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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
4f802da1-9151-4031-92cd-3411e8e08a47 | active-stacking-for-heart-rate-estimation | 1903.10862 | null | http://arxiv.org/abs/1903.10862v1 | http://arxiv.org/pdf/1903.10862v1.pdf | Active Stacking for Heart Rate Estimation | Heart rate estimation from electrocardiogram signals is very important for
the early detection of cardiovascular diseases. However, due to large
individual differences and varying electrocardiogram signal quality, there does
not exist a single reliable estimation algorithm that works well on all
subjects. Every algorithm may break down on certain subjects, resulting in a
significant estimation error. Ensemble regression, which aggregates the outputs
of multiple base estimators for more reliable and stable estimates, can be used
to remedy this problem. Moreover, active learning can be used to optimally
select a few trials from a new subject to label, based on which a stacking
ensemble regression model can be trained to aggregate the base estimators. This
paper proposes four active stacking approaches, and demonstrates that they all
significantly outperform three common unsupervised ensemble regression
approaches, and a supervised stacking approach which randomly selects some
trials to label. Remarkably, our active stacking approaches only need three or
four labeled trials from each subject to achieve an average root mean squared
estimation error below three beats per minute, making them very convenient for
real-world applications. To our knowledge, this is the first research on active
stacking, and its application to heart rate estimation. | ['Chengyu Liu', 'Feifei Liu', 'Dongrui Wu'] | 2019-03-26 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [ 3.06838661e-01 -2.33330384e-01 -4.44546402e-01 -5.36448956e-01
-9.73249912e-01 -2.38237053e-01 -1.42270371e-01 3.78942400e-01
-1.96145386e-01 9.32173669e-01 -4.09274846e-01 -2.08007723e-01
-1.80446461e-01 -4.63302940e-01 5.81413023e-02 -9.51314926e-01
-2.02174738e-01 3.25789452e-01 -8.07011500e-02 2.55303532e-01
2.42650241e-01 2.25344896e-01 -9.95026171e-01 -1.75615057e-01
1.31626141e+00 8.89049888e-01 -1.19522296e-01 7.68570125e-01
1.45271823e-01 9.32443619e-01 -8.88264835e-01 8.26690644e-02
-7.99904764e-02 -8.70798171e-01 -2.17026830e-01 -9.65821967e-02
1.68067858e-01 -2.72062749e-01 2.89760321e-01 4.48573023e-01
1.03134656e+00 7.08577596e-03 5.78147948e-01 -9.92654562e-01
4.61704545e-02 6.00137353e-01 -6.10977173e-01 2.81487286e-01
3.20169032e-01 -6.46421537e-02 6.16954923e-01 -8.56573343e-01
-7.50248134e-02 5.02226233e-01 1.00960922e+00 5.31516135e-01
-1.49108100e+00 -8.92005205e-01 1.00840107e-01 1.90330483e-02
-1.53636920e+00 -7.87163138e-01 8.90038550e-01 -4.03722107e-01
4.28206593e-01 6.11126602e-01 8.35212827e-01 7.32407451e-01
2.40886763e-01 5.23393333e-01 1.30613530e+00 -4.73900527e-01
4.43927556e-01 2.84129530e-01 4.73165274e-01 5.42292058e-01
5.04849732e-01 2.99045276e-02 -5.99867702e-01 -5.17635286e-01
7.35142648e-01 1.88030213e-01 -2.95492917e-01 -3.22031945e-01
-1.18683481e+00 6.13174498e-01 -1.20997094e-02 1.66995332e-01
-4.98053282e-01 5.65080345e-02 3.19247037e-01 2.19887421e-01
6.48666978e-01 5.97481489e-01 -4.96720791e-01 -1.08399920e-01
-1.21273780e+00 1.21668734e-01 9.62876439e-01 4.36798304e-01
5.34219980e-01 3.41549441e-02 -8.42665806e-02 8.77946556e-01
3.41772616e-01 5.15226722e-01 1.88556090e-01 -1.02245629e+00
2.74059623e-01 5.59271097e-01 1.64663941e-01 -8.40160847e-01
-5.22574067e-01 -6.08180761e-01 -9.77331161e-01 1.58708140e-01
4.17091906e-01 -4.76980060e-01 -4.74429637e-01 1.33515108e+00
2.09745869e-01 4.24296886e-01 -2.97430128e-01 7.70798266e-01
8.48002195e-01 2.95466840e-01 2.74216950e-01 -1.03452301e+00
8.99223149e-01 -4.85374004e-01 -7.68891454e-01 -1.25365287e-01
5.35924494e-01 -5.69347143e-01 4.65582728e-01 7.45658338e-01
-9.99193430e-01 -6.11342251e-01 -1.10676634e+00 6.76366210e-01
1.36054337e-01 1.37471080e-01 5.57497323e-01 9.56882000e-01
-5.45665920e-01 7.80117452e-01 -8.70352268e-01 8.91789868e-02
4.06181008e-01 4.74284142e-01 6.34568557e-02 1.87296152e-01
-1.11408973e+00 9.78655636e-01 1.34648636e-01 5.33811271e-01
-3.40921342e-01 -8.27749789e-01 -6.06320024e-01 -3.05604458e-01
3.58566284e-01 -6.91726327e-01 1.09948838e+00 -8.52239907e-01
-1.62588120e+00 4.89190310e-01 -3.87852162e-01 -3.94549370e-01
5.23592055e-01 -4.28522766e-01 -4.23109412e-01 8.67055058e-02
-1.06450833e-01 4.28865701e-02 1.05211771e+00 -9.91974771e-01
-1.26707360e-01 -3.53262365e-01 -5.44485509e-01 2.78706044e-01
8.92455578e-02 9.93769467e-02 -5.00872917e-02 -6.13834858e-01
4.71539706e-01 -9.33038354e-01 -6.21173263e-01 -2.83147186e-01
-2.70617157e-01 -1.93644539e-01 5.22893310e-01 -5.67660511e-01
1.61369288e+00 -2.16826344e+00 9.37742591e-02 4.47574317e-01
7.27867961e-01 1.96259841e-01 4.55086291e-01 6.82942644e-02
-1.05864190e-01 1.67854071e-01 -3.07545394e-01 -1.39881283e-01
-5.84980726e-01 4.81362343e-02 -1.96842209e-01 5.82705259e-01
1.21856993e-02 7.18106747e-01 -9.83742237e-01 -6.57377779e-01
4.43057835e-01 3.03328037e-01 -4.10654843e-02 2.54088402e-01
5.04145920e-01 8.11317503e-01 -4.07928199e-01 5.42668760e-01
3.97353113e-01 -3.71302009e-01 4.14785504e-01 1.47340950e-02
5.27221933e-02 1.10548764e-01 -1.20491028e+00 1.14176559e+00
-3.51311058e-01 5.19072175e-01 -4.87283260e-01 -1.41630542e+00
1.33350527e+00 7.46642649e-01 1.13774717e+00 -3.15883607e-01
-1.00573786e-01 4.58496779e-01 1.54694915e-01 -4.07582313e-01
-3.11281048e-02 -1.44958675e-01 -1.45286947e-01 5.41079700e-01
-1.48833320e-01 2.82707363e-02 1.63447931e-01 2.94827782e-02
9.14294899e-01 -7.30092302e-02 8.62623811e-01 -1.22066282e-01
4.59151059e-01 -4.68955517e-01 1.09725034e+00 1.07807612e+00
-5.45685410e-01 6.84912503e-01 2.37666309e-01 -7.04027951e-01
-5.16426444e-01 -9.98157978e-01 -3.38016689e-01 7.86803484e-01
3.49325910e-02 -5.24716794e-01 -5.51997960e-01 -7.95379460e-01
-1.25296295e-01 3.56828839e-01 -3.88655663e-01 1.23978527e-02
-8.41671884e-01 -1.27550030e+00 4.08998191e-01 5.49936652e-01
3.63935381e-01 -9.10944819e-01 -7.61034846e-01 7.22289503e-01
-4.20350730e-01 -6.43743098e-01 -1.36543199e-01 3.79361570e-01
-1.41521835e+00 -1.15792882e+00 -8.67031217e-01 -2.75102317e-01
6.43165410e-01 6.47224486e-02 1.25777519e+00 2.99747974e-01
-3.22296470e-01 1.89911008e-01 -4.57490712e-01 -8.35445046e-01
-2.78241545e-01 1.82147007e-02 2.01934949e-01 -4.78381254e-02
3.41569960e-01 -4.77484703e-01 -7.57001758e-01 4.15652484e-01
3.81197184e-02 -1.44009784e-01 2.97498822e-01 7.75362492e-01
5.21054268e-01 -6.28057569e-02 1.06366110e+00 -1.12159491e+00
4.62703526e-01 -3.49577278e-01 -2.78923273e-01 3.31898659e-01
-9.25425410e-01 -3.18594396e-01 4.76703703e-01 -4.43282843e-01
-6.37464941e-01 3.98870081e-01 2.35250890e-01 -5.07662743e-02
6.39169589e-02 4.02663141e-01 3.86953831e-01 -4.18316498e-02
9.63414431e-01 6.40582293e-02 1.73016518e-01 -2.91299611e-01
-2.23161820e-02 6.78626120e-01 2.99409449e-01 -2.42373571e-01
5.02539456e-01 2.24466193e-02 3.35374624e-02 -8.99682224e-01
-9.66098130e-01 -6.00831449e-01 -8.17879677e-01 -6.62479877e-01
6.67572856e-01 -8.45871568e-01 -3.86777937e-01 6.09258950e-01
-8.80194187e-01 -2.44325221e-01 -1.60096839e-01 7.96535611e-01
-5.83530307e-01 4.40956593e-01 -2.56193548e-01 -1.17822349e+00
-7.15810776e-01 -7.85163462e-01 5.78229070e-01 2.81852812e-01
-9.01280224e-01 -9.66908455e-01 1.82661742e-01 2.84641832e-01
2.99816221e-01 4.90734935e-01 4.30411130e-01 -6.41163945e-01
-2.24912733e-01 -3.59101832e-01 3.47457558e-01 3.45197886e-01
3.81051123e-01 1.30812034e-01 -9.78341281e-01 -3.42110336e-01
2.39702463e-01 6.00903556e-02 7.88711071e-01 7.77498007e-01
1.16125548e+00 6.09077848e-02 -4.36846882e-01 2.26158723e-01
9.89648044e-01 5.04804611e-01 8.01692665e-01 -2.11762726e-01
4.57762718e-01 1.70990899e-01 4.67338294e-01 4.05322820e-01
3.80778743e-04 5.42224407e-01 -2.41285309e-01 -5.36348641e-01
2.59780228e-01 4.02128696e-01 3.49756271e-01 1.14861047e+00
-5.93682766e-01 1.66640997e-01 -9.23942924e-01 -9.06586275e-02
-1.84655905e+00 -1.15505135e+00 -4.08055276e-01 2.78783846e+00
8.24600637e-01 1.21765479e-01 3.25419456e-01 7.63639688e-01
7.29579031e-01 -4.19838503e-02 -6.58468544e-01 -1.77651942e-01
2.54054338e-01 4.13380444e-01 2.75000244e-01 2.00328439e-01
-1.24367881e+00 -5.56232892e-02 7.43341970e+00 2.54249632e-01
-1.12864900e+00 8.54156762e-02 7.91222036e-01 -2.03787200e-02
3.03797215e-01 1.99615270e-01 -6.43429637e-01 7.42462575e-01
1.07421994e+00 -2.07903072e-01 -3.99591103e-02 7.16552615e-01
4.08655733e-01 -3.96789700e-01 -1.01517308e+00 1.35271716e+00
1.30713254e-01 -9.40056741e-01 -6.21854126e-01 -5.44610135e-02
5.39178014e-01 -3.43034267e-01 -4.00553137e-01 2.59060442e-01
-1.67057127e-01 -9.33623433e-01 1.34240896e-01 9.47687805e-01
7.29286909e-01 -6.15818143e-01 7.53591537e-01 5.07942021e-01
-1.14798057e+00 -1.08627267e-01 -9.78328064e-02 9.37057473e-03
1.68070763e-01 1.17299497e+00 -7.30229318e-01 4.28646445e-01
4.77671325e-01 8.23754907e-01 -4.86779213e-01 1.28352320e+00
-8.12756196e-02 8.52003276e-01 -2.67199069e-01 -3.21773775e-02
-6.84354961e-01 -1.33051217e-01 6.11595571e-01 9.00566697e-01
3.35447073e-01 3.52589697e-01 4.02689785e-01 4.35904384e-01
3.95220786e-01 3.71738672e-01 -5.01300037e-01 3.18162233e-01
7.38822460e-01 1.13387060e+00 -8.68953824e-01 -5.63890517e-01
-4.31148946e-01 7.30270684e-01 9.74778309e-02 2.69426137e-01
-8.16625416e-01 -5.97886741e-01 -1.19869024e-01 1.70234844e-01
-2.80547708e-01 -1.08896978e-01 -7.05067039e-01 -1.11000705e+00
-1.15498483e-01 -9.22069907e-01 3.98556471e-01 -3.88368785e-01
-1.28234422e+00 2.52366543e-01 -2.52758507e-02 -1.66484654e+00
-3.50131065e-01 -2.48983093e-02 -8.19520891e-01 9.25198019e-01
-8.03403437e-01 -9.71820727e-02 -5.44601858e-01 1.72223806e-01
4.80474055e-01 -2.81578958e-01 1.14360023e+00 3.35998535e-01
-9.71734464e-01 5.43036222e-01 -1.38348877e-01 1.80785328e-01
9.64576662e-01 -1.36396003e+00 7.23013729e-02 5.24874926e-01
5.17698713e-02 7.80813575e-01 6.37793362e-01 -5.05856276e-01
-1.03728652e+00 -8.85183156e-01 9.34528828e-01 -6.58865511e-01
1.93985999e-02 6.87264949e-02 -1.04988647e+00 2.52957106e-01
-1.70001313e-01 2.49341443e-01 1.25023198e+00 5.64659238e-01
1.91588297e-01 -5.19479573e-01 -7.97343612e-01 2.88557857e-01
4.85074610e-01 -2.79285192e-01 -4.88559723e-01 6.78479746e-02
-1.41050681e-01 -4.57865953e-01 -1.21015656e+00 7.98265159e-01
8.69090915e-01 -1.02797008e+00 9.17050004e-01 -1.29808441e-01
-1.16966888e-01 -2.98701197e-01 5.19374728e-01 -1.27768588e+00
-4.14041370e-01 -7.70575166e-01 -5.73366880e-01 1.04607105e+00
6.46025658e-01 -8.47157478e-01 5.86370289e-01 7.14967012e-01
1.94233611e-01 -1.01280236e+00 -5.80024123e-01 -6.71770334e-01
-3.16778064e-01 -3.92308265e-01 1.45687878e-01 9.76745188e-01
1.68503955e-01 5.01161814e-01 -5.48232973e-01 -2.55896777e-01
1.07597566e+00 8.74779001e-02 7.99238622e-01 -1.72279751e+00
-3.46514463e-01 -1.31304683e-02 -3.56979370e-01 -8.06248605e-01
-2.60245204e-01 -6.09995484e-01 1.95731372e-01 -1.28164279e+00
2.71274507e-01 -1.05944812e+00 -7.76944518e-01 3.81345719e-01
-7.82465041e-01 5.43134928e-01 4.18150332e-04 4.32205737e-01
-5.50352335e-01 -7.68871158e-02 8.25595677e-01 1.51314855e-01
-9.17288065e-01 5.52040935e-01 -8.17867875e-01 7.86046147e-01
1.23211026e+00 -6.92416191e-01 -5.00244856e-01 2.51254946e-01
2.83843726e-01 4.70498294e-01 1.81483030e-01 -1.09015620e+00
3.13712470e-02 -9.92838442e-02 8.16219211e-01 -5.75107396e-01
7.39793852e-02 -3.95099908e-01 3.60219896e-01 5.05664766e-01
-3.21508616e-01 -2.47390345e-01 -2.30982691e-01 4.44580942e-01
-1.11257955e-01 -2.99106091e-01 9.01913941e-01 -1.92674279e-01
-3.28529060e-01 2.73406386e-01 -5.99212348e-01 1.10996902e-01
1.05839396e+00 -5.33682704e-01 3.77482027e-02 -2.12513000e-01
-9.71450508e-01 1.10968634e-01 -1.12372249e-01 4.19733934e-02
6.16897166e-01 -1.21039116e+00 -8.89294922e-01 1.81428283e-01
4.44507003e-02 -2.71336809e-02 3.02969933e-01 1.42001629e+00
-3.20806533e-01 -5.05001005e-03 7.38630891e-02 -8.90518844e-01
-1.54168427e+00 1.08537808e-01 5.81218421e-01 -1.51888490e-01
-7.63358474e-01 4.41299647e-01 -2.59374589e-01 5.39900325e-02
5.44344373e-02 -8.29660892e-02 -3.53471935e-01 3.13160866e-01
7.46864259e-01 7.48053133e-01 2.35821888e-01 -1.93116099e-01
-4.29771096e-01 7.68271804e-01 -2.78879516e-02 1.79776713e-01
8.97094786e-01 -1.65824756e-01 1.30572235e-02 8.77288520e-01
8.17674816e-01 9.50656831e-02 -8.97793651e-01 -1.15981966e-01
-4.45629209e-02 -2.33163685e-01 1.18626557e-01 -6.02545559e-01
-9.05138552e-01 7.66258836e-01 7.32198000e-01 4.39672589e-01
1.32232392e+00 -5.26357770e-01 3.75799358e-01 3.19748670e-01
3.57510567e-01 -1.21046650e+00 -3.03081237e-02 2.41807457e-02
5.44522166e-01 -1.22964942e+00 6.75427616e-01 -4.83542860e-01
-7.40526795e-01 1.32207203e+00 2.78039187e-01 -1.73452392e-01
7.16409326e-01 1.91618383e-01 3.52073699e-01 5.46769425e-02
-7.20522881e-01 1.74334839e-01 3.33081245e-01 5.49419343e-01
8.06648791e-01 3.10828567e-01 -4.75258559e-01 2.55448818e-01
1.39183849e-01 2.86408156e-01 2.93396145e-01 8.09731066e-01
-5.09215474e-01 -1.22365797e+00 -4.08271790e-01 9.38327491e-01
-5.68120480e-01 1.29284397e-01 -2.57671148e-01 4.91069496e-01
3.08845546e-02 1.19558120e+00 -1.02262581e-02 -2.56922632e-01
2.05097035e-01 4.25898284e-01 6.77432418e-01 -7.16829300e-01
-6.70050025e-01 1.38073877e-01 1.29643887e-01 -3.41263145e-01
-6.04040563e-01 -8.39365482e-01 -1.05173767e+00 -2.02533111e-01
-6.52323365e-01 2.77585745e-01 2.44417951e-01 8.68947148e-01
6.81237802e-02 4.30897206e-01 1.12803030e+00 -5.66897571e-01
-5.05030930e-01 -8.77479374e-01 -7.53885329e-01 7.62597919e-02
4.32437956e-01 -6.25905275e-01 -4.74163711e-01 2.12780297e-01] | [14.283777236938477, 3.296741485595703] |
f30fb240-fae5-4b7e-8a3b-5fdcab4f2e7c | conditional-temporal-variational-autoencoder | 2108.05658 | null | https://arxiv.org/abs/2108.05658v1 | https://arxiv.org/pdf/2108.05658v1.pdf | Conditional Temporal Variational AutoEncoder for Action Video Prediction | To synthesize a realistic action sequence based on a single human image, it is crucial to model both motion patterns and diversity in the action video. This paper proposes an Action Conditional Temporal Variational AutoEncoder (ACT-VAE) to improve motion prediction accuracy and capture movement diversity. ACT-VAE predicts pose sequences for an action clips from a single input image. It is implemented as a deep generative model that maintains temporal coherence according to the action category with a novel temporal modeling on latent space. Further, ACT-VAE is a general action sequence prediction framework. When connected with a plug-and-play Pose-to-Image (P2I) network, ACT-VAE can synthesize image sequences. Extensive experiments bear out our approach can predict accurate pose and synthesize realistic image sequences, surpassing state-of-the-art approaches. Compared to existing methods, ACT-VAE improves model accuracy and preserves diversity. | ['Jiaya Jia', 'Bei Yu', 'LiWei Wang', 'Yi Wang', 'Xiaogang Xu'] | 2021-08-12 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 2.19841778e-01 8.42986032e-02 -2.26029024e-01 -5.51374704e-02
-5.78083932e-01 -1.51180789e-01 7.72029638e-01 -8.84168744e-01
1.54978465e-02 5.89242637e-01 6.33475184e-01 1.45456240e-01
2.90252000e-01 -7.07149267e-01 -9.99723077e-01 -5.72028995e-01
1.65962160e-01 4.02728587e-01 4.77903336e-01 -2.75975645e-01
-7.10718557e-02 3.04474115e-01 -1.69458640e+00 8.42202306e-01
5.08849144e-01 8.68887365e-01 5.74515879e-01 1.01536608e+00
2.89527833e-01 1.59224510e+00 -4.58936602e-01 -2.46768743e-01
3.15113634e-01 -9.86516893e-01 -6.77629828e-01 2.57280827e-01
3.67209166e-01 -8.16601038e-01 -7.11717427e-01 5.39594710e-01
4.39535886e-01 3.82104665e-01 8.29490185e-01 -1.46367943e+00
-6.04855061e-01 4.91402924e-01 -2.95919389e-01 -1.06713943e-01
5.28384447e-01 5.06382823e-01 8.92567754e-01 -6.29580736e-01
1.24216843e+00 1.39872861e+00 5.26095033e-01 9.00733709e-01
-1.05890882e+00 -3.63104403e-01 1.01083793e-01 6.43047631e-01
-1.15568006e+00 -1.35462224e-01 7.45954812e-01 -5.57781637e-01
9.64300334e-01 2.06354633e-01 1.27923119e+00 1.74647558e+00
6.33996785e-01 1.45397091e+00 4.70710754e-01 6.99828938e-02
2.81520575e-01 -7.15875804e-01 -9.62302506e-01 4.70149577e-01
-5.92844605e-01 2.64600009e-01 -7.96936452e-01 2.55350024e-01
1.25147855e+00 1.48007676e-01 -2.13520601e-01 -6.10572875e-01
-1.43448961e+00 6.47287190e-01 2.49754891e-01 2.49856442e-01
-8.33235085e-01 5.84646165e-01 2.04165846e-01 -1.26086092e-02
1.51300892e-01 2.13057354e-01 -1.40713587e-01 -6.84809268e-01
-1.28210664e+00 6.20330989e-01 4.63480562e-01 9.24422383e-01
2.94421583e-01 6.04984343e-01 -6.31060898e-01 4.21164781e-01
2.24576235e-01 5.16185939e-01 5.85250139e-01 -1.56935251e+00
1.01565003e-01 2.45647386e-01 1.49386078e-01 -1.00772202e+00
-7.83129595e-04 -8.59713852e-02 -8.20051312e-01 3.32483828e-01
3.23782396e-03 3.89633141e-02 -9.57470357e-01 1.68824637e+00
2.13706687e-01 5.77150106e-01 -2.30596606e-02 9.37211215e-01
6.50256693e-01 9.78701472e-01 8.11797753e-02 -3.04370463e-01
7.43701041e-01 -1.25951219e+00 -9.70390320e-01 -7.45510831e-02
2.80621767e-01 -5.02045333e-01 7.46081889e-01 4.26316917e-01
-1.39831769e+00 -9.69569921e-01 -8.79765153e-01 -2.00362220e-01
2.53880888e-01 9.79029108e-03 3.27739507e-01 1.35788471e-01
-9.64043319e-01 9.21360970e-01 -1.32239175e+00 -8.05853158e-02
3.17390949e-01 4.24308255e-02 -4.27716911e-01 4.01514880e-02
-1.02102959e+00 7.63723254e-01 4.25847799e-01 -1.20778950e-02
-1.49416494e+00 -5.86402893e-01 -1.14516258e+00 -4.96272221e-02
2.10233867e-01 -1.06401300e+00 1.47178972e+00 -1.14755964e+00
-2.04223824e+00 2.37330586e-01 -1.67396694e-01 -6.06697977e-01
9.81900990e-01 -3.04143995e-01 -3.62140179e-01 3.79255265e-01
5.87453097e-02 1.26163220e+00 9.49973166e-01 -1.18784440e+00
-4.35045570e-01 1.53998524e-01 -1.24565095e-01 2.17652753e-01
6.95855319e-02 -2.81213552e-01 -6.38657331e-01 -8.28535438e-01
-2.27366954e-01 -1.07192624e+00 -3.30589443e-01 1.75234854e-01
-1.91633433e-01 4.18833420e-02 9.66971695e-01 -8.12212825e-01
1.38872969e+00 -1.92673767e+00 7.79214919e-01 -4.21876937e-01
-2.88094152e-02 3.94654304e-01 -3.36892039e-01 6.10261440e-01
-1.44837564e-02 -2.94892609e-01 -1.41083032e-01 -5.78807712e-01
1.18664116e-01 5.68896055e-01 -2.99219280e-01 3.09386551e-01
4.31218259e-02 1.34674799e+00 -9.08107698e-01 -5.67628086e-01
7.07490146e-01 7.15251327e-01 -9.24182177e-01 3.43825161e-01
-7.48740733e-01 8.73182237e-01 -3.21205318e-01 4.62622046e-01
4.11211699e-01 -2.89140999e-01 3.15012723e-01 -2.87871450e-01
-7.62973167e-03 -2.57152200e-01 -1.03701818e+00 2.13815069e+00
-2.38197535e-01 6.88163102e-01 -4.62326080e-01 -6.59080803e-01
6.34043813e-01 4.47536021e-01 9.54793215e-01 -7.50205338e-01
3.04407984e-01 -5.04866168e-02 -2.16617405e-01 -6.41672969e-01
6.59582019e-01 -2.10894179e-03 -6.47284761e-02 1.79070249e-01
1.81887925e-01 -1.12084337e-01 1.43701360e-01 1.13843702e-01
9.46924925e-01 1.05031097e+00 3.07308793e-01 2.27072418e-01
3.64004344e-01 -8.45007673e-02 6.88088298e-01 4.85933840e-01
-2.73281246e-01 9.75390434e-01 2.42289200e-01 -6.25746489e-01
-1.38468182e+00 -1.20114195e+00 3.74041378e-01 7.94179738e-01
2.34431118e-01 -5.53261042e-01 -6.92082644e-01 -3.29495043e-01
-3.07703942e-01 8.35997045e-01 -7.37803817e-01 -2.73660868e-01
-7.10555315e-01 -1.00794146e-02 2.57333547e-01 8.59611094e-01
5.98927677e-01 -1.35347509e+00 -9.28650856e-01 4.55745161e-01
-5.90032041e-01 -1.12593269e+00 -5.87893546e-01 -5.91749609e-01
-6.32457137e-01 -8.29902172e-01 -1.03888655e+00 -6.17928863e-01
7.42821395e-02 -2.25240905e-02 1.18270481e+00 -4.06689435e-01
-1.87410921e-01 3.34797919e-01 -5.43789208e-01 -1.05344735e-01
-7.32000291e-01 -3.86316091e-01 3.71506102e-02 2.79869676e-01
6.33528456e-02 -8.52316737e-01 -9.89751279e-01 3.18396151e-01
-1.05161726e+00 5.89037955e-01 4.48722064e-01 7.53402948e-01
8.25571001e-01 -3.36500764e-01 -2.97501739e-02 -2.03634545e-01
6.40709251e-02 -3.57372403e-01 -2.11383328e-01 1.21343367e-01
-6.60143374e-03 -6.55991063e-02 4.91652757e-01 -6.58524036e-01
-1.19521475e+00 5.14509022e-01 -1.65110156e-01 -1.25990260e+00
1.01219602e-01 3.62388371e-03 -1.80065483e-01 3.86726588e-01
5.98213851e-01 6.72588646e-01 1.19275607e-01 -2.79503733e-01
5.79518020e-01 2.90004313e-01 9.17227328e-01 -1.56784102e-01
4.16530281e-01 6.34227991e-01 9.54924151e-02 -7.18982875e-01
-3.80645603e-01 -1.12572499e-02 -9.00848091e-01 -6.57522142e-01
1.37081146e+00 -1.08559763e+00 -9.71417427e-01 6.73116088e-01
-1.18451250e+00 -6.93781257e-01 -3.34341586e-01 6.61611736e-01
-1.25923789e+00 3.47111166e-01 -7.97179222e-01 -6.51975930e-01
5.08198924e-02 -1.26362574e+00 1.41057229e+00 -6.04834370e-02
-3.38805020e-01 -7.79837430e-01 3.15320551e-01 3.33995610e-01
9.39456932e-03 8.00200939e-01 1.51679277e-01 6.98017180e-02
-1.00204587e+00 9.95974615e-03 4.51882452e-01 2.76916206e-01
-9.95606855e-02 2.71579951e-01 -5.29455006e-01 -1.28702566e-01
-3.05140078e-01 -2.86845952e-01 7.86722124e-01 8.57771218e-01
1.14647210e+00 -3.61067176e-01 -1.33946940e-01 7.21870899e-01
1.12381041e+00 3.24987978e-01 1.13066089e+00 2.36616418e-01
9.28153694e-01 1.70941502e-01 7.70999908e-01 7.07557023e-01
3.92491847e-01 9.79474485e-01 5.91912687e-01 2.42892087e-01
-3.14598233e-01 -8.43401730e-01 5.92338264e-01 7.90702641e-01
-4.11802769e-01 -4.35169101e-01 -5.84962130e-01 5.95613122e-01
-2.19517493e+00 -1.65168011e+00 5.44558018e-02 1.58919513e+00
4.63633180e-01 -1.81247428e-01 4.27018374e-01 -8.67161825e-02
4.36167359e-01 6.21962309e-01 -6.13176763e-01 -3.07622373e-01
-3.85669991e-02 -1.38504744e-01 1.08209692e-01 4.03467476e-01
-9.40442681e-01 1.15468717e+00 6.70980263e+00 9.85247374e-01
-1.08385170e+00 5.60303293e-02 3.72578800e-01 -3.69146347e-01
-3.90571445e-01 -1.71111494e-01 -4.00216520e-01 5.29186785e-01
7.76250005e-01 -7.69313872e-02 2.84657180e-01 1.05532920e+00
4.01874393e-01 5.68505265e-02 -1.10447919e+00 1.01671207e+00
2.78992057e-01 -1.60490012e+00 4.11564082e-01 1.15148827e-01
1.07830667e+00 -3.34798306e-01 1.83018178e-01 4.55822438e-01
4.59283054e-01 -1.06776845e+00 1.12563741e+00 8.91061246e-01
7.49743223e-01 -7.85774052e-01 2.99317628e-01 5.76013923e-01
-1.31747043e+00 -1.24339066e-01 -2.30451912e-01 -1.39719412e-01
8.53475988e-01 -1.69954985e-01 -3.87473732e-01 6.26684308e-01
4.99600351e-01 1.20417559e+00 -2.15050146e-01 6.39565051e-01
-2.46363625e-01 2.96074510e-01 7.23867714e-02 1.14693724e-01
4.30159748e-01 -2.82968469e-02 6.68646038e-01 8.73132765e-01
5.11613488e-01 1.76561683e-01 3.67416382e-01 7.83844829e-01
3.77308697e-01 -2.72268832e-01 -6.21527314e-01 -2.58801162e-01
1.92836195e-01 6.96298301e-01 -3.55434746e-01 -4.87389028e-01
-1.57150596e-01 1.54830110e+00 -6.10238761e-02 2.64388978e-01
-1.16212809e+00 3.39412838e-01 7.37656116e-01 4.09021340e-02
7.44800806e-01 -3.73169929e-01 2.10614979e-01 -1.26224864e+00
-1.63611427e-01 -9.68656600e-01 8.11175406e-02 -1.24553108e+00
-7.31510818e-01 7.16035128e-01 1.88158303e-01 -1.94044638e+00
-1.06738091e+00 -3.85919452e-01 -4.19342518e-01 3.14575106e-01
-7.11798251e-01 -1.50336862e+00 -3.03378582e-01 7.29078889e-01
1.00703347e+00 -3.20557028e-01 6.00382328e-01 1.40383020e-01
-2.62189895e-01 3.29406708e-01 -6.11706860e-02 3.22010852e-02
3.21278423e-01 -9.20573115e-01 6.80655062e-01 9.50937152e-01
4.00409162e-01 1.24204330e-01 9.40764248e-01 -8.66252601e-01
-1.18849134e+00 -1.22026396e+00 6.19541764e-01 -6.25267506e-01
3.20710272e-01 -5.76963127e-02 -5.43387830e-01 9.02884603e-01
3.84541780e-01 -4.15931121e-02 4.56656188e-01 -6.96498632e-01
4.00046632e-02 2.75200635e-01 -6.56131506e-01 9.84178722e-01
1.45746040e+00 -5.01139402e-01 -4.84387577e-01 1.26702070e-01
7.05728889e-01 -7.28650868e-01 -9.34661210e-01 3.54111135e-01
9.50024068e-01 -1.24229157e+00 1.17094851e+00 -5.45496941e-01
1.22312224e+00 -3.87499779e-01 -2.77306914e-01 -1.24511552e+00
-6.38493001e-01 -7.23127723e-01 -5.73980927e-01 6.41273260e-01
-1.73367277e-01 3.52931887e-01 8.81160736e-01 3.52703512e-01
-1.42454192e-01 -7.82376885e-01 -7.75713265e-01 -8.92626524e-01
-8.57076719e-02 -5.70691228e-01 6.19938135e-01 5.88469446e-01
-3.04573625e-01 2.33110562e-02 -1.33493102e+00 -1.36502907e-01
3.98953855e-01 1.28728315e-01 1.12816143e+00 -6.61218226e-01
-7.21757829e-01 -2.85923094e-01 -7.30001271e-01 -1.44852316e+00
2.50745207e-01 -5.10042787e-01 9.55553949e-02 -1.70688748e+00
6.56420663e-02 6.50577605e-01 7.98101723e-02 2.68694937e-01
-7.89806619e-02 3.45323920e-01 6.45421803e-01 2.28799611e-01
-7.67320216e-01 1.24324024e+00 1.82470179e+00 -1.84368908e-01
-2.85252362e-01 -2.22445101e-01 9.52590182e-02 6.59230113e-01
4.11002100e-01 -1.93384334e-01 -7.13698030e-01 -3.23590636e-01
-9.91343409e-02 7.96889424e-01 5.00196278e-01 -1.36439073e+00
2.35064432e-01 -4.96323436e-01 6.47116363e-01 -9.52286899e-01
8.19868803e-01 -6.23207808e-01 9.01693761e-01 7.00990081e-01
-3.62517774e-01 4.75695953e-02 -7.21426634e-03 8.07854652e-01
-3.42688411e-01 4.22238052e-01 6.78011775e-01 -2.59980351e-01
-1.23033106e+00 6.46094203e-01 -5.96760690e-01 -2.18779907e-01
1.22136569e+00 -5.44771612e-01 1.86894372e-01 -8.26289296e-01
-9.99014735e-01 8.60459730e-02 5.34925342e-01 7.79850721e-01
7.24954009e-01 -1.85499716e+00 -7.85562694e-01 1.36194184e-01
7.06717148e-02 -2.22763464e-01 9.65871751e-01 5.37506223e-01
-7.19851494e-01 3.16177040e-01 -8.03174198e-01 -8.24145675e-01
-1.25209928e+00 7.28523076e-01 3.15089434e-01 -2.09637538e-01
-9.20700371e-01 8.10374737e-01 4.37913895e-01 1.91515740e-02
-7.21077546e-02 -1.15317263e-01 -2.78206855e-01 -2.09578335e-01
6.47614896e-01 3.40501428e-01 -7.02375710e-01 -1.08381534e+00
-8.54956880e-02 4.67883736e-01 3.42185199e-01 -5.01230955e-01
1.25617313e+00 -2.72792354e-02 4.23591584e-01 4.51542646e-01
1.14981914e+00 -3.91178697e-01 -2.05324888e+00 9.47530344e-02
-7.10138023e-01 -7.03864455e-01 -1.66085199e-01 -4.52546716e-01
-1.09893227e+00 7.78525233e-01 4.02499348e-01 -4.35983360e-01
1.03014493e+00 -3.06544274e-01 1.07402790e+00 3.03433836e-02
5.69048524e-01 -1.16449606e+00 6.66871548e-01 3.77241731e-01
1.21301913e+00 -8.65345418e-01 -1.41442746e-01 -1.58612803e-01
-1.31424344e+00 9.06668246e-01 7.33290672e-01 -2.61227757e-01
4.56339896e-01 5.85125796e-02 -7.36636072e-02 1.58938244e-01
-9.40690160e-01 -1.12541452e-01 5.29014468e-01 7.41603911e-01
1.49639741e-01 2.81715393e-02 -2.64854848e-01 4.56193626e-01
-2.49167293e-01 4.52128440e-01 4.03563797e-01 8.05804789e-01
-1.64059475e-01 -1.05405796e+00 4.25423408e-04 -1.00619800e-01
-1.12484477e-01 3.26724231e-01 -2.32101172e-01 7.54226625e-01
2.03053758e-01 6.54482007e-01 1.98256895e-01 -7.66816914e-01
1.41360193e-01 -1.05188407e-01 6.45845234e-01 -1.95744142e-01
-2.86938488e-01 2.85294682e-01 -1.38514623e-01 -1.22282529e+00
-5.92232883e-01 -7.01335669e-01 -1.01471007e+00 -3.58433455e-01
4.13758695e-01 -2.65636355e-01 2.06426829e-01 7.15673745e-01
5.35312235e-01 9.09488678e-01 4.29875016e-01 -1.40834224e+00
-3.20393592e-01 -8.28274369e-01 -7.04275727e-01 6.87404811e-01
2.67508030e-01 -7.16558874e-01 1.22426093e-01 5.31224251e-01] | [7.3759236335754395, -0.07743725180625916] |
e6897efb-9fab-4524-aeae-735267de6e4c | leveraging-unsupervised-and-weakly-supervised | 2203.13339 | null | https://arxiv.org/abs/2203.13339v2 | https://arxiv.org/pdf/2203.13339v2.pdf | Leveraging unsupervised and weakly-supervised data to improve direct speech-to-speech translation | End-to-end speech-to-speech translation (S2ST) without relying on intermediate text representations is a rapidly emerging frontier of research. Recent works have demonstrated that the performance of such direct S2ST systems is approaching that of conventional cascade S2ST when trained on comparable datasets. However, in practice, the performance of direct S2ST is bounded by the availability of paired S2ST training data. In this work, we explore multiple approaches for leveraging much more widely available unsupervised and weakly-supervised speech and text data to improve the performance of direct S2ST based on Translatotron 2. With our most effective approaches, the average translation quality of direct S2ST on 21 language pairs on the CVSS-C corpus is improved by +13.6 BLEU (or +113% relatively), as compared to the previous state-of-the-art trained without additional data. The improvements on low-resource language are even more significant (+398% relatively on average). Our comparative studies suggest future research directions for S2ST and speech representation learning. | ['Nobuyuki Morioka', 'Alexis Conneau', 'Yu Zhang', 'Colin Cherry', 'Ankur Bapna', 'Yifan Ding', 'Ye Jia'] | 2022-03-24 | null | null | null | null | ['speech-to-speech-translation'] | ['speech'] | [ 2.73363650e-01 2.02143028e-01 -4.08301055e-01 -4.03477311e-01
-1.69314790e+00 -5.72020233e-01 7.70006001e-01 -2.20447391e-01
-3.58110040e-01 8.20033073e-01 5.17370582e-01 -7.55467892e-01
6.47291541e-01 -1.73954576e-01 -7.99348474e-01 -4.17382240e-01
4.43362594e-01 7.12685287e-01 1.01278849e-01 -7.94273496e-01
-1.13597348e-01 4.07167710e-02 -7.99804330e-01 5.68412364e-01
1.00415170e+00 7.37559080e-01 4.31851268e-01 6.83680177e-01
-2.06450358e-01 4.95766610e-01 -6.78189099e-01 -5.35218537e-01
1.35890111e-01 -8.45579684e-01 -8.05101633e-01 -3.15530747e-01
3.76904339e-01 -2.64292002e-01 -3.25167120e-01 7.24407434e-01
8.39732826e-01 7.45048299e-02 5.76578736e-01 -6.89463377e-01
-1.09010744e+00 1.10650229e+00 -2.84153223e-01 3.44100535e-01
2.54500955e-01 -1.11504523e-02 1.04314113e+00 -1.41509879e+00
5.85990608e-01 1.26110947e+00 4.37857240e-01 7.24814713e-01
-1.12639236e+00 -7.65314639e-01 3.03993039e-02 2.17355043e-03
-1.29728854e+00 -1.01658046e+00 3.83982509e-01 -1.15786806e-01
1.31000161e+00 2.30004862e-01 1.43384531e-01 1.54712594e+00
-1.55320048e-01 1.15601540e+00 1.15121078e+00 -8.25617969e-01
-9.12119299e-02 3.79289627e-01 -1.12878978e-01 4.65951622e-01
-1.51145369e-01 1.78071678e-01 -8.71955633e-01 1.54549658e-01
5.67160785e-01 -4.17784542e-01 -2.28594035e-01 4.39277470e-01
-1.59309673e+00 7.91720569e-01 3.28314155e-01 5.28659821e-01
-1.43082142e-01 -1.45968897e-02 6.45524502e-01 7.99361169e-01
9.87309873e-01 4.33774024e-01 -8.17636013e-01 -4.34695393e-01
-1.12925625e+00 -1.73110515e-01 6.98440492e-01 1.32456279e+00
3.70225608e-01 3.36416811e-01 -3.17676485e-01 9.99918044e-01
1.87380388e-01 9.26551402e-01 8.09172511e-01 -4.37312126e-01
1.31166983e+00 3.01608890e-01 -6.90480769e-02 -2.52619624e-01
1.57358378e-01 -7.09319174e-01 -8.71742547e-01 -4.32483286e-01
2.54836291e-01 -4.23118502e-01 -8.88002455e-01 1.49098182e+00
-1.63393542e-01 -1.73052289e-02 4.78435814e-01 8.27610314e-01
1.02163565e+00 1.22673106e+00 -2.63631433e-01 -4.39188600e-01
8.60437453e-01 -1.55742764e+00 -8.93146694e-01 -2.96686590e-01
9.21618283e-01 -1.22378802e+00 1.31238651e+00 4.42963559e-03
-1.24529231e+00 -6.27699494e-01 -8.95898938e-01 -1.11905053e-01
-7.05622062e-02 6.47630751e-01 3.49272847e-01 5.65172017e-01
-1.45687389e+00 4.44113702e-01 -8.51205647e-01 -6.20618582e-01
-5.03791720e-02 3.93371344e-01 -2.89265871e-01 7.59546412e-03
-1.38822532e+00 1.11526287e+00 -2.66665723e-02 1.11431397e-01
-9.44471121e-01 -5.70852220e-01 -6.35240078e-01 -1.19742930e-01
2.13789806e-01 -5.44723153e-01 1.74136198e+00 -1.01881421e+00
-2.22798657e+00 5.86918592e-01 -5.43679178e-01 -4.51504201e-01
4.99889344e-01 -4.63974386e-01 -5.30987561e-01 4.72610369e-02
5.02818748e-02 5.28080165e-01 6.42710567e-01 -9.28645790e-01
-3.81694615e-01 -1.63541064e-01 -2.99984097e-01 6.25641346e-01
-5.04192412e-01 5.80072105e-01 -4.25998896e-01 -7.97407508e-01
9.89624783e-02 -1.02803457e+00 -2.72828132e-01 -3.58271718e-01
-3.99435967e-01 -4.16729361e-01 5.97606182e-01 -9.52245414e-01
1.13220012e+00 -1.90177488e+00 3.61354738e-01 -4.24228311e-01
-4.40699428e-01 7.06682146e-01 -4.92974907e-01 8.93514514e-01
1.40178248e-01 2.61193067e-01 -6.32750317e-02 -6.26499414e-01
-2.56152570e-01 -1.34477794e-01 -5.38189888e-01 3.60913463e-02
3.28467458e-01 1.12925398e+00 -8.41175556e-01 -1.74672142e-01
-1.35365054e-02 3.98126096e-01 -2.35997260e-01 3.81883740e-01
-6.62781596e-02 6.76819265e-01 -4.28418010e-01 4.15426344e-01
1.66342214e-01 -1.47400245e-01 1.40688181e-01 3.93078804e-01
-1.10377118e-01 1.13860977e+00 -3.22314113e-01 2.05285430e+00
-8.79284382e-01 8.58919442e-01 -2.58866459e-01 -8.70620370e-01
1.23279881e+00 7.98765838e-01 7.59890676e-02 -6.70021713e-01
-1.06283881e-01 6.93979502e-01 1.22654624e-02 -1.59461126e-01
4.81428117e-01 -1.51473150e-01 5.11073507e-02 4.75954235e-01
2.69435138e-01 -4.05141383e-01 -2.15813383e-01 1.89770505e-01
8.38078797e-01 4.44976427e-02 5.98026402e-02 -2.19863281e-01
5.27884483e-01 4.81506698e-02 2.88890004e-01 4.67257231e-01
-1.61410850e-02 6.17887557e-01 5.96328825e-02 -6.73593804e-02
-1.14065790e+00 -9.57885742e-01 3.12944204e-01 1.21643615e+00
-2.23332420e-01 -3.41844022e-01 -1.00544620e+00 -8.46297324e-01
-4.39676017e-01 8.06178093e-01 -1.08666800e-01 -2.50412226e-02
-7.77744770e-01 -4.77951139e-01 9.05625403e-01 6.48645759e-01
4.52389091e-01 -9.15179670e-01 4.36157733e-01 3.99626583e-01
-6.52017117e-01 -1.38192165e+00 -8.46539259e-01 1.21216476e-02
-1.19425464e+00 -1.91865250e-01 -1.14696121e+00 -1.02977014e+00
4.57407713e-01 5.44194877e-01 1.06379414e+00 -1.20024793e-01
5.29029071e-01 -3.77240300e-01 -7.21434295e-01 -1.34950548e-01
-8.61514926e-01 6.57544494e-01 2.94663578e-01 -2.64692873e-01
2.15918571e-01 -3.33314687e-01 -3.25455815e-01 3.80344838e-01
-2.24690959e-01 1.80315688e-01 9.10120368e-01 1.08106136e+00
2.71860331e-01 -6.10777736e-01 9.08034623e-01 -5.83810866e-01
4.96686667e-01 -4.17268962e-01 -3.83973420e-01 2.43228287e-01
-7.26256907e-01 -2.51527317e-02 7.42631555e-01 -4.46838796e-01
-1.25648749e+00 -2.61792272e-01 -3.07837009e-01 -3.08371186e-01
1.48787901e-01 5.87750912e-01 1.55370191e-01 2.85593957e-01
7.67421424e-01 6.00818515e-01 -2.97304895e-02 -6.75445557e-01
4.76701975e-01 1.40486479e+00 2.35643253e-01 -4.76019174e-01
8.28459859e-01 -1.20848119e-01 -7.03853369e-01 -7.50071704e-01
-8.08200836e-01 -4.75059986e-01 -6.06795192e-01 2.16572970e-01
5.58256745e-01 -1.29920506e+00 -1.12078682e-01 2.73936033e-01
-1.41198671e+00 -5.27325690e-01 7.24885166e-02 8.29709470e-01
-4.73539352e-01 2.13051528e-01 -8.63265634e-01 -7.62867808e-01
-7.62824178e-01 -1.35102355e+00 1.27987707e+00 -3.82763296e-01
-1.37435809e-01 -8.33714902e-01 9.67098475e-02 9.94050026e-01
5.25322735e-01 -5.23359060e-01 7.14681745e-01 -9.12482023e-01
-3.44396949e-01 -4.70775068e-02 -1.37730883e-02 7.28146613e-01
2.54179895e-01 -1.77292451e-01 -8.46563756e-01 -5.07529438e-01
-1.86528161e-01 -4.01689202e-01 6.81762338e-01 1.30673558e-01
5.45270383e-01 -4.05534029e-01 -6.35796711e-02 3.10695350e-01
9.94518518e-01 1.52039602e-01 3.87857765e-01 5.51999360e-02
5.71551979e-01 5.75691164e-01 7.03932166e-01 -6.51165247e-02
4.37549770e-01 1.05861235e+00 -1.47874609e-01 -2.55501091e-01
-6.64004505e-01 -5.54471076e-01 1.25511599e+00 1.94918585e+00
-7.54978955e-02 -4.13821429e-01 -9.37966287e-01 6.45783246e-01
-1.69100583e+00 -6.51009262e-01 -3.71829927e-01 2.05556560e+00
1.14274287e+00 -1.20276157e-02 -4.17700224e-02 -6.61335588e-02
1.01522756e+00 2.23182011e-02 -2.96607882e-01 -6.86258137e-01
-7.06191286e-02 2.74417371e-01 3.20344895e-01 4.92732286e-01
-6.69129372e-01 1.69120240e+00 7.02726936e+00 1.00878990e+00
-1.15769327e+00 4.67063934e-01 6.78252578e-01 -8.53530914e-02
-2.58407682e-01 5.06680012e-02 -1.08358896e+00 4.10291731e-01
1.61061788e+00 -1.92188263e-01 5.45899093e-01 6.72845244e-01
4.20510262e-01 4.68215793e-01 -1.01807523e+00 8.26957285e-01
1.52651221e-01 -1.39501953e+00 1.22964144e-01 -1.52735755e-01
1.09914565e+00 5.35038173e-01 1.79131508e-01 5.96091688e-01
4.05989081e-01 -9.15226519e-01 7.69071579e-01 -1.92461401e-01
1.30152905e+00 -5.84448516e-01 8.49076807e-01 4.49882805e-01
-9.73332763e-01 2.34271035e-01 -1.49064690e-01 -1.11831941e-01
2.03304082e-01 4.11343932e-01 -1.34583199e+00 8.90912235e-01
4.84705359e-01 7.56737947e-01 -2.26535454e-01 3.27798337e-01
-5.95173240e-01 1.29481947e+00 -1.84476569e-01 -4.24282938e-01
5.30940115e-01 -2.82591209e-02 3.58669162e-01 1.40401745e+00
5.38147807e-01 -1.58714756e-01 1.93323884e-02 3.40058714e-01
-4.81081486e-01 5.64074516e-01 -6.73040748e-01 -3.87450963e-01
6.39414966e-01 8.37029994e-01 -1.74771965e-01 -6.18766904e-01
-6.17107153e-01 1.23611605e+00 3.72324139e-01 4.39530045e-01
-4.53113109e-01 -1.54648036e-01 5.79964399e-01 -1.71122886e-02
2.65931129e-01 -4.56399113e-01 -2.92096496e-01 -1.29717445e+00
1.33285195e-01 -1.14884794e+00 -4.55259383e-02 -6.83743596e-01
-1.20393360e+00 1.05693185e+00 -4.36952382e-01 -1.37020290e+00
-5.57691514e-01 -4.33384240e-01 -3.91531944e-01 1.08093345e+00
-1.55799758e+00 -1.49075842e+00 3.98967594e-01 4.73437697e-01
1.29833114e+00 -6.58164561e-01 9.56931949e-01 1.79458052e-01
-4.42959368e-01 9.85392630e-01 6.02257192e-01 1.35153875e-01
1.00552392e+00 -8.98930967e-01 1.12390244e+00 9.06031370e-01
4.58882123e-01 6.60385072e-01 3.67380053e-01 -5.47533393e-01
-1.47353184e+00 -1.08575988e+00 1.49709892e+00 -5.70503175e-01
7.97488928e-01 -6.44154310e-01 -6.62507355e-01 7.63542295e-01
5.79936802e-01 -8.65064040e-02 5.21716475e-01 3.07530463e-01
-4.12601084e-01 -1.73480839e-01 -7.42107809e-01 8.51440310e-01
1.28307664e+00 -8.58817935e-01 -7.24878073e-01 4.70778823e-01
1.30247796e+00 -3.64461899e-01 -8.63910735e-01 3.80738497e-01
2.07480595e-01 -3.45489234e-01 7.21405625e-01 -6.74286544e-01
4.91395503e-01 -8.88209343e-02 -4.01530504e-01 -1.74214351e+00
-3.21426950e-02 -1.16373992e+00 6.51806965e-02 1.20604026e+00
9.67014194e-01 -5.69678307e-01 4.10460532e-01 -1.83794424e-01
-6.85208559e-01 -7.31405139e-01 -1.03137505e+00 -1.16556418e+00
5.51364839e-01 -2.16938436e-01 4.79138136e-01 8.83003831e-01
2.32142463e-01 8.24142039e-01 -4.60726857e-01 4.41691354e-02
4.19901945e-02 -1.96064059e-02 8.39197338e-01 -6.79101527e-01
-3.44095856e-01 -2.75340885e-01 -1.03218690e-01 -1.60020065e+00
3.38541895e-01 -1.26359916e+00 2.47417778e-01 -1.46773350e+00
1.51362598e-01 -3.60779554e-01 -3.12991768e-01 4.30552214e-01
-3.31424922e-01 2.85411775e-01 1.08673178e-01 4.39113468e-01
-1.65051624e-01 9.65985835e-01 1.51100934e+00 -1.81779563e-01
-2.22523421e-01 1.97245087e-02 -6.79287076e-01 2.84076184e-01
8.88739347e-01 -3.34793717e-01 -5.36727250e-01 -9.62891698e-01
-1.77703500e-01 3.39513838e-01 -3.16394240e-01 -5.35254121e-01
1.27914503e-01 -1.52131720e-02 -2.40354687e-02 -6.70024633e-01
4.30984706e-01 -2.87426442e-01 -3.16098690e-01 2.61639595e-01
-6.34805322e-01 2.10335925e-01 4.55644615e-02 3.26786011e-01
-4.80953753e-01 9.25034657e-02 4.39814806e-01 -5.57240285e-02
-2.87475735e-01 1.30940154e-01 -5.54089069e-01 1.14936583e-01
5.52492619e-01 3.39629292e-01 -3.13156277e-01 -6.85249031e-01
-3.43858182e-01 -1.62761837e-01 9.43046883e-02 8.12514901e-01
6.63444757e-01 -1.43452680e+00 -1.29270399e+00 6.50905818e-02
1.66030422e-01 -2.66187489e-01 -3.58356178e-01 8.67124557e-01
-8.05454031e-02 6.78747296e-01 2.21216291e-01 -6.24750376e-01
-1.14502132e+00 1.15852937e-01 -7.74846300e-02 -2.94959426e-01
-5.95439553e-01 1.00266302e+00 -7.57928416e-02 -7.44050145e-01
-6.29620552e-02 -2.63777554e-01 1.86974108e-01 -2.51573414e-01
3.03943366e-01 2.84697950e-01 5.31353116e-01 -7.23874509e-01
-3.83659512e-01 3.06392372e-01 -4.88884985e-01 -7.46715248e-01
1.35248721e+00 -1.94719940e-01 9.78657827e-02 5.68160892e-01
1.18722880e+00 2.55384315e-02 -9.80885863e-01 -5.41344523e-01
9.41498578e-02 -2.43795559e-01 8.23389515e-02 -1.14532101e+00
-9.98097897e-01 1.20650303e+00 2.81052291e-01 -3.07821095e-01
9.20850277e-01 1.56058103e-01 1.18060660e+00 7.24064589e-01
4.95706320e-01 -1.04857051e+00 1.53317869e-01 8.53634775e-01
1.21974790e+00 -1.45178390e+00 -4.67921555e-01 -5.18148363e-01
-9.33128893e-01 1.03388202e+00 2.87634164e-01 3.37236822e-02
1.73127562e-01 1.37036011e-01 5.70439339e-01 3.67977798e-01
-9.82171297e-01 -7.98979178e-02 2.96159834e-01 3.92744064e-01
9.43673432e-01 1.68418348e-01 -2.82328695e-01 3.54299009e-01
-4.70614105e-01 -2.68693238e-01 3.46166164e-01 5.11027813e-01
-2.21548736e-01 -1.41646636e+00 -1.14168108e-01 1.16305478e-01
-5.24183333e-01 -7.63875484e-01 -5.54205239e-01 4.59505737e-01
-5.48816681e-01 1.52843058e+00 -2.59667844e-01 -4.99682814e-01
2.05042437e-01 1.69233769e-01 3.24808925e-01 -9.24683630e-01
-8.14183414e-01 1.33350149e-01 5.67553759e-01 -4.03013289e-01
-2.03245223e-01 -6.46740079e-01 -1.09731400e+00 -3.16156805e-01
-6.34064794e-01 2.00104877e-01 1.05421579e+00 9.10002649e-01
4.97156024e-01 3.00628334e-01 1.12895417e+00 -7.27365553e-01
-8.46095085e-01 -1.44334292e+00 3.62479128e-02 -1.33022899e-02
3.69530499e-01 -2.20076382e-01 -2.72962213e-01 1.43697932e-01] | [14.52054500579834, 7.1479902267456055] |
0dfe5a17-348a-4b10-94e2-c37a9b4815ce | supervised-hierarchical-clustering-using | 2302.12716 | null | https://arxiv.org/abs/2302.12716v1 | https://arxiv.org/pdf/2302.12716v1.pdf | Supervised Hierarchical Clustering using Graph Neural Networks for Speaker Diarization | Conventional methods for speaker diarization involve windowing an audio file into short segments to extract speaker embeddings, followed by an unsupervised clustering of the embeddings. This multi-step approach generates speaker assignments for each segment. In this paper, we propose a novel Supervised HierArchical gRaph Clustering algorithm (SHARC) for speaker diarization where we introduce a hierarchical structure using Graph Neural Network (GNN) to perform supervised clustering. The supervision allows the model to update the representations and directly improve the clustering performance, thus enabling a single-step approach for diarization. In the proposed work, the input segment embeddings are treated as nodes of a graph with the edge weights corresponding to the similarity scores between the nodes. We also propose an approach to jointly update the embedding extractor and the GNN model to perform end-to-end speaker diarization (E2E-SHARC). During inference, the hierarchical clustering is performed using node densities and edge existence probabilities to merge the segments until convergence. In the diarization experiments, we illustrate that the proposed E2E-SHARC approach achieves 53% and 44% relative improvements over the baseline systems on benchmark datasets like AMI and Voxconverse, respectively. | ['Sriram Ganapathy', 'Amrit Kaul', 'Prachi Singh'] | 2023-02-24 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.38209239e-01 3.62869084e-01 1.15667075e-01 -6.36500895e-01
-8.36086988e-01 -4.06252980e-01 2.67174274e-01 3.20792675e-01
-2.92230517e-01 -4.32947204e-02 4.91151839e-01 -1.55169874e-01
-3.59215848e-02 -5.76699376e-01 -3.68138969e-01 -9.04417694e-01
-4.48098660e-01 5.67047358e-01 2.29226068e-01 2.04218999e-01
6.98789256e-03 3.23243082e-01 -1.43034661e+00 -2.65665967e-02
1.01801419e+00 7.07104862e-01 -1.06991313e-01 9.64465380e-01
-1.41570687e-01 2.39332542e-01 -6.99106693e-01 -3.80987197e-01
5.26404679e-02 -6.82023227e-01 -6.49561465e-01 3.16762626e-01
2.15244159e-01 3.30640562e-02 -2.80168265e-01 8.89804125e-01
7.91425169e-01 4.44476455e-01 8.59870374e-01 -1.52163243e+00
-3.55872929e-01 1.37108433e+00 -6.44189358e-01 8.49037841e-02
1.75342590e-01 -2.55146563e-01 1.20695508e+00 -8.22589576e-01
1.89210042e-01 1.57240522e+00 5.04977882e-01 6.64427340e-01
-1.35557556e+00 -8.24995816e-01 3.82743031e-01 3.29948366e-01
-1.89869475e+00 -5.30450165e-01 1.17640448e+00 -4.01580751e-01
7.42982268e-01 2.36454323e-01 5.34276009e-01 6.86613739e-01
-2.98826993e-01 7.23794758e-01 2.66615123e-01 -5.41791320e-01
4.96813029e-01 6.73661903e-02 5.10436177e-01 5.01824856e-01
-1.13538466e-02 -2.60274082e-01 -7.67764330e-01 -1.93822622e-01
3.02231222e-01 -1.17913827e-01 -6.52553886e-02 -2.27378771e-01
-8.67376208e-01 9.61328208e-01 6.17556632e-01 2.00569391e-01
-3.47325295e-01 5.07179871e-02 3.90923113e-01 3.76176573e-02
3.91420543e-01 -1.19210668e-01 3.50610971e-01 1.50007963e-01
-1.15695763e+00 -7.40765780e-02 6.36235178e-01 8.72929215e-01
7.58169353e-01 2.60399640e-01 -1.09943964e-01 9.32644665e-01
8.34155500e-01 2.89125979e-01 6.86349452e-01 -5.82796633e-01
5.58976889e-01 5.06811202e-01 -3.69038939e-01 -1.15752327e+00
-1.90030992e-01 -4.10030186e-01 -1.03586364e+00 -2.62866795e-01
-1.06847622e-01 -3.29075694e-01 -1.04414499e+00 1.70701587e+00
6.04086280e-01 6.10622346e-01 1.48094684e-01 6.42035842e-01
9.58409190e-01 9.37631547e-01 1.82992756e-01 -8.48047063e-02
1.30003655e+00 -8.79827380e-01 -8.50081921e-01 -4.67088781e-02
3.99289191e-01 -5.14781475e-01 7.98818648e-01 2.48725563e-01
-8.83277118e-01 -6.92307234e-01 -1.21918082e+00 2.24563152e-01
-1.59401581e-01 -2.45438963e-02 5.58894388e-02 7.83202648e-01
-1.27764928e+00 2.16811851e-01 -1.03182030e+00 -2.98106134e-01
2.36845791e-01 6.40548766e-01 -4.95222285e-02 3.40333164e-01
-1.29839134e+00 1.74143404e-01 5.45664549e-01 1.90262496e-01
-9.12383676e-01 -4.99191105e-01 -9.79181230e-01 4.04441923e-01
-1.02792308e-01 -3.19484591e-01 8.65219116e-01 -6.19643152e-01
-1.71388519e+00 5.49431443e-01 -4.45034325e-01 -7.38906920e-01
2.79254671e-02 1.73448652e-01 -4.42654312e-01 2.77821183e-01
-2.19372138e-01 9.61028099e-01 9.92638409e-01 -1.40570128e+00
-7.19210863e-01 -5.24418592e-01 -4.54052448e-01 5.18694758e-01
-5.99591613e-01 -2.15636808e-02 -9.09668088e-01 -6.31019354e-01
4.29050684e-01 -8.65947306e-01 -2.61450391e-02 -6.74914360e-01
-7.50114143e-01 -6.54036760e-01 9.14092839e-01 -1.01759183e+00
1.86475468e+00 -2.58387089e+00 6.17668748e-01 6.63170636e-01
4.68419939e-01 -1.17331758e-01 2.92861778e-02 3.54822129e-01
1.69515591e-02 8.11103582e-02 -4.72394079e-01 -1.13196039e+00
3.18512142e-01 4.42050174e-02 1.15944661e-01 3.14128190e-01
-1.12292923e-01 4.33500260e-01 -4.16453719e-01 -8.71470869e-01
-3.56007330e-02 7.82047033e-01 -7.65613496e-01 4.04433846e-01
2.59178877e-01 5.33593334e-02 1.06374681e-01 2.49152228e-01
7.45165229e-01 1.85015559e-01 3.77646148e-01 -7.10678399e-02
3.00194710e-01 2.66871214e-01 -1.49931788e+00 1.46578979e+00
-2.90024281e-01 5.49538016e-01 2.00452894e-01 -1.05937672e+00
1.13462031e+00 3.66533428e-01 3.12241495e-01 1.79411508e-02
2.13439241e-01 -2.07946956e-01 -3.59995551e-02 -1.77517712e-01
3.74097347e-01 -1.87510028e-02 -1.89141452e-01 4.24092025e-01
1.84506208e-01 1.03886519e-02 -4.31808196e-02 6.42441034e-01
6.52875781e-01 -6.02595985e-01 -8.78000781e-02 -4.65686433e-02
7.58162856e-01 -5.36154091e-01 7.13049233e-01 3.53704691e-01
-2.98811108e-01 5.84821880e-01 3.87474746e-01 2.26345330e-01
-4.77977365e-01 -1.30650866e+00 1.59502596e-01 1.12791443e+00
4.06509787e-02 -5.87716579e-01 -1.27229631e+00 -7.59355664e-01
-6.79546371e-02 8.81720901e-01 -4.63381171e-01 -3.82567793e-01
-5.20118237e-01 -5.97782314e-01 6.98289454e-01 4.27583158e-01
4.17561769e-01 -8.19325864e-01 7.97529742e-02 1.97481290e-01
-4.43252355e-01 -9.21733797e-01 -1.09436262e+00 2.20462650e-01
-7.36158788e-01 -3.80482227e-01 -4.35590267e-01 -1.40266466e+00
7.64800549e-01 2.05391780e-01 6.03106141e-01 -1.83056772e-01
-1.73266605e-02 2.41130859e-01 -2.88708836e-01 -1.26357883e-01
-5.25753498e-01 4.37291265e-01 2.59354919e-01 5.69795966e-01
5.78243971e-01 -9.01285946e-01 -5.60411513e-01 2.85968989e-01
-8.15548003e-01 -2.32123613e-01 2.57775605e-01 4.63133276e-01
6.06236398e-01 3.04180384e-01 8.30598772e-01 -7.76613057e-01
6.00265861e-01 -3.08019787e-01 -4.84470278e-01 9.70462039e-02
-6.35644257e-01 1.34314671e-01 5.98767638e-01 -4.72461075e-01
-9.09640074e-01 3.15295368e-01 -2.05336392e-01 -4.95966494e-01
-8.81600156e-02 5.20899713e-01 -7.89747655e-01 3.50190371e-01
5.20781100e-01 3.58711630e-01 -1.70334559e-02 -4.38012272e-01
5.62091410e-01 1.23900449e+00 6.36044621e-01 -1.17873564e-01
9.33742464e-01 3.58905829e-02 -6.82812691e-01 -1.07208264e+00
-2.71714091e-01 -7.30642498e-01 -8.68479073e-01 -3.00781071e-01
1.21085954e+00 -9.68233466e-01 -6.47525430e-01 4.59503472e-01
-8.73737872e-01 -2.02033415e-01 -3.81305553e-02 6.77298248e-01
-1.54851466e-01 5.42223334e-01 -4.81269598e-01 -9.74759221e-01
-5.21209240e-01 -1.06537485e+00 9.96883571e-01 3.34824353e-01
-1.04173221e-01 -1.17494977e+00 1.55827120e-01 3.63597691e-01
1.52254850e-01 -5.16410246e-02 9.18741047e-01 -9.88785148e-01
-1.85143828e-01 -1.67256057e-01 1.56563729e-01 3.69809628e-01
3.49190116e-01 5.29999249e-02 -1.11551154e+00 -4.33950573e-01
-3.25531065e-01 2.75018483e-01 7.79968381e-01 4.06735092e-01
1.06990361e+00 -3.58813614e-01 -3.99637192e-01 5.99250436e-01
1.11048377e+00 3.06843191e-01 4.88196045e-01 -2.73773462e-01
1.03641856e+00 5.84218204e-01 5.29525913e-02 3.63932818e-01
6.79377437e-01 6.44163907e-01 8.97046402e-02 -3.34047303e-02
-3.63401681e-01 -4.44903314e-01 6.44159913e-01 1.67074466e+00
3.46022189e-01 -4.64060545e-01 -8.44108462e-01 7.39913404e-01
-1.58181393e+00 -8.46888423e-01 1.55974746e-01 2.37036800e+00
9.40038741e-01 3.07642758e-01 6.17554247e-01 4.17720258e-01
1.33408391e+00 1.64192751e-01 -4.46068555e-01 -3.80144536e-01
1.97337449e-01 -3.76065113e-02 1.34375423e-01 1.04348099e+00
-1.09157383e+00 1.08891690e+00 5.64963722e+00 5.94920933e-01
-9.91420746e-01 -5.05781136e-02 4.26418096e-01 9.87436101e-02
-2.84377784e-01 -2.02974170e-01 -1.04326308e+00 4.45769012e-01
1.26595449e+00 -2.79039621e-01 4.26374078e-01 6.16434336e-01
2.14731559e-01 3.67861509e-01 -1.07553303e+00 1.18359900e+00
4.94391620e-01 -7.40582347e-01 2.31111109e-01 -9.23281014e-02
4.10684049e-01 -2.66858280e-01 -1.77734923e-02 3.30228716e-01
5.52247226e-01 -6.33472383e-01 3.80013257e-01 9.28864703e-02
4.62931156e-01 -1.13832629e+00 5.14607966e-01 1.84660219e-02
-1.55099189e+00 6.74687624e-02 -2.60875463e-01 6.39575124e-01
1.52162999e-01 4.56662357e-01 -1.24210119e+00 5.75382650e-01
6.06108248e-01 5.98412395e-01 -6.57799482e-01 1.03328156e+00
-2.03840405e-01 1.22710156e+00 -4.86425191e-01 -7.09641865e-03
7.42179481e-03 -1.59621507e-01 6.38790548e-01 1.36401880e+00
1.26776516e-01 -2.03821883e-01 1.49071410e-01 4.25001800e-01
-2.84783095e-01 1.39822394e-01 -1.09422363e-01 7.47466758e-02
1.02071857e+00 1.17631483e+00 -8.01205218e-01 -5.39698839e-01
-2.34951265e-02 1.12760329e+00 3.24859977e-01 4.90217805e-01
-1.03250206e+00 -1.09231377e+00 4.89444524e-01 -6.42046100e-03
5.78772604e-01 -2.51289934e-01 -8.83806422e-02 -8.56749058e-01
-1.43145278e-01 -7.57038474e-01 7.40456104e-01 -1.36378706e-01
-1.13028574e+00 7.98186004e-01 1.58388708e-02 -1.00426555e+00
-2.75033832e-01 1.81879804e-01 -1.01800704e+00 7.04996228e-01
-1.23162925e+00 -9.22849119e-01 -3.29283208e-01 6.91740453e-01
4.97482866e-01 -1.67184919e-01 6.75068319e-01 3.64741415e-01
-9.88869965e-01 1.23339486e+00 7.91095793e-02 4.02071834e-01
7.64123082e-01 -1.47988737e+00 6.12514913e-01 1.07358181e+00
3.18112850e-01 5.02219737e-01 6.37096345e-01 -5.37771761e-01
-8.65831852e-01 -1.37458444e+00 1.02311659e+00 7.35762715e-02
5.05750954e-01 -9.71607506e-01 -1.01632011e+00 7.50126183e-01
4.43550855e-01 -4.86024112e-01 1.06795824e+00 2.46366069e-01
-3.55986059e-01 -3.88794601e-01 -8.90616119e-01 4.58949119e-01
7.65659392e-01 -5.58347404e-01 -6.33169472e-01 4.87115085e-02
9.71152127e-01 2.27084123e-02 -7.66636848e-01 -2.13365704e-01
2.53322423e-01 -5.12321055e-01 9.56514835e-01 -3.17563355e-01
-2.20878646e-01 -5.17505288e-01 -2.02163234e-01 -1.51808882e+00
-5.07488787e-01 -7.70038903e-01 -4.37329374e-02 1.83353817e+00
6.71862721e-01 -5.24132788e-01 6.40111327e-01 2.90394396e-01
-1.16065785e-01 -2.27546751e-01 -9.98673618e-01 -5.67701221e-01
-1.05134979e-01 -2.55720675e-01 8.08586240e-01 7.86221504e-01
1.77877337e-01 7.29724824e-01 -1.04214966e-01 8.21139872e-01
9.97436523e-01 1.45159438e-02 6.44440472e-01 -1.36941123e+00
-2.26921305e-01 -3.08204800e-01 -7.22810686e-01 -1.06195176e+00
2.94719726e-01 -1.23284924e+00 4.15281028e-01 -1.72783923e+00
-7.49447271e-02 -2.39043102e-01 -3.31576794e-01 2.65735626e-01
-5.07043421e-01 2.03210507e-02 1.22550279e-01 1.51064843e-01
-7.69597650e-01 7.17872262e-01 4.33778733e-01 -4.25748467e-01
-8.97275865e-01 1.31062105e-01 -6.23534799e-01 3.41924787e-01
9.21351135e-01 -6.45049155e-01 -6.09039009e-01 -3.27365130e-01
-5.27691603e-01 -1.59981161e-01 -4.66805547e-02 -1.14478719e+00
6.14774644e-01 4.49917912e-01 1.07732505e-01 -7.51162708e-01
1.16426252e-01 -6.73403621e-01 -1.32489964e-01 3.14806700e-01
-6.02362275e-01 -9.25402194e-02 3.76499840e-03 7.23331332e-01
-3.68962586e-01 -4.41958979e-02 8.52363467e-01 6.31720066e-01
-2.89327830e-01 3.08194757e-01 -5.66913784e-01 -8.13307986e-02
9.09794152e-01 -7.45535642e-02 3.74453366e-01 -5.94797373e-01
-1.27417994e+00 5.53156018e-01 -9.20575261e-02 3.72934073e-01
6.67595506e-01 -1.59016168e+00 -8.74336958e-01 2.99454778e-01
6.31514639e-02 2.10879400e-01 3.70256573e-01 5.46376944e-01
-2.35466719e-01 -2.74386443e-02 3.10125500e-01 -8.43835294e-01
-1.65310919e+00 5.11691570e-01 1.69666275e-01 -9.39616114e-02
-4.91091192e-01 1.14956927e+00 2.63394862e-01 -4.59619254e-01
9.21583474e-01 -3.34410101e-01 -4.35907334e-01 3.70019108e-01
4.03567135e-01 2.43205935e-01 7.18692690e-02 -7.28424311e-01
-4.95270610e-01 4.39711779e-01 -2.72187591e-01 -3.40418309e-01
1.26305223e+00 -5.46768785e-01 6.41273186e-02 5.46196222e-01
1.32781863e+00 3.55114788e-02 -1.02804732e+00 -3.84088725e-01
-7.19675645e-02 3.73301506e-02 1.20297089e-01 -2.79548794e-01
-1.15372550e+00 1.00044346e+00 7.74221241e-01 3.00053030e-01
1.14136112e+00 1.89399257e-01 8.32741141e-01 1.69976920e-01
-2.76871294e-01 -1.13618553e+00 -8.08232874e-02 4.86251079e-02
6.25154555e-01 -7.54135311e-01 -2.80576795e-01 -3.82749170e-01
-8.32866549e-01 7.67894030e-01 4.74326670e-01 9.17599574e-02
9.90524590e-01 1.34508386e-01 1.64381370e-01 -4.15373668e-02
-3.55287582e-01 -1.05484992e-01 3.26620281e-01 6.14982963e-01
2.74805933e-01 3.64919364e-01 5.31991199e-02 7.40622878e-01
-5.27579963e-01 -8.28387260e-01 2.07657695e-01 4.32382464e-01
-6.15387917e-01 -1.05287051e+00 -4.87615049e-01 6.66840225e-02
5.85178323e-02 -2.54531447e-02 -6.27352357e-01 3.60531926e-01
-1.58549577e-01 1.27431905e+00 2.68664390e-01 -7.70681083e-01
3.30131292e-01 4.30264860e-01 -5.84218614e-02 -5.80235958e-01
-5.73712587e-01 5.25909901e-01 -1.00664720e-01 2.61092577e-02
-1.77332357e-01 -7.00859845e-01 -1.73269999e+00 -1.32429719e-01
-4.85658854e-01 6.43075109e-01 5.88066459e-01 5.36981821e-01
6.32566512e-01 8.08150291e-01 1.19940078e+00 -7.26539671e-01
-2.54599780e-01 -1.13772058e+00 -7.37407804e-01 3.99214327e-01
1.94413245e-01 -4.32146370e-01 -7.87701845e-01 3.02483529e-01] | [14.420991897583008, 6.150274276733398] |
a787af7d-4dd5-411e-8f03-29cb551d7518 | multimodal-across-domains-gaze-target | 2208.10822 | null | https://arxiv.org/abs/2208.10822v1 | https://arxiv.org/pdf/2208.10822v1.pdf | Multimodal Across Domains Gaze Target Detection | This paper addresses the gaze target detection problem in single images captured from the third-person perspective. We present a multimodal deep architecture to infer where a person in a scene is looking. This spatial model is trained on the head images of the person-of- interest, scene and depth maps representing rich context information. Our model, unlike several prior art, do not require supervision of the gaze angles, do not rely on head orientation information and/or location of the eyes of person-of-interest. Extensive experiments demonstrate the stronger performance of our method on multiple benchmark datasets. We also investigated several variations of our method by altering joint-learning of multimodal data. Some variations outperform a few prior art as well. First time in this paper, we inspect domain adaption for gaze target detection, and we empower our multimodal network to effectively handle the domain gap across datasets. The code of the proposed method is available at https://github.com/francescotonini/multimodal-across-domains-gaze-target-detection. | ['Elisa Ricci', 'Cigdem Beyan', 'Francesco Tonini'] | 2022-08-23 | null | null | null | null | ['gaze-target-estimation', 'gaze-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.26073346e-01 -1.68863870e-02 1.51473470e-02 -6.12886965e-01
-6.80873334e-01 -7.77963817e-01 6.13565147e-01 -2.99421638e-01
-4.70049232e-01 4.59836543e-01 2.52109617e-01 1.98243663e-01
1.48025915e-01 -1.55336648e-01 -8.47065091e-01 -7.47157395e-01
3.34747136e-01 1.03171706e-01 1.13659292e-01 -1.47660166e-01
5.23179412e-01 1.48022935e-01 -1.74511135e+00 2.26618007e-01
5.38324714e-01 1.00786793e+00 2.35410318e-01 6.76473439e-01
4.26507115e-01 5.15128851e-01 -4.16431278e-01 -3.75213057e-01
8.85596946e-02 -3.82593900e-01 -6.76981688e-01 1.73297018e-01
1.06400144e+00 -4.04830486e-01 -3.02148998e-01 1.03584182e+00
6.62737072e-01 -5.60113275e-03 6.09359682e-01 -1.61737597e+00
-6.71799362e-01 -5.29838204e-02 -8.89360487e-01 2.53917903e-01
8.11953306e-01 3.44278783e-01 7.76438773e-01 -1.03173649e+00
4.23392922e-01 1.36695755e+00 3.47527385e-01 8.00884068e-01
-9.07226264e-01 -6.76536739e-01 3.66827458e-01 3.59221965e-01
-1.26888156e+00 -8.26782823e-01 7.97510207e-01 -4.04368401e-01
5.04283011e-01 1.76324368e-01 1.62373215e-01 1.66027701e+00
-2.76713192e-01 9.49406147e-01 1.13250673e+00 -5.66615641e-01
-1.74480215e-01 2.37331077e-01 2.21849918e-01 6.48898363e-01
7.16698319e-02 3.14359367e-02 -9.36591864e-01 1.20490696e-02
6.34902000e-01 2.22246960e-01 -5.71203232e-01 -5.77241719e-01
-1.41238475e+00 5.84030092e-01 7.50520408e-01 6.38959324e-03
-2.02741697e-01 4.98950183e-02 8.20731148e-02 -1.23036772e-01
3.42182964e-01 3.84214222e-02 -2.38514543e-01 -1.56735495e-01
-6.85132623e-01 1.85257331e-01 3.50623846e-01 1.17230022e+00
7.20004857e-01 -5.74480414e-01 -2.79985279e-01 5.42421520e-01
5.68839133e-01 7.07360566e-01 2.59283930e-01 -9.69261408e-01
6.36908591e-01 7.14749217e-01 3.77034783e-01 -7.34428883e-01
-5.00278831e-01 5.99070229e-02 -3.25667351e-01 1.26651391e-01
8.40571225e-01 -5.09209633e-01 -8.37216794e-01 1.99299526e+00
4.94387865e-01 7.13985274e-03 -2.63312161e-01 1.34997010e+00
9.31001186e-01 1.87759936e-01 5.59069738e-02 1.47419676e-01
1.72155237e+00 -1.07039368e+00 -7.24219143e-01 -4.51612145e-01
3.09001178e-01 -6.91962063e-01 1.13893187e+00 2.92615473e-01
-1.20466137e+00 -5.01986980e-01 -6.46629751e-01 -2.81492054e-01
-5.49858689e-01 4.56222296e-01 4.07820314e-01 6.12273812e-01
-1.27050722e+00 -1.41413450e-01 -6.13934875e-01 -1.00459862e+00
3.68710011e-01 3.74681324e-01 -5.08204222e-01 1.49220927e-02
-9.04526770e-01 8.10107708e-01 -3.97893004e-02 2.12513953e-01
-9.60475564e-01 -3.68186921e-01 -1.06733620e+00 -1.54425368e-01
4.24251467e-01 -9.45470273e-01 1.48974288e+00 -1.20727730e+00
-1.43778753e+00 1.14742136e+00 -9.65412140e-01 -5.55161014e-02
4.52989995e-01 -5.95074773e-01 -3.13934952e-01 3.45566362e-01
8.05018246e-02 1.01140690e+00 1.01608407e+00 -1.37114346e+00
-7.82743335e-01 -7.21331835e-01 3.03994566e-01 4.78255719e-01
-2.83139735e-01 4.20098245e-01 -6.34134114e-01 1.79751292e-02
-2.35532686e-01 -9.74640191e-01 3.10097575e-01 1.27424076e-01
-6.53464377e-01 -4.01223719e-01 7.96317279e-01 -3.60081017e-01
9.50532079e-01 -2.10459495e+00 1.87213838e-01 -1.80285424e-01
3.02889258e-01 -6.97490349e-02 -7.77438208e-02 4.54885811e-01
-1.64165318e-01 -1.84090689e-01 -9.92273390e-02 -7.41666019e-01
2.02422917e-01 -2.85324991e-01 -1.99290961e-01 8.12889814e-01
2.13851169e-01 9.24132288e-01 -7.05026031e-01 -4.80189234e-01
1.72452703e-01 7.66180336e-01 -2.89637715e-01 3.82460088e-01
-7.65704289e-02 7.47992754e-01 -1.66372001e-01 7.99755394e-01
7.67850399e-01 -4.82894033e-01 -1.53905228e-01 -2.97873497e-01
-1.14046104e-01 2.65787870e-01 -8.71496797e-01 1.96084917e+00
-1.69773668e-01 7.36971021e-01 1.39171705e-01 -3.83675307e-01
4.76651996e-01 1.96854904e-01 1.04336575e-01 -8.61796439e-01
2.46009409e-01 -1.73098266e-01 -9.60774720e-02 -7.61148512e-01
4.06043202e-01 3.76141369e-01 -9.48631838e-02 4.29265678e-01
1.18878230e-01 6.16142213e-01 1.95992380e-01 1.21446155e-01
5.95576525e-01 3.71116400e-01 1.14114843e-01 4.39113788e-02
6.67108357e-01 -2.62592047e-01 2.38195091e-01 5.07561445e-01
-5.73141754e-01 8.93159568e-01 5.49239337e-01 -1.02285646e-01
-6.69920146e-01 -1.01161194e+00 -1.65884659e-01 1.69082177e+00
3.65031600e-01 -2.01998591e-01 -8.81840587e-01 -7.49008954e-01
-1.84178978e-01 5.96400499e-01 -9.80736077e-01 1.76691756e-01
-3.07005107e-01 -3.90295744e-01 3.96724492e-01 3.97299379e-01
4.27256554e-01 -8.39907706e-01 -8.91485989e-01 -7.94842899e-01
-3.87657642e-01 -1.13374281e+00 -6.84952199e-01 -4.03578252e-01
-3.82893205e-01 -1.31158435e+00 -1.07006681e+00 -6.89045250e-01
6.93527997e-01 4.48286086e-01 9.66668665e-01 -3.38344991e-01
7.75683522e-02 9.35630739e-01 -1.44777611e-01 -5.69141805e-01
2.65354216e-01 1.59063578e-01 4.79972772e-02 4.04794276e-01
8.61175179e-01 -2.01136008e-01 -9.70124602e-01 5.05788982e-01
-2.84125566e-01 -7.17999339e-02 4.59146827e-01 4.95106399e-01
2.06755996e-01 -6.33724689e-01 2.42132425e-01 -6.33657455e-01
4.74603146e-01 -5.98423958e-01 -5.29622972e-01 2.72185624e-01
-7.31265619e-02 -3.54799211e-01 -1.87086225e-01 -3.06963444e-01
-1.17525101e+00 2.24425867e-01 3.78206760e-01 -5.78030765e-01
-8.65065157e-01 -4.16536294e-02 -3.45721841e-01 1.09152645e-01
6.20314598e-01 7.54477875e-03 7.82195702e-02 -4.27703291e-01
3.30350608e-01 6.77846491e-01 5.23159862e-01 -3.00579280e-01
4.44248796e-01 8.96592677e-01 -3.39561999e-02 -5.81058085e-01
-9.91842389e-01 -6.30033493e-01 -1.03371119e+00 -4.05969203e-01
9.50974286e-01 -1.20304942e+00 -1.28940415e+00 5.30014157e-01
-1.24680233e+00 -1.39295056e-01 5.18931210e-01 3.64562422e-01
-4.67014611e-01 2.30716288e-01 -1.94337025e-01 -9.46321309e-01
-1.28093913e-01 -9.96008277e-01 1.65735996e+00 6.30664289e-01
-1.16791673e-01 -1.04393899e+00 1.01429492e-01 4.80410188e-01
2.06912205e-01 2.82460749e-01 1.69807434e-01 -4.93457168e-01
-5.73302984e-01 -1.56271189e-01 -4.32668447e-01 -1.85110316e-01
1.06378555e-01 -1.40253112e-01 -1.54658771e+00 -2.79563963e-01
-2.39471972e-01 -3.69231433e-01 7.52153397e-01 5.49754560e-01
9.12081301e-01 -1.97992340e-01 -6.34143710e-01 6.89224541e-01
1.19770014e+00 -2.85591662e-01 4.58288968e-01 4.88745958e-01
9.93188560e-01 9.93971407e-01 7.65455484e-01 4.09145951e-01
9.56080556e-01 8.08007777e-01 8.32752764e-01 -2.29937240e-01
4.27840054e-02 5.59704239e-03 5.18038154e-01 -1.41454861e-01
-2.13132814e-01 -4.50777024e-01 -9.97859120e-01 7.32054591e-01
-1.91452992e+00 -9.44834352e-01 -3.65866423e-01 2.10906506e+00
5.66293776e-01 -3.43237460e-01 7.86894321e-01 -3.96564633e-01
9.81423318e-01 -6.95836032e-03 -7.07055569e-01 -8.32421929e-02
4.06549200e-02 -3.94689322e-01 3.43419224e-01 3.86048824e-01
-1.36184037e+00 8.47083092e-01 5.34621572e+00 1.82573940e-03
-1.14456403e+00 1.76267728e-01 2.27530420e-01 -6.57936156e-01
1.89224645e-01 -2.96039671e-01 -1.15147460e+00 5.66574156e-01
8.53550792e-01 2.58854479e-01 2.59859860e-01 6.77546501e-01
1.55190364e-01 -3.46708477e-01 -1.41882002e+00 1.24321783e+00
5.71051598e-01 -6.94682062e-01 -3.78267437e-01 2.80349106e-01
4.35487717e-01 2.04073936e-01 5.10141671e-01 5.63266054e-02
-1.24951400e-01 -9.86794829e-01 7.78667808e-01 8.52871656e-01
6.60181761e-01 -6.35319352e-01 4.57549840e-01 1.92810029e-01
-9.01081979e-01 -2.16532558e-01 -1.27985716e-01 -8.13120678e-02
2.05651566e-01 2.24293172e-02 -7.86913991e-01 2.27067769e-01
1.09535730e+00 9.20192420e-01 -1.07629871e+00 1.11425042e+00
-3.26058954e-01 2.07857996e-01 -2.54149675e-01 9.47763994e-02
2.00722907e-02 1.83810309e-01 6.16708040e-01 1.19047165e+00
1.95115551e-01 -2.41138309e-01 -2.52263427e-01 8.73972237e-01
6.62227422e-02 -2.90881366e-01 -7.84084201e-01 3.37501526e-01
4.63693589e-01 1.32336557e+00 -7.44749606e-02 4.80299853e-02
-6.08494341e-01 1.05755723e+00 3.38474572e-01 7.90001810e-01
-9.99094903e-01 -3.56305987e-01 7.59673655e-01 2.16615289e-01
4.25735116e-01 -7.49377683e-02 -2.09080517e-01 -1.19915795e+00
1.81755245e-01 -7.23314166e-01 4.80368346e-01 -1.19062304e+00
-1.28086507e+00 5.34734428e-01 5.97390458e-02 -1.32667220e+00
-4.35015827e-01 -8.13522995e-01 -6.63591266e-01 1.19278002e+00
-1.72715914e+00 -1.65031195e+00 -7.81596899e-01 1.18260050e+00
3.28403354e-01 -4.23410013e-02 6.16200089e-01 1.50667638e-01
-7.51531780e-01 8.07903707e-01 -3.08681458e-01 2.01940268e-01
1.39086759e+00 -1.23254418e+00 2.03418851e-01 8.89483035e-01
-1.72519311e-01 9.13668394e-01 6.35501683e-01 -1.77802280e-01
-1.33691216e+00 -7.06130326e-01 9.16360438e-01 -1.23973453e+00
5.42440712e-01 -6.83880031e-01 -6.50747657e-01 1.03690076e+00
8.39593947e-01 -2.02134550e-01 7.09040880e-01 3.98191959e-01
-4.34297740e-01 -2.66590361e-02 -9.61278021e-01 5.34016967e-01
1.05706918e+00 -5.98798871e-01 -6.36358619e-01 3.07654500e-01
4.29303497e-01 -6.74337566e-01 -4.02341694e-01 2.46800125e-01
5.59176743e-01 -1.21972585e+00 1.05678427e+00 -4.87849742e-01
4.09441918e-01 -4.14417803e-01 -4.61457595e-02 -8.92645001e-01
-1.96090117e-01 -3.38325828e-01 -2.89597392e-01 1.32981443e+00
2.62129247e-01 -6.38229072e-01 4.25911099e-01 6.89676702e-01
1.56859025e-01 -2.98427224e-01 -7.42051780e-01 -1.87455699e-01
-1.35337070e-01 -1.21584147e-01 4.74414766e-01 8.21566641e-01
1.77628502e-01 5.20920038e-01 -3.67517442e-01 5.82278132e-01
8.28458011e-01 1.23891115e-01 1.18404186e+00 -1.15730739e+00
-1.28037967e-02 -3.49360317e-01 -3.10469925e-01 -1.09552705e+00
7.22812787e-02 -3.98356944e-01 -1.23649694e-01 -1.35407460e+00
3.33341479e-01 4.07438159e-01 -2.47779161e-01 5.95173359e-01
-1.33707359e-01 4.24545854e-01 2.15137675e-01 3.28362197e-01
-1.06126297e+00 4.46854353e-01 1.15917110e+00 8.11107904e-02
-5.30527458e-02 4.40750159e-02 -8.73947024e-01 8.19802284e-01
7.64908195e-01 -1.78865865e-01 -3.27488720e-01 -7.66192615e-01
3.47885936e-01 -1.70659795e-01 9.34822023e-01 -8.72120082e-01
6.77293837e-01 -4.56999382e-03 6.03494167e-01 -7.40329862e-01
6.96046352e-01 -8.55713606e-01 -4.56987649e-01 -1.57010719e-01
-2.58104175e-01 8.87364224e-02 2.41085991e-01 6.01798117e-01
-3.99613567e-02 5.87148853e-02 6.36605144e-01 -4.59972396e-02
-8.89913678e-01 1.59297079e-01 8.98107141e-03 -7.30347112e-02
1.02207601e+00 -1.37714550e-01 -7.93693781e-01 -4.73127902e-01
-7.61157155e-01 4.71901774e-01 7.96488464e-01 5.92395961e-01
4.52686906e-01 -1.11030042e+00 -4.63974506e-01 1.41675740e-01
6.18377686e-01 -5.57047240e-02 3.32952589e-01 1.23910022e+00
-7.09452257e-02 7.43584394e-01 -2.74656534e-01 -1.05223668e+00
-1.40602219e+00 8.03173482e-01 5.71798146e-01 4.85562593e-01
-1.22820236e-01 1.00906932e+00 8.42556536e-01 -3.88743281e-01
4.36937809e-01 -9.29989144e-02 -4.15900677e-01 1.65921539e-01
7.96992898e-01 2.83971369e-01 -3.05156410e-01 -1.08544850e+00
-6.92949891e-01 8.79539192e-01 -7.54994601e-02 -2.93073088e-01
1.01811111e+00 -8.07187200e-01 -7.25814104e-02 6.47568226e-01
1.08360112e+00 -2.75661070e-02 -1.48932695e+00 -2.92501092e-01
-2.81926423e-01 -6.07223630e-01 -1.81733131e-01 -9.91870821e-01
-8.49961698e-01 1.20473099e+00 8.25518727e-01 -1.06006145e-01
1.32490313e+00 2.74536848e-01 3.21641237e-01 2.44692475e-01
2.14502543e-01 -8.04053009e-01 1.17231317e-01 4.18613315e-01
8.95370781e-01 -1.89445555e+00 -2.83252418e-01 -1.49978563e-01
-1.03916001e+00 8.72484624e-01 1.04533947e+00 2.51380112e-02
5.81902981e-01 -2.68110543e-01 2.71095037e-01 -4.57380235e-01
-7.51665711e-01 -7.42212236e-01 5.37122905e-01 8.13401043e-01
5.52081943e-01 -2.64392108e-01 3.23161721e-01 4.25595999e-01
-1.23899594e-01 -7.50311241e-02 2.44553193e-01 7.55219221e-01
-1.72040805e-01 -6.76310599e-01 -6.48146629e-01 -2.56757706e-01
-3.23104799e-01 -1.31314740e-01 -8.27156246e-01 8.34321201e-01
8.47944468e-02 1.14092195e+00 2.69729942e-01 -1.46306455e-01
4.08961326e-01 2.25205660e-01 6.46664441e-01 -5.01948476e-01
-2.73055196e-01 1.96976796e-01 -2.05227122e-01 -7.66503453e-01
-8.83354962e-01 -1.07744956e+00 -8.03922832e-01 -2.83410132e-01
3.48004624e-02 -3.61157417e-01 6.58577085e-01 7.24178433e-01
5.94047844e-01 1.88770786e-01 4.52118993e-01 -1.31219077e+00
1.05123468e-01 -1.17101121e+00 -3.75875115e-01 3.69047612e-01
8.94397914e-01 -8.15804601e-01 -3.65199775e-01 3.09760004e-01] | [14.094762802124023, 0.05817781016230583] |
6af27df4-4045-4c0a-a4ba-bd5d1e2c3978 | contour-detection-in-unstructured-3d-point | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Hackel_Contour_Detection_in_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Hackel_Contour_Detection_in_CVPR_2016_paper.pdf | Contour Detection in Unstructured 3D Point Clouds | We describe a method to automatically detect contours, i.e. lines along which the surface orientation sharply changes, in large-scale outdoor point clouds. Contours are important intermediate features for structuring point clouds and converting them into high-quality surface or solid models, and are extensively used in graphics and mapping applications. Yet, detecting them in unstructured, inhomogeneous point clouds turns out to be surprisingly difficult, and existing line detection algorithms largely fail. We approach contour extraction as a two-stage discriminative learning problem. In the first stage, a contour score for each individual point is predicted with a binary classifier, using a set of features extracted from the point's neighborhood. The contour scores serve as a basis to construct an overcomplete graph of candidate contours. The second stage selects an optimal set of contours from the candidates. This amounts to a further binary classification in a higher-order MRF, whose cliques encode a preference for connected contours and penalize loose ends. The method can handle point clouds >10^7 points in a couple of minutes, and vastly outperforms a baseline that performs Canny-style edge detection on a range image representation of the point cloud. | ['Jan D. Wegner', 'Timo Hackel', 'Konrad Schindler'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['line-detection', 'contour-detection'] | ['computer-vision', 'computer-vision'] | [ 3.41748714e-01 3.69131193e-02 -1.28874451e-01 -3.45072776e-01
-6.18373632e-01 -8.15147340e-01 4.19572592e-01 5.70961833e-01
-9.76472050e-02 1.00878313e-01 -4.22361612e-01 -2.77986050e-01
2.77005970e-01 -1.05315006e+00 -5.35409153e-01 -4.97905940e-01
-3.94038618e-01 7.59627759e-01 8.39724183e-01 -1.10959940e-01
7.23054051e-01 1.23471034e+00 -1.48745728e+00 4.21857350e-02
8.40643585e-01 1.10714555e+00 9.19762359e-04 6.19211972e-01
-5.42298675e-01 -4.06098403e-02 -2.56588906e-01 -3.28731507e-01
4.14011657e-01 -5.64472079e-02 -4.41826820e-01 5.53075910e-01
8.30247402e-01 5.15129790e-02 1.82879344e-01 1.20998228e+00
-1.22553064e-03 -4.96288650e-02 9.38108504e-01 -1.17685652e+00
-4.95746359e-02 -3.43286693e-01 -9.16529834e-01 -1.17328137e-01
2.81020343e-01 -3.82878870e-01 1.15864170e+00 -1.48884773e+00
5.29498935e-01 9.89836335e-01 8.03416133e-01 1.00930437e-01
-1.20751977e+00 -4.37787116e-01 9.41874236e-02 -3.86787146e-01
-1.57754183e+00 -1.54377967e-01 1.23396671e+00 -5.88815391e-01
5.05306005e-01 6.88545823e-01 1.09042346e+00 -2.34092996e-01
1.07130408e-01 6.67488098e-01 7.53154159e-01 -4.22279596e-01
5.31761467e-01 -3.16543393e-02 5.58460914e-02 9.93218601e-01
2.91608483e-01 -1.61716461e-01 -3.92384261e-01 -5.62389553e-01
1.28894436e+00 3.05010974e-02 -3.55230838e-01 -9.29837286e-01
-9.66624856e-01 9.23759162e-01 5.93488276e-01 -5.04558608e-02
-3.50066423e-01 -1.10356612e-02 -1.89487964e-01 1.66214332e-01
6.72626197e-01 2.06664458e-01 -3.09869617e-01 2.85387844e-01
-1.17257452e+00 2.00075671e-01 8.10838997e-01 1.06017244e+00
1.24984014e+00 -4.36994642e-01 2.44652912e-01 7.64420629e-01
4.47195798e-01 3.66391361e-01 -2.85918564e-01 -9.14962232e-01
3.25275362e-01 1.11015618e+00 1.66440293e-01 -1.27904034e+00
-3.40580344e-01 -1.62882000e-01 -6.63068831e-01 9.02007103e-01
3.71640623e-01 1.14323036e-03 -1.03163266e+00 7.58998334e-01
7.24241972e-01 2.66869545e-01 -4.18150872e-01 8.97184253e-01
4.66467917e-01 7.02076435e-01 -4.49656993e-01 -2.04129174e-01
1.40415394e+00 -6.17558360e-01 -2.82420088e-02 -4.00858939e-01
3.41077089e-01 -9.31685269e-01 8.22745979e-01 4.68443900e-01
-1.14725852e+00 -1.86832979e-01 -1.07134902e+00 -3.45957801e-02
-1.05178870e-01 2.30703056e-01 8.24445546e-01 4.60126311e-01
-1.20977342e+00 6.54456615e-01 -8.08380902e-01 8.46534595e-02
5.45175612e-01 4.49637204e-01 -9.65252891e-02 6.98242038e-02
-3.85086119e-01 3.77597094e-01 -4.76439707e-02 1.36595219e-01
1.30457565e-01 -3.49234045e-01 -1.05662036e+00 8.84543285e-02
3.15163344e-01 -3.28046948e-01 9.06348050e-01 -8.46381962e-01
-1.23231959e+00 1.20918405e+00 -4.77466285e-01 -4.74220030e-02
4.75678593e-01 3.54144815e-03 -5.63391158e-03 3.59054029e-01
1.12348787e-01 7.01347232e-01 1.00496912e+00 -1.58267605e+00
-1.02138209e+00 -4.38786328e-01 -3.90817881e-01 3.94163311e-01
1.90006942e-01 -7.70004764e-02 -1.10711944e+00 -5.55672169e-01
1.06836009e+00 -8.98768127e-01 -5.89727938e-01 6.37692332e-01
-4.92054701e-01 -5.15326560e-01 8.84274721e-01 -1.97982803e-01
9.29680526e-01 -2.07534289e+00 -2.57241398e-01 1.03652728e+00
3.42768162e-01 -1.99540809e-01 2.20736116e-01 1.14440247e-01
7.55644888e-02 1.12346999e-01 -5.70933163e-01 -2.36043975e-01
-2.83571094e-01 -2.07984522e-01 -2.08578378e-01 6.02413476e-01
4.60928023e-01 5.19646525e-01 -8.37892652e-01 -7.26229548e-01
1.79959968e-01 1.45987675e-01 -4.59884882e-01 -6.41566294e-04
-2.58820266e-01 6.75124000e-04 -6.26224458e-01 8.31177831e-01
9.29126620e-01 -3.56383681e-01 -1.52331725e-01 1.25347776e-02
-3.69696558e-01 -1.16363414e-01 -1.63578904e+00 1.33791566e+00
1.25272572e-01 6.69135928e-01 1.55217558e-01 -4.81521875e-01
1.54510808e+00 -2.79165786e-02 8.09669912e-01 -9.03359279e-02
-1.52257964e-01 3.06196153e-01 -3.91420931e-01 2.28260960e-02
5.11768281e-01 -8.12789872e-02 -7.29664192e-02 8.27022195e-02
-6.02818012e-01 -9.41464663e-01 -8.57596919e-02 -5.66084776e-03
9.91131127e-01 -9.50184688e-02 2.13818625e-01 -3.13087583e-01
5.02187610e-01 3.60761672e-01 7.41825700e-01 4.37375128e-01
1.02925543e-02 9.66879427e-01 4.44891572e-01 -6.13343179e-01
-8.84232700e-01 -1.02231896e+00 -2.33364388e-01 5.59191585e-01
5.76021492e-01 -3.48329663e-01 -6.05099797e-01 -5.56159377e-01
2.55444020e-01 2.46015385e-01 -2.64530420e-01 2.56280780e-01
-7.81187773e-01 -3.36150467e-01 -1.91282079e-01 5.01632571e-01
2.51134425e-01 -8.34083855e-01 -7.18630970e-01 5.25314026e-02
2.75640428e-01 -6.76882744e-01 -7.48079062e-01 2.98042715e-01
-1.26666677e+00 -1.11422908e+00 -7.05177784e-01 -1.22319782e+00
1.16311824e+00 6.07351422e-01 1.30166912e+00 5.27403235e-01
-2.92089552e-01 8.94115716e-02 -9.54564065e-02 -2.78974116e-01
2.23136604e-01 -4.22717035e-01 -4.43554223e-01 9.43456665e-02
4.48503077e-01 -2.56735355e-01 -6.67131841e-01 3.63334656e-01
-5.70780277e-01 -4.69813161e-02 5.44329047e-01 6.35488868e-01
1.22140694e+00 9.58758295e-02 -6.17891401e-02 -8.90025318e-01
2.83772916e-01 -1.22481987e-01 -1.12637496e+00 1.24116085e-01
-3.93319964e-01 -3.97648104e-02 1.39861688e-01 -1.52197123e-01
-5.88362873e-01 7.58731186e-01 -6.91269711e-02 -4.08766657e-01
-7.06390217e-02 3.82090032e-01 -1.65108200e-02 -4.43855375e-01
3.74582171e-01 1.45927116e-01 -2.94886500e-01 -2.26149261e-01
3.18460584e-01 4.84890401e-01 5.31506538e-01 -3.82344812e-01
1.13664758e+00 8.83866012e-01 2.13818640e-01 -1.33058727e+00
-4.13101405e-01 -1.08533776e+00 -1.02561557e+00 -3.50495130e-01
6.93849325e-01 -6.92682922e-01 -4.00695115e-01 2.56383121e-01
-1.32498610e+00 -2.83441752e-01 -2.67311931e-01 1.99883237e-01
-5.34415245e-01 4.57331479e-01 -4.58441198e-01 -1.01033175e+00
-1.71719015e-01 -7.04708636e-01 1.50738704e+00 4.45317566e-01
-7.10276738e-02 -8.77398133e-01 9.69222784e-02 -3.03613394e-02
-2.58771569e-01 3.66775572e-01 8.61628592e-01 -2.11422130e-01
-8.52051675e-01 -4.28065509e-01 -1.59333676e-01 -1.26793817e-01
1.26909446e-02 4.59758520e-01 -7.00384259e-01 -2.19368011e-01
-9.35408771e-02 -2.02130750e-02 8.09207380e-01 3.75186503e-01
1.10027325e+00 3.66401598e-02 -7.39391088e-01 7.60672927e-01
1.49648082e+00 1.54217944e-01 5.61697245e-01 5.62440269e-02
5.85638702e-01 4.72387642e-01 7.98302412e-01 2.88246542e-01
2.02799499e-01 5.51336825e-01 3.67011219e-01 -4.91167814e-01
2.29724422e-01 -2.20719233e-01 4.90129925e-03 5.60809255e-01
-1.02950774e-01 1.29282668e-01 -1.04473495e+00 3.99191678e-01
-1.48109949e+00 -5.71570873e-01 -7.05154121e-01 2.43015957e+00
6.30214453e-01 5.42660356e-01 2.20005456e-02 2.29734436e-01
8.58495057e-01 5.22441827e-02 -6.09824896e-01 -1.48049906e-01
-1.67223737e-02 2.40531534e-01 5.66267431e-01 5.87841690e-01
-1.35880756e+00 1.08525038e+00 6.06039667e+00 5.48446953e-01
-1.14384472e+00 -6.22565389e-01 6.11790597e-01 5.36156893e-01
-4.74883914e-01 2.81000972e-01 -6.33942842e-01 1.45134538e-01
1.13013938e-01 7.03636259e-02 -2.14287788e-02 1.04091501e+00
-2.87759043e-02 -2.24196807e-01 -8.57786417e-01 1.14026558e+00
2.37511354e-03 -1.34168458e+00 -1.58504933e-01 1.63928464e-01
6.39468908e-01 1.55373961e-01 -2.39128992e-01 -2.38246918e-01
2.45081067e-01 -8.33240390e-01 7.51280069e-01 3.89073044e-01
9.24726963e-01 -6.68953717e-01 3.47249597e-01 5.17785192e-01
-1.72164929e+00 3.74234319e-01 -6.79162920e-01 1.40199482e-01
6.72129765e-02 8.39171231e-01 -8.71026516e-01 2.83893403e-02
5.24320722e-01 5.21221220e-01 -4.23441947e-01 1.56492949e+00
-1.97590917e-01 4.69874352e-01 -8.16462040e-01 -8.54744241e-02
6.31880611e-02 -8.11440468e-01 5.21313131e-01 1.20993435e+00
3.17571014e-01 5.49696028e-01 3.93479377e-01 7.18806863e-01
-9.14531350e-02 3.75418663e-01 -7.06026316e-01 3.47640961e-01
7.46519566e-01 1.46394646e+00 -1.54445314e+00 -2.49977171e-01
-5.77516437e-01 1.03022206e+00 3.69942635e-01 1.73206791e-01
-2.37759143e-01 -4.66862082e-01 4.37240273e-01 3.02013725e-01
3.68984848e-01 -6.21931434e-01 -7.77707458e-01 -1.02834105e+00
1.93382487e-01 -3.28427136e-01 1.77041695e-01 -6.06103778e-01
-1.06560385e+00 5.45908153e-01 -5.32059968e-01 -1.62915790e+00
1.40955105e-01 -5.15823126e-01 -1.14979589e+00 7.32942522e-01
-1.60615396e+00 -8.69031191e-01 -5.67802191e-01 4.23244119e-01
6.37313843e-01 4.22300577e-01 6.86652184e-01 -1.04918391e-01
-1.35349140e-01 9.46749076e-02 5.04912436e-02 2.40637675e-01
1.76530987e-01 -1.62812710e+00 8.16812158e-01 7.47157097e-01
4.74184841e-01 3.45187396e-01 4.14230555e-01 -9.62099016e-01
-1.22060215e+00 -9.59084690e-01 8.32348049e-01 -1.70217574e-01
2.81885922e-01 -5.25538385e-01 -1.25406480e+00 3.60305786e-01
-4.62185591e-01 4.23020720e-01 4.24610227e-01 -1.71094134e-01
5.79201803e-02 2.73123235e-01 -9.15117860e-01 6.13243341e-01
9.51706469e-01 -2.24371985e-01 -4.80462968e-01 3.19213539e-01
1.66893393e-01 -4.95874524e-01 -7.22518981e-01 2.74952918e-01
2.50007063e-01 -9.74398375e-01 1.11640680e+00 -1.65142089e-01
5.60947973e-03 -5.20728290e-01 3.14619929e-01 -1.10224438e+00
-4.32709336e-01 -7.46511340e-01 2.77954549e-01 8.06573868e-01
4.45384234e-01 -3.65660846e-01 1.61725008e+00 5.48811436e-01
-1.48193642e-01 -8.55799615e-01 -9.20006752e-01 -5.37964761e-01
-1.55655056e-01 -2.82151282e-01 3.78464907e-01 9.09430742e-01
-7.33021647e-02 2.77002007e-01 3.12751442e-01 5.11105537e-01
8.68411362e-01 7.18160927e-01 8.45292211e-01 -2.01308632e+00
2.85273820e-01 -5.86045146e-01 -5.82594872e-01 -1.61964512e+00
-1.44381404e-01 -8.12617004e-01 3.09168160e-01 -1.63192689e+00
-3.01285684e-01 -1.21996284e+00 3.12155902e-01 2.20560148e-01
-2.85175323e-01 2.95237601e-01 -3.43262777e-02 7.02000201e-01
-2.80199289e-01 3.96843940e-01 1.10807741e+00 -3.14116180e-02
-5.45037448e-01 6.04393721e-01 -2.69975930e-01 1.25065863e+00
6.72072172e-01 -2.95982212e-01 -1.22784264e-01 -1.93245754e-01
1.78006753e-01 1.79970205e-01 8.60281214e-02 -9.20287967e-01
4.93236750e-01 -5.44071436e-01 6.62044168e-01 -1.05683291e+00
5.57157874e-01 -9.54838812e-01 -1.88148767e-01 2.44465038e-01
7.66368061e-02 3.27176973e-02 -3.97342928e-02 6.56972051e-01
-1.78619504e-01 -3.48850310e-01 7.87358224e-01 -4.12127003e-02
-7.04596698e-01 5.52384079e-01 3.97250019e-02 -2.27603227e-01
1.23004305e+00 -6.96699381e-01 1.42908752e-01 -1.92358643e-01
-6.29955053e-01 3.74505103e-01 9.92943287e-01 3.33843604e-02
1.19334233e+00 -1.22258115e+00 -6.27724230e-01 5.95235765e-01
2.84186639e-02 8.06775093e-01 -4.15406644e-01 2.50129521e-01
-9.41462576e-01 -4.87008691e-02 8.46835598e-02 -9.83273745e-01
-1.35590017e+00 2.14103103e-01 3.08688879e-01 1.45161778e-01
-1.09102678e+00 1.15193975e+00 1.23839721e-01 -3.24569568e-02
1.08065255e-01 -5.93834519e-01 -2.94394463e-01 -2.83840615e-02
1.78454146e-01 2.34637305e-01 2.35283419e-01 -6.58666551e-01
-4.12597060e-01 1.09881783e+00 1.94414094e-01 -1.20844118e-01
1.24367583e+00 1.50942579e-01 -1.46725968e-01 3.61318707e-01
1.01338995e+00 4.71372902e-01 -1.67000616e+00 -1.14510626e-01
5.07775843e-01 -7.49482691e-01 1.56258196e-01 -8.35681781e-02
-1.07674479e+00 7.79768348e-01 2.73285657e-01 4.14833516e-01
9.48166132e-01 1.99197114e-01 6.57196581e-01 2.37254709e-01
5.56289613e-01 -9.69962180e-01 1.07344929e-02 4.04119730e-01
8.59179139e-01 -1.17892408e+00 1.20753750e-01 -1.25404000e+00
-5.34436166e-01 1.57321203e+00 4.00491983e-01 -4.84375387e-01
8.07276785e-01 4.02144790e-01 1.25213385e-01 -6.66954517e-01
-4.15009111e-01 -1.66401461e-01 5.33362091e-01 4.32853639e-01
2.18331948e-01 1.45465329e-01 -1.04355268e-01 -4.16865163e-02
-3.67245287e-01 -2.75075525e-01 3.79550368e-01 1.17014062e+00
-1.09231842e+00 -7.97951341e-01 -6.80668235e-01 7.59021103e-01
4.66860607e-02 1.02037869e-01 -4.93253738e-01 4.47211564e-01
-9.98250917e-02 6.85800672e-01 6.23845875e-01 -1.81215361e-01
5.64417779e-01 -1.15208477e-01 3.08408700e-02 -8.54841173e-01
-1.77048311e-01 4.09065783e-01 -2.16725111e-01 -5.00221908e-01
-1.15941390e-02 -8.59161377e-01 -1.69118929e+00 2.26967692e-01
-5.52878201e-01 4.10295218e-01 7.98730075e-01 5.68594515e-01
1.42370462e-01 -2.01885521e-01 9.60412323e-01 -1.16627371e+00
-2.23051414e-01 -3.18082899e-01 -6.27438009e-01 3.37971359e-01
2.63393879e-01 -6.14203095e-01 -4.23847884e-01 1.95323676e-01] | [7.998084545135498, -3.0423715114593506] |
39b5d65f-2052-4d45-a3d1-5ffe2ac596d9 | mindgames-targeting-theory-of-mind-in-large | 2305.03353 | null | https://arxiv.org/abs/2305.03353v1 | https://arxiv.org/pdf/2305.03353v1.pdf | MindGames: Targeting Theory of Mind in Large Language Models with Dynamic Epistemic Modal Logic | Theory of Mind (ToM) is a critical component of intelligence, yet accurately measuring it continues to be a subject of debate. Prior research has attempted to apply human ToM assessments to natural language processing models using either human-created standardized tests or rule-based templates. However, these methods primarily focus on simplistic reasoning and require further validation. In this study, we utilize dynamic epistemic logic, which has established overlaps with ToM, to generate more intricate problems. We also introduce novel verbalization techniques to express these problems using natural language. Our findings indicate that certain language model scaling (from 70M to 6B and 350M to 174B) does not consistently yield results better than random chance. While GPT-4 demonstrates improved epistemic reasoning capabilities, there is still room for enhancement. Our code and datasets are publicly available https://github.com/antoinelrnld/modlog https://huggingface.co/datasets/sileod/mindgames | ['Antoine Lernould', 'Damien Sileo'] | 2023-05-05 | null | null | null | null | ['epistemic-reasoning'] | ['miscellaneous'] | [-1.44569904e-01 5.97861886e-01 -1.96157977e-01 -3.10749590e-01
-6.02336287e-01 -4.43205267e-01 6.55716836e-01 2.19693363e-01
-3.25388402e-01 5.93663573e-01 5.09836018e-01 -6.46051407e-01
-3.62138361e-01 -8.43143225e-01 -3.89345378e-01 -1.37906373e-02
3.51877958e-01 8.10928285e-01 -2.42752172e-02 -2.72113532e-01
3.97391826e-01 -1.64789781e-01 -1.22993279e+00 2.75829911e-01
1.06383216e+00 2.62380660e-01 -1.01927429e-01 4.71659601e-01
3.60036045e-02 1.55257618e+00 -4.21646208e-01 -7.07058311e-01
5.31591885e-02 -4.81167436e-01 -1.34776568e+00 -1.69622853e-01
1.07075758e-01 -5.66162705e-01 -5.87000847e-02 1.13182390e+00
1.30851209e-01 2.04990432e-01 4.60870683e-01 -1.07695675e+00
-8.69755208e-01 1.25937212e+00 -2.67403960e-01 1.67039812e-01
1.02289987e+00 6.60517752e-01 1.00328803e+00 -4.70550448e-01
5.34020364e-01 1.49228990e+00 6.10576093e-01 5.20188808e-01
-1.25572097e+00 -7.37453938e-01 -1.03925608e-01 3.81822407e-01
-1.21710944e+00 -3.82872164e-01 6.31584406e-01 -5.06399274e-01
1.10657883e+00 1.16234571e-01 1.03014541e+00 9.81644809e-01
4.54631001e-01 5.48641145e-01 1.81963444e+00 -5.74647129e-01
3.01675647e-01 1.93948478e-01 6.70614719e-01 7.72818983e-01
7.37514675e-01 -2.92145461e-03 -9.32231307e-01 -1.55797273e-01
6.88301563e-01 -5.13732135e-01 4.19406518e-02 1.73499227e-01
-1.13012111e+00 8.76042128e-01 -1.32394120e-01 2.91569054e-01
-5.92116535e-01 3.86941701e-01 3.74547131e-02 1.73037723e-01
2.66811758e-01 9.04010594e-01 -2.13168189e-01 -7.96221495e-01
-9.28236306e-01 6.88184857e-01 8.53079319e-01 5.88299155e-01
3.63268942e-01 3.98260057e-02 -1.61463358e-02 4.97562110e-01
5.93989372e-01 3.99926960e-01 4.13483262e-01 -1.66332591e+00
2.51305133e-01 7.49541342e-01 2.38552868e-01 -1.20216107e+00
-7.34695256e-01 1.62207577e-02 1.16070300e-01 1.81200117e-01
6.42457485e-01 -8.26442838e-02 -5.22992790e-01 1.94262302e+00
3.72466482e-02 -4.23240274e-01 4.53343205e-02 8.89205575e-01
6.07712030e-01 2.72777110e-01 5.07049203e-01 8.57614726e-02
1.45426583e+00 -5.83166957e-01 -6.42925322e-01 -6.33006692e-01
6.90854371e-01 -5.32612085e-01 1.42529333e+00 5.77977836e-01
-1.73643208e+00 1.07272893e-01 -9.30413127e-01 -2.93340921e-01
-8.30517635e-02 -3.76579762e-01 1.33914006e+00 9.01370943e-01
-1.23946476e+00 1.85289890e-01 -1.00909698e+00 -5.22467434e-01
1.96850315e-01 4.02687397e-03 -8.17924440e-02 -1.60420224e-01
-1.44902360e+00 1.33115172e+00 5.52863836e-01 -1.93394691e-01
-7.39024043e-01 -3.41199368e-01 -9.80000734e-01 -2.51711942e-02
5.77271819e-01 -7.30971634e-01 1.59373093e+00 -8.01345348e-01
-1.63216054e+00 9.27552581e-01 -1.19481012e-02 -5.43055713e-01
3.85394037e-01 -7.90922269e-02 -2.41389737e-01 4.21571285e-01
3.25371355e-01 6.15141392e-01 1.56394631e-01 -8.87471735e-01
-9.24123377e-02 -3.29576939e-01 7.38363862e-01 4.31135148e-01
-6.45153150e-02 5.55849075e-01 1.46479875e-01 -4.76729065e-01
3.65841329e-01 -6.48291707e-01 1.93860680e-01 -2.23305658e-01
-1.83921129e-01 -2.72965252e-01 -1.95116162e-01 -6.96303427e-01
1.13927555e+00 -1.61291349e+00 -1.29396170e-01 2.96510477e-02
4.84933913e-01 -1.76025584e-01 2.80342579e-01 5.18016160e-01
1.54831409e-01 4.88656670e-01 -9.36726779e-02 -8.84037465e-03
4.67190355e-01 -1.10244296e-01 2.93107191e-03 4.09668505e-01
-4.35990915e-02 1.03072894e+00 -9.82627213e-01 -8.53932023e-01
1.72202736e-01 1.74265772e-01 -7.28821874e-01 -3.17431599e-01
-4.64464247e-01 -1.92749854e-02 -3.01618308e-01 7.49365866e-01
4.42518175e-01 -5.31453133e-01 4.29446280e-01 4.71112669e-01
-5.37775494e-02 9.51043129e-01 -1.05956542e+00 1.41050756e+00
-3.35365534e-02 3.96884441e-01 2.54083760e-02 -3.49243850e-01
4.71438944e-01 3.11672658e-01 -1.71432849e-02 -4.67426330e-01
4.20080304e-01 -9.25855637e-02 4.12394017e-01 -5.14714956e-01
4.53371257e-01 -6.72136843e-01 -1.49369821e-01 1.00466979e+00
-1.60084412e-01 -6.84820652e-01 3.94864410e-01 4.47322965e-01
1.15234756e+00 2.19653472e-01 4.70557064e-01 -5.00912368e-01
6.50927499e-02 5.40880263e-01 6.52822971e-01 7.46997058e-01
-4.42247480e-01 1.62152395e-01 7.29326010e-01 9.61665288e-02
-6.24426305e-01 -9.74041939e-01 -1.96315825e-01 9.83105659e-01
1.51736945e-01 -5.87111056e-01 -9.20397937e-01 4.58383672e-02
-1.79314569e-01 1.80259967e+00 -3.39343429e-01 -2.77175069e-01
-8.89144391e-02 -5.40713966e-01 9.26220417e-01 3.97762120e-01
5.75655520e-01 -1.03546846e+00 -1.15818286e+00 6.66748434e-02
-2.98591942e-01 -8.28053117e-01 1.29418328e-01 -1.29952133e-01
-6.63718939e-01 -1.09898663e+00 8.76080990e-02 -1.97923228e-01
3.58206481e-01 2.47280911e-01 1.10814941e+00 4.26969558e-01
1.93637937e-01 8.14131200e-01 -5.12137234e-01 -6.51245058e-01
-5.03807843e-01 -4.61852700e-01 2.29221955e-01 -7.96043158e-01
9.12401557e-01 -4.61288631e-01 -2.70718664e-01 2.16610152e-02
-8.11703682e-01 6.41734302e-01 2.28369683e-01 6.48937345e-01
-1.07425332e-01 2.17328474e-01 1.43196329e-01 -8.00773680e-01
1.03741896e+00 -5.71447074e-01 -3.31284106e-01 7.87381679e-02
-9.58801448e-01 -2.57218391e-01 1.66548043e-01 -2.65450388e-01
-1.46488941e+00 -6.80067241e-01 2.50766307e-01 2.43270919e-01
-3.33357543e-01 9.63952065e-01 2.71713555e-01 1.71498105e-01
8.40559781e-01 -6.80526420e-02 2.33973518e-01 7.81893432e-02
2.62479275e-01 3.98564965e-01 2.63952315e-01 -1.29840183e+00
6.70541883e-01 2.98902214e-01 -4.79529023e-01 -3.99080873e-01
-9.27673817e-01 2.77383298e-01 -1.70969665e-01 -4.50265497e-01
8.31595600e-01 -9.13212597e-01 -8.67635906e-01 2.71879852e-01
-8.07693481e-01 -8.43286514e-01 -1.24202900e-01 7.92976320e-01
-7.79824913e-01 3.05766195e-01 -1.09377217e+00 -1.04276371e+00
-2.28237167e-01 -1.14030492e+00 5.31865418e-01 3.42808753e-01
-1.06727386e+00 -9.56911206e-01 -8.56106356e-02 1.17235219e+00
2.39210829e-01 -2.35968363e-02 9.89893258e-01 -7.60071933e-01
-3.97256523e-01 -1.92581177e-01 1.36078417e-01 -6.89635724e-02
-2.68855870e-01 3.81568400e-03 -6.42214000e-01 1.88876346e-01
5.43658257e-01 -9.55491602e-01 2.87374377e-01 1.61897019e-01
5.32749712e-01 -2.57472008e-01 1.97893366e-01 2.63691455e-01
1.15288770e+00 1.28470927e-01 6.33763850e-01 6.79094851e-01
8.40910152e-02 6.91816807e-01 6.92988038e-01 5.21136463e-01
1.08352637e+00 3.33602160e-01 -1.16645060e-01 6.71053112e-01
5.45058586e-02 -2.65475333e-01 6.35869086e-01 5.23750305e-01
-2.30549015e-02 7.35478625e-02 -1.79560518e+00 4.73807752e-01
-1.58766019e+00 -1.25135589e+00 -1.41809210e-01 1.58796954e+00
1.11765158e+00 5.32274067e-01 -7.80242905e-02 1.05737932e-01
4.82815206e-01 1.06525555e-01 -3.62698257e-01 -5.00221074e-01
-1.29211143e-01 1.03972726e-01 6.01941422e-02 7.63035059e-01
-3.00113052e-01 1.28970778e+00 6.58328438e+00 4.20287132e-01
-7.44874239e-01 3.05805892e-01 5.24450183e-01 -3.07942361e-01
-7.98215568e-01 3.72302830e-01 -2.98059940e-01 2.35844180e-01
1.01360440e+00 -6.61044657e-01 8.56917262e-01 3.07301790e-01
4.19186115e-01 -7.66618788e-01 -9.69495893e-01 5.62714636e-01
1.36003450e-01 -9.13409531e-01 -3.99945267e-02 -9.81837139e-02
4.33438897e-01 -3.73552620e-01 2.86422186e-02 5.28628647e-01
8.10158908e-01 -1.24248695e+00 1.35136604e+00 5.37521362e-01
5.15067041e-01 -3.38456780e-01 6.51102245e-01 4.48034316e-01
-4.76467788e-01 -1.01419315e-01 -1.63639054e-01 -1.06658864e+00
1.95487902e-01 3.87532294e-01 -7.22911417e-01 4.67480719e-02
5.27966142e-01 1.86022997e-01 -6.03011727e-01 3.47671330e-01
-6.59661293e-01 9.98369813e-01 -4.32213634e-01 -2.01411963e-01
3.14268582e-02 -1.17788717e-01 4.84837860e-01 7.23169029e-01
-1.24965690e-01 5.75760841e-01 -1.07132792e-01 1.38291609e+00
1.91353977e-01 -7.52770379e-02 -4.44675177e-01 -5.19244730e-01
7.43501246e-01 9.85513210e-01 -9.18346167e-01 -3.31393033e-01
-5.20580888e-01 3.38222176e-01 4.32831377e-01 2.37476438e-01
-1.01413691e+00 3.07495326e-01 5.94432771e-01 1.60470337e-01
-5.02455533e-01 -4.05983359e-01 -9.75446880e-01 -1.23264039e+00
-1.65022224e-01 -1.26046002e+00 2.92727709e-01 -1.21127713e+00
-9.78026390e-01 -1.61059886e-01 4.13323820e-01 -3.60777140e-01
-4.19313967e-01 -5.53697169e-01 -5.81046402e-01 6.90307021e-01
-6.00978076e-01 -1.13647187e+00 -2.43013009e-01 3.38322252e-01
3.97523940e-01 1.87919423e-01 7.81583428e-01 -4.20584410e-01
-5.21404445e-01 4.42758620e-01 -5.77054501e-01 -6.88266531e-02
6.04410470e-01 -1.20773935e+00 3.25523913e-01 9.18469787e-01
-3.61857623e-01 1.21767569e+00 1.02523136e+00 -1.14630473e+00
-1.54832494e+00 -2.93479174e-01 7.91977406e-01 -8.16497028e-01
1.13071370e+00 1.98139530e-03 -8.36742282e-01 1.21417201e+00
4.59688962e-01 -8.83450806e-01 7.65455484e-01 2.18412071e-01
-3.54448378e-01 4.47955817e-01 -1.42943370e+00 1.09249687e+00
1.05209887e+00 -6.26065135e-01 -1.22746801e+00 4.32343334e-01
6.12367690e-01 -4.11794275e-01 -9.63071764e-01 9.06826090e-03
4.19972628e-01 -1.15352845e+00 5.71854830e-01 -3.04765463e-01
6.04627669e-01 -2.63521791e-01 -7.34544098e-02 -1.07934463e+00
-4.34065491e-01 -5.35544395e-01 2.43465304e-02 1.00842917e+00
5.11314452e-01 -9.96344209e-01 6.44092381e-01 1.66495037e+00
4.22835499e-02 -5.04193962e-01 -7.41593063e-01 -5.65087557e-01
5.22008240e-01 -1.05969203e+00 5.92359364e-01 1.09190691e+00
1.06699812e+00 1.06934234e-01 1.91705301e-01 3.50846164e-02
6.05987310e-01 -2.38663360e-01 5.67638576e-01 -1.06453085e+00
-4.48725998e-01 -6.60412312e-01 -2.11857736e-01 -3.97563934e-01
4.07736331e-01 -9.90477324e-01 -2.02448130e-01 -1.96722865e+00
5.47727227e-01 -5.10091633e-02 2.39137888e-01 7.34555483e-01
-5.76268993e-02 -1.64455190e-01 4.82970148e-01 1.41695943e-02
-5.91991246e-01 3.06056827e-01 9.22186971e-01 5.82371168e-02
5.76526895e-02 -6.42912745e-01 -1.43226111e+00 1.14029765e+00
1.47076631e+00 -2.98126787e-01 -6.00636005e-01 -6.02038145e-01
6.65699363e-01 2.43153602e-01 6.16316676e-01 -1.15727532e+00
4.95433807e-01 -7.29393899e-01 2.00346619e-01 3.75486054e-02
3.92594993e-01 -3.16187292e-01 3.64637554e-01 3.35921556e-01
-4.42395002e-01 2.99877286e-01 4.11554575e-01 -1.19934745e-01
2.59406388e-01 -6.04674101e-01 5.08227348e-01 -4.78689671e-01
-4.78717327e-01 -6.52041316e-01 -8.29257846e-01 2.66719133e-01
9.85614717e-01 -2.10068941e-01 -7.96554089e-01 -7.66368151e-01
-4.83873278e-01 4.10557747e-01 9.60929215e-01 -3.88337821e-02
5.81818461e-01 -9.11688924e-01 -6.37970567e-01 -3.37862760e-01
-3.88972573e-02 -2.72058517e-01 2.27533191e-01 9.99504149e-01
-9.05691445e-01 5.85297823e-01 -1.89397797e-01 5.53004816e-02
-7.25311935e-01 4.64209348e-01 4.55849260e-01 1.34946257e-02
-5.94187379e-01 8.30869913e-01 -1.58742398e-01 -3.12311828e-01
-1.50773749e-01 -3.17268133e-01 -1.22033395e-01 -1.25618488e-01
6.29343390e-01 5.84900558e-01 -4.63785619e-01 -4.29898322e-01
-4.33661759e-01 -2.17052493e-02 9.09906533e-03 -7.05650270e-01
1.07153678e+00 -1.56250298e-01 -5.88201225e-01 6.90213501e-01
2.61710957e-02 9.90651920e-02 -8.03896248e-01 1.30566165e-01
1.23222992e-01 -3.93756479e-01 4.89696451e-02 -1.07932925e+00
-3.18527251e-01 3.20148438e-01 6.04134146e-03 2.37827316e-01
7.86119044e-01 2.00386807e-01 2.49463797e-01 5.66640675e-01
9.48439777e-01 -1.29255831e+00 -1.31714210e-01 6.25751019e-01
9.59262908e-01 -9.54451203e-01 3.22303176e-01 -2.65124261e-01
-8.05090904e-01 6.60116494e-01 1.04787171e+00 1.88995659e-01
3.98252368e-01 3.10953856e-01 1.04819909e-01 -5.95997810e-01
-1.12850404e+00 -9.68047231e-02 -4.20064896e-01 3.76792163e-01
7.77281523e-01 3.95303637e-01 -9.33178425e-01 9.79268312e-01
-1.05098999e+00 2.85171628e-01 1.13531280e+00 1.05616522e+00
-4.70810115e-01 -7.18813241e-01 -8.97724807e-01 7.05764174e-01
-4.84289348e-01 -4.42698956e-01 -5.84460735e-01 9.32278931e-01
-3.02697390e-01 1.42061985e+00 -1.57634333e-01 -4.47073370e-01
-6.15854301e-02 4.93782610e-01 5.15687883e-01 -8.05920660e-01
-6.46328986e-01 -3.11804354e-01 5.10238171e-01 -5.16678393e-01
-4.51283664e-01 -1.15083158e+00 -1.57564068e+00 -9.49616134e-01
5.40770814e-02 9.49928388e-02 1.86662942e-01 1.03294313e+00
6.45925850e-02 2.75167823e-01 -5.20016670e-01 -3.99197876e-01
-6.90051138e-01 -1.15316248e+00 -5.55654824e-01 1.93028018e-01
-3.49823743e-01 -9.55776393e-01 -4.54420716e-01 -1.41835898e-01] | [9.655898094177246, 7.449678897857666] |
b347b789-6fca-464e-942a-a70b5682b335 | chae-fine-grained-controllable-story | 2210.05221 | null | https://arxiv.org/abs/2210.05221v1 | https://arxiv.org/pdf/2210.05221v1.pdf | CHAE: Fine-Grained Controllable Story Generation with Characters, Actions and Emotions | Story generation has emerged as an interesting yet challenging NLP task in recent years. Some existing studies aim at generating fluent and coherent stories from keywords and outlines; while others attempt to control the global features of the story, such as emotion, style and topic. However, these works focus on coarse-grained control on the story, neglecting control on the details of the story, which is also crucial for the task. To fill the gap, this paper proposes a model for fine-grained control on the story, which allows the generation of customized stories with characters, corresponding actions and emotions arbitrarily assigned. Extensive experimental results on both automatic and human manual evaluations show the superiority of our method. It has strong controllability to generate stories according to the fine-grained personalized guidance, unveiling the effectiveness of our methodology. Our code is available at https://github.com/victorup/CHAE. | ['Shanlin Zhou', 'Zhihua Wei', 'Han Jiang', 'Xinpeng Wang'] | 2022-10-11 | null | https://aclanthology.org/2022.coling-1.559 | https://aclanthology.org/2022.coling-1.559.pdf | coling-2022-10 | ['story-generation'] | ['natural-language-processing'] | [-4.75008339e-02 2.47288078e-01 -2.82175034e-01 -3.01716417e-01
-2.05818862e-01 -7.31132567e-01 9.28221703e-01 -1.35678556e-02
4.30103354e-02 7.75172293e-01 8.88531148e-01 4.20540333e-01
8.54492188e-02 -9.30619776e-01 -4.16885495e-01 -4.20641601e-01
4.65680540e-01 3.28558743e-01 1.55259714e-01 -5.43662071e-01
5.23598015e-01 1.43610850e-01 -1.65025997e+00 3.92243743e-01
9.75740254e-01 5.61042547e-01 3.80170465e-01 4.65826482e-01
-1.31773904e-01 6.66379869e-01 -7.23128796e-01 -3.60700727e-01
-2.02877283e-01 -6.50824547e-01 -5.95188081e-01 2.72783965e-01
-1.48390546e-01 -1.26004159e-01 -5.11048548e-02 8.61273825e-01
5.70902228e-01 1.70280829e-01 6.06975734e-01 -1.30031729e+00
-8.16122830e-01 1.08910143e+00 -2.27730483e-01 -2.50739247e-01
8.01512361e-01 1.42707556e-01 1.07834458e+00 -4.68418479e-01
8.62487972e-01 1.19977438e+00 2.55430460e-01 6.18805826e-01
-9.54854429e-01 -5.21570563e-01 4.30582821e-01 1.20400511e-01
-1.29121995e+00 -3.72175008e-01 9.44429934e-01 -6.01482093e-01
3.81947130e-01 4.94963914e-01 8.89876366e-01 1.40066111e+00
8.16540793e-02 8.50849450e-01 1.01879847e+00 -4.31945950e-01
1.65114552e-01 4.19294238e-01 -1.86075285e-01 3.29902351e-01
-2.48043481e-02 -2.03052983e-01 -6.60184443e-01 -1.56644243e-03
9.06359255e-01 -2.90958047e-01 -4.91230160e-01 -2.64710516e-01
-1.49859893e+00 9.02600706e-01 -7.09315687e-02 5.26801884e-01
-4.73078638e-01 -2.45930448e-01 4.23411846e-01 -1.11719139e-03
5.51646113e-01 9.40880656e-01 -2.81602055e-01 -6.54866636e-01
-6.61398351e-01 6.16960943e-01 9.33981478e-01 1.29163992e+00
3.60300064e-01 -1.27957135e-01 -9.98523653e-01 8.67867589e-01
-2.74973959e-02 1.50795877e-01 5.08774996e-01 -8.08372498e-01
2.25419790e-01 7.20811248e-01 3.87370378e-01 -1.14328122e+00
-8.67511705e-02 -2.63202310e-01 -7.31312037e-01 -9.21048895e-02
7.04636946e-02 -4.97708470e-01 -5.25658965e-01 1.88286507e+00
3.49382848e-01 -1.67683676e-01 -2.36319393e-01 9.82463598e-01
9.45001185e-01 7.90515304e-01 6.65282458e-02 -3.14501286e-01
1.46778822e+00 -1.04044986e+00 -1.03369021e+00 -5.55615611e-02
3.37209165e-01 -8.42054725e-01 1.55357993e+00 3.61521363e-01
-1.12627780e+00 -4.31590497e-01 -8.42579722e-01 1.11637533e-01
-2.80461073e-01 4.21120673e-01 6.88391328e-01 3.93175721e-01
-8.20326805e-01 4.55099761e-01 -3.62805724e-01 -5.42299330e-01
4.06101719e-03 9.75813996e-03 4.33740020e-02 2.15009809e-01
-1.43153214e+00 7.44485915e-01 6.44289851e-01 -1.97610646e-01
-4.86039430e-01 -5.26574492e-01 -6.95693552e-01 2.93602794e-02
6.61060631e-01 -8.12548518e-01 1.40454507e+00 -1.02708983e+00
-2.10163069e+00 6.23740435e-01 3.12008313e-03 7.89319128e-02
7.63310432e-01 -4.41407859e-01 -1.58635065e-01 -5.81963025e-02
1.88572422e-01 8.71253967e-01 5.77374101e-01 -1.15466464e+00
-5.51870644e-01 1.63751096e-01 3.61256987e-01 5.45245469e-01
-5.26035130e-01 2.40398571e-01 -6.76503718e-01 -1.02591610e+00
-3.81859392e-01 -8.39634657e-01 -2.98430741e-01 -4.05124813e-01
-6.84608877e-01 -3.66749585e-01 5.99359810e-01 -4.60425347e-01
1.61825657e+00 -2.12773490e+00 3.90301555e-01 -2.37194240e-01
-1.05250075e-01 -3.59520949e-02 4.45326641e-02 7.28214741e-01
2.79299349e-01 3.26288521e-01 1.51151717e-01 -9.33998823e-02
2.99321353e-01 -1.25957638e-01 -1.87519878e-01 -3.97986285e-02
1.47726610e-01 8.64216506e-01 -8.31552565e-01 -6.40969157e-01
-6.49258203e-04 5.16348481e-01 -7.22599089e-01 4.33655471e-01
-6.51536882e-01 5.87708712e-01 -9.70397890e-01 3.67765635e-01
1.61380664e-01 -1.16524644e-01 2.10210239e-03 -4.79096919e-02
-3.90770078e-01 1.28416598e-01 -1.14112890e+00 1.58273447e+00
-3.03314626e-01 3.74924481e-01 -2.34804168e-01 -2.35968143e-01
1.02144086e+00 5.65131426e-01 1.90821797e-01 -3.44570965e-01
3.89477819e-01 -2.52058625e-01 -4.04158503e-01 -6.20124519e-01
9.17269468e-01 -6.09689280e-02 -5.56485355e-01 4.90575045e-01
-3.91909868e-01 -6.20590806e-01 5.24909317e-01 3.08190256e-01
5.80500185e-01 4.70751703e-01 4.01560724e-01 -9.80071127e-02
2.64275938e-01 7.51263574e-02 5.12156487e-01 6.65659189e-01
3.75281461e-02 8.12214315e-01 6.63243175e-01 -5.11413738e-02
-9.92027462e-01 -5.16203642e-01 2.53362507e-01 1.12126613e+00
2.47297287e-01 -6.99030459e-01 -1.07529914e+00 -3.14946771e-01
-4.16289687e-01 1.03991544e+00 -6.83270097e-01 9.16490182e-02
-3.43292236e-01 -6.21385634e-01 2.63531387e-01 1.89369380e-01
4.76629287e-01 -1.48602593e+00 -7.52699316e-01 2.35207900e-01
-6.72271550e-01 -1.13279152e+00 -6.78061128e-01 -4.79016960e-01
-3.88537556e-01 -5.51570952e-01 -6.31763518e-01 -4.94060844e-01
4.72555399e-01 -3.97381410e-02 1.07971537e+00 -2.01017428e-02
-8.88809636e-02 2.64133722e-01 -8.26200545e-01 -5.31503856e-01
-3.79926652e-01 3.41092616e-01 -2.86642313e-01 2.40318522e-01
-1.05814748e-01 -5.57860136e-01 -5.48680663e-01 5.10755479e-01
-9.03822064e-01 7.45374322e-01 3.71154338e-01 4.65660423e-01
4.80201870e-01 1.55141890e-01 5.55309296e-01 -9.10100579e-01
1.19377983e+00 -4.26096529e-01 -2.67820090e-01 1.41902551e-01
-3.52105200e-01 -5.53583391e-02 6.25143468e-01 -6.37302458e-01
-1.33962548e+00 -1.71745401e-02 7.27105811e-02 5.73473908e-02
-4.23762232e-01 3.92988086e-01 -5.06393373e-01 4.86377239e-01
3.47666711e-01 1.08238101e-01 -3.59050661e-01 -2.44146302e-01
5.84392369e-01 5.06371617e-01 2.68754154e-01 -8.32697630e-01
6.58460677e-01 2.50544816e-01 -5.56715846e-01 -4.65919852e-01
-7.42287338e-01 -1.18653417e-01 -4.77111131e-01 -5.60742557e-01
9.27976787e-01 -7.88512111e-01 -5.11362851e-01 3.05596054e-01
-1.19226909e+00 -5.22095323e-01 -1.76067024e-01 3.54352593e-01
-7.82362521e-01 -4.80458885e-02 -6.62390828e-01 -6.15322649e-01
-3.40793371e-01 -9.81078625e-01 9.67369020e-01 5.09766042e-01
-9.40501451e-01 -7.69500673e-01 6.41613752e-02 3.01087350e-01
4.09968197e-01 5.75113297e-01 8.12235653e-01 -4.79738027e-01
-5.25047660e-01 -2.82952935e-01 4.77090478e-02 -2.19400406e-01
1.10313334e-01 4.02306825e-01 -7.09218144e-01 1.71042830e-01
-1.55423060e-01 -3.38009983e-01 3.03030491e-01 2.85907596e-01
9.92587209e-01 -5.39763093e-01 -3.33490670e-01 2.87157625e-01
1.19701850e+00 2.22216144e-01 6.83252871e-01 5.05625188e-01
6.63435698e-01 7.09997952e-01 1.12987494e+00 9.96266603e-01
5.36325157e-01 9.07991469e-01 1.47088632e-01 7.79814720e-02
2.08473089e-03 -4.87437546e-01 2.93574512e-01 5.79175115e-01
-3.57087106e-01 -6.23154640e-01 -5.24471521e-01 5.40584624e-01
-2.00899649e+00 -1.10859013e+00 -7.16097355e-02 1.68518889e+00
9.85802531e-01 1.46877900e-01 1.15879990e-01 -4.90172990e-02
8.47531259e-01 4.07304317e-01 -2.71032691e-01 -4.97651905e-01
-1.84886351e-01 -3.24923158e-01 -1.55373752e-01 3.79244417e-01
-9.56854880e-01 1.31912732e+00 6.11161709e+00 9.89481747e-01
-1.12497497e+00 -1.12587303e-01 6.82074666e-01 -3.10907334e-01
-5.91322541e-01 -7.19963461e-02 -7.87009001e-01 5.02452135e-01
3.31629395e-01 -5.10946274e-01 3.75344574e-01 7.38722146e-01
8.93564582e-01 -1.07129356e-02 -9.83011782e-01 5.62549591e-01
1.60864010e-01 -1.29111087e+00 1.93923324e-01 -2.15124056e-01
8.29198122e-01 -6.38460755e-01 -7.94920996e-02 3.24651182e-01
4.21119571e-01 -7.56445646e-01 1.25117183e+00 5.65080822e-01
6.93393111e-01 -9.13516819e-01 5.14490545e-01 4.73037183e-01
-9.54055607e-01 2.54842907e-01 1.27162263e-01 -3.39681268e-01
4.32143062e-01 4.26083595e-01 -7.39961505e-01 3.64310473e-01
5.42375088e-01 5.86383462e-01 -5.85713923e-01 6.41823769e-01
-8.16271544e-01 5.31610966e-01 1.34268776e-01 -4.51729566e-01
2.74502970e-02 -2.07647145e-01 7.29916215e-01 1.52730024e+00
4.88513172e-01 5.13736486e-01 3.50375801e-01 1.05371630e+00
1.48493752e-01 5.40604055e-01 -5.48548579e-01 -2.91747987e-01
5.97945869e-01 1.34502268e+00 -7.74125516e-01 -3.86471838e-01
1.21816672e-01 1.00042617e+00 1.20274320e-01 2.77559817e-01
-9.84695613e-01 -3.85823995e-01 4.12391335e-01 3.03290784e-01
3.31801236e-01 1.86809711e-03 -5.48099995e-01 -1.21097529e+00
-3.74564379e-02 -1.06283176e+00 5.71618639e-02 -9.82905626e-01
-9.64419603e-01 6.68658912e-01 2.04641223e-01 -1.23453057e+00
-3.28668356e-01 2.74612196e-02 -8.74493778e-01 4.61530715e-01
-9.42926764e-01 -1.19038439e+00 -3.60548466e-01 3.93711567e-01
7.80789495e-01 5.07207774e-02 7.41264343e-01 -8.13638195e-02
-6.97511315e-01 4.21138197e-01 -4.14070129e-01 -1.28338769e-01
7.96149492e-01 -1.02675712e+00 2.33466387e-01 7.71348655e-01
-3.05387408e-01 4.50927377e-01 1.14496303e+00 -6.80679202e-01
-1.07747340e+00 -8.47868979e-01 1.15198910e+00 -2.94359893e-01
5.84250152e-01 -4.42364156e-01 -6.53538823e-01 3.69858325e-01
6.04607940e-01 -8.40738893e-01 6.36438012e-01 7.30388686e-02
-8.60505030e-02 3.24856788e-01 -9.65648472e-01 1.13286400e+00
1.02775514e+00 -8.06266963e-02 -4.02092546e-01 3.23563457e-01
1.01406515e+00 -5.18925309e-01 -7.64963210e-01 2.01558948e-01
5.37222505e-01 -1.16621530e+00 6.04673445e-01 -1.97484523e-01
9.06985641e-01 -3.17418128e-01 8.40083957e-02 -1.47249961e+00
-5.93319833e-01 -7.58256853e-01 3.30765426e-01 1.82358432e+00
4.50413525e-01 -3.29700172e-01 4.86111939e-01 7.54493535e-01
-2.31008306e-01 -7.77464271e-01 -1.58439189e-01 -3.83446336e-01
-2.62008309e-01 -3.37359071e-01 8.77214551e-01 9.89297092e-01
3.90619278e-01 4.80159640e-01 -6.13171875e-01 8.79818723e-02
-2.61090416e-02 3.60988915e-01 9.63817179e-01 -1.00664115e+00
-4.19218659e-01 -7.73646891e-01 2.42701583e-02 -8.02881956e-01
3.21925692e-02 -4.38705772e-01 2.87273616e-01 -1.82626736e+00
3.72842818e-01 -1.45589948e-01 8.62290412e-02 4.21698630e-01
-2.94903606e-01 9.63923484e-02 4.64379728e-01 -2.21104920e-02
-7.50843287e-01 6.91846788e-01 1.59164596e+00 5.27609922e-02
-5.04877448e-01 5.63098527e-02 -1.14175880e+00 7.35500932e-01
1.08089507e+00 -3.13290298e-01 -5.56292653e-01 -2.33738601e-01
3.48176926e-01 1.69582576e-01 9.94351432e-02 -7.70117760e-01
8.72834120e-03 -7.74001896e-01 3.93417478e-02 -3.69656295e-01
3.50227565e-01 -3.14460874e-01 3.78523856e-01 1.10399380e-01
-6.20545626e-01 -6.95557669e-02 1.36381835e-01 2.10666925e-01
-3.45506132e-01 -1.52358353e-01 5.16825020e-01 -2.24279791e-01
-5.04794061e-01 2.10473850e-01 -5.04925966e-01 6.57803640e-02
1.19992638e+00 -1.85855910e-01 -7.56236911e-02 -8.19439530e-01
-5.30617237e-01 2.25997686e-01 6.55452013e-01 7.68857419e-01
4.19894427e-01 -1.41941416e+00 -9.32684898e-01 -1.68367848e-01
1.93722069e-01 -4.66908552e-02 1.81737512e-01 4.04565364e-01
-2.10274071e-01 2.93914974e-01 -1.95352107e-01 -2.48913541e-01
-1.25988400e+00 5.93489408e-01 -1.21081270e-01 -2.88472146e-01
-4.36350226e-01 7.19687462e-01 3.20915580e-01 -5.30015752e-02
1.96219012e-01 -2.58846372e-01 -6.68058157e-01 3.45176578e-01
4.82254595e-01 1.05019368e-01 -5.31512260e-01 -5.28880835e-01
2.99250446e-02 4.76622313e-01 -6.08146228e-02 -3.82759988e-01
1.32632077e+00 -3.97943288e-01 -6.80347830e-02 5.51656544e-01
4.83760983e-01 2.12199643e-01 -1.31890094e+00 -3.72671057e-03
-9.55268815e-02 -4.87820804e-01 -2.32369930e-01 -9.08145308e-01
-7.27186620e-01 4.20162976e-01 -1.19962655e-01 3.39352697e-01
1.17282927e+00 1.27291352e-01 6.68142498e-01 -2.52087172e-02
3.47642332e-01 -1.21339273e+00 2.61768490e-01 5.24359703e-01
1.33764374e+00 -8.44759703e-01 -9.53190774e-02 -6.00246131e-01
-1.25601363e+00 8.79018486e-01 7.81548202e-01 4.20556590e-02
2.03391820e-01 1.82189941e-01 5.26496805e-02 -3.34675349e-02
-1.00528944e+00 1.05838530e-01 1.79333061e-01 4.10782278e-01
8.39303672e-01 3.15325946e-01 -7.04188108e-01 9.07441020e-01
-6.87791049e-01 1.83644727e-01 7.29252279e-01 7.71685004e-01
-4.49079961e-01 -1.14851952e+00 -4.09065485e-01 1.23081272e-02
-2.43313238e-01 6.43895147e-03 -7.30886579e-01 6.71300173e-01
2.60038406e-01 1.07577813e+00 -2.38232717e-01 -2.76911706e-01
6.08951688e-01 -8.60923305e-02 1.65476948e-01 -7.87164867e-01
-8.16892743e-01 2.39133954e-01 4.12619650e-01 -3.26519936e-01
-2.56682694e-01 -7.53722847e-01 -1.13052809e+00 -4.05195594e-01
-1.52663976e-01 2.35169336e-01 3.77116114e-01 8.04707170e-01
4.28719550e-01 6.54319167e-01 6.82173371e-01 -9.32333410e-01
-3.32323909e-02 -1.04727089e+00 -4.54646289e-01 5.63118935e-01
-2.06692308e-01 -5.27594328e-01 -1.02226086e-01 2.83896387e-01] | [11.757243156433105, 8.843435287475586] |
26298307-01d0-4e08-ac3a-bdfa5c3346b6 | logicllm-exploring-self-supervised-logic | 2305.13718 | null | https://arxiv.org/abs/2305.13718v2 | https://arxiv.org/pdf/2305.13718v2.pdf | LogicLLM: Exploring Self-supervised Logic-enhanced Training for Large Language Models | Existing efforts to improve logical reasoning ability of language models have predominantly relied on supervised fine-tuning, hindering generalization to new domains and/or tasks. The development of Large Langauge Models (LLMs) has demonstrated the capacity of compressing abundant knowledge into a single proxy, enabling them to tackle multiple tasks effectively. Our preliminary experiments, nevertheless, show that LLMs do not show capability on logical reasoning. The performance of LLMs on logical reasoning benchmarks is far behind the existing state-of-the-art baselines. In this paper, we make the first attempt to investigate the feasibility of incorporating logical knowledge through self-supervised post-training, and activating it via in-context learning, which we termed as LogicLLM. Specifically, we devise an auto-regressive objective variant of MERIt and integrate it with two LLM series, i.e., FLAN-T5 and LLaMA, with parameter size ranging from 3 billion to 13 billion. The results on two challenging logical reasoning benchmarks demonstrate the effectiveness of LogicLLM. Besides, we conduct extensive ablation studies to analyze the key factors in designing logic-oriented proxy tasks. | ['Nancy F. Chen', 'Zhengyuan Liu', 'Aixin Sun', 'Bosheng Ding', 'Shafiq Joty', 'Zhiyang Teng', 'Fangkai Jiao'] | 2023-05-23 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-7.02133775e-02 3.47205222e-01 -3.68597090e-01 -5.75020254e-01
-7.80137062e-01 -4.85001922e-01 8.05547416e-01 -1.93655461e-01
-3.73679429e-01 7.76233077e-01 2.94238567e-01 -8.16658497e-01
1.21335862e-02 -7.47428417e-01 -9.35470819e-01 5.56788258e-02
-3.61435451e-02 4.23265278e-01 1.38074681e-01 -3.10131282e-01
3.66535075e-02 6.67543337e-02 -1.24371326e+00 8.34712625e-01
1.02775300e+00 7.92148173e-01 -5.05436584e-02 5.63755453e-01
-4.74919736e-01 1.70658720e+00 -6.93264008e-01 -8.27521145e-01
-1.66385751e-02 -2.74947345e-01 -1.00031900e+00 -5.56604147e-01
4.90036130e-01 -5.55267572e-01 -2.16501951e-01 6.70265675e-01
1.66802004e-01 4.39105891e-02 2.51613021e-01 -1.17477560e+00
-9.96904373e-01 1.47333801e+00 -1.41166359e-01 2.38363212e-03
3.82856905e-01 2.70107776e-01 1.36794949e+00 -8.70009542e-01
4.38807726e-01 1.59563792e+00 9.36630309e-01 7.10390449e-01
-1.06602025e+00 -6.36008799e-01 4.02231514e-01 2.27800146e-01
-1.32312226e+00 -7.08489656e-01 6.79096162e-01 -1.64068975e-02
1.60506248e+00 6.48421273e-02 3.21274400e-01 1.16836953e+00
1.92277938e-01 1.25339949e+00 1.26962829e+00 -7.09128201e-01
2.97351629e-01 -1.86125021e-02 2.97067285e-01 1.10382593e+00
1.33082956e-01 -1.71679959e-01 -9.41757739e-01 -1.14676133e-01
3.36541891e-01 -4.99569654e-01 2.05648571e-01 -5.36773615e-02
-1.07858264e+00 7.02906907e-01 2.79682934e-01 2.41648272e-01
-5.21595962e-02 6.48805320e-01 5.23400784e-01 5.27434349e-01
2.05427215e-01 7.83942938e-01 -7.66116321e-01 -3.04741293e-01
-8.11360359e-01 4.03547227e-01 9.14556563e-01 9.55308378e-01
2.67855346e-01 -6.75449893e-02 -2.75094062e-01 6.85444474e-01
3.89162332e-01 3.83237392e-01 4.43589240e-01 -1.09977305e+00
8.03104937e-01 8.59768331e-01 -3.75309289e-01 -2.67799824e-01
-4.36069965e-01 -3.02626431e-01 -3.87550920e-01 -9.23457146e-02
3.51162314e-01 -2.64164537e-01 -6.84251249e-01 1.99566376e+00
-2.23967224e-01 1.88510075e-01 6.14271581e-01 4.87978160e-01
7.92042494e-01 4.19626325e-01 3.26642990e-01 1.75561726e-01
1.25600386e+00 -1.27012682e+00 -7.10313737e-01 -6.30849302e-01
1.06622910e+00 -4.25548911e-01 1.86010110e+00 6.44024909e-01
-1.44954383e+00 -2.73318619e-01 -1.19196320e+00 -2.90273190e-01
-4.32330608e-01 6.46623001e-02 1.41475511e+00 6.22015893e-01
-1.07310474e+00 2.05475941e-01 -1.00006354e+00 -1.50383487e-01
3.45540851e-01 8.60087648e-02 7.74937915e-04 -2.11313754e-01
-1.56523693e+00 9.37203050e-01 5.07572949e-01 1.66490182e-01
-8.46053243e-01 -8.10888588e-01 -8.33574176e-01 5.98797016e-02
6.67532206e-01 -8.11615705e-01 1.49437165e+00 -5.57024240e-01
-1.58966041e+00 7.25492656e-01 -1.90813258e-01 -7.45182037e-01
5.29098749e-01 -5.47101736e-01 -4.43480343e-01 -2.24430695e-01
-7.85671696e-02 6.57066226e-01 5.41820467e-01 -8.95130754e-01
-4.55446035e-01 1.91668957e-01 7.79469252e-01 4.15499210e-02
-4.13111866e-01 1.81529447e-02 -6.29242063e-01 -5.79627693e-01
1.14739733e-02 -6.96478963e-01 2.12710872e-01 -4.06185240e-01
-2.69109696e-01 -5.32274604e-01 4.64019150e-01 -4.29971397e-01
1.42991459e+00 -1.89714766e+00 1.18408434e-01 -7.49726668e-02
2.02082321e-01 3.26508820e-01 -2.88144737e-01 2.92700529e-01
2.25907579e-01 2.85096645e-01 -1.14387594e-01 -3.54767770e-01
2.97926188e-01 3.73232067e-01 -6.35741949e-01 -2.23020241e-01
4.82791275e-01 1.47507322e+00 -8.86922538e-01 -6.26277030e-01
-2.19846189e-01 5.33633418e-02 -9.38729107e-01 3.95986848e-02
-1.05685782e+00 -1.78590566e-01 -4.19707865e-01 8.32336962e-01
2.61290312e-01 -6.06014252e-01 3.67498577e-01 4.79044728e-02
2.76164889e-01 8.85918260e-01 -7.20810831e-01 2.01586938e+00
-8.65640819e-01 5.92991650e-01 -3.75137866e-01 -6.28433287e-01
5.07533312e-01 1.40369177e-01 -5.08401589e-03 -1.02986574e+00
-1.72140747e-01 1.39622703e-01 8.40419158e-02 -5.02596200e-01
3.16281676e-01 7.53173931e-03 -2.56299376e-01 5.53421617e-01
-9.03982744e-02 -1.70026317e-01 3.61493319e-01 3.37938696e-01
1.41079748e+00 4.40905929e-01 3.70128900e-02 -6.55191252e-03
8.08387280e-01 -6.36436865e-02 3.40036750e-01 1.05667698e+00
1.02496579e-01 2.79333983e-02 7.64488995e-01 -2.84354150e-01
-5.52282512e-01 -1.08767796e+00 -4.11411449e-02 1.32289898e+00
-1.89899415e-01 -9.60844874e-01 -7.70056963e-01 -8.76790166e-01
1.85722202e-01 1.18013334e+00 -2.23915592e-01 -4.17532742e-01
-6.70111001e-01 -7.95843363e-01 1.33651757e+00 8.89560282e-01
9.76319790e-01 -1.13917637e+00 -4.72116023e-01 1.19770154e-01
-2.42012739e-01 -1.48984277e+00 2.03031644e-01 2.18415290e-01
-7.67198682e-01 -7.60530353e-01 1.64643377e-01 -4.48409349e-01
4.39186722e-01 -1.06755450e-01 1.51111412e+00 3.37683678e-01
1.28461033e-01 2.56337017e-01 -2.39159733e-01 -2.88494170e-01
-5.14151216e-01 3.83763969e-01 4.99798078e-03 -6.83596134e-01
4.53588396e-01 -4.36028957e-01 -2.80575752e-02 -6.05069622e-02
-6.39440417e-01 3.09299409e-01 9.83212590e-01 8.45045507e-01
2.59428531e-01 3.06638777e-01 4.66666758e-01 -1.24203575e+00
1.03720772e+00 -1.59455299e-01 -5.04159212e-01 7.30995417e-01
-8.74304473e-01 5.57395399e-01 7.75319576e-01 -2.64957935e-01
-1.28901601e+00 -4.69246984e-01 -5.20088263e-02 -1.16812311e-01
3.31039220e-01 1.10284841e+00 -1.77611664e-01 1.51477501e-01
6.64511800e-01 3.39623429e-02 -3.11881214e-01 -2.49322459e-01
6.67059600e-01 4.29305375e-01 6.36948347e-01 -1.33711314e+00
7.35271275e-01 8.35369602e-02 -1.71512455e-01 -3.57409745e-01
-1.23667657e+00 2.25818917e-01 -2.74412870e-01 1.89152136e-01
5.36455274e-01 -1.22869265e+00 -8.83289754e-01 4.28699285e-01
-1.00975549e+00 -1.05562174e+00 -7.12560117e-02 2.20658883e-01
-5.74831247e-01 1.43523827e-01 -9.64957416e-01 -6.81991339e-01
-3.22484136e-01 -1.10974622e+00 9.59707379e-01 -1.77044585e-01
-5.08764565e-01 -1.25324035e+00 -3.16647798e-01 8.18952084e-01
4.39794123e-01 -3.67215395e-01 1.49656665e+00 -5.34711897e-01
-8.99834633e-01 -6.91088215e-02 -3.80089104e-01 3.24200988e-01
-2.40512565e-01 -1.84130654e-01 -1.10578060e+00 7.91263655e-02
-1.80871382e-01 -1.00897443e+00 9.54021394e-01 3.76236662e-02
1.34914327e+00 -2.85212308e-01 -4.62450786e-03 8.02572906e-01
1.09107888e+00 -1.58651918e-01 5.62752604e-01 4.61189032e-01
6.33427978e-01 7.50737265e-02 4.68016237e-01 1.11316323e-01
8.95853996e-01 3.32869828e-01 1.66733041e-02 1.82225555e-01
-2.53657043e-01 -6.07930005e-01 6.73287392e-01 7.56115854e-01
-5.67453802e-02 -2.19899386e-01 -1.30969346e+00 6.75843358e-02
-1.96296990e+00 -5.62643588e-01 3.53771776e-01 1.61973321e+00
1.33023787e+00 8.90090644e-01 -3.87859285e-01 -8.02014917e-02
-3.06245744e-01 1.33287817e-01 -6.01328313e-01 -5.52259326e-01
-1.41503692e-01 3.08833897e-01 3.20858151e-01 7.32031286e-01
-8.60511601e-01 1.68910658e+00 6.92361450e+00 7.34319568e-01
-9.26933825e-01 -5.48749566e-02 4.20976251e-01 -3.01220655e-01
-6.37921512e-01 2.62491852e-01 -1.04116809e+00 9.26615372e-02
1.18970430e+00 1.07692992e-02 9.03846323e-01 7.25247204e-01
-9.96882915e-02 -1.96853936e-01 -1.41886091e+00 6.23401284e-01
1.40329748e-01 -1.35767317e+00 1.88515559e-01 -4.07073706e-01
7.13960588e-01 -5.96144125e-02 7.66919032e-02 1.13329709e+00
7.32964635e-01 -1.20308256e+00 9.19017196e-01 5.73953331e-01
6.92841649e-01 -3.94789129e-01 7.57459104e-01 3.86916220e-01
-9.28127527e-01 -2.18055695e-01 -2.56684516e-02 -4.54605222e-01
-1.04116626e-01 3.51991981e-01 -9.18607593e-01 2.43625075e-01
5.57772279e-01 5.85546255e-01 -1.00173640e+00 2.52503902e-02
-7.99056113e-01 8.27455699e-01 -2.48283207e-01 -1.91516757e-01
3.86101544e-01 3.82689238e-01 -9.45075881e-03 1.08961403e+00
-2.59223282e-01 -6.73293248e-02 5.41959591e-02 1.36634457e+00
-4.02239293e-01 -2.82147199e-01 -2.37843379e-01 -5.66704512e-01
5.04902959e-01 8.76083076e-01 -2.77064651e-01 -4.37607914e-01
-8.33041847e-01 5.94211876e-01 8.14509928e-01 4.27545309e-01
-1.10430884e+00 -1.34106383e-01 5.49314976e-01 -2.12080240e-01
-1.02837883e-01 -4.37154889e-01 -6.87365592e-01 -1.40241480e+00
3.42848390e-01 -1.25208032e+00 4.59600329e-01 -8.91810775e-01
-1.20786953e+00 3.15854341e-01 2.59773165e-01 -3.40462297e-01
-3.96953285e-01 -7.45862722e-01 -4.00246352e-01 7.96695054e-01
-1.74482346e+00 -1.50394690e+00 -1.03597924e-01 4.25885171e-01
5.31632364e-01 -5.31656861e-01 8.44826698e-01 1.05531022e-01
-6.61340952e-01 1.07395971e+00 -2.48800203e-01 2.04683885e-01
6.92213237e-01 -1.30775440e+00 7.14138985e-01 8.07128310e-01
6.53188601e-02 1.16718090e+00 4.67032820e-01 -7.10407674e-01
-1.90151525e+00 -8.86449993e-01 1.04100525e+00 -7.93964922e-01
1.05983222e+00 -7.48670995e-01 -8.45241845e-01 1.23023617e+00
-4.17876989e-03 -1.38759106e-01 4.17781651e-01 7.68660903e-01
-6.29353523e-01 -1.81210861e-01 -7.71900594e-01 1.14785933e+00
1.37883270e+00 -7.85693586e-01 -8.48196328e-01 3.74845833e-01
1.00829506e+00 -5.41472852e-01 -7.86068499e-01 5.03920436e-01
5.95775843e-01 -8.79655540e-01 1.12471604e+00 -9.18355703e-01
8.95326912e-01 8.47544521e-02 -2.57454276e-01 -8.64324987e-01
-4.75982465e-02 -5.83060443e-01 -6.57951415e-01 1.23418498e+00
6.83215439e-01 -6.30115807e-01 8.38695228e-01 8.94533515e-01
-2.32087914e-02 -1.01463747e+00 -4.49047953e-01 -7.27935135e-01
3.68545651e-01 -9.57900703e-01 7.51275718e-01 6.89660192e-01
3.37599009e-01 6.43162191e-01 -6.17811605e-02 6.18647151e-02
4.30971771e-01 1.07702442e-01 8.48548710e-01 -7.39171743e-01
-6.52595758e-01 -6.05744302e-01 1.83026027e-02 -1.08376384e+00
9.25023556e-01 -1.23707354e+00 -8.16868469e-02 -1.39975846e+00
8.43324289e-02 -5.79417229e-01 -3.97123158e-01 9.86444831e-01
-3.17056835e-01 -3.52230281e-01 -5.20450855e-03 2.78968969e-03
-8.59955549e-01 4.51497763e-01 1.05613542e+00 -3.50246519e-01
-1.57203078e-01 -3.20176035e-01 -9.53923762e-01 8.42356980e-01
6.49173617e-01 5.49197011e-02 -8.48738313e-01 -9.36062932e-01
7.61522889e-01 -1.13446839e-01 2.99887270e-01 -1.15102398e+00
2.97616720e-01 -2.77057767e-01 1.33337438e-01 -4.53292072e-01
2.40279928e-01 -5.35231352e-01 -4.18565184e-01 2.45395824e-01
-8.13022316e-01 8.91328380e-02 5.67393363e-01 1.48652121e-01
-1.67832226e-01 -2.00574487e-01 2.43428573e-01 -1.74303606e-01
-1.19996929e+00 -2.91543186e-01 -1.56563327e-01 3.83683026e-01
6.31483495e-01 3.40143323e-01 -5.89231372e-01 -1.97886318e-01
-3.04020882e-01 6.13239944e-01 1.85357541e-01 4.83924508e-01
5.27843177e-01 -1.15991199e+00 -5.58930814e-01 3.28949183e-01
2.56817102e-01 9.60874259e-02 -1.06138699e-01 6.69108093e-01
-6.82988882e-01 9.20510769e-01 1.45681038e-01 -3.47624838e-01
-9.60470974e-01 5.09307683e-01 2.92108566e-01 -6.64555252e-01
-6.70231640e-01 1.01840103e+00 -1.22488737e-01 -6.73926115e-01
4.82774019e-01 -8.95551562e-01 2.14185297e-01 -4.82129455e-01
3.89212817e-01 1.07511245e-02 1.79845735e-01 3.42635065e-01
-4.51910704e-01 1.28866315e-01 -2.54605055e-01 -1.62879646e-01
1.16453457e+00 1.97420403e-01 -2.99577177e-01 5.47801435e-01
7.98997402e-01 4.83131744e-02 -1.02696288e+00 -5.31089008e-01
3.97988915e-01 -2.12077126e-02 9.48230028e-02 -1.29597569e+00
-7.07553744e-01 6.91590548e-01 -1.34794608e-01 -2.72779942e-01
1.05817258e+00 -5.75556047e-02 8.69956017e-01 1.18864107e+00
4.35811013e-01 -1.01688516e+00 2.51343697e-01 1.04320288e+00
7.25508034e-01 -1.27325356e+00 1.81615114e-01 -1.69393346e-01
-6.47031486e-01 9.44465339e-01 8.27103138e-01 2.27922887e-01
3.91652524e-01 7.88820863e-01 1.19207829e-01 -1.75412059e-01
-1.52575970e+00 -4.73491708e-03 1.52520612e-01 5.82409389e-02
8.74288678e-01 9.58115049e-03 -1.32167980e-01 8.43936384e-01
-5.69433033e-01 5.03728509e-01 1.15991682e-01 1.18470466e+00
-1.89817194e-02 -1.18035805e+00 -2.13209093e-01 3.96037817e-01
-2.37954915e-01 -5.66850245e-01 -4.87283409e-01 8.69439542e-01
-1.27002120e-01 7.70783544e-01 -1.48624718e-01 -3.77566367e-01
3.05370212e-01 6.14679158e-01 7.56515026e-01 -5.72181821e-01
-4.69511181e-01 -4.75408047e-01 4.46378291e-01 -6.18317068e-01
-1.13308303e-01 -5.27956009e-01 -1.60347390e+00 -5.44300854e-01
1.36249140e-02 -1.17616206e-01 3.03269386e-01 1.18138742e+00
1.38933495e-01 9.29211855e-01 -1.11043006e-01 -7.33146295e-02
-9.60616887e-01 -7.94317722e-01 -7.67986551e-02 1.77938342e-01
1.51925176e-01 -4.81563270e-01 -1.13750376e-01 7.21178725e-02] | [9.647929191589355, 7.405823230743408] |
f618d48a-5925-42b7-a8c3-3e7a27a21d5e | promptonomyvit-multi-task-prompt-learning | 2212.04821 | null | https://arxiv.org/abs/2212.04821v2 | https://arxiv.org/pdf/2212.04821v2.pdf | PromptonomyViT: Multi-Task Prompt Learning Improves Video Transformers using Synthetic Scene Data | Action recognition models have achieved impressive results by incorporating scene-level annotations, such as objects, their relations, 3D structure, and more. However, obtaining annotations of scene structure for videos requires a significant amount of effort to gather and annotate, making these methods expensive to train. In contrast, synthetic datasets generated by graphics engines provide powerful alternatives for generating scene-level annotations across multiple tasks. In this work, we propose an approach to leverage synthetic scene data for improving video understanding. We present a multi-task prompt learning approach for video transformers, where a shared video transformer backbone is enhanced by a small set of specialized parameters for each task. Specifically, we add a set of ``task prompts'', each corresponding to a different task, and let each prompt predict task-related annotations. This design allows the model to capture information shared among synthetic scene tasks as well as information shared between synthetic scene tasks and a real video downstream task throughout the entire network. We refer to this approach as ``Promptonomy'', since the prompts model task-related structure. We propose the PromptonomyViT model (PViT), a video transformer that incorporates various types of scene-level information from synthetic data using the ``Promptonomy'' approach. PViT shows strong performance improvements on multiple video understanding tasks and datasets. | ['Amir Globerson', 'Trevor Darrell', 'Ariel Shamir', 'Leonid Karlinsky', 'Assaf Arbelle', 'Elad Ben-Avraham', 'Ofir Abramovich', 'Roei Herzig'] | 2022-12-08 | null | null | null | null | ['video-understanding'] | ['computer-vision'] | [ 5.50193548e-01 1.21849090e-01 5.56616522e-02 -5.22762716e-01
-8.15624833e-01 -6.30795479e-01 6.88850105e-01 -1.31598324e-01
-1.60578102e-01 4.61459219e-01 3.28372627e-01 7.74000399e-03
1.03287056e-01 -5.16817093e-01 -1.09003687e+00 -4.05945271e-01
1.89698651e-01 4.51702476e-01 6.80063665e-01 -2.76924518e-04
4.59039733e-02 1.81041569e-01 -1.59824300e+00 9.34022427e-01
6.23421609e-01 9.56431091e-01 7.60967970e-01 8.27633977e-01
-1.06437132e-01 1.60283756e+00 -5.75745523e-01 -2.57785201e-01
2.11035863e-01 -2.59257346e-01 -1.07539284e+00 5.22435546e-01
6.40939951e-01 -6.43679380e-01 -4.64529693e-01 5.90208948e-01
1.09973498e-01 2.94471532e-01 3.73043299e-01 -1.61492562e+00
8.70792270e-02 4.58888978e-01 -2.91062713e-01 2.02488884e-01
6.77474141e-01 4.06323999e-01 1.02122939e+00 -8.04324210e-01
8.00443590e-01 1.18575597e+00 4.88025069e-01 4.20239806e-01
-1.19461679e+00 -5.05083144e-01 5.02148509e-01 4.57167327e-01
-1.10225081e+00 -6.59305990e-01 8.77987802e-01 -7.58003414e-01
9.78299618e-01 1.27520978e-01 7.76324213e-01 1.46493447e+00
-3.38750720e-01 1.27785885e+00 7.75691271e-01 -6.72811121e-02
1.89407617e-01 -6.76768795e-02 -6.68145418e-02 7.39637434e-01
-2.18042731e-01 -6.94054067e-02 -9.01806712e-01 2.06430435e-01
9.22162414e-01 -2.44452991e-02 -3.16310704e-01 -4.78579730e-01
-1.48110175e+00 3.05861354e-01 1.61611170e-01 -8.20352063e-02
-3.62844676e-01 3.83122414e-01 5.99787951e-01 1.46972835e-01
4.36263084e-01 3.09433967e-01 -6.57031536e-01 -4.11075056e-01
-7.47213960e-01 2.23699659e-01 7.36709833e-01 1.29141951e+00
8.34213257e-01 1.12939194e-01 -4.73276973e-01 6.82562113e-01
1.64520696e-01 1.38448268e-01 3.78628038e-02 -1.54256749e+00
7.50376999e-01 6.12664342e-01 1.33095697e-01 -8.00929666e-01
-3.94165307e-01 -1.13954611e-01 -6.08006358e-01 -8.76965970e-02
5.25934994e-01 -6.94953799e-02 -8.55652571e-01 1.73927534e+00
3.58054221e-01 6.91979647e-01 -2.74592871e-03 7.85656869e-01
1.00880671e+00 6.93711281e-01 3.29582185e-01 1.60390809e-01
1.32376099e+00 -1.47849703e+00 -4.26643908e-01 -3.17264318e-01
9.53978360e-01 -5.91435730e-01 1.26819062e+00 4.37472790e-01
-9.63191032e-01 -8.81540954e-01 -4.73423392e-01 -2.30324060e-01
-1.52255222e-01 1.76911190e-01 5.99959612e-01 1.91320121e-01
-1.18995130e+00 2.77441382e-01 -8.18907082e-01 -4.43338513e-01
5.70286512e-01 2.93667633e-02 -4.58511949e-01 -3.97757590e-01
-7.36023724e-01 5.02586305e-01 4.53267306e-01 -2.27947369e-01
-1.53313088e+00 -9.51383770e-01 -1.07520437e+00 7.17632025e-02
8.59928012e-01 -9.75611329e-01 1.64805901e+00 -1.02964902e+00
-1.24030602e+00 6.85520649e-01 -2.23244399e-01 -1.95066348e-01
5.08706033e-01 -3.89874160e-01 4.25666757e-02 4.60564077e-01
2.27470234e-01 1.00086546e+00 8.93739998e-01 -1.34180701e+00
-8.45586240e-01 -4.49033491e-02 6.72000289e-01 3.78813028e-01
-6.22654893e-02 2.00117409e-01 -8.77875924e-01 -6.21310234e-01
-1.96152672e-01 -8.40016901e-01 -2.43552223e-01 2.58785427e-01
-4.68773872e-01 -1.78347111e-01 1.15718496e+00 -5.84043860e-01
7.81651497e-01 -2.19443035e+00 3.52876842e-01 -3.06098282e-01
4.07572538e-01 8.24565366e-02 -3.86072457e-01 3.29156429e-01
-1.97730109e-01 4.12953869e-02 -2.19355151e-02 -7.03632474e-01
-8.03232193e-02 4.54909086e-01 -1.46468446e-01 -1.04566678e-01
2.89291352e-01 8.91283631e-01 -1.14194870e+00 -5.20165086e-01
4.98183757e-01 2.54702777e-01 -8.98500264e-01 4.76779401e-01
-7.12046027e-01 7.61176586e-01 -6.81734502e-01 6.40175760e-01
1.93205804e-01 -6.29120648e-01 7.79630765e-02 -6.35942101e-01
5.94430938e-02 3.12553167e-01 -1.02736712e+00 1.96034694e+00
-4.92537707e-01 7.57250845e-01 1.21886581e-01 -1.10180259e+00
5.73339105e-01 5.79633117e-01 8.29329252e-01 -5.29841185e-01
-1.77929193e-01 -3.35284203e-01 -3.70150208e-01 -8.21698248e-01
3.93022776e-01 1.45083800e-01 -2.19852418e-01 6.48259640e-01
1.75185710e-01 -3.21185648e-01 3.48426551e-01 6.59929812e-01
1.52705991e+00 7.73095012e-01 1.27649754e-01 2.26333648e-01
2.77516782e-01 3.96044672e-01 5.87717116e-01 7.43101537e-01
-2.04452708e-01 5.64852118e-01 6.50487423e-01 -6.15583599e-01
-1.09126222e+00 -7.80030549e-01 4.91660863e-01 1.37900257e+00
1.76248387e-01 -6.96602285e-01 -7.38569975e-01 -8.96791458e-01
-2.68116117e-01 5.62864482e-01 -4.55681711e-01 3.05137951e-02
-6.13124669e-01 -2.32017413e-01 6.10029459e-01 8.12977076e-01
7.40087092e-01 -1.06631923e+00 -7.56088316e-01 8.42264667e-02
-7.41133213e-01 -1.78384340e+00 -4.80806828e-01 1.26518399e-01
-7.58168042e-01 -1.41053629e+00 -3.33413154e-01 -5.24429977e-01
6.52960062e-01 6.22342050e-01 1.34074056e+00 -3.19600031e-02
-1.18155569e-01 7.31893897e-01 -6.54585600e-01 -8.04606974e-02
-2.84399480e-01 -1.58035710e-01 -1.09150775e-01 2.03101262e-01
1.07419519e-02 -4.49265599e-01 -3.85108113e-01 5.66083848e-01
-9.45492566e-01 8.00771296e-01 4.45416987e-01 4.61454302e-01
5.92410684e-01 -8.54789615e-02 1.21946514e-01 -9.34214294e-01
1.29094377e-01 -3.67532313e-01 -3.96872669e-01 2.33814478e-01
2.51934052e-01 -1.13274390e-02 6.55238807e-01 -4.70951796e-01
-1.34904134e+00 4.21145707e-01 1.07697770e-01 -8.79509330e-01
-5.03284514e-01 3.09209287e-01 -4.06398267e-01 8.04178193e-02
5.60554564e-01 1.91299751e-01 -3.55725646e-01 -4.90991294e-01
4.29789156e-01 2.35209197e-01 6.51002049e-01 -8.96227658e-01
7.17726231e-01 4.90293533e-01 -1.66763484e-01 -7.24776745e-01
-1.34187019e+00 -6.05605543e-01 -8.62157166e-01 -6.41946018e-01
1.05143499e+00 -1.11760211e+00 -6.34120643e-01 5.32117426e-01
-1.48295307e+00 -1.02612746e+00 -4.41019207e-01 4.30726677e-01
-9.59318161e-01 2.28555933e-01 -4.39747632e-01 -3.79818201e-01
1.67232215e-01 -1.36623585e+00 1.56021821e+00 -1.11675493e-01
-2.26856619e-01 -1.03653026e+00 -1.84745371e-01 7.61497796e-01
8.19162726e-02 2.58102864e-01 7.23442793e-01 -5.63943326e-01
-1.07069278e+00 8.76384228e-02 -6.03410602e-01 3.21501464e-01
9.97648761e-02 -1.07305422e-01 -1.06073380e+00 -6.76281285e-03
-9.33898166e-02 -6.56728864e-01 5.43011844e-01 1.32701203e-01
1.54571187e+00 -2.61778474e-01 -4.35461283e-01 8.08841884e-01
1.10192025e+00 1.30162895e-01 5.70824742e-01 1.46332785e-01
1.19715130e+00 6.62017345e-01 7.38277197e-01 4.00614887e-01
7.25378871e-01 9.81412888e-01 4.85573471e-01 -2.18276545e-01
-3.53504241e-01 -4.29447204e-01 5.54408848e-01 6.60149693e-01
-1.19880617e-01 -4.03925717e-01 -9.47182357e-01 3.47858131e-01
-2.00924158e+00 -1.00804031e+00 -2.96582460e-01 1.79314125e+00
5.76656401e-01 4.77056950e-04 4.88590449e-02 -3.15383337e-02
4.14803892e-01 4.05420870e-01 -5.16033649e-01 -1.66963823e-02
7.08587617e-02 -1.87178120e-01 1.65151089e-01 4.13692713e-01
-1.03801656e+00 1.25375807e+00 6.00957537e+00 6.88766181e-01
-7.76769042e-01 1.46619767e-01 3.90069127e-01 -8.33818391e-02
-1.89042851e-01 2.16875374e-01 -6.39831543e-01 3.12434644e-01
7.39244103e-01 -4.61406931e-02 3.01537573e-01 8.66528153e-01
6.05118394e-01 -2.30085239e-01 -1.49885905e+00 1.02439523e+00
2.08612129e-01 -1.34789968e+00 2.61293769e-01 -1.02235943e-01
6.01112247e-01 2.06756387e-02 -4.13696885e-01 3.84072065e-01
5.50978303e-01 -5.48557878e-01 8.55802298e-01 5.26502609e-01
9.17750001e-01 -5.66881038e-02 2.47347072e-01 5.29123664e-01
-1.55186784e+00 -3.56907696e-02 1.23280607e-01 -1.00778729e-01
4.77110565e-01 2.61114359e-01 -8.79174292e-01 4.88874137e-01
7.47788668e-01 1.11500430e+00 -6.06219888e-01 1.03117955e+00
-3.83627236e-01 4.75834101e-01 -2.95414299e-01 4.76519287e-01
2.60206878e-01 -2.39815805e-02 3.83318186e-01 1.11058068e+00
1.46179691e-01 1.21998087e-01 6.56001091e-01 5.74908197e-01
-6.59953952e-02 -1.89068094e-01 -8.63113999e-01 1.03687264e-01
3.81144464e-01 1.14461625e+00 -6.81707382e-01 -7.27354646e-01
-5.89352071e-01 1.12778521e+00 2.51718700e-01 5.35489440e-01
-9.83757138e-01 1.21255413e-01 6.83797836e-01 1.55021444e-01
1.41941652e-01 -2.47791216e-01 6.04146235e-02 -1.22354960e+00
3.60079445e-02 -9.61796165e-01 3.95256102e-01 -1.42814863e+00
-7.89173245e-01 4.21371847e-01 3.88123482e-01 -1.33029544e+00
-2.56345868e-01 -5.99318683e-01 -5.17986596e-01 3.81476462e-01
-1.36241472e+00 -1.32240248e+00 -9.17764187e-01 9.09959972e-01
1.29063785e+00 6.54490888e-02 5.69937229e-01 2.74357796e-01
-6.22198403e-01 1.32464901e-01 -4.81033027e-01 1.23141281e-01
7.45448768e-01 -1.09413576e+00 4.19334769e-01 7.44275987e-01
2.79579818e-01 -2.87832636e-02 5.47614694e-01 -5.84961534e-01
-1.29659677e+00 -1.43455505e+00 6.03947043e-01 -8.18523109e-01
6.75729096e-01 -4.81311679e-01 -7.21320510e-01 1.18309724e+00
-5.14746308e-02 5.88218123e-02 6.08023405e-01 1.18083386e-02
-4.07130063e-01 -8.91501829e-02 -6.50496364e-01 5.69535971e-01
1.63542318e+00 -7.00183153e-01 -4.35898960e-01 8.41695368e-01
9.94576514e-01 -4.95140433e-01 -7.69327998e-01 4.49648172e-01
3.54908139e-01 -1.02846527e+00 9.92598653e-01 -6.67539537e-01
7.45715380e-01 -3.68057966e-01 -2.03910023e-01 -1.19863677e+00
-2.08803147e-01 -5.62924623e-01 -2.27040455e-01 1.00482881e+00
2.56101429e-01 9.49927494e-02 9.43692505e-01 6.17777288e-01
-5.97160578e-01 -5.15701354e-01 -4.95235354e-01 -6.27289295e-01
-6.98606849e-01 -1.01273143e+00 4.21003103e-01 9.50692534e-01
-2.33728573e-01 5.40163517e-01 -5.62554598e-01 1.07593440e-01
4.39320207e-01 -1.44110173e-01 1.34371209e+00 -1.03875089e+00
-4.63708460e-01 -1.01641439e-01 -2.39292696e-01 -1.50167727e+00
2.53181726e-01 -7.15212822e-01 9.81435403e-02 -1.71027923e+00
2.91791826e-01 -6.50965691e-01 1.07614771e-01 7.20042944e-01
-1.46976233e-01 2.00094074e-01 3.50918114e-01 3.16969067e-01
-1.19622052e+00 4.67617303e-01 1.45129073e+00 -1.65261954e-01
-5.68524040e-02 -7.75114149e-02 -4.68700320e-01 1.04882240e+00
5.32705247e-01 -3.51066381e-01 -8.34447622e-01 -1.04188466e+00
5.53752035e-02 2.72327572e-01 7.30561912e-01 -1.21471763e+00
3.46252561e-01 -3.73300284e-01 1.68992922e-01 -4.36806887e-01
7.99733162e-01 -8.28389585e-01 4.20338601e-01 -3.81228290e-02
-2.91731119e-01 -1.54577389e-01 1.73171312e-01 6.82292879e-01
-3.20080101e-01 -5.56278392e-04 3.35566610e-01 -3.79416227e-01
-1.22132027e+00 5.14695764e-01 -3.19878012e-01 8.69056433e-02
1.30797744e+00 -4.70110685e-01 -4.89166647e-01 -4.93741453e-01
-9.10187066e-01 4.46170330e-01 5.72242618e-01 5.65888047e-01
5.62680721e-01 -1.06474948e+00 -5.27539015e-01 5.30377477e-02
4.11671042e-01 4.45450902e-01 4.29969490e-01 7.72752047e-01
-4.32863235e-01 2.30616033e-01 -9.47325453e-02 -9.45517540e-01
-1.32023990e+00 4.60249752e-01 2.01040357e-01 -3.29940140e-01
-8.13012779e-01 1.02669024e+00 9.12343740e-01 -2.68694967e-01
2.56139606e-01 -3.18436533e-01 -1.78572927e-02 6.99929371e-02
4.30945873e-01 1.66536629e-01 -1.67887792e-01 -7.08324134e-01
-8.25596154e-02 4.91172135e-01 9.53563675e-03 -3.24540026e-02
1.19633579e+00 -9.95060652e-02 1.74829632e-01 3.91747832e-01
9.34444845e-01 -4.53910768e-01 -1.86295319e+00 -4.35633183e-01
-6.19361699e-02 -5.82000077e-01 -1.95076555e-01 -7.10331440e-01
-1.09403300e+00 8.45476091e-01 -3.45179774e-02 6.56406507e-02
1.19536579e+00 2.73120463e-01 5.44037163e-01 5.38301945e-01
6.90642655e-01 -1.02504504e+00 7.13838637e-01 5.25084972e-01
7.75894344e-01 -1.16963363e+00 -2.23410308e-01 -7.61456430e-01
-8.32405031e-01 7.87719965e-01 1.02214050e+00 2.87201047e-01
2.93323040e-01 3.43602598e-01 -7.02836663e-02 -3.27217996e-01
-1.09808302e+00 -2.54033357e-01 8.41468479e-03 8.36266041e-01
1.31378174e-02 -2.36317143e-01 5.42970955e-01 3.34024876e-01
6.97367489e-02 2.92512923e-01 3.99518728e-01 9.40645635e-01
-3.39429110e-01 -1.17372513e+00 -1.93652213e-01 4.99219239e-01
-7.05550388e-02 -1.92689747e-02 -3.44225019e-01 6.12662494e-01
2.28529036e-01 8.80630195e-01 1.61355719e-01 -4.98829544e-01
4.59218651e-01 1.86232124e-02 4.01385546e-01 -1.02519345e+00
-3.49788666e-01 -4.70174700e-02 4.87811923e-01 -1.11688113e+00
-7.11924851e-01 -5.68638563e-01 -9.66093063e-01 -7.83638060e-02
7.67965093e-02 -1.06326535e-01 4.32758987e-01 1.21961641e+00
4.14221644e-01 9.29116189e-01 4.39195126e-01 -1.20286119e+00
1.51616037e-01 -8.58314037e-01 -1.90236464e-01 6.88875973e-01
1.22824172e-02 -7.94501007e-01 1.23289712e-02 8.63032937e-01] | [8.90475082397461, 0.5838494896888733] |
a9bd6fd8-4eaf-40ac-91c7-7423c59183e8 | glad-group-anomaly-detection-in-social-media | 1410.1940 | null | http://arxiv.org/abs/1410.1940v1 | http://arxiv.org/pdf/1410.1940v1.pdf | GLAD: Group Anomaly Detection in Social Media Analysis- Extended Abstract | Traditional anomaly detection on social media mostly focuses on individual
point anomalies while anomalous phenomena usually occur in groups. Therefore it
is valuable to study the collective behavior of individuals and detect group
anomalies. Existing group anomaly detection approaches rely on the assumption
that the groups are known, which can hardly be true in real world social media
applications. In this paper, we take a generative approach by proposing a
hierarchical Bayes model: Group Latent Anomaly Detection (GLAD) model. GLAD
takes both pair-wise and point-wise data as input, automatically infers the
groups and detects group anomalies simultaneously. To account for the dynamic
properties of the social media data, we further generalize GLAD to its dynamic
extension d-GLAD. We conduct extensive experiments to evaluate our models on
both synthetic and real world datasets. The empirical results demonstrate that
our approach is effective and robust in discovering latent groups and detecting
group anomalies. | ['Yu', 'QI', 'Xinran He', 'Yan Liu'] | 2014-10-07 | null | null | null | null | ['group-anomaly-detection'] | ['methodology'] | [-7.41356611e-02 -1.38212740e-01 2.40875646e-01 -3.25297177e-01
-4.20583151e-02 -3.36549997e-01 8.10265064e-01 8.29615235e-01
2.22205520e-01 3.45522076e-01 4.49608415e-02 -2.14071706e-01
-2.16277391e-01 -1.04102695e+00 -3.41413498e-01 -6.04231238e-01
-7.65759706e-01 4.67898458e-01 5.54292023e-01 6.41470961e-03
2.88134485e-01 2.34117389e-01 -1.28148651e+00 4.24832962e-02
1.05104303e+00 4.64826673e-01 -5.13714671e-01 7.46341825e-01
-6.80546984e-02 7.52023220e-01 -7.47210622e-01 -2.07151145e-01
3.17919582e-01 -5.35629928e-01 -3.65667343e-01 6.21876538e-01
1.32973939e-01 -4.64532733e-01 4.87085152e-03 1.07277763e+00
1.58791810e-01 1.08047254e-01 6.90810800e-01 -1.80175304e+00
-1.54665232e-01 4.87457782e-01 -9.76101518e-01 5.27858973e-01
5.24626374e-01 -8.58990327e-02 1.24344039e+00 -3.80897224e-01
6.01393208e-02 1.28866625e+00 7.42430389e-01 2.25273445e-01
-1.39785659e+00 -5.80035746e-01 7.18506217e-01 -2.57114992e-02
-1.18653452e+00 -1.33051693e-01 8.54286253e-01 -7.39857256e-01
6.27765417e-01 3.22844863e-01 7.28136301e-01 8.85087073e-01
1.74007431e-01 7.77900040e-01 7.93757796e-01 -2.98636109e-01
4.27636087e-01 -3.91448081e-01 5.11141598e-01 4.43637818e-01
7.43676722e-01 -2.71908790e-01 -4.45015520e-01 -1.00360334e+00
5.25030792e-01 7.13140309e-01 1.20896742e-01 -2.64739871e-01
-1.03484857e+00 8.92837167e-01 -8.01824182e-02 3.68829221e-01
-5.05752325e-01 -4.38458286e-03 3.66624892e-01 6.19099855e-01
1.13246858e+00 2.01510880e-02 9.22806654e-03 -5.47343306e-02
-8.25952768e-01 4.63575661e-01 7.75166333e-01 6.54834926e-01
6.46881938e-01 -1.66060910e-01 1.39689639e-01 7.81283438e-01
6.27422154e-01 2.07957119e-01 4.61023867e-01 -4.29247975e-01
1.58889607e-01 1.04863954e+00 -5.85499033e-02 -1.57574809e+00
-3.15791726e-01 -1.63389146e-01 -8.35876107e-01 -3.07302564e-01
4.29658860e-01 -1.84398845e-01 -5.85469782e-01 1.55550516e+00
5.48925281e-01 8.24095428e-01 -1.11537464e-01 1.66064560e-01
4.06275690e-01 5.40652752e-01 1.21954158e-02 -4.91499662e-01
7.06216395e-01 -5.09597421e-01 -7.64522195e-01 2.15840545e-02
9.82630372e-01 -4.50794399e-01 6.31693959e-01 1.85421780e-01
-7.44383574e-01 -1.47277147e-01 -6.83518112e-01 8.05527508e-01
-1.18023835e-01 -6.88358784e-01 5.02992213e-01 5.69168270e-01
-7.91326702e-01 6.48495972e-01 -1.23812759e+00 -6.24314964e-01
4.13316399e-01 1.48105487e-01 -1.49005771e-01 -2.80957390e-02
-7.47033894e-01 -2.49848410e-01 3.35212231e-01 -1.32841349e-01
-5.58766305e-01 -3.79781246e-01 -6.66861951e-01 -2.09557876e-01
5.13720512e-01 -4.26017582e-01 1.01985598e+00 -6.90680802e-01
-8.00728083e-01 5.40168285e-01 -3.88173014e-01 -6.06701791e-01
6.71777427e-01 -2.13911936e-01 -7.23337889e-01 -8.38653967e-02
2.70773053e-01 -3.00882846e-01 7.43581414e-01 -1.09033537e+00
-8.15475404e-01 -4.82167155e-01 -2.69602656e-01 -2.88060397e-01
-5.26667595e-01 1.15015104e-01 5.23703434e-02 -7.46370852e-01
5.96802354e-01 -9.83307898e-01 -3.48008066e-01 -5.04067183e-01
-8.04278910e-01 -5.73957860e-01 1.26154077e+00 -4.15653586e-01
1.57791090e+00 -2.36340427e+00 -3.73396605e-01 8.04975808e-01
7.09685266e-01 -1.16937712e-01 3.13536823e-01 7.72371113e-01
-5.62194213e-02 2.19752967e-01 -3.01348627e-01 -6.24704599e-01
-1.91707108e-02 4.62041199e-01 -5.71606874e-01 6.05838954e-01
-4.89371121e-02 4.81449604e-01 -9.09059644e-01 -2.48315394e-01
-2.41232589e-01 -2.09848270e-01 -9.87649918e-01 3.25753748e-01
-8.56827646e-02 4.73659486e-01 -6.48918808e-01 6.62036300e-01
5.78108490e-01 -4.73282158e-01 1.85573190e-01 6.95538461e-01
-7.28937506e-04 -2.36843340e-02 -1.25293791e+00 7.74322689e-01
4.00303632e-01 4.73803312e-01 -3.76744390e-01 -9.97480869e-01
1.03551888e+00 2.83566684e-01 9.30689096e-01 -1.99428257e-02
-2.33310610e-01 1.75063640e-01 2.26389617e-01 -3.91977072e-01
2.24234492e-01 1.53985724e-01 -2.20653594e-01 1.07257903e+00
-3.09526682e-01 6.28805220e-01 3.33600938e-01 3.93822908e-01
1.50023699e+00 -5.29012799e-01 4.62346077e-01 -1.59934491e-01
4.03630018e-01 -4.30330336e-01 7.05082297e-01 1.08290243e+00
-1.14103593e-01 4.86212373e-01 1.08647525e+00 -5.96682727e-01
-7.34234154e-01 -1.08611238e+00 1.37191385e-01 1.01258981e+00
-1.47460559e-02 -9.47792411e-01 -5.39178371e-01 -9.92323637e-01
2.13336214e-01 5.23019433e-01 -6.51073098e-01 -1.27729088e-01
-2.95179456e-01 -1.34614241e+00 1.18236311e-01 1.35498509e-01
3.90010297e-01 -8.44829202e-01 -1.21853597e-01 2.44332612e-01
-6.93461597e-02 -9.79762495e-01 -1.41816482e-01 -5.10705352e-01
-9.20383751e-01 -1.31839848e+00 2.50843577e-02 -3.62139046e-01
9.72642362e-01 3.11193764e-01 9.27254081e-01 4.69191909e-01
-9.30420011e-02 5.53037107e-01 -5.90219259e-01 -8.10666323e-01
-4.66888934e-01 -1.02721177e-01 4.76319075e-01 7.14646995e-01
7.72437692e-01 -1.07075572e+00 -4.25824940e-01 5.51629603e-01
-8.51877928e-01 -3.01586747e-01 8.47395584e-02 3.66081715e-01
3.68394256e-01 6.23724341e-01 5.97948790e-01 -1.21574235e+00
8.51579905e-01 -9.88591373e-01 -5.51265240e-01 -9.00793001e-02
-5.45262456e-01 -4.79419231e-01 1.73203960e-01 -3.87335062e-01
-4.92447466e-01 -3.94315839e-01 9.52557623e-02 -2.43826285e-01
-5.36933243e-01 4.27487433e-01 -9.43679810e-02 3.36234480e-01
3.82696867e-01 2.09091574e-01 6.45439550e-02 -4.46553886e-01
-4.14725244e-01 5.74531734e-01 7.17110783e-02 -3.87001514e-01
9.67125773e-01 7.14195371e-01 -4.60694320e-02 -1.05941427e+00
-6.97351515e-01 -8.20314348e-01 -7.96435118e-01 -2.87682265e-01
5.61314404e-01 -7.37094641e-01 -4.96079266e-01 7.58982897e-01
-7.42758214e-01 -2.43674681e-01 -1.70456469e-01 3.14621240e-01
-1.87322229e-01 6.72365308e-01 -4.93034333e-01 -1.05461276e+00
4.20564786e-02 -5.74528456e-01 8.89949024e-01 -1.84645072e-01
-5.34171760e-01 -1.50189459e+00 4.43720639e-01 -4.33282740e-02
1.61900669e-01 7.60232091e-01 6.94772661e-01 -1.54825830e+00
-4.96881962e-01 -7.16839612e-01 3.17294858e-02 6.55698329e-02
6.29344821e-01 2.30433822e-01 -6.93835795e-01 -3.78232121e-01
-1.05881386e-01 4.56198931e-01 5.69106638e-01 2.01831117e-01
1.26156330e+00 -5.55442691e-01 -5.65780222e-01 1.61473989e-01
9.13977087e-01 6.42307522e-03 2.61475623e-01 4.51041423e-02
9.40643370e-01 6.72480762e-01 4.69180495e-01 8.27166259e-01
4.74496305e-01 4.24489975e-01 3.32257599e-01 9.05418396e-02
6.84671521e-01 -3.77169639e-01 5.24451613e-01 1.13830471e+00
-1.30266786e-01 -4.81881857e-01 -1.18824327e+00 5.44964552e-01
-2.27052450e+00 -1.17612052e+00 -5.57142019e-01 2.17728090e+00
2.31383726e-01 3.66321027e-01 7.42647886e-01 4.87188578e-01
9.31558967e-01 1.14640243e-01 -2.21742451e-01 -1.09817507e-02
7.12751001e-02 -4.42956775e-01 5.26454747e-02 3.40124339e-01
-1.18116868e+00 4.45476085e-01 6.52437973e+00 1.93988219e-01
-6.80974007e-01 -3.14792432e-02 4.40729588e-01 -6.06316049e-03
-1.63279936e-01 7.76490569e-02 -7.21043229e-01 7.45585918e-01
8.63905072e-01 -4.53197896e-01 -1.23394318e-01 5.85448802e-01
3.18549097e-01 1.24208376e-01 -1.15491760e+00 8.26509118e-01
5.83873950e-02 -6.56834841e-01 5.00446409e-02 7.56739259e-01
8.40818942e-01 -7.29730725e-02 -1.57752305e-01 1.00170486e-01
5.58739960e-01 -5.57427943e-01 -1.83371194e-02 6.78405643e-01
-1.75288305e-01 -8.64136755e-01 7.98203707e-01 6.91205144e-01
-9.69497681e-01 -3.16888005e-01 -1.93400942e-02 -3.71994197e-01
2.44947255e-01 9.03999448e-01 -9.85948026e-01 3.59170586e-01
7.57441700e-01 1.35718453e+00 -9.40346122e-01 1.09247267e+00
9.31460634e-02 1.18559098e+00 -5.10029614e-01 2.17555568e-01
1.24462530e-01 -2.84845322e-01 1.04590368e+00 7.32453227e-01
4.44768131e-01 -1.55483454e-01 6.73866034e-01 5.10352194e-01
3.29579473e-01 1.87416285e-01 -8.62766445e-01 -2.59417266e-01
2.69439787e-01 9.01296437e-01 -9.80356395e-01 -2.17786014e-01
-5.54391205e-01 6.12086475e-01 -6.64185826e-03 2.50260562e-01
-4.68682647e-01 -5.65524250e-02 7.06922710e-01 7.07138479e-01
3.22590093e-03 -4.95864868e-01 1.87047139e-01 -1.42013347e+00
-2.34035160e-02 -7.23786056e-01 8.74375820e-01 -6.76079616e-02
-1.88699293e+00 2.57328302e-01 1.97341576e-01 -1.44534194e+00
-6.35622442e-01 -1.06737211e-01 -1.25738895e+00 2.56873071e-01
-9.58196104e-01 -8.81420314e-01 -4.71508265e-01 8.67551565e-01
2.59764105e-01 -3.53188485e-01 7.00888336e-01 1.96843341e-01
-8.86607409e-01 5.44764340e-01 8.03116933e-02 3.52090091e-01
4.88686174e-01 -1.34909141e+00 7.56619096e-01 1.19743848e+00
1.89143687e-01 6.37970805e-01 8.64146411e-01 -1.03229284e+00
-6.36252940e-01 -1.26928568e+00 6.82972848e-01 -5.55724621e-01
1.02065289e+00 -5.89973092e-01 -1.35247815e+00 1.00778961e+00
-2.48993129e-01 2.72052169e-01 1.21703935e+00 4.42811430e-01
-9.80998725e-02 1.03210919e-01 -9.89085078e-01 6.10215664e-01
1.19133461e+00 -2.22226441e-01 -4.32929963e-01 5.91691434e-01
5.21839023e-01 2.09365457e-01 -7.72932768e-01 4.96325701e-01
5.68660535e-02 -1.17528284e+00 5.35235584e-01 -6.46887600e-01
4.95428815e-02 -3.93247515e-01 -9.83337983e-02 -1.22001791e+00
-2.22527355e-01 -6.21239960e-01 -5.87399364e-01 1.45462751e+00
7.14178756e-02 -1.19621575e+00 6.47759020e-01 4.28815186e-01
2.18367547e-01 -3.76054227e-01 -6.77261233e-01 -7.19406188e-01
-4.66797054e-01 -4.99772906e-01 7.78111696e-01 1.27387857e+00
8.79680514e-02 -1.00564370e-02 -3.41832727e-01 6.49018705e-01
9.11764383e-01 1.41390577e-01 1.22971165e+00 -1.99162722e+00
-4.44667190e-01 -3.06641638e-01 -1.01842761e+00 -5.24762690e-01
1.92049742e-02 -4.90133375e-01 -3.21169853e-01 -1.14864469e+00
9.55240428e-03 -4.62893963e-01 -3.60865921e-01 3.10548663e-01
-2.45373920e-01 2.54581034e-01 -3.09902340e-01 5.60517967e-01
-1.00639033e+00 3.87224168e-01 4.44317013e-01 2.72727400e-01
-6.03394687e-01 5.45213640e-01 -5.65350771e-01 1.23628426e+00
8.53464305e-01 -4.93465990e-01 -4.00039256e-01 1.83720633e-01
4.58291173e-01 -4.23983961e-01 4.39923286e-01 -1.01710761e+00
1.53183505e-01 -1.86135039e-01 3.25956754e-02 -6.57041728e-01
-1.85679168e-01 -7.07449675e-01 7.80673139e-03 4.57757413e-01
-9.05899256e-02 3.43116581e-01 -1.46070778e-01 1.10713577e+00
-3.61809105e-01 2.18451545e-01 3.70382726e-01 1.14325881e-01
-4.04132992e-01 6.40649736e-01 -5.14134645e-01 -6.87680170e-02
1.37170446e+00 -8.82413834e-02 -6.13906160e-02 -7.22008467e-01
-1.00032222e+00 5.38910151e-01 4.64985520e-01 4.66178566e-01
5.40064871e-01 -1.45664334e+00 -8.54104996e-01 5.97459495e-01
4.92672116e-01 7.32946694e-02 2.17785791e-01 9.42595601e-01
-3.62416178e-01 -2.48478174e-01 9.05046910e-02 -1.04541922e+00
-1.64251602e+00 4.48182672e-01 1.75442144e-01 -1.93082243e-01
-8.30620468e-01 3.61555934e-01 1.90182537e-01 -2.92314321e-01
2.79235020e-02 9.71411467e-02 -5.03000498e-01 1.83371857e-01
7.63450980e-01 5.01710415e-01 -2.88655043e-01 -7.15287268e-01
-4.41508144e-02 2.75969207e-01 -4.58942592e-01 1.79823980e-01
1.41247737e+00 -2.89305419e-01 -5.22402942e-01 1.01014209e+00
7.14073122e-01 3.85681063e-01 -9.21041489e-01 -6.26871943e-01
4.02552247e-01 -6.95735157e-01 -3.83704573e-01 2.85317093e-01
-7.38106668e-01 6.21061623e-01 2.15499237e-01 1.21294868e+00
1.06602156e+00 1.60712183e-01 5.82330644e-01 1.98445007e-01
2.37765476e-01 -8.44332397e-01 3.10913801e-01 3.07455689e-01
5.34950614e-01 -1.12752187e+00 1.07032336e-01 -6.40997767e-01
-2.06352279e-01 7.38305271e-01 5.80877304e-01 -4.57470208e-01
1.15451777e+00 -1.95406228e-01 -2.29523301e-01 -3.41889262e-01
-7.99930871e-01 8.52201134e-03 1.87696829e-01 3.17777723e-01
2.06572503e-01 7.31435269e-02 -1.39069483e-01 4.33706045e-01
-2.81598598e-01 -5.05010307e-01 6.37238443e-01 1.17488110e+00
-4.69270200e-01 -1.13846099e+00 -4.78994817e-01 8.44089448e-01
-7.62489080e-01 3.69464636e-01 -5.32179832e-01 5.67878485e-01
9.52317044e-02 9.58107352e-01 6.03591979e-01 -3.61973852e-01
2.51512885e-01 2.10758656e-01 -1.88115194e-01 -8.27024519e-01
-1.39153779e-01 1.81325734e-01 -2.05541179e-01 -4.73079354e-01
-4.27470297e-01 -1.18184543e+00 -8.71569991e-01 -5.69624484e-01
-2.24947646e-01 3.50187212e-01 9.84646678e-02 1.12503195e+00
3.77826124e-01 4.89909500e-01 9.48232234e-01 -4.07193840e-01
-9.80867222e-02 -1.03275728e+00 -7.72955000e-01 7.29748249e-01
6.52947962e-01 -6.37066126e-01 -7.28345156e-01 -8.91595632e-02] | [7.452335834503174, 2.7296440601348877] |
5af453b0-ef2a-42e9-bbf5-b5e408d3636c | prior-induced-information-alignment-for-image | 2106.14439 | null | https://arxiv.org/abs/2106.14439v1 | https://arxiv.org/pdf/2106.14439v1.pdf | Prior-Induced Information Alignment for Image Matting | Image matting is an ill-posed problem that aims to estimate the opacity of foreground pixels in an image. However, most existing deep learning-based methods still suffer from the coarse-grained details. In general, these algorithms are incapable of felicitously distinguishing the degree of exploration between deterministic domains (certain FG and BG pixels) and undetermined domains (uncertain in-between pixels), or inevitably lose information in the continuous sampling process, leading to a sub-optimal result. In this paper, we propose a novel network named Prior-Induced Information Alignment Matting Network (PIIAMatting), which can efficiently model the distinction of pixel-wise response maps and the correlation of layer-wise feature maps. It mainly consists of a Dynamic Gaussian Modulation mechanism (DGM) and an Information Alignment strategy (IA). Specifically, the DGM can dynamically acquire a pixel-wise domain response map learned from the prior distribution. The response map can present the relationship between the opacity variation and the convergence process during training. On the other hand, the IA comprises an Information Match Module (IMM) and an Information Aggregation Module (IAM), jointly scheduled to match and aggregate the adjacent layer-wise features adaptively. Besides, we also develop a Multi-Scale Refinement (MSR) module to integrate multi-scale receptive field information at the refinement stage to recover the fluctuating appearance details. Extensive quantitative and qualitative evaluations demonstrate that the proposed PIIAMatting performs favourably against state-of-the-art image matting methods on the Alphamatting.com, Composition-1K and Distinctions-646 dataset. | ['Xin Yang', 'Yong Tang and', 'Yu Qiao', 'Jiake Xie', 'Yuhao Liu'] | 2021-06-28 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 3.79365265e-01 -2.07626432e-01 3.01774353e-01 -3.06935549e-01
-5.01132786e-01 -5.92432544e-02 6.36367738e-01 -1.74480006e-01
-2.66827703e-01 6.09428465e-01 -2.43407279e-01 -1.20441224e-02
-1.44164205e-01 -9.46069539e-01 -8.12461555e-01 -1.16615248e+00
1.92849874e-01 4.94974494e-01 7.23134816e-01 2.09613889e-02
3.00918669e-01 4.70881522e-01 -1.63938451e+00 5.95261991e-01
1.41221535e+00 1.30368340e+00 5.03006279e-01 4.53870058e-01
-5.53640366e-01 8.57258737e-01 -3.97422612e-01 -1.58606425e-01
3.33081663e-01 -3.76829684e-01 -5.04211724e-01 2.31555551e-01
6.05944812e-01 -2.37958640e-01 -1.35368764e-01 1.29655862e+00
2.24030882e-01 -1.03761163e-02 8.13098490e-01 -1.00662434e+00
-4.15503621e-01 2.83183575e-01 -9.83621299e-01 3.49889040e-01
-2.33047858e-01 3.98661494e-01 3.90727460e-01 -7.58888781e-01
4.48085904e-01 1.28577721e+00 4.18745011e-01 3.45401227e-01
-1.28188181e+00 -7.68624902e-01 4.52193648e-01 2.73006886e-01
-1.19739759e+00 -8.53652880e-02 1.03675318e+00 -5.02122998e-01
2.38305449e-01 2.90756673e-01 7.02055573e-01 7.17282832e-01
5.49577415e-01 7.04723358e-01 1.73138893e+00 -1.39477164e-01
2.93844104e-01 1.77296013e-01 2.81241164e-02 6.78763092e-01
2.08776623e-01 9.52484235e-02 -6.92301571e-01 1.10705093e-01
1.26529396e+00 -1.53250456e-01 -3.99412662e-01 -2.46140659e-01
-9.84023690e-01 3.06871355e-01 5.94765484e-01 2.42105886e-01
-4.55378085e-01 -8.83807018e-02 -1.47157639e-01 2.18708575e-01
6.50070369e-01 2.70897239e-01 -3.89452904e-01 2.11186931e-01
-1.21031451e+00 1.96744978e-01 4.04318988e-01 6.18265152e-01
1.38050759e+00 8.70323479e-02 -4.21279639e-01 7.25867927e-01
4.13693517e-01 4.15372550e-01 4.49234992e-01 -9.51664209e-01
3.17858368e-01 8.80021095e-01 1.17623210e-01 -9.96330857e-01
-1.11183181e-01 -4.24750954e-01 -1.08200538e+00 8.76922965e-01
6.20978177e-01 7.16694370e-02 -1.30335057e+00 1.56911993e+00
6.08787715e-01 3.95729005e-01 -2.52917588e-01 9.79483128e-01
7.90643215e-01 7.98897088e-01 -1.23951279e-01 -3.47138286e-01
1.22600305e+00 -1.00294828e+00 -6.55447781e-01 -3.58688653e-01
-2.97614008e-01 -8.44454408e-01 9.63487864e-01 5.58128417e-01
-1.44560385e+00 -9.45979238e-01 -1.09180129e+00 1.20283484e-01
-2.28058413e-01 9.94685013e-03 4.82210666e-01 4.11376119e-01
-1.08600008e+00 7.09224761e-01 -8.99070144e-01 1.16523661e-01
4.58315015e-01 3.89552534e-01 1.51599175e-03 4.41353396e-02
-9.59734082e-01 5.47278047e-01 3.18496615e-01 3.77743602e-01
-1.00640857e+00 -9.81399357e-01 -4.27166849e-01 -1.70112014e-01
2.20734701e-01 -6.39470696e-01 7.55301178e-01 -1.37499118e+00
-1.66964924e+00 7.49052286e-01 -1.29232273e-01 -2.60411859e-01
7.11098075e-01 -7.28050843e-02 -3.80030453e-01 1.80218071e-01
-1.11494146e-01 6.97276056e-01 1.41417885e+00 -1.64898610e+00
-7.38849819e-01 -2.62494028e-01 -9.18391570e-02 4.88960743e-01
-1.57114446e-01 -1.65969521e-01 -5.38679063e-01 -7.57953227e-01
2.68759131e-01 -4.88203943e-01 -1.89901993e-01 2.13819474e-01
-3.03476840e-01 2.71372944e-01 8.30058515e-01 -6.47809029e-01
1.14475167e+00 -2.13804674e+00 2.28970036e-01 1.75105378e-01
3.13596606e-01 3.24014932e-01 3.74347414e-03 -2.33978242e-01
8.83347839e-02 -2.10124895e-01 -5.54387331e-01 -3.20041090e-01
-3.58161777e-01 1.65907145e-01 -2.09036022e-01 4.39978033e-01
3.97117853e-01 7.26562023e-01 -6.98569477e-01 -7.14850843e-01
4.77368772e-01 5.60489774e-01 -3.76012862e-01 4.39533889e-01
-5.28324544e-01 7.47996867e-01 -4.23548579e-01 6.76687360e-01
1.26921988e+00 -2.92938113e-01 -1.33402869e-01 -4.91681933e-01
-4.65075791e-01 -1.10101573e-01 -1.44406319e+00 1.40392351e+00
-2.99434841e-01 3.97274047e-01 5.65055609e-01 -5.34134984e-01
9.32478309e-01 -8.39685723e-02 3.25882435e-01 -6.60383701e-01
2.26231292e-01 1.79960251e-01 1.37756951e-02 -1.48398325e-01
3.69612366e-01 2.21009143e-02 5.21233201e-01 1.94135770e-01
-1.39020324e-01 -1.51331276e-01 -7.45095611e-02 -2.89609153e-02
7.84397185e-01 4.95982990e-02 -1.16012782e-01 -4.67771113e-01
7.02526867e-01 -2.44966686e-01 7.33717740e-01 7.03778565e-01
-8.89842138e-02 9.68533695e-01 3.12502116e-01 -4.17798132e-01
-8.89962494e-01 -1.25946271e+00 -3.97611856e-01 6.48779750e-01
7.31966317e-01 6.06898181e-02 -9.48037267e-01 -4.82258558e-01
-8.24567229e-02 3.94187778e-01 -8.23631525e-01 -7.40462765e-02
-6.61927104e-01 -9.46114302e-01 1.79189316e-03 2.55942047e-01
1.21892095e+00 -1.18202186e+00 -4.12079841e-01 1.40196934e-01
-5.95346279e-02 -7.71350265e-01 -4.41042453e-01 2.10167602e-01
-9.07908797e-01 -8.84036541e-01 -6.26896083e-01 -5.99659503e-01
9.15305376e-01 1.40603110e-01 1.12476683e+00 6.94409534e-02
-1.11108609e-01 -6.91191666e-03 -2.54985001e-02 -1.33356780e-01
-4.82018024e-01 -1.71072751e-01 -3.86087447e-01 5.97640634e-01
-1.14048133e-02 -8.02590728e-01 -9.60350633e-01 5.60549736e-01
-1.11093390e+00 6.65370166e-01 9.27219570e-01 8.11323881e-01
8.53917420e-01 3.28879088e-01 9.98513326e-02 -9.83213127e-01
2.01391831e-01 -2.71254629e-01 -7.43665218e-01 3.12689662e-01
-8.17722023e-01 1.16341896e-01 4.25025254e-01 -6.24135911e-01
-1.65241015e+00 -1.20775118e-01 5.26267365e-02 -5.56833208e-01
-2.19923198e-01 1.22229718e-01 -5.83409309e-01 -4.09343570e-01
4.48178947e-01 3.84842843e-01 2.78576612e-02 -4.37366009e-01
1.43917128e-01 3.28571796e-01 7.28562117e-01 -7.28933096e-01
1.00356328e+00 8.42564046e-01 4.40385425e-03 -5.34806311e-01
-7.62477636e-01 -2.32684702e-01 -5.36265910e-01 -5.69613934e-01
9.72894847e-01 -7.90247858e-01 -3.78300756e-01 1.16000462e+00
-9.87556219e-01 -6.71061397e-01 -2.07711011e-01 2.34926697e-02
-3.53884459e-01 3.79892856e-01 -8.22436154e-01 -6.81777298e-01
-3.18438411e-01 -1.25630677e+00 9.90642369e-01 7.29885221e-01
3.81852955e-01 -8.59202266e-01 -1.25371546e-01 4.88003165e-01
5.48792601e-01 2.96868056e-01 8.93290877e-01 1.18390352e-01
-1.21639347e+00 4.42051739e-01 -5.75158417e-01 4.05423433e-01
2.14105457e-01 9.51569155e-02 -1.14368021e+00 -1.84711844e-01
2.56107390e-01 -8.27478766e-02 1.11200678e+00 7.90720344e-01
1.24181116e+00 -1.26799360e-01 -6.31121695e-02 1.04855943e+00
1.46670842e+00 2.51661301e-01 1.05218518e+00 5.30058682e-01
8.17426741e-01 4.33972806e-01 7.18639553e-01 3.44802976e-01
3.47414725e-02 6.25900924e-01 6.36957705e-01 -4.97980565e-01
-4.88952219e-01 -3.39252464e-02 2.66260684e-01 5.78908682e-01
-2.04988703e-01 -4.93161641e-02 -4.47571576e-01 1.72904059e-01
-1.81936944e+00 -7.14842141e-01 -2.14182675e-01 2.13563418e+00
1.21838963e+00 3.58902246e-01 -1.98070943e-01 -1.08596615e-01
8.10094476e-01 2.90445179e-01 -7.14420319e-01 -2.11977407e-01
-3.81653428e-01 2.36671269e-01 4.79287446e-01 5.48540592e-01
-1.08232355e+00 7.87228763e-01 5.35202694e+00 1.25799668e+00
-1.09734511e+00 2.29970906e-02 1.05850482e+00 3.80487025e-01
-5.61196446e-01 -3.32122445e-02 -8.29144776e-01 8.35749328e-01
3.91343147e-01 2.73984969e-01 4.99408662e-01 5.30710518e-01
4.88782823e-02 -4.58533973e-01 -6.95717275e-01 8.39458466e-01
-9.40765589e-02 -1.40172935e+00 2.99330086e-01 -5.91283739e-02
9.81666267e-01 -1.55177861e-01 3.95969450e-01 -5.72438948e-02
1.46270290e-01 -8.98458242e-01 8.41408670e-01 9.44945753e-01
7.67143726e-01 -5.68995178e-01 5.98895848e-01 2.50042528e-01
-1.46996784e+00 2.21907478e-02 -3.27478617e-01 1.23478077e-01
-2.00597532e-02 9.22896206e-01 -3.30609143e-01 5.84345341e-01
8.04791093e-01 5.04497290e-01 -7.48641670e-01 1.08767557e+00
-2.10030407e-01 5.28292120e-01 -1.53177753e-01 4.31260735e-01
1.17426984e-01 -6.72941923e-01 5.76789796e-01 1.02282107e+00
7.99889639e-02 -1.11702278e-01 -1.90208759e-02 1.30831075e+00
1.92258969e-01 -2.55320370e-01 2.17470184e-01 2.97055364e-01
3.02420884e-01 1.41355908e+00 -9.71848011e-01 -3.41158926e-01
-3.26914549e-01 1.02557719e+00 2.87634730e-01 5.77320933e-01
-8.83397698e-01 3.51473056e-02 6.46749794e-01 4.37088668e-01
2.82573044e-01 1.71692707e-02 -5.76029181e-01 -1.02461219e+00
2.27575842e-02 -8.77517581e-01 1.56013414e-01 -8.44272673e-01
-1.40747273e+00 6.02502167e-01 4.24294882e-02 -1.21735275e+00
2.64404565e-01 -5.99567056e-01 -7.92874396e-01 1.17382872e+00
-1.80194759e+00 -1.15217543e+00 -8.02753508e-01 6.95511281e-01
5.89959323e-01 3.17063592e-02 1.99145183e-01 2.97166556e-01
-5.89639187e-01 3.92785192e-01 1.17232390e-01 -1.72004610e-01
7.01290965e-01 -1.41065907e+00 1.86325312e-01 9.90384936e-01
-1.65754452e-01 4.36103255e-01 7.34803200e-01 -7.82628477e-01
-9.80075359e-01 -1.17583454e+00 1.48867354e-01 -2.02555701e-01
4.97120202e-01 -2.32579410e-01 -1.36047471e+00 2.19793931e-01
1.61402345e-01 9.62723270e-02 -8.60689431e-02 -2.79419273e-01
-1.25680134e-01 -5.36976576e-01 -1.16339910e+00 5.71752667e-01
7.92119145e-01 -2.86302567e-01 -2.25395903e-01 -7.37937838e-02
5.73467553e-01 -6.92369282e-01 -6.19003773e-01 6.53775930e-01
3.66336048e-01 -1.38015616e+00 1.02054143e+00 1.36490047e-01
4.65278774e-01 -7.86740005e-01 1.65833473e-01 -1.01888561e+00
-5.59506118e-01 -6.08787835e-01 -2.44615614e-01 1.52842724e+00
1.24872670e-01 -5.47646701e-01 8.87723327e-01 5.66745758e-01
-1.88528612e-01 -9.33628142e-01 -1.00001061e+00 -4.90061820e-01
-5.14497757e-02 -9.22416598e-02 5.94265759e-01 6.19642735e-01
-8.26927602e-01 -1.34513438e-01 -3.10025454e-01 4.04799432e-01
8.77843618e-01 3.03262621e-01 5.78616261e-01 -1.17435360e+00
-5.34227729e-01 -5.31387448e-01 -1.86861932e-01 -1.06117225e+00
-1.96666047e-01 -4.08480883e-01 1.68978468e-01 -1.37007022e+00
4.30385143e-01 -6.83243215e-01 -5.76315105e-01 4.83719818e-02
-5.31525314e-01 1.12169012e-01 -5.65188564e-02 3.15512121e-01
-4.54700738e-01 5.58505058e-01 1.68065524e+00 -2.38337129e-01
-2.51561850e-01 1.63028419e-01 -4.01491761e-01 7.32357442e-01
6.52739227e-01 -2.38477573e-01 -4.81699586e-01 -3.91565114e-01
1.24094971e-01 -1.39796034e-01 5.42198062e-01 -1.26098704e+00
2.26759061e-01 -4.56204921e-01 6.50089443e-01 -7.57529259e-01
3.37732792e-01 -6.67320192e-01 3.73099536e-01 1.44703239e-01
-1.34341687e-01 -2.75631189e-01 2.21166432e-01 6.09385014e-01
-1.97371811e-01 -8.16708654e-02 1.20424652e+00 -2.72405118e-01
-7.43207276e-01 5.92942834e-01 -2.99382657e-01 1.07117184e-01
7.83201814e-01 -4.90869641e-01 -4.05469269e-01 5.84010333e-02
-3.37525100e-01 8.33540335e-02 7.75800884e-01 2.56917290e-02
6.97076201e-01 -1.16934097e+00 -5.16307414e-01 5.19898295e-01
-2.48434559e-01 4.85572457e-01 5.83834350e-01 7.92365491e-01
-5.78632414e-01 -4.01183814e-01 -3.46686125e-01 -6.96407020e-01
-1.06179476e+00 3.07349294e-01 5.67515492e-01 -4.66372311e-01
-5.22425532e-01 1.15571988e+00 7.52319992e-01 -3.30313817e-02
2.41377532e-01 -4.23660874e-01 -1.66043431e-01 -7.02859834e-02
6.06177092e-01 2.40013376e-01 -3.47389355e-02 -3.40633959e-01
1.99870262e-02 6.31972790e-01 -2.34636948e-01 3.52276460e-04
1.08357167e+00 -2.76310742e-01 -4.31990176e-01 4.87493515e-01
7.30433166e-01 -1.40380368e-01 -2.01526332e+00 -2.49647990e-01
-3.36216867e-01 -6.42639279e-01 2.91126281e-01 -8.44924212e-01
-1.34887457e+00 7.33179867e-01 8.62688124e-01 -1.77930426e-02
1.46603656e+00 -1.35682285e-01 7.61754096e-01 -2.43027151e-01
2.51631171e-01 -1.13127029e+00 2.90849179e-01 1.90842763e-01
7.12534606e-01 -1.12650347e+00 1.21595442e-01 -4.86251742e-01
-4.42698121e-01 1.04170108e+00 1.04591095e+00 -1.30411804e-01
6.39881551e-01 6.41057432e-01 2.00332060e-01 -1.26196757e-01
-6.72420084e-01 -7.30899721e-02 5.15612543e-01 5.09676039e-01
2.94919079e-03 -1.30475670e-01 -1.15593448e-02 4.33422953e-01
2.31319815e-01 -2.48170912e-01 1.46220818e-01 6.37807667e-01
-5.89134812e-01 -9.63489711e-01 -5.69606960e-01 4.34911460e-01
-1.70493543e-01 -2.29401022e-01 -1.54404089e-01 5.51458538e-01
6.31007731e-01 6.82407618e-01 2.29003474e-01 -3.36582541e-01
-4.66136597e-02 -3.64631981e-01 5.16635716e-01 -3.07629734e-01
-6.21194661e-01 2.44752944e-01 -3.23259115e-01 -6.48218691e-01
-5.17139018e-01 -7.18463898e-01 -1.09568536e+00 -1.74913213e-01
-2.71427274e-01 -8.63423645e-02 3.20422560e-01 8.07230890e-01
9.02709737e-02 8.11822772e-01 5.68585217e-01 -1.01794231e+00
-5.22196218e-02 -8.12769115e-01 -7.04649389e-01 3.96839261e-01
2.69147038e-01 -7.79057145e-01 -6.08187675e-01 1.20649554e-01] | [10.664796829223633, -1.0590382814407349] |
2c0f1535-6756-433e-9f5f-d59bd00bddac | does-the-order-of-training-samples-matter | 2102.03554 | null | https://arxiv.org/abs/2102.03554v1 | https://arxiv.org/pdf/2102.03554v1.pdf | Does the Order of Training Samples Matter? Improving Neural Data-to-Text Generation with Curriculum Learning | Recent advancements in data-to-text generation largely take on the form of neural end-to-end systems. Efforts have been dedicated to improving text generation systems by changing the order of training samples in a process known as curriculum learning. Past research on sequence-to-sequence learning showed that curriculum learning helps to improve both the performance and convergence speed. In this work, we delve into the same idea surrounding the training samples consisting of structured data and text pairs, where at each update, the curriculum framework selects training samples based on the model's competence. Specifically, we experiment with various difficulty metrics and put forward a soft edit distance metric for ranking training samples. Our benchmarks show faster convergence speed where training time is reduced by 38.7% and performance is boosted by 4.84 BLEU. | ['Vera Demberg', 'Hui-Syuan Yeh', 'Ernie Chang'] | 2021-02-06 | null | https://aclanthology.org/2021.eacl-main.61 | https://aclanthology.org/2021.eacl-main.61.pdf | eacl-2021-2 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 5.52064836e-01 1.76583961e-01 -1.50544420e-01 -4.87694949e-01
-8.12681437e-01 -6.10336483e-01 8.28967929e-01 3.37501675e-01
-6.41311347e-01 9.27908897e-01 4.75564361e-01 -4.69555140e-01
1.75498366e-01 -7.61790216e-01 -6.19448304e-01 -3.17591608e-01
2.80427605e-01 7.70200849e-01 5.76051027e-02 -5.74350655e-01
5.52229822e-01 -2.27374569e-01 -1.39792907e+00 6.23327971e-01
1.22757018e+00 4.85165209e-01 2.93349594e-01 1.00412965e+00
-5.40531337e-01 4.68322694e-01 -9.34953272e-01 -7.19511390e-01
2.24550858e-01 -8.49443376e-01 -8.55405271e-01 -1.13148227e-01
4.74284947e-01 -2.90359408e-01 -1.48511440e-01 7.98081577e-01
7.22593606e-01 2.88756549e-01 7.02081442e-01 -9.94487762e-01
-7.24926949e-01 1.12784553e+00 -3.46223891e-01 3.11189685e-02
4.58611548e-01 -6.03292063e-02 9.76474047e-01 -9.44703281e-01
7.46917367e-01 1.13393068e+00 4.36102092e-01 9.74254310e-01
-1.16666818e+00 -6.00083113e-01 1.29660964e-01 1.84485361e-01
-8.23677778e-01 -3.40653956e-01 4.20295537e-01 -3.57397527e-01
9.46341157e-01 3.00620705e-01 5.55064142e-01 1.07442224e+00
1.21317074e-01 1.05405796e+00 7.70586610e-01 -8.45014513e-01
1.90012112e-01 1.46789059e-01 2.64881421e-02 4.98715341e-01
2.67203540e-01 5.93234636e-02 -6.04634047e-01 7.28447884e-02
2.85157055e-01 -4.21575129e-01 -1.05008617e-01 -2.12377086e-01
-1.02733994e+00 9.66276765e-01 1.06092244e-01 1.89590573e-01
6.87189549e-02 1.06629618e-02 5.81137002e-01 5.04366696e-01
4.41727579e-01 9.14977074e-01 -4.94594574e-01 -5.77559471e-01
-9.74182487e-01 5.51513135e-01 9.58937764e-01 1.04414093e+00
4.28379893e-01 -3.81756993e-03 -5.92747748e-01 9.85466719e-01
5.20579368e-02 2.48912394e-01 8.24315310e-01 -7.47729957e-01
9.34154630e-01 5.30924618e-01 -4.87472955e-03 -4.74065840e-01
6.67356653e-03 -3.30067724e-01 -6.18076503e-01 3.68686952e-02
4.07941878e-01 -6.19304478e-01 -1.16019404e+00 1.60264623e+00
2.00952977e-01 -2.18229845e-01 5.28253168e-02 5.25453329e-01
4.83756661e-01 7.73835063e-01 -5.69181144e-02 -6.05806969e-02
8.78264427e-01 -1.26907849e+00 -6.09848082e-01 -1.79982901e-01
1.00103307e+00 -9.28924441e-01 1.13685286e+00 4.12958652e-01
-1.43467391e+00 -7.59041190e-01 -1.00482666e+00 3.75111736e-02
-2.90146023e-01 -4.77587655e-02 1.98365420e-01 7.80730844e-01
-1.10157251e+00 8.33460450e-01 -5.40316582e-01 -1.66490838e-01
2.35705405e-01 2.98219442e-01 1.18851453e-01 -2.08440959e-01
-1.10778117e+00 7.75871575e-01 7.68762589e-01 -3.89446527e-01
-7.38760948e-01 -9.57367480e-01 -6.10547721e-01 6.53904453e-02
1.75800994e-01 -9.12546694e-01 1.61875880e+00 -1.15998518e+00
-1.74573076e+00 2.66864389e-01 -8.87613669e-02 -4.92719054e-01
7.48707831e-01 -3.31555873e-01 -1.70315325e-01 -2.80096889e-01
-1.48056030e-01 9.44797695e-01 6.30896091e-01 -1.15793884e+00
-1.03950417e+00 8.34974796e-02 -7.73857310e-02 5.33580780e-01
-6.47214413e-01 -4.52654809e-02 -3.89502317e-01 -8.11256230e-01
-3.53789270e-01 -9.50449705e-01 -3.13856751e-01 -4.36066985e-01
-1.27863199e-01 -3.87464911e-01 5.04800498e-01 -7.37292945e-01
1.36500180e+00 -1.63971317e+00 2.09229812e-01 1.07356505e-02
3.48020457e-02 5.02682686e-01 -4.55384970e-01 7.02187896e-01
-1.57536920e-02 3.36039484e-01 -5.02012596e-02 -4.34241563e-01
1.04055360e-01 -9.28830877e-02 -2.56852299e-01 -3.55914712e-01
2.44712248e-01 8.94841194e-01 -1.20409119e+00 -3.77367258e-01
-1.38867006e-01 1.41726628e-01 -9.32884216e-01 3.39491487e-01
-4.73625511e-01 2.36318409e-01 -2.28952274e-01 1.03450917e-01
2.80276567e-01 -2.86318846e-02 2.25053743e-01 4.82278019e-01
9.93949473e-02 6.45614982e-01 -9.00372028e-01 1.88641620e+00
-5.57183027e-01 8.33870053e-01 -4.43647981e-01 -6.27798676e-01
1.06850839e+00 2.40895897e-01 1.87644541e-01 -8.75769079e-01
-7.61846676e-02 6.86252415e-02 3.00210536e-01 -3.10731471e-01
1.12836134e+00 1.48891866e-01 7.12316185e-02 5.91365278e-01
-1.62452117e-01 -8.75239074e-02 7.46657610e-01 3.14382553e-01
1.03564596e+00 3.00177276e-01 -4.41141650e-02 3.37191634e-02
2.94504344e-01 2.13727549e-01 2.37790972e-01 8.41057956e-01
2.57986128e-01 4.27341819e-01 3.92300069e-01 -2.90570259e-01
-1.38170958e+00 -9.24259841e-01 2.23045588e-01 1.41239083e+00
-2.71431625e-01 -7.77422667e-01 -1.06137812e+00 -9.02761161e-01
-5.10358587e-02 1.05283058e+00 -5.37923515e-01 -3.49769682e-01
-8.48850787e-01 -5.52683592e-01 6.02543056e-01 6.25350416e-01
1.93969697e-01 -1.15317619e+00 -4.31993127e-01 4.35166985e-01
-2.57822394e-01 -5.50237596e-01 -7.81182289e-01 1.56873524e-01
-1.13564706e+00 -5.78689814e-01 -8.46626282e-01 -8.33532751e-01
8.89932513e-01 5.96791171e-02 1.31514478e+00 2.78138909e-02
-1.24533772e-01 -1.77224666e-01 -7.64385581e-01 -5.78876376e-01
-8.60178947e-01 6.77256167e-01 -1.94427520e-01 -5.57769060e-01
1.14289433e-01 -1.98799044e-01 -4.47917193e-01 2.55922794e-01
-8.45221102e-01 3.81591856e-01 8.03517759e-01 1.09700167e+00
2.50357538e-01 -2.87140101e-01 7.04084098e-01 -1.04762435e+00
1.12077558e+00 -5.05067229e-01 -3.11735719e-01 3.27475727e-01
-1.04143798e+00 5.05762219e-01 8.17276716e-01 -4.41656262e-01
-1.19394267e+00 -5.07342182e-02 -1.51252300e-01 -9.65583473e-02
-8.79087150e-02 6.37521446e-01 1.30584881e-01 3.80036056e-01
8.05821657e-01 2.89731324e-01 -7.70585015e-02 -2.55014002e-01
4.63869095e-01 7.65308857e-01 2.70948112e-01 -6.81724846e-01
7.36777365e-01 -2.78117985e-01 -5.02342880e-01 -3.02438796e-01
-7.25833595e-01 -2.23375648e-01 -4.65162843e-01 -1.94658935e-01
5.31283975e-01 -6.88689888e-01 -3.04012954e-01 2.27660298e-01
-9.57623005e-01 -8.42306018e-01 -3.04322630e-01 2.40938008e-01
-4.23056155e-01 2.57485211e-01 -5.17148077e-01 -5.33682287e-01
-6.60780072e-01 -9.76758897e-01 9.25491154e-01 2.92919219e-01
-5.48893690e-01 -1.05697262e+00 3.19733351e-01 4.05444026e-01
4.86772895e-01 -9.06156301e-02 9.65617299e-01 -7.23115444e-01
-3.06310385e-01 -1.31120071e-01 1.04835629e-01 2.69484788e-01
1.21398903e-01 -2.95955930e-02 -5.52130461e-01 -4.63677555e-01
-4.75178689e-01 -4.03555512e-01 7.62724876e-01 7.05072805e-02
1.30895674e+00 -2.56250054e-01 -3.13658565e-01 4.50590611e-01
1.16303301e+00 3.87266368e-01 5.74695885e-01 3.13306689e-01
5.93411446e-01 6.08182371e-01 7.67819643e-01 4.50191081e-01
2.95280308e-01 6.84175193e-01 5.86817786e-02 6.44027889e-02
-2.57824928e-01 -6.44570112e-01 5.02869785e-01 1.03910685e+00
1.96653649e-01 -5.59728861e-01 -9.09545422e-01 6.57251060e-01
-1.79944801e+00 -9.27343428e-01 -5.21804718e-03 2.23639774e+00
1.35999656e+00 3.99100900e-01 1.34339035e-01 5.21534085e-02
6.05265558e-01 -7.97447711e-02 -4.03775096e-01 -6.95359230e-01
2.73514390e-01 4.43705857e-01 3.70800644e-01 6.00833178e-01
-6.97226584e-01 1.00479126e+00 6.65376186e+00 8.50204587e-01
-9.40706372e-01 -2.61420190e-01 6.68364346e-01 -3.96233916e-01
-5.29784262e-01 -1.58445656e-01 -1.17607379e+00 6.82494164e-01
1.26053154e+00 -4.61914122e-01 3.72146994e-01 7.71340728e-01
2.05111533e-01 9.47970711e-03 -1.12644136e+00 5.85001886e-01
1.76621392e-01 -1.44659758e+00 2.69078463e-01 1.14933439e-02
1.20944440e+00 -1.48243427e-01 -6.52763098e-02 5.73845208e-01
7.30684221e-01 -1.15321732e+00 6.60169661e-01 3.80898714e-01
7.13591695e-01 -1.05888689e+00 7.31079459e-01 3.96280885e-01
-8.57206762e-01 1.38325579e-02 -4.22254801e-01 -3.37861218e-02
1.43338874e-01 4.12239492e-01 -1.60165632e+00 5.81445932e-01
3.81519973e-01 4.60348219e-01 -6.05491698e-01 1.12423468e+00
-2.19459578e-01 7.52614915e-01 1.05794340e-01 -5.49951673e-01
2.53942758e-01 -1.10641412e-01 2.25908279e-01 1.43790579e+00
3.14009517e-01 -2.55734414e-01 1.96349844e-01 6.41847849e-01
-3.06009054e-01 1.32255733e-01 -3.37163925e-01 -5.18885702e-02
5.47559202e-01 1.09209776e+00 -4.24184978e-01 -4.79810774e-01
-5.94767295e-02 1.08757567e+00 3.34334671e-01 -2.08119918e-02
-7.29831934e-01 -7.45274067e-01 5.45867205e-01 1.00555606e-02
4.83931512e-01 -1.52049407e-01 -5.30381143e-01 -7.46012807e-01
-2.64892355e-02 -1.08355308e+00 1.19703457e-01 -5.68225443e-01
-1.05900693e+00 5.37984848e-01 -4.66021895e-02 -1.03327084e+00
-6.07149720e-01 -4.02655154e-01 -7.29433656e-01 9.57508564e-01
-1.27412820e+00 -5.37696123e-01 -3.51841807e-01 1.99427977e-01
1.05041742e+00 -2.17445880e-01 7.80685425e-01 2.82818824e-01
-4.55174893e-01 1.08687353e+00 6.37595475e-01 1.75262749e-01
1.02091324e+00 -1.59907258e+00 1.02211690e+00 6.83247328e-01
1.65964574e-01 6.36717558e-01 7.60393679e-01 -7.91308999e-01
-1.09567690e+00 -1.19891632e+00 1.16077054e+00 -4.88604933e-01
3.84838581e-01 -4.89470780e-01 -6.85672402e-01 2.65578330e-01
4.98279750e-01 -6.39851332e-01 7.02649653e-01 2.27743432e-01
-1.77688301e-01 5.96632138e-02 -7.85635591e-01 8.62884045e-01
1.02000403e+00 -1.84787348e-01 -5.40236771e-01 3.29881489e-01
8.08126211e-01 -7.96253383e-01 -8.64462852e-01 -1.90005470e-02
4.78592575e-01 -5.83428085e-01 5.50578535e-01 -8.31679583e-01
8.65249336e-01 -6.12311959e-02 2.36674473e-01 -1.89918077e+00
-2.54632622e-01 -7.06308126e-01 -1.63650930e-01 1.16702819e+00
7.18484402e-01 -1.00282550e-01 1.18816471e+00 3.57224613e-01
-4.49448705e-01 -9.30718839e-01 -3.61627728e-01 -6.83644950e-01
4.32186604e-01 -1.08456403e-01 6.00454748e-01 7.56333053e-01
2.21111387e-01 6.90503061e-01 -2.82322854e-01 -5.15442014e-01
2.73496896e-01 -6.90528005e-02 8.55988562e-01 -1.08182847e+00
-3.89920413e-01 -6.14635468e-01 2.27904901e-01 -1.31715190e+00
-4.24749143e-02 -1.00248754e+00 3.42449903e-01 -1.62254596e+00
1.40237585e-01 -3.31494361e-01 -8.55168477e-02 3.72969538e-01
-7.30014801e-01 -1.45613059e-01 3.26281190e-01 -1.30208343e-01
-5.87402403e-01 5.61225474e-01 1.21642542e+00 -3.69551033e-02
-3.26552361e-01 -4.65481244e-02 -7.93888390e-01 7.67590478e-02
1.21210921e+00 -5.49137235e-01 -6.84411108e-01 -7.29794621e-01
3.33304644e-01 6.03398830e-02 -3.62560958e-01 -9.23251748e-01
2.28576541e-01 -1.11106306e-01 4.57548648e-01 -6.15943074e-01
-6.28598128e-03 -2.76668042e-01 -1.38402492e-01 6.58455491e-01
-9.33676720e-01 3.80061209e-01 2.09364966e-01 4.33789611e-01
-7.87459835e-02 -5.66501081e-01 6.26410604e-01 -7.71749328e-05
-4.20009464e-01 3.65382545e-02 -3.34716469e-01 3.46821815e-01
7.99522221e-01 -3.00699741e-01 -3.57112408e-01 -4.52244610e-01
-3.82890284e-01 4.68611270e-01 3.25786531e-01 6.56800032e-01
4.59279686e-01 -1.25407386e+00 -1.07074869e+00 5.82307130e-02
1.54031500e-01 -6.69283494e-02 -1.51914686e-01 3.26826364e-01
-4.88474965e-01 6.35065615e-01 -3.21105897e-01 -3.28610688e-01
-1.48005998e+00 1.38389349e-01 2.25445718e-01 -6.76979601e-01
-1.12078294e-01 1.02122104e+00 -3.81758392e-01 -5.57731390e-01
4.11176115e-01 -2.26412192e-01 -3.68705541e-02 8.66295397e-02
6.18305743e-01 5.51133573e-01 2.14545324e-01 1.19121693e-01
2.11446017e-01 9.25619528e-02 -6.99487865e-01 -3.48226279e-01
1.12417090e+00 1.68861479e-01 3.08426797e-01 1.88736737e-01
1.08379173e+00 -1.27023906e-01 -1.27580464e+00 -1.87032819e-01
1.93997532e-01 -3.93820614e-01 -2.53181815e-01 -1.17840385e+00
-7.74539411e-01 7.64933169e-01 4.38655108e-01 1.18271805e-01
8.10292661e-01 -3.77632976e-01 9.59107459e-01 6.66021466e-01
1.22807942e-01 -1.42949283e+00 4.51867789e-01 9.89015996e-01
7.01538026e-01 -1.15841544e+00 -2.06881776e-01 -2.48263646e-02
-7.40584612e-01 1.17242491e+00 8.89156222e-01 5.38741313e-02
1.38778180e-01 2.24079832e-01 -1.15193436e-02 3.23881567e-01
-1.14019573e+00 1.62022173e-01 3.53696436e-01 5.62017143e-01
9.37350929e-01 1.07164616e-02 -4.92418051e-01 2.75341302e-01
-6.77988172e-01 -8.89585540e-03 5.11626184e-01 9.32096064e-01
-6.70255542e-01 -1.71280277e+00 -1.06201448e-01 7.12808132e-01
-2.93825001e-01 -3.65495622e-01 -5.33240139e-01 5.26419699e-01
-5.04004508e-02 9.60928023e-01 1.98251352e-01 -5.48224688e-01
4.10060793e-01 3.56013238e-01 4.58316892e-01 -8.48115802e-01
-1.08469558e+00 -2.95985520e-01 3.39139938e-01 -2.10446596e-01
1.16602711e-01 -7.52647579e-01 -1.23501778e+00 -4.08071190e-01
-3.21220160e-01 5.03289104e-01 8.22170138e-01 8.24208140e-01
3.78903866e-01 8.42321277e-01 6.89362705e-01 -4.77505386e-01
-9.09993172e-01 -1.10490286e+00 8.25139359e-02 2.68314451e-01
3.62917110e-02 -7.82769993e-02 8.70174021e-02 2.46174201e-01] | [11.858701705932617, 9.026329040527344] |
555f5d50-b415-48c4-9dc3-455c8a4eabc3 | image-augmentation-with-conformal-mappings | 2212.05258 | null | https://arxiv.org/abs/2212.05258v1 | https://arxiv.org/pdf/2212.05258v1.pdf | Image augmentation with conformal mappings for a convolutional neural network | For augmentation of the square-shaped image data of a convolutional neural network (CNN), we introduce a new method, in which the original images are mapped onto a disk with a conformal mapping, rotated around the center of this disk and mapped under such a M\"obius transformation that preserves the disk, and then mapped back onto their original square shape. This process does not result the loss of information caused by removing areas from near the edges of the original images unlike the typical transformations used in the data augmentation for a CNN. We offer here the formulas of all the mappings needed together with detailed instructions how to write a code for transforming the images. The new method is also tested with simulated data and, according the results, using this method to augment the training data of 10 images into 40 images decreases the amount of the error in the predictions by a CNN for a test set of 160 images in a statistically significant way (p-value=0.0360). | ['Riku Klén', 'Matti Vuorinen', 'Mohamed M. S. Nasser', 'Oona Rainio'] | 2022-12-10 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 3.86078745e-01 9.06656206e-01 3.68784547e-01 -3.16458344e-01
2.10554332e-01 -4.61160034e-01 6.66848958e-01 -1.86739787e-02
-6.97651505e-01 6.29558802e-01 -1.81997061e-01 -2.09488571e-01
1.34650499e-01 -8.37134063e-01 -1.17071927e+00 -6.49042368e-01
3.36812027e-02 4.59868610e-01 4.80465293e-01 -2.48026401e-01
1.68920442e-01 8.72773945e-01 -1.62402606e+00 1.68923363e-01
5.17506599e-01 9.47103202e-01 1.54938295e-01 5.49617290e-01
6.07364066e-02 1.64706588e-01 -5.24174213e-01 -2.99298406e-01
5.84649086e-01 -3.82353842e-01 -7.45855451e-01 1.91498816e-01
5.30569792e-01 -2.23626629e-01 -4.13394153e-01 1.12856233e+00
3.93039510e-02 -8.21599644e-03 8.57515275e-01 -9.65305328e-01
-7.35077083e-01 3.25985551e-01 -3.66809338e-01 2.73560546e-02
-1.08474307e-01 -2.51189888e-01 2.14726582e-01 -9.17649031e-01
7.62807250e-01 1.09659100e+00 7.89628506e-01 4.62950796e-01
-1.26028204e+00 -4.00770545e-01 -4.19613123e-01 -9.55941901e-02
-1.19931889e+00 -8.85936990e-02 4.69680548e-01 -5.20896852e-01
8.22862983e-01 3.31430227e-01 8.21746349e-01 7.10003734e-01
5.00806272e-01 9.59414467e-02 1.01632786e+00 -7.50238240e-01
1.29409388e-01 6.41649425e-01 -8.12332854e-02 7.32799053e-01
4.25260574e-01 9.57406610e-02 1.81658879e-01 2.09350005e-01
1.02122331e+00 -1.76425532e-01 1.36452436e-01 -4.57898885e-01
-1.11022162e+00 8.10952485e-01 8.15138817e-01 5.55994987e-01
-2.33133078e-01 -7.29330406e-02 2.03753918e-01 1.49292260e-01
4.79008019e-01 6.25112891e-01 -3.60104114e-01 5.46222985e-01
-5.42126894e-01 2.12976769e-01 5.45241654e-01 7.86635756e-01
8.43005896e-01 1.13585196e-01 1.06231757e-01 5.63013911e-01
4.45660530e-03 4.08208370e-01 7.16690898e-01 -7.88511097e-01
1.13846749e-01 8.64277363e-01 -7.25039095e-02 -1.11946845e+00
-5.86778998e-01 -3.82970273e-01 -1.05090690e+00 6.94146693e-01
5.79996943e-01 -2.72283465e-01 -1.03544188e+00 1.70448005e+00
1.19941674e-01 -5.83997667e-02 1.61502242e-01 7.51664460e-01
6.33386791e-01 7.14322209e-01 -2.10415408e-01 5.77433556e-02
1.37543225e+00 -8.48463237e-01 -6.29500926e-01 -1.78921491e-01
5.34906447e-01 -6.89037025e-01 9.73866999e-01 3.64387214e-01
-1.02660108e+00 -9.15599525e-01 -1.55041552e+00 -5.73613457e-02
-7.87432790e-01 3.46278548e-01 2.22949713e-01 5.06414890e-01
-1.24467778e+00 7.46831536e-01 -5.48181534e-01 -5.20129442e-01
3.36147696e-01 5.82702994e-01 -7.78045356e-01 4.07383442e-01
-1.07748449e+00 1.24491763e+00 6.83592200e-01 -1.77721158e-02
-6.00657701e-01 -5.51346779e-01 -7.06993818e-01 2.31131986e-01
-2.49381006e-01 -2.97270656e-01 7.51265883e-01 -1.51063836e+00
-1.21990371e+00 1.07093966e+00 1.75538570e-01 -6.00610971e-01
5.53991020e-01 6.56239763e-02 -3.27205509e-01 -1.48802623e-01
-8.73753801e-02 1.30734754e+00 1.00694227e+00 -1.28635693e+00
-3.66848975e-01 -4.25572693e-01 -2.83806510e-02 -6.47857711e-02
-5.40277183e-01 -1.22412212e-01 -2.47656301e-01 -7.91251957e-01
1.93932474e-01 -1.16610169e+00 -3.05965781e-01 7.50049576e-02
-2.31438771e-01 2.65568078e-01 8.11276317e-01 -6.95432246e-01
6.30510986e-01 -2.47949624e+00 7.97314104e-03 3.91726285e-01
3.86119336e-02 3.57064664e-01 -2.48566061e-01 2.20087245e-01
-7.11179316e-01 1.04604557e-01 -4.48691517e-01 -2.66646683e-01
-3.87121171e-01 1.31401703e-01 -2.51527488e-01 4.52184767e-01
3.53529364e-01 7.01383471e-01 -4.48849648e-01 -5.21393120e-02
8.26684013e-02 6.05307937e-01 -3.72136086e-01 1.72225937e-01
-8.75402838e-02 3.29989672e-01 2.82606073e-02 -3.85881573e-01
8.53007793e-01 8.32522064e-02 -2.01960623e-01 -8.17789435e-02
-8.49576592e-02 -2.29603320e-01 -1.04913366e+00 1.23318899e+00
-2.03479469e-01 1.05349243e+00 -2.52081156e-01 -1.00306618e+00
1.43458617e+00 3.34404469e-01 2.44790062e-01 -6.42341971e-01
2.74709940e-01 2.47032344e-01 1.77387863e-01 -3.49203110e-01
2.76280940e-01 -1.10804111e-01 2.68975496e-01 4.99182880e-01
9.56753045e-02 -2.54595846e-01 8.76217857e-02 -1.62411109e-01
7.68388152e-01 -1.84682216e-02 1.79167375e-01 -6.03503585e-01
6.89664543e-01 9.01109129e-02 1.32653981e-01 3.33312899e-01
2.65388310e-01 7.99539268e-01 6.86773777e-01 -9.87077594e-01
-1.79878199e+00 -6.35719478e-01 -3.67522717e-01 5.28469861e-01
-2.07477525e-01 4.40996811e-02 -1.51340520e+00 -4.10185158e-01
-8.75081420e-02 5.06018877e-01 -1.18342650e+00 -3.97975773e-01
-4.78627026e-01 -7.14685857e-01 5.47138453e-01 4.30647016e-01
6.16673052e-01 -1.44511473e+00 -6.96024418e-01 -1.96834773e-01
2.93305725e-01 -1.11951923e+00 -5.83236180e-02 3.90479177e-01
-1.09504747e+00 -8.25462580e-01 -6.50258899e-01 -1.06339872e+00
1.27400124e+00 -6.10679798e-02 6.30187035e-01 1.13251999e-01
-2.33851492e-01 -2.77122140e-01 -1.57980755e-01 -6.28216088e-01
-7.40331948e-01 7.25071356e-02 1.35412782e-01 1.28918752e-01
2.29224071e-01 -3.68713230e-01 -3.62472773e-01 3.68994862e-01
-1.04414499e+00 1.84826329e-01 7.02803016e-01 6.75928831e-01
3.76088887e-01 4.85189781e-02 4.50350404e-01 -7.73130059e-01
4.72255856e-01 -2.22253680e-01 -8.95049453e-01 -2.30731025e-01
-6.62464201e-01 2.27452099e-01 9.24930811e-01 -4.89325851e-01
-7.18110979e-01 3.44988763e-01 4.14409582e-03 -3.63743603e-01
-3.91651124e-01 1.11810327e-01 5.67811802e-02 -4.37650651e-01
1.12573671e+00 2.03928724e-02 4.22673047e-01 -3.99677068e-01
2.29404077e-01 4.87881601e-01 6.50012195e-01 7.99204707e-02
1.02658141e+00 4.59324807e-01 3.65462154e-01 -6.46008492e-01
-4.24985379e-01 1.74950838e-01 -1.06620777e+00 -8.48396271e-02
1.08963788e+00 -4.16544169e-01 -4.60052341e-01 3.54106188e-01
-1.27736890e+00 -3.43660265e-01 -5.41986942e-01 4.72638875e-01
-6.42012596e-01 9.64801386e-02 -2.50087172e-01 -2.97208518e-01
-1.10971920e-01 -1.03924763e+00 5.39196610e-01 4.28851217e-01
-4.96931188e-02 -9.53900576e-01 1.02394201e-01 -7.57052377e-02
4.37675625e-01 6.00316264e-02 1.23740125e+00 -9.30473924e-01
-1.81398600e-01 -4.92773682e-01 -3.74157250e-01 7.61682928e-01
1.12278081e-01 1.35868108e-02 -1.14752209e+00 -2.04047948e-01
1.65276423e-01 -1.54815316e-02 5.90142131e-01 3.38554472e-01
1.19549739e+00 -1.71072468e-01 -2.30374441e-01 7.08696425e-01
1.34120858e+00 5.14549196e-01 9.07012939e-01 5.75748265e-01
3.75592291e-01 6.61470532e-01 4.51134086e-01 5.12623191e-02
-2.77019471e-01 7.03587413e-01 6.46433473e-01 -6.63744271e-01
-1.18349500e-01 -4.28361110e-02 1.63771257e-01 3.11838776e-01
-2.16359064e-01 9.42613035e-02 -7.85797834e-01 3.83752078e-01
-1.44893074e+00 -6.06180370e-01 -2.91047782e-01 2.39351273e+00
5.88154912e-01 2.72827715e-01 -1.23109855e-01 2.90387064e-01
8.14623773e-01 -1.69417828e-01 -2.71569729e-01 -9.14792180e-01
-1.64307021e-02 3.57276827e-01 6.42237127e-01 7.35258281e-01
-1.12301040e+00 8.26872528e-01 7.28253937e+00 4.80020404e-01
-1.29457390e+00 -5.09121157e-02 7.89740741e-01 3.79111469e-01
-4.27153260e-02 -1.40860811e-01 -5.31473875e-01 3.36683184e-01
1.02969420e+00 7.75505826e-02 5.81555665e-01 6.87097967e-01
9.55856442e-02 -2.84074426e-01 -9.42153633e-01 7.24333286e-01
3.26817483e-01 -1.23768103e+00 1.54534385e-01 1.86122820e-01
7.86195576e-01 -7.92302191e-02 1.79944053e-01 7.70619093e-03
-2.32665002e-01 -1.35179889e+00 6.69698417e-01 5.93316078e-01
7.67650545e-01 -6.77334666e-01 9.30453718e-01 3.74384135e-01
-6.13890052e-01 -4.80124168e-02 -7.72223353e-01 -1.16858505e-01
-4.34686542e-01 3.12919557e-01 -1.16520584e+00 2.49922380e-01
6.63250208e-01 6.44246861e-02 -9.23700929e-01 9.06376600e-01
2.18063712e-01 2.42910132e-01 -2.77545959e-01 4.35102992e-02
3.75492096e-01 -6.61614299e-01 4.25782263e-01 9.81553078e-01
4.96076703e-01 -1.41612619e-01 -7.01039076e-01 9.90171313e-01
3.77585031e-02 2.34191626e-01 -1.01308477e+00 2.00769812e-01
6.96321651e-02 1.37535512e+00 -1.00894213e+00 -2.86034673e-01
-2.97355771e-01 8.58437538e-01 1.61652163e-01 2.14222208e-01
-7.41711199e-01 -4.99483824e-01 1.77036107e-01 4.50619996e-01
2.47679844e-01 7.15223402e-02 -4.53521520e-01 -4.18110520e-01
1.81015670e-01 -6.91196382e-01 -1.74496084e-01 -1.16786993e+00
-5.98824263e-01 1.11109281e+00 6.97084889e-02 -1.26379716e+00
-1.52041242e-01 -1.16647482e+00 -4.87067968e-01 1.03929615e+00
-7.80954242e-01 -8.22947025e-01 -6.20763719e-01 3.26369852e-01
1.53707430e-01 -6.31727159e-01 1.04834449e+00 2.11110443e-01
-1.38081416e-01 3.69097024e-01 7.24128932e-02 2.41326213e-01
6.30848646e-01 -1.20598507e+00 7.54272163e-01 7.06723750e-01
1.89367719e-02 3.97969395e-01 7.94313252e-01 -3.67147028e-01
-7.34454095e-01 -1.03145111e+00 8.68299127e-01 -3.75250965e-01
3.56674552e-01 -4.31056261e-01 -1.07320595e+00 8.54791462e-01
5.17551601e-01 2.42995158e-01 3.06307375e-01 -4.04155493e-01
-1.63899675e-01 -2.30089009e-01 -1.20711803e+00 5.50816834e-01
6.42036974e-01 -1.53453439e-01 -6.43108368e-01 2.71693081e-01
6.22398853e-01 -3.74978423e-01 -8.41614842e-01 4.06032711e-01
4.43857938e-01 -1.13177669e+00 6.99569166e-01 -6.84124053e-01
4.60502476e-01 -3.04738790e-01 1.44139558e-01 -1.55684566e+00
-3.04707944e-01 -1.87212139e-01 4.92100179e-01 6.10240638e-01
5.50317824e-01 -8.25363159e-01 6.79978967e-01 5.98048531e-02
-2.87323117e-01 -6.70089722e-01 -9.36004937e-01 -5.95864296e-01
2.67213911e-01 2.59961281e-02 5.19931555e-01 8.02734077e-01
-1.22782387e-01 7.24495724e-02 1.15607686e-01 -5.06595187e-02
1.55292869e-01 -5.59590518e-01 7.45908141e-01 -1.43141186e+00
6.38917908e-02 -3.93044442e-01 -9.31237221e-01 -5.08751690e-01
7.95263872e-02 -7.39853621e-01 -1.13440335e-01 -1.19193017e+00
1.26628444e-01 -4.93970782e-01 3.20502333e-02 5.24878383e-01
2.06634879e-01 6.13679051e-01 3.88919860e-02 -3.54691297e-02
2.69036204e-01 3.00993145e-01 1.33483887e+00 1.24287955e-01
4.21333313e-02 -3.71307991e-02 -5.06667137e-01 8.53746474e-01
7.29523420e-01 -4.93989527e-01 -4.90967721e-01 -2.23322883e-01
3.60939264e-01 -2.52946377e-01 4.03874844e-01 -1.34421897e+00
-2.13800743e-01 3.19473982e-01 7.82697082e-01 -3.75700980e-01
1.97039112e-01 -1.18288875e+00 5.31215608e-01 7.25960314e-01
-2.98588991e-01 2.98841238e-01 5.64732790e-01 8.59759971e-02
-2.73863554e-01 -6.69508696e-01 1.14265192e+00 1.79356873e-01
-2.35633850e-01 8.56292471e-02 -3.50812614e-01 -5.57113171e-01
1.34087682e+00 -3.73565435e-01 -1.49255112e-01 -3.60101432e-01
-1.09880066e+00 -4.73992527e-01 6.53791070e-01 4.95609015e-01
1.67322934e-01 -1.46686471e+00 -4.51230764e-01 8.62352669e-01
-8.68611634e-02 -1.09398076e-02 1.29895220e-02 7.89473891e-01
-1.16299224e+00 5.52813768e-01 -9.14062023e-01 -4.56875086e-01
-1.16490781e+00 9.24841106e-01 8.33185136e-01 8.46770629e-02
-6.71110213e-01 4.42140311e-01 4.41990137e-01 -4.36288923e-01
5.48659032e-03 -5.24660408e-01 -6.11087918e-01 -2.04692796e-01
3.50324929e-01 -1.30764081e-03 4.51767206e-01 -6.59562230e-01
2.48823743e-02 7.73268580e-01 2.37213876e-02 -3.20569783e-01
1.24289763e+00 3.75799149e-01 -3.28668863e-01 2.49703676e-01
1.13710999e+00 -5.64035773e-02 -1.27802312e+00 1.07183360e-01
-3.18822145e-01 -2.54380673e-01 -2.06687540e-01 -4.97409165e-01
-1.14658821e+00 8.53569210e-01 9.73304927e-01 5.46188414e-01
1.07144153e+00 -1.67981833e-01 6.58927038e-02 4.15036500e-01
5.53654246e-02 -7.89507329e-01 -7.90646449e-02 5.53181529e-01
1.18337762e+00 -9.01696861e-01 -2.54639417e-01 -3.47667187e-01
-5.03148854e-01 1.38442552e+00 7.26391077e-01 -5.70788920e-01
7.53820121e-01 3.89659226e-01 1.34602666e-01 -7.63929933e-02
-4.01799172e-01 2.08000690e-01 2.52233416e-01 7.84125030e-01
2.93021619e-01 -1.05951093e-01 -3.65554422e-01 2.58160442e-01
-4.47249860e-01 -9.15889367e-02 7.82733202e-01 4.52092320e-01
-4.90693331e-01 -8.39656293e-01 -6.90934658e-01 2.11468592e-01
-2.82257199e-01 9.95556414e-02 -4.55733150e-01 1.11589897e+00
4.27550554e-01 2.85198063e-01 8.56292725e-01 -3.89355481e-01
3.84928554e-01 8.74464437e-02 4.67800349e-01 -4.17219132e-01
-4.01028544e-01 -1.78717360e-01 -2.65394300e-01 -2.11607754e-01
-1.53341487e-01 -4.99437094e-01 -1.41586614e+00 -2.27324039e-01
2.41573229e-02 9.96209085e-02 9.01762843e-01 8.86893272e-01
2.12670222e-01 4.94372904e-01 7.04709291e-01 -9.87984240e-01
-3.43966842e-01 -1.34249461e+00 -4.41555649e-01 4.84102249e-01
1.42148837e-01 -6.23396039e-01 -4.39829141e-01 2.94696659e-01] | [9.232810020446777, 2.312486410140991] |
0142572a-70a0-4aad-a962-42b5c6ec0628 | dynamics-aware-adversarial-attack-of-3d | 2112.09428 | null | https://arxiv.org/abs/2112.09428v2 | https://arxiv.org/pdf/2112.09428v2.pdf | Dynamics-aware Adversarial Attack of 3D Sparse Convolution Network | In this paper, we investigate the dynamics-aware adversarial attack problem in deep neural networks. Most existing adversarial attack algorithms are designed under a basic assumption -- the network architecture is fixed throughout the attack process. However, this assumption does not hold for many recently proposed networks, e.g. 3D sparse convolution network, which contains input-dependent execution to improve computational efficiency. It results in a serious issue of lagged gradient, making the learned attack at the current step ineffective due to the architecture changes afterward. To address this issue, we propose a Leaded Gradient Method (LGM) and show the significant effects of the lagged gradient. More specifically, we re-formulate the gradients to be aware of the potential dynamic changes of network architectures, so that the learned attack better "leads" the next step than the dynamics-unaware methods when network architecture changes dynamically. Extensive experiments on various datasets show that our LGM achieves impressive performance on semantic segmentation and classification. Compared with the dynamic-unaware methods, LGM achieves about 20% lower mIoU averagely on the ScanNet and S3DIS datasets. LGM also outperforms the recent point cloud attacks. | ['Jiwen Lu', 'Jie zhou', 'Haowen Sun', 'Pengliang Ji', 'Ziyi Wu', 'He Wang', 'Yueqi Duan', 'An Tao'] | 2021-12-17 | null | null | null | null | ['3d-classification'] | ['computer-vision'] | [-1.66227624e-01 -1.40046597e-01 2.58815251e-02 -2.28607282e-01
-2.46441327e-02 -9.10232484e-01 4.04548049e-01 -4.75532889e-01
-5.81907094e-01 5.00474870e-01 -4.03944105e-01 -6.82526946e-01
2.75134183e-02 -7.74749219e-01 -1.07492387e+00 -9.31951940e-01
-2.75725216e-01 3.53411138e-01 8.69939566e-01 -4.01860654e-01
1.92172542e-01 8.01211596e-01 -7.65478492e-01 -5.54140285e-02
6.24176145e-01 7.65785277e-01 -1.23855069e-01 6.00087881e-01
-1.82397693e-01 6.85503900e-01 -9.17558670e-01 -5.60466647e-01
7.01564848e-01 -1.40976757e-01 -6.94015443e-01 -4.03898656e-01
4.49847788e-01 -4.89308000e-01 -8.19799542e-01 1.64181948e+00
5.32831311e-01 -1.13916747e-01 1.57692507e-01 -1.63786542e+00
-1.72376603e-01 1.00201535e+00 -6.56343102e-01 6.28595352e-01
-5.56600869e-01 3.65031034e-01 3.99102986e-01 -2.73219526e-01
4.97851044e-01 1.42132020e+00 6.57769203e-01 1.06564510e+00
-8.06427419e-01 -1.29589224e+00 8.97845209e-01 1.41727880e-01
-1.05305374e+00 -1.06407866e-01 1.05898929e+00 -1.34320542e-01
7.22867250e-01 2.67817557e-01 4.43136811e-01 1.45496023e+00
4.69842017e-01 7.51541853e-01 9.22088087e-01 2.91690916e-01
3.50504875e-01 -2.11727217e-01 9.56614316e-02 5.61968207e-01
4.40879881e-01 3.60280454e-01 -2.14335188e-01 -1.57448620e-01
7.67133594e-01 -1.14055067e-01 -2.62978405e-01 -2.44866535e-01
-8.85117292e-01 7.63367116e-01 6.32362008e-01 2.14230284e-01
-9.90405679e-02 6.16529346e-01 7.48340905e-01 5.00177383e-01
4.37217772e-01 2.31163204e-01 -7.85982132e-01 -5.98558262e-02
-7.28735507e-01 3.93760949e-01 8.73270214e-01 8.02278817e-01
4.21162933e-01 5.44162571e-01 3.29881519e-01 1.37992695e-01
1.81676835e-01 3.70720118e-01 3.14784080e-01 -7.60230720e-01
6.22791052e-01 2.57936567e-01 -2.22488478e-01 -1.28833091e+00
-4.31733370e-01 -5.93653381e-01 -1.10306227e+00 5.45060098e-01
5.36653161e-01 -5.42668939e-01 -1.17254889e+00 1.96958792e+00
5.64598203e-01 9.51416552e-01 9.40743461e-02 8.80309165e-01
5.01356781e-01 4.73083436e-01 5.50713250e-03 3.99938636e-02
6.93881094e-01 -9.81486022e-01 -5.68618536e-01 -2.19054043e-01
4.47316527e-01 -4.51322705e-01 8.96079838e-01 3.17410469e-01
-7.51244783e-01 -3.40125740e-01 -1.26498830e+00 8.08326483e-01
-3.37094992e-01 -6.56049967e-01 6.28593683e-01 9.93021429e-01
-9.05178845e-01 9.54112053e-01 -1.36964750e+00 9.35454071e-02
6.81207538e-01 5.03068864e-01 1.24344975e-02 2.75294125e-01
-1.33658326e+00 4.71861035e-01 3.85232985e-01 2.60868222e-01
-1.13221800e+00 -1.01384664e+00 -4.09915030e-01 -2.15582788e-01
4.25587505e-01 -3.91772091e-01 1.06046581e+00 -1.02244604e+00
-1.68917656e+00 4.26304013e-01 3.43830258e-01 -8.17212701e-01
1.14418328e+00 -4.12477016e-01 -2.23783806e-01 1.88060030e-01
-3.58430564e-01 3.91199589e-01 1.02412438e+00 -1.24825752e+00
-3.45875829e-01 -4.27389383e-01 4.94250834e-01 1.05586454e-01
-4.06546086e-01 -1.48008302e-01 -4.46051508e-01 -7.90664434e-01
1.29492238e-01 -1.34218025e+00 -5.24999440e-01 8.89632031e-02
-5.63954175e-01 1.66303664e-01 1.55042827e+00 -1.82939366e-01
9.70023751e-01 -2.16641140e+00 9.21892598e-02 3.26003194e-01
4.02058274e-01 6.69926822e-01 -1.27964586e-01 -4.04587612e-02
-1.57109782e-01 3.29725981e-01 -6.47885278e-02 -1.93516135e-01
-2.33843699e-01 5.19858658e-01 -6.50817573e-01 5.77380538e-01
6.97772279e-02 8.94550860e-01 -7.75010169e-01 -1.98872447e-01
7.49156028e-02 3.31689715e-01 -7.07020462e-01 -7.08626136e-02
-2.11557165e-01 6.71821892e-01 -8.21996570e-01 5.52144885e-01
1.13078821e+00 -4.64885719e-02 -1.41867008e-02 -2.30260447e-01
1.68893963e-01 -4.94039245e-02 -1.07134378e+00 1.46791863e+00
-1.26717582e-01 4.38779324e-01 1.50337219e-01 -1.12622249e+00
7.71900594e-01 1.13375187e-01 4.24670666e-01 -3.77894014e-01
3.84727180e-01 1.50442570e-01 3.28856111e-01 -1.32773012e-01
6.24852329e-02 4.40114625e-02 3.00646964e-02 2.21338183e-01
-4.08719778e-01 -1.12167977e-01 -3.49358737e-01 3.84211630e-01
1.32157433e+00 -1.00113586e-01 -3.93470854e-01 -4.19582427e-01
3.64984661e-01 1.12654939e-01 8.82301867e-01 1.03341532e+00
-7.48083234e-01 2.10533530e-01 8.52252722e-01 -6.15764856e-01
-6.41250014e-01 -9.31740820e-01 8.01271126e-02 6.20682061e-01
6.80538416e-01 -8.15829113e-02 -1.02447474e+00 -1.30133820e+00
2.92237550e-02 4.32934970e-01 -7.13465393e-01 -6.28259778e-01
-9.17180479e-01 -9.88815367e-01 1.05187929e+00 5.14779747e-01
1.02006006e+00 -8.40074539e-01 -5.29903769e-01 1.20271720e-01
1.93735063e-01 -1.03023672e+00 -4.89346951e-01 4.22623008e-02
-1.05155468e+00 -1.05921125e+00 -3.02165657e-01 -5.23775995e-01
6.45036757e-01 1.24644518e-01 7.89923131e-01 2.45548069e-01
7.55590200e-02 -5.18139489e-02 -2.77340710e-01 -5.95744431e-01
-4.78704542e-01 2.63589174e-01 3.00666392e-01 -9.27895531e-02
3.50395776e-02 -8.32729638e-01 -6.60896003e-01 4.62495923e-01
-9.77032304e-01 -2.89595634e-01 3.26564550e-01 8.08793008e-01
4.74737942e-01 5.31497002e-01 4.66306150e-01 -1.21140134e+00
4.63048577e-01 -3.72670710e-01 -8.13766420e-01 5.47102354e-02
-6.58116341e-01 -1.73041467e-02 9.32387173e-01 -1.08532190e+00
-7.51174331e-01 -1.99512675e-01 -3.99911493e-01 -1.04775405e+00
1.36536527e-02 1.33742183e-01 -5.56597769e-01 -5.93475103e-01
5.67308664e-01 -1.16665270e-02 -2.58323029e-02 -2.83524364e-01
2.18258556e-02 -8.28981847e-02 4.68601882e-01 -7.66015470e-01
1.41669273e+00 6.66013241e-01 2.18218505e-01 -3.69585544e-01
-6.12680912e-01 1.76930696e-01 -2.71795481e-01 -1.60402864e-01
5.99187255e-01 -6.49171829e-01 -8.40596199e-01 1.20681596e+00
-1.06743062e+00 -6.58758700e-01 -1.32993400e-01 2.36049518e-01
-1.96760267e-01 2.90430814e-01 -8.88267756e-01 -3.97952199e-01
-4.13792491e-01 -1.54308963e+00 4.10019904e-01 3.01604927e-01
3.27149540e-01 -1.23661411e+00 -1.66020498e-01 -1.35325775e-01
5.58832228e-01 6.11802757e-01 7.93378055e-01 -8.81631255e-01
-6.36660755e-01 -1.78858876e-01 -5.68156987e-02 4.29935038e-01
-1.17730044e-01 1.07321225e-01 -8.32574308e-01 -6.22645080e-01
3.80200833e-01 -9.56459716e-02 7.08503604e-01 2.90482551e-01
1.41938734e+00 -4.52468127e-01 -3.46930832e-01 1.16910100e+00
1.39378297e+00 5.00301540e-01 5.58496475e-01 5.99290967e-01
1.10206532e+00 7.57050887e-02 5.01594424e-01 7.58542567e-02
-1.01625532e-01 2.85833567e-01 1.11704826e+00 -3.48505229e-02
1.77713916e-01 -1.72534376e-01 2.74169326e-01 6.72989011e-01
2.22308829e-01 -3.72778684e-01 -8.94884706e-01 1.34727255e-01
-1.76871896e+00 -7.39490211e-01 -2.68572625e-02 1.72953558e+00
6.07280910e-01 7.73194253e-01 -1.45599231e-01 -3.50697935e-02
5.81903696e-01 5.46699643e-01 -1.17548358e+00 -3.51666838e-01
-3.45980451e-02 2.47503787e-01 9.99661505e-01 3.49204242e-01
-1.12950754e+00 1.23616421e+00 6.17960358e+00 8.72992814e-01
-1.47693574e+00 2.55284339e-01 6.76197171e-01 -1.22087963e-01
-2.21042097e-01 4.48544435e-02 -8.47744584e-01 5.91605067e-01
7.34582961e-01 -1.15214869e-01 3.31953108e-01 9.18504715e-01
-3.75565775e-02 3.60710025e-01 -6.78350389e-01 6.13602519e-01
-2.98584461e-01 -1.34090626e+00 1.78914174e-01 -4.07162011e-02
7.40665495e-01 2.68923372e-01 3.25966120e-01 1.71689138e-01
5.67577720e-01 -7.29539156e-01 7.40690470e-01 5.14331087e-02
4.58695501e-01 -1.21913958e+00 6.19601548e-01 5.09640038e-01
-1.04764223e+00 -9.92454514e-02 -5.25925100e-01 2.24476263e-01
4.29551452e-02 4.11979198e-01 -3.85224819e-01 4.42517668e-01
9.38929558e-01 5.31001627e-01 -3.23945224e-01 6.01925910e-01
-3.16768616e-01 9.27424252e-01 -4.34323698e-01 5.89529425e-02
6.89051449e-01 -2.37409230e-02 1.05083239e+00 8.60070646e-01
-6.57813400e-02 -7.12004825e-02 1.99289992e-01 6.39215052e-01
-1.35932967e-01 -3.32322985e-01 -5.04598260e-01 7.29066432e-02
5.92336833e-01 8.62962723e-01 -9.12640512e-01 -5.36410883e-02
-1.98475420e-01 1.09420192e+00 1.37878835e-01 4.07619387e-01
-1.26634741e+00 -3.35896850e-01 1.20937920e+00 -1.27984613e-01
1.14298619e-01 -4.27323431e-01 -3.12108576e-01 -1.10193610e+00
-3.65715921e-02 -1.03671420e+00 2.46828005e-01 -8.62897336e-02
-1.15156138e+00 7.74278820e-01 -1.12054139e-01 -9.73815799e-01
1.88782588e-01 -5.87344289e-01 -1.04778314e+00 7.11524427e-01
-1.41612840e+00 -9.00568545e-01 -9.27137062e-02 9.37080920e-01
5.59440970e-01 -3.42420667e-01 3.78398031e-01 3.94468158e-01
-1.01534534e+00 1.13162875e+00 2.74522696e-02 4.03793603e-01
5.24771750e-01 -1.05995703e+00 9.07300532e-01 1.35589778e+00
-8.74518827e-02 5.76286554e-01 8.54490876e-01 -8.82009566e-01
-1.29015803e+00 -1.36309278e+00 9.27710254e-03 -3.47851455e-01
1.00927055e+00 -2.59516090e-01 -1.14665246e+00 5.96004128e-01
-2.38437846e-01 2.30134279e-01 9.50906649e-02 -3.59612972e-01
-4.75321293e-01 -2.30797693e-01 -1.30542815e+00 8.46717715e-01
1.31540823e+00 -1.95736170e-01 -1.63506716e-01 3.52543861e-01
1.34332824e+00 -9.20726120e-01 -5.05133033e-01 6.00737035e-01
2.43736386e-01 -7.82676041e-01 1.09342349e+00 -8.53360891e-01
1.25076160e-01 -2.19701871e-01 -1.52308987e-02 -1.26992095e+00
-2.25736443e-02 -9.05310988e-01 -2.85498053e-01 9.52353358e-01
1.01300143e-01 -1.09079850e+00 1.18322158e+00 5.48287630e-01
-8.44797045e-02 -8.01665068e-01 -1.19887638e+00 -1.04100668e+00
4.91311938e-01 -4.54825550e-01 8.86685610e-01 1.09862840e+00
-7.86191106e-01 -3.03463817e-01 -3.70587319e-01 8.03253293e-01
7.90229559e-01 -2.43386641e-01 8.15547347e-01 -7.91290343e-01
-3.93206269e-01 -5.26611030e-01 -6.43535733e-01 -9.86187220e-01
4.23988998e-01 -7.18650579e-01 -2.21620262e-01 -7.10298181e-01
-2.42666170e-01 -6.77312016e-01 -4.58342135e-01 4.57461506e-01
-3.31707686e-01 5.20494394e-02 2.68281966e-01 3.34305972e-01
-2.46615872e-01 3.05084467e-01 1.41479802e+00 -2.27814481e-01
-2.67248422e-01 3.60589623e-01 -5.34115613e-01 8.51804554e-01
1.07377088e+00 -8.48262072e-01 -7.56371021e-01 -6.38623059e-01
-6.69105053e-02 -2.59662300e-01 5.26067555e-01 -1.08690798e+00
2.53804564e-01 -4.65706289e-01 9.90113243e-02 -5.42868078e-01
-1.56561993e-02 -9.56244349e-01 1.15048252e-01 1.09566116e+00
-3.81567539e-03 2.53254533e-01 4.84241813e-01 7.70445108e-01
7.34004565e-03 -2.68009473e-02 1.09521508e+00 3.30557227e-02
-6.95170045e-01 8.93435717e-01 -2.78216833e-03 2.71014601e-01
1.07692957e+00 -6.89712018e-02 -5.25603592e-01 -1.26701742e-01
-6.42182529e-01 3.16177964e-01 3.50089818e-01 6.03599131e-01
4.92686719e-01 -1.12141347e+00 -4.06670421e-01 1.89812064e-01
-5.14509201e-01 4.40470695e-01 3.41335863e-01 6.10898077e-01
-8.16219091e-01 -1.42172083e-01 -1.40170112e-01 -5.41506052e-01
-1.17354119e+00 5.98298550e-01 7.73518920e-01 -3.97105634e-01
-7.54847467e-01 1.20655155e+00 3.36796135e-01 -3.43600899e-01
4.86164212e-01 4.25020754e-02 1.69297531e-01 -3.99975419e-01
3.67374688e-01 1.84388876e-01 -1.27623588e-01 -2.10291788e-01
-4.20614958e-01 3.15747052e-01 -4.68978107e-01 1.02044322e-01
1.32445121e+00 1.51319221e-01 -1.59639139e-02 1.49254963e-01
1.13759899e+00 -1.46852344e-01 -1.68539059e+00 -1.87027752e-01
-2.67745972e-01 -4.24735457e-01 -4.08253036e-02 -4.73736286e-01
-1.85808182e+00 8.84496987e-01 7.01633394e-01 -3.21587175e-02
1.18862414e+00 -3.96419823e-01 1.04047632e+00 5.09928465e-01
5.27690887e-01 -7.97557831e-01 2.20907554e-01 6.95701659e-01
4.95335758e-01 -8.17324996e-01 -1.51300266e-01 -4.66063529e-01
-3.76349449e-01 8.76623690e-01 1.15735602e+00 -5.20574749e-01
1.06562579e+00 6.86316788e-01 2.77963400e-01 -1.60526469e-01
-4.79242593e-01 3.67716253e-01 -1.76815718e-01 4.89668429e-01
-4.85406041e-01 -9.98554006e-02 -2.44300086e-02 4.21515167e-01
-1.65824831e-01 -5.48696041e-01 3.70898753e-01 1.04495203e+00
-1.13395818e-01 -1.14952731e+00 -2.16393396e-01 1.72827899e-01
-7.22749650e-01 1.52716845e-01 -3.39954197e-01 1.11039495e+00
5.29768094e-02 5.58192730e-01 5.01725590e-03 -7.60498881e-01
3.79504412e-01 -1.94179699e-01 2.43297786e-01 -1.51684880e-01
-7.61404037e-01 -2.31279999e-01 -3.46941739e-01 -8.88025105e-01
-1.65856972e-01 -4.79931951e-01 -1.31360042e+00 -7.49465883e-01
-3.70626956e-01 -6.32065013e-02 6.91758037e-01 8.54739904e-01
1.85062036e-01 8.36875618e-01 8.92872155e-01 -7.49627709e-01
-9.73464966e-01 -4.36411291e-01 -3.57279480e-01 3.53143960e-01
2.66234010e-01 -6.00890219e-01 -7.82583237e-01 -3.98619384e-01] | [5.567728519439697, 7.886961936950684] |
e989c4c9-cea0-41c3-ab39-35127cc3e4d1 | insertionnet-2-0-minimal-contact-multi-step | 2203.01153 | null | https://arxiv.org/abs/2203.01153v1 | https://arxiv.org/pdf/2203.01153v1.pdf | InsertionNet 2.0: Minimal Contact Multi-Step Insertion Using Multimodal Multiview Sensory Input | We address the problem of devising the means for a robot to rapidly and safely learn insertion skills with just a few human interventions and without hand-crafted rewards or demonstrations. Our InsertionNet version 2.0 provides an improved technique to robustly cope with a wide range of use-cases featuring different shapes, colors, initial poses, etc. In particular, we present a regression-based method based on multimodal input from stereo perception and force, augmented with contrastive learning for the efficient learning of valuable features. In addition, we introduce a one-shot learning technique for insertion, which relies on a relation network scheme to better exploit the collected data and to support multi-step insertion tasks. Our method improves on the results obtained with the original InsertionNet, achieving an almost perfect score (above 97.5$\%$ on 200 trials) in 16 real-life insertion tasks while minimizing the execution time and contact during insertion. We further demonstrate our method's ability to tackle a real-life 3-step insertion task and perfectly solve an unseen insertion task without learning. | ['Dotan Di Castro', 'Vladimir Tchuiev', 'Oren Spector'] | 2022-03-02 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 0.21884915 0.02903462 -0.06221942 -0.10276025 -0.7377117 -0.48379192
0.22770186 0.04218391 -0.8637588 0.7048397 -0.27078566 -0.31097347
-0.47961485 -0.52099186 -1.0097891 -0.50255364 -0.28379476 0.7618171
0.4431591 -0.79125196 0.43466803 0.45906407 -1.8359853 0.06966661
0.993759 0.94593513 0.77934134 0.64626515 0.21794434 0.5135994
-0.41331175 -0.14303586 0.55657315 0.15671322 -0.72409445 -0.15974058
0.18051082 -0.42913514 -0.05576836 0.68809116 0.7105883 0.55602336
0.4607604 -1.0741318 -0.05118607 0.462595 -0.39342773 -0.11922657
0.80031425 0.5682352 0.8564145 -0.784696 0.72461 1.1191003
0.7415511 0.50886 -1.1542635 -0.409249 -0.0797768 0.34685925
-0.936013 -0.2235437 0.6982255 -0.31938434 1.3177307 0.01922228
0.9545374 1.3364685 0.21702893 0.9797498 0.8325535 -0.50574976
0.42429855 -0.01011625 -0.36595795 0.6336263 0.05934666 0.38330403
-0.36268708 0.2482401 0.85746205 0.14352223 -0.12742135 -1.3009684
-1.2922018 0.6219922 0.7336256 0.18530326 -0.5110258 0.22644775
0.57003456 0.51163965 -0.05938807 0.8518017 -0.79310405 -0.5849416
-0.3985186 0.42364037 1.0082301 1.0111358 0.7169591 -0.15138122
0.21774717 0.8554663 -0.08176585 0.38191926 0.32902464 -1.149974
0.7319978 0.45786944 0.44968745 -0.87863684 -0.7114746 -0.28764322
-0.33927155 0.7130668 0.42614186 -0.12964171 -0.8424165 1.5340418
0.36200586 -0.33251926 -0.07751504 0.89367104 0.30393398 0.19048396
-0.25799316 0.12910046 0.8286128 -1.1533116 -0.5032016 -0.43715894
0.7107114 -0.46267092 1.4072454 0.9923398 -1.0690656 -0.52775085
-1.1394532 0.07592415 -0.4257978 -0.08768563 0.7670563 0.24109854
-0.8826429 1.0587629 -0.99349236 -0.34694558 0.18393584 0.8792998
-0.71333414 -0.11521934 -0.81551147 1.3516445 0.70099705 0.1682519
-0.89019513 -0.32793146 -0.9654064 -0.23275325 1.0649868 -0.47816163
1.4500352 -0.66581523 -1.8872943 0.682159 0.573592 -0.5002852
0.72705007 -0.7768124 0.3415187 0.05710875 -0.08207397 0.7983803
0.7853962 -1.2509353 -0.4803475 -0.17425394 0.43256044 0.3700705
-0.23000059 -0.45491493 -0.30631557 -0.33145773 -0.060568 -1.1262858
-0.4929476 0.22654328 -0.12046257 -0.15018395 0.584095 -0.42789435
0.4473934 -2.06175 0.9001196 0.10695805 0.00855322 0.2984551
-0.27717146 0.73457617 -0.02390221 -0.4898597 -0.15600544 -0.25288072
0.07051824 0.60056096 0.14712608 0.33131608 -0.10835511 0.93899673
-1.1767337 -0.20272008 0.73588824 0.19577049 -0.7351434 0.43233252
-0.29111943 0.46645695 -0.06900097 0.54127157 0.27055573 0.14601573
0.13197586 0.04147173 0.07515562 -0.07001841 -1.2381611 2.50106
-0.9558981 0.16268978 0.2822847 -1.0511113 1.0327435 0.01328569
0.5890888 -0.771232 0.3781567 0.5870156 0.08764412 -0.8075019
0.39543083 0.03819343 -0.2984852 0.28028956 0.42789766 -0.62787545
0.15180579 -0.10187955 1.2013979 0.74822617 0.35038802 -0.18380848
0.32225487 -0.06806051 0.12739371 0.8507863 -0.17091607 0.44955048
0.15975931 -0.6466868 -1.0882154 -1.122701 0.5108504 1.3044792
0.37473908 -0.05278035 -0.53216183 -0.54626125 0.06449527 0.7422475
-0.6522836 -0.36207816 -0.73685163 -0.07495231 0.02658841 0.7817584
0.19031204 -1.5540649 -1.3302593 0.33038074 -0.0971944 -0.94425297
-0.03794891 0.6775708 -0.8431906 -1.1634135 -0.6162274 -0.9591761
0.57891357 0.15622015 0.97497296 -0.08782655 -0.6828634 0.44661218
-0.64084566 -0.3207115 -0.18249711 0.17102075 0.09812044 -0.6886537
-0.1445018 -0.7890783 -0.5035042 0.28212926 -0.5993487 0.03678045
0.9556479 1.0929332 0.12307165 -0.14654192 0.3295373 -0.48195887
0.77149475 -0.10902251 -0.55899405 0.07428665 -0.3510698 0.11570583
0.49169132 -0.6507817 -0.79315436 0.39867175 -0.2764768 -0.50211024
-0.1587474 0.21204473 0.11695413 -0.4145736 0.79836136 0.11877695
0.33589628 -0.475991 0.5955739 0.3622754 0.8134767 -0.8097993
0.68765646 0.22289737 -0.07381063 -0.33620545 -0.5033608 -0.5403585
-1.0727516 -0.16908878 0.4530119 -0.6290359 -1.1831378 0.54300934
-1.011593 -0.666618 -0.46314606 0.45872077 -1.3379196 0.34406418
-0.6151983 -0.855039 -0.21719223 -1.1537143 1.1233441 0.08250364
-0.16154446 -0.647141 0.12627798 0.13121521 0.49987662 0.33416435
0.6464068 -0.62547886 -0.418946 -0.3793513 0.08038194 0.23872761
0.06137532 -0.38730863 -0.48838413 -0.6418192 -0.16595928 -0.78582627
0.6469177 0.00929572 1.027535 -0.05599226 -0.15986483 0.37522763
1.39298 0.29337355 0.60366935 0.699452 0.514148 0.82828325
0.96505696 0.46494052 0.12581015 0.9325632 1.076116 0.09051739
0.32781065 -0.3411547 0.12349565 0.50357246 -0.35020643 -0.04775129
-0.8844577 0.570061 -2.0797749 -0.726963 0.44141605 2.3302662
0.68074435 0.4398771 0.28193474 0.34798366 0.20397665 -0.21459404
-0.64894193 -0.53350073 0.5339047 0.51595926 0.44391587 0.3200272
-1.0153921 0.9948183 6.5951667 0.71955293 -0.95210457 -0.32531965
-0.23037471 -0.13646838 0.19434576 -0.29157153 -0.2383089 0.16386911
0.4372505 0.12038447 0.6162331 1.1286857 -0.0074809 -0.4601831
-1.2090605 0.93035245 0.03030742 -0.86543596 -0.5657436 -0.14930247
0.3454982 -0.00820267 0.02482815 0.74825454 0.3341605 -0.9202016
0.57453036 0.24446826 0.6314221 -0.7848653 0.87200594 0.79504323
-0.8585502 -0.5187432 -0.18485084 -0.40148288 0.13345507 -0.01001612
-1.0416719 0.6707692 0.7571167 0.7070619 -0.1934818 1.3154784
-0.29540104 -0.27373463 -0.38149962 -0.47735757 0.27444288 0.02102967
0.55461514 0.8570139 0.18144669 -0.25557455 0.3885972 0.4536574
0.4539194 0.16090956 -0.92167056 0.2939449 0.15479903 1.101946
-0.63382787 0.05322646 -0.03101037 1.1369526 0.6270599 0.07944767
-0.76023054 -0.715414 0.26610112 -0.07910331 0.5720925 -0.6372087
-0.00869738 -0.79790044 0.20062083 -0.89847046 -0.05779781 -0.8500585
-0.96132755 0.44870993 0.11464354 -1.4452547 -0.6214873 -0.9640692
-0.4804744 0.4793873 -1.1579932 -0.83270735 -0.36029562 0.4975728
0.72493815 -0.09785073 1.0088111 0.14004503 -0.16755146 0.29336387
-0.09415726 -0.35229543 0.5848843 -1.3623741 0.4014817 0.20377053
-0.1845006 0.5813658 1.0076557 -0.3801637 -1.6771085 -0.30183196
0.23307829 -0.3336261 0.6855095 -0.6040901 -0.60629636 0.49042812
-0.01448558 -0.16546524 0.20575488 0.3173385 -0.10432652 -0.06467002
-1.3162324 0.64380366 1.4423264 -0.11512358 -1.0555689 0.3052255
0.8443877 -0.90840554 -0.6513215 0.64916193 0.791047 -0.9916333
1.2081189 -0.5083047 0.48239985 0.04117091 -0.02583264 -1.4487331
-0.2590072 -0.55479765 -0.05621959 0.26853603 0.27540612 -0.5314532
0.8445761 0.2316516 -0.24904564 -1.0722171 -0.99505866 -0.9112969
-0.33270735 -0.52068526 0.08684006 0.2331281 0.28748855 0.01384065
-0.64516634 -0.34853172 0.46709806 0.04572422 1.1096652 -1.1169498
-0.5263021 -0.21629915 -0.5309161 -1.087945 0.12778567 -0.43816906
0.70814174 -1.5095723 -0.06299996 -0.38862944 -0.30781555 0.5803931
0.08033894 0.01772526 0.5054534 0.00964696 -0.79257244 0.6152276
1.4704719 -0.0510362 -0.32444936 0.20707214 -0.31830072 0.6832772
0.7958961 -0.23263532 -0.5729751 -0.265308 0.2769047 0.2347388
0.40516403 -1.2661456 0.19717593 -0.12187785 0.13116294 -0.56311464
0.7024474 -1.1469232 -0.04150788 0.88076276 -0.24038616 0.03123082
0.3647229 0.64803773 0.15067923 -0.35799822 0.31236744 -0.405307
-1.1308566 -0.09237081 -0.18647552 -0.08686381 1.3782228 -0.5261836
0.10214987 -0.38322604 -1.0566607 0.4391523 0.5367024 0.7697889
0.72217196 -1.0061469 -0.23841669 0.4104584 0.13819219 0.18132305
0.24452293 0.70809317 -0.7091393 0.2669946 -0.7361983 -0.69453216
-1.018976 0.68027794 0.29749835 -0.25856143 -0.9468192 1.0066588
-0.27193725 -0.94439304 0.67971015 -0.2223044 -0.06065377 -0.31351158
0.12435896 0.38276434 0.10026002 -0.01338745 -0.36741486 0.74598813
-0.03307454 -0.16474696 1.5966582 0.13127704 0.3158982 0.56151575
0.92273295 -0.3697722 -1.6036903 -0.03106852 -0.02301672 -0.5856345
-0.5381502 -1.1458027 -0.7823239 0.77647537 0.48702452 -0.2438826
0.91312873 -0.21969187 0.7939086 1.022834 1.2267683 -1.2830831
0.63149816 0.5542486 1.1715862 -1.3690596 -0.08098195 -0.353213
-0.6105345 1.1945623 0.92389035 -0.43653104 0.17356989 0.3652006
-0.00936756 -0.10393964 -0.7439544 -0.04253822 0.09234323 0.86790866
-0.01635786 -0.02034528 -0.18209456 0.18048157 -0.26911378 0.10002022
0.32918358 1.4540278 -0.6894652 -1.0982215 -0.13426216 0.23853901
0.0067976 0.39211342 -0.19751959 0.9892343 0.05794296 0.6608267
-0.22935635 -0.6157091 0.7095256 -0.2038333 0.9933738 -0.74015975
-0.6526442 -0.14028437 0.01806518 -1.0631641 -0.20892681 -0.49230263
-1.223152 -0.08465166 -0.3122415 -0.2869389 0.76333123 0.87270844
0.10342504 0.60772586 0.4241886 -1.6900818 -0.97279 -0.83327806
-0.3777087 0.3846564 0.23408496 -1.1386526 -0.23229264 -0.50973773] | [4.62382173538208, 0.8026793599128723] |
fc6d591f-8f65-47f4-b6d1-7c059bbb54cd | time-to-green-predictions-for-fully-actuated | 2208.11344 | null | https://arxiv.org/abs/2208.11344v1 | https://arxiv.org/pdf/2208.11344v1.pdf | Time-to-Green predictions for fully-actuated signal control systems with supervised learning | Recently, efforts have been made to standardize signal phase and timing (SPaT) messages. These messages contain signal phase timings of all signalized intersection approaches. This information can thus be used for efficient motion planning, resulting in more homogeneous traffic flows and uniform speed profiles. Despite efforts to provide robust predictions for semi-actuated signal control systems, predicting signal phase timings for fully-actuated controls remains challenging. This paper proposes a time series prediction framework using aggregated traffic signal and loop detector data. We utilize state-of-the-art machine learning models to predict future signal phases' duration. The performance of a Linear Regression (LR), a Random Forest (RF), and a Long-Short-Term-Memory (LSTM) neural network are assessed against a naive baseline model. Results based on an empirical data set from a fully-actuated signal control system in Zurich, Switzerland, show that machine learning models outperform conventional prediction methods. Furthermore, tree-based decision models such as the RF perform best with an accuracy that meets requirements for practical applications. | ['Anastasios Kouvelas', 'Monica Menendez', 'Lukas Ambühl', 'Kaidi Yang', 'Michail A. Makridis', 'Alexander Genser'] | 2022-08-24 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 3.91601920e-01 -2.95100689e-01 -9.02534187e-01 -7.66694784e-01
-9.76695836e-01 5.36211953e-02 5.74946642e-01 -5.44726476e-02
-3.61602783e-01 9.57075834e-01 2.98185963e-02 -9.20604348e-01
-1.96324632e-01 -8.73799205e-01 -4.65865433e-01 -3.65660906e-01
-3.95385712e-01 6.25366151e-01 6.29569650e-01 -4.92319077e-01
1.48315340e-01 9.66252327e-01 -1.62559211e+00 3.69355857e-01
4.83652979e-01 1.14002502e+00 -4.65573482e-02 9.45764720e-01
1.12913959e-01 1.04491103e+00 -7.18186140e-01 2.48044133e-01
2.87937433e-01 1.47957236e-01 -3.22440237e-01 -5.62282801e-01
3.08668911e-01 -3.10389072e-01 -9.61542010e-01 -8.66797194e-02
3.77368242e-01 2.75600314e-01 6.41738117e-01 -1.79004943e+00
4.72960293e-01 4.84200925e-01 -2.26550981e-01 4.70187068e-01
2.26843089e-01 5.85136950e-01 7.28333175e-01 -3.01913284e-02
4.67709661e-01 1.01375890e+00 8.53927255e-01 1.42478064e-01
-1.53542769e+00 -9.87675846e-01 9.77150649e-02 5.65715373e-01
-1.40606713e+00 -4.77294326e-01 8.13707530e-01 -6.16921127e-01
1.27844691e+00 2.87044168e-01 7.19468176e-01 9.53554213e-01
8.10585916e-01 8.36012900e-01 8.17868829e-01 -2.16894075e-01
1.90091953e-02 -1.20744206e-01 5.03363721e-02 3.32089901e-01
-2.18903050e-01 6.79059863e-01 -3.91339213e-01 1.12150423e-01
3.30842435e-01 -2.68066168e-01 5.42535260e-02 -6.19858764e-02
-9.34316039e-01 6.32938802e-01 3.36329192e-01 2.55495906e-01
-5.27072132e-01 4.66865718e-01 4.54253137e-01 4.55889434e-01
4.14487392e-01 1.98988214e-01 -5.87117553e-01 -4.95821804e-01
-1.35574412e+00 7.55915284e-01 7.02398598e-01 9.99438941e-01
7.54961014e-01 5.83071470e-01 -5.95210910e-01 5.23260653e-01
1.98739991e-01 7.53202081e-01 -3.32976468e-02 -1.01116312e+00
7.13091016e-01 1.54410437e-01 2.11511210e-01 -1.19800341e+00
-8.57966781e-01 -5.01037896e-01 -4.72471803e-01 3.12325060e-01
5.11308074e-01 -4.66784805e-01 -7.58680224e-01 1.54001176e+00
-4.97948043e-02 3.87778640e-01 -2.76002020e-01 5.66909134e-01
6.87285960e-02 9.10600066e-01 2.30694234e-01 -8.98411572e-02
6.32267416e-01 -7.70505309e-01 -7.17347682e-01 -4.43644077e-01
8.21953297e-01 -5.82448244e-01 3.44128877e-01 2.89358944e-01
-7.87623644e-01 -1.01833153e+00 -1.19033027e+00 4.07886058e-01
-4.13812816e-01 -2.33309921e-02 4.74117637e-01 7.64002204e-01
-7.93282390e-01 6.39669836e-01 -8.79726291e-01 -8.34555849e-02
2.48697415e-01 5.65011084e-01 2.59641260e-01 1.64680392e-01
-1.36326408e+00 1.32674170e+00 -9.04107019e-02 1.83384448e-01
-8.30330551e-01 -9.70223069e-01 -7.76333034e-01 -3.76009762e-01
2.29889050e-01 -4.25087839e-01 1.68706512e+00 -4.39807028e-01
-1.52735829e+00 1.48201182e-01 -4.66103613e-01 -1.18410885e+00
5.92243135e-01 -1.67521492e-01 -1.11330414e+00 -1.33118957e-01
1.83602437e-01 6.84915304e-01 6.63038433e-01 -1.11654878e+00
-1.29184663e+00 1.67880684e-01 -3.06002349e-01 -3.08815002e-01
5.52699685e-01 -1.76569074e-01 -1.41527727e-01 -4.93039191e-01
-4.48079258e-01 -1.13611042e+00 -5.64976454e-01 -4.40290540e-01
-3.30700576e-01 -1.09969057e-01 9.03415799e-01 -6.29901230e-01
1.84801340e+00 -1.49229610e+00 -5.14126837e-01 6.78603470e-01
-3.63834918e-01 2.27940962e-01 1.07304439e-01 6.09461188e-01
-6.63580745e-02 -3.27289939e-01 1.52693063e-01 -3.52940708e-02
7.72582134e-03 4.23173130e-01 -6.11199141e-01 3.57062906e-01
1.21795990e-01 7.26731360e-01 -6.26848936e-01 -4.97634619e-01
8.83924544e-01 1.86456423e-02 -2.85268933e-01 1.77768692e-02
-3.15487146e-01 6.48590982e-01 -6.32524788e-01 4.34029579e-01
2.19178319e-01 4.97913986e-01 -1.10493585e-01 -1.85005605e-01
-6.00795686e-01 4.16727662e-01 -7.79383361e-01 1.18877518e+00
-9.27079856e-01 1.24336672e+00 -1.81363210e-01 -1.00464201e+00
1.18976998e+00 2.18250856e-01 9.31381345e-01 -1.08105218e+00
1.89867944e-01 1.43418595e-01 -6.31913692e-02 -4.81356770e-01
7.07215607e-01 2.41717547e-02 -4.08093721e-01 1.39719874e-01
-4.00798053e-01 -1.86154082e-01 3.87170643e-01 -1.81080878e-01
1.26830912e+00 8.50550756e-02 -1.70814618e-01 1.36911660e-01
5.43777347e-01 4.64641452e-01 5.46439648e-01 6.92701757e-01
-3.28543037e-01 3.40612195e-02 2.09748447e-01 -7.02837944e-01
-6.93035364e-01 -8.23152483e-01 -4.84270006e-02 1.13968945e+00
5.31189516e-02 -1.89622670e-01 -3.19831431e-01 -2.11656585e-01
2.78768122e-01 1.47935843e+00 -2.08473668e-01 -1.55754343e-01
-1.09636199e+00 -2.40626663e-01 6.39857471e-01 6.26033962e-01
1.61235601e-01 -5.92944443e-01 -6.17726088e-01 7.26771414e-01
-9.34573486e-02 -1.24375939e+00 -2.30824694e-01 1.00680493e-01
-5.73455572e-01 -8.90203536e-01 -5.90317585e-02 -1.73465505e-01
-2.37168577e-02 1.17720276e-01 9.47156787e-01 -3.35527420e-01
-1.13326728e-01 2.31837958e-01 1.20991111e-01 -6.31670415e-01
-4.82682317e-01 1.79084599e-01 2.56358813e-02 1.14031866e-01
4.97189850e-01 -4.43727553e-01 -3.60639453e-01 8.94368649e-01
-1.57995686e-01 6.69894964e-02 5.82466006e-01 2.86717683e-01
4.63539630e-01 1.16835520e-01 7.49522448e-01 -3.20783973e-01
4.30345505e-01 -4.77288306e-01 -6.69192135e-01 -2.00930685e-01
-8.20281684e-01 -5.28634451e-02 5.58411241e-01 -2.05829844e-01
-1.09218872e+00 3.84089425e-02 -3.04329485e-01 -2.89549917e-01
-3.72594237e-01 2.19039321e-01 4.81935740e-02 1.16048669e-02
6.03033245e-01 1.16181687e-01 -3.42783448e-03 -1.61604017e-01
1.67656571e-01 9.17063475e-01 5.06538033e-01 -3.71034056e-01
8.85279298e-01 4.47331101e-01 3.29418689e-01 -1.22901046e+00
-5.11089861e-01 -3.63328516e-01 -6.45173311e-01 -1.01585913e+00
6.21323824e-01 -7.82669961e-01 -7.50919163e-01 1.93713099e-01
-1.01132119e+00 -5.00930309e-01 -2.54843712e-01 7.22987115e-01
-1.00808954e+00 -3.10862869e-01 -3.91852230e-01 -1.04853606e+00
2.55108058e-01 -1.02582836e+00 9.66198206e-01 5.25462031e-02
-7.03414857e-01 -9.25473750e-01 1.59082323e-01 5.56195796e-01
7.13166833e-01 5.95891058e-01 7.60997593e-01 -4.61519510e-01
-7.82579839e-01 -6.62135303e-01 1.76024973e-01 2.99393497e-02
-2.39851847e-01 8.73453245e-02 -7.49019921e-01 7.28921220e-02
-6.12710297e-01 2.02162638e-01 8.21398973e-01 1.01908410e+00
9.19946611e-01 -3.85139957e-02 -1.02624464e+00 2.00164974e-01
9.44634080e-01 7.14202881e-01 3.79950613e-01 4.80041236e-01
1.47790164e-01 6.63594961e-01 1.05332208e+00 3.06066215e-01
6.80488706e-01 1.07826853e+00 3.57110263e-03 -6.54133707e-02
-1.79625988e-01 -6.38256729e-01 3.09037119e-01 2.49074236e-01
7.00017810e-03 -3.27943444e-01 -1.04163325e+00 5.30011833e-01
-1.79886019e+00 -1.23644483e+00 -6.27700388e-01 2.24339652e+00
2.80766010e-01 9.43291724e-01 4.37781960e-01 4.50944602e-01
6.63058102e-01 2.70291239e-01 -4.22967732e-01 -8.44340920e-01
3.00124079e-01 7.47653916e-02 1.20931625e+00 5.02820551e-01
-1.20275939e+00 9.29413974e-01 7.31269598e+00 1.21325850e+00
-1.20214748e+00 -2.07266867e-01 4.25253183e-01 -1.82138234e-01
1.92447975e-01 -6.19349815e-03 -1.10538065e+00 5.60116768e-01
1.81854045e+00 -1.30028829e-01 9.73658115e-02 6.73507571e-01
1.14582717e+00 -2.09094986e-01 -1.05013943e+00 5.44956028e-01
-4.58811700e-01 -1.54066658e+00 -2.75775284e-01 -1.14683574e-02
2.55483121e-01 3.56978655e-01 -6.77678809e-02 8.55989635e-01
2.93440729e-01 -9.62206364e-01 7.55516648e-01 1.01731479e+00
5.88184714e-01 -9.10173118e-01 6.71214700e-01 3.08958918e-01
-1.51338446e+00 -3.33118916e-01 1.99995995e-01 -2.76731491e-01
1.07871294e+00 3.85895044e-01 -9.86483097e-01 7.64149308e-01
3.42587590e-01 7.76298285e-01 -3.25087011e-01 1.05003524e+00
1.22635186e-01 1.00016439e+00 -5.59086144e-01 -1.71704963e-01
4.00380611e-01 2.49101132e-01 6.47617459e-01 1.42946291e+00
6.88700005e-02 -2.62701571e-01 4.28384900e-01 3.06420475e-01
7.06458092e-01 -3.68610561e-01 -8.98898482e-01 2.95933455e-01
4.48009670e-01 7.43136764e-01 -2.51223266e-01 -6.21531606e-02
-4.35978383e-01 -1.42155856e-01 -3.43805611e-01 5.84555328e-01
-1.21796501e+00 -5.22221923e-01 8.51965964e-01 6.78309500e-01
2.08475426e-01 -5.90325236e-01 -3.44964027e-01 -2.37148866e-01
-3.79993618e-01 -3.00438583e-01 2.33880773e-01 -6.40525162e-01
-8.40428352e-01 5.58203340e-01 5.95190346e-01 -1.82111609e+00
-7.88124323e-01 -4.78210658e-01 -7.27428675e-01 6.24400139e-01
-1.60597837e+00 -1.15247333e+00 1.06563479e-01 3.25043768e-01
7.49678850e-01 -3.49259824e-01 2.80281514e-01 5.48100352e-01
-6.31375492e-01 3.10826361e-01 -2.04045385e-01 -9.72475577e-03
4.30836260e-01 -6.95856392e-01 3.52833837e-01 4.30671662e-01
-2.46836185e-01 -1.57469865e-02 1.12080181e+00 -7.18950033e-01
-1.32344389e+00 -1.43618691e+00 9.68449771e-01 -3.22622001e-01
8.20296049e-01 1.55885160e-01 -4.11087871e-01 9.11101997e-01
-3.81328799e-02 -2.27918312e-01 6.98059738e-01 -6.91269413e-02
3.80824864e-01 -7.02000439e-01 -7.60449111e-01 6.26836956e-01
7.07931578e-01 -1.25445858e-01 -6.36217833e-01 7.37525225e-02
3.40674788e-01 -1.77895546e-01 -8.62029612e-01 6.61193132e-01
8.16363990e-01 -7.73525417e-01 1.02679777e+00 -5.49756706e-01
-2.47848079e-01 -3.70657384e-01 -1.68532342e-01 -1.27271807e+00
-2.56498665e-01 -8.88707459e-01 -1.99147061e-01 7.56417155e-01
7.02308178e-01 -5.62423766e-01 9.01768327e-01 6.70259833e-01
-4.07853574e-01 -8.04893672e-01 -1.35980773e+00 -8.91026795e-01
-1.77409261e-01 -1.45626581e+00 5.62069774e-01 1.39797837e-01
-2.13364229e-01 6.17643833e-01 -6.06651604e-01 3.28162163e-02
4.80728388e-01 1.67239070e-01 1.21377635e+00 -1.21393657e+00
1.33546889e-01 -5.79638481e-01 -6.49901450e-01 -1.49654627e+00
3.60155821e-01 -6.55291438e-01 7.56651089e-02 -1.40890205e+00
-9.44637716e-01 -6.88310742e-01 -4.14880067e-02 3.68616760e-01
5.69946706e-01 -2.26214215e-01 -1.18992321e-01 -6.27189949e-02
-3.27746004e-01 6.59352124e-01 8.67007017e-01 -3.85147870e-01
-4.15466100e-01 8.22476745e-01 1.80641770e-01 6.55278802e-01
1.05616260e+00 -4.50199664e-01 -2.58732855e-01 -1.20679915e-01
-3.35634142e-01 5.60121298e-01 1.53252959e-01 -1.55726182e+00
6.50587320e-01 -4.53900725e-01 2.84494489e-01 -1.36269653e+00
6.46696866e-01 -9.85751271e-01 2.07402378e-01 7.56263852e-01
-4.80943203e-01 1.70622338e-02 5.21460116e-01 9.33674634e-01
-2.00455874e-01 2.75351346e-01 4.50433016e-01 5.59883952e-01
-8.12950015e-01 2.98167467e-01 -1.33868301e+00 -3.09517860e-01
1.40511250e+00 -5.81633449e-01 -3.72505635e-02 -6.73694074e-01
-5.72477341e-01 7.52479792e-01 -2.39011496e-01 7.75708735e-01
4.89232987e-01 -1.25335836e+00 -6.18596256e-01 3.43804240e-01
1.34107415e-02 -4.85922903e-01 2.77270049e-01 9.63696241e-01
-4.61953998e-01 1.21698022e+00 -1.87862918e-01 -7.90091932e-01
-8.98066521e-01 2.82296330e-01 3.51708204e-01 -4.12677258e-01
-5.00782967e-01 6.53188080e-02 -7.72033036e-01 -4.26605284e-01
1.60870001e-01 -6.28413737e-01 -2.65286207e-01 -1.47527596e-02
3.72532070e-01 9.36098576e-01 3.37449345e-03 -9.82901692e-01
-2.80192614e-01 5.98651826e-01 3.64168435e-01 -1.90760747e-01
9.83213305e-01 -1.38529196e-01 6.85911059e-01 6.78828180e-01
9.62568164e-01 -3.54672492e-01 -1.26233017e+00 -1.15198866e-01
3.24937314e-01 -3.52586597e-01 4.71711606e-01 -6.07600808e-01
-9.27328885e-01 6.60697103e-01 6.06429577e-01 2.48291895e-01
9.53693092e-01 -5.14596879e-01 1.25511968e+00 3.37928981e-01
7.84284890e-01 -1.17490458e+00 -7.56138682e-01 5.12986064e-01
5.13837636e-01 -9.75771129e-01 -3.02852660e-01 -1.66469947e-01
-6.51799917e-01 1.12974596e+00 5.25943458e-01 6.00907058e-02
9.77563798e-01 6.15752041e-01 -2.58638691e-02 2.37645656e-01
-1.26730514e+00 -3.40451926e-01 2.96483278e-01 6.29194081e-01
1.45700444e-02 2.24033967e-01 -1.33103326e-01 3.10307890e-01
-3.38744462e-01 3.25114131e-01 2.11993754e-01 1.04106164e+00
-7.21505046e-01 -8.86259735e-01 -2.97664344e-01 7.31479764e-01
-5.82135990e-02 5.17912686e-01 9.19975713e-02 1.08696830e+00
2.30228733e-02 1.54685724e+00 3.28026205e-01 -8.85807931e-01
8.57790351e-01 -2.15458702e-02 1.84386969e-02 -1.85087681e-01
-4.77292776e-01 -2.60532260e-01 7.30288088e-01 -8.98995757e-01
-1.61750212e-01 -6.68506682e-01 -1.33780432e+00 -4.27187949e-01
2.14065779e-02 1.79928586e-01 7.58067608e-01 1.05812240e+00
3.30504119e-01 8.97804797e-01 8.29484940e-01 -1.15515280e+00
-1.77140519e-01 -9.01703417e-01 -2.60881513e-01 -2.18800277e-01
5.22954166e-01 -9.76061404e-01 -1.54539853e-01 -1.59988508e-01] | [5.753003120422363, 1.0844441652297974] |
a72034cd-2016-4f5b-8d2e-3f6a6618987e | nicts-unsupervised-neural-and-statistical | null | null | https://aclanthology.org/W19-5330 | https://aclanthology.org/W19-5330.pdf | NICT's Unsupervised Neural and Statistical Machine Translation Systems for the WMT19 News Translation Task | This paper presents the NICT{'}s participation in the WMT19 unsupervised news translation task. We participated in the unsupervised translation direction: German-Czech. Our primary submission to the task is the result of a simple combination of our unsupervised neural and statistical machine translation systems. Our system is ranked first for the German-to-Czech translation task, using only the data provided by the organizers ({``}constraint{'}{''}), according to both BLEU-cased and human evaluation. We also performed contrastive experiments with other language pairs, namely, English-Gujarati and English-Kazakh, to better assess the effectiveness of unsupervised machine translation in for distant language pairs and in truly low-resource conditions. | ['Rui Wang', 'Masao Utiyama', 'Kehai Chen', 'Atsushi Fujita', 'Haipeng Sun', 'Benjamin Marie', 'Eiichiro Sumita'] | 2019-08-01 | null | null | null | ws-2019-8 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 1.46885708e-01 2.50273287e-01 -2.92038649e-01 -4.70536560e-01
-1.46854937e+00 -7.72402763e-01 9.60359454e-01 2.14800891e-02
-8.20954084e-01 1.31464326e+00 5.82896531e-01 -7.57838547e-01
1.41173610e-02 -3.45101744e-01 -5.40456414e-01 -3.49216729e-01
4.65364605e-01 1.24538255e+00 -3.10025960e-01 -5.71365058e-01
1.22318096e-01 1.86374575e-01 -4.55187500e-01 5.70621908e-01
9.79477644e-01 2.40511894e-01 9.88432094e-02 5.09633601e-01
1.76116705e-01 3.81605178e-01 -3.86411816e-01 -1.01665556e+00
3.69238466e-01 -9.71364498e-01 -1.19445038e+00 -3.83355260e-01
3.18707138e-01 -7.96409696e-02 -2.59706497e-01 8.39981079e-01
7.63363063e-01 9.82449204e-02 6.44929171e-01 -4.59808797e-01
-7.96773136e-01 1.25297010e+00 -1.51484594e-01 4.08884406e-01
2.48115227e-01 -1.86050963e-02 1.16223204e+00 -1.25469875e+00
1.02339065e+00 8.55451345e-01 4.63621140e-01 4.23177212e-01
-1.05002034e+00 -3.76605511e-01 -2.98063666e-01 3.63990366e-02
-1.11056888e+00 -8.78223240e-01 4.03173149e-01 -1.97327495e-01
1.55874252e+00 4.38627511e-01 3.29945624e-01 1.29252934e+00
-7.76563361e-02 8.50670993e-01 1.45786703e+00 -9.51519549e-01
-1.14183016e-01 2.96776712e-01 -2.14488834e-01 3.15677136e-01
-6.99221790e-02 2.72300929e-01 -7.33011484e-01 -1.94690689e-01
4.38510239e-01 -7.65467465e-01 -1.24488145e-01 3.59929740e-01
-1.97239089e+00 6.76135719e-01 -1.36686666e-02 7.14369714e-01
-3.57741684e-01 -1.92584351e-01 4.51497734e-01 9.38842595e-01
9.09574151e-01 7.38455534e-01 -1.07517350e+00 -5.24970829e-01
-1.00380707e+00 -5.70883602e-02 9.78012681e-01 1.10111284e+00
3.56310368e-01 1.11650184e-01 -1.31213263e-01 1.21868300e+00
-2.88127996e-02 6.17864013e-01 4.74603236e-01 -5.97793162e-01
9.43440676e-01 -1.33767232e-01 1.08924545e-01 -4.67064261e-01
-9.47436690e-02 -5.47719419e-01 -8.65230262e-01 -4.61500794e-01
5.61249197e-01 -4.44135308e-01 -7.63864934e-01 1.62403321e+00
-2.85091192e-01 -6.25137150e-01 4.91647780e-01 9.19996083e-01
7.23025620e-01 7.48167396e-01 -3.25656682e-01 -7.38614678e-01
1.09671593e+00 -1.22860682e+00 -6.87374532e-01 -1.43680677e-01
9.02812123e-01 -1.37398815e+00 9.52355266e-01 5.08036688e-02
-1.42385828e+00 -3.61700892e-01 -7.20882952e-01 -1.52562261e-02
-4.46202248e-01 4.89554822e-01 4.62700635e-01 2.92114317e-01
-1.32008255e+00 6.22801840e-01 -6.39963567e-01 -8.89115155e-01
-8.91406611e-02 4.24366027e-01 -4.92598116e-01 2.19089761e-02
-1.58347678e+00 1.46157503e+00 3.14880610e-01 2.53554761e-01
-3.71204346e-01 -5.67639880e-02 -4.68041509e-01 -4.55109149e-01
-1.42748952e-01 -6.18326962e-01 1.18542814e+00 -1.16843116e+00
-1.67989540e+00 1.47354043e+00 -1.26258329e-01 -4.83167201e-01
1.00511479e+00 -2.49060839e-01 -5.29229045e-01 -1.96158499e-01
2.58134693e-01 4.87728000e-01 1.73825189e-01 -8.32735598e-01
-4.13549423e-01 -3.23895305e-01 -2.58192897e-01 5.20769894e-01
9.42610111e-03 8.42541397e-01 -4.06315267e-01 -1.04610634e+00
1.40724525e-01 -1.12651742e+00 -2.34581932e-01 -8.78433347e-01
-5.52943945e-01 -9.44167152e-02 1.75886527e-02 -1.25399578e+00
8.73728275e-01 -1.83083355e+00 4.57239270e-01 1.18938141e-01
-2.80938327e-01 6.60221577e-02 -1.73585162e-01 8.49354088e-01
-2.14069132e-02 2.23558247e-01 -3.74294221e-01 -4.25104946e-01
-3.38090882e-02 2.37922490e-01 -2.26429939e-01 4.34548944e-01
8.72866139e-02 1.11310637e+00 -9.01017487e-01 -3.70973706e-01
-2.61177778e-01 1.43004293e-02 -2.44885869e-03 -7.65431896e-02
-4.66425419e-02 6.14764214e-01 -1.27876267e-01 5.02915621e-01
1.00859061e-01 3.18688780e-01 2.69241840e-01 2.06266105e-01
-3.30612987e-01 1.02182829e+00 -1.95186868e-01 1.81250942e+00
-3.31062675e-01 8.51717830e-01 -1.27953216e-01 -7.54479468e-01
9.05443430e-01 6.56919301e-01 1.70033678e-01 -8.53448212e-01
1.70153126e-01 9.67824817e-01 4.08434927e-01 -1.00293741e-01
6.19817436e-01 -9.13441330e-02 -2.58830905e-01 6.24009967e-01
2.18672574e-01 -3.53955686e-01 2.10771322e-01 -2.52716839e-02
6.74223006e-01 4.54048991e-01 1.09665386e-01 -8.01969290e-01
5.98115847e-02 3.37875903e-01 5.15838146e-01 7.55214095e-01
-7.71675333e-02 8.45395267e-01 1.27505690e-01 -5.46028972e-01
-1.43480623e+00 -9.37172592e-01 -9.87348631e-02 1.14440775e+00
-4.61170018e-01 -2.18602553e-01 -7.62204051e-01 -6.28109932e-01
-6.78712845e-01 8.98102701e-01 -4.48061645e-01 2.01197118e-01
-9.39980507e-01 -9.61566448e-01 1.07700944e+00 1.63562030e-01
3.72235447e-01 -1.13815033e+00 5.93387596e-02 2.85691231e-01
-7.86379695e-01 -1.17044437e+00 -5.50779462e-01 4.16091323e-01
-8.02612782e-01 -2.89907336e-01 -1.04665995e+00 -1.28341544e+00
2.92049348e-01 -3.00294757e-01 1.37931395e+00 -2.26679310e-01
3.53081018e-01 -2.55135119e-01 -4.41214770e-01 -4.17721450e-01
-6.84675455e-01 5.62364340e-01 3.51951391e-01 -4.28024977e-01
5.33126533e-01 -5.70074737e-01 -4.30868715e-02 2.90627658e-01
-3.13073277e-01 3.55762661e-01 7.20968604e-01 1.00245261e+00
5.02498806e-01 -6.26967609e-01 4.44077522e-01 -9.97031033e-01
8.53613198e-01 -9.37516391e-02 -3.08785588e-01 3.09585541e-01
-5.22112370e-01 -2.45267019e-01 5.55969477e-01 -1.71970487e-01
-8.46082926e-01 -1.54856429e-01 -2.79764235e-01 2.81033188e-01
-1.78663507e-01 6.91049695e-01 -2.07276225e-01 4.24619406e-01
8.33571613e-01 4.43413705e-01 -2.84025133e-01 -6.28558576e-01
3.92593741e-01 9.25721288e-01 7.56591678e-01 -6.77234173e-01
7.93679714e-01 -2.08936974e-01 -6.52301908e-01 -5.81385732e-01
-2.75934309e-01 -2.66786009e-01 -9.89508450e-01 1.24519311e-01
7.88378596e-01 -9.64819312e-01 1.30529940e-01 4.90815789e-01
-1.58916247e+00 -4.42977905e-01 -1.38539836e-01 1.08266163e+00
-6.66119874e-01 6.27840608e-02 -1.27199113e+00 -3.81624192e-01
-6.69051230e-01 -1.15762448e+00 8.58634293e-01 -3.36600333e-01
-6.08180702e-01 -9.73555624e-01 3.07733685e-01 4.34605449e-01
4.64849263e-01 -8.36112872e-02 1.05821431e+00 -1.07829535e+00
-6.95004761e-02 -1.53965071e-01 -1.16811976e-01 2.57793993e-01
4.08247374e-02 -8.37501436e-02 -6.62968516e-01 -2.07064763e-01
-2.55310386e-01 -4.50888813e-01 7.97148049e-01 1.66822318e-02
1.14787994e-02 -5.30540884e-01 1.11736596e-01 1.39283106e-01
1.08507073e+00 7.37515017e-02 6.40935361e-01 5.15297592e-01
4.16285157e-01 6.64775729e-01 4.26611125e-01 -1.15229167e-01
3.18132848e-01 8.93841326e-01 -4.11449790e-01 -4.04667795e-01
-6.46157935e-02 -1.41765565e-01 6.82850718e-01 1.80089056e+00
-6.24904692e-01 -2.80578375e-01 -1.04281640e+00 8.13975155e-01
-1.96572936e+00 -7.39788175e-01 -3.11245769e-01 2.08302951e+00
1.20849240e+00 2.44266093e-01 1.75728783e-01 -4.03953165e-01
6.82486117e-01 -9.55848545e-02 2.90335417e-01 -1.02430892e+00
-4.81047511e-01 4.29260343e-01 4.64615077e-01 8.14205706e-01
-8.06356430e-01 1.60278928e+00 6.86841631e+00 9.37681913e-01
-8.57138276e-01 3.44356149e-01 6.44986212e-01 -3.20568420e-02
-3.44129831e-01 2.14884266e-01 -4.43910748e-01 3.95028204e-01
1.15549910e+00 -1.34334728e-01 6.88226044e-01 2.78966516e-01
3.24887246e-01 1.24037199e-01 -1.13483787e+00 5.95447659e-01
1.20081708e-01 -1.25571764e+00 2.54706144e-02 5.93549795e-02
9.70405817e-01 7.92061567e-01 -9.19919387e-02 3.56718034e-01
6.92348838e-01 -9.11816597e-01 6.69612467e-01 3.32182169e-01
9.62624848e-01 -6.99810565e-01 1.03922391e+00 4.99913067e-01
-4.71266419e-01 8.14260781e-01 -3.83167833e-01 -6.18067048e-02
1.35175303e-01 5.27410388e-01 -8.72175574e-01 8.36396337e-01
3.90673339e-01 4.85660583e-01 -3.64722103e-01 5.84051073e-01
-4.14752990e-01 9.38544989e-01 -3.13891560e-01 -2.77207255e-01
4.69737500e-01 -5.17467856e-01 9.28958356e-01 1.69052327e+00
1.85967550e-01 -8.29954818e-02 9.19166654e-02 2.98089832e-01
-3.44622195e-01 6.42913342e-01 -6.54450536e-01 -2.23491222e-01
1.09511442e-01 8.48307550e-01 -7.33563781e-01 -5.05623221e-01
-1.22999132e-01 1.42472124e+00 5.09889662e-01 5.23300588e-01
-3.92552167e-01 -4.18832541e-01 3.55314732e-01 -2.33449966e-01
-2.19467416e-01 -4.28421021e-01 -4.85238343e-01 -1.29461646e+00
1.02245674e-01 -1.08729720e+00 2.62090806e-02 -7.67720640e-01
-1.28663099e+00 1.30274916e+00 -2.17667267e-01 -1.05368841e+00
-5.31560779e-01 -4.40904379e-01 -2.80560851e-01 1.27390683e+00
-1.04991853e+00 -1.32316422e+00 7.41408229e-01 3.30911547e-01
7.03662813e-01 -6.88770533e-01 1.33706641e+00 4.01502013e-01
-2.80302435e-01 7.78876722e-01 5.61540544e-01 5.20297050e-01
1.14040673e+00 -1.13208711e+00 9.33266342e-01 9.64995742e-01
5.45808136e-01 5.58064520e-01 8.37300956e-01 -5.24304152e-01
-1.01336467e+00 -9.04010713e-01 2.13982868e+00 -8.11481774e-01
6.66667998e-01 -5.08718550e-01 -3.13766986e-01 7.09377646e-01
8.89969349e-01 -6.38457537e-01 7.33140588e-01 3.91441554e-01
-1.43548056e-01 2.40807712e-01 -8.18332911e-01 7.49768734e-01
1.06037736e+00 -8.65733504e-01 -7.97197282e-01 8.77045453e-01
7.06618845e-01 -3.68842691e-01 -8.38446617e-01 4.51162040e-01
5.26978850e-01 -4.36431378e-01 4.75444764e-01 -1.19315720e+00
8.81982803e-01 4.70655859e-02 -3.90930951e-01 -1.66072166e+00
-4.14027244e-01 -8.71807873e-01 6.60747588e-01 1.17686379e+00
1.29300177e+00 -4.76422995e-01 3.74782801e-01 6.38557449e-02
-2.41807908e-01 -5.34329891e-01 -1.16307819e+00 -8.24018419e-01
5.21921933e-01 -2.63889551e-01 2.36197099e-01 1.44309318e+00
2.79974073e-01 7.22765803e-01 -5.89259803e-01 -3.49541694e-01
1.96122050e-01 1.01414584e-02 5.59494138e-01 -8.49412024e-01
-7.13737369e-01 -5.64306855e-01 -4.72793169e-02 -7.72053957e-01
6.31174669e-02 -1.30163777e+00 3.02782327e-01 -1.66137028e+00
2.27133110e-01 -2.01277077e-01 -2.41123274e-01 5.84645748e-01
-1.01655096e-01 7.95539558e-01 4.69642319e-02 6.32151127e-01
-3.27993274e-01 1.87158689e-01 1.07863867e+00 -1.93147883e-01
-1.21404439e-01 6.00417592e-02 -4.95106846e-01 4.01160419e-01
8.86116266e-01 -5.34432769e-01 2.10973233e-01 -1.14554822e+00
4.13842589e-01 2.75230855e-01 -2.99059600e-01 -3.15309405e-01
3.12641561e-02 -1.52988896e-01 2.67468274e-01 -3.77591580e-01
1.93657219e-01 -2.95205653e-01 4.08097217e-03 3.41745198e-01
-5.69514751e-01 5.80301762e-01 1.33520260e-01 -2.05623791e-01
-3.42162311e-01 1.56402737e-01 6.14416122e-01 -4.59431112e-01
-9.84134898e-02 6.00551479e-02 -5.96310198e-01 1.41919285e-01
3.72580409e-01 4.70517464e-02 -1.70688868e-01 -4.44098234e-01
-7.91366100e-01 -1.37829380e-02 4.24468905e-01 4.31806117e-01
1.40249252e-01 -1.33354509e+00 -1.64711332e+00 -1.33912653e-01
2.65628010e-01 -5.75494885e-01 -5.47259629e-01 1.11669648e+00
-6.67422950e-01 5.00632226e-01 -2.80770987e-01 -2.44737729e-01
-8.98315012e-01 3.78348753e-02 1.00061260e-01 -6.06357813e-01
-2.95629263e-01 9.11011815e-01 -5.35921931e-01 -1.14203513e+00
-1.79054067e-02 2.50114232e-01 1.82900012e-01 -5.36460951e-02
6.35543838e-02 1.80735812e-01 4.95318800e-01 -9.28171754e-01
-3.53141814e-01 -8.77856463e-03 -2.69461751e-01 -9.48461831e-01
1.16534460e+00 -1.00988120e-01 -5.04540026e-01 5.56288958e-01
1.13001478e+00 3.52660507e-01 -1.38336912e-01 -4.56200093e-01
2.38207221e-01 -7.97762349e-02 -2.55846024e-01 -1.49119270e+00
-4.18945938e-01 6.95687413e-01 1.91035748e-01 -3.89609098e-01
7.00402141e-01 -1.54543892e-01 8.10205758e-01 7.28850424e-01
5.08913457e-01 -1.49152708e+00 -7.61452079e-01 1.11442089e+00
8.64765584e-01 -1.16436565e+00 -2.56239295e-01 -7.39185736e-02
-5.07010281e-01 9.85949934e-01 1.10983953e-01 8.15173239e-02
1.57548189e-01 4.01207209e-02 4.88122404e-01 2.58491218e-01
-6.61515892e-01 -1.42311856e-01 3.18848491e-01 4.22851503e-01
8.56118381e-01 3.66039395e-01 -1.08426273e+00 4.71861720e-01
-6.67387545e-01 -1.89281181e-01 2.74868071e-01 7.09695637e-01
-3.30713093e-02 -1.38725424e+00 -5.59434406e-02 2.85722941e-01
-7.81786859e-01 -7.62182295e-01 -1.27421653e+00 7.21430421e-01
-1.89707890e-01 9.41566706e-01 -2.10321680e-01 -4.65050191e-01
2.34199271e-01 2.78383642e-01 5.35825968e-01 -6.01567507e-01
-1.02298212e+00 4.18411463e-01 8.05546641e-01 2.34130230e-02
-3.66395414e-01 -5.92330754e-01 -7.55514801e-01 -4.45252389e-01
-2.87530631e-01 6.11340404e-01 6.43301845e-01 1.18354475e+00
-1.68189686e-02 6.46461993e-02 5.68127453e-01 -6.17126465e-01
-3.70963961e-01 -1.43066311e+00 -1.43999591e-01 2.23937452e-01
-1.74414262e-01 2.71309793e-01 -1.25237390e-01 2.97764018e-02] | [11.556175231933594, 10.392269134521484] |
7b783119-370d-446b-83bf-11a2587895aa | se-ornet-self-ensembling-orientation-aware | 2304.05395 | null | https://arxiv.org/abs/2304.05395v1 | https://arxiv.org/pdf/2304.05395v1.pdf | SE-ORNet: Self-Ensembling Orientation-aware Network for Unsupervised Point Cloud Shape Correspondence | Unsupervised point cloud shape correspondence aims to obtain dense point-to-point correspondences between point clouds without manually annotated pairs. However, humans and some animals have bilateral symmetry and various orientations, which lead to severe mispredictions of symmetrical parts. Besides, point cloud noise disrupts consistent representations for point cloud and thus degrades the shape correspondence accuracy. To address the above issues, we propose a Self-Ensembling ORientation-aware Network termed SE-ORNet. The key of our approach is to exploit an orientation estimation module with a domain adaptive discriminator to align the orientations of point cloud pairs, which significantly alleviates the mispredictions of symmetrical parts. Additionally, we design a selfensembling framework for unsupervised point cloud shape correspondence. In this framework, the disturbances of point cloud noise are overcome by perturbing the inputs of the student and teacher networks with different data augmentations and constraining the consistency of predictions. Extensive experiments on both human and animal datasets show that our SE-ORNet can surpass state-of-the-art unsupervised point cloud shape correspondence methods. | ['Zhe Zhang', 'Jiyang Yu', 'Tianzhu Zhang', 'Jianfeng He', 'Jiahao Lu', 'Chuxin Wang', 'Jiacheng Deng'] | 2023-04-10 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Deng_SE-ORNet_Self-Ensembling_Orientation-Aware_Network_for_Unsupervised_Point_Cloud_Shape_Correspondence_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Deng_SE-ORNet_Self-Ensembling_Orientation-Aware_Network_for_Unsupervised_Point_Cloud_Shape_Correspondence_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-dense-shape-correspondence'] | ['computer-vision'] | [-2.40132123e-01 1.46163687e-01 4.63059247e-02 -6.15891635e-01
-1.75091714e-01 -5.33070624e-01 4.76193666e-01 2.54096538e-01
-5.64059056e-02 2.35081285e-01 -3.23065728e-01 5.26093468e-02
-8.90959799e-02 -7.92735577e-01 -9.66289103e-01 -4.92280453e-01
2.72616625e-01 9.64696527e-01 3.67605180e-01 -2.34867662e-01
3.40261251e-01 8.37178290e-01 -1.54858875e+00 -5.52206822e-02
1.16285002e+00 9.57381070e-01 8.50213468e-02 1.97677445e-02
-4.35852259e-01 3.93877476e-02 -2.87096262e-01 -2.72899121e-01
5.70876062e-01 8.66286233e-02 -3.85169744e-01 2.41665572e-01
8.57202291e-01 -2.86204189e-01 -5.06373495e-02 1.08400774e+00
2.11977750e-01 9.15153846e-02 8.07939291e-01 -1.66285419e+00
-6.33339047e-01 2.32368857e-01 -6.91327751e-01 -2.72409499e-01
-6.52583539e-02 1.16743192e-01 8.86897743e-01 -9.92230296e-01
3.84915113e-01 1.21866894e+00 1.00144112e+00 3.02923709e-01
-1.37687659e+00 -9.48566556e-01 2.63411224e-01 -1.17017373e-01
-1.53355563e+00 -2.83390790e-01 1.20473540e+00 -5.94250739e-01
6.80754542e-01 4.10803407e-02 7.49465942e-01 7.65565932e-01
-3.19585294e-01 4.93487537e-01 5.84662616e-01 -2.12615374e-02
2.66630650e-01 5.02419025e-02 -1.56234384e-01 4.16045994e-01
2.85151839e-01 1.94699153e-01 -2.59460837e-01 -2.44107500e-01
1.36103427e+00 4.45766896e-01 1.52980201e-02 -1.19586563e+00
-1.24078572e+00 4.76875722e-01 8.98029625e-01 9.44385678e-02
-3.45133185e-01 -1.09095732e-02 7.30665103e-02 5.97248487e-02
5.46318412e-01 6.11171186e-01 -4.05826062e-01 2.12643906e-01
-8.01318705e-01 3.91794354e-01 4.59346771e-01 1.39907920e+00
1.13537145e+00 -8.71047452e-02 2.04938605e-01 7.47766912e-01
4.32982981e-01 3.98183316e-01 1.69766292e-01 -7.96533823e-01
5.49155116e-01 1.02163291e+00 -3.99121083e-02 -1.44221163e+00
-3.16008329e-01 -5.47473848e-01 -1.08886385e+00 2.72392631e-01
2.29178041e-01 1.94270507e-01 -8.36551964e-01 1.45395982e+00
5.35713255e-01 5.39369941e-01 -2.67092258e-01 1.07885945e+00
7.17010438e-01 3.22289497e-01 2.19388884e-02 4.37616587e-01
9.66562927e-01 -7.47330844e-01 -1.44273326e-01 -3.42556417e-01
2.86232919e-01 -5.72528481e-01 9.81693327e-01 -4.79240017e-03
-8.67540896e-01 -6.00069404e-01 -9.96827424e-01 -3.77855375e-02
-2.33542815e-01 6.90900311e-02 5.27952075e-01 6.28014952e-02
-5.95089853e-01 9.05040860e-01 -9.81253028e-01 -1.19638078e-01
6.03529453e-01 5.52554846e-01 -4.70695436e-01 1.93917498e-01
-5.04546404e-01 6.25418901e-01 3.73108327e-01 1.99194223e-01
-2.06771359e-01 -1.06983471e+00 -7.92234898e-01 -1.97368637e-02
1.53206706e-01 -8.48273516e-01 9.61519599e-01 -8.54527950e-01
-1.37490845e+00 7.14652717e-01 8.57215971e-02 -1.97661906e-01
6.88293099e-01 -1.07466824e-01 7.41432933e-03 -3.66165861e-02
2.79763937e-01 1.13028228e+00 8.25185895e-01 -1.42472398e+00
-4.08408672e-01 -5.68531811e-01 -3.46873164e-01 3.50562215e-01
-2.42763981e-01 -4.41775680e-01 -4.00081098e-01 -5.66700339e-01
8.36936533e-01 -9.16530490e-01 -3.03730905e-01 4.70621884e-01
-3.15446496e-01 -3.78919214e-01 9.32615221e-01 -1.67178854e-01
4.53600675e-01 -2.27015781e+00 2.09674928e-02 4.95549530e-01
4.37777668e-01 -1.67247012e-01 -2.93722957e-01 1.85010493e-01
-3.45588058e-01 1.41667286e-02 -3.44419986e-01 -4.37793314e-01
-6.91835731e-02 5.50744951e-01 -3.91179681e-01 5.33356130e-01
5.87421060e-01 6.93659067e-01 -9.40061927e-01 -3.83869946e-01
3.41592789e-01 4.34596628e-01 -8.17426741e-01 3.28798711e-01
-3.34579825e-01 7.36465693e-01 -5.48666239e-01 5.32562733e-01
1.17223060e+00 -3.62010688e-01 -7.42370039e-02 -3.61459941e-01
-1.58537835e-01 3.59080821e-01 -1.14354336e+00 1.71199429e+00
-2.41199598e-01 2.21368045e-01 -2.04948887e-01 -9.50058758e-01
1.31454599e+00 1.39438331e-01 6.69701636e-01 -3.37712675e-01
-7.36998096e-02 2.39969000e-01 -4.81214523e-02 -4.05190401e-02
3.06317061e-01 -9.74776372e-02 3.18655372e-01 1.10990278e-01
4.25112508e-02 -5.58111846e-01 -3.52077663e-01 -8.25086311e-02
5.72247207e-01 3.01246673e-01 7.27864131e-02 -1.77046806e-01
3.14808995e-01 3.81560065e-02 7.59132385e-01 3.41590315e-01
2.97770333e-02 9.80222821e-01 3.54543686e-01 -5.67240000e-01
-1.21702373e+00 -1.23004532e+00 -2.09572479e-01 7.28610277e-01
5.25796115e-01 -2.68553972e-01 -4.72456872e-01 -6.23275697e-01
3.94762665e-01 4.11904722e-01 -2.71812439e-01 -2.35478487e-02
-5.75220525e-01 -2.28525266e-01 2.40957826e-01 9.43599403e-01
4.81409431e-01 -8.04395020e-01 -2.48055935e-01 -4.90700789e-02
-6.68054074e-02 -1.21621919e+00 -3.62796068e-01 1.26112565e-01
-1.25925922e+00 -1.01017845e+00 -4.46442485e-01 -9.25714016e-01
1.31256187e+00 4.15000647e-01 1.11796391e+00 1.39459714e-01
1.82989597e-01 3.84253054e-03 -1.21159993e-01 -4.92743611e-01
-2.28026927e-01 1.37997210e-01 2.95167983e-01 -2.85875380e-01
6.09404981e-01 -1.23668981e+00 -6.49709046e-01 7.00625837e-01
-6.79583848e-01 1.60917073e-01 5.26200950e-01 7.00001895e-01
9.82050598e-01 -1.43792287e-01 2.32077807e-01 -6.82424068e-01
2.37692714e-01 -3.35470766e-01 -9.15733755e-01 -4.19383682e-02
-5.39098680e-01 1.15728669e-01 7.35528588e-01 -5.59327185e-01
-4.66697127e-01 4.96690512e-01 -2.82881521e-02 -1.07348645e+00
-4.26000834e-01 1.67368688e-02 -1.89459488e-01 -3.95626694e-01
6.23995721e-01 1.06840220e-04 2.07883492e-01 -6.19180739e-01
3.64453852e-01 4.58588928e-01 8.90948713e-01 -7.35836148e-01
1.36207020e+00 4.62335825e-01 3.97826843e-02 -6.97386980e-01
-4.88462985e-01 -6.73518002e-01 -1.06753159e+00 7.73152933e-02
5.47842979e-01 -1.04792571e+00 -7.94083536e-01 3.90682936e-01
-1.29716790e+00 8.70009735e-02 -3.57667953e-01 2.89070964e-01
-5.74184418e-01 4.79970247e-01 -1.91046014e-01 -3.90747100e-01
-2.71153271e-01 -1.12460804e+00 1.20339096e+00 1.99072018e-01
-1.68552697e-01 -6.64697886e-01 -7.36583397e-02 1.69245359e-02
7.17904493e-02 1.83333963e-01 8.38878989e-01 -8.50165069e-01
-7.37368524e-01 -1.83150887e-01 -3.92193109e-01 2.87190557e-01
2.96141386e-01 4.38429639e-02 -8.01480114e-01 -1.05282515e-01
-6.52470738e-02 -3.46105434e-02 3.36314231e-01 6.02417067e-02
1.46737421e+00 -2.33391732e-01 -4.25933480e-01 9.15493488e-01
1.08655250e+00 -2.13168249e-01 2.96257287e-01 3.07356149e-01
1.10614145e+00 5.79328299e-01 4.84560072e-01 3.29461217e-01
5.06989062e-01 6.24726057e-01 7.65397012e-01 -1.61914125e-01
2.91916341e-01 -7.60121107e-01 -3.45780879e-01 7.95521200e-01
-6.15619719e-02 2.58902222e-01 -1.06065094e+00 5.94511271e-01
-1.82775700e+00 -5.45558631e-01 -9.27324407e-03 2.18837571e+00
6.60541952e-01 -2.20183451e-02 5.21873422e-02 -7.43631199e-02
7.47836769e-01 -6.34719059e-02 -6.83966756e-01 1.87485442e-01
2.73012996e-01 1.04301311e-02 5.69497705e-01 -7.01141078e-03
-1.07707548e+00 1.01825249e+00 5.65124798e+00 4.05800134e-01
-1.06051493e+00 -2.94210255e-01 2.52659410e-01 1.82582885e-01
-3.46268415e-01 1.62300587e-01 -6.12134516e-01 4.38258499e-01
1.60641804e-01 4.41105887e-02 3.45683426e-01 1.17395532e+00
-5.34245893e-02 5.44727623e-01 -1.45023394e+00 1.09929001e+00
-1.70099393e-01 -1.31220186e+00 2.22400595e-02 -2.57187872e-03
7.69863009e-01 2.87254661e-01 -5.87837175e-02 7.88735300e-02
2.47112677e-01 -7.62845874e-01 6.97884023e-01 4.29148883e-01
6.31095052e-01 -7.72798955e-01 5.55565953e-01 6.76875770e-01
-1.13589382e+00 2.86434740e-01 -7.86040425e-01 5.38063943e-02
-1.58193275e-01 4.84025389e-01 -1.03490078e+00 5.28072357e-01
7.77936459e-01 9.50570643e-01 -5.00991166e-01 1.19705284e+00
-2.63975352e-01 7.90637955e-02 -8.83556068e-01 2.43063584e-01
5.58271147e-02 -5.30360818e-01 6.50281131e-01 6.12373710e-01
2.55002528e-01 6.94135278e-02 3.20197612e-01 1.42522860e+00
-7.48335198e-02 -1.30721182e-01 -7.16136873e-01 7.76966587e-02
9.02138650e-01 1.20896423e+00 -5.92317641e-01 -1.02926239e-01
-1.51253507e-01 7.50499427e-01 5.09082019e-01 2.01572508e-01
-5.35115480e-01 -1.84500664e-01 9.09398556e-01 4.33513224e-01
2.32309848e-01 -4.08257663e-01 -6.20551229e-01 -1.13300920e+00
1.84996635e-01 -6.53564155e-01 -1.44348547e-01 -7.72776842e-01
-1.69541216e+00 4.25783008e-01 9.44363475e-02 -1.81903613e+00
-1.47781044e-01 -3.56158853e-01 -9.06528831e-01 7.88623393e-01
-1.36804473e+00 -1.11182642e+00 -6.09536409e-01 3.48202825e-01
2.41263002e-01 -2.77821980e-02 4.82671708e-01 1.36512712e-01
-2.58270711e-01 6.30266190e-01 -7.82792866e-02 2.69722193e-01
7.03714430e-01 -1.25371802e+00 9.19981837e-01 5.16406476e-01
1.93790928e-01 8.80572557e-01 5.99905729e-01 -7.22965360e-01
-1.14590347e+00 -1.36366963e+00 5.87881148e-01 -4.64510471e-01
5.37005544e-01 -4.99912053e-01 -1.23917758e+00 6.65924072e-01
-3.35673571e-01 3.32356900e-01 3.49799603e-01 1.43347681e-01
-4.64521229e-01 -2.12514132e-01 -1.06677330e+00 6.99948430e-01
1.23677063e+00 -4.63638693e-01 -6.94034636e-01 4.61673290e-01
6.48877740e-01 -6.86293304e-01 -9.28801835e-01 6.57165051e-01
2.94523418e-01 -8.50599349e-01 1.22471166e+00 -5.19208610e-01
5.43327034e-01 -4.82464582e-01 1.09171547e-01 -1.48863542e+00
-3.32306325e-01 -3.06232691e-01 1.32696480e-01 1.22341073e+00
3.13996002e-02 -7.76763260e-01 1.09315979e+00 5.98530948e-01
-1.74282193e-01 -7.03071177e-01 -8.01494360e-01 -8.48181844e-01
2.60174811e-01 -2.65479833e-01 1.07062161e+00 1.17131722e+00
-3.39367419e-01 2.34920859e-01 4.43539619e-02 6.86194241e-01
8.21784854e-01 2.63617486e-01 1.12330532e+00 -1.62653375e+00
-3.43342870e-02 -4.72062230e-01 -7.74297476e-01 -1.40793133e+00
3.09077203e-01 -8.22071135e-01 2.30801985e-01 -1.17880094e+00
-2.67446429e-01 -8.39121938e-01 1.37524247e-01 6.89726353e-01
2.45016906e-02 1.33424953e-01 1.48720324e-01 5.14558315e-01
-1.98013023e-01 8.72892618e-01 1.09628701e+00 -7.06773177e-02
-2.89954424e-01 1.79651380e-01 -4.67485189e-01 1.08718216e+00
6.91318631e-01 -5.58055043e-01 -3.71590644e-01 -7.37364233e-01
7.52386153e-02 -2.21607432e-01 7.72767186e-01 -1.04958236e+00
5.34740567e-01 -2.30631620e-01 6.25263453e-01 -1.09408557e+00
2.48328134e-01 -1.30702543e+00 5.55753522e-02 -2.34507248e-02
-1.56066298e-01 2.24458814e-01 1.51130676e-01 5.89460194e-01
-1.54139146e-01 -1.09073962e-03 7.36415327e-01 -3.60959359e-02
-3.20319355e-01 6.08283281e-01 2.54627287e-01 -2.16596574e-01
7.66603529e-01 -5.09175599e-01 -1.60372943e-01 -6.19746782e-02
-3.56635332e-01 5.52780986e-01 8.47371221e-01 5.65004349e-01
7.42335260e-01 -1.40687311e+00 -3.22141439e-01 7.35081196e-01
3.54660451e-01 1.04623234e+00 -3.92844267e-05 6.29579604e-01
-5.97373307e-01 3.10410722e-03 -2.58462101e-01 -1.17046213e+00
-9.95490193e-01 2.29939148e-01 3.80476266e-01 2.41450205e-01
-8.43114614e-01 7.14272976e-01 3.18132907e-01 -1.14674723e+00
1.40627354e-01 -5.60970783e-01 -5.58597073e-02 -4.33480293e-01
-8.99930149e-02 1.21677786e-01 8.14431980e-02 -6.55813217e-01
-4.31664497e-01 8.88262689e-01 -1.25831425e-01 4.03973520e-01
1.40939236e+00 1.19813405e-01 -1.80565611e-01 3.00246805e-01
1.10516882e+00 -1.04939252e-01 -1.41360188e+00 -2.52590030e-01
-2.02508718e-01 -5.79813540e-01 -5.83961643e-02 -3.74885172e-01
-1.06657457e+00 8.56633902e-01 4.42266315e-01 -6.24182858e-02
8.27597737e-01 -1.04394592e-01 6.68236971e-01 5.09351969e-01
4.07430887e-01 -7.71422267e-01 -1.79846659e-01 4.48264688e-01
8.32493186e-01 -1.35405171e+00 7.87543878e-02 -7.40293503e-01
-4.60622281e-01 1.12097764e+00 9.98346329e-01 -4.85903561e-01
5.70694625e-01 1.47061989e-01 -1.58994216e-02 -2.69914538e-01
-3.76398295e-01 1.17250778e-01 5.64392805e-01 6.99281752e-01
1.30236492e-01 -8.06661323e-02 1.38574287e-01 3.01412821e-01
-5.85396886e-01 -2.08287060e-01 -1.91407111e-02 7.40978718e-01
-2.58823514e-01 -1.00508809e+00 -4.44079220e-01 1.56526074e-01
1.52067184e-01 7.15581849e-02 -5.83886147e-01 7.65866756e-01
1.93273630e-02 4.49933678e-01 6.74117327e-01 -4.70342308e-01
4.79055554e-01 -2.44429708e-01 2.25162774e-01 -5.59340179e-01
-5.33663094e-01 8.34730044e-02 -6.37969732e-01 -4.60908532e-01
-2.75498122e-01 -2.89579421e-01 -1.43605220e+00 -2.61069149e-01
-3.05079281e-01 2.18670920e-01 6.41485274e-01 8.65204513e-01
6.88941479e-01 2.63175368e-01 7.26643026e-01 -1.14601719e+00
-8.25327456e-01 -8.03950846e-01 -3.42495322e-01 6.18675232e-01
2.48581469e-01 -7.92350650e-01 -4.12012517e-01 -1.84638515e-01] | [8.046256065368652, -3.363560199737549] |
f82eb974-dd46-48ec-ab8d-0fbd216d2f6d | transformer-based-visual-segmentation-a | 2304.09854 | null | https://arxiv.org/abs/2304.09854v2 | https://arxiv.org/pdf/2304.09854v2.pdf | Transformer-Based Visual Segmentation: A Survey | Visual segmentation seeks to partition images, video frames, or point clouds into multiple segments or groups. This technique has numerous real-world applications, such as autonomous driving, image editing, robot sensing, and medical analysis. Over the past decade, deep learning-based methods have made remarkable strides in this area. Recently, transformers, a type of neural network based on self-attention originally designed for natural language processing, have considerably surpassed previous convolutional or recurrent approaches in various vision processing tasks. Specifically, vision transformers offer robust, unified, and even simpler solutions for various segmentation tasks. This survey provides a thorough overview of transformer-based visual segmentation, summarizing recent advancements. We first review the background, encompassing problem definitions, datasets, and prior convolutional methods. Next, we summarize a meta-architecture that unifies all recent transformer-based approaches. Based on this meta-architecture, we examine various method designs, including modifications to the meta-architecture and associated applications. We also present several closely related settings, including 3D point cloud segmentation, foundation model tuning, domain-aware segmentation, efficient segmentation, and medical segmentation. Additionally, we compile and re-evaluate the reviewed methods on several well-established datasets. Finally, we identify open challenges in this field and propose directions for future research. The project page can be found at https://github.com/lxtGH/Awesome-Segmentation-With-Transformer. We will also continually monitor developments in this rapidly evolving field. | ['Chen Change Loy', 'Ziwei Liu', 'Kai Chen', 'Guangliang Cheng', 'Jiangmiao Pang', 'Haobo Yuan', 'Wenwei Zhang', 'Henghui Ding', 'Xiangtai Li'] | 2023-04-19 | null | null | null | null | ['point-cloud-segmentation'] | ['computer-vision'] | [ 3.76853466e-01 -4.47952263e-02 -2.92477757e-01 -3.97762299e-01
-7.32058227e-01 -5.42143881e-01 2.14142099e-01 -2.75854170e-02
-2.70459890e-01 1.20517910e-01 -2.45697215e-01 -3.95668596e-01
1.76407397e-01 -5.73965132e-01 -6.56634629e-01 -6.03401721e-01
1.16886690e-01 5.04667163e-01 4.28482920e-01 2.12582462e-02
2.57500559e-01 6.59349203e-01 -1.43354654e+00 1.83199078e-01
7.77570188e-01 1.30080354e+00 3.50264579e-01 6.01904988e-01
-5.06692767e-01 4.65765238e-01 -5.94549775e-01 -3.37311804e-01
2.38477767e-01 -1.44073337e-01 -1.06956422e+00 4.60787028e-01
5.03367305e-01 -1.57976195e-01 -3.32545400e-01 9.66016829e-01
4.04958636e-01 -7.32932985e-02 4.71041024e-01 -1.27125061e+00
-7.04548597e-01 3.83062571e-01 -6.12415552e-01 3.03456306e-01
-4.23473716e-02 2.05675572e-01 6.79493189e-01 -7.92772412e-01
4.44931269e-01 1.03965318e+00 9.34142292e-01 5.90987921e-01
-8.35334539e-01 -4.60525662e-01 4.86359954e-01 2.34052971e-01
-1.15464985e+00 -1.62055239e-01 8.10301006e-01 -4.83225733e-01
1.10628092e+00 1.69579685e-01 1.07407749e+00 8.10029566e-01
1.45316124e-01 1.38548887e+00 6.69818699e-01 -2.42384568e-01
1.61343273e-02 -2.64432520e-01 3.15844774e-01 8.92869771e-01
-1.11293932e-02 -2.58058429e-01 -8.15867037e-02 1.46464109e-01
1.05125034e+00 1.12486355e-01 -1.70248792e-01 -6.06201649e-01
-1.41779888e+00 6.95385635e-01 6.23891354e-01 3.58554065e-01
-2.27041602e-01 2.93655634e-01 5.64865649e-01 -3.67961004e-02
5.47435224e-01 3.56622010e-01 -5.13873279e-01 8.48280638e-02
-1.07849288e+00 1.03554599e-01 4.01321858e-01 1.23256743e+00
7.68608630e-01 2.08754286e-01 -2.13750049e-01 9.23664629e-01
1.82424456e-01 3.62323135e-01 4.59330380e-01 -1.23189640e+00
1.29640341e-01 5.83889067e-01 -2.65294999e-01 -7.71604240e-01
-5.54904580e-01 -3.99972737e-01 -9.64889050e-01 -1.73204858e-02
8.93373564e-02 3.54983360e-02 -1.52802050e+00 1.24157703e+00
3.39963943e-01 3.57416779e-01 -2.20161214e-01 8.80748510e-01
1.26684749e+00 5.27121902e-01 3.88693698e-02 2.80361213e-02
1.47462142e+00 -1.31503820e+00 -5.43845475e-01 -3.63389671e-01
1.27228513e-01 -7.27868438e-01 9.36659336e-01 3.98098826e-01
-1.23696947e+00 -6.43718719e-01 -9.06244993e-01 -5.33827364e-01
-4.98460054e-01 2.95987148e-02 7.61319518e-01 3.99736553e-01
-1.31896198e+00 5.79948723e-01 -1.27272117e+00 -6.05603099e-01
9.93629932e-01 4.60436314e-01 1.16694763e-01 1.33037224e-01
-8.48440826e-01 6.12808526e-01 1.54601797e-01 3.54119807e-01
-8.75083625e-01 -7.04890490e-01 -8.83621573e-01 -2.77417123e-01
3.26879472e-01 -1.05454481e+00 1.77001548e+00 -9.77814317e-01
-1.46586859e+00 1.32152057e+00 -3.31941634e-01 -6.61898136e-01
4.08113122e-01 -2.51150459e-01 1.17757879e-02 3.23451966e-01
1.53527021e-01 1.13927460e+00 7.09650457e-01 -1.14588261e+00
-8.87572646e-01 -2.21338794e-01 8.19020346e-02 2.87936002e-01
-9.33517143e-03 -5.59571311e-02 -1.23535752e+00 -6.51523888e-01
2.05817237e-01 -7.94389248e-01 -6.91986084e-01 2.56251454e-01
-5.98187387e-01 -3.17586035e-01 8.57879639e-01 -2.10887820e-01
8.91983986e-01 -2.05543113e+00 1.19019553e-01 -1.21271893e-01
6.13218725e-01 3.28839004e-01 -6.18111761e-03 1.90547496e-01
-7.47813806e-02 1.40265852e-01 -6.71838880e-01 -7.04348922e-01
-1.53623715e-01 2.40410879e-01 -2.50950694e-01 5.47239184e-01
1.93240300e-01 1.39806628e+00 -7.10622191e-01 -7.33636975e-01
7.24856436e-01 4.46138173e-01 -1.73699155e-01 -2.81835273e-02
-3.40648204e-01 5.10610282e-01 -4.60073143e-01 1.01729131e+00
6.81220233e-01 -6.21306181e-01 -2.19186395e-01 -3.30777019e-01
-2.25415260e-01 6.51735738e-02 -6.25436544e-01 1.95358610e+00
-1.78680509e-01 8.59293580e-01 2.09713712e-01 -1.42178035e+00
7.76634216e-01 1.85389191e-01 9.67319191e-01 -5.88795960e-01
3.46601695e-01 1.80205196e-01 -4.91480738e-01 -4.99668509e-01
5.46602547e-01 4.89916690e-02 4.69806790e-02 4.61736396e-02
-3.37738805e-02 -4.72506672e-01 1.84186488e-01 6.47723861e-03
7.51070917e-01 5.97465709e-02 3.31167519e-01 5.75799607e-02
4.36556309e-01 4.54126477e-01 5.92684627e-01 6.44749224e-01
-7.01406777e-01 9.52340126e-01 2.44763434e-01 -5.53520799e-01
-6.87661707e-01 -1.13633478e+00 -2.72129238e-01 6.41479671e-01
6.29486263e-01 -1.84900135e-01 -8.33187938e-01 -5.09091020e-01
-7.76425228e-02 1.59339830e-01 -6.48105621e-01 1.66086927e-01
-6.10598445e-01 -6.66107893e-01 6.74643636e-01 8.05433750e-01
7.49092877e-01 -1.14762998e+00 -7.56424844e-01 1.09854564e-01
-4.11679596e-01 -1.29379559e+00 -3.00130576e-01 7.59256110e-02
-1.25354350e+00 -1.04447067e+00 -1.05908191e+00 -1.27167428e+00
4.04538006e-01 5.75060964e-01 1.39387143e+00 1.26509413e-01
-4.09859776e-01 6.94082499e-01 -2.30275288e-01 -4.95542020e-01
6.92470893e-02 2.81599283e-01 -3.50591332e-01 -3.15625131e-01
5.54769039e-01 -3.10341209e-01 -7.07471728e-01 1.66316792e-01
-8.41443896e-01 1.88448086e-01 4.64396328e-01 5.48720241e-01
1.13781583e+00 -2.52143621e-01 2.10353568e-01 -8.26927483e-01
3.94482046e-01 -2.68135339e-01 -5.96208513e-01 2.72546947e-01
-3.01323920e-01 -3.63988549e-01 3.35263073e-01 -2.00383097e-01
-6.55587614e-01 2.16023326e-01 -4.94996428e-01 -1.01097345e+00
-3.88344437e-01 4.06558484e-01 6.47277609e-02 -1.67424232e-01
3.60387355e-01 3.10988992e-01 8.79026353e-02 -4.26260740e-01
5.04304409e-01 5.89869142e-01 7.70884633e-01 -4.58774865e-01
5.18380642e-01 8.12872887e-01 -2.78894186e-01 -9.84215081e-01
-7.89978027e-01 -7.42367029e-01 -6.86315596e-01 -1.79995924e-01
1.18927014e+00 -7.86023140e-01 -5.58973014e-01 8.47961843e-01
-1.35804462e+00 -6.43927574e-01 -3.81618798e-01 1.16778739e-01
-8.22500169e-01 6.06221437e-01 -7.52036929e-01 -4.35809135e-01
-5.70308328e-01 -1.57971728e+00 1.33792019e+00 4.38901961e-01
7.37713277e-02 -1.19394767e+00 -9.69790295e-02 5.47279716e-01
2.98222750e-01 3.71972620e-01 7.04422772e-01 -3.87150198e-01
-8.10392499e-01 1.66094854e-01 -3.09630573e-01 3.10139298e-01
1.43573686e-01 2.37393692e-01 -8.97998273e-01 -1.59984186e-01
-1.19893186e-01 -1.66705295e-01 1.03261364e+00 9.85748172e-01
1.61886680e+00 1.93635181e-01 -8.04696918e-01 1.08930957e+00
1.24360454e+00 2.85497576e-01 4.68891919e-01 2.72255599e-01
8.79313111e-01 4.40927267e-01 5.10059476e-01 3.74367796e-02
6.24453604e-01 4.94124562e-01 5.39187610e-01 -5.72072804e-01
-2.06853852e-01 1.53354749e-01 -4.01618592e-02 8.94611895e-01
-4.54702834e-03 -3.65989447e-01 -1.15605724e+00 7.75202036e-01
-1.86752248e+00 -6.31404459e-01 -1.32557034e-01 1.72432923e+00
5.19392610e-01 8.47257003e-02 1.58944964e-01 -3.48984450e-02
8.44568193e-01 1.82780325e-01 -1.01115334e+00 -3.36525381e-01
-6.96001062e-03 3.53699625e-01 6.14267528e-01 2.95433521e-01
-1.63975775e+00 1.34458041e+00 6.75280333e+00 7.33591855e-01
-1.47906601e+00 -1.18439633e-03 6.97928011e-01 1.61698997e-01
-1.01910666e-01 -3.59285444e-01 -5.72982728e-01 1.54734120e-01
5.25629640e-01 -1.31728724e-02 2.42621094e-01 8.95213246e-01
1.27552778e-01 4.95607778e-02 -9.46050942e-01 1.26626718e+00
1.10324524e-01 -1.71235228e+00 1.44003918e-02 -1.04505539e-01
5.49231052e-01 7.26853967e-01 3.35093349e-01 1.19751677e-01
1.53430417e-01 -9.56905603e-01 7.21724093e-01 2.82282352e-01
9.35472429e-01 -5.18636346e-01 5.22881567e-01 7.84219876e-02
-1.51351523e+00 1.26906648e-01 -3.16341311e-01 1.87800527e-01
3.78155202e-01 6.38246119e-01 -5.36724627e-01 6.61655188e-01
1.12417269e+00 1.27385008e+00 -4.17746365e-01 1.25860584e+00
2.51501389e-02 4.40231621e-01 -3.33429217e-01 2.57427275e-01
5.37977993e-01 -2.60181665e-01 4.56776798e-01 1.38397098e+00
4.12075855e-02 7.20250085e-02 3.07087094e-01 8.56806338e-01
-2.91466825e-02 -1.63891137e-01 -5.25108755e-01 -2.78360676e-02
4.46795166e-01 1.24114525e+00 -1.27376044e+00 -5.82007527e-01
-6.07663512e-01 9.96782959e-01 1.03456087e-01 5.64817011e-01
-1.03013361e+00 -5.02862453e-01 9.47335541e-01 -1.71757609e-01
3.58094603e-01 -4.21215326e-01 -6.81202292e-01 -1.05730116e+00
-1.76002994e-01 -7.83246040e-01 2.83714086e-01 -7.41242588e-01
-1.15795982e+00 7.11839020e-01 9.93418247e-02 -1.33868444e+00
5.88110015e-02 -6.44792020e-01 -4.42249149e-01 4.53899950e-01
-1.66335464e+00 -1.15163910e+00 -4.43853050e-01 5.39210618e-01
1.04526901e+00 1.01179764e-01 4.56129640e-01 2.96440452e-01
-7.68248796e-01 3.55218112e-01 3.82391848e-02 4.55312401e-01
4.49222565e-01 -1.18578839e+00 9.17089045e-01 8.86977077e-01
1.13092817e-01 5.12295961e-01 2.21044421e-01 -4.71973270e-01
-1.15972078e+00 -1.37665141e+00 5.02095520e-01 -4.19413686e-01
2.86137164e-01 -3.91124874e-01 -8.95807505e-01 9.41662192e-01
2.92680532e-01 3.93381901e-02 4.01185691e-01 -1.65056303e-01
-2.84334589e-02 -6.02701791e-02 -1.09591877e+00 6.32521987e-01
1.14140582e+00 -3.00516158e-01 -4.73476112e-01 3.77461016e-01
9.00922239e-01 -8.97442341e-01 -7.21042752e-01 5.19550860e-01
3.86707097e-01 -1.00723100e+00 1.18250251e+00 -2.25471422e-01
2.56370574e-01 -3.34594220e-01 1.04642421e-01 -9.80331659e-01
-4.11674887e-01 -5.16830146e-01 3.39919887e-02 7.53137589e-01
-1.85420434e-03 -5.61941504e-01 1.11552238e+00 3.58901560e-01
-7.69791305e-01 -1.12989080e+00 -8.96265030e-01 -5.29389501e-01
3.18494380e-01 -6.62960768e-01 4.99683022e-01 7.63211787e-01
-3.81076336e-01 2.07473055e-01 1.62619501e-01 1.12265937e-01
6.46440029e-01 3.47251028e-01 5.77171683e-01 -1.00535309e+00
1.15972862e-01 -8.70063365e-01 -4.76782233e-01 -1.70542288e+00
4.26047705e-02 -8.85787904e-01 1.96215272e-01 -2.13688540e+00
-7.92668983e-02 -3.50710452e-01 -4.41976339e-02 6.61362648e-01
9.47671011e-02 6.01463318e-01 2.19909593e-01 2.79147059e-01
-6.84425056e-01 3.19935381e-01 1.45891476e+00 -5.01913190e-01
-2.33411193e-01 1.44883558e-01 -6.39999449e-01 8.53265107e-01
1.03355408e+00 -5.97532615e-02 -5.41584373e-01 -8.79573524e-01
-1.61817178e-01 -1.53290048e-01 3.23441058e-01 -1.06343281e+00
3.76917362e-01 -2.18453452e-01 2.22961932e-01 -9.91402924e-01
4.54188704e-01 -7.03172028e-01 -9.14054811e-02 3.94168735e-01
-4.14383262e-02 2.01268464e-01 4.14373457e-01 3.58939648e-01
-3.91713768e-01 4.46495488e-02 9.83326018e-01 -5.07205725e-01
-1.18078017e+00 6.95135117e-01 -5.46840668e-01 8.74629319e-02
1.15087819e+00 -7.43226826e-01 -1.10497773e-01 -6.65919036e-02
-7.10679173e-01 5.58470488e-01 4.66515720e-01 5.52349687e-01
8.66013050e-01 -8.43856454e-01 -4.86067086e-01 1.04467943e-01
8.83313790e-02 6.51600838e-01 3.86728793e-01 9.75502610e-01
-8.90132725e-01 6.07955575e-01 -1.33237839e-01 -1.13888443e+00
-1.15809500e+00 5.87541819e-01 7.13200331e-01 4.95536365e-02
-8.54653299e-01 9.58897948e-01 5.16964674e-01 -4.25658166e-01
4.33841407e-01 -9.28847432e-01 -2.26628244e-01 -2.01543137e-01
1.25511870e-01 1.97771519e-01 1.72756910e-01 -5.35872936e-01
-6.46932960e-01 1.04519558e+00 -8.56204480e-02 2.37776816e-01
1.06055808e+00 -3.24005634e-01 -6.47970736e-02 5.39311588e-01
1.16211653e+00 -5.56367695e-01 -1.23565960e+00 -1.81676269e-01
-2.07111165e-01 -1.94596741e-02 9.47991014e-03 -4.79426742e-01
-1.55729306e+00 1.24130619e+00 5.22648156e-01 1.81801349e-01
1.37740612e+00 2.84449100e-01 1.05592966e+00 1.87146932e-01
3.05798650e-01 -9.38162506e-01 4.70311905e-04 6.58725500e-01
6.31981432e-01 -1.26698899e+00 -7.02776015e-02 -5.62289655e-01
-6.02315128e-01 1.06664884e+00 6.30923331e-01 -8.16872492e-02
8.70699108e-01 2.60443658e-01 4.72415447e-01 -4.46525484e-01
-3.96609902e-01 -3.98216188e-01 1.38964787e-01 9.43892479e-01
4.15281683e-01 -4.35865000e-02 1.48241177e-01 2.74238408e-01
-1.59968197e-01 1.64943859e-01 3.32096577e-01 7.81389654e-01
-4.03999239e-01 -8.67115259e-01 -2.81812072e-01 5.77090621e-01
-2.59582102e-01 -1.69634763e-02 -3.76513422e-01 6.97074831e-01
1.55967444e-01 8.00056159e-01 5.17417967e-01 -3.34166884e-01
2.77229697e-01 -2.94803858e-01 4.01818722e-01 -5.97327769e-01
-5.86284041e-01 1.91869974e-01 -2.97827065e-01 -7.05860078e-01
-8.13364089e-01 -5.95134258e-01 -1.43168640e+00 -2.37881273e-01
-8.22669268e-02 -3.52822989e-02 4.92662162e-01 8.27510238e-01
4.73710597e-01 6.56602919e-01 1.91374257e-01 -1.13229477e+00
5.25959954e-02 -5.50407946e-01 -2.34148338e-01 -1.22717358e-01
5.18838167e-01 -6.48989022e-01 1.42570948e-02 2.23419294e-01] | [9.470132827758789, 0.09020749479532242] |
f4d26d3f-1492-46bf-9ef0-6880f7454bf2 | efficient-eigen-updating-for-spectral-graph | 1301.1318 | null | http://arxiv.org/abs/1301.1318v4 | http://arxiv.org/pdf/1301.1318v4.pdf | Efficient Eigen-updating for Spectral Graph Clustering | Partitioning a graph into groups of vertices such that those within each
group are more densely connected than vertices assigned to different groups,
known as graph clustering, is often used to gain insight into the organisation
of large scale networks and for visualisation purposes. Whereas a large number
of dedicated techniques have been recently proposed for static graphs, the
design of on-line graph clustering methods tailored for evolving networks is a
challenging problem, and much less documented in the literature. Motivated by
the broad variety of applications concerned, ranging from the study of
biological networks to the analysis of networks of scientific references
through the exploration of communications networks such as the World Wide Web,
it is the main purpose of this paper to introduce a novel, computationally
efficient, approach to graph clustering in the evolutionary context. Namely,
the method promoted in this article can be viewed as an incremental eigenvalue
solution for the spectral clustering method described by Ng. et al. (2001). The
incremental eigenvalue solution is a general technique for finding the
approximate eigenvectors of a symmetric matrix given a change. As well as
outlining the approach in detail, we present a theoretical bound on the quality
of the approximate eigenvectors using perturbation theory. We then derive a
novel spectral clustering algorithm called Incremental Approximate Spectral
Clustering (IASC). The IASC algorithm is simple to implement and its efficacy
is demonstrated on both synthetic and real datasets modelling the evolution of
a HIV epidemic, a citation network and the purchase history graph of an
e-commerce website. | ['Stéphan Clémençon', 'Romaric Gaudel', 'Charanpal Dhanjal'] | 2013-01-07 | null | null | null | null | ['spectral-graph-clustering'] | ['graphs'] | [ 2.02065676e-01 2.08337218e-01 6.91211596e-02 2.71890372e-01
1.54880121e-01 -8.56401801e-01 5.12661338e-01 4.69965130e-01
-9.42587033e-02 4.38680023e-01 -4.64744791e-02 -4.94359285e-01
-8.21118534e-01 -7.23611832e-01 -3.27277005e-01 -7.34248340e-01
-6.74025834e-01 6.35050654e-01 3.71048331e-01 -5.52992858e-02
2.61329621e-01 6.91718698e-01 -1.23735046e+00 -2.28600100e-01
7.61733711e-01 4.94872481e-01 6.21774532e-02 5.77987075e-01
-2.50172138e-01 8.11638907e-02 -4.13690299e-01 -3.78155053e-01
3.78192186e-01 -8.17762613e-01 -9.06608462e-01 4.66986001e-01
-3.13862950e-01 7.37155378e-01 -1.89513877e-01 1.21080399e+00
4.80334044e-01 2.00387150e-01 6.66121662e-01 -1.29842234e+00
-2.22537667e-01 7.06395388e-01 -8.17532003e-01 2.24090055e-01
2.92628288e-01 -3.21815163e-01 9.52880442e-01 -2.99418390e-01
9.89018679e-01 1.08628583e+00 6.87983155e-01 1.18946485e-01
-1.77195394e+00 -1.70616075e-01 -3.87094356e-02 2.33490467e-01
-1.48995483e+00 -1.09555028e-01 9.09630656e-01 -5.16123772e-01
4.85364974e-01 4.23653096e-01 9.28297400e-01 5.75115919e-01
-3.55922058e-02 3.15293044e-01 9.55940187e-01 -4.61274832e-01
5.48567176e-01 -1.02296546e-01 7.79045075e-02 4.60267305e-01
5.06332099e-01 -2.92444766e-01 -2.03206003e-01 -6.32528126e-01
4.80002165e-01 -5.26618622e-02 -3.06485981e-01 -9.79064822e-01
-9.81019735e-01 8.81177545e-01 4.15286452e-01 6.22719109e-01
-3.91666353e-01 -1.10366665e-01 4.19169068e-01 4.77597207e-01
4.95534897e-01 3.73210907e-01 -6.10969663e-02 -7.64713669e-03
-9.31355774e-01 1.18543513e-01 1.13497579e+00 6.40582860e-01
6.98944569e-01 -1.70940667e-01 6.33249819e-01 6.78105414e-01
3.20023865e-01 2.19470151e-02 2.80114591e-01 -8.63046288e-01
-1.16513520e-01 1.01876271e+00 -3.51812452e-01 -1.38771379e+00
-5.56401074e-01 -4.77423400e-01 -1.18949842e+00 -1.05034769e-01
5.02270460e-01 -3.73317450e-02 -3.91918272e-01 1.52594626e+00
5.33855557e-01 1.83196396e-01 -2.01253012e-01 4.97868180e-01
2.60636717e-01 5.08574903e-01 -3.87161523e-01 -7.27498949e-01
1.00001276e+00 -3.91995102e-01 -3.62948686e-01 2.40476385e-01
4.02278036e-01 -6.20375514e-01 3.72658908e-01 2.83652604e-01
-8.66273224e-01 -8.33750218e-02 -6.67820215e-01 6.66598320e-01
-3.82153541e-01 -6.06808245e-01 4.99979824e-01 6.73491716e-01
-1.45276427e+00 7.38025725e-01 -7.96335518e-01 -1.14960277e+00
1.84132263e-01 3.10556412e-01 -3.23484480e-01 8.15111399e-03
-8.02289724e-01 4.86680388e-01 4.02070820e-01 -9.84866768e-02
-1.05896145e-02 -3.28596205e-01 -3.29932779e-01 1.65145606e-01
6.58939421e-01 -5.92457056e-01 5.07097721e-01 -1.24605393e+00
-1.18137145e+00 7.50452578e-01 -8.51382613e-02 -4.52695251e-01
4.17625934e-01 6.49704278e-01 -3.98061156e-01 1.91697478e-01
-7.86722451e-02 1.81006379e-02 5.51524162e-01 -1.35581958e+00
-1.61969379e-01 -5.26887298e-01 -3.77922088e-01 -1.92197300e-02
-1.90791577e-01 4.26684655e-02 -4.28196788e-01 -6.37705445e-01
2.95376271e-01 -1.20084882e+00 -4.91825551e-01 -3.66915256e-01
-4.51593459e-01 -1.90021276e-01 7.19209433e-01 -4.05586690e-01
1.56932247e+00 -2.09937763e+00 8.16988885e-01 1.08970201e+00
4.77827042e-01 1.37960896e-01 8.35945159e-02 1.10861135e+00
-3.50442111e-01 3.56898665e-01 -6.17698431e-01 2.73390263e-01
-1.94499671e-01 1.82170138e-01 4.61813025e-02 6.91083252e-01
-2.40388587e-01 6.40367866e-01 -1.06076765e+00 -2.31582433e-01
6.31712889e-03 2.59682894e-01 -4.15578514e-01 -1.77087128e-01
1.09635964e-01 2.77840458e-02 -1.39975712e-01 3.05301458e-01
4.60706055e-01 -5.36278248e-01 8.03120494e-01 -8.44413508e-03
-1.45925432e-01 -4.06629533e-01 -1.43477917e+00 1.08693290e+00
3.59725475e-01 6.51118159e-01 4.37840313e-01 -1.25335836e+00
8.00301015e-01 2.51763582e-01 9.12467599e-01 -1.58009633e-01
1.16620108e-01 1.35610595e-01 4.27660137e-01 -4.34781089e-02
1.82006225e-01 -1.53210042e-02 9.06282961e-02 7.77889848e-01
-1.92875057e-01 1.90208703e-01 6.10031903e-01 7.36487329e-01
1.42968655e+00 -3.13474089e-01 6.97240710e-01 -5.38422167e-01
4.33338255e-01 1.09475575e-01 1.19928837e-01 3.85007083e-01
-2.78682083e-01 3.12549710e-01 9.18052137e-01 -2.60816067e-01
-1.14494193e+00 -9.68584120e-01 -2.35848054e-02 5.25691211e-01
-5.36226202e-03 -7.36622632e-01 -8.85142982e-01 -2.21918330e-01
7.15493634e-02 1.06069058e-01 -7.08628535e-01 -8.58269706e-02
-2.26256326e-01 -8.72715652e-01 1.41584456e-01 -2.35209465e-01
1.47103041e-01 -9.07195091e-01 -3.59197617e-01 2.61576802e-01
4.67858389e-02 -7.30658650e-01 -3.23719233e-01 4.88741463e-03
-9.30801213e-01 -1.52456427e+00 -4.62734640e-01 -7.11621523e-01
7.89315283e-01 2.88340092e-01 9.43360448e-01 3.09613526e-01
-4.86505985e-01 7.31422305e-01 -4.56389070e-01 -7.61426389e-02
-5.60709119e-01 2.43621096e-01 1.64601952e-01 3.75131279e-01
1.95814222e-01 -8.46123934e-01 -3.44811261e-01 3.43867511e-01
-1.16028571e+00 -1.70610547e-01 4.33243960e-01 5.81933379e-01
4.85816866e-01 3.89505327e-01 3.89315307e-01 -8.80463481e-01
1.02875948e+00 -8.58656585e-01 -5.04642546e-01 2.40708470e-01
-8.35287511e-01 5.14027523e-03 6.58026099e-01 -2.99920976e-01
-3.14348102e-01 2.03823894e-01 3.60758454e-01 -3.01859498e-01
4.20560539e-02 8.85131299e-01 5.72000481e-02 -3.62192959e-01
5.23338497e-01 3.19356233e-01 4.74278301e-01 -4.18773323e-01
4.87013042e-01 4.86110628e-01 2.96483546e-01 -2.63518333e-01
1.03756702e+00 5.31456530e-01 4.33535039e-01 -1.19965887e+00
1.90659985e-02 -8.27897370e-01 -8.63463759e-01 -5.48002958e-01
4.96888578e-01 -1.57908306e-01 -8.27851176e-01 1.57722101e-01
-6.86025262e-01 -1.79912746e-01 -1.51156142e-01 -3.63165326e-02
-4.76846009e-01 7.74562478e-01 -3.03337961e-01 -8.11317980e-01
1.10302925e-01 -6.88735783e-01 3.98319095e-01 -2.53167488e-02
-4.21256751e-01 -1.37889826e+00 6.01779699e-01 -1.00087471e-01
2.68149406e-01 7.33958960e-01 1.11605227e+00 -6.77324533e-01
-3.90929788e-01 -1.73309192e-01 -3.98678221e-02 -1.37124598e-01
1.68286398e-01 4.68967915e-01 -2.14179948e-01 -5.45523107e-01
-2.18008876e-01 4.22856063e-01 5.12240648e-01 3.10031265e-01
6.50138795e-01 -2.18840569e-01 -6.99191391e-01 3.55970770e-01
1.51514196e+00 2.00072527e-01 3.91321093e-01 2.24528879e-01
5.91310143e-01 8.82884443e-01 2.81813089e-02 3.74257237e-01
5.28960489e-02 6.60267711e-01 2.76336789e-01 1.57125577e-01
1.44571885e-01 1.54486984e-01 6.12873212e-02 1.17813241e+00
-3.35049033e-01 -2.33854845e-01 -1.16604352e+00 6.06962323e-01
-2.14153028e+00 -1.22592819e+00 -3.58176917e-01 2.26747561e+00
3.72639418e-01 -1.08527273e-01 7.33481050e-01 2.90089071e-01
1.16248894e+00 -1.05937056e-01 -4.11656320e-01 -4.59858388e-01
-2.11502239e-01 -6.59457594e-03 4.09379631e-01 5.10869205e-01
-5.22539735e-01 5.29308736e-01 6.44595242e+00 6.31800056e-01
-6.62966311e-01 -1.71268895e-01 5.67002356e-01 3.40021141e-02
-1.59691900e-01 2.37244859e-01 7.35293552e-02 5.19531965e-01
1.15423250e+00 -6.17606223e-01 8.81895483e-01 5.29459000e-01
2.86651582e-01 -1.48298711e-01 -6.68514371e-01 8.98523688e-01
-1.68677136e-01 -1.20306134e+00 -1.03092775e-01 5.67647815e-01
7.30046630e-01 -1.35636544e-02 -2.26328537e-01 -3.03936213e-01
4.52760458e-01 -8.37754190e-01 2.85648406e-01 3.14218640e-01
4.96101201e-01 -1.01408613e+00 3.63203436e-01 2.38642797e-01
-1.47336137e+00 -6.50998130e-02 -8.05601105e-02 -2.98910178e-02
2.57596821e-01 6.46761119e-01 -6.30125821e-01 8.09420764e-01
6.33359194e-01 7.17880309e-01 -5.79457283e-01 1.30684602e+00
3.50841343e-01 7.19292521e-01 -3.55188370e-01 -1.32702589e-01
2.27270484e-01 -7.35918701e-01 8.76839578e-01 8.36302221e-01
2.30293661e-01 6.96499720e-02 -1.67158656e-02 6.40778303e-01
1.25204669e-02 3.57985556e-01 -6.94305837e-01 -4.27374035e-01
4.23632026e-01 1.32930195e+00 -1.53395879e+00 2.94844378e-02
-2.43022382e-01 8.75312865e-01 3.81043792e-01 4.57104146e-01
-2.80869871e-01 -5.09302914e-01 4.47758734e-01 4.74737227e-01
3.52729708e-01 -3.63184333e-01 6.91995025e-02 -7.74542153e-01
-3.00940394e-01 -8.49737346e-01 5.07588863e-01 -3.85703593e-01
-1.25836241e+00 5.15345991e-01 1.03030890e-01 -9.69647467e-01
-2.34721631e-01 -3.48779172e-01 -5.81414819e-01 5.36567211e-01
-7.10816324e-01 -4.58435893e-01 -1.54672757e-01 5.85583448e-01
-3.34565528e-02 -1.02754876e-01 5.11896074e-01 9.15514082e-02
-5.82279384e-01 -1.39021590e-01 8.91613543e-01 -8.18323493e-02
3.53893995e-01 -1.29779184e+00 3.77092987e-01 9.15035069e-01
1.90784886e-01 6.76059425e-01 8.16121578e-01 -8.90793443e-01
-1.37465286e+00 -7.94654489e-01 7.57425785e-01 -1.47554293e-01
1.05648327e+00 -4.16788906e-01 -9.25425291e-01 4.75462824e-01
2.96968609e-01 -2.38018438e-01 7.76579797e-01 2.03798413e-02
4.85922769e-02 8.04294925e-03 -1.05753827e+00 8.51954103e-01
1.22740197e+00 -2.57733732e-01 -2.38593388e-02 3.21630359e-01
2.10763022e-01 2.68838197e-01 -9.97953951e-01 -6.44230396e-02
3.26566726e-01 -1.15364778e+00 7.96016932e-01 -4.42414224e-01
-1.46419108e-01 -5.30941367e-01 2.14875728e-01 -1.50530195e+00
-7.89038599e-01 -1.14798474e+00 -9.14302170e-02 1.11658239e+00
7.27954507e-02 -7.70110667e-01 8.32847178e-01 8.50739777e-02
3.24233800e-01 -6.51093841e-01 -1.05545092e+00 -7.29322851e-01
-2.50924557e-01 4.12420146e-02 2.77542382e-01 1.15014160e+00
4.52600330e-01 4.51200604e-01 -4.47153114e-02 -1.18571132e-01
8.81183028e-01 1.77474812e-01 6.26699090e-01 -1.82726550e+00
-3.52412492e-01 -9.68016088e-01 -8.27195406e-01 -3.67997617e-01
2.41792817e-02 -1.03117681e+00 -4.22020674e-01 -1.55911577e+00
2.49588504e-01 -2.64440387e-01 5.12466878e-02 -4.53555658e-02
3.73650715e-02 2.00833946e-01 1.46132320e-01 4.85876232e-01
-5.69524050e-01 9.35739502e-02 7.07850814e-01 9.98755544e-02
-3.11180353e-01 6.75587803e-02 -6.15741253e-01 3.55055928e-01
6.34875894e-01 -4.57110047e-01 -5.49631655e-01 4.96434480e-01
6.21645808e-01 -3.72150280e-02 8.71570557e-02 -6.97806001e-01
3.54859740e-01 3.19218859e-02 -1.15819737e-01 -2.73171246e-01
-2.69844770e-01 -9.72377300e-01 9.66895223e-01 7.59167671e-01
-6.99531436e-02 3.78161937e-01 -7.63875097e-02 1.08139789e+00
-1.32460192e-01 -1.10275164e-01 7.31223106e-01 -2.28955090e-01
-4.26325738e-01 2.13873640e-01 -5.58949769e-01 -3.06730885e-02
1.29878747e+00 -3.91848296e-01 -8.29005986e-02 -5.35800755e-01
-8.80436420e-01 1.87004581e-01 8.64880383e-01 1.01997837e-01
3.52230519e-01 -1.14486289e+00 -5.90130150e-01 -1.58433810e-01
-1.63529754e-01 -2.92472720e-01 2.43534651e-02 9.99645770e-01
-5.49507856e-01 2.23715186e-01 -1.14397369e-01 -6.02216542e-01
-1.57866693e+00 9.19956923e-01 2.40413129e-01 -2.58670568e-01
-3.99564385e-01 3.34625661e-01 -1.59533441e-01 -2.24040270e-01
-1.20707430e-01 3.07468057e-01 -1.76332042e-01 2.13942438e-01
1.42106861e-01 7.65434682e-01 -1.76117003e-01 -9.19046640e-01
-4.17034268e-01 6.11069500e-01 2.99568981e-01 4.75694314e-02
1.47487724e+00 -4.33724850e-01 -7.66426504e-01 6.58869445e-01
1.07087815e+00 2.88876239e-02 -6.49096489e-01 -1.89239994e-01
4.91615683e-01 -1.75668254e-01 -3.37540001e-01 -3.09243679e-01
-9.63780999e-01 3.23259860e-01 2.85355568e-01 1.20568120e+00
1.19980204e+00 2.59273261e-01 2.27370873e-01 2.22032875e-01
2.02669695e-01 -9.11435366e-01 -1.98015235e-02 2.04212889e-01
5.82002699e-01 -7.33421087e-01 1.73928663e-01 -6.45598531e-01
-2.81556785e-01 1.19924879e+00 -9.20964405e-02 -2.44497508e-01
8.69195580e-01 7.80888349e-02 -4.20856953e-01 -4.74544615e-01
-5.19071758e-01 -2.11701035e-01 1.98475212e-01 6.35106027e-01
2.78355688e-01 -2.68797111e-03 -5.16175449e-01 -3.29673663e-02
-2.29723006e-02 -4.06583220e-01 6.43701494e-01 7.80289412e-01
-4.35336083e-01 -1.21121001e+00 -3.03615719e-01 5.76646686e-01
-1.84304088e-01 1.51374983e-03 -1.15738893e+00 7.31296062e-01
-2.55298376e-01 7.57768452e-01 -1.59757491e-03 -2.79184937e-01
1.18775621e-01 2.23779783e-01 3.63921463e-01 -4.30434644e-01
-5.64563453e-01 5.63670546e-02 1.36710638e-02 -3.62740338e-01
-7.97850013e-01 -9.72794056e-01 -9.26526248e-01 -7.45025873e-01
-2.32073069e-01 5.77958882e-01 4.82316315e-01 6.45307064e-01
5.57486296e-01 4.53624338e-01 6.32008374e-01 -7.29999304e-01
-2.25850567e-02 -5.88569641e-01 -1.03670371e+00 4.72538173e-01
-4.49485555e-02 -4.36100334e-01 -6.33612275e-01 -1.84595045e-02] | [7.057108402252197, 5.250741958618164] |
f30e6af8-b9fd-4cca-adca-eab4e9edd098 | linking-garment-with-person-via-semantically | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yan_Linking_Garment_With_Person_via_Semantically_Associated_Landmarks_for_Virtual_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yan_Linking_Garment_With_Person_via_Semantically_Associated_Landmarks_for_Virtual_CVPR_2023_paper.pdf | Linking Garment With Person via Semantically Associated Landmarks for Virtual Try-On | In this paper, a novel virtual try-on algorithm, dubbed SAL-VTON, is proposed, which links the garment with the person via semantically associated landmarks to alleviate misalignment. The semantically associated landmarks are a series of landmark pairs with the same local semantics on the in-shop garment image and the try-on image. Based on the semantically associated landmarks, SAL-VTON effectively models the local semantic association between garment and person, making up for the misalignment in the overall deformation of the garment. The outcome is achieved with a three-stage framework: 1) the semantically associated landmarks are estimated using the landmark localization model; 2) taking the landmarks as input, the warping model explicitly associates the corresponding parts of the garment and person for obtaining the local flow, thus refining the alignment in the global flow; 3) finally, a generator consumes the landmarks to better capture local semantics and control the try-on results.Moreover, we propose a new landmark dataset with a unified labelling rule of landmarks for diverse styles of garments. Extensive experimental results on popular datasets demonstrate that SAL-VTON can handle misalignment and outperform state-of-the-art methods both qualitatively and quantitatively. The dataset is available on https://modelscope.cn/datasets/damo/SAL-HG/summary. | ['Chengjun Xie', 'HUI ZHANG', 'Tingwei Gao', 'Keyu Yan'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['virtual-try-on'] | ['computer-vision'] | [-3.52029830e-01 -4.25427184e-02 -2.36921757e-01 -2.51592398e-01
-5.11285365e-01 -4.47398096e-01 5.35649538e-01 -2.03550726e-01
6.33962974e-02 2.36342147e-01 4.92838353e-01 3.88162613e-01
-1.40915841e-01 -6.85137272e-01 -5.55332720e-01 -6.62447512e-01
2.98303012e-02 5.97518623e-01 1.05367221e-01 -3.21903020e-01
7.48001561e-02 3.75924259e-01 -8.70181143e-01 -2.21683994e-01
5.89186370e-01 8.44987631e-01 2.52931923e-01 3.16655159e-01
1.85892910e-01 3.76710176e-01 -3.13205361e-01 -4.50805455e-01
6.32093906e-01 -6.57247841e-01 -6.88021600e-01 3.97797883e-01
5.88691831e-01 8.92135501e-03 -4.32289392e-01 1.04667377e+00
4.63339716e-01 4.96254981e-01 3.41959327e-01 -1.35559571e+00
-6.68791771e-01 5.43278158e-01 -6.51911914e-01 -7.83278793e-02
3.96739244e-01 2.27962092e-01 1.03301430e+00 -1.01048017e+00
9.14516687e-01 1.53377867e+00 5.84626079e-01 4.36801136e-01
-1.06150079e+00 -4.18975055e-01 4.39233959e-01 1.00695796e-01
-1.54807818e+00 -2.53788441e-01 1.18149221e+00 -4.01637226e-01
2.27534801e-01 2.82644570e-01 7.61918187e-01 9.05189991e-01
-1.22762369e-02 6.81753099e-01 5.97930014e-01 -2.67405450e-01
-3.82395945e-02 -3.85241002e-01 -1.76406354e-01 1.07569456e+00
-3.16135809e-02 6.95636347e-02 -5.28207004e-01 2.21279204e-01
1.17031586e+00 1.53789401e-01 -1.36265308e-01 -6.48031592e-01
-1.51748860e+00 5.51009119e-01 8.29041004e-01 3.39845210e-01
-4.70088571e-01 2.13521585e-01 3.74336451e-01 -1.07921630e-01
4.18458253e-01 3.15959990e-01 -2.77141005e-01 9.25107673e-02
-7.72087395e-01 3.27481210e-01 5.66719532e-01 1.14787018e+00
9.72844541e-01 -2.16631070e-01 -1.88652068e-01 7.06045806e-01
6.20331287e-01 3.60217929e-01 2.41473988e-02 -9.25197303e-01
7.78562427e-01 8.63440573e-01 1.49167791e-01 -1.37175393e+00
-5.20782053e-01 -4.99944121e-01 -8.14716339e-01 -1.67439714e-01
5.46224415e-01 1.01735398e-01 -9.58519876e-01 1.59871769e+00
5.25222242e-01 6.32609308e-01 -3.19719493e-01 1.27591288e+00
7.29613602e-01 3.88854206e-01 1.53802335e-01 3.38428676e-01
1.62613952e+00 -1.47323322e+00 -6.59887254e-01 -4.83446002e-01
4.11835611e-01 -9.96615469e-01 1.03224945e+00 -2.03037232e-01
-1.02262378e+00 -9.16864395e-01 -9.30294216e-01 -2.30172798e-01
-1.03868358e-01 5.48145235e-01 4.38692927e-01 1.96030974e-01
-7.46153057e-01 5.68922937e-01 -9.97713864e-01 -4.95019346e-01
2.69494444e-01 1.61978062e-02 -4.13068980e-01 -1.26753524e-01
-9.94798183e-01 7.74554431e-01 2.80391723e-01 7.28583753e-01
-8.10983658e-01 -5.95520496e-01 -1.32212913e+00 -1.75032392e-01
6.24244034e-01 -8.87135088e-01 6.69562936e-01 -7.98863590e-01
-1.36414385e+00 9.39570367e-01 -2.27266073e-01 -1.07583683e-02
9.51308072e-01 -4.01527464e-01 -3.13583493e-01 8.61231014e-02
6.18153393e-01 6.39622271e-01 7.10161984e-01 -1.52889848e+00
-4.83044058e-01 -3.14061165e-01 -1.00543842e-01 4.06736702e-01
3.45093250e-01 -1.03435535e-02 -1.15370083e+00 -9.37333941e-01
3.94340724e-01 -1.07870901e+00 -1.60139531e-01 -6.28011823e-02
-9.02775705e-01 -1.60835177e-01 6.98643506e-01 -9.98619735e-01
1.18680358e+00 -2.31036687e+00 6.23275995e-01 5.22142887e-01
2.12842226e-01 -4.29469757e-02 -1.88522846e-01 3.13280702e-01
1.93232822e-03 -1.03917509e-01 -1.64710343e-01 -8.86061013e-01
1.32884398e-01 2.69457579e-01 -1.91154242e-01 7.18733311e-01
-2.87145972e-02 1.20095336e+00 -1.17487383e+00 -5.81370831e-01
7.08601475e-01 4.78712261e-01 -2.41095096e-01 3.24017107e-01
1.04841098e-01 8.56565058e-01 -6.33243024e-01 1.03577602e+00
4.62879419e-01 -5.21652400e-02 8.13527107e-02 -6.16836667e-01
6.69554323e-02 5.47607206e-02 -1.45938981e+00 2.10974574e+00
-2.16642141e-01 9.23849046e-02 2.81050298e-02 -6.08338892e-01
1.25015795e+00 1.52759776e-01 6.74515843e-01 -4.72779602e-01
1.86642766e-01 1.71399221e-01 -3.70615900e-01 -5.17654538e-01
4.36152220e-01 1.47614539e-01 -3.97414416e-01 6.58193752e-02
1.11597262e-01 4.36789989e-02 2.45885491e-01 3.07029136e-03
4.43235159e-01 6.81757629e-01 3.65807526e-02 -2.39350930e-01
6.65837348e-01 -1.96161479e-01 7.20782995e-01 3.26553106e-01
-3.58835936e-01 7.64360309e-01 2.99909860e-01 -7.14353681e-01
-1.02800167e+00 -1.14244819e+00 2.03361571e-01 7.82200515e-01
8.38105619e-01 -6.81831360e-01 -8.22206140e-01 -7.11265028e-01
4.69497703e-02 3.92691821e-01 -6.59421921e-01 -1.51520625e-01
-9.23135936e-01 -3.34979087e-01 3.31546903e-01 6.48841083e-01
7.83354104e-01 -9.75384772e-01 -2.95655191e-01 7.23362565e-02
-5.08122027e-01 -1.15446603e+00 -1.10468268e+00 -5.07814765e-01
-5.74856699e-01 -1.16910005e+00 -6.34612024e-01 -8.24861884e-01
9.75570023e-01 8.63490924e-02 9.02718365e-01 2.14669064e-01
-1.67993933e-01 1.73975274e-01 -2.59803891e-01 2.04285994e-01
-9.96359512e-02 7.14132786e-02 -5.61604016e-02 4.45317984e-01
4.47860882e-02 -4.00118440e-01 -8.26389432e-01 8.05188596e-01
-3.79660100e-01 2.93006122e-01 2.53525943e-01 6.06493473e-01
8.03538322e-01 -1.74182020e-02 1.55297816e-01 -3.99655998e-01
1.79809943e-01 -3.44231844e-01 -2.12869883e-01 1.27459019e-01
-1.05969608e-01 -5.61797507e-02 4.15373325e-01 -3.02073091e-01
-7.73371637e-01 1.04724981e-01 -9.06779617e-02 -6.44861698e-01
-7.04824403e-02 1.09016530e-01 -5.68324387e-01 -1.45970192e-03
1.47107184e-01 -9.43910256e-02 -5.68780079e-02 -6.93135619e-01
6.49463296e-01 1.87433213e-02 9.67012525e-01 -6.81947649e-01
9.99047697e-01 6.61737382e-01 1.27168804e-01 -3.25354338e-01
-8.61263812e-01 -6.13563001e-01 -1.00205481e+00 -5.52725315e-01
1.15560353e+00 -7.77523220e-01 -5.00058651e-01 5.21087289e-01
-9.73710895e-01 -3.82766902e-01 -3.91389310e-01 5.41366577e-01
-6.04611039e-01 2.68599272e-01 -5.45508921e-01 -3.67395639e-01
-1.75820455e-01 -1.42080212e+00 1.37124860e+00 4.18675274e-01
-3.81082714e-01 -1.34629846e+00 -7.50414236e-03 4.27361250e-01
-1.91177037e-02 6.60001338e-01 4.69421268e-01 -3.32341433e-01
-6.72769010e-01 -1.29937172e-01 -8.18010941e-02 1.04520313e-01
4.12124515e-01 -8.07665214e-02 -5.53356349e-01 -2.67781317e-01
-4.45566028e-01 2.10540414e-01 6.86908960e-01 4.84762132e-01
6.89839005e-01 -6.40709698e-02 -4.36242759e-01 8.74658942e-01
1.27387822e+00 -8.60685632e-02 6.15762353e-01 4.55972940e-01
1.34159088e+00 6.59680486e-01 9.25320446e-01 1.32998109e-01
6.34168506e-01 9.41663742e-01 4.32703972e-01 -3.89793128e-01
-4.90149885e-01 -8.74782562e-01 2.37078130e-01 8.15360665e-01
-4.83449161e-01 -1.09934369e-02 -7.41266727e-01 6.62089169e-01
-2.19110370e+00 -6.83771908e-01 -6.37945831e-02 2.06509447e+00
4.18147475e-01 -3.73862199e-02 1.48347408e-01 -2.80813843e-01
9.54246759e-01 5.10790229e-01 -2.96000779e-01 2.04180464e-01
1.48571074e-01 -9.51932892e-02 4.35899019e-01 7.54009128e-01
-1.33374476e+00 1.06509829e+00 5.22021341e+00 5.58250964e-01
-8.47525775e-01 7.27870613e-02 4.28916544e-01 1.50176108e-01
7.27173835e-02 1.93194389e-01 -8.48132849e-01 5.05628049e-01
8.70467871e-02 -4.62055113e-03 3.04172903e-01 8.18230748e-01
4.61640120e-01 1.99071676e-01 -1.02661288e+00 8.19176197e-01
1.81726798e-01 -1.10269713e+00 -6.98555913e-03 -1.41283467e-01
6.53165042e-01 -4.34999228e-01 -2.16854066e-01 -7.21662790e-02
7.98281208e-02 -6.70988202e-01 1.18754840e+00 8.18225861e-01
7.05926299e-01 -7.66103327e-01 6.53473139e-01 1.50356153e-02
-1.83557630e+00 2.84858942e-01 -5.55905886e-02 2.79301196e-01
5.74114621e-01 2.28025019e-01 -3.38882595e-01 1.09970272e+00
5.69274247e-01 1.15528178e+00 -5.11978507e-01 8.72267544e-01
-6.96000099e-01 1.32910132e-01 -1.32455960e-01 5.53185821e-01
2.45340690e-01 -6.00271106e-01 7.63602078e-01 1.06264293e+00
3.10350806e-01 -3.17276716e-01 8.06734860e-01 9.98506427e-01
-7.97343161e-03 -3.12271323e-02 -2.89573580e-01 3.31066072e-01
5.43083370e-01 1.44668865e+00 -8.97558153e-01 -2.82836199e-01
-9.71109420e-02 1.09022045e+00 8.53199363e-02 4.53248084e-01
-9.85330462e-01 -1.25452787e-01 8.61354053e-01 1.69238880e-01
1.04344755e-01 -3.89710099e-01 -2.34467611e-01 -1.12167454e+00
1.88656360e-01 -4.47213709e-01 3.15849274e-01 -7.75858462e-01
-1.19558823e+00 4.39942420e-01 9.42259505e-02 -1.24898255e+00
-6.05538487e-04 -1.82537004e-01 -7.32662678e-01 1.04050958e+00
-1.32818902e+00 -1.68786561e+00 -5.91188371e-01 6.67239368e-01
5.61596334e-01 6.26102239e-02 5.53669095e-01 4.11791235e-01
-8.46891463e-01 3.92824411e-01 -6.31423235e-01 4.66922432e-01
8.00736010e-01 -1.15468681e+00 6.72811568e-01 1.22240293e+00
1.53846964e-01 7.35908031e-01 4.56264228e-01 -9.57381845e-01
-1.18438041e+00 -1.39475536e+00 8.52995455e-01 -5.30207396e-01
5.75717449e-01 -4.49148566e-01 -7.33936191e-01 1.09032559e+00
-1.55804724e-01 1.48004666e-01 2.12735400e-01 -1.61784917e-01
-1.32021144e-01 -8.77179429e-02 -9.27152693e-01 5.68398416e-01
1.35504246e+00 -4.35703725e-01 -6.21400893e-01 2.74362832e-01
5.78535974e-01 -9.72648501e-01 -8.74937654e-01 9.88348499e-02
4.29412425e-01 -6.57979608e-01 1.38656378e+00 -2.34730944e-01
4.10783291e-01 -7.51669884e-01 8.99418965e-02 -1.38918424e+00
-5.08594394e-01 -7.81402886e-01 3.44415419e-02 1.45240176e+00
-7.94841349e-02 -4.13232863e-01 4.72508818e-01 5.28941631e-01
-2.75634617e-01 -6.54703081e-01 -8.56308639e-01 -5.70295990e-01
-6.00484729e-01 -3.26858968e-01 7.53169656e-01 1.01865053e+00
-4.38179344e-01 8.65344480e-02 -6.65850699e-01 2.91640520e-01
7.36345291e-01 9.20116156e-02 9.02782202e-01 -8.73870969e-01
-9.95257273e-02 -3.12647551e-01 -6.78218067e-01 -1.11111474e+00
1.73132211e-01 -1.01462185e+00 4.68557999e-02 -1.63579297e+00
-1.37044743e-01 -6.69827580e-01 -7.90188760e-02 7.17370212e-01
-4.25559014e-01 4.53247845e-01 4.93144661e-01 3.78025919e-01
-6.03979349e-01 4.19221193e-01 1.65315735e+00 -2.66333222e-02
-3.66385490e-01 -1.36953697e-01 -5.73117614e-01 8.89236987e-01
5.53806782e-01 -1.83493540e-01 -1.78653583e-01 -3.09327245e-01
-1.02193005e-01 -1.29965916e-01 8.67658973e-01 -7.54833758e-01
1.22988559e-01 -1.58077300e-01 3.07955384e-01 -2.73790985e-01
2.22568393e-01 -8.35592330e-01 3.99836540e-01 4.09678340e-01
-1.80127650e-01 1.12311706e-01 -1.30774483e-01 5.19901395e-01
-1.60097823e-01 1.51204839e-01 7.36455739e-01 -9.18192714e-02
-9.99083340e-01 5.07717609e-01 5.09400070e-01 -1.77709572e-02
1.19613719e+00 1.29109100e-02 -1.43702388e-01 -1.04877435e-01
-1.06300688e+00 6.57342494e-01 5.91377258e-01 7.93110669e-01
5.42922854e-01 -1.64716494e+00 -7.38636017e-01 4.84580755e-01
3.27351987e-02 3.64955455e-01 3.22496235e-01 9.67985630e-01
-6.79271698e-01 4.18279432e-02 -1.70087010e-01 -6.91506326e-01
-1.17941666e+00 3.14619362e-01 5.87034881e-01 -3.63362879e-01
-9.90886033e-01 8.05527210e-01 3.12365890e-01 -4.13951099e-01
3.46854366e-02 -3.68520051e-01 -7.27793947e-02 -6.54542223e-02
3.77493143e-01 5.76569974e-01 -2.77237177e-01 -1.16053438e+00
-5.36357045e-01 1.03337967e+00 2.48244628e-01 4.48784195e-02
1.15733564e+00 -4.71143484e-01 -2.77411461e-01 2.40068465e-01
1.12488687e+00 1.75353989e-01 -1.63124418e+00 -2.80127108e-01
-2.01401591e-01 -8.56392801e-01 -2.50293165e-01 -4.90603507e-01
-1.42934561e+00 4.23254848e-01 2.98694551e-01 -1.37467608e-01
1.02805436e+00 3.69314402e-01 9.11451995e-01 -4.06646132e-01
4.91071135e-01 -8.70851636e-01 6.25941157e-02 1.16811879e-01
1.02180302e+00 -8.15383554e-01 -2.78881732e-02 -7.74156332e-01
-8.25545728e-01 9.84416068e-01 5.43722868e-01 -5.62645853e-01
4.01147604e-01 3.29579525e-02 2.04574138e-01 -3.40429425e-01
6.25331849e-02 -5.28477021e-02 5.88041186e-01 4.62096035e-01
-8.32485706e-02 4.07512009e-01 -1.28940538e-01 6.08724833e-01
-3.48807454e-01 -1.75597921e-01 -9.31334794e-02 7.67732263e-01
1.18592363e-02 -1.17006791e+00 -4.68444079e-01 -1.41313553e-01
6.23096377e-02 3.00485998e-01 -3.21962118e-01 8.76053274e-01
6.17951989e-01 7.28374124e-01 1.56436726e-01 -4.70157653e-01
7.05129921e-01 -7.08770528e-02 4.14283782e-01 -4.02994841e-01
-5.98239422e-01 3.82959396e-01 3.90355960e-02 -9.93985355e-01
-6.18079484e-01 -5.85715473e-01 -1.41377914e+00 -3.65903616e-01
1.21806618e-02 -4.66098338e-02 2.88954109e-01 9.92602825e-01
3.93364578e-01 7.00290203e-01 5.67074478e-01 -1.31428707e+00
-9.94332880e-02 -7.63008535e-01 -6.86303437e-01 9.29117799e-01
2.14244485e-01 -9.65132535e-01 -2.38322139e-01 3.75742346e-01] | [11.876921653747559, -0.8745721578598022] |
e8d9ab76-f6a0-461e-894c-56f417b5a59b | structural-stability-of-infinite-order | 1911.08637 | null | https://arxiv.org/abs/1911.08637v4 | https://arxiv.org/pdf/1911.08637v4.pdf | Robust Inference on Infinite and Growing Dimensional Time Series Regression | We develop a class of tests for time series models such as multiple regression with growing dimension, infinite-order autoregression and nonparametric sieve regression. Examples include the Chow test and general linear restriction tests of growing rank $p$. Employing such increasing $p$ asymptotics, we introduce a new scale correction to conventional test statistics which accounts for a high-order long-run variance (HLV) that emerges as $ p $ grows with sample size. We also propose a bias correction via a null-imposed bootstrap to alleviate finite sample bias without sacrificing power unduly. A simulation study shows the importance of robustifying testing procedures against the HLV even when $ p $ is moderate. The tests are illustrated with an application to the oil regressions in Hamilton (2003). | ['Myung Hwan Seo', 'Abhimanyu Gupta'] | 2019-11-20 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [-2.40860984e-01 -2.75574446e-01 -5.94748974e-01 -3.00851703e-01
-7.67107844e-01 -6.99857116e-01 5.14839053e-01 -2.15023607e-01
-4.15635407e-01 1.09749007e+00 -1.31505728e-03 -1.07356000e+00
-6.43853724e-01 -9.33140635e-01 -6.59279525e-01 -5.29741228e-01
-7.07762778e-01 5.16015440e-02 8.53364915e-02 1.32618830e-01
4.99449044e-01 3.22178215e-01 -1.31684899e+00 -6.09738171e-01
1.16613519e+00 8.54651511e-01 -3.78329903e-01 6.18217766e-01
4.13046956e-01 3.45868796e-01 -5.65625608e-01 -3.11298966e-01
5.93572021e-01 -4.03670877e-01 5.27741760e-02 -8.07166770e-02
-6.47990480e-02 -5.98616481e-01 -3.30805704e-02 1.01874650e+00
5.03338158e-01 2.34501809e-01 9.09688890e-01 -1.38469291e+00
-7.25809991e-01 7.89921343e-01 -9.76921320e-01 4.21416789e-01
1.18958294e-01 1.32290155e-01 8.97966504e-01 -9.28978384e-01
6.40003979e-02 1.20960784e+00 9.55345273e-01 -1.64487615e-01
-1.42925274e+00 -1.02691031e+00 -1.75711006e-01 -4.14087951e-01
-1.49681962e+00 -1.88162640e-01 5.18615842e-01 -5.88009596e-01
7.39409149e-01 2.76738822e-01 4.23718095e-01 7.36858606e-01
4.04294074e-01 -3.12984996e-02 1.56697810e+00 -3.52931947e-01
2.78511941e-01 -1.86894894e-01 2.71409094e-01 4.06229258e-01
8.91312301e-01 5.49574435e-01 -4.22683507e-02 -7.33739734e-01
1.41442215e+00 -2.19792366e-01 2.59509325e-01 -1.11784458e-01
-8.30723941e-01 1.36438966e+00 -3.35727692e-01 1.81629807e-01
-4.91952121e-01 1.94933027e-01 2.79604822e-01 6.64330006e-01
8.48220170e-01 1.98929459e-01 -9.84808862e-01 -7.44057307e-03
-1.07822490e+00 3.93142551e-01 8.80533218e-01 7.62678266e-01
2.98308402e-01 6.61915362e-01 -2.75043935e-01 6.63450122e-01
2.31939524e-01 1.15728939e+00 5.13951182e-01 -1.14457595e+00
2.25018770e-01 1.60443202e-01 4.04911101e-01 -5.87447941e-01
-3.77190620e-01 -6.39459491e-01 -7.61775792e-01 2.27327436e-01
6.78842366e-01 -5.14589727e-01 -5.59622347e-01 1.84343982e+00
-1.78222936e-02 2.24913821e-01 -2.15781540e-01 2.25630671e-01
8.79641902e-03 4.34812605e-01 9.93977413e-02 -9.44008589e-01
1.01252401e+00 -2.68089890e-01 -4.87551540e-01 -1.64841767e-03
6.89827800e-01 -3.49008828e-01 1.12901962e+00 3.88023317e-01
-1.30608141e+00 -1.86274484e-01 -6.94779456e-01 5.56496501e-01
-9.21802148e-02 -2.71317065e-01 5.85221529e-01 8.19553852e-01
-1.01481736e+00 7.43125260e-01 -5.83661556e-01 4.61038500e-02
3.83856110e-02 4.39568192e-01 1.27780423e-01 2.08165899e-01
-1.25788021e+00 6.80485010e-01 -2.88325548e-01 -6.39338344e-02
-5.69780767e-01 -9.50627029e-01 -5.90040743e-01 3.17725658e-01
1.60074130e-01 -3.21871996e-01 1.05308616e+00 -6.69175863e-01
-1.21547484e+00 3.75189990e-01 -2.40748413e-02 -4.32365954e-01
7.26617098e-01 2.01279130e-02 -6.52522981e-01 -3.03306431e-01
4.51796234e-01 -3.49729270e-01 7.59043336e-01 -6.80891633e-01
-3.53981614e-01 -4.19776231e-01 -4.45810795e-01 -3.72272551e-01
-7.09299073e-02 4.05186564e-01 4.90420967e-01 -8.62676561e-01
1.11483067e-01 -7.81407714e-01 -5.12462437e-01 -9.11556542e-01
-7.79916123e-02 -1.83287486e-01 1.92817599e-01 -9.62300301e-01
1.68189335e+00 -1.92716622e+00 -3.38018775e-01 8.47733200e-01
-2.56698728e-01 -5.22696912e-01 -1.37986630e-01 2.28295341e-01
-3.30472946e-01 5.36915541e-01 -2.58828402e-01 5.69261789e-01
2.30235428e-01 -1.21731706e-01 -3.80234212e-01 7.60413468e-01
8.85603391e-03 6.42301798e-01 -4.45119679e-01 3.63596380e-02
-1.97978631e-01 -1.19664572e-01 -6.79238796e-01 -2.50005454e-01
2.25338012e-01 1.25339851e-01 -2.34106749e-01 6.43260181e-01
6.45078897e-01 -3.77526969e-01 6.16291799e-02 4.53329295e-01
-7.49413192e-01 5.55370823e-02 -1.04008055e+00 7.37228692e-01
7.32032210e-03 2.75253683e-01 -1.35020781e-02 -1.28590441e+00
1.04470515e+00 1.71744421e-01 3.44722599e-01 -5.63237309e-01
7.90238231e-02 6.50013268e-01 4.17548209e-01 -1.46183237e-01
1.76518679e-01 -5.31352460e-01 -4.08063114e-01 2.41771162e-01
-2.68752903e-01 -2.93235540e-01 3.30434471e-01 -3.94494236e-01
1.28317022e+00 -4.84080724e-02 6.63766146e-01 -1.03564739e+00
2.29671478e-01 -2.60507345e-01 7.32489109e-01 1.03182578e+00
-8.21042061e-02 -5.61100468e-02 9.60413158e-01 1.25746533e-01
-1.20843446e+00 -1.17545843e+00 -6.39092207e-01 1.07956839e+00
-6.34981930e-01 2.70340204e-01 -1.78712711e-01 -3.73506457e-01
7.42566168e-01 9.69410300e-01 -7.32185245e-01 -3.29875164e-02
-3.72656435e-01 -1.17361283e+00 4.79507655e-01 6.40809536e-01
4.16772723e-01 -5.12209296e-01 -4.27778602e-01 1.68788448e-01
2.74015307e-01 -4.03322995e-01 -2.25981519e-01 2.01351270e-01
-1.31585610e+00 -9.01818156e-01 -9.47669625e-01 -2.42550984e-01
3.57581228e-01 -6.87982738e-02 1.03408051e+00 -3.92903239e-01
1.34344935e-01 4.25610155e-01 -7.09566334e-03 -6.13568544e-01
-1.75751418e-01 -4.32415754e-01 2.66679645e-01 -6.19940519e-01
4.75988597e-01 -7.29511023e-01 -7.10317254e-01 5.29533744e-01
-5.96914053e-01 -9.12248969e-01 3.95335346e-01 9.69567537e-01
2.55623251e-01 1.56119585e-01 1.20284486e+00 -4.97217655e-01
7.65565932e-01 -8.32246423e-01 -1.23506737e+00 1.79116055e-01
-9.54948008e-01 -7.90233687e-02 2.62878865e-01 -9.27358627e-01
-6.85415983e-01 -9.74747121e-01 4.10147697e-01 -2.48361841e-01
5.48732221e-01 1.20604503e+00 4.29956466e-01 3.23957130e-02
5.91323733e-01 -1.48316234e-01 6.19571954e-02 -4.51774210e-01
-1.65345799e-02 1.73378095e-01 3.03236634e-01 -5.59733272e-01
9.02470112e-01 1.93049461e-02 5.56946397e-01 -7.42564499e-01
-3.38089138e-01 -7.23782927e-02 -2.99539030e-01 7.11064488e-02
3.95127505e-01 -1.12403917e+00 -8.39597881e-01 1.16323873e-01
-4.34831560e-01 -7.28367031e-01 -5.51989913e-01 1.13510370e+00
-7.05690444e-01 8.71178508e-02 -5.64260960e-01 -1.24774528e+00
1.11432448e-01 -6.53067827e-01 3.59109730e-01 4.32873182e-02
-2.23849446e-01 -1.03979266e+00 2.72738695e-01 -3.55259955e-01
4.94342387e-01 5.69396555e-01 1.12037420e+00 -9.59070981e-01
-1.45020336e-02 -4.38290626e-01 -5.71407378e-02 2.90077627e-01
-1.56559080e-01 4.16579276e-01 -3.08841914e-01 -4.06478673e-01
2.75617599e-01 4.06909585e-02 4.72788334e-01 1.32870150e+00
9.59834874e-01 -4.91480142e-01 1.59704193e-01 3.47352028e-01
1.34970331e+00 5.05925119e-01 5.33604443e-01 2.01669008e-01
-1.57356635e-02 4.93907571e-01 6.00417912e-01 1.14646280e+00
6.15864359e-02 4.39072698e-02 -9.22999159e-02 -4.54028361e-02
6.40844285e-01 -1.58164129e-01 6.04849935e-01 9.06673968e-01
-4.26113486e-01 1.05974838e-01 -7.89862394e-01 3.92606705e-01
-1.50924146e+00 -1.17838991e+00 -4.06356364e-01 2.70117283e+00
7.30396628e-01 1.46658033e-01 6.04704618e-01 1.02553323e-01
7.22715318e-01 -8.93038735e-02 -7.33638227e-01 -4.65176046e-01
-3.24470937e-01 4.94230628e-01 1.17253077e+00 4.76113677e-01
-6.95960343e-01 4.20410246e-01 8.33943939e+00 6.52356267e-01
-6.04655266e-01 1.12704888e-01 8.14800620e-01 -6.31265119e-02
-5.05291402e-01 3.37090611e-01 -6.79080486e-01 4.22922701e-01
1.56199431e+00 -9.33649838e-01 1.18197352e-01 8.17227721e-01
7.96340466e-01 -4.16582376e-01 -6.01805568e-01 3.26100916e-01
-3.46106708e-01 -6.37736201e-01 -4.77573663e-01 5.38089514e-01
9.12819803e-01 -9.22839940e-02 3.48581165e-01 6.62450016e-01
8.77388358e-01 -9.51390564e-01 4.62750435e-01 4.12975878e-01
1.02168357e+00 -9.72414613e-01 5.87778389e-01 2.51647174e-01
-8.60389650e-01 -5.62620401e-01 -5.73634267e-01 -6.54608250e-01
-3.68708037e-02 9.73641753e-01 -2.21480593e-01 2.69761831e-01
4.35851216e-01 3.22594523e-01 -3.18777084e-01 9.32956576e-01
-4.33719531e-02 9.38464046e-01 -3.62950504e-01 1.25121415e-01
3.23097073e-02 -7.20815063e-01 4.09000844e-01 7.27480471e-01
1.33681595e+00 4.31772888e-01 -2.33418778e-01 6.65015042e-01
2.76540935e-01 2.41121382e-01 -8.59561801e-01 -1.09024778e-01
6.18972182e-01 5.33874750e-01 -6.98218644e-01 -3.41560870e-01
-9.50996518e-01 1.83009729e-01 -3.40275526e-01 6.36364222e-01
-8.37521911e-01 -1.67805105e-01 3.07499111e-01 2.67465591e-01
2.87087470e-01 -4.41860616e-01 -4.09456521e-01 -1.18627620e+00
-3.33432704e-01 -8.15389931e-01 7.18270004e-01 -5.57535052e-01
-1.27085376e+00 -4.40481097e-01 5.95939517e-01 -1.01551795e+00
-6.36172950e-01 -6.54258668e-01 -6.61739588e-01 9.67913806e-01
-1.13750577e+00 -4.91035223e-01 5.20141721e-01 4.55250591e-01
1.91283554e-01 -1.44890472e-01 5.51385701e-01 3.46539170e-02
-6.05395973e-01 4.63455796e-01 6.36449754e-01 -2.87440807e-01
6.55168474e-01 -1.07741964e+00 1.64258648e-02 8.92771363e-01
-6.76269650e-01 8.40190887e-01 1.03311503e+00 -1.07177401e+00
-1.08381212e+00 -6.39913619e-01 7.20043957e-01 -2.18715638e-01
1.36632276e+00 4.19602431e-02 -7.73962736e-01 8.72002482e-01
-1.29015654e-01 -2.23741811e-02 9.19425845e-01 5.40819824e-01
-3.48817945e-01 -2.18948901e-01 -1.23272479e+00 5.87494850e-01
7.57025898e-01 -3.16116780e-01 -3.36651504e-01 1.03116676e-01
4.98930871e-01 3.79123807e-01 -1.42361152e+00 9.31818187e-01
9.03727710e-01 -9.19879496e-01 7.74320364e-01 -6.21794283e-01
2.40463555e-01 2.47823089e-01 -2.68218040e-01 -1.15445340e+00
-5.72958708e-01 -7.72218347e-01 1.69341773e-01 1.13514161e+00
5.48656702e-01 -1.00800014e+00 2.98724771e-01 7.89926827e-01
2.04913750e-01 -9.23709124e-02 -1.11807644e+00 -1.30161834e+00
6.86335862e-01 -3.09944034e-01 5.36307633e-01 1.24482381e+00
2.90407658e-01 -2.05389202e-01 -4.82432604e-01 3.66942808e-02
8.52620721e-01 -3.15024368e-02 6.46521986e-01 -1.61927342e+00
-4.99666631e-01 -3.63458097e-01 -5.93749620e-02 -5.90851486e-01
1.93191662e-01 -4.46984172e-01 -1.97252944e-01 -5.97657204e-01
3.73439729e-01 -1.68955684e-01 -3.81073087e-01 2.51473766e-02
3.40838768e-02 -8.09065998e-02 -3.12004715e-01 -2.80804671e-02
2.29185551e-01 3.01973850e-01 8.92273366e-01 4.23965603e-01
-3.71399254e-01 2.44752333e-01 -8.15195382e-01 8.64809632e-01
9.16652203e-01 -5.11872768e-01 -4.37920392e-01 2.57507473e-01
4.56548810e-01 8.42018664e-01 5.47446549e-01 -4.41477209e-01
-1.68547571e-01 -6.18807852e-01 4.64102536e-01 -5.76452196e-01
-4.10239458e-01 -5.41556895e-01 3.94677460e-01 6.95484161e-01
-3.20657164e-01 7.82850981e-01 2.54197568e-01 3.61119658e-01
4.28047590e-02 -1.65940866e-01 6.40197277e-01 1.31230012e-01
6.02855384e-02 4.55498062e-02 -3.81581485e-01 1.47738248e-01
7.98190773e-01 1.51209325e-01 -5.09602726e-01 -6.06528759e-01
-5.89057088e-01 3.35472435e-01 2.41888940e-01 -1.26396313e-01
1.84839219e-01 -1.46566701e+00 -9.44615841e-01 2.98606336e-01
-3.56310338e-01 -8.34248424e-01 1.32497147e-01 1.19667029e+00
-3.31409462e-02 3.27761412e-01 6.64357021e-02 -4.05509025e-02
-8.33333015e-01 4.85686958e-01 2.01233849e-02 -3.93936455e-01
-2.29038507e-01 4.54025745e-01 3.28701884e-01 7.56081492e-02
-3.27246904e-01 -5.30803442e-01 1.01590991e-01 1.64829761e-01
3.99966687e-01 1.00437689e+00 -5.89488983e-01 -2.75543422e-01
-2.02132747e-01 5.31372070e-01 5.22358239e-01 -5.44181287e-01
1.44134462e+00 -2.81847775e-01 -2.42568478e-01 9.86317158e-01
7.99247682e-01 3.17917049e-01 -1.31617665e+00 4.62320112e-02
3.70148569e-01 -4.05015737e-01 -3.19538057e-01 -5.43478370e-01
-4.89255428e-01 2.34017313e-01 3.40256095e-01 6.30226254e-01
9.60709810e-01 -3.04980516e-01 -2.77702719e-01 1.45562381e-01
3.25002611e-01 -1.17550945e+00 -2.78087974e-01 5.24493754e-01
9.34333563e-01 -1.10435975e+00 4.14666384e-01 -6.37522191e-02
-2.84586966e-01 8.23194742e-01 5.86302318e-02 -6.49070919e-01
1.14972186e+00 4.04343039e-01 -4.97435749e-01 5.64070605e-02
-7.01168180e-01 -1.61544420e-02 5.77830411e-02 2.42451712e-01
5.68003953e-01 2.42627546e-01 -1.40481496e+00 9.74212229e-01
-2.20689639e-01 5.81781603e-02 6.81777537e-01 5.62901855e-01
-6.23093903e-01 -7.38598883e-01 -5.56873918e-01 9.19474006e-01
-9.02170360e-01 -2.52464294e-01 2.50330895e-01 1.51620567e+00
-6.34751320e-01 1.04941416e+00 3.33599895e-01 -7.54426494e-02
2.66248673e-01 4.70924050e-01 1.50578395e-01 -5.19396923e-02
-2.33761102e-01 8.72138143e-01 1.29045501e-01 -3.78196806e-01
-5.68455994e-01 -1.15821052e+00 -6.53134227e-01 -6.79681301e-01
-4.55058306e-01 4.03403103e-01 2.74026692e-01 7.52547085e-01
-1.14378519e-01 1.91597953e-01 1.01148438e+00 -4.97400373e-01
-1.26805139e+00 -1.26578319e+00 -1.45447600e+00 -8.28852654e-02
1.85587198e-01 -7.46556759e-01 -1.20336783e+00 -4.02791411e-01] | [6.3974289894104, 4.217240810394287] |
649c89cf-d62a-4951-b667-4f25e9de6fc3 | table-to-text-generation-by-structure-aware | 1711.09724 | null | http://arxiv.org/abs/1711.09724v1 | http://arxiv.org/pdf/1711.09724v1.pdf | Table-to-text Generation by Structure-aware Seq2seq Learning | Table-to-text generation aims to generate a description for a factual table
which can be viewed as a set of field-value records. To encode both the content
and the structure of a table, we propose a novel structure-aware seq2seq
architecture which consists of field-gating encoder and description generator
with dual attention. In the encoding phase, we update the cell memory of the
LSTM unit by a field gate and its corresponding field value in order to
incorporate field information into table representation. In the decoding phase,
dual attention mechanism which contains word level attention and field level
attention is proposed to model the semantic relevance between the generated
description and the table. We conduct experiments on the \texttt{WIKIBIO}
dataset which contains over 700k biographies and corresponding infoboxes from
Wikipedia. The attention visualizations and case studies show that our model is
capable of generating coherent and informative descriptions based on the
comprehensive understanding of both the content and the structure of a table.
Automatic evaluations also show our model outperforms the baselines by a great
margin. Code for this work is available on
https://github.com/tyliupku/wiki2bio. | ['Lei Sha', 'Baobao Chang', 'Zhifang Sui', 'Tianyu Liu', 'Kexiang Wang'] | 2017-11-27 | null | null | null | null | ['table-to-text-generation'] | ['natural-language-processing'] | [ 3.02357674e-01 4.32438672e-01 -1.97919935e-01 -4.16679978e-01
-9.70195472e-01 -6.33374751e-01 5.71669579e-01 4.47582334e-01
-4.65610325e-02 1.34806526e+00 9.39086854e-01 -2.30383910e-02
2.48646840e-01 -1.08825529e+00 -1.02813804e+00 -4.61436898e-01
2.20868662e-01 5.62648058e-01 -1.54362589e-01 -4.92948413e-01
4.80940938e-01 -1.50971755e-01 -1.29759574e+00 7.99568355e-01
8.96261394e-01 7.89677918e-01 6.29668474e-01 6.82518184e-01
-5.20031810e-01 7.79713094e-01 -8.74457717e-01 -6.22364581e-01
-5.87685853e-02 -7.85460353e-01 -9.39808428e-01 -1.62506357e-01
3.26278806e-01 -2.42630661e-01 -4.70387727e-01 9.91566718e-01
6.83874965e-01 1.98294342e-01 4.50743496e-01 -8.85296941e-01
-9.23134148e-01 1.16684699e+00 -1.08502418e-01 1.88671380e-01
4.36272502e-01 1.95181444e-01 1.27849162e+00 -5.35857856e-01
9.72969115e-01 1.12956977e+00 1.70019791e-01 6.93183839e-01
-8.14326406e-01 -6.01487458e-01 1.88383162e-01 3.06576908e-01
-1.22766244e+00 -3.50995749e-01 3.31852049e-01 -1.99563816e-01
1.23138094e+00 2.17348561e-01 7.42609680e-01 1.08961225e+00
5.80346704e-01 6.65447712e-01 4.09311205e-01 -1.13666773e-01
1.21604860e-01 1.96849462e-02 -8.82360637e-02 7.85551727e-01
4.19093519e-01 -5.21209419e-01 -8.68088722e-01 1.21859394e-01
8.04852068e-01 -3.08145702e-01 -2.38322333e-01 2.45535225e-01
-1.50783169e+00 6.75122142e-01 5.48260689e-01 1.50203556e-01
-5.37634134e-01 4.18849736e-01 6.18749261e-01 -1.30518660e-01
1.58678472e-01 6.59894228e-01 -2.84557194e-01 -6.31923899e-02
-4.37848806e-01 4.05912042e-01 8.47802103e-01 1.43759823e+00
5.00054717e-01 -4.11040410e-02 -1.00469863e+00 4.71101671e-01
1.91436266e-03 2.56324351e-01 5.80653787e-01 -7.07062840e-01
8.30356300e-01 7.88876235e-01 1.67889446e-01 -6.47218227e-01
-9.03976634e-02 -4.49580908e-01 -1.08920479e+00 -5.46798229e-01
5.72238024e-03 -2.25780249e-01 -9.98997211e-01 1.74594653e+00
7.28980675e-02 -9.73072052e-02 2.66866595e-01 5.98244667e-01
1.44168591e+00 1.02595639e+00 1.08473554e-01 1.07308999e-01
1.63273978e+00 -9.30511117e-01 -1.24707758e+00 -1.55692175e-01
6.43613815e-01 -4.61837530e-01 9.00300503e-01 -2.98050791e-02
-1.34487891e+00 -5.98999321e-01 -1.14289606e+00 -6.13599956e-01
-7.01363325e-01 2.33728990e-01 6.01785123e-01 -3.02750766e-02
-8.85087013e-01 4.99205083e-01 -4.03369755e-01 -2.52771407e-01
4.70600605e-01 2.54697442e-01 -3.76276046e-01 -1.36767020e-02
-1.63072741e+00 5.58773041e-01 8.56420755e-01 2.20458150e-01
-8.47221136e-01 -6.70338631e-01 -1.12833619e+00 6.00041926e-01
3.44141841e-01 -1.01704657e+00 9.41316962e-01 -2.35969737e-01
-1.04768932e+00 6.77277267e-01 -3.41039568e-01 -5.71566403e-01
4.01196778e-02 9.69768241e-02 -1.93907902e-01 -2.37199329e-02
3.50741655e-01 9.87754107e-01 3.31045240e-02 -7.52895892e-01
-5.61512172e-01 -1.12210378e-01 1.50858626e-01 3.88975531e-01
1.30670756e-01 -3.50052059e-01 -7.95703530e-01 -7.87520528e-01
-2.78933257e-01 -4.75085646e-01 -1.63412333e-01 -4.39028263e-01
-9.50053632e-01 -1.89998776e-01 -2.48516742e-02 -1.04106569e+00
1.58462584e+00 -1.71053886e+00 7.30556771e-02 -7.80427456e-02
9.52912793e-02 5.50468974e-02 2.35420410e-02 8.49583387e-01
4.69993800e-02 4.97618943e-01 -5.26002526e-01 6.06106874e-03
7.46483654e-02 5.97632257e-04 -3.72363687e-01 -1.18480176e-01
4.16761905e-01 1.23442996e+00 -1.04855418e+00 -3.67917299e-01
-3.10762584e-01 5.55595577e-01 -6.26251578e-01 3.04477543e-01
-7.70002007e-01 2.96635419e-01 -3.97891730e-01 3.19712758e-01
3.47190142e-01 -6.11481190e-01 3.98782492e-01 -2.73896903e-01
1.05234934e-02 8.27495396e-01 -8.58976424e-01 1.97707403e+00
-3.25675368e-01 4.51175749e-01 -3.06397378e-01 -3.77113372e-01
1.00376570e+00 4.03132707e-01 -9.98982191e-02 -8.27517509e-01
3.90548185e-02 5.36255091e-02 7.83721954e-02 -2.41297618e-01
7.24653304e-01 1.97831020e-01 -4.31506485e-01 3.85486573e-01
1.80181697e-01 8.85174870e-02 8.17478657e-01 5.79089284e-01
1.07437527e+00 3.36720258e-01 3.97695392e-01 -3.88906777e-01
7.02232242e-01 5.56030497e-02 6.42598152e-01 6.64125562e-01
5.25127172e-01 5.94580054e-01 9.05007660e-01 -3.91122997e-01
-1.10624146e+00 -6.24855995e-01 1.32462949e-01 8.22148561e-01
1.54075529e-02 -8.62542331e-01 -7.98619628e-01 -4.93166536e-01
-1.17099874e-01 9.81175959e-01 -1.05643129e+00 -3.01025271e-01
-6.06307626e-01 -7.04792202e-01 4.47579920e-01 7.54549265e-01
7.83648670e-01 -1.39775300e+00 -2.85073310e-01 3.55929494e-01
-5.98131478e-01 -8.51709843e-01 -9.26551402e-01 3.83521855e-01
-4.18656230e-01 -7.65060008e-01 -4.90692765e-01 -8.67278934e-01
7.27506876e-01 -4.99644756e-01 1.26121211e+00 3.39643583e-02
-1.27914011e-01 -3.97029102e-01 -2.22775295e-01 -4.87859815e-01
-5.07158995e-01 4.46055382e-01 -6.17817342e-01 -1.26916841e-01
1.16811804e-01 -1.81807891e-01 -5.34155428e-01 -1.66925505e-01
-9.48367715e-01 6.43879354e-01 6.60896659e-01 9.81015980e-01
8.74437511e-01 -2.91515172e-01 7.76137054e-01 -1.21775675e+00
7.81927347e-01 -7.18826711e-01 -6.15862787e-01 3.89202714e-01
-4.33614582e-01 7.37130165e-01 8.74388933e-01 2.11854488e-01
-1.21029401e+00 -7.14894632e-05 -2.95563430e-01 2.10999832e-01
2.22589925e-01 7.17707336e-01 -6.38797641e-01 8.83327603e-01
2.96957850e-01 6.13454878e-01 -1.78038120e-01 -4.37165767e-01
2.56852806e-01 4.85692620e-01 7.84213722e-01 -3.36111814e-01
3.26846302e-01 1.02148943e-01 -4.48724702e-02 -1.87620521e-01
-8.87565434e-01 -1.48039814e-02 -5.87382615e-01 7.60147870e-02
1.03308260e+00 -9.49878275e-01 -7.89304852e-01 9.36626196e-02
-1.32751203e+00 -3.94360721e-01 -1.63015261e-01 1.12106830e-01
-5.53353190e-01 -2.00068846e-01 -6.78161621e-01 -3.81230593e-01
-7.65325904e-01 -9.34420705e-01 1.00350976e+00 2.74668396e-01
-8.92979130e-02 -1.03814054e+00 -1.29982159e-01 1.80037498e-01
2.28012145e-01 3.95998746e-01 1.25722206e+00 -7.68711984e-01
-8.49163771e-01 2.12476198e-02 -1.57508731e-01 -1.06486782e-01
-1.18053323e-02 -1.77927330e-01 -7.24625587e-01 1.36568815e-01
-6.06671035e-01 -2.94934779e-01 1.13837242e+00 1.12096950e-01
1.45637739e+00 -7.56592155e-01 -2.56973594e-01 6.21589541e-01
1.54896200e+00 4.19810981e-01 9.42778528e-01 2.39722192e-01
8.00614297e-01 5.73869586e-01 4.98388380e-01 6.44137621e-01
6.42788947e-01 5.26427567e-01 3.16962153e-01 1.98312446e-01
-4.37706888e-01 -7.44450569e-01 2.02461272e-01 7.13462770e-01
3.59901279e-01 -9.53809679e-01 -8.32635641e-01 6.76331043e-01
-1.70195973e+00 -1.12467635e+00 -1.97014306e-02 2.12639403e+00
1.07172620e+00 1.86581939e-01 -2.68363267e-01 -2.06695452e-01
9.02265251e-01 5.59976213e-02 -5.08235574e-01 -6.74566269e-01
-9.97933969e-02 8.87045544e-03 4.10091192e-01 5.65154731e-01
-5.81444860e-01 1.01428127e+00 5.49535799e+00 7.24736512e-01
-8.01143467e-01 -3.36744189e-01 9.85698998e-01 -1.29610345e-01
-7.43536532e-01 -3.01638365e-01 -1.23584628e+00 6.23000801e-01
1.15931892e+00 -6.33501768e-01 3.03707153e-01 3.25117528e-01
1.38922319e-01 2.14487892e-02 -1.08263731e+00 6.59439325e-01
1.10380195e-01 -1.92432582e+00 6.59527004e-01 -1.69979315e-02
5.70844591e-01 -2.68435985e-01 -1.54323027e-01 4.16412175e-01
2.85620719e-01 -1.28693438e+00 6.74726188e-01 7.65764236e-01
1.10646081e+00 -7.64030874e-01 9.49967563e-01 -7.68828671e-03
-1.24286222e+00 1.86640605e-01 -3.44865799e-01 1.45488694e-01
1.59897149e-01 3.45603257e-01 -1.10570276e+00 5.62007546e-01
5.05541980e-01 8.01677585e-01 -5.53546309e-01 7.85793841e-01
-4.30502802e-01 3.26552272e-01 1.55723125e-01 -2.77182937e-01
3.01410139e-01 -2.75878757e-02 1.81773394e-01 1.38382912e+00
2.69371361e-01 2.11902380e-01 -1.43582091e-01 1.22074056e+00
-6.61923587e-01 1.05553724e-01 -4.44823742e-01 -3.89018536e-01
5.62121868e-01 1.28557038e+00 -6.68942928e-01 -8.02441239e-01
-1.39231667e-01 8.45566154e-01 3.38227332e-01 7.91123584e-02
-8.39642644e-01 -8.48174036e-01 4.25671458e-01 7.14560598e-02
5.94612300e-01 2.90929466e-01 -4.91066873e-01 -1.03499877e+00
1.25253245e-01 -7.12637246e-01 5.29124498e-01 -1.05859697e+00
-7.14227080e-01 9.37725127e-01 -2.91459095e-02 -8.03105533e-01
-5.16425610e-01 -8.91838297e-02 -6.24852240e-01 1.29361784e+00
-1.20219755e+00 -9.46556270e-01 -4.48999226e-01 1.04912564e-01
6.70940340e-01 -9.40535218e-02 8.65397871e-01 1.19324237e-01
-7.26319730e-01 5.50224900e-01 -9.05905757e-03 3.92894834e-01
6.27835214e-01 -1.46262443e+00 1.14197421e+00 7.03381956e-01
-2.41899431e-01 8.90259564e-01 5.34096122e-01 -1.08596635e+00
-1.20298624e+00 -1.39164507e+00 1.43946159e+00 -2.28653908e-01
2.46513858e-01 -9.08884406e-01 -1.05285311e+00 7.57569134e-01
5.86706817e-01 -4.04586792e-01 6.81393206e-01 -4.05051976e-01
-1.84532195e-01 -7.47781917e-02 -9.82332587e-01 6.07819319e-01
1.05500984e+00 -1.81327075e-01 -3.57299984e-01 3.96764666e-01
1.05228949e+00 -6.72867894e-01 -1.02175748e+00 6.94292411e-02
4.81150150e-01 -5.19831181e-01 5.22218406e-01 -6.01868808e-01
9.37299430e-01 -3.41562301e-01 4.53894548e-02 -1.26955938e+00
-5.95110476e-01 -5.97227812e-01 -1.75225660e-01 1.36923826e+00
7.92904198e-01 -9.87251401e-02 6.29363298e-01 2.71465898e-01
-4.68550146e-01 -1.01480508e+00 -4.36223447e-01 -2.99045026e-01
-1.09514400e-01 2.16743380e-01 9.55478549e-01 5.14067590e-01
6.49157837e-02 8.22287798e-01 -3.43923986e-01 -2.65171736e-01
2.39403069e-01 1.74548984e-01 3.08597982e-01 -7.20027804e-01
-1.41886249e-01 -2.67155468e-01 1.20881267e-01 -9.15598989e-01
6.31381050e-02 -1.30693102e+00 1.90253302e-01 -2.26814198e+00
4.30556238e-01 8.74196067e-02 -8.01522657e-02 5.84990203e-01
-4.69990999e-01 2.00226717e-02 1.03150949e-01 -2.14053378e-01
-6.90506339e-01 6.09039545e-01 1.44491291e+00 -1.53087407e-01
-5.26645631e-02 -5.61303675e-01 -1.11538732e+00 -9.82665420e-02
9.84759271e-01 -3.54464382e-01 -4.29875493e-01 -5.65947115e-01
4.39546198e-01 4.13789809e-01 2.47236919e-02 -9.42476630e-01
2.34531105e-01 -3.93709540e-02 6.50679469e-01 -7.05744505e-01
1.73424274e-01 -9.75380093e-02 1.49537683e-01 6.82838321e-01
-1.06987190e+00 4.15510327e-01 4.27540272e-01 4.43318635e-01
-1.50996625e-01 -1.34349585e-01 3.08778375e-01 -6.76021636e-01
-7.18932211e-01 4.88734514e-01 -4.59254026e-01 3.45829129e-01
4.70370919e-01 1.95890084e-01 -6.30803347e-01 -4.81955230e-01
-4.05703396e-01 6.22318923e-01 2.02654272e-01 4.37763482e-01
6.15133941e-01 -1.34893930e+00 -9.94306743e-01 1.72669992e-01
3.25125784e-01 1.95272967e-01 2.75222957e-01 1.84946582e-01
-6.59689367e-01 9.68526006e-01 -6.27916336e-01 1.41158894e-01
-8.29240501e-01 5.19160569e-01 9.59639996e-02 -5.78392327e-01
-2.95697033e-01 8.90851140e-01 4.28316414e-01 -2.39427388e-01
1.92057729e-01 -4.57257181e-01 -4.57151264e-01 1.06947608e-02
9.11485970e-01 9.92057621e-02 1.71511397e-01 -3.49945188e-01
-2.16305241e-01 -1.25165090e-01 -3.61764312e-01 2.60789841e-02
1.25977945e+00 3.26099098e-02 -1.92421779e-01 3.43361914e-01
1.07984972e+00 -1.09460853e-01 -1.02827740e+00 1.70686871e-01
-1.23753592e-01 -1.87116843e-02 -1.72820032e-01 -1.37012494e+00
-9.10724819e-01 7.24780679e-01 -1.06528603e-01 -3.32881480e-01
9.63646352e-01 -1.00929685e-01 9.81992304e-01 4.22147691e-01
-7.95280188e-02 -7.85660982e-01 5.56394979e-02 6.55025840e-01
1.19301295e+00 -9.56590235e-01 -2.29804114e-01 -4.14084762e-01
-8.24962676e-01 9.86185431e-01 9.48843777e-01 1.43934071e-01
1.58307850e-02 1.82093352e-01 -3.17499250e-01 -1.11700840e-01
-1.45036042e+00 -2.84311891e-01 3.01431686e-01 4.37983543e-01
8.20766866e-01 -2.38472775e-01 -4.64756638e-01 8.10845733e-01
-3.69522691e-01 -6.49695024e-02 7.52724946e-01 6.80436909e-01
-3.81889850e-01 -1.07763863e+00 3.81304021e-03 7.52664387e-01
-4.76866513e-01 -6.47646070e-01 -3.91900420e-01 4.39433575e-01
1.20204404e-01 5.25053144e-01 2.89882302e-01 -4.07446295e-01
1.68909445e-01 2.39048064e-01 1.84012130e-01 -9.77575123e-01
-1.00287104e+00 -7.79665560e-02 3.66646111e-01 -4.56119955e-01
1.71977267e-01 -3.74592960e-01 -1.69699240e+00 -3.97939920e-01
2.73307502e-01 4.61778283e-01 4.13954169e-01 4.41716164e-01
6.78643644e-01 1.03780746e+00 1.42974118e-02 -1.73781097e-01
-1.74879208e-01 -1.07831109e+00 -4.15061384e-01 3.18744481e-01
1.23624638e-01 -2.09881768e-01 1.77024558e-01 2.56938994e-01] | [11.608440399169922, 8.776432037353516] |
b0b79a40-40c2-49f0-99cc-1567572366fa | bunji-at-semeval-2017-task-3-combination-of | null | null | https://aclanthology.org/S17-2058 | https://aclanthology.org/S17-2058.pdf | bunji at SemEval-2017 Task 3: Combination of Neural Similarity Features and Comment Plausibility Features | This paper describes a text-ranking system developed by bunji team in SemEval-2017 Task 3: Community Question Answering, Subtask A and C. The goal of the task is to re-rank the comments in a question-and-answer forum such that useful comments for answering the question are ranked high. We proposed a method that combines neural similarity features and hand-crafted comment plausibility features, and we modeled inter-comments relationship using conditional random field. Our approach obtained the fifth place in the Subtask A and the second place in the Subtask C. | ['Yuta Koreeda', 'Takuya Hashito', 'Yoshiki Niwa', 'Misa Sato', 'Kohsuke Yanai', 'Toshihiko Yanase', 'Kenzo Kurotsuchi'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['question-similarity'] | ['natural-language-processing'] | [-2.26177365e-01 5.49308062e-01 3.60754952e-02 -5.75038373e-01
-1.17820263e+00 -5.00288129e-01 9.56063151e-01 9.80946600e-01
-5.34941077e-01 9.08814609e-01 9.91742551e-01 -3.22402447e-01
-2.95175523e-01 -3.76654148e-01 -4.17083859e-01 1.47687152e-01
1.85443938e-01 6.70018017e-01 6.98265016e-01 -3.81129086e-01
8.75393987e-01 -5.79822958e-01 -1.23857260e+00 1.18735695e+00
1.15449691e+00 1.06222630e+00 1.84672654e-01 1.20685887e+00
-4.28561330e-01 1.57614696e+00 -6.37155235e-01 -9.03464317e-01
-3.66999060e-01 -2.24537179e-01 -1.81940520e+00 -4.41227436e-01
8.58034551e-01 2.58592833e-02 -8.63976553e-02 6.76753461e-01
1.16853394e-01 3.97550672e-01 1.02441490e+00 -1.04513323e+00
-1.16712952e+00 1.13077986e+00 7.96482563e-02 5.53997457e-01
8.24030399e-01 -4.36331749e-01 1.99513018e+00 -1.15886819e+00
5.48752069e-01 1.35711288e+00 6.49214864e-01 4.64488447e-01
-8.81310701e-01 -3.59625995e-01 1.50732502e-01 7.21677005e-01
-8.34841311e-01 3.10139135e-02 6.05539858e-01 -9.96376336e-01
1.03291011e+00 3.07685405e-01 5.41789122e-02 9.76575196e-01
3.05256564e-02 8.26817751e-01 9.53613877e-01 -3.09910327e-01
1.36537790e-01 1.42823800e-01 1.08314157e+00 2.64019370e-01
-9.21963006e-02 -6.60887122e-01 -4.63683754e-01 -8.54566157e-01
-3.99369210e-01 -1.26521856e-01 -2.60831892e-01 4.59838331e-01
-9.34630036e-01 1.17000735e+00 7.52916873e-01 2.81773567e-01
-3.64145100e-01 -2.22233888e-02 3.46817493e-01 1.88729644e-01
6.85445905e-01 9.06149328e-01 -6.62280679e-01 4.83124182e-02
-9.37869787e-01 7.40284741e-01 1.08387899e+00 6.21885717e-01
7.85385013e-01 -1.03495526e+00 -1.02526164e+00 1.01615691e+00
5.28159618e-01 7.26515707e-03 1.82621047e-01 -1.12913847e+00
7.49722660e-01 7.84850717e-01 3.45962107e-01 -1.07116365e+00
-3.44386637e-01 -3.10038835e-01 -5.65740228e-01 -5.56658685e-01
3.23309958e-01 -2.71018326e-01 -3.54611635e-01 1.16879654e+00
-1.14392415e-01 -1.15025705e-02 3.57989930e-02 5.53185880e-01
1.63027930e+00 6.78552806e-01 2.36838624e-01 1.55709773e-01
1.27064121e+00 -1.50655210e+00 -6.12562895e-01 1.68578878e-01
4.44991887e-01 -1.04127181e+00 9.97006893e-01 5.86787239e-02
-8.89485061e-01 -5.81962764e-01 -6.48960173e-01 -4.31763500e-01
-2.17875987e-01 8.67016092e-02 7.17481747e-02 -4.56154682e-02
-1.27899146e+00 5.80277920e-01 1.02597177e-01 -3.29052061e-01
2.18704790e-01 -1.20918691e-01 -1.84665732e-02 7.86034465e-02
-1.63419962e+00 8.39219034e-01 2.87969828e-01 -2.35295519e-01
-8.09306324e-01 -8.08728755e-01 -3.81179899e-01 2.23889053e-01
-5.59854470e-02 -9.23590600e-01 1.73429775e+00 -7.81107485e-01
-1.25833464e+00 9.48125005e-01 -3.64267439e-01 -4.59850311e-01
1.05302989e-01 -3.66435468e-01 -5.67004681e-01 3.45358968e-01
4.45352167e-01 4.26475972e-01 6.61998987e-01 -1.08510458e+00
-8.58515084e-01 -2.72623897e-02 5.17358303e-01 1.26495361e-02
-3.28316629e-01 5.04809082e-01 -8.22106153e-02 -3.33495319e-01
-3.42360944e-01 -5.71902752e-01 -1.01377383e-01 -5.23356318e-01
-3.95877421e-01 -1.36012912e+00 4.41519469e-01 -1.00153303e+00
1.53798807e+00 -1.45216537e+00 -1.53872862e-01 1.57471746e-02
7.20912993e-01 1.34815145e-02 -2.48106897e-01 8.59743237e-01
-8.12366256e-04 5.24196029e-01 1.16234645e-01 -3.19450915e-01
-1.62386373e-01 -3.38280231e-01 -6.76193357e-01 -2.11910635e-01
2.17243686e-01 9.54664171e-01 -1.21945298e+00 -6.22376382e-01
-7.52864301e-01 -1.68153629e-01 -6.84871912e-01 5.80705941e-01
-5.70353270e-01 1.44894794e-01 -6.81229711e-01 1.63261443e-01
3.22269022e-01 -9.63878572e-01 -4.26826060e-01 5.82017414e-02
-3.13322954e-02 9.69972253e-01 -4.25012738e-01 1.15925419e+00
-3.88973475e-01 8.39499354e-01 -1.45421084e-02 -5.11485755e-01
9.84740555e-01 4.38490212e-01 1.80643611e-02 -3.87098163e-01
2.66288947e-02 2.20933557e-01 -2.25525990e-01 -8.43050122e-01
8.89192581e-01 3.00293297e-01 1.97930783e-02 6.11662984e-01
-5.10096131e-03 -2.26416454e-01 3.44482034e-01 1.02206957e+00
1.29059565e+00 -3.28391016e-01 1.31370485e-01 -6.12583339e-01
9.66673970e-01 2.42430195e-02 -1.09748214e-01 1.04771936e+00
-8.66109282e-02 8.55896533e-01 6.44273937e-01 -3.01964074e-01
-7.94133008e-01 -7.49273837e-01 5.67945912e-02 1.58941245e+00
-7.53295496e-02 -8.44853878e-01 -2.77818620e-01 -1.09764802e+00
1.75517038e-01 9.92735624e-01 -1.11757112e+00 1.20945677e-01
-3.59356850e-01 -1.23821804e-02 5.14563501e-01 2.96192259e-01
2.44800538e-01 -7.83267736e-01 -2.92283017e-02 1.71225905e-01
-9.30502474e-01 -7.72666633e-01 -7.88851738e-01 1.15972131e-01
-6.14005208e-01 -1.20056355e+00 -9.12478030e-01 -1.09945726e+00
4.40540493e-01 1.30324334e-01 1.76541197e+00 4.66658145e-01
3.45678240e-01 3.88990909e-01 -8.32162440e-01 -2.95278996e-01
-3.30171853e-01 1.82767913e-01 -4.00944293e-01 -5.87759912e-02
2.74835080e-01 -1.44078702e-01 -7.16815054e-01 2.88836956e-01
-5.32458782e-01 -3.48545223e-01 -1.31520964e-02 8.99520397e-01
3.44404094e-02 -5.31252921e-01 1.04171073e+00 -8.57148230e-01
1.40141904e+00 -9.27075326e-01 9.43008885e-02 5.60217977e-01
-5.48000276e-01 1.32897601e-01 7.34000802e-01 -2.53891766e-01
-1.14174378e+00 -5.48300982e-01 -4.89173383e-01 3.81062925e-01
3.49019617e-01 8.27395678e-01 2.63412982e-01 5.98889410e-01
1.14105737e+00 -4.02537376e-01 -7.07806706e-01 -6.97777510e-01
4.29644853e-01 1.00031698e+00 3.48948747e-01 -7.34554172e-01
5.80954075e-01 -4.06153612e-02 -6.57037556e-01 -4.88319069e-01
-1.77333522e+00 -1.24557519e+00 -3.21597904e-01 -4.84976381e-01
1.01849329e+00 -8.85958314e-01 -8.74241829e-01 2.50609696e-01
-1.71428382e+00 1.03228703e-01 -2.97553223e-02 9.33535919e-02
-2.73603909e-02 5.00004888e-01 -6.94594979e-01 -6.60939753e-01
-7.04658628e-01 -4.16038513e-01 8.46067548e-01 3.96851122e-01
-5.50791264e-01 -1.02781236e+00 5.16291022e-01 1.15949726e+00
5.75346112e-01 -2.49713048e-01 1.09724414e+00 -1.27751434e+00
-9.08352658e-02 -8.36233571e-02 -4.70879048e-01 3.57963234e-01
-3.16290081e-01 1.67341828e-01 -9.94214594e-01 7.90936500e-02
-2.94448376e-01 -8.68445873e-01 1.36564636e+00 2.07397744e-01
1.30850840e+00 -6.86418355e-01 -1.19530581e-01 -6.18695915e-01
8.01675558e-01 -5.81376255e-01 3.21686745e-01 1.49437740e-01
4.02608424e-01 9.09402251e-01 3.81407470e-01 2.62693167e-01
1.12840593e+00 5.54832935e-01 3.96163881e-01 4.69617397e-01
-1.03053197e-01 -5.33061028e-01 2.31813386e-01 1.15980697e+00
2.23321542e-01 -6.07798517e-01 -1.02759802e+00 8.74500155e-01
-2.05387378e+00 -1.17735350e+00 -1.08532250e+00 1.60948992e+00
1.10262752e+00 1.09182991e-01 5.17918654e-02 -3.20023835e-01
8.62227261e-01 1.17956996e-01 4.66427095e-02 -4.31282192e-01
-1.34213790e-01 2.00114660e-02 -1.74459040e-01 9.03839767e-01
-9.50895429e-01 4.72908467e-01 6.24253941e+00 6.63326323e-01
-1.35340959e-01 4.35810059e-01 7.47621953e-01 6.03505969e-01
-9.19875562e-01 2.65084654e-01 -9.32408631e-01 3.41472328e-01
1.19676197e+00 -2.60172278e-01 -4.01312739e-01 7.55281031e-01
7.13173905e-03 -2.73473650e-01 -1.19669235e+00 4.54907775e-01
5.52749753e-01 -1.65219069e+00 2.49556363e-01 -5.21669805e-01
1.13760340e+00 1.02247663e-01 -2.96415448e-01 7.48975575e-01
6.70052111e-01 -9.71602857e-01 6.33192956e-01 8.27276051e-01
7.80696142e-03 -2.68841356e-01 1.10392642e+00 5.81928909e-01
-9.28891242e-01 -1.07419990e-01 -3.73042256e-01 -3.49117517e-01
1.64217010e-01 9.34037149e-01 -1.10978556e+00 3.45130771e-01
9.32142258e-01 7.83600032e-01 -9.57683921e-01 1.24867618e+00
-4.96650517e-01 1.20467317e+00 1.72314465e-01 -8.25741529e-01
2.71131128e-01 3.91485393e-01 5.45004427e-01 1.28184247e+00
-1.56449035e-01 8.46325047e-03 2.45838404e-01 8.11354816e-01
-7.29850173e-01 4.37991202e-01 2.04972878e-01 7.70806074e-02
3.89222085e-01 1.23244679e+00 -1.45243898e-01 -5.26848495e-01
-2.04574868e-01 8.04431975e-01 8.12343836e-01 2.29577973e-01
-4.39817280e-01 -5.27997553e-01 -2.28725657e-01 8.60672146e-02
3.54373530e-02 2.81243324e-01 -3.59200329e-01 -1.14367568e+00
-7.08825067e-02 -4.75254178e-01 7.71252573e-01 -9.03891563e-01
-2.24332452e+00 8.40506136e-01 -1.57262206e-01 -7.54515588e-01
-1.92009777e-01 -4.12276387e-01 -1.03069973e+00 1.09422433e+00
-1.69771540e+00 -1.04171729e+00 -4.35745746e-01 5.62278748e-01
6.92780972e-01 -1.43159211e-01 6.90198004e-01 2.66294092e-01
-9.44859013e-02 4.22935337e-01 2.05696583e-01 1.35571495e-01
9.46057320e-01 -1.58377540e+00 3.58299524e-01 4.17447239e-01
1.31743094e-02 6.64875209e-01 6.96288943e-01 -7.65778601e-01
-1.73670083e-01 -9.41907406e-01 2.17285442e+00 -1.06068695e+00
1.14594817e+00 2.43790392e-02 -9.87788856e-01 3.34652930e-01
8.16933513e-01 -4.96709496e-01 9.70053315e-01 6.83924019e-01
-6.54128373e-01 2.33287483e-01 -6.77868128e-01 2.85182208e-01
2.23084614e-01 -9.92025435e-01 -1.52245736e+00 8.14628899e-01
1.00354266e+00 7.28812069e-02 -7.65829384e-01 1.29381537e-01
1.45530790e-01 -6.15685105e-01 7.42082119e-01 -1.07188475e+00
1.07687378e+00 -2.76933312e-01 -1.35580683e-02 -1.17048073e+00
-5.55004597e-01 -5.48772097e-01 -2.64852852e-01 1.37783301e+00
1.13143027e+00 -1.05980434e-01 5.46580195e-01 5.91304243e-01
-3.05472732e-01 -5.25760412e-01 -7.19704747e-01 -2.01591536e-01
4.93870974e-01 -1.21535748e-01 5.27667701e-02 7.92970061e-01
3.96888703e-01 1.04823864e+00 -1.95765302e-01 -2.04331681e-01
1.60650104e-01 1.78615078e-01 3.02538306e-01 -1.89094913e+00
-1.68616548e-01 -5.89539111e-01 3.74592990e-01 -1.37746143e+00
4.53214824e-01 -1.28484368e+00 3.50421429e-01 -2.11269021e+00
8.42784524e-01 -7.28754327e-02 -2.88672179e-01 3.15704159e-02
-9.79386687e-01 -9.11026448e-02 -3.71475443e-02 4.23213303e-01
-1.35181785e+00 6.12772822e-01 1.03996611e+00 -3.42544615e-01
2.72214502e-01 4.30230200e-01 -9.42529738e-01 6.29831791e-01
6.04096949e-01 -6.27287269e-01 -1.30952597e-01 -4.40355122e-01
9.98113692e-01 -2.05979615e-01 4.04829711e-01 -5.35886765e-01
6.07438803e-01 2.30792146e-02 -4.34965231e-02 -9.75431323e-01
-8.23330358e-02 -4.57342535e-01 -6.03250265e-01 1.63677871e-01
-1.46703696e+00 1.55477941e-01 -5.82832873e-01 8.28513145e-01
-3.91154826e-01 -7.52239704e-01 4.60775018e-01 -1.01277344e-02
-2.81453848e-01 5.62273115e-02 -6.76633954e-01 8.63423288e-01
4.27722991e-01 4.41119701e-01 -8.34805429e-01 -9.09764528e-01
-5.93882322e-01 6.68327510e-01 -2.67055124e-01 8.71893108e-01
6.43321872e-01 -1.15384531e+00 -1.53856218e+00 -8.18237484e-01
6.99741602e-01 -2.84305811e-01 3.36472660e-01 6.31065965e-01
-2.68319815e-01 6.27744138e-01 3.00838113e-01 -3.26648563e-01
-1.17154229e+00 -3.25758569e-02 9.07664225e-02 -9.51074064e-01
-5.69092073e-02 1.40219426e+00 -3.37659955e-01 -6.21795297e-01
2.76087284e-01 -2.73086697e-01 -1.13603497e+00 4.09007698e-01
8.84520710e-01 6.53448164e-01 3.66931371e-02 -4.71991658e-01
-2.68411130e-01 7.71247968e-02 -3.69919866e-01 7.44283292e-03
1.13557410e+00 -4.12688434e-01 -5.48882723e-01 3.48959088e-01
1.37513196e+00 2.21702650e-01 -5.42457283e-01 -4.89805877e-01
7.19216704e-01 -2.92707179e-02 -7.71373883e-03 -1.21950758e+00
-2.25105122e-01 8.85406613e-01 2.93727275e-02 5.93259811e-01
3.39402437e-01 5.69162250e-01 8.68846118e-01 7.98906624e-01
-2.48739421e-01 -1.34043527e+00 8.00693393e-01 1.45317781e+00
1.29968750e+00 -1.31038868e+00 -6.88333586e-02 -9.33095813e-02
-8.72250557e-01 1.17007029e+00 6.98665082e-01 -2.35538036e-02
8.57246578e-01 -5.13712585e-01 -4.40488793e-02 -3.91961187e-01
-1.24565673e+00 -1.15045108e-01 1.19452596e+00 3.95337671e-01
7.91543067e-01 -4.40671816e-02 -7.32229948e-01 1.06562459e+00
-2.37798482e-01 -3.59776676e-01 6.37675703e-01 3.40106457e-01
-8.09277296e-01 -7.90229797e-01 -2.05957033e-02 7.38095343e-01
-6.73345089e-01 -3.56365234e-01 -9.68421459e-01 -1.85392037e-01
-9.77054760e-02 1.86821127e+00 -1.27757028e-01 -6.06074512e-01
3.96844149e-02 1.80172354e-01 -4.39829350e-01 -8.28859329e-01
-1.40950000e+00 -6.24582350e-01 7.14699507e-01 -4.27856781e-02
-6.49110794e-01 -3.77768487e-01 -1.10195637e+00 5.35435155e-02
-4.38515186e-01 8.92557204e-01 2.80077845e-01 1.03433776e+00
4.68992144e-01 2.66129017e-01 6.80301845e-01 -3.48839462e-02
-6.48212314e-01 -1.50595224e+00 -1.42093673e-01 7.06148446e-01
5.12255132e-01 -2.07374573e-01 -5.39982617e-01 1.00598633e-01] | [11.4056396484375, 8.004109382629395] |
52d048b4-02ae-40be-a437-3f9c03d22a57 | a-free-lunch-from-vit-adaptive-attention | 2110.01240 | null | https://arxiv.org/abs/2110.01240v2 | https://arxiv.org/pdf/2110.01240v2.pdf | A free lunch from ViT:Adaptive Attention Multi-scale Fusion Transformer for Fine-grained Visual Recognition | Learning subtle representation about object parts plays a vital role in fine-grained visual recognition (FGVR) field. The vision transformer (ViT) achieves promising results on computer vision due to its attention mechanism. Nonetheless, with the fixed size of patches in ViT, the class token in deep layer focuses on the global receptive field and cannot generate multi-granularity features for FGVR. To capture region attention without box annotations and compensate for ViT shortcomings in FGVR, we propose a novel method named Adaptive attention multi-scale Fusion Transformer (AFTrans). The Selective Attention Collection Module (SACM) in our approach leverages attention weights in ViT and filters them adaptively to correspond with the relative importance of input patches. The multiple scales (global and local) pipeline is supervised by our weights sharing encoder and can be easily trained end-to-end. Comprehensive experiments demonstrate that AFTrans can achieve SOTA performance on three published fine-grained benchmarks: CUB-200-2011, Stanford Dogs and iNat2017. | ['Weiqian Chen', 'Feng Ling', 'Zhiyi Wang', 'Xiangcheng Liu', 'Ling Zhang', 'Jian Cao', 'Yuan Zhang'] | 2021-10-04 | null | null | null | null | ['fine-grained-visual-recognition'] | ['computer-vision'] | [-5.49191516e-03 -1.70072690e-02 -1.78126171e-01 -4.74405438e-01
-7.84387827e-01 -3.74365866e-01 6.82387054e-01 -3.07000488e-01
-3.31200838e-01 4.13327605e-01 5.08060753e-01 1.50160640e-01
1.98401794e-01 -6.94862068e-01 -1.02568543e+00 -4.99996752e-01
3.57646108e-01 2.28477135e-01 5.45354903e-01 -7.34754056e-02
-3.15808840e-02 6.25866771e-01 -1.53327167e+00 7.75885165e-01
6.86745286e-01 1.41329908e+00 4.66904163e-01 6.48408055e-01
-8.30771867e-04 1.13444579e+00 -4.68884110e-01 -3.96278143e-01
5.21624565e-01 -1.07545499e-02 -8.18935454e-01 1.12546362e-01
1.11025631e+00 -6.77661836e-01 -6.86063647e-01 9.57584620e-01
3.38530272e-01 2.67566647e-02 6.54927969e-01 -1.05179310e+00
-1.22705555e+00 7.49462664e-01 -9.23010290e-01 5.50389290e-01
-2.06319243e-01 4.21272606e-01 1.16865706e+00 -1.18276060e+00
2.75936514e-01 1.51018834e+00 6.26790404e-01 5.01476347e-01
-1.18564701e+00 -5.30070007e-01 5.62597632e-01 2.41458297e-01
-1.32838094e+00 -4.23775017e-01 4.42004561e-01 -4.64071512e-01
1.17081761e+00 2.05172272e-03 5.80236912e-01 1.08268046e+00
1.12944663e-01 9.98678446e-01 1.06221426e+00 -8.90005827e-02
7.62054836e-03 -3.84469301e-01 2.99004942e-01 8.90189290e-01
8.66428241e-02 2.16059089e-01 -4.41261023e-01 1.09494247e-01
1.16360736e+00 4.86761332e-01 -1.05211794e-01 -2.53526986e-01
-1.48060763e+00 7.55107939e-01 1.19988048e+00 1.11140432e-02
-5.70504189e-01 6.36777937e-01 3.40853035e-01 3.45638692e-02
3.17936540e-01 1.34139583e-01 -6.15406692e-01 4.23491448e-01
-7.09266961e-01 1.07176073e-01 -6.99334964e-03 1.02102828e+00
7.95367599e-01 5.85374190e-03 -1.07203937e+00 9.52561617e-01
2.50685155e-01 5.16866326e-01 5.05451083e-01 -8.11000049e-01
3.09909701e-01 7.84997702e-01 -3.21730524e-02 -4.34840411e-01
-4.94798310e-02 -5.40820003e-01 -1.01743412e+00 1.89534679e-01
2.98582911e-01 2.30088770e-01 -1.61996770e+00 1.51224887e+00
3.44871670e-01 2.32786164e-01 -2.08298311e-01 1.25178242e+00
1.19432187e+00 3.32194388e-01 3.62016559e-01 3.93470019e-01
1.64125192e+00 -1.56422353e+00 -1.85562953e-01 -5.03806531e-01
6.09894283e-02 -5.38309336e-01 1.18972874e+00 -2.05163937e-02
-9.85539079e-01 -9.08588052e-01 -8.02614570e-01 -6.39937282e-01
-3.16446185e-01 5.07259846e-01 7.03496754e-01 -3.07561215e-02
-1.01976991e+00 2.53229022e-01 -8.14984024e-01 -1.15809113e-01
1.00193906e+00 1.59632176e-01 -4.88474160e-01 -1.93671837e-01
-8.49987030e-01 7.04804242e-01 1.23957194e-01 1.29068479e-01
-1.41761005e+00 -9.08340752e-01 -8.61502409e-01 3.57410103e-01
8.83331895e-02 -1.08692539e+00 1.18742859e+00 -9.93751287e-01
-1.15499508e+00 9.19828832e-01 -2.21262813e-01 -7.27317214e-01
4.05648887e-01 -2.44822636e-01 2.28947997e-01 2.49304324e-01
4.50483501e-01 1.09104788e+00 1.24444020e+00 -8.24989796e-01
-8.58286679e-01 -4.55508053e-01 2.46693671e-01 -1.92114323e-01
-5.38077764e-02 1.28546178e-01 -5.70801258e-01 -9.96908069e-01
-1.96260810e-01 -5.80467761e-01 -3.16825628e-01 1.60904303e-01
-2.21914738e-01 -6.12244904e-01 6.84395909e-01 -4.30812389e-01
8.56097043e-01 -2.12007928e+00 8.73592421e-02 -3.34923953e-01
4.82975155e-01 3.08877110e-01 -3.47244799e-01 -4.83398102e-02
3.01718619e-02 -1.50028095e-01 -9.27007347e-02 -3.29497665e-01
1.05936013e-01 1.05552576e-01 -5.49137592e-01 3.05410475e-01
6.12992346e-01 1.50409687e+00 -7.57709265e-01 -4.25923735e-01
4.62321758e-01 6.50344074e-01 -7.22743452e-01 2.92761296e-01
-2.94730842e-01 -2.12229192e-02 -7.12533414e-01 9.48446512e-01
7.64576495e-01 -6.12166762e-01 -3.62180382e-01 -7.69624233e-01
-2.30745990e-02 -3.71967815e-02 -6.01109028e-01 1.72942233e+00
-3.78746510e-01 3.05257201e-01 -2.47609131e-02 -6.96949303e-01
7.09002793e-01 -4.86197099e-02 -1.18074403e-03 -8.47685277e-01
1.64645806e-01 -1.33335397e-01 -2.14152053e-01 -1.89166009e-01
4.63357389e-01 2.67045766e-01 -1.50970563e-01 6.33763149e-02
5.15859127e-01 2.79423624e-01 6.74381629e-02 2.62339920e-01
1.16534173e+00 2.15601489e-01 3.91737103e-01 -2.87232280e-01
3.23370278e-01 -1.00339830e-01 6.62621021e-01 8.20152938e-01
-5.52885711e-01 8.29511881e-01 2.07728054e-02 -7.30469346e-01
-9.68082190e-01 -1.10203278e+00 -1.20618765e-03 1.55795717e+00
1.37747303e-01 -2.93213218e-01 -5.92421651e-01 -1.05956638e+00
3.95995319e-01 7.50060678e-02 -1.23903561e+00 -1.30489215e-01
-4.98404264e-01 -2.55790591e-01 6.16505265e-01 1.18535960e+00
7.58671939e-01 -1.34229243e+00 -7.19340146e-01 1.28538936e-01
-1.49289012e-01 -1.18074048e+00 -8.70884895e-01 2.67305255e-01
-5.13411283e-01 -1.07867169e+00 -8.54887724e-01 -7.72632599e-01
6.52197897e-01 7.46950805e-01 1.22325432e+00 -1.39433354e-01
-6.03741944e-01 2.28706226e-01 -3.29595715e-01 -3.26945364e-01
2.40705043e-01 -2.93914527e-02 -3.96922022e-01 1.55391276e-01
4.43830073e-01 -2.00814754e-01 -1.07079184e+00 3.63360465e-01
-6.60828710e-01 -2.48017963e-02 9.70348418e-01 1.19165993e+00
1.01771832e+00 -6.04237437e-01 5.79196572e-01 -6.87659025e-01
9.64754298e-02 -3.21657330e-01 -5.95954895e-01 5.16695559e-01
-3.85825485e-02 1.36704400e-01 6.47624433e-01 -3.48106384e-01
-1.07025552e+00 1.85774952e-01 -3.40867601e-02 -1.13366699e+00
-2.75032103e-01 -1.47863477e-01 -6.30288795e-02 -2.90973276e-01
5.27263045e-01 1.64523646e-01 -4.83704329e-01 -6.47371769e-01
6.77037537e-01 3.82707328e-01 7.53279269e-01 -4.78528947e-01
5.55279315e-01 4.99877334e-01 -4.38316435e-01 -4.17688400e-01
-1.24430156e+00 -5.46423733e-01 -4.91615713e-01 1.35986015e-01
1.04631126e+00 -1.44815528e+00 -7.61095524e-01 4.01082963e-01
-1.02785039e+00 -4.86564606e-01 -5.11707485e-01 2.68983662e-01
-4.93520975e-01 -1.58084124e-01 -7.50550389e-01 -3.89015883e-01
-7.16075242e-01 -1.12731266e+00 1.64912319e+00 5.03698528e-01
1.12450279e-01 -5.53462327e-01 -2.58912623e-01 4.85036880e-01
6.66131735e-01 7.03725219e-02 5.21522701e-01 -8.41591060e-02
-1.01067126e+00 2.63400525e-01 -9.09131050e-01 2.25270003e-01
1.06659733e-01 -1.39044762e-01 -1.32790303e+00 -3.64314049e-01
-4.74965036e-01 -7.09167778e-01 1.66366792e+00 5.79929411e-01
1.53147161e+00 -2.72318631e-01 -3.12669158e-01 9.04156983e-01
1.50099766e+00 -3.61160040e-01 4.63179410e-01 1.04405433e-01
1.04513907e+00 2.63273060e-01 6.24073982e-01 2.72811502e-01
6.16571248e-01 5.79839647e-01 6.34079874e-01 -2.42016420e-01
-8.09691310e-01 -3.14109474e-01 3.76500368e-01 1.39447004e-01
-6.03575222e-02 1.47506371e-01 -4.50707823e-01 8.96088243e-01
-1.85378993e+00 -1.04118872e+00 3.18796188e-01 1.72112620e+00
7.32480586e-01 -5.82177490e-02 1.22595069e-04 -4.09456432e-01
5.51115513e-01 2.34181389e-01 -8.04959238e-01 -2.29375497e-01
-3.11767846e-01 3.42880011e-01 6.25135958e-01 2.00750738e-01
-1.25784576e+00 1.29490757e+00 5.47991323e+00 1.01160491e+00
-1.09814346e+00 2.79630393e-01 7.71857619e-01 -9.97743383e-02
-9.61223096e-02 -2.99480110e-01 -1.01471162e+00 3.18900377e-01
3.70430231e-01 2.67316699e-01 4.46341038e-01 9.83182132e-01
-1.10919647e-01 2.32560292e-01 -1.06064892e+00 9.66401517e-01
-3.14816348e-02 -1.51602292e+00 2.44260073e-01 -2.26110995e-01
8.77749205e-01 6.99931741e-01 1.42400965e-01 5.12946248e-01
5.02748370e-01 -1.12819791e+00 9.43906248e-01 6.29168928e-01
1.07945168e+00 -4.97797161e-01 4.56712335e-01 9.75401048e-03
-1.59369600e+00 -2.89468735e-01 -8.08865190e-01 2.86484450e-01
-2.35083744e-01 4.78331059e-01 -6.91036046e-01 2.44139358e-01
1.20402515e+00 7.96799183e-01 -7.86376476e-01 9.09640312e-01
-1.59904405e-01 5.02353966e-01 -1.29081994e-01 3.27640116e-01
3.95120889e-01 3.83953929e-01 1.47212029e-01 1.23901594e+00
1.96607206e-02 1.01997085e-01 2.15632603e-01 1.03082752e+00
-4.33228165e-01 -4.37166184e-01 -1.71363309e-01 -3.55768502e-02
3.62962127e-01 1.56074524e+00 -5.73980331e-01 -4.60263938e-01
-5.05930185e-01 9.61625516e-01 8.39799643e-01 3.28940809e-01
-8.51067185e-01 -2.24191815e-01 1.00622189e+00 6.14185743e-02
1.17512429e+00 2.77935296e-01 -2.15806544e-01 -1.35900974e+00
-5.62575832e-02 -8.61674905e-01 4.91824031e-01 -7.88276076e-01
-1.73794377e+00 9.06402111e-01 -3.59437466e-01 -9.33129132e-01
1.96702749e-01 -5.24786055e-01 -6.08097613e-01 8.72515500e-01
-1.83482909e+00 -1.89835584e+00 -6.06804967e-01 1.03960145e+00
8.57300103e-01 -1.26190379e-01 6.96703017e-01 1.77517548e-01
-4.17216241e-01 9.08258915e-01 -4.28207636e-01 2.70956904e-01
7.31649518e-01 -1.37091339e+00 8.33317757e-01 7.66211987e-01
3.73642482e-02 6.72182858e-01 1.63222909e-01 -5.28224826e-01
-1.22604597e+00 -1.80612755e+00 5.99665523e-01 -5.00856698e-01
5.51761389e-01 -3.65179688e-01 -8.40091705e-01 8.33656311e-01
2.67494798e-01 9.43349063e-01 8.88020694e-02 -1.07808486e-01
-1.00658143e+00 -2.93839633e-01 -1.05946648e+00 2.13562205e-01
1.18460190e+00 -6.72992945e-01 -7.19052196e-01 2.12142110e-01
1.03166580e+00 -3.17759693e-01 -9.01347816e-01 5.58251917e-01
4.73560721e-01 -9.76571500e-01 1.17735386e+00 -6.04281664e-01
4.61627960e-01 -6.69767857e-01 -3.23750377e-01 -1.16377497e+00
-9.75549638e-01 -1.16581224e-01 -3.45275760e-01 1.13723612e+00
5.06435521e-02 -4.82581019e-01 6.73429310e-01 8.71232003e-02
-3.99882585e-01 -8.88976157e-01 -6.83369339e-01 -3.09572607e-01
-1.00274891e-01 2.16694865e-02 9.24085855e-01 5.15682220e-01
-6.82394147e-01 3.96514267e-01 -2.24992871e-01 2.10975185e-01
9.25394535e-01 5.67274511e-01 6.84143901e-01 -1.07268274e+00
-4.99901801e-01 -5.12565911e-01 -4.05318230e-01 -1.26356959e+00
-9.87272635e-02 -7.33964384e-01 1.02188841e-01 -1.62035596e+00
5.21990836e-01 -4.88688409e-01 -6.78159237e-01 8.81779373e-01
-4.68746454e-01 8.33871722e-01 3.51321042e-01 2.15184823e-01
-9.62230325e-01 6.64765656e-01 1.47832155e+00 -4.11687225e-01
3.63601446e-01 -3.57389450e-01 -8.91559899e-01 6.97119474e-01
3.86172205e-01 -1.08249702e-01 -2.54606545e-01 -7.45702267e-01
-3.30717176e-01 -2.30191037e-01 8.57353508e-01 -7.81931698e-01
2.24157616e-01 -1.58528611e-01 9.20103490e-01 -6.83645964e-01
1.63621977e-01 -6.04119241e-01 -2.44978681e-01 1.25932053e-01
-3.90746623e-01 -1.52085215e-01 1.21418834e-01 5.62484086e-01
-3.14702153e-01 5.70575535e-01 9.76346791e-01 -1.60467595e-01
-1.09308875e+00 9.39654052e-01 1.33232221e-01 1.81743830e-01
9.03404415e-01 -6.28424808e-02 -6.61366940e-01 3.14380497e-01
-5.56783617e-01 3.80982816e-01 4.55976367e-01 6.74197674e-01
7.28613317e-01 -1.48851907e+00 -8.88199747e-01 4.14493889e-01
5.56103587e-01 3.59714657e-01 7.13397026e-01 7.19723523e-01
-1.48765877e-01 5.10608494e-01 -4.16035295e-01 -7.05593944e-01
-1.00084782e+00 7.53934383e-01 5.29697776e-01 -2.37925485e-01
-7.96375275e-01 1.42173290e+00 1.08685720e+00 -7.50392023e-03
2.91033953e-01 -7.23552167e-01 -2.24931672e-01 -2.54157603e-01
8.39083433e-01 -1.01975258e-02 1.11534081e-01 -7.17173994e-01
-4.60432470e-01 5.96359432e-01 -5.21023154e-01 5.49229205e-01
1.27889717e+00 -1.11983202e-01 9.64117646e-02 -1.16886022e-02
1.11287367e+00 -1.76680952e-01 -1.91278493e+00 -4.91191536e-01
-5.04088819e-01 -5.23711145e-01 2.03054056e-01 -9.61733222e-01
-1.37586820e+00 8.93389761e-01 5.38667321e-01 -1.95136681e-01
1.25714231e+00 3.26603830e-01 9.24299777e-01 1.51445165e-01
4.38164890e-01 -6.72805607e-01 1.06471956e-01 5.58797777e-01
1.15009916e+00 -1.40051746e+00 -3.64658773e-01 -1.37964889e-01
-7.36610115e-01 7.82125413e-01 9.50838864e-01 -6.65157616e-01
4.95400965e-01 3.39282006e-01 -3.97351868e-02 -7.70756379e-02
-1.09748352e+00 -4.82998759e-01 5.84202528e-01 6.16867483e-01
2.73774296e-01 2.43919328e-01 1.89315274e-01 8.19433153e-01
5.20945899e-02 5.25917225e-02 -2.46515274e-02 5.21239102e-01
-5.58443367e-01 -5.87251782e-01 -2.25212887e-01 5.96200228e-01
-4.39661324e-01 -2.15739965e-01 -4.12002027e-01 3.34862739e-01
4.74338621e-01 6.87474847e-01 1.77461430e-01 -2.50070870e-01
3.79319876e-01 -4.82160300e-01 7.38716304e-01 -5.54833233e-01
-9.03053820e-01 1.02795340e-01 -2.55846500e-01 -1.02621031e+00
-4.03341681e-01 -2.25266248e-01 -1.19836068e+00 -6.87835664e-02
-1.57030851e-01 -5.96573055e-02 1.50385872e-01 9.43540096e-01
6.18262470e-01 9.61263537e-01 4.10475641e-01 -9.22830343e-01
-5.80190718e-01 -1.17227232e+00 -3.49810988e-01 3.57796878e-01
6.51686370e-01 -8.03922474e-01 6.89833164e-02 5.55135123e-02] | [9.580404281616211, 2.0093352794647217] |
c34e6b10-4546-45c5-8b55-17f2f77faffc | adaptive-ensemble-q-learning-minimizing-1 | 2306.11918 | null | https://arxiv.org/abs/2306.11918v1 | https://arxiv.org/pdf/2306.11918v1.pdf | Adaptive Ensemble Q-learning: Minimizing Estimation Bias via Error Feedback | The ensemble method is a promising way to mitigate the overestimation issue in Q-learning, where multiple function approximators are used to estimate the action values. It is known that the estimation bias hinges heavily on the ensemble size (i.e., the number of Q-function approximators used in the target), and that determining the `right' ensemble size is highly nontrivial, because of the time-varying nature of the function approximation errors during the learning process. To tackle this challenge, we first derive an upper bound and a lower bound on the estimation bias, based on which the ensemble size is adapted to drive the bias to be nearly zero, thereby coping with the impact of the time-varying approximation errors accordingly. Motivated by the theoretic findings, we advocate that the ensemble method can be combined with Model Identification Adaptive Control (MIAC) for effective ensemble size adaptation. Specifically, we devise Adaptive Ensemble Q-learning (AdaEQ), a generalized ensemble method with two key steps: (a) approximation error characterization which serves as the feedback for flexibly controlling the ensemble size, and (b) ensemble size adaptation tailored towards minimizing the estimation bias. Extensive experiments are carried out to show that AdaEQ can improve the learning performance than the existing methods for the MuJoCo benchmark. | ['Junshan Zhang', 'Sen Lin', 'Hang Wang'] | 2023-06-20 | adaptive-ensemble-q-learning-minimizing | http://proceedings.neurips.cc/paper/2021/hash/cfa45151ccad6bf11ea146ed563f2119-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/cfa45151ccad6bf11ea146ed563f2119-Paper.pdf | neurips-2021-12 | ['q-learning'] | ['methodology'] | [ 3.58436368e-02 -2.90916353e-01 -1.75572038e-01 -4.11466882e-02
-7.74345875e-01 -5.90866268e-01 1.99530214e-01 1.09112076e-01
-3.96653801e-01 9.52031374e-01 -1.61785722e-01 -4.38640326e-01
-4.11066473e-01 -6.78209186e-01 -6.04267955e-01 -1.10839593e+00
8.33560079e-02 1.99145768e-02 -2.78909337e-02 -4.88309175e-01
5.42757928e-01 1.71494663e-01 -1.48747230e+00 -1.45838916e-01
1.53378510e+00 1.31166768e+00 -1.16911165e-01 6.54589236e-01
2.44548559e-01 6.00956023e-01 -8.20820212e-01 -1.94944978e-01
3.08366150e-01 -7.68599212e-01 -2.98266560e-01 -1.62058264e-01
1.00741394e-01 -1.78034231e-01 1.08170994e-01 9.64089572e-01
7.44208872e-01 4.90296632e-01 6.55516863e-01 -1.26247191e+00
-1.14786565e-01 4.98619497e-01 -2.06370175e-01 3.33089203e-01
-2.98358928e-02 3.68848294e-01 8.97541761e-01 -7.09880948e-01
5.37355207e-02 1.07847369e+00 6.43037736e-01 6.11200392e-01
-1.26504123e+00 -8.89682770e-01 3.27841550e-01 8.57411623e-02
-1.39728475e+00 -3.83404791e-01 6.36127949e-01 -5.36324859e-01
4.35274482e-01 1.77244484e-01 6.64359510e-01 7.93502569e-01
6.67187512e-01 3.62399489e-01 1.25960410e+00 -5.63899875e-01
8.41598451e-01 2.52035290e-01 -1.47689030e-01 4.03745711e-01
2.19435155e-01 5.90648055e-01 -3.85352463e-01 -2.79996634e-01
8.18013430e-01 -3.17315966e-01 -2.85765260e-01 -3.45899820e-01
-7.41484165e-01 9.52956438e-01 2.10369363e-01 2.13742871e-02
-5.66754043e-01 2.43670985e-01 5.93062937e-01 5.73743045e-01
5.34108102e-01 8.43459189e-01 -6.11509204e-01 -2.95046747e-01
-4.62598383e-01 5.52131891e-01 7.08968997e-01 4.30176616e-01
5.20901203e-01 3.32106590e-01 -5.03362596e-01 7.62075424e-01
9.37743187e-02 5.74142635e-01 4.79509771e-01 -1.11994016e+00
5.45444787e-01 7.28771389e-01 5.19110084e-01 -7.98855841e-01
1.97009183e-02 -7.74515867e-01 -6.61088467e-01 4.21833068e-01
6.48752987e-01 -5.53885400e-01 -5.63467503e-01 1.83569026e+00
5.66456199e-01 3.09675150e-02 -7.63428882e-02 8.12678337e-01
-5.56027666e-02 4.42880005e-01 9.40738842e-02 -5.77435613e-01
1.02344286e+00 -6.37230039e-01 -7.48549879e-01 -1.98591985e-02
5.53128183e-01 -3.59374762e-01 1.02923000e+00 5.50444961e-01
-9.39280510e-01 -5.60033560e-01 -1.10160553e+00 8.27638209e-01
-1.16949610e-01 -1.40717611e-01 1.71620920e-01 8.05580616e-01
-6.84383690e-01 9.11634743e-01 -6.04248941e-01 2.12017283e-01
1.82898134e-01 5.72231352e-01 2.34413847e-01 3.31066251e-01
-1.42475152e+00 9.13612783e-01 6.26328588e-01 2.66396075e-01
-6.51967645e-01 -7.31783688e-01 -3.93609703e-01 5.61643057e-02
5.75562477e-01 -7.15424955e-01 1.21538031e+00 -1.45227909e+00
-1.91726816e+00 -5.19912802e-02 1.86055169e-01 -4.61853951e-01
6.44001424e-01 -1.99263304e-01 -2.49845192e-01 -2.24821255e-01
-3.77622843e-01 2.01411143e-01 1.27610326e+00 -1.10356963e+00
-7.87730694e-01 -4.37472284e-01 1.08859716e-02 3.88513058e-01
-4.60111499e-01 -2.91193128e-01 1.72829807e-01 -7.04515159e-01
-1.85366288e-01 -1.02852738e+00 -4.72350389e-01 -3.20984870e-01
-3.23859677e-02 -3.78055125e-01 5.16221941e-01 -4.45016474e-01
1.74567389e+00 -1.85206103e+00 1.58990294e-01 5.09953380e-01
-3.51665318e-02 3.68472219e-01 -9.26945657e-02 5.09081662e-01
1.39903858e-01 2.40844414e-02 -2.47180089e-01 2.04182133e-01
-1.24914318e-01 1.35422215e-01 -3.06850404e-01 2.04981461e-01
1.43994540e-01 6.13198757e-01 -1.08447778e+00 -1.84674859e-01
2.66055595e-02 2.63361573e-01 -7.23899186e-01 5.33467412e-01
-2.27786049e-01 8.70429218e-01 -7.37156212e-01 2.96847820e-01
3.06927353e-01 -7.05097169e-02 3.98479700e-01 -1.18287995e-01
-1.90882877e-01 -3.34823653e-02 -1.53496480e+00 8.25163305e-01
-5.73978424e-01 1.53585047e-01 9.03796256e-02 -1.16930854e+00
1.07582498e+00 3.34086746e-01 5.89596510e-01 -6.05761826e-01
3.17184091e-01 3.52939963e-01 2.41622284e-01 -3.36327493e-01
2.62229860e-01 -1.85163721e-01 4.17335145e-02 4.05802995e-01
-2.67820179e-01 -1.61897540e-01 1.93297684e-01 -4.06124949e-01
7.73516238e-01 1.99102297e-01 4.73914236e-01 -4.72339094e-01
9.12085354e-01 -3.40630591e-01 8.58302474e-01 7.69244611e-01
-3.60051900e-01 -4.14749384e-02 4.33718562e-01 -3.70812058e-01
-1.03174484e+00 -7.30715334e-01 -1.85882628e-01 1.29863656e+00
4.19016965e-02 -2.02870965e-01 -8.24067116e-01 -8.09804916e-01
2.93593705e-01 7.81909883e-01 -5.70997000e-01 -6.33983016e-01
-7.09869444e-01 -7.52740979e-01 1.51547551e-01 7.54609823e-01
3.83758008e-01 -7.34533131e-01 -6.68767095e-01 4.47585702e-01
6.77067265e-02 -5.81917346e-01 -7.32032478e-01 2.07071081e-01
-1.07657540e+00 -1.06450927e+00 -4.85103637e-01 -1.16974108e-01
6.42544568e-01 -2.82571673e-01 1.03599048e+00 8.96646306e-02
3.42185169e-01 5.03512084e-01 -2.87365258e-01 -7.03386843e-01
-7.01527715e-01 7.15785548e-02 1.75327271e-01 1.16718583e-01
-1.24475919e-01 -4.51081753e-01 -8.06627572e-01 4.82074618e-01
-8.24924052e-01 -2.72916794e-01 4.23435092e-01 1.29246974e+00
5.23108780e-01 1.18655853e-01 1.14269495e+00 -7.90370464e-01
9.39071536e-01 -5.02006471e-01 -9.25842047e-01 4.08598810e-01
-1.12825608e+00 2.64052242e-01 9.98434544e-01 -7.46167719e-01
-8.60037625e-01 -1.58554509e-01 2.12141518e-02 -4.77753907e-01
4.63288307e-01 4.29035813e-01 6.39158636e-02 -2.74795145e-01
6.04389906e-01 2.97416866e-01 2.88765103e-01 -3.57786007e-02
2.44373992e-01 6.88073635e-01 1.23596393e-01 -8.55732620e-01
5.71611345e-01 -9.46036801e-02 1.67456657e-01 -3.79541755e-01
-7.26665556e-01 -1.41130611e-01 -4.61011291e-01 -4.91973013e-01
4.07273382e-01 -6.97000384e-01 -1.04222155e+00 2.46455356e-01
-6.59613311e-01 -4.57635999e-01 -2.57368445e-01 3.33554745e-01
-7.78791308e-01 -7.66503885e-02 -1.47704303e-01 -1.30880976e+00
-6.79890513e-01 -1.06456983e+00 5.67837775e-01 3.12900037e-01
-5.73693998e-02 -1.02750850e+00 1.76035196e-01 1.34478971e-01
5.37528098e-01 3.69100004e-01 8.52542996e-01 -6.79092288e-01
-2.80826300e-01 -1.43651038e-01 2.69559890e-01 7.70814598e-01
-1.35004312e-01 -3.43712531e-02 -5.98165512e-01 -6.37799323e-01
8.65677148e-02 -2.10354328e-01 3.93499851e-01 4.30518985e-01
1.22120821e+00 -5.02599537e-01 -5.29937111e-02 2.18113363e-01
1.40562713e+00 5.03556490e-01 3.20610523e-01 3.22016358e-01
3.84064645e-01 3.61519843e-01 6.89321816e-01 8.38438869e-01
2.53933251e-01 7.07157254e-01 4.50106502e-01 4.67749864e-01
4.95013118e-01 -2.85563737e-01 4.89490181e-01 8.73195529e-01
-2.51290113e-01 -8.32286999e-02 -6.82075560e-01 1.17937468e-01
-1.94460464e+00 -6.95328057e-01 3.16643804e-01 2.70502830e+00
9.48426545e-01 1.06899738e-01 4.27972764e-01 2.26849511e-01
6.85981989e-01 -1.55437410e-01 -1.00787604e+00 -4.88595515e-01
3.69141638e-01 9.79943871e-02 4.83691663e-01 3.83224666e-01
-9.27558243e-01 3.04770768e-01 6.70000219e+00 9.19750512e-01
-1.25277054e+00 -1.44883916e-01 7.85193145e-01 1.13035724e-01
-2.12399483e-01 -1.11039884e-01 -7.80741096e-01 8.55363727e-01
1.22229958e+00 -3.49679023e-01 6.77734256e-01 7.93680191e-01
5.04705548e-01 1.52501790e-02 -9.25552547e-01 5.76257229e-01
-4.22706068e-01 -1.00607061e+00 -2.41394654e-01 3.50391679e-02
9.97513413e-01 -5.49879909e-01 -2.08199378e-02 7.75116801e-01
3.70650925e-02 -8.67695570e-01 6.51087165e-01 6.60700202e-01
4.38832283e-01 -1.13947535e+00 8.90265644e-01 6.89046204e-01
-1.05150270e+00 -8.00007820e-01 -1.41009122e-01 -1.71437263e-02
-3.20377916e-01 6.45589709e-01 -5.85925639e-01 4.02914941e-01
3.39268684e-01 2.23492980e-01 -2.14365870e-01 9.58542347e-01
-1.35857537e-01 7.37625480e-01 -1.72499269e-01 -3.25491518e-01
-4.64695431e-02 -5.82109034e-01 6.31447017e-01 6.52946770e-01
3.54039669e-01 1.78216591e-01 2.50346065e-01 8.99472415e-01
7.15595856e-02 2.09908530e-01 -2.08675519e-01 -6.23032451e-02
9.58975673e-01 9.38772976e-01 -1.87170923e-01 -1.41551942e-01
3.42019126e-02 3.89097661e-01 3.60226244e-01 4.47074026e-01
-8.14042687e-01 -3.24892223e-01 5.81049323e-01 -1.48761749e-01
3.67487609e-01 -4.33473364e-02 -2.36914739e-01 -8.27603221e-01
-2.02358123e-02 -1.17697847e+00 3.81191283e-01 -1.86698690e-01
-1.27285767e+00 3.12343478e-01 -3.23415101e-02 -1.43891752e+00
-5.39572358e-01 -4.20965225e-01 -7.17639387e-01 7.72766113e-01
-1.24726319e+00 -4.76719767e-01 1.79991517e-02 1.52120724e-01
3.40566397e-01 -7.67198298e-03 6.31550610e-01 1.11226641e-01
-9.48272705e-01 8.25195014e-01 6.52268767e-01 -2.90897757e-01
5.49724340e-01 -1.33396268e+00 -3.32118273e-01 4.41442817e-01
-5.24158418e-01 4.06768501e-01 8.61932397e-01 -2.60193169e-01
-1.62687302e+00 -1.03748775e+00 3.82682443e-01 -5.26370883e-01
5.73331892e-01 -1.31512702e-01 -1.00283027e+00 1.88913792e-01
-2.01250389e-01 1.48551241e-01 4.68777955e-01 -5.31462990e-02
4.60251980e-02 -5.77037632e-01 -1.05832136e+00 5.05792737e-01
5.92030883e-01 -2.36522943e-01 -2.38061458e-01 -3.18819992e-02
5.46608627e-01 -5.27606130e-01 -1.37257159e+00 6.20870292e-01
7.31992066e-01 -8.62774551e-01 7.46306419e-01 -6.23115063e-01
2.84807622e-01 -9.91386771e-02 -5.15006902e-03 -1.78366923e+00
-1.35229006e-01 -6.72560573e-01 -6.56795263e-01 1.10286236e+00
1.35763407e-01 -9.24735844e-01 6.17976904e-01 6.33307099e-01
1.72700360e-01 -1.50375760e+00 -9.92342651e-01 -9.67940271e-01
3.94740105e-01 -1.67483822e-01 6.74690127e-01 5.71667850e-01
-1.62031874e-01 2.44209424e-01 -2.28535205e-01 -5.47777973e-02
5.34489870e-01 2.75034253e-02 7.56247163e-01 -1.03217900e+00
-3.92581373e-01 -6.03965342e-01 7.09825531e-02 -9.36414719e-01
1.34955700e-02 -2.32976258e-01 2.39271015e-01 -6.88440561e-01
-1.81858122e-01 -6.49527073e-01 -7.91806936e-01 5.33364601e-02
-7.84779906e-01 -3.31545234e-01 3.07904243e-01 2.03902438e-01
-5.29216409e-01 8.77744496e-01 1.51709735e+00 1.91528901e-01
-5.45267940e-01 5.59876561e-01 -6.78298533e-01 3.35667312e-01
7.65375853e-01 -4.14080471e-01 -5.23025632e-01 2.53149241e-01
3.25275391e-01 4.45645273e-01 1.19487643e-01 -8.57342064e-01
3.71514377e-03 -4.94511962e-01 2.47596830e-01 -2.94328153e-01
-1.55980647e-01 -7.33770132e-01 -1.91471830e-01 5.98985434e-01
-5.85225284e-01 1.65076360e-01 1.84327632e-01 8.27692032e-01
-9.44143236e-02 -2.11186484e-01 8.58638287e-01 1.78325966e-01
-2.95057982e-01 2.55998015e-01 -3.94650877e-01 2.03099638e-01
9.37188506e-01 -1.13685355e-01 6.56184927e-02 -3.44402224e-01
-3.59103948e-01 5.80988824e-01 6.05138466e-02 2.41765961e-01
2.63545901e-01 -1.34994090e+00 -8.13404441e-01 2.98813760e-01
-1.85871776e-02 -2.56047368e-01 1.16185509e-01 9.95503187e-01
1.55590951e-01 2.94595957e-01 1.24592287e-02 -4.33748513e-01
-9.31096375e-01 3.51056486e-01 8.22683811e-01 -5.52269161e-01
-1.31784022e-01 4.78341222e-01 -1.45945892e-01 -3.45879465e-01
2.87475973e-01 -8.10859501e-02 -2.03083724e-01 -5.08539639e-02
4.01787728e-01 6.99085236e-01 6.54191598e-02 -9.49180052e-02
-1.74312472e-01 4.28329617e-01 1.97471634e-01 1.87514380e-01
1.04102886e+00 -2.57388502e-01 1.63093194e-01 5.34674764e-01
7.53777802e-01 -2.97692299e-01 -1.62501097e+00 -5.88150769e-02
5.23360521e-02 -3.07398766e-01 2.46348813e-01 -8.63587022e-01
-9.08926487e-01 5.08242488e-01 9.04216349e-01 3.97593141e-01
1.44726253e+00 -7.35405028e-01 6.51545882e-01 3.17481518e-01
3.28557163e-01 -1.52419269e+00 1.05022386e-01 5.40268481e-01
9.59200203e-01 -1.20996213e+00 1.15744799e-01 5.55615611e-02
-6.28177106e-01 1.22082984e+00 9.10449684e-01 -2.31853172e-01
5.32496214e-01 3.79346162e-02 -3.21773477e-02 2.33121619e-01
-1.09200001e+00 1.83458865e-01 5.23068428e-01 2.60278732e-01
4.49807167e-01 8.33063573e-02 -6.34443343e-01 6.27580225e-01
1.10601611e-01 6.39569238e-02 4.85530868e-03 8.13027382e-01
-4.74658877e-01 -1.09361517e+00 -6.82172477e-01 4.57913995e-01
-2.74844050e-01 2.80104965e-01 -1.44645227e-02 7.22429037e-01
2.05910057e-01 8.94395769e-01 6.07223772e-02 -3.37825298e-01
5.44609070e-01 1.33741066e-01 4.92863595e-01 -2.21711233e-01
-8.86114359e-01 -1.21962570e-01 -4.21484821e-02 -5.11682749e-01
-2.35981286e-01 -3.59141052e-01 -1.00641310e+00 -2.16822773e-01
-8.17734718e-01 3.90816838e-01 4.57892686e-01 1.26382422e+00
3.41812938e-01 5.73081315e-01 1.20424128e+00 -7.03866839e-01
-1.53822386e+00 -8.85797799e-01 -4.11131710e-01 3.55386317e-01
3.70604545e-01 -8.88670921e-01 -6.06293797e-01 -4.36974317e-01] | [4.214423656463623, 2.408907651901245] |
419870f9-cb77-4219-ae38-c69a67eb0d58 | chatgpt-vs-human-authored-text-insights-into | 2306.07799 | null | https://arxiv.org/abs/2306.07799v1 | https://arxiv.org/pdf/2306.07799v1.pdf | ChatGPT vs Human-authored Text: Insights into Controllable Text Summarization and Sentence Style Transfer | Large-scale language models, like ChatGPT, have garnered significant media attention and stunned the public with their remarkable capacity for generating coherent text from short natural language prompts. In this paper, we aim to conduct a systematic inspection of ChatGPT's performance in two controllable generation tasks, with respect to ChatGPT's ability to adapt its output to different target audiences (expert vs. layman) and writing styles (formal vs. informal). Additionally, we evaluate the faithfulness of the generated text, and compare the model's performance with human-authored texts. Our findings indicate that the stylistic variations produced by humans are considerably larger than those demonstrated by ChatGPT, and the generated texts diverge from human samples in several characteristics, such as the distribution of word types. Moreover, we observe that ChatGPT sometimes incorporates factual errors or hallucinations when adapting the text to suit a specific style. | ['Vera Demberg', 'Dongqi Pu'] | 2023-06-13 | null | null | null | null | ['style-transfer', 'text-summarization'] | ['computer-vision', 'natural-language-processing'] | [-8.60466883e-02 4.35383022e-01 1.72208428e-01 -1.85173392e-01
-7.69177556e-01 -8.53061438e-01 1.03896582e+00 2.31003195e-01
-2.68154472e-01 7.32367754e-01 7.68029332e-01 -3.21539372e-01
2.63054729e-01 -4.34098452e-01 -1.70453370e-01 -2.15893552e-01
5.32974064e-01 7.63713717e-01 -8.76634642e-02 -5.45764446e-01
6.08016133e-01 -2.30451003e-02 -1.08492982e+00 5.09759605e-01
1.12604451e+00 2.28808790e-01 3.96300405e-01 7.90170968e-01
-4.07255918e-01 9.30563509e-01 -1.19640470e+00 -7.43726432e-01
-2.01085597e-01 -7.27493823e-01 -9.49129164e-01 6.57846928e-02
3.02007318e-01 -4.11890037e-02 -2.28257954e-01 8.25363696e-01
5.91980100e-01 1.65407225e-01 8.55929673e-01 -8.80900979e-01
-1.07371151e+00 1.02850068e+00 -9.09725502e-02 6.87761679e-02
6.86297894e-01 6.07280791e-01 8.34577560e-01 -6.49590850e-01
8.51293325e-01 1.52251363e+00 6.59854650e-01 8.77250969e-01
-1.32286012e+00 -5.02011478e-01 5.48474044e-02 -2.92221099e-01
-1.24179935e+00 -5.23653924e-01 5.47119558e-01 -5.63034356e-01
7.94644058e-01 2.70064354e-01 4.96124893e-01 1.76789570e+00
3.98368746e-01 6.18539214e-01 1.20517600e+00 -4.67751443e-01
3.26846033e-01 6.53150260e-01 -1.79362997e-01 2.30332837e-01
5.85941225e-02 -1.41266555e-01 -7.23953187e-01 -2.95614988e-01
5.41276336e-01 -6.11366749e-01 -3.82692516e-01 4.38493729e-01
-1.44794261e+00 9.78381634e-01 -4.26365882e-02 5.17175972e-01
-2.73433357e-01 -1.21602692e-01 3.90272290e-01 2.73649246e-01
6.95291936e-01 1.16883230e+00 -3.47852886e-01 -5.54462552e-01
-7.33817995e-01 7.51060128e-01 1.19803548e+00 1.24462306e+00
2.98331350e-01 7.55208507e-02 -6.30775869e-01 1.01321721e+00
1.11646190e-01 7.24319756e-01 1.04827416e+00 -8.81233931e-01
6.93651259e-01 3.49633515e-01 3.21576148e-01 -1.32058930e+00
-1.84669331e-01 -2.55595893e-01 -5.71783841e-01 -2.27989689e-01
6.33789361e-01 -5.46903014e-01 -3.00473541e-01 1.73361313e+00
-1.60803631e-01 -8.28904390e-01 1.79691583e-01 6.09593391e-01
8.30591440e-01 7.75885165e-01 3.90425533e-01 -1.92228615e-01
1.12347519e+00 -6.09450042e-01 -6.51040137e-01 -5.58683455e-01
7.27083206e-01 -1.06466901e+00 1.58273232e+00 1.85184151e-01
-1.27132094e+00 -5.71330428e-01 -6.17688596e-01 1.00637697e-01
-1.29088655e-01 -1.40439207e-02 2.51541525e-01 6.29900277e-01
-9.81789470e-01 5.80644250e-01 -2.41857678e-01 -6.41271651e-01
5.45171574e-02 -3.45071107e-01 -1.22724310e-01 2.99880207e-01
-1.28011799e+00 1.15005600e+00 3.43292475e-01 -2.14167118e-01
-4.50843692e-01 -3.33579123e-01 -7.63268352e-01 -1.06067710e-01
2.42185742e-01 -6.75846040e-01 1.92259657e+00 -1.10938609e+00
-1.75248957e+00 7.74979055e-01 -5.86648174e-02 -2.58006722e-01
7.37703443e-01 8.57897326e-02 -4.10203815e-01 1.00313291e-01
2.43738070e-01 5.34393013e-01 8.31781149e-01 -1.01633179e+00
-3.17504585e-01 2.26150434e-02 -8.58242586e-02 3.00558835e-01
-1.70139417e-01 2.26283774e-01 2.33907610e-01 -1.11856949e+00
-3.30118150e-01 -7.65648782e-01 -3.05866152e-01 -3.26498210e-01
-7.22143233e-01 -4.77634370e-01 3.61285269e-01 -5.61935127e-01
1.34786904e+00 -1.93719494e+00 -7.46400654e-02 3.86621766e-02
3.64153892e-01 1.96504250e-01 -2.16075763e-01 1.13612473e+00
3.59750062e-01 5.96367359e-01 9.85527262e-02 -3.79893601e-01
2.37840384e-01 -2.15527117e-02 -3.34058553e-01 -7.71485344e-02
6.09464012e-02 9.27568793e-01 -1.10913789e+00 -5.92280090e-01
-2.76539266e-01 4.36763763e-02 -5.38111031e-01 3.80864531e-01
-5.19271135e-01 4.86063063e-01 -6.00148618e-01 1.60518304e-01
-3.80600244e-03 -4.31308597e-01 8.20866972e-02 4.16368604e-01
-2.89657682e-01 5.35032570e-01 -4.76291448e-01 1.17565393e+00
-6.78087831e-01 9.54046786e-01 -3.02132934e-01 -4.66383062e-02
9.20900106e-01 5.75140297e-01 -4.32596058e-01 -5.63560307e-01
2.43543357e-01 1.42978370e-01 2.96614319e-01 -6.82214499e-01
9.34127390e-01 -4.52287227e-01 -3.65504056e-01 9.45768476e-01
-7.07868710e-02 -5.25705576e-01 3.45372438e-01 4.23721313e-01
7.36957431e-01 -3.10837507e-01 3.49826992e-01 -4.05135751e-01
1.52999178e-01 1.33593440e-01 -9.32401493e-02 1.15211809e+00
8.42670426e-02 6.34914458e-01 6.70836151e-01 -1.27485484e-01
-8.95927608e-01 -7.63821661e-01 2.57526398e-01 1.03359461e+00
-1.64364249e-01 -5.73128819e-01 -1.06102872e+00 -5.14910340e-01
-3.58032644e-01 1.41716242e+00 -5.36491692e-01 -2.99660981e-01
-3.62807989e-01 -4.19793189e-01 7.65628517e-01 3.32503676e-01
3.15760672e-01 -1.56294930e+00 -8.26683581e-01 4.17854160e-01
-6.10027194e-01 -1.00375628e+00 -8.22591126e-01 -3.07151675e-01
-6.13584757e-01 -6.72563553e-01 -7.61851907e-01 -6.38523579e-01
5.71695864e-01 -3.59696373e-02 1.32239211e+00 1.13037033e-02
2.15748757e-01 4.58262682e-01 -7.49728143e-01 -6.28644407e-01
-1.23818314e+00 4.31859374e-01 -1.36531815e-01 -3.09704006e-01
8.19659084e-02 -2.64831930e-01 -1.92751214e-01 3.84960085e-01
-1.00151598e+00 4.43194747e-01 5.69586754e-01 6.04324877e-01
-3.21996540e-01 -2.91674107e-01 6.45392299e-01 -1.08786464e+00
1.56036794e+00 -4.94095802e-01 -8.20098668e-02 6.65452629e-02
-5.31669080e-01 1.03075802e-01 7.46893287e-01 -7.78015137e-01
-1.39212024e+00 -6.50517881e-01 -8.23722929e-02 3.72177601e-01
-3.14817846e-01 6.52055740e-01 1.91867575e-01 2.44481087e-01
1.10717273e+00 2.69220918e-01 4.89609921e-03 -4.73300785e-01
4.46438670e-01 9.34668899e-01 4.34054404e-01 -8.39307368e-01
9.44002807e-01 -2.01248974e-01 -8.95274222e-01 -9.76980507e-01
-7.75372922e-01 8.92517269e-02 -2.03596100e-01 -1.74784556e-01
6.09995127e-01 -6.46316230e-01 -4.21870381e-01 6.43171787e-01
-1.52769911e+00 -7.56648540e-01 -3.36047918e-01 2.05854133e-01
-4.74662364e-01 2.22342551e-01 -7.85373151e-01 -7.54266024e-01
-3.95031303e-01 -9.12662625e-01 8.44943643e-01 3.50404620e-01
-1.27594912e+00 -1.17841268e+00 2.38370880e-01 2.40880996e-01
6.73676074e-01 9.01562646e-02 1.08468974e+00 -1.17861021e+00
-1.06545061e-01 -3.52616936e-01 -2.12612767e-02 1.54331222e-01
8.17629099e-02 6.85452521e-02 -7.17494369e-01 1.43649757e-01
6.27244636e-02 -3.81308675e-01 1.07262209e-01 -6.02986887e-02
5.69977105e-01 -1.05685115e+00 -2.54632942e-02 9.29379612e-02
8.32833290e-01 1.10507883e-01 4.63395059e-01 1.71684712e-01
3.86156023e-01 8.55590403e-01 3.60678881e-01 4.85641241e-01
2.30916381e-01 4.89296854e-01 -2.12597460e-01 3.09524536e-01
-2.38980167e-02 -8.32941830e-01 4.62038100e-01 8.06552351e-01
6.22660108e-02 -7.75478423e-01 -8.36012244e-01 3.65927041e-01
-1.66413260e+00 -1.02312422e+00 -1.10151976e-01 1.95478427e+00
1.11696291e+00 1.96228832e-01 8.88384506e-02 -2.75392354e-01
6.93591297e-01 3.04648519e-01 -1.20313771e-01 -7.20742226e-01
-1.46447033e-01 -1.15620811e-02 1.28477827e-01 4.07753468e-01
-2.43153006e-01 8.93610895e-01 7.47019243e+00 7.75110960e-01
-1.01848257e+00 -1.48423553e-01 7.83613026e-01 1.51109084e-01
-4.53351498e-01 -2.69687861e-01 -5.25692284e-01 6.95788264e-01
1.18888617e+00 -7.36573339e-01 3.47648650e-01 5.67806542e-01
4.47796404e-01 -2.50214022e-02 -1.09679019e+00 5.76411843e-01
2.80185312e-01 -1.23086512e+00 3.12687904e-01 -2.19673604e-01
6.80109501e-01 -3.16211104e-01 3.01287603e-03 3.84118021e-01
4.46522951e-01 -1.10812140e+00 1.17847180e+00 3.34325701e-01
5.96416116e-01 -3.94291699e-01 7.33825445e-01 7.66545057e-01
-4.49235022e-01 2.13832736e-01 -1.23700678e-01 -5.13904333e-01
2.72439957e-01 1.19968154e-01 -1.16149521e+00 -1.15406662e-01
2.64298320e-01 1.56674698e-01 -8.89105380e-01 5.14139175e-01
-4.98882592e-01 7.54315197e-01 1.64172336e-01 -6.63934708e-01
1.88128665e-01 -3.40736215e-03 6.93271041e-01 1.38224661e+00
3.39074463e-01 5.96823804e-02 1.20242484e-01 1.12492085e+00
-2.44957097e-02 4.70446140e-01 -5.53963482e-01 -6.81221843e-01
5.93958795e-01 1.02733910e+00 -5.16341031e-01 -6.04904830e-01
-1.62609980e-01 9.85185385e-01 1.72148526e-01 4.46356356e-01
-4.23816264e-01 -4.14473981e-01 2.12463379e-01 3.76624554e-01
-1.57434225e-01 -8.95912051e-02 -4.41661924e-01 -1.00620902e+00
-1.86715171e-01 -1.36123896e+00 -8.32166299e-02 -1.26917303e+00
-1.60251820e+00 9.10887122e-01 1.47520944e-01 -8.17474186e-01
-6.11358881e-01 -3.57036114e-01 -8.95479262e-01 1.01634598e+00
-5.60412347e-01 -9.48895574e-01 -1.44341558e-01 2.72939086e-01
8.51751089e-01 -1.95982948e-01 6.64791763e-01 -3.01740944e-01
-2.88350940e-01 6.74054980e-01 -3.38092148e-01 1.75204784e-01
8.86785567e-01 -1.13786077e+00 8.14451396e-01 5.30905664e-01
3.01632099e-02 8.75905931e-01 1.34731221e+00 -6.85375810e-01
-7.40383029e-01 -9.15410280e-01 1.50834918e+00 -7.23815978e-01
8.72154832e-01 -2.41244540e-01 -8.05047691e-01 7.25621879e-01
6.58644140e-01 -8.03714514e-01 6.50674284e-01 3.50420065e-02
-3.20834070e-01 5.63724458e-01 -1.10430861e+00 1.15066767e+00
9.26906586e-01 -5.19350410e-01 -9.61295128e-01 5.04980147e-01
7.73347676e-01 -3.98071438e-01 -4.26957130e-01 -4.07376766e-01
3.97803098e-01 -9.78667021e-01 4.56293106e-01 -7.71042705e-01
8.65106285e-01 1.12993762e-01 1.45109341e-01 -1.83046818e+00
-4.19351965e-01 -1.04983234e+00 4.26237524e-01 1.42402911e+00
6.40130997e-01 -8.24404657e-01 2.48156235e-01 9.07098711e-01
-2.39067554e-01 -4.45374638e-01 -3.69776338e-01 -5.24812818e-01
2.65037566e-01 -1.57183617e-01 4.28049028e-01 7.68361151e-01
5.72687387e-01 7.77792215e-01 -3.25153768e-01 -3.88692468e-01
3.75143662e-02 -1.45911247e-01 9.51814532e-01 -1.21135819e+00
-5.35941839e-01 -6.00133002e-01 9.83721390e-02 -9.48207140e-01
6.42073825e-02 -6.98929846e-01 4.01938409e-01 -1.23197973e+00
1.82488903e-01 -1.01692677e-01 5.26262164e-01 2.67566264e-01
-3.96973819e-01 1.43823609e-01 4.94427383e-01 2.26802379e-01
-3.61417890e-01 4.21109647e-01 1.40989578e+00 6.36890009e-02
-4.68675613e-01 2.64058679e-01 -1.40368700e+00 8.89956415e-01
9.26246822e-01 -4.48594540e-01 -4.31565315e-01 -4.02448416e-01
5.66397190e-01 2.05491677e-01 2.08421990e-01 -7.72093534e-01
-4.04784232e-02 -3.89221966e-01 2.78115362e-01 1.66592747e-01
-1.72484919e-01 -8.68792906e-02 8.63066763e-02 1.86519891e-01
-9.54625070e-01 4.78891611e-01 1.97937220e-01 3.47454369e-01
-1.90983154e-02 -3.99095744e-01 5.79481602e-01 -4.61068422e-01
-1.19065464e-01 -3.22603993e-02 -1.23674822e+00 6.29876494e-01
5.91299713e-01 -2.61758626e-01 -4.71460968e-01 -9.27981436e-01
-3.30464125e-01 8.33859481e-03 6.77638650e-01 5.24633348e-01
3.88606608e-01 -1.05257142e+00 -8.99043620e-01 -7.20187426e-02
3.21573079e-01 -2.47058734e-01 1.15206510e-01 6.99716330e-01
-4.01847005e-01 2.82678276e-01 1.68657210e-02 -1.15400225e-01
-9.26232219e-01 2.30085760e-01 6.87651485e-02 -3.01614255e-01
-5.82742214e-01 7.13911653e-01 2.94213057e-01 -3.28371614e-01
-1.34484172e-01 -3.83711159e-01 -8.61284584e-02 9.95951220e-02
6.93154812e-01 2.64403254e-01 -3.54476064e-01 -5.86789191e-01
1.64894849e-01 2.95217019e-02 -3.02129358e-01 -4.53369737e-01
7.71773219e-01 -1.72722235e-01 2.22310573e-01 7.16090143e-01
8.26021314e-01 3.34524304e-01 -9.18166816e-01 -1.44818932e-01
4.38966788e-02 -3.81968886e-01 -5.97364068e-01 -1.07598472e+00
-3.00479472e-01 5.41581869e-01 -3.13851058e-01 5.35869360e-01
4.33644533e-01 2.39968017e-01 8.88584912e-01 5.71285129e-01
3.90584350e-01 -1.21288919e+00 4.07880813e-01 7.05738664e-01
1.23095894e+00 -8.87373328e-01 -3.12478244e-01 1.94516871e-02
-1.11728454e+00 1.02230048e+00 5.69698453e-01 1.96111351e-01
-4.29841205e-02 -1.79494381e-01 5.31885087e-01 -6.34043440e-02
-1.11563325e+00 2.59473294e-01 2.27960512e-01 7.85157800e-01
7.62482524e-01 1.17437296e-01 -4.69133496e-01 6.32376075e-01
-9.39718604e-01 -2.42443815e-01 9.72273290e-01 5.25464356e-01
-5.62868118e-01 -8.02880943e-01 -3.68006468e-01 3.58134180e-01
-2.21584767e-01 -2.55177945e-01 -8.94255340e-01 7.12421596e-01
-2.21183449e-01 1.19199955e+00 -9.92551073e-02 -1.87668040e-01
3.42768222e-01 2.64040381e-01 1.95852235e-01 -1.14121842e+00
-9.60988939e-01 -1.33534074e-01 4.28279102e-01 -1.33005053e-01
1.02116100e-01 -7.70611048e-01 -8.68124306e-01 -3.92382294e-01
-1.17887117e-01 5.01755178e-01 3.14153343e-01 1.18023741e+00
1.33316025e-01 1.81067601e-01 3.89528841e-01 -7.49525547e-01
-8.11786413e-01 -1.49220884e+00 -4.27731097e-01 5.28444171e-01
4.82561737e-02 4.56124777e-03 -5.36868215e-01 5.75797968e-02] | [11.768999099731445, 8.947772026062012] |
cba50433-edab-47ba-abae-a6991030c8af | multidimensional-evaluation-for-text-style | 2304.13462 | null | https://arxiv.org/abs/2304.13462v1 | https://arxiv.org/pdf/2304.13462v1.pdf | Multidimensional Evaluation for Text Style Transfer Using ChatGPT | We investigate the potential of ChatGPT as a multidimensional evaluator for the task of \emph{Text Style Transfer}, alongside, and in comparison to, existing automatic metrics as well as human judgements. We focus on a zero-shot setting, i.e. prompting ChatGPT with specific task instructions, and test its performance on three commonly-used dimensions of text style transfer evaluation: style strength, content preservation, and fluency. We perform a comprehensive correlation analysis for two transfer directions (and overall) at different levels. Compared to existing automatic metrics, ChatGPT achieves competitive correlations with human judgments. These preliminary results are expected to provide a first glimpse into the role of large language models in the multidimensional evaluation of stylized text generation. | ['Malvina Nissim', 'Antonio Toral', 'Huiyuan Lai'] | 2023-04-26 | null | null | null | null | ['style-transfer', 'text-style-transfoer'] | ['computer-vision', 'natural-language-processing'] | [ 2.11283088e-01 5.20529523e-02 -2.46859007e-02 -3.14479321e-01
-9.82322633e-01 -9.24877107e-01 9.00300622e-01 1.54598475e-01
-5.91864228e-01 5.24006009e-01 7.87891448e-01 -2.66863227e-01
-2.00462788e-02 -3.20654064e-01 1.69211984e-01 -2.17040092e-01
5.63079715e-01 8.42822433e-01 -5.54255843e-02 -4.20733899e-01
5.95579088e-01 2.28795782e-01 -1.15506256e+00 3.07367474e-01
1.03100955e+00 6.40087306e-01 1.18052371e-01 8.19786727e-01
-2.36434281e-01 6.60084724e-01 -9.84072745e-01 -8.93550992e-01
-2.34205842e-01 -6.72670722e-01 -1.07969189e+00 -3.81234474e-02
5.64279139e-01 -2.22945362e-01 2.17676729e-01 8.47194493e-01
8.52856457e-01 1.52447522e-01 9.76880312e-01 -8.33563268e-01
-8.83957505e-01 7.90594041e-01 -2.67193109e-01 3.92967463e-01
6.04267836e-01 3.35540235e-01 1.19142687e+00 -8.02907884e-01
8.24988008e-01 1.56581664e+00 6.74499393e-01 6.54480934e-01
-1.51819503e+00 -4.03493851e-01 -7.74210468e-02 -2.40574449e-01
-7.51263678e-01 -6.55004621e-01 5.46926916e-01 -7.22893715e-01
9.10464525e-01 2.58724272e-01 2.50516623e-01 1.41129720e+00
1.21582709e-01 7.09768236e-01 1.67292070e+00 -3.90574247e-01
4.78162356e-02 3.88612956e-01 3.22425991e-01 3.66655737e-01
-3.41148004e-02 9.04724076e-02 -1.01175082e+00 -5.90919657e-03
5.39411545e-01 -9.50966775e-01 -1.71172619e-01 1.39882013e-01
-1.43776655e+00 6.64169431e-01 -1.74663812e-01 5.38552046e-01
-2.03899086e-01 -2.00661153e-01 7.55568802e-01 5.98641872e-01
7.87707090e-01 1.01366186e+00 -5.30266821e-01 -9.05721843e-01
-9.81714606e-01 3.67718607e-01 8.78049493e-01 8.83852601e-01
3.83522630e-01 1.83700055e-01 -8.67141902e-01 1.24386728e+00
-1.47173017e-01 5.79381883e-01 7.24524140e-01 -1.04021716e+00
6.69720054e-01 2.86721736e-01 8.92272517e-02 -7.36877739e-01
-4.73809123e-01 -3.24239850e-01 -3.52506071e-01 1.96714804e-01
7.41909564e-01 -4.39383984e-01 -2.26924077e-01 2.02222991e+00
2.07303632e-02 -7.17582822e-01 -2.38819003e-01 6.74849510e-01
6.85466290e-01 3.92481267e-01 2.57007897e-01 -3.11172694e-01
1.26868939e+00 -7.55789399e-01 -6.34719729e-01 -2.29933545e-01
9.35837090e-01 -1.22191572e+00 1.85309207e+00 3.67230833e-01
-1.46688235e+00 -7.23859012e-01 -7.03253090e-01 -1.83410525e-01
-1.22132197e-01 1.11402534e-01 3.31215769e-01 8.95834744e-01
-1.18616521e+00 7.57208765e-01 -3.83238226e-01 -4.13384795e-01
-1.76012982e-02 -7.27790669e-02 -1.54257178e-01 3.12103331e-01
-9.70768452e-01 1.25596535e+00 3.76770422e-02 -4.13826585e-01
-3.95333439e-01 -8.00213754e-01 -6.04243040e-01 -1.25609338e-01
-2.24636607e-02 -6.73685312e-01 1.65577805e+00 -9.09679055e-01
-1.94497395e+00 1.21174538e+00 -3.39436121e-02 -2.75368448e-02
5.92243552e-01 -3.68197441e-01 -2.73232967e-01 -8.00714344e-02
2.73645997e-01 6.27890110e-01 4.96442080e-01 -9.47046757e-01
-4.50564235e-01 -2.13500440e-01 -1.29582420e-01 5.04826844e-01
-5.94870508e-01 5.23117661e-01 6.57642111e-02 -1.02665794e+00
-4.11823720e-01 -1.02347171e+00 1.79960191e-01 -4.25438762e-01
-2.90835202e-01 -4.42013890e-01 3.49846750e-01 -9.36350822e-01
1.35138392e+00 -1.92457962e+00 4.76308167e-01 -1.29902110e-01
2.09413558e-01 2.32154623e-01 -3.42321754e-01 4.41136360e-01
2.38247693e-01 2.66315907e-01 -5.68112582e-02 -6.64949834e-01
3.53034526e-01 -3.58886987e-01 1.09330721e-01 -5.42280171e-03
2.17665553e-01 1.19082487e+00 -8.88360977e-01 -5.96574426e-01
-2.30221357e-02 1.89995125e-01 -5.51402390e-01 1.74628362e-01
-1.31127819e-01 4.21168000e-01 -1.46380499e-01 1.69795021e-01
1.86251700e-01 -3.85986269e-02 8.09940547e-02 1.30573705e-01
-3.36619198e-01 6.34275734e-01 -5.93274713e-01 1.74554634e+00
-7.09816217e-01 8.75823975e-01 -1.22374319e-01 -1.76871181e-01
1.06617665e+00 3.34754735e-01 -3.68299261e-02 -7.21216619e-01
2.76552230e-01 1.05953038e-01 3.17316681e-01 -1.76749751e-01
8.09161901e-01 -4.90405291e-01 -2.93411136e-01 9.84872997e-01
3.80825102e-01 -4.42807734e-01 3.44909728e-01 1.98272303e-01
9.26579773e-01 1.35310546e-01 2.55499244e-01 -7.48557150e-01
2.25101888e-01 -7.73492455e-02 7.39357099e-02 6.43390179e-01
-2.59047449e-01 5.37273169e-01 7.35592604e-01 1.00393407e-01
-1.06416631e+00 -1.06412721e+00 7.87488073e-02 1.63364947e+00
-3.06717515e-01 -5.43077290e-01 -1.12852192e+00 -7.34889328e-01
-3.03103030e-01 1.14627993e+00 -6.15044355e-01 -2.31966645e-01
-4.60214227e-01 -5.86264610e-01 7.48095572e-01 6.13491714e-01
2.21526220e-01 -1.30729151e+00 -4.95717585e-01 7.68409595e-02
-4.15918529e-01 -1.02007139e+00 -7.79959202e-01 -9.49961618e-02
-7.79173970e-01 -5.94570458e-01 -9.46982205e-01 -7.74952173e-01
-1.17136352e-01 2.98781674e-02 1.69040072e+00 -1.58648759e-01
1.53770760e-01 5.50788045e-01 -3.93485010e-01 -4.05172020e-01
-8.60553980e-01 4.57866073e-01 5.23163239e-04 -4.77714777e-01
3.33945900e-01 -5.29636085e-01 -4.68890280e-01 4.72171992e-01
-5.28715074e-01 2.01585144e-01 5.90166986e-01 6.45968199e-01
-7.31522515e-02 -5.49263656e-01 7.80521631e-01 -8.66831779e-01
1.50367188e+00 -1.23403698e-01 -7.97313377e-02 1.54157415e-01
-7.79647708e-01 1.10637257e-02 4.06877577e-01 -4.46573436e-01
-1.45964217e+00 -5.70501745e-01 -9.30895656e-02 2.18354482e-02
-3.19658399e-01 3.42586964e-01 1.54438557e-03 3.97606082e-02
9.77693379e-01 -1.40289381e-01 6.92245439e-02 -4.87485707e-01
5.93774855e-01 6.87916756e-01 3.56169760e-01 -9.65645492e-01
6.94437444e-01 -1.76068589e-01 -3.53325874e-01 -8.71746838e-01
-8.91716123e-01 -1.67718440e-01 -8.05481076e-01 -2.18802527e-01
8.86061192e-01 -6.46090746e-01 -5.76515973e-01 4.37344581e-01
-1.23758483e+00 -8.32491159e-01 -3.31621349e-01 4.64861721e-01
-9.68460381e-01 2.99137384e-01 -1.01286805e+00 -5.73058605e-01
-5.42009115e-01 -9.19271886e-01 1.17142463e+00 -1.39338389e-01
-1.19358623e+00 -1.35755694e+00 4.83745396e-01 5.16658783e-01
5.22291660e-01 8.19699392e-02 1.11195827e+00 -7.84110188e-01
1.80154234e-01 3.44327278e-02 -1.97219878e-01 5.08449495e-01
-2.52401624e-02 -1.25188604e-01 -9.68157828e-01 -4.18496206e-02
2.42289796e-01 -7.36206472e-01 4.94417131e-01 2.11545125e-01
5.96729159e-01 -9.84481275e-02 2.86809921e-01 2.16493234e-01
9.24430132e-01 1.73260383e-02 4.36439276e-01 1.86244383e-01
6.53929114e-01 8.83347988e-01 6.18050098e-01 3.90500605e-01
3.02279055e-01 8.05899382e-01 -4.00461584e-01 5.66421486e-02
-4.17738348e-01 -2.21449584e-01 5.07305086e-01 1.22627342e+00
-1.41923800e-01 -4.88342375e-01 -9.21596467e-01 1.81321725e-01
-1.33608317e+00 -8.79023254e-01 -4.26141471e-02 2.16362333e+00
9.43543851e-01 2.32434466e-01 5.45281649e-01 1.71204045e-01
7.50367641e-01 2.02234417e-01 -1.98029399e-01 -9.27734494e-01
-2.28766307e-01 3.91795993e-01 -2.18097884e-02 5.90172350e-01
-6.74658298e-01 1.22102702e+00 7.43379068e+00 8.97203386e-01
-9.00033355e-01 1.50876090e-01 7.42477000e-01 -1.60397381e-01
-5.13048291e-01 -3.10378969e-01 -6.28022373e-01 3.48926634e-01
9.97059822e-01 -3.99450183e-01 6.34526253e-01 4.55499887e-01
2.48764277e-01 5.22373579e-02 -1.35810184e+00 7.18171120e-01
1.94548145e-01 -8.00289690e-01 1.46043347e-02 -8.73319954e-02
8.62644076e-01 -8.94637406e-02 2.91542441e-01 4.69249278e-01
3.60004008e-01 -8.66359949e-01 9.60490346e-01 3.50748926e-01
1.07920229e+00 -4.73503351e-01 4.31104332e-01 1.92793816e-01
-6.79348409e-01 2.06241578e-01 -4.57255431e-02 -2.96548218e-01
4.00347263e-01 2.50379920e-01 -7.88018286e-01 1.33319229e-01
2.12014169e-01 4.82029915e-01 -7.34863162e-01 5.26163578e-01
-2.23039761e-01 8.23769510e-01 1.55756250e-01 -2.79152066e-01
1.14274055e-01 -7.79540464e-02 6.23319089e-01 1.58816862e+00
3.32642317e-01 -2.28246357e-02 -1.16166890e-01 8.58811855e-01
5.01911808e-03 5.92235625e-01 -5.11500657e-01 -2.27418408e-01
4.09137070e-01 1.28195739e+00 -6.98267281e-01 -3.96056443e-01
-3.20106626e-01 1.06939507e+00 6.17048264e-01 2.82482356e-01
-4.94462818e-01 -3.77239645e-01 5.83025336e-01 -4.44115512e-02
-1.95270315e-01 -3.72805417e-01 -9.30447996e-01 -8.90405238e-01
-1.27274066e-01 -1.04088485e+00 1.78705156e-01 -9.60482359e-01
-1.46440101e+00 6.51796520e-01 -1.28729073e-02 -9.21725154e-01
-4.52739775e-01 -7.54841328e-01 -8.31117868e-01 1.13045239e+00
-7.94194996e-01 -9.34096813e-01 -2.39171356e-01 3.50677550e-01
7.34866560e-01 -2.35143498e-01 8.58557224e-01 -5.36656231e-02
-3.85431617e-01 8.44674468e-01 -1.70503408e-01 -1.00870185e-01
1.08199728e+00 -1.54457080e+00 7.53594637e-01 4.14910585e-01
1.34762183e-01 5.08866131e-01 8.17295253e-01 -6.46019101e-01
-7.34158337e-01 -5.95847845e-01 9.58075047e-01 -8.14488292e-01
9.08414423e-01 -3.19174916e-01 -6.71227455e-01 3.90602201e-01
6.13557279e-01 -1.01781499e+00 8.82156372e-01 7.08858907e-01
-2.61278152e-01 3.53541613e-01 -9.05187905e-01 7.95899510e-01
1.36792171e+00 -7.41868436e-01 -7.22443581e-01 2.48478472e-01
7.77619600e-01 -2.48392045e-01 -1.17156291e+00 1.20088253e-02
6.05417907e-01 -1.18910444e+00 5.94122589e-01 -6.50668025e-01
7.09255457e-01 3.92760515e-01 5.67854382e-02 -1.95006239e+00
-7.01749086e-01 -9.03160930e-01 4.92930233e-01 1.63630128e+00
6.99619114e-01 -4.10601676e-01 4.69921321e-01 5.85774541e-01
-3.14512521e-01 -5.24690211e-01 -6.89774096e-01 -7.91390479e-01
6.50209129e-01 -3.50798279e-01 3.95809710e-01 9.27700043e-01
2.76499957e-01 1.13296235e+00 -3.05890352e-01 -7.03368604e-01
3.82075697e-01 -1.94640622e-01 6.96618974e-01 -1.45625877e+00
-3.15552890e-01 -1.08448660e+00 -7.70824850e-02 -6.71621621e-01
3.41992319e-01 -1.00935876e+00 -9.57299769e-02 -1.27242649e+00
1.42784074e-01 -2.29412466e-01 -7.26273805e-02 -4.45720740e-02
-3.76368761e-01 2.38277078e-01 5.44465780e-01 1.12063453e-01
-5.26458442e-01 4.69044596e-01 1.57076263e+00 1.86787054e-01
-3.79966229e-01 -1.28087392e-02 -1.03140020e+00 6.43682539e-01
1.27662003e+00 -1.15231045e-01 -4.97820437e-01 -6.88377798e-01
1.68869719e-01 2.12615237e-01 -3.28951180e-02 -9.78572786e-01
-2.41545618e-01 -1.93240419e-01 1.72805801e-01 -5.63193411e-02
1.69299930e-01 -3.19819674e-02 -3.35289687e-01 -8.87255892e-02
-8.04215670e-01 5.39004326e-01 2.20293179e-01 2.27391664e-02
-2.81610452e-02 -2.55395114e-01 7.79937148e-01 -1.73009336e-01
-1.79829270e-01 -1.01029031e-01 -5.19308507e-01 7.14089870e-01
3.10765117e-01 -2.20049426e-01 -5.77500641e-01 -4.84187037e-01
-6.35434389e-01 -7.95445591e-02 6.10780180e-01 5.78898609e-01
1.99012265e-01 -1.42336202e+00 -8.66368294e-01 -2.34894335e-01
1.76688790e-01 -6.77117765e-01 6.30602613e-02 8.96222353e-01
-1.71062633e-01 3.33886683e-01 -3.83000672e-01 -3.03971052e-01
-1.23583806e+00 4.05123323e-01 4.37269025e-02 -4.07599270e-01
-1.83548465e-01 6.98251486e-01 4.71913338e-01 -2.94274658e-01
3.56681012e-02 -1.88284218e-01 -1.96883783e-01 2.46898532e-01
5.61899304e-01 7.84546077e-01 1.43188015e-02 -6.28839791e-01
1.02315344e-01 4.68684644e-01 1.85230970e-02 -9.56492841e-01
7.95689404e-01 -2.86466062e-01 -7.01237395e-02 9.98824298e-01
1.01009059e+00 1.68311536e-01 -9.05848980e-01 -6.97814152e-02
2.10000098e-01 -2.89420009e-01 -2.48891070e-01 -1.19622219e+00
-3.92230272e-01 9.42205131e-01 3.21616292e-01 2.12899476e-01
7.41555989e-01 -3.25211175e-02 6.20838344e-01 3.56160939e-01
7.75477290e-02 -1.46848297e+00 5.33049762e-01 7.63186276e-01
1.05618465e+00 -9.99731004e-01 -3.50750089e-01 -1.44136786e-01
-1.21038413e+00 9.56024051e-01 7.37844348e-01 2.21105009e-01
2.41338253e-01 1.16920397e-01 3.15891355e-01 -1.18434504e-01
-7.93762088e-01 -1.31929711e-01 5.27358115e-01 6.08641803e-01
1.15032673e+00 2.32881874e-01 -5.87100089e-01 4.90639359e-01
-9.02544737e-01 -9.11777169e-02 2.05229908e-01 4.81164664e-01
-5.06881833e-01 -1.07859790e+00 -1.37001663e-01 5.61096072e-01
-3.68376762e-01 -3.53796989e-01 -8.84403884e-01 6.16365492e-01
-3.08173805e-01 1.05125415e+00 -3.06608304e-02 -5.33563673e-01
4.90494460e-01 4.20143962e-01 5.44225454e-01 -7.24668324e-01
-9.72540021e-01 4.52808551e-02 6.25710607e-01 -2.75628209e-01
-1.08373426e-01 -9.53912437e-01 -5.01038790e-01 -4.11367863e-01
-1.29784286e-01 1.26144484e-01 3.73662293e-01 1.00371027e+00
1.75900292e-02 3.26824456e-01 4.67532992e-01 -9.35985386e-01
-5.69697797e-01 -1.50385666e+00 -3.98469269e-01 7.09829032e-01
-1.30888879e-01 -4.53171253e-01 -3.32818598e-01 -1.26460299e-01] | [11.481846809387207, 9.649511337280273] |
17164e0e-2366-4de0-9a1b-a7991613f6be | overflow-putting-flows-on-top-of-neural | 2211.06892 | null | https://arxiv.org/abs/2211.06892v2 | https://arxiv.org/pdf/2211.06892v2.pdf | OverFlow: Putting flows on top of neural transducers for better TTS | Neural HMMs are a type of neural transducer recently proposed for sequence-to-sequence modelling in text-to-speech. They combine the best features of classic statistical speech synthesis and modern neural TTS, requiring less data and fewer training updates, and are less prone to gibberish output caused by neural attention failures. In this paper, we combine neural HMM TTS with normalising flows for describing the highly non-Gaussian distribution of speech acoustics. The result is a powerful, fully probabilistic model of durations and acoustics that can be trained using exact maximum likelihood. Experiments show that a system based on our proposal needs fewer updates than comparable methods to produce accurate pronunciations and a subjective speech quality close to natural speech. Please see https://shivammehta25.github.io/OverFlow/ for audio examples and code. | ['Gustav Eje Henter', 'Éva Székely', 'Jonas Beskow', 'Harm Lameris', 'Ambika Kirkland', 'Shivam Mehta'] | 2022-11-13 | null | null | null | null | ['normalising-flows', 'text-to-speech-synthesis'] | ['methodology', 'speech'] | [-2.54145619e-02 1.69158012e-01 2.49816738e-02 -5.06346285e-01
-1.04249001e+00 -4.45211619e-01 3.42493504e-01 -2.04306111e-01
-1.83636874e-01 6.92898750e-01 3.03155899e-01 -8.55485678e-01
3.52876902e-01 -1.20373711e-01 -6.47037983e-01 -7.07009912e-01
1.10568486e-01 4.36793566e-01 4.54328567e-01 4.50539626e-02
-6.35028854e-02 3.41720939e-01 -1.56916726e+00 1.21516310e-01
4.42369252e-01 9.46301281e-01 6.37339115e-01 1.35710418e+00
-1.22740433e-01 7.81485915e-01 -9.04773831e-01 -1.68528125e-01
-9.25487429e-02 -5.87142885e-01 -3.99731278e-01 -2.75692850e-01
2.50621021e-01 -3.20178270e-01 -6.08985066e-01 8.73875558e-01
8.20906997e-01 3.47587466e-01 6.67398453e-01 -9.33028221e-01
-4.68223035e-01 8.91628444e-01 7.69356564e-02 3.93377692e-01
3.25887591e-01 2.59733260e-01 7.42936909e-01 -6.76476955e-01
-5.81028759e-02 1.44589376e+00 5.89682937e-01 7.59491682e-01
-1.10037792e+00 -7.62429714e-01 -1.72579810e-01 2.43939430e-01
-1.29514229e+00 -1.24728811e+00 3.52453381e-01 -1.20379113e-01
1.38067281e+00 5.76445222e-01 3.31638098e-01 1.48220563e+00
1.99714094e-01 8.19532275e-01 6.97329044e-01 -7.52786100e-01
2.81280369e-01 4.27341014e-02 -1.13658339e-01 3.67018253e-01
-5.00812232e-01 3.59232783e-01 -7.04779923e-01 -1.83172241e-01
6.22015357e-01 -4.55276132e-01 -3.29768270e-01 3.60743791e-01
-8.85574520e-01 6.59209013e-01 -3.23120713e-01 3.55769634e-01
-4.27851915e-01 5.13814509e-01 5.48004746e-01 1.85629353e-01
4.20079350e-01 9.46518183e-02 -6.93733454e-01 -8.56083989e-01
-1.01059747e+00 9.75281671e-02 9.79979753e-01 1.23830509e+00
2.03196809e-01 8.72222185e-01 -7.13432059e-02 1.21430969e+00
4.50891376e-01 8.80824685e-01 8.91594052e-01 -1.22678077e+00
1.79941148e-01 -5.69894671e-01 -1.39868796e-01 -3.93421769e-01
-1.90508857e-01 -2.36801237e-01 -6.30138099e-01 -3.00414767e-02
4.27384943e-01 -2.60399163e-01 -1.20890987e+00 1.72767019e+00
-1.04839087e-01 1.94302723e-01 -8.94472189e-03 3.35383385e-01
5.20684659e-01 1.33166492e+00 -8.99813399e-02 -4.33488846e-01
1.25911796e+00 -7.52191842e-01 -1.27242684e+00 -2.40318000e-01
4.38049704e-01 -1.07232130e+00 1.14883053e+00 5.01890004e-01
-1.46606624e+00 -7.17732310e-01 -8.58392000e-01 3.00009623e-02
-2.41876662e-01 -5.56733534e-02 2.09277645e-01 9.96587276e-01
-1.39464772e+00 6.35853767e-01 -1.02064204e+00 -9.74497572e-02
-1.42060444e-01 3.17862779e-01 2.36584455e-01 3.84796679e-01
-1.29444993e+00 9.37508345e-01 5.68827808e-01 -3.16870399e-02
-8.12545538e-01 -4.67757344e-01 -8.51439118e-01 2.79969543e-01
2.93300122e-01 -2.31339201e-01 2.30994272e+00 -5.63315988e-01
-2.30707479e+00 3.35801244e-02 -6.50654435e-01 -6.72326684e-01
-2.00493876e-02 -3.40556502e-01 -6.80600822e-01 -6.94087893e-02
-4.43988293e-01 6.86985493e-01 9.49711204e-01 -8.12574983e-01
-6.38135970e-01 2.49497652e-01 -8.63026619e-01 -4.64698672e-03
-2.72180140e-01 2.06640363e-01 -5.19089997e-01 -6.60552979e-01
-1.71450123e-01 -8.03410113e-01 -6.97572976e-02 -5.98272502e-01
-4.89764482e-01 -4.84805763e-01 7.47294486e-01 -1.00494683e+00
1.46199179e+00 -2.09230018e+00 -1.79993808e-01 -7.64699578e-02
-3.15801233e-01 6.08939409e-01 -5.73598146e-02 5.47305048e-01
8.97722766e-02 1.74315080e-01 -9.55770090e-02 -7.44794667e-01
3.65341365e-01 3.47680390e-01 -4.87899601e-01 1.93400964e-01
-3.63912918e-02 8.59919250e-01 -6.99812949e-01 -4.52933431e-01
3.76681596e-01 5.27593553e-01 -2.25855082e-01 5.61747730e-01
-2.85098672e-01 2.89784163e-01 1.62600771e-01 3.16487074e-01
1.46762043e-01 3.27323735e-01 -2.83271261e-03 1.85557052e-01
-1.32850543e-01 9.20625269e-01 -9.06767726e-01 1.52529752e+00
-6.08098388e-01 1.07511365e+00 7.21219555e-02 -5.59679747e-01
8.95442247e-01 1.00643218e+00 -9.47243422e-02 -6.12249732e-01
4.21153992e-01 3.96000832e-01 2.32863024e-01 -4.04434800e-01
4.92437661e-01 -2.87008613e-01 9.40248966e-02 2.68086225e-01
3.46381456e-01 -6.81103706e-01 9.53366309e-02 -8.02842304e-02
8.65408242e-01 8.62763077e-02 1.79325610e-01 4.89924476e-02
1.24425320e-02 -6.99773192e-01 4.03093398e-01 8.52080882e-01
-2.26250682e-02 5.65158367e-01 1.63010761e-01 1.87451303e-01
-1.57270992e+00 -1.08587193e+00 -3.40243317e-02 1.18165016e+00
-6.92081571e-01 -3.79994690e-01 -1.01149666e+00 1.00617260e-01
-4.45667714e-01 1.34901381e+00 -1.68184061e-02 -3.90664227e-02
-6.73970640e-01 -2.33610049e-01 1.00350761e+00 4.64597344e-01
-4.49574590e-01 -1.32775724e+00 -1.54366583e-01 5.59220552e-01
-1.65932968e-01 -1.24064732e+00 -7.89802611e-01 5.72390616e-01
-7.72499800e-01 -1.14564613e-01 -9.93571281e-01 -7.15607882e-01
-1.39388010e-01 -2.03762114e-01 8.85728598e-01 -2.44516298e-01
1.45217210e-01 -1.75194927e-02 -2.92052835e-01 -8.03108633e-01
-1.29593015e+00 7.66649395e-02 5.03631353e-01 -2.06261680e-01
2.44059443e-01 -7.80171931e-01 -5.33165857e-02 1.97465822e-01
-8.06725383e-01 5.79066128e-02 5.33834875e-01 6.30981505e-01
1.87721997e-01 7.03774169e-02 7.10729480e-01 -4.18203294e-01
7.78507411e-01 -3.05213392e-01 -6.29576266e-01 -2.08658203e-01
-6.20515227e-01 8.35927725e-02 8.61309111e-01 -8.06802630e-01
-1.06197047e+00 -2.95292344e-02 -8.45668852e-01 -6.50395691e-01
-7.17343628e-01 3.15560251e-01 -1.51554376e-01 5.90235174e-01
5.24581909e-01 4.12301451e-01 6.17761351e-02 -7.26893663e-01
4.28542912e-01 1.26444685e+00 8.15582454e-01 -2.61975884e-01
5.55309594e-01 -2.52480954e-01 -5.54996073e-01 -1.14481068e+00
-3.25201988e-01 -4.90079403e-01 -4.53666478e-01 -6.77091330e-02
4.96579319e-01 -6.02272093e-01 -7.99208283e-01 6.87465310e-01
-1.44822574e+00 -6.35205269e-01 -3.38225633e-01 8.54031622e-01
-8.99670959e-01 4.95223612e-01 -1.01638794e+00 -1.44047856e+00
-3.03896993e-01 -9.13071215e-01 8.03673089e-01 2.79341072e-01
-4.87715542e-01 -8.89112294e-01 1.26433581e-01 -9.97284129e-02
6.61059797e-01 -4.16730523e-01 6.89752042e-01 -7.88168132e-01
-1.20178245e-01 -2.41481021e-01 3.43168736e-01 8.80413353e-01
2.17225730e-01 2.54576772e-01 -1.47369730e+00 -7.54989386e-02
1.67211711e-01 9.33581789e-04 3.96008253e-01 8.16222787e-01
1.08156705e+00 -7.21681178e-01 -5.49822450e-02 3.35975468e-01
8.09488893e-01 8.68546665e-01 6.69288933e-01 -3.93744588e-01
5.31066358e-01 4.73423719e-01 2.29249835e-01 4.33423102e-01
1.00768894e-01 7.44828224e-01 8.07992294e-02 -2.86298972e-02
-2.01342970e-01 -2.44960681e-01 8.22789550e-01 1.74655008e+00
3.22912157e-01 -7.43278980e-01 -6.95454359e-01 7.22956002e-01
-1.58110249e+00 -9.92103636e-01 -2.25512087e-01 2.18394279e+00
1.17838550e+00 5.72950840e-01 3.19844961e-01 2.62172818e-01
8.24665189e-01 6.32988885e-02 -2.88930476e-01 -1.13262522e+00
7.49920458e-02 4.98450577e-01 5.77001452e-01 8.53185833e-01
-6.15370393e-01 1.06226528e+00 6.64497757e+00 1.36991525e+00
-9.38488841e-01 2.71735787e-01 3.83518428e-01 -2.72860974e-01
-2.14470699e-01 -1.66115031e-01 -8.87495637e-01 6.13987148e-01
2.30986500e+00 2.84698755e-02 4.84962821e-01 4.91731972e-01
7.08015740e-01 7.07835630e-02 -1.08311367e+00 8.75265300e-01
-1.31226301e-01 -1.11719573e+00 -1.96629420e-01 -5.46771064e-02
2.91034967e-01 2.24128023e-01 1.32608682e-01 3.20235521e-01
4.28605437e-01 -1.11323154e+00 1.16091025e+00 4.91634727e-01
8.53541613e-01 -7.46292830e-01 6.33220971e-01 6.26668990e-01
-1.00556576e+00 8.19881186e-02 -3.04882586e-01 9.04423743e-03
7.86787808e-01 5.13385117e-01 -1.38717294e+00 2.08553791e-01
4.51199353e-01 1.09820426e-01 2.86390353e-02 1.16379023e+00
-2.03781083e-01 1.32688296e+00 -5.73177993e-01 -4.61955726e-01
2.72458136e-01 3.63803029e-01 6.34323835e-01 1.76943398e+00
6.29328430e-01 -5.83633296e-02 -2.62575895e-01 5.37835419e-01
1.99025705e-01 3.45969647e-02 -5.41663289e-01 -3.70202452e-01
9.04530883e-01 6.80126309e-01 -3.80396873e-01 -5.39655089e-01
-1.05551280e-01 9.76650894e-01 -1.43529683e-01 5.40458202e-01
-8.96180093e-01 -6.05347514e-01 5.03301561e-01 -1.45715997e-01
5.43992937e-01 -5.05945623e-01 -1.15761897e-02 -5.53569794e-01
-2.85511196e-01 -9.59462702e-01 -2.02685505e-01 -1.01074660e+00
-9.68846262e-01 8.83471429e-01 7.09467903e-02 -8.73751104e-01
-1.08081889e+00 -6.13317668e-01 -5.30270278e-01 1.19877350e+00
-1.09914947e+00 -5.73353291e-01 4.91885692e-01 1.16332285e-01
1.05063760e+00 -6.64962083e-02 1.06420720e+00 2.88594335e-01
-4.42174256e-01 6.44563377e-01 2.55010158e-01 -1.98659033e-01
4.99762356e-01 -1.36072719e+00 1.01887786e+00 8.93092930e-01
3.20765346e-01 3.36248040e-01 1.15717709e+00 -4.78836656e-01
-1.00470233e+00 -7.58054376e-01 1.13510764e+00 -4.67985541e-01
6.19272411e-01 -6.15509391e-01 -1.18346035e+00 6.01213992e-01
5.04149497e-01 -6.42660499e-01 5.90406299e-01 -4.18004133e-02
-3.32324207e-02 1.36676282e-01 -5.33511519e-01 6.67108297e-01
5.28086901e-01 -8.38870525e-01 -5.27745128e-01 1.68487415e-01
1.09503162e+00 -5.71593821e-01 -7.81209171e-01 -1.53110623e-02
5.47339320e-01 -8.84617925e-01 6.03123009e-01 -1.89291060e-01
-2.43094698e-01 -1.88645542e-01 -2.94005245e-01 -1.38959372e+00
-2.88495779e-01 -1.33298457e+00 -4.65514094e-01 1.28671443e+00
7.39333868e-01 -5.67275703e-01 4.42333013e-01 2.86526352e-01
-6.87039435e-01 -4.66509044e-01 -1.24842846e+00 -1.14758742e+00
2.88747922e-02 -1.04170704e+00 5.48541307e-01 4.42937016e-01
7.74257481e-02 4.49853152e-01 -7.29558706e-01 3.82890612e-01
3.49106997e-01 -7.12485969e-01 4.71113622e-01 -7.23242164e-01
-6.21291220e-01 -6.71070993e-01 3.33166798e-03 -1.49247873e+00
7.57765695e-02 -4.34726268e-01 7.92895496e-01 -1.24263406e+00
-4.36658680e-01 -1.14524730e-01 -8.20489302e-02 4.33033794e-01
6.64854795e-02 -9.65187792e-03 1.81533232e-01 -1.46085799e-01
-1.11040756e-01 6.66506231e-01 8.45415533e-01 2.56952971e-01
-2.78800100e-01 3.84496212e-01 -1.77319840e-01 5.54879308e-01
1.08110511e+00 -8.13248217e-01 -4.00511295e-01 -2.19141185e-01
-2.88115889e-01 4.84880447e-01 -2.61679245e-03 -1.00241041e+00
3.05147916e-01 -3.48714627e-02 -4.86657098e-02 -6.20489061e-01
8.54145527e-01 -5.15892744e-01 2.41227925e-01 3.75379026e-01
-4.03961569e-01 9.65090916e-02 5.02102852e-01 3.64526540e-01
-1.58062190e-01 -5.45790613e-01 7.28620112e-01 -1.26027148e-02
-3.48659456e-01 9.20523107e-02 -1.20372379e+00 -1.79852322e-01
2.72919476e-01 -1.76081970e-01 4.86349314e-02 -9.79048729e-01
-6.54482484e-01 -3.12734127e-01 7.02967718e-02 5.02329707e-01
5.20126581e-01 -1.12707508e+00 -7.15977430e-01 2.11626261e-01
-3.49834293e-01 -1.72520727e-01 2.22571850e-01 6.47211909e-01
-2.35775098e-01 9.49875653e-01 2.81189829e-01 -5.27635455e-01
-1.20209765e+00 2.62127101e-01 4.50502664e-01 1.38142809e-01
-4.33789641e-01 9.76228118e-01 -2.79795136e-02 -3.76860082e-01
7.23171651e-01 -7.28947759e-01 1.44821718e-01 -3.30371767e-01
5.71659625e-01 2.95429200e-01 1.20596312e-01 -5.68588674e-01
-2.31648624e-01 -6.22882880e-02 9.25630853e-02 -8.06207001e-01
8.90901148e-01 -3.59409034e-01 3.80862504e-01 1.13557696e+00
1.13469923e+00 4.76740934e-02 -1.24036705e+00 -2.29552761e-01
6.75957426e-02 -4.93016914e-02 2.00989813e-01 -6.99019134e-01
-4.42336351e-01 1.16359460e+00 6.31064773e-01 7.63130128e-01
9.15278494e-01 1.96198318e-02 1.18475986e+00 2.49137089e-01
-4.65469211e-02 -1.11529839e+00 -2.33912066e-01 9.60657239e-01
7.18566656e-01 -6.75040781e-01 -6.23549342e-01 1.01454467e-01
-5.29770315e-01 1.13271618e+00 2.53381878e-01 3.28801632e-01
6.39745653e-01 7.62535930e-01 4.01412100e-01 3.25930119e-01
-1.07625711e+00 -1.19260319e-01 2.32487023e-01 7.93053567e-01
4.50620651e-01 2.88247824e-01 2.68039942e-01 3.47820044e-01
-6.99868083e-01 -3.44070435e-01 6.43638492e-01 6.53910637e-01
-7.03717589e-01 -1.14487994e+00 -4.91023153e-01 2.70960689e-01
-7.21607208e-01 -5.27133167e-01 4.52715755e-02 2.81297565e-01
-4.42235112e-01 1.16006398e+00 2.36541316e-01 -4.44088757e-01
3.44801843e-01 6.14797294e-01 2.72174329e-01 -6.80297792e-01
-3.48840475e-01 7.31911778e-01 3.62144589e-01 -2.47974530e-01
9.98615623e-02 -6.35120511e-01 -1.32007766e+00 -4.29864377e-01
-6.72674119e-01 4.16329086e-01 1.12890518e+00 9.39595759e-01
1.76083893e-01 8.08594465e-01 5.42620003e-01 -1.11744857e+00
-7.62733698e-01 -1.54508686e+00 -6.70941591e-01 -2.42984474e-01
6.61337137e-01 -5.50254062e-02 -5.49889028e-01 2.57447302e-01] | [14.807901382446289, 6.598591327667236] |
b9884956-e6f2-4f5d-8bf6-5b0b39f66231 | benchmarking-person-re-identification-1 | 2212.09981 | null | https://arxiv.org/abs/2212.09981v1 | https://arxiv.org/pdf/2212.09981v1.pdf | Benchmarking person re-identification datasets and approaches for practical real-world implementations | Recently, Person Re-Identification (Re-ID) has received a lot of attention. Large datasets containing labeled images of various individuals have been released, allowing researchers to develop and test many successful approaches. However, when such Re-ID models are deployed in new cities or environments, the task of searching for people within a network of security cameras is likely to face an important domain shift, thus resulting in decreased performance. Indeed, while most public datasets were collected in a limited geographic area, images from a new city present different features (e.g., people's ethnicity and clothing style, weather, architecture, etc.). In addition, the whole frames of the video streams must be converted into cropped images of people using pedestrian detection models, which behave differently from the human annotators who created the dataset used for training. To better understand the extent of this issue, this paper introduces a complete methodology to evaluate Re-ID approaches and training datasets with respect to their suitability for unsupervised deployment for live operations. This method is used to benchmark four Re-ID approaches on three datasets, providing insight and guidelines that can help to design better Re-ID pipelines in the future. | ['Joris Guerin', 'Esteban Clua', 'Luigy Machaca', 'Felix O. Sumari', 'Jose Huaman'] | 2022-12-20 | null | null | null | null | ['person-re-identification', 'pedestrian-detection'] | ['computer-vision', 'computer-vision'] | [-9.07071307e-02 -3.33425939e-01 1.21960059e-01 -6.49846852e-01
-1.10137247e-01 -6.88979506e-01 6.76687539e-01 3.06438297e-01
-6.84008420e-01 6.27224565e-01 2.49082536e-01 2.57045388e-01
1.64900482e-01 -6.90027297e-01 -4.09599394e-01 -4.94232178e-01
1.15870712e-02 6.83332205e-01 2.24618852e-01 -5.36175929e-02
1.75199658e-01 6.79744840e-01 -1.99724138e+00 1.14342168e-01
3.67857993e-01 5.32958329e-01 8.71400386e-02 6.19148076e-01
1.40456080e-01 4.38630760e-01 -9.56299365e-01 -8.40142787e-01
4.27179068e-01 -2.75782496e-01 -7.26657391e-01 2.20307082e-01
7.90033281e-01 -5.70424974e-01 -3.01886916e-01 1.11766565e+00
7.74020135e-01 2.15582401e-01 5.23391962e-01 -1.20862877e+00
-5.48291564e-01 3.02034825e-01 -3.65293115e-01 4.80078787e-01
6.60857916e-01 -2.58251298e-02 5.78761220e-01 -6.19657874e-01
6.74110293e-01 1.13973403e+00 7.72637486e-01 6.91204011e-01
-1.10913217e+00 -6.95945978e-01 3.61976512e-02 4.34190810e-01
-1.66950619e+00 -4.85291064e-01 5.59324086e-01 -4.86897171e-01
5.32586694e-01 2.95213580e-01 6.58020020e-01 1.15503466e+00
-5.72080851e-01 5.77664495e-01 9.04216290e-01 -5.08817673e-01
7.25583285e-02 8.17820013e-01 7.85462558e-02 2.25929618e-01
5.35968721e-01 -1.38827011e-01 -3.47500950e-01 -5.83437234e-02
4.45854187e-01 2.02357218e-01 4.14905772e-02 -3.88776898e-01
-1.13153148e+00 5.83089888e-01 2.65479952e-01 4.49695796e-01
-2.03115270e-01 -3.78700465e-01 4.94342744e-01 9.03756544e-02
3.42551976e-01 3.04023206e-01 -1.33329540e-01 -1.49295285e-01
-9.10277605e-01 5.05953133e-01 6.80750608e-01 8.76582146e-01
7.84071147e-01 -3.77253056e-01 1.84683502e-02 9.48264003e-01
4.17077690e-02 4.75163102e-01 5.45149922e-01 -6.72094226e-01
4.20560390e-01 8.05859685e-01 3.72770935e-01 -1.35310781e+00
-5.10594547e-01 -2.40211561e-01 -7.82042444e-01 -6.69869035e-02
6.17245257e-01 -1.59690768e-01 -5.20117044e-01 1.43635070e+00
4.24293250e-01 1.59373716e-01 4.09489982e-02 1.03790331e+00
7.21674263e-01 3.89250249e-01 1.77369878e-01 1.94917589e-01
1.65293396e+00 -7.15489328e-01 -2.61596799e-01 -3.55742186e-01
3.38908464e-01 -7.13651121e-01 5.07663369e-01 1.98290884e-01
-7.44497359e-01 -9.24260557e-01 -7.21570671e-01 3.83784205e-01
-6.56555593e-01 1.78918555e-01 1.85940504e-01 1.21740389e+00
-9.91307735e-01 4.26445603e-01 -2.39999846e-01 -1.23434067e+00
2.09127918e-01 3.34825218e-01 -5.51823080e-01 -2.11067170e-01
-1.08833027e+00 8.39450955e-01 4.75633889e-01 -1.76040027e-02
-7.39929497e-01 -4.27525580e-01 -5.78422070e-01 -2.35982522e-01
4.20379303e-02 -4.80577469e-01 9.17398870e-01 -1.21460485e+00
-8.66454184e-01 1.39754093e+00 -9.62855741e-02 -5.04434645e-01
5.88650048e-01 -2.49278490e-02 -8.74603570e-01 2.27376595e-02
4.14856553e-01 6.06016159e-01 7.75777519e-01 -1.26583850e+00
-1.15672553e+00 -5.40201485e-01 1.97088927e-01 1.76296338e-01
-6.28085017e-01 6.80315077e-01 -6.90463543e-01 -6.60758734e-01
-4.94983822e-01 -1.03271651e+00 -8.11018720e-02 -4.43424463e-01
-1.26537040e-01 -1.81841701e-01 5.91921031e-01 -9.10066724e-01
1.26986861e+00 -2.18968010e+00 8.50184448e-03 1.79457754e-01
1.19043581e-01 4.86305624e-01 4.10057604e-02 3.40421170e-01
-2.14879528e-01 2.22248688e-01 2.16982856e-01 -3.47836971e-01
-1.72090188e-01 5.57631114e-03 1.74701676e-01 4.57159698e-01
-1.51788592e-01 3.68160993e-01 -8.09820056e-01 -5.25124490e-01
4.10636872e-01 2.86437094e-01 -1.86614409e-01 3.55770230e-01
4.45940346e-01 5.92466235e-01 -1.96412325e-01 7.57410944e-01
5.79570889e-01 1.69881701e-01 3.10388785e-02 -3.14297050e-01
-2.75263011e-01 -3.81619871e-01 -1.42585003e+00 1.13362503e+00
-7.59390518e-02 7.05937386e-01 -1.35309935e-01 -1.01543796e+00
9.86015379e-01 2.92486489e-01 7.06825912e-01 -6.59688294e-01
1.08857147e-01 3.88669968e-02 -2.70936400e-01 -5.38623393e-01
7.72929430e-01 1.13836475e-01 -2.73128480e-01 5.25016189e-01
-1.82810739e-01 5.93983650e-01 5.69730043e-01 -2.38982048e-02
7.12269425e-01 -1.90884933e-01 2.15883464e-01 -4.28691171e-02
8.40035915e-01 1.75740033e-01 4.71975803e-01 8.60835850e-01
-6.58353150e-01 7.54792154e-01 -8.70052353e-02 -1.01706576e+00
-1.36594057e+00 -6.84077919e-01 -2.48091817e-02 1.12781942e+00
2.16118783e-01 -2.90885657e-01 -9.65270638e-01 -6.55050159e-01
-4.70189303e-02 3.29688609e-01 -4.96513963e-01 2.87552401e-02
-5.58423519e-01 -1.06047821e+00 7.75875330e-01 3.28895211e-01
9.26084578e-01 -8.97226274e-01 -7.02120364e-01 1.58631295e-01
-5.13994932e-01 -1.24739051e+00 -4.55216378e-01 -4.12913173e-01
-4.52808470e-01 -1.23858261e+00 -1.21498501e+00 -8.22754920e-01
7.98285186e-01 6.12187982e-01 1.19022763e+00 3.28818113e-01
-4.75928336e-01 7.92417884e-01 -6.50084734e-01 -4.26617354e-01
-2.86481470e-01 1.32239759e-01 4.56437618e-01 5.54399788e-01
8.63929272e-01 -1.20573930e-01 -6.21094406e-01 8.25816035e-01
-6.51913464e-01 -1.49874151e-01 1.38179511e-01 4.28505421e-01
1.24474965e-01 5.51984727e-01 2.44197249e-01 -6.82180107e-01
5.08275151e-01 -4.81902182e-01 -3.60853314e-01 4.50410843e-01
-3.38189602e-01 -3.94641161e-01 4.51186597e-01 -4.21790987e-01
-1.05104256e+00 1.35541141e-01 7.48799369e-02 1.01818759e-02
-8.34775984e-01 5.93128689e-02 -2.24325031e-01 -2.12712232e-02
6.54355109e-01 5.75439893e-02 -1.71998337e-01 -6.50371850e-01
-8.68937969e-02 9.54159558e-01 6.73480153e-01 -4.21040416e-01
1.01802099e+00 5.00191092e-01 -5.02187014e-01 -9.62772667e-01
-7.23026097e-01 -8.60577941e-01 -1.01067948e+00 -6.14937544e-01
1.00657499e+00 -1.03506827e+00 -3.70776325e-01 8.07132244e-01
-9.56490159e-01 8.93906802e-02 -1.53451979e-01 3.30983311e-01
2.02169027e-02 4.74604458e-01 -4.57757600e-02 -8.24298263e-01
-6.77990913e-02 -8.91892016e-01 7.58199155e-01 6.46426380e-01
-3.60841244e-01 -1.01978314e+00 -2.24273466e-02 6.48809135e-01
3.63157392e-01 2.15711072e-01 3.21438193e-01 -7.58422315e-01
-1.50451437e-01 -5.83442867e-01 -4.40605730e-01 4.11748827e-01
2.66035140e-01 1.26097293e-03 -1.04299462e+00 -4.62411970e-01
-4.24318939e-01 9.41160768e-02 3.31656337e-01 7.65774548e-02
8.32032919e-01 -1.67119205e-01 -5.17745078e-01 4.30088580e-01
1.27674603e+00 1.03789605e-01 5.03787637e-01 8.50409746e-01
6.49818182e-01 1.08002985e+00 5.77842772e-01 7.15989590e-01
7.11131692e-01 9.39808905e-01 3.05332094e-01 -2.33050555e-01
-1.12517685e-01 1.32136354e-02 3.35890412e-01 1.42689601e-01
-6.31200850e-01 -1.91778585e-01 -1.11617780e+00 5.93578398e-01
-1.71874809e+00 -1.21017647e+00 -1.49399638e-01 2.45636678e+00
3.21765363e-01 -2.30590671e-01 8.45373333e-01 6.59032837e-02
1.24065375e+00 -1.52524367e-01 -4.13621008e-01 -3.47233303e-02
-2.41043523e-01 -2.99151957e-01 8.17215264e-01 -7.85542876e-02
-1.30385113e+00 6.49440229e-01 6.34655285e+00 3.89060080e-01
-8.68689775e-01 -2.90432237e-02 6.84213221e-01 1.99197233e-01
3.41375440e-01 -1.74195200e-01 -1.26961982e+00 7.52159178e-01
9.00785208e-01 -1.86458275e-01 5.85901678e-01 8.74168634e-01
8.43527839e-02 -2.31293499e-01 -8.80960643e-01 1.44463551e+00
5.53387880e-01 -8.97367954e-01 -1.48999259e-01 1.04525700e-01
5.46099722e-01 -1.75903082e-01 -1.92180648e-01 2.09267050e-01
2.33002067e-01 -8.03143978e-01 9.11432266e-01 5.13080180e-01
7.83351839e-01 -8.48608196e-01 1.05748689e+00 2.87712753e-01
-1.27644765e+00 -3.38581741e-01 -4.82434779e-01 -8.60622227e-02
1.88306347e-01 1.96382672e-01 -8.43884051e-01 5.17353177e-01
1.39906323e+00 6.65494502e-01 -1.28192103e+00 1.19035900e+00
1.17303856e-01 2.67023504e-01 -2.93455362e-01 6.90187439e-02
-2.31063664e-01 -2.57893819e-02 3.11711490e-01 1.28574800e+00
4.33922857e-01 -1.93273067e-01 2.02119261e-01 2.86602348e-01
1.24127977e-01 1.08012013e-01 -4.55344766e-01 8.48548114e-02
3.32504213e-01 1.32952011e+00 -7.56654084e-01 -4.02781367e-01
-6.86953366e-01 1.22162747e+00 1.03232235e-01 2.97628045e-01
-9.22310650e-01 -4.79234643e-02 9.61345494e-01 6.01755619e-01
-3.12654525e-02 -1.47115320e-01 2.25051373e-01 -1.04101276e+00
-1.44093245e-01 -9.68453526e-01 7.53139615e-01 -6.17079616e-01
-1.45227587e+00 5.86804748e-01 3.20953190e-01 -1.33489430e+00
-1.74299076e-01 -6.95597768e-01 -2.87028253e-01 7.51556039e-01
-1.30479705e+00 -1.20092487e+00 -8.03932607e-01 6.85793698e-01
3.83708775e-01 -7.17588544e-01 7.34546542e-01 8.15517724e-01
-9.04277444e-01 5.66407323e-01 6.76321462e-02 6.22888565e-01
1.08611345e+00 -9.80282307e-01 4.61587250e-01 1.01665545e+00
1.14604026e-01 4.72060740e-01 7.21694827e-01 -5.90154469e-01
-9.55791295e-01 -1.09699440e+00 9.62388277e-01 -8.36275697e-01
3.41276467e-01 -2.69203246e-01 -5.47149122e-01 5.62339008e-01
5.33210486e-03 -1.13364905e-01 8.57952178e-01 2.01747287e-02
-1.25444874e-01 -4.52993751e-01 -1.31331515e+00 4.43703771e-01
1.03939283e+00 -2.72858143e-01 -3.76731128e-01 2.43176430e-01
-8.15588832e-02 -7.63622448e-02 -7.75795281e-01 8.87055621e-02
5.97364604e-01 -1.29022217e+00 1.25741518e+00 -4.54212457e-01
-1.73684791e-01 -4.09947723e-01 -6.92378432e-02 -1.04531789e+00
-4.68320101e-01 -2.42807716e-01 3.96863043e-01 1.79966676e+00
-5.61953895e-02 -5.82898855e-01 7.56450772e-01 1.07416129e+00
4.98189300e-01 2.16596499e-01 -6.66911006e-01 -7.13213742e-01
-3.28004539e-01 -2.57575363e-01 1.02426744e+00 8.44055533e-01
-3.86112869e-01 3.74727175e-02 -6.36292815e-01 3.21728170e-01
6.36981130e-01 -2.15679601e-01 1.32607377e+00 -1.57577920e+00
1.72730342e-01 -2.87277222e-01 -7.51456678e-01 -6.19803250e-01
-5.31847663e-02 -6.19898438e-01 -2.41499767e-01 -1.32370853e+00
5.47052979e-01 -6.82874143e-01 -3.47015634e-02 2.85404384e-01
-1.30044341e-01 6.23840272e-01 2.78854877e-01 5.28177619e-01
-6.14201307e-01 -9.00862291e-02 5.35496652e-01 -2.36129701e-01
-9.55812111e-02 2.02755019e-01 -7.13198185e-01 7.71965384e-01
9.94122863e-01 -3.38197052e-01 2.73843743e-02 -4.72181439e-01
1.52020469e-01 -6.22442842e-01 6.80701494e-01 -1.57489240e+00
4.25441206e-01 -3.23764943e-02 6.80461049e-01 -4.53676552e-01
8.02548453e-02 -1.06540585e+00 4.90653783e-01 2.21008211e-01
1.18836343e-01 4.08385605e-01 -1.60142090e-02 3.63197982e-01
-1.98124409e-01 -5.98626673e-01 7.51673222e-01 -4.42160964e-01
-1.17146158e+00 2.42340967e-01 -4.31643665e-01 -6.02798387e-02
1.31784117e+00 -6.99667573e-01 -8.31783749e-03 -1.82435364e-01
-5.38472831e-01 8.90367925e-02 9.38648939e-01 6.47745371e-01
2.55832940e-01 -1.20028687e+00 -8.64708543e-01 3.31915170e-01
4.28198904e-01 -3.52954477e-01 4.56170827e-01 3.79222900e-01
-6.79431498e-01 2.03803793e-01 -5.07264495e-01 -5.09110808e-01
-1.68860281e+00 6.41569018e-01 3.21692765e-01 -1.63610294e-01
-4.90373969e-01 5.46147764e-01 -9.39615369e-02 -4.01219815e-01
2.85605133e-01 3.57051104e-01 -6.04178250e-01 4.95681643e-01
8.52713883e-01 5.75312078e-01 -7.89027810e-02 -1.33599246e+00
-4.63835001e-01 5.70667326e-01 -5.97015135e-02 4.16307189e-02
1.16158175e+00 -4.66918230e-01 -9.13596991e-03 8.11523795e-02
7.62468815e-01 -2.79979631e-02 -1.11841285e+00 -3.44368875e-01
1.82255104e-01 -6.71271324e-01 -4.05051976e-01 -5.16308725e-01
-9.99980032e-01 5.34210384e-01 1.03793585e+00 2.93733001e-01
1.08857834e+00 2.31301542e-02 6.48234487e-01 1.06821358e-01
6.61296248e-01 -1.49861097e+00 -3.05874553e-02 2.35135481e-01
4.53763694e-01 -1.45209479e+00 1.04309462e-01 -8.58172849e-02
-7.15182066e-01 1.09643328e+00 4.40521181e-01 3.39246750e-01
5.63882589e-01 -7.86599070e-02 1.09054297e-01 1.24384940e-01
2.10278332e-01 -4.80915099e-01 1.06053881e-01 1.06888056e+00
2.10976794e-01 6.70091659e-02 1.36185125e-01 5.50939202e-01
-2.79789180e-01 -1.72895715e-01 4.07429576e-01 6.73683584e-01
-2.60825008e-01 -1.32467401e+00 -9.01627839e-01 2.86899775e-01
-3.44884634e-01 3.34293216e-01 -5.23647130e-01 7.43804574e-01
6.75475359e-01 1.03494871e+00 2.06875488e-01 -3.74758810e-01
5.95329642e-01 -2.81559210e-02 3.20580542e-01 -6.68568134e-01
-7.65578389e-01 -5.24779797e-01 2.01414257e-01 -1.00337319e-01
-8.77941310e-01 -1.02126431e+00 -6.17305994e-01 -6.81064367e-01
-1.10676801e-02 1.89723130e-02 6.32511914e-01 6.70795202e-01
1.08458199e-01 1.12512127e-01 5.68453610e-01 -9.97642398e-01
2.05330905e-02 -8.18707228e-01 -5.66522598e-01 9.04049039e-01
-4.56448160e-02 -6.45576119e-01 -1.99210063e-01 5.09994268e-01] | [14.526824951171875, 1.039085030555725] |
e9d16df4-ddad-4db8-bf23-1e96c4625c2e | stance-detection-and-open-research-avenues | 2210.12383 | null | https://arxiv.org/abs/2210.12383v1 | https://arxiv.org/pdf/2210.12383v1.pdf | Stance Detection and Open Research Avenues | This tutorial aims to cover the state-of-the-art on stance detection and address open research avenues for interested researchers and practitioners. Stance detection is a recent research topic where the stance towards a given target or target set is determined based on the given content and there are significant application opportunities of stance detection in various domains. The tutorial comprises two parts where the first part outlines the fundamental concepts, problems, approaches, and resources of stance detection, while the second part covers open research avenues and application areas of stance detection. The tutorial will be a useful guide for researchers and practitioners of stance detection, social media analysis, information retrieval, and natural language processing. | ['Fazli Can', 'Dilek Küçük'] | 2022-10-22 | null | null | null | null | ['stance-detection'] | ['natural-language-processing'] | [ 2.63600081e-01 3.40626776e-01 -1.00471973e+00 -2.55538464e-01
-4.31865990e-01 -6.58952415e-01 7.19243884e-01 3.82874429e-01
-6.50418503e-03 8.72903287e-01 4.37526345e-01 -9.11612883e-02
-7.48314783e-02 -9.76153135e-01 1.81947544e-01 -6.12275064e-01
-9.04761031e-02 5.54461718e-01 3.58518749e-01 -7.24795341e-01
9.89887059e-01 -3.51856440e-01 -1.82856119e+00 5.43564439e-01
8.33227336e-01 6.80056036e-01 -3.17864984e-01 3.10173541e-01
-3.30717027e-01 7.34072506e-01 -1.25664389e+00 -4.13626552e-01
-1.68668777e-01 -8.04444253e-01 -8.77454519e-01 3.18409801e-02
3.24883640e-01 1.85035899e-01 9.81581211e-02 1.28676248e+00
6.65338516e-01 -1.72032475e-01 6.78356171e-01 -1.36198938e+00
-5.12892187e-01 1.13730073e+00 -5.82465649e-01 7.58416951e-01
9.09558892e-01 -4.16106552e-01 1.37475264e+00 -7.60146499e-01
8.49207938e-01 1.33785033e+00 6.23635292e-01 2.72045374e-01
-4.38794464e-01 -4.87963200e-01 7.07836807e-01 5.01811802e-01
-5.80180585e-01 -3.92400026e-02 1.06019604e+00 -6.02493942e-01
4.81236428e-01 5.40492296e-01 1.04699826e+00 1.36044300e+00
3.25230539e-01 1.13942480e+00 1.24351263e+00 -6.49335623e-01
1.72969535e-01 1.27594978e-01 1.14509153e+00 2.50363410e-01
4.04780954e-01 -1.13354795e-01 -7.51646996e-01 -5.48889399e-01
2.30367221e-02 -4.61818367e-01 -1.00943893e-02 6.77292123e-02
-1.28324866e+00 1.41100371e+00 1.40234875e-02 6.80916488e-01
-2.96187043e-01 -6.29411459e-01 8.84132504e-01 6.29756749e-01
9.43114638e-01 3.85720968e-01 7.69892382e-03 -2.96579242e-01
-1.05476558e+00 8.31136465e-01 1.02444303e+00 7.19097197e-01
-4.73534726e-02 6.57114759e-02 -3.40833306e-01 8.71664166e-01
5.78053832e-01 6.41949892e-01 5.64119339e-01 -7.86596417e-01
4.47553098e-01 7.28494227e-01 -1.39346957e-01 -1.02226388e+00
-2.48921067e-01 -4.26924467e-01 -8.38839486e-02 2.88243175e-01
2.26426184e-01 -1.31597325e-01 -2.52585620e-01 1.20828450e+00
6.55372977e-01 -6.28887832e-01 -6.13722429e-02 5.47150910e-01
1.60753870e+00 3.89513969e-01 -1.89888194e-01 -9.13584709e-01
1.75327957e+00 -1.06897199e+00 -1.19092941e+00 -4.37299907e-01
2.83093721e-01 -1.47295856e+00 9.67895985e-01 1.41844168e-01
-1.45626128e+00 -1.02013193e-01 -1.26455653e+00 3.96861643e-01
-5.63603759e-01 -3.76648903e-01 2.96210319e-01 7.12788463e-01
-4.63924736e-01 4.89512026e-01 -1.95110977e-01 -5.02332330e-01
2.96785623e-01 -2.26936042e-02 6.81810856e-01 4.60259169e-01
-1.73558092e+00 1.07518423e+00 6.96230531e-02 -4.71550703e-01
-2.47713178e-01 -5.98843932e-01 -7.14649022e-01 -8.76166046e-01
4.40310746e-01 -4.78310317e-01 1.36770082e+00 -8.45058560e-01
-1.39140892e+00 1.44563043e+00 -2.31252909e-01 -4.01521891e-01
6.28056705e-01 -2.35485956e-01 -5.33675730e-01 1.23905554e-01
8.05821419e-01 -2.64072508e-01 9.26722825e-01 -7.67423987e-01
-6.49116397e-01 -2.42453441e-01 2.82457203e-01 5.77071130e-01
-2.37110421e-01 9.37920630e-01 4.54373956e-01 -7.04698801e-01
3.42057347e-01 -6.48634255e-01 2.47513950e-01 -4.92535830e-01
-2.38459304e-01 -7.89052606e-01 1.39922094e+00 -3.63091856e-01
1.78222072e+00 -1.48112285e+00 -7.03718066e-02 1.36409879e-01
3.94704789e-01 2.52048999e-01 4.60424751e-01 8.38343859e-01
-1.73377112e-01 -8.66421238e-02 8.18364993e-02 2.62575477e-01
1.59562111e-01 -1.96581453e-01 -4.75723326e-01 8.44500005e-01
-6.98976219e-01 8.94239426e-01 -1.17123330e+00 -7.73624480e-01
1.10865034e-01 -1.58932373e-01 3.61465625e-02 -5.38178384e-02
-3.31394851e-01 8.21892470e-02 -5.43409348e-01 9.56642747e-01
2.47625813e-01 -3.91574055e-01 4.15411443e-01 -6.16597831e-02
-4.83616948e-01 9.93150055e-01 -8.74141574e-01 6.83513045e-01
1.35296613e-01 8.85354042e-01 -6.82393461e-02 -1.05938840e+00
8.73286784e-01 5.09134233e-01 7.30998039e-01 -7.10271537e-01
4.02917981e-01 4.87410486e-01 1.13484919e-01 -4.11657363e-01
4.55766469e-01 6.11404218e-02 -3.48142207e-01 1.22998953e+00
-6.53289795e-01 2.47248888e-01 7.59873688e-01 5.44932857e-02
6.83844924e-01 4.26181927e-02 9.88377571e-01 -6.18291855e-01
8.44116449e-01 1.77532315e-01 1.84138790e-01 4.72349584e-01
-5.02747238e-01 -8.03069249e-02 3.74677986e-01 -5.31886756e-01
-7.58367002e-01 -6.45003915e-01 -3.39087576e-01 1.72751307e+00
1.89116195e-01 -6.88289225e-01 -8.04501057e-01 -8.40238988e-01
-2.66068336e-02 2.59270221e-01 -9.92325068e-01 3.73619437e-01
-9.48333859e-01 -7.72063196e-01 3.18726093e-01 3.06807518e-01
7.07894862e-01 -1.46013308e+00 -8.19884896e-01 1.90274306e-02
-1.00142825e+00 -7.68832326e-01 -1.83268860e-01 -2.62404144e-01
-8.54622483e-01 -1.63448381e+00 -6.52928829e-01 -1.00916934e+00
4.38807130e-01 6.39919341e-01 1.39439642e+00 3.33266139e-01
2.38304794e-01 1.48749366e-01 -5.90976417e-01 -5.61118484e-01
-5.50531924e-01 8.53970200e-02 2.25828126e-01 -5.74141860e-01
7.83304811e-01 -3.91113073e-01 -4.69067812e-01 5.93918145e-01
-2.86763877e-01 -3.94852519e-01 -9.65662599e-02 6.15069509e-01
2.91854113e-01 -2.77319532e-02 7.17772186e-01 -1.35627437e+00
1.30830991e+00 -4.76066858e-01 -2.90600687e-01 2.30290443e-02
-9.10479426e-01 -7.04407871e-01 -1.22435443e-01 -3.82688731e-01
-7.98190534e-01 -1.19605112e+00 -1.55929267e-01 5.48087776e-01
1.75461948e-01 8.73362541e-01 -2.07412783e-02 3.54690328e-02
1.12048078e+00 -1.43449172e-01 5.96110933e-02 -2.62298435e-01
1.11025922e-01 1.04853618e+00 -1.42105306e-02 -5.28257966e-01
3.94265503e-01 5.55070460e-01 -5.18045664e-01 -8.62341583e-01
-1.61850357e+00 -7.38332927e-01 -4.95314181e-01 -8.28089118e-01
4.93727386e-01 -5.26109517e-01 -5.30717194e-01 4.41727310e-01
-8.54126573e-01 -9.47090834e-02 -1.35078505e-01 -4.46629263e-02
-6.03120267e-01 5.02464712e-01 -7.13875055e-01 -5.96155405e-01
-9.77090061e-01 -7.11571753e-01 4.63789493e-01 1.50590569e-01
-1.37753320e+00 -1.33927822e+00 4.11690444e-01 9.27485049e-01
1.99058410e-02 6.59627080e-01 7.87317097e-01 -6.75372124e-01
2.90933192e-01 -1.44857287e-01 3.07824373e-01 -1.64718583e-01
3.32779884e-01 3.35143864e-01 -7.00248957e-01 -2.49562219e-01
3.84757847e-01 -3.16528708e-01 4.24903661e-01 7.37408638e-01
3.88309926e-01 -5.15710831e-01 -4.77339625e-01 -3.53771210e-01
7.35463858e-01 1.42496020e-01 3.56323570e-01 1.19284308e+00
-6.80367947e-02 9.08164859e-01 1.55574334e+00 3.37834209e-01
2.20779583e-01 4.36433762e-01 2.21287757e-01 3.49329770e-01
-3.21433008e-01 2.04861566e-01 5.44836640e-01 1.07060850e+00
-3.74071598e-01 -4.81142551e-02 -8.50100517e-01 4.91628498e-01
-2.10003853e+00 -1.43206227e+00 -5.93357384e-01 1.76850259e+00
9.87237275e-01 8.49311173e-01 8.85882914e-01 7.64122546e-01
8.70287538e-01 6.75597489e-01 -3.24605286e-01 -5.57709515e-01
-2.19269991e-01 9.49668419e-03 -1.28815010e-01 6.06334627e-01
-1.35464227e+00 7.86233723e-01 8.42487907e+00 4.35162723e-01
-8.17681372e-01 1.60844713e-01 -4.37279195e-02 1.98756158e-02
-2.11370252e-02 -1.92446157e-01 -9.29326892e-01 5.42500317e-01
8.90595973e-01 -8.60151589e-01 -4.19544399e-01 1.10084558e+00
4.41797227e-01 -3.77996922e-01 -8.69695246e-01 5.11609435e-01
3.70104939e-01 -1.63954866e+00 -1.97673008e-01 1.77640319e-01
8.94858062e-01 -1.38190389e-01 2.85935313e-01 1.78899020e-01
3.17865580e-01 -3.92789423e-01 9.12139654e-01 -1.54602453e-01
3.87798250e-01 -5.18267155e-01 8.47327054e-01 3.75271887e-01
-1.10094154e+00 -4.50313278e-02 -2.08655655e-01 -7.39142179e-01
1.86568215e-01 7.32636154e-01 -3.80491585e-01 2.21588850e-01
8.11631143e-01 1.28614342e+00 -1.22903109e-01 8.10465217e-01
-6.93045616e-01 6.51895881e-01 1.77052960e-01 -6.29486561e-01
2.50813961e-01 -2.32248187e-01 1.09931648e+00 1.21614122e+00
-5.05066514e-01 1.31023839e-01 4.45126206e-01 4.09225672e-01
3.77512664e-01 2.55013794e-01 -4.84738588e-01 -1.56624690e-01
6.51357889e-01 8.71587813e-01 -1.07059515e+00 -6.84416711e-01
-1.81880593e-01 1.01381302e-01 -2.27411389e-01 -1.29485533e-01
-8.22225392e-01 -1.51514143e-01 8.25315595e-01 3.89274925e-01
-8.87658596e-02 1.13741554e-01 -8.47797155e-01 -7.81709552e-01
-1.63070709e-01 -1.58447397e+00 8.26979101e-01 -6.03050403e-02
-1.27217424e+00 3.19906950e-01 3.70048910e-01 -1.49930477e+00
-2.63470829e-01 -3.64401609e-01 -9.30119395e-01 4.94413108e-01
-1.16284811e+00 -9.89171028e-01 -2.94161856e-01 2.60187358e-01
9.84405518e-01 -3.32455337e-01 8.62506986e-01 5.75907193e-02
-3.13932955e-01 2.95750976e-01 -2.59647787e-01 1.08472101e-01
8.86010587e-01 -1.02149308e+00 4.55864787e-01 3.99201363e-01
-3.67040873e-01 7.24015176e-01 1.36922741e+00 -1.00754738e+00
-5.05056024e-01 -6.29988015e-01 1.32011759e+00 -5.09958744e-01
9.85930026e-01 1.30208865e-01 -5.79927802e-01 5.31082273e-01
6.43230081e-01 -9.43788230e-01 1.35345531e+00 5.62458813e-01
-2.69245327e-01 5.14315188e-01 -1.09527361e+00 8.36899042e-01
1.10374272e+00 -2.37553835e-01 -1.21109796e+00 8.57774317e-01
2.49849752e-01 -7.99926639e-01 -7.77916610e-01 1.86884612e-01
6.68596447e-01 -9.98081744e-01 1.05960286e+00 -5.17874539e-01
7.36948609e-01 -7.16919601e-02 -2.32097600e-02 -8.27771366e-01
-4.33446497e-01 -6.53648078e-01 -7.05963135e-01 7.56499112e-01
2.64391273e-01 -5.13461471e-01 8.96592081e-01 1.10019306e-02
-7.29963109e-02 -9.21096206e-01 -4.93052125e-01 -4.12832767e-01
3.34065616e-01 -1.72843739e-01 3.46111059e-01 1.15108883e+00
8.69601727e-01 8.85896683e-01 -1.68709174e-01 -4.01575863e-01
9.16683495e-01 9.62636471e-01 6.75169408e-01 -1.49996758e+00
1.74878299e-01 -8.93237352e-01 -1.37935877e-01 -1.10141313e+00
1.29643366e-01 -5.25917828e-01 -4.68665332e-01 -1.86805570e+00
1.83805421e-01 1.55938007e-02 3.80991958e-02 1.29311755e-01
-9.34254378e-02 4.75775152e-01 -1.00639082e-01 6.04678690e-01
-8.32947671e-01 2.78128684e-02 1.23400474e+00 -4.35574412e-01
-2.22153872e-01 7.34542191e-01 -1.10096836e+00 1.20056677e+00
8.63189936e-01 -4.61331338e-01 -4.37546551e-01 3.77255410e-01
5.67774236e-01 -3.41130614e-01 -2.53738582e-01 -3.66548508e-01
7.36053064e-02 -5.77339053e-01 -1.87671408e-01 -1.45088875e+00
1.72253549e-01 -8.24843198e-02 -4.44803178e-01 7.27374852e-01
-2.23609895e-01 1.11865684e-01 -3.62816364e-01 2.22788632e-01
-3.34391475e-01 -4.49675709e-01 1.02905202e+00 -4.30838257e-01
-6.28185570e-01 -1.74743503e-01 -8.51485968e-01 4.55427885e-01
1.37376094e+00 -6.27001762e-01 -6.44970894e-01 -5.57518244e-01
-8.10066879e-01 2.72624165e-01 3.58628124e-01 6.85579658e-01
6.35583580e-01 -1.51544714e+00 -9.58347142e-01 -2.53214210e-01
2.53760338e-01 -7.58683741e-01 -3.02178502e-01 9.89949703e-01
-2.82462895e-01 4.48243141e-01 -1.47182345e-01 -5.01972258e-01
-2.19053173e+00 2.06916824e-01 2.05923140e-01 -3.86231840e-01
-5.21468103e-01 8.06622624e-01 -4.00604069e-01 -2.47122064e-01
2.86487162e-01 1.30165070e-01 -1.00476503e+00 6.15605116e-01
9.74052310e-01 6.99837744e-01 4.93831560e-02 -9.67436850e-01
-5.60093224e-01 4.39518362e-01 -1.86480269e-01 9.32565704e-02
9.18973505e-01 -3.04626852e-01 -5.68103075e-01 9.97965455e-01
6.34211957e-01 9.95885134e-02 -2.82075882e-01 -3.97527575e-01
3.11084658e-01 -3.16584617e-01 -4.54728365e-01 -5.93841732e-01
-4.30324227e-01 3.28742802e-01 8.52297768e-02 8.31165373e-01
4.57245797e-01 2.48321906e-01 7.23181963e-01 -2.05106642e-02
3.56881648e-01 -1.85430086e+00 7.05902934e-01 1.05881929e+00
9.10945475e-01 -1.11195898e+00 6.38111174e-01 -6.23878837e-01
-3.86363924e-01 1.12404788e+00 2.95165509e-01 -2.36732826e-01
1.09685683e+00 9.58160833e-02 6.33233011e-01 -7.23763227e-01
-5.12077272e-01 5.62742352e-02 3.94401222e-01 6.00385010e-01
9.96525109e-01 1.80958644e-01 -1.26325476e+00 4.01357979e-01
-8.81523728e-01 -5.85007966e-01 4.74254549e-01 1.38466156e+00
-1.07810283e+00 -1.63668883e+00 -7.26569355e-01 4.30881381e-01
-6.54131591e-01 5.74632324e-02 -1.04145730e+00 6.78581119e-01
2.72663068e-02 1.40715754e+00 -3.08050513e-01 -3.59735906e-01
3.81393373e-01 -8.24220479e-02 4.25274342e-01 -9.55063164e-01
-9.18046892e-01 4.75424975e-02 6.18442833e-01 -3.62406820e-01
-1.20416391e+00 -1.20839918e+00 -1.11628854e+00 -9.84409392e-01
-3.80207360e-01 4.15944993e-01 2.49011591e-01 1.13857448e+00
-3.53375793e-01 6.00496411e-01 4.99777675e-01 -3.86872292e-01
-4.58447397e-01 -1.03888035e+00 -3.81188512e-01 1.30712345e-01
8.83117616e-02 -1.08826709e+00 -3.19791406e-01 -5.44858351e-02] | [8.932502746582031, 9.986995697021484] |
68fb76da-dcd6-4ee9-89de-baa4e407c4c4 | sleep-stage-classification-based-on-multi | 1711.00629 | null | http://arxiv.org/abs/1711.00629v1 | http://arxiv.org/pdf/1711.00629v1.pdf | Sleep Stage Classification Based on Multi-level Feature Learning and Recurrent Neural Networks via Wearable Device | This paper proposes a practical approach for automatic sleep stage
classification based on a multi-level feature learning framework and Recurrent
Neural Network (RNN) classifier using heart rate and wrist actigraphy derived
from a wearable device. The feature learning framework is designed to extract
low- and mid-level features. Low-level features capture temporal and frequency
domain properties and mid-level features learn compositions and structural
information of signals. Since sleep staging is a sequential problem with
long-term dependencies, we take advantage of RNNs with Bidirectional Long
Short-Term Memory (BLSTM) architectures for sequence data learning. To simulate
the actual situation of daily sleep, experiments are conducted with a resting
group in which sleep is recorded in resting state, and a comprehensive group in
which both resting sleep and non-resting sleep are included.We evaluate the
algorithm based on an eight-fold cross validation to classify five sleep stages
(W, N1, N2, N3, and REM). The proposed algorithm achieves weighted precision,
recall and F1 score of 58.0%, 60.3%, and 58.2% in the resting group and 58.5%,
61.1%, and 58.5% in the comprehensive group, respectively. Various comparison
experiments demonstrate the effectiveness of feature learning and BLSTM. We
further explore the influence of depth and width of RNNs on performance. Our
method is specially proposed for wearable devices and is expected to be
applicable for long-term sleep monitoring at home. Without using too much prior
domain knowledge, our method has the potential to generalize sleep disorder
detection. | ['Eric I-Chao Chang', 'Yan Xu', 'Yubo Fan', 'Xin Zhang', 'Weixuan Kou', 'He Gao'] | 2017-11-02 | null | null | null | null | ['sleep-staging', 'automatic-sleep-stage-classification'] | ['medical', 'medical'] | [ 3.39249641e-01 -2.95311213e-01 -3.04277092e-01 -4.35940385e-01
-2.35622421e-01 2.29713656e-02 5.82687892e-02 -1.91594839e-01
-6.78158700e-01 8.78810048e-01 2.99560815e-01 -9.23679769e-02
-1.63065448e-01 -5.35756409e-01 2.66100336e-02 -7.31158197e-01
-3.97364289e-01 -2.15220064e-01 -4.39712480e-02 -1.27388105e-01
4.31926586e-02 1.86281785e-01 -1.63669932e+00 1.64352506e-01
6.82112217e-01 1.28659832e+00 2.81763583e-01 6.64784014e-01
1.28325939e-01 6.04658365e-01 -7.42345214e-01 3.57676923e-01
-4.36318070e-02 -6.61757708e-01 -5.46560407e-01 -4.09051590e-02
-3.08504462e-01 1.81567475e-01 -4.31941412e-02 6.48908496e-01
8.13886404e-01 3.61038029e-01 2.66306490e-01 -9.03041959e-01
-2.41908878e-01 2.13875309e-01 -4.73346323e-01 9.14932668e-01
4.76778686e-01 1.29582062e-01 7.35607147e-01 -5.74252427e-01
-1.25629436e-02 4.66947377e-01 9.64876533e-01 9.07078147e-01
-1.09004736e+00 -7.90267408e-01 6.48935698e-03 1.71872333e-01
-1.36206937e+00 -5.66156030e-01 6.80018902e-01 -2.55448192e-01
1.19107521e+00 2.74910748e-01 9.92688537e-01 1.11187589e+00
7.76259899e-01 2.19843522e-01 1.30658352e+00 -2.19482347e-01
3.03459048e-01 -9.34671462e-02 4.82055664e-01 6.28371477e-01
1.23395428e-01 1.08275548e-01 -7.97418058e-01 -2.88036168e-02
3.83596390e-01 6.69936299e-01 -2.33799770e-01 3.70703399e-01
-9.62277591e-01 5.07775784e-01 3.46646309e-01 8.02335858e-01
-6.01892233e-01 -4.73779216e-02 4.17944789e-01 2.91898787e-01
5.60316265e-01 7.17282742e-02 -4.16809261e-01 -5.33000529e-01
-1.52475786e+00 -4.05053347e-01 7.12960958e-01 4.87397373e-01
5.96140623e-01 2.09117621e-01 -2.37674087e-01 8.88210237e-01
3.54965985e-01 5.78397274e-01 1.35365176e+00 -7.59066105e-01
7.58139119e-02 7.27430463e-01 -8.14967901e-02 -6.80634618e-01
-1.06899428e+00 -5.78377545e-01 -1.19676197e+00 -3.39382499e-01
-1.51230365e-01 -1.37482926e-01 -7.11275876e-01 1.59974813e+00
-9.38103124e-02 3.46449584e-01 9.51286182e-02 6.82327211e-01
9.23276842e-01 4.79428619e-01 1.16022704e-02 -7.02878952e-01
1.50359762e+00 -7.03628182e-01 -9.34487224e-01 -3.38386685e-01
2.91196346e-01 -3.10609281e-01 1.14658046e+00 2.10323960e-01
-1.10283184e+00 -8.65319371e-01 -1.22364390e+00 1.10355578e-01
-1.77117884e-01 2.48248607e-01 2.30880827e-01 7.82417357e-01
-1.25642002e+00 7.96876967e-01 -1.42187679e+00 -6.06233001e-01
2.22043291e-01 8.69597793e-01 -1.98957343e-02 4.44780290e-01
-1.16520047e+00 6.98872268e-01 -6.39127791e-02 5.39434910e-01
-7.00073421e-01 -3.07159334e-01 -7.05186605e-01 -1.81708448e-02
-3.08741152e-01 -8.36881638e-01 9.55027163e-01 -6.88191712e-01
-1.50472105e+00 8.34370434e-01 -6.30749404e-01 -5.76835394e-01
-2.08056271e-01 -1.45900786e-01 -9.32032108e-01 1.03594840e-01
4.56416160e-02 8.90462175e-02 7.54774988e-01 -4.51783925e-01
-5.00978529e-01 -6.90514684e-01 -3.84939522e-01 1.39551889e-02
-4.76228386e-01 4.22045253e-02 1.76557258e-01 -3.06788802e-01
5.10361455e-02 -9.14308190e-01 -1.25967890e-01 -3.89892638e-01
-1.14187531e-01 -1.02897406e-01 4.54474092e-01 -5.56235552e-01
1.57904673e+00 -2.30754304e+00 -2.92700589e-01 1.33350030e-01
3.33227605e-01 2.63711661e-01 2.54908919e-01 2.57816672e-01
1.81272238e-01 -3.10927741e-02 -5.74789979e-02 -3.87563437e-01
-3.80266547e-01 4.19568837e-01 1.47002846e-01 5.31600952e-01
-1.26742989e-01 7.19152808e-01 -6.67217135e-01 -2.80893117e-01
1.13706119e-01 5.48966825e-01 -1.82776213e-01 1.98497280e-01
5.87148666e-01 4.33598697e-01 -3.30060124e-01 5.33180296e-01
3.33223119e-02 -4.20322686e-01 1.04709470e-03 -1.44749656e-01
-3.14432532e-01 6.30229771e-01 -6.57446861e-01 1.74348330e+00
-7.65767157e-01 4.92036134e-01 -1.86401397e-01 -9.05618906e-01
1.19651818e+00 4.34083313e-01 4.86229539e-01 -9.94734228e-01
2.42087722e-01 3.54322698e-03 1.60722211e-01 -7.81274557e-01
3.04713309e-01 -6.26296937e-01 -5.77884093e-02 5.79847395e-01
1.22211501e-01 6.61036193e-01 9.98893082e-02 -3.05684716e-01
1.50343227e+00 -1.60771996e-01 5.68078160e-01 -3.84636253e-01
6.05332196e-01 -7.53696799e-01 8.37872684e-01 6.51493251e-01
-5.06960690e-01 3.06812048e-01 -9.96222161e-03 -5.47999442e-01
-3.57693344e-01 -9.86531973e-01 8.89307559e-02 1.12073433e+00
9.16596353e-02 -5.49910724e-01 -3.12493831e-01 -2.67527938e-01
-4.25525516e-01 4.52243805e-01 -7.69332647e-01 -7.29784310e-01
-3.82440776e-01 -1.08117986e+00 6.28630102e-01 6.75228000e-01
6.70669436e-01 -1.43844140e+00 -1.03903592e+00 1.95539817e-01
-2.56884366e-01 -8.65409136e-01 -4.71139967e-01 4.59459871e-01
-1.30242217e+00 -8.90733004e-01 -4.29462552e-01 -6.52560294e-01
3.37752819e-01 2.15194136e-01 1.01634967e+00 6.11043200e-02
-3.14883024e-01 2.59522855e-01 -2.45449692e-01 -3.96178849e-02
1.18191659e-01 2.74392009e-01 4.83756095e-01 1.56161174e-01
6.24892116e-01 -1.26216257e+00 -1.09963918e+00 4.47354555e-01
-3.59682739e-01 -3.05153042e-01 6.92328334e-01 7.02080727e-01
5.72296977e-01 -1.17014498e-01 7.43906260e-01 -4.22207803e-01
6.51129425e-01 -5.35464704e-01 3.19742449e-02 -2.14173347e-02
-9.73472893e-01 -9.66443196e-02 6.70854568e-01 -3.85296255e-01
-7.61958182e-01 -1.79637030e-01 -3.84628534e-01 -3.05629149e-02
-1.25874266e-01 3.05797696e-01 5.08793863e-03 3.08875591e-01
6.47199690e-01 7.41506517e-01 3.22832055e-02 -4.02680576e-01
-3.86493742e-01 9.45002556e-01 4.44921583e-01 -1.12454340e-01
2.68737972e-01 4.52455133e-01 -2.20999628e-01 -1.17049384e+00
-1.01092339e+00 -6.53133571e-01 -4.98623848e-01 -8.20991546e-02
1.02026391e+00 -1.05143452e+00 -9.26616728e-01 1.64960161e-01
-4.45045680e-01 -4.75105733e-01 -4.75705922e-01 6.62012577e-01
-5.15512109e-01 5.07348068e-02 -6.27421558e-01 -1.00890899e+00
-1.10600698e+00 -7.04983354e-01 8.86823118e-01 4.55608964e-01
-6.20834410e-01 -1.03448677e+00 2.67724127e-01 2.65756428e-01
5.68103909e-01 1.23637520e-01 4.66497779e-01 -6.21455789e-01
2.92076230e-01 1.01786284e-02 2.99064696e-01 4.28413838e-01
5.87763846e-01 -5.03616869e-01 -9.68113482e-01 -4.17435408e-01
6.56256914e-01 -1.54320806e-01 6.50215149e-01 4.56620485e-01
8.89325023e-01 -1.58876956e-01 -2.83262551e-01 5.72018564e-01
1.14095080e+00 4.67615515e-01 6.12379014e-01 2.36742541e-01
3.10107023e-01 9.78743657e-02 1.92903996e-01 4.82629597e-01
4.51551437e-01 3.88913155e-01 7.70999044e-02 7.41345882e-02
-7.47599751e-02 1.51445627e-01 7.93859065e-01 1.15570319e+00
-2.87042916e-01 2.26958320e-01 -6.64407790e-01 4.72684056e-01
-1.46174324e+00 -1.20888841e+00 1.32152498e-01 2.24153638e+00
8.36383462e-01 3.12284708e-01 5.11525691e-01 4.63386327e-01
6.52681828e-01 1.61484808e-01 -5.60433626e-01 -5.04140198e-01
3.07816267e-01 5.49841344e-01 1.27781436e-01 -3.65621932e-02
-7.56362557e-01 3.09181154e-01 6.16725397e+00 3.37161809e-01
-1.23046577e+00 3.44497204e-01 3.82962823e-01 -5.95742285e-01
3.33285511e-01 -3.47707391e-01 -8.94806385e-01 9.77041066e-01
1.83625698e+00 4.19843867e-02 5.46296835e-01 4.34872687e-01
7.51598418e-01 -2.13345125e-01 -8.11526299e-01 1.22730613e+00
6.28833920e-02 -7.79843271e-01 -6.68884456e-01 5.85123757e-03
2.94118375e-01 2.71135211e-01 -1.52395427e-01 5.00390947e-01
-3.33200872e-01 -8.97873938e-01 1.02785714e-01 8.53540719e-01
8.72565150e-01 -6.52977586e-01 7.98735619e-01 3.83649349e-01
-1.36498249e+00 -2.41105124e-01 -2.00220048e-01 -5.33040702e-01
7.15473294e-02 6.58469498e-01 -3.52147907e-01 2.66712546e-01
9.75515783e-01 1.13262427e+00 -5.99927843e-01 7.88333535e-01
-9.34477970e-02 9.23613667e-01 -3.21778625e-01 -2.76304930e-01
-1.88867748e-01 -9.51844454e-02 -1.01034716e-02 1.01622164e+00
4.64505970e-01 1.35681227e-01 3.74031882e-03 6.49897456e-01
5.56477979e-02 -3.65499824e-01 -5.45031846e-01 1.30325407e-01
2.18373597e-01 1.43482590e+00 -8.57686162e-01 -3.59327234e-02
-4.52076465e-01 1.00655591e+00 -5.33049516e-02 1.32975101e-01
-7.04619050e-01 -5.13233483e-01 6.11492574e-01 5.41786104e-02
1.82361409e-01 -6.93748295e-02 -2.90532708e-01 -1.17226934e+00
6.08953089e-02 -5.42297840e-01 4.31996435e-01 -5.79512477e-01
-1.22452724e+00 8.28270555e-01 -3.48034263e-01 -1.25369096e+00
-2.28803813e-01 -5.49825616e-02 -9.24841762e-01 6.80383861e-01
-1.21074581e+00 -7.27769136e-01 -4.59197521e-01 8.56663048e-01
5.11810184e-01 -3.67616792e-03 8.81744087e-01 4.58426118e-01
-8.00888598e-01 6.76262796e-01 -3.60098928e-02 -5.68013340e-02
2.61079580e-01 -1.02073205e+00 4.36230078e-02 5.31382143e-01
-5.60348481e-02 9.80451286e-01 4.96449172e-01 -3.66468281e-01
-1.20523763e+00 -1.19781888e+00 1.02671897e+00 -2.02343374e-01
4.78463471e-01 -5.02496123e-01 -6.03863657e-01 5.34718156e-01
-9.27177966e-02 1.77897483e-01 1.29295766e+00 3.23359698e-01
3.38875145e-01 -6.65421426e-01 -1.24063325e+00 2.18551829e-01
1.00925052e+00 -7.21990764e-01 -7.84059882e-01 -2.91974582e-02
2.23881900e-01 -6.56724349e-02 -1.07629049e+00 3.36099207e-01
9.39100802e-01 -1.08946931e+00 6.64263189e-01 -7.44588375e-02
8.14265013e-02 -1.23188399e-01 -2.73360014e-02 -1.02271020e+00
-2.54493654e-01 -6.66536748e-01 -4.48362112e-01 1.14945984e+00
3.29826415e-01 -6.90480769e-01 6.80102706e-01 3.28947663e-01
-3.40550363e-01 -9.10113633e-01 -9.83308315e-01 -8.68859291e-01
-5.08419275e-01 -3.44724804e-01 2.51975894e-01 3.58821750e-01
1.50611430e-01 7.90734768e-01 -6.18019640e-01 -1.25506565e-01
1.36621684e-01 4.17907774e-01 2.14834243e-01 -1.33000147e+00
-1.64392710e-01 2.23808046e-02 -4.88612086e-01 -6.85543954e-01
1.23739466e-01 -8.11461270e-01 1.81822255e-02 -1.53971267e+00
2.92862117e-01 -5.92380054e-02 -1.11086822e+00 6.15884483e-01
1.22070182e-02 6.27477646e-01 -2.50911862e-01 2.19926000e-01
-8.54656458e-01 6.84904993e-01 7.76836634e-01 1.72214180e-01
-7.24853873e-01 6.64597571e-01 -6.85463548e-01 7.01049924e-01
1.19972157e+00 -6.28440380e-01 -5.99425673e-01 4.53608856e-02
9.49869379e-02 1.56158343e-01 9.91921872e-02 -1.44371581e+00
2.66618073e-01 2.78774530e-01 7.63476610e-01 -4.54134256e-01
5.60184002e-01 -6.84901655e-01 7.87425414e-02 8.62559676e-01
-2.74090499e-01 3.19174558e-01 1.08675994e-01 5.26828408e-01
6.93400353e-02 5.96288517e-02 5.70083082e-01 -1.88872561e-01
-4.70154434e-01 1.39272049e-01 -7.42463231e-01 3.50461937e-02
6.45678937e-01 -6.03830397e-01 -1.20065622e-02 -2.51198798e-01
-1.21439421e+00 8.70018154e-02 1.17542066e-01 3.40715855e-01
8.13723147e-01 -1.18093979e+00 -1.37143090e-01 6.18506014e-01
3.75057757e-02 -5.60397208e-01 2.90429085e-01 1.33769190e+00
4.09410261e-02 2.98718601e-01 -4.14063126e-01 -5.32739818e-01
-1.20034051e+00 3.28728944e-01 4.34727132e-01 -4.39997673e-01
-6.77680194e-01 4.99683231e-01 -3.05372089e-01 6.99254265e-03
4.60542776e-02 -6.89729095e-01 -3.52393329e-01 4.77348231e-02
6.98970318e-01 5.87944090e-01 2.80332655e-01 -5.78435898e-01
-6.28168046e-01 6.08776093e-01 1.47312418e-01 2.81313211e-01
1.34259748e+00 -4.44379866e-01 -2.08640322e-01 1.16902316e+00
1.10948718e+00 -2.27074876e-01 -7.41961360e-01 1.16595782e-01
2.46916533e-01 3.00520331e-01 -8.31020847e-02 -7.09969461e-01
-1.15152192e+00 9.09290433e-01 1.28688431e+00 2.81720847e-01
1.53963435e+00 -1.59348816e-01 1.12250233e+00 4.41469222e-01
3.30900639e-01 -8.51183355e-01 -2.41196435e-02 2.77816832e-01
3.53449225e-01 -9.44701970e-01 1.11614726e-01 4.89847809e-01
-3.36988777e-01 9.79443133e-01 4.29786086e-01 -3.49076211e-01
7.88307190e-01 1.35698140e-01 -3.34700011e-02 -4.20620918e-01
-8.78243208e-01 -2.75585890e-01 2.17531249e-01 4.05317277e-01
6.19020760e-01 -7.60869086e-02 -5.16947150e-01 9.62666452e-01
-3.85045856e-01 3.46132487e-01 2.98147082e-01 9.57254469e-01
-4.59325880e-01 -9.01163340e-01 2.71577388e-02 9.69441354e-01
-9.34075296e-01 -2.53316462e-02 1.07360959e-01 5.79188883e-01
1.56824723e-01 1.35971546e+00 1.46090493e-01 -7.58470356e-01
2.62395412e-01 3.82223696e-01 2.49841213e-01 -8.15907419e-01
-9.23519135e-01 -2.79705098e-04 -1.28042445e-01 -7.59328425e-01
-7.48088896e-01 -4.89481509e-01 -1.38745570e+00 1.62447728e-02
-1.21534392e-01 3.45254481e-01 5.09996474e-01 1.07980967e+00
6.20422959e-01 7.63237238e-01 6.15197659e-01 -8.37189019e-01
-2.08901450e-01 -1.21656728e+00 -9.14444864e-01 4.45246398e-02
7.91582406e-01 -5.59960127e-01 -4.09700006e-01 5.44300489e-02] | [13.550385475158691, 3.469374895095825] |
3290ec5d-688e-4651-b3a0-406f047ee29a | a-commonsense-reasoning-framework-for | 2101.04017 | null | https://arxiv.org/abs/2101.04017v5 | https://arxiv.org/pdf/2101.04017v5.pdf | A Commonsense Reasoning Framework for Explanatory Emotion Attribution, Generation and Re-classification | We present DEGARI (Dynamic Emotion Generator And ReclassIfier), an explainable system for emotion attribution and recommendation. This system relies on a recently introduced commonsense reasoning framework, the TCL logic, which is based on a human-like procedure for the automatic generation of novel concepts in a Description Logics knowledge base. Starting from an ontological formalization of emotions based on the Plutchik model, known as ArsEmotica, the system exploits the logic TCL to automatically generate novel commonsense semantic representations of compound emotions (e.g. Love as derived from the combination of Joy and Trust according to Plutchik). The generated emotions correspond to prototypes, i.e. commonsense representations of given concepts, and have been used to reclassify emotion-related contents in a variety of artistic domains, ranging from art datasets to the editorial contents available in RaiPlay, the online platform of RAI Radiotelevisione Italiana (the Italian public broadcasting company). We show how the reported results (evaluated in the light of the obtained reclassifications, the user ratings assigned to such reclassifications, and their explainability) are encouraging, and pave the way to many further research directions. | ['Rossana Damiano', 'Viviana Patti', 'Stefano Zoia', 'Gian Luca Pozzato', 'Antonio Lieto'] | 2021-01-11 | null | null | null | null | ['commonsense-knowledge-base-construction', 'causal-emotion-entailment', 'novel-concepts'] | ['knowledge-base', 'natural-language-processing', 'reasoning'] | [ 2.22074613e-01 8.48241448e-01 -1.69828549e-01 -5.55835426e-01
1.72346517e-01 -5.82127392e-01 9.24008667e-01 2.67488480e-01
1.24235637e-01 7.64910817e-01 3.25793028e-01 1.23997994e-01
-5.44532478e-01 -8.79091918e-01 -4.69164252e-01 -2.22432941e-01
2.07930394e-02 5.50978243e-01 -2.27664798e-01 -7.34225750e-01
1.58140898e-01 5.14281750e-01 -2.28103971e+00 6.34523630e-01
9.94861543e-01 1.34512711e+00 -3.57324004e-01 3.88875186e-01
-1.47096172e-01 1.14357734e+00 -6.02724433e-01 -8.80713701e-01
-4.68785107e-01 -7.74437904e-01 -9.45837021e-01 -3.57943773e-02
-2.07487851e-01 4.59012687e-01 1.78707317e-01 1.19921815e+00
-5.64999320e-02 4.35609579e-01 8.35199773e-01 -1.81852531e+00
-1.07449484e+00 1.14717507e+00 6.38141036e-01 -4.55701947e-01
6.76119387e-01 -4.48912054e-01 1.25309634e+00 -6.81651175e-01
1.05468166e+00 1.32243943e+00 4.45549846e-01 8.57905984e-01
-1.04286134e+00 -5.08433342e-01 -1.74480334e-01 7.51208186e-01
-1.23987865e+00 -3.14460605e-01 8.89214754e-01 -3.57657313e-01
8.14138412e-01 6.85623407e-01 9.27896440e-01 1.28893602e+00
-3.07559483e-02 5.16570687e-01 1.28014004e+00 -8.00634503e-01
7.36707389e-01 1.03779936e+00 2.47598305e-01 4.52273279e-01
1.80089474e-01 4.21510413e-02 -7.23001838e-01 -1.42940000e-01
2.33637586e-01 -3.45088929e-01 -2.81297743e-01 -1.66071281e-01
-9.35229659e-01 9.13511813e-01 4.49866831e-01 6.94476664e-01
-5.87548792e-01 3.39392610e-02 6.58587277e-01 3.35628241e-01
4.50461954e-01 9.23167646e-01 -2.90705562e-01 -6.67150989e-02
-3.46912920e-01 3.35450232e-01 1.02107179e+00 9.36660171e-01
5.30413330e-01 -4.19104621e-02 3.32557708e-02 6.39993429e-01
3.34432483e-01 3.05283368e-01 7.13419795e-01 -1.16657686e+00
-6.33453190e-01 8.14202428e-01 8.89671743e-02 -1.39487231e+00
-5.26519895e-01 -3.08108866e-01 -5.01202404e-01 1.06526963e-01
-3.13500702e-01 -5.10191545e-03 -2.98275322e-01 1.69754541e+00
1.27064884e-01 -1.67407235e-03 6.80310071e-01 9.20101941e-01
1.17672360e+00 5.01619220e-01 1.41927183e-01 -4.37399238e-01
1.23292029e+00 -5.49559236e-01 -1.16018522e+00 2.40159765e-01
5.11480212e-01 -3.14320326e-01 1.20278752e+00 9.21096087e-01
-8.13674152e-01 -3.36110204e-01 -1.06167471e+00 2.01458573e-01
-9.72696066e-01 6.81885108e-02 1.19818723e+00 5.31478703e-01
-9.03357446e-01 7.40728736e-01 -6.14770651e-02 -6.77733779e-01
1.94763154e-01 -8.13858733e-02 -3.14090401e-01 3.30714434e-01
-1.73588169e+00 1.03481281e+00 6.18057311e-01 -9.72773060e-02
-4.30690646e-01 -3.46729994e-01 -9.46164548e-01 1.24863684e-01
4.13569421e-01 -4.25837398e-01 7.68029630e-01 -1.68265390e+00
-1.76408243e+00 1.14925086e+00 3.82625401e-01 -5.89081943e-01
8.54648650e-02 6.59211166e-03 -1.14660037e+00 2.75064260e-01
-9.62575451e-02 6.72795355e-01 6.66221082e-01 -1.37215436e+00
-3.26497018e-01 -1.11715659e-01 4.86680597e-01 -1.14058824e-02
-5.66996336e-01 8.41043070e-02 1.85455084e-01 -6.96631193e-01
-2.09406808e-01 -9.61557806e-01 5.95143959e-02 -2.77311981e-01
-4.75181818e-01 -3.07117373e-01 3.07880431e-01 -2.78276116e-01
1.22815919e+00 -2.12205768e+00 4.02326584e-01 6.15950882e-01
3.66142355e-02 -3.54525566e-01 2.72676140e-01 2.61873603e-01
-4.67096776e-01 2.50085890e-01 -4.52731876e-03 1.44456431e-01
5.71304679e-01 3.16107929e-01 -6.67502999e-01 -1.07025571e-01
-1.97612703e-01 6.24910414e-01 -1.04147816e+00 -3.49484891e-01
2.76527792e-01 3.85699719e-01 -5.49159527e-01 -1.39780873e-02
-5.68095863e-01 -6.94833472e-02 -3.22426558e-01 4.51940894e-01
1.07664451e-01 1.50103390e-01 1.53539538e-01 -4.62518722e-01
1.11136794e-01 1.96170822e-01 -1.01877439e+00 1.46894610e+00
-6.23853087e-01 3.53283793e-01 -6.33219838e-01 -7.19991803e-01
1.42935979e+00 5.39553165e-01 2.68755674e-01 -3.54667604e-01
4.74486440e-01 1.94355235e-01 -4.11032915e-01 -6.08635485e-01
7.30825305e-01 -8.02367985e-01 -4.91969049e-01 4.66836154e-01
3.07280868e-01 -5.37775993e-01 2.74753451e-01 3.18038136e-01
7.38735378e-01 3.32488328e-01 5.15364170e-01 -2.59258062e-01
6.81922615e-01 2.76500732e-01 2.22509727e-01 3.37456852e-01
1.08976386e-01 8.52407739e-02 5.45762956e-01 -4.96914476e-01
-6.23449266e-01 -7.74906635e-01 -1.54219866e-01 1.05223358e+00
1.19068071e-01 -6.76000714e-01 -7.28082240e-01 -4.01216507e-01
-3.11530232e-01 1.63303494e+00 -8.62048030e-01 -6.40008569e-01
4.45556164e-01 -3.08971941e-01 6.72590733e-01 9.83525217e-02
1.26738042e-01 -1.79372263e+00 -8.32952678e-01 8.05096626e-02
-2.89481401e-01 -1.14585388e+00 4.98466581e-01 3.24547201e-01
-3.38387430e-01 -9.39225137e-01 1.77946672e-01 -2.55438894e-01
2.90979505e-01 -5.07972300e-01 1.29452288e+00 1.01526506e-01
-3.48197907e-01 6.16755486e-01 -1.06040931e+00 -6.93126023e-01
-8.68350744e-01 -4.60202932e-01 1.87401026e-01 2.14424103e-01
3.27961683e-01 -6.18712664e-01 1.43771410e-01 2.95680220e-04
-1.23141634e+00 2.10259214e-01 -1.41937569e-01 5.45986354e-01
4.71825480e-01 4.23402369e-01 7.47999609e-01 -1.04088032e+00
8.42304051e-01 -5.54652750e-01 -8.56049135e-02 3.56490612e-01
-8.27119768e-01 4.32496630e-02 5.92830777e-01 -4.39230114e-01
-1.10821700e+00 -4.39685881e-01 4.43527065e-02 -4.03554320e-01
-3.14465523e-01 5.14583945e-01 -3.23713899e-01 7.01767579e-02
9.83061135e-01 -8.62316936e-02 -3.61205995e-01 -3.23082656e-02
9.58429456e-01 7.89482057e-01 7.34385192e-01 -8.15906107e-01
3.27080280e-01 5.04563272e-01 4.06975932e-02 -4.41672087e-01
-8.98917317e-01 1.32793397e-01 -5.74499853e-02 -7.39136577e-01
8.18132520e-01 -4.57437664e-01 -9.86570776e-01 -3.03562969e-01
-1.09462452e+00 2.35799197e-02 -9.57919776e-01 5.41883707e-01
-1.07903004e+00 -2.28628412e-01 -3.03880572e-01 -9.76899803e-01
-3.60246539e-01 -4.15373862e-01 6.55439496e-01 2.97087461e-01
-9.15372133e-01 -9.80624676e-01 -1.58782810e-01 1.19614623e-01
1.95462018e-01 6.44711196e-01 1.43801498e+00 -8.91229510e-01
2.64044225e-01 -2.55834758e-01 2.88839161e-01 5.33463776e-01
-2.01557115e-01 3.17407042e-01 -9.82279837e-01 6.19942486e-01
-7.04358593e-02 -6.25220180e-01 3.33135337e-01 -1.28134236e-01
9.83145237e-01 -4.53989714e-01 -3.06016486e-02 9.79928225e-02
1.37584102e+00 2.15493172e-01 9.39696610e-01 5.16161382e-01
1.13782197e-01 8.87852311e-01 8.27095687e-01 9.31728125e-01
3.54503661e-01 8.80869091e-01 5.00386178e-01 2.68058509e-01
9.07016098e-02 -1.69071853e-01 4.48299080e-01 5.01314819e-01
-4.69200015e-01 -8.67789835e-02 -4.13863331e-01 3.51241440e-01
-1.93372321e+00 -1.30893111e+00 -1.39501184e-01 1.88886476e+00
8.88208032e-01 -1.92322414e-02 -1.04588896e-01 2.48831391e-01
6.97065711e-01 -2.34372944e-01 -8.35813060e-02 -1.31664681e+00
-5.02287984e-01 4.45577294e-01 -3.06559652e-01 4.98003423e-01
-5.57287276e-01 1.13066185e+00 5.43534994e+00 6.94612861e-01
-7.02302277e-01 1.14984579e-01 1.33004397e-01 6.02483973e-02
-7.31447160e-01 1.03018163e-02 -1.31962016e-01 3.97697613e-02
1.01598120e+00 -7.09788382e-01 9.04254854e-01 1.03934169e+00
-7.77977854e-02 1.65636852e-01 -1.03374243e+00 8.66079807e-01
6.18073225e-01 -1.28312492e+00 2.52059340e-01 -4.59860563e-01
5.64187884e-01 -7.65296519e-01 -1.90079570e-01 7.08129287e-01
1.41960487e-01 -8.43192220e-01 1.15252614e+00 8.96188259e-01
6.95747137e-01 -1.05033076e+00 8.26428533e-01 -4.64552306e-02
-5.24383128e-01 4.63555604e-02 -2.94184059e-01 -2.95462012e-01
-6.61763176e-02 7.08997965e-01 -5.31951666e-01 6.38984144e-01
7.35358179e-01 8.16880047e-01 -4.13379282e-01 2.39305556e-01
-6.68254256e-01 2.87627578e-01 8.60287026e-02 -2.38562584e-01
-2.73128599e-01 -8.32194388e-02 7.17138469e-01 1.10347021e+00
3.90717417e-01 2.66027898e-01 -4.70718831e-01 1.15105247e+00
-1.77927405e-01 3.74131382e-01 -4.54926878e-01 -2.50393271e-01
2.48191655e-01 1.48493755e+00 -7.04962611e-01 -6.68411911e-01
1.80362150e-01 1.12487984e+00 -1.56307314e-02 1.77292049e-01
-1.05443740e+00 -5.86479843e-01 3.95126700e-01 -1.49567798e-01
-1.51538298e-01 8.13243449e-01 -1.86545864e-01 -1.13128591e+00
-3.26757729e-01 -8.65908861e-01 4.45186257e-01 -1.77416718e+00
-1.46810710e+00 1.22776043e+00 3.28810483e-01 -1.02587306e+00
-4.49331641e-01 -5.59019029e-01 -3.37582380e-01 2.22696126e-01
-8.85418653e-01 -1.08660960e+00 -4.52155799e-01 5.56855559e-01
6.84629604e-02 -1.71297193e-02 1.62077975e+00 -1.76904455e-01
3.85242142e-02 2.22689122e-01 -4.93800759e-01 -6.22882783e-01
6.52124524e-01 -1.17309034e+00 -4.24733788e-01 2.27055520e-01
1.90069094e-01 3.90187770e-01 1.27970910e+00 -4.36090618e-01
-7.61812508e-01 -8.95602167e-01 1.28555453e+00 -3.80948812e-01
8.11017036e-01 -1.44900441e-01 -7.67345488e-01 7.11895049e-01
2.20801711e-01 -3.10679466e-01 1.03148305e+00 1.74791247e-01
-4.71987993e-01 7.38763437e-02 -1.33308649e+00 7.12876201e-01
9.03409779e-01 -3.24539214e-01 -1.01851296e+00 4.18589592e-01
9.24372911e-01 -7.07364976e-02 -9.93133903e-01 2.66856223e-01
5.73608875e-01 -1.17622566e+00 6.14504457e-01 -8.57339799e-01
8.13861668e-01 -3.32944751e-01 -5.10501444e-01 -1.44492888e+00
-1.61429778e-01 -6.21783137e-01 -1.54241116e-03 1.33812094e+00
3.92258793e-01 -3.57050627e-01 1.20061971e-01 8.16009343e-01
-2.74800122e-01 -4.52817202e-01 -8.10187280e-01 -7.78709650e-01
-5.13935804e-01 -9.85266507e-01 8.73862147e-01 1.39997852e+00
9.57790852e-01 5.27574182e-01 -2.23245978e-01 -2.68534034e-01
2.96548158e-01 2.66666025e-01 3.87914598e-01 -1.50241518e+00
-4.19965893e-01 -4.69673902e-01 -7.79987454e-01 2.26649404e-01
7.37321258e-01 -1.25794196e+00 -5.30159920e-02 -1.21965110e+00
8.32597688e-02 -3.78952652e-01 -3.92636269e-01 7.28238165e-01
3.58433187e-01 4.60105151e-01 3.26918215e-01 -2.35233102e-02
-7.29638338e-01 6.34880245e-01 8.50946844e-01 8.81136283e-02
-2.17198536e-01 -3.11830640e-01 -1.10983074e+00 1.10757852e+00
7.46733010e-01 -4.71230000e-01 -2.84362137e-01 4.92729962e-01
1.11376023e+00 -2.28414610e-01 6.22274876e-01 -9.75821376e-01
-3.59814674e-01 -3.88410479e-01 -6.46223053e-02 7.80512169e-02
3.36726934e-01 -1.08887458e+00 8.21341097e-01 1.64144278e-01
-8.24351549e-01 -3.73806357e-01 1.50810793e-01 1.62742496e-01
-2.60150403e-01 -5.56805253e-01 5.41386366e-01 1.50906116e-01
-9.38223362e-01 -5.43878734e-01 -3.18947285e-01 -1.44794896e-01
1.12195349e+00 -1.29189879e-01 -2.11938888e-01 -5.75215399e-01
-1.11349416e+00 -5.67276657e-01 4.75526452e-01 6.30381942e-01
6.20800197e-01 -1.62246346e+00 -4.45789963e-01 -1.72135442e-01
7.95472741e-01 -8.79957676e-01 2.52750844e-01 4.89567518e-01
-1.90788016e-01 8.28631520e-02 -4.59697664e-01 1.14035979e-01
-8.52722764e-01 8.27264011e-01 2.18039989e-01 5.20775542e-02
-5.33712745e-01 6.51481271e-01 -1.87011406e-01 -2.89975077e-01
3.12949866e-02 -3.16042811e-01 -6.66419864e-01 1.45531386e-01
4.88754302e-01 2.32830271e-01 -6.66626021e-02 -7.58614779e-01
-5.56626320e-01 1.14627212e-01 5.10040402e-01 -2.65537798e-01
1.30744243e+00 7.23176748e-02 -5.02705336e-01 9.39125896e-01
7.27735698e-01 7.17683434e-02 -2.04207927e-01 1.68163687e-01
-1.23935156e-01 -7.69427791e-02 -2.09953412e-02 -1.37627065e+00
-5.58039367e-01 4.29802418e-01 3.14212739e-01 4.40228432e-01
1.22639608e+00 3.64213198e-01 2.96727002e-01 3.91876906e-01
5.78307927e-01 -1.16073132e+00 -1.25726610e-01 3.42729688e-01
1.44215882e+00 -6.20555937e-01 -8.28275178e-03 -6.43577456e-01
-1.24616146e+00 1.36227143e+00 3.44482303e-01 1.00097291e-01
4.58961129e-01 -8.46191794e-02 -1.49637163e-01 -4.49881613e-01
-9.38576818e-01 -1.96315318e-01 2.57099330e-01 7.09096670e-01
4.62764204e-01 6.02584660e-01 -6.57260299e-01 1.68725717e+00
-4.73574907e-01 3.16976696e-01 9.28913653e-01 4.15163308e-01
-3.95194381e-01 -7.90794313e-01 -4.26706582e-01 1.02384821e-01
-6.61426783e-02 -5.48018254e-02 -9.41285491e-01 6.63307011e-01
6.21937215e-01 1.09097052e+00 -1.41545966e-01 -6.55365229e-01
5.62125802e-01 1.20065264e-01 4.83081996e-01 -6.53033614e-01
-7.39345372e-01 -5.55725276e-01 4.85289842e-01 -4.77824420e-01
-8.05052280e-01 -2.11242095e-01 -1.73960197e+00 -3.37712288e-01
-2.72284776e-01 5.40043950e-01 8.84314716e-01 1.11454260e+00
2.53584236e-01 6.22260749e-01 4.54213828e-01 -6.91676378e-01
9.01168808e-02 -7.72176087e-01 -9.23926055e-01 8.63911688e-01
-6.13473713e-01 -8.79499137e-01 -5.92349350e-01 4.14671838e-01] | [9.924309730529785, 7.4331207275390625] |
81e096d2-9607-456f-89a7-41e39ac21976 | structured-domain-adaptation-for-unsupervised | 2003.06650 | null | https://arxiv.org/abs/2003.06650v3 | https://arxiv.org/pdf/2003.06650v3.pdf | Structured Domain Adaptation with Online Relation Regularization for Unsupervised Person Re-ID | Unsupervised domain adaptation (UDA) aims at adapting the model trained on a labeled source-domain dataset to an unlabeled target-domain dataset. The task of UDA on open-set person re-identification (re-ID) is even more challenging as the identities (classes) do not have overlap between the two domains. One major research direction was based on domain translation, which, however, has fallen out of favor in recent years due to inferior performance compared to pseudo-label-based methods. We argue that the domain translation has great potential on exploiting the valuable source-domain data but existing methods did not provide proper regularization on the translation process. Specifically, previous methods only focus on maintaining the identities of the translated images while ignoring the inter-sample relations during translation. To tackle the challenges, we propose an end-to-end structured domain adaptation framework with an online relation-consistency regularization term. During training, the person feature encoder is optimized to model inter-sample relations on-the-fly for supervising relation-consistency domain translation, which in turn, improves the encoder with informative translated images. The encoder can be further improved with pseudo labels, where the source-to-target translated images with ground-truth identities and target-domain images with pseudo identities are jointly used for training. In the experiments, our proposed framework is shown to achieve state-of-the-art performance on multiple UDA tasks of person re-ID. With the synthetic-to-real translated images from our structured domain-translation network, we achieved second place in the Visual Domain Adaptation Challenge (VisDA) in 2020. | ['Hongsheng Li', 'Xiaogang Wang', 'Rui Zhao', 'Dapeng Chen', 'Yixiao Ge', 'Feng Zhu'] | 2020-03-14 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 4.45313156e-01 1.17916808e-01 -2.63469577e-01 -7.00705171e-01
-8.12431455e-01 -5.19407988e-01 8.68950188e-01 -6.24079645e-01
-4.07508254e-01 1.01268554e+00 3.83472234e-01 3.65007967e-01
3.48936558e-01 -4.19757485e-01 -7.75729299e-01 -4.98246372e-01
5.53313673e-01 1.08620322e+00 -1.60314336e-01 -1.48100361e-01
-4.37959075e-01 -4.06800918e-02 -1.43825603e+00 4.17760134e-01
9.32831764e-01 7.41520107e-01 1.40558602e-02 1.81413800e-01
6.42940998e-02 3.52882683e-01 -5.47754824e-01 -9.82881784e-01
6.15113735e-01 -8.92475426e-01 -9.18095946e-01 4.01218772e-01
7.86102653e-01 -3.73770267e-01 -3.78824234e-01 1.18431997e+00
7.36340702e-01 1.47483662e-01 8.30880642e-01 -1.33215439e+00
-1.38816774e+00 1.68430284e-01 -5.28093278e-01 -2.21480668e-01
4.73093331e-01 3.04763541e-02 6.12467587e-01 -1.11916816e+00
1.11860085e+00 1.41169488e+00 5.73528886e-01 1.27051830e+00
-1.22949481e+00 -8.33855391e-01 7.10861832e-02 4.40529019e-01
-1.50200236e+00 -5.61119139e-01 7.40976453e-01 -5.01368880e-01
5.45153439e-01 -3.93645950e-02 3.86100292e-01 1.85883200e+00
-4.76146251e-01 5.93598485e-01 1.26382625e+00 -6.42319381e-01
-9.86022726e-02 7.27203608e-01 -1.86442971e-01 3.25186938e-01
2.35741884e-01 3.05137575e-01 -5.41989684e-01 9.83705148e-02
7.36749709e-01 -1.91757783e-01 -1.02935158e-01 -6.65929794e-01
-1.31248021e+00 6.61031365e-01 2.39803120e-01 5.38251214e-02
-2.72586942e-02 -5.37252247e-01 5.38735628e-01 3.67614031e-01
5.81238091e-01 1.57087296e-01 -2.39651904e-01 1.94619462e-01
-7.67328560e-01 3.56931239e-01 4.80139881e-01 1.39696586e+00
6.73324049e-01 -1.37816146e-01 -4.33840603e-01 1.28665590e+00
4.44726609e-02 8.35045457e-01 5.76270282e-01 -7.44325936e-01
5.70276439e-01 5.61646760e-01 2.77595878e-01 -5.21503806e-01
7.18879551e-02 -4.23112750e-01 -1.11726558e+00 8.15968513e-02
8.44862580e-01 -5.27284369e-02 -9.72697794e-01 2.11026382e+00
3.12142342e-01 1.28043503e-01 3.06639850e-01 1.15379918e+00
8.48790646e-01 2.67740458e-01 1.71956420e-01 -8.57024193e-02
1.52786946e+00 -1.14923942e+00 -7.29194343e-01 -3.89787138e-01
2.45611206e-01 -6.79424524e-01 1.11818194e+00 -7.35687390e-02
-8.66041899e-01 -1.00269842e+00 -1.00364149e+00 -2.16951385e-01
-3.25631887e-01 4.83622789e-01 -1.94692407e-02 7.10591316e-01
-8.44516397e-01 5.14723472e-02 -3.95599604e-01 -9.99861479e-01
4.85093206e-01 3.05670619e-01 -7.20808804e-01 -4.20538306e-01
-1.44780207e+00 1.07626605e+00 2.90713876e-01 -3.79415303e-01
-1.05442750e+00 -5.09768963e-01 -8.25219750e-01 -3.78689796e-01
2.89133519e-01 -1.02435732e+00 1.01005137e+00 -1.58326662e+00
-1.42839086e+00 1.67524779e+00 -3.51304561e-01 -5.98932207e-01
7.78503656e-01 -9.99409109e-02 -9.05877411e-01 -1.16483569e-02
6.16090178e-01 8.20788503e-01 9.51496124e-01 -1.47709918e+00
-6.73306644e-01 -6.06000543e-01 -2.11553887e-01 4.10474241e-01
-4.94347453e-01 1.97616950e-01 -5.54619610e-01 -8.37634683e-01
-3.40650916e-01 -1.14860260e+00 2.17594281e-01 -6.20263722e-03
-2.91283816e-01 -5.45081049e-02 6.74124837e-01 -1.00290656e+00
7.13455558e-01 -2.03456879e+00 2.42598087e-01 -7.54231736e-02
-4.96373549e-02 3.82647783e-01 -1.65270954e-01 8.79322644e-04
-3.31453472e-01 -4.01535034e-01 -3.39244127e-01 -6.51965678e-01
4.88126129e-02 1.43558800e-01 -2.19271347e-01 4.68679279e-01
1.05672754e-01 8.48517835e-01 -8.51967275e-01 -5.80939353e-01
-9.00661759e-03 5.00809014e-01 -3.30936283e-01 5.02367556e-01
2.83626225e-02 1.05151570e+00 -3.63828167e-02 5.81654966e-01
8.77580345e-01 -2.54260480e-01 2.49601424e-01 -1.75994933e-01
2.73187250e-01 -6.21108897e-02 -9.87448215e-01 1.89185202e+00
-2.08559230e-01 4.65418428e-01 -6.73056319e-02 -1.04136264e+00
1.00926352e+00 3.80477935e-01 5.15156031e-01 -1.03548241e+00
-1.03449419e-01 2.31886789e-01 -2.45565817e-01 -2.84222752e-01
3.19036812e-01 -2.46685594e-01 -2.12926835e-01 4.29200530e-01
3.74756902e-01 5.07200062e-01 6.66982979e-02 8.57708231e-02
3.97853255e-01 7.43030846e-01 3.06641191e-01 -1.24913692e-01
8.40639472e-01 2.70367622e-01 6.97274864e-01 4.02174413e-01
-5.33089042e-01 9.52176869e-01 -1.26375511e-01 -3.94198835e-01
-1.47672272e+00 -1.28760231e+00 -1.30286440e-01 1.11941206e+00
3.18671942e-01 -5.03530204e-02 -1.04045510e+00 -9.48253393e-01
-2.01009512e-01 6.88105583e-01 -7.81027317e-01 -2.55135030e-01
-5.83752096e-01 -4.65643585e-01 7.94828653e-01 3.95643085e-01
9.49611723e-01 -8.06473613e-01 6.54067844e-02 2.97105741e-02
-8.28008831e-01 -1.52590096e+00 -9.42626953e-01 -2.93254048e-01
-3.70862544e-01 -9.64506745e-01 -1.41835260e+00 -1.25120664e+00
1.04949534e+00 2.43044093e-01 1.14825106e+00 -7.06079006e-01
1.11540392e-01 4.55900460e-01 -3.11606109e-01 -2.29211301e-01
-5.90707719e-01 -1.36004508e-01 4.69352126e-01 5.30031860e-01
8.99951279e-01 -4.24275398e-01 -4.21360314e-01 7.54541218e-01
-4.05000269e-01 2.08331987e-01 4.42301571e-01 1.23953569e+00
7.55280674e-01 -2.83541262e-01 6.82650805e-01 -1.02871418e+00
4.11634147e-01 -2.00614274e-01 -1.90583795e-01 6.01973295e-01
-6.35466218e-01 -3.26451845e-02 7.33293474e-01 -7.55996585e-01
-1.61459243e+00 2.62691289e-01 2.50107765e-01 -6.22636735e-01
-4.07718867e-01 -2.22534925e-01 -7.00213730e-01 8.79685357e-02
9.58501518e-01 4.56370920e-01 1.48341423e-02 -4.60072935e-01
4.85240221e-01 8.82011950e-01 1.08472288e+00 -6.94656670e-01
1.00163710e+00 6.24986410e-01 -3.60784084e-01 -3.74680877e-01
-7.84946561e-01 -5.26438594e-01 -9.90853786e-01 -4.02940474e-02
9.42782938e-01 -1.51022875e+00 1.44271059e-02 5.34335613e-01
-1.16907692e+00 -1.16817787e-01 -4.25878137e-01 3.81884098e-01
-6.43749952e-01 4.23179060e-01 -2.13472262e-01 -4.54023868e-01
-2.19961479e-01 -9.76189911e-01 9.25554335e-01 2.16573641e-01
-2.99506843e-01 -9.10198390e-01 1.09914839e-01 6.63362086e-01
1.57387063e-01 1.36611938e-01 5.80524087e-01 -8.22897434e-01
-1.66517437e-01 -1.00974748e-02 -4.45302248e-01 4.70117390e-01
3.22276950e-01 -8.72228146e-01 -1.07033122e+00 -4.91737694e-01
-1.91573918e-01 -4.16073382e-01 5.44530511e-01 1.07642472e-01
6.33027613e-01 -2.27135330e-01 -3.58721435e-01 7.94046998e-01
1.14706683e+00 -1.90114617e-01 5.53757012e-01 3.78711849e-01
8.54768693e-01 8.01957786e-01 8.84431601e-01 3.29296023e-01
7.17377484e-01 1.22382057e+00 -1.26261324e-01 -2.43554443e-01
-8.54751229e-01 -7.13064671e-01 5.18763065e-01 3.04092139e-01
-3.20106268e-01 3.24528627e-02 -6.82890654e-01 8.19878399e-01
-1.77745700e+00 -1.05828810e+00 9.17939693e-02 2.51707363e+00
9.42890584e-01 -2.61366665e-01 5.54831445e-01 -4.68416035e-01
1.19440544e+00 -3.29858154e-01 -7.25308299e-01 3.74907479e-02
-5.33398390e-01 -8.65966156e-02 5.66811800e-01 2.99338996e-01
-1.18201840e+00 1.04644632e+00 5.30732489e+00 7.96479285e-01
-7.30518162e-01 5.80274880e-01 4.73939776e-01 6.83942661e-02
8.45427997e-03 -1.30245134e-01 -1.05536366e+00 5.42868853e-01
8.57328832e-01 -2.77248770e-01 6.08945310e-01 9.82105374e-01
-7.90150184e-03 3.51230323e-01 -1.37275040e+00 1.38588798e+00
5.34230828e-01 -8.38618159e-01 2.24911228e-01 1.68635800e-01
1.19066191e+00 -3.96519005e-01 1.54849172e-01 4.84192818e-01
3.56319636e-01 -8.41876924e-01 7.17830002e-01 3.35228294e-01
1.65508699e+00 -6.90764189e-01 7.08393216e-01 2.20570400e-01
-1.03692687e+00 5.68569861e-02 -5.96677721e-01 1.85236618e-01
1.72779009e-01 4.91086207e-02 -1.10055912e+00 7.34762549e-01
7.50626564e-01 8.71546388e-01 -6.63760126e-01 5.07788897e-01
-1.53163031e-01 2.56990671e-01 6.37300238e-02 5.63628495e-01
-2.64072120e-01 -3.01084042e-01 5.16163588e-01 1.17843306e+00
3.30515891e-01 -2.51066983e-01 5.35785593e-02 9.46429551e-01
-2.51777560e-01 -4.74354252e-02 -6.91973746e-01 1.93393752e-01
3.46900195e-01 7.77332067e-01 -1.80010572e-01 -5.74854851e-01
-6.91010714e-01 1.65913022e+00 3.47360253e-01 5.73586702e-01
-9.52304423e-01 2.84247901e-02 7.54256248e-01 3.98984432e-01
6.31941333e-02 2.04028636e-01 -3.39177459e-01 -1.50962651e+00
1.12759419e-01 -1.13006425e+00 5.58492422e-01 -5.91533720e-01
-1.80154920e+00 6.36437654e-01 1.27993807e-01 -1.66638076e+00
-4.74339783e-01 -4.09024000e-01 3.26030217e-02 1.18950832e+00
-1.37645268e+00 -1.91246176e+00 -3.70678723e-01 1.17810440e+00
6.30820930e-01 -7.50780940e-01 1.04577184e+00 6.22428179e-01
-4.03927475e-01 1.26179159e+00 2.77117819e-01 4.99559045e-01
1.65612042e+00 -1.03812540e+00 5.13803184e-01 1.16873026e+00
-1.19917430e-02 5.31036258e-01 5.97100317e-01 -7.91190863e-01
-8.00283253e-01 -1.40414751e+00 1.11459589e+00 -7.07022607e-01
8.81738588e-02 -6.60173595e-01 -7.38063514e-01 9.62362647e-01
2.15006024e-01 1.66812435e-01 6.70813799e-01 -6.90794876e-03
-5.57377279e-01 -1.01955153e-01 -1.36862731e+00 5.21758318e-01
1.55850816e+00 -6.76266789e-01 -6.54755473e-01 4.38839406e-01
4.90589440e-01 -3.42348933e-01 -7.31662095e-01 1.60997942e-01
5.43113589e-01 -6.99940324e-01 1.33919907e+00 -7.55428195e-01
2.32677430e-01 -3.79442692e-01 -1.55343577e-01 -1.30431414e+00
-5.14626920e-01 -4.16915953e-01 7.14109913e-02 1.64020622e+00
7.40409344e-02 -5.51261663e-01 7.69641340e-01 7.49154985e-01
1.34004399e-01 8.41428936e-02 -9.58943546e-01 -1.04645324e+00
2.00103179e-01 1.01643100e-01 6.39937580e-01 1.17731261e+00
-2.95345187e-01 5.34109652e-01 -9.43994224e-01 1.54217079e-01
1.00407028e+00 2.76538581e-02 1.15257382e+00 -1.11421704e+00
-3.38715822e-01 -9.71328374e-03 -2.16259733e-01 -1.10172880e+00
4.02503133e-01 -1.11263049e+00 -2.01101989e-01 -1.24479890e+00
5.75780213e-01 -2.56602883e-01 -1.09875053e-01 3.15882206e-01
-2.22824261e-01 4.59021449e-01 1.44227877e-01 5.57781100e-01
-6.43371463e-01 5.66392481e-01 1.39551330e+00 -4.31237727e-01
4.82430011e-02 -1.37626380e-02 -7.92594194e-01 4.22061294e-01
4.73698050e-01 -4.62793231e-01 -4.40890253e-01 -4.57547575e-01
-4.70759481e-01 -1.86267227e-01 5.39238274e-01 -1.07108009e+00
1.73927158e-01 -4.22096588e-02 7.13506579e-01 -1.52458668e-01
3.42795491e-01 -8.74702334e-01 3.47686380e-01 3.45385335e-02
-4.90863413e-01 -2.91966081e-01 -1.46975189e-01 6.43358588e-01
-2.18583211e-01 6.44364879e-02 1.08436155e+00 -2.15111926e-01
-1.00027966e+00 3.75312656e-01 3.13287079e-01 3.10362667e-01
1.04595959e+00 -4.88917172e-01 -1.05135679e-01 -3.86853486e-01
-9.16122317e-01 5.97392060e-02 9.07804072e-01 6.19033039e-01
4.10896242e-01 -1.83664334e+00 -1.17695594e+00 3.74708474e-01
6.53722405e-01 -2.44556993e-01 3.97449970e-01 4.28910524e-01
2.19732039e-02 3.45268488e-01 -7.11001635e-01 -5.32941341e-01
-1.31985021e+00 8.04695904e-01 4.35996264e-01 -4.47357178e-01
-4.93664145e-01 8.51481140e-01 6.24208868e-01 -7.12329149e-01
1.41346544e-01 5.19327462e-01 -1.35037601e-01 -1.19773142e-01
6.51886940e-01 2.45296285e-01 -1.33720189e-01 -1.27397168e+00
-4.57314938e-01 7.10935235e-01 -1.57865152e-01 -3.05865437e-01
9.49040115e-01 -5.40405095e-01 2.25378066e-01 5.02344035e-02
1.07079387e+00 -1.18920498e-01 -1.55928671e+00 -8.27431023e-01
-2.68704385e-01 -6.13031447e-01 -7.09973156e-01 -1.28258789e+00
-7.41516352e-01 8.06884110e-01 9.96977270e-01 -6.04239464e-01
1.13391054e+00 -3.16006057e-02 8.69109750e-01 1.15452369e-03
6.08275712e-01 -1.36467052e+00 1.13222815e-01 3.22167277e-01
9.41685021e-01 -1.65789592e+00 -1.26021072e-01 -3.66040319e-01
-1.23096275e+00 7.45498359e-01 8.92647684e-01 8.11606944e-02
1.31513745e-01 -2.56657243e-01 1.51611686e-01 4.35301006e-01
-1.48511484e-01 -3.84548903e-01 4.20750260e-01 1.52955091e+00
1.32542267e-01 2.83708066e-01 1.41915074e-03 9.55706000e-01
-1.66659459e-01 1.31123036e-01 1.11354895e-01 2.85846412e-01
1.05972521e-01 -1.57413208e+00 -6.87739551e-01 -3.77723761e-02
-1.91565245e-01 1.46277606e-01 -6.00567400e-01 7.01965392e-01
5.80128312e-01 8.73680830e-01 -1.71611220e-01 -3.61577302e-01
4.76380050e-01 2.92088032e-01 6.76075935e-01 -5.64729214e-01
-2.92957187e-01 -1.41540408e-01 2.17128024e-01 -2.18484655e-01
-5.20378590e-01 -8.98425519e-01 -8.91224146e-01 -2.19203830e-01
2.22329497e-01 -5.42920306e-02 2.29451612e-01 7.72896111e-01
5.05786121e-01 1.93980768e-01 3.95864338e-01 -5.70680499e-01
-5.30424178e-01 -1.09507990e+00 -4.17152256e-01 1.26873279e+00
1.11762606e-01 -9.13685262e-01 2.26791538e-02 6.90368354e-01] | [14.726445198059082, 1.0641043186187744] |
ca1ef739-c8a2-42db-980e-90d553db38ef | mathsf-g-2retro-two-step-graph-generative | 2206.04882 | null | https://arxiv.org/abs/2206.04882v3 | https://arxiv.org/pdf/2206.04882v3.pdf | $\mathsf{G^2Retro}$ as a Two-Step Graph Generative Models for Retrosynthesis Prediction | Retrosynthesis is a procedure where a target molecule is transformed into potential reactants and thus the synthesis routes can be identified. Recently, computational approaches have been developed to accelerate the design of synthesis routes. In this paper, we develop a generative framework $\mathsf{G^2Retro}$ for one-step retrosynthesis prediction. $\mathsf{G^2Retro}$ imitates the reversed logic of synthetic reactions. It first predicts the reaction centers in the target molecules (products), identifies the synthons needed to assemble the products, and transforms these synthons into reactants. $\mathsf{G^2Retro}$ defines a comprehensive set of reaction center types, and learns from the molecular graphs of the products to predict potential reaction centers. To complete synthons into reactants, $\mathsf{G^2Retro}$ considers all the involved synthon structures and the product structures to identify the optimal completion paths, and accordingly attaches small substructures sequentially to the synthons. Here we show that $\mathsf{G^2Retro}$ is able to better predict the reactants for given products in the benchmark dataset than the state-of-the-art methods. | ['Xia Ning', 'Huan Sun', 'James R. Fuchs', 'Oluwatosin R. Ayinde', 'Ziqi Chen'] | 2022-06-10 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 4.28300649e-01 1.61619306e-01 -3.80045772e-01 -3.65588404e-02
-4.18785930e-01 -1.05784726e+00 5.70188344e-01 3.29721212e-01
2.63574898e-01 8.54732275e-01 8.80013183e-02 -6.45617187e-01
2.55645383e-02 -1.13102794e+00 -8.78884375e-01 -9.16026890e-01
-1.96835831e-01 4.95195717e-01 2.60698736e-01 -4.07200247e-01
5.88134885e-01 7.80656219e-01 -1.38062632e+00 2.52067924e-01
1.00088477e+00 9.71720755e-01 4.18179244e-01 5.13917446e-01
-3.78997236e-01 6.18400097e-01 -3.86614323e-01 -6.39822781e-01
3.26543570e-01 -7.92054057e-01 -7.44790912e-01 -3.68209958e-01
-9.73990485e-02 2.49070108e-01 -1.86276019e-01 8.46966684e-01
1.11149214e-01 2.12994009e-01 1.30001032e+00 -8.98658395e-01
-4.23245788e-01 8.59777272e-01 -1.10625319e-01 -9.79123265e-02
3.53922725e-01 -1.17199095e-02 1.30093539e+00 -1.26301694e+00
8.28872442e-01 9.44335103e-01 2.00084031e-01 3.40041161e-01
-1.17218018e+00 -9.60621655e-01 1.67377114e-01 -1.28199279e-01
-1.23582625e+00 -3.30343723e-01 5.98467827e-01 -6.95851505e-01
1.01992345e+00 4.43352133e-01 6.27090752e-01 7.88486600e-01
4.95525777e-01 4.24486220e-01 5.48211634e-01 -3.06743711e-01
4.51383293e-01 -2.57415056e-01 -6.94031864e-02 9.13425922e-01
8.05960095e-04 3.17747086e-01 -4.43616807e-01 1.31563127e-01
5.80186486e-01 3.56595367e-01 7.52424076e-02 -6.40327781e-02
-1.17258418e+00 8.98727477e-01 7.46339738e-01 2.31373951e-01
-2.25662649e-01 -8.99629761e-03 1.27715975e-01 -2.94314057e-01
7.44964601e-03 1.12146139e+00 -3.65237862e-01 2.74997234e-01
-7.37046301e-01 5.29871769e-02 7.03554749e-01 1.34377277e+00
1.07483113e+00 4.30460423e-02 1.61853015e-01 4.24980521e-01
1.89209491e-01 9.54725146e-02 -2.02994034e-01 -7.95939684e-01
5.69791973e-01 7.68075883e-01 2.04123944e-01 -6.08429849e-01
-5.39878368e-01 -5.91139682e-02 -7.42422223e-01 -2.63637573e-01
2.47265548e-01 -1.92104444e-01 -1.01337135e+00 1.43460429e+00
3.81000787e-01 -3.27645361e-01 8.97242781e-03 4.16065961e-01
8.93789649e-01 1.41951478e+00 2.48367190e-01 -6.25240982e-01
9.26765382e-01 -1.19728518e+00 -1.29753143e-01 7.56429508e-02
7.70072997e-01 -9.69432235e-01 5.10539651e-01 5.15984774e-01
-1.20670938e+00 -4.68995810e-01 -1.11563861e+00 9.57657695e-02
-6.20261669e-01 1.58325434e-01 8.47803771e-01 6.36079252e-01
-5.64909399e-01 1.10810018e+00 -5.10690212e-01 1.21420108e-01
8.24321508e-02 6.03479326e-01 -1.85484856e-01 -3.70654352e-02
-9.38820839e-01 7.88256109e-01 8.01569700e-01 1.69932440e-01
-1.71906126e+00 -9.61816490e-01 -8.59631300e-01 -7.80030489e-02
7.25370347e-01 -5.07538795e-01 8.75374794e-01 -6.87647820e-01
-1.48119843e+00 4.50859070e-01 -2.25960538e-02 6.61421120e-02
8.76254216e-02 5.37836671e-01 -4.92594630e-01 -8.31169188e-02
1.90417171e-01 8.04208934e-01 6.55451298e-01 -1.11920524e+00
-6.86705947e-01 -9.89033654e-02 -4.10684720e-02 -9.81300697e-03
3.43958110e-01 3.67004685e-02 -2.13114172e-01 -6.58719361e-01
3.50387692e-01 -9.95380700e-01 -4.97197628e-01 -5.78135729e-01
-8.69131565e-01 -3.20129067e-01 -6.56270161e-02 -1.75179571e-01
1.22052801e+00 -1.85786116e+00 7.12266445e-01 5.63362479e-01
2.98853129e-01 6.73300400e-02 -1.67635549e-02 1.09522891e+00
-5.53634107e-01 2.16011837e-01 -8.53020549e-02 1.50291041e-01
-1.19961701e-01 -2.75205463e-01 -2.21093267e-01 2.20516250e-01
-5.16852476e-02 5.89829266e-01 -1.03454888e+00 -8.20162073e-02
1.87213346e-01 1.34146631e-01 -5.35410345e-01 1.75101772e-01
-8.83817375e-01 4.05820608e-01 -7.29309201e-01 8.22812140e-01
3.51948947e-01 -1.15169920e-01 5.94222963e-01 -3.67577851e-01
-6.34161770e-01 6.51010156e-01 -8.42329621e-01 1.51382244e+00
-6.32298961e-02 -1.49733961e-01 -4.92314935e-01 -8.96124482e-01
1.25905967e+00 2.37989277e-01 6.27454400e-01 -4.32392776e-01
2.25983784e-01 4.76524860e-01 -2.55793445e-02 -2.46586949e-02
4.31439787e-01 -4.84550446e-01 -1.99796930e-01 2.98496544e-01
1.10123649e-01 -3.96007627e-01 8.98662508e-01 3.33072335e-01
7.75456667e-01 3.98959488e-01 5.01732171e-01 -2.59029806e-01
6.78154230e-01 1.40812069e-01 4.17152435e-01 5.03961742e-01
4.60284919e-01 3.16888243e-01 5.07808983e-01 -6.67096853e-01
-1.36693501e+00 -1.28098822e+00 4.31891918e-01 1.29630458e+00
1.11333549e-01 -9.13682699e-01 -7.39004612e-01 -7.14718997e-01
-4.61828202e-01 7.88898110e-01 -4.32396263e-01 -3.21099162e-01
-4.42375779e-01 -4.86572951e-01 4.29583311e-01 5.41015446e-01
7.22035626e-03 -1.02778757e+00 1.61151677e-01 4.83773410e-01
4.27447446e-02 -3.37576777e-01 -8.84032488e-01 5.02613902e-01
-5.72632670e-01 -1.25196958e+00 -3.12181264e-01 -8.68526340e-01
9.04026985e-01 1.47924885e-01 6.43859804e-01 -7.73039013e-02
-4.81605887e-01 -4.50374931e-01 -2.65751719e-01 -4.49493766e-01
-8.45195651e-01 1.99382156e-02 -2.26144329e-01 -1.89157024e-01
-3.23054045e-01 -6.97625220e-01 -7.56393373e-01 3.48855346e-01
-8.16480875e-01 3.28263342e-01 6.02852762e-01 4.55843568e-01
1.05028784e+00 2.89569438e-01 4.50514168e-01 -9.40631986e-01
6.37586191e-02 -5.77353477e-01 -9.08137321e-01 4.72622424e-01
-6.74858272e-01 5.67024589e-01 1.40684903e+00 -2.87204593e-01
-8.83928299e-01 4.25134748e-01 -2.19527677e-01 -1.24703810e-01
-7.64236087e-03 6.00150287e-01 -4.81016755e-01 3.03627431e-01
7.27391541e-01 5.14284670e-01 -4.58233386e-01 -4.28406954e-01
7.00793564e-01 3.52626331e-02 3.30991298e-01 -9.31661248e-01
4.39710081e-01 1.14822708e-01 6.44242167e-01 -6.54640436e-01
-7.31561899e-01 -2.34417364e-01 -5.99760234e-01 2.32025124e-02
8.02171826e-01 -5.80208063e-01 -1.08626819e+00 -9.28362980e-02
-9.98119533e-01 -6.38491660e-02 -1.42521009e-01 2.22682998e-01
-7.33453929e-01 2.89321840e-01 -4.72867012e-01 -4.43993539e-01
-4.58432436e-01 -1.67874789e+00 6.65159225e-01 5.27286753e-02
-2.73365140e-01 -7.72782683e-01 7.24515021e-02 2.21627563e-01
-3.45046043e-01 2.80049562e-01 1.56967735e+00 -8.25890899e-01
-9.49368656e-01 3.80480364e-02 1.40525684e-01 -6.87305555e-02
3.74594271e-01 4.36104596e-01 -2.39352793e-01 -6.05223626e-02
-6.18431747e-01 1.06687784e-01 8.05111051e-01 1.97699726e-01
9.86346066e-01 -5.04749775e-01 -6.68126702e-01 4.48072106e-01
1.09459126e+00 1.14044774e+00 6.72302365e-01 9.14073959e-02
6.99409902e-01 6.21768296e-01 6.22362852e-01 4.31221604e-01
1.42620593e-01 2.83150822e-01 6.10980272e-01 1.60096109e-01
1.73320055e-01 -8.24178576e-01 5.53849816e-01 3.32272738e-01
-3.06213200e-01 -5.63138425e-01 -6.98171020e-01 -1.98373813e-02
-1.33476388e+00 -8.63534272e-01 -1.85449660e-01 2.07136154e+00
8.95479620e-01 1.62289105e-02 4.40523207e-01 2.56482720e-01
7.37198055e-01 2.08621621e-01 -4.99003202e-01 -4.36508566e-01
3.22883248e-01 8.27699006e-01 4.87264037e-01 4.64641899e-01
-7.70835996e-01 1.32774401e+00 6.10082245e+00 9.10004616e-01
-1.14879692e+00 -4.92776304e-01 4.79751408e-01 1.75154544e-02
-5.60000360e-01 5.39915502e-01 -1.03213668e+00 5.33944368e-01
7.97521651e-01 -1.80411041e-01 6.54612005e-01 6.61951900e-01
2.56340891e-01 2.01427504e-01 -1.61308980e+00 5.21294534e-01
-3.27325761e-01 -2.21466899e+00 4.36854690e-01 -1.65749267e-01
8.37977588e-01 -5.63889742e-01 -5.01752719e-02 -2.39199288e-02
5.69283783e-01 -1.01992130e+00 1.15533507e+00 2.31059968e-01
8.31752896e-01 -9.43826675e-01 -1.97052330e-01 2.34393582e-01
-1.48954165e+00 -2.23521158e-01 -1.29314780e-01 2.85073429e-01
-3.04415058e-02 2.83125043e-01 -1.42999613e+00 9.67424989e-01
1.01359829e-01 9.12771046e-01 -1.69191837e-01 5.33205628e-01
-6.08611584e-01 2.64612108e-01 -6.28163218e-02 -4.13448215e-01
3.66804838e-01 -8.35032046e-01 2.30175778e-01 9.88224626e-01
6.57883108e-01 4.07198817e-01 2.49380738e-01 1.05166543e+00
-3.42359930e-01 3.35757673e-01 -4.52091575e-01 -6.20671391e-01
5.55600226e-01 9.86633301e-01 -1.18469334e+00 -3.29405546e-01
4.56005847e-03 5.50229609e-01 1.06132306e-01 3.04113060e-01
-7.41473496e-01 -4.15025383e-01 5.42700529e-01 2.58936375e-01
4.35514897e-01 -2.07955301e-01 1.62853777e-01 -8.49604487e-01
-4.72518355e-01 -8.69899035e-01 6.46111548e-01 -7.95950055e-01
-7.05950201e-01 3.92142564e-01 -1.99503466e-01 -9.43908691e-01
1.49474889e-01 -5.95146656e-01 -7.40891814e-01 6.40547454e-01
-9.85579908e-01 -8.59914601e-01 2.12197825e-01 4.87286538e-01
7.65688717e-01 -6.04185686e-02 8.11913371e-01 -1.08641341e-01
-7.51755238e-01 2.73579270e-01 3.53403717e-01 -2.86896557e-01
5.19700825e-01 -9.69338059e-01 2.74469405e-01 7.85633683e-01
6.22731894e-02 8.25927615e-01 6.27835274e-01 -8.48153830e-01
-1.74857795e+00 -1.26318479e+00 6.61853194e-01 -1.74228102e-01
5.57065248e-01 -5.70416212e-01 -1.49287254e-01 7.59868443e-01
-8.58862177e-02 -5.09635746e-01 9.00307536e-01 -3.01596999e-01
-5.49558640e-01 -2.53675848e-01 -9.45803761e-01 9.88854349e-01
1.12039578e+00 -3.39806899e-02 -1.99006721e-01 6.11544549e-01
9.56943810e-01 -3.35945278e-01 -1.12862480e+00 -3.87727208e-02
3.14788640e-01 -7.92136014e-01 1.23086047e+00 -7.69214451e-01
7.52770960e-01 -3.36943835e-01 -1.03010632e-01 -1.10438371e+00
-4.44179684e-01 -1.04437721e+00 2.65158236e-01 8.09443057e-01
1.13607824e+00 -4.15674984e-01 8.36292565e-01 4.18012351e-01
-7.60633230e-01 -5.31019628e-01 -5.21527767e-01 -5.45253634e-01
1.54852644e-01 -1.06982039e-02 7.67357051e-01 5.60440481e-01
2.92342454e-01 5.56368887e-01 -1.82343811e-01 3.23328748e-02
2.92080790e-01 4.42995191e-01 4.68464851e-01 -9.65545058e-01
-5.89496382e-02 -5.26122451e-01 1.40051901e-01 -9.98584807e-01
4.29118015e-02 -1.43203294e+00 -1.09958328e-01 -1.42129111e+00
-5.05754091e-02 -9.17409360e-01 -1.61900073e-01 5.42628944e-01
1.39530703e-01 -3.31791848e-01 1.57555237e-01 1.19319357e-01
-4.79183704e-01 4.07368720e-01 1.31891680e+00 -3.69878501e-01
-4.19870734e-01 5.04676327e-02 -9.28880394e-01 4.58705992e-01
5.91147184e-01 -4.94382799e-01 -3.92357349e-01 2.95972764e-01
8.95727277e-01 5.11158109e-01 -3.58451188e-01 -4.70932692e-01
1.81091174e-01 -7.03355908e-01 3.40439439e-01 -8.74039829e-01
3.82705659e-01 -5.61032534e-01 8.53820920e-01 5.72626829e-01
-3.15356880e-01 -1.78425372e-01 -1.99030802e-01 3.70315850e-01
2.22918555e-01 -4.53343272e-01 6.02247953e-01 -5.76728702e-01
-2.28211477e-01 7.96663880e-01 -5.49227595e-01 -3.95718664e-01
1.29915082e+00 -1.14378735e-01 -2.19861507e-01 1.32899910e-01
-1.03957951e+00 7.04147965e-02 3.63646001e-01 2.07012981e-01
5.30952930e-01 -9.72047746e-01 -5.64235635e-02 8.24233964e-02
1.99474907e-03 3.41249436e-01 1.75491925e-02 6.00060999e-01
-8.91953826e-01 6.26256108e-01 1.11643553e-01 -1.44301698e-01
-9.83666241e-01 1.17951691e+00 3.74859333e-01 -1.82593957e-01
1.22905806e-01 1.00133467e+00 4.71146375e-01 -2.34592661e-01
5.43928146e-02 -4.19007480e-01 -7.23692477e-02 1.95076719e-01
2.77833402e-01 5.48244119e-01 5.59758730e-02 -2.77590781e-01
-2.32159629e-01 3.67090374e-01 -3.89910161e-01 3.33538324e-01
1.41642809e+00 3.10310721e-01 -3.26143086e-01 -1.55574024e-01
1.18817782e+00 -1.14163987e-01 -1.12343395e+00 2.20461190e-01
-6.39309213e-02 -6.05700016e-02 -5.03872752e-01 -6.66517556e-01
-8.83431315e-01 7.01188624e-01 -2.89288372e-01 2.98419576e-02
8.96228790e-01 1.94281772e-01 5.17363906e-01 3.84030938e-01
4.51326132e-01 -7.95132041e-01 4.19541836e-01 3.26968431e-01
6.43926382e-01 -6.52792811e-01 -2.78424956e-02 -9.23153520e-01
-4.02399808e-01 1.37994516e+00 2.43800238e-01 1.76991791e-01
4.51350898e-01 -1.00903660e-01 -6.95547342e-01 -2.83899397e-01
-4.96699989e-01 1.41665563e-01 2.68553384e-04 1.77735597e-01
3.63171995e-01 1.36861503e-01 -5.32781422e-01 5.93831480e-01
-4.22709644e-01 -3.56597424e-01 3.82078677e-01 1.03367770e+00
-5.13690948e-01 -1.63624263e+00 -2.34659269e-01 -3.73044759e-02
-1.52728140e-01 -3.88018638e-01 -7.90837407e-01 5.75923204e-01
5.88827610e-01 9.08575773e-01 -4.04380113e-01 -4.09027308e-01
4.22100306e-01 1.24123402e-01 7.54751682e-01 -8.49149287e-01
-7.78734803e-01 3.08794200e-01 1.79738760e-01 -3.40776771e-01
-2.07100525e-01 -5.47099352e-01 -1.77904415e+00 -3.87011915e-01
-3.36661369e-01 7.73428082e-01 6.20127857e-01 8.42577517e-01
4.34357166e-01 3.33148807e-01 1.24899435e+00 -7.04311728e-01
1.73395425e-02 -4.27562326e-01 -4.41458732e-01 -1.25971854e-01
-1.93044871e-01 -5.49376070e-01 -1.63514987e-01 1.04072511e-01] | [4.501950740814209, 6.105088710784912] |
0187cbdf-182f-49d2-b019-5796e580d8be | meshnet-mesh-neural-network-for-3d-shape | 1811.11424 | null | http://arxiv.org/abs/1811.11424v1 | http://arxiv.org/pdf/1811.11424v1.pdf | MeshNet: Mesh Neural Network for 3D Shape Representation | Mesh is an important and powerful type of data for 3D shapes and widely
studied in the field of computer vision and computer graphics. Regarding the
task of 3D shape representation, there have been extensive research efforts
concentrating on how to represent 3D shapes well using volumetric grid,
multi-view and point cloud. However, there is little effort on using mesh data
in recent years, due to the complexity and irregularity of mesh data. In this
paper, we propose a mesh neural network, named MeshNet, to learn 3D shape
representation from mesh data. In this method, face-unit and feature splitting
are introduced, and a general architecture with available and effective blocks
are proposed. In this way, MeshNet is able to solve the complexity and
irregularity problem of mesh and conduct 3D shape representation well. We have
applied the proposed MeshNet method in the applications of 3D shape
classification and retrieval. Experimental results and comparisons with the
state-of-the-art methods demonstrate that the proposed MeshNet can achieve
satisfying 3D shape classification and retrieval performance, which indicates
the effectiveness of the proposed method on 3D shape representation. | ['Yutong Feng', 'Yue Gao', 'Yifan Feng', 'Xibin Zhao', 'Haoxuan You'] | 2018-11-28 | null | null | null | null | ['3d-shape-retrieval', '3d-shape-representation'] | ['computer-vision', 'computer-vision'] | [-3.50410491e-01 -4.25235420e-01 -3.17593925e-02 -1.72775820e-01
-2.11971164e-01 -1.65033773e-01 4.05945003e-01 9.01070461e-02
9.17459205e-02 2.77769119e-01 -1.15305245e-01 -1.48362905e-01
-3.60318780e-01 -1.21456468e+00 -3.72645408e-01 -4.92346346e-01
9.41948369e-02 8.04949224e-01 2.78155506e-01 -2.96728998e-01
3.91104162e-01 1.23150599e+00 -1.81486535e+00 -1.56163853e-02
5.65386653e-01 1.38105977e+00 7.62577280e-02 -7.26777762e-02
-7.45477796e-01 -1.48204684e-01 -5.03529131e-01 -8.84644762e-02
3.18604529e-01 9.20432582e-02 -5.05404055e-01 2.58916885e-01
3.49726379e-01 -2.52363861e-01 -3.45375150e-01 9.28888440e-01
9.06289041e-01 7.16613233e-02 1.09937859e+00 -9.55392659e-01
-8.97104323e-01 -1.92782283e-01 -8.40060771e-01 -2.58291494e-02
1.51252791e-01 -4.25533503e-01 5.73121846e-01 -1.44766676e+00
5.83723128e-01 1.62604022e+00 7.12893963e-01 2.85280854e-01
-6.31567836e-01 -6.61729157e-01 -4.19621915e-03 1.80375539e-02
-1.72119379e+00 -6.20339671e-03 1.06351578e+00 -3.97710681e-01
8.70425761e-01 2.79887229e-01 9.06301856e-01 3.93472970e-01
2.96705633e-01 7.33501136e-01 9.19186294e-01 -2.42670998e-01
-6.44888803e-02 -2.16746312e-02 -5.50898351e-02 8.19131792e-01
3.28292370e-01 -3.19673009e-02 8.10864344e-02 -3.06757867e-01
1.37583053e+00 3.00831765e-01 -1.66457698e-01 -5.57699025e-01
-8.08221340e-01 6.14466131e-01 5.29513955e-01 5.37663400e-01
-4.10082102e-01 -1.12508088e-01 5.61169505e-01 1.75811529e-01
1.03724742e+00 -2.59462923e-01 -1.91095427e-01 1.36065423e-01
-6.15994334e-01 2.98309714e-01 5.86805522e-01 1.21409583e+00
5.62957108e-01 4.46967572e-01 1.27870023e-01 1.07781720e+00
5.12061656e-01 8.76585603e-01 2.55676180e-01 -4.11300927e-01
3.12468618e-01 1.22011888e+00 -2.32164994e-01 -1.62574863e+00
-3.94267380e-01 -2.69207299e-01 -1.21853864e+00 3.11346710e-01
-2.66447604e-01 4.66330945e-01 -1.00202596e+00 8.51463616e-01
6.42819464e-01 1.73978746e-01 -9.25025493e-02 9.58181739e-01
1.71923828e+00 7.31853902e-01 -2.79034674e-01 -9.33548734e-02
1.15080738e+00 -4.77375984e-01 -5.07536888e-01 4.82249141e-01
1.67694196e-01 -8.60579491e-01 7.52841175e-01 8.73271972e-02
-1.23868525e+00 -8.15756440e-01 -8.65327179e-01 3.17385085e-02
-3.88298988e-01 2.00588569e-01 4.89984304e-01 3.49606037e-01
-7.98648655e-01 5.37882864e-01 -5.28869390e-01 -3.87584448e-01
5.90901852e-01 3.90639663e-01 -3.75609159e-01 -1.25975013e-01
-1.00564158e+00 9.03997421e-01 4.53979224e-01 2.58438975e-01
-4.57628548e-01 -5.02461553e-01 -7.94847846e-01 1.34259805e-01
2.35619433e-02 -7.60498285e-01 8.06700468e-01 -2.49402419e-01
-1.16509533e+00 1.05945480e+00 -5.79744112e-03 3.29991728e-01
2.07589462e-01 2.78875887e-01 -3.90497267e-01 1.90286208e-02
-1.12106092e-01 2.65387893e-01 8.02243888e-01 -1.48969364e+00
-3.74414206e-01 -7.63680220e-01 1.44004403e-03 4.01814371e-01
-2.63975412e-01 -3.05571914e-01 -6.30446911e-01 -5.60100436e-01
6.30991280e-01 -5.43287933e-01 2.41971724e-02 2.77881801e-01
2.88943052e-02 -8.13864887e-01 1.30121839e+00 -3.53854239e-01
9.57013547e-01 -2.12401485e+00 9.54449773e-02 5.24695694e-01
2.38429546e-01 4.32819337e-01 -4.08250690e-02 4.49936837e-01
1.06696732e-01 2.17540964e-01 -1.46827728e-01 -2.10574523e-01
-1.38632774e-01 3.62042904e-01 -1.28157362e-01 4.81898576e-01
1.59353733e-01 8.57594669e-01 -5.10717988e-01 -6.13882661e-01
5.08721113e-01 8.39608073e-01 -2.82410681e-01 1.30084008e-01
-7.17655942e-02 3.73334050e-01 -9.00385439e-01 1.06105053e+00
1.12200773e+00 -2.01544777e-01 -3.47901523e-01 -4.82123584e-01
-1.51103303e-01 -3.29110950e-01 -1.55187452e+00 1.59970582e+00
-4.95410293e-01 8.68223011e-02 3.21519941e-01 -1.06659198e+00
1.53418553e+00 3.75710040e-01 5.93710124e-01 -7.15785384e-01
4.94737148e-01 5.75897992e-01 -4.33969676e-01 -5.97778976e-01
2.06923425e-01 -1.47336528e-01 3.22487235e-01 2.91719764e-01
-1.97193101e-01 -6.54003441e-01 -2.14942470e-01 -3.88417780e-01
3.69655967e-01 1.71328753e-01 2.10079372e-01 -3.18333447e-01
1.00575912e+00 -5.30645587e-02 4.04480517e-01 1.27176195e-01
2.52830029e-01 5.71743488e-01 -5.90083748e-02 -9.59779680e-01
-1.14524448e+00 -7.33155906e-01 -4.54514593e-01 3.15797478e-01
4.89178061e-01 -8.60440135e-02 -3.51989865e-01 -3.32847387e-01
5.04841745e-01 6.22929893e-02 -3.51498008e-01 1.16795704e-01
-8.11419249e-01 -2.51866728e-01 4.09041524e-01 4.13052768e-01
9.72363710e-01 -1.17162168e+00 -3.00930440e-01 1.22838050e-01
3.24402034e-01 -7.67990828e-01 -3.53347808e-01 -6.33029044e-01
-1.23343420e+00 -1.12498486e+00 -1.05413210e+00 -1.13295257e+00
7.93369532e-01 6.71124458e-01 1.06886888e+00 6.90389156e-01
-3.26085091e-01 6.11103415e-01 -3.49214792e-01 -5.14303863e-01
-2.16041431e-01 8.32280964e-02 8.97915736e-02 2.55161431e-02
1.28700063e-01 -8.43827784e-01 -5.38504004e-01 3.62754226e-01
-1.03735471e+00 -2.27854326e-01 6.29776001e-01 7.38537788e-01
1.04594064e+00 2.05218017e-01 6.06522083e-01 -6.46598577e-01
7.68391252e-01 -3.51824284e-01 -6.96712852e-01 1.64031148e-01
-5.39022565e-01 -3.92669886e-01 4.87733305e-01 -1.42787948e-01
-7.11207807e-01 -2.37998366e-01 -5.25255799e-01 -1.00909042e+00
-3.77282500e-01 5.58729529e-01 -1.88933492e-01 -5.99962413e-01
2.06127509e-01 6.13619208e-01 4.62310195e-01 -8.11726213e-01
1.01427779e-01 9.85373437e-01 -2.92655975e-02 -6.94746256e-01
6.67838871e-01 5.31239331e-01 4.95521247e-01 -1.12101960e+00
-3.99047017e-01 -4.42284584e-01 -5.47487497e-01 -3.37311000e-01
7.08847582e-01 -8.21900487e-01 -9.32360768e-01 4.81350303e-01
-1.34548938e+00 4.89741951e-01 -4.96575236e-02 2.72938281e-01
-5.97767174e-01 5.27360737e-01 -3.47942978e-01 -8.49302411e-01
-7.97747076e-01 -1.20691037e+00 1.40941024e+00 1.45414412e-01
4.73100364e-01 -1.06594551e+00 -1.77732348e-01 6.28215373e-02
6.74793899e-01 4.30531710e-01 1.26323473e+00 -2.74608135e-01
-6.98540449e-01 -2.62981504e-01 -4.51105982e-01 1.77729517e-01
3.38191897e-01 -1.21866390e-01 -6.90306425e-01 -3.88001412e-01
1.72168717e-01 -2.71325737e-01 5.04690170e-01 3.04741263e-01
1.53050816e+00 4.86016907e-02 -4.07270610e-01 5.84935725e-01
1.59964967e+00 4.43878978e-01 3.67097616e-01 -7.22883940e-02
6.82495654e-01 4.98402953e-01 4.28431362e-01 3.32463354e-01
2.71688998e-01 5.94133019e-01 6.16644025e-01 -3.44066828e-01
-1.06042750e-01 -1.45522729e-01 -5.62057376e-01 1.35266542e+00
-4.06304270e-01 -3.46275754e-02 -9.18084264e-01 2.77866781e-01
-1.59627402e+00 -8.29074621e-01 -2.46604651e-01 1.80940378e+00
3.13667059e-01 -5.55329164e-03 -1.48495585e-01 1.93474546e-01
7.62383163e-01 1.66850582e-01 -5.59602678e-01 -3.01522911e-01
-1.19280726e-01 2.52601296e-01 -1.96266547e-02 1.84122145e-01
-1.01668143e+00 6.73187792e-01 5.60615301e+00 1.37624848e+00
-1.23286092e+00 -1.41796827e-01 3.54069263e-01 5.26790559e-01
-4.31845456e-01 -5.73888600e-01 -7.49687791e-01 1.95089117e-01
3.49768475e-02 -1.73872799e-01 2.36125886e-01 8.84036362e-01
2.38396917e-02 3.59316558e-01 -7.64046431e-01 1.59225023e+00
2.44498819e-01 -1.46865880e+00 7.02240229e-01 8.28060359e-02
5.86871028e-01 -2.40504473e-01 -1.95823740e-02 1.99382231e-01
-4.90648180e-01 -1.11988807e+00 3.81296545e-01 5.58838546e-01
1.13015079e+00 -8.48326385e-01 9.00709331e-01 5.80445826e-01
-1.69850922e+00 3.17604572e-01 -8.42210472e-01 2.85587776e-02
-3.91902681e-03 5.72889686e-01 -4.39912081e-01 9.62756574e-01
5.01498282e-01 9.20021415e-01 -1.92659169e-01 1.39448237e+00
3.95805180e-01 -4.52446789e-02 -3.47365409e-01 -2.82696009e-01
2.17081115e-01 -5.16595006e-01 5.65654218e-01 7.88201272e-01
5.74200451e-01 6.23951077e-01 5.22858083e-01 9.97184038e-01
-3.51382606e-02 6.45011485e-01 -1.24790585e+00 1.78623781e-01
4.90998328e-01 1.10131013e+00 -7.88248777e-01 -3.08853418e-01
-4.75270391e-01 3.35311115e-01 1.15029469e-01 2.77563948e-02
-4.14916039e-01 -3.16789925e-01 2.33263060e-01 2.19445527e-01
3.04630578e-01 -4.56102520e-01 -3.59853923e-01 -8.62137318e-01
9.23017636e-02 -5.59661627e-01 1.37144893e-01 -7.36920357e-01
-1.58146834e+00 7.18877733e-01 1.87529087e-01 -1.49089825e+00
3.86164188e-01 -7.54347503e-01 -5.77782631e-01 1.03458738e+00
-1.54413676e+00 -1.30958962e+00 -4.50199187e-01 5.94287694e-01
7.02585042e-01 -4.55073267e-01 1.00310731e+00 5.33867121e-01
2.43865903e-02 2.83395231e-01 2.00715631e-01 2.04676278e-02
1.65129080e-01 -7.81345427e-01 3.33047360e-01 -1.63352296e-01
-1.70069840e-02 3.91076446e-01 -1.16640478e-02 -7.22842693e-01
-1.75024617e+00 -1.03686047e+00 4.17606294e-01 1.08930534e-02
1.16648404e-02 -1.46125123e-01 -9.91665542e-01 1.21108428e-01
-8.60544071e-02 1.23876780e-01 3.22604626e-01 -2.69076884e-01
4.18089963e-02 -1.24994196e-01 -1.39444232e+00 1.69995457e-01
1.24297941e+00 -2.47454643e-01 -7.06529379e-01 3.51295710e-01
5.37719607e-01 -7.35473871e-01 -1.24149287e+00 9.18691337e-01
5.26706159e-01 -7.38121629e-01 1.15807307e+00 -4.17549700e-01
1.39404133e-01 -2.90185601e-01 -2.11619213e-01 -1.19811034e+00
-4.66377944e-01 3.09521705e-01 8.77818675e-04 9.56515670e-01
-7.07438886e-02 -5.71485341e-01 8.89248848e-01 8.63422602e-02
-3.79269361e-01 -1.28683519e+00 -1.12403488e+00 -7.18046606e-01
2.87200838e-01 -6.30150139e-02 9.26443696e-01 8.67539525e-01
-7.34978557e-01 1.94734126e-01 2.77105309e-02 -5.17271692e-03
5.45724690e-01 6.75820589e-01 6.25943422e-01 -1.83324254e+00
3.63862872e-01 -6.10708058e-01 -7.43460178e-01 -1.32912898e+00
2.17649028e-01 -1.13111472e+00 -5.44640243e-01 -1.89659417e+00
-1.85858663e-02 -8.55195403e-01 -1.57727338e-02 9.39845592e-02
2.74505854e-01 3.03648531e-01 1.94792524e-01 3.61111850e-01
-2.41239652e-01 9.81679976e-01 1.99648571e+00 -4.89705175e-01
-9.98662785e-02 4.97768261e-02 -1.60768166e-01 7.53132999e-01
6.58772945e-01 -1.97342839e-02 -4.47485447e-01 -7.15058029e-01
-1.05838403e-01 3.24368894e-01 2.22003788e-01 -7.60111451e-01
2.57473618e-01 1.17585575e-02 5.52826643e-01 -1.37381387e+00
6.35949075e-01 -1.30174470e+00 2.34792739e-01 3.00808966e-01
4.16791230e-01 3.11529458e-01 3.89959157e-01 4.74614829e-01
-3.92432898e-01 -2.44194403e-01 5.68443179e-01 -5.08115649e-01
-5.41340947e-01 8.58727336e-01 5.84062226e-02 -4.70381491e-02
1.02001297e+00 -4.84225273e-01 8.46482515e-02 3.25390585e-02
-8.05778384e-01 2.03635499e-01 3.94086540e-01 4.45036590e-01
1.36484456e+00 -2.00985575e+00 -7.19533920e-01 6.60095572e-01
-4.84246761e-02 3.65837783e-01 3.06852311e-01 2.82843441e-01
-7.59924412e-01 4.88677084e-01 -3.09126496e-01 -8.30624223e-01
-1.20431924e+00 3.86389524e-01 4.78125572e-01 1.71439156e-01
-7.53389955e-01 5.36095738e-01 6.64290562e-02 -7.18439221e-01
3.71902913e-01 -2.34133720e-01 -5.54116607e-01 -8.82239565e-02
1.53797686e-01 4.39141899e-01 2.76120663e-01 -8.31334114e-01
-2.79545069e-01 1.51561570e+00 2.65458643e-01 4.73411113e-01
1.39576519e+00 1.71950668e-01 -5.05257487e-01 4.19467062e-01
1.24694479e+00 -2.63149977e-01 -4.37635392e-01 -1.55744001e-01
-4.02883261e-01 -6.95364356e-01 1.28047671e-02 -2.24195749e-01
-1.23989165e+00 1.00976598e+00 6.92339957e-01 4.20834690e-01
7.76087880e-01 -6.35809153e-02 9.98402715e-01 4.17901725e-01
8.89755011e-01 -6.31769598e-01 -8.23748261e-02 7.73987830e-01
1.33512807e+00 -1.01407158e+00 8.59860405e-02 -7.25750625e-01
1.45902976e-01 1.35503745e+00 7.83592522e-01 -3.82002443e-01
1.26507246e+00 -1.08411126e-01 -1.79364607e-01 -6.84001625e-01
-2.50097275e-01 -2.74592843e-02 5.05637288e-01 4.66682076e-01
2.26411134e-01 -1.09839626e-01 -5.02781391e-01 2.15803787e-01
1.25682518e-01 -1.41765863e-01 -4.58295718e-02 8.69912207e-01
-6.01809025e-01 -1.09396267e+00 -5.95970929e-01 7.02353954e-01
-1.34294525e-01 2.57483840e-01 -2.53414869e-01 9.65051651e-01
2.33674526e-01 4.59813535e-01 2.24672511e-01 -3.45017940e-01
9.03046548e-01 -5.68251647e-02 5.29574513e-01 -5.89141369e-01
-2.10836992e-01 9.09224078e-02 -3.16539049e-01 -2.34948903e-01
-6.08040750e-01 -2.26626679e-01 -1.23424387e+00 -4.10333425e-01
-4.87610608e-01 2.46657655e-01 7.43846834e-01 5.86048424e-01
5.19580543e-01 2.64807254e-01 8.08906198e-01 -1.24498153e+00
-6.25648379e-01 -8.38602543e-01 -7.83836722e-01 4.67960864e-01
-1.09103210e-01 -1.26241660e+00 -1.69903785e-01 -4.58262295e-01] | [8.118633270263672, -3.79814076423645] |
00bf78c4-88fd-407d-9bdf-55d8d3556f9c | spliceradar-a-learned-method-for-blind-image | 1906.11663 | null | https://arxiv.org/abs/1906.11663v1 | https://arxiv.org/pdf/1906.11663v1.pdf | SpliceRadar: A Learned Method For Blind Image Forensics | Detection and localization of image manipulations like splices are gaining in importance with the easy accessibility of image editing softwares. While detection generates a verdict for an image it provides no insight into the manipulation. Localization helps explain a positive detection by identifying the pixels of the image which have been tampered. We propose a deep learning based method for splice localization without prior knowledge of a test image's camera-model. It comprises a novel approach for learning rich filters and for suppressing image-edges. Additionally, we train our model on a surrogate task of camera model identification, which allows us to leverage large and widely available, unmanipulated, camera-tagged image databases. During inference, we assume that the spliced and host regions come from different camera-models and we segment these regions using a Gaussian-mixture model. Experiments on three test databases demonstrate results on par with and above the state-of-the-art and a good generalization ability to unknown datasets. | ['Terrance E. Boult', 'Aurobrata Ghosh', 'Zheng Zhong', 'Maneesh Singh'] | 2019-06-27 | null | null | null | null | ['image-manipulation-detection', 'image-forensics'] | ['computer-vision', 'computer-vision'] | [ 4.93707210e-01 -1.21519931e-01 -1.46198481e-01 -1.83792233e-01
-1.21643484e+00 -8.13690484e-01 4.85780686e-01 -7.39187449e-02
-3.67619127e-01 3.05191517e-01 -2.01839566e-01 -3.50279808e-01
4.09185678e-01 -4.07341599e-01 -1.33628631e+00 -7.46325970e-01
2.43765160e-01 1.87630594e-01 2.47914404e-01 4.25563395e-01
3.84487987e-01 4.85420257e-01 -1.26502705e+00 6.25939488e-01
3.59473467e-01 9.29775596e-01 1.70349613e-01 9.05196309e-01
3.13745439e-01 9.44752455e-01 -8.15874934e-01 -7.11874306e-01
5.15574694e-01 -1.58697009e-01 -4.99282062e-01 5.91045737e-01
7.99481511e-01 -5.77873170e-01 -4.11797613e-01 1.23454022e+00
5.37421256e-02 -3.68250012e-01 7.25255549e-01 -1.25626898e+00
-5.05663037e-01 2.31680468e-01 -9.05671120e-01 3.33056480e-01
2.69222587e-01 4.02786255e-01 7.80641973e-01 -8.22156012e-01
7.40723193e-01 8.77414703e-01 8.02900016e-01 2.21285775e-01
-1.49065781e+00 -5.55735350e-01 -2.29319170e-01 2.52566963e-01
-1.50193155e+00 -4.76955742e-01 7.26724505e-01 -6.64161384e-01
6.40011609e-01 9.05131325e-02 3.80812585e-01 1.36776638e+00
-1.43039431e-02 9.49571252e-01 1.01333666e+00 -5.80738187e-01
3.37966643e-02 3.27976555e-01 -2.32740641e-02 9.30434942e-01
1.27076104e-01 1.37045324e-01 -5.44251680e-01 -8.09779763e-02
6.00920439e-01 -1.54043194e-02 -3.65606070e-01 -5.07792830e-01
-1.06244922e+00 4.83737230e-01 1.48901582e-01 -1.92760732e-02
-3.47925901e-01 2.75261253e-01 1.88907921e-01 1.73178181e-01
3.00294816e-01 2.49051347e-01 -4.76276338e-01 -6.90166503e-02
-1.32714188e+00 -1.03090219e-01 6.31434500e-01 9.32222009e-01
9.15064871e-01 -8.48538205e-02 8.06768239e-02 4.86691982e-01
3.05412933e-02 4.79789227e-01 2.13449672e-02 -1.01007164e+00
5.70323288e-01 4.85999525e-01 1.83485523e-01 -7.81188786e-01
9.27455574e-02 -4.42210734e-01 -5.93168736e-01 4.40003544e-01
5.93605459e-01 -1.20474966e-02 -9.74718451e-01 1.42533517e+00
1.13216519e-01 4.50487018e-01 -1.38472408e-01 7.56632507e-01
1.19396687e-01 5.65891683e-01 -2.74452329e-01 6.85064793e-02
1.42298019e+00 -8.33188415e-01 -2.49031812e-01 -3.39335501e-01
5.16854942e-01 -9.69280064e-01 8.97239149e-01 8.32514226e-01
-8.25730026e-01 -5.29052377e-01 -1.02013910e+00 -1.10496312e-01
-3.96095157e-01 6.34869218e-01 3.99383962e-01 9.00631487e-01
-8.69967639e-01 3.47852260e-01 -6.82188034e-01 -3.82867485e-01
7.52605021e-01 3.24151218e-01 -6.56852365e-01 -2.66562313e-01
-5.25906742e-01 8.45851183e-01 1.97607532e-01 1.99759845e-02
-1.34117067e+00 -5.76763988e-01 -6.71910107e-01 9.36440751e-02
6.57068133e-01 -4.83924955e-01 1.01298797e+00 -1.11646247e+00
-9.83456671e-01 1.38233984e+00 -3.21861237e-01 -6.49923444e-01
7.88191736e-01 -3.83785695e-01 -2.54506052e-01 4.38313812e-01
1.69765726e-01 6.25805557e-01 1.40213358e+00 -1.47157574e+00
-7.18510389e-01 -3.59151840e-01 -1.08351484e-01 -2.76378810e-01
-9.91608277e-02 1.72100276e-01 -8.29569638e-01 -5.63475966e-01
-2.18512505e-01 -8.95626128e-01 2.88850218e-02 2.39149690e-01
-9.02482271e-01 4.28823084e-01 1.15411568e+00 -9.03241098e-01
8.67221713e-01 -2.33515739e+00 -2.04275802e-01 2.79043466e-01
9.69321206e-02 3.66970003e-01 -2.11732224e-01 3.46138537e-01
-2.21857160e-01 1.66317001e-01 -1.59213662e-01 -4.69064564e-01
-1.73178673e-01 2.36526001e-02 -5.68484068e-01 9.14749205e-01
3.84295017e-01 5.87238550e-01 -5.04592359e-01 -3.87261420e-01
4.21311527e-01 3.60803843e-01 -4.05176848e-01 3.12386453e-01
-1.72983810e-01 2.03166753e-01 -2.15472952e-02 8.54129791e-01
8.42436671e-01 -2.02004723e-02 -1.59333590e-02 -3.62638414e-01
2.04576507e-01 -2.03530341e-01 -1.20519185e+00 1.46574795e+00
-2.82163888e-01 8.79081190e-01 2.50018001e-01 -8.51178110e-01
6.55765533e-01 3.60294692e-02 1.16498269e-01 -2.36921906e-01
5.22248894e-02 -3.34934071e-02 -2.78932750e-01 -7.57689595e-01
5.00514686e-01 1.61490843e-01 7.33136460e-02 5.04389226e-01
2.70202667e-01 -3.19046751e-02 8.19932073e-02 3.09839606e-01
1.22088289e+00 1.35997593e-01 9.25026536e-02 1.09421961e-01
3.17066491e-01 -2.47069791e-01 4.14065093e-01 1.03512979e+00
-1.01506457e-01 8.63232970e-01 8.44127774e-01 -2.73512363e-01
-1.14906132e+00 -1.19389176e+00 1.99889138e-01 5.99219561e-01
6.84469491e-02 -3.71281862e-01 -1.07122564e+00 -9.38818693e-01
-3.68427262e-02 7.92633533e-01 -6.80711031e-01 -1.30936489e-01
-2.71000415e-01 -3.21718603e-01 6.50960863e-01 4.38579232e-01
3.07119817e-01 -5.22761405e-01 -3.63110811e-01 -2.62791365e-01
-2.15048805e-01 -1.30479896e+00 -5.26452184e-01 3.63959610e-01
-5.74209273e-01 -1.54179120e+00 -3.48157763e-01 -6.85556412e-01
9.40672398e-01 3.36436659e-01 1.01681304e+00 1.22212628e-02
-6.86959922e-01 4.22651857e-01 -5.30208182e-03 -2.25582302e-01
-6.15840256e-01 -3.04810911e-01 -2.19022408e-01 3.06016684e-01
3.62187833e-01 -1.42147824e-01 -4.07641262e-01 2.92536408e-01
-9.51683760e-01 -2.42403686e-01 7.50019729e-01 7.80576885e-01
4.65199858e-01 3.94793570e-01 -2.74031192e-01 -1.02523875e+00
7.18094334e-02 -3.11836123e-01 -8.91187489e-01 3.89594555e-01
-3.17410767e-01 1.75142381e-02 3.51657957e-01 -3.77385408e-01
-9.26380634e-01 4.78959262e-01 1.42694741e-01 -8.78322959e-01
-6.08806014e-01 -3.41302082e-02 -2.96695799e-01 -1.83393717e-01
4.56378579e-01 2.73956895e-01 1.66213829e-02 -3.74975502e-01
3.86883825e-01 5.50799966e-01 8.92610729e-01 -3.77694905e-01
9.38088834e-01 7.16010451e-01 -9.02121291e-02 -8.21792126e-01
-6.17498159e-01 -7.37635553e-01 -6.21018112e-01 -2.33174816e-01
8.63616109e-01 -9.39913929e-01 -6.70012474e-01 7.58886576e-01
-1.33791506e+00 -2.15025723e-01 2.87024844e-02 2.59811729e-01
-4.86399174e-01 6.20843887e-01 -6.38795435e-01 -9.05106008e-01
1.02202371e-01 -1.31712770e+00 1.47059572e+00 1.25593627e-02
-5.20323776e-02 -7.30794251e-01 -4.59151387e-01 7.45536089e-01
-1.14792287e-02 9.97733101e-02 8.03816557e-01 -6.04451478e-01
-1.03535128e+00 -6.52010202e-01 -2.05471545e-01 7.17560828e-01
-2.42498234e-01 3.69709730e-01 -1.30718637e+00 -5.06874770e-02
9.49384123e-02 -1.51002675e-01 1.10012519e+00 2.89350033e-01
1.32382369e+00 -2.59986341e-01 -3.66283387e-01 6.20615244e-01
1.61490500e+00 -1.22009411e-01 8.80699873e-01 3.09344947e-01
5.93377769e-01 4.71713811e-01 4.53707814e-01 3.22798103e-01
-4.44355514e-03 5.93096077e-01 7.91069210e-01 -4.10750687e-01
5.41516244e-02 -3.59883875e-01 5.93354881e-01 -1.12631693e-01
2.70428777e-01 -5.30740201e-01 -7.78797925e-01 5.96917510e-01
-1.60162389e+00 -1.14540672e+00 -4.01856571e-01 2.34311485e+00
3.79525334e-01 2.81357169e-01 6.39259489e-03 3.24554779e-02
1.04166961e+00 5.36681749e-02 -3.93101960e-01 -8.75820369e-02
-2.67342597e-01 8.17625374e-02 9.27095413e-01 4.36328411e-01
-1.42177558e+00 9.51497853e-01 6.55921078e+00 8.83660018e-01
-1.00442564e+00 6.90646321e-02 9.00739312e-01 -4.21762615e-02
1.70356363e-01 1.49698481e-01 -7.43562460e-01 5.62459290e-01
6.67118251e-01 3.40562999e-01 3.84740591e-01 8.92497063e-01
3.27752143e-01 -4.60813433e-01 -1.23948514e+00 1.09181213e+00
5.11448741e-01 -1.32756925e+00 -1.38863489e-01 2.78308451e-01
5.52406251e-01 2.36755814e-02 2.50777692e-01 -2.85284687e-02
2.29986608e-01 -7.23447680e-01 1.12502062e+00 4.80315834e-01
6.46496892e-01 -5.62116742e-01 7.99654007e-01 3.39214891e-01
-6.12329662e-01 -1.52668163e-01 -2.68839717e-01 1.70733020e-01
8.82786512e-03 7.06411064e-01 -1.08889806e+00 2.30670944e-01
6.16341829e-01 6.68459952e-01 -8.33287299e-01 9.77218688e-01
-7.08494604e-01 8.28818917e-01 -2.12046087e-01 5.90496302e-01
2.17258371e-02 -2.85676509e-01 5.31422675e-01 1.26365221e+00
3.02225858e-01 -4.77770686e-01 1.19802393e-01 1.12874174e+00
-2.97826707e-01 -3.82562220e-01 -8.49067092e-01 -1.68483481e-01
3.44901055e-01 1.32239318e+00 -8.57432127e-01 -1.83738083e-01
-5.28091371e-01 1.43399763e+00 1.14583448e-01 4.48815852e-01
-1.03608477e+00 -1.07534662e-01 5.47575414e-01 1.54436588e-01
8.15575361e-01 -1.16227232e-01 -1.54206380e-01 -1.22301495e+00
2.41713107e-01 -9.87111390e-01 2.91575938e-01 -1.10386574e+00
-1.02830863e+00 1.20540485e-01 -3.92561138e-01 -1.30343211e+00
-1.02902949e-01 -8.45789492e-01 -6.76515639e-01 5.50169647e-01
-1.30237806e+00 -1.56539154e+00 -3.03662904e-02 4.60318416e-01
4.00440693e-01 -1.39703959e-01 5.93593240e-01 1.89890549e-01
-4.96095717e-01 5.68547249e-01 1.99494198e-01 5.67966461e-01
8.99090886e-01 -1.23841751e+00 3.20963025e-01 1.48258579e+00
5.37675917e-01 6.34317160e-01 6.67424023e-01 -3.14263135e-01
-1.58878672e+00 -1.06728709e+00 5.69372118e-01 -6.83098257e-01
7.13436007e-01 -7.49991298e-01 -6.77480221e-01 8.03748012e-01
3.68834198e-01 9.90242809e-02 4.84571904e-01 3.03854216e-02
-6.41481340e-01 -2.59531319e-01 -1.02281129e+00 3.36980730e-01
6.28286004e-01 -8.92860115e-01 -3.31293792e-01 5.25700867e-01
3.81338820e-02 -2.16953769e-01 -3.50372165e-01 -4.81749028e-02
3.94475311e-01 -1.30670381e+00 9.54554915e-01 -4.10618931e-01
4.71660823e-01 -5.17734945e-01 -1.99999347e-01 -1.01896822e+00
-7.00907595e-03 -6.26648903e-01 1.10068880e-01 1.55087590e+00
4.52344209e-01 -2.21258029e-01 1.01976848e+00 6.04182541e-01
2.56806277e-02 -2.97989666e-01 -7.37478375e-01 -7.21712649e-01
-3.45816404e-01 -6.68921113e-01 7.11766109e-02 7.77449548e-01
-4.02716845e-01 1.32477239e-01 -5.68676770e-01 6.41621590e-01
8.04609120e-01 -8.60566199e-02 1.17188811e+00 -7.59760618e-01
-5.95396996e-01 -3.02897513e-01 -8.49476635e-01 -8.72721434e-01
4.26695734e-01 -5.77467144e-01 1.27586573e-01 -9.78709519e-01
5.68446517e-01 1.62644744e-01 -3.31257395e-02 4.39582407e-01
-8.23970139e-02 3.88110638e-01 7.34534413e-02 1.34721175e-01
-7.47633100e-01 -1.55387059e-01 4.27799702e-01 -3.52618098e-01
3.64544123e-01 -6.49079978e-02 -4.70898271e-01 7.93381512e-01
5.42770088e-01 -6.96195185e-01 -1.22864321e-01 -5.50150156e-01
5.88420816e-02 3.57243940e-02 8.74641776e-01 -1.11492884e+00
2.23471180e-01 3.36184353e-01 5.79328001e-01 -5.45006752e-01
2.77751535e-01 -1.13419282e+00 1.26530334e-01 3.89819324e-01
-4.23328102e-01 -1.88230723e-01 7.42345676e-02 9.74451244e-01
-2.09529355e-01 -3.46021473e-01 9.77457345e-01 -1.50370792e-01
-9.08529222e-01 4.33213152e-02 -5.71112275e-01 -2.45009333e-01
1.22839820e+00 -3.43822449e-01 -4.05779183e-01 -4.17817026e-01
-6.08161628e-01 -2.14377925e-01 9.10693109e-01 2.91193485e-01
5.26300848e-01 -6.23385072e-01 -4.14555490e-01 5.42948008e-01
2.17020810e-01 -3.59186739e-01 2.20168754e-01 8.57888222e-01
-5.86417437e-01 -4.09881165e-03 5.24312817e-02 -8.40716720e-01
-1.64646101e+00 9.00538623e-01 2.68239260e-01 -7.36120418e-02
-3.18687409e-01 8.78590584e-01 2.59099275e-01 -5.06345145e-02
2.03687057e-01 -4.34883595e-01 4.12601024e-01 -9.00539085e-02
5.07712960e-01 2.27933168e-01 1.39563635e-01 -6.47859871e-01
-3.46535861e-01 3.52244914e-01 -2.57873297e-01 -1.39305722e-02
1.12541473e+00 -4.92862612e-02 -8.95173773e-02 2.61111379e-01
1.32492948e+00 2.30997562e-01 -1.37394738e+00 -3.75467651e-02
-2.93045565e-02 -7.89622903e-01 1.42233849e-01 -8.08976889e-01
-1.13641226e+00 1.04658151e+00 6.49500072e-01 1.64384078e-02
1.00079930e+00 6.93356618e-02 4.29972261e-01 2.71247715e-01
3.51809382e-01 -9.87889707e-01 6.32164031e-02 8.84535089e-02
4.07969385e-01 -1.42339826e+00 -4.71766703e-02 -5.32327652e-01
-5.89000344e-01 1.23051405e+00 3.84809881e-01 -1.24324135e-01
3.69773507e-01 6.07748568e-01 1.64608791e-01 1.39995450e-02
-6.16551876e-01 -1.15583643e-01 -1.54441297e-01 7.81339824e-01
1.32257700e-01 -2.26533473e-01 4.80954379e-01 2.73819208e-01
1.97833344e-01 -3.27976532e-02 6.96171939e-01 6.87301874e-01
-3.07004094e-01 -8.54474485e-01 -6.43318832e-01 3.87961298e-01
-6.94511056e-01 -2.20775858e-01 -7.40461290e-01 8.03336442e-01
2.27068782e-01 9.46233511e-01 4.31486405e-03 -3.29595953e-01
4.34943959e-02 2.41649244e-02 3.93122077e-01 -5.09799123e-01
-4.38330948e-01 1.60130665e-01 6.88510686e-02 -7.22081721e-01
-3.67419600e-01 -8.75470221e-01 -6.49459779e-01 -8.64198357e-02
-6.90918684e-01 1.16028283e-02 7.89947033e-01 8.49727392e-01
4.47495043e-01 1.54543847e-01 5.19972444e-01 -7.16950893e-01
-4.93919581e-01 -7.13141501e-01 -8.96413803e-01 6.18501663e-01
4.13859606e-01 -3.21418732e-01 -4.92333978e-01 7.65970409e-01] | [12.32193374633789, 1.0213125944137573] |
549ad968-e600-4dea-8183-16430605e3d9 | pairwise-sequence-alignment-at-arbitrarily | 2207.12543 | null | https://arxiv.org/abs/2207.12543v1 | https://arxiv.org/pdf/2207.12543v1.pdf | Pairwise sequence alignment at arbitrarily large evolutionary distance | Ancestral sequence reconstruction is a key task in computational biology. It consists in inferring a molecular sequence at an ancestral species of a known phylogeny, given descendant sequences at the tip of the tree. In addition to its many biological applications, it has played a key role in elucidating the statistical performance of phylogeny estimation methods. Here we establish a formal connection to another important bioinformatics problem, multiple sequence alignment, where one attempts to best align a collection of molecular sequences under some mismatch penalty score by inserting gaps. Our result is counter-intuitive: we show that perfect pairwise sequence alignment with high probability is possible in principle at arbitrary large evolutionary distances - provided the phylogeny is known and dense enough. We use techniques from ancestral sequence reconstruction in the taxon-rich setting together with the probabilistic analysis of sequence evolution models involving insertions and deletions. | ['Sebastien Roch', 'Brandon Legried'] | 2022-07-25 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 8.02523732e-01 -1.92205846e-01 -1.81122750e-01 -2.23493174e-01
-6.80871010e-01 -9.76090193e-01 2.18493044e-01 5.89373887e-01
-6.74229205e-01 1.09915352e+00 -4.61732633e-02 -6.71916485e-01
-1.30504638e-01 -4.65038478e-01 -8.33895802e-01 -1.13269281e+00
-3.03948671e-01 7.49709129e-01 3.65471214e-01 -1.45688191e-01
4.90613133e-01 5.48437059e-01 -1.23830986e+00 -4.50801570e-04
8.35742414e-01 1.77175224e-01 5.43945670e-01 1.11131942e+00
-1.87199801e-01 -8.96892548e-02 -3.43550563e-01 -7.30901122e-01
1.66286647e-01 -7.85043955e-01 -1.06181228e+00 -1.81913421e-01
2.91135848e-01 1.42491266e-01 1.47454977e-01 1.11268675e+00
2.56408483e-01 -1.17783852e-01 5.07413983e-01 -9.46992576e-01
-7.04881251e-02 3.74396563e-01 -6.88649178e-01 2.61611670e-01
2.88394749e-01 8.06936994e-02 1.38694465e+00 -3.31058145e-01
9.31033969e-01 1.04263294e+00 1.00345075e+00 2.69144565e-01
-1.68904376e+00 -8.11251774e-02 -3.29803228e-01 2.84788221e-01
-1.25207543e+00 9.14264172e-02 2.22723305e-01 -5.77918172e-01
8.87126744e-01 4.83405590e-01 6.20428383e-01 9.31464851e-01
4.02904451e-01 3.15659463e-01 7.45760322e-01 -4.04985934e-01
2.18900308e-01 -5.16441584e-01 4.17357013e-02 5.33841908e-01
2.44976789e-01 -2.45716378e-01 -3.53200972e-01 -9.84116912e-01
5.58973432e-01 -5.93378432e-02 -3.65189493e-01 -3.82235974e-01
-1.30587745e+00 9.18954790e-01 5.75273857e-02 2.53616065e-01
-4.16705191e-01 4.14866582e-02 3.93155634e-01 4.71966080e-02
7.67883509e-02 4.44644094e-01 -7.14638829e-01 -3.58412236e-01
-8.26913774e-01 1.49196878e-01 9.28794682e-01 6.75806463e-01
8.47923279e-01 -4.30767477e-01 6.87853634e-01 6.83186233e-01
-1.46254078e-01 1.34955809e-01 2.68177301e-01 -9.99658883e-01
8.92419815e-02 1.98424414e-01 2.17655182e-01 -6.43105209e-01
-1.17703229e-02 -2.18555734e-01 -7.79772758e-01 -5.62810013e-03
7.65216231e-01 2.21108511e-01 -4.62307990e-01 1.99010825e+00
7.25653887e-01 5.36987543e-01 -1.11354917e-01 4.27153438e-01
-1.52518854e-01 7.62062252e-01 1.12699017e-01 -6.77103341e-01
1.25572145e+00 -2.63431966e-01 -1.68706119e-01 -9.13100764e-02
2.74822980e-01 -8.18203270e-01 6.76815152e-01 4.32955861e-01
-6.93756282e-01 -7.13402852e-02 -8.26959193e-01 5.34437178e-03
2.60436416e-01 -4.82995719e-01 5.17516673e-01 5.54285824e-01
-7.95936108e-01 1.05869639e+00 -8.07213545e-01 -6.11937106e-01
5.43044023e-02 2.25809753e-01 -6.51948869e-01 -5.06686652e-03
-6.46901369e-01 9.50430512e-01 5.91462374e-01 -6.76441491e-02
-5.47538221e-01 -5.77429295e-01 -3.78089964e-01 -7.27851018e-02
3.20878923e-01 -9.01271999e-01 1.15627980e+00 -9.87329364e-01
-1.08020580e+00 9.91182566e-01 -6.62126780e-01 -5.59551835e-01
4.05871093e-01 -4.41188961e-02 1.94648087e-01 1.39760226e-01
-1.73265383e-01 2.37105295e-01 3.23053211e-01 -6.80758059e-01
-5.36793113e-01 -6.71186924e-01 -5.04617810e-01 -9.02688578e-02
4.08951491e-01 6.00646548e-02 -1.02640629e-01 -4.06985402e-01
4.08344507e-01 -1.11817765e+00 -8.57414782e-01 1.51243106e-01
-1.07352547e-01 -1.38141021e-01 5.29435687e-02 -1.01596344e+00
7.64981925e-01 -1.96142483e+00 7.51663685e-01 1.25961348e-01
9.99971479e-02 1.58425957e-01 -5.54581210e-02 9.77644265e-01
-2.87588775e-01 1.80637613e-01 -9.71257627e-01 1.60178006e-01
-4.03003782e-01 5.62469423e-01 -4.59276676e-01 8.77494097e-01
1.20628797e-01 6.25491142e-01 -9.60153520e-01 -4.02081937e-01
-7.32791573e-02 3.45800757e-01 -6.51850283e-01 2.53370643e-01
-3.51727277e-01 5.27301967e-01 -2.63150156e-01 2.64911085e-01
7.58098423e-01 -2.95821786e-01 9.09590185e-01 3.10282230e-01
-2.74119228e-01 4.63557541e-01 -8.71969104e-01 1.80100822e+00
-9.35487598e-02 4.41382438e-01 7.12224245e-02 -1.10943210e+00
6.35361850e-01 1.79229394e-01 5.57347000e-01 2.08213151e-01
-8.61986056e-02 4.23756659e-01 5.00091791e-01 -3.03793937e-01
4.47482407e-01 -6.53815866e-01 1.99476212e-01 1.17896557e-01
-4.29993793e-02 5.49075194e-02 1.81290228e-02 1.06520794e-01
1.10651064e+00 2.94689775e-01 1.04970706e+00 -2.94682026e-01
3.13408256e-01 1.76094636e-01 9.88957822e-01 6.23486280e-01
-1.49187267e-01 6.21603787e-01 6.30488515e-01 -3.95009160e-01
-1.74751389e+00 -1.19944572e+00 -4.11580712e-01 8.73459041e-01
1.97311640e-02 -4.05367315e-01 -6.43968344e-01 -1.62467405e-01
-5.66459121e-03 4.10633922e-01 -4.90126640e-01 8.62979144e-03
-6.50887370e-01 -1.09982860e+00 6.70208156e-01 1.17812388e-01
-2.38365829e-01 -6.42643034e-01 -6.24970019e-01 5.75639486e-01
-4.46850717e-01 -7.82521963e-01 -2.64286846e-01 4.96047080e-01
-1.20860696e+00 -1.26310253e+00 -7.74585366e-01 -4.57070142e-01
3.58336955e-01 1.90899506e-01 1.01373518e+00 2.71025538e-01
-6.37338877e-01 -2.27460280e-01 -7.95906559e-02 -2.04096492e-02
-7.43882358e-01 3.23322229e-02 2.02062368e-01 -3.23015213e-01
2.77817369e-01 -1.08761322e+00 -4.78067636e-01 3.55457991e-01
-9.11543846e-01 -2.97233071e-02 3.53572011e-01 1.10738862e+00
6.58115625e-01 -2.56307781e-01 9.32504237e-02 -7.44282603e-01
1.33684248e-01 -7.10630119e-01 -8.29849422e-01 6.92901373e-01
7.92870969e-02 2.30101839e-01 6.89313591e-01 -3.03735256e-01
-6.31331265e-01 2.39566386e-01 -8.06111753e-01 1.03490040e-01
-1.62433565e-01 4.20977652e-01 -3.21194947e-01 6.10154271e-02
6.58042550e-01 6.33568287e-01 1.72684684e-01 -7.13916361e-01
2.25995928e-01 5.48526883e-01 8.00723016e-01 -8.49321961e-01
2.81696916e-01 5.01268148e-01 4.68737900e-01 -1.25698519e+00
-3.80961150e-01 -7.61919558e-01 -1.08935213e+00 3.66501689e-01
5.24996936e-01 -4.54245597e-01 -1.25011921e+00 2.82309592e-01
-1.19272852e+00 -7.20879659e-02 1.11881442e-01 4.20732230e-01
-8.13945591e-01 1.16552889e+00 -4.71410602e-01 -8.06545615e-01
-4.01780047e-02 -1.03019047e+00 8.88482511e-01 -8.91540796e-02
-3.28273803e-01 -8.35853279e-01 6.84977472e-01 6.18384890e-02
-1.63384393e-01 4.88831639e-01 1.23928308e+00 -6.82900846e-01
-3.76741111e-01 8.34162310e-02 3.04862231e-01 -3.85938734e-02
1.61662921e-01 6.01716399e-01 -4.93333220e-01 -2.15075418e-01
1.83481187e-01 1.30069599e-01 7.47186720e-01 3.42068225e-01
9.23045397e-01 -2.52754688e-01 -3.68416458e-01 6.52323723e-01
1.66732454e+00 7.91025981e-02 4.60778117e-01 2.79715478e-01
2.66900182e-01 8.11257899e-01 5.02507389e-01 4.63659525e-01
-2.57785738e-01 7.98621416e-01 3.65214378e-01 4.96345848e-01
4.45313931e-01 -3.67137253e-01 -1.19801089e-02 5.38653374e-01
3.05792764e-02 -1.97992817e-01 -1.15077853e+00 4.83156443e-01
-1.85290456e+00 -1.22529113e+00 -4.08591539e-01 2.75305867e+00
1.02644980e+00 -2.17972219e-01 4.02711213e-01 1.86667033e-02
9.12357688e-01 -2.32256979e-01 -8.08150709e-01 -4.25428152e-01
-4.23600495e-01 2.11007163e-01 7.64517426e-01 6.67470217e-01
-4.58434135e-01 5.65044940e-01 7.69875860e+00 6.66303217e-01
-8.54614973e-01 -2.83723772e-01 4.62001622e-01 1.91309229e-01
-5.30243218e-01 5.05767524e-01 -7.53449857e-01 6.16964579e-01
1.20095682e+00 -3.68551672e-01 3.54088396e-01 9.32708383e-01
-4.10204791e-02 -2.96153277e-01 -1.08838761e+00 7.15152204e-01
-3.85530710e-01 -1.53434503e+00 -1.04749121e-01 3.73952895e-01
2.98270971e-01 3.96700412e-01 -2.55690306e-01 -6.19560897e-01
6.04546309e-01 -8.41537833e-01 3.14746141e-01 2.93092057e-02
5.83303154e-01 -8.66064370e-01 3.94169599e-01 8.33975554e-01
-7.46913731e-01 2.78903395e-01 -8.29485059e-01 -8.87122154e-02
5.54300070e-01 5.92802703e-01 -1.18717515e+00 6.28408968e-01
2.62640476e-01 3.81319463e-01 -2.02640821e-03 1.36114645e+00
-2.10082710e-01 8.66111398e-01 -6.47835672e-01 1.22415937e-01
1.35796726e-01 -7.09918678e-01 7.30810881e-01 1.16617489e+00
3.65026653e-01 2.06067488e-01 7.85878375e-02 3.30181658e-01
2.05412894e-01 1.27205759e-01 -5.29249787e-01 -1.47762825e-04
5.18627882e-01 9.08931077e-01 -7.01000512e-01 -2.22756267e-01
-4.96348888e-02 1.19846714e+00 4.52922076e-01 -7.47572184e-02
-6.75676048e-01 -1.58024147e-01 1.20649612e+00 -8.13262314e-02
4.91492122e-01 -5.49715638e-01 1.24010500e-02 -1.10871506e+00
-1.01167828e-01 -8.94691467e-01 5.69635987e-01 -4.14899021e-01
-1.31462896e+00 3.85201037e-01 -2.18075931e-01 -8.05287421e-01
-5.96504509e-01 -5.66224754e-01 -4.18111384e-01 1.12564611e+00
-1.18774915e+00 -5.37368119e-01 4.33096498e-01 -2.70189159e-02
5.33767343e-01 3.46367210e-01 8.28073263e-01 -4.11968157e-02
-5.64080536e-01 2.58707523e-01 8.19050848e-01 -3.42780650e-01
4.07049119e-01 -1.24988151e+00 1.00717223e+00 1.00555396e+00
2.15918034e-01 1.06527090e+00 1.23112226e+00 -7.61988401e-01
-1.48324609e+00 -5.41444063e-01 9.23182726e-01 -2.98241585e-01
8.74914110e-01 -3.74081880e-01 -1.37004745e+00 8.71693492e-01
-9.79276896e-02 -3.57844651e-01 1.15652990e+00 4.26930673e-02
-5.58260918e-01 4.09833908e-01 -1.07445383e+00 5.05562365e-01
1.05107737e+00 -4.78500277e-01 -7.29646802e-01 2.80944258e-01
6.44930482e-01 -4.70950864e-02 -9.36425567e-01 1.05437860e-01
8.67452323e-01 -8.86723042e-01 1.17005563e+00 -1.23797870e+00
2.76067972e-01 -5.33863425e-01 -4.68842566e-01 -9.88124311e-01
-3.51249933e-01 -9.42169368e-01 5.26898563e-01 9.84857321e-01
1.39798313e-01 -5.46299636e-01 8.91862273e-01 1.50072977e-01
2.34629419e-02 -4.05976564e-01 -1.39329731e+00 -1.09521580e+00
2.58469433e-01 -1.31165618e-02 4.82648164e-01 9.68130648e-01
-1.91040728e-02 2.50819564e-01 -5.13264894e-01 3.66474688e-01
8.36219072e-01 2.22955465e-01 8.51973355e-01 -1.34408808e+00
-6.42700911e-01 -2.51023084e-01 -5.54776907e-01 -1.13590455e+00
1.47196844e-01 -7.64787436e-01 3.04836839e-01 -1.00726187e+00
4.42959756e-01 -3.76648813e-01 -1.02816429e-02 5.89441657e-02
-5.11921763e-01 -3.91850024e-02 -1.99126914e-01 4.02700007e-01
-3.21138203e-02 2.88011521e-01 4.70767319e-01 4.74173605e-01
3.40922713e-01 1.86659366e-01 -3.47679824e-01 7.46810257e-01
7.20471501e-01 -8.74868631e-01 1.16137691e-01 -1.19847365e-01
4.33508813e-01 2.94309556e-01 8.26798901e-02 -3.83853018e-01
4.65883268e-03 -5.55364907e-01 -1.29220158e-01 -5.42570412e-01
2.24333882e-01 -8.11799586e-01 8.24340999e-01 1.04902124e+00
-2.15175629e-01 4.08672988e-01 -7.03982487e-02 7.86818087e-01
1.71225622e-01 -7.35943615e-01 9.98450100e-01 -2.52275229e-01
-7.20307350e-01 1.78179249e-01 -3.53202641e-01 -1.65507823e-01
8.79897892e-01 -3.28671634e-01 -1.53364152e-01 8.30783248e-02
-4.39711422e-01 -1.89409718e-01 1.09740138e+00 -8.53861719e-02
2.93958098e-01 -6.56386733e-01 -9.45218146e-01 -5.79224229e-02
2.10889891e-01 -4.18933868e-01 2.19464555e-01 7.43313968e-01
-9.62888300e-01 1.00616105e-01 -3.96800995e-01 -9.89067078e-01
-1.89868057e+00 8.33100200e-01 1.19388103e-01 -4.14771177e-02
-5.98027587e-01 9.10681725e-01 2.05541730e-01 -4.53625202e-01
-1.63246110e-01 2.50001401e-01 4.90377247e-01 -3.44221026e-01
6.51366174e-01 2.08510607e-01 -1.43630803e-01 -7.53574073e-01
-4.39408839e-01 5.73626101e-01 -1.22694433e-01 -1.93191394e-01
1.59930611e+00 -1.34631112e-01 -5.79101443e-01 5.96825957e-01
1.28948855e+00 1.40934527e-01 -1.04253364e+00 -2.68424988e-01
5.42765379e-01 -8.06121290e-01 -8.36452365e-01 -3.21645826e-01
-1.16140097e-01 9.23520327e-01 7.36092217e-03 -1.44801095e-01
7.24604130e-01 6.95786476e-02 6.35598361e-01 5.71535945e-01
4.70648646e-01 -3.15483958e-01 -4.74102616e-01 2.49925822e-01
4.41246212e-01 -9.93879020e-01 1.03787489e-01 -5.12800097e-01
-2.62437642e-01 1.08628440e+00 -1.36966735e-01 2.99770460e-02
-1.45782441e-01 1.42008245e-01 -3.78210068e-01 1.62464723e-01
-8.25786471e-01 7.38945454e-02 -8.15665796e-02 4.56849992e-01
4.96650964e-01 -2.73827165e-02 -7.04280078e-01 -1.91214070e-01
-3.34798485e-01 -3.04984987e-01 5.32527804e-01 9.10182476e-01
-8.43890786e-01 -1.70105910e+00 -2.95798361e-01 -4.11908478e-02
-7.41113603e-01 -4.00071025e-01 -6.37255907e-01 6.36385262e-01
-2.44289055e-01 4.47306037e-01 -1.05231889e-02 2.17203945e-02
-2.29434341e-01 2.92198449e-01 7.65542269e-01 -4.99739915e-01
-5.37016869e-01 4.13456224e-02 2.90526054e-03 -1.69525757e-01
-1.21440642e-01 -1.07585597e+00 -9.10093546e-01 -6.42659903e-01
-3.81944478e-01 7.35860944e-01 9.13103878e-01 7.62078166e-01
2.25899205e-01 -1.02723472e-01 6.44178331e-01 -3.01211745e-01
-8.43275487e-01 -5.90641558e-01 -5.09179175e-01 2.44169772e-01
4.07555938e-01 -8.38284343e-02 -3.27105582e-01 1.67642459e-01] | [4.843857288360596, 5.15723180770874] |
209a5163-d009-4974-9b85-8d6b2957fd3c | multimodality-and-dialogue-act-classification | null | null | https://aclanthology.org/W13-4031 | https://aclanthology.org/W13-4031.pdf | Multimodality and Dialogue Act Classification in the RoboHelper Project | null | ['Lin Chen', 'Barbara Di Eugenio'] | 2013-08-01 | null | null | null | ws-2013-8 | ['dialogue-act-classification'] | ['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] | [-7.292131423950195, 3.794543504714966] |
8e24dead-f0f8-4836-b733-c593ef7e67b3 | referring-segmentation-in-images-and-videos | 2102.04762 | null | https://arxiv.org/abs/2102.04762v1 | https://arxiv.org/pdf/2102.04762v1.pdf | Referring Segmentation in Images and Videos with Cross-Modal Self-Attention Network | We consider the problem of referring segmentation in images and videos with natural language. Given an input image (or video) and a referring expression, the goal is to segment the entity referred by the expression in the image or video. In this paper, we propose a cross-modal self-attention (CMSA) module to utilize fine details of individual words and the input image or video, which effectively captures the long-range dependencies between linguistic and visual features. Our model can adaptively focus on informative words in the referring expression and important regions in the visual input. We further propose a gated multi-level fusion (GMLF) module to selectively integrate self-attentive cross-modal features corresponding to different levels of visual features. This module controls the feature fusion of information flow of features at different levels with high-level and low-level semantic information related to different attentive words. Besides, we introduce cross-frame self-attention (CFSA) module to effectively integrate temporal information in consecutive frames which extends our method in the case of referring segmentation in videos. Experiments on benchmark datasets of four referring image datasets and two actor and action video segmentation datasets consistently demonstrate that our proposed approach outperforms existing state-of-the-art methods. | ['Yang Wang', 'Xiaoqin Zhang', 'Zhi Liu', 'Mrigank Rochan', 'Linwei Ye'] | 2021-02-09 | null | null | null | null | ['referring-expression-segmentation'] | ['computer-vision'] | [ 3.28150272e-01 -1.07101195e-01 -4.06038791e-01 -4.73946214e-01
-8.36712956e-01 -2.42700607e-01 3.73014957e-01 -9.23779905e-02
-4.89756167e-01 4.74790305e-01 5.45574427e-01 3.15196842e-01
6.85053319e-02 -5.58095634e-01 -8.34076822e-01 -5.88100374e-01
1.93071678e-01 -2.28486121e-01 5.81340551e-01 -1.62793949e-01
3.34288508e-01 2.98563659e-01 -1.56699598e+00 7.55314648e-01
6.85049057e-01 1.10008335e+00 4.70620930e-01 6.36690855e-01
-4.72731233e-01 1.27803552e+00 -4.99748707e-01 -2.61581689e-01
-1.45809740e-01 -9.27178800e-01 -1.28142786e+00 6.50577247e-01
5.27461290e-01 -1.52358785e-01 -3.92699093e-01 1.19894457e+00
3.20277244e-01 6.73144519e-01 3.56449455e-01 -1.25841296e+00
-9.47928369e-01 5.36440849e-01 -8.01696062e-01 9.17989850e-01
5.96894741e-01 2.68957287e-01 7.89167523e-01 -7.23968029e-01
8.31341803e-01 1.69400108e+00 1.24889672e-01 5.24086952e-01
-8.09817910e-01 -3.45338404e-01 7.89189219e-01 5.43496072e-01
-1.21372175e+00 -3.79830301e-01 1.08839881e+00 -4.17946577e-01
7.68432379e-01 3.65208611e-02 7.43012726e-01 8.79628062e-01
2.45406345e-01 1.31332958e+00 8.79826367e-01 -2.37188548e-01
6.66353479e-02 -2.20937565e-01 2.02125847e-01 7.82972574e-01
-5.76899588e-01 -3.04694921e-01 -7.07685113e-01 2.02917650e-01
9.07089889e-01 2.55699068e-01 -4.04822290e-01 -1.72059219e-02
-1.26227522e+00 7.31527448e-01 6.09921455e-01 7.37709939e-01
-8.26656818e-01 3.20783287e-01 4.75799829e-01 3.35451378e-03
4.27113771e-01 -6.83237016e-02 -4.09123808e-01 -4.64772508e-02
-7.53715217e-01 -2.77258694e-01 2.26791248e-01 1.12792850e+00
8.18508148e-01 8.28371570e-02 -7.25818932e-01 7.01910734e-01
1.83901653e-01 2.05347106e-01 5.59304893e-01 -1.34645712e+00
3.52359414e-01 7.43471742e-01 -1.70684427e-01 -1.06454432e+00
-2.37641528e-01 -4.73982431e-02 -6.33742750e-01 -3.78218621e-01
6.56773001e-02 1.34029314e-01 -9.16749418e-01 1.84707212e+00
5.03373504e-01 2.77937293e-01 1.29707858e-01 1.10831273e+00
1.29423440e+00 8.44592750e-01 5.91149807e-01 -5.74059665e-01
1.62949657e+00 -1.27203298e+00 -1.09056115e+00 -4.22896564e-01
2.92145848e-01 -8.03809464e-01 1.12749100e+00 -8.43523145e-02
-1.46795082e+00 -1.06536412e+00 -5.12118816e-01 -4.94473994e-01
-4.24419612e-01 -5.02166599e-02 4.91034210e-01 -1.70438722e-01
-8.89239132e-01 2.55534679e-01 -5.06934941e-01 -5.68090260e-01
6.09120667e-01 2.65644550e-01 -4.91130680e-01 -1.20529912e-01
-1.27658117e+00 5.09907544e-01 5.02191067e-01 6.81855157e-02
-6.61253810e-01 -4.39827800e-01 -1.05741751e+00 -5.25063574e-02
6.62550688e-01 -7.82504380e-01 9.70819473e-01 -1.83379090e+00
-1.18564034e+00 1.04684114e+00 -5.86346805e-01 -2.73288995e-01
1.02582201e-01 -3.22852015e-01 -3.70198280e-01 7.57027030e-01
4.15161341e-01 1.22068942e+00 9.14878607e-01 -1.27423990e+00
-8.43716681e-01 -4.03627366e-01 5.01343191e-01 4.39890742e-01
5.63678406e-02 3.91220748e-01 -1.11082172e+00 -8.11196089e-01
9.38346833e-02 -4.07791704e-01 5.98737225e-02 -2.50811934e-01
-2.72146642e-01 -5.45118928e-01 1.11803877e+00 -6.61258459e-01
1.33815217e+00 -2.28127122e+00 5.10597408e-01 -5.09261727e-01
-1.21514266e-02 8.88926163e-02 -2.49852583e-01 -4.21152301e-02
-1.43874496e-01 1.21882252e-01 -1.42294854e-01 -1.01165861e-01
-3.65838557e-01 5.05389988e-01 -3.91492173e-02 4.74150866e-01
4.99340534e-01 1.17790270e+00 -1.07503009e+00 -1.19405782e+00
4.79421645e-01 5.64523518e-01 -4.04378742e-01 3.80455136e-01
-1.72821820e-01 7.09752321e-01 -8.93309414e-01 9.20654833e-01
2.96936572e-01 -2.33060032e-01 -4.52085882e-01 -6.82925701e-01
6.54102350e-03 -3.16886634e-01 -8.66211414e-01 1.95554960e+00
-4.21441138e-01 4.72260088e-01 5.98527938e-02 -1.17971838e+00
5.86182892e-01 4.98685054e-02 5.95712006e-01 -8.55265439e-01
5.24727046e-01 -3.21480930e-01 -2.32525036e-01 -1.06171548e+00
3.51150423e-01 4.75442074e-02 -2.71828175e-01 1.81709185e-01
5.14506280e-01 5.31984009e-02 3.60735714e-01 3.60385597e-01
5.53568125e-01 3.37020516e-01 5.45098782e-01 -2.48133108e-01
1.17427170e+00 -2.52073944e-01 7.26647615e-01 4.77585524e-01
-7.48031616e-01 4.80606407e-01 5.05486965e-01 -3.78900617e-01
-5.74026227e-01 -6.50797367e-01 1.85462043e-01 1.60562789e+00
5.80736041e-01 -1.01778440e-01 -8.68535995e-01 -6.99466288e-01
-2.85152048e-01 6.10481083e-01 -1.05107200e+00 -2.74022341e-01
-6.58010244e-01 -1.42603576e-01 -9.51906946e-03 8.36418211e-01
8.92585039e-01 -1.55612481e+00 -7.49574125e-01 1.02605343e-01
-4.70119923e-01 -1.44309032e+00 -1.02383816e+00 7.27079064e-02
-5.98827660e-01 -1.06946051e+00 -5.92567861e-01 -9.96616364e-01
7.47356594e-01 3.92066479e-01 1.19239485e+00 -5.53650111e-02
-1.82530373e-01 8.86751413e-01 -5.30351639e-01 1.95055902e-01
-1.70799159e-03 -2.37949714e-01 -4.23260778e-01 4.04546559e-01
4.50737864e-01 -1.89440958e-02 -6.55680537e-01 3.55173171e-01
-1.10192561e+00 -8.24642461e-03 3.32717538e-01 9.00887430e-01
1.02496958e+00 -2.37752706e-01 5.09275973e-01 -6.20772600e-01
3.01283032e-01 -5.49881756e-01 -8.88342112e-02 4.91090953e-01
3.11200112e-01 -8.73571262e-02 4.29948300e-01 -4.87549126e-01
-1.11550617e+00 1.35660887e-01 6.92915544e-02 -8.59205127e-01
-2.84435391e-01 1.58877850e-01 -5.55038095e-01 1.67503089e-01
-2.74351574e-02 3.32469612e-01 -3.42516214e-01 2.20958684e-02
8.16753805e-01 4.88147438e-01 8.68720651e-01 -7.36255109e-01
2.44452152e-02 5.56021810e-01 -2.15770528e-01 -6.38568461e-01
-1.34511423e+00 -7.29092717e-01 -1.07941389e+00 -5.26559114e-01
1.57610619e+00 -9.33797956e-01 -7.95504093e-01 2.49660403e-01
-1.29961681e+00 -5.48897944e-02 -5.63964248e-01 2.37374768e-01
-8.75115991e-01 4.24957037e-01 -4.74706948e-01 -7.00344205e-01
-2.34473750e-01 -1.48127282e+00 1.53932333e+00 7.68770933e-01
2.15186384e-02 -1.10756004e+00 -4.92718846e-01 5.71298122e-01
9.99796912e-02 3.55153412e-01 5.74063182e-01 -5.59513271e-01
-6.21670246e-01 1.29261807e-01 -4.55299705e-01 3.48531127e-01
2.58055627e-01 2.74797883e-02 -7.11275518e-01 9.75303072e-03
-2.00945400e-02 -4.54874933e-01 1.10683405e+00 7.27112710e-01
1.35446501e+00 -1.45468324e-01 -2.60202169e-01 5.97213328e-01
1.14797235e+00 4.73827481e-01 4.75748032e-01 3.03025693e-02
9.21519578e-01 7.44243145e-01 9.72898901e-01 2.14502126e-01
4.59155887e-01 5.72180271e-01 3.83349836e-01 -3.22557628e-01
-2.64515042e-01 -9.05737281e-02 3.29700291e-01 6.15011156e-01
-1.42412651e-02 -2.96761960e-01 -4.49002862e-01 7.26127863e-01
-2.07288742e+00 -1.33419859e+00 -6.34724349e-02 1.56720841e+00
7.41011560e-01 -1.32231683e-01 1.26002908e-01 -2.33909518e-01
1.01786745e+00 3.04440767e-01 -8.90785515e-01 -3.73675317e-01
-4.12030488e-01 -1.50013641e-01 1.77142993e-01 4.05215919e-01
-1.58344722e+00 1.35046756e+00 5.52417231e+00 1.05256367e+00
-8.70125949e-01 3.67286414e-01 9.70495045e-01 4.75028194e-02
-5.03598712e-02 -2.12210253e-01 -5.29628992e-01 4.79343146e-01
5.80965877e-01 -2.12259173e-01 2.76445508e-01 6.88027322e-01
2.86611557e-01 -4.74391401e-01 -9.49480176e-01 9.83378232e-01
5.76461256e-01 -1.06279218e+00 2.43691221e-01 -4.91217762e-01
7.79975414e-01 -3.44572306e-01 7.86200017e-02 2.65087247e-01
-9.67682451e-02 -6.75267994e-01 7.76526809e-01 8.95175457e-01
6.02328539e-01 -9.53072488e-01 7.49094427e-01 -4.22064215e-02
-1.52347827e+00 -1.58683792e-01 -5.18045425e-02 3.31604391e-01
3.70290220e-01 -3.39647825e-03 1.54229760e-01 5.30306280e-01
9.15414870e-01 1.18916798e+00 -5.20585239e-01 4.15770680e-01
-2.50432134e-01 2.38723248e-01 5.31345159e-02 3.04740608e-01
7.37525880e-01 -1.26893058e-01 2.68507689e-01 1.28793287e+00
-1.15808442e-01 6.48679137e-01 3.62857401e-01 8.68572891e-01
-3.57836150e-02 2.53724396e-01 -3.82274747e-01 6.64833263e-02
2.23182619e-01 1.23003292e+00 -9.38900948e-01 -7.44525731e-01
-7.18330562e-01 1.29774857e+00 1.19429402e-01 5.48008859e-01
-1.04903638e+00 -1.72841743e-01 4.45141613e-01 -1.51801661e-01
5.23164093e-01 1.15036823e-01 2.19522417e-01 -1.05622911e+00
-1.68519422e-01 -6.54813945e-01 7.60104358e-01 -1.20323491e+00
-1.25268757e+00 7.48996973e-01 4.70637120e-02 -1.18330967e+00
-3.95243951e-05 -1.54050663e-01 -5.36778033e-01 7.90438533e-01
-1.55639637e+00 -1.31485796e+00 -4.52928185e-01 1.15885651e+00
1.23974264e+00 3.62697281e-02 1.57711014e-01 1.68089181e-01
-6.57542884e-01 1.21664472e-01 -4.91074234e-01 1.52435705e-01
6.16583824e-01 -9.27484334e-01 -1.96057275e-01 1.06624186e+00
5.38130961e-02 6.80917978e-01 3.60220343e-01 -6.55715585e-01
-1.18503714e+00 -1.25665224e+00 5.17949343e-01 -1.83802247e-01
6.90394998e-01 1.95415735e-01 -1.00019825e+00 7.28628695e-01
6.11256838e-01 4.62589085e-01 5.07896125e-01 -5.37948906e-01
-2.76889130e-02 1.90117955e-01 -1.12888801e+00 3.48075390e-01
1.13030601e+00 -7.30450809e-01 -8.63475025e-01 3.28539312e-01
1.13016200e+00 -3.87980878e-01 -8.24827850e-01 3.97962928e-01
2.12555304e-01 -9.96997178e-01 1.09304130e+00 -8.51706445e-01
5.90918124e-01 -4.93030488e-01 -4.98288661e-01 -7.89256454e-01
-3.02136242e-01 -4.82642591e-01 -3.72009277e-02 1.66619396e+00
-2.94284701e-01 -3.33450176e-02 2.02659428e-01 5.66447020e-01
-2.60798335e-01 -6.87510371e-01 -9.03477073e-01 -2.92754173e-01
-2.19832242e-01 -3.55146646e-01 1.59359917e-01 9.24476266e-01
-6.11561313e-02 4.80933875e-01 -3.41254354e-01 -2.92599164e-02
3.59690547e-01 3.57193232e-01 4.23712969e-01 -6.84553683e-01
5.57879917e-02 -5.29181600e-01 -6.14263117e-01 -1.06275892e+00
8.45889568e-01 -6.92348480e-01 3.19453813e-02 -1.51708531e+00
5.28940797e-01 3.76385033e-01 -4.67606217e-01 4.79482800e-01
-6.92328870e-01 1.45832166e-01 3.66590619e-01 -2.25317515e-02
-1.31134129e+00 6.40718281e-01 1.63179851e+00 -2.76389509e-01
-1.75425336e-01 -3.17400813e-01 -6.06291711e-01 9.24838901e-01
3.38691920e-01 -1.51906878e-01 -3.63416016e-01 -3.08792472e-01
-4.25336033e-01 3.33600700e-01 5.02242148e-01 -6.17179453e-01
3.37309420e-01 -2.79773176e-01 4.17374581e-01 -8.97428930e-01
2.75694579e-01 -7.93948650e-01 -2.82321602e-01 1.74907237e-01
-5.89993477e-01 1.46344230e-01 1.92931727e-01 6.73290133e-01
-6.74727619e-01 3.42711210e-02 1.04280949e+00 -3.71333301e-01
-1.35254502e+00 3.57226074e-01 -2.04291373e-01 3.84014666e-01
1.41346025e+00 -2.77416587e-01 -2.35742733e-01 -2.90751159e-01
-1.04702997e+00 5.33569694e-01 3.01189542e-01 5.95298946e-01
8.30434322e-01 -1.35475659e+00 -3.63186002e-01 7.43475780e-02
2.25500658e-01 -1.39652908e-01 8.47045958e-01 1.04161918e+00
-5.74117377e-02 3.46699089e-01 -3.73126775e-01 -7.62056351e-01
-1.33692777e+00 1.06181848e+00 4.28150386e-01 1.73509300e-01
-5.49010873e-01 1.24312043e+00 6.76931679e-01 3.33646297e-01
1.72902003e-01 -5.17192364e-01 -5.52097619e-01 4.14192945e-01
5.67826390e-01 9.99543294e-02 -5.35783410e-01 -1.66788173e+00
-4.76624221e-01 1.13771009e+00 5.77725247e-02 2.41499588e-01
9.79419172e-01 -6.82859182e-01 -1.72853142e-01 7.01797426e-01
1.47028375e+00 -4.05859709e-01 -1.39005792e+00 -5.31878769e-01
-3.09462011e-01 -5.53724051e-01 2.49024883e-01 -3.70687634e-01
-1.58709502e+00 9.10257936e-01 5.11456311e-01 -8.43776315e-02
1.63960123e+00 5.61399639e-01 7.77499676e-01 -1.39057294e-01
1.33240551e-01 -1.12584555e+00 4.85743672e-01 2.36699656e-01
9.15452480e-01 -1.33960545e+00 1.68945146e-04 -2.92392433e-01
-1.26204586e+00 1.00323200e+00 8.48036945e-01 -1.98749185e-01
4.19540733e-01 1.76115264e-03 2.03287210e-02 -3.20618063e-01
-7.75180161e-01 -8.47710609e-01 4.84744996e-01 4.54681456e-01
4.96404141e-01 -3.77091855e-01 -3.50796521e-01 7.43486583e-01
5.35307229e-01 -1.42964907e-02 2.77543843e-01 8.40515852e-01
-4.38169330e-01 -6.38370156e-01 -3.48629057e-01 6.17442019e-02
-4.89254683e-01 3.30528081e-03 -4.11447614e-01 7.75295138e-01
4.56164330e-01 1.02202821e+00 3.98631960e-01 -1.75834611e-01
3.86119664e-01 1.03267975e-01 3.47759664e-01 -3.41509879e-01
-8.09395194e-01 5.30827463e-01 -2.76075244e-01 -9.90089118e-01
-1.42221725e+00 -9.19852376e-01 -1.62569308e+00 2.36798093e-01
-1.19271792e-01 5.35593070e-02 3.92368697e-02 1.08979142e+00
3.43022317e-01 1.08402312e+00 4.99406129e-01 -9.64307249e-01
4.08347845e-01 -7.28090346e-01 -5.56213737e-01 6.41822934e-01
3.75803918e-01 -8.85797024e-01 -1.58624262e-01 7.22431660e-01] | [10.004650115966797, 0.8482334613800049] |
6196f4a2-f506-4157-9031-f5bc0f2ef522 | discriminative-models-still-outperform | null | null | https://openreview.net/forum?id=BB1pmTFXqOS | https://openreview.net/pdf?id=BB1pmTFXqOS | Discriminative Models Still Outperform Generative Models in Aspect Based Sentiment Analysis In Cross-Domain and Cross-Lingual Settings | Aspect-based Sentiment Analysis (ABSA) helps to explain customers' opinions towards products and services. In the past, ABSA models were discriminative, but more recently generative models have been used to generate aspects and polarities directly from text. In contrast, discriminative models first select aspects from the text, and then classify the aspect's polarity. Previous results showed that generative models outperform discriminative models on several English ABSA datasets. Here, we rigorously contrast discriminative and generative models in several settings. We compare both model types in cross-lingual, cross-domain, and cross- lingual and domain, to understand generalizability in settings other than mono-lingual English in-domain. Our more thorough evaluation shows that, contrary to previous studies, discriminative models still clearly outperform generative models in almost all settings. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-1.11595690e-01 1.12294026e-01 -4.22798991e-01 -9.31778312e-01
-1.01943243e+00 -1.08140934e+00 8.59358728e-01 6.35989606e-02
-1.55402794e-01 4.71854150e-01 6.08897865e-01 -4.42205340e-01
8.76108408e-02 -8.82974863e-01 -3.83437604e-01 -4.15589124e-01
5.96526980e-01 8.06239426e-01 -3.57285321e-01 -5.71802676e-01
1.80203900e-01 8.52991175e-03 -1.31723678e+00 6.65706813e-01
7.24608660e-01 8.05835187e-01 -1.89484730e-01 2.67260939e-01
-5.81402302e-01 3.78910154e-01 -6.49172604e-01 -1.27908063e+00
7.10488390e-03 -5.00973225e-01 -6.27566755e-01 2.09867656e-01
2.08231479e-01 1.10102214e-01 2.34860793e-01 7.45279610e-01
4.74796951e-01 -2.53856748e-01 8.64819944e-01 -1.16000581e+00
-1.13147652e+00 8.22775185e-01 -4.65378076e-01 -1.04928121e-01
7.15562105e-01 -1.30590200e-01 1.66794229e+00 -8.73773813e-01
7.70784140e-01 1.38009274e+00 9.76291001e-01 3.44481587e-01
-1.54499114e+00 -4.16586727e-01 6.33409441e-01 -1.28530592e-01
-9.55274880e-01 -9.99243408e-02 9.26855803e-01 -2.78105706e-01
1.10110033e+00 3.29961181e-01 1.01197648e+00 1.51822543e+00
2.25275919e-01 1.22900844e+00 1.71348715e+00 -3.31393629e-01
1.44249350e-01 6.81145012e-01 1.95116669e-01 -1.50173023e-01
5.05726874e-01 -3.98611665e-01 -6.48478746e-01 -3.47828507e-01
1.45235568e-01 -2.07646459e-01 1.09381586e-01 -1.47675246e-01
-7.89263844e-01 1.39700842e+00 -2.89417565e-01 4.09075767e-01
-5.76229274e-01 -4.79520798e-01 2.62474865e-01 3.93109024e-01
8.88972759e-01 6.25520647e-01 -9.42008257e-01 -5.96162736e-01
-8.30978215e-01 5.51771700e-01 1.40750945e+00 1.18798602e+00
9.17658746e-01 1.63512573e-01 -8.48564729e-02 9.99036252e-01
5.17767668e-01 8.25945020e-01 6.68583512e-01 -3.54195327e-01
3.77949357e-01 6.70112908e-01 -1.17413059e-01 -1.09626496e+00
-2.21368790e-01 -4.90209430e-01 -1.96028158e-01 -3.14873815e-01
1.51357725e-01 -1.71856269e-01 -9.76684928e-01 1.48376775e+00
2.39192657e-02 -7.40957439e-01 8.56537148e-02 6.59093261e-01
8.53715360e-01 4.97935772e-01 1.63905933e-01 -5.52344732e-02
1.68880296e+00 -5.12897372e-01 -7.76661634e-01 -7.36129463e-01
6.99182570e-01 -1.30641997e+00 1.39100075e+00 5.21626234e-01
-8.37890506e-01 -3.70828539e-01 -7.19304979e-01 -1.09063193e-01
-8.39696348e-01 7.74088576e-02 1.16216302e+00 1.17502737e+00
-9.35967684e-01 -1.08086132e-01 -4.28074539e-01 -4.84461427e-01
3.67516816e-01 1.86016500e-01 -2.11546570e-01 -1.79528907e-01
-1.17846751e+00 7.38192141e-01 -1.45327285e-01 -2.10668012e-01
-4.30959463e-01 -6.71387196e-01 -1.19254196e+00 -2.81964660e-01
5.10122851e-02 -6.95438266e-01 1.37939680e+00 -1.30774426e+00
-1.38193214e+00 1.09067512e+00 -6.27669275e-01 -2.92879969e-01
-1.52562186e-02 -4.54732239e-01 -6.42449558e-01 -5.39507926e-01
4.16017681e-01 5.08617878e-01 5.58683276e-01 -1.35550416e+00
-6.58073246e-01 -4.81219977e-01 4.03098583e-01 2.00438946e-01
-1.22746445e-01 9.95129868e-02 -4.69598442e-01 -6.83709979e-01
4.11827192e-02 -1.12435853e+00 -2.27761477e-01 -9.26036358e-01
-4.32952821e-01 -4.12976980e-01 7.48331189e-01 -3.71429503e-01
1.10658002e+00 -1.95724583e+00 -3.30794424e-01 2.55091846e-01
-1.27730653e-01 -9.45151746e-02 -6.96972525e-03 5.82267642e-01
-6.41449615e-02 1.45432487e-01 -8.44492391e-03 -6.77624047e-01
5.33723295e-01 4.03561860e-01 -4.75822449e-01 -1.27097499e-02
2.81542331e-01 1.21320975e+00 -4.78157699e-01 -3.26368600e-01
-2.39727087e-03 7.95741677e-01 -1.03166687e+00 -2.00561196e-01
-2.56751865e-01 3.12403291e-02 -5.85451663e-01 9.13207173e-01
6.94206238e-01 -1.54143423e-01 4.53633010e-01 -1.11402929e-01
9.38003585e-02 9.16920006e-01 -7.67901599e-01 1.34313858e+00
-9.17805016e-01 8.99244010e-01 -2.72494137e-01 -7.86555588e-01
8.93743932e-01 6.12177402e-02 3.77445102e-01 -9.05859411e-01
1.13584228e-01 2.27147549e-01 3.26284254e-03 -2.30864868e-01
9.51973438e-01 -5.78195989e-01 -5.76389015e-01 4.99160528e-01
-1.87300183e-02 -6.42626345e-01 3.66902679e-01 2.18740091e-01
3.04891735e-01 1.63131863e-01 4.10892069e-01 -3.14532161e-01
2.15680584e-01 2.18993023e-01 5.22718906e-01 4.55191344e-01
1.30704135e-01 8.41057360e-01 9.01229739e-01 -3.48472446e-01
-7.37702370e-01 -8.16461325e-01 -3.10346186e-01 1.29455674e+00
-1.59577444e-01 -9.75474954e-01 -4.49595481e-01 -9.67541397e-01
1.80811703e-01 1.28953576e+00 -9.36121285e-01 1.56440109e-01
-2.59739608e-01 -1.16524673e+00 2.40646768e-02 6.58620894e-01
1.13888634e-02 -8.97711575e-01 1.00828752e-01 1.50824398e-01
-1.51151374e-01 -1.18938768e+00 -2.33560771e-01 2.75968730e-01
-5.50997019e-01 -6.17789090e-01 -5.24636269e-01 -7.48080075e-01
3.64579469e-01 1.79221798e-02 1.84018898e+00 -7.11155653e-01
7.33569488e-02 6.13805830e-01 -5.07090867e-01 -7.24375546e-01
-3.07513058e-01 3.69159132e-01 -2.77873516e-01 -6.57472108e-03
1.35931396e+00 -4.83853966e-01 -4.44413722e-01 2.74988234e-01
-7.88054466e-01 -2.85132021e-01 7.04729676e-01 4.97766256e-01
7.75193572e-01 -1.85333952e-01 5.43831348e-01 -1.60881948e+00
9.40228701e-01 -5.62376738e-01 -1.92258045e-01 -1.36385784e-01
-1.03146768e+00 -2.23202094e-01 3.21716249e-01 -2.73959219e-01
-1.05132020e+00 -3.26427966e-01 -5.27299881e-01 2.06713334e-01
-3.69058996e-01 8.08284461e-01 -3.05403650e-01 5.39134502e-01
3.30762714e-01 2.10521519e-01 -3.31628472e-01 -2.82555699e-01
3.84101808e-01 7.38240898e-01 -3.40687096e-01 -4.15442169e-01
4.35527742e-01 5.18673658e-01 -5.84884644e-01 -7.25174606e-01
-1.06582355e+00 -3.72191489e-01 -4.01372075e-01 4.92038913e-02
6.56500340e-01 -1.11186719e+00 -3.49079639e-01 2.22976759e-01
-8.35014343e-01 -5.31325005e-02 -5.29851973e-01 5.30808806e-01
-4.27125335e-01 -9.69246924e-02 -4.52020139e-01 -6.46494985e-01
-2.03113794e-01 -1.17968655e+00 1.33642018e+00 3.19292754e-01
-8.93810987e-01 -1.49496782e+00 2.58527428e-01 5.20896792e-01
5.96935272e-01 -7.45759457e-02 9.02735472e-01 -1.14013076e+00
4.13070023e-02 -4.40802544e-01 1.50439918e-01 4.29462940e-01
4.96711999e-01 5.48938960e-02 -1.05471480e+00 5.18269427e-02
2.36988485e-01 -1.56310126e-01 5.83010674e-01 6.07079744e-01
5.33859968e-01 -3.09352070e-01 -2.34776750e-01 3.32905620e-01
1.14755249e+00 3.00034493e-01 6.88600838e-01 5.39567709e-01
5.85828185e-01 7.32364595e-01 8.46506238e-01 3.65030736e-01
9.86562014e-01 5.65287352e-01 -1.74279556e-01 -1.42891020e-01
4.26225401e-02 -3.25944185e-01 6.79683805e-01 1.02962852e+00
1.66278213e-01 -2.82576025e-01 -6.54943049e-01 8.16555202e-01
-1.34922850e+00 -7.44933128e-01 -2.99793333e-01 1.71955597e+00
9.33124661e-01 3.21589768e-01 4.25117403e-01 -1.71957895e-01
1.44851819e-01 3.94682437e-01 -2.53692687e-01 -9.98923421e-01
-6.03515089e-01 2.79719472e-01 1.93769619e-01 3.67295712e-01
-9.70908582e-01 9.86392140e-01 7.39831257e+00 8.31634045e-01
-1.13388383e+00 1.15851820e-01 9.85732555e-01 -1.09808318e-01
-1.35790575e+00 2.48681784e-01 -1.05507016e+00 4.04933840e-01
7.85705209e-01 -2.00713322e-01 -6.90876245e-02 1.36352563e+00
-1.08435936e-01 -2.72049066e-02 -1.08870673e+00 7.36780465e-01
4.79986370e-01 -8.33421648e-01 2.35821322e-01 2.71766424e-01
1.07630134e+00 -6.76633939e-02 5.63218951e-01 7.31329262e-01
5.41657209e-01 -9.76278901e-01 6.21006250e-01 -1.08301952e-01
4.42466736e-01 -9.58128393e-01 1.03262615e+00 -2.49059394e-01
-8.38156462e-01 3.99862558e-01 -1.95732906e-01 6.50502294e-02
4.45407212e-01 8.50641549e-01 -7.66263664e-01 5.36531568e-01
7.78839588e-01 9.29463625e-01 -7.48888195e-01 1.72970951e-01
-2.92688847e-01 9.54716921e-01 -4.56133224e-02 -1.34126619e-01
5.31546354e-01 -6.44242346e-01 3.84010255e-01 1.49542034e+00
2.50739604e-01 -3.39913487e-01 -1.73764810e-01 7.20289350e-01
2.40867943e-01 5.02021611e-01 -8.05034518e-01 -5.94675422e-01
-3.35875377e-02 1.30436254e+00 -7.28345156e-01 -3.72810662e-01
-9.22393084e-01 7.70365357e-01 -1.71396092e-01 4.16496694e-01
-7.81675279e-01 -1.67406619e-01 8.60598862e-01 3.65217894e-01
6.07085943e-01 -1.60433084e-01 -7.23854780e-01 -1.15634954e+00
-2.65846979e-02 -1.17600822e+00 4.23272073e-01 -7.36606538e-01
-1.55492759e+00 8.00242245e-01 -1.26695424e-01 -1.10365570e+00
-7.59107172e-01 -8.06551754e-01 -3.17357123e-01 9.02489364e-01
-1.48500860e+00 -1.46953440e+00 8.98831859e-02 3.98895085e-01
7.40586996e-01 -1.96852282e-01 8.35033953e-01 2.42671967e-01
-1.95764035e-01 5.22619128e-01 -4.10767198e-02 -5.62367514e-02
9.51693118e-01 -1.48989427e+00 7.00207293e-01 5.12463152e-01
4.85359371e-01 1.16896462e+00 9.57864881e-01 -6.09243989e-01
-1.30582321e+00 -6.12721980e-01 1.41598833e+00 -8.03941131e-01
7.68846214e-01 -5.94948769e-01 -5.58067322e-01 1.04447365e+00
7.24924386e-01 -6.59754753e-01 1.32832527e+00 1.06390560e+00
-6.33948922e-01 -1.16730645e-01 -8.57840180e-01 6.33349836e-01
6.21057570e-01 -8.06657374e-01 -5.81356466e-01 1.37783289e-01
4.52374518e-01 -1.46778032e-01 -8.44723761e-01 3.23362887e-01
8.45520377e-01 -1.22060263e+00 5.98332167e-01 -6.13050759e-01
5.68948746e-01 -4.17562760e-03 -5.52206159e-01 -1.54250276e+00
-2.38387033e-01 -5.73995709e-01 4.54754531e-01 1.65481627e+00
1.03220487e+00 -9.76398647e-01 7.42865860e-01 7.50678301e-01
-1.68743983e-01 -8.88571203e-01 -4.53547686e-01 -4.88264441e-01
3.17684025e-01 -9.36275840e-01 7.89638877e-01 8.67660999e-01
-1.15864500e-01 5.63027620e-01 1.05695419e-01 -2.82688200e-01
-7.51620382e-02 7.32193232e-01 7.77223706e-01 -8.27335477e-01
-4.99477148e-01 -6.55706644e-01 -4.12685543e-01 -1.06934857e+00
3.41641873e-01 -6.66812539e-01 -1.92238316e-01 -1.62480652e+00
2.97312289e-01 -4.27499592e-01 4.47794683e-02 3.69773805e-01
-2.60404110e-01 5.10608017e-01 -1.39068663e-01 -8.21844339e-02
-2.83762723e-01 4.63445514e-01 1.08612466e+00 -1.43521532e-01
-2.04021320e-01 4.61291701e-01 -1.61672270e+00 7.26013958e-01
7.56675184e-01 -4.28116530e-01 -6.19254708e-01 -2.83624202e-01
8.60506594e-01 -8.13352466e-01 -1.18689395e-01 -1.73274055e-01
-4.25291628e-01 -1.69800416e-01 3.73591691e-01 -6.58990622e-01
4.93082523e-01 -5.10969162e-01 -6.31607473e-02 -2.04116404e-01
-1.47109598e-01 4.17704701e-01 3.68588179e-01 2.77909219e-01
-6.37252450e-01 -6.54237941e-02 1.69194832e-01 9.64487270e-02
-4.10764426e-01 -3.89586426e-02 -6.76728904e-01 3.40382755e-01
6.80033326e-01 -4.22322065e-01 -1.89636841e-01 -7.52584577e-01
-6.11852586e-01 -3.23448032e-01 5.21488667e-01 6.48973227e-01
2.40573928e-01 -1.36240363e+00 -5.77874720e-01 4.23821986e-01
3.57454300e-01 -2.93748796e-01 2.32136607e-01 1.00340462e+00
-1.43152848e-01 8.26503396e-01 1.50564969e-01 -5.31365275e-01
-1.03644276e+00 3.45476747e-01 -7.08420649e-02 -3.18763286e-01
-4.76839542e-02 7.30881691e-01 5.05382001e-01 -6.63810670e-01
-5.24669111e-01 -4.22731400e-01 -2.85886705e-01 7.17681646e-01
8.55481476e-02 -9.20207351e-02 2.15342864e-01 -1.06829274e+00
-4.78808701e-01 5.52319884e-01 -4.49412733e-01 -5.41580498e-01
1.31948948e+00 -2.46603757e-01 7.49105215e-02 8.82477939e-01
1.23672163e+00 6.43859744e-01 -5.99164188e-01 1.77107495e-03
-1.60908356e-01 -3.62581253e-01 -1.05433673e-01 -8.71702254e-01
-9.99104500e-01 6.90706789e-01 -1.11738760e-02 7.28399634e-01
9.68691587e-01 3.77993107e-01 5.11041522e-01 -1.00668162e-01
1.66458920e-01 -1.18840957e+00 -2.10252717e-01 4.42994088e-01
6.83991551e-01 -1.23797858e+00 -9.99939721e-03 -4.58929151e-01
-1.47871542e+00 6.63788319e-01 4.68183368e-01 1.22075200e-01
8.17603350e-01 3.46040159e-01 4.35258538e-01 -3.84812951e-01
-8.02216291e-01 -3.80839318e-01 4.96968895e-01 7.23285437e-01
1.09620297e+00 2.45386079e-01 -2.95833647e-01 1.14741981e+00
-9.77646053e-01 -4.96693015e-01 3.88138711e-01 9.07727122e-01
2.98438042e-01 -1.39539504e+00 -4.57092412e-02 5.93779981e-01
-7.57679701e-01 -4.33054328e-01 -7.37003744e-01 1.03182662e+00
-1.85548142e-02 1.11769748e+00 1.27240658e-01 -3.56023133e-01
4.74327356e-01 2.30008408e-01 2.44940922e-01 -7.15366900e-01
-6.98186457e-01 4.71998304e-01 5.17804027e-01 -4.74754065e-01
-4.70837444e-01 -1.26563787e+00 -6.22524798e-01 -2.09740400e-01
-3.10009658e-01 3.16087902e-01 8.76718283e-01 1.01951933e+00
6.06235802e-01 2.62750894e-01 6.08197391e-01 -2.86758482e-01
2.16404814e-02 -1.14413929e+00 -7.49789953e-01 4.66616660e-01
-4.57475409e-02 -3.24597180e-01 -5.52464187e-01 9.12579000e-02] | [11.42342758178711, 6.722738742828369] |
9c9dad5d-91cb-4552-ad2e-3e68fc8bee43 | prompt-learning-to-mitigate-catastrophic | 2305.07393 | null | https://arxiv.org/abs/2305.07393v2 | https://arxiv.org/pdf/2305.07393v2.pdf | Prompt Learning to Mitigate Catastrophic Forgetting in Cross-lingual Transfer for Open-domain Dialogue Generation | Dialogue systems for non-English languages have long been under-explored. In this paper, we take the first step to investigate few-shot cross-lingual transfer learning (FS-XLT) and multitask learning (MTL) in the context of open-domain dialogue generation for non-English languages with limited data. We observed catastrophic forgetting in both FS-XLT and MTL for all 6 languages in our preliminary experiments. To mitigate the issue, we propose a simple yet effective prompt learning approach that can preserve the multilinguality of multilingual pre-trained language model (mPLM) in FS-XLT and MTL by bridging the gap between pre-training and fine-tuning with Fixed-prompt LM Tuning and our hand-crafted prompts. Experimental results on all 6 languages in terms of both automatic and human evaluations demonstrate the effectiveness of our approach. Our code is available at https://github.com/JeremyLeiLiu/XLinguDial. | ['Jimmy Xiangji Huang', 'Lei Liu'] | 2023-05-12 | null | null | null | null | ['dialogue-generation', 'cross-lingual-transfer', 'dialogue-generation'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [-8.83100554e-02 3.68316650e-01 1.64583027e-02 -4.20103937e-01
-1.42111325e+00 -7.84974337e-01 8.47366333e-01 -1.37958914e-01
-6.86463058e-01 1.35721445e+00 4.32316273e-01 -7.02271104e-01
2.44074866e-01 -3.16721499e-01 -5.77706695e-01 -2.61054933e-01
1.05438791e-01 6.94792688e-01 1.92154035e-01 -7.10725069e-01
-6.69045523e-02 -3.15892428e-01 -8.57533872e-01 4.46547270e-01
1.28022599e+00 1.98252097e-01 3.77856106e-01 7.73647428e-01
-2.52405941e-01 7.42290139e-01 -5.84854186e-01 -6.20731890e-01
9.22161341e-02 -6.37710989e-01 -1.20923150e+00 -1.62918642e-01
4.00708109e-01 -2.85769373e-01 -8.41696486e-02 6.31913722e-01
8.14354002e-01 3.11112195e-01 4.52188313e-01 -1.03689206e+00
-4.97601986e-01 9.31483984e-01 -3.69621992e-01 4.67893630e-02
3.96287948e-01 4.02693331e-01 9.72851336e-01 -1.10117638e+00
7.05197215e-01 1.57977557e+00 6.78756595e-01 8.84099424e-01
-1.27625144e+00 -5.05724728e-01 8.03416744e-02 3.17961201e-02
-1.17813730e+00 -9.32826579e-01 4.22292024e-01 -2.43709505e-01
1.29871356e+00 1.99668705e-02 1.11599267e-01 1.42477524e+00
2.67472476e-01 9.38786030e-01 1.36392999e+00 -1.00589371e+00
-2.34480463e-02 4.94731188e-01 -2.12863475e-01 7.53159523e-01
-1.98163897e-01 -1.94402993e-01 -7.07918644e-01 -1.64630204e-01
4.77606386e-01 -6.93012714e-01 -1.44214794e-01 -3.10393814e-02
-1.31006670e+00 1.02205765e+00 -3.40450555e-01 5.22391558e-01
2.74754036e-02 -2.42969185e-01 7.73418069e-01 6.85821533e-01
9.39744294e-01 5.37829876e-01 -8.57416570e-01 -3.32359642e-01
-7.38814294e-01 -3.81513834e-02 1.05205047e+00 1.06766295e+00
5.72291315e-01 1.71892166e-01 -4.82042342e-01 1.37253523e+00
-1.14659838e-01 4.16635156e-01 7.40554929e-01 -6.92439616e-01
7.20065832e-01 6.92051426e-02 1.24825403e-01 -3.32813449e-02
-1.64651185e-01 1.51258320e-01 -4.64528203e-01 -1.27222493e-01
5.77863157e-01 -7.51421571e-01 -6.15549266e-01 1.96096790e+00
2.16196269e-01 -1.77919388e-01 5.48102617e-01 4.92876798e-01
5.05855918e-01 9.05431509e-01 2.66397804e-01 -4.33697164e-01
1.21415186e+00 -1.30429578e+00 -8.98129344e-01 -3.60639960e-01
1.09841979e+00 -1.02830577e+00 1.65504646e+00 1.07204638e-01
-1.10283637e+00 -5.24097383e-01 -7.51132429e-01 -2.86116511e-01
-3.24663341e-01 2.68700808e-01 3.88424546e-01 4.37218159e-01
-1.14972663e+00 3.97994041e-01 -7.23923087e-01 -7.24245965e-01
-1.37946233e-01 6.45494163e-02 -3.26202214e-01 -9.41289067e-02
-1.69695604e+00 1.33463120e+00 4.75253016e-01 -9.96929407e-02
-8.98159266e-01 -4.59752232e-01 -8.25990260e-01 -3.38325053e-01
5.45151293e-01 -4.08771247e-01 1.79177737e+00 -7.61647880e-01
-1.86732709e+00 8.73105049e-01 -1.42991215e-01 -3.30537379e-01
7.33868420e-01 -5.37545979e-01 -3.00694138e-01 -2.61551559e-01
2.50182062e-01 7.41579056e-01 5.45648098e-01 -1.07063544e+00
-8.34401429e-01 7.59131238e-02 -3.75432260e-02 5.74281514e-01
-4.29673135e-01 1.42085418e-01 -1.95140079e-01 -5.64562380e-01
-7.04599917e-01 -9.46395338e-01 -1.25615284e-01 -6.00387454e-01
-3.26871783e-01 -6.84323370e-01 5.91323733e-01 -1.04922998e+00
1.25021172e+00 -1.88392091e+00 5.24333306e-02 -5.28174996e-01
-3.00411075e-01 6.92065477e-01 -4.53272045e-01 9.26130652e-01
3.28349888e-01 1.06700331e-01 -3.30914557e-01 -6.40161633e-01
-4.19232510e-02 2.17471927e-01 -2.56894976e-01 9.39541310e-02
2.55827725e-01 9.92043197e-01 -1.14414978e+00 -5.67582369e-01
2.21392944e-01 3.30123693e-01 -2.41183564e-01 4.28214520e-01
-6.37392521e-01 8.21213424e-01 -1.92239136e-01 3.77612889e-01
2.05979988e-01 1.10190079e-01 3.16754013e-01 1.86491087e-01
-4.22335982e-01 6.73043728e-01 -6.09007955e-01 2.08463407e+00
-9.13603723e-01 3.99959505e-01 -1.48517480e-02 -3.37961435e-01
7.89031744e-01 6.88165605e-01 4.50594500e-02 -8.11730981e-01
5.96870435e-03 5.26937306e-01 1.41819984e-01 -4.46929336e-01
7.72966385e-01 -2.60720998e-01 -4.80187207e-01 8.09207082e-01
6.35780692e-01 -1.89377800e-01 3.28376442e-01 2.65811414e-01
6.71991050e-01 3.74044180e-01 6.11651599e-01 -4.26725566e-01
4.51454580e-01 7.26317102e-03 3.53214771e-01 8.83213222e-01
-2.65370607e-01 2.17913464e-01 1.70006156e-01 -1.06657045e-02
-1.13644731e+00 -8.69243681e-01 6.84029087e-02 1.77506709e+00
-4.34744567e-01 -4.47526664e-01 -8.17498982e-01 -7.83082962e-01
-2.46512815e-01 1.44233477e+00 -1.21265605e-01 -1.12120636e-01
-7.15360463e-01 -5.60921848e-01 9.07384455e-01 5.85846938e-02
3.40993255e-01 -1.25303662e+00 -3.45701277e-01 5.20562530e-01
-5.62141657e-01 -1.17055559e+00 -8.07604253e-01 1.91573948e-01
-5.90261519e-01 -5.90760231e-01 -7.79040277e-01 -9.10957754e-01
2.31130689e-01 4.73852903e-02 1.07872367e+00 -2.07503751e-01
-4.05926369e-02 2.77681619e-01 -3.79383147e-01 -2.13292822e-01
-9.35777545e-01 5.81764221e-01 2.96992064e-01 -1.55726492e-01
3.25634807e-01 -3.60504329e-01 3.07748299e-02 8.39903951e-02
-5.41081131e-01 3.09000850e-01 5.03572583e-01 1.06534910e+00
1.36579618e-01 -5.62883079e-01 8.66975725e-01 -1.15910006e+00
1.23570347e+00 -3.57851923e-01 -3.92481714e-01 6.91176653e-01
-6.13660574e-01 1.71686783e-01 5.74863315e-01 -5.58294475e-01
-1.58252442e+00 -2.20417649e-01 -2.31927663e-01 -9.74443555e-02
-1.55791014e-01 4.58186060e-01 -3.66431586e-02 2.62740761e-01
5.68363726e-01 1.54327407e-01 -2.02258170e-01 -5.82297564e-01
7.83549488e-01 9.05274332e-01 5.22172809e-01 -8.62079203e-01
4.15895700e-01 -7.47214779e-02 -7.98131645e-01 -8.61754954e-01
-9.59828854e-01 -1.81951821e-01 -8.81152987e-01 -2.27655768e-01
7.77876914e-01 -1.15053046e+00 -1.94838658e-01 5.77788591e-01
-1.36902213e+00 -1.01051283e+00 -2.05285668e-01 3.35293591e-01
-5.64257681e-01 3.16080630e-01 -9.46196854e-01 -8.73094559e-01
-6.96443677e-01 -8.51354957e-01 8.99955809e-01 1.54088393e-01
-5.52401125e-01 -1.47005248e+00 4.87978786e-01 2.33361825e-01
5.39019942e-01 -2.61117578e-01 9.68053699e-01 -1.07362998e+00
-1.41559780e-01 2.61454970e-01 1.05908744e-01 3.00719202e-01
2.21517473e-01 -4.06620950e-01 -9.86815691e-01 -6.95420802e-01
-5.45862354e-02 -9.91689026e-01 5.44298172e-01 7.30752200e-02
1.66580871e-01 -5.32841444e-01 -6.72552139e-02 3.15775156e-01
1.19281077e+00 9.03751403e-02 2.98542202e-01 3.42057049e-01
6.14727974e-01 7.75495708e-01 8.79996896e-01 4.05251145e-01
7.35772729e-01 5.84251881e-01 -4.88452584e-01 -1.77145585e-01
-2.82807887e-01 -5.64985633e-01 7.59294331e-01 1.36093807e+00
2.73716241e-01 -4.26276535e-01 -1.01818573e+00 8.52332771e-01
-1.88217080e+00 -5.86564124e-01 8.02912265e-02 2.21477437e+00
1.53597224e+00 1.11828446e-01 -6.41205087e-02 -6.85474873e-01
7.23012686e-01 7.35581368e-02 -3.96752596e-01 -6.96023822e-01
-1.37897253e-01 1.67186633e-01 3.40975851e-01 1.14866543e+00
-8.95569682e-01 1.72191405e+00 5.42133331e+00 9.01226401e-01
-1.00661385e+00 7.22997546e-01 5.23140967e-01 -3.82989086e-02
-2.05532506e-01 2.36797437e-01 -1.09332991e+00 2.94493526e-01
1.34683216e+00 -4.97843415e-01 4.01271552e-01 4.30405676e-01
2.86911786e-01 -2.07213834e-01 -1.07679605e+00 4.28479314e-01
1.16271473e-01 -7.41113663e-01 -8.43902230e-02 -2.53535956e-01
7.24485695e-01 2.74676919e-01 -2.72069454e-01 9.26261306e-01
7.26655483e-01 -6.83121145e-01 5.79239845e-01 2.33609915e-01
1.14803123e+00 -6.09426379e-01 4.75194395e-01 6.30064607e-01
-7.51000166e-01 3.57485205e-01 -2.45369479e-01 -3.30219790e-02
4.91139531e-01 1.60660774e-01 -1.34707046e+00 5.44142842e-01
3.27132106e-01 3.22785467e-01 -5.57281911e-01 5.27735651e-01
-5.12648642e-01 7.91111827e-01 -1.64290756e-01 8.97454545e-02
2.68001258e-01 -2.11228207e-02 6.35114074e-01 1.56187129e+00
2.41256624e-01 -1.44806713e-01 4.85291630e-01 4.72443223e-01
-6.91225454e-02 3.94383997e-01 -6.83670700e-01 -1.03891112e-01
5.63894928e-01 1.18126142e+00 -3.15179467e-01 -4.30798441e-01
-4.94653285e-01 1.26550496e+00 6.70637727e-01 2.09203094e-01
-6.17351174e-01 -4.14629728e-01 2.60278195e-01 -2.17458099e-01
1.48666754e-01 -3.94804031e-01 1.19089941e-02 -1.37235415e+00
-1.56421065e-01 -1.08273637e+00 4.31099951e-01 -4.03183609e-01
-1.45117307e+00 9.63762522e-01 7.07517564e-02 -8.61665428e-01
-8.41332614e-01 -2.15318292e-01 -5.06788313e-01 1.03585625e+00
-1.56228256e+00 -1.42029524e+00 3.69105518e-01 5.96969306e-01
1.16269124e+00 -3.09027731e-01 1.12782288e+00 2.38659874e-01
-5.65811634e-01 7.95651197e-01 2.66823292e-01 -4.10720073e-02
1.46920359e+00 -1.19184411e+00 7.50646472e-01 9.27168310e-01
2.37085018e-02 4.82489586e-01 7.31037796e-01 -8.35795403e-01
-9.67829168e-01 -1.05738461e+00 1.62708259e+00 -5.65009415e-01
8.46791387e-01 -8.15783620e-01 -1.03522658e+00 8.85452986e-01
1.04493761e+00 -5.96154273e-01 7.18821824e-01 3.57907861e-01
-1.56574950e-01 3.05234969e-01 -8.21063280e-01 8.03128302e-01
7.19727695e-01 -6.87144399e-01 -7.50816882e-01 6.10657632e-01
9.09152865e-01 -4.75591153e-01 -9.05114233e-01 1.42926857e-01
2.61247724e-01 -5.37208855e-01 4.79490846e-01 -5.11069536e-01
1.83222204e-01 5.99823073e-02 1.64426602e-02 -1.85901248e+00
5.75894676e-02 -1.04802954e+00 1.52479991e-01 1.53898156e+00
7.22652078e-01 -5.83898365e-01 -5.80356177e-03 4.36562479e-01
-2.25087762e-01 -3.62065166e-01 -9.84388113e-01 -7.91684985e-01
4.31408823e-01 -7.80470371e-02 3.03698168e-03 1.08759475e+00
2.95348257e-01 1.10573566e+00 -9.35602069e-01 -2.16022998e-01
4.71254319e-01 -1.36740163e-01 8.12798738e-01 -8.06140065e-01
-3.72082978e-01 -1.10570136e-02 6.50025964e-01 -8.41020525e-01
4.48956966e-01 -9.19941008e-01 4.70779985e-01 -1.32793283e+00
1.12560857e-02 -3.49314213e-01 4.20098156e-02 8.69265378e-01
-3.21051538e-01 -1.93116531e-01 2.59651393e-01 2.11353153e-01
-8.08716118e-01 8.74167264e-01 1.05981052e+00 1.87132329e-01
-4.79339391e-01 -9.97868404e-02 -5.85823894e-01 4.42943901e-01
9.80114400e-01 -5.67028105e-01 -4.10833418e-01 -6.55313313e-01
-3.98442090e-01 2.14494675e-01 -1.95643932e-01 -6.49467528e-01
2.37107530e-01 -2.71296483e-02 -1.61140695e-01 -2.33112842e-01
1.77948028e-01 -1.52321577e-01 -3.49346608e-01 3.13127398e-01
-5.81968963e-01 7.78244734e-02 5.66585124e-01 7.53023997e-02
-8.56391117e-02 -2.80492723e-01 7.64468908e-01 -4.14554268e-01
-8.20764780e-01 -6.21705838e-02 -5.26243210e-01 4.68672603e-01
8.16557884e-01 4.17265505e-01 -4.97079730e-01 -5.61904788e-01
-5.31636894e-01 6.03183448e-01 2.69670516e-01 7.77009964e-01
2.91073918e-01 -1.03678107e+00 -1.13917410e+00 4.79369350e-02
1.82277247e-01 -3.37965727e-01 1.53062999e-01 7.25185215e-01
-5.96373752e-02 8.28304827e-01 -3.19385111e-01 -3.06610644e-01
-1.10515511e+00 2.08083928e-01 3.19268018e-01 -6.86982810e-01
-3.81366849e-01 7.97011077e-01 -8.59206095e-02 -1.09133279e+00
1.84611738e-01 1.55408144e-01 1.09839447e-01 1.25241563e-01
2.84713477e-01 1.25135526e-01 2.87891962e-02 -5.57739735e-01
-2.74770558e-02 -1.40573099e-01 -6.67271018e-01 -6.70786381e-01
1.13417947e+00 -5.93029320e-01 -3.82839181e-02 1.02982867e+00
9.33055520e-01 1.84400454e-01 -1.22635603e+00 -4.71899837e-01
3.54483008e-01 -1.62944347e-01 -2.51971722e-01 -1.19885182e+00
-1.54545233e-01 8.86041045e-01 2.78829992e-01 -2.05574587e-01
7.02063024e-01 -3.70810851e-02 1.02512491e+00 5.35440266e-01
6.12475693e-01 -1.37720156e+00 3.51572968e-02 1.13468766e+00
8.89360726e-01 -1.35717535e+00 -4.02492106e-01 -9.71264485e-03
-1.05938435e+00 8.14028263e-01 1.01211226e+00 3.25484663e-01
1.63450316e-01 1.65247798e-01 4.75769013e-01 1.87587097e-01
-1.31798780e+00 -1.63042039e-01 -1.03522554e-01 4.40531909e-01
1.01613593e+00 7.77649730e-02 -4.82310474e-01 4.94510919e-01
-2.13122398e-01 -1.48801804e-01 5.47171414e-01 1.03350711e+00
-3.59396964e-01 -1.56512189e+00 -8.19297656e-02 7.30927810e-02
-3.83561730e-01 -5.50772190e-01 -4.78078514e-01 8.40722620e-01
-1.64780959e-01 9.51220751e-01 -2.40002006e-01 -2.17455432e-01
3.65671128e-01 7.35643864e-01 5.13679802e-01 -9.47522044e-01
-7.44362473e-01 3.82803679e-01 5.78576267e-01 -1.74730822e-01
-1.92509472e-01 -7.90381074e-01 -9.25017953e-01 -1.26511589e-01
-4.18621480e-01 3.53607774e-01 3.02241951e-01 1.01759839e+00
2.28546783e-01 2.58579701e-01 4.71275687e-01 -5.73796391e-01
-8.22542369e-01 -1.57425582e+00 -3.26382488e-01 1.32419586e-01
2.86724687e-01 -2.93079406e-01 -2.91713148e-01 1.48181200e-01] | [12.382258415222168, 8.553850173950195] |
e795e292-77a5-4d93-a98a-df18ea0206b6 | passage-ranking-with-weak-supervsion | 1905.05910 | null | https://arxiv.org/abs/1905.05910v2 | https://arxiv.org/pdf/1905.05910v2.pdf | Passage Ranking with Weak Supervision | In this paper, we propose a \textit{weak supervision} framework for neural ranking tasks based on the data programming paradigm \citep{Ratner2016}, which enables us to leverage multiple weak supervision signals from different sources. Empirically, we consider two sources of weak supervision signals, unsupervised ranking functions and semantic feature similarities. We train a BERT-based passage-ranking model (which achieves new state-of-the-art performances on two benchmark datasets with full supervision) in our weak supervision framework. Without using ground-truth training labels, BERT-PR models outperform BM25 baseline by a large margin on all three datasets and even beat the previous state-of-the-art results with full supervision on two of the datasets. | ['Xiaofei Ma', 'Peng Xu', 'Bing Xiang', 'Ramesh Nallapati'] | 2019-05-15 | null | https://openreview.net/forum?id=S1ltj47xdE | https://openreview.net/pdf?id=S1ltj47xdE | iclr-workshop-lld-2019 | ['passage-ranking'] | ['natural-language-processing'] | [ 1.72669902e-01 -1.58878416e-02 -5.87113678e-01 -6.09636605e-01
-1.08167958e+00 -4.30168897e-01 1.12208366e+00 2.21981093e-01
-7.62684286e-01 7.70364702e-01 4.90205735e-01 -1.74189404e-01
-4.87382084e-01 -5.81626236e-01 -9.25364137e-01 -3.19375128e-01
-2.43719518e-02 8.49616051e-01 6.99855089e-01 -5.98401129e-01
1.31697372e-01 -2.30214298e-01 -1.75544667e+00 6.70500875e-01
9.07378674e-01 9.99714434e-01 1.62692949e-01 3.72289151e-01
5.12876213e-02 1.20408344e+00 -3.12376082e-01 -8.91566813e-01
1.97738811e-01 1.01110004e-01 -1.10337675e+00 -8.69042993e-01
6.19797587e-01 -3.83783191e-01 -4.78436500e-01 8.51570904e-01
5.30762672e-01 3.80874217e-01 6.85143769e-01 -8.21925163e-01
-8.23324561e-01 1.13116050e+00 -2.79989511e-01 2.06587180e-01
1.64097890e-01 -3.36413860e-01 1.76981258e+00 -9.45100486e-01
4.72231239e-01 9.68749404e-01 5.49982965e-01 6.92591190e-01
-1.40191126e+00 -6.18068933e-01 1.71517625e-01 1.97584927e-01
-1.03885579e+00 -3.30965251e-01 6.60670578e-01 -2.36578256e-01
9.28617179e-01 2.16397524e-01 -2.73753162e-02 1.45566869e+00
-4.16529983e-01 1.19715464e+00 1.08750868e+00 -5.25048852e-01
-1.98702421e-02 5.12484089e-02 8.99370611e-01 6.09746993e-01
5.40728159e-02 2.43650556e-01 -9.63301420e-01 -7.53502473e-02
3.06741834e-01 2.14763671e-01 -1.43222734e-01 -2.29288451e-02
-1.37265742e+00 6.33967221e-01 6.64087415e-01 3.05726588e-01
4.47610347e-03 2.79068559e-01 5.60742319e-01 5.85366249e-01
8.51599872e-01 6.19744241e-01 -8.66435468e-01 -1.69815384e-02
-8.79322767e-01 1.00326017e-01 6.16934419e-01 9.35750365e-01
5.63325942e-01 -3.66000444e-01 -6.75793350e-01 1.27060890e+00
3.31151903e-01 3.73020470e-01 5.56190670e-01 -8.61863494e-01
7.23397672e-01 4.08785194e-01 -9.59611312e-03 -2.95633733e-01
-3.77189070e-01 -7.57133126e-01 -7.07090855e-01 -2.78389186e-01
4.17842686e-01 1.15626782e-01 -1.09257233e+00 1.91436398e+00
-2.24114984e-01 3.53853345e-01 2.28762254e-03 7.31064260e-01
1.29474258e+00 3.42948616e-01 -1.69974368e-03 1.63746610e-01
8.68157804e-01 -1.72301173e+00 -2.10742816e-01 -1.58467531e-01
7.64085531e-01 -1.48544669e-01 1.54867303e+00 6.58043742e-01
-8.40424418e-01 -5.40023386e-01 -1.21337700e+00 -3.39035720e-01
-4.04684365e-01 1.92460492e-01 7.51770735e-01 2.05492184e-01
-1.36923790e+00 1.02160478e+00 -5.88257253e-01 -1.05367467e-01
2.35451952e-01 3.90445799e-01 -2.36316025e-01 -4.13299531e-01
-1.60057569e+00 9.41260815e-01 3.57182026e-01 1.43321585e-02
-1.30150902e+00 -7.65398622e-01 -6.22335434e-01 -5.66638224e-02
3.20716619e-01 -7.18512356e-01 1.31335652e+00 -6.59190476e-01
-1.58005118e+00 1.05107260e+00 1.12230718e-01 -6.54765725e-01
4.82067496e-01 -7.37164199e-01 -3.16791356e-01 2.63238116e-03
1.34083450e-01 5.62903523e-01 2.33700201e-01 -1.07502985e+00
-3.43994290e-01 -3.46575558e-01 2.35910401e-01 1.17046006e-01
-6.09511256e-01 1.84013635e-01 -6.30009294e-01 -5.99720001e-01
-2.74787366e-01 -6.89267874e-01 -2.13212520e-01 -5.68669200e-01
-4.99960274e-01 -6.42875910e-01 -3.83101664e-02 -3.21476817e-01
1.24385428e+00 -1.87324190e+00 2.88951993e-01 5.08510172e-02
2.03623652e-01 2.42912844e-01 -4.92578685e-01 3.21598977e-01
1.87816367e-01 2.87030160e-01 -8.57945299e-04 -8.51238608e-01
2.04686746e-01 2.41563782e-01 -6.88681424e-01 5.97714446e-03
1.08885080e-01 9.59367990e-01 -1.25052202e+00 -3.14313412e-01
-2.87394196e-01 2.61223435e-01 -6.21621132e-01 1.66648656e-01
-4.94894177e-01 2.55280763e-01 -5.16446948e-01 5.65241575e-01
2.36739442e-02 -5.21368146e-01 1.01312913e-01 1.29252881e-01
3.27585012e-01 8.70752215e-01 -6.49799585e-01 2.10086322e+00
-4.28170919e-01 2.23010466e-01 -4.87937778e-01 -9.76433575e-01
7.71581411e-01 1.92166656e-01 2.84814000e-01 -7.49530494e-01
-1.74376905e-01 4.45568770e-01 -3.01813602e-01 8.27896595e-02
4.63051289e-01 1.03921920e-01 -1.17042124e-01 4.12887692e-01
5.70712984e-01 1.16926320e-01 3.27703804e-01 4.31396395e-01
1.53299940e+00 3.13065022e-01 -2.23053917e-01 -3.78950626e-01
3.67813379e-01 -2.95933455e-01 4.89015043e-01 1.16383791e+00
4.94465791e-02 8.38476419e-01 3.03748727e-01 -3.35874379e-01
-5.06806254e-01 -1.19675696e+00 -3.01003575e-01 2.09923172e+00
1.54245039e-02 -7.04010606e-01 -2.53103137e-01 -1.07663155e+00
5.82524687e-02 6.78555787e-01 -8.53052497e-01 -8.81420523e-02
-3.59135598e-01 -7.76000977e-01 7.76344419e-01 5.95783710e-01
3.90737861e-01 -1.14949727e+00 1.17846660e-01 1.20430335e-01
-1.68198466e-01 -1.00187850e+00 -1.48851454e-01 6.75010026e-01
-1.10745013e+00 -6.75247133e-01 -7.72548378e-01 -6.56642616e-01
4.09330040e-01 1.76027313e-01 1.85042059e+00 3.92151773e-01
3.17690700e-01 -1.67779289e-02 -4.92261767e-01 -4.02671278e-01
1.80901483e-01 4.19033319e-01 1.49053052e-01 -4.28344429e-01
3.58096838e-01 -6.46796942e-01 -4.87071097e-01 5.26895642e-01
-7.45931149e-01 1.25765041e-01 4.88591075e-01 1.10208297e+00
6.64120317e-01 -3.17730188e-01 8.37175250e-01 -1.24744606e+00
4.97000396e-01 -5.42707860e-01 -3.89686167e-01 4.66599911e-01
-7.73311913e-01 4.19927180e-01 7.96585679e-01 -1.42710805e-01
-8.60066175e-01 -4.50039953e-01 6.19099364e-02 -4.05387461e-01
2.11312860e-01 7.67512262e-01 -2.00407300e-02 3.83894950e-01
9.36249793e-01 -1.02726687e-02 -5.90402067e-01 -8.51305902e-01
4.99319255e-01 4.93401945e-01 6.15301430e-01 -1.08440900e+00
5.80474675e-01 2.39832386e-01 -2.20240310e-01 -1.16882324e-01
-1.89514530e+00 -7.18807876e-01 -7.69651949e-01 1.65849015e-01
6.23485148e-01 -1.07864237e+00 -2.40136653e-01 3.46809953e-01
-1.02290261e+00 -6.74819827e-01 -2.00693727e-01 3.43987882e-01
-3.89746875e-01 -4.54358682e-02 -8.96746576e-01 -3.04275334e-01
-5.10735691e-01 -9.57535505e-01 1.09968662e+00 -9.39942822e-02
2.21389472e-01 -9.22153294e-01 3.53855699e-01 6.65657520e-01
3.75846565e-01 -3.37338373e-02 7.80587196e-01 -1.12261319e+00
-4.85429615e-01 1.79303624e-02 -1.92136228e-01 4.95385081e-01
-3.21877480e-01 -1.50162905e-01 -1.24375641e+00 -3.66784967e-02
-5.14917016e-01 -9.24971640e-01 1.75115526e+00 4.73260880e-02
1.30138993e+00 -8.61224905e-02 -2.17492685e-01 5.74517250e-01
1.21179461e+00 -5.35390198e-01 3.98656577e-01 4.74275798e-01
7.03966439e-01 6.27383709e-01 4.24815714e-01 1.28894687e-01
6.37383282e-01 7.24541068e-01 3.99025261e-01 6.04713783e-02
-3.90309952e-02 -5.47189593e-01 5.68835855e-01 1.07891369e+00
-2.95478225e-01 -1.30766362e-01 -1.04771650e+00 6.00292683e-01
-2.15794110e+00 -8.59367907e-01 -1.12977259e-01 2.42319393e+00
1.36299312e+00 5.41356206e-01 1.45708188e-01 -6.87769055e-02
4.79040980e-01 2.89197445e-01 -5.24864852e-01 -1.01365343e-01
-1.80656284e-01 3.12423259e-01 3.98034751e-01 3.86155307e-01
-1.12938356e+00 1.02376878e+00 6.49435329e+00 8.68527055e-01
-6.84978247e-01 3.44963789e-01 8.29455912e-01 -3.74397516e-01
-5.75277090e-01 -1.54651895e-01 -9.89284813e-01 4.59568352e-01
1.26858139e+00 -8.56178254e-02 4.90824014e-01 7.36253977e-01
-4.19428617e-01 2.07811400e-01 -1.47041047e+00 6.35313809e-01
9.18330327e-02 -1.13692939e+00 6.51754811e-02 -1.05680317e-01
1.05465925e+00 1.07953441e+00 1.38666391e-01 9.44386065e-01
1.19089377e+00 -1.15228653e+00 5.02275944e-01 4.91963923e-01
8.74683797e-01 -4.78252649e-01 9.03215706e-01 4.76502419e-01
-8.10287952e-01 -1.34667978e-01 -4.33973193e-01 -6.66411370e-02
-9.13892686e-02 9.74499941e-01 -3.95830065e-01 8.00653756e-01
9.57818031e-01 1.16414785e+00 -8.71117175e-01 7.94909418e-01
-8.25660467e-01 1.16326523e+00 -3.44280720e-01 -1.48061424e-01
1.37288928e-01 7.78800920e-02 2.47692212e-01 1.23333740e+00
-2.31772825e-01 -4.69447114e-02 2.80856252e-01 5.59996247e-01
-7.13340282e-01 -4.54440676e-02 -5.39593756e-01 1.41816229e-01
3.15236241e-01 1.04137313e+00 -2.17112914e-01 -4.61930931e-01
-3.34625661e-01 7.52159655e-01 1.02677476e+00 4.21764523e-01
-6.42916381e-01 -3.23085040e-01 2.22864926e-01 -2.57504862e-02
1.20079763e-01 1.09156273e-01 -2.94385254e-01 -1.45704436e+00
-1.13401543e-02 -6.17757559e-01 8.12129319e-01 -6.46758199e-01
-1.82115531e+00 6.29573047e-01 1.35465413e-01 -1.01441264e+00
-2.34954327e-01 -7.30118275e-01 -6.46804392e-01 5.44571877e-01
-2.08815861e+00 -1.24213517e+00 3.87593210e-02 5.74869156e-01
2.07518905e-01 -2.71285474e-01 8.15165639e-01 3.78677130e-01
-5.07049263e-01 8.40834856e-01 5.75024962e-01 3.11761677e-01
1.05482411e+00 -1.67858922e+00 2.47748464e-01 4.48463142e-01
6.21414065e-01 9.21732426e-01 3.53520781e-01 -2.94295192e-01
-1.11488402e+00 -9.07253861e-01 1.08388615e+00 -9.65780318e-01
8.00953329e-01 -4.13792044e-01 -6.69862568e-01 7.34313071e-01
1.49118707e-01 3.49889934e-01 8.51903141e-01 9.91760075e-01
-8.71055067e-01 -4.06944990e-01 -7.79301226e-01 3.70403707e-01
1.43122911e+00 -6.17002547e-01 -1.04228842e+00 5.38953185e-01
1.18931949e+00 -1.22556321e-01 -8.38369787e-01 7.49497533e-01
5.53441346e-01 -8.64503145e-01 1.18069363e+00 -1.05923891e+00
1.08206820e+00 -5.08148670e-02 -3.52389216e-01 -1.41817939e+00
-3.43197405e-01 -5.91670871e-01 -3.51399273e-01 1.31895995e+00
8.25601995e-01 -5.00186265e-01 5.11413574e-01 4.89174336e-01
-2.43129954e-01 -7.64277399e-01 -7.71277606e-01 -8.95207703e-01
3.45798492e-01 -5.39688587e-01 4.63202566e-01 8.92010510e-01
2.07292974e-01 7.73107350e-01 -3.37474048e-01 -2.27361009e-01
6.39783502e-01 2.13347338e-02 5.59263229e-01 -1.62248087e+00
-5.85287333e-01 -5.36243916e-01 -1.57982260e-01 -1.24514687e+00
5.86353600e-01 -1.53469622e+00 7.18582720e-02 -1.67962718e+00
5.82289875e-01 -4.23412651e-01 -1.45582044e+00 8.14571023e-01
-4.03721064e-01 3.65387708e-01 8.45241845e-02 4.87760216e-01
-1.43488765e+00 8.01172495e-01 9.29107785e-01 -1.94518685e-01
1.24397911e-01 7.99287558e-02 -9.53382909e-01 5.98832607e-01
4.27849382e-01 -5.62912166e-01 -4.60227311e-01 -1.18433917e+00
6.31209195e-01 -2.64402390e-01 1.63358152e-01 -8.85759771e-01
3.42772096e-01 1.33299664e-01 1.42546728e-01 -4.80861396e-01
1.36315435e-01 -4.76250768e-01 -5.26226997e-01 -4.00475189e-02
-1.08142984e+00 -5.65496273e-02 -2.54337490e-03 8.74725759e-01
-2.65906930e-01 -2.70580858e-01 3.06357533e-01 3.89061612e-03
-4.66151863e-01 2.32603133e-01 3.20423126e-01 5.64243674e-01
2.40460277e-01 4.61458832e-01 -9.22440708e-01 -1.10027641e-01
-4.52275813e-01 6.38912559e-01 1.83708355e-01 5.23318887e-01
2.41019145e-01 -1.41173851e+00 -9.78912532e-01 -3.00488919e-01
4.72768813e-01 1.00424238e-01 -2.06073895e-01 6.64070487e-01
1.88052468e-02 6.17909968e-01 1.50713667e-01 -4.44162726e-01
-8.40066016e-01 2.86745399e-01 3.85264158e-02 -1.04376984e+00
-3.81672502e-01 1.02611995e+00 4.20051366e-02 -1.08362174e+00
4.69649106e-01 -1.93216786e-01 -5.15518546e-01 -4.49904129e-02
5.73843658e-01 3.26305628e-01 2.94047385e-01 -7.97065869e-02
-3.34848017e-01 2.07417592e-01 -5.04514992e-01 -2.64245003e-01
1.79678047e+00 3.77577275e-01 -2.18514297e-02 5.95007360e-01
1.22854984e+00 -6.06662668e-02 -1.06045425e+00 -8.24261665e-01
5.84729493e-01 -2.24984422e-01 2.26025745e-01 -1.47638571e+00
-8.60438287e-01 9.48558867e-01 2.65820865e-02 -1.57500759e-01
8.32438052e-01 1.71532288e-01 5.34307897e-01 1.00077450e+00
6.42279506e-01 -1.09271741e+00 3.34971100e-01 9.83720362e-01
8.90183985e-01 -1.41851509e+00 -7.26172701e-03 1.00647658e-01
-7.33346164e-01 7.27120936e-01 7.23935246e-01 -1.94740370e-01
7.46074975e-01 -1.03685886e-01 -1.59641996e-01 -2.94924766e-01
-1.30921626e+00 -3.82376403e-01 8.56400430e-01 2.82426894e-01
7.55645335e-01 1.49833634e-02 -3.34349334e-01 1.06352890e+00
-1.60390437e-01 1.59461603e-01 -4.58241440e-03 6.59005165e-01
-3.71336669e-01 -1.27906787e+00 3.32452983e-01 7.66995370e-01
-6.83037937e-01 -6.57456398e-01 -5.75605989e-01 3.78133148e-01
-3.21773261e-01 1.01078355e+00 -3.60777408e-01 -6.23304784e-01
3.40171576e-01 3.28433476e-02 2.11423248e-01 -9.48819995e-01
-8.80623519e-01 -3.54938179e-01 3.55932564e-01 -5.80269694e-01
-4.59373415e-01 -3.39493454e-01 -1.12699795e+00 1.39120996e-01
-4.42636400e-01 1.85386404e-01 4.64040577e-01 9.29461539e-01
4.06908005e-01 4.35385257e-01 6.01161480e-01 -6.58809304e-01
-9.65205312e-01 -1.19334078e+00 -4.44709837e-01 5.90909302e-01
8.88810083e-02 -7.24191606e-01 -5.16799927e-01 -3.81514668e-01] | [11.411267280578613, 7.530455589294434] |
1bf987c8-cd98-480a-82aa-7a7eef1cf225 | diffgrad-an-optimization-method-for | 1909.11015 | null | https://arxiv.org/abs/1909.11015v4 | https://arxiv.org/pdf/1909.11015v4.pdf | diffGrad: An Optimization Method for Convolutional Neural Networks | Stochastic Gradient Decent (SGD) is one of the core techniques behind the success of deep neural networks. The gradient provides information on the direction in which a function has the steepest rate of change. The main problem with basic SGD is to change by equal sized steps for all parameters, irrespective of gradient behavior. Hence, an efficient way of deep network optimization is to make adaptive step sizes for each parameter. Recently, several attempts have been made to improve gradient descent methods such as AdaGrad, AdaDelta, RMSProp and Adam. These methods rely on the square roots of exponential moving averages of squared past gradients. Thus, these methods do not take advantage of local change in gradients. In this paper, a novel optimizer is proposed based on the difference between the present and the immediate past gradient (i.e., diffGrad). In the proposed diffGrad optimization technique, the step size is adjusted for each parameter in such a way that it should have a larger step size for faster gradient changing parameters and a lower step size for lower gradient changing parameters. The convergence analysis is done using the regret bound approach of online learning framework. Rigorous analysis is made in this paper over three synthetic complex non-convex functions. The image categorization experiments are also conducted over the CIFAR10 and CIFAR100 datasets to observe the performance of diffGrad with respect to the state-of-the-art optimizers such as SGDM, AdaGrad, AdaDelta, RMSProp, AMSGrad, and Adam. The residual unit (ResNet) based Convolutional Neural Networks (CNN) architecture is used in the experiments. The experiments show that diffGrad outperforms other optimizers. Also, we show that diffGrad performs uniformly well for training CNN using different activation functions. The source code is made publicly available at https://github.com/shivram1987/diffGrad. | ['Bidyut Baran Chaudhuri', 'Swalpa Kumar Roy', 'Snehasis Mukherjee', 'Soumendu Chakraborty', 'Satish Kumar Singh', 'Shiv Ram Dubey'] | 2019-09-12 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [-5.88741243e-01 -3.10973525e-01 1.45051211e-01 -6.41427100e-01
-2.28671581e-01 -4.49720562e-01 3.71422708e-01 1.34784114e-02
-9.79459107e-01 1.03102160e+00 -3.06055814e-01 -3.59038264e-01
2.15950739e-02 -7.05035329e-01 -7.47891963e-01 -9.76826191e-01
-5.45249842e-02 4.71101478e-02 1.29602551e-01 -4.03361768e-01
3.30259681e-01 5.48231781e-01 -1.17789209e+00 -2.18625709e-01
8.21949899e-01 1.29325676e+00 1.40615283e-02 7.18759596e-01
-1.37095034e-01 5.99843025e-01 -4.82143462e-01 -3.93490046e-01
7.31763244e-01 -5.50194263e-01 -4.78205830e-01 -2.70452410e-01
2.45276928e-01 -3.29307169e-01 -3.33020777e-01 1.23242569e+00
6.61075115e-01 3.47393602e-01 5.65219760e-01 -9.78957951e-01
-4.67251867e-01 4.91145164e-01 -5.66988826e-01 5.29535770e-01
-1.85625732e-01 7.78896585e-02 8.77109051e-01 -1.00260019e+00
3.26443642e-01 1.05837703e+00 7.64248490e-01 5.06670177e-01
-7.54284859e-01 -6.02374852e-01 4.04208809e-01 3.01575989e-01
-1.26581860e+00 -1.43346652e-01 7.25924253e-01 -1.96895525e-01
7.69003630e-01 1.54168919e-01 6.05058551e-01 5.71412265e-01
5.28749406e-01 6.62104785e-01 1.09875691e+00 -4.08076376e-01
5.13821125e-01 3.45531225e-01 2.88592368e-01 8.96205246e-01
1.90369368e-01 -7.17001781e-02 -2.40025952e-01 -5.96396998e-02
6.27631545e-01 1.16594538e-01 -3.67100716e-01 -5.89491092e-02
-7.06880331e-01 9.22875762e-01 6.61634982e-01 3.96655709e-01
-4.58708167e-01 2.24382177e-01 5.03455520e-01 5.37326872e-01
7.43623853e-01 3.52949977e-01 -6.41665816e-01 -3.21763873e-01
-8.18720639e-01 2.77127713e-01 8.61009181e-01 4.43509758e-01
7.39859760e-01 2.95355380e-01 -1.07979842e-01 9.65198278e-01
2.99085528e-01 4.03911650e-01 1.08220029e+00 -6.51751697e-01
4.79465663e-01 5.56798041e-01 4.08995524e-02 -1.26404214e+00
-3.73491973e-01 -6.65209174e-01 -8.78855646e-01 7.46966660e-01
5.14739215e-01 -6.15634680e-01 -9.31591272e-01 1.64173198e+00
4.68233377e-01 -2.57064383e-02 2.46433541e-02 1.03097224e+00
5.62801838e-01 7.32506454e-01 -2.50632793e-01 -1.54717743e-01
8.34856093e-01 -1.02096856e+00 -5.34563899e-01 -7.20958859e-02
7.11351633e-01 -5.99398971e-01 9.92151797e-01 3.18524867e-01
-9.72426355e-01 -5.19823849e-01 -1.31290531e+00 2.42975265e-01
-6.08093679e-01 4.01835591e-01 4.86509681e-01 6.42998874e-01
-1.05184877e+00 1.15484846e+00 -1.02152085e+00 7.96725154e-02
4.13908422e-01 3.45675021e-01 -2.12608531e-01 3.65059853e-01
-1.06641400e+00 8.23370278e-01 6.09991908e-01 5.51193893e-01
-6.68736577e-01 -7.01022029e-01 -4.89771098e-01 -2.15588920e-02
-1.02272972e-01 -3.06726217e-01 1.18056190e+00 -1.50807023e+00
-1.98650885e+00 7.19741523e-01 9.03573558e-02 -7.80853510e-01
1.06134820e+00 -2.99784690e-01 1.32268533e-01 -1.49969742e-01
-4.86277819e-01 3.94775063e-01 9.47147548e-01 -6.68167233e-01
-6.21519327e-01 -3.75046462e-01 -1.45867243e-02 3.91575158e-01
-5.53748846e-01 -1.14112154e-01 8.53291377e-02 -4.02497411e-01
1.30781800e-01 -9.29471731e-01 -9.26268175e-02 9.54592153e-02
-2.75358588e-01 -1.54309884e-01 6.77244425e-01 -6.42572105e-01
1.16365027e+00 -2.07593513e+00 -5.45444824e-02 1.53221607e-01
9.65560749e-02 5.18392146e-01 2.30438948e-01 1.21087626e-01
-8.96215141e-02 8.76446143e-02 -3.59984845e-01 -2.06928119e-01
2.15988290e-02 1.45753426e-02 -1.26399145e-01 6.60009682e-01
-9.01302323e-02 6.05162680e-01 -7.16645420e-01 -1.05522074e-01
3.72272287e-03 4.30112034e-01 -4.38983917e-01 6.80924132e-02
1.12080753e-01 7.94987530e-02 -4.35134321e-01 3.24135154e-01
9.05923009e-01 -7.74114430e-02 -9.68719348e-02 -2.36012831e-01
-4.03916001e-01 1.69409122e-02 -1.38020539e+00 1.17482829e+00
-4.93992448e-01 7.96524644e-01 -4.03122157e-02 -1.20501399e+00
1.10899186e+00 2.33336836e-01 2.27973923e-01 -4.58332598e-01
3.77389371e-01 5.18198252e-01 -2.48182956e-02 -1.80213541e-01
1.64550483e-01 8.14354233e-03 5.27081609e-01 2.42922768e-01
-8.09615552e-02 7.22689927e-02 3.22110146e-01 -8.45983475e-02
9.01463449e-01 -8.21924210e-02 2.46996596e-01 -3.18045735e-01
8.41153979e-01 -1.10935815e-01 6.24152064e-01 7.16447651e-01
-3.25309753e-01 5.82390249e-01 3.03533524e-01 -7.27895737e-01
-1.01598930e+00 -7.24337935e-01 -1.33188605e-01 9.69040334e-01
-4.98189852e-02 1.37816414e-01 -8.34571183e-01 -7.43665874e-01
-3.11064776e-02 4.88118380e-01 -5.11987746e-01 -2.28483289e-01
-6.53957367e-01 -1.20693123e+00 4.57602799e-01 3.13594550e-01
1.23105550e+00 -1.00525475e+00 -6.07810140e-01 2.56056488e-01
2.64597714e-01 -5.43398261e-01 -4.96440411e-01 1.65533215e-01
-1.20682657e+00 -9.37506557e-01 -1.03944516e+00 -7.42410839e-01
8.03686738e-01 -2.54248250e-02 7.34136403e-01 7.41332248e-02
-1.03674591e-01 1.17795691e-01 -3.21270734e-01 -4.72141504e-01
-1.71932131e-01 1.69421986e-01 1.46057650e-01 2.03397900e-01
7.48885272e-04 -6.52293622e-01 -8.64844322e-01 2.19982505e-01
-7.21344769e-01 -1.42808944e-01 5.54352820e-01 9.73382533e-01
6.16798937e-01 9.31603909e-02 4.19291526e-01 -8.66677821e-01
9.82805729e-01 -3.91294897e-01 -8.71742547e-01 1.70803621e-01
-1.01228034e+00 2.69373626e-01 1.07094371e+00 -5.56672931e-01
-1.09156358e+00 -1.86860412e-01 -2.94136763e-01 -3.92521173e-01
2.52335846e-01 4.41851467e-01 2.65325278e-01 -3.98294836e-01
6.93804562e-01 3.23106229e-01 -1.96703319e-02 -4.31777388e-01
1.05507880e-01 5.29462039e-01 9.47077423e-02 -3.31450343e-01
6.51847184e-01 3.93904001e-01 -1.25155956e-01 -5.41261137e-01
-7.49561131e-01 -1.83807179e-01 -2.31065959e-01 -2.56384254e-01
5.81384480e-01 -5.90402365e-01 -6.85930133e-01 1.11786819e+00
-9.34638739e-01 -4.25726652e-01 -1.36833593e-01 6.66560948e-01
-2.44582027e-01 4.49157581e-02 -6.80632412e-01 -7.85526037e-01
-8.85064125e-01 -9.62210238e-01 2.57179886e-01 7.67416239e-01
3.23768288e-01 -1.33873582e+00 2.00092178e-02 -1.75601408e-01
6.80584073e-01 3.48884761e-01 4.94307905e-01 -8.15157533e-01
-7.79830813e-02 -3.24353665e-01 -4.72217761e-02 1.00234878e+00
1.38272077e-01 1.73534289e-01 -7.33905792e-01 -4.99831319e-01
3.99909645e-01 -2.36929953e-01 7.54540324e-01 4.42564726e-01
9.71507728e-01 -6.12309933e-01 4.78433371e-02 9.34162676e-01
1.63503134e+00 5.24547935e-01 4.96790498e-01 6.79631889e-01
5.53254962e-01 -8.86674691e-03 5.08334816e-01 4.00933295e-01
-2.33718660e-02 2.27512613e-01 4.22290117e-01 7.93101615e-04
1.75602973e-01 -5.46509735e-02 4.23274726e-01 8.51450861e-01
-1.07335106e-01 -2.75542922e-02 -7.59711325e-01 3.06358963e-01
-1.86908793e+00 -8.03076744e-01 4.44160216e-02 2.25601006e+00
9.45516407e-01 3.61813515e-01 -1.31928146e-01 2.08480880e-01
6.50640249e-01 1.28184125e-01 -8.21172893e-01 -7.95080304e-01
-4.02165204e-02 1.44123495e-01 7.87150681e-01 6.00409508e-01
-1.03096056e+00 7.72013128e-01 5.20333719e+00 9.01209652e-01
-1.55289519e+00 3.63868214e-02 9.49561596e-01 -5.21240979e-02
2.27906212e-01 -2.08373800e-01 -1.00768769e+00 7.05210149e-01
7.06875384e-01 -1.46707192e-01 5.79685986e-01 1.21692562e+00
3.07406723e-01 -1.13450065e-01 -7.49623775e-01 1.10008895e+00
-2.29420915e-01 -1.19846797e+00 -2.20151171e-01 -2.63873935e-01
8.37891757e-01 2.59563386e-01 9.13443565e-02 3.37557822e-01
1.63281709e-01 -7.64224052e-01 7.18008220e-01 4.51452911e-01
1.29547566e-01 -8.27947259e-01 1.03380692e+00 3.38264674e-01
-8.12581778e-01 -1.59402952e-01 -6.74943507e-01 -1.41494274e-01
-1.81618735e-01 9.61015701e-01 -8.77875209e-01 2.47359946e-01
6.95341468e-01 5.29945910e-01 -2.89271057e-01 1.22094572e+00
-3.68384540e-01 8.51152718e-01 -4.91691917e-01 -5.83418548e-01
4.58952010e-01 -5.92961252e-01 6.69837058e-01 1.21101749e+00
2.11708069e-01 -6.64192140e-02 -2.43756950e-01 7.42706537e-01
-2.29816481e-01 3.91407728e-01 -1.45596340e-01 1.41546041e-01
3.91822964e-01 1.28817904e+00 -6.17151141e-01 -2.35686556e-01
-1.33400259e-03 9.26113188e-01 4.19849306e-01 3.86803329e-01
-7.89305747e-01 -8.19517434e-01 6.86687291e-01 -1.94410533e-01
2.99042463e-01 -2.87606061e-01 -3.09546918e-01 -9.17200148e-01
3.69378984e-01 -6.99564517e-01 3.53892416e-01 -4.08168733e-01
-1.07644808e+00 7.05574036e-01 -2.17688575e-01 -9.85978603e-01
-2.14941233e-01 -8.04305613e-01 -8.79272282e-01 9.23795938e-01
-1.66457629e+00 -3.29973608e-01 -3.39228123e-01 5.36779583e-01
5.91547906e-01 -1.70807630e-01 3.58370334e-01 2.82839626e-01
-7.13433027e-01 8.49338472e-01 7.26360440e-01 2.67512709e-01
2.89063275e-01 -1.17240667e+00 1.37482256e-01 7.62358069e-01
-1.45798787e-01 4.39433426e-01 8.50535452e-01 -2.41197318e-01
-1.15619862e+00 -8.60852361e-01 5.56096733e-01 3.05019438e-01
5.88824749e-01 -1.43427193e-01 -9.28960383e-01 4.53433931e-01
-1.28218621e-01 3.45243156e-01 2.91439500e-02 -2.07402587e-01
6.65700212e-02 -6.31262004e-01 -1.36084962e+00 4.95742768e-01
6.67038143e-01 -6.79262802e-02 -2.34800041e-01 4.35613722e-01
4.05553997e-01 -5.65684974e-01 -5.71950316e-01 2.11844057e-01
6.05473876e-01 -1.16609430e+00 5.88383257e-01 -5.32268107e-01
2.51206875e-01 -1.94233611e-01 9.99540612e-02 -1.71319151e+00
-3.31309177e-02 -6.19475424e-01 -1.28833205e-01 1.06569207e+00
5.68517029e-01 -1.34736073e+00 7.28772998e-01 6.03919148e-01
-4.37882245e-02 -1.35957873e+00 -9.40798283e-01 -8.26147199e-01
2.01826215e-01 3.08938641e-02 2.45769233e-01 6.37078524e-01
-3.21029782e-01 -1.96851969e-01 -7.18736947e-02 -6.54023141e-02
3.59535933e-01 -1.66303605e-01 5.75365603e-01 -8.02932560e-01
-3.36610407e-01 -5.35141766e-01 -6.36524498e-01 -1.00475872e+00
-1.09539710e-01 -7.30233669e-01 -2.82734670e-02 -1.25408483e+00
-1.55279562e-01 -4.92683232e-01 -5.68152606e-01 4.35881048e-01
-2.80242234e-01 -2.26169918e-02 1.68585122e-01 1.59405902e-01
-2.86063313e-01 7.02511430e-01 1.20373917e+00 -6.13951311e-02
-5.19627035e-01 2.55765021e-01 -3.73144507e-01 8.36462319e-01
1.35194957e+00 -6.49695396e-01 -3.91222209e-01 -4.85885471e-01
1.78556934e-01 -4.63337928e-01 9.80504379e-02 -1.07970667e+00
2.64464647e-01 4.65080477e-02 4.43198860e-01 -1.66056216e-01
1.49692059e-01 -4.18931931e-01 -1.76087931e-01 6.47841990e-01
-3.59355897e-01 4.87413347e-01 1.27633199e-01 3.45188320e-01
-2.65814364e-01 -5.64307451e-01 1.03430653e+00 -2.42983997e-01
-4.10013109e-01 4.17243928e-01 -9.46435034e-02 1.69868946e-01
8.77311826e-01 -1.73540205e-01 -1.93856992e-02 -3.63442600e-01
-5.03732622e-01 1.43071428e-01 4.72028740e-02 1.64087057e-01
6.05128646e-01 -1.18213952e+00 -8.35238874e-01 -2.15690639e-02
-4.99147356e-01 -1.36965916e-01 -2.82237884e-02 9.91561353e-01
-9.80338216e-01 1.96108058e-01 -7.98326731e-02 -2.76773095e-01
-1.16571605e+00 1.16865546e-01 9.24300849e-01 -3.77438128e-01
-5.15242100e-01 1.11273909e+00 -1.21629648e-01 -4.14679348e-01
3.87478322e-01 -3.23447347e-01 -8.61934647e-02 -1.24999583e-01
6.40530705e-01 5.94242275e-01 3.44848484e-01 -8.93752798e-02
-2.29971796e-01 4.47206020e-01 -3.28995526e-01 -1.46236643e-01
1.53636515e+00 3.43063399e-02 -1.01545580e-01 5.41805387e-01
1.56347334e+00 -2.24963084e-01 -1.50999618e+00 -1.19740553e-02
-1.59709901e-01 -3.58672142e-01 3.35460663e-01 -6.93281353e-01
-1.45733070e+00 6.39166713e-01 1.08246136e+00 1.77832216e-01
1.14585602e+00 -7.42533684e-01 7.12827563e-01 5.85301220e-01
2.45062083e-01 -1.34142613e+00 -1.64734691e-01 6.18502855e-01
8.28737617e-01 -1.33716297e+00 -2.36332677e-02 2.78064430e-01
-4.32209134e-01 1.41751790e+00 4.95822042e-01 -5.01359761e-01
8.73984694e-01 9.28460360e-02 2.18220845e-01 7.92107955e-02
-5.14694214e-01 2.29393706e-01 7.68999085e-02 -3.90989743e-02
4.29426819e-01 -8.73892307e-02 -9.00301218e-01 1.40931800e-01
-2.44121447e-01 -5.41093424e-02 3.79250795e-01 9.39170539e-01
-4.79525715e-01 -8.15672159e-01 -2.61074662e-01 3.13468665e-01
-7.86042213e-01 -1.17774844e-01 -1.34159788e-01 6.84448242e-01
-1.09939598e-01 7.18678653e-01 -1.44782126e-01 -1.91693380e-01
1.96675092e-01 2.01205183e-02 3.41062218e-01 -3.63945812e-02
-7.22077429e-01 -6.33697689e-01 -2.73104578e-01 -3.58435571e-01
-2.40596514e-02 -5.46978235e-01 -1.55851817e+00 -3.08992267e-01
-3.66460085e-01 3.50673616e-01 1.24371350e+00 9.02161419e-01
1.18450359e-01 2.72674710e-01 1.05330575e+00 -6.99909687e-01
-1.15571690e+00 -9.21157777e-01 -4.64130044e-01 3.43462914e-01
3.21991891e-01 -4.94088471e-01 -9.23101306e-01 -2.43402317e-01] | [7.7345685958862305, 3.634995222091675] |
4fe4c629-6d43-4406-8b83-131b5f5427bb | deep-movement-primitives-toward-breast-cancer | 2202.09265 | null | https://arxiv.org/abs/2202.09265v1 | https://arxiv.org/pdf/2202.09265v1.pdf | Deep Movement Primitives: toward Breast Cancer Examination Robot | Breast cancer is the most common type of cancer worldwide. A robotic system performing autonomous breast palpation can make a significant impact on the related health sector worldwide. However, robot programming for breast palpating with different geometries is very complex and unsolved. Robot learning from demonstrations (LfD) reduces the programming time and cost. However, the available LfD are lacking the modelling of the manipulation path/trajectory as an explicit function of the visual sensory information. This paper presents a novel approach to manipulation path/trajectory planning called deep Movement Primitives that successfully generates the movements of a manipulator to reach a breast phantom and perform the palpation. We show the effectiveness of our approach by a series of real-robot experiments of reaching and palpating a breast phantom. The experimental results indicate our approach outperforms the state-of-the-art method. | ['Amir M. Ghalamzan E.', 'Kiyanoush Nazari', 'Pablo C. Lopez-Custodio', 'Muhammad Arshad Khan', 'Giorgio Bonvicini', 'Oluwatoyin Sanni'] | 2022-02-14 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [ 6.34688661e-02 4.95256603e-01 -2.64835775e-01 -2.43455067e-01
-4.59321707e-01 -4.16889995e-01 1.74074635e-01 1.82411686e-01
-2.23439083e-01 4.86333728e-01 -5.23980021e-01 -5.87315500e-01
-4.70926642e-01 -4.67159420e-01 -1.06473482e+00 -6.56441748e-01
-2.86135286e-01 7.84885228e-01 4.24337611e-02 -2.56773263e-01
1.51319221e-01 1.06498027e+00 -7.29875267e-01 -1.15581900e-01
6.85447454e-01 6.48753166e-01 9.24103796e-01 8.69292319e-01
1.34266168e-01 4.89085257e-01 -1.76040903e-01 1.52256489e-01
2.45482773e-01 -2.02053502e-01 -7.02045858e-01 2.63712909e-02
-2.68917024e-01 -4.04618114e-01 -2.23027781e-01 9.70119953e-01
5.58120847e-01 -4.07572001e-01 6.82987392e-01 -1.49628866e+00
-1.51675567e-01 5.11087716e-01 -5.81539392e-01 -5.47292769e-01
2.82820582e-01 2.31982097e-01 4.75933142e-02 -5.33313036e-01
1.12932456e+00 1.45251179e+00 8.38507116e-01 6.14047706e-01
-1.21009731e+00 -4.64799315e-01 -4.38738883e-01 1.20325819e-01
-9.54629600e-01 4.76605073e-02 8.04658413e-01 -3.44678909e-01
3.18242639e-01 3.13979894e-01 7.94668436e-01 9.94604230e-01
9.18332815e-01 7.33756125e-01 1.05413985e+00 -4.22196656e-01
2.35442504e-01 5.34563921e-02 -7.16487169e-02 1.08950603e+00
2.20381021e-01 2.82629073e-01 1.85917020e-01 -1.43405572e-01
1.35738051e+00 -1.55670255e-01 -1.85167551e-01 -9.40076649e-01
-1.40597737e+00 7.86055982e-01 7.95445740e-01 2.23935708e-01
-6.49866104e-01 6.62441194e-01 2.56907433e-01 2.04741061e-01
-5.85348547e-01 5.46071410e-01 -2.64341474e-01 -9.05383825e-02
-3.40772867e-01 4.52714652e-01 1.19128609e+00 1.37935209e+00
2.27935135e-01 -4.03950721e-01 -2.17136458e-01 3.46078128e-01
6.36774838e-01 6.47538126e-01 2.63238978e-02 -1.19645214e+00
1.83578730e-01 4.59010929e-01 4.32532668e-01 -1.26153302e+00
-1.15490162e+00 7.53647089e-02 -8.34444463e-01 4.04488772e-01
7.08103180e-01 -1.46245793e-01 -9.28489268e-01 1.09259975e+00
6.26203597e-01 -5.41045964e-01 9.67121720e-02 1.09245050e+00
9.37160850e-01 5.08133769e-01 5.72769456e-02 -6.73264340e-02
1.14023352e+00 -1.00310051e+00 -7.91562140e-01 4.08302667e-03
6.03109121e-01 -4.81895626e-01 7.57580638e-01 5.33668578e-01
-9.41690326e-01 -5.63252032e-01 -7.66451240e-01 3.50096337e-02
1.18665360e-01 4.56073374e-01 1.00954294e+00 3.74817252e-01
-7.89165974e-01 7.48421609e-01 -1.47147667e+00 -6.66307867e-01
4.25312281e-01 4.70398188e-01 -3.58083367e-01 -2.92168677e-01
-6.01275444e-01 1.21441472e+00 3.89590889e-01 3.31588984e-01
-1.26279819e+00 -7.68399596e-01 -7.90817738e-01 -3.36659849e-01
2.17349574e-01 -3.58483702e-01 1.54467702e+00 -3.30876678e-01
-1.82682848e+00 6.90126657e-01 3.04410607e-01 -2.14232177e-01
9.72547174e-01 -1.04593569e-02 3.02589417e-01 3.51552129e-01
-6.75545633e-02 1.10773039e+00 4.96561140e-01 -1.44438720e+00
-1.69054329e-01 -2.46946305e-01 -1.19130693e-01 2.65007436e-01
1.64549068e-01 -2.40664333e-01 -5.32425225e-01 -1.81756079e-01
4.31615740e-01 -1.26366723e+00 -8.56533051e-01 1.10392523e+00
-6.10611677e-01 -8.34535882e-02 7.49615371e-01 -7.51022220e-01
3.02661061e-01 -2.17848277e+00 4.52048570e-01 8.87969956e-02
-1.65545598e-01 -2.36979395e-01 -1.10497773e-01 6.00165844e-01
3.04786175e-01 -1.33212343e-01 -5.12885973e-02 1.56483695e-01
4.14801063e-04 5.15117049e-01 6.09388016e-02 7.00121522e-01
-1.06928423e-01 9.01907027e-01 -1.33758724e+00 -8.44157755e-01
2.18070269e-01 2.64669508e-01 -1.90441787e-01 7.01291263e-02
-4.75231320e-01 9.56327200e-01 -9.19110000e-01 8.97220254e-01
8.16032887e-01 1.01610534e-01 3.33794117e-01 -2.01665983e-01
-7.74372965e-02 -4.98340666e-01 -8.29962194e-01 2.18756962e+00
-2.48584077e-01 3.18645984e-01 5.48028827e-01 -9.33402181e-01
1.07576919e+00 1.68688446e-01 6.63254082e-01 -3.68784696e-01
2.50807911e-01 5.14269829e-01 3.29435915e-01 -1.13813639e+00
1.77619144e-01 -2.31528981e-03 -1.98640808e-01 1.10303380e-01
-3.95307094e-01 -1.04590070e+00 -6.97298795e-02 -6.30510822e-02
1.20146930e+00 6.00017488e-01 3.44130695e-01 -3.00751388e-01
1.26703456e-01 8.56832206e-01 2.36099914e-01 7.44346261e-01
-2.16203451e-01 1.50637537e-01 6.45382404e-01 -2.42905021e-01
-1.01542425e+00 -9.81959403e-01 -2.39939079e-01 3.69583488e-01
6.28660500e-01 3.52480024e-01 -5.45870781e-01 -4.08236712e-01
3.76878828e-01 6.84295297e-01 -3.91104251e-01 1.42760456e-01
-7.78419733e-01 -4.05453116e-01 5.56908071e-01 4.64860350e-01
3.17605495e-01 -1.50602019e+00 -1.28337181e+00 2.39119485e-01
7.93353766e-02 -1.00826561e+00 1.66310802e-01 1.74357176e-01
-1.06680572e+00 -1.23935735e+00 -9.04064298e-01 -1.18961787e+00
1.06390035e+00 -4.28339913e-02 4.87448335e-01 -2.29996741e-01
-1.06708908e+00 4.89341140e-01 -2.53930211e-01 -7.46650219e-01
-1.17505956e+00 -2.16366485e-01 -4.45170611e-01 -8.56472850e-01
-4.49228108e-01 -2.87335753e-01 -4.61100549e-01 4.94653970e-01
-5.67984879e-01 4.00798291e-01 1.05872142e+00 6.64726913e-01
6.79041266e-01 1.24775924e-01 2.24239469e-01 -4.35539484e-01
7.04068899e-01 -3.76976013e-01 -7.31874883e-01 7.91290551e-02
6.51717484e-02 -1.91069581e-02 3.04698884e-01 -8.83042932e-01
-8.00218463e-01 7.97325909e-01 2.23385379e-01 -3.54068428e-01
-1.72336429e-01 5.75664699e-01 1.50317565e-01 -4.38215137e-01
7.42930830e-01 5.26465848e-02 5.06958127e-01 -4.19869155e-01
3.52084905e-01 4.27467346e-01 7.28245676e-01 -5.53398311e-01
5.60797513e-01 5.46320081e-01 7.73428261e-01 -1.00898266e+00
1.09085329e-01 -1.74285859e-01 -7.46486306e-01 -4.62587178e-01
7.19604611e-01 -2.07539603e-01 -1.27644479e+00 4.45683032e-01
-1.51307619e+00 -7.73893535e-01 -1.27588451e-01 5.31608522e-01
-9.23386872e-01 3.89044076e-01 -7.08442390e-01 -6.66266501e-01
-2.48813510e-01 -1.36581993e+00 1.03586769e+00 2.05117270e-01
-1.57098219e-01 -6.11043692e-01 -1.38262331e-01 -2.43902266e-01
4.40315604e-01 1.04515982e+00 1.05701375e+00 5.32750972e-03
-6.64874196e-01 -4.02490109e-01 -3.12391937e-01 -3.88756931e-01
-4.43705507e-02 -2.51134038e-01 -3.10514629e-01 -3.06577653e-01
1.10215724e-01 -5.08182645e-01 1.06569774e-01 6.42749846e-01
1.34069943e+00 -3.94899398e-02 -1.12174356e+00 3.33262295e-01
1.20338893e+00 3.75818700e-01 4.84495074e-01 2.15318307e-01
5.36184311e-01 7.45874286e-01 1.28149021e+00 2.97663540e-01
8.98589715e-02 6.93761528e-01 9.22772765e-01 -1.77786108e-02
6.52450621e-02 -1.86514050e-01 9.92573649e-02 1.80520236e-01
1.22421971e-02 2.20559448e-01 -1.19317412e+00 3.86627138e-01
-1.87462926e+00 -3.83070230e-01 -5.05662382e-01 1.81798089e+00
8.22122037e-01 -2.52052277e-01 -3.55008245e-01 9.20444578e-02
4.67458457e-01 -7.15324521e-01 -8.25243533e-01 -3.77153963e-01
7.52202928e-01 -1.88865345e-02 8.41759682e-01 3.56442004e-01
-8.75479639e-01 6.53095961e-01 5.95101738e+00 5.66058755e-01
-1.20139384e+00 -8.03524181e-02 1.98943377e-01 6.18279517e-01
1.42410249e-01 -1.93097576e-01 -4.84453470e-01 -3.57658528e-02
4.06473368e-01 1.13646634e-01 2.87058711e-01 1.07506490e+00
3.18174928e-01 -5.54409862e-01 -1.51015401e+00 1.02147973e+00
-2.62068838e-01 -9.57572103e-01 -5.58595300e-01 -5.95461391e-03
2.54671007e-01 -2.90760666e-01 -3.25000256e-01 3.31126273e-01
1.48233175e-01 -1.11698639e+00 8.20641696e-01 7.91057527e-01
7.72092402e-01 -4.20777947e-01 3.19107115e-01 1.00715554e+00
-7.03258693e-01 -2.14170292e-01 -4.84374732e-01 2.69240826e-01
3.51095527e-01 2.53144473e-01 -1.63580596e+00 5.01278043e-01
5.34947038e-01 2.18286067e-01 -2.63284326e-01 1.36924279e+00
-1.58961028e-01 -4.78354581e-02 -4.68653440e-01 -6.90013826e-01
1.87736422e-01 6.61646947e-02 7.75560617e-01 9.47409570e-01
4.61869150e-01 -2.60741096e-02 4.93890673e-01 1.11278415e+00
3.37538362e-01 4.55845967e-02 -6.92434371e-01 -1.57392710e-01
2.18883291e-01 1.44496477e+00 -8.94229412e-01 2.78063148e-01
3.26363802e-01 9.66870725e-01 -5.24192452e-02 8.45483169e-02
-1.15499556e+00 -4.04187262e-01 -2.68054932e-01 -2.65553361e-03
7.78590445e-04 -4.37311828e-01 -1.15198530e-01 -1.45590916e-01
1.61632448e-01 -5.71505487e-01 -2.54713714e-01 -9.93788481e-01
-7.50971675e-01 1.96263179e-01 4.21099424e-01 -1.22460830e+00
-1.40945762e-01 -9.17554438e-01 -1.89778209e-01 5.88776648e-01
-1.35898709e+00 -1.35340023e+00 -8.34635556e-01 1.43437341e-01
5.40768802e-01 3.49332809e-01 1.06939125e+00 -1.83908537e-01
1.07350759e-02 1.16418004e-02 -4.65783551e-02 -9.34019685e-02
4.20564860e-01 -1.08345330e+00 -1.25163063e-01 -1.45739719e-01
-7.78979778e-01 1.18946731e-01 8.20455313e-01 -8.51091564e-01
-2.34277272e+00 -9.14475322e-01 2.49824300e-01 -4.95257340e-02
5.75237572e-01 -6.80466294e-02 -4.40698743e-01 5.72957039e-01
-1.64004922e-01 1.76869109e-01 -2.22417861e-01 -7.51516581e-01
6.25543952e-01 7.29381144e-02 -1.66747797e+00 6.43312633e-01
7.72878230e-01 3.57083470e-01 -1.64605066e-01 8.02098215e-01
5.38366437e-01 -1.13690543e+00 -8.39204848e-01 6.13057137e-01
7.43160844e-01 -4.00948644e-01 9.61330414e-01 -1.71558902e-01
5.28486192e-01 -7.44631141e-02 2.14429915e-01 -1.21518922e+00
-2.20386758e-02 -7.70204723e-01 6.70746490e-02 5.27276456e-01
1.55396253e-01 -4.04411227e-01 8.35047960e-01 4.26259637e-01
-2.63287306e-01 -9.75318491e-01 -9.38114583e-01 -8.96248341e-01
9.53767970e-02 -1.89734191e-01 1.77139398e-02 5.69650650e-01
3.01176786e-01 -2.95702428e-01 3.35672125e-02 3.55306417e-01
7.10199416e-01 7.61479652e-03 1.03926349e+00 -1.12129426e+00
-3.71444881e-01 -1.50826037e-01 -5.01456857e-01 -9.22167242e-01
-1.35721534e-01 -1.09367669e+00 6.77682519e-01 -1.84576559e+00
1.55315936e-01 -9.95834351e-01 6.46195889e-01 5.41511834e-01
5.05559683e-01 -4.26226884e-01 2.34828249e-01 1.29701734e-01
9.08501372e-02 1.22635223e-01 2.05928636e+00 -2.89860964e-01
-4.60063547e-01 1.26929075e-01 8.29843283e-02 8.51292908e-01
8.37343752e-01 -7.21187890e-01 -1.22772455e-01 -2.06385419e-01
-6.63316995e-02 9.94036436e-01 5.75902522e-01 -8.66137445e-01
4.42484558e-01 -3.28801990e-01 2.91047215e-01 -7.09258914e-01
5.08954346e-01 -1.23966169e+00 3.96159559e-01 1.54761434e+00
-2.60115862e-01 -2.66984671e-01 5.16473651e-01 7.37138927e-01
4.07875061e-01 -5.61225414e-01 7.55527675e-01 -1.99999318e-01
-4.23030436e-01 -1.66350082e-01 -4.18762535e-01 -7.16323137e-01
1.48696768e+00 -3.24908867e-02 8.71911123e-02 -1.83256015e-01
-9.22192276e-01 5.63047767e-01 3.38174164e-01 2.78291017e-01
7.21687913e-01 -1.05828309e+00 -6.33295834e-01 -4.11672592e-02
-1.81496158e-01 4.26734298e-01 -9.00821462e-02 1.13136554e+00
-1.30912387e+00 3.59964132e-01 -4.91277397e-01 -1.04241979e+00
-1.21151233e+00 5.68729579e-01 4.10736412e-01 -4.45180148e-01
-8.29534769e-01 4.23102915e-01 -2.11830720e-01 -8.88859928e-01
5.04254758e-01 -7.07261324e-01 -1.68864746e-02 -6.90750599e-01
-1.24739945e-01 5.66789210e-01 -4.09720331e-01 -2.72817891e-02
-1.09576568e-01 5.55805922e-01 3.31472307e-01 -1.85767487e-01
1.31289399e+00 3.32176208e-01 -2.51523763e-01 2.17223346e-01
6.77044153e-01 -5.77455342e-01 -1.17378080e+00 3.37309092e-01
6.26172870e-02 -2.51119256e-01 -3.18000972e-01 -8.50515485e-01
-6.70948446e-01 7.70239830e-01 6.02576375e-01 -1.51834279e-01
6.87802672e-01 1.18054181e-01 3.85429293e-01 9.62658465e-01
9.34108555e-01 -7.09169447e-01 3.22382987e-01 1.31205902e-01
1.68863797e+00 -1.23120880e+00 9.79897305e-02 -1.07201207e+00
-4.18912262e-01 1.47115469e+00 5.42213857e-01 -2.59286374e-01
6.94016218e-01 5.06432593e-01 3.02888379e-02 -4.08332318e-01
1.31698266e-01 6.28819466e-01 -6.03698529e-02 7.79288948e-01
6.69342130e-02 1.98107928e-01 -6.92449689e-01 3.44284922e-01
-2.57445574e-02 4.83550847e-01 5.17909348e-01 1.51374578e+00
-3.82405668e-01 -8.95743489e-01 -6.81659877e-01 7.32480437e-02
1.30457962e-02 6.79817796e-01 -8.69308934e-02 1.32095635e+00
-2.22109824e-01 5.49247086e-01 -2.65049696e-01 4.18900967e-01
3.55214238e-01 -2.75777221e-01 1.02415884e+00 -4.68141556e-01
-1.70808464e-01 -5.86869046e-02 -8.65589157e-02 -7.61718869e-01
-1.06549092e-01 -4.57637459e-01 -1.89628577e+00 3.43653411e-01
-1.89218685e-01 -2.82663167e-01 1.43826962e+00 3.28890324e-01
1.18470065e-01 5.43554485e-01 3.91538799e-01 -1.64956558e+00
-1.15690219e+00 -1.01552665e+00 -5.22990942e-01 8.73806849e-02
8.04227367e-02 -6.36617720e-01 2.00482294e-01 1.60140898e-02] | [5.895063400268555, -0.7316150069236755] |
67c8bb47-1c6b-4548-886c-8c37b1be25c4 | lambdabeam-neural-program-search-with-higher | 2306.02049 | null | https://arxiv.org/abs/2306.02049v1 | https://arxiv.org/pdf/2306.02049v1.pdf | LambdaBeam: Neural Program Search with Higher-Order Functions and Lambdas | Search is an important technique in program synthesis that allows for adaptive strategies such as focusing on particular search directions based on execution results. Several prior works have demonstrated that neural models are effective at guiding program synthesis searches. However, a common drawback of those approaches is the inability to handle iterative loops, higher-order functions, or lambda functions, thus limiting prior neural searches from synthesizing longer and more general programs. We address this gap by designing a search algorithm called LambdaBeam that can construct arbitrary lambda functions that compose operations within a given DSL. We create semantic vector representations of the execution behavior of the lambda functions and train a neural policy network to choose which lambdas to construct during search, and pass them as arguments to higher-order functions to perform looping computations. Our experiments show that LambdaBeam outperforms neural, symbolic, and LLM-based techniques in an integer list manipulation domain. | ['Charles Sutton', 'Kevin Ellis', 'Wen-Ding Li', 'Hanjun Dai', 'Kensen Shi'] | 2023-06-03 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 8.57765302e-02 1.46620711e-02 -8.42090905e-01 -4.18245465e-01
-3.99088264e-01 -7.06460476e-01 4.52813655e-01 2.10895672e-01
-1.46581873e-01 5.24271190e-01 2.79670507e-01 -1.04987466e+00
-2.84913424e-02 -1.26785564e+00 -7.41225481e-01 -6.44838810e-02
-2.50561625e-01 3.12330216e-01 3.63515705e-01 -3.66708815e-01
6.93016410e-01 8.48646939e-01 -1.67652571e+00 5.51638961e-01
8.58804226e-01 5.93557239e-01 2.49515638e-01 8.78496408e-01
-6.01110041e-01 8.70866776e-01 -8.66290510e-01 -1.13068558e-01
1.21348754e-01 -3.37742716e-01 -7.45956421e-01 -5.92744112e-01
-6.87547773e-02 -3.68985981e-01 -1.05730407e-01 1.01584280e+00
1.71572298e-01 2.48382866e-01 2.61976123e-01 -1.38175857e+00
-4.94863927e-01 1.29906321e+00 1.06471479e-02 1.10538051e-01
3.82100046e-01 4.31538135e-01 1.06900942e+00 -3.78903002e-01
5.44829249e-01 1.48982775e+00 6.99319780e-01 5.92626750e-01
-1.53078568e+00 -6.75982118e-01 8.04251656e-02 -1.58145115e-01
-1.16729546e+00 -3.84254575e-01 6.62566900e-01 -3.27119857e-01
1.60600746e+00 5.88253736e-01 7.16836333e-01 7.08258450e-01
4.08100933e-01 8.17674041e-01 5.94208479e-01 -6.19960666e-01
3.77066612e-01 -2.07177028e-01 1.70808598e-01 9.62481797e-01
2.42945962e-02 4.45215255e-01 -4.38078195e-02 -6.82871699e-01
9.54639196e-01 -3.00148875e-01 -1.69310540e-01 -3.01118970e-01
-1.33160913e+00 1.05214381e+00 5.20422697e-01 3.58645290e-01
-2.62241781e-01 7.71926165e-01 7.81287134e-01 4.01170194e-01
-2.12469161e-01 1.27238631e+00 -6.99652016e-01 -3.92931223e-01
-8.38266790e-01 5.95886469e-01 1.12043440e+00 8.91209662e-01
8.83325815e-01 4.03961450e-01 -4.31855440e-01 6.15849674e-01
1.22917436e-01 1.23688042e-01 4.08950716e-01 -1.16330862e+00
2.26382032e-01 9.65092182e-01 -1.12389401e-01 -9.71332252e-01
-4.70610797e-01 -2.88450807e-01 -2.71802157e-01 5.05830884e-01
1.44515589e-01 -1.56100422e-01 -7.08080113e-01 1.71122539e+00
-1.07539063e-02 -2.91392803e-01 8.64000395e-02 5.40306926e-01
4.92236018e-01 9.81861055e-01 6.98357522e-02 1.32072559e-02
8.90441895e-01 -1.13142443e+00 -3.77242178e-01 -3.09522986e-01
1.17103195e+00 -4.50250298e-01 1.22566998e+00 2.99487352e-01
-1.31948352e+00 -3.74258697e-01 -1.05272746e+00 1.87779516e-01
-3.35934192e-01 9.94389579e-02 1.36041415e+00 5.89963436e-01
-1.24405253e+00 6.59509718e-01 -9.82597589e-01 1.66892573e-01
1.51359048e-02 6.52082503e-01 2.32207954e-01 4.56385583e-01
-9.70565617e-01 9.19533372e-01 8.07719648e-01 -1.82348743e-01
-8.61606121e-01 -7.67697632e-01 -9.10885930e-01 3.74322742e-01
3.20102572e-01 -5.70615113e-01 1.55804980e+00 -1.13285661e+00
-1.47463667e+00 4.31156367e-01 -1.14801183e-01 -5.52986860e-01
-1.78444222e-01 1.67310700e-01 -3.19093227e-01 -4.01440322e-01
-2.50048935e-01 6.58035994e-01 3.88071418e-01 -1.11557317e+00
-6.07989132e-01 -3.77080590e-02 6.39627218e-01 5.07088192e-02
-4.56861965e-02 2.33770847e-01 -4.13014621e-01 -4.49156612e-01
-1.32827178e-01 -9.85810518e-01 -4.33252007e-01 -1.45581260e-01
-3.96826982e-01 -2.87856162e-01 5.39036095e-01 -3.46459955e-01
1.71056581e+00 -2.07690835e+00 3.43171418e-01 3.94642472e-01
3.79581228e-02 1.51755989e-01 -7.03174621e-02 4.64213073e-01
-2.37008017e-02 2.17741981e-01 4.23176326e-02 5.61887562e-01
3.37697566e-01 2.22080722e-01 -4.53635484e-01 -1.48347244e-02
-7.55572617e-02 1.04680455e+00 -9.62140858e-01 -3.29699069e-01
-2.00890042e-02 6.15251958e-02 -1.14882338e+00 2.39636928e-01
-1.07792246e+00 -5.85023314e-02 -6.20776892e-01 7.20841289e-01
1.76260397e-01 -3.19800526e-01 2.13216782e-01 -2.08353788e-01
-3.85950834e-01 6.30014062e-01 -9.19005334e-01 1.51805532e+00
-8.92891347e-01 6.20252848e-01 3.07917595e-02 -8.15769672e-01
1.08525956e+00 -1.36243068e-02 1.09787621e-02 -6.15582645e-01
5.33803888e-02 2.99997032e-01 1.70217901e-01 -2.89374590e-01
5.66767812e-01 4.31452543e-01 -4.01653588e-01 5.27783990e-01
-4.26184058e-01 -4.57695305e-01 4.10310119e-01 -1.46292701e-01
1.31112325e+00 3.25636894e-01 2.83205658e-01 -2.67585456e-01
6.49883986e-01 3.56597185e-01 4.98699993e-01 9.27372098e-01
3.47528964e-01 -1.66769668e-01 9.02820289e-01 -6.52159333e-01
-9.34633553e-01 -7.44200885e-01 2.42176101e-01 1.79394972e+00
-2.33023509e-01 -6.56040668e-01 -8.04398358e-01 -6.11756802e-01
-4.87790704e-02 1.41279685e+00 -3.35043877e-01 -3.91133457e-01
-1.15551591e+00 -4.53192621e-01 9.45468605e-01 6.62085116e-01
2.61506677e-01 -1.29957807e+00 -1.16794038e+00 4.07955766e-01
2.01321632e-01 -3.34401578e-01 -5.49097538e-01 6.21075213e-01
-9.84267473e-01 -8.31031203e-01 -1.07605889e-01 -1.01419973e+00
7.87949145e-01 -5.06579995e-01 1.34460115e+00 3.94568056e-01
-1.74518496e-01 -2.17626039e-02 2.03192890e-01 5.94917648e-02
-1.02140403e+00 3.39094281e-01 -6.38676465e-01 -7.43393958e-01
1.97268844e-01 -3.69147360e-01 -2.99362361e-01 2.19305381e-01
-8.18245530e-01 1.63433492e-01 5.52069843e-01 9.72284973e-01
3.58340502e-01 2.42427304e-01 3.18638325e-01 -8.15467894e-01
1.23214710e+00 -1.97188303e-01 -1.26023316e+00 3.70827764e-01
-6.39602363e-01 6.93829417e-01 7.97449887e-01 -5.85937142e-01
-8.70654225e-01 1.80411279e-01 -2.00680465e-01 -2.36257657e-01
3.04314960e-02 8.26438308e-01 3.96547280e-02 -1.76817328e-01
1.05397463e+00 2.43154317e-01 -1.78176630e-02 -6.35809004e-02
3.61465067e-01 1.72103301e-01 5.57093561e-01 -1.23899055e+00
4.15171087e-01 -1.45746827e-01 -4.41199616e-02 -2.55120575e-01
5.87666370e-02 2.81605273e-01 4.36773598e-02 1.58621237e-01
4.67512935e-01 -2.41045296e-01 -1.03069258e+00 4.29279357e-02
-1.25175130e+00 -7.36874521e-01 -1.18221372e-01 -7.37269549e-03
-8.67830575e-01 -9.80107561e-02 -6.56925857e-01 -6.34422958e-01
-3.03756624e-01 -1.87067902e+00 7.58917451e-01 1.85696825e-01
-8.51331234e-01 -1.00590003e+00 -9.46674570e-02 -5.14071763e-01
7.38042772e-01 4.55731042e-02 1.77737069e+00 -7.38752842e-01
-8.98641944e-01 -8.57782289e-02 2.63709184e-02 -1.20592996e-01
-8.89598951e-02 2.65394628e-01 -2.71191984e-01 -3.54172662e-02
-2.96031773e-01 -2.65495211e-01 3.72969687e-01 2.86217093e-01
1.43984425e+00 -5.86035788e-01 -7.24315286e-01 9.72242177e-01
1.29682136e+00 7.28106380e-01 4.15367305e-01 6.63490176e-01
2.77347147e-01 3.71039152e-01 4.07087505e-01 3.32726866e-01
1.75645411e-01 5.14116526e-01 5.10138631e-01 1.64475605e-01
1.25180855e-01 -3.77099872e-01 3.25176954e-01 1.99215963e-01
2.23343581e-01 -1.01396114e-01 -1.23114848e+00 5.07989705e-01
-1.73694289e+00 -5.76274455e-01 3.06058079e-01 2.07155585e+00
1.11687648e+00 3.09588790e-01 7.30516911e-02 8.74156505e-02
5.14466882e-01 -3.18751559e-02 -5.59496760e-01 -1.04298365e+00
3.63674283e-01 4.58073348e-01 7.19696760e-01 5.65695524e-01
-7.11591244e-01 1.21342480e+00 7.32411194e+00 5.43952107e-01
-1.41076386e+00 -4.77883369e-01 3.43084514e-01 1.72096655e-01
-7.15751708e-01 2.04132214e-01 -9.37101066e-01 5.20842597e-02
1.01721275e+00 -4.94561791e-01 1.06070304e+00 1.11848891e+00
-1.87189311e-01 1.32208876e-02 -1.51400995e+00 4.58406270e-01
-2.74474174e-01 -1.76014245e+00 -5.67579977e-02 -3.42854589e-01
5.41848123e-01 -1.52260453e-01 -4.22293134e-02 6.54809535e-01
7.07964957e-01 -1.12387514e+00 4.13184077e-01 2.63966680e-01
4.63029504e-01 -1.01451433e+00 1.18531466e-01 3.19198251e-01
-1.01372182e+00 -4.11108136e-01 5.59831560e-02 -6.12476766e-02
-1.30257443e-01 1.76531803e-02 -1.05209696e+00 -2.20836282e-01
3.62090558e-01 1.18615583e-03 -5.04490733e-01 9.74774301e-01
-2.32377410e-01 3.67154121e-01 -3.75224769e-01 -5.55852413e-01
3.60029668e-01 2.40772992e-01 3.13827515e-01 1.34887266e+00
3.41607213e-01 -3.13480139e-01 3.01935017e-01 1.51353347e+00
1.83565721e-01 -1.57457709e-01 -6.30544245e-01 -4.89981234e-01
6.25621557e-01 8.09845150e-01 -9.50578094e-01 -3.83853555e-01
-4.19201106e-01 4.56454843e-01 3.36858243e-01 5.63546240e-01
-9.54342186e-01 -7.41834342e-01 6.44539773e-01 -3.71708274e-02
3.04388016e-01 -3.99108976e-01 -6.16513014e-01 -7.45618165e-01
-1.20463006e-01 -1.41015148e+00 2.22856522e-01 -6.34962559e-01
-3.71542156e-01 6.57709956e-01 4.21331674e-01 -3.17047328e-01
-8.66181433e-01 -4.89856005e-01 -6.31147683e-01 1.03309917e+00
-9.88500953e-01 -6.66204691e-01 1.41567260e-01 2.16306835e-01
4.90413636e-01 -3.41322184e-01 1.06598437e+00 -5.80079742e-02
-1.94686562e-01 6.47364736e-01 -2.68472403e-01 -1.06703706e-01
1.77961960e-03 -9.90633845e-01 7.77721345e-01 5.93910217e-01
-4.31446433e-01 1.11365640e+00 7.17689574e-01 -6.72943652e-01
-1.87993765e+00 -6.11620724e-01 7.94167399e-01 -8.56359974e-02
6.98036909e-01 -1.99694887e-01 -9.02041137e-01 9.88496244e-01
-1.07360691e-04 -2.76598155e-01 2.22419798e-01 1.93268031e-01
-3.46723646e-01 1.75671875e-01 -9.20724392e-01 1.03994429e+00
8.86093676e-01 -5.51961005e-01 -4.89036292e-01 1.34279683e-01
9.16533411e-01 -8.15922081e-01 -9.34808969e-01 5.17145634e-01
4.55421716e-01 -8.82619441e-01 1.17174876e+00 -5.62020540e-01
4.31916237e-01 -3.69666696e-01 -2.21117198e-01 -1.24987626e+00
-2.90862948e-01 -6.49469554e-01 -3.78559023e-01 8.00587416e-01
6.35971069e-01 -7.39400983e-01 9.07809138e-01 7.42659330e-01
-3.14251155e-01 -8.48062515e-01 -3.48085821e-01 -5.98753452e-01
1.91023454e-01 -4.06282991e-01 1.20128095e+00 7.11505949e-01
1.85427308e-01 2.07124993e-01 6.30166903e-02 5.77100404e-02
3.87600921e-02 5.51968873e-01 6.70805454e-01 -8.25027466e-01
-7.98969507e-01 -1.23203337e+00 2.88382322e-02 -1.23024225e+00
5.54232776e-01 -1.17644000e+00 1.39272302e-01 -1.26660848e+00
-4.34017658e-01 -7.35735118e-01 7.98460841e-02 8.26331854e-01
2.22269252e-01 -4.99437988e-01 -9.13212523e-02 -7.37086609e-02
-9.67217982e-02 1.53571859e-01 1.12758446e+00 -3.31029564e-01
-5.25272369e-01 4.54570204e-02 -6.49268568e-01 7.11327434e-01
7.52696872e-01 -2.43267402e-01 -5.33262432e-01 -5.17396152e-01
7.29155779e-01 6.62956655e-01 1.53408483e-01 -9.36145246e-01
3.73659968e-01 -5.85224211e-01 1.25788406e-01 -5.28598249e-01
3.08031738e-02 -6.75543189e-01 2.40150020e-01 8.86517107e-01
-1.03330719e+00 5.55859923e-01 6.26850069e-01 -9.41339806e-02
-9.77613479e-02 -6.01518273e-01 5.06753325e-01 -2.77860969e-01
-8.20708513e-01 -8.59422982e-02 -6.17548823e-01 -9.36373472e-02
7.57101834e-01 -1.44025713e-01 -1.18717410e-01 -1.01646967e-01
-6.12421870e-01 2.27434069e-01 6.71877503e-01 5.96812129e-01
4.50116366e-01 -1.19130838e+00 -1.31358132e-01 5.61494648e-01
5.47822416e-02 -5.53199053e-02 -5.89349151e-01 1.86220169e-01
-9.80154514e-01 6.19913757e-01 -2.24223018e-01 -3.35629284e-01
-1.05128849e+00 7.99263597e-01 3.97371709e-01 -3.07365626e-01
-4.68695641e-01 6.77388072e-01 1.43784270e-01 -8.79873097e-01
5.15450895e-01 -8.41611028e-01 1.24777228e-01 -6.16656959e-01
4.04522240e-01 3.24695021e-01 -1.11728124e-01 1.37818260e-02
-3.12030643e-01 2.36634880e-01 1.63882136e-01 -6.54399693e-02
1.01762927e+00 6.02257490e-01 -4.98185754e-01 2.68212855e-01
1.09566450e+00 -1.00829095e-01 -7.19436705e-01 -4.34388742e-02
3.24121416e-01 -3.19451839e-01 1.37138113e-01 -9.11655724e-01
-8.58124197e-01 4.65187281e-01 9.27335992e-02 2.06289232e-01
1.11757588e+00 -1.25060022e-01 5.72185516e-01 8.98919344e-01
4.46949512e-01 -8.95788968e-01 -2.17871070e-01 8.73354971e-01
7.78226018e-01 -5.03179252e-01 -2.26767719e-01 -2.73719758e-01
-9.89679694e-02 1.58740783e+00 8.73362184e-01 -2.41462111e-01
2.22679138e-01 7.83586025e-01 -3.77447963e-01 -9.98993143e-02
-9.26142514e-01 2.94843704e-01 1.44957021e-01 3.88312787e-01
7.09124327e-01 -6.17916286e-02 -4.01578486e-01 1.51131451e-01
-3.06093335e-01 3.96099567e-01 2.55455434e-01 1.32823706e+00
-4.59634423e-01 -1.40002990e+00 -5.46274722e-01 5.43693781e-01
-3.75948191e-01 -2.39514291e-01 -1.01152316e-01 8.22284639e-01
-8.08343068e-02 3.91272902e-01 1.24309778e-01 -3.59525323e-01
2.25101948e-01 2.24429250e-01 6.05136573e-01 -6.42879188e-01
-1.12092459e+00 -2.05666035e-01 4.63468492e-01 -7.34683931e-01
3.43451262e-01 -3.98608208e-01 -1.67419231e+00 -3.55893314e-01
-2.52848119e-01 2.32007533e-01 7.60870278e-01 4.93351191e-01
4.13561612e-01 9.26594436e-01 1.25619605e-01 -7.65994489e-01
-6.72800958e-01 -1.36091903e-01 2.50113100e-01 -4.27554250e-02
3.44800889e-01 -4.23160076e-01 -7.52400979e-02 -3.31361502e-01] | [8.279458045959473, 7.3221845626831055] |
75b77fca-6f4a-4685-80e1-225cfa0addf8 | interpreting-arabic-transformer-models | 2201.07434 | null | https://arxiv.org/abs/2201.07434v2 | https://arxiv.org/pdf/2201.07434v2.pdf | Interpreting Arabic Transformer Models | Arabic is a Semitic language which is widely spoken with many dialects. Given the success of pre-trained language models, many transformer models trained on Arabic and its dialects have surfaced. While these models have been compared with respect to downstream NLP tasks, no evaluation has been carried out to directly compare the internal representations. We probe how linguistic information is encoded in Arabic pretrained models, trained on different varieties of Arabic language. We perform a layer and neuron analysis on the models using three intrinsic tasks: two morphological tagging tasks based on MSA (modern standard Arabic) and dialectal POS-tagging and a dialectal identification task. Our analysis enlightens interesting findings such as: i) word morphology is learned at the lower and middle layers ii) dialectal identification necessitate more knowledge and hence preserved even in the final layers, iii) despite a large overlap in their vocabulary, the MSA-based models fail to capture the nuances of Arabic dialects, iv) we found that neurons in embedding layers are polysemous in nature, while the neurons in middle layers are exclusive to specific properties. | ['Hassan Sajjad', 'Fahim Dalvi', 'Nadir Durrani', 'Ahmed Abdelali'] | 2022-01-19 | null | null | null | null | ['morphological-tagging'] | ['natural-language-processing'] | [-1.78739712e-01 -1.21351937e-02 1.48676470e-01 -3.91799122e-01
-1.72152996e-01 -1.07496452e+00 6.64937139e-01 4.08165127e-01
-6.50483251e-01 1.53425708e-01 4.75931793e-01 -4.46663588e-01
1.38065219e-01 -9.01017606e-01 -4.59600329e-01 -7.72712290e-01
-2.66604275e-01 5.89969456e-01 4.36250605e-02 -7.85788894e-01
1.68500215e-01 5.76596558e-01 -1.08411336e+00 4.18557167e-01
1.03207219e+00 4.37821716e-01 1.09184764e-01 2.56911576e-01
-1.13811061e-01 5.40065110e-01 -3.75245154e-01 -5.95336795e-01
1.13914825e-01 -4.76784080e-01 -9.57535982e-01 -1.39949113e-01
6.08621359e-01 -2.12593555e-01 -2.17900038e-01 1.17555165e+00
2.11432129e-01 -1.75746933e-01 9.80304778e-01 -5.76948702e-01
-1.38010883e+00 1.11236835e+00 -1.66053742e-01 3.06129932e-01
1.34179935e-01 1.88445628e-01 1.30558348e+00 -1.09005475e+00
6.08679831e-01 1.53445768e+00 9.75994825e-01 3.87455344e-01
-1.30929482e+00 -1.70908228e-01 2.82246619e-01 3.93452942e-02
-1.39305913e+00 -5.24737298e-01 4.38742250e-01 -3.85365486e-01
1.12570822e+00 -8.53700712e-02 7.12420583e-01 7.61943460e-01
1.72679797e-01 6.46020293e-01 1.45400918e+00 -7.22389817e-01
-1.51242703e-01 2.60712445e-01 5.60454488e-01 9.52647090e-01
2.29606628e-01 -8.41848850e-02 -3.86339337e-01 1.26638919e-01
7.74181008e-01 -4.32160079e-01 -1.29170194e-01 4.36843783e-02
-1.24183047e+00 9.13632393e-01 5.56473672e-01 7.25677669e-01
-3.87306124e-01 -2.94710696e-01 4.59846079e-01 7.01562881e-01
2.04502530e-02 5.58081508e-01 -6.63895726e-01 2.98052073e-01
-6.39544487e-01 -2.41611585e-01 8.67449045e-01 6.04113936e-01
7.20977724e-01 5.04664242e-01 3.49858314e-01 1.29721141e+00
5.66746950e-01 7.01943874e-01 7.81507432e-01 -4.85071629e-01
4.87483114e-01 6.29546106e-01 -3.62948924e-01 -9.92217839e-01
-5.06975114e-01 2.97195502e-02 -4.59357291e-01 1.97948024e-01
9.10855770e-01 -2.72456318e-01 -7.54562914e-01 1.86709130e+00
-3.22744399e-01 -7.89698243e-01 3.09655100e-01 6.68846071e-01
5.95158517e-01 8.00757527e-01 3.70307773e-01 2.42860794e-01
1.55574298e+00 -6.67404175e-01 -4.43528980e-01 -5.88840008e-01
7.86102414e-01 -8.74189377e-01 1.44778144e+00 2.37309650e-01
-1.29851758e+00 -5.09477794e-01 -1.26653230e+00 -2.34076604e-01
-1.04785347e+00 1.76996753e-01 5.61180055e-01 1.06144309e+00
-1.45274937e+00 4.62925613e-01 -7.34827995e-01 -7.50831544e-01
2.38803253e-01 4.98508126e-01 -6.35295093e-01 -1.02537766e-01
-1.33242512e+00 1.65021265e+00 4.75536615e-01 1.76872090e-01
-7.77021468e-01 -1.47860229e-01 -1.16067910e+00 1.48406392e-02
-4.67053622e-01 -6.55679479e-02 8.43055546e-01 -1.51215553e+00
-1.36451125e+00 1.35691500e+00 -8.60297307e-02 -6.33373335e-02
-4.10736233e-01 5.49104251e-02 -6.70907199e-01 9.87138003e-02
-1.72873914e-01 7.43761241e-01 4.09320056e-01 -1.29947412e+00
-5.79056501e-01 -7.48417735e-01 9.76460874e-02 4.10294503e-01
-4.65869188e-01 3.69021028e-01 1.51640013e-01 -8.64402771e-01
3.93973619e-01 -9.52532768e-01 6.87198043e-02 -3.85664046e-01
8.48975629e-02 -2.30832800e-01 5.93759567e-02 -1.23963237e+00
1.14091051e+00 -1.95952332e+00 1.26531988e-01 3.70274782e-01
-2.25980103e-01 3.75153929e-01 -4.72997248e-01 6.51317477e-01
5.01318090e-03 1.90109044e-01 -3.81900609e-01 1.96083650e-01
3.66106421e-01 3.17121595e-01 -3.20400685e-01 7.34279990e-01
6.37267411e-01 9.34428692e-01 -6.58808231e-01 -1.16044998e-01
-9.96638462e-02 3.74871582e-01 -4.89052624e-01 -2.10022405e-01
-5.30763008e-02 -3.89306396e-02 3.67113709e-01 8.84422243e-01
7.34714448e-01 3.04554045e-01 6.70781076e-01 -1.18225813e-01
-2.16116473e-01 9.10860837e-01 -8.93934309e-01 1.43802786e+00
-4.86796826e-01 6.36182249e-01 3.61239821e-01 -8.58997405e-01
6.79236770e-01 3.47986758e-01 -3.48346204e-01 -5.94870210e-01
3.54720891e-01 6.34926677e-01 8.33805680e-01 -2.20484495e-01
4.56526250e-01 -3.68299901e-01 -2.94088274e-01 6.47556663e-01
6.39327466e-01 1.42739490e-01 1.78761140e-01 -1.29296750e-01
5.35712242e-01 1.12941504e-01 4.29831654e-01 -8.44369829e-01
6.74966335e-01 1.71290159e-01 1.58029780e-01 4.18153405e-01
-1.91071540e-01 4.35004920e-01 4.55485851e-01 -3.96315098e-01
-6.95836306e-01 -1.31866550e+00 -5.53190589e-01 1.73090839e+00
-1.43788904e-01 -1.04585074e-01 -9.33303416e-01 -6.99379981e-01
-1.95237875e-01 7.42370903e-01 -7.65741527e-01 -1.16839468e-01
-9.66843724e-01 -1.32883990e+00 1.01983297e+00 5.29833496e-01
1.43774137e-01 -1.55959070e+00 -1.36065245e-01 2.16629893e-01
1.55282214e-01 -6.50686622e-01 -3.74070376e-01 5.49965918e-01
-8.64554524e-01 -9.23605859e-01 -5.31236827e-01 -1.63154280e+00
6.95787847e-01 -1.85156494e-01 1.25731719e+00 1.13882199e-01
4.70831215e-01 2.79948711e-01 -3.18416089e-01 -4.35381979e-01
-8.25595260e-01 3.25810283e-01 1.48695499e-01 -4.72434461e-02
1.01572549e+00 -3.80569130e-01 -8.75987634e-02 6.25079051e-02
-9.55497563e-01 -7.13212729e-01 7.91437626e-01 5.71036279e-01
6.80076331e-02 -4.34810460e-01 7.05187142e-01 -1.00832486e+00
4.98281419e-01 -5.69321990e-01 -3.49218190e-01 2.85227537e-01
-4.22978759e-01 1.25990659e-01 5.82323074e-01 -3.78238678e-01
-8.68033290e-01 -1.09425478e-01 -7.16253519e-01 7.78392076e-01
-5.02224028e-01 7.54054368e-01 -4.53704596e-01 -5.68148792e-02
7.84098387e-01 4.06252712e-01 2.86662787e-01 -3.89429331e-01
4.12018180e-01 6.79381788e-01 3.45494092e-01 -4.98970509e-01
6.40015781e-01 3.27707052e-01 -5.60182989e-01 -1.12780595e+00
-4.15391564e-01 -5.08621670e-02 -1.29080737e+00 2.39418656e-01
9.98658538e-01 -9.81719851e-01 -4.89770651e-01 8.60848963e-01
-1.10463965e+00 -5.56477129e-01 -6.97177723e-02 3.02388281e-01
-2.75754750e-01 3.56653005e-01 -1.20100117e+00 -3.25132012e-01
-6.78923503e-02 -1.00559747e+00 3.25036258e-01 -1.51519537e-01
-4.97643113e-01 -1.56983149e+00 1.53282389e-01 -3.12316477e-01
4.09901381e-01 -3.72033834e-01 1.79027402e+00 -1.02745533e+00
-2.59365905e-02 7.77864307e-02 -1.59395441e-01 4.38172460e-01
2.78352946e-01 -8.50313231e-02 -1.11208487e+00 -2.07291082e-01
3.70212570e-02 -4.84535515e-01 8.94879699e-01 1.21410608e-01
2.18646958e-01 -4.22236919e-01 1.50068909e-01 4.35091734e-01
1.38864887e+00 1.90734521e-01 5.99685848e-01 4.59502310e-01
6.95570171e-01 1.09907651e+00 4.67276759e-02 -4.25576448e-01
7.08427250e-01 1.99302241e-01 2.05482185e-01 1.16128482e-01
-3.15514475e-01 -7.56954849e-02 1.25244868e+00 1.51638293e+00
1.39438778e-01 -3.13826986e-02 -1.17046368e+00 1.04991055e+00
-1.26669002e+00 -5.59831560e-01 -1.13602296e-01 2.17787313e+00
1.32353842e+00 -9.24637839e-02 1.03818357e-01 -2.12142561e-02
4.10295129e-01 2.12341964e-01 -4.67401855e-02 -8.94477189e-01
-7.51886785e-01 3.97777706e-01 1.72273591e-01 7.49166191e-01
-1.01113498e+00 1.53562605e+00 6.65947914e+00 4.59732234e-01
-1.24628675e+00 1.46408424e-01 3.62089247e-01 5.90423942e-01
-2.84306884e-01 -2.29359388e-01 -9.47186768e-01 1.93754688e-01
1.30429661e+00 5.48140645e-01 5.35317838e-01 3.00607562e-01
-3.62425596e-01 1.42770723e-01 -1.18416095e+00 3.55608433e-01
4.27380323e-01 -9.86106277e-01 5.19006729e-01 -8.63580853e-02
3.47238600e-01 3.60460937e-01 1.59471869e-01 5.23663104e-01
7.16080904e-01 -1.44772530e+00 9.64078248e-01 4.01156843e-02
6.93562388e-01 -7.74321198e-01 6.77265286e-01 1.06075391e-01
-6.84762359e-01 -5.81520088e-02 -6.50092185e-01 -2.33999401e-01
-1.12668976e-01 -3.32690001e-01 -6.85897827e-01 -8.90569761e-02
4.20638502e-01 5.29688418e-01 -1.00052261e+00 4.62066293e-01
-4.69152421e-01 9.80948925e-01 -1.70316786e-01 -6.83557391e-02
5.94132841e-01 -5.41071415e-01 2.77516723e-01 1.67193711e+00
2.14867860e-01 -2.06620649e-01 -9.37964246e-02 5.17043591e-01
1.60792157e-01 5.19253075e-01 -5.32523751e-01 -2.78371483e-01
2.99649656e-01 1.12092245e+00 -8.22114766e-01 1.46113662e-02
-7.17883766e-01 1.00084770e+00 6.39443457e-01 3.38627845e-01
-2.55458802e-01 -4.54581141e-01 7.73653746e-01 7.71617666e-02
1.90401703e-01 -5.44801295e-01 -5.76959372e-01 -1.05133498e+00
-4.47256774e-01 -1.04985273e+00 4.74206448e-01 -4.70949709e-01
-1.64096797e+00 8.48875523e-01 -6.42607987e-01 -5.79424441e-01
-2.25842804e-01 -1.53101707e+00 -4.05144095e-01 1.27278543e+00
-1.35844815e+00 -1.54558480e+00 3.64500195e-01 6.24815166e-01
1.91793084e-01 -5.85984170e-01 1.55119824e+00 2.08145946e-01
-4.17018980e-01 6.88301265e-01 2.35558048e-01 9.10892844e-01
9.13857102e-01 -1.49489093e+00 3.35719377e-01 8.24936807e-01
6.41561985e-01 1.02828145e+00 1.52840808e-01 -3.51013780e-01
-1.08184183e+00 -8.10798109e-01 1.45471931e+00 -9.76480782e-01
9.35172558e-01 -5.75339973e-01 -1.19834757e+00 1.12902296e+00
8.75562966e-01 -4.68497455e-01 9.65291500e-01 3.78278434e-01
-7.64932096e-01 -9.29644778e-02 -8.28240931e-01 8.07556570e-01
6.84271038e-01 -7.03498006e-01 -1.17785645e+00 6.84728026e-02
3.86290312e-01 2.01115116e-01 -7.64293194e-01 -3.32881175e-02
4.92012590e-01 -6.97159111e-01 7.85061955e-01 -8.80590796e-01
3.60871911e-01 -1.08219482e-01 -4.82862175e-01 -1.70277178e+00
-5.62070310e-01 -6.27465472e-02 4.48988527e-01 1.36754739e+00
8.66052389e-01 -6.59291208e-01 1.83603019e-01 7.36486465e-02
-3.10340613e-01 -2.29025051e-01 -6.00117803e-01 -6.29800737e-01
1.00910616e+00 -8.83624628e-02 4.15370435e-01 1.48698509e+00
2.47442111e-01 6.67888224e-01 4.77208465e-01 3.50357115e-01
-2.75113676e-02 -2.68104747e-02 -9.67581570e-02 -1.36074460e+00
1.11912355e-01 -8.73597026e-01 -1.84493840e-01 -6.14410937e-01
4.55824494e-01 -1.54706800e+00 6.59797117e-02 -1.34269607e+00
-2.83956856e-01 -6.89715743e-01 -3.13243508e-01 8.38843822e-01
-2.12680306e-02 6.05418980e-01 1.12573579e-01 1.13596015e-01
-4.32199985e-02 1.75443187e-01 6.72679663e-01 -8.71981159e-02
-1.34000227e-01 -3.27530116e-01 -9.46642697e-01 1.11648941e+00
8.92566621e-01 -3.48133892e-01 -2.88330279e-02 -1.01293921e+00
5.77058077e-01 -5.55497229e-01 6.10749424e-02 -7.67960429e-01
-1.61771402e-01 2.82734111e-02 6.20582759e-01 -2.57016756e-02
1.15241051e-01 -8.29679012e-01 -7.44601071e-01 5.91054261e-01
-8.85070190e-02 5.43501854e-01 4.54581201e-01 -9.46679264e-02
-3.15871775e-01 -6.67327464e-01 1.14328849e+00 -4.11458284e-01
-9.02521551e-01 2.98499577e-02 -1.11964750e+00 1.58914819e-01
3.28536123e-01 -2.17810586e-01 -2.03261524e-01 8.57489035e-02
-9.23988879e-01 -2.50828177e-01 6.52158678e-01 4.37130481e-01
1.59787357e-01 -1.23986220e+00 -9.39386368e-01 4.40704882e-01
1.27920121e-01 -6.13640726e-01 -3.17530870e-01 7.69711256e-01
-9.12142754e-01 3.76087099e-01 -7.41699278e-01 -1.77269921e-01
-8.15885842e-01 3.22762012e-01 3.16335827e-01 -2.61876918e-02
4.63171862e-02 8.45734119e-01 1.23104483e-01 -1.16573238e+00
-8.92354548e-02 -1.30598977e-01 -8.11771810e-01 8.50437105e-01
3.27742875e-01 2.28035405e-01 2.01043412e-01 -1.32147253e+00
-3.89861256e-01 4.59775209e-01 -1.83849677e-01 -3.37758988e-01
1.44251990e+00 -2.15077862e-01 -7.84252763e-01 7.39122629e-01
8.06435943e-01 7.41070211e-01 -7.59703159e-01 -1.76694825e-01
4.70264256e-01 3.12329531e-01 -3.61289561e-01 -1.16255975e+00
-7.50608146e-01 1.32080162e+00 4.28914994e-01 1.67657495e-01
8.31353664e-01 -1.82039946e-01 5.16276419e-01 4.64955389e-01
1.02661110e-01 -1.08507848e+00 -3.50029767e-01 1.51555085e+00
8.54916036e-01 -1.08251286e+00 -6.32955730e-01 -8.23867321e-02
-7.70782471e-01 1.17571628e+00 5.23078740e-01 -4.85796273e-01
7.61344731e-01 1.52270168e-01 6.46170139e-01 -1.51390210e-01
-2.11236417e-01 -3.52952540e-01 3.09287190e-01 8.95188868e-01
9.39035058e-01 -1.00920498e-02 -4.24937785e-01 9.65775371e-01
-5.69419265e-01 -9.09504175e-01 6.77529871e-01 6.23995125e-01
-4.82660621e-01 -1.13024294e+00 -4.60641414e-01 1.49044260e-01
-5.72175026e-01 -7.45386183e-01 -6.56963229e-01 1.01911652e+00
3.91647309e-01 7.13117599e-01 3.01995814e-01 -6.47323281e-02
5.12006432e-02 7.73678541e-01 7.31141567e-01 -9.58305240e-01
-1.13365030e+00 -2.05669031e-01 1.36820704e-01 5.39870001e-03
-2.72173017e-01 -9.32952344e-01 -1.41031492e+00 -2.06531450e-01
3.01540699e-02 -4.99349050e-02 4.01090145e-01 9.32766140e-01
-5.38940094e-02 3.64088826e-02 2.34884918e-02 -7.65160322e-01
-5.77377319e-01 -1.27789879e+00 -1.04135311e+00 3.73345166e-01
2.60960609e-01 -2.19062969e-01 -1.27760842e-01 2.04479143e-01] | [10.567975044250488, 10.090438842773438] |
efdf1269-5c46-4628-92fe-a26f1421ca70 | predicting-above-sentence-discourse-structure | 2112.06196 | null | https://arxiv.org/abs/2112.06196v1 | https://arxiv.org/pdf/2112.06196v1.pdf | Predicting Above-Sentence Discourse Structure using Distant Supervision from Topic Segmentation | RST-style discourse parsing plays a vital role in many NLP tasks, revealing the underlying semantic/pragmatic structure of potentially complex and diverse documents. Despite its importance, one of the most prevailing limitations in modern day discourse parsing is the lack of large-scale datasets. To overcome the data sparsity issue, distantly supervised approaches from tasks like sentiment analysis and summarization have been recently proposed. Here, we extend this line of research by exploiting distant supervision from topic segmentation, which can arguably provide a strong and oftentimes complementary signal for high-level discourse structures. Experiments on two human-annotated discourse treebanks confirm that our proposal generates accurate tree structures on sentence and paragraph level, consistently outperforming previous distantly supervised models on the sentence-to-document task and occasionally reaching even higher scores on the sentence-to-paragraph level. | ['Giuseppe Carenini', 'Linzi Xing', 'Patrick Huber'] | 2021-12-12 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [ 3.38047236e-01 7.35871732e-01 -5.33309639e-01 -4.30129230e-01
-1.35448015e+00 -8.96593273e-01 1.03903902e+00 7.35933304e-01
-2.87715018e-01 1.22771454e+00 1.00581014e+00 -4.10049349e-01
-1.08641982e-02 -3.74111861e-01 -5.08038044e-01 -5.31336486e-01
1.27204165e-01 4.67493385e-01 4.99975026e-01 -4.14177090e-01
6.42244637e-01 -1.32142812e-01 -1.20077002e+00 7.83920765e-01
9.33106899e-01 5.74085593e-01 2.53912389e-01 5.81666946e-01
-5.89250028e-01 1.16714036e+00 -9.66811240e-01 -7.41912007e-01
-4.91429865e-01 -6.58842444e-01 -1.33710575e+00 2.48167902e-01
3.25094908e-01 8.99072140e-02 3.65305841e-01 9.70772803e-01
1.30263895e-01 -1.80299878e-01 4.99935687e-01 -4.69621569e-01
-1.48250133e-01 9.76254225e-01 -4.77720767e-01 2.74276018e-01
7.69897997e-01 -3.10415000e-01 1.64506221e+00 -5.42262852e-01
1.05291748e+00 1.29693294e+00 4.31431442e-01 3.71720254e-01
-1.11814034e+00 1.37174865e-02 2.75837749e-01 3.07675805e-02
-5.66096187e-01 -5.36080241e-01 1.01706159e+00 -5.31953692e-01
8.72795224e-01 4.02415782e-01 4.24917638e-01 1.39108133e+00
-1.50999621e-01 1.25067878e+00 1.35760641e+00 -6.04815722e-01
2.15895832e-01 2.00447038e-01 3.27004135e-01 2.84648746e-01
1.23332635e-01 -9.16301370e-01 -5.88645875e-01 -2.86586761e-01
1.27078533e-01 -7.55075574e-01 -4.76894468e-01 1.69230700e-01
-1.17609060e+00 9.30159867e-01 -2.02728063e-02 9.44178164e-01
-3.35284293e-01 -2.88065493e-01 9.02029395e-01 2.21067339e-01
7.71861553e-01 7.38425553e-01 -5.61200798e-01 -6.62891448e-01
-8.73073220e-01 3.81539285e-01 1.21136439e+00 5.12005925e-01
1.90785185e-01 -3.75457138e-01 -2.59115070e-01 8.28088641e-01
1.81450933e-01 9.29939896e-02 4.57433790e-01 -1.15351093e+00
1.09289372e+00 8.11249614e-01 1.48802146e-01 -9.58883584e-01
-3.55393797e-01 -2.40736082e-01 -6.12868607e-01 -2.98158973e-01
8.36901367e-01 -2.12187737e-01 -1.66151479e-01 1.65468633e+00
4.85149652e-01 -5.85667253e-01 2.81519979e-01 7.25143373e-01
8.79551232e-01 7.24581897e-01 2.54690230e-01 -6.21243238e-01
1.68850529e+00 -8.29967678e-01 -8.57233524e-01 -3.88868779e-01
7.41721570e-01 -9.34418440e-01 1.08859503e+00 3.87002051e-01
-1.20621085e+00 -1.97638780e-01 -8.53819966e-01 -3.74356598e-01
-1.23588424e-02 1.13601319e-01 4.17899907e-01 5.33610463e-01
-7.81124592e-01 5.42666614e-01 -5.47614396e-01 -2.36513868e-01
5.94936013e-01 -2.60132521e-01 -2.25388497e-01 7.23153651e-02
-1.12645483e+00 8.20617676e-01 5.87435246e-01 -2.44046450e-02
-1.87913477e-01 -6.05627120e-01 -9.02048230e-01 2.90736295e-02
7.48435140e-01 -4.05950785e-01 1.41099966e+00 -9.88456428e-01
-1.71582127e+00 1.08274662e+00 -4.71298128e-01 -6.66266024e-01
6.14649475e-01 -4.74004656e-01 1.59911126e-01 3.21406603e-01
2.47156069e-01 2.58846611e-01 5.78846455e-01 -1.21241581e+00
-6.71900809e-01 -3.96048903e-01 2.61051446e-01 3.99331272e-01
-2.60174334e-01 3.93896908e-01 8.77455771e-02 -4.86868531e-01
-1.76240299e-02 -5.06881714e-01 -2.02078685e-01 -5.66145480e-01
-6.92219794e-01 -7.62641191e-01 5.70964694e-01 -1.04553890e+00
1.31492174e+00 -1.81185043e+00 5.64108074e-01 -4.90335286e-01
9.14316773e-02 2.40170985e-01 3.42606544e-01 8.10441494e-01
1.62427604e-01 1.13471463e-01 -4.83485460e-01 -5.87639868e-01
2.94386595e-02 1.91687271e-01 -5.85571408e-01 2.04370767e-01
5.40348172e-01 9.81634319e-01 -9.94869411e-01 -5.73992670e-01
-7.29088187e-02 1.58937663e-01 -2.69403100e-01 8.58516917e-02
-8.64569366e-01 8.12470198e-01 -7.07471371e-01 3.44981402e-01
-5.53090237e-02 -3.90724540e-01 6.57169640e-01 2.04486623e-01
-2.60543227e-01 1.10306096e+00 -6.83937609e-01 1.73045313e+00
-2.53358245e-01 9.34123814e-01 3.34131271e-01 -1.27817440e+00
5.11695683e-01 5.29708326e-01 2.02139065e-01 -5.29537082e-01
2.96043456e-01 3.58203560e-01 1.19539373e-01 -6.53778613e-01
9.35696125e-01 -3.40654552e-01 -4.99213487e-01 3.00287664e-01
-2.11787432e-01 -1.42530039e-01 4.20058757e-01 2.63034731e-01
8.97950232e-01 6.78958222e-02 6.98335767e-01 -4.69831914e-01
8.55938792e-01 4.40941751e-01 4.20825899e-01 3.25093657e-01
-1.06302043e-02 7.07497895e-01 1.33533418e+00 9.19564534e-03
-9.75344539e-01 -4.55987126e-01 -3.88871253e-01 1.14263606e+00
-2.64307946e-01 -4.44509566e-01 -9.92568493e-01 -7.09936321e-01
-2.70280898e-01 7.43827701e-01 -5.78997731e-01 8.00988078e-01
-1.09438062e+00 -7.69258797e-01 5.11082947e-01 2.83211887e-01
2.78203338e-01 -1.11787891e+00 -5.37556946e-01 4.76071507e-01
-5.64223766e-01 -1.49664176e+00 2.52110571e-01 -2.71069128e-02
-8.76690984e-01 -1.32247102e+00 -7.54283488e-01 -6.47903085e-01
2.60650158e-01 -1.30555406e-01 1.23457050e+00 -7.96671361e-02
1.71879813e-01 -6.63972497e-02 -4.79311705e-01 -4.43696022e-01
-9.29115653e-01 4.92043346e-01 -4.87501591e-01 -2.70114362e-01
1.59933399e-02 -3.59615326e-01 -2.35693157e-01 -3.05918396e-01
-6.77813232e-01 8.79069343e-02 2.55891174e-01 8.48515093e-01
1.17287524e-01 -2.99941212e-01 9.40658748e-01 -1.47427022e+00
9.21958268e-01 -5.11351705e-01 -3.18904012e-01 1.98057577e-01
-9.81382951e-02 -1.18855454e-01 5.94083726e-01 1.27500519e-02
-1.37510121e+00 -6.14943087e-01 -5.02867937e-01 6.09967113e-01
-4.21715885e-01 7.26752698e-01 -2.50216693e-01 8.54736865e-01
5.40452480e-01 2.51273774e-02 -8.23646262e-02 -4.84795630e-01
3.94989610e-01 6.24127269e-01 3.37896675e-01 -7.15493679e-01
5.51897407e-01 4.53647077e-01 -2.32657567e-01 -1.25646722e+00
-1.58105397e+00 -5.45820713e-01 -7.06727147e-01 7.96118602e-02
1.19819891e+00 -8.66125405e-01 -4.41656470e-01 1.95347875e-01
-1.46588349e+00 -2.63389349e-01 -4.12702739e-01 5.31589314e-02
-3.30001861e-01 5.59515536e-01 -7.61133671e-01 -8.80898178e-01
-2.33888939e-01 -7.55769432e-01 1.22178960e+00 1.95015550e-01
-8.14589381e-01 -1.34737110e+00 1.75852031e-01 8.66525114e-01
1.05991021e-01 6.03473008e-01 1.06092930e+00 -9.25210118e-01
-4.05353129e-01 -2.82895360e-02 -1.09234586e-01 2.28919253e-01
3.88761796e-02 -1.34279236e-01 -1.05993569e+00 1.10304337e-02
4.18850452e-01 -5.47283590e-01 7.77709126e-01 3.09825361e-01
5.25775075e-01 -4.02893931e-01 -1.25350699e-01 -3.53653222e-01
1.09819865e+00 -2.68687278e-01 3.97193015e-01 5.48488259e-01
4.67793792e-01 1.19381809e+00 6.74157798e-01 1.46676898e-01
6.70648396e-01 4.45874989e-01 6.86116889e-02 1.79678530e-01
-2.08760872e-01 -1.53516799e-01 4.15043920e-01 1.15669036e+00
3.65062431e-02 -2.53574222e-01 -8.28470528e-01 8.89525890e-01
-1.98702717e+00 -1.09360158e+00 -5.61599791e-01 1.69699538e+00
1.12593329e+00 5.33206463e-01 1.47817776e-01 4.30157155e-01
4.17967111e-01 8.52804065e-01 2.14976743e-02 -4.33629900e-01
-4.65057284e-01 -5.31860664e-02 -2.65303463e-01 6.33370638e-01
-1.09320199e+00 9.19106543e-01 5.58873415e+00 7.71820784e-01
-7.69685626e-01 1.49883598e-01 7.34165430e-01 1.43742055e-01
-4.37788099e-01 1.17720276e-01 -8.22877049e-01 4.01723415e-01
8.86929572e-01 -1.18683040e-01 -3.25617552e-01 8.30737770e-01
2.37421900e-01 -5.59069276e-01 -1.16978717e+00 3.00053239e-01
1.60928562e-01 -1.68315911e+00 -2.31754899e-01 1.15845958e-02
8.44188213e-01 -6.38940558e-02 -3.38721603e-01 1.85281813e-01
7.56109804e-02 -8.22727203e-01 6.61843896e-01 -7.18044341e-02
3.89321864e-01 -2.58966625e-01 8.53349566e-01 6.95377529e-01
-7.71518707e-01 1.14266180e-01 -7.67434090e-02 -4.98234659e-01
6.76974952e-01 6.24880970e-01 -8.57620597e-01 7.53600597e-01
3.75531197e-01 7.55142510e-01 -3.45721960e-01 4.79625612e-01
-6.84825242e-01 1.02988875e+00 -1.10907689e-01 -3.52582783e-01
4.98632163e-01 -9.24426690e-02 8.19530606e-01 1.34734750e+00
-2.30326399e-01 2.47245193e-01 1.58288643e-01 5.73101163e-01
-1.72191665e-01 4.03175086e-01 -4.39590722e-01 -1.28144473e-01
2.47084916e-01 1.14271140e+00 -8.32342625e-01 -4.78223979e-01
-3.50849837e-01 6.07667089e-01 5.39767206e-01 1.46354586e-01
-3.35565984e-01 3.96422967e-02 3.12175512e-01 4.92488220e-02
3.71136397e-01 -3.60879093e-01 -5.01293480e-01 -1.09891129e+00
4.52904552e-01 -9.61901367e-01 3.98777813e-01 -1.22975677e-01
-1.27939641e+00 4.74364132e-01 -2.37507448e-02 -7.92275310e-01
-4.49918747e-01 -5.08928061e-01 -7.41154432e-01 6.22710526e-01
-1.93306375e+00 -1.02972901e+00 2.08102629e-01 9.02077742e-03
1.08323157e+00 2.18570475e-02 8.24949741e-01 -1.09261476e-01
-5.08000314e-01 -3.76863629e-02 8.38905200e-02 2.76862592e-01
5.09113789e-01 -1.49528396e+00 2.49518052e-01 6.85711741e-01
2.80101657e-01 4.95609730e-01 1.07168400e+00 -5.12677729e-01
-1.13800299e+00 -4.64760631e-01 1.39835644e+00 -7.52089381e-01
1.13784468e+00 -4.45623904e-01 -1.33745897e+00 4.86274004e-01
7.16042399e-01 -7.99168169e-01 8.34006906e-01 5.04304528e-01
-1.38776854e-01 4.88060892e-01 -7.10534155e-01 5.39763868e-01
7.39460111e-01 -6.55991137e-01 -1.40353286e+00 5.09766340e-01
7.23603249e-01 -7.34404624e-01 -8.95353436e-01 1.65195122e-01
1.93030864e-01 -1.13028884e+00 5.78587413e-01 -6.52221560e-01
1.15955234e+00 1.08665034e-01 -4.09969650e-02 -9.17906463e-01
3.24549943e-01 -7.66314387e-01 -6.73900917e-02 1.72602534e+00
5.90438724e-01 -4.26931113e-01 8.93032968e-01 3.63053471e-01
-2.03878522e-01 -7.35180557e-01 -1.10457277e+00 -4.21544939e-01
5.31166553e-01 -3.97144169e-01 1.41977742e-02 1.01141739e+00
5.23333609e-01 9.62362647e-01 9.14751086e-04 -2.09327608e-01
4.43975449e-01 4.60064620e-01 7.54966736e-01 -1.43673110e+00
-3.13611746e-01 -8.37434292e-01 9.58306566e-02 -1.20883584e+00
5.65557837e-01 -7.80377269e-01 7.54410177e-02 -1.77898622e+00
2.23438084e-01 -1.60974801e-01 4.16808903e-01 -4.59085628e-02
-3.77749592e-01 -7.47049972e-02 3.57682891e-02 3.42491657e-01
-7.72495270e-01 5.92417836e-01 1.15535426e+00 1.29845351e-01
-2.08465800e-01 -5.77748939e-02 -9.15528595e-01 1.14972746e+00
5.96093893e-01 -3.85342747e-01 -1.38237402e-01 -4.24374521e-01
3.16658169e-01 2.59350359e-01 -3.23363766e-02 -3.62303466e-01
4.70571220e-02 -4.10287306e-02 -9.70435962e-02 -5.75428367e-01
3.67791623e-01 -3.27861398e-01 -5.33107340e-01 2.49261469e-01
-5.58918893e-01 -1.88419461e-01 1.19707972e-01 5.05596876e-01
-5.95741391e-01 -4.79012817e-01 3.39453429e-01 -2.27237314e-01
-3.05569828e-01 -5.44299781e-01 -4.31112915e-01 6.83270037e-01
9.14386213e-01 -7.75581896e-02 -7.18757629e-01 -2.30866119e-01
-4.92364168e-01 1.99695632e-01 3.72646302e-01 3.89510751e-01
2.03944564e-01 -7.19283700e-01 -9.94065523e-01 -3.65388066e-01
-1.86034620e-01 4.41805601e-01 9.64460801e-03 9.21070755e-01
-2.62478769e-01 8.40246558e-01 3.35635126e-01 -7.54309595e-01
-1.34497786e+00 -5.21608479e-02 -2.84632683e-01 -9.36763406e-01
-7.58144081e-01 7.09052742e-01 3.68015207e-02 -1.00722991e-01
9.13901478e-02 -2.25532934e-01 -5.86413622e-01 6.37962222e-01
5.42039156e-01 1.90887183e-01 -1.18350469e-01 -8.12777042e-01
-1.93383619e-01 3.40212762e-01 -1.71648800e-01 -2.13992387e-01
1.60076237e+00 -3.65321815e-01 -4.69110996e-01 8.32139134e-01
8.81266475e-01 4.19198722e-01 -1.10890937e+00 -2.32874185e-01
1.07001662e+00 -1.12218358e-01 -4.33654428e-01 -6.58702672e-01
-1.26214385e-01 1.05638909e+00 -5.90737283e-01 8.29843879e-01
5.39818347e-01 4.98371214e-01 8.93076062e-01 3.52610677e-01
1.93974137e-01 -1.11566257e+00 1.95299894e-01 7.45013654e-01
1.03863275e+00 -1.55231571e+00 2.65536100e-01 -5.99363029e-01
-8.24567139e-01 1.05520999e+00 1.51085213e-01 -4.20999080e-02
1.92310154e-01 1.99149057e-01 1.05425581e-01 -3.87682289e-01
-7.52669632e-01 -1.14430273e-04 2.38711998e-01 2.68037409e-01
8.83160114e-01 -1.14206791e-01 -7.67476559e-01 6.89946830e-01
-3.73903424e-01 -4.59284633e-01 6.96405113e-01 8.32067370e-01
-6.34105980e-01 -1.40369833e+00 -2.87275493e-01 8.72651935e-02
-1.05509984e+00 -8.98123719e-03 -5.94031751e-01 8.42587531e-01
-5.42854309e-01 9.66013312e-01 -4.50540669e-02 4.92293686e-01
2.15246245e-01 2.86006212e-01 4.47413534e-01 -8.93669605e-01
-8.88749242e-01 1.93865746e-01 9.50907230e-01 -3.21511596e-01
-9.70305920e-01 -1.14037633e+00 -1.29353404e+00 1.39522240e-01
-1.91152439e-01 4.52879190e-01 5.98672450e-01 1.41707218e+00
8.82028937e-02 5.85201144e-01 3.04443508e-01 -7.55364895e-01
-5.82406521e-01 -1.15051031e+00 -1.39597207e-01 5.16474843e-01
3.37031901e-01 -2.07509086e-01 -2.44752377e-01 7.49117434e-02] | [10.87222957611084, 9.386451721191406] |
f8a7eadc-70e3-4052-a251-8641aae2c39e | data-augmentation-for-skin-lesion-analysis | 1809.01442 | null | http://arxiv.org/abs/1809.01442v1 | http://arxiv.org/pdf/1809.01442v1.pdf | Data Augmentation for Skin Lesion Analysis | Deep learning models show remarkable results in automated skin lesion
analysis. However, these models demand considerable amounts of data, while the
availability of annotated skin lesion images is often limited. Data
augmentation can expand the training dataset by transforming input images. In
this work, we investigate the impact of 13 data augmentation scenarios for
melanoma classification trained on three CNNs (Inception-v4, ResNet, and
DenseNet). Scenarios include traditional color and geometric transforms, and
more unusual augmentations such as elastic transforms, random erasing and a
novel augmentation that mixes different lesions. We also explore the use of
data augmentation at test-time and the impact of data augmentation on various
dataset sizes. Our results confirm the importance of data augmentation in both
training and testing and show that it can lead to more performance gains than
obtaining new images. The best scenario results in an AUC of 0.882 for melanoma
classification without using external data, outperforming the top-ranked
submission (0.874) for the ISIC Challenge 2017, which was trained with
additional data. | ['Fábio Perez', 'Sandra Avila', 'Eduardo Valle', 'Cristina Vasconcelos'] | 2018-09-05 | null | null | null | null | ['skin-cancer-classification', 'skin-lesion-classification'] | ['medical', 'medical'] | [ 6.22987747e-01 1.86318874e-01 4.94052842e-02 -1.12458199e-01
-6.57563329e-01 -4.78646189e-01 7.47577548e-01 2.38799199e-01
-9.73500192e-01 7.20166087e-01 2.00476244e-01 -4.48613048e-01
1.70280889e-01 -6.82502031e-01 -5.27166486e-01 -8.26516569e-01
2.03471594e-02 4.83752191e-02 2.23489150e-01 -1.71741158e-01
-1.07608281e-01 6.06888056e-01 -1.19064260e+00 4.16047722e-01
8.59361410e-01 9.28922653e-01 -2.56487250e-01 8.75338972e-01
-6.90917894e-02 3.76672268e-01 -5.83628476e-01 -6.63150668e-01
4.59859937e-01 -1.39327615e-01 -6.52114928e-01 9.60545316e-02
5.49353182e-01 -3.17018092e-01 -1.99234292e-01 6.69473827e-01
7.08818376e-01 -3.58990610e-01 4.31143135e-01 -1.02772927e+00
-4.49390590e-01 2.73278117e-01 -6.95468366e-01 1.98872581e-01
-4.19678651e-02 5.39355636e-01 4.21739012e-01 -6.29598200e-01
9.17817295e-01 6.19129300e-01 1.15375531e+00 8.76346171e-01
-1.19570088e+00 -4.22105521e-01 -2.09954128e-01 1.24076875e-02
-1.22067606e+00 -2.20394880e-01 2.76443958e-01 -3.88776630e-01
8.44908118e-01 6.63989902e-01 7.36556709e-01 1.28841591e+00
2.48248316e-02 5.40624857e-01 1.47645664e+00 -6.30736649e-01
8.46326947e-02 2.56925166e-01 -1.73488036e-02 5.31790555e-01
2.84421831e-01 7.48913661e-02 -3.69657576e-01 -1.46314770e-01
5.48588932e-01 -1.34278193e-01 -2.28482574e-01 1.29562721e-01
-1.06550002e+00 5.80997765e-01 5.83219528e-01 2.68687487e-01
-4.41143870e-01 1.15446880e-01 5.76952398e-01 2.42217734e-01
6.96741402e-01 5.88534892e-01 -4.09314960e-01 8.17872584e-02
-1.02052224e+00 2.81512998e-02 4.26703811e-01 2.34535590e-01
2.09413096e-01 -2.97381990e-02 -3.54877293e-01 8.47317159e-01
-1.53772652e-01 2.64315605e-01 6.30427003e-01 -2.60490060e-01
1.98977277e-01 7.67880142e-01 -2.97449082e-01 -4.26192015e-01
-8.03424299e-01 -7.67278075e-01 -1.08306766e+00 2.95746535e-01
6.86904073e-01 -4.41692233e-01 -1.72524834e+00 1.65702224e+00
2.46173322e-01 1.26731306e-01 1.00015067e-01 7.56645679e-01
8.92867744e-01 1.35530666e-01 4.40543145e-01 1.73160657e-01
1.29691923e+00 -9.89496291e-01 -5.42386472e-01 -1.33386916e-02
1.00021970e+00 -7.78186262e-01 1.00156534e+00 4.50433582e-01
-9.57703948e-01 -3.49621743e-01 -8.98032904e-01 -1.79147795e-02
-7.07829416e-01 2.85765201e-01 7.43225515e-01 8.46871614e-01
-1.45434511e+00 3.77118021e-01 -8.80270541e-01 -8.12171340e-01
6.56103134e-01 5.26271284e-01 -7.87332833e-01 -2.61611283e-01
-9.65956569e-01 1.05575180e+00 2.74189860e-01 5.23811914e-02
-6.32545829e-01 -1.05772734e+00 -5.50029039e-01 -3.85057002e-01
1.63502559e-01 -6.35287106e-01 7.84281015e-01 -1.30852926e+00
-1.05414748e+00 1.00318384e+00 1.00933082e-01 -7.75633633e-01
9.71507430e-01 4.01375853e-02 -5.16800582e-01 8.94383118e-02
-3.11395913e-01 1.15885055e+00 5.82685292e-01 -9.90680039e-01
-4.80040044e-01 -2.97275007e-01 -1.01881452e-01 -7.64390361e-03
-7.32698023e-01 -1.57308459e-01 -4.51805890e-01 -6.51499748e-01
-3.91260713e-01 -1.11839736e+00 -5.63008547e-01 2.70575762e-01
-6.28083706e-01 3.37604374e-01 6.28363788e-01 -1.20352292e+00
9.59886491e-01 -2.12082100e+00 -1.53380021e-01 4.43057507e-01
3.30635190e-01 6.46789968e-01 -5.09915888e-01 2.14701861e-01
-4.13376212e-01 6.14173532e-01 -2.26554260e-01 -3.26060653e-01
-3.70863467e-01 -3.60332355e-02 2.65361965e-01 2.69152880e-01
6.86021626e-01 1.13257039e+00 -4.37227488e-01 -3.92709851e-01
1.99577674e-01 8.15979779e-01 -3.63328248e-01 -2.88209796e-01
-1.92817040e-02 2.37863302e-01 7.10136518e-02 8.01199913e-01
7.29885042e-01 -4.02041793e-01 4.45301458e-02 -2.76383430e-01
4.59560603e-02 -2.98724622e-01 -7.49977052e-01 1.54624283e+00
-3.81553948e-01 9.22575951e-01 -1.10826276e-01 -3.92079383e-01
5.93445122e-01 2.96511531e-01 5.14949381e-01 -6.07214928e-01
3.32187042e-02 1.28740966e-01 4.94841248e-01 -4.67260957e-01
4.08451259e-01 3.85748260e-02 4.01096582e-01 2.26838514e-01
1.29301706e-02 1.37043536e-01 3.73866647e-01 1.15972891e-01
1.53480840e+00 -1.67118981e-01 1.65524855e-01 -4.31779176e-02
2.77430385e-01 1.54541567e-01 1.71577796e-01 5.06580889e-01
-1.11213736e-01 6.42658174e-01 6.24907315e-01 -5.07284701e-01
-1.20811391e+00 -8.13720405e-01 -2.58306414e-01 7.27862656e-01
-4.12581265e-01 -2.66935021e-01 -8.83632183e-01 -8.93499076e-01
4.50373627e-02 3.84506285e-01 -1.38401663e+00 -1.36005968e-01
-2.18632892e-01 -1.41282213e+00 8.92014682e-01 5.89185536e-01
6.37193739e-01 -8.23086917e-01 -4.89557356e-01 -2.24206209e-01
-1.52661614e-02 -1.00537157e+00 -1.48854002e-01 2.26700619e-01
-7.26617813e-01 -1.24299228e+00 -1.07670236e+00 -4.53677475e-01
1.11195374e+00 -2.88464427e-01 7.16193676e-01 3.23122621e-01
-8.35972130e-01 2.35457346e-01 -4.42155212e-01 -6.66394770e-01
-4.53043550e-01 3.98912460e-01 -3.01256478e-01 9.89603400e-02
3.54588479e-01 -1.08523145e-01 -6.56349540e-01 6.87154904e-02
-1.30664778e+00 3.22798848e-01 1.12891984e+00 9.18388605e-01
4.95052725e-01 -2.46450260e-01 4.82011676e-01 -1.10148454e+00
7.59597898e-01 -2.18471140e-01 -9.23935175e-02 3.74807566e-01
-5.21743476e-01 -2.01234743e-01 4.14658248e-01 -5.65055609e-01
-8.44361007e-01 1.86381042e-01 -3.50075692e-01 -2.09670961e-01
-3.26657832e-01 6.07463419e-01 3.84517819e-01 -4.98615235e-01
1.11130536e+00 5.82293374e-03 4.88307476e-01 -2.73592681e-01
-4.59224498e-03 3.74383330e-01 2.69862890e-01 -7.80456141e-03
8.37609708e-01 5.81914246e-01 1.84506342e-01 -7.56374776e-01
-5.83427548e-01 -2.38430396e-01 -7.03202367e-01 -2.15246007e-01
9.38537717e-01 -5.89527071e-01 -1.27993196e-01 8.53713334e-01
-7.67691374e-01 -6.59414411e-01 -6.82168603e-01 2.93561548e-01
1.21660948e-01 1.98303252e-01 -5.91710985e-01 -4.12491500e-01
-4.45047081e-01 -1.03501666e+00 8.12373102e-01 3.64018649e-01
-1.98135659e-01 -1.21584427e+00 2.82897949e-02 3.96211922e-01
8.37752223e-01 8.82973194e-01 7.29650557e-01 -8.91956508e-01
-9.93270874e-02 -7.14455605e-01 -3.63933235e-01 4.15130466e-01
3.03392351e-01 2.80292511e-01 -1.20312810e+00 -4.80236530e-01
-7.54894793e-01 -5.06016314e-01 1.24924958e+00 3.96167368e-01
1.31592393e+00 2.09938595e-03 -3.40371251e-01 6.49977088e-01
1.48370218e+00 2.43208744e-02 1.04662824e+00 3.75860602e-01
4.40205872e-01 6.96974277e-01 1.19566463e-01 -1.85844824e-02
-7.60945200e-04 2.58085221e-01 5.89029014e-01 -1.03645968e+00
-5.57953954e-01 1.01717904e-01 -1.77029401e-01 2.04603568e-01
-5.13292849e-01 -2.31094912e-01 -1.16767502e+00 5.27953506e-01
-1.22041166e+00 -4.11280483e-01 -1.46975234e-01 2.26902461e+00
9.10044611e-01 2.21022934e-01 2.13981956e-01 1.34154990e-01
6.54681623e-01 -1.13662757e-01 -6.68590486e-01 -3.42171043e-01
-3.56903851e-01 6.98132217e-01 7.86165476e-01 2.13787466e-01
-1.11829507e+00 7.32519686e-01 6.79506445e+00 7.01745450e-01
-1.48154461e+00 9.81402919e-02 1.03840375e+00 -1.73128888e-01
-8.60009938e-02 -3.83043796e-01 -4.49942112e-01 3.19477946e-01
9.42402422e-01 2.49373734e-01 -5.24112023e-04 3.35984915e-01
-9.23120230e-02 -3.16659003e-01 -7.41908610e-01 5.56647539e-01
1.04169317e-01 -1.49334717e+00 1.04863748e-01 5.55646300e-01
7.61221468e-01 1.44293249e-01 2.87968457e-01 3.08547504e-02
9.72499996e-02 -1.29661179e+00 -1.35704549e-02 7.59737253e-01
1.26495922e+00 -4.51069236e-01 1.21379459e+00 -2.35081509e-01
-5.61581671e-01 7.34316409e-02 2.34358963e-02 3.34996223e-01
-1.75018460e-01 4.92543519e-01 -1.30618405e+00 5.19429266e-01
3.99542183e-01 3.73546302e-01 -1.30169392e+00 1.36891091e+00
1.49856955e-01 8.31284583e-01 -3.43454331e-01 -1.45599350e-01
3.12583089e-01 4.12008792e-01 1.83022082e-01 1.31360233e+00
1.07058167e-01 -2.58643508e-01 -2.00393334e-01 3.35974574e-01
-6.40849322e-02 2.50024408e-01 -2.87862092e-01 -1.29994556e-01
-6.40287343e-03 1.59396017e+00 -7.60330081e-01 -6.54310212e-02
-2.95259148e-01 8.49940956e-01 -2.14240700e-02 3.47138017e-01
-7.66770542e-01 -2.45039150e-01 2.31867358e-01 4.72686946e-01
-1.12391025e-01 1.73805982e-01 -3.12226653e-01 -6.67145789e-01
-1.08186133e-01 -8.47603858e-01 5.39651513e-01 -6.96310043e-01
-1.28141546e+00 6.87672496e-01 -3.60590398e-01 -1.02759326e+00
-1.79520510e-02 -8.11633825e-01 -7.12684989e-01 9.71002877e-01
-1.62496352e+00 -1.65810049e+00 -7.34131575e-01 4.20778960e-01
1.51676089e-01 -2.80945718e-01 1.00906432e+00 2.08900839e-01
-4.98246163e-01 1.08269131e+00 -1.41663641e-01 3.61354887e-01
8.91602755e-01 -1.27278018e+00 3.27189922e-01 7.67136991e-01
-1.85260624e-01 2.14151189e-01 2.87193567e-01 -6.20733678e-01
-9.46954668e-01 -1.23645914e+00 4.21081394e-01 -2.40100220e-01
6.51102245e-01 -1.63089320e-01 -7.95884490e-01 4.21029419e-01
3.65083456e-01 1.57966495e-01 1.19009650e+00 1.69350579e-01
-3.64211082e-01 -9.17829648e-02 -1.49275625e+00 6.60802901e-01
6.47714376e-01 -9.81624350e-02 2.21197322e-01 4.89491343e-01
5.37961960e-01 -5.17672598e-01 -1.08134151e+00 6.97004080e-01
5.55623412e-01 -6.57696068e-01 7.78265834e-01 -9.46802676e-01
7.87603199e-01 3.88571918e-02 6.23877794e-02 -1.39257646e+00
-1.45626619e-01 -1.66557029e-01 3.45243603e-01 9.06184971e-01
8.74748528e-01 -8.53541136e-01 1.10624754e+00 7.82187819e-01
-1.12075970e-01 -1.15430534e+00 -7.87295878e-01 -4.80164081e-01
3.04340839e-01 -1.55226111e-01 3.79486144e-01 1.15325475e+00
-4.28387076e-01 -1.97286338e-01 -1.20237157e-01 -1.19110338e-01
3.32438916e-01 -7.20537484e-01 7.55775213e-01 -9.97419298e-01
-8.87289345e-02 -5.02873838e-01 -7.74111152e-01 4.41062413e-02
-1.49013683e-01 -9.17046010e-01 -5.94657063e-01 -1.61929262e+00
3.35241169e-01 -5.50852478e-01 -4.65299875e-01 9.87314820e-01
-3.23894024e-01 8.26549232e-01 1.37427434e-01 -8.20838511e-02
-1.67337298e-01 -1.04123496e-01 1.30586946e+00 -2.62695312e-01
-2.50095338e-01 -1.77821010e-01 -7.37219453e-01 4.80776340e-01
1.08806551e+00 1.17344216e-01 -1.82582125e-01 -4.91339982e-01
1.48794994e-01 -4.33457494e-01 7.22892523e-01 -1.17830181e+00
2.22416982e-01 9.18004513e-02 9.36849594e-01 -3.97532374e-01
3.66538048e-01 -6.03868127e-01 2.32148319e-01 7.70436943e-01
-4.88866717e-01 -6.22991659e-02 8.15857232e-01 2.86291391e-01
-8.81695077e-02 -1.45195067e-01 9.52107608e-01 -5.29022738e-02
-5.57502091e-01 2.81662613e-01 -2.13436753e-01 -1.63948894e-01
1.28399074e+00 -4.73338187e-01 -6.46719813e-01 -6.07386455e-02
-1.15025198e+00 3.45287882e-02 4.63943273e-01 2.84111291e-01
3.99851799e-01 -1.17131495e+00 -9.72320199e-01 1.39957637e-01
1.56764507e-01 -2.14517623e-01 4.28562820e-01 1.14285147e+00
-7.96970487e-01 1.84511736e-01 -4.81432706e-01 -6.32633090e-01
-1.59193063e+00 1.86528891e-01 6.35742903e-01 -5.85132837e-01
-3.48296762e-01 9.88791227e-01 -1.64133489e-01 -4.13875371e-01
1.42268345e-01 -2.07614884e-01 -2.00213660e-02 -5.81412278e-02
4.29610848e-01 2.12247476e-01 4.95673001e-01 -1.35761380e-01
-1.34439707e-01 2.13896513e-01 -5.69395959e-01 -1.05191901e-01
1.25963104e+00 6.05644882e-01 -8.83744657e-02 -1.10616669e-01
9.25277591e-01 -5.07285111e-02 -1.02272499e+00 -8.93570259e-02
-2.20969483e-01 -3.68793994e-01 1.23850033e-01 -1.52923560e+00
-1.26437533e+00 8.07487726e-01 1.11782026e+00 1.67482793e-01
1.42840183e+00 -3.05315077e-01 6.03648543e-01 1.17209196e-01
-7.76880085e-02 -1.07813871e+00 7.90809840e-02 -9.37437918e-03
8.15441072e-01 -1.15647054e+00 6.30695447e-02 -3.87262374e-01
-5.97002625e-01 9.78098631e-01 8.44558597e-01 1.10827118e-01
5.75765789e-01 5.26015580e-01 2.37695456e-01 -5.27412295e-02
-7.59417236e-01 -3.26032698e-01 3.35923910e-01 8.01927149e-01
5.03319383e-01 7.77955400e-03 -3.47779185e-01 1.63013354e-01
-1.15435220e-01 1.00735858e-01 6.20870829e-01 8.56090844e-01
1.10520095e-01 -1.15070367e+00 -2.24945143e-01 1.19569719e+00
-6.55688286e-01 -2.77076453e-01 -9.29586947e-01 1.32597160e+00
3.63684118e-01 4.67346847e-01 2.24485636e-01 -4.16940719e-01
2.29752302e-01 8.87618661e-02 5.07015526e-01 -3.84530395e-01
-9.34823394e-01 -1.13482267e-01 2.36106589e-01 -2.19787940e-01
-3.68256301e-01 -5.94607770e-01 -7.86768675e-01 -2.10892096e-01
-4.06024694e-01 -1.59386396e-01 1.02973950e+00 6.87538207e-01
3.07541966e-01 8.31554592e-01 2.61947244e-01 -4.07881528e-01
-2.49444902e-01 -1.20589399e+00 -3.41245502e-01 3.27386409e-01
2.58873135e-01 -3.48112792e-01 -2.67620146e-01 2.10702911e-01] | [15.549935340881348, -2.8469510078430176] |
cb80a615-560e-4726-87b2-48a152b74a22 | why-we-should-report-the-details-in | 2306.02044 | null | https://arxiv.org/abs/2306.02044v1 | https://arxiv.org/pdf/2306.02044v1.pdf | Why We Should Report the Details in Subjective Evaluation of TTS More Rigorously | This paper emphasizes the importance of reporting experiment details in subjective evaluations and demonstrates how such details can significantly impact evaluation results in the field of speech synthesis. Through an analysis of 80 papers presented at INTERSPEECH 2022, we find a lack of thorough reporting on critical details such as evaluator recruitment and filtering, instructions and payments, and the geographic and linguistic backgrounds of evaluators. To illustrate the effect of these details on evaluation outcomes, we conducted mean opinion score (MOS) tests on three well-known TTS systems under different evaluation settings and we obtain at least three distinct rankings of TTS models. We urge the community to report experiment details in subjective evaluations to improve the reliability and interpretability of experimental results. | ['Hung-Yi Lee', 'Wei-Ping Huang', 'Cheng-Han Chiang'] | 2023-06-03 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 6.10176735e-02 -1.67774707e-01 -4.93669920e-02 -6.28500760e-01
-1.31004226e+00 -9.84773517e-01 4.67522442e-01 2.88735423e-03
-4.98409599e-01 7.67775536e-01 5.19759476e-01 -7.37337649e-01
1.80458948e-02 2.10766256e-01 -2.58204639e-01 -2.57929295e-01
-3.74121740e-02 1.30592659e-01 1.99745938e-01 -3.44782442e-01
4.25771892e-01 3.77281815e-01 -1.40320230e+00 2.98167974e-01
5.66577971e-01 8.48651290e-01 1.41123738e-02 9.00737286e-01
2.28995904e-01 4.32451248e-01 -1.58644426e+00 -7.71843433e-01
-4.33215313e-02 -3.62112105e-01 -7.18443573e-01 1.81168355e-02
3.51438999e-01 -1.88924775e-01 -1.90686867e-01 9.38956976e-01
1.15643263e+00 7.42479637e-02 4.89045352e-01 -1.17838478e+00
-5.15088201e-01 7.68663406e-01 3.95054705e-02 5.83873868e-01
5.01663387e-01 3.37995082e-01 1.11716104e+00 -8.27779233e-01
3.30157071e-01 1.19192719e+00 5.01168847e-01 3.07782561e-01
-9.59835529e-01 -7.21605420e-01 1.47755742e-01 1.16885006e-01
-1.35030985e+00 -1.10656822e+00 3.63763630e-01 -3.63024265e-01
8.75234544e-01 5.98855495e-01 2.96250284e-01 1.15110290e+00
-2.38214552e-01 6.36974573e-01 1.05493808e+00 -5.84579051e-01
2.80744255e-01 3.11786771e-01 1.35859713e-01 1.95397601e-01
5.49066849e-02 1.85476691e-01 -7.55045474e-01 -2.58564204e-01
4.03833419e-01 -1.03830075e+00 -4.16292429e-01 4.65269163e-02
-1.28006828e+00 5.41110218e-01 -1.72753930e-01 4.14394438e-01
3.73442993e-02 1.40668362e-01 6.90090001e-01 5.68456829e-01
3.04048389e-01 6.34576738e-01 -3.33805203e-01 -6.83957338e-01
-9.34287846e-01 2.12535799e-01 8.59829485e-01 1.04839957e+00
-8.38256404e-02 4.63668048e-01 -3.54175240e-01 1.12317050e+00
2.22576484e-01 6.95128977e-01 3.34332496e-01 -1.44375873e+00
7.55074024e-01 -2.81294107e-01 6.41206324e-01 -7.00348556e-01
6.44132420e-02 -5.55174112e-01 1.07830957e-01 1.89871088e-01
1.82977930e-01 -7.07364500e-01 -9.34468865e-01 1.74890804e+00
-4.31807905e-01 -5.74364364e-01 -9.24786404e-02 8.19514751e-01
7.21488237e-01 4.40637738e-01 1.54759109e-01 -5.28945088e-01
1.23158848e+00 -7.27266908e-01 -1.17922854e+00 -2.04203218e-01
5.26739657e-01 -1.46802557e+00 1.41703963e+00 3.97177100e-01
-1.54648709e+00 -4.85991329e-01 -1.33574557e+00 3.77919316e-01
-1.92330554e-01 3.64416867e-01 1.97037920e-01 1.23831069e+00
-1.18090868e+00 4.57513571e-01 -6.12026572e-01 -4.74564761e-01
-3.06034207e-01 5.13716221e-01 7.51629695e-02 4.85444367e-01
-1.37999427e+00 1.09726560e+00 -2.14934483e-01 1.80432037e-01
-3.80604863e-01 -2.26065800e-01 -4.84981835e-01 -1.72205437e-02
2.19127193e-01 -2.33582497e-01 2.03358698e+00 -5.58030784e-01
-1.62701726e+00 6.27284586e-01 -3.70981187e-01 -2.25442071e-02
3.72866124e-01 -7.79831558e-02 -1.14126456e+00 -2.66335290e-02
-1.21092061e-02 3.57297182e-01 5.14401853e-01 -1.16081190e+00
-9.15090442e-01 1.97867498e-01 -8.68125409e-02 4.67138618e-01
-2.15967670e-01 6.54708147e-01 -5.31586826e-01 -8.80208731e-01
-9.87682194e-02 -9.39094067e-01 9.64905396e-02 -6.52973592e-01
-2.75295079e-01 -1.08256936e-01 3.64573687e-01 -6.83481038e-01
1.73544860e+00 -2.41293645e+00 -2.58404195e-01 1.33827358e-01
-1.13022570e-02 3.84492189e-01 -8.67589563e-02 5.50324738e-01
2.93495934e-02 5.35782754e-01 2.80100495e-01 -5.72838962e-01
4.49532568e-01 -1.84076771e-01 -3.78494591e-01 3.07550728e-01
-1.39161333e-01 5.57024539e-01 -6.69828832e-01 -4.27097708e-01
3.01957846e-01 2.42357865e-01 -3.17616343e-01 1.84744731e-01
2.89459169e-01 2.34201625e-01 -3.16960663e-01 7.30713129e-01
6.91271201e-02 6.86333776e-02 5.73074855e-02 -2.59197533e-01
-4.46959078e-01 9.97676253e-01 -1.05137813e+00 1.24199402e+00
-3.67801219e-01 1.15244830e+00 1.57808080e-01 -4.25476253e-01
5.77994108e-01 8.48222554e-01 1.02455700e-02 -6.66053712e-01
2.50457525e-01 7.37385273e-01 3.83685946e-01 -3.01449627e-01
8.32598984e-01 -4.10203226e-02 -6.07480928e-02 4.48154271e-01
-5.20566553e-02 -4.19331580e-01 4.30336326e-01 1.69017851e-01
5.99549115e-01 -5.84143877e-01 -1.63508460e-01 -3.09278190e-01
1.74248248e-01 -2.36652538e-01 2.31713653e-01 8.07706773e-01
-6.61253333e-01 7.08552837e-01 2.01236695e-01 2.08818421e-01
-9.74484980e-01 -1.13341737e+00 -2.17695847e-01 1.00664556e+00
-1.49075210e-01 -6.80657625e-01 -7.65491545e-01 -2.73486435e-01
-4.37967986e-01 9.86790419e-01 -2.36340672e-01 1.41065583e-01
-4.13086414e-01 -5.18550217e-01 7.90428162e-01 5.82958162e-01
-4.05196361e-02 -9.64276671e-01 -3.41478765e-01 4.41039465e-02
-6.32481515e-01 -1.19374132e+00 -7.68254280e-01 2.88592875e-01
-3.81192118e-01 -5.36131263e-01 -9.74113107e-01 -9.28749979e-01
1.95665315e-01 1.19737521e-01 8.72253716e-01 -1.04856454e-01
2.02604800e-01 4.59247977e-01 -3.16632390e-01 -5.48125148e-01
-8.21375251e-01 2.31544435e-01 2.15462700e-01 -5.16971409e-01
2.66828775e-01 -4.01821911e-01 -5.62118292e-01 6.97533727e-01
-2.39156470e-01 -5.13669431e-01 4.16227221e-01 4.21324879e-01
1.84036456e-02 4.23292965e-02 6.55456781e-01 -5.49227715e-01
1.30132246e+00 3.18453796e-02 -5.02504706e-01 4.79387969e-01
-7.62768328e-01 -1.18008263e-01 1.43031001e-01 -3.02716911e-01
-9.32429850e-01 -6.30711555e-01 -1.91336289e-01 1.51141152e-01
2.83272844e-02 5.32394826e-01 1.19051887e-02 -1.33041054e-01
6.09545350e-01 -1.89703658e-01 -3.78404438e-01 -2.71345854e-01
-6.95175305e-02 1.19046271e+00 5.29700994e-01 -4.94052798e-01
6.23851955e-01 -3.18438560e-01 -6.51068926e-01 -7.02338696e-01
-2.73233175e-01 -3.65295470e-01 2.84227245e-02 -2.38950953e-01
4.87940729e-01 -9.50687587e-01 -6.01482332e-01 3.34756732e-01
-1.01038635e+00 -2.87524223e-01 -6.83536679e-02 9.90998149e-01
-5.36644220e-01 2.98308015e-01 -7.92777300e-01 -9.65584278e-01
1.20113425e-01 -1.67166996e+00 8.69264722e-01 2.31007338e-02
-9.82363582e-01 -9.77149785e-01 -5.09912446e-02 6.34384155e-01
5.95060706e-01 -3.87904048e-01 7.27107108e-01 -5.12052894e-01
-1.52934417e-01 -2.84816414e-01 1.74589068e-01 4.75164592e-01
1.77631482e-01 4.38543350e-01 -1.18366957e+00 -1.70694992e-01
-1.44478753e-01 -3.71766776e-01 1.34852588e-01 7.53327310e-01
6.38695419e-01 1.23412227e-02 -1.30068243e-01 -9.39953327e-02
8.64837766e-01 5.33521533e-01 4.05218780e-01 2.03070223e-01
1.75709739e-01 8.28757584e-01 6.51467144e-01 4.08336490e-01
4.43781950e-02 8.19488883e-01 -2.93726593e-01 -1.33134827e-01
-1.46706298e-01 -1.03456341e-01 5.49348474e-01 1.28360152e+00
-5.15515283e-02 -8.98765624e-01 -5.58837593e-01 5.27049482e-01
-1.29498839e+00 -8.31072927e-01 1.01794995e-01 2.35302615e+00
7.62940824e-01 4.45362210e-01 4.49375927e-01 4.53788877e-01
1.14297140e+00 3.14056426e-01 -5.17407246e-02 -9.58816946e-01
-2.96804041e-01 3.73919122e-02 6.25395954e-01 8.61026704e-01
-6.34018362e-01 6.68738186e-01 8.91172886e+00 9.27452564e-01
-1.05021083e+00 -1.76999688e-01 6.27407730e-01 -2.64396042e-01
-4.63496029e-01 -9.78242680e-02 -5.74172139e-01 3.40514660e-01
1.19262314e+00 -5.41893542e-01 4.34558123e-01 2.48160586e-01
6.87207878e-01 -3.13977078e-02 -1.11361611e+00 7.98745334e-01
-1.71174988e-01 -1.02771592e+00 -3.18552911e-01 4.22578827e-02
4.91930097e-01 9.21560749e-02 5.31221211e-01 3.31722289e-01
2.71034300e-01 -9.78888333e-01 1.20519245e+00 -2.65239656e-01
9.03452039e-01 -4.83556926e-01 5.58002055e-01 -2.84139067e-01
-9.58574712e-01 -3.98042239e-02 9.54915956e-02 -1.40159816e-01
6.05454862e-01 2.20947102e-01 -7.97243536e-01 3.51608753e-01
4.55389261e-01 -8.84773210e-02 -3.50136966e-01 1.01381254e+00
-2.01168999e-01 1.14057875e+00 -1.56315163e-01 -1.40607744e-01
8.55037421e-02 2.63969749e-01 5.76235771e-01 1.30550063e+00
3.43598634e-01 5.33288941e-02 -2.53555626e-01 4.03112203e-01
-5.09245358e-02 1.90329283e-01 -2.18421638e-01 -4.86650288e-01
8.98785174e-01 7.67122269e-01 -6.82561815e-01 -2.88995326e-01
-5.90337396e-01 5.96497416e-01 -4.19751704e-01 7.90664971e-01
-8.01305294e-01 -6.61946893e-01 5.51071405e-01 6.86051622e-02
2.30320632e-01 -3.07879120e-01 -5.03838003e-01 -7.44117916e-01
1.25915468e-01 -1.24757493e+00 8.53840709e-02 -9.69491601e-01
-8.30083728e-01 7.91383147e-01 2.08335504e-01 -1.13765061e+00
-3.25938791e-01 -3.75143170e-01 -3.95899743e-01 1.16590834e+00
-8.97750318e-01 -3.08556646e-01 2.98768461e-01 6.29694462e-02
6.46790206e-01 -3.70222896e-01 6.54841959e-01 7.48164296e-01
-4.50828612e-01 1.11325204e+00 9.47456285e-02 -1.74568817e-01
1.19741702e+00 -1.15586925e+00 8.22408617e-01 6.71453536e-01
1.35455266e-01 8.61350238e-01 1.26659584e+00 -3.16249281e-01
-8.32544029e-01 -3.70638788e-01 1.29352820e+00 -6.36835217e-01
6.86858475e-01 -3.49384874e-01 -2.99364060e-01 5.68340659e-01
5.37099719e-01 -7.44153678e-01 7.49420822e-01 4.93184745e-01
1.85069591e-01 8.39018598e-02 -6.44860744e-01 1.13497972e+00
9.26639438e-01 -7.47176349e-01 -4.24759299e-01 1.10029094e-01
7.21022844e-01 -5.25284111e-01 -8.31412673e-01 4.86905634e-01
8.15359175e-01 -8.90674055e-01 6.42326951e-01 -2.34958723e-01
-1.35978192e-01 -2.07867071e-01 -4.63991046e-01 -1.34706545e+00
-1.65170282e-01 -9.68689919e-01 4.90757644e-01 1.39713216e+00
1.01937640e+00 -5.09023845e-01 4.80178505e-01 8.31132114e-01
-4.73163038e-01 -4.31280375e-01 -5.89754224e-01 -8.37638617e-01
1.24258241e-02 -6.05639458e-01 4.38192815e-01 5.66976309e-01
3.47657710e-01 5.11002839e-01 -3.60262483e-01 1.20885596e-01
4.45427597e-02 -3.10721785e-01 5.41426778e-01 -6.31320953e-01
-2.73926497e-01 -5.24113595e-01 -2.33394593e-01 -1.06809556e+00
-1.77511454e-01 -1.74857482e-01 1.32952496e-01 -1.16094530e+00
-1.65693238e-01 -2.10421965e-01 -3.80895585e-01 -1.34568438e-01
-1.24479964e-01 4.00724053e-01 3.31602484e-01 7.49809891e-02
-4.96372789e-01 2.59733438e-01 1.04626656e+00 -3.17439325e-02
-2.05721885e-01 3.67231011e-01 -8.25264156e-01 4.43117291e-01
7.45773852e-01 -2.11399913e-01 -7.55157053e-01 -6.62239671e-01
-2.83004567e-02 1.60990164e-01 -8.20520371e-02 -9.43462372e-01
1.77578643e-01 -2.10091740e-01 8.59967694e-02 -5.11000514e-01
5.01635849e-01 -4.45544958e-01 5.62121160e-02 3.34212095e-01
-8.19220901e-01 5.52815199e-01 4.04845297e-01 1.33734375e-01
-5.39772689e-01 -2.76680827e-01 3.21765006e-01 1.81038871e-01
-5.36163032e-01 -1.78401694e-01 -1.01620877e+00 2.88280249e-01
4.93326634e-01 -4.38042104e-01 -1.12754188e-01 -1.01498103e+00
-7.47951448e-01 1.64967954e-01 3.75223845e-01 4.89469796e-01
1.39796818e-02 -1.13674355e+00 -5.45532286e-01 -1.63255364e-01
1.93691179e-01 -9.16414022e-01 6.75250683e-03 7.47309089e-01
-2.91338414e-01 7.04336405e-01 1.45282328e-01 -3.14203888e-01
-1.44843185e+00 -1.08822770e-01 2.70898581e-01 8.65116939e-02
3.37555744e-02 7.47685194e-01 1.27354696e-01 -6.58300659e-03
8.51490915e-01 -4.22337979e-01 1.31270266e-03 -2.97445040e-02
1.67985529e-01 5.26500106e-01 4.83313620e-01 -6.05240226e-01
-5.21830678e-01 2.57035077e-01 -1.39873503e-02 -1.16844547e+00
4.60675091e-01 -6.69314325e-01 6.05391622e-01 9.29584563e-01
1.04676628e+00 5.69346964e-01 -8.11715841e-01 1.63263410e-01
-2.71452963e-02 -3.80434275e-01 2.14794464e-02 -1.12624407e+00
-6.95699036e-01 9.04629290e-01 6.07711136e-01 3.37416232e-01
9.59868312e-01 -2.77216136e-01 6.92936957e-01 2.24065587e-01
1.30705416e-01 -1.57099676e+00 -9.12712589e-02 4.22060817e-01
9.36138511e-01 -1.00890338e+00 -2.06767976e-01 -5.23878992e-01
-9.04421806e-01 7.88904250e-01 1.72466636e-01 5.87407827e-01
3.88704747e-01 3.86444271e-01 6.10928476e-01 1.10962704e-01
-8.34518671e-01 9.73052382e-02 1.43215582e-01 6.14277244e-01
1.06450534e+00 1.76859587e-01 -7.10932434e-01 4.67365354e-01
-6.05569124e-01 -2.22908705e-01 4.60067600e-01 9.28038061e-01
-2.89651275e-01 -1.52405465e+00 -4.28901583e-01 1.94865808e-01
-7.61246443e-01 -2.21946344e-01 -6.98790669e-01 7.01708674e-01
-3.66361231e-01 1.57515991e+00 -3.20620030e-01 -5.73880732e-01
6.81271970e-01 8.81565586e-02 4.37685698e-01 -5.95995069e-01
-7.33444452e-01 4.12338108e-01 8.65102768e-01 -1.84751809e-01
-3.12658638e-01 -8.14146280e-01 -8.03637505e-01 -6.63036227e-01
-6.32322729e-01 7.32283175e-01 9.06232655e-01 9.48686600e-01
1.99428141e-01 7.01818943e-01 7.25202799e-01 -4.39733297e-01
-9.82214153e-01 -1.02908731e+00 -6.89832211e-01 1.06314570e-01
3.95364463e-01 -5.12160957e-01 -6.37125552e-01 8.73908550e-02] | [14.775620460510254, 6.507643699645996] |
90b675bf-c33b-48c7-9d42-818606e9dedd | dependency-grammar-induction-with-a-neural | 1811.05889 | null | http://arxiv.org/abs/1811.05889v1 | http://arxiv.org/pdf/1811.05889v1.pdf | Dependency Grammar Induction with a Neural Variational Transition-based Parser | Dependency grammar induction is the task of learning dependency syntax
without annotated training data. Traditional graph-based models with global
inference achieve state-of-the-art results on this task but they require
$O(n^3)$ run time. Transition-based models enable faster inference with $O(n)$
time complexity, but their performance still lags behind. In this work, we
propose a neural transition-based parser for dependency grammar induction,
whose inference procedure utilizes rich neural features with $O(n)$ time
complexity. We train the parser with an integration of variational inference,
posterior regularization and variance reduction techniques. The resulting
framework outperforms previous unsupervised transition-based dependency parsers
and achieves performance comparable to graph-based models, both on the English
Penn Treebank and on the Universal Dependency Treebank. In an empirical
comparison, we show that our approach substantially increases parsing speed
over graph-based models. | ['Frank Keller', 'Yang Liu', 'Jianpeng Cheng', 'Bowen Li'] | 2018-11-14 | null | null | null | null | ['dependency-grammar-induction'] | ['natural-language-processing'] | [ 8.63995254e-02 7.91216969e-01 -3.47704321e-01 -6.09695494e-01
-1.41464019e+00 -4.96968716e-01 8.88673514e-02 3.87584180e-01
-5.25134981e-01 7.77467132e-01 3.81175918e-03 -1.00606847e+00
9.89247635e-02 -9.33707774e-01 -8.63479376e-01 -4.22911823e-01
-4.97217000e-01 8.82633150e-01 1.83999151e-01 -7.33201578e-02
7.52716660e-02 1.60667315e-01 -9.54234064e-01 -1.54352769e-01
8.66269171e-01 5.58619142e-01 1.61916465e-01 1.07471776e+00
-4.93783444e-01 9.80880141e-01 -2.48843044e-01 -7.40392327e-01
-1.89233303e-01 -5.09687543e-01 -1.05175769e+00 -5.06991863e-01
3.96572828e-01 -8.83932933e-02 -2.64978021e-01 1.23519385e+00
2.64083147e-01 1.98207423e-01 2.82662690e-01 -5.86208344e-01
-5.07310927e-01 1.35673642e+00 -1.97057053e-01 3.80424052e-01
5.21832667e-02 -2.84117252e-01 1.70043385e+00 -4.38665301e-01
6.10227406e-01 1.55199730e+00 6.48808360e-01 6.58552468e-01
-1.58012223e+00 -5.89136302e-01 5.66858172e-01 -1.72719061e-01
-8.72598410e-01 -3.05871189e-01 7.92860389e-01 -1.88377857e-01
1.53298223e+00 -2.50845641e-01 2.69082040e-01 8.66794169e-01
1.09269537e-01 8.85529876e-01 8.39055121e-01 -8.61548126e-01
2.72173703e-01 -4.30946380e-01 6.78473175e-01 1.35958552e+00
2.89006770e-01 -1.04459040e-01 -4.46737707e-01 -1.72400766e-03
7.34229028e-01 -6.76868498e-01 2.66001940e-01 7.05705136e-02
-4.82082456e-01 1.38748121e+00 1.14052609e-01 2.84948945e-01
-1.04636133e-01 6.72163904e-01 5.76118112e-01 2.46850744e-01
7.42340505e-01 5.84825017e-02 -9.90070522e-01 -2.12141365e-01
-9.00876164e-01 7.26656467e-02 1.21197855e+00 1.06300044e+00
8.89819503e-01 2.69135445e-01 2.35186726e-01 8.42165232e-01
5.38040817e-01 4.71201867e-01 7.25240931e-02 -1.09569156e+00
8.68689835e-01 3.19257826e-01 -3.13860685e-01 -3.07295978e-01
-6.48975074e-01 -1.20248713e-01 -5.77500463e-01 1.23067714e-01
6.48730934e-01 -6.37094676e-01 -1.03204179e+00 2.07408237e+00
1.98655963e-01 -1.33643299e-01 1.16457358e-01 8.80857334e-02
5.40372431e-01 5.99319398e-01 4.65662390e-01 -4.06267077e-01
1.27207541e+00 -7.36840963e-01 -7.43440568e-01 -4.30829227e-01
1.18912101e+00 -2.88426518e-01 9.20365691e-01 3.52151662e-01
-1.37976348e+00 -3.13247919e-01 -6.86822236e-01 -1.84089825e-01
-1.61442891e-01 -1.26909302e-03 1.06104052e+00 1.02590442e+00
-1.20611012e+00 8.05934966e-01 -1.38176250e+00 -6.30763099e-02
3.84929657e-01 7.12707043e-01 -2.61877090e-01 -9.28842947e-02
-1.10695732e+00 8.69738758e-01 4.49529976e-01 1.41419217e-01
-5.49599290e-01 -4.64408040e-01 -1.42966723e+00 -5.06652966e-02
3.43153924e-01 -4.11195099e-01 1.44887888e+00 -3.92081767e-01
-1.94657826e+00 5.95053852e-01 -3.68754089e-01 -7.03432739e-01
-2.16083214e-01 -3.80889386e-01 1.39570147e-01 -6.36620354e-03
6.69269487e-02 4.64493215e-01 5.63595414e-01 -8.57975662e-01
-6.51832581e-01 -4.66750801e-01 2.01651901e-01 -2.15947717e-01
1.97061449e-01 2.48795137e-01 -3.71582985e-01 -1.56384647e-01
1.53126672e-01 -8.83103549e-01 -6.27927363e-01 -8.50639343e-01
-6.25815213e-01 -7.91390717e-01 3.10966253e-01 -9.44926143e-01
1.29133141e+00 -1.68377376e+00 1.46245167e-01 7.77363330e-02
-7.39896744e-02 1.20520733e-01 -1.98084548e-01 2.59537280e-01
1.58724904e-01 3.96990776e-01 -6.63199246e-01 -6.38053358e-01
2.02858716e-01 7.71068513e-01 1.53130919e-01 1.57701090e-01
4.34002161e-01 9.72239852e-01 -9.34751749e-01 -8.05958152e-01
2.94518215e-03 3.09898764e-01 -1.03387082e+00 1.30490318e-01
-6.46182954e-01 2.84711719e-01 -5.68754494e-01 5.96077919e-01
3.84097070e-01 -1.37810215e-01 1.01984012e+00 3.88936698e-01
-8.21342170e-02 8.31980824e-01 -8.00685227e-01 2.06377411e+00
-9.37292755e-01 4.42914635e-01 2.84849048e-01 -1.45823586e+00
8.09713185e-01 1.74216971e-01 -4.90872301e-02 -3.75859171e-01
6.53908700e-02 1.78782880e-01 2.71140009e-01 -4.28837001e-01
8.25945735e-02 -3.24529350e-01 -5.24369419e-01 3.82017553e-01
6.53309464e-01 -2.20018476e-01 5.45600295e-01 2.50857532e-01
1.42545104e+00 7.03365505e-01 1.40024349e-01 -3.91960472e-01
1.92568421e-01 -1.04463018e-01 7.55262733e-01 9.48705256e-01
5.01812957e-02 -3.70497592e-02 1.02998471e+00 -3.72681528e-01
-6.77613378e-01 -1.08551633e+00 -7.23871067e-02 1.54231131e+00
-5.62976420e-01 -5.19077063e-01 -1.12703717e+00 -1.12463248e+00
-3.89790565e-01 9.81172919e-01 -6.71165287e-01 4.19569254e-01
-1.36932790e+00 -1.03599453e+00 6.25920594e-01 7.98158824e-01
1.03793889e-01 -1.34140861e+00 -4.14554715e-01 6.77726746e-01
-1.20930336e-01 -1.37199700e+00 -1.36131616e-02 9.48484659e-01
-1.33653951e+00 -8.23178828e-01 1.33307680e-01 -1.11095130e+00
6.49937272e-01 -9.42476332e-01 1.56049860e+00 -3.69670801e-02
-1.88493058e-01 6.32871985e-02 -3.42594475e-01 -3.63897234e-01
-7.44606674e-01 4.03571904e-01 -2.30011895e-01 -6.83404863e-01
3.95695716e-01 -7.73216248e-01 -4.56527211e-02 -6.07722402e-01
-4.38306153e-01 -3.67887288e-01 6.23874009e-01 1.03096759e+00
4.82821018e-01 -1.61197037e-01 4.68345255e-01 -1.52047253e+00
4.86067951e-01 -2.12853536e-01 -1.01243424e+00 4.51767445e-02
-6.74917459e-01 6.91263497e-01 8.13937008e-01 8.75870660e-02
-1.25884187e+00 3.91114354e-01 -7.51202941e-01 1.59866050e-01
-2.57852793e-01 8.09917927e-01 -1.46906525e-01 1.76820084e-01
3.55575383e-01 -1.35839656e-01 -3.21135610e-01 -7.67234325e-01
5.23239315e-01 8.32164064e-02 3.45085979e-01 -8.71127605e-01
5.53367436e-01 -2.50074407e-03 2.92429179e-01 -7.48769343e-01
-1.07333708e+00 -2.86688674e-02 -1.13615799e+00 4.15594608e-01
1.18953359e+00 -7.06508577e-01 -5.46138763e-01 -9.21227597e-03
-1.38557553e+00 -8.31832528e-01 -1.28695861e-01 5.60696721e-01
-5.63390970e-01 4.99421120e-01 -1.15987384e+00 -1.12122178e+00
-3.31310034e-01 -9.15933967e-01 9.76389706e-01 -6.07150123e-02
-6.26156777e-02 -1.45126641e+00 3.82328957e-01 1.57246739e-01
-2.93426178e-02 8.40042159e-02 1.38960493e+00 -7.72178292e-01
-5.66754162e-01 -1.82749704e-01 -2.30144057e-02 3.74015063e-01
-1.44046918e-01 -4.14284468e-02 -7.80361652e-01 6.33160472e-02
-2.92329222e-01 -2.80510187e-01 1.06571352e+00 8.89296234e-01
7.45933712e-01 -2.32666880e-01 -2.60902107e-01 5.72090268e-01
1.55923712e+00 1.25701681e-01 4.11631197e-01 -3.82830463e-02
9.17935073e-01 6.30384862e-01 3.06742579e-01 1.13695517e-01
4.98275578e-01 1.23728417e-01 4.08531815e-01 2.07750887e-01
4.16386239e-02 -3.05367649e-01 5.49658954e-01 1.24467003e+00
-3.39902312e-01 -1.35065198e-01 -9.59202170e-01 7.15511143e-01
-1.78400326e+00 -5.48174262e-01 -3.00093085e-01 1.88592231e+00
9.47968185e-01 5.68996191e-01 3.23042162e-02 1.02635071e-01
6.30589128e-01 1.04773916e-01 -1.50709718e-01 -1.24595690e+00
3.42350334e-01 1.27670443e+00 6.38611853e-01 1.07807386e+00
-1.18260837e+00 1.53482354e+00 6.88010311e+00 5.11061192e-01
-4.94028270e-01 3.48675221e-01 5.38077295e-01 2.07273573e-01
-1.75322697e-01 3.88392478e-01 -1.12705863e+00 -1.48709621e-02
1.67797720e+00 4.02773470e-01 3.23500186e-01 8.61581028e-01
-1.60146892e-01 -9.24268365e-02 -1.17290342e+00 3.71923834e-01
-2.08850503e-01 -1.24467909e+00 -2.86803812e-01 9.30460840e-02
3.29047382e-01 4.28858876e-01 -3.67063314e-01 6.24627173e-01
1.27100754e+00 -1.06192553e+00 1.85742378e-01 -8.19931179e-02
6.34012043e-01 -9.81385410e-01 7.29484379e-01 4.47785676e-01
-1.29905117e+00 -2.50934213e-02 -4.58926231e-01 -3.38338107e-01
5.58776081e-01 7.06749380e-01 -5.52742600e-01 5.60380697e-01
6.32531345e-01 2.36205429e-01 -1.72723547e-01 4.76397425e-02
-8.24091434e-01 1.21963787e+00 -5.48457444e-01 -2.90437728e-01
4.59642321e-01 -3.25840265e-01 1.24155812e-01 1.67414021e+00
-5.31557016e-02 1.64297849e-01 2.12582797e-01 6.42964005e-01
-3.25642377e-01 2.00841442e-01 -5.86152196e-01 -6.79494664e-02
2.51091301e-01 1.03574169e+00 -7.74676800e-01 -3.42908472e-01
-6.47431672e-01 7.37822413e-01 1.00974274e+00 8.86908174e-02
-7.32889593e-01 -6.11787021e-01 2.52982706e-01 -4.03333485e-01
8.73817623e-01 -6.00706518e-01 -2.08343551e-01 -9.72001255e-01
-1.89882383e-01 -3.90172720e-01 4.86744672e-01 -2.85600312e-03
-1.12514162e+00 5.72839558e-01 2.27997988e-01 -1.73886903e-02
-7.45556056e-01 -1.03837347e+00 -5.24114966e-01 8.57475936e-01
-1.55864084e+00 -9.98021185e-01 5.74552953e-01 1.30941018e-01
6.00625336e-01 1.36622950e-01 1.33866978e+00 -9.09457430e-02
-7.62663782e-01 6.45528316e-01 -1.68716624e-01 7.18003094e-01
-6.59579262e-02 -1.77742231e+00 1.11566222e+00 1.08850169e+00
5.53091645e-01 5.58279634e-01 5.96310258e-01 -5.09786248e-01
-1.45273054e+00 -1.04526734e+00 1.38013721e+00 -5.86511433e-01
8.17029774e-01 -5.81318259e-01 -8.44585776e-01 1.17547238e+00
2.52290010e-01 2.00356692e-01 7.43871093e-01 8.05926442e-01
-4.83843178e-01 2.07755029e-01 -7.34727621e-01 3.39579403e-01
1.20712996e+00 -4.83311862e-01 -6.96968973e-01 3.12611580e-01
8.53615463e-01 -2.19208017e-01 -1.08945322e+00 1.99861109e-01
4.25698578e-01 -9.40564334e-01 6.61729038e-01 -6.65401101e-01
3.71396363e-01 4.63152438e-01 -1.46185428e-01 -1.13234937e+00
-3.86360914e-01 -9.95624065e-01 -2.82839477e-01 1.08039308e+00
1.01146519e+00 -6.20052338e-01 1.07343519e+00 4.33153123e-01
-2.56312251e-01 -6.65654898e-01 -1.14164650e+00 -6.45215571e-01
8.31784129e-01 -6.65167928e-01 1.29229063e-02 6.45665824e-01
3.14037263e-01 7.22083390e-01 9.77256820e-02 1.80486321e-01
9.49040294e-01 4.93664853e-02 4.20043290e-01 -1.42808306e+00
-5.34103453e-01 -2.42683992e-01 -4.47788537e-02 -1.01393139e+00
9.57910717e-01 -1.03511989e+00 5.14992595e-01 -1.66797411e+00
7.05950288e-03 -4.81247872e-01 -2.26587281e-01 6.07180417e-01
-3.31251115e-01 -1.26485810e-01 -1.15467094e-01 -5.03724158e-01
-4.35030758e-01 1.41941281e-02 8.03617418e-01 1.97999209e-01
-2.36294746e-01 -1.54887989e-01 -5.19358575e-01 1.03237820e+00
1.07039201e+00 -9.27484095e-01 -2.57630557e-01 -6.45544112e-01
3.21873367e-01 4.83534694e-01 -2.05919310e-01 -6.16433322e-01
7.99440071e-02 1.63472936e-01 -2.07550645e-01 -3.78692120e-01
1.27386212e-01 -1.79178566e-02 -6.24442160e-01 5.18635571e-01
-3.85153383e-01 2.03013048e-01 2.79737592e-01 6.01204813e-01
-1.21750936e-01 -6.45610631e-01 6.80187404e-01 -5.75218976e-01
-2.81092942e-01 3.63931417e-01 -6.49102390e-01 5.00091374e-01
1.92918345e-01 3.37918669e-01 1.37782454e-01 -1.21160612e-01
-1.08799839e+00 -1.46250829e-01 -2.03665897e-01 -2.05710039e-01
2.68939674e-01 -5.47474146e-01 -8.73966634e-01 9.01623368e-02
-4.95575935e-01 3.97857755e-01 -1.04329608e-01 5.13496459e-01
-5.75721562e-01 5.28607607e-01 3.01713496e-01 -4.61844146e-01
-1.04686582e+00 3.08304667e-01 -7.27712223e-03 -1.09316707e+00
-6.99600399e-01 1.41275156e+00 -1.88965481e-02 -7.72212982e-01
9.10509080e-02 -7.99296856e-01 -2.04777960e-02 -1.43817887e-01
3.30380276e-02 -9.11137760e-02 -3.93772870e-02 -2.55215675e-01
-4.38760012e-01 5.01251400e-01 -8.67121443e-02 -4.10333395e-01
1.31681073e+00 2.01819420e-01 -2.47683540e-01 5.12541294e-01
1.13129723e+00 1.39542252e-01 -1.22330821e+00 -4.53851335e-02
6.18144691e-01 3.28228325e-01 1.65746957e-01 -5.59760273e-01
-8.07703197e-01 1.18446577e+00 8.27156752e-02 3.01220387e-01
7.87374735e-01 2.93935597e-01 8.36524963e-01 8.21867883e-01
4.01653916e-01 -1.19064963e+00 -3.87901485e-01 9.43542302e-01
5.80563843e-02 -1.21799886e+00 -1.20712839e-01 -6.67868137e-01
-2.11958602e-01 1.13529837e+00 2.39272639e-01 -6.78108394e-01
8.48213255e-01 7.44394243e-01 2.38740444e-02 -6.35530129e-02
-9.31405604e-01 -4.90398228e-01 -2.58483380e-01 7.13521063e-01
8.56029928e-01 1.99320331e-01 -3.18854988e-01 7.16368496e-01
-2.77524114e-01 -3.22483778e-01 2.94208556e-01 1.12604225e+00
-5.94889581e-01 -1.62721407e+00 2.00605646e-01 2.77522802e-01
-1.11543024e+00 -8.43140244e-01 1.75961982e-02 1.04806161e+00
-2.77982891e-01 9.99587238e-01 6.10136762e-02 9.57881436e-02
4.87612858e-02 7.21072674e-01 1.09994328e+00 -9.76573706e-01
-5.65406561e-01 4.18049581e-02 7.67408431e-01 -6.20393693e-01
-4.93339777e-01 -8.21419239e-01 -1.61297381e+00 2.91329697e-02
-4.94904399e-01 4.22640026e-01 9.48738158e-01 1.14029205e+00
-3.46934539e-03 5.44730783e-01 2.94769347e-01 -7.81716883e-01
-5.80275416e-01 -1.06177771e+00 -5.29340684e-01 -1.06542774e-01
1.26643971e-01 -2.94434845e-01 -4.33026075e-01 2.64356941e-01] | [10.369688034057617, 9.639059066772461] |
940fa26e-265b-4c85-85f6-973222348f18 | transformer-based-language-models-for | 2204.03214 | null | https://arxiv.org/abs/2204.03214v2 | https://arxiv.org/pdf/2204.03214v2.pdf | Transformer-Based Language Models for Software Vulnerability Detection | The large transformer-based language models demonstrate excellent performance in natural language processing. By considering the transferability of the knowledge gained by these models in one domain to other related domains, and the closeness of natural languages to high-level programming languages, such as C/C++, this work studies how to leverage (large) transformer-based language models in detecting software vulnerabilities and how good are these models for vulnerability detection tasks. In this regard, firstly, a systematic (cohesive) framework that details source code translation, model preparation, and inference is presented. Then, an empirical analysis is performed with software vulnerability datasets with C/C++ source codes having multiple vulnerabilities corresponding to the library function call, pointer usage, array usage, and arithmetic expression. Our empirical results demonstrate the good performance of the language models in vulnerability detection. Moreover, these language models have better performance metrics, such as F1-score, than the contemporary models, namely bidirectional long short-term memory and bidirectional gated recurrent unit. Experimenting with the language models is always challenging due to the requirement of computing resources, platforms, libraries, and dependencies. Thus, this paper also analyses the popular platforms to efficiently fine-tune these models and present recommendations while choosing the platforms. | ['Surya Nepal', 'Josef Pieprzyk', 'Seyit Camtepe', 'Muhammad Ejaz Ahmed', 'Seung Ick Jang', 'Chandra Thapa'] | 2022-04-07 | null | null | null | null | ['code-translation', 'vulnerability-detection'] | ['computer-code', 'miscellaneous'] | [-1.72680646e-01 -4.02522057e-01 -3.95007402e-01 -8.18333775e-02
-8.20362210e-01 -7.27428436e-01 1.86789826e-01 3.30184132e-01
-2.05866415e-02 7.37064630e-02 1.99225917e-01 -1.02328467e+00
8.86436254e-02 -8.62141550e-01 -5.43152332e-01 -1.45389736e-01
-4.98362690e-01 -4.38614368e-01 4.20618057e-01 -4.49373573e-01
8.19045067e-01 1.20929740e-01 -1.48701918e+00 3.25064003e-01
7.03853726e-01 5.09527504e-01 6.47369251e-02 6.04184389e-01
-5.01746476e-01 1.13518214e+00 -5.49535036e-01 -5.95896184e-01
-1.60723165e-01 6.76949322e-02 -7.58858860e-01 -7.60708392e-01
-3.00719365e-02 -8.10421184e-02 -1.53032735e-01 1.25355971e+00
4.06264782e-01 -3.80227327e-01 1.97126925e-01 -1.07556200e+00
-8.83872092e-01 1.28674781e+00 -7.52221942e-01 5.15000045e-01
4.61173296e-01 -4.16708402e-02 1.03427386e+00 -6.98981106e-01
3.49471003e-01 1.21324229e+00 8.21628630e-01 3.34974140e-01
-6.84049368e-01 -4.18936014e-01 3.67571801e-01 1.89680696e-01
-1.47725666e+00 -1.38223901e-01 6.37216985e-01 -7.73439109e-01
1.78122091e+00 1.34831518e-01 -1.31169200e-01 1.21580493e+00
5.33872902e-01 3.23734730e-01 7.20602095e-01 -6.83741093e-01
1.30100191e-01 3.14494222e-01 1.01205456e+00 7.28658915e-01
1.84885174e-01 2.06399977e-01 -3.27871859e-01 -5.95715642e-01
1.55437976e-01 1.39144380e-02 -7.42105991e-02 2.52739131e-01
-8.23364854e-01 7.89789140e-01 2.09394276e-01 7.54258990e-01
-1.93066485e-02 1.05570011e-01 1.06831110e+00 4.73099232e-01
2.21520051e-01 3.90117824e-01 -4.74524617e-01 -5.13558209e-01
-9.63018596e-01 -2.66232610e-01 6.80303514e-01 1.07919502e+00
4.01530087e-01 3.55179548e-01 -8.76174308e-03 4.66739416e-01
3.53245348e-01 4.00089175e-01 6.94786131e-01 -7.29939267e-02
8.30808461e-01 9.05268312e-01 -4.65630949e-01 -1.19319081e+00
-4.60063726e-01 -4.18292582e-01 -6.57315195e-01 1.55291688e-02
-5.17508341e-03 1.25955522e-01 -2.78180748e-01 1.77160120e+00
-2.66644597e-01 -7.69398883e-02 3.75938058e-01 2.25156248e-01
6.36108279e-01 6.29737258e-01 2.21493497e-01 -5.34881949e-02
1.40457821e+00 -5.89541733e-01 -3.56095165e-01 -2.22906157e-01
1.14719033e+00 -7.82543778e-01 1.22411287e+00 1.52495176e-01
-6.63161635e-01 -4.14697677e-01 -9.37579989e-01 7.66946152e-02
-6.67387545e-01 8.28240812e-02 6.16965055e-01 1.18358648e+00
-1.10578179e+00 3.01939517e-01 -9.51367021e-01 -3.32091421e-01
-1.88762650e-01 7.74602126e-03 5.85249513e-02 4.64226335e-01
-1.35457921e+00 7.41688073e-01 2.04658628e-01 6.73775449e-02
-8.97055209e-01 -6.50201142e-01 -8.79189909e-01 2.11856678e-01
3.32660139e-01 -4.10405360e-02 9.14500296e-01 -4.76545930e-01
-1.09054637e+00 5.67381859e-01 -1.21364281e-01 -4.91607040e-01
-2.62927592e-01 -1.99251369e-01 -5.50978720e-01 -4.31145132e-01
-2.08436027e-01 -3.18443269e-01 6.46064520e-01 -5.51093340e-01
-3.32018971e-01 -3.18612099e-01 3.63648802e-01 -6.05774522e-01
-9.75636899e-01 8.77177596e-01 -2.07679659e-01 -6.51684940e-01
-4.22015816e-01 -8.20730686e-01 -1.55452162e-01 -7.39890158e-01
-5.30310154e-01 -2.03526482e-01 5.18793166e-01 -8.93972814e-01
2.12787056e+00 -2.41904449e+00 1.43028826e-01 5.38175285e-01
2.15705454e-01 1.49272576e-01 -1.48967668e-01 5.70864141e-01
-3.92428100e-01 6.36284351e-01 -2.64380991e-01 1.08696073e-01
1.93162516e-01 -3.28670651e-01 -1.02224863e+00 2.51413703e-01
-8.79255533e-02 9.05026495e-01 -4.89754677e-01 -3.04374009e-01
-9.55787748e-02 4.30421561e-01 -4.85037684e-01 1.95722967e-01
-2.25546348e-04 -3.40643287e-01 -6.95932388e-01 9.04118180e-01
2.23391101e-01 -1.25472739e-01 8.37354437e-02 1.69736654e-01
-4.93531078e-01 5.15239239e-01 -7.43551910e-01 1.32193804e+00
-9.32714462e-01 5.29123604e-01 -2.45916918e-01 -7.92097867e-01
1.16077948e+00 4.46097761e-01 -1.74161926e-01 -8.37329626e-01
-3.10830884e-02 2.74500787e-01 -8.96176994e-02 -6.16407812e-01
4.51808572e-01 5.15274107e-01 -5.69809437e-01 7.79581726e-01
-3.15338075e-01 9.58838314e-02 2.07121402e-01 2.16400057e-01
1.24022365e+00 -7.33152479e-02 2.72458017e-01 -3.63351703e-01
8.30459177e-01 -3.23879540e-01 3.56049389e-01 6.69333100e-01
9.41162482e-02 1.60964966e-01 9.47267413e-01 -2.99641997e-01
-8.09491515e-01 -6.26924753e-01 2.17665974e-02 1.25031245e+00
-3.61443222e-01 -9.17322397e-01 -8.86961699e-01 -5.16631961e-01
-4.43435878e-01 7.70036936e-01 -6.85107827e-01 -6.16692781e-01
-8.38713288e-01 -9.87607777e-01 1.20726216e+00 9.22402143e-01
1.25473127e-01 -1.01105487e+00 -8.34562898e-01 1.27382398e-01
-2.89691687e-02 -8.36394906e-01 -5.57055235e-01 2.61868954e-01
-8.27753901e-01 -9.50158775e-01 -1.29927188e-01 -6.12345278e-01
2.40696907e-01 5.90464508e-04 1.32986093e+00 3.76025259e-01
-3.66315633e-01 5.19977137e-02 -6.35680556e-01 1.77475661e-02
-5.26115239e-01 2.49074325e-01 -1.40848473e-01 -4.74182785e-01
4.59816307e-01 -5.63917458e-01 2.08636671e-01 3.68829787e-01
-1.02787828e+00 -5.71657360e-01 4.41998184e-01 5.98770618e-01
1.59530774e-01 3.27803344e-01 1.53068766e-01 -7.04135656e-01
8.33204865e-01 -6.70325518e-01 -9.78360415e-01 6.27525568e-01
-7.08733082e-01 2.91404784e-01 7.83310115e-01 -3.84162515e-01
-9.37530398e-01 -3.89597386e-01 -3.01386893e-01 -1.02125406e-01
2.96068609e-01 1.09848809e+00 -2.12788418e-01 -1.29445061e-01
9.30334568e-01 3.14793020e-01 -4.93883133e-01 -4.13277298e-01
1.25512287e-01 4.62298691e-01 3.37578267e-01 -9.97645199e-01
8.43314946e-01 -1.51540235e-01 -5.61888278e-01 -7.18616366e-01
-2.03537822e-01 -4.94162887e-02 -4.30313766e-01 2.93980181e-01
6.04470372e-01 -5.67999184e-01 -5.15732348e-01 5.09401858e-01
-1.22533834e+00 -1.26469508e-01 2.70599067e-01 9.53743160e-02
-3.70540917e-02 6.71940506e-01 -8.30152750e-01 -9.32866096e-01
-6.39292419e-01 -1.64406121e+00 8.70639324e-01 -1.04652129e-01
-3.61905277e-01 -1.16590059e+00 1.63132876e-01 -1.02101408e-01
8.93612444e-01 -7.70238116e-02 1.67793822e+00 -4.11875725e-01
-4.71135050e-01 -2.47935340e-01 -5.77780828e-02 2.16156662e-01
-2.57380426e-01 5.00698268e-01 -7.54245937e-01 -2.18597740e-01
6.83674142e-02 -1.59456089e-01 6.10201180e-01 5.16334251e-02
1.24012101e+00 -1.74208343e-01 -2.97983140e-01 6.89310610e-01
1.47006655e+00 2.07783699e-01 7.10718572e-01 5.43194294e-01
6.63643658e-01 6.46060944e-01 3.19999307e-01 3.23611259e-01
3.76915276e-01 6.89961672e-01 4.28400099e-01 3.79537404e-01
3.87103200e-01 -1.24663927e-01 1.03832746e+00 1.10242748e+00
1.54973641e-01 -5.87560497e-02 -1.38838255e+00 4.51314300e-01
-1.48257649e+00 -7.62008727e-01 -1.24602132e-01 2.29032779e+00
7.61835217e-01 4.06484544e-01 1.07066274e-01 2.68183798e-01
6.67908788e-01 4.31962050e-02 -2.64385819e-01 -9.36444998e-01
-9.14350301e-02 1.84967309e-01 3.80111635e-01 4.13077891e-01
-8.14254045e-01 9.78829682e-01 6.44956303e+00 8.88956249e-01
-1.44887412e+00 1.27028883e-01 3.63878578e-01 2.55505085e-01
-5.07262230e-01 1.47032261e-01 -1.03276575e+00 4.49247211e-01
1.65923464e+00 -4.85250354e-01 4.46766466e-01 9.64944661e-01
-9.11204051e-03 4.06659722e-01 -1.12961674e+00 6.95995569e-01
1.07502081e-02 -1.04048669e+00 -1.41097484e-02 -1.05717756e-01
5.33923618e-02 4.06195074e-02 3.01514417e-01 6.37902141e-01
2.07980365e-01 -1.09893990e+00 8.58374000e-01 2.39168718e-01
6.32431686e-01 -8.02585602e-01 9.21483397e-01 2.51327455e-01
-1.49268270e+00 -4.50791270e-01 -3.27014506e-01 -6.03361391e-02
-9.91329476e-02 6.16450608e-01 -3.46129715e-01 5.13626277e-01
1.03239882e+00 7.11117983e-01 -1.12990034e+00 5.63525021e-01
-2.40763202e-01 7.52314448e-01 -1.09268218e-01 -2.04661787e-01
2.58129239e-02 2.15836361e-01 2.22647801e-01 1.47340739e+00
3.98900300e-01 -3.44845682e-01 6.89395517e-02 1.12972355e+00
3.16007495e-01 2.34294355e-01 -5.91572285e-01 -4.36995119e-01
6.10112309e-01 1.00196266e+00 -5.40912986e-01 -6.21250831e-02
-7.31242239e-01 2.11884424e-01 3.28782856e-01 3.12327713e-01
-1.05078769e+00 -6.05176508e-01 5.35968125e-01 -6.60169870e-02
-3.49072827e-04 -2.59212643e-01 -3.40167850e-01 -1.30423474e+00
4.59507614e-01 -1.09410357e+00 4.53725874e-01 -4.29478377e-01
-1.15462053e+00 1.23936820e+00 1.83168322e-01 -1.01794648e+00
-4.38770682e-01 -6.28182948e-01 -8.48738253e-01 9.02690411e-01
-1.47186697e+00 -9.24336374e-01 1.52697906e-01 4.88976479e-01
1.69911936e-01 -5.57248294e-01 9.82547760e-01 4.96282876e-01
-1.06062555e+00 1.18185925e+00 -7.44097978e-02 2.33935967e-01
3.69411081e-01 -9.90576506e-01 1.07595468e+00 1.24238575e+00
-1.59164190e-01 1.45378327e+00 4.82022762e-01 -5.69868207e-01
-1.73532832e+00 -1.02283490e+00 8.79606962e-01 -6.00344360e-01
1.13536310e+00 -5.50549746e-01 -1.23784792e+00 6.49463654e-01
-8.33997205e-02 -1.16704658e-01 7.02499926e-01 2.39708185e-01
-1.09390271e+00 1.59298688e-01 -7.26030350e-01 5.41378677e-01
7.57135868e-01 -1.04283285e+00 -5.22038937e-01 5.77005148e-02
8.71878088e-01 -1.19269468e-01 -9.45185244e-01 2.24871263e-01
3.90753210e-01 -9.62583244e-01 1.11671305e+00 -3.80680531e-01
3.91930193e-01 -6.12100586e-02 -5.37477672e-01 -8.02359879e-01
-2.01686397e-01 -6.07609212e-01 -1.51607946e-01 1.42451406e+00
7.75319278e-01 -8.64331365e-01 2.38200188e-01 4.07996893e-01
2.42114533e-03 -5.35155058e-01 -5.78293979e-01 -7.99760044e-01
5.61360121e-01 -7.99251258e-01 8.33479881e-01 8.83318663e-01
4.74254310e-01 1.36827037e-01 -1.93518683e-01 3.69303554e-01
2.70483285e-01 3.20266247e-01 4.18614119e-01 -1.03427804e+00
-5.84826231e-01 -7.34811246e-01 -3.98514479e-01 -4.41842973e-01
6.68927133e-01 -9.49458241e-01 -4.01103348e-01 -6.19800687e-01
3.29524606e-01 -4.23067153e-01 -3.89364719e-01 7.67199576e-01
-1.94376171e-01 -1.37436166e-01 -9.43098664e-02 1.31987676e-01
-3.33735853e-01 1.30285159e-01 1.01218186e-01 -3.04777473e-01
-1.88363940e-01 -8.34209099e-02 -8.65077436e-01 7.30160654e-01
7.03546047e-01 -5.08332431e-01 -3.63997370e-01 -6.59354150e-01
8.31695378e-01 2.94849664e-01 1.52708456e-01 -8.13843489e-01
1.72260299e-01 7.19782943e-03 -4.50737804e-01 -1.74918368e-01
-2.11859852e-01 -4.69655991e-01 -1.32173330e-01 4.78616148e-01
-5.67088366e-01 7.22421169e-01 5.16828775e-01 1.54148713e-01
-3.09661120e-01 -5.55188775e-01 4.70422626e-01 -8.44236612e-02
-9.43131208e-01 9.97935012e-02 -5.36476910e-01 1.48141593e-01
8.55227292e-01 5.01974337e-02 -5.71438968e-01 6.14755191e-02
-2.81010479e-01 -9.91360396e-02 5.46557963e-01 1.03026700e+00
7.29706883e-01 -7.60089159e-01 -5.94560742e-01 3.86063069e-01
4.22486871e-01 -8.51419806e-01 1.65313601e-01 7.35933244e-01
-2.35006422e-01 6.65029585e-01 -9.30815935e-02 -2.10016042e-01
-1.41024184e+00 1.08712828e+00 2.19319120e-01 -3.88520360e-01
-5.86620986e-01 7.60241508e-01 2.22578689e-01 -5.66083491e-01
2.85735548e-01 -5.31921625e-01 -3.73080790e-01 -2.97519296e-01
8.51434171e-01 3.12865525e-01 3.61449838e-01 -5.02440691e-01
-8.07053983e-01 8.15072656e-01 -2.44482428e-01 2.44537786e-01
1.15883029e+00 1.26384556e-01 -8.03790569e-01 4.83040303e-01
1.27021873e+00 1.94552094e-01 -2.64798462e-01 -2.29628131e-01
5.70904136e-01 -1.12694241e-01 6.70088530e-02 -6.61699772e-01
-1.10582519e+00 1.17641330e+00 5.12595117e-01 2.40524948e-01
1.10390103e+00 -1.53437495e-01 6.52357757e-01 4.39822793e-01
7.27303088e-01 -4.72351670e-01 -2.09789410e-01 8.64016771e-01
5.71936607e-01 -6.67995334e-01 -3.38359684e-01 -2.92102396e-01
-2.22942159e-01 1.20553088e+00 5.81326902e-01 1.06783144e-01
6.95535362e-01 9.51753676e-01 -4.52314727e-02 -1.20662436e-01
-8.30079854e-01 3.09901565e-01 6.28737211e-02 2.62536526e-01
9.49014187e-01 -5.01000360e-02 -1.69889167e-01 6.83511138e-01
-3.88116241e-01 -2.71666974e-01 9.01193082e-01 1.14664316e+00
-1.58241063e-01 -1.32223666e+00 -5.94715118e-01 2.15810806e-01
-7.55197406e-01 -6.00185275e-01 -3.29938978e-01 4.32593793e-01
-4.24212605e-01 1.09907544e+00 -5.25178075e-01 -6.06871426e-01
3.03370923e-01 3.16515751e-02 -8.21512379e-03 -7.42613912e-01
-1.05723274e+00 -3.66918504e-01 -1.71529993e-01 -7.50781596e-01
2.60012329e-01 -3.91749531e-01 -1.08427286e+00 -3.65901679e-01
-3.11110049e-01 3.35816443e-01 5.80143273e-01 7.87983596e-01
4.43832278e-01 6.59890354e-01 4.29862827e-01 -4.00627762e-01
-7.89834499e-01 -6.73313200e-01 -3.02220166e-01 -8.33861381e-02
2.66345471e-01 -4.79301900e-01 -4.49045062e-01 -4.76797409e-02] | [7.081683158874512, 7.7787766456604] |
cbdd53fd-e0b2-43db-a2ea-8bd3892d9f54 | controlling-keywords-and-their-positions-in | 2304.09516 | null | https://arxiv.org/abs/2304.09516v1 | https://arxiv.org/pdf/2304.09516v1.pdf | Controlling keywords and their positions in text generation | One of the challenges in text generation is to control generation as intended by a user. Previous studies have proposed to specify the keywords that should be included in the generated text. However, this is insufficient to generate text which reflect the user intent. For example, placing the important keyword beginning of the text would helps attract the reader's attention, but existing methods do not enable such flexible control. In this paper, we tackle a novel task of controlling not only keywords but also the position of each keyword in the text generation. To this end, we show that a method using special tokens can control the relative position of keywords. Experimental results on summarization and story generation tasks show that the proposed method can control keywords and their positions. We also demonstrate that controlling the keyword positions can generate summary texts that are closer to the user's intent than baseline. We release our code. | ['Yasuhiro Sogawa', 'Osamu Imaichi', 'Hiroaki Ozaki', 'Terufumi Morishita', 'Yuichi Sasazawa'] | 2023-04-19 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 4.50289041e-01 2.72308230e-01 -2.97654122e-01 -7.15066940e-02
-6.73564911e-01 -7.47786105e-01 7.80951142e-01 5.16599536e-01
-3.84631127e-01 7.45417833e-01 8.49059999e-01 -2.40423903e-01
2.18892992e-01 -7.01323509e-01 -5.05146742e-01 -3.58187526e-01
4.42385316e-01 3.26294988e-01 3.82218003e-01 -4.39323604e-01
7.43278503e-01 1.61998972e-01 -1.31934357e+00 5.76059282e-01
1.07280684e+00 3.65423650e-01 7.25706875e-01 6.22699380e-01
-2.50501066e-01 4.00941849e-01 -1.21821129e+00 -2.72477381e-02
-5.61329499e-02 -7.73891628e-01 -7.24650145e-01 1.30242944e-01
3.71819526e-01 -4.46015626e-01 -5.79995997e-02 8.02636087e-01
6.00676358e-01 1.44367799e-01 8.57728422e-01 -9.96180356e-01
-5.39867699e-01 1.35158205e+00 -3.89332682e-01 1.26029223e-01
5.79727232e-01 -2.23015875e-01 1.16999936e+00 -5.11213779e-01
7.13187933e-01 1.09724915e+00 2.51072407e-01 6.69083834e-01
-1.10252106e+00 -4.24692601e-01 5.23121595e-01 -2.15015799e-01
-1.11765993e+00 -5.87909937e-01 8.01780164e-01 -4.00842458e-01
6.77887499e-01 5.32018065e-01 4.80430245e-01 1.02530527e+00
3.18616405e-02 8.73813808e-01 5.71821511e-01 -8.43417406e-01
1.50380015e-01 7.83015862e-02 4.65896307e-03 1.84411272e-01
1.75237760e-01 -5.62448680e-01 -5.74003339e-01 -2.04864755e-01
4.38808441e-01 -3.35548609e-01 -6.25992715e-01 1.20487377e-01
-1.53643799e+00 9.66748476e-01 -3.82782519e-02 4.35882002e-01
-3.76660854e-01 1.29855856e-01 3.67964834e-01 -1.70188352e-01
4.59351271e-01 9.64995384e-01 -3.07270527e-01 -2.76219219e-01
-1.10149062e+00 6.89269245e-01 8.73785377e-01 1.20918024e+00
3.14658195e-01 -1.99294552e-01 -1.01403272e+00 8.00353110e-01
3.25725712e-02 2.06933275e-01 5.90306640e-01 -7.79671907e-01
7.05434084e-01 4.70092177e-01 5.98653316e-01 -9.97618258e-01
4.04095910e-02 -4.20969993e-01 -6.37752473e-01 -2.95669168e-01
4.98844832e-01 -5.46260774e-01 -7.79825628e-01 1.72943258e+00
3.16545926e-02 -1.96514085e-01 4.09269296e-02 6.44563854e-01
6.40102923e-01 1.05519664e+00 -2.42164791e-01 -4.55580890e-01
1.46716380e+00 -9.55404639e-01 -1.02997506e+00 -3.82010460e-01
7.00257301e-01 -9.76712644e-01 1.06160998e+00 7.01751634e-02
-1.12737370e+00 -4.34877664e-01 -9.72973585e-01 6.63899109e-02
3.05499323e-02 5.34081280e-01 2.15139002e-01 3.66467386e-01
-7.72662878e-01 4.34197336e-01 -4.63228583e-01 -2.08924189e-01
-4.64626066e-02 6.67890627e-03 2.49047861e-01 3.55010480e-01
-1.33858979e+00 6.06988013e-01 4.95730132e-01 -1.05943233e-01
-1.52924404e-01 -6.57691181e-01 -7.75031865e-01 2.59959728e-01
5.60425997e-01 -7.11019456e-01 1.76277614e+00 -8.41303408e-01
-1.36180830e+00 1.83669224e-01 -4.67656136e-01 -3.57801944e-01
5.35952389e-01 -4.11239624e-01 1.99282899e-01 2.01096952e-01
3.14176828e-01 8.90090048e-01 6.10135972e-01 -1.26339924e+00
-9.28759158e-01 6.16348423e-02 1.70762822e-01 4.40537959e-01
-4.32967633e-01 -3.65495831e-02 -6.20415747e-01 -1.10857046e+00
-6.80714697e-02 -9.62236583e-01 -4.18791950e-01 -4.45916951e-01
-9.19610262e-01 -4.31578249e-01 7.20291376e-01 -5.78692794e-01
1.75749850e+00 -1.89680111e+00 -2.70806849e-02 4.92571108e-02
-1.03330284e-01 2.57753015e-01 -2.97315717e-02 9.50974703e-01
2.53997594e-01 5.62003672e-01 -2.46441606e-02 -2.35040128e-01
1.40349537e-01 -4.83481511e-02 -6.84040487e-01 -1.34449065e-01
-5.21811508e-02 7.89077163e-01 -9.91108298e-01 -4.83479589e-01
-2.58443952e-01 2.85906911e-01 -8.10983062e-01 2.67575294e-01
-6.48052216e-01 1.84932053e-01 -6.95704103e-01 2.48981230e-02
2.90935993e-01 -1.43690363e-01 3.26137058e-02 -2.04566434e-01
-2.81623393e-01 7.86635995e-01 -1.08727920e+00 1.25195515e+00
-3.87680084e-01 5.53147733e-01 -2.72477150e-01 -3.86479944e-01
7.60219038e-01 3.37595254e-01 2.04493150e-01 -4.11211610e-01
-2.95135230e-01 4.77471389e-02 2.11953178e-01 -2.49235451e-01
1.00910258e+00 1.20825402e-01 -2.58471280e-01 6.90604031e-01
-7.65567780e-01 -4.25417632e-01 7.35795975e-01 5.80438912e-01
7.71656632e-01 -3.85293998e-02 4.25821573e-01 -4.38724793e-02
2.98072219e-01 5.14376797e-02 4.48572248e-01 1.14356494e+00
4.34004784e-01 7.71308899e-01 9.18729961e-01 1.51887804e-01
-1.00693905e+00 -4.47071344e-01 2.63884813e-01 1.07625091e+00
-6.30514100e-02 -9.61770475e-01 -9.39188957e-01 -6.97227240e-01
-1.85992628e-01 1.28001559e+00 -4.75430131e-01 1.18293069e-01
-8.27095747e-01 -3.17692220e-01 4.21366602e-01 4.02905017e-01
1.42673582e-01 -1.18729258e+00 -8.56782019e-01 3.37077051e-01
-7.73715436e-01 -1.01230311e+00 -1.38093913e+00 -1.45960987e-01
-4.78257269e-01 -6.69693768e-01 -7.60951936e-01 -7.46245205e-01
9.30374563e-01 3.34386379e-01 7.71287680e-01 2.06558466e-01
3.13668758e-01 2.24574283e-02 -7.46775687e-01 -5.41693866e-01
-7.45052516e-01 5.93192339e-01 -3.58481318e-01 -2.15432733e-01
-2.43445486e-01 -2.51715928e-01 -5.91790915e-01 2.86962241e-01
-1.15619838e+00 6.07247591e-01 3.61159295e-01 6.77567899e-01
2.68759549e-01 2.20858883e-02 7.10257888e-01 -1.03912234e+00
1.28356326e+00 -2.22700343e-01 -2.34921128e-01 2.24253595e-01
-4.73578095e-01 3.47244143e-01 8.83347929e-01 -5.84913969e-01
-9.35041964e-01 -7.10000098e-02 -1.25352472e-01 3.26988995e-01
-7.02013075e-02 6.29018784e-01 -2.37974748e-01 7.20492899e-01
4.45405602e-01 4.21342254e-01 -2.74273336e-01 -2.67589360e-01
2.76879996e-01 7.45268106e-01 1.36971772e-01 -6.03808403e-01
5.90650797e-01 -9.93310958e-02 -3.30745757e-01 -6.78840399e-01
-7.41214395e-01 -3.58550787e-01 -3.60719085e-01 1.65924150e-02
5.17627895e-01 -6.03931248e-01 -2.66900837e-01 2.13529766e-01
-1.62158012e+00 -3.40009749e-01 -4.70960140e-02 1.62680402e-01
-4.00772572e-01 4.72043902e-01 -4.82708424e-01 -7.71565020e-01
-5.41432500e-01 -1.22315097e+00 1.22708380e+00 2.07306087e-01
-8.73347342e-01 -7.22170115e-01 -1.89430431e-01 5.42595163e-02
3.57761741e-01 9.57652852e-02 9.61649656e-01 -8.31261218e-01
-3.43559355e-01 -1.54711857e-01 9.73238498e-02 -9.06649157e-02
4.73685950e-01 2.67729759e-01 -4.28976059e-01 1.29818209e-02
-3.62139761e-01 1.70542181e-01 8.13825250e-01 3.42588395e-01
1.03030622e+00 -9.48179603e-01 -5.67917109e-01 2.18869060e-01
8.78375769e-01 3.75060529e-01 6.40212595e-01 3.11824769e-01
6.22140884e-01 7.61328936e-01 7.50873923e-01 6.45518124e-01
4.08676147e-01 1.03550577e+00 1.43100172e-01 1.30134299e-01
5.95456995e-02 -6.02768719e-01 3.59615147e-01 5.67547202e-01
2.89842844e-01 -9.30773675e-01 -6.93786919e-01 5.80182016e-01
-1.84009862e+00 -1.07485461e+00 -3.71617496e-01 2.05691028e+00
1.17881024e+00 2.94585973e-01 3.57118994e-02 2.19013035e-01
8.32144380e-01 3.31512988e-01 -8.56438205e-02 -4.92708474e-01
2.00982049e-01 -1.03130348e-01 2.91629195e-01 6.13135099e-01
-7.43361890e-01 1.07667851e+00 6.13581419e+00 7.45379090e-01
-1.09157813e+00 -3.22979718e-01 5.23928165e-01 -1.46720037e-01
-7.21636474e-01 1.57592908e-01 -1.21502268e+00 7.75475264e-01
6.49050772e-01 -5.32742321e-01 2.49100979e-02 5.22205353e-01
7.79970109e-01 -2.78661668e-01 -1.14573622e+00 4.88525093e-01
1.40643582e-01 -1.29055190e+00 5.43552577e-01 4.23430987e-02
6.15392506e-01 -9.03209507e-01 -1.60804927e-01 1.50105819e-01
-2.17623860e-02 -7.38459170e-01 1.06485128e+00 3.56922239e-01
4.55395818e-01 -7.54743397e-01 3.81030589e-01 5.81006289e-01
-8.90738428e-01 1.92769989e-01 -1.67550936e-01 -1.16071291e-01
4.22841161e-01 6.91681862e-01 -1.32298660e+00 2.26161182e-01
6.16119504e-02 3.01050007e-01 -4.47453767e-01 8.48290980e-01
-6.31638467e-01 7.80721843e-01 -8.91709477e-02 -4.58038598e-01
2.24428996e-01 -6.11761808e-02 6.67919576e-01 1.43053234e+00
6.81907833e-01 2.11910099e-01 3.64562422e-01 7.51204550e-01
-1.10098071e-01 4.79766935e-01 -4.82917696e-01 -3.78886133e-01
7.23052979e-01 9.91958082e-01 -6.92799568e-01 -4.47401166e-01
6.50033280e-02 1.08039916e+00 1.57286495e-01 3.70701283e-01
-7.96822786e-01 -6.77550137e-01 3.55313897e-01 5.22554219e-01
5.26806772e-01 -1.82547614e-01 -2.41391882e-01 -8.65253091e-01
2.55605131e-01 -1.05788958e+00 7.39518851e-02 -8.76900256e-01
-7.10314155e-01 4.35599297e-01 2.25371107e-01 -9.71349001e-01
-5.93697131e-01 -3.63882743e-02 -8.53895366e-01 9.98494387e-01
-1.13791585e+00 -7.56368637e-01 1.42730828e-02 -1.12944715e-01
9.86616492e-01 2.38186836e-01 6.33913636e-01 -1.24200292e-01
-3.84125054e-01 5.68377495e-01 -3.06820214e-01 1.21719033e-01
8.82531941e-01 -1.29157019e+00 6.38729215e-01 9.67554808e-01
-7.34459534e-02 9.82792914e-01 1.21406412e+00 -1.07056117e+00
-1.13944542e+00 -9.77831841e-01 1.45931590e+00 -1.38129517e-01
4.10775781e-01 -3.19608241e-01 -8.45350981e-01 4.60614502e-01
5.32095015e-01 -7.21561015e-01 5.24642944e-01 -1.09467857e-01
1.12773426e-01 1.82650805e-01 -7.19008684e-01 1.15456271e+00
8.46003473e-01 8.37861001e-03 -6.34460032e-01 4.28007394e-01
9.53017652e-01 -6.79963768e-01 -2.70373762e-01 1.58625320e-01
3.53616118e-01 -7.18504906e-01 5.90815485e-01 -4.00480270e-01
7.36807227e-01 -3.71039748e-01 2.39571184e-01 -1.74253535e+00
-2.79788673e-01 -9.03574049e-01 -5.51157333e-02 1.50210524e+00
6.84140623e-01 -2.98713624e-01 5.55662155e-01 6.43974125e-01
-3.30857188e-01 -6.33524299e-01 -3.62201124e-01 -3.15700352e-01
-1.52007341e-01 -1.29450381e-01 5.34378052e-01 5.35931408e-01
3.60205829e-01 5.87765455e-01 -4.47519660e-01 -1.27952278e-03
7.00909793e-02 8.05162415e-02 7.65428364e-01 -8.61090422e-01
-2.18656600e-01 -6.31514907e-01 2.56441414e-01 -1.54729998e+00
2.44732257e-02 -6.15591586e-01 4.42121148e-01 -2.08120394e+00
1.96802363e-01 -3.43114048e-01 3.52100164e-01 4.54412431e-01
-7.66871095e-01 -2.52113104e-01 3.78772736e-01 1.05509520e-01
-3.00384998e-01 5.32925248e-01 1.38598669e+00 -1.15616560e-01
-5.29230773e-01 3.14061880e-01 -1.20607996e+00 3.46527398e-01
1.05561042e+00 -4.51766044e-01 -6.08306170e-01 -2.67473668e-01
2.80336708e-01 1.14759207e-01 1.11396397e-02 -5.94766974e-01
3.67700130e-01 -4.70723122e-01 2.28512466e-01 -8.56490731e-01
3.46788839e-02 -3.50774437e-01 1.24907074e-02 3.57914209e-01
-8.53124559e-01 3.88118654e-01 1.37689173e-01 2.64902413e-01
-3.20648625e-02 -6.30658686e-01 3.77894491e-01 -1.92003682e-01
-1.15389794e-01 -4.13062833e-02 -7.77663648e-01 1.68585464e-01
8.51942599e-01 -1.27840057e-01 -2.47669846e-01 -6.96710587e-01
-1.22106247e-01 3.06949824e-01 4.83979315e-01 4.36167121e-01
5.92374504e-01 -1.19610059e+00 -8.93499196e-01 -2.09709898e-01
2.47468997e-04 5.30221127e-02 -2.67273158e-01 4.98276621e-01
-2.95456856e-01 5.90543926e-01 3.27981919e-01 -1.50714055e-01
-1.33513987e+00 3.25561255e-01 -4.71507944e-03 -3.75090122e-01
-5.12560070e-01 4.81208026e-01 1.87310398e-01 -1.20597361e-02
2.91390061e-01 -4.96900529e-01 -4.64852780e-01 4.08289641e-01
8.42738926e-01 1.72648299e-02 -1.47231579e-01 -4.02457327e-01
-2.64754537e-02 1.26332149e-01 -4.50732589e-01 -4.86076862e-01
1.12394035e+00 -1.81631938e-01 -3.13339271e-02 4.13420707e-01
7.05742836e-01 7.41819561e-01 -9.88982260e-01 -5.07716509e-03
6.82836995e-02 -5.14135122e-01 -1.13115832e-01 -7.51498044e-01
-6.51627600e-01 4.91303623e-01 -2.77291745e-01 4.75472361e-01
9.55480516e-01 -5.01174890e-02 9.58196044e-01 3.00266445e-01
5.50503023e-02 -1.10093844e+00 1.25510782e-01 6.81156278e-01
1.05815208e+00 -7.34621227e-01 -1.50858071e-02 -6.52970910e-01
-8.92852962e-01 1.20939040e+00 8.36572886e-01 3.34946275e-01
2.05775071e-02 1.57577276e-01 5.68267591e-02 1.63419411e-01
-8.93418908e-01 -7.97253102e-02 3.83515328e-01 4.07246381e-01
9.40781534e-01 -4.48345691e-02 -8.84902954e-01 5.38133204e-01
-5.02108097e-01 -3.62316132e-01 9.83805120e-01 9.17414427e-01
-6.54519439e-01 -1.38663173e+00 -4.96147633e-01 5.44393122e-01
-7.31168926e-01 -4.15178865e-01 -6.02718592e-01 4.85040218e-01
-9.87820178e-02 1.26430094e+00 -1.24835260e-02 -5.36723249e-02
4.40818608e-01 7.99125656e-02 2.14648008e-01 -1.07213449e+00
-6.81957960e-01 3.30788046e-01 3.83907795e-01 7.02748969e-02
-2.13120170e-02 -7.40984321e-01 -1.39722443e+00 4.19729874e-02
-3.98534954e-01 5.45498312e-01 4.39232618e-01 8.24260771e-01
3.17875832e-01 6.11357212e-01 5.82945287e-01 -6.52354956e-01
-6.34826124e-01 -1.22639358e+00 -6.58286735e-02 3.23014736e-01
1.91397041e-01 -1.97827116e-01 -2.90420264e-01 2.19324887e-01] | [11.926097869873047, 8.985147476196289] |
52f4a5d4-13da-48b7-a9a7-e9fcfae4ed66 | annotating-targets-of-opinions-in-arabic | null | null | https://aclanthology.org/W15-3210 | https://aclanthology.org/W15-3210.pdf | Annotating Targets of Opinions in Arabic using Crowdsourcing | null | ['Kathy Mckeown', 'Noura Farra', 'Nizar Habash'] | 2015-07-01 | null | null | null | ws-2015-7 | ['fine-grained-opinion-analysis', 'subjectivity-analysis'] | ['natural-language-processing', '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] | [-7.411048889160156, 3.7760651111602783] |
9edea27b-8487-4b2a-a2a7-61f2ba027d3f | machine-learning-in-sports-a-case-study-on | 2206.09258 | null | https://arxiv.org/abs/2206.09258v1 | https://arxiv.org/pdf/2206.09258v1.pdf | Machine Learning in Sports: A Case Study on Using Explainable Models for Predicting Outcomes of Volleyball Matches | Machine Learning has become an integral part of engineering design and decision making in several domains, including sports. Deep Neural Networks (DNNs) have been the state-of-the-art methods for predicting outcomes of professional sports events. However, apart from getting highly accurate predictions on these sports events outcomes, it is necessary to answer questions such as "Why did the model predict that Team A would win Match X against Team B?" DNNs are inherently black-box in nature. Therefore, it is required to provide high-quality interpretable, and understandable explanations for a model's prediction in sports. This paper explores a two-phased Explainable Artificial Intelligence(XAI) approach to predict outcomes of matches in the Brazilian volleyball League (SuperLiga). In the first phase, we directly use the interpretable rule-based ML models that provide a global understanding of the model's behaviors based on Boolean Rule Column Generation (BRCG; extracts simple AND-OR classification rules) and Logistic Regression (LogReg; allows to estimate the feature importance scores). In the second phase, we construct non-linear models such as Support Vector Machine (SVM) and Deep Neural Network (DNN) to obtain predictive performance on the volleyball matches' outcomes. We construct the "post-hoc" explanations for each data instance using ProtoDash, a method that finds prototypes in the training dataset that are most similar to the test instance, and SHAP, a method that estimates the contribution of each feature on the model's prediction. We evaluate the SHAP explanations using the faithfulness metric. Our results demonstrate the effectiveness of the explanations for the model's predictions. | ['Tirtharaj Dash', 'Aditya Jain', 'Apoorv Singh', 'Aman Saraiya', 'Abhinav Lalwani'] | 2022-06-18 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 7.04236748e-03 4.97107804e-01 -5.78913033e-01 -4.91182208e-01
-2.50974298e-01 -2.69343853e-01 1.32478386e-01 2.08716124e-01
2.15225205e-01 8.40262949e-01 2.48571917e-01 -5.86721599e-01
-7.64064312e-01 -1.11482048e+00 -9.85099375e-01 -2.99584270e-01
-2.13752031e-01 8.21998298e-01 -7.93434978e-02 -4.74860936e-01
4.14739519e-01 6.55631602e-01 -2.08232141e+00 8.43922317e-01
7.93718457e-01 1.09755325e+00 -4.12562132e-01 5.26240408e-01
2.01015789e-02 1.13578999e+00 -5.37399769e-01 -5.29886901e-01
2.09944323e-01 -4.01595533e-01 -7.31355369e-01 -2.69522280e-01
1.34382978e-01 -8.60943794e-02 -2.22933143e-01 5.77788889e-01
-1.03682458e-01 2.17962754e-03 6.82518244e-01 -1.57436943e+00
-6.28279150e-01 9.01104331e-01 -2.22879704e-02 -6.02764897e-02
3.85739297e-01 4.13938999e-01 1.12508297e+00 -8.76644194e-01
4.35223222e-01 1.13347411e+00 6.38648689e-01 4.81879801e-01
-1.15556443e+00 -6.94693089e-01 4.83858660e-02 7.14976728e-01
-1.02963877e+00 -2.57690698e-02 5.45135379e-01 -5.51319599e-01
9.51517284e-01 7.33010471e-01 1.00200582e+00 7.11992145e-01
5.62559366e-01 6.59234166e-01 8.99664283e-01 -4.60390598e-01
3.72779548e-01 9.52729359e-02 4.37957764e-01 5.18762350e-01
5.25503933e-01 6.08984590e-01 -1.09206855e+00 7.97820985e-02
5.77893317e-01 -4.68131453e-02 1.23081930e-01 2.31819469e-02
-7.50310361e-01 1.02431524e+00 4.34729069e-01 -5.39108180e-02
-5.87723136e-01 2.45279878e-01 1.60691887e-01 2.86973715e-01
2.04036698e-01 1.02053821e+00 -6.49551213e-01 -2.00282723e-01
-6.13800406e-01 7.87955761e-01 9.21228707e-01 4.96500909e-01
7.42338777e-01 1.76251978e-01 -2.16727182e-01 3.87953430e-01
1.95135415e-01 -4.86774184e-02 3.20380598e-01 -8.63702238e-01
1.48732722e-01 1.19810009e+00 -1.62268549e-01 -1.11648190e+00
-4.11473066e-01 -7.68078923e-01 -6.39725566e-01 6.29357576e-01
4.33307499e-01 7.34994859e-02 -9.43713069e-01 1.48100185e+00
1.25944078e-01 -1.59516379e-01 -2.97127035e-03 1.08697283e+00
9.54389274e-01 4.45294350e-01 -3.26776616e-02 6.43071383e-02
9.90500271e-01 -7.27501929e-01 -4.35792625e-01 -3.29043269e-01
5.87478995e-01 -3.38829279e-01 1.05256021e+00 9.28349555e-01
-9.92505968e-01 -8.83472264e-01 -1.21262777e+00 4.59904782e-02
-3.00534159e-01 2.21529584e-02 1.04952073e+00 4.02473718e-01
-5.15456915e-01 9.29127991e-01 -5.81666708e-01 1.33105084e-01
1.77445486e-01 7.87958682e-01 -3.13797086e-01 7.97520876e-02
-1.33267653e+00 1.11537468e+00 5.83615839e-01 2.44403556e-01
-8.74921024e-01 -7.33551979e-01 -7.73822963e-01 4.26754206e-01
5.86123705e-01 -5.56308925e-01 7.58216560e-01 -1.15573764e+00
-1.08219767e+00 4.96332079e-01 4.94910986e-04 -6.52145386e-01
3.63663256e-01 -2.65524209e-01 -3.80577654e-01 -4.03471559e-01
5.67255914e-02 5.08018851e-01 6.19371980e-02 -1.27751994e+00
-6.35802448e-01 -3.56416166e-01 2.35567078e-01 -1.60291821e-01
-1.17720123e-02 -2.45415583e-01 -1.68732420e-01 -4.47596550e-01
4.10505950e-01 -8.85961533e-01 -3.15218657e-01 -1.84112534e-01
-8.44743848e-01 -3.70609134e-01 2.48046294e-01 -7.48391092e-01
1.53855979e+00 -1.89459264e+00 9.98521969e-02 7.25721180e-01
2.74970770e-01 -2.46817945e-03 1.32911175e-01 3.82721394e-01
-5.01856506e-01 2.57651806e-01 1.93843558e-01 2.55891711e-01
1.20715335e-01 4.95028436e-01 -2.70551205e-01 -2.19636813e-01
3.56958628e-01 7.10743129e-01 -3.59768569e-01 -3.34468275e-01
1.61255732e-01 -1.04133576e-01 -8.61451507e-01 1.29976302e-01
-3.85363877e-01 2.41682157e-01 -4.17002082e-01 5.72275877e-01
1.41415417e-01 1.38975149e-02 4.20637310e-01 -2.41490364e-01
-9.36414823e-02 2.36323535e-01 -1.32051790e+00 1.05966949e+00
-1.09359510e-01 6.33912981e-01 -7.92927623e-01 -1.17076802e+00
1.28550959e+00 7.39652887e-02 2.75616527e-01 -6.52714849e-01
8.52527395e-02 2.77101159e-01 4.42983091e-01 -8.05150092e-01
4.66335773e-01 -4.03311610e-01 -3.10372084e-01 3.07600915e-01
-1.75818786e-01 2.98416525e-01 3.29867214e-01 -1.43037647e-01
1.05036569e+00 2.63632715e-01 4.46764231e-01 -8.55713859e-02
2.89503872e-01 5.95511794e-01 9.56783533e-01 9.64251757e-01
4.26266849e-01 6.10309422e-01 8.11569571e-01 -1.32523525e+00
-9.76955235e-01 -8.22382689e-01 1.15521193e-01 7.07820058e-01
5.82068190e-02 -2.27043241e-01 -5.40932834e-01 -4.61096168e-01
3.10440451e-01 1.23854876e+00 -9.21590507e-01 -5.70260763e-01
-5.59386134e-01 -4.14292723e-01 2.68444747e-01 8.20591271e-01
1.59347847e-01 -1.14894867e+00 -4.60473478e-01 2.38126695e-01
-2.59348363e-01 -5.01067221e-01 2.76374280e-01 4.55123812e-01
-8.19125354e-01 -1.34272718e+00 3.40987414e-01 -2.54593164e-01
4.38343346e-01 -5.36854804e-01 1.23609662e+00 6.16030753e-01
-3.38543594e-01 -3.46963555e-01 -3.63298744e-01 -6.80517435e-01
-5.81661582e-01 -5.18868305e-02 2.99644291e-01 -1.35660470e-01
7.35010743e-01 -4.08650815e-01 -4.88709837e-01 4.84831423e-01
-7.62469113e-01 3.47318143e-01 7.50423253e-01 7.65504599e-01
6.95153356e-01 1.18057273e-01 5.98446846e-01 -7.17118919e-01
6.74649358e-01 -5.40322244e-01 -1.97706074e-01 5.23752868e-01
-9.24029231e-01 3.98330331e-01 6.95881605e-01 -4.86973196e-01
-6.23281300e-01 8.17669705e-02 -1.83311701e-01 -2.93648869e-01
-2.49600321e-01 9.38741386e-01 -1.52070850e-01 2.73535907e-01
1.12640238e+00 5.39663173e-02 -2.67768241e-02 -4.74093825e-01
4.67191674e-02 5.12061834e-01 5.45573652e-01 -7.51476467e-01
7.75150955e-01 -2.00394411e-02 1.32488191e-01 -1.15136847e-01
-9.18963313e-01 1.91284642e-01 -3.91158044e-01 -7.45974004e-01
8.52287412e-01 -4.10902292e-01 -1.31621945e+00 -2.07955614e-01
-9.39691901e-01 8.95164441e-03 -5.90457082e-01 7.61869192e-01
-5.77619255e-01 -2.03530699e-01 -3.45861346e-01 -8.28598201e-01
-5.51836044e-02 -1.09728646e+00 5.21004021e-01 3.53792906e-01
-9.54782426e-01 -4.80290681e-01 -1.74676999e-01 7.28219807e-01
-1.51010618e-01 5.83426654e-01 1.49033749e+00 -1.16908562e+00
-6.56152010e-01 -4.05548006e-01 1.51014119e-01 3.96812826e-01
-3.48257661e-01 1.54574633e-01 -7.76029706e-01 3.57666492e-01
-3.33190173e-01 -2.00622961e-01 4.20028329e-01 6.18829548e-01
1.48531091e+00 -4.02279198e-01 -1.01517685e-01 3.97051573e-01
1.19158781e+00 4.29633439e-01 8.15294385e-01 6.58609688e-01
3.10965776e-01 8.58419240e-01 9.07095671e-01 3.14803511e-01
2.87088335e-01 6.77853703e-01 6.67610765e-01 -7.71133304e-02
9.92265530e-03 -5.88767588e-01 1.86209708e-01 3.43461633e-01
-6.19524956e-01 -5.59442639e-02 -9.96182382e-01 3.18839997e-01
-2.09778333e+00 -9.55179036e-01 -5.53799570e-01 2.08942699e+00
4.66795444e-01 5.81686676e-01 1.60728190e-02 7.68902600e-01
4.05828506e-01 -6.37704670e-01 -8.07304680e-01 -9.14441407e-01
-9.23081934e-02 3.44573498e-01 2.18522057e-01 2.75923043e-01
-5.25042117e-01 6.78366125e-01 6.24186468e+00 7.73575068e-01
-8.60448182e-01 -3.73462260e-01 8.93842101e-01 -2.55888343e-01
-5.37146568e-01 2.34149352e-01 -6.24664366e-01 1.73047200e-01
6.96069837e-01 -6.41729534e-02 3.17488104e-01 1.17390275e+00
4.35024470e-01 -1.34633407e-01 -1.57708490e+00 4.49380875e-01
-2.26955920e-01 -1.81753457e+00 4.05955583e-01 2.31270596e-01
4.63004500e-01 -7.56210744e-01 -6.55363351e-02 5.56202412e-01
3.56253594e-01 -1.46145022e+00 1.15884805e+00 8.52075219e-01
4.53755885e-01 -8.55377376e-01 8.45335960e-01 5.29509544e-01
-6.89095318e-01 -5.65666854e-01 -2.11904272e-01 -7.99656034e-01
-1.19554155e-01 6.05055332e-01 -9.52531934e-01 5.01062810e-01
8.88854980e-01 3.26392174e-01 -3.39836717e-01 7.80782759e-01
-4.20025557e-01 8.52910161e-01 -7.72838071e-02 -2.98969060e-01
-8.23635012e-02 8.97214711e-02 4.28728789e-01 4.91356283e-01
3.66087407e-01 2.59163529e-01 -4.98022772e-02 1.22819197e+00
3.70297670e-01 -8.94198269e-02 -3.20478290e-01 1.42556757e-01
3.12146574e-01 7.44168818e-01 -4.86683279e-01 -3.07343274e-01
3.25083211e-02 1.39944643e-01 1.53816909e-01 1.19106717e-01
-8.22601140e-01 -4.79254238e-02 7.49538481e-01 5.60820043e-01
-2.41572574e-01 2.29344815e-01 -1.06124854e+00 -8.23047280e-01
4.24034260e-02 -1.32716918e+00 4.84988660e-01 -1.06285357e+00
-1.11521637e+00 5.26421726e-01 4.75968495e-02 -1.36334312e+00
-5.54584503e-01 -6.57383859e-01 -6.67447746e-01 8.26295018e-01
-8.34230304e-01 -1.03753614e+00 -1.38541520e-01 2.74797887e-01
5.10822654e-01 -3.04813296e-01 6.42811716e-01 3.38840745e-02
-4.83264953e-01 4.75433707e-01 -4.08428431e-01 8.79850909e-02
-7.43762765e-04 -1.03764534e+00 1.02933809e-01 5.89436114e-01
8.32847878e-02 6.59294128e-01 1.24250281e+00 -7.50126362e-01
-1.25312519e+00 -5.90532541e-01 1.07822812e+00 -2.60330498e-01
3.63347083e-01 1.10786885e-01 -7.66853750e-01 7.07337379e-01
-4.19922858e-01 -6.06368959e-01 1.04292309e+00 6.58367157e-01
9.62847024e-02 -2.81469256e-01 -8.84215236e-01 6.52266562e-01
9.83335257e-01 1.01497211e-02 -7.97147453e-01 2.24514097e-01
3.99752885e-01 -3.14597011e-01 -9.16390955e-01 5.99507093e-01
8.61382902e-01 -1.24967480e+00 8.57050657e-01 -1.35273540e+00
1.19322312e+00 -3.84754300e-01 1.34919537e-03 -1.11968112e+00
-3.83398324e-01 -5.75581491e-02 8.11744779e-02 7.57437289e-01
8.57168078e-01 -3.31039220e-01 1.14561009e+00 1.23235643e+00
-4.10249442e-01 -1.19512892e+00 -8.53554189e-01 -5.56360543e-01
-1.55090332e-01 -8.13586116e-01 9.80098426e-01 7.98692286e-01
2.49629021e-01 2.24342942e-02 -4.79190230e-01 2.57173687e-01
4.82434094e-01 3.17207038e-01 1.00180590e+00 -1.48580933e+00
-6.25136733e-01 -2.95975327e-01 -8.21211278e-01 -5.44026732e-01
-4.57920656e-02 -8.24137986e-01 -6.76047727e-02 -1.51354551e+00
2.18600988e-01 -3.87747735e-01 -1.43286437e-01 8.10437024e-01
5.10698892e-02 -3.45943421e-02 1.77770928e-01 9.17864665e-02
-4.27969359e-02 1.84160814e-01 1.10190284e+00 -3.28405127e-02
-9.00298730e-02 1.59332171e-01 -9.16789532e-01 8.62437546e-01
7.14436114e-01 -7.30288148e-01 -2.07895607e-01 -2.50604898e-01
8.11098039e-01 4.73714411e-01 6.03779316e-01 -1.05219090e+00
-1.98413786e-02 -5.58125734e-01 5.94264150e-01 -3.91311377e-01
2.15347156e-01 -5.62878013e-01 6.93527281e-01 8.01517427e-01
-7.62679279e-01 -1.37066978e-04 1.16449542e-01 2.41430759e-01
-4.61811453e-01 -1.89858690e-01 2.30887696e-01 4.95490097e-02
-6.29903138e-01 -7.29555823e-03 -4.34182137e-01 -4.25538450e-01
9.92445588e-01 -6.89080119e-01 -1.39316797e-01 -4.62655038e-01
-1.29335761e+00 1.49553105e-01 -2.04549450e-02 5.26165366e-01
7.97559440e-01 -1.26819217e+00 -7.57437170e-01 2.20272645e-01
6.69877529e-02 -1.40029639e-01 2.91972280e-01 4.52902675e-01
-7.09417641e-01 2.93150008e-01 -4.72629607e-01 -3.83309066e-01
-1.15341878e+00 3.30303580e-01 4.70784009e-01 -4.41395998e-01
-2.75827318e-01 6.98576450e-01 -3.94777767e-02 -5.44280350e-01
-6.18488789e-02 -4.58031535e-01 -4.08719778e-01 -4.92238998e-01
2.28287861e-01 4.38156158e-01 -5.63514791e-02 -1.72702610e-01
-2.20102906e-01 2.11392626e-01 2.55901784e-01 1.58635765e-01
1.55697119e+00 6.27505839e-01 7.54082873e-02 3.97057295e-01
4.82884377e-01 -4.39509571e-01 -9.01411831e-01 2.40709096e-01
6.20446801e-02 -3.74294788e-01 -2.27501199e-01 -1.09124577e+00
-1.02274811e+00 8.51048589e-01 3.85946900e-01 3.18364739e-01
9.83585477e-01 -3.55593190e-02 3.63131821e-01 3.74546111e-01
2.43963450e-01 -1.17058384e+00 -1.27506971e-01 6.71765804e-02
1.15475571e+00 -1.00141919e+00 1.35799766e-01 -3.96914035e-01
-8.12559187e-01 1.42509377e+00 8.99738848e-01 -1.09524466e-01
4.86562371e-01 2.15843737e-01 9.35064442e-03 -4.72768068e-01
-9.61916029e-01 1.81437999e-01 6.73152924e-01 1.90189168e-01
2.92271703e-01 3.69485795e-01 -5.38268268e-01 1.31639755e+00
-7.03179598e-01 3.87849063e-01 3.62931907e-01 6.19892240e-01
-5.32868862e-01 -1.02929544e+00 -5.56096017e-01 8.50297868e-01
-1.18958548e-01 7.08257258e-02 -6.54980481e-01 1.05114698e+00
5.50051272e-01 1.05975664e+00 -1.11702934e-01 -1.16034830e+00
6.21331632e-01 6.44164309e-02 5.43490052e-02 -4.21756417e-01
-9.62446988e-01 -5.20957053e-01 4.85731930e-01 -6.31005287e-01
-1.13513637e-02 -5.43049335e-01 -1.48250949e+00 -7.02842593e-01
-2.69690514e-01 2.91728914e-01 6.83627665e-01 1.30946469e+00
2.17636839e-01 6.01905406e-01 3.52553517e-01 -4.87366259e-01
-4.23666835e-01 -7.06463933e-01 -4.33712751e-01 6.42534077e-01
-3.22370380e-01 -8.47389638e-01 -2.51927435e-01 -5.24731241e-02] | [8.852115631103516, 5.9402384757995605] |
7ef910d9-4773-4e6d-96fc-d1a401c40521 | knowledgebra-an-algebraic-learning-framework | 2204.07328 | null | https://arxiv.org/abs/2204.07328v1 | https://arxiv.org/pdf/2204.07328v1.pdf | Knowledgebra: An Algebraic Learning Framework for Knowledge Graph | Knowledge graph (KG) representation learning aims to encode entities and relations into dense continuous vector spaces such that knowledge contained in a dataset could be consistently represented. Dense embeddings trained from KG datasets benefit a variety of downstream tasks such as KG completion and link prediction. However, existing KG embedding methods fell short to provide a systematic solution for the global consistency of knowledge representation. We developed a mathematical language for KG based on an observation of their inherent algebraic structure, which we termed as Knowledgebra. By analyzing five distinct algebraic properties, we proved that the semigroup is the most reasonable algebraic structure for the relation embedding of a general knowledge graph. We implemented an instantiation model, SemE, using simple matrix semigroups, which exhibits state-of-the-art performance on standard datasets. Moreover, we proposed a regularization-based method to integrate chain-like logic rules derived from human knowledge into embedding training, which further demonstrates the power of the developed language. As far as we know, by applying abstract algebra in statistical learning, this work develops the first formal language for general knowledge graphs, and also sheds light on the problem of neural-symbolic integration from an algebraic perspective. | ['Pengyu Hong', 'Jan Engelbrecht', 'Long Sha', 'Yifei Wang', 'Tong Yang'] | 2022-04-15 | null | null | null | null | ['general-knowledge', 'abstract-algebra'] | ['miscellaneous', 'reasoning'] | [-3.51254284e-01 6.17585957e-01 -4.00803477e-01 -3.20098668e-01
7.01033371e-03 -4.54844415e-01 5.65876663e-01 4.17998701e-01
-3.79737318e-02 5.05376935e-01 1.77768350e-01 -4.76040423e-01
-6.57619298e-01 -1.23604941e+00 -9.61840749e-01 -5.15267372e-01
-4.42347765e-01 5.27706206e-01 -9.88717079e-02 -3.82836133e-01
-2.23134071e-01 4.72449869e-01 -1.40569043e+00 7.28286663e-03
8.92112315e-01 9.56209898e-01 -2.30940923e-01 1.16040565e-01
-3.42952371e-01 1.08599067e+00 -1.54825345e-01 -1.06869090e+00
3.21384706e-02 -1.58136070e-01 -9.96643186e-01 -4.16504502e-01
3.86376709e-01 9.28200185e-02 -6.74227893e-01 1.01826072e+00
3.73337604e-03 1.32789880e-01 8.12412679e-01 -1.48497593e+00
-1.33921051e+00 1.17901850e+00 1.91238180e-01 -2.14923814e-01
2.76874006e-01 -3.59664887e-01 1.66935122e+00 -9.82598364e-01
9.58823681e-01 1.12937999e+00 8.09355140e-01 2.94296205e-01
-1.43895197e+00 -3.66214991e-01 -5.81657477e-02 6.02279007e-01
-1.73630154e+00 -1.57557756e-01 8.10607135e-01 -5.78090549e-01
1.29112637e+00 7.27638528e-02 8.98146033e-01 8.47064078e-01
1.77130941e-02 7.44805217e-01 7.18620718e-01 -7.39199996e-01
1.00278117e-01 3.57993156e-01 5.47727525e-01 1.32423830e+00
5.73973894e-01 -7.00290278e-02 -6.30117059e-01 -1.48062930e-01
6.64293766e-01 -3.54056656e-01 -2.88100004e-01 -9.40539122e-01
-1.23115695e+00 1.13873315e+00 6.75184071e-01 4.96565253e-01
-1.73773929e-01 3.12254697e-01 4.58702177e-01 4.33753252e-01
2.10309833e-01 4.92812425e-01 -4.57646072e-01 2.54151076e-01
-3.53727072e-01 2.02374369e-01 1.28854656e+00 1.02679563e+00
8.73658299e-01 7.44739100e-02 5.99767342e-02 5.59016943e-01
3.47471148e-01 3.68814260e-01 2.40702763e-01 -7.06743717e-01
2.55067259e-01 1.03874004e+00 -3.88894081e-01 -1.41221905e+00
-3.09576780e-01 -5.13460517e-01 -8.02802145e-01 -1.41625494e-01
3.36957663e-01 3.50338429e-01 -3.36883783e-01 1.94888067e+00
1.41615599e-01 3.57020468e-01 3.14290941e-01 5.49535930e-01
8.02783549e-01 3.37770671e-01 -3.72388363e-02 -4.32680547e-02
1.22011256e+00 -5.35564303e-01 -8.57465684e-01 3.70571733e-01
1.07285726e+00 6.40938282e-02 8.58488679e-01 2.46170253e-01
-7.67330885e-01 -2.81915277e-01 -1.19976735e+00 -1.40451998e-01
-9.51449275e-01 1.20117757e-02 1.40662837e+00 5.95118940e-01
-9.45929706e-01 7.48706400e-01 -7.30634212e-01 -4.29415405e-01
4.34347570e-01 3.54520470e-01 -8.03194821e-01 -1.38501031e-02
-1.66910601e+00 1.18746662e+00 8.55232954e-01 3.41660380e-01
-4.03893113e-01 -7.02867091e-01 -1.37356436e+00 9.42843035e-02
5.46564102e-01 -1.04485881e+00 6.31620884e-01 -3.34046394e-01
-1.30931735e+00 8.49563181e-01 1.91962972e-01 -6.24912262e-01
-7.72811398e-02 -8.26112330e-02 -6.49114549e-01 -1.79520417e-02
-2.07262367e-01 1.24566950e-01 6.81107223e-01 -9.93652761e-01
3.40754129e-02 -4.20391351e-01 3.81073654e-01 -1.41526535e-01
-6.06476963e-01 -3.02999049e-01 -1.94872320e-01 -4.63542759e-01
9.35017467e-02 -6.40038669e-01 -5.33863604e-02 -2.35835359e-01
-5.01163304e-01 -5.26148260e-01 4.33116972e-01 -6.12034619e-01
1.53094792e+00 -1.93436885e+00 7.17879057e-01 7.70911813e-01
5.59846044e-01 3.12610298e-01 1.26762956e-01 7.31100678e-01
-2.24708810e-01 2.50401378e-01 -6.69121444e-02 -2.71818358e-02
5.30109525e-01 5.77058315e-01 -6.95932567e-01 3.68510365e-01
4.01749432e-01 1.37987173e+00 -1.12098897e+00 -5.53509891e-01
1.56608447e-01 4.07974511e-01 -5.94647884e-01 1.06804781e-01
-4.05577242e-01 -2.37892583e-01 -3.48355860e-01 6.00181818e-01
3.33092272e-01 -3.75930518e-01 5.52452981e-01 -5.90416372e-01
3.55905592e-01 1.07863605e-01 -1.32618439e+00 1.87035871e+00
-3.55179191e-01 3.36643815e-01 -3.98462176e-01 -1.46450794e+00
8.99976969e-01 1.36660859e-01 2.79805362e-01 -1.64093778e-01
1.19168967e-01 1.65746823e-01 -4.44998667e-02 -4.71132725e-01
5.30209601e-01 -1.50173798e-01 -1.01382963e-01 1.60233781e-01
5.80874264e-01 -1.24159351e-01 2.34108493e-01 5.29789031e-01
1.12906671e+00 2.22265929e-01 4.93279785e-01 -1.40981853e-01
4.44369286e-01 -7.04209059e-02 3.03014874e-01 5.73061466e-01
2.29305387e-01 -2.18646511e-01 8.01224589e-01 -4.04281825e-01
-5.70524395e-01 -1.42151332e+00 -2.57833749e-01 7.33924448e-01
-6.96628839e-02 -1.03360569e+00 -4.07656968e-01 -6.87547743e-01
4.63714957e-01 7.11650312e-01 -7.66764224e-01 -3.94712389e-01
-1.48088813e-01 -3.83576095e-01 9.27046001e-01 5.70559680e-01
2.49482617e-01 -8.07357132e-01 1.51850432e-01 1.26312748e-01
2.48531148e-01 -1.16029894e+00 5.49387634e-02 4.40147847e-01
-7.62772202e-01 -1.42216277e+00 -5.62324375e-02 -8.28539491e-01
6.56747580e-01 -3.29183310e-01 1.11482692e+00 1.43522294e-02
-2.85694808e-01 6.09882116e-01 -2.93961823e-01 -1.01167068e-01
-4.24648285e-01 -7.76852742e-02 4.37558264e-01 -9.15397424e-03
3.55470866e-01 -8.03344309e-01 2.13328358e-02 -9.66569781e-02
-9.38898504e-01 -1.16631649e-01 4.43620384e-01 1.12997413e+00
5.02273440e-01 3.71104419e-01 6.30534887e-01 -9.15144622e-01
7.23189473e-01 -3.66113633e-01 -5.88161230e-01 5.63924253e-01
-7.61762679e-01 6.87083125e-01 7.11974680e-01 -1.88446194e-01
-7.17326045e-01 -1.43627375e-01 1.95905164e-01 -6.38183355e-01
3.19356263e-01 1.20966077e+00 -2.36523524e-01 -2.05076173e-01
6.98874533e-01 9.89252627e-02 5.98787777e-02 -2.45852441e-01
1.23477018e+00 2.19157353e-01 5.33134282e-01 -9.84260321e-01
9.99710679e-01 2.73436844e-01 4.50169057e-01 -8.34146261e-01
-7.89532125e-01 -2.49847934e-01 -8.71512592e-01 3.38159651e-01
6.27564669e-01 -8.38729620e-01 -9.54576313e-01 3.50463763e-02
-1.16078520e+00 -1.44003913e-01 -5.05169630e-01 5.32674074e-01
-4.87200111e-01 6.34616196e-01 -6.75505817e-01 -6.50892794e-01
-1.36673659e-01 -4.95448709e-01 6.66380405e-01 -3.68454278e-01
-3.14834386e-01 -1.53339124e+00 1.45825490e-01 3.60819921e-02
1.87712550e-01 2.76368409e-01 1.54527426e+00 -7.83689141e-01
-7.18050539e-01 -1.08737305e-01 -1.68341741e-01 5.89471340e-01
-1.55147329e-01 2.96383761e-02 -7.67892599e-01 1.18606709e-01
-3.97104383e-01 -4.24483180e-01 9.17249441e-01 -9.78010595e-02
9.60363865e-01 -3.11792463e-01 -3.66447270e-01 7.01248288e-01
1.55932724e+00 -4.71256167e-01 5.78936636e-01 1.03967302e-01
9.37122583e-01 4.22247618e-01 2.85488904e-01 1.88685104e-01
4.91130918e-01 6.90655529e-01 2.46014982e-01 3.13853621e-01
2.52517518e-02 -7.27231860e-01 3.25179458e-01 1.34641814e+00
-3.11389148e-01 2.81860709e-01 -8.40715885e-01 5.22814035e-01
-1.96919429e+00 -1.03475213e+00 -1.20595001e-01 1.98797703e+00
1.15033436e+00 1.38149261e-02 -2.07172111e-01 1.63489997e-01
3.58585685e-01 -5.68542443e-02 -8.10530782e-02 -4.02245224e-01
-3.34732413e-01 6.95657790e-01 4.07006979e-01 5.61832488e-01
-9.10238743e-01 1.14035153e+00 5.93812752e+00 8.33483458e-01
-6.62397861e-01 5.77348247e-02 -3.38362247e-01 4.97470737e-01
-6.49834752e-01 2.30523974e-01 -7.67686605e-01 -6.70618117e-02
8.47862363e-01 -4.49157685e-01 5.86487889e-01 9.58120942e-01
-4.98559296e-01 3.57244700e-01 -1.55425179e+00 9.88871217e-01
1.49016893e-02 -1.58052146e+00 3.20040286e-01 -9.67988092e-03
5.77087045e-01 -3.20853055e-01 -2.31637537e-01 7.66342640e-01
4.97171789e-01 -1.29643273e+00 5.00027239e-01 7.56375134e-01
8.19023907e-01 -5.84892809e-01 7.70129561e-01 1.55722558e-01
-1.27897990e+00 9.29508135e-02 -4.40968722e-01 -1.94583461e-01
-2.80327629e-03 6.67432427e-01 -1.01069283e+00 1.30495119e+00
9.37074944e-02 9.67668414e-01 -6.77108526e-01 4.31620270e-01
-6.74002707e-01 4.70301270e-01 -2.14039207e-01 -2.00640298e-02
-1.10916279e-01 -2.92170644e-01 1.97324768e-01 1.20596230e+00
1.18546091e-01 -1.22329906e-01 -6.71893507e-02 1.31741250e+00
-2.44692013e-01 1.01564668e-01 -1.16972041e+00 -5.42574763e-01
3.43329191e-01 1.18301249e+00 -2.25140542e-01 -2.06916213e-01
-4.88974005e-01 7.12225497e-01 9.59789038e-01 4.73060489e-01
-7.43998468e-01 -5.83768904e-01 5.17969191e-01 -2.60160267e-01
4.38895494e-01 -4.06894237e-01 -1.71599492e-01 -1.58906353e+00
1.85472980e-01 -5.35586178e-01 4.64718312e-01 -6.56809986e-01
-1.49467206e+00 2.30455533e-01 2.89679170e-01 -7.60261118e-01
-3.35202634e-01 -1.11644471e+00 -2.15446785e-01 8.45170259e-01
-1.30983734e+00 -1.28314996e+00 -5.78661039e-02 7.57734776e-01
-6.09946728e-01 -2.90413111e-01 1.52745032e+00 1.61359161e-01
-4.53815460e-01 7.60745227e-01 -3.21544744e-02 4.07120019e-01
2.58780718e-01 -1.44163966e+00 7.74085447e-02 5.47568560e-01
7.06862032e-01 1.29785824e+00 4.29744840e-01 -5.33081591e-01
-2.03328133e+00 -9.76095378e-01 1.06302392e+00 -7.24334300e-01
1.28611386e+00 -3.63100886e-01 -1.03536284e+00 1.05837643e+00
-1.97976753e-01 2.96494722e-01 8.78112793e-01 8.05962801e-01
-9.46399808e-01 -1.89321324e-01 -7.19859004e-01 5.62837720e-01
1.31050444e+00 -1.03042412e+00 -1.04031909e+00 2.12573901e-01
9.33413804e-01 -2.68210709e-01 -1.44363320e+00 4.29483205e-01
5.55826545e-01 -5.29653132e-01 1.10229158e+00 -1.12133837e+00
4.14045602e-01 -3.74832243e-01 -4.70370114e-01 -1.44686234e+00
-3.65711153e-01 -4.64287847e-01 -1.06005704e+00 1.16220844e+00
3.14944714e-01 -7.09712505e-01 6.04740500e-01 4.22091514e-01
-1.41417831e-01 -9.84381378e-01 -7.97510564e-01 -1.14243150e+00
5.72958328e-02 -7.70196021e-01 4.33164716e-01 1.40669489e+00
7.04798281e-01 5.23157954e-01 -1.40033305e-01 2.87026703e-01
6.03008270e-01 1.38926998e-01 7.87641168e-01 -1.37037396e+00
-4.60666478e-01 -4.33694065e-01 -1.13934147e+00 -6.03622437e-01
8.17346156e-01 -1.88131189e+00 -6.53098941e-01 -1.65179753e+00
-5.94553128e-02 -3.91343862e-01 -4.76704866e-01 7.87959397e-01
1.39906570e-01 -2.06428975e-01 -1.78735629e-02 -1.12268284e-01
-4.69083786e-01 7.50975132e-01 9.07842994e-01 -2.21783519e-01
2.66253948e-02 -5.38651347e-01 -7.85796881e-01 5.70477068e-01
3.69264722e-01 -1.43008113e-01 -4.94383842e-01 -1.38810918e-01
8.54905248e-01 -1.91834286e-01 6.95902050e-01 -6.38953924e-01
3.77819657e-01 1.74010207e-03 -5.75751476e-02 -1.06068358e-01
3.16463470e-01 -9.54058528e-01 3.66737485e-01 2.47488379e-01
-2.12275013e-01 -4.66587096e-01 1.59699738e-01 7.79194772e-01
-3.20050925e-01 -2.57981330e-01 7.65918195e-02 2.32280344e-01
-1.14920676e+00 2.74931908e-01 3.94591361e-01 1.49641305e-01
8.21049929e-01 9.03575495e-03 -4.01684582e-01 -1.01453632e-01
-8.40407550e-01 1.20447502e-01 2.46828675e-01 2.03589082e-01
6.72624528e-01 -1.72147238e+00 -3.15689743e-01 2.60243118e-01
5.15445173e-01 -2.37832353e-01 -9.36704203e-02 9.74846780e-01
-3.37872803e-01 5.51529408e-01 2.19719205e-02 -3.29623580e-01
-8.45785797e-01 7.31087267e-01 1.99708253e-01 -4.40686762e-01
-7.74770141e-01 7.54404783e-01 -2.07301885e-01 -6.36965275e-01
3.94130826e-01 -7.61588454e-01 -2.53953133e-02 9.24702436e-02
2.32677847e-01 2.44735345e-01 1.41693547e-01 -3.93078834e-01
-4.81777072e-01 4.86615449e-01 1.66424900e-01 1.27284348e-01
1.29946291e+00 3.68884802e-01 -5.59984028e-01 6.20118380e-01
1.23824000e+00 7.99744390e-03 -4.16511565e-01 -6.34936512e-01
1.65455431e-01 -2.65789628e-01 -1.69855431e-01 -5.04465282e-01
-6.70741975e-01 7.89487600e-01 -1.36492699e-01 3.97938699e-01
6.42418087e-01 3.43157589e-01 3.53428006e-01 1.10677564e+00
5.25146544e-01 -8.58927786e-01 -1.23419605e-01 8.68641019e-01
9.44211543e-01 -9.51404333e-01 1.46066174e-01 -6.86893046e-01
-4.93333638e-01 1.34159422e+00 1.17521435e-01 -8.00388306e-02
9.04215217e-01 -9.46635231e-02 -5.36029696e-01 -5.99999547e-01
-6.54015899e-01 -4.67709571e-01 5.72802186e-01 8.45529377e-01
2.34702900e-01 3.69600326e-01 -1.40188083e-01 1.07122910e+00
-4.70475167e-01 2.58894593e-01 1.80591166e-01 5.34337580e-01
-1.43929914e-01 -1.13531756e+00 1.20190807e-01 4.35854256e-01
-2.03067381e-02 -2.21367747e-01 -3.49904776e-01 1.06538498e+00
2.21564606e-01 4.63974655e-01 -4.69162762e-01 -6.41497135e-01
3.75779063e-01 5.26312768e-01 9.39765513e-01 -8.48652661e-01
-5.24251424e-02 -8.84202123e-01 3.59345704e-01 -4.06939685e-01
-4.57021058e-01 -1.93353131e-01 -1.37389636e+00 -3.96240741e-01
-4.30439770e-01 2.12551549e-01 4.84840065e-01 9.82486069e-01
2.98650205e-01 4.45368022e-01 1.08658157e-01 -2.91708052e-01
-8.39573801e-01 -6.78843260e-01 -1.19448447e+00 5.71421146e-01
-1.41134694e-01 -9.90203083e-01 -3.94745857e-01 1.35373488e-01] | [8.817604064941406, 7.706940174102783] |
af26ab36-9a59-421d-a01b-2b02ae59d32f | distributed-maximization-of-submodular-plus | 1903.08351 | null | http://arxiv.org/abs/1903.08351v2 | http://arxiv.org/pdf/1903.08351v2.pdf | Distributed Maximization of Submodular plus Diversity Functions for Multi-label Feature Selection on Huge Datasets | There are many problems in machine learning and data mining which are
equivalent to selecting a non-redundant, high "quality" set of objects.
Recommender systems, feature selection, and data summarization are among many
applications of this. In this paper, we consider this problem as an
optimization problem that seeks to maximize the sum of a sum-sum diversity
function and a non-negative monotone submodular function. The diversity
function addresses the redundancy, and the submodular function controls the
predictive quality. We consider the problem in big data settings (in other
words, distributed and streaming settings) where the data cannot be stored on a
single machine or the process time is too high for a single machine. We show
that a greedy algorithm achieves a constant factor approximation of the optimal
solution in these settings. Moreover, we formulate the multi-label feature
selection problem as such an optimization problem. This formulation combined
with our algorithm leads to the first distributed multi-label feature selection
method. We compare the performance of this method with centralized multi-label
feature selection methods in the literature, and we show that its performance
is comparable or in some cases is even better than current centralized
multi-label feature selection methods. | ['Mehrdad Ghadiri', 'Mark Schmidt'] | 2019-03-20 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 1.96531996e-01 4.07959335e-02 -4.18352157e-01 -4.20928687e-01
-8.37281704e-01 -4.52516109e-01 -6.63684160e-02 6.15316868e-01
-2.20348179e-01 7.46646345e-01 1.50048554e-01 2.20034644e-01
-8.35491359e-01 -7.86158502e-01 -4.43704456e-01 -9.11819339e-01
-4.20126691e-02 8.67615104e-01 -1.38775548e-02 -1.24383807e-01
3.09784085e-01 3.70925009e-01 -1.77488697e+00 1.30453840e-01
9.21206176e-01 1.08267605e+00 2.07793608e-01 5.12623429e-01
4.63383794e-02 3.29177022e-01 -4.11497295e-01 -1.49506912e-01
5.56503296e-01 -3.44713718e-01 -9.86068130e-01 3.60076427e-01
3.72194976e-01 1.42945563e-02 1.78367585e-01 1.03733313e+00
7.22354054e-01 2.82127947e-01 5.50917506e-01 -1.54679358e+00
-3.41327488e-01 5.84092498e-01 -8.96273434e-01 1.10124409e-01
2.80794889e-01 -6.65341794e-01 1.34662509e+00 -7.23471761e-01
6.63490832e-01 1.08069062e+00 3.96349549e-01 2.61017829e-01
-1.17988205e+00 -1.68274507e-01 2.34225541e-01 2.36407691e-03
-1.34691310e+00 -3.63966435e-01 2.84947902e-01 -2.45778263e-01
6.35600209e-01 6.88498378e-01 6.13611817e-01 -2.26242281e-02
3.01640481e-01 7.21896350e-01 6.12892032e-01 -5.08897603e-01
4.67833847e-01 2.43468389e-01 4.96433586e-01 7.00094998e-01
6.44741416e-01 -4.14915949e-01 -6.44533992e-01 -8.14624846e-01
2.84214988e-02 3.67372274e-01 -2.82142848e-01 -5.26122451e-01
-1.09188247e+00 1.15136099e+00 -8.03966299e-02 1.07065685e-01
-6.87501192e-01 6.66677207e-02 2.63564885e-01 8.21738601e-01
6.70855880e-01 6.10908926e-01 -5.73276520e-01 2.64009178e-01
-1.02600873e+00 2.50710607e-01 1.22867525e+00 1.07544255e+00
8.69945824e-01 -3.78416359e-01 -6.88068420e-02 8.21940720e-01
1.60200074e-01 4.75316674e-01 3.09112549e-01 -1.00833952e+00
2.81250119e-01 6.81989372e-01 1.73634127e-01 -9.65728223e-01
-8.07374299e-01 -4.42784071e-01 -7.05509484e-01 -1.91061258e-01
8.11245292e-03 -4.26684320e-01 -1.32690668e-01 1.66480148e+00
7.99579442e-01 -3.27467769e-01 -1.29567549e-01 8.55066836e-01
5.23091733e-01 7.78294265e-01 -3.42358649e-01 -1.28758729e+00
1.18581200e+00 -8.84073734e-01 -7.40103483e-01 2.86711931e-01
7.83129036e-01 -7.29352117e-01 5.02995849e-01 6.99995697e-01
-1.14925539e+00 1.42117530e-01 -7.99926579e-01 2.53523499e-01
3.36913355e-02 -5.26626632e-02 8.18746269e-01 5.11003673e-01
-1.14999950e+00 5.43623149e-01 -2.80487984e-01 -6.95220113e-01
1.69038530e-02 7.88186193e-01 -2.82482773e-01 -1.03058457e-01
-6.92901373e-01 5.92098355e-01 2.39628211e-01 -3.46654862e-01
-3.75724673e-01 -6.18425548e-01 -4.01042879e-01 3.98429222e-02
8.17658365e-01 -1.04060256e+00 1.16482151e+00 -9.92838323e-01
-1.18526816e+00 6.68424487e-01 -1.54421568e-01 -1.64440930e-01
1.66491777e-01 6.23207316e-02 -1.49312258e-01 1.85583174e-01
1.71136826e-01 7.52546191e-02 5.74027359e-01 -1.03889668e+00
-1.18913746e+00 -7.24853456e-01 1.10450171e-01 4.42165047e-01
-6.06321812e-01 1.50476784e-01 -1.21098191e-01 -2.66017288e-01
1.45516664e-01 -1.05179894e+00 -5.65920711e-01 -4.07454908e-01
-3.78779531e-01 -5.05243361e-01 5.06353438e-01 -1.23909190e-01
1.43779743e+00 -2.05994487e+00 4.10456777e-01 5.07516563e-01
3.84909779e-01 -2.67885536e-01 -1.53813541e-01 8.88624251e-01
4.16326493e-01 2.02601641e-01 1.35267470e-02 -2.51443237e-01
-7.32063055e-02 8.32096040e-02 7.84241781e-02 1.03788209e+00
-2.91840762e-01 3.53284121e-01 -8.27598989e-01 -4.87849772e-01
-3.56162608e-01 -1.14077874e-01 -6.22807741e-01 -8.52307007e-02
-1.17710911e-01 -3.03576671e-04 -7.03670979e-01 7.40066350e-01
6.08964264e-01 -4.40645278e-01 4.04996544e-01 8.65336135e-02
-1.78310364e-01 -3.07562232e-01 -1.73038018e+00 1.34869850e+00
-3.82318407e-01 8.27417001e-02 3.18743080e-01 -1.05334163e+00
7.33941734e-01 2.48789072e-01 1.23856688e+00 -1.99542165e-01
5.78640103e-02 3.70710552e-01 -4.07852918e-01 -4.56782907e-01
7.42366970e-01 -2.43945077e-01 -2.24338859e-01 8.84087741e-01
-1.16488084e-01 3.33609104e-01 3.65030259e-01 4.22969729e-01
1.17577398e+00 -8.00533414e-01 6.72456861e-01 -6.68542683e-01
3.06905866e-01 6.69924840e-02 8.59585345e-01 7.82894492e-01
3.16688329e-01 5.06419182e-01 7.03053236e-01 -3.82191211e-01
-8.44804823e-01 -2.98551410e-01 -7.04868808e-02 1.45071912e+00
3.72069538e-01 -6.59648538e-01 -4.15507555e-01 -5.87951779e-01
4.15224969e-01 4.27626193e-01 -5.53017378e-01 -1.09676696e-01
-3.21197689e-01 -9.14236128e-01 -2.62872964e-01 -5.00530899e-02
-7.64475614e-02 -6.08813286e-01 -5.57944715e-01 3.57818931e-01
-2.04828940e-02 -5.81366777e-01 -7.38774538e-01 2.71971852e-01
-8.58247161e-01 -1.08024645e+00 -3.96181077e-01 -6.71366274e-01
6.61961615e-01 7.13618875e-01 1.08017683e+00 1.93981364e-01
-1.43390328e-01 4.34916288e-01 -5.81758320e-01 -4.85152215e-01
3.08264084e-02 2.47403517e-01 3.59905899e-01 3.77085716e-01
4.97844778e-02 -3.12016368e-01 -4.71986979e-01 3.22849005e-01
-1.05894899e+00 -3.27386349e-01 3.49220634e-01 8.40693653e-01
1.09079778e+00 2.69806981e-01 1.07701063e+00 -1.27268136e+00
7.50615537e-01 -9.28588212e-01 -5.65049767e-01 5.59895635e-01
-1.07360113e+00 1.39326036e-01 6.13161862e-01 -2.24791154e-01
-6.16371751e-01 3.85810316e-01 3.28262389e-01 6.43804744e-02
3.31079543e-01 7.00369656e-01 -3.75651836e-01 -1.02413714e-01
3.67317140e-01 5.23351654e-02 3.44015425e-03 -5.60966849e-01
3.72411847e-01 8.86647522e-01 -1.64274871e-01 -2.38945276e-01
3.16490889e-01 5.70287645e-01 3.97416204e-01 -7.64424860e-01
-8.48811448e-01 -1.00025833e+00 -4.18354422e-01 -1.17039554e-01
3.40062767e-01 -6.94849133e-01 -8.60223114e-01 -2.65832115e-02
-6.95831597e-01 2.54840583e-01 -6.38193965e-01 3.58101964e-01
-7.52430499e-01 2.90443391e-01 -1.88091934e-01 -8.57322812e-01
-6.84664369e-01 -7.58073032e-01 9.75612819e-01 2.06659973e-01
1.74627900e-01 -6.18051291e-01 2.61029840e-01 1.11325204e-01
2.57542908e-01 3.12624574e-01 8.71441662e-01 -9.87430990e-01
-3.64176989e-01 -3.45696867e-01 1.92288548e-01 -1.36069745e-01
2.01084074e-02 -1.98676363e-01 -3.14298034e-01 -6.46711409e-01
3.50207812e-03 -2.76126623e-01 8.77958298e-01 5.34723103e-01
8.95440638e-01 -5.20755410e-01 -4.43402112e-01 5.20565987e-01
1.64595830e+00 9.03135573e-04 -1.49849340e-01 1.64382845e-01
4.97343093e-01 7.12524116e-01 1.03419244e+00 1.22126877e+00
3.91507953e-01 7.44036436e-01 3.04783911e-01 2.15171456e-01
4.18379456e-01 2.97732443e-01 1.02055244e-01 8.53736579e-01
1.68354586e-01 -5.43811023e-01 -6.25371039e-01 5.13069630e-01
-2.32277298e+00 -8.60044599e-01 -3.09993267e-01 2.57944179e+00
6.54634595e-01 -5.69986045e-01 6.31549776e-01 7.80148506e-02
8.19346011e-01 -2.86353648e-01 -6.81861579e-01 -6.12222970e-01
-3.31729710e-01 -2.68851191e-01 6.67659879e-01 2.56506413e-01
-1.11821675e+00 4.74620670e-01 6.35047817e+00 6.37098014e-01
-8.39283884e-01 3.22654277e-01 4.39470649e-01 -7.37342656e-01
-4.90527809e-01 -9.69034880e-02 -9.41999435e-01 3.27363223e-01
1.07819271e+00 -6.64025843e-01 4.08489764e-01 7.62802303e-01
2.01155245e-01 -3.10349852e-01 -1.10068357e+00 9.94138300e-01
1.51955232e-01 -1.06336510e+00 -1.26203626e-01 6.25973642e-01
1.08636034e+00 -1.10001184e-01 -9.48465765e-02 -1.74612477e-01
2.65527338e-01 -6.14725769e-01 5.15483618e-01 4.82721210e-01
5.98585069e-01 -1.29084277e+00 7.00725198e-01 5.71729600e-01
-1.12621677e+00 -6.26423240e-01 -6.73701942e-01 1.38694242e-01
4.34241205e-01 1.20344520e+00 -2.46480465e-01 8.29287589e-01
3.90094101e-01 4.62851137e-01 -3.31869781e-01 1.54565990e+00
5.71726322e-01 2.72430986e-01 -5.82254052e-01 -2.04641402e-01
7.72701874e-02 -2.98664927e-01 5.89418828e-01 9.72507656e-01
5.39983630e-01 3.20357651e-01 6.42033160e-01 -1.76745966e-01
-2.74041861e-01 8.74734163e-01 -6.37340486e-01 2.46174678e-01
5.49377799e-01 1.40631461e+00 -7.27383018e-01 -2.82699972e-01
-4.77452934e-01 7.68578827e-01 3.17541033e-01 -1.11452430e-01
-3.42544883e-01 -4.33972657e-01 6.41808033e-01 -5.95969195e-03
2.37470761e-01 2.87779152e-01 -3.31260473e-01 -1.08792078e+00
1.39432037e-02 -7.73898661e-01 9.82877076e-01 5.43363728e-02
-1.29640925e+00 4.27156329e-01 -1.30075261e-01 -1.11017025e+00
-2.22944632e-01 3.21953185e-02 -1.53908476e-01 3.76095951e-01
-1.30646420e+00 -6.73494995e-01 -3.55573036e-02 6.96148634e-01
3.80546719e-01 -2.76209116e-01 6.83943093e-01 4.50371653e-01
-5.20166576e-01 6.05733216e-01 7.79112279e-01 -9.32712257e-01
6.87087715e-01 -1.34791684e+00 -4.37392563e-01 5.26190341e-01
-1.71591207e-01 2.07449615e-01 6.71536028e-01 -5.87138176e-01
-1.92888153e+00 -1.09418869e+00 1.24284887e+00 4.38069068e-02
2.40702674e-01 -1.22187994e-01 -5.11713147e-01 3.21205467e-01
4.99491990e-02 3.50745544e-02 1.21212292e+00 3.26999247e-01
1.76226273e-01 -3.51585269e-01 -1.46357739e+00 1.11069214e-02
1.11862183e+00 2.59448767e-01 6.63466528e-02 9.12502885e-01
6.27050638e-01 4.69509177e-02 -1.10377789e+00 2.26195097e-01
4.40034807e-01 -6.75758123e-01 6.46070957e-01 -7.92137623e-01
4.55974601e-02 -1.13214716e-01 -4.77300167e-01 -1.40867686e+00
-6.40735090e-01 -7.78572857e-01 -2.09277272e-01 1.18521202e+00
4.33658957e-01 -5.82296193e-01 7.35082507e-01 4.41215724e-01
1.23617306e-01 -1.04947186e+00 -8.99111390e-01 -7.96340346e-01
-2.03650758e-01 2.40906239e-01 6.16352558e-01 8.58834267e-01
2.87484080e-01 4.74686146e-01 -6.77382708e-01 4.09885235e-02
8.02920282e-01 8.22140872e-01 4.45311069e-01 -1.72854507e+00
-3.56668323e-01 -1.46793321e-01 -1.94964290e-01 -7.50690997e-01
3.60842124e-02 -1.08565414e+00 -1.47583738e-01 -1.67284191e+00
6.38685465e-01 -6.30990684e-01 -3.12024742e-01 2.83135235e-01
2.32760585e-03 -2.46539384e-01 2.73855329e-01 4.90667433e-01
-1.21400869e+00 3.44700456e-01 8.78982961e-01 7.76455775e-02
-5.42888522e-01 4.82569844e-01 -1.25775945e+00 4.12629008e-01
7.53153324e-01 -8.42108607e-01 -3.65456164e-01 -1.65012509e-01
6.05464816e-01 4.87903237e-01 -4.87987936e-01 -5.67645609e-01
4.87196654e-01 -5.72748721e-01 -1.86389402e-01 -6.08071685e-01
6.14993200e-02 -9.05969799e-01 2.70279199e-01 5.18959403e-01
-4.53393519e-01 3.16215187e-01 -5.73753536e-01 8.38309169e-01
-8.16577747e-02 -5.62680960e-01 7.45974481e-01 -1.17977589e-01
-3.91065508e-01 5.99136174e-01 -4.46976990e-01 2.74235159e-02
1.40993917e+00 5.70514090e-02 -2.28962541e-01 -3.76639873e-01
-5.49571276e-01 6.43123031e-01 4.19127792e-01 2.18838617e-01
4.89388883e-01 -1.29265523e+00 -1.01403928e+00 -1.75467491e-01
1.86074704e-01 -1.49067327e-01 1.69016510e-01 1.01774263e+00
-1.36084616e-01 3.69415760e-01 4.05696407e-02 -2.69593656e-01
-1.52931905e+00 6.67542219e-01 -7.48260394e-02 -3.07957560e-01
-3.37012708e-01 6.91729426e-01 -1.35518327e-01 -2.77614802e-01
1.76588342e-01 2.51752108e-01 -2.58241922e-01 6.61715746e-01
6.37184799e-01 7.69943416e-01 2.22429886e-01 -5.30992150e-01
-5.04492640e-01 4.25987452e-01 -1.33214474e-01 -1.24015985e-02
1.67360234e+00 -4.65746582e-01 -5.67436278e-01 3.39192599e-01
1.22093689e+00 7.00286478e-02 -4.53400016e-01 -5.28356850e-01
6.28578216e-02 -4.71653074e-01 5.67013696e-02 -4.92832601e-01
-1.21402276e+00 1.06695876e-01 3.95715624e-01 6.63351715e-01
1.47085333e+00 2.05919430e-01 6.85449541e-01 6.49834991e-01
6.32656872e-01 -1.36702967e+00 -1.39411733e-01 3.52969497e-01
6.99732900e-01 -1.18311906e+00 2.88934439e-01 -4.15434271e-01
-7.54893184e-01 1.00605738e+00 2.79243112e-01 -8.53290781e-02
9.21300411e-01 2.41446614e-01 -3.63020688e-01 -2.96971411e-01
-1.26770484e+00 -2.73307413e-01 1.42387643e-01 8.44294131e-02
2.28626117e-01 3.38669270e-01 -1.08710158e+00 7.33520627e-01
-7.85795078e-02 -1.22829482e-01 6.33731067e-01 1.13106704e+00
-1.01267993e+00 -1.18400180e+00 -3.53793740e-01 1.14298546e+00
-7.40932047e-01 1.53915018e-01 -4.06398177e-01 1.25213325e-01
8.34652260e-02 1.32915068e+00 -1.89694643e-01 -4.50911313e-01
4.11216706e-01 -5.06275557e-02 1.49303213e-01 -8.83755863e-01
-8.35061073e-01 3.06862921e-01 7.13699758e-02 -4.46244478e-01
-3.27074260e-01 -8.71857584e-01 -1.13122356e+00 -5.12400568e-01
-9.00136650e-01 5.34435868e-01 7.92979360e-01 7.28560328e-01
5.93736827e-01 1.78718925e-01 1.21968007e+00 -3.78353029e-01
-1.05269754e+00 -5.94467342e-01 -1.28616178e+00 1.22225769e-01
2.56246954e-01 -3.90132129e-01 -1.84453025e-01 -3.12493235e-01] | [6.6453094482421875, 4.9015326499938965] |
129e8a42-adcb-43e5-9b6e-76f4cb411720 | physics-based-motion-retargeting-from-sparse | 2307.01938 | null | https://arxiv.org/abs/2307.01938v1 | https://arxiv.org/pdf/2307.01938v1.pdf | Physics-based Motion Retargeting from Sparse Inputs | Avatars are important to create interactive and immersive experiences in virtual worlds. One challenge in animating these characters to mimic a user's motion is that commercial AR/VR products consist only of a headset and controllers, providing very limited sensor data of the user's pose. Another challenge is that an avatar might have a different skeleton structure than a human and the mapping between them is unclear. In this work we address both of these challenges. We introduce a method to retarget motions in real-time from sparse human sensor data to characters of various morphologies. Our method uses reinforcement learning to train a policy to control characters in a physics simulator. We only require human motion capture data for training, without relying on artist-generated animations for each avatar. This allows us to use large motion capture datasets to train general policies that can track unseen users from real and sparse data in real-time. We demonstrate the feasibility of our approach on three characters with different skeleton structure: a dinosaur, a mouse-like creature and a human. We show that the avatar poses often match the user surprisingly well, despite having no sensor information of the lower body available. We discuss and ablate the important components in our framework, specifically the kinematic retargeting step, the imitation, contact and action reward as well as our asymmetric actor-critic observations. We further explore the robustness of our method in a variety of settings including unbalancing, dancing and sports motions. | ['Alexander Winkler', 'Michiel Van de Panne', 'Yuting Ye', 'Jungdam Won', 'Daniele Reda'] | 2023-07-04 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [ 1.19403312e-02 2.79749423e-01 1.20199122e-01 3.17149252e-01
-3.68816555e-01 -8.48174751e-01 4.93109286e-01 -4.36118811e-01
-4.87247795e-01 6.59649014e-01 -1.38450623e-01 6.58421144e-02
3.43665689e-01 -4.40055966e-01 -7.13674724e-01 -4.51287150e-01
-1.05666161e-01 7.07116365e-01 6.35819614e-01 -6.54766798e-01
3.60769518e-02 6.84220195e-01 -1.46171236e+00 -2.61991739e-01
5.72457314e-01 4.68900084e-01 3.28779250e-01 1.37426591e+00
6.22969508e-01 6.30813420e-01 -6.47351980e-01 -9.22768936e-02
5.64065874e-01 -4.62909967e-01 -4.68517095e-01 3.47346455e-01
4.59551811e-01 -3.70280534e-01 -4.35211897e-01 5.99504530e-01
7.98087597e-01 6.63873851e-01 3.74291450e-01 -1.23709595e+00
7.39378333e-02 2.87064344e-01 -6.19675457e-01 -6.54629543e-02
8.76299679e-01 5.63839197e-01 7.18380928e-01 -2.64188439e-01
1.11512661e+00 1.14806080e+00 7.73308694e-01 1.12924755e+00
-1.18022752e+00 -5.60590804e-01 1.08382024e-01 -4.90518361e-02
-1.25891268e+00 -3.88910174e-01 6.82293892e-01 -5.15359581e-01
5.51934481e-01 2.70806253e-01 1.21414697e+00 1.66316390e+00
3.92290875e-02 8.09677780e-01 6.55836523e-01 -1.66878998e-01
3.96554828e-01 -2.13354185e-01 -6.46429539e-01 8.12666178e-01
-1.39836192e-01 2.23160192e-01 -2.94052333e-01 -3.75240773e-01
1.57904208e+00 -1.96347699e-01 -2.56983429e-01 -9.24609780e-01
-1.36188567e+00 7.27440059e-01 6.19499423e-02 -1.05435014e-01
-4.81758684e-01 6.65424407e-01 1.66753888e-01 2.48918794e-02
-4.58621830e-01 7.84930646e-01 -5.21481633e-01 -6.94119811e-01
-5.08860171e-01 9.63139117e-01 1.06504345e+00 1.00057101e+00
9.06457379e-02 5.16967595e-01 2.32564092e-01 4.20864940e-01
8.26474652e-02 4.79059458e-01 4.63519573e-01 -1.62504256e+00
2.84285963e-01 -1.08458884e-02 4.94622260e-01 -8.51352930e-01
-5.57206690e-01 4.22057398e-02 -3.71407270e-01 7.57541656e-01
6.49832845e-01 -6.58521414e-01 -8.94287944e-01 1.90886283e+00
6.75799608e-01 5.11381209e-01 -1.30029097e-01 1.35703063e+00
3.55716914e-01 3.77521217e-01 -4.19187509e-02 1.58644468e-01
1.21122611e+00 -8.44003797e-01 -3.43789697e-01 -2.37177879e-01
3.24537158e-01 -4.37578768e-01 1.19907105e+00 4.83400464e-01
-1.33044982e+00 -4.90934223e-01 -9.53460813e-01 3.40892822e-01
2.43524253e-01 -2.61852920e-01 6.05822027e-01 4.82698202e-01
-6.43562973e-01 1.00570905e+00 -1.35129082e+00 -5.42214453e-01
-2.10138604e-01 6.13335550e-01 -3.70181590e-01 5.36103785e-01
-1.03851998e+00 9.73367929e-01 2.77889688e-02 -1.25857189e-01
-1.17299998e+00 -5.13516665e-01 -9.93150771e-01 -2.41218746e-01
6.02715850e-01 -9.48361754e-01 1.59909689e+00 -1.02315402e+00
-2.21546435e+00 5.79038501e-01 4.16665643e-01 -1.38324723e-01
1.07332242e+00 -3.26584905e-01 -3.76163125e-02 1.07475571e-01
-7.05524832e-02 5.65170169e-01 8.10032010e-01 -1.31348145e+00
-4.70157325e-01 -1.84915096e-01 1.52031884e-01 5.61312556e-01
3.27588469e-01 -3.37592900e-01 -6.82396889e-01 -7.37854302e-01
-3.74547839e-02 -1.43613410e+00 -6.37694001e-01 2.50476241e-01
-3.75065625e-01 3.73482972e-01 8.38309944e-01 -2.04080343e-01
6.24076605e-01 -1.75147426e+00 7.24273860e-01 2.62960136e-01
3.99890766e-02 1.22952290e-01 -1.23266391e-01 3.52527440e-01
1.63357228e-01 -1.74353719e-01 5.43214530e-02 -2.56576449e-01
-1.45141885e-01 4.16862965e-01 -6.90189749e-02 5.77104926e-01
-2.29114443e-01 8.41739833e-01 -1.08351421e+00 -5.15245378e-01
1.40550286e-01 5.39526582e-01 -7.22168922e-01 4.06060696e-01
-3.20005774e-01 1.20635366e+00 -7.28983581e-01 5.18430531e-01
6.08785897e-02 9.16149989e-02 3.04820091e-01 1.91801116e-01
-8.22160915e-02 -1.15419529e-01 -1.64028752e+00 1.94364786e+00
-3.57485473e-01 3.27068329e-01 5.38591266e-01 -5.49759805e-01
5.45082510e-01 4.33142126e-01 6.20205283e-01 -3.66055191e-01
3.49606425e-01 -5.85956452e-03 1.49021834e-01 -6.69975460e-01
6.06225669e-01 -3.79925907e-01 -2.95187116e-01 4.23120797e-01
-2.66145825e-01 -6.09295130e-01 -7.15207756e-02 -4.92254756e-02
1.33067918e+00 6.53768122e-01 1.63424611e-01 1.48237631e-01
1.28746539e-01 6.59693331e-02 6.84853971e-01 7.23859489e-01
-1.85249910e-01 9.44627941e-01 3.32547694e-01 -2.94935435e-01
-1.41340089e+00 -1.16796958e+00 5.49732268e-01 1.10296392e+00
3.68625522e-01 -1.93721950e-01 -6.27667546e-01 -1.90680936e-01
9.76702124e-02 4.34032172e-01 -6.16241574e-01 -1.67527840e-01
-1.18729222e+00 -2.22520635e-01 6.68602645e-01 6.49598002e-01
5.52426092e-02 -1.05452168e+00 -1.48612845e+00 2.95003504e-01
3.40022030e-03 -1.23296142e+00 -5.16695023e-01 -2.14822903e-01
-7.00664461e-01 -1.02139318e+00 -8.37301850e-01 -4.63961720e-01
3.12416971e-01 -1.74449742e-01 9.42388058e-01 -9.63334143e-02
-3.91784549e-01 7.08858550e-01 -3.40372980e-01 -1.51760131e-01
-6.34952068e-01 -6.66842163e-02 4.26274776e-01 -2.97131985e-01
-7.23454654e-01 -7.98449337e-01 -7.07923234e-01 5.45880914e-01
-5.08946180e-01 1.70946777e-01 1.26305923e-01 5.85419536e-01
4.51484054e-01 -4.43423122e-01 -1.23425148e-01 -7.32134402e-01
5.66183686e-01 -2.07713038e-01 -5.99889398e-01 -2.16000676e-01
6.49739504e-02 -1.03675894e-01 6.67113841e-01 -1.24059248e+00
-6.82718635e-01 6.94481254e-01 -1.63667917e-01 -7.04344988e-01
-5.85068949e-02 -1.39494061e-01 -9.85883102e-02 -1.96756810e-01
8.71411800e-01 -9.17160138e-02 6.43188804e-02 -2.74026722e-01
4.74793494e-01 1.14031829e-01 9.98412073e-01 -9.57497656e-01
1.00739729e+00 5.14941692e-01 1.55647323e-01 -9.47035670e-01
-3.51062380e-02 -1.77939668e-01 -6.85630083e-01 -3.81416708e-01
8.88717294e-01 -8.00648332e-01 -1.49288583e+00 3.36323828e-01
-1.10126936e+00 -8.32606733e-01 -6.14317715e-01 6.49885535e-01
-1.28032053e+00 4.02896047e-01 -7.60109067e-01 -7.59521782e-01
5.32456376e-02 -1.33738685e+00 1.12877476e+00 4.17516530e-01
-7.50625968e-01 -8.01871359e-01 4.59285021e-01 1.51410103e-01
1.64196819e-01 9.16711569e-01 2.73825705e-01 -3.29063714e-01
-4.10950065e-01 -2.52138108e-01 6.23607397e-01 -3.21239054e-01
-3.34823281e-02 1.30601868e-01 -5.87580442e-01 -5.17498910e-01
-3.31071883e-01 -4.99928802e-01 1.47483215e-01 3.09965074e-01
7.47320175e-01 -2.04982311e-01 -2.71770358e-01 6.32635713e-01
9.79632676e-01 2.40868464e-01 5.01358569e-01 3.67496014e-01
1.04635823e+00 5.35321712e-01 7.21848965e-01 7.32815802e-01
3.03556740e-01 1.13758075e+00 6.70556486e-01 2.99189091e-02
1.41935468e-01 -4.72866297e-01 3.66158962e-01 3.39906216e-01
-6.91478729e-01 -2.11485758e-01 -8.02581072e-01 3.16464305e-01
-1.95782185e+00 -9.61607575e-01 9.60399583e-02 2.19847131e+00
6.16015971e-01 5.65726347e-02 5.67468226e-01 -6.21431507e-02
6.18003309e-01 -1.60365209e-01 -8.10154855e-01 -3.21520746e-01
2.44521692e-01 2.33812064e-01 5.32889903e-01 6.29139125e-01
-8.80277038e-01 1.15942419e+00 6.48723316e+00 2.67402470e-01
-1.13175321e+00 -1.82499066e-01 -6.77094087e-02 -2.73236215e-01
-9.56905782e-02 7.87876621e-02 -4.38946009e-01 2.82596588e-01
6.95966899e-01 5.67259453e-02 4.66502368e-01 9.65622306e-01
4.96180505e-01 -5.94957434e-02 -1.29134774e+00 8.63769352e-01
-5.18031232e-02 -1.08295906e+00 -3.21178913e-01 -2.09079124e-02
5.41134596e-01 -2.37966537e-01 4.72693564e-03 2.47421339e-01
8.06560278e-01 -1.06105244e+00 9.16819155e-01 4.10774261e-01
6.87376797e-01 -5.84909499e-01 1.00929879e-01 5.63639820e-01
-1.10582292e+00 4.32881601e-02 -2.68197078e-02 -1.71324447e-01
4.68481243e-01 -6.21820986e-01 -7.24646509e-01 -1.10289706e-02
3.26824516e-01 2.60656923e-01 -3.70638371e-02 9.92022097e-01
-1.85824871e-01 1.49854735e-01 -5.40922642e-01 -4.55896556e-02
3.01692896e-02 -1.31274536e-01 9.79582012e-01 8.48953426e-01
2.19718203e-01 4.35682833e-01 4.82384264e-01 7.48310030e-01
3.30432981e-01 -2.79857010e-01 -5.05564749e-01 1.95559055e-01
1.91054001e-01 1.19365656e+00 -7.05664217e-01 -4.50149626e-02
7.94433430e-02 1.31557167e+00 9.40300599e-02 3.05044144e-01
-1.08135211e+00 -1.81333080e-01 9.65018570e-01 5.22796512e-01
3.45019668e-01 -7.05022573e-01 1.23123370e-01 -1.30947649e+00
-2.16300026e-01 -1.10655200e+00 1.42914832e-01 -8.52465749e-01
-7.50744224e-01 3.94049346e-01 3.46705355e-02 -1.38182735e+00
-8.98057818e-01 -2.60930240e-01 -6.04214847e-01 5.56961298e-01
-5.34263849e-01 -9.58269656e-01 -4.76513863e-01 7.01191545e-01
6.04291975e-01 1.60172004e-02 6.86795413e-01 -5.24858758e-03
-4.03914452e-01 3.93757999e-01 -3.59460354e-01 1.68702707e-01
5.45664549e-01 -1.20105326e+00 7.21105874e-01 5.00737906e-01
-4.64395992e-02 3.36680025e-01 1.34406447e+00 -7.16339231e-01
-1.71556425e+00 -5.59593558e-01 -1.31823689e-01 -7.53355980e-01
7.40725279e-01 -4.69711900e-01 -7.48410165e-01 8.50111783e-01
-3.14396888e-01 2.04713330e-01 1.81088880e-01 -2.93227911e-01
4.46667485e-02 4.20903921e-01 -1.00933695e+00 1.15230262e+00
1.09233916e+00 -8.84402990e-02 -4.84822184e-01 4.16973270e-02
4.63085413e-01 -1.23676038e+00 -7.29120135e-01 2.02366695e-01
9.89735842e-01 -7.59276330e-01 1.09357250e+00 -8.74369979e-01
1.81187496e-01 -3.87640238e-01 1.44208651e-02 -1.31642020e+00
-1.87975466e-01 -1.07618904e+00 -2.64885694e-01 5.76930404e-01
2.96393819e-02 -5.91650158e-02 1.34024382e+00 8.84083390e-01
2.55353570e-01 -5.39995849e-01 -8.33282948e-01 -8.05022120e-01
7.26221800e-02 -3.70849311e-01 2.60158956e-01 8.25316250e-01
5.53808734e-02 3.60609084e-01 -8.14522445e-01 2.88783282e-01
4.54951435e-01 -5.71349487e-02 1.36973441e+00 -1.07533634e+00
-9.55009162e-01 -1.86837941e-01 -6.58949435e-01 -1.12324607e+00
1.60964429e-01 -3.20032954e-01 1.99668929e-01 -1.09469831e+00
-4.41653281e-03 -4.56280887e-01 5.47876596e-01 2.46387541e-01
-4.83986512e-02 9.37314928e-02 4.87479538e-01 2.69861519e-01
-4.89200503e-01 4.92886275e-01 1.65923548e+00 2.82172740e-01
-6.84448063e-01 3.67002934e-01 -1.02746241e-01 1.08894348e+00
6.43686056e-01 -3.72525215e-01 -4.35615361e-01 -3.22982877e-01
1.09700695e-01 7.52775908e-01 6.29732966e-01 -1.13405454e+00
2.58117616e-01 -4.51693386e-01 4.55700248e-01 7.61771873e-02
8.19519460e-01 -7.96512187e-01 4.99422520e-01 7.86640406e-01
-2.25036070e-01 3.87796223e-01 1.82719171e-01 6.66562974e-01
4.32197154e-01 -3.25724781e-02 8.84786367e-01 -4.87901449e-01
-6.10753596e-01 3.55897963e-01 -6.51316702e-01 2.21455187e-01
1.30984354e+00 -4.58796144e-01 2.78301924e-01 -8.44042599e-01
-1.11044657e+00 2.54249990e-01 9.15386140e-01 4.52029377e-01
5.54222584e-01 -1.12361014e+00 -5.59864938e-01 2.27841642e-02
-2.32735202e-01 5.22728004e-02 1.62336364e-01 4.99909848e-01
-9.50641096e-01 -2.51377910e-01 -5.57941914e-01 -6.35478079e-01
-1.37138391e+00 3.94776940e-01 5.85585535e-01 -2.40792683e-03
-9.06198204e-01 5.28776944e-01 1.47775650e-01 -4.89780843e-01
1.35931343e-01 -1.16328202e-01 -1.59852982e-01 -3.88790429e-01
9.26114097e-02 5.33470154e-01 -4.79306400e-01 -7.34020174e-01
-1.52473867e-01 8.05116415e-01 2.68048614e-01 -7.06943512e-01
1.10951042e+00 5.48874885e-02 6.53307140e-01 4.46990162e-01
6.63967371e-01 2.39947423e-01 -1.77794921e+00 2.93541044e-01
-3.80576521e-01 -5.18449545e-01 -6.19565547e-01 -2.76572764e-01
-9.75942254e-01 5.70881069e-01 4.83141065e-01 -2.24775538e-01
5.53113818e-01 -2.37518437e-02 8.58441114e-01 4.26465124e-01
5.99703670e-01 -1.18309546e+00 5.30094385e-01 3.75890762e-01
8.60248029e-01 -9.22892869e-01 8.65889899e-03 -2.79062599e-01
-1.06754696e+00 9.41952527e-01 9.02791560e-01 -5.39033532e-01
1.66785091e-01 5.93810558e-01 2.60223955e-01 -5.25846928e-02
-5.82050979e-01 -7.89722651e-02 1.18692489e-02 7.87166297e-01
7.48932734e-02 -1.29571892e-02 6.43703714e-02 3.43278855e-01
-5.86973608e-01 -2.52048910e-01 9.99301016e-01 1.25491381e+00
-2.91001856e-01 -1.17925966e+00 -5.71172595e-01 -2.18812853e-01
-4.71412390e-01 5.14131188e-01 -4.21256989e-01 1.19802320e+00
-9.17804763e-02 5.03667414e-01 6.03756905e-02 -5.24441123e-01
7.64199436e-01 -3.69277805e-01 8.10711265e-01 -7.56514907e-01
-6.66061223e-01 1.75886080e-01 -9.02094692e-02 -9.19470072e-01
-4.86900620e-02 -8.77500236e-01 -1.58559215e+00 -4.54948246e-01
-5.05723469e-02 -6.39535040e-02 5.78265071e-01 4.47800249e-01
9.00321528e-02 4.04183507e-01 3.43431890e-01 -1.55838490e+00
-5.14233291e-01 -5.39350867e-01 -5.68794072e-01 5.89384675e-01
4.41348732e-01 -7.57777750e-01 -1.01015262e-01 3.69228274e-02] | [5.106180667877197, 0.7020900249481201] |
2f8c8587-233b-4df6-bcd5-f57ac23fce82 | tgrl-an-algorithm-for-teacher-guided | 2307.03186 | null | https://arxiv.org/abs/2307.03186v1 | https://arxiv.org/pdf/2307.03186v1.pdf | TGRL: An Algorithm for Teacher Guided Reinforcement Learning | Learning from rewards (i.e., reinforcement learning or RL) and learning to imitate a teacher (i.e., teacher-student learning) are two established approaches for solving sequential decision-making problems. To combine the benefits of these different forms of learning, it is common to train a policy to maximize a combination of reinforcement and teacher-student learning objectives. However, without a principled method to balance these objectives, prior work used heuristics and problem-specific hyperparameter searches to balance the two objectives. We present a $\textit{principled}$ approach, along with an approximate implementation for $\textit{dynamically}$ and $\textit{automatically}$ balancing when to follow the teacher and when to use rewards. The main idea is to adjust the importance of teacher supervision by comparing the agent's performance to the counterfactual scenario of the agent learning without teacher supervision and only from rewards. If using teacher supervision improves performance, the importance of teacher supervision is increased and otherwise it is decreased. Our method, $\textit{Teacher Guided Reinforcement Learning}$ (TGRL), outperforms strong baselines across diverse domains without hyper-parameter tuning. | ['Pulkit Agrawal', 'Aviv Tamar', 'Zhang-Wei Hong', 'Idan Shenfeld'] | 2023-07-06 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 2.83448786e-01 5.44170439e-01 -5.07943094e-01 -4.23888564e-01
-7.80929506e-01 -5.16794384e-01 6.41495109e-01 2.20005006e-01
-9.76729870e-01 1.15599608e+00 -1.01340488e-01 -6.42469764e-01
-4.75935578e-01 -7.45780051e-01 -7.11711407e-01 -9.57676411e-01
2.39356279e-01 6.90524578e-01 1.52693629e-01 -1.69983774e-01
4.97904181e-01 1.17708459e-01 -1.51545882e+00 -2.31538579e-01
1.02830458e+00 8.20042133e-01 3.00219357e-01 6.01022601e-01
-9.95317623e-02 1.05549705e+00 -7.81293035e-01 -2.23499939e-01
3.52823585e-01 -6.71430051e-01 -7.79687822e-01 5.62434718e-02
1.35846958e-01 -6.37050748e-01 5.38751483e-02 1.04076064e+00
6.24417424e-01 4.85828251e-01 6.22611225e-01 -1.28775942e+00
-4.55644816e-01 8.18885088e-01 -8.88782978e-01 2.87772298e-01
9.74948555e-02 5.65083206e-01 1.02728808e+00 -3.86063568e-02
2.76053160e-01 1.32483065e+00 -8.16650540e-02 6.40795231e-01
-1.57022631e+00 -7.68266737e-01 5.71986794e-01 -3.19380909e-02
-6.30303144e-01 -3.49288315e-01 6.42803013e-01 -3.17952484e-01
9.26816285e-01 -6.20369129e-02 6.38848186e-01 8.75730038e-01
-4.59249355e-02 9.03840840e-01 1.49310076e+00 -6.91124320e-01
5.59211552e-01 3.94986510e-01 -1.54058710e-01 4.62848365e-01
1.71879336e-01 7.40499616e-01 -3.91793996e-01 -1.43382356e-01
8.28578115e-01 -1.97723895e-01 4.39586639e-02 -5.15066862e-01
-8.96109402e-01 1.02621734e+00 1.64224833e-01 -1.02270685e-01
-6.38680279e-01 5.37871778e-01 2.64624119e-01 6.62327170e-01
9.65795964e-02 8.76188993e-01 -6.33434057e-01 -3.97773325e-01
-7.53715038e-01 6.70133471e-01 4.25699472e-01 5.12597799e-01
9.15969133e-01 4.36528832e-01 -3.09295446e-01 9.62486863e-01
3.53006780e-01 6.08342171e-01 7.66026735e-01 -1.43284547e+00
3.74914974e-01 4.30503190e-01 6.22867048e-01 -3.33198905e-01
-1.65985584e-01 -2.82855868e-01 -5.38615324e-02 7.28525698e-01
6.93296969e-01 -8.22960377e-01 -8.51803482e-01 2.00713134e+00
5.84737659e-01 1.79641657e-02 1.19205765e-01 7.93636620e-01
3.21840733e-01 4.69841838e-01 2.57977009e-01 -4.74168807e-01
9.02611494e-01 -9.41302598e-01 -4.53678548e-01 -3.74836773e-01
6.71652973e-01 -5.88971317e-01 1.10103917e+00 5.54055870e-01
-1.37953043e+00 -3.24354321e-01 -8.23419809e-01 6.23240769e-01
8.77899230e-02 -2.52187192e-01 4.92377967e-01 5.85209966e-01
-9.43182826e-01 8.01536679e-01 -8.19070458e-01 6.70968071e-02
3.12441558e-01 5.70700824e-01 2.73831815e-01 2.25804955e-01
-1.00353158e+00 9.67761874e-01 4.27399457e-01 -3.89549375e-01
-1.28048062e+00 -5.44535637e-01 -7.54956901e-01 1.33894622e-01
9.54452276e-01 -3.89884293e-01 1.91723621e+00 -1.29612148e+00
-2.13913012e+00 4.73322123e-01 2.82389820e-01 -5.73294699e-01
4.53702718e-01 -1.76737294e-01 2.32202172e-01 5.42452885e-03
3.89041118e-02 9.12883937e-01 9.83544230e-01 -1.26907265e+00
-9.77122307e-01 -2.02054694e-01 3.51249486e-01 6.23829663e-01
-1.36432186e-01 2.33838838e-02 2.98623681e-01 -4.46465641e-01
-3.85250181e-01 -1.02337170e+00 -5.69795966e-01 -5.21958768e-01
-7.61098340e-02 -7.17693388e-01 4.05021310e-01 4.88018291e-03
9.82452810e-01 -1.90422714e+00 -1.21340573e-01 2.69161463e-01
3.51902694e-02 2.64781445e-01 -3.34960520e-01 2.08547994e-01
-1.07388705e-01 -2.78577097e-02 -2.35362668e-02 1.52919635e-01
1.20552808e-01 2.40036964e-01 -1.74794242e-01 1.49139717e-01
9.64313447e-02 4.88809794e-01 -1.21737456e+00 -3.08909893e-01
1.53831646e-01 -1.29026577e-01 -8.66687179e-01 5.33141315e-01
-6.42326236e-01 3.03433985e-01 -6.93591356e-01 2.21820414e-01
2.74329394e-01 -2.65521761e-02 4.54622179e-01 6.41031086e-01
-1.42044351e-01 8.11805665e-01 -1.49826217e+00 9.65889037e-01
-3.06461722e-01 2.20183253e-01 1.08874291e-01 -1.34462643e+00
9.01043653e-01 3.80345851e-01 4.75269228e-01 -8.95105541e-01
8.00387859e-02 1.67085439e-01 3.01725775e-01 -4.51581240e-01
1.62934989e-01 -3.05779189e-01 2.18813315e-01 1.05042446e+00
8.13874453e-02 -5.32797873e-01 3.25343132e-01 2.99094133e-02
1.09620464e+00 6.28697038e-01 4.45353419e-01 -1.94855303e-01
1.56503394e-01 2.80535240e-02 7.13887215e-01 1.00341892e+00
-3.73326302e-01 -1.56372085e-01 8.32540393e-01 -3.66163194e-01
-7.34589636e-01 -9.34155107e-01 2.14216709e-01 1.68470347e+00
-1.49623632e-01 1.76103130e-01 -6.68901503e-01 -1.02569425e+00
2.83315629e-01 1.06722033e+00 -5.82238615e-01 -1.74021751e-01
-6.41668200e-01 -7.40732729e-01 1.01152984e-02 3.06294501e-01
3.90841216e-01 -1.43783224e+00 -1.06086648e+00 4.48516786e-01
1.98048219e-01 -4.71874326e-01 -4.95510727e-01 5.58165610e-01
-7.72729695e-01 -9.72961903e-01 -6.22394085e-01 -2.64635056e-01
6.30726635e-01 1.68441236e-01 1.15674567e+00 6.96300492e-02
1.32209986e-01 4.72753465e-01 -1.35354534e-01 -7.65960038e-01
-3.29766333e-01 -2.18837634e-01 3.49628329e-01 -3.65388453e-01
3.11539441e-01 -6.68292344e-01 -7.25169063e-01 3.13249201e-01
-7.87375271e-01 -2.25840002e-01 6.48769438e-01 9.42493439e-01
4.83438969e-01 -3.82475406e-02 9.66984987e-01 -9.47893977e-01
9.67706859e-01 -3.29484463e-01 -1.02053797e+00 1.91717610e-01
-1.23451841e+00 6.49627388e-01 6.57090187e-01 -9.38973367e-01
-1.16530025e+00 -1.96587235e-01 2.14063227e-01 -4.07439232e-01
-2.70088524e-01 2.34403327e-01 1.79960445e-01 2.75728643e-01
7.93787181e-01 1.31340340e-01 5.65250441e-02 -1.59857765e-01
2.26499766e-01 3.37868094e-01 7.84650072e-02 -1.21785903e+00
5.06990433e-01 -1.48752391e-01 -2.51915514e-01 -2.19906241e-01
-9.06362176e-01 1.26026586e-01 2.35901605e-02 -2.18167335e-01
5.40323138e-01 -6.71233177e-01 -1.01632130e+00 3.28913666e-02
-5.84040165e-01 -9.74715471e-01 -8.01425755e-01 6.87566817e-01
-9.80830729e-01 -6.92230314e-02 -1.00018881e-01 -1.19155204e+00
-6.51921555e-02 -1.36583364e+00 4.56282377e-01 7.33813882e-01
-1.37355626e-01 -8.46608579e-01 1.66535959e-01 3.95493478e-01
3.93728614e-01 3.20426598e-02 8.50820720e-01 -8.67344081e-01
-3.36815029e-01 3.76099169e-01 1.38653353e-01 3.08330715e-01
1.53723985e-01 -3.25260460e-02 -6.43526793e-01 -4.82345700e-01
-2.31988486e-02 -1.04709053e+00 4.58618253e-01 5.76952755e-01
9.38036621e-01 -7.91572332e-01 1.11541390e-01 1.43809155e-01
1.13707793e+00 6.64997995e-01 3.00659508e-01 7.58325756e-01
2.39148848e-02 7.02508092e-01 6.79287612e-01 6.67120039e-01
4.35025543e-01 5.52059114e-01 5.14449894e-01 3.20700437e-01
3.31229687e-01 -3.92966628e-01 6.97326183e-01 -8.79425853e-02
-1.53069928e-01 9.82338041e-02 -7.53419697e-01 3.77090722e-01
-1.81542337e+00 -1.05223083e+00 5.58490515e-01 2.32842231e+00
1.20995820e+00 5.10612369e-01 5.91070414e-01 -1.25425503e-01
4.74751741e-01 4.90426132e-03 -1.01916373e+00 -8.67717743e-01
4.32317734e-01 3.64860475e-01 3.59089345e-01 6.08091354e-01
-6.46062195e-01 9.85254645e-01 6.03831053e+00 8.52067769e-01
-1.06496584e+00 -2.40105495e-01 7.31988430e-01 -5.04951000e-01
-3.23129714e-01 1.90395474e-01 -8.47749770e-01 3.32372546e-01
1.16053832e+00 -3.41371000e-01 7.84534872e-01 9.32836354e-01
5.06949127e-01 -5.75652719e-01 -1.23222625e+00 4.03431088e-01
-5.10344803e-01 -9.87471759e-01 -2.74588436e-01 1.54659197e-01
8.49269390e-01 -2.08061472e-01 1.79219753e-01 7.39109993e-01
1.47571707e+00 -9.09351885e-01 6.72148287e-01 3.50202285e-02
4.90575731e-01 -8.35652471e-01 4.72556591e-01 7.37434149e-01
-5.39153337e-01 -3.01078647e-01 -2.21853890e-02 -3.61678272e-01
-3.27343255e-01 1.99847370e-01 -1.13164854e+00 1.26625329e-01
6.77698612e-01 1.51760146e-01 -1.08448997e-01 6.77200258e-01
-7.49299169e-01 8.58008802e-01 -3.00008357e-01 -3.22800249e-01
5.76845169e-01 -3.71333510e-01 3.83602917e-01 8.73518884e-01
-5.25110811e-02 3.44673097e-01 5.75397670e-01 7.71775424e-01
8.38764831e-02 2.50735730e-02 -3.69821250e-01 -8.16301256e-02
5.93207002e-01 9.27664459e-01 -5.56662202e-01 -3.54136080e-01
1.45516340e-02 2.47830331e-01 6.02364242e-01 4.76787359e-01
-5.72388232e-01 -2.53175169e-01 6.09553933e-01 3.08666676e-02
3.48076969e-01 -4.27483469e-02 -1.54147238e-01 -6.58997476e-01
-3.39942962e-01 -1.21120310e+00 5.41625500e-01 -4.62143481e-01
-9.32453275e-01 9.07521024e-02 3.54092896e-01 -9.70617473e-01
-7.49196649e-01 -2.08845899e-01 -6.21548891e-01 8.67692769e-01
-1.78581679e+00 -2.87655532e-01 5.06758749e-01 4.73769188e-01
5.58407366e-01 -8.27175155e-02 4.98382568e-01 -3.47916275e-01
-5.46229959e-01 7.09320784e-01 2.50224680e-01 -1.65026620e-01
7.59350657e-01 -1.48396003e+00 -3.06906313e-01 2.62264043e-01
-2.65837431e-01 3.75025868e-01 9.23945129e-01 -3.50272477e-01
-9.52507734e-01 -5.21031797e-01 4.68190700e-01 -1.60530522e-01
6.48699880e-01 1.02349520e-01 -6.73532844e-01 6.24930799e-01
4.52452660e-01 -3.46969426e-01 4.91154999e-01 1.18380561e-01
-1.75675288e-01 -2.67260402e-01 -1.39973783e+00 7.11736083e-01
4.71554995e-01 6.19654320e-02 -8.21198761e-01 2.25345433e-01
7.63477862e-01 -3.71147126e-01 -6.71532035e-01 2.35240668e-01
3.27565253e-01 -8.73333037e-01 6.97899699e-01 -7.95693576e-01
5.55981874e-01 -5.14936149e-02 9.93069187e-02 -1.79327190e+00
-1.66437432e-01 -8.63749504e-01 -3.62431332e-02 9.60457563e-01
5.79172611e-01 -6.61343098e-01 8.48080754e-01 7.98158348e-01
1.92046553e-01 -1.00515616e+00 -7.92865634e-01 -7.26787686e-01
3.12649101e-01 -7.54486071e-03 4.86987919e-01 8.14252138e-01
-1.03500737e-02 2.14330062e-01 -2.63783574e-01 -4.47142012e-02
6.68679416e-01 9.54586565e-02 7.39285469e-01 -8.04862857e-01
-6.43942237e-01 -6.53409302e-01 3.47066760e-01 -1.19862854e+00
8.78041089e-02 -4.61306721e-01 2.27324873e-01 -1.26810610e+00
4.11999077e-02 -6.94862425e-01 -5.81374705e-01 7.51699686e-01
-4.44012940e-01 -6.30850434e-01 4.07041013e-01 -4.07214696e-03
-7.06922591e-01 6.35649323e-01 1.41432679e+00 1.07309110e-01
-6.85329258e-01 2.74374038e-01 -8.49647045e-01 7.42590904e-01
8.76692772e-01 -7.74475038e-01 -5.22568524e-01 -3.16539884e-01
2.13042110e-01 5.44353545e-01 1.38221141e-02 -4.96539950e-01
1.74814239e-01 -1.10288739e+00 2.10141852e-01 -1.12439111e-01
2.78848931e-02 -5.83083332e-01 -4.57955509e-01 5.87286949e-01
-9.00924742e-01 2.07608223e-01 1.06246851e-01 4.93603557e-01
1.01402186e-01 -5.55493891e-01 1.04392457e+00 -4.15000528e-01
-1.45373821e-01 -2.54954193e-02 -6.33119166e-01 3.40440243e-01
1.09487510e+00 -1.97191071e-02 -2.44489878e-01 -5.38103938e-01
-5.51469088e-01 7.93108642e-01 6.82789907e-02 1.77540317e-01
4.50805336e-01 -9.83551800e-01 -7.16182113e-01 -1.19196564e-01
-2.57339597e-01 2.02620536e-01 -8.61157626e-02 5.39865673e-01
7.50512257e-02 3.97807598e-01 -1.18394502e-01 -2.84267962e-01
-8.82959008e-01 5.15092671e-01 6.27488315e-01 -8.12357485e-01
-1.07720926e-01 7.77510464e-01 1.03168078e-01 -6.28062725e-01
5.75613141e-01 -3.30722481e-02 -2.54158080e-01 4.18560281e-02
4.40059751e-01 4.10607815e-01 -2.09962830e-01 2.41395518e-01
5.24746068e-02 7.26055354e-02 -4.00599152e-01 -6.00364506e-01
1.47177589e+00 1.03085048e-01 3.22864413e-01 3.20301712e-01
4.27193254e-01 -2.47909904e-01 -1.89010119e+00 -4.89450842e-01
1.73427075e-01 -5.74864209e-01 1.26323923e-01 -1.18632948e+00
-9.22511220e-01 5.86299479e-01 4.95155156e-01 3.72459769e-01
1.06018245e+00 -3.05591404e-01 2.41889909e-01 6.56025589e-01
3.05466890e-01 -1.57670689e+00 5.39187431e-01 3.51898640e-01
4.97086585e-01 -1.33389926e+00 1.57690212e-01 3.85584861e-01
-9.94030833e-01 8.78869951e-01 1.08360207e+00 -3.28265369e-01
3.48173052e-01 1.58681035e-01 1.47299364e-01 -1.79557619e-03
-1.20066953e+00 -1.81504503e-01 -1.22127838e-01 4.82425362e-01
3.46967459e-01 1.18615955e-01 -3.76716584e-01 1.22005038e-01
-2.53824055e-01 1.32917007e-02 5.18378615e-01 1.22513008e+00
-8.10324669e-01 -1.35330439e+00 -5.76847434e-01 6.37722075e-01
-5.45751572e-01 1.11185208e-01 -2.13274822e-01 7.07579315e-01
1.56212673e-01 1.11911952e+00 -4.66715954e-02 8.78492743e-02
3.45839947e-01 4.30645272e-02 4.73011166e-01 -6.46360099e-01
-8.85720670e-01 3.25382859e-01 3.03284428e-03 -4.13907528e-01
-5.52860796e-01 -8.09538305e-01 -1.37421513e+00 -1.29082680e-01
-2.66140729e-01 4.36174363e-01 5.01516879e-01 9.98260319e-01
-4.42590155e-02 4.55287844e-01 1.05031192e+00 -6.99252546e-01
-1.40088129e+00 -7.28463054e-01 -4.74152744e-01 1.35125905e-01
3.99008006e-01 -8.15258622e-01 -4.74378735e-01 -3.90896946e-01] | [4.088977813720703, 1.9883264303207397] |
816207e4-922e-4014-a2e9-28e6d7cff65a | bindsnet-a-machine-learning-oriented-spiking | 1806.01423 | null | http://arxiv.org/abs/1806.01423v2 | http://arxiv.org/pdf/1806.01423v2.pdf | BindsNET: A machine learning-oriented spiking neural networks library in Python | The development of spiking neural network simulation software is a critical
component enabling the modeling of neural systems and the development of
biologically inspired algorithms. Existing software frameworks support a wide
range of neural functionality, software abstraction levels, and hardware
devices, yet are typically not suitable for rapid prototyping or application to
problems in the domain of machine learning. In this paper, we describe a new
Python package for the simulation of spiking neural networks, specifically
geared towards machine learning and reinforcement learning. Our software,
called BindsNET, enables rapid building and simulation of spiking networks and
features user-friendly, concise syntax. BindsNET is built on top of the PyTorch
deep neural networks library, enabling fast CPU and GPU computation for large
spiking networks. The BindsNET framework can be adjusted to meet the needs of
other existing computing and hardware environments, e.g., TensorFlow. We also
provide an interface into the OpenAI gym library, allowing for training and
evaluation of spiking networks on reinforcement learning problems. We argue
that this package facilitates the use of spiking networks for large-scale
machine learning experimentation, and show some simple examples of how we
envision BindsNET can be used in practice. BindsNET code is available at
https://github.com/Hananel-Hazan/bindsnet | ['Robert Kozma', 'Hava T. Siegelmann', 'Hananel Hazan', 'Daniel J. Saunders', 'Hassaan Khan', 'Darpan T. Sanghavi'] | 2018-06-04 | null | null | null | null | ['neural-network-simulation'] | ['computer-code'] | [-3.10896814e-01 -3.58995408e-01 4.65347946e-01 -2.31885254e-01
1.36874497e-01 -6.40732050e-01 3.61239940e-01 -1.29712075e-01
-6.67331755e-01 6.17064953e-01 -2.53867716e-01 -3.81546825e-01
2.72804260e-01 -8.10793281e-01 -6.18972898e-01 -8.72231185e-01
-2.57394016e-01 1.33457735e-01 6.16984427e-01 -4.48518574e-01
1.66922852e-01 8.17184687e-01 -2.03802872e+00 2.11788282e-01
4.30454127e-02 7.68480599e-01 3.65144372e-01 8.25397789e-01
5.00044301e-02 5.67794561e-01 -5.79595029e-01 1.92376301e-01
1.68723196e-01 -2.52023160e-01 -4.28621173e-01 -5.47089219e-01
-5.59735037e-02 -1.10804342e-01 -3.32319558e-01 6.93072438e-01
8.80607426e-01 -2.20992014e-01 2.18408369e-02 -1.48304117e+00
9.36415195e-02 6.06564760e-01 2.59812713e-01 5.78104079e-01
-1.91370770e-01 6.46809876e-01 4.10431802e-01 -5.45509160e-01
4.99576598e-01 1.04311419e+00 7.30944276e-01 8.81240845e-01
-1.40031838e+00 -1.02446425e+00 -1.45555466e-01 -2.19872281e-01
-1.17100251e+00 -7.66617000e-01 3.08072597e-01 -3.93681139e-01
1.36014366e+00 1.32946821e-03 1.10042715e+00 1.17840755e+00
2.86833405e-01 5.97070277e-01 1.10697007e+00 4.97907065e-02
7.90149689e-01 -2.89307892e-01 3.22492957e-01 3.82740021e-01
1.20718792e-01 3.52345496e-01 -6.20824039e-01 -2.46680290e-01
1.46407533e+00 -2.52568692e-01 1.71051808e-02 -1.19342938e-01
-1.30723143e+00 4.53425288e-01 6.64857864e-01 4.78880733e-01
-1.77819803e-01 8.28689039e-01 7.39107072e-01 2.51824558e-01
-2.34039992e-01 4.48750287e-01 -3.87304753e-01 -3.43442380e-01
-5.33742368e-01 7.85429239e-01 1.10560524e+00 8.63297939e-01
6.25551701e-01 4.63983059e-01 2.90339261e-01 7.39867926e-01
3.27462196e-01 4.24814746e-02 5.17123580e-01 -1.44072235e+00
-4.83085841e-01 2.73315281e-01 -1.66426271e-01 -2.02966943e-01
-7.21292317e-01 -3.39516759e-01 -5.51904023e-01 8.43339682e-01
5.10066926e-01 -3.01562399e-01 -8.53430688e-01 1.62012148e+00
8.29858407e-02 3.66905451e-01 1.36008531e-01 9.63989735e-01
1.00804830e+00 4.00424451e-01 2.96169013e-01 2.77266651e-01
1.16947818e+00 -5.23769379e-01 -1.89762399e-01 -3.04062515e-01
6.62299573e-01 -4.51535881e-01 7.12552667e-01 2.22150683e-01
-1.24382365e+00 -1.32154718e-01 -1.02413464e+00 9.17717665e-02
-6.57187045e-01 -4.33120787e-01 1.08222783e+00 4.72889692e-01
-1.50920510e+00 7.39070833e-01 -1.51666522e+00 -8.16828609e-01
6.10994518e-01 7.22605884e-01 -1.22472130e-01 2.96050578e-01
-8.17484915e-01 9.25742209e-01 4.68774676e-01 -1.81533486e-01
-9.98679578e-01 -5.27304947e-01 -5.53262651e-01 1.02877067e-02
-4.35937166e-01 -7.74547279e-01 1.55661929e+00 -8.75162303e-01
-1.52329111e+00 8.92161131e-01 1.05891265e-01 -6.62186980e-01
3.15346383e-02 4.78563398e-01 -6.29917681e-02 -2.74247140e-01
-2.54285842e-01 1.14289582e+00 1.03955068e-01 -8.95968139e-01
-1.54144138e-01 -2.90949434e-01 -2.90378765e-03 6.28265589e-02
1.17658377e-01 3.45793754e-01 -2.80833125e-01 -4.93466407e-01
-1.50675103e-01 -7.83414543e-01 -4.89285201e-01 5.52880943e-01
2.36815080e-01 4.32248451e-02 7.20418870e-01 8.02570134e-02
4.30265903e-01 -2.26402473e+00 -2.58508444e-01 1.09550491e-01
9.81056467e-02 3.77462268e-01 -1.89126432e-01 4.06642944e-01
4.21647392e-02 -1.80772468e-01 -3.05993766e-01 7.03063384e-02
-1.40745610e-01 4.10801619e-01 -9.52244699e-02 3.01855862e-01
1.57034710e-01 7.88103938e-01 -7.25819409e-01 -2.46169101e-02
2.08198547e-01 6.18483961e-01 -6.07626379e-01 3.08458190e-02
-2.00049803e-01 6.18652761e-01 -1.71873018e-01 6.54360652e-01
4.77716982e-01 -2.27860495e-01 2.19454523e-02 1.73356608e-01
-6.41441047e-01 5.72736323e-01 -1.11439955e+00 1.81458461e+00
-2.26473674e-01 6.64704800e-01 3.85836691e-01 -7.81680703e-01
1.10542262e+00 2.21793801e-01 2.21429884e-01 -5.47262192e-01
5.11349201e-01 4.64754134e-01 4.50112283e-01 -2.85669919e-02
-2.28234515e-01 8.29100795e-03 2.42647961e-01 6.03651226e-01
4.73274738e-01 -4.10263121e-01 3.80409509e-01 -6.41783932e-03
1.37905037e+00 3.61186415e-01 7.09726885e-02 -6.03647470e-01
-1.14294074e-01 2.10577562e-01 4.24161047e-01 6.41890347e-01
-4.44076359e-01 1.42011508e-01 3.46109957e-01 -4.98803496e-01
-1.38034344e+00 -1.24233532e+00 -5.49496472e-01 1.14219260e+00
-5.66558763e-02 -2.96999514e-01 -9.17619467e-01 5.74429989e-01
5.77997267e-02 3.11236829e-01 -3.50003898e-01 5.02068289e-02
-3.54247272e-01 -8.97819996e-01 1.10256481e+00 5.65205038e-01
4.20406908e-01 -1.74954867e+00 -1.17522860e+00 5.04123092e-01
5.76350987e-01 -9.03864443e-01 2.86466271e-01 8.19372237e-01
-1.00929582e+00 -7.06651807e-01 -3.57019126e-01 -1.02841485e+00
2.53586650e-01 -4.71756794e-02 7.94080019e-01 4.07312959e-01
-5.96875131e-01 5.40398359e-02 7.78213441e-02 -6.44372046e-01
-2.47563630e-01 -2.33990271e-02 7.22869709e-02 -8.13239634e-01
4.61447179e-01 -1.16156828e+00 -5.11677325e-01 2.72366703e-01
-9.69494522e-01 3.60966533e-01 4.22023721e-02 6.65463448e-01
4.88612026e-01 -6.00016952e-01 6.59630001e-01 -4.21867937e-01
6.32971942e-01 -4.32314187e-01 -9.53731775e-01 -3.31084371e-01
-1.86047256e-01 5.30478247e-02 5.83059967e-01 -3.19920868e-01
-2.94981301e-01 2.44400173e-01 -6.55170500e-01 -1.59326613e-01
-5.31123102e-01 3.38746727e-01 3.05125237e-01 -5.30573905e-01
8.74821365e-01 3.20291877e-01 3.74308705e-01 -2.04253852e-01
-2.41651163e-01 4.86141235e-01 5.19521356e-01 -5.82208872e-01
9.10719037e-02 4.96114105e-01 -1.60785943e-01 -6.77308798e-01
1.61214888e-01 -1.59244031e-01 -3.27196449e-01 -3.55950534e-01
4.13975000e-01 -9.15001392e-01 -1.07406664e+00 1.02656043e+00
-1.07768464e+00 -1.03077233e+00 -1.89584512e-02 2.77245671e-01
-8.98307979e-01 -3.72389793e-01 -9.68354583e-01 -4.44682389e-01
-4.74517643e-01 -1.25487542e+00 6.90503657e-01 4.92895097e-01
-4.58440363e-01 -9.32741463e-01 3.57163608e-01 -3.95727456e-01
8.06850016e-01 2.43890360e-01 5.11842251e-01 -4.86010432e-01
-5.78587413e-01 1.17294043e-01 -1.02980405e-01 -1.28026232e-01
-3.84748369e-01 4.66434181e-01 -1.34952676e+00 -2.51061022e-01
-3.57642621e-01 -6.58243239e-01 6.43123090e-01 5.03092051e-01
1.14703226e+00 9.22532901e-02 -4.05787796e-01 8.06069911e-01
1.48261869e+00 1.19189337e-01 7.64099181e-01 6.09075546e-01
2.18665689e-01 4.81016815e-01 -1.20656997e-01 6.41151965e-01
2.77384520e-01 4.48613971e-01 7.65691042e-01 -1.28704142e-02
1.60082638e-01 2.86340803e-01 2.44520172e-01 7.13434100e-01
8.22586007e-03 3.16879988e-01 -1.21965909e+00 4.58569944e-01
-1.89710987e+00 -9.65798199e-01 -1.25443473e-01 1.96927607e+00
8.26909363e-01 1.38553411e-01 4.24734116e-01 4.26280499e-02
6.28183901e-01 -3.91278952e-01 -8.27860892e-01 -7.53898323e-01
-1.60508797e-01 6.30458355e-01 4.32019025e-01 2.57605821e-01
-6.99982285e-01 1.08744323e+00 7.07652760e+00 3.59034836e-01
-1.57703841e+00 5.81280999e-02 2.19582006e-01 -2.86284864e-01
7.28111863e-02 5.86330816e-02 -8.58784854e-01 4.74432409e-01
1.35093772e+00 -2.53919274e-01 9.43246186e-01 7.23166525e-01
3.36374730e-01 -9.58683789e-02 -9.52612519e-01 9.03200269e-01
-7.24409997e-01 -1.87003064e+00 -4.55391258e-01 -1.01203382e-01
1.95148364e-01 9.79293883e-01 -3.54502827e-01 2.12475240e-01
6.78576589e-01 -1.02706981e+00 7.45167553e-01 2.28398979e-01
3.95206124e-01 -6.59794748e-01 3.77316624e-01 4.57758933e-01
-8.77705812e-01 1.11072622e-01 -4.10626084e-01 -6.06349647e-01
-1.83652818e-01 2.78128684e-01 -6.13054216e-01 -4.04508621e-01
1.17807090e+00 8.21504712e-01 -3.98742676e-01 1.56040263e+00
2.41049215e-01 4.31512117e-01 -7.20413208e-01 -5.19751072e-01
1.35115966e-01 1.58744052e-01 2.30312094e-01 1.38251889e+00
1.33705750e-01 1.46527931e-01 -2.26709202e-01 1.20258582e+00
8.32295492e-02 -2.73447514e-01 -6.99017406e-01 -9.28811207e-02
8.15746307e-01 1.42491472e+00 -1.12282610e+00 -1.65030405e-01
-2.57060260e-01 4.33407426e-01 4.31248516e-01 2.67122269e-01
-6.85596168e-01 -5.30097544e-01 1.21564317e+00 -8.19470175e-03
1.44946471e-01 -4.34699357e-01 -7.63547599e-01 -8.44474554e-01
-4.57107067e-01 -9.07497227e-01 -1.35781989e-01 -1.12195039e+00
-9.47667658e-01 6.87848449e-01 -3.64814728e-01 -7.99786210e-01
-1.62706599e-01 -1.01072776e+00 -9.80105996e-01 8.30982745e-01
-9.47840214e-01 -6.21617973e-01 -4.92794961e-01 8.37642968e-01
1.51511058e-01 -4.99568395e-02 1.22319198e+00 1.84671298e-01
-5.68746746e-01 8.13784897e-02 2.25909613e-02 1.76066399e-01
3.42487842e-01 -8.77111375e-01 9.58521426e-01 5.17687559e-01
-1.69114232e-01 7.67049611e-01 9.15932059e-01 -2.17768505e-01
-1.38565588e+00 -9.08691108e-01 3.14950258e-01 -3.61256674e-02
9.65014398e-01 -6.67656541e-01 -9.46322918e-01 7.75747001e-01
2.09573045e-01 2.53057420e-01 6.40825272e-01 -2.66991079e-01
-3.65699440e-01 1.66853610e-02 -1.04869616e+00 9.98362362e-01
9.63186145e-01 -4.50936019e-01 -7.60774985e-02 1.49889529e-01
2.55038738e-01 -3.35631490e-01 -8.92175376e-01 2.36522108e-01
7.47088909e-01 -1.11809039e+00 5.60278833e-01 -1.92118093e-01
8.24147314e-02 -5.64428627e-01 -3.92684639e-02 -1.20615149e+00
-2.35807523e-01 -5.44044673e-01 2.00913772e-01 8.04709733e-01
3.83094013e-01 -8.96767557e-01 6.69039309e-01 4.28075790e-01
-3.25970531e-01 -6.98730171e-01 -9.57684934e-01 -6.14258170e-01
2.55493015e-01 -5.79521000e-01 5.20126581e-01 6.16080105e-01
4.95735049e-01 -2.09118146e-02 5.41309655e-01 8.81216899e-02
4.65606391e-01 -1.66836157e-01 7.30051219e-01 -9.99055743e-01
-4.07568902e-01 -8.99539709e-01 -9.77179646e-01 -5.58782995e-01
-6.65019676e-02 -9.41006422e-01 2.13108659e-01 -1.47373283e+00
-1.24669038e-01 -5.53836882e-01 -1.74134687e-01 9.68974411e-01
5.73727667e-01 4.26427901e-01 1.41420886e-01 2.56612450e-01
-1.41365051e-01 1.58451557e-01 9.79558527e-01 3.98495227e-01
6.27432484e-03 -2.23111913e-01 -4.20920521e-01 7.07953036e-01
1.23404026e+00 -6.32638633e-01 3.79720591e-02 -4.56302643e-01
-1.78187504e-01 -2.92126983e-01 9.26990688e-01 -1.67548966e+00
5.50586343e-01 2.77178306e-02 3.66051137e-01 -1.27718812e-02
4.01051849e-01 -5.39956391e-01 3.26859176e-01 8.96871150e-01
-3.86460960e-01 1.56153947e-01 7.77758837e-01 -1.94210708e-01
9.17559117e-02 -9.73940641e-03 1.06816518e+00 -3.35139632e-01
-7.35483587e-01 2.11810887e-01 -1.08038557e+00 -1.80319145e-01
9.76702809e-01 -3.79585236e-01 -7.80805409e-01 1.68057550e-02
-9.73783791e-01 1.61794633e-01 8.31467330e-01 2.16711000e-01
5.05225480e-01 -1.04977441e+00 -3.17373246e-01 4.58811492e-01
1.73337236e-01 -2.35323653e-01 -2.49028087e-01 7.18194604e-01
-1.01401305e+00 1.45269886e-01 -1.07789099e+00 -6.78784251e-01
-1.07422841e+00 1.43178597e-01 8.36352587e-01 3.98936957e-01
-5.08490384e-01 8.24310660e-01 1.10797491e-02 -6.34620845e-01
2.79616058e-01 -4.75561231e-01 1.07704565e-01 -6.19325280e-01
6.55772746e-01 -1.00503089e-02 2.17022225e-01 -1.87287375e-01
-4.99007285e-01 4.97621447e-02 1.20089777e-01 -3.07071120e-01
1.57620597e+00 4.09201562e-01 -3.87588590e-01 7.54164100e-01
6.83757126e-01 -8.49943042e-01 -1.49877715e+00 2.95663416e-01
8.82846490e-03 3.68008554e-01 -6.22864068e-02 -7.23824322e-01
-9.20416772e-01 1.05560255e+00 7.60421038e-01 -5.03611676e-02
9.74677384e-01 -9.27153677e-02 3.79545450e-01 5.38923979e-01
8.92289877e-01 -6.81492388e-01 -2.82400548e-01 8.74258578e-01
7.34097600e-01 -6.29781246e-01 -4.72759187e-01 -3.25889885e-02
-3.64277661e-01 1.32623887e+00 9.16007102e-01 -7.86108911e-01
8.19453061e-01 1.32568848e+00 2.98003614e-01 -1.47679329e-01
-1.27510035e+00 -2.42402539e-01 -6.87238932e-01 7.84451127e-01
7.38289535e-01 2.03183386e-02 8.11202079e-02 3.53346258e-01
-3.73665094e-01 5.93413293e-01 6.50521934e-01 1.31023717e+00
-7.75650859e-01 -9.24209654e-01 -3.14327091e-01 4.26620424e-01
-4.01322454e-01 -3.43649209e-01 -2.55178601e-01 4.69250470e-01
1.04845069e-01 4.18556094e-01 4.16954845e-01 -2.42517442e-01
7.51162618e-02 -5.62436692e-02 7.63877749e-01 -6.48661971e-01
-1.16481137e+00 6.08768649e-02 -2.20611785e-02 -5.94059289e-01
-2.13524476e-01 -5.17690778e-01 -1.83676147e+00 -7.90933669e-01
9.54233557e-02 -1.35439798e-01 1.10931993e+00 7.63270676e-01
5.90428650e-01 4.58791971e-01 -9.43226218e-02 -1.31559241e+00
4.16612811e-02 -6.12510204e-01 -7.58786321e-01 -2.48559758e-01
1.06309883e-01 -5.01826584e-01 -1.73573405e-01 4.34198454e-02] | [8.151185035705566, 2.6167449951171875] |
bb68862e-bc0c-48aa-acc8-416f9b82268a | behaviour-discriminator-a-simple-data | 2301.11734 | null | https://arxiv.org/abs/2301.11734v1 | https://arxiv.org/pdf/2301.11734v1.pdf | Behaviour Discriminator: A Simple Data Filtering Method to Improve Offline Policy Learning | This paper studies the problem of learning a control policy without the need for interactions with the environment; instead, learning purely from an existing dataset. Prior work has demonstrated that offline learning algorithms (e.g., behavioural cloning and offline reinforcement learning) are more likely to discover a satisfactory policy when trained using high-quality expert data. However, many real-world/practical datasets can contain significant proportions of examples generated using low-skilled agents. Therefore, we propose a behaviour discriminator (BD) concept, a novel and simple data filtering approach based on semi-supervised learning, which can accurately discern expert data from a mixed-quality dataset. Our BD approach was used to pre-process the mixed-skill-level datasets from the Real Robot Challenge (RRC) III, an open competition requiring participants to solve several dexterous robotic manipulation tasks using offline learning methods; the new BD method allowed a standard behavioural cloning algorithm to outperform other more sophisticated offline learning algorithms. Moreover, we demonstrate that the new BD pre-processing method can be applied to a number of D4RL benchmark problems, improving the performance of multiple state-of-the-art offline reinforcement learning algorithms. | ['Stephen J. Redmond', 'Francisco Roldan Sanchez', "Noel E. O'Connor", 'Kevin McGuinness', 'David Cordova Bulens', 'Robert McCarthy', 'Qiang Wang'] | 2023-01-27 | null | null | null | null | ['d4rl'] | ['robots'] | [ 2.37283587e-01 -3.04098092e-02 -2.15305164e-01 -2.20164269e-01
-7.46786892e-01 -8.95838857e-01 6.55941904e-01 2.62799174e-01
-8.96052778e-01 9.32081640e-01 -2.93999583e-01 -3.61063004e-01
-5.97719252e-01 -3.54751140e-01 -9.72088516e-01 -6.10702872e-01
-4.36830223e-01 8.71377051e-01 3.55579019e-01 -4.22136307e-01
4.40826356e-01 5.66586733e-01 -2.03769112e+00 2.43620649e-01
1.04494369e+00 1.07202864e+00 5.76415479e-01 5.63502729e-01
3.08895409e-01 9.37453568e-01 -7.31825054e-01 1.19461656e-01
7.89914429e-01 -4.49643433e-01 -9.18838501e-01 -1.32157117e-01
2.83249795e-01 -3.58384907e-01 3.52270529e-02 9.06582832e-01
4.88524258e-01 4.47032392e-01 4.67060328e-01 -1.32145095e+00
-1.92728370e-01 7.12819695e-01 -1.03610292e-01 -1.96385924e-02
4.80456531e-01 6.69535220e-01 8.42101038e-01 -2.69676119e-01
8.72972429e-01 1.30313480e+00 4.81578141e-01 5.86569369e-01
-1.41830611e+00 -5.44921517e-01 3.12988222e-01 3.57869238e-01
-7.75756061e-01 -1.88571468e-01 5.65627337e-01 -5.28896868e-01
9.68694389e-01 -1.34379044e-01 7.19659030e-01 1.53979051e+00
1.82081804e-01 1.13877928e+00 1.71085513e+00 -2.38685027e-01
6.47828698e-01 -1.72963813e-02 -2.01379016e-01 7.00404286e-01
8.11575949e-02 1.02607059e+00 -4.10881460e-01 -1.55663803e-01
6.55518711e-01 -2.58023858e-01 -1.48890510e-01 -9.80878353e-01
-1.31589603e+00 6.25808477e-01 2.16088757e-01 1.67791411e-01
-5.29348612e-01 -2.63786856e-02 5.09864330e-01 9.11925733e-01
-8.55121836e-02 1.24910736e+00 -9.48625147e-01 -7.45286047e-01
-3.82286370e-01 6.65095329e-01 1.12403619e+00 8.05931628e-01
6.29216850e-01 -1.18214719e-01 -9.73713547e-02 6.17496848e-01
-5.40596172e-02 1.99171424e-01 6.48210466e-01 -1.35152352e+00
3.22148949e-01 6.55746520e-01 4.97036636e-01 -4.97705966e-01
-5.21985650e-01 -2.25929171e-01 -2.40789741e-01 8.25441003e-01
8.43023360e-01 -2.37102717e-01 -9.24304843e-01 1.66463935e+00
3.96672994e-01 -1.38253802e-02 2.24955857e-01 1.05685556e+00
2.50223994e-01 2.41009146e-01 -1.01227127e-02 -1.47745967e-01
7.37879813e-01 -1.05151939e+00 -4.05544102e-01 -2.72149265e-01
7.29006231e-01 -3.63104790e-01 1.21885002e+00 1.14783621e+00
-8.77082527e-01 -7.38606811e-01 -1.13254011e+00 4.12084192e-01
-4.24549013e-01 1.12736955e-01 7.57205606e-01 3.44747543e-01
-6.78752899e-01 1.19485080e+00 -9.46614325e-01 -1.36103958e-01
2.30378866e-01 4.31725711e-01 -4.66528386e-01 -1.54053628e-01
-1.00256741e+00 1.20380068e+00 8.05386186e-01 -2.70738244e-01
-1.66160440e+00 -6.89721406e-01 -7.13131189e-01 -3.73060912e-01
1.00372374e+00 -1.41007438e-01 1.65210068e+00 -1.34606838e+00
-2.10524583e+00 5.20475864e-01 6.25596821e-01 -7.14985132e-01
7.43940175e-01 -4.49385375e-01 -1.35322377e-01 9.62307230e-02
-8.94692317e-02 6.48638248e-01 1.12973738e+00 -1.37455523e+00
-9.33828533e-01 -2.00944424e-01 3.32820445e-01 1.48049265e-01
2.03338027e-01 -4.30929244e-01 1.59297571e-01 -4.68553692e-01
-3.72584134e-01 -1.15697026e+00 -3.51457149e-01 -2.39784181e-01
1.76132321e-01 -5.74594259e-01 5.21850228e-01 -4.30529684e-01
7.24009454e-01 -2.18548393e+00 5.91712773e-01 2.55021393e-01
5.90370111e-02 4.31781024e-01 -4.76603568e-01 4.11818832e-01
-9.11663249e-02 -2.65260845e-01 -1.85342744e-01 1.04823425e-01
2.99278229e-01 5.21311998e-01 -1.98355153e-01 4.60640043e-01
1.88095182e-01 5.81395984e-01 -1.42637849e+00 -4.69561741e-02
1.90172315e-01 -3.18068475e-01 -5.85789680e-01 4.86201614e-01
-6.59322083e-01 6.84530258e-01 -3.44018191e-01 5.18083215e-01
1.23269916e-01 2.34055281e-01 2.74232745e-01 2.52935529e-01
-1.56044543e-01 2.06579372e-01 -1.21121430e+00 2.14693379e+00
-4.85169798e-01 2.82883883e-01 1.25242308e-01 -1.22828257e+00
9.62845683e-01 1.08720548e-01 6.25416696e-01 -8.74777615e-01
2.41943359e-01 4.51191515e-01 4.45021808e-01 -5.90802193e-01
3.32670450e-01 1.84455305e-01 -2.49836177e-01 4.16837722e-01
4.28197026e-01 -3.32686722e-01 7.63178051e-01 -1.21147379e-01
1.55473363e+00 6.39800489e-01 2.13793784e-01 -2.78679878e-01
2.97114104e-01 5.44815600e-01 6.88378930e-01 1.21078193e+00
-5.33495963e-01 3.69715653e-02 4.46866244e-01 -3.79906267e-01
-7.99901903e-01 -9.10395324e-01 2.81335384e-01 1.29832292e+00
1.23723798e-01 -4.36781645e-01 -5.63832223e-01 -1.05687118e+00
4.75706458e-01 7.76561439e-01 -6.05485082e-01 -5.23319304e-01
-4.27281499e-01 -1.99646056e-01 3.63506615e-01 5.72632909e-01
3.50665092e-01 -1.34862268e+00 -1.12173855e+00 4.92089748e-01
1.75834849e-01 -9.69186127e-01 -2.54731756e-02 7.00601935e-01
-7.87502229e-01 -1.45672226e+00 -2.20276549e-01 -7.17919827e-01
5.14608800e-01 -9.91003290e-02 1.13803577e+00 9.63852834e-03
-3.26468378e-01 7.51644969e-01 -7.84300804e-01 -4.47320640e-01
-7.98000395e-01 4.54905368e-02 5.03895879e-01 -3.99612308e-01
2.96505064e-01 -3.37380290e-01 -2.43261606e-01 5.18849373e-01
-6.50427639e-01 -1.60755247e-01 1.06920898e+00 1.22160399e+00
4.34786767e-01 4.35588509e-01 5.62826931e-01 -6.78558111e-01
9.76251245e-01 -3.15073252e-01 -9.22041237e-01 1.53765783e-01
-8.78096759e-01 4.78817314e-01 8.72272074e-01 -9.36584234e-01
-8.74702215e-01 2.90338904e-01 3.46552938e-01 -4.90233660e-01
-4.55872446e-01 4.93409872e-01 9.34773088e-02 -5.68535402e-02
9.22217011e-01 1.76964939e-01 4.41081315e-01 -5.76993942e-01
3.95474017e-01 3.89871389e-01 6.03615940e-01 -1.13996017e+00
6.57768130e-01 -9.87665206e-02 -3.41050588e-02 -2.47139916e-01
-5.63367188e-01 -5.07654428e-01 -6.63976669e-01 -2.60201335e-01
5.38739204e-01 -6.80506468e-01 -1.05613744e+00 6.87707841e-01
-6.33256376e-01 -1.11054111e+00 -4.05119270e-01 6.47740960e-01
-1.17777979e+00 4.91720140e-02 -3.88526112e-01 -7.12429404e-01
2.74053395e-01 -1.42172098e+00 8.12907517e-01 1.28862366e-01
-2.39717171e-01 -6.90541565e-01 2.56836623e-01 1.10867254e-01
2.16399297e-01 2.62672335e-01 8.05768728e-01 -9.75837171e-01
-1.81208134e-01 -4.04186770e-02 2.52052754e-01 6.10053897e-01
8.57875422e-02 -2.22337589e-01 -5.05832851e-01 -7.68687546e-01
-1.61513388e-02 -1.16107225e+00 5.68784535e-01 -1.78433642e-01
1.13983536e+00 -1.52206838e-01 -4.75096181e-02 1.21601589e-01
1.19211006e+00 4.49243516e-01 3.10559094e-01 7.59974718e-01
3.31644446e-01 6.20057166e-01 1.18644929e+00 5.07246792e-01
2.37306934e-02 5.33610463e-01 4.69412953e-01 3.64910066e-01
1.54165387e-01 -3.74798477e-01 6.80444121e-01 4.39691275e-01
-1.18725657e-01 1.76487789e-01 -9.71223831e-01 5.12004018e-01
-2.13930583e+00 -7.91206837e-01 4.95889783e-01 2.08237672e+00
1.20150387e+00 3.61295104e-01 4.23418969e-01 2.08038315e-01
2.98928916e-01 -3.79036516e-01 -9.39477980e-01 -3.68456036e-01
2.04032585e-01 4.44890469e-01 5.53265929e-01 3.68099362e-02
-1.17206585e+00 7.20846415e-01 5.95197630e+00 9.68681097e-01
-8.65605235e-01 -2.62939215e-01 1.97541099e-02 -3.80183198e-02
2.72490740e-01 -8.48915055e-02 -5.79730749e-01 3.71575147e-01
1.00638270e+00 9.65741053e-02 1.06400728e+00 1.25243878e+00
4.68110889e-02 -5.34888148e-01 -1.58272529e+00 9.15944457e-01
-1.89971343e-01 -8.55282903e-01 -3.81874830e-01 -1.30939230e-01
7.33922124e-01 2.44820211e-02 -1.19791850e-01 1.03140926e+00
9.73880768e-01 -9.42498267e-01 9.34445798e-01 4.20639426e-01
3.18089366e-01 -7.56782234e-01 5.78429639e-01 8.37167382e-01
-7.37218201e-01 -7.34519243e-01 -1.34614006e-01 -1.73828036e-01
-3.33662271e-01 3.56087610e-02 -8.54134738e-01 7.54766524e-01
8.46029460e-01 8.37567031e-01 -7.64258921e-01 1.09631658e+00
-3.12346071e-01 5.04478276e-01 -2.86269248e-01 -2.77946711e-01
5.12441754e-01 -1.80003464e-01 5.27044892e-01 5.37279904e-01
-2.97951996e-02 -9.08752531e-02 5.96381783e-01 6.47671103e-01
3.05898190e-01 -2.48255447e-01 -5.16300559e-01 -5.07609606e-01
3.15456361e-01 1.02312601e+00 -4.98632044e-01 -7.11420923e-02
-1.38070345e-01 7.19401062e-01 6.02282882e-01 5.77239059e-02
-6.78097606e-01 -4.89177853e-01 5.17906845e-01 -2.61167109e-01
4.18956608e-01 -4.92384940e-01 2.55580783e-01 -7.29466736e-01
-1.13254882e-01 -1.84648967e+00 2.79962003e-01 -5.99783182e-01
-1.51602340e+00 2.16998771e-01 1.83622003e-01 -1.53478694e+00
-4.46710765e-01 -1.03450859e+00 -1.27449349e-01 3.33854854e-01
-1.31172895e+00 -4.06799853e-01 -3.03565353e-01 5.52123308e-01
5.73251367e-01 -5.61332285e-01 7.53548980e-01 3.80141549e-02
-3.63943666e-01 2.13022545e-01 2.50715196e-01 -7.72500858e-02
8.67872179e-01 -1.56834137e+00 -4.71843965e-02 3.86970133e-01
3.42703797e-02 6.19376063e-01 6.61636531e-01 -6.35642409e-01
-1.88291192e+00 -8.60352457e-01 -9.84777287e-02 -4.85124707e-01
7.74497867e-01 -3.33595604e-01 -7.26268768e-01 4.33859259e-01
4.74601649e-02 -1.65223509e-01 3.12037081e-01 5.27819581e-02
-8.62313583e-02 -1.02825485e-01 -1.07022047e+00 5.26356518e-01
1.20965552e+00 -2.62184858e-01 -1.12235653e+00 1.38901472e-01
6.50883675e-01 -7.24520504e-01 -9.50876355e-01 6.59396589e-01
5.10844290e-01 -6.66265905e-01 6.94204569e-01 -9.85600054e-01
3.26560080e-01 -2.72556484e-01 1.53578877e-01 -1.82825506e+00
-1.70820296e-01 -9.32086051e-01 -2.69949764e-01 8.74947190e-01
2.56257266e-01 -5.47050536e-01 5.13696313e-01 2.84017354e-01
-1.57098070e-01 -7.22372234e-01 -7.13888407e-01 -1.36142862e+00
-4.23281938e-02 -3.38981658e-01 5.27270436e-01 7.37262607e-01
1.25761658e-01 4.03887499e-03 -1.62610888e-01 -2.88602948e-01
5.29741883e-01 1.51195243e-01 1.13417435e+00 -1.37150609e+00
-6.14612937e-01 -5.34468174e-01 -2.10256159e-01 -1.18134522e+00
4.62203205e-01 -8.83511186e-01 5.43114066e-01 -1.09795296e+00
-4.00572300e-01 -6.69807553e-01 -3.10904503e-01 6.21941447e-01
8.48593116e-02 -5.64840436e-01 1.03026040e-01 3.42326686e-02
-8.94051969e-01 7.72300005e-01 1.47616053e+00 -1.82517111e-01
-3.85059148e-01 6.15447797e-02 -2.87640899e-01 6.19419277e-01
8.88917506e-01 -6.10725343e-01 -6.48500323e-01 -1.68806016e-01
2.21501157e-01 -6.56505525e-02 3.90164912e-01 -1.12109101e+00
2.42110953e-01 -3.64574134e-01 3.16892207e-01 -2.20336020e-01
-1.65189169e-02 -1.04656565e+00 -1.03112735e-01 7.51114964e-01
-5.01284242e-01 6.41441420e-02 4.23193425e-01 8.30803514e-01
-3.06453742e-02 -3.93099070e-01 5.36160707e-01 -4.12579685e-01
-1.23052824e+00 -5.75949512e-02 -6.60932958e-01 2.93377876e-01
1.19454145e+00 -3.31845172e-02 -2.80366421e-01 6.08682632e-03
-7.28989661e-01 5.33586144e-01 3.75697106e-01 8.02618265e-01
3.70397329e-01 -1.09214079e+00 -3.64663512e-01 2.52165914e-01
1.88077912e-01 -4.60338481e-02 -2.17739284e-01 8.94703150e-01
-3.50168616e-01 1.77031353e-01 -7.05175936e-01 -6.62813842e-01
-7.40036190e-01 6.36541128e-01 2.63222635e-01 -4.45048273e-01
-5.79820096e-01 6.01270854e-01 -6.08678818e-01 -9.56171095e-01
5.72727740e-01 -5.84896326e-01 -6.31795526e-02 -6.41068071e-02
2.04847172e-01 4.72301036e-01 1.72682986e-01 4.60110344e-02
-2.79023379e-01 3.23245488e-02 -5.90910688e-02 -1.77735820e-01
1.38083839e+00 4.62272793e-01 1.63676843e-01 4.85388637e-01
7.11755931e-01 -5.47421753e-01 -1.71961021e+00 -1.74232498e-01
5.04382610e-01 -3.90497625e-01 -1.07948750e-01 -1.20561111e+00
-6.74427748e-01 3.73818666e-01 7.03433633e-01 1.73599571e-01
1.00277495e+00 -3.04094553e-01 3.86816949e-01 1.01896405e+00
1.14843655e+00 -1.74191248e+00 4.85662580e-01 7.65318513e-01
1.02405083e+00 -1.46056116e+00 -1.16445646e-01 2.50561982e-01
-7.38209784e-01 1.19069493e+00 1.00873220e+00 -2.70409733e-01
2.99838394e-01 1.94777802e-01 -9.36035737e-02 4.47794358e-04
-1.02436340e+00 -4.14842010e-01 1.31308213e-01 7.38319159e-01
-3.17357957e-01 -7.60344788e-02 -1.93181291e-01 7.56850243e-01
-2.08999217e-01 1.71499163e-01 3.29894662e-01 1.47698641e+00
-3.79618615e-01 -1.31624174e+00 -1.34543210e-01 4.20870513e-01
-2.90522482e-02 4.66426253e-01 -5.97763062e-01 8.95252705e-01
2.14735851e-01 9.48970795e-01 -1.57276675e-01 -6.03507936e-01
5.94399750e-01 -3.95508148e-02 1.03066230e+00 -6.78705513e-01
-8.03994238e-01 -3.53783160e-01 2.23831922e-01 -1.09780931e+00
-4.79584724e-01 -7.43646801e-01 -1.23795521e+00 4.37434129e-02
-1.68112993e-01 2.69698113e-01 5.95366538e-01 9.63448763e-01
3.16554189e-01 5.63572764e-01 6.43388152e-01 -1.05418980e+00
-1.22980535e+00 -9.33638215e-01 -5.23484588e-01 6.93250656e-01
2.18740046e-01 -1.36139905e+00 -4.23388660e-01 -1.24614738e-01] | [4.2136454582214355, 1.5421786308288574] |
c4358ac4-c66f-4e52-921f-8ac72322ba57 | qasc-a-dataset-for-question-answering-via | 1910.11473 | null | https://arxiv.org/abs/1910.11473v2 | https://arxiv.org/pdf/1910.11473v2.pdf | QASC: A Dataset for Question Answering via Sentence Composition | Composing knowledge from multiple pieces of texts is a key challenge in multi-hop question answering. We present a multi-hop reasoning dataset, Question Answering via Sentence Composition(QASC), that requires retrieving facts from a large corpus and composing them to answer a multiple-choice question. QASC is the first dataset to offer two desirable properties: (a) the facts to be composed are annotated in a large corpus, and (b) the decomposition into these facts is not evident from the question itself. The latter makes retrieval challenging as the system must introduce new concepts or relations in order to discover potential decompositions. Further, the reasoning model must then learn to identify valid compositions of these retrieved facts using common-sense reasoning. To help address these challenges, we provide annotation for supporting facts as well as their composition. Guided by these annotations, we present a two-step approach to mitigate the retrieval challenges. We use other multiple-choice datasets as additional training data to strengthen the reasoning model. Our proposed approach improves over current state-of-the-art language models by 11% (absolute). The reasoning and retrieval problems, however, remain unsolved as this model still lags by 20% behind human performance. | ['Ashish Sabharwal', 'Tushar Khot', 'Peter Jansen', 'Peter Clark', 'Michal Guerquin'] | 2019-10-25 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 2.46813238e-01 6.07507527e-01 -1.07076801e-01 -1.03370063e-01
-1.59624350e+00 -8.67855430e-01 6.72546983e-01 7.44004369e-01
-3.27190995e-01 1.02834558e+00 4.97209638e-01 -4.37970221e-01
-4.51328158e-01 -7.65420675e-01 -8.19119632e-01 -9.33063924e-02
2.92332411e-01 9.31304812e-01 8.86438072e-01 -8.58838618e-01
1.96418762e-01 -1.34524584e-01 -1.68676007e+00 9.57822740e-01
1.21145058e+00 1.11999559e+00 1.85355216e-01 6.24867141e-01
-8.32378507e-01 1.19920790e+00 -6.79391384e-01 -7.61552215e-01
1.18036792e-02 -3.19191426e-01 -1.74471974e+00 -2.99167842e-01
7.84162879e-01 -1.39954194e-01 1.55927271e-01 8.83658946e-01
1.60594821e-01 1.10577904e-01 4.20634627e-01 -9.12133753e-01
-7.19751120e-01 9.43608046e-01 -3.88046876e-02 2.87908763e-01
1.07766664e+00 -3.67810935e-01 1.71950984e+00 -7.15519845e-01
7.97294319e-01 1.16449344e+00 3.89195085e-01 5.22063673e-01
-7.31854677e-01 -1.35712132e-01 2.99485862e-01 4.25616175e-01
-1.28557336e+00 -4.07872885e-01 5.23432970e-01 -9.96139199e-02
1.34929848e+00 7.06890583e-01 2.99773604e-01 7.11626053e-01
-3.13441455e-01 8.94784212e-01 7.14370608e-01 -7.36735761e-01
3.68874744e-02 2.07524717e-01 4.51465517e-01 7.65306771e-01
2.68067747e-01 -7.70659685e-01 -6.05372131e-01 -3.85764718e-01
-1.00183241e-01 -3.77088785e-01 -4.03515846e-01 2.16520607e-01
-9.62173760e-01 6.36072755e-01 1.95914999e-01 4.31903243e-01
-4.62506354e-01 -1.80818453e-01 2.81105965e-01 4.57150072e-01
8.73441696e-02 1.04614091e+00 -8.82322013e-01 1.71845138e-01
-7.72880852e-01 5.94012737e-01 1.26013625e+00 1.00661564e+00
7.12388813e-01 -7.91820645e-01 -4.63863254e-01 7.84033895e-01
2.72478938e-01 3.61253053e-01 4.19499040e-01 -1.08473396e+00
9.89092469e-01 1.07161856e+00 2.76371837e-01 -9.98522580e-01
-2.30514884e-01 -3.92637670e-01 -3.05304229e-01 -8.18670630e-01
4.96685892e-01 1.42660841e-01 -3.44408542e-01 1.55829215e+00
4.58529890e-01 -8.99236277e-02 5.23162127e-01 6.90577507e-01
1.22258747e+00 5.40989935e-01 -4.48789820e-02 -2.11714461e-01
1.73558474e+00 -1.11006916e+00 -7.56996572e-01 -3.19980025e-01
7.56397665e-01 -8.39441419e-01 1.09399319e+00 2.37076774e-01
-1.20381427e+00 -3.04755181e-01 -9.35501575e-01 -4.67962652e-01
-4.75045353e-01 -7.89429918e-02 3.99554372e-01 2.74422437e-01
-9.70979035e-01 1.65914148e-01 -1.57173812e-01 -2.74377614e-01
1.25979800e-02 8.30934346e-02 -3.30176391e-02 -5.35950303e-01
-1.75493324e+00 9.49775517e-01 4.70315784e-01 -2.46498063e-01
-3.12372476e-01 -8.33729029e-01 -7.36858606e-01 3.09368908e-01
1.11764920e+00 -9.90757346e-01 1.32780123e+00 -4.14586991e-01
-1.00817835e+00 6.42971337e-01 -4.51075226e-01 -5.06207287e-01
1.11285314e-01 -4.24126893e-01 -6.79373503e-01 6.99315012e-01
5.82243979e-01 4.10139203e-01 3.87735754e-01 -1.35984588e+00
-9.21823680e-01 -2.20307857e-01 7.44056523e-01 3.36835831e-01
-2.39666000e-01 1.87138077e-02 -7.38008976e-01 -2.12909445e-01
1.94697618e-01 -5.17532468e-01 1.17830738e-01 -3.66860181e-01
-3.69281441e-01 -7.23181725e-01 5.97041428e-01 -7.13593960e-01
1.43897724e+00 -1.53141177e+00 -3.32401022e-02 8.10164586e-02
3.42536420e-01 2.29976431e-01 -2.50782609e-01 7.89542973e-01
2.82463193e-01 2.88669616e-01 -4.63702902e-02 -1.22412898e-01
7.34446794e-02 3.81870300e-01 -7.99876988e-01 -5.22139609e-01
3.92314434e-01 9.97802556e-01 -1.09240699e+00 -7.68106163e-01
-4.91843760e-01 -4.16586511e-02 -5.53922355e-01 9.25827473e-02
-8.96493316e-01 1.87801704e-01 -5.76370716e-01 7.02951074e-01
3.76291215e-01 -5.13511777e-01 1.89977169e-01 -4.07286435e-01
4.40127522e-01 8.16418886e-01 -1.15341485e+00 1.58548772e+00
-4.03290898e-01 1.90653317e-02 -2.67479002e-01 -9.86564040e-01
6.29954159e-01 4.68529433e-01 3.96677464e-01 -8.59924793e-01
-2.04817265e-01 4.72727239e-01 -2.52803385e-01 -1.02474856e+00
7.15964079e-01 -2.08836094e-01 -1.16141558e-01 5.01434624e-01
1.50667518e-01 -3.60188752e-01 8.42264473e-01 7.35037327e-01
1.41973376e+00 -2.27730602e-01 1.77161515e-01 -1.55219615e-01
1.03370249e+00 4.00485724e-01 3.73592645e-01 8.45628083e-01
3.48685354e-01 2.97400415e-01 4.85341996e-01 -3.02299559e-01
-5.10069191e-01 -7.99367309e-01 2.62134582e-01 1.02880824e+00
1.98175013e-01 -9.09168422e-01 -4.68444735e-01 -9.46506798e-01
-7.49095529e-02 8.02332163e-01 -3.27092767e-01 9.77178570e-03
-6.55982137e-01 -1.92219272e-01 7.67731309e-01 3.69534731e-01
4.60493475e-01 -7.00040758e-01 -4.93424565e-01 2.87057519e-01
-1.04060841e+00 -1.60149539e+00 -1.96544185e-01 -5.38899489e-02
-5.38745701e-01 -1.62471521e+00 -2.30251864e-01 -7.64971793e-01
5.53992331e-01 3.36376518e-01 1.57849956e+00 8.14857066e-01
1.07985362e-01 6.89660847e-01 -8.04806054e-01 -2.65647054e-01
-4.14802104e-01 1.79126218e-01 -2.71847576e-01 -2.19768837e-01
3.84583622e-01 -2.14654088e-01 -3.73634040e-01 1.20941781e-01
-1.17613900e+00 -7.36839846e-02 4.79444325e-01 7.58989573e-01
6.23729646e-01 4.43517298e-01 6.27986729e-01 -9.86492574e-01
1.05163825e+00 -6.29377902e-01 -3.83443266e-01 1.05121493e+00
-4.52870339e-01 4.62866485e-01 7.87755847e-01 -2.19359234e-01
-1.21123600e+00 -3.74537200e-01 -2.60605663e-01 1.60129890e-01
1.22621335e-01 1.05630565e+00 9.33860987e-02 2.82622039e-01
7.85350680e-01 8.50313157e-02 -3.79673421e-01 -4.23034370e-01
5.90269804e-01 3.83800656e-01 5.21852791e-01 -1.21088588e+00
7.65507281e-01 2.12742314e-01 -6.70944825e-02 -5.61399996e-01
-1.67780435e+00 -8.68182659e-01 -5.91197610e-01 -1.68905053e-02
6.47581518e-01 -8.75003457e-01 -7.58125484e-01 -3.74653518e-01
-1.42039084e+00 9.98724476e-02 -3.47508878e-01 2.86062178e-03
-1.61489949e-01 5.04788697e-01 -4.48076814e-01 -8.61287951e-01
-5.81354558e-01 -7.51027644e-01 1.22868741e+00 6.93971068e-02
-4.02028471e-01 -8.60732734e-01 -6.89362064e-02 1.12415719e+00
2.17805550e-01 -1.82209477e-01 1.43269479e+00 -9.35203254e-01
-6.87080443e-01 -2.23783672e-01 -9.51823071e-02 1.27543867e-01
3.92564982e-02 -2.81830877e-01 -6.06635213e-01 6.54376745e-02
3.32708694e-02 -7.79423594e-01 7.87232816e-01 -3.86902213e-01
8.81842613e-01 -5.89144051e-01 -3.31153348e-02 -3.70430619e-01
1.24923956e+00 -1.04337171e-01 3.60412478e-01 2.83353925e-01
3.40068847e-01 9.39676881e-01 6.06498122e-01 5.71697168e-02
1.03905249e+00 4.81782824e-01 1.49591282e-01 3.12888086e-01
-2.67452210e-01 -5.16906440e-01 2.94138268e-02 1.21353924e+00
1.84146285e-01 -1.53762653e-01 -1.11888909e+00 6.37011111e-01
-1.82734931e+00 -1.05097795e+00 -4.11865860e-01 1.79878485e+00
1.34191227e+00 9.17854905e-02 -1.56301394e-01 3.12200069e-01
1.57781988e-01 -1.87692791e-01 -3.50231946e-01 -8.45476612e-02
-3.06086034e-01 3.59764546e-01 -7.78005421e-02 7.66273916e-01
-6.93823755e-01 1.06407404e+00 6.05057096e+00 8.12492073e-01
-5.69127262e-01 3.28417458e-02 9.72796232e-02 1.90016776e-01
-8.01466048e-01 3.68549645e-01 -1.07327580e+00 -7.62100425e-03
6.59962714e-01 -1.25031978e-01 3.91367495e-01 3.62190425e-01
-4.98287559e-01 -4.34937596e-01 -1.04192448e+00 4.79703963e-01
5.29876947e-01 -1.51383364e+00 5.67205966e-01 -5.83079278e-01
5.76897502e-01 -3.43581170e-01 -2.75059670e-01 6.27705634e-01
3.70642364e-01 -7.04085350e-01 7.00521708e-01 5.04707038e-01
4.09275383e-01 -3.72745097e-01 8.37326288e-01 6.58562601e-01
-1.23555052e+00 -1.90275520e-01 -2.52743781e-01 -5.81106097e-02
2.46872932e-01 5.46964169e-01 -7.79099524e-01 1.14334571e+00
6.74026847e-01 2.04178885e-01 -8.97322953e-01 6.51670337e-01
-5.99960446e-01 4.86433864e-01 -4.04969931e-01 -2.04105094e-01
2.42347851e-01 3.07578415e-01 3.66894871e-01 8.73231471e-01
1.88856795e-01 4.89542902e-01 1.44702569e-01 8.38816524e-01
-2.97909796e-01 1.76312923e-01 -1.16812848e-01 -4.77762334e-02
5.32537401e-01 9.63543713e-01 -3.94866377e-01 -6.90501511e-01
-5.82489252e-01 7.22259521e-01 6.80172980e-01 3.81561607e-01
-3.98942232e-01 -3.31449687e-01 1.86132938e-01 -1.06198557e-01
1.40712485e-01 2.15963628e-02 -7.19405757e-03 -1.49538124e+00
7.01587737e-01 -1.09447372e+00 9.15053844e-01 -8.70418608e-01
-1.45257199e+00 6.47813499e-01 3.82455252e-02 -8.83753240e-01
-2.14590803e-01 -3.43301654e-01 -1.43315762e-01 7.65554726e-01
-2.06310010e+00 -1.06240869e+00 -9.04181078e-02 6.76425755e-01
6.33998454e-01 1.11146413e-01 9.70800579e-01 4.46182489e-01
-1.00457571e-01 3.62828523e-01 -5.28423607e-01 3.69083434e-02
7.04846323e-01 -1.30500269e+00 -5.07982895e-02 8.73852074e-01
4.62212622e-01 9.94324863e-01 6.94374681e-01 -7.67476618e-01
-1.63171220e+00 -8.03508341e-01 1.52734804e+00 -7.88829625e-01
9.06076968e-01 -5.47756366e-02 -1.22828877e+00 5.30976772e-01
2.48254418e-01 -1.92836836e-01 7.92887866e-01 4.07543272e-01
-7.35638857e-01 -9.51996744e-02 -8.58163714e-01 6.15703046e-01
8.53909075e-01 -6.80599391e-01 -1.39602757e+00 5.23336172e-01
1.23412359e+00 -4.44858134e-01 -7.59339273e-01 3.66358250e-01
2.93786973e-01 -5.57441056e-01 9.49967802e-01 -9.49802935e-01
6.55624390e-01 -5.10964632e-01 -4.18820590e-01 -9.63661849e-01
1.77926391e-01 -5.09778678e-01 -5.50953686e-01 1.23328769e+00
1.04607642e+00 -4.26872998e-01 4.40846503e-01 9.10979748e-01
-5.58759421e-02 -8.39143395e-01 -9.00823355e-01 -7.38603175e-01
1.44433469e-01 -4.72372591e-01 8.59621286e-01 9.29769397e-01
4.37993169e-01 9.06188369e-01 1.55773398e-03 3.33512396e-01
1.85071096e-01 3.43843907e-01 5.69990575e-01 -1.20609713e+00
-2.57937670e-01 -1.92866281e-01 2.45689616e-01 -1.45672977e+00
1.68836638e-01 -9.97143090e-01 1.87467095e-02 -2.06172299e+00
1.57370627e-01 -5.84344566e-01 -7.83982873e-02 6.70879185e-01
-6.57494009e-01 -2.74154097e-01 2.22643271e-01 2.03655541e-01
-1.18407261e+00 4.98890191e-01 1.25435412e+00 -3.28075141e-01
1.33353829e-01 -1.64737046e-01 -1.14788151e+00 4.33896512e-01
5.28950632e-01 -3.68404359e-01 -7.17202425e-01 -6.99690878e-01
9.68929231e-01 4.86527652e-01 1.62620485e-01 -7.97357976e-01
8.43548298e-01 -7.39002153e-02 -2.20328137e-01 -6.16002083e-01
4.20519769e-01 -9.79067802e-01 -4.12378639e-01 2.03520596e-01
-6.24505758e-01 1.56822786e-01 1.37517288e-01 4.33741689e-01
-5.61367095e-01 -4.03392017e-01 1.34811597e-02 -2.76388437e-01
-5.52585423e-01 -1.07593141e-01 -2.19320476e-01 7.92477846e-01
4.76950377e-01 3.65802139e-01 -9.41890597e-01 -4.16093171e-01
-3.85185599e-01 6.99639201e-01 -2.00247571e-01 5.15254676e-01
6.23753309e-01 -1.12774658e+00 -7.28518248e-01 -3.36465299e-01
5.03465891e-01 2.22785145e-01 2.42752567e-01 7.87603915e-01
-2.56534100e-01 7.27100611e-01 3.75458628e-01 -1.36316687e-01
-1.18687093e+00 6.39811397e-01 4.83713523e-02 -8.35794210e-01
-4.92907882e-01 8.97232771e-01 -5.05842686e-01 -4.91382807e-01
1.71836913e-01 -5.00713646e-01 -5.80985963e-01 2.27957383e-01
7.36955106e-01 3.47729512e-02 3.72845262e-01 -3.84219766e-01
-2.57981241e-01 4.10546422e-01 -1.44482121e-01 -1.33067727e-01
1.14048433e+00 -3.29412252e-01 -5.18262327e-01 1.94372207e-01
9.00916755e-01 2.31185362e-01 -3.30190390e-01 -8.15948546e-01
5.34015238e-01 -1.43535137e-01 -3.09754908e-01 -1.28470790e+00
-4.92105752e-01 5.17789304e-01 -3.50405991e-01 5.64057529e-01
1.02663338e+00 4.05966461e-01 1.27722037e+00 9.13306892e-01
2.77503043e-01 -1.11786604e+00 3.41283560e-01 9.34981108e-01
1.08261192e+00 -1.21568179e+00 -3.25463638e-02 -6.83561206e-01
-5.70922017e-01 1.09414113e+00 6.35514677e-01 4.43128020e-01
3.64621401e-01 -2.12368462e-02 1.46350145e-01 -6.57661498e-01
-1.17704773e+00 -5.75985909e-01 7.38207161e-01 9.00522172e-02
2.83725411e-01 -2.95530558e-01 -4.68149096e-01 9.13448453e-01
-4.16469604e-01 -1.89525083e-01 1.61572307e-01 1.01230133e+00
-5.69092929e-01 -1.16077638e+00 -3.08810413e-01 5.15321016e-01
-5.13972461e-01 -3.61124307e-01 -6.88957274e-01 3.30544353e-01
-1.07419938e-02 1.55319989e+00 -5.52268982e-01 -1.85345963e-01
5.52104235e-01 4.28041428e-01 3.07930738e-01 -7.61304200e-01
-6.25990927e-01 -4.91383225e-01 6.57888651e-01 -3.86374623e-01
-6.26178026e-01 -2.83160836e-01 -1.35636735e+00 7.33825266e-02
-4.83839154e-01 6.99188411e-01 2.81477898e-01 1.53653693e+00
4.33203965e-01 5.88464916e-01 1.08622357e-01 2.99686521e-01
-5.73327601e-01 -7.18470216e-01 1.51512830e-03 4.87959772e-01
3.81655127e-01 -4.40707892e-01 -2.05251932e-01 5.55204451e-02] | [10.684418678283691, 7.942838191986084] |
e538348e-5cff-4279-9310-616314ef1215 | jigsawgan-self-supervised-learning-for | 2101.07555 | null | https://arxiv.org/abs/2101.07555v3 | https://arxiv.org/pdf/2101.07555v3.pdf | JigsawGAN: Auxiliary Learning for Solving Jigsaw Puzzles with Generative Adversarial Networks | The paper proposes a solution based on Generative Adversarial Network (GAN) for solving jigsaw puzzles. The problem assumes that an image is divided into equal square pieces, and asks to recover the image according to information provided by the pieces. Conventional jigsaw puzzle solvers often determine the relationships based on the boundaries of pieces, which ignore the important semantic information. In this paper, we propose JigsawGAN, a GAN-based auxiliary learning method for solving jigsaw puzzles with unpaired images (with no prior knowledge of the initial images). We design a multi-task pipeline that includes, (1) a classification branch to classify jigsaw permutations, and (2) a GAN branch to recover features to images in correct orders. The classification branch is constrained by the pseudo-labels generated according to the shuffled pieces. The GAN branch concentrates on the image semantic information, where the generator produces the natural images to fool the discriminator, while the discriminator distinguishes whether a given image belongs to the synthesized or the real target domain. These two branches are connected by a flow-based warp module that is applied to warp features to correct the order according to the classification results. The proposed method can solve jigsaw puzzles more efficiently by utilizing both semantic information and boundary information simultaneously. Qualitative and quantitative comparisons against several representative jigsaw puzzle solvers demonstrate the superiority of our method. | ['Bing Zeng', 'Guanghui Liu', 'Guangfu Wang', 'Shuaicheng Liu', 'Ru Li'] | 2021-01-19 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 5.01185119e-01 1.24903254e-01 -4.58773747e-02 2.71435887e-01
-7.60864377e-01 -1.05470741e+00 3.69195670e-01 -7.37013340e-01
1.81238294e-01 7.63379157e-01 5.97449318e-02 -8.65132883e-02
-1.21750042e-01 -1.18781292e+00 -7.61274934e-01 -1.04167664e+00
5.23114800e-01 6.89333975e-01 6.06007501e-02 -3.70490342e-01
5.08802474e-01 3.60641152e-01 -1.03039312e+00 7.61260033e-01
1.02285647e+00 8.87308836e-01 4.76492047e-01 7.60167599e-01
-3.35543782e-01 1.03502333e+00 -1.00188947e+00 -3.51166099e-01
6.46032989e-01 -1.13156950e+00 -8.77121091e-01 3.09739947e-01
2.72001314e-04 -1.79707184e-01 -3.36303234e-01 1.03152180e+00
1.33847311e-01 -1.63271166e-02 5.70957780e-01 -1.54940355e+00
-9.51234102e-01 2.52750397e-01 -5.41818798e-01 -1.28722548e-01
4.81312752e-01 6.56695426e-01 9.83846247e-01 -2.91449100e-01
9.28519487e-01 8.89640212e-01 4.73983496e-01 5.88091493e-01
-1.17104769e+00 -7.01172233e-01 -2.50399947e-01 4.83192563e-01
-9.49379027e-01 -3.69799845e-02 1.22673893e+00 -5.07837951e-01
4.13636595e-01 3.72821480e-01 7.94445395e-01 1.18288219e+00
2.97804266e-01 6.13718450e-01 1.24641943e+00 -2.98718661e-01
3.56923252e-01 -1.69711784e-01 -1.89246193e-01 7.37780750e-01
-4.05171216e-02 1.47993624e-01 -3.45978528e-01 5.87070212e-02
1.06639421e+00 -4.10507381e-01 -3.95033717e-01 -3.74901682e-01
-9.90298808e-01 1.01082337e+00 6.28189266e-01 2.95365661e-01
-4.76794541e-01 -6.69629946e-02 3.02627664e-02 4.06269282e-01
-7.20061585e-02 9.45868671e-01 -1.95846319e-01 1.39823347e-01
-9.40751135e-01 2.97651172e-01 7.54880011e-01 8.61930251e-01
8.94417048e-01 1.67736486e-01 -4.02671844e-01 5.48730493e-01
-2.04645447e-03 2.58088291e-01 5.04073203e-01 -1.11628735e+00
6.58807755e-01 6.82687104e-01 -1.28916465e-02 -1.35226226e+00
1.03329783e-02 -3.28846991e-01 -8.35399389e-01 6.08187914e-01
5.85265160e-01 1.51517898e-01 -1.05833375e+00 1.62088382e+00
1.49286911e-01 9.07939449e-02 7.08271191e-02 1.24376965e+00
9.30098176e-01 8.68289649e-01 -5.49472213e-01 1.43628120e-01
1.41635668e+00 -1.46156311e+00 -5.87153494e-01 -4.28374916e-01
1.87615424e-01 -7.99597383e-01 1.16631341e+00 4.76625204e-01
-9.69823599e-01 -5.17564118e-01 -1.30584800e+00 -2.51710206e-01
-9.66814011e-02 2.94972628e-01 5.53377151e-01 5.96374989e-01
-1.02862787e+00 5.42307079e-01 -3.96502912e-01 1.12500094e-01
5.94035447e-01 3.20012003e-01 -4.28451151e-01 -2.37257361e-01
-8.67177546e-01 5.50952733e-01 4.01302934e-01 1.58110335e-01
-1.11064363e+00 -4.67882276e-01 -7.90742695e-01 1.81635827e-01
3.91322076e-01 -8.67880940e-01 5.76605618e-01 -1.49024200e+00
-1.85518301e+00 9.43382084e-01 2.48596355e-01 -8.02667737e-02
6.03365839e-01 6.25982046e-01 2.31542945e-01 2.66360432e-01
3.36965770e-01 6.66299820e-01 1.27962840e+00 -1.53675985e+00
-5.07421970e-01 -2.78064966e-01 3.74159575e-01 2.49803532e-02
4.57346380e-01 -4.50942039e-01 -4.10576642e-01 -9.49011207e-01
2.47211009e-01 -8.39068353e-01 -3.41349654e-03 -1.87653199e-01
-7.54101276e-01 2.30195269e-01 9.68687475e-01 -1.15263867e+00
5.02554476e-01 -2.03336596e+00 7.03385174e-01 1.75414950e-01
1.84890643e-01 1.33357406e-01 -6.95895731e-01 2.99208313e-01
-2.19218686e-01 -1.28475815e-01 -4.37331021e-01 -1.27177507e-01
-1.68453231e-01 2.33853236e-01 -4.92421567e-01 2.65235007e-01
2.05365121e-01 1.27286553e+00 -8.42259347e-01 -1.23974472e-01
-1.49372499e-02 1.14142619e-01 -5.05440354e-01 6.18534088e-01
-3.71463776e-01 9.07318354e-01 -4.43878591e-01 5.88205397e-01
8.73191357e-01 -8.94432813e-02 1.16176255e-01 -3.48234385e-01
1.92015409e-01 5.76316230e-02 -9.02959466e-01 1.87964237e+00
-3.88311028e-01 5.02814233e-01 1.89965978e-01 -1.24762988e+00
1.25282836e+00 -6.25630096e-02 1.74238458e-01 -6.29080296e-01
1.50759235e-01 1.74528018e-01 5.08834645e-02 -5.36378920e-01
3.30786586e-01 -7.00808764e-02 -2.41808176e-01 7.00528860e-01
1.25981569e-01 -8.14896643e-01 2.51759857e-01 3.40139456e-02
1.23453128e+00 4.60794389e-01 9.44951773e-02 1.08573996e-02
5.42862415e-01 3.70082170e-01 7.32900858e-01 5.24242282e-01
2.38195419e-01 1.01990616e+00 9.17962730e-01 -7.93222010e-01
-1.24413323e+00 -1.04790831e+00 7.34401464e-01 4.29241896e-01
4.40106064e-01 -4.49582149e-04 -1.03259826e+00 -8.80194306e-01
-3.09791982e-01 6.58674479e-01 -8.87621641e-01 -3.68650049e-01
-5.99440932e-01 -5.19514680e-01 3.72559011e-01 1.85077220e-01
8.93080294e-01 -1.48262835e+00 -4.99826908e-01 -1.19457670e-01
-4.71603662e-01 -9.07152474e-01 -6.55969977e-01 -7.82801583e-02
-4.82085168e-01 -1.42142546e+00 -5.92811346e-01 -1.05752552e+00
1.01811278e+00 3.34526956e-01 7.87988603e-01 1.60849378e-01
-3.14172953e-01 9.61788148e-02 -5.56398869e-01 2.81180352e-01
-5.84563971e-01 1.01208620e-01 -7.25950599e-01 1.72991067e-01
-2.35735849e-01 -7.49000251e-01 -4.92440641e-01 5.54504931e-01
-1.03483295e+00 5.30352473e-01 6.74985349e-01 1.10873485e+00
5.52477181e-01 2.32466340e-01 1.41368374e-01 -8.21276307e-01
5.20069718e-01 -3.35325658e-01 -6.36026561e-01 3.53825301e-01
-4.14758399e-02 1.86726317e-01 9.71130252e-01 -5.36858916e-01
-9.43486810e-01 1.41704500e-01 1.07695274e-01 -5.29252052e-01
3.52023207e-02 4.22690138e-02 -6.64534807e-01 -3.08671325e-01
3.53771806e-01 6.62166893e-01 6.50129169e-02 -1.63766742e-01
3.56977046e-01 3.68557811e-01 7.87037790e-01 -6.93974674e-01
9.66849029e-01 5.09493828e-01 4.47871573e-02 -3.35978687e-01
-4.64716762e-01 2.61286408e-01 -5.68497479e-01 -1.58957943e-01
9.89215970e-01 -2.92083353e-01 -7.13954329e-01 6.87069416e-01
-1.45345223e+00 -5.62511504e-01 -5.84115207e-01 6.29187189e-03
-8.77112806e-01 3.28530818e-01 -7.70422399e-01 -3.40075821e-01
-2.29445428e-01 -1.47714126e+00 8.04809809e-01 4.08836603e-01
1.44673996e-02 -6.16691351e-01 1.22542694e-01 9.64855075e-01
2.26360008e-01 3.59894127e-01 1.33050609e+00 -2.85966337e-01
-1.00296593e+00 -8.43542367e-02 -3.33549857e-01 1.82668716e-01
3.63733381e-01 -2.93781728e-01 -7.02087283e-01 -2.67966837e-01
4.09756750e-01 -2.79032826e-01 7.81105697e-01 3.72800797e-01
1.07381666e+00 -4.42432404e-01 1.78268608e-02 9.50843155e-01
1.32169092e+00 7.30549514e-01 1.28533947e+00 4.88994032e-01
8.19335520e-01 5.80187380e-01 3.30572456e-01 1.04133815e-01
1.52947709e-01 6.77383065e-01 6.73710167e-01 -1.03744060e-01
-6.01717412e-01 -5.92435062e-01 2.80544668e-01 7.37122416e-01
-1.00619800e-01 -3.78678501e-01 -4.24824536e-01 3.38544071e-01
-1.60816348e+00 -8.48360717e-01 -3.37625518e-02 1.87927222e+00
5.76371193e-01 -2.24796474e-01 -2.61954129e-01 2.33603626e-01
8.40046048e-01 3.57335269e-01 -6.29376709e-01 -3.68072301e-01
-2.78702319e-01 3.64346355e-01 2.72642374e-01 4.70217735e-01
-7.45251477e-01 1.03010273e+00 5.39749813e+00 1.19910741e+00
-9.84208822e-01 8.24019685e-02 8.19043100e-01 2.83138812e-01
-5.99588811e-01 4.02849019e-01 -8.94616172e-02 6.56148255e-01
-2.03223899e-01 2.29491979e-01 1.05318022e+00 5.42787552e-01
-3.53715606e-02 -2.68834859e-01 -7.89883494e-01 9.27164495e-01
4.73895252e-01 -1.45631933e+00 3.08343887e-01 2.81068366e-02
9.68804598e-01 -8.60780001e-01 1.32867947e-01 5.30526191e-02
3.32140058e-01 -1.15684974e+00 7.28584349e-01 5.27484119e-01
7.91574657e-01 -6.40233636e-01 6.86545193e-01 3.59752476e-01
-1.01157653e+00 -1.43761337e-01 -3.39050621e-01 -5.71900569e-02
8.02153498e-02 3.23935002e-01 -6.82814896e-01 8.88173223e-01
1.88789845e-01 4.54957813e-01 -2.70137489e-01 7.10462093e-01
-9.51484442e-01 2.62869507e-01 2.02960074e-01 3.79216582e-01
2.89099723e-01 -8.48609447e-01 6.74734771e-01 3.84678602e-01
4.90186900e-01 1.61198109e-01 -4.03750734e-03 1.43312812e+00
-3.33266258e-02 -2.64809042e-01 -5.20821333e-01 -9.29636806e-02
2.53901392e-01 1.29175389e+00 -1.06613314e+00 -5.16519696e-02
8.91842507e-03 1.56266570e+00 2.20301762e-01 3.50959867e-01
-8.63919795e-01 -5.87397873e-01 2.92736888e-01 -9.35252532e-02
4.09123987e-01 -1.61243886e-01 -6.21307909e-01 -1.01708674e+00
8.24692380e-03 -1.12197566e+00 3.07883650e-01 -1.21431077e+00
-1.19336414e+00 8.24928463e-01 -5.12528181e-01 -1.18255293e+00
-1.15261331e-01 -6.15156949e-01 -8.19841325e-01 9.20226276e-01
-1.22214746e+00 -1.40955555e+00 -6.20008171e-01 4.89603192e-01
8.44067514e-01 -4.56921935e-01 6.76347017e-01 -7.10438788e-02
-4.86461282e-01 3.66994321e-01 -1.34632707e-01 3.96123469e-01
2.58324295e-01 -1.03021514e+00 3.64814460e-01 8.66175294e-01
2.70590693e-01 -1.31562576e-01 5.71148276e-01 -7.21020758e-01
-1.52536213e+00 -7.37794816e-01 4.03143018e-01 -1.23757429e-01
2.97532111e-01 -2.36607522e-01 -7.64247239e-01 5.92631578e-01
2.31627718e-01 -2.29327962e-01 2.36507744e-01 -6.29720747e-01
-4.56105858e-01 -3.40454169e-02 -1.48031580e+00 3.61509413e-01
9.44510043e-01 -2.17784226e-01 -6.37127161e-01 1.87230006e-01
5.98351419e-01 -5.24558961e-01 -2.22610071e-01 1.05392337e-01
4.14708674e-01 -9.90717292e-01 9.30384040e-01 -6.09139085e-01
1.15946603e+00 -5.27716637e-01 3.29745524e-02 -1.73079133e+00
-5.77493131e-01 -5.17012537e-01 3.81776392e-01 1.00251031e+00
1.30190507e-01 -5.64334989e-01 1.16102815e+00 2.78238565e-01
-8.06563646e-02 -3.72304469e-01 -1.06011343e+00 -7.28112817e-01
-1.57189518e-01 4.56905961e-02 8.77492011e-01 1.11861932e+00
-2.55109459e-01 2.46989742e-01 -5.95952928e-01 2.28649247e-02
4.31682289e-01 6.97871327e-01 9.96520042e-01 -6.59123123e-01
-6.72284722e-01 -4.87186253e-01 -4.63210851e-01 -9.62055206e-01
2.45362028e-01 -9.46579993e-01 1.56838447e-01 -1.60662115e+00
2.88319498e-01 -4.87868667e-01 2.30336905e-01 6.24719262e-01
-3.26318629e-02 5.27785122e-01 3.96294534e-01 2.48973057e-01
-7.33174980e-02 6.41011238e-01 1.78318238e+00 -5.82848907e-01
-2.15624511e-01 -9.57409814e-02 -5.95752656e-01 4.44287747e-01
8.37722957e-01 -4.97178257e-01 -3.98716927e-01 -7.19019711e-01
1.43118784e-01 3.32209200e-01 4.44839418e-01 -8.29272270e-01
1.77857801e-01 -3.42209905e-01 2.55675375e-01 -9.77866724e-02
3.78638387e-01 -7.59886324e-01 6.04809165e-01 5.66371024e-01
-1.17301285e-01 -4.50940989e-02 -1.63318124e-02 2.80296713e-01
-1.46598756e-01 -5.18644392e-01 6.18097126e-01 -3.84653091e-01
-4.64945465e-01 7.98694789e-02 -2.64342546e-01 2.05266178e-02
1.22928166e+00 -3.80691588e-01 -4.66485918e-01 -5.35228550e-01
-5.39503753e-01 -1.26872838e-01 5.79220831e-01 2.41942525e-01
7.49224305e-01 -1.51629353e+00 -7.52828062e-01 6.20728076e-01
-2.97370315e-01 6.60563633e-03 5.10140717e-01 4.63416755e-01
-9.14444566e-01 1.21118754e-01 -7.72779346e-01 -2.46136174e-01
-1.04569781e+00 6.80702507e-01 2.86717355e-01 -6.64846301e-01
-3.99073124e-01 7.88552880e-01 4.71187979e-01 -4.32663351e-01
-2.98192382e-01 -1.51875429e-02 -5.17487563e-02 2.98348404e-02
3.85654658e-01 3.52567405e-01 -1.04877926e-01 -5.82129955e-01
4.53932844e-02 7.45986044e-01 1.86388656e-01 -1.10282591e-02
1.36936235e+00 3.37386847e-01 -4.72683668e-01 -1.00830838e-01
9.38424528e-01 1.22459091e-01 -1.33890378e+00 1.79600343e-01
-4.84914452e-01 -8.57747197e-01 -2.55139798e-01 -1.01828134e+00
-1.69832838e+00 6.68742001e-01 5.42200729e-02 -1.02934847e-02
1.51866710e+00 -1.42556787e-01 9.95373726e-01 -2.63093799e-01
6.97111189e-01 -6.19557202e-01 4.04151231e-01 3.48171085e-01
1.17878330e+00 -6.96992278e-01 -2.23548949e-01 -6.79934204e-01
-7.22443283e-01 1.27802777e+00 6.08893573e-01 -6.21818483e-01
-8.59910026e-02 2.90057093e-01 -2.46994235e-02 -1.71305090e-01
-3.49632293e-01 9.16681364e-02 1.49326012e-01 8.46478939e-01
-5.70154369e-01 1.45526435e-02 -3.17626297e-01 1.06769896e+00
-4.04351979e-01 -4.27350886e-02 4.93946642e-01 5.10221362e-01
-1.27447426e-01 -1.27185214e+00 -8.24374437e-01 1.20765857e-01
1.08068302e-01 1.24825299e-01 -6.10086560e-01 5.96900165e-01
5.29976726e-01 8.16636801e-01 1.81031793e-01 -6.45501912e-01
2.29772598e-01 -3.24726403e-01 6.99899793e-01 -3.64194065e-01
-6.98333979e-01 -4.34177935e-01 -2.30159432e-01 -5.87838054e-01
-6.37382567e-02 -2.26162821e-01 -8.94034445e-01 -2.89596707e-01
6.98994249e-02 2.45343372e-01 4.66095626e-01 9.21181977e-01
3.96548182e-01 6.42920673e-01 9.63587701e-01 -1.00018871e+00
-3.80156264e-02 -7.43178427e-01 -5.55752993e-01 4.39276755e-01
-2.97560077e-02 -4.87958342e-01 -4.47091103e-01 4.03563157e-02] | [11.557024955749512, -0.5976397395133972] |
fcd018b0-f156-4054-9425-6a3b253d1b79 | evaluation-of-output-embeddings-for-fine | 1409.8403 | null | http://arxiv.org/abs/1409.8403v2 | http://arxiv.org/pdf/1409.8403v2.pdf | Evaluation of Output Embeddings for Fine-Grained Image Classification | Image classification has advanced significantly in recent years with the
availability of large-scale image sets. However, fine-grained classification
remains a major challenge due to the annotation cost of large numbers of
fine-grained categories. This project shows that compelling classification
performance can be achieved on such categories even without labeled training
data. Given image and class embeddings, we learn a compatibility function such
that matching embeddings are assigned a higher score than mismatching ones;
zero-shot classification of an image proceeds by finding the label yielding the
highest joint compatibility score. We use state-of-the-art image features and
focus on different supervised attributes and unsupervised output embeddings
either derived from hierarchies or learned from unlabeled text corpora. We
establish a substantially improved state-of-the-art on the Animals with
Attributes and Caltech-UCSD Birds datasets. Most encouragingly, we demonstrate
that purely unsupervised output embeddings (learned from Wikipedia and improved
with fine-grained text) achieve compelling results, even outperforming the
previous supervised state-of-the-art. By combining different output embeddings,
we further improve results. | ['Bernt Schiele', 'Scott Reed', 'Zeynep Akata', 'Honglak Lee', 'Daniel Walter'] | 2014-09-30 | evaluation-of-output-embeddings-for-fine-1 | http://openaccess.thecvf.com/content_cvpr_2015/html/Akata_Evaluation_of_Output_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Akata_Evaluation_of_Output_2015_CVPR_paper.pdf | cvpr-2015-6 | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 3.83803211e-02 -6.29256666e-02 -2.97684610e-01 -8.07138503e-01
-6.29515707e-01 -6.45864308e-01 9.15014505e-01 4.00326461e-01
-7.65421808e-01 5.36885202e-01 3.65698785e-01 3.15825939e-01
-8.77030864e-02 -7.63448417e-01 -5.69411218e-01 -4.99986142e-01
-1.45837024e-01 6.29963696e-01 1.74506754e-01 -7.84750357e-02
2.32643202e-01 3.19770664e-01 -2.02320385e+00 3.86431843e-01
5.32616556e-01 1.32898378e+00 1.39813140e-01 5.42708337e-01
-2.07120970e-01 5.25085449e-01 -3.22281301e-01 -3.55892956e-01
3.56533647e-01 2.14041919e-02 -8.48880708e-01 1.43684521e-01
1.18126965e+00 -3.57948303e-01 -2.97015667e-01 1.01348674e+00
1.41905576e-01 -2.39812005e-02 1.07987535e+00 -1.56521237e+00
-1.32904196e+00 2.02841222e-01 -2.86611944e-01 1.66633539e-02
-5.37534989e-02 -1.98998880e-02 1.43053782e+00 -1.05158055e+00
6.33230746e-01 1.06945527e+00 6.52594805e-01 5.87943554e-01
-1.63319504e+00 -6.08054817e-01 1.42041445e-01 8.15388709e-02
-1.47919202e+00 -2.05584824e-01 3.72680053e-02 -8.95525217e-01
1.26331401e+00 -2.65400648e-01 4.55429524e-01 8.10860217e-01
-1.01821885e-01 2.58056194e-01 1.06550574e+00 -4.86699671e-01
6.35392889e-02 3.43181551e-01 3.53129715e-01 1.04170907e+00
2.61341304e-01 1.74512357e-01 -3.86903048e-01 -3.45645398e-02
5.48607767e-01 3.21442962e-01 1.19966753e-01 -8.48246992e-01
-1.50983715e+00 1.09320343e+00 8.74211192e-01 2.71896809e-01
-1.13100717e-02 2.40749851e-01 3.70918959e-01 3.64499062e-01
5.05467832e-01 7.99839616e-01 -5.62272847e-01 1.07564270e-01
-8.22835445e-01 2.39651278e-01 5.92116594e-01 1.12796044e+00
1.15485406e+00 -1.70393094e-01 -1.90354049e-01 1.04911518e+00
2.08493531e-01 2.62633145e-01 6.88616097e-01 -1.00308514e+00
6.52260557e-02 6.49238646e-01 -2.22979411e-02 -6.90136731e-01
-2.54940063e-01 -4.51626927e-01 -7.38568962e-01 5.69906056e-01
4.94422555e-01 2.12912604e-01 -1.27188683e+00 1.75130188e+00
5.60537949e-02 -1.50617927e-01 -2.76554544e-02 8.26170743e-01
7.16644228e-01 5.59511364e-01 4.38713759e-01 4.40328807e-01
1.45378447e+00 -1.28948224e+00 -2.92054772e-01 -3.00211966e-01
7.06455588e-01 -4.77001250e-01 1.13790631e+00 2.78474241e-01
-4.09780502e-01 -9.32457864e-01 -1.28797579e+00 -3.14724445e-01
-9.61008012e-01 1.92714453e-01 7.30915844e-01 5.22689939e-01
-1.05518651e+00 6.34949982e-01 -4.62957799e-01 -7.58131683e-01
6.54450655e-01 3.56635630e-01 -1.00031102e+00 -2.45345369e-01
-8.80249083e-01 1.10754585e+00 3.60584468e-01 -4.53597546e-01
-1.08317876e+00 -7.70099103e-01 -1.01047373e+00 3.06931064e-02
-6.66707230e-04 -5.97544789e-01 1.01828158e+00 -7.14410484e-01
-7.80654073e-01 1.42222035e+00 2.83110350e-01 -4.82160330e-01
4.39977013e-02 -2.67069876e-01 -2.82849193e-01 1.26113981e-01
5.79850078e-01 1.49941456e+00 9.18203592e-01 -1.12181437e+00
-9.20690238e-01 -4.67656225e-01 2.18008846e-01 2.97281370e-02
-9.08115745e-01 -1.47061169e-01 -1.36664912e-01 -4.91638184e-01
-1.75860077e-01 -1.01137102e+00 -2.21224770e-01 5.67556202e-01
5.48278689e-02 -6.23219907e-01 6.34196877e-01 -9.15573463e-02
9.22564447e-01 -2.36649394e+00 1.92992166e-01 -3.87720883e-01
3.77456665e-01 -2.36879997e-02 -4.35251057e-01 2.18366429e-01
-1.17837660e-01 3.32406193e-01 -1.30174425e-03 -2.65426040e-01
4.37742263e-01 3.53657663e-01 -1.79608047e-01 4.21234608e-01
6.18113160e-01 6.95321918e-01 -8.63050580e-01 -5.72920203e-01
2.40409955e-01 9.01401341e-02 -7.87931561e-01 4.39519882e-01
-5.04902042e-02 -1.09888129e-01 -2.64695287e-01 5.37720203e-01
1.97191492e-01 -5.66962481e-01 -1.30796537e-01 -6.49910450e-01
-8.93596932e-02 -2.37521589e-01 -8.47596586e-01 1.87260199e+00
-5.09946465e-01 5.67580044e-01 -3.47659349e-01 -1.02852941e+00
8.35602582e-01 6.43413365e-02 7.88729414e-02 -5.38446128e-01
-1.21428862e-01 1.55943260e-01 3.19019072e-02 -3.66541147e-01
7.03575313e-01 -1.93703532e-01 -5.82564533e-01 3.54885250e-01
9.18092787e-01 -4.20407027e-01 3.25074494e-01 3.38768035e-01
9.06004488e-01 2.46044755e-01 3.35095912e-01 -5.07508337e-01
3.96616831e-02 3.29615712e-01 2.44127393e-01 7.69883096e-01
-3.84919673e-01 5.97837210e-01 1.57246634e-01 -6.28037214e-01
-1.54187012e+00 -1.12577713e+00 -5.43220341e-01 1.71923101e+00
-1.71012655e-02 -5.78789771e-01 -6.91319108e-01 -7.40823269e-01
4.70352173e-01 2.87549376e-01 -1.19352019e+00 2.30476372e-02
1.50341643e-02 -4.63789850e-01 5.09783268e-01 7.86792934e-01
3.46291184e-01 -8.32510114e-01 -2.94481754e-01 5.75496405e-02
1.76743522e-01 -1.15642262e+00 -2.98175573e-01 6.16154611e-01
-6.83206737e-01 -8.90577018e-01 -6.07778847e-01 -1.18385065e+00
7.26714075e-01 2.25302264e-01 1.23306692e+00 1.97446067e-02
-8.08927417e-01 3.00473124e-01 -4.42405343e-01 -1.55391052e-01
-8.31766799e-02 8.84521976e-02 3.23722482e-01 -1.65077165e-01
5.15856564e-01 -1.43032983e-01 -4.39421833e-01 4.04305816e-01
-8.39371145e-01 -2.01985389e-01 5.95899403e-01 1.24692035e+00
6.61063313e-01 -1.23939514e-01 7.40013540e-01 -9.76273298e-01
2.54046261e-01 -4.00769174e-01 -4.54506963e-01 2.00121656e-01
-6.93929195e-01 4.43877786e-01 7.82369316e-01 -4.28637594e-01
-7.52495527e-01 7.14480057e-02 -4.53026667e-02 -3.78515899e-01
-6.77521765e-01 1.08748987e-01 2.69322038e-01 -4.01035398e-02
9.71379220e-01 -2.24102318e-01 -2.06464797e-01 -4.32747155e-01
9.50197279e-01 9.28773642e-01 6.83114350e-01 -6.64419293e-01
6.95521355e-01 3.86909336e-01 -2.84310609e-01 -7.07683206e-01
-1.27592289e+00 -7.01560140e-01 -1.05527616e+00 1.76911473e-01
1.28744948e+00 -1.15751052e+00 -2.23103538e-01 3.13725352e-01
-7.24166453e-01 -2.67514378e-01 -4.49093074e-01 4.73187327e-01
-5.85770130e-01 -3.97984684e-02 -6.15146577e-01 -1.18639544e-01
2.25617923e-02 -9.38799143e-01 1.22755456e+00 1.72274470e-01
-3.33113074e-01 -8.21594417e-01 -1.42166065e-02 3.13324034e-01
4.58489567e-01 -3.14249620e-02 1.21497834e+00 -8.56030941e-01
-5.01609206e-01 -2.63573140e-01 -6.61685586e-01 4.22818750e-01
1.49087027e-01 -1.65638745e-01 -1.17440856e+00 -4.13282394e-01
-6.53252006e-01 -1.06634641e+00 1.01112235e+00 1.23396423e-02
1.11054635e+00 1.79635696e-02 -4.51556325e-01 5.67271292e-01
1.67313898e+00 -3.57219458e-01 2.68796414e-01 4.94717836e-01
7.86586285e-01 6.72421932e-01 7.34522164e-01 3.18304151e-01
2.86022216e-01 6.36787117e-01 4.08438295e-01 -5.81165068e-02
-3.73947859e-01 -3.82723153e-01 -1.20057814e-01 6.23763025e-01
1.66701272e-01 -8.48690048e-02 -6.92597270e-01 8.58853102e-01
-1.60723996e+00 -8.36597085e-01 1.80952415e-01 2.07827616e+00
8.91776323e-01 2.65865892e-01 -8.78814906e-02 -1.08561814e-01
5.82759261e-01 9.99597907e-02 -5.18729031e-01 -3.35369319e-01
3.21383178e-02 3.49037737e-01 5.52409172e-01 2.25387648e-01
-1.48734808e+00 1.03615832e+00 6.75730085e+00 8.33421350e-01
-7.71629751e-01 2.09269181e-01 4.47071016e-01 1.76495016e-01
1.81205664e-02 -6.46460205e-02 -1.05524004e+00 1.52301028e-01
8.65653694e-01 -4.18180898e-02 2.14086771e-01 1.17649198e+00
-6.13974810e-01 9.91309658e-02 -1.53529119e+00 9.48926747e-01
3.08454692e-01 -1.43085885e+00 1.57440528e-01 1.52568251e-01
8.57881188e-01 3.79596293e-01 -2.08934955e-02 6.01150095e-01
6.07362866e-01 -1.18203914e+00 7.48737693e-01 2.67875850e-01
1.33739328e+00 -2.46568888e-01 6.61890090e-01 1.10825993e-01
-1.24741030e+00 -2.69294381e-01 -7.64108300e-01 -1.64254963e-01
-2.47510940e-01 2.52092749e-01 -7.15588093e-01 1.38216645e-01
9.86943126e-01 9.14879978e-01 -9.34458494e-01 8.75023067e-01
-1.83134079e-01 2.56111711e-01 -7.16708824e-02 -3.88255604e-02
3.92791480e-01 1.82161048e-01 -2.06759632e-01 1.35554731e+00
1.18067272e-01 -2.12429482e-02 6.43704653e-01 6.08361304e-01
-4.38404024e-01 -1.48678929e-01 -8.32640767e-01 -2.52548367e-01
4.50855643e-01 1.64228392e+00 -5.35400867e-01 -6.14537716e-01
-6.70499980e-01 9.65426505e-01 8.24726164e-01 1.45130157e-02
-5.33983946e-01 -5.62455535e-01 9.12036419e-01 -1.92564707e-02
4.44798380e-01 -1.75984710e-01 -6.89718947e-02 -1.10078621e+00
-5.75274289e-01 -4.87413704e-01 5.12717366e-01 -8.90356243e-01
-1.86013937e+00 7.95997918e-01 -1.53085619e-01 -1.28123367e+00
-2.08156869e-01 -9.71051455e-01 1.84932500e-02 5.76723039e-01
-1.48016715e+00 -1.25887704e+00 -5.86014509e-01 4.19202149e-01
5.68278611e-01 -2.69210309e-01 1.51126671e+00 3.99484605e-01
-7.20728114e-02 6.83004022e-01 1.81805179e-01 4.27748203e-01
1.10139263e+00 -1.58316624e+00 2.61354595e-01 3.26498389e-01
3.43749106e-01 4.17060286e-01 4.44676340e-01 -2.35806510e-01
-1.01328254e+00 -1.22033215e+00 8.43949199e-01 -7.23823249e-01
1.01865602e+00 -5.16333640e-01 -6.37685180e-01 7.06915259e-01
2.61670440e-01 6.48460925e-01 8.03399980e-01 2.98109949e-01
-1.10043108e+00 -1.68268666e-01 -1.12637305e+00 2.68232107e-01
1.11963046e+00 -7.68189013e-01 -9.49024022e-01 2.60075331e-01
8.31871450e-01 1.11312553e-01 -1.23528409e+00 2.77017355e-01
8.63170981e-01 -6.71364367e-01 9.76202786e-01 -9.13016677e-01
5.57526052e-01 -4.69195604e-01 -6.26803041e-01 -1.59208238e+00
-8.44202936e-01 3.13834935e-01 4.94919449e-01 1.09713256e+00
4.61911291e-01 -4.20935094e-01 6.43452406e-01 3.97907436e-01
-2.69227326e-01 -4.98949528e-01 -4.98280823e-01 -9.48862553e-01
2.21105367e-01 4.74078991e-02 3.07597041e-01 9.78380680e-01
-3.70895229e-02 6.26095593e-01 -1.26174718e-01 6.22455217e-02
8.10792983e-01 3.85899425e-01 8.44592571e-01 -1.54084635e+00
-1.28980443e-01 -2.16266274e-01 -1.02351069e+00 -9.23389971e-01
2.24227577e-01 -1.04134095e+00 3.39148045e-01 -1.63636327e+00
4.44532424e-01 -7.84691989e-01 -5.34118116e-01 8.78831625e-01
9.85946432e-02 1.06787860e+00 3.38885427e-01 2.78116852e-01
-8.96289408e-01 3.03006858e-01 8.92147541e-01 -3.61528426e-01
5.33152103e-01 -8.69605303e-01 -6.29277706e-01 7.25923061e-01
5.32149315e-01 -5.89242458e-01 -1.75607294e-01 -6.50574267e-01
-2.19048792e-03 -4.28544194e-01 3.09884220e-01 -1.12866306e+00
1.68663755e-01 7.13585541e-02 7.25478947e-01 -2.37664342e-01
4.75224257e-01 -7.59071708e-01 -9.58344564e-02 2.53455013e-01
-6.61265075e-01 -1.37000442e-01 1.34451568e-01 6.45727932e-01
-3.95345986e-01 -3.90728563e-01 1.01081467e+00 -1.91792920e-01
-1.28594542e+00 4.80787516e-01 -1.55599520e-01 2.90504038e-01
1.01840341e+00 -1.69217214e-01 -6.98530018e-01 1.36037230e-01
-9.77408528e-01 9.64458436e-02 7.85317719e-01 6.09115243e-01
3.36470157e-01 -1.62151849e+00 -7.26871789e-01 3.92514199e-01
9.79604125e-01 -4.02606696e-01 -5.46320491e-02 1.91730291e-01
-2.47122645e-01 3.56086284e-01 -7.44220853e-01 -7.68016040e-01
-1.12893641e+00 7.02419639e-01 2.43883077e-02 -1.18293919e-01
-3.21616769e-01 1.01993608e+00 5.25964797e-01 -6.89577281e-01
1.94806963e-01 -3.30862015e-01 -2.96121776e-01 4.19687867e-01
5.03050566e-01 -7.87269138e-03 1.07819945e-01 -6.11953676e-01
-3.28842133e-01 8.94725025e-01 -2.73339957e-01 1.19446665e-01
1.54752338e+00 -5.79576381e-02 1.26037478e-01 3.91179442e-01
1.49562061e+00 -2.87598640e-01 -1.37055552e+00 -3.16629857e-01
-1.18839629e-01 -6.88978136e-01 -1.74789783e-02 -9.31281745e-01
-6.49030745e-01 1.07439518e+00 9.09886658e-01 2.96158433e-01
7.16881633e-01 3.71920854e-01 3.84872049e-01 5.81564605e-01
6.04316175e-01 -1.03995514e+00 4.32921767e-01 4.86621171e-01
5.97082794e-01 -1.44575024e+00 6.10486940e-02 -2.10455820e-01
-5.49821854e-01 1.11490870e+00 7.62985229e-01 -4.22959149e-01
7.87128806e-01 1.13029532e-01 6.75581694e-02 -1.92651927e-01
-7.33855247e-01 -4.53788817e-01 2.51556367e-01 7.59052753e-01
5.78419924e-01 2.02526659e-01 1.13801539e-01 3.51821631e-01
-1.05687141e-01 -9.87814963e-02 3.97775203e-01 7.73228884e-01
-7.88156569e-01 -1.11670589e+00 -1.92599464e-02 9.24249589e-01
-3.28979522e-01 -3.37122619e-01 1.29155321e-02 6.18622363e-01
3.66029739e-01 8.04730892e-01 4.27364200e-01 -3.08173716e-01
3.06696385e-01 6.63847998e-02 7.61550009e-01 -1.19860363e+00
-3.36865991e-01 -3.07907611e-01 2.87302732e-01 -4.43764746e-01
-5.03062487e-01 -2.90827066e-01 -1.06876314e+00 -8.75076056e-02
-3.87546539e-01 1.20191723e-01 7.04870462e-01 7.77680397e-01
3.46461982e-01 2.27718234e-01 4.60147709e-01 -1.13367045e+00
-6.63214922e-01 -1.03605938e+00 -7.21136987e-01 8.37696254e-01
3.11121851e-01 -1.00838292e+00 -4.21127319e-01 4.27836478e-01] | [9.83305549621582, 2.15319561958313] |
a185b880-2f12-436f-bfd7-675c1c298f5d | temporally-aware-feature-pooling-for-action | 2104.06779 | null | https://arxiv.org/abs/2104.06779v1 | https://arxiv.org/pdf/2104.06779v1.pdf | Temporally-Aware Feature Pooling for Action Spotting in Soccer Broadcasts | Toward the goal of automatic production for sports broadcasts, a paramount task consists in understanding the high-level semantic information of the game in play. For instance, recognizing and localizing the main actions of the game would allow producers to adapt and automatize the broadcast production, focusing on the important details of the game and maximizing the spectator engagement. In this paper, we focus our analysis on action spotting in soccer broadcast, which consists in temporally localizing the main actions in a soccer game. To that end, we propose a novel feature pooling method based on NetVLAD, dubbed NetVLAD++, that embeds temporally-aware knowledge. Different from previous pooling methods that consider the temporal context as a single set to pool from, we split the context before and after an action occurs. We argue that considering the contextual information around the action spot as a single entity leads to a sub-optimal learning for the pooling module. With NetVLAD++, we disentangle the context from the past and future frames and learn specific vocabularies of semantics for each subsets, avoiding to blend and blur such vocabulary in time. Injecting such prior knowledge creates more informative pooling modules and more discriminative pooled features, leading into a better understanding of the actions. We train and evaluate our methodology on the recent large-scale dataset SoccerNet-v2, reaching 53.4% Average-mAP for action spotting, a +12.7% improvement w.r.t the current state-of-the-art. | ['Bernard Ghanem', 'Silvio Giancola'] | 2021-04-14 | null | null | null | null | ['action-spotting'] | ['computer-vision'] | [ 5.04575903e-03 -7.89771527e-02 -2.91422158e-01 -2.71301478e-01
-8.00354660e-01 -6.30738735e-01 6.44224286e-01 1.81814596e-01
-7.05806434e-01 4.54033107e-01 9.48354483e-01 5.48218369e-01
-1.02078840e-01 -8.13066900e-01 -7.66704798e-01 -6.68113172e-01
-1.12190172e-02 9.74150822e-02 6.89033806e-01 -5.14824450e-01
5.95952235e-02 2.42945582e-01 -1.63828385e+00 9.04258251e-01
1.89017221e-01 1.11665320e+00 4.98517424e-01 5.57083786e-01
3.21204625e-02 1.11600065e+00 -8.01391244e-01 -4.77269232e-01
1.59877971e-01 -6.43812478e-01 -8.58940721e-01 1.32802978e-01
2.92413950e-01 -1.95879601e-02 -7.18113601e-01 6.73330188e-01
3.49803329e-01 5.29747069e-01 8.24446529e-02 -9.42911088e-01
-1.74794495e-01 9.71119463e-01 -2.88927704e-01 4.39336717e-01
3.85060161e-01 2.30570912e-01 1.34254742e+00 -5.91940224e-01
9.82329369e-01 1.00394082e+00 4.59794402e-01 4.45716590e-01
-1.15046203e+00 -5.51956773e-01 6.01309419e-01 6.84273303e-01
-1.44520485e+00 -3.35372329e-01 8.74137998e-01 -4.60065454e-01
8.49799037e-01 3.96237463e-01 1.21164727e+00 1.14498270e+00
3.82817507e-01 1.14280355e+00 7.89265275e-01 -1.53749466e-01
2.91529983e-01 -3.55986059e-01 1.70295492e-01 1.90623939e-01
-3.31564039e-01 -8.55230987e-02 -1.40430915e+00 4.09307122e-01
6.36200726e-01 8.30688924e-02 -2.23695919e-01 -2.16644779e-01
-1.38423836e+00 6.69606924e-01 3.61457884e-01 4.26656157e-01
-6.48051620e-01 4.28724200e-01 6.13556027e-01 1.01697437e-01
4.52666312e-01 4.85957831e-01 -4.36074048e-01 -6.16468847e-01
-1.03019536e+00 6.75762355e-01 5.40108800e-01 7.44414389e-01
6.69136882e-01 -3.03746134e-01 -7.43493259e-01 6.23992741e-01
-1.71266839e-01 1.65140126e-02 3.83018821e-01 -1.03217888e+00
5.63554108e-01 5.87190211e-01 5.88321313e-02 -8.92949998e-01
-5.20599306e-01 -5.16815782e-01 -2.08199054e-01 -1.35025099e-01
5.80213428e-01 -5.01691923e-02 -7.18376219e-01 1.92615151e+00
2.21059531e-01 6.02268577e-01 -3.11602931e-02 1.13041556e+00
6.95722222e-01 7.21233964e-01 1.90450922e-01 4.53265831e-02
1.79609966e+00 -1.04931808e+00 -7.74417937e-01 -5.01456857e-01
4.30705488e-01 -5.38444459e-01 8.89450490e-01 4.88884330e-01
-9.09726202e-01 -6.66916609e-01 -1.12793970e+00 -1.48936749e-01
-1.86150864e-01 -2.66354010e-02 7.12548673e-01 1.32417038e-01
-6.27072871e-01 7.25159645e-01 -1.04868007e+00 -3.00890267e-01
2.39306703e-01 7.06576258e-02 -4.51886952e-01 8.73302892e-02
-1.42727911e+00 7.40479469e-01 6.63893938e-01 8.60799924e-02
-1.00389922e+00 -8.48600447e-01 -8.59701991e-01 1.54407203e-01
9.27958012e-01 -2.88740844e-01 1.39714158e+00 -1.00709629e+00
-1.50502920e+00 6.77550495e-01 -5.49577512e-02 -7.39138424e-01
3.71554971e-01 -5.68271637e-01 -3.02741468e-01 1.82048813e-01
1.77659228e-01 6.23736799e-01 5.21452010e-01 -6.43619657e-01
-9.97906685e-01 -2.09424302e-01 5.19764423e-01 4.39917326e-01
-1.38256490e-01 1.81685418e-01 -1.01431835e+00 -9.07779276e-01
2.06360906e-01 -8.25410008e-01 -1.13875501e-01 -1.80780426e-01
8.40167850e-02 -1.24791376e-01 4.89908814e-01 -9.12651479e-01
1.51666820e+00 -2.28404546e+00 5.75861096e-01 -1.62899807e-01
1.03899188e-01 -1.21754311e-01 -1.20424619e-02 5.23908734e-01
1.38620779e-01 -3.09064716e-01 7.07518458e-02 -3.68211478e-01
8.76276642e-02 5.03406107e-01 -3.53364736e-01 4.38869238e-01
1.11682512e-01 9.93939757e-01 -1.11376524e+00 -3.33115846e-01
1.20740287e-01 3.32006425e-01 -8.96918833e-01 -7.52888620e-02
-2.88879812e-01 5.58541715e-01 -5.00574231e-01 3.38616163e-01
1.56691805e-01 8.97830203e-02 2.60236800e-01 -2.58900881e-01
-2.27539212e-01 4.55439568e-01 -1.34463823e+00 2.47728610e+00
-4.58184034e-01 7.24718869e-01 -1.27655104e-01 -1.04303741e+00
6.82035089e-01 3.14061403e-01 7.58254766e-01 -9.53791320e-01
6.80454150e-02 -1.49354264e-01 -1.01168916e-01 -5.14234185e-01
8.13232541e-01 -2.09183201e-01 -7.62151778e-01 1.97417885e-01
6.94096148e-01 6.48503453e-02 4.76950020e-01 2.68127203e-01
1.25286627e+00 5.23606122e-01 2.16016024e-01 -8.20512846e-02
2.45510206e-01 -6.24913462e-02 8.85895252e-01 7.76495516e-01
-2.62537301e-01 7.95050502e-01 6.43781781e-01 -5.72061956e-01
-5.60750484e-01 -9.47183132e-01 4.31761026e-01 1.45532966e+00
5.01572311e-01 -8.87971461e-01 -5.56458473e-01 -6.93930328e-01
-2.16301963e-01 5.97085774e-01 -8.40341091e-01 -2.96620756e-01
-1.01828098e+00 -5.76668918e-01 3.68998200e-01 6.94717109e-01
4.51066971e-01 -1.14994872e+00 -9.00368214e-01 5.65852523e-01
-6.51708841e-01 -1.34119058e+00 -7.57868946e-01 1.64233416e-01
-4.14407760e-01 -9.51451123e-01 -5.12408793e-01 -3.70788872e-01
-1.34622410e-01 -1.13172784e-01 1.30311346e+00 -3.81937504e-01
-1.40005782e-01 3.31881642e-01 -6.76777184e-01 -2.82016367e-01
1.99158266e-02 1.15875587e-01 -2.09803522e-01 3.23044837e-01
3.52696389e-01 -5.37493646e-01 -7.77414799e-01 3.49363565e-01
-7.05104530e-01 3.88308495e-01 4.70222741e-01 5.91348171e-01
6.79625094e-01 -1.22464381e-01 2.27440313e-01 -3.94617796e-01
2.11488962e-01 -4.35881704e-01 -3.38125676e-01 1.99204370e-01
1.73490331e-01 4.68951464e-02 2.02122122e-01 -6.22149885e-01
-1.09573424e+00 1.61809310e-01 -1.84706599e-01 -2.29550451e-01
4.85761426e-02 3.74166816e-01 -2.40433231e-01 4.44267452e-01
6.07875943e-01 4.30785447e-01 -4.35454011e-01 -5.24975419e-01
5.13550043e-01 1.41868293e-01 6.83584809e-01 -7.29382813e-01
3.15655679e-01 6.65267825e-01 -2.32428163e-01 -6.29009724e-01
-1.03634191e+00 -7.29396284e-01 -6.81952238e-01 -6.42806470e-01
1.19554186e+00 -1.10051644e+00 -8.20576608e-01 4.24872488e-01
-1.07928836e+00 -3.72345269e-01 -8.34521532e-01 6.24267399e-01
-8.91527355e-01 5.38920127e-02 -5.80537617e-01 -3.46989989e-01
1.81318954e-01 -1.17805982e+00 1.15977359e+00 2.40765482e-01
-3.90709102e-01 -5.72005332e-01 2.32224047e-01 4.30132836e-01
7.41007105e-02 1.86759323e-01 1.87274203e-01 -7.32199192e-01
-5.86490810e-01 -1.14202261e-01 1.10499382e-01 2.31568754e-01
-1.94055624e-02 -5.40344357e-01 -9.24728453e-01 3.54965664e-02
2.46180291e-03 -1.12078935e-01 1.00930309e+00 3.84063780e-01
9.64249074e-01 -1.15338273e-01 -7.25038797e-02 5.43360353e-01
1.18845379e+00 1.49591103e-01 6.55067086e-01 2.60048956e-01
5.34100175e-01 6.37401879e-01 8.95376563e-01 6.28232598e-01
3.92725170e-01 1.28466797e+00 3.10044408e-01 2.89236039e-01
-4.67439890e-01 -5.56529582e-01 5.77921391e-01 5.84839165e-01
-2.70189822e-01 -6.24657236e-02 -5.41288137e-01 7.79098392e-01
-2.23833084e+00 -1.18498778e+00 3.21852744e-01 1.85463810e+00
9.39756751e-01 2.35849783e-01 2.93056160e-01 -7.95205459e-02
6.15649343e-01 6.99909806e-01 -2.19093129e-01 -1.97897822e-01
-1.14862725e-01 4.84609038e-01 5.24030030e-01 3.34966004e-01
-1.35263884e+00 1.20688248e+00 5.37164497e+00 1.24307573e+00
-1.10097265e+00 4.58608031e-01 2.04878271e-01 -7.16102839e-01
2.58012451e-02 2.00791061e-01 -8.33188355e-01 3.82169932e-01
6.83093488e-01 -9.80930403e-02 3.48069847e-01 5.45017898e-01
3.64523560e-01 -2.07222521e-01 -1.19327080e+00 8.91129076e-01
1.69333950e-01 -1.57458472e+00 -2.97974572e-02 -2.54900813e-01
5.85335553e-01 -8.50522295e-02 -3.69418800e-01 6.29014015e-01
2.24238873e-01 -7.10815549e-01 1.53764987e+00 7.89205074e-01
2.80386448e-01 -7.25717604e-01 5.63471377e-01 2.26322651e-01
-1.54212534e+00 -2.39104494e-01 1.22147679e-01 -3.84389728e-01
5.91382563e-01 2.07724333e-01 -2.98102379e-01 7.76045799e-01
8.68124843e-01 9.23069596e-01 -3.83677691e-01 8.77851367e-01
-4.99058187e-01 5.28640687e-01 -1.98035359e-01 8.29471275e-02
3.66023362e-01 9.15605128e-02 6.99735820e-01 1.19797659e+00
1.79521561e-01 2.93846548e-01 4.55487669e-01 7.21857369e-01
1.18961716e-02 2.36776425e-03 -1.30542472e-01 -3.20616737e-02
1.52410530e-02 1.00138140e+00 -8.17781568e-01 -3.04215610e-01
-4.56348836e-01 1.08665156e+00 1.81314141e-01 3.08385432e-01
-1.16725969e+00 -2.05806553e-01 1.03060591e+00 2.06502557e-01
6.06163919e-01 -2.37120509e-01 -1.48840293e-01 -1.28619421e+00
1.76990479e-01 -7.02793241e-01 5.55065811e-01 -5.67448020e-01
-6.89162850e-01 3.89250785e-01 2.93496311e-01 -1.22642255e+00
-1.48642987e-01 -3.31927657e-01 -4.78178859e-01 6.07866764e-01
-1.04609907e+00 -1.21850610e+00 -1.59125030e-01 4.63283151e-01
1.09055400e+00 2.18621179e-01 3.78546774e-01 2.49508962e-01
-2.51390159e-01 3.47584188e-01 -4.92920995e-01 1.73010603e-01
5.86008966e-01 -1.00610411e+00 2.33500004e-01 1.02915478e+00
6.67512953e-01 1.59740746e-01 9.03378129e-01 -4.89183217e-01
-1.11958063e+00 -8.19809914e-01 9.55216169e-01 -4.85785723e-01
9.22195077e-01 -7.28647947e-01 -6.93217099e-01 7.34464407e-01
2.51172669e-02 3.36474814e-02 3.94304723e-01 1.81995496e-01
-1.65971354e-01 -3.76707911e-01 -6.41312063e-01 6.84599876e-01
1.37205029e+00 -6.13887072e-01 -8.47718060e-01 1.77641720e-01
8.16875339e-01 -5.84270477e-01 -8.27712357e-01 2.32444301e-01
7.34311044e-01 -9.00347233e-01 1.00663328e+00 -6.58618450e-01
4.38620180e-01 -3.51043016e-01 -3.27356517e-01 -1.11948621e+00
-2.03509986e-01 -5.85376680e-01 -6.05471171e-02 1.03745556e+00
1.53241366e-01 -5.95904961e-02 8.58934879e-01 3.45143467e-01
-3.99961531e-01 -5.58515370e-01 -1.23529077e+00 -6.42621040e-01
-2.39447445e-01 -9.20661151e-01 6.38577342e-01 6.14582002e-01
3.19166452e-01 -3.48844044e-02 -5.97245157e-01 7.32446611e-02
1.37598038e-01 2.20623568e-01 7.76124775e-01 -8.52637947e-01
-6.74530625e-01 -4.28884208e-01 -7.90337741e-01 -1.39435697e+00
1.35344803e-01 -8.01478088e-01 2.73056895e-01 -1.31610751e+00
2.61725605e-01 1.34922387e-02 -6.28399134e-01 6.82134748e-01
-3.31697613e-02 3.49455774e-01 5.58419049e-01 -7.64468610e-02
-1.00107217e+00 6.39246523e-01 1.30079412e+00 -3.41666490e-01
-2.69475758e-01 -3.58579345e-02 -6.15158141e-01 6.09068871e-01
5.03029644e-01 -4.96160209e-01 -3.41458857e-01 -4.41095233e-01
1.56891644e-01 2.24755362e-01 5.15590787e-01 -1.20784044e+00
3.34029138e-01 -3.81651074e-01 -3.78844105e-02 -3.32101047e-01
6.87039256e-01 -5.73759139e-01 2.70005435e-01 2.28782237e-01
-4.68228579e-01 -2.60456145e-01 2.97648579e-01 6.85011089e-01
-6.21503472e-01 -8.17559287e-03 3.47087532e-01 -1.87884197e-01
-1.40080583e+00 1.96125820e-01 -4.23699230e-01 5.95214888e-02
1.12086964e+00 -1.26643598e-01 -6.39035776e-02 -3.97985816e-01
-1.34647179e+00 1.91559151e-01 1.59952939e-01 8.99941146e-01
3.55199367e-01 -1.30789196e+00 -6.60476565e-01 -4.80278879e-02
2.90528983e-01 -2.72787333e-01 8.49941969e-01 9.22880411e-01
-3.52578878e-01 2.28476182e-01 -3.07600021e-01 -5.87160110e-01
-1.16947246e+00 3.32186580e-01 2.72099316e-01 -5.90458155e-01
-8.36089790e-01 9.32721913e-01 3.99457216e-01 1.00884371e-01
1.55004695e-01 -6.17732167e-01 -4.90196824e-01 3.99485588e-01
6.47262394e-01 2.42739320e-01 8.40899125e-02 -8.74894023e-01
-3.59473079e-01 4.17523950e-01 3.20644639e-02 -3.69726717e-01
1.35578740e+00 -3.47903520e-02 1.76791862e-01 6.08908296e-01
8.60788345e-01 5.34541756e-02 -1.71407521e+00 -3.93196911e-01
8.73454735e-02 -4.38067794e-01 4.35291156e-02 -7.03861773e-01
-1.25666904e+00 6.52404130e-01 3.37262183e-01 -9.73300412e-02
1.17882466e+00 4.10257220e-01 7.21803904e-01 6.57895207e-02
4.80426341e-01 -1.28420222e+00 1.66588798e-01 4.98837560e-01
1.01186335e+00 -8.46499026e-01 3.61584243e-03 -2.21223101e-01
-9.81612802e-01 8.62808764e-01 4.49055433e-01 -3.31626922e-01
5.60166061e-01 8.95601958e-02 -1.70973271e-01 -3.38640600e-01
-7.78375030e-01 -3.89193267e-01 5.86473286e-01 1.79368287e-01
5.34904040e-02 2.09783509e-01 -4.87152934e-01 1.27901840e+00
-1.49664730e-01 9.01064500e-02 3.14677000e-01 1.12581491e+00
-3.39725733e-01 -1.06718934e+00 -1.17338896e-01 -4.33856621e-03
-4.77061808e-01 8.59741569e-02 -1.31088525e-01 7.57594287e-01
6.47894800e-01 7.77352452e-01 2.73643792e-01 -5.40213287e-01
7.35637486e-01 1.92387238e-01 4.88799959e-01 -7.92879999e-01
-8.47481549e-01 7.42714927e-02 2.02028751e-01 -1.06090295e+00
-6.12987995e-01 -7.99676597e-01 -1.13689089e+00 8.99095163e-02
-3.67990583e-02 6.17434233e-02 3.81414145e-01 1.11623216e+00
1.86048076e-01 8.41833651e-01 2.00269669e-01 -1.10547602e+00
-1.04214676e-01 -8.24019074e-01 -6.56548738e-01 6.73570454e-01
1.02782786e-01 -8.44437957e-01 -2.30829138e-02 1.80626139e-01] | [8.033080101013184, 0.2271987646818161] |
9cf6da07-ff55-4ec6-9a60-13958d3261c7 | combining-word-embeddings-with-bilingual | null | null | https://aclanthology.org/2020.coling-main.531 | https://aclanthology.org/2020.coling-main.531.pdf | Combining Word Embeddings with Bilingual Orthography Embeddings for Bilingual Dictionary Induction | Bilingual dictionary induction (BDI) is the task of accurately translating words to the target language. It is of great importance in many low-resource scenarios where cross-lingual training data is not available. To perform BDI, bilingual word embeddings (BWEs) are often used due to their low bilingual training signal requirements. They achieve high performance, but problematic cases still remain, such as the translation of rare words or named entities, which often need to be transliterated. In this paper, we enrich BWE-based BDI with transliteration information by using Bilingual Orthography Embeddings (BOEs). BOEs represent source and target language transliteration word pairs with similar vectors. A key problem in our BDI setup is to decide which information source {--} BWEs (or semantics) vs. BOEs (or orthography) {--} is more reliable for a particular word pair. We propose a novel classification-based BDI system that uses BWEs, BOEs and a number of other features to make this decision. We test our system on English-Russian BDI and show improved performance. In addition, we show the effectiveness of our BOEs by successfully using them for transliteration mining based on cosine similarity. | ['Hinrich Sch{\\"u}tze', 'Alexander Fraser', 'Viktor Hangya', 'Silvia Severini'] | 2020-12-01 | null | null | null | coling-2020-8 | ['transliteration'] | ['natural-language-processing'] | [-1.42541528e-01 -5.79776287e-01 -5.18753707e-01 -3.67456853e-01
-5.49941301e-01 -6.54462099e-01 7.81965137e-01 2.76118428e-01
-7.11619854e-01 7.77417541e-01 5.00888526e-01 -8.84945214e-01
1.16492687e-02 -8.68714690e-01 -5.37118852e-01 -3.95534843e-01
4.21410888e-01 8.40971172e-01 -1.87323213e-01 -6.58472478e-01
-1.83548667e-02 4.81088132e-01 -1.16241121e+00 -3.04404963e-02
1.21679509e+00 3.83641034e-01 3.73819709e-01 1.09087005e-01
-4.24673319e-01 9.89121869e-02 -4.42028075e-01 -6.27893925e-01
3.63651514e-01 -5.49879551e-01 -8.22429359e-01 -3.18981647e-01
1.07755728e-01 9.96305048e-03 -2.60261506e-01 1.18014967e+00
6.13242686e-01 5.01744747e-02 8.26478422e-01 -8.48746002e-01
-8.93633902e-01 8.38791847e-01 -2.20206007e-01 3.68507147e-01
3.88210446e-01 -3.41088653e-01 1.36560321e+00 -1.41687584e+00
8.41695428e-01 1.12863541e+00 4.13354456e-01 3.07826847e-01
-1.08335400e+00 -9.49326277e-01 -3.39195989e-02 4.93352562e-01
-1.52565539e+00 -3.43052894e-01 4.56126004e-01 -3.23553562e-01
1.20649064e+00 2.02711180e-01 3.89786631e-01 1.10487998e+00
2.56463706e-01 6.29748583e-01 1.33093786e+00 -8.85125339e-01
-2.47973800e-01 5.89296818e-01 3.65693629e-01 4.45610732e-01
4.24815863e-01 2.35382050e-01 -5.34107864e-01 1.44194603e-01
3.51087362e-01 -6.04448374e-04 -3.60033721e-01 2.08278969e-01
-1.54025769e+00 1.09439290e+00 2.50436723e-01 7.75308967e-01
-1.48467347e-01 -5.06122231e-01 3.70850742e-01 7.31914997e-01
2.64666229e-01 6.90854192e-01 -4.89245683e-01 -4.47020642e-02
-6.52478874e-01 6.29613455e-03 7.12939680e-01 1.19261301e+00
9.06368971e-01 -1.81898996e-01 -2.94979960e-02 1.32575524e+00
1.32374197e-01 6.99889183e-01 1.05891883e+00 1.89241916e-01
6.77445292e-01 5.90784490e-01 -8.44529644e-02 -8.84858489e-01
-3.45759779e-01 -2.82426000e-01 -7.54696250e-01 -3.18870395e-01
8.69663283e-02 8.57547275e-04 -6.48041427e-01 1.56944489e+00
3.55703533e-01 -4.08724040e-01 2.34204695e-01 9.77177382e-01
8.90831411e-01 9.03976381e-01 -3.25246304e-01 -1.64694324e-01
1.42293930e+00 -8.85567844e-01 -7.61455774e-01 -2.80837059e-01
8.77619803e-01 -1.24379408e+00 1.25042856e+00 5.71978465e-02
-5.41514456e-01 -7.10579395e-01 -9.46380734e-01 -6.05088584e-02
-7.50756621e-01 4.09878224e-01 5.05017698e-01 6.66202545e-01
-6.53111219e-01 4.87545043e-01 -4.33352739e-01 -6.58278942e-01
-1.80457518e-01 4.44600672e-01 -7.03995287e-01 -3.76044929e-01
-1.62751901e+00 1.27427089e+00 4.33824748e-01 -7.28838593e-02
-4.36082870e-01 -3.63852918e-01 -1.09357440e+00 -1.38861924e-01
8.22898895e-02 -3.15853566e-01 6.43354475e-01 -8.25541317e-01
-1.36921918e+00 1.00190914e+00 -3.65631223e-01 -1.42488092e-01
8.15323517e-02 1.70223102e-01 -1.00136995e+00 -3.08094203e-01
3.59182715e-01 3.34136188e-01 6.37104332e-01 -9.01303589e-01
-7.83171773e-01 -4.10098046e-01 -3.87125164e-02 2.55455643e-01
-8.01840246e-01 4.72474992e-01 -4.81414706e-01 -8.59669030e-01
1.93070278e-01 -1.17861855e+00 1.17888056e-01 -5.94229698e-01
-1.97554976e-01 -5.52301407e-01 3.76162320e-01 -9.16247785e-01
1.48403358e+00 -1.89092815e+00 2.87047505e-01 2.36925021e-01
-7.63727948e-02 5.97407281e-01 -3.44683737e-01 6.05615377e-01
-1.13709323e-01 5.28938994e-02 4.30657342e-02 -4.88437079e-02
-3.19926813e-02 5.43636322e-01 -2.61062860e-01 4.41704035e-01
3.30260724e-01 8.39755177e-01 -9.04558539e-01 -4.47776109e-01
1.99317455e-01 1.82130665e-01 -3.17887187e-01 2.28020191e-01
3.13133091e-01 4.67688948e-01 -1.00988992e-01 6.36577129e-01
5.59014976e-01 2.30430961e-01 3.89317274e-01 -2.73733169e-01
-3.24227750e-01 7.65409946e-01 -1.11160219e+00 1.34358847e+00
-1.13413465e+00 6.10028148e-01 -4.37879741e-01 -1.18003869e+00
1.13942027e+00 2.03762054e-01 1.52304918e-01 -9.58554626e-01
2.57733256e-01 7.31964231e-01 3.18964899e-01 -2.67394066e-01
6.28179729e-01 -3.93728316e-01 -2.37453341e-01 6.82029366e-01
3.45296651e-01 -8.21750984e-02 5.61542153e-01 -1.41341940e-01
8.06163371e-01 -1.33867292e-02 6.45166695e-01 -5.28055847e-01
7.24914193e-01 1.69701591e-01 7.83251464e-01 2.84271955e-01
9.22613814e-02 4.06598806e-01 -2.85645910e-02 -4.36212122e-01
-9.43426907e-01 -8.76423180e-01 -3.71834129e-01 1.07807660e+00
3.10295820e-01 -6.36974156e-01 -2.36434400e-01 -8.12789321e-01
-1.28077418e-01 7.71196365e-01 -2.67579496e-01 -1.04513243e-01
-7.94840574e-01 -7.99931109e-01 3.23331088e-01 4.63709772e-01
3.33371721e-02 -9.82285678e-01 9.92893875e-02 4.05092865e-01
-5.20099580e-01 -1.03152227e+00 -8.82430673e-01 4.06396687e-01
-6.63004339e-01 -6.28890634e-01 -7.15356529e-01 -1.25777030e+00
7.65724301e-01 5.35799742e-01 1.20755482e+00 -1.93680733e-01
-8.92038941e-02 -1.19572289e-01 -7.66012371e-01 -3.41585934e-01
-6.41747713e-01 3.04723501e-01 6.16079628e-01 -1.12147786e-01
8.29207540e-01 -3.08378130e-01 -1.06651448e-01 7.55504549e-01
-8.99393380e-01 -1.34865776e-01 5.25123119e-01 1.17825985e+00
7.55298257e-01 -1.49508826e-02 4.34926897e-01 -8.51602793e-01
7.18489349e-01 -4.78028983e-01 -5.90040803e-01 5.09640813e-01
-7.33265221e-01 2.27081865e-01 1.00524080e+00 -6.38857245e-01
-6.58264637e-01 -3.46694976e-01 -5.03898442e-01 -2.08761469e-01
6.93907514e-02 8.34306359e-01 -2.40841180e-01 4.29295078e-02
6.88086212e-01 2.17436716e-01 -4.23335165e-01 -5.72258413e-01
2.72251129e-01 1.21998942e+00 1.42915085e-01 -6.00801408e-01
7.91416347e-01 7.11987447e-03 -3.73640954e-01 -7.76561141e-01
-5.35669029e-01 -7.74361312e-01 -7.41087019e-01 2.73290575e-01
5.58315873e-01 -1.01258588e+00 -1.38327092e-01 -9.60073397e-02
-1.12312806e+00 1.33117720e-01 -1.44081682e-01 9.81289268e-01
-1.12833954e-01 3.55403244e-01 -3.83643687e-01 -2.78966099e-01
-3.63945395e-01 -1.50042534e+00 7.38204360e-01 -1.44454151e-01
-5.49533546e-01 -1.06851494e+00 2.84906328e-01 2.02515915e-01
3.25495511e-01 -3.75112385e-01 1.24890232e+00 -9.21569645e-01
-1.21650435e-02 -9.29707885e-02 -1.53003797e-01 5.39751530e-01
5.64775348e-01 -2.71275789e-01 -5.42991102e-01 -3.39637130e-01
-2.21243516e-01 -1.60590917e-01 5.98895788e-01 -2.94236355e-02
4.98887897e-01 -1.92237884e-01 -1.43696368e-01 7.09933937e-01
1.32929158e+00 8.77954513e-02 2.57864118e-01 3.85733306e-01
6.82539940e-01 5.05971432e-01 8.17313492e-01 1.59411266e-01
4.81238484e-01 1.00566864e+00 -1.97676301e-01 -1.31932623e-03
-3.78751934e-01 -3.37407142e-01 8.26680839e-01 1.81915128e+00
1.05539709e-01 -1.83809802e-01 -1.04096878e+00 9.01526988e-01
-1.42787015e+00 -5.57450056e-01 -2.84277767e-01 2.48136497e+00
1.21867740e+00 -2.23178163e-01 -9.20535251e-02 1.89419687e-01
6.16960526e-01 -1.75057411e-01 -3.92474383e-02 -7.11124599e-01
-3.14977765e-01 7.88777173e-01 5.21310091e-01 6.37827218e-01
-8.14929724e-01 1.38052380e+00 4.79516983e+00 9.93881822e-01
-1.26638556e+00 4.21022445e-01 -3.38805187e-03 4.97854084e-01
-6.68390572e-01 9.40880552e-02 -1.35367417e+00 2.85614640e-01
7.84828603e-01 -2.61871815e-01 6.01956964e-01 4.79363501e-01
-2.95419544e-02 1.47205681e-01 -1.19638085e+00 1.17479706e+00
3.38900298e-01 -1.06576192e+00 2.48166159e-01 -1.01838164e-01
9.00015056e-01 8.04324746e-02 -1.73418388e-01 5.90696454e-01
3.66040081e-01 -9.22595978e-01 5.88095307e-01 -1.85234785e-01
1.23276627e+00 -7.13692009e-01 8.51489186e-01 2.12772503e-01
-1.16087937e+00 2.43842050e-01 -7.88354039e-01 1.12207390e-01
-1.22292219e-02 6.59860969e-01 -8.95442665e-01 8.93386006e-01
4.86046731e-01 8.21003556e-01 -3.29338729e-01 6.41878724e-01
-6.03230655e-01 6.46933079e-01 -3.57743740e-01 -2.08208919e-01
2.85038799e-01 -7.30589092e-01 4.03830826e-01 1.38298452e+00
6.95780456e-01 -4.00263518e-01 9.23016593e-02 5.14925003e-01
-2.95095235e-01 8.31694841e-01 -7.50596762e-01 -3.02060723e-01
5.19089818e-01 1.12162793e+00 -5.72778940e-01 -3.43146026e-01
-6.87163472e-01 1.31399035e+00 3.34734410e-01 1.94111109e-01
-6.65285945e-01 -5.85977614e-01 9.90253747e-01 -1.55100510e-01
2.85990119e-01 -3.81536007e-01 -6.47647828e-02 -1.54684532e+00
1.42786235e-01 -1.25760174e+00 3.36746663e-01 -2.24356934e-01
-1.43442619e+00 8.69077682e-01 -2.44127572e-01 -1.55224228e+00
-1.31571159e-01 -8.28806818e-01 -3.84150773e-01 1.18922603e+00
-1.78029954e+00 -1.10255814e+00 1.26100248e-02 7.41499722e-01
6.88615501e-01 -4.09701437e-01 1.12219441e+00 7.51961291e-01
-6.19880259e-01 1.02444744e+00 4.36508775e-01 1.98686138e-01
1.10106790e+00 -1.06983411e+00 4.95241255e-01 8.10285211e-01
9.15518463e-01 9.07534003e-01 4.10309255e-01 -5.93412161e-01
-1.43884265e+00 -1.26088464e+00 1.97990823e+00 -3.27080399e-01
8.14207971e-01 -4.07323629e-01 -6.36351943e-01 6.89236403e-01
1.00437388e-01 -1.05295874e-01 9.85667586e-01 4.79977995e-01
-4.36712444e-01 -2.02282920e-01 -6.83192313e-01 8.82423222e-01
1.13430142e+00 -6.29225731e-01 -7.81750441e-01 6.03719413e-01
5.83849430e-01 -1.43569425e-01 -8.30541909e-01 2.56336570e-01
3.80697608e-01 -4.41694260e-01 8.64777684e-01 -7.77194858e-01
3.58440906e-01 -4.16601330e-01 -2.96301991e-01 -1.93205750e+00
-2.31238052e-01 -3.58294487e-01 4.14886445e-01 1.20450521e+00
7.00243652e-01 -7.11815536e-01 6.54467335e-03 -1.67143777e-01
-1.14990950e-01 -4.68986988e-01 -9.79112625e-01 -1.26265609e+00
2.33614042e-01 -3.83709073e-01 8.18973839e-01 1.23265517e+00
2.01950476e-01 7.48683751e-01 -3.93622130e-01 -2.32826825e-03
-3.74469794e-02 5.32667816e-01 7.47010946e-01 -8.85305703e-01
-2.55951434e-01 -3.51487875e-01 -5.14366388e-01 -1.08621037e+00
4.92017478e-01 -1.63417804e+00 -6.82118461e-02 -1.45940399e+00
-6.83936030e-02 -5.92323661e-01 -2.99480319e-01 3.93866181e-01
-3.12681705e-01 3.20329547e-01 -5.25789000e-02 1.93031475e-01
-5.09026460e-02 7.10707724e-01 9.41921711e-01 -3.38479996e-01
-3.71158391e-01 -4.67728749e-02 -4.82878834e-01 4.13849115e-01
7.81886280e-01 -7.28363216e-01 -1.24095723e-01 -8.11848819e-01
2.37788141e-01 -2.71423280e-01 -4.03276891e-01 -6.53034627e-01
-1.37843732e-02 -1.09421991e-01 5.52009381e-02 -2.52227724e-01
9.64896288e-03 -9.07309115e-01 1.05604436e-02 3.20400655e-01
-3.82779241e-02 5.77495575e-01 3.66423116e-03 1.15429737e-01
-6.15400553e-01 -4.81116831e-01 6.47378862e-01 7.60433152e-02
-6.56163692e-01 3.38527113e-01 -4.51739401e-01 6.38610795e-02
7.74873972e-01 1.00708172e-01 2.22452000e-01 -5.42676710e-02
-4.06057686e-01 -6.40307069e-02 4.43072200e-01 7.72709966e-01
5.20586967e-01 -1.76930130e+00 -8.65650713e-01 6.16179824e-01
7.79282093e-01 -5.67954183e-01 -4.20270860e-01 9.05165851e-01
-5.77959120e-01 4.13045198e-01 -1.95616990e-01 -2.11824432e-01
-1.31857324e+00 5.04609466e-01 -9.22103077e-02 -4.22328949e-01
-4.61318493e-01 7.23825514e-01 -3.38557698e-02 -1.00563514e+00
-2.62646517e-03 -3.65030557e-01 -2.92395681e-01 3.53042901e-01
3.74126047e-01 1.31855682e-01 5.15368104e-01 -1.10167122e+00
-4.84895915e-01 6.06616020e-01 -2.34610707e-01 -2.06368357e-01
1.20056129e+00 -1.51652068e-01 -3.28206003e-01 3.66796732e-01
1.39321423e+00 4.06644136e-01 -2.41049975e-02 -8.04077566e-01
1.68240786e-01 -7.16523290e-01 -9.66021642e-02 -4.39589441e-01
-8.26436162e-01 1.12132609e+00 4.71927613e-01 -3.82775575e-01
8.97018552e-01 -1.19232297e-01 1.10900486e+00 4.16274071e-01
5.22396684e-01 -1.37757468e+00 -4.67227131e-01 8.56603205e-01
7.13009596e-01 -1.46705377e+00 -1.91621974e-01 -1.57817423e-01
-6.14191473e-01 1.25267780e+00 2.24981934e-01 1.69855505e-01
7.06193209e-01 -9.78224874e-02 3.44338268e-01 6.41555786e-02
-8.77599344e-02 -5.44774950e-01 4.21285421e-01 4.42442507e-01
6.02568686e-01 3.50473702e-01 -1.08780754e+00 6.24201000e-01
-4.62103963e-01 -4.99509007e-01 2.31159657e-01 5.98414361e-01
-7.22469985e-02 -1.82229328e+00 -3.45788181e-01 4.29871082e-01
-2.78886467e-01 -5.33721685e-01 -2.61168510e-01 5.97739398e-01
2.50921965e-01 1.09043741e+00 -1.87150940e-01 -6.06567860e-01
3.38052541e-01 1.12624615e-01 4.42454219e-01 -8.34426522e-01
-6.77264094e-01 -2.13940144e-01 8.12824592e-02 -3.62720817e-01
-2.32576087e-01 -5.35143018e-01 -9.24515188e-01 -4.76875484e-01
-5.69424391e-01 2.73380131e-01 8.43898237e-01 1.02513802e+00
7.29015991e-02 1.69183657e-01 1.02799511e+00 -2.69631326e-01
-5.60190856e-01 -9.77311730e-01 -5.78611434e-01 6.63877904e-01
-7.24293664e-02 -5.75766265e-01 -1.01502173e-01 -1.67929083e-01] | [11.04175853729248, 10.025375366210938] |
5df4db35-558c-4d00-9bae-8eca3cc45718 | fast-underwater-image-enhancement-for | 1903.09766 | null | https://arxiv.org/abs/1903.09766v3 | https://arxiv.org/pdf/1903.09766v3.pdf | Fast Underwater Image Enhancement for Improved Visual Perception | In this paper, we present a conditional generative adversarial network-based model for real-time underwater image enhancement. To supervise the adversarial training, we formulate an objective function that evaluates the perceptual image quality based on its global content, color, local texture, and style information. We also present EUVP, a large-scale dataset of a paired and unpaired collection of underwater images (of `poor' and `good' quality) that are captured using seven different cameras over various visibility conditions during oceanic explorations and human-robot collaborative experiments. In addition, we perform several qualitative and quantitative evaluations which suggest that the proposed model can learn to enhance underwater image quality from both paired and unpaired training. More importantly, the enhanced images provide improved performances of standard models for underwater object detection, human pose estimation, and saliency prediction. These results validate that it is suitable for real-time preprocessing in the autonomy pipeline by visually-guided underwater robots. The model and associated training pipelines are available at https://github.com/xahidbuffon/funie-gan. | ['Md Jahidul Islam', 'Youya Xia', 'Junaed Sattar'] | 2019-03-23 | null | null | null | null | ['underwater-image-restoration'] | ['computer-vision'] | [ 3.01695198e-01 3.54336321e-01 9.32911992e-01 -5.75617015e-01
-7.72683680e-01 -6.28323734e-01 2.90298164e-01 -1.84104949e-01
-9.28155839e-01 5.44992447e-01 5.48954271e-02 1.17621504e-01
7.19671398e-02 -9.53855097e-01 -1.16697311e+00 -8.37047577e-01
-4.62472230e-01 8.11602250e-02 2.74839699e-01 -5.15562236e-01
1.35682896e-01 2.81434447e-01 -1.73490751e+00 1.95868742e-02
1.00973940e+00 8.64478469e-01 5.82425952e-01 1.16094553e+00
7.03799546e-01 8.30426276e-01 -5.42663872e-01 -5.86235166e-01
6.15092158e-01 -4.06553239e-01 -5.02626300e-01 -2.24142805e-01
7.73277521e-01 -8.03870976e-01 -4.55302835e-01 1.53189540e+00
8.81593466e-01 3.54490459e-01 3.88301134e-01 -1.08194327e+00
-7.36106277e-01 4.08140004e-01 -1.21818319e-01 1.33049920e-01
2.37749383e-01 3.66739243e-01 7.26431668e-01 -9.16387320e-01
5.35248339e-01 1.21436977e+00 6.75307691e-01 6.74882114e-01
-6.76791430e-01 -5.83194792e-01 -1.83581263e-01 2.65617102e-01
-8.21995497e-01 -5.56745112e-01 5.13777554e-01 -2.75476128e-01
4.13549811e-01 2.21234798e-01 8.71553361e-01 8.16808760e-01
2.84347922e-01 8.09572875e-01 1.21914113e+00 -2.51252323e-01
3.06323081e-01 -8.89324322e-02 -6.27681017e-01 9.10748243e-01
2.76396573e-02 5.24203181e-01 -6.68455124e-01 1.67010903e-01
7.16556907e-01 -1.42565519e-01 -6.94524586e-01 -2.63624359e-02
-7.99306631e-01 7.47418523e-01 7.61975408e-01 -4.61766630e-01
-1.59158424e-01 1.68283001e-01 5.46761565e-02 4.26522642e-01
3.51272643e-01 6.30584419e-01 -3.96386117e-01 8.81114081e-02
-5.78672051e-01 1.23682186e-01 6.29190564e-01 8.74987006e-01
9.40088212e-01 4.04843092e-01 2.60933816e-01 7.76848316e-01
6.77998662e-01 1.26009285e+00 3.13383192e-01 -1.34452546e+00
-2.92619057e-02 6.52919114e-02 1.91849083e-01 -9.11891282e-01
-4.61651444e-01 2.44042221e-02 -5.12197375e-01 9.10781264e-01
1.18146114e-01 -5.26214719e-01 -9.77397025e-01 1.49415576e+00
7.67566487e-02 1.32288367e-01 6.49457693e-01 1.20607102e+00
1.25444961e+00 7.11359799e-01 -1.86867893e-01 2.75696546e-01
1.00164652e+00 -1.08261025e+00 -6.03219151e-01 -6.15316153e-01
1.06715122e-02 -5.57019174e-01 7.63344169e-01 2.43822739e-01
-1.18428695e+00 -4.57778871e-01 -1.12501204e+00 -1.11705385e-01
-2.00075418e-01 2.93410085e-02 4.24664676e-01 4.59665358e-01
-1.39455771e+00 5.19796014e-01 -8.84094417e-01 -1.17538936e-01
2.17722446e-01 1.05006486e-01 -5.34686267e-01 -2.75026143e-01
-1.22914970e+00 9.52553868e-01 1.01444572e-01 5.17900646e-01
-1.71348464e+00 -4.78254884e-01 -1.31627440e+00 -2.63710529e-01
-2.17726916e-01 -2.88305253e-01 1.35889280e+00 -9.48072672e-01
-1.52282894e+00 7.43127465e-01 3.62761557e-01 -4.81362134e-01
5.50423682e-01 -5.41802883e-01 -9.07868370e-02 5.76177299e-01
8.85622650e-02 9.38924909e-01 7.77980328e-01 -1.63933790e+00
-8.23522806e-01 -3.21670979e-01 2.59408474e-01 6.80510104e-01
-6.94484264e-02 -2.70579699e-02 -5.36195517e-01 -4.99979883e-01
4.85216416e-02 -8.36426318e-01 -2.99693435e-01 5.75540304e-01
-1.64315000e-01 6.82208121e-01 7.94497788e-01 -8.99054885e-01
4.91221063e-02 -2.05088305e+00 3.24476689e-01 -1.66026860e-01
-1.09988384e-01 1.92518622e-01 -3.84565920e-01 2.15662062e-01
3.83932680e-01 -2.07897067e-01 -5.28955102e-01 -6.21180832e-01
-8.30409378e-02 3.70121241e-01 3.13171856e-02 8.49512458e-01
8.63552690e-02 7.76217937e-01 -1.20869720e+00 -3.38312805e-01
3.83923143e-01 3.19500118e-01 -5.66113412e-01 8.00938368e-01
1.34042919e-01 7.45329618e-01 4.80471589e-02 9.04568791e-01
9.15385664e-01 5.73025048e-01 -1.50257185e-01 -2.45104030e-01
-2.18091771e-01 -4.61912185e-01 -8.26819539e-01 1.55772352e+00
-4.47481036e-01 1.11096692e+00 7.49283671e-01 -4.52110946e-01
1.00827146e+00 -8.55808035e-02 -8.94289836e-03 -6.87415779e-01
3.24656069e-01 1.01754032e-01 -3.19274485e-01 -8.19936216e-01
8.60348046e-01 -2.51639724e-01 1.00596182e-01 1.80072971e-02
2.92996466e-01 -5.90074062e-01 8.57556984e-03 1.91433042e-01
8.38646948e-01 2.31025010e-01 -3.30046892e-01 -2.88305461e-01
8.27827305e-02 -2.90661484e-01 5.40775716e-01 7.27306724e-01
-3.14618826e-01 9.57172513e-01 -1.35612279e-01 -1.97847188e-01
-1.07569349e+00 -1.04038322e+00 -1.81014165e-01 1.26875937e+00
8.78919363e-01 3.98938477e-01 -6.50404274e-01 -1.83987737e-01
-2.47993946e-01 2.99744993e-01 -9.81597781e-01 -2.28370667e-01
-1.83798507e-01 -7.33844101e-01 7.92127311e-01 4.91493225e-01
6.24694467e-01 -1.40385556e+00 -8.75905991e-01 -4.83017415e-02
-2.38004446e-01 -1.08213961e+00 -2.29533643e-01 3.43734294e-01
-5.57946861e-01 -1.03641653e+00 -8.52233052e-01 -1.04443586e+00
7.90355444e-01 2.47816756e-01 9.16690469e-01 1.66717350e-01
-1.00558341e-01 4.25441891e-01 -9.07910407e-01 -5.69717467e-01
-5.10075212e-01 -8.51893246e-01 1.30876541e-01 -1.98069021e-01
-3.55781287e-01 -3.88749093e-01 -8.68841112e-01 4.58746165e-01
-1.17807460e+00 -5.00886887e-02 6.42050564e-01 9.47399974e-01
2.52466172e-01 -1.92967862e-01 1.19484261e-01 -2.19220713e-01
1.11205555e-01 -3.44003052e-01 -6.93594277e-01 1.67981070e-02
-1.08508594e-01 -2.77827531e-01 2.34730512e-01 -5.80973737e-02
-1.15532291e+00 1.21505946e-01 -4.95824635e-01 -3.08590561e-01
-9.24825221e-02 4.93216246e-01 -1.14531524e-01 -6.33302927e-01
6.85614109e-01 4.30618703e-01 2.90204316e-01 -2.21211150e-01
1.00670122e-01 5.89285553e-01 9.69774902e-01 -2.22611994e-01
1.03555393e+00 8.32474768e-01 -4.22143728e-01 -9.95144546e-01
-6.34999037e-01 -3.05091351e-01 -4.02249873e-01 -6.93595231e-01
1.00162518e+00 -1.41530037e+00 -5.53777158e-01 1.07754016e+00
-9.03013408e-01 -9.82043207e-01 -4.71697636e-02 4.33413774e-01
-4.37456578e-01 5.75025141e-01 -8.94781888e-01 -9.28410232e-01
-5.99840045e-01 -1.25146925e+00 1.27730858e+00 7.48177350e-01
6.69124126e-01 -1.00734556e+00 9.60993841e-02 3.04020226e-01
4.24374819e-01 3.86021942e-01 -2.48903662e-01 -3.48362066e-02
-4.21936929e-01 5.55066690e-02 -1.70396015e-01 7.08998203e-01
-1.60104215e-01 1.43710360e-01 -1.00981545e+00 -6.73733532e-01
-2.44612992e-01 -7.16753244e-01 9.49977338e-01 3.07551056e-01
5.49956381e-01 -2.72493482e-01 3.55829060e-01 1.10222960e+00
1.43174636e+00 6.26891106e-03 8.63709033e-01 6.26360714e-01
6.62401736e-01 7.59519279e-01 8.90526295e-01 5.84804773e-01
5.19224882e-01 2.72307605e-01 1.31057811e+00 -5.53153813e-01
8.99487212e-02 -1.66211948e-01 7.24525273e-01 3.37010235e-01
-3.42247665e-01 -4.22730774e-01 -5.93692482e-01 7.13495314e-01
-1.43117619e+00 -8.60705614e-01 -8.79995674e-02 1.81022000e+00
6.49761975e-01 -3.76186609e-01 -3.07493657e-01 -2.98846632e-01
6.42658710e-01 4.71474370e-04 -3.65226746e-01 -9.45661142e-02
-5.36202252e-01 -1.79461211e-01 9.64965701e-01 6.26404405e-01
-1.26473057e+00 9.56015050e-01 5.75133324e+00 1.65884063e-01
-9.37231719e-01 -1.19935550e-01 4.59853321e-01 3.54053825e-01
-2.59939313e-01 -3.48862708e-01 -4.80805904e-01 3.93348426e-01
6.77817225e-01 2.27998927e-01 3.23575735e-01 9.52274084e-01
2.75503844e-01 -2.72108495e-01 -6.52130961e-01 5.44905663e-01
3.55103195e-01 -1.06234491e+00 -1.37725174e-01 -1.35421574e-01
1.03042042e+00 5.58147311e-01 1.31940171e-01 9.57842171e-02
6.44338012e-01 -7.84769952e-01 1.29745924e+00 6.62149429e-01
8.60606551e-01 -5.51153064e-01 1.31436503e+00 7.99197555e-02
-8.62346649e-01 -1.18644901e-01 -5.95606744e-01 -8.93632881e-03
3.45985711e-01 -5.38013503e-02 -4.28802550e-01 3.82618010e-01
1.47066653e+00 7.42649436e-01 -7.41070807e-01 1.30547202e+00
-6.41271234e-01 3.77379894e-01 -3.31669331e-01 3.43027450e-02
1.54566556e-01 -2.60990381e-01 6.21018946e-01 1.24089563e+00
5.33432186e-01 3.20440978e-01 -1.60493478e-01 4.12143350e-01
-9.09770578e-02 -1.41855165e-01 -4.90897894e-01 5.78360140e-01
4.23005521e-01 1.21708775e+00 -4.21781361e-01 -1.00742273e-01
-1.91416070e-01 1.08704507e+00 -6.20583221e-02 2.35087886e-01
-8.54979455e-01 -3.90586883e-01 5.56241632e-01 -5.94832122e-01
2.66034395e-01 -1.91907696e-02 -1.87066719e-02 -1.06638408e+00
-3.24171513e-01 -8.31331849e-01 5.50711751e-02 -1.28735900e+00
-1.15239298e+00 8.74785244e-01 -2.04890564e-01 -1.57593203e+00
2.63432354e-01 -6.89095080e-01 -7.05555081e-01 5.86484373e-01
-1.80977952e+00 -1.19530749e+00 -1.04782093e+00 4.16896969e-01
5.36079824e-01 -2.02472564e-02 6.73914850e-01 2.70896882e-01
-2.17337117e-01 4.95282620e-01 3.50152642e-01 4.52993810e-01
6.93237722e-01 -1.48760903e+00 1.58488065e-01 1.37423122e+00
-3.72776389e-01 8.70810449e-02 1.22202480e+00 -4.01847601e-01
-1.44832325e+00 -1.27924573e+00 -7.25962594e-02 -1.18786298e-01
5.14436841e-01 1.60421818e-01 -8.85696232e-01 5.55861175e-01
5.00243664e-01 2.03592423e-02 5.26157737e-01 -7.57062912e-01
2.52387941e-01 1.22802174e-02 -1.15146542e+00 3.12321514e-01
5.71928024e-01 -7.05543682e-02 -4.34935570e-01 2.07367897e-01
6.80469871e-01 -7.99697459e-01 -9.85655069e-01 6.68830574e-01
5.30199111e-01 -1.25842559e+00 8.54034305e-01 1.08772077e-01
9.06165898e-01 -5.97135425e-01 -3.58384222e-01 -1.54085159e+00
2.01664753e-02 -2.50123173e-01 4.99948293e-01 7.69872189e-01
5.32200098e-01 -3.96044016e-01 5.05883932e-01 1.96830675e-01
-8.16118777e-01 -1.56974375e-01 -8.35598350e-01 -4.16814357e-01
-3.39698605e-02 -1.32539272e-01 1.43401250e-01 6.40812099e-01
-2.34400168e-01 -1.06128104e-01 -9.04366493e-01 1.01360965e+00
1.02145720e+00 6.79610595e-02 9.02114511e-01 -9.09047842e-01
-3.52185845e-01 -1.54970428e-02 -6.19905472e-01 -1.01248217e+00
-1.72998279e-01 -3.14483345e-01 1.10719717e+00 -1.70440042e+00
2.68420409e-02 -8.58588144e-02 3.59858871e-02 6.26399755e-01
-2.59958059e-01 1.01163030e+00 4.12430651e-02 9.29930955e-02
-6.57992005e-01 1.01909292e+00 1.32751739e+00 -1.88239425e-01
1.85855880e-01 -1.26818329e-01 -3.33107293e-01 6.53746784e-01
8.83134365e-01 -2.19304144e-01 1.46766320e-01 -8.83115768e-01
1.70322195e-01 1.60001174e-01 6.03757501e-01 -1.18106198e+00
3.58219326e-01 4.99811508e-02 2.43061036e-01 -9.50139239e-02
5.04579306e-01 -5.51761687e-01 -6.88505396e-02 5.78919828e-01
-2.04345256e-01 -1.81616649e-01 2.83711523e-01 7.33908415e-01
-3.75581235e-01 -4.41880316e-01 1.26471412e+00 -1.60890177e-01
-1.33195853e+00 1.32679403e-01 -4.10125375e-01 -1.17760152e-01
7.98761129e-01 -6.70793802e-02 -6.50577307e-01 -7.45426595e-01
-6.12708688e-01 4.13625002e-01 8.69206786e-01 3.13959301e-01
1.17938662e+00 -5.60554326e-01 -1.08982015e+00 9.11964476e-02
3.02626163e-01 2.75492072e-01 5.98504305e-01 6.26873612e-01
-1.41777325e+00 -7.47235179e-01 -7.00305164e-01 -4.25927103e-01
-1.30872607e+00 -9.11795571e-02 5.86725593e-01 4.58127707e-01
-5.02147973e-01 1.41230738e+00 1.89347386e-01 -6.79905415e-01
1.24300979e-01 7.30048027e-03 -3.72399628e-01 -3.23967189e-01
6.27049088e-01 4.36725199e-01 -2.26081267e-01 -8.92768025e-01
-1.27912164e-01 4.40569550e-01 5.29466927e-01 -2.45096371e-01
1.69094563e+00 -5.06358981e-01 -2.30493993e-01 -6.80553392e-02
9.47114229e-01 8.79996121e-02 -2.00673103e+00 3.68730398e-03
-8.20215821e-01 -4.96998817e-01 1.90277994e-01 -7.43615210e-01
-1.41809666e+00 6.98378563e-01 9.64328825e-01 -5.34542911e-02
1.24640048e+00 9.34784114e-02 4.93260473e-01 4.03403163e-01
4.48126376e-01 -8.55018556e-01 1.74765199e-01 6.80456042e-01
1.08013296e+00 -1.62932110e+00 -1.09344712e-02 -8.17476884e-02
-9.14247513e-01 8.95907283e-01 7.51141369e-01 -1.09294251e-01
2.47482672e-01 4.96167898e-01 8.30462694e-01 -1.13658093e-01
-2.31383473e-01 -3.01186174e-01 3.64585109e-02 8.14354241e-01
-2.46977717e-01 -9.18869376e-02 2.86000103e-01 2.37140775e-01
-4.06172812e-01 -6.76557243e-01 1.08543193e+00 8.71206403e-01
-7.82245338e-01 -3.91325414e-01 -4.99991983e-01 -7.60041759e-04
-4.16074574e-01 -3.04959387e-01 1.16812168e-02 3.85390967e-01
7.40749463e-02 1.18491006e+00 -1.58406347e-01 -6.08607769e-01
2.92619795e-01 -8.68034959e-01 1.80195555e-01 -4.14765120e-01
-2.21075639e-01 -1.36017412e-01 1.45168584e-02 -3.71043622e-01
-8.09550285e-01 -6.08507395e-01 -1.17073607e+00 3.95611376e-02
-2.77631342e-01 4.59008336e-01 7.77794600e-01 4.89291787e-01
-1.33010641e-01 4.62041080e-01 7.77721763e-01 -1.78096473e+00
-1.14767201e-01 -1.39464831e+00 -7.50358641e-01 2.79156536e-01
4.70561028e-01 -5.57986438e-01 -9.86309469e-01 5.02694309e-01] | [10.689448356628418, -3.5447585582733154] |
eb62696e-076c-4974-940f-60fdab205c08 | cross-modal-progressive-comprehension-for | 2105.07175 | null | https://arxiv.org/abs/2105.07175v1 | https://arxiv.org/pdf/2105.07175v1.pdf | Cross-Modal Progressive Comprehension for Referring Segmentation | Given a natural language expression and an image/video, the goal of referring segmentation is to produce the pixel-level masks of the entities described by the subject of the expression. Previous approaches tackle this problem by implicit feature interaction and fusion between visual and linguistic modalities in a one-stage manner. However, human tends to solve the referring problem in a progressive manner based on informative words in the expression, i.e., first roughly locating candidate entities and then distinguishing the target one. In this paper, we propose a Cross-Modal Progressive Comprehension (CMPC) scheme to effectively mimic human behaviors and implement it as a CMPC-I (Image) module and a CMPC-V (Video) module to improve referring image and video segmentation models. For image data, our CMPC-I module first employs entity and attribute words to perceive all the related entities that might be considered by the expression. Then, the relational words are adopted to highlight the target entity as well as suppress other irrelevant ones by spatial graph reasoning. For video data, our CMPC-V module further exploits action words based on CMPC-I to highlight the correct entity matched with the action cues by temporal graph reasoning. In addition to the CMPC, we also introduce a simple yet effective Text-Guided Feature Exchange (TGFE) module to integrate the reasoned multimodal features corresponding to different levels in the visual backbone under the guidance of textual information. In this way, multi-level features can communicate with each other and be mutually refined based on the textual context. Combining CMPC-I or CMPC-V with TGFE can form our image or video version referring segmentation frameworks and our frameworks achieve new state-of-the-art performances on four referring image segmentation benchmarks and three referring video segmentation benchmarks respectively. | ['Guanbin Li', 'Bo Li', 'Yunchao Wei', 'Shaofei Huang', 'Tianrui Hui', 'Si Liu'] | 2021-05-15 | null | null | null | null | ['referring-expression-segmentation'] | ['computer-vision'] | [ 4.07391310e-01 2.14917898e-01 -2.04765081e-01 -2.90031314e-01
-6.87815607e-01 -4.89912152e-01 6.16469085e-01 1.46679476e-01
-3.83723706e-01 3.15702736e-01 1.89443842e-01 -6.73256069e-03
-5.40825315e-02 -7.53700614e-01 -6.22328758e-01 -6.63161576e-01
3.97024989e-01 1.52471349e-01 6.29052043e-01 -3.02321762e-01
3.26839477e-01 3.45212907e-01 -1.49830997e+00 5.75325251e-01
8.90824974e-01 9.46117997e-01 3.94891232e-01 3.65401387e-01
-5.48004508e-01 9.78913784e-01 -3.41230571e-01 -4.69819933e-01
-5.99765889e-02 -5.05063891e-01 -9.47247624e-01 5.08299112e-01
1.63961962e-01 -1.45922661e-01 -1.89455569e-01 1.31215501e+00
2.10727394e-01 3.50265086e-01 5.03699124e-01 -1.47203386e+00
-7.91753829e-01 6.48593485e-01 -8.76262724e-01 9.68795866e-02
8.67377877e-01 1.80101454e-01 9.21112657e-01 -9.42306936e-01
8.96193445e-01 1.48303986e+00 6.48761466e-02 4.92822140e-01
-1.03832829e+00 -4.58147764e-01 8.98017943e-01 4.09140944e-01
-1.51536286e+00 -2.43550539e-01 1.14142811e+00 -2.61675566e-01
6.42010093e-01 4.14271235e-01 7.89339721e-01 7.18290567e-01
-1.07145138e-01 1.13668382e+00 1.00631714e+00 -2.78179646e-01
1.03503101e-01 1.86932907e-01 2.72192210e-01 8.96659911e-01
-3.50272208e-01 -2.86197543e-01 -7.28444815e-01 1.55017123e-01
8.22116435e-01 -5.95014682e-03 -5.26187956e-01 -2.93648273e-01
-1.28496182e+00 5.70022643e-01 6.19527042e-01 5.20666003e-01
-6.27376497e-01 -2.41720732e-02 2.32539013e-01 -6.01518266e-02
8.27334821e-02 3.32532451e-02 -2.17185020e-01 2.08193213e-01
-9.95848238e-01 1.05077639e-01 4.71620858e-01 1.23611176e+00
9.78298008e-01 -3.23312432e-01 -6.63735449e-01 6.27946973e-01
5.37484646e-01 3.68538320e-01 1.71597421e-01 -1.10795248e+00
4.67929512e-01 1.01116836e+00 -7.37747103e-02 -1.45533824e+00
-1.58842906e-01 -2.40666375e-01 -6.78207159e-01 -1.30703062e-01
1.54320166e-01 3.07177693e-01 -9.93107915e-01 1.94014859e+00
5.68951905e-01 5.39108589e-02 2.34993268e-02 1.18825781e+00
1.27382410e+00 7.90917218e-01 4.84677881e-01 -4.56013829e-01
1.67884183e+00 -1.11392200e+00 -9.32551980e-01 -2.43837699e-01
5.36204755e-01 -6.95989251e-01 1.08278215e+00 1.86078981e-01
-1.10178220e+00 -6.03898466e-01 -6.78727269e-01 -4.30026591e-01
-4.28386092e-01 1.32297054e-01 4.20051664e-01 4.64564785e-02
-1.04882181e+00 1.10277534e-01 -3.42723459e-01 -4.81871635e-01
4.37812299e-01 2.73995250e-01 -5.40082693e-01 -1.59297854e-01
-1.24114633e+00 5.42730093e-01 4.64129567e-01 4.10059273e-01
-5.85853338e-01 -2.78183967e-01 -9.40015674e-01 -4.93112467e-02
6.83303893e-01 -7.72481441e-01 8.10355604e-01 -1.40269303e+00
-1.19748735e+00 1.13556719e+00 -4.34596509e-01 -1.22050836e-03
4.91933107e-01 -6.97620064e-02 -3.89897078e-01 6.51823580e-01
3.35503727e-01 1.26627839e+00 7.98315406e-01 -1.62891710e+00
-8.97378564e-01 -3.55424196e-01 3.59370261e-01 6.30217493e-01
1.41125590e-01 1.14219114e-01 -1.39227974e+00 -5.18534601e-01
5.07422626e-01 -6.81156099e-01 -1.65669695e-02 1.47200450e-01
-6.24955952e-01 -5.02962053e-01 1.04772949e+00 -6.94967389e-01
1.25748062e+00 -2.16687155e+00 6.82626784e-01 2.33766690e-01
4.16833341e-01 8.78236219e-02 -2.19300032e-01 1.59226224e-01
-2.93761995e-02 2.51795262e-01 -2.34351918e-01 -2.39521623e-01
-1.38105631e-01 3.12532991e-01 -1.47842139e-01 1.96406767e-01
2.90434301e-01 1.08225203e+00 -9.69212174e-01 -1.11437070e+00
3.75022382e-01 3.42425227e-01 -3.26434016e-01 1.74839780e-01
-3.82770211e-01 5.90903282e-01 -7.39363372e-01 9.29363072e-01
7.29404867e-01 -2.76459873e-01 -3.47647928e-02 -7.89097071e-01
-6.56525493e-02 -4.43168461e-01 -1.16003346e+00 1.71311712e+00
-1.07101863e-02 2.84302533e-01 1.90077543e-01 -8.79458547e-01
8.25470150e-01 -1.52462255e-02 4.79116023e-01 -8.83312166e-01
2.19691992e-01 -2.46057454e-02 -3.27660501e-01 -8.13497186e-01
3.52168053e-01 2.24261731e-01 -1.74665332e-01 9.82520282e-02
4.41855304e-02 -6.46028072e-02 2.93703496e-01 6.93034768e-01
5.89382589e-01 5.25109887e-01 3.23919743e-01 2.34109480e-02
1.16779673e+00 8.20336416e-02 6.13999963e-01 6.53537810e-01
-4.06684965e-01 4.49127704e-01 6.41762972e-01 -3.60233001e-02
-4.24961001e-01 -9.11319911e-01 2.50754595e-01 1.13550496e+00
8.26979935e-01 -4.20715004e-01 -8.80540609e-01 -7.27633178e-01
-3.38955045e-01 7.58757889e-01 -7.44536161e-01 -2.84512318e-03
-4.68185604e-01 -1.42591342e-01 2.02208355e-01 3.90700281e-01
8.93866658e-01 -1.26038980e+00 -5.12171388e-01 -6.15746193e-02
-4.35692459e-01 -1.31000531e+00 -7.38749623e-01 5.68024516e-02
-3.98616672e-01 -1.09263575e+00 -6.37754500e-01 -9.40273523e-01
8.83207619e-01 3.58866960e-01 9.14350212e-01 3.30513328e-01
8.63873959e-02 6.92861915e-01 -5.21363437e-01 1.32366806e-01
-7.89755285e-02 -3.17784131e-01 -4.53480273e-01 3.96726727e-01
3.27432960e-01 -2.83827454e-01 -7.04170346e-01 4.13891733e-01
-1.13901913e+00 6.35444880e-01 5.79824328e-01 5.96053421e-01
9.86958504e-01 -2.55832464e-01 5.92541173e-02 -7.27876604e-01
4.42197770e-01 -3.80904078e-01 -3.42217028e-01 7.22131371e-01
-2.34045789e-01 7.72032095e-03 2.56230295e-01 -5.90516329e-01
-1.24746740e+00 3.40693295e-01 9.54189077e-02 -6.66600525e-01
-9.09983069e-02 5.52946925e-01 -6.35381043e-01 -7.52986781e-03
2.08347887e-01 4.76982206e-01 -4.34955686e-01 3.73349451e-02
7.53811657e-01 3.90836328e-01 7.65559077e-01 -7.30292737e-01
6.95033252e-01 3.67623687e-01 -4.19097096e-02 -5.70384324e-01
-8.22650850e-01 -4.58962172e-01 -8.65836084e-01 -6.14285171e-01
1.34479249e+00 -8.52780461e-01 -8.05562139e-01 3.71912479e-01
-1.36899936e+00 2.75639500e-02 -1.71141207e-01 1.46038886e-02
-5.05194068e-01 5.26872516e-01 -4.17355508e-01 -7.26942599e-01
-1.91822007e-01 -1.42180967e+00 1.51424921e+00 6.23161912e-01
-8.69835764e-02 -7.43366063e-01 -5.58391511e-01 4.49954957e-01
5.20955510e-02 1.65694728e-01 1.05971420e+00 -5.17883301e-01
-9.86249983e-01 7.22954348e-02 -7.21234381e-01 -1.35735780e-01
5.97722493e-02 2.02698812e-01 -7.66310573e-01 1.83995947e-01
-1.14455923e-01 4.36335504e-02 8.10930789e-01 3.01684260e-01
9.99019027e-01 -1.53373376e-01 -5.10869026e-01 5.89023411e-01
1.24844694e+00 3.76416445e-01 6.54078960e-01 1.77217975e-01
9.37467337e-01 8.97229493e-01 9.95871305e-01 -5.11793680e-02
5.83782375e-01 7.12009847e-01 5.37700415e-01 -3.62386584e-01
-1.75152794e-01 -3.86609018e-01 2.72770703e-01 6.43116593e-01
-1.93107519e-02 -3.23597431e-01 -7.21879065e-01 4.11429167e-01
-2.15235281e+00 -8.42326880e-01 -3.55659962e-01 1.73599529e+00
6.59463465e-01 -7.88017288e-02 -1.78230837e-01 -1.18240751e-01
9.78225946e-01 1.61532164e-01 -5.87343752e-01 -9.45486054e-02
-4.51327264e-01 -2.86317915e-01 7.96774775e-02 4.55709666e-01
-1.02233398e+00 1.27772629e+00 4.54634333e+00 1.04062557e+00
-1.01302922e+00 1.67970493e-01 6.62085950e-01 2.58886337e-01
-5.58233738e-01 2.57506937e-01 -5.38871109e-01 1.67344138e-01
2.49099601e-02 -9.92111862e-02 3.78412127e-01 5.23740232e-01
2.77304232e-01 -5.74729681e-01 -1.14774752e+00 1.22874296e+00
1.69966817e-01 -1.12430298e+00 5.30875206e-01 -2.82761246e-01
4.49073315e-01 -5.85060716e-01 4.39885892e-02 2.31181934e-01
-2.97867537e-01 -7.81284928e-01 1.03680909e+00 8.70526731e-01
6.71426713e-01 -5.29615641e-01 6.24101520e-01 1.82030946e-01
-1.63937950e+00 5.34708016e-02 4.07200716e-02 4.08042014e-01
3.74389559e-01 1.40219271e-01 -1.13779619e-01 8.53830099e-01
7.24648178e-01 8.29552472e-01 -7.70265222e-01 6.25200272e-01
-5.45755148e-01 9.09876749e-02 -1.48561165e-01 7.34163299e-02
3.87296140e-01 -3.49080741e-01 6.83015764e-01 1.08147311e+00
1.04085445e-01 4.86456543e-01 2.40724728e-01 1.26703978e+00
2.11238991e-02 4.11867917e-01 -3.63369077e-01 5.79805113e-02
3.69916111e-01 1.38228858e+00 -1.01374257e+00 -4.17304575e-01
-5.35096824e-01 1.31654990e+00 2.93676913e-01 7.03853607e-01
-9.85327423e-01 -2.10536346e-01 5.52262291e-02 -2.32065711e-02
2.60274053e-01 -2.73225568e-02 1.10577114e-01 -1.11419106e+00
2.38382109e-02 -8.58274817e-01 4.82016325e-01 -1.42115521e+00
-1.01261640e+00 6.83134198e-01 1.52524203e-01 -1.21548271e+00
1.40698090e-01 -3.75737429e-01 -4.82204407e-01 9.61371541e-01
-1.49865079e+00 -1.43531442e+00 -6.36612356e-01 1.04390001e+00
6.40157938e-01 3.09658140e-01 4.81992632e-01 1.19645514e-01
-6.46508098e-01 1.36681214e-01 -8.30258191e-01 1.08693704e-01
5.24956048e-01 -9.35702085e-01 -2.75726050e-01 9.58823085e-01
2.34907910e-01 7.79823303e-01 5.85537791e-01 -8.73300314e-01
-1.25253439e+00 -1.03107285e+00 7.50571251e-01 -6.27314895e-02
4.29279834e-01 -3.67854983e-02 -1.02777362e+00 6.22529745e-01
3.88352066e-01 1.07093044e-02 1.51778162e-01 -3.83478850e-01
-1.61353201e-01 5.00959717e-02 -1.04327035e+00 9.45546806e-01
1.27965665e+00 -7.29153335e-01 -6.72646999e-01 1.82262376e-01
9.83054936e-01 -5.35194993e-01 -6.01977527e-01 4.29098099e-01
1.75414786e-01 -9.11754012e-01 8.59063208e-01 -3.82359326e-01
4.45159614e-01 -7.40680873e-01 -3.02277178e-01 -7.02449441e-01
-1.39102682e-01 -5.18442810e-01 1.51238799e-01 1.63904655e+00
3.00138503e-01 -2.71127880e-01 4.12674457e-01 7.51782894e-01
-4.09857333e-02 -6.46730781e-01 -7.44629681e-01 -8.90184641e-02
-4.47445422e-01 -5.59535325e-01 3.09751481e-01 9.20516908e-01
-5.97395152e-02 2.94304371e-01 -1.49651170e-01 3.49004954e-01
3.82252157e-01 3.19042981e-01 6.21188164e-01 -8.95920813e-01
2.46119201e-02 -5.87524116e-01 -3.58145565e-01 -1.35342371e+00
3.08379143e-01 -8.71372581e-01 4.20876779e-02 -1.64454377e+00
4.02320772e-01 -5.74466512e-02 -3.05153221e-01 5.41125715e-01
-3.71311516e-01 -4.60077301e-02 4.74656940e-01 2.35773891e-01
-1.09665811e+00 4.63864982e-01 1.53754938e+00 -3.04248273e-01
-3.87848914e-01 -4.36066449e-01 -6.71729863e-01 7.96708345e-01
3.18611652e-01 -2.74107367e-01 -4.20575023e-01 -3.61993819e-01
1.49344802e-01 4.31118101e-01 6.59064174e-01 -4.90901262e-01
6.97435915e-01 -3.45416874e-01 4.14884865e-01 -9.24788773e-01
3.06961834e-01 -9.24960732e-01 1.04250804e-01 1.02183565e-01
-3.22719157e-01 4.88840416e-02 1.20651223e-01 7.52911687e-01
-5.05871475e-01 -2.10092124e-02 4.20445383e-01 -2.28548065e-01
-1.20704722e+00 2.55555660e-01 -3.70428085e-01 -3.01226359e-02
1.27996898e+00 -4.61092710e-01 -4.00925994e-01 -5.02315879e-01
-9.83477354e-01 6.55158937e-01 4.36101437e-01 3.82719219e-01
9.87141252e-01 -1.25773513e+00 -3.14065009e-01 6.57848120e-02
3.63783181e-01 8.71867463e-02 6.58176005e-01 1.25386679e+00
-1.83403924e-01 2.06578493e-01 -1.07178889e-01 -9.61638451e-01
-1.39576888e+00 7.85913825e-01 5.01377106e-01 1.33132963e-02
-3.88538361e-01 9.38152552e-01 6.51215434e-01 -8.45904127e-02
2.47752100e-01 -3.16357285e-01 -6.37614667e-01 2.91926056e-01
2.35129267e-01 -9.73044932e-02 -4.58154142e-01 -1.28617024e+00
-5.67225158e-01 1.09514880e+00 9.50845331e-02 -1.57082245e-01
8.43033373e-01 -6.59806848e-01 -4.11043704e-01 3.16442758e-01
9.87516642e-01 3.14086601e-02 -1.03471494e+00 -4.34701025e-01
-6.05933629e-02 -2.76256531e-01 1.12575576e-01 -7.25443184e-01
-1.30293167e+00 7.32716620e-01 4.69242275e-01 -4.48697470e-02
1.62252295e+00 3.35129768e-01 4.10842001e-01 -1.33598298e-02
2.25876465e-01 -9.72828686e-01 2.10825667e-01 1.45872667e-01
1.03395855e+00 -1.15869296e+00 2.42368281e-02 -9.39344645e-01
-1.05948424e+00 1.19341886e+00 8.23333561e-01 2.17917860e-01
3.51219773e-01 -6.81900308e-02 5.23439571e-02 -4.94025260e-01
-5.42560041e-01 -6.68050468e-01 6.00772798e-01 4.74867582e-01
2.04563946e-01 -1.29540801e-01 -4.88762081e-01 7.59151518e-01
3.91112983e-01 -1.04312211e-01 1.09022394e-01 7.10699677e-01
-3.54801208e-01 -9.10189629e-01 -3.51500273e-01 -2.69151945e-02
-6.31115288e-02 -1.49559468e-01 -6.37960315e-01 8.54754388e-01
3.40197533e-01 1.18636119e+00 -3.81347463e-02 -3.75808120e-01
3.95836651e-01 5.43655716e-02 2.86031157e-01 -3.95482600e-01
-6.14905715e-01 4.21084464e-01 -8.39327276e-02 -9.05511379e-01
-9.65630412e-01 -6.50689602e-01 -1.74709976e+00 4.89539765e-02
-2.52031088e-01 3.29284444e-02 2.45761663e-01 1.13786387e+00
2.16572195e-01 7.59111881e-01 2.78444201e-01 -7.48411298e-01
2.48542443e-01 -4.56255257e-01 -4.41983849e-01 7.17524648e-01
1.20602861e-01 -6.41111076e-01 -2.31958494e-01 2.20299363e-01] | [10.312397956848145, 1.2337366342544556] |
7f37725d-c408-4416-a9fa-332470bd2cda | evaluation-and-optimization-of-gradient | 2306.08881 | null | https://arxiv.org/abs/2306.08881v1 | https://arxiv.org/pdf/2306.08881v1.pdf | Evaluation and Optimization of Gradient Compression for Distributed Deep Learning | To accelerate distributed training, many gradient compression methods have been proposed to alleviate the communication bottleneck in synchronous stochastic gradient descent (S-SGD), but their efficacy in real-world applications still remains unclear. In this work, we first evaluate the efficiency of three representative compression methods (quantization with Sign-SGD, sparsification with Top-k SGD, and low-rank with Power-SGD) on a 32-GPU cluster. The results show that they cannot always outperform well-optimized S-SGD or even worse due to their incompatibility with three key system optimization techniques (all-reduce, pipelining, and tensor fusion) in S-SGD. To this end, we propose a novel gradient compression method, called alternate compressed Power-SGD (ACP-SGD), which alternately compresses and communicates low-rank matrices. ACP-SGD not only significantly reduces the communication volume, but also enjoys the three system optimizations like S-SGD. Compared with Power-SGD, the optimized ACP-SGD can largely reduce the compression and communication overheads, while achieving similar model accuracy. In our experiments, ACP-SGD achieves an average of 4.06x and 1.43x speedups over S-SGD and Power-SGD, respectively, and it consistently outperforms other baselines across different setups (from 8 GPUs to 64 GPUs and from 1Gb/s Ethernet to 100Gb/s InfiniBand). | ['Bo Li', 'Xiaowen Chu', 'Shaohuai Shi', 'Longteng Zhang', 'Lin Zhang'] | 2023-06-15 | null | null | null | null | ['quantization'] | ['methodology'] | [-3.12419444e-01 -7.66898096e-01 -9.41089541e-02 -3.55348170e-01
-8.21933985e-01 -2.20132858e-01 4.98661876e-01 2.67291695e-01
-4.98632371e-01 3.91780496e-01 4.65164840e-01 -9.50340986e-01
4.36329171e-02 -8.60505521e-01 -6.19678020e-01 -5.68289399e-01
-4.36723888e-01 3.90746772e-01 4.10060167e-01 -2.18830600e-01
9.29640904e-02 3.26073796e-01 -1.38348269e+00 3.92676860e-01
8.99595857e-01 1.07348931e+00 2.69155771e-01 9.35158789e-01
1.77347008e-02 8.69880080e-01 -5.36179304e-01 -4.78728652e-01
5.25019050e-01 -1.63404286e-01 -7.83242583e-01 -4.89322960e-01
5.56122422e-01 -9.51582491e-01 -5.34116089e-01 8.59484017e-01
8.19561243e-01 5.01453131e-02 7.34330760e-03 -1.12641454e+00
1.82242513e-01 6.05512321e-01 -7.55836487e-01 3.74195486e-01
8.61763433e-02 8.85466635e-02 9.47380841e-01 -7.70321727e-01
3.46734434e-01 1.37559581e+00 9.53448296e-01 5.59457615e-02
-1.31868613e+00 -6.15656435e-01 -2.06586812e-02 5.49699068e-02
-1.47922659e+00 -4.10408378e-01 2.05098048e-01 8.55397657e-02
1.31639671e+00 5.41016638e-01 7.84911096e-01 4.52562958e-01
2.30876669e-01 1.10869157e+00 1.21418130e+00 -2.35447720e-01
4.71315444e-01 -3.89872402e-01 3.65144610e-01 9.15444493e-01
5.24786949e-01 5.80432685e-03 -1.03207910e+00 -7.82203972e-01
4.96165842e-01 5.14250547e-02 -1.36172473e-01 2.28078321e-01
-1.32261860e+00 7.15618491e-01 1.94349065e-01 1.10270634e-01
-2.33652294e-01 4.64479327e-01 1.03611553e+00 5.77406824e-01
7.99391448e-01 -9.07882117e-03 -7.24049151e-01 -7.11622655e-01
-1.35591471e+00 4.94553894e-01 1.22666585e+00 9.75943506e-01
8.02203715e-01 1.21940449e-01 -4.93450314e-01 6.25265062e-01
-2.83581410e-02 6.66930318e-01 6.89910889e-01 -9.67100441e-01
9.71170306e-01 3.90104860e-01 -4.23959017e-01 -1.12512434e+00
-4.67495263e-01 -6.37856245e-01 -1.37600017e+00 -2.29844525e-01
-1.17292427e-01 -3.28552186e-01 -4.14620787e-01 1.36250341e+00
6.20855689e-01 4.92365837e-01 -9.72669125e-02 9.28495347e-01
6.12584114e-01 7.85569906e-01 -3.07522804e-01 -7.21718818e-02
1.16695714e+00 -1.47427499e+00 -3.49579006e-01 -2.43967384e-01
1.46413076e+00 -8.01269412e-01 1.34163725e+00 4.13884640e-01
-1.12312162e+00 -3.98575276e-01 -1.23702574e+00 -3.24676663e-01
1.82969004e-01 2.88265586e-01 1.17313516e+00 6.98054254e-01
-1.49539638e+00 1.03232777e+00 -1.60012794e+00 1.41265914e-01
3.10002238e-01 4.92551863e-01 -1.93345308e-01 -1.74191579e-01
-7.41482794e-01 2.38566637e-01 1.68667167e-01 -2.03146502e-01
-6.86929762e-01 -8.74087036e-01 -3.54233742e-01 2.32073531e-01
7.49779269e-02 -9.57876205e-01 1.25111818e+00 -1.55985445e-01
-1.68109012e+00 4.76955652e-01 -2.75807053e-01 -6.88624918e-01
3.90642226e-01 -5.45110047e-01 1.10693648e-01 7.70700797e-02
-1.48343578e-01 1.21198162e-01 5.42333066e-01 -3.75553817e-01
-6.88237250e-01 -6.41272187e-01 -1.66236222e-01 4.41685259e-01
-9.80341256e-01 -3.05257067e-02 -7.71571100e-01 -6.32255018e-01
2.49514565e-01 -1.07525134e+00 -3.76477093e-01 -2.26237357e-01
-4.74750459e-01 2.48576961e-02 9.24993813e-01 -5.73638141e-01
1.62230575e+00 -2.09988356e+00 1.10026680e-01 1.44031122e-01
5.86824596e-01 5.20408213e-01 -3.76483314e-02 4.26215827e-01
5.89147270e-01 -1.64558664e-01 -4.70347889e-02 -9.16526079e-01
2.22851075e-02 5.28364778e-01 -3.56303036e-01 4.97416496e-01
-6.18049741e-01 6.51981235e-01 -8.32764864e-01 -5.05040824e-01
-1.36491373e-01 5.68457544e-01 -1.01367915e+00 1.07548699e-01
8.06458592e-02 -1.08194659e-02 -5.40179968e-01 4.71632361e-01
9.33339000e-01 -5.94143629e-01 3.76809031e-01 -2.82653302e-01
-4.52564545e-02 7.79936731e-01 -1.34468520e+00 1.75635874e+00
-6.14888668e-01 4.14663285e-01 3.88393104e-01 -8.68985474e-01
5.35558105e-01 -1.34844229e-01 3.62959296e-01 -6.25000238e-01
-1.13180831e-01 6.08607233e-01 -4.69731927e-01 1.30372643e-01
8.29280198e-01 5.33421397e-01 1.46329179e-01 9.01960969e-01
-1.38262644e-01 -2.49341249e-01 4.52409267e-01 7.96936154e-01
1.54623199e+00 -2.14956015e-01 -2.19540387e-01 -4.58871067e-01
3.76265675e-01 -2.34299272e-01 4.77826297e-01 8.68607342e-01
9.74740982e-02 1.43356651e-01 4.81594235e-01 -5.58734655e-01
-7.92423606e-01 -6.53732181e-01 -8.36213827e-02 1.50270700e+00
-1.43116135e-02 -1.44279540e+00 -8.38317990e-01 -4.77428794e-01
1.40862495e-01 3.88400137e-01 9.35671702e-02 -7.50930980e-02
-6.82023287e-01 -1.29027820e+00 6.39559388e-01 2.35596448e-01
1.06004179e+00 -5.13321161e-02 -5.88143110e-01 2.51760304e-01
1.45049533e-02 -1.08806825e+00 -7.87859738e-01 1.17544100e-01
-1.52202344e+00 -5.88839769e-01 -3.16454887e-01 -3.63345176e-01
5.25704324e-01 7.21907258e-01 1.40347254e+00 4.02024508e-01
2.35872297e-03 2.05533542e-02 -2.79139727e-01 1.28390923e-01
-2.28464097e-01 4.76627797e-01 1.13140933e-01 -3.28049153e-01
1.51783004e-01 -9.77805197e-01 -9.68392313e-01 1.78218946e-01
-7.21211791e-01 4.43190575e-01 5.17274380e-01 9.65423763e-01
8.82364273e-01 3.69126536e-02 -3.92204493e-01 -1.07425916e+00
8.26728821e-01 -2.44494185e-01 -6.11669838e-01 -7.29428977e-02
-1.13194752e+00 2.77463138e-01 9.70588386e-01 -2.05218270e-01
-6.69736981e-01 -1.80692181e-01 -3.94888729e-01 -4.20174778e-01
5.78853369e-01 5.10417104e-01 3.11017573e-01 -3.81468773e-01
5.44748008e-01 4.94895130e-01 1.11632384e-01 -6.63999617e-01
2.83427089e-01 7.11634278e-01 3.64244521e-01 -1.00587034e+00
4.29093152e-01 6.53011382e-01 2.74865597e-01 -6.16445363e-01
-8.92497003e-01 -2.86869735e-01 -8.05551466e-03 5.02464712e-01
1.39032021e-01 -1.27751482e+00 -7.17554390e-01 8.25170457e-01
-8.69793773e-01 -7.08323419e-01 -1.25017270e-01 6.17772698e-01
-2.74157971e-01 6.70114398e-01 -1.42591882e+00 -2.75939286e-01
-1.23173809e+00 -1.27745962e+00 1.40603662e+00 -1.50877297e-01
1.57310501e-01 -8.05162728e-01 -7.77468979e-02 1.98483795e-01
9.19390500e-01 -1.05954997e-01 5.87400317e-01 -4.52840626e-01
-6.26216948e-01 -1.17449127e-01 -3.45552266e-01 6.20596170e-01
-3.44371855e-01 -3.95736963e-01 -4.03565109e-01 -1.00809574e+00
3.32249254e-01 -3.55390817e-01 7.65527129e-01 2.23845243e-01
1.33024776e+00 -7.30403364e-01 -4.08944637e-01 1.39443886e+00
1.43593216e+00 -4.99280155e-01 3.94257784e-01 1.88352630e-01
7.35236287e-01 -3.13234657e-01 5.17758548e-01 8.99330556e-01
5.89214921e-01 7.19527125e-01 2.68722892e-01 -1.80651486e-01
-2.28231058e-01 -1.86746255e-01 6.22500896e-01 1.80054820e+00
-1.52912766e-01 1.93457440e-01 -6.78779781e-01 6.33242726e-02
-1.83313894e+00 -2.58748382e-01 -4.88980293e-01 2.40045094e+00
1.11013997e+00 1.19615197e-01 6.70021027e-02 2.42242545e-01
2.19238132e-01 2.55487949e-01 -5.37456930e-01 -4.05972660e-01
-5.58573380e-02 4.48971242e-01 8.42192292e-01 4.04089957e-01
-8.64656746e-01 8.54964614e-01 5.76672029e+00 1.30906379e+00
-1.15709400e+00 6.16088212e-01 8.95008206e-01 -5.69229007e-01
5.22544309e-02 1.18569039e-01 -1.40569890e+00 7.28397787e-01
1.22485077e+00 -2.95832783e-01 7.96058774e-01 1.24247384e+00
-8.82431343e-02 -2.38079131e-02 -1.05244589e+00 1.22358644e+00
-1.32420257e-01 -1.41036606e+00 -6.43444359e-02 1.98518202e-01
9.56434071e-01 6.66483283e-01 -6.20111115e-02 3.86018038e-01
4.93211418e-01 -4.84142452e-01 3.59556645e-01 -2.30737314e-01
8.37794423e-01 -8.19685340e-01 7.91700363e-01 3.25942010e-01
-1.23186219e+00 1.58557847e-01 -6.32896483e-01 -3.63605410e-01
2.09615782e-01 1.34445572e+00 -2.65790254e-01 4.16090816e-01
1.00459003e+00 6.29360855e-01 -6.63688958e-01 6.29243612e-01
-8.44196081e-02 1.03124344e+00 -9.37863052e-01 -2.80672252e-01
3.55748922e-01 -4.26067501e-01 4.76676047e-01 1.37296093e+00
5.16013682e-01 -4.25643055e-03 2.19840363e-01 1.57247826e-01
-1.44819573e-01 2.40327060e-01 -4.25788686e-02 2.58020192e-01
7.21907794e-01 1.23280966e+00 -4.20585930e-01 -8.36057901e-01
-5.81953466e-01 1.23966134e+00 5.78204930e-01 2.65667252e-02
-7.54196167e-01 -4.67777401e-01 1.00243139e+00 1.19420514e-01
3.24607998e-01 -6.47342801e-01 -2.60887891e-01 -1.33793044e+00
3.29698801e-01 -1.05566037e+00 4.01415825e-01 -1.84457272e-01
-1.00800753e+00 6.97959185e-01 -1.61114946e-01 -7.69879878e-01
-2.96162330e-02 -3.04787040e-01 -2.87400842e-01 6.07466757e-01
-1.35349357e+00 -7.09146261e-01 -4.08240080e-01 7.62131155e-01
4.79988158e-02 -1.61416262e-01 8.48870218e-01 6.66850984e-01
-5.27588546e-01 9.43019748e-01 7.58247256e-01 -3.01875919e-01
4.55819279e-01 -1.01333535e+00 8.73578548e-01 7.88869143e-01
9.93817002e-02 8.23658347e-01 5.35932779e-01 -5.44835687e-01
-2.24465799e+00 -1.19480348e+00 8.47539902e-01 1.83300272e-01
6.33369803e-01 -4.13067251e-01 -7.40057409e-01 4.62044507e-01
-9.63807255e-02 4.93339539e-01 5.84535003e-01 2.54943460e-01
-3.56707811e-01 -6.20454550e-01 -9.07405615e-01 7.01790929e-01
1.33464086e+00 -5.02333701e-01 1.56408340e-01 1.16354585e+00
9.02216136e-01 -9.78979111e-01 -1.07363927e+00 8.57618004e-02
4.74030674e-02 -1.38687325e+00 1.07345247e+00 -2.85660386e-01
4.03478980e-01 -1.52289718e-01 -3.23333979e-01 -1.13175607e+00
-2.39143923e-01 -1.03620994e+00 -7.73247182e-01 8.06458831e-01
-3.81307527e-02 -8.06331515e-01 1.05860531e+00 4.06821191e-01
-3.50608289e-01 -1.14755929e+00 -1.06309330e+00 -9.37740982e-01
-1.35681376e-01 -3.96172166e-01 8.59089732e-01 7.00922787e-01
-8.61449465e-02 4.21203852e-01 -3.99153054e-01 -1.41154498e-01
7.26956129e-01 3.19814116e-01 1.18564820e+00 -5.40732324e-01
-9.23069239e-01 -2.77612418e-01 -2.79785812e-01 -1.80790651e+00
-1.62909850e-01 -1.14332318e+00 -5.98385096e-01 -1.19230008e+00
2.05108240e-01 -6.43821657e-01 -5.94516732e-02 6.06560826e-01
-3.53633076e-01 2.22474575e-01 1.72720626e-01 4.63867396e-01
-7.91737676e-01 6.14618778e-01 1.07327533e+00 1.61709949e-01
-4.45644230e-01 -1.35595843e-01 -5.62470496e-01 5.02671182e-01
6.35486364e-01 -4.00810510e-01 -2.53874689e-01 -8.86086881e-01
4.90423143e-01 8.54229778e-02 -1.16416831e-02 -1.32613802e+00
5.31456530e-01 3.65287989e-01 -2.07953274e-01 -5.80591381e-01
2.28108123e-01 -4.22162682e-01 9.88776311e-02 8.10959876e-01
8.34067911e-02 4.82755095e-01 1.98512763e-01 2.83853471e-01
-3.81035239e-01 1.75333098e-01 4.25147951e-01 7.88006485e-02
-2.95271367e-01 4.79632497e-01 -1.22225501e-01 1.51112735e-01
4.54963535e-01 2.56364584e-01 -5.56700408e-01 -2.53524363e-01
1.04460502e-02 2.86292374e-01 2.97605485e-01 -2.36364111e-01
4.89804149e-01 -1.38287199e+00 -8.41769755e-01 3.70161086e-01
-5.57588458e-01 1.30081743e-01 4.12247777e-01 1.14955235e+00
-1.04109871e+00 4.07090127e-01 3.90968233e-01 -8.19052279e-01
-1.56338668e+00 3.19417477e-01 -1.29445478e-01 -9.12464857e-01
-8.76611412e-01 1.19104946e+00 -2.46954978e-01 -4.65427309e-01
2.16774628e-01 -3.75540465e-01 7.63668776e-01 -5.89630127e-01
6.26758695e-01 9.30835485e-01 6.03149652e-01 -2.27367923e-01
-1.15697615e-01 2.86382616e-01 -2.18185872e-01 2.44019672e-01
1.42968309e+00 -1.03797883e-01 -6.85909629e-01 -1.25726968e-01
1.66277456e+00 9.13297012e-03 -1.06792533e+00 -5.16233623e-01
-4.15943354e-01 -6.57571852e-01 6.73047721e-01 -2.27934748e-01
-1.49293387e+00 7.69133508e-01 5.08868814e-01 -9.81089398e-02
1.33763289e+00 -4.33600247e-01 1.58564758e+00 6.53239846e-01
8.25388968e-01 -1.19637346e+00 -5.62326983e-02 7.50952482e-01
3.70079070e-01 -8.70268703e-01 7.23580003e-01 -4.79306996e-01
-3.12481761e-01 8.75178456e-01 4.34945554e-01 -4.57900539e-02
8.16969454e-01 6.36096418e-01 -4.51782823e-01 -1.36290684e-01
-1.07893574e+00 3.34740102e-01 -1.38608381e-01 -6.59148619e-02
4.25876170e-01 1.86034262e-01 -7.38699019e-01 4.68752772e-01
-7.38693655e-01 1.10143095e-01 6.87321052e-02 1.13368070e+00
-1.15540989e-01 -1.19874835e+00 -2.57478148e-01 1.00182199e+00
-4.50639516e-01 -4.44960743e-01 1.97726592e-01 3.26815665e-01
-3.10201168e-01 7.24556029e-01 8.29614699e-02 -8.53297234e-01
2.56069638e-02 -4.61905599e-01 2.79375732e-01 -3.30369174e-01
-1.07084656e+00 -5.57408966e-02 1.64290249e-01 -1.10381365e+00
1.18532307e-01 -4.50062752e-01 -1.06964433e+00 -1.16584146e+00
-2.28333861e-01 2.90235728e-01 9.88839507e-01 3.34220290e-01
1.09061456e+00 3.01064730e-01 6.16220951e-01 -8.40917826e-01
-1.10321951e+00 -7.34112203e-01 -5.89181483e-01 2.43517980e-01
-1.49941832e-01 -2.93201376e-02 -6.24942839e-01 -4.25513685e-01] | [8.52370834350586, 3.421203136444092] |
93710241-16c8-4cab-ba60-f45b1f4a13a6 | learning-long-range-spatial-dependencies-with-1 | 1805.08315 | null | https://arxiv.org/abs/1805.08315v4 | https://arxiv.org/pdf/1805.08315v4.pdf | Learning long-range spatial dependencies with horizontal gated-recurrent units | Progress in deep learning has spawned great successes in many engineering applications. As a prime example, convolutional neural networks, a type of feedforward neural networks, are now approaching -- and sometimes even surpassing -- human accuracy on a variety of visual recognition tasks. Here, however, we show that these neural networks and their recent extensions struggle in recognition tasks where co-dependent visual features must be detected over long spatial ranges. We introduce the horizontal gated-recurrent unit (hGRU) to learn intrinsic horizontal connections -- both within and across feature columns. We demonstrate that a single hGRU layer matches or outperforms all tested feedforward hierarchical baselines including state-of-the-art architectures which have orders of magnitude more free parameters. We further discuss the biological plausibility of the hGRU in comparison to anatomical data from the visual cortex as well as human behavioral data on a classic contour detection task. | ['Junkyung Kim', 'Drew Linsley', 'Thomas Serre', 'Vijay Veerabadran'] | 2018-05-21 | null | null | null | neurips-2018 | ['contour-detection'] | ['computer-vision'] | [ 5.29170036e-01 1.50989905e-01 7.84687102e-02 -4.24505889e-01
-5.62074661e-01 -3.80673289e-01 6.81837082e-01 -1.80589175e-03
-6.40203357e-01 4.75465596e-01 2.87508905e-01 -3.55139226e-01
6.11221306e-02 -4.35051650e-01 -6.93479538e-01 -8.44015658e-01
-4.69088614e-01 -1.14670448e-01 5.24605215e-01 -2.00909674e-01
4.54016089e-01 5.89592040e-01 -1.42189610e+00 6.46607220e-01
1.05111480e-01 1.21957469e+00 1.14959069e-01 7.80828357e-01
2.73014873e-01 5.98626256e-01 -3.17340404e-01 -1.70160875e-01
2.36275062e-01 -3.93123239e-01 -7.28092432e-01 -1.95935935e-01
8.70015085e-01 -7.96966627e-02 -3.52896184e-01 7.47032642e-01
6.58990860e-01 -1.75986782e-01 7.81371057e-01 -6.07662916e-01
-1.01624525e+00 3.90770763e-01 -6.86187267e-01 5.13899982e-01
-2.03560546e-01 -1.29955143e-01 1.28797281e+00 -1.10569012e+00
7.07711458e-01 1.06845975e+00 9.73885655e-01 6.01217330e-01
-1.60597944e+00 -4.51156795e-01 3.00318331e-01 -6.79736137e-02
-1.03163719e+00 -4.30665106e-01 4.87193853e-01 -6.80677354e-01
1.57069480e+00 -1.93035141e-01 8.50275815e-01 1.15760887e+00
7.63771653e-01 7.63663054e-01 1.08212805e+00 -4.22299892e-01
-1.08270779e-01 -6.76675320e-01 2.89846689e-01 8.60140204e-01
2.80299693e-01 3.36221933e-01 -6.01685941e-01 1.19129471e-01
1.40317297e+00 -1.95627823e-01 -4.04311210e-01 -3.94608468e-01
-1.27292109e+00 8.79341781e-01 1.14928234e+00 4.72947598e-01
-3.32567751e-01 4.99893695e-01 3.51216257e-01 2.09931403e-01
9.74052995e-02 5.16326785e-01 -4.45982993e-01 4.54782665e-01
-1.05329895e+00 -5.83962351e-02 4.72031862e-01 5.87415338e-01
5.91301322e-01 4.55392510e-01 -5.57869151e-02 7.83684671e-01
2.00397849e-01 -1.04423337e-01 6.43179059e-01 -6.53939545e-01
4.95728590e-02 4.85122710e-01 -4.03384715e-01 -8.97926569e-01
-8.98660243e-01 -8.63853395e-01 -1.36148620e+00 6.60530984e-01
6.50673747e-01 -7.29003474e-02 -1.32280290e+00 1.63359475e+00
-4.60561633e-01 -2.85863966e-01 -2.92734683e-01 9.61014211e-01
6.85033619e-01 4.08177555e-01 1.09572567e-01 -4.71057259e-02
1.21206439e+00 -6.91430032e-01 -2.98822314e-01 -5.49827754e-01
2.82769173e-01 -5.61885178e-01 7.46575594e-01 4.77115929e-01
-1.07808483e+00 -7.71565199e-01 -1.28380430e+00 -2.78077900e-01
-5.18597007e-01 2.09238216e-01 9.07503545e-01 4.08691585e-01
-1.41805410e+00 8.84792447e-01 -7.52568662e-01 -4.11152333e-01
5.66888094e-01 6.55890942e-01 -4.93242919e-01 3.33955735e-01
-9.03957069e-01 9.02815521e-01 1.78585216e-01 4.35228020e-01
-9.63670731e-01 -5.96255004e-01 -6.68431818e-01 1.35997385e-01
-1.38599753e-01 -6.40724242e-01 1.20042038e+00 -9.76534307e-01
-1.10553718e+00 1.17553246e+00 -5.53545989e-02 -6.94573939e-01
1.20109387e-01 -2.43990630e-01 -8.31377506e-02 1.56574324e-01
-2.22473517e-01 1.32352138e+00 8.07469070e-01 -1.32239139e+00
-5.33627689e-01 -5.52700520e-01 -4.63298500e-01 -1.24955913e-02
7.30307028e-02 -5.52414134e-02 8.10901970e-02 -8.38941455e-01
5.31191707e-01 -7.17641890e-01 -3.56999159e-01 1.96173370e-01
-3.80744189e-01 -1.27736285e-01 5.79616904e-01 -5.11973262e-01
8.85228992e-01 -2.04530883e+00 3.17276955e-01 1.62234202e-01
4.04397219e-01 1.74910679e-01 -8.73555318e-02 2.40523845e-01
-4.57053930e-01 2.26447135e-01 -5.37513077e-01 -4.22240831e-02
-3.37032557e-01 1.73070446e-01 -3.81371886e-01 5.59223294e-01
4.69013333e-01 1.07985377e+00 -4.30641472e-01 -7.38773867e-02
-1.58050284e-01 6.76027000e-01 -2.42349431e-01 -1.03168823e-01
-5.71467020e-02 2.13419929e-01 1.53955519e-01 4.14258897e-01
2.52318621e-01 -3.86924744e-01 1.40086174e-01 -1.23447314e-01
-4.31566387e-01 1.76643878e-01 -5.54071248e-01 1.64245939e+00
-7.72058815e-02 1.14282620e+00 6.98196813e-02 -9.96461987e-01
8.75255466e-01 2.50776947e-01 1.26875371e-01 -8.18875253e-01
2.89124638e-01 2.80621171e-01 6.75428152e-01 7.93023482e-02
9.35837701e-02 -1.00133054e-01 2.59900451e-01 1.21766627e-01
5.87621152e-01 2.39447221e-01 -9.37927067e-02 -6.99586943e-02
1.03829551e+00 1.71627641e-01 4.66359437e-01 -6.60498381e-01
2.09062681e-01 -3.09508890e-01 4.51747268e-01 7.94980645e-01
-3.23991179e-01 1.13043845e+00 7.18932927e-01 -7.59320080e-01
-1.20906985e+00 -1.02379179e+00 -4.45789784e-01 1.36504734e+00
-3.04187953e-01 -1.40715465e-01 -6.49067760e-01 -1.98744446e-01
1.95210259e-02 4.33667265e-02 -1.23424053e+00 -9.07978937e-02
-5.89856803e-01 -8.09293449e-01 8.28578234e-01 1.00960279e+00
6.16669297e-01 -1.50652540e+00 -1.27812207e+00 3.48340988e-01
6.58146799e-01 -8.39360476e-01 -8.04280862e-03 9.47522998e-01
-8.44081342e-01 -7.53778338e-01 -1.03862214e+00 -1.17162538e+00
5.23091972e-01 1.35484636e-01 1.04078639e+00 8.88053700e-02
-8.88730586e-01 3.73914912e-02 4.20890898e-02 -2.19416499e-01
2.74073809e-01 2.35595599e-01 -3.42781782e-01 -1.53703585e-01
8.97690505e-02 -5.89922667e-01 -7.61258483e-01 1.35457516e-01
-6.90511644e-01 1.32479295e-01 8.92076612e-01 1.24915385e+00
4.43995208e-01 -6.19351923e-01 7.91738153e-01 -9.76295233e-01
4.91157353e-01 -3.52608413e-02 -5.07030427e-01 1.27996296e-01
-3.49078327e-01 1.49173960e-01 3.96851689e-01 -1.84190527e-01
-8.43159080e-01 2.52694339e-01 -2.83395916e-01 -3.22429761e-02
-2.83577293e-01 5.09990036e-01 3.93449515e-01 -4.12127137e-01
9.44681287e-01 1.60539523e-01 -3.43897104e-01 -2.21701905e-01
3.05293858e-01 -1.53736332e-02 7.94812024e-01 -3.09077978e-01
2.36366570e-01 4.33430761e-01 1.46979362e-01 -1.16707623e+00
-8.07632923e-01 -4.56786782e-01 -1.17723715e+00 -1.01865180e-01
1.09463048e+00 -6.49006188e-01 -6.05710387e-01 6.36354387e-01
-1.33840096e+00 -5.86907744e-01 1.16888382e-01 2.53206581e-01
-5.98020911e-01 -1.22253865e-01 -1.20083129e+00 -7.19906628e-01
-2.54539937e-01 -9.21324432e-01 9.86454904e-01 3.04531068e-01
-1.08617470e-01 -1.00482392e+00 1.06833994e-01 -4.11601394e-01
5.32374263e-01 3.22261482e-01 1.10053098e+00 -3.56881291e-01
-2.19672188e-01 4.82278019e-02 -6.58325136e-01 1.03708930e-01
-1.75103724e-01 2.42318630e-01 -1.24442041e+00 -3.51170182e-01
-3.82824063e-01 -7.65711427e-01 1.73368704e+00 7.46508360e-01
1.12266636e+00 1.08157083e-01 -2.57862687e-01 7.11080611e-01
1.37879753e+00 9.05430540e-02 6.92400157e-01 2.68444687e-01
5.41721702e-01 6.82775080e-01 -3.43238920e-01 8.36017728e-02
-1.79917976e-01 2.79685318e-01 3.01629871e-01 -5.27763546e-01
-3.80419433e-01 -6.48855343e-02 -1.46583542e-02 6.29520476e-01
-4.92734522e-01 2.23158896e-01 -1.08031404e+00 5.52343011e-01
-1.77175772e+00 -9.63581145e-01 -9.21809226e-02 2.10574722e+00
6.88802719e-01 5.87494850e-01 9.62693989e-02 1.17377818e-01
5.10178924e-01 2.59080410e-01 -6.55534923e-01 -5.16503513e-01
-5.07953286e-01 5.37825167e-01 5.27493715e-01 3.75660568e-01
-1.36454642e+00 1.03088987e+00 7.79422712e+00 2.55736560e-01
-1.33294725e+00 -1.30789310e-01 8.89061213e-01 3.80618751e-01
7.24377483e-02 -1.23872556e-01 -7.78712749e-01 -4.58301842e-01
8.74984741e-01 3.69518459e-01 5.06202877e-02 7.20744312e-01
-4.43631947e-01 -1.48691712e-02 -1.08987045e+00 8.40861917e-01
5.26809730e-02 -1.63920045e+00 8.96078870e-02 1.93677962e-01
7.29885399e-01 4.93662566e-01 3.63267094e-01 1.72657058e-01
2.44425580e-01 -1.53578675e+00 5.84088027e-01 2.94917524e-01
8.97213101e-01 -5.20080805e-01 4.94790465e-01 1.08658738e-01
-1.34405696e+00 -1.63655683e-01 -8.07909846e-01 -1.77847415e-01
-1.17441647e-01 2.74268150e-01 -4.49373305e-01 1.53028145e-01
7.78966069e-01 7.34224677e-01 -9.23372269e-01 1.29083073e+00
-1.47916779e-01 5.57422340e-01 -1.83057323e-01 1.03301637e-01
6.88899100e-01 2.83310384e-01 -5.75024597e-02 1.77967989e+00
-9.65457112e-02 2.70470709e-01 -1.05403647e-01 8.50006282e-01
-1.55282259e-01 2.27076150e-02 -8.72748077e-01 7.95295238e-02
-6.94267750e-02 1.40836394e+00 -1.08444595e+00 -7.78337056e-03
-4.57351685e-01 9.48699772e-01 8.68435800e-01 6.41239643e-01
-3.47024232e-01 -4.61055875e-01 3.67565185e-01 -9.69142094e-02
3.66267711e-01 -4.81534153e-01 -4.30840880e-01 -9.18604076e-01
-3.04995447e-01 -5.32869041e-01 2.93818533e-01 -6.82337344e-01
-1.26462770e+00 9.58105385e-01 -3.23476195e-01 -6.52858317e-01
-1.92681924e-01 -1.29801834e+00 -6.32906795e-01 7.36252844e-01
-1.24247634e+00 -1.13513100e+00 -7.52119422e-02 5.44217527e-01
3.98055077e-01 -3.45054604e-02 9.93148625e-01 -1.24659382e-01
-4.30453300e-01 4.94533330e-01 -2.64989287e-01 6.06704712e-01
4.60247934e-01 -1.26770222e+00 7.46051013e-01 7.37273812e-01
5.45480430e-01 7.43132889e-01 3.08814019e-01 -4.32236671e-01
-8.83134723e-01 -6.65939748e-01 5.88474929e-01 5.11892438e-02
5.51768363e-01 -8.03966463e-01 -1.25931418e+00 9.36508596e-01
7.01717973e-01 3.81619215e-01 5.43726742e-01 2.53106952e-01
-8.56031001e-01 1.84559494e-01 -6.13008022e-01 5.23654044e-01
1.20927966e+00 -5.73472857e-01 -7.34617352e-01 -4.36522886e-02
3.38582426e-01 -2.69191079e-02 -7.35870719e-01 6.50663614e-01
1.02296150e+00 -1.38626873e+00 8.36498201e-01 -8.26526523e-01
4.12985444e-01 6.81985095e-02 -2.32652366e-01 -1.32262480e+00
-8.90167594e-01 -4.57176954e-01 2.86525160e-01 5.94001710e-01
6.43697202e-01 -4.31396395e-01 7.73268163e-01 1.40699401e-01
-3.59735876e-01 -9.61907506e-01 -1.02886975e+00 -4.28019971e-01
5.48899055e-01 1.06631741e-02 -5.33879876e-01 7.08429694e-01
1.82983041e-01 7.94869900e-01 -7.86884502e-02 -9.66559201e-02
6.47328973e-01 1.18790396e-01 1.74273979e-02 -1.43587887e+00
-9.51414928e-02 -1.03780353e+00 -7.36339927e-01 -1.02701235e+00
-1.15608377e-03 -8.09305906e-01 2.27374524e-01 -1.64638710e+00
4.65887077e-02 2.97887754e-02 -4.22033548e-01 7.15468884e-01
1.14045948e-01 4.72302496e-01 1.16540708e-01 9.83653665e-02
-2.57231444e-01 3.45881194e-01 1.19021106e+00 8.02033693e-02
1.94782346e-01 -3.38445634e-01 -4.32202607e-01 8.16402197e-01
8.27291965e-01 -1.67756021e-01 -9.69160497e-02 -4.56173450e-01
2.14661896e-01 -1.14759207e-01 4.70444947e-01 -1.30432010e+00
4.95350391e-01 3.71799231e-01 1.16982651e+00 -3.85711998e-01
2.33455077e-01 -4.33810472e-01 -4.30847079e-01 6.67099237e-01
-7.98266292e-01 2.74240941e-01 3.97475183e-01 3.84344250e-01
-2.80565888e-01 8.22083429e-02 1.09245288e+00 -3.21770638e-01
-9.54510868e-01 1.12699233e-01 -7.88916290e-01 -9.93571430e-02
6.64747059e-01 -2.64182717e-01 -6.15071535e-01 -2.17312783e-01
-9.57125068e-01 -1.93694487e-01 -5.24339676e-02 2.83840060e-01
9.36287522e-01 -1.07928085e+00 -7.47672796e-01 4.75719541e-01
9.90195549e-04 -3.36440146e-01 2.86880136e-03 6.63286567e-01
-5.13570011e-01 6.70427859e-01 -8.72153163e-01 -7.20994651e-01
-8.30024302e-01 3.98086041e-01 6.21578634e-01 -8.85361582e-02
-8.58750284e-01 1.22137773e+00 7.45056093e-01 -2.11332723e-01
3.91546398e-01 -5.07499218e-01 -2.63653010e-01 1.41573697e-01
4.15611684e-01 -2.45286494e-01 1.61792532e-01 -4.75883663e-01
-2.40294799e-01 8.36325228e-01 -1.40612960e-01 -1.49354458e-01
1.38572097e+00 2.86861897e-01 -1.96443032e-02 8.15440893e-01
1.18241262e+00 -4.54898953e-01 -1.59867978e+00 -7.42136911e-02
3.66203904e-01 1.09865293e-01 1.33434266e-01 -8.59782100e-01
-1.16159356e+00 1.53740454e+00 6.73520327e-01 1.89669311e-01
9.44910765e-01 -1.47129774e-01 2.70731360e-01 5.69421768e-01
3.11298043e-01 -8.21214616e-01 1.87346011e-01 6.53703868e-01
1.26666307e+00 -1.23097861e+00 -2.02928528e-01 -7.60379210e-02
-6.18146360e-01 1.44227242e+00 7.45976031e-01 -7.45574474e-01
8.15214455e-01 4.33556527e-01 1.33662373e-01 -2.99421906e-01
-1.11008835e+00 -4.56056237e-01 4.87518281e-01 5.12503147e-01
1.15718257e+00 -8.31308961e-02 4.37973104e-02 2.89258927e-01
2.71238070e-02 -4.00919139e-01 3.14722240e-01 8.93224597e-01
-8.72742176e-01 -6.92350924e-01 1.11269765e-02 5.53718209e-01
-5.80547810e-01 -3.29172075e-01 -7.14631319e-01 1.02698565e+00
-1.77963719e-01 4.30210292e-01 1.66703999e-01 -3.13716501e-01
2.50171989e-01 1.97908282e-01 7.42083788e-01 -5.92997193e-01
-9.62539017e-01 5.79595387e-01 -8.77001584e-02 -5.49596667e-01
-3.77236336e-01 -5.25065720e-01 -1.47847879e+00 1.30948648e-01
-1.56835392e-01 -2.37978920e-01 5.04152119e-01 6.09158218e-01
1.05455764e-01 6.83140874e-01 -2.66512558e-02 -1.41179550e+00
-3.75269115e-01 -1.09203863e+00 -6.93161011e-01 1.07857145e-01
5.03720164e-01 -6.24867916e-01 -1.40856683e-01 1.32997811e-01] | [9.573237419128418, 2.4122469425201416] |
1f0762ce-9043-4d36-ab81-89ae5224959b | sentube-a-corpus-for-sentiment-analysis-on | null | null | https://aclanthology.org/L14-1188 | https://aclanthology.org/L14-1188.pdf | SenTube: A Corpus for Sentiment Analysis on YouTube Social Media | In this paper we present SenTube -- a dataset of user-generated comments on YouTube videos annotated for information content and sentiment polarity. It contains annotations that allow to develop classifiers for several important NLP tasks: (i) sentiment analysis, (ii) text categorization (relatedness of a comment to video and/or product), (iii) spam detection, and (iv) prediction of comment informativeness. The SenTube corpus favors the development of research on indexing and searching YouTube videos exploiting information derived from comments. The corpus will cover several languages: at the moment, we focus on English and Italian, with Spanish and Dutch parts scheduled for the later stages of the project. For all the languages, we collect videos for the same set of products, thus offering possibilities for multi- and cross-lingual experiments. The paper provides annotation guidelines, corpus statistics and annotator agreement details. | ['ro', 'Agata Rotondi', 'Olga Uryupina', 'Barbara Plank', 'Aliaksei Severyn', 'Aless Moschitti'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['spam-detection'] | ['natural-language-processing'] | [-1.33732632e-01 -5.16499244e-02 -6.84257865e-01 -4.94903147e-01
-6.80604935e-01 -1.01735306e+00 8.01693082e-01 5.95754325e-01
-4.17644262e-01 4.88328218e-01 6.74461246e-01 -5.28398342e-03
2.90622473e-01 -9.91592184e-02 -2.68029511e-01 -4.76801157e-01
-1.10546267e-02 1.21544562e-01 3.84621292e-01 8.86325762e-02
7.45554864e-01 1.27848372e-01 -1.92512703e+00 1.00705945e+00
3.32090437e-01 1.30274224e+00 2.13026330e-01 7.47874498e-01
-1.87867418e-01 1.03551662e+00 -4.06167805e-01 -9.00683522e-01
-8.99719598e-04 -4.42933619e-01 -9.17727053e-01 2.28291720e-01
5.97543776e-01 1.59109741e-01 4.23238128e-02 1.12076247e+00
2.21028611e-01 -9.10624489e-02 5.64675450e-01 -1.31343138e+00
-2.74067909e-01 6.20145023e-01 -5.10396063e-02 3.54116231e-01
8.79878759e-01 -2.23907575e-01 1.36035967e+00 -7.92800725e-01
1.50367260e+00 9.04695988e-01 4.80296016e-01 3.61559540e-01
-4.97773319e-01 -2.62289703e-01 -8.89195944e-04 3.43760043e-01
-1.10976481e+00 -3.76110703e-01 4.91456181e-01 -9.15717959e-01
9.24992323e-01 3.88273209e-01 8.64574373e-01 1.45610607e+00
-8.17937553e-02 1.16551828e+00 7.79503942e-01 -4.25301999e-01
1.36946827e-01 9.72459137e-01 1.24957934e-01 3.46672952e-01
-1.43123925e-01 -6.74700856e-01 -7.80153632e-01 -1.02873199e-01
-3.04289293e-02 -3.96093041e-01 -3.29605728e-01 -2.95943588e-01
-1.18890452e+00 9.64326024e-01 -3.77203338e-02 6.42585278e-01
-2.50748098e-01 -3.34971100e-01 1.38728356e+00 5.84250391e-01
9.63244796e-01 3.89989913e-01 -6.86191857e-01 -6.29133940e-01
-9.49585795e-01 2.77476728e-01 1.29320765e+00 1.25033200e+00
4.56764013e-01 -3.74072284e-01 -2.21494306e-02 8.26660633e-01
3.25617343e-01 5.00703640e-02 8.88898492e-01 -7.64679790e-01
7.89709449e-01 6.83250666e-01 1.30654678e-01 -1.14698982e+00
-2.55155206e-01 3.21042687e-01 -7.75880367e-02 -3.13939124e-01
2.99617708e-01 -2.27603301e-01 -2.49391422e-01 1.01658046e+00
1.70364752e-01 -3.07165533e-01 -1.00025475e-01 8.05141211e-01
1.32624876e+00 6.58780277e-01 -2.58596963e-03 -6.19001091e-01
1.56827676e+00 -9.61841643e-01 -8.20278168e-01 2.78473794e-01
1.02839220e+00 -1.12657583e+00 1.08208406e+00 5.91171622e-01
-9.78752077e-01 -4.78559881e-01 -6.46457732e-01 -9.64078456e-02
-8.21029603e-01 4.37384933e-01 2.63238907e-01 6.88112199e-01
-8.08975875e-01 3.34853888e-01 -2.30420783e-01 -8.04015100e-01
4.02514249e-01 2.93670259e-02 -6.11034155e-01 1.47258461e-01
-1.00891149e+00 4.93890822e-01 2.82369286e-01 -4.52871799e-01
-3.13623011e-01 -4.03274298e-01 -8.29442978e-01 -3.36445838e-01
1.97413594e-01 2.72097886e-01 1.23319304e+00 -1.78962731e+00
-1.28965080e+00 1.51422966e+00 -3.06629688e-01 -5.21209896e-01
4.22713995e-01 -3.82500201e-01 -6.15210235e-01 4.06106949e-01
3.60692322e-01 7.19141185e-01 8.93485248e-01 -7.12597489e-01
-9.64302361e-01 -2.92263210e-01 8.81449953e-02 2.09712163e-01
-7.46261001e-01 6.79706275e-01 -7.39790440e-01 -8.41727138e-01
-4.51369613e-01 -1.05519915e+00 4.68984962e-01 -2.79293925e-01
-3.07448298e-01 -6.38000488e-01 1.10091746e+00 -8.82309020e-01
1.41479075e+00 -2.43031311e+00 1.41407609e-01 -1.15732223e-01
-8.77207741e-02 1.59313470e-01 -6.97980681e-03 7.31361449e-01
-9.71725807e-02 3.44345987e-01 4.42514926e-01 -2.80319303e-01
-1.56599417e-01 -1.44953877e-01 -1.67609826e-01 6.20122671e-01
-1.23064443e-01 6.45677447e-01 -8.35689485e-01 -7.62193143e-01
-6.85242265e-02 1.49175674e-01 -5.31183481e-01 -1.31342083e-01
-2.97851950e-01 3.26828092e-01 -5.82962096e-01 8.25520039e-01
1.25448465e-01 1.17991097e-01 6.24240749e-02 -3.55625838e-01
-4.28247452e-01 5.80830216e-01 -8.36270154e-01 1.43951857e+00
-4.52921569e-01 1.45708144e+00 1.50353298e-01 -8.70776534e-01
7.05249190e-01 5.34900844e-01 7.80188739e-01 -5.12318730e-01
2.99800396e-01 3.64537448e-01 -4.32725430e-01 -1.06014442e+00
7.77885973e-01 4.76309597e-01 -2.16369137e-01 3.97811145e-01
2.91327327e-01 8.60018730e-02 9.74651098e-01 3.40532511e-01
6.87622964e-01 2.07166597e-01 3.00028205e-01 -4.57011610e-01
1.01429904e+00 2.01793358e-01 8.67721885e-02 1.58621758e-01
-2.82458931e-01 4.42265391e-01 1.04501987e+00 -4.87996399e-01
-1.04959714e+00 -2.34861091e-01 -2.35298648e-01 1.26840162e+00
-4.38440628e-02 -1.03843057e+00 -8.31580579e-01 -1.00456095e+00
-1.19624399e-01 4.19405729e-01 -6.33108974e-01 4.00171936e-01
-3.44455391e-01 -2.72653282e-01 9.46205854e-02 1.43241733e-01
2.06953064e-01 -1.34673786e+00 -4.22532141e-01 -1.75569326e-01
-5.38228869e-01 -1.33979702e+00 -4.36481595e-01 -1.39820445e-02
-5.20113945e-01 -1.41250610e+00 -5.16600072e-01 -9.78598297e-01
3.63914102e-01 9.40268934e-02 1.17350674e+00 -6.46456629e-02
-1.42473415e-01 7.54718840e-01 -1.13207531e+00 -4.13593352e-01
-5.39152801e-01 3.61381024e-01 -4.17620987e-02 9.96134877e-02
7.77947426e-01 1.94354862e-01 -2.47243434e-01 6.60828650e-01
-7.35080600e-01 -5.21217763e-01 2.97913738e-02 4.72660929e-01
2.43360907e-01 -4.45635058e-02 3.74702215e-01 -9.85121667e-01
7.03235865e-01 -6.26206398e-01 -5.37925601e-01 -1.43764392e-01
-2.45766297e-01 -7.57747948e-01 4.73083407e-01 -1.87762618e-01
-9.97314215e-01 2.66572148e-01 -1.67722479e-01 -3.74569222e-02
-2.93480039e-01 5.74551463e-01 -2.63139084e-02 1.67349577e-02
6.73374236e-01 -2.36394852e-01 -2.70816624e-01 -3.89419645e-01
1.96533084e-01 1.32573152e+00 5.95223084e-02 6.60311952e-02
7.92098939e-02 3.87253314e-01 -5.01038432e-01 -1.23473585e+00
-9.06567216e-01 -1.21443224e+00 -9.09915566e-01 -6.59604132e-01
1.06978118e+00 -1.05278039e+00 -5.56126773e-01 2.58598804e-01
-9.45737660e-01 -1.33493006e-01 -3.96937072e-01 5.22260964e-01
-4.94999588e-01 4.01016742e-01 -6.75255954e-01 -5.89530051e-01
-5.34015149e-02 -1.19628954e+00 1.07088542e+00 -1.54411331e-01
-6.85176790e-01 -1.12278259e+00 6.92532808e-02 7.64922619e-01
-1.66422293e-01 9.52457357e-03 1.68358162e-01 -1.01933110e+00
-1.33484555e-02 -8.19955051e-01 -1.58724993e-01 6.26251042e-01
-3.14044446e-01 4.49790478e-01 -8.58316481e-01 -1.26428515e-01
-6.17348589e-02 -7.32112169e-01 6.64037883e-01 3.29801470e-01
1.10275280e+00 -4.71923620e-01 -4.03811872e-01 -4.65127602e-02
1.12842560e+00 -8.32196623e-02 7.08545506e-01 5.93358099e-01
5.19968688e-01 1.05673814e+00 1.18675411e+00 6.46081507e-01
1.79200187e-01 7.83145905e-01 5.05323946e-01 6.46530509e-01
1.19613796e-01 2.27760933e-02 7.17263162e-01 8.15274894e-01
-1.55765578e-01 -6.94949567e-01 -5.88377893e-01 7.88011909e-01
-1.80185413e+00 -1.11626816e+00 -7.71399915e-01 1.94047356e+00
3.80548716e-01 1.09031305e-01 6.46246254e-01 1.13993980e-01
6.49857044e-01 2.13496253e-01 1.66371256e-01 -5.71381390e-01
-7.46332854e-02 -4.36066419e-01 3.98584515e-01 5.16541526e-02
-1.43894982e+00 7.44875252e-01 5.97438097e+00 9.30212855e-01
-1.09743869e+00 3.91634822e-01 6.08606637e-01 -1.46685466e-01
9.98937041e-02 -1.57456607e-01 -1.05999386e+00 8.14959884e-01
1.07436073e+00 2.01569811e-01 -2.50189602e-02 1.24106336e+00
3.13896775e-01 -4.10950720e-01 -8.78753483e-01 8.12735021e-01
6.10347927e-01 -1.51024377e+00 -2.18072623e-01 -1.00781985e-01
7.96114743e-01 4.68503833e-01 -2.76264459e-01 1.61957756e-01
-4.72759873e-01 -3.65813196e-01 1.03742492e+00 -1.04502887e-02
5.70008874e-01 -5.09334445e-01 1.14857852e+00 1.71263605e-01
-9.33206618e-01 -2.42563367e-01 -1.61544949e-01 1.12303421e-01
3.02992821e-01 5.71608543e-01 -3.84601742e-01 2.40232944e-01
1.23742211e+00 1.56515002e+00 -6.47748470e-01 8.83122146e-01
-2.01575279e-01 4.71513599e-01 -1.39947951e-01 -6.51895106e-01
2.96860725e-01 -3.13097984e-01 7.45891988e-01 1.68461621e+00
3.05882066e-01 -3.65015686e-01 -1.32506853e-02 1.01537660e-01
-1.56443700e-01 7.15332270e-01 -5.67454398e-01 -3.35304230e-01
1.66181505e-01 1.65590954e+00 -1.20293450e+00 -3.07456911e-01
-7.42902040e-01 9.55933034e-01 -1.74458101e-02 -5.05985990e-02
-6.91754818e-01 -4.68657762e-01 4.84869689e-01 4.67675179e-01
5.51380575e-01 1.34595530e-02 5.44365421e-02 -1.35881841e+00
1.84365913e-01 -8.17207813e-01 4.15281862e-01 -8.71974289e-01
-1.10386932e+00 7.03077316e-01 2.98320707e-02 -1.59111571e+00
-2.66504526e-01 -8.82222712e-01 -3.84688079e-01 1.88144803e-01
-1.07639074e+00 -9.41266954e-01 -2.81700343e-01 5.84213316e-01
9.80727077e-01 -5.80269217e-01 4.88108397e-01 6.98201478e-01
-4.11539495e-01 1.19145125e-01 -2.71640345e-02 1.91574544e-02
1.02188468e+00 -1.03015554e+00 -1.55460566e-01 3.15755516e-01
2.93401778e-01 1.42092437e-01 7.40317345e-01 -6.79067552e-01
-9.87299621e-01 -9.09533441e-01 1.43826485e+00 -7.26083338e-01
1.23233759e+00 -3.37413758e-01 -3.14761996e-01 5.96813202e-01
5.11819780e-01 -2.43405417e-01 9.51233745e-01 -1.34490766e-02
-1.01510130e-01 1.65582016e-01 -9.15188372e-01 3.30825597e-01
7.23092675e-01 -6.92869723e-01 -1.54867202e-01 1.12905896e+00
2.31229082e-01 -3.27308327e-01 -8.45116913e-01 -2.43712917e-01
6.42823696e-01 -1.16737664e+00 5.07256806e-01 -4.55330521e-01
8.13058436e-01 6.96709529e-02 -8.08145106e-03 -9.91577744e-01
1.50038943e-01 -6.52744591e-01 6.83715343e-02 1.49279690e+00
6.28832400e-01 -1.91920727e-01 7.70772278e-01 3.32582623e-01
-2.46739507e-01 -3.68087143e-01 -7.62566328e-01 -4.29469824e-01
-3.79218251e-01 -7.56646693e-01 -1.81226194e-01 9.19273496e-01
6.65806830e-01 4.04576540e-01 -3.01533580e-01 -6.37032926e-01
-1.88994348e-01 -1.40873834e-01 7.58947909e-01 -1.31802583e+00
1.40556842e-01 -6.00053251e-01 -6.62816584e-01 -8.18618357e-01
4.60932642e-01 -6.94719136e-01 -1.06218718e-01 -1.13674474e+00
8.28169435e-02 4.40188088e-02 4.77205396e-01 1.70724899e-01
4.39664274e-01 7.38567710e-01 6.44153431e-02 4.64228094e-01
-1.04594553e+00 -2.17654616e-01 8.50943208e-01 7.05395788e-02
-3.17379162e-02 2.29439080e-01 -4.20805633e-01 8.98555756e-01
7.56441176e-01 -4.69195932e-01 -2.29117215e-01 1.33807778e-01
8.47089708e-01 -3.46320361e-01 9.10199061e-02 -6.84930563e-01
-1.08742014e-01 1.83402568e-01 -1.54365338e-02 -5.65441132e-01
2.38133222e-01 -9.14635003e-01 -3.01928133e-01 3.63323152e-01
-5.92783749e-01 5.78056984e-02 -6.36179820e-02 5.79369724e-01
-7.10234761e-01 -9.30056095e-01 4.42533791e-01 -2.21071228e-01
-8.22753310e-01 -5.14670946e-02 -9.16026771e-01 8.09343755e-02
1.27388144e+00 -3.63600761e-01 -1.97377831e-01 -6.51572108e-01
-8.96666110e-01 4.44847606e-02 5.55339217e-01 8.31268072e-01
5.46814092e-02 -1.04051709e+00 -6.22815549e-01 5.22673083e-03
5.12019217e-01 -7.21730769e-01 1.82269663e-02 9.81813192e-01
-7.14394510e-01 6.47592306e-01 -1.66838124e-01 -4.76355016e-01
-1.83084095e+00 5.70551038e-01 -2.85957724e-01 -1.33912593e-01
-1.96633413e-01 7.53913820e-01 -2.14676023e-01 -2.74285246e-02
2.10340276e-01 -1.36314839e-01 -8.56094360e-01 1.11738646e+00
5.15000403e-01 4.00745422e-01 2.14738652e-01 -1.28651726e+00
-2.87682742e-01 3.58929962e-01 8.90827272e-03 1.43537298e-01
1.15824223e+00 -4.36846137e-01 -2.81975329e-01 6.66229844e-01
1.63130009e+00 1.82874143e-01 -6.93823934e-01 3.39765936e-01
2.92789191e-01 -5.43923676e-01 1.53347880e-01 -4.24382240e-01
-1.07978344e+00 6.14824831e-01 3.03512156e-01 1.09974456e+00
8.51883173e-01 1.67270184e-01 6.21118724e-01 2.22970009e-01
5.35346083e-02 -1.60649049e+00 1.51609614e-01 4.77727056e-01
9.66169178e-01 -1.46335292e+00 2.29736164e-01 -7.00422406e-01
-1.11805987e+00 1.38065863e+00 3.01705420e-01 3.54430228e-02
9.24035490e-01 -6.19630031e-02 7.80218234e-03 -3.86795640e-01
-5.94623685e-01 -1.19467132e-01 5.67110658e-01 5.12340844e-01
8.91407013e-01 -3.70305955e-01 -8.79889727e-01 3.64553928e-01
-4.81862091e-02 6.14929199e-02 7.59522974e-01 7.31158376e-01
-2.44176790e-01 -1.05871105e+00 2.14364771e-02 6.47694707e-01
-9.91384268e-01 8.33473206e-02 -7.61160135e-01 9.34706390e-01
1.58501998e-01 9.19635594e-01 1.50665939e-01 -4.80446666e-02
1.56187043e-01 -3.34067941e-02 7.75387883e-02 -6.81626558e-01
-1.02143109e+00 2.96613336e-01 7.98987865e-01 -6.80352330e-01
-1.16314900e+00 -1.29653490e+00 -5.56855321e-01 -7.31953233e-02
-4.52118248e-01 4.32814360e-01 1.02928519e+00 7.99231589e-01
3.05586249e-01 1.79401059e-02 4.67661470e-01 -1.04533982e+00
1.55415222e-01 -1.03420901e+00 -6.87335491e-01 5.86574852e-01
-1.40594579e-02 -4.09895837e-01 -6.63120270e-01 6.48585021e-01] | [12.93708324432373, 5.261788845062256] |
bd0cb3aa-757f-456d-b9f5-4b047376668a | damo-nlp-at-semeval-2023-task-2-a-unified | 2305.03688 | null | https://arxiv.org/abs/2305.03688v3 | https://arxiv.org/pdf/2305.03688v3.pdf | DAMO-NLP at SemEval-2023 Task 2: A Unified Retrieval-augmented System for Multilingual Named Entity Recognition | The MultiCoNER \RNum{2} shared task aims to tackle multilingual named entity recognition (NER) in fine-grained and noisy scenarios, and it inherits the semantic ambiguity and low-context setting of the MultiCoNER \RNum{1} task. To cope with these problems, the previous top systems in the MultiCoNER \RNum{1} either incorporate the knowledge bases or gazetteers. However, they still suffer from insufficient knowledge, limited context length, single retrieval strategy. In this paper, our team \textbf{DAMO-NLP} proposes a unified retrieval-augmented system (U-RaNER) for fine-grained multilingual NER. We perform error analysis on the previous top systems and reveal that their performance bottleneck lies in insufficient knowledge. Also, we discover that the limited context length causes the retrieval knowledge to be invisible to the model. To enhance the retrieval context, we incorporate the entity-centric Wikidata knowledge base, while utilizing the infusion approach to broaden the contextual scope of the model. Also, we explore various search strategies and refine the quality of retrieval knowledge. Our system\footnote{We will release the dataset, code, and scripts of our system at {\small \url{https://github.com/modelscope/AdaSeq/tree/master/examples/U-RaNER}}.} wins 9 out of 13 tracks in the MultiCoNER \RNum{2} shared task. Additionally, we compared our system with ChatGPT, one of the large language models which have unlocked strong capabilities on many tasks. The results show that there is still much room for improvement for ChatGPT on the extraction task. | ['Yong Jiang', 'Fei Huang', 'Pengjun Xie', 'Kewei Tu', 'Yueting Zhuang', 'Weiming Lu', 'Yinghui Li', 'Jiong Cai', 'Zixia Jia', 'Shen Huang', 'Zeqi Tan'] | 2023-05-05 | null | null | null | null | ['named-entity-recognition-ner', 'multilingual-named-entity-recognition'] | ['natural-language-processing', 'natural-language-processing'] | [-4.76772219e-01 -1.35393351e-01 -8.65060985e-02 -7.27487653e-02
-1.40208614e+00 -9.96452808e-01 4.51031417e-01 -1.83043584e-01
-8.90285134e-01 1.03146684e+00 3.63163531e-01 -5.25600672e-01
-2.92105377e-01 -4.43198234e-01 -5.68278253e-01 -3.31584185e-01
2.79220611e-01 6.43411279e-01 3.55285376e-01 -6.84982657e-01
-1.03303544e-01 1.09815113e-01 -1.11935413e+00 3.63598764e-01
8.99382174e-01 4.55249697e-01 5.90527952e-01 5.53806901e-01
-2.92065799e-01 7.10505068e-01 -6.54978335e-01 -7.82720506e-01
1.11452907e-01 1.32238418e-01 -1.19223189e+00 -8.61285746e-01
4.91209239e-01 1.11881264e-01 -5.45083463e-01 1.07918572e+00
8.92915368e-01 9.88490507e-02 2.14904279e-01 -9.41876411e-01
-9.73899662e-01 1.28744876e+00 -2.00875461e-01 3.53065819e-01
3.72193575e-01 -2.00037852e-01 1.23852921e+00 -1.26714194e+00
1.04674196e+00 9.71553862e-01 9.59022701e-01 6.25617087e-01
-7.88969159e-01 -9.68535066e-01 2.54818350e-01 6.00810535e-02
-1.83783126e+00 -8.00091326e-01 3.46188657e-02 -1.20455272e-01
1.30444467e+00 4.80996490e-01 8.15420300e-02 1.23103666e+00
-3.28365862e-01 8.35234940e-01 1.08546698e+00 -4.87548918e-01
-2.95217037e-01 1.19062304e-01 3.90915006e-01 5.22686362e-01
2.97089249e-01 -7.40542263e-02 -6.21906519e-01 -8.07440653e-02
5.29034555e-01 -2.22728297e-01 -4.85416800e-01 3.71398956e-01
-1.38563180e+00 2.72887617e-01 1.40446365e-01 6.08235598e-01
-1.47077143e-01 -1.15557976e-01 3.75832945e-01 4.15106505e-01
4.30799693e-01 6.90367877e-01 -1.22014868e+00 -2.56961673e-01
-9.40597475e-01 1.11592554e-01 9.99831438e-01 1.39322817e+00
8.05043459e-01 -3.23024422e-01 -2.43532732e-01 1.17675900e+00
3.10321510e-01 6.91589117e-01 4.06906515e-01 -8.04461122e-01
9.61020350e-01 5.52204609e-01 1.81602374e-01 -4.15920049e-01
-5.62087297e-01 -5.79151571e-01 -5.24891138e-01 -5.27922690e-01
4.86367434e-01 -5.33416808e-01 -8.26956809e-01 1.72751617e+00
2.53054261e-01 -7.40742013e-02 3.01715374e-01 7.20563471e-01
1.22100997e+00 4.11257952e-01 3.78908783e-01 1.14215903e-01
1.64424646e+00 -1.05014062e+00 -8.74720752e-01 -2.17556134e-01
1.21887445e+00 -1.11825728e+00 9.02970672e-01 1.72290027e-01
-7.85287738e-01 -3.54768693e-01 -6.88599586e-01 -3.19253564e-01
-9.12788987e-01 5.06109178e-01 4.99286383e-01 5.50697803e-01
-1.15218306e+00 3.60063255e-01 -5.09799242e-01 -5.75280607e-01
-3.55153263e-01 9.91658866e-02 -6.03411734e-01 -3.49808306e-01
-1.97442067e+00 1.10676157e+00 7.69085050e-01 4.57295209e-01
-4.41677392e-01 -6.18067384e-01 -9.34787154e-01 -1.10303327e-01
6.04682326e-01 -5.18737614e-01 1.22294164e+00 -1.33864060e-01
-1.07619441e+00 6.75635695e-01 -1.50566027e-01 -4.53631356e-02
2.54406422e-01 -4.94611502e-01 -7.01591611e-01 -3.02065670e-01
4.91258025e-01 4.71146613e-01 8.25125957e-04 -1.00976753e+00
-8.48960578e-01 -3.06022108e-01 7.49608725e-02 3.25033158e-01
-2.15747863e-01 4.77193534e-01 -8.55449259e-01 -8.32879066e-01
-2.14871801e-02 -1.03625357e+00 -9.66162607e-02 -1.13689375e+00
-5.57766736e-01 -3.49497765e-01 3.13427329e-01 -1.00284278e+00
1.84912848e+00 -1.91455722e+00 -1.82969775e-02 2.55327835e-03
1.44022629e-02 3.93021196e-01 -2.02591419e-01 8.18207204e-01
2.65035201e-02 7.36064434e-01 4.97481711e-02 -2.79351383e-01
1.48201913e-01 1.92161620e-01 -2.17124596e-01 -5.20285070e-02
-1.01004891e-01 1.01406121e+00 -1.03635490e+00 -5.72185278e-01
-2.85504311e-01 3.80568683e-01 -2.05477402e-01 -1.01654500e-01
-3.18138413e-02 3.56139451e-01 -5.59151292e-01 8.61159563e-01
4.66512799e-01 -1.74238741e-01 3.74785691e-01 -1.95930466e-01
-4.81772214e-01 6.30949795e-01 -1.45581400e+00 1.91581917e+00
-4.80524004e-01 2.53610998e-01 2.17999458e-01 -1.24165192e-01
6.54454768e-01 7.00017333e-01 1.37311384e-01 -7.11975753e-01
-2.37970516e-01 7.14810073e-01 -1.88607678e-01 -3.66749495e-01
1.05280006e+00 2.11947739e-01 -4.39742744e-01 2.50842690e-01
3.16792995e-01 3.67342740e-01 2.98817128e-01 3.72411221e-01
1.38024187e+00 4.47575927e-01 2.43928239e-01 -1.36145011e-01
3.63409728e-01 2.04756126e-01 8.47929299e-01 1.07729912e+00
-1.83654591e-01 4.69666123e-01 -6.44226521e-02 -1.07006364e-01
-7.67491460e-01 -7.07147956e-01 -3.48183632e-01 1.73643255e+00
-1.39682114e-01 -8.38601172e-01 -5.96917093e-01 -8.54812264e-01
-2.94593006e-01 6.43002272e-01 -3.23081732e-01 4.51134890e-01
-9.01003540e-01 -1.00052834e+00 1.45470321e+00 4.55995351e-01
3.09827328e-01 -1.12937713e+00 1.75126046e-01 3.40943158e-01
-7.50803053e-01 -1.32996154e+00 -6.09367728e-01 3.76918167e-01
-4.17537451e-01 -9.01975811e-01 -6.98925197e-01 -7.00931847e-01
1.57089472e-01 9.80881900e-02 1.25370932e+00 1.19344778e-01
6.59788698e-02 3.76009762e-01 -7.96189427e-01 -2.47439414e-01
-1.50074646e-01 8.55619729e-01 3.13745700e-02 -6.94426417e-01
5.83409071e-01 -2.12560639e-01 -4.09320444e-01 5.67904532e-01
-7.60682344e-01 -1.92765623e-01 7.03262031e-01 8.89807582e-01
5.67829907e-01 -2.71610111e-01 9.38029408e-01 -9.84745741e-01
6.70474052e-01 -7.20475674e-01 -3.84481311e-01 6.66682720e-01
-5.88411987e-01 8.42890963e-02 3.63137603e-01 -8.86282623e-02
-1.26952839e+00 -1.81553781e-01 -4.35246199e-01 -4.75081876e-02
-2.14250296e-01 8.49380851e-01 -3.35740894e-01 3.12977165e-01
7.31133640e-01 1.52186275e-01 -8.44050109e-01 -1.21155632e+00
6.88899934e-01 9.23832715e-01 3.23456138e-01 -9.55832601e-01
5.48536897e-01 -6.57151490e-02 -7.84823060e-01 -6.24633372e-01
-1.04692864e+00 -9.59573925e-01 -5.96844316e-01 1.40218481e-01
7.48651266e-01 -1.31078029e+00 -4.11581844e-01 3.21148515e-01
-1.28591704e+00 -3.46561939e-01 6.61444664e-02 6.19767725e-01
-7.65691474e-02 2.51851499e-01 -9.76460516e-01 -6.52524292e-01
-5.07599115e-01 -9.99388635e-01 1.23737192e+00 6.06305376e-02
-9.82229188e-02 -1.03187227e+00 2.89252907e-01 7.30739415e-01
3.42899084e-01 -3.61580640e-01 7.53767073e-01 -1.29186845e+00
-4.82529074e-01 4.55468660e-03 -2.31119201e-01 1.30892247e-01
-1.14480391e-01 -2.87743032e-01 -9.47346807e-01 -3.55852783e-01
-5.54536641e-01 -2.76672691e-01 8.15707862e-01 -2.29178295e-01
6.24425530e-01 -1.93099573e-01 -4.34645891e-01 4.02557492e-01
1.30702019e+00 -6.73280563e-03 4.95686859e-01 7.90531099e-01
8.34171593e-01 5.76845229e-01 6.77989304e-01 4.78241332e-02
8.69015396e-01 6.94602311e-01 -1.28444145e-02 3.18946168e-02
-2.82514453e-01 -2.50449896e-01 5.14503658e-01 1.49758768e+00
-3.09864879e-01 -2.71352857e-01 -1.20992625e+00 8.48839521e-01
-1.97688067e+00 -8.38630736e-01 -2.15077892e-01 1.94250953e+00
1.21075547e+00 -2.34707400e-01 -3.33658308e-01 -5.00315189e-01
8.28800559e-01 3.12560648e-02 -1.63010985e-01 -6.77336454e-02
-4.19892460e-01 2.36077160e-01 7.78127015e-01 5.26299119e-01
-9.21717525e-01 1.57410336e+00 5.44904518e+00 1.19706869e+00
-6.70004666e-01 4.84433562e-01 -5.18358871e-02 1.14644840e-01
-2.49074876e-01 2.84311086e-01 -1.76948667e+00 3.34374607e-01
1.04848993e+00 -4.02731588e-03 3.57849032e-01 6.30052924e-01
-1.71010390e-01 1.59122035e-01 -9.27140594e-01 5.80792189e-01
-1.94355339e-01 -1.08371413e+00 -4.29292955e-02 1.57431886e-02
4.89705056e-01 7.58028269e-01 -2.56412894e-01 1.00791681e+00
7.52917051e-01 -9.24112797e-01 6.40174687e-01 5.39831400e-01
8.99819314e-01 -4.20058131e-01 1.03588414e+00 5.19866467e-01
-1.47922194e+00 2.38031000e-01 -2.15210348e-01 2.67098218e-01
3.87983739e-01 4.23694849e-01 -8.02125037e-01 1.11958754e+00
9.96224344e-01 2.72049159e-01 -6.86950326e-01 8.69634032e-01
-3.94754618e-01 5.81788838e-01 -3.50060582e-01 1.64473936e-01
1.78334817e-01 1.23292990e-01 8.05684745e-01 1.82303178e+00
3.01576167e-01 -1.19585562e-02 2.78072059e-01 3.80310833e-01
-5.47913909e-01 4.67389435e-01 -5.32483757e-01 1.08874433e-01
9.42906976e-01 1.33187747e+00 -2.67632663e-01 -2.49638319e-01
-4.11662221e-01 7.57800341e-01 6.98161244e-01 5.45914173e-01
-5.79268277e-01 -7.76782632e-01 6.00566387e-01 -1.65823355e-01
2.75400609e-01 -2.61049837e-01 1.14137679e-01 -1.52024209e+00
9.11208615e-02 -9.77407157e-01 7.15451002e-01 -6.89533591e-01
-1.43889439e+00 8.67493510e-01 -6.23231344e-02 -7.27491379e-01
-1.07672706e-01 -6.03931129e-01 8.52755606e-02 9.97208297e-01
-1.89792287e+00 -1.41669405e+00 2.42288664e-01 7.75788128e-01
4.90518004e-01 -4.30257358e-02 1.04051483e+00 9.98949230e-01
-7.47044265e-01 8.57114732e-01 2.21089333e-01 7.81781256e-01
1.35422349e+00 -1.39721501e+00 3.90030026e-01 8.37598503e-01
2.61070997e-01 1.29501867e+00 3.53089601e-01 -9.87743735e-01
-1.23386633e+00 -1.22062564e+00 1.48432541e+00 -1.00432408e+00
1.03980494e+00 -4.74434048e-01 -8.79020631e-01 1.10942352e+00
3.50933760e-01 -1.52211353e-01 8.64173114e-01 6.90215230e-01
-5.50084889e-01 1.56474516e-01 -8.00082147e-01 6.96283698e-01
1.28077757e+00 -7.44965553e-01 -1.00519788e+00 1.39620930e-01
1.16164446e+00 -4.60227817e-01 -1.45911109e+00 3.98041040e-01
4.68049109e-01 -3.37920725e-01 7.78276563e-01 -6.87566459e-01
-2.49256313e-01 -4.23744261e-01 -4.23260957e-01 -1.36854398e+00
-2.24378869e-01 -6.70218647e-01 6.64222687e-02 1.78615284e+00
1.09067047e+00 -7.22074449e-01 6.73471317e-02 4.39888149e-01
-3.33543777e-01 -2.54519910e-01 -1.06123328e+00 -7.75755823e-01
2.50220865e-01 -6.55522883e-01 6.71412885e-01 1.36243236e+00
5.94808161e-02 5.83930790e-01 -3.04813206e-01 3.76229763e-01
3.47149491e-01 -3.36217761e-01 4.78719920e-01 -1.13905394e+00
-7.22155273e-02 5.71194775e-02 3.02390307e-01 -9.78092551e-01
1.44913852e-01 -1.21687722e+00 1.70984510e-02 -1.61088145e+00
1.29309490e-01 -1.07548583e+00 -7.09456563e-01 1.02873003e+00
-2.36260056e-01 1.32010430e-01 3.67209464e-01 5.09460092e-01
-1.22314143e+00 1.89372599e-01 9.11345124e-01 3.04088473e-01
-1.98999748e-01 -3.34673733e-01 -9.30113614e-01 5.65634608e-01
5.37751913e-01 -8.45969141e-01 1.32398590e-01 -9.86470878e-01
8.99859250e-01 -1.07329851e-02 -1.31756812e-02 -6.55869603e-01
7.16273904e-01 1.47671521e-01 1.06619060e-01 -5.33225834e-01
9.80390459e-02 -5.63754618e-01 1.21100076e-01 -2.23591253e-01
-1.59326017e-01 4.10086930e-01 1.91045016e-01 2.69824147e-01
-3.02001774e-01 -2.30681583e-01 1.68025687e-01 -4.64756727e-01
-8.61385763e-01 2.23807499e-01 -3.68520528e-01 6.23424172e-01
2.33219385e-01 2.71480441e-01 -7.52760410e-01 1.84249580e-01
-9.34628248e-01 5.39854050e-01 1.99004576e-01 7.09055305e-01
1.49903968e-01 -1.22704566e+00 -8.88947845e-01 -3.36950988e-01
3.72906625e-01 -1.14015127e-02 2.14323834e-01 9.54602838e-01
-1.63711116e-01 8.22696745e-01 1.41458064e-01 -1.11718506e-01
-9.28505421e-01 9.55873504e-02 2.30124548e-01 -8.28544199e-01
-2.63013691e-01 6.88784361e-01 -1.54737622e-01 -1.22328973e+00
1.05456635e-01 -6.66546077e-02 -4.27112401e-01 3.18609536e-01
5.09488225e-01 5.99980652e-01 5.02704501e-01 -7.97381043e-01
-5.24131298e-01 3.65199953e-01 -4.69830424e-01 -2.77669221e-01
1.37717927e+00 -5.18984079e-01 -3.26862216e-01 3.10216904e-01
9.07439768e-01 4.95356530e-01 -4.45796251e-01 -5.32870293e-01
6.03063703e-01 1.74140602e-01 -7.12372884e-02 -1.40775537e+00
-6.62182331e-01 4.42339867e-01 2.87665755e-01 -5.72887138e-02
7.37210512e-01 1.41867399e-01 9.63802636e-01 8.36161017e-01
7.13145256e-01 -1.40667319e+00 -6.59485042e-01 1.29673886e+00
6.52525008e-01 -1.20922995e+00 -2.93201029e-01 -1.54440939e-01
-5.07198036e-01 7.75717616e-01 6.65814817e-01 5.46506464e-01
6.17274344e-01 3.17530364e-01 4.84743208e-01 -3.75743151e-01
-7.32597291e-01 -6.40533030e-01 1.43585265e-01 3.39484692e-01
7.23109841e-01 9.86547396e-02 -4.91444230e-01 1.05318594e+00
-3.66184533e-01 -2.50149071e-02 1.32450432e-01 9.56591189e-01
-2.70872384e-01 -1.33258390e+00 -2.87889749e-01 -4.60979976e-02
-1.02616334e+00 -9.20502484e-01 -3.23898971e-01 1.04920805e+00
2.80212075e-01 9.98014569e-01 -6.11624479e-01 -2.95965940e-01
6.74328208e-01 5.28906286e-01 -4.30216901e-02 -7.55088389e-01
-1.11914957e+00 5.51633611e-02 5.88213384e-01 -4.58424300e-01
-1.36609048e-01 -4.56423134e-01 -1.21957242e+00 -5.79236150e-02
-5.94108582e-01 5.12046039e-01 6.88170016e-01 8.24160635e-01
5.69998443e-01 2.58034110e-01 8.47212970e-02 -2.49272346e-01
-4.06333864e-01 -1.23021865e+00 -4.14648235e-01 1.04251474e-01
-9.04936194e-02 -4.37439740e-01 -3.10452670e-01 -2.65839726e-01] | [9.708683013916016, 9.554638862609863] |
54113c34-3613-451b-b1b4-f3b7e8017f68 | t-leap-occlusion-robust-pose-estimation-of | 2104.08029 | null | https://arxiv.org/abs/2104.08029v2 | https://arxiv.org/pdf/2104.08029v2.pdf | T-LEAP: Occlusion-robust pose estimation of walking cows using temporal information | As herd size on dairy farms continues to increase, automatic health monitoring of cows is gaining in interest. Lameness, a prevalent health disorder in dairy cows, is commonly detected by analyzing the gait of cows. A cow's gait can be tracked in videos using pose estimation models because models learn to automatically localize anatomical landmarks in images and videos. Most animal pose estimation models are static, that is, videos are processed frame by frame and do not use any temporal information. In this work, a static deep-learning model for animal-pose-estimation was extended to a temporal model that includes information from past frames. We compared the performance of the static and temporal pose estimation models. The data consisted of 1059 samples of 4 consecutive frames extracted from videos (30 fps) of 30 different dairy cows walking through an outdoor passageway. As farm environments are prone to occlusions, we tested the robustness of the static and temporal models by adding artificial occlusions to the videos.The experiments showed that, on non-occluded data, both static and temporal approaches achieved a Percentage of Correct Keypoints (PCKh@0.2) of 99%. On occluded data, our temporal approach outperformed the static one by up to 32.9%, suggesting that using temporal data was beneficial for pose estimation in environments prone to occlusions, such as dairy farms. The generalization capabilities of the temporal model was evaluated by testing it on data containing unknown cows (cows not present in the training set). The results showed that the average PCKh@0.2 was of 93.8% on known cows and 87.6% on unknown cows, indicating that the model was capable of generalizing well to new cows and that they could be easily fine-tuned to new herds. Finally, we showed that with harder tasks, such as occlusions and unknown cows, a deeper architecture was more beneficial. | ['Gert Kootstra', 'Rik van der Tol', 'Helena Russello'] | 2021-04-16 | null | null | null | null | ['animal-pose-estimation'] | ['computer-vision'] | [-8.77805725e-02 2.30088860e-01 -1.38728330e-02 -6.05414152e-01
-5.78520931e-02 -3.85528922e-01 -4.76895049e-02 4.68026608e-01
-3.77772629e-01 4.14333373e-01 -4.46894884e-01 1.45740539e-01
-1.14971861e-01 -6.67629778e-01 -1.32864821e+00 -6.64190948e-01
-9.77780044e-01 4.32309568e-01 5.96134961e-01 -3.12468648e-01
-2.18076110e-01 9.00702238e-01 -1.93622160e+00 4.56850559e-01
2.55199552e-01 7.99581885e-01 4.69528049e-01 7.68844306e-01
2.87973672e-01 4.72802967e-01 -6.05425775e-01 -3.80646646e-01
1.82513207e-01 -1.31620646e-01 -5.02155483e-01 4.21783291e-02
6.99015260e-01 -6.84213758e-01 -9.87676904e-02 5.51300228e-01
2.80181646e-01 -5.74438013e-02 2.92916864e-01 -1.12885344e+00
-2.10657567e-02 2.93514997e-01 -7.77503192e-01 3.02812964e-01
2.93542087e-01 1.15779385e-01 2.80934960e-01 -2.75336623e-01
1.02613842e+00 1.46480894e+00 1.13250375e+00 2.56491780e-01
-1.46630716e+00 -6.70943439e-01 1.72630306e-02 3.99574518e-01
-9.65380192e-01 -3.83720070e-01 2.80918747e-01 -5.71575046e-01
7.90806830e-01 -1.27292434e-02 8.29431474e-01 9.11256194e-01
7.28293836e-01 6.70108318e-01 8.08341265e-01 -2.65327930e-01
-9.06867720e-03 -4.11623763e-03 -1.00038216e-01 6.97740912e-01
6.61869228e-01 4.75906461e-01 -2.75334328e-01 -3.80381085e-02
9.17327166e-01 -2.75622308e-01 -7.91232660e-02 -1.15652955e+00
-1.21175158e+00 8.87487411e-01 5.31884432e-01 1.73298389e-01
-5.48025668e-01 1.84212416e-01 7.71516025e-01 3.30034405e-01
4.08605039e-01 6.12973630e-01 -8.43661189e-01 1.15592659e-01
-6.96492016e-01 3.80624473e-01 8.95889938e-01 8.62442315e-01
8.17310750e-01 -2.31568187e-01 -1.91559084e-02 5.14506102e-01
5.46636879e-02 8.47542882e-01 1.12220481e-01 -1.11771929e+00
-8.59303102e-02 3.03545117e-01 1.92716062e-01 -1.59311283e+00
-9.37062144e-01 -3.98726255e-01 -3.86335403e-01 1.34628773e-01
6.21940196e-01 -1.59417227e-01 -1.04635978e+00 1.65578210e+00
7.40233004e-01 3.32519598e-02 -2.25823075e-02 8.67523849e-01
8.46345425e-01 4.83399630e-01 1.00298151e-01 -3.43264520e-01
1.52856708e+00 -4.54337806e-01 -9.78094220e-01 -3.63482505e-01
8.99469972e-01 -8.79355073e-01 2.93817639e-01 3.30650598e-01
-7.97870100e-01 -8.06195796e-01 -1.12612760e+00 3.48568559e-01
-3.23214918e-01 2.81101942e-01 7.39992142e-01 3.80585998e-01
-9.07252908e-01 8.93442869e-01 -1.46415007e+00 -8.22574794e-01
2.42247507e-01 4.41602677e-01 -9.30307150e-01 -1.81876406e-01
-1.02981615e+00 1.37303627e+00 4.42662299e-01 4.21143472e-01
-8.17857444e-01 -6.38424039e-01 -1.10142767e+00 -3.22980493e-01
3.48651618e-01 -3.09332728e-01 1.18802023e+00 -8.59282553e-01
-1.33884728e+00 1.17177331e+00 -1.53761342e-01 -8.71228337e-01
9.00644422e-01 -8.00796032e-01 -2.46642217e-01 5.65261960e-01
1.19680434e-01 8.31759989e-01 7.57191241e-01 -1.27312529e+00
-5.64977586e-01 -3.92251670e-01 1.49629414e-01 -2.14575291e-01
2.72587776e-01 -1.69056058e-01 -2.43076444e-01 -2.14575917e-01
4.26979542e-01 -1.29788816e+00 -8.94699320e-02 3.46482366e-01
4.19781268e-01 3.62094223e-01 1.02932823e+00 -8.85314465e-01
7.17561364e-01 -2.01516509e+00 -1.02181725e-01 -1.23379439e-01
-1.35927498e-01 3.74943554e-01 -2.29900837e-01 2.98678815e-01
-1.09974653e-01 -6.30357444e-01 2.27350369e-01 2.45999306e-01
-3.73373568e-01 4.79931533e-01 3.24978054e-01 8.97066772e-01
4.22702640e-01 7.96984196e-01 -1.04036748e+00 -5.78494728e-01
5.28324783e-01 5.61909795e-01 -6.06889009e-01 1.76445365e-01
-2.51914095e-02 3.64000469e-01 -8.43115151e-02 6.33170784e-01
9.65715289e-01 1.18088655e-01 4.14049894e-01 -6.03863776e-01
9.96357128e-02 -4.96238977e-01 -1.17184114e+00 1.28790414e+00
-3.76801908e-01 8.07055831e-01 3.42571020e-01 -1.08466506e+00
8.15072060e-01 2.38849685e-01 7.90922523e-01 -3.95675182e-01
1.15925848e-01 1.83217674e-01 1.38578191e-01 -1.00612915e+00
3.75871599e-01 1.80926397e-01 8.78784060e-02 -2.50191122e-01
3.44368637e-01 -2.37275492e-02 6.15647137e-01 -3.59083086e-01
1.08057833e+00 6.23499990e-01 2.35415190e-01 -4.32838231e-01
1.35428816e-01 2.89850861e-01 4.38511044e-01 5.96242905e-01
-2.56214291e-01 5.50754488e-01 5.47625184e-01 -7.68402100e-01
-7.30107248e-01 -8.32352817e-01 -4.77622271e-01 1.03390288e+00
2.44179517e-01 -7.14844614e-02 -7.41729558e-01 -4.17945772e-01
5.22410572e-01 2.76380032e-01 -8.31879735e-01 -3.32755566e-01
-9.63987112e-01 -8.37148488e-01 2.16521934e-01 5.86391509e-01
4.42972243e-01 -9.73951638e-01 -1.41193461e+00 5.55120230e-01
-2.91967809e-01 -1.40020204e+00 2.95974165e-01 4.23757851e-01
-1.17949021e+00 -1.21304905e+00 -7.50485361e-01 -8.26057851e-01
5.33925176e-01 2.14577287e-01 9.93720591e-01 -2.12656543e-01
-5.38865089e-01 6.64731786e-02 -4.37561959e-01 -5.07220149e-01
-6.05962157e-01 -1.56202152e-01 1.08816549e-01 -2.73183078e-01
4.88816023e-01 -1.03805020e-01 -5.70435882e-01 6.80163026e-01
-7.20399618e-01 -2.34229773e-01 7.33155251e-01 9.74926770e-01
4.26733464e-01 -1.84108928e-01 2.30829343e-01 -7.78516591e-01
-3.26628327e-01 -3.98458153e-01 -7.25191712e-01 2.62256321e-02
1.36609077e-01 -4.13755402e-02 4.92748879e-02 -8.71915638e-01
-7.47873187e-01 3.76586586e-01 6.65957406e-02 -4.13711995e-01
-4.39062566e-01 3.98353010e-01 7.40027279e-02 -2.29797751e-01
6.55223906e-01 -5.24792075e-01 3.79725218e-01 -5.32835782e-01
-9.36682075e-02 3.35624032e-02 6.02727532e-01 -2.83146113e-01
4.42681432e-01 5.06215930e-01 1.41024560e-01 -1.09690452e+00
-4.26235795e-01 -4.92831528e-01 -8.12850416e-01 -4.56313372e-01
8.86073411e-01 -9.11982656e-01 -9.96772528e-01 6.63449526e-01
-1.08462965e+00 -3.98544341e-01 8.50176662e-02 1.13883984e+00
-7.43556440e-01 2.16282055e-01 -7.96226919e-01 -5.40751338e-01
1.56999111e-01 -1.00087106e+00 1.22413397e+00 -2.24428419e-02
-5.20309865e-01 -6.63444221e-01 -2.25096047e-01 2.55966872e-01
2.70712435e-01 8.70969892e-01 7.92679965e-01 -4.03566658e-01
-1.71058789e-01 -5.48360944e-01 -5.78273721e-02 1.11765392e-01
2.11931124e-01 2.57105619e-01 -9.06219780e-01 -5.80465138e-01
1.02250986e-01 -1.78142682e-01 6.40012860e-01 9.37082529e-01
6.50663733e-01 -6.86732829e-02 -6.90454364e-01 4.80422169e-01
1.15754652e+00 2.83891618e-01 5.18238485e-01 5.05509973e-01
3.61460149e-01 1.20314622e+00 1.30598712e+00 1.94742665e-01
-9.25728008e-02 8.28380704e-01 7.21778452e-01 -3.16466451e-01
1.13015041e-01 8.32338035e-02 3.54293704e-01 2.52363622e-01
-2.36286357e-01 -3.86577211e-02 -8.67769718e-01 6.79322600e-01
-1.70380378e+00 -8.63057971e-01 -2.25250572e-01 2.03070426e+00
5.33424973e-01 3.00115556e-01 -2.83696223e-02 -7.98123423e-03
7.94347823e-01 -1.64138839e-01 -4.75875378e-01 -4.85059202e-01
1.19243421e-01 1.59623593e-01 9.81123984e-01 -2.15303595e-03
-1.47544718e+00 6.51650846e-01 6.19072247e+00 3.33757997e-01
-1.22498393e+00 -1.98840007e-01 2.71545738e-01 -1.26779014e-02
8.32897246e-01 -3.81254733e-01 -7.64188886e-01 2.24466711e-01
8.59548688e-01 4.05585468e-01 -2.83081293e-01 1.07725430e+00
3.00687343e-01 -6.00412130e-01 -1.24846864e+00 3.91280413e-01
-2.23304793e-01 -8.42400968e-01 -4.21206027e-01 -1.83561623e-01
6.48384511e-01 -1.30168810e-01 -3.18752527e-01 3.60113293e-01
5.66489026e-02 -9.07553494e-01 6.24675274e-01 2.06686586e-01
4.96644825e-01 -5.84147155e-01 1.49522865e+00 3.33765864e-01
-1.21059239e+00 1.23040669e-01 -3.89556855e-01 1.98244154e-01
7.14711323e-02 3.20017189e-01 -8.64486992e-01 4.59407449e-01
1.08042848e+00 5.90261579e-01 -7.02197492e-01 1.10277069e+00
2.43226498e-01 1.82976812e-01 -7.47005403e-01 2.57509679e-01
1.80611014e-01 1.69029936e-01 3.50292534e-01 1.52698112e+00
3.78646970e-01 -5.45989443e-03 3.84366848e-02 2.05406189e-01
5.14627695e-01 -3.23556364e-02 -6.97008014e-01 1.53126657e-01
8.90868977e-02 1.04413605e+00 -1.00385499e+00 8.06929693e-02
-1.57039940e-01 6.24712884e-01 -1.49389118e-01 1.88148618e-01
-1.09530973e+00 -2.69773602e-01 6.33843660e-01 4.37692344e-01
6.72105789e-01 -8.70514512e-02 3.93228799e-01 -6.06302857e-01
1.05878077e-01 -7.30712414e-01 2.20217973e-01 -6.57641828e-01
-7.69375384e-01 3.12449366e-01 7.19940722e-01 -1.30536604e+00
-2.90651053e-01 -8.41504872e-01 2.32450459e-02 3.24961036e-01
-1.21454287e+00 -1.08523238e+00 -4.95769739e-01 -1.47779346e-01
3.54950339e-01 4.17552561e-01 7.37201869e-01 2.85405666e-01
-5.38956001e-02 4.88273859e-01 1.49912909e-01 -5.84755056e-02
7.24267900e-01 -7.12534666e-01 1.36394009e-01 5.44306159e-01
-3.93826693e-01 4.18254733e-01 1.23581874e+00 -7.63316274e-01
-1.46549392e+00 -1.13483548e+00 7.23454058e-01 -1.39513597e-01
4.40192163e-01 -1.38064399e-01 -1.15297639e+00 7.25463212e-01
-2.53508955e-01 2.25382432e-01 1.31843269e-01 -3.21344465e-01
2.61510104e-01 -1.77306086e-01 -1.38980889e+00 5.42952940e-02
9.63816762e-01 1.21472426e-01 -5.16250849e-01 2.64781415e-01
4.59463388e-01 -9.82822239e-01 -1.01635253e+00 1.15698254e+00
9.78907049e-01 -9.15578842e-01 1.04250205e+00 -4.59522963e-01
5.09666443e-01 -5.00976667e-02 4.98646647e-02 -9.97019053e-01
-2.00362965e-01 -1.20641790e-01 -2.56148458e-01 7.89719760e-01
-1.23515181e-01 -5.56261003e-01 7.48534203e-01 -3.12033556e-02
1.15470355e-02 -6.61395371e-01 -6.90297723e-01 -9.64083731e-01
-2.60283768e-01 -2.26313397e-02 3.62012118e-01 8.59817326e-01
-3.65212768e-01 -2.41419882e-01 -1.99174494e-01 4.90892023e-01
3.32557350e-01 1.37681007e-01 8.91962767e-01 -1.35368812e+00
-1.06543548e-01 1.04696371e-01 -9.94493842e-01 -6.91815495e-01
-2.64922399e-02 -4.42338884e-01 4.17975426e-01 -1.40707207e+00
-2.52948910e-01 -1.41635075e-01 1.68397903e-01 4.37611192e-01
-1.94391385e-02 2.21036330e-01 -6.61844388e-02 -1.87019363e-01
-6.56965524e-02 -1.56254113e-01 1.28364098e+00 1.00558773e-01
-3.28593910e-01 2.62127519e-01 2.92598963e-01 9.18023586e-01
5.91350377e-01 -5.15760541e-01 -8.47935826e-02 -3.29536021e-01
-1.39131412e-01 4.33833033e-01 4.74954545e-01 -1.13134551e+00
-3.74186099e-01 -3.92277502e-02 6.27791047e-01 -8.33549500e-01
5.37343264e-01 -9.97187793e-01 6.64726853e-01 1.13416374e+00
-4.47128601e-02 -1.47424549e-01 7.86161065e-01 5.26774466e-01
-9.82872993e-02 -1.07946716e-01 8.68185818e-01 -1.29074201e-01
-9.26137924e-01 -1.95445955e-01 -4.73673284e-01 -4.04448420e-01
1.35161746e+00 -5.92827916e-01 -2.85331067e-02 1.80943348e-02
-9.33897138e-01 3.88815105e-01 5.36366045e-01 6.53418958e-01
2.34149367e-01 -9.63799596e-01 -8.20387661e-01 4.91944820e-01
2.11845234e-01 2.46983152e-02 3.34318131e-01 1.17858088e+00
-1.32096529e+00 4.87135917e-01 -5.63895941e-01 -1.32333970e+00
-1.60571837e+00 5.94809175e-01 3.48241389e-01 3.79909901e-03
-5.76283991e-01 7.30269730e-01 3.89465004e-01 -2.51945853e-01
3.40206116e-01 -8.47478688e-01 -5.09523809e-01 4.54683095e-01
2.05404267e-01 4.78556693e-01 4.74125862e-01 -8.54418933e-01
-5.14706850e-01 8.41503918e-01 -2.98329312e-02 5.40126681e-01
1.45995355e+00 2.04163156e-02 2.72838563e-01 4.89696145e-01
1.34435964e+00 -4.29724246e-01 -1.27279890e+00 2.56328937e-02
1.82000667e-01 -4.99477297e-01 -2.81689376e-01 -3.07030737e-01
-1.23877084e+00 7.15391457e-01 1.14033508e+00 -1.81881562e-02
8.69129062e-01 -4.11268510e-02 5.12441754e-01 4.92688298e-01
6.84312761e-01 -7.18884826e-01 -6.34055287e-02 2.92740256e-01
7.76020348e-01 -1.45954573e+00 1.65523469e-01 -5.12515068e-01
-6.22848570e-01 1.23493123e+00 6.51081383e-01 -1.77992374e-01
4.30879921e-01 3.51978868e-01 1.77280188e-01 -3.30649108e-01
-5.43971121e-01 -6.14509471e-02 1.56678870e-01 8.67256701e-01
4.63638097e-01 5.50547540e-02 -3.62364829e-01 9.92560759e-02
-2.26363376e-01 2.31448010e-01 4.78664398e-01 1.36192369e+00
-4.11361694e-01 -4.10017073e-01 -4.93214786e-01 3.59242171e-01
-4.45871621e-01 6.23822451e-01 1.93086490e-01 1.21602786e+00
3.53119969e-01 7.80767798e-01 1.70693532e-01 5.21151023e-03
6.88024879e-01 -1.23018660e-01 6.67007506e-01 -6.67735279e-01
-5.15796542e-01 8.86773765e-02 1.72310531e-01 -8.61312628e-01
-6.28707945e-01 -6.97115183e-01 -1.09200370e+00 -4.05900985e-01
-6.62264466e-01 4.56057163e-03 6.50541723e-01 5.43374002e-01
-1.81047115e-02 6.11426413e-01 2.35046506e-01 -1.31732655e+00
-3.94105583e-01 -9.56131756e-01 -5.82908571e-01 2.16282412e-01
6.72171116e-01 -1.10330856e+00 -3.14223081e-01 5.31426907e-01] | [7.678747177124023, -0.9594832062721252] |
ff1e56be-5bae-4056-8cd9-afb93ed124b2 | kest-kernel-distance-based-efficient-self | 2306.10414 | null | https://arxiv.org/abs/2306.10414v1 | https://arxiv.org/pdf/2306.10414v1.pdf | KEST: Kernel Distance Based Efficient Self-Training for Improving Controllable Text Generation | Self-training (ST) has come to fruition in language understanding tasks by producing pseudo labels, which reduces the labeling bottleneck of language model fine-tuning. Nevertheless, in facilitating semi-supervised controllable language generation, ST faces two key challenges. First, augmented by self-generated pseudo text, generation models tend to over-exploit the previously learned text distribution, suffering from mode collapse and poor generation diversity. Second, generating pseudo text in each iteration is time-consuming, severely decelerating the training process. In this work, we propose KEST, a novel and efficient self-training framework to handle these problems. KEST utilizes a kernel-based loss, rather than standard cross entropy, to learn from the soft pseudo text produced by a shared non-autoregressive generator. We demonstrate both theoretically and empirically that KEST can benefit from more diverse pseudo text in an efficient manner, which allows not only refining and exploiting the previously fitted distribution but also enhanced exploration towards a larger potential text space, providing a guarantee of improved performance. Experiments on three controllable generation tasks demonstrate that KEST significantly improves control accuracy while maintaining comparable text fluency and generation diversity against several strong baselines. | ['Xing Xie', 'Laks V. S. Lakshmanan', 'Xiaoyuan Yi', 'Yuxi Feng'] | 2023-06-17 | null | null | null | null | ['text-generation'] | ['natural-language-processing'] | [ 3.57497662e-01 3.76815110e-01 -2.70790756e-01 -1.83714807e-01
-9.18826878e-01 -6.35046601e-01 7.78398693e-01 -7.64911398e-02
-2.40935862e-01 1.13959110e+00 4.79090720e-01 -7.33974949e-02
1.99624047e-01 -7.11067617e-01 -6.04807734e-01 -6.23309731e-01
3.06376755e-01 7.64564812e-01 -2.81193256e-01 -3.56148094e-01
7.21666403e-03 -9.51442793e-02 -1.41389394e+00 1.91563591e-02
1.61629462e+00 5.96814215e-01 5.53295612e-01 4.71588820e-01
-1.19994804e-01 8.27591121e-01 -7.36406863e-01 -4.89108801e-01
4.46534678e-02 -7.94627488e-01 -7.74728119e-01 -1.06449530e-01
3.49337190e-01 -2.64934301e-01 -1.09605402e-01 9.27010655e-01
5.78280389e-01 3.34336728e-01 7.48396397e-01 -1.09357417e+00
-8.04807842e-01 1.27283049e+00 -1.93376124e-01 -2.35507205e-01
1.06428608e-01 4.55166876e-01 1.01111233e+00 -5.71839154e-01
6.56244040e-01 1.36537790e+00 5.18111408e-01 9.85958040e-01
-1.63017988e+00 -8.30203831e-01 3.76668833e-02 -5.45705616e-01
-1.21484542e+00 -5.30028224e-01 6.56540871e-01 -3.94985020e-01
8.96663249e-01 9.78095680e-02 6.71995401e-01 1.53929543e+00
2.24609211e-01 1.05787790e+00 1.17584252e+00 -3.77703607e-01
2.90830463e-01 3.76259983e-01 -2.63103545e-01 8.10926139e-01
1.14827834e-01 2.95902848e-01 -7.50446677e-01 -7.10355863e-02
6.44876719e-01 -5.38879514e-01 -4.09094334e-01 -5.44880450e-01
-1.39710605e+00 1.02184677e+00 2.39817291e-01 3.35453898e-01
-1.83293551e-01 2.92881638e-01 4.41470891e-01 2.46573284e-01
6.20822549e-01 9.32120681e-01 -4.92081076e-01 -5.43559670e-01
-1.02591431e+00 3.58504534e-01 6.19251490e-01 9.94425178e-01
6.69143081e-01 5.58231294e-01 -6.43712580e-01 8.82848501e-01
9.85050425e-02 7.33134568e-01 1.04776549e+00 -6.81868672e-01
7.61129975e-01 5.15889764e-01 -2.08056741e-03 -4.88356978e-01
-2.91325599e-01 -7.88632333e-01 -1.05733514e+00 1.17867710e-02
5.02753079e-01 -3.90010178e-01 -9.56063509e-01 2.41210365e+00
1.38292074e-01 -1.39334336e-01 1.18212834e-01 3.65539283e-01
2.08052903e-01 4.79749650e-01 2.84232050e-01 -2.16186389e-01
8.28568101e-01 -1.12507689e+00 -6.64769351e-01 -3.28141361e-01
8.55990410e-01 -4.87445772e-01 1.47999918e+00 2.98093319e-01
-1.22578657e+00 -6.25453234e-01 -8.39593351e-01 -7.44392201e-02
-2.28914604e-01 3.03416729e-01 7.13570416e-01 7.13425279e-01
-1.17576718e+00 7.88218141e-01 -8.47880244e-01 3.81285883e-03
5.14980018e-01 3.08352917e-01 -1.71303287e-01 2.96088099e-01
-1.34219384e+00 8.41808975e-01 8.16414177e-01 -2.41063356e-01
-7.03723133e-01 -9.87660170e-01 -1.10990834e+00 8.83043632e-02
3.92939717e-01 -1.06019831e+00 1.42822754e+00 -8.28113914e-01
-2.01258564e+00 3.84327710e-01 -5.91624901e-02 -6.83241904e-01
9.19672668e-01 -3.41579318e-01 -1.12538978e-01 -2.30575800e-01
2.40119219e-01 1.02755296e+00 1.20555294e+00 -9.70710814e-01
-2.96812922e-01 -1.57602087e-01 -3.95857841e-01 4.22462165e-01
-5.38017213e-01 -6.70834303e-01 -5.65819927e-02 -9.63461161e-01
-3.37438762e-01 -1.02516890e+00 -3.29258293e-01 -4.78718579e-01
-6.45965815e-01 -3.30663800e-01 3.94576639e-01 -4.21791315e-01
1.32275856e+00 -1.76836562e+00 2.87006617e-01 -5.00570275e-02
3.56054097e-01 4.46416318e-01 -1.79467261e-01 2.69961655e-01
1.65846378e-01 3.29876721e-01 -2.59833872e-01 -6.63335562e-01
2.22011283e-01 2.94262543e-02 -6.82534993e-01 -9.46800262e-02
1.52826935e-01 1.33983076e+00 -1.28079331e+00 -4.08363432e-01
2.18186304e-01 5.12160480e-01 -6.45197809e-01 2.61617918e-02
-6.89828038e-01 6.10972106e-01 -4.10004348e-01 1.36689723e-01
2.27472901e-01 -3.19763422e-01 1.39675243e-03 1.81163445e-01
1.87565893e-01 3.40192586e-01 -8.11395586e-01 1.96262956e+00
-9.59282160e-01 4.68344420e-01 -4.15854365e-01 -7.23725915e-01
1.00835502e+00 7.24888891e-02 8.71492922e-02 -6.51134133e-01
1.19928330e-01 2.25798681e-01 -2.35368311e-01 -8.25019702e-02
7.38814831e-01 -8.09853971e-02 -1.80623129e-01 6.98089480e-01
1.47498921e-01 -6.14513397e-01 3.39987040e-01 2.40742803e-01
6.91879451e-01 5.10916591e-01 2.76816875e-01 -2.32586861e-01
2.74754107e-01 -5.10317832e-02 2.85835952e-01 8.84803534e-01
1.14115827e-01 3.07732493e-01 3.77667367e-01 1.37483448e-01
-1.09123862e+00 -1.27272558e+00 3.51762474e-02 9.81432855e-01
-2.00721875e-01 -5.34041584e-01 -8.64336610e-01 -8.06018829e-01
2.69390643e-02 1.22353697e+00 -4.48300451e-01 -6.49717033e-01
-6.05330765e-01 -5.60160458e-01 6.14449441e-01 5.69269121e-01
5.23944378e-01 -1.14297652e+00 -4.36583608e-01 4.69174713e-01
-3.54200125e-01 -8.16599548e-01 -8.77254546e-01 1.84132412e-01
-9.69809949e-01 -3.86104494e-01 -8.12470734e-01 -6.55097008e-01
6.84588253e-01 -5.34384139e-02 1.30286670e+00 -2.94341266e-01
-1.30427882e-01 7.11788833e-02 -1.31745964e-01 -2.20795229e-01
-9.90148306e-01 5.88798106e-01 1.01264782e-01 -3.40682626e-01
-3.49980667e-02 -4.90521908e-01 -4.02728796e-01 2.25244518e-02
-7.39719748e-01 5.77819586e-01 6.40930414e-01 1.28797793e+00
4.58830923e-01 1.12709478e-01 8.28520954e-01 -1.00127149e+00
1.00239539e+00 -2.70167857e-01 -5.68764210e-01 2.15210423e-01
-1.04907811e+00 6.82268143e-01 9.55357313e-01 -6.07482493e-01
-1.30813062e+00 2.38801483e-02 1.78977549e-01 -3.66340339e-01
1.33979067e-01 3.86850655e-01 4.37184609e-02 4.42125738e-01
8.00550103e-01 5.82855642e-01 1.63734376e-01 -2.46807352e-01
8.27673078e-01 5.29831529e-01 5.25864899e-01 -6.80602491e-01
8.96276236e-01 5.38289286e-02 -2.58677065e-01 -4.80171323e-01
-9.22968626e-01 -3.14185135e-02 -3.77687573e-01 1.88072771e-01
6.70470178e-01 -9.55006123e-01 -4.50974822e-01 5.56078970e-01
-8.02254379e-01 -9.27757204e-01 -6.09848976e-01 3.53547513e-01
-8.75813723e-01 6.09757043e-02 -5.38474143e-01 -7.92012572e-01
-6.38647795e-01 -1.08890522e+00 1.32805145e+00 1.14267811e-01
-4.92528647e-01 -1.17176640e+00 1.39366359e-01 3.92056584e-01
6.25481069e-01 7.61974007e-02 9.33151960e-01 -5.78877807e-01
-3.62077385e-01 -3.33894901e-02 1.58293977e-01 4.12557453e-01
3.49642128e-01 -3.20553511e-01 -8.72154713e-01 -3.78082395e-01
-1.29810274e-01 -6.57693744e-01 9.25055087e-01 1.83562070e-01
1.09108090e+00 -4.82688814e-01 -2.93593585e-01 5.80415487e-01
9.25592661e-01 -1.33763775e-01 2.96146721e-01 3.20175141e-02
7.64632225e-01 4.14894462e-01 4.60730433e-01 3.31169933e-01
3.24265122e-01 6.22863770e-01 -1.69444799e-01 -5.78253809e-03
-3.47088546e-01 -9.68051493e-01 6.23327374e-01 7.79010296e-01
4.46732670e-01 -5.18675208e-01 -7.56629348e-01 3.63186389e-01
-1.77670169e+00 -8.94760072e-01 2.91988522e-01 2.39335704e+00
1.53535557e+00 2.64414072e-01 1.14862978e-01 -3.86878885e-02
5.95767379e-01 4.34547774e-02 -8.75099421e-01 -2.57108003e-01
-1.47830173e-01 4.26509380e-01 2.53317356e-01 6.25102043e-01
-9.33118999e-01 1.42120206e+00 6.12015390e+00 1.17738712e+00
-1.15960801e+00 -4.79330905e-02 7.36808300e-01 -1.87522307e-01
-5.17598689e-01 -2.36582369e-01 -9.09829140e-01 6.44088507e-01
1.00588858e+00 -5.73784351e-01 6.74839497e-01 9.92644370e-01
1.82464138e-01 8.20244029e-02 -1.09565628e+00 7.97811389e-01
1.17592663e-02 -1.23661280e+00 3.94075334e-01 -1.08409934e-02
1.07250750e+00 4.43815690e-04 3.94440383e-01 6.77651644e-01
8.17701578e-01 -9.89854574e-01 8.93671453e-01 1.82669058e-01
1.08861029e+00 -7.16596961e-01 2.15001598e-01 5.63700914e-01
-9.53865230e-01 9.60020535e-03 -1.74949244e-01 2.75816679e-01
2.13159248e-01 6.61915958e-01 -1.18068302e+00 2.96487838e-01
-1.34925395e-02 4.83196229e-01 -5.14999211e-01 4.78607535e-01
-5.50395012e-01 6.04337752e-01 -9.11736637e-02 -1.86558232e-01
1.99409589e-01 -1.70484304e-01 6.27842665e-01 1.02732074e+00
3.49610299e-01 -4.96631354e-01 2.71746755e-01 1.42555845e+00
-1.61424547e-01 7.25808293e-02 -5.86797178e-01 -5.27519584e-01
5.13583720e-01 1.02605855e+00 -4.37178850e-01 -4.99196291e-01
2.22981483e-01 1.23688841e+00 5.56112707e-01 2.85430133e-01
-7.63938725e-01 -4.35900569e-01 3.89928430e-01 -1.47388637e-01
5.22204302e-02 -3.55938643e-01 -3.60871494e-01 -1.22103024e+00
-1.02867857e-01 -9.09401894e-01 -2.44966336e-02 -5.77886343e-01
-1.09348273e+00 8.14246118e-01 -1.84761547e-02 -1.02005041e+00
-9.97896135e-01 -1.51139781e-01 -2.86204278e-01 1.03616524e+00
-1.31828642e+00 -1.23970044e+00 -1.18161760e-01 3.29620093e-01
7.77620673e-01 -3.09017420e-01 7.63480902e-01 -7.81905353e-02
-5.37099898e-01 1.07790494e+00 1.67396441e-01 -2.86727071e-01
8.28807414e-01 -1.63579977e+00 7.84319043e-01 5.23582757e-01
2.39349067e-01 6.62545145e-01 6.51087642e-01 -9.00616288e-01
-1.03067899e+00 -1.31321836e+00 9.10082638e-01 -4.94233906e-01
6.77091360e-01 -7.06542075e-01 -7.02888310e-01 5.71737885e-01
-7.93534294e-02 -5.03258944e-01 4.34976071e-01 8.84965807e-03
-3.61805022e-01 1.40279546e-01 -9.76868510e-01 9.83959854e-01
1.20916510e+00 -5.76486528e-01 -3.53590161e-01 4.26131994e-01
1.00726306e+00 -5.74531734e-01 -5.91738939e-01 1.68845475e-01
3.29818606e-01 -6.35250151e-01 7.78244078e-01 -5.22341788e-01
5.02377212e-01 5.16023114e-02 2.71984726e-01 -1.94464493e+00
-1.31014824e-01 -1.31806517e+00 -2.90247798e-01 1.38262343e+00
6.40281737e-01 -7.19290733e-01 7.60938048e-01 4.92903590e-01
-2.45242521e-01 -6.94185436e-01 -6.24322474e-01 -1.06868076e+00
4.84684587e-01 -2.93521971e-01 7.12454796e-01 7.19949722e-01
2.14078739e-01 6.07320011e-01 -4.66777265e-01 -4.96869981e-01
5.50168693e-01 -6.00533523e-02 7.97654569e-01 -1.03538549e+00
-5.05919516e-01 -7.54618049e-01 1.23463646e-01 -1.24301517e+00
5.23542225e-01 -1.18709683e+00 2.99456865e-01 -1.23740649e+00
1.21521905e-01 -7.01878428e-01 1.17028542e-01 4.29501981e-01
-6.03421092e-01 7.15174526e-02 1.01495445e-01 3.97514030e-02
-4.66670007e-01 1.02196395e+00 1.55701733e+00 -1.57507762e-01
-6.20788038e-01 1.72819986e-04 -8.95502627e-01 4.88462806e-01
1.08855021e+00 -1.81735009e-01 -9.01571989e-01 -4.17669326e-01
1.90197036e-01 -5.94162475e-03 5.60740158e-02 -9.28551555e-01
-6.03314899e-02 -4.87716347e-02 3.19805056e-01 -2.87682056e-01
2.54567564e-01 -1.43354371e-01 -1.45476356e-01 5.58606029e-01
-7.53366649e-01 -1.39197886e-01 2.58417279e-01 6.39760256e-01
1.11245140e-01 -8.94484520e-02 7.46233463e-01 -1.01453073e-01
-2.02888012e-01 2.89040357e-01 -2.34889477e-01 4.34555501e-01
7.55534053e-01 2.08359249e-02 -3.22442621e-01 -4.58127201e-01
-5.71398795e-01 3.72770727e-01 3.50942224e-01 6.60164297e-01
2.25666195e-01 -1.39964128e+00 -7.07619131e-01 3.79936069e-01
3.17275822e-02 1.10753052e-01 2.12560132e-01 3.78896773e-01
-5.54718543e-03 5.80344617e-01 1.54720068e-01 -5.35892069e-01
-7.97886610e-01 4.07729059e-01 1.50283799e-01 -7.05912113e-01
-6.36238873e-01 8.81000042e-01 2.44342133e-01 -4.52449530e-01
1.83708400e-01 -4.30901706e-01 7.70963281e-02 -9.33545828e-02
4.58473951e-01 3.33798289e-01 -1.04466558e-01 -2.58079857e-01
2.90335655e-01 2.50049055e-01 -3.56250316e-01 -4.09795344e-01
9.07898009e-01 -1.99193582e-01 2.49431595e-01 4.52912748e-01
9.32716548e-01 -7.23570772e-03 -1.51710021e+00 -3.47318709e-01
-1.01397708e-01 -2.57262528e-01 2.19815910e-01 -1.15875411e+00
-8.33016753e-01 6.49513483e-01 3.38047683e-01 -1.26913693e-02
7.96546698e-01 -1.39738202e-01 1.09591699e+00 3.46920907e-01
3.93117309e-01 -1.20711982e+00 4.27010328e-01 5.34339070e-01
9.05923009e-01 -1.13518846e+00 -4.12449628e-01 -2.11467788e-01
-9.75236714e-01 7.71674156e-01 7.68595695e-01 3.10406238e-01
1.91552535e-01 1.63296416e-01 -1.09702952e-01 2.03259915e-01
-8.72355103e-01 -4.08824682e-02 2.67356396e-01 6.68809354e-01
5.06888211e-01 2.21222475e-01 -1.00835323e-01 4.32996780e-01
-8.95158172e-01 -7.33095314e-03 3.36370617e-01 4.71088469e-01
-3.24618250e-01 -1.37062490e+00 4.00440767e-02 6.44269347e-01
-1.17103525e-01 -4.26576465e-01 -3.02088439e-01 5.84154367e-01
-5.65199517e-02 9.04924214e-01 1.67208370e-02 -2.12306827e-01
-4.44373898e-02 2.37343743e-01 4.80487078e-01 -7.29730666e-01
-4.42150295e-01 -7.83762056e-03 1.42112570e-02 -4.07062799e-01
2.28949621e-01 -6.28493130e-01 -1.31247580e+00 -1.18915088e-01
-6.06697977e-01 1.99660227e-01 5.65328538e-01 8.23630333e-01
5.08089662e-01 6.30471230e-01 6.83289766e-01 -6.19564652e-01
-1.23516667e+00 -1.05258262e+00 -3.82887781e-01 4.11182642e-01
2.92835116e-01 -6.74829245e-01 -3.28582883e-01 -1.95648931e-02] | [11.831191062927246, 9.187782287597656] |
dfed8df9-6dcc-494e-a49d-9594f9e83b77 | knowledge-aware-bayesian-deep-topic-model | 2209.14228 | null | https://arxiv.org/abs/2209.14228v1 | https://arxiv.org/pdf/2209.14228v1.pdf | Knowledge-Aware Bayesian Deep Topic Model | We propose a Bayesian generative model for incorporating prior domain knowledge into hierarchical topic modeling. Although embedded topic models (ETMs) and its variants have gained promising performance in text analysis, they mainly focus on mining word co-occurrence patterns, ignoring potentially easy-to-obtain prior topic hierarchies that could help enhance topic coherence. While several knowledge-based topic models have recently been proposed, they are either only applicable to shallow hierarchies or sensitive to the quality of the provided prior knowledge. To this end, we develop a novel deep ETM that jointly models the documents and the given prior knowledge by embedding the words and topics into the same space. Guided by the provided knowledge, the proposed model tends to discover topic hierarchies that are organized into interpretable taxonomies. Besides, with a technique for adapting a given graph, our extended version allows the provided prior topic structure to be finetuned to match the target corpus. Extensive experiments show that our proposed model efficiently integrates the prior knowledge and improves both hierarchical topic discovery and document representation. | ['Mingyuan Zhou', 'Bo Chen', 'Chaojie Wang', 'Zhibin Duan', 'Miaoge Li', 'Yishi Xu', 'Dongsheng Wang'] | 2022-09-20 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-7.86655396e-02 5.56540430e-01 -4.89913791e-01 -4.71041024e-01
-5.85123956e-01 -3.23807508e-01 7.59045720e-01 4.02107418e-01
2.26566680e-02 4.90649968e-01 5.51537573e-01 -5.10197692e-02
-4.25194800e-01 -1.16042554e+00 -3.42010140e-01 -4.66988564e-01
-5.51530123e-02 9.81300890e-01 5.06341875e-01 -1.97621454e-02
3.24992567e-01 -2.81739533e-01 -1.37774980e+00 7.76596740e-02
1.32904053e+00 6.20257437e-01 6.13084435e-01 2.41794810e-01
-4.37301010e-01 3.65963846e-01 -6.14530861e-01 -2.38715515e-01
-2.25786150e-01 -2.91331530e-01 -7.75671363e-01 5.36217570e-01
7.15877786e-02 -1.59051254e-01 -3.29342365e-01 9.04239535e-01
5.37431892e-03 2.46121883e-01 8.65907550e-01 -1.20699012e+00
-4.55603778e-01 1.05722785e+00 -6.94693565e-01 -5.78704942e-03
3.46947424e-02 -4.05962527e-01 1.35764349e+00 -6.45175099e-01
5.87642372e-01 1.40470028e+00 2.83897907e-01 7.84143582e-02
-1.31066918e+00 -6.62026227e-01 7.54722178e-01 2.37078935e-01
-1.45375431e+00 6.81267679e-02 1.11445844e+00 -4.16197807e-01
6.89846635e-01 -1.04228474e-01 7.77803004e-01 1.08661675e+00
1.21633537e-01 8.23939681e-01 9.33912575e-01 -3.45427811e-01
5.27463198e-01 5.46198666e-01 5.91970444e-01 4.51005071e-01
6.17175519e-01 -7.06408143e-01 -7.58656859e-01 -5.39399445e-01
7.17016399e-01 1.74138367e-01 -3.31890494e-01 -7.93561041e-01
-9.74775672e-01 1.21781981e+00 2.16150120e-01 5.29274940e-01
-5.62491179e-01 -1.02734238e-01 2.14083970e-01 -1.77320078e-01
8.02866399e-01 4.01466101e-01 -2.57427990e-01 1.36048660e-01
-1.38567197e+00 2.53515303e-01 9.02983308e-01 1.17662358e+00
1.11414361e+00 -2.01443791e-01 -1.98005810e-01 7.66404152e-01
6.68367684e-01 3.89151052e-02 5.80303013e-01 -3.84858966e-01
3.88740391e-01 9.62270737e-01 4.03164551e-02 -1.36655962e+00
-2.76624888e-01 -7.19459355e-01 -6.78085208e-01 -5.50218225e-01
1.81172006e-02 5.45196906e-02 -9.78960633e-01 1.66121697e+00
5.06594002e-01 1.03744231e-01 -1.87033534e-01 5.64642012e-01
5.15712023e-01 9.40406680e-01 1.75005883e-01 -1.74412578e-01
1.58718634e+00 -7.28620291e-01 -9.51989889e-01 -2.65804350e-01
3.55445683e-01 -4.07568723e-01 9.68605578e-01 3.58499229e-01
-8.25774848e-01 -2.92924225e-01 -9.05306458e-01 -1.33081928e-01
-2.73361683e-01 1.96353775e-02 5.80709457e-01 6.66412771e-01
-9.80850220e-01 1.63498029e-01 -1.09002984e+00 -5.59837103e-01
3.35087091e-01 2.66861737e-01 2.03519478e-01 -6.15649857e-02
-1.39447343e+00 4.80326772e-01 7.40644813e-01 -2.81454504e-01
-1.06434989e+00 -5.82325697e-01 -6.45951271e-01 5.33244014e-01
8.13819349e-01 -1.03138947e+00 1.00748634e+00 -1.85575753e-01
-1.47721124e+00 2.82098323e-01 -5.11530161e-01 -6.13056183e-01
2.91178543e-02 -2.86746502e-01 -9.90218818e-02 2.15348706e-01
2.97283322e-01 6.02663517e-01 9.00253713e-01 -1.37265062e+00
-7.27505028e-01 -5.20081103e-01 1.61402106e-01 3.24203938e-01
-9.77322340e-01 -2.34223753e-01 -7.20617652e-01 -7.12447822e-01
4.64494318e-01 -6.46199942e-01 -3.19349796e-01 -4.82680112e-01
-6.99736536e-01 -6.64983749e-01 1.03213573e+00 -4.48834777e-01
1.63490069e+00 -1.77381289e+00 2.96742558e-01 3.46085757e-01
5.42010367e-01 -3.96914601e-01 3.28174055e-01 6.76224649e-01
4.49542791e-01 3.48586231e-01 -1.85229495e-01 -4.73909110e-01
1.96294203e-01 1.87996507e-01 -6.33943856e-01 1.82688266e-01
-2.13779584e-01 5.84412813e-01 -8.19558203e-01 -8.29578996e-01
-2.04625316e-02 5.44588983e-01 -8.21952641e-01 1.25626594e-01
-4.96372342e-01 6.45028055e-02 -7.35578358e-01 2.08721399e-01
5.57077408e-01 -7.26727128e-01 5.14857233e-01 -8.87746960e-02
2.59935617e-01 7.40647316e-01 -1.25166333e+00 1.66579556e+00
-4.64764059e-01 4.72183049e-01 -6.74176263e-03 -8.66341174e-01
1.21741927e+00 4.06758308e-01 4.87185478e-01 4.41348180e-02
-1.80106491e-01 -2.33905926e-01 -1.78741559e-01 -3.69376619e-03
9.69872773e-01 -6.83989376e-02 4.56495807e-02 6.67224407e-01
3.14908832e-01 -1.37853235e-01 2.66291767e-01 7.15805888e-01
6.63638651e-01 -1.00139305e-01 5.60058355e-01 -5.31516016e-01
6.78757653e-02 9.03813988e-02 5.80821276e-01 8.26262236e-01
3.14759344e-01 3.73963565e-01 5.91239512e-01 -4.16468307e-02
-8.55257750e-01 -9.14874554e-01 -2.55577594e-01 1.27378571e+00
1.98513240e-01 -1.04652822e+00 -5.39661825e-01 -6.69600308e-01
-2.51728088e-01 8.47890496e-01 -7.22863197e-01 2.45996639e-02
-1.62421823e-01 -8.07147563e-01 5.28155081e-02 4.24229443e-01
3.98604333e-01 -4.63887960e-01 -4.68867183e-01 4.77894753e-01
-3.88776004e-01 -8.45987022e-01 -3.90432119e-01 1.70328721e-01
-1.16080403e+00 -7.86634624e-01 -7.61522412e-01 -5.16335368e-01
5.99827588e-01 3.96246791e-01 1.03968561e+00 -2.93433934e-01
2.31301770e-01 4.08546597e-01 -4.70240384e-01 -3.93166810e-01
-1.65538326e-01 5.88999987e-01 -1.52384639e-01 8.36566538e-02
4.79519129e-01 -8.17243040e-01 -6.22394145e-01 4.10027027e-01
-1.07769525e+00 2.36230135e-01 4.28512990e-01 7.76087165e-01
2.67783672e-01 5.67025006e-01 5.84586084e-01 -1.05185735e+00
7.49137342e-01 -7.06216872e-01 -5.62639594e-01 2.03505650e-01
-1.00489771e+00 1.60394400e-01 1.58090562e-01 -4.71391112e-01
-1.39138389e+00 -5.13374329e-01 3.54405463e-01 -2.43232965e-01
-2.20111117e-01 9.20706570e-01 -3.67049277e-01 6.90576494e-01
3.04906100e-01 2.58539349e-01 -5.43291092e-01 -6.43301010e-01
5.20305693e-01 5.28182030e-01 9.84382182e-02 -7.66556382e-01
6.11498713e-01 6.83176339e-01 -2.94274390e-01 -7.81960309e-01
-9.13984001e-01 -8.37380171e-01 -6.90489233e-01 1.53639942e-01
6.49115860e-01 -9.87920105e-01 -7.75376931e-02 -7.96496049e-02
-1.21607244e+00 8.10443535e-02 -5.08143343e-02 4.68186468e-01
-2.89035022e-01 5.45379102e-01 -3.93031359e-01 -9.69021618e-01
-2.66762078e-01 -9.85476434e-01 1.27131820e+00 1.37643367e-01
-4.17517155e-01 -1.44410276e+00 2.23004982e-01 1.15591690e-01
3.47856998e-01 -1.54096708e-01 1.25809503e+00 -1.05839980e+00
-7.14427352e-01 -7.72323310e-02 -2.54189253e-01 -3.35519642e-01
4.51907694e-01 -2.14237228e-01 -7.45944560e-01 -2.54102260e-01
1.10411741e-01 4.06508334e-02 1.10676968e+00 4.86711144e-01
9.62707222e-01 -6.56883657e-01 -8.76050413e-01 2.24869922e-01
1.16645312e+00 3.59090343e-02 3.31757337e-01 4.89581108e-01
6.10995471e-01 7.83022225e-01 3.95378888e-01 6.76625907e-01
8.06100011e-01 7.63485968e-01 2.60873407e-01 1.25078157e-01
2.68222153e-01 -4.99219507e-01 9.02811214e-02 9.74001527e-01
1.50341153e-01 -4.65802073e-01 -9.71904278e-01 9.52021360e-01
-1.85402834e+00 -7.41606355e-01 6.03246354e-02 1.97331071e+00
1.08225286e+00 2.74846405e-01 1.62003532e-01 -1.07662402e-01
8.59293997e-01 3.38434696e-01 -4.37165618e-01 2.04702422e-01
6.81985766e-02 -8.64729285e-02 7.29560060e-03 3.89513195e-01
-9.38542962e-01 1.08959079e+00 5.85335112e+00 9.45530117e-01
-7.20742881e-01 1.21765897e-01 3.57961804e-01 1.50719345e-01
-8.12871337e-01 4.49112892e-01 -1.12103403e+00 2.39236578e-01
6.74018979e-01 -6.68312311e-01 -2.48137176e-01 1.11452472e+00
1.74004629e-01 -1.04879446e-01 -1.00843322e+00 3.99556011e-01
-3.22800092e-02 -1.13715029e+00 5.43542206e-01 4.85739052e-01
8.84762824e-01 -3.74041528e-01 2.07545608e-02 4.11426127e-01
5.73725104e-01 -5.74131131e-01 3.95156354e-01 3.65730941e-01
1.77808598e-01 -7.12674320e-01 5.40651023e-01 4.93421823e-01
-1.16196501e+00 1.00817628e-01 -6.69318914e-01 1.59453541e-01
2.76707351e-01 9.25279140e-01 -1.31357634e+00 7.32447565e-01
6.19190991e-01 5.87928891e-01 -5.31188130e-01 1.04041529e+00
-2.37152249e-01 1.12579679e+00 -3.50043833e-01 -7.29377568e-02
3.05155575e-01 -1.14390828e-01 7.71120250e-01 1.19882369e+00
3.48481119e-01 -1.36713788e-01 4.95174497e-01 1.08147132e+00
1.19177438e-01 3.97342443e-01 -3.59253258e-01 -1.48098350e-01
5.67651272e-01 1.25863767e+00 -1.09584701e+00 -6.40648842e-01
-3.36568922e-01 4.81905282e-01 2.59860158e-01 4.88861680e-01
-4.81796920e-01 -3.16582203e-01 4.07175869e-01 3.84256721e-01
5.18068314e-01 -4.28814203e-01 -3.43759835e-01 -1.34145117e+00
-3.03195089e-01 -4.96428341e-01 6.14994824e-01 -4.84652787e-01
-1.22445166e+00 6.28615081e-01 6.24589562e-01 -9.01423156e-01
-3.66566420e-01 -9.80981663e-02 -5.65519631e-01 6.67452753e-01
-1.35835290e+00 -1.15610743e+00 -1.47043288e-01 4.10667896e-01
8.06649685e-01 -2.15261411e-02 5.89346468e-01 -2.65166312e-01
-4.31077152e-01 1.83570474e-01 7.16372803e-02 -3.16714764e-01
7.77399421e-01 -1.42365980e+00 2.92116761e-01 6.93701446e-01
3.24795395e-01 1.16027606e+00 9.82652366e-01 -1.05108380e+00
-8.46323133e-01 -1.13442016e+00 7.72187531e-01 -1.93955511e-01
8.72624755e-01 -6.05819762e-01 -1.33510232e+00 7.72229016e-01
4.90698010e-01 -7.18841910e-01 8.88389468e-01 8.41775894e-01
-3.65612894e-01 1.46894768e-01 -5.78320742e-01 4.89945769e-01
7.60392487e-01 -3.38224024e-01 -9.29667056e-01 3.28750104e-01
9.86504078e-01 -7.60297477e-02 -8.48427773e-01 1.84078738e-01
3.04067820e-01 -5.63973784e-01 5.91809452e-01 -4.96003360e-01
3.32128316e-01 -2.93445021e-01 -4.67982963e-02 -1.38087392e+00
-4.93435085e-01 -5.30325294e-01 -4.22266454e-01 1.48428392e+00
4.48553741e-01 -5.58929801e-01 9.56804395e-01 5.27562857e-01
-3.16408426e-02 -6.65346324e-01 -7.08461702e-01 -6.03442430e-01
-5.07501289e-02 -3.91352534e-01 7.00727224e-01 9.52616394e-01
4.76811141e-01 5.88870227e-01 -4.53667641e-01 4.05211240e-01
6.55545413e-01 5.01303494e-01 8.51959348e-01 -1.70590079e+00
-2.80358106e-01 -3.85956883e-01 -1.52359411e-01 -1.43196034e+00
3.30351735e-03 -7.45181739e-01 -8.18153843e-02 -2.00980806e+00
6.59188926e-01 -3.73275608e-01 -2.71286309e-01 4.02935356e-01
-5.06441414e-01 -3.67706150e-01 -4.20386977e-02 4.29133356e-01
-8.97648156e-01 9.11057472e-01 9.21375394e-01 -1.13568209e-01
-4.46771353e-01 4.58303019e-02 -1.02310681e+00 7.22865582e-01
6.79039359e-01 -6.83002651e-01 -7.50070989e-01 -2.74200857e-01
3.39985251e-01 -3.74147631e-02 1.64367378e-01 -6.17204368e-01
6.69732094e-01 -1.83169886e-01 -7.21747475e-03 -1.01953101e+00
2.57342219e-01 -6.40121400e-01 -1.38154821e-02 -1.08519584e-01
-5.45183301e-01 -3.23604077e-01 2.39769518e-02 1.18392813e+00
-3.81368011e-01 -1.47538856e-01 2.57545650e-01 -7.02029094e-02
-4.05658633e-01 3.69284868e-01 -5.61725795e-01 -7.35816881e-02
6.04950607e-01 -1.37294859e-01 -2.24904060e-01 -7.89808273e-01
-7.44208634e-01 4.78721797e-01 4.13011014e-01 3.79039109e-01
4.73817319e-01 -1.10946143e+00 -5.83509564e-01 -3.04457814e-01
2.56131917e-01 3.54599953e-01 2.72948503e-01 4.93320882e-01
3.96302223e-01 9.30321813e-01 2.63086319e-01 -7.27232456e-01
-8.24337006e-01 5.88614821e-01 -2.52702475e-01 -8.10786843e-01
-6.07569337e-01 4.51220959e-01 9.31423903e-01 -1.78518936e-01
2.69319355e-01 -3.94137800e-01 -4.16593760e-01 1.88637927e-01
2.59431243e-01 3.48732561e-01 -1.62072659e-01 -2.17570290e-01
-9.59860831e-02 2.65312046e-01 -5.04746616e-01 -3.11357737e-01
1.40562356e+00 -5.68372846e-01 -8.60451609e-02 6.64690852e-01
7.35140324e-01 1.52672082e-02 -1.06061804e+00 -7.40078211e-01
4.04748052e-01 -2.67916262e-01 3.13133866e-01 -3.85987341e-01
-6.72534645e-01 8.50278735e-01 -8.29735324e-02 6.37150228e-01
9.88218367e-01 3.95841241e-01 5.38318515e-01 3.32683653e-01
2.71327913e-01 -8.68454218e-01 2.61278898e-01 3.69611651e-01
6.57212615e-01 -9.25402403e-01 2.98462927e-01 -7.86682546e-01
-4.71193254e-01 1.08680463e+00 5.98189473e-01 1.05862513e-01
7.57806182e-01 9.69735556e-05 -3.78270000e-01 -5.08280933e-01
-1.00714934e+00 -2.16633976e-01 5.76819301e-01 5.37804425e-01
4.92066145e-01 -5.98191060e-02 -3.48030359e-01 8.30760777e-01
-4.16629314e-01 -2.36820802e-01 4.07014638e-01 5.90696990e-01
-8.34801137e-01 -1.09613585e+00 -4.30227578e-01 5.06321847e-01
-3.08898896e-01 -7.59298280e-02 -2.53249884e-01 7.67889559e-01
-2.68566996e-01 8.95160973e-01 7.27340877e-02 -4.44559902e-02
-1.83444396e-01 2.12728232e-01 -5.53100407e-02 -1.10048950e+00
-3.06797903e-02 6.95558846e-01 -6.37752935e-02 -1.13719836e-01
-2.97848761e-01 -7.05561936e-01 -8.62077534e-01 2.05582961e-01
-8.15403938e-01 6.84370399e-01 6.49657011e-01 9.97347295e-01
5.39691210e-01 6.99522913e-01 3.24796170e-01 -4.88546968e-01
-2.16853887e-01 -1.34021723e+00 -7.74388969e-01 -1.32291630e-01
-3.31577808e-02 -9.69316721e-01 -1.40383825e-01 1.44662887e-01] | [10.3707857131958, 6.966981410980225] |
4505d6b8-a47b-4636-a57b-8aba052cf8a7 | query-adaptive-late-fusion-for-image | 1810.13103 | null | http://arxiv.org/abs/1810.13103v1 | http://arxiv.org/pdf/1810.13103v1.pdf | Query Adaptive Late Fusion for Image Retrieval | Feature fusion is a commonly used strategy in image retrieval tasks, which
aggregates the matching responses of multiple visual features. Feasible sets of
features can be either descriptors (SIFT, HSV) for an entire image or the same
descriptor for different local parts (face, body). Ideally, the to-be-fused
heterogeneous features are pre-assumed to be discriminative and complementary
to each other. However, the effectiveness of different features varies
dramatically according to different queries. That is to say, for some queries,
a feature may be neither discriminative nor complementary to existing ones,
while for other queries, the feature suffices. As a result, it is important to
estimate the effectiveness of features in a query-adaptive manner. To this end,
this article proposes a new late fusion scheme at the score level. We base our
method on the observation that the sorted score curves contain patterns that
describe their effectiveness. For example, an "L"-shaped curve indicates that
the feature is discriminative while a gradually descending curve suggests a bad
feature. As such, this paper introduces a query-adaptive late fusion pipeline.
In the hand-crafted version, it can be an unsupervised approach to tasks like
particular object retrieval. In the learning version, it can also be applied to
supervised tasks like person recognition and pedestrian retrieval, based on a
trainable neural module. Extensive experiments are conducted on two object
retrieval datasets and one person recognition dataset. We show that our method
is able to highlight the good features and suppress the bad ones, is resilient
to distractor features, and achieves very competitive retrieval accuracy
compared with the state of the art. In an additional person re-identification
dataset, the application scope and limitation of the proposed method are
studied. | ['Liang Zheng', 'Shengjin Wang', 'Zhongdao Wang'] | 2018-10-31 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 6.49123117e-02 -5.42313099e-01 -1.38875201e-01 -4.68477458e-01
-8.17291975e-01 -5.25564253e-01 7.05926597e-01 4.27864164e-01
-4.32788491e-01 5.34068346e-01 3.16824764e-02 3.21822792e-01
-3.46273482e-01 -6.92962408e-01 -3.37676048e-01 -9.54588056e-01
1.02486968e-01 3.00688148e-01 3.84499937e-01 -2.66118914e-01
1.93933219e-01 7.46623755e-01 -2.08816981e+00 2.55006880e-01
4.71912384e-01 1.15701187e+00 4.98180836e-02 2.60634184e-01
1.19516559e-01 3.95518631e-01 -7.17121184e-01 -4.28246558e-01
1.92188174e-01 -1.99439421e-01 -4.30625111e-01 1.64994046e-01
5.55285871e-01 -2.66811758e-01 -2.35632807e-01 1.02647483e+00
5.64333498e-01 2.65235174e-02 7.94419110e-01 -1.25808179e+00
-4.33609098e-01 1.69637397e-01 -3.64827186e-01 6.79662302e-02
6.82626009e-01 1.87053960e-02 9.56657350e-01 -1.06633151e+00
5.21801054e-01 1.12680244e+00 4.10888493e-01 1.71965018e-01
-1.01899934e+00 -5.12327850e-01 1.89570889e-01 3.23290944e-01
-1.64724219e+00 -4.18811888e-01 8.00178885e-01 -4.19295996e-01
4.75378871e-01 5.11638522e-01 6.62956834e-01 8.50569606e-01
1.27351701e-01 9.73986268e-01 1.13003004e+00 -4.01421845e-01
1.08989723e-01 3.34129333e-01 3.22302520e-01 5.63503146e-01
2.79241264e-01 1.55467615e-01 -5.04470825e-01 -1.81979373e-01
2.00629279e-01 3.44318300e-01 -3.12174946e-01 -5.22627532e-01
-9.27690566e-01 6.01255298e-01 5.41167259e-01 6.65259004e-01
-4.58083093e-01 -1.61398724e-01 2.56651282e-01 4.57277775e-01
1.97309732e-01 4.72119339e-02 -1.19668014e-01 1.82479441e-01
-9.87083435e-01 5.65599442e-01 6.08562529e-01 6.52001619e-01
1.05856931e+00 -2.98769295e-01 -6.28469884e-01 8.74079883e-01
2.16115415e-01 4.26538587e-01 5.41153014e-01 -4.34729278e-01
-6.04106225e-02 8.16776693e-01 2.76662529e-01 -1.16458499e+00
-3.40165764e-01 -3.78826320e-01 -8.01940322e-01 3.99808995e-02
4.49722826e-01 4.21913594e-01 -8.61142337e-01 1.71666598e+00
2.70967513e-01 -3.46110873e-02 -6.20051771e-02 1.09419763e+00
1.03747702e+00 4.06058699e-01 -3.63688692e-02 -1.85895368e-01
1.54710269e+00 -7.45580614e-01 -5.94230592e-01 4.80253734e-02
2.66402215e-01 -9.24158156e-01 7.41496086e-01 2.95387179e-01
-7.98881173e-01 -9.18481886e-01 -9.92213964e-01 1.15938999e-01
-6.93496644e-01 3.27392191e-01 3.73382211e-01 7.63355315e-01
-1.00778949e+00 4.89949971e-01 -3.56606990e-01 -5.12269318e-01
7.40900449e-03 4.62530792e-01 -6.65986240e-01 -3.26217949e-01
-1.16367865e+00 7.50032485e-01 4.31782842e-01 1.14752837e-01
-6.46764696e-01 -2.23009154e-01 -6.46715760e-01 2.62089819e-01
2.83253103e-01 -6.46493077e-01 8.54789317e-01 -1.06427646e+00
-1.09493911e+00 8.78914773e-01 -2.27467522e-01 -2.35456482e-01
3.92194539e-01 -1.07792079e-01 -5.61742663e-01 2.67817855e-01
1.05399854e-01 4.70456272e-01 1.16119087e+00 -1.25859356e+00
-7.75949776e-01 -6.09139562e-01 1.92079097e-01 1.40668154e-01
-5.15858412e-01 2.14085713e-01 -6.43349826e-01 -6.03599489e-01
1.70207163e-03 -9.26410258e-01 1.70812726e-01 9.91890281e-02
-7.03903958e-02 -4.42112058e-01 8.62753749e-01 -4.37042505e-01
1.22549176e+00 -2.28414679e+00 6.88318461e-02 5.06682754e-01
9.29246936e-03 2.78602004e-01 -1.43576160e-01 4.99042243e-01
5.28111681e-02 -2.46621817e-01 -4.55199443e-02 -2.51153260e-01
-7.89312422e-02 1.12144183e-02 5.49197420e-02 5.96355736e-01
3.44507515e-01 8.23269486e-01 -7.88707554e-01 -5.81666589e-01
3.42374057e-01 6.15570903e-01 -2.19492912e-01 1.74155518e-01
3.72330606e-01 2.45509431e-01 -5.69882035e-01 1.01014030e+00
6.49618685e-01 -2.93490943e-02 -2.39707250e-02 -5.66694021e-01
-2.33652741e-01 -3.03752869e-01 -1.47888744e+00 1.36658704e+00
-5.82190789e-02 4.28659260e-01 -1.10186242e-01 -1.15688860e+00
1.08035147e+00 1.91550747e-01 5.22736311e-01 -8.47442627e-01
-2.07613106e-03 4.13987666e-01 -6.30218610e-02 -4.24826682e-01
7.84852326e-01 1.09229766e-01 -1.85624108e-01 2.60442168e-01
9.53117162e-02 1.66304961e-01 3.94372463e-01 -8.61802548e-02
8.62291455e-01 -1.92855131e-02 5.35680175e-01 -8.92800540e-02
9.80440199e-01 -2.42866829e-01 2.59830087e-01 9.60220873e-01
-2.72419959e-01 6.10512555e-01 1.03308648e-01 -4.32046324e-01
-7.42699444e-01 -9.20746803e-01 -2.71075130e-01 1.10688174e+00
4.16820168e-01 -4.68226641e-01 -4.94202286e-01 -6.26985788e-01
2.22709432e-01 1.01320878e-01 -5.18161416e-01 -3.47545177e-01
-4.06205922e-01 -4.16638851e-01 5.12279987e-01 3.69553208e-01
6.63795829e-01 -9.17456865e-01 -6.18682146e-01 1.02203637e-01
-9.43151712e-02 -8.84326220e-01 -3.76951605e-01 -4.21105251e-02
-4.13376778e-01 -1.11760139e+00 -1.05844629e+00 -7.02759862e-01
5.88676214e-01 6.73480928e-01 9.35210407e-01 3.67973447e-01
-1.40349150e-01 7.53851235e-01 -6.40129983e-01 -2.62087524e-01
4.36381102e-02 -1.38458669e-01 1.29317805e-01 5.23537397e-01
4.30952817e-01 -9.74566638e-02 -7.23651171e-01 4.61008698e-01
-9.36121583e-01 -5.24369895e-01 6.02505863e-01 1.15460217e+00
6.04598284e-01 1.09725073e-01 2.97693998e-01 -2.88777471e-01
7.13377416e-01 -2.03598201e-01 -2.88466752e-01 6.94356024e-01
-3.47742885e-01 1.11010045e-01 5.24804235e-01 -5.66997707e-01
-7.44624496e-01 2.23466963e-01 2.50440650e-02 -4.15199280e-01
-2.17791572e-01 5.34469366e-01 -2.97709465e-01 -3.06869537e-01
4.10013855e-01 4.33349252e-01 -1.84523109e-02 -4.65004414e-01
1.23469308e-01 7.54890680e-01 3.98892790e-01 -6.16312981e-01
7.47029841e-01 4.59000885e-01 1.90652050e-02 -8.00876677e-01
-4.58011419e-01 -7.56518364e-01 -5.68959832e-01 -5.22696912e-01
4.14131254e-01 -8.28277409e-01 -6.70918822e-01 6.02473795e-01
-9.35857534e-01 2.85680443e-01 -2.46530950e-01 4.06825721e-01
-2.10055500e-01 5.44368684e-01 -2.08991379e-01 -8.87687147e-01
-3.52368295e-01 -1.38215482e+00 1.35572922e+00 5.78206241e-01
-3.91575284e-02 -6.33675039e-01 -7.87854120e-02 1.42514557e-01
3.95681590e-01 -2.10441705e-02 6.66131198e-01 -9.00543153e-01
-4.16062474e-01 -7.01314449e-01 -2.90008664e-01 3.00231129e-01
3.34478855e-01 1.38915479e-01 -1.04122376e+00 -6.32186472e-01
-2.09456444e-01 -2.80641079e-01 1.11722350e+00 1.68875426e-01
9.23546910e-01 -1.14214301e-01 -3.96301061e-01 1.64480120e-01
1.28820622e+00 1.57952532e-01 6.12658143e-01 2.71922499e-01
3.45020026e-01 7.21028864e-01 8.23394239e-01 3.70054096e-01
3.13526779e-01 1.08288980e+00 1.75967112e-01 -1.60272241e-01
-6.17349744e-02 -2.51188446e-02 3.76142412e-01 2.92660803e-01
-2.17703283e-01 -1.43622339e-01 -5.47359407e-01 3.57907057e-01
-1.85493183e+00 -1.04731393e+00 6.52502105e-02 2.47040796e+00
3.13506305e-01 -6.94268048e-02 3.06952029e-01 3.49550605e-01
8.41863394e-01 7.13207424e-02 -2.68791944e-01 -1.14074573e-01
-3.70961249e-01 1.43625319e-01 6.30383790e-02 2.31982186e-01
-1.23831213e+00 7.35140502e-01 5.62362909e+00 1.09120715e+00
-1.26086092e+00 -4.29874510e-02 2.46587336e-01 1.96823433e-01
-1.59048676e-01 3.97105813e-02 -7.68725455e-01 4.66860741e-01
4.40854818e-01 -2.18677387e-01 8.37899223e-02 7.19297349e-01
-1.43544793e-01 -2.79872149e-01 -1.09030545e+00 1.27752113e+00
3.22952718e-01 -7.73185372e-01 1.36852175e-01 -5.05932830e-02
2.64269561e-01 -4.32985961e-01 1.21821225e-01 4.49280679e-01
-3.34881306e-01 -8.27679634e-01 7.70986140e-01 7.56582737e-01
5.29490113e-01 -7.00202107e-01 1.02684426e+00 2.94356495e-01
-1.48473966e+00 -1.39507949e-01 -3.35454375e-01 1.64131030e-01
-2.64231414e-02 4.31641161e-01 -4.68913913e-01 9.63776588e-01
7.07249403e-01 5.95719278e-01 -1.01622021e+00 1.37546968e+00
-3.47305499e-02 -3.53934653e-02 -3.53979141e-01 -1.22500442e-01
5.16553968e-02 -6.39428869e-02 4.91764367e-01 1.23651767e+00
4.98065770e-01 -5.46228737e-02 4.17075455e-01 4.25811499e-01
1.73042744e-01 4.83334392e-01 -7.37580836e-01 1.73496529e-01
3.69410902e-01 1.41309941e+00 -6.34504557e-01 -5.19009888e-01
-4.85907793e-01 1.00907433e+00 1.15901679e-01 3.68669987e-01
-4.78126436e-01 -2.50257760e-01 5.02877772e-01 1.75376106e-02
3.74486476e-01 1.88528046e-01 1.22050598e-01 -1.11379898e+00
2.32206434e-01 -8.30873370e-01 5.32861412e-01 -6.77199006e-01
-1.31279564e+00 6.39542937e-01 1.11724064e-01 -1.52030826e+00
-3.94768745e-01 -5.47480941e-01 -2.58282989e-01 6.98446393e-01
-1.45403659e+00 -1.38180315e+00 -5.34845829e-01 9.03595328e-01
3.16130400e-01 -3.72630447e-01 7.80019999e-01 4.79273498e-01
-4.18690294e-01 6.86180472e-01 -2.12428176e-07 4.70687682e-03
9.58331108e-01 -8.03724170e-01 -2.91997880e-01 8.02726209e-01
3.34206104e-01 7.40393817e-01 7.13796794e-01 -5.27311087e-01
-1.33541632e+00 -6.94949746e-01 8.87144566e-01 -1.23456933e-01
2.65708268e-01 4.78811488e-02 -9.67863739e-01 1.79608345e-01
-2.01330371e-02 1.33110061e-01 6.27524495e-01 4.18820381e-02
-4.75354612e-01 -4.96729732e-01 -1.19091296e+00 4.01432484e-01
6.53288305e-01 -4.80489731e-01 -4.50857282e-01 1.36111811e-01
6.54729903e-02 -1.51358321e-01 -8.45795572e-01 6.84271276e-01
1.06482637e+00 -1.32356000e+00 1.05029213e+00 -3.59542966e-01
2.47663050e-03 -4.70554471e-01 -3.56908381e-01 -9.83525097e-01
-4.57705170e-01 1.22353742e-02 -1.23726666e-01 1.31599271e+00
-4.99954782e-02 -4.59718138e-01 3.89319658e-01 5.04760146e-01
1.07364148e-01 -6.79343402e-01 -1.05156076e+00 -8.49643111e-01
-3.55574012e-01 -2.28111982e-01 6.12346649e-01 6.47895992e-01
-3.12294006e-01 7.11235777e-02 -4.69782352e-01 -2.02153102e-02
5.22484958e-01 3.00572097e-01 8.85212541e-01 -1.49069858e+00
-1.06629118e-01 -5.79725504e-01 -7.93568432e-01 -8.69105518e-01
1.18229777e-01 -7.54479766e-01 -1.22347675e-01 -1.19836712e+00
4.58899170e-01 -4.04303163e-01 -6.69388771e-01 5.23409545e-01
-2.92041868e-01 4.23808515e-01 5.00738919e-01 4.19054151e-01
-5.96818984e-01 3.62610877e-01 8.48896444e-01 -5.12885094e-01
-2.11714387e-01 8.80180001e-02 -5.83090067e-01 4.62441206e-01
5.87824523e-01 -2.65846342e-01 -1.63983218e-02 1.56172393e-02
-3.78480963e-02 -1.16878822e-01 7.42785454e-01 -1.05251527e+00
4.81902540e-01 1.19491220e-01 5.17166078e-01 -6.18337691e-01
5.59306860e-01 -9.14374709e-01 2.03298792e-01 6.26794934e-01
3.05726705e-03 7.98366405e-03 -2.75620855e-02 3.90041649e-01
-6.02894783e-01 -2.53151119e-01 6.04141295e-01 -6.25431761e-02
-9.74727213e-01 1.94633082e-01 -1.73538834e-01 -4.85135198e-01
9.23829198e-01 -4.93422300e-01 -1.51516467e-01 -3.78220618e-01
-4.07609642e-01 3.45652141e-02 5.04112065e-01 5.39114594e-01
9.38858986e-01 -1.49887800e+00 -7.87479997e-01 3.20001990e-01
6.07019782e-01 -6.70849979e-01 3.32135499e-01 8.26544106e-01
1.03006609e-01 4.44341689e-01 -2.79362023e-01 -7.82012045e-01
-1.60050213e+00 6.57086849e-01 9.69905332e-02 -2.88317472e-01
-2.90985465e-01 5.12095213e-01 7.82670528e-02 4.67238575e-02
2.44044811e-01 1.63113382e-02 -5.52358270e-01 5.01968145e-01
7.55785048e-01 2.14747280e-01 1.66568086e-01 -1.09236550e+00
-5.68804204e-01 8.54170144e-01 -2.45274603e-01 7.76089951e-02
1.04623497e+00 -1.25796169e-01 -1.52621999e-01 4.04735774e-01
1.13994718e+00 5.97853102e-02 -7.97228277e-01 -4.30291027e-01
-4.48381007e-02 -5.79678178e-01 -1.57061294e-01 -5.08459687e-01
-1.11198080e+00 5.65769017e-01 1.05796218e+00 2.83078939e-01
1.38182080e+00 4.02120985e-02 3.33275467e-01 3.81668121e-01
6.68004036e-01 -9.81029928e-01 1.10530242e-01 1.48549274e-01
9.28146720e-01 -1.33271599e+00 9.78246257e-02 -3.08333665e-01
-4.82290149e-01 1.30501497e+00 3.08553219e-01 -1.55711964e-01
4.72284883e-01 -2.61271950e-02 -2.03054652e-01 -7.06534311e-02
-4.42864716e-01 -6.96019053e-01 7.50894964e-01 4.63898689e-01
3.50664049e-01 2.52104308e-02 -6.27464116e-01 4.61450547e-01
1.05016977e-01 9.28119421e-02 -4.11977433e-02 7.28793263e-01
-3.63350481e-01 -1.28954053e+00 -6.69227362e-01 4.14834142e-01
-2.28864536e-01 1.42514393e-01 -5.64153671e-01 9.04819548e-01
2.50308156e-01 9.35058951e-01 -5.75824007e-02 -5.42268157e-01
4.90973443e-01 1.08557150e-01 5.81540465e-01 -1.99050456e-01
-8.01489532e-01 1.40968040e-01 -1.88476712e-01 -4.57136601e-01
-8.65355015e-01 -7.27968812e-01 -7.72078335e-01 -3.80807966e-02
-3.97783399e-01 -3.08920746e-03 4.18975651e-01 1.03746891e+00
1.63176686e-01 1.95068225e-01 6.21392071e-01 -7.99438298e-01
-3.68686140e-01 -8.22189748e-01 -4.76702034e-01 7.59426355e-01
2.98709065e-01 -1.07034326e+00 -2.34226272e-01 -1.31830543e-01] | [10.846480369567871, 0.6857770681381226] |
f8e66df5-6ed0-408e-b793-5f40e132d929 | rego-reference-guided-outpainting-for-scenery | 2106.10601 | null | https://arxiv.org/abs/2106.10601v4 | https://arxiv.org/pdf/2106.10601v4.pdf | ReGO: Reference-Guided Outpainting for Scenery Image | We aim to tackle the challenging yet practical scenery image outpainting task in this work. Recently, generative adversarial learning has significantly advanced the image outpainting by producing semantic consistent content for the given image. However, the existing methods always suffer from the blurry texture and the artifacts of the generative part, making the overall outpainting results lack authenticity. To overcome the weakness, this work investigates a principle way to synthesize texture-rich results by borrowing pixels from its neighbors (i.e., reference images), named \textbf{Re}ference-\textbf{G}uided \textbf{O}utpainting (ReGO). Particularly, the ReGO designs an Adaptive Content Selection (ACS) module to transfer the pixel of reference images for texture compensating of the target one. To prevent the style of the generated part from being affected by the reference images, a style ranking loss is further proposed to augment the ReGO to synthesize style-consistent results. Extensive experiments on two popular benchmarks, NS6K \cite{yangzx} and NS8K \cite{wang}, well demonstrate the effectiveness of our ReGO. Our code will be made public available. | ['Yi Yang', 'Li Zhu', 'Xueming Qian', 'Yunchao Wei', 'Yaxiong Wang'] | 2021-06-20 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 5.31967282e-01 8.34665000e-02 7.80752599e-02 -1.48691565e-01
-6.70059919e-01 -3.48396063e-01 4.01914060e-01 -5.55770040e-01
7.60786161e-02 1.03488219e+00 7.35054538e-02 8.79527181e-02
1.16562702e-01 -7.62336791e-01 -1.05555236e+00 -1.00144887e+00
5.73600590e-01 -1.55104786e-01 2.56363694e-02 -3.05166274e-01
1.93366602e-01 2.62539327e-01 -1.20929825e+00 3.53829116e-01
1.31915879e+00 1.12455571e+00 5.93050778e-01 3.77745718e-01
-1.33570507e-01 9.74914014e-01 -8.65410089e-01 -4.55822825e-01
5.17025352e-01 -9.00050640e-01 -4.77072388e-01 2.80037105e-01
6.08350396e-01 -5.14472187e-01 -6.41829669e-01 1.44933057e+00
4.39032406e-01 1.48940399e-01 3.96864444e-01 -1.27877283e+00
-1.00903904e+00 3.94298166e-01 -7.61953235e-01 -3.46101224e-02
9.82323512e-02 3.23809743e-01 4.84796047e-01 -8.42181861e-01
7.50971437e-01 1.41765797e+00 3.62220615e-01 7.22810209e-01
-1.15123546e+00 -8.70531976e-01 1.34802148e-01 1.41019955e-01
-1.41974509e+00 -5.03773451e-01 1.23188913e+00 2.85441857e-02
9.66409072e-02 6.26323879e-01 3.65130931e-01 1.25158525e+00
3.30738991e-01 9.36582863e-01 1.48883891e+00 -2.29661494e-01
4.85945754e-02 1.89749047e-03 -4.85540330e-01 5.77368259e-01
1.14599779e-01 1.70024902e-01 -4.41432863e-01 1.79471493e-01
1.06621432e+00 1.26400217e-01 -7.12931573e-01 -3.79449539e-02
-1.11804247e+00 4.12106156e-01 6.02648675e-01 -1.96128730e-02
-2.77788073e-01 6.61084726e-02 2.75182575e-01 3.04904431e-01
6.69863939e-01 3.88037473e-01 -2.00430825e-01 2.41098538e-01
-9.42023218e-01 3.19582373e-01 2.09193572e-01 1.35272872e+00
8.69218528e-01 4.08606529e-01 -5.86950779e-01 1.09416676e+00
-1.05463974e-01 5.66844463e-01 4.87568706e-01 -1.10869777e+00
6.86012149e-01 3.72808278e-01 1.61759228e-01 -1.07982612e+00
3.39863688e-01 -2.89633840e-01 -1.20316172e+00 3.41172308e-01
1.60688996e-01 -2.12515637e-01 -9.19893801e-01 1.51807809e+00
2.02684969e-01 1.98937103e-01 -2.07097474e-02 9.67339873e-01
8.61018062e-01 1.00272751e+00 7.04235723e-03 -2.21112981e-01
9.14090216e-01 -1.32837760e+00 -8.52618635e-01 -3.56400043e-01
8.43072608e-02 -1.06650734e+00 1.03073144e+00 4.83148426e-01
-1.22571266e+00 -9.09154832e-01 -1.14777458e+00 -1.73296824e-01
7.46743828e-02 2.95641068e-02 2.98542500e-01 3.68943036e-01
-8.76848102e-01 6.16298437e-01 -3.58027637e-01 9.10308659e-02
7.09055126e-01 -7.29299262e-02 -1.78353444e-01 -3.58044803e-01
-1.18516338e+00 5.24057508e-01 5.70385277e-01 2.86002845e-01
-8.68090808e-01 -6.55854106e-01 -6.98013663e-01 -1.41586438e-01
4.22908932e-01 -5.96264362e-01 8.04921031e-01 -1.70928288e+00
-1.54046464e+00 6.36745512e-01 7.30460957e-02 -1.52760610e-01
9.12604988e-01 -2.43185982e-01 -4.87051547e-01 -2.10881000e-03
2.83772945e-01 7.61477530e-01 1.33613324e+00 -1.63608181e+00
-4.74424392e-01 -1.47356868e-01 -1.28996506e-01 2.74248987e-01
-1.28748596e-01 -2.47624785e-01 -7.24996865e-01 -1.57651150e+00
-5.35672158e-02 -8.34396780e-01 -4.87062745e-02 1.49301201e-01
-7.40719497e-01 2.80042797e-01 1.10600543e+00 -9.61817741e-01
9.83728886e-01 -2.34388661e+00 2.28584915e-01 -5.80757298e-02
2.38809548e-03 4.20896232e-01 -3.00999552e-01 1.23416774e-01
-1.12737626e-01 -2.72246040e-02 -4.63154346e-01 -3.76462132e-01
-4.34651911e-01 3.75006765e-01 -6.04205072e-01 3.06248844e-01
2.11066991e-01 9.90421534e-01 -8.37718248e-01 -5.65369129e-01
3.16842586e-01 3.63819301e-01 -4.39757198e-01 4.45809156e-01
-4.45474744e-01 7.63303220e-01 -5.93904257e-01 6.36280656e-01
1.04183495e+00 1.04710512e-01 -3.15889537e-01 -3.25091660e-01
1.04164109e-01 -3.15067470e-01 -1.05891085e+00 1.87846315e+00
-4.72991645e-01 3.38763237e-01 2.16236547e-01 -7.16191173e-01
1.14470482e+00 -2.25981642e-02 2.60587960e-01 -9.73388076e-01
1.25302836e-01 2.09601507e-01 -3.21524978e-01 -4.68845725e-01
5.77933967e-01 -3.77885960e-02 1.22387007e-01 5.03223948e-02
-2.96201974e-01 -2.82679856e-01 -1.28841907e-01 5.30765764e-02
6.09433949e-01 4.97718722e-01 -1.46263078e-01 -1.48742199e-01
6.95591211e-01 -1.22799806e-01 7.19215333e-01 6.27195537e-01
-1.26627952e-01 1.10300350e+00 2.88633823e-01 -3.00017804e-01
-1.22009814e+00 -9.86937582e-01 3.93990390e-02 6.55857205e-01
6.70613229e-01 -8.69360119e-02 -9.29826438e-01 -7.12481797e-01
-3.15372735e-01 7.14849591e-01 -6.24545455e-01 -4.13286120e-01
-6.83858335e-01 -5.41626871e-01 5.57548344e-01 4.16736096e-01
1.10127699e+00 -1.33242810e+00 -6.98662326e-02 2.50684291e-01
-4.87416685e-01 -9.70672905e-01 -9.74580467e-01 -3.58044505e-01
-6.43877089e-01 -7.07381368e-01 -1.31681442e+00 -8.05285990e-01
7.73710668e-01 2.25965217e-01 9.76256371e-01 1.54339686e-01
-1.32009506e-01 -2.94407904e-01 -4.71599847e-01 -2.95698404e-01
-6.49963379e-01 -3.33705336e-01 -4.08232868e-01 3.15348625e-01
-3.40843379e-01 -4.77299392e-01 -9.48878765e-01 3.91686618e-01
-1.34448612e+00 4.91305381e-01 7.07592309e-01 1.10388458e+00
8.86151612e-01 3.03459644e-01 3.98980767e-01 -8.90695333e-01
4.27686810e-01 -2.01531649e-01 -2.55895853e-01 2.30889857e-01
-5.67207754e-01 -1.11870848e-01 1.06243789e+00 -5.25204957e-01
-1.39374435e+00 -1.51198998e-01 -1.64821580e-01 -8.27588856e-01
-1.00919157e-01 -1.13913663e-01 -6.29447401e-01 -3.23575698e-02
3.55879724e-01 8.05304110e-01 -3.27339917e-02 -4.33307737e-01
3.16596925e-01 5.52107573e-01 7.59987295e-01 -5.73809862e-01
9.27927434e-01 5.43203473e-01 -1.14379331e-01 -6.57753646e-01
-7.94334471e-01 1.45816058e-01 9.11991484e-03 -1.91212937e-01
8.70173395e-01 -9.64313328e-01 -2.30326802e-01 8.41755927e-01
-1.09900832e+00 -4.01479781e-01 -4.08866853e-01 -1.93783492e-02
-6.52669430e-01 6.48249626e-01 -4.93868291e-01 -4.06847507e-01
-5.77699602e-01 -1.25580800e+00 1.26196682e+00 4.71637398e-01
1.26466766e-01 -5.39789557e-01 -2.75001943e-01 5.11186481e-01
3.71141553e-01 3.53528023e-01 8.47468495e-01 6.31032735e-02
-7.08399296e-01 1.16664812e-01 -4.30135369e-01 8.27826977e-01
2.39385530e-01 -1.76248729e-01 -9.14503396e-01 -3.03385049e-01
1.67485595e-01 -1.65683255e-01 8.51122618e-01 1.69973746e-01
1.59397936e+00 -5.59634924e-01 -3.76811363e-02 9.09350574e-01
1.53600729e+00 3.57533872e-01 1.19645154e+00 3.36077154e-01
1.05428040e+00 3.26976210e-01 7.65294731e-01 2.53864110e-01
-5.01357159e-03 6.91675186e-01 4.75074291e-01 -3.42993468e-01
-6.29502654e-01 -5.67659676e-01 3.81558239e-01 4.29516494e-01
-1.37781659e-02 -4.90309864e-01 -1.48689851e-01 2.57582277e-01
-1.67319548e+00 -9.40221667e-01 -1.25170127e-01 1.94558179e+00
1.06182051e+00 1.82526857e-02 -2.75493115e-01 5.32441251e-02
8.83254170e-01 3.96650195e-01 -6.29158974e-01 -1.11554258e-01
-4.63687122e-01 2.74944484e-01 5.25815427e-01 2.81529129e-01
-8.92933547e-01 9.54072773e-01 4.63811302e+00 1.52388453e+00
-1.17799580e+00 4.92021181e-02 1.27325320e+00 1.65764987e-01
-4.19750065e-01 -6.19228669e-02 -3.69885981e-01 1.03654814e+00
9.72414687e-02 5.69348596e-02 6.11251593e-01 7.56748796e-01
3.17660511e-01 -1.29948646e-01 -5.94891846e-01 1.08134818e+00
1.43226743e-01 -1.25146532e+00 3.63029450e-01 -2.81614810e-01
1.11911488e+00 -5.59908390e-01 2.52444327e-01 1.98984593e-01
3.62279126e-03 -9.95394170e-01 9.75506723e-01 7.58711278e-01
1.13242280e+00 -8.88996065e-01 4.54689503e-01 2.09588721e-01
-9.47441399e-01 1.27498746e-01 -4.32189763e-01 2.73190767e-01
6.31356910e-02 7.52836108e-01 -1.00086361e-01 8.98487628e-01
8.16976547e-01 7.48469830e-01 -7.18270481e-01 6.85212553e-01
-4.13610727e-01 3.80169004e-01 1.11974686e-01 3.93428534e-01
8.68936479e-02 -4.05744314e-01 7.54681766e-01 7.93376088e-01
3.27419072e-01 1.45984158e-01 1.51697204e-01 1.19728863e+00
-2.72312731e-01 1.50550455e-01 -3.88841480e-01 2.13780507e-01
1.91815868e-01 1.00652039e+00 -6.79633021e-01 -4.59394276e-01
-1.49018094e-01 1.64530432e+00 -1.73517894e-02 4.61747587e-01
-9.24367785e-01 -5.18569291e-01 4.45266217e-01 9.48857889e-02
3.65935951e-01 1.84412226e-01 -4.69529837e-01 -1.22916436e+00
2.89729506e-01 -1.15940702e+00 1.26489680e-02 -1.14195871e+00
-1.36596370e+00 6.67441368e-01 -1.57773316e-01 -1.50707304e+00
2.36642033e-01 -2.94088796e-02 -6.88877463e-01 1.20067060e+00
-1.32820797e+00 -1.10490870e+00 -4.96851504e-01 6.97162211e-01
9.19408202e-01 -1.29925415e-01 2.94646144e-01 4.24407482e-01
-6.38323486e-01 6.75480425e-01 2.56270647e-01 1.73500013e-02
8.89642358e-01 -1.00498772e+00 2.27399945e-01 1.02316034e+00
-2.91435957e-01 3.37851971e-01 7.98348665e-01 -9.26447034e-01
-1.37186182e+00 -1.47498417e+00 2.90296644e-01 1.41811520e-02
9.87935811e-02 -1.61410436e-01 -9.37009871e-01 3.65544856e-01
3.82265776e-01 7.76013955e-02 -8.89981985e-02 -8.55585098e-01
-2.41672695e-01 -3.34257513e-01 -1.33698213e+00 9.15231109e-01
1.05172408e+00 -1.85242057e-01 -3.90816212e-01 1.72996461e-01
8.19526970e-01 -5.56999266e-01 -6.61048174e-01 4.46174800e-01
3.06436896e-01 -1.20357978e+00 1.04338038e+00 5.40756583e-02
1.01865125e+00 -5.59135020e-01 6.09695874e-02 -1.19674778e+00
-1.45579919e-01 -7.60293961e-01 5.81607558e-02 1.59500659e+00
-2.36477509e-01 -4.00427729e-01 6.92796409e-01 2.68411070e-01
-1.82020739e-01 -5.75953007e-01 -5.80874443e-01 -7.49560654e-01
2.03068480e-01 1.87597889e-02 7.24460185e-01 9.34533954e-01
-7.62334645e-01 5.84520027e-02 -8.92296374e-01 -4.81234826e-02
6.66269064e-01 2.91726649e-01 8.19306135e-01 -5.42497694e-01
-4.94199216e-01 -3.84044617e-01 -4.45792712e-02 -1.25056183e+00
-5.52605651e-02 -7.08512187e-01 2.05313146e-01 -1.20407856e+00
8.82453471e-02 -5.98537982e-01 -7.87142366e-02 3.03095639e-01
-5.74550986e-01 5.44188440e-01 3.70223075e-01 3.50235403e-01
-3.31109077e-01 8.52107584e-01 2.13948584e+00 -3.63741636e-01
5.99278957e-02 -1.48406520e-01 -7.73538649e-01 4.77192342e-01
6.92061782e-01 -3.61502767e-01 -4.83120978e-01 -4.56313282e-01
-2.39644349e-01 2.98554331e-01 6.07543409e-01 -9.85002935e-01
-2.01355323e-01 -2.56241709e-01 7.69486189e-01 -5.87975085e-01
2.21765965e-01 -7.63538182e-01 4.20739919e-01 3.77507687e-01
-2.53704578e-01 -1.00894637e-01 1.38195679e-01 7.05114365e-01
-4.23232883e-01 -1.19608581e-01 1.13626766e+00 -2.30263531e-01
-8.05188596e-01 5.00533104e-01 1.20645434e-01 1.87190950e-01
1.00180161e+00 -2.26824641e-01 -3.16077381e-01 -3.95487934e-01
-2.91784495e-01 4.96883243e-02 7.80924916e-01 5.57212114e-01
7.29182839e-01 -1.37696147e+00 -7.77142882e-01 3.22051644e-01
-2.80512706e-03 3.20036978e-01 7.08142936e-01 4.41279143e-01
-7.25975811e-01 -1.35277033e-01 -4.39924806e-01 -2.21972004e-01
-9.44622099e-01 8.38656187e-01 2.24173576e-01 -1.29790008e-01
-8.00007522e-01 8.84388745e-01 6.09522283e-01 3.78079116e-02
1.03089690e-01 6.51534554e-03 1.24336489e-01 -3.67367387e-01
5.30424654e-01 4.19841915e-01 -9.76954773e-02 -6.23616159e-01
3.23774256e-02 2.93557793e-01 -2.88325757e-01 1.31835371e-01
1.11573374e+00 -3.33154023e-01 -2.84094870e-01 -5.74961640e-02
1.23866296e+00 9.59705189e-02 -1.64812231e+00 -3.44276801e-02
-6.31344974e-01 -7.38679647e-01 -1.16004996e-01 -8.44961226e-01
-1.53637540e+00 5.09197652e-01 7.19242573e-01 -1.31544560e-01
1.51259744e+00 -3.70792180e-01 1.25844836e+00 -1.80182099e-01
3.07518303e-01 -9.42726791e-01 1.25802770e-01 8.84180218e-02
1.14972818e+00 -1.01579082e+00 -3.42019349e-02 -5.11694431e-01
-7.38914251e-01 9.19477701e-01 7.24833369e-01 -4.42023218e-01
2.53575772e-01 1.05130032e-01 9.51391309e-02 1.85426816e-01
-2.23281711e-01 1.64151356e-01 1.49536148e-01 5.15487611e-01
4.88840044e-02 -6.46253973e-02 -4.68778074e-01 4.05269206e-01
-7.29672089e-02 3.38654369e-02 4.61801440e-01 8.11705649e-01
-2.70319767e-02 -1.17008460e+00 -6.97308958e-01 2.52552152e-01
-4.92264092e-01 -2.72926688e-01 -3.18809450e-01 5.80441713e-01
5.33123076e-01 8.06617916e-01 -2.34335676e-01 -3.02488953e-01
2.69483000e-01 -2.00616002e-01 3.70803416e-01 -1.91578820e-01
-3.99849683e-01 3.90317589e-01 -3.53269041e-01 -6.02421343e-01
-2.97736138e-01 -3.96421254e-01 -9.38935995e-01 -2.69599766e-01
-9.42420959e-02 -1.93254650e-01 1.65160730e-01 6.90805435e-01
2.11016774e-01 7.67067194e-01 8.54529858e-01 -7.21761942e-01
-3.88137072e-01 -7.87481546e-01 -6.62671447e-01 8.18932116e-01
2.63783216e-01 -3.80883336e-01 -2.49994233e-01 2.04811275e-01] | [11.391535758972168, -1.1937392950057983] |
9f650975-151b-4072-bd2d-3b4f7eacfa71 | viser-visual-self-regularization | 1802.02568 | null | http://arxiv.org/abs/1802.02568v1 | http://arxiv.org/pdf/1802.02568v1.pdf | VISER: Visual Self-Regularization | In this work, we propose the use of large set of unlabeled images as a source
of regularization data for learning robust visual representation. Given a
visual model trained by a labeled dataset in a supervised fashion, we augment
our training samples by incorporating large number of unlabeled data and train
a semi-supervised model. We demonstrate that our proposed learning approach
leverages an abundance of unlabeled images and boosts the visual recognition
performance which alleviates the need to rely on large labeled datasets for
learning robust representation. To increment the number of image instances
needed to learn robust visual models in our approach, each labeled image
propagates its label to its nearest unlabeled image instances. These retrieved
unlabeled images serve as local perturbations of each labeled image to perform
Visual Self-Regularization (VISER). To retrieve such visual self regularizers,
we compute the cosine similarity in a semantic space defined by the penultimate
layer in a fully convolutional neural network. We use the publicly available
Yahoo Flickr Creative Commons 100M dataset as the source of our unlabeled image
set and propose a distributed approximate nearest neighbor algorithm to make
retrieval practical at that scale. Using the labeled instances and their
regularizer samples we show that we significantly improve object categorization
and localization performance on the MS COCO and Visual Genome datasets where
objects appear in context. | ['Hamid Izadinia', 'Pierre Garrigues'] | 2018-02-07 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 1.77942351e-01 -2.37342007e-02 -5.49730003e-01 -6.04065895e-01
-9.73888397e-01 -1.08243525e+00 4.79222089e-01 -6.55547082e-02
-4.38232690e-01 3.22212130e-01 1.76289484e-01 -2.51343977e-02
1.49800643e-01 -3.82058948e-01 -1.16030788e+00 -5.94670832e-01
2.27759212e-01 4.39517081e-01 4.34373803e-02 3.58565375e-02
3.75535369e-01 5.93335986e-01 -1.60907865e+00 6.51975513e-01
4.77313936e-01 1.14701140e+00 2.45709568e-01 5.77893257e-01
-5.62047288e-02 1.11110532e+00 -1.70256674e-01 3.61832753e-02
6.86646342e-01 -1.00863144e-01 -9.20347095e-01 4.87577617e-01
1.14926815e+00 -1.95215151e-01 -5.41058660e-01 1.13963413e+00
4.09215778e-01 4.60649341e-01 1.12315905e+00 -1.35069346e+00
-1.60352111e+00 4.54149336e-01 -7.77983606e-01 2.27416456e-01
-3.94275859e-02 1.53335690e-01 8.17494273e-01 -1.33567405e+00
9.78829622e-01 1.06792593e+00 4.47440743e-01 5.90953648e-01
-1.22206461e+00 -5.11611581e-01 1.75918370e-01 2.22400218e-01
-1.78641176e+00 -5.84743261e-01 9.28364336e-01 -3.64155263e-01
7.49482036e-01 1.02976365e-02 2.16911554e-01 9.91603851e-01
-5.45869708e-01 8.34273815e-01 9.66192067e-01 -6.06161714e-01
4.26105797e-01 4.29564983e-01 3.63299936e-01 7.86754072e-01
-2.23109219e-02 3.28704804e-01 -3.64996374e-01 -1.31476700e-01
4.60398436e-01 4.19353396e-01 -1.48628995e-01 -6.81548476e-01
-9.09003079e-01 8.51602018e-01 1.11443329e+00 1.63136005e-01
-2.94193551e-02 2.53995210e-01 2.11680889e-01 3.80285233e-01
4.14810359e-01 2.17606306e-01 -2.31840596e-01 7.33190656e-01
-1.01546979e+00 -2.50835210e-01 3.19312006e-01 1.13657820e+00
1.18987274e+00 -5.03734909e-02 -3.71193558e-01 1.13966012e+00
4.21314538e-01 7.17306376e-01 7.42310643e-01 -1.18454230e+00
3.13160449e-01 8.05916131e-01 8.05669948e-02 -1.08086491e+00
1.43765450e-01 -2.75563389e-01 -8.09878349e-01 6.00698218e-02
1.93095297e-01 5.94144881e-01 -1.41097200e+00 1.61879766e+00
3.73831660e-01 4.09293473e-01 1.78018078e-01 9.37016547e-01
6.60515904e-01 5.62748492e-01 -6.95159063e-02 -7.03134611e-02
8.99765372e-01 -1.52061272e+00 -1.16806202e-01 -2.54527837e-01
5.96096098e-01 -8.08144033e-01 1.23216045e+00 1.05042569e-01
-7.04198062e-01 -7.37252474e-01 -9.84891176e-01 -3.68650049e-01
-6.74357712e-01 1.87977135e-01 3.59542489e-01 1.64531887e-01
-1.04774356e+00 3.44836771e-01 -5.43637872e-01 -5.19127786e-01
7.42492199e-01 6.95431307e-02 -6.87512875e-01 -7.17296422e-01
-4.73902225e-01 7.04668760e-01 4.58943039e-01 -1.43043309e-01
-1.38568151e+00 -5.35262227e-01 -8.81457925e-01 -1.57434285e-01
1.76986232e-01 -8.61172080e-02 8.31397831e-01 -1.34298265e+00
-7.06954718e-01 1.29334176e+00 -1.52108580e-01 -3.65146309e-01
2.72566557e-01 1.07271247e-01 -3.55363727e-01 5.48565507e-01
3.63777280e-01 1.04943538e+00 1.42848122e+00 -1.61413956e+00
-3.87487411e-01 -5.19353211e-01 -2.95348525e-01 2.98900940e-02
-4.18369502e-01 -2.15939373e-01 -7.34120488e-01 -6.99169993e-01
3.17658961e-01 -1.23811615e+00 -1.96275726e-01 1.90118149e-01
-3.86874646e-01 -9.87600833e-02 1.03348672e+00 -2.30602250e-01
8.00349653e-01 -2.19412971e+00 -2.24471875e-02 4.23780411e-01
1.21094622e-01 3.43243986e-01 -7.14664638e-01 2.56413311e-01
-3.00674736e-01 6.37968779e-02 -1.38275683e-01 -3.93105805e-01
-2.54006684e-01 3.12907249e-01 -6.02254570e-01 7.37260997e-01
9.27006900e-02 1.27014160e+00 -8.59485388e-01 -5.15985072e-01
2.99119532e-01 5.20166099e-01 -5.94317198e-01 3.97233605e-01
-1.82088658e-01 3.89643699e-01 -3.14347982e-01 8.49431992e-01
7.89073050e-01 -6.60671413e-01 -2.20184654e-01 -4.38353360e-01
3.35978001e-01 -4.56534296e-01 -1.13378549e+00 1.87991929e+00
-2.51062602e-01 4.11262780e-01 -2.08844706e-01 -1.25198698e+00
7.88702905e-01 -1.03295378e-01 2.41314605e-01 -6.25637531e-01
-6.63912892e-02 1.02876440e-01 -6.37824178e-01 -5.26136279e-01
2.41901696e-01 1.06228828e-01 1.84864163e-01 7.36288011e-01
3.76123518e-01 9.83866900e-02 1.21930771e-01 6.77927077e-01
9.71180260e-01 -8.96881148e-03 1.39097601e-01 -9.54359397e-02
4.16544169e-01 1.26442850e-01 1.88516304e-01 9.04479444e-01
-3.04819167e-01 8.63300204e-01 -3.69789928e-01 -6.92113698e-01
-1.31985927e+00 -1.04027331e+00 -2.82806158e-01 1.50653219e+00
2.17961252e-01 -1.95542738e-01 -4.38335657e-01 -1.00134254e+00
3.46824735e-01 3.03687602e-01 -9.21758890e-01 -4.58729833e-01
-2.29501486e-01 -3.34549010e-01 5.64601243e-01 5.90260684e-01
3.39767486e-01 -1.06250811e+00 -2.67227720e-02 -3.00689876e-01
-6.13063313e-02 -8.96401584e-01 -8.18992078e-01 4.04141456e-01
-6.48903668e-01 -1.15452230e+00 -6.08795762e-01 -1.20645654e+00
1.29292941e+00 8.03379476e-01 9.16520000e-01 2.51249492e-01
-6.11370087e-01 7.48442769e-01 -4.03341413e-01 5.68878315e-02
-4.07734424e-01 -2.68757135e-01 2.91001499e-01 2.07242012e-01
5.62534392e-01 -3.94141018e-01 -7.26838887e-01 3.03209662e-01
-1.27022648e+00 -4.05389339e-01 4.59488988e-01 1.00861192e+00
9.77686167e-01 -3.96490335e-01 5.13104379e-01 -8.86141539e-01
8.27106833e-02 -6.18280530e-01 -6.05078936e-01 5.89526951e-01
-7.37202644e-01 3.76591593e-01 5.93832076e-01 -7.98122525e-01
-6.36608779e-01 6.85586512e-01 5.86431682e-01 -1.14443421e+00
-1.95398480e-01 2.64131576e-01 2.73548752e-01 -5.35820544e-01
1.15332949e+00 2.85099655e-01 -8.31063613e-02 -4.45162594e-01
1.16017520e+00 8.63072336e-01 6.67622447e-01 -4.59240854e-01
1.15402198e+00 6.64239943e-01 -2.73119211e-01 -6.08539879e-01
-1.30024838e+00 -8.51362467e-01 -8.92213523e-01 -8.14843923e-03
6.17818534e-01 -1.23891270e+00 -2.75376707e-01 -1.22622788e-01
-7.68093109e-01 -1.39255926e-01 -3.92413825e-01 1.60318449e-01
-5.27656972e-01 4.61167514e-01 -3.98594946e-01 -5.17107368e-01
-3.26122433e-01 -8.73301089e-01 1.25932264e+00 -2.43800282e-02
2.15995759e-01 -7.27122009e-01 1.85215011e-01 5.79643488e-01
3.61110002e-01 -1.26096994e-01 7.14693904e-01 -8.62472415e-01
-7.90387988e-01 -3.06267798e-01 -4.78023589e-01 6.98096097e-01
5.26908040e-02 -1.74329177e-01 -1.26322997e+00 -7.33744323e-01
-3.02070916e-01 -1.21018863e+00 1.23823810e+00 1.30158842e-01
1.19594491e+00 -4.41693395e-01 -3.48815203e-01 8.85819852e-01
1.66126359e+00 -4.73435462e-01 1.89115211e-01 1.43672563e-02
1.04126668e+00 4.20018882e-01 4.81116891e-01 8.41782019e-02
2.41655707e-01 2.61733651e-01 4.22136635e-01 -4.66496721e-02
-2.50740170e-01 -5.81210613e-01 9.17310193e-02 7.28712916e-01
2.73168206e-01 1.42006889e-01 -8.45360756e-01 7.42103755e-01
-1.93313849e+00 -9.33114946e-01 3.56325060e-01 2.24155164e+00
9.47779596e-01 -4.52613473e-01 -2.58286536e-01 -2.93590426e-01
7.52554059e-01 1.02170765e-01 -8.15921187e-01 1.82685912e-01
-2.98349857e-01 -9.77097005e-02 6.54709339e-01 5.53337157e-01
-1.24157250e+00 1.27700233e+00 6.12186575e+00 7.08675027e-01
-1.20156527e+00 4.41754721e-02 7.61584282e-01 -2.32650161e-01
-1.00481518e-01 -1.23786397e-01 -6.56019449e-01 1.61738828e-01
7.35241830e-01 -1.05324397e-02 8.53690207e-01 1.15733802e+00
-8.09715316e-02 1.58568993e-01 -1.40482056e+00 1.34728646e+00
5.18459320e-01 -1.46685302e+00 5.81855893e-01 -1.29291683e-01
1.24674761e+00 4.91836011e-01 3.79306674e-01 2.94052601e-01
4.96946186e-01 -1.31148648e+00 4.55040276e-01 3.84468794e-01
8.66432726e-01 -5.64929962e-01 2.23984435e-01 2.82688022e-01
-1.06412923e+00 -3.30059081e-01 -9.63309824e-01 4.00414705e-01
-4.65837359e-01 1.87051818e-01 -8.37632775e-01 -2.28698831e-02
8.36337209e-01 9.20558274e-01 -1.15708184e+00 8.44438255e-01
6.21105470e-02 4.76561189e-01 -3.39862317e-01 5.37838936e-01
3.78752232e-01 -7.20107779e-02 1.48936927e-01 8.13144505e-01
-6.41602352e-02 3.56286094e-02 5.19776285e-01 9.09794629e-01
-5.29939175e-01 2.54123330e-01 -1.03206646e+00 -7.14360848e-02
7.82659173e-01 1.33249772e+00 -6.42232239e-01 -5.20748615e-01
-2.97157705e-01 1.26300848e+00 1.02132857e+00 7.77756870e-01
-4.80289340e-01 6.05901442e-02 1.04951493e-01 -1.16840564e-01
3.84160548e-01 1.92165345e-01 1.95696831e-01 -1.48023224e+00
-4.90753306e-03 -8.86424661e-01 5.58934212e-01 -9.71985638e-01
-1.75416434e+00 4.93134141e-01 -2.42039442e-01 -1.31346524e+00
-1.38572395e-01 -5.15931010e-01 -3.25577706e-01 6.50566995e-01
-1.54814136e+00 -1.51039398e+00 -3.96823496e-01 1.07972825e+00
4.36583698e-01 -3.55120033e-01 7.77868330e-01 1.76325604e-01
-3.06899399e-01 6.77833378e-01 4.63825852e-01 2.94987231e-01
1.17849278e+00 -1.07191741e+00 -7.19865486e-02 7.97823906e-01
8.17607880e-01 8.61747444e-01 2.32069820e-01 -3.99710655e-01
-1.51204622e+00 -1.63854492e+00 4.28185910e-01 -6.94211006e-01
5.52181005e-01 -2.66350210e-01 -8.98781240e-01 1.13923037e+00
2.27671470e-02 1.05850625e+00 7.12798953e-01 -4.13845688e-01
-9.96759593e-01 -9.78379399e-02 -1.28904998e+00 4.53530550e-01
1.00535941e+00 -1.02497840e+00 -7.29581833e-01 7.34019339e-01
7.93603361e-01 7.12827966e-02 -4.86926645e-01 2.89783329e-01
2.43760109e-01 -5.30581117e-01 1.24236214e+00 -1.06329668e+00
2.11396217e-01 -6.51169360e-01 -7.06267595e-01 -1.13889563e+00
-3.28739017e-01 -1.56359226e-01 8.10636114e-03 1.17066526e+00
3.62773150e-01 -9.49533507e-02 8.48352730e-01 6.35024846e-01
2.95139313e-01 -3.60282004e-01 -6.82730019e-01 -7.11336672e-01
1.12343900e-01 -3.94256324e-01 2.37129666e-02 1.24150658e+00
-2.54020512e-01 3.38331848e-01 -4.34089690e-01 3.76368225e-01
1.05951524e+00 4.13188487e-01 6.18108571e-01 -8.87178183e-01
-1.41432434e-01 2.15182781e-01 -4.15878713e-01 -9.06155348e-01
4.32844132e-01 -1.57711911e+00 -4.45858575e-02 -9.81199801e-01
5.64826310e-01 -4.36743557e-01 -6.82498574e-01 8.22048008e-01
-8.16196352e-02 1.07633960e+00 1.60939276e-01 6.19536042e-01
-1.18847728e+00 5.16496778e-01 8.59723091e-01 -5.78637183e-01
-2.22108383e-02 -4.85562563e-01 -7.25344419e-01 7.35887945e-01
4.16301876e-01 -5.87520361e-01 -5.56988716e-01 -5.76717734e-01
-7.14584216e-02 -4.43651706e-01 4.73596454e-01 -7.00092316e-01
4.62103814e-01 1.08760204e-02 8.36367965e-01 -4.83422875e-01
8.38404149e-02 -1.05903757e+00 -2.95242161e-01 5.80376834e-02
-8.90455127e-01 -1.79831818e-01 -1.27802327e-01 1.04679024e+00
-2.74795026e-01 -2.93760717e-01 1.14419770e+00 -2.45507851e-01
-9.02172625e-01 5.58945179e-01 1.19497456e-01 2.70309180e-01
1.13009739e+00 1.73633862e-02 -4.18615401e-01 -2.47880191e-01
-8.48585010e-01 3.21739197e-01 8.87109339e-01 5.56627870e-01
8.23064268e-01 -1.74013376e+00 -3.49718660e-01 4.20260310e-01
8.70983124e-01 -2.09337607e-01 1.55331373e-01 3.01997870e-01
-4.09978896e-01 1.97103366e-01 -1.70340404e-01 -8.28214407e-01
-1.09217656e+00 1.37254822e+00 1.59772530e-01 2.72286475e-01
-5.26062727e-01 9.37242210e-01 1.66123763e-01 -6.98589027e-01
5.69897175e-01 -5.84267899e-02 3.28897405e-03 -7.14092404e-02
8.40182185e-01 7.51582207e-03 -5.73464595e-02 -1.00118375e+00
-3.80449057e-01 7.32151985e-01 -2.83839554e-01 1.14871122e-01
1.28839386e+00 -4.62362468e-01 -1.71572611e-01 3.67895067e-01
1.95952332e+00 -1.43219798e-03 -1.33294225e+00 -7.37939119e-01
1.60528362e-01 -4.97127980e-01 2.47842725e-02 -7.40584612e-01
-1.22426140e+00 6.81406856e-01 9.37085748e-01 -3.28859478e-01
8.63890052e-01 4.02172774e-01 2.67718613e-01 9.26800549e-01
2.80838311e-01 -1.09463716e+00 5.17744780e-01 3.88975859e-01
8.18018675e-01 -1.91472077e+00 -9.95963439e-02 -8.24028552e-02
-8.45565736e-01 7.90714562e-01 5.97213268e-01 -5.50113082e-01
7.53460169e-01 -8.72635543e-02 4.75941062e-01 -1.92305773e-01
-7.02342689e-01 -2.72602946e-01 6.13743961e-01 7.56397307e-01
1.13671429e-01 -1.74373507e-01 3.88056129e-01 1.58417478e-01
2.55509079e-01 -1.23404764e-01 1.70069888e-01 8.45284581e-01
-5.61482012e-01 -9.41305935e-01 -3.25505108e-01 3.04804564e-01
-2.12076098e-01 -3.94975603e-01 -5.02572119e-01 3.09942156e-01
-1.08307578e-01 8.86124253e-01 -6.04950562e-02 -2.18926117e-01
-8.94747376e-02 2.39312023e-01 3.62839818e-01 -8.26015353e-01
-2.75214583e-01 -1.54524729e-01 -5.00754893e-01 -7.18896508e-01
-6.00604117e-01 -4.93618986e-03 -1.10922718e+00 1.42515123e-01
-1.87937483e-01 8.63913596e-02 3.96028608e-01 7.40773141e-01
5.10052204e-01 -2.16756985e-01 8.73469293e-01 -1.02269208e+00
-8.13228667e-01 -8.62154186e-01 -7.37433434e-01 1.08388114e+00
4.75138754e-01 -4.61606592e-01 -4.43426728e-01 5.11980236e-01] | [10.041766166687012, 1.9345341920852661] |
3911df60-409b-466d-a1cb-a759be858506 | functional-regularisation-for-continual | 1901.11356 | null | https://arxiv.org/abs/1901.11356v4 | https://arxiv.org/pdf/1901.11356v4.pdf | Functional Regularisation for Continual Learning with Gaussian Processes | We introduce a framework for Continual Learning (CL) based on Bayesian inference over the function space rather than the parameters of a deep neural network. This method, referred to as functional regularisation for Continual Learning, avoids forgetting a previous task by constructing and memorising an approximate posterior belief over the underlying task-specific function. To achieve this we rely on a Gaussian process obtained by treating the weights of the last layer of a neural network as random and Gaussian distributed. Then, the training algorithm sequentially encounters tasks and constructs posterior beliefs over the task-specific functions by using inducing point sparse Gaussian process methods. At each step a new task is first learnt and then a summary is constructed consisting of (i) inducing inputs -- a fixed-size subset of the task inputs selected such that it optimally represents the task -- and (ii) a posterior distribution over the function values at these inputs. This summary then regularises learning of future tasks, through Kullback-Leibler regularisation terms. Our method thus unites approaches focused on (pseudo-)rehearsal with those derived from a sequential Bayesian inference perspective in a principled way, leading to strong results on accepted benchmarks. | ['Yee Whye Teh', 'Jonathan Schwarz', 'Michalis K. Titsias', 'Alexander G. de G. Matthews', 'Razvan Pascanu'] | 2019-01-31 | null | https://openreview.net/forum?id=HkxCzeHFDB | https://openreview.net/pdf?id=HkxCzeHFDB | iclr-2020-1 | ['sequential-bayesian-inference'] | ['time-series'] | [ 3.82532984e-01 3.12757492e-01 2.04141080e-01 -3.75204295e-01
-6.77177668e-01 -1.46536097e-01 1.09524751e+00 2.27078289e-01
-8.45504165e-01 8.89922321e-01 3.12799066e-01 7.35124126e-02
-4.05641288e-01 -7.08507955e-01 -1.12568891e+00 -1.07486463e+00
7.98910204e-03 9.29846823e-01 1.44177929e-01 3.56588989e-01
4.37487245e-01 2.45562971e-01 -1.71891761e+00 1.51186183e-01
5.38055420e-01 9.32575762e-01 5.71332335e-01 4.80161607e-01
-7.81611279e-02 7.92814553e-01 -5.65175891e-01 -4.22210366e-01
-4.93796654e-02 -1.59496054e-01 -9.71135199e-01 2.98771381e-01
-9.47692990e-03 2.18035262e-02 -1.68181702e-01 1.04011512e+00
4.90274839e-02 7.17437387e-01 1.04790008e+00 -7.58526742e-01
-5.96671045e-01 6.41185045e-01 -1.52377039e-01 6.07736502e-03
2.06975877e-01 4.31612968e-01 8.21444333e-01 -1.21148956e+00
3.97011071e-01 1.25114608e+00 8.98077190e-01 3.48022789e-01
-1.95792556e+00 -1.33146435e-01 3.33844185e-01 -5.35863526e-02
-1.26152790e+00 -4.97997165e-01 6.34148538e-01 -4.72721875e-01
9.90744233e-01 -1.01765625e-01 9.10350025e-01 1.35680437e+00
6.81214035e-01 7.98282504e-01 1.12606597e+00 -4.93149191e-01
8.48311245e-01 2.92615473e-01 1.48449734e-01 4.50185388e-01
-8.78810734e-02 5.47841005e-02 -8.93993556e-01 -3.16267222e-01
7.59228289e-01 1.56267822e-01 -2.05608055e-01 -5.97283125e-01
-1.26304317e+00 7.35164523e-01 2.76342660e-01 2.93422192e-01
-1.15655744e+00 4.35268253e-01 3.64995837e-01 2.63510883e-01
4.90553439e-01 2.10230380e-01 -5.00262558e-01 -8.43483582e-02
-1.07064104e+00 4.11950052e-01 8.43082547e-01 5.87127268e-01
1.10478032e+00 5.85201271e-02 -5.35344303e-01 6.97620392e-01
4.05081689e-01 1.76820502e-01 6.06090426e-01 -1.15134227e+00
-3.77322994e-02 -1.03966214e-01 3.21508884e-01 -5.61988175e-01
1.25574591e-02 -5.47164917e-01 -6.57311499e-01 2.11844981e-01
3.08549255e-01 -1.55148476e-01 -9.24121857e-01 1.94458830e+00
1.09071448e-01 1.98919579e-01 1.74671244e-02 2.56107599e-01
-1.19644739e-01 7.67686546e-01 4.46841687e-01 -5.59475064e-01
1.10064495e+00 -6.68542206e-01 -6.68613076e-01 -3.14515561e-01
6.58988506e-02 -3.58817488e-01 1.12606227e+00 9.04648662e-01
-1.43084979e+00 -9.06495810e-01 -7.88410544e-01 2.16357544e-01
-1.78036690e-01 -2.82562464e-01 5.00366628e-01 4.02606785e-01
-1.49559069e+00 9.36402917e-01 -9.67259467e-01 2.68873218e-02
4.90913182e-01 2.71931469e-01 -1.94651127e-01 6.90862723e-03
-9.85664129e-01 1.02087367e+00 8.77510428e-01 5.97190000e-02
-1.46361387e+00 -7.81309485e-01 -7.95658171e-01 1.41587541e-01
2.38729045e-01 -1.14897490e+00 1.47212803e+00 -9.86163437e-01
-1.73930359e+00 6.06999576e-01 -5.09331882e-01 -7.66966403e-01
3.60743403e-01 -3.87337863e-01 1.48444667e-01 -1.79364875e-01
-1.06078245e-01 8.22807550e-01 1.69207776e+00 -1.45963132e+00
-6.02499664e-01 -4.12471265e-01 -1.76607236e-01 4.25904870e-01
-1.07398123e-01 -4.08982277e-01 -3.93040270e-01 -5.15386045e-01
2.12549046e-01 -8.72655511e-01 -2.03997910e-01 -2.18225658e-01
-9.41054523e-02 -6.41543090e-01 3.48755449e-01 -5.59188247e-01
1.00389040e+00 -2.31647754e+00 5.63158572e-01 2.14141324e-01
2.58071780e-01 -1.55154094e-01 1.43284932e-01 7.93785229e-02
-1.27499685e-01 -4.25796241e-01 -4.46393460e-01 -1.05926907e+00
3.89535017e-02 4.42851245e-01 -5.52439272e-01 6.11589372e-01
2.12888010e-02 8.24855626e-01 -1.09068441e+00 -4.69240062e-02
3.03220123e-01 5.85429013e-01 -3.79611045e-01 1.80074856e-01
-4.94975030e-01 2.43783951e-01 -1.26657054e-01 -3.38259265e-02
3.21825206e-01 -1.24291465e-01 -3.23882923e-02 8.19795430e-02
-6.87557906e-02 2.62378722e-01 -1.18887222e+00 1.96178830e+00
-6.28041387e-01 2.95970470e-01 -1.06947429e-01 -1.26460075e+00
1.03962576e+00 5.60672581e-01 2.82464087e-01 -3.95468250e-02
-8.75828415e-02 -9.26244445e-03 -4.51367259e-01 -6.76953942e-02
1.88918307e-01 -6.17016375e-01 2.14929000e-01 5.46664000e-01
7.68492162e-01 -3.77866924e-01 -2.76357140e-02 1.20798023e-02
9.34716165e-01 4.83218580e-01 3.00743580e-01 -4.07047153e-01
5.00617266e-01 -3.87624919e-01 2.90893167e-01 1.08271277e+00
-6.05415180e-02 3.89501423e-01 3.26274961e-01 -5.67164779e-01
-9.23541009e-01 -1.53612506e+00 -2.17800617e-01 1.30245960e+00
-5.78529000e-01 -1.86721742e-01 -5.71813464e-01 -4.67932910e-01
1.38489932e-01 1.42971754e+00 -1.08923185e+00 -4.89010841e-01
-2.74557441e-01 -7.06926107e-01 -7.70975044e-03 3.24895054e-01
2.73752272e-01 -1.45969546e+00 -6.31509423e-01 4.67575908e-01
2.33768523e-01 -3.68358165e-01 -2.21958071e-01 9.43546593e-01
-1.18126082e+00 -4.01831537e-01 -8.63254011e-01 -8.06101918e-01
6.13369644e-01 -1.62927970e-01 1.16635871e+00 -6.46651983e-01
7.42083713e-02 7.26061404e-01 2.79016346e-01 -5.10966837e-01
-3.40533227e-01 -2.84476161e-01 2.58033097e-01 4.33410019e-01
3.10196429e-01 -8.51223648e-01 -3.98218155e-01 -3.55760366e-01
-8.37118268e-01 -6.78320378e-02 7.14437842e-01 7.85413742e-01
8.07535946e-01 3.56176406e-01 5.02564430e-01 -1.00689816e+00
9.22148824e-01 -5.38908124e-01 -3.98202449e-01 2.02553928e-01
-7.47940838e-01 5.02257347e-01 3.84426773e-01 -6.90788329e-01
-1.58538485e+00 1.45444825e-01 1.00920327e-01 -5.50604761e-01
-3.96285236e-01 4.41552758e-01 1.45054400e-01 4.46130425e-01
9.29279268e-01 8.41071427e-01 2.16286659e-01 -4.29828584e-01
5.67390323e-01 1.92056939e-01 7.31346130e-01 -9.30314958e-01
4.95707244e-01 4.91078824e-01 -9.34723690e-02 -5.50575256e-01
-1.04737020e+00 -3.14333856e-01 -7.63647735e-01 -2.59962857e-01
8.68930697e-01 -8.32298517e-01 -6.75009966e-01 7.24568069e-01
-1.29300869e+00 -5.94865739e-01 -9.23968136e-01 5.65705180e-01
-1.07222998e+00 4.47903154e-03 -5.83103538e-01 -9.70948875e-01
-7.83994868e-02 -9.14244711e-01 1.02838731e+00 -2.73743924e-02
-4.98828620e-01 -1.43641150e+00 2.99423724e-01 -1.42386392e-01
6.15900576e-01 -2.62821734e-01 9.59316611e-01 -6.67153716e-01
-2.26889357e-01 -1.86847016e-01 2.20512390e-01 7.44668424e-01
1.67501047e-02 -4.76297706e-01 -1.22038448e+00 -2.66277850e-01
8.61203492e-01 -4.47823972e-01 1.06126273e+00 6.58554852e-01
1.28614914e+00 -2.72719562e-01 -2.23297164e-01 2.20888123e-01
1.24966502e+00 -7.62446821e-02 5.44396818e-01 2.28904217e-01
3.14620018e-01 7.27634311e-01 1.38178453e-01 5.64342082e-01
9.73666534e-02 1.20684944e-01 1.08137943e-01 6.18132830e-01
3.27924751e-02 -3.84391338e-01 4.46507096e-01 4.84008670e-01
-1.44075230e-01 3.73070717e-01 -8.77236784e-01 5.62770963e-01
-1.95114267e+00 -9.35287893e-01 3.48272145e-01 2.45702100e+00
1.41036689e+00 5.07257760e-01 -1.33985132e-01 1.12525478e-01
6.28624976e-01 3.62569233e-03 -6.74322367e-01 -1.65286273e-01
2.71999300e-01 3.83662671e-01 -2.87274510e-04 7.05600321e-01
-8.82680118e-01 8.40922773e-01 6.48551083e+00 9.74306107e-01
-8.01742017e-01 1.58851862e-01 5.98792970e-01 9.16183144e-02
-3.22959840e-01 1.36475280e-01 -9.55754697e-01 3.27738285e-01
1.34296024e+00 -2.48032793e-01 6.94844246e-01 8.98912787e-01
-6.05264455e-02 -3.57149243e-01 -1.46065080e+00 8.32204044e-01
-7.85104260e-02 -1.14573610e+00 2.56687015e-01 -1.00914948e-01
8.36977005e-01 -8.79042596e-02 4.14828986e-01 7.60518372e-01
5.93225479e-01 -9.05349433e-01 1.00232661e+00 1.26708269e+00
4.54350829e-01 -6.45538568e-01 4.06285614e-01 8.43507767e-01
-6.24448419e-01 -1.07084468e-01 -5.27393043e-01 2.58820970e-02
2.33645607e-02 8.81893873e-01 -8.28847468e-01 8.69403332e-02
5.21661818e-01 6.06209517e-01 -3.54331851e-01 1.01219583e+00
-5.09371638e-01 7.51667559e-01 -2.30282024e-01 1.80577785e-01
1.66280836e-01 -2.24016950e-01 6.06859565e-01 1.17075777e+00
2.15941280e-01 -5.32049358e-01 3.95117253e-02 1.05455780e+00
1.31866038e-01 -1.18992999e-01 -6.12743258e-01 4.55460489e-01
3.65752608e-01 9.61257577e-01 -5.71227491e-01 -4.92820114e-01
-1.24800205e-01 1.09425628e+00 6.17014706e-01 7.32341409e-01
-5.38948119e-01 1.50787514e-02 1.01760149e-01 -4.25158329e-02
4.80102748e-01 -3.02312464e-01 -2.89530247e-01 -7.13454843e-01
-1.56084508e-01 -5.35443604e-01 2.39864409e-01 -8.75101209e-01
-1.33111799e+00 3.33977193e-01 2.91992307e-01 -5.04242837e-01
-4.93893743e-01 -2.86116570e-01 -6.41552329e-01 1.31114316e+00
-1.15894473e+00 -7.05170929e-01 1.37199730e-01 7.97633231e-01
8.44814956e-01 -2.19253004e-01 1.10444689e+00 -5.09861350e-01
-8.18334073e-02 1.94888916e-02 1.89615816e-01 -5.78379333e-01
5.26256204e-01 -1.67806852e+00 1.59692794e-01 3.23825240e-01
2.77021796e-01 1.07823026e+00 8.84093344e-01 -6.52200758e-01
-1.16199625e+00 -1.01430118e+00 9.39806759e-01 -6.56108916e-01
7.96398818e-01 -4.56389874e-01 -1.33420515e+00 9.42794263e-01
1.66222259e-01 -1.31332636e-01 4.07725066e-01 3.80903870e-01
-1.12682441e-02 -1.70967355e-01 -1.01252246e+00 4.44075525e-01
6.98295057e-01 -6.00529850e-01 -1.10150576e+00 5.36986530e-01
6.47970736e-01 -4.27821884e-03 -7.52714515e-01 1.25295939e-02
2.64539331e-01 -9.05394793e-01 8.89101803e-01 -5.18609166e-01
-5.30038364e-02 -1.27235189e-01 -7.37972483e-02 -1.65605307e+00
-6.53804779e-01 -8.37008893e-01 -5.59724152e-01 8.62174332e-01
1.63802624e-01 -6.79584444e-01 5.78990817e-01 6.65936291e-01
-1.40623629e-01 -7.43809700e-01 -9.61485684e-01 -7.07377732e-01
1.31771401e-01 -7.24994540e-01 2.92973388e-02 3.69524956e-01
-3.60062778e-01 6.14294708e-01 -1.15500212e-01 -1.73049003e-01
9.99773920e-01 -4.23397571e-01 4.38945442e-01 -1.47137272e+00
-5.79718709e-01 -4.37532037e-01 4.81692590e-02 -1.16337168e+00
4.55890089e-01 -8.49701285e-01 4.81533170e-01 -1.23920596e+00
1.70640543e-01 -2.03264907e-01 -4.12573993e-01 3.99837315e-01
-2.21855357e-01 -3.65770847e-01 -1.32821739e-01 5.42180479e-01
-7.79780030e-01 9.17920053e-01 1.02174771e+00 2.11330354e-01
-2.98646301e-01 4.10531312e-01 -5.98167419e-01 9.09151435e-01
6.10677063e-01 -6.25532687e-01 -6.05322659e-01 -2.12814763e-01
2.70935535e-01 8.72824043e-02 4.64352548e-01 -1.17923844e+00
5.49236059e-01 1.07294373e-01 6.80062711e-01 -3.63529325e-01
5.01183212e-01 -6.01277769e-01 2.33542174e-01 5.03693223e-01
-6.32634103e-01 -2.66984731e-01 8.72810334e-02 1.18416929e+00
9.19493847e-03 -6.65186763e-01 7.42487967e-01 -3.58962744e-01
-4.52268332e-01 8.74403268e-02 -7.50584662e-01 -2.74625510e-01
9.14805353e-01 -1.43282443e-01 3.41295421e-01 -3.40004802e-01
-1.41488516e+00 1.79847367e-02 1.30506828e-01 1.65862609e-02
6.98148608e-01 -1.12818205e+00 -6.80979550e-01 2.77876109e-01
-2.21953541e-01 3.64392251e-01 8.88082087e-02 7.41748095e-01
2.00506747e-01 4.80903506e-01 -7.07466304e-02 -7.13403165e-01
-5.30812204e-01 6.57718420e-01 1.83986753e-01 -3.59401971e-01
-6.23929739e-01 1.07074308e+00 3.29751223e-01 -1.85191259e-01
5.86430371e-01 -2.06758782e-01 -1.94857880e-01 1.93679646e-01
5.31885862e-01 2.12086469e-01 1.99079335e-01 -1.70613453e-01
2.11585864e-01 3.90910404e-03 -2.11821079e-01 -6.99077308e-01
1.54017234e+00 -1.95289597e-01 -3.19772720e-01 1.34639490e+00
9.53488767e-01 -4.77422297e-01 -1.98098648e+00 -7.22285569e-01
1.35781363e-01 -2.33145013e-01 2.45114908e-01 -5.99289834e-01
-3.39793146e-01 8.18912745e-01 2.58707136e-01 1.69924423e-01
7.82197177e-01 3.63018587e-02 1.66241229e-01 6.80158973e-01
4.54732269e-01 -1.10701871e+00 4.64214325e-01 7.62350619e-01
1.09845936e+00 -7.62684107e-01 -1.74022630e-01 4.21956688e-01
-6.61705554e-01 1.22891307e+00 1.07310079e-01 -3.50075781e-01
1.05098605e+00 -5.60681820e-02 -6.02377057e-01 -2.55075544e-01
-1.00460005e+00 1.56328723e-01 3.85951608e-01 7.01496005e-01
2.58630872e-01 -8.78445655e-02 1.27265230e-01 6.56545341e-01
-1.64544716e-01 2.11103305e-01 3.33412476e-02 9.62296903e-01
-7.49268413e-01 -7.64920175e-01 -4.25854862e-01 5.85739136e-01
-2.15381756e-01 -2.11474761e-01 2.58594632e-01 3.27148855e-01
7.46805593e-02 3.96500736e-01 2.19479576e-01 1.61928922e-01
3.40944715e-02 8.91027689e-01 6.56360924e-01 -1.10134053e+00
-2.88015783e-01 1.31823197e-01 -2.68013030e-01 -3.59421551e-01
-1.55659184e-01 -1.16202164e+00 -1.04436421e+00 -4.56584021e-02
7.63739422e-02 1.68888524e-01 7.86606967e-01 1.09658432e+00
-8.57117213e-03 5.16915560e-01 3.70666504e-01 -1.25895548e+00
-1.05637121e+00 -1.10917711e+00 -6.93371594e-01 4.32032526e-01
2.28405029e-01 -8.33808124e-01 -4.88102227e-01 4.47572261e-01] | [7.233018398284912, 3.7820887565612793] |
d64d9696-f536-4ed8-8da0-a651b1d35494 | in-silico-identification-of-potential-natural | 2006.00652 | null | https://arxiv.org/abs/2006.00652v1 | https://arxiv.org/pdf/2006.00652v1.pdf | In silico identification of potential natural product inhibitors of human proteases key to SARS-CoV-2 infection | Presently, there are no approved drugs or vaccines to treat COVID-19 which has spread to over 200 countries and is responsible for over 3,65,000 deaths worldwide. Recent studies have shown that two human proteases, TMPRSS2 and cathepsin L, play a key role in host cell entry of SARS-CoV-2. Importantly, inhibitors of these proteases were shown to block SARS-CoV-2 infection. Here, we perform virtual screening of 14010 phytochemicals produced by Indian medicinal plants to identify natural product inhibitors of TMPRSS2 and cathepsin L. We built a homology model of TMPRSS2 as an experimentally determined structure is not available. AutoDock Vina was used to perform molecular docking of phytochemicals against TMPRSS2 model structure and cathepsin L crystal structure. Potential phytochemical inhibitors were filtered by comparing their docked binding energies with those of known inhibitors of TMPRSS2 and cathepsin L. Further, the ligand binding site residues and non-covalent protein-ligand interactions were used as an additional filter to identify phytochemical inhibitors that either bind to or form interactions with residues important for the specificity of the target proteases. We have identified 96 inhibitors of TMPRSS2 and 9 inhibitors of cathepsin L among phytochemicals of Indian medicinal plants. The top inhibitors of TMPRSS2 are Edgeworoside C, Adlumidine and Qingdainone, and of cathepsin L is Ararobinol. Interestingly, several herbal sources of identified phytochemical inhibitors have antiviral or anti-inflammatory use in traditional medicine. Further in vitro and in vivo testing is needed before clinical trials of the promising phytochemical inhibitors identified here. | ['Areejit Samal', 'Himansu S. Biswal', 'Nithin Rajan', 'Abhijit Rana', 'R. P. Vivek-Ananth'] | 2020-06-01 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 1.90273598e-01 -3.85955483e-01 -4.54088032e-01 -5.24860248e-02
-6.92299664e-01 -9.48005736e-01 6.82511702e-02 8.32695365e-01
-5.05497456e-01 1.12913430e+00 1.96815711e-02 -5.57066202e-01
3.49922121e-01 -4.25944746e-01 -5.80465317e-01 -6.72443032e-01
-3.68979424e-01 4.15723294e-01 6.91665933e-02 -2.60680288e-01
1.51502073e-01 8.96710575e-01 -4.78500038e-01 1.49680525e-01
1.36403358e+00 2.70647198e-01 2.33581111e-01 4.10138816e-01
5.25343195e-02 2.02892452e-01 -3.67104918e-01 1.10979527e-01
1.44300863e-01 -5.71137905e-01 -4.86412257e-01 -4.89395559e-01
-5.09205507e-04 -2.97952712e-01 5.47054410e-01 7.66511202e-01
1.74533471e-01 -1.38159692e-01 9.70812142e-01 -5.97073495e-01
-5.29491067e-01 -4.40883368e-01 -7.65643835e-01 -5.94417984e-03
5.89415789e-01 3.35431956e-02 6.15967214e-01 -1.26855922e+00
7.24873662e-01 1.00506377e+00 5.98288000e-01 6.30285501e-01
-1.31019676e+00 -6.94556892e-01 -3.62535864e-01 1.62410006e-01
-1.31915188e+00 -1.66550592e-01 2.44968861e-01 -5.92932820e-01
1.45548403e+00 2.47745305e-01 9.28116262e-01 4.58747029e-01
8.13533604e-01 4.35710609e-01 1.19127440e+00 7.68538788e-02
3.96956980e-01 2.43248463e-01 3.97544205e-02 6.29589617e-01
4.93178546e-01 -1.49866641e-01 -2.91813582e-01 -1.11342037e+00
6.43309712e-01 4.35699195e-01 -3.40642750e-01 -3.08337212e-01
-5.84440827e-01 1.15909135e+00 1.67870343e-01 2.39790514e-01
-5.04692733e-01 -3.21302772e-01 5.41792035e-01 -2.42894202e-01
4.57923599e-02 5.53540468e-01 -9.74352717e-01 5.27865648e-01
-3.55975449e-01 2.84868348e-02 6.93003833e-01 1.13255292e-01
5.79091311e-01 -2.14111105e-01 4.33155596e-01 5.80437660e-01
5.21021783e-01 8.45111609e-01 1.48916314e-03 -4.18504566e-01
-1.92334995e-01 5.41567922e-01 4.63662446e-01 -8.67987037e-01
-6.21073484e-01 4.70572263e-01 -3.02066773e-01 -6.88872337e-02
4.53136683e-01 -2.84239352e-01 -7.55539000e-01 1.71420157e+00
4.99839932e-01 1.96056813e-01 2.15744093e-01 9.34416354e-01
5.99559367e-01 8.44584823e-01 7.26031184e-01 -9.54128802e-01
1.88362348e+00 -4.32745248e-01 -4.47675943e-01 1.52542233e-01
5.05698979e-01 -9.15296674e-01 5.39490938e-01 3.77861053e-01
-8.37472081e-01 1.59415767e-01 -9.83844876e-01 4.51176763e-01
-4.33942288e-01 4.46813330e-02 6.25217915e-01 8.48734081e-01
-4.37810689e-01 4.95708257e-01 -1.40815806e+00 -6.88630342e-01
1.05191067e-01 6.93861961e-01 -3.64302129e-01 2.08408847e-01
-1.29422998e+00 1.11798418e+00 1.74677595e-01 -9.69738960e-02
-7.29170442e-01 -6.17638648e-01 -5.43441117e-01 -1.56729877e-01
7.14627514e-03 -8.53680432e-01 7.46272862e-01 -4.86194462e-01
-1.40893912e+00 6.56819522e-01 -4.94180650e-01 -4.25745606e-01
-4.05266881e-01 -4.09122378e-01 -4.59759235e-01 5.89945018e-01
1.61188170e-01 1.61370993e-01 1.55014679e-01 -9.55862522e-01
-3.16439450e-01 -6.84359789e-01 -3.24157476e-01 1.76210642e-01
2.40701377e-01 5.73836148e-01 2.06434932e-02 -5.03010213e-01
-4.60589193e-02 -9.82691705e-01 -6.93743706e-01 -4.77422982e-01
-3.98705155e-01 -1.55204520e-01 5.67278445e-01 -5.78809381e-01
4.95772630e-01 -1.48287570e+00 -2.85224438e-01 5.12125492e-01
3.84210655e-03 5.59346139e-01 -1.33799478e-01 9.84626949e-01
6.71351627e-02 5.59055865e-01 7.04614744e-02 9.87252235e-01
-6.54520750e-01 -2.67032772e-01 -4.11606342e-01 1.04312003e+00
2.11904615e-01 6.20153308e-01 -9.52015877e-01 -2.19653279e-01
6.45330176e-02 8.06297243e-01 -5.40004611e-01 -2.07780018e-01
-5.26980281e-01 8.81939352e-01 -1.36349976e+00 1.04395998e+00
1.01690471e+00 -9.73857567e-02 8.62305939e-01 -2.28806764e-01
-1.16363205e-01 4.15901244e-01 -4.29803491e-01 1.26845837e+00
2.91343629e-01 -1.61697417e-01 -1.81906506e-01 -2.88202316e-01
6.40930295e-01 6.52554035e-01 6.25326037e-01 -6.33872747e-01
1.81816354e-01 2.76823819e-01 -3.03967357e-01 -7.47201517e-02
1.10456496e-01 -4.77857590e-01 2.79064029e-01 -2.66819179e-01
-4.07788604e-01 5.26639462e-01 2.75496870e-01 2.92068124e-01
7.98283815e-01 1.49738848e-01 6.82548702e-01 -5.55525243e-01
7.45896399e-01 5.74011803e-01 8.56825650e-01 1.81027383e-01
-6.09855284e-04 -3.02348614e-01 2.80052900e-01 -3.53504479e-01
-7.85447657e-01 -1.07784605e+00 -4.09274548e-01 5.87001920e-01
1.08425692e-01 -6.13288105e-01 -6.81176305e-01 -2.77972937e-01
-3.04872692e-01 5.44486642e-01 -4.46821541e-01 3.73383254e-01
-4.30283815e-01 -8.35197270e-01 6.58804715e-01 -1.12071648e-01
1.55372590e-01 -5.40991068e-01 -5.29223859e-01 5.05287170e-01
6.39419928e-02 -7.55456328e-01 -4.38635528e-01 1.62821054e-01
-1.15135944e+00 -1.73235667e+00 -5.18607438e-01 -7.10039616e-01
4.92214203e-01 1.37719139e-01 2.78162897e-01 -1.33227021e-01
-2.30802312e-01 -1.04386009e-01 -3.13843414e-02 -7.96313643e-01
-2.71243960e-01 -4.97610927e-01 5.14881909e-01 -6.61701798e-01
7.26215482e-01 -4.36315417e-01 -9.98782396e-01 3.54623079e-01
-2.95832515e-01 -2.16175288e-01 4.36723739e-01 5.83483934e-01
9.79402184e-01 -3.59598756e-01 5.60331523e-01 -1.19805419e+00
5.91334403e-01 -6.98325574e-01 -6.35405421e-01 6.29226491e-02
-4.13430601e-01 -3.57286125e-01 8.50902975e-01 -3.20000589e-01
-1.05761445e+00 3.17529261e-01 -1.94555357e-01 7.22763455e-03
-1.02264158e-01 5.71152747e-01 -2.63640285e-01 -2.58377399e-02
6.28838539e-01 2.15934843e-01 2.59078965e-02 -5.97815275e-01
-1.89365014e-01 1.58896055e-02 2.27459922e-01 -5.57784021e-01
6.48718834e-01 6.37982547e-01 1.97568893e-01 -1.18697810e+00
-1.81231841e-01 -6.06343806e-01 1.94083810e-01 5.19563377e-01
1.29436934e+00 -1.07187474e+00 -1.36299348e+00 1.73724368e-01
-1.25332248e+00 1.77697569e-01 6.14599824e-01 8.63566697e-01
-1.62767991e-01 5.14523029e-01 -9.48432386e-01 -4.75632221e-01
-7.68198192e-01 -1.06398022e+00 3.59402984e-01 4.23896819e-01
-9.20382068e-02 -6.89582169e-01 9.47553754e-01 2.92405367e-01
4.26431485e-02 7.68005192e-01 1.34181988e+00 -1.22376788e+00
-4.66053486e-01 -7.97226131e-02 1.50205829e-04 -1.49524346e-01
3.01188260e-01 1.57427162e-01 -4.29332078e-01 -3.42944682e-01
7.02098198e-03 -2.06014350e-01 4.55498219e-01 6.20948315e-01
1.77503213e-01 -6.48239493e-01 -8.52737129e-01 5.43938816e-01
1.59308279e+00 1.03453672e+00 7.89345682e-01 3.22354198e-01
5.02400219e-01 3.67890000e-01 9.68861401e-01 4.82940733e-01
-1.28001153e-01 5.14773130e-01 3.34024578e-01 -2.07544208e-01
9.98330295e-01 -1.21678539e-01 3.17338169e-01 3.52013744e-02
-7.06336617e-01 3.65436338e-02 -7.10925937e-01 2.46229693e-01
-1.25559354e+00 -9.58567560e-01 -7.06858039e-01 2.27180576e+00
1.04044223e+00 -4.38015372e-01 1.95558965e-01 -7.78599858e-01
3.40557784e-01 -3.94218862e-01 -6.76016629e-01 -6.24806821e-01
-6.65737828e-03 6.77579820e-01 9.17785108e-01 3.74548584e-01
-8.34584594e-01 9.73940372e-01 6.53263235e+00 5.59312224e-01
-8.84047627e-01 -2.35108912e-01 8.14793333e-02 1.36725545e-01
-1.99307039e-01 7.20674872e-01 -6.65232062e-01 1.04784526e-01
9.33888435e-01 -1.43158391e-01 2.61315197e-01 6.14158869e-01
7.44829655e-01 -3.42080951e-01 -7.09731698e-01 5.61904192e-01
-5.40217936e-01 -1.65677619e+00 -3.10198851e-02 5.03571630e-01
5.76434553e-01 2.57339701e-02 -3.29474509e-01 -2.86920607e-01
2.00771257e-01 -9.15390790e-01 -9.86384526e-02 4.00812358e-01
4.88683403e-01 -1.22570145e+00 4.24081355e-01 5.01426589e-03
-1.20388138e+00 4.31204587e-01 -4.10720676e-01 4.59342837e-01
2.96434611e-01 1.11317441e-01 -1.03518093e+00 8.49724337e-02
3.33431989e-01 2.83996582e-01 -2.54092515e-01 8.04143727e-01
-2.90821701e-01 3.80024046e-01 -2.91781813e-01 7.74870291e-02
9.73096788e-02 -5.39614379e-01 4.51042742e-01 1.00333631e+00
-3.60685408e-01 8.78174365e-01 5.27720809e-01 5.62853634e-01
1.60649881e-01 7.94947565e-01 -3.84532452e-01 -4.12857205e-01
2.42577702e-01 9.88936901e-01 -7.69063592e-01 -3.74646276e-01
-5.75008035e-01 8.54091048e-01 -3.67545933e-01 5.81423759e-01
-6.39880419e-01 -1.22307196e-01 1.26573133e+00 1.45831615e-01
2.39547938e-01 1.14512935e-01 3.50136191e-01 -7.01744080e-01
-9.12178576e-01 -1.17862511e+00 5.89512110e-01 -3.55747581e-01
-9.61021721e-01 1.44020036e-01 -9.85668879e-03 -9.17100847e-01
8.47824141e-02 -3.61105144e-01 -5.88304758e-01 1.11089444e+00
-1.01973259e+00 -1.22760308e+00 4.91180688e-01 5.08081973e-01
3.29759330e-01 1.54581308e-01 1.40655422e+00 -3.20196986e-01
-2.89526194e-01 -6.05293512e-02 4.88112986e-01 -2.62604982e-01
9.27935779e-01 -8.09914589e-01 -4.64810207e-02 2.16653138e-01
-6.07201040e-01 1.54429293e+00 1.04574537e+00 -1.30326009e+00
-1.44525290e+00 -8.81697774e-01 8.42053950e-01 -1.89260900e-01
4.77589041e-01 3.48526798e-02 -8.64899158e-01 6.64187610e-01
2.75961280e-01 -7.26518333e-01 1.64578342e+00 -3.23212445e-01
-4.57555026e-01 7.29156017e-01 -1.46610010e+00 7.40170598e-01
1.36890322e-01 -4.21124965e-01 -4.57487464e-01 5.94945431e-01
4.81004030e-01 -3.54778558e-01 -1.17750251e+00 4.66637850e-01
9.21061277e-01 -7.24940240e-01 1.19225109e+00 -9.24253821e-01
-1.05060145e-01 -9.21776831e-01 5.08756004e-02 -8.72451723e-01
-2.21810862e-01 -5.17476559e-01 1.64920121e-01 5.76212049e-01
2.65513241e-01 -5.18587351e-01 5.94909549e-01 3.90181810e-01
2.36428320e-01 -7.35093474e-01 -4.66683716e-01 -5.62811136e-01
-2.23672111e-02 2.93178082e-01 1.77324086e-01 9.50219512e-01
3.26248735e-01 6.32309079e-01 -3.43335599e-01 1.10291198e-01
5.78572392e-01 1.94523618e-01 7.17656791e-01 -1.17606056e+00
-2.08522588e-01 1.69030905e-01 4.74891774e-02 -4.03721660e-01
-2.06036717e-01 -6.44400418e-01 -8.11178207e-01 -1.56353402e+00
4.79556650e-01 1.87119722e-01 -1.10970855e-01 4.63388056e-01
3.27782482e-01 5.63675612e-02 1.60253402e-02 5.49476027e-01
6.10003341e-03 1.77933291e-01 9.00670886e-01 -1.05919436e-01
-1.03013873e+00 -2.40219727e-01 -9.35529828e-01 1.03339350e+00
9.94585812e-01 -6.81893170e-01 -4.12351578e-01 5.99450529e-01
3.20537001e-01 1.22948699e-01 -8.51302072e-02 5.66403307e-02
-3.36516052e-01 -9.13901687e-01 6.26953006e-01 -1.04383636e+00
3.65709752e-01 -7.46688724e-01 7.31339276e-01 1.07993305e+00
5.28732657e-01 -2.47151390e-01 5.02695084e-01 6.13042295e-01
4.14306194e-01 1.01889521e-01 1.04636157e+00 -2.57973790e-01
-4.05020028e-01 -2.10557152e-02 -1.39255381e+00 -9.41620767e-02
9.05149937e-01 -6.80670917e-01 -6.81023151e-02 2.58198738e-01
-8.54848802e-01 -5.13680018e-02 1.03916609e+00 -2.47354303e-02
6.22293890e-01 -7.90096641e-01 -4.92972821e-01 -2.46346533e-01
3.44814211e-01 -7.30022848e-01 4.43140626e-01 1.21215260e+00
-9.85371113e-01 8.87275696e-01 -5.01987815e-01 -5.65059483e-01
-1.68697226e+00 1.11165595e+00 3.47819835e-01 -2.64178485e-01
-7.69040510e-02 3.30221087e-01 7.01887906e-01 1.46095306e-02
-2.57521123e-01 -4.77156527e-02 -4.98886526e-01 -1.32093027e-01
7.00909734e-01 3.66878301e-01 -1.91664636e-01 -9.08685327e-01
-1.08348644e+00 5.54781675e-01 -2.83787042e-01 7.40537822e-01
1.38061285e+00 9.33137909e-02 -4.59588081e-01 -1.94408774e-01
1.06477201e+00 6.16053820e-01 -4.75203961e-01 7.67188072e-02
-6.77950382e-02 -3.06492805e-01 -6.01401687e-01 -8.03718626e-01
-4.29270118e-01 7.90479407e-02 8.79612684e-01 -7.29993343e-01
1.10158300e+00 7.74854794e-03 1.00694537e+00 2.08824322e-01
5.24356365e-01 -7.96707094e-01 -2.04345927e-01 1.21313021e-01
5.52830219e-01 -7.98387289e-01 4.27173972e-02 -5.54611802e-01
-6.46339417e-01 6.28625810e-01 4.34052497e-01 -2.41729960e-01
2.63584703e-01 -8.76891464e-02 1.48430038e-02 -6.04672015e-01
-8.79016757e-01 2.29608729e-01 2.54989177e-01 5.55580139e-01
1.01729417e+00 3.10955811e-02 -1.48194957e+00 4.89042550e-01
4.61666584e-01 2.42808461e-02 4.26511586e-01 1.07203305e+00
-5.38625836e-01 -1.52716660e+00 -7.08250165e-01 6.74360618e-02
-1.05699801e+00 -2.66785383e-01 -8.31736028e-01 8.92613649e-01
-3.97133604e-02 8.23496044e-01 -4.49634373e-01 1.04886085e-01
2.60547191e-01 8.80814269e-02 5.41896522e-01 -4.76546019e-01
-7.24214375e-01 1.19974041e+00 2.47207671e-01 -3.75998080e-01
-5.97134352e-01 -3.67244273e-01 -2.06439924e+00 -3.73161465e-01
-5.17995298e-01 6.42552435e-01 6.18193686e-01 7.50738144e-01
1.56975091e-01 -3.45819533e-01 5.75435817e-01 -3.84429038e-01
-9.48878974e-02 -5.55918813e-01 -1.02089393e+00 3.75664271e-02
5.15587293e-02 -1.77860469e-01 1.83235198e-01 2.83879817e-01] | [4.647504806518555, 5.0618414878845215] |
9a804d52-b485-4c8a-bba1-609264429ca6 | data-free-sketch-based-image-retrieval | 2303.07775 | null | https://arxiv.org/abs/2303.07775v1 | https://arxiv.org/pdf/2303.07775v1.pdf | Data-Free Sketch-Based Image Retrieval | Rising concerns about privacy and anonymity preservation of deep learning models have facilitated research in data-free learning (DFL). For the first time, we identify that for data-scarce tasks like Sketch-Based Image Retrieval (SBIR), where the difficulty in acquiring paired photos and hand-drawn sketches limits data-dependent cross-modal learning algorithms, DFL can prove to be a much more practical paradigm. We thus propose Data-Free (DF)-SBIR, where, unlike existing DFL problems, pre-trained, single-modality classification models have to be leveraged to learn a cross-modal metric-space for retrieval without access to any training data. The widespread availability of pre-trained classification models, along with the difficulty in acquiring paired photo-sketch datasets for SBIR justify the practicality of this setting. We present a methodology for DF-SBIR, which can leverage knowledge from models independently trained to perform classification on photos and sketches. We evaluate our model on the Sketchy, TU-Berlin, and QuickDraw benchmarks, designing a variety of baselines based on state-of-the-art DFL literature, and observe that our method surpasses all of them by significant margins. Our method also achieves mAPs competitive with data-dependent approaches, all the while requiring no training data. Implementation is available at \url{https://github.com/abhrac/data-free-sbir}. | ['Anjan Dutta', 'Yi-Zhe Song', 'Ayan Kumar Bhunia', 'Abhra Chaudhuri'] | 2023-03-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chaudhuri_Data-Free_Sketch-Based_Image_Retrieval_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chaudhuri_Data-Free_Sketch-Based_Image_Retrieval_CVPR_2023_paper.pdf | cvpr-2023-1 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 1.33373484e-01 -2.82613546e-01 -4.11779225e-01 -4.26889598e-01
-1.39542961e+00 -1.10342550e+00 1.02678585e+00 -5.72887585e-02
-6.29568219e-01 5.74019492e-01 7.54417032e-02 -2.51479805e-01
-2.93340445e-01 -3.91893506e-01 -8.79441023e-01 -5.03158391e-01
3.59958336e-02 2.51803607e-01 -2.71909535e-01 -1.14211254e-02
1.85939327e-01 6.43655062e-01 -1.44247615e+00 3.69032592e-01
1.55374125e-01 1.09001637e+00 -4.47602332e-01 5.59338748e-01
1.23141065e-01 5.42976439e-01 -5.76190427e-02 -1.07542801e+00
8.11401188e-01 -3.68413143e-02 -8.11326087e-01 -2.44882718e-01
1.30302465e+00 -9.06242192e-01 -7.75394499e-01 6.79712474e-01
6.55923128e-01 1.46392688e-01 7.85867810e-01 -1.88086414e+00
-1.01923382e+00 2.90025979e-01 -5.08258581e-01 -4.35769498e-01
1.81789681e-01 -8.60965103e-02 1.42612147e+00 -1.16903806e+00
8.41835380e-01 1.17234576e+00 5.23242950e-01 8.75496387e-01
-1.51248658e+00 -8.65587890e-01 -1.44251943e-01 8.04330111e-02
-1.62650049e+00 -8.08982253e-01 7.98483074e-01 -3.75603318e-01
3.23168486e-01 5.14488041e-01 5.97777404e-02 1.31920004e+00
-4.96832103e-01 1.23295629e+00 1.05447793e+00 -4.83016670e-01
1.58239052e-01 3.49372000e-01 -4.97837253e-02 6.80412233e-01
1.12048529e-01 1.27681836e-01 -7.25968242e-01 -5.00469744e-01
6.20067954e-01 2.48045236e-01 -2.43496835e-01 -1.12704015e+00
-1.22944534e+00 7.55340695e-01 5.02155423e-01 1.84251778e-02
2.13144377e-01 3.54980230e-01 6.93681717e-01 7.38985181e-01
4.56230730e-01 2.29931548e-01 -4.27674115e-01 2.56237864e-01
-1.06265116e+00 3.67152423e-01 6.58170998e-01 1.18962288e+00
6.78431749e-01 -5.57232141e-01 -1.27184451e-01 9.54380751e-01
4.45061326e-02 5.63576460e-01 -1.68780908e-02 -1.10038996e+00
4.96277839e-01 1.29564136e-01 1.46276340e-01 -8.04391026e-01
2.81027973e-01 2.60731012e-01 -9.06389058e-01 3.48637789e-01
7.01031685e-01 2.92907774e-01 -6.30485713e-01 1.93388617e+00
2.81689595e-02 -1.57780290e-01 2.41981316e-02 8.16973507e-01
6.31083131e-01 3.39621782e-01 -5.86573444e-02 1.79764405e-01
1.10756826e+00 -7.46865392e-01 -2.63240933e-01 1.86224967e-01
6.11428261e-01 -7.30168462e-01 1.43325651e+00 4.74724829e-01
-8.86024833e-01 -1.86879858e-01 -9.86528456e-01 -7.21364319e-01
-6.43431842e-01 3.10505509e-01 6.28595948e-01 6.73169553e-01
-1.24951994e+00 5.56348979e-01 -4.97231543e-01 -4.90199655e-01
1.03931606e+00 1.89045846e-01 -1.16388023e+00 -6.68749034e-01
-8.45919967e-01 4.93203342e-01 -1.52329534e-01 -1.61957815e-01
-9.59935546e-01 -8.82916451e-01 -5.94956756e-01 -1.57389939e-01
3.51765871e-01 -5.23159385e-01 1.02482080e+00 -9.00478184e-01
-1.03732133e+00 1.37103903e+00 9.75466371e-02 -7.68653303e-02
1.02537489e+00 -2.03630209e-01 -2.47933473e-02 1.76423475e-01
-1.08309947e-01 8.95733953e-01 9.82806742e-01 -1.39318824e+00
-1.73698604e-01 -4.93170828e-01 3.60306144e-01 -4.59925495e-02
-7.13308036e-01 -1.24508746e-01 -6.90098524e-01 -4.14472044e-01
-4.12536353e-01 -1.10513997e+00 2.31398284e-01 1.09176540e+00
-4.26169991e-01 -2.50821918e-01 8.45799685e-01 -5.10801971e-01
8.34510505e-01 -2.46292257e+00 -8.63681436e-02 1.98033348e-01
2.78369278e-01 3.47799510e-01 -5.76869249e-01 6.79486513e-01
8.80180567e-04 2.83615768e-01 -1.34499818e-01 -8.75918806e-01
4.77083087e-01 7.63992742e-02 -6.31443024e-01 6.50128722e-01
7.78528154e-02 8.47893834e-01 -7.20166743e-01 -5.46302438e-01
2.63528198e-01 5.58453321e-01 -5.70839465e-01 1.78335771e-01
-9.68748108e-02 1.86449111e-01 -2.21694231e-01 7.36811936e-01
1.00343263e+00 -1.80640832e-01 1.40864462e-01 -2.56932139e-01
2.32251257e-01 -2.32793093e-02 -1.05940497e+00 2.11041498e+00
-5.28264761e-01 7.69584477e-01 1.12755641e-01 -8.70131910e-01
6.81458056e-01 2.97730088e-01 2.63668686e-01 -6.66387439e-01
-1.60619259e-01 3.01714242e-01 -6.64065003e-01 -2.86054790e-01
3.11494410e-01 -1.88851897e-02 -1.86553851e-01 6.92825437e-01
1.10002942e-01 5.15076406e-02 -8.06555673e-02 5.32480419e-01
1.03347290e+00 1.46032631e-01 -9.85467359e-02 -5.07776141e-02
3.95740002e-01 -4.55055594e-01 1.80296719e-01 8.24799657e-01
-2.21692950e-01 7.85005689e-01 3.50861996e-01 -4.88113105e-01
-1.23156977e+00 -1.34282398e+00 -1.89690262e-01 1.40240598e+00
6.09429646e-03 -4.99361783e-01 -3.52525592e-01 -8.71915042e-01
3.73188287e-01 4.15346444e-01 -6.85029328e-01 7.73025602e-02
-2.11339369e-01 -1.33386850e-01 7.42117763e-01 3.05236310e-01
2.98873603e-01 -6.75933421e-01 -6.23550676e-02 -4.48710650e-01
3.45982276e-02 -8.88792813e-01 -6.85107112e-01 -2.87766725e-01
-5.02976120e-01 -9.54350233e-01 -8.92080605e-01 -4.82812047e-01
6.75403237e-01 4.33541834e-01 1.09062886e+00 2.01030254e-01
-5.54652750e-01 8.82080495e-01 -1.19230926e-01 -2.81069279e-01
-2.19435859e-02 1.67769462e-01 -1.37228191e-01 1.36219561e-01
3.80168885e-01 -7.22094953e-01 -8.63754690e-01 2.67914236e-01
-1.17652488e+00 -1.86714888e-01 5.40738285e-01 8.83180320e-01
4.84611452e-01 -5.21461368e-01 4.64592636e-01 -8.00967932e-01
3.76326084e-01 -3.02668840e-01 -6.26273274e-01 6.53370976e-01
-4.33534354e-01 -2.84579229e-02 4.36965317e-01 -4.89862531e-01
-6.02919638e-01 1.03227310e-01 2.56078273e-01 -9.78080928e-01
-2.38430556e-02 1.16549820e-01 -4.72383738e-01 -2.72134781e-01
6.17176354e-01 4.84031774e-02 2.55009651e-01 -5.73604047e-01
8.70724678e-01 7.98189878e-01 5.14844954e-01 -8.88077617e-01
8.90801966e-01 7.21465528e-01 7.47036040e-02 -6.27584696e-01
-5.87224424e-01 -3.65850568e-01 -6.62877560e-01 1.05468137e-02
4.59845424e-01 -1.11113381e+00 -8.41959000e-01 3.27067256e-01
-9.63003039e-01 -1.16808504e-01 -2.20735773e-01 2.02086687e-01
-6.98267102e-01 4.66548741e-01 -4.72864747e-01 -9.73011017e-01
-5.25357068e-01 -7.26423562e-01 1.16360438e+00 -2.73510575e-01
-2.31770352e-02 -8.54412138e-01 1.08944029e-01 4.47403461e-01
3.79107803e-01 2.02420071e-01 8.76173198e-01 -7.92735577e-01
-1.03041947e+00 -5.57235897e-01 -5.37543058e-01 5.46835065e-01
1.15797736e-01 -1.87757201e-02 -1.48854518e+00 -5.89644372e-01
-6.69921339e-01 -1.06159949e+00 8.85313690e-01 -2.15303585e-01
1.50915408e+00 -4.93424684e-01 -2.24253401e-01 7.24614739e-01
1.55276072e+00 -4.67491597e-01 6.63098931e-01 2.09187828e-02
7.39099562e-01 5.18432736e-01 4.17863309e-01 4.06444997e-01
3.83362889e-01 6.60318255e-01 2.49325573e-01 -9.39449817e-02
-5.90173155e-02 -6.36237323e-01 4.54242006e-02 2.32711464e-01
2.19861224e-01 -2.06114754e-01 -6.24090552e-01 8.00629556e-01
-1.86074829e+00 -8.53215933e-01 4.11318600e-01 2.59035420e+00
9.69023347e-01 -4.74607468e-01 2.14626417e-01 -2.50676777e-02
3.71814221e-01 2.73324490e-01 -6.39620900e-01 -1.30648792e-01
-1.10445380e-01 2.35753283e-01 5.63671470e-01 4.01307106e-01
-1.32320201e+00 8.27323735e-01 5.46385479e+00 1.05344999e+00
-1.08363986e+00 1.57704934e-01 6.29653037e-01 -6.38397753e-01
-5.25839269e-01 -1.05607279e-01 -3.60420048e-01 1.71700522e-01
5.85454345e-01 -1.73611283e-01 7.72880077e-01 9.65905726e-01
-4.22073126e-01 2.77703702e-01 -1.82404625e+00 1.49172747e+00
1.41471371e-01 -1.56843448e+00 2.22366080e-01 1.86702698e-01
4.71254438e-01 1.60108149e-01 3.91743332e-01 1.89435050e-01
2.70624876e-01 -1.02475584e+00 5.54770947e-01 4.11026657e-01
1.48437095e+00 -6.22497439e-01 2.87785202e-01 1.71280131e-01
-8.77850413e-01 -3.96834128e-02 -5.52854300e-01 2.53532648e-01
-3.86151403e-01 4.13148552e-01 -3.98675025e-01 6.36609972e-01
6.97380841e-01 7.30080545e-01 -5.32799542e-01 7.67533600e-01
1.59743220e-01 1.75655410e-01 -4.97770667e-01 2.36025915e-01
-3.12881768e-02 -3.86121087e-02 1.82473600e-01 1.06389809e+00
2.40081280e-01 -1.24407955e-01 -5.75015247e-02 7.76980758e-01
-6.71525896e-01 1.81910366e-01 -1.18320656e+00 -2.42809147e-01
7.30341136e-01 1.36791956e+00 6.08020183e-03 -9.17440578e-02
-5.11228085e-01 1.11382484e+00 5.61171174e-01 2.16633692e-01
-4.09931213e-01 -2.65570313e-01 8.51296961e-01 1.25141278e-01
3.42721999e-01 -1.77036583e-01 -7.47756958e-02 -1.36351883e+00
3.07040513e-01 -9.35175061e-01 6.87431276e-01 -7.61407614e-01
-1.88712668e+00 2.93159813e-01 -3.88526358e-02 -1.18704200e+00
-3.67361121e-02 -6.04695797e-01 -2.75313705e-01 8.74624789e-01
-1.64755940e+00 -1.70539641e+00 -2.94307061e-02 1.11059165e+00
3.76111157e-02 -2.67496109e-01 1.04362440e+00 5.91911077e-01
-2.95101017e-01 1.39516890e+00 5.05497575e-01 4.02260959e-01
1.25999200e+00 -1.01225317e+00 2.44694903e-01 5.04557848e-01
4.98865634e-01 7.33928382e-01 1.68218598e-01 -1.92321584e-01
-1.78383315e+00 -9.70307231e-01 8.60762954e-01 -9.35264409e-01
6.06318355e-01 -1.00325823e+00 -7.42440760e-01 9.04744804e-01
-3.01375296e-02 5.34510672e-01 9.10623252e-01 2.20784903e-01
-1.29587865e+00 -4.88936752e-01 -1.39767253e+00 5.98346591e-01
1.19865298e+00 -1.21211636e+00 -1.92852795e-01 4.36350375e-01
4.57882375e-01 -1.89446472e-02 -8.37174594e-01 -6.48084134e-02
9.58971798e-01 -8.47297072e-01 1.36639500e+00 -7.50434458e-01
4.86719221e-01 -7.59568810e-02 -5.98239124e-01 -7.49583662e-01
3.12540755e-02 -7.79687703e-01 2.24466234e-01 1.49656785e+00
2.77604043e-01 -5.55355489e-01 7.70122468e-01 1.24867976e+00
5.75080633e-01 -3.71643931e-01 -1.05129063e+00 -9.77131486e-01
4.09082294e-01 -3.09870601e-01 5.76857150e-01 1.17716694e+00
-1.78572074e-01 6.95901141e-02 -7.37952650e-01 3.51433121e-02
9.58764195e-01 1.84569523e-01 1.22632980e+00 -1.17422402e+00
-1.07812315e-01 -2.59598583e-01 -3.64149213e-01 -8.03083301e-01
3.73300612e-01 -1.17145658e+00 -3.07888776e-01 -1.18187559e+00
3.60740006e-01 -8.13667119e-01 -4.55444634e-01 9.72641110e-01
1.99812979e-01 6.42557263e-01 5.72086394e-01 4.19292033e-01
-6.86304927e-01 5.02239227e-01 9.12863314e-01 -2.35720083e-01
3.33751798e-01 -2.86104530e-01 -7.01136649e-01 2.00807214e-01
4.70031112e-01 -3.95237803e-01 -7.41656899e-01 -4.80696499e-01
2.05483347e-01 -2.49257132e-01 7.64519989e-01 -5.90103090e-01
2.20038727e-01 -3.71017009e-02 2.91611761e-01 -3.38902116e-01
5.53726614e-01 -1.10650337e+00 1.60161316e-01 -1.15331840e-02
-8.78699005e-01 -3.01989615e-01 1.59022629e-01 6.20562792e-01
4.18966934e-02 6.66302964e-02 6.10933840e-01 2.28247866e-02
-4.24930185e-01 7.58625448e-01 3.71251225e-01 9.63985547e-02
7.29578972e-01 9.56090614e-02 -5.72312534e-01 -5.74423015e-01
-3.49010468e-01 1.58645678e-02 7.13638008e-01 4.94183540e-01
4.85792011e-01 -1.55569172e+00 -6.74477398e-01 2.91236967e-01
6.70351863e-01 -1.93265602e-01 3.31147939e-01 5.36042392e-01
-2.18385994e-01 2.74637401e-01 -2.52706707e-01 -3.70737374e-01
-1.39849448e+00 9.63625312e-01 1.71356741e-02 8.48521516e-02
-5.40893257e-01 9.89770472e-01 3.86282563e-01 -7.51021028e-01
5.29391527e-01 2.14326337e-01 7.04826474e-01 3.37256603e-02
6.85267270e-01 1.79751396e-01 3.64640844e-03 -1.72942922e-01
-4.47320223e-01 3.54414701e-01 -3.46689343e-01 -2.81344533e-01
1.37119055e+00 -1.17512606e-01 -8.81841555e-02 3.70926172e-01
1.81552839e+00 -3.98772359e-02 -1.32586420e+00 -5.53741634e-01
-1.79667801e-01 -1.03762734e+00 -4.14143968e-03 -9.88988876e-01
-1.19310260e+00 1.15534270e+00 6.77176952e-01 -1.06749028e-01
9.84174252e-01 3.31188813e-02 7.62686014e-01 6.75714791e-01
5.32077372e-01 -8.95000279e-01 3.80277634e-02 -5.03153801e-02
1.20218801e+00 -1.43838286e+00 1.06939971e-01 6.45325892e-03
-5.51174104e-01 8.64749193e-01 2.48778984e-01 -1.05962761e-01
7.79865861e-01 7.58167505e-02 1.81308776e-01 8.54380727e-02
-7.57460773e-01 2.15905726e-01 3.13928425e-01 6.49169207e-01
1.17657654e-01 -2.79514324e-02 1.62779406e-01 3.27893943e-01
1.13601916e-01 3.32664222e-01 1.18901707e-01 1.08715022e+00
3.05658966e-01 -1.48331118e+00 -9.09404084e-02 3.15655023e-01
-2.00221851e-01 -1.20171182e-01 -7.54028857e-01 8.74792516e-01
-2.40204990e-01 4.89796281e-01 8.95772576e-02 -2.24461421e-01
1.41052067e-01 1.96009517e-01 6.53161228e-01 -4.29575630e-02
-2.39825100e-01 -2.84056962e-01 5.41757159e-02 -7.79345393e-01
-2.80456811e-01 -6.61143541e-01 -4.84370023e-01 -8.18802774e-01
-2.03305408e-02 -2.96291292e-01 8.38336289e-01 5.56765258e-01
7.48951435e-01 -5.90104401e-01 7.78269529e-01 -9.12323892e-01
-7.68918872e-01 -4.15759176e-01 -6.45565271e-01 6.88495100e-01
6.64862394e-01 -4.36575502e-01 -5.12557626e-01 7.70280976e-03] | [11.55176830291748, 0.6575353145599365] |
4492ead1-b119-4703-b055-b3fe495aaa36 | a-large-scale-english-multi-label-twitter | null | null | https://aclanthology.org/2021.woah-1.16 | https://aclanthology.org/2021.woah-1.16.pdf | A Large-Scale English Multi-Label Twitter Dataset for Cyberbullying and Online Abuse Detection | In this paper, we introduce a new English Twitter-based dataset for cyberbullying detection and online abuse. Comprising 62,587 tweets, this dataset was sourced from Twitter using specific query terms designed to retrieve tweets with high probabilities of various forms of bullying and offensive content, including insult, trolling, profanity, sarcasm, threat, porn and exclusion. We recruited a pool of 17 annotators to perform fine-grained annotation on the dataset with each tweet annotated by three annotators. All our annotators are high school educated and frequent users of social media. Inter-rater agreement for the dataset as measured by Krippendorff’s Alpha is 0.67. Analysis performed on the dataset confirmed common cyberbullying themes reported by other studies and revealed interesting relationships between the classes. The dataset was used to train a number of transformer-based deep learning models returning impressive results. | ['Yulan He', 'Jo Lumsden', 'Semiu Salawu'] | null | null | null | null | acl-woah-2021-8 | ['abuse-detection'] | ['natural-language-processing'] | [-3.74203980e-01 4.18531537e-01 -2.09928498e-01 -3.08055609e-01
-8.38384926e-01 -5.70513010e-01 4.35327888e-01 9.00051177e-01
-6.31285012e-01 7.00905621e-01 5.77587128e-01 -3.65933366e-02
1.11563072e-01 -6.31680012e-01 -3.05655032e-01 -3.29499394e-01
-6.58674166e-02 2.14022934e-01 7.86821842e-02 -5.07494628e-01
6.01880550e-01 1.73234090e-01 -1.16832113e+00 4.38578188e-01
5.74633777e-01 7.00865507e-01 -7.46324301e-01 5.80014110e-01
6.42684996e-02 1.04304659e+00 -9.31578457e-01 -1.20259178e+00
-3.10569406e-01 -3.90071385e-02 -1.18934035e+00 -3.78021479e-01
6.48900390e-01 -3.27351928e-01 -3.41507465e-01 9.35829043e-01
5.52536905e-01 1.13317549e-01 3.95441115e-01 -1.26256847e+00
-3.14021587e-01 7.74268806e-01 -6.35397434e-01 6.03777528e-01
8.31242263e-01 -6.25882894e-02 9.97414947e-01 -8.34476352e-01
9.89131749e-01 1.32591116e+00 1.07277036e+00 3.53524745e-01
-1.18021905e+00 -1.19606090e+00 -6.27293587e-01 1.73282683e-01
-1.45318449e+00 -3.53076011e-01 7.76374161e-01 -7.03122437e-01
7.77578950e-01 3.21767032e-01 7.46796906e-01 1.73216140e+00
2.39767060e-02 3.76086801e-01 9.56231236e-01 -9.43275765e-02
-8.61240923e-02 5.22279978e-01 3.89727533e-01 4.74182218e-01
2.23835573e-01 -5.40011346e-01 -7.52966940e-01 -9.02013958e-01
-1.29943982e-01 -4.54035580e-01 3.20305914e-01 5.71141958e-01
-5.67396343e-01 1.45483124e+00 4.03156728e-01 4.57781762e-01
-2.92397738e-01 -1.10182814e-01 1.18503010e+00 1.42160684e-01
1.00474250e+00 6.08506262e-01 1.41978934e-02 -5.00590622e-01
-7.77611077e-01 3.62270325e-01 8.87857676e-01 2.04225793e-01
5.19353271e-01 -3.01939249e-01 9.41507146e-02 1.26872504e+00
1.05615966e-01 2.14160502e-01 3.39206129e-01 -8.00055087e-01
6.63652182e-01 3.88769090e-01 -2.41965130e-01 -1.90788674e+00
-6.74427032e-01 9.98800024e-02 -3.86504859e-01 -3.40174556e-01
4.12910819e-01 -4.84695107e-01 -3.82291317e-01 1.49690962e+00
6.23078585e-01 1.94979116e-01 -6.65572941e-01 7.65193760e-01
9.15027976e-01 5.00951946e-01 7.81992674e-01 -1.26167685e-01
1.34863698e+00 -2.54534423e-01 -8.84109735e-01 9.16709974e-02
8.87118042e-01 -8.62168372e-01 8.57654989e-01 5.60899436e-01
-8.75315547e-01 -8.67007077e-02 -6.28138959e-01 -2.15461150e-01
-7.81729043e-01 -7.53149331e-01 4.78324533e-01 1.31296778e+00
-7.86073685e-01 4.42940503e-01 -3.36541593e-01 -6.85717821e-01
6.53368354e-01 7.31022656e-02 -6.28299892e-01 3.18242252e-01
-1.41252542e+00 9.97405827e-01 4.57982302e-01 -2.62637228e-01
-6.33001328e-01 -7.28070557e-01 -7.73026943e-01 -5.68553209e-01
2.03243479e-01 2.99434751e-01 9.81917918e-01 -7.08538771e-01
-8.56658459e-01 1.80601335e+00 1.09611914e-01 -3.72873604e-01
3.26013774e-01 -5.67358136e-01 -6.21085346e-01 3.93703789e-01
6.62236214e-01 1.37035251e-01 9.30737257e-01 -9.01874781e-01
-3.47255856e-01 -4.33393985e-01 1.25568449e-01 -2.97625780e-01
-8.29846442e-01 9.90175903e-01 3.93544912e-01 -4.88514245e-01
-1.83706596e-01 -6.71559691e-01 3.04248840e-01 -6.36394322e-01
-1.11879337e+00 -5.41861117e-01 9.79648352e-01 -9.56366599e-01
1.63712907e+00 -2.16286874e+00 -4.69041288e-01 3.19390118e-01
6.55610263e-01 3.68758917e-01 3.30548674e-01 1.09586525e+00
-1.42568439e-01 6.52835369e-01 1.28338933e-01 -3.43674898e-01
2.38916174e-01 1.29834581e-02 -3.86539072e-01 8.88352334e-01
3.44414823e-02 4.98610497e-01 -1.09455216e+00 -7.72919178e-01
3.51153500e-03 2.15982243e-01 -5.86067140e-01 2.45083556e-01
4.62026715e-01 2.70726353e-01 -4.73805130e-01 5.16054511e-01
6.61214948e-01 4.57533658e-01 -1.84409857e-01 1.35627147e-02
-3.52104545e-01 5.34287751e-01 -4.59547997e-01 8.69902790e-01
-2.85977632e-01 1.09378695e+00 4.97758448e-01 -5.87060511e-01
8.59946907e-01 3.37595344e-01 3.96675229e-01 -4.60247844e-01
6.63984478e-01 1.68497697e-01 -1.86510876e-01 -1.06472790e+00
7.09093511e-01 -2.44221970e-01 -6.93086624e-01 8.01894367e-01
1.81273654e-01 -1.07010975e-02 9.67544839e-02 5.22818387e-01
9.07200694e-01 -5.53214490e-01 1.51709229e-01 -2.07232744e-01
2.57875592e-01 7.22114071e-02 1.57446951e-01 7.03240812e-01
-4.81961787e-01 2.24863097e-01 8.74491274e-01 -7.08127022e-01
-1.13606179e+00 -7.65134156e-01 -4.72066283e-01 1.77308786e+00
-2.32646987e-01 -6.45311832e-01 -9.26792622e-01 -6.47834659e-01
-1.53392151e-01 8.78357887e-01 -8.07734549e-01 -3.23475868e-01
-4.80614722e-01 -7.30642676e-01 1.25131023e+00 -2.80034810e-01
4.81572747e-01 -1.06141508e+00 -5.05046248e-01 1.93952292e-01
-3.97715420e-01 -1.31260359e+00 -7.19247609e-02 -3.55847031e-01
8.14429000e-02 -1.37574184e+00 -3.44436541e-02 -3.71463120e-01
1.04074955e-01 -1.48709863e-01 8.80407393e-01 4.28666353e-01
-3.79371285e-01 1.49184138e-01 -7.84773052e-01 -5.42166591e-01
-5.69518864e-01 3.95732552e-01 2.37061560e-01 1.26800254e-01
7.56209314e-01 -5.23909688e-01 -2.06398323e-01 -7.78483525e-02
-9.69419360e-01 -5.85556507e-01 -5.54674506e-01 4.29591715e-01
-4.87377465e-01 -1.10935748e-01 4.69582915e-01 -1.09652340e+00
1.04128385e+00 -1.27330852e+00 2.93311328e-01 -6.54790521e-01
5.13527617e-02 -1.05365050e+00 2.67577291e-01 -5.96039832e-01
-6.66321397e-01 -5.53053319e-01 -6.58680439e-01 -4.78369035e-02
-6.55161500e-01 4.60071951e-01 4.14166987e-01 1.24784812e-01
7.92312860e-01 -3.03143859e-01 -3.07445973e-01 -4.46234375e-01
-5.02690487e-02 9.70811427e-01 2.20004380e-01 -3.99183691e-01
8.85928035e-01 1.82243064e-01 -3.61606777e-01 -1.50045502e+00
-1.35067594e+00 -6.32474065e-01 -6.31585300e-01 -5.68232179e-01
1.22167611e+00 -5.85793674e-01 -1.08224285e+00 7.88248420e-01
-1.18736136e+00 -1.05797194e-01 3.96240145e-01 1.44336566e-01
7.63810575e-02 1.87914431e-01 -8.71000051e-01 -1.05027974e+00
-4.85436738e-01 -4.69514310e-01 6.95514619e-01 9.51010268e-03
-9.92268562e-01 -1.13409340e+00 3.70912015e-01 6.83792531e-01
3.68065566e-01 1.02235782e+00 7.35094190e-01 -1.58089745e+00
7.03000367e-01 -6.60765886e-01 -2.30472490e-01 -5.46579761e-03
-2.82815814e-01 2.72007465e-01 -1.02299559e+00 6.18298724e-02
-1.39519572e-01 -1.17335260e+00 4.59636033e-01 -3.09040755e-01
1.01356399e+00 -9.32087541e-01 -1.60561115e-01 3.99267673e-03
1.07866061e+00 -3.08538377e-01 4.50529963e-01 7.30144083e-01
8.75927031e-01 7.64494121e-01 2.41084680e-01 7.14280128e-01
3.59182537e-01 4.98456299e-01 5.50185978e-01 2.16711670e-01
5.28507710e-01 -5.18742442e-01 2.24097505e-01 2.89280385e-01
1.24714680e-01 9.48987603e-02 -1.42028892e+00 6.62803292e-01
-1.31948853e+00 -1.34523821e+00 -6.61106646e-01 1.60505235e+00
7.62103975e-01 1.91561997e-01 6.43100679e-01 3.12069446e-01
8.97915363e-01 5.12274921e-01 -2.10501328e-02 -1.14798319e+00
2.78715193e-01 4.40957755e-01 1.23386696e-01 6.03598773e-01
-1.33938074e+00 1.11219847e+00 6.58198261e+00 1.01451993e+00
-9.46157098e-01 5.17539382e-01 6.69194818e-01 -2.48265684e-01
1.59241349e-01 -5.23505449e-01 -6.42841816e-01 6.69049144e-01
1.38846719e+00 -9.59147587e-02 8.78536329e-02 7.94394314e-01
3.17912430e-01 -2.51236975e-01 -4.26721305e-01 6.89864218e-01
4.12480503e-01 -9.33095574e-01 -5.52051246e-01 -2.43911892e-02
3.82294834e-01 2.03226522e-01 -1.46278739e-02 5.81556559e-01
1.90904185e-01 -1.08278167e+00 7.43021846e-01 7.83095211e-02
4.09430206e-01 -9.14665520e-01 7.63770759e-01 3.86384726e-01
-2.97893971e-01 -2.16577202e-01 -4.32877131e-02 -3.05598140e-01
2.35216707e-01 5.19936740e-01 -8.53851497e-01 -1.16750844e-01
9.65040863e-01 6.65214896e-01 -6.38774395e-01 6.33184969e-01
-2.94827759e-01 1.10152435e+00 -5.85324764e-01 -2.31888741e-01
5.29277444e-01 1.07751988e-01 8.02907586e-01 1.76844513e+00
-2.44784623e-01 1.58884063e-01 -1.50250793e-02 6.09016597e-01
-4.43904176e-02 1.94599494e-01 -8.51435959e-01 -1.99241638e-01
6.42002583e-01 1.62258005e+00 -5.81496954e-01 -5.82949966e-02
5.19332774e-02 5.41358829e-01 6.94319963e-01 2.53643133e-02
-8.67559731e-01 -5.53264618e-01 5.81994355e-01 6.19233549e-01
-3.16186607e-01 -1.64546400e-01 -1.37412712e-01 -8.17160428e-01
-5.81170976e-01 -6.58444881e-01 6.24627054e-01 -5.00612557e-01
-1.56814170e+00 5.24929762e-01 3.24172527e-01 -5.17189801e-01
-6.97295293e-02 -2.91159570e-01 -8.96248043e-01 4.36045349e-01
-7.44490802e-01 -1.17284596e+00 -9.42782536e-02 5.74085593e-01
2.15575933e-01 1.10036470e-01 7.20414221e-01 6.30448043e-01
-9.95773077e-01 6.70495331e-01 -6.37167394e-01 7.57216692e-01
6.29101992e-01 -7.30257034e-01 -1.77289799e-01 2.15034470e-01
-3.44840318e-01 5.78977823e-01 9.61195886e-01 -6.31567657e-01
-7.96481133e-01 -9.47843313e-01 1.28018761e+00 -6.60159826e-01
1.55623293e+00 -5.03777444e-01 -9.77198482e-01 6.93837106e-01
4.44685847e-01 -4.16571826e-01 1.47381079e+00 4.07154471e-01
-7.18412399e-01 3.92552078e-01 -1.44128227e+00 1.95952132e-01
7.44985342e-01 -7.14359462e-01 -7.51418591e-01 8.22783649e-01
4.49510753e-01 -3.88740540e-01 -9.68924582e-01 -3.44097048e-01
6.18145406e-01 -1.06965065e+00 6.32391632e-01 -1.16172338e+00
9.18801963e-01 8.09481025e-01 1.56944960e-01 -1.02322805e+00
-1.80636242e-01 -6.56772792e-01 2.97056913e-01 1.64742935e+00
3.16116273e-01 -5.87805510e-01 2.97748774e-01 1.04068804e+00
-1.40225768e-01 -5.05956531e-01 -1.34555960e+00 -1.39247283e-01
4.35392380e-01 -8.04279566e-01 1.12043232e-01 1.68581057e+00
6.81939065e-01 3.65887880e-01 -4.48385268e-01 -1.26109332e-01
5.41290164e-01 -5.71389854e-01 8.67987514e-01 -1.18806159e+00
4.64367926e-01 -5.81570804e-01 -4.08601850e-01 1.47366047e-01
5.66499829e-01 -7.06749737e-01 -4.60107327e-01 -9.84265387e-01
3.46983433e-01 -2.32749104e-01 1.95758596e-01 6.06154382e-01
6.73184842e-02 1.11236894e+00 5.95754236e-02 -8.64186510e-02
-8.56073081e-01 1.65936351e-01 6.52915597e-01 5.77505864e-02
-9.24097896e-02 -3.19268644e-01 -6.08770311e-01 1.04215801e+00
9.66810226e-01 -8.75205457e-01 2.82588273e-01 -5.32290675e-02
8.99325848e-01 -4.00440246e-01 5.29639304e-01 -5.95370591e-01
-3.22064817e-01 -1.23285852e-01 2.91214697e-02 -3.09532225e-01
4.88953501e-01 -5.74748218e-01 -2.95231372e-01 3.10823083e-01
-5.43068469e-01 -1.65483341e-01 2.21521839e-01 1.58904597e-01
-9.72592011e-02 -3.53782922e-01 8.85340035e-01 8.85569900e-02
-9.14806053e-02 1.56872258e-01 -9.17803288e-01 5.00042975e-01
1.30240560e+00 -1.42680004e-01 -4.59366053e-01 -7.92773843e-01
-9.47954118e-01 -9.51491222e-02 1.47000859e-02 4.90282059e-01
4.11271513e-01 -1.17327833e+00 -9.48034167e-01 -1.74125433e-01
1.66304652e-02 -7.26310551e-01 1.23170890e-01 9.99916971e-01
-4.01822507e-01 5.07615060e-02 -2.93724686e-01 -9.48394164e-02
-1.17428458e+00 2.08522812e-01 -2.92334836e-02 7.51795322e-02
-3.68088484e-01 8.43201160e-01 -8.10082734e-01 -4.15264517e-01
-1.81222126e-01 5.89854598e-01 -4.83791500e-01 7.81938970e-01
5.92839599e-01 7.88654745e-01 -2.43518963e-01 -1.55006087e+00
-3.93010497e-01 1.73995107e-01 -8.42917114e-02 2.72928681e-02
1.45080888e+00 -1.23909429e-01 -5.82617581e-01 5.84385276e-01
1.76701713e+00 2.76279628e-01 -8.45219642e-02 1.42365575e-01
2.08872035e-01 -7.10487962e-01 -5.71299754e-02 -3.62849861e-01
-6.41808093e-01 7.25127280e-01 -9.10388678e-02 1.21267056e+00
4.80615556e-01 3.03326279e-01 1.15031326e+00 5.09560369e-02
1.69646621e-01 -1.22616482e+00 3.16895545e-01 7.96647787e-01
9.76396382e-01 -1.22562945e+00 5.93418740e-02 -4.57968980e-01
-6.87090814e-01 1.00134480e+00 8.01056564e-01 -5.00377297e-01
9.23322439e-01 -9.32880640e-02 6.68908060e-02 -7.63880074e-01
-2.96258628e-01 1.45053566e-01 6.02489188e-02 4.85516280e-01
5.68160713e-01 -2.05237810e-02 -8.17565262e-01 5.09933233e-01
-8.95323515e-01 -5.71944654e-01 6.63935900e-01 4.30341631e-01
-4.98923898e-01 -4.71814632e-01 -6.40353441e-01 3.95490080e-01
-1.32597125e+00 2.45224342e-01 -8.53431582e-01 1.07011187e+00
2.58668393e-01 1.24571598e+00 2.26250932e-01 -6.02998316e-01
1.16817467e-01 1.28542215e-01 -2.37088263e-01 -6.33000553e-01
-1.29829657e+00 -3.40667844e-01 8.99830520e-01 -6.11409307e-01
-1.67913154e-01 -6.88710630e-01 -8.92320752e-01 -7.63082385e-01
-8.18314329e-02 2.62494564e-01 5.84137082e-01 1.02971864e+00
1.02159008e-01 -3.05483341e-01 6.70018077e-01 -8.56816173e-01
8.50706249e-02 -1.44918716e+00 -5.37040532e-01 9.72353339e-01
5.17009795e-01 -6.28508329e-01 -4.74285692e-01 -2.99758047e-01] | [8.707881927490234, 10.501469612121582] |
5c7ec035-f896-42ec-842b-b7da034e07a1 | multimodal-material-classification-for-robots | 2004.01160 | null | https://arxiv.org/abs/2004.01160v2 | https://arxiv.org/pdf/2004.01160v2.pdf | Multimodal Material Classification for Robots using Spectroscopy and High Resolution Texture Imaging | Material recognition can help inform robots about how to properly interact with and manipulate real-world objects. In this paper, we present a multimodal sensing technique, leveraging near-infrared spectroscopy and close-range high resolution texture imaging, that enables robots to estimate the materials of household objects. We release a dataset of high resolution texture images and spectral measurements collected from a mobile manipulator that interacted with 144 household objects. We then present a neural network architecture that learns a compact multimodal representation of spectral measurements and texture images. When generalizing material classification to new objects, we show that this multimodal representation enables a robot to recognize materials with greater performance as compared to prior state-of-the-art approaches. Finally, we present how a robot can combine this high resolution local sensing with images from the robot's head-mounted camera to achieve accurate material classification over a scene of objects on a table. | ['Sonia Chernova', 'Eliot Xing', 'Zackory Erickson', 'Charles C. Kemp', 'Bharat Srirangam'] | 2020-04-02 | null | null | null | null | ['material-classification', 'material-recognition'] | ['computer-vision', 'computer-vision'] | [ 8.92921388e-01 -1.96034104e-01 4.40345071e-02 -4.67540175e-01
-7.42653131e-01 -4.05871302e-01 2.57158250e-01 -3.98104578e-01
-2.61566401e-01 4.59714055e-01 -3.06203514e-01 1.90152720e-01
-4.27557945e-01 -8.17764044e-01 -9.33930576e-01 -8.03111315e-01
1.03325985e-01 7.98663557e-01 -1.20879142e-02 -1.45668313e-01
2.90534496e-01 8.19754422e-01 -1.84905565e+00 5.68563044e-01
2.07599640e-01 1.37245774e+00 9.48644638e-01 4.61384714e-01
2.58245409e-01 6.16243005e-01 -9.99333337e-02 3.73093992e-01
4.58644688e-01 2.62025267e-01 -8.62534404e-01 4.42127794e-01
6.99376762e-01 -7.35056162e-01 -4.87540334e-01 9.45258319e-01
-1.20149634e-03 2.82035977e-01 1.10698056e+00 -8.01908672e-01
-1.00356245e+00 5.15590727e-01 -5.63862085e-01 -3.02487195e-01
5.12494504e-01 4.90304120e-02 5.49379051e-01 -8.33154738e-01
5.68431020e-01 1.35342419e+00 6.98287725e-01 4.82483417e-01
-1.14685977e+00 -2.35359401e-01 1.51597084e-02 3.04064751e-01
-1.21353292e+00 -6.03851378e-01 9.01152730e-01 -6.22186244e-01
1.16426945e+00 -3.84223089e-02 3.71248990e-01 1.17049849e+00
4.68457639e-01 3.05671126e-01 9.06090975e-01 -4.61607009e-01
-2.87222918e-02 -2.20017418e-01 -2.50749588e-01 9.89663124e-01
4.48907495e-01 6.83815703e-02 -6.83767080e-01 6.16781935e-02
9.41980779e-01 3.06333095e-01 -1.44959083e-02 -7.79034674e-01
-1.71096087e+00 1.62187129e-01 6.66645288e-01 1.37914181e-01
-7.75106907e-01 4.01671022e-01 -1.59386992e-01 1.79793522e-01
1.87782809e-01 7.38089144e-01 -4.08567697e-01 2.97105722e-02
3.92988808e-02 -1.40133336e-01 6.56053126e-01 1.06673896e+00
1.00368369e+00 -3.89703102e-02 4.81743246e-01 1.00215518e+00
5.53463221e-01 1.17786300e+00 6.33985102e-02 -1.49814498e+00
3.94849718e-01 1.91861674e-01 4.50157583e-01 -7.48116136e-01
-6.72280312e-01 2.51994282e-01 -4.17784244e-01 4.17194396e-01
2.91107833e-01 -5.80023900e-02 -7.37076223e-01 1.32623684e+00
1.31253794e-01 -5.29407799e-01 3.61019939e-01 1.14217377e+00
5.19652724e-01 4.35391039e-01 -3.22286606e-01 2.41340429e-01
1.31212270e+00 -7.92526245e-01 -4.68247414e-01 -5.71294427e-01
2.76701480e-01 -6.65854990e-01 7.25384831e-01 5.12702823e-01
-7.74371207e-01 -4.70576525e-01 -1.27376175e+00 -1.84728727e-01
-6.37645543e-01 3.79701585e-01 9.34065044e-01 2.69011319e-01
-7.53981352e-01 7.80768812e-01 -1.13861048e+00 -7.30678082e-01
2.42695495e-01 6.75281763e-01 -5.74465275e-01 -4.83157754e-01
-3.49906057e-01 1.16169834e+00 4.37210262e-01 3.45459074e-01
-6.94458723e-01 -2.57538408e-01 -9.68700886e-01 -4.76068139e-01
5.35163939e-01 -7.39914298e-01 1.36162150e+00 -4.90877897e-01
-1.94855440e+00 8.56732666e-01 1.83362767e-01 1.95961043e-01
-1.51929960e-01 -2.77646840e-01 -1.49459630e-01 5.37553787e-01
-4.40016463e-02 7.49155045e-01 8.92761409e-01 -1.50119221e+00
-3.72368038e-01 -7.74273276e-01 -6.96199713e-03 4.63958144e-01
-1.71880513e-01 -3.15402806e-01 1.18236691e-01 1.39757350e-01
9.06571448e-01 -1.07409644e+00 7.74633959e-02 2.82773852e-01
-5.09788096e-01 7.39292800e-02 1.01677644e+00 -3.62973750e-01
-3.50307077e-01 -1.97378802e+00 3.55224878e-01 2.60104448e-01
-2.25092862e-02 -6.51454389e-01 -3.70446414e-01 3.17923158e-01
1.62591875e-01 -5.31073928e-01 -1.17164198e-02 -8.44368637e-02
9.26425904e-02 1.24411874e-01 3.39069813e-02 7.99277544e-01
2.13320106e-01 7.37725258e-01 -5.97818553e-01 1.87763348e-01
4.21949714e-01 4.65249360e-01 -1.40558153e-01 9.47423875e-02
-2.80953109e-01 5.08506238e-01 -5.04647136e-01 1.15255749e+00
6.47545576e-01 -1.90276027e-01 2.97499776e-01 -7.15827107e-01
-1.64706588e-01 6.53097592e-03 -8.31854641e-01 2.11465740e+00
-3.54202062e-01 4.80781436e-01 7.10717082e-01 -1.06503534e+00
1.01619184e+00 -1.48578525e-01 7.88428307e-01 -7.78713942e-01
3.41621190e-01 2.06600189e-01 -4.31750596e-01 -9.57939148e-01
6.86152518e-01 -1.60194993e-01 -2.08032429e-02 5.00949502e-01
-1.06457569e-01 -5.74090481e-01 -2.02671185e-01 -4.73906964e-01
1.07569432e+00 7.05322206e-01 -3.03048015e-01 -2.33624548e-01
8.07665512e-02 2.74713665e-01 -1.32162288e-01 7.49015033e-01
2.52724290e-01 3.61868948e-01 -5.21800339e-01 -5.35673738e-01
-1.14870322e+00 -1.39929104e+00 2.40383460e-03 1.55729592e+00
6.00672841e-01 4.97244596e-01 -3.33522707e-01 2.37698331e-01
3.62437487e-01 1.16927981e-01 -4.64769900e-01 -1.85729101e-01
-4.58588213e-01 -4.68093961e-01 -1.23303615e-01 5.95046520e-01
5.40348828e-01 -1.05681467e+00 -8.92065406e-01 2.45816051e-03
-2.57881880e-01 -1.37378073e+00 1.98229179e-01 5.76173007e-01
-7.78121710e-01 -1.11878932e+00 -5.08249640e-01 -1.11169469e+00
6.94447339e-01 8.03690553e-01 6.13912046e-01 -4.90927279e-01
-5.63647866e-01 1.01874542e+00 -2.96164036e-01 -3.99239004e-01
-2.09579602e-01 1.29137650e-01 4.54589754e-01 -1.06959008e-01
2.70215839e-01 -5.26977897e-01 -2.07383618e-01 2.74115860e-01
-4.50488478e-01 1.51537642e-01 1.06776619e+00 5.51417828e-01
4.81703669e-01 1.81886092e-01 4.07114252e-02 -4.44936335e-01
2.18314633e-01 -3.21346492e-01 -5.98327041e-01 2.70701647e-01
-1.30537212e-01 2.35652775e-01 1.45831928e-01 -7.05934346e-01
-1.21569896e+00 3.89315307e-01 4.00268763e-01 -2.84866154e-01
-4.83411163e-01 5.13659954e-01 -5.71424626e-02 -4.31513578e-01
8.49358141e-01 -1.28261775e-01 2.73602366e-01 -4.33702618e-01
2.63253957e-01 9.54859078e-01 1.03744924e+00 -1.12697768e+00
7.53477812e-01 7.30948210e-01 3.05270642e-01 -1.17201829e+00
-5.63065588e-01 -4.05169159e-01 -1.07024288e+00 -2.62483954e-01
9.00125742e-01 -1.01175642e+00 -1.21881473e+00 6.81346953e-01
-8.75376225e-01 -7.53260076e-01 1.15514785e-01 8.70584011e-01
-1.09017253e+00 6.69660419e-02 -7.62291372e-01 -9.86364484e-01
6.46015629e-02 -1.05943644e+00 1.71882486e+00 -1.98034500e-03
-3.33927497e-02 -5.24638474e-01 -4.38193411e-01 7.29062259e-01
4.47724849e-01 2.50117451e-01 8.17235947e-01 2.66639441e-01
-9.63641047e-01 3.15250037e-03 -3.57667357e-01 -2.90078342e-01
5.65588772e-01 -2.39125535e-01 -1.09193981e+00 -1.91515312e-01
-4.66199629e-02 -6.25430763e-01 8.10997605e-01 2.92228490e-01
9.81995761e-01 -9.36404243e-02 -4.06797856e-01 6.00774467e-01
1.18580925e+00 2.64395028e-01 3.31819683e-01 7.08165586e-01
9.75514472e-01 9.62808669e-01 7.73612559e-01 2.35672683e-01
2.68238246e-01 6.65991724e-01 6.58652186e-01 2.55676866e-01
1.27711013e-01 2.65690535e-01 4.78017986e-01 3.66811007e-01
-2.46972978e-01 1.27982691e-01 -7.89959192e-01 7.49417618e-02
-1.70127749e+00 -8.25829864e-01 1.61932796e-01 1.74513471e+00
3.31248432e-01 -2.39850640e-01 -1.52562305e-01 -2.13999867e-01
6.83560848e-01 -1.60033241e-01 -1.10527229e+00 -3.88912531e-03
-9.10948813e-02 -2.41242096e-01 7.38213360e-01 3.39945704e-01
-1.11463881e+00 7.58823514e-01 7.15520191e+00 5.20203933e-02
-1.07207000e+00 -2.72540241e-01 2.66392902e-02 1.69547610e-02
-2.32116692e-02 -6.89320326e-01 -2.16752470e-01 -2.64206320e-01
8.57371151e-01 5.64659119e-01 1.29775739e+00 8.53970826e-01
-2.12708190e-01 -5.13799429e-01 -1.45516384e+00 1.34053302e+00
2.51566887e-01 -1.05917192e+00 -3.18654090e-01 1.74678981e-01
4.86969471e-01 4.41298723e-01 3.60709727e-01 -1.86179146e-01
2.68410176e-01 -8.10156167e-01 1.12826669e+00 9.14199293e-01
8.12432110e-01 -1.89275846e-01 3.64512056e-01 3.27435404e-01
-1.14936972e+00 -5.68003595e-01 -6.49210811e-01 -2.41738930e-01
-1.53965682e-01 3.22212517e-01 -9.38561618e-01 4.43984985e-01
9.28972900e-01 6.98878467e-01 -2.82004595e-01 3.84641588e-01
2.77007341e-01 -3.71556908e-01 -5.17744541e-01 -1.15135312e-01
-2.00725004e-01 -4.29529160e-01 5.78280799e-02 6.55181348e-01
5.84210038e-01 1.87386453e-01 4.64890480e-01 1.20072496e+00
-2.49215122e-03 -4.79871184e-01 -1.07223320e+00 -2.31282994e-01
2.98923999e-01 1.28079832e+00 -7.27295756e-01 -2.03705862e-01
-3.38922203e-01 9.48575199e-01 2.86715299e-01 5.85960627e-01
-1.10265568e-01 -1.24089532e-01 5.01346111e-01 -2.25333869e-01
1.89008355e-01 -9.45359230e-01 -2.96084672e-01 -9.37783241e-01
2.42586881e-01 -4.63370115e-01 -2.84507185e-01 -1.70840728e+00
-1.58421326e+00 1.64962366e-01 2.92988271e-02 -9.58697736e-01
-1.08555384e-01 -1.38740420e+00 2.30367720e-01 7.09382236e-01
-1.16988266e+00 -1.42192292e+00 -7.48423398e-01 3.67650867e-01
3.31900269e-01 -2.09833160e-01 9.99109507e-01 -2.97847599e-01
-1.78956971e-01 -3.55549425e-01 3.31086993e-01 -2.80707665e-02
7.23945022e-01 -9.26047921e-01 -2.01037154e-01 2.43052304e-01
-1.78463638e-01 4.93645877e-01 6.03173971e-01 -7.82873690e-01
-2.44020414e+00 -8.03580999e-01 -1.97436854e-01 -4.99320030e-01
6.32795393e-01 -3.61816585e-01 -4.79750127e-01 7.92325318e-01
-2.90165134e-02 -1.38252065e-01 4.23634142e-01 2.09440246e-01
-4.23671305e-01 -2.58913010e-01 -1.37493777e+00 3.84843409e-01
1.28022373e+00 -1.09251678e+00 -5.60403705e-01 6.45188689e-01
4.40247267e-01 -6.42397344e-01 -1.17408609e+00 4.89227682e-01
1.16316307e+00 -6.13256454e-01 1.12784433e+00 -9.40751731e-02
2.49448597e-01 -2.25494429e-01 -7.88262427e-01 -1.01105154e+00
-6.04360282e-01 -1.31770492e-01 1.83815747e-01 5.11910677e-01
2.45313495e-01 -7.53302395e-01 8.62509251e-01 7.57465839e-01
-4.47576106e-01 -6.09536283e-02 -5.80247343e-01 -9.63445961e-01
-2.43895411e-01 -2.74639845e-01 5.25085390e-01 7.43985057e-01
1.45293236e-01 3.44549894e-01 -1.44050449e-01 6.64443672e-01
8.98639441e-01 3.15949887e-01 5.77717125e-01 -1.57590139e+00
-3.52141820e-02 -1.98221147e-01 -3.98478657e-01 -8.83874178e-01
4.94114608e-01 -6.42575145e-01 6.88937902e-01 -1.62518203e+00
4.88442361e-01 -4.68340933e-01 -4.36766632e-02 6.14244163e-01
5.97974002e-01 4.45300490e-01 -5.83477020e-02 3.73572886e-01
-6.36180639e-01 4.44792837e-01 1.13946545e+00 -7.50888288e-01
-9.75758806e-02 -3.98598373e-01 -4.65246737e-01 5.82812607e-01
7.67143071e-01 -3.58805545e-02 -9.98787433e-02 -9.69888747e-01
2.73327976e-01 3.90858278e-02 3.98155570e-01 -1.39071345e+00
3.55920434e-01 -4.36092854e-01 7.27944970e-01 -6.28346622e-01
1.02596176e+00 -1.30384564e+00 3.29239279e-01 3.36863488e-01
-2.75559038e-01 -1.75498277e-01 2.09456354e-01 5.38283527e-01
3.77562702e-01 -5.13186939e-02 5.01292944e-01 -3.71984452e-01
-1.00046182e+00 3.67252491e-02 -7.37193763e-01 -1.03611171e+00
7.86934137e-01 -2.67029285e-01 -7.77646542e-01 -1.26773655e-01
-7.10337818e-01 -1.70384198e-01 9.07158494e-01 4.04414535e-01
7.58747935e-01 -1.50337970e+00 -1.72046930e-01 4.17452365e-01
3.72489065e-01 5.29573411e-02 1.94992542e-01 7.91616201e-01
-7.08046615e-01 3.24577153e-01 -6.90044165e-01 -1.00262952e+00
-9.99152184e-01 5.17494977e-01 3.47760767e-01 6.89455092e-01
-6.64502501e-01 4.66434389e-01 7.83339515e-02 -8.43644500e-01
3.63874249e-02 -6.69469059e-01 2.49223799e-01 -4.99707520e-01
3.14207882e-01 3.26879233e-01 8.48504379e-02 -8.08582664e-01
-2.61031449e-01 1.15145183e+00 4.89717424e-01 -2.73051262e-01
1.69867182e+00 -5.07913470e-01 -4.81798410e-01 9.21570003e-01
1.07287455e+00 -4.48421091e-01 -1.45504487e+00 -4.08357173e-01
-1.80738881e-01 -2.59173930e-01 -8.56814906e-02 -7.83663630e-01
-7.44242311e-01 6.54948950e-01 7.25470722e-01 7.61127006e-03
9.39464331e-01 3.47881556e-01 3.80000502e-01 1.76107228e+00
9.62501347e-01 -1.27298343e+00 4.80872899e-01 7.18437552e-01
8.68569374e-01 -1.55927134e+00 1.90155074e-01 -6.82932675e-01
-2.15156689e-01 1.49507153e+00 3.74085426e-01 1.27499849e-01
3.57688636e-01 3.57906997e-01 1.42528594e-01 -4.91202503e-01
-3.74882013e-01 -1.16148382e-01 2.83394575e-01 1.14535511e+00
-6.84735402e-02 2.31706619e-01 1.11906171e+00 -9.28101242e-02
-2.89749324e-01 -1.83471039e-01 3.81181210e-01 1.24365306e+00
-1.04376483e+00 -6.72144830e-01 -8.97838533e-01 4.38069642e-01
1.57002583e-01 3.76506388e-01 -4.02656943e-01 3.01551431e-01
-1.12842925e-01 1.11579907e+00 1.99597433e-01 -6.47465229e-01
1.70456260e-01 1.47556365e-01 1.19882691e+00 -5.80638349e-01
4.02434587e-01 3.47693264e-02 2.88905799e-01 -8.46333683e-01
-9.74186957e-01 -7.48657227e-01 -1.23381674e+00 -1.12064220e-01
-1.24757089e-01 -4.85973358e-01 1.21873212e+00 1.13554490e+00
9.90540609e-02 5.85227251e-01 3.86476696e-01 -1.92577648e+00
-2.39941165e-01 -1.10738266e+00 -9.98504817e-01 1.54229924e-01
5.39220810e-01 -1.11719143e+00 3.36295441e-02 2.52021194e-01] | [5.920146465301514, -0.9078185558319092] |
5c94b21e-19b4-42d3-b883-e765a9202f0e | blocking-bandits | 1907.11975 | null | https://arxiv.org/abs/1907.11975v1 | https://arxiv.org/pdf/1907.11975v1.pdf | Blocking Bandits | We consider a novel stochastic multi-armed bandit setting, where playing an arm makes it unavailable for a fixed number of time slots thereafter. This models situations where reusing an arm too often is undesirable (e.g. making the same product recommendation repeatedly) or infeasible (e.g. compute job scheduling on machines). We show that with prior knowledge of the rewards and delays of all the arms, the problem of optimizing cumulative reward does not admit any pseudo-polynomial time algorithm (in the number of arms) unless randomized exponential time hypothesis is false, by mapping to the PINWHEEL scheduling problem. Subsequently, we show that a simple greedy algorithm that plays the available arm with the highest reward is asymptotically $(1-1/e)$ optimal. When the rewards are unknown, we design a UCB based algorithm which is shown to have $c \log T + o(\log T)$ cumulative regret against the greedy algorithm, leveraging the free exploration of arms due to the unavailability. Finally, when all the delays are equal the problem reduces to Combinatorial Semi-bandits providing us with a lower bound of $c' \log T+ \omega(\log T)$. | ['Rajat Sen', 'Sujay Sanghavi', 'Soumya Basu', 'Sanjay Shakkottai'] | 2019-07-27 | blocking-bandits-1 | http://papers.nips.cc/paper/8725-blocking-bandits | http://papers.nips.cc/paper/8725-blocking-bandits.pdf | neurips-2019-12 | ['product-recommendation'] | ['miscellaneous'] | [ 2.09494397e-01 5.10270357e-01 -6.63755059e-01 2.75437981e-02
-8.96673262e-01 -1.07990789e+00 -2.40407035e-01 -2.55394708e-02
-6.20717406e-01 9.94679749e-01 -4.62072730e-01 -1.11438549e+00
-1.00144625e+00 -8.21985543e-01 -1.18988335e+00 -9.06820357e-01
-4.64158803e-01 1.07494628e+00 -1.31087914e-01 -1.07970931e-01
1.79117516e-01 6.57399952e-01 -1.03033113e+00 -1.46858782e-01
2.66609609e-01 1.38227773e+00 1.55946478e-01 1.08240879e+00
-1.79242611e-01 4.63932663e-01 -3.67611855e-01 -3.34925056e-01
8.65678728e-01 -3.59576136e-01 -1.04826820e+00 3.92285198e-01
-3.43620688e-01 -3.94965857e-01 -1.59271270e-01 9.38295007e-01
-9.06713456e-02 4.18879628e-01 6.32261261e-02 -1.34865057e+00
-8.15399662e-02 1.21069217e+00 -1.24298394e+00 3.75142723e-01
1.29381409e-02 -1.44498155e-01 1.16615570e+00 3.83314878e-01
3.78949553e-01 9.99150157e-01 4.62972336e-02 3.74862909e-01
-1.24693024e+00 -5.57394147e-01 6.08500719e-01 -3.22280854e-01
-7.08045959e-01 -2.21777216e-01 5.12627900e-01 2.81061903e-02
6.76080108e-01 9.30843472e-01 5.07715464e-01 2.29921952e-01
-2.83920646e-01 8.13372195e-01 1.08460236e+00 -8.78932238e-01
4.06074703e-01 -5.91311604e-02 2.76890367e-01 3.59335363e-01
5.86884677e-01 4.24318820e-01 -2.12859824e-01 -2.25551516e-01
5.37212133e-01 4.24668223e-01 1.10626683e-01 -2.07007319e-01
-8.22284341e-01 7.75163829e-01 1.48347884e-01 9.45438296e-02
-6.48078084e-01 6.69144034e-01 4.51467521e-02 6.51475489e-01
2.33774871e-01 4.25621837e-01 -7.05198646e-01 -1.85679436e-01
-8.35124075e-01 2.05546096e-01 7.35058665e-01 1.29044688e+00
5.44862211e-01 -2.93604076e-01 -2.74480760e-01 3.35671693e-01
-1.64351180e-01 6.35987401e-01 -2.15958744e-01 -1.04256523e+00
9.76266563e-01 -8.40157494e-02 1.17791545e+00 -4.02514726e-01
-6.37110293e-01 -6.07561231e-01 -4.27209973e-01 1.64445955e-02
7.33364940e-01 -5.78210711e-01 -6.82435572e-01 1.83359182e+00
2.04579443e-01 -2.46557668e-01 -2.93839455e-01 7.80811787e-01
-5.86854577e-01 5.21439195e-01 -3.47973377e-01 -9.82966304e-01
1.23872530e+00 -1.10035217e+00 -5.33575237e-01 -5.72148979e-01
5.35486877e-01 -7.07786679e-01 4.49716210e-01 6.30836666e-01
-1.66094303e+00 2.31978685e-01 -7.42385745e-01 6.99411154e-01
4.93976772e-02 -2.50160396e-01 1.11914551e+00 1.07889628e+00
-7.37305820e-01 5.58079243e-01 -9.10902441e-01 7.86031857e-02
1.99999839e-01 5.94263494e-01 1.14543937e-01 -2.20071360e-01
-5.26839316e-01 4.39735353e-01 2.02384427e-01 3.98131520e-01
-7.06262171e-01 -1.52791038e-01 -1.36818483e-01 2.22270772e-01
1.09516263e+00 -3.92502367e-01 1.70331967e+00 -1.07278717e+00
-1.35722840e+00 4.19313937e-01 -1.55591220e-01 -6.40902877e-01
7.05932260e-01 7.85656571e-02 1.23256266e-01 -7.59172998e-03
1.80905312e-01 -6.69512451e-02 6.20249689e-01 -8.34703505e-01
-1.09834075e+00 -6.40068054e-01 7.40469158e-01 2.50406917e-02
-3.97154912e-02 3.31002697e-02 -1.31704450e-01 -1.44401640e-01
3.70256990e-01 -1.44950104e+00 -7.14697778e-01 -4.63582724e-01
-5.02107978e-01 -3.35680284e-02 -1.25310481e-01 -3.97799201e-02
9.81378734e-01 -1.92374790e+00 -2.57581085e-01 6.04046226e-01
-8.13796073e-02 -4.01245922e-01 -1.71874344e-01 6.29602909e-01
4.17703763e-03 2.88104028e-01 4.76420939e-01 -1.82072580e-01
2.37320900e-01 2.46623993e-01 -2.66836584e-01 6.54522955e-01
-7.24040747e-01 5.77890515e-01 -6.65979803e-01 8.35411102e-02
-2.89967835e-01 -9.62115586e-01 -4.71660554e-01 -2.25649282e-01
-5.14538467e-01 -5.20548895e-02 -9.50533569e-01 5.33247888e-01
9.71747994e-01 -5.07593930e-01 4.82961029e-01 7.21261382e-01
-2.16106921e-01 4.12592106e-02 -1.44192219e+00 1.20425117e+00
-6.21894598e-01 1.15025818e-01 5.63809097e-01 -1.14426494e+00
1.28973853e-02 1.13544911e-01 4.77366358e-01 -6.12929881e-01
2.49754176e-01 3.47864002e-01 -8.47201794e-02 -1.68120474e-01
4.50462461e-01 -4.02105302e-01 -1.59456298e-01 1.03642178e+00
-5.79273462e-01 4.97634679e-01 1.32291481e-01 2.16791883e-01
1.43763590e+00 -4.27079797e-01 -1.48109585e-01 -4.75552902e-02
-3.01666319e-01 2.69443750e-01 5.19389212e-01 1.51605582e+00
-1.03292838e-01 -1.09003440e-01 7.44864106e-01 -3.30582082e-01
-8.66628468e-01 -7.58108914e-01 3.48130256e-01 1.50131714e+00
4.52136040e-01 3.25882196e-01 -3.28094244e-01 -5.42818069e-01
3.75663131e-01 7.07962036e-01 -9.04357791e-01 3.56684774e-01
-1.03478335e-01 -7.32674778e-01 -2.12028041e-01 3.75600368e-01
1.25935838e-01 -6.18297577e-01 -7.82263219e-01 5.39178073e-01
-1.33384699e-02 -8.22012842e-01 -6.99577451e-01 7.54856050e-01
-9.37706888e-01 -8.70153964e-01 -5.05890310e-01 -1.91658780e-01
8.68181646e-01 9.14089143e-01 7.71384597e-01 -4.30213548e-02
5.36685623e-03 3.10954541e-01 -3.54170710e-01 -6.32468283e-01
2.01786116e-01 -6.60057068e-02 -7.09343180e-02 -2.66132534e-01
6.05730526e-02 -2.47742280e-01 -8.39481115e-01 3.14217120e-01
-7.13588476e-01 -8.96090455e-03 6.20504081e-01 7.74761617e-01
6.66206241e-01 4.30267394e-01 5.08529603e-01 -1.12802827e+00
3.82278979e-01 -5.01585364e-01 -1.21646762e+00 4.12828028e-01
-5.74956119e-01 1.94927394e-01 6.56806231e-01 -6.10968769e-01
-8.02600503e-01 1.63641483e-01 5.80776811e-01 -2.15466306e-01
2.87702262e-01 5.23249865e-01 3.69675249e-01 -2.46069785e-02
3.27635616e-01 -8.90783779e-03 -1.49459407e-01 -5.28998554e-01
2.64022708e-01 4.68507439e-01 7.05789924e-02 -8.06105375e-01
5.30631304e-01 4.94041950e-01 5.05953491e-01 -1.79054290e-01
-1.02009022e+00 -5.24712861e-01 3.82158220e-01 4.79167961e-02
-5.69038093e-02 -4.79748756e-01 -1.62699831e+00 -1.57497957e-01
-7.71117032e-01 -3.92812371e-01 -3.11895370e-01 5.33717752e-01
-1.01711750e+00 1.76177323e-01 -3.44781756e-01 -1.69088948e+00
-8.45076665e-02 -7.89527774e-01 4.54057992e-01 3.12454164e-01
2.64277369e-01 -3.26231748e-01 -3.79849464e-01 6.18758798e-01
4.17080671e-01 1.12204365e-02 7.96607971e-01 -5.99086285e-01
-1.01451874e+00 -3.79043132e-01 1.12191476e-02 -3.52343023e-01
-1.04289636e-01 -5.39626777e-01 -3.26531559e-01 -5.53332806e-01
-2.92906135e-01 -1.72613114e-01 4.71724421e-01 8.51137996e-01
1.28539681e+00 -9.12759721e-01 -6.39041305e-01 1.50095657e-01
1.45858645e+00 7.22275078e-01 1.59470186e-01 6.95373178e-01
-3.12597930e-01 2.24313930e-01 1.03008032e+00 8.88975203e-01
-1.44226596e-01 5.08811772e-01 8.31623435e-01 2.37259552e-01
8.78025889e-01 5.29220961e-02 -2.20433008e-02 -2.63478547e-01
-1.02755040e-01 -4.26005274e-01 -3.78363699e-01 8.31262231e-01
-2.01024604e+00 -8.20818961e-01 1.50242165e-01 2.67414951e+00
8.16869438e-01 7.59500146e-01 5.99871099e-01 -2.38454938e-02
7.34142721e-01 -3.81889671e-01 -8.96834493e-01 -9.90069866e-01
2.36529633e-01 2.69202948e-01 1.60620070e+00 3.81616890e-01
-7.06658185e-01 5.23528636e-01 5.66944265e+00 9.40885186e-01
-5.68616271e-01 2.46520787e-01 9.72135842e-01 -8.59353721e-01
-2.11761534e-01 3.12471658e-01 -5.53143144e-01 5.64216733e-01
1.08755720e+00 -4.25125808e-01 1.24789965e+00 1.05248117e+00
1.74852163e-01 -5.42296350e-01 -1.07582963e+00 7.01532722e-01
-7.67807782e-01 -1.35130310e+00 -8.35531831e-01 4.78855193e-01
8.45074713e-01 -1.80553526e-01 1.03906639e-01 8.50643218e-02
9.82235432e-01 -6.95970237e-01 9.53739107e-01 1.34562984e-01
5.13555288e-01 -1.23277080e+00 5.40689170e-01 8.64482999e-01
-7.56127596e-01 -6.37808621e-01 -2.71424085e-01 -2.40653098e-01
1.77153334e-01 5.63330889e-01 -6.32107496e-01 6.43074512e-01
5.35013139e-01 -6.08289123e-01 4.29028839e-01 1.29836917e+00
2.77392507e-01 3.82155865e-01 -9.41916645e-01 -3.69499892e-01
6.70178175e-01 -4.43459094e-01 3.96109670e-01 6.96793079e-01
4.05172110e-01 4.77476954e-01 2.61191398e-01 3.34909499e-01
-1.23505923e-03 -1.25174597e-01 -1.90350443e-01 -4.28027302e-01
5.65395355e-01 1.00170159e+00 -1.06323469e+00 1.50790727e-02
-1.96002707e-01 7.77895689e-01 7.34779909e-02 3.86995524e-01
-8.00631881e-01 -4.55751002e-01 4.64257568e-01 -3.13356072e-02
6.24001801e-01 -1.05861187e-01 -4.53567594e-01 -3.34702790e-01
2.22294733e-01 -2.79897124e-01 5.90915084e-01 -5.02512038e-01
-1.02351320e+00 2.25132689e-01 -1.43179297e-01 -7.71661878e-01
-2.83203512e-01 -4.09564227e-01 -2.68776685e-01 8.62797618e-01
-1.37493515e+00 -5.54194927e-01 5.93373716e-01 2.92209059e-01
2.62699693e-01 1.55291989e-01 4.16649938e-01 -2.38825053e-01
-5.18383861e-01 6.43398046e-01 6.06434703e-01 -4.69042003e-01
-2.32588593e-02 -1.18127930e+00 -1.42810076e-01 7.50574350e-01
-2.05195531e-01 5.67218065e-01 1.02614653e+00 -1.86593860e-01
-1.98901415e+00 -7.78115213e-01 5.44553339e-01 6.34506196e-02
8.02104235e-01 1.80348810e-02 2.11537674e-01 9.99774456e-01
-1.31970748e-01 -1.57593638e-02 6.14447951e-01 4.92681831e-01
1.16438389e-01 -3.76955748e-01 -1.21816397e+00 5.30726552e-01
1.11114979e+00 1.47190513e-02 5.78783229e-02 7.37206221e-01
7.13662863e-01 -4.88158643e-01 -3.85627866e-01 -5.64919449e-02
5.87983549e-01 -5.57245672e-01 4.69686151e-01 -1.12575889e+00
-1.06100418e-01 1.12191342e-01 -6.23351000e-02 -1.02038538e+00
-4.23022717e-01 -1.36339319e+00 -1.37056217e-01 6.79751337e-01
8.11358273e-01 -6.93969727e-01 1.21207130e+00 1.05719161e+00
2.25173831e-01 -6.56609178e-01 -1.44247997e+00 -1.13434255e+00
-2.39503002e-04 -4.36610579e-01 7.91182756e-01 3.86420459e-01
3.16041410e-01 -8.40742961e-02 -5.31804264e-01 2.33951628e-01
4.53052610e-01 9.56668794e-01 4.58921969e-01 -7.77697146e-01
-1.04640830e+00 -4.79655683e-01 5.28664768e-01 -1.46060622e+00
-1.34551033e-01 -5.24811625e-01 7.91752934e-02 -1.26688361e+00
5.37199438e-01 -1.10473263e+00 -6.58575177e-01 6.35543883e-01
3.29989821e-01 -3.79772455e-01 2.66765684e-01 -3.23827453e-02
-1.03589368e+00 -2.19259918e-01 1.32948577e+00 -1.29503906e-01
-3.12111050e-01 9.56215918e-01 -1.15318584e+00 1.81998715e-01
7.72333205e-01 -7.04618752e-01 -3.42413694e-01 -3.78013670e-01
7.43271351e-01 1.15828729e+00 -1.63582757e-01 -1.31372154e-01
2.73906201e-01 -8.87776315e-01 -1.04811452e-01 -6.89722121e-01
2.26217762e-01 -1.22062802e+00 4.19842333e-01 5.61735451e-01
-5.87117374e-01 9.86824334e-02 1.22938171e-01 1.12615025e+00
6.67158902e-01 -6.74151123e-01 3.85770679e-01 -3.00877839e-01
2.29505017e-01 4.50134635e-01 -3.77927482e-01 -2.70806104e-01
1.26049531e+00 -2.73210406e-01 -3.53897005e-01 -5.85406363e-01
-9.66654658e-01 6.97053015e-01 6.60360977e-02 -1.18127257e-01
-9.21308249e-02 -8.56736600e-01 -1.14165872e-01 -2.05861136e-01
-3.01993638e-01 -2.48902693e-01 5.70137024e-01 8.86216938e-01
-1.00081831e-01 8.37023616e-01 6.21953905e-02 -1.48952469e-01
-1.04877627e+00 1.07458377e+00 1.60766974e-01 -5.89693725e-01
-5.55994809e-02 9.82029319e-01 -2.74704963e-01 4.57322687e-01
3.37583274e-01 -1.58704504e-01 6.25379026e-01 -1.26141906e-01
3.73226345e-01 5.22780895e-01 1.84117153e-01 3.60139728e-01
-1.95762247e-01 -9.48967636e-02 -5.12081027e-01 -5.80558777e-01
1.39719641e+00 -3.42858762e-01 -1.26692131e-01 -4.69672456e-02
6.45715773e-01 1.33128017e-01 -1.22960460e+00 -2.75221556e-01
4.02334854e-02 -8.45584035e-01 4.67396192e-02 -9.33309734e-01
-1.20636714e+00 3.12624842e-01 3.15427333e-01 9.31716442e-01
1.26654661e+00 -3.38266082e-02 5.39902508e-01 5.52413344e-01
1.23668766e+00 -1.25454104e+00 -4.12947118e-01 2.06125632e-01
3.88772398e-01 -7.80862033e-01 -1.78332835e-01 -2.68964946e-01
-3.27941507e-01 9.52397406e-01 1.32805198e-01 3.55155692e-02
4.99014229e-01 2.34286278e-01 -6.88476384e-01 1.66422665e-01
-9.78874207e-01 -2.86757439e-01 -4.66117620e-01 4.59969454e-02
-6.93291798e-02 6.86292708e-01 -6.52032256e-01 8.97974074e-01
-1.33160025e-01 7.78087303e-02 7.02736974e-01 1.24868858e+00
-6.68936431e-01 -1.18731546e+00 -5.18372953e-01 8.44917297e-01
-1.02449179e+00 1.66696385e-01 2.65824478e-02 5.18950641e-01
-3.54060531e-01 1.39915395e+00 8.52460563e-02 1.43188536e-01
2.20693186e-01 -1.14002906e-01 7.46201813e-01 -4.86298621e-01
-3.25201869e-01 4.95571911e-01 1.86060086e-01 -3.88545126e-01
-2.40295902e-01 -5.35253286e-01 -9.53798354e-01 -6.87543929e-01
-8.15509021e-01 5.40772617e-01 7.40316093e-01 9.67496932e-01
2.49134526e-01 3.40299845e-01 1.25177288e+00 -4.01882321e-01
-1.17321944e+00 -6.39543891e-01 -1.19194198e+00 -2.23124161e-01
4.23950255e-01 -5.61100364e-01 -4.15291876e-01 -5.52632451e-01] | [4.556836128234863, 3.345365524291992] |
79a5ddeb-9cdb-4f59-b5a0-0590a2ecf593 | extracting-opinion-expressions-with-semi | null | null | https://aclanthology.org/D12-1122 | https://aclanthology.org/D12-1122.pdf | Extracting Opinion Expressions with semi-Markov Conditional Random Fields | null | ['Bishan Yang', 'Claire Cardie'] | 2012-07-01 | null | null | null | emnlp-2012-7 | ['fine-grained-opinion-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.303305149078369, 3.7040557861328125] |
4e405e8a-35a0-4a09-94a3-c963e8b25709 | face-image-reflection-removal | 1903.00865 | null | http://arxiv.org/abs/1903.00865v1 | http://arxiv.org/pdf/1903.00865v1.pdf | Face Image Reflection Removal | Face images captured through the glass are usually contaminated by
reflections. The non-transmitted reflections make the reflection removal more
challenging than for general scenes, because important facial features are
completely occluded. In this paper, we propose and solve the face image
reflection removal problem. We remove non-transmitted reflections by
incorporating inpainting ideas into a guided reflection removal framework and
recover facial features by considering various face-specific priors. We use a
newly collected face reflection image dataset to train our model and compare
with state-of-the-art methods. The proposed method shows advantages in
estimating reflection-free face images for improving face recognition. | ['Ling-Yu Duan', 'Boxin Shi', 'Alex C. Kot', 'Haoliang Li', 'Renjie Wan'] | 2019-03-03 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 7.47560441e-01 9.71128717e-02 4.63026196e-01 -7.35420167e-01
-6.98384106e-01 7.58381411e-02 4.58251417e-01 -1.28929079e+00
-2.07666792e-02 5.03033102e-01 2.81240612e-01 3.63729954e-01
1.49548769e-01 -6.03683889e-01 -6.85307860e-01 -1.03245747e+00
4.95124280e-01 -7.15130987e-03 -3.58114749e-01 -3.17610204e-02
1.17982551e-01 7.86664426e-01 -1.84297252e+00 6.15558028e-01
4.89749759e-01 7.34226823e-01 -3.25808534e-03 9.79830623e-02
1.95689537e-02 5.25865376e-01 -4.53777611e-01 -2.84639597e-01
4.43430305e-01 -4.68217254e-01 -1.63047999e-01 7.30350077e-01
9.23912108e-01 -7.46207654e-01 -4.29557323e-01 8.15351605e-01
2.58011818e-01 8.41352344e-02 7.13742197e-01 -8.88593495e-01
-7.78715312e-01 1.92964394e-02 -1.18195021e+00 -2.45705694e-01
5.96671641e-01 -3.26934248e-01 1.80330083e-01 -1.31096590e+00
4.51309890e-01 1.65983522e+00 7.43140042e-01 9.25213039e-01
-8.68263304e-01 -8.61825168e-01 1.22509062e-01 1.32261887e-01
-1.40279675e+00 -1.27859294e+00 1.15955055e+00 -5.83596788e-02
5.34814894e-01 2.93182850e-01 3.51262063e-01 1.02365386e+00
4.26714681e-02 4.96374011e-01 1.17825913e+00 -5.11915445e-01
-3.87366891e-01 7.59344250e-02 -1.28834233e-01 8.70551169e-01
3.16249728e-01 2.33930871e-01 -6.16332889e-01 -1.68914765e-01
7.32141376e-01 6.50839746e-01 -4.53070551e-01 2.73922741e-01
-3.73490363e-01 4.10975844e-01 -4.69546951e-02 -3.21155158e-03
-3.13351393e-01 -1.03492856e-01 -4.42378432e-01 3.75869066e-01
9.11670089e-01 -4.67830628e-01 -1.40596926e-01 6.18590474e-01
-1.14204037e+00 4.67464291e-02 6.34047627e-01 8.10092866e-01
8.90531480e-01 2.98648924e-01 -1.07783675e-01 1.30294669e+00
8.46726298e-01 9.58325565e-01 -2.52108648e-02 -9.57702160e-01
7.27991909e-02 1.57699019e-01 -3.57807092e-02 -8.69267106e-01
-2.44657710e-01 -8.79674554e-02 -8.31717014e-01 2.24048033e-01
8.06955919e-02 -4.05757055e-02 -1.07958257e+00 1.45281911e+00
4.76661474e-01 6.80337667e-01 1.29149441e-04 9.31864083e-01
1.17711186e+00 6.34534538e-01 -5.04415929e-01 -6.27357304e-01
1.38537729e+00 -8.82449329e-01 -9.39286113e-01 -2.86764473e-01
-1.79553539e-01 -1.27601695e+00 5.61292827e-01 6.65244401e-01
-1.02195704e+00 -4.53349710e-01 -7.59327054e-01 -5.95102124e-02
5.79542875e-01 3.61533314e-01 3.38898003e-01 9.96000111e-01
-8.97802651e-01 2.09642887e-01 -6.85766518e-01 -1.39370328e-02
6.61150575e-01 2.56893486e-01 -6.66487038e-01 -7.62753844e-01
-5.01758695e-01 6.41268909e-01 -7.75062621e-01 7.82348931e-01
-1.10832703e+00 -5.60894430e-01 -9.15192604e-01 -3.22259068e-01
4.02068585e-01 -4.34913695e-01 9.73219514e-01 -8.41529608e-01
-1.93318260e+00 9.41234887e-01 -6.95512593e-01 4.71876919e-01
3.51141900e-01 -5.23774147e-01 -4.66175050e-01 2.70548701e-01
-4.11696970e-01 -1.38260931e-01 1.61470973e+00 -1.48121941e+00
1.78136721e-01 -7.90262222e-01 -3.80928904e-01 5.93509749e-02
-1.97868884e-01 3.31744373e-01 -6.08489037e-01 -3.66945386e-01
4.87753451e-01 -9.88280296e-01 2.02140972e-01 1.16295725e-01
-3.32482249e-01 2.97384053e-01 1.34461176e+00 -8.22765112e-01
3.96120489e-01 -2.05762982e+00 -3.04028422e-01 7.25006685e-02
1.15070809e-02 3.51759642e-01 -5.63493669e-01 3.27177107e-01
-1.18706264e-01 -4.16661739e-01 -3.14398378e-01 -9.31897223e-01
-4.58375007e-01 5.30337691e-02 -2.71609008e-01 1.04039967e+00
2.41781905e-01 3.33294958e-01 -2.23808989e-01 -1.95218652e-01
1.24689907e-01 1.52787519e+00 -6.32628441e-01 3.63686025e-01
1.92505136e-01 8.64766955e-01 -4.77224559e-01 9.09166396e-01
1.58947122e+00 3.89397353e-01 1.02890030e-01 -2.64596432e-01
3.31949554e-02 -1.56722702e-02 -1.07769167e+00 1.30493140e+00
-5.62045455e-01 4.64604497e-01 5.03223121e-01 -7.18136728e-01
1.29199624e+00 3.45272124e-01 5.69976091e-01 -6.43988967e-01
9.32143107e-02 -9.04170722e-02 -4.17251021e-01 -7.37206817e-01
8.84205922e-02 -4.70603526e-01 6.76901579e-01 5.65853775e-01
-1.15048066e-01 -9.82793868e-02 -4.66949582e-01 -3.10437351e-01
8.86345685e-01 5.21556497e-01 -1.41537353e-01 -1.94746554e-02
8.15368295e-01 -9.46771979e-01 7.32012630e-01 2.87089556e-01
2.72701651e-01 1.04155433e+00 -7.39281327e-02 -3.97940427e-01
-4.50747937e-01 -9.47562873e-01 -3.60205650e-01 8.13343883e-01
-2.07988679e-01 -2.34091192e-01 -9.81583238e-01 -3.78996998e-01
-8.89245942e-02 2.64887184e-01 -4.53458995e-01 -1.44541845e-01
-7.26385355e-01 -8.15818906e-01 2.96606928e-01 8.51511508e-02
6.64110780e-01 -1.03926098e+00 -9.25166607e-02 -1.48199677e-01
-2.88185298e-01 -1.28152585e+00 -4.09135193e-01 -6.76485598e-01
-8.43826532e-01 -1.23564780e+00 -8.61283958e-01 -6.64728880e-01
1.18006229e+00 9.46159422e-01 9.92035031e-01 5.52626848e-01
-8.07246327e-01 6.17132783e-01 -9.60290357e-02 -4.25554335e-01
-4.05254960e-02 -7.65429080e-01 -7.63672963e-02 5.69212317e-01
4.54317153e-01 -4.82949764e-01 -7.77214229e-01 4.94272530e-01
-7.90909588e-01 -2.66200095e-01 5.75154722e-01 7.13456810e-01
4.74283397e-01 7.86452740e-02 1.77793145e-01 -1.38414812e+00
2.03266904e-01 -1.34342313e-01 -6.13383651e-01 2.30412539e-02
-1.50559753e-01 -2.26922438e-01 3.88994604e-01 -1.92343742e-01
-1.94857240e+00 6.93192929e-02 -1.84365690e-01 -4.01327461e-01
-1.69575229e-01 -9.17106494e-02 -4.57471520e-01 -4.45341051e-01
2.55836040e-01 1.62364319e-01 1.59585446e-01 -7.60628223e-01
8.05131271e-02 5.56245267e-01 2.40715519e-01 -5.13815105e-01
1.00582123e+00 1.00183165e+00 2.60387570e-01 -1.59122539e+00
-9.20416236e-01 -2.44709209e-01 -5.13389170e-01 -2.61339396e-01
3.48857105e-01 -9.84159172e-01 -9.85975981e-01 7.63290882e-01
-1.03907228e+00 -6.54456094e-02 2.29208201e-01 5.14849365e-01
-3.79503787e-01 5.43134749e-01 -6.65033460e-01 -1.30009544e+00
-2.81236291e-01 -1.13248634e+00 1.21281457e+00 2.35431582e-01
5.41876972e-01 -5.51072359e-01 8.01268741e-02 7.36694276e-01
3.67277622e-01 -7.76915625e-02 1.56707749e-01 2.75561422e-01
-8.27333987e-01 -3.33745740e-02 -3.12877804e-01 6.42556429e-01
5.35324454e-01 1.39355168e-01 -1.60746455e+00 -4.68128413e-01
6.85664237e-01 -1.93095393e-02 1.25803518e+00 4.39014941e-01
1.21232796e+00 -2.30724543e-01 -2.21188456e-01 1.03654277e+00
1.23370874e+00 -5.74832549e-03 1.07372391e+00 -1.90708026e-01
6.56752527e-01 9.59430277e-01 5.18323123e-01 6.40426040e-01
6.07260801e-02 6.09281719e-01 3.04790348e-01 -1.44457519e-01
-3.31663609e-01 1.23827837e-01 5.23057103e-01 6.13156259e-01
-4.84299600e-01 -2.26538554e-01 -3.22440863e-01 7.11555779e-02
-1.47246706e+00 -1.05466652e+00 -2.53030270e-01 2.12015200e+00
5.64062119e-01 -7.13735998e-01 -3.59548360e-01 -1.47321131e-02
4.37944025e-01 2.48980075e-01 -1.88627407e-01 -4.40874286e-02
1.54559026e-02 6.14691794e-01 1.56023830e-01 8.12690377e-01
-6.93683028e-01 7.69645452e-01 6.81268263e+00 2.46131554e-01
-9.52910006e-01 1.09860480e-01 4.72724676e-01 -2.72390544e-01
-4.59301710e-01 -1.02625184e-01 -1.11657012e+00 2.60355137e-02
5.64817965e-01 4.45897162e-01 5.26340604e-01 3.28033656e-01
3.38750899e-01 -6.88034073e-02 -9.60238278e-01 1.37542355e+00
8.01611960e-01 -5.53330004e-01 -2.27220599e-02 8.05453062e-02
6.21716261e-01 -3.78797680e-01 3.28524530e-01 -1.82986811e-01
-1.35963857e-01 -1.29115701e+00 1.62644595e-01 1.04978037e+00
7.38276958e-01 -7.10676610e-01 3.09673876e-01 -8.14175755e-02
-8.56436253e-01 2.05259025e-01 -7.56159842e-01 6.64439425e-02
7.57280923e-03 1.05435956e+00 -5.30269802e-01 4.71108377e-01
6.38744533e-01 8.16767097e-01 -4.20837626e-02 6.23760879e-01
-4.15864348e-01 4.90539700e-01 -4.93592680e-01 7.14585304e-01
-4.13672060e-01 -7.97172964e-01 4.74084944e-01 8.87101054e-01
5.94665825e-01 4.72197026e-01 -9.42326635e-02 8.55058253e-01
-2.61641860e-01 -1.75413847e-01 -7.05953002e-01 3.22037816e-01
1.25231713e-01 1.43038487e+00 -2.75728375e-01 1.47020444e-01
-8.53242040e-01 1.05955112e+00 -4.23324034e-02 6.97290480e-01
-6.81602299e-01 8.26833956e-03 8.44050050e-01 3.55973274e-01
3.24449211e-01 -1.07891843e-01 3.78180057e-01 -1.34203982e+00
3.08918804e-01 -7.20501423e-01 -6.84824139e-02 -6.99049592e-01
-1.19862640e+00 6.88729107e-01 -1.00982361e-01 -8.84333491e-01
1.37562081e-01 -6.76987827e-01 -6.54535055e-01 9.14168000e-01
-2.23343349e+00 -1.28823853e+00 -6.81634665e-01 1.05731845e+00
3.90557200e-01 -1.69936493e-01 7.60052919e-01 5.84748328e-01
-6.41044080e-01 6.61328793e-01 9.23957229e-02 -1.24801017e-01
1.02182066e+00 -2.12223992e-01 3.49282585e-02 7.09459126e-01
8.96695107e-02 8.29584360e-01 6.09849811e-01 -4.85313952e-01
-1.90103173e+00 -9.71616209e-01 3.68990541e-01 -7.11013302e-02
-1.86961249e-01 -4.00794089e-01 -8.48302722e-01 9.72207665e-01
3.67731787e-02 3.62711847e-01 9.11312401e-01 2.02612616e-02
-7.18693972e-01 -4.79370058e-01 -1.56232154e+00 3.32071811e-01
1.34323752e+00 -4.73880857e-01 -3.41861993e-01 3.87497902e-01
-9.76777002e-02 -1.22296728e-01 -4.02647734e-01 3.56500238e-01
8.78916979e-01 -1.22751439e+00 1.17757785e+00 1.15857890e-03
1.59256995e-01 -1.27214044e-01 -2.63144463e-01 -1.20981753e+00
-1.23683497e-01 -9.40451562e-01 1.77374303e-01 1.42004824e+00
-2.92070433e-02 -9.81188715e-01 9.81271565e-01 4.85812813e-01
-5.30763343e-02 -5.68608642e-02 -6.14506483e-01 -5.26663303e-01
-4.28083032e-01 -1.37257934e-01 5.85099339e-01 5.97746134e-01
-5.09223163e-01 -1.16081469e-01 -8.35254490e-01 1.82699740e-01
1.36030459e+00 2.72172600e-01 8.70128274e-01 -1.29184306e+00
-1.68414459e-01 3.31766605e-01 -6.97554424e-02 -8.84028316e-01
5.60339868e-01 -4.66826946e-01 2.32233703e-01 -1.15649974e+00
4.73537356e-01 -1.68912277e-01 2.49249619e-02 4.54063594e-01
2.83505563e-02 8.73966277e-01 2.23336555e-02 9.48424041e-02
-4.23908606e-02 7.53750682e-01 1.48224258e+00 3.55859548e-02
6.61085919e-02 7.83875883e-02 -8.35684240e-01 1.06779742e+00
5.41488409e-01 -5.07971287e-01 -4.95686024e-01 -6.74113095e-01
-1.03205271e-01 1.06327899e-01 2.06666097e-01 -8.10262442e-01
-1.17149629e-01 -2.44993463e-01 6.09743416e-01 -4.53548223e-01
1.03372037e+00 -1.02413249e+00 3.69600803e-01 4.72671874e-02
1.29793718e-01 -6.04407609e-01 1.72304258e-01 5.17201304e-01
-8.02739114e-02 -1.43317655e-01 1.09227252e+00 -2.74009019e-01
7.19119655e-03 5.67873597e-01 -3.63995910e-01 -3.53550792e-01
7.13051140e-01 -2.65375972e-01 -4.54015553e-01 -4.43227798e-01
-6.22963130e-01 -4.77783948e-01 3.69616151e-01 2.56780535e-01
1.23853004e+00 -1.07019436e+00 -1.08981252e+00 1.04941964e+00
-1.17113359e-01 -1.71294227e-01 5.24421155e-01 7.82692134e-01
-3.82136405e-01 -4.42045229e-03 -1.91971093e-01 -4.49700147e-01
-1.82038283e+00 1.40932590e-01 2.22193807e-01 4.00784731e-01
-9.41223204e-01 8.86614025e-01 6.63803697e-01 -3.63846302e-01
-2.10230816e-02 1.58023760e-01 -1.43091485e-01 -2.69325614e-01
1.11723101e+00 3.63198847e-01 2.26522803e-01 -8.86471152e-01
-3.88930529e-01 1.15960646e+00 -3.46692979e-01 9.04357433e-02
1.75936615e+00 -2.06732973e-01 -4.02153373e-01 -6.15649931e-02
1.30385745e+00 3.45043272e-01 -1.26173258e+00 -4.16978210e-01
-6.25828385e-01 -1.17014956e+00 2.43538380e-01 -3.89446467e-01
-1.84737527e+00 8.39069664e-01 4.26661074e-01 -5.12415349e-01
1.25256336e+00 -2.30629444e-01 6.60466671e-01 3.96393239e-01
2.96912402e-01 -6.19589627e-01 2.20767185e-01 2.66686350e-01
1.11824155e+00 -1.08644938e+00 3.90887082e-01 -1.06983387e+00
-1.59554079e-01 1.22185552e+00 4.11969662e-01 -4.12683815e-01
1.03294432e+00 5.78517079e-01 2.33788893e-01 -2.49533728e-01
-5.56090653e-01 5.87892439e-03 2.28612363e-01 8.66568744e-01
7.24314451e-01 -3.57905895e-01 -3.67139354e-02 1.57953247e-01
1.21496769e-03 1.29113302e-01 5.23992836e-01 8.37696314e-01
-2.93762475e-01 -1.21944332e+00 -8.81049097e-01 2.21624151e-01
-6.78248346e-01 3.39714140e-02 -2.44270027e-01 4.28833574e-01
-2.52609313e-01 1.32483542e+00 -1.08123444e-01 6.12374581e-02
1.94860935e-01 -3.29136364e-02 1.18461895e+00 -6.86008573e-01
-1.09705165e-01 5.03556728e-01 -8.06833580e-02 -6.30343258e-01
-7.63450265e-01 -7.56469071e-01 -9.59092677e-01 -4.33322608e-01
-4.96927023e-01 -1.96287870e-01 7.25268602e-01 7.66822517e-01
2.95648962e-01 3.82937461e-01 9.47515011e-01 -1.15408885e+00
-2.19173893e-01 -9.96842921e-01 -1.01588309e+00 5.06621957e-01
6.26912594e-01 -8.98545265e-01 -6.37785673e-01 1.00594558e-01] | [12.909052848815918, -0.05766501650214195] |
22b9b74c-109d-4a67-aa0f-463239ad4492 | signals-to-spikes-for-neuromorphic-regulated | 2106.11169 | null | https://arxiv.org/abs/2106.11169v4 | https://arxiv.org/pdf/2106.11169v4.pdf | Signals to Spikes for Neuromorphic Regulated Reservoir Computing and EMG Hand Gesture Recognition | Surface electromyogram (sEMG) signals result from muscle movement and hence they are an ideal candidate for benchmarking event-driven sensing and computing. We propose a simple yet novel approach for optimizing the spike encoding algorithm's hyper-parameters inspired by the readout layer concept in reservoir computing. Using a simple machine learning algorithm after spike encoding, we report performance higher than the state-of-the-art spiking neural networks on two open-source datasets for hand gesture recognition. The spike encoded data is processed through a spiking reservoir with a biologically inspired topology and neuron model. When trained with the unsupervised activity regulation CRITICAL algorithm to operate at the edge of chaos, the reservoir yields better performance than state-of-the-art convolutional neural networks. The reservoir performance with regulated activity was found to be 89.72% for the Roshambo EMG dataset and 70.6% for the EMG subset of sensor fusion dataset. Therefore, the biologically-inspired computing paradigm, which is known for being power efficient, also proves to have a great potential when compared with conventional AI algorithms. | ['Jean Rouat', 'Fabien Alibart', 'Dominique Drouin', 'Yann Beilliard', 'Ismael Balafrej', 'Nikhil Garg'] | 2021-06-09 | null | null | null | null | ['emg-gesture-recognition'] | ['medical'] | [ 8.36315036e-01 -2.33449385e-01 2.71051198e-01 1.79641008e-01
-3.25006783e-01 -1.85320422e-01 5.37473619e-01 -6.94895685e-02
-7.16517627e-01 9.21099305e-01 3.34671955e-03 2.52538621e-01
-1.44815862e-01 -5.94227433e-01 -9.08049941e-01 -1.28111184e+00
-4.61566180e-01 2.23173663e-01 2.85309911e-01 -3.35973889e-01
6.39382362e-01 3.64536524e-01 -2.09726930e+00 5.88741124e-01
6.47060335e-01 1.41724777e+00 3.81734043e-01 7.09508538e-01
-7.39537030e-02 6.80422723e-01 -5.70413411e-01 3.45268548e-01
7.21105188e-02 -8.61853063e-01 -2.55169302e-01 -6.74369454e-01
-1.54433563e-01 3.29928964e-01 -3.76384139e-01 8.29531670e-01
8.77340853e-01 -8.18328932e-02 5.59095085e-01 -6.93915844e-01
-2.98998564e-01 8.14347327e-01 1.36391938e-01 4.81322080e-01
1.96039289e-01 5.73387027e-01 3.14642787e-01 -6.29365802e-01
8.38261425e-01 3.58490020e-01 7.10743129e-01 8.13075423e-01
-1.31008136e+00 -5.65281451e-01 -6.67155206e-01 1.40966520e-01
-1.19042599e+00 -2.46106640e-01 7.65510619e-01 -3.39187413e-01
1.46836555e+00 8.54796097e-02 1.05092454e+00 1.41221118e+00
8.87830913e-01 4.35899615e-01 1.58089125e+00 -2.52054900e-01
8.79712105e-01 -3.71756583e-01 3.51241902e-02 4.04250890e-01
3.59199315e-01 6.38713062e-01 -1.31476367e+00 4.92353365e-02
8.49804461e-01 -5.39812893e-02 -1.50976092e-01 1.95307136e-01
-1.33229995e+00 2.46854126e-01 5.82703590e-01 7.23336041e-01
-9.54243302e-01 9.07938957e-01 3.35688263e-01 3.63239169e-01
-1.56072259e-03 7.46252000e-01 -2.28950270e-02 -7.22639143e-01
-1.07698393e+00 4.12780195e-01 8.56586993e-01 2.32871652e-01
1.66082904e-01 4.16177988e-01 -1.69418141e-01 2.08466038e-01
4.74779792e-02 4.15285707e-01 8.10756028e-01 -9.35009301e-01
-1.32408634e-01 7.72053242e-01 -1.77245840e-01 -4.10632700e-01
-3.94576728e-01 -3.45445216e-01 -1.03634953e+00 5.87305009e-01
5.89675248e-01 4.34521325e-02 -1.08382738e+00 1.30800569e+00
-4.10659283e-01 4.70521867e-01 2.86047190e-01 9.69732285e-01
5.51120162e-01 4.86804992e-01 -8.49636197e-02 -2.07751736e-01
1.19230390e+00 -9.69417617e-02 -6.71771944e-01 -3.87047827e-02
1.69011489e-01 -3.02467891e-03 7.69470930e-01 5.77234566e-01
-1.21104562e+00 -2.85458595e-01 -1.35503209e+00 5.43697655e-01
-5.14503121e-01 -3.89021605e-01 6.77229881e-01 3.10492814e-01
-1.04061174e+00 1.24823928e+00 -1.27405775e+00 -4.49170828e-01
5.47915459e-01 6.43649817e-01 -4.67037931e-02 5.52959383e-01
-8.89475524e-01 9.98200178e-01 4.32584375e-01 2.55674385e-02
-7.81149447e-01 -3.53037894e-01 7.10889548e-02 1.43530406e-02
-4.06393111e-01 -5.21133661e-01 7.86294699e-01 -4.80651140e-01
-1.80887127e+00 1.00099778e+00 -1.87030416e-02 -1.09754121e+00
7.88605511e-02 3.44543129e-01 -7.87547603e-02 2.12672532e-01
-4.85357225e-01 4.48796928e-01 7.97630608e-01 -7.81861663e-01
-3.68107669e-02 -5.61479032e-01 -7.84663141e-01 -2.74314523e-01
-3.26319486e-01 9.65652615e-03 3.85809869e-01 -5.90920091e-01
3.15004587e-01 -8.01463306e-01 -2.00084150e-01 -2.13339224e-01
1.58381671e-01 -2.63490558e-01 3.10763061e-01 -2.84848452e-01
1.05398607e+00 -1.97209144e+00 3.87104720e-01 2.29890913e-01
-8.29374650e-04 2.57282406e-01 9.04146507e-02 6.83512032e-01
1.05481356e-01 -2.52376109e-01 -6.02994561e-01 3.01919878e-01
-2.44520083e-01 2.24350587e-01 -2.26867303e-01 3.45984638e-01
5.96914113e-01 1.30004978e+00 -7.54839659e-01 -9.52626169e-02
-3.64669506e-03 5.70189238e-01 -1.71639398e-01 2.16657087e-01
-2.50882655e-01 8.58623326e-01 -3.09155107e-01 9.27359760e-01
6.77372664e-02 -1.21581018e-01 -9.93020087e-02 1.26357272e-01
-3.06315631e-01 3.30927700e-01 -9.10066545e-01 2.27644110e+00
6.36184290e-02 7.42379189e-01 -1.87991530e-01 -1.07704961e+00
1.37598217e+00 3.94946575e-01 6.71130300e-01 -1.13661385e+00
5.07616282e-01 9.03222263e-01 3.75824690e-01 -5.17016530e-01
-1.27381846e-01 -2.79223382e-01 3.22438441e-02 4.89766061e-01
3.51490706e-01 -1.12337522e-01 1.04342654e-01 -3.17986101e-01
1.75471580e+00 4.17175293e-01 -2.59984225e-01 -5.36238611e-01
1.60267338e-01 2.07616314e-01 1.21258333e-01 7.51711786e-01
-1.52887464e-01 4.74076182e-01 1.20493382e-01 -5.62587798e-01
-9.59123433e-01 -1.19808078e+00 -2.80725509e-01 6.54596090e-01
7.29243830e-02 1.82691827e-01 -9.59907591e-01 7.22813487e-01
1.09751888e-01 2.24026263e-01 -5.75808406e-01 -3.16993386e-01
-9.59884524e-01 -7.55647719e-01 9.86642003e-01 7.57499933e-01
4.26986635e-01 -1.81141245e+00 -1.58567202e+00 9.97847557e-01
2.30197117e-01 -8.16982687e-01 4.79904473e-01 9.32967544e-01
-1.35402787e+00 -7.71565795e-01 -9.07044053e-01 -6.58946872e-01
1.48893297e-01 -7.03544855e-01 7.37185299e-01 1.48648480e-02
-7.66270936e-01 1.21568395e-02 -2.64456540e-01 -6.67559266e-01
-3.08286577e-01 -1.41630983e-02 1.19081549e-01 -2.10119769e-01
7.10299253e-01 -1.18220425e+00 -7.07564294e-01 -2.94296145e-01
-7.62831688e-01 -1.78296685e-01 6.32273018e-01 9.66664553e-01
8.34583759e-01 -4.00438696e-01 6.03875875e-01 -2.49906834e-02
8.28119338e-01 -2.53828675e-01 -1.44116402e-01 -1.65221095e-01
-6.88030064e-01 4.36683238e-01 5.24786174e-01 -6.72263682e-01
-3.85822684e-01 3.20774436e-01 1.67271867e-01 -2.85596192e-01
-7.35059381e-02 3.52606446e-01 6.42984629e-01 -2.84151018e-01
1.02179182e+00 7.97288716e-01 4.16661561e-01 -2.33972967e-01
-3.08536053e-01 7.96116889e-01 1.01075220e+00 -5.14979064e-01
1.36225656e-01 5.59782565e-01 3.16469640e-01 -6.89131856e-01
3.92463326e-01 -2.29393750e-01 -4.43948835e-01 -4.22699630e-01
8.13728988e-01 -4.40443099e-01 -1.14293587e+00 1.12150645e+00
-1.18242848e+00 -3.69176626e-01 -5.95826864e-01 5.02835631e-01
-1.07741427e+00 -3.58062506e-01 -8.72554004e-01 -1.43593693e+00
-7.03014791e-01 -7.10107386e-01 8.28246295e-01 3.99982989e-01
-2.25852698e-01 -3.34244072e-01 2.73829073e-01 -3.04842025e-01
9.36873734e-01 9.23004448e-01 3.96971375e-01 -4.19202566e-01
-5.63906372e-01 -2.22848833e-01 3.14310908e-01 1.37817204e-01
-3.03276896e-01 -2.63105154e-01 -1.23568273e+00 -1.61629096e-01
2.57746667e-01 -3.38358909e-01 1.26020789e+00 3.74943256e-01
8.63686085e-01 1.71425298e-01 -4.26578850e-01 5.14718235e-01
1.62795281e+00 3.78610641e-01 1.01623964e+00 3.37253809e-01
-4.95977663e-02 3.65829319e-01 -6.24729395e-02 3.11265975e-01
-2.44345427e-01 3.49241406e-01 4.31291312e-01 3.06407332e-01
-6.85388818e-02 -1.71378806e-01 5.33039868e-01 9.71891463e-01
-7.33201563e-01 -4.91973087e-02 -8.68593574e-01 4.85140145e-01
-1.77702641e+00 -1.16193545e+00 -5.32409176e-02 2.05486536e+00
7.41869390e-01 1.88208923e-01 1.34460390e-01 5.75388134e-01
4.53563333e-01 -1.86782598e-01 -9.16552544e-01 -6.10992372e-01
-5.98261952e-01 1.18054664e+00 5.27714729e-01 -4.67870772e-01
-4.86559778e-01 5.57575285e-01 6.27082825e+00 2.11730316e-01
-1.38459921e+00 1.44364163e-01 -7.20370784e-02 -4.54721421e-01
1.75579429e-01 -3.66251767e-01 -3.18098038e-01 1.01176977e+00
1.74395382e+00 9.40234140e-02 1.01028943e+00 1.84029520e-01
9.62942392e-02 -3.65068704e-01 -9.42342877e-01 1.20354486e+00
-1.69483706e-01 -1.83402574e+00 -4.09961939e-01 2.30974644e-01
4.79219466e-01 6.15887523e-01 -2.54297763e-01 -1.46791890e-01
-4.37430054e-01 -1.47566295e+00 6.91889584e-01 1.13727248e+00
7.21302867e-01 -3.96878034e-01 6.58072412e-01 3.36094290e-01
-1.11027956e+00 -3.02932084e-01 -2.64112353e-01 -4.47368026e-01
1.56746268e-01 2.58541495e-01 -8.06277916e-02 -5.56091517e-02
1.05666471e+00 5.95567167e-01 -1.25364646e-01 1.00212944e+00
3.03473085e-01 7.14415193e-01 -7.25872278e-01 -8.61364305e-01
1.35069475e-01 -2.11837608e-02 5.88474572e-01 1.16230536e+00
5.73427022e-01 1.98107883e-01 -7.55519927e-01 1.45064712e+00
1.11549303e-01 -3.85966301e-01 -8.68253350e-01 -7.62878180e-01
5.27019799e-01 6.89163864e-01 -7.86060989e-01 -1.80705652e-01
3.64589393e-01 1.07848024e+00 -9.04816240e-02 8.51795375e-02
-4.57656860e-01 -6.61064863e-01 4.35983390e-01 1.20908856e-01
3.98602813e-01 -4.11983848e-01 -7.31747508e-01 -8.04132402e-01
3.68515790e-01 -4.33960110e-01 -1.96427554e-01 -7.36408412e-01
-1.00740647e+00 6.63750768e-01 -4.73663896e-01 -1.15624118e+00
-4.77212846e-01 -1.16792417e+00 -7.83577740e-01 7.40301132e-01
-1.13070762e+00 -4.23985809e-01 -3.47315043e-01 5.41946411e-01
5.12750223e-02 -2.46250227e-01 1.22534120e+00 -1.05651073e-01
-2.46528257e-02 1.07752815e-01 3.44534963e-01 -1.10771298e-01
-2.14838528e-05 -1.02508116e+00 3.80603462e-01 5.99348664e-01
6.49616420e-02 4.29591894e-01 6.09529138e-01 -5.04710734e-01
-2.12638807e+00 -4.07145739e-01 7.48302042e-01 -1.54873177e-01
6.75740063e-01 -4.22376603e-01 -1.13692153e+00 -1.32531956e-01
3.05354387e-01 2.09598988e-02 3.61686409e-01 -6.86748922e-01
3.83726768e-02 -1.29414499e-01 -1.36784172e+00 2.36608863e-01
1.40334618e+00 -4.50672656e-01 -7.96776891e-01 2.88086850e-03
7.41925910e-02 -2.95013011e-01 -1.22391319e+00 4.78182852e-01
1.00364113e+00 -9.41807032e-01 6.83814585e-01 -4.53399479e-01
4.17765260e-01 -2.21598402e-01 -2.00242519e-01 -1.10421777e+00
-1.10947974e-02 -8.69170785e-01 -6.28255844e-01 6.86927497e-01
1.67125210e-01 -1.05281079e+00 8.97594273e-01 2.16393232e-01
-2.05584839e-01 -1.08033371e+00 -1.52768481e+00 -1.11673760e+00
3.94592769e-02 -2.45931208e-01 4.21476245e-01 2.15831965e-01
4.83668536e-01 -2.28089914e-01 1.61947161e-01 -5.60066104e-01
7.79718816e-01 -9.58498195e-03 1.62789840e-02 -1.35261226e+00
-1.01573266e-01 -7.64129221e-01 -1.00364900e+00 -2.74337828e-01
-3.14032674e-01 -1.05170310e+00 4.38752323e-01 -1.25013685e+00
-1.50701940e-01 -1.69732764e-01 -1.05570757e+00 3.03187370e-01
5.11158109e-01 5.58210313e-01 2.49765545e-01 4.55265731e-01
1.21016063e-01 2.27922961e-01 9.53004003e-01 -6.25608563e-02
-2.98150420e-01 -4.03127939e-01 -2.86706746e-01 1.27516761e-01
9.07646120e-01 -6.13876343e-01 3.11261535e-01 -5.03741391e-02
6.16364516e-02 1.06685221e-01 7.07488000e-01 -1.79440737e+00
8.26601684e-01 2.00940609e-01 4.08132881e-01 -4.30458665e-01
4.77848798e-01 -4.72481638e-01 3.46020579e-01 1.22831535e+00
-4.04526889e-01 2.23558508e-02 2.63327044e-02 5.39831877e-01
-1.49545342e-01 9.30064917e-02 6.38483107e-01 -3.43377948e-01
-6.84536040e-01 -2.07137063e-01 -8.58252704e-01 -1.94522381e-01
8.36460292e-01 -1.01838410e+00 -3.15702200e-01 3.96884382e-01
-5.96671462e-01 -3.54666620e-01 4.54745650e-01 2.87015080e-01
8.43230367e-01 -1.14743340e+00 -7.14852452e-01 6.76741004e-01
4.56486903e-02 -5.24032474e-01 -2.38075465e-01 1.01429439e+00
-3.72272998e-01 4.31992114e-01 -9.98353183e-01 -1.02498758e+00
-4.45816606e-01 2.89913826e-02 4.25194949e-01 -1.47096999e-02
-7.21782386e-01 7.63727784e-01 -1.14023006e+00 3.38094495e-02
2.79386312e-01 -5.54937482e-01 1.11712001e-01 -2.58663297e-01
4.58020419e-01 5.05148053e-01 2.58200198e-01 -1.93958715e-01
-6.46254718e-01 6.46418929e-01 8.22616458e-01 -2.26248726e-01
1.70484316e+00 5.94649255e-01 -2.04882428e-01 8.71809185e-01
7.65848577e-01 -8.36696029e-01 -1.04636741e+00 1.65994525e-01
3.03686351e-01 2.11904883e-01 6.87405281e-03 -1.12954152e+00
-7.21742630e-01 1.05077672e+00 1.24933684e+00 9.58446786e-02
1.28847897e+00 -2.39272609e-01 9.91631210e-01 5.78624487e-01
1.00266254e+00 -1.30263364e+00 7.28894770e-02 3.85975242e-01
9.21410441e-01 -8.94913316e-01 -4.59498286e-01 2.81678706e-01
-1.17678672e-01 1.45092654e+00 2.08749771e-01 -9.51134562e-01
4.45654243e-01 8.60858381e-01 -1.98617592e-01 -3.40957403e-01
-7.94813991e-01 -2.41645917e-01 -6.89275935e-02 6.33549035e-01
4.84358221e-01 2.28784472e-01 -7.10299790e-01 5.55488110e-01
5.03860088e-03 9.41219449e-01 1.72133431e-01 1.25818682e+00
-6.52150571e-01 -7.22744823e-01 -1.51285276e-01 7.71655977e-01
-2.73817718e-01 -6.83514103e-02 -4.28768456e-01 3.89943957e-01
-4.25759256e-02 5.79852998e-01 2.87174374e-01 -7.08477139e-01
1.49687439e-01 4.40185547e-01 9.26268697e-01 -2.83198714e-01
-1.24494767e+00 -2.55773932e-01 -3.48203987e-01 -8.56777430e-01
-6.41485274e-01 -4.99291450e-01 -2.03970742e+00 -8.24520066e-02
-1.39045611e-01 -2.07582667e-01 1.38004065e+00 8.95312428e-01
7.26625741e-01 6.07683003e-01 2.82394201e-01 -1.32171464e+00
-6.72261834e-01 -1.07089913e+00 -7.63751447e-01 2.38170281e-01
-8.04750472e-02 -7.40855575e-01 -2.64519632e-01 -5.93516463e-03] | [8.264547348022461, 2.454026699066162] |
4a914cae-6c52-4a75-bd06-8689e6a92e7b | scaling-semantic-parsers-with-on-the-fly | null | null | https://aclanthology.org/D13-1161 | https://aclanthology.org/D13-1161.pdf | Scaling Semantic Parsers with On-the-Fly Ontology Matching | null | ['Luke Zettlemoyer', 'Tom Kwiatkowski', 'Eunsol Choi', 'Yoav Artzi'] | 2013-10-01 | null | null | null | emnlp-2013-10 | ['ontology-matching'] | ['knowledge-base'] | [-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] | [-7.342721939086914, 3.7508883476257324] |
d08ee114-6a72-4be3-a13b-66a499f90827 | understanding-dynamic-scenes-using-graph | 2005.04437 | null | https://arxiv.org/abs/2005.04437v5 | https://arxiv.org/pdf/2005.04437v5.pdf | Understanding Dynamic Scenes using Graph Convolution Networks | We present a novel Multi-Relational Graph Convolutional Network (MRGCN) based framework to model on-road vehicle behaviors from a sequence of temporally ordered frames as grabbed by a moving monocular camera. The input to MRGCN is a multi-relational graph where the graph's nodes represent the active and passive agents/objects in the scene, and the bidirectional edges that connect every pair of nodes are encodings of their Spatio-temporal relations. We show that this proposed explicit encoding and usage of an intermediate spatio-temporal interaction graph to be well suited for our tasks over learning end-end directly on a set of temporally ordered spatial relations. We also propose an attention mechanism for MRGCNs that conditioned on the scene dynamically scores the importance of information from different interaction types. The proposed framework achieves significant performance gain over prior methods on vehicle-behavior classification tasks on four datasets. We also show a seamless transfer of learning to multiple datasets without resorting to fine-tuning. Such behavior prediction methods find immediate relevance in a variety of navigation tasks such as behavior planning, state estimation, and applications relating to the detection of traffic violations over videos. | ['K. Madhava Krishna', 'Anoop Namboodiri', 'Priyesh Vijayan', 'Mahtab Sandhu', 'Balaraman Ravindran', 'Sravan Mylavarapu'] | 2020-05-09 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [-4.52051908e-02 8.21871236e-02 -4.01647955e-01 -5.70936322e-01
-3.73872370e-01 -2.71730810e-01 6.97285950e-01 -2.28992347e-02
-4.08195376e-01 3.05343419e-01 1.90261409e-01 -5.62624037e-01
-2.65792668e-01 -8.63063097e-01 -1.14609051e+00 -4.99774843e-01
-4.81641203e-01 4.84996378e-01 8.06606233e-01 -3.62335622e-01
8.29460844e-02 7.42385447e-01 -1.51812243e+00 3.04881096e-01
2.46130913e-01 9.86949861e-01 2.82439768e-01 8.58847976e-01
3.17192793e-01 1.47029245e+00 -5.77484779e-02 -2.82860130e-01
2.13430136e-01 1.76885158e-01 -8.00997138e-01 1.29506141e-01
6.88583791e-01 -5.17331243e-01 -1.12329292e+00 8.34918439e-01
-2.17053369e-01 6.81628168e-01 3.77314448e-01 -1.72927415e+00
-4.52736855e-01 2.02071115e-01 -2.78137088e-01 6.65369034e-01
3.60860318e-01 3.87184620e-01 1.10178113e+00 -5.07341683e-01
8.47846925e-01 1.37233591e+00 3.17070991e-01 2.98767865e-01
-7.81713724e-01 -4.08730805e-01 7.57251620e-01 1.01194227e+00
-1.10350275e+00 -5.74990571e-01 6.88437819e-01 -5.98191082e-01
1.28047800e+00 7.89641365e-02 6.34075046e-01 9.55228388e-01
2.28706390e-01 8.64760697e-01 3.62486392e-01 3.83262128e-01
6.01448715e-02 -2.94078887e-01 4.57507670e-01 1.06422806e+00
-9.20164958e-02 1.58025458e-01 -5.63988209e-01 2.47064698e-02
5.69233358e-01 1.44023061e-01 6.63940981e-02 -7.49279797e-01
-1.05460441e+00 5.51576436e-01 6.90497339e-01 -1.84173673e-01
-3.51276636e-01 6.65166318e-01 3.33811313e-01 2.60396689e-01
3.30314845e-01 -1.83348343e-01 -3.52344751e-01 -1.40795335e-02
-3.06238353e-01 2.22022295e-01 2.97552288e-01 1.20136237e+00
1.06401670e+00 -6.29536882e-02 -2.51930118e-01 4.84431356e-01
4.02838379e-01 2.98893273e-01 -1.63509086e-01 -1.26961708e+00
8.21789324e-01 7.81165361e-01 1.47449881e-01 -1.23792744e+00
-5.22566199e-01 4.49581072e-02 -4.46413398e-01 1.24865703e-01
7.43514672e-02 1.42786875e-01 -8.39875698e-01 1.79893613e+00
5.07100761e-01 8.58749866e-01 -6.36796653e-02 9.81101274e-01
1.00230551e+00 7.68755317e-01 1.08849749e-01 5.34502380e-02
1.15181565e+00 -1.30926311e+00 -5.87823868e-01 -3.55529755e-01
7.81921029e-01 -1.69680696e-02 6.87981844e-01 -4.68337834e-02
-9.52281713e-01 -6.26968503e-01 -6.83369517e-01 -2.69075572e-01
-5.58923602e-01 -2.08763555e-01 5.25978506e-01 1.01266438e-02
-1.43471467e+00 2.24132270e-01 -1.06766069e+00 -4.33716118e-01
4.79445010e-01 4.90697742e-01 -6.60078526e-01 -1.81583092e-01
-1.10009408e+00 9.78706896e-01 7.11282566e-02 4.84846443e-01
-1.41832459e+00 -2.51669019e-01 -1.31288075e+00 1.97592434e-02
7.12445259e-01 -6.13440096e-01 1.07622337e+00 -5.92743099e-01
-1.24758303e+00 7.29984641e-01 -4.35190082e-01 -6.19325042e-01
3.24331939e-01 -2.44115144e-01 -5.64645529e-01 2.86013454e-01
2.07759246e-01 6.72500193e-01 5.43394923e-01 -1.12977004e+00
-9.01911676e-01 -3.09366375e-01 5.61319828e-01 4.42780793e-01
1.49982348e-01 6.50163218e-02 -9.77285624e-01 -1.58768967e-02
-6.13626465e-02 -1.02797365e+00 -3.96612108e-01 6.16393518e-03
-2.55697817e-01 -3.65847260e-01 1.28667951e+00 -3.91164750e-01
9.23757195e-01 -2.06581903e+00 2.58299857e-01 1.59033671e-01
3.36836874e-01 2.52612442e-01 -4.24149036e-01 4.15159732e-01
5.43233305e-02 -1.23928532e-01 2.01311558e-01 -2.75493383e-01
-1.00897096e-01 4.65287626e-01 -1.41345695e-01 7.58993268e-01
1.75623432e-01 1.14164376e+00 -1.20801020e+00 -3.71434301e-01
7.27801383e-01 4.72699106e-01 -4.27034706e-01 3.81473780e-01
-3.24020118e-01 5.13707638e-01 -6.75382137e-01 3.99893045e-01
4.08392370e-01 -2.59589046e-01 -3.14862132e-02 -1.51413560e-01
-9.97956023e-02 4.36435163e-01 -8.35484803e-01 1.39288437e+00
-2.79443681e-01 9.18358445e-01 -2.48859990e-02 -1.03992975e+00
5.23706377e-01 6.39034286e-02 5.72425246e-01 -8.97720218e-01
-1.10279329e-01 -5.31097710e-01 -8.33925009e-02 -9.36550379e-01
5.48445642e-01 5.47928393e-01 -1.10493720e-01 1.44906074e-01
7.98786804e-02 3.58570516e-01 4.22431171e-01 4.89639312e-01
1.48119724e+00 2.14268818e-01 1.07971676e-01 -2.53291074e-02
7.00861156e-01 2.20272858e-02 6.43313527e-01 7.15661049e-01
-4.96744186e-01 1.07064389e-01 7.67174304e-01 -8.89173985e-01
-6.66649878e-01 -8.93493772e-01 4.26663280e-01 1.41142273e+00
6.91560805e-01 -3.98008138e-01 -3.34117204e-01 -6.19380593e-01
-6.64507225e-02 6.25882626e-01 -7.80863225e-01 -2.61299908e-01
-8.97659183e-01 -2.12824598e-01 3.27257544e-01 5.96482694e-01
4.85162318e-01 -1.00066233e+00 -7.99978614e-01 4.81386334e-02
-2.06673890e-01 -1.76243591e+00 -4.66994524e-01 -2.22080499e-01
-5.07515788e-01 -1.28947306e+00 1.55368224e-01 -7.07830369e-01
5.59716105e-01 6.61873639e-01 9.72981036e-01 1.25219434e-01
9.39472839e-02 6.37193918e-01 -2.49312505e-01 2.20140174e-01
-3.32614899e-01 -2.23490093e-02 -1.15196876e-01 5.09815633e-01
1.63141966e-01 -4.37056303e-01 -7.19689012e-01 6.26951337e-01
-4.31020170e-01 2.66268820e-01 2.09386304e-01 2.85465419e-01
4.94230241e-01 -1.03588261e-01 1.03553869e-01 -7.49123573e-01
1.84898321e-02 -7.34289408e-01 -9.09306288e-01 2.06807569e-01
8.97024572e-02 -4.78889644e-02 4.99785900e-01 -1.51281476e-01
-8.97583425e-01 7.87757337e-03 1.07403388e-02 -7.33111918e-01
-4.31842834e-01 4.30025637e-01 -1.67944118e-01 -3.84065066e-03
1.97748333e-01 1.45536372e-02 -1.87828898e-01 1.17048793e-01
5.27785420e-01 7.24012256e-02 5.26376605e-01 -2.56205708e-01
5.43654978e-01 6.25043094e-01 2.56721884e-01 -8.56061578e-01
-5.67345142e-01 -6.92510366e-01 -6.25744104e-01 -6.42471969e-01
1.20837498e+00 -1.08062208e+00 -1.22887599e+00 4.13955599e-01
-1.10255170e+00 -9.98018563e-01 1.97649986e-01 3.73883545e-01
-6.51146114e-01 3.09191614e-01 -7.04273105e-01 -6.33103371e-01
3.92709613e-01 -1.39888120e+00 1.31871104e+00 -3.75549588e-03
2.23708585e-01 -1.12191904e+00 -6.80336282e-02 5.50214708e-01
1.08774729e-01 1.79334760e-01 8.52911294e-01 -3.99836868e-01
-1.36845148e+00 6.88608922e-03 -2.92452931e-01 -1.80838376e-01
2.06175931e-02 4.51272801e-02 -7.58900762e-01 -1.71689779e-01
-5.14890373e-01 -1.57556698e-01 1.00175202e+00 4.67905223e-01
1.22928631e+00 -3.10599893e-01 -7.27562964e-01 6.51785374e-01
1.12425780e+00 3.98430467e-01 5.47328413e-01 1.67047948e-01
1.22096395e+00 8.01985502e-01 6.81547105e-01 -2.46689618e-02
1.02408576e+00 1.05761027e+00 1.00593781e+00 -4.36784439e-02
-6.59828037e-02 -2.38191471e-01 5.58320463e-01 2.99624711e-01
-5.65825477e-02 -6.43687785e-01 -1.06072056e+00 7.04811931e-01
-2.37763572e+00 -1.33839083e+00 -3.62626553e-01 1.90076005e+00
-7.77059272e-02 2.28110537e-01 2.08658278e-01 -4.43537474e-01
7.45107770e-01 6.02929235e-01 -7.40922213e-01 -2.93256491e-01
1.90324292e-01 -5.29762387e-01 7.09597111e-01 9.45132911e-01
-1.27785742e+00 1.22477067e+00 5.88963270e+00 3.05002689e-01
-8.40920925e-01 1.43366024e-01 6.72061086e-01 -3.95160690e-02
-1.60347283e-01 -7.76420459e-02 -7.86026657e-01 1.08075216e-01
1.01596892e+00 3.20539415e-01 5.50183952e-01 6.42660916e-01
5.17306745e-01 -1.88691095e-01 -1.31634450e+00 8.25380385e-01
-5.85654601e-02 -1.37508917e+00 9.05851796e-02 6.12110719e-02
5.63717961e-01 5.32660425e-01 1.32064924e-01 3.45817477e-01
6.76591933e-01 -1.00062025e+00 7.28660464e-01 4.54243541e-01
3.88258547e-01 -5.63552856e-01 4.07468230e-01 3.49759251e-01
-1.63494253e+00 -3.14806938e-01 -4.00447696e-02 -9.97034535e-02
5.28515697e-01 -3.80538846e-03 -8.43131423e-01 4.83460128e-01
6.06089294e-01 1.12610495e+00 -4.87266213e-01 7.28129625e-01
-1.19445622e-01 4.92618740e-01 -1.05210990e-01 1.85160324e-01
5.69723964e-01 -3.49602938e-01 6.47100270e-01 1.18027294e+00
-1.67021956e-02 2.69074529e-01 2.42034778e-01 5.78033686e-01
1.40268639e-01 -4.89035159e-01 -1.12015438e+00 3.79426867e-01
2.11180076e-01 1.10398209e+00 -6.59263432e-01 -4.10509050e-01
-8.44867349e-01 6.09656274e-01 7.22595632e-01 7.65402377e-01
-1.28258240e+00 3.03703636e-01 1.16660357e+00 1.48622200e-01
5.20277500e-01 -4.31990981e-01 2.12070376e-01 -9.97775614e-01
-2.71185320e-02 -4.04084057e-01 4.11352634e-01 -9.65950608e-01
-6.84300959e-01 6.11423790e-01 3.49585980e-01 -1.15497530e+00
-2.51420408e-01 -7.38981307e-01 -4.32501227e-01 4.60316271e-01
-1.67673504e+00 -1.25521302e+00 -3.98042828e-01 9.24829185e-01
6.55561686e-01 -1.49263069e-01 2.71110922e-01 3.62373948e-01
-7.46184170e-01 1.30073875e-01 -5.98633349e-01 3.30992311e-01
6.50274754e-02 -1.01284194e+00 8.60131323e-01 1.09477866e+00
1.62502959e-01 2.87723869e-01 4.96856958e-01 -5.81856847e-01
-1.50314891e+00 -1.60006344e+00 7.85103977e-01 -6.70373201e-01
7.17018187e-01 -4.04739499e-01 -8.30818415e-01 1.28244817e+00
2.03523040e-01 5.74508548e-01 2.28760168e-01 -4.73638885e-02
-3.35471034e-01 -2.89343387e-01 -5.95492184e-01 7.32197762e-01
1.61213887e+00 -7.32558966e-01 -3.13059479e-01 3.75719428e-01
6.33146524e-01 -7.13909686e-01 -2.67078310e-01 4.87832338e-01
3.25465649e-01 -1.06562269e+00 1.04036105e+00 -9.12518263e-01
2.61301368e-01 -4.87275690e-01 -2.35960364e-01 -9.29820001e-01
-4.56512630e-01 -4.26202506e-01 -2.58562267e-01 5.15507638e-01
3.51463884e-01 -4.24555093e-01 7.26477981e-01 6.97342992e-01
-6.42544210e-01 -6.57965779e-01 -9.80425715e-01 -3.49129617e-01
-6.15760326e-01 -7.78795600e-01 4.30847555e-01 6.59620225e-01
-2.07132056e-01 3.90066385e-01 -5.03861129e-01 8.15953672e-01
4.38349009e-01 -4.46770489e-02 1.04852068e+00 -8.82916391e-01
-5.14754057e-02 -3.36130589e-01 -9.21013236e-01 -1.60376441e+00
5.48175693e-01 -7.96238959e-01 9.79013294e-02 -1.78270984e+00
1.61133125e-01 -1.98256522e-01 -3.50695938e-01 4.95173424e-01
1.78274810e-02 -7.01740906e-02 1.22566588e-01 -5.41103780e-02
-1.29801667e+00 5.12925386e-01 1.25588059e+00 -2.26104930e-01
-9.56208035e-02 4.19090949e-02 -1.00222878e-01 6.95158184e-01
4.70460445e-01 -2.36324981e-01 -8.92913222e-01 -7.49849379e-01
3.64351094e-01 6.82320595e-01 8.44227314e-01 -8.65085661e-01
5.27400613e-01 -4.86331612e-01 -2.78218001e-01 -7.86474884e-01
6.02674961e-01 -1.04727709e+00 3.75892013e-01 1.48580864e-01
-4.74488646e-01 3.73806953e-01 8.98922309e-02 9.81688142e-01
-1.66382819e-01 4.68252569e-01 4.66324449e-01 3.23324129e-02
-1.41251564e+00 7.70227492e-01 -7.26180553e-01 -1.24903068e-01
1.35929120e+00 -2.58326411e-01 -3.69106710e-01 -5.61197579e-01
-9.58829403e-01 7.02126324e-01 1.92643702e-01 7.28527129e-01
7.26277888e-01 -1.30367064e+00 -3.88040811e-01 1.81230024e-01
1.43151179e-01 3.09877172e-02 4.39083427e-01 7.70820439e-01
-5.37859201e-01 5.69074035e-01 -2.94308215e-01 -9.48944747e-01
-1.23262978e+00 6.13060594e-01 6.86722636e-01 -2.78785825e-01
-7.89795578e-01 7.42201626e-01 6.04032755e-01 -3.87364507e-01
1.61410123e-01 -6.48431778e-01 -4.11145717e-01 -1.84641078e-01
2.69851655e-01 4.86863136e-01 5.92106096e-02 -1.32142949e+00
-5.21879554e-01 5.77938139e-01 -4.12756056e-02 -9.07941908e-02
1.22536683e+00 -4.63674664e-01 4.99506816e-02 4.92213309e-01
1.27900553e+00 -5.22596180e-01 -1.85257661e+00 -2.66666353e-01
-9.01572406e-02 -4.06735867e-01 -5.47147803e-02 -2.05355465e-01
-1.30802584e+00 6.81692243e-01 2.81879395e-01 1.61606073e-01
7.39799500e-01 3.08895826e-01 6.07943416e-01 6.29861832e-01
5.84088326e-01 -8.00648153e-01 2.18797147e-01 7.50320077e-01
6.60486102e-01 -1.29425895e+00 -3.28282684e-01 -4.76663530e-01
-7.19446003e-01 8.00297320e-01 8.36753666e-01 -1.53725594e-01
7.56805182e-01 3.49222012e-02 -6.27469569e-02 -5.51614046e-01
-1.25078368e+00 -5.43767154e-01 2.60149151e-01 6.23857558e-01
-1.46775439e-01 -3.59419920e-02 4.37195033e-01 -2.08290488e-01
2.44606584e-01 -2.72171021e-01 4.17885542e-01 5.03527462e-01
-2.96428114e-01 -5.93329906e-01 1.29481524e-01 3.31506938e-01
4.33548465e-02 3.13875437e-01 -4.83505912e-02 8.78570795e-01
1.16208762e-01 1.09008324e+00 3.70907515e-01 -5.26289821e-01
4.98640865e-01 -2.50988126e-01 1.75645441e-01 -6.16233349e-01
-2.03926221e-01 -2.41215497e-01 3.55514288e-01 -1.14000034e+00
-7.63642311e-01 -8.76443505e-01 -1.22126412e+00 -2.44224489e-01
2.01606318e-01 -1.94933027e-01 9.02740136e-02 1.10585451e+00
5.01694858e-01 5.80603838e-01 5.43501735e-01 -9.67962623e-01
2.75857002e-01 -5.38303733e-01 -8.11041966e-02 6.13108277e-01
9.56368864e-01 -8.61750722e-01 6.89444989e-02 1.31423041e-01] | [6.074680805206299, 0.7894836068153381] |
b72effc7-7de9-4c08-b816-507d6351238a | adaptive-and-personalized-exercise-generation | 2306.02457 | null | https://arxiv.org/abs/2306.02457v1 | https://arxiv.org/pdf/2306.02457v1.pdf | Adaptive and Personalized Exercise Generation for Online Language Learning | Adaptive learning aims to provide customized educational activities (e.g., exercises) to address individual learning needs. However, manual construction and delivery of such activities is a laborious process. Thus, in this paper, we study a novel task of adaptive and personalized exercise generation for online language learning. To this end, we combine a knowledge tracing model that estimates each student's evolving knowledge states from their learning history and a controlled text generation model that generates exercise sentences based on the student's current estimated knowledge state and instructor requirements of desired properties (e.g., domain knowledge and difficulty). We train and evaluate our model on real-world learner interaction data from Duolingo and demonstrate that LMs guided by student states can generate superior exercises. Then, we discuss the potential use of our model in educational applications using various simulations. These simulations show that our model can adapt to students' individual abilities and can facilitate their learning efficiency by personalizing learning sequences. | ['Mrinmaya Sachan', 'Peng Cui'] | 2023-06-04 | null | null | null | null | ['knowledge-tracing'] | ['miscellaneous'] | [ 1.35548666e-01 2.75142789e-01 -2.09989905e-01 -3.82608384e-01
-4.14986044e-01 -9.13085938e-01 2.00359508e-01 3.69367301e-01
-9.32649821e-02 7.83060789e-01 9.69443992e-02 -5.28450310e-01
-2.94852138e-01 -1.02529204e+00 -4.45286453e-01 -1.02398992e-01
1.61587983e-01 2.41851315e-01 5.73780477e-01 -4.85312551e-01
4.56035107e-01 5.18359959e-01 -2.03311324e+00 2.60245651e-01
1.52264977e+00 2.24466369e-01 5.46817243e-01 1.05085886e+00
-4.83781099e-01 1.05435240e+00 -8.88618648e-01 1.43682119e-02
-3.23455036e-01 -9.66350973e-01 -9.42563534e-01 1.79155886e-01
2.82099783e-01 -2.47431517e-01 -1.20264322e-01 7.82776833e-01
5.77743351e-01 5.71213663e-01 5.32941341e-01 -1.08864224e+00
-4.69173700e-01 1.05320549e+00 1.29912481e-01 1.99871361e-01
5.57934344e-01 1.65824845e-01 3.27323765e-01 -2.76011318e-01
4.33279753e-01 6.73104227e-01 4.43703473e-01 1.00217772e+00
-9.67911720e-01 -4.12205100e-01 3.19569111e-01 2.93324023e-01
-1.18251419e+00 -1.88046843e-01 7.29002893e-01 -5.58838606e-01
5.17685950e-01 1.74836934e-01 1.00938475e+00 6.52388275e-01
2.32101306e-01 1.03137112e+00 8.49671960e-01 -7.04708874e-01
4.65389609e-01 5.14992416e-01 6.79842308e-02 8.14295053e-01
3.75313982e-02 -3.78224552e-01 -8.37792754e-01 8.27514976e-02
6.07397079e-01 -2.77692109e-01 -3.08715612e-01 -3.28661323e-01
-5.80498040e-01 5.29991686e-01 -3.00977051e-01 2.92429000e-01
-1.25580817e-01 -1.81764103e-02 1.57243442e-02 4.84945476e-01
5.91056757e-02 4.53026354e-01 -7.69054949e-01 -6.51991367e-01
-7.78951824e-01 2.38347247e-01 1.17396498e+00 1.26453781e+00
5.34187496e-01 2.14883804e-01 -5.19870579e-01 8.80887985e-01
3.72821927e-01 3.22883815e-01 9.20982301e-01 -9.29655254e-01
1.55145422e-01 8.92927527e-01 1.81973845e-01 -3.09475571e-01
-1.21534109e-01 -4.08443809e-01 -9.90175530e-02 1.65608570e-01
4.75850224e-01 -6.67953312e-01 -6.52587354e-01 1.90189970e+00
3.55088472e-01 5.75488687e-01 2.26177529e-01 1.64067969e-01
6.47259176e-01 4.78187948e-01 4.04067844e-01 -5.43146908e-01
1.06026196e+00 -7.55214274e-01 -9.11827743e-01 -6.87441900e-02
8.70572329e-01 -6.26217186e-01 1.27239811e+00 5.99484444e-01
-1.48597586e+00 -7.00802386e-01 -7.09436417e-01 4.64355558e-01
-1.15186945e-01 -1.32187620e-01 2.26148248e-01 1.11747754e+00
-8.99686992e-01 8.37795794e-01 -9.04852927e-01 -1.78292960e-01
7.37370253e-02 2.94676691e-01 5.46235144e-01 3.86639237e-01
-1.12869716e+00 4.63762492e-01 3.27368200e-01 -4.57307905e-01
-8.12497079e-01 -8.89700174e-01 -6.25874221e-01 5.52014053e-01
3.47581029e-01 -6.89989805e-01 1.86408508e+00 -9.50703442e-01
-2.48252726e+00 2.32184201e-01 1.18352979e-01 -1.47902770e-02
4.17170852e-01 -3.68047416e-01 3.38353664e-02 -1.20767869e-01
-4.58143830e-01 2.59283155e-01 2.93869108e-01 -1.07075238e+00
-8.55081439e-01 -2.49935761e-02 2.04974070e-01 6.69245005e-01
-9.98032153e-01 -4.36363012e-01 -3.40843201e-01 -2.61410028e-01
-1.90494865e-01 -7.86468983e-01 -3.65486532e-01 -4.78384584e-01
8.18177760e-02 -5.85441530e-01 5.84120810e-01 -3.00512999e-01
1.93748140e+00 -1.74339473e+00 5.78665771e-02 3.94108832e-01
-9.10374373e-02 3.59647274e-01 -1.34595186e-01 5.89991570e-01
2.50546962e-01 2.69442976e-01 2.58250922e-01 1.79206535e-01
-2.50362717e-02 4.93320040e-02 1.60531908e-01 -3.71237278e-01
-3.64324868e-01 6.07212663e-01 -1.26532793e+00 -7.00367510e-01
1.61700413e-01 2.62486577e-01 -8.26092839e-01 7.10633039e-01
-4.65882540e-01 3.68915528e-01 -8.34210753e-01 2.26818562e-01
-5.38366400e-02 -2.39632636e-01 6.23483658e-01 4.89843398e-01
-2.45167524e-01 2.78288066e-01 -1.37048209e+00 1.51894879e+00
-1.01238716e+00 4.15412515e-01 -2.70125002e-01 -5.63656271e-01
9.76251662e-01 3.79229903e-01 4.11142409e-01 -4.55313534e-01
4.94575128e-03 -1.17715865e-01 1.36968419e-01 -9.02441442e-01
6.63819849e-01 2.71495670e-01 1.65849626e-01 8.49146903e-01
9.93963331e-02 -4.66140509e-01 4.40798134e-01 3.40923905e-01
1.05080748e+00 5.81074357e-01 3.17504883e-01 -1.93834439e-01
5.96936047e-01 -2.70258754e-01 3.12618226e-01 9.78503823e-01
-2.49816164e-01 -2.22750045e-02 3.16501290e-01 6.09365590e-02
-5.83717525e-01 -8.99561644e-01 3.84926617e-01 1.55667400e+00
-1.36120439e-01 -5.38301706e-01 -1.12804914e+00 -7.56915092e-01
-3.83486718e-01 1.16072071e+00 -1.14560507e-01 -4.82797027e-01
-6.19386733e-01 -2.01785177e-01 1.34614438e-01 4.78510708e-01
3.44494700e-01 -1.25961435e+00 -8.74790311e-01 5.11236250e-01
-2.56784171e-01 -5.95109224e-01 -7.84294188e-01 -1.66972399e-01
-9.42202985e-01 -8.17335606e-01 -2.71196395e-01 -1.09439206e+00
8.42567682e-01 5.27294166e-02 1.19699156e+00 4.18952674e-01
-4.55477415e-03 1.19516790e+00 -3.93226951e-01 -5.24209619e-01
-7.64063239e-01 3.27919960e-01 8.62733424e-02 -3.68884087e-01
1.23950504e-01 -6.94358110e-01 -7.25275815e-01 2.29142025e-01
-9.83171999e-01 3.45850825e-01 2.66123801e-01 5.89845896e-01
3.67547572e-01 3.75773519e-01 9.31132436e-01 -1.11034060e+00
1.33602405e+00 -4.16432858e-01 -6.25684202e-01 7.62662530e-01
-1.03683186e+00 3.28673840e-01 7.51521647e-01 -8.14552128e-01
-1.67692161e+00 3.46109122e-01 -9.99834612e-02 -3.89916971e-02
-2.62830675e-01 7.34341085e-01 -1.01515114e-01 1.42740365e-02
8.81241560e-01 5.75643182e-01 -2.51167417e-01 -2.14894980e-01
1.83083951e-01 8.35975885e-01 2.48935863e-01 -1.11706924e+00
6.64102793e-01 -5.93829334e-01 -2.94915318e-01 -6.01579249e-01
-1.03778291e+00 -1.15605913e-01 -5.58100224e-01 -7.87976921e-01
2.74016768e-01 -5.61799407e-01 -1.26349080e+00 4.41816181e-01
-5.84085763e-01 -1.22220457e+00 -6.26691639e-01 3.94378543e-01
-6.86656892e-01 1.56935558e-01 -6.38602495e-01 -1.14534831e+00
-2.27516904e-01 -9.56626832e-01 2.31577307e-01 1.17459953e+00
-3.36868256e-01 -1.32584560e+00 2.49150142e-01 4.11755264e-01
4.31209058e-01 -2.34467894e-01 7.41147876e-01 -6.94758773e-01
-5.92869341e-01 1.11291334e-01 7.03765571e-01 1.80173606e-01
2.94834644e-01 3.29750448e-01 -6.14477694e-01 -8.17550719e-02
-2.26747081e-01 -4.27437931e-01 -6.59557432e-02 9.81004685e-02
1.52603579e+00 -5.90763390e-01 -1.18464775e-01 1.58172160e-01
1.28242922e+00 3.99225414e-01 4.05847222e-01 2.96989232e-01
1.80721015e-01 6.80278659e-01 7.53093839e-01 6.76133215e-01
4.59975004e-01 5.31241179e-01 -7.84712359e-02 8.15190256e-01
-1.58995092e-01 -5.35400927e-01 4.89647359e-01 1.23113382e+00
-2.67600179e-01 -5.62275529e-01 -1.03646123e+00 6.11635208e-01
-1.78513861e+00 -9.55419779e-01 -2.58370885e-03 2.25647688e+00
1.53796029e+00 1.97168365e-01 -5.67938611e-02 -1.65652126e-01
4.16408777e-01 -4.03015465e-01 -6.03567064e-01 -5.45912981e-01
6.83113158e-01 6.80624127e-01 1.48878694e-01 7.77312756e-01
-8.50417763e-02 9.89144564e-01 6.07950783e+00 8.55849922e-01
-7.23342180e-01 -1.74394622e-01 4.39935416e-01 -8.75107273e-02
-5.68927348e-01 -3.08434844e-01 -1.05568016e+00 4.41827923e-01
1.42526984e+00 -9.68561888e-01 5.10268211e-01 7.99680591e-01
4.42794651e-01 -1.18072085e-01 -9.85817969e-01 3.41442347e-01
-5.86792380e-02 -1.21419811e+00 -8.01420882e-02 -3.14729750e-01
1.05894732e+00 -7.84243107e-01 -4.58990186e-02 7.44621217e-01
8.51275742e-01 -5.68993390e-01 4.38308448e-01 7.05807328e-01
4.91636455e-01 -1.03918338e+00 3.28214049e-01 8.08906615e-01
-9.96259332e-01 -1.29790783e-01 1.80155247e-01 -1.75961033e-02
-5.91793545e-02 -1.16116822e-01 -1.24972713e+00 2.26688921e-01
3.82390589e-01 9.83193591e-02 -6.21621072e-01 1.03569877e+00
-5.78249454e-01 1.10168171e+00 -9.05991420e-02 -5.86972594e-01
-3.68443310e-01 -8.71072151e-03 4.11133803e-02 9.85053480e-01
7.66590714e-01 6.99341714e-01 3.08693320e-01 6.07583404e-01
2.16112053e-03 3.62443835e-01 -2.00444877e-01 -1.24731302e-01
9.14170742e-01 1.16173637e+00 -6.10408843e-01 -4.06950176e-01
-9.82162803e-02 8.73558760e-01 3.13944459e-01 3.81904602e-01
-6.26260698e-01 -4.91033971e-01 5.70217788e-01 2.70674407e-01
1.42581630e-02 -3.58872533e-01 -2.82978386e-01 -9.51672435e-01
-5.77374041e-01 -8.23595345e-01 2.29751915e-01 -8.26284885e-01
-7.14915812e-01 2.34935254e-01 1.72696933e-02 -1.26304734e+00
-5.68055153e-01 -2.44672045e-01 -9.31799114e-01 6.81441724e-01
-1.34383500e+00 -2.59721667e-01 -7.68789530e-01 5.34519732e-01
8.47718835e-01 -1.31472766e-01 6.49854243e-01 -2.04417505e-04
-6.07859850e-01 8.69669974e-01 -6.42274600e-03 -2.78267533e-01
5.90016365e-01 -1.35704994e+00 1.50364824e-02 5.98227441e-01
-5.86665682e-02 5.33205926e-01 9.77605283e-01 -8.08216214e-01
-1.37019253e+00 -9.42428827e-01 7.74367332e-01 -4.03943539e-01
5.46149135e-01 1.71517298e-01 -1.04663467e+00 6.35291517e-01
2.96110511e-01 -4.52964544e-01 1.12791288e+00 -5.56131676e-02
4.18509185e-01 1.02797322e-01 -1.06407356e+00 1.02864277e+00
1.23375535e+00 -4.06035967e-02 -4.17117298e-01 3.91839176e-01
5.54014802e-01 -9.05948460e-01 -1.19149935e+00 -5.95276617e-03
4.52867240e-01 -7.53831983e-01 6.63797021e-01 -7.37309873e-01
3.17284465e-01 -4.41787182e-04 5.32208085e-01 -1.57502019e+00
-2.22554505e-01 -7.73690343e-01 -3.41260970e-01 1.46601605e+00
3.04350048e-01 -2.32478052e-01 1.19940341e+00 1.08229613e+00
-3.13880891e-01 -7.25430369e-01 -1.84287488e-01 -6.44763649e-01
-5.83843142e-02 -2.35267878e-01 6.13719523e-01 8.67850363e-01
4.49123710e-01 2.51196593e-01 -1.17109027e-02 1.41264111e-01
5.29335916e-01 1.16564475e-01 7.31889486e-01 -1.22693050e+00
-4.38275009e-01 -3.54463428e-01 2.06185013e-01 -1.31386018e+00
1.51673585e-01 -5.20073354e-01 2.18625695e-01 -1.37064600e+00
3.88682857e-02 -4.06852931e-01 -1.89473882e-01 5.11696398e-01
-4.84590232e-01 -5.55618942e-01 4.24165539e-02 -3.75704199e-01
-8.31141889e-01 3.57232630e-01 1.58085203e+00 2.58064359e-01
-1.11462712e+00 4.17641789e-01 -7.48446345e-01 6.71010852e-01
1.14471149e+00 -3.17741126e-01 -1.00047052e+00 -4.95295413e-02
4.27721679e-01 4.37469274e-01 -5.02596736e-01 -1.05887294e+00
6.68283224e-01 -9.55971777e-01 2.00990081e-01 -1.22435264e-01
-2.65318125e-01 -5.97501874e-01 -1.74728230e-01 5.97166479e-01
-7.52431154e-01 -1.92348674e-01 3.68991226e-01 4.43590462e-01
7.96355903e-02 -7.71806836e-01 5.22053182e-01 -8.52002874e-02
-4.81888413e-01 1.08279191e-01 -9.35929239e-01 5.42740673e-02
1.23912907e+00 -3.02963167e-01 -2.50896543e-01 -6.73238218e-01
-9.48003054e-01 6.62185490e-01 3.64993773e-02 3.48496914e-01
7.72231579e-01 -1.06856441e+00 -5.48733175e-01 9.26555917e-02
2.11371537e-02 3.78869697e-02 4.31972653e-01 2.48967141e-01
-4.61167634e-01 1.34135664e-01 -1.76429033e-01 -3.41441840e-01
-1.76427054e+00 2.32094437e-01 4.16691899e-01 -4.84644860e-01
-8.56608972e-02 6.99765801e-01 -3.23182106e-01 -5.52943647e-01
3.49839032e-01 -2.63532937e-01 -7.70701468e-01 -6.39907867e-02
6.67504549e-01 3.28125745e-01 -1.59289822e-01 3.57194871e-01
3.20837110e-01 1.83191553e-01 -6.50223996e-03 -2.30509345e-03
1.12910557e+00 -4.28926855e-01 5.93202412e-01 4.34403837e-01
2.60634780e-01 3.63684118e-01 -1.40885520e+00 -2.48195484e-01
1.60779059e-01 -2.03608111e-01 -2.32395604e-01 -1.02832258e+00
-7.59008586e-01 4.58655089e-01 6.35290623e-01 2.07262486e-01
1.14903450e+00 -2.07772672e-01 4.93509978e-01 6.35453761e-01
2.95995981e-01 -1.43526959e+00 5.77003956e-01 6.49339795e-01
4.11956549e-01 -7.51040578e-01 -2.75077462e-01 -3.50706995e-01
-4.72853303e-01 1.35671616e+00 1.40188730e+00 5.09574831e-01
4.66315687e-01 1.56529009e-01 1.05797917e-01 1.76455587e-01
-1.23739076e+00 -5.41942660e-04 2.06462264e-01 4.44867790e-01
9.99389410e-01 3.08751944e-03 -4.91788089e-01 6.58749044e-01
-4.16572660e-01 4.22515392e-01 1.16123176e+00 1.36824644e+00
-9.31249261e-01 -1.45083475e+00 -2.35399500e-01 4.06533241e-01
-2.11437061e-01 2.45243199e-02 -1.97551653e-01 4.31212008e-01
-5.56802079e-02 9.14750099e-01 -2.14411825e-01 -2.93318242e-01
5.43543339e-01 4.88040656e-01 7.73051083e-01 -1.18342245e+00
-1.00745332e+00 -3.59998316e-01 -1.67138800e-02 -7.27734342e-02
-1.46878973e-01 -7.14389443e-01 -1.59350133e+00 -1.82262540e-01
-3.50410074e-01 5.73467791e-01 6.69373989e-01 8.20044219e-01
2.37948120e-01 1.10816109e+00 5.55432975e-01 -2.44296983e-01
-8.20004821e-01 -8.29580069e-01 -3.86934310e-01 1.88273102e-01
-1.65652603e-01 -2.10228905e-01 -1.69307694e-01 5.63052714e-01] | [10.168664932250977, 7.192203521728516] |
7e311d68-c8a9-48a6-8197-b02d378616a9 | l2-constrained-remnet-for-camera-model | 2009.05379 | null | https://arxiv.org/abs/2009.05379v2 | https://arxiv.org/pdf/2009.05379v2.pdf | L2-Constrained RemNet for Camera Model Identification and Image Manipulation Detection | Source camera model identification (CMI) and image manipulation detection are of paramount importance in image forensics. In this paper, we propose an L2-constrained Remnant Convolutional Neural Network (L2-constrained RemNet) for performing these two crucial tasks. The proposed network architecture consists of a dynamic preprocessor block and a classification block. An L2 loss is applied to the output of the preprocessor block, and categorical crossentropy loss is calculated based on the output of the classification block. The whole network is trained in an end-to-end manner by minimizing the total loss, which is a combination of the L2 loss and the categorical crossentropy loss. Aided by the L2 loss, the data-adaptive preprocessor learns to suppress the unnecessary image contents and assists the classification block in extracting robust image forensics features. We train and test the network on the Dresden database and achieve an overall accuracy of 98.15%, where all the test images are from devices and scenes not used during training to replicate practical applications. The network also outperforms other state-of-the-art CNNs even when the images are manipulated. Furthermore, we attain an overall accuracy of 99.68% in image manipulation detection, which implies that it can be used as a general-purpose network for image forensic tasks. | ['Md. Kamrul Hasan', 'Jonathan Wu', 'Abdul Muntakim Rafi'] | 2020-09-10 | null | null | null | null | ['image-manipulation-detection', 'image-forensics'] | ['computer-vision', 'computer-vision'] | [ 4.63355333e-01 -1.76437899e-01 6.34798855e-02 -8.84637982e-02
-5.39923072e-01 -2.41229445e-01 3.00649881e-01 4.84145880e-02
-8.58092666e-01 1.36841461e-01 -5.72432935e-01 -2.79983759e-01
-1.03922170e-02 -6.22360885e-01 -9.08452690e-01 -9.07251835e-01
-1.60683692e-01 -8.76811668e-02 9.01599228e-02 4.53721255e-01
4.60493803e-01 8.88841212e-01 -1.53719807e+00 3.27307761e-01
7.20565557e-01 1.44870985e+00 1.85727850e-01 6.48485780e-01
2.37135217e-01 1.09778488e+00 -7.74444699e-01 -7.47936130e-01
4.03083652e-01 4.08086553e-03 -2.18481094e-01 5.64865768e-02
6.31702065e-01 -6.22029424e-01 -5.55043638e-01 1.32359684e+00
7.02517211e-01 8.83493014e-03 4.80026811e-01 -1.07173848e+00
-5.51831186e-01 1.86321378e-01 -7.86463261e-01 4.08170372e-01
-2.80474097e-01 3.70021135e-01 4.65372026e-01 -8.49939466e-01
3.74282956e-01 1.22853708e+00 6.03231668e-01 3.19324225e-01
-9.86456454e-01 -1.06559467e+00 -2.95646220e-01 4.74817842e-01
-1.20637751e+00 -5.63596189e-01 1.06211710e+00 -3.34088594e-01
7.22815454e-01 -1.34778827e-01 1.01173989e-01 8.69571567e-01
3.96380603e-01 8.59674931e-01 6.90753043e-01 -4.14205700e-01
2.11875588e-02 2.96429545e-01 1.88294753e-01 6.00949705e-01
4.40452963e-01 2.76826769e-01 -5.06840229e-01 1.54609635e-01
5.50281703e-01 2.07668711e-02 -2.37420425e-01 -8.03660303e-02
-6.83134675e-01 7.45394349e-01 4.18506563e-01 1.16646945e-01
-4.85273361e-01 2.29674146e-01 6.88238382e-01 2.18619600e-01
3.93576890e-01 2.05751136e-01 -4.63509113e-02 1.21332109e-01
-1.18746233e+00 7.47665465e-02 3.00038368e-01 7.06707776e-01
4.74830002e-01 2.86611617e-01 -1.72401458e-01 8.66077662e-01
-4.17865664e-02 4.73493338e-01 2.81807274e-01 -8.86153400e-01
6.95417523e-01 6.17952287e-01 -1.91333935e-01 -1.20331776e+00
-2.24371310e-02 -4.07331795e-01 -9.80587900e-01 5.89240491e-01
4.34613764e-01 7.66702220e-02 -6.41502678e-01 1.44366121e+00
1.60215050e-01 2.27165837e-02 -9.20987725e-02 6.89689398e-01
5.28016567e-01 3.95906448e-01 2.28233054e-01 -6.78283572e-02
1.30511582e+00 -6.70519471e-01 -6.19085073e-01 -1.33257076e-01
2.34028146e-01 -7.42119372e-01 8.09945524e-01 8.32982361e-01
-1.23735046e+00 -7.65765667e-01 -1.24343693e+00 -2.21523106e-01
-3.18878502e-01 7.26620674e-01 1.75387576e-01 8.18741500e-01
-7.83364594e-01 7.72471249e-01 -6.39276206e-01 1.51311591e-01
9.31958377e-01 4.98959482e-01 -3.87959301e-01 1.12062329e-02
-8.47118914e-01 5.87032199e-01 6.09671295e-01 3.32273513e-01
-8.52983713e-01 -7.01768160e-01 -8.22287560e-01 2.67110944e-01
1.48048878e-01 5.75112738e-02 6.27749920e-01 -8.96192312e-01
-1.18575513e+00 1.01449585e+00 3.02335352e-01 -7.37618566e-01
7.76100874e-01 -1.57493427e-01 -3.36481035e-01 4.93090004e-01
3.62876803e-02 4.48888093e-01 1.35985088e+00 -1.02093792e+00
-5.51413059e-01 -4.42531049e-01 -2.88623184e-01 -2.74620891e-01
-7.46762097e-01 4.02761877e-01 -7.44503915e-01 -8.17109048e-01
-4.94577795e-01 -6.82565689e-01 4.26738083e-01 3.10152233e-01
-5.81130624e-01 6.97083622e-02 1.24632394e+00 -1.14071631e+00
1.18211985e+00 -2.49084020e+00 -3.69490504e-01 -4.81455363e-02
9.27220955e-02 7.89143503e-01 -2.46825606e-01 -1.63480952e-01
-4.40825522e-01 -1.63244992e-01 -3.49504143e-01 -7.06413686e-01
-2.54330754e-01 -4.93825227e-01 -3.33287179e-01 1.02179980e+00
5.88692844e-01 6.53190553e-01 -5.78430474e-01 -4.07838821e-01
5.94893098e-01 6.40821934e-01 -2.84848392e-01 3.39495778e-01
1.13848835e-01 1.86608374e-01 -6.65949956e-02 9.07362640e-01
1.14185047e+00 2.46385694e-01 -8.98000374e-02 -3.69150788e-01
4.09237146e-02 -2.90324599e-01 -9.96250153e-01 1.28365672e+00
-4.60389644e-01 1.12880218e+00 3.31785470e-01 -1.08319294e+00
8.24873924e-01 -4.84366529e-02 3.10529381e-01 -8.90564322e-01
4.47636575e-01 6.87088221e-02 -4.08348404e-02 -6.75902247e-01
4.70082492e-01 1.97263837e-01 2.48693358e-02 3.92472774e-01
8.56169537e-02 5.06818116e-01 2.23646313e-01 1.36551768e-01
9.58032727e-01 -1.25692517e-01 -2.24735200e-01 -1.01089031e-01
7.70704448e-01 -3.48210067e-01 4.24992114e-01 5.95499218e-01
-3.44387650e-01 4.34475213e-01 5.83357513e-01 -4.85542566e-01
-1.27275765e+00 -7.87456691e-01 -1.35353789e-01 6.62187576e-01
2.53555477e-01 8.10294226e-02 -9.40552235e-01 -8.08347225e-01
-1.26054930e-02 8.26756299e-01 -5.99613428e-01 -4.22123373e-01
-9.12794769e-01 -6.62068665e-01 8.80876660e-01 5.76684833e-01
8.72344077e-01 -1.30195606e+00 -7.36646891e-01 5.84174320e-02
3.61011922e-02 -1.24923217e+00 -5.54836929e-01 1.32189021e-01
-6.46783054e-01 -1.38663721e+00 -3.04941118e-01 -5.33575535e-01
6.42004132e-01 3.79678607e-01 6.88628376e-01 2.50129372e-01
-6.26933932e-01 5.98606095e-02 1.85067728e-02 -4.33012694e-01
-4.68198240e-01 -2.35147074e-01 -1.67342514e-01 4.30055082e-01
3.56638193e-01 -3.73630494e-01 -6.38503730e-01 1.30244344e-01
-1.26379192e+00 -3.91922623e-01 6.10320568e-01 9.48058188e-01
4.73751962e-01 5.83278120e-01 2.35606730e-01 -6.21865928e-01
4.05967981e-01 -3.05836737e-01 -8.84206891e-01 2.15727702e-01
-3.65893483e-01 -1.22349165e-01 8.09917092e-01 -7.57574499e-01
-1.01381063e+00 2.28014868e-02 9.17754397e-02 -9.93671656e-01
4.04268280e-02 3.91477011e-02 -5.57155013e-01 -2.71329373e-01
2.80365467e-01 2.73807913e-01 1.54530674e-01 -4.66691315e-01
3.93154174e-02 8.50683630e-01 8.36937606e-01 -3.03571641e-01
8.34003747e-01 4.68155593e-01 2.16070756e-01 -1.07224381e+00
-3.97435397e-01 -3.70767146e-01 -3.73483479e-01 -4.01900560e-01
9.67744470e-01 -8.11454296e-01 -1.06951523e+00 1.08881629e+00
-1.31311095e+00 -1.36415735e-01 -1.39859170e-01 2.87217170e-01
-2.45006621e-01 6.11692905e-01 -3.99035245e-01 -1.15056026e+00
-6.97963357e-01 -1.39704108e+00 1.07183051e+00 2.56161273e-01
3.48140359e-01 -6.60542250e-01 -7.43689954e-01 4.69453573e-01
2.41682395e-01 4.09288466e-01 9.16696548e-01 -5.98987460e-01
-5.46561599e-01 -5.85965216e-01 -6.66302025e-01 1.00273108e+00
-2.88819999e-01 -4.85439487e-02 -1.30346894e+00 -5.94799757e-01
2.83986092e-01 -4.03703481e-01 1.09341276e+00 3.18059862e-01
1.65474927e+00 -1.44340843e-01 4.65379395e-02 9.20125961e-01
1.52056432e+00 3.70350271e-01 1.07151031e+00 6.44252241e-01
7.42794037e-01 5.83094597e-01 3.59584093e-01 2.10492745e-01
-1.00305334e-01 5.48236430e-01 7.81348586e-01 4.11629602e-02
-3.98859739e-01 7.28316382e-02 5.62529504e-01 1.46264613e-01
3.70526493e-01 -3.40571165e-01 -7.55982041e-01 4.46122617e-01
-1.44969249e+00 -1.18314111e+00 -2.17901960e-01 2.48196435e+00
4.52727556e-01 2.13759661e-01 -1.13091074e-01 3.25553745e-01
1.11394083e+00 2.20173150e-01 -8.18701208e-01 -2.33812749e-01
-9.12864059e-02 3.00104588e-01 7.35153675e-01 1.56244382e-01
-1.45892489e+00 7.48558521e-01 4.79095268e+00 1.40834904e+00
-1.31068218e+00 3.37603427e-02 8.71576011e-01 -1.37623146e-01
5.38590908e-01 -3.37462306e-01 -7.28039920e-01 9.11032200e-01
1.03692901e+00 4.04911280e-01 3.76635820e-01 9.46920395e-01
2.01368913e-01 -3.38640064e-01 -8.23911428e-01 1.13656557e+00
2.56685078e-01 -1.15509319e+00 -1.10158108e-01 2.69442737e-01
1.58207834e-01 -2.40787894e-01 4.93955493e-01 1.65704608e-01
-3.28646481e-01 -8.93015146e-01 9.12394762e-01 2.46848047e-01
1.17954051e+00 -1.22865021e+00 6.58707321e-01 3.41902047e-01
-9.21855927e-01 -5.20603180e-01 -4.33276683e-01 4.77606505e-01
8.54448043e-03 8.35991800e-01 -4.69179004e-01 4.02801067e-01
7.24806488e-01 3.82659525e-01 -5.54965913e-01 9.75758493e-01
-2.35412151e-01 5.47351182e-01 -1.55755267e-01 4.57635492e-01
9.15177837e-02 -1.60747692e-01 6.59420252e-01 1.25104451e+00
2.19569817e-01 -3.78150135e-01 -3.44435930e-01 9.58471775e-01
-5.22997558e-01 -2.38945961e-01 -3.81901324e-01 -8.24615508e-02
4.59719718e-01 1.24193251e+00 -7.04526544e-01 -2.22452760e-01
-1.17990077e-01 1.13398492e+00 3.34149003e-01 7.57550523e-02
-9.43550587e-01 -8.29557240e-01 5.18519580e-01 -6.61224052e-02
6.65824533e-01 1.50919454e-02 -2.06384018e-01 -1.08099771e+00
3.64744127e-01 -7.76768148e-01 2.50042081e-01 -5.53149700e-01
-1.06514764e+00 6.29780591e-01 -1.19830161e-01 -1.30430925e+00
5.12275137e-02 -9.08833981e-01 -8.71795118e-01 8.45516205e-01
-1.79346669e+00 -1.13785017e+00 -3.88198942e-01 5.29980361e-01
5.08406579e-01 -5.28770208e-01 3.18458617e-01 6.72424734e-01
-1.05342197e+00 1.01738715e+00 2.42947742e-01 5.40528774e-01
6.46230817e-01 -9.20634687e-01 2.80202180e-01 1.23393595e+00
-3.36665452e-01 7.03424096e-01 3.35095435e-01 -6.44177914e-01
-1.34870660e+00 -1.49569011e+00 5.37932277e-01 -1.31502107e-01
3.96204829e-01 -4.12949771e-01 -9.09220994e-01 2.01470718e-01
1.84872344e-01 1.40844882e-01 4.19263154e-01 -6.68375492e-01
-5.92199445e-01 -2.15735868e-01 -1.52650142e+00 3.63160521e-02
4.91737485e-01 -5.21169007e-01 -1.23335585e-01 1.88528657e-01
4.15318668e-01 -1.59042135e-01 -6.06078804e-01 2.12619856e-01
5.96932471e-01 -1.10116267e+00 1.13303661e+00 -1.69901997e-01
5.05104542e-01 -7.89189935e-02 4.20113839e-02 -6.47501588e-01
9.82988104e-02 -5.35236657e-01 -3.79858464e-01 1.35906124e+00
-1.75467700e-01 -3.46402735e-01 7.01278329e-01 4.45839614e-01
4.61034440e-02 -6.56895041e-01 -9.64086294e-01 -9.04297590e-01
-3.48305069e-02 -5.88432431e-01 3.44803810e-01 5.57400584e-01
-7.40569592e-01 -1.22399248e-01 -4.77515280e-01 2.13087320e-01
1.26898241e+00 -1.99344888e-01 6.33848131e-01 -8.82812023e-01
-2.25406364e-01 -5.29168606e-01 -5.30443966e-01 -7.53150403e-01
3.37723076e-01 -5.90437293e-01 -7.98013359e-02 -5.81655502e-01
3.37254643e-01 -1.88748717e-01 -1.88926503e-01 3.03259134e-01
-1.94952458e-01 4.29335088e-01 3.73280197e-01 3.44653547e-01
-2.85763472e-01 3.96519750e-01 7.90730655e-01 -4.42215323e-01
1.65226385e-02 -2.46880800e-02 -5.11870086e-01 6.10947728e-01
7.58397698e-01 -5.95506907e-01 -2.04224974e-01 -4.67782736e-01
-3.78576964e-01 -2.48601019e-01 6.90751791e-01 -1.15780485e+00
2.84201384e-01 2.62079000e-01 6.84026241e-01 -8.66072476e-01
2.91751176e-01 -8.64336014e-01 -6.74995705e-02 4.31543410e-01
-1.57720655e-01 -1.89934507e-01 2.81960756e-01 5.15215755e-01
-1.75566003e-01 -3.96423519e-01 1.34221756e+00 1.30308717e-01
-4.49121237e-01 3.68452638e-01 -1.72821522e-01 -3.10846150e-01
1.11235523e+00 -4.69097406e-01 -3.49594384e-01 1.44435987e-02
-2.33918160e-01 -9.09698755e-02 4.96609628e-01 1.29561171e-01
7.98238575e-01 -1.12772822e+00 -6.39339089e-01 5.09299338e-01
-6.98664039e-02 -7.55907670e-02 4.71114486e-01 7.27991521e-01
-7.63447285e-01 1.78725928e-01 -3.50564331e-01 -4.38976794e-01
-1.32138741e+00 7.64566243e-01 2.48122573e-01 -3.35757613e-01
-5.42491257e-01 8.12347770e-01 5.13839014e-02 -7.18777105e-02
5.63413262e-01 1.47305489e-01 -4.30425182e-02 1.83239579e-01
8.91362667e-01 7.02840924e-01 2.50548095e-01 -8.40223312e-01
-3.26777220e-01 2.73184210e-01 -3.09559703e-01 2.18366563e-01
1.57420337e+00 2.10661948e-01 -1.38684124e-01 -3.30251157e-01
1.65580642e+00 -2.61750549e-01 -1.37207639e+00 -6.36477321e-02
-2.20763683e-01 -7.19667912e-01 5.40615320e-01 -5.15635669e-01
-1.51801717e+00 1.18881679e+00 8.98880482e-01 -6.97463527e-02
1.53614438e+00 -5.16530335e-01 9.77017820e-01 1.17260285e-01
-1.80744771e-02 -1.06727064e+00 3.14660072e-01 1.93591565e-01
7.39582539e-01 -1.04692912e+00 8.60078037e-02 3.97722796e-02
-5.27906418e-01 1.36967719e+00 4.75715190e-01 -3.08057129e-01
3.33482295e-01 2.94866741e-01 -2.82900810e-01 -1.83141261e-01
-2.33568504e-01 1.87412873e-01 1.01313919e-01 6.95890069e-01
-3.94004211e-02 -3.50514680e-01 -3.69498394e-02 4.35795844e-01
1.62002653e-01 -2.31382221e-01 3.49098086e-01 7.69448996e-01
-1.15262315e-01 -8.05659890e-01 -5.79554677e-01 3.86670440e-01
-7.64956117e-01 1.00255562e-02 -1.74343884e-01 7.82533526e-01
5.32377124e-01 9.62799728e-01 1.39640749e-01 -4.78799194e-01
3.05743963e-01 -3.01093042e-01 2.67216027e-01 6.39561331e-04
-7.55165339e-01 -1.76788211e-01 -2.50889570e-01 -7.07867742e-01
-2.43220210e-01 -7.86465049e-01 -6.87513351e-01 -4.89805549e-01
-7.29792774e-01 -2.85866052e-01 9.90867138e-01 6.16382420e-01
2.50866234e-01 4.22219038e-01 7.74603426e-01 -1.11495495e+00
-8.08289945e-01 -7.88055837e-01 -5.66579819e-01 4.88301456e-01
5.42728186e-01 -6.46762311e-01 -3.42103958e-01 2.24061757e-01] | [12.403790473937988, 0.9979224801063538] |
6675b952-adf9-4c74-92c9-6560c4835ca4 | physically-guided-disentangled-implicit | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhang_Physically-Guided_Disentangled_Implicit_Rendering_for_3D_Face_Modeling_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhang_Physically-Guided_Disentangled_Implicit_Rendering_for_3D_Face_Modeling_CVPR_2022_paper.pdf | Physically-Guided Disentangled Implicit Rendering for 3D Face Modeling | This paper presents a novel Physically-guided Disentangled Implicit Rendering (PhyDIR) framework for high-fidelity 3D face modeling. The motivation comes from two observations: widely-used graphics renderers yield excessive approximations against photo-realistic imaging, while neural rendering methods are highly entangled to perceive 3D-aware operations. Hence, we learn to disentangle the implicit rendering via explicit physical guidance, meanwhile guarantee the properties of (1) 3D-aware comprehension and (2) high-reality imaging. For the former one, PhyDIR explicitly adopts 3D shading and rasterizing modules to control the renderer, which disentangles the lighting, facial shape and view point from neural reasoning. Specifically, PhyDIR proposes a novel multi-image shading strategy to compensate the monocular limitation, so that the lighting variations are accessible to the neural renderer. For the latter one, PhyDIR learns the face-collection implicit texture to avoid ill-posed intrinsic factorization, then leverages a series of consistency losses to constrain the robustness. With the disentangled method, we make 3D face modeling benefit from both kinds of rendering strategies. Extensive experiments on benchmarks show that PhyDIR obtains superior performance than state-of-the-art explicit/implicit methods, on both geometry/texture modeling. | ['Dongjin Huang', 'Zhifeng Xie', 'Chengjie Wang', 'Xiaoming Huang', 'Hao Tang', 'Kunlin Liu', 'Renwang Chen', 'Weijian Cao', 'Ying Tai', 'Yanhao Ge', 'Zhenyu Zhang'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['3d-face-modeling'] | ['computer-vision'] | [-3.78959402e-02 2.35046551e-01 6.43973723e-02 -2.55989224e-01
-4.65991527e-01 -4.63986397e-01 7.05686033e-01 -9.43783045e-01
3.17023695e-01 3.87345701e-01 2.36990675e-01 -2.18678609e-01
-1.06349535e-01 -8.27507913e-01 -7.95627058e-01 -1.08524346e+00
2.52760381e-01 3.63988757e-01 -4.55858707e-01 -1.81475013e-01
-5.06037213e-02 5.51378191e-01 -1.58868253e+00 1.90627217e-01
8.87832463e-01 1.04013371e+00 -1.47814125e-01 5.55359304e-01
-4.68032323e-02 6.36123300e-01 -2.01739799e-02 -4.81122851e-01
5.40787935e-01 -2.25160152e-01 -2.34594449e-01 3.34399760e-01
8.27566028e-01 -8.78708541e-01 -7.33797848e-01 8.92169833e-01
4.24225390e-01 -8.03559124e-02 4.84255493e-01 -1.20545912e+00
-1.12571979e+00 -3.44542949e-03 -8.18179429e-01 -3.40775281e-01
3.41859907e-01 4.73033607e-01 7.49794245e-01 -1.09842229e+00
3.61310661e-01 1.60555947e+00 2.70634115e-01 7.07866490e-01
-1.38883221e+00 -9.21822190e-01 3.97269219e-01 -3.19465101e-01
-1.39349437e+00 -7.65999794e-01 1.00779474e+00 -4.89755541e-01
4.75015253e-01 6.85398877e-01 7.27349877e-01 1.16836214e+00
4.62883115e-02 4.50993270e-01 1.47805500e+00 -2.85879467e-02
9.48145911e-02 1.57363750e-02 -7.69095169e-03 1.21276462e+00
2.83851296e-01 5.09504616e-01 -9.80320454e-01 -4.26665097e-01
1.42389035e+00 -5.27173989e-02 -5.10074735e-01 -5.37849665e-01
-1.03750443e+00 4.42609310e-01 2.98442483e-01 -4.91670966e-01
-2.09078386e-01 7.69002810e-02 1.31659225e-01 9.83415917e-02
7.73605883e-01 2.35314772e-01 -3.32234472e-01 1.88135147e-01
-6.81033731e-01 4.08234864e-01 7.24772334e-01 8.77028883e-01
8.00175786e-01 2.03656271e-01 -1.27599478e-01 5.28454304e-01
6.99388623e-01 7.79229045e-01 -2.36150533e-01 -1.20073760e+00
4.17691320e-01 4.55689490e-01 4.54343520e-02 -1.24782169e+00
3.12928967e-02 -2.94079751e-01 -1.15242851e+00 5.45838714e-01
1.43443927e-01 2.55615741e-01 -8.93686831e-01 1.87955689e+00
6.07308030e-01 6.18447721e-01 -2.09383562e-01 1.32718813e+00
8.86474133e-01 4.25715864e-01 -2.17056885e-01 -2.00735241e-01
1.41629410e+00 -7.98571944e-01 -6.06509089e-01 4.77607623e-02
8.61887634e-02 -6.53671384e-01 1.13630641e+00 5.29476821e-01
-1.34644413e+00 -3.40169400e-01 -1.03549302e+00 -5.23435891e-01
3.14139724e-01 1.37093529e-01 1.07298040e+00 7.55401611e-01
-1.03868377e+00 2.73682773e-01 -9.94545341e-01 3.05939227e-01
5.65018177e-01 2.82945752e-01 -4.90813851e-01 -3.02414119e-01
-8.05487633e-01 3.36415619e-01 -3.03691357e-01 4.12177503e-01
-9.05053914e-01 -9.97872114e-01 -7.95549273e-01 1.86638925e-02
3.85108262e-01 -1.39384615e+00 7.58614421e-01 -7.03083575e-01
-1.96452439e+00 1.10799551e+00 -1.72263861e-01 3.62967163e-01
7.14683473e-01 -2.95678288e-01 -6.74465299e-02 1.90326795e-01
-3.06904614e-01 3.58012497e-01 1.12960505e+00 -1.63760459e+00
2.93917298e-01 -6.77883685e-01 3.85509342e-01 4.62877750e-01
-1.45033807e-01 -3.93405110e-01 -6.54241204e-01 -5.44085801e-01
6.47137105e-01 -8.40745568e-01 6.82037324e-02 6.41011596e-01
-5.57651043e-01 3.41295898e-01 7.15246856e-01 -6.95266306e-01
5.53790510e-01 -2.14460278e+00 4.90940392e-01 2.08723266e-02
9.57879245e-01 -5.40144108e-02 -2.12897077e-01 -9.22978111e-03
-3.03640842e-01 1.90649956e-01 1.71196431e-01 -1.02032197e+00
1.80324391e-01 4.00125235e-01 -6.78201914e-01 7.08250701e-01
8.98317322e-02 9.59528685e-01 -6.97668195e-01 -2.64977634e-01
2.93016911e-01 8.77976656e-01 -1.10149550e+00 4.78478462e-01
-1.06562093e-01 9.95824218e-01 -7.29884684e-01 4.55413759e-01
1.37055898e+00 -3.33632678e-01 7.42368475e-02 -5.51450849e-01
-2.64798068e-02 4.48969185e-01 -8.97738516e-01 1.99775445e+00
-5.22574604e-01 1.33969232e-01 4.93824422e-01 -3.22664320e-01
6.53992891e-01 3.11536252e-01 3.14254284e-01 -7.69522130e-01
8.62442926e-02 -4.56178598e-02 -4.48072195e-01 -4.31006670e-01
3.77143413e-01 -2.20242441e-01 3.99322242e-01 3.85711163e-01
-2.78233558e-01 -4.32926387e-01 -7.54672229e-01 3.62900347e-01
7.82583475e-01 5.20280004e-01 -2.62620300e-01 -4.20472980e-01
3.42027515e-01 -7.94433117e-01 6.11367822e-01 3.37068975e-01
8.15823451e-02 7.93276191e-01 6.46559119e-01 -4.73448128e-01
-7.88466334e-01 -1.46804214e+00 5.99495545e-02 7.01964617e-01
3.30695599e-01 -3.39092880e-01 -7.02578485e-01 -2.08355755e-01
9.96854082e-02 5.16168416e-01 -7.59836078e-01 -1.60716593e-01
-5.71793497e-01 -4.97555822e-01 4.66902047e-01 3.46309319e-02
5.62861800e-01 -2.22528994e-01 -3.03421527e-01 -5.43209314e-01
-7.85794389e-03 -1.03417552e+00 -4.32755351e-01 -4.70223308e-01
-8.03097248e-01 -1.03155839e+00 -4.44991499e-01 -6.23049848e-02
9.44169939e-01 7.56696582e-01 1.14700913e+00 5.27077496e-01
-2.85841394e-02 3.89316916e-01 1.54823646e-01 2.14242801e-01
-5.80987660e-03 -4.34624970e-01 2.60624081e-01 2.58311778e-01
-3.69164854e-01 -1.11059415e+00 -8.46034765e-01 3.58019739e-01
-8.10965717e-01 8.82415175e-01 2.24712282e-01 8.66437197e-01
6.13145471e-01 -3.75351638e-01 -1.78701341e-01 -9.48310494e-01
2.60090292e-01 -3.98688734e-01 -7.15629339e-01 2.66946137e-01
-3.76955897e-01 1.52668282e-01 3.68670553e-01 -5.59354007e-01
-1.43018162e+00 -2.98521280e-01 -8.68344530e-02 -9.50683594e-01
9.37648956e-03 1.51516090e-03 -8.48932087e-01 -4.66806114e-01
3.18166763e-01 1.99144199e-01 -1.01430319e-01 -4.94233757e-01
5.96260369e-01 1.15517944e-01 4.74526793e-01 -1.20951760e+00
1.05463433e+00 8.84271920e-01 2.01163009e-01 -6.42252982e-01
-8.64550650e-01 3.16137344e-01 -3.63286018e-01 -1.78540826e-01
8.43458414e-01 -1.14293444e+00 -1.20297158e+00 6.77221954e-01
-1.27713656e+00 -3.03857952e-01 -8.96351039e-02 2.22758278e-01
-6.18318379e-01 3.70079041e-01 -6.62815928e-01 -9.20009136e-01
-3.70473787e-02 -1.43198287e+00 1.63783216e+00 7.58291185e-02
1.53802946e-01 -6.79490149e-01 -2.10370377e-01 6.75951838e-01
1.99036077e-01 3.16988945e-01 1.07334018e+00 3.51860672e-01
-1.38750696e+00 3.35161507e-01 -5.74169338e-01 2.07592715e-02
1.16956964e-01 -2.97944546e-02 -1.50119936e+00 -1.87221318e-01
4.70314085e-01 -3.36308718e-01 6.33155167e-01 1.19544946e-01
1.52830136e+00 -4.92191106e-01 9.46199943e-05 1.51280379e+00
1.01238573e+00 -2.84008890e-01 7.89551795e-01 -4.77680862e-01
1.26064932e+00 5.25656879e-01 8.92709792e-02 6.23053968e-01
7.52543151e-01 6.70899928e-01 7.19864786e-01 -1.94715992e-01
-1.59995973e-01 -5.96421242e-01 1.84803501e-01 1.17784393e+00
-2.03916103e-01 -1.51079118e-01 -3.72970104e-01 -4.01310861e-01
-1.66339135e+00 -6.09974563e-01 -1.22988448e-01 2.24084711e+00
6.31349027e-01 -2.04930291e-01 -4.16771740e-01 -2.25340262e-01
1.59997597e-01 5.39563835e-01 -8.59723926e-01 -3.46571510e-03
-3.13338280e-01 1.19247451e-01 2.09249377e-01 6.32198334e-01
-5.69241583e-01 8.00360143e-01 5.25478792e+00 7.29028046e-01
-1.17182672e+00 3.69968414e-01 8.52177024e-01 -3.81540865e-01
-1.09722114e+00 1.62343845e-01 -4.75010127e-01 3.25869650e-01
2.10163683e-01 2.51564085e-01 9.00173724e-01 3.95843446e-01
3.08287412e-01 9.83195826e-02 -1.25174141e+00 1.36135042e+00
1.45295173e-01 -1.41280746e+00 3.41836452e-01 4.71325666e-01
6.05192721e-01 -4.39640880e-01 4.91058111e-01 3.10281999e-02
1.72423854e-01 -1.19006586e+00 1.00961101e+00 8.52227271e-01
1.09583580e+00 -4.87199664e-01 -2.06437767e-01 4.20107991e-01
-9.59865272e-01 4.12163317e-01 -1.40497848e-01 -1.98094562e-01
2.01254040e-01 8.01290452e-01 1.89287305e-01 7.55580425e-01
4.33239698e-01 5.03587604e-01 -1.74838647e-01 2.76380062e-01
-3.39857876e-01 3.70032668e-01 -2.44096190e-01 5.97816944e-01
-1.60560116e-01 -6.43321931e-01 7.58259356e-01 3.13293517e-01
1.79730132e-01 8.07724416e-01 9.35363770e-02 1.50567043e+00
-9.84976217e-02 -3.52916479e-01 -5.44916868e-01 5.42440899e-02
2.66010255e-01 1.15187919e+00 -3.67399454e-01 -1.10724394e-03
-1.48347706e-01 1.28000927e+00 4.68551010e-01 7.64851391e-01
-9.00228500e-01 5.20340860e-01 1.34033430e+00 1.18917093e-01
-1.99825823e-01 -6.75991774e-01 -6.40255868e-01 -1.79351902e+00
1.99002296e-01 -8.26435804e-01 -2.09781453e-01 -1.11788368e+00
-1.29352653e+00 5.11081755e-01 -2.18946159e-01 -7.92886257e-01
3.49304646e-01 -7.16596901e-01 -3.45720142e-01 1.25811327e+00
-1.73084080e+00 -1.44420671e+00 -5.03154159e-01 5.88548124e-01
1.54765368e-01 2.93072462e-01 8.61821532e-01 3.67789686e-01
-7.36880183e-01 6.73482239e-01 -4.11972225e-01 -2.01372996e-01
5.23393512e-01 -8.79539788e-01 4.95243073e-01 6.10359788e-01
4.25637849e-02 1.06083441e+00 3.14838201e-01 -6.16491199e-01
-2.34011054e+00 -6.76955402e-01 1.37023315e-01 -5.40821612e-01
3.41325253e-01 -7.77803540e-01 -9.24519539e-01 5.13316393e-01
-4.54745702e-02 3.27922463e-01 5.98216355e-01 9.23264846e-02
-1.08655763e+00 3.27667557e-02 -1.12498772e+00 8.87455702e-01
1.59820437e+00 -9.83697116e-01 -4.70666215e-02 4.09524918e-01
9.96987820e-01 -9.34833884e-01 -6.42392457e-01 1.89061031e-01
7.78150022e-01 -1.44789565e+00 1.27084243e+00 -5.33938348e-01
6.85353518e-01 -2.42415592e-01 -3.78887087e-01 -9.88019705e-01
-2.73135394e-01 -1.11369967e+00 -4.71101999e-01 1.09424567e+00
-3.30492318e-01 -1.02289915e+00 8.23910654e-01 1.01315022e+00
-1.78998366e-01 -9.89066005e-01 -8.83934021e-01 -3.45312536e-01
-8.56239069e-03 -4.72594589e-01 9.67771113e-01 1.09367228e+00
-6.58044279e-01 3.25111710e-02 -6.96174800e-01 6.81850791e-01
8.34678888e-01 1.99174449e-01 8.93346608e-01 -1.06330621e+00
-6.55505538e-01 -3.66532326e-01 -6.48953319e-02 -1.66910827e+00
4.01272446e-01 -6.54273212e-01 -3.24928552e-01 -9.04432178e-01
1.96035832e-01 -8.99071574e-01 1.30044684e-01 3.15765113e-01
-4.01853919e-02 2.29756147e-01 3.04738194e-01 2.77199864e-01
-1.33168906e-01 1.02476335e+00 1.86104631e+00 1.54058516e-01
1.64749786e-01 -3.02749038e-01 -9.54954267e-01 7.67118633e-01
5.23399562e-02 -1.39886573e-01 -7.18178153e-01 -1.11749101e+00
6.38026118e-01 4.25637454e-01 9.14763212e-01 -4.22456294e-01
9.28704534e-03 -3.89762402e-01 3.55939329e-01 -3.99580061e-01
9.54712629e-01 -7.35111117e-01 3.96463633e-01 2.25796942e-02
-7.50227645e-02 -1.14302985e-01 6.70471489e-02 5.68238676e-01
2.36303538e-01 5.33515036e-01 6.17826521e-01 -1.20942621e-02
-1.58550188e-01 7.90121019e-01 3.34997803e-01 -4.72375713e-02
4.93785501e-01 -1.27539873e-01 -3.98781329e-01 -3.41835678e-01
-4.15623248e-01 4.47799824e-02 8.90948415e-01 2.82465965e-01
6.73947513e-01 -1.35617876e+00 -5.71186960e-01 8.12599659e-01
-1.21982589e-01 3.83030415e-01 6.80826306e-01 8.67680669e-01
-5.76716542e-01 -8.64139013e-03 6.31019101e-02 -6.99929059e-01
-1.02808702e+00 3.21935087e-01 4.21018243e-01 -5.68055883e-02
-6.91430807e-01 9.87730563e-01 1.13301253e+00 -6.71985805e-01
-7.90591240e-02 -2.74459451e-01 4.83058333e-01 -3.89272898e-01
5.73963523e-01 2.47909203e-01 -4.05842351e-04 -6.19373143e-01
-1.07742764e-01 6.98161185e-01 1.66274905e-01 -1.03785269e-01
1.17322803e+00 -3.01043242e-01 -4.23581630e-01 2.56145298e-01
1.15160835e+00 3.62002492e-01 -1.70566666e+00 -4.51629944e-02
-9.40904081e-01 -1.03211749e+00 2.49091431e-01 -6.75992608e-01
-1.36631751e+00 1.12206495e+00 2.28259742e-01 -2.09700197e-01
1.24486339e+00 -4.01421160e-01 7.58741975e-01 4.08348553e-02
5.68579853e-01 -1.54526830e-01 1.80251971e-02 4.80529487e-01
1.02025759e+00 -9.12462592e-01 1.91987425e-01 -1.01268423e+00
-3.50118488e-01 8.49515498e-01 7.52362728e-01 -2.68782705e-01
8.07431579e-01 3.64078701e-01 -1.20600633e-01 -4.52977002e-01
-6.98804080e-01 4.86586154e-01 5.56403995e-01 3.22668463e-01
1.72333032e-01 1.89814821e-01 3.64193946e-01 5.60419202e-01
-1.56740978e-01 -1.94580406e-01 9.54194441e-02 3.27505529e-01
3.31923515e-01 -9.13402975e-01 -3.09528738e-01 7.49653950e-02
1.38397679e-01 -2.11371422e-01 -3.17967683e-01 6.06777906e-01
2.21912637e-01 4.73425239e-01 7.10946172e-02 -2.92500496e-01
1.37751415e-01 -2.13480785e-01 9.30479825e-01 -4.96272802e-01
-2.54257441e-01 1.79361552e-01 -2.83880442e-01 -1.05340612e+00
-8.73713642e-02 -4.09052908e-01 -9.79083478e-01 -7.01694310e-01
-2.84985691e-01 -1.63267657e-01 5.86747289e-01 9.30942059e-01
7.68972576e-01 4.76032794e-01 7.43423998e-01 -1.25107408e+00
-4.39498425e-01 -2.70030975e-01 -6.12468600e-01 2.59060413e-01
4.88755614e-01 -1.01457262e+00 -4.56822842e-01 -2.33204693e-01] | [12.865384101867676, -0.24175941944122314] |
bca18210-1832-4b6f-83d9-b78a074dc9c7 | unsupervised-sampling-promoting-for | 2304.04298 | null | https://arxiv.org/abs/2304.04298v1 | https://arxiv.org/pdf/2304.04298v1.pdf | Unsupervised Sampling Promoting for Stochastic Human Trajectory Prediction | The indeterminate nature of human motion requires trajectory prediction systems to use a probabilistic model to formulate the multi-modality phenomenon and infer a finite set of future trajectories. However, the inference processes of most existing methods rely on Monte Carlo random sampling, which is insufficient to cover the realistic paths with finite samples, due to the long tail effect of the predicted distribution. To promote the sampling process of stochastic prediction, we propose a novel method, called BOsampler, to adaptively mine potential paths with Bayesian optimization in an unsupervised manner, as a sequential design strategy in which new prediction is dependent on the previously drawn samples. Specifically, we model the trajectory sampling as a Gaussian process and construct an acquisition function to measure the potential sampling value. This acquisition function applies the original distribution as prior and encourages exploring paths in the long-tail region. This sampling method can be integrated with existing stochastic predictive models without retraining. Experimental results on various baseline methods demonstrate the effectiveness of our method. | ['Kun Zhang', 'Shunxing Fan', 'Zhenhao Chen', 'Guangyi Chen'] | 2023-04-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_Unsupervised_Sampling_Promoting_for_Stochastic_Human_Trajectory_Prediction_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_Unsupervised_Sampling_Promoting_for_Stochastic_Human_Trajectory_Prediction_CVPR_2023_paper.pdf | cvpr-2023-1 | ['trajectory-prediction'] | ['computer-vision'] | [ 1.82764485e-01 -5.75491972e-02 -5.22425950e-01 -2.50856251e-01
-7.24496663e-01 -4.04543519e-01 7.09999323e-01 -4.63936716e-01
-1.81422442e-01 9.48539734e-01 5.60717881e-01 -3.08190167e-01
-1.98391467e-01 -1.03122723e+00 -7.59825468e-01 -7.71955848e-01
1.80000931e-01 5.40361881e-01 7.24152088e-01 2.09213331e-01
2.24630088e-01 2.43746743e-01 -1.30911911e+00 -4.02335450e-02
1.07514501e+00 4.91867423e-01 4.73783284e-01 6.43006146e-01
-1.00848012e-01 5.60527265e-01 -2.12227419e-01 -1.10292770e-01
-1.01976907e-02 -4.24158365e-01 -4.55077350e-01 1.29216751e-02
-4.28904623e-01 -5.99450529e-01 -4.19787377e-01 1.03253520e+00
1.94860682e-01 5.54310679e-01 9.43720579e-01 -1.23916101e+00
-2.89080083e-01 6.54690564e-01 -5.08512020e-01 6.95581809e-02
4.54753280e-01 5.19803226e-01 6.05768919e-01 -7.69496977e-01
4.66325581e-01 1.37229180e+00 6.69263601e-01 4.48023766e-01
-9.16237772e-01 -5.13163567e-01 4.40285861e-01 1.65912867e-01
-1.35975218e+00 -2.83369541e-01 8.57813478e-01 -4.34921175e-01
2.86286861e-01 1.48350984e-01 8.92509520e-01 1.49753284e+00
2.21771747e-01 1.20629168e+00 7.63303220e-01 7.67767942e-03
5.24989903e-01 -3.16263139e-02 -1.67044878e-01 4.38273549e-01
7.28459097e-03 4.25641447e-01 -4.63052660e-01 -5.39181292e-01
8.34689140e-01 2.83316404e-01 -2.65564919e-01 -1.81171030e-01
-1.23493373e+00 6.16068602e-01 -1.38335794e-01 -1.76350400e-01
-4.74084377e-01 3.90525669e-01 3.68542485e-02 -4.14795935e-01
3.30311179e-01 -2.27634922e-01 -3.23097318e-01 -6.36155665e-01
-1.09979093e+00 7.19743073e-01 6.15051329e-01 1.14520335e+00
7.04761386e-01 -1.25080988e-01 -5.64043880e-01 4.12366778e-01
7.59758413e-01 7.52922714e-01 3.25657040e-01 -1.12387931e+00
5.00179946e-01 2.24758074e-01 6.90874577e-01 -9.14044857e-01
-8.85559525e-03 -1.23375706e-01 -7.57746100e-01 -6.45222589e-02
5.69159329e-01 -4.07062024e-01 -7.50858426e-01 1.64917076e+00
6.47366643e-01 6.94525778e-01 -1.99730799e-01 8.35636437e-01
1.34771049e-01 1.05462515e+00 1.38535827e-01 -3.61488193e-01
9.11705256e-01 -9.06657755e-01 -7.72834957e-01 1.23512372e-02
2.50727117e-01 -3.78777862e-01 1.28203464e+00 4.54707026e-01
-8.04309189e-01 -5.21520138e-01 -7.92999566e-01 4.43672270e-01
1.51217446e-01 1.65051296e-01 5.16422689e-01 6.87009811e-01
-6.22103095e-01 7.57801414e-01 -1.26177132e+00 6.37764169e-04
6.76115930e-01 -2.13774174e-01 3.03353965e-01 -1.02564156e-01
-1.19502139e+00 3.53861630e-01 5.12273550e-01 1.85128197e-01
-1.22898376e+00 -5.73355734e-01 -5.08531332e-01 -1.09056897e-01
6.26269698e-01 -7.60934114e-01 1.12185180e+00 -3.72963667e-01
-1.86001730e+00 -1.25544533e-01 -7.52025008e-01 -2.58615553e-01
8.80122244e-01 -4.11347032e-01 -3.22514325e-01 -1.96785294e-02
1.90542951e-01 4.26825166e-01 9.82494771e-01 -1.29014599e+00
-7.39845693e-01 -1.53402025e-02 -2.40259051e-01 2.77416438e-01
-6.54288828e-02 -3.90217721e-01 -7.76593626e-01 -5.79772413e-01
1.02226280e-01 -1.03210485e+00 -7.78926909e-01 -1.55044541e-01
-5.82936645e-01 -2.86153287e-01 6.88400805e-01 -4.97810870e-01
1.55163419e+00 -2.00673461e+00 -6.40247837e-02 5.06774604e-01
-1.25153199e-01 -1.96910322e-01 2.89105713e-01 3.74554902e-01
6.23368263e-01 4.13747765e-02 -5.29395282e-01 -2.43610919e-01
-6.13146229e-04 2.24641919e-01 -7.09635913e-01 3.04088354e-01
-1.15677826e-01 5.97924531e-01 -1.38505459e+00 -6.39165878e-01
2.50596166e-01 1.09796546e-01 -5.49313426e-01 3.06895018e-01
-6.03995204e-01 7.69928634e-01 -1.12191486e+00 5.30594945e-01
8.38292122e-01 -3.03746343e-01 7.38289533e-03 1.61960706e-01
-4.89823595e-02 -1.02773719e-01 -1.34422779e+00 1.71978974e+00
5.56978099e-02 1.56679377e-01 -4.90023226e-01 -4.30117846e-01
9.15031493e-01 1.89360514e-01 4.82414007e-01 2.60128498e-01
-1.40063509e-01 8.78418386e-02 -3.72165233e-01 -7.18651235e-01
8.06532741e-01 -1.55454427e-01 5.01338728e-02 3.77877057e-01
-5.67881584e-01 -1.21626267e-02 -9.29184332e-02 1.58755139e-01
1.00218332e+00 9.85752165e-01 9.99880508e-02 6.81860186e-03
4.20900315e-01 1.48584872e-01 9.01538014e-01 1.02537310e+00
-4.23360527e-01 5.84584713e-01 4.12750095e-01 -1.28231779e-01
-9.58790898e-01 -1.40810478e+00 1.66346077e-02 7.46037424e-01
5.10848105e-01 -3.55130941e-01 -6.54675484e-01 -7.10896134e-01
-2.31293350e-01 1.02900457e+00 -3.68637532e-01 -2.39358470e-01
-6.73412859e-01 -9.66527104e-01 3.85259897e-01 5.79371870e-01
4.21037972e-01 -1.01641250e+00 -6.33893371e-01 4.51283157e-01
-4.72233325e-01 -7.74711132e-01 -5.68993032e-01 -4.27897543e-01
-8.98050070e-01 -7.42033362e-01 -6.65343463e-01 -1.35070935e-01
4.29644585e-01 1.03255577e-01 6.93533421e-01 -2.32149735e-01
3.20461243e-01 3.39097202e-01 -3.36265445e-01 -2.98497111e-01
-3.30791444e-01 -6.67353570e-02 -5.32017238e-02 1.09392524e-01
2.71027803e-01 -6.16682053e-01 -7.67229319e-01 3.78769934e-01
-6.39485478e-01 1.59786299e-01 4.87481177e-01 7.29844153e-01
7.59921968e-01 3.68620723e-01 6.61748827e-01 -6.71775877e-01
6.00237012e-01 -9.61420119e-01 -5.07744491e-01 1.43701047e-01
-4.55770314e-01 1.30807430e-01 5.55632234e-01 -8.21829855e-01
-1.58307624e+00 6.68621883e-02 -4.58660126e-02 -4.65773016e-01
-3.84723455e-01 5.67818642e-01 -4.55470920e-01 5.26464462e-01
3.41344297e-01 5.33963382e-01 -1.02928355e-01 -3.17941278e-01
4.70114231e-01 3.96169990e-01 4.23382938e-01 -9.05433118e-01
6.69146359e-01 8.16096067e-01 1.58357978e-01 -6.43922985e-01
-7.38203228e-01 -3.19239110e-01 -4.50866193e-01 -6.33045733e-01
7.65276253e-01 -7.55255461e-01 -7.55408823e-01 4.78863835e-01
-1.10896182e+00 -3.47495556e-01 -1.23510793e-01 8.38102281e-01
-8.83280933e-01 6.30463779e-01 -2.76633501e-01 -1.38863206e+00
1.05362371e-01 -1.11547577e+00 1.11340094e+00 4.56518441e-01
-2.63937324e-01 -9.23456788e-01 2.14399174e-01 5.61267808e-02
1.42843693e-01 2.81497002e-01 5.74926257e-01 -2.74013489e-01
-1.10776138e+00 -1.60723731e-01 9.82860327e-02 -1.47742555e-01
-8.11942294e-02 1.70901194e-01 -6.50578976e-01 8.02529380e-02
-5.79795800e-02 4.82886024e-02 6.93512261e-01 6.34934783e-01
1.49045134e+00 -1.91681996e-01 -9.48385000e-01 3.95640254e-01
1.15398991e+00 2.71970302e-01 7.78625488e-01 2.16406614e-01
6.23514116e-01 3.64883423e-01 1.05464661e+00 8.38379204e-01
5.01497030e-01 3.77274901e-01 2.70089686e-01 6.77644372e-01
3.67539197e-01 -1.00095916e+00 4.95998114e-01 4.69887137e-01
-1.56645477e-01 -3.37586075e-01 -8.67320240e-01 6.45261049e-01
-2.18720865e+00 -1.33953607e+00 -1.72304869e-01 2.21491790e+00
9.30754662e-01 4.24040496e-01 2.55173057e-01 -1.82689473e-01
7.03091025e-01 2.27904513e-01 -7.12684214e-01 3.95898551e-01
2.25383162e-01 -4.65224564e-01 5.47278345e-01 5.60294449e-01
-9.02827442e-01 9.96790290e-01 6.65860653e+00 1.35065818e+00
-5.67810059e-01 -7.88221583e-02 4.73728299e-01 3.41296531e-02
-7.54619420e-01 4.42187011e-01 -1.21375823e+00 1.00535846e+00
9.06120241e-01 -1.24985673e-01 2.90500104e-01 8.27256918e-01
8.19516540e-01 -3.41150582e-01 -8.24433863e-01 6.76084697e-01
-4.68007565e-01 -1.24817216e+00 5.95280640e-02 1.34614617e-01
6.71640635e-01 -2.58073717e-01 -7.06609488e-02 3.04421335e-01
7.02102482e-01 -9.17798460e-01 9.04932380e-01 1.39634156e+00
2.97406316e-01 -7.15111554e-01 3.24901253e-01 9.69020545e-01
-1.23202968e+00 -1.45625081e-02 -3.40529203e-01 5.34633771e-02
9.42053080e-01 8.15967858e-01 -9.46363628e-01 5.40446639e-01
4.61882055e-01 8.11986804e-01 -7.39423186e-02 1.23363793e+00
-2.77560353e-01 1.26569486e+00 -5.88066638e-01 -3.54409456e-01
2.75058031e-01 -6.55107856e-01 1.05298305e+00 9.72162783e-01
6.51420832e-01 3.77328880e-02 4.34334159e-01 1.07570195e+00
5.13055086e-01 -1.82696730e-01 -4.73476052e-01 5.22165038e-02
1.03353298e+00 7.69595265e-01 -6.45563066e-01 -4.08744037e-01
-1.28323704e-01 7.23706722e-01 6.88009039e-02 6.91462815e-01
-1.23014677e+00 -8.68254825e-02 1.35475531e-01 5.54328114e-02
3.94607186e-01 -4.47989076e-01 -4.50928688e-01 -9.48546946e-01
9.57142040e-02 -5.36554575e-01 2.03748778e-01 -6.94906652e-01
-1.27901125e+00 1.12037137e-01 4.38347191e-01 -1.52046704e+00
-4.82318759e-01 -6.66394308e-02 -9.20318902e-01 9.38431263e-01
-1.06527233e+00 -1.08084321e+00 -6.78601637e-02 4.40916866e-01
5.49447119e-01 -9.11331270e-03 4.48938668e-01 -5.42512760e-02
-4.11698133e-01 2.42118120e-01 2.32321233e-01 -2.86485612e-01
3.02705109e-01 -1.06505072e+00 1.82247207e-01 8.62281740e-01
-2.95852989e-01 6.38662338e-01 1.01374030e+00 -1.18682313e+00
-1.08868933e+00 -1.12136102e+00 3.52320105e-01 -6.66821718e-01
6.69236720e-01 3.22764590e-02 -1.00218928e+00 6.58569813e-01
-2.55445600e-01 -2.59656489e-01 6.48756981e-01 5.31973783e-03
2.98792422e-01 1.50619164e-01 -9.50199246e-01 9.76029932e-01
1.16110218e+00 1.84260476e-02 -6.29274309e-01 1.94084436e-01
8.34162831e-01 -2.95188546e-01 -7.53319263e-01 4.76080358e-01
7.01472580e-01 -6.54375672e-01 9.45592821e-01 -4.00368690e-01
3.71225536e-01 -6.86863065e-01 -8.70244429e-02 -1.11794245e+00
-2.51524985e-01 -8.40186059e-01 -5.78728676e-01 1.21473408e+00
3.81222963e-01 -4.65074569e-01 1.11625123e+00 6.50489748e-01
-5.83782606e-02 -9.85056281e-01 -1.00134754e+00 -7.91550756e-01
-9.78842452e-02 -9.19988632e-01 1.00868332e+00 4.04895663e-01
-2.43428469e-01 -1.66852340e-01 -7.78363824e-01 4.16257679e-01
8.98436785e-01 1.29197255e-01 1.11762941e+00 -7.96749651e-01
-5.95144272e-01 -2.76270062e-01 1.39185265e-01 -1.88790166e+00
6.23099767e-02 -3.15123737e-01 5.52962422e-01 -1.40327787e+00
3.02611530e-01 -6.22910976e-01 8.62997770e-02 -1.66871697e-01
-6.29842281e-01 -6.93980873e-01 -1.73846960e-01 5.03567159e-01
-6.63098574e-01 1.15958893e+00 1.53081107e+00 2.14144900e-01
-4.11212474e-01 5.92505157e-01 -3.59580994e-01 1.08118010e+00
5.18287241e-01 -5.77072203e-01 -7.65485287e-01 4.93718572e-02
1.25109211e-01 4.49446678e-01 3.67380679e-01 -9.67218399e-01
3.00308824e-01 -8.87360871e-01 3.49655062e-01 -1.39875388e+00
3.36370528e-01 -6.65458679e-01 4.38273132e-01 2.36965492e-01
-2.65346766e-01 -3.12519401e-01 -1.98610172e-01 1.45846939e+00
1.16427764e-01 -3.22296202e-01 3.23636472e-01 -1.41142190e-01
-6.32065475e-01 6.50130510e-01 -6.89488351e-01 -1.51725309e-02
9.89090264e-01 -3.02286536e-01 1.64758056e-01 -3.99423540e-01
-6.86833620e-01 6.14536107e-01 3.34632635e-01 1.74649268e-01
8.49798918e-01 -1.56585848e+00 -4.06203866e-01 -1.59210250e-01
-1.48005471e-01 3.66746992e-01 3.92912716e-01 6.79168522e-01
-2.41297096e-01 -8.38560313e-02 3.31665039e-01 -7.76649237e-01
-5.84029257e-01 4.92128581e-01 1.04802921e-01 -4.56685185e-01
-7.88395464e-01 5.94636381e-01 -1.10687494e-01 -4.65617061e-01
9.89263318e-03 -1.89031541e-01 -3.36509466e-01 -3.97161692e-01
5.11015654e-01 7.64104486e-01 -9.02146697e-01 -3.20085734e-01
-3.32793072e-02 3.00099373e-01 2.02246085e-01 -6.99951172e-01
9.18653727e-01 -5.70336103e-01 1.99738368e-01 9.56285298e-01
6.06678188e-01 1.66757897e-01 -1.91998541e+00 -8.87175426e-02
-2.34613437e-02 -7.49383330e-01 -2.24647641e-01 -4.53273863e-01
-3.83035868e-01 5.23309290e-01 1.91205800e-01 2.15335917e-02
7.65361488e-01 -1.60355806e-01 1.14813662e+00 1.90116704e-01
5.90727985e-01 -1.18413901e+00 -1.46813497e-01 4.70835865e-01
4.75452662e-01 -1.09661257e+00 9.72792041e-04 -5.05823612e-01
-7.47180343e-01 9.64257777e-01 6.15587592e-01 -3.27377468e-02
1.00170982e+00 5.73088862e-02 -5.24258435e-01 7.28230104e-02
-6.76225901e-01 1.05694041e-01 2.01678410e-01 6.10582829e-01
-1.41323015e-01 2.57973880e-01 -3.69418651e-01 9.88569200e-01
-1.30269513e-01 5.54583609e-01 4.28750515e-01 8.84198666e-01
-6.73473060e-01 -8.66806090e-01 -5.26979327e-01 4.43862647e-01
-6.15062751e-02 7.83005953e-02 4.11794752e-01 3.80736321e-01
8.93760771e-02 8.99549484e-01 -1.65323362e-01 -3.55667204e-01
-1.45863481e-02 -2.98762657e-02 1.18834861e-01 -1.86083376e-01
2.57629782e-01 2.25829840e-01 2.82234084e-02 -5.34060538e-01
-3.23209792e-01 -1.20622921e+00 -1.33042026e+00 -1.83577895e-01
-2.93228269e-01 9.57688615e-02 1.28071874e-01 1.18759215e+00
2.75711566e-01 4.19719189e-01 5.55106580e-01 -8.19337904e-01
-9.48221922e-01 -1.00787568e+00 -5.34730375e-01 2.46296778e-01
1.14981383e-01 -8.91180694e-01 -3.17791194e-01 1.62361965e-01] | [6.5062408447265625, 0.8591596484184265] |
e38d2acb-4ced-46fc-9b71-a2fadaaf7366 | carla-a-python-library-to-benchmark | 2108.00783 | null | https://arxiv.org/abs/2108.00783v1 | https://arxiv.org/pdf/2108.00783v1.pdf | CARLA: A Python Library to Benchmark Algorithmic Recourse and Counterfactual Explanation Algorithms | Counterfactual explanations provide means for prescriptive model explanations by suggesting actionable feature changes (e.g., increase income) that allow individuals to achieve favorable outcomes in the future (e.g., insurance approval). Choosing an appropriate method is a crucial aspect for meaningful counterfactual explanations. As documented in recent reviews, there exists a quickly growing literature with available methods. Yet, in the absence of widely available opensource implementations, the decision in favor of certain models is primarily based on what is readily available. Going forward - to guarantee meaningful comparisons across explanation methods - we present CARLA (Counterfactual And Recourse LibrAry), a python library for benchmarking counterfactual explanation methods across both different data sets and different machine learning models. In summary, our work provides the following contributions: (i) an extensive benchmark of 11 popular counterfactual explanation methods, (ii) a benchmarking framework for research on future counterfactual explanation methods, and (iii) a standardized set of integrated evaluation measures and data sets for transparent and extensive comparisons of these methods. We have open-sourced CARLA and our experimental results on Github, making them available as competitive baselines. We welcome contributions from other research groups and practitioners. | ['Gjergji Kasneci', 'Tobias Richter', 'Johannes van den Heuvel', 'Sascha Bielawski', 'Martin Pawelczyk'] | 2021-08-02 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 6.35951981e-02 5.70007145e-01 -1.07364583e+00 -4.95027095e-01
-5.81050038e-01 -3.95674318e-01 1.01205111e+00 2.31520981e-01
-3.77896726e-02 1.33013344e+00 8.86149883e-01 -9.38918650e-01
-2.62974143e-01 -5.62609017e-01 -6.84545100e-01 -1.17089137e-01
-7.93137401e-03 3.71096402e-01 -5.82215488e-01 -9.22335777e-03
5.89802563e-01 1.04314752e-01 -1.71830714e+00 2.09708840e-01
1.25207627e+00 3.95079911e-01 -2.74826735e-01 2.76305348e-01
-2.40029190e-02 5.84381819e-01 -1.98575959e-01 -8.79357636e-01
2.39376456e-01 -6.68980122e-01 -9.42588091e-01 -1.97427034e-01
1.90828398e-01 -4.01625007e-01 1.22340858e-01 6.18253589e-01
4.51955169e-01 1.48504108e-01 7.74122357e-01 -1.63120413e+00
-1.00406027e+00 1.12112904e+00 -3.28701615e-01 1.01616830e-01
5.74331403e-01 5.43851614e-01 1.07773566e+00 -5.84136963e-01
5.39988160e-01 1.30817735e+00 5.30393243e-01 8.60143840e-01
-1.44567168e+00 -8.81376863e-01 4.97556180e-01 2.63060510e-01
-6.07380629e-01 -4.77443963e-01 7.39220679e-01 -2.79803574e-01
1.15292418e+00 8.34165096e-01 9.46492910e-01 1.36192334e+00
2.78323859e-01 7.03049779e-01 1.32444680e+00 -4.75755960e-01
4.47711378e-01 1.07922100e-01 1.70527667e-01 2.83151064e-02
7.93253303e-01 8.31060231e-01 -3.79601866e-01 -5.70711017e-01
4.12784874e-01 2.69486606e-01 -4.12413806e-01 -5.37447393e-01
-1.08097303e+00 1.10235643e+00 4.16205525e-01 9.25655011e-03
-5.85165977e-01 4.13105190e-01 3.44813645e-01 9.86724645e-02
5.25685489e-01 7.16762245e-01 -7.33541131e-01 -2.68215030e-01
-7.90927827e-01 1.00266838e+00 7.06339121e-01 4.39409554e-01
3.60649884e-01 7.80026019e-02 -2.60045856e-01 5.42671144e-01
4.85220343e-01 4.01399761e-01 7.51551211e-01 -1.10140419e+00
5.45420110e-01 5.49432933e-01 6.09909415e-01 -5.43787837e-01
-3.88684839e-01 -2.14388400e-01 -5.49597442e-01 2.15821549e-01
4.18637574e-01 -4.67453450e-01 -5.18527091e-01 1.88035834e+00
2.16696754e-01 2.41963863e-01 7.15694502e-02 1.02917361e+00
5.50727308e-01 1.03711307e-01 6.22278154e-01 -5.01220226e-01
1.12193155e+00 -6.68710709e-01 -7.26979554e-01 -3.98205549e-01
7.21092999e-01 -6.82371855e-01 1.12238801e+00 -1.48538109e-02
-1.12410593e+00 -9.82299745e-02 -8.54783833e-01 1.59374118e-01
-3.98279965e-01 -3.88861686e-01 1.62425590e+00 8.69823515e-01
-5.69956243e-01 7.36502111e-01 -5.43664873e-01 -4.15477157e-01
4.53323990e-01 2.30528355e-01 -8.63293409e-02 1.46361500e-01
-1.26593745e+00 8.66054118e-01 2.94066787e-01 -3.03954810e-01
-6.16975188e-01 -1.26429820e+00 -9.97193754e-01 2.73195595e-01
2.56701022e-01 -1.31050611e+00 1.54227591e+00 -1.34436274e+00
-1.19642544e+00 5.71837664e-01 -2.07095649e-02 -1.02226961e+00
8.72852147e-01 -2.00444534e-01 -4.88045841e-01 -5.97765028e-01
4.87514466e-01 7.38994837e-01 2.96425313e-01 -1.05312741e+00
-6.36550725e-01 -3.31052333e-01 2.65627354e-01 3.66113096e-01
1.78049594e-01 8.33438486e-02 5.14186800e-01 -8.09360623e-01
-3.92400324e-01 -7.81452179e-01 -6.17301643e-01 -3.19320381e-01
-7.56743193e-01 -2.50730008e-01 3.51706773e-01 -3.19288284e-01
1.23667598e+00 -1.48447680e+00 -6.00704074e-01 -2.32247800e-01
-1.92844287e-01 8.14443827e-03 2.66247720e-01 3.76297176e-01
-7.91644990e-01 7.08209157e-01 -3.34858119e-01 -2.23979995e-01
3.49524200e-01 -1.70431361e-01 -8.01467478e-01 4.27995920e-01
-6.72149658e-02 1.03086495e+00 -9.93064165e-01 -3.00283015e-01
6.01057351e-01 4.32760678e-02 -7.27765203e-01 -8.71895030e-02
-1.34254336e-01 1.87479585e-01 -3.11100364e-01 4.61485028e-01
6.40381217e-01 -1.48099542e-01 2.52794325e-01 4.17921156e-01
-3.94126177e-01 1.04071581e+00 -1.06241870e+00 1.12713361e+00
-5.70974946e-01 4.45736468e-01 -7.02474236e-01 -8.08448911e-01
3.44040155e-01 5.37827313e-01 2.15642422e-01 -3.76082659e-01
-7.81928450e-02 3.77828240e-01 2.72527754e-01 -1.91253364e-01
3.50313574e-01 -5.58120668e-01 -1.13280207e-01 7.30121374e-01
-5.10942221e-01 -1.81226104e-01 9.96164158e-02 -4.57067005e-02
6.22331083e-01 2.39279553e-01 1.09876978e+00 -4.44567293e-01
9.67610478e-02 2.45784625e-01 6.30318105e-01 8.73910010e-01
-2.36417845e-01 5.92433989e-01 3.82393628e-01 -8.45679641e-01
-8.61633837e-01 -1.03472126e+00 -3.89809251e-01 5.24881959e-01
-2.66587764e-01 -5.10298386e-02 -4.45476234e-01 -9.15267408e-01
5.43565691e-01 1.68320608e+00 -9.87805247e-01 3.01356353e-02
-1.66002229e-01 -1.06895423e+00 1.44844338e-01 6.62703156e-01
5.38734913e-01 -1.09218359e+00 -7.33712316e-01 -1.82888249e-03
-2.42357373e-01 -1.45038068e-01 -4.04136688e-01 -3.07285070e-01
-1.16409457e+00 -1.31697845e+00 -5.84586978e-01 1.29723951e-01
3.29468578e-01 4.34465975e-01 1.30465937e+00 2.36223400e-01
2.46401787e-01 2.23778680e-01 4.21584658e-02 -1.04460001e+00
-1.15717508e-01 -4.29113388e-01 2.24746987e-01 -5.23356318e-01
4.27586973e-01 -5.76444685e-01 -9.75480855e-01 -1.61229894e-02
-5.20677507e-01 2.85819560e-01 2.64933825e-01 7.64705181e-01
3.43465030e-01 -5.39616883e-01 9.16332603e-01 -1.15901613e+00
8.15685809e-01 -8.84048760e-01 -6.50317669e-01 1.08167805e-01
-1.32378423e+00 1.81712359e-02 5.77313423e-01 -9.28684101e-02
-1.49081862e+00 -4.21442896e-01 3.85273471e-02 2.15562627e-01
-3.87733549e-01 4.84936982e-01 -8.97036679e-03 6.95110083e-01
8.05655837e-01 -2.69363642e-01 -2.42438212e-01 -5.03950417e-01
5.91005087e-01 4.71381545e-01 2.67482996e-01 -3.06155682e-01
4.12108660e-01 6.24883115e-01 -3.18094313e-01 -1.26989678e-01
-7.70104468e-01 -1.78959370e-01 -1.50033329e-02 7.19695091e-02
4.80723739e-01 -7.21956313e-01 -8.39648843e-01 -1.38560593e-01
-9.77665246e-01 -5.74711442e-01 -4.90093380e-01 8.45381618e-01
-7.39623666e-01 3.54515798e-02 6.16768785e-02 -1.01880455e+00
-3.45955521e-01 -7.54321218e-01 5.07777274e-01 4.48579639e-01
-9.24145639e-01 -1.36210835e+00 2.78409094e-01 5.33039212e-01
2.68843472e-01 4.80281413e-01 7.30747163e-01 -8.02644908e-01
-3.83625686e-01 -1.15679190e-01 -4.68804389e-02 -3.72704566e-01
1.55228093e-01 2.08298028e-01 -9.12566423e-01 1.48392633e-01
-2.36578673e-01 -6.61455989e-02 9.53262925e-01 1.05877566e+00
1.23063231e+00 -8.51857007e-01 -7.19498813e-01 3.00499469e-01
1.18407643e+00 2.77780518e-02 5.45381069e-01 5.70212245e-01
9.37728509e-02 7.77633309e-01 8.79500508e-01 5.87874830e-01
7.65408456e-01 7.50647724e-01 4.77157772e-01 -1.62495822e-01
-7.05251172e-02 -4.85201806e-01 2.81793356e-01 -2.85010189e-01
-4.04665679e-01 -8.37742835e-02 -7.57637262e-01 9.26715434e-01
-2.00622559e+00 -1.46255255e+00 -4.25413907e-01 2.57273388e+00
4.49698567e-01 -1.43775484e-02 2.11431637e-01 -9.51606482e-02
4.18228567e-01 1.42141014e-01 -7.31122494e-01 -8.87545526e-01
-2.62480266e-02 -2.23424256e-01 4.63814050e-01 5.91176152e-01
-8.52018237e-01 4.46424186e-01 7.16142988e+00 2.83852160e-01
-9.34034288e-01 1.21753730e-01 9.86921012e-01 -3.97208482e-01
-1.05372274e+00 5.73805869e-01 -3.18976641e-01 5.83558261e-01
1.09723401e+00 -1.03118253e+00 1.40694812e-01 1.01294982e+00
9.40972924e-01 -9.27209258e-02 -1.26682711e+00 5.52787542e-01
-3.89351726e-01 -1.81615341e+00 1.24360777e-01 1.44067228e-01
1.16716707e+00 -1.00710019e-01 2.04404563e-01 4.34239388e-01
8.20073962e-01 -1.08151507e+00 9.03646231e-01 2.66779602e-01
5.05351126e-01 -6.80859566e-01 9.57884610e-01 1.70674577e-01
-6.70170963e-01 -3.08180451e-01 -1.82456449e-01 -7.68554688e-01
1.36187792e-01 6.42144322e-01 -8.78055096e-01 7.64376163e-01
5.70385873e-01 7.05789030e-01 -1.49189383e-01 9.94717777e-01
-7.15504110e-01 8.83454323e-01 5.95163181e-02 4.28647734e-02
8.89521092e-03 1.49484962e-01 4.75644916e-01 9.17791665e-01
4.11357433e-01 3.38239014e-01 -3.67855757e-01 1.21656084e+00
-9.28551480e-02 2.24091142e-01 -9.52499211e-01 1.55242950e-01
6.62001133e-01 7.55714536e-01 -3.72489482e-01 -4.13743824e-01
-5.28746724e-01 4.32871670e-01 -4.75732312e-02 3.65785807e-01
-9.41038847e-01 2.08177447e-01 1.12712371e+00 2.35520586e-01
-3.77341241e-01 6.36966169e-01 -9.80984092e-01 -1.27827370e+00
-3.24600309e-01 -9.27400529e-01 9.65950966e-01 -8.23176086e-01
-1.18726575e+00 -1.87544391e-01 3.49814057e-01 -1.13276911e+00
-8.20379317e-01 -2.35544920e-01 -1.20280778e+00 1.00767314e+00
-1.43047273e+00 -9.40687716e-01 1.17689185e-01 2.21275892e-02
6.54331028e-01 1.88173234e-01 8.17952871e-01 -2.40263060e-01
-5.87388098e-01 3.93592000e-01 -5.20008132e-02 -3.93982351e-01
6.32666111e-01 -1.58432889e+00 7.50685096e-01 8.33872914e-01
2.62358766e-02 1.04538846e+00 1.18803883e+00 -8.37854087e-01
-6.89125896e-01 -9.11527753e-01 1.22236550e+00 -6.38948917e-01
4.23319489e-01 1.27574429e-01 -5.61821282e-01 1.15632224e+00
3.16836923e-01 -4.43333715e-01 9.11956668e-01 7.03053713e-01
1.15588158e-01 1.27642334e-01 -1.29791415e+00 1.01473916e+00
9.53211129e-01 1.23397363e-02 -9.12607551e-01 2.96282470e-01
6.39213502e-01 -1.54961184e-01 -4.68142837e-01 1.62230775e-01
7.73650348e-01 -1.41271818e+00 9.55605447e-01 -9.74550664e-01
7.96774685e-01 2.84116771e-02 1.47655904e-01 -1.56067479e+00
-1.04839012e-01 -6.93411827e-01 4.64852974e-02 1.22527719e+00
7.72710800e-01 -1.17875028e+00 5.36517739e-01 1.46596408e+00
-5.23863658e-02 -8.52030039e-01 -8.90892565e-01 -6.29922867e-01
2.62392163e-01 -7.90755451e-01 1.37847173e+00 1.22546113e+00
5.11272907e-01 -4.94067520e-02 -3.00233811e-01 -5.66211380e-02
6.96174383e-01 7.37003863e-01 9.65023100e-01 -9.60039556e-01
-1.96477011e-01 -6.27865911e-01 1.95834085e-01 -5.23222506e-01
2.13444740e-01 -8.84396195e-01 -4.76087064e-01 -1.64737904e+00
8.05253565e-01 -2.01667443e-01 -2.19187647e-01 6.39872551e-01
-5.55456042e-01 -7.52633438e-02 1.21342413e-01 1.88481554e-01
-9.29891244e-02 7.18506873e-01 9.22907531e-01 1.40240476e-01
-3.36732119e-01 3.46171230e-01 -1.41941524e+00 1.01638746e+00
9.27399397e-01 -6.30107284e-01 -5.07873416e-01 -9.66511387e-03
8.02914612e-03 1.09389387e-01 8.92436624e-01 -4.74786520e-01
-4.53734607e-01 -9.40595090e-01 4.13551778e-01 -3.02515030e-01
-1.07313521e-01 -4.08801854e-01 3.78777444e-01 8.05946290e-01
-5.79804182e-01 -3.53179337e-03 3.69186729e-01 5.43839753e-01
1.26446992e-01 -1.37764066e-01 2.79029518e-01 -1.41712382e-01
-4.79407996e-01 3.89540009e-02 -1.20109208e-01 -2.56328583e-02
9.41724598e-01 -2.67885923e-01 -6.21897757e-01 -7.24186420e-01
-4.05738533e-01 4.24597204e-01 6.72799110e-01 6.24508798e-01
3.23969305e-01 -1.40623164e+00 -8.82019162e-01 -3.09688717e-01
1.20381832e-01 -6.64017260e-01 3.02951425e-01 7.29875028e-01
9.31701809e-02 1.01289225e+00 -2.31603324e-01 1.91958711e-01
-9.45245206e-01 6.44836426e-01 4.01251823e-01 -4.20490563e-01
-3.95530611e-01 3.16067129e-01 4.56277698e-01 -5.99527717e-01
-5.25594950e-01 -4.47931945e-01 -1.91616371e-01 -3.12208086e-01
7.47680783e-01 6.11744046e-01 -2.99453437e-01 -1.24170177e-01
-5.26498437e-01 -3.90319884e-01 9.85442773e-02 -1.92844257e-01
1.46568859e+00 -2.38638371e-01 1.92303047e-01 5.41840672e-01
4.63047206e-01 4.98076566e-02 -1.26787567e+00 5.36489248e-01
5.68203516e-02 -8.56596828e-01 -1.65027142e-01 -1.13974535e+00
-4.87946302e-01 4.68284994e-01 4.86992449e-01 3.57873857e-01
8.91113341e-01 -1.38546228e-01 8.67762864e-02 -1.87010422e-01
2.64415443e-01 -8.88930261e-01 -5.38325191e-01 -2.15227753e-01
1.42347741e+00 -1.50643694e+00 4.77904856e-01 -2.41780043e-01
-7.39978969e-01 6.05726600e-01 4.42537934e-01 5.15112281e-02
3.91872346e-01 -3.57837290e-01 1.02702845e-02 -1.00284785e-01
-1.13506353e+00 -2.01739762e-02 3.39859843e-01 6.69259071e-01
1.03904653e+00 6.32208228e-01 -1.00696671e+00 1.07540429e+00
-6.08031750e-01 1.83003455e-01 6.31845236e-01 1.23204380e-01
1.25144467e-01 -1.13349438e+00 -5.62269747e-01 9.29713547e-01
-7.31823564e-01 -2.56364614e-01 -3.92605752e-01 1.41684067e+00
-9.76444874e-03 1.21215475e+00 1.61908404e-03 3.17714095e-01
3.64444882e-01 5.33870198e-02 6.33199364e-02 -5.87950587e-01
-5.77321947e-01 -5.50264835e-01 4.99283254e-01 -7.17556119e-01
-4.41963553e-01 -1.26056993e+00 -1.16563487e+00 -6.32812023e-01
-2.21743539e-01 2.13294834e-01 4.88922656e-01 1.05878150e+00
4.38306600e-01 2.56035477e-01 3.44522327e-01 -7.89830267e-01
-6.31972492e-01 -9.58422422e-01 -3.49674761e-01 6.20115995e-01
1.44157156e-01 -8.21445107e-01 -4.98563021e-01 -2.51631498e-01] | [8.680935859680176, 5.662644863128662] |
847982b1-6a1d-4485-a6ae-97e3f273f9d4 | an-efficient-framework-for-zero-shot-sketch | 2102.04016 | null | https://arxiv.org/abs/2102.04016v1 | https://arxiv.org/pdf/2102.04016v1.pdf | An Efficient Framework for Zero-Shot Sketch-Based Image Retrieval | Recently, Zero-shot Sketch-based Image Retrieval (ZS-SBIR) has attracted the attention of the computer vision community due to it's real-world applications, and the more realistic and challenging setting than found in SBIR. ZS-SBIR inherits the main challenges of multiple computer vision problems including content-based Image Retrieval (CBIR), zero-shot learning and domain adaptation. The majority of previous studies using deep neural networks have achieved improved results through either projecting sketch and images into a common low-dimensional space or transferring knowledge from seen to unseen classes. However, those approaches are trained with complex frameworks composed of multiple deep convolutional neural networks (CNNs) and are dependent on category-level word labels. This increases the requirements on training resources and datasets. In comparison, we propose a simple and efficient framework that does not require high computational training resources, and can be trained on datasets without semantic categorical labels. Furthermore, at training and inference stages our method only uses a single CNN. In this work, a pre-trained ImageNet CNN (e.g., ResNet50) is fine-tuned with three proposed learning objects: domain-aware quadruplet loss, semantic classification loss, and semantic knowledge preservation loss. The domain-aware quadruplet and semantic classification losses are introduced to learn discriminative, semantic and domain invariant features through considering ZS-SBIR as object detection and verification problem. ... | ['Clinton Fookes', 'Ethan Goan', 'Sridha Sridharan', 'Simon Denman', 'Osman Tursun'] | 2021-02-08 | null | null | null | null | ['sketch-based-image-retrieval', 'content-based-image-retrieval'] | ['computer-vision', 'computer-vision'] | [ 3.71705681e-01 -4.71924841e-01 -4.80258197e-01 -2.87717193e-01
-7.90071666e-01 -3.76577139e-01 6.84805036e-01 -5.29429391e-02
-6.11012220e-01 4.99524713e-01 -1.93550661e-01 1.29309535e-01
-5.33779263e-01 -1.00299895e+00 -5.09515047e-01 -5.17386913e-01
4.20332372e-01 3.61574680e-01 4.87608910e-01 -2.88894296e-01
2.78779566e-01 5.33815265e-01 -1.87256348e+00 6.36074245e-02
5.96528709e-01 1.54060972e+00 2.67342448e-01 3.19085240e-01
-3.96162003e-01 6.45255744e-01 -4.33861822e-01 -5.03310621e-01
3.30900580e-01 -4.24581617e-02 -5.48802316e-01 -1.39356375e-01
6.56946659e-01 -4.86751199e-01 -6.33707464e-01 1.11285400e+00
6.84622169e-01 4.41948652e-01 7.58571625e-01 -1.57141244e+00
-1.25078106e+00 -1.34559339e-02 -2.96335548e-01 -2.20071543e-02
-7.96720758e-02 -1.30226254e-01 1.03070819e+00 -1.12457287e+00
7.36675799e-01 1.31352222e+00 6.00274026e-01 7.05788016e-01
-9.76865053e-01 -9.37256515e-01 1.04333393e-01 6.63851261e-01
-1.71675587e+00 -1.39400065e-01 1.00237072e+00 -3.67226690e-01
5.87756753e-01 4.59806621e-02 2.29283392e-01 1.22347569e+00
-4.02059644e-01 7.93700635e-01 7.20245600e-01 -3.53218645e-01
4.75448281e-01 2.30619729e-01 2.02029899e-01 4.94506299e-01
1.80418774e-01 -6.10347092e-02 -4.35910016e-01 -4.18107323e-02
9.19032037e-01 7.32279539e-01 -3.70703340e-02 -7.79959381e-01
-9.07067955e-01 1.03952980e+00 7.14169383e-01 4.87590969e-01
-5.35439467e-03 2.24622786e-01 6.69923246e-01 3.04966092e-01
4.41382170e-01 1.53666615e-01 -4.18470711e-01 3.20117623e-01
-9.66566443e-01 1.28862306e-01 4.08752799e-01 1.15402377e+00
8.97335470e-01 -5.45393154e-02 -3.61545622e-01 1.49059165e+00
3.68997119e-02 5.74450612e-01 6.18099093e-01 -6.19716585e-01
1.93226218e-01 7.13763773e-01 -1.08109720e-01 -1.20794046e+00
-2.49394868e-02 -2.49975428e-01 -1.06754005e+00 1.52036816e-01
3.53869162e-02 2.49326393e-01 -1.17105913e+00 1.71976471e+00
-3.14728953e-02 2.77162552e-01 1.67582054e-02 1.05062354e+00
1.16972494e+00 5.56050956e-01 1.51289150e-01 3.13138694e-01
1.21153963e+00 -1.06791604e+00 -6.01512790e-01 -8.26895516e-03
8.34082142e-02 -6.36892378e-01 1.23248839e+00 1.41631797e-01
-8.27353299e-01 -6.96567655e-01 -1.24500978e+00 -4.72209036e-01
-1.00035524e+00 2.04159901e-01 4.82003182e-01 4.67774928e-01
-9.27049160e-01 4.07641053e-01 -3.02517682e-01 -5.36558807e-01
6.11798584e-01 1.81446224e-01 -3.79741877e-01 -4.59829897e-01
-1.34036791e+00 7.39319384e-01 5.51201284e-01 -1.48972631e-01
-8.81538033e-01 -6.80171788e-01 -1.03682137e+00 3.56331915e-01
6.02073789e-01 -4.97793138e-01 9.64255989e-01 -1.07744288e+00
-1.43226409e+00 7.84963667e-01 1.67148575e-01 -1.49167359e-01
4.88029718e-01 -3.82604897e-02 -3.87603521e-01 2.06374347e-01
8.86907056e-02 7.40064681e-01 9.75570321e-01 -1.22148383e+00
-4.11259145e-01 -4.18156564e-01 3.09428215e-01 -2.64223199e-03
-8.42872202e-01 -5.54036675e-03 -8.02626550e-01 -8.84456992e-01
8.58326331e-02 -7.62022257e-01 1.64245982e-02 6.40224874e-01
-4.36679786e-03 -4.65638995e-01 1.13211060e+00 -3.95242512e-01
1.08122253e+00 -2.19300032e+00 -2.90166177e-02 6.90623522e-02
-3.31302322e-02 7.84997940e-01 -4.19313699e-01 3.78377616e-01
1.91929147e-01 -3.57357524e-02 -3.45442683e-01 -1.91918001e-01
1.50441572e-01 2.38514468e-01 -2.94711113e-01 1.34401783e-01
2.86355853e-01 1.10953486e+00 -9.10735965e-01 -5.68697155e-01
5.54033458e-01 6.14038050e-01 -5.65705180e-01 1.19024508e-01
-1.81354165e-01 3.77120636e-02 -3.14101696e-01 7.83713579e-01
8.15434873e-01 -4.07775372e-01 2.61685271e-02 -2.79807389e-01
1.95887223e-01 -3.60116422e-01 -1.18649709e+00 2.10588360e+00
-7.70734727e-01 4.94448692e-01 -2.38081902e-01 -1.33880985e+00
1.15910137e+00 1.95578128e-01 2.53164381e-01 -1.08473945e+00
9.01488885e-02 2.33054772e-01 -6.47466123e-01 -3.69873494e-01
4.32392269e-01 -2.60262936e-01 -2.29640469e-01 1.37091190e-01
4.84658539e-01 1.45200059e-01 -3.04760784e-02 1.51876897e-01
7.09358275e-01 7.55480900e-02 1.91896647e-01 1.72406062e-02
7.53647327e-01 -1.18645310e-01 5.61183751e-01 7.48980045e-01
-1.14553578e-01 8.14501762e-01 7.69496262e-02 -5.30309558e-01
-1.17500496e+00 -1.18234015e+00 -1.85923263e-01 1.36215878e+00
7.36503363e-01 -9.86612141e-02 -2.43974805e-01 -4.91499573e-01
2.18897328e-01 4.16448534e-01 -6.01534545e-01 -3.93758446e-01
-3.08992207e-01 -2.12573275e-01 4.79545683e-01 6.03627205e-01
6.83938622e-01 -1.14739001e+00 -3.87221962e-01 1.00447595e-01
-2.81418674e-02 -1.09267581e+00 -3.38689476e-01 -2.68767178e-01
-5.90069115e-01 -1.16723812e+00 -1.18250620e+00 -1.07980967e+00
5.86437762e-01 6.10408306e-01 9.11324143e-01 2.45388299e-02
-7.72228360e-01 5.22253990e-01 -5.05129457e-01 -3.58422101e-01
2.53805995e-01 -2.94339135e-02 -1.69323459e-01 1.89027444e-01
5.36088765e-01 -4.80330735e-01 -8.11996400e-01 4.68497664e-01
-1.17916417e+00 -3.01323324e-01 6.55526757e-01 1.22535336e+00
5.70254207e-01 -1.52818754e-01 7.70095170e-01 -6.62020445e-01
4.96173590e-01 -3.45053434e-01 -5.98868191e-01 6.66872323e-01
-5.67903221e-01 -9.46512967e-02 6.35635734e-01 -5.83910763e-01
-1.01424038e+00 -2.40247101e-01 6.40669540e-02 -1.09889483e+00
-5.67065850e-02 2.70970762e-01 -1.50645465e-01 -2.26378500e-01
4.59137797e-01 3.87154698e-01 8.07795301e-03 -4.98474240e-01
5.90838730e-01 8.72313917e-01 4.23221588e-01 -4.36538547e-01
7.92999744e-01 5.02029240e-01 -9.16922763e-02 -7.92403221e-01
-9.13513839e-01 -8.67686868e-01 -4.60428953e-01 1.07640743e-01
8.22500587e-01 -9.80081737e-01 -6.27127469e-01 5.10720015e-01
-1.15743923e+00 1.29623562e-01 -2.10955605e-01 2.43350551e-01
-4.70596671e-01 3.80061209e-01 -3.13335240e-01 -6.85075939e-01
-6.52526379e-01 -9.50596869e-01 1.16464365e+00 1.05011605e-01
3.82075101e-01 -8.15532684e-01 -7.19863325e-02 2.06050918e-01
6.77382708e-01 7.93861970e-02 1.00514197e+00 -6.44706190e-01
-6.90820634e-01 -3.60592365e-01 -9.30701256e-01 7.14756131e-01
7.64304399e-02 -3.08014780e-01 -9.64056611e-01 -4.10874456e-01
-4.01203573e-01 -6.84314787e-01 1.02761424e+00 1.18899442e-01
1.65843499e+00 -3.24274004e-01 -1.67503014e-01 5.88254094e-01
1.89195323e+00 2.49781683e-01 4.97817993e-01 1.87582105e-01
6.68686926e-01 4.92849588e-01 7.23421276e-01 4.03745532e-01
2.57032394e-01 8.56846869e-01 3.14370453e-01 -1.06698848e-01
-3.46607924e-01 -4.16796833e-01 -1.77580461e-01 5.67245662e-01
1.72491476e-01 -1.12154767e-01 -6.85282350e-01 8.20386112e-01
-2.03376079e+00 -1.06446421e+00 6.25023246e-01 2.22009921e+00
5.44892311e-01 -2.80112147e-01 -1.49325073e-01 1.52757689e-01
9.11584795e-01 2.30646253e-01 -7.54701793e-01 -1.06247365e-01
6.48723766e-02 4.52045053e-01 3.89595360e-01 8.29977989e-02
-1.21929979e+00 1.08025122e+00 5.05663872e+00 1.29773307e+00
-1.30608809e+00 4.26243514e-01 3.22193176e-01 2.14554220e-02
-9.75009277e-02 -3.43572974e-01 -6.39513671e-01 3.99560630e-01
3.48609000e-01 -2.34874532e-01 3.78471196e-01 1.16633058e+00
-3.98038238e-01 3.39803785e-01 -1.06214702e+00 1.49656498e+00
4.49629456e-01 -1.40104651e+00 5.07716477e-01 -3.32974195e-01
6.97741330e-01 -1.27418578e-01 1.94062710e-01 6.50311887e-01
1.26256779e-01 -9.90940630e-01 6.40192389e-01 4.87082869e-01
1.33585560e+00 -7.40895152e-01 6.45485520e-01 1.56777278e-01
-1.33016610e+00 -3.73544902e-01 -8.77113640e-01 1.62744567e-01
-2.11599812e-01 1.37494102e-01 -4.09969121e-01 6.01197124e-01
7.16778815e-01 8.29265654e-01 -3.58926982e-01 1.08140910e+00
4.04228307e-02 6.67160749e-02 -5.77855706e-02 -5.67207597e-02
3.19370300e-01 -1.01915449e-01 1.44007072e-01 9.86035287e-01
3.51870835e-01 5.57631776e-02 3.77436101e-01 8.95138025e-01
-2.70796448e-01 1.42653450e-01 -7.42597759e-01 8.76943991e-02
5.18990993e-01 1.34311795e+00 -4.34993327e-01 -5.09225130e-01
-5.65663934e-01 1.00470150e+00 3.12976509e-01 3.85774076e-01
-5.99420846e-01 -8.32170010e-01 7.54226863e-01 -7.42229298e-02
5.10293305e-01 -2.84384098e-02 -1.24554232e-01 -1.35726511e+00
-2.45627221e-02 -4.46266621e-01 4.29744840e-01 -6.72701955e-01
-1.70992017e+00 3.86602491e-01 -1.11488342e-01 -1.45413291e+00
7.46485442e-02 -8.51476908e-01 -4.52870727e-01 7.14840353e-01
-2.02297354e+00 -1.55246997e+00 -6.66036844e-01 1.07244408e+00
7.64186621e-01 -3.42299998e-01 8.46539080e-01 7.89525568e-01
-3.51035953e-01 9.95687544e-01 3.19908500e-01 4.02583361e-01
7.83450544e-01 -8.89484227e-01 1.37660339e-01 3.66222560e-01
7.83587061e-03 6.26216471e-01 1.02351680e-01 -2.85072446e-01
-1.42965138e+00 -1.37763441e+00 6.14059985e-01 -9.25067067e-02
6.26501203e-01 -5.40397406e-01 -7.88008928e-01 3.10721546e-01
-3.41435999e-01 4.43566293e-01 4.93856579e-01 4.55277115e-02
-8.73326421e-01 -4.14186686e-01 -1.30146182e+00 4.48048800e-01
1.24012673e+00 -8.76127660e-01 -5.03547549e-01 2.70547718e-01
7.39456892e-01 5.91826513e-02 -7.72229493e-01 5.25098324e-01
8.61726284e-01 -6.36752725e-01 1.44019854e+00 -6.94107771e-01
4.63067591e-01 -1.72054514e-01 -4.38174754e-01 -8.34442377e-01
-3.34586978e-01 2.31847346e-01 2.83934981e-01 1.09323895e+00
-1.64921116e-02 -5.86898208e-01 6.87981546e-01 4.07597154e-01
1.51963279e-01 -4.74507987e-01 -8.83545518e-01 -1.09175742e+00
1.35443330e-01 -2.02040091e-01 5.15817463e-01 1.01590550e+00
-3.59309852e-01 2.99196243e-01 -5.78101635e-01 -7.64031634e-02
7.32438564e-01 3.59698504e-01 6.84714913e-01 -1.55137694e+00
2.25107698e-03 -3.95628065e-01 -7.65097141e-01 -8.11594725e-01
2.17190966e-01 -1.08309019e+00 -7.86371827e-02 -1.52707446e+00
3.89787763e-01 -7.25070894e-01 -8.90442133e-01 6.22454047e-01
1.02930352e-01 5.58854878e-01 5.34790754e-01 2.80837387e-01
-9.55433249e-01 8.57404232e-01 1.14869094e+00 -5.85485876e-01
1.63131922e-01 -2.64826566e-01 -4.30049211e-01 5.47255695e-01
5.22294283e-01 -2.63479114e-01 -6.66829646e-01 -4.38190877e-01
-4.34529446e-02 -4.06059548e-02 7.89094627e-01 -9.49008942e-01
4.29992318e-01 -1.36005536e-01 3.49226415e-01 -5.95174670e-01
6.37879968e-01 -1.02560377e+00 -6.74730241e-02 3.00747514e-01
-4.26020354e-01 -4.68363464e-01 -7.81445764e-03 7.56402075e-01
-5.07027507e-01 -3.10713381e-01 9.08168375e-01 -2.99095273e-01
-1.29743445e+00 7.40969360e-01 2.72545040e-01 -7.68757612e-02
1.03334987e+00 -3.35451603e-01 -3.30476910e-01 -1.42935276e-01
-4.82757896e-01 7.27944449e-02 4.14671808e-01 8.10574055e-01
9.16991055e-01 -1.53392959e+00 -3.76355588e-01 2.00571388e-01
7.24138737e-01 -6.30025268e-02 5.58418751e-01 1.71886325e-01
-3.35871458e-01 6.27155483e-01 -5.31889260e-01 -4.94660646e-01
-1.01772106e+00 9.52417195e-01 -3.28649301e-04 -3.33648399e-02
-6.72746658e-01 7.77635753e-01 4.85439777e-01 -7.31367350e-01
4.24385250e-01 1.41545668e-01 -1.52531222e-01 1.18363537e-01
5.47582924e-01 4.57288265e-01 5.08703552e-02 -4.85108584e-01
-3.77129942e-01 9.28980827e-01 -1.43718049e-01 1.28325492e-01
1.19352913e+00 6.01928346e-02 -1.84526611e-02 3.58174175e-01
1.56569803e+00 -7.51911581e-01 -9.68285978e-01 -8.32215071e-01
-1.90493036e-02 -6.29202724e-01 1.50152296e-01 -8.77443433e-01
-1.10214686e+00 1.13373768e+00 8.72139573e-01 -2.63467073e-01
1.06870520e+00 -8.55149627e-02 1.03902602e+00 6.30279124e-01
6.94795072e-01 -1.25818741e+00 4.20013815e-01 4.59493995e-01
9.15051341e-01 -1.48666477e+00 -2.61731952e-01 -2.34875351e-01
-4.15684998e-01 9.41557705e-01 6.66531563e-01 -3.25897306e-01
8.05040836e-01 -4.26918507e-01 -1.22276425e-01 -4.87691425e-02
-4.52343166e-01 -2.71093637e-01 3.91959161e-01 5.47038972e-01
-7.22066835e-02 -4.16234285e-02 -3.50230873e-01 6.51669562e-01
3.94229621e-01 3.24620813e-01 -7.15433434e-02 8.23803961e-01
-4.80319530e-01 -1.02914476e+00 -4.20489311e-02 4.25472438e-01
-1.85629278e-01 -1.15110591e-01 -1.27201274e-01 5.58604181e-01
2.49060690e-01 7.58919239e-01 1.03917383e-01 -1.82503536e-01
3.83746564e-01 1.29796704e-02 3.13423038e-01 -5.59568286e-01
-1.18855380e-01 -4.33280349e-01 -5.31997144e-01 -4.64511275e-01
-4.02340055e-01 -4.92401496e-02 -8.07009578e-01 -1.56010956e-01
-3.70038778e-01 -1.29127920e-01 8.89568269e-01 6.87461972e-01
4.46978301e-01 3.08373988e-01 6.34226680e-01 -7.68796325e-01
-6.85676098e-01 -7.35241711e-01 -6.03902340e-01 7.15522230e-01
1.25104144e-01 -1.09766257e+00 -1.36674032e-01 -2.34053701e-01] | [11.528831481933594, 0.7692456841468811] |
06edc7e2-8bb2-435f-8ce0-7f469b525c36 | adversarial-representation-learning-for-3 | 2305.00011 | null | https://arxiv.org/abs/2305.00011v1 | https://arxiv.org/pdf/2305.00011v1.pdf | Adversarial Representation Learning for Robust Privacy Preservation in Audio | Sound event detection systems are widely used in various applications such as surveillance and environmental monitoring where data is automatically collected, processed, and sent to a cloud for sound recognition. However, this process may inadvertently reveal sensitive information about users or their surroundings, hence raising privacy concerns. In this study, we propose a novel adversarial training method for learning representations of audio recordings that effectively prevents the detection of speech activity from the latent features of the recordings. The proposed method trains a model to generate invariant latent representations of speech-containing audio recordings that cannot be distinguished from non-speech recordings by a speech classifier. The novelty of our work is in the optimization algorithm, where the speech classifier's weights are regularly replaced with the weights of classifiers trained in a supervised manner. This increases the discrimination power of the speech classifier constantly during the adversarial training, motivating the model to generate latent representations in which speech is not distinguishable, even using new speech classifiers trained outside the adversarial training loop. The proposed method is evaluated against a baseline approach with no privacy measures and a prior adversarial training method, demonstrating a significant reduction in privacy violations compared to the baseline approach. Additionally, we show that the prior adversarial method is practically ineffective for this purpose. | ['Tuomas Virtanen', 'Konstantinos Drossos', 'Diep Luong', 'Minh Tran', 'Shayan Gharib'] | 2023-04-29 | null | null | null | null | ['sound-event-detection'] | ['audio'] | [ 9.08704758e-01 3.54906440e-01 4.17553037e-01 -1.56918690e-01
-7.16221571e-01 -8.41598630e-01 4.53467190e-01 2.44084448e-01
-4.33117777e-01 4.43807721e-01 7.63598979e-02 -7.74661079e-02
2.08644181e-01 -8.98123026e-01 -7.17343748e-01 -9.83480573e-01
1.22163016e-02 -1.87601432e-01 1.31278858e-01 1.75433800e-01
-8.42969492e-02 7.07909882e-01 -1.76113737e+00 3.59689057e-01
5.42618036e-01 9.70830858e-01 -4.15911257e-01 7.64013231e-01
2.05121204e-01 5.72128654e-01 -1.17571676e+00 -3.94431233e-01
5.28931022e-01 -2.44383827e-01 -3.49263757e-01 -7.76461437e-02
3.38774383e-01 -3.54640245e-01 -4.19283956e-01 1.32999206e+00
6.32675350e-01 2.83644915e-01 2.36415371e-01 -1.34170043e+00
-2.52912104e-01 3.37606579e-01 7.44826123e-02 6.04203567e-02
5.22126496e-01 -9.81161185e-03 6.06023431e-01 -4.13857430e-01
1.04754813e-01 9.22197044e-01 4.31854188e-01 6.72414124e-01
-1.05646908e+00 -1.12757206e+00 1.51383072e-01 -6.60151988e-02
-1.58386791e+00 -9.20948327e-01 9.89540458e-01 -2.65036196e-01
6.24148190e-01 9.93797719e-01 3.31865728e-01 1.27355349e+00
-1.05304629e-01 4.62778538e-01 9.64613259e-01 -4.06779647e-01
6.07903063e-01 6.32424712e-01 -2.01250434e-01 3.28180581e-01
6.53881161e-03 4.00972188e-01 -5.20887494e-01 -8.46366346e-01
2.35048160e-01 1.19841062e-01 -3.72035950e-01 -1.97052345e-01
-7.56760418e-01 6.14201963e-01 -9.87126864e-03 3.15964848e-01
-3.60751748e-01 -1.80506676e-01 3.97947818e-01 2.99716502e-01
6.46268308e-01 2.68384963e-01 -8.69341642e-02 -1.63826551e-02
-8.14420760e-01 -1.07748486e-01 8.45463037e-01 8.82979631e-01
3.09153855e-01 3.24080795e-01 -2.20518529e-01 5.53650379e-01
6.42382279e-02 5.75612009e-01 6.02140188e-01 -4.74417597e-01
4.18054730e-01 3.15186411e-01 5.28219044e-02 -1.42530465e+00
1.24513499e-01 -4.36369658e-01 -7.16026843e-01 1.27779663e-01
1.10265844e-01 -2.99182653e-01 -4.89057600e-01 1.62344885e+00
3.86561543e-01 6.96887791e-01 3.85350168e-01 5.82913816e-01
5.03905773e-01 7.17588246e-01 1.56304594e-02 -4.97925311e-01
9.92023289e-01 -5.52237749e-01 -9.29655373e-01 -2.37074256e-01
1.39756814e-01 -6.10123575e-01 9.11623299e-01 4.60651189e-01
-7.15751052e-01 -3.42904747e-01 -1.15636861e+00 4.98310953e-01
-5.28327882e-01 -6.41582459e-02 3.23921204e-01 1.20627177e+00
-6.50340438e-01 2.13281319e-01 -5.46325803e-01 -1.26213608e-02
3.85183036e-01 5.74383497e-01 -4.73081976e-01 3.02171379e-01
-1.39305806e+00 2.68826723e-01 1.92902192e-01 1.36522070e-01
-1.00072610e+00 -3.03488880e-01 -8.29283893e-01 2.83711016e-01
2.18414709e-01 -7.82859549e-02 9.45569932e-01 -1.18770480e+00
-1.50264323e+00 6.13155127e-01 2.15786118e-02 -7.08961189e-01
7.15826750e-01 -1.18631683e-01 -9.03311849e-01 1.24278605e-01
-1.39496788e-01 -1.37160588e-02 1.69886887e+00 -1.10846925e+00
-4.46371794e-01 -3.65448058e-01 -6.49946928e-02 -1.23294823e-01
-7.91773558e-01 1.57275245e-01 -1.62314419e-02 -8.46639752e-01
-6.95810616e-02 -8.90372515e-01 -8.11005086e-02 -1.18400929e-02
-3.08805317e-01 3.88925016e-01 1.11756206e+00 -5.36821783e-01
1.16377902e+00 -2.74498868e+00 -4.85711634e-01 5.03974259e-01
-3.19664292e-02 5.53058147e-01 1.82286389e-02 1.58759207e-01
-3.06064576e-01 1.78521276e-01 -4.22174126e-01 -4.92538840e-01
-2.19763234e-01 9.41410586e-02 -7.67765045e-01 6.58086836e-01
1.50003448e-01 1.58652261e-01 -7.87452400e-01 -1.41479611e-01
2.29604140e-01 5.65087497e-01 -4.42050189e-01 4.96231705e-01
6.62700385e-02 4.69695300e-01 -3.89623493e-01 4.26134646e-01
7.91478276e-01 5.13111472e-01 4.07423861e-02 2.44605601e-01
8.58052745e-02 3.09725851e-01 -1.25873864e+00 9.71265793e-01
-5.98013163e-01 5.69058239e-01 3.38735193e-01 -9.25440907e-01
1.07388544e+00 8.16785216e-01 2.44901881e-01 -3.41582924e-01
2.16300651e-01 -1.70736805e-01 -2.31243342e-01 -4.80187327e-01
2.28763714e-01 7.62211755e-02 -2.39172727e-01 3.80289704e-01
-2.49748409e-01 1.79901674e-01 -7.13845074e-01 -8.90498161e-02
1.18248081e+00 -5.70831656e-01 1.16506904e-01 3.05832088e-01
6.87476099e-01 -6.56568348e-01 6.76336646e-01 7.62982845e-01
-3.67973596e-01 3.93412709e-01 1.44740343e-01 -1.49638042e-01
-5.32303154e-01 -9.56924498e-01 -1.68366820e-01 9.71730649e-01
-1.16149150e-01 -2.56495267e-01 -9.14594412e-01 -1.07918739e+00
-4.96219285e-02 8.35978627e-01 -6.49584532e-01 -6.43975198e-01
-2.95612425e-01 -4.84330088e-01 1.05681741e+00 1.25534266e-01
4.98953432e-01 -1.01547062e+00 -4.90010053e-01 1.52347445e-01
-7.82187581e-02 -1.18749404e+00 -4.83020574e-01 1.81614771e-01
-4.79304343e-01 -7.32811391e-01 -1.88239649e-01 -4.89115655e-01
8.54656160e-01 1.52055889e-01 3.20373952e-01 -2.81203501e-02
-2.45378748e-01 6.24595940e-01 -4.54635143e-01 -8.33233714e-01
-8.98681939e-01 -2.60698020e-01 3.13323110e-01 8.38132679e-01
4.06087905e-01 -7.36991405e-01 -3.04283321e-01 2.89419651e-01
-1.05864787e+00 -6.58121586e-01 2.00796038e-01 6.20587528e-01
3.40618789e-01 4.65215981e-01 7.39365995e-01 -1.04432917e+00
7.12828219e-01 -6.22525275e-01 -5.63238442e-01 1.37179807e-01
-3.07502240e-01 -2.67444342e-01 8.50632310e-01 -8.24892282e-01
-8.86912942e-01 4.40323979e-01 -2.28993207e-01 -4.38271999e-01
-4.21751648e-01 -2.08273709e-01 -6.32070005e-01 -3.11145633e-01
6.49371862e-01 4.72719252e-01 -8.34492147e-02 -3.62331659e-01
3.86376493e-02 1.20228934e+00 4.83149201e-01 -3.72763760e-02
9.91076052e-01 5.04063964e-01 -2.71755248e-01 -1.10478711e+00
-4.30336714e-01 -5.02329588e-01 -3.36127311e-01 -1.71853796e-01
5.90890765e-01 -7.41725862e-01 -6.16840601e-01 4.35628384e-01
-1.23067594e+00 4.34741348e-01 -4.09025371e-01 4.82677162e-01
-2.18522102e-01 5.04335165e-01 -6.87924623e-02 -1.46398437e+00
-4.19650018e-01 -9.00269032e-01 9.15818632e-01 -1.48318827e-01
-2.62322783e-01 -6.04671836e-01 -3.44930775e-03 3.86972725e-01
2.78546423e-01 4.72516477e-01 5.31659484e-01 -1.30943322e+00
-3.09727073e-01 -8.47115993e-01 3.68420690e-01 7.56729424e-01
4.89646524e-01 -1.48493782e-01 -1.74500573e+00 -3.88587207e-01
6.64385319e-01 1.01399831e-01 4.01890963e-01 -2.25573257e-01
1.21260226e+00 -9.71279621e-01 -8.08621421e-02 5.70311725e-01
8.83041799e-01 5.76326430e-01 6.23619020e-01 2.76828688e-02
4.04893935e-01 6.46411180e-01 3.91521990e-01 5.35885930e-01
-4.12835479e-01 5.47702670e-01 5.99381030e-01 -4.11240309e-02
4.59344953e-01 -3.72563779e-01 6.87981188e-01 4.96600389e-01
2.83914089e-01 -3.22106540e-01 -4.11539048e-01 4.64903742e-01
-1.55235708e+00 -1.15558815e+00 2.84541398e-01 2.60076308e+00
6.41244948e-01 1.27823830e-01 -1.34114563e-01 6.69062495e-01
8.76433492e-01 2.61474669e-01 -4.65529740e-01 -6.69408321e-01
2.06452757e-01 5.04099071e-01 5.21561682e-01 3.30436438e-01
-1.51407194e+00 6.96048737e-01 6.25399494e+00 5.98539412e-01
-1.23827088e+00 2.49895543e-01 1.70634195e-01 -2.27993950e-01
-1.93658262e-01 -2.87483901e-01 -3.47223788e-01 5.48942208e-01
1.00797343e+00 -7.82216266e-02 4.70333308e-01 1.02566004e+00
2.58962780e-01 3.66883546e-01 -9.99957681e-01 9.67435956e-01
2.85982251e-01 -9.33314800e-01 8.23007002e-02 1.17076524e-01
2.18995944e-01 -5.00195146e-01 3.13186258e-01 8.22237283e-02
-7.46173179e-03 -8.15933883e-01 6.12513423e-01 4.08814818e-01
6.21151268e-01 -1.06470037e+00 6.96741164e-01 6.24704182e-01
-9.45950687e-01 -2.52724081e-01 -2.35561728e-01 -3.30777131e-02
-2.39528760e-01 4.40310538e-01 -1.16110456e+00 3.33767176e-01
7.00289428e-01 4.92732450e-02 -4.31605607e-01 6.06470704e-01
-2.04708204e-01 9.13691521e-01 -2.88522571e-01 -2.24439390e-02
-1.98910505e-01 -1.01697110e-02 9.95522559e-01 1.11681759e+00
2.72023082e-01 -6.17234036e-02 1.05764642e-02 6.67204976e-01
-1.63300291e-01 2.44322345e-01 -1.16333449e+00 -1.92814067e-01
7.41855741e-01 8.99188936e-01 -2.75122613e-01 4.82141487e-02
-1.14478208e-01 1.08708811e+00 -5.69095872e-02 1.54236540e-01
-7.62433887e-01 -5.50947785e-01 6.59656465e-01 1.47464156e-01
1.17514201e-01 3.48894335e-02 6.21365756e-02 -8.78664851e-01
2.97770321e-01 -9.56522942e-01 3.54528844e-01 -4.47570801e-01
-1.12429750e+00 9.28197443e-01 -2.96164453e-01 -1.49979651e+00
-3.43402952e-01 1.50809914e-01 -6.41042471e-01 7.39243329e-01
-1.11195862e+00 -9.56312299e-01 -8.10501575e-02 9.00166929e-01
2.19640806e-01 -5.01592875e-01 1.44156206e+00 3.34441215e-01
-5.23400128e-01 1.14462829e+00 1.76188707e-01 4.76121575e-01
5.29050887e-01 -7.62622237e-01 3.99630994e-01 1.20940995e+00
5.09014845e-01 7.05158234e-01 6.46288276e-01 -5.64485610e-01
-1.00565767e+00 -1.41901588e+00 7.44489789e-01 -2.88759589e-01
3.54181617e-01 -9.02267098e-01 -9.95121896e-01 5.24102628e-01
-7.13465139e-02 1.80838808e-01 1.24089062e+00 -2.42258981e-01
-5.19233286e-01 -2.74758995e-01 -1.74595463e+00 3.13240826e-01
5.30955970e-01 -1.01665163e+00 -5.77776313e-01 2.83021778e-01
9.28969860e-01 -2.75831390e-02 -4.70209688e-01 6.94874525e-02
4.84890848e-01 -6.13420904e-01 8.72067630e-01 -5.59280097e-01
-3.91258359e-01 -3.41934413e-01 -1.73899174e-01 -8.73857081e-01
1.57152310e-01 -1.06304908e+00 -2.06547037e-01 1.64447832e+00
1.98044419e-01 -9.81757998e-01 7.76363730e-01 7.08301902e-01
2.71636873e-01 -6.47100583e-02 -1.33322501e+00 -7.61259735e-01
-4.41014498e-01 -6.38876021e-01 8.89808476e-01 1.10546565e+00
-2.49533877e-01 -1.34607345e-01 -4.90230829e-01 9.44964170e-01
4.85286951e-01 -3.56118321e-01 7.81113207e-01 -1.02704048e+00
-3.23462158e-01 3.27500999e-01 -8.22036386e-01 -5.60955465e-01
2.64162511e-01 -8.03719521e-01 1.97207406e-01 -5.57803512e-01
-4.77643192e-01 -3.88061434e-01 -3.98975849e-01 5.23258686e-01
7.56829306e-02 -3.04348115e-03 3.86642702e-02 -3.19594122e-03
-2.02680044e-02 5.25186181e-01 5.30088484e-01 -3.47693592e-01
-3.18320751e-01 8.08696985e-01 -5.94168663e-01 7.47095227e-01
9.01963651e-01 -1.00380051e+00 -5.62340260e-01 2.26276517e-02
-1.10807121e-01 -8.65016729e-02 6.02179706e-01 -1.14783132e+00
2.47777283e-01 5.58558330e-02 1.35552168e-01 -1.43055901e-01
4.54180747e-01 -1.41652250e+00 1.56362966e-01 2.12695077e-01
-5.22509933e-01 -3.49249661e-01 3.33021611e-01 9.03293490e-01
-1.83805779e-01 -3.61768901e-01 6.23370409e-01 1.30585179e-01
-1.03080139e-01 1.22696504e-01 -7.47760952e-01 -2.90678591e-01
1.21880591e+00 -3.19315970e-01 7.33414432e-03 -5.85239291e-01
-9.42674339e-01 -3.21609259e-01 1.50894865e-01 4.86537397e-01
6.79142654e-01 -1.07120109e+00 -4.57611978e-01 7.84946203e-01
6.61902577e-02 -1.78312704e-01 2.38997132e-01 2.29858741e-01
-8.08929354e-02 -2.33154520e-02 1.15443327e-01 -3.32858056e-01
-1.73869908e+00 8.99984360e-01 2.37699404e-01 1.24580748e-01
-5.38043082e-01 6.64347529e-01 6.98312223e-02 -3.39635581e-01
7.19251931e-01 -2.47520015e-01 -1.59004733e-01 -5.26813455e-02
6.86579227e-01 3.16963345e-01 2.91006237e-01 -7.27601349e-01
-4.38724607e-01 2.98302229e-02 5.42394854e-02 -1.19933799e-01
1.04131973e+00 6.49709925e-02 1.07286379e-01 2.62287199e-01
1.35942280e+00 7.47745574e-01 -8.43258381e-01 -7.29899108e-02
-2.96645015e-01 -7.66292810e-01 -8.08812492e-03 -4.55630302e-01
-9.59831774e-01 9.10780668e-01 1.04783463e+00 6.86375141e-01
1.37136638e+00 -4.19555694e-01 6.67258203e-01 3.94760132e-01
4.10107523e-01 -9.80665922e-01 -2.34842032e-01 7.08284080e-02
8.90990257e-01 -8.58172238e-01 -2.88140029e-01 -4.93731529e-01
-6.32042170e-01 7.74690747e-01 1.84249595e-01 2.37666499e-02
7.81947553e-01 4.88137513e-01 1.25279501e-01 2.30189160e-01
-4.61433053e-01 1.98053777e-01 9.49680656e-02 9.79190350e-01
-2.74259925e-01 6.53583705e-02 2.81045139e-01 9.20533121e-01
-4.04461175e-01 -3.12377095e-01 3.62008154e-01 1.12426126e+00
-1.38835251e-01 -1.04612947e+00 -5.95089853e-01 1.41178459e-01
-6.57048643e-01 -6.81694895e-02 -8.18098128e-01 2.30740830e-01
3.27689111e-01 1.36799228e+00 1.30382687e-01 -6.50242925e-01
5.66469729e-01 3.00782710e-01 -1.76003367e-01 -6.64374173e-01
-9.91886735e-01 -1.10198304e-01 9.17822123e-03 -3.11267585e-01
-2.52526820e-01 -8.14445794e-01 -9.86119092e-01 3.48253138e-02
-4.71574962e-01 4.46094781e-01 6.74979806e-01 6.98924899e-01
4.67501909e-01 5.91317534e-01 1.37332392e+00 -2.86694497e-01
-6.76568866e-01 -6.14263833e-01 -4.37350959e-01 3.07658792e-01
7.03597844e-01 -3.80355090e-01 -8.36490631e-01 1.39381707e-01] | [13.98108196258545, 5.822329044342041] |
649eca70-954d-4415-9fde-5a5b3f8669f4 | an-entity-based-claim-extraction-pipeline-for | 2304.05268 | null | https://arxiv.org/abs/2304.05268v1 | https://arxiv.org/pdf/2304.05268v1.pdf | An Entity-based Claim Extraction Pipeline for Real-world Biomedical Fact-checking | Existing fact-checking models for biomedical claims are typically trained on synthetic or well-worded data and hardly transfer to social media content. This mismatch can be mitigated by adapting the social media input to mimic the focused nature of common training claims. To do so, Wuehrl & Klinger (2022) propose to extract concise claims based on medical entities in the text. However, their study has two limitations: First, it relies on gold-annotated entities. Therefore, its feasibility for a real-world application cannot be assessed since this requires detecting relevant entities automatically. Second, they represent claim entities with the original tokens. This constitutes a terminology mismatch which potentially limits the fact-checking performance. To understand both challenges, we propose a claim extraction pipeline for medical tweets that incorporates named entity recognition and terminology normalization via entity linking. We show that automatic NER does lead to a performance drop in comparison to using gold annotations but the fact-checking performance still improves considerably over inputting the unchanged tweets. Normalizing entities to their canonical forms does, however, not improve the performance. | ['Roman Klinger', 'Lara Grimminger', 'Amelie Wührl'] | 2023-04-11 | null | null | null | null | ['entity-linking'] | ['natural-language-processing'] | [ 4.21449721e-01 6.89918637e-01 -1.83121413e-01 6.89677373e-02
-1.18992281e+00 -6.07080698e-01 6.40144229e-01 9.49811697e-01
-7.79364824e-01 9.83891904e-01 3.70120734e-01 -4.31713670e-01
6.62187040e-02 -7.90325701e-01 -5.03769040e-01 -7.68148974e-02
3.84886891e-01 4.68827307e-01 5.12888670e-01 -2.80585557e-01
1.42078876e-01 2.44011194e-01 -1.16352963e+00 4.92325425e-01
1.01549065e+00 5.09554267e-01 -2.14980692e-01 3.66452038e-01
-4.11207259e-01 9.61926043e-01 -9.89791036e-01 -1.07793558e+00
5.54496609e-02 -4.19368744e-01 -1.05298746e+00 -1.77365720e-01
2.44278893e-01 7.82252625e-02 -4.98792231e-02 1.12099302e+00
3.47299904e-01 -4.34979111e-01 7.20022440e-01 -8.42963338e-01
-6.20320857e-01 8.80571842e-01 -2.80379653e-01 1.75987910e-02
5.23608267e-01 -2.71718591e-01 1.02329302e+00 -6.54407084e-01
1.17004347e+00 5.51378727e-01 1.17822635e+00 5.75723588e-01
-1.25162673e+00 -3.93005013e-01 -2.39685878e-01 -8.46785456e-02
-1.35759246e+00 -4.56329852e-01 4.92357850e-01 -6.14950657e-01
8.34360480e-01 4.50316995e-01 3.83299291e-01 1.08047879e+00
-2.02736825e-01 5.36472380e-01 1.10027456e+00 -6.30281031e-01
5.58728650e-02 4.00396556e-01 5.83741739e-02 5.99277794e-01
6.52132392e-01 -5.20805657e-01 -2.49840319e-01 -5.00585377e-01
1.24323383e-01 -3.67503762e-01 -2.21839175e-01 -4.98671317e-03
-1.36265576e+00 7.41397083e-01 1.21865116e-01 8.15922081e-01
-6.24767542e-01 -2.91900188e-01 6.74941540e-01 1.89434752e-01
6.44012570e-01 9.04185593e-01 -4.90194559e-01 1.60060935e-02
-1.40526187e+00 3.76346320e-01 1.03227293e+00 7.82793403e-01
4.12617862e-01 -4.94392037e-01 -3.24899137e-01 6.23623013e-01
7.66713638e-03 3.96270633e-01 3.91496599e-01 -5.39010942e-01
5.88763297e-01 9.22238052e-01 2.92286754e-01 -1.31068873e+00
-6.61280453e-01 -4.85824704e-01 -4.69948649e-01 -8.54738876e-02
8.13179433e-01 -2.87493974e-01 -7.02858210e-01 1.56577837e+00
3.72015268e-01 -1.85870957e-02 3.88712198e-01 5.12208462e-01
9.77617025e-01 -4.95083481e-02 4.75519449e-01 -2.54714638e-01
1.70630705e+00 -5.08309662e-01 -1.03785694e+00 -1.52414531e-01
9.63654101e-01 -1.12514985e+00 7.23936081e-01 1.83385089e-02
-9.97792423e-01 9.60766152e-02 -9.92547631e-01 -1.10015906e-01
-6.24512911e-01 -5.94288344e-03 3.92178863e-01 9.82870877e-01
-5.86897075e-01 5.54415226e-01 -8.69709074e-01 -6.00101352e-01
5.14486015e-01 2.20109865e-01 -4.82486010e-01 2.86916316e-01
-1.50466824e+00 1.08786976e+00 5.42337298e-01 -4.54839975e-01
2.03388244e-01 -8.82094920e-01 -8.90766263e-01 -1.71605796e-01
5.09038508e-01 -7.20320880e-01 1.05486953e+00 -8.56602252e-01
-9.82055664e-01 1.20927215e+00 5.65034635e-02 -6.57468081e-01
7.34035492e-01 4.64769378e-02 -8.21846485e-01 1.95206538e-01
5.47764063e-01 1.23711124e-01 2.51011223e-01 -9.69645321e-01
-4.56009477e-01 -4.46924120e-02 1.64179262e-02 -1.69227526e-01
-4.38912600e-01 4.28663969e-01 -2.61960834e-01 -9.91189718e-01
-7.58312047e-02 -9.15844977e-01 -3.29094142e-01 -3.65292102e-01
-6.61557913e-01 3.11348364e-02 3.38688731e-01 -7.83799648e-01
1.41166520e+00 -1.91893315e+00 -5.40903509e-01 1.94428027e-01
3.22726935e-01 2.91717798e-01 9.44883376e-02 5.75449526e-01
-3.63239087e-03 5.54105699e-01 -3.82222414e-01 -2.51090676e-01
-7.56801590e-02 -5.12644127e-02 -2.18195438e-01 3.49323094e-01
4.97158855e-01 9.25691247e-01 -1.04160440e+00 -8.34033549e-01
-3.74094695e-01 3.19684565e-01 -6.91306412e-01 -2.57998437e-01
-9.90782082e-02 1.85034186e-01 -4.34387982e-01 3.70369941e-01
5.33632517e-01 -3.50085825e-01 4.46805596e-01 -3.94727975e-01
-1.37826428e-02 6.80416942e-01 -1.02022636e+00 1.56701362e+00
-2.49907508e-01 5.16222715e-01 -2.66898811e-01 -6.99240506e-01
5.40389895e-01 5.83320677e-01 8.20441484e-01 -5.33345103e-01
1.23561807e-02 5.90222895e-01 -1.67595088e-01 -6.84079111e-01
7.70383596e-01 -5.55544555e-01 -3.42230767e-01 4.99525994e-01
-1.73828959e-01 1.38835251e-01 3.19800049e-01 3.13910365e-01
1.36468709e+00 -3.17102810e-03 5.31902313e-01 -1.05031051e-01
4.02779818e-01 5.28363466e-01 6.59020662e-01 6.56201124e-01
8.34060647e-03 6.76012099e-01 5.94043970e-01 -3.46100121e-03
-9.97814059e-01 -6.50890350e-01 -2.52579361e-01 6.23497546e-01
-1.99213341e-01 -7.10735798e-01 -1.05303705e+00 -1.10397732e+00
-6.25820830e-02 7.18029559e-01 -7.15011716e-01 1.46234095e-01
-5.92221797e-01 -8.82998288e-01 1.13402236e+00 1.84592128e-01
1.39991269e-01 -8.79337192e-01 -7.96161234e-01 4.97940123e-01
-5.35134017e-01 -1.41048205e+00 -2.14951262e-01 -1.53272137e-01
-4.81968433e-01 -1.34008956e+00 -7.73043871e-01 -4.27406102e-01
6.16656780e-01 -3.22868854e-01 1.17066932e+00 2.07350314e-01
-7.09409714e-02 3.43676880e-02 -3.49653631e-01 -7.17804432e-01
-8.95265341e-01 4.25907701e-01 -1.31684572e-01 -2.09438041e-01
6.07179165e-01 -2.35938326e-01 -4.03650075e-01 2.73928996e-02
-1.10338843e+00 -8.13634247e-02 5.63535690e-01 6.89176857e-01
3.46871048e-01 -2.71329939e-01 8.24137390e-01 -1.58634257e+00
7.14856625e-01 -4.75941986e-01 -8.10593292e-02 3.57923329e-01
-8.00241411e-01 2.51716703e-01 3.98684621e-01 -5.34308314e-01
-7.86888838e-01 -3.57347950e-02 -3.81045938e-01 2.68202782e-01
-1.07769221e-01 7.70515203e-01 1.56017989e-01 2.69611090e-01
9.25735354e-01 -2.46655211e-01 -1.73977956e-01 -4.78402734e-01
3.75819743e-01 8.24076176e-01 4.28853780e-01 -3.60274762e-01
9.33768809e-01 4.79766488e-01 -2.51581669e-01 -4.28977400e-01
-1.21531057e+00 -4.35083330e-01 -4.97886777e-01 2.51187205e-01
1.07007122e+00 -7.63580441e-01 -5.23697019e-01 1.76662311e-03
-1.22905934e+00 1.51807731e-02 -4.52729762e-01 5.42660534e-01
-2.41480663e-01 5.95433295e-01 -7.49363482e-01 -6.28335953e-01
-2.99900264e-01 -7.85921812e-01 9.35326338e-01 -2.75212169e-01
-7.56826878e-01 -9.00535583e-01 2.62358785e-01 6.23609781e-01
2.62395382e-01 7.37677157e-01 9.54332590e-01 -1.18266404e+00
9.62892100e-02 -4.30582196e-01 -2.92967588e-01 1.69221200e-02
3.16615343e-01 -1.67663932e-01 -8.83477271e-01 1.56513423e-01
-9.27212909e-02 -1.86276436e-01 6.26224816e-01 -2.44469926e-01
3.53717536e-01 -5.88761032e-01 -3.69194865e-01 -1.87607288e-01
1.20266473e+00 -2.81390995e-01 5.04574239e-01 7.57140100e-01
5.46490252e-01 7.88765848e-01 3.85197461e-01 8.63403752e-02
5.39937496e-01 6.22605622e-01 2.67553069e-02 -2.61484653e-01
-1.79916322e-01 -2.50687838e-01 1.10408939e-01 6.92571878e-01
4.31464687e-02 -1.14198141e-01 -1.11480045e+00 7.82518089e-01
-1.63701010e+00 -1.16631222e+00 -6.26717031e-01 2.05756497e+00
1.36735237e+00 4.83212143e-01 2.48672098e-01 3.36244613e-01
6.57858372e-01 -1.89746574e-01 4.76046605e-03 -1.00127541e-01
-3.47811490e-01 2.52327234e-01 7.48528421e-01 2.76597261e-01
-9.47193861e-01 6.21990323e-01 5.89761353e+00 6.85769498e-01
-9.67488170e-01 5.94012201e-01 2.48479664e-01 1.79995701e-01
-4.69553143e-01 7.23204911e-02 -5.97468615e-01 4.14557576e-01
1.11710751e+00 -1.39589027e-01 -2.22032234e-01 4.93450314e-01
-1.08859045e-02 2.38827005e-01 -7.38635182e-01 7.37247586e-01
2.46157154e-01 -1.47916389e+00 -2.22314268e-01 1.75506279e-01
6.72830462e-01 1.32649010e-02 -1.76383585e-01 2.93224871e-01
1.34936184e-01 -6.56951189e-01 1.01921177e+00 5.24633348e-01
7.62954950e-01 -4.54400778e-01 1.07339454e+00 2.16692001e-01
-6.86764061e-01 3.06336999e-01 -8.63367021e-02 2.72749543e-01
3.85540217e-01 8.56182635e-01 -1.22674882e+00 8.06180537e-01
1.36774331e-01 1.57693699e-01 -7.56613553e-01 9.43519175e-01
-2.35437989e-01 6.44201517e-01 -4.15383756e-01 6.53000623e-02
1.14995688e-01 4.31536704e-01 5.49524963e-01 1.64422369e+00
2.31606930e-01 -1.00123562e-01 1.29168108e-01 7.58991480e-01
-4.29696798e-01 6.90534174e-01 -4.69793081e-01 -3.34372103e-01
3.02207589e-01 1.19532382e+00 -9.72371936e-01 -5.30837715e-01
-6.14618957e-01 8.68759036e-01 3.18316042e-01 -8.28262419e-03
-7.38023281e-01 -5.17224193e-01 2.82443643e-01 5.92938364e-01
1.77860126e-01 1.63619354e-01 -3.80893171e-01 -1.29237413e+00
1.63544595e-01 -9.91236925e-01 5.74407876e-01 -3.19706321e-01
-1.27485597e+00 8.23929727e-01 -1.96088389e-01 -1.27239251e+00
-7.38183185e-02 -3.32396090e-01 -7.87022635e-02 6.65540755e-01
-1.43618941e+00 -1.27393627e+00 1.13163717e-01 2.59026021e-01
1.04724087e-01 2.82251894e-01 9.68739986e-01 6.70942724e-01
-4.59483266e-01 8.27139616e-01 -2.77576119e-01 5.20190299e-01
1.11768854e+00 -1.28352153e+00 4.41189021e-01 7.98081636e-01
2.36972272e-01 9.76831317e-01 1.01423371e+00 -1.00508976e+00
-6.97903454e-01 -1.08110070e+00 1.68371642e+00 -1.00131369e+00
1.02104962e+00 -2.36921534e-01 -1.14684892e+00 3.88407826e-01
2.74650514e-01 -2.19802380e-01 1.14379668e+00 1.64483726e-01
-6.72386706e-01 5.56835592e-01 -1.23544872e+00 6.17574453e-01
8.32497656e-01 -7.13235140e-01 -9.21112537e-01 5.47298372e-01
5.96752286e-01 -3.88155609e-01 -1.26043165e+00 3.08466554e-01
3.16804618e-01 -6.06485128e-01 5.97009897e-01 -8.09370160e-01
3.60355020e-01 -3.65420699e-01 1.00123793e-01 -7.60707915e-01
8.08445290e-02 -5.16005814e-01 7.75592625e-02 1.60217345e+00
1.13530266e+00 -5.79562724e-01 6.20157301e-01 8.71626377e-01
1.46857157e-01 -5.14238298e-01 -7.79503703e-01 -7.18263924e-01
1.10460296e-01 -5.34227788e-01 6.80738747e-01 1.52388871e+00
6.12192631e-01 3.84865880e-01 -6.42597303e-02 7.70018473e-02
2.39325210e-01 -6.96155056e-02 4.52903956e-01 -1.31304848e+00
-2.55511373e-01 -3.09801936e-01 -4.06613737e-01 -2.90923357e-01
8.04610103e-02 -9.84404683e-01 -5.81450649e-02 -1.62431085e+00
3.14194560e-01 -7.65185118e-01 -2.48021483e-01 6.75107479e-01
-4.42270070e-01 6.35404170e-01 8.60411301e-02 4.72012013e-01
-6.16559446e-01 -1.20425999e-01 6.60100460e-01 -9.27860215e-02
-1.87183112e-01 -2.38408729e-01 -9.99123991e-01 7.49083459e-01
7.22439528e-01 -1.04209435e+00 4.32339907e-02 -1.80177584e-01
8.38439703e-01 -2.44829118e-01 2.71155775e-01 -8.10680270e-01
2.76736975e-01 6.95331469e-02 -5.41948341e-02 -1.01106852e-01
-1.82763949e-01 -6.74080610e-01 2.93853909e-01 4.71847326e-01
-4.23442602e-01 5.63221760e-02 2.08874032e-01 6.10431731e-01
-9.12583843e-02 -4.95116353e-01 5.18252254e-01 -1.09015815e-01
7.75366463e-03 -3.11250925e-01 -5.82739294e-01 3.13488066e-01
7.71900237e-01 -2.38801733e-01 -5.23517191e-01 -1.16116896e-01
-5.61035454e-01 -3.24559808e-01 7.74391651e-01 1.47572488e-01
-6.29297122e-02 -1.12445772e+00 -8.45290005e-01 -3.27100486e-01
4.06269610e-01 -3.08534682e-01 -5.08529842e-02 1.27204645e+00
-5.14103472e-01 2.03159645e-01 2.27835402e-01 -3.78447771e-02
-1.22054231e+00 7.61967123e-01 5.52107394e-02 -7.16367364e-01
-5.69446981e-01 3.57404917e-01 -2.62803286e-01 -3.82688940e-01
-2.29465783e-01 -3.41783851e-01 -4.99470323e-01 4.12361354e-01
5.01626313e-01 8.73112977e-02 4.78298545e-01 -8.66794229e-01
-5.08974612e-01 1.54553205e-01 -1.18270822e-01 -2.65366703e-01
1.28804100e+00 1.30732097e-02 -9.81131848e-03 2.92778224e-01
9.23871219e-01 9.75017786e-01 -4.49307621e-01 -2.41820872e-01
7.71908581e-01 4.79156710e-02 -2.61023551e-01 -7.58373260e-01
-7.63840497e-01 2.50933141e-01 1.45558551e-01 6.10323846e-01
7.95480311e-01 1.15571089e-01 8.96822035e-01 3.68938446e-01
9.44052264e-02 -1.05551732e+00 -3.94693226e-01 3.84641081e-01
4.97719646e-01 -1.10787785e+00 8.84415954e-02 -6.56223774e-01
-7.49207973e-01 9.72015381e-01 -7.35205561e-02 2.16682643e-01
4.91093218e-01 4.08714086e-01 3.28136444e-01 -5.74808419e-01
-2.83704609e-01 -4.91946459e-01 3.41207296e-01 4.38070804e-01
7.47298479e-01 -2.47729585e-01 -9.00549352e-01 6.44946516e-01
-3.15142661e-01 -2.84031592e-03 6.62782788e-01 9.51472700e-01
3.43352295e-02 -1.33410788e+00 -3.19755226e-01 5.74877977e-01
-1.45612180e+00 -4.23906475e-01 -5.40972412e-01 7.88826823e-01
3.98547053e-01 1.07313275e+00 -3.75776559e-01 -2.15682641e-01
6.79661810e-01 2.68788755e-01 1.23623714e-01 -9.56252158e-01
-1.06335020e+00 -7.50225484e-02 6.46117032e-01 -6.84318468e-02
-9.21878815e-01 -7.08208442e-01 -1.30220175e+00 -1.07383713e-01
-6.12360239e-01 5.45553625e-01 5.74552178e-01 1.10058951e+00
4.70627218e-01 4.32644397e-01 -8.71768370e-02 4.08783183e-02
-2.77187675e-01 -9.20425177e-01 -7.65372738e-02 8.36591184e-01
1.27196983e-01 -4.54350024e-01 -2.61117578e-01 3.93344998e-01] | [8.694701194763184, 8.9251070022583] |
87bb69ff-4d97-455b-a1d3-b732d0d0a484 | multi-frame-super-resolution-reconstruction | 1812.09375 | null | http://arxiv.org/abs/1812.09375v1 | http://arxiv.org/pdf/1812.09375v1.pdf | Multi-Frame Super-Resolution Reconstruction with Applications to Medical Imaging | The optical resolution of a digital camera is one of its most crucial
parameters with broad relevance for consumer electronics, surveillance systems,
remote sensing, or medical imaging. However, resolution is physically limited
by the optics and sensor characteristics. In addition, practical and economic
reasons often stipulate the use of out-dated or low-cost hardware.
Super-resolution is a class of retrospective techniques that aims at
high-resolution imagery by means of software. Multi-frame algorithms approach
this task by fusing multiple low-resolution frames to reconstruct
high-resolution images. This work covers novel super-resolution methods along
with new applications in medical imaging. | ['Thomas Köhler'] | 2018-12-21 | null | null | null | null | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 8.51981521e-01 -4.58186448e-01 3.42501774e-02 -1.37003556e-01
-7.51108050e-01 -2.07590923e-01 2.69574344e-01 -1.17402755e-01
-4.85560745e-01 1.06111550e+00 -1.87792897e-01 1.63694888e-01
-2.96933353e-01 -7.30611682e-01 -1.61451161e-01 -9.35907602e-01
1.57557517e-01 4.32194993e-02 4.75851953e-01 -1.12000823e-01
1.60363868e-01 8.06797683e-01 -1.59598851e+00 2.15855390e-01
6.88151836e-01 9.07167196e-01 6.75974965e-01 6.67682648e-01
1.44643098e-01 3.25552225e-01 -3.38367760e-01 5.67818992e-02
2.11181380e-02 -5.92082262e-01 -3.46958190e-01 3.76770735e-01
1.77752912e-01 -6.17872357e-01 2.64278173e-01 1.13775599e+00
3.78430098e-01 -1.32227272e-01 2.39624336e-01 -2.02618763e-01
-3.16789418e-01 -3.95159759e-02 -8.23348522e-01 5.91936111e-01
6.95333838e-01 -1.17095917e-01 3.06997418e-01 -7.64504492e-01
5.47179461e-01 7.88709641e-01 4.56909865e-01 4.62958157e-01
-1.33884013e+00 -2.68503278e-01 -4.09627944e-01 -4.99597657e-03
-1.36780310e+00 -5.98353982e-01 6.39950633e-01 -3.95571470e-01
4.66879487e-01 4.40177321e-01 2.84565836e-01 6.76904440e-01
6.11564994e-01 -1.24137864e-01 1.50521290e+00 -4.03198719e-01
1.38067394e-01 2.02617779e-01 -1.29833370e-01 2.09959686e-01
7.08283901e-01 2.86490381e-01 -4.20205593e-01 -1.77745193e-01
1.41061008e+00 4.37080562e-02 -6.10153556e-01 -2.64018960e-03
-1.03560805e+00 4.86518741e-01 -6.31958712e-04 8.26110542e-01
-7.26356745e-01 -3.64999920e-01 -3.16698179e-02 -1.07886922e-02
4.89039242e-01 4.52386022e-01 -1.18066818e-01 4.48836498e-02
-8.44479918e-01 -2.73728430e-01 1.04367457e-01 4.88602996e-01
5.01008511e-01 8.15999955e-02 2.67953098e-01 6.93145394e-01
9.58845988e-02 3.78216356e-01 4.09089386e-01 -1.15410089e+00
-2.43907794e-01 1.55748844e-01 3.16902280e-01 -5.76114297e-01
-3.95387083e-01 -3.34386051e-01 -1.07284534e+00 4.51877534e-01
1.55371651e-01 1.87205374e-01 -5.73453486e-01 1.02432215e+00
5.41839123e-01 1.47648871e-01 2.59694248e-01 1.26195550e+00
8.31430912e-01 5.60196221e-01 -1.23443201e-01 -8.69946897e-01
1.60844922e+00 -2.24114195e-01 -1.05199611e+00 -2.30253458e-01
-2.89847732e-01 -9.16291714e-01 5.05623341e-01 6.03041530e-01
-1.28658211e+00 -4.45529103e-01 -9.88230824e-01 -1.03623625e-02
3.06592345e-01 6.29745349e-02 2.85431504e-01 6.26863003e-01
-6.84233546e-01 5.78930020e-01 -8.39544237e-01 -3.04275692e-01
2.30089709e-01 2.52035201e-01 -3.41310710e-01 -2.15297297e-01
-8.52691710e-01 8.65214169e-01 2.63405025e-01 7.00123608e-02
-2.68838406e-01 -6.51848435e-01 -6.68396711e-01 -2.62740254e-01
2.59280264e-01 -6.85866714e-01 9.54061866e-01 -7.36163139e-01
-1.73404920e+00 9.67267752e-01 -2.08923221e-01 -2.77456552e-01
2.96819359e-01 -1.63578540e-01 -7.16277122e-01 8.22742581e-01
-7.97074586e-02 -1.09779857e-01 1.11593342e+00 -9.65107620e-01
-6.70923710e-01 -3.80101621e-01 -1.60255328e-01 2.03455120e-01
1.19722351e-01 2.74702400e-01 -8.54284503e-03 -2.36890987e-01
2.52473086e-01 -5.79573393e-01 -3.95768344e-01 -9.58676189e-02
7.05845207e-02 4.28176463e-01 8.69243443e-01 -6.05405152e-01
9.11302567e-01 -2.14473820e+00 5.61618432e-02 -3.71211231e-01
1.10636681e-01 3.30727160e-01 3.17200243e-01 7.38440547e-03
9.72185358e-02 -3.48391116e-01 -3.03943694e-01 2.58779526e-01
-1.09461403e+00 -7.27773532e-02 4.58120629e-02 8.05070341e-01
7.50544146e-02 2.77295858e-01 -7.84183502e-01 -7.66637564e-01
8.15641284e-01 9.81729627e-01 1.55471936e-02 9.24072787e-02
1.66649178e-01 1.19781709e+00 -6.73257172e-01 6.81974173e-01
7.86405444e-01 -3.92230421e-01 2.09750146e-01 -4.75392103e-01
-6.50845468e-01 -1.45314962e-01 -1.17110515e+00 1.57554746e+00
-3.57643664e-01 2.93212205e-01 4.74446148e-01 -5.10149121e-01
6.92640185e-01 5.14878631e-01 8.64975452e-01 -8.63436222e-01
-2.07121484e-03 3.92714173e-01 -4.36735421e-01 -6.99923575e-01
6.67986691e-01 -6.89559400e-01 4.20222014e-01 6.14305623e-02
-4.84327406e-01 -2.35952646e-01 -6.58159554e-02 -3.80574107e-01
6.53630912e-01 1.79623038e-01 8.21362615e-01 -2.82380641e-01
7.23483324e-01 9.27238241e-02 5.11523485e-01 1.83216840e-01
4.08184603e-02 8.89594316e-01 -1.82546109e-01 -3.63642514e-01
-1.28948402e+00 -7.11320281e-01 -8.52123499e-01 3.55002552e-01
2.26667523e-01 1.95050761e-01 -5.95273674e-01 3.36169749e-01
-4.12730575e-01 2.48179018e-01 -2.78583497e-01 3.25783074e-01
-5.88490427e-01 -8.31301391e-01 -2.20357448e-01 2.80208588e-02
5.20860612e-01 -6.59252942e-01 -1.45654821e+00 3.69708836e-01
-2.25993276e-01 -1.74500680e+00 3.00104031e-03 -1.69245884e-01
-1.15761065e+00 -1.11755717e+00 -1.04980052e+00 -2.80222118e-01
5.60568571e-01 6.09716177e-01 1.02011502e+00 2.16456112e-02
-5.03512740e-01 3.45359594e-01 -3.25031042e-01 -1.10089801e-01
-3.99937183e-01 -2.95779765e-01 5.25657982e-02 2.39297241e-01
-2.37023868e-02 -4.53573316e-01 -7.11784720e-01 3.36496711e-01
-1.04449093e+00 2.23000824e-01 7.77516484e-01 7.16343403e-01
1.10047650e+00 4.32505757e-01 1.74466759e-01 -7.73056805e-01
2.72437513e-01 -4.54323366e-04 -9.37761784e-01 1.37852743e-01
-3.09664249e-01 -6.05726279e-02 4.68037784e-01 -3.13856691e-01
-1.49051046e+00 2.54962921e-01 4.24029455e-02 -3.52916658e-01
-4.04087007e-01 2.86676824e-01 -1.13749355e-02 -2.45131284e-01
8.03065181e-01 1.96127564e-01 1.16314590e-01 -4.69825447e-01
-2.05585063e-01 6.17534518e-01 8.26084077e-01 -1.15902983e-01
5.64021826e-01 9.82153773e-01 4.64996010e-01 -1.57484293e+00
-1.04560089e+00 -4.43247080e-01 -6.95720792e-01 -2.57247269e-01
1.13556218e+00 -9.44579363e-01 -5.00226796e-01 2.90135950e-01
-8.61148655e-01 1.17442295e-01 -1.22109570e-01 9.84816432e-01
-4.42904592e-01 5.23185432e-01 -5.88253438e-01 -9.57351089e-01
-2.27281421e-01 -9.87429261e-01 1.05052400e+00 4.34439123e-01
-3.58647183e-02 -7.34906733e-01 -1.03899851e-01 4.55936790e-01
4.31533635e-01 6.47927642e-01 2.31198788e-01 5.42804480e-01
-8.97448599e-01 2.28620153e-02 -2.18932986e-01 1.64091617e-01
5.11384666e-01 -1.90380722e-01 -1.09620595e+00 -2.71188378e-01
6.28066242e-01 3.57277654e-02 4.85031217e-01 1.08164430e+00
8.45755458e-01 1.72419354e-01 -2.88489014e-01 7.45122731e-01
1.91522861e+00 3.11914861e-01 7.22607493e-01 2.82811046e-01
3.57873321e-01 5.47672689e-01 9.60132837e-01 4.10469711e-01
-3.10022414e-01 8.56462121e-01 1.85872883e-01 -5.77945337e-02
-7.70425871e-02 3.84276181e-01 -1.02089226e-01 2.70322442e-01
-5.67615747e-01 2.63631016e-01 -4.62438971e-01 2.43446857e-01
-1.40477514e+00 -1.38519204e+00 -6.15243554e-01 2.52924204e+00
6.36772990e-01 -1.67946100e-01 -2.14526296e-01 2.40226865e-01
9.51155066e-01 -6.59922957e-02 -5.38998365e-01 -1.34973705e-01
-3.10125947e-01 1.05816819e-01 6.69576168e-01 5.19277751e-01
-9.83121753e-01 2.82172471e-01 6.93843603e+00 5.29759765e-01
-1.17731047e+00 1.91384241e-01 3.86340946e-01 -1.13097787e-01
-1.16657004e-01 -3.21292281e-01 -7.86733985e-01 3.97073239e-01
9.30948496e-01 -2.10759044e-01 3.28631729e-01 4.22395825e-01
5.40805340e-01 -8.85544300e-01 -7.56980419e-01 1.28469205e+00
-1.50968526e-02 -1.17201245e+00 -1.84661821e-01 2.44540811e-01
6.71598792e-01 -2.82854795e-01 1.71759307e-01 -7.91281343e-01
-4.20365959e-01 -8.93691421e-01 2.61833191e-01 4.96278286e-01
1.30830061e+00 -7.02959776e-01 5.37848532e-01 2.79377162e-01
-1.13175344e+00 1.35237604e-01 -5.52894771e-01 -1.94186345e-03
6.88159168e-01 9.98445690e-01 -3.44972730e-01 6.72514200e-01
7.70965755e-01 4.78860259e-01 -4.71399445e-03 8.97604465e-01
-5.72410598e-02 -8.06466863e-03 -2.72599697e-01 5.21840274e-01
-1.68902382e-01 -6.12893045e-01 5.83723426e-01 7.82195568e-01
5.33462167e-01 8.86064947e-01 -1.27478793e-01 6.94876909e-01
4.35223281e-01 -2.98572987e-01 -5.28570414e-01 4.46652025e-02
2.34315515e-01 1.43552709e+00 -7.61146843e-01 -1.94408625e-01
-6.63499594e-01 8.83071065e-01 -4.82247949e-01 1.27493083e-01
-5.16376495e-01 5.29476330e-02 6.11218810e-01 6.41709805e-01
2.13006735e-01 -3.35021347e-01 -2.51028359e-01 -1.06199861e+00
-1.23801813e-01 -5.59172451e-01 3.77437055e-01 -8.90189588e-01
-7.45781541e-01 6.45429969e-01 9.67218578e-02 -1.58144367e+00
-2.49654785e-01 -3.88805628e-01 1.35608792e-01 1.11189544e+00
-1.88491464e+00 -9.37542796e-01 -5.74475050e-01 6.41689062e-01
5.95537305e-01 1.97795659e-01 9.10674572e-01 3.09436440e-01
-2.48736501e-01 -3.66144836e-01 2.94671267e-01 -3.61820847e-01
5.02305388e-01 -7.29503870e-01 -2.82754838e-01 9.65445220e-01
-4.46066886e-01 1.68340161e-01 1.02265239e+00 -6.24317646e-01
-1.33158672e+00 -6.38191104e-01 5.98599315e-01 -1.34029388e-01
2.24932164e-01 2.62454987e-01 -1.11089230e+00 2.97466159e-01
-3.94439660e-02 2.75404513e-01 5.88296831e-01 -4.16415274e-01
3.18856835e-01 -1.04032412e-01 -1.56152081e+00 6.88576400e-02
4.62745398e-01 -5.05015194e-01 -3.66203398e-01 9.67897996e-02
3.88837934e-01 -7.33707547e-01 -1.05700672e+00 3.03802639e-01
7.40512908e-01 -1.36897326e+00 1.08685017e+00 3.22626561e-01
3.24149489e-01 -4.97349769e-01 -5.53987399e-02 -8.77608776e-01
-4.81791526e-01 -3.57499301e-01 1.95760965e-01 7.18510509e-01
-1.39287144e-01 -6.23422563e-01 5.51383317e-01 5.68791270e-01
8.10737759e-02 -1.94171578e-01 -9.18176889e-01 -7.50589013e-01
-5.52628696e-01 5.12834452e-02 3.21820170e-01 8.13295603e-01
-3.43265355e-01 3.61404091e-01 -5.21240950e-01 4.70101267e-01
1.28260612e+00 3.06387037e-01 1.56711042e-01 -1.47455621e+00
-2.02820882e-01 -2.74810344e-02 -3.84848475e-01 -4.88004506e-01
-4.02852178e-01 -3.10687022e-03 -2.51999468e-01 -1.43241775e+00
3.07308793e-01 -1.14252359e-01 6.65585697e-02 -3.94589901e-01
1.23366572e-01 5.98262966e-01 -2.44442672e-01 5.83899856e-01
-1.27364129e-01 -1.04488969e-01 1.31719923e+00 2.89986551e-01
-3.27469409e-01 1.33594453e-01 -4.09144431e-01 6.83028042e-01
7.64709473e-01 -2.30152637e-01 -2.92267680e-01 -3.77615333e-01
1.97966784e-01 5.57356000e-01 3.99491370e-01 -9.55477238e-01
2.27130562e-01 -4.78040904e-01 4.56648111e-01 -5.09555578e-01
6.40074134e-01 -1.20640087e+00 9.23368096e-01 5.83017528e-01
3.46367359e-02 -2.70570576e-01 -1.69096589e-01 4.35504347e-01
-3.20469856e-01 -2.78938740e-01 1.55363917e+00 -5.41847587e-01
-8.28447580e-01 9.12022144e-02 -3.88258040e-01 -4.82818633e-01
1.21215343e+00 -7.20784724e-01 -1.89604491e-01 -6.04755282e-02
-6.41083419e-01 -3.16535443e-01 7.13523507e-01 -1.29600078e-01
8.49019587e-01 -1.02789044e+00 -6.94397211e-01 2.77471274e-01
-6.57125935e-03 2.69878268e-01 7.42066324e-01 1.05221093e+00
-7.37083316e-01 4.10543382e-01 -4.68160629e-01 -8.13933015e-01
-1.59744620e+00 5.84635496e-01 5.31856775e-01 -4.65530939e-02
-1.08200657e+00 2.95291007e-01 1.42272919e-01 5.80967784e-01
-4.63769138e-01 -7.17662182e-03 -4.37028885e-01 -1.81529969e-01
1.18975210e+00 5.32471180e-01 4.31383029e-02 -8.11443865e-01
-3.75001609e-01 1.11557460e+00 2.32795298e-01 -1.69209346e-01
1.27760983e+00 -5.51272333e-01 -6.13943450e-02 5.51822424e-01
7.26167977e-01 -7.55539536e-02 -1.21130073e+00 -2.50753224e-01
-1.37057453e-01 -6.47106767e-01 4.81431454e-01 -3.97177786e-01
-8.58901381e-01 5.73323131e-01 5.76198816e-01 2.04399928e-01
1.62931275e+00 6.74923509e-02 4.86274600e-01 -3.46609280e-02
7.29006171e-01 -1.15657115e+00 -2.09964037e-01 -1.45505503e-01
6.68941557e-01 -1.22224700e+00 7.15921283e-01 -6.98326707e-01
-3.36059272e-01 1.19174075e+00 5.96973822e-02 2.02092320e-01
6.03909075e-01 5.70803106e-01 1.56797320e-01 -1.51161700e-01
-4.96566474e-01 -2.46338874e-01 1.17003962e-01 6.79549932e-01
7.09375381e-01 -1.08302608e-01 -6.11104965e-01 -8.09594765e-02
2.47126088e-01 4.40583825e-01 8.10199022e-01 8.57908249e-01
-6.22191429e-01 -8.30724597e-01 -9.83020186e-01 3.68807554e-01
-9.07258630e-01 2.90390849e-01 3.13484430e-01 4.30121958e-01
-4.47448269e-02 1.00970864e+00 5.91702722e-02 1.50425166e-01
2.07208961e-01 -4.64135110e-01 7.27146029e-01 -4.11692917e-01
-1.80509314e-01 5.34684300e-01 -2.48279646e-02 -6.36302471e-01
-1.09622717e+00 -8.29718828e-01 -1.09111643e+00 -2.30274826e-01
-2.38172412e-01 4.17097658e-02 8.56497407e-01 6.78147078e-01
1.20117694e-01 4.27633882e-01 5.89954197e-01 -7.97377944e-01
-1.16561361e-01 -6.66532576e-01 -1.16694474e+00 8.26439857e-02
6.86945319e-01 -3.69728386e-01 -2.88949072e-01 2.68499792e-01] | [11.110040664672852, -2.3406996726989746] |
e4de7791-6f25-4a3a-8be0-5710811c92f8 | large-age-gap-face-verification-by-feature | 1602.06149 | null | http://arxiv.org/abs/1602.06149v1 | http://arxiv.org/pdf/1602.06149v1.pdf | Large age-gap face verification by feature injection in deep networks | This paper introduces a new method for face verification across large age
gaps and also a dataset containing variations of age in the wild, the Large
Age-Gap (LAG) dataset, with images ranging from child/young to adult/old. The
proposed method exploits a deep convolutional neural network (DCNN) pre-trained
for the face recognition task on a large dataset and then fine-tuned for the
large age-gap face verification task. Finetuning is performed in a Siamese
architecture using a contrastive loss function. A feature injection layer is
introduced to boost verification accuracy, showing the ability of the DCNN to
learn a similarity metric leveraging external features. Experimental results on
the LAG dataset show that our method is able to outperform the face
verification solutions in the state of the art considered. | ['Simone Bianco'] | 2016-02-19 | null | null | null | null | ['age-invariant-face-recognition'] | ['computer-vision'] | [ 8.92817825e-02 -1.44495564e-02 4.89789620e-02 -6.00865960e-01
-5.28368831e-01 -2.52238810e-01 7.98926592e-01 -1.46862283e-01
-5.38355291e-01 6.27046168e-01 -1.59781575e-01 6.00617416e-02
-1.90300629e-01 -6.37116432e-01 -6.71609700e-01 -7.27998435e-01
-3.75370800e-01 5.06055057e-01 -3.04167092e-01 -3.33550088e-02
1.40548959e-01 7.59250045e-01 -1.88230443e+00 -1.09830067e-01
8.81409287e-01 1.57564163e+00 -4.80468094e-01 4.20547962e-01
4.13199127e-01 1.57088548e-01 -7.44166493e-01 -9.71603572e-01
9.67364669e-01 -2.30017990e-01 -5.43820977e-01 -4.44588289e-02
1.56329083e+00 -2.99872339e-01 -3.54027748e-01 9.65944469e-01
8.30279171e-01 5.24705127e-02 4.78640854e-01 -1.70267177e+00
-7.91495204e-01 4.68926102e-01 -4.61650461e-01 1.19911477e-01
3.59059483e-01 1.60752058e-01 7.17385828e-01 -9.75235164e-01
5.41089892e-01 1.47666228e+00 1.02923930e+00 1.06144559e+00
-9.33741927e-01 -1.22189355e+00 -1.68356925e-01 2.06245095e-01
-1.55312312e+00 -5.42673290e-01 6.77304566e-01 -3.38307559e-01
6.23857796e-01 -1.95618689e-01 4.84395802e-01 1.43185806e+00
9.90066752e-02 2.25399882e-01 1.20831978e+00 -2.29574829e-01
1.46339506e-01 -1.04163811e-01 -5.60705550e-02 8.78954113e-01
3.52611631e-01 6.84254050e-01 -7.98504472e-01 -2.44339760e-02
3.95394146e-01 -6.52391985e-02 -1.58928990e-01 -6.29972994e-01
-8.97683144e-01 7.29676843e-01 5.74845970e-01 2.00068265e-01
-9.23222527e-02 8.98272544e-03 5.59277654e-01 6.72403216e-01
4.59899813e-01 4.29115772e-01 -7.73329854e-01 1.13572799e-01
-1.34393358e+00 4.18056309e-01 7.19965398e-01 5.57823539e-01
6.31771266e-01 5.47673926e-02 -4.83762562e-01 6.22943580e-01
2.05573037e-01 5.43860078e-01 5.62546372e-01 -8.85351777e-01
5.20416558e-01 5.86099744e-01 -2.74724960e-01 -5.84336877e-01
-2.51066864e-01 -3.64233077e-01 -9.25376892e-01 7.55537808e-01
7.93357551e-01 -8.93288255e-02 -1.20201051e+00 2.15682745e+00
5.02394736e-01 2.90655911e-01 1.04810847e-02 4.47048903e-01
7.98013628e-01 -1.38138533e-02 7.59916455e-02 -6.34395927e-02
1.36137450e+00 -8.90326977e-01 -3.91580909e-01 3.14366743e-02
1.08003154e-01 -5.05585968e-01 7.13087499e-01 3.26377124e-01
-1.13398433e+00 -8.70852828e-01 -1.22279084e+00 1.42030194e-01
-9.08582091e-01 3.23956013e-01 3.67865741e-01 1.19090486e+00
-1.46326602e+00 9.79375482e-01 -5.27308643e-01 -5.25401413e-01
9.75487888e-01 9.43993926e-01 -9.40141559e-01 -1.76560372e-01
-1.13169098e+00 8.76347423e-01 2.75032103e-01 2.15495378e-01
-1.06210291e+00 -1.24102139e+00 -1.23649096e+00 -7.09019508e-03
-1.05584241e-01 -7.82688141e-01 7.34957159e-01 -1.14666963e+00
-1.48164773e+00 1.25235534e+00 2.04126865e-01 -7.01663375e-01
8.96525562e-01 -1.74744770e-01 -5.26937366e-01 1.97553471e-01
-3.64194922e-02 9.13033307e-01 1.51784074e+00 -8.24071884e-01
-2.75670856e-01 -7.90470243e-01 -1.63564414e-01 -4.58453953e-01
-4.81619924e-01 3.50061744e-01 -8.12971890e-02 -7.86555707e-01
-4.70614284e-01 -8.93406689e-01 2.59516567e-01 5.64007103e-01
-7.67661482e-02 -3.58221292e-01 1.09380496e+00 -8.83679867e-01
1.05011618e+00 -1.97813010e+00 2.85337836e-01 3.32903564e-01
1.07521236e-01 4.80496168e-01 -5.66896081e-01 1.65177420e-01
-5.16738713e-01 -2.21239522e-01 -4.55712259e-01 -8.47032905e-01
4.94464561e-02 -1.01680011e-02 2.62215793e-01 6.44199252e-01
4.94918048e-01 8.15827370e-01 -6.75486028e-01 -4.47013378e-01
-1.57685846e-01 8.07790577e-01 -4.92088258e-01 3.61388624e-01
1.75110012e-01 3.49681914e-01 1.59272999e-01 9.45661843e-01
1.26361489e+00 1.74204022e-01 -2.64093339e-01 -1.31420180e-01
1.38974681e-01 -3.36606503e-01 -9.48994935e-01 1.76257777e+00
-5.70869684e-01 2.90566206e-01 3.20671320e-01 -8.87808204e-01
9.96864915e-01 2.55383551e-01 3.77605885e-01 -5.12600958e-01
3.79916012e-01 3.34686726e-01 1.80114955e-01 -9.76701155e-02
2.54336298e-01 3.03650703e-02 1.41186953e-01 3.21392089e-01
6.21414065e-01 1.53035726e-02 3.90486658e-01 -4.93018553e-02
1.00960958e+00 -2.06780173e-02 -9.31789130e-02 -3.74843299e-01
1.18522894e+00 -8.90937030e-01 6.17558181e-01 3.55324656e-01
-6.90701842e-01 7.74439454e-01 3.72678548e-01 -6.01382494e-01
-1.07104409e+00 -1.17670059e+00 -3.86780024e-01 9.28440690e-01
-4.05268192e-01 -3.23559254e-01 -1.04531586e+00 -1.41180480e+00
5.00649989e-01 -1.21739253e-01 -1.46308589e+00 -3.57465446e-01
-6.77366078e-01 -4.92429167e-01 6.97358251e-01 6.26135349e-01
9.51423466e-01 -1.11942720e+00 -2.83692986e-01 -3.88397038e-01
5.47406971e-01 -1.19452703e+00 -6.63428187e-01 -1.61181629e-01
-5.93215108e-01 -1.33176148e+00 -1.01427197e+00 -8.35817337e-01
8.74816120e-01 -6.20071054e-01 1.14293909e+00 2.65442282e-01
-5.55465162e-01 4.99843538e-01 -7.26016164e-02 -2.87988842e-01
-4.74332958e-01 1.91802770e-01 3.93979698e-01 4.39746678e-01
2.83732891e-01 -5.98020673e-01 -6.88644528e-01 2.95602500e-01
-6.71184659e-01 -7.36256480e-01 3.96073043e-01 1.25009751e+00
-7.44088665e-02 -3.12807411e-01 6.19677901e-01 -4.70582813e-01
2.99645901e-01 -1.26127928e-01 -9.65184212e-01 3.37731302e-01
-8.57356250e-01 2.28592008e-01 4.77316767e-01 -6.03663743e-01
-7.47419596e-01 1.48692518e-01 -1.91414595e-01 -7.38779545e-01
-4.48308885e-02 -2.88217217e-01 -4.21731085e-01 -6.80571496e-01
4.52642530e-01 -2.04111245e-02 3.56317461e-01 -4.35687214e-01
1.14409730e-01 2.38979816e-01 7.07782209e-01 -6.37266994e-01
1.37060070e+00 2.39836723e-01 4.96357739e-01 -2.72694737e-01
-5.80528021e-01 8.35229382e-02 -8.77965748e-01 -1.26018673e-01
7.29895711e-01 -8.84794831e-01 -8.59263778e-01 1.08221126e+00
-7.21337020e-01 -2.35069185e-01 -4.39661980e-01 2.70208687e-01
-3.32979977e-01 2.91176319e-01 -5.60931921e-01 -3.97717118e-01
-8.23919892e-01 -1.00252092e+00 1.31006122e+00 4.67241764e-01
1.41993687e-01 -1.00640750e+00 1.00435957e-01 3.38318348e-01
6.77427113e-01 4.15612280e-01 4.25631702e-01 -7.84844995e-01
-2.11934209e-01 -2.68681705e-01 -3.87384593e-01 7.94054389e-01
1.48362905e-01 2.44446084e-01 -1.26366997e+00 -7.00348020e-01
-5.34318030e-01 -6.72016561e-01 9.93048668e-01 1.01509713e-01
1.35259402e+00 -1.62642360e-01 -1.79814901e-02 8.84029031e-01
1.16330659e+00 -3.30297172e-01 6.52127028e-01 1.53633235e-02
5.01526535e-01 5.81021726e-01 4.21487033e-01 3.97545248e-01
2.74072111e-01 7.67320096e-01 5.27892947e-01 5.61439916e-02
-4.58249569e-01 -1.20060310e-01 3.80722374e-01 9.47856903e-02
-1.27707392e-01 4.20169741e-01 -8.53730321e-01 5.75909734e-01
-1.31952047e+00 -9.83977318e-01 6.55586362e-01 2.31560779e+00
7.98587441e-01 -1.17609665e-01 2.60190964e-01 4.40441787e-01
7.24497557e-01 3.35614383e-02 -4.03161615e-01 -6.69551253e-01
-1.79091752e-01 1.00525248e+00 2.67245650e-01 2.96681911e-01
-1.34284902e+00 7.92007267e-01 6.41393423e+00 7.26462007e-01
-1.27500510e+00 1.75263047e-01 5.40424049e-01 5.42249084e-02
4.83428389e-01 -5.67637622e-01 -9.66458678e-01 5.76030433e-01
8.84696960e-01 -8.39205310e-02 4.22525406e-01 7.09743857e-01
-3.45844984e-01 2.63273776e-01 -1.43038464e+00 1.11834013e+00
6.11826241e-01 -8.79039168e-01 -5.51042147e-02 6.39592558e-02
9.23377395e-01 -2.85349995e-01 7.12036967e-01 3.98059547e-01
8.92345458e-02 -1.21824479e+00 5.50040007e-01 3.14413488e-01
1.04676950e+00 -7.66342998e-01 7.21330106e-01 -3.02933067e-01
-1.33134007e+00 -4.78242397e-01 -1.96938500e-01 2.06604809e-01
-3.12651545e-01 3.15048605e-01 -6.81177497e-01 5.21515250e-01
8.75636101e-01 7.46749043e-01 -1.18249273e+00 9.92761433e-01
-2.94241067e-02 6.37941882e-02 -1.56746060e-01 6.17952704e-01
-6.19998574e-02 4.07428890e-02 1.76677421e-01 9.40067887e-01
6.35879040e-01 -6.40470624e-01 -2.17934996e-01 6.11960590e-01
-6.11168861e-01 -1.98088244e-01 -5.69542825e-01 1.76876456e-01
3.07915002e-01 1.49028873e+00 -1.45734981e-01 -2.34462693e-01
-9.09269974e-02 1.08111465e+00 3.68414819e-01 -5.60323820e-02
-7.20266283e-01 -4.37099993e-01 1.03730047e+00 -9.45748910e-02
7.80360103e-01 1.76087454e-01 -7.62694050e-03 -1.06753433e+00
1.26800388e-01 -1.00873411e+00 6.83635354e-01 -4.76696014e-01
-1.58158195e+00 8.49043489e-01 -1.24370314e-01 -1.02126193e+00
-4.76592064e-01 -9.03406441e-01 -8.29863369e-01 9.26695943e-01
-1.63342154e+00 -1.62237549e+00 -4.58302408e-01 9.07011986e-01
1.43368840e-01 -6.42544866e-01 7.05791891e-01 6.11979604e-01
-6.84899926e-01 1.54982984e+00 -3.03656399e-01 3.98791134e-01
9.73852515e-01 -1.23754466e+00 4.18569475e-01 8.60097766e-01
-2.36500293e-01 6.58600211e-01 4.91176456e-01 -3.01397502e-01
-1.17904127e+00 -1.24462354e+00 1.07727778e+00 -5.18021226e-01
4.55381572e-01 -5.85728943e-01 -7.40113378e-01 4.36053097e-01
4.10432667e-01 7.89816201e-01 4.56387609e-01 -4.31885757e-02
-1.20854819e+00 -6.53854311e-01 -2.05621624e+00 1.17513992e-01
1.27401507e+00 -6.62853301e-01 -4.09274280e-01 1.06257666e-02
3.46322685e-01 -2.31931403e-01 -1.28043616e+00 7.76694775e-01
9.12495792e-01 -9.36814487e-01 1.22487545e+00 -5.72825313e-01
4.39628899e-01 6.18108623e-02 1.03720747e-01 -1.29106462e+00
1.29352257e-01 -6.79351270e-01 -4.89146322e-01 1.49941492e+00
2.07711488e-01 -6.63505197e-01 9.04503703e-01 4.41832244e-01
2.16134235e-01 -7.75907636e-01 -1.35798728e+00 -1.11459291e+00
4.97730315e-01 1.56781957e-01 1.09866083e+00 7.72304237e-01
-5.83464742e-01 -1.89134836e-01 -1.80083185e-01 2.34792560e-01
1.02593553e+00 -7.60215968e-02 6.41626537e-01 -1.55843329e+00
1.12817422e-01 -5.21495163e-01 -9.29654956e-01 -6.63634539e-02
7.51817763e-01 -8.07079196e-01 -3.02107960e-01 -6.59326613e-01
2.59736069e-02 -3.03070366e-01 -4.80160296e-01 4.86936271e-01
-1.50057316e-01 7.82408834e-01 2.61221945e-01 -6.53397799e-01
-2.51856148e-01 6.41930878e-01 1.03733146e+00 -4.92200077e-01
2.93442577e-01 1.11465670e-01 -3.23000103e-01 3.33709031e-01
7.10378945e-01 -2.35605672e-01 -1.30188186e-02 -5.16294921e-03
-2.14646012e-01 -4.22636598e-01 4.59235668e-01 -1.50179946e+00
1.33059338e-01 3.96260202e-01 7.82108903e-01 -3.47481251e-01
4.97253478e-01 -8.54373753e-01 -9.44597200e-02 7.11600900e-01
-2.17773676e-01 3.74306262e-01 2.25541785e-01 2.30414812e-02
-2.56686300e-01 8.78990069e-02 1.21239829e+00 3.34218979e-01
-3.25112969e-01 1.04755902e+00 3.95664632e-01 2.26367861e-01
1.04171181e+00 -2.82323599e-01 -2.06853002e-01 8.69072229e-02
-6.75116718e-01 2.20847279e-01 6.45045280e-01 7.26951241e-01
5.01384854e-01 -1.65411222e+00 -9.27053571e-01 7.62894332e-01
4.44368124e-01 -6.65514290e-01 1.64148480e-01 6.63245976e-01
-3.12238783e-01 9.40403566e-02 -8.41916740e-01 -4.29065168e-01
-1.75070572e+00 7.19678402e-01 6.94977224e-01 -4.78091419e-01
-2.06885189e-01 1.16020775e+00 -1.08502910e-01 -7.25440681e-01
6.16401851e-01 -1.65492371e-01 -1.88462228e-01 2.16905639e-01
8.26851070e-01 2.98557878e-01 2.65453637e-01 -8.66881192e-01
-4.56926525e-01 9.85394835e-01 -3.49404220e-03 1.41503364e-01
1.22658384e+00 2.36793786e-01 -8.92604440e-02 -3.09806108e-01
1.55964828e+00 -1.51617423e-01 -1.34825301e+00 -1.30787611e-01
-1.14670798e-01 -6.69370055e-01 -2.75352597e-01 -9.06689823e-01
-1.84356666e+00 8.85311902e-01 1.18794918e+00 -2.31186152e-01
1.30649328e+00 -1.82462916e-01 6.92782521e-01 2.12338284e-01
2.71363735e-01 -1.05187631e+00 1.86673805e-01 3.68143678e-01
1.13545346e+00 -1.49305856e+00 -2.09458731e-02 -4.95591126e-02
-1.32079646e-01 1.08563256e+00 8.37731183e-01 -1.74271137e-01
1.03214693e+00 1.45538181e-01 9.11155120e-02 5.72016351e-02
-4.78932709e-01 -1.58670664e-01 3.68745893e-01 9.53162014e-01
1.01951681e-01 -3.09971720e-01 -2.85865486e-01 2.75575340e-01
-3.72444808e-01 3.27302851e-02 -1.41055332e-02 5.59148431e-01
3.72928768e-01 -1.61258578e+00 -4.18404281e-01 3.23544085e-01
-6.65376186e-01 1.94375917e-01 -4.92291152e-01 7.75972962e-01
7.73732424e-01 5.62212169e-01 1.93814546e-01 -2.44904563e-01
1.55652553e-01 2.63687551e-01 9.72075820e-01 -1.76127106e-01
-1.33380973e+00 -9.52216268e-01 -2.26094440e-01 -9.45900917e-01
-3.83548439e-01 -8.37074757e-01 -5.39777637e-01 -3.80288571e-01
-4.97613922e-02 -5.71393990e-04 7.84035385e-01 5.53574204e-01
2.45967150e-01 2.76096970e-01 9.28415775e-01 -9.38431621e-01
-7.56549358e-01 -1.00608635e+00 -6.51612639e-01 6.55232251e-01
7.06206203e-01 -9.72094178e-01 -4.93951142e-01 -1.16483942e-01] | [13.32540225982666, 0.6297290921211243] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.