76809788e-01 -9.32381570e-01 2.06190661e-01 -3.32293995e-02
9.12389040e-01 -3.34820330e-01 -2.47603536e-01 3.22794855e-01
6.37814760e-01 -1.03303003e+00 6.89244747e-01 3.28775257e-01
6.73561931e-01 -6.48019373e-01 1.05030990e+00 3.98982316e-01
-4.74800795e-01 1.51640952e-01 -2.77624369e-01 -2.68856227e-01
2.53707707e-01 6.65187538e-01 -1.46064007e+00 4.51713353e-01
7.03024715e-02 6.06991164e-02 -6.44937277e-01 7.26762056e-01
-2.46401131e-01 4.57159430e-01 1.84078366e-02 -3.81329715e-01
1.16476752e-01 -1.46221250e-01 5.05584955e-01 1.46877897e+00
5.08482397e-01 -5.01363762e-02 8.36692229e-02 5.49321115e-01
-3.70351858e-02 4.81540561e-01 -6.62419736e-01 -1.70908287e-01
4.41431820e-01 1.30157840e+00 -5.54333687e-01 -6.53321445e-01
9.84452218e-02 9.17823136e-01 4.22579914e-01 1.77649304e-01
-5.59459746e-01 -4.98315632e-01 1.38156697e-01 -8.41543823e-02
-8.62108096e-02 -3.33368629e-02 -4.68285829e-01 -1.09175885e+00
7.14511648e-02 -1.08195901e+00 1.75889701e-01 -8.16570759e-01
-1.20582688e+00 8.89029086e-01 1.05721503e-01 -1.18682647e+00
-1.04350388e+00 -5.38569212e-01 -6.91238225e-01 9.12186682e-01
-9.69717145e-01 -5.90066612e-01 -2.70399898e-01 -3.36151593e-03
5.52568078e-01 -1.39520198e-01 1.04420626e+00 -6.47058189e-02
-3.02405983e-01 6.65538549e-01 -3.22512724e-02 -2.40192562e-01
1.11605835e+00 -1.83288407e+00 4.42070931e-01 6.46825075e-01
5.74145094e-03 5.96938252e-01 1.12718058e+00 -4.33126420e-01
-7.58849919e-01 -8.85679662e-01 1.46446550e+00 -6.82009876e-01
4.96669590e-01 -4.38242182e-02 -6.73537016e-01 1.65509731e-01
4.68190968e-01 -8.82234454e-01 7.96644449e-01 2.14382187e-01
-2.23952219e-01 4.06455547e-01 -1.02249396e+00 7.42279708e-01
6.23106658e-01 -4.11344975e-01 -8.97775412e-01 5.59735954e-01
5.25478065e-01 -2.96069086e-01 -8.39967668e-01 2.85365373e-01
6.39143705e-01 -8.11076462e-01 2.95453876e-01 -5.29516697e-01
6.65775537e-01 -2.17650220e-01 -1.29217088e-01 -1.65195239e+00
-1.82341710e-01 -8.37951601e-01 8.81099924e-02 1.36010301e+00
9.77379084e-01 -4.05197859e-01 3.36440176e-01 9.57934439e-01
-5.86530976e-02 -5.07718146e-01 -4.38304991e-01 -4.33212042e-01
6.60417080e-02 -1.11227162e-01 3.91745299e-01 9.02509928e-01
4.24101353e-01 7.57881224e-01 -2.29896963e-01 -3.51154298e-01
3.83641839e-01 -6.87120259e-02 9.31955338e-01 -1.35906994e+00
-2.30415717e-01 -6.85993075e-01 -2.77110219e-01 -6.16241395e-01
5.16707674e-02 -5.29362857e-01 4.52719390e-01 -1.63705969e+00
3.39708149e-01 -1.52820215e-01 -5.95407598e-02 2.48718813e-01
-5.26848853e-01 2.88480043e-01 1.48760572e-01 2.77166694e-01
-6.63458288e-01 2.93314010e-01 8.63541484e-01 7.18716234e-02
-3.08775514e-01 8.32170472e-02 -1.08094919e+00 5.83187819e-01
8.14428449e-01 -2.59636402e-01 -4.14555490e-01 -4.60906476e-01
4.67826843e-01 -6.76880330e-02 5.37024587e-02 -6.78578317e-01
1.94153175e-01 -1.23202279e-01 8.70707035e-02 -2.61296600e-01
-6.80788904e-02 -2.30554312e-01 -5.64899929e-02 5.17450124e-02
-7.64726579e-01 4.79066789e-01 -2.07693323e-01 1.93155557e-01
-3.81403565e-01 -5.69927871e-01 4.44909304e-01 -5.92157133e-02
-1.26922783e-02 -3.09908450e-01 -4.68156934e-01 2.59742767e-01
5.67584097e-01 -2.96885055e-02 -4.43405479e-01 -7.89589822e-01
-3.46518874e-01 2.09818751e-01 5.75520456e-01 3.45584422e-01
1.85822248e-01 -1.05654120e+00 -1.00297546e+00 -3.80460382e-01
7.82180801e-02 -1.48530915e-01 -2.29268461e-01 7.67676055e-01
-4.58483726e-01 5.22660017e-01 -2.88812667e-02 -4.47508335e-01
-1.18018663e+00 3.55310813e-02 1.35653108e-01 -6.71426356e-01
-1.02002984e-02 3.22729439e-01 7.80598074e-03 -3.60289395e-01
9.10164118e-02 -3.71379316e-01 -2.48981580e-01 2.22154558e-01
5.64712167e-01 5.44915318e-01 3.37138712e-01 -4.51400578e-01
2.87363282e-03 2.54973710e-01 -2.83033103e-01 -8.18561256e-01
8.39806259e-01 -1.81450471e-01 4.06480134e-02 7.54598737e-01
7.06658006e-01 8.76107812e-02 -7.22051084e-01 1.22266024e-01
1.32442474e-01 -1.41175568e-01 -1.83119506e-01 -1.09867370e+00
-1.69365048e-01 7.25667238e-01 9.21459422e-02 6.69423044e-01
8.29718709e-01 -3.77712786e-01 4.63035256e-01 7.52262771e-01
2.92845428e-01 -1.36695898e+00 -2.26885244e-01 5.73297441e-01
1.05761135e+00 -1.27150011e+00 1.21588401e-01 -3.07774276e-01
-1.10275924e+00 1.00168800e+00 4.31637973e-01 1.20815873e-01
-6.85229003e-02 -1.30465820e-01 2.39606455e-01 3.23059000e-02
-1.12952816e+00 1.33099750e-01 3.84238303e-01 5.36868334e-01
1.05257940e+00 1.89669132e-01 -8.23551059e-01 5.11649191e-01
-7.23468900e-01 -2.41753876e-01 8.94107461e-01 8.08951080e-01
-5.01782656e-01 -1.17566204e+00 -1.13374479e-01 4.30274963e-01
-6.13083243e-01 -1.52383238e-01 -1.11679351e+00 5.55533171e-01
-5.72681546e-01 1.35150766e+00 -2.17036843e-01 -4.53224480e-01
4.70684171e-01 4.22508419e-01 4.04200792e-01 -7.77006209e-01
-9.68711078e-01 -1.50665596e-01 6.61264956e-01 -2.03810394e-01
-3.91303331e-01 -6.02175772e-01 -9.12926972e-01 -2.31566861e-01
-6.46443009e-01 7.15625882e-01 6.31162226e-01 9.67029810e-01
4.06932831e-01 1.97922543e-01 8.29890132e-01 -8.40242267e-01
-1.12145936e+00 -1.51740670e+00 -2.15108082e-01 7.81168163e-01
1.40702516e-01 -6.03311837e-01 -6.42091036e-01 2.37453189e-02] | [11.810128211975098, 8.964162826538086] |
05f80221-150e-439f-8874-47da56f57bd4 | towards-evaluating-gaussian-blurring-in | 2002.00140 | null | https://arxiv.org/abs/2002.00140v2 | https://arxiv.org/pdf/2002.00140v2.pdf | Towards Evaluating Gaussian Blurring in Perceptual Hashing as a Facial Image Filter | With the growth in social media, there is a huge amount of images of faces available on the internet. Often, people use other people's pictures on their own profile. Perceptual hashing is often used to detect whether two images are identical. Therefore, it can be used to detect whether people are misusing others' pictures. In perceptual hashing, a hash is calculated for a given image, and a new test image is mapped to one of the existing hashes if duplicate features are present. Therefore, it can be used as an image filter to flag banned image content or adversarial attacks --which are modifications that are made on purpose to deceive the filter-- even though the content might be changed to deceive the filters. For this reason, it is critical for perceptual hashing to be robust enough to take transformations such as resizing, cropping, and slight pixel modifications into account. In this paper, we would like to propose to experiment with effect of gaussian blurring in perceptual hashing for detecting misuse of personal images specifically for face images. We hypothesize that use of gaussian blurring on the image before calculating its hash will increase the accuracy of our filter that detects adversarial attacks which consist of image cropping, adding text annotation, and image rotation. | ['Yigit Alparslan', 'Mannika Kshettry', 'Ken Alparslan', 'Louis Kratz'] | 2020-02-01 | null | null | null | null | ['image-cropping', 'text-annotation'] | ['computer-vision', 'natural-language-processing'] | [ 3.55337530e-01 -1.62956506e-01 3.02783459e-01 -4.59082007e-01
-1.34722441e-01 -8.59174550e-01 4.35739756e-01 3.42972517e-01
-4.76420879e-01 3.87171298e-01 -4.52371947e-02 -1.81258291e-01
5.08151650e-01 -9.40184116e-01 -6.11793280e-01 -6.75022960e-01
3.99617366e-02 -1.48674399e-01 6.43231392e-01 -3.19562219e-02
5.05922496e-01 7.84143925e-01 -1.53118753e+00 3.75215262e-01
2.20153838e-01 8.26125205e-01 -9.83729735e-02 7.45761991e-01
1.04713403e-01 4.97473508e-01 -8.44835639e-01 -8.98102701e-01
5.55311620e-01 -1.72386363e-01 -5.62893391e-01 1.45665631e-01
4.11452770e-01 -7.46222854e-01 -5.57290137e-01 1.66683280e+00
4.55376238e-01 -1.95940241e-01 3.70935053e-01 -1.48693419e+00
-7.24904537e-01 3.51642191e-01 -6.38980210e-01 3.63929600e-01
5.96840918e-01 4.54936355e-01 4.22914803e-01 -8.41550231e-01
3.67194444e-01 1.41444957e+00 3.60628128e-01 3.99945557e-01
-1.03110456e+00 -9.35893059e-01 -5.75064719e-01 5.28319299e-01
-1.66416514e+00 -6.02096140e-01 5.81047773e-01 -1.48405939e-01
3.32734346e-01 4.63964731e-01 3.91703755e-01 7.00838685e-01
4.45097625e-01 8.07212815e-02 1.12519133e+00 -4.80790079e-01
1.61656022e-01 5.73815227e-01 -1.29879802e-01 6.78709626e-01
3.86328250e-01 1.83640420e-02 -2.76061296e-01 -5.10152996e-01
5.10281503e-01 1.01073049e-01 -2.63756901e-01 5.87681644e-02
-7.62302876e-01 7.96383917e-01 4.17043298e-01 3.28935646e-02
-3.06407660e-01 -9.07086805e-02 3.49212408e-01 3.51869255e-01
-3.77898291e-02 3.95783931e-01 1.60603330e-01 2.31718197e-01
-8.22430730e-01 1.22143045e-01 6.60954595e-01 5.36344171e-01
1.01504385e+00 -3.79898310e-01 -8.45011249e-02 7.66536951e-01
3.21815163e-01 6.27134025e-01 6.73123062e-01 -6.79145694e-01
8.36314075e-03 5.69319546e-01 -2.91916616e-02 -1.53919101e+00
1.15045279e-01 2.72535801e-01 -6.48319900e-01 2.86297262e-01
3.31205487e-01 2.07430273e-01 -8.40362549e-01 1.36310315e+00
4.04186785e-01 2.73508400e-01 -5.76555803e-02 8.30733359e-01
5.20446062e-01 5.61167300e-01 4.70877849e-02 -6.92872554e-02
1.82475722e+00 -2.93882310e-01 -6.28499866e-01 -1.10309944e-01
1.61796153e-01 -1.26161516e+00 8.60356033e-01 2.88546979e-01
-1.03335476e+00 -7.53058076e-01 -1.14268863e+00 1.02860004e-01
-7.74117529e-01 -3.02009791e-01 5.84814698e-02 1.26161671e+00
-9.37504411e-01 4.87731516e-01 -1.50418714e-01 -5.11238515e-01
2.06295654e-01 4.97593969e-01 -6.33148193e-01 -3.17925125e-01
-1.45155740e+00 8.03323805e-01 3.07738215e-01 4.27372493e-02
-7.70270944e-01 -5.25460206e-02 -7.08822906e-01 9.96987894e-02
6.31125495e-02 -7.10811093e-02 8.86364758e-01 -1.00444424e+00
-7.55304396e-01 1.04158080e+00 -3.96487005e-02 -5.21982729e-01
4.07656252e-01 2.11345568e-01 -9.05153871e-01 4.49449301e-01
1.16241649e-01 8.60239804e-01 1.70624375e+00 -9.67580974e-01
-5.04155397e-01 -5.66214502e-01 1.00368097e-01 -8.29532966e-02
-5.94897449e-01 2.82238632e-01 -4.40308452e-01 -6.30434632e-01
-1.42389685e-01 -1.05656838e+00 2.89963067e-01 -4.53418866e-02
-4.11114961e-01 2.16496140e-01 1.44226623e+00 -8.77772689e-01
1.26945853e+00 -2.60926199e+00 -5.24558127e-01 5.48524737e-01
-3.59697081e-02 5.71429670e-01 -8.17135572e-02 3.86206895e-01
-5.77196814e-02 3.44616771e-01 -8.10258761e-02 5.50523810e-02
-3.36749077e-01 1.66695580e-01 -3.73481899e-01 7.72226334e-01
2.37790316e-01 4.21989381e-01 -5.43185592e-01 -5.56112945e-01
1.39423624e-01 4.78686333e-01 -3.64817351e-01 2.04606533e-01
2.35910118e-01 -3.43315932e-03 6.46078810e-02 5.89676023e-01
1.21760857e+00 2.73367614e-01 -2.57327147e-02 -3.64419013e-01
9.62780714e-02 -4.70584631e-02 -1.41230416e+00 4.97484565e-01
8.07636082e-02 6.35826468e-01 -1.11151598e-02 -5.68240702e-01
8.14676464e-01 3.05638939e-01 -1.08062029e-01 -4.50476408e-01
1.19706377e-01 -2.84424163e-02 7.38378018e-02 -6.22580588e-01
5.91039896e-01 9.83984396e-02 -1.23755066e-02 4.84476298e-01
-3.48448813e-01 -4.94847447e-02 3.92764509e-02 1.69616118e-01
1.11398947e+00 -6.10372961e-01 1.97248757e-01 1.92163102e-02
1.02393627e+00 -5.36236823e-01 3.38143229e-01 5.86381972e-01
-3.87673885e-01 4.68270093e-01 2.09535539e-01 -4.32057470e-01
-1.19669223e+00 -9.31643963e-01 -7.08846897e-02 9.42112148e-01
2.96411723e-01 -2.77930766e-01 -9.12002623e-01 -7.79629171e-01
2.70374864e-01 3.16308886e-01 -4.56252366e-01 -4.45435822e-01
-3.76301050e-01 -5.81516862e-01 8.53902936e-01 7.52180219e-02
1.01562762e+00 -1.17193413e+00 -6.74489796e-01 6.19546026e-02
-1.43038062e-02 -9.30488467e-01 -7.65571952e-01 -4.23224449e-01
-3.65843832e-01 -1.25405610e+00 -6.31799340e-01 -7.52385139e-01
8.15143406e-01 7.76121199e-01 3.91931266e-01 4.44718182e-01
-4.92429525e-01 3.60972971e-01 -4.60875809e-01 -2.92542279e-01
-7.82165408e-01 -3.43621612e-01 7.83820823e-02 6.08817399e-01
5.06767035e-01 -1.44637048e-01 -7.56147146e-01 5.90532303e-01
-1.40623844e+00 -4.94869888e-01 4.63378400e-01 4.08444554e-01
-1.93042502e-01 7.85292566e-01 1.23385124e-01 -7.46475101e-01
7.66664743e-01 -3.30696017e-01 -2.21369222e-01 2.18513787e-01
-2.29620248e-01 -1.43675767e-02 5.95861197e-01 -6.00597799e-01
-6.65614188e-01 -9.70070288e-02 -1.71806701e-02 -4.26433057e-01
-3.85780543e-01 8.89045838e-03 -1.59475893e-01 -4.70468014e-01
5.77021956e-01 3.37122411e-01 1.80042908e-01 -1.25492170e-01
1.49954960e-01 1.09688556e+00 5.12060881e-01 1.36815105e-02
1.18460906e+00 5.02208471e-01 -1.03075139e-01 -1.12620890e+00
-1.08864002e-01 -4.84274358e-01 -3.87128830e-01 -4.97350097e-01
7.90681720e-01 -5.40074229e-01 -8.18812847e-01 7.97581315e-01
-1.15436482e+00 3.73990774e-01 4.04963225e-01 5.80993369e-02
2.22922415e-01 9.15887833e-01 -5.76400936e-01 -8.17081213e-01
-1.44953996e-01 -1.25152802e+00 6.48513019e-01 3.41933787e-01
-2.18930095e-01 -5.55537581e-01 -3.40018719e-01 4.04761523e-01
4.45081711e-01 -7.05615506e-02 7.74758875e-01 -5.67599297e-01
-2.88844377e-01 -4.62937117e-01 -3.29609603e-01 6.46110892e-01
3.61156076e-01 3.03591967e-01 -1.10960674e+00 -5.61031461e-01
1.34956777e-01 -3.92250046e-02 5.25976062e-01 -2.00738281e-01
1.08274865e+00 -7.60760725e-01 -1.83429971e-01 1.06893487e-01
1.20789051e+00 4.71326113e-01 1.21766984e+00 2.10889161e-01
3.97388488e-01 7.03281701e-01 4.46708769e-01 4.26505536e-01
5.86834699e-02 6.54514134e-01 5.01444936e-01 1.95379369e-02
2.32990846e-01 -1.67583734e-01 5.39099634e-01 9.13907513e-02
3.02270800e-01 -1.03628077e-01 -6.28496766e-01 2.48869494e-01
-1.20901334e+00 -1.22335613e+00 6.74046353e-02 2.61035514e+00
7.22912192e-01 2.19140500e-01 1.37647107e-01 4.13591802e-01
1.32231581e+00 1.09206803e-01 -4.11208689e-01 -6.07638776e-01
8.30579177e-02 8.59386548e-02 7.63118744e-01 4.01519954e-01
-1.23914456e+00 7.92847931e-01 5.76924562e+00 5.07770240e-01
-1.32143044e+00 -4.03503776e-02 7.81736910e-01 4.16270047e-01
1.40719980e-01 -4.02551405e-02 -8.19408059e-01 1.00399411e+00
8.00067902e-01 1.43252328e-01 6.20178401e-01 6.36279166e-01
3.15527022e-02 -3.52280527e-01 -6.06481552e-01 9.83398736e-01
4.59164381e-01 -6.93557978e-01 9.08021703e-02 2.34098881e-01
2.68122405e-01 -7.31833756e-01 3.22341383e-01 -3.00688837e-02
3.41863297e-02 -8.60089362e-01 5.71233690e-01 5.43087460e-02
4.47255582e-01 -8.48795712e-01 8.22203040e-01 3.62478755e-02
-1.05691803e+00 -1.35688514e-01 -7.46178627e-01 8.57423246e-02
-1.82999566e-01 3.18253398e-01 -1.31262207e+00 -1.75506532e-01
7.30541468e-01 4.04612608e-02 -1.02236450e+00 1.10450315e+00
-1.59977183e-01 3.49068969e-01 -2.76395053e-01 8.79980475e-02
2.39119120e-02 6.09039813e-02 4.47466254e-01 1.12747800e+00
3.91349882e-01 -8.52992758e-02 -9.88653451e-02 4.89370733e-01
-2.14461967e-01 1.54610306e-01 -8.86549175e-01 -4.80395481e-02
7.34791577e-01 1.24102139e+00 -7.61041045e-01 -2.85727650e-01
-3.46156657e-01 1.52662277e+00 -2.27294087e-01 9.27859098e-02
-9.82543886e-01 -4.80328977e-01 7.10274041e-01 3.65153342e-01
2.90421486e-01 -1.02333732e-01 2.80081421e-01 -5.92823267e-01
-1.47546619e-01 -1.05017531e+00 4.17290807e-01 -8.22991312e-01
-1.06724930e+00 3.44922483e-01 -2.45426059e-01 -1.04485500e+00
1.44337356e-01 -2.41457194e-01 -5.53832233e-01 6.64841831e-01
-1.17601609e+00 -7.39125192e-01 -2.88118362e-01 7.80013025e-01
2.38443747e-01 -1.81768402e-01 5.16098797e-01 2.60881305e-01
-3.09565187e-01 7.57876039e-01 -2.47814640e-01 4.96710449e-01
1.13549280e+00 -7.15494931e-01 3.48774701e-01 1.11310863e+00
7.57944286e-02 7.33939588e-01 8.34156692e-01 -8.80495727e-01
-1.26547229e+00 -9.32167470e-01 8.11554193e-01 -4.45361324e-02
3.07810247e-01 -2.84781963e-01 -1.12192309e+00 4.22302306e-01
2.43933663e-01 -1.34653285e-01 7.17212319e-01 -6.78586185e-01
-6.21661007e-01 -3.75109971e-01 -1.75981164e+00 5.29501140e-01
2.63423949e-01 -7.97876596e-01 -4.82647955e-01 1.28181562e-01
2.60176808e-01 1.15586519e-01 -6.06648147e-01 1.15803666e-01
6.84032798e-01 -1.18013740e+00 1.07836103e+00 -7.00251162e-02
2.74227541e-02 -5.82054853e-01 -1.00833833e-01 -9.76986408e-01
-3.07759672e-01 -4.84831870e-01 3.73812407e-01 1.39714301e+00
-1.40417982e-02 -7.84010172e-01 6.03676975e-01 7.25224972e-01
6.72755897e-01 3.11725199e-01 -7.51234114e-01 -5.65550089e-01
-5.66875517e-01 2.24704891e-02 7.74524212e-01 9.96848702e-01
1.83960088e-02 -4.14888263e-02 -5.58506727e-01 7.01875150e-01
6.69725597e-01 -6.57730162e-01 8.61021638e-01 -9.69226122e-01
-1.18592426e-01 -7.36516714e-02 -9.21883941e-01 -4.06971723e-01
-1.97939560e-01 -4.33052272e-01 -2.51511395e-01 -5.30880392e-01
1.01965703e-01 -1.78695753e-01 -7.86404759e-02 4.58399415e-01
-2.16187671e-01 9.68063295e-01 4.03484881e-01 3.85815322e-01
-1.27717093e-01 -2.16384195e-02 8.35958719e-01 -2.67345965e-01
6.47463351e-02 -1.74021833e-02 -5.06846488e-01 5.76208472e-01
9.09739971e-01 -4.96298403e-01 -1.25109211e-01 1.04647867e-01
9.44751501e-02 -1.95771664e-01 6.24751210e-01 -1.29618990e+00
3.02101940e-01 1.05059415e-01 6.31584525e-01 -3.10572654e-01
2.62092620e-01 -1.08800817e+00 2.67861277e-01 7.48111844e-01
-1.30615413e-01 4.42675799e-01 9.96744782e-02 4.25112426e-01
-1.86573893e-01 -6.95151806e-01 1.21060920e+00 -3.44055533e-01
-8.92024338e-01 4.83414829e-02 -5.45538425e-01 -5.38653433e-01
1.34642553e+00 -4.33662206e-01 -4.50514674e-01 -5.09261966e-01
-4.19112831e-01 -2.29675263e-01 9.03993070e-01 3.67553651e-01
8.68325889e-01 -1.10314417e+00 -4.58206534e-01 8.18192661e-01
1.84091944e-02 -7.27985680e-01 2.48244137e-01 2.74721146e-01
-8.33471537e-01 -2.00811643e-02 -5.39901912e-01 -2.65156746e-01
-1.96375227e+00 9.54771876e-01 -1.36409812e-02 3.42407763e-01
-1.54915661e-01 6.05328381e-01 -1.00977957e-01 1.39674500e-01
7.72088990e-02 6.17751367e-02 -1.40886620e-01 9.69628617e-02
1.03436875e+00 3.56860638e-01 3.45838293e-02 -9.58121240e-01
-3.90212893e-01 3.71382147e-01 -3.57419938e-01 -1.28168210e-01
6.86859846e-01 -2.04385251e-01 -5.05499363e-01 -2.41550207e-02
1.41645932e+00 2.83030450e-01 -8.05030465e-01 -7.90329129e-02
-1.19303182e-01 -9.96999204e-01 -9.98719335e-02 -4.24748152e-01
-9.97994244e-01 6.98596537e-01 9.32800293e-01 8.25560749e-01
1.12998998e+00 -1.99952453e-01 7.36018658e-01 8.39730278e-02
4.28744644e-01 -9.28597152e-01 1.55059516e-01 2.56412625e-02
6.50168180e-01 -1.22455239e+00 3.24150510e-02 -4.06287700e-01
-6.16855502e-01 1.17033887e+00 2.92454451e-01 -3.19141299e-02
6.47975326e-01 7.37381130e-02 -5.52880540e-02 2.74109282e-03
-1.75423995e-01 5.99942915e-02 3.14884447e-02 5.65826416e-01
-6.09079599e-02 -7.01267496e-02 -1.02247111e-01 -3.56801063e-01
-2.39515886e-01 -1.84721291e-01 6.87102497e-01 9.75666463e-01
-8.06697369e-01 -1.11901140e+00 -1.02232599e+00 3.26150328e-01
-6.58767164e-01 2.05482673e-02 -4.98233378e-01 3.75883162e-01
3.76122892e-01 1.10669339e+00 1.48967847e-01 -5.28255224e-01
9.30867195e-02 2.38931432e-01 2.36229986e-01 -4.27619249e-01
-5.70687234e-01 -1.87319562e-01 -3.06723535e-01 -3.66539687e-01
-3.12251300e-01 -6.68384910e-01 -8.69491756e-01 -7.64163554e-01
-1.41352057e-01 -8.66857637e-03 7.57118464e-01 5.66807747e-01
1.49073198e-01 -2.19972491e-01 9.69819486e-01 -5.44512689e-01
-4.35161829e-01 -7.97911227e-01 -5.59565723e-01 8.61265302e-01
2.71000177e-01 -5.42833269e-01 -5.45423388e-01 2.77675778e-01] | [12.675508499145508, 1.0716829299926758] |
0f6e7690-24f6-4107-9d70-13ea9c69af8e | a-close-up-comparison-of-the | 1907.11505 | null | https://arxiv.org/abs/1907.11505v1 | https://arxiv.org/pdf/1907.11505v1.pdf | A close-up comparison of the misclassification error distance and the adjusted Rand index for external clustering evaluation | The misclassification error distance and the adjusted Rand index are two of the most commonly used criteria to evaluate the performance of clustering algorithms. This paper provides an in-depth comparison of the two criteria, aimed to better understand exactly what they measure, their properties and their differences. Starting from their population origins, the investigation includes many data analysis examples and the study of particular cases in great detail. An exhaustive simulation study allows inspecting the criteria distributions and reveals some previous misconceptions. | ['José E. Chacón'] | 2019-07-26 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [-4.78232503e-02 -2.30610490e-01 -3.52894247e-01 -5.19673705e-01
-1.70656800e-01 -4.87314016e-01 4.97509629e-01 3.91689211e-01
-4.43550646e-01 7.97758162e-01 -1.45752832e-01 -4.77694452e-01
-9.35072899e-01 -7.01594472e-01 8.93505216e-02 -9.49677110e-01
-3.49085093e-01 9.14561808e-01 6.59202933e-02 1.08038180e-01
7.04345226e-01 4.48323220e-01 -1.91192329e+00 -7.32859895e-02
1.24824393e+00 5.72538018e-01 3.70348655e-02 4.45683300e-01
4.37790230e-02 3.54724020e-01 -1.13949895e+00 -6.70748949e-01
-6.71190545e-02 -3.96861315e-01 -8.56411815e-01 2.75082439e-01
-3.43195349e-01 2.08774447e-01 1.17456026e-01 9.91487384e-01
5.06306171e-01 1.26455203e-01 1.18164217e+00 -1.56824172e+00
-6.24907851e-01 9.15684283e-01 -3.67566198e-01 4.13167328e-01
4.47844684e-01 -3.37794609e-02 5.66854119e-01 -3.28288823e-01
2.43228003e-01 1.14492917e+00 9.22764421e-01 1.53001681e-01
-1.18479908e+00 -4.37269002e-01 -3.08652490e-01 2.50347912e-01
-1.83787274e+00 1.10751912e-01 4.72428143e-01 -4.48993772e-01
5.79533219e-01 7.24322796e-01 5.28045356e-01 4.45845515e-01
7.48353973e-02 2.94248730e-01 1.40339220e+00 -6.82092547e-01
2.31446862e-01 6.08785570e-01 7.31739104e-01 2.68068224e-01
8.65829766e-01 1.80413932e-01 2.87863076e-01 -5.11263072e-01
4.37138855e-01 5.70412353e-02 -1.81784645e-01 -5.17978609e-01
-6.96114659e-01 9.68513191e-01 8.01940486e-02 8.65769684e-01
5.04834689e-02 -3.52955341e-01 2.29307801e-01 1.94086194e-01
2.42664337e-01 4.11782473e-01 -2.71311522e-01 -1.61117122e-01
-8.90393317e-01 1.27873793e-01 5.67570150e-01 5.89003325e-01
4.46261257e-01 -8.62415433e-02 2.46785820e-01 9.76422608e-01
2.01036260e-01 4.82616961e-01 5.09853065e-01 -1.04402781e+00
1.65427789e-01 6.12008631e-01 1.20650001e-01 -1.27440488e+00
-5.58980763e-01 -2.50645489e-01 -7.81068921e-01 6.26752600e-02
5.77299058e-01 -1.39187798e-01 -5.20662010e-01 1.39011025e+00
-3.98167260e-02 -4.19455707e-01 -1.84027687e-01 4.31693107e-01
4.97777432e-01 4.79493588e-01 6.17204839e-03 -6.60875618e-01
1.09032917e+00 -3.39969337e-01 -9.12901878e-01 3.33719641e-01
5.04376411e-01 -7.31805444e-01 8.08150649e-01 5.12595832e-01
-1.03235996e+00 -5.33395827e-01 -9.04389024e-01 6.00444019e-01
-7.16389120e-01 -1.80739146e-02 6.11559033e-01 1.38755953e+00
-7.94460833e-01 7.65253246e-01 -8.21417689e-01 -9.46218371e-01
-6.39829114e-02 3.62466156e-01 2.54656941e-01 2.42886059e-02
-8.53302300e-01 8.99272621e-01 3.42122525e-01 -2.12779105e-01
-1.51760980e-01 -3.95266324e-01 -4.63156223e-01 5.33847604e-03
2.53566980e-01 -2.61596948e-01 6.05871439e-01 -6.45270288e-01
-8.24132323e-01 8.23156178e-01 -9.09818150e-03 -1.75449044e-01
4.49620008e-01 4.63022858e-01 -8.94695580e-01 -4.77674901e-02
1.13055289e-01 -1.50558844e-01 1.50187939e-01 -1.57812870e+00
-6.55569017e-01 -4.85230029e-01 -3.81384104e-01 1.99141614e-02
-1.14878170e-01 3.27821463e-01 -2.79586643e-01 -1.07850266e+00
6.22951314e-02 -6.58119917e-01 -3.65387768e-01 -7.93697119e-01
-2.36488655e-01 -4.95852113e-01 3.47329974e-01 -2.10614219e-01
1.91444802e+00 -2.22835517e+00 -2.81199515e-01 9.47077394e-01
3.30681279e-02 1.44833792e-02 4.15077835e-01 8.64319384e-01
-1.98685467e-01 3.57418031e-01 -5.74416697e-01 1.35894090e-01
4.99849990e-02 3.26043606e-01 1.22841127e-01 7.46939301e-01
-2.75575459e-01 1.98727816e-01 -9.02721405e-01 -5.26464164e-01
4.06207651e-01 3.91923696e-01 -1.15551531e-01 -3.68134081e-02
7.77421892e-01 -1.74191028e-01 -2.46775359e-01 5.93711674e-01
7.55496800e-01 -8.93599242e-02 3.34733605e-01 7.65202865e-02
-1.94362566e-01 5.74058257e-02 -1.48703289e+00 5.83101451e-01
8.24208781e-02 5.46165049e-01 -2.44106084e-01 -1.11299956e+00
9.37313676e-01 8.74248222e-02 6.32626772e-01 -3.62043351e-01
3.18845451e-01 4.12808657e-01 2.61830479e-01 -4.79049772e-01
3.66034746e-01 -7.46302977e-02 -3.04200560e-01 7.36671209e-01
-3.07801038e-01 3.43892090e-02 4.74725217e-01 1.67570822e-02
7.87247777e-01 -6.64132237e-01 6.53369486e-01 -8.43767583e-01
5.68504751e-01 2.56801873e-01 5.18265605e-01 8.21553469e-01
-3.00375998e-01 5.90456247e-01 3.55215967e-01 -4.03267056e-01
-7.49706268e-01 -1.18145967e+00 -4.65605527e-01 5.15093982e-01
3.20621938e-01 -3.53125334e-01 -9.79821205e-01 -5.28977394e-01
2.59716272e-01 8.53251874e-01 -9.15735066e-01 -5.24787717e-02
-2.07577005e-01 -1.42955983e+00 6.22925699e-01 3.87144387e-01
3.66524786e-01 -6.18582726e-01 -7.38367021e-01 -2.78262496e-01
-1.71210796e-01 -4.48630750e-01 1.12047195e-01 1.62027463e-01
-9.80499327e-01 -1.51998091e+00 -5.24304152e-01 -5.93871653e-01
9.59881067e-01 3.92146915e-01 1.11949825e+00 5.81347585e-01
-4.15552497e-01 3.19277525e-01 -5.38992941e-01 -1.12531655e-01
-4.86188233e-01 -7.22747901e-03 3.51832330e-01 -4.22624946e-01
1.07204807e+00 -4.30109560e-01 -2.44246110e-01 8.95097911e-01
-6.72675371e-01 -1.13974679e+00 4.60318357e-01 6.57304704e-01
1.75115392e-01 8.65814209e-01 5.40956914e-01 -1.08593452e+00
9.84686375e-01 -4.62419301e-01 -3.86372715e-01 4.72608775e-01
-1.31241548e+00 -1.14006802e-01 2.15139076e-01 -2.12131694e-01
-1.00552785e+00 -5.02931237e-01 1.12768739e-01 2.57670939e-01
-4.29572761e-01 1.79188073e-01 -1.28752798e-01 1.35004565e-01
6.94147289e-01 -2.13733185e-02 -3.28627117e-02 -5.89687169e-01
-1.50638387e-01 9.54226613e-01 4.76064146e-01 -4.33290124e-01
6.10805273e-01 4.35501099e-01 -2.69017398e-01 -8.97052050e-01
-2.47414798e-01 -4.76279408e-01 -6.68252826e-01 -3.13805431e-01
4.81005847e-01 -2.11969808e-01 -8.07624340e-01 4.10235733e-01
-5.57226539e-01 -7.82733709e-02 -2.15635598e-01 4.87346411e-01
-4.64730233e-01 5.28646410e-01 -3.71209681e-01 -1.02507615e+00
2.60041386e-01 -1.22467101e+00 1.89017847e-01 8.86940435e-02
-4.81873482e-01 -1.29103482e+00 8.34123045e-02 3.86012375e-01
2.64311820e-01 3.10186833e-01 9.54143941e-01 -8.71575952e-01
3.36178988e-01 -2.95001328e-01 -1.28826708e-01 2.46587358e-02
3.74513626e-01 5.21087170e-01 -6.31668925e-01 -4.11823124e-01
1.82040140e-01 3.86444956e-01 5.48159957e-01 5.08216083e-01
1.17090118e+00 -1.63169116e-01 -6.82619572e-01 3.47781897e-01
1.83836770e+00 7.75825500e-01 7.06571281e-01 4.65108663e-01
-1.93830058e-02 9.19406116e-01 7.32262492e-01 5.59282243e-01
2.55470872e-01 5.02605677e-01 1.32637352e-01 1.24242336e-01
1.01080164e-01 2.34986439e-01 -9.61607844e-02 8.48109901e-01
-4.32422340e-01 -3.38695288e-01 -1.15218592e+00 5.45208395e-01
-1.49286389e+00 -1.22350931e+00 -6.41354084e-01 2.55387306e+00
3.40107828e-01 1.10529199e-01 6.12786651e-01 8.44018221e-01
1.18788314e+00 -2.18624756e-01 5.57956696e-02 -6.96260154e-01
-1.83438629e-01 -2.42457181e-01 6.33465052e-01 5.18606067e-01
-7.66151905e-01 1.89318523e-01 8.71606159e+00 1.06412494e+00
-4.66962665e-01 -1.99287146e-01 7.32920766e-01 1.97332561e-01
6.44734828e-03 -1.22065522e-01 -4.67602700e-01 8.07630241e-01
8.89687836e-01 -2.80042052e-01 1.37707591e-01 3.97242397e-01
2.25059003e-01 -7.05204308e-01 -6.48897290e-01 8.26714814e-01
1.08667783e-01 -8.37622046e-01 2.84339450e-02 1.70823663e-01
5.31851649e-01 -5.49229085e-01 7.23400190e-02 -7.43861571e-02
4.86851841e-01 -9.62822735e-01 3.38695437e-01 3.87761414e-01
4.66315061e-01 -1.18896425e+00 1.17188680e+00 -3.63721810e-02
-7.39261448e-01 -4.30786759e-01 -3.48073304e-01 -3.12038571e-01
-2.23553181e-03 8.36456895e-01 -3.35688382e-01 6.82055235e-01
1.15074039e+00 1.93067700e-01 -7.72417188e-01 1.24118030e+00
1.39720142e-01 7.47925639e-01 -4.42254037e-01 -3.71315628e-01
2.67580692e-02 -5.18659234e-01 3.18111211e-01 1.24523938e+00
2.72586286e-01 2.24582791e-01 -5.19566655e-01 7.51145899e-01
6.45418167e-01 7.58466199e-02 -5.60761392e-01 2.73946226e-01
9.06462133e-01 8.38180065e-01 -1.11775577e+00 -3.63893539e-01
-3.47342342e-01 4.47313309e-01 -1.32921398e-01 1.20774329e-01
-8.36181819e-01 -9.25467432e-01 5.05870223e-01 1.61815539e-01
3.54207051e-03 -8.92280266e-02 -7.03121722e-01 -5.77124953e-01
-3.52450401e-01 -7.66545951e-01 7.62490273e-01 -3.79846066e-01
-1.42865646e+00 4.23957437e-01 5.24699211e-01 -1.13229775e+00
-1.71771303e-01 -4.98677790e-01 -7.60784626e-01 5.46384931e-01
-5.60089290e-01 -2.40042359e-01 -3.42277110e-01 4.47747886e-01
1.43021718e-01 -3.29709411e-01 4.55889881e-01 3.81906867e-01
-9.08414245e-01 8.07221949e-01 7.76187479e-01 1.71832412e-01
4.95054752e-01 -1.22302902e+00 -2.65632123e-01 5.37870347e-01
-4.45402443e-01 8.06632161e-01 1.17796135e+00 -4.75428671e-01
-5.27444422e-01 -5.12730658e-01 7.92412758e-01 -6.12008989e-01
2.52864331e-01 4.67127264e-02 -7.45256543e-01 2.72879392e-01
3.51857722e-01 -1.03549945e+00 1.01112354e+00 1.42403424e-01
1.67651251e-01 -1.02425583e-01 -1.48617637e+00 4.66363013e-01
7.68440247e-01 2.48570740e-01 -8.04230809e-01 5.24493828e-02
9.64867696e-03 4.28272396e-01 -1.16641605e+00 4.37590837e-01
4.58756030e-01 -1.68946552e+00 9.66419280e-01 -1.73421100e-01
-1.48248598e-01 -8.37004110e-02 -1.02453507e-01 -1.14595449e+00
-4.47947562e-01 -3.52623999e-01 6.57797933e-01 1.60319865e+00
5.03083467e-01 -9.17565167e-01 6.61584735e-01 5.94344556e-01
3.22228998e-01 -6.10196114e-01 -6.90214515e-01 -1.04416025e+00
1.17573570e-02 -2.43203729e-01 1.09460890e+00 1.32525039e+00
5.12369990e-01 1.52193652e-02 1.12307459e-01 -5.65367229e-02
1.04427338e+00 1.86578259e-01 4.00353968e-01 -1.45956481e+00
1.77422151e-01 -9.06398535e-01 -5.70671260e-01 -3.17891300e-01
-2.19206944e-01 -4.84741420e-01 -2.83596188e-01 -1.33238256e+00
3.26382577e-01 -7.33796418e-01 -1.93701297e-01 -6.04265071e-02
-2.07536459e-01 2.21843019e-01 -1.66879728e-01 3.02515894e-01
-2.56811261e-01 3.43614295e-02 4.14988369e-01 1.04079731e-01
-1.79611981e-01 4.36147034e-01 -6.79365218e-01 8.42400074e-01
1.18067110e+00 -5.87450743e-01 -4.17852998e-01 -1.13457464e-01
1.73905507e-01 -3.09869081e-01 -6.67756125e-02 -1.21514416e+00
1.64959103e-01 -4.25648272e-01 2.51849771e-01 -8.71050596e-01
-5.47634438e-02 -1.13477015e+00 4.69098240e-01 7.79977322e-01
-3.06633890e-01 7.34433830e-01 -1.63225695e-01 4.59293514e-01
-2.47142836e-01 -6.52997017e-01 8.28748763e-01 -1.68631338e-02
-4.45715845e-01 -3.23314250e-01 -7.59675920e-01 -8.25504214e-02
1.30149186e+00 -6.10535979e-01 -2.32856676e-01 -2.39018574e-01
-7.69313574e-01 4.06474590e-01 8.51765215e-01 2.88643152e-03
3.88489246e-01 -1.29589760e+00 -4.52957958e-01 1.22141607e-01
1.38351852e-02 -5.45882702e-01 4.54044044e-02 1.02093828e+00
-8.72411788e-01 5.13361156e-01 -3.37936103e-01 -4.75872666e-01
-1.39164221e+00 1.04387629e+00 4.55687076e-01 -1.33211911e-01
2.28659734e-02 3.41900289e-01 -2.38068119e-01 -4.32388425e-01
3.11556160e-01 -5.79944476e-02 -3.25527787e-01 1.47888288e-01
4.04035687e-01 1.39082038e+00 -1.19415402e-01 -6.35183990e-01
-6.29509866e-01 8.36407721e-01 4.97041583e-01 -3.31764184e-02
1.05448449e+00 -5.71073949e-01 -5.62435925e-01 5.70123911e-01
9.27653551e-01 -3.66830565e-02 -3.60285789e-01 2.08991483e-01
3.07273775e-01 -7.24536896e-01 -3.20590585e-01 -7.56307304e-01
-9.22279775e-01 6.71385467e-01 6.80024266e-01 9.93171871e-01
1.28261220e+00 1.35579305e-02 -1.75073557e-03 1.66699678e-01
3.58160913e-01 -1.49035525e+00 -3.13349456e-01 1.97987468e-03
4.10591125e-01 -8.63296688e-01 3.80695373e-01 -8.28538835e-01
-5.33034980e-01 9.84800458e-01 3.09104949e-01 -2.24805996e-02
8.84465456e-01 1.94022104e-01 1.90793741e-02 -1.98093548e-01
-3.87649149e-01 -8.57034996e-02 -5.63431084e-02 1.00665629e+00
5.38640380e-01 2.20390916e-01 -1.19011426e+00 6.87757552e-01
-3.00022781e-01 -2.90269315e-01 7.39676356e-01 7.41975665e-01
-4.61708456e-01 -9.83184040e-01 -9.00729299e-01 6.65698826e-01
-6.20043814e-01 3.87253582e-01 -6.70207918e-01 1.11643422e+00
2.54566103e-01 1.43134665e+00 3.49278986e-01 -5.82710683e-01
3.02832454e-01 -3.21936049e-02 2.99663246e-01 -9.43658799e-02
-7.42391229e-01 -1.95738494e-01 1.03455633e-01 -2.58239985e-01
-8.70810688e-01 -1.05324948e+00 -9.96464014e-01 -9.64447796e-01
-6.41545773e-01 9.88876820e-01 3.95591497e-01 5.49278855e-01
4.04381566e-02 2.59706378e-01 8.56907129e-01 -3.58295560e-01
-3.21019471e-01 -8.81704986e-01 -1.09720099e+00 5.20114422e-01
-6.15927055e-02 -7.99319208e-01 -9.08303082e-01 -2.87349254e-01] | [7.6774516105651855, 4.494951248168945] |
2f49128b-2e3c-4874-937d-d003183957eb | deep-hierarchical-planning-from-pixels | 2206.04114 | null | https://arxiv.org/abs/2206.04114v1 | https://arxiv.org/pdf/2206.04114v1.pdf | Deep Hierarchical Planning from Pixels | Intelligent agents need to select long sequences of actions to solve complex tasks. While humans easily break down tasks into subgoals and reach them through millions of muscle commands, current artificial intelligence is limited to tasks with horizons of a few hundred decisions, despite large compute budgets. Research on hierarchical reinforcement learning aims to overcome this limitation but has proven to be challenging, current methods rely on manually specified goal spaces or subtasks, and no general solution exists. We introduce Director, a practical method for learning hierarchical behaviors directly from pixels by planning inside the latent space of a learned world model. The high-level policy maximizes task and exploration rewards by selecting latent goals and the low-level policy learns to achieve the goals. Despite operating in latent space, the decisions are interpretable because the world model can decode goals into images for visualization. Director outperforms exploration methods on tasks with sparse rewards, including 3D maze traversal with a quadruped robot from an egocentric camera and proprioception, without access to the global position or top-down view that was used by prior work. Director also learns successful behaviors across a wide range of environments, including visual control, Atari games, and DMLab levels. | ['Pieter Abbeel', 'Ian Fischer', 'Kuang-Huei Lee', 'Danijar Hafner'] | 2022-06-08 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 5.07107470e-03 3.50325972e-01 -1.83415160e-01 -4.93297242e-02
-4.36004400e-01 -6.55689299e-01 5.32699883e-01 -2.25265175e-01
-4.53648031e-01 1.09385049e+00 3.07504207e-01 -1.63850754e-01
-3.09203237e-01 -6.68793082e-01 -5.48431158e-01 -9.27071571e-01
-4.97638017e-01 7.57844925e-01 1.39157414e-01 -2.59568781e-01
5.45849621e-01 2.99024940e-01 -1.54235971e+00 5.89959286e-02
7.75053203e-01 5.45520306e-01 6.90500617e-01 9.63838577e-01
1.64940849e-01 1.19144404e+00 -3.84952813e-01 5.76518834e-01
5.78742981e-01 -4.67892677e-01 -9.54017043e-01 3.52003604e-01
1.02333188e-01 -5.34091473e-01 -2.89775819e-01 9.42742467e-01
2.82020390e-01 5.51375449e-01 5.12015700e-01 -1.25116813e+00
-4.87044930e-01 4.29897040e-01 -3.56969208e-01 2.74562743e-02
4.22425598e-01 8.12065423e-01 1.06481516e+00 -1.49684101e-01
7.72177935e-01 1.31787777e+00 4.68529798e-02 6.59350872e-01
-1.45698607e+00 -3.67798775e-01 4.55027401e-01 1.36828020e-01
-7.62829185e-01 -1.37043998e-01 3.90985817e-01 -5.52520037e-01
1.50822413e+00 -7.04998374e-02 9.69603360e-01 1.14240062e+00
5.33345044e-01 7.41505623e-01 1.50513387e+00 -7.99529180e-02
7.25291669e-01 -4.24407601e-01 -2.39922225e-01 9.35846329e-01
-7.99827129e-02 5.89928031e-01 -8.08627784e-01 -5.27565666e-02
1.41300440e+00 -1.73922271e-01 -2.58679450e-01 -1.15484154e+00
-1.59969938e+00 8.21149945e-01 5.48248410e-01 -1.25739696e-02
-6.92636132e-01 4.37881440e-01 -1.26444800e-02 6.32714510e-01
-2.46549025e-01 1.32642400e+00 -6.04847491e-01 -4.08645749e-01
-4.22806919e-01 5.58505297e-01 8.63243699e-01 9.85215366e-01
7.71493375e-01 4.28549379e-01 7.71379797e-04 2.73579121e-01
7.37336352e-02 3.37675869e-01 5.12745917e-01 -1.81482852e+00
4.33444291e-01 3.85913938e-01 5.21293759e-01 -7.91995406e-01
-8.36978555e-01 -3.55367064e-01 -3.63224000e-01 1.20991325e+00
6.28997147e-01 -3.20884705e-01 -1.09110546e+00 1.74812627e+00
2.77115166e-01 -3.67186993e-01 2.69637108e-01 1.21749210e+00
1.38354868e-01 6.87873840e-01 8.52867439e-02 -1.00687586e-01
8.87484372e-01 -1.31405640e+00 -4.35620725e-01 -6.75461769e-01
7.32272744e-01 3.83518003e-02 1.34679997e+00 7.67153263e-01
-1.25113046e+00 -3.76691788e-01 -8.92723143e-01 -7.00972155e-02
-1.08636603e-01 -3.40198219e-01 9.13599968e-01 8.95659178e-02
-1.32561612e+00 5.92195332e-01 -1.20155549e+00 -3.57190818e-01
1.30724907e-01 4.56239313e-01 -3.74510318e-01 1.58021897e-01
-8.01640511e-01 1.26085055e+00 5.95550656e-01 -3.38996142e-01
-1.79183447e+00 -1.86522320e-01 -9.14113045e-01 1.32875264e-01
7.18890011e-01 -8.47638667e-01 1.43657911e+00 -7.52326131e-01
-1.82930183e+00 5.95892370e-01 2.20559940e-01 -5.15367270e-01
3.50308269e-01 -3.21089029e-01 5.58405578e-01 1.79043040e-01
3.05190921e-01 1.23777604e+00 8.14946532e-01 -1.00649905e+00
-7.03809917e-01 -4.71675396e-01 4.63158548e-01 1.13651764e+00
-5.92988543e-02 -5.43383956e-01 -1.36523947e-01 -1.97111100e-01
3.29959571e-01 -1.05168056e+00 -8.28161538e-01 -1.19071752e-01
-1.93115041e-01 -2.15356424e-01 3.97872478e-01 -4.38241929e-01
4.78321880e-01 -1.77370942e+00 9.12308872e-01 -1.38593718e-01
2.83644468e-01 -6.38714314e-01 -4.11099615e-03 2.47421548e-01
2.23176911e-01 -2.04005927e-01 1.00674041e-01 -1.27622765e-02
9.51019898e-02 4.26067114e-01 -2.76813984e-01 3.37315738e-01
-3.04337829e-01 8.41595173e-01 -1.29141366e+00 -2.86635041e-01
3.08294207e-01 -1.28806889e-01 -6.77238941e-01 3.23201597e-01
-7.52084374e-01 8.77417326e-01 -7.39029288e-01 4.58427370e-01
-1.18113637e-01 -4.34526622e-01 3.93270671e-01 6.26404345e-01
-2.82042921e-01 5.36771655e-01 -1.02055573e+00 2.03084898e+00
-1.97594330e-01 4.62336898e-01 2.62267560e-01 -8.65712166e-01
7.25571156e-01 2.44411901e-01 6.21565819e-01 -8.00535977e-01
-1.65253237e-01 -4.89158854e-02 6.09096736e-02 -5.00207663e-01
4.16815937e-01 -3.99122713e-03 -2.25150138e-01 4.76733178e-01
9.80450734e-02 -6.39773428e-01 2.90438235e-01 5.12233004e-02
1.47090387e+00 8.52646053e-01 4.12043333e-01 -1.57495946e-01
-2.79517829e-01 7.49290943e-01 3.74180436e-01 1.17127573e+00
-2.80645728e-01 3.41434121e-01 7.05749989e-01 -6.46413863e-01
-1.14377034e+00 -1.34465754e+00 4.22930300e-01 1.50440121e+00
3.03968161e-01 -1.39455229e-01 -5.83481610e-01 -4.41430598e-01
-1.97155833e-01 1.00065243e+00 -6.39720142e-01 1.00731194e-01
-6.49151921e-01 -3.86124462e-01 -5.93730137e-02 4.98029828e-01
4.73319888e-01 -1.66160715e+00 -1.52267146e+00 2.45769501e-01
2.05190340e-03 -7.43793607e-01 -1.56769216e-01 7.55952477e-01
-1.22764993e+00 -1.12856400e+00 -6.37703896e-01 -7.43026078e-01
6.93564832e-01 1.65222898e-01 1.24256337e+00 -3.44662696e-01
-3.36006492e-01 6.05108976e-01 -1.53324336e-01 -1.00363724e-01
6.08498305e-02 -5.79444356e-02 2.09566265e-01 -9.47157025e-01
1.58301950e-01 -7.29682863e-01 -4.95416075e-01 1.80222452e-01
-2.10524797e-01 7.11134672e-01 7.67033696e-01 1.02918839e+00
4.89199281e-01 9.07932222e-02 1.34237662e-01 -3.41472447e-01
6.38183594e-01 -4.07102197e-01 -9.50470626e-01 -1.56776935e-01
-4.42558557e-01 4.26070243e-01 4.43964094e-01 -4.33376402e-01
-1.07457018e+00 2.70685524e-01 4.19875294e-01 -2.97045082e-01
-4.29185003e-01 2.93545812e-01 1.65647104e-01 3.48471165e-01
9.46113110e-01 3.45455408e-01 -7.55026042e-02 -1.24353608e-02
5.69563150e-01 -9.63090062e-02 3.98229241e-01 -7.62024820e-01
3.79442394e-01 3.68726462e-01 3.94509081e-03 -7.46951699e-01
-7.22101390e-01 -1.16354913e-01 -6.53368056e-01 -2.28354365e-01
1.18465114e+00 -8.26214552e-01 -1.03178179e+00 1.21500209e-01
-7.54176021e-01 -1.13034189e+00 -5.42841315e-01 6.01682246e-01
-1.55979824e+00 -9.12850499e-02 -6.20799482e-01 -7.85360515e-01
2.11901203e-01 -1.29429531e+00 8.56886804e-01 3.10796529e-01
-4.46753561e-01 -7.91137457e-01 4.53489721e-02 8.26672018e-02
2.34559298e-01 3.22573364e-01 9.30786908e-01 1.86021887e-02
-1.07978928e+00 5.80686867e-01 1.86417028e-01 -4.51644927e-01
9.25727487e-02 -7.42370665e-01 -4.36162412e-01 -5.00746369e-01
6.52367845e-02 -1.07429481e+00 6.51779711e-01 6.37482524e-01
1.03819466e+00 -2.74917722e-01 -3.30684960e-01 5.65613031e-01
1.06300294e+00 4.36157316e-01 3.75760496e-01 7.29431927e-01
3.92419994e-01 8.48024189e-01 8.47022295e-01 4.60528791e-01
3.96386921e-01 5.27206182e-01 1.00756490e+00 2.47636974e-01
2.28103280e-01 -3.24243486e-01 5.48549235e-01 2.12086380e-01
-1.76828116e-01 8.24113712e-02 -9.64355409e-01 5.04665554e-01
-1.97824228e+00 -9.47232783e-01 4.06125695e-01 2.00416136e+00
8.79425228e-01 5.52715242e-01 2.05656327e-02 -4.47925597e-01
6.64418638e-02 2.06642091e-01 -1.19522560e+00 -2.74479955e-01
2.97282755e-01 -2.76843935e-01 4.74266320e-01 7.31784463e-01
-8.87893140e-01 1.37411547e+00 7.20430231e+00 2.85066664e-01
-8.89828265e-01 -1.43380299e-01 4.03077722e-01 -4.88664508e-01
-3.26267332e-02 9.67827290e-02 -7.44716465e-01 -9.70423371e-02
4.88483310e-01 4.65368852e-02 1.01214886e+00 1.21419728e+00
2.95486480e-01 -5.17465651e-01 -1.29767513e+00 9.24727917e-01
-3.13620806e-01 -1.05092812e+00 -4.19701338e-01 2.46798888e-01
8.70009482e-01 1.70797870e-01 2.93286830e-01 7.37150669e-01
1.36179900e+00 -1.51816607e+00 5.72527468e-01 4.86485124e-01
4.86533940e-01 -6.07837856e-01 -5.73746823e-02 9.56004620e-01
-6.87983990e-01 -4.65300292e-01 -5.40719807e-01 -6.93763614e-01
6.40760064e-02 -2.34226063e-01 -9.51846123e-01 -3.14185202e-01
8.61627281e-01 5.61086595e-01 -2.06741139e-01 7.13746786e-01
-5.93242049e-01 2.24971741e-01 -3.38157266e-01 -3.39643389e-01
7.27041960e-01 -3.28853637e-01 6.35849893e-01 4.88581568e-01
2.40847319e-01 3.56279045e-01 6.81280375e-01 8.90026033e-01
5.59102535e-01 -3.14304680e-01 -8.87106478e-01 -2.38982230e-01
1.00767002e-01 1.01647568e+00 -7.66718268e-01 -2.05363527e-01
1.68362409e-02 9.41162109e-01 8.27383757e-01 6.53046191e-01
-5.80351353e-01 2.04818603e-02 7.40154684e-01 -1.00545667e-01
9.88814011e-02 -8.71472001e-01 -5.78232944e-01 -1.03736222e+00
-3.45775217e-01 -1.02350724e+00 2.46078327e-01 -1.25600505e+00
-6.03339434e-01 3.22623551e-01 1.03218898e-01 -9.55872118e-01
-7.51979291e-01 -6.91044927e-01 -2.56813556e-01 8.30257058e-01
-1.06508756e+00 -7.24439025e-01 -4.08072323e-01 5.39064527e-01
9.88183260e-01 -2.90939659e-01 1.00802565e+00 -7.81854212e-01
4.52764668e-02 -3.30887586e-01 2.95857545e-02 -3.09091926e-01
4.15234715e-01 -1.82813478e+00 3.84726167e-01 3.69674355e-01
-6.52026460e-02 4.71778065e-01 9.90693927e-01 -8.43898475e-01
-1.34534860e+00 -4.15602922e-01 5.86490966e-02 -5.15436888e-01
5.15279233e-01 -2.97871888e-01 -4.89449769e-01 9.03241992e-01
4.11458641e-01 -4.26524580e-01 2.00238660e-01 3.35092783e-01
1.97318736e-02 4.36378092e-01 -8.16587687e-01 1.15902734e+00
1.21724570e+00 -4.10198160e-02 -8.32010806e-01 4.27103311e-01
6.74374640e-01 -8.09983850e-01 -4.62885559e-01 6.25861511e-02
5.68545580e-01 -1.03297663e+00 8.93145740e-01 -9.84399736e-01
4.79415387e-01 -3.44603747e-01 1.69119388e-02 -1.72420394e+00
-6.39833570e-01 -7.32509494e-01 -2.53623724e-01 6.54261336e-02
2.82162905e-01 -3.09279829e-01 1.16033399e+00 5.18548548e-01
-1.52777240e-01 -6.73259974e-01 -5.85239768e-01 -5.96445024e-01
5.49867749e-02 -1.09356709e-01 1.15754969e-01 7.13545442e-01
3.23688418e-01 3.39159667e-01 -5.86245000e-01 2.05189481e-01
8.58210206e-01 3.14596742e-01 9.65712488e-01 -8.52496505e-01
-6.09815955e-01 -4.49779809e-01 8.99213925e-02 -1.60767341e+00
7.53819793e-02 -5.58120012e-01 4.34930921e-01 -1.83582461e+00
-4.64560911e-02 -5.08024573e-01 -1.14810551e-02 7.49304235e-01
1.63388774e-01 -2.85354614e-01 2.69314766e-01 2.73640364e-01
-8.63571227e-01 6.00065112e-01 1.76181209e+00 -6.54319078e-02
-6.43193841e-01 -2.50536084e-01 -6.52391195e-01 9.22764480e-01
1.07464981e+00 -3.59758973e-01 -8.11266482e-01 -5.80844998e-01
4.40569639e-01 4.90883380e-01 2.97068506e-01 -1.05610478e+00
3.94691348e-01 -8.07707012e-01 6.97599709e-01 -4.09997702e-01
6.19797885e-01 -5.43693244e-01 -1.15767457e-01 7.95255184e-01
-6.90434933e-01 1.19616613e-01 -5.51945008e-02 6.35092556e-01
1.95674479e-01 -1.08645365e-01 6.12721562e-01 -7.90257573e-01
-1.12169492e+00 2.56077945e-02 -8.74087632e-01 1.56256899e-01
1.31994140e+00 -4.02974308e-01 -2.26513356e-01 -6.73473716e-01
-1.38925147e+00 9.40464556e-01 6.52584314e-01 3.65144759e-01
6.25449836e-01 -1.01296997e+00 -4.35092121e-01 -1.69257596e-02
-2.02497229e-01 2.42023945e-01 5.01448028e-02 6.20823383e-01
-6.49550617e-01 4.53792870e-01 -8.23982477e-01 -7.17670202e-01
-1.01539934e+00 6.09300613e-01 4.16068465e-01 -5.76045752e-01
-9.16747153e-01 7.96775222e-01 6.80372715e-01 -5.91894329e-01
4.30481911e-01 -2.75260657e-01 -4.29908633e-01 -3.35024744e-01
3.54312539e-01 3.13999414e-01 -6.79798782e-01 -2.73298454e-02
1.63807407e-01 2.91047841e-01 7.76516423e-02 -5.36576509e-01
1.37048697e+00 -1.24515809e-01 1.31260142e-01 5.66854775e-01
2.94463784e-01 -4.76751626e-01 -2.21535635e+00 1.65577963e-01
-4.71981689e-02 -4.10232276e-01 -1.24861680e-01 -8.59272122e-01
-3.84704918e-01 1.03886545e+00 3.39500427e-01 1.73169792e-01
8.09278548e-01 -9.06318501e-02 2.42104918e-01 1.09816074e+00
8.82183373e-01 -1.33323967e+00 8.81433666e-01 9.75935161e-01
1.06393337e+00 -1.24960864e+00 -9.11586881e-02 2.03325227e-01
-1.00278854e+00 1.05975199e+00 1.18885481e+00 -2.96824545e-01
1.07186534e-01 2.91439146e-01 -7.84663558e-02 -3.12752485e-01
-1.11676753e+00 -2.58347690e-01 -2.34855920e-01 9.12798584e-01
-6.32178113e-02 9.20721889e-02 2.10569844e-01 -1.08730361e-01
-5.10054767e-01 -2.83948183e-01 5.85111320e-01 1.16204286e+00
-1.12831461e+00 -6.19518995e-01 -4.22071964e-01 4.05477852e-01
-5.50107770e-02 1.64926901e-01 -1.07718475e-01 7.06673443e-01
-2.65163928e-01 6.91862106e-01 6.75773919e-02 -3.81619968e-02
-6.88323677e-02 -7.00398460e-02 8.15036952e-01 -9.82330561e-01
-2.31077999e-01 2.35267095e-02 -2.41680592e-02 -1.09507787e+00
-6.33951575e-02 -7.85712361e-01 -1.61866260e+00 4.86171804e-02
1.85801953e-01 9.45435390e-02 4.35701579e-01 8.07796896e-01
8.16836432e-02 6.44901574e-01 1.78712502e-01 -1.27830958e+00
-8.74741912e-01 -7.86451280e-01 -5.58751702e-01 1.72843099e-01
4.32703137e-01 -8.55167866e-01 -1.71648651e-01 8.50108117e-02] | [4.159060001373291, 1.3018606901168823] |
a65c7198-625a-4787-b958-98de98eab7e9 | targeted-attention-attack-on-deep-learning | 2010.04331 | null | https://arxiv.org/abs/2010.04331v3 | https://arxiv.org/pdf/2010.04331v3.pdf | Targeted Physical-World Attention Attack on Deep Learning Models in Road Sign Recognition | Real world traffic sign recognition is an important step towards building autonomous vehicles, most of which highly dependent on Deep Neural Networks (DNNs). Recent studies demonstrated that DNNs are surprisingly susceptible to adversarial examples. Many attack methods have been proposed to understand and generate adversarial examples, such as gradient based attack, score based attack, decision based attack, and transfer based attacks. However, most of these algorithms are ineffective in real-world road sign attack, because (1) iteratively learning perturbations for each frame is not realistic for a fast moving car and (2) most optimization algorithms traverse all pixels equally without considering their diverse contribution. To alleviate these problems, this paper proposes the targeted attention attack (TAA) method for real world road sign attack. Specifically, we have made the following contributions: (1) we leverage the soft attention map to highlight those important pixels and skip those zero-contributed areas - this also helps to generate natural perturbations, (2) we design an efficient universal attack that optimizes a single perturbation/noise based on a set of training images under the guidance of the pre-trained attention map, (3) we design a simple objective function that can be easily optimized, (4) we evaluate the effectiveness of TAA on real world data sets. Experimental results validate that the TAA method improves the attack successful rate (nearly 10%) and reduces the perturbation loss (about a quarter) compared with the popular RP2 method. Additionally, our TAA also provides good properties, e.g., transferability and generalization capability. We provide code and data to ensure the reproducibility: https://github.com/AdvAttack/RoadSignAttack. | ['DaCheng Tao', 'Wei Liu', 'Shengli Zhang', 'Weifeng Liu', 'Xinghao Yang'] | 2020-10-09 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 3.83557007e-02 -1.66192994e-01 -5.97767904e-02 -1.54488087e-01
-7.76044548e-01 -5.03318191e-01 6.22544885e-01 -7.15883791e-01
-3.93096179e-01 6.80872679e-01 8.23553354e-02 -4.89251941e-01
1.56045035e-01 -8.86932969e-01 -9.10369933e-01 -7.88501501e-01
1.33290276e-01 1.16804965e-01 6.99028313e-01 -5.40388286e-01
2.09083661e-01 7.06159294e-01 -1.38712621e+00 -3.45578068e-03
1.17985082e+00 8.42997313e-01 -2.09371001e-01 7.09167540e-01
5.85950390e-02 7.79705882e-01 -7.85402715e-01 -6.72741055e-01
6.24976575e-01 -4.52408820e-01 -4.19040293e-01 -5.34292340e-01
6.20271921e-01 -6.07537389e-01 -8.36128235e-01 1.37853634e+00
6.92506194e-01 1.15806080e-01 5.47176182e-01 -1.72195923e+00
-5.72659314e-01 3.28081399e-01 -5.33578694e-01 1.96555585e-01
-1.94342032e-01 9.09631491e-01 5.51216662e-01 -6.54729545e-01
3.93974006e-01 1.36436844e+00 5.55213213e-01 9.98683095e-01
-6.01515591e-01 -1.20072472e+00 1.46119565e-01 6.58119678e-01
-1.12737536e+00 -4.76396203e-01 7.82411098e-01 -1.31925017e-01
5.03593683e-01 3.67541581e-01 4.18509424e-01 1.38918972e+00
-9.10460670e-03 9.28502023e-01 8.20513427e-01 4.65384349e-02
6.64075837e-02 1.39976735e-03 -3.13787237e-02 5.69586933e-01
2.46051624e-01 5.38860440e-01 1.20228581e-01 3.18164937e-02
4.41075593e-01 -1.36467665e-01 -2.65708953e-01 -1.51419237e-01
-1.04139161e+00 8.20600927e-01 6.83609247e-01 -4.91888709e-02
-2.30453402e-01 4.77488309e-01 5.60546100e-01 1.85675412e-01
-1.21974900e-01 2.56910801e-01 -1.97707489e-01 -2.51933545e-01
-3.38543206e-01 3.40234816e-01 5.88225722e-01 6.86199248e-01
6.76735759e-01 5.36927223e-01 -4.07518357e-01 5.78354597e-01
1.43777609e-01 1.05538893e+00 2.97611982e-01 -8.31879020e-01
7.67739892e-01 2.23171353e-01 7.08874986e-02 -1.16662550e+00
-7.63376206e-02 -8.78257453e-02 -8.60357285e-01 7.58708537e-01
4.91121262e-01 -4.45283413e-01 -1.18017876e+00 1.77055669e+00
3.15689117e-01 5.01211882e-01 8.87721181e-02 9.43516970e-01
8.11336219e-01 5.92412114e-01 8.66399035e-02 2.29469433e-01
1.08065009e+00 -1.01198316e+00 -6.47410989e-01 -2.03445658e-01
3.96819800e-01 -5.72456121e-01 1.26926792e+00 1.07248291e-01
-7.67539322e-01 -4.82336372e-01 -1.19028306e+00 3.39713573e-01
-6.29028857e-01 -3.04268058e-02 4.01365578e-01 1.03881073e+00
-5.59542596e-01 3.55462879e-01 -6.90290093e-01 -3.83395478e-02
8.62504721e-01 2.75205523e-01 -2.20075518e-01 -1.46994293e-01
-1.45855176e+00 1.05064011e+00 1.31764576e-01 3.52487028e-01
-1.22946942e+00 -8.00158799e-01 -7.00704575e-01 -1.48922816e-01
3.88290703e-01 -4.63628739e-01 9.63389218e-01 -8.37786853e-01
-1.87397325e+00 3.79249007e-01 -1.36168497e-02 -5.59530914e-01
8.94505024e-01 -3.13870758e-01 -5.22093415e-01 5.80739044e-02
-1.54415637e-01 7.24110603e-01 8.21871281e-01 -1.27633393e+00
-4.86712396e-01 6.13567792e-02 1.95668235e-01 -4.11856472e-02
-2.73858637e-01 1.86085254e-01 -4.13758308e-01 -7.22507060e-01
-3.63226980e-01 -9.77960944e-01 -2.70682067e-01 2.29066610e-01
-5.86447716e-01 1.71927646e-01 1.34832811e+00 -5.75605094e-01
1.10878158e+00 -2.09577179e+00 -2.58536428e-01 3.11918437e-01
8.69372934e-02 1.08485806e+00 -3.14614534e-01 1.61235914e-01
-5.01809120e-02 3.94568443e-01 -3.72867018e-01 1.04513340e-01
1.84099063e-01 2.23796204e-01 -7.09119856e-01 3.95167083e-01
3.50199431e-01 1.26660085e+00 -8.80804718e-01 -1.87863737e-01
4.00908262e-01 4.49361473e-01 -5.22405028e-01 2.07021069e-02
-9.77334306e-02 3.84244233e-01 -5.84699690e-01 6.45755410e-01
8.79434943e-01 4.29653257e-01 -4.78087127e-01 -3.83760989e-01
9.58027095e-02 3.41449305e-02 -1.05547619e+00 7.79105067e-01
-1.85365438e-01 8.57657135e-01 -1.74982190e-01 -8.25176299e-01
8.93346488e-01 1.24954157e-01 1.04965590e-01 -8.00205171e-01
3.36867005e-01 1.77343503e-01 2.24882036e-01 -6.59337938e-01
1.66198581e-01 6.79158941e-02 -4.66118194e-03 2.35213786e-01
-4.48005348e-01 -7.41160661e-02 6.72965348e-02 1.47130311e-01
1.18764651e+00 -1.33857742e-01 -1.76030755e-01 2.60792881e-01
6.87003076e-01 -1.40554801e-01 7.79213786e-01 6.97722614e-01
-7.49650002e-01 5.56625962e-01 5.98157048e-01 -4.50944424e-01
-9.74980414e-01 -9.26212788e-01 2.02032238e-01 6.12433076e-01
5.15818059e-01 9.17031467e-02 -7.31212497e-01 -9.83090758e-01
6.97307959e-02 8.05109560e-01 -6.81115925e-01 -6.05878592e-01
-9.87229943e-01 -7.24550664e-01 1.10464656e+00 6.29830122e-01
1.11276615e+00 -1.18235600e+00 -2.77490824e-01 -2.39877224e-01
-9.17565227e-02 -1.09580362e+00 -5.93053222e-01 -6.18873179e-01
-3.53086233e-01 -1.19692051e+00 -6.04660809e-01 -5.55792868e-01
7.41414487e-01 3.67763430e-01 4.98131573e-01 1.07728563e-01
-1.87559918e-01 8.15649256e-02 -3.22755963e-01 -5.32216489e-01
-5.66530228e-01 -2.43842795e-01 6.51747361e-02 2.02711806e-01
3.26870024e-01 -5.07303059e-01 -6.48587644e-01 6.85608447e-01
-7.84601390e-01 -3.04040700e-01 7.25631475e-01 7.60646820e-01
2.44895115e-01 -5.73403724e-02 5.80013514e-01 -6.81117535e-01
5.93591332e-01 -3.24235916e-01 -7.92931497e-01 4.18825001e-02
-2.78858721e-01 -2.47096345e-02 7.65557587e-01 -7.29095876e-01
-8.94422710e-01 -2.47379020e-01 -3.39865923e-01 -6.89958930e-01
-2.34981820e-01 -4.87613119e-02 -5.01159608e-01 -6.03003502e-01
8.30355704e-01 2.10769758e-01 2.84957856e-01 2.77532358e-02
4.73872095e-01 5.50788343e-01 6.17165446e-01 -4.13317442e-01
1.46468711e+00 5.46011090e-01 7.55703971e-02 -5.45193493e-01
-3.67556393e-01 1.27672285e-01 -9.72651020e-02 -5.15884101e-01
7.15057731e-01 -5.28316736e-01 -1.06558025e+00 7.71563053e-01
-9.78815615e-01 -5.31779051e-01 -1.02655776e-01 4.94596243e-01
-3.36927772e-01 6.65636718e-01 -4.30429429e-01 -7.51425624e-01
-3.77124250e-01 -1.25400889e+00 5.39243758e-01 5.88770628e-01
1.99973330e-01 -6.15188897e-01 -1.75113171e-01 4.55230355e-01
8.54154825e-01 6.44881189e-01 6.66561007e-01 -5.99352837e-01
-8.88011873e-01 -4.13007200e-01 -4.60931897e-01 4.79076594e-01
-4.65192739e-03 3.45936179e-01 -8.99701774e-01 -1.13490112e-01
-4.00545120e-01 -3.66437346e-01 8.43256652e-01 2.39459738e-01
1.13764465e+00 -4.43169564e-01 -3.18661809e-01 7.26337731e-01
1.07820058e+00 3.19502413e-01 1.16976976e+00 4.97877866e-01
8.89093161e-01 3.11219096e-01 5.66148341e-01 -3.73504013e-02
2.13535383e-01 6.35257781e-01 8.58477890e-01 -1.35973528e-01
-2.83564210e-01 -2.94864088e-01 6.98942304e-01 3.00336152e-01
5.51236270e-04 -1.94460794e-01 -7.37539470e-01 5.85489690e-01
-1.74591243e+00 -1.27420771e+00 -1.13604330e-01 2.20951891e+00
5.54296851e-01 5.21180809e-01 2.33435690e-01 8.29389021e-02
8.05863380e-01 1.81284875e-01 -7.99012840e-01 -2.67118931e-01
-9.86434668e-02 1.01045102e-01 7.52466917e-01 5.58600605e-01
-1.14934325e+00 1.23498058e+00 5.69414234e+00 1.01551640e+00
-1.40709114e+00 -2.73880828e-02 4.51709569e-01 -7.32676312e-02
-1.81582332e-01 -1.89272776e-01 -7.33896315e-01 7.53379345e-01
7.06564605e-01 -3.24631274e-01 3.49187165e-01 8.72164249e-01
3.31393957e-01 4.54685032e-01 -5.86266816e-01 7.25777149e-01
5.17551228e-02 -1.12710047e+00 2.70729065e-01 2.61298027e-02
6.59210205e-01 1.87310815e-01 3.47179800e-01 6.52850986e-01
5.99831998e-01 -1.01556516e+00 4.94228572e-01 3.44193012e-01
7.39796519e-01 -9.09340799e-01 7.29188621e-01 8.35385248e-02
-9.50383484e-01 -1.54766321e-01 -3.75928700e-01 3.80146176e-01
3.27701569e-01 1.02872878e-01 -4.63459283e-01 4.32578892e-01
4.81944501e-01 4.49526876e-01 -4.23487544e-01 1.32210135e+00
-6.75259769e-01 8.74829888e-01 -4.03579891e-01 -1.09960526e-01
4.76180375e-01 -1.65823653e-01 9.57586646e-01 1.01634002e+00
1.27331421e-01 -3.24216336e-02 -1.88367814e-01 8.42447281e-01
-1.24585610e-02 -1.55232161e-01 -5.64779580e-01 2.69966185e-01
5.59804559e-01 1.03049350e+00 -2.04377860e-01 -1.87851980e-01
-2.14254752e-01 7.87706912e-01 -2.92689092e-02 6.36849463e-01
-1.49732232e+00 -9.35687244e-01 1.12365127e+00 -1.59698188e-01
5.54636538e-01 2.22976189e-02 -1.33353204e-01 -9.23992097e-01
2.37470001e-01 -1.05781245e+00 1.58767343e-01 -7.24331677e-01
-1.06711423e+00 6.18434668e-01 -1.23340949e-01 -1.56782353e+00
1.18728392e-02 -7.10225284e-01 -1.14069939e+00 7.58792698e-01
-1.68701422e+00 -1.04499447e+00 -5.15518606e-01 6.98960364e-01
3.80039573e-01 -3.80652159e-01 4.24936891e-01 5.73674202e-01
-1.00515139e+00 1.19199753e+00 -1.21079192e-01 5.03344715e-01
5.97455859e-01 -8.15197825e-01 6.60891533e-01 1.24638116e+00
-2.21294120e-01 3.04795682e-01 6.37375176e-01 -4.88612235e-01
-1.20878887e+00 -1.35234666e+00 3.49263728e-01 -4.53068316e-01
7.14793503e-01 -9.73814502e-02 -8.42409730e-01 4.67051864e-01
-1.10141076e-01 2.48211965e-01 1.54277712e-01 -5.64054966e-01
-5.25424600e-01 -4.78666931e-01 -1.32343614e+00 1.10163188e+00
9.29975390e-01 -2.84407109e-01 -2.99220890e-01 1.98394418e-01
6.87219799e-01 -6.24240994e-01 -3.37994516e-01 5.22686481e-01
5.68657756e-01 -8.30038786e-01 9.86761451e-01 -7.39599407e-01
3.19998741e-01 -7.22750843e-01 1.09644100e-01 -1.10541821e+00
-1.41787753e-01 -8.82509112e-01 -2.24509388e-01 1.13247550e+00
4.71825570e-01 -1.10275066e+00 8.57052982e-01 6.94763780e-01
-2.40258813e-01 -8.30239952e-01 -9.28252280e-01 -8.72854173e-01
1.14029288e-01 -5.05612314e-01 7.66628802e-01 7.40602970e-01
-4.98351157e-01 -1.92846090e-01 -6.28696978e-01 5.53440630e-01
5.77583134e-01 -3.77354085e-01 1.30307829e+00 -6.20039225e-01
-7.75184715e-03 -7.38464773e-01 -8.42527986e-01 -1.08638811e+00
7.32084662e-02 -5.39886117e-01 8.75346884e-02 -1.20659542e+00
-2.45287284e-01 -4.44247246e-01 -3.07993710e-01 6.30776882e-01
-4.98741925e-01 2.87685037e-01 3.60654444e-01 7.61724412e-02
-3.45271051e-01 7.11937904e-01 1.25765896e+00 -3.48615497e-01
8.57437029e-02 2.72825420e-01 -7.03413069e-01 6.15241468e-01
9.30127025e-01 -5.18964767e-01 -2.86554098e-01 -3.40788096e-01
-2.24987596e-01 -4.69462126e-01 5.57630479e-01 -1.21659112e+00
2.05766112e-01 -3.46573919e-01 -4.52509932e-02 -4.02155846e-01
2.43072569e-01 -6.94033444e-01 -1.78346723e-01 7.21549332e-01
2.32196264e-02 -3.14581335e-01 4.20976460e-01 5.80004215e-01
-2.04767838e-01 -3.93266641e-02 1.03936052e+00 3.04772466e-01
-8.54826629e-01 6.06687427e-01 -1.75239071e-01 1.96936339e-01
1.22466934e+00 -2.96122342e-01 -6.42883062e-01 -6.13253057e-01
-1.17996842e-01 5.27869463e-01 2.19182268e-01 6.00250185e-01
5.41189551e-01 -1.48105407e+00 -8.17531049e-01 2.62247354e-01
-1.96174625e-03 -1.51705623e-01 2.53302485e-01 6.84654832e-01
-8.61974001e-01 2.56822795e-01 -2.96948314e-01 -3.44564408e-01
-1.20589185e+00 5.53213179e-01 6.04373038e-01 4.19047512e-02
-6.12996936e-01 8.41419995e-01 2.31833637e-01 -4.25035477e-01
3.31068903e-01 -1.52833566e-01 -1.92343459e-01 -4.58443612e-01
7.59852707e-01 5.09728372e-01 -2.90748626e-01 -7.79982865e-01
-3.59452397e-01 7.28145003e-01 -2.23166555e-01 5.29618002e-02
1.00378883e+00 3.17887157e-01 3.29260647e-01 -3.48020107e-01
1.04154468e+00 9.37923864e-02 -1.54854429e+00 -6.89555937e-03
-4.76337761e-01 -7.31420040e-01 -1.73684090e-01 -6.68414712e-01
-1.43688452e+00 7.96141803e-01 5.35985768e-01 -8.04582797e-03
1.07385790e+00 -4.69226003e-01 1.21612501e+00 3.81338567e-01
4.54427563e-02 -8.06606352e-01 4.13378663e-02 4.84704643e-01
9.35709715e-01 -1.19058836e+00 -2.67601937e-01 -3.12126577e-01
-8.61503124e-01 8.61405075e-01 9.53791916e-01 -2.81270772e-01
5.19542396e-01 1.51309505e-01 3.32036287e-01 8.46557245e-02
-5.53922951e-01 -2.77927905e-01 1.16454698e-01 7.56154656e-01
-3.61123890e-01 -1.10168688e-01 -1.71771079e-01 2.74937004e-01
-2.22048387e-01 -1.04973242e-01 4.23204690e-01 6.16569221e-01
-3.07771116e-01 -9.78017211e-01 -3.31603050e-01 2.45806322e-01
-3.19478214e-01 5.95260262e-02 -2.85606533e-01 8.57792020e-01
9.75295305e-02 8.38037133e-01 -3.66099328e-01 -6.13786817e-01
7.03181982e-01 -1.81918755e-01 4.05416787e-02 1.96858533e-02
-3.98619533e-01 -6.56407595e-01 3.01135983e-02 -7.58137941e-01
8.15503150e-02 -3.83197337e-01 -1.12583435e+00 -5.61534047e-01
-3.47201347e-01 -1.17663711e-01 6.22394145e-01 8.97532821e-01
5.40310323e-01 4.58991766e-01 8.31403971e-01 -8.90911222e-01
-6.61742389e-01 -8.32683444e-01 4.37699147e-02 5.41113555e-01
3.43972921e-01 -7.19218791e-01 -6.88272655e-01 -1.97233945e-01] | [5.39818000793457, 7.8810954093933105] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.